Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2019 Nov 1;317(5):G625-G639.
doi: 10.1152/ajpgi.00014.2019. Epub 2019 Sep 23.

Proteomics Analysis of the Matrisome From MC38 Experimental Mouse Liver Metastases

Affiliations
Free PMC article

Proteomics Analysis of the Matrisome From MC38 Experimental Mouse Liver Metastases

Arseniy E Yuzhalin et al. Am J Physiol Gastrointest Liver Physiol. .
Free PMC article

Abstract

Dissemination of primary tumors to distant anatomical sites has a substantial negative impact on patient prognosis. The liver is a common site for metastases from colorectal cancer, and patients with hepatic metastases have generally much shorter survival, raising a need to develop and implement novel strategies for targeting metastatic disease. The extracellular matrix (ECM) is a meshwork of highly crosslinked, insoluble high-molecular-mass proteins maintaining tissue integrity and establishing cell-cell interactions. Emerging evidence identifies the importance of the ECM in cancer cell migration, invasion, intravasation, and metastasis. Here, we isolated the ECM from MC38 mouse liver metastases using our optimized method of mild detergent solubilization followed by biochemical enrichment. The matrices were subjected to label-free quantitative mass spectrometry analysis, revealing proteins highly abundant in the metastatic matrisome. The resulting list of proteins upregulated in the ECM significantly predicted survival in patients with colorectal cancer but not other cancers with strong involvement of the ECM component. One of the proteins upregulated in liver metastatic ECM, annexin A1, was not previously studied in the context of cancer-associated matrisome. Here, we show that annexin A1 was markedly upregulated in colon cancer cell lines compared with cancer cells of other origin and also over-represented in human primary colorectal lesions, as well as hepatic metastases, compared with their adjacent healthy tissue counterparts. In conclusion, our study provides a comprehensive ECM characterization of MC38 experimental liver metastases and proposes annexin A1 as a putative target for this disease.NEW & NOTEWORTHY Here, the authors provide an extensive proteomics characterization of murine colorectal cancer liver metastasis matrisome (the ensemble of all extracellular matrix molecules). The findings presented in this study may enable identification of therapeutic targets or biomarkers of hepatic metastases.

Keywords: S100-A11; annexin A1; colorectal cancer; extracellular matrix; liver metastasis; matrisome.

Conflict of interest statement

No conflicts of interest, financial or otherwise, are declared by the authors.

