Imperfect repair in aging: senescent cells and the hepatic fibrotic niche

Abstract

Aging proceeds in a nonuniform spatiotemporal manner across tissues. While metabolic stress and chronic inflammation are implicated, the underlying mechanisms remain elusive. Here, we propose that imperfect wound healing—a failure of full resolution—creates and sustains pathological niches that drive progressive age-related dysfunction. Using the liver as a model system, we deconstruct this ‘imperfect repair’. We posit that it is driven by a pro-fibrotic, non-resolving microenvironment sustained by complex crosstalk between functionally heterogeneous senescent cells and non-senescent scar-associated cell (SAC) populations (including macrophages, endothelial cells (ECs), and hepatic stellate cells (HSCs)). This pathological ecosystem is further shaped by the spatial context of hepatic zonation collapse, and the dysregulation of core signaling hubs, like WNT, Transforming Growth Factor (TGF)−β, and YAP and TAZ (YAP/TAZ). Viewing aging through the lens of imperfect repair provides a unifying framework linking senescence, inflammation, and fibrosis. This perspective shifts the therapeutic paradigm from targeting single senescent cells toward engineering the pathological niche itself, and redirects focus from end-stage disease, to the sub-clinical, spatial origins of tissue vulnerability.

INTRODUCTION

The process of aging is not a uniform, linear decline dictated solely by chronological time. Instead, it manifests as a spatially and temporally heterogeneous process across, and within, tissues (1, 2). This heterogeneity is driven by the cumulative impact of various stressors that include metabolic insults, chronic inflammation, and mechanical strain, which trigger local tissue repair responses (3). While acute wound healing is a highly orchestrated process that restores tissue integrity (4), its resolution often fails with advancing age. This fundamental failure of resolution, which we term ‘imperfect repair’, creates and perpetuates pathological niches that underpin progressive organ dysfunction (5).

We posit that imperfect repair is a primary driver of the aging process. Within this framework, the persistent activation of repair programs represents a long-term adaptation to chronic stress. However, this initially protective response can transition into a chronic, maladaptive state, which we describe as damaging adaptation, where beneficial mechanisms ultimately cause harm. This progression is characterized by the sustained presence of senescent cells, unresolved inflammation, and the progressive accumulation of a fibrotic extracellular matrix (ECM) (6, 7). Together, the constituents of this pathological niche create a self-perpetuating pathological cycle that actively drives tissue decline.

Here, we explore this concept through the lens of the liver, an organ renowned for its remarkable regenerative capacity, yet highly susceptible to age-related fibrotic diseases. We first discuss how cellular senescence, initially a pro-regenerative signal, becomes a pivotal orchestrator of the chronic, non-resolving wound. Subsequently, we dissect the cellular and molecular anatomy of the resulting fibrotic niche. Finally, we examine the core signaling networks that sustain this pathological state, providing a molecular basis for the unhealing wound. By viewing aging through the prism of imperfect repair, we provide a unifying framework that connects interrelated hallmarks of aging, and reveals new therapeutic avenues to enhance tissue resilience.

CORE OF THE PATHOLOGICAL TRANSITION: SENESCENT CELL PERSISTENCE AND FUNCTIONAL HETEROGENEITY

Cellular senescence, a hallmark of aging and a driver of related diseases (3, 8), is defined by a stable cell cycle arrest that is driven by the upregulation of cyclin-dependent kinase (CDK) inhibitors, such as p16^Ink4a (9) and p21^Cip1 (10, 11). Paradoxically, a transient form of this state is essential for acute physiological wound healing, where its appearance orchestrates an orderly regenerative response (12). This paradoxical function reflects a highly dynamic, multi-phasic program, rather than a static state. In response to diverse triggers, p21 often spearheads the initial arrest (13). This involves both the upregulation of p21 mRNA, and the rapid translation of pre-existing p21 mRNA, to deploy immediate signals (14). The subsequent dynamics of p21 expression, namely its amplitude and duration, serve as a critical determinant of cell fate, distinguishing between a transient arrest, and a stable senescence decision (15). This initial p21–mediated decision point appears to initiate a complex regulatory relay. This relay governs the progression towards a stable senescent state (often characterized by p16 upregulation) through a dynamic interplay of various epigenetic regulators, such as the bromodomain protein BRD4 (16), and the acetyltransferase KAT7 (17), and key transcriptional drivers that include AP−1 (18), GATA4 (19), and C/EBPβ (20).

This complex relay dictates the temporal evolution of the Senescence-Associated Secretory Phenotype (SASP) (21, 22). The secretome is not monolithic; it emerges in distinct waves, which are hypothesized to mediate different functions. The process may begin with an early, p21–dependent secretome, termed PASP, involved in repair, which can then transition into a canonical, NF–κB–maintained SASP (23). Concurrently, other inputs, such as the cGAS−STING pathway, can establish an antiviral, interferon-driven secretome (24, 25). This mature, persistent SASP is responsible for generating the significant pro-inflammatory and pro-fibrotic signals, such as IL−6, IL−1α, and IL−8, that drive pathology (26, 27). Moreover, the influence of senescent cells extends beyond this paracrine secretome. Senescent cells also engage in juxtacrine signaling via direct cell-to-cell contact, particularly through the Notch signaling pathway, which propagates the senescent state and drives local tissue dysfunction (28). While the precise mechanisms governing this process are still being elucidated, the inherently dynamic nature of the senescence program is hypothesized to be a primary contributor to the functional heterogeneity observed in vivo.

