Comparative proteomic analysis of the ECM composition of the human omentum and mesentery, the main sites of ovarian cancer metastasis

Summary

Due to its limited symptoms, high-grade serous ovarian cancer (HGSOC) has frequently metastasized extensively throughout the peritoneal cavity prior to its diagnosis, resulting in an overall five-year survival rate of less than 50%. The omentum and mesentery are two of the most common metastatic sites in HGSOC. However, the mechanisms underlying HGSOC metastatic tropism remain unknown. The extracellular matrix (ECM) is a meshwork of proteins that provides biochemical and mechanical signals to cells and has been shown to drive the dissemination of several cancer types to preferential distant sites. Here, combining histological assessment and proteomics, we defined the architectural features and composition of the ECM of paired omentum and mesentery samples from disease-free adult women. We found that over 90% of the proteins detected were shared between the omentum and mesentery, but also identified tissue-specific ECM proteins, including collagen XII and tenascin-C.

Graphical abstract

Highlights

• •Histological and proteomic analyses of the ECM of paired human omentum and mesentery samples
• •Higher collagen content in the mesothelial layer than in the adipose region in the omentum
• •Identification of ECM protein signatures distinguishing the omentum and mesentery

Cancer; Proteomics

Introduction

Due to its limited symptoms, high-grade serous ovarian cancer (HGSOC) has frequently metastasized extensively throughout the peritoneal cavity prior to its diagnosis, resulting in an overall five-year survival rate of less than 50%.^1,2,3,4 The greater omentum and the small bowel mesentery are two of the most common metastatic sites in advanced HGSOC.⁵ Nearly 80% of patients have a metastasis on the omentum, which is commonly removed by omentectomy during a debulking.⁶ In contrast, approximately 30% of patients present with tumors on the small bowel mesentery,⁷ with disease in this site primarily treated by chemotherapy. The difference in treatment approach stems from their unique anatomical roles. Both are components of the visceral peritoneum forming double-layered tissues that cover and provide structure for the abdominal organs, and it has been suggested that the omentum develops when a “bubble”-like structure forms in the upper mesentery.⁸ At maturity, the greater omentum is suspended loosely between the intestines and the peritoneal wall and has non-essential roles in fat storage and inflammation.⁹ In contrast, the mesentery surrounds the intestines and is heavily vascularized to provide blood supply for these critical organs.¹⁰ Mesenteric involvement can prevent complete cytoreduction,^11,12 which is associated with a worse outcome.¹³ Therefore, identifying the mechanisms that support metastasis to the mesentery could provide treatments that would improve HGSOC patient prognosis.

Due to its more common involvement and frequent removal during surgical debulking, the greater omentum has been extensively studied to identify factors that support HGSOC metastasis.^14,15,16 Studies have demonstrated a loss of adipocytes as tumors progress in the omentum and identified mechanisms by which tumor cells utilize lipids stored in adipocytes for energy to support proliferation.¹⁷ Another common hypothesis for the involvement of the omentum in HGSOC is the presence of “milky spots,”¹⁸ clusters of leukocytes and macrophages that may secrete factors to attract tumor cells¹⁹ or alter the tissue microenvironment to support tumor cell adhesion.²⁰ However, neither of these observations would explain why the omentum is more commonly impacted in HGSOC relative to the mesentery, as both tissues have adipose and milky spots. The peritoneal cavity is subject to fluid flow resulting from a complex mix of factors such as breathing, digestion, and physical movement²¹; therefore, it is possible that the location or accessibility of these tissues influences which sites are impacted.

The extracellular matrix (ECM) is a complex and dynamic meshwork of proteins that provides biochemical and mechanical signals governing cellular phenotypes.^22,23 The ECM plays key roles in promoting tumor metastasis,²⁴ including in HGSOC.^16,25,26 It can also guide metastatic tropism, i.e., the preferential dissemination and/or outgrowth of secondary tumors in specific organs, as originally proposed by Stephen Paget in 1889 when he formulated the “seed and soil” theory.^27,28,29,30 Proteomics has been broadly applied to characterize changes in the ECM during different stages of ovarian cancer progression, including tumor initiation and metastatic dissemination to the omentum.^{31,32,33,34,35,36,37,38,39,40} In addition, in-vitro studies have identified roles for ECM proteins in tumor cell adhesion during their spread to peritoneal tissues.^26,41,42 However, much less is known about the ECM of the mesentery compared to the omentum.

Here, we hypothesized that there might be differences in ECM composition, abundance, or organization between the greater omentum and small bowel mesentery that could explain the preferential dissemination of HGSOC to the omentum. To examine this hypothesis, we combined histological analysis and ECM-focused proteomics to characterize the ECMs of donor-matched omentum and mesentery samples. Histological analysis revealed a higher fibrillar collagen content in the mesothelial layer of the omentum as compared to the mesentery. Using a stringent quantitative proteomic pipeline, we found that the ECM composition – or matrisome – of the omentum and mesentery is highly similar, with >90% of the proteins shared between the two tissues and no significant quantitative differences among the proteins represented in both matrisomes. However, our analysis also revealed small subsets of tissue-specific ECM components, which we further validated by histological analyses when reliable antibodies were available. Altogether, our data show that the human omentum and mesentery have a highly similar ECM profiles and that there are thus likely additional factors explaining the preferential dissemination of HGSOC metastasis to the omentum. Last, our dataset, shared publicly with ample metadata to facilitate its reuse, can also serve as a reference ECM atlas of the human omentum and mesentery for future studies.

