Category: 3. Business

Opening panel remarks by Christine Lagarde, President of the ECB, at the annual Economic Policy Symposium “The policy implications of labour market transition” organised by the Federal Reserve Bank of Kansas City in Jackson Hole

Jackson Hole, 23 August 2025

Alexis de Tocqueville – one of the keenest observers of early American democracy – once wrote: “History is a gallery of pictures in which there are few originals and many copies.”

In monetary policy, too, we often look to past cycles for guidance, expecting familiar patterns to repeat themselves. But this cycle has proven to be original in striking ways.

Major central banks have undertaken the most aggressive tightening in a generation. At the outset, there were understandable concerns about how such a rapid and substantial adjustment would affect labour markets.

Historically, disinflation has come at a cost. Since the 1960s, the “sacrifice ratio” has typically been around 1.^[1] In practice, this means that permanently lowering inflation by 1 percentage point has cost about 1% of GDP in forgone output.

And given Europe’s well-known structural rigidities, it was reasonable to assume that a sharp tightening could lead to rising unemployment^[2], which might then become entrenched through hysteresis effects (Slide 2).^[3]

Even in the United States – with its more flexible labour market – many feared that a significant rise in unemployment would be required to bring inflation under control.

Instead, we find ourselves in a very different position from what many expected: in both the euro area and the United States, inflation has fallen sharply, and at a remarkably low cost in terms of employment.^[4]

In fact, in the euro area we have seen the opposite of hysteresis: employment growth has been significantly stronger than historical patterns would have predicted.

Traditionally, Okun’s law suggests that employment tends to grow at roughly half the pace of real GDP. Yet between the end of 2021 and mid-2025, cumulative employment rose by 4.1% – an increase of 6.3 million of people in employment – while real GDP increased by 4.3%. That implies an employment elasticity nearly twice as high as Okun’s relationship would suggest.

For monetary policymakers, the key question is why this atypical employment response has occurred – and whether it signals a broader shift in how inflation will respond to different types of shocks.

Part of the answer lies in global factors. Monetary tightening helped bring inflation back to target, but it coincided with other forces that supported activity: an easing of supply constraints worldwide, a steep drop in energy prices and proactive fiscal policies – all of which help explain the unusually low sacrifice ratio.^[5]

At the same time, Europe’s experience reflects distinctive domestic drivers. Three features have shaped its labour market performance.

First, a delayed wage response to inflation that supported higher employment; second, a reduction in hours worked, driven by labour hoarding and changing preferences; and third, an expansion in labour supply that kept pace with rising demand.

The response of real wages to inflation

Although the euro area has faced a complex mix of shocks in recent years, the dominant force was a major negative supply shock, as post-pandemic bottlenecks coincided with the cut-off of Russian gas.^[6]

Historically, supply shocks of this scale would have quickly passed through to nominal wages, with real wage growth often outpacing productivity.^[7]

For example, after the oil shocks of the 1970s, real wages^[8] in Europe rose by around 20% between 1972 and 1976, while productivity increased by only 15% (Slide 3, left panel). A similar pattern occurred during major demand shocks such as the global financial crisis (Slide 3, middle panel).

This time, however, the response was different – which was the first distinctive feature of this episode. Real wages fell by nearly 2% between late 2021 and early 2023, and only gradually caught up with cumulative productivity growth early last year (Slide 3, right panel).

This unusual pattern reflects a European labour market that has become more flexible in some respects, while remaining rigid in others.

Most notably, formal automatic indexation of wages to inflation has all but disappeared: in the 1970s it covered around half of all private sector employees, whereas today it applies to only about 3%. For more than half of private sector workers, inflation now plays no formal or automatic role in wage setting.^[9]

At the same time, nearly 60% of workers remain covered by multi-year collective agreements which take inflation into account but adjust only gradually – a nominal rigidity that created a lag in wage adjustment relative to prices.^[10] Research also suggests that as the workforce ages, union priorities are shifting, with greater emphasis on pensions and employment protection relative to wage growth.^[11]

ECB analysis confirms that this delayed real wage response acted as a shock absorber. By widening the gap between productivity and labour costs, it eased unit labour cost pressures and supported firms’ profitability, while also making labour relatively more attractive than capital. Both dynamics encouraged firms to expand hiring.

For example, the “factor substitution” effect is estimated to have accounted for around a quarter of total employment growth since the end of 2019, with most of this impact occurring after the onset of the energy crisis.^[12]

This effect was particularly important in manufacturing, which was hit harder than services by negative shocks. That divergence helps explain some of the cross-country heterogeneity in employment growth in the euro area.^[13] Yet, at the aggregate level, manufacturing employment still remained well above what Okun’s law would predict (Slide 4, left panel).

A key factor was firms’ ability to pass on higher input costs, which boosted profit margins and led to a steeper fall in sectoral real wages (Slide 4, right panel). Real wages in industry, measured using sectoral value-added deflators, fell by almost 11% at the trough.^[14]

For comparison, in euro area countries where automatic wage indexation remains in place, the decline in real wages was more limited, and the link between output and employment was notably weaker than for the euro area aggregate.^[15]

The reaction of hours worked

However, the increase in employment in the euro area has been accompanied by a decline in average hours worked – the second distinctive feature of the labour market.

Average hours remain 1% below their pre-pandemic level, equivalent to about four hours fewer per worker per quarter, or a reduction in labour input of roughly 1.3 million full-time jobs (Slide 5).

Two factors help explain how employment could rise even as hours fell.^[16]

The first is labour hoarding, which curbed job losses in firms facing weaker demand – particularly those hit by the energy crisis – but at the cost of fewer hours worked.

The ECB’s labour hoarding indicator rose to almost 30% in the third quarter of 2022 – nearly double its pre-pandemic average – and climbed even higher in manufacturing (Slide 6, left panel).

This behaviour reflected broader labour market tightness: survey evidence suggests employers viewed hoarding as less costly and less risky than rehiring later in an even more competitive market. Fears of future labour shortages – probably reinforced by Europe’s demographic outlook – also played a role.^[17]

The fall in sectoral real wages, together with unusually high profit margins, in turn made it easier for firms to sustain this strategy.^[18]

The second factor is a shift in worker preferences towards shorter hours, which constrained firms’ ability to raise hours per employee and left them more reliant on hiring.^[19]

Average hours worked in Europe have been in long-term decline, driven roughly two-thirds by an increase in part-time employment, much of which is voluntary.^[20] Since late 2021, however, the decline has stemmed mostly from a fall in the number of long hours worked^[21] and from reduced overtime among full-time workers, especially in industry.

While part of this shift is cyclical, it also has a structural component. Over the past decade, preferences for long working hours have declined in parallel with the recorded drop in overtime (Slide 6, right panel).^[22]

Increasing labour supply

Still, for these two features – lower real wages and fewer hours worked – to be compatible with higher employment, labour supply had to respond.

This is where the third feature comes in: the surge in the labour force in recent years.

On demographics alone, Europe’s capacity to expand its labour supply is already constrained. By 2040, the working-age population^[23] is projected to shrink by around 3.4 million. Since 2002, the number of people over 60 has risen by 28 million, while that of those aged 15–60 has fallen by 2.4 million, and of those under 14 by 2.8 million.

Yet after a brief dip during the lockdowns, the labour force was back to its pre-pandemic level by the end of 2021 – and has since grown by about six million people.

This reflects continued increases in participation and employment, particularly among women and older individuals, extending trends already in motion before the pandemic. ECB analysis suggests that without the compositional shift towards older workers – who often enter the labour market directly into employment – the unemployment rate today would be around 6.6% rather than 6.3%.^[24]

Even more important, however, has been the rise in both the number and participation rate of foreign workers.

Although they represented only around 9% of the total labour force in 2022, foreign workers have accounted for half of its growth over the past three years.^[25] Without this contribution, labour market conditions could be tighter^[26] and output lower.

In Germany, for example, GDP would be around 6% lower than in 2019 without the contribution of foreign workers (assuming no behavioural changes among domestic workers). Spain’s strong post-pandemic GDP performance – which has helped support the euro area aggregate – also owes much to the contribution of foreign labour.^[27]

Implications going forward

Looking ahead, it is difficult to say with confidence whether the patterns of recent years will persist, given the complex interplay of cyclical and structural forces. Drawing conclusions about future sacrifice ratios from current developments could therefore be misleading.

But it is worth taking a broader look at some underlying trends.

First, the demographic trend is likely to continue. And this is not just a European story: new research suggests that 2023 was likely to have been the first year in human history when the global fertility rate fell below the replacement rate.^[28]

Migration could, in principle, play a crucial role in easing labour supply constraints in selected regions. But in all plausible scenarios – even those assuming high migration – the euro area’s working-age population will continue to shrink (see illustration for the 20-64 age group in Slide 7, left panel).

Moreover, political economy pressures may increasingly limit inflows, and even when migration is significant, its impact on easing labour shortages depends on how closely migrants’ skills match vacancies in key sectors.^[29]

Second, labour hoarding could persist as a feature of the employment landscape. As demographic trends constrain hiring and preferences shift towards shorter hours, firms may find it harder to increase labour input during upswings. This, in turn, could strengthen incentives to hoard labour during downturns.

Third, these same forces could weigh on labour productivity. In Europe, productivity growth has historically displayed a pronounced cyclical pattern (Slide 7, right panel), in part because firms tend to reduce hours rather than shed workers in downturns.^[30]

If lower job turnover continues to slow labour reallocation, it is likely to reduce the efficiency of job matching. By contrast, the stronger post-pandemic productivity growth in the United States has been linked to higher labour market churn.^[31] An ageing population is also found to slow productivity growth.^[32]

In such a scenario, Europe might escape the unemployment hysteresis that plagued past cycles, but at the cost of a decline in productivity.

However, this is of course not the only possible path. This view focuses solely on labour market dynamics and overlooks the potential for automation and artificial intelligence to boost productivity and investment, which may well also be spurred by a shrinking population.

Conclusion

Let me conclude.

The European labour market has come through recent shocks in unexpectedly good shape, helped by a mix of global tailwinds and domestic strengths.

But we should be cautious in assuming that this unique constellation of forces will last. To borrow from de Tocqueville, we should not expect copies of past cycles to guide us through original ones.

By understanding the sources of recent resilience, we can be better prepared for the next shock, whatever shape it may take.

By Philip van Doorn

Various strategies for tracking the aerospace and defense industry have made good money for investors over the past 10 years

Archer Aviation Inc.’s sales are expected to ramp up after the company finishes building its first batch of vertical takeoff and landing aircraft.

This has been an amazing year for aerospace and defense stocks so far. A longer look back highlights three exchange-traded funds that have performed well by tracking this industry group in various ways.

This is also a good time to look at individual stocks to see which companies analysts expect to grow their businesses most rapidly as defense spending ramps up and pent-up demand for commercial aircraft, parts and maintenance services plays out.

A quick look at three aerospace and defense ETFs

It is difficult for any portfolio manager to outperform the S&P 500 SPX. Low-cost index funds that track the index can come close. Then there are index funds that track sectors or smaller industry groups. It may surprise you to see how well three exchange-traded funds focusing on the aerospace and defense industries have performed over the past decade.

Here is a 10-year chart comparing total returns, with dividends reinvested, for the three ETFs with those of the S&P 500 aerospace and defense (A&D) industry group and that of the full S&P 500. The ETFs are listed by how well they have performed:

Ten-year total returns with dividends reinvested through Aug. 20, 2025.

The S&P 500 is weighted by market capitalization, as are most broad stock indexes. So is its A&D subset. But the three ETFs use various methods to track the space and weight the stocks in their portfolios.

The Invesco Aerospace & Defense ETF PPA has been the best 10-year performer among the three exchange-traded funds listed here. It and the SPDR S&P Aerospace & Defense ETF XAR have outperformed the S&P 500 and its A&D industry group. The iShares U.S. Aerospace & Defense ETF ITA has also performed well, beating the 10-year return of the S&P 500 A&D industry group but trailing the return of the full S&P 500.

Here’s how the three ETFs track the aerospace and defense industries:

— The Invesco Aerospace & Defense ETF PPA was established in October 2005. It tracks the SPADE Defense Index and holds 55 stocks. Its annual expenses total 0.57% of assets under management. This means that annual expenses will be $57 for a $10,000 investment. This ETF is rated five stars (the highest rating) within Morningstar’s “U.S. Fund Industrials” category. The underlying index is reconstituted and rebalanced quarterly and weighted by market cap, with a limitation of 10% for each stock. According to FactSet, PPA’s “broadly defined” approach to A&D means it might “include firms in non-defense industries.”

— The SPDR S&P Aerospace & Defense ETF XAR holds 38 stocks as it tracks the S&P Aerospace & Defense Select Industry Index. The index is rebalanced quarterly with a modified equal-weighted approach, so it holds stocks of large-cap, midcap and small-cap companies. It has an expense ratio of 0.35%. This fund was established in September 2011 and has a five-star rating from Morningstar. According to FactSet, XAR’s portfolio “tends to reach outside our definition of the [A&D] segment into other sectors.”

— The iShares U.S. Aerospace & Defense ETF ITA tracks the Dow Jones U.S. Select Aerospace & Defense Index of 37 stocks. The index follows a modified cap-weighting methodology that includes a 22.5% limit on individual stocks, but limits components with weightings higher than 4.5% to a total index weighting of 45%. The fund’s expense ratio is 0.38% and it is rated four stars by Morningstar. According to FactSet, this fund’s concentration risk “remains high despite the caps, reflecting the undiversified nature of the industry.”

Different weighting strategies

The largest three companies in the S&P 500 aerospace and defense industry group are GE Aerospace (GE), RTX Corp. (RTX) and Boeing Co. (BA). Together, these three companies make up 54% of the industry group’s combined market cap, according to FactSet.

According to the ETFs’ own published portfolio lists, these three stocks together make up 24.2% of the PPA portfolio, 10.6% of the XAR portfolio and 45.2% of the ITA portfolio.

In an interview with MarketWatch, State Street Investment Management Global Head of Research Strategists Matthew Bartolini explained his preference for the less-concentrated approach followed by the SPDR S&P Aerospace & Defense ETF.

A heavy cap weighting when “trying to harness the rise in defense spending” would mean that an investor’s exposure would “be less dictated by industry trends [and tied more to] the fortunes and failures of one specific firm,” he said.

Bartolini continued: “With a modified equal-weighted view, you will still have exposure to the large firms, but also to a greater part of the ecosystem, harnessing industry trends.”

Through Wednesday, the S&P 500 A&D industry group had returned 31.5% this year, while XAR had returned 29.3%. When asked about prospects for the underlying stocks after such large gains, Bartolini expressed confidence for the space long term, based on plans for increased defense spending by the U.S., Germany and other European countries.

“The geopolitical risk is elevated, which continues to call for aerospace and defense spending and solutions,” he said.

Stock screen-aerospace and defense

Combining the three ETFs’ holdings gives us a total of 64 stocks. To have a good sampling of estimates, we have cut the list to 50 companies covered by at least five analysts polled by FactSet and for which consensus sales estimates were available through 2027. The data was adjusted by FactSet to match calendar years, accounting for companies whose fiscal reporting periods don’t match the calendar.

Among the remaining 50 companies, these 10 are expected to show the highest compound annual growth rates (CAGR) for sales from 2025 through 2027:

   Company                                    Ticker   Two-year estimated sales CAGR through 2027  Market cap ($mil)Held by 
   Archer Aviation Inc.                      ACHR                                        1,287.5%              5,954XAR, ITA 
   Rocket Lab Corp.                          RKLB                                           42.4%             19,629XAR, PPA, ITA 
   Palantir Technologies Inc.                PLTR                                           35.1%            354,808PPA 
   Intuitive Machines Inc. Class A           LUNR                                           32.5%              1,019XAR 
   BlackSky Technology Inc.                  BKSY                                           31.4%                579PPA 
   Redwire Corp.                             RDW                                            29.6%              1,253XAR 
   AeroVironment Inc.                        AVAV                                           24.1%             11,459XAR, PPA, ITA 
   Axon Enterprise Inc.                      AXON                                           22.5%             59,733XAR, PPA, ITA 
   C3.ai Inc.                                AI                                             21.2%              2,217PPA 
   Kratos Defense & Security Solutions Inc.  KTOS                                           17.7%             10,848XAR, PPA, ITA 
                                                                                                                    Source: FactSet 

In comparison, the projected revenue CAGR for the S&P 500 from 2025 through 2027 is 5.4%, according to FactSet.

The projected CAGR for Archer Aviation (ACHR) reflects lofty expectations as the company nears completion of its first “Midnight” electric vertical takeoff and landing aircraft, designed for military and commercial use. The company had no revenue during 20024, and the consensus estimates are for sales totaling $2 million this year, increasing to $84 million in 2026 and $320 million in 2027.

Here’s more data, this time showing projected sales growth rates for the largest four companies by market cap held by any of the above ETFs in the aerospace and defense industry group, as defined by Standard & Poor’s:

   Company                         Ticker  Two-year estimated sales CAGR through 2027  Market cap ($mil)Held by 
   GE Aerospace                  GE                                             10.2%            282,543XAR, PPA, ITA 
   RTX Corp.                     RTX                                             5.9%            209,602XAR, PPA, ITA 
   Boeing Co.                    BA                                             12.3%            170,604XAR, PPA, ITA 
   Honeywell International Inc.  HON                                             4.7%            137,773PPA 
                                                                                                        Source: FactSet 

Click on the tickers for more about each company.

Read: Tomi Kilgore’s detailed guide to the information available on the MarketWatch quote page

Don’t miss: The S&P 500 is at its most expensive by this measure. These stocks have bucked the trend.

-Philip van Doorn

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

  (END) Dow Jones Newswires

  08-23-25 1226ET
Copyright (c) 2025 Dow Jones & Company, Inc.

Europe’s labor market has proved surprisingly resilient in the face of a once-in-a-generation inflation shock and aggressive interest-rate hikes, according to European Central Bank President Christine Lagarde.

Employment expanded by 4.1% between the end of 2021 and mid-2025, nearly as much as the economy and roughly twice as much as an established economic rule would suggest, Lagarde said Saturday in Jackson Hole, Wyoming, where she’s attending the Federal Reserve’s annual symposium. She added that both global tailwinds and domestic strengths helped deliver such an outcome.

Introduction

Lung cancer has become one of the most fatal cancers at present,^1,2 which leads to a 5-year survival rate of less than 15%.³ Traditional treatments such as surgery, chemotherapy, and radiation therapy showed limited therapeutic effects on lung cancer.⁴ It is in great demand to develop novel strategies to improve the therapeutic effects. Phototherapy, including photothermal therapy (PTT) and photodynamic therapy (PDT), has shown powerful effects on cancer therapy, which is highlighted by the advantages such as high specificity, non-invasiveness, and negligible drug resistance compared to conventional cancer treatments.⁵ The Jablonski diagram indicates that when a photosensitizer molecule is excited by absorbing a photon, the excited electrons in the photosensitizer are unstable and dissipate energy through fluorescence, thermal, and intersystem crossing transition with energy releasing, which generates phosphorescence or interacts with surrounding substances like oxygen and H₂O to generate ROS including various cytotoxic oxygen radicals (such as superoxide anion radicals (•O- 2), hydroxyl radicals (•OH), and hydrogen peroxide (H₂O₂)) and the singlet oxygen (¹O₂).^6,7 The effects can be used for fluorescence imaging, PTT, or PDT, respectively.⁸ The effective phototherapy is largely dependent on the photosensitizers.⁹ However, conventional photosensitizers are limited to their aggregation-caused quenching behavior, resulting in weakened fluorescence emission and reduced ROS generation in the aggregated form.¹⁰ The disadvantages have seriously restricted the application of PTT in cancer diagnosis and treatment.

The aggregation-induced emission luminogens (AIEgens) were considered favorable photosensitizers for diagnosis and treatment, which have recently become popular in biomedical applications.^11–13 Different from the conventional photosensitizers, AIEgens exhibit intense emission and enhanced ROS generation when aggregated.^14,15 Beneficial from this unique photophysical phenomenon, AIEgens show better imaging or phototherapy performances than conventional photosensitizers under light irradiation, which is useful in the theranostics of cancer.¹⁶ Recently, a novel AIE active zwitterionic compound C₄₁H₃₇N₂O₃S₂ (named BITT) was successfully synthesized.¹⁷ BITT nanoparticles exhibited excellent aggregation-induced emission (AIE) characteristics, good photothermal stability, and effective reactive oxygen species generation capacity, which could provide superior fluorescence imaging and photothermal/photodynamic performance on cancer diagnosis and treatment.¹⁸ Nevertheless, several problems might still need to be solved for further application. BITT nanoparticles for phototherapy might be limited by their hydrophobic nanostructure and poor tumor accumulation.¹⁹ The limited tissue permeability might also decrease the fluorescence imaging ability and photothermal/photodynamic effect. Thus, the improvement of the surface properties may endow the AIEgen with enhanced tumor-targeting ability and high tissue permeability, which is beneficial for the treatment of lung cancer.

Cell membrane coating technology has been a useful strategy to improve the surface function of nanoparticles for precise cancer therapy.^20–22 These nanoparticles functionalized with cell membranes leveraged the intrinsic characteristics of nanoparticles and unique functionalities of natural cells, facilitating their stability, long circulation, and improved accumulation in the tumors.^23–25 CD8⁺ T cells are considered as one of the T lymphocytes for mediating adaptive immune responses, which play important roles in immunotherapy.^26–28 However, the programmed cell death protein 1 (PD-1), which is highly expressed on the activated CD8⁺ T cells, can specifically target programmed cell death ligand 1 (PD-L1) expressed on tumor cells and help tumor cells evade immune surveillance.^29–31 As a result, the T cell-mediated cancer-killing effect might be suppressed by the interaction between PD-1 and PD-L1.³² Furthermore, CD8⁺ T cell membrane’s phospholipid bilayer inherently improves nanoparticle hydrophobicity by providing an amphiphilic interface: its outer leaflet, enriched with hydrophilic phosphorylcholine headgroups and glycolipids, masks the nanoparticle’s core hydrophobicity through biomimetic surface hydration, while the inner hydrophobic fatty acid tail region interfaces directly with hydrophobic nanoparticle surfaces via lipophilic interactions. This structural duality simultaneously enhances water compatibility through the membrane’s hydrophilic exterior and stabilizes hydrophobic nanoparticle cores via the lipid tails’ affinity, effectively shielding hydrophobic domains and improving colloidal stability without compromising membrane protein functionality like PD-1/PD-L1 targeting.^33,34 Inspired by the facts mentioned above, the biomimetic nanoparticles coating with PD-1 overexpressed CD8⁺ T cell membrane could specifically target the tumor cells by blocking the immune checkpoint ligand PD-L1. Biomimetic nanoparticles decorated with CD8⁺ T cell membrane might accumulate on tumor cells with the PD-1 and PD-L1 interaction and reactivate the endogenous T cell-mediated immune response by immune checkpoint blockade. Hence, the CD8⁺ T cell membrane decoration on BITT might endow good tumor-targeting ability and high tissue permeability for BITT, and enhance superior photothermal efficiency and antitumor immunity in lung cancer therapy.

In the present work, a type of biomimetic AIEgen (termed TB) was developed for efficient phototherapy and immunotherapy on lung cancer based on the AIE and ICB strategies. In this biomimetic platform, the AIEgen BITT nanoparticles were used as the core and severed as an excellent photosensitizer for photothermal and photodynamic efficacy. To amplify the photothermal/photodynamic efficacy and activate the antitumor immunity, BITT nanoparticles were camouflaged with CD8⁺ T cells membrane (Scheme 1). TB was expected to accumulate in the lung cancer cells with the PD-1 and PD-L1 interaction, then would enhance the photothermal effects and ROS efficacies with the AIE characteristic, and synergistically exert intrinsic immune activation by immune checkpoint blockade (Scheme 1). Furthermore, the AIEgen BITT developed in this study exhibited both photodynamic and photothermal properties, allowing synergistic phototherapy without the need for multiple agents or complex formulations. This research not only eliminated the primary tumor but also activates the immune response in the tumor microenvironment, thereby further expanding its therapeutic potential.

Scheme 1 The preparation of TB and the cancer therapy induced by TB. TB, CD8+ T cells membsranes/BITT. Abbreviations: PDT, photodynamic therapy; PTT, photothermal therapy; ICB, immune checkpoint blockade.

Materials and Methods

Materials

2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) was purchased from Sigma-Aldrich (St. Louis, USA). Fetal bovine serum (FBS), Roswell Park Memorial Institute (RPMI-1640) medium, Dulbecco’s modified Eagle medium (DMEM) medium, and trypsin were purchased from Gibco. The antibiotics (penicillin and streptomycin), phosphate-buffered saline (PBS), CCK-8, Calcein-AM/PI double stain kit, Actin-488, and Hoechst 33342 were purchased from Beyotime Biotechnology. BITT was obtained from Xi’anruixi Biological Technology Co., Ltd. All other reagents were purchased from commercial suppliers without further purification.

Extraction and Identification of CD8⁺ T Cells

Male C57BL/6 mice were anesthetized with 5% isoflurane and sacrificed. The spleens were collected from mice and washed with HBSS (Hank′s balanced salt solution). Then the spleen was chopped and digested with the digestive enzymes, shaken at 37°C for 60 min, and filtered through a 300-mesh filter. Finally, the filtered mixture was collected and resuspended in a complete RPMI 1640 medium. The mouse T lymphocyte cells were isolated from the spleens of normal C57BL/6 mice with Lymphoprep (STEMCELL Technologies Inc., Vancouver, BC, Canada), following the manufacturer′s instructions. Mouse T lymphocyte cells were maintained in RPMI 1640 medium supplemented with 10% FBS and 1% antibiotics (penicillin/streptomycin). CD8⁺ T cells were purified by EasySep™ Mouse CD8+ T Cell Isolation Kit (STEMCELL Technologies Inc., Vancouver, BC, Canada). The harvested T cells were stained with fluorochrome-conjugated antibodies against CD3, CD8, PD-1, and CD45, and analyzed on a flow cytometry analyzer (BD LSR Fortessa, BD, CA). Furthermore, CD8⁺ T cells were incubated with Phorbol-12-myristate-13-acetate (50 ng/mL) and Ionomycin (1 μg/mL) at 37°C for 4 h. The cells were collected and stained with flow cytometry antibodies including CD8 and CD69. To detect the PD-L1 on the LLC cells, the flow cytometry antibodies PD-L1 and TTF1 were used to stain the LLC cells.

Preparation of BITT Nanoparticles

BITT nanoparticles were prepared according to the previous study with minor modifications.¹⁷ In brief, the BITT solution (1.0 mg/mL in DMSO, 2 mL) was added dropwise into 18 mL of deionized water under sonication (Scientz-IID sonicator, Scientz Biotechnology Co., Ltd., Ningbo, China) for 5 min. Then, free BITT was removed by dialysis (MWCO:14000 Da) in deionized water for 24 h. Finally, BITT nanoparticles were concentrated by ultrafiltration and stored at 4°C before use.

Preparation of TB

T cell membrane-coated BITT (TB) was prepared according to the previous study.¹⁹ Firstly, the T cell membrane (CM) was extracted from T cells by a membrane protein extraction kit, following the instructions from the manufacturer (Biyotime, China). The obtained T cell membrane was suspended in PBS and sonicated by a Scientz-IID ultrasonicater (Scientz Biotechnology, China). The bicinchoninic acid (BCA) method was applied to determine the concentration. The cell membranes were stored at −80°C until use.

