Could Aging Human Skin Use a Connective Tissue Growth Factor

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector commentary

pattern in human ridged skin as an Vassar R, Coulombe PA, Degenstein L et al. (1991) localization for CTGF, TGF-β, and adaptation to high physical stress. Br J Mutant keratin expression in transgenic mice type I procollagen-α1, as well as Dermatol 139:767–75 causes marked abnormalities resembling a human genetic skin disease. Cell 64:365–80 CTGF’s effects on TGF-β signaling. Takahashi K, Paladini RD, Coulombe PA (1995) Using samples of young (21–30 years Cloning and characterization of multiple Wilson NJ, Messenger AG, Leachman SA et al. human genes and cDNAs encoding highly (2010) Keratin K6c mutations cause focal of age) and aged (80 or more years related type 2 keratin 6 isoforms. J Biol palmoplantar keratoderma. J Invest Dermatol of age) but otherwise normal human Chem 270:18581–92 130:425–9 skin, and laser-capture microdissect- ed cells, the authors show that TGF-β and CTGF are normally expressed and produced in skin and skin fibroblasts. See related article on pg 415 Comparing samples from young and aged subjects, Quan et al. found that Could Aging Human Skin Use a in aged skin and fibroblasts, TGF-β, CTGF, and type I procollagen mRNA Connective Tissue Growth Factor and protein levels were coordinately reduced. Using transfected normal human dermal fibroblasts, knock- Boost to Increase Collagen Content? down of CTGF was associated with Noelynn Oliver1, Mark Sternlicht1, Karin Gerritsen2 decreased type I procollagen promoter and Roel Goldschmeding1,2 activity (COL1A2), mRNA, and protein content, whereas overexpression The roles of connective tissue growth factor (CTGF) and transforming growth of V5-tagged human CTGF increased factor-β (TGF-β), both well-known collagen production stimulators, were the same readouts. Three approaches examined in skin aging. Aged skin and fibroblasts exhibited a coordinate to block signaling due to endogenous decrease in CTGF, TGF-β, and type I procollagen expression and content. CTGF TGF-β were used to assess molecular knockdown and TGF-β blockade in normal dermal fibroblasts reduced procol- mechanisms by which CTGF modu- lagen expression, whereas overexpressing CTGF increased procollagen by a lates type I procollagen: (i) a specific TGF-β/Smad signaling–dependent mechanism without involving Smad2/3. TGF-βRI kinase inhibitor SB431542, Journal of Investigative Dermatology (2010) 130, 338–341. doi:10.1038/jid.2009.331 (ii) Smad4 knockdown, and (iii) overexpression of inhibitory Smad7. All approaches completely blocked the Background reported disease associations (Leask CTGF-mediated increase in type I pro- Type I collagen is a major struc- et al., 2009; Cicha et al., 2009), collagen expression and production, tural protein in human skin and, by potential macromolecular interac- indicating that TGF-β receptor and mass, the most abundant protein in tions, and complex gene regulation Smad signaling are required. However, the human body. The association of that also involves disease-associated no effects of CTGF knockdown or over- age-dependent collagen loss with processes and factors (Figure 1). The expression were observed on TGF-β- thinning and fragility of elderly skin evolution of a purely pathogenic fac- dependent Smad2/3 phosphorylation has long been appreciated, yet the tor is unlikely without invoking an or Smad3 transcriptional activity. underlying mechanisms are not well extreme selfish-gene concept, and the On the basis of these findings, the understood. A recent study (Quan results of Quan et al., which address authors conclude that endogenous et al., 2010, this issue) suggests that CTGF function in normal young and production of both TGF-β and CTGF diminished expression of connective aging skin, provide an alternative in human dermal fibroblasts normally tissue growth factor (CTGF), together view of CTGF’s physiological signifi- acts to modulate type I procollagen with diminished transforming growth cance in tissue homeostasis. expression in skin. They propose that factor (TGF)-β/Smad signaling, is this involves a TGF-β/Smad/CTGF responsible for this progressive loss of CTGF in aging skin: new findings axis that is operated by interdepen- dermal collagen. CTGF is a secreted, By means of associative and mecha- dent, yet distinct, mechanisms to matricellular protein, and, like that nistic in vivo and in vitro studies, regulate type I procollagen produc- of other such proteins, its function is Quan et al. (2010) probe the impor- tion. Decreased expression of TGF-β thought to be regulatory, not struc- tance of CTGF and TGF-β for col- and CTGF by aged skin fibroblasts is tural. CTGF is generally considered a lagen loss in aging skin. They report proposed to underlie age-associated pathogenic factor because of its many findings on expression, content, and downregulation of the TGF-β/Smad/ CTGF axis, thereby leading to reduced type I procollagen expression. These 1Therapeutics Research, FibroGen, Inc., San Francisco, California, USA and 2Department of Pathology, University Medical Center Utrecht, Utrecht, The Netherlands results suggest beneficial effects of Correspondence: Noelynn Oliver, FibroGen, Inc., 409 Illinois Street, San Francisco, California 94080, USA. CTGF and provide novel insight into E-mail: [email protected] the mechanisms of skin aging.

