What Explains the Pattern of Fibrosis in Chronic Graft‐Versus‐Host Disease?

ABSTRACT

Fibrosis is common in diverse conditions but has different patterns across tissues and organs despite shared cellular and molecular pathways reasons for which are mostly unexplained. In this review we discuss chronic graft‐Versus‐host disease (GvHD), progressive systemic sclerosis (PSS), idiopathic pulmonary fibrosis (IPF), primary myelofibrosis (PMF), chronic lung allograft dysfunction (CLAD) and VEXAS syndrome to explore potential mechanisms underlying these patterns of fibrosis. We propose the distribution of fibrosis reflects a complex interplay between aetiology, initiating stimuli, local immune response, tissue‐specific micro‐environment and host susceptibility. Distinguishing systemic and local drivers of fibrosis is needed to understand and explain these patterns. A deeper understanding of early, tissue‐specific events governing fibrotic responses may allow development of preventive and therapeutic strategies.

1. Introduction

Fibrosis is a prominent feature of chronic graft‐Versus‐host disease (GvHD) and other disorders including progressive systemic sclerosis (PSS), idiopathic pulmonary fibrosis (IPF), primary myelofibrosis (PMF), chronic lung allograft dysfunction (CLAD) and VEXAS syndrome. Curiously, chronic GvHD and PSS are systemic but have different patterns of fibrosis. Why is uncertain [1, 2, 3, 4, 5]. We discuss mechanism(s) underlying end‐tissue and ‐organ fibrosis and the potential role of initiating stimuli which may explain, at least part, the pattern of fibrosis in chronic GvHD Versus other fibrotic diseases.

Chronic GvHD predominately affects the skin, eyes, mouth and sometimes the lungs but rarely the bone marrow, kidneys or heart. In contrast, in PSS fibrosis occurs in several sites including the skin, lungs, kidneys and heart. Neither chronic GvHD nor PSS affect the bone marrow whereas in PMF there is bone marrow and sometimes spleen fibrosis but no fibrosis in the skin or elsewhere despite most of the blood within tissues and organs where fibrosis occurs in other diseases. In IPF fibrosis is confined to the lungs despite the same widespread blood flow [3, 4, 5]. These distinct fibrotic patterns are also evident in animal models of these diseases [6, 7, 8, 9].

These diseases share similar cellular and molecular pathways in the late stages of fibrosis but their aetiologies and initiating mechanisms, when known, differ [1, 10]. Interrogating these initial events may help explain why fibrosis occurs in different patterns across tissues and organs.

The initial inflammatory response underlying fibrosis is likely influenced by the nature of the triggering stimulus and the unique micro‐environment of each tissue and organ [11, 12]. Consequently, some tissues and organs may develop fibrosis when exposed to specific stimuli whilst others similarly exposed do not [13, 14]. Other factors such as prior injury may contribute to these responses [15, 16].

2. Wound Healing and Fibrosis

Wound healing and fibrotic diseases share several common features (Figure 1). In both an injury triggers a series of repair processes at the site which typically begins with inflammation when damaged cells up‐regulate inflammatory mediators [13]. This pro‐inflammatory environment recruits immune cells to the site of injury. Subsequently, fibroblasts and other mesenchymal cells are activated and differentiate into myo‐fibroblasts because of increased expression of fibrogenic cytokines like fibroblast growth factors (FGFs) and platelet‐derived growth factors (PDGFs) which then secrete components of the extra‐cellular matrix (ECM) including reticulin and collagen [17].

FIGURE 1.

Controlled repair in wound healing versus dysregulated repair in fibrosis.

