ADPKD as a Temporally Dynamic Disease: Implications for Biomarkers, Trial Design, and Therapeutic Targeting

Abstract

Autosomal dominant polycystic kidney disease (ADPKD) has traditionally been conceptualized as a structurally progressive disease, with cyst growth driving loss of kidney function. This view has led to the widespread use of total kidney volume for risk stratification and clinical trial design. However, longitudinal human studies and experimental models indicate that the relationship between structure, function, and underlying biology changes over the disease course. Earlier disease is characterized by prominent cystic, proliferative, and metabolic activity together with early inflammatory responses; inflammatory and injury-repair processes become increasingly prominent during intermediate disease, whereas fibrosis and loss of functional parenchyma assume a greater role later as kidney function declines. These processes overlap rather than define discrete stages, and their timing and relative contributions likely vary among individuals. This review proposes that ADPKD be considered as a temporally dynamic disease and examines the implications of this framework for biomarker development, clinical trial design, and therapeutic targeting. Such a framework emphasizes the need to complement structural measures with biomarkers of evolving biology, to identify biological transitions prospectively, and to align therapeutic targets and end points with the processes contributing meaningfully to progression in the population being studied.

ADPKD is one of the most common inherited kidney diseases and a major cause of kidney failure worldwide.^1,2 It is a lifelong disease process characterized by progressive cyst development and kidney enlargement, leading to a decline in kidney function later in the individual life.^3,4 Over the past 2 decades, the field has made substantial progress in defining structural correlates of progression. Longitudinal imaging studies established total kidney volume (TKV), particularly height-adjusted TKV (htTKV), as the dominant biomarker for risk stratification and patient enrichment in clinical trials.5, 6, 7, 8 The Consortium for Radiologic Imaging Studies of Polycystic Kidney Disease showed that kidney enlargement begins early, progresses throughout the disease course, and is strongly associated with later decline in kidney function.⁵ Subsequent work demonstrated that baseline htTKV and kidney growth rate are powerful predictors of chronic kidney disease progression and led to the development and validation of the use of imaging-based systems, such as the Mayo Imaging Classification (MIC).^6,9

These advances have been transformative and have appropriately placed structural progression at the center of clinical risk assessment in ADPKD. However, a predominantly structural model leaves several observations incompletely explained. Kidney function often remains preserved for decades despite substantial kidney enlargement, whereas later functional decline may accelerate even as structural growth slows.¹⁰ Patients with similar structural burden may follow different functional trajectories, and therapeutic effects on kidney growth do not necessarily translate into proportional preservation of kidney function. Moreover, although TKV remains the strongest prognostic imaging biomarker, it does not fully capture active disease biology or therapeutic response.^11,12

These discrepancies suggest that ADPKD progression is not fully captured by cumulative structural burden alone. Rather, multiple biological processes, including epithelial proliferation, metabolic reprogramming, inflammation, injury-repair responses, and fibrosis, contribute to disease progression, but their relative importance changes over time. The structural model, therefore, remains essential, but its clinical consequences may be better understood when structural change is interpreted in the context of the evolving biology of the kidney.

Why Does the Current Structural Model Remain Incomplete Even When TKV Performs Well?

The argument for a temporally dynamic model does not depend on diminishing the value of TKV. TKV is a well-validated prognostic biomarker; baseline htTKV and kidney growth rate are associated with subsequent chronic kidney disease progression, and age-adjusted htTKV underlies the MIC used for clinical risk stratification.^5,6,9 Long-term analyses further demonstrate that kidney function decline is nonlinear and strongly associated with MIC class.¹⁰

Its limitation is therefore not prognostic usefulness, but biological resolution. TKV primarily captures accumulated cystic burden and does not directly indicate which biological processes are the strongest contributors to progression at the time of assessment or whether these processes are changing before measurable structural effects occur. The fact that biochemical biomarkers can provide information beyond TKV further suggests that disease dimensions exist that are not captured by structural burden alone.^11,12

Recent multivariable approaches combining structural, clinical, and biochemical information reinforce this point—prediction can improve when complementary dimensions of disease are considered together.¹³ The key question is therefore not whether TKV predicts outcomes, but what biologically relevant information remains outside the structural measurement.

Evidence That ADPKD Progression Is Not Biologically Uniform Over Time

Although ADPKD has not been explicitly described as a temporally dynamic disease, multiple lines of evidence from developmental biology, clinical observation, longitudinal imaging, and molecular studies support a temporally dynamic model in which the relative contribution of different disease processes changes over time.

Developmental studies provided an early indication of this pattern. Piontek et al.¹⁴ showed that the effect of Pkd1 loss depends strongly on timing, with a developmental switch profoundly altering the kinetics of cyst formation. This established an important principle—biological timing can change disease behavior even with the same underlying disease-causing genetic variant.

Clinical and conceptual work also pointed in this direction early on. Chapman’s 2007 review, "Time for a change?”¹⁵, argued that advances in imaging and disease biology were already changing how ADPKD should be approached clinically. Grantham et al.³ further emphasized that cyst growth contributes to later kidney failure through secondary processes such as tubular obstruction, vascular compromise, and fibrosis, implying that the relative contribution of disease processes evolves over time rather than remaining constant.

Human longitudinal studies provide particularly strong support for this view. Analyses of long-term kidney function trajectories in ADPKD demonstrate that decline is nonlinear, with prolonged periods of relative stability followed by accelerated decline, and that the shape of those trajectories is strongly associated with baseline imaging class.¹⁰ These findings indicate that progression is not biologically uniform across adulthood and that the relationship between structure and function changes over time.

