FIBROSIS: FROM MECHANISMS TO MEDICINES

PREFACE

Fibrosis can affect any organ and is responsible for up to 45% of all deaths in the industrialized world. Long thought to be relentlessly progressive and irreversible, both pre-clinical models and clinical trials in various organ systems have shown that fibrosis is a highly dynamic process. This has clear implications for therapeutic interventions designed to capitalize on this inherent plasticity. However, despite significant progress in our understanding of the pathobiology of fibrosis, a translational gap remains between identification of putative antifibrotic targets and conversion into effective patient therapies. Here we discuss the transformative experimental strategies that are being leveraged to dissect the key cellular and molecular mechanisms regulating fibrosis, and the translational approaches which are enabling the emergence of precision medicine-based therapies for patients with fibrosis.

Fibrosis is not a disease but rather an outcome of the tissue repair response that becomes dysregulated following many types of tissue injury, most notably during chronic inflammatory disorders. The formation of fibrotic tissue, which is defined by the excessive accumulation of extracellular matrix (ECM) components such as collagen and fibronectin, is in fact a normal and important phase of tissue repair in all organs. When tissues are injured, local tissue fibroblasts become activated, increasing their contractility, secretion of inflammatory mediators, and synthesis of ECM components that together initiate the wound healing response. When damage is minor or non-repetitive, the wound healing response is efficient, resulting in only a transient accumulation of excess ECM components that is then quickly eliminated, facilitating the restoration of normal tissue architecture. However, when the injury is repetitive or severe, ECM components continue to accumulate, which can lead to disruption of tissue architecture, organ dysfunction and ultimately organ failure. Genetic influences, aging, the response to invading microbes, and the changing character of the inflammatory response over time influence whether wound healing responses lead to progressive fibrosis or end in efficient repair. In this review, we provide an update on recent research investigating the mechanisms of fibrosis and discuss how this information is enabling the development of novel anti-fibrotic treatments.

FIBROSIS DECONVOLVED BY SINGLE CELL GENOMICS

Single cell multi-omics approaches are transforming our understanding of disease pathogenesis across medicine, allowing interrogation of cell populations in health and disease at unprecedented resolution. This ‘resolution revolution’ allows the powerful, unbiased exploration of cell states and types at single-cell level, resulting in unexpected novel insights into tissue biology and disease mechanisms.

These cutting-edge single-cell approaches have already been avidly adopted by the fibrosis research community to deepen our understanding of the complex, multi-cellular interplay driving lung fibrosis (Figure 1)¹. Mesenchymal cells are the key source of pathological ECM deposition during lung fibrosis, ultimately leading to architectural disruption and reduced lung function. Spatial transcriptional maps of the mouse lung mesenchyme have been generated by combining single-cell RNA sequencing (scRNAseq) and signaling lineage reporters². Each mesenchymal lineage demonstrated a distinct spatial address and transcriptome conferring unique niche regulatory functions. Examples include the mesenchymal alveolar niche cells being Pdgfrα⁺, Wnt responsive, and critical for alveolar epithelial cell growth and self-renewal. In contrast, Axin2⁺ myofibrogenic progenitor cells preferentially generated pathologically deleterious myofibroblasts following lung injury². Further studies using the mouse bleomycin-injury model have also identified lung mesenchymal cell heterogeneity in healthy and fibrotic mouse lungs^3,4.

Figure 1: Deconvolving fibrosis using multi-modal single-cell approaches.

Cutting-edge single-cell approaches are transforming our understanding of the complex cellular and molecular mechanisms regulating fibrosis, allowing assessment of the transcriptome, genome, epigenome and proteome at single-cell level, in addition to spatial profiling. Furthermore, combined readouts from the same single cell are now possible (for example, the simultaneous profiling of transcriptome and chromatin accessibilty), and integration of these multi-modal single-cell omics readouts has allowed ever more powerful, comprehensive assessments of cell state, ontogeny, phenotype and function during human fibrotic disease. The new biological insights gained from these integrated approaches should enable the identification of novel and tractable therapeutic targets to treat patients across a broad range of fibrotic diseases.

Analyzing over 70,000 multi-lineage cells from eight human pulmonary fibrosis lung explants (of varying etiologies) and eight healthy donor lung samples⁵, identified a distinct, novel population of profibrotic alveolar macrophages restricted to patients with fibrosis which had previously been identified in mice^5,6. This study and others^5,7 have suggested a pathological role for alveolar type (AT) 2 cells, which secrete pulmonary surfactant and serve as alveolar stem cells, uncovering a distinct AT2 cell population in fibrotic lungs⁸, and establishing a direct mechanistic link between elevated mechanical tension-activated TGF-β signaling caused by impaired alveolar regeneration and progressive lung fibrosis. Decreasing mechanical tension on alveoli represents a potentially novel therapeutic approach to treat progressive lung fibrosis⁸. Profiling 312,928 cells from 32 IPF, 29 healthy control and 18 chronic obstructive pulmonary disease (COPD) lungs as a disease-control Adams et.al. identified a novel population of IPF-enriched aberrant basaloid epithelial cells located at the edge of myofibroblast foci in the IPF lung⁹. Within the vascular endothelial cell compartment an expanded cell population was identified in IPF samples that is transcriptomically identical to vascular endothelial cells normally restricted to the bronchial circulation. Furthermore, diffusion map and pseudotemporal trajectory analyses (novel computational techniques used in single-cell transcriptomics to determine the pattern of a dynamic process experienced by cells, and then arrange cells based on their progression through the process), also allowed inference of activated IPF myofibroblast origins⁹.

ScRNAseq studies have also comprehensively profiled the cellular and molecular landscape in liver homeostasis and regeneration^10–14. Since their discovery as major collagen-producing cells in the liver¹⁵, hepatic stellate cells (HSC) have been considered a homogenous population, with the potential to transition to the activated, myofibroblast phenotype thought to be equally distributed across all HSCs. ScRNAseq has uncovered functional zonation of murine HSC, allowing high-resolution identification of the critical pathogenic collagen-producing cells following centrilobular liver injury¹⁶. Pseudotemporal trajectory and RNA velocity, another computational approach that predicts the future state of individual cells on a timescale, demonstrated that central vein-associated HSCs are the dominant source of pathogenic collagen-producing cells following centrilobular liver injury¹⁶. Furthermore, using scRNAseq to interrogate retinol-positive myofibroblasts isolated from fibrotic mouse liver has also demonstrated heterogeneity and functional diversity within liver myofibroblasts¹⁷.

