Bone morphogenetic protein-7 (BMP7) exerted potent anti-fibrotic actions in vitro
[30], and in pre-clinical fibrosis models
[31][32][33][34]. BMP7 activation of Smad1/5/8 results in the competitive binding of Smad4, thereby inhibiting the TGF-β1/Smad2/3 pathway’s transcriptional activities
[34]. Early reports indicated that BMP7 failed to prevent bleomycin-induced skin fibrosis
[35]. However, a later study showed that BMP7 treatment at the point of thermal injury prevented hypertrophic scar formation
[34]. Thus, BMP7 therapies require further optimization and testing to determine their effectiveness in the prevention of dermal fibrosis. Interferon (IFN)-γ induced the expression of Smad7, which antagonised interactions between Smad2/3 and TGFβRI/II
[36]. The effects of IFN-γ on fibroblasts in vitro are conflicting; exogenous IFN-γ was suggested to abrogate TGF-β1-induced proliferation, migration, and differentiation to myofibroblasts
[37]; but T-cell secretion of IFN-γ, found to be prominent in SSc, aggravated fibrosis by promoting fibroblast proliferation and collagen synthesis
[38]. Early pre-clinical and small clinical studies suggested that IFN-γ was effective in the attenuation of lung, renal, and liver fibrosis
[39][40][41], and oral submucosal fibrosis
[39][40][41][42]. There is a current lack of studies describing definitive IFN-γ mechanisms involved in the regulation of myofibroblasts. Thus, whether IFN-γ delivery would be effective in attenuating dermal fibrosis remains unclear.
Alternative approaches include molecular inhibitors of non-canonical and mechanotransduction pathways to regulate ECM composition and disrupt the myofibroblast phenotype. Hepatocyte growth factor (HGF) was first implicated in liver regeneration
[43], but has since been shown to prevent fibrosis initiation and progression in animal models
[44][45]. HGF treatment was shown to attenuate collagen production by fibroblasts in multiple tissues
[46][47][48]. Elevated HGF expression in fibrosis
[49][50][51][52], but also in differentiation-resistant fibroblasts
[53][54], suggests a duality of HGF actions which may be a consequence of truncated isoforms possessing a variable number of kringle domains (HGF/NK1-4) with differential signalling activities. Indeed, human oral mucosal fibroblast resistance to TGF-β1-induced differentiation was dependent on heightened HGF and HGF/NK1 expression, whereas HGF/NK2 was preferentially expressed by dermal fibroblasts
[55]. Recently, HGF/NK1 gene therapy was shown to exert potent anti-fibrotic effects through attenuated collagen types I, III, and IV deposition in mouse models of renal fibrosis
[32]. In addition, the HGF inhibition of collagen synthesis and promotion of MMP production has shown promise for therapeutic applications in reducing the pro-fibrotic activity of keloid fibroblasts in vitro and in a keloid heterograft mouse model
[56][57][58]. Small molecule inhibitors of DDR1 have shown promise in reducing collagen types I and IV deposition in bleomycin-induced renal and lung fibrosis
[59][60][61]. Ongoing research into DDR1-specific inhibitors have yielded promising results in models of renal fibrosis
[60].
HA bioactivity is dependent on its molecular weight, enzymatic synthesis, and endogenous versus exogenous application
[62][63][64]. The disruption or prevention of HA pericellular coat synthesis using hyaluronidase enzymes
[65], HAS inhibitors
[66][67] or exogenous HA oligosaccharides
[68] results in the failure of TGF-β1-stimulated myofibroblast phenotypic acquisition. The inhibition of HAS2 activity
[69] or global HAS synthesis of HA
[66][67] have demonstrated preventative effects in various models of fibrosis. However, given the ubiquitous role of HAS-synthesised HA in tissues, and the lack of specific HAS isoenzyme inhibitors, it is unclear at present whether the inhibition of HA synthesis would be beneficial in a clinical setting. An additional action of BMP7 involved the induction of the nuclear translocation of HYAL2 and the subsequent splicing of CD44 mRNA, which resulted in the upregulated expression of the variant isoform, CD44v7/8. Fibroblasts with upregulated expression of CD44v7/8 exhibited ‘HA-phage’ activity, wherein the HA pericellular coat was rapidly internalised and broken down, resulting in the destabilization of the myofibroblast and phenotypic reversion to fibroblasts
[30][70]. Whether CD44v7/8-dependent actions convey anti-fibrotic effects in dermal fibrosis is currently unknown.