At the molecular level, studies of human tissue and experimental models also support the notion of temporal heterogeneity. Work from the Peters group identified strong enrichment of injury-repair biology across PKD expression profiles.¹⁶ Later work has highlighted the overlap between canonical renal injury–repair mechanisms and pathways involved in ADPKD progression. In this context, processes that normally support repair may become persistently activated, contributing to disease progression.¹⁷ Early-stage analyses tend to emphasize proliferative and metabolic programs,18, 19, 20, 21, 22 although inflammatory signaling is already detectable early in disease. With progression, inflammatory and injury-repair pathways become increasingly prominent, followed by greater extracellular matrix remodeling and fibrosis.^16,23, 24, 25 Although many of these datasets are cross-sectional, together they support a model in which the relative contribution of biological processes changes across the disease course.

In our recent longitudinal human–mouse study, we extend this concept by showing that kidney growth varies not only across ADPKD severity classes but also with age within the more severe groups, slowing even as functional decline continues. These patterns were accompanied by shifts in transcriptomic and metabolic programs.²⁶

An important implication of these observations is that differences in disease trajectory may begin to emerge much earlier in life than is typically appreciated. Longitudinal studies in children with ADPKD have demonstrated progressive cyst development and kidney enlargement while glomerular filtration rate (GFR) generally remains preserved, except in individuals with unusually severe early-onset disease.^27,28 Children with greater kidney enlargement or hypertension also show more rapid structural progression, indicating that heterogeneity in disease severity is already evident during childhood.^27,28 More recent pediatric imaging studies have similarly demonstrated substantial variability in age-adjusted kidney volume and have extended structural risk stratification to the pediatric population.²⁹ These observations complement adult longitudinal studies, in which kidney growth varies with age and eventually decelerates in more severe disease even as functional decline becomes increasingly apparent.^10,26 Taken together, the data suggest that substantial structural disease may accumulate during childhood, adolescence, and early adulthood, well before the onset of measurable loss of kidney function. Whether early structural acceleration determines later biological progression remains uncertain, but pediatric studies provide an important opportunity to define the earliest detectable divergence in disease trajectory and to identify biomarkers of active disease before functional decline occurs.

Across the available evidence, ADPKD progression does not appear to follow a biologically uniform course. Rather, multiple processes contribute throughout the disease, with their relative importance changing over time. The proposed temporally dynamic model is intended to organize these observations and examine their implications for disease assessment, biomarkers, and therapeutic targeting.

A Conceptual Framework: A Temporally Dynamic Model of ADPKD

Evidence from experimental models and longitudinal human studies supports a model in which multiple biological processes contribute across the course of ADPKD, but their relative prominence changes over time (Figure 1). For conceptual purposes, this evolution can be organized into 3 broad, overlapping patterns characterized predominantly by (1) proliferative biology, (2) inflammatory and/or remodeling biology, and (3) fibrotic and/or decline-associated biology. These patterns are not intended to represent discrete biological stages, and the underlying processes coexist and overlap throughout disease progression.

Figure 1.

Conceptual model of temporally dynamic ADPKD progression and its implications for biomarker development and therapeutic targeting. Schematic representation of the proposed temporal evolution of autosomal dominant polycystic kidney disease (ADPKD). The model illustrates three broad, overlapping periods characterized by changing relative contributions of proliferative, inflammatory/remodeling, metabolic, vascular, and fibrotic processes rather than discrete biological stages. During early disease, epithelial proliferation, cyst development and expansion, metabolic reprogramming, growth-promoting pathways, and early inflammatory responses are active, resulting in rapid increases in total kidney volume (TKV) despite preserved glomerular filtration rate (GFR). During the intermediate period, inflammatory signaling, adaptive immune activation, injury-repair pathways, and extracellular matrix remodeling become increasingly prominent, while proliferative and metabolic processes remain active. Later in the disease course, inflammatory and repair-related signaling persist, but fibrosis, loss of functional parenchyma, microenvironmental disruption, and declining metabolic capacity become increasingly important contributors to functional decline. The solid curves illustrate conceptual trajectories of major disease features. TKV (solid blue line) increases rapidly earlier in disease and may continue to increase despite relative deceleration of growth later. GFR (solid green line) remains relatively preserved during early disease but declines as injury, remodeling, and fibrosis accumulate. Fibrosis (solid red line) is limited earlier in disease but increases progressively and becomes particularly prominent later. The dashed curves represent hypothetical trajectories of temporally informative biomarkers rather than established biomarkers. They illustrate the type of progressive increase or decrease that an ideal biomarker might exhibit if it tracked an evolving biological process across the disease course. Such trajectories are conceptual and may not exist as single measurable analytes. In practice, a limited combination of complementary structural, physiological, molecular, and functional biomarkers may be required to estimate the relative contribution of different biological processes at a given time and to identify processes that are disproportionately altered relative to structural burden or kidney function. The lower conceptual framework illustrates the translational implications of this model. Structural burden should be interpreted together with selected biological and physiological measurements to generate a multidimensional assessment of current disease state and anticipated trajectory. Therapeutic targeting should similarly be aligned with biological processes that meaningfully contribute to progression in the population or individual being treated, although the optimal window for modifying those processes remains to be defined. Importantly, the time at which a biological process becomes detectable, the period during which it contributes most strongly to progression, and the optimal therapeutic window for modifying that process may not coincide. Preventive intervention may therefore need to precede the period of greatest phenotypic expression of the targeted pathology. The timing of the illustrated periods, curves, and transitions is conceptual rather than absolute, and the underlying processes coexist and overlap throughout disease progression. Their timing, magnitude, and persistence are expected to vary among individuals as a consequence of genotype, genetic and nongenetic modifiers, environmental exposures, comorbidities, and other factors. The framework is intended to generate hypotheses for prospective testing rather than establish stage-specific biomarker or treatment recommendations.

