Skip to main page content
U.S. flag

An official website of the United States government

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 Jan;10(1):40-58.
doi: 10.1016/j.molonc.2015.08.001. Epub 2015 Aug 11.

Fibroblast activation protein-α, a stromal cell surface protease, shapes key features of cancer associated fibroblasts through proteome and degradome alterations

M M Koczorowska  1 , S Tholen  1 , F Bucher  2 , L Lutz  3 , J N Kizhakkedathu  4 , O De Wever  5 , U F Wellner  6 , M L Biniossek  1 , A Stahl  2 , S Lassmann  7 , O Schilling  8
Affiliations

Fibroblast activation protein-α, a stromal cell surface protease, shapes key features of cancer associated fibroblasts through proteome and degradome alterations

M M Koczorowska et al. Mol Oncol. 2016 Jan.

Abstract

Cancer associated fibroblasts (CAFs) constitute an abundant stromal component of most solid tumors. Fibroblast activation protein (FAP) α is a cell surface protease that is expressed by CAFs. We corroborate this expression profile by immunohistochemical analysis of colorectal cancer specimens. To better understand the tumor-contextual role of FAPα, we investigate how FAPα shapes functional and proteomic features of CAFs using loss- and gain-of function cellular model systems. FAPα activity has a strong impact on the secreted CAF proteome ("secretome"), including reduced levels of anti-angiogenic factors, elevated levels of transforming growth factor (TGF) β, and an impact on matrix processing enzymes. Functionally, FAPα mildly induces sprout formation by human umbilical vein endothelial cells. Moreover, loss of FAPα leads to a more epithelial cellular phenotype and this effect was rescued by exogenous application of TGFβ. In collagen contraction assays, FAPα induced a more contractile cellular phenotype. To characterize the proteolytic profile of FAPα, we investigated its specificity with proteome-derived peptide libraries and corroborated its preference for cleavage carboxy-terminal to proline residues. By "terminal amine labeling of substrates" (TAILS) we explored FAPα-dependent cleavage events. Although FAPα acts predominantly as an amino-dipeptidase, putative FAPα cleavage sites in collagens are present throughout the entire protein length. In contrast, putative FAPα cleavage sites in non-collagenous proteins cluster at the amino-terminus. The degradomic study highlights cell-contextual proteolysis by FAPα with distinct positional profiles. Generally, our findings link FAPα to key aspects of CAF biology and attribute an important role in tumor-stroma interaction to FAPα.

Keywords: Angiogenesis; CAFs; Degradome; FAPα; Secretome; TGFβ.

