Identification of a myofibroblast-specific expression signature in skin wounds

Article MATBIO-1350; No. of pages: 16; 4C:

Identification of a myofibroblastspecific expression signature in skin wounds Vera Bergmeier a, b, 1 , Julia Etich a, b, 1 , Lena Pitzler a, b , Christian Frie a, b , Manuel Koch b, c , Matthias Fischer d, e , Gunter Rappl e, f , Hinrich Abken e, f , James J. Tomasek g and Bent Brachvogel a, b a - Department of Pediatrics and Adolescent Medicine, Experimental Neonatology, Medical Faculty, University of Cologne, Cologne, Germany b - Center for Biochemistry, Medical Faculty, University of Cologne, Cologne, Germany c - Institute for Dental Research and Oral Musculoskeletal Biology, Medical Faculty, University of Cologne, Cologne 50931, Germany d - Department of Experimental Pediatric Oncology, University Children's Hospital, Cologne, Germany e - Center for Molecular Medicine Cologne (CMMC), University of Cologne, Cologne, Germany f - Department I of Internal Medicine, Tumorgenetics, Medical Faculty, University of Cologne, Germany g - Department of Cell Biology, University of Oklahoma Health Sciences Center, Oklahoma City, OK, United States

Correspondence to Bent Brachvogel: Department of Pediatrics and Adolescent Medicine, Center for Biochemistry, Medical Faculty, University of Cologne, Joseph-Stelzmann-Str. 52, D-50931 Cologne, Germany. [email protected] http://dx.doi.org/10.1016/j.matbio.2017.07.005

Abstract After skin injury fibroblasts migrate into the wound and transform into contractile, extracellular matrix-producing myofibroblasts to promote skin repair. Persistent activation of myofibroblasts can cause excessive fibrotic reactions, but the underlying mechanisms are not fully understood. We used SMA-GFP transgenic mice to study myofibroblast recruitment and activation in skin wounds. Myofibroblasts were initially recruited to wounds three days post injury, their number reached a maximum after seven days and subsequently declined. Expression profiling showed that 1749 genes were differentially expressed in sorted myofibroblasts from wounds seven days post injury. Most of these genes were linked with the extracellular region and cell periphery including genes encoding for extracellular matrix proteins. A unique panel of core matrisome and matrisome-associated genes was differentially expressed in myofibroblasts and several genes not yet known to be linked to myofibroblast-mediated wound healing were found (e.g. Col24a1, Podnl1, Bvcan, Tinagl1, Thbs3, Adamts16, Adamts19, Cxcl's, Ccl's). In addition, a complex network of G protein-coupled signaling events was regulated in myofibroblasts (e.g. Adcy1, Plbc4, Gnas). Hence, this first characterization of a myofibroblast-specific expression profile at the peak of in situ granulation tissue formation provides important insights into novel target genes that may control excessive ECM deposition during fibrotic reactions. © 2017 Elsevier B.V. All rights reserved.

Introduction Fibroblasts are central effector cells in skin repair. Due to changes in the extracellular matrix (ECM) and the release of inflammatory signals after injury, dermal fibroblasts transform from quiescent cells embedded within an organized ECM in the intact skin into migrating, proliferating cells leaving the ECM niche. These cells enter the wound and differentiate into contractile, alpha smooth muscle 0022-2836/© 2017 Elsevier B.V. All rights reserved.

actin + (SMA +) myofibroblasts [1]. Here, myofibroblasts pull the wound edges together to close the wound [2] and lay down a substantial amount of ECM consisting of collagens, fibronectin, and other matrix components to stabilize the granulation tissue [3,4]. In addition, myofibroblasts are a source for signal molecules that coordinate the recruitment and activation of other cells within the wound. Later, during the remodeling phase of wound repair, myofibroblasts undergo apoptosis or revert to an Matrix Biol. (2017) xx, xxx–xxx

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

2

Identification of a myofibroblast-specific expression signature in skin wounds

inactive phenotype [5,6]. Hence, myofibroblasts are crucial to restore skin integrity after injury. Any imbalances in myofibroblast homeostasis can result in abnormal healing responses and ECM deposition as seen in fibrosis and hypertrophic scarring [7]. There have been earlier attempts to identify gene networks expressed in differentiated skin myofibroblast in order to determine the molecular mechanisms of normal and pathological myofibroblast activation. Previously, genome scale gene expression profiling of cultured human dermal and chicken myofibroblasts [8–10] or skin biopsies from scleroderma patients [11] was performed to outline conserved gene networks. However, ex vivo cultures do not mimic the complex microenvironment in the wound and the heterogenous cellular composition of scleroderma biopsies may hamper the identification of myofibroblast-specific networks. Approaches to identify a myofibroblastspecific transcriptome in skin wounds have been lacking.

Here, we used transgenic SMA-GFP mice [12] to determine the course of myofibroblast differentiation in skin wounds and to define the transcriptome at the peak of myofibroblast formation. We found that myofibroblasts are particular abundant in wounds between five and seven days after injury when they express a unique gene signature linked to ECM production, cell-matrix interactions, cell proliferation and communication. This unique gene expression network therefore defines the transcriptional activation of skin wound myofibroblasts and identifies factors and pathways that may be targeted to inhibit myofibroblast activation, ECM production and fibrotic reactions in vivo.

Results To determine the distribution of myofibroblasts after skin injury full thickness wounds were inflicted on the

Fig. 1. Characterization of the SMA-GFP expression in wound repair. Representative unprocessed cryosections of full thickness wounds from the back skin of SMA-GFP transgenic mice one (D1), three (D3), five (D5), seven (D7), ten (D10) and 14 (D14) days post injury were analyzed by light (top row) and fluorescence (central row) microscopy. Higher magnifications of the wound are shown (bottom row). Three images were merged to generate an overview of an individual wound. The granulation tissue (white line) and the magnified area (red box) are marked. Bar: 400 μm (central row), 100 μm (bottom row). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

Identification of a myofibroblast-specific expression signature in skin wounds

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

Fig. 2. Kinetics of cell recruitment during wound healing. (a) Four color flow cytometry of CD45, CD31 and SMA-GFP expression in cell suspensions isolated from full thickness wounds one (D1) and seven (D7) post injury of eight weeks old transgenic mice. The cell suspension was stained with 7AAD to detect dead cells. Analyzed parameters are given. FSC - forward scatter, SSC - sideward scatter. (b) Percentage of CD45 + hematopoietic cells, CD31 + endothelial cells and SMA-GFP + myofibroblasts at the different time points of flow cytometry analysis were given with standard deviation (the number of animals used for each time point are indicated).