Profiling of more than 100,000 single human liver cells yielded molecular definitions for non-parenchymal cell types found in healthy and cirrhotic human liver, allowing identification of a novel scar-associated triggering receptor expressed on a myeloid cell (TREM)2⁺CD9⁺ subpopulation of macrophages which expands in liver fibrosis, differentiates from circulating monocytes and is pro-fibrogenic¹⁸. Disease-associated atypical chemokine receptor (ACKR1)⁺ and plasmalemma vesicle-associated protein (PLVAP)⁺ endothelial cell subpopulations which are topographically restricted to the fibrotic niche and enhance the transmigration of leucocytes were also defined. Using multi-lineage modelling^19,20 of ligand-receptor interactions between the scar-associated macrophages, endothelial cells and PDGFRα⁺ collagen-producing mesenchymal cells revealed intra-scar activity of several pro-fibrogenic pathways including TNF receptor superfamily (TNFRSF) 12A, PDGFR and NOTCH signalling, providing a conceptual framework for the discovery of rational therapeutic targets in liver cirrhosis¹⁸. Of note, murine liver injury-associated macrophages show significant overlap of marker genes with those observed for human scar-associated macrophages, with TREM2 and CD9 demonstrating conservation across species¹⁸. Unbiased cross-species mapping of scRNAseq data using canonical correlation analysis (CCA)²¹ confirmed that murine and human scar-associated macrophages represent corollary populations. This demonstrates the utility of scRNAseq approaches in defining ‘core’ fibrotic injury-induced populations and therapeutic targets across species, thereby increasing precision in the interrogation of putative targets across the translational pipeline, from pre-clinical rodent models to human liver primary cell / organoid-based systems.

In the gastrointestinal tract²², a single-cell census of the colonic mesenchyme revealed four subsets of fibroblasts in addition to pericytes and myofibroblasts, and identification of a fibroblast subpopulation proximal to the colonic crypt niche which expressed SOX6, F3 (CD142), and WNT genes essential for colonic epithelial stem cell function. In colitis, this niche became dysregulated with emergence of an activated mesenchymal population expressing TNFSF14, fibroblastic reticular cell-associated genes, interleukin (IL)-33, and lysyl oxidases (LOX) leading to impaired epithelial proliferation and maturation, thus illustrating how the colonic mesenchyme remodels to drive inflammation and barrier dysfunction in inflammatory bowel disease (IBD)²². In the context of arthritis, deletion of fibroblast activation protein-α (FAPα)⁺ fibroblasts suppressed inflammation and bone erosions in mouse models of resolving and persistent arthritis, and scRNASeq identified two anatomically distinct fibroblast subsets within the FAPα⁺ population: FAPα⁺ thymus cell antigen (THY1)⁺ immune ‘effector’ fibroblasts located in the synovial sub-lining, and FAPα⁺THY1⁻ ‘destructive’ fibroblasts restricted to the synovial lining layer. Adoptive transfer of FAPα⁺THY1⁻ fibroblasts into the joint selectively mediated bone and cartilage damage with little effect on inflammation, whereas transfer of FAPα⁺ THY1⁺ fibroblasts resulted in a more severe and persistent inflammatory arthritis, with minimal effect on bone and cartilage. Discovery of these anatomically discrete, functionally distinct fibroblast subsets with non-overlapping functions clearly has important implications in the rational design of therapies aimed at precisely modulating inflammation, fibrosis and tissue repair²³.

FIBROBLAST HETEROGENEITY AND PLASTICITY

Functional fibroblast heterogeneity

Increasingly sophisticated experimental approaches have allowed discovery of significant diversity and functional heterogeneity within the fibroblast population during organ fibrosis^22–25 . A combination of fate mapping and live imaging, revealed that a specialised subset of fibroblasts, fascia fibroblasts, rise to the surface of the skin following wounding²⁶. These fascia fibroblasts gather their surrounding extracellular matrix (including blood vessels, macrophages and peripheral nerves) to form the provisional matrix, and ablation of these fibroblasts inhibited matrix homing into wounds leading to defective scars. Intriguingly, placement of an impermeable film beneath the skin (preventing upward migration of fascia fibroblasts) led to chronic open wounds. Thus, fascia contains a specialized prefabricated kit of sentry fibroblasts, which are embedded within a movable sealant. Whether similar fibroblast subpopulations exist in other organs and use analogous mechanisms to promote wound healing remains to be determined. Significant myofibroblast functional diversity has also been identified during skin injury and ageing²⁷. Lineage tracing and flow cytometry identified distinct subsets of wound bed myofibroblasts, including CD26-expressing adipocyte precursors and also a CD29^High subpopulation. Adipocyte precursors were significantly reduced and CD29^High cells more abundant in wound beds from aged mice and bleomycin-induced fibrotic mouse skin, suggesting that the fibrotic microenvironment impacts myofibroblast composition.

Recent studies have implicated a range of mesenchymal progenitor cells (MPs) in the initiation and propagation of fibrosis^28–30. In particular two studies, focussing on populations of hypermethylated in cancer protein (Hic1)⁺Pdgfrα⁺ lymphocyte antigen 6 complex (LY6A)⁺ MP populations in heart and skeletal muscle, have elegantly demonstrated MP hierarchy, diversity in the pathophysiological roles of their progeny, and how fate determination of MP is context dependent^31,32. Conditional genetic inactivation of Hic1 in mice led to activation and expansion of MPs in both heart and skeletal muscle, demonstrating that Hic1 is required for maintaining MP quiescence. In the heart, Hic1 deficiency (in Pdgfrα-expressing cells) led to activation of MPs and accumulation of cardiac fibroadipogenic progenitor cells, with epicardial thickening, interstitial fibrosis and fibrofatty depositions resulting in major pathological features pathognomonic in arrhythmogenic cardiomyopathy³². However, although Hic1 inactivation in skeletal muscle also resulted in a marked expansion of Pdgfrα⁺LY6A⁺ cells during homeostasis, this did not increase skeletal muscle fibrosis and had no substantial effect on skeletal muscle regeneration³¹, highlighting the diverse pathophysiological roles of these cells in different organs.

Fibrosis, secondary to age-associated chronic low-grade inflammation, is increasingly recognised as an important cause of morbidity and mortality³³. Increasing age is a driver of fibroblast functional heterogeneity, with ‘old’ fibroblasts demonstrating variability in their ability to reprogram and heal wounds³⁴ . Old mice exhibited variability in wound healing rates in vivo, and scRNASeq identified distinct subpopulations of fibroblasts with differing cytokine expression profiles in the wounds of old mice with slow versus fast healing rates. This increased variability in wound healing with increasing age may reflect distinct stochastic ageing trajectories between individuals, which will need to be considered when designing personalized antifibrotic therapies for the elderly population³⁴.