PubMed Disclaimer

Figures

Figure 1
Figure 1
FAPα is expressed in fibroblasts within the stroma of colorectal cancers. Immunohistochemical stainings were performed for tissue specimens of colorectal cancer patients, showing strong FAPα staining in 14/19 cases of NOS (A–D*) or mucinous (G) histotype of colorectal cancers in fibroblasts adjacent to the tumor glands. Slight or negligible (E–F, H*) FAPα expression was observed in 5/19 cases of colorectal cancers. Magnification = 20× throughout. Note D* and H* are from the same patient with NOS (D*) and mucinous (H*) differentiation.
Figure 2
Figure 2
(a) Western blot detection of CAFs markers αSMA and FAPα in the CT5.3 cell models. The expression of both markers confirms the cancer associated fibroblast origin of CT5.3 cell line. Human recombinant FAPα was loaded as a positive control. (b) FAPα expression level was manipulated in CT5.3 cells by to create loss‐of‐function and gain‐of‐function systems. Knock‐down was made by using FAPα – specific shRNA (CT5.3shFAP) and the respective scrambled RNA control (CT5.3shctr). In the overexpression of active FAPα (CT5.3FAPact) and FAPα deprived of the enzymatic activity by introducing mutation S624→A in the active site (FAPα active site mutant – CT5.3FAPasm) similar FAPα levels were achieved. (c) CT5.3shctr cells display unchanged FAPα expression level, comparing to the CT5.3 wild type cells.
Figure 3
Figure 3
Quantitative secretome profiling. (a) Overlap of 1713 proteins (around 80%) which were consistently identified and quantified in both experiments. (b) Distribution of fold‐change values (log2 of light/heavy ratios) for the secretome comparison of CT5.3shctr/CT5.3shFAP, and CT5.3FAPact/CT5.3FAPasm. (c) In total 2420 and 2140 proteins respectively were identified and quantified in each experiment. The proteins with Fc > 0.58 or Fc < −0.58 were considered as significantly altered.
Figure 4
Figure 4
Western blot detection of several proteins identified as altered in the LC–MS/MS quantitative secretome profiling in CT5.3 cell conditioned medium: tPA, PEDF, VEGFC, LAP, mature TGFβ dimer, CCL2, MMP1, and OPN. The electrophoresis was run on 12.5% PAA gel and 15–25 μg of CCM was loaded. To detect TGFβ, runs in non‐reducing conditions of electrophoresis were applied, for the rest – reducing conditions (10 mM DTT).
Figure 5
Figure 5
(a) STRING (Search Tool for the Retrieval of Interacting Genes/Proteins) analysis of proteins affected by altered FAPα activity. The cluster represents overall connectivity of around 50% of significantly and consistently altered proteins in the secretome profiling analysis. Lines represent associations between proteins, line thickness reflects number of evidences for each association. Dashed circles indicate proteins involved in TGFβ signaling (centrally in the network), angiogenesis and maintaining ECM structure (p < 0.05) (the full list of clusters in Suppl. Tab. S2a). (b) Heat‐map illustrating the Upstream Regulators predicted by Ingenuity Pathway Analysis (IPA) software. The values represent Z‐score calculated by the software; the positive Z‐score indicates predicted positive regulation, the negative Z‐score – negative regulation (the full list of predicted signaling in Suppl. Tab. S2a).
Figure 6
Figure 6
(a) Light microscopic image of low density CT5.3 cells expressing differential FAPα levels. FAPα knock‐down cells (CT5.3shFAP) and the control cells (CT5.3shctr) do not display any morphological differences. (b) Light microscopic image of 100% confluent CT5.3 cells. FAPα knock‐down cells (CT5.3shFAP) loose their fibroblastic spindle shape, comparing to the control cells (CT5.3shctr), in favor of more epithelial‐like morphology. The fibroblastic phenotype is rescued by treating CT5.3shFAP cells with recombinant TGFβ. (c) In TGFβ treated CT5.3shFAP cells the αSMA level is elevated, comparing to the untreated control.
Figure 7
Figure 7
Collagen contraction assay shows differential ability of CT5.3 cells s to process collagen depending on FAPα activity. After 48 h, CAFs embedded in collagen matrix processed collagen which was observed as shrinkage. The difference in collagen shrinkage was visible and consistent in all experiments (n = 3). CAFs overexpressing active, but not inactive FAPα displayed much higher capacity of processing collagen. In CAFs with knocked‐down FAPα this ability was much lower when compared to the control. (a) Representative picture of collagen rings after 48 h. (b) Columns represent quantification of the collagen area after 48 h using ImageJ software. The columns show relative area, setting the samples with lower FAPα level 100%. The error bars represent standard deviation calculated from replicates in all 3 experiments. The differences in both loss‐of‐function and gain‐of‐function systems are significant at p ≤ 0.05 according to two‐tail T‐Student test.
Figure 8
Figure 8
The effect of CAFs co‐culture on sprouting from endothelial cell spheroids (HUVEC). The pro‐angiogenic effect was determined by number and length of sprouts. The graphs represent the relative sprouting rate of each condition normalized to the sample containing CT5.3 cells (a) loss‐of‐function system and (b) gain‐of‐function system with higher FAPα level set to 100%. Error bars show S.E.M. The average number of sprouts for CT5.3shctr, CT5.3shFAP, CT5.3FAPasm and CT5.3FAPact was 2.24 ± 0.24, 1.91 ± 0.19, 1.26 ± 0.10 and 1.81 ± 0.29, and the length of sprouts (in pixel) 61.70 ± 5.30, 46.36 ± 5.91, 31.97 ± 8.63 and 27.85 ± 5.92 respectively.
Figure 9
Figure 9
Identification of FAPα substrates. (a) FAPα specificity PICS profile performed using recombinant FAPα and E. coli peptide library. (b) Amino acid enrichment in nonprime peptides with proline in P1 position, corresponding to the P1Pro N‐termini with the fold change value ≥0.58. The collagenase fingerprint is clearly overrepresented which appears as GlyPro sequence motif in P2–P1 position of non‐prime peptides (P2GlyP1Pro).
Figure 10
Figure 10
(a) Distribution of fold‐change values (log2 of light/heavy ratios) of CT5.3shFAP/CT5.3shctr, and CT5.3FAPact/CT5.3FAPasm in N‐terminome profiling of secretome using TAILS. In total, respectively, 2982 and 3968 free N‐termini were identified in each experiment. (b) Distribution of fold change values among P1Pro N‐termini. The bars indicate the frequency of identifications (axis y) of peptides with quantified, binned Fc values (axis x). The enrichment of positive Fc values in the samples with higher FAPα activity comparing to the lower is indicated with green arrow.
Figure 11
Figure 11
(a) Positional clustering of all P1Pro N termini with Fc ≥ 0.58 found in collagens. The random distribution of identified peptides in the collagen sequences showed endopeptidase activity of FAPα. (b) Positional clustering of all P1Pro N‐termini with Fc ≥ 0.58 found in proteins other than collagen. The strong enrichment of almost half of all considered peptides mapped in the first 10% of the protein sequence represented amino‐peptidase activity of FAPα. (c) Selected FAPα dependent cleavage sites. The identified P1Pro cleavage sites are distributed along entire main chain of COL1A2. Moreover, all internal cleavages follow the motive P2GlyP1Pro. (d) In non‐collagenous proteins, the identified P1Pro cleavage sites cluster close to N‐termini of the protein, as shown for sample proteins, represent aminopeptidase activity of FAPα.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Acharya, P.S. , Zukas, A. , 2006. Fibroblast activation protein: a serine protease expressed at the remodeling interface in idiopathic pulmonary fibrosis. Hum. Pathol. 37, (3) 352–360. - PubMed
    1. Aertgeerts, K. , Levin, I. , 2005. Structural and kinetic analysis of the substrate specificity of human fibroblast activation protein alpha. J. Biol. Chem. 280, (20) 19441–19444. - PubMed
    1. Aggarwal, S. , Brennen, W.N. , 2008. Fibroblast activation protein peptide substrates identified from human collagen I derived gelatin cleavage sites. Biochemistry. 47, (3) 1076–1086. - PMC - PubMed
    1. Aoyama, A. , Chen, W.T. , 1990. A 170-kDa membrane-bound protease is associated with the expression of invasiveness by human malignant melanoma cells. Proc. Natl. Acad. Sci. U. S. A. 87, (21) 8296–8300. - PMC - PubMed
    1. Ariga, N. , Sato, E. , 2001. Stromal expression of fibroblast activation protein/seprase, a cell membrane serine proteinase and gelatinase, is associated with longer survival in patients with invasive ductal carcinoma of breast. Int. J. Cancer. 95, (1) 67–72. - PubMed

Publication types