Category: 3. Business

(Bloomberg) — The pitch went like this: Good, safe drinking water has become such a scarce resource that Americans will pay to fill up jugs — 30 or 40 cents to the gallon — at dispensers all across the country.

Most Read from Bloomberg

Hundreds of investors bought thousands of units, believing in the vision laid out by Ryan Wear, founder of a startup called WaterStation Management. They plunked down $8,500 for each vending machine and then waited for the dispensers to throw off a steady stream of cash. Among those lured in was a product manager in Oregon in 2021 and then, several months later, a dentist in Illinois. Each bought dozens of water dispensers, which Wear’s team would install and operate.

What they didn’t know is that at least one of those Hylyte-branded machines wound up being sold to both of them: serial number 101962, wedged in between a liquor store and a yoga studio in a strip mall in the Los Angeles suburb of Torrance, according to court records. To make matters worse, a machine with that same serial number was also pledged as collateral to back WaterStation bonds that were sold in April 2022 to the investment bank Jefferies Financial Group Inc.

Serial number 101962, faded and rusted in pictures submitted to court, is now gone from that strip mall, WaterStation is in bankruptcy, and Wear, 49, is the target of legal action by federal prosecutors, the Securities and Exchange Commission, a state banking regulator, Jefferies and scores of small-time investors, all of whom claim the company’s business was largely an illusion.

“This case involves a massive Ponzi scheme,” Assistant U.S. Attorney Justin Rodriguez told a federal judge in Manhattan Wednesday after Wear pleaded not guilty to criminal charges. A few feet away sat Jordan Chirico, a former Jefferies fund manager who prosecutors allege also committed fraud, directing his fund to purchase more WaterStation bonds after Wear admitted many water machines didn’t exist. He, too, pleaded not guilty. Lawyers for both declined to comment.

Voluminous legal filings describe a business that drew in military veterans, stock traders, pharmacists, salespeople and retirees by leveraging growing fears about contaminated tap water and microplastics and offering a lucrative solution that’d churn out annual returns as high as 20%, year after year. That desire to make easy money, coupled with clever marketing and alleged diligence lapses, kept the sputtering business going until the money finally stopped flowing around June 2023, according to court documents.

“These schemes never arise in a vacuum,” said John Bender, a lawyer who is representing a committee of WaterStation creditors. “The tragic consequences of the WaterStation affair could have been avoided had it not been for a cabal of insiders and institutions that prioritized their greed over doing the right thing even if it meant devastating the lives of a lot of people.”

Earlier this month, the committee asked a judge overseeing the proceedings to rule that Wear’s business meets the legal definition of a Ponzi scheme — a type of operation that uses new money to pay returns of existing investors or other creditors, with promoters usually promising high returns for little risk. If granted, franchisees would likely get tax relief and advisers would have an easier time clawing back funds from entities that profited off WaterStation’s business, which “will lay the groundwork for future recoveries on behalf of victims of this Ponzi,” according to the committee’s filing.”

Machine Mismatch

Formed in 2016 in Everett, Washington, WaterStation Management sold upwards of 21,000 machines and raised more than $380 million over the course of some seven years. Growth was fueled by a handful of banks that issued loans backed by the US Small Business Administration, as well as about $100 million in bonds bought by a fund run by Jefferies. Wear claimed in a 2023 deposition that his business had more than 500 employees.

In actuality, Wear’s firm only deployed roughly 2,100 machines, many likely never existed at all and most of the money WaterStation raised paid other costs or payments to existing franchisees, according to court records. Many dispensers that do exist were sold to multiple buyers — like the one that was in Torrance — or were promised as Jefferies’ collateral. Serial numbers on other machines don’t correspond with addresses investors were given by the company, court papers indicate. Determining who owns each machine and who bears responsibility for the alleged scheme is being fought over in federal court.

To locate machines that were sold more than once, Bloomberg News analyzed thousands of serial numbers submitted in court by creditors with claims to WaterStation machines and more than a dozen investor lawsuits that have piled up. Restructuring advisers say in court papers that more than 10,000 water machines were sold to multiple purchasers.

Falling Behind

WaterStation had trouble paying franchisees for years before monthly payments stopped completely two years ago, according to Becky Yang O’Malley, a GlassRatner Advisory Services managing director and certified fraud examiner retained by a committee representing WaterStation franchisees and other creditors. Franchisees have sued WaterStation and Wear, who along with Chirico, was sued by the Jefferies fund, while banks have sued franchisees who have fallen behind on business loans, according to court records. Jefferies has also sued one lender, First Fed Bank, alleging it helped keep WaterStation afloat after becoming aware of the alleged fraud in order to prioritize repayment of debt it was owed.

Wear in a sworn statement in April 2024 said franchisees’ allegations that machines don’t exist were untrue and their claims of fraud “are baseless, inflammatory and false.” Although WaterStation had occasionally experienced cash-flow issues, the business was legitimate and profits derived from water machines “were historically paid to plaintiffs,” Wear said at the time.

A First Fed spokesman said the bank isn’t able to comment on specific aspects of Jefferies’ lawsuit “as this is an ongoing legal matter,” but that the lender did nothing wrong. The bank will be submitting a formal response to Jefferies’ complaint next month, “which will provide additional clarity at that time,” he said.

‘Financially Devastated’

Chirico, 41, has also denied wrongdoing. His lawyer has said Chirico is also a victim of the WaterStation fraud and that Jefferies has “tried to scapegoat our client for an alleged scheme that deceived him along with hundreds of other investors and major institutions.”

Jefferies’ 352 Capital fund, once managed by Chirico, filed a civil lawsuit against Chirico in New York state court after a federal judge in May dismissed an earlier complaint. The bond transactions and their risks “were no secret” to the firm and other institutions, Chirico’s lawyers said in a motion to dismiss the latest lawsuit. Chirico sought to protect the fund by removing Wear as manager and attempted “to stabilize the collateral so the possibility of a restructuring or refinancing could be explored,” according to his Aug. 14 motion.

Restructuring advisers face a daunting task of trying to return money to franchisees who face substantial losses after Wear’s businesses went bankrupt last year. The situation is worse for those who took out loans to buy machines because even though the business was an alleged fraud, franchisees are still responsible for the debt and certain banks have sued borrowers who have fallen behind on payments. Some investors contend banks that partnered with Wear’s business should have uncovered the alleged scheme earlier because they had access to machine lists with duplicate serial numbers.

“My family has been financially devastated by the WaterStation scheme,” one Indiana franchisee noted in a sworn statement. He said he spent $3.3 million on machines and took out loans from two banks to fund his investment, pushing his monthly loan payments to $35,000. He said WaterStation’s assurances that it would buy back machines and that the financing was “SBA-approved” made him believe the business was more profitable and secure than it actually was.

Bank Loans

WaterStation was listed on the SBA’s database of franchises eligible for agency-approved loans starting in 2018. It gained momentum two years later, when Wear hired former bank-loan officer Kevin Nooney to help forge ties with banks and build a financing program to boost machine sales. The arrangement brought in new investor cash as the pandemic triggered a plunge in interest rates that motivated Americans to pile into a raft of alternative investments during lockdowns.

First Fed and fellow regional lenders Unibank and Celtic Bank were among the institutions that provided the most financing to investors, according to papers filed by a committee representing WaterStation creditors. Nooney said in a 2024 court filing that one of his former colleagues knew First Fed’s vice president of commercial lending, and that he also had “long-standing personal relationships” with Unibank’s former chief credit officer and a former loan officer.

Unibank and Celtic participate in the SBA’s preferred lending program which lets private banks administer SBA-backed loans with minimal agency review. Preferred lenders approved 28,875 SBA loans worth nearly $30 billion in fiscal 2021, roughly 55% of all loans approved in the SBA’s flagship lending program, according to a 2022 congressional report.

Unibank and Celtic didn’t respond to requests for comment.

From the start, though, the machines that Wear’s business was built on never made enough money to pay investors or cover WaterStation’s other costs. Instead, Wear relied on investors’ money and other loans to pay returns he promised franchisees “and to perpetuate the illusion of a legitimate business,” according to Yang O’Malley’s report.

As new money rolled in, people who already purchased machines got payouts they thought were their cut of the money generated from the vending business, according to court documents. WaterStation paid out $31.5 million in investor returns in 2021, about double what it paid in 2020, and more than $44 million in 2022, according to Yang O’Malley’s report.

But cracks were already forming as soon as August 2021, when Nooney learned that WaterStation purchases could constitute a security, according to a complaint brought by Washington’s banking regulator in May. The company responded by altering how it pitched the opportunity and paid returns, and these changes had the effect of curtailing new purchases, the complaint said. A lawyer for Nooney didn’t return messages seeking comment.

There was another problem with Wear’s business. The company pitched its machines as a way to make passive income, even though SBA rules say the loans WaterStation benefited from can only be used to fund actively managed franchises, according to the complaint. The state regulator also said WaterStation exploited the SBA’s preferred lender program.

The SBA was “left in the dark” and relied on lenders to verify that funds for the WaterStation loans were being used for approved purposes, Washington authorities said.

Enter Jefferies

In need of fresh capital, Wear turned to the bond market. In 2022, a Jefferies hedge fund called 352 Capital purchased roughly $100 million in WaterStation bonds earmarked for machine purchases. The fund was run by Chirico, who had bought hundreds of machines prior to joining 352 Capital as portfolio manager, according to federal prosecutors. Chirico didn’t fully disclose to Jefferies his personal stake in WaterStation, according to the indictment, which he disputes.

The bonds have spawned a separate Jefferies lawsuit against First Fed, which the firm claims became aware in the summer of 2022 that many machines didn’t exist. The lender, a unit of First Northwest Bancorp, had serial numbers for machines purchased with loans it gave franchisees, as well as machines WaterStation claimed ownership of that served as collateral for the bonds, “and hundreds of machines appeared on both lists,” according to Jefferies’ suit.

First Fed has denied wrongdoing and last year sought a receiver to take over Wear’s business. In a July bankruptcy settlement, the bank also agreed to pay $2.87 million to creditors and make additional payments and concessions to benefit franchisees, according to court papers.

First Fed in a statement this month said the bankruptcy settlement will benefit creditors because the bank released claims against WaterStation as well as liens on properties owned by an affiliate company. Proceeds from those assets will “become available for ratable distribution” to creditors, the bank said.

‘Going to Jail’

As for Chirico, prosecutors allege he had “learned of serious issues at WaterStation” by the summer of 2023. Then, in a phone call the following January, Wear admitted that thousands of machines supporting the bonds didn’t exist. But instead of telling Jefferies, Chirico allegedly directed the fund to purchase more WaterStation bonds which Wear partly used to repay a debt to Chirico, according to the indictment.

At Wednesday’s arraignment, Rodriguez, the prosecutor, told a federal judge that a “lengthy recorded phone call” is among the evidence law enforcement collected along with messages from Wear’s business email account. An investor who was also on the recorded phone call told Wear this was the “largest franchise fraud case in the history of the United States” and that he was “going to jail,” according to the indictments.

As the legal process plays out, borrowers are still responsible for SBA loans even if they are victims of an alleged scam, said Paul Midzak, a lawyer who advises small business owners. However, borrowers like the WaterStation franchisees can raise the alleged fraud as a defense against their loans and challenge lenders in court, Midzak said. The SBA, meanwhile, can be slow in responding to borrowers and working with the agency can be “like dealing with a woolly mammoth,” he said.

An SBA spokesman directed Bloomberg News to the Department of Justice. The DOJ in its press release announcing the criminal charges this month said that the SBA’s Office of Inspector General was among the federal agencies that assisted law enforcement in the criminal investigation.

Outside court proceedings, the physical markers that remain of Wear’s business include the water dispensers that actually were deployed in places like Pahrump, Nevada, where a Hylyte machine sits near the Lakeside Casino & RV Park, serving water to travelers who stop at the casino for a $12 plate of steak and eggs before heading east to Las Vegas or west to Death Valley.

Another 4,000-plus machines are sitting idle in two dozen abandoned warehouses stretching from Everett to Missoula, Montana, and Fort Meyers, Florida. Liquidation firm TAGeX Brands was hired to inventory machines. It discovered that people have broken into the warehouses to harvest copper wiring from the walls and that some packaged food left inside has attracted rats, according to court papers.

The facilities “stand out as among the most disorganized warehouses I have encountered in my 38-year career,” TAGeX Chief Executive Officer Neal Sherman said in an August court filing. The machines lack insulation and pipes inside dispensers left in cooler climates often burst, making them impossible to sell.

“The water machines were in poor condition,” Sherman said, “and were poorly made.”

–With assistance from Nicola M White and Katanga Johnson.

Most Read from Bloomberg Businessweek

©2025 Bloomberg L.P.

Participants

We recruited 36 young participants from southern Taiwan through advertisements on the Internet and bulletin boards. We conducted a power analysis (G*Power 3.1.9.7¹², power = 0.95, effect size f = 0.25, α = 0.05, within-between subjects’ design, correlation among repeat measures = 0.5). The analysis indicated that a sample size of 36 participants would be sufficient to detect an estimated medium effect size. All participants were right-handed and without evidence of neurological or psychiatric disorders based on self-reports. This study was reviewed and approved by the Human Research Ethics Committee at National Cheng Kung University (NCKU), Tainan, Taiwan, R.O.C., authorized by the Ministry of Education, Taiwan. All experimental procedures and the informed consent were obtained from all the participants and were approved under Approval No. NCKU HREC-E-112-120-2. All research was performed in accordance with the relevant guidelines and regulations, including the Declaration of Helsinki. Upon completion of all experiments — including computerized span tasks conducted outside the MRI scanner and a visual discrimination task performed inside the MRI scanner — participants received compensation of 1,500 New Taiwan Dollars (NTD). Detailed demographic characteristics are presented in Table 1.

Span tasks outside an MRI scanner

We assessed participants’ span capacity (individual differences) using computerized span tasks developed by Stone and Towse (2015) in JAVA¹³. The tasks included three complex span tasks and corresponding simple span tasks in both the verbal and visuo-spatial domains.

Verbal domain

Operation span task (complex span)

The Operation Span task was chosen as the complex span task for the verbal domain (see Supplementary Figure S1). This task involved a repetitive sequence of memory and processing components. In each trial, participants were presented with an integer to memorize and recall in its original serial position at the end of the trial. Following each memory element (the integer), there was a processing phase where participants encountered a mathematical operation, such as ‘9 + 9 = 27’. They had to determine whether the presented answer was correct. Digits and operations were generated randomly for each trial, with digits ranging from 1 to 99. Each operation had an equal 50% chance of being correct, and the types of operations (multiplication, division, addition, subtraction) each had a 25% probability, ensuring a diverse range of operation types requiring both correct and incorrect responses. Digit span (simple span). Digit span corresponded to the simple span task for Operation Span. Essentially, it was Operation Span without the processing phase. Participants only needed to remember the digits and recall them in sequence at the end of the trial.