These complex molecular dynamics are exemplified in vivo during acute, physiological repair. To illustrate how these principles operate in a biological system, the liver, with its remarkable regenerative capacity, provides an exemplary context for this well-orchestrated beneficial program (29). Following acute insults, such as a single dose of carbon tetrachloride (CCl4) and partial hepatectomy, hepatocytes and HSCs undergo transient senescence (30, 31). Senescent HSCs actively secrete inflammatory and fibrogenic factors through the SASP, which includes IL−6 and ligands for CXCR2. These SASP factors are crucial to promote hepatocyte proliferation; IL−6 activates STAT3 and induces YAP activation via SRC family kinases, and synergizes with CXCL2 to activate ERK1/2 (32). Furthermore, senescent HSCs upregulate cell surface ligands for the NKG2D receptor, marking them for elimination by natural killer (NK) cells (Fig. 1) (33). This coordinated immune-mediated clearance is a critical step to prevent the accumulation of these cells, and acts as an effective brake on excessive scarring, thereby facilitating orderly tissue regeneration.

Fig. 1.

Dual roles of cellular senescence in acute wound healing and chronic hepatic fibrosis. Schematic illustrating the context-dependent roles of hepatic senescent cells. After acute injury (left panel), transient senescence promotes tissue repair and the immune clearance of senescent cells. However, under chronic damage (right panel), persistent senescence drives chronic inflammation, extracellular matrix (ECM) remodeling, and the formation of a fibrotic niche, leading to fibrosis progression. This figure was created with BioRender.com.

The pathological shift from this beneficial to detrimental role occurs when this tightly regulated program fails in its resolution phase. This failure is often actively driven by systemic age-related immunosenescence, where the capacity of the immune system to clear early PASP–driven senescent cells is markedly compromised (34, 35). This allows senescent cells to persist, mature into the NF–κb driven state, and acquire immune-evasive properties. These properties include the upregulation of immune checkpoint ligands, such as PD−L1, and non-classical MHC molecules, like HLA−E, alongside the secretion of immunosuppressive factors (e.g., TGF−β), which collectively impair cytotoxic T–cell and NK–cell function and establish a pathological feed-forward loop (36, 37). The pathogenic role of these persistent cells is supported by studies showing that senolytic removal of these cells improves tissue function and mitigates age-related fibrosis (38). It is this failure of resolution that marks the transition from an acute, adaptive response to a chronic, maladaptive state, turning cellular senescence from an orchestrator of acute repair, into the persistent engine of imperfect repair (Fig. 1).

However, a fundamental, unanswered question remains: what precisely dictates this systemic failure of resolution, and why does a beneficial, transient program become detrimental and persistent in chronic contexts? The challenge is amplified by the fact that the consequence of this persistence is not uniformly pathogenic, but rather reveals a striking functional heterogeneity. For example, recent lineage-tracing studies in liver fibrosis models highlight this paradox: the selective ablation of p16–positive senescent macrophages ameliorates disease, confirming their pro-fibrotic role. In stark contrast, ablating p16–positive senescent ECs from the same niche worsens fibrosis and impairs regeneration, indicating that even in a chronic setting, they retain a reparative function (39). This paradox reveals that the status of a single marker, such as p16 expression, is insufficient to predict the niche’s functional state. It suggests that the answer to the field’s central question lies not in the senescent cell in isolation, but rather in an integrated understanding of its complex, reciprocal interactions within the pathological niche.

THE PATHOLOGICAL ECOSYSTEM: CELLULAR DRIVERS OF THE FIBROTIC NICHE

Although the well-established regenerative capacity of liver has led to the long-held belief that it was resilient to the insults of aging, this power is not limitless; the aging liver exhibits a subtle yet progressive functional decline, characterized by reduced hepatic blood flow (40), impaired drug detoxification, and the accumulation of non-functional macromolecular aggregates, such as intracellular lipofuscin, and modified proteins, like Advanced Glycation End Products (AGEs) (41, 42). These acute repair mechanisms fail to resolve under conditions of chronic challenge, such as persistent metabolic insults, viral infection, or toxins This failure of resolution results in imperfect repair—a deviation from homeostatic regeneration where functional tissue is replaced by scar tissue. This transitions the response from a transient, reversible state, to a chronic, irreversible structural pathology (5). The persistent engine of this process, the senescent cell, fundamentally reshapes the local tissue through its complex SASP. The SASP mediates a dual pathological process of chronic inflammation and pathological ECM remodeling, locking the tissue into a non-resolving, pro-fibrotic state that builds the scar (43, 44). This is achieved by simultaneously promoting new collagen deposition (e.g., via TGF−β1), while fundamentally disrupting the delicate balance between matrix degradation (via matrix metalloproteinases (MMPs)) and inhibition (via tissue inhibitors of metalloproteinases (TIMPs)) (45-47). This disruption is clearly exemplified by the high-level secretion of PAI−1 (SERPINE1), which potently inhibits the uPA/uPAR system (a key pathway for ECM breakdown) (48). This suppression of matrix degradation, coupled with increased matrix synthesis, creates a self-perpetuating, pathological microenvironment: the fibrotic niche. The clinical and pathological endpoint of this unhealing wound in the liver is cirrhosis, an irreversible state defined by both extensive scar tissue and a substantial loss of core metabolic and synthetic functions.