The most promising integrin targets that have demonstrated roles in the pathogenesis of fibrosis are α4-containing integrins (α4β1/7)
[71][72] and αv-containing integrins (αvβ1/3/5/6)
[73][74][75]. Mechanotransduction by integrin αvβ6 promoted traction from proliferating liver cholangiocytes to FN/LAP, which subsequently released active TGF-β1 and initiated the differentiation of surrounding hepatic stellate cells to myofibroblasts
[73][76]. A small peptidomimetic, EMD527040, mimics the RGD-binding sites of αvβ6 and αvβ1; orally administered EMD527040 attenuated biliary and non-biliary fibrogenesis
[73]. The expression of integrin αvβ6 was elevated in keratinocytes during wound healing and fibrosis
[77], but the detailed mechanistic roles of integrin αvβ6 and epithelial–mesenchymal crosstalk during dermal fibrosis have yet to be elucidated. Research has identified that integrin interactions with the EDGIHEL motif of EDA–FN was causative of downstream profibrotic responses
[72][78][79]. The IST-9
[78][80], F8
[81][82], and vaccine-generated
[83] antibodies that target EDA–FN or integrin α4 have demonstrated prevention of fibroblast–myofibroblast differentiation
[78][80]. Therapeutic applications of antibodies can be limited by the cost and complexity of production, off-target immune activation or unspecific protein masking of critical protein–protein interactions. To address these issues, Zhang et al. developed a small blocking polypeptide to bind and block EDA–FN interactions with the α4β1 binding cleft, with high specificity
[84]. The polypeptide, AF38Pep, was designed to mimic the integrin α4β1 receptor site for the EDGIHEL motif of EDA–FN and was shown to interfere with TGF-β1-stimulated fibroblast–myofibroblast formation by the specific blockade of integrin α4β1 signalling, the inhibition of FAK activation, and the prevention of profibrotic gene transcription
[84]. The aforementioned first generation of small blocking polypeptides have revealed the renewed promise of the specific interruption of integrin-mediated mechanotransduction and myofibroblast formation. The implementation of integrin receptor peptidomimetics for the prevention of skin fibrosis will become more apparent in the coming years when the in vitro research progresses into pre-clinical models.
Another potential peptide therapy is the N-terminal amino acid sequence of α-SMA, Ac-EEED, which is important for the tropomyosin-1.6/7-stabilized incorporation of α-SMA into cytoplasmic stress fibres
[85]. Interestingly, the cytoplasmic delivery of the Ac-EEED peptide resulted in the loss of α-SMA from β-cytoplasmic actin stress fibres and inhibited G-actin polymerization into F-actin
[86][87], thereby reducing myofibroblast contraction in wound healing
[88]. Ac-EEED peptide therapy has yet to be evaluated in skin fibrosis models, which may be related to the current lack of available dermal myofibroblast-specific targeting moieties that operate by endosomal uptake and the cytoplasmic release of payloads.
3. Immunomodulating Biomolecules for Fibrosis Attenuation
Immunoregulatory interventions aim to control the inflammatory phase in postnatal skin healing to attenuate scar formation
[89][90]. Studies in transgenic mice provided early indicators that the absence of neutrophils and macrophages led to scar-free healing
[91]. Targeted repression of the gap junction and inflammatory mediator protein, connexin-43, supported these findings
[90][92]. Additionally, connexin-43 was shown to mediate cardiac fibroblast–myofibroblast differentiation
[93] and promote aberrant cardiomyocyte–myofibroblast functional coupling
[94]. Certain interleukin (IL) cytokines are implicated in the activation of inflammatory cascades. IL-8 production is a chemoattractant to neutrophils, whereas IL-6 secretion by fibroblasts activates macrophages and monocyte chemotaxis. Both IL-6 and IL-8 exhibit the rapid induction of expression following tissue injury, resulting in the recruitment of circulating inflammatory cells. The expression levels of IL-6 and IL-8 are elevated and maintained for longer in adult skin, compared to scarless scar-free foetal repair. The inhibition of phosphodiesterase 4 (PDE4) reduced scar formation in skin fibrosis models by interfering with the release of IL-6 from M2 macrophages
[95]. Thus, IL-6 and IL-8 are considered pro-fibrotic mediators, whereas IL-10 antagonised their activity
[89][96]. IL-10 gene therapy resulted in reduced inflammation and the promotion of scarless healing in mouse wound healing studies
[97]. The exogenous addition and macrophage paracrine production of IL-10 were shown to induce myofibroblast reversal to fibroblastic phenotypes in vitro
[98][99]. More recently, IL-10 induced myofibroblast–fibroblast dedifferentiation was shown to alter dynamic interactions with the surrounding fibrillar matrix with a demonstratable loss of contractility in IL-10 treated myofibroblasts
[100]. Research into inflammation-induced fibrosis has revealed additional potential candidate ILs that may be targetable in skin fibrosis, including IL-11
[101], IL-16
[102], and IL-33
[103]. Keratinocytes have suggested roles in the regulation of the myofibroblast phenotype and profibrotic ECM during wound healing
[104][105][106]. More recently, the keratinocyte secretion of IL-1α was demonstrated to restrain the myofibroblast phenotype, dependent on fibroblast integrin α4β1 expression and Cox-2/Nrf2 signalling
[107][108].