Fibrosis can be considered a form of uncontrolled repair in response to injury [18, 19]. In normal wound healing activated myo‐fibroblasts undergo apoptosis and the repair response stops [20, 21]. However, in fibrotic disorders severe and/or repetitive injury and subsequent structural changes in the affected tissues and organs result in persistent myo‐fibroblast activation and progression of fibrosis [22, 23]. Also, persistent and widespread trans‐differentiation to myo‐fibroblasts including epithelial, endothelial and adipocyte, as well as phenotypic conversion of circulating mononuclear cells, pericytes and telocytes are reported [24]. Continuous accumulation of ECM disrupts normal tissue architecture, impairs tissue and organ function and may ultimately result in organ failure [25, 26].

3. Fibrotic Diseases

To frame the question of why the pattern of fibrosis in chronic GvHD differs from several other fibrotic diseases we discuss several examples (Table 1). Skin fibrosis is the hallmark of chronic GvHD. However, other tissues and organs are often involved including the lungs and eye and oral glands but not typically the kidneys or heart [27, 28, 29, 30, 31]. In chronic GvHD, the initiating factor for fibrosis is immune activation resulting from histo‐compatibility discordances between donor and recipient after an allogeneic haematopoietic cell transplant [13]. T‐ and B‐cells are involved to varying degrees in mediating chronic GvHD [3]. Abnormalities in central and peripheral immune tolerance mechanisms may explain, in part, chronic GvHD pathophysiology [32, 33, 34, 35, 36, 37, 38]. Exposure to ionizing radiation is a risk factor [13]. Skin involvement in chronic GvHD can occur in areas of previous damage from UV and radiation therapy, viral infections and trauma [15, 16]. Although skin involvement in chronic GvHD and PSS is similar there are differences in micro‐vascular abnormalities which may reflect distinct aetiologies and/or early pathogenic events despite sharing common downstream fibrotic pathways [39, 40, 41].

TABLE 1.

Primary sites of fibrosis in various fibrotic diseases.

Disorder	How fibrosis relates to the disease	Main process that may initiate fibrosis	Main sites of fibrosis
Chronic GvHD	Develops downstream of tissue injury and immune dysregulation	Transplant‐related tissue injury followed by donor alloimmune responses and impaired immune tolerance	Skin/fascia Lungs
IPF	A major defining manifestation	Repeated injury to alveolar epithelial cells, cellular senescence and defective tissue repair in a susceptible lung	Lungs
PSS	A major disease manifestation, although vascular and immune abnormalities may precede it	Small‐vessel injury, autoimmunity and abnormal activation of stromal cells	Skin Lungs Gastrointestinal tract Heart
PMF	Defining in overt PMF; advanced fibrosis is not present in pre‐fibrotic PMF	Signals from abnormal clonal megakaryocytic and myeloid cells activate marrow stromal cells	Bone marrow
CLAD	Develops after repeated allograft injury and alloimmunity	Repeated epithelial injury and recipient alloimmune responses against the transplanted lung	Transplanted lung
AIMF	Develops as part of an autoimmune process	Immune‐mediated activation of marrow stromal cells	Bone marrow
VEXAS syndrome	Not a defining feature; marrow fibrosis is reported in a subset	Inflammation produced by UBA1‐mutant myeloid cells; its direct relationship to fibrosis remains uncertain	Bone marrow (in a subset of patients)
Morphea/ lichen sclerosus	Develops after localized cutaneous inflammation and injury	Local immune and vascular injury followed by fibroblast activation	Skin and underlying soft tissues

Abbreviations: AIMF, autoimmune myelofibrosis; CLAD, chronic lung allograft dysfunction; GvHD, graft‐Versus‐host disease; IPF, idiopathic pulmonary fibrosis; PDGF, platelet‐derived growth factor; PMF, primary myelofibrosis; PSS, progressive systemic sclerosis; TGF‐β, transforming growth factor beta; UBA1, ubiquitin‐like modifier activating enzyme 1; VEXAS, vacuoles, E1 enzyme, X‐linked, autoinflammatory, somatic.