3

4

Identification of a myofibroblast-specific expression signature in skin wounds

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

Identification of a myofibroblast-specific expression signature in skin wounds

back of eight weeks old SMA-GFP mice. The distribution of SMA-GFP + cells was recorded during the progression of skin repair by immunofluorescence microscopy (Fig. 1). SMA-GFP + cells were not present in the wound on day one, i.e. in the early phase of wound repair, but after three days SMA-GFP + cells were detected. The area with SMA-GFP + cells increased with the expansion of the granulation tissue seven and ten days post injury. At day 14 the number of SMA-GFP + cells strongly declined. Next the proportions of SMA-GFP + cells, CD45 + hematopoietic and PECAM1 + endothelial cells were determined by multicolor flow cytometry [13]. One day after injury the majority of cells were CD45 + hematopoietic cells, whereas PECAM1 + endothelial or SMA-GFP + cells were not present (Fig. 2a). SMA-GFP + cells were first detected three days after injury, when PECAM1 + endothelial cells migrate into the wound and number of CD45 + hematopoietic cells start to decline (Fig. 2b). At seven days the proportion of CD45 + hematopoietic cells had decreased to 30% of the infiltrating cell population. Approximately 10% of the cells were PECAM + endothelial cells and 55% SMA-GFP + cells (Fig. 2a, b). Later, at 10 and 14 days after injury, number of CD45 + hematopoietic, PECAM1 + endothelial and SMA-GFP + cells were reduced (Fig. 2b). Hence, CD45 + hematopoietic cells were transiently increased in the early phase of inflammation. SMA-GFP + cells were abundant in the granulation tissue at the peak of neoangiogenesis where they form the majority of cells in the wounds. To define the global transcriptome of woundinfiltrating SMA-GFP + cells in wounds seven days post injury, non-hematopoietic/non-endothelial cell populations with low expression of SMA-GFP from dermis (SMA-GFP low) and with high expression from wounds (SMA-GFP high) were isolated by cell sorting (Fig. 3a) and assayed by microarray analysis. The PCA plot analysis confirmed the identity of the transcriptomes from three independent isolations according to their origin (Fig. 3b). Data sets generated from SMA-GFP low cells (dermis) were substantially different from data sets obtained from SMA-GFP high cells (wound) pointing to significant differences in the transcriptome between SMA-GFP low and SMA-GFP high cells. Bioinformatic analysis identified regulated genes in SMA-GFP high cells. Entities that show signal intensity

5

above background noise and a significant change in relative expression levels (FC ≥ 3, p-value ≤ 0.01) were considered to be differentially expressed between SMA-GFP low and SMA-GFP high cells. Using such selection criteria 819 upregulated and 930 downregulated mRNAs were identified in SMAGFP high cells when compared with SMA-GFP low cells (Supplemental Table 1). Hierarchical cluster analysis identified three clusters of upregulated and four clusters of downregulated genes (Fig. 3c). Smooth muscle actin (Acta2) was within the cluster of the most strongly upregulated genes in SMA-GFP high cells. To identify regulated gene networks that are linked to the myofibroblast phenotype of SMA-GFP high cells 819 upregulated and 930 downregulated genes were imported into the FunRich database (V3) [14] and Gene Ontology (GO) analysis was performed (Fig. 3d, Supplemental Tables 2, 3). The majority of the regulated genes were associated with the extracellular space/region and cell periphery. Interestingly, the upregulated genes were linked to the ECM (cellular component), the positive regulation of cell proliferation and to cell migration and adhesion (biological processes). The downregulated genes were mainly found at the cell surface and plasma membrane (cellular component) and associated with the WNT receptor signaling pathway, the negative regulation of cell proliferation and the positive regulation of gene expression (biological processes). The genes with chemokine, cytokine or growth factor activity were enriched in the cluster of upregulated genes, whereas genes with RNA polymerase II promotor and sequence-specific DNA binding activity were enriched in the cluster of downregulated genes (molecular function). Hence, myofibroblasts express a complex network of genes associated with ECM production, cell-matrix interaction and cell communication. To gain a detailed understanding of these processes a comprehensive characterization of the matrisome and matrisome-associated ([15]) expression signature of myofibroblast was performed. Among the 274 genes of the core matrisome and the 836 genes of the matrisome-associated cluster (http://matrisomeproject. mit.edu) 77 and 157 genes were differentially expressed in myofibroblasts, respectively (Fig. 4a). Within the core matrisome collagens form the largest group of regulated ECM genes. 28% of the collagen genes were upregulated in myofibroblasts and among those Col8a1 and Col5a3 showed the highest fold

Fig. 3. Transcriptome profiling and GO cellular components analysis. (a) The gating scheme illustrates the cell sorting strategy to isolate SMA-GFP low dermal cells from skin and SMA-GFP high myofibroblasts from wounds. (b) PCA-plot of transcriptome analysis. Each circle represents an individual independent isolation. SMA-GFP low dermal cells (red) and SMA-GFP high myofibroblasts (blue) are shown. (c) Non-averaged hierarchical clustered intensity plot (distant metrics euclidean, linkage rule - ward's) of differentially expressed genes is shown. Entities number of the colored clusters are given and relative low (blue) and high (red) expression values are highlighted. SMA (Acta2) expression values are marked. (d) GO term analysis of expressed and regulated mRNAs using the FunRich analysis tool. The highest ranked categories are shown and percentage of enriched genes and p-values are given. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

6

Identification of a myofibroblast-specific expression signature in skin wounds

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

Identification of a myofibroblast-specific expression signature in skin wounds

changes and expression values (Fig. 4b). The increase in Col5a3 was accompanied by an increase of Col5a1 and Col5a2 pointing to a high fibrillar collagen V production in SMA-GFP high myofibroblasts. Moreover, two collagen VI chains (Col6a4, Col6a5) were increased and the fibril-associated collagens with interrupted triple helices (FACIT) collagens XII (Col12a1), XIV (Col14a1) and XXII (Col22a1) as well as collagen XVIII (Col18a1), XXIV (Col24a1) and XXVII (Col27a1) were upregulated. In contrast, the expression of the membrane associated collagen XXIII (Col23a1), which is found on the surface of basal keratinocytes [16], was decreased in SMA-GFP high myofibroblasts. Proteoglycans are essential components of the skin which stabilize the ECM and produce hydrated structures that maintain skin integrity [17]. Several distinct proteoglycans are produced by myofibroblasts and within the proteoglycan matrisome cluster 8 of 36 genes were differentially expressed in SMA-GFP high myofibroblasts (Fig. 4a). Podocan like 1 gene (Podnl1) was highly upregulated while brevican (Bcan), lubricin (Prg4) and the hyaluronan and proteoglycan link protein 4 (Hpln4) were moderately increased (Fig. 4b). Fibromodulin (Fmod), podocan (Podn), nyctalopin (Nyx) and proline arginine-rich end leucine-rich repeat (Prelp) were moderately decreased. Proteoglycans and collagens act in concert with glycoproteins to regulate ECM assembly, ECM-cell interaction and signaling. Within the core matrisome glycoproteins represent the largest cluster and a unique glycoprotein signature of glycoproteins is expressed in SMA-GFP high myofibroblasts. Some of the most upregulated glycoproteins are collagen triple helix repeat containing 1 (Cthrc1), tubulointerstitial nephritis antigen-like 1 (Tinagl1), tenascin C (Tnc) and fibrillin 2 (Fbn2), whereas SPARC related modular calcium binding 1 (Smoc1), laminin γ3 (Lamc3) and gliomedin (Gldn) are among the most downregulated genes in skin. Only few of the upregulated glycoproteins are known to contribute to fibrosis (e.g. Tnc, Fbn2) and presumably to promote myofibroblast-mediated skin wound healing. Several proteins that associate with the ECM, but are not commonly viewed as ECM proteins, are included in the group of “ECM-affiliated” proteins of the matrisomeassociated cluster [18]. Here only 7% of the corresponding genes were up- and 6% downregulated in SMA-GFP high myofibroblasts (Fig. 4a). Interestingly, the transmembrane anchored proteoglycans syndecan 1 (Sdc1) and syndecan 4 (Sdc4) are differentially expressed with Sdc1 being increased and Sdc4 being decreased in these cells (Fig. 4b). The ECM serves as