338 Journal of Investigative Dermatology (2010), Volume 130 commentary

TGF-β Complexity of CTGF action: binding Angiotensin II CAATbox partners, metabolism, and expression VEGF ? CAESAR CTGF is a modular protein composed Sp1 of four domains that have homol- TNF-α Ets1 CTGF gene E5 Nitric oxide 3′ UTR ogy with certain protein motifs. Each TATA E1 CTGF domain has been reported to SMADs BCE-1 E2 E4 35 kDa interact with biologically active fac- E3 mRNA NF- B stabilizilizer tors and binding partners to regulate HIF-1 their effects (Figure 2). Of note, in AP1 HRE 5.2 Sp3 HRE 5.1 lung epithelial cells CTGF has been Aldosterone, (mouse) (mouse) mechanical reported to enhance binding of TGF-β ? Polymorphic stress C/EBP Sp1/Sp3 site with TGF-β receptor complex, there- ? Glucocorticoids by increasing Smad2 phosphorylation Hypoxia Endothelin-1 and Smad-dependent gene transcrip- Bradykinin Thrombin, factor Xa, factor VIIa, high glucose, tion (Abreu et al., 2002). However, advanced glycation end products Quan et al. (2010) demonstrate that, although CTGF knockdown or overexpression affected type I procolla- Figure 1. CTGF gene regulation is complex. The human connective tissue growth factor (CTGF) gen expression in a TGF-β-dependent promoter contains multiple control elements that are regulated by a variety of inducing factors (green manner, it did not alter Smad2/3 arrows), inhibitory factors (red dashed lines), and mechanical forces or hypoxic stimuli. Many of these phosphorylation and had no effect regulators are disease associated. CTGF gene transcription is strongly induced by transforming growth on TGF-β/Smad3-dependent reporter factor-β (TGF-β) and also by angiotensin II, endothelin-1, and advanced glycation end products, at gene transcription. This observation least partially through TGF-β and SMAD-independent pathways. Posttranscriptional control elements (CAESAR, cis-acting element of structure-anchored repression) and a functional, scleroderma-associated is concordant with an independent polymorphism (Sp1/Sp3 site) have also been observed. BCE, basal control element; E1–E5, exons 1–5; observation (Pannu et al., 2007) that HRE, hypoxia-responsive element; NF-κB, nuclear factor kappa light-chain enhancer of activated B cells; TGF-β-dependent type 1 procollagen TNF-α, tumor necrosis factor-α; VEGF, vascular endothelial growth factor. production is mediated by Alk1/ Smad1 and ERK1/2, but not Smad 2/3, in normal dermal fibroblasts. Together, these data do not support a simple model of CTGF regulating type I pro- Relationship of CTGF’s physiologic received little attention compared collagen expression through direct and pathogenic functions with the extensive literature on the augmentation of the TGF-β/Smad2/3 The relationship of CTGF’s patho role and association of CTGF in dis- pathway and are consistent with the physiologic actions to its normal ease states involving fibrosis and, idea that different mechanisms oper- physiological functions is unknown; more recently, cancer (Leask et al., ate in human dermal fibroblasts and therefore, a better understanding of 2009; Cicha and Goppelt-Struebe, lung epithelial cells. The CTGF– CTGF’s biological significance for 2009; Abraham, 2008). To summa- vascular endothelial growth factor normal and aging skin will be gained rize, CTGF expression and protein (VEGF) relationship and its impact by collecting mechanistic informa- are elevated in nearly all fibrotic tis- on angiogenesis further exemplify tion on both activities. CTGF is essen- sues examined. Exceptions appear to the potential complexity and con- tial in embryonic development, in involve inactive disease processes in text dependence of CTGF action. In which it is important for normal car- which affected tissue has progressed this case, over and above all the pos- tilage and bone formation. In adults, to inert scar tissue. Increased CTGF sible physical interactions, CTGF and however, normal physiologic CTGF concentrations in blood and other bio- VEGF can each induce expression of levels appear dispensable because logical fluids have also been reported the other. Cicha and Goppelt-Struebe, animal studies testing CTGF block- for many fibrotic diseases. As a result, (2009) give an excellent review of ade via CTGF antisense or anti-CTGF there is active interest in CTGF’s util- CTGF and angiogenesis, detailing antibody have not reported adverse ity as a diagnostic and prognostic bio- evidence for pro-, anti-, and neutral effects (Guha et al., 2007; Ponticos marker (Abraham, 2008; Leask et al., effects of CTGF and emphasizing et al., 2009), and mice carrying a 