Several potential factors might explain the distribution of fibrosis in chronic GvHD including differences in T‐cell migration patterns, varying degrees of self‐antigen expression across different organs and pre‐existing damage and inflammation from pre‐transplant interventions. Sacirbegovic et al. reported GvHD is primarily maintained locally in target tissues with limited reliance on the continuous recruitment of allo‐reactive blood‐derived T‐cells. In mice, clonal analyses are reported to show distinct T‐cell populations across different organs which become more pronounced with time. Mathematical modeling indicates a small pool of local progenitor T‐cells expands within each tissue to sustain a larger population of effector T‐cells, largely independent of ongoing input from the blood [42]. Ng et al. suggested 3 distinct biologic sub‐types of chronic GvHD, none of which were closely associated with specific organ involvement [43].

Acute and chronic GvHD are clinically distinguishable but biologically related conditions. NIH criteria distinguish acute and chronic GvHD by clinical features rather than by time posttransplant [44]. Acute GvHD is one of the strongest predictive co‐variates for chronic GvHD. However, not everyone surviving acute GvHD develops chronic GvHD nor does everyone with chronic GvHD seemingly have had acute GvHD. Also, severities of acute and chronic GvHD are correlated [45, 46]. Chronic GvHD may develop during active acute GvHD (progressive onset), after resolution of acute GvHD (quiescent onset) or seemingly without prior acute GvHD (new onset) [47]. However, the latter is confounded by difficulty in accurately diagnosing acute GvHD [46]. These categories describe the clinical course rather than discrete biological mechanisms.

The patho‐phyisology of chronic GvHD may reflect epithelial and stromal cell injuries from acute GvHD and ongoing inflammation. Continued tissue damage, impaired thymic and peripheral tolerance, dysregulated posttransplant immune reconstitution, macrophage activation and abnormal fibroblast‐mediated repair may shift the predominant phenotype from acute tissue inflammation towards chronic immune dysregulation, tissue remodeling and fibrosis. In this model acute GvHD does not simply transform into chronic GvHD. Rather, the tissue injury and immune abnormalities associated with acute GvHD create conditions promoting chronic GvHD [3, 13].

In CLAD after lung transplants histo‐compatibility discordances are limited to the lungs illustrating local Versus systemic effects of the immune responses. Fibrosis in CLAD is restricted to the lungs where it resembles lung involvement in chronic GvHD [48]. Also, apart from acute GvHD, risk factors for CLAD and chronic GvHD are similar [48].

PMF is characterized by variants in JAK2, CALR or MPL in most cases resulting in constitutive activation of the JAK/STAT signaling pathway. The characteristic feature of PMF is the presence of immature and abnormal clonal megakaryocytes and subsequent bone marrow fibrosis [49, 50, 51]. Increased inflammatory cytokines and alterations in the bone marrow micro‐environment are reportedly driven by cytokine‐mediated mechanisms [52, 53, 54]. Neoplastic monocyte‐derived fibroblasts are increased in the bone marrow in PMF compared with normals and are thought to be important in bone marrow fibrosis by producing TGF‐β1, collagen and reticulin [55, 56, 57, 58, 59]. JAK2‐mutated endothelial cells are reported in people with myelo‐proliferative neoplasms (MPNs) which may contribute to disease pathophysiology and influence the proliferative advantage of JAK2‐mutated haematopoietic stem and progenitor cells [60, 61, 62, 63, 64]. Transforming growth factor beta‐1 (TGF‐β1), an important mediator of fibrosis, is increased 5‐10‐fold in bone marrow megakaryocytes but not in the blood perhaps because of a short half‐life [7, 65]. Plasma levels of other inflammatory cytokines are increased in PMF but do not cause fibrosis at other sites [66]. In people with an MPN, bone marrow TGF‐β1 concentration is independent of whether or not they develop bone marrow fibrosis [67]. These observations question the role of inflammatory cytokines like TGF‐β1 in the development of fibrosis in PMF and suggest possible contributions of other factors such as megakaryocytes and micro‐environmental cells [64]. However, PMF can be cured by a haematopoietic cell transplant making an important role for the bone marrow micro‐environment unlikely although some cells comprising the micro‐environment may also be altered or replaced by donor‐derived cells [68].