Fibroblast reprogramming

Fibroblasts, in addition to displaying significant functional heterogeneity, are capable of remarkable degrees of plasticity and phenotype switching during progression and regression of fibrosis^30,35,36. For example, the transcription factor PU.1, whose expression is silenced in fibroblasts during tissue homeostasis but upregulated in fibrotic disease, has been shown to play a major role in fibroblast polarization and fibrogenesis³⁷. PU.1 both polarized resting fibroblasts and repolarized ECM-degrading inflammatory fibroblasts to an ECM-producing fibrotic phenotype. Furthermore, genetic and pharmacological inactivation of PU.1 enabled reprogramming of fibrotic fibroblasts into resting fibroblasts, leading to regression of fibrosis in several organs³⁷. In addition, a remarkable inter-lineage plasticity between myofibroblasts and other cell types during fibrosis exists. During cutaneous wound healing in mice, adipocytes were regenerated from myofibroblasts. This reprogramming required neogenic hair follicles, which triggered bone morphogenetic protein (BMP) signaling and activation of adipocyte transcription factors expressed during development³⁸. Furthermore, adipocytes were generated from human keloid fibroblasts when treated with BMP in vitro, or when placed with human hair follicles³⁸. In the liver, viral vector-mediated expression of specific transcription factors in myofibroblasts has been used to reprogram myofibroblasts into hepatocyte-like cells in fibrotic mouse livers thereby reducing liver fibrosis and increasing liver function^39,40. The ability to selectively target scar-producing myofibroblasts during fibrosis and reprogram these cells into other lineages which support organ function opens up exciting new avenues for antifibrotic and pro-regenerative therapies.

Taken together, these studies highlight that profound fibroblast diversity, functional heterogeneity and plasticity exists during fibrosis, both within and between organs (Figure 2). More precise delineation of fibroblast heterogeneity and phenotype in different disease settings should facilitate the design of rational, highly-targeted antifibrotic therapies, ultimately allowing specific inhibition and reprogramming of pathologic fibroblast subpopulations whilst preserving essential, homeostatic fibroblast function. One of the first trials to explore fibroblast cellular therapy was a phase 2 study using allogeneic human dermal fibroblasts to remodel contracted scars (NCT01564407; TABLE 1).

Figure 2: Functional fibroblast heterogeneity and plasticity.

Recent studies have uncovered significant functional heterogeneity and plasticity within fibroblast populations during fibrosis. In the context of arthritis, scRNASeq combined with adoptive transfer experiments identified two anatomically distinct fibroblast subsets within the FAPα⁺ population: FAPα⁺ thymus cell antigen (THY1)⁺ immune ‘effector’ fibroblasts located in the synovial sub-lining, and FAPα⁺THY1⁻ ‘destructive’ fibroblasts restricted to the synovial lining layer. Studies of mesenchymal progenitor (MP) populations (Hic1⁺Pdgfrα⁺ LY6A⁺) in heart and skeletal muscle have demonstrated MP hierarchy, diversity in the pathophysiological roles of their progeny, and how fate determination of MP is tissue-dependent. Fibroblasts are also capable of remarkable degrees of plasticity and phenotype switching during progression and regression of fibrosis. Myofibroblasts can revert to a quiescent state in the absence of ongoing injury, or may undergo full lineage switching with adipocytes which is observed during cutaneous wound healing in mice. Furthermore, genetic and pharmacological inactivation of the transcription factor PU.1 can reprogramme fibrotic fibroblasts into resting fibroblasts, resulting in regression of fibrosis in several organs. Finally, viral vector-mediated expression of specific transcription factors in myofibroblasts in the liver has been used to reprogram myofibroblasts into hepatocyte-like cells in fibrotic mouse livers, thereby reducing liver fibrosis and increasing liver function.

Table 1.

Drugs tested in phase 2 or phase 3 clinical trials for antifibrotic therapy with relevance to this review article.

Target	Type of agent	Drug ID	Manufacturer	Disease	Phase	NCT identifier
Anti αvβ6	Small molecule	GSK3008348	Glaxo Smith Kline	IPF	2	03069989
Anti αvβ6	Antibody	STX-100	Biogen	IPF	2	01371305
Anti αvβ1/αvβ6	Small molecule	PLN-74809	Pliant	IPF	2	04072315
Anti αvβ1/αvβ3/αvβ6	Small molecule	IDL-2965	Indalo	IPF	2	03949530
Anti-IL-6	Antibody	Tocilizumab	Genentech	Scleroderma	3	02453256
Anti IL-13	Antibody	QAX576	Novartis	IPF, skin keloids	2	01266135; 0987545
Anti IL-13	Antibody	Tralokinumab	AstraZeneca	IPF	2	01629667
Anti IL-13	Antibody	Lebrikizumab	Roche	IPF	2	01872689
Anti IL-4 / Anti IL-13	Antibody	SAR156597	Sanofi	IPF, scleroderma	2	01529853
Anti CCL2	Antibody	Carlumab	Centocor	IPF	2	0786201
Anti CCR2 / Anti CCR5	Antibody	Cenicriviroc	Allergan	NASH, liver cirrhosis	3	02217475; 03028740; 03059446
Anti TLR4	Antibody	NI-0101	Light Chain Bioscience	RA	2	03241108
Pentraxin 2	Recombinant protein	PRM-151	Promedior	IPF, myelofibrosis	2	02550873; 01981850
Allogeneic dermal fibroblasts	Allogeneic fibroblasts	ICX-RHY-013	Intercytex	Skin scars	2	01564407

Abbreviations: IPF: Idiopathic pulmonary fibrosis; NASH: Non-alcoholic steatohepatitis; RA: Rheumatoid arthritis.

METABOLIC REGULATION OF MESENCHYMAL CELLS

It is now apparent that cells involved in the progression and resolution of fibrosis are metabolically ‘reprogrammed’ to perform distinct functions during tissue repair. The impact of metabolism has been explored extensively in the context of NASH-driven fibrosis, in which dysregulated hepatic lipid metabolism serves as a key driver of liver injury and cirrhosis⁴¹. As discussed earlier, fibroblasts are the key source of ECM deposition during fibrosis. Hence, metabolic alterations of local tissue mesenchymal cells may offer future therapeutic avenues that include the major carbohydrate, amino acid and lipid metabolism pathways (Figure 3).

Figure 3: Metabolomic reprogramming of activated fibroblasts.

Profibrotic fibroblasts increase glycolysis via hexokinase leading to ehanced pyruvate and lactate. Lactate decreases extracellular pH and activates latent TGF-β1. Pyruvate also feeds the TCA cycle after conversion to acetyl-CoA increasing succinate. Both mechanisms lead to increase in a-SMA, collagen production and proliferation. Glutaminase activity is increased converting glutamate to glutamine which gets converted into α-KG via the TCA cycle and decreases apoptosis and enhances collagen stabilization. Abbreviations: α-KG: alpha-ketoglutarate; a-SMA: alpha smooth muscle actin; LDH: Lactate hydrogenase; P: Phosphate; R: Receptor; TCA:tricarboxylic acid; TGF: Transforming growth factor.