7

an extracellular signaling network to provide temporal and spatial signals to various cell types and to orchestrate the skin wound healing process. The presentation of the signal can depend on ECM remodeling and within the large group of “regulators” enzymes crosslinking collagen and elastin (Loxl2, Loxl4), processing procollagens (Bmp1, Adamts3, Adamts14), degrading the ECM (Mmp12, Mmp13, Mmp15, Mmp16, Mmp27, Adam12) were differentially expressed in SMA-GFP high myofibroblasts (Fig. 4b). Moreover, these cells show increased expression of various chemokines (Ccl1, Ccl2, Cxcl2, Cxcl3, Cxcl5) and growth factors (Pdgfs, Ngf, Hgf, Fgfs) important for the recruitment and activation of myeloid cells and for the proliferation and differentiation of mesenchymal cells as listed in the “secreted factors” column of the matrisome-associated cluster. The results show that SMA-GFP high myofibroblasts express a unique panel of core matrisome and matrisome-associated genes to adjust the extracellular and cellular microenvironment in the wound and promote wound repair. To validate the expression profile and characterize the dynamics of gene expression during skin wound healing we studied the expression of selected genes by qPCR and immunoblot analysis. Here, total extracts of skin wounds harvested at various time points after injury were used as the amount of cell-sorted material that could be obtained was limited. The expression of several genes of the core matrisome and the matrisome-associated cluster was first characterized by qPCR. In general, the collagen expression was predominately increased between day five and day 14, when myofibroblasts enter the wound. Col5a1 and Col5a3 expression was increased already three days post wounding. Col23a1 was not upregulated, while thrombospondin (Thbs) 3 and Thbs4 were slightly increased in the granulation tissue. In accordance with the microarray results we detected increased levels of Tnc, Clec7a, Mmp13 and Cxcl3 mRNA in healing wounds compared to normal skin. Interestingly, the expression of Tnc, Clec7a and Mmp13 was increased throughout skin wound healing, while Cxcl3 levels decreased with time. Bmp4 and Bmp7 levels were reduced between day three and day seven. Next, the expression of collagen XII, tenascin C, collagen XIV and latent transforming growth factor β binding protein (LTBP) 4 was characterized by immunoblot analysis. Here, collagen XII, collagen XIV and tenascin C (TNC) proteins were found in wounds 3 or 5 days to 14 days after injury. LTBP4 was detected in normal skin and in the later phase of wound repair, 10 and

Fig. 4. Matrisome analysis of differentially expressed genes. (a) The proportion of entities within the differentially expressed genes that are found in the core matrisome or matrisome-associated data set are shown (Venn diagram). The number and percentages of genes found in subcategories are listed. (b) Non-averaged hierarchical clustered intensity plot (distant metrics - euclidean, linkage rule - ward's) of differentially expressed genes (gene level experiment) is shown and the fold change (FC) is given.

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

8

Identification of a myofibroblast-specific expression signature in skin wounds

14 days post injury, but not in wounds seven days after injury. Hence, the results from qPCR and immunoblot analysis resembled the gene expression data from microarray analysis and individual genes were differentially expressed during skin wound healing. Finally, the SMA-GFP high myofibroblasts gene expression signature was used to conduct pathway analysis and determine myofibroblast-specific regulatory networks. Therefore, a FunRich background database search was performed to define genes linked to signaling transduction processes followed by Reactome [19] or KEGG pathway enrichment analysis using the String database [20]. Among the differentially expressed genes 418 were associated with signal transduction processes and according to Reactome pathway analysis G α-protein modulated (s + i + q) signaling events are highly overrepresented (Fig. 6a). KEGG pathway analysis ranked the neuroactive-ligand-receptor interaction pathway highest among the enriched pathways and this curated pathway shares 29 of 33 genes with the G α-protein signaling pathway of Reactome (Fig. 6). The RAF/MAP kinase and PI3K/AKT signaling pathways were enriched in Reactome and KEGG analysis, respectively, whereas adrenoreceptor, RET-, Rap1-, calcium and Wnt signaling pathways were identified in either the Reactome or in the KEGG database. Moreover, the integrin-cell surface interaction and proteoglycans in cancer pathways were enriched G α-protein-mediated signaling events were among the highest ranked signaling pathways and the expression of genes within the pathway was studied in detail. Several G protein-coupled receptors and mediators of cell growth activating the adenylate cyclase (Fig. 6c, G α (s)), were up(Adora2a, Gnas, Gipr, Gpr20, Gpr84, Gpr176, Pthlh) or downregulated (Adcyap1r1, Calcr, Rxfp1, Adrbk2, Vipr2, Adrb3, Adrb1, Arrb1, Ramp2, Ghrh). Interestingly, the adenylate cyclase 1 (Adcy1) was decreased in SMA-GFP high myofibroblasts. Regulated genes that are involved in the inhibition of the adenylate cyclase dependent cAMP production (Fig.6c G α (i)) were mainly receptors (13 of 20 listed genes) or act as hormones (Agt, Ppy) or chemokines (Cxcl16, Ccl5, Cxcl5, Ccl19). Receptors and mediators in the signaling route of G protein-dependent activation of protein kinase C were also enriched (Fig. 6c G α (q)) and here phospholipase C β-4 (Plcb4) was significantly upregulated as well as potent regulators of G protein-dependent smooth muscle contraction (Nts) and blood pressure and body fluid homeostasis (Agt). Hence, a complex interactome of G protein coupled signaling events was differentially expressed in SMA-GFP high myofibroblasts and this network may control the production of a unique ECM and set of matrix-associated factors to stabilize the granulation tissue in wounds seven days post injury.