2009; Gressner and Gressner, 2009). that microenvironment strongly influ- single functional CTGF allele exhibit Associations between CTGF and can- ences outcome. Because translation no obvious phenotype. Furthermore, cer are less consistent, with both posi- of in vitro findings to whole-organism results of clinical trials using a human tive and negative correlations reported physiology has limitations, bet- anti-CTGF monoclonal antibody, (Chu et al., 2008; Cicha and Goppelt- ter insight into context dependence FG-3019, show that FG-3019 has Struebe, 2009). This variation could of CTGF–VEGF relationships might been well tolerated in all studies to reflect a context-dependent action of be provided by evaluation of CTGF date. The adult function of CTGF has CTGF, as described below. in clinical studies of therapeutics

www.jidonline.org 339 commentary

targeting VEGF. In summary, CTGF as partners. Hence, if aging is considered include (i) feedback from other CCN gene a matricellular protein may normally pathogenic—and our thinking certainly family members and (ii) epigenetic modi- integrate and communicate extra comes and goes on this topic—reduced fication affecting CTGF gene expression. cellular matrix (ECM)-transmitted CTGF production could be said to cre- Very little is known about the influence molecular signals, thereby coordinating ate an environment that permits inap- of the other CCN gene family proteins cellular function. Besides the possible propriate responses. and whether they affect the interactions cell- or tissue-specific mechanisms, Data consistent with the idea that between CTGF and its binding partners its numerous binding partners provide regulated CTGF levels are important or alter CTGF gene expression. Quan et opportunities for context-dependent for tissue homeostasis were reported in al. (2010) report that CTGF knockdown CTGF action because local factor con- a recent publication by Nguyen et al. in normal adult dermal fibroblasts had tent and availability, as well as CTGF (2008). They found that recombinant no effect on any of the other five CCN content and form, will contribute to the human CTGF injected intraperitone- gene family mRNA levels, and they resultant interaction. ally in mice was first briefly detected in point out that this indicates that expres- the circulation, followed by evidence sion of other CCN family members is not of reduced BMP7 signaling in the kid- dependent on or controlled by CTGF, at CTGF is a key player ney. This outcome was not surprising least not in dermal fibroblasts. However, in the declining because CTGF–BMP4 interaction has this may not accurately reflect the situa- been reported to block BMP4 from tion in vivo. There is evidence for coor- collagen content of binding to its receptor and triggering dinate up- or downregulation of CTGF |aging skin. signaling in lung epithelial cells (Abreu and Cyr61 in primary gliomas (Xie et et al., 2002), but it confirmed a similar al., 2004) and lung cancers (Chen et relationship for CTGF–BMP7 interac- al., 2007), whereas in animal models of In fibrotic disease, dysregulated tion and, importantly, demonstrated renal fibrotic disease, CTGF is elevated CTGF expression results in abnor- that intact biological systems can be and Cyr61 is reduced (Sawai et al., 2007; mally elevated CTGF content that in sensitive to changes in CTGF levels, Mason, 2009). Epigenetic changes affect- turn could favor aberrant interaction even when these occur in different tis- ing CTGF gene expression may also dynamics between CTGF and its bind- sue compartments. This finding also explain reduced CTGF in aging skin. ing partners. This situation could create emphasizes the need to understand There is emerging evidence for this idea, an environment where the fibrogenic, details of CTGF metabolism in vivo. especially as it applies to the variability inflammatory, or angiogenic activity If suitable models were identified, it of CTGF and cancer associations; this of other factors that are usually neu- would be interesting to test whether this topic is a focus of a recent review (Cicha tralized or accumulated are instead approach could restore young skin’s and Goppelt-Struebe, 2009). Given enhanced, promoting inappropriate type I procollagen expression levels to the link between epigenetic effects and responses or outcomes. Similarities aged skin. aging, it seems rational to explore this could be expected for the opposite In addition to binding partners, area further. situation, in which abnormally low microenvironment, and metabolism, levels of CTGF could lead to aberrant mechanisms that