Following tissue injury, mesenchymal cells undergo profound metabolic changes in order to facilitate energy consuming cellular functions such as proliferation and protein synthesis⁴². Fibroblast aerobic glycolysis is increased by upregulating rate-limiting glycolytic enzymes⁴³. In addition to providing a rapid energy-generating mechanism compared to oxidative phosphorylation, glycolysis produces by-products such as lactate that are also important in the context of fibrosis. Lowering extracellular pH combined with an increase in lactic acid promotes myofibroblast differentiation via activation of TGF-β1 and lactate itself may serve as an additional energy source promoting a synthetic phenotype in mesenchymal cells^44–46. Fibroblast activation also involves increased activity of key glycolytic enzymes, such as hexokinase 2 and lactate dehydrogenase (LDH)⁴⁷, which in turn increase cell proliferation⁴⁷ and collagen synthesis⁴⁸. During enhanced glycolysis, increased amounts of pyruvate are converted to acetyl-CoA in the mitochondrial matrix⁴⁷ before entering the citric acid cycle. This yields intermediate metabolites, such as succinate, which promote fibrosis⁴⁹. Increased glycolysis has been implicated in experimental lung, liver and kidney fibrosis models and inhibition of glycolysis reduces ECM accumulation^50–52. During fibrogenesis, mesenchymal cells also exploit changes in amino acid metabolism through glutaminolytic reprogramming. Glutaminolysis and levels of the key enzyme glutaminase are increased in TGF-β1 stimulated fibroblasts⁵³. This leads to enhanced conversion from glutamine to glutamate, which confers resistance to apoptosis⁵⁴ and regulates collagen production by promoting its stabilization⁵³ via mTOR signaling. In vivo, inhibition of glutaminase 1 ameliorates bleomycin- and TGF-β1 induced pulmonary fibrosis⁵⁵. As a further mechanism, changes in fatty acid oxidation have also been linked to fibrogenesis. Intracellular fatty acid oxidation is downregulated in tubulointerstitial fibrosis in mice and humans, and its restoration was fibroprotective⁵⁶. Of note, upregulation of glycolysis has been reported as a compensatory mechanism for reduced fatty acid oxidation during kidney injury, which could result in enhanced progression to fibrosis⁵² (Figure 3).

Taken together, key metabolic pathways such as increased glycolysis, upregulation of glutaminolysis and enhanced fatty acid oxidation are emerging as important drivers of fibroblast activation. To date no drugs targeting these metabolic pathways have reached the clinic as antifibrotic therapies. However drugs targeting metabolism pathways are approved or in clinical trials for treatment of cancer with known safety profiles, fueling the hope for future application in fibrotic diseases⁵⁷.

MACROPHAGE-MEDIATED REGULATION OF FIBROSIS

Inflammatory monocytes and tissue-resident macrophages are key regulators of tissue fibrosis, playing major roles in the initiation, maintenance and resolution of tissue injury^58–60. Furthermore, monocytes and macrophages are capable of remarkable functional plasticity, displaying diverse phenotypes during wound healing dependent on multiple cues including the environmental niche^61,62 and the temporal stage of tissue injury and repair^63–66. Tissue macrophages are also important producers of T cell- and fibroblast-recruiting chemokines that orchestrate the development of the fibrotic niche⁶⁷.

Functional Heterogeneity of Monocytes and Macrophages

Several recent studies have identified subpopulations of monocytes and macrophages which have the capacity to regulate fibrosis and tissue remodelling^{18,61,68–71}. For example, a newly identified population of atypical monocytes carcinoembryonic antigen-related adhesion molecules (Ceacam)1⁺Msr1⁺Ly6C⁻F4/80⁻Mac1⁺ monocytes)⁶⁸ termed segregated-nucleus-containing atypical monocytes (SatM) and sharing granulocyte characteristics, have been shown to play a key role in lung fibrogenesis. SatM were found to be regulated by CCAAT/enhancer binding protein β (C/EBPβ), with Cebpb deficiency leading to a complete deficiency of SatM. Bleomycin-induced fibrosis, but not inflammation, was inhibited in chimaeric mice with Cebpb^−/− haematopoietic cells, and adoptive transfer of SatM into Cebpb^−/− mice resulted in fibrosis. Interestingly, SatM were derived from Ly6C⁻FcεRI⁺ granulocyte/macrophage progenitors, but not from macrophage/dendritic cell progenitors. ScRNAseq approaches have also been used to investigate macrophage heterogeneity and function in the context of lung fibrosis. Although the major tissue-resident macrophage populations have been intensively studied, much less is known regarding the role of interstitial macrophages (IMs) during tissue homeostasis and injury. Two independent IM subpopulations that are conserved across lung, fat, heart, and dermis were identified: (Lyve1^loMHCII^hiCX3CR1^hi (Lyve1^loMHCII^hi) and Lyve1^hiMHCII^loCX3CR1^lo (Lyve1^hiMHCII^lo) monocyte-derived IMs. Using a novel mouse model of inducible macrophage depletion (Slco2b1^flox/DTR), demonstrated that the absence of Lyve1^hiMHCII^lo IMs exacerbates experimental lung fibrosis, thereby showing that two independent populations of IMs coexist across tissues with conserved niche-dependent functional programs⁶¹. In addition, a pathological subgroup of transitional macrophages that are required for the fibrotic response to injury was indentified in murine bleomycin-induced lung fibrosis. Using a novel computational approach which allows annotation of scRNAseq by reference to bulk transcriptomes (SingleR) enabled macrophage subclustering and uncovered a disease-associated subpopulation with a transitional gene expression profile intermediate between monocyte-derived and alveolar macrophages. These CX3CR1⁺SiglecF⁺ transitional macrophages localized to the fibrotic niche and were profibrotic in vivo. This has been linked to human disease because human orthologs of genes expressed by these transitional macrophages were upregulated in samples from patients with idiopathic pulmonary fibrosis⁷⁰.

Research examining the regulatory roles of monocytes and macrophages during tissue injury and repair has largely focussed on blood-derived monocytes and macrophages. However, emerging evidence has also identified resident cavity macrophages as key contributors to fibrosis and tissue remodelling. Previous work demonstrated that a reservoir of mature F4/80^HiGata6⁺ peritoneal cavity macrophages rapidly invade the liver via direct (avascular) recruitment across the mesothelium in response to sterile liver injury⁷². These recruited macrophages dismantle necrotic cell nuclei, releasing DNA and forming a cover across the site of injury⁷². Building on this observation, the same group has investigated the contribution of resident cavity macrophages located in the pericardial space adjacent to the site of cardiac injury. Following myocardial infarction in mice, Gata6⁺ macrophages in mouse pericardial fluid invaded the epicardium and lost Gata6 expression but maintained anti-fibrotic properties, and loss of this macrophage population enhanced interstitial fibrosis post-ischemic injury. Gata6⁺ macrophages were also present in human pericardial fluid, suggesting that this immune cardioprotective role for the pericardial tissue compartment may be relevant in human disease⁷³.