The early proliferative period is characterized by epithelial proliferation, cyst development, and expansion, together with activation of growth-related signaling pathways such as cyclic adenosine monophosphate signaling, metabolic reprogramming, and early immune responses, whereas renal function remains preserved.^2,18, 19, 20, 21, 22^,26 Structural growth is most prominent during this period, and the disease may be more responsive to interventions targeting epithelial proliferation, cyst development, and expansion.

During the intermediate period, structural growth begins to decelerate relative to early disease, whereas inflammatory signaling becomes more prominent and is accompanied by emerging adaptive immune activation and increasing activity of injury-repair pathways, including extracellular matrix remodeling.^{16,17,23,24,26} Early microscopic evidence of fibrosis may emerge during this period, often while kidney function remains relatively preserved. Proliferative and metabolic processes remain active, but their relative contribution to disease progression may decrease as inflammatory and remodeling processes assume a greater role. This period may therefore represent an important transition between predominantly cystic development and the increasing contribution of processes associated with fibrotic remodeling and functional decline.

Later in the disease course, incremental structural growth slows further. Inflammatory and repair-related signaling persist, but fibrosis, loss of functional parenchyma, and microenvironmental disruption increasingly contribute to functional decline. Metabolic activity declines, and kidney function deteriorates more rapidly. Processes such as matrix remodeling, tubular dysfunction, and alterations in the local microenvironment are increasingly important drivers of progression.^16,23, 24, 25, 26^,30,31 In this context, therapies directed solely at cyst expansion may have diminishing effects unless these downstream processes are also targeted.

These processes are not confined to fixed compartments, and their relative contributions are expected to vary both over time and among individuals. Genetics is likely to be an important determinant of this variability. Pathogenic variants in PKD1 are generally associated with more severe kidney disease than pathogenic variants in PKD2, whereas differences among pathogenic PKD1 variants further contribute to variability in disease severity.^32,33 Genetic background also contributes to phenotypic variability beyond the primary disease-causing variant.^33,34 The primary genotype and additional genetic modifiers may therefore influence the timing, magnitude, or persistence of the biological processes that contribute to progression. Modifier effects may operate at different levels of disease biology, including the expression or functional consequences of the polycystin defect, proximal signaling pathways, or processes such as proliferation, metabolic dysfunction, inflammation, vascular injury, and fibrotic remodeling. Environmental exposures and comorbidities may similarly modify these processes and contribute to interindividual differences in their timing and severity. The temporal course of ADPKD can therefore be viewed as the product of the underlying polycystin defect interacting with genetic and nongenetic modifiers that shape how individual disease processes are expressed over time.

A temporally dynamic model of ADPKD should not be understood simply as the observation that “the disease changes over time,” which is true of nearly every chronic disease. It proposes that the relative contribution of pathophysiologic processes is not constant—proliferative programs and cyst initiation and expansion may contribute more strongly early on; inflammatory and repair responses increasingly accompany progression; and fibrosis and loss of functional parenchyma become progressively more consequential for functional decline. Because these processes overlap and evolve, the biological meaning of a biomarker or therapeutic target may also change over the disease course.

This type of evolution helps explain why a single snapshot can be misleading. A biomarker that may be highly informative during an early proliferative period may lose explanatory power later, not because it was “wrong,” but because the biological context has changed. Similarly, therapies targeting proliferation may be more effective when epithelial expansion contributes strongly to disease progression but less effective as fibrosis and remodeling assume a greater role in functional decline. This temporal dependence may help explain why some therapies show effects on TKV growth without proportional changes in long-term kidney function, and why certain biochemical markers correlate imperfectly with TKV as the disease advances. It also suggests that disagreement between markers should not automatically be interpreted as biomarker failure; in some cases, it may instead indicate that the markers are capturing different biological dimensions of the disease.

The structural model of ADPKD has been extraordinarily useful, but it has also shaped how the disease is conceptualized. By emphasizing cyst growth and kidney enlargement, it may favor an interpretation of progression as predominantly continuous structural accumulation. However, longitudinal human studies and experimental models indicate that this interpretation does not fully capture the evolving relationship among structure, function, and underlying biology.

Taken together, these observations support a temporally dynamic model in which structural burden reflects accumulated cystic disease, whereas the relative contributions of the biological processes driving progression and functional decline evolve over time.

Testable Predictions of a Temporally Dynamic Model

For a temporally dynamic model to be useful, it should generate predictions that can be tested prospectively and potentially falsified. Several observations would support the model proposed here.

First, biomarkers reflecting different biological processes should not be expected to change uniformly over the disease course. Their trajectories and relationships with TKV and kidney function should shift as the relative contributions of the underlying processes change. This possibility is already suggested by longitudinal ADPKD biomarker studies in which metabolic biomarkers associated with disease severity cross-sectionally did not necessarily track structural progression within individuals over time.³⁵ Conversely, the model would be weakened if markers representing different biological processes maintain stable relationships with one another, TKV, and kidney function throughout the disease course and simply change in parallel with increasing disease severity.

Second, similar age and structural burden should not necessarily imply equivalent underlying biological or physiological states. Patients with comparable htTKV or MIC classification may differ in cystic architecture, parenchymal integrity, fibrosis, microvascular function, inflammation, metabolism, or other processes not fully captured by structural imaging. If these differences are biologically important, measurements reflecting these processes should account for part of the residual variability in subsequent functional decline among patients with similar structural classification. This prediction would be weakened if physiological, molecular, or tissue-level measurements fail to provide meaningful information about subsequent progression beyond established clinical, structural, and genetic predictors.