Visuo-spatial domain

Symmetry span task (complex span)

The Symmetry Span task is a type of visuo-spatial complex span task where participants were required to recall grid locations in a 4 × 4 grid in the correct serial order (see Supplementary Figure S2). Following the presentation of each To-Be-Remembered (TBR) grid, participants engaged in a processing operation where they judged whether the presented pattern was symmetrical along the vertical axis, using the left/right arrow keys. Patterns were displayed on an 8 × 8 grid for this assessment. The recall phase began once the required number of storage-processing elements had been completed for a trial. Participants were prompted to recall by presenting them with the 4 × 4 grid, allowing them to click on the boxes in the order they remembered seeing them. Upon selection, a box turned blue, helping participants keep track of their responses. Matrix span (simple span). The matrix span task served as the simple span counterpart to the symmetry span task. Its procedure closely mirrored that of the symmetry span task, except for omitting the processing element.

Rotation span task (complex span)

The Rotation Span task was another visuo-spatial complex span task (see Supplementary Figure S3). The To-Be-Remembered (TBR) stimuli consisted of arrows characterized by two features: length (long or short) and angle of rotation (0°, 45°, 90°, 135°, 180°, 225°, 270°, or 315°). Participants were tasked with remembering these arrows presented in their correct serial order during the storage phase. In this complex span task, the processing operation involved presenting participants with a letter (F, G, or R) that could appear in its standard form or as a mirror image, and it could also be rotated at one of the 45-degree angles. Participants had to mentally rotate the image to determine whether the letter was presented normally or as a mirror image, using the left/right keys for their judgment. During the recall phase, participants were shown a 2 × 8 grid displaying all 16 possible arrows. The top row displayed all the long arrows, while the bottom row displayed all the short arrows. Participants used the mouse to select the arrows they recalled seeing in the correct sequential order. Arrow span (simple span). The processing phase was omitted in the arrow span task, which served as the memory span equivalent to the rotation span task. Consequently, the arrow span task focused on the participant’s ability to remember the arrows in their correct serial positions.

Visual discrimination tasks with difficulty manipulation in an MRI scanner

Participants engaged in a visual discrimination task featuring four distinct conditions, as outlined in Fig. 1. The task was programmed using OpenSesame¹⁴. Inside the MRI scanner, the stimulus display was projected onto a mirror affixed to the head coil. The task design was adapted from difficulty manipulations reported by Crittenden and Duncan (2014). These tasks varied in cognitive demands, ranging from simple perceptual discrimination to complex rule-based tasks¹¹.

Each trial began with a uniform gray screen in the baseline 4-line (4 L) condition. Participants were required to select the one odd line out of four based on length, considered the baseline or least demanding condition involving simple perceptual discrimination. At the start of each trial, a small fixation cross appeared at the center of the screen for 200 ms. Subsequently, four vertical lines were briefly presented (100 ms), aligned along the middle of the screen with their midpoints distributed symmetrically on either side of the fixation cross (total width 8.3° visual angle). Among these lines, three were equal in length (13.4°), while the fourth was consistently 50% shorter (6.7°). Participants indicated the position of the shorter line by pressing the corresponding key on an 8-button response box (e.g., Fig. 1, leftmost line shortest, response with the left middle finger). Responses were recorded only within the time window of the fixation cross, which persisted for 1000 ms after the lines disappeared from the display. Following a response, there was a jittered interstimulus interval of 500 to 1500 ms before the onset of the subsequent trial.

The remaining conditions exhibited similarities with some alterations as described below. In the 8 L condition, similar to the 4 L condition but with eight lines instead of four, two additional vertical lines were presented on either side of the original four lines (total display width 16.7°; see Fig. 1). This required participants to use each hand’s little and ring fingers for a response. The 8 L condition increased the perceptual load by doubling the number of lines, thereby raising the cognitive demand compared to the 4 L condition.

In the fine discrimination (FD) condition, similar to the 4 L condition, there were still four lines, but the shortest line was reduced to only 10% shorter than the other three lines (see Fig. 1). Participants had to discriminate between lines of very similar lengths, necessitating fine perceptual judgments. This condition heightened the cognitive demand further by requiring precise discrimination of line lengths, adding to the perceptual load.

In the mapping switch (MS) condition, the stimulus-response mapping was modified from the natural one to the alternative illustrated in Fig. 1. Participants were still required to discriminate between lines of very similar lengths, making fine perceptual judgments. This condition was considered the most demanding, as participants had to select the right-most odd line but press the left-most button. This complex rule introduced higher-order cognitive processing involving working memory and response inhibition. It placed significant demands on executive functions and was generally regarded as the most challenging^15,16. Participants practiced this condition until their accuracy rate reached 70% or higher.

Trials were grouped into blocks, each dedicated to one task condition. Each block displayed a schematic similar to Fig. 1 indicating the upcoming condition in the middle of the screen for 2,000 ms. Following the cue, there was a 3,800 ms pause before the onset of the first trial. Each block consisted of 8 trials, with a total duration of 18,400 milliseconds. There was a 10-second interval between blocks. To maintain task engagement, the accuracy of each block was displayed at the end.

The experiment was divided into three scanning sessions, each separated by a 30-second break. Within each session, there were 20 task blocks, comprising five blocks for each of the four conditions (4 L, 8 L, FD, and MS). The sequence of blocks was arranged in a pseudorandom order.

Behavioral data analysis

Span task performance

For each span task, we calculated their Full-Trial Accuracy (FTA) score. In the case of simple span tasks, points were awarded only when all Target-to-Be-Remembered (TBR) stimuli within a trial were correctly recalled. The score for each trial was determined by the loading of that particular trial, and the sum of scores across all trials constituted the FTA score for that task. In the case of complex span tasks, the calculation method was similar to that of simple span tasks, with the distinction that points were awarded only for trials where the response to the processing component was correct.

Grouping participants based on a median split of the FTA score

Our study employed the complex span task total scores to perform a median split, grouping participants into high-span and low-span groups based on individual differences in span task performances. This decision was based on the higher complex span scores compared to simple span scores, making them more effective for distinguishing individual differences in span capacity. However, we acknowledge that this approach may inadvertently capture differences in perceptual abilities, as evidenced by significant differences in simple span performance between groups (p <.005; see Results). Complex span tasks, which involve both storage and processing components, are known to be more sensitive measures of working memory capacity compared to simple span tasks that only require storage¹⁷. Given that our participant sample consisted of young adults, the complex span tasks were particularly effective in capturing individual differences in working memory capacity, which is critical for examining PFC activation patterns under varying cognitive demands^18,19.

Additionally, using the median split method allowed for a clear and balanced division of participants into two groups, facilitating the analysis of how these differences associate with neural activation patterns during task performance²⁰. Specifically, using a median split ensured an equal distribution of participants into high- and low-span groups, facilitating balanced and well-built statistical comparisons. This approach also mitigated potential biases that could arise from uneven group sizes, enhancing the validity and reliability of our findings²¹. By employing a median split on complex span scores, we aimed to provide a clear delineation of how varying levels of individual differences impact neural activation patterns and cognitive control processes. However, we acknowledge that treating individual differences as a continuous covariate in a regression analysis might offer greater statistical power and a more profound understanding of its relationship with PFC activation. To address this, we conducted an additional regression analysis using complex span FTA scores as a continuous covariate, confirming that the observed patterns of PFC activation remained significant (p <.001; see Figure S4 in Supplementary Information).

Visual discrimination task performance

Behavioral performance (reaction time [RT] and accuracy) was measured separately for each of the four conditions (4 L, 8 L, FD, MS). Subsequently, we conducted a one-way analysis of variance (ANOVA) on the four conditions to determine if the behavioral performance replicated previous findings reported by Duncan and colleagues¹¹. To test the hypothesis regarding whether individual differences would be associated with primary contrasts between conditions of interest in task performance, we initially compared the 8 L, FD, and MS conditions with the 4 L condition to identify any increase in reaction time (RT) and/or decrease in accuracy resulting from manipulation difficulty. Tasks with higher task demands, such as increased perceptual load in the 8 L condition, finer discrimination in the FD condition, or complex rule mapping in the MS condition, require greater working memory resources. High-span individuals typically exhibit superior cognitive control in these contexts²². In contrast, low-span individuals exhibited broader neural recruitment, which may reflect the additional engagement of brain regions under increased task demands²³.

Subsequently, we contrasted the conditions of interest. We used the 4 L condition as a baseline to evaluate how the 8 L, FD, and MS conditions demanded greater performance costs. The three contrast pairs were as follows:

8 L–4 L (Perceptual Load): The primary difference lies in the number of items to be processed. The 8 L condition has double the items of the 4 L condition, increasing perceptual load and cognitive demand.

FD-4 L (Precision of Discrimination): While both conditions involve selecting an odd line, the FD condition requires finer perceptual judgments, thereby increasing cognitive load compared to the broader discrimination required in the 4 L condition.

MS-4 L (Complexity of Rule Application): The MS condition introduces a complex rule that reverses the typical response mapping, significantly increasing cognitive demands compared to the straightforward perceptual discrimination in the 4 L condition.

We then employed mixed-design repeated-measures 2 (high- and low-span groups) x 3 (paired-contrasts: 8 L–4 L, FD-4 L, MS-4 L) ANOVAs on RT and accuracy, respectively, to examine the effects of the three contrasted conditions and determine if they interacted across two groups of individuals with high versus low span performances. Following the initial statistical testing, post hoc analyses were conducted using the Holm correction to further explore and compare the significant differences identified among the conditions.

Imaging acquisition and analysis for the visual discrimination task

The imaging data was gathered utilizing a General Electric (GE) Discovery MR750 3 Tesla scanner (General Electric Medical Systems, Milwaukee, USA) equipped with a 32-channel receive-only phased-array head coil at the Mind Research Imaging Center, National Cheng Kung University. High-resolution structural images were obtained using a fast-SPGR sequence comprising 166 axial slices (TR/TE/flip angle 7.6 ms/3.3 ms/12°; field of view (FOV) 22.4 × 22.4 cm²; matrix size 224 × 224; slice thickness 1 mm). Functional EPI images were acquired through an interleaved T2* weighted gradient-echo planar imaging (EPI) pulse sequence (TR/TE/flip angle, 2000 ms/30 ms/78°; matrix size, 64 × 64; FOV, 22 × 22 cm²; slice thickness, 3 mm; voxel size, 3.4375 × 3.4375 × 3 mm). Each run comprised 368 volumes, with the initial eight being dummy scans discarded to mitigate T1 equilibrium effects.

fMRI imaging preprocessing

Functional imaging data were analyzed using the FMRIB Software Library (FSL)²⁴ software. The analysis process comprised several specific steps at the 1 st level: Initially, preprocessing involved correcting head motion artifacts using the Motion Correction FMRIB’s Linear Image Registration Tool (MCFLIRT)^25,26,27. Subsequently, the brain extraction tool eliminated non-brain tissue from the preprocessed MR images (BET²⁷. The FSL Motion Outliers tool^25,27 accessible at https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FSLMotionOutliers, was then employed to detect outlier volumes based on frame displacement between volumes (exceeding the 75th percentile + 1.5 times the interquartile range). The results of this process were utilized to reduce the influence of those volumes in subsequent analyses. Individual brain functional images underwent registration to the high-resolution T1 structural image via linear transformation, followed by registration of the individual structural image to the standard MNI152 template via linear transformation²⁸. The first-level General Linear Model (GLM) in FEAT tool^29,30,31 was then established, incorporating a 9 mm full-width half-maximum (FWHM) Gaussian kernel for spatial smoothing.

Statistical analysis: fMRI blocked analyses for the visual discrimination task

A general linear model (GLM) was used to estimate parameter values reflecting the mean difference between experimental conditions of the visual discrimination task. Contrasts were performed to identify the regions recruited more for the 8 L, FD, and MS conditions relative to the 4 L condition.

The start time of each condition block’s stimulus to the endpoint of the block was captured and used to generate the onset file. The onset files for the four conditions were incorporated into the model as EVs (explanatory variables) and convolved with the double gamma hemodynamic response function. The six head motion parameters and the motion outlier data obtained in the previous step were included as covariates in the model for control.

The second-level analysis integrated data from the three runs, and the results of the three paired-contrasts (8 L–4 L, FD-4 L, MS-4 L) obtained at the first level were averaged separately.

To investigate our hypotheses regarding the high- and low-span groups, the subject-level files of the two groups were compared by averaging them separately according to the three paired-contrasts. We further examined the significant clusters for the combinations with simple main effects based on the ANOVA results of behavioral data, separately contrasting high span > low span and low span > high span.

Whole-brain univariate analysis

All group-level analyses involved computing the activation level across the whole brain region for each participant and submitting each of those to a group-level t-test, treating the participant as a random effect. We identified clusters of activity that were significant at a cluster-level rate of 0.01, using a 3.1 z-threshold to define contiguous clusters³². Subsequently, the estimated significance level of each cluster (derived from Gaussian Random Field theory) was compared with the probability threshold³³. To account for potential confounding factors, weperformed a partial regression analysis to control for gender effects between the high- and low-span groups. This step aimed to isolate the specific contributions of individual differences observed in neural activation patterns between the two groups.

fMRI preprocessing and GLM for multivoxel pattern analysis (MVPA)

The fMRI data were preprocessed using Statistical Parametric Mapping (SPM) 12^34,35 implemented in MATLAB (The MathWorks, Inc., Natick, MA). The preprocessing steps included slice-time correction and realignment to correct head motion using a rigid-body transformation³⁶. The T1 image was co-registered to the mean EPI image, and then both the T1 image and functional volumes were normalized to the MNI template. All images were resliced to a 2 × 2 × 2 mm voxel size, resulting in a data cube of 79 × 95 × 79 voxels. The onset files, marked with the start times for three runs and four conditions, were input into the GLM model. Head motion parameters obtained from realignment were included as regressors. After preprocessing, we obtained the beta maps and SPM.mat files for subsequent MVPA analysis.