The importance and therapeutic potential of macrophages in the wound healing process has been highlighted in recent years
[109][110][111]. Crosstalk between myofibroblasts and macrophages during skin repair was reported
[112]. Growth factors secreted by CD301b
+ M2-type macrophages were shown to selectively stimulate the proliferation of adipocyte precursor (AP)-derived myofibroblasts only. In aged mice wounds and experimentally induced mouse skin fibrosis, AP-derived myofibroblasts and CD301b
+ macrophages were reduced, and a CD29
+ myofibroblast pool was increased. In keloids, CD301b
+ macrophages and AP-derived myofibroblasts were also increased
[113]. In fibrotic lung tissues, cadherin-11 mediated the adhesion between macrophages and myofibroblasts, promoting pro-fibrotic myofibroblast activity via the paracrine release of TGF-β1 by the macrophages
[114]. Experimental research has suggested roles of mast cells in scar formation
[115]; reduced numbers of activated mast cells were reported to improve healing and minimize scarring
[116][117][118], whereas mast cell hyperplasia was causally linked to myofibroblast hyperplasia
[119][120][121]. Increased neuropeptide activity and the presence of substance-P (SP) were found in hypertrophic scar samples
[122][123]. Mast cells were previously suggested to be a major source of neuropeptide SP-stimulated inflammation and increased myofibroblast activity
[124]. These studies showed that myofibroblast activity could be mediated through regulated leukocyte–myofibroblast interactions and suggest that targeting certain leukocyte subpopulations may serve as anti-fibrotic strategies.
Targeting inflammatory meditators has shown promise in alleviating the magnitude of fibrosis in various animal models. Follistatin-like-1 (FSTL1) was found to be elevated in serum from patients with silicosis and in mouse lung fibrosis models. FSTL1-induced IL-1β production by macrophages and positively regulated TGF-β1 signalling in fibroblasts. The inhibition of FSTL1 expression or activity protected against lung injury and fibrosis
[125][126]. Recently, FSTL1 neutralizing antibodies were shown to exhibit potent anti-inflammatory actions, attenuate bleomycin-induced IPF and dermal fibrosis in vivo, and downregulate TGF-β1-driven fibrosis in human skin ex vivo
[127].
The transmembrane serine protease and collagenase, fibroblast activation protein (FAP) is prominently expressed by activated fibroblasts and myofibroblasts during tissue remodelling and fibrosis
[128]. FAP-cleaved collagen binds to the scavenger receptor (SR)-A, recruiting SR-A
+ macrophages to sites of collagen turnover
[129]. The liver expression of FAP in cirrhosis was shown to correlate with the severity of fibrosis but was not exclusively expressed by α-SMA
+ myofibroblasts, suggesting that FAP marks a differentially activated fibroblast state
[130]. Lines of research have started to establish the mechanistic actions of FAP in fibroblast heterogeneity and governance over the pro-fibrotic ECM
[131]. Treatment with anti-FAP antibody reduced collagen type I production by fibro-stenotic intestinal myofibroblasts
[132]. The FAP and dipeptidyl peptidase IV inhibitor, talabostat mesylate (PT100), was used to treat bleomycin-induced IPF murine models. Treatment with PT100 showed anti-fibro-proliferative activity but increased macrophage activation, with no effect on collagen expression
[133]. Therefore, the present scope for specific FAP inhibition in dermal fibrosis models remains unclear, until more mechanistic information is reported.