To coat the BITT nanoparticles with CD8⁺ T cell membranes, BITT nanoparticles were mixed with cell membranes at different weight ratios, and sonicated for 1 h by a Scientz IID ultrasonicater (Scientz Biotechnology, China). Finally, the mixture was extruded through the 200 and 100 nm polycarbonate membrane using a mini extruder (Avanti, Canada) 20 times.

Characterization of TB

The particle size and morphology of TB were observed on a JEM-100XII transmission electron microscope (TEM, JEOL, Japan). Particle sizes, size distribution, and zeta potential of TB were analyzed on a Malvern Zeta sizer Nano ZS90 measurement (Malvern, UK) at room temperature via the dynamic light scattering method (DLS). UV absorption spectra of BITT and TB (20 μg/mL BITT in water) were measured on a UV-2600 UV-Vis spectrophotometer (SHIMADZU, Japan). Fluorescence spectra of BITT and TB (BITT at a dosage of 20 μg/mL) were measured with an emission wavelength of 534 nm on a fluorescence spectrophotometer (Hitachi, Japan). Fluorescence spectra of BITT and TB (20 μg/mL BITT in water) were also measured with an excitation wavelength of 730 nm on a fluorescence spectrophotometer (Hitachi, Japan). SDS-PAGE gel electrophoresis assay was displayed for the protein characterization.

Detection of TB Nanoparticles by Singlet Oxygen Sensor Green (SOSG) Probe

Different formulations (BITT equivalent to 20 μg/mL) were incubated with singlet oxygen sensor green (SOSG) probe (5 μM) and exposed to 658 nm laser for 5 min with the power of 2 W∙cm⁻². PBS was used as control. The emission spectrums range were scanning at Ex = 504 nm with fluorescence spectrophotometer (Hitachi, Japan).

Photothermal Performance and Photothermal Stability of TB

To evaluate the photothermal conversion behavior, the water solutions of BITT and TB (BITT at a dosage of 20 μg/mL) were continuously exposed to a 658 nm laser irradiation with a power density of 1 or 2 W∙cm⁻² for 6 min, respectively. The temperature was measured every 10s and stopped until the temperature nearly arrived at a plateau. For the photothermal stability studies, the water solutions of BITT and TB (BITT at a dosage of 20 μg/mL) were irradiated under a 658 nm laser irradiation at 2 W∙cm⁻², respectively. The temperatures of the samples were measured in six circles of heating and cooling process. During one heating and cooling cycle, the samples were irradiated with a 658 nm laser for the first 5 min to reach a steady state. Following the removal of the laser irradiation, the samples were allowed to cool down naturally to the ambient temperature for 5 min.

Thermal Imager Observed the Temperature of BITT in the Solution

Different formulations (BITT equivalent to 20 μg/mL) were exposed to 658 nm laser for 5 min with the power of 2 W∙cm⁻². The temperatures were monitor by FLUKE Ti480u thermal infrared imager.

Cell Culture

Mouse Lewis lung carcinoma (LLC) cells were purchased from the Shanghai Institute of Cell Biology, Chinese Academy of Sciences (Shanghai, China). LLC cells were cultivated in RPMI1640 supplement with 10% fetal bovine serum and 1% antibiotics (penicillin/streptomycin) and maintained at 37°C in a humidified atmosphere with 5% CO₂.

Binding Efficiency

LLC cells with a density of 1×10⁵ cells/well were seeded in confocal imaging dishes and cultured for 24 h. Then, LLC cells were incubated for 6 h with TB at various BITT concentrations (5, 10, 15, and 20 μg∙mL⁻¹), respectively. LLC cells treated with TB (20 μg∙mL⁻¹) were also incubated for different hours (1, 3, 6, and 9 h). After that, the treated LLC cells were washed with PBS three times. For confocal imaging, the treated cells were stained with Hoechst 33342 for 30 min. The cellular images were observed by a Zeiss LSM 880 confocal microscope (Zeiss, USA). For FACS analysis, the treated cells were trypsinized, collected, and analyzed with a flow cytometer (Becton Dickinson, USA). LLC cells without treatment were chosen as the negative control.

PD-L1 Inhibitor BMS-1 Verify the Binding of PD-1/PD-L1

LLC cells with a density of 5×10⁴ cells per well were seeded into confocal imaging dishes and cultured for 24 h. And then added the 5 μM BMS-1 to LLC cells and co-culture at 37°C for 2 h. Subsequently, the nanoparticles containing BITT from different groups were co-cultured for 9 h. After washing with PBS, the fluorescence of BITT was observed using confocal microscopy.

Cytotoxicity Study

The cytotoxicity of TB was performed on LLC cells by the CCK-8 method. In brief, LLC cells with a density of 5×10³ cells per well were seeded into 96-well plates and cultured for 24 h. Then, the LLC cells were incubated with different concentrations of BITT in fresh media for 9 h. Afterward, the BITT-treated cells were exposed to laser irradiation (2 W∙cm⁻²) for 5 min. Meanwhile, the BITT-treated cells without light irradiation were also conducted for the dark cytotoxicity study. After further incubation for 2 h, the treated cells were washed with PBS and then incubated with fresh medium containing 10% CCK-8 at 37°C for another 4 h. Finally, the absorbance was measured at a wavelength of 450 by a microplate reader (Thermo, USA). The relative cell viability was calculated after different treatments. The cells without any treatment were chosen as the control.

Live/Dead Staining

LLC cells with at a density of 1×10⁵ cells per well were seeded into 24-well plates and cultured for 24 h. Then, LLC cells were incubated with PBS, BITT, and TB (BITT, 20 μg∙mL⁻¹) for 9 h, respectively. After that, the treated cells were exposed to the laser irradiation (2 W∙cm⁻²) for 5 min. Meanwhile, the treated cells without light irradiation were also conducted as the negative controls. After 2 h incubation, the cells were stained with a Calcein-AM/PI Double Stained Kit (Biyuntian, China) for 10 min. Finally, the treated cells were washed with PBS and imaged by an inverted fluorescent microscope (Leica, Japan).

Intracellular ROS Generation

LLC cells with at a density of 1×10⁵ cells per well were seeded into confocal imaging dishes and cultured for 24 h. Then, the cells were incubated with PBS, BITT, and TB (20 µg∙mL⁻¹ BITT) in the dark for 6 h, respectively. DCFH-DA was added, and the cells were incubated at 37°C for 20 min. After washing with PBS, the treated cells were exposed to laser irradiation (2 W∙cm⁻²) for 5 min, followed by CLSM imaging. The treated cells without laser irradiation were chosen as the negative controls.

Penetration of TB Into Lung 3D in vitro Tumor Spheroid Model

The in vitro lung 3D tumor spheroid model was established according to the previous method.¹⁹ When the lung 3D tumor spheroid reached a size of 100 μm, PBS, BITT, and TB (20 µg∙mL⁻¹ BITT) were added and penetrated the tumor spheroids in the dark at 37°C for 6 h, respectively. After that, the treated tumor spheroids were exposed to laser irradiation (2 W∙cm⁻²) for 5 min. After 2 h incubation, the tumor spheroids were stained with a Calcein-AM/PI Double-Stained Kit (Beyotime Biotechnology, China) for 10 min. Finally, the tumor spheroids were washed with PBS and imaged by CLSM. The treated tumor spheroids without laser irradiation were chosen as the negative controls.

Animals and Tumor-Bearing Mouse Model

The male C57BL/6 mice at 6–8 weeks old (20 g weight per mouse) were obtained from Shanghai Slac Laboratory Animal Co. Ltd (Shanghai, China). We performed the animal studies according to the Guide for the Care and Use of Laboratory Animals published by US National Institutes of Health (No. 85–23, 1996). All the animal experiments were performed by the protocols approved by the Institutional Animal Care and Use Committee of Guangzhou Medical University (GY2022-024). To establish the LLC tumor-bearing mice model, LLC cells (2×10⁶ cells suspended in 100 μL PBS buffer per mouse) were injected subcutaneously into the right flanks of C57BL/6 mice. When the tumor volume reached about 100 mm³, the LLC tumor-bearing mice were used for the in vivo experiments.

In vivo Fluorescence Imaging

The LLC tumor-bearing C57BL/6 mice were anesthetized with 5% isoflurane. Then, the mice were administrated with BITT and TB with a BITT dose of 1 mg/kg via intravenous injections, respectively (n=3). The in vivo fluorescence imaging was captured on the in vivo imaging system (PerkinElmer, USA) at different time points (1, 4, 8, 12, 24, and 48 h) after injection. To evaluate the tissue distribution of TB, the mice were sacrificed at 48 h postinjection. Tumors and major organs including heat, liver, spleen, lung, and kidney were harvested, followed by fluorescence imaging and quantitative analysis via the in vivo imaging system (PerkinElmer, USA). To investigate the pharmacokinetics of TB, the plasma was collected from the TB or BITT-treated mice after post-injection at different time points (0, 1, 2, 4, 6, 8, 12, and 24 h). The relative fluorescence intensity was also measured by the in vivo imaging system (PerkinElmer, USA). Mice with PBS treatment were chosen as the negative control.

In vivo Therapeutic Study

To evaluate the in vivo therapeutic efficacy of TB, the LLC tumor-bearing mice were randomly divided into five groups (n=4), followed by the treatments of PBS, PBS with laser irradiation (PBS+L), BITT, BITT with laser irradiation (BITT+L), and TB with laser irradiation (TB+L), respectively. For PBS and BITT groups, the LLC tumor-bearing mice were intravenously injected with PBS and BITT (1 mg/kg) without the subsequent laser irradiation, respectively. For PBS+L, BITT+L, and TB+L groups, the tumor-bearing mice were administrated with PBS, BITT, and TB (1 mg/kg BITT) via intravenous injections, respectively. At 12 h after administration, the tumor sites were exposed to laser irradiation (658 nm, 2 W∙cm⁻²) for 10 min. The treatments were carried out every three days during the 15-day study, and the tumor volume and body weight were both recorded. The tumor volume (V) was measured via a vernier caliper and calculated by the equation (1):

(1)

Where V is the tumor volume, a is the tumor length and b is the tumor width.

The relative tumor volume (RTV) was calculated with the equation (2):

(2)

Where RTV is the relative tumor volume, V is the tumor volume and V0 is the initial tumor volume on day 0.

The relative body weight (RBW) was calculated by the equation (3):

(3)

Where RBT is the relative body weight, W is the body weight and W0 is the initial body weight on day 0.

Histological and Immunohistochemical Analysis

After the treatment of 15 days, mice after different treatments were euthanized. Blood samples of the treated mice were collected and tested for the hematology analysis. Then, the mice were sacrificed. The tumors and major organs (including hearts, livers, spleens, lungs, and kidneys) were collected and fixed in 4% formalin saline for 24 h. The fixed tissues were embedded in paraffin and sectioned at 5 μm thickness. The slices of tumors were stained with H&E, TUNEL, CD4, and CD8 staining, respectively. The stained slices were photographed by an inverted fluorescence microscope (Leica, Japan). For the in vivo biosafety evaluation, the collected major organs were also sectioned and stained with H&E analysis.

Statistical Analysis

Experiments were performed on at least three replicates. All the results were expressed as the mean standard deviation (SD). Statistical significance was analyzed by one-way analysis of variance (ANOVA) with a Tukey post-hoc test using SPSS 13.0 software. A p-value of less than 0.05 (p < 0.05) was considered statistically significant.

Results and Discussion

Synthesis and Characterization of BITT and TB

The preparation of TB was illustrated in Scheme 1. The BITT nanoparticles were prepared by the self-assembly method in deionized water, followed by the sonication of the nanoparticles in the suspension. For the T cell membrane coating on BITT, T cells were harvested from the spleens in mice. The collected T cells expressed high levels of CD8, CD45, and CD3 (Figure S1A–S1C in the supplementary materials file), which indicated that the CD8⁺ T cells were successfully collected. After the induction with Phorbol-12-myristate-13-acetate and Ionomycin, PD-1 was detected to be highly expressed on the cell surface by flow cytometry (FACS) analysis (Figure S2 in the supplementary materials file), which was an important receptor on T cells for immune checkpoint therapy and served as a key targeting ligand for active tumor accumulation. The cell membranes were collected and used to camouflage BITT nanoparticles through sonication and physical extrusion. We investigated the physicochemical properties of BITT and TB. The transmission electron microscope (TEM) images indicated that both BITT and TB were uniform and spherical (Figure 1A and B). Dynamic light scattering (DLS) measurement indicated that BITT nanoparticles showed an average hydrodynamic diameter of 130.0 ± 1.7 nm with a PDI of 0.15, and the one of TB was 199.3 ± 4.1 nm with a PDI of 0.12, consistent with the TEM observations (Figure 1A and B). DLS measurement and TEM imaging demonstrated that the particle sizes of BITT and TB might be beneficial for improving the circulation lifetime with an appropriate size of less than 200 nm.²⁴ To confirm the CD8⁺ T cell membrane decoration on BITT, the profiles of membrane proteins on TB were performed via sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The SDS-PAGE indicated that TB displayed a protein profile similar to those of T cell membranes and T cell lysate (Figure 1C). The data mentioned above demonstrated that TB was successfully prepared by coating BITT with CD8⁺ T cell membranes.

Figure 1 Characterization of the nanoparticles. (A) TEM and DLS analysis of BITT. (B) TEM and DLS analysis of TB. (C) SDS-PAGE analysis of T cell proteins (TP), T cell membrane (TMP), and TB. (D) Zeta potential of BITT and TB. (E) Excitation spectra of BITT and TB with an emission wavelength of 730 nm (20 μg∙mL−1 BITT). (F) Emission spectra of BITT and TB with an excitation wavelength of 658 nm (20 μg∙mL−1 BITT). (G) UV absorption spectra of BITT and TB (20 μg∙mL−1 BITT). (H) Photothermal effect of TB under 658 nm laser irradiation with a powder density of 1 or 2 W∙cm−2, respectively. (I) Photothermal stability of BITT and TB (20 μg∙mL−1 BITT) under laser irradiation (658 nm, 2 W∙cm−2) during 6 cycles of heating and natural cooling processes.

We further analyzed the parameters, such as zeta potential, fluorescence/UV-Vis spectra, and thermal properties. Zeta potential analysis indicated that BITT was −15 mV and TB was −30 mV (Figure 1D), respectively, implying that the camouflage with T cell membranes induced a higher negative charge. The fluorescence spectra analysis indicated that BITT and TB in water displayed a maximum excitation wavelength at 534 nm and the fluorescence emission one at 730 nm (Figure 1E and F), implying that the camouflage with cell membranes did not change the fluorescent characteristics of BITT significantly. The Ultraviolet-visible spectroscopy (UV-Vis) analysis also indicated that the maximum absorption peak was at ~550 nm (Figure 1G). To investigate the photothermal conversion behavior of TB, the temperature changes were recorded under laser irradiation. The photothermal heating curves of TB in water revealed power density-dependent temperature variation behavior under the 658 nm laser irradiation (Figure 1H). The photothermal properties of BITT were also studied under 658 nm laser irradiation. The temperature of TB was significantly increased with the laser irradiation power density increased. When the irradiation power density was 2 W∙cm⁻², the temperature of TB reached 50°C after irradiation for 5 min. At the same time, to further evaluate the photodynamic performance of TB nanoparticles in solution, we used SOSG probe to monitor the changes in BITT before and after light irradiation (Figure S3 in the supplementary materials file). The result showed that, compared with the control group, both the TB+L and BITT+L groups exhibited significantly enhanced SOSG fluorescence, indicating efficient singlet oxygen generation. This suggests that the photodynamic performance of TB nanoparticles in solution remains unaffected. The photothermal stability of TB was also investigated. Under a 658 nm laser irradiation (2 W∙cm⁻²), the maximum solution temperatures of BITT and TB (20 μg∙mL⁻¹ BITT) showed no obvious change during the 6 cycles of the natural cooling process, suggesting the excellent photostability of both BITT and TB (Figure 1I). The parameters indicated that TB might be useful in phototherapy. To eliminate the influence of TB solution on the photodynamic performance, we used a thermal infrared imager to test the different components containing BITT (Figure S4 in the supplementary materials file). From the result, it can be seen that except for the BITT group, the other control groups did not show any temperature changes whether in the presence or absence of irradiation. That indicated that the temperature increase of the TB nanoparticles was caused by BITT rather than the solvent.

The results indicated that the T cell membrane coating did not affect the excellent optical performance of BITT. For example, TB showed excellent light-to-heat conversion ability and photothermal stability after a 658 nm laser irradiation, which might be applied to tumor photothermal therapy. All these characteristics demonstrated that we successfully prepared TB, which might be useful in the treatment of lung cancer.

The Binding to LLC Cells

We evaluated the binding to the cells in vitro by the nanoparticles. To investigate the binding, confocal laser microscopy (CLSM) and flow cytometry (FACS) analysis was performed. The BITT-positive cells displayed a BITT concentration and time-dependent manner (Figure 2A and B). FACS analysis showed that when the concentration of BITT increased to 20 μg∙mL⁻¹, the BITT-positive LLC cells were up to 93.6% after 6 h incubation, which was a 15-fold increase compared with those in low concentration (BITT, 5 μg∙mL⁻¹) (Figure 2A). TB bound to LLC cells within 1 h, and stronger fluorescence could be observed with the incubation time extended (Figure 2B). The BITT-positive cells were increased to 92.4% with the incubation time extended to 9 h. Moreover, the fluorescence intensity of TB was stronger than that of the negative control and BITT group (Figure S5 in the supplementary materials file). The data indicated TB showed a high-level binding efficiency in the optimized conditions including the dosage of 20 μg∙mL⁻¹ and incubation for 9 h with TB, which was used for further experiments. TB exhibited excellent specific binding ability to LLC cells in vitro, which might be due to the interactions between the overexpressed PD-1 receptors on TB and the overexpressed PD-L1 ligand on LLC cells (Figure S6 in the supplementary materials file). To verify that the binding of PD-1/PD-L1 effectively promotes the uptake of BITT by LLC cells, we treated the LLC cells with the PD-L1 inhibitor (Figure S7 in the supplementary materials file). From the result, the fluorescence intensity of the TB group with the inhibitor was similar to that of the BITT group. Clearly, the TB group had the strongest fluorescence due to the mutual binding of PD-1 and PD-L1. CLSM and FACS analysis demonstrated that TB bound to LLC cells effectively. The in vitro phototherapeutic effect of TB will be further evaluated.

Figure 2 The binding efficiency of TB to LLC cells. (A) CLSM and FACS analysis of LLC cells treated with TB at various BITT concentrations for 6 h. LLC cells were treated with TB at various BITT concentrations (5, 10, 15, and 20 μg∙mL−1) for 6 h. (B) CLSM and FACS analysis of LLC cells treated with TB for 1, 3, 6, and 9 h, respectively. For CLSM analysis, the cells were stained with Actin-488 (green), and the nuclei were stained with Hoechst (blue). The cells without any treatment were chosen as the negative control (NC).

In vitro Phototherapy Effects

As an AIEgen-base photosensitizer, TB was supposed to show excellent photothermal capacity and efficient intercellular ROS generation ability for cancer therapy. ROS generation induced by AIEgen-base photosensitizer might play an important role in cancer cell inhibition. The intracellular ROS generation of TB was evaluated by DCFH-DA as an indicator. LLC cells treated with BITT and TB displayed a strong green fluorescence signal upon 658 nm laser irradiation for 5 min, but no obvious green fluorescence signal was observed in the negative controls (PBS, PBS+L, or BITT without laser irradiation) (Figure 3C). These phenomena demonstrated that both BITT and TB with laser irradiation aroused intracellular ROS generation efficiently. The green fluorescence signal of TB was significantly stronger than that of BITT, indicating that the CD8⁺ T cell membrane coating might enhance the intracellular ROS generation of TB.

Figure 3 In vitro phototherapeutic effects on LLC cells treated with different formulations. (A) Live/dead assay of LLC cells by Calcein-AM/PI staining after different treatments. (B) Cell viability of LLC cells measured by CCK-8 method after treatment with different BITT concentrations with or without laser irradiation. (C) Intracellular ROS level in LLC cells indicated by DCFH-DA after different treatments. (D) Live/Dead assay of LLC 3D tumor spheroid stained by Calcein-AM/PI after different treatments. The laser irradiation was performed in the parameters including 658 nm, 2 W∙cm−2, and 300 s). Scanning images of the LLC 3D tumor spheroid by CLSM were shown at the depth from 68 μm to 101 μm. For A and D, the live and dead cells were distinguished with Calcein-AM/ PI. **P<0.01; ***P<0.001; ns, no significant differences.

The Live/Dead assay was also carried out by Calcein-AM/Propidium Iodide (PI) staining to differentiate live (green) and dead (red) cells. As illustrated in Figure 3A, the strong green fluorescence was found in laser irradiation or BITT without laser irradiation group. However, the strong red one was observed in cells with BITT or TB with laser irradiation, which indicated that BITT or TB with laser irradiation could induce the death of lung cancer cells effectively. Compared to BITT, the red fluorescence intensity of TB was significantly stronger, which indicated that TB exhibited enhanced tumor inhibition by the coating with CD8⁺ T cell membranes.

To evaluate the in vitro cell inhibition, the Cell counting kit-8 (CCK-8) assay was performed by the treatment with TB with or without laser irradiation on LLC cells. As shown in Figure 3B, the cell viabilities of LLC cells were retained up to 90% for 48 h, with BITT concentrations ranging from 0 to 40 μg/mL, indicating the good biocompatibility of TB. After the exposure to 658 nm laser irradiation, the cell viabilities of LLC cells were gradually decreased with BITT concentration increased, which was lower than 10% with BITT increasing to 20 μg/mL. The CCK-8 results indicated that TB led to effective cell death in the presence of laser irradiation, which demonstrated that TB showed excellent cell inhibition effect in the presence of laser irradiation.

To evaluate the potential tumor inhibition, the three-dimensional (3D) tumor spheroid models generated with LLC cells were constructed to evaluate the penetration ability and the phototherapeutic effect of TB in vitro. The 3D tumor spheroid of LLC cells treated with different formulations was also stained by Calcein-AM/PI assay. CLSM analysis indicated that the 3D tumor spheroids showed significant red fluorescence after the treatment with BITT upon laser irradiation (Figure 3D), implying that most of the cancer cells were dead with PI as an indicator. By contrast, the green fluorescence but without an obvious red fluorescence signal could be found in the spheroids treated with the control groups including PBS, PBS plus Laser irradiation, and BITT, which meant that most of the cancer cells were alive. The results indicated that after the AIEgens (BITT and TB) penetrated the LLC 3D tumor spheroids, cell death was achieved in the presence of laser irradiation. Moreover, TB improved the penetration and phototherapeutic effects within the LLC 3D tumor spheroids compared to BITT (Figure 3D), which might be beneficial from the CD8⁺ T cell membrane coating. The in vitro 3D tumor spheroid study indicated that TB with laser irradiation might show great potential in lung cancer therapy.

The results showed that TB enhanced tumor accumulation and promoted cell inhibition in the presence of laser irradiation, which might be attributed to the good tumor-targeting ability via the CD8⁺ T membrane coating and BITT-induced efficient intercellular ROS generation and photothermal effect. Thus, TB exhibited excellent phototherapeutic effect on lung cancer cells in vitro, which might be useful for the in vivo application of tumor phototherapy.

In vivo Tracking

To evaluate the phototherapy, the tumor accumulation and circulation properties of the nanodrugs are critical parameters. We investigated the above factors of TB in LLC-tumor-bearing mice. As illustrated in Figure 4A, significant fluorescence signals were captured on tumor sites after the intravenous injection of TB for 1 h. Then, the fluorescence intensity in the tumor area was gradually increased and reached the highest level at 8 h post-injection, implying that TB could be efficiently accumulated in the tumors. Compared with TB, BITT showed much less fluorescence in tumor sites. The results implied that TB exhibited much higher tumor accumulation than BITT, which might be ascribed to the active tumor targeting ability via CD8⁺ T cell membrane coating. Notably, the fluorescence intensity of BITT was significantly decreased after 12 h, implying that the clearance of the AIEgens from the body. However, the fluorescence intensity remained higher in TB than in BITT at 48 h post-injection, which might be attributed to the strong binding through the interactions between PD-1 on the CD8⁺ T cell membrane and PD-L1 on the lung cancer cells. The fluorescence images and semi-quantitative fluorescence intensity of the ex vivo organs from mice were detected in Figure 4B and C. The fluorescence intensity of the tumors treated with TB was 2-fold higher than those treated with BITT (P<0.001), which was consistent with the in vivo fluorescence imaging. Obviously, TB also showed significant accumulation in the liver and lungs. Nanoparticles accumulate highly in the liver primarily due to its rich blood supply, filtration function, and efficient capture by Kupffer cells (liver macrophages) of the reticuloendothelial system (RES). In the lungs, high accumulation occurs because of their dense capillary network trapping particles, initial distribution after intravenous injection or inhalation, and active uptake by alveolar macrophages.^35,36 The results demonstrated that TB showed enhanced accumulation on the tumors and excellent fluorescence for in vivo imaging, which might contribute to good phototherapeutic effect and real-time tracking.

Figure 4 In vivo imaging of TB on mice bearing xenograft LLC tumors. (A) Fluorescence images of LLC tumor-bearing mice at different time points (1, 4, 8, 12, 24, and 48 h) after intravenous injection of BITT and TB (BITT: 1 mg/kg). Mice treated with PBS were chosen as the negative control. (B) Fluorescence images of ex vivo organs on mice at 48 h after intravenous injection of TB and BITT (BITT: 1 mg/kg). Mice treated with PBS were chosen as the negative control. (C) Fluorescence intensity of ex vivo tissues on mice at 48 h after intravenous injection of TB and BITT (BITT: 1 mg/kg). **P<0.01.

The pharmacokinetic behavior of TB was also evaluated in healthy mice to investigate the circulation lifetime of TB. The results indicated that the half-time (T_1/2) of TB in blood was approximately 8 h, which was 1.6 times longer than that of BITT (T_1/2 of BITT: ~5 h, P<0.05) (Figure S8 in the supplementary materials file). Similarly, the corresponding pharmacokinetic studies demonstrated that TB effectively enhanced the long circulation after CD8⁺ T cell membrane decoration. It might be implied that both the favorable particle size of BITT and the coating with CD8⁺ T cell membrane improved the circulation lifetime significantly, which was beneficial for the accumulation of the tumors.

The in vivo tracking and pharmacokinetic indicated that TB could accumulate in the tumor sites effectively and show good performance on the imaging. Based on the camouflage with CD8⁺ T cell membranes, TB was endowed with an active targeting effect and binding to tumor tissues effectively. Furthermore, the good performance of AIEgens facilitated the real-time detection of the biodistribution of nanoparticles. Thus, TB could be quite beneficial to the in vivo phototherapy for lung cancer.