Results and discussion

Histological characterization of the extracellular matrix of human mesentery and omentum samples

The extracellular matrix is found in distinct locations within the mesentery and omentum. An interstitial-type ECM²² is found on the basal side of the mesothelial layer, referred to as the tissue edge. In contrast, a basement-membrane-type ECM^22,43 is found surrounding the vascular endothelium and around adipocytes (Figure 1A). One challenge in comparing the ECM of two tissues is the potential for variations within the tissues due to the underlying health of the donor, as conditions such as obesity result in the inflammation and fibrosis of adipose tissues.⁴⁴ Therefore, we opted to obtain matched samples of the omentum and mesentery. However, the omentum and mesentery are not commonly biopsied from patients with benign conditions, and the need for non-fixed samples for proteomic analysis prevented us from using archived samples. Therefore, we collected tissue from female decedents. To assess the overall distribution of the ECM in human mesentery and omentum, we stained tissue sections with picrosirius red, a dye that binds specifically to fibrillar collagens (primarily collagens I and III) (Figure 1B). We found that the overall thickness of the collagen band between the adipose and the mesothelial layer was similar between the two tissues (Figure 1C, left panel). While the ECM around the adipocytes is often described as a basement-membrane type ECM (lacking fibrillar collagen and containing the network-forming collagen IV), we noted in these samples the presence of fibrillar collagen between adipocytes (Figure S1). We next quantified the mean signal intensity corresponding to fibrillar collagen staining and observed a significantly higher collagen intensity for the region along the tissue edge relative to the adipose-rich interior region, but only for the omentum (Figure 1C, right panel).

Figure 1.

Histological characterization of the ECM of the human omentum and mesentery

(A) Annotated hematoxylin and eosin staining of a histological section of the human omentum (left) and mesentery (right). Scale bars: 200 μm.

(B) Picrosirius-red-stained images of the omentum and mesentery were analyzed for the thickness of the collagen band underlying the mesothelial layer (red line) and the mean signal intensity of the collagen located within the edge and first layer of adipocytes (green dotted line). Scale bars: 50 μm.

(C) Bar graphs represent the thickness of the tissue edge (left) and the mean signal intensity of collagens in the edge (circle) or adipocyte-rich interior of the tissue (triangle). Each dot represents the data for a single donor (N = 7); in addition, the mean value (black bar) and standard error of the mean are depicted. t test; ∗: p < 0.05. See also Figure S1.

Proteomic pipeline to characterize the extracellular matrix composition of human mesentery and omentum samples

To characterize the molecular composition of the human mesenteric and omental ECMs, we obtained paired healthy omentum and mesentery samples from four donors (Figure 2A) and submitted these samples to a five-step decellularization process, extracting sequentially proteins of high to intermediate solubility, and resulting in the enrichment of highly insoluble ECM proteins.⁴⁵ This approach is necessary to specifically capture those proteins that are, in fact, incorporated into the ECM scaffold. We evaluated the efficiency of the decellularization using western blotting and found that intracellular proteins were completely (GAPDH) or at least partially (actin) extracted in the intermediate fractions generated through the decellularization protocol. Importantly, we observed that the ECM protein marker, collagen I, was not extracted during the procedure and was retained in the ECM-enriched fractions for both mesentery and omentum samples (Figures 2B and 2C).

Figure 2.

Experimental workflow to characterize the ECM of the human omentum and mesentery

(A) Schematic representation of the experimental design of this study, created using BioRender.

(B and C) Western blots show the sequential depletion of intracellular components (actin, GAPDH) during the 4-step decellularization process of human mesentery (B) and omentum (C) samples (fractions 1 to 4) and the resulting enrichment of ECM proteins such as collagen I.

(D–E) Donut charts represent the proportion of the total precursor ion intensity (left) and total spectral count (right) of the core matrisome (dark blue), matrisome-associated (orange), and non-matrisome (gray) components in a representative set of matched human mesentery (D) and omentum (E) ECM-enriched protein samples. See also Tables S2A and S2B.

Next, we submitted these ECM-enriched samples to proteolytic digestion to generate peptides for proteomic analysis using liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS). We analyzed the LC-MS/MS output with stringent criteria, accepting peptide and protein identifications if they could be established with a false-discovery rate (FDR) of less than 1.0% and requiring the detection of at least 2 unique peptides to further accept protein identification (Table S2A). We first sought to assess the efficiency of the ECM enrichment strategy. To do so, we determined the proportion of the total precursor ion intensities and spectral counts attributable to matrisome vs. non-matrisome proteins (Table S2B). These two metrics serve as proxies for protein abundance. Our results confirmed that we achieved a significant enrichment of ECM components for both mesentery and omentum samples, as the majority of the total precursor ion intensities (>99%; Figures 2D and 2E, left panels) and spectral count (>95%; Figures 2D and 2E, right panels), were attributable to core matrisome proteins, i.e., the proteins forming the structure of the ECM scaffold (Table S2B).