4. Targeted Myofibroblast Apoptosis
In physiological wound healing, myofibroblasts disappeared following wound closure and resolution
[134][135], predominantly by apoptosis
[136]. Despite the elevated production of reactive oxygen species (ROS) by myofibroblasts during fibrosis, the persistence of TGF-β1 expression, ECM deposition, and accumulative stress-induced FAK promotes pro-survival and anti-apoptotic myofibroblast phenotypes
[137][138]. The susceptibility of myofibroblasts to nitric oxide (NO)-induced apoptosis has been reported in vitro
[139]. Therefore, a combination of reduced profibrotic growth factor expression, increased ECM turnover, and increased NO generation may set the stage for triggering myofibroblast apoptosis during the resolution of tissue repair and remodelling
[140][141].
A single chain antibody (C1-3) specifically targets synaptophysin
+ liver myofibroblasts without co-localising with liver monocytes or macrophages
[142], thus demonstrating that the identification of unique markers of myofibroblasts in fibrosis could serve as targeting devices. The researchers showed that C1-3-gliotoxin conjugates induced non-parenchymal cell apoptosis and depleted liver myofibroblasts without affecting monocytes or macrophages, resulting in the reduced severity of fibrosis
[142]. The anticancer drug, Elesclomol, was found to selectively induce apoptosis in activated fibroblasts and myofibroblasts isolated from scar tissue samples. Elesclomol upregulated intracellular levels of ROS, caspase-3, and cytochrome-c proteins, resulting in reduced myofibroblast numbers and a lower scar elevation in vivo
[143].
The mechanical tension-stimulated myofibroblast differentiation increased mitochondrial priming and death signalling proteins, such as the pro-apoptotic BH3-only protein BIM
[144][145]. The anti-apoptotic protein BCL-X
L sequesters BIM and ensures myofibroblast survival. Lagares et al. showed that myofibroblasts were susceptible to apoptosis induced by the BCL-2 inhibitor and the BH3 mimetic drug, ABT-263 (Navitoclax), which inhibited BCL-X
L and allowed BIM to activate myofibroblast apoptosis in mouse models of scleroderma dermal fibrosis
[146]. These results were recapitulated in rabbit ear hypertrophic scar models, wherein ABT-263 improved scar appearance and collagen arrangement
[147]. Future studies into the physiological triggers for time-appropriate myofibroblast apoptosis could potentially lead to the identification of novel treatments with improved therapeutic indexes for scarless wound healing.
5. Antioxidant Therapeutics
Another proposed regulator of normal and pathological scarring in numerous tissues is oxidative stress, referring to the overproduction of ROS via such mechanisms as NADPH oxidases (NOXs) and the mitochondria, at the expense of cellular and tissue antioxidant defences
[148][149][150][151][152]. The induction of the myofibroblast phenotype is accompanied by depleted cellular antioxidants, leading to ROS generation and the implication of ROS in multiple signalling pathways associated with myofibroblast differentiation. Thus, an emergent area of research is focusing on evaluating the efficacious index of antioxidants against fibrosis and in restoring cellular redox balance. The liposomal delivery of copper/zinc (Cu/Zn) superoxide dismutase (SOD)-attenuated TGF-β1, α-SMA and collagen type I expression in dermal fibroblasts, although myofibroblast apoptosis remained unaffected
[153]. Similarly, SOD1-containing fusion proteins alleviated oxidative stress in cardiac myofibroblasts via the reduced expression of TGF-β1, α-SMA, and collagen types I and III—whilst restoring MMP-1 and attenuating MMP inhibitor (TIMP-1) secretion
[154]. Small molecule inhibitors of NOXs, such as GKT136901 and GKT137831, also exhibited therapeutic potential by reducing murine liver fibrosis
[155][156]. In addition to the progress made with such promising findings, the development of nanoparticles with inherent antioxidant activities, such as cerium oxide, fullerene, and mesoporous silica, have also been explored as potential therapeutic options for fibrosis, which may also be combined with payloads of pharmaceuticals, genes, or proteins
[157].
Natural compounds have a long history of use in wound healing, and the cellular mechanisms of action are beginning to be delineated as active components are extracted and assessed
[158][159][160]. The exploitation of the aromatic nature and antioxidant capabilities of various naturally sourced polyphenolic compounds and their extracts have been demonstrated in numerous in vitro and in vivo systems, with desirable biocompatible and anti-fibrotic effects in models of pulmonary
[161][162][163]; renal
[164][165]; myocardial
[166][167]; and skin fibrosis
[168][169]. The clinical potential of natural compound-based therapies is often restricted by the lack of clarity in their multifaceted bioactivities. Hence, delineating cellular responses to the more specific and potent actions of natural compound extracts has become a popularised concept towards their clinical translation.