In vivo Therapeutic Effects

To evaluate the in vivo therapeutic effect of TB, the antitumor effects were investigated on LLC tumor-bearing mice (Figure 5A). The thermal imaging analysis indicated that the treatment with TB showed a temperature of 58.7°C in the presence of 658 nm laser irradiation (2 W∙cm⁻²) for 10 min (Figure S9 in the supplementary materials file). In contrast, the treatment with TB only induced a temperature of 38.5°C without laser irradiation. The treatment with BITT induced a temperature of 47.2°C in the presence of laser irradiation, implying a weaker thermal effect, compared with the TB-treated ones. Correspondingly, the tumor growth curves showed that TB with the laser irradiation exhibited the strongest antitumor effect in all the treatment groups (P<0.05 vs BITT+L, P<0.001 vs the negative controls) (Figure 5B), which was consistent with the extracted tumors (Figure 5D). However, the relative tumor volumes were increased gradually and showed no significant difference (P>0.05) in all control groups (PBS, PBS with laser irradiation, and BITT without irradiation) (Figure 5D). Although BITT with laser irradiation showed some tumor inhibition compared with the single BITT treatment (P<0.001), the tumor inhibition was much lower than the one of TB with laser irradiation (Figure 5B and C). The effective tumor inhibition induced by TB with laser irradiation indicated that not only the favorable nanostructure and aggregated-induced emission properties but also the CD8⁺ T cell membrane coating promoted the phototherapeutic effects.

Figure 5 In vivo therapeutic efficacy of TB on LLC tumor-bearing mice. (A) Experimental timeline for in vivo study. (B) Tumor growth curves of LLC tumor-bearing mice after different treatments. (C) Photograph of tumors extracted from mice on day 14 after different treatments (BITT: 1 mg/kg). (D) Body weight changes of LLC tumor-bearing mice after different treatments. (BITT: 1 mg/kg). (E) Blood routine assay of LLC tumor-bearing mice on day 14 after various treatments. *P<0.05; **P<0.01.

Biosafety is an important precondition for lung cancer therapy. We examined the body weight every other day. The relative body weight showed no significant changes in all of the administration (P>0.05) (Figure 5D), indicating that no obvious side effects were found after different treatments. After the treatment for 14 days, the blood from mice was collected for the blood routine assay, which also indicated that the routine blood indexes in all the treatment groups were presented in the normal range, implying that no apparent systemic toxicity of TB with laser irradiation treatment (Figure 5E). Hematoxylin and eosin (H&E) staining images of the major organs (including heart, liver, lung, spleen, and kidney) also showed that TB with laser irradiation had no obvious organ damages or inflammatory lesions compared with the other treatments (Figure 6A), further revealing the good biocompatibility and biosafety of TB.

Figure 6 Histological staining analysis after different treatments. (A) H&E staining images of major organ sections from LLC tumor-bearing mice on the 14th day of treatment. (B) CD4 (green) and CD8 (red) staining analysis of tumor sections from LLC tumor-bearing mice on the 14th day of treatment. (BITT: 1 mg/kg). (C) FACS analysis of the infiltration of T cells. (D) H&E and TUNEL (red) staining analysis of tumor sections from LLC tumor-bearing mice on the 14th day of treatment (BITT: 1 mg/kg). The nucleus was stained with Hoechst (blue). **P<0.01; ***P<0.001.

T cell activation plays an important role in lung cancer immunotherapy. The activation of CD4⁺ and CD8⁺ T cells in tumors potentially remodeled the tumor microenvironment to overcome tumor-associated immune suppression, which provided a promising strategy to enhance immunotherapy for lung cancer. Herein, the novel AIEgen TB camouflaged with CD8⁺ T cell membranes, not only endowed the improved accumulation by the active targeting to LLC cells by the interactions between PD-1 and PD-L1 but facilitated the immunotherapy by the activations of T cells for lung cancer. To validate the therapeutic effects of TB, the tumors were harvested on day 14 after different treatments and then evaluated with H&E, TdT-mediated dUTP Nick-End Labeling (TUNEL), and immunofluorescence staining. As expected, the immunofluorescence staining in Figure 6B showed that significant fluorescence was found in BITT and TB with laser irradiation, which indicated that AIEgens with laser irradiation increased the levels of CD4⁺ and CD8⁺ T cells. Moreover, TB with laser irradiation exhibited stronger fluorescence compared to BITT in the presence of laser irradiation. FACS analysis indicated that the treatment with TB with laser irradiations reduced the populations of Treg cells and increased the activation of CD8⁺ T cells, as was represented by the reduction of FOXP3 and CD4 signals and increase of CD69 and CD8 ones (Figure 6C). These results demonstrated that TB amplified the T cell response by the camouflage with T cell membranes. Meanwhile, H&E staining showed that the treatment with TB and laser irradiation induced significant tumor suppression with severe tumor cell necrosis or apoptosis (Figure 6D). Consistently, the TUNEL staining showed that the enhanced apoptosis of cancer cells represented by the red fluorescence in the group of TB with laser irradiation, which confirmed that TB induced the most effective apoptosis of cancer cells in the presence of laser irradiation. These results convincingly demonstrated that TB could facilitate synergistic tumor inhibition by excellent immunotherapy and phototherapeutic effects, which opened a new avenue for lung cancer therapy.

Conclusion

In conclusion, a type of biomimetic AIE aggregation (TB) was rationally designed and successfully fabricated. The prepared TB possessed improved binding efficiency, photothermal effects, and ROS generation ability to kill the lung cancer cells. Moreover, TB also showed improved circulation lifetime and excellent tumor targeting ability, which induced effective phototherapy and immunotherapy in vivo based on BITT and the CD8⁺ T cell-derived membranes. The work provided a new perspective on the design of biomimetic AIEgens and offered an encouraging strategy for precise lung cancer therapy. However, future studies should investigate the effects of this nanoplatform on memory T cell subsets, including central memory T cells, to assess long-term immunotherapeutic efficacy. The clinical translation still faces challenges requiring further investigation, such as stability and storage conditions, biocompatibility, and toxicology profiles in human systems.

Acknowledgments

This work was financially supported by National Natural Science Foundation of China (82072047, 81370299, and 32171312), Major Project of Guangzhou National Laboratory (GZNL2024A01013), Medical Scientific Research Foundation of Guangdong Province (A2023047), the Guangzhou Science and Technology Project (2024A03J0891, 202201011593, and SL2022A03J00613), Plan on enhancing scientific research in GMU (2025SRP003 and 02-410-2302068XM), The Open research funds from The Sixth Affiliated Hospital of Guangzhou Medical University, Qingyuan Peoples Hospital (202201-303 and 202301-304).

Disclosure

The authors report no conflicts of interest in this work.

References

1. Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. CA Cancer J Clin. 2024;74(1):12–49. doi:10.3322/caac.21820

2. Zhang L, Lin Y, Li S, et al. In situ reprogramming of tumor-associated macrophages with internally and externally engineered exosomes. Angew Chem Int Ed. 2023;62(11):e202217089. doi:10.1002/anie.202217089

3. Leiter A, Veluswamy RR, Wisnivesky JP. The global burden of lung cancer: current status and future trends. Nat Rev Clin Oncol. 2023;20(9):624–639. doi:10.1038/s41571-023-00798-3

4. Niu Z, Jin R, Zhang Y, et al. Signaling pathways and targeted therapies in lung squamous cell carcinoma: mechanisms and clinical trials. Signal Transduct Target Ther. 2022;7(1):353. doi:10.1038/s41392-022-01200-x

5. Overchuk M, Weersink RA, Wilson BC, et al. Photodynamic and photothermal therapies: synergy opportunities for nanomedicine. ACS Nano. 2023;17(9):7979–8003. doi:10.1021/acsnano.3c00891

6. Luo H, Gao S. Recent advances in fluorescence imaging-guided photothermal therapy and photodynamic therapy for cancer: from near-infrared-I to near-infrared-II. J Control Release. 2023;362:425–445. doi:10.1016/j.jconrel.2023.08.056

7. Xie Z, Fan T, An J, et al. Emerging combination strategies with phototherapy in cancer nanomedicine. Chem Soc Rev. 2020;49(22):8065–8087. doi:10.1039/D0CS00215A

8. Zhang Z, Du Y, Shi X, et al. NIR-II light in clinical oncology: opportunities and challenges. Nat Rev Clin Oncol. 2024;21(6):449–467. doi:10.1038/s41571-024-00892-0

9. Chen T, Yao T, Peng H, et al. An Injectable Hydrogel for simultaneous photothermal therapy and photodynamic therapy with ultrahigh efficiency based on carbon dots and modified cellulose nanocrystals. Adv Funct Mater. 2021;31(45):2106079. doi:10.1002/adfm.202106079

10. Qian J, Tang BZ. AIE luminogens for bioimaging and theranostics: from organelles to animals. Chem. 2017;3(1):56–91. doi:10.1016/j.chempr.2017.05.010

11. Wang Y, Nie J, Fang W, et al. Sugar-based aggregation-induced emission luminogens: design, structures, and applications. Chem Rev. 2020;120(10):4534–4577. doi:10.1021/acs.chemrev.9b00814

12. Zhao Z, Zhang H, Tang BZ, et al. Aggregation-induced emission: new vistas at the aggregate level. Angew Chem Int Ed. 2020;59(25):9888–9907. doi:10.1002/anie.201916729

13. Kang M, Zhang Z, Chen X, et al. Aggregation-enhanced theranostics: AIE sparkles in biomedical field. Aggregate. 2020;1(1):80–106. doi:10.1002/agt2.7

14. Cen P, Huang J, Zhang H, et al. Aggregation-induced emission luminogens for in vivo molecular imaging and theranostics in cancer. Aggregate. 2023;4(5):e352. doi:10.1002/agt2.352

15. Feng X, Tong B, Dong Y, et al. Recent progress of aggregation-induced emission luminogens (AIEgens) for bacterial detection and theranostics. Mater Chem Front. 2021;5(3):1164–1184. doi:10.1039/D0QM00753F

16. Zhang L, Wang Z, Zhao Y, et al. Multi-stimuli-responsive and cell membrane camouflaged aggregation-induced emission nanogels for precise chemo-photothermal synergistic therapy of tumors. ACS Nano. 2023;17(24):25205–25221. doi:10.1021/acsnano.3c08409

17. Zhu W, Kang M, Li C, et al. Zwitterionic AIEgens: rational molecular design for NIR-II fluorescence imaging-guided synergistic phototherapy. Adv Funct Mater. 2021;31(3):2007026. doi:10.1002/adfm.202007026

18. Ding K, Wang L, Zhang W, et al. Photo-enhanced chemotherapy performance in bladder cancer treatment via albumin coated AIE aggregates. ACS Nano. 2022;16(5):7535–7546. doi:10.1021/acsnano.1c10770

19. Lin Y, Yi M, Li S, et al. “Two birds with one stone” strategy for the lung cancer therapy with bioinspired AIE aggregates. J Nanobiotechnol. 2023;21(1):49. doi:10.1186/s12951-023-01799-1

20. Fang RH, Kroll AV, Zhang L, et al. Cell membrane coating nanotechnology. Adv Mater. 2018;30(23):1706759. doi:10.1002/adma.201706759

21. Celadon A, Sun H, Zhang G, et al. Batteries for electric vehicles: technical advancements, environmental challenges, and market perspectives. SusMat. 2024;4(5):e234. doi:10.1002/sus2.234

22. Li H, Li S, Huang X, et al. Artificial exosomes mediated spatiotemporal-resolved and targeted delivery of epigenetic inhibitors. J Nanobiotechnol. 2021;19(1):364. doi:10.1186/s12951-021-01107-9

23. Liang L, Cen H, Wang S, et al. The reversion of DNA methylation-induced miRNA silence via biomimetic nanoparticles-mediated gene delivery for efficient lung adenocarcinoma therapy. Mol Cancer. 2022;21(1):186. doi:10.1186/s12943-022-01651-4

24. Zhang Y, Yang L, Chen S, et al. Bioinspired metal–organic frameworks mediated efficient delivery of siRNA for cancer therapy. Chem Engin J. 2021;426:131926. doi:10.1016/j.cej.2021.131926

25. Yang L, Lin Y, Miao Y, et al. Biomimetic metal–organic frameworks navigated biological bombs for efficient lung cancer therapy. J Colloid Interf Sci. 2022;625:532–543. doi:10.1016/j.jcis.2022.06.008

26. Gu X, Wei H, Chen L, et al. Itaconate promotes hepatocellular carcinoma progression by epigenetic induction of CD8⁺ T-cell exhaustion. Nat Commun. 2023;14(1):8154. doi:10.1038/s41467-023-43988-4

27. Lv Z, Li Z, Zhang R, et al. A smart DNA nanoassembly containing multivalent aptamers enables controlled delivery of CRISPR/Cas9 for cancer immunotherapy. Adv Funct Mater. 2024;34(12):2311069. doi:10.1002/adfm.202311069

28. Lin YX, Wang Y, Yu M, et al. Reactivation of the tumor suppressor PTEN by mRNA nanoparticles enhances antitumor immunity in preclinical models. Sci Transl Med. 2021;13(599):eaba9772. doi:10.1126/scitranslmed.aba9772

29. Luo Z, He T, Liang X, et al. Self-adjuvanted molecular activator (SeaMac) nanovaccines promote cancer immunotherapy. Adv Healthc Mater. 2021;10(7):2002080. doi:10.1002/adhm.202002080

30. Liu L, Pan Y, Rao L, et al. Boosting checkpoint immunotherapy with biomaterials. ACS Nano. 2023;17(4):3225–3258. doi:10.1021/acsnano.2c11691

31. Zhang Y, Qin Y, Yang L, et al. Artificial platelets for efficient siRNA delivery to clear “Bad Cholesterol”. ACS Appl Mater Interfaces. 2020;12(25):28034–28046. doi:10.1021/acsami.0c07559

32. Zhang HT, Peng R, Tang W, et al. Versatile nano-PROTAC-induced epigenetic reader degradation for efficient lung cancer therapy. Adv Sci. 2022;9(29):2202039. doi:10.1002/advs.202202039

33. Chi SY, Zuo MM, Liu ZH, et al. Loading drugs in natural phospholipid bilayers of cell membrane shells to construct biomimetic nanocomposites for enhanced tumor therapy. ACS Appl Mater Interfaces. 2022;14(25):28671–28682. doi:10.1021/acsami.2c08587

34. Märkl S, Schroter A, Hirsch T. Small and bright water-protected upconversion nanoparticles with long-time stability in complex, aqueous media by phospholipid membrane coating. Nano Lett. 2020;20(12):8620–8625. doi:10.1021/acs.nanolett.0c03327

35. Zhang D, Lin ZG, Liu XL. Ultrasound-driven biomimetic nanosystem suppresses tumor growth and metastasis through sonodynamic therapy, CO therapy, and indoleamine 2,3-dioxygenase inhibition. ACS Nano. 2020;14(7):8985–8999. doi:10.1021/acsnano.0c03833

36. Zhang W, Gong CY, Gao J, Li M, Li Y, Gao J. Tumor microenvironment-activated cancer cell membrane-liposome hybrid nanoparticle-mediated synergistic metabolic therapy and chemotherapy for non-small cell lung cancer. J Nanobiotechnol. 2021;19(1):339. doi:10.1186/s12951-021-01085-y

The Board of Directors of easypaisa digital bank approved the financial statements for the half year ended June 30, 2025. The Bank reported a profit before tax of PKR 3.64 billion, compared to PKR 2.61 billion in the same period last year, reflecting a 39.4% increase.

This performance was driven by higher markup income and robust fee-based revenues from digital lending and payments, despite a decline in the SBP discount rate from 20% to 11% in line with easing inflation.

Net markup income rose by 15.6%, supported by growth in digital lending. At the same time, non-markup income surged by 60.5% on the back of increased transaction volumes, particularly in cash deposits and withdrawals, as well as higher earnings from load and bundle products, commissions on corporate disbursements and collections, and insurance products.

Operating expenses rose by 9.6%, reflecting continued investments in technology, talent, and customer acquisition costs to support future growth and partially offset by a reversal of accruals for compensation costs. Cost to income ratio improved from 80.5% last year to 66.9%.

The Bank’s digital ecosystem continued to strengthen, with monthly active users (MAUs) reaching 18.2 million. Customer deposits stood at PKR 94.7 billion, marking a 41.3% increase over June 2024, underpinned by strong customer confidence following easypaisa’s transition to a digital retail bank. The CASA ratio remained exceptional at 98.1%, with the cost of deposits among the lowest in the industry at 1.57%.

Total advances stood at PKR 27.7 billion, with a loan-to-deposit ratio of 25.0%. Non-performing loans (NPLs) were reported at 16.1%, with a healthy coverage ratio of 91.4%. The Bank’s equity was recorded at PKR 16.8 billion, while the Capital Adequacy Ratio (CAR) remained strong at 20.52%.

Jahanzeb Khan, President & CEO, easypaisa digital bank, stated, “Our robust profitability during the first half of 2025 signals our intent to innovate Pakistan’s digital banking space. We are committed to ensuring and empowering access to easy and convenient digital financial services.

With a range of innovative products and services and more in the pipeline, easypaisa is all set to lead Pakistan’s digital banking space through its customer-centric approach. I am thankful for the support of our partners, board members, industry stakeholders, and the State Bank of Pakistan in this journey of cashless economy.”

Commenting on the results, Amin Sukhiani, Chief Financial Officer, easypaisa digital bank, said, “easypaisa digital bank is well-positioned to accelerate its growth journey by expanding its product suite, including foreign exchange, Islamic products, credit cards, remittances, and Buy Now Pay Later offerings. We are also investing in merchant expansion under the government’s digital cashless initiative and strengthening the insurance marketplace to improve our ecosystem further and create value for our customers.”

With over 55 million registered users and as the country’s first digital bank to commence commercial operations, easypaisa remains aligned with the State Bank of Pakistan’s vision to drive inclusive economic growth. We are committed to expanding our range of financial products and services, not just for our current users, but also for the millions who remain unbanked or underbanked.

MUMBAI: India’s national investigative bureau opened a criminal case against tycoon Anil Ambani after receiving a complaint from the country’s biggest bank alleging fraud, the agency said Saturday.

Anil, the younger sibling of Asia’s richest man Mukesh Ambani, has business interests that range from power to defence.

The State Bank of India (SBI) alleged Anil Ambani and his former telecom firm Reliance Communications “misappropriated” bank funds by entering into transactions that were in violation of the terms of the loans.

SBI claims it was hit with a loss of INR 29.29 billion ($335.4 million) as a result of their actions.

The Central Bureau of Investigation said it had registered a case and that the bank’s complaint would be subjected to “thorough investigation”.

The agency searched premises linked to Reliance Communications and Anil Ambani’s residence on Saturday.

Ambani’s spokesperson said the tycoon “strongly denies all allegations and charges” and “will duly defend himself”.

“The complaint filed by State Bank of India (SBI) pertains to matters dating back more than 10 years. At the relevant time, Mr. Ambani was a Non-Executive Director of the company, with no involvement in the day-to-day management,” the spokesperson said.

“It is pertinent to note that SBI, by its own order, has already withdrawn proceedings against five other Non-Executive Directors. Despite this, Mr. Ambani has been selectively singled out.”

Anil Ambani was last public spotlight seven years ago after Indian politician Rahul Gandhi accused Indian Prime Minister Narendra Modi and him of dodgy dealings related to the purchase of Rafale jets from France – allegations that both denied.

India’s Supreme Court in December 2018 dismissed calls for an investigation into the controversial jet deal, saying it did “not find any substantial material on record to show that this is a case of commercial favouritism to any party by the Indian government”.

Introduction

The spleen is the largest secondary lymphoid organ, and an extensive meshwork of nerve fibers is distributed throughout the splenic compartments including the splenic nodules, periarteriolar lymphoid sheath, marginal zones, trabeculae, white pulp and red pulp.^1,2 Even more striking, the nerve fibers are co-localized with various subsets of dendritic cells (DCs), macrophages (Mφ), and lymphocytes (ie B cells and T cells).^2,3 As early as 30 years ago, some studies demonstrated that sympathetic nerve terminals in the spleen secrete norepinephrine (NE) interacting with T cells in response to stimulation, while the content of splenic norepinephrine significantly decreases following chemical or surgical sympathectomy.^4,5 It was not until around 2000 that the cholinergic anti-inflammatory pathway (CAP) concept was formally proposed and extensively studied because of its vital function in modulating the mammalian immune response.^6–10 The CAP mechanism relies on the parasympathetic vagus nerve transmitting signals from the brain to the adrenergic splenic nerve that interacts with immune cells (Supplementary Figure 1). Munyaka et al¹¹ revealed that CAP is activated after central cholinergic activation by intracerebroventricular infusion of the M1 muscarinic acetylcholine (ACh) receptor agonist in mice but is suppressed by vagotomy or splenic neurectomy. Moreover, Ji et al¹² proved that central cholinergic activation induced by the acetylcholinesterase inhibitor galantamine or the muscarinic ACh receptor agonist relieves colitis in mice, which is inhibited after vagotomy, splenic neurectomy or splenectomy. These reports indicate that spleen is a vital target of the CAP.

In the past decade, many researchers applied physical methods to trigger the splenic neural-immune reflex for regulating the immune response to chronic inflammatory diseases.¹³ Inoue et al¹⁴ utilized a bipolar silver wire electrode stimulating the isolated left cervical vagus nerve to attenuate kidney ischemia-reperfusion injury (IRI), suggesting that vagus nerve stimulation (VNS)-mediated IRI treatment and systemic inflammation depend on α7 nicotinic acetylcholine receptors (α7nAChR)-positive splenocytes. Koopman et al^15,16 reported that stimulation of the CAP by VNS to treat rheumatoid arthritis (RA) in experimental models and clinical trials exerts a significant mitigation effect, whilst the symptoms aggravated following unilateral cervical vagotomy or α7nAChR knockout in mice. Besides, Meregnani et al¹⁷ demonstrated that the symptoms of colonic colitis in rats induced by trinitrobenzene sulphonic acid was significantly reduced after VNS performed with electrode cuffs for 5 days. Moreover, electroacupuncture intervention eliciting splenic efferent vagus nerve activity not only reduced tumor proliferation in breast tumor-bearing mice by alleviating inflammation and enhancing antitumor immunity,¹⁸ but also could alleviate the severity of chemotherapy-induced nausea and vomiting in patients with advanced cancer.^19,20 In addition, bioelectronic devices for VNS in clinical application have already been approved by the European Medicines Agency and the US Food and Drug Administration (FDA) for the treatment of drug-refractory epilepsy and depression.^13,21,22

Unsatisfactorily, nerve stimulation using permanently implanted electrodes, transcutaneous bioelectronic devices, or slightly injurious electroacupuncture was limited to stimulate large nerves or superficial nerves.²³ Peripheral organs, not just the spleen, are extensively traversed by peripheral nervous system, which makes it difficult to selectively stimulate subsets of axons that terminate in spleen and uniquely modulate the immune response to inflammatory diseases.²³ For example, cervical nerve stimulation usually activates the extensive efferent and afferent neural pathways that elicit neural responses from non-target organs.^{13,16,24–26} Although scientists are committed to developing more advanced, miniaturized and sophisticated electrode designs for implantation on specific nerves near the target organ,^27,28 invasive implantation easily causes damage to the delicate nerve fibers of the target organ and it is difficult to stimulate nerves throughout the organ. Thus, new nonintrusive strategies are imperative to trigger specific nerves, especially the splenic nerve, and mobilize the splenic immune function against diseases.

Notably, recent studies showed that noninvasive ultrasound successfully elicits peripheral nerve activation in the spleen to treat chronic inflammatory diseases (Supplementary Figure 1). Gigliotti JC and co-workers²⁹ used an ultrasound imaging probe to successfully irradiate the neural innervation in the spleen mitigating IRI in mice, whose mechanism is mainly attributable to the CAP activation.^6,9,16,24 Later, Cotero et al²³ reported that ultrasound modulating splenic nerve reduces endotoxin-induced cytokine production at a level comparable with traditional VNS performed with electrodes. Zachs et al³⁰ applied ultrasound stimulation targeting the spleen, which significantly reduces inflammation severity in RA mouse models. These studies provide ample evidence on the potential of precise splenic ultrasound stimulation to replace bioelectronic devices for the translation of peripheral neuromodulation-based therapies for inflammatory disease. However, the potential roles and underlying mechanisms of splenic ultrasound stimulation in cancer management have been rarely reported and superficially defined. Upon validation of its significant antitumor efficacy, splenic ultrasound stimulation would represent a groundbreaking clinical advancement by circumventing multiple limitations of conventional cancer therapies: surgical resection-associated recurrence and visceral injury compromising quality of life, chemotherapy-induced systemic toxicity, off-target risks of immunotherapy, and the technical complexities of personalized gene therapy.³¹

Further investigation is needed to establish whether ultrasound directly regulates the immunomodulatory function of the spleen by interfering with splenic immune cells and microenvironment. Ultrasound targeting the spleen is crucial in achieving conspicuously therapeutic effects, since the stimulation of other body locations or the lack of immune cells in mice is noneffective, which suggests the indispensability and importance of immune cells.^29,30 Cotero et al²³ and Zachs et al³⁰ concluded that ultrasound activating CAP indirectly exerts an influence on splenic immune cells or cytokines to reduce the severity of lipopolysaccharide acute inflammation and RA, but they mainly focused on the analysis of neuro-immune mechanism and did not perform a systematic research on the changes of splenic immune cells in response to ultrasonic stimulation. Generally, ultrasonic irradiation on cells directly irritates the expression of multiple key genes/signaling pathways or interferes with the secretion of cytokines to regulate cell proliferation, differentiation and migration, such as stimulating the expression of vascular endothelial growth factor in several cells including endothelial cells, neural cells and ischemic cells.^32–35 Therefore, we hypothesized that ultrasound could effectively immunomodulated on the splenic immune cells responding to cancer cell antigens in splenic microenvironment, and enhance the proliferation, activation, migration, and information dissemination of immune cells during cancer immunotherapy.

Accordingly, this study was designed to explore whether focused ultrasound precisely stimulating spleen (FUS sti. spleen) was effective in suppressing tumor proliferation, and the underlying mechanisms regulating splenic ultrasound stimulation in cancer management, including the activation of splenic nerve-related CAP, and splenic immune cells responding directly to ultrasonic capabilities. The therapeutic efficacy of FUS sti. spleen was first assessed on various tumors under specific parameters screened before. Furthermore, the splenic immunomodulation through FUS sti. spleen was evaluated based on the changes of splenic immune cell population and cytokine levels. Subsequently, nerve blockade, immune cell clearance and RNA sequencing were performed to identify the underlying mechanisms of splenic ultrasound stimulation in cancer management, such as splenic nerve-related CAP activation and FUS directly modulating immune cells, and the related molecular mechanism.

Methods

Cell Lines and Animal Models

Hepatocellular carcinoma (HCC) cell lines (H22 and hepa1-6) and 4T1 breast cancer cell line were purchased from the American Type Culture Collection (ATCC) and used for subsequent analysis. H22 cells were reproduced by ascites after intraperitoneal injection in C57black/6 mice (0.5–1.5*10^7 cells/mL; 200 µL per mouse). Hepa1-6 cells and 4T1 cells were cultured in Dulbecco’s modified Eagle’s medium (Cat. No. 11320033, Gibco, Invitrogen, Carlsbag, USA) supplemented with 10% fetal bovine serum (FBS; Cat. No. A5669701, Gibco, Invitrogen, Carlsbag, USA) and 1% penicillin-streptomycin solution (Cat. No. 15140122, Gibco, Invitrogen, Carlsbag, USA), and incubated at 37 °C under 5% CO₂ and 100% humidity.

