Category: 3. Business

China will exempt some Nexperia chips from an export ban that was imposed amid an escalating row with the Dutch government, officials said on Saturday.

“We will comprehensively consider the actual situation of enterprises and grant exemptions to exports that meet the criteria,” the Chinese Commerce Ministry said in a statement.

Nexperia produces components in Europe, sends them to China for finishing and then re-exports them back to customers in Europe.

The Netherlands-based company is owned by China’s Wingtech Technology. But the Dutch government invoked a Cold War-era law to effectively take control of the semiconductor maker in September, citing security concerns.

This prompted China to announce export controls on the chips in October.

China, EU and US talk export controls

The Wall Street Journal, citing unnamed sources, said the exemption for Nexperia chips came after a meeting between US President Donald Trump and Chinese President Xi Jinping in South Korea.

The Dutch government refused to comment on the reports and said it remained in contact with Chinese authorities “to work toward a constructive solution that restores balance to the chip supply chain and that is good for Nexperia and our economies.”

Meanwhile, Chinese and European Union officials also held talks on export controls more broadly.

“China confirmed that the suspension of the October export controls applies to the EU. Both sides reaffirmed commitment to continue engagement on improving the implementation of export control policies,” EU Trade Commissioner Maros Sefcovic said in a post on X.

Why are Nexperia semiconductors important?

Nexperia components are mainly found in cars, with the company supplying 49% of the electronic components used in the European automotive industry, according to German business newspaper Handelsblatt.

Although the components are technically replaceable, establishing alternative supply chains poses a major challenge for European automakers and other Nexperia customers.

“Without these chips, European automotive suppliers cannot build the parts and components needed to supply vehicle manufacturers and this therefore threatens production stoppages,” European auto lobby ACEA warned last month.

In its statement on Saturday, China’s Commerce Ministry placed blame on “the Dutch government’s improper intervention in the internal affairs of enterprises” for causing “the current chaos in the global supply chain.”

Edited by: Srinivas Mazumdaru 

Company: Advanced Drainage Systems (WMS)

Business: Advanced Drainage Systems is a manufacturer of stormwater and onsite wastewater solutions. The company and its subsidiary, Infiltrator Water Technologies, provide stormwater drainage and onsite wastewater products used in a wide variety of markets and applications, including commercial, residential, infrastructure and agriculture, while delivering customer service. Its pipe segment manufactures and markets thermoplastic corrugated pipe throughout the United States. Its infiltrator segment is a provider of plastic leachfield chambers and systems, septic tanks and accessories, primarily for use in residential applications. Its international segment manufactures and markets products in regions outside the United States, with a strategy focused on its owned facilities in Canada and those markets serviced through its joint ventures in Mexico and South America. Its Allied Products segment manufactures a range of products which are complementary to their pipe products.

Stock Market Value: : $11.98 billion ($144.10 per share)

Activist: Impactive Capital

Ownership: 2.14%

Average Cost: n/a

Activist Commentary: Impactive Capital is an activist hedge fund founded in 2018 by Lauren Taylor Wolfe and Christian Alejandro Asmar. Impactive Capital is an active ESG investor that launched with a $250 million investment from CalSTRS and now has approximately $3 billion. In just seven years, they have made quite a name for themselves as AESG investors. Wolfe and Asmar realized that there was an opportunity to use tools, notably on the social and environmental side, to drive returns. Impactive focuses on positive systemic change to help build more competitive, sustainable businesses for the long run. Impactive will use traditional operational, financial and strategic tools that activists use, but will also implement ESG change that they believe is material to the business and drives profitability of the company and shareholder value. Impactive looks for high quality businesses that are usually complex and mispriced, where they can underwrite a minimum of a high teens or low 20% internal rate of return over a three- to five-year holding period, and have active engagement with management to set up multiple ways to win.

What’s happening

On Oct. 21, Impactive said they had taken a position in Advanced Drainage Systems.