This functional collapse is driven by a complex multicellular ecosystem. To understand the anatomy of this fibrotic niche, we focus on the specific pathological contributions of two critical cellular axes: (1) a diverse population of persistent senescent cells, and (2) non-senescent, pro-fibrotic SACs. Although single-cell technologies have revealed that the accumulation of persistent senescent cells is a hallmark, this phenotype is highly heterogeneous, shaped by both the originating cell’s identity, and the specific damage context (49). In the parenchymal compartment, senescence in hepatocytes is triggered by accumulating metabolic insults, such as lipotoxicity, DNA damage, and excessive iron accumulation leading to ferroptotic stress (50, 51). These result in profound metabolic failures driven by mitochondrial dysfunction that severely impair core liver functions, and a potent pro-inflammatory SASP. A major upstream driver for this inflammatory secretome is the activation of the cGAS−STING innate immune pathway (52). This pathway senses mitochondrial DNA released into the cytosol during abortive apoptosis, triggering factors (e.g., IL−6, IL−1β), alongside an imbalanced ECM-remodeling secretome (MMPs, TIMPs) (53). Furthermore, the impact of senescent hepatocytes is not confined to their cell-autonomous dysfunction; they act as key mediators of both local paracrine signaling and systemic aging. For example, the SASP secreted by senescent hepatocytes, exemplified by factors like GDF15, acts in a paracrine manner on various adjacent non-parenchymal cells, fueling the pro-fibrotic microenvironment (54). Concurrently, acute hepatocellular senescence can induce a cascading senescence response in distant organs, including the kidney and lung, via circulating TGF−β (55). This highlights that the liver’s parenchymal compartment serves as a hub that exports its pathological state, accelerating systemic aging.

Although non-parenchymal cells (NPCs) are numerically overshadowed by hepatocytes, their pathological burden in aging and fibrotic disease is disproportionately high (56). Liver sinusoidal endothelial cells (LSECs) are a critical population that are frequently driven into senescence by their unique scavenger function, which ensures constant exposure to gut-derived microbial products (e.g., pathogen-associated molecular patterns (PAMPs), like lipopolysaccharide (LPS)), and metabolic damage-associated molecular patterns (DAMPs) (e.g., oxidized LDL) (57). However, this senescent state is defined by a fundamental functional paradox. While this p16–positive population is known to be indispensable for sinusoidal homeostasis, performing essential scavenger functions like the clearance of oxidized lipoproteins, studies employing lineage tracing and the targeted, cell-type-specific elimination of p16–positive populations are now unraveling the molecular basis of their dysfunctional reparative role in fibrosis (58, 59). This dysfunctional state, which transforms these reparative cells into key architects of the fibrotic niche, manifests along two primary axes. First, they undergo a critical loss of their unique endothelial identity. This is characterized by VEGFR2 downregulation and a marked impairment of endothelial nitric oxide synthase (eNOS) activity, which reduces Nitric Oxide bioavailability (60). Second, they transform into active inflammatory mediators. This is evidenced by the expression of immune cell adhesion molecules (e.g., VCAM−1), and the upregulation of PLVAP to promote monocyte transmigration (61, 62). They also secrete neutrophil chemo-attractants, such as CXCL1, which drive their pathogenic trafficking into the tissue (63). This influx of neutrophils is particularly damaging, as they significantly enforce the senescent state by releasing reactive oxygen species, which induces paracrine telomere dysfunction in adjacent cells, and locks in the imperfect repair state. Collectively, this dysfunction drives a pathological capillarization of the sinusoid—a process that includes the formation of a dense subendothelial basement membrane, directly contributing to the physical ECM scaffold of the fibrotic niche. This leads to defenestration, compromised substrate exchange between blood and hepatocytes, and increased intrahepatic vascular resistance, a primary driver of portal hypertension.

HSCs are the central architects of the fibrotic scar, serving as the primary source of pathological ECM (64). While their transient senescence following acute injury can be beneficial and pro-regenerative, it is their persistent dysregulation in chronic disease that is fundamental to the imperfect repair process. In this chronic fibrotic context, they are often driven by sustained pro-fibrotic signaling (such as TGF−β) or excessive oxidative stress, and adopt a hyper-fibrogenic state (65). Recent single-cell studies in metabolic dysfunction-associated steatotic liver disease (MASLD) contexts have revealed that these senescent HSCs acquire a unique pathogenic phenotype, distinct from conventionally activated HSCs, characterized by the aberrant expression of markers traditionally associated with macrophages, such as the mannose receptor CD206 (MRC1) (66). They serve as a primary source of ECM components, and secrete pro-fibrotic SASPs (e.g., TGF−β1, PDGF, CTGF) and the key ECM–stabilizer PAI−1, which actively recruit other immune cells and perpetuate the fibrogenic loop.

While the orchestrated recruitment of macrophages to clear senescent cells is an essential process for tissue resolution, this fundamental interaction is compromised within the fibrotic niche. Senescent macrophages accumulate in the fibrotic liver, exhibiting a downregulation of phagocytic gene signatures, representing a significant impairment of their reparative function. Instead of clearing debris and senescent cells, they become central instigators of the niche. They contribute to the inflammatory milieu by secreting their own pro-inflammatory and pro-fibrotic SASP, including CCL5 and MMP2, which further amplify the pathological environment (58). This pro-fibrotic, dysfunctional state makes senescent macrophages a pathological driver and a potential target for senolytic interventions, as their selective removal has been shown to ameliorate fibrosis (39).

