Integrin α7 Defines a Profibrotic and Proinflammatory Adipose Stromal Population Targeted for Nanoparticle PAI-1 Gene Silencing in Obesity

Abstract

Obesity-associated metabolic dysfunction is driven not only by adipocyte expansion but also by pathological remodeling of adipose tissue mediated by distinct stromal cell subpopulations. However, the lack of strategies to selectively interrogate and therapeutically modulate these pathogenic cells has limited clinical progress. Here, we identify integrin α7 (ITGA7)-expressing adipose stromal cells (ASCs) as a proinflammatory and profibrotic population that contributes to adipose tissue dysfunction in obesity and introduce a targeted nanotherapy to selectively reprogram their inflammatory signaling. ITGA7^high ASCs from obese white adipose tissue display a profibrotic, proinflammatory phenotype with impaired adipogenic capacity, exacerbated by inflammatory stress. To selectively target this subset, we engineered lipid-coated mesoporous silica nanoparticles functionalized with anti-ITGA7 antibodies, which exhibit preferential uptake by ITGA7^high ASCs in vitro and in vivo. Leveraging this platform, we delivered siRNA against plasminogen activator inhibitor-1 (PAI-1), a key regulator of obesity-associated inflammation and fibrosis. In obese mice, ITGA7-targeted siRNA nanotherapy selectively accumulates in white adipose tissue and effectively silences PAI-1 expression at both the tissue level and within ITGA7^high ASCs. Together, these findings establish ITGA7 as a therapeutically accessible marker of pathogenic ASC and define a molecular therapy framework for cell-selective genetic intervention within stromal niches in obesity.

Graphical Abstract

Obesity expands a pathogenic ITGA7^high adipose stromal cell population that promotes fibrosis. We engineered ITGA7-targeted lipid-coated mesoporous silica nanoparticles to selectively deliver PAI-1 siRNA, suppress profibrotic signaling, and restore a healthier adipose microenvironment.

Introduction

Obesity, a chronic imbalance between energy intake and expenditure, represents a global public health crisis and remains one of the most pressing medical challenges worldwide. In obesity, white adipose tissue (WAT) expands, and the resulting pro-inflammatory state contributes to the development of metabolic syndrome and insulin resistance^1–3. While the early research has focused on adipocytes and immune cells, the role of resident stromal cells in WAT remains less well understood. Adipose stromal cells (ASCs) are now increasingly recognized as a heterogenous population with cells contributing to obesity-associated fibrosis and inflammation, processes that exacerbate adipose tissue dysfunction⁴. The emerging evidence demonstrates the significant role of ASCs in sustaining chronic inflammation within WAT and highlights the therapeutic potential of targeting proinflammatory stromal subpopulations to restore healthy adipose tissue function.

The fate of ASCs is largely determined by biochemical and mechanical cues from the extracellular matrix (ECM), which are sensed and transmitted by integrins^5–7. Integrins are the main substrate adhesion molecules that cells use to mediate intracellular signaling in response to the ECM^8,9. However, little is known about the ASC integrin expression in pro-fibrotic and pro-inflammatory environments. Our recent findings demonstrate an increase in expression of integrin alpha 7 (ITGA7) in obese mice compared to lean controls¹⁰. Based on this observation, we hypothesize that ASCs expressing high levels of ITGA7 (ITGA7^high ASCs) represent a distinct proinflammatory and profibrotic stromal subpopulation that contributes to pathological adipose tissue remodeling in obesity. Leveraging ITGA7 as a surface-accessible marker therefore provides an opportunity to selectively interrogate and therapeutically modulate this pathogenic ASC subset using targeted nanotechnology, offering a novel strategy to attenuate inflammation and fibrotic signaling within obese adipose tissue.

Plasminogen activator inhibitor-1 (PAI-1) is a proinflammatory adipokine strongly associated with fibrosis, insulin resistance, and metabolic syndrome in obesity^11,12, and genetic or pharmacologic reduction of PAI-1 confers protection against obesity-associated metabolic dysfunction¹³. ASCs are the primary source of PAI-1 within adipose tissue¹⁴, where PAI-1 regulates adipogenesis¹⁵ and contributes to chronic inflammation during high-fat diet (HFD)-induced obesity¹⁶. In this study, we characterize ITGA7^high ASCs and exploit ITGA7 surface expression to enable selective delivery of siRNA targeting PAI-1. Specifically, we leverage the substantial promise of our previously reported lipid-coated mesoporous silica nanoparticle (LCMSN) platform^17,18 as a gene delivery and cell-targeting vehicle^19–21. LCMSNs combine key properties required for stromal cell-directed molecular therapy, including high adipose tissue retention, biocompatibility, high cargo capacity, and modular incorporation of targeting ligands^22–24. Targeting a profibrotic ASC subpopulation within white adipose tissue to modulate PAI-1 secretion thus represents a cell-selective molecular therapy approach²⁵ with potential to address obesity-associated adipose tissue dysfunction.

Results

ITGA7^high ASC Are Enriched in White Adipose Tissue in Obesity

To identify changes in ITGA7 expression that occur in adipose tissue with obesity, we used ob/ob genetic mice, which are morbidly obese resulting from a loss-of-function mutation in the gene encoding the satiety hormone leptin. Results were compared to age-matched wild-type (WT lean) mice. We first stained ITGA7 in white subcutaneous adipose tissue (sWAT) and analyzed its distribution and expression using confocal images. ITGA7 staining was observed in cells present in the stromal tissue of obese mice and control mice (Fig 1A). Obese mice exhibited a marked expansion of stromal content and increased cellular staining within the sWAT stroma. Staining localize to stromal cells with no observable staining of differentiated adipocytes. The area occupied by ITGA7 within a defined region of interest, independent of adipocyte size^26–28 was five times greater in the sWAT of obese mice compared to lean controls (237.8 ± 223.7 μm² in lean vs. 1666 ± 1077 μm² in obese; p = 0.016) (Fig. 1B). This immunofluorescence analysis suggested that obesity enhanced levels of stromal cells with high expression of ITGA7 (ITGA7^high) in sWAT. We next quantified ITGA7 expression in sWAT from lean and obese mice. ITGA7 gene expression was significantly elevated in the sWAT of obese mice compared to lean controls (Fig. 1C). To validate this enrichment at the cellular level, ASCs were isolated from lean and obese sWAT and sorted based on ITGA7 surface expression (Fig. 1D). Obese mice exhibited a greater than two-fold increase in the frequency of ITGA7^high ASCs relative to lean controls, indicating obesity-associated expansion of this stromal cell subset.

Fig 1:

ITGA7 is upregulated in subcutaneous adipose tissue of obese mice. (A) Two representative confocal images of subcutaneous adipose tissue from obese (ob/ob) and two from lean mice. Tissues were stained with antibodies against ITGA7 (red). Lipids were counterstained with BODIPY (green) and nuclei with DAPI (blue). n = 10 images analyzed from 3 animals per group. (B) Quantification of ITGA7-positive area reveals a ~600% increase in obese compared to lean tissue (237.8 ± 223.7 μm² in lean vs. 1666 ± 1077 μm² in obese; p = 0.016). (C) ITGA7 mRNA expression levels in subcutaneous adipose tissue of obese and lean mice. Values represent mean ± SEM from n = 4 animals per group, each run was done in triplicate. Statistical analysis by unpaired t-test; *p < 0.005, ***p < 0.001. (D,E) Flow cytometric quantification of ITGA7^high adipose stromal cells (ASCs) isolated from whole tissue of subcutaneous fat. The proportion of ITGA7^high ASCs was doubled in obese mice compared to lean controls.

ITGA7^highASC have lower adipogenic potential and poor metabolism

ASC isolated from WAT were sorted into ITGA7^high and ITGA7^low and exposed to white and beige adipocyte differentiation procedures. Differential lipid loading was observed between the ITGA7^low and ITGA7^high ASCs for both white and beige mature adipocytes (Fig 2A–B). ITGA7^low cells exhibited a robust pattern of lipid loading under both conditions consistent with differentiation observed in mixed ASC populations. Beige differentiation protocols resulted in cells with many small lipid vacuoles, while white adipose differentiation led to cells with larger lipid vacuoles. ITGA7^high cells, on the other hand, showed limited lipid loading in white and beige differentiation when compared to the ITGA7^low cells. Genetic markers of beige (UCP1 and Adiponectin) and white (PPARG and Adiponectin) differentiation were assessed between the two different subpopulations (Fig 2C–D). ITGA7 ^low cells had a significantly higher expression of Adiponectin and PPRG following exposure to white adipogenic media (WAM) compared to ITGA7^high cells (Fig 2C); as well as Adiponectin and UCP1 following beige adipogenic media (BAM) (Fig 2D). These results are consistent with the phenotypic differences observed between ITGA7^high and ITGA7^low cells, indicating that ITGA7^low cells had a greater capacity to become adipocytes compared to their ITGA7^high counterparts (Fig 2).