Discussion Myofibroblasts are a major source of ECM material and here we were able to demonstrate that a unique panel of matrisome and matrisome-associated molecules is expressed at higher levels by SMA-GFP high myofibroblasts during wound healing. A specific set of collagens was increased including the heterotrimeric fibrillar collagen V. Interestingly, the main fibrillar collagens in skin, collagen I and collagen III were not upregulated in SMA-GFP high myofibroblasts seven days post injury. This points to the need for a continuous supply of collagen I and III from skin fibroblasts and wound myofibroblast, whereas collagen V expression is specifically increased in myofibroblasts to adjust the fibrillar collagen repertoire according to the local needs. The variations in expression of individual chains of the heterotrimeric collagen V indicate that specific collagen isoforms are produced by myofibroblast to modulate the mechanical properties of the ECM [21]. Collagen V is known to act as a dominant regulator of collagen fibril assembly [22,23] in vitro and in vivo. Mice deficient for collagen V showed impaired wound healing and abnormal collagen fibril formation, while higher expression of collagen V correlates with skin thickening and stiffening in scleroderma [24]. Therefore, the increased collagen V expression may be initiated in myofibroblasts to transform the provisional collagen fibril network into a more mature and stable structure and to prevent excessive and uncontrolled collagen fibril deposition. This is in line with the increase in FACIT collagens XII, XIV and XXII that are needed to regulate fibril growth for the organization of the ECM suprastructure [25,26] and for fibroblast adhesion to the ECM [27] as well as with the strong increase in collagen VIII. This homotrimeric collagen can form hexagonal networks that stabilize the ECM [28] and may provide signals to stimulate myofibroblast formation [29]. Interestingly, only collagen XXIII was downregulated in myofibroblasts. This membrane bound collagen is mainly expressed by basal keratinocytes and found in fibroblasts close to the basement membrane to promote keratinocyte adhesion and migration [16], but may be dispensable for myofibroblast-matrix interactions in skin wounds. Collagen XVIII and XXIV were also upregulated in wound-derived myofibroblasts. The function of collagen XXIV in wound healing is not yet defined, but in skeletal development collagen XXIV has been reported to promote osteoblastic differentiation and mineralization through TGF-β/Smads signaling pathway [30], whereas products of collagen XVIII known as endostatins may have antiangiogenic and antifibrotic activity [31]. Lack of collagen XVIII in mice leads to a higher capillary density in wounds and to accelerated wound healing [32]. Collagen XVIII expression may be increased in myofibroblasts to normalize neoangiogenesis and prevent excessive fibrotic reactions. The

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

Identification of a myofibroblast-specific expression signature in skin wounds

9

Fig. 5. Characterization of gene expression at different stages of skin wound repair. (a) Expression analysis of core matrisome genes (collagens: Col5a1, Col5a3, Col8a1, Col12a1, Col14a1, Col18a1, Col23a1; proteoglycans: Podnl1; glycoproteins: Thbs3, Thbs4, Tnc) and matrisome-associated genes (ECM affiliated: Clec7a, regulators: Mmp13, secreted factors: Bmp4, Bmp7, Cxcl3) in unwounded skin and wounds 1 to 14 days post injury (D1–D14) using qPCR. The values were normalized to RNU6B and Log2-transformed fold changes relative to unwounded skin (Log2-FC) are shown with standard deviations. The non-transformed fold changes (FC) are marked at the y-axis. For Podnl1 (skin and D1) and Mmp13 (skin) a single ct value could be determined in skin due to low expression values. n = 3 biological replicates per time point. Statistical significances are shown in the Supplemental Table 5. (b) Immunoblot analysis of collagen XII (upper left), collagen XIV (lower left), tenascin C (upper right) and latent TGFβ binding protein 4 (LTBP4, lower right) in unwounded skin (skin) and at 1 to 14 days post injury (D1–D14). Actin was used as loading control. Films of two different exposure intensities are shown and the predicted molecular weight (kDa) is indicated for each protein. n = 3 biological replicates per time point.

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

10

Identification of a myofibroblast-specific expression signature in skin wounds

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

Identification of a myofibroblast-specific expression signature in skin wounds

11

Fig. 7. Summary.

collagen VI chains (α4 and α5) were increasingly expressed in myofibroblasts and for Col6a5 a restricted expression in the granulations tissue was previously demonstrated [33]. Col6a4 and Col6a5 assemble with Col6a1 and Col6a2 in heterotrimeric collagen VI molecules that then form homotetramers [34]. These assemble to heterotypic microfibrils and changes in microfibril composition may alter the mechanical properties of the microfibrils and of the granulation tissue. Matricellular proteins are often upregulated after tissue injury [35] to provide extracellular matrix signals critical for tissue repair processes [36,37]. The matricellular genes Thbs3, Thbs4 and Tnc of the core matrisome were strongly upregulated in myofibroblasts. The function of Thbs3 and Thbs4 during skin wound healing has not yet been described, but Tnc is known to drive fibroblast-to-myofibroblast transformation and fibrotic responses in skin [38]. Interestingly, the expression of the Tnc interacting protein SMOC1 [39] is decreased in myofibroblast. This protein promotes TGF-β-dependent proliferation of endothelial cells [40] and its downregulation in

myofibroblast seven days post injury may support vessel remodeling at this stage of wound repair. Hence, the induction of matricellular genes of the core matrisome in myofibroblasts may provide important extracellular ECM signals for mesenchymal and endothelial cell differentiation and proliferation. Genes encoding noncollagenous basement membrane proteins and proteins of elastic fibers were also differentially expressed in SMA-GFP high myofibroblasts seven days post injury. Lama2 and Lamc3 were downregulated, whereas Nid2 and Fbn2 were upregulated in myofibroblasts. Network-forming laminins together with nidogens stabilize and modulate the properties of basement membranes [41] and changes in the expression of laminins or nidogen isoforms may be needed to orchestrate the ECM assembly in skin wounds. The increased expression of the microfibril component fibrillin 2 and the decreased expression of LTBP4, needed to promote elastic fiber assembly, may act in a similar way to organize the elastic fiber compartment. The membrane-bound proteoglycan syndecan1 [42] and the glycoprotein vitronectin were also

Fig. 6. Identification of regulated signal pathways in myofibroblasts. (a) The Reactome and KEGG database was used to determine enriched signaling pathways within the cluster of genes linked to signaling transduction processes in FunRich enrichment analysis. (*) Hypergeometric tests were used to identify overrepresented entities and for KEGG pathway analysis p-values are corrected for multiple testing using the method of Benjamini and Hochberg. (b) The Venn diagram illustrates the proportion of entities within the KEGG Neuroactive ligand receptor interaction pathway that are also found in the G alpha (s) + (i) + (q) Reactome pathways. (c) String database network analysis of the regulated entities within the G alpha (s) + (i) + (q) Reactome pathway using a medium confidence score. Lines indicate associations between proteins: pink - experimentally determined, blue - curated from databases, green - textmining. Small nodes - proteins with unknown 3d structure, large nodes - proteins with known or predicted 3D structure [20]. Entities that are not connected in the network have not yet been described in String database to interact with any of the displayed entities. Non-averaged hierarchical clustered intensity plot (distant metrics - euclidean, linkage rule - ward's) of the regulated genes are shown. The fold change (FC) and color encoded expression values are given. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