Behind the scenes

Advanced Drainage Systems is the market share leader in plastic stormwater and onsite septic wastewater management solutions. The company is a pioneer in the development and manufacturing of plastic drainage products, primarily utilizing high-density polyethylene (HDPE) and polypropylene. Recycled materials made up 46% of WMS’ purchased inputs in fiscal year 2025, making it one of the largest recyclers in North America. The company has three primary business lines: (i) Pipe – storm and drainage pipe, 56% of FY25 revenue; (ii) Allied Products – complementary products to its pipe offerings like storm chambers, structures and fittings, 26%; and (iii) Infiltrator – chambers, tanks and advanced wastewater treatment solutions, 18%. Between its three segments, the company has a $15 billion addressable market and is the clear industry leader with 75% to 95% market share across its segments.

There is a lot to like about WMS, as it is an extremely high-quality and well-run company with a long history of compounding growth and secular tailwinds. As a result, WMS has an impressive track record, having grown earnings per share almost 10x since its initial public offering, and has a 28% EPS compound annual growth rate with returns on invested capital consistently above 20%. Management is also very focused on shareholder value and are great capital allocators, increasing dividends and launching buybacks in most years where it does not see a compelling M&A opportunity.

Despite this, the company’s share price performance has been lackluster over the past 1- and 3-year periods, underperforming the Russell 2000, and its stock has re-rated down to a P/E multiple in the low-to-mid 20s. The reason for this is twofold: investor fears regarding the cyclicality of construction spending and margin compression. However, Impactive Capital believes that both concerns appear to be overblown or misplaced and that management has built this business to protect its top line from market cyclicality and make margin expansion structural, not cyclical.

As to the cyclicality of construction spending, construction spending is down 3% year to date as higher interest rates and affordability concerns have dampened residential and non-residential construction spending, setting this up to be the worst year for construction in the past two decades aside from the global financial crisis. But company revenue has not been declining and is not expected to decline for several reasons.

First, plastic pipes have been stealing market share from concrete and steel. Only about 20% of the market in 2010, plastic now exceeds 40% due to it being 20% cheaper than alternatives and offering superior performance.

Second, with the 2019 acquisition of Infiltrator and the upcoming acquisition of National Diversified Sales, WMS has increased its exposure to the residential repair and remodel end-market, adding resiliency to its revenue streams. This should also make WMS a natural beneficiary of the reversion in existing home sales, which are currently at a 15-year low.

Third, billion-dollar storm events have quintupled since the 1980s, necessitating increased investment in resiliency and more complex stormwater infrastructure. The company also has a wide moat, enabled by its high brand loyalty from contractors, its vertical integration and excellent distribution network.

As for margin concerns, there are fears that weakness in construction will lead to margin compression. However, this is something else that management has taken a lot of steps and adopted many initiatives to avoid. Over the past six years, the company has been diversifying its business toward its higher-margin Allied Product and Infiltrator offerings, both of which have adjusted operating margins in the mid-50s, whereas pipe is around 30%.

Additionally, one of its largest input costs are oil and resin, and WMS has a unique way to mitigate these costs. The company toggles between recycled and virgin resins depending on the price of oil. So, when oil spikes, they use recycled resin, and when it drops, they switch to virgin resin and capture better margins. WMS is the only one of its competitors who can do this at scale. Moreover, when construction is weak, oil and resin prices tend to decline. So, loss to the top line can be made up on the bottom line as the decline in resin prices is more than enough to offset end-market weaknesses (i.e., construction spending is down about 3% YTD, resin prices are down 15% to 20%). As a result, pipe and Allied Products adjusted EBITDA margins have expanded by about 8 percentage points since 2020, but some fear that this will eventually normalize.

However, Impactive believes that this shift is structural, not cyclical and WMS will not only avoid margin compression but could see gross margin expand by 100 bps over the next 12-24 months; something that is not factored into forward consensus estimates.

As a result of this confluence of factors, Impactive models that WMS will return to mid-teens EPS growth and projects a base case three-year total return and IRR of 69% and 19%, respectively, and an upside case of 146% and 34%, respectively.

Ken Squire is the founder and president of 13D Monitor, an institutional research service on shareholder activism, and the founder and portfolio manager of the 13D Activist Fund, a mutual fund that invests in a portfolio of activist investments.

The expected reduction in tariffs between the US and China, combined with declining global demand, has led to a renewed downward trend in gold prices.