MVPA

MVPA was conducted using the Decoding Toolbox (TDT, version 3.999 F) implemented in MATLAB, following standard preprocessing procedures. A whole-brain searchlight analysis was performed to identify regions whose activation patterns allowed classification between higher difficulty conditions and the 4 L condition used as the baseline³⁷. The input data for the classifier were the beta maps obtained from preprocessing and GLM analysis, normalized to MNI standard space. For every voxel in the brain, a sphere with a radius of 5 voxels centered on that voxel was used to train and test a linear support vector machine (SVM) using leave-one-run-out cross-validation within each participant^38,39. The classification accuracy of each sphere was assigned to the center voxel, resulting in a subject-level accuracy map. Accuracy maps were then entered into group-level analysis to identify regions where decoding accuracy was significantly above chance (50%). To ensure independence between training and testing phases, cross-validation was performed within each participant’s functional space. This method avoids assumptions of voxel-wise anatomical correspondence across participants^38,40.

Introduction

Hepatocellular carcinoma (HCC) is a leading global health challenge, ranking as the sixth most common cancer and the third leading cause of cancer-related deaths worldwide. In China, the burden of HCC is particularly high, with over 370,000 new cases and 320,000 deaths annually.¹ The majority of patients are diagnosed at advanced stages, precluding surgical resection. Even for those eligible for surgery, the five-year recurrence rate exceeds 60%,² underscoring the urgent need for more effective systemic and multimodal therapeutic strategies.

Interventional therapies, such as transarterial chemoembolization (TACE) and hepatic arterial infusion chemotherapy (HAIC), remain pivotal in the management of unresectable HCC (uHCC).³ These techniques leverage the dual blood supply of liver, delivering targeted therapy directly to the tumor while sparing healthy tissue. Although effective in controlling tumor growth, these approaches often yield limited long-term benefits when used as monotherapies, highlighting the importance of combining them with systemic therapies to improve outcomes. Donafenib, a deuterated derivative of sorafenib, represents an important advancement in the systemic treatment of advanced HCC. As a multi-target tyrosine kinase inhibitor (TKI), donafenib inhibits several critical signaling pathways involved in tumor proliferation and angiogenesis, and has demonstrated improved pharmacokinetics, safety, and superior overall survival compared with sorafenib.^4,5 Notably, donafenib is the only molecular targeted drug that has demonstrated superior overall survival (OS) compared to sorafenib in a head-to-head trial.⁵

Recent studies have further established the value of combining locoregional and systemic therapies in advanced HCC. FOLFOX-based HAIC combined with targeted agents such as sorafenib or lenvatinib has been shown to significantly improve response rates, survival outcomes, and surgical conversion rates compared to monotherapy, with objective response rate (ORR) exceeding 40% and conversion rates up to 12.8% for HAIC plus sorafenib.^6,7 Similarly, large randomized trials have demonstrated that TACE combined with targeted agents, including sorafenib or lenvatinib, confers significant benefits in progression-free survival (PFS), OS, and ORR compared to monotherapy.^8,9 Further, triple therapy regimens incorporating HAIC, targeted agents, and immune checkpoint inhibitors (ICIs) have yielded even higher response and conversion rates in retrospective and prospective studies. A retrospective study showed that, compared with lenvatinib monotherapy, the combination of HAIC, lenvatinib, and toripalimab achieved a significantly higher ORR (59.2% vs 9.3%) and a higher surgical conversion rate (12.7% vs 0).¹⁰ Furthermore, Phase II clinical studies reported at ASCO 2022 demonstrated favorable efficacy of such triple regimens: the TRIPLET study of HAIC combined with apatinib and camrelizumab reported an ORR of 70.96%,¹¹ while another phase II study of HAIC combined with sintilimab and the bevacizumab biosimilar IBI305 in initially unresectable HCC achieved an ORR of 66.7% and a surgical conversion rate of 66.7%.¹²

Despite these encouraging advances, evidence from real-world clinical practice regarding the efficacy and safety of triple therapy combining interventional treatments, donafenib, and anti-PD-1 monoclonal antibodies in patients with unresectable HCC remains limited. Therefore, this study aims to supplement and expand the real-world evidence regarding the use of triple therapy in patients with unresectable HCC in Chinese patients.

Materials and Methods

Study Design

This study was conducted as a retrospective, real-world clinical investigation to evaluate the effectiveness and safety of donafenib combined with anti-PD-1 monoclonal antibodies and interventional surgery as a first-line therapy for patients with unresectable HCC. The study retrospectively included patients who received this combination therapy at Yunnan Cancer hospital between March 2022 and December 2023. This study adheres to the STROBE guidelines for the reporting of observational studies.¹³ Ethical approval for this study was obtained from the Ethics committee of Yunnan Cancer hospital (Approval No.: KYLX2025-01) and the need for individual patient consent was waived due to the retrospective nature of the study.

Study Population

Patients were eligible for inclusion if they were between 18 and 80 years of age, irrespective of gender, and had been diagnosed with HCC according to the 2024 Guidelines for Diagnosis and Treatment of Primary Liver Cancer¹⁴ or via histological or cytological confirmation. Additional inclusion criteria included the presence of unresectable HCC, no prior systemic or local therapy, and at least one measurable lesion defined by the modified Response Evaluation Criteria in Solid Tumors (mRECIST).¹⁵ Patients were required to have an Eastern Cooperative Oncology Group Performance Status (ECOG PS) score of 0–1, and a Child-Pugh score of ≤8. To ensure reliable data analysis, only patients with complete baseline data and at least one follow-up evaluation was included. Exclusion criteria were histological confirmation of fibrolamellar or sarcomatoid HCC or cholangiocarcinoma components, a Child-Pugh score >8 after hepatoprotective therapy, or an ECOG PS score ≥2. Patients with severe comorbidities affecting critical organs or systems or incomplete data records were also excluded.

Treatment Procedures

All patients included in this study received the triple therapy regimen consisted of interventional therapy (TACE or HAIC) combined with anti-PD-1 monoclonal antibody (tislelizumab or sintilimab) and donafenib. Donafenib was administered orally at an initial dose of 200 mg twice daily. The choice of interventional regimen (TACE or HAIC) was determined by the treating physician based on each patient’s clinical characteristics. The criteria for selecting TACE or HAIC were as follows: (1) HAIC was administered to patients with large, localized tumors or tumors accompanied by surrounding satellite lesions; it was also used as a subsequent local therapy following initial TACE and HAIC sessions, as well as for patients with diffuse hepatocellular carcinoma or those with portal vein tumor thrombus. (2) TACE was chosen for patients with well-demarcated, localized lesions exhibiting a rich arterial blood supply, or for those with multiple tumors adjacent to the liver capsule and a relatively low overall tumor burden but a high risk of tumor rupture. (3) Combined HAIC and TACE was employed in cases where numerous tumors were distributed across different liver lobes with blood supply from multiple arteries. In these situations, TACE was performed on the non-dominant feeding arteries, while HAIC was administered through the dominant feeding artery. The interval between interventional therapy and the initiation of anti-PD-1 monoclonal antibodies plus donafenib did not exceed one month. All treatment decisions were made in accordance with routine clinical practice and were tailored to optimize therapeutic outcomes for each individual patient.

Data Collection and Follow-Up

Data were retrospectively extracted from electronic medical records, including demographic details, clinical history, laboratory test results, imaging findings, and treatment outcomes. Tumor characteristics, including size, number, and vascular invasion, were recorded. Liver function was assessed using the Child-Pugh score, and disease staging was determined according to the BCLC staging system.

Follow-up assessments were conducted every 12 weeks until disease progression, initiation of a new anticancer therapy, withdrawal of consent, loss to follow-up, death, or study termination. Regular assessments included imaging evaluations (CT/MRI), laboratory tests and ECOG PS scoring. Safety follow-up included monitoring adverse events (AEs) and serious adverse events (SAEs) for up to 30 days after the last dose or until initiation of new therapy, with continued monitoring for unresolved events.

Outcomes

The primary endpoint was the objective response rate (ORR), defined as the proportion of patients achieving a complete response (CR) or partial response (PR) according to mRECIST criteria. Secondary endpoints included progression-free survival (PFS), overall survival (OS), and disease control rate (DCR; including CR, PR, and stable disease [SD]). Progression-free survival was defined as the time from initiation of combination therapy to documented disease progression or death. Patients lost to follow-up without confirmed radiological progression or death were censored at the date of last follow-up with no evidence of progression. Patients who were alive without tumor progression at the data cutoff date were censored at the date of the last imaging assessment. Overall survival was defined as the time from initiation of combination therapy to death from any cause; patients still alive at the data cutoff were censored at their last known follow-up. Safety was evaluated based on the incidence and severity of treatment-emergent adverse events (TEAEs), which were documented and graded according to the Common Terminology Criteria for Adverse Events (CTCAE) version 5.0.

Statistical Analysis

Descriptive statistics were used to summarize baseline demographic and clinical characteristics. Continuous variables, such as age, were described as means, medians, and standard deviations, while categorical variables were reported as frequencies and percentages. Survival analyses were conducted using the Kaplan-Meier method, with differences analyzed using the Log rank test. All variables with p < 0.2 during univariate analyses were included in multivariate analyses, which included a Cox regression analysis to identify factors independently associated with PFS and OS. The proportional hazards assumption was verified using Schoenfeld residuals andglobal tests. The missing data in this study were missing at random, attributable to loss to follow-up. All statistical tests were two-sided, with a significance level set at 0.05. Safety analyses were descriptive, focusing on the type, incidence, and severity of adverse events during the treatment period.

Results

Patient Characteristics

Between March 2022 and December 2023, 171 patients with advanced HCC were screened. A total of 82 patients were excluded due to prior alternative therapies, poor adherence, or incomplete data, resulting in 89 patients with uHCC being included in the final analysis (Figure 1). The median age of the cohort was 56 years (range, 34–79 years), and 87.6% were male. Most patients were classified as Barcelona Clinic Liver Cancer (BCLC) stage C (51.7%) or stage B (34.8%). The majority of patients (91.0%) had a Child-Pugh class A liver function, and 75.3% had an Eastern Cooperative Oncology Group performance status (ECOG PS) score of 0. Hepatitis B virus (HBV) infection was the predominant etiology, accounting for 83.1% of cases. Notably, 52.8% of patients presented with more than three lesions, and 82.0% had tumors with a maximum diameter greater than 5 cm. Extrahepatic metastases were observed in 11.2% of patients (Table 1). Among them, 88 patients were treated with tislelizumab and 1 patient with sintilimab. Regarding interventional modalities, 10 patients underwent hepatic arterial infusion chemotherapy (HAIC), 42 received transarterial chemoembolization (TACE), and 37 were treated with a combination of TACE and HAIC.

Table 1 Patient Demographics and Baseline Characteristics

Figure 1 Flowchart of patients selection. Poor adherence: the anti-PD-1 monoclonal antibody was used for less than two times or was not followed up regularly.

Treatment Efficacy

The combination therapy demonstrated substantial efficacy. The best overall response (BOR) rates were as follows: complete response (CR) in 14.6% of patients, partial response (PR) in 60.7%, and stable disease (SD) in 24.7%, leading to an ORR of 75.3% and a disease control rate (DCR) of 100% (Figure 2). The median time to response (TTR) was 2.3 months (95% CI: 2.0–3.2), and the median time to progression (TTP) was 12.0 months (95% CI: 10.9–13.9). The median duration of response (DOR) was 11.1 months (95% CI: 9.1–11.7). Additionally, 15.7% of patients underwent successful conversion surgery, with 57.1% achieving a major pathological response (MPR) and 40.0% achieving a pathological complete response (pCR) (Table 2).

Table 2 Treatment Efficacy Outcomes of Patients

Figure 2 Best Percentage Change from Baseline in the Sum of Diameters of Target Lesions This waterfall plot illustrates the best percentage change from baseline in the sum of diameters of target lesions among evaluable patients treated with the study regimen. Bars are color-coded to indicate the best overall response: complete response (CR, green), partial response (PR, Orange), and stable disease (SD, purple). The dashed lines at −30% and +20% denote the thresholds for partial response (PR) and progressive disease (PD), respectively, as defined by RECIST criteria. Negative values represent tumor shrinkage, while positive values reflect tumor growth.

Subgroup analyses revealed significant variations in the objective response rate (ORR) among different clinical and tumor-related characteristics (Figure 3). Patients with an ECOG PS of 0 exhibited a significantly higher ORR of 83.4% (95% CI: 72.5–91.5) compared to 50.0% (95% CI: 28.2–71.7) in those with a PS of 1 (χ² = 10.037, P = 0.002). Similarly, the BCLC stage significantly influenced ORR, with patients in stages A or B achieving 88.4% (95% CI: 74.9–96.1) compared to 63.0% (95% CI: 47.5–76.8) in stage C (χ² = 7.662, P = 0.006). Baseline alpha-fetoprotein (AFP) levels were also predictive of response, as patients with AFP ≤400 μg/mL demonstrated an ORR of 84.7% (95% CI: 71.1–93.7), which was significantly higher than the 69.2% (95% CI: 49.4–79.0) observed in those with AFP >400 μg/mL (χ² = 4.619, P = 0.032). The presence or absence of tumor thrombus further impacted ORR. Patients without tumor thrombus achieved an ORR of 85.4% (95% CI: 72.2–93.9), whereas those with tumor thrombus had a lower ORR of 63.4% (95% CI: 46.7–77.9) (χ² = 5.752, P = 0.016). Additionally, extrahepatic metastases were associated with significantly lower efficacy, as patients without metastases had an ORR of 79.7% (95% CI: 62.9–88.0), compared to 40.0% (95% CI: 12.1–73.8) in those with metastases (χ² = 5.551, P = 0.018). In contrast, no significant differences in ORR were observed when stratified by sex, age, HBV infection status, Child-Pugh grade, number of target lesions, or the maximum diameter of target lesions (P > 0.05 for all). These findings suggest that clinical characteristics such as ECOG PS of 0, BCLC stage A or B, AFP ≤400 μg/mL, and absence of tumor thrombus or extrahepatic metastases are strong predictors of superior therapeutic response in patients with unresectable HCC.

Figure 3 Subgroup Analysis of Objective Response Rate (ORR). Forest plot summarizing the subgroup characteristics analysis of objective response rate (ORR) in patients treated with triple therapy. Subgroups include sex, age, hepatitis B virus (HBV) infection status, Child-Pugh classification, Eastern Cooperative Oncology Group (ECOG) performance status (PS), Barcelona Clinic Liver Cancer (BCLC) stage, number of target lesions, maximum diameter of target lesions, baseline alpha-fetoprotein (AFP) levels, presence of tumor vascular invasion, extrahepatic metastases, and interval from diagnosis to treatment. Significant differences in ORR were observed in subgroups stratified by ECOG PS, BCLC stage, AFP levels, tumor vascular invasion, and extrahepatic metastases (P < 0.05). Data are presented as ORR percentages with 95% confidence intervals.