Notably, the fibrotic niche is not composed solely of senescent cells; it is invariably defined by a deeply intertwined ecosystem of pro-fibrotic, non-senescent SAC populations. This distinction is crucial, as these SACs often derive from the same cellular lineages as their senescent counterparts, but have adopted a purely scar-promoting state, rather than a growth-arrested one. Landmark single-cell studies have identified three core populations: (1) Scar-Associated Macrophages (SAM) (52), (2) Scar-Associated Endothelial Cells (SAECs (VWF^＋ WNT9B^＋)) (67), and (3) a heterogeneous population of activated Myofibroblasts, namely activated HSCs and PDGFRα^＋ Portal Fibroblasts. Rather than existing in isolation, these populations co‑localize within fibrotic septa to form spatially organized niches. Among these, the SAM population is central to the imperfect repair process, with the OLR1^＋ subset showing the most pro‑fibrogenic program. This pathogenic population is highly heterogeneous, defined by a constellation of markers that include TREM2, CD9, SPP1, and OLR1 (68). This TREM2^＋ SAM population is often considered the scar-associated counterpart to the lipid-associated macrophages (LAMs; TREM2^＋ APOE^＋ GPNMB^＋ LGALS3^＋) also found in metabolic dysfunction-associated steatohepatitis (MASH) and MASLD; TREM2^＋ macrophage states adopt contrasting roles, with pro‑fibrogenic states and LAMs supporting reparative lipid‑clearing and ECM‑degrading functions during regression, underscoring context-dependent roles (69-71).

Functionally, this SAC population constitutes an integrated pathogenic hub that is maintained by maladaptive intercellular crosstalk within the fibrotic niche. Although the regulation of tissue repair involves a multifaceted signaling network, SACs sustain imperfect repair through three major interacting mechanisms : (1) ADAM17 activation by TNF–α and IL−1β causes TREM2 shedding, reducing membrane TREM2–mediated efferocytosis and lipid handling, and biasing macrophages toward pathological SAM states (72); (2) loss of pro‑resolution SAEC−HSC crosstalk (for example, Wnt9b–Sfrp2 disruption in fibrotic septa) (67); and (3) reduced local stromal IL−6 restraint that normally limits differentiation toward CD9^＋TREM2^＋ SAM‑like states (73). These changes reinforce myofibroblast activation through SAM‑derived PDGFB and TGF−β1, consolidating a pro‑fibrogenic circuit.

Thus, the fibrotic niche is now defined as a complex cellular ecosystem that is characterized by the co-existence of persistent senescent cells and diverse SAC populations. This ecosystem, sustained by pathogenic feed-forward loops, and the active dismantling of resolution pathways, represents the functional and structural manifestation of imperfect repair. Having identified these key cellular players (the ‘who’), the critical next step is to understand their spatial context (the ‘where’). The liver’s intricate metabolic architecture, governed by hepatic zonation, provides the spatial blueprint that dictates where and how these pathological niches first emerge.

THE SPATIAL DIMENSION OF IMPERFECT REPAIR: ZONATION, VULNERABILITY, AND NICHE DEFINITION

While the liver lobule may appear histologically uniform, its function is governed by an intricate spatial division of labor known as hepatic zonation. This architecture is not static, but is actively patterned by the hemodynamic gradient of blood flowing from the portal triad (periportal, Zone 1) to the central vein (pericentral, Zone 3). As blood traverses this axis, steep gradients of oxygen, nutrients, and gut-derived factors, including microbial products, are generated (74). In turn, this microenvironmental gradient orchestrates the differential expression of key signaling pathways—most notably the Wnt/β–catenin pathway. The activity of this pathway is established and maintained by HSCs, particularly pericentral HSCs acting as a niche component, which secrete R–spondin 3 (RSPO3) to drive high Wnt activity, specifically in adjacent Zone 3 hepatocytes (75). This mechanism results in a steep functional gradient where the pathway is actively suppressed in Zone 1 while highly active in Zone 3, thereby imparting distinct, often opposing, metabolic identities to hepatocytes in each zone. The oxygen-rich periportal zone (Zone 1), marked by Gpc, is specialized for oxidative processes, including gluconeogenesis, the urea cycle (Cps1, Arg1), and amino acid catabolism. In contrast, the hypoxic pericentral zone (Zone 3) handles glycolysis, lipogenesis, and xenobiotic (drug) metabolism, marked by Cyp family enzymes like Cyp2e1, and is uniquely defined by this high Wnt activity, and its target, glutamine synthetase (Glul) (76).

However, this intricate metabolic blueprint is highly vulnerable to the aging process, manifesting as a zonal drift or collapse, where the sharp, defined borders between metabolic zones become blurred (77). Specifically, this drift involves two consistent features across mouse and human datasets. First, the pericentral (Zone 3) identity expands and diffuses, as evidenced by an increased glutamine synthetase (GS)-marked territory, and the emergence of bi-zonal (ASS1^＋-GS^＋ co-expressing) hepatocytes (77). Second, the periportal (Zone 1) compartment contracts; however, this zone simultaneously undergoes the most extensive transcriptional remodeling, harboring the largest number of age-related differentially expressed genes (DEGs) (78). This disruption of spatial identity is not confined to the parenchyma; LSEC zonation is also compromised by aging, exemplified by the age-dependent inactivation of pericentral endothelium-derived C–kit, an early, spatially-defined driver of pathology (79).