In the international bullion market, the price of gold fell by $16 per ounce to reach $4,002, which also impacted local markets.

On Saturday, the price of 24-carat gold per tola dropped by Rs1,600 to Rs422,562, while the price per 10 grams fell by Rs1,372 to Rs362,278.

Read: Gold shines again as global prices surge $53 per ounce

Similarly, the price of silver per tola decreased by Rs65 to Rs5,127, and the price per 10 grams fell by Rs56 to Rs4,395.

On October 31, spot gold fell 0.6% to $4,001.74 per ounce at 1:49 pm ET (1749 GMT) and was on track for a 3.7% gain this month.

Elsewhere, spot silver fell 0.4% to $48.73 per ounce, platinum lost 1.7% to $1,583.41, and palladium fell 0.4% to $1,440.02.

People view a Chinese BYD vehicle at the Take Charge Expo in Christchurch, New Zealand, Nov. 1, 2025. The on-going expo showcases the latest in battery electric vehicles with Chinese brands drawing attention from international buyers. (Xinhua/Long Lei)

People experience a Chinese BYD vehicle at the Take Charge Expo in Christchurch, New Zealand, Nov. 1, 2025. The on-going expo showcases the latest in battery electric vehicles with Chinese brands drawing attention from international buyers. (Xinhua/Long Lei)

A staff member introduces a Chinese ZEEKR vehicle at the Take Charge Expo in Christchurch, New Zealand, Nov. 1, 2025. The on-going expo showcases the latest in battery electric vehicles with Chinese brands drawing attention from international buyers. (Xinhua/Long Lei)

A man learns about a Chinese BYD vehicle at the Take Charge Expo in Christchurch, New Zealand, Nov. 1, 2025. The on-going expo showcases the latest in battery electric vehicles with Chinese brands drawing attention from international buyers. (Xinhua/Long Lei)

People view a Chinese BYD vehicle at the Take Charge Expo in Christchurch, New Zealand, Nov. 1, 2025. The on-going expo showcases the latest in battery electric vehicles with Chinese brands drawing attention from international buyers. (Xinhua/Long Lei)

Characterization of glial and vascular cell populations in perioperative hippocampus

The hippocampi of mice in both control and surgery groups were dissected 24 h postoperatively for single-cell RNA sequencing (scRNA-seq, Fig. 1A). Sequencing yielded a total of 20,684 hippocampal cells. Following classification, 10 cell populations were identified including microglia, astrocytes, ECs, neurons, and others (Fig. 1B). Given their essential biological functions within BBB damage, CNS inflammation and homeostasis, as well as their observed transcriptional alterations, we focused on two glial cell populations and one vascular cell population: microglia (marked by C1qc, 3,814 and 4,776 cells in control and surgery groups, Fig. 1C), astrocytes (marked by Aqp4, 1,183 and 1,072 cells, Fig. 1D), and ECs (marked by Flt, 150 and 236 cells, Fig. 1E). Heatmap displayed the top 10 differentially expressed genes (DEGs) for each cell type (Fig. 1F). For microglia, top DEGs such as C1qc and Ccl4 are involved in the classical complement cascade and chemokine-mediated immune recruitment, contributing to microglial inflammatory activation [24], while Cx3cr1 mediates neuron-microglia signaling and synaptic maintenance. For astrocytes, DEG Slc1a2 regulates glutamate clearance to prevent excitotoxicity, Aldoc participates in glycolysis, and Plpp3 modulates phospholipid metabolism and astrocyte development [25]. For ECs, DEG Cldn5 is a key tight junction (TJ) protein essential for BBB integrity, Flt1 mediates vascular remodeling, and Igfbp7 influences endothelial permeability and angiogenic balance (Fig. 1G) [26].