Long-Term Survival outcomes

The median follow-up duration was 13.7 months (range: 3.6–27.5). The median PFS was 18.5 months (95% CI: 15.0–NA), while the median overall survival (OS) had not yet been reached. The PFS rates at 6, 12, and 18 months were 87.6%, 72.4%, and 52.7%, respectively, and the corresponding OS rates were 93.3%, 81.6%, and 72.4% (Figure 4). Univariate survival analyses using Kaplan-Meier curves and Log rank tests revealed that several clinical factors were significantly associated with survival outcomes. As shown in Supplementary Figure 1, patients with ECOG PS 1 had significantly worse OS than those with ECOG PS 0 (P = 0.0055), and patients with baseline AFP ≥400 ng/mL also had worse OS than those with AFP <400 ng/mL (P = 0.0012). Supplementary Figure 2 further demonstrated that PFS was significantly shorter in patients with ECOG PS 1 compared to those with ECOG PS 0 (P = 0.024), and in patients with extrahepatic metastases compared to those without (P = 0.044). The detrimental effect of ECOG PS 1 and elevated AFP on OS was consistently observed across analyses. Importantly, as shown in Supplementary Figure 3, there were no statistically significant differences in PFS or OS among patients who received different interventional modalities (TACE, HAIC, or the combination of TACE and HAIC). The median PFS was 13.0 months for HAIC, not reached for TACE, and 15.0 months for the combination group (P = 0.16). The median OS was 16.3 months for HAIC, not reached for TACE, and not reached for the combination group (P = 0.063).

Figure 4 Kaplan-Meier Curves for Progression-Free Survival (PFS) and Overall Survival (OS). (A) The Kaplan-Meier curve illustrates the progression-free survival (PFS) of patients treated with the combination therapy of donafenib, Anti-PD-1 monoclonal antibodies, and TACE/HAIC. The median PFS was 18.5 months (95% CI: 15.0–NA). (B) The Kaplan-Meier curve shows the overall survival (OS) of the same cohort. The median OS was not reached during the follow-up period. Shaded areas represent 95% confidence intervals. The number at risk is shown below each time point.

Prognostic Factors

The proportional hazards assumption was satisfied for all covariates (Supplementary Tables 1 and 2), and visual inspection of Schoenfeld residual plots confirmed no systematic deviation from proportionality over time (Supplementary Figures 4 and 5). In univariate Cox regression analysis, several factors were found to be significantly associated with poorer PFS, including a higher number of target lesions (>3 vs ≤3; HR: 2.06, 95% CI: 1.03–4.13, P = 0.041), ECOG performance status (PS) of 1 versus 0 (HR: 2.19, 95% CI: 1.09–4.42, P = 0.028), and the presence of extrahepatic metastases (HR: 2.41, 95% CI: 1.00–5.83, P = 0.051). For overall survival (OS), univariate analysis showed that a higher number of target lesions (>3; HR: 2.58, 95% CI: 1.00–6.65, P = 0.05), ECOG PS of 1 (HR: 3.23, 95% CI: 1.35–7.72, P = 0.008), the presence of extrahepatic metastases (HR: 3.32, 95% CI: 1.20–9.15, P = 0.02), and baseline AFP ≥400 ng/mL (HR: 4.62, 95% CI: 1.67–12.77, P = 0.003) were associated with worse outcomes (Table 3).

Table 3 Univariate and Multivariate Regression Analysis of Progression-Free Survival (PFS) and Overall Survival (OS)

In multivariate Cox regression analysis, ECOG PS of 1 (HR: 2.51, 95% CI: 1.14–5.51, P = 0.022) and the presence of extrahepatic metastases (HR: 2.66, 95% CI: 1.04–6.79, P = 0.04) remained independent prognostic factors for shorter PFS. For OS, ECOG PS of 1 (HR: 2.78, 95% CI: 1.08–7.17, P = 0.034) and baseline AFP ≥400 ng/mL (HR: 5.21, 95% CI: 1.81–14.98, P = 0.002) were identified as independent predictors of poorer survival (Table 3). Although extrahepatic metastases and the number of target lesions were significant in univariate analysis for OS, they did not retain significance in the multivariate model. These findings highlight the prognostic value of baseline performance status, tumor burden, and AFP level in patients with unresectable HCC treated with triple therapy.

Safety

TEAEs occurred in 97.8% of patients, with 30.3% experiencing grade 3 or higher TEAEs. The most common TEAEs (≥10%) included elevated alanine aminotransferase (50.6%), hand-foot skin reaction (49.4%), elevated aspartate aminotransferase (47.2%), hyperbilirubinemia (44.9%), thrombocytopenia (37.1%), diarrhea (21.3%), rash (12.4%), and leukocytosis (11.2%). Severe TEAEs (≥3 grade) included thrombocytopenia (7.9%), elevated aspartate transaminase AST (6.7%), hyperbilirubinemia (5.6%), rash (5.6%), hand-foot skin reaction (4.5%), and elevated alanine transaminase (ALT) (3.4%). No treatment-related deaths or grade 5 events were reported (Table 4).

Table 4 Incidence of Treatment-Emergent Adverse Events (TEAEs)

Discussion

In this real-world retrospective study, triple therapy with interventional treatment, donafenib, and anti-PD-1 monoclonal antibodies demonstrated notable efficacy and safety in patients with uHCC. The ORR was 75.3% and DCR reached 100%. Median PFS was 18.5 months, while median OS was not reached after a median follow-up of 13.7 months. Conversion surgery was achieved in 15.7% of patients, with most resected cases showing major or complete pathological response. Grade ≥3 TRAEs occurred in 30.3% of patients, with no treatment-related deaths. These findings suggest that triple therapy can achieve sustained tumor control and provide surgical opportunities for a subset of patients with advanced HCC in real-world clinical practice.

This study’s findings are consistent with prior research demonstrating the efficacy of triple therapies involving TACE, TKIs, and ICIs in uHCC. The observed ORR of 75.3% and DCR of 100% align closely with Shang et al, who reported an ORR of 70.5% and a DCR of 95.1% in a multicenter study using TACE, donafenib, and ICIs.¹⁶ Similarly, Li et al demonstrated the benefit of adding ICIs to systemic therapies, with their TACE+donafenib+PD-1 inhibitors (TACE+DP) group achieving an ORR of 50.6%, significantly higher than the 41.4% observed in the TACE+donafenib (TACE+D) group.¹⁷ In terms of survival outcomes, this study observed a median PFS of 18.5 months, which significantly exceeds the 12.7 months reported by Shang et al and the 10.6 months achieved in Li et al’s TACE+DP group.^16,17 This may be attributed to a greater proportion of patients with Child-Pugh A liver function and ECOG PS 0, earlier tumor stage, and fewer with extrahepatic metastasis or elevated AFP. Individualized multidisciplinary management may have also contributed to these favorable results. Conversion surgery remains a critical endpoint for evaluating the potential curative impact of triple therapies. In this study, 15.7% of patients underwent conversion surgery, a rate slightly lower than the 19.7% reported by Shang et al.¹⁶ However, the pCR rate of 40% among resected patients in this study demonstrates effective tumor cytoreduction. Recent research using HAIC-based regimens reported conversion rates ranging from 12.8% to 38.9%, depending on the systemic agents used. For example, in a real-world study of lenvatinib, tislelizumab, and HAIC, the ORR reached 94.4%, with a conversion rate of 38.9%.^18,19

Other recent studies provide additional context for evaluating triple therapy strategies. Wu et al reported an ORR of 76.4% and a conversion rate of 52.7% with lenvatinib, camrelizumab, and TACE in a multicenter prospective study, highlighting the versatility of triple therapy regimens across different systemic agents.²⁰ Zhang et al observed an ORR of 53.6% and a PFS of 8.9 months with lenvatinib and Anti-PD-1 monoclonal antibodies alone, emphasizing the added value of integrating TACE to achieve higher response rates and extended survival outcomes.²¹ Furthermore, Qi et al, focusing on an adjuvant setting with donafenib, TACE, and tislelizumab, reported favorable results in reducing recurrence risk in high-risk patients following resection, further underscoring the utility of triple therapy across various disease stages.²² The safety profile observed in this study aligns with findings from related research. TEAEs ≥ Grade 3 occurred in 30.3% of patients, lower than the 36.1% reported by Shang et al and comparable to the 33.8% in Li et al’s TACE+DP group.^16,17 The absence of Grade 5 TEAEs in this study further supports the tolerability of triple therapy regimens, comparable to that observed in recent trials such as LAUNCH, where lenvatinib combined with TACE also demonstrated an acceptable safety profile alongside significant survival benefits.⁹

The triple therapy used in this study combines TACE, donafenib, and anti-PD-1 antibodies to integrate locoregional and systemic approaches for synergistic treatment of HCC. TACE induces tumor necrosis and immunogenic cell death, which enhances antigen presentation and immune activation.²³ Donafenib inhibits VEGF signaling, resulting in vascular normalization, improved immune cell infiltration, and reduced recruitment of immunosuppressive cells.^24,25 Anti-PD-1 antibodies restore T-cell function by blocking the PD-1/PD-L1 pathway.²⁶ The inhibition of VEGF further enhances anti-tumor immunity and works synergistically with PD-1 blockade, while TACE-induced hypoxia upregulates VEGF and PD-L1 expression, providing a mechanistic basis for this combination.^27–29 Overall, these three modalities complement each other at both local and systemic levels.

Multivariate analysis in our study demonstrated that ECOG PS 1, extrahepatic metastasis, and elevated baseline AFP (>400 ng/mL) were independent adverse prognostic factors for both PFS and OS among patients with unresectable HCC receiving triple therapy. This aligns with recent large-scale studies and prognostic models in this field. For example, multiple retrospective and multicenter studies have consistently found that patients with impaired performance status, the presence of extrahepatic spread, or high baseline AFP have significantly worse survival outcomes after triple or combination therapies.^30–32 Elevated AFP is recognized as a marker of aggressive tumor biology, while extrahepatic metastasis reflects advanced disease burden and is repeatedly associated with reduced benefit from locoregional and systemic therapies. Similarly, an ECOG PS above 0 indicates decreased functional reserve, which is linked to both lower treatment tolerance and poorer prognosis. These findings reinforce the importance of comprehensive risk assessment before initiating triple therapy, and suggest that individualized treatment strategies and more intensive monitoring may be needed for patients with these high-risk features in clinical practice.

In real-world clinical practice, patients with unresectable HCC often present with heterogeneous tumor burdens, necessitating individualized selection of local treatment modalities (such as TACE or HAIC) based on comprehensive clinical assessment. In this study, different interventional approaches (TACE, HAIC, or their combination) were employed, which may have introduced variability in the treatment procedures. However, subgroup analysis revealed no significant differences in treatment efficacy among the different interventional modalities (TACE vs HAIC vs TACE combined with HAIC) (P > 0.05; see Supplementary Figure 3). Therefore, despite variations in interventional approaches, the choice of local treatment modality did not appear to affect efficacy outcomes in this cohort.

This study has several limitations that should be acknowledged. First, as a retrospective analysis without a control arm, it is subject to inherent biases, including potential selection bias in patient inclusion, which may limit the generalizability of the findings. Importantly, the absence of a comparator arm (such as TACE plus TKI or TKI plus ICI) precludes direct comparison of efficacy and safety outcomes with other commonly used treatment regimens. As a result, it is not possible to determine whether the observed outcomes are superior to those of alternative therapeutic strategies, and efficacy claims should therefore be interpreted with caution. Second, the relatively small sample size might influence the robustness of the observed outcomes. Third, the current follow-up duration may not be sufficient to fully assess long-term efficacy and safety trends. Ongoing follow-up will be conducted to provide more comprehensive data on long-term efficacy and survival in the future. While the promising efficacy and safety profiles of the triple therapy were demonstrated, larger randomized controlled trials are needed to validate these findings and establish the definitive role of this approach in the management of unresectable HCC. Furthermore, the identification and validation of predictive biomarkers could refine patient selection and guide individualized treatment strategies, ultimately improving therapeutic outcomes and minimizing unnecessary toxicities. Future research should focus on addressing these gaps to optimize the clinical utility of this promising therapeutic approach.

Conclusion

This real-world study demonstrates that triple therapy with interventional treatment, donafenib, and anti-PD-1 antibodies achieves high response rates, favorable survival, and manageable safety in unresectable HCC. ECOG PS 1, extrahepatic metastasis, and elevated AFP independently predict poorer outcomes. These findings warrant further investigation to confirm the clinical utility of this approach and its potential role in future treatment strategies.

Data Sharing Statement

The datasets used and analyzed in this study are available from the corresponding author upon request.

Ethics Approval and Informed Consent

The study was approved by the Ethics committee of Yunnan Cancer hospital (Approval No.: KYLX2025-01) and complied with the Declaration of Helsinki 1975. Informed consent was waived due to the retrospective nature of the study, with assurance that data was either anonymized or kept confidential.

Funding

This study was funded by Yunnan Provincial Science and Technology Department Basic Research Special Project – General Program (Grant No.: 202301AY070001-096); Yunnan Provincial Science and Technology Department – Kunming Medical University Joint Special Project – General Program (Grant No.: 202501AT070128); Yunnan Province Science and Technology Talents and Platform Plan Project (Grant No.: 202305AF150067); The Scientific and technological innovation team building program of Yunnan Provincial Education Department (Grant No.: K1322149); and Yunnan Province Talent Support Program (Grant No.: CZ0096-901264).

Disclosure

The authors declare that there is no conflict of interest in this work.