12

Identification of a myofibroblast-specific expression signature in skin wounds

upregulated. Earlier studies showed syndecan1 to be induced in granulation tissue fibroblasts [43] and vitronectin is deposited in the provisional matrix during wound repair to attenuate fibronectin-encoded stimulatory signals [44]. Both are known to organize fibroblast-matrix interactions by modulating αvcontaining integrins (syndecan1) [42] or α5β1 integrin signaling events in fibroblasts (vitronectin) [44] and integrin mediated cell surface interactions are enriched in our data set (Fig. 6). Vitronectin and syndecan1 could be increased in myofibroblasts to sense changes in the ECM microenvironment and to translate those via integrins into appropriate myofibroblast cellular responses. Syndecan4 was also expressed in myofibroblasts, but its expression was at a lower level. This proteoglycan is essential to incorporate TGF-ß/ERK signals to stimulate the concentration of dermal fibroblasts [45], but at seven days post injury wounds were almost closed [46], which may explain the decrease in expression of this syndecan. We detected a significant and strong upregulation of enzymes crosslinking, processing and degrading the wound ECM. The lysyl oxidases (Loxl) are amine oxidases needed to introduce intermolecular crosslinks into collagen and elastin fibrils. Loxl2 and Loxl3 were significantly upregulated in myofibroblasts. LOXL2 is capable of utilizing collagen I [47], collagen IV [48] and elastin [49] as substrates, whereas LOXL3 shows a wide substrate specificity for elastin and seven collagen types (I, II, III, IV, VI, VIII, and X) [50]. Hence, both enzymes may be increased to crosslink and stabilize the granulation tissue ECM to stiffen the wound. Similarly, procollagen processing enzymes bone morphogenic protein 1 (BMP1), a disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS) 3 and ADAMTS 14 may be increased to promote the maturation of procollagens. BMP1 can cleave procollagen I and V [51], while ADAMTS 14 processes procollagen I (α1 and α2 chains). Few reports address the substrate specificity of ADAMTS 3, but this enzyme could be involved in the proteolytic release of active VEGF-C to promote angiogenesis [52]. The functions and substrate specificities of several other upregulated ADAMTS enzymes are unknown (see Fig. 4). The increased expression of matrix metalloprotease (MMP)12, MMP13, Mmp16, Mmp27 and ADAM12, may coordinate tissue remodeling in the wound. These enzymes cleave multiple collagens and non-collagenous ECM proteins and are needed to balance the excessive ECM production by myofibroblasts and avoid the onset and progression of fibrosis (reviewed in [53]). Tissue inhibitors of metalloproteases (TIMPs) inhibit the proteolytic activity of MMPs and Timp1 is strongly expressed in myofibroblasts. Therefore, proteases and their inhibitor contribute to the remodeling of the provisional collagen matrix in the wound (Fig. 5). In addition, proteases can regulate

myofibroblast differentiation. MMP13 may promote myofibroblast-dependent granulation tissue formation [54] and ADAM12 may modulate transdifferentiation of mesenchymal cells into myofibroblast in skin wounds [55]. Syndecans and integrins sense cell-ECM interactions and interact with G-proteins to mediate cellular responses. G α subunit-mediated signaling events were highly enriched in our data set, which emphasize the importance of G protein signaling pathways for the myofibroblast phenotype. Here, the adenylate cyclase 1 was among the most strongly downregulated genes. This central molecule in the G-protein mediated signaling pathway negatively regulates fibroblast-myofibroblast transformation via cAMP [56]. However, the complex interactions between the various regulated signal molecules of the pathway and their importance for myofibroblast homeostasis need to be defined in future experiments. According to our microarray analysis myofibroblasts are a major source of secreted factors in wounds seven days post injury. Here, the genes encoding platelet-derived growth factors (Pdgfs), nerve growth factor (Ngf), hepatocyte growth factor (Hgf) and several fibroblast growth factors (Fgfs) were upregulated. Such growth factors can regulate proliferation, migration and collagen synthesis [57–61] and may act in para- and autocrine fashion to modulate fibroblast recruitment, myofibroblast formation and ECM production and remodeling in skin wounds. Chemokines and growth factors play diverse roles in inflammatory and homeostatic processes but are commonly not associated with the myofibroblast phenotype. We identified a unique myofibroblastspecific set of genes, which are upregulated in myofibroblasts on day seven post injury and are critical for the influx of non-resident monocytes (Ccl1, Ccl2), neutrophils (Cxcl1, Cxcl2, Cxcl3 and Cxcl5) and myofibroblast/macrophage interaction (Csf2/Gm-Csf) [62] in granulation tissue. In summary, we present a gene network specifically expressed in myofibroblasts at the peak of granulation tissue formation in skin wounds (Fig. 7). This network is likely to contribute to the deposition of a specific ECM and to increase the production of unique growth factor/cytokine and chemokine components to orchestrate reepithelialization, neoangiogenesis and granulation tissue homeostasis.

Experimental procedures Wound healing experiments Full thickness wounds were inflicted on the back of SMA-GFP mice [12] as described [13]. Wounds were embedded in tissue tek (Sakura Finetek Europe,

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

Identification of a myofibroblast-specific expression signature in skin wounds

Leiden, Netherlands) 1, 3, 5, 7, 10 and 14 days post injury, sectioned (Leica Cryotome CM3050, Wetzlar, Germany) and analyzed by fluorescence microscopy (Nikon Europe Eclipse TE2000-U Microscope, Amsterdam, Netherlands). Flow cytometry For flow cytometry isolated dermis and wounds were digested as described [13]. Briefly, the tissue was treated with 0.5% (w/v) Dispase® II (Roche) in PBS at 4 °C overnight, subcutaneous tissue, epidermis and vascular structures removed and dermis and wounds incubated in 0.2% (w/v) collagenase type 1 (Worthington, Lakewood, NJ, USA) in DMEM, 10% FCS for 1.5 h at 37 °C. After washing with 5% FCS/PBS cells were passed through 100, 70 and 40 μm cell strainer (BD, Franklin lakes, NJ, USA), incubated with mouse-Fc-Block (BD) on ice and stained with antibodies specific for CD31 and CD45 as well as with 7-actinomycin (7AAD) (BD) followed by flow cytometry (FACSCantoII, BD) or cell sorting (FACSAriaIII, BD). Microarray analysis For microarray analysis phenol-choloroform extracted non-degraded RNA (50 ng) was amplified, labeled and hybridized to a Whole Mouse Genome Microarray 4x44K v2 and a SurePrint G3 Mouse GE 8x60K Microarray (Agilent, Santa Clara, CA, USA) using the protocol of Agilent (Low input Quick Amp Labeling kit, Agilent). After scanning (G2595C scanner, Agilent), data were extracted, combined and processed (Genespring 14.5 Multi-Omic software, Agilent). Data separation was confirmed by PCA plot analysis. A fold change cut off (FC ≥ 3) and moderate t-test cut off (p-value ≤ 0.01) was used to define differentially expressed mRNA. Hierarchical clustering (distant metrics - euclidean, linkage rule - ward's) was applied to determine relationships among the expression levels. GO-term analysis [63] was used to link differentially expressed genes to biological processes, molecular functions and cellular components. KEGG [64] and Reactome analysis [19] was performed to determine enriched signaling pathways and the FunRich tool [14] was used to generate graphical representations of the enrichment analyses. Associated protein interactions between the regulated entities were assessed by String database analysis Version 10 [20]. RNA isolation and quantification by qPCR Skin and wound tissue samples were immediately frozen in liquid nitrogen and RNA isolation and quality assessment was performed as described [65]. RNA (1 μg) was reversely transcribed using

13

miScript HiFlex Buffer from the miScript II RT Kit (Qiagen) to obtain cDNA that contains reversely transcribed small non-coding RNA and mRNA. cDNA (5 ng/5 μl) was used for the miScript SYBR Green PCR assays (Qiagen) in a total volume of 25 μl in clear FrameStar 96 semi-skirted plates (4titude, UK). Thermal cycling parameters were 10 min at 95 °C followed by 40 cycles of 15 s at 95 °C and 1 min at 60 °C. RNU6B-specific primers were used in combination with the miScript Universal Primer (Qiagen) for detection of RNU6B (miScript PCR Controls, Qiagen) or gene-specific primers (Sigma Aldrich, Germany) were used for detection of selected genes. qPCR Analysis was performed with the StepOnePlus™ Real-Time PCR System (Applied Biosystems, CA, USA) and specificity of gene amplification was confirmed by melting curve analysis. Primers used for gene expression analysis are listed in Supplemental Table 4. Western Blot analysis Skin and wound tissue samples were immediately frozen in liquid nitrogen and minced with a pre-cooled stainless steel pestle in a stainless steel mortar. Pulverized tissue was denatured for 10 min at 95 °C in SDS sample buffer (2% SDS, 10% glycerin, 0.4% Bromophenol blue, 0.5% β-Mercaptoethanol, 62.5 mM Tris, pH 6.8) and similar amounts were separated by 8% SDS-PAGE and transferred onto nitrocellulose (Whatman). Primary antibodies detecting actin (Millipore), collagen XII, collagen XIV, tenascin C or LTBP4 (R&D Systems) were detected with corresponding secondary antibodies coupled to horseradish peroxidas (DAKO) and visualized using SuperSignal™ West Pico Chemoluminescent Substrate (Thermo Fisher Scientific).

