Keloid disorder: Fibroblast differentiation and gene expression profile in fibrotic skin diseases

Keloid disorder: Fibroblast differentiation and gene expression profile in fibrotic skin diseases132 | Experimental Dermatology. 2021;30:132–145.wileyonlinelibrary.com/journal/exd
Received: 28 July 2020 | Revised: 10 November 2020 | Accepted: 13 November 2020
DOI: 10.1111/exd.14243
R E V I E W A R T I C L E
Keloid disorder: Fibroblast differentiation and gene expression profile in fibrotic skin diseases
Edward J. Macarak | Peter J. Wermuth | Joel Rosenbloom | Jouni Uitto
© 2020 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Abbreviations: Akt, Protein kinase B (PKB); Akt-mTORC1, Akt-mammalian target of rapamycin complex 1; ALK-5, TGF-β receptor type 1; BMMSC, bone marrow mesenchymal stem cell; BMP, bone morphogenetic protein; c-Abl, Abelson non-receptor tyrosine kinase; CD44, cluster of differentiation 44 cell surface marker; Cdc42, cell division control protein 42; cFN, cellular fibronectin; c-Jun, c-Jun N-terminal kinase; COL1A1, collagen, type I, alpha1; COL3A1, collagen, type III, alpha1; ECM, extracellular matrix; EMT, Epithelial Mesenchymal Transition; EndoMT, Endothelial-Mesenchymal Transition; ERKs, extracellular signal-regulated kinases; ESC, embryonic stem cell; F-actin, polymeric fibrillar actin; FN, fibronectin; G-actin, monomeric globular actin; GLI-1, glioma-associated oncogene homolog 1 or zinc finger protein; GSK-3B, glycogen synthase kinase 3B; IL-1A, interleukin 1A; IL-6, interleukin 6; IPF, Idiopathic Pulmonary Fibrosis; JNK, stress-activated protein kinase JNK or c-JUN kinase; KALT, keloid associated lymphoid tissue; KPC, keloid precursor cell; MAPKs, mitogen- activated protein kinases; MEK, MAP kinase or ERK kinase; MKL-1, megakaryoblastic leukaemia 1 (MRTF-A); MMP-19, matrix metalloproteinase 19; MRTF-A, myocardin-related transcription factor A (MKL-1); NANOG, homeobox transcription factor in embryonic stem cells; NEDD4, neural precursor cell expressed developmentally down-regulated protein 4; OCT-4, octamer-binding transcription factor 4; pFN, plasma fibronectin; PI3K, phosphoinosotide 3 kinase; PKC-γ, protein kinase C-γ; proα2(1), procollagen α2, type I collagen polypeptide; RGD, arginine-glycine-aspartate; RT-PCR, reverse transcriptase-polymerase chain reaction; SMADs, signal transducers for TGF-β; SRF, serum response factor; SSEA-4, stage-specific embryonic antigen-4; TBRII, TGF-β, receptor serine/threonine protein kinase; TF, transcription factor; TGF-β, transforming growth factor-β; TIMP, tissue inhibitor of metalloproteinases; VEGF, vascular endothelial growth factor; YAP-TAZ, transcriptional coactivators: “yes-associated protein-tafazzin.