Perioperative inflammation associated microglia expanded over 14-fold

The initiation of CNS inflammation is closely associated with the activation of hippocampal microglia, which serve as the resident immune cells of the brain [27]. To further investigate the heterogeneity of microglial activation in this context, a secondary clustering analysis was performed on 8,590 hippocampal microglia, identifying eight transcriptionally distinct subpopulations with divergent activation states and transcriptional profiles. Among them, 42% were homeostatic microglia (HM). 27% of microglia expressed both immediate early genes (IEGs, e.g. Fos, Jun, Egr1) and inflammatory genes (e.g. Irf1, Icam1), representing a transcriptional state intermediate between homeostasis and full activation. They were annotated as transition state microglia (TSM). One subpopulation (accounting for 12%) expressed a significant number of inflammation related genes and was defined as inflammation associated microglia (IAM), and another subpopulation (6%) expressed a set of neurodegenerative disease related genes and was defined as disease associated microglia (DAM). Three subpopulations with anti-inflammatory and neuroprotective gene expression patterns were identified as inflammation inhibitory microglia (IIM, 6%), metabolism associated inhibitory microglia (MIM, 2%), and disease associated inhibitory microglia (DIM, 1%). The remaining 4% of cells were macrophages (MΦ, Fig. 2A, B). Heatmap illustrated the top 10 DEGs in each microglial subpopulation, highlighting the distinct transcriptional profiles (Fig. 2C). Representative functional genes were selected from each subpopulation to show their specificity. In HM, DEG P2ry12 maintained microglial homeostasis and supported synaptic pruning, while Glul encoded glutamine synthetase, essential for ammonia detoxification and neurotransmitter recycling [28]. In TSM, DEG Egr1 and Jun were IEGs responsive to neuronal stimuli and immune signals, facilitating the transcription of inflammation-related pathways. In IAM, Tnf was specifically upregulated (Fig. 2D, scaled average expression: IAM = 1.99, IIM=−1.15, other subpopulations < 0.28), indicating its key role in proinflammatory signaling; Lgals3bp, another highly expressed gene, promoted microglial activation and was associated with neuroinflammatory progression [29]. In DAM, DEG Malat1 contributed to RNA metabolism and was implicated in neurodegeneration, while Plp1, encoding myelin proteolipid protein, reflected alterations in myelin integrity. The top DEGs of IIM, MIM, and DIM included Rplp1, S100a6, Cd74, etc. T-SNE plots (Fig. 2E) showed the top DEGs of each subpopulation in different spatial distributions, with HM, TSM, IAM, DAM, and IIM being the primary subpopulations.

The variation of microglial subpopulations between control group (preoperative period) and surgery group (postoperative period) was further analyzed. In control group, microglia were in a homeostatic state, and HM accounted for 81% of total microglia. TSM were scarcely detected (1.4%). In surgery group, HM decreased from preoperative 81% to 9.8% of total microglia (decreased by 88%). In contrast, TSM increased from preoperative 1.4% to 47.8% (increased by 3343%, or over 33-fold), which suggested that a large proportion of microglia in surgery group was in a transitional state, preparing for further differentiation. Notably, IAM increased from preoperative 1.3% to 21% (increased by 1474%, or over 14-fold), while changes in other subpopulations were relatively insignificant (Fig. 3A). These findings indicated a profound disruption of hippocampal microglial homeostasis, characterized by a shift toward transition state and inflammation-associated activation state after surgery.

Given the significant alterations of microglial subpopulations, their functions were further explored (Fig. 3D). For HM, enrichment of terms included regulation of neurogenesis and positive regulation of cellular catabolic process. For TSM, enriched terms included myeloid leukocyte differentiation and mononuclear cell differentiation. IAM exhibited significant enrichment in regulation of inflammatory response, and cytokine-mediated signaling pathways. In contrast, IIM showed downregulation of inflammation-related pathways. DAM exhibited enrichment in autophagy and regulation of protein stability, while DIM exhibited reduced enrichment in disease-related pathways. Lastly, MIM exhibited enrichment in metabolism-associated pathways and downregulation of inflammation-related terms. The DEGs associated with inflammation and neurodegeneration were illustrated in Fig. 3E. The dashed lines were utilized to approximately delineate the projected distributions of IAM and DAM. In IAM, the high expression of Tnf and Tnfrsf1a (Tnf receptor) suggested its role in inflammatory cascades. C5ar1, a complement receptor, mediated chemotactic responses, and Ier3 modulated cellular resistance to inflammatory and oxidative stress. In DAM, elevated Apod expression reflected lipid dysregulation and oxidative imbalance, while Irf1 was a transcriptional regulator of autophagy and immune response.