Data analysis Threshold cycles (C(t)) were determined using StepOne™ software v2.3 and the relative expression levels (fold change) were calculated using the delta-delta C(t) method. Data were normalized to the expression of the snRNA RNU6B [65]. Fold changes were log2-transformed and displayed on a log2-scaled axis. Biological triplicates (individual mice) were analyzed at each time point and standard deviations were indicated. Statistical analysis of the log2-transformed data for the biological triplicates was performed with One-way ANOVA with post-hoc Tukey HSD test. p-values b 0.05 (*), b 0.01 (**) and b 0.001 (***) were considered to be statistically significant. Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.matbio.2017.07. 005.

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

14

Identification of a myofibroblast-specific expression signature in skin wounds

Acknowledgments

[12]

DFG (DFG 2304/5-3, 2304/7-1, 2304/9-1), DFG (SFB829-B04).

[13]

Received 17 March 2017; Received in revised form 1 July 2017; Accepted 31 July 2017 Available online xxxx Keywords: Extracellular matrix; Myofibroblast; Fibrosis; Collagen; Growth factors; Cytokines

[14]

[15]

[16]

1Both contributed equally. [17]

References [18] [1] S.A. Eming, B. Brachvogel, T. Odorisio, M. Koch, Regulation of angiogenesis: wound healing as a model, Prog. Histochem. Cytochem. 42 (2007) 115–170. [2] G. Majno, G. Gabbiani, B.J. Hirschel, G.B. Ryan, P.R. Statkov, Contraction of granulation tissue in vitro: similarity to smooth muscle, Science 173 (1971) 548–550. [3] D. Chester, A.C. Brown, The role of biophysical properties of provisional matrix proteins in wound repair, Matrix Biol. (2016). [4] F. Klingberg, B. Hinz, E.S. White, The myofibroblast matrix: implications for tissue repair and fibrosis, J. Pathol. 229 (2013) 298–309. [5] L. Hecker, R. Jagirdar, T. Jin, V.J. Thannickal, Reversible differentiation of myofibroblasts by MyoD, Exp. Cell Res. 317 (2011) 1914–1921. [6] T. Kisseleva, M. Cong, Y. Paik, D. Scholten, C. Jiang, C. Benner, K. Iwaisako, T. Moore-Morris, B. Scott, H. Tsukamoto, S.M. Evans, W. Dillmann, C.K. Glass, D.A. Brenner, Myofibroblasts revert to an inactive phenotype during regression of liver fibrosis, Proc. Natl. Acad. Sci. U. S. A. 109 (2012) 9448–9453. [7] A. Wells, A. Nuschke, C.C. Yates, Skin tissue repair: matrix microenvironmental influences, Matrix Biol. 49 (2016) 25–36. [8] B. Eckes, P. Zigrino, D. Kessler, O. Holtkotter, P. Shephard, C. Mauch, T. Krieg, Fibroblast-matrix interactions in wound healing and fibrosis, Matrix Biol. 19 (2000) 325–332. [9] D. Kessler, S. Dethlefsen, I. Haase, M. Plomann, F. Hirche, T. Krieg, B. Eckes, Fibroblasts in mechanically stressed collagen lattices assume a “synthetic” phenotype, J. Biol. Chem. 276 (2001) 36575–36585. [10] J. Kosla, M. Dvorak, V. Cermak, Molecular analysis of the TGF-beta controlled gene expression program in chicken embryo dermal myofibroblasts, Gene 513 (2013) 90–100. [11] J.M. Mahoney, J. Taroni, V. Martyanov, T.A. Wood, C.S. Greene, P.A. Pioli, M.E. Hinchcliff, M.L. Whitfield, Systems level analysis of systemic sclerosis shows a network of

[19]

[20]

[21]

[22]

[23]

[24]

[25]

immune and profibrotic pathways connected with genetic polymorphisms, PLoS Comput. Biol. 11 (2015), e1004005. J.-Y. Tsai, Z.-X. Yu, E.F. Wawrousek, Generation and characterization of SMAA-GFP mice, Invest. Ophthalmol. Vis. Sci. 42 (2001) 2734. J. Etich, V. Bergmeier, C. Frie, S. Kreft, L. Bengestrate, S. Eming, C. Mauch, B. Eckes, H. Ulus, F.E. Lund, G. Rappl, H. Abken, M. Paulsson, B. Brachvogel, PECAM1(+)/Sca1(+)/ CD38(+) vascular cells transform into myofibroblast-like cells in skin wound repair, PLoS One 8 (2013), e53262. M. Pathan, S. Keerthikumar, C.S. Ang, L. Gangoda, C.Y. Quek, N.A. Williamson, D. Mouradov, O.M. Sieber, R.J. Simpson, A. Salim, A. Bacic, A.F. Hill, D.A. Stroud, M.T. Ryan, J.I. Agbinya, J.M. Mariadason, A.W. Burgess, S. Mathivanan, FunRich: an open access standalone functional enrichment and interaction network analysis tool, Proteomics 15 (2015) 2597–2601. A. Naba, K.R. Clauser, H. Ding, C.A. Whittaker, S.A. Carr, R.O. Hynes, The extracellular matrix: tools and insights for the “omics” era, Matrix Biol. 49 (2016) 10–24. G. Veit, D. Zwolanek, B. Eckes, S. Niland, J. Kapyla, M.C. Zweers, A. Ishada-Yamamoto, T. Krieg, J. Heino, J.A. Eble, M. Koch, Collagen XXIII, novel ligand for integrin alpha2beta1 in the epidermis, J. Biol. Chem. 286 (2011) 27804–27813. M.M. Smith, J. Melrose, Proteoglycans in normal and healing skin, Adv. Wound Care 4 (2015) 152–173. R.O. Hynes, A. Naba, Overview of the matrisome—an inventory of extracellular matrix constituents and functions, Cold Spring Harb. Perspect. Biol. 4 (2012) a004903. A. Fabregat, K. Sidiropoulos, P. Garapati, M. Gillespie, K. Hausmann, R. Haw, B. Jassal, S. Jupe, F. Korninger, S. McKay, L. Matthews, B. May, M. Milacic, K. Rothfels, V. Shamovsky, M. Webber, J. Weiser, M. Williams, G. Wu, L. Stein, H. Hermjakob, P. D'Eustachio, The reactome pathway knowledgebase, Nucleic Acids Res. 44 (2016) D481–487. D. Szklarczyk, A. Franceschini, S. Wyder, K. Forslund, D. Heller, J. Huerta-Cepas, M. Simonovic, A. Roth, A. Santos, K.P. Tsafou, M. Kuhn, P. Bork, L.J. Jensen, C. von Mering, STRING v10: protein-protein interaction networks, integrated over the tree of life, Nucleic Acids Res. 43 (2015) D447–452. D.E. Birk, Type V collagen: heterotypic type I/V collagen interactions in the regulation of fibril assembly, Micron 32 (2001) 223–237. M. Sun, S. Chen, S.M. Adams, J.B. Florer, H. Liu, W.W. Kao, R.J. Wenstrup, D.E. Birk, Collagen V is a dominant regulator of collagen fibrillogenesis: dysfunctional regulation of structure and function in a corneal-stroma-specific Col5a1-null mouse model, J. Cell Sci. 124 (2011) 4096–4105. R.J. Wenstrup, J.B. Florer, E.W. Brunskill, S.M. Bell, I. Chervoneva, D.E. Birk, Type V collagen controls the initiation of collagen fibril assembly, J. Biol. Chem. 279 (2004) 53331–53337. P. Martin, W.R. Teodoro, A.P. Velosa, J. de Morais, S. Carrasco, R.B. Christmann, C. Goldenstein-Schainberg, E.R. Parra, M.L. Katayama, M.N. Sotto, V.L. Capelozzi, N.H. Yoshinari, Abnormal collagen V deposition in dermis correlates with skin thickening and disease activity in systemic sclerosis, Autoimmun. Rev. 11 (2012) 827–835. P. Agarwal, D. Zwolanek, D.R. Keene, J.N. Schulz, K. Blumbach, D. Heinegard, F. Zaucke, M. Paulsson, T. Krieg, M. Koch, B. Eckes, Collagen XII and XIV, new partners of cartilage oligomeric matrix protein in the skin extracellular matrix suprastructure, J. Biol. Chem. 287 (2012) 22549–22559.