The Joan and Joel Rosenbloom Center for Fibrotic Diseases, and the Jefferson Institute of Molecular Medicine, Department of Dermatology and Cutaneous Biology, Sidney Kimmel Medical College at Thomas Jefferson University, Philadelphia, PA, USA
Correspondence Jouni Uitto, Department of Dermatology and Cutaneous Biology, Thomas Jefferson University, 233 S. 10th Street, Suite 450 BLSB, Philadelphia, PA 19107, USA. Email: [email protected]
Funding information The Joan and Joel Rosenbloom Center for Fibrotic Diseases
Abstract Keloid disorder, a group of fibroproliferative skin diseases, is characterized by unre- mitting accumulation of the extracellular matrix (ECM) of connective tissue, primar- ily collagen, to develop cutaneous tumors on the predilection sites of skin. There is a strong genetic predisposition for keloid formation, and individuals of African and Asian ancestry are particularly prone. The principal cell type responsible for ECM ac- cumulation is the myofibroblast derived from quiescent resident skin fibroblasts ei- ther through trans-differentiation or from keloid progenitor stem cells with capacity for multi-lineage differentiation and self-renewal. The biosynthetic pathways leading to ECM accumulation are activated by several cytokines, but particularly by TGF-β sig- nalling. The mechanical properties of the cellular microenvironment also play a criti- cal role in the cell's response to TGF-β, as demonstrated by culturing of fibroblasts derived from keloids and control skin on substrata with different degrees of stiffness. These studies also demonstrated that culturing of fibroblasts on tissue culture plastic in vitro does not reflect their biosynthetic capacity in vivo. Collectively, our current understanding of the pathogenesis of keloids suggests a complex network of inter- acting cellular, molecular and mechanical factors, with distinct pathways leading to myofibroblast differentiation and activation. Keloids can serve as a model system of fibrotic diseases, a group of currently intractable disorders, and deciphering of the critical pathogenetic steps leading to ECM accumulation is expected to identify tar- gets for pharmacologic intervention, not only for keloids but also for a number of other, both genetic and acquired, fibrotic diseases.
K E Y W O R D S extracellular matrix, fibroblast differentiation, fibrotic diseases, keloid disorder, myofibroblast
16000625, 2021, 1, D ow
nloaded from https://onlinelibrary.w
iley O nline L
s and C onditions (https://onlinelibrary.w
iley.com /term
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
1 | CLINIC AL AND GENETIC FE ATURES OF KELOIDS
Keloid disorder, a recently introduced term, collectively refers to a group of fibroproliferative skin diseases, including keloids, hyper- trophic scars, keloidalis nuchae and acne keloidalis.1 Keloids mani- fest clinically as solid papules, nodules and plaques following minor trauma, with primary predilection sites on the ear lobes, face, chest and upper back (Figure 1). The keloid lesions cause considerable morbidity owing to their propensity to infections and associated pruritus, often with cosmetic concern with devastating psychosocial impact. Unlike normal scars, which stabilize temporally after normal healing of the original wound, keloid growth is indeterminate and the lesions can increase in size well beyond the margins of the initial trauma. Since a keloid exhibits tumor-like properties, it is prudent to examine which genes are differentially expressed in keloid tissue ver- sus normal skin to understand how a keloid cell has changed from its normal predecessor. While a variety of approaches have been used to characterize keloids, there are currently no animal models that accurately recapitulate the pathobiology of human keloid lesions.1–3 Given this limitation, investigations have been largely limited to de- fining differences between normal and keloid tissues pathologically, genetically or biologically.
A number of observations suggest that there is a strong genetic contribution to keloid formation. Firstly, there is often a familial clustering, and some studies have suggested an autosomal dominant
mode of inheritance with incomplete penetrance. Secondly, there is an increased prevalence of keloids in individuals with African ances- try, estimated to affect 4%–16% of individuals with dark skin colour, an incidence up to 15 times higher than in Caucasian populations. Individuals of Asian ancestry also have an elevated risk of keloids. Thirdly, a number of genomic susceptibility loci for keloid formation have been identified by genome-wide association studies in both Chinese Han and Japanese populations, and NEDD4 has been sug- gested to be a candidate gene.4–6 NEDD4 enhances proliferation and invasiveness of fibroblasts coupled with activation of TGF-β/ catenin transcriptional activity. Furthermore, altered expression of microRNAs has been suggested to contribute to keloid pathogene- sis by affecting a number of signalling pathways pertinent to wound healing and scar formation.7–11 In spite of these clues, no pathogenic mutations in candidate genes have been reported, and specifically, a number of genes, including those encoding TGF-β1-3, TGF-β recep- tors I, II and III, as well as proα2(I) collagen, have been excluded as pathogenic in keloids.