Shift in microglial gene expression patterns from HM to IAM and IIM

As shown in dot plot (Fig. 4A), the expression of homeostatic genes (Fcrls, P2ry12, etc.) was higher in HM and TSM (p < 0.0001). In contrast, inflammation-related genes (Tnf, Ccl2, etc.) were highly expressed in IAM (p < 0.0001), expressed at lower levels in HM and TSM, and nearly absent in IIM. Pseudotime analysis (Fig. 4B) revealed that microglia underwent a dynamic transition from HM, through an intermediate transcriptional state characterized by immediate early gene expression (TSM), and subsequently bifurcated into two distinct branches, marked by either pro-inflammatory (IAM) or anti-inflammatory (IIM) gene expression patterns.

Gene co-expression network analysis was employed to identify key regulatory genes responsible for the functions of HM, TSM, IAM and IIM. Modules with the highest number of co-expressed genes and hub genes were identified as the representatives for these subpopulations (Fig. 4C). HM hub genes P2ry12, Itm2b and Sparc were involved in maintaining microglial persistence and homeostatic function [28, 30]. Co-expressed genes, such as Cx3cr1 and Picalm, were also involved in the regulation of neurogenesis [31, 32]. The hub genes Jun and Ier5 identified in the TSM module were associated with rapid responses to stimuli, inflammatory responses, and CNS functions, particularly those involved in neuronal plasticity [33]. Co-expressed genes, such as Cebpb and Hhex, played roles in the development, differentiation, and proliferation of microglia [34]. IAM hub gene Ddx5 influenced the inflammation balance by inhibiting IL-10 secretion [35]. Junb and Egr1 promoted the production of various pro-inflammatory cytokines, thereby facilitating the propagation of CNS inflammation [36]. Numerous co-expressed inflammatory genes (e.g., Nfkbia, Nlrp3, Tnfrsf1a) defined the proinflammatory signature of IAM, shaping the postoperative hippocampal inflammatory milieu. IIM hub gene Tyrobp interacted with immune receptors and was linked to AD development [37]. The downregulation of Tyrobp in IIM, along with complement system genes C1qa and C1qb, suggested neuroprotective effects. Analysis of the four co-expression modules revealed enriched terms that characterize the perioperative functional reprogramming of microglia, suggesting a gradual shift from a homeostatic state to an inflammation-related state.

TNF signaling activation in postoperative hippocampal IAM

Based on the perioperative changes in microglial subpopulations, microglial activation-related signaling pathways were further established (Fig. 5A). TSM, originating from HM, could transform into M1 microglia via IEG pathway (Egr1, Jun and Fos) or into M2 microglia via TGF-β pathway and related mechanisms. In IAM, both Tnf and its receptor Tnfrsf1a were upregulated. As a central regulator of immune responses, Tnf contributed to the pathogenesis of various diseases, such as sepsis and multiple sclerosis, by triggering pro-inflammatory signaling cascades and promoting programmed cell death [38]. The inflammatory cytokine Il1b was upregulated in IAM, possibly via TNF and NF-κB pathway, leading to a pro-inflammatory reaction. Ccl2 and Tspo were upregulated in IAM, while Hk2 and Prkcd were elevated in DAM. These genes were associated with immune signaling, mitochondrial activity, and metabolic processes [39, 40]. In IIM and DIM subpopulations, genes such as Nlrp3, Il1b, and Casp8 were downregulated, indicating suppressed inflammasome and apoptosis-related pathways. However, the proportional changes of these regulatory subpopulations were relatively limited. A significant increase of Tnf + IBA1 + cells (IAM) in the hippocampus of surgery group was confirmed by IF/FISH double staining (p = 0.0074, Fig. 5B). As inflammatory marker genes of IAM, the fluorescence intensities of Tnf, Tnfrsf1a, and Il1b mRNA were significantly elevated in hippocampal microglia of surgery group. As chemokine and mitochondrial marker genes of IAM, the fluorescence intensities of Ccl2 and Tspo mRNA were also elevated in hippocampal microglia of surgery group (Fig. 5C). These findings suggested the role of microglial activation and the associated regulatory network in postoperative hippocampal inflammation.