Seven-week (weighted ~17 g) C57black/6 mice, purchased from the Beijing Animal Experiment Center (Chinese Academy of Sciences, Beijing, China), were housed at 24±2 °C under a 12-h light/dark cycle and acclimatized for at least 1 week before the experiments. All animals had access to sterilized food and water ad libitum. An amount of 0.2 mL H22 cancer cells (approximately 1–3*10^6 cells/mL) or 4T1 cancer cells (1–3*10^6 cells/mL) was hypodermically injected into the dorsal hindlimb region of the mouse to establish a subcutaneous tumor model. Regarding the in situ xenograft HCC models, an incision of approximately 1 cm in length was made at the upper end of the midabdominal line after the mouse was anesthetized with isohalothane and fixed on the anatomic stage. A total of 25 µL H22 cancer cells (approximately 0.5–1.5*10^7 cells/mL) or 50 µL hepa1-6 cancer cells (approximately 4–6*10^6 cells/mL) was injected into left hepatic lobe with an insulin syringe, and the incision was closed after applying pressure on the pinhole with a medical cotton swab for 3–4 minutes.

The tumor volume was calculated as follows: volume = 0.5*L*W^2, where L and W were the length and width of the tumor, respectively, measured by a caliper. The weight of spleen, tumor and mice was assessed by electronic scales. The splenic index was calculated by the ratio of spleen weight to mouse weight.

Ultrasonic Platform Set-up and Application for Spleen Stimulation

A diagram of the FUS system is shown in Supplementary Figure 2A. A function generator (Cat. No. 33120A, Agilent, Santa Clara, USA) produced a pulsed sinusoidal waveform triggering the power amplifier (Cat. No. AG1019, California, USA; or RPR-4000, RITEC Inc., Warwick, RI, USA) to drive a 1.04 MHz FUS transducer with a 100 mm aperture and 65 mm focus, whose focal region was ~1.4*1.4*8.6 mm³ (Supplementary Figure 2B and C). The acoustic pressure and spatial beam profile of the FUS transducer were measured using a hydrophone (Cat. No. HNR-1000, ONDA, Videlles, France). The FUS transducer was mounted on a XYZ motorized positioning stage to control ultrasonic duration, and the position of the FUS focus relative to the mice was adjusted under a B-mode imaging guidance with a 3.2 MHz phased array positioned at the center of the FUS transducer.

The mice were anesthetized with 2–3% inhaled isoflurane and placed on a manual translation stage equipped with heating pad. Then, a centrifuged coupling gel was immediately applied to the shaved skin, and the manual translation stage was adjusted to allow the spleen (or tumor) site to be in tight contact with the bottom of the water tank under the B-mode imaging guidance (Supplementary Figure 2D). The FUS transducer was moved across the whole spleen (or tumor) through the XYZ programmable logical controller (Supplementary Figure 2E).

Flow Cytometry

Single-cell suspensions from the spleen, peripheral blood, tumor, and para-carcinoma tissue were obtained after red blood cells were lysed using lysis buffer (Cat. No. 555899, BD Bioscience, USA) according to the manufacturer’s instructions. Cell suspensions were washed with sterilized PBS, and then incubated with the antibodies listed in Supplementary Table 1, and the scheme of antibody labeling for each immune cell is shown in Supplementary Table 2. Next, the immune cells were counted by flow cytometry (FCM) using CytoFLEX LX (Beckman Coulter Life Sciences, USA) after filtration through a 70-μm nylon cell strainer (Cat. No. 352350, Corning, USA). Data were analyzed using the FlowJo software (FlowJo 10, LLC, Ashland, OR).

Cytokine Measurement

Luminex Assay

Cytokine/chemokine quantification in plasma was performed by Luminex xMAP technology using a magnetic Luminex assay (R&D Systems, Minneapolis, MN, USA). The quantification was carried out using a Luminex^® 200 Flow Cytometry System (Cat. No. x-200, Thermo Fisher Scientific, MA, USA) and Milliplex Analyst software (Version 5.1, Merck Millipore, MA, USA).

ELISA

The concentration of NE, ACh, granzyme and perforin in plasma, respectively, detected by enzyme-linked immunosorbent assay (ELISA). The operation processes were carried out in strict accordance with the instruction manual from Mouse Perforin ELISA Kit (Cat. No. F30718-A, FANKEW, Shanghai, China), Mouse Granzyme-B ELISA Kit (Cat. No. F3214-A, FANKEW, Shanghai, China), Mouse ACh ELISA Kit (Cat. No. MU30072, Bioswamp, Wuhan, China) and Mouse NE ELISA Kit (Cat. No. MU30372, Bioswamp, Wuhan, China), respectively.

Histological and Immunohistochemical Staining

Hematoxylin eosin (HE) staining was applied to evaluate the pathological changes of the spleen after FUS irradiation. Transferase-mediated deoxyuridine triphosphatebiotin nick end labeling (TUNEL) staining was performed to observe splenic cell apoptosis through the colorimetric TUNEL Apoptosis Assay Kit (Cat. No. C1091, Beyotime, Shanghai, China). The primary antibodies of anti-Ki67 (Cat. No. RM9106S1, Thermo Fisher Scientific, MA, USA; dilution 1:200), anti-NK1.1 (Cat. No. 108759, BioLegend, California, USA), anti-CD8a (Cat. No. ab4055, Abcam, Cambridge, UK; dilution 1:1000), and anti-c-Fos (Cat. No. ab222699, Abcam, Cambridge, UK; dilution 1:2000) were used for immunohistochemical staining to detect the protein expression of cyclin-D1, Ki67, NK1.1, CD8a, and c-Fos respectively. At least 5–10 different regions in each section were randomly selected for image acquisition using a fluorescence microscopy (observer3, Carl Zeiss, Jena, Germany), and the positive area ratio was quantified using the Image Pro Plus 6.0 software (Media Cybernetics, CA, United States).

RNA Sequencing

Approximately 1*10⁷ CD8 T cells (CD45⁺ CD3⁺ CD8a⁺) and 5*10⁶ NK cells (CD45⁺ CD3⁻ NK1.1⁺) were sorted from the spleen using the Beckman Kurt MoFlo Astrios ultra high-speed flow cytometry sorting system (MoFlo Astrios EQ, Beckman Coulter, lnc, USA). Subsequently, general transcriptome sequencing was performed when the positive cell rate was greater than 90% detected by flow cytometry (CytoFLEX LX, Beckman Coulter Life Sciences, USA). The cDNA library construction and sequencing of all RNA samples were performed by Shanghai OE Biotech Co., LTD. Transcriptome sequencing was performed to screen differentially expressed genes (DEGs). Then, gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were applied to annotate the function of the DEGs. The DEG screening criteria were as follows: P < 0.05 and differential fold change |logFC| ≥ 1.

Nerve Signal Blockade

Propranolol Hydrochloride Blocking CAP

Propranolol hydrochloride (Cat. No. abs44129405, Absin, Shanghai, China), a kind of adrenergic receptor antagonist, was intraperitoneally administered at a dose of 3 mg/kg to block CAP by directly binding to the adrenergic receptor at 1.5 hours before FUS exposure to spleen according to previous reports,^36–38 which ensured FUS sti. spleen at the optimal time point of CAP blockade.

Absolute Ethanol-Induced Splenic Nerve Block

Absolute ethanol was used for splenic nerve block according to previous studies.^39–44 Briefly, the mice were anesthetized with the isoflurane, and disinfection with iodophor was performed on the shaved skin. A 7–10 mm incision was made next to the left side of the spleen. The spleen was isolated with tweezers to keep it away from the peritoneal cavity for a clear exposure of the three main blood supply vascular trees. A wet cotton was used to protect the peritoneal cavity and other organs. Subsequently, absolute ethanol was repeatedly applied to those vascular trees with a cotton tip for 5–10 seconds each time, with an interval of five seconds, for a total of seven times to block the splenic nerve fibers. Excessive ethanol dripping from the cotton tip was carefully avoided, to avoid significant vasospasm, which could lead to permanent damage to blood vessels, consequently leading to spleen necrosis and organ complete absorption. As regards the sham-operated mice, the same entire surgical procedure was performed except for the repeated application of anhydrous ethanol that was replaced by normal saline (pH = 7.4).

In vivo Immune Cell Depletion

The experimental C57BL/6 mice were treated with an intravenous injection of Ultra-LEAF™ purified anti-mouse NK1.1 antibody (Cat. No. 108759, BioLegend, California, USA) or Ultra-LEAF™ purified anti-mouse CD8a antibody (Cat. No. 100763, BioLegend, California, USA) at a dose of 50 μg/mouse on day 0, 3, 7, 12 and 17 to deplete NK cell or CD8 T cell. As regards MDSC depletion, the Ultra-LEAF™ Purified anti-mouse Ly-6G antibody (Cat. No. 127650; BioLegend, California, USA) was intraperitoneally administered to C57BL/6 mice at a dose of 0.25 mg/mouse on day 0, 6, 12, and 18. The cell clearance efficacy was assessed by FCM.

Calcium Detection and in vitro Anticancer Experiment of FUS Stimulating Splenic CD8 T Cells or NK Cells Enhanced by Calcium

Von Kossa Staining

The splenic specimens were cut into 5 μm-thick sections and subjected to VON KOSSA Calcium Staining Kit (Cat. No. JM1519; HPBIO, Shanghai, China) for the histological visualization of calcium deposits (mineralization). Calcium deposition in splenic cells was assessed by fluorescence microscopy (Zeiss observer3), where mass deposits appeared black, while dispersed deposits appeared gray. The positive area ratio was quantified using the Image Pro Plus 6.0 software.

FCM Detection of Fluo-4 AM Labeled Cells

The splenic cell suspensions were washed twice with sterilized PBS, and then incubated with Fluo-4 AM dye (1 mL dye/10⁶ cells; Cat. No. S1061M, Beyotime, Shanghai, China) for 30 min at 37 °C according to the manufacturer’s instructions. Next, the fluorescent cells were monitored by FCM using the CytoFLEX LX (Ex/Em = 490/525 nm).

In vitro Experiments Verifying Calcium-Strengthened FUS Stimulating Splenic NK Cells Against Tumor

Approximately 1.2*10⁶ NK cells were sorted from the spleen using the Beckman Kurt MoFlo Astrios ultra high-speed flow cytometry sorting system, and then subjected to FUS stimulation with or without high calcium concentration (300 nM). Then, NK cells were co-cultured with pEGFP-C1 plasmid-transfected Hepa1-6 cancer cells at a ratio of 1:1, where GFP fluorescent protein was used to distinguish cancer cells from splenic cells. Transwell devices (Corning Incorporated, NY, USA) with 3-µm diameter holes were used to distinguish immune cells in the superstratum and cancer cells in the substratum. The cells were co-cultured for 48 h before GFP fluorescence observation with the Zeiss observer3, CCK-8 assay (Cat. No. ABS50003-500T, Univ, China), and crystal violet (CV) staining to assess cancer cell suppression and immune cell proliferation. The tumoral background was mimicked by adding ultrasound shattered cancer cell suspension to the NK cells solution during FUS stimulation.

Statistical Analysis

Statistical analysis was performed using the statistical product and service solutions software (SPSS, USA). Statistical significance was analyzed using Student’s t-test for parametric data and Mann–Whitney U-test for nonparametric comparisons, with Bonferroni correction (specially for multiple comparisons), and the results were expressed as mean ± SEM. A value of p < 0.05 was considered statistically significant (*p < 0.05, **p < 0.01, and ***p < 0.001).

Results

Screening and Optimization of Ultrasonic Parameters for Spleen Stimulation

In view of the different ultrasonic parameters and platforms applied in previous studies of splenic ultrasound stimulation, it is pseudoscientific to indiscriminately imitate their ultrasonic parameters or methods to this study for regulating splenic immunotherapy to anti-cancer. Based on the pre-experiment results, 11 processing groups were set-up: G1-G11 group (detailed in Table 1), for the screening and optimization of ultrasonic parameters, and FUS sti. spleen in subsequent experiments was performed once every other day starting from the 2nd day after cancer cell implantation, with a total experimental duration of 28 days (Figure 1A). At the experimental endpoint (day 29), animals underwent terminal anesthesia followed by comprehensive tissue harvesting for other experimental analyses (eg, HE and TUNEL staining). The degree of subcutaneous H22 tumor suppression and spleen injury (assessed by HE and TUNEL staining) was used as the decisive criterion to establish the appropriate ultrasonic parameters.

Table 1 Scheme of Different Ultrasonic Parameters for FUS Sti. Spleen

Figure 1 Inhibitory effect of FUS sti. spleen on subcutaneous H22 tumor under various ultrasonic parameters. (A) experimental flow diagram. (B) images of spleen and tumor in G1-G11 groups. (C) tumor growth curves. (D) tumor weight. (E) mice weight. (F) spleen weight. (G) spleen index. (H) HE and TUNEL staining of the spleen irradiated with FUS under different ultrasonic parameters; the red arrow indicates the area of the magnified image, and the green arrow indicates the TUNEL-positive cells. (I) analysis of TUNEL positive area. (n = 5; * p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; mean ± SEM). G1-G11 group details are shown in Table 1.

The phenotypic results showed that FUS sti. spleen significantly suppressed the proliferation of subcutaneous H22 tumor, especially in the G4, G5, G8 and G11 group (Figure 1B–D). The weight of the mice subjected to FUS sti. spleen was lower than that of the non-FUS mice (Figure 1E), and FUS sti. spleen showed a tendency to decrease splenomegaly (Figure 1F and G). By comparison, the increase of the duty cycle or the FUS exposure time at 1.8 MPa did not improve the anti-tumor effect (G3, G6, G9 and G10 group; Figure 1B–D). However, the high peak negative pressure of 2.8 MPa induced a severe spleen injury, easily leading to the death of mice (G5 group; Figure 1B, H and I). In contrast, FUS sti. spleen effectively suppressed tumor proliferation without damaging the splenic cells at 2.3 MPa (Figure 1B–D, H and I). However, the antitumor effect was significantly reduced when the FUS exposure time of 20s was shortened to 10s (G4 and G7 group; Figure 1B–D). FUS damaged a few splenic cells once increasing the exposure time of 20s to 40s or the duty cycle of 1% to 10% (G4, G8 and G11 group; Figure 1B–D, H–I). The above splenic stimulation results demonstrated a clear “dose response” to ultrasound with a distinct power level required for the specific regulation of biophysical effects on splenic immune cells, which was similar to previous studies.^23,29,30 In conclusion, the ultrasonic parameters of the G4 group (2.3 MPa, 1%, 20s) were selected for subsequent experiments of splenic immunomodulation for cancer therapy.

FUS Universally Activated Spleen Immunological Function to Suppress the Proliferation of Various Tumors

FUS Activated Splenic Immune Cells to Suppress the Proliferation of the Xenograft HCC in situ

The mouse model of HCC in situ, which was better to resemble the situation of primary liver cancer, was used to assess the anticancer efficiency of FUS sti. spleen. The orthotopic HCC mice were subjected to FUS sti. spleen from day 0 after cancer cell implantation, and once every two days for a total experimental duration of 20 days (Figure 2A). At the experimental endpoint (day 21), animals underwent terminal anesthesia followed by comprehensive tissue harvesting for other experimental analyses (eg, FCM).

Figure 2 FUS sti. spleen suppressed the proliferation of the xenograft HCC in situ. (A) experimental flow diagram of FUS sti. spleen after HCC cancer cell implantation. (B–I) results of FUS sti. spleen suppressing the proliferation of the xenograft H22 HCC in situ. (B) images of spleen and xenograft H22 HCC in situ. (C and D) tumor volume and tumor weight, respectively. (E–G) liver weight, mice weight and spleen weight, respectively. (H) spleen index calculated by the ratio of spleen weight to mouse weight. (I) survival curves. S1, control group of xenograft H22 HCC in situ; S2, FUS sti. spleen started on day 0 after H22 cancer cell implantation. (J–Q) results of FUS sti. spleen suppressing the proliferation of the xenograft Hepa1-6 HCC in situ. (J) images of spleen and xenograft Hepa1-6 HCC in situ. (K and L), tumor volume and tumor weight, respectively. (M–O, liver weight, mice weight and spleen weight, respectively. (P) spleen index calculated by the ratio of spleen weight to mouse weight. (Q) survival curves. (n = 10; * p ≤ 0.05; ***p ≤ 0.001; mean ± SEM). L1, control group of xenograft Hepa1-6 HCC in situ; L2, FUS sti. spleen started on day 0 after Hepa1-6 cancer cell implantation.

Twenty orthotopic H22 HCC mice were randomly divided into two groups: the S1 group was used as the control group, and the S2 group was subjected to FUS sti. spleen. The results of visual observation, tumor volume and tumor weight demonstrated that FUS sti. spleen significantly suppressed the proliferation of xenograft H22 carcinoma in situ, with a tumor inhibition rate of up to ~70% as compared with the control group (S2 vs S1, p<0.001; Figure 2B–D). Although the mean value of liver weight (total weight of the tumor and liver parenchyma) showed a decreasing trend in the S2 group compared to the S1 group, no significant difference was observed between the two groups (Figure 2E). In addition, the weight of the mice in the S2 group was lower than that of the mice in the S1 group (p<0.001; Figure 2F). The spleen weight and spleen index showed a decreasing trend of splenomegaly in the S2 group as compared with the S1 group, but without statistically significant difference (Figure 2G and H). More importantly, the survival statistics proved that the orthotopic H22 HCC mice treated with FUS sti. spleen had a longer survival time than the control group (Figure 2I).

Sixteen orthotopic Hepa1-6 HCC mice were randomly divided into two groups: L1 group was performed as the control group, and L2 group was administered to FUS sti. spleen. The results of visual observation, tumor volume and tumor weight showed that FUS sti. spleen remarkablely inhibited the proliferation of xenograft Hepa1-6 carcinoma in situ, with an antitumor ratio of up to ~80% (L2 vs L1, p<0.001; Figure 2J and L). Additionally, the weight of the liver and mice in the L2 group was lower than that of the mice in the L1 group (p<0.05; Figure 2M and N). No significant difference in spleen weight and spleen index was observed, although there was a decreasing trend shown in the L2 group as compared with the L1 group (Figure 2O and P). It was noteworthy that the survival time of the mice in the L2 group was significantly longer than that of the mice in the L1 group (Figure 2Q).

Thereafter, FCM results revealed that the proportion of tumor suppressor related immune cells in the spleen, blood, tumor and para-carcinoma tissue was significantly increased in the S2 group as compared with the S1 group, such as Th2 cells, NK cells, CD8 T cells, Mφ (Mφ2) and DC1 in the spleen; B cells, NK cells, CD8 T cells, Mφ1 and DC1 in the blood; CD4 T cells, NK cells, CD8 T cells, Mφ (Mφ1) and DC1 in the tumors; Th1 cells, NK cells, CD8 T cells and Mφ (Mφ1) in the para-carcinoma tissue (Table 2, Supplementary Figures 3–6). Moreover, the proportion of negative immune cells was significantly reduced, such as Treg cells and MDSCs (M-MDSCs and PMN-MDSCs) in both spleen and tumor; Th17, PMN-MDSCs in the blood; and PMN-MDSCs in the para-carcinoma tissue (Table 2, Supplementary Figures 3–6).

Table 2. Statistical Significance of FCM Results of Immune Cells in the Spleen, Blood, Tumor and Para-Carcinoma Tissue From Orthotopic HCC Mice Subjected to FUS Sti. Spleen or Non-Treated Mice

As regards to orthotopic Hepa1-6 HCC mice, FCM results showed that the number of positive anticancer immune cells in the spleen, blood, tumor and para-carcinoma tissue was significantly increased in the L2 group than in the L1 group, such as Th2 cells, NK cells, CD8 T cells, Mφ and DCs in the spleen; CD4 T cells, Th2 cells, NK cells, CD8 T cells, Mφ1, Mφ2 and DCs in the blood; NK cells, CD8 T cells, Mφ and DCs in the tumor; and NK cells, CD8 T cells and Mφ (Mφ1 and Mφ2) in the para-carcinoma tissue (Table 2, Supplementary Figures 7–10). The number of negative immune cells was significantly reduced, such as Treg cells and PMN-MDSCs in the spleen; Th17 and MDSCs in the blood; Th17 and PMN-MDSCs in the tumor; and PMN-MDSC in the para-carcinoma tissue (Table 2, Supplementary Figures 7–10).

These results fully demonstrated the significant effect of FUS on splenic immunomodulation for cancer immunotherapy. Particularly, FUS sti. spleen remarkably increased the proportion of NK cells and CD8 T cells in the spleen, blood, tumor and para-carcinoma tissue, followed by Mφ and DCs, whilst it reduced the number of PMN-MDSCs in all tissues with one accord, and then Treg cells.

Additionally, we assessed the therapeutic efficacy of FUS sti. spleen on subcutaneous 4T1 breast tumor, and further quantitatively evaluated the alterations in immune cell proportions following FUS intervention. Obviously, FUS sti. spleen demonstrated potent antitumor efficacy, achieving 70% tumor growth inhibition and significantly prolonging survival. FCM revealed that FUS elicited significant immunomodulatory effects, particularly enhancing NK cell and CD8+ T cell populations while suppressing immunosuppressive cells. The complete results are available in Supplementary File 1.

FUS Sti. Spleen Altered the Cytokine Levels

In addition to the modulation of cellular immunity by FUS sti. spleen, humoral immunity may also be significantly regulated, which represents one dominant factor in the anticancer process during FUS irradiation. Herein, we selectively detected the cytokine levels of GM-CSF, TNF-alpha, IL-12, CCL2, IL-1beta, IL-2, IL-4, IL-6, IL-10, IL-13, IL-17, IFN-gamma, CXCL10, M-CSF, IL-1alpha, CCL4, CXCL12, IL-27, perforin and granzyme in the peripheral blood using the Luminex xMAP technology. Interestingly, the concentration levels of TNF-alpha, IFN-γ, perforin and granzyme were significantly increased in the FUS sti. spleen group as compared with the control group in both H22 and Hepa1-6 HCC in situ models (Figure 3A–C), which indicated the activation of cytotoxic CD8 T cells and NK cells against tumor. Usually, the naive CD8 T cells in the immune process undergo activation and clone expansion, which in turn produce the effector cytokine TNF-alpha, IFN-gamma, perforin and granzyme.^45,46 In addition, the secretion level of IFN-gamma, perforin and granzyme specifically indicated the activation of NK cells to fight cancer.⁴⁷

Figure 3 Cytokine detection and bioinformatics analysis. (A and B), heatmap of cytokine levels in orthotopic H22 and Hepa1-6 HCC mice subjected to FUS sti. spleen. (n = 5; * p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; mean ± SEM). (C) Venn diagram of up- and down-regulated cytokines in both orthotopic H22 and Hepa1-6 HCC mice subjected to FUS sti. spleen. (D) protein-protein network analysis of cytokines and target genes based on STRING database. (E) KEGG pathway enrichment analysis of cytokines and target genes.

The concentration levels of the pro-inflammatory chemokines CCL2, CCL4 and CXCL10 were significantly elevated in the FUS sti. spleen group of both orthotopic H22 and Hepa1-6 HCC models compared with those in the control group (Figure 3A–C). These pro-inflammatory chemokines promote immune cells especially effector T cells successfully migrating into metastatic tumor sites to perform an anti-tumor activity.^48–50 CXCL12 is a homeostatic chemokine involved in physiological processes like embryogenesis, neurogenesis, cardiogenesis, hematopoiesis, leukocyte homing, and angiogenesis, by inducing the migration and activation of hematopoietic progenitor and stem cells, endothelial cells, and most leukocytes.^51,52 In this study, the CXCL12 level was significantly increased in the FUS sti. spleen group (Figure 3A–C). As a final level of activity regulation, CXCL12 synergizes with CXCL8 but also with CXCL9, CXCL10, CXCL11, and multiple CC chemokines to attract B- and T-lymphocytes, DCs, monocytes, and CD34⁺ progenitor cells.⁵³ Moreover, CXCL12 binding to its receptor CXCR4 mediates the biological behavior of tumor cells such as movement, chemotaxis, adhesion, secretion (such as MMPs and VEGF), angiogenesis, growth, and proliferation by activating various signaling pathways.^54,55

Besides, GM-CSF and M-CSF are 2 important cytokines that regulate macrophage population and function. Many experimental loads of evidence suggest that the aberrant expression of GM-CSF and its receptors are present in many cancers such as glioblastoma, small cell carcinoma, skin carcinoma, meningiomas, colon cancer, lung cancer, and particularly HCC, promoting cancer progression by regulating the tumor microenvironment involving Mφ and MDSCs, and promoting epithelial to mesenchymal transition, angiogenesis, and expression of immune checkpoint molecules.^56–59 The concentration level of GM-CSF in this study was significantly decreased as both orthotopic H22 and Hepa1-6 HCC mouse models underwent FUS sti. spleen (Figure 3A–C). In addition, the M-CSF level was increased after orthotopic HCC mice were subjected to FUS sti. spleen (Figure 3A–C), which regulated the survival, proliferation and differentiation of the monocyte-macrophage lineage from progenitors to mature cells and activated several important functions of mature tissular Mφ.^60,61

The proinflammatory cytokine IL-1beta showed an association with tumor progression,⁶² but interestingly, although the inhibition of IL-1beta signaling has potent anti-metastatic effects, the inhibition of the activity of this cytokine has contradictory effects on primary tumors, sometimes reducing but often promoting their growth.⁶³ Moreover, anti-cancer treatments are able to promote IL-1beta production by cancer or immune cells, with opposite effects on cancer progression,⁶⁴ which resembles the tumor-suppressing effect of FUS sti. spleen. However, beyond that, the cytokine levels of IL-1alpha, IL-17 and IL-27 showed no significant differences in consistency as orthotopic H22 and Hepa1-6 HCC mice subjected to FUS sti. spleen (Figure 3A–C). In addition, IL-2 is a pleiotropic cytokine required for both effector lymphocyte proliferation/differentiation and regulatory T cell expansion/survival,⁶⁵ which was significantly secreted as spleen responded to FUS stimulation (Figure 3A–C). Additionally, IL-4 mainly secreted by T cells and NK cells induces Mφ2 activation, and regulates the expression of IL-6, IL-10, IL-12 and IL-13.⁶⁶ Some studies also reported that polarized Th1 cells produced IL-2, IL-12 and IFN-gamma, and polarized Th2 cells secreted IL-4, IL-6, IL-10 and IL-13 that promoted tumor proliferation⁶⁷ (Figure 3A–C). In this study, IL-6 level was increased and IL-13 level was decreased when the orthotopic HCC mice were subjected to FUS sti. spleen (Figure 3A–C). Nevertheless, the cytokine levels of IL-4, IL-10 and IL-12 were altered between orthotopic H22 and Hepa1-6 HCC mice models subjected to FUS sti. spleen (Figure 3A–C).

Protein–protein networks of these cytokines and target genes, and KEGG pathway enrichment analysis showed that FUS stimulating splenic immune cells to regulate cytokines against tumor was associated with several signaling pathways, especially JAK-STAT signaling pathway and PI3K-AKT signaling pathway (Figure 3D and E). It was evident that these cytokines characterized the tumor cytotoxicity of NK cells and CD8 T cells, mainly through cytokine-cytokine receptor interaction pathway to obtain an anticancer effect (Figure 3D and E).