Figures

None
Fig. 1.
Fig. 1.
Study design and preparation of decellularized matrices for proteomics analysis. A: a cartoon illustrating the model for experimental mouse liver metastases. Briefly, MC38 mouse colon cancer cells were injected into the spleen parenchyma, and tumor cells traveled to the liver via the portal vein. Splenectomy was performed to exclude cancer cell formation in the splenic bed. Mice were humanely culled, and hepatic metastases were excised 12–14 days postoperation. B: study workflow describing extracellular matrix (ECM) isolation and enrichment protocol with the following label-free proteomics analysis [Yuzhalin et al. (59)]. C: representative image of a mouse liver bearing MC38 metastasis. Metastatic lesions are outlined in yellow. D: representative images of mouse livers during decellularization over 72 h. E: decellularized or intact mouse livers were cryosectioned and stained for the indicated ECM proteins (green) or with hematoxylin and eosin (H&E). All sections were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; blue). Original scale bars, 100 μm. Isotype control staining is provided in bottom-right corners. LC-MS/MS, liquid chromatography tandem mass spectrometry.
Fig. 2.
Fig. 2.
Label-free quantitative proteomics of intact and MC38 metastasis-bearing mouse livers. A: total protein was isolated from intact (left), decellularized (middle), and decellularized plus biochemically enriched (right) mouse livers and resolved by SDS-PAGE with the following silver staining. High and low molecular mass areas were considered <100 kDa and >100 kDa, respectively. B: intact livers or MC38 metastasis dissected from mouse livers were decellularized and extracellular matrix (ECM) enriched. After analysis by liquid chromatography tandem mass spectrometry with label-free quantitation (n = 3 biological replicates), the principal component analysis of relative protein abundances between metastasis and intact livers was computed (P < 0.05). Percentage of variance is displayed in parentheses. Gray “clouds” indicate individual proteins, whereas black circles represent replicates. C: proteomics analysis revealed 2,328 proteins in the ECM (both groups were considered). Of these, 140 proteins were classified as the matrisome in accordance with the categorization proposed by Naba et al. (35). Twenty-seven of 140 proteins were significantly different between groups (two-way ANOVA, P < 0 0.01) after restriction to a fold-change threshold of >3 and identification of at least 1 peptide. n.s., not significant.
Fig. 3.
Fig. 3.
A heat map displaying 27 extracellular matrix (ECM) proteins with significantly different abundance between intact and MC38 metastasis-bearing mouse livers (two-way ANOVA, P < 0.01). These proteins were selected based on identification of at least 1 peptide and a fold-change threshold of >3 and ranked in accordance with their corresponding ECM category and fold change. #Technical replicate.
Fig. 4.
Fig. 4.
Overall survival of patients with colorectal adenocarcinoma (A), breast-invasive carcinoma (B), and pancreatic adenocarcinoma (C) who had an alteration in the 13-protein combination provided in Fig. 3 (alteration here means significant overexpression or underexpression). Log rank test. CPTAC, Clinical Proteomic Tumor Analysis Consortium; RPPA, reverse-phase protein array; TCGA, Tissue Cancer Genome Atlas.
Fig. 5.
Fig. 5.
Annexin A1 is overexpressed in the extracellular matrix (ECM) from colorectal cancer liver metastases and might serve as a useful prognostication marker. A: annexin A1 expression in different cancer cell lines. B: Oncomine bioinformatics analysis of 3 different studies [see Skrzypczak et al. (47a) (left), Gaspar et al. (15a) (middle), and Kaiser et al. (23a) (right)] evaluating annexin A1 expression in patients with colorectal cancer. Mann-Whitney test. Small circles indicate range, error bars indicate median, whiskers indicate 95% confidence interval, box bounds indicate 25th and 75th quartiles. C: normalized annexin A1 abundance in the decellularized and enriched ECM fraction from colorectal cancer liver metastases or adjacent unaffected liver tissues (n = 5 per group). Circles and squares indicate biological replicates, error bars indicate mean, whiskers indicate SE. Extracted from Yuzhalin et al. (59). D: immunoblotting for annexin A1 in 8 resected colorectal cancer liver metastases (labeled as M) and adjacent unaffected liver tissue specimens (labeled as N). Whole tissue lysate was used for immunoblotting. GAPDH was used as a loading control. E: densitometry analysis of bands from the experiment in D. Wilcoxon signed-rank test. Error bars indicate median, whiskers indicate range, box bounds indicate 25th and 75th quartiles. F: average staining intensity (left) and representative microphotographs (right) of resected colorectal cancer liver metastases (n = 21) and normal hepatic tissue specimens (n = 6) immunostained for annexin A1 (red). All sections were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; blue). Original scale bars, 100 μm. Mann-Whitney test. Circles and squares indicate biological replicates, error bars indicate mean, whiskers indicate SE. G: ELISA for annexin A1 in serum from healthy blood donors (n = 30), patients with primary colon cancer (n = 30), and patients with liver metastases from colon cancer (n = 40). Kruskal-Wallis test with Dunn’s multiple comparison posttest. Circles, squares, and triangles indicate biological replicates, error bars indicate mean, whiskers indicate SE. H: overall survival of patients with colorectal adenocarcinoma who had an alteration in annexin A1 protein (alteration here means significant overexpression or underexpression). Log rank test. n.s., not significant; RPPA, reverse-phase protein array; TCGA, Tissue Cancer Genome Atlas.
Fig. 6.
Fig. 6.
In the tumor microenvironment, annexin A1 is produced by both cancer and stromal cells. A: scheme illustrating the experimental pipeline. Human colon cancer cells HT-29 were injected intrasplenically into immunocompromised mice to establish hepatic metastases. Resultant liver tumors were decellularized and enriched, as described previously, with the following liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis of human and mouse sequences. B: semiquantitative calculation of the exponentially modified protein abundance index (emPAI) (23) for estimation of a relative proportion of human-derived (i.e., cancer cell-derived; Hu) and mouse-derived (i.e., stroma-derived; Ms) annexin A1 (4 biological replicates per group). The pie chart reflects an averaged emPAI score for both human and mouse annexin A1. C: mouse liver metastases generated using MC38 and HT-29 cells were cryosectioned and stained for annexin A1 (red) and common bone marrow-derived cell marker CD45 (green). All sections were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; blue). Original scale bars, 100 μm. ECM, extracellular matrix.
Fig. 7.
Fig. 7.
Immunostaining for annexin A1 (red) and immune cell markers CD11b, Ly6G, and CD3 (all green) displays their colocalization in the MC38 tumor microenvironment. All sections were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; blue). Arrowheads indicate annexin A1 staining associated with immune markers. Original scale bars, 100 μm.
Fig. 8.
Fig. 8.
Annexin A1 is coexpressed with S100-A11 in liver metastasis extracellular matrix (ECM) and across other cancer types. A: normalized abundance of annexin A1 and S100-A11 in the quantitative proteomics data set from this study. Each circle indicates a replicate. B and C: coexpression of annexin A1 and S100-A11 in the Tissue Cancer Genome Atlas (TCGA) data set (colorectal adenocarcinoma), as assessed by mRNA (B) and protein (C) levels. Each circle represents a patient. D: coexpression of annexin A1 and S100-A11 in the Cancer Cell Line Encyclopedia data set. Each circle represents a human cancer cell line. ALL, acute lymphoblastic leukemia; AML, acute myelogenous leukemia; CML, chronic myelogenous leukemia; DLBCL, diffuse large B cell lymphoma; NSCLC, nonsmall cell lung carcinoma. A–D: Spearman rank correlation was measured. E: immunostaining for annexin A1 and S100-A11 in the decellularized and cryosectioned ECM from MC38 murine liver metastasis. Negative IgG control staining is shown to the left. Original scale bars, 100 μm.
Fig. 9.
Fig. 9.
Annexin A1 reduces proliferation and promotes migration of cancer cells in vitro and diminishes xenograft growth in vivo. A and B: PCR evaluation of Anxa1 (A) or S100-A11 (B) transcripts in Lewis lung carcinoma (LLC) cells transfected with empty vector [control (CTL)] or shRNA targeting Anxa1. Mann-Whitney U-test was used. C: proliferation rate of control and Anxa1-deficient cells. Two-way ANOVA was used. D: photograph of the Boyden chamber (left) and microphotographs (right) of crystal violet-stained control and Anxa1-deficient LLC cells subjected to 48 h Transwell migration assay. E: density quantification of migrated cells in microphotographs from the experiment in D. Each circle and square represents 1 field of view. Mann-Whitney U test was used. F: tumor growth curves of control and Anxa1-deficient LLC cells implanted into flanks of C57BL/6 mice. Two-way ANOVA was used. a.u., arbitrary units; n.s., not significant.