Third, longitudinal biomarker trajectories should provide the strongest evidence of temporal biological change, but a clinically useful framework cannot depend exclusively on serial sampling. A complementary approach would be to identify a limited set of well-validated biomarkers representing distinct biological domains and interpret them jointly with structural and functional measures. Rather than assigning disease state based on any single biomarker, a multidimensional biological profile could estimate the relative contributions of different processes at the time of assessment and identify processes that appear disproportionately altered relative to the patient's structural burden and functional state. Such profiles may be particularly informative in individuals whose functional course deviates from conventional structural predictions. Candidate biomarkers will also differ in temporal informativeness, and their value within a multidimensional profile may depend not only on how they change over time but also on the biological dimension they represent and the information they provide relative to other structural, functional, and biological measurements. Experimental models in which molecular profiles can be examined across defined disease states may therefore help prioritize candidate biomarkers for subsequent human validation. Ultimately, prospective longitudinal studies would be required to determine whether such cross-sectional profiles can reliably estimate the current biological state and improve the prediction of subsequent trajectories.

Fourth, therapeutic efficacy should depend, in part, on whether the targeted biological process is an important contributor to progression at the time of intervention. Therapies directed at cystogenesis, proliferation, metabolism, inflammation, vascular dysfunction, or fibrosis should therefore not be expected to have uniform effects across the disease course. A treatment may have greater efficacy when its target makes a substantial contribution to current disease activity and less effect once other processes have become more important determinants of progression. This prediction could be tested through treatment-by-disease-state interactions in clinical or experimental studies and would be weakened if therapeutic effects remain independent of biological context and timing of intervention.

Together, these predictions distinguish a temporally dynamic model from a simpler model in which the relative contribution of biological processes remains largely constant over time and progression primarily reflects increasing cystic burden and disease severity. Testing these predictions will require longitudinal studies that integrate structural, functional, molecular, and physiological measurements across a sufficiently broad spectrum of disease.

An important implication of this framework is that the timing of biological activity, biomarker informativeness, and therapeutic opportunity may not coincide. A biological process may become abnormal before its structural or functional consequences are apparent, and the optimal opportunity to modify that process may precede the period in which its phenotypic effects are greatest. Conversely, a biomarker may become most informative only after the underlying process has been active for some time. Figure 1 illustrates these distinctions conceptually rather than assigning specific biomarkers or therapies to fixed disease periods. These temporal relationships can begin to be defined by integrating existing longitudinal cohorts, experimental models sampled across disease states, and prospective studies incorporating complementary structural, functional, molecular, and physiological measurements. Longitudinal observations will be particularly informative, but cross-sectional multidimensional profiles may also help identify candidate biological states for subsequent validation. Together, these approaches can help determine when individual processes become active, when they contribute meaningfully to progression, when they can be reliably detected, and when they remain therapeutically modifiable.

Lessons From Other Diseases: Timing Changes Meaning

The temporally dynamic model proposed for ADPKD reflects a broader principle observed across chronic diseases—the relative contribution of biological processes can change during progression, altering the meaning of biomarkers and therapeutic targets.^36,37 Oncology provides a familiar example. Clonal evolution, stromal interactions, immune escape, and treatment resistance acquire different relevance during tumor evolution, such that therapeutic efficacy depends partly on the biological context in which a target is engaged.³⁸

Similar temporal changes occur outside oncology. In chronic kidney disease, compensatory and repair responses can evolve toward persistent inflammation, fibrosis, and tubular atrophy,^39,40 whereas sepsis can transition from hyperinflammation toward immune suppression, with major implications for therapeutic timing.⁴¹ These examples differ fundamentally from ADPKD, but they illustrate the same principle that a process may be adaptive, pathogenic, measurable, or therapeutically relevant at different points in a disease trajectory.

Alzheimer disease provides a particularly relevant biomarker example. Dynamic models propose that biomarkers of distinct pathophysiologic processes become abnormal at different times, with early molecular abnormalities preceding later neurodegeneration, structural atrophy, and clinical decline.^42,43 Their trajectories are nonlinear, and substantial biological change can occur while clinical function remains preserved. This framework demonstrates how biomarkers can be interpreted as temporally ordered indicators of different biological processes rather than parallel measures of overall disease severity, a concept directly relevant to ADPKD.

General Adaptation Syndrome as a Biological Analogy

The General Adaptation Syndrome, described by Hans Selye, provides a useful, though imperfect, analogy for this framework. Persistent stress was conceptualized as eliciting alarm, resistance, and eventually exhaustion responses.^44,45 In ADPKD, the persistent stressor is cyst development and expansion arising from the underlying polycystin defect. Cystogenic programs and metabolic reprogramming support early growth, while inflammatory, immune, and cellular stress responses emerge in parallel. With persistent structural and microenvironmental stress, inflammatory and repair-like responses are sustained, and fibrosis and functional decline eventually become increasingly prominent.

The value of the analogy lies in illustrating how the biological meaning of a pathway can change over time. Responses initially engaged to support adaptation or survival may become maladaptive, insufficient, or exhausted under persistent stress.^46,47 This may be particularly relevant to metabolic and oxidative stress pathways, where early compensatory activation could later give way to loss of adaptive capacity. Thus, the same pathway may not represent the same biomarker or therapeutic opportunity throughout the disease course.

Dynamic Metabolic Abnormalities Across the Disease Course

Metabolic abnormalities are among the strongest arguments against a static view of ADPKD. However, it is important not to oversimplify them as belonging only to “early disease” or “late disease.” A more accurate interpretation is that metabolic dysregulation is present throughout the disease course, but its form and function change over time.