FUS Directly Activated Splenic Immune Cells for Cancer Immunotherapy

FUS Promoted Splenic Positive Immune Cell Proliferation and Activation to Combat Tumor

According to previous studies, splenic ultrasound stimulation activated CAP to inhibit arthritis, colitis, hyperglycemia and other inflammatory diseases, but they paid more attention to the modulation of inflammatory factors such as TNF-α, IL-6 and IL-10, as well as CAP validation, but not to the changes of immune cells.^23,68,69 In this study, FUS sti spleen significantly altered the proportion of immune cells in the spleen, blood, and tumor, especially NK cells and CD8 T cells that increased uniformly in various tumor models (Table 2, Supplementary File 1), thus arousing our special attention. Inexplicably, the previous studies based on inflammatory disease models (eg, pneumonia, myocarditis, and colitis) proved that splenic ultrasound stimulation activates CAP to reduce Mφ and inhibits the secretion of pro-inflammatory cytokines (eg TNF-α), but in this study based on tumor models, FUS sti. spleen significantly increased Mφ and promoted TNF-α secretion. Therefore, our speculation was that FUS directly stimulated splenic immune cells stressing tumor signals, which promoted the proliferation and activation of positive immune cells (ie NK cells and CD8⁺ T cells) to suppress cancer cell proliferation. To further confirm this hypothesis, splenic nerve blockade was performed with propranolol hydrochloride targeting CAP to explore the impact of CAP deficiency on tumor suppression during FUS sti. spleen. Subsequently, NK cells and CD8 T cells were depleted with Ultra-LEAF™ purified anti-mouse NK-1.1 antibody and Ultra-LEAF™ purified anti-mouse CD8a antibody, respectively, to further explore their role and importance in the tumor-suppressing process of FUS sti. spleen. Forty-eight orthotopic H22 HCC mice were randomly divided into eight groups: the Q1 group was used as the control group; the Q2 group was subjected to FUS sti. spleen; the Q3 group was treated with an intraperitoneal injection of propranolol hydrochloride to block CAP; the Q4 group was subjected to FUS sti. spleen after an intraperitoneal injection of propranolol hydrochloride; the Q5 group was subjected to NK cell clearance; the Q6 group was subjected to FUS sti. spleen after NK cell depletion; the Q7 group was subjected to CD8 T cell clearance; the Q8 group was subjected to FUS sti. spleen after CD8 T cell depletion. The operation of FUS sti. spleen and experimental timeline were identical to those described in Figure 2A.

The therapeutic results once again demonstrated the significant tumor suppression of FUS sti. spleen (Q2 vs Q1, p < 0.001; Figure 4A–C). It is noteworthy that specifically CAP blockage did not significantly attenuate the therapeutic efficacy of FUS sti. spleen on the tumor (Q4 vs Q3, p < 0.001; Figure 4A–C), although the anti-tumor rate was slightly reduced in the Q4 group as compared to the Q2 group. It confirmed our previous speculation that CAP did not play a dominant role in the process of FUS sti. spleen for tumor suppression. We supposed that cancer is different from inflammatory diseases such as arthritis, colitis, and pneumonia, and FUS sti spleen to inhibit inflammatory diseases mainly relied on CAP affecting humoral immunity, while it suppressed tumor proliferation mainly depending on the direct modulation of cellular immunity.

Figure 4 Therapeutic effect of FUS sti. spleen on xenograft H22 carcinoma in situ after CAP blockage, NK cells depletion and CD8 T cell clearance. (A) images of spleen and orthotopic H22 tumor. (B and C) tumor volume and tumor weight, respectively. (D–F) liver weight, mice weight and spleen weight, respectively. G, splenic index calculated by the ratio of spleen weight to mouse weight. (H–K) concentration of NE, ACh, granzyme and perforin, respectively, detected by ELISA. (n = 6; * p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; mean ± SEM). Q1, control group of xenograft H22 HCC in situ; Q2, FUS sti. spleen on day 2 after H22 cancer cell implantation; Q3, injection of propranolol hydrochloride to block CAP; Q4, FUS sti. spleen after injection of propranolol hydrochloride; Q5, NK cell clearance; Q6, FUS sti. spleen after NK cell depletion; Q7, CD8 T cell clearance; Q8, FUS sti. spleen after CD8 T cell depletion.

Furthermore, the tumor proliferation was significantly promoted after NK cell depletion (Q5 vs Q1, p < 0.001; Figure 4A–C). Interestingly, FUS sti. spleen did not show any tumor suppressive effect after depleting NK cells (Q6 vs Q5; Figure 4A–C). Moreover, CD8 T cell clearance promoted tumor proliferation (Q7 vs Q1, p<0.05; Figure 4A–C), but as opposite as NK cell clearance, did not reduce the therapeutic efficacy of FUS sti. spleen on tumor (Q8 vs Q2; Figure 4A–C). These results proved that NK cells were much more indispensable and important in cancer suppression in response to FUS stimulation than CD8 T cells. The statistical results of liver weight, mice weight, spleen weight and spleen index also showed that NK cell depletion had a significant effect on tumor proliferation and splenomegaly, and elucidated the attenuation of FUS sti. spleen on tumor suppression and splenomegaly (Figure 4D–G).

The concentration of NE and ACh in the plasma was significantly increased after FUS sti. spleen (Q2 vs Q1, Q6 vs Q5, and Q8 vs Q7, p < 0.05; Figure 4H and I), which indicated that FUS sti. spleen indeed activated splenic CAP. In addition, propranolol hydrochloride significantly blocked the binding of NE to adrenergic receptors, that is, effectively blocking CAP, thus increasing the NE content (Q3 vs Q1, p<0.01; Figure 4H) but no significant changes in ACh (Q3 vs Q1; Figure 4I). Therefore, NE content increased after FUS sti. spleen (Q4 vs Q3; Figure 4H), while ACh did not significantly change (Q4 vs Q3; Figure 4I). Nevertheless, it is important to note that the results of granzyme and perforin detection by ELISA further confirmed the significance of splenic NK cells activation and their secreting cytotoxic molecules in response to FUS stimulation for an anti-cancer effect, but not affected by CAP (Figure 4J and K).

FCM results proved that Ultra-LEAF™ purified anti-mouse NK-1.1 antibody and Ultra-LEAF™ purified anti-mouse CD8a antibody effectively eliminated NK cells and CD8 T cells in the body, respectively (Q5/6 and Q7/8 vs Q1/2; Tables 3–6, Supplementary Figures 11–14). In the case of CAP blockage and NK or CD8 T cell clearance, the proportion of Mφ, DCs and MDSCs in the body was also affected, especially the number of Mφ and DCs that was generally reduced after depleting NK cells and CD8 T cells (Q3/4, Q5/6 and Q7/8 vs Q1/2; Tables 3–6, Supplementary Figures 11–14). However, FUS sti. spleen consistently increased the number of NK cells and CD8 T cells in the body regardless of CAP blockade and NK or CD8 T cell clearance, followed by Mφ1 and DC1 (Q4 vs Q3, Q6 vs Q5, and Q8 vs Q7; Tables 3–6, Supplementary Figures 11–14), while the variation tendency of other immune cells (including PMN-MDSCs) was not consistent in the spleen, blood, tumor and para-carcinoma tissue (Tables 3–6, Supplementary Figures 11–14). Clearly, changes in the proportion of immune cells (especially NK cells) with or without modulation by splenic ultrasound stimulation were closely related to the anti-tumor effect.

Table 3. FCM Results of Immune Cells in the Spleen

Table 4. FCM Results of Immune Cells in the Blood

Table 5. FCM Results of Immune Cells in the Tumor

Table 6. FCM Results of Immune Cells in Para-Carcinoma Tissue

The results of the above CAP blockage experiments demonstrated that CAP did not play a dominant role in tumor suppression by FUS sti. spleen; however, it could not exclude the impact of splenic nerve or other nerve signals. Thereafter, we applied absolute ethanol to denervate the splenic nerve to explore the role of the splenic nerve on spleen immunomodulation and anti-cancer effect by FUS sti. spleen. Moreover, the spleen was removed to exclude the possibility of immune regulation and tumor suppression by FUS irradiation on non-splenic organs or tissues. Thus, forty-two orthotopic H22 HCC mice were randomly divided into six groups: H1 group was used as the control group; H2 group was subjected to FUS sti. spleen; H3 group was subjected to splenic nerve denervation by absolute ethanol; H4 group was subjected to FUS sti. spleen after splenic nerve denervation; H5 group was subjected to the removal of the spleen; and H6 group was exposed to FUS on spleen position after splenectomy.

The results showed that splenic nerve denervation promoted tumor proliferation to some extent, although not significant (H3 vs H1, p = 0.3430 in tumor volume and p = 0.1590 in tumor weight), suggesting that the splenic nerve had a certain role in the anti-tumor effect. However, more importantly, splenic nerve denervation did not attenuate the tumor suppression efficacy of FUS sti. spleen, which was up to ~75% (H4 vs H3, p<0.001; Figure 5A–C) that was superior to or consistent with the tumor inhibition rate of FUS sti. spleen without splenic nerve denervation (~70%, H2 vs H1, p<0.001; Figure 5A–C). It further ruled out the indispensable role of splenic nerve on tumor suppression by FUS sti. spleen, which was also direct evidence that FUS directly stimulated splenic immune cells to regulate immunity against cancer.

Figure 5 Therapeutic effect of FUS sti. spleen on tumor suppression after splenic nerve denervation and splenectomy. (A) images of spleen and orthotopic H22 tumor. (B and C) tumor volume and tumor weight, respectively. (D–F) liver weight, mice weight and spleen weight respectively. (G) splenic index calculated by the ratio of spleen weight to mouse weight. (n = 7; * p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; mean ± SEM). (H and I) concentration of NE and ACh, respectively, detected by ELISA. (J) c-Fos immunohistochemical staining and statistical result of positive area in section; the red arrows indicate the c-Fos-positive cells. (K and L) concentration of granzyme and perforin, respectively, detected by ELISA. (n = 5; * p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; mean ± SEM). H1, control group of xenograft H22 HCC in situ; H2, FUS sti. spleen on day 2 after H22 cancer cell implantation; H3, splenic nerve denervation by absolute ethanol; H4, FUS sti. spleen after splenic nerve denervation; H5, removal of the spleen; H6, FUS stimulated the splenic position after splenectomy.

Liver weight and mice weight also indirectly reflected the anti-cancer effect of FUS sti. spleen after splenic nerve denervation (Figure 5D and E). However, it must be noted that spleen weight and spleen index clearly indicated that splenic nerve denervation dramatically promoted tumor-induced splenomegaly (H3 vs H1, p < 0.01; Figure 5F and G), while FUS sti. spleen significantly reduced splenomegaly (H4 vs H3, p < 0.01; Figure 5F and G). The comprehensive analysis of the changes in splenic characteristics and tumor proliferation showed that the splenic nerve played an important role in the anti-cancer effect of the spleen under the natural state but not the splenic ultrasound stimulation.

ELISA results of NE and ACh showed that splenic neurons were not activated by FUS under splenic nerve denervation; thus, there was no significant change in the concentration of CAP-related NE and Ach (H4 vs H3; Figure 5H and I). The c-Fos immunohistochemical staining also proved that few splenic neurons responded to FUS stimulation after splenic nerve denervation (Figure 5J). However, splenic nerve denervation did not affect the response of splenic immune cells to FUS stimulation, especially the NK cells that secreted significantly higher levels of perforin and granzyme (H4 vs H3, p<0.01; Figure 5K and L). Particularly, after splenectomy, FUS irradiation in splenic position did not show any anti-tumor effect (H6 vs H5; Figure 5A–E, K-L), which demonstrated that the spleen irradiated with FUS was the only cause of the anti-cancer effect, and directly showed that the integrity of spleen was critical during FUS sti. spleen for tumor suppression. In summary, comprehensive analysis of splenectomy and splenic nerve denervation experiments clearly confirmed that FUS activated splenic anticancer immune function through the direct FUS stimulation of immune cells, thus further suppressing tumor proliferation.

FCM results proved that splenic nerve denervation did not affect cell proportion (H3 vs H1; Supplementary Figures 15–18), particularly not weaken the modulation properties of FUS sti. spleen on the increase of the positive immune cells (especially NK cells, CD8 T cells, Mφ1 and DC1) in the body, and decrease of the negative immune cells PMN-MDSCs (H4 vs H2/3; Supplementary Figures 15–18). It fully demonstrated that FUS directly stimulated the splenic immune cells to promote their proliferation and activated the anti-cancer capability. After the removal of the spleen, the proportion of most immune cells in the peripheral blood was significantly decreased (H5/6 vs H1/2; Supplementary Figure 16), which indicated the importance of spleen for the body’s immune system. However, after splenectomy, the spleen location irradiated by FUS did not show any impact on the number of immune cells (H6 vs H5; Supplementary Figures 16–18), which demonstrated that the spleen integrity or the presence of immune cells was essential for responding to FUS stimulation in tumor immunotherapy.

Association of Splenic Positive and Negative Immune Cells Responding to FUS Irradiation

Our previous study demonstrated that the proportion of PMN-MDSCs in the spleen and blood from tumor-bearing mice was significantly increased, which suppressed NK cells and CD8 T cells.⁷⁰ After FUS sti. spleen, the proportion of positive immune cells such as NK cells and CD8 T cells in the body was significantly increased, while the proportion of negative immune cells such as PMN-MDSCs was significantly decreased (Table 2, Supplementary File 1). Why did FUS, as a non-specific physical method, cause the number of positive and negative immune cells to change in two directions? The depletion of NK cells and CD8 T cells did not show consistent changes of PMN-MDSCs in the spleen and blood (Q5/6 and Q7/8 vs Q1/2; Tables 3 and 4); thus we proposed to reversely demonstrate the association of NK cells and CD8 T cells to PMN-MDSCs by depleting PMN-MDSCs. Forty orthotopic H22 HCC mice were randomly divided into four groups: R1 group was used as the control group; R2 group was subjected to FUS sti. spleen; R3 group was subjected to PMN-MDSCs depletion; and R4 group was subjected to FUS sti. spleen after PMN-MDSCs clearance.

PMN-MDSCs depletion inhibited tumor proliferation to a certain extent (R3 vs R1; Supplementary Figure 19A–D) and the subsequent FUS sti. spleen further enhanced tumor suppression (R4 vs R3; Supplementary Figure 19A–D). However, the tumor suppressive effect of FUS sti. spleen after PMN-MDSCs clearance was slightly weaker than that without PMN-MDSCs depletion (R4 vs R2; Supplementary Figure 19A–C). Mice weight did not clearly show a significant difference among the four groups (Supplementary Figure 19E), but spleen weight and spleen index showed a decreased trend in the R4 group as compared to the R3 group (Supplementary Figure 19Fand G).

FCM results showed that Ultra-LEAF™ purified anti-mouse Ly-6G antibody effectively eliminated PMN-MDSCs from the spleen, blood, tumor, and para-carcinoma tissue (R3 vs R1; Supplementary Figures 19H–K and 20–23). Interestingly, it significantly increased the proportion of Mφ, especially Mφ2, in the spleen and blood (R3 vs R1; Supplementary Figures 19H and I, 20 and 21). Moreover, PMN-MDSC depletion significantly increased NK cells and CD8 T cells only in the blood (R3 vs R1; Supplementary Figures 19I and 21). These results explained the reason that PMN-MDSCs clearance suppressed tumor proliferation to a certain extent (Supplementary Figure 19A–D). Even more important, PMN-MDSC depletion in spleen did not promote the proliferation of splenic NK cells and CD8 T cells (R3 vs R1; Supplementary Figures 19H and 20). Combined with the observation that the depletion of NK cells and CD8 T cells in the spleen did not increase the proportion of splenic PMN-MDSCs (Q5 and Q7 vs Q1; Table 3), it proved that there was no direct interaction between PMN-MDSCs and NK cells or CD8 T cells. Accordingly, it is reasonable to deduce that FUS sti. spleen increased the number of NK cells and CD8 T cells while the decrease of PMN-MDSCs was not directly correlated. PMN-MDSC decrease may be attributed to the increase of other immune cells in the spleen such as Mφ, especially Mφ2 mentioned above, or even to the rise of total tumor suppressor related immune cells in the spleen and blood. Of course, this mechanism has yet to be explored in the light of more direct and conclusive evidence.

FUS Regulating Calcium-Related Signaling Pathways to Modulate Splenic Immunity

The above results revealed that splenic NK cells and CD8 T cells were critical for suppressing tumor growth in response to FUS stimulation. Thereafter, RNA sequencing was performed to verify their changes at the molecular level and investigate the biological mechanism underlying this phenomenon.

According to the differential gene expression clustering heatmap and volcano plot, 655 upregulated genes and 344 downregulated genes were found in NK cells (Supplementary Figure 24A–C), as well as 1869 upregulated genes and 877 downregulated genes in CD8 T cells (Supplementary Figure 24A–C) in the FUS sti. spleen group compared with the control group (Supplementary Figure 24D–F). Both GO and KEGG enrichment analysis showed that the DEGs were significantly enriched in pathways of cell adhesion, cell migration, immune process (ie ECM-receptor interaction, antigen processing and presentation, and intestinal immune network for IgA production) and signal transduction (ie PI3K-AKT signaling pathway, Rap1 signaling pathway, Hippo signaling pathway and calcium signaling pathway) in the FUS sti. spleen group compared with the control group (Supplementary Figure 24G–N).

Further analysis revealed that genes related to NK cell immune function, especially activation, proliferation, and motility, were significantly up-regulated after FUS sti. spleen (Supplementary Figure 25A–C). Genes involved in calcium-related signaling pathways were also significantly highly expressed (Supplementary Figure 25D). According to GO functional annotation analysis, these upregulated genes were not only involved in the positive regulation of cell proliferation, migration and kinase activity, but also mainly involved in integrin-mediated signaling pathway and epidermal growth factor receptor signaling pathway (Supplementary Figure 26A). Moreover, MHC class II protein complex was suppressed (Supplementary Figure 26B), which promoted NK cell activation in mice as reported by Li et al.⁷¹ Particularly, KEGG enrichment analysis of these upregulated genes showed the upregulation of PI3K-Akt signaling pathway, Hippo signaling pathway, ErbB signaling pathway, and Ras signaling pathway (Supplementary Figure 26C), which were closely associated with cell proliferation, differentiation and activation, especially under ultrasound stimulation.^72–74 Downregulated genes were significantly enriched in the intestinal immune network for IgA production (Supplementary Figure 26D), which might be beneficial to tumor suppression as previous studies reported that IgA inhibits NK activity of NK cell-enriched lymphoid cells and gamma-interferon-treated effector cells.^75,76

More importantly, protein-protein networks revealed that among these significantly different signaling pathways, it is likely that the calcium signaling pathway played a central role and affected the expression of other signals to promote NK cell proliferation, migration and activation (Supplementary Figure 26E). Furthermore, GSEA analysis confirmed that FUS stimulation significantly regulated various signaling pathways of NK cells, such as up-regulating ECM-receptor interaction, PI3K-AKT signaling pathway, Hippo signaling pathway and calcium signaling pathways (Supplementary Figure 26F).

Additionally, the genes related to CD8 T cell activation, proliferation, and motility were significantly up-regulated in the FUS sti. spleen group compared to the control group (Supplementary Figure 27E–G). Genes involved in calcium-related signaling pathways were also highly expressed (Supplementary Figure 27H). GO and KEGG enrichment and network analysis of the differential genes showed significant enrichment of pathways in signal transduction (ie Rap1 signaling pathway, MAPK/ERK signaling pathway, and calcium-related signaling pathways) and immune process (ie IL-6 production, IgA production, Fc epsilon RI signaling pathway and Fc gamma R-mediated phagocytosis) (Supplementary Figure 28A–D). Particularly, the calcium signaling pathway was involved in regulating most of the others and it was probably the initial response to FUS stimulation (Supplementary Figure 28E). GSEA analysis proved that these signaling pathways involving CD8 T cells proliferation, activation and migration, such as ECM-receptor interaction, PI3K-AKT signaling pathway, Rap1 signaling pathway and calcium signaling pathways were significantly upregulated (Supplementary Figure 28F). Although the GO and KEGG enrichment analysis of the downregulated genes showed the downregulation of MAPK signaling pathway, TNF signaling pathway and others (Supplementary Figure 28D), GSEA analysis revealed that these signaling pathways were highly expressed when the tumor-bearing mice were subjected to FUS sti. spleen (Supplementary Figure 28G).

These functional annotations suggested that FUS first activated calcium-related signaling pathways and then regulated other signaling pathways to promote the proliferation, activation, and migration of NK cells and CD8 T cells. Many previous studies confirmed that ultrasound affects the proliferation, activation, and differentiation of various cells by altering intracellular calcium-related signaling pathways to achieve specific biological effects.^77–79 The further exploration of splenic calcium changes in response to FUS stimulation by the Von Kossa staining and FCM detecting Fluo-4 AM labeled cells revealed that the splenic cell population of Von Kossa stained and Fluo-4 AM labeled were observably increased in the FUS sti. spleen group as compared with the control group (Figure 6A and B). Calcium deposits in spleen can promote cytokine production, such as Mφ secreting IL-1β and TNF-α.^80,81 Particularly, the results of in vitro experiments (Figure 6C) fully demonstrated that FUS stimulation promoted NK cell proliferation and activation to suppress cancer cell deterioration (note: H22 cancer cells were unable to adhere to the bottom and difficult to separate from NK cells that migrated from the upper layer of the transwell to the bottom layer, which led to a significant error in the subsequent detection results. Thus, Hepa1-6 cancer cells were chosen for the in vitro study). Moreover, calcium strengthened the impact of FUS stimulation on splenic NK cells to suppress the proliferation of cancer cells in vitro (Figure 6C). Besides, ultrasound shattered cancer cell suspension further enhanced the anticancer effect of calcium-involved FUS simulation on splenic NK cells. Additionally, we also verified that calcium reinforced the effect of FUS stimulation on splenic CD8 T cells to suppress cancer cell deterioration in vitro (not supplying the data). Overall, the above experimental results succinctly proved that FUS altered calcium-related signaling pathways in NK cells and CD8 T cells to reinforce the antitumor effect. However, in the future, it is worthwhile comprehensively and systematically analyzing the specific mechanism of splenic immune cells in response to FUS stimulation by spatial region-resolved proteome and spatial single-cell sequencing.

Figure 6 Calcium strengthened FUS-stimulated splenic NK cells against Hepa1-6 cancer cells. (A) Von Kossa staining of calcium deposition in the spleen with or without FUS stimulation. (B) FCM detected Fluo-4 AM labeled cells in the spleen with or without FUS stimulation. (n = 4; * p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; mean ± SEM). (C) results of calcium strengthened impact of FUS stimulation on splenic NK cells to suppress cancer cells in vitro, including GFP fluorescence imaging, CV staining and CCK-8 assay to analyze the proliferation and activity of Hepa1-6 cancer cells, and CCK-8 assay to characterize the proliferation and cytoactivity of NK cells. (n = 3; * p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; mean ± SEM). T1, Hepa1-6 cancer cells; T2, NK cells; T3, NK cells cocultured with Hepa1-6 cancer cells in the transwell; T4, FUS stimulated NK cells; T5, NK cells subjected to FUS irradiation, and then cocultured with Hepa1-6 cancer cells in the transwell; T6, NK cells mixed with Hepa1-6 lysate and cocultured with Hepa1-6 cancer cells in the transwell; T7, NK cells mixed with Hepa1-6 lysate, subjected to FUS irradiation, and then cocultured with Hepa1-6 cancer cells in the transwell; T8, NK cells mixed with Hepa1-6 lysate in 300 nM calcium culture medium and cocultured with Hepa1-6 cancer cells in the transwell; T9, NK cells in 300 nM calcium culture medium subjected to FUS irradiation, and then cocultured with Hepa1-6 cancer cells in the transwell; T10, NK cells mixed with Hepa1-6 lysate in 300 nM calcium culture medium, subjected to FUS irradiation, and then cocultured with Hepa1-6 cancer cells in the transwell.

Therapeutic Application of FUS Sti. Spleen on Intermediate and Advanced Tumors

The above studies in which FUS sti. spleen was performed on the second day after cancer cells implanted into the mice were equivalent to the start of the treatment in the early stage of cancer (before tumor formation). However, the anticancer effect of FUS sti. spleen after tumor formation is one of the worthiest explorations in clinical transformation and application. Therefore, a comparative study was performed to investigate the therapeutic application of FUS sti. spleen on 5^th and 9^th day after H22 cancer cell transplantation into the liver deemed to be tumorigenesis (Figure 7A). Four experimental groups were set as follows: K1 group was used as the control group of xenograft H22 HCC in situ; K2-K4 groups were subjected to FUS sti. spleen on days 0, 4 and 8 after H22 cancer cell implantation, and once every two days (Figure 7A). At the experimental endpoint (day 21), animals underwent terminal anesthesia followed by comprehensive tissue harvesting for other experimental analyses (eg, FCM).

Figure 7 FUS sti. spleen suppressed the proliferation of the xenograft H22 HCC in situ. (A) experimental flow diagram of FUS sti. spleen on day 0, 4 and 8 after H22 cancer cell implantation. (B) images of spleen and xenograft H22 HCC in situ. (C) tumor volume. (D) tumor weight. (E) liver weight. (F) curve of mice weight. (G) spleen weight. (H) spleen index calculated by the ratio of spleen weight to mouse weight. (n = 7; * p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; mean ± SEM). (I–L), FCM results of NK cells, CD8 T cells and PMN-MDSCs in the spleen, blood, tumor, and para-carcinoma tissues, respectively. (n = 4; * p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; mean ± SEM). K1, control group of xenograft H22 HCC in situ; K2-K4, FUS sti. spleen on day 0, 4 and 8, respectively, after H22 cancer cell implantation. Representative FCM plots and statistical diagram of immune cells in K1-K4 groups are shown in Supplementary Figures 29–32. The results shown in Figures 7 and 5 were from the same batch of experiments. With animal welfare in mind, only one control group and FUS sti. spleen group were set up; thus, the K1 and K2 group were the same as the H1 and H2 group respectively.

The results showed that FUS sti. spleen on day 0, 4 and 8 significantly suppressed the proliferation of xenograft H22 HCC carcinoma in situ, with an anti-tumor rate of up to ~70%, ~67% and ~48% respectively (K2/3/4 vs K1; Figure 7B–D). The results tautologically demonstrated that FUS sti. spleen exerted an evident tumor suppression, regardless of being performed before or after tumorigenesis, but the later ultrasound irradiation on the spleen was performed after the onset of the tumor, with a smaller effect on tumor suppression. Additionally, the statistical results of liver weight and mice weight showed a decreasing trend in the K2-K4 groups compared to the K1 group (Figure 7E and F). The spleen weight and spleen index also demonstrated that FUS sti. spleen attenuated the symptoms of splenomegaly caused by cancer, as same as in Figure 7G–H.