References

1. Bray F, Laversanne M, Sung H. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. Ca Cancer J Clin. 2024;74:229–263. doi:10.3322/caac.21834

2. Akateh C, Black SM, Conteh L, et al. Neoadjuvant and adjuvant treatment strategies for hepatocellular carcinoma. World J Gastroenterol. 2019;25:3704–3721. doi:10.3748/wjg.v25.i28.3704

3. Si T, Huang Z, Khorsandi SE, Ma Y, Heaton N. Hepatic arterial infusion chemotherapy versus transarterial chemoembolization for unresectable hepatocellular carcinoma: a systematic review with meta-analysis. Front Bioeng Biotechnol. 2022;10:1010824. doi:10.3389/fbioe.2022.1010824

4. Keam SJ, Duggan S. Donafenib: first approval. Drugs. 2021;81:1915–1920. doi:10.1007/s40265-021-01603-0

5. Qin S, Bi F, Gu S. Donafenib versus sorafenib in first-line treatment of unresectable or metastatic hepatocellular carcinoma: a randomized, open-label, parallel-controlled Phase II-III trial. J Clin Oncol. 2021;39:3002–3011. doi:10.1200/JCO.21.00163

6. He M, Li Q, Zou R, et al. Sorafenib plus hepatic arterial infusion of oxaliplatin, fluorouracil, and leucovorin vs sorafenib alone for hepatocellular carcinoma with portal vein invasion: a randomized clinical trial. JAMA Oncol. 2019;5:953–960. doi:10.1001/jamaoncol.2019.0250

7. Zheng K, Zhu X, Fu S, Cao G, Li WQ. sorafenib plus hepatic arterial infusion chemotherapy versus sorafenib for hepatocellular carcinoma with major portal vein tumor thrombosis: a randomized trial. Radiology. 2022;303:455–464. doi:10.1148/radiol.211545

8. Kudo M, Ueshima K, Ikeda M, et al. Randomised, multicentre prospective trial of transarterial chemoembolisation (TACE) plus sorafenib as compared with TACE alone in patients with hepatocellular carcinoma: TACTICS trial. Gut. 2020;69:1492–1501. doi:10.1136/gutjnl-2019-318934

9. Peng Z, Fan W, Zhu B. Lenvatinib combined with transarterial chemoembolization as first-line treatment for advanced hepatocellular carcinoma: a phase III, randomized clinical trial (LAUNCH). J Clin Oncol. 2023;41:117–127. doi:10.1200/JCO.22.00392

10. He MK, Liang RB, Zhao Y, et al. Lenvatinib, toripalimab, plus hepatic arterial infusion chemotherapy versus lenvatinib alone for advanced hepatocellular carcinoma. Therapeut Adv Med Oncol. 2021;13:17588359211002720. doi:10.1177/17588359211002720

11. Zhang TQ, Geng ZJ, Zuo MX, Li JB. Camrelizumab (a PD-1 inhibitor) plus apatinib (an VEGFR-2 inhibitor) and hepatic artery infusion chemotherapy for hepatocellular carcinoma in Barcelona Clinic Liver Cancer stage C (TRIPLET): a phase II study. Signal Trans a Targeted Ther. 2023;8:413. doi:10.1038/s41392-023-01663-6

12. Liu D, Mu H, Liu C, et al. Hepatic artery infusion chemotherapy (HAIC) combined with sintilimab and bevacizumab biosimilar (IBI305) for initial unresectable hepatocellular carcinoma (HCC): a prospective, single-arm phase II trial. J clin oncol. 2022;40:4073. doi:10.1200/JCO.2022.40.16_suppl.4073

13. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370:1453–1457. doi:10.1016/S0140-6736(07)61602-X

14. Department of Medical Administration, National Health Commission of the People’s Republic of China. Guidelines for diagnosis and treatment of primary liver cancer (2024 edition). Chin J Gen Surg. 2024;33:475–530 doi:10.7659/j.issn.1005-6947.2024.04.001.

15. Lencioni R, Llovet JM. Modified RECIST (mRECIST) assessment for hepatocellular carcinoma. Seminars in liver disease. Seminars Liver Dis. 2010;30:52–60. doi:10.1055/s-0030-1247132

16. Shang C, Zhong J, Luo X, et al. Efficacy and safety of transarterial therapy combined with donafenib plus immune checkpoint inhibitors for unresectable hepatocellular carcinoma: a multicenter retrospective study. J Clin Oncol. 2024;42:e16133–e. doi:10.1200/JCO.2024.42.16_suppl.e16133

17. Li H, Wang J, Zhang G, et al. Transarterial chemoembolization combined donafenib with/without PD-1 for unresectable HCC in a multicenter retrospective study. Front Immunol. 2023;14:1277329. doi:10.3389/fimmu.2023.1277329

18. Tan K, He X, Zhang H et al. Chicago, 2023 June 2-6, 2023 ASCO Annual Meeting . Efficacy and safety of tislelizumab combined with lenvatinib and FOLFOX4-HAIC in conversion therapy of middle-advanced stage hepatocellular carcinoma (HCC): a real-world retrospective study. J Clin Oncol. 2023 ;41: e16137-e16137. doi:10.1200/JCO.2023.41.16_suppl.e16137

19. Dong W, Zhang S, Huo Z et al. Chicago, 2023 June 2-6, 2023 ASCO Annual Meeting, Lenvatinib in combination with PD-1 inhibitor and hepatic arterial infusion chemotherapy (HAIC) for patients with potentially resectable hepatocellular carcinoma: a retrospective analysis. J Clin Oncol. 2023;41:e16160-e16160: doi:10.1200/JCO.2023.41.16_suppl.e16160

20. Wu XK, Yang LF, Chen YF, et al. Transcatheter arterial chemoembolisation combined with lenvatinib plus camrelizumab as conversion therapy for unresectable hepatocellular carcinoma: a single-arm, multicentre, prospective study. EClinicalMedicine. 2024;67:102367. doi:10.1016/j.eclinm.2023.102367

21. Zhang W, Tong S, Hu B, et al. Lenvatinib plus anti-PD-1 antibodies as conversion therapy for patients with unresectable intermediate-advanced hepatocellular carcinoma: a single-arm, phase II trial. J ImmunoTher Cancer. 2023;11.

22. Qi W, Peng W, Qi X, Qiu Z, Wen T, Li C. TIDE: adjuvant tislelizumab plus donafenib combined with transarterial chemoembolization for high-risk hepatocellular carcinoma after surgery: protocol for a prospective, single-arm, phase II trial. Front Oncol. 2023;13:1138570. doi:10.3389/fonc.2023.1138570

23. Kroemer G, Galluzzi L, Kepp O, Zitvogel L. Immunogenic cell death in cancer therapy. Ann Rev Immunol. 2013;31:51–72. doi:10.1146/annurev-immunol-032712-100008

24. Viallard C, Larrivée B. Tumor angiogenesis and vascular normalization: alternative therapeutic targets. Angiogenesis. 2017;20:409–426. doi:10.1007/s10456-017-9562-9

25. Yang J, Yan J, Liu B. Targeting VEGF/VEGFR to Modulate Antitumor Immunity. Front Immunol. 2018;9:978. doi:10.3389/fimmu.2018.00978

26. Hato T, Goyal L, Greten TF, Duda DG, Zhu AX. Immune checkpoint blockade in hepatocellular carcinoma: current progress and future directions. Hepatology. 2014;60:1776–1782. doi:10.1002/hep.27246

27. Li L, Wen Q, Ding R. Therapeutic targeting of VEGF and/or TGF-β to enhance anti-PD-(L)1 therapy: the evidence from clinical trials. Front Oncol. 2022;12:905520. doi:10.3389/fonc.2022.905520

28. Montasser A, Beaufrère A, Cauchy F. Transarterial chemoembolisation enhances programmed death-1 and programmed death-ligand 1 expression in hepatocellular carcinoma. Histopathology. 2021;79:36–46. doi:10.1111/his.14317

29. Liu K, Min XL, Peng J, Yang K, Yang L, Zhang XM. The Changes of HIF-1α and VEGF Expression After TACE in Patients With Hepatocellular Carcinoma. J Clin Med Res. 2016;8:297–302. doi:10.14740/jocmr2496w

30. Liu DY, Li YN, Wu J-Y. Efficacy and safety of transcatheter arterial chemoembolization combined with lenvatinib plus anti-PD-1 inhibitors for hepatocellular carcinoma patients with extrahepatic metastases: a multicenter retrospective study. J Hepatocell Carcinoma. 2024;11:2339–2349. doi:10.2147/JHC.S480958

31. Zeng ZX, Wu JY, Wu J-Y, et al. The TAE score predicts prognosis of unresectable HCC patients treated with TACE plus lenvatinib with PD-1 inhibitors. Hepatol Internat. 2024;18:651–660. doi:10.1007/s12072-023-10613-x

32. Li X, Sun W, Ding X, Li W, Chen J. Prognostic model of immune checkpoint inhibitors combined with anti-angiogenic agents in unresectable hepatocellular carcinoma. Front Immunol. 2022;13:1060051. doi:10.3389/fimmu.2022.1060051

Introduction

Interstitial cystitis/bladder pain syndrome (IC/BPS) is a chronic bladder condition characterized by severe pelvic pain and frequent urination, significantly impairing patients’ quality of life.¹ The prevalence of IC/BPS varies widely, from 0.01% to 17.3%, with women exhibiting notably higher rates.^2,3 Currently, the European Society for the Study of Interstitial Cystitis (ESSIC) distinguishes two IC/BPS subtypes based on Hunner’s lesions: ESSIC Type 3 (with Hunner’s lesions) and ESSIC Types 1 and 2 (without Hunner’s lesions).⁴ Despite substantial research efforts, the etiology and pathogenesis of IC/BPS remain elusive, complicating accurate diagnosis and treatment.¹ Current clinical management focuses on pain control, symptom relief, and quality-of-life enhancement for patients.⁵ Several studies have demonstrated a close relationship between IC/BPS and chronic inflammation, glycosaminoglycan barrier deficiency, autoimmune dysregulation, uroepithelial dysfunction, infectious cystitis, and urinary cytotoxicity,^6,7 however, none of these factors fully explain the condition’s complex clinical profile or sufficiently inform treatment strategies. Thus, identifying specific mechanisms underlying IC/BPS and developing novel therapeutic options remain essential. Oxidative stress, characterized by an imbalance between the production and clearance of reactive oxygen species (ROS), is pivotal in numerous inflammatory disorders.^8–10 Recent findings indicate that elevated ROS levels may exacerbate IC/BPS progression by altering bladder permeability and impairing normal bladder function.^11,12 Nuclear factor E2-related factor 2 (NRF2), a critical modulator of oxidative stress, regulates the transcription of antioxidant protein-encoding genes, thereby mitigating oxidative damage.¹³ Among these genes, Heme oxygenase-1 (HO-1) and NAD(P)H dehydrogenase [quinone] 1 (NQO1) serve as key NRF2-targeted antioxidant proteins across various diseases.^14,15 Although evidence increasingly supports the roles of oxidative stress and NRF2 in IC/BPS development,^11,16 the precise mechanisms remain insufficiently understood, as previous studies primarily focused on alleviating symptoms in animal models through pharmacological NRF2 enhancement in vivo without thoroughly exploring the underlying mechanisms;¹⁷ consequently, this study investigates the molecular mechanisms of oxidative stress imbalance by collecting bladder mucosal specimens from IC/BPS patients.

The ubiquitin-proteasome system plays a pivotal role in modulating post-translational protein degradation and transport.¹⁸ E3 ubiquitin ligase dysfunction has been implicated in the onset and progression of numerous inflammatory diseases.^19,20 Tripartite motif-containing 21 (TRIM21), a RING E3 ubiquitin ligase within the TRIM protein family, has been linked to tumor progression through its disruption of tumor suppressor protein expression.²¹ Recent studies also indicate that TRIM21 regulates glucose uptake, glycolysis, and oxidative stress, underscoring its role in various inflammatory conditions.^22,23 Nonetheless, the involvement of ubiquitination modifications and TRIM21 in inflammatory bladder disorders, particularly IC/BPS, remains unclarified.

This study demonstrates that TRIM21-mediated NRF2 ubiquitination is essential for bladder epithelial cells to counter inflammation and preserve oxidative stress equilibrium. Additionally, our findings indicate that the N6-methyladenosine (M6A) modification recognized by IGF2BP2 stabilizes TRIM21 mRNA, thereby enhancing its expression. This research unveils a novel IGF2BP2/TRIM21/NRF2 regulatory axis in IC/BPS, proposing potential new therapeutic strategies for this condition.

Methods and Materials

Patients and Tissues

This study was conducted in accordance with the Declaration of Helsinki. With the approval of the Ethics Committee and informed consent from patients, we collected paired lesional and non-lesional bladder mucosa specimens from 16 IC/BPS patients (same individual per patient) at the Second Affiliated Hospital of Nanjing Medical University between 2020 and 2022 (2017KY-102). According to the guidelines of the American Urological Association, the study participants are female patients aged 18 to 70 who were diagnosed with IC/BPS in the absence of infection.^24–26 The inclusion criteria and exclusion criteria for IC/BPS patients were shown in Box S1 and full information about these patients are summarized in Table S1.

Establishment of Animal Cystitis Model

8–10 weeks female wild-type (WT) and gene knockout (KO) C57BL/6J mice were chosen for intraperitoneal injection of cyclophosphamide (CYP, Sigma, USA) at a dose of 150mg/kg body weight.²⁷ The control group was injected with an equal dose of saline solution. All mice were subjected to the next step of the experiment 24 hours after injection. All animal experiments strictly adhered to the Guidelines for the Ethical Review of Animal Welfare (GB/T 35892–2018) and were approved by the Animal Ethics Committee of Nanjing Medical University (IACUC-2210019).

Void Spot Assays (VSA)

Urine frequency and volume were studied using VSA 24 hours after CYP injection.²⁸ Each mouse was gently placed into a circular metabolic cage (Yuyan, Shanghai, China) padded with filter paper (Whatman No. 1). The metabolic cage was placed in a quiet room and the mice were provided with food and water during the experiment. The filter paper was recovered after 2 hours and dried before being imaged.

Measurement of Pelvic Pain Sensitivity

Pelvic pain sensitivity measurements were performed 24 hours after CYP injection. Each mouse was individually acclimatized in metal (Yuyan, Shanghai, China) with a porous bottom for at least 1 hour. Mechanical stimulation was applied vertically to the pelvic region near the bladder and using Dixon’s up and down method to calculate the 50% withdrawal threshold.²⁹ We selected 0.008g, 0.02g, 0.04g, 0.07g, 0.16g, 0.4g, 1g, and 2g fiber filaments (North Coast, America) as a series for the experiment, and abdominal retraction or movement of position or licking of the stimulated site after stimulation was considered as a single positive response. We started the test with 0.07g of fiber, and if the mice showed a negative response to fiber stimulation, they were stimulated with a stronger fiber, and if they showed a positive response, they were stimulated with a smaller fiber.

Urodynamic Measurements

Mice anaesthetized with isoflurane (2%) nebulized 24 hours after CYP injection. After opening the lower abdomen to expose the bladder of the mice, a puncture needle connected to a pressure transducer (Taimeng, Chengdu, China) and a micro syringe pump (Silugao, Beijing, China) was threaded into the bladder. Saline was injected into the bladder at a rate of 2 mL/h to simulate the process of bladder filling and contraction, and the mice’s micturition curves were recorded in real time by the software (Taimeng, Chengdu, China).³⁰

Cell Culture and Transfection

Human normal bladder epithelial cell line (SV-HUC-1) and HEK293T cells were obtained from the Cell Bank of the Shanghai Institute of the Chinese Academy of Sciences. All these cell lines were cultured in a medium containing 10% FBS (Gibco, USA) and 1% antibiotics (Gibco, USA). At the cellular level, we established an in vitro inflammatory model by stimulating SV-HUC-1 cells with lipopolysaccharide (LPS, Sigma,USA) at 1μg/mL for 24 hours.