Fig 2: ITGA7 expression stratifies WAT adipose stromal cells by adipogenic, metabolic, and functional potential.

A-B Representative contrast phase images of adipocytes differentiated for 7 days in white adipogenic media (WAM) or beige adipogenic media (BAM) from (A) ITGA7^high and (B) ITGA7^low ASCs isolated from subcutaneous white adipose tissue (sWAT) of obese mice. (C-D). Relative mRNA expression of adipogenic and metabolic genes in differentiated adipocytes derived from ITGA7^high and ITGA7^low ASCs isolated from sWAT of obese mice (n = 3). (E-F). Insulin-stimulated glucose uptake measured in adipocytes differentiated from ITGA7^high and ITGA7^low ASCs following exposure to WAM or BAM for 7 days, assessed in the presence or absence of insulin. (G-H). Glycerol release measured as an index of lipolytic activity in differentiated adipocytes derived from ITGA7^high and ITGA7^low ASCs following WAM or BAM differentiation. Data represent n = 3 biological replicates with 3 technical replicates. *p < 0.05, **p < 0.005, ***p < 0.0005, ****p < 0.0001.

Given the distinct phenotypic and genetic profiles between ITGA7^high and ITGA7^low cells, we next evaluated metabolic function. Glucose uptake and glycerol metabolism were assessed following differentiation into white and beige adipocytes (Fig 2-E–H). White ITGA7^low adipocytes exhibited significantly higher glucose uptake compared to ITGA7^high adipocytes and robust response to insulin stimulation. In contrast, white ITGA7^high adipocytes showed minimal glucose uptake and were unresponsive to insulin (Fig 2E). A similar insulin resistance was observed in beige ITGA7^high adipocytes, whereas their ITGA7^low counterparts demonstrated increased glucose uptake (Fig 2F). As expected, glucose uptake was generally higher in beige adipocytes compared to white adipocytes. To further evaluate metabolic functionality, we measured glycerol release as an indicator of lipolysis. ITGA7^low adipocytes demonstrated much greater glycerol production than white ITGA7^high adipocytes under following both beige and white differentiation (Fig 2 G–H). Collectively, these results indicate that ITGA7^low cells exhibit enhanced adipogenic potential and metabolic function, while ITGA7^high cells are characterized by limited adipogenic differentiation capacity.

ITGA7^high ASC exhibit a pro-inflammatory phenotype

To determine whether ITGA7^high ASCs represent a proinflammatory stromal subset, we evaluated their inflammatory profile and responsiveness to proinflammatory stimuli. ITGA7^high ASCs were isolated by fluorescence-activated cell sorting (FACS) from the stromal vascular fraction of sWAT of 12–15-week-old ob/ob mice. Inflammation markers were evaluated and compared in the two different ASCs populations. Isolated ITGA7^high ASC had greater mRNA levels of the cytokines TNFα (Fig 3A) and IL6 (Fig 3B) and well-described proinflammatory chemokine CCL2 (Fig 3C). To evaluate how these ASC respond to pro-inflammatory signals we stimulated both ITGA7^high and ITGA7^low ASC with lipopolysaccharide (LPS) and examined changes in mRNA expression of proinflammatory markers. LPS stimulation in vitro induces chronic inflammation and reduces insulin signaling in adipocytes²⁹. 24 hours of LPS stimulation led to a rapid and robust induction in mRNA levels of TNFα (Fig 3D), IL6 (Fig 3E) and CCL2 (Fig 3F) in ITGA7^high ASC compared to ITGA7^low.

Fig. 3:

ITGA7^high ASCs are pro-inflammatory and respond to inflammatory stimulus. (A-C) Gene-expression analysis of ITGA7^high and ITGA7^low: mRNA levels of TNFα (A), IL6 (B), CCL2 (C). (D-F) mRNA levels of TNFα (D), IL6 (E), CCL2 (F) in ITGA7^high and ITGA7^low after stimulation with lipopolysaccharide (LPS) for 24 h. G-H: Representative data from two experiments with 3 replicates per experiment are shown. *p < 0.05, **p < 0.01, and ****p < 0.0001 by one-way ANOVA. In all panels, bars represent mean + SEM. n= 3 replicates. (G) Exposure to conditioned media from ITGA7^high ASCs increases pro-inflammatory gene expression in macrophages. Quantitative RT-PCR analysis of IL6 and TNFα mRNA expression in macrophages exposed to conditioned media from ITGA7^high versus ITGA7^low ASC. Macrophages cultured with ITGA7^high ASC-conditioned media showed significantly elevated expression of both pro-inflammatory cytokines compared to those exposed to ITGA7^low ASC-conditioned media. Data are presented as mean ± SEM. p < 0.05 by unpaired two-tailed t-test; n = 6 replicates.

Conditioned media was used to investigate whether ITGA7^high ASCs, influence macrophage activation. Macrophage were exposed to conditioned media derived from ITGA7^high or ITGA7^low ASCs isolated from obese subcutaneous adipose tissue (Fig 3G). Macrophage treated with ITGA7^high ASC-conditioned media exhibited significantly increased expression of the pro-inflammatory cytokines IL6 and TNFα, compared to those treated with ITGA7^low ASC-conditioned media (Fig 3G). These results suggest that ITGA7^high ASC secrete soluble factors capable of modulating macrophage polarization toward a pro-inflammatory state. These findings are consistent with prior reports in other fibrotic models showing that alterations in stromal cell mitochondrial metabolism or phenotype can influence adipose tissue macrophage composition and inflammatory tone³⁰, supporting a model in which ITGA7^high ASCs contribute to immune dysregulation in obese adipose tissue.

ITGA7^high ASCs are enriched for profibrotic gene expression

Our next goal was to assess whether these ITGA7^high ASCs are profibrotic. Adipocyte progenitors with high levels of CD9 (CD9^high) controls obesity-induced adipose tissue fibrosis^31,32. ASCs from sWAT from ob/ob and WT mice were isolated and fluorescently labeled with ITGA7 monoclonal antibody (MA5–23608, Invitrogen) conjugated with R-phycoerythrin, and CD9 monoclonal antibody conjugated with FITC (124807, BioLegend). Flow cytometry was then used to analyze ASC co-expression of ITGA7 and CD9. Results showed that 88% of the ITGA7^high ASCs also expressed high levels of CD9^high in ASCs from ob/ob mice (Fig 4A). The percentage of ITGA7^high cells that expressed high levels of CD9 in ASC isolated from lean animals was 47.2% (Fig 4B). These results suggest that ITGA7^high cells exhibit a more fibrotic phenotype in an obese microenvironment. Fluorescent images of ITGA7^high cells show high fluorescence in CD9 simultaneously, compared to the ITGA7^low cells that exhibit little to none fluorescence for CD9. We then tested the expression of other genes associated with fibrosis including collagen 1A1 (COL1A1, Fig 4D), collagen 3A1 (COL3A1, Fig 4E) and PAI-1 (Fig 4F). Expression of these markers was elevated in ITGA7^high ASC subpopulation compared to ITGA7^low, with statistically significant differences in COL3A1 and PAI-1. Together these results suggest that in obesity the increase levels of ASC expressing ITGA7 are correlated with an increase in pro-fibrotic cells.