Statistic FCM results showed that the proportion of NK cells and CD8 T cells was significantly increased in the spleen, blood, tumor, and para-carcinoma tissue of the K2-K4 groups compared with the K1 group, and the proportion of PMN-MDSCs was significantly reduced (Figure 7I–L, Supplementary Figures 29–32). However, no significant difference was found in the number of immune cells among the K2-K4 groups except that NK cells were significantly higher in the K2 group (Figure 7I–L, Supplementary Figures 29–32). In addition, the proportion of Mφ1 and DC1 in four tissues of the K2-K4 groups was significantly increased (Supplementary Figures 29–32). What needs to be highlighted is the inconsistency in the changes in the cell proportion of Mφ2, DC2 or other immune cells in the spleen, blood, tumor, and para-carcinoma tissue (Supplementary Figures 29–32). Our hypothesis was that this phenomenon was mainly attributed to the migration of splenic immune cells to tumor lesions after responding to ultrasonic stimulation and stressing the tumor microenvironment, which was fully demonstrated in our previous work that CD11b⁺ CD43^hi Ly6C^lo splenocyte-derived Mφ migrate into the lesion in a liver fibrosis model.⁸²

Special attention should be paid to the fact that the biological effect of ultrasound on spleen varies in different disease models. As reported by Nunes et al⁶⁸ and Liu et al,⁸³ FUS sti. spleen mainly affects Mφ to inhibit the secretion of inflammatory cytokines in colitis and myocarditis. Also related to RA, CD 4 T cell, CD8 T cell, B cell and myeloid cell populations were significantly changed in response to ultrasound, specifically the CD8 T cells show an evident correlation with FUS sti. spleen based therapy.⁸⁴ However, this study fully confirmed that FUS sti. spleen significantly promoted the proliferation of NK cells, CD8 T cells, Mφ1 and DC1 in various tumor models, and reduced PMN-MDSCs, but mainly suppressed cancer progression by stimulating NK cells. Additionally, the biological effects of ultrasound on different types of cells are also varied, including enhancing MC3T3-E1 osteoblasts and hematopoietic stem cell ingrowth, proliferation, and early differentiation,^85,86 inducing neural progenitor cell polarization,⁷² promoting proliferation and migration of HaCaT keratinocytes,⁷³ facilitating the extracellular matrix synthesis of degenerative human nucleus pulposus cells,⁸⁷ and motivating the expression of brain-derived neurotrophic factor in astrocyte.⁷⁷ However, the difference in biological effects of ultrasound irradiation on various immune cells enriched in the spleen is not yet known and needs to be detailedly studied in the future.

In conclusion, although FUS sti. spleen non-specifically activated spleen immunological function to universally suppress tumor proliferation, it was effective on the early stage of cancer or cancer prevention. Nevertheless, it might be necessary to combine FUS sti. spleen with other treatments in patients with advanced cancer, such as applying FUS qualitatively and quantitatively intervention on the tumor (FUS int. tumor).

FUS Sti. Spleen & FUS Int. Tumor Synergistically Suppress Tumor Proliferation

Previous studies reported that ultrasonic intervention on tumor lesions enhances tumor suppression by promoting the proliferation, activation, and infiltration of immune cells.^88,89 In this study, comparing and discussing FUS sti. spleen and FUS int. tumors – two strategies for immunomodulation to counteract the swelling by interfering from immune organs and tumor foci, respectively – are crucial. Five experimental groups were considered: #1 group, normal mice (mice without any treatment); #2 group, control group of subcutaneous H22 tumor mice; #3 group, FUS sti. spleen on day 0 after H22 cancer cell implantation, and once every two days; #4 group, FUS int. tumor; #5 group, FUS sti. spleen and FUS int. tumor synergistic treatment. At the experimental endpoint (day 29), animals underwent terminal anesthesia followed by comprehensive tissue harvesting for other experimental analyses (eg, FCM). According to pre-experiments and our previous results,⁹⁰ FUS int. tumor was performed every 4 days from day 9 after the subcutaneous inoculation of H22 cancer cells (Figure 8A), and the ultrasonic parameters were set to 3.3 MPa, 1%, 20s to achieve controlled disruption of tumor’s extracellular matrix barrier and tumor microenvironment, thereby enhancing immunomodulatory effects without excessively compromising tissue architecture or inducing cellular necrosis.

Figure 8 FUS sti. spleen and FUS int. tumor synergistically or individually suppressed the proliferation of the subcutaneous H22 HCC tumor. (A) experimental flow diagram of FUS int. tumor every four days started on day 8 after H22 cancer cell implantation. (B) image of spleen and tumor. (C) tumor volume. (D) tumor weight. (E) curve of mice weight. (F) spleen weight. (G) spleen index, the ratio of spleen weight to mouse weight. (n = 8; * p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; mean ± SEM). (H–J) FCM results of NK cells, CD8 T cells and PMN-MDSCs in the spleen, blood, and tumor, respectively. (n = 4; * p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001; mean ± SEM). (K and L) schematic illustration of the mechanism during FUS int. tumor to enhance anticancer immune response of FUS sti. spleen. #1, normal mice group; #2, control group of subcutaneous H22 tumor mice; #3, FUS sti. spleen on the second day after H22 cancer cell implantation; #4, FUS int. tumor; #5, FUS sti. spleen and FUS int. tumor synergistic effect.

The results of #3 vs #2 group again confirmed the ability and application prospect of FUS sti. spleen in suppressing tumor proliferation (Figure 8B–D). Moreover, FUS int. tumor showed an effective tumor suppression (#4 vs #2 group; Figure 8B–D), but it did not show any significant difference as compared with the #3 group (Figure 8B–D). It was noteworthy that FUS sti. spleen and FUS int. tumor synergistically induced the best tumor suppression (Figure 8B–D). In addition, FUS sti. spleen and FUS int. tumor synergistically or either alone reduced the mice weight (Figure 8E), but the combined application significantly exacerbated splenomegaly (Figure 8F and G).

FCM results revealed that the proportion of NK cells and CD8 T cells in the spleen, blood and tumor was significantly and synergistically increased compared to the effect of each of the two taken individually (Figure 8H–J). The comparison of the #3-#5 groups revealed that the number of NK cells and CD8 T cells in the spleen and blood was significantly higher after H22 tumor-bearing mice were subjected to the synergistic effect of FUS sti. spleen and FUS int. tumor compared to the effect of each of the two taken individually (Figure 8H and I). However, such a trend in the number of immune cells was not present in tumor tissues, which was mainly attributed to the fact that FUS int. tumor not only disrupted the structure of tumor stromal barrier and cancer cells, but also killed immune cells. Additionally, FUS sti. spleen, FUS int. tumor or both synergistically significantly increased the proportion of Mφ1 or DC1 in the spleen, blood, and tumor (Supplementary Figures 33–35). However, consistent with the FCM results in Table 2, FUS sti. spleen had no significant consistent effect on CD4 T cells or their subtype in the spleen, blood, and tumor from subcutaneous H22 tumor-bearing mice.

FUS has been gradually developed into a clinical strategy approach for targeted tumor disruption through thermal mechanisms.^91–93 In essence, FUS at high intensity generated focal temperatures exceeding 60 °C that instantaneously caused coagulative necrosis of the target soft tissue (Figure 8K).^94,95 In contrast, non-thermomechanical disruption may have a more relevant role in low intensity FUS ablation, particularly in terms of immunomodulatory effects. As opposite to thermal ablation causing protein coagulation necrosis, the mechanical disruption of FUS int. tumor converts immunologically “cold” (poorly infiltrated) tumors into immunologically “hot” (well-infiltrated) ones by damaging cancer cells to release antigens, collapsing the tumor matrix barrier to enhance antigens captured by immune cells, and especially intensifying the infiltration of immune cells and cytokines into the tumor lesions, which strengthened the immunotherapeutic responses to cancer (Figure 8K and L).^88,96 Although FUS int. tumor alone exerted an anti-tumor effect, to a certain extent, the tumor matrix damage would increase the risk of cancer cell spread and metastasis.⁹⁷ However, in this study, the synergistic use of FUS sti. spleen and FUS int. tumor was much more effective, since tumor disruption by FUS provided an antigen source for activating splenic immune cells during the process of FUS sti. spleen to suppress tumor proliferation (Figure 8L).

Discussion

The comparison of tumor inhibition and spleen injury under different ultrasonic parameters of FUS sti. spleen revealed that the spleen (or splenic cells) had significant ultrasound dose dependence and tolerance. This was also reported in previous studies; for example, Liu et al⁸³ successfully screened an effective ultrasonic parameter of 0.35 MPa with a 1s on /5s off duty cycle to significantly alleviate autoimmune myocarditis, and regulate the proportion and function of Tregs and Mφ by activating CAP. Besides, Cotero et al²³ discovered that non-invasive splenic ultrasound stimulation at 0.83 MPa effectively alleviates the severity of arthritis and reduces cytokine response to endotoxin through CAP modulation based on CD4⁺ T cells. However, according to a summary of studies (Supplementary Table 3), although splenic ultrasound stimulation can effectively suppress various inflammatory diseases, the ultrasonic parameters applied for spleen stimulation are extremely variable. Some studies only referred to the ultrasonic parameters mentioned in previous reports but lacked screening and optimization; for example, Hu et al⁸⁴ and Morton et al⁹⁸ used the parameters of 0.35 MPa and 1-s on/5-s off reported by Zachs et al³⁰ for splenic immunomodulation to treat inflammatory diseases. However, in practical applications, different ultrasound platforms lead to significant differences in ultrasonic focal region. More importantly, although some studies proved that the mechanism of splenic ultrasound stimulation for inflammatory disease therapy is based on CAP activation,^23,29,69 some other studies demonstrated the direct action of ultrasound on splenic immune cells.^30,84 Therefore, the dependence, specificity, and mechanism of spleen response to ultrasonic stimulation is also different, which may be one of the reasons for the different ultrasonic parameters applied in the above studies. This study also demonstrated that FUS sti. spleen to modulate immune function for tumor suppression was not mainly by activating CAP but depended on the direct stimulation of the proliferation and activation of NK cells, as well as the biological effects of ultrasonic stimulation on CD8 T cells, Mφ, DCs and MDSCs.

At present, there is a lack of studies on splenic ultrasound stimulation to regulate immunotherapy for various diseases; thus, ultrasonic parameters are not objectively consistent, which seriously restricts the translational progress of clinical application. Even if the ultrasonic parameters were screened and optimized in this study, there are still some shortcomings. For example, when the ultrasonic intensity is less than 2.3 MPa, the effect of a longer duration of ultrasound irradiation on splenic immune modulation for tumor suppression needs to be further studied. Fortunately, there are two studies of splenic ultrasound stimulation for immunotherapy on inflammatory diseases registered in http://www.clinicaltrials.gov (NCT03690466, and NCT03548116), but their clinical results are not yet officially published. We look forward to their positive news, and expect that FUS sti. spleen can be widely employed in clinical treatment to alleviate patient suffering.

Conclusion

Combined under optimized ultrasound parameters (2.3 MPa, 1%, and 20s), FUS sti. spleen has been demonstrated to directly activate splenic immune cells, particularly NK cells and CD8⁺ T lymphocytes, through altering calcium-dependent signaling but without significant involvement of cholinergic neuroimmune modulation, thereby effectively suppressing early-to-intermediate stage malignancies. With perspective, concurrent FUS sti. spleen and FUS int. tumor significantly achieve superior therapeutic outcomes, indicating substantial clinical translation potential that warrants further validation through animal and clinical trials.

Abbreviations

IRI, ischemia-reperfusion injury; VNS, vagus nerve stimulation; α7nAChR, α7 nicotinic acetylcholine receptors; FUS, focused ultrasound; DC, dendritic cell; NK, natural killer; MDSC, myeloid-derived suppressor cell; M-MDSC, monocytic myeloid-derived suppressor cell; PMN-MDSC, polymorphonucler myeloid-derived suppressor cell; Mφ, macrophage; Treg, regulatory T cells; FCM, flow cytometry; HCC, hepatocellular carcinoma; TNF-α, tumor necrosis factor α; CAP, cholinergic anti-inflammatory pathway; NE, norepinephrine; ACh, acetylcholine; IgA, immunoglobulin A.

Data Sharing Statement

Raw and analyzed sequencing data in this study have been deposited in the NCBI’s Gene Expression Omnibus (under series accession code GSE267237 and GSE267445).

Ethical Approval

All animal experiments were performed in accordance with the institutional guidelines and approved by the Animal Experimentation Ethics Committee of Xi’an Jiaotong University. Additionally, all methodological protocols were designed and reported in accordance with the ARRIVE guidelines for the welfare of the laboratory animals.

Acknowledgments

We thank Hongwei Tian, Gaixia He and Haiyan Chen for providing technical support in preparation of animal models. We thank Pro. Pengfei Liu, Pro. Guangyao Kong, Pro. Yujin Zong, Pro. Shemin Lv and Pro. Tielin Yang for their careful guidance on the design of experimental scheme. In addition, Wei Dong thanks the care and support from Yinggang Zhang, Shoufei Qu, Yameng Wei, Shaoying Zhang, Pro. Jun Li and Pro. Fanpu Ji in the past 2 years.

We would like to thank MogoEdit (https://www.mogoedit.com) for its English editing during the preparation of this manuscript. And we sincerely thank Shanghai OE Biotech Co., LTD (https://www.oebiotech.com/) for the RNA sequencing and data analysis.

This paper has been uploaded to ResearchSquare and bioRxiv as a preprint: https://www.researchsquare.com/article/rs-4639146/v1; https://www.biorxiv.org/content/10.1101/2025.03.31.646454v1.

Funding

This work was supported by the National Natural Science Foundation of China (No. 12204370), the Innovation Ability Supporting Program of Shaanxi Province (No. 2023WGZJ-ZD-09), and the Basic-Clinical Integration Innovation Project of Xi ‘an Jiaotong University (No. YXJLRH2022092).

Disclosure

The authors declare that they have no competing interests in this work.

References

1. Verlinden TJM, van Dijk P, Hikspoors J, Herrler A, Lamers WH, Köhler SE. Innervation of the human spleen: a complete hilum-embedding approach. Brain Behav Immun. 2019;77:92–100. doi:10.1016/j.bbi.2018.12.009

2. Hu D, Al-Shalan HAM, Shi Z, et al. Distribution of nerve fibers and nerve-immune cell association in mouse spleen revealed by immunofluorescent staining. Scientific Reports. 2020;10(1):9850. doi:10.1038/s41598-020-66619-0

3. Anagnostou VK, Doussis-Anagnostopoulou I, Tiniakos DG, et al. Ontogeny of intrinsic innervation in the human thymus and spleen. J Histochem Cytochem. 2007;55(8):813–820. doi:10.1369/jhc.6A7168.2007

4. Elenkov IJ, Vizi ES. Presynaptic modulation of release of noradrenaline from the sympathetic nerve terminals in the rat spleen. Neuropharmacology. 1991;30(12a):1319–1324. doi:10.1016/0028-3908(91)90029-B

5. Sudo A. Decrease in Adrenaline content of various organs of the rat after 6-hydroxydopamine. Eur J Pharmacol. 1985;114(1):79–83. doi:10.1016/0014-2999(85)90524-2

6. Pavlov VA, Tracey KJ. The vagus nerve and the inflammatory reflex–linking immunity and metabolism. Nat Rev Endocrinol. 2012;8(12):743–754. doi:10.1038/nrendo.2012.189

7. Tracey KJ. Reflex control of immunity. Nat Rev Immunol. 2009;9(6):418–428. doi:10.1038/nri2566

8. Tracey KJ. Reflexes in Immunity. Cell. 2016;164(3):343–344. doi:10.1016/j.cell.2016.01.018

9. Borovikova LV, Ivanova S, Zhang M, et al. Vagus nerve stimulation attenuates the systemic inflammatory response to endotoxin. Nature. 2000;405(6785):458–462. doi:10.1038/35013070

10. Bai A, Guo Y, Lu N. The effect of the cholinergic anti-inflammatory pathway on experimental colitis. Scand J Immunol. 2007;66(5):538–545. doi:10.1111/j.1365-3083.2007.02011.x

11. Munyaka P, Rabbi MF, Pavlov VA, Tracey KJ, Khafipour E, Ghia JE. Central muscarinic cholinergic activation alters interaction between splenic dendritic cell and CD4+CD25- T cells in experimental colitis. PLoS One. 2014;9(10):e109272. doi:10.1371/journal.pone.0109272

12. Ji H, Rabbi MF, Labis B, Pavlov VA, Tracey KJ, Ghia JE. Central cholinergic activation of a vagus nerve-to-spleen circuit alleviates experimental colitis. Mucosal Immunol. 2014;7(2):335–347. doi:10.1038/mi.2013.52

13. Bonaz B, Sinniger V, Pellissier S. Vagus nerve stimulation: a new promising therapeutic tool in inflammatory bowel disease. J Intern Med. 2017;282(1):46–63. doi:10.1111/joim.12611

14. Inoue T, Abe C, Sung SS, et al. Vagus nerve stimulation mediates protection from kidney ischemia-reperfusion injury through α7nAChR+ splenocytes. J Clin Invest. 2016;126(5):1939–1952. doi:10.1172/JCI83658

15. Koopman FA, Schuurman PR, Vervoordeldonk MJ, Tak PP. Vagus nerve stimulation: a new bioelectronics approach to treat rheumatoid arthritis? Best Pract Res Clin Rheumatol. 2014;28(4):625–635. doi:10.1016/j.berh.2014.10.015

16. Koopman FA, Chavan SS, Miljko S, et al. Vagus nerve stimulation inhibits cytokine production and attenuates disease severity in rheumatoid arthritis. Proc Natl Acad Sci U S A. 2016;113(29):8284–8289. doi:10.1073/pnas.1605635113

17. Meregnani J, Clarençon D, Vivier M, et al. Anti-inflammatory effect of vagus nerve stimulation in a rat model of inflammatory bowel disease. Auton Neurosci. 2011;160(1–2):82–89. doi:10.1016/j.autneu.2010.10.007

18. Zhang Z, Yu Q, Zhang X, et al. Electroacupuncture regulates inflammatory cytokines by activating the vagus nerve to enhance antitumor immunity in mice with breast tumors. Life Sci. 2021;272:119259. doi:10.1016/j.lfs.2021.119259

19. Li QW, Yu MW, Yang GW, et al. Effect of acupuncture in prevention and treatment of chemotherapy-induced nausea and vomiting in patients with advanced cancer: study protocol for a randomized controlled trial. Trials. 2017;18(1):185. doi:10.1186/s13063-017-1927-2

20. Li QW, Yu MW, Wang XM, et al. Efficacy of acupuncture in the prevention and treatment of chemotherapy-induced nausea and vomiting in patients with advanced cancer: a multi-center, single-blind, randomized, sham-controlled clinical research. Chin Med. 2020;15:57. doi:10.1186/s13020-020-00333-x

21. Rush AJ, Marangell LB, Sackeim HA, et al. Vagus nerve stimulation for treatment-resistant depression: a randomized, controlled acute phase trial. Biol Psychiatry. 2005;58(5):347–354. doi:10.1016/j.biopsych.2005.05.025

22. Rush AJ, Sackeim HA, Marangell LB, et al. Effects of 12 months of vagus nerve stimulation in treatment-resistant depression: a naturalistic study. Biol Psychiatry. 2005;58(5):355–363. doi:10.1016/j.biopsych.2005.05.024

23. Cotero V, Fan Y, Tsaava T, et al. Noninvasive sub-organ ultrasound stimulation for targeted neuromodulation. Nature Communications. 2019;10(1):952. doi:10.1038/s41467-019-08750-9

24. Olofsson PS, Tracey KJ. Bioelectronic medicine: technology targeting molecular mechanisms for therapy. J Intern Med. 2017;282(1):3–4. doi:10.1111/joim.12624

25. Bassi GS, Dias DPM, Franchin M, et al. Modulation of experimental arthritis by vagal sensory and central brain stimulation. Brain Behav Immun. 2017;64:330–343. doi:10.1016/j.bbi.2017.04.003

26. Martelli D, McKinley MJ, McAllen RM. The cholinergic anti-inflammatory pathway: a critical review. Auton Neurosci. 2014;182:65–69. doi:10.1016/j.autneu.2013.12.007

27. Mei H, Irazoqui PP. Miniaturizing wireless implants. Nat Biotechnol. 2014;32(10):1008–1010. doi:10.1038/nbt.3038

28. Fattahi P, Yang G, Kim G, Abidian MR. A review of organic and inorganic biomaterials for neural interfaces. Adv Mater. 2014;26(12):1846–1885. doi:10.1002/adma.201304496

29. Gigliotti JC, Huang L, Ye H, et al. Ultrasound prevents renal ischemia-reperfusion injury by stimulating the splenic cholinergic anti-inflammatory pathway. J Am Soc Nephrol. 2013;24(9):1451–1460. doi:10.1681/ASN.2013010084

30. Zachs DP, Offutt SJ, Graham RS, et al. Noninvasive ultrasound stimulation of the spleen to treat inflammatory arthritis. Nat Commun. 2019;10(1):951. doi:10.1038/s41467-019-08721-0

31. Dong W, Li Y, Fei Q, et al. Targeted spleen modulation: a novel strategy for next-generation disease immunotherapy. Theranostics. 2025;15(10):4416–4445. doi:10.7150/thno.111116

32. Yahata K, Kanno H, Ozawa H, et al. Low-energy extracorporeal shock wave therapy for promotion of vascular endothelial growth factor expression and angiogenesis and improvement of locomotor and sensory functions after spinal cord injury. J Neurosurg Spine. 2016;25(6):745–755. doi:10.3171/2016.4.SPINE15923

33. Ito K, Fukumoto Y, Shimokawa H. Extracorporeal shock wave therapy for ischemic cardiovascular disorders. Am J Cardiovasc Drugs. 2011;11(5):295–302. doi:10.2165/11592760-000000000-00000

34. Vezeridis PS, Semeins CM, Chen Q, Klein-Nulend J. Osteocytes subjected to pulsating fluid flow regulate osteoblast proliferation and differentiation. Biochem Biophys Res Commun. 2006;348(3):1082–1088. doi:10.1016/j.bbrc.2006.07.146

35. Fung CH, Cheung WH, Pounder NM, Harrison A, Leung KS. Osteocytes exposed to far field of therapeutic ultrasound promotes osteogenic cellular activities in pre-osteoblasts through soluble factors. Ultrasonics. 2014;54(5):1358–1365. doi:10.1016/j.ultras.2014.02.003

36. Shin VY, Jin HC, Ng EK, et al. Nicotine and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone induce cyclooxygenase-2 activity in human gastric cancer cells: involvement of nicotinic acetylcholine receptor (nAChR) and beta-adrenergic receptor signaling pathways. Toxicol Appl Pharmacol. 2008;233(2):254–261. doi:10.1016/j.taap.2008.08.012

37. Maccari S, Profumo E, Saso L, Marano G, Buttari B. Propranolol promotes monocyte-to-macrophage differentiation and enhances macrophage anti-inflammatory and antioxidant activities by NRF2 activation. Int J Mol Sci. 2024;25(7):3683. doi:10.3390/ijms25073683

38. Chai GS, Wu JJ, Gong J, et al. Activation of β2-adrenergic receptor ameliorates amyloid-β-induced mitophagy defects and tau pathology in mice. Neuroscience. 2022;505:34–50. doi:10.1016/j.neuroscience.2022.09.020

39. Shionoya Y, Sunada K, Shigeno K, Nakada A, Honda M, Nakamura T. Can nerve regeneration on an artificial nerve conduit be enhanced by ethanol-induced cervical sympathetic ganglion block? PLoS One. 2017;12(12):e0189297. doi:10.1371/journal.pone.0189297

40. Liu M, Ni H, Tao J, Xie K. Lumbar sympathetic nerve modulation using absolute ethanol for the treatment of primary lower-extremity hyperhidrosis: a dose-effect pilot study. Med Sci Monit. 2021;27:e928209. doi:10.12659/MSM.928209

41. Shionoya Y, Sunada K, Tsujimoto G, Shigeno K, Nakamura T. Ethanol-induced cervical sympathetic ganglion block applications for promoting canine inferior alveolar nerve regeneration using an artificial nerve. J Vis Exp. 2018;141.

42. Lu X, Li B, Liu S, Zhu J, Li J. The dose and volume-response relationship of an ethanol-induced tibial nerve block in a rabbit model. Am J Phys Med Rehabil. 2014;93(2):138–145. doi:10.1097/PHM.0b013e3182a92b57

43. Pham-Marcou TA, Gentili M, Asehnoune K, Fletcher D, Mazoit JX. Effect of neurolytic nerve block on systemic carrageenan-induced inflammatory response in mice. Br J Anaesth. 2005;95(2):243–246. doi:10.1093/bja/aei150

44. Zhang X, Lei B, Yuan Y, et al. Brain control of humoral immune responses amenable to behavioural modulation. Nature. 2020;581(7807):204–208. doi:10.1038/s41586-020-2235-7

45. Tuomela K, Ambrose AR, Davis DM. Escaping death: how cancer cells and infected cells resist cell-mediated cytotoxicity. Front Immunol. 2022;13:867098. doi:10.3389/fimmu.2022.867098

46. Kuerten S, Nowacki TM, Kleen TO, Asaad RJ, Lehmann PV, Tary-Lehmann M. Dissociated production of perforin, granzyme B, and IFN-gamma by HIV-specific CD8(+) cells in HIV infection. AIDS Res Hum Retroviruses. 2008;24(1):62–71. doi:10.1089/aid.2007.0125

47. Hodge G, Barnawi J, Jurisevic C, et al. Lung cancer is associated with decreased expression of perforin, granzyme B and interferon (IFN)-γ by infiltrating lung tissue T cells, natural killer (NK) T-like and NK cells. Clin Exp Immunol. 2014;178(1):79–85. doi:10.1111/cei.12392

48. Chen S, Jiao J, Jiang D, et al. T-box transcription factor Brachyury in lung cancer cells inhibits macrophage infiltration by suppressing CCL2 and CCL4 chemokines. Tumour Biol. 2015;36(8):5881–5890. doi:10.1007/s13277-015-3260-2

49. Ottaviani C, Nasorri F, Bedini C, de Pità O, Girolomoni G, Cavani A. CD56brightCD16(-) NK cells accumulate in psoriatic skin in response to CXCL10 and CCL5 and exacerbate skin inflammation. Eur J Immunol. 2006;36(1):118–128. doi:10.1002/eji.200535243

50. Ghafouri-Fard S, Shahir M, Taheri M, Salimi A. A review on the role of chemokines in the pathogenesis of systemic lupus erythematosus. Cytokine. 2021;146:155640. doi:10.1016/j.cyto.2021.155640

51. Janssens R, Struyf S, Proost P. The unique structural and functional features of CXCL12. Cell Mol Immunol. 2018;15(4):299–311. doi:10.1038/cmi.2017.107

52. Janssens R, Struyf S, Proost P. Pathological roles of the homeostatic chemokine CXCL12. Cytokine Growth Factor Rev. 2018;44:51–68. doi:10.1016/j.cytogfr.2018.10.004

53. Gouwy M, Schiraldi M, Struyf S, Van Damme J, Uguccioni M. Possible mechanisms involved in chemokine synergy fine tuning the inflammatory response. Immunol Lett. 2012;145(1–2):10–14. doi:10.1016/j.imlet.2012.04.005

54. Zhou W, Guo S, Liu M, Burow ME, Wang G. Targeting CXCL12/CXCR4 axis in tumor immunotherapy. Curr Med Chem. 2019;26(17):3026–3041. doi:10.2174/0929867324666170830111531

55. Yang Y, Li J, Lei W, et al. CXCL12-CXCR4/CXCR7 axis in cancer: from mechanisms to clinical applications. Int J Biol Sci. 2023;19(11):3341–3359. doi:10.7150/ijbs.82317

56. Urdinguio RG, Fernandez AF, Moncada-Pazos A, et al. Immune-dependent and independent antitumor activity of GM-CSF aberrantly expressed by mouse and human colorectal tumors. Cancer Res. 2013;73(1):395–405. doi:10.1158/0008-5472.CAN-12-0806

57. Bretscher V, Andreutti D, Neuville P, et al. GM-CSF expression by tumor cells correlates with aggressivity and with stroma reaction formation. J Submicrosc Cytol Pathol. 2000;32(4):525–533.