2 | E X TR ACELLUL AR MATRIX PATHOLOGY IN KELOIDS
Histopathologic examination of mature keloid lesions reveals ac- cumulation of ECM of connective tissue, particularly collagen with compact fibre architecture (Figure 1). Previous studies have shown
F I G U R E 1 Clinical and histopathological features of keloids. Note the predilection sites on face, chest and upper back. Histopathology reveals tightly packed collagen fibres into which a number of fibroblastic cells are embedded in some areas of the lesion. (Clinical pictures are courtesy of Michael Tirgan, MD, President of Keloid Research Foundation, New York, NY.) Histopathology staining is either by haematoxylin-eosin (left and right panels) or by trichrome (middle panel) stains. The bar = 100 µm. Modified with permission from Uitto J, Tirgan M: J Invest Dermatol 140:515–518, 20201
16000625, 2021, 1, D ow
nloaded from https://onlinelibrary.w
iley O nline L
s and C onditions (https://onlinelibrary.w
iley.com /term
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
134 | MACARAK et Al.
histological differences between keloid lesions and hypertrophic scars. Collagen fibres in keloids were shown to be thick and arranged in a parallel fashion without nodules, while hypertrophic scars were characterized by the presence of nodules containing a high density of α-smooth muscle actin (α-SMA)–positive myofibroblast cells and thin fibrillar collagen.12 These observations confirm and extend
the earlier histological observations of Blackburn and Cosman.13 In addition, hypertrophic scars exhibit a changed morphology as the wound matures. Early stages of maturation exhibit the nodules noted above with α-SMA–positive cells; however, in older wounds, the cellular components are fewer in number and are α-SMA–nega- tive. Conversely, keloid lesions exhibit a relatively constant histology
F I G U R E 2 Genetic, cellular and mechanical factors that contribute to the trans-differentiation of precursor cells into myofibroblasts which constitutively express the pro-fibrotic genes resulting in accumulation the extracellular matrix, primarily collagen, as the principal component of keloids
16000625, 2021, 1, D ow
nloaded from https://onlinelibrary.w
iley O nline L
s and C onditions (https://onlinelibrary.w
iley.com /term
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
| 135MACARAK et Al.
with few α-SMA–positive cells.14 Keloids also express high levels of chondroitin sulphate proteoglycans and low-density dermatan sul- phate proteoglycans, while hypertrophic scars contain high levels of low-density dermatan sulphate.15 Keloids also exhibit a higher proliferative capacity and apoptosis resistance, difference in MMP- 19 expression, difference in relative amounts of type I and type III collagens and greater infiltration of immune cells as compared to hy- pertrophic scars.14 Together with clinical observations, these data suggest that it would be pertinent to recognize keloids and hyper- trophic scars as separate entities.16–18
The principal genetically distinct collagens in keloids are types I, III and V, but type VI is also abundantly present particularly in the early stages of development.19,20 Consequently, it has been sug- gested that upregulation of type VI collagen gene expression may serve as an early biomarker of the fibrotic process. Another promi- nent protein found in the keloid ECM and that of systemic sclerosis is periostin, a matricellular protein which has been suggested to pro- mote abnormal scar formation by promoting TGF-β1 secretion by ac- tivation of the RhoA pathway. Periostin is not only highly expressed in healing wounds but is also involved in angiogenesis, a prominent feature of keloids and most likely in all fibrotic diseases.21–24
A number of fibroblastic cells can be noted to be embedded into the collagenous matrix (see Figure 1), often with myofibro- blastic characteristics with defined expression of marker proteins, specifically α-SMA and vimentin. These cells also prominently dis- play rough endoplasmic reticulum and Golgi apparatus, suggesting their activated synthetic and secretory capacity. In addition to my- ofibroblasts, there are populations of cells with myofibroblast-like appearance with lesser degree of differentiation, attesting to the heterogeneity of the phenotype of myofibroblasts and implying plasticity of fibroblasts.25 A number of other cell types have also been identified in keloids, including mast cells26 which are often in close proximity to myofibroblasts and exhibit direct cell-cell con- tacts (Figure 2). Mast cell products, such as histamine, heparin and various cytokines, can stimulate fibroblasts and initiate biochemical events affecting scarring and keloid development.27 There is also a vascular component with the presence of factor VIII–immuno-pos- itive endothelial cells,28 and keloids express vascular endothelial growth factor (VEGF) at high levels which has been suggested to originate from epidermis. Epidermal keratinocytes can also be a source of other modulating factors, such as hypoxia-induced factor 1a (Figure 2). Thus, there is considerable cellular heterogeneity in keloids, with an apparent network of interactive pathways, yet fibro- blastic cells, particularly myofibroblasts, appear to be the principal cell type responsible for ECM production in these fibroproliferative lesions and most likely in all fibrotic diseases.