After transfection with Lipomaster 3000 (Vazyme, Nanjing, China) for 48–72 hours, the cells were collected for the next experiments. All plasmids used in the study were generated by Genebay (Nanjing, China). The shRNA sequences used are shown in Table S2.

RNA Extraction and Real-Time qPCR (RT-qPCR)

Total RNA was extracted using the FreeZol kit (Vazyme, Nanjing, China) and was reverse transcribed into cDNA for the next RT-PCR using HiScript II SuperMix (Vazyme, Nanjing, China) according to the instructions. Expression of the corresponding genes was analyzed and calculated by the comparative threshold cycling (2-ΔΔCt) method after normalization with β-Actin. The primers used for RT-PCR are shown in Table S3.

Western Blot

Cells and tissues were lysed for proteins using RIPA lysis buffer containing PMSF and protease inhibitors (NCM Biotech, Suzhou, China). Extracted proteins were subjected to SDS-PAGE and then transferred to PVDF membranes (Millipore, USA), and the membrane was incubated with primary antibodies at 4°C. After incubation with secondary antibodies at room temperature, membranes were treated with an ECL kit (Vazyme, Nanjing, China) for visualization. The antibodies used in the study are shown in Table S4.

ROS, Apoptosis and GSH Assay

After the treated cells were collected by centrifugation, they were incubated for 20 min at 37°C (Beyotime, Shanghai, China). Cells were washed and resuspended in phosphate-buffered saline and cellular ROS levels were assessed by flow cytometry.

For cell apoptosis analysis, cells were stained with Annexin V-FITC and PI (Vazyme, Nanjing, China) according to the supplier’s protocol and were assessed by flow cytometry.

Relative GSH concentrations were measured spectrophotometrically at 412 nm using a GSH assay kit (Beyotime, Shanghai, China) according to the manufacturer’s instructions.

Immunoprecipitation (IP)

Cell supernatants lysed in IP Lysis Buffer (Beyotime, Shanghai, China) were incubated with the target antibody in a rotating culture at 4°C overnight. Protein A/G beads (Vazyme, Nanjing, China) were added and then continued to be spun at 4°C for 6 h. After thoroughly washing the A/G beads three times with wash buffer (1 mL), 1X SDS-PAGE buffer was added to the samples and incubated at 100°C for 10 minutes.

RNA Stability Assay

Briefly, 5 μg/mL of actinomycin D (Abmole, USA) was added to cells 48h after transfection with either plasmid, and RNA was collected at the indicated times. Specific RNA expression levels were analyzed and normalized to β-Actin.

Immunohistochemistry (IHC) and Immunofluorescence (IF)

Sections or cells were blocked with 5% BSA for 30 minutes. After incubation with the primary antibody overnight at 4°C, they were incubated with the secondary antibody for 1 h at room temperature, protected from light. IHC was performed using DAB staining, followed by staining of nuclei with hematoxylin, and IF was performed using DAPI to stain the nuclei. The antibodies used in the study are shown in Table S4.

Dot Blot Assay

In brief, the extracted RNA concentrations were diluted to 100ng/μL and 200ng/μL and added to a nylon membrane (Beyotime, Shanghai, China). It was then exposed to UV light for 30 min and blocked with 5% nonfat for 1h. When the membrane was incubated overnight with an M6A antibody, it was incubated in a secondary antibody for 1h and then exposed for imaging. Finally, the amount of total RNA was shown using 0.1% methylene blue.

Statistical Analysis

All data were analyzed using GraphPad Prism 9.0. Data are expressed as mean ± SD of at least three independent experiments. Differences between groups were measured by Student’s t-test (for two-group comparisons) or one-way ANOVA followed by Tukey’s post hoc test (for multi-group comparisons), with P < 0.05 considered statistically significant.

Results

NRF2 Protein Expression Is Significantly Decreased in the Lesional Bladder Mucosa of Patients with IC/BPS

To determine the association of NRF2 expression with IC/BPS, NRF2 levels were initially examined in the bladder mucosa of patients with IC/BPS. IHC and Western blotting analyses demonstrated a significantly lower expression of NRF2 in the lesional mucosa compared to non-lesional mucosa distal to the affected area (Figures 1A and B, S1A and S3A).

Figure 1 Effect of NRF2 expression on bladder function in the bladder inflammation model. (A and B) NRF2 protein levels in the lesional and non-lesional bladder mucosa of patients with IC/BPS (Each L and N constituted a paired sample.). (C) Urinary blots from WT and NRF2 KO mice, with each bright dot representing a urination event. (D) Diagram of the urodynamic examination, where each peak signifies a urination. (E) 50% withdrawal threshold of WT and NRF2 KO mice. (F) Representative bladder images. (G) HE staining of bladder tissues. ** P < 0.01.

In Mice and Cellular Models of Inflammation, Lower Levels of NRF2 Expression Clearly Correlate with Poorer Performance

NRF2 function in bladder inflammation was further evaluated in vivo by comparing NRF2 KO and WT mice post-modeling. VSA results indicated that 24 hours following CYP injection, NRF2 KO mice exhibited increased urination frequency and decreased urine volume (Figure 1C). Urodynamic measurements revealed a notably shorter bladder filling phase and reduced urination pressure in NRF2 KO mice compared to WT mice following CYP injection (Figure 1D). The 50% withdrawal threshold assessment indicated heightened sensitivity to pelvic stimulation in the modeled NRF2 KO mice, suggesting more pronounced pelvic discomfort (Figure 1E). Observations of the bladder further revealed that modeled NRF2 KO mice displayed increased redness and swelling relative to WT mice (Figure 1F), and HE staining showed not only increased swelling but also structural damage in KO mouse bladders post-CYP injection (Figure 1G).

To elucidate the role of NRF2 in bladder inflammation in vitro, an inflammatory model was developed by stimulating the bladder epithelial cell line SV-HUC-1 with lipopolysaccharide (LPS). Based on prior studies linking NRF2 with oxidative stress and apoptosis,^31,32 the effects of NRF2 knockdown and overexpression on related phenotypes were examined. Results indicated that NRF2 knockdown led to excessive ROS production and significantly reduced glutathione (GSH) levels in the bladder inflammation model, while NRF2 overexpression produced opposite effects (Figure 2A and B). Flow cytometry further revealed an increase in apoptosis upon NRF2 knockdown, whereas NRF2 overexpression reduced apoptotic cell proportions (Figure 2C). Western blotting analysis demonstrated that NRF2 overexpression enhanced HO-1 and NQO1 protein expression, thereby mitigating cellular inflammation and apoptosis, whereas NRF2 knockdown had the reverse impact (Figures 2D,E,S3B and C). Collectively, these results suggest that diminished NRF2 expression compromises its anti-inflammatory capacity at both cellular and organismal levels.

Figure 2 Effects of NRF2 expression on oxidative stress and inflammation in the cellular inflammation model. (A) ROS levels in cells transfected with vector, NRF2, shCtrl, or shNRF2. (B) Relative GSH/GSSG ratio of cells transfected as indicated. (C) Apoptosis rate of cells under the specified transfections. (D and E) Western blotting analysis of oxidative stress inflammatory and apoptotic indicators in SV-HUC-1 cells following indicated transfections and treatments. ** P < 0.01.

TRIM21 Interacts with NRF2

Given NRF2’s critical role in mouse and cellular models of bladder inflammation, further exploration of its specific molecular mechanisms was pursued. Unlike protein levels, NRF2 mRNA levels showed no significant difference between lesional and non-lesional bladder mucosa (Figure S1B), suggesting that the reduction of NRF2 in lesional bladder mucosa was not due to altered gene expression but likely stemmed from post-translational protein modification. Previous research has indicated that changes in NRF2 protein expression across various diseases are largely influenced by ubiquitination modifications,^33–35 though less is known regarding inflammatory bladder conditions. To probe this, NRF2 protein synthesis was inhibited in SV-HUC-1 cells using cycloheximide (CHX), and protein expression was evaluated via Western blotting analysis. Findings revealed decreased NRF2 protein expression, which significantly increased following MG132 treatment (Figure S1C), supporting the hypothesis that NRF2 is regulated post-translationally by the ubiquitin-proteasome system.

To identify potential ubiquitin ligases associated with NRF2 degradation, IP and mass spectrometry (MS) analyses were performed. Silver staining and MS identified TRIM21 as a possible direct interactor with NRF2 in bladder epithelial cells (Figure 3A and B). Co-IP and Western blotting analyses further validated this interaction (Figure 3C and D), which was confirmed by fluorescence co-localization (Figure 3E). Molecular docking predictions using the ZDOCK platform suggested that the SPRY domain of TRIM21 was essential for binding (Figure 3F).³⁶ To verify this, various truncated mutants of Myc-TRIM21 were designed (Figure 3G). Transfection of these truncated mutants into HEK293T cells, followed by co-IP experiments, demonstrated that the SPRY domain is indeed critical for the TRIM21-NRF2 interaction (Figure 3H).

Figure 3 TRIM21 interacts with NRF2. (A) Silver staining of SDS-PAGE was performed through IP assays using either IgG or NRF2 antibodies. Red colored items: silver-stained protein bands (B) MS analysis confirmed the interaction between TRIM21 and NRF2. y-type ions: C-terminal fragments (blue peaks); b-type ions: N-terminal fragments (red peaks); Modified fragments: y+-H2O: Water loss peaks; y+-NH₃: Ammonia loss peaks. (C and D) Cell lysates from SV-HUC-1 cells were subjected to IP using antibodies against TRIM21 and NRF2. (E) IF images demonstrate the colocalization of TRIM21 and NRF2 in SV-HUC-1 cells. (F) Structural domains illustrating the binding of TRIM21 and NRF2 are shown through molecular docking analysis. (G) Schematic representation of full-length (FL) Myc-labeled TRIM21 along with various deletion mutants. (H) HEK293T cells were co-transfected with the indicated plasmids, and cell lysates were analyzed by IP using Myc beads, followed by Western blotting analysis with specific antibodies.

TRIM21 Regulates the Ubiquitination of NRF2

To investigate TRIM21’s role as a ubiquitinase affecting NRF2 ubiquitination, HEK293T cells were transfected with Myc-tagged wild-type or ΔRING mutant TRIM21. The results indicated that wild-type TRIM21, but not the ΔRING mutant, reduced NRF2 protein levels in a dose-dependent manner, implying that TRIM21 mediates NRF2 regulation via ubiquitination (Figures S1D,4A and S3D). Additionally, deletion of TRIM21 led to an increase in NRF2 expression (Figure 4B and S3E).

Figure 4 TRIM21 ubiquitinates NRF2. (A) HEK293T cells expressing vector or Myc-TRIM21 were treated with CHX (40 μg/mL) and collected at the indicated time, and analyzed by Western blotting. (B) SV-HUC-1 cells transfected with control shRNA (shCtrl) or TRIM21 shRNA were treated with CHX, collected at specified times, and subjected to Western blotting analysis. (C) HEK293T cells were cotransfected and treated as indicated; cell lysates underwent IP followed by Western blotting analysis. (D) SV-HUC-1 cells were cotransfected with the indicated shRNA and HA-ubiquitin (HA-Ub); cell lysates were subjected to IP and subsequent Western blotting analyses. (E and F) SV-HUC-1 cells were cotransfected as indicated, and NRF2 ubiquitination linkage was analyzed. (G) SV-HUC-1 cells transfected with wild-type Ub or Ub-Lys48R were cultured with either control vector or TRIM21; cell lysates were analyzed by Western blotting using specific antibodies. “+”: Treated; “-”: Untreated.

Building on evidence of TRIM21’s critical role in NRF2 degradation, further analysis assessed whether TRIM21 directly ubiquitinates NRF2. Co-transfection of HA-UB and Flag-NRF2 with either Myc-TRIM21 or Myc-TRIM21-ΔRING in HEK293T cells, followed by immunoprecipitation after MG132 treatment, revealed that wild-type TRIM21, but not the mutant, significantly elevated NRF2 ubiquitination levels (Figure 4C). Furthermore, knockdown of TRIM21 using two independent shRNAs in bladder epithelial cells led to a marked reduction in NRF2 ubiquitination (Figure 4D), identifying TRIM21 as a specific ubiquitinase targeting NRF2 in bladder epithelial cells. Lysine 48 (K48) ubiquitination primarily targets proteins for proteasomal degradation, while Lysine 63 (K63) ubiquitination is not linked to proteasome-mediated degradation pathways.³⁷ Thus, a series of ubiquitin mutants with single Lys residues was generated for further analysis. Co-transfection of Flag-NRF2, Myc-TRIM21, HA-UB, and its mutants demonstrated that TRIM21 overexpression selectively enhanced Lys48 (K48)-linked ubiquitination of NRF2 without affecting Lys63 (K63)-linked ubiquitination (Figure 4E). Moreover, in cells overexpressing TRIM21, forced expression of a Lys48-resistant (K48R) ubiquitin variant led to increased NRF2 levels (Figures 4F, G and S3F). These results indicate that TRIM21 promotes NRF2 degradation through the formation of K48-linked polyubiquitin chains.

TRIM21 Exacerbates Inflammation in Cellular Models

Given the finding that TRIM21 mediates NRF2 degradation through ubiquitination, the role of TRIM21 in bladder inflammation models was further explored. In the bladder cell inflammation model, TRIM21 knockdown reduced excess ROS and increased GSH levels following LPS stimulation, while TRIM21 overexpression led to opposite outcomes (Figure 5A and B). Furthermore, TRIM21 knockdown in this cellular model decreased apoptosis rates induced by LPS stimulation, whereas TRIM21 overexpression increased apoptosis (Figure 5C). Western blotting analysis revealed that TRIM21 knockdown partially restored protein levels of HO-1 and NQO1, thereby attenuating cellular inflammation and apoptosis, whereas TRIM21 overexpression exacerbated these effects (Figures 5D–E,S3G and H). Collectively, these results indicate that TRIM21 knockout mitigates the adverse effects of bladder inflammation on bladder function.

Figure 5 Effect of TRIM21 expression on oxidative stress, and inflammation in the cellular inflammation model. (A) ROS levels in cells transfected with vector, TRIM21, shCtrl, or shTRIM21. (B) Relative GSH/GSSG ratio of SV-HUC-1 cells transfection as indicated. (C) Apoptosis rates of cells transfected as indicated. (D and E) Western blot detected expression levels of oxidative stress inflammatory and apoptotic indicators in SV-HUC-1 after transfection and treated as indicated. * P < 0.05.