58. Hong IS. Stimulatory versus suppressive effects of GM-CSF on tumor progression in multiple cancer types. Exp Mol Med. 2016;48(7):e242. doi:10.1038/emm.2016.64

59. Thorn M, Guha P, Cunetta M, et al. Tumor-associated GM-CSF overexpression induces immunoinhibitory molecules via STAT3 in myeloid-suppressor cells infiltrating liver metastases. Cancer Gene Ther. 2016;23(6):188–198. doi:10.1038/cgt.2016.19

60. Sweet MJ, Hume DA. CSF-1 as a regulator of macrophage activation and immune responses. Arch Immunol Ther Exp. 2003;51(3):169–177.

61. Hume DA, MacDonald KP. Therapeutic applications of macrophage colony-stimulating factor-1 (CSF-1) and antagonists of CSF-1 receptor (CSF-1R) signaling. Blood. 2012;119(8):1810–1820. doi:10.1182/blood-2011-09-379214

62. Malkova AM, Gubal AR, Petrova AL, et al. Pathogenetic role and clinical significance of interleukin-1β in cancer. Immunology. 2023;168(2):203–216. doi:10.1111/imm.13486

63. Zhou J, Tulotta C, Ottewell PD. IL-1β in breast cancer bone metastasis. Expert Rev Mol Med. 2022;24:e11. doi:10.1017/erm.2022.4

64. Rébé C, Ghiringhelli F. Interleukin-1β and Cancer. Cancers. 2020;12(7):1791. doi:10.3390/cancers12071791

65. Mizui M. Natural and modified IL-2 for the treatment of cancer and autoimmune diseases. Clin Immunol. 2019;206:63–70. doi:10.1016/j.clim.2018.11.002

66. Corrêa LH, Corrêa R, Farinasso CM, de Sant’Ana Dourado LP, Magalhães KG. Adipocytes and macrophages interplay in the orchestration of tumor microenvironment: new implications in cancer progression. Front Immunol. 2017;8:1129. doi:10.3389/fimmu.2017.01129

67. Becker Y. Molecular immunological approaches to biotherapy of human cancers–a review, hypothesis and implications. Anticancer Res. 2006;26(2a):1113–1134.

68. Nunes NS, Chandran P, Sundby M, et al. Therapeutic ultrasound attenuates DSS-induced colitis through the cholinergic anti-inflammatory pathway. EBioMedicine. 2019;45:495–510. doi:10.1016/j.ebiom.2019.06.033

69. Ahmed U, Graf JF, Daytz A, et al. Ultrasound neuromodulation of the spleen has time-dependent anti-inflammatory effect in a pneumonia model. Front Immunol. 2022;13:892086. doi:10.3389/fimmu.2022.892086

70. Li BH, Jiang W, Zhang S, et al. The spleen contributes to the increase in PMN-MDSCs in orthotopic H22 hepatoma mice. Mol Immunol. 2020;125:95–103. doi:10.1016/j.molimm.2020.07.002

71. Li Y, Wang X, Cheng C, Jiang S, Ma T, Xu L. Absence of MHC class II molecules promotes natural killer cells activation in mice. Int Immunopharmacol. 2020;88:106888. doi:10.1016/j.intimp.2020.106888

72. Lee IC, Wu HJ, Liu HL. Dual-frequency ultrasound induces neural stem/progenitor cell differentiation and growth factor utilization by enhancing stable cavitation. ACS Chem Neurosci. 2019;10(3):1452–1461. doi:10.1021/acschemneuro.8b00483

73. Leng X, Shang J, Gao D, Wu J. Low-intensity pulsed ultrasound promotes proliferation and migration of HaCaT keratinocytes through the PI3K/AKT and JNK pathways. Braz J Med Biol Res. 2018;51(12):e7862. doi:10.1590/1414-431×20187862

74. Xing J, Singh S, Zhao Y, et al. Increasing vaccine production using pulsed ultrasound waves. PLoS One. 2017;12(11):e0187048. doi:10.1371/journal.pone.0187048

75. Komiyama K, Crago SS, Itoh K, Moro I, Mestecky J. Inhibition of natural killer cell activity by IgA. Cell Immunol. 1986;101(1):143–155. doi:10.1016/0008-8749(86)90193-0

76. Moţa G, Galatiuc C, Popescu I, Hirt M, Cialâcu V, Sulică A. IgA monoclonal and polyclonal proteins as regulatory factors of the NK cytotoxic activity. Rom J Virol. 1999;50(1–4):17–31.

77. Liu SH, Lai YL, Chen BL, Yang FY. Ultrasound enhances the expression of brain-derived neurotrophic factor in astrocyte through activation of TrkB-Akt and calcium-CaMK signaling pathways. Cereb Cortex. 2017;27(6):3152–3160. doi:10.1093/cercor/bhw169

78. Burks SR, Lorsung RM, Nagle ME, Tu TW, Frank JA. Focused ultrasound activates voltage-gated calcium channels through depolarizing TRPC1 sodium currents in kidney and skeletal muscle. Theranostics. 2019;9(19):5517–5531. doi:10.7150/thno.33876

79. Yoon CW, Lee NS, Koo KM, et al. Investigation of ultrasound-mediated intracellular Ca(2+) oscillations in HIT-T15 pancreatic β-cell line. Cells. 2020;9(5). doi:10.3390/cells9051129

80. Dautova Y, Kapustin AN, Pappert K, et al. Calcium phosphate particles stimulate interleukin-1β release from human vascular smooth muscle cells: a role for spleen tyrosine kinase and exosome release. J Mol Cell Cardiol. 2018;115:82–93. doi:10.1016/j.yjmcc.2017.12.007

81. de Water R, Leenen PJ, Noordermeer C, et al. Cytokine production induced by binding and processing of calcium oxalate crystals in cultured macrophages. Am J Kidney Dis. 2001;38(2):331–338. doi:10.1053/ajkd.2001.26098

82. Zhang S, Wan D, Zhu M, et al. CD11b + CD43 hi Ly6C lo splenocyte-derived macrophages exacerbate liver fibrosis via spleen-liver axis. Hepatology. 2023;77(5):1612–1629. doi:10.1002/hep.32782

83. Liu T, Fu Y, Shi J, et al. Noninvasive ultrasound stimulation to treat myocarditis through splenic neuro-immune regulation. J Neuroinflammation. 2023;20(1):94. doi:10.1186/s12974-023-02773-2

84. Hu X, Li F, Zeng J, et al. Noninvasive low-frequency pulsed focused ultrasound therapy for rheumatoid arthritis in mice. Research. 2022;2022:0013. doi:10.34133/research.0013

85. Wu L, Lin L, Qin YX. Enhancement of cell ingrowth, proliferation, and early differentiation in a three-dimensional silicon carbide scaffold using low-intensity pulsed ultrasound. Tissue Eng Part A. 2015;21(1–2):53–61. doi:10.1089/ten.tea.2013.0597

86. Xu P, Gul-Uludag H, Ang WT, et al. Low-intensity pulsed ultrasound-mediated stimulation of hematopoietic stem/progenitor cell viability, proliferation and differentiation in vitro. Biotechnol Lett. 2012;34(10):1965–1973. doi:10.1007/s10529-012-0984-6

87. Zhang X, Hu Z, Hao J, Shen J. Low intensity pulsed ultrasound promotes the extracellular matrix synthesis of degenerative human nucleus pulposus cells through FAK/PI3K/Akt pathway. Spine. 2016;41(5):E248–54. doi:10.1097/BRS.0000000000001220

88. Qu S, Worlikar T, Felsted AE, et al. Non-thermal histotripsy tumor ablation promotes abscopal immune responses that enhance cancer immunotherapy. J Immunother Cancer. 2020;8(1). doi:10.1136/jitc-2019-000200

89. Mauri G, Nicosia L, Xu Z, et al. Focused ultrasound: tumour ablation and its potential to enhance immunological therapy to cancer. Br J Radiol. 2018;91(1083):20170641. doi:10.1259/bjr.20170641

90. Dong W, Wu P, Zhou D, et al. Ultrasound-mediated gene therapy of hepatocellular carcinoma using pre-microRNA plasmid-loaded nanodroplets. Ultrasound Med Biol. 2020;46(1):90–107. doi:10.1016/j.ultrasmedbio.2019.09.016

91. Hoogenboom M, Eikelenboom D, den Brok MH, Heerschap A, Fütterer JJ, Adema GJ. Mechanical high-intensity focused ultrasound destruction of soft tissue: working mechanisms and physiologic effects. Ultrasound Med Biol. 2015;41(6):1500–1517. doi:10.1016/j.ultrasmedbio.2015.02.006

92. Izadifar Z, Babyn P, Chapman D. Mechanical and biological effects of ultrasound: a review of present knowledge. Ultrasound Med Biol. 2017;43(6):1085–1104. doi:10.1016/j.ultrasmedbio.2017.01.023

93. Cheung TT, Chu FS, Jenkins CR, et al. Tolerance of high-intensity focused ultrasound ablation in patients with hepatocellular carcinoma. World J Surg. 2012;36(10):2420–2427. doi:10.1007/s00268-012-1660-7

94. Hoogenboom M, Eikelenboom D, den Brok MH, et al. In vivo MR guided boiling histotripsy in a mouse tumor model evaluated by MRI and histopathology. NMR Biomed. 2016;29(6):721–731. doi:10.1002/nbm.3520

95. Zhou YF. High intensity focused ultrasound in clinical tumor ablation. World J Clin Oncol. 2011;2(1):8–27. doi:10.5306/wjco.v2.i1.8

96. Sheybani ND, Price RJ. Perspectives on recent progress in focused ultrasound immunotherapy. Theranostics. 2019;9(25):7749–7758. doi:10.7150/thno.37131

97. Curley CT, Sheybani ND, Bullock TN, Price RJ. Focused ultrasound immunotherapy for central nervous system pathologies: challenges and opportunities. Theranostics. 2017;7(15):3608–3623. doi:10.7150/thno.21225

98. Morton C, Cotero V, Ashe J, Ginty F, Puleo C. Accelerating cutaneous healing in a rodent model of type II diabetes utilizing non-invasive focused ultrasound targeted at the spleen. Front Neurosci. 2022;16:1039960.

Introduction

As a benign disease, keloids are featured by the overgrowth of dense fibrous tissue on the affected skin. Although benign, they often cause troublesome symptoms including itching, pain, and pruritus. Furthermore, their high recurrence rates make them particularly challenging in clinical practice.¹ Recommended treatments for keloids include compression therapy, intralesional injections, laser therapy, topical treatments, surgical excision, postoperative radiotherapy, and cryotherapy.² However, the nonspecific nature of these therapies increases the risk of damaging surrounding tissues. Intralesional corticosteroid injections, a standard first-line treatment, frequently result in side effects such as localized pain and changes in skin appearance.³ The combination of corticosteroids and 5-Fluorouracil has demonstrated potential in improving treatment outcomes, yet persistent side effects remain a concern.⁴ Furthermore, irregular post-operative radiotherapy, or intralesional injections, and high-risk sites have also been implicated in the recurrence of keloids.^5–7 Therefore, targeting the distinct genetic and protein expression profiles of keloids compared to normal skin necessitates the identification of novel therapeutic targets and corresponding ligands to advance keloid treatment.⁸

Previous studies on keloid therapy have primarily focused on the transforming growth factor-beta (TGF-β)/Smad signaling pathway,^9,10 which plays a crucial role in the prolonged stimulation of fibroblasts and myofibroblasts, leading to excessive collagen production in keloids.⁹ However, considering the involvement of TGF-β in multiple biological processes, therapies targeting this signaling may cause systemic cytotoxic effects, limiting the advancement of TGF-β-based treatments for fibrotic diseases.¹⁰ Therefore, further research is warranted to refine targeted therapies for keloids.

In our previous study, we identified several potential targets for keloid therapy through Mendelian randomization and single-cell sequencing analyses. However, these studies did not reveal any functional targets. Notably, proteins from the BAG family have attracted our interest due to their overexpression in both tumors and, to some extent, keloids, highlighting the tumor-like properties of these lesions.¹¹ Among them, BAG cochaperone 2 (BAG2) has emerged as a promising target for keloid therapy according to the single-cell sequencing analysis. Previous research has shown that the high BAG2 expression among tumor-associated fibroblasts correlated with the poor prognosis in breast cancer and its anti-apoptotic characteristics,^12,13 suggesting that BAG2 may play a similar role in keloid progression. Mechanistically, BAG2 negatively regulates the chaperone-associated ubiquitin ligase, C terminus of Hsc70-interacting protein (CHIP), which facilitates the ubiquitin-mediated degradation of misfolded proteins. This suggests that BAG2 may inhibit the degradation of specific proteins, including overexpressed collagen in keloids. Additionally, BAG2 interacts with the MAPK signaling pathway, influencing downstream cellular proliferation.^14,15 In keloids, BAG2 may inhibit collagen degradation and promote abnormal collagen accumulation, warranting further investigation.

This study aimed to validate the role of BAG2 in keloid progression and identify compounds targeting BAG2, employing high-throughput screening technology tailored for keloid patients. The experimental setup is illustrated in Figure S1. Additionally, we explored the intrinsic connection between BAG2 and the progression of keloids to provide insights for more precise keloid management.

Materials and Methods

Ethics Statement

Keloid biopsies (10 patients with 4 men and 6 women age ranging from 26 to 42 years old, with keloids from operative excision of previous abdominal incisionor chest incision, having the lesion of over one year without reduction, without previous surgical treatment or radiotherapy, diagnosed by 2 experienced clinical experts with lesions extending beyond the wound boundary into the normal skin and other clinical features) and normal skin samples (3 female patients, aged 38, 41 and 42, with skin excision from facial plastic surgery as brow lifting) were obtained from the Shanghai Jiao Tong University affiliated ninth people’s hospital in accordance with the institutional review board (SH9H-2024-TK561). All patients provided formal, informed and written consent to supply a biopsy for this study. The study complied with the Declaration of Helsinki.

Keloid Fibroblasts (KFs) Isolation and Culture

Ten keloid samples from four men and six women were obtained after surgical excision. The sterile samples were rinsed in PBS, cut into pieces, and digested with 0.2% collagenase IV for 4 hours at 37 °C. After centrifugation, the sedimentary cells were resuspended in DMEM supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 µg/mL streptomycin, and incubated at 37°C in a 5% CO₂ incubator. The spindle keloid fibroblasts (KFs) from at least three patients at passages 2 to 3 were mixed for usage in further experiments.¹⁶

Single-Cell Data Analysis

To explore BAG2 gene expression in different cell groups, we downloaded the single-cell dataset GSE181297 of keloid patients from the Gene Expression Omnibus (GEO) database.¹⁷ This dataset was generated via the Illumina NovaSeq 6000 platform for scRNA-seq including keloid and normal skin samples from human subjects. The data processing included normalizing, filtering low-quality cells, de-batching, selecting highly variable genes for dimensionality reduction, and analyzing BAG2 expression in different cell clusters.

Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR)

qRT-PCR was performed as previously described.¹⁸ Briefly, total RNA was extracted from cells after treatment using an RNA isolation kit (Takara Bio, Shiga, Japan). RNA purity was assessed by measuring the A260/A280 ratio, with acceptable values ranging from 1.8 to 2.0. The primers used for gene amplification are listed in Table 1. The results from three independent reactions were used to determine relative gene expression, normalized against β-actin expression.

Table 1 Primer Sequence

Cell Transfection

Small interference RNA (si-RNA) targeting BAG2 for BAG2 knockdown and its negative control (si-NC) were synthesized by Sangon Biotech (Shanghai, China). The sequence of si-BAG2 was listed in Supplementary Table 1. The non-transfected KFs were set as controls. Cell transfection was conducted using RNATransMate transfection reagents (Sangon Biotech, Shanghai, China) according to the manufacturer’s protocol.

Lentiviral Production and Infection

DNA oligonucleotides encoding primers for obtaining the target gene BAG2 (sequences: BAG2-1, 5’-AGGTCGACTCTAGAGGATCCCGCCACCATGGCTCAGGCGAAGATCAACGCTAAAG-3’, BAG2-2, 5’-TCCTTGTAGTCCATGGATCCATTGAATCTGCTTTCAGCATTTTG-3’), or a negative control were inserted into the lentiviral vector GV703 (Genechem, Shanghai, China). Human mRNA-BAG2 cDNA (synthesized by Genechem, Shanghai, China) was inserted into the BamHI/Age I sites of the lentiviral vector GV703. For lentiviral production, 293T cells were transfected with the lentiviral vector along with packaging plasmids using Lipofectamine 2000 (Life Science, USA) according to the manufacturer’s instructions. At 48 h and 72 h after transfection, culture media was collected, pooled and filtered. Human Fibroblast cell line (HHF1) obtained from The Institute of Cell Biology, Chinese Academy of Sciences was infected with the indicated lentivirus, and BAG2 expression was determined by WB at 48 h after infection.

Ex-Vivo Explant Culture of Keloid Tissue

After harvesting aseptic keloid tissue, and removing epidermis, the remaining dermis of keloids was cut into about 3×2 × 2 mm pieces by a scalpel. The dermal fragments were divided into different groups, and cultured for 3 days in DMEM containing 10% FBS as previously described.¹⁹ After tissue attachment, the medium was replaced for the control, siRNA-treated, and compounds-treated groups, which were then incubated for an additional five and seven days. The explants were collected after treatment.

Cell Proliferation Assay

The KFs transfected with si-BAG2 and si-NC (2×10³ cells/well) were seeded into a 96-well plate. Following the manufacturer’s protocol, 10 μL of CCK-8 reagent (Beyotime, Shanghai, China) was added to each well and incubated for 2 hours at 37°C. Subsequently, cell proliferation was assessed daily from day 1 to day 5. The number of cells was quantified by measuring absorbance at 450 nm using a microplate reader. Each treatment group was evaluated in triplicate to ensure reliable results.

Cell Cycle Analysis

KFs were treated with si-BAG2, C16-PAF (Targetmol, #T21547, 10 nM), or si-BAG2 alone for 24 and 48 hours, along with the si-NC control. After treatment, the cells were rinsed once with PBS, and fixed with 70% ethanol overnight. The cell cycle analyses were performed according to the instructions of the Cell Cycle Kit (Qihai Biotechnology, Shanghai, China), and flow cytometric analyses were performed by a flow cytometer (Beckman Coulter) equipped with ModiFit LT v2.0 software.

Cell Migration Assay

To evaluate the effect of BAG2 knockdown on the KFs migration, a scratch assay was performed. KFs were seeded in multi-well plates and cultured to confluence. The culture medium was then removed, and the cell monolayers were scratched using a 200 µL pipette tip. The cells were rinsed with PBS and subsequently treated with either si-BAG2, si-NC, or control. Images of cell migration were captured at 0 and 24 hours using an inverted light microscope (TE2000 Nikon, Japan) at 40×magnification at 0 and 24 hours. These experiments were performed in triplicate, and migration areas were quantified using ImageJ analysis software (National Institutes of Health, Bethesda, MD).

Histological, Immunohistochemical and Immunofluorescence Analyses

Keloid tissue and explant samples were fixed overnight in 4% paraformaldehyde at 4°C, embedded in paraffin, and sectioned to 5 μm thickness. The sections were subsequently stained with hematoxylin and eosin (H&E) and Masson’s trichrome for histological examination. Furthermore, keloid tissue sections were also treated with antibodies specific to BAG2 and α-SMA at dilutions from 1:2000 to 1:100. Antibody binding was visualized using 3,3’-diaminobenzidine (DAB) chromogen (Dako, Glostrup, Denmark) and counterstained with hematoxylin for immunohistochemical analyses. For immunofluorescence staining, after overnight incubation with primary antibodies, specimens were incubated with secondary antibodies for 1 hour. Cell nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI). Finally, digital images were obtained using a V10-ASW 4.2 computerized image analysis system (Olympus, Tokyo, Japan). The antibodies used were: BAG2 antibody (Affinity, #DF2650, 1:100), Goat Anti-Rabbit IgG (H+L) HRP (Jackson, #111-035-045, 1:200), α-SMA antibody (Proteintech, #67735-1-Ig, 1:100), Goat Anti-Mouse IgG (H+L) Red (Jackson, #115-295-003, 1:200), and Goat Anti-Rabbit IgG (H+L) FITC (Jackson, #111-095-003, 1:200). The positive rate of BAG2 was defined as the proportion of BAG2-positive cells among all the cells in the field.

Western Blot

Western blot was performed on tissue samples or cultured cells as indicated. Total protein was extracted using RIPA lysis buffer, as previously described.¹⁸ The protein concentration of each lysate was determined using a BCA protein assay kit. Protein samples were separated by SDS-PAGE electrophoresis and transferred onto polyvinylidene fluoride (PVDF) membranes. The membranes were then blocked with 5% skimmed milk and immunoblotted with specific primary antibodies, including COL1, COL3, α-SMA, MEK, p-MEK and β-actin, diluted between 1:2000 and 1:1000 in TBST. After overnight incubation with the primary antibodies and three subsequent TBST washes, the membranes were incubated with the appropriate secondary antibodies for 1 hour at room temperature. Protein bands were detected using an enhanced chemiluminescence (ECL) kit (Amersham Biosciences, Chalfont St. Giles, UK). The intensity of each protein band was normalized against the β-actin band for comparison. The primary antibodies used in this study were: COL3 (Abcam, #Ab184993, 1:1000), COL1 (Huabio, #HA722517, 1:1000), α-SMA (Affinity, #AF1032, 1:1000), BAG2 (Affinity, #DF2650, 1:1000), MEK (Affinity, #AF3385, 1:1000), p-MEK (Affinity, #AF6385, 1:1000), and β-actin (Affinity, #AF7018, 1:2000).

Surface Plasmon Resonance (SPR) for Affinity Screening and Affinity Determination

SPR experiments were performed in at 25 °C on a BIAcore T200 with CM5 sensor chips, and data were analyzed with BIAcore T200 Evaluation software (GE Healthcare), following the manufacturer’s instruction. BIAcore T200 optical biosensor was used to screen for BAG2 affinity and to measure equilibrium dissociation constant (K_D) values for protein–ligand interactions. A cell on the CM5 sensor chip was activated using a mixture of 200 μM 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide and 50 μM N-hydroxysuccinimide at a flow rate of 10 μL/min for 420 seconds. Subsequently, BAG2 protein (50 μL) was mixed with 180 μL of 10 mM sodium acetate solution at pH 5.0 and immobilized on the cell surface at the same flow rate and duration for two cycles as previously described.²⁰ The cell was then blocked using 1 M ethanolamine. A neighboring channel, used as a reference, underwent the same activation and blocking steps, except immobilization was carried out with PBS adjusted to pH 5.0. Both channels were equilibrated with PBS afterward. Grouped SPR single concentration screening: Fifty compounds were grouped, with each compound contributing 1 μL, combined in an EP tube to total 50 μL. EP tubes were labeled from 1 to 54, and 950 μL of PBS was added to achieve a screening concentration of 10 μM for each compound. The DMSO content was maintained at 5% within the correction range. Each run flowed at 10 μL/min for 150 seconds. After each run, the chip was regenerated for 5 minutes with 10 mM glycine-HCl (pH 2.0), repeating until all groups (54 in total) were tested. Sequential SPR single concentration screening: The two groups with the highest RU values from the first screening round (100 compounds total) were selected. Each compound was taken as 2 μL from the plate and mixed with 198 μL of PBS to achieve a screening concentration of 100 μM. Each run flowed at 10 μL/min for 150 seconds. At the end of each run, the chip was regenerated for 5 minutes with 10 mM glycine-HCl (pH 2.0) solution, repeating this process until all compounds (100 in total) were completed. Data were collected from the sample cell using BIAcore T200 Control software (v. 2.0, GE Healthcare) and normalized against the reference cell. Association and dissociation constants were calculated using global fitting to a 1:1 Langmuir binding model with BIAcore T200 Evaluation software (v. 2.0, GE Healthcare). Final figures were generated using Origin 7 software (v. 7.0552, OriginLab).

To determine the K_D value for protein-ligand interactions, each compound was diluted 6 times from 12 μM to 15.625 nM. The compounds were flowed over a chip with immobilized BAG2 protein from lowest to highest concentration at 10 μL/min for 150 seconds. After each concentration, the chip was regenerated with 10 mM glycine-HCl (pH 2.0) for 5 minutes at the same flow rate. Data were recorded in real time, with molecular weight adjustment and solvent correction applied to address non-specific binding and signal drift molecular effects. Data analysis was conducted using BIAcore T200 software (GE Healthcare) following the manufacturer’s guidelines.

Statistical Analysis

All data are presented as means ± SD and analyzed using GraphPad Prism version 8 software (GraphPad Software Incorporation, La Jolla, California, United States). The cell migration area data were analyzed using Image J software. Sample sizes were determined based on practical and experimental considerations, without predetermined statistical methods. Differences among multiple groups were analyzed using analysis of variance (ANOVA), followed by Tukey’s post hoc test for comparisons between groups (α = 0.05). Differences among two groups were analyzed using t-test. Statistically significant differences were indicated by asterisks, with a two-tailed p-value of less than 0.05 considered significant.

Results

Higher BAG2 Expression in KF Clusters Compared to Normal Skin and Higher BAG2 Protein Expression in Keloid Tissue

After quality control of the GSE181297 dataset, we conducted dimensionality reduction and clustering analysis. Unique molecular identifiers between 200 and 6000 were selected. Cell clusters were categorized based on original molecular markers. These clusters included endothelial cell (EC), fibroblast (FB), proliferating myofibroblast (MFB), keratinocyte (KC), immune cells (lymphocytes, macrophages, and mast cells), lymphatic endothelial cell (LEC), and neural cell (Figure S2A and B). Besides, elevated BAG2 gene expression in KFs was observed in differential expression analysis (Figure S2C), with no significant difference in BAG2 expression between ECs and LECs in keloid and normal skin samples. Immunohistochemical staining confirmed BAG2 expression around fibrous nodules (Figure 1A). Additionally, immunofluorescent staining showed colocalization of BAG2 with α-SMA-positive cells, while BAG2 expression was also observed in α-SMA-negative cells, consistent with single-cell sequencing results (Figure 1B). Western blot analysis further demonstrated that BAG2 protein levels were significantly higher in keloid tissue than in normal skin samples (Figure 1C). The images of individual channel of immunofluorescence staining of keloid samples were showed in Figure S3A. The negative controls of immunohistochemical staining for keloid and normal skin, and the Hematoxylin and Eosin Staining of the keloid tissue was showed in Figure S3B and Figure S4.

Figure 1 The expression of BAG2 in keloid and normal skin samples. (A) Immunohistochemical analysis reveals enhanced positive staining for BAG2 in and around the fibrous nodules of keloid tissues compared to normal skin samples. (Scale bar = 100 μm at 200× magnification; Scale bar = 50 μm at 400× magnification); (B) The statistical analysis of BAG2 protein expression in keloid tissues and normal skin samples, as detected by immunohistochemical staining; (C) Immunofluorescence staining of a keloid sample slide shows nuclear staining with DAPI (blue) and colocalization of BAG2 (green) with α-SMA (red) positive cells; BAG2 expression is also visible in α-SMA negative cells. (Scale bar = 100 μm); (D) BAG2 protein levels in keloid tissue were significantly elevated compared to normal skin samples. (K: Keloid; (N) Normal Skin). ***P<0.001.