Early disease is associated with metabolic reprogramming that supports proliferation, cyst growth, and biosynthetic demand.^18,20 Changes in mitochondrial content, tricarboxylic acid-cycle intermediates, and bioenergetic handling are compatible with a state of increased cellular proliferation and growth, consistent with a shift toward glycolytic and anabolic metabolism observed in cystic epithelia.^19,21,26,48 At a conceptual level, this resembles the metabolic reprogramming observed in proliferative conditions such as cancer, in which metabolism is reorganized to support growth and stress tolerance.³⁸

A particularly relevant example comes from fumarate hydratase-deficient associated renal neoplasia, in which fumarate accumulation promotes activation of NRF2-dependent antioxidant responses and widespread metabolic rewiring.^49,50 Experimental models of fumarate hydratase deficiency also develop proliferative renal cysts, underscoring the biological overlap among metabolic reprogramming, cystic growth, and stress-response signaling in proliferative renal disease.^22,50 Although ADPKD is biologically distinct from fumarate hydratase-deficient neoplasia, these observations reinforce the concept that metabolic and oxidative stress pathways can support proliferative growth and cellular survival in chronically stressed tissues.

With disease progression, however, the metabolic programs that initially supported cyst growth and proliferation appear to become less efficient and increasingly maladaptive. Mitochondrial abnormalities become more pronounced, cellular energy production declines, and the balance between adenosine triphosphate to adenosine diphosphate worsens. Metabolic processes that initially supported proliferation and growth may shift toward energetic exhaustion and impaired repair, reflecting progressive mitochondrial dysfunction and loss of metabolic flexibility.²⁶ In this latter setting, metabolic abnormalities are not merely markers of proliferation; they may become markers of declining cellular resilience, fibrosis, and impending organ failure, a transition from proliferative metabolic reprogramming to metabolic insufficiency.

This temporal distinction has biomarker implications. Metabolites associated with proliferative metabolic reprogramming earlier in disease may not retain the same meaning as mitochondrial dysfunction becomes more pronounced and metabolic flexibility declines with progression. Longitudinal observations further suggest that metabolic biomarkers associated with disease severity cross-sectionally may not necessarily track disease progression in the same manner within individuals.³⁵ Metabolism should therefore be viewed not as an exclusively early or late phenomenon, but as a continuously present process whose configuration and biomarker implications change with disease progression.

Repair Pathways and Maladaptive Repair in ADPKD

Repair biology provides a second major pillar of the temporally dynamic model. In many chronic diseases, pathways that would normally support repair can become drivers of maladaptive remodeling when persistently activated. ADPKD appears to fit this pattern closely.

This concept is supported by the “futile repair” hypothesis proposed in the context of mechanistic target of rapamycin (mTOR) signaling in ADPKD. In normal kidneys, mTOR activity is low under baseline conditions but is transiently activated in response to injury to support proliferative and hypertrophic repair processes. In ADPKD, disruption of polycystin signaling leads to activation of mTOR, resulting in a persistent repair-like state in the absence of a resolving injury.³⁰ In this model, cystic epithelial cells behave as if they are continuously responding to injury, leading to sustained proliferation, remodeling, and cyst growth, a form of maladaptive or “futile” repair.³⁰

Subsequent transcriptomic and integrative analyses have reinforced this concept. Meta-analysis of polycystic kidney disease expression profiles identified strong involvement of injury-repair programs, including pathways related to proliferation, inflammatory signaling, growth factor activity, and extracellular matrix remodeling.¹⁶ Later review and systems-biology work emphasized the substantial overlap between canonical renal injury-repair mechanisms and pathways implicated in ADPKD progression.¹⁷ Together, these observations suggest that progression is not simply the consequence of genetic cystogenesis plus passive structural distortion but also involves chronic activation of injury and repair responses that evolve over time, including profibrotic signaling and matrix-producing cell populations.

Periostin is a particularly useful example. It is overexpressed in ADPKD cyst epithelial cells and functions as an autocrine mitogen through αV-integrin signaling, promoting epithelial proliferation.⁵¹ Experimental studies have further shown that periostin contributes to both cyst growth and interstitial fibrosis in PKD models,^25,52 placing it at the intersection of proliferative and fibrotic pathways. More broadly, periostin is part of a class of matricellular proteins that regulate tissue repair, matrix interaction, and remodeling.⁵³ This makes it a compelling illustration of how a pathway with “repair-like” properties can become chronically maladaptive in a persistent stress environment.

Together, these observations illustrate how repair pathways activated early in response to tissue stress can become sustained mediators of remodeling and fibrosis as that stress persists. Thus, the biological consequences of repair pathway activation may change over the disease course.

Why Static Structural Biomarkers are Inherently Limited

TKV remains indispensable because it captures one of ADPKD's defining features— structural cystic burden. As a macrostructural measure, it is highly informative, correlates closely with total cyst volume and its change over time, and predicts long-term risk.^5,54 However, TKV and total cyst volume remain fundamentally structural measures. Thus, the key limitations of TKV are not related to inaccuracy, but to incompleteness. In the context of a temporally dynamic disease, this incompleteness has the following 2 distinct dimensions: TKV does not directly capture how the biological processes contributing to progression change over time, and it does not fully characterize the condition of the kidney underlying a given structural burden.

More detailed analysis of conventional kidney imaging can refine this structural assessment. Measures of cyst number, cyst volume, cyst-parenchyma surface area, hemorrhagic cyst burden, parenchymal volume, image texture, and cyst distribution have shown additional prognostic value beyond TKV.55, 56, 57, 58 These approaches extract additional information from conventional kidney imaging and may improve structural phenotyping and risk prediction. Importantly, however, they remain predominantly anatomical measures of disease. Although they can characterize cystic architecture and the structural composition of the kidney more precisely than TKV alone, they do not directly measure the evolving physiological and molecular processes that may determine how structural disease ultimately translates into functional decline.