This evidence of zonal collapse raises a fundamental question: what is the underlying insult that initiates and persistently drives this spatial disorganization? We posit that this driver is the imperfect repair process itself, which in turn establishes fibrotic niches. However, identifying the origin of these events remains a major methodological challenge. First, the precise insult initiating imperfect repair is poorly understood, making it inherently difficult to define a discrete start amidst the slow, continuous progression of aging (5). This challenge in defining a start is compounded by the limitations of conventional study designs, which typically rely on cross-sectional snapshots at a few discrete time points. While this approach is often well-suited to capturing the established outcomes of pathology, it struggles to capture this poorly defined initiation phase. Second, the initial imperfect repair niche is likely a rare, highly localized spatial event. Even with modern high-resolution spatial transcriptomics, identifying such a rare niche de novo is technically challenging (80); standard analysis pipelines are not designed to detect these subtle, nascent events, which are often obscured by technical bias and a lack of a priori new knowledge of the niche’s precise features. This methodological gap has limited our understanding of how and where the age-related imperfect repair niche is first established.

This methodological gap highlights the need for new approaches that can capture traces of imperfect repair, particularly those that can isolate these rare, physically defined niches. We recently developed Fibrotic Niche enrichment Sequencing (FiNi–seq), a quasi-spatial method that physically enriches for the niche based on tissue stiffness (81). This enrichment revealed that the fibrotic niche in aged mice is characterized by an enrichment of SACs, p21＋ senescent ECs, Smoc1＋ fibroblasts, and Exhausted T–cells; subsequent spatial validation localized this niche to the periportal region. The identification of this established scar-associated niche in aged, yet non-cirrhotic, livers is notable, as it suggests imperfect repair can establish a persistent pathological environment without overt disease. Furthermore, in studies of overt fibrotic disease, this same periportal region has been independently identified as a critical fibrogenic hub. This spatial convergence raises a critical question: whether the mechanisms driving clinically apparent disease leverage or amplify the same imperfect repair processes initiated sub-clinically during aging (Fig. 2).

Fig. 2.

Dynamic changes in the hepatic cellular microenvironment. (A) The healthy young liver maintains homeostasis, with resident Kupffer cells and quiescent fibroblasts in the space of Disse. (B) The response to acute injury in a young liver (‘damaged young liver’) involves the recruitment of immune cells (macrophages, T–cells), and the activation of reparative Wif1＋ fibroblasts. (C) The aged-damaged liver microenvironment includes monocyte-derived macrophages, PD−1＋ T–cells (interacting with PD−L1 on senescent endothelial cells (ECs)), and Smoc1＋ fibroblasts. (D) Advanced fibrosis is characterized by a fibrotic niche composed of Trem2＋ scar-associated macrophages, VWF＋WNT9B＋ scar-associated ECs, and scar-associated fibroblasts, which promote irreversible ECM deposition. This figure was created with BioRender.com.

The FiNi–seq study also profiled niches from young mice, revealing that nascent fibrotic niches are present even early in life. However, these young niches were compositionally distinct; while characterized by Wif1＋ fibroblasts, they notably lacked the full pathogenic assembly (e.g., p21＋ senescent ECs and Exhausted T–cells) observed in the aged niche (Fig. 2). This comparison provides a distinct temporal snapshot, suggesting a progression of imperfect repair from a basal, perhaps still reversible, state, into a persistent, non-resolving pathological phenotype. Resolving this spatial organization and understanding the signaling pathways that govern the initiation of the young niche and its progression into the aged, pathogenic state is therefore a key future direction.

SIGNALING HUBS THAT LOCK-IN IMPERFECT REPAIR

Recent investigations into fibrotic mechanisms have solidified the consensus around a set of core signaling hubs that govern pathological progression across different organs. These pathways form an interconnected network that translates chronic tissue damage, such as the persistent metabolic and hypoxic stress, into pathological cellular responses, like myofibroblast activation, unresolved inflammation, and ECM deposition (82). Among these core hubs, the TGF−β signaling pathway serves as the master regulator of the fibrotic response. This pathway is intricately linked to mechanotransduction pathways, particularly the YAP/TAZ pathway, which senses and responds to complex changes in the matrix mechanics of the niche, including viscoelasticity, as well as stiffness (83). Complementing these are key developmental and inflammatory pathways, such as WNT signaling, which dictates metabolic and spatial reprogramming, and Notch signaling, which drives unresolved inflammation and pathological cell fate decisions. Finally, pathways like Hedgehog (HH) signaling, fibroblast Growth Factor (FGF) signaling, and bone morphometric proteins (BMP) signaling act as critical context-dependent modulators that determine whether repair resolves, or persists.

Indeed, dysregulation of these core pathways begins during the aging process, long before overt disease. Although compared to advanced fibrosis, the aged tissue environment is phenotypically mild, this pathogenic program is already activated by specific cellular drivers, in particular senescent cells. The heterogeneous SASP from these cells is known to include a complex cocktail of pro-fibrotic factors, including TGF−β ligands, WNT proteins, and Notch ligands, thereby sustaining a low-level, chronic activation of these core pathways. Concurrently, aging itself has been shown to directly dysregulate key developmental pathways, such as HH signaling, which impairs liver resiliency and regenerative capacity. However, paradoxically, a primary dysregulation observed in the general aged stroma is not the hyper-activation of all pathways, but rather a critical decline in the activity of the mechanosensors YAP/TAZ (84). This is in contrast to the YAP/TAZ hyper-activation seen in stiff, advanced fibrotic lesions (85). This paradoxical decline in YAP/TAZ activity—which promotes cGAS−STING-driven inflammation, accelerates senescence, and creates a feed-forward loop where the resulting SASP directly promotes pathological ECM deposition by stromal cells—may represent the critical initiating event that creates the imperfect repair state, establishing the mosaic of pathogenic niches.