Perioperative astrocytic subpopulations remained relatively stable

Microglia have been shown to influence astrocytic function through the release of specific cytokines or direct cell-cell interactions [41]. Given the role of astrocytes in maintaining CNS homeostasis and coordinating inflammatory responses, we next explored astrocytic heterogeneity in the perioperative hippocampus. Clustering analysis identified six distinct astrocytic subpopulations (Fig. 6A, B). 36% were homeostatic astrocytes (HA). A small subpopulation (accounting for 3%) expressed genes related to brain blood flow regulation and was defined as angiogenesis associated astrocytes (AAA). Another small subpopulation (accounting for 4%) expressed a significant number of metabolism related genes and was defined as metabolism associated astrocytes (MAA). 32% expressed both neurodegeneration and metabolism related genes and were defined as disease and metabolism associated astrocytes (DMA). The remaining two clusters exhibited neuroprotective functions and were defined as disease associated inhibitory astrocytes (DIA, 9%) and metabolism associated inhibitory astrocytes (MIA, 16%). Heatmap illustrated the top 10 DEGs (Fig. 6C), and dot plot showed the representative functional genes of each subpopulation (Fig. 6D). In HA, DEG Htra1 encoded a serine protease that modulated extracellular matrix remodeling and TGF-β signaling, potentially affecting astrocyte reactivity, while Hacd2 was involved in very long-chain fatty acid elongation, supporting membrane lipid balance [42]. In AAA, DEG Agt played a central role in the renin-angiotensin system, influencing cerebral perfusion and vascular tone [43]. In MAA and DMA, DEG Dlk1 functioned in the regulation of adipocyte differentiation and neurodevelopment, and Aldoc, a glycolytic enzyme, contributed to astrocytic energy metabolism and was involved in maintaining redox homeostasis under stress conditions [44]. In DIA and MIA, the top DEGs included Kif5a, Son, and others. T-SNE plots (Fig. 6E) showed the top DEGs of each subpopulation in distinct spatial distributions.

Compared to microglia, perioperative changes in astrocytes were less pronounced. In surgery group, HA decreased from preoperative 39.6% to 32.6% of total astrocytes (decreased by 17.5%), and MIA increased from 14.1% to 17.9% of total astrocytes (increased by 26.9%). And the changes of other subpopulations were more insignificant (Fig. 7A). The functions of astrocytic subpopulations were further explored (Fig. 7D). For HA, enriched GO terms included glial cell development and regulation of neuron differentiation, indicating their role in supporting neuronal maturation and maintaining homeostatic glial functions. For AAA, enriched terms included regulation of blood circulation and regulation of angiogenesis, indicating their role in vascular homeostasis and neurovascular coupling. For MAA and DMA, enriched terms included membrane lipid metabolic process and ATP metabolic process, indicating their role in energy and lipid metabolism under stress or disease conditions. In contrast, MIA exhibited downregulation of the metabolic pathways, indicating bioenergetic dysfunction or suppressed metabolic state potentially linked to pathological processes. The DEGs associated with metabolism and neurodegeneration were illustrated (Fig. 7E). The dashed lines in t-SNE plots roughly outlined the projected locations of DMA and MIA. Ptgds, Jun and Fos were expressed in DMA, with increased levels in surgery group, indicating an enhanced metabolic and neurotoxic function. Dio2 was expressed in MIA, with decreased levels in surgery group, indicating suppressed thyroid hormone metabolism.