The first limitation is temporal. The distinction between structural burden and the biological processes active at a given time becomes apparent when structure and function become uncoupled. In early disease, TKV may increase substantially whereas estimated GFR (eGFR) remains preserved.^5,10 Later in the disease course, eGFR may decline more rapidly even as structural growth decelerates. This is consistent with the temporally dynamic model, in which the relative contribution of different biological processes changes over time. TKV primarily provides a cumulative measure of kidney enlargement associated with cyst development and expansion over time, but does not directly indicate which biological processes are the strongest contributors to progression at the time of measurement. A given TKV may therefore reflect the cumulative effects of epithelial proliferation and cyst development and expansion over many years without indicating the current relative activity of proliferation, inflammation, metabolic dysfunction, fibrotic remodeling, or loss of functional parenchyma. Changes in the underlying biology and changes in TKV need not occur in parallel. The relative contribution of biological processes may shift without a corresponding change in the rate of kidney enlargement. In this sense, TKV provides a cumulative structural record of cystic disease rather than a direct readout of its current biological activity.

The second limitation concerns the biological information contained within a given structural burden. TKV alone cannot distinguish kidneys with different cyst number, size, distribution, or cyst-parenchyma relationships, and recent image-derived biomarkers have begun to address this limitation.^55,56,58 However, even more detailed anatomical characterization does not necessarily define the biological or physiological condition of the remaining parenchyma. Two patients of similar age and with comparable htTKV may differ in the extent of fibrosis, tubular injury, microvascular impairment, inflammation, metabolic dysfunction, or other processes affecting the remaining kidney tissue. Such differences could contribute to variation in subsequent functional decline despite apparently similar structural burden.

The residual variability observed within the current structural risk classifications provides a clinical context in which this hypothesis can be tested. The MIC partly addresses structural heterogeneity by first distinguishing typical (Class 1) from atypical (Class 2) patterns of cystic involvement and then, within Class 1 disease, incorporating age and htTKV to estimate expected structural progression.⁹ However, subsequent validation studies have demonstrated variability in both kidney growth and functional decline within individual MIC classes.⁵⁹ This heterogeneity is particularly evident in MIC 1C, where patients exhibit a broad range of disease progression despite similar structural classification.⁶⁰ These observations do not establish the biological basis for this variability, but they demonstrate that age and structural burden alone do not fully account for subsequent progression. Whether differences in cystic architecture, parenchymal integrity, fibrosis, microvascular function, or other biological processes explain part of this residual variability is therefore an important testable question.

These 2 limitations are complementary. TKV provides an essential measure of accumulated cystic burden, and more detailed structural imaging biomarkers can further characterize cystic architecture and improve structural phenotyping. However, neither necessarily identifies which biological processes are the strongest contributors to progression at a given time nor fully captures the kidney's biological and functional condition associated with that structural burden. A temporally dynamic framework therefore does not diminish the value of TKV or more detailed structural imaging biomarkers, but places this information within a broader biological context in which both the processes active at a given time and the integrity of the remaining parenchyma may influence subsequent progression.

Implications for Biomarkers

If ADPKD is temporally dynamic, biomarkers should be evaluated not only by their association with structural or functional disease severity at a given time, but also by whether they capture active biological processes, anticipate changes in trajectory, and respond to therapeutic intervention. No single biomarker is likely to capture these complementary dimensions of disease; the objective is instead to integrate a limited set of measurements representing distinct aspects of structure, biology, and function.

TKV provides the structural backbone of such a framework, whereas complementary biomarkers may capture injury, inflammation, hypoxia, metabolic dysfunction, fibrosis, microvascular impairment, or loss of functional reserve.^16,17 The relative informativeness of these measurements is unlikely to remain constant over time and cannot yet be assigned to specific disease periods. Some biological or physiological abnormalities may precede detectable structural or functional consequences, whereas others may become more informative as tissue injury and remodeling progress. Defining these temporal relationships is therefore an important goal of longitudinal biomarker studies.

One practical approach toward this goal is to complement TKV with imaging and molecular biomarkers that interrogate the noncystic kidney parenchyma and provide information beyond cumulative structural burden. Quantitative and functional imaging techniques are particularly attractive because they can repeatedly assess physiological and tissue-level processes in vivo. Blood oxygen level-dependent magnetic resonance imaging can provide information related to renal oxygenation, arterial spin labeling, and other perfusion-based approaches can assess renal hemodynamics, and diffusion-weighted and magnetization-transfer imaging may provide information related to parenchymal microstructure and remodeling.^61,62 More recently, ultrasound localization microscopy has enabled direct in vivo visualization and quantification of the intrarenal microvasculature in ADPKD, providing measures of microvascular density, flow, and perfusion that may complement conventional structural imaging.⁶³ Rather than expecting any single modality to replace TKV, a more informative strategy may be to combine structural burden with selected physiological, imaging, and molecular biomarkers that capture complementary aspects of disease biology. The objective would be to identify a limited, complementary set of measurements capable of representing these biological dimensions across the disease course. The set of measurements need not change over time; rather, their values and relationships to one another may change, generating a multidimensional profile of the patient's evolving biological state. Longitudinal studies will be important to determine whether such profiles identify biological transitions and improve prediction of subsequent progression beyond established structural and clinical measures.