Fig 4:

ITGA7^high ASCs identify pro-fibrotic cell subsets with increased CD9 expression. (A-B): Quantification of the percentage of ITGA7^high ASCs that are also CD9^high. (C-D): Representative images captured by the Amnis ImageStreamX Flow Cytometer of ITGA7 and CD9 staining in ITGA7^high (C) and ITGA7^low (D) ASC (derived from subcutaneous fat). First column shows brightfield images of the cells, second column shows images of fluorescence of CD9, third column shows ITGA7 fluorescence, and fourth column represents side scatter signal. (E-G) Relative mRNA expression of fibrotic genes COL1A1 (E), COL3A1 (F), PAI-1(G) measured in ITGA7^high and ITGA7^low ASCs isolated from iWAT of ob/ob mice. *p < 0.05, ***p < 0.001. Data are expressed as mean ± SEM. n=3 biological replicates with 3 technical replicates per group.

PAI-1 Silencing Attenuates Inflammatory Programs in ITGA7^high ASC

Given the elevated expression of fibrotic and inflammatory markers in ITGA7^high ASCs, we sought to determine whether these cells also exhibited functional differences in cytokine secretion. One of the most significantly upregulated pro-fibrotic genes in ITGA7^high ASCs was PAI-1 (Fig. 4G), a serine protease inhibitor known to play a critical role in extracellular matrix remodeling and fibrosis. To confirm that ITGA7^high cells from sWAT secrete higher levels of PAI-1, we quantified PAI-1 concentrations in the secretome using ELISA. The results showed that PAI-1 secretion was nearly twice as high in ITGA7^high cells compared to ITGA7^low cells (Fig 5A). Importantly, PAI-1 has also been implicated in promoting chronic inflammation in adipose tissue by modulating immune cell recruitment and activity¹⁶. These dual roles make PAI-1 a compelling candidate for further investigation as a mediator of the dysfunctional inflammatory and fibrotic phenotype of ITGA7^high ASCs. These findings support the role of PAI-1 as a potential therapeutic target to mitigate the pro-inflammatory and pro-fibrotic phenotype associated with ITGA7^high cells.

Fig 5.

Effect of PAI-1 knockdown on ITGA7^high ASC inflammatory response. (A) PAI-1 secreted protein found in the secretome of both ITGA7^high and ITGA7^low ASCs measured by ELISA (B) Efficacy of PAI-1 siRNA in knocking down levels of PAI-1 in both ITGA7^high and ITGA7^low ASC (****p < 0.0001). (C) IL6 and (D) TNFα mRNA expression in PAI-1 knockdown cells with and without LPS stimulus. n = 3 experiments with 3 replicates per group per experiment. Values are means ± standard errors, from one-way ANOVA analysis followed with Tukey’s multiple comparisons test.

To evaluate the role of PAI-1 in modulating inflammation in ITGA7^high ASCs, we silenced PAI-1 using Lipofectamine-mediated transfection of PAI-1 siRNA, followed by stimulation with LPS. Both ITGA7^high and ITGA7^low ASCs were transfected with PAI-1 siRNA for 48 hours. Transfection resulted in a 75% reduction in PAI-1 mRNA levels (Fig. 5B). Following transfection, cells were stimulated with LPS for 24 hours. As previously observed (Fig. 5), LPS treatment significantly upregulated IL6 and TNFα expression in ITGA7^high ASCs compared to ITGA7^low ASCs (Fig. 5C–D). Notably, silencing PAI-1 markedly reduced IL6 expression in ITGA7^high ASCs, reducing it to levels comparable to ITGA7^low cells (Fig. 5C). Similarly, TNFα expression was significantly decreased in PAI-1-silenced ITGA7^high ASCs following LPS stimulation (Fig. 5D), suggesting that PAI-1 plays a key role in mediating the heightened inflammatory response of ITGA7^high ASCs. To further establish the translational relevance of our findings, primary human adipose stromal cells (hASCs) isolated from obese donors were evaluated for PAI-1 expression and susceptibility to PAI-1 silencing. hASCs exhibited robust PAI-1 expression at both the mRNA and protein levels. Transfection with PAI-1 siRNA significantly reduced PAI-1 gene expression, which was accompanied by a marked decrease in PAI-1 protein as determined by immunofluorescence staining (Fig. S1). These findings demonstrate that PAI-1 can be effectively suppressed in primary human obese ASCs, supporting the translational relevance of targeting this pathway in obesity-associated adipose tissue dysfunction.

Engineered ITGA7-Targeted LCMSNs Support Targeted PAI-1 siRNA Delivery

The fabrication of siRNA-loaded, anti ITGA7 LCMSN was carried out according to the schematic in Fig 6A. MSN were first designed with a pore size centered at 5 nm and a positively charged surface using APTES as a co-condensation moiety (Fig 6B). In parallel, liposomes (DPPC/DMPG/cholesterol/DSPE PEG2K) were made via solvent evaporation followed by sonication-assisted hydration in PBS, characterized and freshly used to form the LCMSN. The liposomes size was 208 ± 4 nm with a PDI = 0.29 ± 0.02 and ζ = −29 ± 3 mV. The hydrodynamic size of the MSN and the subsequent construct is shown in Fig 6D and shows an increase of the size from 88 nm to 148 and 183 nm due to the incorporation of the lipid bilayer and antibody. The hydrodynamic size increase surpassed the theoretical increase upon addition of siRNA, lipids and antibody, this was explained by TEM evidence shows several instances of lipid coating 2–4 MSNs with a physical size up to 172 nm (Fig S2). The surface modification is further supported by a zeta potential evolution from highly positive MSN to ~neutrally charged surface upon PAI-1 siRNA addition and lipid fusion, as well as anti ITGA7 conjugation.

Fig 6:

Fabrication and Characterization of nanoparticles. (A) Schematic depicting the fabrication of the LCMSN construct with siRNA loading, liposomes fusion with neutravidin, ITGA7 antibody conjugation. (B) TEM images of implemented MSN and (C) negatively stained particles upon construct formation (scale bar = 100 nm) D: Hydrodynamic size, and zeta potential evolution of MSN, loaded MSN with lipids and full construct LCMSN with ITGA7 antibody (AB). E: siRNA release % from lipid-free and lipid-coated MSN.

The negatively stained LCMSNs are shown in Fig 6C where the halo/outer layer is indicated by white arrows and emphasizing the soft lipid shell comparable to our recent observations^33,34. The siRNA loading was assessed using a cy3-siRNA and data shows that siRNA added (2% wt) was 100% loaded. Release data were recorded upon incubation of siRNA-loaded nanoparticles with PBS and following our previously reported procedure^22–24. Data show that the siRNA is gradually released within 4–7 hours. To evaluate the durability of PAI-1 silencing with ITGA7-LCMSN delivery, ITGA7^high ASCs were treated with ITGA7-LCMSN-siPAI-1 and PAI-1 gene expression was assessed over time. One-way ANOVA revealed significant differences among time points (F(3,32) = 23.79, p < 0.0001). Post hoc Tukey analysis demonstrated significant reductions in PAI-1 expression at 48 h, 72 h, and 96 h compared with the 6 h time point (all p < 0.0001), whereas no significant differences were observed among the 48 h, 72 h, and 96 h groups. These findings indicate that ITGA7-LCMSN-mediated siRNA delivery produces sustained PAI-1 suppression for at least four days following a single treatment (Fig. S3). To provide complementary protein-level evidence of siRNA activity in obese ASCs, stromal cells isolated from ob/ob adipose tissue were treated with fluorescently labeled LCMSN-siPAI-1 and analyzed by immunofluorescence staining (Fig. S4). Treated cells exhibited reduced PAI-1 protein expression compared with untreated controls, while fluorescent nanoparticles were readily detected intracellularly (Fig. S4A,B). Quantification of PAI-1 immunofluorescence intensity confirmed a reduction in PAI-1 protein levels following treatment (Fig. S4C). To further validate these findings in vivo, adipose tissue sections from ob/ob mice treated with LCMSN-siPAI-1 were analyzed by immunofluorescence staining. Consistent with the in vitro and RT-qPCR results, treated animals exhibited reduced PAI-1 protein expression within adipose tissue compared with control mice (Fig. S4D–E), providing tissue-level confirmation of effective PAI-1 silencing following nanoparticle treatment. Together, these findings demonstrate that LCMSN-mediated siRNA delivery effectively suppresses PAI-1 protein expression both in cultured obese ASCs and in adipose tissue in vivo.