IAM dominated postoperative glial network communication

Given the roles of microglia and astrocytes in amplifying or modulating neuroinflammation [45], the communications between glial cells and ECs after surgery were explored. CellChat analysis showed increased intercellular communications after surgery, with significantly enhanced signaling from microglia to microglia, astrocytes to astrocytes, and astrocytes to microglia. Their interaction numbers increased by 8.5%, 26.8%, and 84.0%, and interaction strengths rose by 30.6%, 15.6%, and 17.5%, respectively (Fig. 8A). In microglial subpopulations, IAM and TSM exhibited the highest total differential outgoing and incoming interaction numbers (IAM = 91, TSM = 85) and interaction strengths (both IAM and TSM = 2.7). In astrocytic subpopulations, HA and MAA exhibited the highest total differential outgoing and incoming interaction numbers (HA = 135, MAA = 111) and strengths (HA = 4.4, MAA = 4.6). Additionally, ECs also exhibited increased differential interaction numbers (60) and strengths (2.4, Fig. 8B, C). Given the perioperative changes of microglia, the signaling pathway alterations between control and surgery groups of microglia were further compared. Stacked bar plots (Fig. 8D) showed that the majority of signaling pathways were upregulated in surgery group. Notably, the signaling pathways related to extracellular matrix interaction (LAMININ), immune regulation (CD22, CD25), and vascular remodeling (ANGPTL) exhibited increased activity. This upregulation was observed in both signal-emitting and signal-receiving cells, indicating a bidirectional amplification of glial cell communication. Subsequent analysis of the IAM subpopulation revealed that among all differential pathways, the TNF signaling exhibited the largest increase in both incoming and outgoing interaction strengths, further highlighting its role in postoperative glial communication (Fig. 8E).

The major signaling pathways upregulated in microglia of surgery group were further analyzed. For TNF signaling pathway, chord diagram (Fig. 9A) showed that the interaction signals existed exclusively between microglial subpopulations. In control group, several subpopulations including HM, TSM, and IAM participated with moderate signaling activity. However, after surgery, IAM became the principal regulator of TNF signaling, functioning as the primary source of TNF ligands, the main recipient of TNF-related signals, and a key intermediary in signal propagation (red box, Fig. 9B). This shift suggested a central role of IAM in postoperative inflammatory response. Compared to control group, Tnf-Tnfrsf1a ligand-receptor (L-R) pair communication increased significantly in the IAM of surgery group (red box, Fig. 9C). Microglia-drived Tnf interacted with its receptor Tnfrsf1a, mediating various biological processes, including neuroinflammation and BBB damage [46, 47]. For BAFF pathway, chord diagram (Fig. 9D) showed that in control group, IAM primarily served as signaling initiators. After surgery, IAM displayed enhanced involvement by simultaneously producing BAFF ligands, responding to BAFF-related cues, and participating in signal relay within the microglial network (Fig. 9E). Compared to control group, Tnfrsf13b-Tnfrsf13b L-R pair communication increased in the IAM-dominated subpopulations of surgery group (red box, Fig. 9F), suggesting the involvement of other TNF-related signaling in IAM crosstalk. CHEMERIN is an adipokine with immunomodulatory properties. For the CHEMERIN pathway (Fig. 9G-H), astrocytes (HA, AAA, and MAA) acted as major signal emitters in control group, primarily targeting HM and TSM. After surgery, activated IAM emerged as the dominant recipients of astrocyte-derived CHEMERIN signals, implying a shift in astrocyte-microglia communication toward proinflammatory microglial states. Compared to control group, Rarres2-Cmklr1 L-R pair communication increased in surgery group from HA, AAA and MAA to IAM (red box, Fig. 9I). For APP pathway, chord diagram (Fig. 9J) showed that the interaction signals existed mainly from ECs to microglia. ECs served as the primary source of APP-related signals, targeting DIM as major recipients. IAM were actively involved as signal intermediates, suggesting their role in integrating vascular signals into inflammatory responses, potentially linking endothelial dysfunction with microglial activation (Fig. 9K). App-Cd74 L-R pair exhibited strong interactions from ECs to DIM, with a noticeable increasing trend in surgery group (Fig. 9L).