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

Identification of a myofibroblast-specific expression signature in skin wounds

[26] H.L. Ansorge, X. Meng, G. Zhang, G. Veit, M. Sun, J.F. Klement, D.P. Beason, L.J. Soslowsky, M. Koch, D.E. Birk, Type XIV collagen regulates fibrillogenesis: premature collagen fibril growth and tissue dysfunction in null mice, J. Biol. Chem. 284 (2009) 8427–8438. [27] M. Koch, J. Schulze, U. Hansen, T. Ashwodt, D.R. Keene, W.J. Brunken, R.E. Burgeson, P. Bruckner, L. BrucknerTuderman, A novel marker of tissue junctions, collagen XXII, J. Biol. Chem. 279 (2004) 22514–22521. [28] S. Stephan, M.J. Sherratt, N. Hodson, C.A. Shuttleworth, C.M. Kielty, Expression and supramolecular assembly of recombinant alpha1(viii) and alpha2(viii) collagen homotrimers, J. Biol. Chem. 279 (2004) 21469–21477. [29] B. Skrbic, K.V. Engebretsen, M.E. Strand, I.G. Lunde, K.M. Herum, H.S. Marstein, I. Sjaastad, P.K. Lunde, C.R. Carlson, G. Christensen, J.L. Bjornstad, T. Tonnessen, Lack of collagen VIII reduces fibrosis and promotes early mortality and cardiac dilatation in pressure overload in mice, Cardiovasc. Res. 106 (2015) 32–42. [30] W. Wang, D. Olson, G. Liang, R.T. Franceschi, C. Li, B. Wang, S.S. Wang, S. Yang, Collagen XXIV (Col24alpha1) promotes osteoblastic differentiation and mineralization through TGF-beta/Smads signaling pathway, Int. J. Biol. Sci. 8 (2012) 1310–1322. [31] Y. Yamaguchi, T. Takihara, R.A. Chambers, K.L. Veraldi, A.T. Larregina, C.A. Feghali-Bostwick, A peptide derived from endostatin ameliorates organ fibrosis, Sci. Transl. Med. 4 (2012), 136ra171. [32] L. Seppinen, R. Sormunen, Y. Soini, H. Elamaa, R. Heljasvaara, T. Pihlajaniemi, Lack of collagen XVIII accelerates cutaneous wound healing, while overexpression of its endostatin domain leads to delayed healing, Matrix Biol. 27 (2008) 535–546. [33] S. Lettmann, W. Bloch, T. Maass, A. Niehoff, J.N. Schulz, B. Eckes, S.A. Eming, P. Bonaldo, M. Paulsson, R. Wagener, Col6a1 null mice as a model to study skin phenotypes in patients with collagen VI related myopathies: expression of classical and novel collagen VI variants during wound healing, PLoS One 9 (2014), e105686. [34] T. Maass, C.P. Bayley, M. Morgelin, S. Lettmann, P. Bonaldo, M. Paulsson, C. Baldock, R. Wagener, Heterogeneity of collagen VI microfibrils: structural analysis of noncollagenous regions, J. Biol. Chem. 291 (2016) 5247–5258. [35] L.E. Tracy, R.A. Minasian, E.J. Caterson, Extracellular matrix and dermal fibroblast function in the healing wound, Adv. Wound Care 5 (2016) 119–136. [36] T.R. Kyriakides, S. Maclauchlan, The role of thrombospondins in wound healing, ischemia, and the foreign body reaction, J. Cell. Commun. Signal. 3 (2009) 215–225. [37] J.E. Murphy-Ullrich, E.H. Sage, Revisiting the matricellular concept, Matrix Biol. 37 (2014) 1–14. [38] S. Bhattacharyya, W. Wang, L. Morales-Nebreda, G. Feng, M. Wu, X. Zhou, R. Lafyatis, J. Lee, M. Hinchcliff, C. FeghaliBostwick, K. Lakota, G.R. Budinger, K. Raparia, Z. Tamaki, J. Varga, Tenascin-C drives persistence of organ fibrosis, Nat. Commun. 7 (2016) 11703. [39] F. Brellier, S. Ruggiero, D. Zwolanek, E. Martina, D. Hess, M. Brown-Luedi, U. Hartmann, M. Koch, A. Merlo, M. Lino, R. Chiquet-Ehrismann, SMOC1 is a tenascin-C interacting protein over-expressed in brain tumors, Matrix Biol. 30 (2011) 225–233. [40] K. Awwad, J. Hu, L. Shi, N. Mangels, R. Abdel Malik, N. Zippel, B. Fisslthaler, J.A. Eble, J. Pfeilschifter, R. Popp, I. Fleming, Role of secreted modular calcium-binding protein 1

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

[52]

[53]

[54]

[55]