ITGA7-LCMSN selectively target ITGA7^high ASCs in vivo

Targeted and untargeted LCMSN were injected into the subcutaneous adipose tissue of mice. The tissue was harvested 48 h after injection, and ITGA7^high and ITGA7^low ASC populations isolated. Flow cytometry analysis revealed significantly higher nanoparticle fluorescence (NP+) in ITGA7^high ASCs compared to ITGA7^low ASCs (Fig. 7A). Two-way ANOVA revealed a significant interaction between cell population and nanoparticle formulation (F(1,8) = 9.03, p = 0.017), indicating that the effect of targeting depended on ITGA7 expression levels (Fig. 7B). Significant main effects of both cell population (F(1,8) = 498.9, p < 0.0001) and nanoparticle formulation (F(1,8) = 17.05, p = 0.0033) were also observed. Post hoc Fisher’s LSD analysis demonstrated that targeted LCMSNs exhibited significantly greater uptake than untargeted nanoparticles in ITGA7^high ASCs (p = 0.0010), whereas no significant difference was detected in ITGA7^low ASCs (p = 0.4491). 80% of ITGA7^high ASCs internalized ITGA7-LCMSN, whereas less than 20% of ITGA7^low ASCs showed detectable uptake (Fig. 7B). Furthermore, uptake of targeted LCMSNs was significantly greater in ITGA7^high ASCs than in ITGA7^low ASCs (p < 0.0001), supporting selective nanoparticle internalization by ITGA7-expressing cells. Representative images obtained using flow cytometry further confirmed that ITGA7-LCMSN uptake was specific to ITGA7^high ASCs, with strong colocalization of ITGA7 and LCMSN fluorescence signals (Fig. 7C). These data demonstrate that ITGA7-LCMSN are internalized by ITGA7^high ASCs in vivo, supporting our system as an effective targeting mechanism for ASC pathologic subpopulation. To evaluate the biocompatibility of the nanoparticle formulations used in this study, primary ITGA7^high ASCs isolated from obese mice were treated with targeted and untargeted LCMSNs loaded with PAI-1 siRNA, scramble siRNA, or unloaded nanoparticles. Cytotoxicity was assessed by LDH release assay. All nanoparticle formulations exhibited minimal cytotoxicity, with cell death remaining below 6% across all treatment groups (Fig. S5), demonstrating excellent cellular compatibility.

Fig 7.

ITGA7-LCMSN preferentially accumulate in adipose tissue and selectively target ITGA7^high adipose stromal cells (ASCs) after in vivo injection. A: Flow cytometry analysis showing ITGA7-LCMSN uptake in ITGA7^high and ITGA7^low ASCs isolated from adipose tissue after in vivo injection. B: Percentage of ITGA7^high and ITGA7^low ASCs positive for LCMSN uptake following treatment with targeted or untargeted LCMSNs. Two-way ANOVA revealed significant effects of cell population (F(1,8) = 498.9, p < 0.0001), nanoparticle formulation (F(1,8) = 17.05, p = 0.0033), and a significant interaction between factors (F(1,8) = 9.03, p = 0.017), indicating that the effect of targeting depended on ITGA7 expression levels. Fisher’s LSD post hoc analysis demonstrated significantly greater uptake of targeted LCMSNs compared to untargeted LCMSNs in ITGA7^high ASCs (***p = 0.001), whereas no significant difference was observed in ITGA7^low ASCs (ns, p = 0.4491). Uptake of both targeted and untargeted LCMSNs was significantly higher in ITGA7^high ASCs than in ITGA7^low ASCs (****p < 0.0001). n= 3 biological replicates C: Representative images acquired using the Amnis ImageStreamX Flow Cytometer showing colocalization of ITGA7 (green) and LCMSN (red) fluorescence, confirming that nanoparticle uptake is specific to ITGA7^high ASCs. Data indicate that ITGA7-LCMSN effectively bind to and are internalized by ITGA7-expressing ASCs. D: Schematic of DyLight 633-labeled ITGA7-LCMSN administration in obese mice via dual subcutaneous (s.c.) and intraperitoneal (i.p.) injection routes. E: Whole-body fluorescence imaging acquired 72 h after s.c. and i.p. administration using the Kino imaging system, demonstrating persistent fluorescence signal in adipose-rich regions for both delivery routes. F: Quantification of whole-body fluorescence intensity (arbitrary units), showing significantly elevated fluorescence following ITGA7-LCMSN administration. G: Ex vivo fluorescence imaging of major organs and adipose depots collected 48 h post-injection. Strong fluorescence is observed in subcutaneous and visceral adipose tissues, with minimal signal detected in liver and negligible fluorescence in other organs. H: Quantification of ex vivo fluorescence intensity (% of total signal), confirming preferential accumulation of ITGA7-LCMSNs in adipose tissue with limited off-target uptake (n = 3 biological replicates).

ITGA7-Targeted LCMSNs Preferentially Accumulate in Adipose Tissue and Enable Efficient In Vivo PAI-1 Gene Silencing

ITGA7-LCMSN were injected both subcutaneously (s.c.) and intraperitoneally (i.p.) into ob/ob mice to evaluate the in vivo biodistribution and fat-targeting capacity of fluorescently labeled ITGA7-LCMSN (Fig. 7D). Intraperitoneal injections were used to specifically access visceral adipose depots. Whole-body fluorescence was assessed 48 hours post-injection using the Kino in vivo imaging system. As shown in Fig. 7E, the combined administration routes resulted in detectable nanoparticle signal throughout the body; however, the most prominent accumulation was observed in adipose-rich regions. Notably, mice displayed strong, localized fluorescence in the abdominal cavity consistent with visceral fat uptake, as well as intense signal at peripheral adipose depots. Quantification of whole-body fluorescence (Fig. 7F) confirmed persistent retention of nanoparticles at 48 hours. To further determine tissue-specific biodistribution, major organs and adipose depots were collected at 48 hours post-injection and imaged ex vivo using the Kino system (Fig. 7G). Fluorescence was markedly enriched in both subcutaneous (42.10%) and visceral (45.94%) adipose tissues. In contrast to classical biodistribution patterns observed for many nanoparticle platforms, only a relatively low fluorescence signal (9.13%) was detected in the liver, and negligible signal was observed in the spleen, kidney, lungs, or heart. Quantification of ex vivo fluorescence efficiency (Fig. 7H) corroborated these findings, demonstrating that adipose tissues were the primary sites of ITGA7-LCMSN retention. Collectively, these results demonstrate that ITGA7-LCMSN using these administration routes, preferentially accumulate in adipose tissue following both s.c. and i.p. administration, with limited off-target uptake by filtering organs. To further assess the temporal biodistribution and retention of ITGA7-LCMSNs, fluorescently labeled nanoparticles were administered subcutaneously to ob/ob mice and tissue distribution was evaluated at 4 h and 4 days post-injection (Fig. 8 A–D). Whole-body imaging revealed persistent nanoparticle localization within adipose-rich regions at both time points (Fig. 8A–B). Ex vivo analysis demonstrated that adipose tissue remained the primary site of nanoparticle accumulation, whereas minimal fluorescence was detected in classical clearance organs, including the liver, spleen, kidney, lungs, and heart (Fig. 8 C–D). Notably, the adipose-enriched biodistribution profile was maintained at 4 days following administration, indicating prolonged retention within adipose tissue and limited redistribution to off-target organs over time. These findings further support the favorable tissue specificity and retention properties of the ITGA7-LCMSN platform for adipose tissue-targeted therapeutic delivery. To determine whether ITGA7-targeted nanoparticles can selectively modulate gene expression in ITGA7^high ASC, we loaded the ITGA7-LCMSN with siPAI-1 and evaluated gene silencing efficiency in vivo after 48 h of subcutaneous injection. ITGA7^high cells isolated from mice treated with ITGA7-LCMSN -siPAI-1 exhibited a robust and significant reduction in PAI-1 mRNA levels compared to untreated control group (Fig 8E), demonstrating effective and cell-specific knockdown within the targeted stromal population. Importantly, this targeted silencing translated into tissue-level effects: whole adipose tissue lysates from treated mice showed a significant decrease in total PAI-1 expression (Fig 8F). Together, these results demonstrate that ITGA7-targeted LCMSNs achieve efficient and selective in vivo PAI-1 gene silencing within adipose tissue, supporting their potential as a molecular therapy for obesity-associated adipose dysfunction.