Proteomic profiling of the extracellular matrix of human mesentery samples

We next compared the proteomic profile of each of the four mesentery samples and found a substantial overlap with 22 proteins detected in all four samples, and another 27 detected in at least two independent samples (Figures 3A and S2A; Table S2C). We defined the ECM profile of the human mesentery as the ensemble of 49 proteins detected in at least two mesentery samples. The ECM profile of the healthy human mesentery is composed of 42 core matrisome proteins, including 21 ECM glycoproteins, 16 collagens, and five proteoglycans (Figure 3B, Tables 1 and S2C). Our analysis identified distinct protein chains of functional ECM multimers. For example, we detected the α4, α5, β1, β2, and γ1 laminin chains, encoded by the LAMA4, LAMA5, LAMB1, LAMB2, and LAMC1 genes, respectively. Laminin chains are known to assemble selectively to form functional trimers.⁴⁶ Our results point to the presence of four functional assemblies, the α4β1γ1, α4β2γ1, α5β1γ1, and α5β2γ1 trimers, in the basement membrane compartment of the mesentery.

Figure 3.

Proteomic profiling of the ECM of human mesentery and omentum samples

(A) Venn diagram represents the number of matrisome proteins detected in each of the four mesentery samples profiled.

(B) Bar graph shows the distribution of matrisome proteins identified in at least two mesentery samples across the different matrisome categories. This ensemble of 49 proteins constitutes the ECM profile of the human mesentery; see also Tables 1 and S2C.

(C) Venn diagram represents the number of matrisome proteins detected in the omentum samples for each donor.

(D) Bar graph shows the distribution of matrisome proteins identified in at least two omentum samples across the different matrisome categories. This ensemble of 48 proteins constitutes the ECM profile of the human omentum; see also Tables 2 and S2D.

(E) Bar graph shows the distribution of proteins of the ECM profiles of the human mesentery (left) and omentum (right) across the different categories of matrisome components.

Table 1.

ECM profile of the human mesentery

Core matrisome			Matrisome-associated
ECM Glycoproteins	Collagens	Proteoglycans	ECM-affiliated	ECM Regulators	Secreted Factors
EFEMP1	COL1A1	ASPN	ANXA1	AMBP	ANGPTL2
ELN	COL1A2	BGN	LGALS1	LOXL1	S100A10
EMILIN1	COL2A1	DCN		TGM2
FBLN2	COL3A1	HSPG2
FBN1	COL4A1	LUM
FGA	COL4A2
FGB	COL4A5
FGG	COL5A1
FN1	COL5A2
IGFBP7	COL5A3
LAMA4	COL6A1
LAMA5	COL6A2
LAMB1	COL6A3
LAMB2	COL8A1
LAMC1	COL12A1
NID1	COL14A1
NID2
POSTN
TNC
VTN
VWF

The ECM profile of the human mesentery is composed of 49 proteins identified in at least two of the four mesentery samples profiled, related to Figures 3A and 3B and Table S2C.

Interestingly, we identified two collagen IV chains, α1 (COL4A1) and α2 (COL4A2), that assemble to form the functional ([α1]₂;α2) collagen IV trimer found in basement membranes. We also detected the α5 chain of collagen IV (COL4A5), which has been reported to assemble with either the α3 and α4 chains of collagen IV to form the function (α3;α4;α5) trimer or with the two α6 chains of collagen IV to form the ([α5]₂;α6).^43,47 However, neither the α3, α4, nor α6 chains were detected here. It is unlikely that this is due to a lack of detection due to insufficient abundance, since these proteins assemble with well-characterized 1:1:1 or 2:1 stoichiometries. It could be due to a limit in the sensitivity of our analytical pipeline that is unable to distinguish confidently highly homologous proteins with the stringent criteria imposed, or, last, it is possible that the α5 chain of collagen IV engages with other partners not identified before.

In addition to these basement membrane components, we identified collagen components of the interstitial ECM, such as the three collagen VI chains, α1 (COL6A1), α2 (COL6A2), and α3 (COL6A3), known to assemble to form the functional triple helical collagen VI.^47,48 Collagen VI has been previously identified in omental and mesenteric metastases and correlates with a poor prognosis at the transcriptional level.⁴⁹ While prior studies have focused on the impact of collagen VI produced by tumor cells,^49,50,51 these findings support that this protein may be involved in early metastasis and become dysregulated in the tumor microenvironment. We also identified the two collagen I chains, α1 (COL1A1) and α2 (COL1A2), that must assemble to form a functional collagen I trimer ([α1]₂;α2)⁴⁷. Other core matrisome proteins of the mesentery ECM profile also include fibrillar glycoproteins such as fibronectin (FN1), fibrillin 1 (FBN1), proteins forming elastic fibers such as elastin (ELN) and the elastic microfibril interface located protein-1 (EMILIN1), the matricellular proteins^52,53 fibulins 2 and 3 (FBLN2 and EFEMP1), periostin (POSTN) and tenascin-C (TNC), and ECM proteins associated with hemostasis⁵⁴ such as the three fibrinogen chains (FGA, FGB, FGG) and the von Willebrand factor (VWF).

The ECM profile of the human mesentery also included seven matrisome-associated proteins, including two ECM-affiliated proteins, three ECM regulators (including two ECM crosslinking enzymes, LOXL1 and TGM2), and two secreted factors (Figure 3B, Tables 1 and S2C). Of note, this study provides the first proteomic characterization of the ECM of the human mesentery of disease-free individuals and can serve as a baseline for future studies focused on this tissue.