Macrophage and fibroblast cross-talk

Irrespective of the mode of monocyte and macrophage recruitment into areas of tissue injury, pro-fibrotic macrophages commonly coordinate scar formation via a range of interactions with fibroblasts, the major cellular source of pathological extracellular matrix deposition during fibrosis^{18,27,74–78}. For example, macrophage-derived amphiregulin (an epidermal growth factor receptor ligand) has recently been shown to induce the differentiation of mesenchymal stromal cells into myofibroblasts via integrin αv-mediated activation of TGF-β⁷⁹. Previous work has shown that proximity is crucial to allow crosstalk between macrophages and contractile fibroblasts^27,77,78, however until recently it remained elusive as to how proximity between these two cell types is established. In an elegant study, contracting fibroblasts were shown to generate deformation fields in fibrillar collagen matrix that provided far-reaching physical cues to macrophages⁸⁰. Within the collagen deformation fields created by fibroblasts or actuated microneedles, macrophages migrated towards the force source from distances of several hundred micrometres, and the presence of a dynamic force source within the matrix was critical to initiate and direct macrophage migration. Interestingly, and disrupting traditional views on how macrophages migrate within fibrotic tissues, the authors proposed that macrophages mechanosense the velocity of local displacements of their substrate, allowing contractile fibroblasts to attract macrophages over distances that exceed the range of chemotactic gradients⁸⁰.

INTEGRIN MEDIATED TGF-BETA ACTIVATION

Secreted TGF-β1 is a major pro-fibrogenic cytokine⁸¹, and therefore potentially represents an attractive anti-fibrotic target. Sustained systemic inhibition of TGF-β1, however, has been shown to have undesired effects including cardiac valve problems, and TGF-β1 knockout mice develop systemic autoimmunity⁸². This has relevance to all mucosal surfaces, especially the intestine, where TGF-β1 activity is believed to control tissue homeostasis^83–85. In addition, pan-TGF-β1 blockade has been found to induce carcinogenesis, perhaps owing to the role of TGF-β1 as an anti-proliferative mediator for most epithelial cell types. Strategies to avoid these deleterious effects could invovle choosing the correct magnitude or duration of inhibition, following the yet to be proven hypothesis that low levels of TGF-β1 control immune homeostasis while high levels mediate fibrosis, co-administering anti-inflammatory therapies, or inhibiting TGF-β1 locally at specific sites in the tissue by blocking integrins and other locally acting mediators that activate latent TGF-β1.

The pericellular fibrotic matrix is a remarkably dynamic environment, which exerts profound influences on cell behavior, and many of the key cell–cell and cell–matrix interactions which regulate fibrosis are mediated by members of the integrin family (noncovalent α/β heterodimers from 18 different α subunits and 8 β subunits resulting in 24 known members in humans)^86,87. Importantly, integrins can mediate the translation of spatially-fixed extracellular signals into a wide variety of changes in cell behavior including cell adhesion, migration, proliferation, differentiation and apoptosis^86,87. Of key relevance to fibrosis, integrins can also potentiate signals from soluble pro-fibrogenic growth factors such as TGF-β1. Nearly all TGF-β1 is secreted and bound to the ECM in a latent form, and therefore the majority of the regulation of TGF-β function during fibrosis is dependent on site-specific regulation of TGF-β activation, rather than synthesis or secretion⁸⁸.

The most intensively studied activation mechanism for TGF-β1 with demonstrated relevance in vivo has been the interaction of the TGF-β1 latent complex with the αv-containing subset of integrins. Specifically, the integrins αvβ1, αvβ3, αvβ5, αvβ6 and αvβ8 have all been shown to bind to an N-terminal fragment of the TGFβ1 gene product called the latency associated peptide (LAP), which normally forms a noncovalent complex (latent complex) with the active cytokine, preventing TGF-β from binding to its cognate receptors and inducing biological effects^89–93. When a mechanical force is applied to the latent complex by contraction of αvβ6 integrin-expressing cells, the resultant conformational change results in release of active TGF-β1^94–96. Interestingly, a recent study has shown that in contrast to this αvβ6-mediated mechanism of TGF-β activation, αvβ8-dependent TGFβ activation can occur independently of actin-cytoskeletal force and does not require release of mature TGF-β⁹⁷, further highlighting the complexity of αv integrin-mediated TGF-β activation and signaling in different contexts.

There is now abundant pre-clinical data across a range of fibrotic disease models demonstrating critical regulatory roles for αv-containing integrins expressed on various different cell lineages. Mice lacking the αvβ6 integrin are protected in mouse models of lung, kidney and biliary fibrosis^90,98–102. This protection appears to be secondary to local inhibition of TGF-β, and antibody-mediated inhibition of α_vβ₆-mediated TGF-β1 activation decreased lung fibrosis in pre-clinical models^99,103. α_vβ₈-dependent TGF-β1 activation represents a further potential therapeutic target^104,105. Conditional depletion of lung fibroblast αvβ8 inhibited experimental airway fibrosis¹⁰⁵, and mice genetically engineered to replace murine β₈ with its human ortholog demonstrated that a blocking antibody against human α_vβ₈ blocks TGF-β1 activation and protects against allergic airway inflammation and remodeling induced by cigarette smoke¹⁰⁶. Furthermore, depletion of the αv integrin subunit on mesenchymal cells has also proven effective in inhibiting fibrosis in models of liver, lung and kidney fibrosis¹⁰⁷. Using Pdgfrβ-Cre mice to deplete αv integrins on hepatic myofibroblasts protected mice from carbon tetrachloride-induced hepatic fibrosis, whereas global loss of β3, β5 or β6 integrins or conditional loss of β8 integrins in myofibroblasts did not, highlighting context dependency of the various αv-containing integrins in the regulation of fibrosis across different organs. Pharmacological blockade of αv-containing integrins by a small molecule (CWHM 12) attenuated both liver and lung fibrosis, even when fibrosis was already established¹⁰⁷. Tissue fibroblasts can express four αv-containing integrins, αvβ1, αvβ3, αvβ5 and αvβ8. However, as the αvβ1 integrin is composed of α and β subunits that are both present in multiple heterodimers, it has been challenging to generate heterodimer-specific antibodies or to deduce function from gene knockout studies. Therefore, small molecule inhibitors of αvβ1 have been developed to help interrogate the role of this integrin in fibrosis, with recent studies demonstrating an anti-fibrotic effect of αvβ1 blockade in models of lung and liver fibrosis⁸⁹.