IAM intervention with TNF inhibitor restored PNCI in mice

To further explore potential cellular communications in the hippocampus following surgery, a hypothetical regulatory network was established based on identified microglial subpopulations and intercellular communication patterns (Fig. 10A). This network showed that IAM played a pivotal role in initiating inflammation through elevated TNF signaling, thereby influencing both astrocytic responses and endothelial dysfunction. Inflammatory genes Egr1 and C1qa were upregulated in astrocytic subpopulations (HA, DMA or MIA). It was hypothesized that TNF signaling in IAM targeted astrocytes, promoting reactive astrogliosis and an inflammatory astrocyte phenotype [48], characterized by the upregulation of Egr1 and C1qa. These changes were accompanied by a pro-inflammatory milieu, potentially impairing endothelial function via Cirbp upregulation and inflammasome activation [49]. In IAM, upregulated inflammatory genes included Tnf, Il1b, and Ccl2. Concurrently, ECs showed an increased expression of Icam1 and Ifitm3, suggesting endothelial responses to a pro-inflammatory environment. These transcriptional alterations reflected potential mechanisms under perioperative BBB damage. The ligands Csf1 (from MIA) and Rarres2 (from HA), along with their respective receptors Csf1r and Cmklr1 in IAM, were upregulated. The Csf1-Csf1r axis could promote microglial proliferation, driving IAM expansion, while the Rarres2-Cmklr1 axis could facilitate IAM migration via the MAPK pathway. Transmission electron microscopy revealed ultrastructural changes in hippocampal vasculature following surgery (Fig. 10B). In control group, ECs exhibited intact morphology, with continuous TJ and basement membranes (BM). In surgery group, ECs exhibited marked cytoplasmic vacuolization and BM fragmentation (red arrows), indicating endothelial barrier damage. Microglia were also observed in close anatomical proximity to ECs, accompanied by expanded rough endoplasmic reticulum, indicating increased protein synthesis activity associated with microglial activation. Collectively, these findings showed a putative inflammation-related intercellular network communication centered on IAM, providing a conceptual framework for understanding glial-mediated neuroinflammation in PNCI.

Etanercept, a clinically available TNF inhibitor, was selected to target IAM-driven inflammation. ELISA confirmed that hippocampal TNF-α levels were significantly increased in surgery group, and were effectively inhibited following etanercept treatment (Fig. 11B). Next, the expressions of major functional genes of regulatory network were validated with IF/FISH. The inflammatory genes Tnf, Ccl2, and Il1b were significantly upregulated in the hippocampal microglia of surgery group, while TNF inhibitor reduced their expressions to baseline levels (Fig. 10C). Consistently, immunofluorescence confirmed that TNF-α protein expression was also increased in the microglia of surgery group, and TNF inhibitor reduced their expressions to baseline. Egr1 and C1qa were significantly upregulated in astrocytes, and Icam1 and Ifitm3 were significantly upregulated in ECs along blood vessels of surgery group. TNF inhibitor reduced their expressions to baseline (Fig. 11A). Together, these findings confirmed TNF dominant glial network communications, and provided a potential therapeutic strategy for PNCI-related neuroinflammation.

Finally, the cognitive function of aged mice after surgery and TNF inhibition was assessed with Morris water maze (MWM) and fear conditioning test (FCT). In the MWM test, escape latency decreased significantly with ongoing training in control group, and this trend appeared less pronounced in surgery group, which could be rescued by TNF inhibitor treatment (p < 0.05 or p < 0.01, Two-way repeated-measures ANOVA followed by Bonferroni post-hoc test, Fig. 11C). There was no difference in swimming speed among the groups, excluding motor impairment (p > 0.05, Fig. 11D). Spatial memory, assessed by platform crossings and time spent in the target quadrant, was impaired after surgery (p < 0.05 or p < 0.01), and could be rescued by TNF inhibitor treatment (p < 0.05, Fig. 11E, F). Representative probe trial tracks were shown in Fig. 11G. In the FCT, hippocampus-dependent (contextual) memory was significantly impaired at both 1 day (p < 0.05, Fig. 11H) and 5 days (p < 0.01, Fig. 11I) after surgery, and TNF inhibitor treatment rescued these context cognitive impairments (p < 0.05 or p < 0.01). For hippocampus-independent (tone) cognition, two-way ANOVA analysis revealed no significant interaction between treatment (etanercept vs. saline) and condition (control vs. surgery, p > 0.05, Fig. 11J, K). Overall, these findings confirmed the establishment of the PNCI model, and the rescue effects of TNF inhibitor, highlighting the therapeutic potential of postoperative IAM and TNF signaling.