Fig 8:

ITGA7-targeted LCMSNs achieve adipose tissue retention, functional siRNA delivery, and modulation of profibrotic remodeling in vivo. A-D: Fluorescently labeled ITGA7-LCMSNs were administered subcutaneously to ob/ob mice and biodistribution was evaluated at 4 h and 4 days post-injection. A: Representative whole-body fluorescence images and B: quantification of fluorescence signal at 4 h and 4 days following subcutaneous administration of fluorescently labeled ITGA7-LCMSNs, demonstrating sustained nanoparticle retention within adipose-rich regions over time. C: Ex vivo fluorescence distribution across major organs and adipose tissue depots at 4 h and D: at 4 days post-injection. Adipose tissue remained the predominant site of nanoparticle accumulation at both time points, whereas minimal fluorescence was detected in classical clearance organs including the liver, spleen, kidney, lungs, and heart. E: PAI-1 mRNA expression in ITGA7^high ASCs isolated by flow cytometric sorting 48 h after in vivo administration of ITGA7-LCMSN-siPAI-1, showing a significant reduction compared with untreated controls. F: Whole-tissue PAI-1 mRNA expression in adipose depots harvested 48 h post-treatment; qPCR of bulk lysates shows a significant decrease in mice receiving ITGA7-LCMSN-siPAI-1, indicating that cell-selective stromal silencing translates to reduced tissue-level PAI-1. G: Whole-tissue PAI-1 mRNA expression after long-term treatment (twice weekly for four weeks) with targeted or untargeted LCMSN-siPAI-1; both formulations reduced adipose PAI-1 expression. H-J: mRNA expression of the profibrotic markers CD9 (H), COLA1 (I), and COL3A1 (J) following long-term treatment. ITGA7-targeted LCMSN-siPAI-1 produced a greater reduction in all three markers than untargeted nanoparticles, supporting enhanced modulation of the ITGA7high profibrotic stromal population. Statistical analysis for (E, F) was performed using an unpaired t-test; **p < 0.01. n = 3 biological replicates. For (G-J), data are mean ± SEM, n = 3 per group, analyzed by one-Way ANOVA **p < 0.01

To determine whether PAI-1 silencing translates into functional changes in adipose tissue remodeling, we performed a long-term proof-of-concept study in ob/ob mice treated twice weekly for four weeks with targeted or untargeted LCMSN-siPAI-1. Consistent with the short-term studies, both formulations reduced adipose tissue PAI-1 expression (Fig. 8G). Because PAI-1 is expressed by multiple cell types within obese adipose tissue, including adipocytes, macrophages, endothelial cells, and stromal populations, whole-tissue PAI-1 expression may not fully reflect selective modulation of the ITGA7^high ASC population targeted by our system. Importantly, when evaluating markers associated with the profibrotic ITGA7^high ASC phenotype, targeted nanoparticles demonstrated greater efficacy. ITGA7-targeted LCMSN-siPAI-1 produced a stronger reduction in CD9, Col1a1, and Col3a1 expression compared with untargeted nanoparticles (Fig. 8H–J), supporting enhanced modulation of fibrosis-associated pathways linked to the pathogenic stromal cell population identified in this study. Assessment of glucose levels throughout the treatment period revealed trends toward improved glucose regulation in treated animals (Fig. S6), although the study was not powered to detect metabolic endpoints.

Discussion

Adipose tissue dysfunction is a major contributor to the metabolic consequences associated with obesity. Fibrotic remodeling is now recognized as a key driver of WAT impairment and is implicated in the progression of obesity-related metabolic complications³⁵. Excessive WAT fibrosis can diminish the benefits of interventions such as bariatric surgery^36,37. Thus, engineering innovative platforms capable of selectively modifying profibrotic pathways in WAT may yield promising therapeutic strategies.

Our findings reveal a previously unknown role for ITGA7-expressing ASCs as key cellular effectors driving ECM remodeling and inflammatory dysfunction within adipose tissue. We demonstrate that ITGA7^high ASCs represent a distinct, pathogenic stromal subset characterized by elevated secretion of profibrotic mediators such as PAI-1 and the ability to potentiate macrophage pro-inflammatory activation. These results position ITGA7^high ASCs as critical contributors to the maladaptive fibrotic rewiring of adipose tissue that accompanies obesity and contributes to systemic metabolic decline. Importantly, the increased prevalence and intensified activity of ITGA7^high ASCs relative to ITGA7^low populations suggest that specific stromal cell subsets within the adipogenic niche differentially regulate matrix deposition, immune recruitment, and tissue homeostasis. This highlights the need to move beyond treating ASCs as a uniform therapeutic target and instead recognize their functional heterogeneity and the subset-specific roles they play in disease progression.

A recent study demonstrated that CD9⁺ adipose progenitors strongly correlate with adipose fibrosis and impaired glycemic control in humans³⁸. CD9⁺ ASCs remained tightly associated with insulin resistance and visceral fat accumulation, suggesting their involvement in the pathological link between visceral obesity and type 2 diabetes. Notably, in patients with T2D undergoing bariatric surgery, elevated CD9⁺ ASCs predicted reduced BMI loss and diminished diabetes remission, indicating that their accumulation may reflect underlying adverse tissue remodeling that limits metabolic improvement³⁸. Consistent with this, our results show that more than 80% of ASCs with high ITGA7 expression also exhibit high CD9 expression, further supporting a shared pathogenic phenotype. Rebiere et al. also demonstrated that pharmacologic inhibition using the FDA-approved drugs nintedanib and celecoxib reduced CD9⁺ ASC expansion and fibrotic activation, preserved visceral white adipose tissue architecture, and improved metabolic outcomes in obese mice³⁸. However, these conventional anti-fibrotic therapies exhibited depot-specific limitations: while they reduced fibrosis, it failed to regenerate the CD9⁻ progenitor pool in visceral WAT once fibrosis was already established, indicating that this approach does not restore adipogenic capacity in that depot³⁸. Also, the secondary effects of these drugs when administered systemically, and without targeted delivery mechanisms, remain a significant concern. Our platform addresses these limitations by enabling cell-subset specific targeting, allowing genetic modulation directly to profibrotic ASCs while minimizing off-target exposure. Consistent with this concept, long-term treatment studies revealed that both targeted and untargeted LCMSN-siPAI-1 formulations reduced whole-adipose tissue PAI-1 expression. This finding is not unexpected, as PAI-1 is produced by multiple adipose-resident cell populations, including adipocytes, macrophages, endothelial cells, and stromal cells. Consequently, whole-tissue PAI-1 expression may represent a relatively nonspecific readout of selective targeting of ITGA7^high ASCs. In contrast, markers such as CD9, Col1a1, and Col3a1 are more directly associated with the profibrotic stromal phenotype identified in this study. Notably, ITGA7-targeted nanoparticles produced greater suppression of these fibrosis-associated genes than untargeted nanoparticles, suggesting enhanced modulation of the pathogenic ITGA7^high ASC population. These findings indicate that the therapeutic benefit of ITGA7-directed delivery extends beyond PAI-1 suppression alone and lies in its ability to selectively remodel the profibrotic stromal niche that contributes to adipose tissue dysfunction in obesity. While larger studies will be required to fully define the metabolic consequences of this approach, the observed reduction in fibrosis-related genes supports further investigation of ITGA7-targeted therapies as a strategy to improve adipose tissue health and limit obesity-associated adipose remodeling. Our targeted therapy offers an opportunity to directly re-engineer the fate of pathogenic ASC subsets. By precisely modulating pro-fibrotic and inflammatory ASC populations, our platform has the potential to genetically modify these cells toward a more regenerative, preadipocyte-like state and expand the progenitor pool in a controlled, depot-specific manner. Thus, beyond its anti-fibrotic and anti-inflammatory effects, this targeted approach opens new avenues to promote functional adipose tissue remodeling, improve metabolic health, and restore adipogenic plasticity where systemic pharmacologic treatments fall short. Collectively, these findings highlight the emerging importance of ITGA7^high ASC and the therapeutic potential of precision molecular targeting profibrotic cells in obesity and T2D. While our findings demonstrate that ITGA7^high ASCs are enriched in obesity and exhibit a distinct pro-inflammatory and pro-fibrotic phenotype within the obese adipose microenvironment, future studies will be needed to determine how this population emerges during obesity progression and how its molecular characteristics compare to rare ITGA7^high cells present under lean conditions. Such studies may further clarify the obesity-specific mechanisms that drive the expansion and pathogenic activity of ITGA7high ASCs.