Given the abundance of pre-clinical data, this remains a very active area of research and development in the fibrosis field, with multiple small molecule and antibody-based approaches undergoing assessment in clinical trials, including inhibitors designed to selectively target multiple αv-containing integrins simultaneously. This includes phase 2 trials inhibiting αvβ6 (NCT01371305), αvβ1/αvβ6 (NCT04072315) and αvβ1/αvβ3/αvβ6 (NCT03949530), all in pulmonary fibrosis (TABLE 1). Patient safety will be an important consideration in these trials as Biogen recently terminated their IPF trial of a selective anti-αvβ6 antibody due to unspecified safety concerns.

CYTOKINE-MEDIATED REGULATION OF FIBROSIS

Other than TGF-β several additional cytokines secreted from multiple cellular sources have been identified as triggers of fibrosis¹⁰⁸. The pro-inflammatory cytokine IL-17A was identified as a key inducer of fibrosis in several different organ systems, including the lung, liver, kidney, heart, and skin^109–114. In a study of bleomycin-induced pulmonary fibrosis, IL-17A produced by γδ and CD4+ T cells induced severe lung inflammation, neutrophil recruitment, and production of the pro-fibrotic cytokine TGF-β¹¹⁰. Neutrophils and mast cells were also identified as important sources of IL-17A¹¹⁵. Experiments with IL-17A- and IL-17RA-deficient mice, as well as therapeutic studies with IL-17A neutralizing monoclonal antibodies confirmed a role for IL-17A signaling in the development of fibrosis in multiple tissues^{110–112,116–118}. In addition to promoting TGF-β production^116,119, IL-17A increases and stabilizes TGF-βRII expression on fibroblasts, thereby enhancing their sensitivity to TGF-β¹²⁰. The Th17-associated cytokine IL-22 similarly enhanced TGF-β signaling in fibroblasts¹¹⁵. TGF-β in turn induced the expression of IL-17A when produced concurrently with the pro-inflammatory cytokines IL-1, IL-6, or TNF^110,121,122, suggesting a feedforward mechanism involving acute phase cytokines, IL-17A, and TGF-β is responsible for the development of fibrosis following acute tissue injury^109,110,123 (Figure 4). Importantly, IL-17A exhibited similar pro-fibrotic activity in both animal studies and in human cells^124,125.

Figure 4: Divergent cytokine pathways drive fibrosis.

The Innate acute phase pro-inflammatory cytokines IL-1, IL-6, and TNF, together with TGF beta promote the development of IL-17 secreting cells. IL-17A potentiates neutrophil responses that contribute to tissue injury through ROS production, while increasing TGF-beta receptor expression on fibroblasts that facilitates ECM production in response to TGF-beta. TGF-beta, activated locally through integrin-mediated mechanisms, serves as a key driver of fibrosis. A second and distinct cytokine-mediated pathway that can promote fibrosis independently of TGF-beta is the type-2 cytokine axis. Here, the alarmin cytokines IL-25, IL-35, and TSLP, secreted by epithelial cells and other damaged tissues drives the expansion and activation of type-2 innate lymphoid cells (ILC2) that secrete large amounts of IL-5 and IL-13. IL-5 in turn drives the recruitment and activation of local tissue eosinophils, which provide an additional source of type 2 cytokines and other pro-fibrotic mediators. IL-13, derived from eosinophils, CD4+ Th2 cells, and ILC2s exhibits potent pro-fibrotic activity that is TGF-beta independent. Finally, the cytokine IL-11, produced by activated myofibroblasts has been found to stimulate ECM production by myofibroblasts in response to multiple pro-fibrotic mediators, including TGF beta and IL-13.

Caspase-1, NOD, LRR and pyrin domain-containing (NLRP) 3 inflammasome, and NFkB were identified as important upstream mediators contributing to the activation of the IL-17A-TGF-β axis¹²². The exact mechanisms responsible for the sustained activation of NFkB and NLRP3 inflammasome signaling remain unclear, although commensal microbe stimulation of Toll-like receptors on myeloid cells and tissue fibroblasts has been hypothesized as an important activating mechanism, with the resulting pro-inflammatory cytokine and chemokine production exacerbating inflammation and the progression of fibrosis^126,127. Of note, stimulation of TLR4/NFkB in hepatic stellate cells enhances TGF-β signaling by directly downregulating the TGF-β pseudoreceptor BMP and the activin membrane-bound inhibitor Bambi¹²⁷. A related study showed that sustained activation of the NLRP3 inflammasome is associated with increased chemokine expression, recruitment of neutrophils and macrophages, and persistent production of IL-17A and TNF¹²². Thus, activation of the pro-fibrotic TGF-β signaling pathway is driven by several collaborating mechanisms, with the pro-inflammatory cytokine IL-17A playing a prominent role.

While the TGF-β signaling pathway is a well-known driver of fibrosis in many organs and tissues, the type 2-associated cytokines IL-4 and IL-13 have also emerged as distinct but important inducers of fibrosis. Here, the fibrotic response is associated with a predominant eosinophil and M2-like macrophage infiltrate, rather than the neutrophil and M1-like monocyte/macrophage phenotype that characterizes the IL-1-IL-17A-TGF-β axis¹²⁸. Also, instead of acute phase cytokines serving as co-inducers, the alarmin cytokines thymic stromal lymphopoietin, IL-25, and IL-33 function as key initiators of type 2-dependent fibrosis by triggering the production of IL-4 and IL-13 in innate lymphoid cells, T cells, eosinophils, and other type 2-associated leukocytes^129–132. Although IL-13 can induce and activate TGF-β in macrophages¹³³, it may promote fibrosis independently of TGF-β¹³⁴ in part by directly targeting stromal and parenchymal cells, including epithelial populations and collagen-producing myofibroblasts¹³⁵. IL-13, IL-4R, and IL-13Rβ1-deficient mice as well as animals treated with neutralizing antibodies to IL-13 or IL-4R display reduced fibrosis following many different types of tissue injury^136–140, confirming a critical role for type 2 cytokine signaling in the progression of fibrosis (Figure 4).