Proteomic profiling of the extracellular matrix of human omental samples

We next compared the proteomic profile of each of the four omentum samples and found a very large overlap with 32 proteins detected in all four samples, and another 16 detected in at least two independent samples (Figures 3C and S2C; Table S2D). Therefore, we define the ECM profile of the human omentum as the ensemble of 48 proteins detected in at least two omentum samples. This profile is composed of 39 core matrisome proteins, including 20 ECM glycoproteins, 14 collagens, and five proteoglycans (Figure 3D; Tables 2 and S2D), a distribution similar to what we observed for the mesentery (Figure 3E). Our analysis identified the same laminin and collagen chains and trimer as those detected in the mesentery. The omentum ECM profile also included nine matrisome-associated proteins, including four ECM-affiliated proteins, three ECM regulators (including two ECM crosslinking enzymes, LOXL1 and TGM2, also found in the mesentery matrisome), and two secreted factors (Figure 3D; Tables 2 and S2D).

Table 2.

ECM profile of the human omentum matrisome

Core matrisome			Matrisome-associated
ECM Glycoproteins	Collagens	Proteoglycans	ECM-affiliated	ECM Regulators	Secreted Factors
EFEMP1	COL1A1	ASPN	ANXA1	AMBP	ANGPTL2
ELN	COL1A2	BGN	ANXA2	LOXL1	S100A10
EMILIN1	COL2A1	DCN	ANXA8	TGM2
EMILIN2	COL3A1	HSPG2	LGALS1
FBLN2	COL4A1	LUM
FBLN5	COL4A2
FBN1	COL4A5
FGA	COL5A1
FGB	COL5A2
FGG	COL5A3
FN1	COL6A1
LAMA4	COL6A2
LAMA5	COL6A3
LAMB1	COL14A1
LAMB2
LAMC1
NID1
NID2
POSTN
VTN

The ECM profile of the human omentum is composed of 48 matrisome proteins identified in at least two of the four omentum samples profiled, related to Figures 3C and 3D and Table S2D.

While there is no prior study on the ECM composition of the human mesentery, we and others have reported the composition of the ECM of the human omentum, and these studies have typically identified a larger number of proteins.^31,33,36 This can be explained by the fact that we focused here on the identification of proteins of the insoluble ECM scaffold after sample decellularization, as opposed to previous studies that have performed proteomic analyses on the bulk omental samples or urea-soluble omental samples. We have also applied stringent criteria to accept protein identification, including strict false-discovery rate thresholds and strict cohort-level cut-offs (see STAR Methods). Last, since we focused on the most insoluble ECM protein fraction, fibrillar collagens that contribute hundreds of spectra, tend to saturate the MS signal, decreasing our ability to identify proteins present in lower abundance. To alleviate this, we attempted to fractionate samples using basic reverse-phase liquid chromatography, a method commonly employed to achieve greater depth in proteomic studies; however, because of the dominance of collagen-derived peptides, this approach fell short. To increase protein identification and capture ECM proteins of higher insolubility, future studies could adopt alternative protocols using different ECM enrichment protocols, profile protein fractions of intermediate solubilities to capture more soluble ECM proteins, or use different methods to digest proteins into peptides.^{55,56,57,58,59}

Comparative analysis of the human omentum and mesentery extracellular matrix profiles

We first compared the ECM profiles of paired omentum and mesentery samples and found a very large overlap at the individual donor level (Figure S2C). Moreover, the comparative analysis of the matrisomes of the human mesentery and omentum reveals an overall very similar ECM composition, both in terms of the representation of the different categories of matrisome components (Figure 3E) and in overall protein abundances (Figure 4A, Tables S2B and S2E), with fibrillar collagens (COL1A1, COL1A2, COL3A1) contributing to the majority of the protein signal and dominating ECM composition by two orders of magnitude over the rest of the matrisome (Figure 4A and Table S2E). This result, in combination with the PSR staining demonstrating a thicker collagen band surrounding the adipose with a thin network between individual adipocytes, suggests that fibrillar collagens play an essential role in the overall organization of these tissues. Collagen I is further increased in tissues with metastases,³² where it remains the dominant component.^31,33

Figure 4.

Comparison of the ECM profiles of the human omentum and mesentery

(A) Heatmap of normalized total precursor ion intensities of the 44 proteins detected in both human mesentery and omentum ECM profiles.

(B) Proportional Venn diagram represents the overlap between the mesentery and omentum ECM profiles.

(C) Volcano plot represents the fold-change average protein abundance inferred from total precursor ion intensity between the omentum and mesentery; see also STAR Methods and Table S2E. Dotted lines along the x axis indicate a 25% lower (left quadrant) or higher (right quadrant) protein abundance in the omentum as compared to the mesentery. The dotted line along the y axis depicts the statistical significance cut-off set to p < 0.05 determined using a Student’s t test.

The two tissues share 44 matrisome proteins (Figure 4B) and notably share the same proteoglycanome composed of asporin (ASPN), biglycan (BGN), decorin (DCN), lumican (LUM), and the basement-membrane-associated proteoglycan perlecan (HSPG2). The quantitative analysis of the abundance of these 44 proteins, using normalized total precursor ion intensities, did not reveal any statistically significant differences (Figure 4C and Table S2E).