3 | CELL BIOLOGY IN KELOIDS
It has been suggested, based upon the body locations of keloids, that while the initial stimulus may be some form of injury, the ex- acerbation of the keloid growth is related to the degree of tension
in tissue surrounding the initial lesion, that is, the mechanical force environment. Typically, keloids occur at specific sites including the ears, face, chest and shoulders (see Figure 1). These locations rep- resent regions where the skin is exposed to external forces, some of which can be characterized as abnormal, but the common feature is that the keloid tissue exists in an altered mechanical environment. Interestingly, keloids are absent in body locations that do not sustain either tensional or contractile mechanical force, such as the lower legs, eyelids and other body regions where the skin is perpetually relaxed. Despite these associations, little information is currently available regarding how the mechanical forces associated with the uncontrolled growth of keloids are transduced into molecular signals that transform the normal phenotype of resident fibroblasts into ke- loid cells.
The direction of keloid growth results in characteristic shapes, which depend on their location. For example, keloids on the ante- rior chest grow horizontally in a “crab claw”-like pattern suggesting that they may be related to the tension caused by muscle contrac- tion and/or chest movements which occur during normal breathing. Keloids on the shoulders grow vertically, and this response is likely caused by growth and/or contractile movements of the deltoid mus- cle. Collectively, these observations suggest that stimuli or local fac- tors initiate the process of wound healing in keloid tissue, but that these signals do not terminate once the initial repair is completed, and, consequently, permit the keloid to continue to grow. Such growth is perpetuated by unremitting influence of resident inflam- matory cells releasing cytokine factors, such as transforming growth factor beta (TGF-β), interleukin-1a and interleukin-6 (IL-1a, IL-6) and VEGF (Figure 2).29 Thus, importantly, these locations experience repetitive bouts of mechanical deformation which perpetuates the initial injury by continually exposing existing cells to a mechanically challenging environment.
4 | TR ANS -DIFFERENTIATION OF FIBROBL A STS TO MYOFIBROBL A STIC PHENOT YPE AND MYOFIBROBL A ST AC TIVATION
Essentially every organ and tissue in the human body, including the dermis, can be involved by a reparative process and by pathologic fibrotic reactions. Normally, these repair processes are self-limiting and constitute an important reparative mechanism to restore the functional integrity of injured tissues. However, under pathologic circumstances, the normal homeostatic regulatory mechanisms fail and an uncontrolled fibrotic process occurs, characterized by the progressive accumulation of excessive amounts of connective tis- sue that disrupts the normal organ architecture and function.30–32 Despite considerable understanding of the pathogenesis of the fi- brotic processes attained recently,30–33 effective disease-modifying therapy is currently limited. Presently, idiopathic pulmonary fibrosis (IPF) is treated with one of two newly FDA-approved anti-fibrotic agents (pirfenidone and nintedanib), which are thought to inhibit
16000625, 2021, 1, D ow
nloaded from https://onlinelibrary.w
iley O nline L
s and C onditions (https://onlinelibrary.w
iley.com /term
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
136 | MACARAK et Al.
ECM production by myofibroblasts.34,35 While the approval of these drugs was initially met with enthusiasm, real-world clinical practice has tempered this enthusiasm given their marginal impact on disease mortality.