The mechanisms that dictate whether the IL-1-IL-17A-TGF-β axis or type 2 cytokine response dominate as the key driver of fibrosis remain unclear although the type of cellular damage or duration of the injury likely play key roles. For example, studies with the commonly used “single hit” bleomycin pulmonary fibrosis model revealed a prominent role for the IL-1-IL-17A-TGF-β axis but little to no contribution for type 2 cytokine signaling, despite significant upregulation of IL-4 and IL-13 in the lung¹¹⁰. Nevertheless, a modified version of this model in which bleomycin is injected intra-dermally rather than intra-tracheally over several weeks uncovered a substantial role for IL-4R signaling in the development of pulmonary fibrosis¹⁴¹. Different stimuli or types of injuries that lead to the preferential production of alarmin cytokines versus activation of NF-kB and inflammasome signaling likely play decisive roles as well. For example, integrin receptors that interact with the extracellular matrix were recently shown to preferentially activate TGF-β signaling and production of IL-17A while antagonizing the production of type 2 cytokines¹⁴². Consistent with these observations, several studies have revealed substantial cross-regulation between IL-17A and IL-13^110,143, with marked upregulation of the opposing pathway when only one mechanism was targeted therapeutically^{136,142,144,145}. Consequently, a successful anti-fibrotic strategy will either target the dominant mechanism or reduce both pathways simultaneously. Interestingly, IL-11, a member of the IL-6/gp130 cytokine family, may be one such target as it was recently shown to integrate pro-fibrotic signals emanating from both pathways^146–149.

Not surprisingly, given the robust preclinical data, inhibitors to IL-13 alone (NCT01266135; NCT00987545; NCT00581997; NCT01872689; NCT01629667) or the combination of IL-4/IL-13 (NCT02921971; NCT01529853) were tested in phase 2 trials for pulmonary fibrosis, skin keloids, and systemic sclerosis. Additional cytokine, chemokine or growth factor inhibitors in development for fibrosis are anti-CCR2/CCR5 (NCT02217475; NCT03028740; NCT03059446; NCT02330549) for liver fibrosis and NASH, anti IL-1 (NCT01538719) for systemic sclerosis, anti IL-6 (NCT02453256) for scleroderma, and anti CCL2 (NCT00786201) for pulmonary fibrosis (Table 1).

CONTRIBUTION OF THE MICROBIOME TO FIBROSIS

Environmental factors such as pollutants, chemicals, and microbes can trigger or protect from fibrosis. While the exact environmental agents remain poorly characterized, recent work has focused on host / microbiome interactions as some microbe populations have been found to regulate the progression of fibrosis. Immune and non-immune cells sense microbe-derived, pathogen-associated molecular patterns (PAMPs) through pattern recognition receptors (PRRs), including Nod-like receptors (NLRs) and Toll-like receptors (TLRs), driving innate and adaptive immune responses¹⁵⁰. Extensive studies have investigated microbial sensing in the context of immunity and epithelial cell biology, but studies reporting a direct link between the microbiota and fibrogenesis are now emerging.

The vast majority of human-associated microbes are found in the gut, and during homeostasis these microbial populations are essential in maintaining gut health. However, when the balance between healthy and pathogenic microbes shifts towards pathogenic subsets disease can ensue. Inflammatory bowel disease comprising Crohn’s disease (CD) and ulcerative colitis (UC)) represents a prototypical pathology where dysbiosis is thought to be a key driver of disease pathogenesis, with evidence suggesting a strong link between the microbiota and the development of fibrosis. For example, CD patients carrying variants of the NOD2 gene, an intracellular pattern recognition receptor, are at increased risk of stricture formation which is the major manifestation of intestinal fibrosis¹⁵¹. Furthermore, serologic anti-microbial antibodies are common in patients with CD and are associated with and predictive of intestinal strictures^152,153, and almost all murine models of intestinal fibrosis are influenced by the microbiota¹⁵⁴. For instance, global deletion of the bacterial signaling adaptor molecule MyD88 (a critical signaling adaptor molecule for all TLRs, except TLR3¹⁵⁵) reduced intestinal fibrosis in a model of Salmonella-induced colitis¹⁵⁶. Microbiota driven intestinal fibrosis may be mediated by induction of the IL-33 receptor ST2 on epithelial cells¹⁵⁷ or the pro-fibrotic action of tumor necrosis factor-like cytokine 1A (TL1A)¹⁵⁸. On a cellular level, TLR2 or TLR4 ligands induce cytokine and chemokine secretion from cultured intestinal myofibroblasts¹⁵⁹. Interestingly, although intestinal mesenchymal cells express multiple TLRs and NLRs, a pro-fibrogenic phenotype was triggered exclusively by flagellin, a broad activator of innate and adaptive immunity and a TLR5 ligand. This occurred in a TGF-β1 independent manner and via post-transcriptional regulation¹⁶⁰. The role of myofibroblasts directly sensing PAMPs in intestinal fibrosis has been confirmed in vivo, as selective deletion of MyD88 in αSMA positive cells ameliorated intestinal fibrosis¹⁶⁰.

Dysbiosis in the gut has been shown to influence liver fibrosis as well. Translocation of bacteria and their products across the intestinal barrier due to intestinal barrier disruption is common in patients with chronic liver disease. Increased levels of the microbe-derived ligand lipopolysaccaride (LPS) in the portal vein or translocation of whole bacteria or their products to the liver activates inflammation that leads to fibrosis^161,162. Blocking TLR4 signaling in mice or reducing hepatic exposure to intestinal microbes by reducing microbial load with antibiotics ameliorates experimental liver fibrosis¹²⁷. HSCs express all known human TLRs and respond to TLR4 ligands^127,163, which also downregulate a TGF-β1 decoy receptor sensitizing HSC to the action of TGF-β1¹²⁷. A comparable mechanism has been described in rat pancreatic fibrosis¹⁶⁴. TLR4 signaling includes an additional signaling adaptor, called TRIF. Deletion of TRIF in a diet-induced NASH mouse model reduced hepatic steatosis but increased hepatic fibrosis, and TRIF^−/− HSCs expressed higher levels of CXCL1 and C-C motif chemokine ligands in response to LPS, highlighting a potential mechanism for this unexpected effect¹⁶⁵. Conversely, distinct gut microbiota may be hepatoprotective in the context of liver fibrosis. Increased liver ECM deposition was observed in germ free mice compared to conventionally housed mice¹⁶⁶, which was also observed in MyD88/TRIF deficient mice. In the kidney, pericytes, a myofibroblast precursor, activate a TLR2/TLR4/MyD88-dependent proinflammatory program in response to tissue injury¹⁶⁷. The downstream kinase IRAK4 controlled pericyte myofibroblast conversion in vitro and pharmacological inhibition of MyD88 signaling with an IRAK4 inhibitor reduced fibrosis by attenuating tissue injury in vivo¹⁶⁷. Global TLR4 knockout¹⁶⁸ and a small molecule inhibiting MyD88 ameliorated renal fibrosis in mice¹⁶⁹. In systemic sclerosis, TLR4 and its co-receptors lymphocyte antigen 96 (MD2) and CD14 are overexpressed in lesional skin and chronic dermal LPS exposure led to overexpression of TGF-β signature genes¹⁷⁰.