However, our analysis led to the identification of five matrisome proteins in the ECM profile of the mesentery but not of the omentum (Figure 4B, Tables S2C and S2D). These include the network-forming collagen VIII (COL8A1), the fibril-associated collagen with interrupted triple helix collagen XII (COL12A1), the insulin-like growth factor binding protein (IGFBP7), the matricellular protein tenascin-C (TNC),⁶⁰ and the von Willebrand factor (VWF). Conversely, we identified four matrisome proteins characteristic of the omentum matrisome (Figure 4B). These include two components of the core matrisome, the elastic microfibril interface located protein-2 (EMILIN2), and the matricellular protein fibulin 5 (FBLN5), neither of which has been clearly associated with HGSOC metastasis. We also detected two ECM-affiliated proteins, annexins A2 and A8 (ANXA2, ANXA8), in the omentum but not in the mesentery. Annexin A2 has previously been shown to be expressed in the mesothelial layer of the omentum, and weekly treatments with a blocking antibody against annexin A2 resulted in slower tumor growth in a xenograft model of SKOV3 cells.⁶¹

Collagen XII and tenascin C present tissue-specific distribution patterns

Two of the proteins detected in the mesenteric but not omental ECM profile, collagen XII and tenascin C, have previously been associated with tumor metastasis and metastatic tropism across several cancer types, including breast and pancreatic cancers.^{62,63,64,65,66,67} Somewhat surprisingly, our analysis found them associated with the ECM of the mesentery, the tissue that is less frequently a site of metastatic dissemination in HGSOC. We thus sought to validate the proteomic data using immunohistochemistry. We found that the abundance of collagen XII (Figure 5A) and tenascin C (Figure 5B) supported the proteomic results, with a higher abundance in mesentery samples as compared to omentum samples. Specifically, collagen XII, which is a fibril-associated collagen with interrupted triple helices,⁴⁷ presented a fibrillar pattern between adipocytes of the mesentery characterized by short and distinct structures (Figure 5A, left panel), in agreement with its known association with fibrillar collagens, collagens I and III. This pattern was absent in the omentum samples analyzed. In contrast, the matricellular protein tenascin C presents a more homogeneous pattern surrounding adipocytes, suggesting a possible association with the basement membrane that continuously surrounds these cells. We also observed an abundant staining of the large vessels of the mesentery but not of the omentum in two of the three pairs of samples analyzed (Figure 5B, left panel). Last, it is important to note that very low levels of collagen XII and tenascin C were still detected in the omental samples (Figures S5A and S5B, right panels). This can be attributable to the fact that our proteomic pipeline focused on the insoluble ECM, while immunohistochemistry will detect proteins regardless of their solubility or level of incorporation in the ECM. For tissues with a more robust ECM, this could, for example, be addressed by decellularizing the tissue prior to embedding and sectioning.

Figure 5.

Comparison of the pattern of distribution of collagen XII and tenascin C in human mesentery and omentum samples

(A) Left: Representative images of 5 μm-thick tissue sections stained with an anti-collagen XII antibody confirming the proteomic data showing the presence of collagen XII in the mesenteric ECM and its absence in the omental ECM (scale bars: 100 μm). Right: Dot plot represents the proportion of collagen XII-positive tissue area in three paired omentum and mesentery samples. Each dot represents data from a single donor (N = 3), and the plot depicts the mean and standard deviation.

(B) Left: Representative images of 5 μm-thick tissue sections stained with an anti-tenascin C antibody confirming the proteomic data showing the presence of tenascin C in the mesenteric ECM and its absence in the omental ECM (scale bars: 100 μm). Right: Dot plot represents the proportion of tenascin C-positive tissue area in three paired omentum and mesentery samples. Each dot represents data from a single donor (N = 3), and the plot depicts the mean and standard deviation.

Altogether, our data show that the protein compositions of the ECM of the human omentum and mesentery are highly similar. This is consistent with recent studies that support the idea that the mesentery is a continuous organ and that the omentum develops when a “bubble”-like structure forms in the upper mesentery.⁸ Given the heterogeneity observed in the tissue structures, a limitation of our study is the use of bulk proteomics that may mask spatial variations captured by methods such as immunostaining. Future work should explore the use of matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MSI), a modality successfully applied to the study of the tumor ECM.^68,69,70,71 Improvements with this technique would allow for more detailed characterization of early tumors or metastases, where changes in the tumor site may be too small to register in the bulk analysis. For example, the precursor tumor for HGSOC, serous tubal intraepithelial carcinoma (STIC), is only identifiable by pathological examination.⁴⁰ This would also enable the analysis of a much larger number of samples, since the retrospective analysis of fixed tissue is not possible with the digestion protocol used here.

While the similarities in the matrisome of healthy tissues do not provide an obvious explanation for the predilection for omental metastases over mesentery metastases, future work should aim to test the possible contribution of the ECM proteins presenting a tissue-specific expression to guiding the metastatic tropism of HGSOC. This can be accomplished using tissue engineering strategies that provide control over the composition of the ECM.¹⁵ While prior studies have focused on the impact of proteins that change drastically in abundance during metastasis, analysis of the matrisome abundance supports the use of fibrillar collagens as the base material for in-vitro models of the omentum and mesentery to study the initial metastasis of ovarian cancer cells to these sites.^26,42 For example, we have recently demonstrated that the fibrillar structure of collagen I provides adhesion cues for ovarian cancer cells.²⁶ In addition, it is well known that the omental matrisome changes drastically during tumor progression.^{25,31,33,36,42,72} Therefore, it remains possible that the tissues diverge during this process to provide distinct pre-metastatic and metastatic microenvironments. Future work should thus aim to test the possible functional contributions of the ECM proteins found in this study to be present in differential abundance in the healthy omentum and mesentery to HGSOC metastatic tropism.