However, understanding the initiation of fibrosis within this subtle, rare imperfect repair niche presents a unique challenge. While single-cell omics technologies are effective to identify cellular composition and predict pathway activity from gene expression, this molecular-level prediction alone often cannot capture the full biochemical activation and functional consequences in a sub-clinical state (80). Furthermore, the inherent scarcity and transient nature of this pathogenic niche make direct biochemical validation via gold-standard methods (e.g., Western blotting, functional assays) challenging. Therefore, to derive comprehensive insight into the birth and progression of this aged niche, it is essential to integrate and cross-analyze the predictive power of omics signatures with the robust, biochemically validated mechanisms established in advanced disease (Fig. 3). Building upon this comparative framework, we now focus on these representative signaling pathways that act as the master-switches and stabilizers of the fibrotic niche ecosystem.

Fig. 3.

Major signaling pathways implicated in hepatic fibrosis. A summary heatmap illustrating the general effects of key signaling pathways (X–axis) on different processes of fibrosis pathogenesis (Y–axis). The color scale indicates whether a pathway is generally anti-fibrotic (yellow), pro-fibrotic (red), or has mixed/context–dependent effects (orange).

TGF−β signaling: Master regulator of liver fibrosis

As a multifunctional cytokine, TGF−β regulates a wide range of cellular processes, such as proliferation, migration, and epithelial to mesenchymal transition (EMT), and holds pivotal roles in tissue homeostasis, oncogenesis, and immune responses (86). When TGF−β homodimer binds to two molecules of its high-affinity receptor TGFβRII, a constitutively active receptor serine/threonine kinase, it recruits two molecules of TGFβRI, forming a heterotetramer complex. In the canonical (SMAD–dependent) pathway, the phosphorylation of TGFβRI by TGFβRII exposes its kinase domain, allowing SMAD2/3 to interact and be phosphorylated. Phosphorylated SMAD2/3 detaches from TGFβRI and binds to SMAD4 to form a heterotrimer complex, which translocates into the nucleus and induces the transcription of target genes, such as Serpine1, Acta2, and Col1a1. In the non-canonical (SMAD–independent) pathway, TGF−β signaling activates other cascades, including MAPKs, PI3K/AKT, and Rho GTPases, which are involved in cell proliferation, migration, and survival (87).

Upon repetitive hepatocyte injuries, unresolved scars recruit SAM, which secrete TGF−β to activate HSCs (88, 89). Upon TGF−β induction, HSC activates the proliferation cycle via the SMAD–independent pathway, and expresses α–SMA and fibrillar collagens via the SMAD–dependent pathway to transdifferentiate HSCs into the myofibroblast. Meanwhile, hepatic injuries attract and activate immune cells, such as Kupffer cells and circulating macrophages, to generate an inflammatory environment. Inflamed macrophages produce reactive oxygen species (ROS) and activate HSCs (90). To resolve the inflammation, immune cells secrete TGF−β, the immune breaker, which induces macrophage differentiation toward the immunomodulatory M2 type (91). However, the persistent secretion of TGF−β in the chronic inflammatory environment further activates HSCs and drives the fibrotic processes.

TGF-β acts not only on HSCs and macrophages but also on hepatocytes. During hepatic fibrosis, hepatocytes lose epithelial properties and acquire mesenchymal cell features (92). This EMT expands the pool of fibroblast-like cells, which actively secrete fibrotic matrix into the extracellular space (93). Furthermore, TGF-β shifts the homeostatic balance of the ECM by reducing the expression of MMPs while enhancing the expression of TIMPs, thereby driving net ECM accumulation (94).

Taken together, within the fibrotic niche, sustained TGF-β signaling is primarily propagated by persistent senescent cells and SAMs. This signaling hub acts pleiotropically to stabilize the fibrotic ECM scaffold: it drives quiescent HSCs into a proliferative, myofibroblastic state; polarizes macrophages toward a pro-fibrotic phenotype; and induces EMT in hepatocytes. Collectively, these coordinated actions lock the tissue into a self-perpetuating cycle of imperfect repair.

Hippo signaling: Mechanotransducer of liver fibrosis

The Hippo signaling pathway plays a central role as a mechano-sensing signal that transforms the information about ECM stiffness to the tension of the actomyosin cytoskeleton. The core components of Hippo signaling are MST1/2 and LATS1/2 kinases, which phosphorylate and inactivate the transcriptional co-activators YAP and TAZ. When Hippo signaling is on, YAP/TAZ are phosphorylated, retained in the cytoplasm, and degraded by the proteasome. When Hippo signaling is off, YAP/TAZ are dephosphorylated, translocated into the nucleus, and bound to the TEAD family of transcription factors to induce the expression of target genes, such as Ccn2, and Ankrd1, which are involved in cell proliferation, survival, and differentiation.

The activity of YAP/TAZ is tightly regulated by the mechanical properties of the ECM. In the fibrotic liver, the stiff ECM activates YAP/TAZ in HSCs, which in turn promotes the expression of fibrotic genes, such as Acta2 and Col1a1. Myofibroblast–specific knockout or overexpression of YAP/TAZ attenuates or exacerbates liver fibrosis, respectively, highlighting the role of Hippo signaling in hepatic fibrosis (95). Thus, Hippo signaling functions as a mechanotransductive integrator of the fibrotic niche, coupling pathological ECM stiffness to sustained myofibroblast activation and senescence-associated inflammatory signaling.