In PSS auto‐immunity, inflammation and vascular injury are hypothesized to initiate and sustain a chain of events ending in fibrosis; the sequence of these events is uncertain [69, 70]. Parvovirus B19 and cytomegalovirus (CMV) infections are proposed as potential triggering events but this is controversial and most people with these infections do not develop PSS [71]. Regardless of the trigger micro‐vascular involvement is central to the development and progression of fibrosis in PSS and may explain the widespread tissue and organ involvement. However, sometimes fibrosis in PSS is limited to the skin or may not develop for years despite the presence of systemic sclerosis‐specific auto‐antibodies and vascular injury [69, 72, 73, 74].

In IPF fibrosis is restricted to the lungs. The underlying mechanism(s) is not understood. Environmental factors and recurrent local injuries to the alveolar epithelium are proposed as important but this is controversial [2]. Repeated injury to the alveolar epithelium is thought to disrupt normal tissue repair resulting in fibrosis [75, 76, 77]. The most affected cells are alveolar epithelial type 2 cells (AEC2s) which are progenitors of type‐1 alveolar epithelial cells (AEC1s) and are important in maintaining alveolar epithelial cell renewal [12, 78]. Interestingly, targeted deletion of telomeric repeat–binding factor 1 (Trf1) in surfactant protein C–positive AEC2s in mice results in progressive pulmonary fibrosis and loss of lung progenitor cells [78, 79]. Variants in TERT, TERC, PARN and RTEL1 are risk factors [80]. In familial IPF the single nucleotide variant (SNV) rs35705950 in the promoter region of mucin 5B (MUC5B) is associated with an increased risk of IPF [76]. Interestingly, there is an increased risk of developing rheumatoid arthritis in people with IPF suggesting a possible contribution of auto‐immunity [81]. IPF does not respond to most anti‐inflammatory therapies putting in question the role of inflammation but this likely reflects delayed therapy [4, 82, 83]. A recent study of nerandomilast, an anti‐inflammatory drug, reported slowing of progression of fibrosis [84]. There is also improvement in mice after transplants of adult lung cell [78].

Autoimmune myelofibrosis (AIMF) is a rare complication of autoimmune disorders like systemic lupus erythematosus (SLE). Sometimes it arises spontaneously. Fibrosis in AIMF is typically mild and restricted to reticulin deposition [85, 86, 87]. Bone marrow biopsies show aggregates of T‐cells, poly‐clonal B‐cells, increased cellularity with normal appearing megakaryocytes (in contrast to PMF) [85, 87, 88, 89]. Unlike PMF, AIMF responds to corticosteroids as do some cases of chronic GvHD and IPF [85, 88]. The aetiology and mechanism(s) of AIMF remain unknown.

VEXAS syndrome (Vacuoles, E1 enzyme, X‐linked, Autoinflammatory, Somatic) is a rare autoinflammatory–haematologic disease caused by somatic mutations in UBA1. Clonal haematopoietic stem and progenitor cells (HSPCs) expand and outcompete wild type cells driving disease development [90, 91]. VEXAS is a multi‐system inflammatory disease involving skin, joints, lungs, heart, kidneys, genitourinary tract, eyes and the bone marrow [92]. Concordant haematologic disorders include myelodysplastic neoplasm (MDS), plasma cell myeloma or monoclonal gammopathy of undetermined significance (MGUS) [93, 94]. Activation of inflammatory pathways, release of pro‐inflammatory cytokines, proliferation of UBA1‐mutant myeloid cells is reported [95, 96, 97]. Bone marrow fibrosis is common but nowhere else. The bone marrow shows increased myelopoiesis and dysplastic megakaryocytes [96, 98, 99, 100, 101, 102, 103]. About one‐quarter of people with VEXAS have skin vasculitis but do not develop fibrosis [104, 105].