Limitations of the study

The limitations of this study include a small sample size composed of only female samples from an older population. While appropriate for considering ovarian cancer metastasis, this bias means these results may not be generalizable to younger patients. Our analysis focused on the most insoluble ECM fraction, which is dominated by fibrillar collagens that may have decreased our ability to identify less abundant proteins. Additionally, we could not perform immunohistochemistry to confirm all the differences identified by proteomics due to the lack of reliable antibody availability. Finally, the use of bulk proteomics means that there remains the potential for spatial variation between the omental and mesentery ECM.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Dr. Alexandra Naba (anaba@uic.edu).

Materials availability

This study did not generate new, unique reagents.

Data and code availability

• •Data availability: Raw mass spectrometry data and accompanying metadata file in the Sample and Data Relationship Format (SDRF)⁷³ have been deposited to the ProteomeXchange Consortium⁷⁴ via the PRIDE partner repository⁷⁵ with the dataset identifier PXD061586.
• •Code availability: This study did not develop new code. Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments

This work was supported in part by the National Cancer Institute (NCI) of the National Institutes of Health (NIH) [R01CA232517 to PKK and AN, R01CA290693 to PKK and AN, and R21CA261642 to AN]. The authors thank the University of Wisconsin Translational Research Initiatives in Pathology Laboratory, supported by the UW Department of Pathology and Laboratory Medicine UWCCC (P30 CA014520) and the Office of the Director of the NIH (S10 OD023526), for use of its facilities and services. Proteomics services were provided by the UIC Research Resources Center Mass Spectrometry Core, which was established in part with a grant from The Searle Funds at the Chicago Community Trust to the Chicago Biomedical Consortium and is supported in part by an NIH shared instrumentation grant [S10OD027016]. INT was the recipient of a UIC Honors College Undergraduate Research Grant. INT and CG were recipients of awards from the Liberal Arts and Sciences Undergraduate Research Initiative (LASURI) at UIC.

The authors would like to thank Dr. Hui Chen and Lasanthi Jayathilaka from the UIC Mass Spectrometry Core facility and Dr. Maria Sverdlov and Dr. Ryan Easton from the UIC Research Histology and Tissue Imaging Core for their technical assistance.

Author contributions

J.C.: investigation, methodology, formal analysis, visualization, and writing. C.G.: investigation, methodology, and formal analysis. D.P.: formal analysis and writing of revised article. N.Y.: investigation, methodology, and formal analysis. I.N.T.: investigation, methodology, and formal analysis. J.S.: investigation and formal analysis. E.G.B.: Resources. P.K.K: conceptualization, methodology, resources, formal analysis, visualization, writing, supervision, funding acquisition, and project administration. A.N.: conceptualization, methodology, resources, formal analysis, visualization, writing, supervision, funding acquisition, and project administration.

Declaration of interests

The funders had no role in the study design, the decision to publish, or the preparation of the article. The Naba laboratory holds a sponsored research agreement with Boehringer-Ingelheim for work not related to the content of this article.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Collagen I	Sigma	Cat#AB765P; RRID: AB9229
GAPDH	Sigma	Cat#MAB374, RRID:AB_2107445
Actin	Dr. Richard O Hynes, MIT, USA	N/A
Collagen XII	Abcam	Cat#ab196619
Tenascin-C	Abcam	Cat#ab108930, RRID:AB_10865908
Biological samples
Human mesentery and omentum samples	Autopsy Service, Department of Pathology and Laboratory Medicine, University of Wisconsin Hospital and Clinics	N/A
Chemicals, peptides, and recombinant proteins
PNGaseF	New England Biolabs	Cat#P0704L
LysC	Thermo Scientific	Cat#90307
Trypsin	Thermo Scientific	Cat#90058
Critical commercial assays
Subcellular protein fractionation kit	Millipore	Cat#2145
Quantitative Colorimetric Peptide Assay kit	Pierce	Cat#23275
BOND Polymer Refine Detection Kit	Leica	Cat#DS9800
Deposited data
Raw mass spectrometry data	This paper	PRIDE repository, Identifier#PXD061586
Software and algorithms
Scaffold Q+S, v5.2.0	Proteome Software	www.proteomesoftware.com/products/scaffold-qs
HALO, v3.6.4134	Indica labs	www.indicalab.com/halo/
Area quantification 2.4.3 algorithm	HALO, Indica labs	www.indicalab.com/halo/
Prism 10	GraphPad	www.graphpad.com
BioVenn		www.biovenn.nl
UpSetR	Gehlenborg Lab, Harvard Medical School, USA	https://gehlenborglab.shinyapps.io/upsetr/
ImageScope, v12.4.6.5003	Leica	www.leicabiosystems.com/us/digital-pathology/manage/aperio-imagescope/
Other
Orbitrap Q Exactive HF mass spectrometer	Thermo Fisher Scientific	N/A
Axio Observer.Z1 inverted microscope	Zeiss	N/A
Aperio AT2	Leica	RRID# SCR_021256

Experimental model and study participant details

De-identified normal omental and mesentery samples were collected from deceased females over the age of 21 during autopsy (Table S1) and procured from the autopsy service, Department of Pathology and Laboratory Medicine, University of Wisconsin Hospital and Clinics. Tissues were classified as non-human-subjects research by the University of Wisconsin-Madison internal review board, and as such were exempt from ethical approval or the obtention of consent. Samples were excluded if there was a known history of gynecological cancer. Samples were placed in cold PBS and transported to the laboratory, where they were cut into pieces (1-2 cm²) and snap frozen in liquid nitrogen for proteomic analysis or fixed in 4% paraformaldehyde for paraffin embedding and sectioning at the Translational Research Initiatives in Pathology (TRIP) laboratory at the University of Wisconsin Carbone Cancer Center.