might contribute to CTGF sources, forms, and research tools interaction dynamics with its binding reduced CTGF expression in aging skin Cellular sources of CTGF have been determined to include multiple mesen- chymal and epithelial cell types as well as N-terminus C-terminus Protease- Domain 3: vascular smooth muscle and endothelial Domain 2: sensitive Thrombospondin Von Willebrand factor “hinge” type 1 repeat cells. CTGF appears to be constitutively Domain 1: type C repeat Domain 4: IGF-binding protein “Cys knot” present at low levels in adult tissues domain Integrin v 3 6 1 Integrin Wnt and biological fluids, and elevated IGF 1/2 Fibronectin BMP 4, 7 TGF- 1, 2 LRP 1 VEGF 165 HSPG LRP 5/6 expression is thought to be restricted to Integrins Fibronectin FRZ processes involving injury, tissue repair, and the generation of connective tissue. BMP target Focal adhesion and -Catenin genes cytoskeletal remodeling Tissue content could derive from local TGF- target Myofibroblast genes production or deposition from the circu- activation Wnt target (e.g., CTGF) Cell polarization genes lation. Thus, evaluation of tissue CTGF and motility protein is the best way to determine the ECM deposition Angiogenesis and vascular functional importance of CTGF at a par- permeability Cell adhesion Morpho- ECM remodeling and migration genesis ticular site. Currently, the details of CTGF trafficking and metabolism that are need- Figure 2. CTGF domain structure, reported binding partners, and signaling effects. CTGF is a multimodular matricellular protein that binds, modulates signaling, and coordinates activities of ed to understand the sources of CTGF biologically active factors. BMP, bone morphogenetic protein; Cys, cystine; ECM, extracellular matrix; are lacking. HSPG, heparan sulfate proteoglycan; IGF, insulin-like growth factor; LRP, lipoprotein receptor–related Quan et al. (2010) report that nor- protein; TGF-β, transforming growth factor-β; VEGF, vascular endothelial growth factor. mal skin CTGF immunolocalization,

340 Journal of Investigative Dermatology (2010), Volume 130 commentary

primarily to vascular and interstitial CTGF and the other is an Escherichia biomarker or molecular target for anti-fibrotic cells, is consistent with expectations; coli–derived human CTGF domain 4 therapy in SSc? Rheumatology 47(Suppl 5):v8–9 however, the authors point out that (11.2 kDa). The discrepant findings on Abreu JG, Ketpura NI, Reversade B et al. (2002) localization appears to be intracellular, CTGF’s role in inflammation could rep- Connective-tissue growth factor (CTGF) modulates cell signaling by BMP and TGF-β. possibly because the detecting antibody resent context-dependent differences Nat Cell Biol 4:599–604 recognized the CTGF C-terminal half. in the experimental systems used, but Chen Y, Segarini P, Raoufi F et al. (2001) Connective As expected for a matricellular pro- reagent differences deserve consid- tissue growth factor is secreted through the tein, newly synthesized CTGF has been eration as well. It is possible that this Golgi and degraded in the endosome. Exp Cell reported to be quantitatively secreted activity is fragment specific, a situa- Res 271:109–17 in cell culture (Chen et al., 2001), so a tion for which there is much biological Chen P-P, Li W-J, Wang Y et al. (2007) Expression of Cyr61, CTGF, and WISP-1 correlates with lack of extracellular staining in skin is precedent (e.g., endostatin). Indeed, an clinical features of lung cancer. PLoS One surprising. Secreted CTGF is likely to be obvious complicating outcome is the 2:e534 complexed with macromolecular bind- confusion that can result when some Chu CY, Chang CC, Prakash E et al. (2008) ing partners, including proteoglycans investigators use fragments and others Connective tissue growth factor (CTGF) and and glycosaminoglycans in the ECM. use full-length CTGF. cancer progression. J Biomed Sci 15:675–85 Therefore, extracellular CTGF detec- Cicha I, Goppelt-Struebe M (2009) Connective tissue growth factor: context-dependent functions and tion may require “unmasking” of CTGF Concluding remarks mechanisms of regulation. BioFactors 35:200–8 epitopes. Alternatively, because CTGF The results reported by Quan et al. Gressner OA, Gressner AM (2008) Connective is readily cleaved at the central hinge (2010) provide new insight into skin tissue growth factor: a fibrogenic master switch to N- and C-terminal half fragments aging and identify CTGF as a key play- in fibrotic liver diseases. Liver Int 28:1065–79 and the C-terminal half fragment is er for the decline in collagen content. Guha M, Xu Z-G, Tung D et