IGF2BP2 Maintains TRIM21 mRNA Stability Through m6A Modification

M6A modifications have been established as significant contributors to various inflammatory and bladder-related diseases.^38–40 However, their involvement in IC/BPS remains undocumented. In this study, dot blot experiments demonstrated elevated levels of M6A modifications in the lesional bladder mucosa of patients with IC/BPS (Figure 6A), accompanied by significant differences in TRIM21 mRNA levels between lesional and non-lesional tissues (Figure S2A). Further analysis utilizing the GEPIA database revealed a positive correlation between the expression of IGF2BP2 and YTHDF1, members of the M6A reader family, and TRIM21 at the mRNA level (Figure S2B and C). Detailed examination of IGF2BP2 and YTHDF1 mRNA levels in lesional versus non-lesional bladder mucosa indicated a significant difference for IGF2BP2, while YTHDF1 showed no such difference (Figure S2D and E). This observation led to the hypothesis that M6A modifications recognized by IGF2BP2 play a pivotal role in this disease. Subsequent experiments demonstrated that IGF2BP2 knockdown in SV-HUC-1 cells resulted in reduced mRNA and protein levels of TRIM21 (Figures 6B,C and S3I). Conversely, IGF2BP2 overexpression significantly elevated both mRNA and protein levels of TRIM21 (Figures 6D,E and S3J). To further assess TRIM21 mRNA stability, SV-HUC-1 cells were treated with actinomycin D. The results indicated that IGF2BP2 knockdown decreased TRIM21 mRNA stability, whereas overexpression of IGF2BP2 enhanced it (Figure 6F). Additionally, the decrease in NRF2 protein expression due to IGF2BP2 overexpression could be reversed by TRIM21 knockdown (Figure 6G and S3K). Predictive analysis of m6A binding sites using the RGB suite and SRAMP website suggested that IGF2BP2 likely binds to two specific sites in the first segment of the 3′-UTR of TRIM21. Based on this prediction, two mutants were successfully designed (Figure S2F–H). Comparison with the wild-type group revealed that luciferase activity for Mut1 was significantly reduced, while Mut2 remained unchanged, indicating that site 1 is the most critical binding site (Figure 6H). Finally, RT-qPCR analysis confirmed a positive correlation between TRIM21 and IGF2BP2 expression in the lesional bladder mucosa of patients with IC/BPS (Figure 6I). These results suggest that M6A modifications play a vital role in maintaining TRIM21 expression through IGF2BP2-dependent stabilization of TRIM21 mRNA.

Figure 6 IGF2BP2 maintains TRIM21 mRNA stability through m6A modification. (A) Dot blot assay results from lesional and non-lesional bladder mucosa of patients with IC/BPS. (B and C) RT-qPCR and Western blotting analysis of TRIM21 expression following IGF2BP2 silencing in SV-HUC-1 cells. (D and E) RT-qPCR and Western blotting analysis of TRIM21 expression after IGF2BP2 overexpression in SV-HUC-1 cells. (F) IGF2BP2 enhances TRIM21 stability; SV-HUC-1 cells were treated with Actinomycin D at specified times, and the mRNA half-life of TRIM21 was assessed by RT-qPCR normalized to β-Actin. (G) Western blotting analysis of NRF2 following IGF2BP2 overexpression, with or without shTRIM21. (H) Luciferase activities of WT and mutated TRIM21 plasmids. (I) Positive correlation between IGF2BP2 and TRIM21 expression in the lesional bladder mucosa of patients with IC/BPS. “+”: Treated; “-”: Untreated. * P < 0.05, ** P < 0.01, *** P < 0.001. Abbreviation: ns, not significant.

Correlation Between IGF2BP2, TRIM21 and NRF2

The clinical relevance of our findings was evaluated through IHC detection of IGF2BP2, TRIM21, and NRF2. The results indicated a significant overexpression of both IGF2BP2 and TRIM21, accompanied by a marked reduction in NRF2 levels within the lesional bladder mucosa of patients with IC/BPS. In contrast, these expression patterns were reversed in the non-lesional bladder mucosa (Figure 7A). Collectively, these results demonstrate that IGF2BP2-dependent M6A modifications lead to elevated TRIM21 expression in patients with IC/BPS, with TRIM21 facilitating the progression of IC/BPS by ubiquitinating NRF2 (Figure 7B).

Figure 7 Correlation between IGF2BP2, TRIM21, and NRF2. (A) Representative IHC images of IGF2BP2, TRIM21, and NRF2 of clinical patients with IC/BPS. (B) The mechanistic scheme of our study.

Discussion

IC/BPS, while classified as a benign bladder condition, has garnered increasing attention from both clinicians and patients due to its debilitating pelvic pain and the absence of a specific treatment.²⁵ Despite significant research efforts, few advancements have translated into tangible benefits for patients. Oxidative stress is known to play a critical role in various inflammatory diseases, including IC/BPS.¹¹ However, the precise molecular mechanisms underlying the development of IC/BPS due to oxidative stress imbalance remain unclear. This study identifies the oxidative stress-related protein NRF2 as essential for IC/BPS development, revealing that M6A-modified TRIM21 interacts with NRF2 to regulate its post-translational modification. Notably, this research marks the first identification of the involvement of the E3 ubiquitin ligase TRIM21 and the M6A reader IGF2BP2 in IC/BPS.

Dysregulation of NRF2 signaling has been strongly linked to numerous oxidative stress-related and inflammatory diseases, including liver fibrosis, neurodegeneration, cancer, and IC/BPS.^41–43 Previous studies have primarily focused on alleviating symptoms in animal models of IC/BPS by pharmacologically enhancing NRF2 expression in vivo, yet the underlying mechanisms remain largely unexplored,^17,44 however, in other inflammatory conditions or bladder pathologies, transcriptional regulation and post-translational modifications of NRF2 have been observed to exert pivotal influences on its expression, which are critically involved in disease pathogenesis and progression.^45,46 In this study, specimens collected from patients with clinical IC/BPS revealed significantly decreased NRF2 protein expression in the lesional bladder mucosa, while higher expression levels were noted in non-lesional tissues. This disparity likely contributes to the lesional bladder mucosa’s diminished capacity to withstand damage from harmful substances, leading to scattered areas of bleeding and ulceration. Furthermore, through the establishment of mouse and cellular models of bladder inflammation, it was demonstrated that NRF2 plays a pivotal role in maintaining redox homeostasis and enhancing resistance to bladder inflammation. Additionally, this study elucidated that TRIM21, a key ubiquitinase, interacts with NRF2, subsequently reducing its protein expression and impairing its function in sustaining redox balance. Ubiquitination is a post-translational modification process that governs protein activity and function, playing a critical role in the progression of various inflammatory diseases.^47,48 TRIM21, a member of the TRIM protein family and a RING E3 ubiquitin ligase, regulates the ubiquitination of multiple proteins and has been implicated in several oxidative stress-related and inflammatory conditions. For instance, TRIM21 ubiquitinates NOX2 to amplify oxidative inflammation in the heart, modifies TRPM2 to contribute to airway inflammation, and interacts with SIRT5 to modulate colonic inflammation.^49–51 This study identifies TRIM21 as a novel regulator of NRF2, facilitating NRF2 degradation through ubiquitination and thereby accelerating the development of IC/BPS.

As the most prevalent mRNA modification in humans, M6A modification plays a key role in regulating RNA structure and stability, significantly impacting the progression of various inflammatory diseases.^38,52 However, the role of M6A modification in inflammatory bladder diseases, particularly IC/BPS, has been underexplored. This study establishes a link between M6A modification and IC/BPS, revealing that IGF2BP2 is highly expressed in the bladder mucosa of patients with IC/BPS. IGF2BP2 enhances TRIM21 expression by stabilizing TRIM21 mRNA, resulting in decreased NRF2 levels.

However, our study has several limitations. First, the etiology and pathogenesis of IC/BPS remain incompletely understood, precluding the availability of disease-specific animal or cellular models. Consequently, we adopted surrogate inflammation models – CYP-induced cystitis in vivo and LPS stimulation in vitro – to validate the protective role of NRF2. While CYP-induced cystitis effectively recapitulates key IC/BPS clinical features (urinary frequency, pelvic pain, and bladder hyperemia),⁵³ and both CYP/LPS models induce ROS overproduction mirroring human disease mechanisms,¹¹ they primarily represent acute inflammatory states. Chronic IC/BPS involves complex neuroimmune interactions and tissue remodeling processes not fully captured here.⁵⁴ Our study focused on elucidating the IGF2BP2-TRIM21-NRF2 regulatory axis – a pathway conserved across acute/chronic inflammation – as validated by human tissue analyses. Future work will employ chronic models (eg, repeated low-dose CYP or autoimmune cystitis) to examine long-term adaptations of this axis. Additionally, the limited cohort size (n=16 patients) reflects inherent challenges in studying IC/BPS: low prevalence rates, stringent inclusion/exclusion criteria, and ethically constrained recruitment. We will continue collecting IC/BPS specimens for further validation.

In conclusion, our findings demonstrate a significant reduction in NRF2 expression in the lesional bladder mucosa of patients with IC/BPS, identifying NRF2 as a critical regulator in cellular and mouse models of bladder inflammation. Furthermore, this research marks the first identification of the roles of ubiquitination and M6A modifications in IC/BPS. Specifically, the M6A modification recognized by IGF2BP2 increases the stability and expression of TRIM21 mRNA, leading to excessive TRIM21-mediated degradation of NRF2 protein through ubiquitination. These insights provide new perspectives on the pathogenesis of IC/BPS and suggest potential strategies for therapeutic intervention.

Data Sharing Statement

All data necessary for confirming the conclusions are included in this article. The data employed in this investigation are accessible from the corresponding author upon justified request.

Ethics Approval and Consent to Participate

This study was conducted in accordance with the Declaration of Helsinki. The collection of specimens was approved by the Ethics Committee of the second Affiliated Hospital of Nanjing Medical University (2017KY-102), and the samples involved in this study were obtained with written informed consent from patients or their families. All animal experiments strictly adhered to the Guidelines for the Ethical Review of Animal Welfare (GB/T 35892-2018) and were approved by the Animal Ethics Committee of Nanjing Medical University (IACUC-2210019).

Acknowledgments

We thank the Second Clinical Medical School of Nanjing Medical University for providing the experimental platforms. We thank members of our teams for helpful discussion.

Author Contributions

Z.F., Q.G., and B.N. contributed equally to this work. 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.

Funding

This work was supported by grants from the National Natural Science Foundation of China (NO.82370781 and 82470808), Key Research and Development Plan (Social Development) of Science and Technology, Department of Jiangsu Province (NO. BE2020724), the National Key Technology R&D Program of China (NO. 2023YFC23606003 and NO. 2018YFC2002204) and Ningguo city Science and Technology Development Project (NO.2023-43).

Disclosure

The authors declare no conflicts of interest in this work.

References

1. van de Merwe JP, Nordling J, Bouchelouche P, et al. Diagnostic criteria, classification, and nomenclature for painful bladder syndrome/interstitial cystitis: an ESSIC proposal. Eur Urol. 2008;53(1):60–67. doi:10.1016/j.eururo.2007.09.019

2. Davis NF, Brady CM, Creagh T. Interstitial cystitis/painful bladder syndrome: epidemiology, pathophysiology and evidence-based treatment options. Eur J Obstet Gynecol Reprod Biol. 2014;175:30–37. doi:10.1016/j.ejogrb.2013.12.041

3. Homma Y, Akiyama Y, Tomoe H, et al. Clinical guidelines for interstitial cystitis/bladder pain syndrome. Int J Urol. 2020;27(7):578–589. doi:10.1111/iju.14234

4. Akiyama Y, Homma Y, Maeda D. Pathology and terminology of interstitial cystitis/bladder pain syndrome: a review. Histol Histopathol. 2019;34(1):25–32. doi:10.14670/HH-18-028

5. Lackner JM, Clemens JQ, Radziwon C, et al. Cognitive behavioral therapy for chronic pelvic pain: what is it and does it work? J Urol. 2024;211(4):539–550. doi:10.1097/JU.0000000000003847

6. Duh K, Funaro MG, DeGouveia W, et al. Crosstalk between the immune system and neural pathways in interstitial cystitis/bladder pain syndrome. Discov Med. 2018;25(139):243–250.

7. Patnaik SS, Lagana AS, Vitale SG, et al. Etiology, pathophysiology and biomarkers of interstitial cystitis/painful bladder syndrome. Arch Gynecol Obstet. 2017;295(6):1341–1359. doi:10.1007/s00404-017-4364-2

8. Gu J, Zhang P, Li H, et al. Cerium-luteolin nanocomplexes in managing inflammation-related diseases by antioxidant and immunoregulation. ACS Nano. 2024;18(8):6229–6242. doi:10.1021/acsnano.3c09528

9. Pokharel MD, Garcia-Flores A, Marciano D, et al. Mitochondrial network dynamics in pulmonary disease: bridging the gap between inflammation, oxidative stress, and bioenergetics. Redox Biol. 2024;70:103049. doi:10.1016/j.redox.2024.103049

10. Ueda Y, Nakazawa D, Nishio S, et al. Transcription factor Nrf2 activation regulates NETosis, endothelial injury, and kidney disease in myeloperoxidase-positive antineutrophil cytoplasmic antibody-associated vasculitis. Kidney Int. 2024;105:1291–1305. doi:10.1016/j.kint.2024.03.010

11. Mohammad A, Laboulaye MA, Shenhar C, Dobberfuhl AD. Mechanisms of oxidative stress in interstitial cystitis/bladder pain syndrome. Nat Rev Urol. 2024;21:433–449. doi:10.1038/s41585-023-00850-y

12. Ener K, Keske M, Aldemir M, et al. Evaluation of oxidative stress status and antioxidant capacity in patients with painful bladder syndrome/interstitial cystitis: preliminary results of a randomised study. Int Urol Nephrol. 2015;47(8):1297–1302. doi:10.1007/s11255-015-1021-1

13. Baird L, Yamamoto M. The molecular mechanisms regulating the KEAP1-NRF2 pathway. Mol Cell Biol. 2020;40(13). doi:10.1128/MCB.00099-20

14. Ryter SW, Choi AM. Targeting heme oxygenase-1 and carbon monoxide for therapeutic modulation of inflammation. Transl Res. 2016;167(1):7–34. doi:10.1016/j.trsl.2015.06.011

15. Morgenstern C, Lastres-Becker I, Demirdogen BC, et al. Biomarkers of NRF2 signalling: current status and future challenges. Redox Biol. 2024;72:103134. doi:10.1016/j.redox.2024.103134