Developing such an approach would require several complementary strategies. Candidate biomarkers should first be linked to defined biological processes and characterized across the full disease course, including childhood and early adulthood, when structural progression may already be evident despite preserved kidney function. Longitudinal studies should determine how these markers change within individuals, whether their trajectories precede structural or functional transitions, and whether they provide information beyond established predictors such as MIC and eGFR. However, testing this concept need not require following individual patients across the entire natural history of ADPKD. Existing cohorts and repositories spanning different ages and levels of disease severity may allow retrospective assessment of candidate biomarkers when appropriate stored biospecimens, imaging, or other data are available. These data, particularly when shorter longitudinal observations are available within individuals, could be used to reconstruct candidate biomarker trajectories across the disease course. A similar strategy was used in the Consortium for Radiologic Imaging Studies of Polycystic Kidney Disease to reconstruct long-term GFR trajectories across adulthood from patients observed over different portions of the disease course, with subsequent validation in the HALT-PKD trial.¹⁰ For biomarkers or physiological measurements not available in existing repositories, prospective acquisition across strategically selected portions of the disease spectrum could provide complementary information without requiring decades of follow-up.

Biomarkers that provide complementary information could then be evaluated together to determine whether integrating different dimensions of disease improves assessment of progression or therapeutic response. Importantly, the objective should not be to maximize the number of biomarkers measured, but to identify a limited combination that captures biologically distinct aspects of disease and provides reproducible information beyond established measures. Prospective studies could then test and refine these profiles and determine whether they improve disease characterization, progression prediction, and assessment of therapeutic response.

Longitudinal measurements remain particularly valuable because slope, acceleration, and inflection points may reveal biological change that is not apparent from absolute biomarker levels. At the same time, the complementary strategies outlined above may allow candidate temporal patterns to be identified without requiring complete longitudinal trajectories from every individual. Ultimately, a limited set of complementary measurements obtained at a single time point may provide an estimate of the relative contribution of different biological processes at that time, but such profiles would need to be derived and validated against longitudinal outcomes. Dynamic biomarker frameworks in Alzheimer disease similarly demonstrate that the relevance of individual markers can vary across the disease course.^42,43

Existing ADPKD biomarker studies already suggest that structural, inflammatory, injury, metabolic, and concentrating markers capture partly distinct dimensions of disease biology.^10,11 Rather than viewing these differences as inconsistencies, a temporally dynamic framework provides a rationale for determining how the informativeness and relationships of these signals change over time and whether integrating complementary measurements into a multidimensional profile improves prediction or therapeutic-response assessment beyond established structural and clinical measures.

Implications for Clinical Trial Design

The study to assess glucosylceramide synthase inhibitor efficacy in ADPKD trial (NCT03523728) illustrates how evolving end point strategies are beginning to influence ADPKD trial design.⁶⁴ By combining prognostic enrichment with a 2-stage design using TKV growth in stage 1 and eGFR slope in stage 2, the study recognizes that structural and functional end points evolve differently over the course of disease and can be sequenced to improve trial efficiency.⁶⁴ Although developed primarily as a statistical and regulatory strategy, this design is broadly consistent with the view that different end points may be most informative at different points along the disease course.

Traditional clinical trial designs in ADPKD have largely assumed that disease progression can be captured by a single dominant end point applied uniformly across all patients and disease stages. A temporally dynamic framework challenges this assumption. If the relative contribution of biological processes changes over time, then no single end point can be expected to be equally informative across the entire disease course. Instead, trial design must account for the possibility that structural, functional, and biological end points may provide different information depending on the biological context of the population being studied.

The same principle should apply more broadly. Trial design in ADPKD may improve when end points and patient selection are matched not only to feasibility and statistical power, but also to the biological processes expected to contribute most strongly to progression in the population being studied. A therapy targeting a specific biological process may be most appropriately evaluated in patients in whom that process is expected, or ideally demonstrated, to make a meaningful contribution to current disease activity. The corresponding end points should likewise be selected based on the targeted biological process and the anticipated consequences of its modification, rather than assuming that structural or functional measures will be equally informative for every intervention. For example, interventions targeting proliferative, metabolic, inflammatory, injury-repair, vascular, or fibrotic mechanisms may require different biological, physiological, structural, or functional readouts, depending on the mechanism targeted and the population studied. For many of these processes, the most informative response markers and their optimal timing remain to be established. Future trials may therefore benefit from time- and biology-informed designs incorporating improved biological stratification, biomarker-guided enrichment, and sequential or adaptive end point strategies aligned with evolving disease biology.

Taken together, these considerations support a trial framework in which patient selection and end point choice are aligned with the biological processes targeted by the intervention, rather than applying the same end point strategy uniformly across the ADPKD population.

Therapeutic Implications: Why Timing May Be as Important as Target

A temporally dynamic model implies that treatment efficacy is likely to depend on the relative contribution of the targeted biological process. Therapeutic timing, however, need not coincide with the period in which that process or its phenotypic consequences are most prominent. Interventions may be most effective when the targeted process is already contributing meaningfully to disease activity, but in other settings, earlier intervention may be required to prevent the development of irreversible consequences of processes that become increasingly important later. This principle may apply across proliferative, metabolic, inflammatory, injury-repair, vascular, fibrotic, and other mechanisms contributing to ADPKD progression. In this framework, apparent treatment failure or partial efficacy may reflect not only limitations of the therapeutic target, but also a mismatch between the mechanism of action and the biological context in which the therapy is administered, including whether the targeted process is active, emerging, or already associated with irreversible downstream consequences.