Hedgehog (HH) signaling: Profibrotic HSC activator

Known for its crucial role in embryonic development, the HH signaling pathway also governs adult tissue homeostasis (96). During hepatic fibrosis, hedgehog signaling mainly performs two roles: HSC expansion, and metabolic reprogramming in activated HSCs. Chronic liver injuries induce HH expression in hepatocytes, as well as in activated HSCs. These paracrine or autocrine HHs activate hedgehog signaling in activated HSCs (97, 98), leading to expansion of HSCs and transdifferentiation to ECM-secreting myofibroblasts. HH and YAP boost metabolic reprogramming (glutaminolysis) in the activated HSCs (99). They regulate the expression of glutaminase, the rate-limiting enzyme in glutaminolysis, thereby producing alpha–ketoglutarate to provide the energy for HSCs to proliferate and transdifferentiate to myofibroblasts.

BMP signaling: Context-dependent modulator

The BMP signaling pathway, a critical component of the TGF−β superfamily, is involved in numerous biological processes that include organ fibrosis. In hepatic fibrosis, BMP signaling shows both pro- and antifibrotic features, depending on the context. BMP1/2/4 have shown the HSC–activating effect. The activated HSCs secrete BMP−1, and function in hepatocytes to activate EGFR signaling, inducing hepatocyte EMT (100). BMP−2 significantly increased α –SMA level in HSC without affecting its proliferation, suggesting its role in the transdifferentiation of HSC to myofibroblast (101). BMP−4 expression is induced upon liver injuries, and activates HSCs through both the canonical and non-canonical BMP signaling pathways (102). In contrast, BMP6/7/9 have shown the anti-fibrotic effect. The expression of BMP6 is increased upon hepatic steatosis and inhibits HSC activation, thereby reducing proinflammatory cytokine and matrix protein secretion (103). BMP−7 protein level shows the increase in the first 2 days upon CCl4–induced liver injury, but decreases after 2 weeks of hepatic injury. BMP−7 indeed does have anti-fibrotic activity by counteracting the effect of TGF–β on α–SMA and Col–I expression in HSC. However, TGF–β blocks the expression of BMP−7. Since TGF–β expression keeps augmenting during CCl4–induced liver injury, it can be postulated that in the initial phase of hepatic injury, secreted BMP−7 tried to counteract TGF–β signaling; but as TGF–β expression exceeded the threshold, BMP−7 expression was inhibited, and HSC activation accelerated the fibrotic procedure (104). Within the fibrotic niche, BMP signaling exhibits strong context dependency, reflecting its regulation by distinct cellular components. Thus, BMP signaling functions as a niche-modulated pathway in which the balance between pro- and anti-fibrotic outputs is dictated by the relative abundance and functional state of senescent cells, activated fibroblasts, and the surrounding ECM.

FGF signaling: Diverse roles and functions

FGF signaling orchestrates fundamental biological processes, from embryonic development to adult tissue homeostasis and repair. In liver fibrosis, FGF ligands are secreted by multiple cell types, and control liver fibrosis directly or indirectly. FGF2 helps the progression of fibrosis by inducing the proliferation of HSC, as well as ColI and α–SMA production in HSC (105). FGF18 is mainly produced in HSCs, and FGFR3C, its primary receptor, is also expressed in HSCs. This autocrine FGF signal transduction induces Ccnd1 expression, leading to the proliferation of HSCs. Moreover, FGF18 promotes profibrotic gene expression, such as Col1a1, Col3a1, and Acta2, resulting in liver fibrosis (106). In an acute liver injury mouse model, exogenous FGF7 acts in hepatocytes and promotes their survival via PI3K/AKT and RAS/MAPK pathways. Increased survival of hepatocytes leads to the reduced infiltration of macrophages, thereby mitigating inflammation and preventing HSC from activation (107). The ectopic administration of FGF21 also blocks HSC activation via reducing TGF–β gene expression and the phosphorylation level of SMAD2/3 (108). Moreover, FGF21 inhibits HSC activation by downregulating the expression of leptin, a profibrotic cytokine, followed by the inactivation of MAPK and STAT3, which is the downstream of leptin signaling (109).

FGF signaling operates as a complex regulator within the fibrotic niche, exerting divergent effects dictated by the ligand source. Pro-fibrotic FGFs, such as FGF2 and FGF18, are derived in part from activated HSCs and act to sustain myofibroblast proliferation, whereas ligands like FGF7 and FGF21 provide protective signals to hepatocytes or suppress HSC activation. This duality positions FGF signaling as a biphasic modulator, where the net impact on hepatic fibrosis is determined by the specific balance of ligands present in the injured microenvironment.

WNT signaling: Spatial and metabolic reprogrammer

WNT signaling pathway, known for its complexity, regulates cell proliferation, differentiation, and cell fate decisions (110). WNT signaling is also involved in liver zonation, metabolism, and regeneration. In the quiescent liver, canonical WNT signaling is high in the perivenous (central) zone and low in the periportal zone, which determines the zonal expression of metabolic enzymes (111). Upon liver injury, WNT signaling is activated, and promotes liver regeneration by inducing hepatocyte proliferation. However, its sustained activation in chronic liver injury contributes to hepatic fibrosis by promoting the activation of HSCs and the development of hepatocellular carcinoma (HCC). WNT signaling in HSCs induces the expression of fibrotic genes, such as Acta2 and Col1a1, and promotes their proliferation and survival (112).