Chronic cutaneous fibrosis can complicate inflammatory dermatoses such as morphea and lichen sclerosus. The aetiologies and initiating events of these disorders are poorly understood. In morphea, an early phase of lymphocyte and macrophage infiltration and micro‐vascular injury may precede fibroblast activation, collagen deposition and sclerosis. Local trauma, radiation and infections are proposed as potential triggers but their causal roles are uncertain. In lichen sclerosus, chronic interface inflammation is associated with tissue remodeling and scarring. Auto‐immunity and local tissue injury including trauma and friction are proposed but what initiates and sustains the local immune response is unclear. In both disorders, activation of fibroblasts and myofibroblasts is likely a downstream response to persistent local inflammation rather than the initiating event. These dermatoses reflect how chronic local inflammation may result in fibrosis but do not explain why inflammation develops or why fibrosis remains restricted to particular sites [106, 107].

As discussed above, these diseases differ in their initiating processes. Another distinction is the timing and prominence of fibrosis. In some disorders, fibrosis is a major defining or presenting feature whereas in others fibrosis develops after a period of persistent tissue injury, inflammation or immune dysregulation (Table 1) [1, 10]. Both groups of disorders converge on partially shared downstream pathways including persistent immune‐stromal signaling, pro‐fibrotic mediators like TGF‐β and PDGF, activation of extra‐cellular‐matrix‐producing mesenchymal cells and failure of normal tissue repair. Matrix‐stiffness feedback may then sustain fibrosis after the initiating stimulus has diminished. Importantly, inflammation is not restricted to the latter group. Inflammatory processes may also precede or accompany fibrosis in fibrosis‐dominant disorders even when fibrosis is the first recognized clinical manifestation [1, 10].

4. Conclusion

The reason(s) for the distinct pattern of fibrosis in chronic GvHD, a systemic disorder, compared with other fibrotic disorders some of which, like PSS, are also systemic is unclear and likely reflect an interplay between aetiology, site, nature and severity of the initial stimulus, different cellular interactions and micro‐environmental factors. Genetic and/or epigenetic modifiers and prior damage and inflammation influence the threshold for uncontrolled local tissue repair. It is important to distinguish between systemic and local immune and inflammatory processes. Local stimuli recruit and activate immune cells at the site of injury without systemic effects. Pro‐inflammatory cytokines like TGF‐β may have different concentrations and effects at different sites including the skin, blood, bone marrow and lungs. We suggest systemic fibrotic diseases like PSS and chronic GvHD may reflect the underlying etiology and the resulting widespread distribution of the initial injury whereas other fibrotic disease like IPF and CLAD are site‐restricted. Features of the affected tissue or organ are also important. In summary, although some of these diseases are systemic, the development of fibrosis appears primarily driven by local tissue injury.

Future research should explore the relationship between the characteristics of the initiating injury—including its intensity, chronicity and nature—and the corresponding inflammatory response, particularly in early‐stage fibrosis. For instance, what underlies the higher propensity for skin fibrosis in chronic GvHD and PSS? Is it driven by donor T‐cell homing and higher local antigen expression in chronic GvHD, an inherent predisposition of the skin to develop fibrosis or a combination of these factors? A comparison of local mediators of inflammation and immune cell activity Versus systemic effects could be informative and identify tissue‐ and organ‐specific mechanisms enabling development of more effective prevention and therapy.

Author Contributions

Mohammadamin Noorafrooz, Hoda Kavosi and Robert Peter Gale authors developed the study and the typescript, take responsibility for the content and agreed to submit for publication.

Funding

The authors has nothing to report.

Conflicts of Interest

Robert Peter Gale is a consultant to Antengene Biotech LLC; Medical Director, FFF Enterprises Inc.; A speaker for Janssen Pharma and Hengrui Pharma; Board of Directors: Russian Foundation for Cancer Research Support and Scientific Advisory Board, StemRad Ltd.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.