Our work further demonstrates that these pathogenic ITGA7^high ASCs can be precisely targeted in vivo using engineered antibody-bearing nanoparticles ITGA7-LCMSN, providing proof-of-concept that integrin-guided nanomedicine can selectively engage specific stromal cell populations to modulate pharmacologically. An important consideration for translational nanoparticle therapies is the ability to achieve efficient intracellular delivery while maintaining cellular viability. Here, all LCMSN formulations displayed excellent biocompatibility, inducing less than 6% cytotoxicity in primary ITGA7^high ASCs. In addition, nanoparticle-associated fluorescence and the successful silencing of PAI-1 in adipose-derived ASCs provide functional evidence of nanoparticle internalization and intracellular siRNA release. These findings support the safety and efficacy of the LCMSN platform for selectively modulating pathogenic ASC populations within obese adipose tissue. We have recently shown that lipid-coated silica nanoparticles are retained for at least five days in the adipose tissue, but without identifying the target cell of accumulation. Herein, the strong retention in visceral fat is attributed to i.p. route delivery highlights a promising route for depot-specific nanoparticle targeting in future therapeutic applications. The design of a targeting system that accumulates specifically in subcutaneous and visceral fat (>80%), while avoiding classical MPS sequestration, demonstrates the feasibility of using nanotechnology to deliver molecular therapeutic cargo directly to dysfunctional ASC subsets. This allowed the delivery of siRNA that targets the ITGA7^high subset and reduce the PAI-1 gene associated with fibrotic behavior. Additional biodistribution studies performed at short (4 h) and long (4 d) times following administration demonstrated that adipose tissue remained the dominant site of nanoparticle retention over time, with minimal accumulation in classical clearance organs. These findings suggest that the favorable biodistribution profile of ITGA7-LCMSNs is maintained beyond the 48 h endpoint and supports prolonged residence within adipose tissue. Together, these findings establish ITGA7^high ASCs as both mechanistic drivers of adipose fibrosis and inflammation and as actionable molecular therapeutic targets, offering new opportunities to reprogram the extracellular matrix and restore metabolic health in obesity-associated diseases.

The accumulation of immune cells within adipose tissue is a central driver of obesity-associated inflammation³⁹. WAT expansion in obesity leads to the accumulation of proinflammatory macrophage, contributing to metabolic syndrome and insulin resistance^1–3. While extensive work has characterized how immune populations such as T cells and macrophages contribute to this process⁴⁰, the stromal cues that orchestrate their infiltration remain less well understood. A recent study shows the role of ASCs in regulating T cell accumulation in WAT during the early phases of obesity⁴¹. Also, single-cell RNA sequencing data identified a subpopulation of stromal cells which promote inflammation termed fibroadipogenic inflammatory precursors (FIPs)⁴. Further investigation showed FIPs’ critical role in modulating macrophage levels in WAT, particularly under metabolic inflammation. Mouse models revealed that FIPs’ activation sustains macrophage accumulation in obesity. These findings highlight ASC’s significant role in chronic adipose tissue inflammation and suggest targeting proinflammatory stromal cells as a potential therapeutic approach for healthier adipose tissue function. In our study, we demonstrate that ITGA7^high ASCs function as potent pro-inflammatory effectors, capable of altering macrophage phenotype and promoting a shift toward a more inflammatory state. These findings suggest that dysfunctional ASC subsets can actively shape the immune landscape of obese adipose tissue, suggesting that ECM composition may also regulate the extent and localization of immune cell recruitment. Understanding how ECM remodeling influences immune cell infiltration, and how immune-stroma feedback loops further exacerbating matrix deposition, will be an important direction for future studies. Together, these results highlight a mechanistic link between pathogenic ASC subsets, ECM dynamics, and immune activation during obesity progression.

The beginning of single-cell transcriptomics has revolutionized our view of adipose tissue, uncovering an unexpected diversity of stromal, immune, and progenitor populations and revealing how their transcriptional states drive tissue plasticity. Although this complexity presents major challenges, it also opens extraordinary opportunities to decode the cellular and molecular programs that preserve adipose health, and to identify those that collapse in pathologic obesity. Leveraging these insights, our work introduces what is, to our knowledge, the first example of a nanoparticle-based platform specifically engineered to target fibroadipogenic adipose stromal cell subsets for genetic manipulation. This approach goes beyond observation to active, cell-type-specific reprogramming, offering unprecedented precision in modulating disease-driving stromal populations. Such engineered tools hold transformative potential: they not only deepen our mechanistic understanding of adipose tissue biology but also project a path toward next-generation therapies capable of reshaping dysfunctional adipose niches and treating a spectrum of metabolic diseases.

Conclusion

Together, these studies establish ITGA7^high ASC as central components of the fibrotic, inflammatory, and metabolic deterioration that characterizes obesity, positioning them as a previously unrecognized but highly actionable molecular therapeutic target. By defining their pathogenic phenotype, elucidating their role in reshaping the ECM and immune microenvironment, and demonstrating that they can be selectively engaged in vivo using integrin-guided nanomedicine, our work introduces a new conceptual and technological framework for reprogramming dysfunctional stromal niches in adipose tissue. This precision-targeted platform overcomes key limitations of systemic anti-fibrotic therapies and opens a path toward restoring adipogenic plasticity, attenuating inflammation, and improving metabolic function through cell-subset-specific intervention. As single-cell technologies continue to reveal the heterogenity of adipose tissue architecture, our findings provide a foundation for next-generation therapeutic strategies that modulate disease-driving stromal populations with specificity. Ultimately, the ability to re-engineer pathogenic ASC subsets represents a transformative opportunity to reshape adipose tissue biology and develop innovative treatments for obesity, type 2 diabetes, and related metabolic disorders.

Materials and Methods

Animal model

Adult female and male mice ages 12–15 weeks, B6.Cg-Lep^ob/J homozygous (ob/ob, obese) and heterozygous (control lean) were acquired from Jackson Laboratory. The mice were fed standard chow diet ad libitum, and all animal procedures were approved by the IACUC at the University of Texas at San Antonio.

Tissue harvesting and stem cell isolation

Mice were sacrificed and subcutaneous white adipose tissue (WAT) was harvested from obese and lean animals. The WAT was minced and enzymatically digested using Collagenase type 1 (Worthington Biochemical Co.) in accordance with previously published protocol^17,18. The mature adipocyte and lipid top layer were discarded, and the stromal vascular fraction was isolated through various centrifugation with Phosphate Buffer Solution (PBS) washes and filtration steps.

Tissue Confocal Imaging and Analysis

Inguinal white adipose tissue (iWAT) from WT lean and obese ob/ob mice was fixed overnight at 4°C in 4% paraformaldehyde (PFA), followed by three 10-minute PBS washes. Tissues were permeabilized in 0.5% Triton X-100 in PBS for 1 hour at room temperature and subsequently blocked for 2 hours in 10% donkey serum prepared in 0.5% Triton X-100. Samples were then incubated with an anti-ITGA7 polyclonal antibody (ThermoFisher, BS-1816R; 1:100) or anti-PAI-1 antibody (No. ab222754, Abcam, 1:250) for 24 hours at 4°C. After primary antibody incubation, tissues were washed and incubated with Alexa Fluor 594-conjugated goat anti-rabbit IgG (ThermoFisher, A-11012; 1:100), for 1 hour at room temperature in the dark. Lipid droplets were labeled with BODIPY (ThermoFisher; 1:100), and nuclei were counterstained with DAPI (ThermoFisher; 1:100) for an additional hour. Imaging was performed on a Leica TCS SP8 confocal microscope (Leica Microsystems, Deerfield, IL) using a 20× dry objective (NA 0.75). Z-stack images were collected at 1-μm intervals at 1024 × 1024 resolution with a line averaging of 6 to ensure consistent visualization of tissue architecture. Imaging parameters, including laser settings, pinhole size, and detector gain, were held constant across all samples. Z-stacks were reconstructed and analyzed using Leica LAS X software (version 3.5.2). ITGA7-positive area (μm²) was quantified using the 3D analysis module. PAI-1 immunofluorescence intensity was quantified by measuring the mean fluorescence intensity within standardized fields of view.