Of note, several TLRs are promiscuous and can also sense damage associated molecular patterns, lipids or ECM. This has been elegantly shown in lung fibrosis, where TLR4 and the glycosaminoglycan hyaluronan are important for type 2 alveolar epithelial cell renewal, limiting lung injury and fibrosis¹⁷¹. Interestingly, TLR4 was protective in the lung, as opposed to its pathogenic effect in gut and liver fibrosis^127,159. Nasal polyposis is a disease characterized by remodeling of the sinonasal mucosa. Short single stranded DNA molecules, so called CpG oligonucleotides, can activate fibroblasts derived from nasal polyposis patients via TLR9 stimulation, providing an additional example whereby multiple PRRs, activated by distinct ligands, can contribute to aberrant wound healing and fibrosis.

The pathophysiologic relevance of TLR4 in inflammation has led to clinical development programs for rheumatoid arthritis using a TLR4 inhibitor (NCT03241108). Furthermore, pentraxin 2 (PTX2), also known as serum amyloid protein 2, has demonstrated anti-inflammatory and anti-fibrotic properties in multiple preclinical fibrosis models^172,173 and recombinant PTX2 (PRM-151) has entered phase 2 trials for pulmonary fibrosis and myelofibrosis (NCT02550873; NCT01981850) (Table 1).

FUTURE DIRECTIONS

Fibrosis is a major global healthcare burden. Consequently, the discovery of key therapeutic targets with high relevance to human fibrotic disease and the subsequent development of effective antifibrotic therapies directed against these targets continues to be a research priority. Single-cell genomics methodologies have already yielded multiple new discoveries in fibrotic disease that would previously have been unattainable. This field continues to evolve rapidly, and emerging technologies are now able to measure multiple “omic” readouts (genomes, epigenomes, transcriptomes and proteomes) in single cells^174–176. Spatially resolved molecular profiling is expanding our understanding of how these unique populations interact in situ^177–179. The convergence and integration of these multi-modal single-cell technologies^20,180, alongside global initiatives such as the Human Cell Atlas¹⁸¹, represent an extraordinary opportunity to decode the cellular and molecular mechanisms of fibrosis at unprecedented resolution, which should in turn help drive a new era of precision medicine in the treartment of fibrotic disease. Novel therapeutics developed for one fibrotic disorder may be applicable to a wide range of fibrotic diseases due to shared pathways across organs that are uncovered by this work. Drug repositioning efforts may also be impacted by these studies¹⁸².

Despite impressive progress over the past few years in our understanding of the pathogenesis of fibrosis, multiple challenges need to be overcome to translate this information to effective anti-fibrotic therapies (Figure 5). Prognostic animal models and ex vivo primary human tissue culture systems need to be developed that allow better translation of novel mechanisms from the bench to the bedside. Accurate and validated predictors of fibrotic disease progression are needed to stratify patients into high-risk populations. Patient heterogeneity together with the fact that fibrosis progression is typically slow makes selection of patients for clinical trials difficult. Hence, accurate and validated predictors of fibrotic disease progression are needed to stratify patients into high-risk populations prior to inclusion. Currently used trial endpoints are highly variable and often lack the sensitivity needed to predict favorable responses over a short period of time, which necessitates the inclusion of large numbers of patients in clinical trials. Consequently, endpoints for fibrosis clinical trials continue to evolve and may require a more global approach involving scientists, industry leaders, patients and regulatory partners, as recently shown for liver fibrosis and intestinal fibrosis^183,184. Ideally, non-invasive endpoints that better correlate with clinically meaningful outcomes are needed. Recent research in the field has been fueled by the discovery of robust biomarkers and cutting-edge imaging modalities such as PET imaging of collagen and molecular imaging of fibrosis^185,186, which allows fast, non-invasive, and whole-organ-quantitative and longitudinal readouts of drug efficacy in antifibrotic clinical trials. Furthermore, combining molecular imaging of fibrosis¹⁸⁵ with cutting edge omic approaches such as single cell genomics^174–176 could markedly improve patient diagnostics, staging, prognostication, stratification and cohort enrichment, which would in turn optimize clinical trial design and maximize the number of trials that could be run quickly and efficiently^177–179. Innovative approaches to trial design are being developed that allow incorporation of adaptive strategies and use of ‘bucket’ trials that include patients with different types of fibrosis, as well as inclusion of ‘real-world’ evidence into the regulatory approval process.

Figure 5:

Challenges and solutions in the translation of antifibrotic mechanisms into drugs.

Similar to the major advances seen in cancer therapy or the successful treatment of HIV and viral hepatitis, we will very likely see increasing numbers of clinical trials testing combinations of drugs, as fibrosis is increasingly recognized as a highly complex disorder, with multiple mechanisms collaborating to drive disease progression. These antifibrotic drug cocktails will likely target a variety of orthogonal mechanisms, including a range of receptors, signaling pathways, and cell types that have been shown to function as core drivers of fibrosis in multiple disease states. These multi-faceted approaches should pave the way towards the delivery of effective antifibrotic therapies in the future.

Abbreviations

ACKR

Atypical chemokine receptor

AT

Alveolar type

BMP

Bone morphogenetic protein

CCA

Canonical correlation analysis

CD

Crohn’s disease

CEACAM

Carcinoembyonic antigen-related adhesion molecules

C/EBPβ

CCAAT/enhancer binding protein β

ECM

Extracellular matrix

FAP

Fibroblast activation protein

Hic

Hypermethylated in cancer protein

HSC

Hepatic stellate cells

Ly

Lymphocyte antigen

Lyve 1

Lymphatic vessel endothelial hyaluronan receptor 1

IL

Interleukin

IBD

Inflammatory Bowel Disease

LOX

Lysyl oxidases

LPS

lipopolysaccaride

MP

Mesenchymal progenitor cells

NLR

Nod-like receptors

NLRP

NOD, LRR and pyrin domain-containing protein 3

NOD

Nucleotide-binding oligomerization domain-containing protein

PAMP

Pathogen-associated molecular patterns

PDGF

Platelet derived growth factor

PLVAP

Plasmalemma vesicle-associated protein

PRR

Pattern recognition receptors

scRNAseq

Single cell RNA sequencing

SatM

Segregated-nucleus-containing atypical monocytes

Siglec

Sialic acid-binding immunoglobulin-type lectins

SMA

Smooth muscle actin

SOX

SRY-box

TGF

Transforming growth factor

THY

Thymus cell antigen

TLR

Toll-like receptors

TNFRSF

Tumor necrosis factor receptor superfamily

TREM

Triggering receptor expressed on myeloid cells

UC

Ulcerative colitis