Method details

ECM enrichment from tissue samples

150 to 300 mg of tissue was homogenized for 2 minutes at a power setting of 12 using a Bullet Blender and stainless-steel beads of varying diameters (Navy bead lysis kit, Next Advance, Averill Park, NY) following the manufacturer’s instructions. ECM enrichment was achieved through the sequential extraction of intracellular proteins in order of decreasing protein solubility, using a subcellular protein fractionation kit (Millipore, #2145) as previously described.⁴⁵ The efficiency of the sequential extraction of intracellular components and concomitant ECM-protein enrichment was monitored by western blot analysis, probing for collagen I (Sigma, #AB765P, used at a 1 μg/mL concentration), GAPDH (Sigma, #MAB374, used at a 5 μg/mL concentration), and actin using a serum containing anti-actin antibodies (used at a 1/5000 dilution) kindly gifted by the Hynes lab at the Massachusetts Institute of Technology, Cambridge, MA, USA.

Protein sample preparation for mass spectrometry

The ECM-enriched protein samples were subsequently solubilized and digested into peptides as previously described.⁴⁵ In brief, proteins were solubilized in an 8M urea solution prepared in 100mM NH₄HCO₃, and protein disulfide bonds were reduced using 10mM dithiothreitol. Reduced disulfide bonds were then alkylated with 25mM iodoacetamide for 30 minutes at room temperature in the dark. The urea concentration was brought to 2M, and proteins were then deglycosylated with PNGaseF (New England Biolabs, #P0704L) for 2 hours at 37°C and digested sequentially, first with Lys-C (Thermo Scientific, #90307) for 2 hours at 37°C, and then with trypsin (Thermo Scientific, #90058), overnight at 37°C. A fresh aliquot of trypsin was added the following day, and samples were incubated for another 2 hours at 37°C. All incubations were performed under agitation. The samples were acidified with 50% trifluoroacetic acid (TFA) and desalted using Peptide Desalting Spin Columns (Pierce #89852). Peptides were reconstituted in a buffer containing 95% HPLC-grade water, 5% acetonitrile, and 0.1% formic acid, and the concentration of the peptide solution was measured using the Quantitative Colorimetric Peptide Assay kit (Pierce, #23275).

Peptide analysis by LC-MS/MS

Approximately 1μg of desalted peptides of each sample was analyzed at the UIC Mass Spectrometry Core facility on a quadrupole Orbitrap mass spectrometer Q Exactive HF mass spectrometer coupled to an UltiMate 3000 RSLC nano system with a Nanospray Frex Ion Source (Thermo Fisher Scientific). In brief, samples were loaded into a PepMap C18 cartridge (0.3 x 5mm, 5μm particle) trap column and then a 75 μm x 150 mm Waters BEH C18 (130A, 1.7μm, 75μm x 15cm) and separated at a flow rate of 300 nL/min. Solvent A was 0.1% formic acid (FA) in water and solvent B was 0.1% FA, 80% acetonitrile (ACN) in water. The solvent gradient of LC was 5% B in 0-3 min, 8% B in 3.5min, 8-40% B in 85 min, 40-95% B in 90min, wash 95% in 95 min, followed by 5% B equilibration until 105 min. Full MS scans were acquired in the Q-Exactive mass spectrometer over 350-1400 m/z range with a resolution of 120,000 from 5 min to 105 min. The AGC target value was 3.00E+06 for the full scan. The 20 most intense peaks with charge states 2, 3, 4, 5 were fragmented in the HCD collision cell with a normalized collision energy of 28%; these peaks were then excluded for 45s within a mass window of 1.2 m/z. A tandem mass spectrum was acquired in the mass analyzer with a resolution of 30,000. The AGC target value was 1.00E+05. The ion selection threshold was 2.00E+04 counts, and the maximum allowed ion injection time was 50 ms for full scans and 120 ms for fragment ion scans.

Database searching

Spectra were searched against the UniProt human database using MaXQuant (version 2.0.3.0) with the following parameters: parent mass tolerance of 20 ppm, fragment ion mass tolerance of 20ppm, and constant modification on cysteine alkylation, variable modifications on methionine oxidation, deamidation of asparagine, glutamine, N-terminal glutamine to pyroglutamate, and hydroxylation of lysine and proline, the latter two being characteristic post-translational modifications of ECM proteins, in particular collagens and collagen-domain-containing proteins, as we previously reported.³³

Picrosirius red staining

Tissue sections were deparaffinized and rehydrated in a graded ethanol series according to standard procedures. Deparaffinized sections were incubated with picrosirius red (PSR) solution at room temperature for 1 hour. The PSR solution was prepared by dissolving 0.5 g of Direct Red 80 (Sigma-Aldrich) in 500 mL of saturated picric acid (Sigma-Aldrich). Subsequently, sections were washed with 0.5% (v/v) acetic acid in water, dehydrated in 100% ethanol, cleared in xylene, and mounted with Richard-Allan toluene-based mounting medium (Thermo Fisher Scientific). The stained sections were imaged on a Zeiss Axio Observer.Z1 inverted microscope with a Plan-Apochromat 20x/0.8 Ph2 M27 objective using the filter set for mCherry (Zeiss, Oberkochen, Germany).