Flow Cytometry and Sorting

Pelleted adipose derived stromal cells were resuspended in 0.1 mL 2% Fetal Bovine Serum (FBS) in PBS and incubated with 10 uL (per million cells/g fat) of integrin alpha 7 (ITGA7) monoclonal antibody (MA5–23608, Invitrogen) conjugated with R-phycoerythrin and/or 2 uL (per million cells/g fat) of CD9 monoclonal antibody conjugated with FITC (124807, BioLegend) at 4°C for 15 minutes protected from light. After incubation cells were washed and centrifuged 600g first with 0.5 mL, following 0.250 mL, finally resuspending with 0.5 mL of 2% FBS/PBS. Appropriate negative fluorescence controls were made for each antibody and a no fluorescence control of the ASC. ASC labeled with fluorescent antibodies were sorted into two different groups using a BD FACSAria-II cell sorter (BD Biosciences), ITGA7 ^high and ITGA7 ^low subpopulations into 1 mL 100% FBS. Analysis was performed using the FlowJo software version 10. Each subpopulation type was propagated in a flask 25 cm² in 5 mL of complete media (Dulbecco’s Modified Eagle Medium/F12 10%FBS, 1% Antibiotic-antimycotic) in a 37°C incubator with 5% CO₂· until 100% confluency was reached. Fluorescence ITGA7 sorting was repeated three times in obese animals as well as flow cytometry analysis, total of n=3 for each. The ASC visualized in the Amnis ImageStreamX Flow Cytometer were fixed in 4% paraformaldehyde in PBS in 4°C overnight, washed in PBS and resuspended in 0.1 mL PBS.

ASC differentiation

ITGA7^high and ITGA7^low ASCs at passage 1 were seeded in 24-well tissue culture plates and maintained in complete media (DMEM F12 supplemented with 10% FBS and 1% antibiotic–antimycotic) until reaching ~90% confluency, after which they were induced to differentiate into either white or beige adipocytes. Beige adipocyte differentiation: Cells were cultured in beige induction medium for the first 4 days consisting of DMEM complete medium supplemented with 5 μg/mL insulin, 10 μM forskolin, 2 μg/mL dexamethasone, 125 μM indomethacin, 0.5 μM rosiglitazone, and 1 nM triiodothyronine (T3). From day 4 onward, cells were maintained in DMEM complete medium containing 5 μg/mL insulin, 10 μM forskolin, 1 μM rosiglitazone, and 1 nM T3. White adipocyte differentiation: Cells were cultured in white adipogenic induction medium composed of complete DMEM supplemented with 5 μg/mL insulin, 10 μM forskolin, 2 μg/mL dexamethasone, and 125 μM indomethacin. On day 4, the medium was replaced with maintenance medium containing complete DMEM supplemented with 5 μg/mL insulin. For both differentiation protocols, media were changed every other day. On day 7, cells were collected for qPCR analysis to assess adipogenic differentiation.

Glucose Uptake and Lipolysis Assay

Glucose Uptake:

Glucose uptake was assessed in mature adipocytes differentiated from ITGA7^high ASC and ITGA7^low ASC in 96-well plates. Glucose uptake was quantified using the Glucose Uptake-Glo^™ assay (Promega) according to the manufacturer’s instructions. Briefly, culture media were replaced with DMEM containing RealTime-Glo reagent and 50 μM cytochalasin B and incubated for 30 min at 37°C. Insulin (10 μM) was then added for 1 h to stimulate glucose uptake, followed by the addition of 1 mM 2-deoxyglucose. After a 10-min incubation, the 2-deoxyglucose-6-phosphate detection reagent was added, and luminescence was recorded using a Synergy 2 microplate reader (BioTek).

Lipolysis Assay:

Differentiated adipocytes were incubated with 100 nM isoproterenol, and glycerol release was measured using the MAK211 Lipolysis Colorimetric Assay Kit (Sigma-Aldrich). Absorbance at 570 nm was recorded using a Synergy 2 microplate reader, and glycerol concentration was used as a readout of lipolytic activity.

Inflammation Assessment, LPS Stimulation and PAI-1-silencing

Sorted ITGA7^high and ITGA7^low ASCs at passage 1 were seeded in 24-well plates and cultured in complete media until approximately 90% confluency. Basal inflammatory marker expression was first evaluated by culturing both cell populations in complete media without stimulation, followed by qPCR analysis. To assess inflammatory responsiveness, ITGA7^high and ITGA7^low ASCs were stimulated with 100 ng/mL lipopolysaccharide (LPS; eBioscience, 00-4976-93) for 24 hours at 37°C and 5% CO₂. After treatment, cells were collected for qPCR analysis of inflammatory gene expression. PAI-1 silencing experiments were performed using Lipofectamine RNAiMAX (Invitrogen, 13778075) and siRNA targeting mouse PAI-1 (Santa Cruz Biotechnology, sc-36180) following the manufacturer’s protocol. Briefly, Lipofectamine and PAI-1 siRNA were mixed in Opti-MEM and incubated for 20 minutes at room temperature to allow complex formation. Complexes were then diluted in complete medium and added to ASCs for 48 hours to achieve gene silencing. In designated groups, LPS stimulation (100 ng/mL, 24 hours) was performed following PAI-1 knockdown. Cells were harvested after treatment for qPCR analysis of inflammatory markers.

Macrophage stimulation

RAW 264.7 macrophages were cultured in complete RPMI medium supplemented with 10% FBS and 1× antibiotic–antimycotic. Cells were seeded in 24-well plates and allowed to reach approximately 90% confluency. At that point, the medium was replaced with conditioned media consisting of a 1:1 mixture of complete RPMI and the secretome collected from sorted obese ITGA7^high or ITGA7^low ASC. Macrophages were incubated for 6 hours at 37°C and 5% CO₂, after which cells were harvested for RNA extraction and qPCR analysis of inflammatory gene expression.

RNA isolation and quantitative RT-PCR

Total RNA was isolated from ASC (ITGA7^high and ITGA7^low subpopulations) using the QIAGEN RNeasy Mini Kit (Valencia, CA). RNA concentration was quantified using a Take3 microvolume plate reader (BioTek, Winooski, VT), and 100 ng of RNA was reverse-transcribed into cDNA using the iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA). Gene expression analysis was performed by real-time quantitative PCR (qPCR) using the CFX96 Touch Real-Time PCR Detection System and iTaq Universal SYBR Green Supermix (Bio-Rad). Target genes included markers of beige adipogenesis (UCP1, ADIPONECTIN), white adipogenesis (PPARG, ADIPONECTIN), inflammation (IL6, TNFα, CCL2), and fibrosis (CD9, PAI-1, COL1A1, COL3A1). Primer sequences used are in Table S1 in Supplemental Material. Fold-change expression was calculated using the 2^−ΔΔCt method, normalizing to ITGA7^low controls and using 18S rRNA as the internal reference gene.

ELISA for mouse PAI-1

Culture media from passage 0 ITGA7^high and ITGA7^low ASC (propagated in T25 flasks) was collected prior to cell splitting and stored at −80°C until analysis. Secreted mouse PAI-1 levels were quantified using the DuoSet ELISA kit (R&D Systems, DY3828–05) following the manufacturer’s instructions. Briefly, 96-well plates were coated with capture antibody, sealed, and incubated overnight at room temperature. Plates were washed three times and blocked for 1 hour at room temperature. After an additional three washes, samples and standards were added in duplicate, sealed, and incubated for 2 hours at room temperature. Wells were washed and incubated with the detection antibody for 2 hours, followed by streptavidin-horseradish peroxidase for 20 minutes while protected from light. The reaction was developed and stopped according to kit instructions, and optical density was measured immediately at 450 nm with wavelength correction at 540 nm using a microplate reader (BioTek Synergy 2 or specify model). A 7-point standard curve was run in duplicate, and all samples were analyzed in triplicate.

Obese human adipose stromal cell studies.