Notch signaling: Inflammatory activator

Notch signaling facilitates cell–cell communication to regulate cell fate decisions. The binding of Notch ligands (Delta-like, Jagged) to Notch receptors (Notch1−4) induces the cleavage of the Notch receptor by ADAM and γ–secretase, releasing the Notch intracellular domain (NICD), which then translocates into the nucleus, and binds to the transcription factor RBP−J and Mastermind-like (MAML) co-activators. This complex activates the transcription of target genes such as Hes and Hey, regulating cellular differentiation programs, and generally maintaining undifferentiated states or promoting differentiation toward specific cell fates.

Notch signaling promotes hepatic fibrosis largely by two mechanisms: by activating HSCs, and provoking inflammation. The expression of Jagged–1 and Notch3 is strongly increased in myofibroblasts to promote the myofibroblastic markers, α–SMA and ColI (113). Notch signaling promotes HSC activation via its representative target gene, Hes1, which seems to increase the α–SMA promoter activity (114). Notch signaling can also be activated independent of its ligands. ROS released from the inflamed macrophages upon liver injury enhances the activity of ADAM17, which cleaves the Notch receptor to generate NICD (115). Moreover, Notch signaling contributes to hepatic fibrosis by generating an inflammatory milieu during the process. In macrophages during hepatic fibrotic progression, the activated Notch signaling induces NICD binding to RBP−J to repress the expression of CYLD, a negative regulator of NF–κB, thereby increasing the inflammation (116). In addition, Notch signaling induces the polarization of M1 macrophages and aggravates the inflammation. Notch signaling in HSCs derives M1 macrophages, while M1 macrophages, in turn, activate HSCs, forming a vicious cycle (117). Functioning as a juxtacrine amplifier within the fibrotic niche, Notch signaling reinforces these inflammatory macrophage–fibroblast interactions, thereby locking cells into a non-resolving state.

CONCLUSION AND PERSPECTIVE: TARGETING IMPERFECT REPAIR

This review has contextualized imperfect repair primarily within the aging liver, culminating in the spatial identification of the rare, pathogenic niches that orchestrate this process. However, the fundamental principles—the failure of wound resolution, the paradoxical dysregulation of core signaling hubs, the persistence of senescent cells, and the establishment of a self-perpetuating pathological ecosystem—are not unique to the liver. Indeed, fibrosis is a conserved pathophysiological process that drives progressive dysfunction across the lung, kidney, gut, and skin. Regardless of the specific organ, this process is universally characterized by the persistent activation of myofibroblasts, the excessive deposition of ECM, and the dysregulation of the same core signaling hubs (e.g., TGF−β and YAP/TAZ) that create a self-perpetuating cycle. While the underlying grammar of damaging adaptation is likely universal, the specific manifestation of this imperfect repair is context dependent. This context-dependency is critical, as the precise pathological outcome is dictated by a complex interplay of local factors, including the inherent tissue micro-architecture and its differential vulnerability to insults, the intrinsic state and acquired epigenetic adaptations of the resident cells, and their continuous crosstalk. The FiNi–seq-defined periportal niche represents one such specific manifestation. Future cross-organ comparison of these niches will be an essential next step in validating this framework as a central mechanism of aging.

Understanding imperfect repair as the sub-clinical inception of a disease process, rather than its end-stage, has significant therapeutic implications. Current therapeutic strategies, particularly for MASH, are largely focused on resolving established inflammation and steatosis, or reversing overt, advanced fibrosis. The imperfect repair framework suggests a paradigm shift: a move towards niche management aimed at the initial traces of the sub-clinical wound. This strategy is critical, because advanced fibrosis presents significant barriers to intervention. As the disease progresses, excessive elastin accumulation, ECM crosslinking (e.g., by tissue transglutaminase), and subsequent compaction limit drug penetration, while also establishing a point of no return, where hepatocyte loss becomes irreversible.

Addressing this sub-clinical niche, especially as it progresses towards this irreversible threshold, demands a strategy beyond targeting single cell types. Conventional Senolytics, for example, function as a blunt instrument, unable to distinguish detrimental from beneficial cell populations. The therapeutic future lies in holistic niche engineering: simultaneously reprogramming the pathological activity of multiple cell types (e.g., via Senomorphics and YAP/TAZ activators), while dismantling the fibrotic scaffold. Achieving such complex niche reprogramming requires sophisticated, integrated therapeutic modalities. This requires advanced drug delivery systems, informed by high-resolution spatial and single-cell atlases, to ensure precision. Such tools include engineered nanoparticles or therapeutic exosomes (e.g., loaded with NF–κB inhibitors to target innate immune cells) (118), and Antibody–Drug Conjugates (ADCs) (e.g., designed to target myofibroblast–specific antigens like CD248) (119). Furthermore, integrating these delivery systems with next-generation payloads, such as Targeted Protein Degraders (TPDs) (120), offers a pathway to neutralize the core drivers of imperfect repair with potentially greater efficacy and safety than current cytotoxic approaches (121) . Ultimately, the spatial and molecular dissection of the imperfect repair niche provides a practical roadmap, moving beyond the management of symptoms to the precise resolution of the sub-clinical wounds that drive aging.