16. Ni B, Chen Z, Shu L, et al. Nrf2 pathway ameliorates bladder dysfunction in cyclophosphamide-induced cystitis via suppression of oxidative stress. Oxid Med Cell Longev. 2021;2021:4009308. doi:10.1155/2021/4009308

17. Kuret T, Kreft ME, Romih R, Veranic P. Cannabidiol as a promising therapeutic option in IC/BPS: in vitro evaluation of its protective effects against inflammation and oxidative stress. Int J Mol Sci. 2023;24(5):5055. doi:10.3390/ijms24055055

18. Hoeller D, Hecker CM, Dikic I. Ubiquitin and ubiquitin-like proteins in cancer pathogenesis. Nat Rev Cancer. 2006;6(10):776–788. doi:10.1038/nrc1994

19. Gitlin AD, Maltzman A, Kanno Y, et al. N4BP1 coordinates ubiquitin-dependent crosstalk within the IkappaB kinase family to limit toll-like receptor signaling and inflammation. Immunity. 2024;57(5):973–986e7. doi:10.1016/j.immuni.2024.04.004

20. Huang CK, Chen Z, Zhou Z, et al. RNF149 destabilizes IFNGR1 in macrophages to favor post-infarction cardiac repair. Circ Res. 2024;135:518–536. doi:10.1161/CIRCRESAHA.123.324023

21. Liu J, Zhang C, Xu D, et al. The ubiquitin ligase TRIM21 regulates mutant p53 accumulation and gain of function in cancer. J Clin Invest. 2023;133(6). doi:10.1172/JCI164354

22. Gu M, Tan M, Zhou L, et al. Protein phosphatase 2Acalpha modulates fatty acid oxidation and glycolysis to determine tubular cell fate and kidney injury. Kidney Int. 2022;102(2):321–336. doi:10.1016/j.kint.2022.03.024

23. Yang C, Wang Z, Kang Y, et al. Stress granule homeostasis is modulated by TRIM21-mediated ubiquitination of G3BP1 and autophagy-dependent elimination of stress granules. Autophagy. 2023;19(7):1934–1951. doi:10.1080/15548627.2022.2164427

24. Hanno PM, Erickson D, Moldwin R, Faraday MM, American Urological A. Diagnosis and treatment of interstitial cystitis/bladder pain syndrome: AUA guideline amendment. J Urol. 2015;193(5):1545–1553. doi:10.1016/j.juro.2015.01.086

25. Shen SH, Peng L, Zeng X, Zhang J, Shen H, Luo DY. Intravesical interferon therapy vs hyaluronic acid for pain among female individuals with interstitial cystitis: a randomized clinical trial. JAMA Netw Open. 2024;7(4):e244880. doi:10.1001/jamanetworkopen.2024.4880

26. Peng L, Jin X, Li BY, et al. Integrating single-cell RNA sequencing with spatial transcriptomics reveals immune landscape for interstitial cystitis. Signal Transduct Target Ther. 2022;7(1):161. doi:10.1038/s41392-022-00962-8

27. Tay C, Grundy L. Animal models of interstitial cystitis/bladder pain syndrome. Front Physiol. 2023;14:1232017. doi:10.3389/fphys.2023.1232017

28. Wegner KA, Abler LL, Oakes SR, et al. Void spot assay procedural optimization and software for rapid and objective quantification of rodent voiding function, including overlapping urine spots. Am J Physiol Renal Physiol. 2018;315(4):F1067–F1080. doi:10.1152/ajprenal.00245.2018

29. Chaplan SR, Bach FW, Pogrel JW, Chung JM, Yaksh TL. Quantitative assessment of tactile allodynia in the rat paw. J Neurosci Methods. 1994;53(1):55–63. doi:10.1016/0165-0270(94)90144-9

30. Bjorling DE, Wang Z, Vezina CM, et al. Evaluation of voiding assays in mice: impact of genetic strains and sex. Am J Physiol Renal Physiol. 2015;308(12):F1369–78. doi:10.1152/ajprenal.00072.2015

31. Anandhan A, Dodson M, Shakya A, et al. NRF2 controls iron homeostasis and ferroptosis through HERC2 and VAMP8. Sci Adv. 2023;9(5):eade9585. doi:10.1126/sciadv.ade9585

32. Huang W, Wu D, Cai C, et al. Inhibition of MST1 ameliorates neuronal apoptosis via GSK3beta/beta-TrCP/NRF2 pathway in spinal cord injury accompanied by diabetes. Redox Biol. 2024;71:103104. doi:10.1016/j.redox.2024.103104

33. Li L, Xie D, Yu S, et al. WNK1 Interaction with KEAP1 promotes NRF2 stabilization to enhance the oxidative stress response in hepatocellular carcinoma. Cancer Res. 2024;84:2776–2791. doi:10.1158/0008-5472.CAN-23-1167

34. Tanaka T, Okuda H, Isonishi A, et al. Dermal macrophages set pain sensitivity by modulating the amount of tissue NGF through an SNX25-Nrf2 pathway. Nat Immunol. 2023;24(3):439–451. doi:10.1038/s41590-022-01418-5

35. Lu W, Cui J, Wang W, et al. PPIA dictates NRF2 stability to promote lung cancer progression. Nat Commun. 2024;15(1):4703. doi:10.1038/s41467-024-48364-4

36. Pierce BG, Wiehe K, Hwang H, Kim BH, Vreven T, Weng Z. ZDOCK server: interactive docking prediction of protein-protein complexes and symmetric multimers. Bioinformatics. 2014;30(12):1771–1773. doi:10.1093/bioinformatics/btu097

37. Kulathu Y, Komander D. Atypical ubiquitylation – the unexplored world of polyubiquitin beyond Lys48 and Lys63 linkages. Nat Rev Mol Cell Biol. 2012;13(8):508–523. doi:10.1038/nrm3394

38. Wang JN, Wang F, Ke J, et al. Inhibition of METTL3 attenuates renal injury and inflammation by alleviating TAB3 m6A modifications via IGF2BP2-dependent mechanisms. Sci Transl Med. 2022;14(640):eabk2709. doi:10.1126/scitranslmed.abk2709

39. Zhang L, Li Y, Zhou L, et al. The m6A reader YTHDF2 promotes bladder cancer progression by suppressing RIG-I-mediated immune response. Cancer Res. 2023;83(11):1834–1850. doi:10.1158/0008-5472.CAN-22-2485

40. Han X, Liu L, Huang S, et al. RNA m(6)A methylation modulates airway inflammation in allergic asthma via PTX3-dependent macrophage homeostasis. Nat Commun. 2023;14(1):7328. doi:10.1038/s41467-023-43219-w

41. Blas-Garcia A, Apostolova N. Novel therapeutic approaches to liver fibrosis based on targeting oxidative stress. Antioxidants. 2023;12(8):1567. doi:10.3390/antiox12081567

42. Brackhan M, Arribas-Blazquez M, Lastres-Becker I. Aging, NRF2, and TAU: a perfect match for neurodegeneration? Antioxidants. 2023;12(8). doi:10.3390/antiox12081564

43. Hecht F, Zocchi M, Alimohammadi F, Harris IS. Regulation of antioxidants in cancer. Mol Cell. 2024;84(1):23–33. doi:10.1016/j.molcel.2023.11.001

44. Gao Q, Zhao Y, Luo R, et al. Intrathecal umbilical cord mesenchymal stem cells injection alleviates neuroinflammation and oxidative stress in the cyclophosphamide-induced interstitial cystitis rats through the Sirt1/Nrf2/HO-1 pathway. Life Sci. 2023;331:122045. doi:10.1016/j.lfs.2023.122045

45. Tang L, Yang C, Peng Y, et al. YAP/Nrf2 suppresses ferroptosis to alleviate acute lung injury induced by intestinal ischemia/reperfusion. Redox Biol. 2025;86:103811. doi:10.1016/j.redox.2025.103811

46. Fan L, Guo D, Zhu C, et al. LRRC45 accelerates bladder cancer development and ferroptosis inhibition via stabilizing NRF2 by competitively KEAP1 interaction. Free Radic Biol Med. 2025;226:29–42. doi:10.1016/j.freeradbiomed.2024.11.001

47. Mevissen TET, Komander D. Mechanisms of deubiquitinase specificity and regulation. Annu Rev Biochem. 2017;86:159–192. doi:10.1146/annurev-biochem-061516-044916

48. Zhao Y, Fan S, Zhu H, et al. Podocyte OTUD5 alleviates diabetic kidney disease through deubiquitinating TAK1 and reducing podocyte inflammation and injury. Nat Commun. 2024;15(1):5441. doi:10.1038/s41467-024-49854-1

49. Yao P, Chen T, Jiang P, Li L, Du W. Functional skewing of TRIM21-SIRT5 interplay dictates IL-1beta production in DSS-induced colitis. EMBO Rep. 2022;23(9):e54391. doi:10.15252/embr.202154391

50. Liu X, Zhang W, Luo J, et al. TRIM21 deficiency protects against atrial inflammation and remodeling post myocardial infarction by attenuating oxidative stress. Redox Biol. 2023;62:102679. doi:10.1016/j.redox.2023.102679

51. Li X, Wang W, Shao Y, et al. LncTRPM2-AS inhibits TRIM21-mediated TRPM2 ubiquitination and prevents autophagy-induced apoptosis of macrophages in asthma. Cell Death Dis. 2021;12(12):1153. doi:10.1038/s41419-021-04437-6

52. Olazagoitia-Garmendia A, Zhang L, Mera P, et al. Gluten-induced RNA methylation changes regulate intestinal inflammation via allele-specific XPO1 translation in epithelial cells. Gut. 2022;71(1):68–76. doi:10.1136/gutjnl-2020-322566

53. de Oliveira MG, Monica FZ, Passos GR, et al. Selective pharmacological inhibition of NOX2 by GSK2795039 improves bladder dysfunction in cyclophosphamide-induced cystitis in mice. Antioxidants. 2022;12(1):92. doi:10.3390/antiox12010092

54. Jhang JF, Yu WR, Jiang YH, Kuo HC. Pathophysiology and potential multimodal therapeutic strategies for IC/BPS. Nat Rev Urol. 2025. doi:10.1038/s41585-025-01044-4

Home security is no longer a luxury. In today’s fast-changing world, a reliable security camera can give peace of mind, whether you’re away on vacation or simply resting at home. Modern cameras don’t just record footage. They provide smart alerts, crystal-clear video, night vision, two-way audio, and easy integration with other smart devices. For buyers in Pakistan, the latest options bring the best global features ever.

From indoor units that let you check on pets to outdoor cameras guarding your gate, the choices are wide. Some even come with floodlights, solar charging, and smart motion detection. Whether you’re looking for one camera or building a complete system, 2025 offers strong contenders for every budget and home in Pakistan.

Google Nest Cam (Wired)

Indoor/Outdoor: Indoor
Power source: Wired
Resolution: 1080p
Night vision: Infrared
Storage: Subscription only

Google Nest Cam remains one of the most user-friendly options in 2025. Setting it up is simple using the Google Home app, which offers clear menus and step-by-step guidance. The design is sleek, with adjustable angles for flexible placement. Video quality is excellent during the day and clear at night with infrared.

However, without a Nest Aware subscription, you only get three hours of event history. Paid plans unlock longer storage and advanced features like familiar face alerts. For homes in Pakistan already using Google devices, this camera integrates smoothly with Nest Hub and Chromecast.

Ecobee SmartCamera

Indoor/Outdoor: Indoor
Power source: Wired
Resolution: 1080p
Compatibility: Alexa, Apple HomeKit
Subscription: Optional

The Ecobee SmartCamera is best for Apple HomeKit users in Pakistan. Setup is surprisingly quick, and the device stays connected reliably. Its wide 170-degree field of view captures large spaces, and the warmer picture tone adds comfort when checking in remotely.

It also doubles as an Alexa speaker, making it versatile. Without a subscription, Apple’s iCloud+ can store 10 days of footage. With Ecobee’s own plan, you can add professional monitoring and smoke alarm detection.

Blink Outdoor 4

Indoor/Outdoor: Both
Power source: Battery (AA)
Resolution: 1080p
Storage: Cloud or local USB
Best for: Alexa users

For Alexa households, Blink Outdoor 4 is the most convenient choice. It supports voice commands through Echo devices and streams directly to Alexa displays. The system is easy to expand; new cameras can be added anytime.

The cameras run on AA batteries, lasting up to two years. Video quality is sharp, with effective motion detection. Local storage with a USB drive avoids subscription costs, while paid plans offer advanced detection and longer cloud storage.

Blink Mini 2

Indoor/Outdoor: Both
Power source: Wired
Resolution: 1080p
Special feature: Spotlight for color night vision

Blink Mini 2 is a budget-friendly option often sold for under PKR 19,000 in Pakistan. It’s compact, easy to install, and offers solid features for the price. Motion alerts are quick, and the optional spotlight delivers color video at night.

While it lacks automatic arming and advanced AI detection without a subscription, it remains a strong starter camera. With a Sync Module 2, you can also save recordings locally.

Reolink Atlas PT Ultra

Indoor/Outdoor: Outdoor
Power source: Rechargeable battery (20,000 mAh)
Resolution: 4K Ultra HD
Features: Auto-tracking, solar panel support

For larger homes or offices in Pakistan, the Reolink Atlas PT Ultra is an excellent outdoor camera. It delivers 4K video quality and supports auto-tracking with full pan and tilt control. The huge battery lasts months, and solar charging adds convenience.

The companion app is responsive, allowing fine control of motion zones, recording settings, and pre-recording options. Local microSD storage is available, but a cloud plan can be added for extended backup.

What to Consider Before Buying a Security Camera in Pakistan

1. Indoor vs. Outdoor: Choose weatherproof models for gates, garages, or outdoor spaces.
2. Power source: Decide between wired, rechargeable battery, or solar-powered cameras.
3. Video quality: Most households will find 1080p enough, while businesses may benefit from 4K.
4. Storage options: Local microSD is safer for privacy, while cloud storage allows remote access.
5. Smart features: Look for motion detection, night vision, and voice assistant compatibility.
6. Budget: Prices in Pakistan vary widely. Blink offers affordability, while Reolink and Nest target premium buyers.

Security cameras in Pakistan are evolving quickly. Whether you want a simple indoor unit or a feature-rich outdoor system, 2025 brings excellent options at different price points. Google Nest Cam and Ecobee offer seamless smart home integration, Blink covers affordability, and Reolink delivers professional-grade outdoor security.

Before buying, consider your home’s layout, power needs, and whether you prefer local or cloud storage. With the right choice, you can secure your home with smart technology and peace of mind.

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.