This perspective helps reinterpret previous clinical trial results in ADPKD. For example, mTOR inhibitors demonstrated effects on structural end points in some studies but did not consistently improve kidney function. In a randomized trial of everolimus, treatment slowed the increase in TKV but did not significantly alter the decline in kidney function.⁶⁵ Importantly, the investigators noted that slowing kidney enlargement did not necessarily translate into preservation of kidney function and that treatment effects may vary depending on disease stage.⁶⁵ Similar observations in sirolimus studies highlight the broader challenge of linking structural and functional outcomes over relatively short time frames.⁶⁶ Within a temporally dynamic framework, these findings can be interpreted not simply as therapeutic failure but as suggesting that the effectiveness of mTOR inhibition may depend on the biological context in which the pathway is targeted, including the relative contribution of mTOR-dependent proliferative and growth-related processes and the extent to which less reversible downstream processes have already become established. Once substantial fibrosis and loss of functional parenchyma are present, modifying proliferative and growth-related signaling alone may be insufficient to produce proportional preservation of kidney function.

The mTOR story also illustrates a broader translational challenge in ADPKD. In multiple rodent models, mTOR inhibition with sirolimus or everolimus reduced cystic disease and kidney enlargement,67, 68, 69, 70 supporting the biological relevance of mTOR-dependent growth and proliferative signaling in disease progression. However, these structural effects did not translate consistently into clinical benefit, highlighting the challenge of translating modulation of a biologically relevant pathway in experimental models into meaningful outcomes in human disease.

This discrepancy may reflect important differences between preclinical models and human disease. Many experimental studies use early, rapidly progressive models in which proliferative and cystic processes contribute prominently, and treatment is initiated early in cyst development. Human ADPKD, in contrast, evolves over decades, during which the relative contribution of inflammatory, repair, vascular, fibrotic, and other processes may change substantially. Moreover, preclinical studies frequently rely on cyst burden or kidney volume as primary efficacy readouts and are often limited to relatively short treatment and observation periods. Although these studies are essential for establishing biological and structural efficacy, an early reduction in cystic burden or kidney enlargement does not necessarily establish durable preservation of kidney function or delayed kidney failure. Early suppression of cystic disease may reduce or delay downstream injury and remodeling, particularly to the extent that these processes arise as consequences of cyst development and expansion. However, this relationship cannot be assumed to be proportional or complete, particularly if other disease processes may also be influenced by the underlying polycystin defect or its downstream biological consequences. Even polycystin re-expression remains sensitive to disease stage—experimental re-expression of Pkd1 or Pkd2 can reverse cystic and associated molecular changes, but recovery is less complete when re-expression occurs at more advanced stages, with persistent fibrosis, tubular atrophy, inflammation, and less complete functional recovery.⁷¹ When an intervention slows rather than reverses cystic disease, short-term structural improvement may therefore represent a delay in progression rather than a durable alteration of the subsequent disease trajectory. Where feasible, longer follow-up in slower-progressing models could help distinguish these possibilities and determine whether early structural efficacy translates into sustained functional benefit. Thus, differences in underlying biology, timing of intervention, duration of observation, and end point selection may all contribute to the gap between preclinical efficacy and clinical benefit.

The mTOR example adds another layer to this translational complexity because of the pleiotropic nature of mTOR signaling. mTOR regulates not only proliferation but also apoptosis and autophagy, and the effects of its inhibition vary across models, doses, and cellular contexts.⁷² In experimental systems, mTOR inhibition can have divergent effects on apoptosis and tissue remodeling depending on the underlying biological context, underscoring that its impact is not uniform across the disease course. This interpretation is also consistent with the concept that ADPKD may represent a state of persistent or “futile” repair, in which pathways such as mTOR that normally support adaptive responses to injury become chronically activated.³⁰ In this context, the therapeutic effects of pathway inhibition may depend not only on the target itself, but on whether the underlying repair program remains reversible at the time of intervention.

Similar considerations may apply to pathways involved in oxidative stress responses and cellular adaptation. Early in the disease, activation of antioxidant and stress-response programs may help sustain proliferation, metabolic reprogramming, and cellular survival under conditions of persistent cystic stress. With progression, however, these same pathways may become insufficient, dysregulated, or exhausted. This raises the possibility that therapeutic modulation of stress-response pathways could have different effects depending on the biological context in which they are targeted. For example, NRF2-dependent antioxidant responses have been explored as therapeutic targets in kidney disease, including ADPKD, but the consequences of further augmenting these pathways may differ depending on whether they are already highly active, compensating, or failing at the time of intervention.

These considerations suggest that therapeutic strategies in ADPKD may benefit from time- and biology-informed targeting, in which interventions are selected according to the biological processes relevant to progression and timed according to when those processes remain therapeutically modifiable. Importantly, the optimal therapeutic window need not coincide with the period of greatest biological or phenotypic prominence; for some processes, intervention may need to occur earlier to prevent downstream changes that become increasingly difficult to reverse. In this sense, a temporally dynamic model aligns ADPKD with broader precision-medicine thinking—the right target, in the right patient, at the right time.

Conclusion

The available literature supports a view of ADPKD in which multiple biological processes contribute throughout the lifespan, but their relative importance changes over time. Developmental timing, nonlinear structural and functional trajectories, metabolic reprogramming, injury-repair biology, and progressive fibrosis can therefore be integrated within a temporally dynamic model.

This framing does not replace structural assessment. TKV, MIC, genotype, and kidney function remain essential, but they provide only part of the information needed to characterize the biological processes contributing to progression at a given time. The next challenge is to identify complementary measurements that detect meaningful biological change, determine how the contribution of specific processes evolves over time, and identify when those processes remain therapeutically modifiable. If successful, this approach could enable more precise alignment of biomarkers, patient selection, therapeutic targets, and timing across the ADPKD disease course.

Disclosure

The author declared no competing interests.