15

(SMOC1) in transforming growth factor beta signalling and angiogenesis, Cardiovasc. Res. 106 (2015) 284–294. D. Breitkreutz, I. Koxholt, K. Thiemann, R. Nischt, Skin basement membrane: the foundation of epidermal integrity—BM functions and diverse roles of bridging molecules nidogen and perlecan, Biomed. Res. Int. 2013 (2013) 179784. R.A. Jurjus, Y. Liu, S. Pal-Ghosh, G. Tadvalkar, M.A. Stepp, Primary dermal fibroblasts derived from sdc-1 deficient mice migrate faster and have altered alphav integrin function, Wound Repair Regen. 16 (2008) 649–660. K. Elenius, S. Vainio, M. Laato, M. Salmivirta, I. Thesleff, M. Jalkanen, Induced expression of syndecan in healing wounds, J. Cell Biol. 114 (1991) 585–595. C.D. Gildner, D.C. Roy, C.S. Farrar, D.C. Hocking, Opposing effects of collagen I and vitronectin on fibronectin fibril structure and function, Matrix Biol. 34 (2014) 33–45. Y. Chen, X. Shi-Wen, J. van Beek, L. Kennedy, M. McLeod, E.A. Renzoni, G. Bou-Gharios, S. Wilcox-Adelman, P.F. Goetinck, M. Eastwood, C.M. Black, D.J. Abraham, A. Leask, Matrix contraction by dermal fibroblasts requires transforming growth factor-beta/activin-linked kinase 5, heparan sulfate-containing proteoglycans, and MEK/ERK: insights into pathological scarring in chronic fibrotic disease, Am. J. Pathol. 167 (2005) 1699–1711. S. Kreft, A.R. Klatt, J. Strassburger, E. Poschl, R.J. Flower, S. Eming, C. Reutelingsperger, A. Brisson, B. Brachvogel, Skin wound repair is not altered in the absence of endogenous AnxA1 or AnxA5, but pharmacological concentrations of AnxA4 and AnxA5 inhibit wound hemostasis, Cells Tissues Organs 201 (2016) 287–298. H.M. Rodriguez, M. Vaysberg, A. Mikels, S. McCauley, A.C. Velayo, C. Garcia, V. Smith, Modulation of lysyl oxidase-like 2 enzymatic activity by an allosteric antibody inhibitor, J. Biol. Chem. 285 (2010) 20964–20974. C. Anazco, A.J. Lopez-Jimenez, M. Rafi, L. Vega-Montoto, M.Z. Zhang, B.G. Hudson, R.M. Vanacore, Lysyl oxidaselike-2 cross-links collagen IV of glomerular basement membrane, J. Biol. Chem. 291 (2016) 25999–26012. Y.M. Kim, E.C. Kim, Y. Kim, The human lysyl oxidase-like 2 protein functions as an amine oxidase toward collagen and elastin, Mol. Biol. Rep. 38 (2011) 145–149. J.E. Lee, Y. Kim, A tissue-specific variant of the human lysyl oxidase-like protein 3 (LOXL3) functions as an amine oxidase with substrate specificity, J. Biol. Chem. 281 (2006) 37282–37290. A.M. Muir, D. Massoudi, N. Nguyen, D.R. Keene, S.J. Lee, D.E. Birk, J.M. Davidson, M.P. Marinkovich, D.S. Greenspan, BMP1-like proteinases are essential to the structure and wound healing of skin, Matrix Biol. 56 (2016) 114–131. M. Bekhouche, A. Colige, The procollagen N-proteinases ADAMTS2, 3 and 14 in pathophysiology, Matrix Biol. 44-46 (2015) 46–53. C. Bonnans, J. Chou, Z. Werb, Remodelling the extracellular matrix in development and disease, Nat. Rev. Mol. Cell Biol. 15 (2014) 786–801. M. Toriseva, M. Laato, O. Carpen, S.T. Ruohonen, E. Savontaus, M. Inada, S.M. Krane, V.M. Kahari, MMP-13 regulates growth of wound granulation tissue and modulates gene expression signatures involved in inflammation, proteolysis, and cell viability, PLoS One 7 (2012), e42596. P. Cipriani, P. Di Benedetto, P. Ruscitti, V. Liakouli, O. Berardicurti, F. Carubbi, F. Ciccia, G. Guggino, F. Zazzeroni, E. Alesse, G. Triolo, R. Giacomelli, Perivascular cells in

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005

16

[56]

[57]

[58]

[59]

Identification of a myofibroblast-specific expression signature in skin wounds

diffuse cutaneous systemic sclerosis overexpress activated ADAM12 and are involved in myofibroblast transdifferentiation and development of fibrosis, J. Rheumatol. 43 (2016) 1340–1349. J.S. Swaney, H.H. Patel, U. Yokoyama, B.P. Head, D.M. Roth, P.A. Insel, Focal adhesions in (myo)fibroblasts scaffold adenylyl cyclase with phosphorylated caveolin, J. Biol. Chem. 281 (2006) 17173–17179. J.C. Chen, B.B. Lin, H.W. Hu, C. Lin, W.Y. Jin, F.B. Zhang, Y.A. Zhu, C.J. Lu, X.J. Wei, R.J. Chen, NGF accelerates cutaneous wound healing by promoting the migration of dermal fibroblasts via the PI3K/Akt-Rac1-JNK and ERK pathways, Biomed. Res. Int. 2014 (2014) 547187. A. Joannes, S. Brayer, V. Besnard, J. Marchal-Somme, M. Jaillet, P. Mordant, H. Mal, R. Borie, B. Crestani, A.A. Mailleux, FGF9 and FGF18 in idiopathic pulmonary fibrosis promote survival and migration and inhibit myofibroblast differentiation of human lung fibroblasts in vitro, Am. J. Phys. Lung Cell. Mol. Phys. 310 (2016) L615–629. N. Lian, T. Li, Growth factor pathways in hypertrophic scars: molecular pathogenesis and therapeutic implications, Biomed Pharmacother 84 (2016) 42–50.

[60] C. Shimbori, P.S. Bellaye, J. Xia, J. Gauldie, K. Ask, C. Ramos, C. Becerril, A. Pardo, M. Selman, M. Kolb, Fibroblast growth factor-1 attenuates TGF-beta1-induced lung fibrosis, J. Pathol. 240 (2016) 197–210. [61] M. Toyoda, H. Takayama, N. Horiguchi, T. Otsuka, T. Fukusato, G. Merlino, H. Takagi, M. Mori, Overexpression of hepatocyte growth factor/scatter factor promotes vascularization and granulation tissue formation in vivo, FEBS Lett. 509 (2001) 95–100. [62] S. Vyalov, A. Desmouliere, G. Gabbiani, GM-CSF-induced granulation tissue formation: relationships between macrophage and myofibroblast accumulation, Virchows Arch. B Cell. Pathol. Incl. Mol. Pathol. 63 (1993) 231–239. [63] C. Gene Ontology, Gene ontology consortium: going forward, Nucleic Acids Res. 43 (2015) D1049–1056. [64] M. Kanehisa, Y. Sato, M. Kawashima, M. Furumichi, M. Tanabe, KEGG as a reference resource for gene and protein annotation, Nucleic Acids Res. (2015). [65] J. Etich, V. Bergmeier, L. Pitzler, B. Brachvogel, Identification of a reference gene for the quantification of mRNA and miRNA expression during skin wound healing, Connect. Tissue Res. 58 (2017) 196–207.

Please cite this article as: V. Bergmeier, et al., Identification of a myofibroblast-specific expression signature in skin wounds, Matrix Biol (2017), http://dx.doi.org/10.1016/j.matbio.2017.07.005