Fundamentals of dermatologic drug discovery

1 Introduction

Dermatology has emerged as one of the most fertile areas for mechanism-based therapeutic innovation. Over the past two decades, advances in cutaneous immunology, human genetics, and molecular pharmacology have transformed the treatment landscape for inflammatory, autoimmune, infectious, adnexal, and neoplastic skin diseases. Targeted biologics, oral small molecules, improved topical therapies, and device-enabled delivery approaches have demonstrated that skin disease is not only tractable, but especially well-suited to translationally informed drug development (Kamat et al., 2024).

Successful translation from biologic insight to clinically meaningful therapy nevertheless remains difficult. The skin imposes constraints that differ in important ways from those of other therapeutic areas. Barrier function, skin-compartment biology, immune redundancy, disease heterogeneity, and the nature of dermatologic endpoints all influence development strategy. Success depends not only on identifying a biologically relevant target, but also on aligning that target with the intended site of action, therapeutic modality, delivery strategy, translational model package, and early clinical-development plan. Programs may fail despite persuasive biology because of inadequate cutaneous exposure, poor local tolerability, insensitive endpoints, or a benefit-risk profile misaligned with disease severity and expected duration of use.

Many drug-discovery frameworks are disease-agnostic or adapted from oncology and rheumatology. These frameworks remain useful, but they may underemphasize issues that are especially consequential in dermatology, including local tolerability, compartment-specific delivery, the distinction between rapid symptom control and deeper disease modification, and the unusual opportunity to assess tissue-level pharmacodynamic effects directly in humans. For example, an oncology-derived benefit-risk calculus that accepts substantial toxicity in exchange for survival benefit does not transfer to a chronic, frequently non-life-threatening dermatosis, where tolerability and monitoring burden over years of use often determine success. A development logic borrowed from systemic drug development, in which plasma exposure is assumed to predict activity at the target site, can likewise mislead, because systemic exposure does not reliably predict concentration in the stratum corneum, viable epidermis, or dermis where the relevant cutaneous biology resides. A rheumatology-style default toward systemic biologic blockade, appropriate for severe systemic disease, can over-treat anatomically localized skin disease for which a topical or locally targeted approach would achieve the goal with less systemic risk. A topical program may fail despite strong target potency if adequate local exposure cannot be achieved without irritation. A systemic program may demonstrate biologic activity yet still fail if its safety burden is disproportionate to the severity and chronicity of the intended indication.

Figure 1 provides the organizing framework for this Primer, linking target product profile definition, target validation, pathway positioning, modality and delivery selection, translational evidence generation, and early clinical proof-of-concept through iterative decision gates. A glossary of key terms is provided in the Supplementary Material.

1.1 Matching route of administration to disease biology

Route selection in dermatologic drug development should begin with the therapeutic objective. Early in a program, this means defining the site of action, disease extent, target accessibility, acceptable systemic burden, and intended clinical use. Each administration strategy (topical application, local injection or procedural delivery, oral therapy, and parenteral systemic therapy) carries distinct requirements for exposure, tolerability, monitoring, adherence, and setting of care. The aim is to select the approach most likely to achieve meaningful target engagement at the site of disease while preserving an acceptable benefit-risk profile.

Many inflammatory skin diseases have systemic dimensions. Psoriasis is associated with cardiometabolic, hepatic, psychiatric, and joint comorbidities; atopic dermatitis reflects systemic type 2 immune dysregulation and atopic comorbidity; and hidradenitis suppurativa is linked to broader metabolic and inflammatory burden. In these settings, systemic therapy may be the most rational development path from the outset, particularly when disease is extensive, refractory, comorbidity-associated, or accompanied by extracutaneous manifestations, or when it depends on immune mechanisms that cannot be adequately modulated through local delivery.

Other dermatologic conditions are predominantly localized or compartment-specific. Adnexal disorders, focal pigmentary conditions, contact and irritant dermatitis, superficial infections, localized inflammatory lesions, and cosmetically-relevant conditions may be better served by topical delivery, including follicular-targeted and formulation- or device-enabled approaches, or by intralesional injection. Even in systemic immune-mediated diseases, regional manifestations such as scalp psoriasis, hand or eyelid dermatitis, or follicular-predominant involvement in hidradenitis suppurativa may call for local or adjunctive treatment rather than systemic dose escalation.

Dermatologic care is therefore tiered by clinical need. Topical therapies remain foundational for many mild-to-moderate diseases; selected oral agents may be appropriate depending on mechanism and severity; and biologics or other systemic therapies are generally reserved for patients whose disease extent, refractoriness, extracutaneous involvement, or comorbidity burden justifies broader immunomodulation. Patient context shapes this decision further, including age, pregnancy potential, polypharmacy, infection or malignancy risk, monitoring feasibility, treatment preference, and willingness to accept chronic systemic exposure.

The development question, then, is not whether topical or systemic therapy is preferable in principle, but which approach best matches the underlying biology, the target population, the intended use, and the required benefit-risk profile.