Immunohistochemical staining

Tissue sections were deparaffinized and rehydrated in a graded ethanol series according to standard procedures. Immunohistochemical staining of paired human omentum and mesentery sections (N=3) was performed at the UIC Research Histology and Tissue Imaging Core. The following antibodies were used: anti-COL12A1 antibody (Abcam #ab196619; used at a final concentration of 10mg/mL), anti-TNC antibody (Abcam #ab108930; used at a final concentration of 10mg/mL, after heat-induced epitope retrieval performed using a Tris solution at pH 9). Staining was performed with the BOND Polymer Refine Detection Kit (Leica, #DS9800) on the BOND RX automated stainer (Leica Biosystems) according to a standard preset protocol. In brief, endogenous peroxidase activity and non-specific binding sites were blocked by sequentially treating samples with peroxidase block (BOND Polymer Refine Detection Kit) and protein block (Background Sniper, Biocare Medical, #BS966) for 15 min. at room temperature. Sections were then incubated with the primary antibody for 30 min at room temperature. After several washes, signal detection was performed with a rabbit anti-horseradish peroxidase (HRP) antibody and 3,3′-diaminobenzidine (DAB; BOND Polymer Refine Kit) by incubating the sections for 15 min. and 10 min. at room temperature, respectively. Sections were counterstained with hematoxylin. All slides were dehydrated in an Autostainer XL and mounted with Micromount (Leica Biosystems). Stained sections were scanned at a resolution of 40X using a Leica Aperio AT2 slide scanner.

Quantification and statistical analysis

Fibrillar collagen quantification

While PSR is commonly imaged with colorimetric or polarized light, its fluorescent signal is as sensitive as other gold-standard collagen fiber imaging modalities.⁷⁶ For each tissue, five fields of view were randomly selected at the mesothelial edge of the tissue. The width of the ECM band from the mesothelial layer to the first adipocytes was measured in ImageJ⁷⁷ in three locations per image and averaged across all measurements for each sample. The mean fluorescence intensity of the mesothelial layer and the first row of adipocytes was measured in each field of view and averaged across all fields of view for that donor. Autofluorescence signal for unstained tissues was minimal and was subtracted to calculate the mean signal intensity (MSI). An additional five fields of view were randomly selected in the adipocyte-rich bulk, and the MSI was measured using ImageJ. Measurements were averaged across the five fields of view for each donor. As we used donor-matched tissues throughout, samples were compared by a paired t-test; p < 0.05 was considered statistically significant.

Collagen XII and tenascin-C quantification

The proportion of collagen XII-positive and tenascin-C-positive area (%) was quantified from the entire section in an unbiased way to account for staining variability within each tissue and sample. Images were manually annotated using the Aperio ImageScope version 12.4.6.5003 to mark tissue boundaries and exclude artefacts such as tissue folding. The annotations were imported to Halo 3.6.4134 (Indica Labs) and analyzed by an experimenter blinded to the proteomic results using the area quantification 2.4.3 algorithm. In brief, the two-color channels (DAB: brown and hematoxylin: blue) were manually assigned from single-stained regions in the tissue. The threshold to detect signal from both the color channels was manually assigned for each staining condition. Next, two phenotypes were defined: collagen XII-positive or tenascin C-positive regions where the staining (brown color) is positive, but the counterstain (blue) is negative (phenotype 1), and a nucleus region where the counterstain is positive (phenotype 2). The area of phenotype 1 (collagen XII or tenascin-C) was obtained and the positive area % is calculated by dividing the area of phenotype 1 by the total tissue area.

Label-free protein quantification from proteomic dataset

Search results were entered into Scaffold Q+S software (v5.2.0, Proteome Software, Portland, OR) for compilation, normalization, and comparison of total precursor ion intensities.

Peptide identifications were accepted if they could be established with a false-discovery rate (FDR) less than 1.0% by the Scaffold Local FDR algorithm. Protein identifications were accepted if they could be established with a false-discovery rate (FDR) less than 1.0% and were identified with at least 2 distinct peptides. Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. Proteins sharing significant peptide evidence were grouped into clusters.

Mass spectrometry output was further annotated to identify ECM and non-ECM components. Specifically, matrisome components are classified as core-matrisome or matrisome-associated components and further categorized into groups based on structural or functional features: ECM glycoproteins, collagens, or proteoglycans for core matrisome components; and ECM-affiliated proteins, ECM regulators, or secreted factors for matrisome-associated components (Table S2). Relative protein quantification was performed using normalized precursor intensities (Table S2E). Raw mass spectrometry data and accompanying metadata file in the Sample and Data Relationship Format (SDRF) using lesSDRF⁷³ have been deposited to the ProteomeXchange Consortium⁷⁴ via the PRIDE partner repository⁷⁵ with the dataset identifier PXD061586.

Statistical analysis

Dot plots represent mean and standard error of mean or standard deviation unless specified. Paired t-test was performed to evaluate statistical significance and p < 0.05 was considered statistically significant. Statistical analysis and data visualization were performed using Prism (GraphPad). Proportional Venn diagrams were generated with the BioVenn application: https://www.biovenn.nl/.⁷⁸ UpSet plots were generated using UpSetR.^79,80

Published: December 17, 2025