al. (2007) Specific internalized (Chen et al., 2001), ECM- The findings confirm the complexity of down-regulation of connective tissue growth factor attenuates progression of nephropathy in sequestered CTGF might not be the mechanisms whereby CTGF and TGF-β mouse models of type 1 and type 2 diabetes. full-length molecule and therefore may collaborate and act independently and FASEB J 21:3355–68 not be recognized by the detecting anti- raise new questions regarding these Leask A, Parapuram SK, Xu S-W et al. (2009) body used by Quan et al. Quan et al. pathways and how they are affected Connective tissue growth factor (CTGF, CCN2) report that dermal fibroblasts expressed by aging. Although opposite effects on gene regulation: a potent clinical bio-marker of fibroproliferative disease? J Cell Commun Signal primarily 38- and 35-kDa forms of regulation of the TGF-β/CTGF/collagen 3:89–94 CTGF, but that other small CTGF frag- axis in aging and fibrosis are not surpris- Mason RM (2009) Connective tissue growth ments were also detected, indicating ing, the contrast offers an opportunity factor(CCN2), a pathogenic factor in diabetic that the major cellular source in skin is to define the underlying mechanisms. nephropathy. What does it do? How does it do capable of CTGF fragment generation. Identifying the overlapping and unique it? J Cell Commun Signal 3:95-104 Thus, immunohistochemical evaluation components that drive these oppos- Nguyen TQ, Roestenberg P, van Nieuwenhoven FA et al. (2008) CTGF inhibits BMP-7 signalling using a C-terminal reactive antibody to ing processes could provide insight for in diabetic nephropathy. J Am Soc Nephrol detect CTGF is likely to represent main- reinstating balance and thus strategies 19:2098–107 ly newly synthesized or endocytosed to control or reverse fibrosis, as well Pannu J, Nakerakanti S, Smith E et al. (2007) CTGF and, potentially, an incomplete as to improve properties of aging skin. Transforming growth factor-beta receptor picture of abundance and localization Undoubtedly, the research on CTGF’s type I-dependent fibrogenic gene program is mediated via activation of Smad1 and ERK1/2 in the skin. physiological significance will build pathways. J Biol Chem 282:10405–13 Although context-dependent action on the results reported by Quan et al., Ponticos M, Holmes AM, Xu S-W et al. (2009) of CTGF could explain discrepant find- with the potential for beneficial appli- Pivotal role of connective tissue growth factor in ings in the literature on CTGF activity cations of CTGF and its regulatory pro- lung fibrosis: MAPK-dependent transcriptional and functional significance, more triv- gram. In addition to improving aging activation of type I collagen. Arthritis Rheum ial, but important, explanations should skin, the results might have implica- 60:2142–55 also be considered—specifically, tions for some disorders of fragile skin Quan TH, Shao Y, He T et al. (2010) Reduced expression of connective tissue growth factor reagent suitability and properties. Two (or bone) and impaired wound healing; (CTGF/CCN2) mediates collagen loss in recent publications reporting the effects or, perhaps one should even consider chronologically aged human skin. J Invest of CTGF on inflammation illustrate the what CTGF might do for wrinkles. Dermatol 130:415–24 importance of using comparable CTGF Sanchez-Lopez E, Rayego S, Rodrigues-Dıez R et CONFLICT OF INTEREST materials. One study reports that CTGF al. (2009) CTGF promotes inflammatory cell Noelynn Oliver has been an employee and infiltration of the renal interstitium by activating has no effect on inflammation (Ponticos consultant for FibroGen, Inc. Mark Sternlicht is an NF-kappa B. J Am Soc Nephrol 20:1513–26 employee at FibroGen, Inc. Roel Goldschmeding et al., 2009); the other paper shows that Sawai K, Mukoyama M, Mori K et al. (2007) has received research funding from FibroGen, CTGF stimulates NF-κB signaling, con- Expression of CCN1 (CYR61) in developing, Inc., and has been an employee and consultant. normal, and diseased human kidney. Am J sistent with CTGF’s promoting inflam- Karin Gerritsen states no conflict of interest. mation (Sanchez-Lopez et al. 2009). Physiol Renal Physiol 293:F1363–72 The CTGFs used in these studies differ Xie D, Yin D, Wang H-J et al. (2004) Levels of REFERENCES expression of CYR61 and CTGF are prognostic significantly: one is a flag-tagged, mam- Abraham D (2008) Connective tissue growth factor: for tumor progression and survival of individuals malian cell–derived full-length human growth factor, matricellular organizer, fibrotic with gliomas. Clin Cancer Res 10:2072–81