At the cellular level, it is generally accepted that myofibroblasts are the cells ultimately responsible for the pathologic fibrotic pro- cess.36–39 Myofibroblasts comprise a distinctive population of mes- enchymal cells characterized by α-SMA expression and increased production of fibrillar collagens (types I, III, V and VI) as well as other ECM macromolecules. Increased ECM production is coupled with an increased expression of tissue inhibitor of metalloproteinases (TIMP) leading to inhibition of ECM-degradative enzymes further exacer- bating the fibrosis.40 Furthermore, myofibroblasts induce changes in the biomechanical properties of the affected tissues causing a progressive increase in tissue stiffness, which in itself may act as a potent pro-fibrotic stimulus.41–44
While the origin of myofibroblasts may vary depending upon the tissue reaction, there are several potential sources, including (a) recruitment of fibroblast precursor cells (fibrocytes) from the bone marrow; (b) trans-differentiation of other cell types, including pericytes, adipocytes and epithelial, mesothelial and endothelial cells, into a mesenchymal phenotype; and (c) proliferation and ac- tivation of quiescent tissue-resident fibroblasts into a myofibro- blast phenotype (Figure 2). The current preponderance of opinion is that the activation of tissue-resident fibroblasts is the major source of activated myofibroblasts, although epithelial-to-mesen- chymal transition (EMT), endothelial-to-mesenchymal transition (EndoMT) or pericyte-to-myofibroblast transition may play a role under specific circumstances.45,46 Additionally, recent studies have demonstrated that perivascular Gli1+ progenitor mesenchymal stem-like cells may be an important source of myofibroblasts.47,48 It should be emphasized that even though the trans-differentiation of some of the cell types may not be the predominant source of myofibroblasts, these cells could produce and secrete pro-fibrotic factors, including TGF-β, which play an important role in the fi- brotic process (Figure 2).
In order to transition to bona fide myofibroblasts, cells require the presence of a fibronectin (FN) splice variant, FN-EDA, and TGF- β, or another cytokine, as well as some form of mechanical stress, such as that caused by alterations in the stiffness or elastic constant of the ECM in which the cells are embedded. Matrix stiffness can be defined as a measure of ECM resistance to mechanical deformation and could be viewed as an active component of the trans-differentia- tion process of precursor cells. Once these cells become established, they can only be removed either by apoptosis49,50 or by being de-ac- tivated.51–53 The apoptosis hypothesis was proposed by Gabbiani and co-workers who suggested that growth factor-addicted myofi- broblasts are present during the early phases of wound repair which, upon depletion of these same growth factors, self-induce apopto- sis ending the wound healing response.37 Other mechanisms have been proposed: (a) mechanical loading promotes decreased cellular apoptosis initiating hypertrophic scar formation54 and (b) selective triggering of apoptosis either by activation of cell death signalling
pathways or by inhibition of pro-survival pathways.55 As will be dis- cussed below, the elastic constant or stiffness of the ECM to which a cell is attached can act as a promoter of myofibroblast trans-differ- entiation, whereas release from ECM stiffening results in a loss of contractile function and reduced myofibroblast activation.56,57
The persistent activation of myofibroblasts is the hallmark of tissue and organ fibrosis. Tissue stiffness regulates the subcellular localization of transcription factors (TFs) MRTF-A and B (also known as MKL1 and MKL2). On soft matrices, localization of these TFs is in the cytoplasm; however, increasing matrix stiffness promotes their movement to the cell nucleus with subsequent activation of pro-fi- brotic genes.58,59 Thus, the TF subcellular localization is regulated by the state of the actin cytoskeleton which itself is responsive to the tension in the surrounding ECM. When actin is in the monomeric state (G-actin), the TFs are predominantly localized in the cytoplasm. When actin becomes polymerized to form filamentous actin (F-actin) and stress fibres, the levels of G-actin are depleted and the unbound TFs migrate to the nucleus where they bind to and activate pro-fi- brotic gene promoters (Figure 2).
Upon nuclear translocation, MRTF-A becomes a coactivator of serum response factor (SRF), another TF that plays a central role in regulating the expression of pro-fibrotic genes, including α-SMA, a marker for myofibroblast differentiation. SRF has over 200 target genes many of which involve a variety of functions, such as cyto- skeletal-actin dynamics, integrin-cytoskeletal coupling and cellular growth.60 Thus, SRF is a master regulator of the cellular cytoskel- eton. These same TFs also play a role in controlling cell migratory and mitotic behaviour as well as pro-fibrotic gene expression of trans-differentiated and pathologically activated myofibroblasts. ECM stiffening and TF movement to the nuclei of fibroblasts and epithelial cells initiate the trans-differentiation of a formerly quies- cent, static precursor cell into that of a motile, mitotic and contrac- tile myofibroblast and occurs in fibroblasts found in kidney, heart, liver, dermis and lung and potentially in all organs. This activation mechanism caused by…