Inhibition of BAG2 Reduced Collagen Synthesis and Deposition of KFs, and in an Ex-Vivo Model

The relative gene expression of collagen type I (COL1), which is associated with collagen synthesis and excessive deposition in keloids,²¹ and Tissue inhibitor of metalloproteinase 1 (TIMP1), which inhibits matrix metalloproteinases and subsequently promotes COL1 deposition in keloids,¹⁹ was significantly decreased following BAG2 inhibition (Figure 2A). Additionally, Western Blot analysis showed reduced protein expression of BAG2, COL1, collagen type III (COL3), and α-SMA in si-BAG2-treated KFs, highlighting the role of BAG2 in abnormal collagen deposition (Figure 2B). Consistent with these findings, collagen structures in si-BAG2-treated ex-vivo keloid explants appeared thinner and more degraded compared to the control group (Figure 2C). Furthermore, COL1 and COL3 protein levels were significantly reduced in ex-vivo explants treated with si-BAG2 (Figure 2D).

Figure 2 Inhibition of BAG2 reduced keloid collagen synthesis and deposition. (A) The relative gene expression of COL1A1 and TIMP1 in si-BAG2-treated keloid fibroblasts compared to control groups. Gene expression related to collagen synthesis and the inhibition of collagen degradation was significantly reduced following BAG2 inhibition; (B) Western blot analysis showing reduced protein expression of BAG2, COL1, COL3, and α-SMA in si-BAG2-treated keloid fibroblasts; (C) Masson’s trichrome staining of keloid explants from different groups. Collagen structures in the si-BAG2-treated groups appeared thinner compared to control groups; (D) Western blot analysis demonstrates decreasing protein expression of COL1 and COL3 in si-BAG2-treated ex-vivo keloid explants. The reduction in COL1 and COL3 protein expression was statistically significant. Statistical analysis of relative protein expression is provided (n=3). **P<0.01, **** P<0.0001.

Inhibition of BAG2 Reduced Migration and Proliferation of KFs, Correlating with the MEK Pathway

BAG2 inhibition significantly reduced the migration and proliferation of KFs (Figure 3A and B). Additionally, the relative gene expression of TGF-β, an indicator of the proliferation and migration of KFs, was significantly decreased in si-BAG2-treated KFs (Figure 3C).²² In addition, Western Blot further demonstrated decreased levels of phosphorylated MEK (p-MEK) following si-BAG2 treatment (Figure 3D). Furthermore, si-BAG2 treatment significantly increased KFs’ percentage within the G0/G1 phase, suggesting cell cycle arrest. This effect was reversed by C16-PAF, an MEK activator, which restored G0/G1 phase percentages in si-BAG2-treated KFs (Figure 3E).

Figure 3 Inhibition of BAG2 reduced KF migration and proliferation via the MEK signaling pathway. (A) Representative images and bar graph depict the migration of KFs treated with or without si-BAG2 at 0 and 24 hours after scratching (n=3). Scale bar = 500 μm; (B) CCK-8 assays were performed on control and si-BAG2-treated KFs, showing a significant reduction in the proliferation of si-BAG2-treated cells (n=3); (C) Relative gene expression of TGF-β in si-BAG2-treated and control KFs, with significantly decreased TGF-β expression observed following si-BAG2 treatment (n=3); (D) Western blot analysis revealing decreased p-MEK in si-BAG2-treated cells; (E) Representative cell cycle profiles of KFs from various groups, assessed by flow cytometry, accompanied by statistical analysis (n=3). *P<0.05, ** P<0.01, ***P<0.001, **** P<0.0001. KF, keloid fibroblast.

Screening of Compounds with High‑affinity to BAG2 via Surface Plasmon Resonance (SPR) and Their Cell Inhibition Rates

BAG2 protein was found to play a significant role in collagen deposition in keloid tissues and the proliferation of KFs, making it a potential target for keloid treatment. To explore compounds with high-affinity to BAG2, the structure of BAG2 predicted by AlphaFold3 is shown in Figure S5. SPR was employed as an affinity screening technology to identify potential high-affinity compounds for BAG2.

A total of 2,732 compounds from the FDA and anti-fibrosis libraries were screened (Supplementary Table 2). In the first round, 2 groups of compounds were identified, and 6 compounds were identified as putative BAG2 ligands in the second round (Figure 4A and B). The primary cell inhibition rates of these 6 compounds on KFs were determined via CCK8 assays on Day 3 (Supplementary Table 3). Three compounds (Saikosaponin B1, Bazedoxifene acetate, and Ponesimod) showed significant inhibition of KFs and were selected for further investigation. Subsequently, the affinity index (K_D values) of these compounds with BAG2 was determined using SPR, and their inhibitory effects on KFs were evaluated at varying concentrations. The results indicated that Saikosaponin B1, Bazedoxifene acetate, and Ponesimod had K_D values of 6.35E−7 M, 2.84E−6 M, and 5.87E−6 M, respectively. Additionally, the fitted curves demonstrating the effects of different compound concentrations on the relative cell viability of KFs revealed IC50 values of 9.32 μM for Bazedoxifene acetate and 24.96 μM for Ponesimod, indicating that both drugs exhibit strong binding affinity and inhibitory effects (Figure 4C and D). Although Saikosaponin B1 exhibited relatively high binding affinity to BAG2, its inhibitory effect on KFs was below 40% at a concentration of 50 μM (Figure S6), suggesting it is not an optimal candidate for BAG2 targeting. The association rate constants (Ka) and dissociation rate constants (Kd) of these compounds are listed in Supplementary Table 4. Overexpressing BAG2 in human fibroblasts, and adding Bazedoxifene (10μM) acetate or Ponesimod (25μM) for 48h, it was found that these two compounds could inhibit BAG2-mediated cell cycle transition and cell proliferation (Figure 4E). The overexpression of BAG2 was validated by Western blot (Figure S7).

Figure 4 SPR screening of high-affinity compounds with BAG2 and validation in vitro. (A) Initial screening of 54 compound groups via SPR identified two groups demonstrating high affinity for BAG2 (shown in blue), compared to the positive control, BAG2 rabbit monoclonal antibody (mAb) from ABclonal Technology (A8775) (shown in red); (B) A second screening via SPR of two groups comprising 100 compounds revealed six compounds with high affinity for BAG2 (shown in blue), using the same BAG2 rabbit mAb as a positive control (shown in red); (C-D) SPR assays were used to analyze the binding affinities of Bazedoxifene acetate and Ponesimod to human BAG2 protein. Additionally, dose-response curves displaying the response of KFs to Bazedoxifene acetate and Ponesimod treatment over 72 hours are presented (n=6); (E) Representative cell cycle profiles of human fibroblasts from control, BAG2 overexpressed (oe-BAG2), oe-BAG2 treated with Bazedoxifene acetate or Ponesimod groups, assessed by flow cytometry, accompanied by statistical analysis (n=3). **** P<0.0001. KF, keloid fibroblast.

Compounds with High-Affinity to BAG2 Exhibited Inhibition of Collagen Deposition in Keloid Tissue

To further evaluate the therapeutic effects of the identified compounds on keloids, the effects of Bazedoxifene acetate and Ponesimod were assessed through protein expression analysis of COL1 and COL3, as well as histological evaluation of ex-vivo keloid explants. The experiments revealed that treatment with Bazedoxifene acetate and Ponesimod for five days or longer significantly reduced both the protein expression of collagen and the histological deposition of collagen in the ex-vivo explants (Figure 5). Additionally, Bazedoxifene acetate showed a superior inhibitory capacity for the expression of COL1 and COL3. These findings confirmed that compounds with high-affinity to BAG2 could markedly inhibit collagen deposition in ex-vivo keloid explants, indicating their clinical potential as therapeutic agents for keloids.

Figure 5 Masson’s trichrome of keloid explants in different groups. (A) In ex-vivo keloid explants treated with Bazedoxifene acetate and Ponesimod, the collagen structures appeared thinner compared to control groups; (B) Western blot analysis showed reduced protein expression of COL1 and COL3 in ex-vivo keloid explants treated with Bazedoxifene acetate and Ponesimod. Notably, Bazedoxifene acetate demonstrated a superior inhibitory effect on collagen deposition; (C) Statistical analysis of the relative protein expression is included (n=3). *P<0.05, **P<0.01.

Discussion

Keloids, a benign fibrogenic skin disease, share numerous characteristics with tumors, including the absence of spontaneous regression, excessive proliferation, and high recurrence rates.²³ Consequently, it is critical to explore therapeutic agents that target specific mechanisms involved in keloid formation. In this study, we identified BAG2 as a novel therapeutic target and screened thousands of compounds to find high-affinity BAG2 ligands. Our findings include the identification of Ponesimod and Bazedoxifene acetate as high-affinity ligands to BAG2, whose inhibitory effects were evaluated on keloid explants.

The present study observed an upregulation of BAG2 in keloid tissues, meanwhile, decreasing BAG2 with siBAG2 inhibited the proliferation of KFs, which correlated with MEK signaling. Consistent with the alterations of TGF-β following BAG2 inhibition, TGF-β modulates cell proliferation through the MEK pathway via a Smad-independent signaling mechanism (Derynck et al, 2003). This is consistent with prior findings linking MEK pathway activation to the proliferation of KFs,²⁴ and to the proliferative and migratory behaviors of cancer cells.^25,26 As evidenced by our findings, si-BAG2 suppressed KFs’ progression at the G0/G1 phase, a process that could be reversed by MEK activation. This indicates that BAG2 facilitates KF proliferation, which is correlated with the MEK signaling pathway, as an early indicator of a long-term process involving ECM remodeling. In accordance with previous studies, inhibiting the MEK pathway could induce G0/G1 phase arrest.^27,28 Additionally, the initial phosphorylation site Ser20, located within the curled helical domain near the amino terminus of BAG2, is phosphorylated by MAPKAPK-2. The p38 MAPK-MAPKAPK-2-BAG2 phosphorylation cascade, activated in response to extracellular stress, influences cell proliferation activities,²⁹ and may explain BAG2’s role in activating the MEK pathway. The BAG2-Heat shock protein (HSP) complex maintains basement membrane integrity and regulates ECM and cell junctions in glial cells, suggesting a similar role in keloids.³⁰ BAG2 also has anti-apoptotic effects that promote cell survival and, in some cases, enhances TIMP synthesis and activity by boosting intracellular anti-apoptotic factors. Since TIMPs inhibit MMP activity to control ECM degradation, BAG2-mediated upregulation of TIMPs may reduce MMP-driven ECM breakdown, preventing excessive degradation.¹³

While the proliferation and migration of KFs are well-documented, accelerated collagen accumulation also plays a significant role in keloid pathology.¹⁹ Additionally, mechanical stress at keloid sites has been recognized as a risk factor for the initiation and progression of this condition.³¹ The stiffness of the extracellular matrix (ECM) in keloids, largely due to collagen accumulation, augments the effects of mechanical stress, which, in turn increases the production of ECM by KFs.³² Therefore, halting the collagen deposition cycle is vital for effective keloid therapy.³³ Traditional therapies for keloids fail to adequately target the essential pathological processes of collagen synthesis and degradation.³⁴ In this study, altering BAG2 activity reduced both collagen synthesis and deposition in vitro and ex vivo. Additionally, our findings demonstrate that drugs with high-affinity to BAG2 can significantly reduce collagen accumulation, highlighting their potential as novel targeted therapies for keloids that inhibit collagen deposition and reduce the viability of KFs. In pulmonary fibrosis, TGF-β treatment promotes α-SMA expression through acetylated CCAAT/enhancer binding protein β (C/EBP-β), and a similar mechanism may exist in keloids, which warrants further investigations.³⁵

Several skin diseases, including psoriasis, vitiligo, dermatitis, and skin cancers, have been successfully treated with targeted therapies.^36–39 Drugs such as vemurafenib and trametinib, which target the BRAF mutation and MEK respectively, have significantly advanced the field of targeted therapy, although many such drugs are still in the preclinical research stage. Specifically, drugs targeting keloids, due to their localized presence on the body surface, could offer more precise treatment options. Nevertheless, there are currently no specific inhibitors targeting BAG2, a keloid-associated marker identified through this study. In this study, several high-affinity BAG2 ligands with were identified, including Saikosaponin B1, Ponesimod, and Bazedoxifene acetate, and verified their capability to inhibit collagen expression and deposition in ex-vivo keloid tissue. Saikosaponin B1, a bioactive molecule found in Radix Bupleuri, has been shown to exhibit anticancer properties⁴⁰ and antifibrotic activity in liver fibrosis.⁴¹ Ponesimod, a Sphingosine 1-phosphate (S1P) modulator, has been investigated as a treatment for Multiple Sclerosis, focusing on its ability to regulate S1P activity in lymphocytes.⁴² Bazedoxifene acetate, a selective Estrogen Receptor modulator, was examined for its role as a BAG2 ligand and its capacity to inhibit the viability of KFs in this study.⁴³ Interestingly, tamoxifen, another Estrogen Receptor modulator, has also been shown to reduce keloid collagen fibers.⁴⁴ Like bazedoxifene acetate, tamoxifen’s impact on keloids appears to extend beyond simple Estrogen Receptor modulation. Bazedoxifene acetate appears to influence keloid pathology by correlating with the inhibition of BAG2. The mechanisms through which bazedoxifene acetate and ponesimod act on keloids necessitate further detailed exploration.

Despite the lack of prior reports on these drugs as keloid inhibitors, their strong affinity for BAG2 and effectiveness in inhibiting KFs suggest their potential as promising candidates for targeted keloid therapy. Bazedoxifene acetate, in particular, stands out for its lower K_D, lower IC₅₀, and higher binding affinity. As an FDA-approved medication, bazedoxifene acetate may manage antifibrotic activity by affecting both collagen deposition and the viability of KFs. Moreover, due to their relatively high SLogP values, as detailed in Supplementary Table 5, bazedoxifene acetate and ponesimod were more likely to facilitate transdermal absorption, potentially reducing side effects associated with systemic administration. Topical application of these compounds may provide a viable therapeutic option for keloid treatment.

Despite the promising potential of drugs targeting BAG2 in keloid therapy, treatment with these agents may induce unintended off-target effects on other proteins. Therefore, further functional analyses and the development of systems to predict potential side effects are warranted. In this study, we found that inhibiting BAG2 could reduce collagen deposition. Additionally, the system of MMPs and TIMPs was also affected, warranting further investigation. Furthermore, the direct interaction between BAG2 and MEK signaling requires further investigation to elucidate the role of BAG2 in keloid formation. In addition, proliferation is a confounding factor for the scratch assay, the effect of BAG2 in the migration of KFs should be further investigated. Further, in vivo models are warrant for validation of the role of BAG2 in keloids. Compared to the ex-vivo model, the in-vivo model provides more comprehensive information on physiological responses and systemic pathological changes. In addition, as indicated by SPR, the binding affinity of Bazedoxifene acetate and Ponesimod to BAG2 is relatively higher than that of the other tested drugs. To further validate the effect of Bazedoxifene acetate and Ponesimod on keloids via BAG2, silencing experiments should be conducted in future studies. In summary, this study highlighted BAG2’s impact on collagen deposition and the proliferation of KFs, which are critical to the pathogenesis of keloids, benign skin tumors with high recurrence rates. Additionally, we identified compounds with high-affinity ligands to BAG2 that could serve as potential therapeutic targets. Among the compounds tested, Bazedoxifene acetate demonstrated superior affinity for BAG2 and more effective inhibition of keloid fibroblasts and keloid tissue compared to other ligands, suggesting novel approaches for the precise management of keloid disease.

Conclusion

This study revealed the pathogenic role of BAG2 in keloid and identified its high-affinity ligands, Bazedoxifene acetate and Ponesimod. The therapeutic capabilities of these compounds demonstrated their potential to improve targeting therapy for keloids.

Resource Availability

The data underlying this article are available in the article and in its online supplementary material. Single cell sequencing data is available in [GEO database] at [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE181297]. Further data underlying this article will be shared on reasonable request to the corresponding author.

Acknowledgments

We thank all the patients who provided samples for the experiment for their support. This work was supported by National Nature Youth Foundation of China (Grant No. 82402938 & 82102319) Science and Technology Commission of Shanghai Municipality (Grant No. 22MC1940300) and Wuxi Taihu Lake Talent Plan, Supports for Leading Talents in Medical and Health Profession. The funding body did not play a role in the study design, data collection, analyses, interpretation, manuscript preparation, and in the decision to submit the manuscript.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors have declared that no conflict of interest exists. The patents associated with this article include patent application number 202411342096.X, credited to Yinmin Wang, Lin Lu, and Jun Yang.

References

1. Naik PP. Novel targets and therapies for keloid. Clin Exp Dermatol. 2022;47(3):507–515. doi:10.1111/ced.14920

2. Walsh LA, Wu E, Pontes D, et al. Keloid treatments: an evidence-based systematic review of recent advances. Syst Rev. 2023;12(1):42. doi:10.1186/s13643-023-02192-7

3. Kim SW. Management of keloid scars: noninvasive and invasive treatments. Arch Plast Surg. 2021;48(2):149–157. doi:10.5999/aps.2020.01914

4. Jiang ZY, Liao X-C, Liu M-Z, et al. Efficacy and safety of intralesional triamcinolone versus combination of triamcinolone with 5-fluorouracil in the treatment of keloids and hypertrophic scars: a systematic review and meta-analysis. Aesthetic Plast Surg. 2020;44(5):1859–1868. doi:10.1007/s00266-020-01721-2

5. Ekstein SF, Wyles SP, Moran SL, et al. Keloids: a review of therapeutic management. Int J Dermatol. 2021;60(6):661–671. doi:10.1111/ijd.15159

6. Deng K, Xiao H, Liu X, et al. Strontium-90 brachytherapy following intralesional triamcinolone and 5-fluorouracil injections for keloid treatment: a randomized controlled trial. PLoS One. 2021;16(3):e0248799. doi:10.1371/journal.pone.0248799

7. Katano A, Minamitani M, Yamashita H. Risk factors for local recurrence of keloids and hypertrophic scars after postoperative electron beam radiotherapy. J Cancer Res Ther. 2024;20(1):163–166. doi:10.4103/jcrt.jcrt_1861_22

8. Deng CC, Hu Y-F, Zhu D-H, et al. Single-cell RNA-seq reveals fibroblast heterogeneity and increased mesenchymal fibroblasts in human fibrotic skin diseases. Nat Commun. 2021;12(1):3709. doi:10.1038/s41467-021-24110-y

9. Zhang T, Wang X-F, Wang Z-C, et al. Current potential therapeutic strategies targeting the TGF-β/Smad signaling pathway to attenuate keloid and hypertrophic scar formation. Biomed Pharmacother. 2020;129:110287. doi:10.1016/j.biopha.2020.110287

10. Peng D, Fu M, Wang M, et al. Targeting TGF-β signal transduction for fibrosis and cancer therapy. Mol Cancer. 2022;21(1):104. doi:10.1186/s12943-022-01569-x

11. Mariotto E, Viola G, Zanon C, et al. A BAG’s life: every connection matters in cancer. Pharmacol Ther. 2020;209:107498. doi:10.1016/j.pharmthera.2020.107498

12. Yoon CI, Ahn S-G, Cha Y-J, et al. Metastasis risk assessment using BAG2 expression by cancer-associated fibroblast and tumor cells in patients with breast cancer. Cancers. 2021;13(18):4654. doi:10.3390/cancers13184654

13. Hou M, Yue M, Han X, et al. Comparative analysis of BAG1 and BAG2: insights into their structures, functions and implications in disease pathogenesis. Int Immunopharmacol. 2024;143(Pt 1):113369. doi:10.1016/j.intimp.2024.113369

14. Arndt V, Daniel C, Nastainczyk W, et al. BAG-2 acts as an inhibitor of the chaperone-associated ubiquitin ligase CHIP. Mol Biol Cell. 2005;16(12):5891–5900. doi:10.1091/mbc.e05-07-0660

15. Sun L, Chen G, Sun A, et al. BAG2 promotes proliferation and metastasis of gastric cancer via ERK1/2 signaling and partially regulated by miR186. Front Oncol. 2020;10:31. doi:10.3389/fonc.2020.00031

16. Gu JJ, Deng -C-C, An Q, et al. Lactate promotes collagen expression, proliferation and migration through H3K18 lactylation-dependent stimulation of LTBP3/TGF-β1 axis in keloid fibroblasts. J Invest Dermatol. 2025. doi:10.1016/j.jid.2025.04.044

17. Shim J, Oh SJ, Yeo E, et al. Integrated analysis of single-cell and spatial transcriptomics in keloids: highlights on fibrovascular interactions in keloid pathogenesis. J Invest Dermatol. 2022;142(8):2128–2139.e11. doi:10.1016/j.jid.2022.01.017

18. Wang Y, Wang X, Zhou X, et al. Suppressive effect mediated by human adipose-derived stem cells on T cells involves the activation of JNK. Int J Mol Med. 2019;43(1):177–184. doi:10.3892/ijmm.2018.3953

19. Aoki M, Matsumoto NM, Dohi T, et al. Direct delivery of apatite nanoparticle-encapsulated siRNA targeting TIMP-1 for intractable abnormal scars. Mol Ther Nucleic Acids. 2020;22:50–61. doi:10.1016/j.omtn.2020.08.005

20. Ma T, Tian X, Zhang B, et al. Low-dose metformin targets the lysosomal AMPK pathway through PEN2. Nature. 2022;603(7899):159–165. doi:10.1038/s41586-022-04431-8

21. Wang X, Ma Y, Gao Z, et al. Human adipose-derived stem cells inhibit bioactivity of keloid fibroblasts. Stem Cell Res Ther. 2018;9(1):40. doi:10.1186/s13287-018-0786-4

22. Luo Y, Wang D, Yuan X, et al. Oleanolic acid regulates the proliferation and extracellular matrix of keloid fibroblasts by mediating the TGF -β1/ SMAD signaling pathway. J Cosmet Dermatol. 2023;22(7):2083–2089. doi:10.1111/jocd.15673

23. Macarak EJ, Wermuth PJ, Rosenbloom J, et al. Keloid disorder: fibroblast differentiation and gene expression profile in fibrotic skin diseases. Exp Dermatol. 2021;30(1):132–145. doi:10.1111/exd.14243

24. Liu F, Yu T, Liu J, et al. IGFBP-7 secreted by adipose-derived stem cells inhibits keloid formation via the BRAF/MEK/ERK signaling pathway. J Dermatol Sci. 2023;111(1):10–19. doi:10.1016/j.jdermsci.2023.05.004

25. Zheng D, Wei Z, Zhang C, et al. ZNF692 promotes osteosarcoma cell proliferation, migration, and invasion through TNK2-mediated activation of the MEK/ERK pathway. Biol Direct. 2024;19(1):28. doi:10.1186/s13062-024-00472-3

26. Zhao H, Bo Q, Wu Z, et al. KIF15 promotes bladder cancer proliferation via the MEK–ERK signaling pathway. Cancer Manag Res. 2019;11:1857–1868. doi:10.2147/CMAR.S191681

27. Xu DQ, Yuan X-J, Toyoda H, et al. Anti-tumor effect of Huaier extract against neuroblastoma cells in vitro. Int J Med Sci. 2021;18(4):1015–1023. doi:10.7150/ijms.48219

28. Liu Y, Cheng Y, Huang G, et al. Preclinical characterization of tunlametinib, a novel, potent, and selective MEK inhibitor. Front Pharmacol. 2023;14:1271268. doi:10.3389/fphar.2023.1271268

29. Ueda K, Kosako H, Fukui Y, et al. Proteomic identification of Bcl2-associated athanogene 2 as a novel MAPK-activated protein kinase 2 substrate. J Biol Chem. 2004;279(40):41815–41821. doi:10.1074/jbc.M406049200

30. Coraggio F, Bhushan M, Roumeliotis S, et al. Age-progressive interplay of HSP-proteostasis, ECM-cell junctions and biomechanics ensures C. elegans astroglial architecture. Nat Commun. 2024;15(1):2861. doi:10.1038/s41467-024-46827-2

31. Ogawa R, Okai K, Tokumura F, et al. The relationship between skin stretching/contraction and pathologic scarring: the important role of mechanical forces in keloid generation. Wound Repair Regen. 2012;20(2):149–157. doi:10.1111/j.1524-475X.2012.00766.x

32. Hsu CK, Lin -H-H, Harn HI, et al. Caveolin-1 controls hyperresponsiveness to mechanical stimuli and fibrogenesis-associated RUNX2 activation in keloid fibroblasts. J Invest Dermatol. 2018;138(1):208–218. doi:10.1016/j.jid.2017.05.041

33. Feng F, Liu M, Pan L, et al. Biomechanical regulatory factors and therapeutic targets in keloid fibrosis. Front Pharmacol. 2022;13:906212. doi:10.3389/fphar.2022.906212

34. Huang C, Liu L, You Z, et al. Managing keloid scars: from radiation therapy to actual and potential drug deliveries. Int Wound J. 2019;16(3):852–859. doi:10.1111/iwj.13104

35. Ding H, Chen J, Qin J, et al. TGF-β-induced α-SMA expression is mediated by C/EBPβ acetylation in human alveolar epithelial cells. Mol Med. 2021;27(1):22. doi:10.1186/s10020-021-00283-6

36. Rendon A, Schäkel K. Psoriasis pathogenesis and treatment. Int J Mol Sci. 2019;20(6):1475. doi:10.3390/ijms20061475

37. Feng Y, Lu Y. Advances in vitiligo: update on therapeutic targets. Front Immunol. 2022;13:986918. doi:10.3389/fimmu.2022.986918

38. Brunner PM, Guttman-Yassky E, Leung DY. The immunology of atopic dermatitis and its reversibility with broad-spectrum and targeted therapies. J Allergy Clin Immunol. 2017;139(4s):S65–s76. doi:10.1016/j.jaci.2017.01.011

39. Guo W, Wang H, Li C. Signal pathways of melanoma and targeted therapy. Signal Transduct Target Ther. 2021;6(1):424. doi:10.1038/s41392-021-00827-6

40. Luo J, Wang J, Yang J, et al. Saikosaponin B1 and Saikosaponin D inhibit tumor growth in medulloblastoma allograft mice via inhibiting the Hedgehog signaling pathway. J Nat Med. 2022;76(3):584–593. doi:10.1007/s11418-022-01603-8

41. Dong H, Shao M, Dong H, et al. Nanodrug rescues liver fibrosis via synergistic therapy with H(2)O(2) depletion and Saikosaponin b1 sustained release. Commun Biol. 2023;6(1):184. doi:10.1038/s42003-023-04473-2

42. Kar SS, Gharai SR, Sahu SK, et al. The current landscape in the development of small-molecule modulators targeting sphingosine-1-phosphate receptors to treat neurodegenerative diseases. Curr Top Med Chem. 2024;24(28):2431–46.

43. Komm BS, Kharode YP, Bodine PVN, et al. Bazedoxifene acetate: a selective estrogen receptor modulator with improved selectivity. Endocrinology. 2005;146(9):3999–4008. doi:10.1210/en.2005-0030

44. Soares-Lopes LR, Soares-Lopes IM, Filho LL, et al. Morphological and morphometric analysis of the effects of intralesional tamoxifen on keloids. Exp Biol Med. 2017;242(9):926–929. doi:10.1177/1535370217700524