Primary human adipose stromal cells (hASCs) isolated from an obese donor (BMI 31; Lonza, PT-5006) were cultured in DMEM/F12 supplemented with 10% fetal bovine serum and 1% antibiotic-antimycotic at 37°C and 5% CO₂. Cells were treated for 48 h with PAI-1 (SERPINE1) siRNA using Lipofectamine RNAiMAX according to the manufacturer’s instructions or with the PAI-1 inhibitor tiplaxtinin (PAI-039). PAI-1 gene expression was quantified by RT-qPCR using 18S as the housekeeping gene. In parallel, cells were fixed and immunostained overnight with PAI-1 antibody (Abcam, ab222754) followed by Alexa Fluor-conjugated secondary antibody and DAPI nuclear staining. Confocal images were acquired using a Leica TCS SP8 microscope under identical imaging settings for all groups, and PAI-1 immunofluorescence was quantified using Leica LAS X software. RT-qPCR analyses were performed using three independent biological replicates, each analyzed in triplicate, while immunofluorescence quantification was performed from a minimum of 10 randomly selected fields per biological replicate.

Nanoparticle design, loading and antibody conjugation.

The LCMSN were implemented following our previous reports^17,18. Dendritic MSN with a net positive charge were made via the biphase reaction with a delayed addition co-condensation of aminopropyltriethoxysilane as primary amine. The siRNA loading (PAI-1/SERPINE1, sc-36180, Santa Cruz) was carried out by incubating siRNA with positive MSN in water (2% wt) for 30 min, followed by isolation of MSN-siRNA by centrifugation (15 min, 21 Krcf). Upon re-suspension in water, liposomes were added under brief sonication (5~10 seconds). The resulting system was centrifuged and washed once with buffer. Liposomes were made via the film hydration method in presence of neutravidin. Liposomes (5 mg) were made by mixing DPPC:DMPG:cholesterol:DSPE-PEG-NHS (all from Avanti) at a molar ratio of (59: 23: 10: 8) and the chloroform was evaporated under vacuum for 10 minutes. In parallel, neutravidin (0.2 mg) was solubilized in PBS (1 mL) and the solution is added to the lipid film that was sonicated for 20 minutes to create liposomes with conjugated neutravidin upon reaction with the NHS terminated DSPE-PEG. Upon lipid fusion, ITGA7-biotin (10 μM, LifeSpan BioSciences, Inc) was added to the LCMSN-neutravidin suspension in PBS (60 μg Ab for each mg LCMSN) at room temperature for 1 h before washing once by PBS.

siRNA loading assessment and release experiment.

To confirm the total loading of siRNA we have utilized a cy3-negative control siRNA surrogate (ThermoFisher). The fluorescence of the siRNA-loaded MSN suspension was compared to the one from supernatant. All siRNA added was loaded (100% equivalent to 20 μg per mg MSN). For the release, the particles lipid-free or coated siRNA-MSN (250 μg) were suspended in PBS at 37 °C and at selected time points (0.5h-24h) the particles were isolated by centrifugation and the supernatant’s fluorescence due to cy3-siRNA release was recorded and compared with the whole suspension at each timepoint.

In vitro treatment of ITGA7^high ASC with LCMSN loaded with PAI-1 siRNA

ITGA7^high ASCs were cultured under standard conditions and treated with LCMSNs loaded with PAI-1 siRNA at a nanoparticle concentration of 50 μg/μL. Cells were incubated with nanoparticles for 6 h, after which the treatment medium was removed and replaced with fresh culture medium. Cells were then collected at 6 and 48h after which the treatment medium was removed and replaced with fresh culture medium. Cells were then collected at 48, 72, and 96 h after the start of treatment to assess the durability of PAI-1 silencing. Total RNA was isolated from each time point, and PAI-1 gene expression was quantified by RT-qPCR. Gene expression was normalized to the housekeeping gene and analyzed relative to the 6 h time point.

In vitro Cytotoxicity assessment of LCMSN

ITGA7^high ASCs were cultured under standard conditions and treated for 48 hours with 50 μg/mL of different control groups of ITGA7-LCMSNs loaded with siRNA Pai1, siRNA scramble, LCMSNs siRNA Pai1, siRNA scramble, ITGA7-LCMSN and LCMSN (empty core). Lactate dehydrogenase (LDH) was measured according to manufacturer’s protocol (Invitrogen, C20300) to assess cytotoxicity of the LCMSNs to the cells. Briefly, spontaneous LDH release was measured by transferring 50 μL of each sample medium to a 96-well flat bottom plate in triplicates. 10 μL of 10X Lysis buffer was added to each well to measure the maximum LDH release. Plate was incubated for 45 minutes at room temperature. 50 μL of the reaction mixture was transferred to each sample well and incubated at room temperature for 30 minutes protected from light. Reaction was stopped by adding 50 uL of stop solution to each sample well and absorbance was measured immediately at 490 nm and 680 nm wavelength. A positive control was used to assess reproducibility of assay. % Cytotoxicity was calculated according to kit.

In vitro Immunofluorescence Staining and Confocal Microscopy.

Adipose stromal cells (ASCs) isolated from ob/ob mice were treated with fluorescently labeled LCMSN-siPAI-1, fixed with 4% paraformaldehyde, permeabilized with 0.2% Triton X-100, and blocked with 10% goat serum. Cells were incubated overnight with anti-PAI-1 primary antibody (Abcam, ab222754; 1:250), followed by Alexa Fluor 594-conjugated secondary antibody (Invitrogen, A-11037; 1:500) and DAPI for nuclear staining. Images were acquired using a Leica TCS SP8 confocal microscope with identical acquisition settings for all groups. PAI-1 fluorescence intensity was quantified using Leica LAS X software by measuring background-corrected fluorescence within cell-containing regions of interest and normalized to the number of DAPI-positive nuclei. A minimum of 10 randomly selected fields per biological replicate were analyzed.

In vivo Biodistribution of ITGA7-LCMSN

Biodistribution of ITGA7-LCMSNs was evaluated in male ob/ob mice (16 weeks old; Jackson Laboratory). Mice received two doses (one at the beginning and one after 24 hours of first injection) with subcutaneous injection near the inguinal white adipose tissue (iWAT). Treatment groups received 200 μg of ITGA7-LCMSNs loaded with PAI-1 siRNA and labeled with DyLight 633, suspended in PBS. Control animals received 200 μL of PBS. At the study endpoint, 72 hours, whole-body and organ-level nanoparticle distribution was assessed using KINO, a non-invasive optical imaging system (Kino, Spectral Instruments Imaging). Following euthanasia, major organs, including iWAT, visceral WAT (vWAT), liver, heart, kidney, lung, and spleen, were imaged to quantify ITGA7-LCMSN retention and fluorescence intensity. An additional study was performed to assess the temporal biodistribution and retention of ITGA7-LCMSNs, in which, fluorescently labeled nanoparticles were administered subcutaneously to ob/ob mice at 200 μg, and tissue distribution was evaluated at 4 h and 4 days post-injection. Whole-body and major organs imaging were measured and evaluated. All animal procedures were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) at The University of Texas at San Antonio.

In Vivo Efficacy of ITGA7-LCMSNs

To evaluate the functional efficacy of ITGA7-LCMSN, mice were treated as described above and tissues were collected 48 hours after the final injection. Inguinal white adipose tissue (iWAT) was dissected, and ASC were isolated for fluorescence-activated cell sorting to obtain ITGA7^high ASCs. Total RNA was extracted from both sorted ITGA7^high ASCs and whole iWAT tissue. Gene expression analysis was performed by quantitative PCR to assess PAI-1 silencing. Expression levels were normalized to housekeeping genes using the 2^−ΔΔCt method.

In vivo Fasting Glucose Effect of Extended ITGA7-LCMSNs treatment

To evaluate the metabolic effect of prolonged PAI-1 siRNA treatment, ITGA7-targeted and untargeted LCMSN loaded with PAI-1 siRNA were administered to male ob/ob mice (16 weeks old; Jackson Laboratory) twice per week. Each dose consisted of 100 μg of nanoparticles delivered subcutaneously and 100 μg delivered intraperitoneally (200 μg total per administration). Control mice received 200 μL of PBS by the same routes. Fasting glucose (5-hour fast) was measured at baseline and at the two- and four-week timepoints. Adipose tissue was collected to perform RT-PCR (as described above) for gene expression analysis. For each mouse, nanoparticles were administered into distinct regions of the adipose depot. At endpoint, tissue from different sites were collected separately for gene expression analysis. Injection-site measurements were analyzed to assess regional treatment responses within each animal.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.