Similar articles

See all similar articles

References

    1. Aoi W, Naito Y, Takagi T, Tanimura Y, Takanami Y, Kawai Y, Sakuma K, Hang LP, Mizushima K, Hirai Y, Koyama R, Wada S, Higashi A, Kokura S, Ichikawa H, Yoshikawa T. A novel myokine, secreted protein acidic and rich in cysteine (SPARC), suppresses colon tumorigenesis via regular exercise. Gut 62: 882–889, 2013. doi:10.1136/gutjnl-2011-300776. - DOI - PubMed
    1. Aran D, Camarda R, Odegaard J, Paik H, Oskotsky B, Krings G, Goga A, Sirota M, Butte AJ. Comprehensive analysis of normal adjacent to tumor transcriptomes. Nat Commun 8: 1077, 2017. doi:10.1038/s41467-017-01027-z. - DOI - PMC - PubMed
    1. Barrow H, Rhodes JM, Yu LG. The role of galectins in colorectal cancer progression. Int J Cancer 129: 1–8, 2011. doi:10.1002/ijc.25945. - DOI - PubMed
    1. Boudhraa Z, Bouchon B, Viallard C, D’Incan M, Degoul F. Annexin A1 localization and its relevance to cancer. Clin Sci (Lond) 130: 205–220, 2016. doi:10.1042/CS20150415. - DOI - PubMed
    1. Brown RJ, Mallory C, McDougal OM, Oxford JT. Proteomic analysis of Col11a1-associated protein complexes. Proteomics 11: 4660–4676, 2011. doi:10.1002/pmic.201100058. - DOI - PMC - PubMed

Publication types

MeSH terms

Feedback