1.2 Defining the target product profile

A target product profile (TPP) is a foundational tool in dermatologic drug discovery. It specifies the minimum characteristics a successful therapy must meet, including indication, target population, route of administration, efficacy threshold, onset and durability of effect, safety and tolerability expectations, dosing frequency, monitoring burden, co-therapy and comorbidity context, and practical considerations related to adherence, access, storage, distribution, and setting of care (Boccardi et al., 2024; Tebbey and Rink, 2009).

In dermatology, TPP definition requires explicit attention to skin-specific needs: acceptable local tolerability, feasibility of chronic or intermittent use, rapid symptom relief when clinically relevant, and suitability for topical or localized delivery when appropriate. For systemically administered therapies, acceptable safety and monitoring burdens must be judged against disease severity, available alternatives, comorbidity burden, and expected treatment duration.

An explicit TPP sharpens strategic discipline from the outset. It helps filter targets, guide modality and formulation choices, and structure development decisions around the characteristics a therapy must ultimately deliver. Without it, programs may advance candidates that show biologic activity yet are poorly positioned to meet clinical, regulatory, operational, or adoption requirements. As a program matures, the TPP should be revisited at predefined decision points to incorporate emerging data on target biology, exposure feasibility, and evolving standards of care. These revisions should be prospectively governed to avoid unnecessary development drift and preserve regulatory coherence.

2 Target identification and validation

The first task in dermatologic drug discovery is not simply to identify biologic activity, but to identify biology that is sufficiently relevant, actionable, and therapeutically tractable to justify development. In practice this means moving beyond association and building a coherent target rationale that can withstand translational attrition. Three filters are especially useful at this stage: disease relevance, therapeutic index, and actionability (Copeland, 2026).

2.1 Human relevance as the foundation of target selection

Successful dermatologic drug discovery begins with targets causally implicated in human disease. Targets supported by human genetic evidence, disease-associated pathway activation, tissue-level expression patterns, and pharmacodynamic modulation during clinical response are substantially more likely to translate into approved therapies than targets supported by a single line of preclinical evidence (Minikel et al., 2024; Nelson et al., 2015; Finan et al., 2017; King et al., 2019). In dermatology, this principle is particularly salient due to the accessibility of diseased tissue and the availability of lesional and non-lesional skin for molecular profiling.

Human genetics has played a pivotal role in validating immune and barrier-related pathways implicated in inflammatory skin disease. Both common variant association studies (GWAS) and rare monogenic disorders have helped distinguish causal drivers from downstream epiphenomena (Finan et al., 2017). Importantly, not every disease-associated pathway represents an optimal therapeutic target. Redundancy within cutaneous immune networks and compensatory mechanisms can attenuate the clinical effect of single-node inhibition (Minikel et al., 2024; Finan et al., 2017).

Animal, phenotypic, and in vitro evidence remains valuable, particularly when matched to the question being asked. These approaches answer questions that human-relevant systems cannot, including chronic dosing, integrated physiology, formulation tolerability, and selected delivery problems. Targets nominated primarily through animal or phenotypic discovery should subsequently be anchored to human disease biology before substantial development investment, because programs that lack human-relevance support face systematically lower probabilities of approval (Minikel et al., 2024; Nelson et al., 2015; Finan et al., 2017; King et al., 2019). The point is not to dismiss preclinical evidence but to insist that human relevance is the anchor against which other evidence types are interpreted.

Animal and porcine models have contributed positively to dermatologic translation when used for questions they can address directly, including barrier penetration, local tolerability, wound repair, systemic safety, repeated or long-term dosing, and device-enabled delivery. Porcine skin is particularly useful for topical and transdermal development because its architecture and permeability more closely approximate human skin than many small-animal models (Montagna and Yun, 1964; Bartek et al., 1972). Classic domestic pig studies showed that maintaining a moist, covered wound environment accelerates re-epithelialization compared with allowing wounds to dry and scab (Winter 1962). Rabbit ear ischemic wounds and murine excisional or diabetic-wound models have provided mechanistic insight and enabled candidate comparison (Ahn and Mustoe, 1990; Galiano et al., 2004; Fang et al., 2010). However, these models also illustrate limits of prediction: rodent wounds often close predominantly by contraction, and db/db mouse wounds incompletely reflect the polygenic, vascular, neuropathic, infectious, and inflammatory complexity of human diabetic ulcers (Galiano et al., 2004; Fang et al., 2010). Accordingly, improvement in animal-model outcomes, such as faster wound closure or reduced inflammation, should be interpreted as supportive preclinical evidence rather than definitive proof of human clinical benefit unless corroborated by human disease biology, exposure, pharmacodynamic evidence, or clinical data.

Tissue specificity is another important consideration. Targets broadly expressed across organ systems may introduce systemic risk without proportional dermatologic benefit, particularly when local or skin-restricted modulation could achieve the desired effect. Pathways enriched or dysregulated in skin-resident immune cells, keratinocytes, fibroblasts, neurons, or adnexal structures offer better opportunities for selective intervention. Table 1 summarizes a practical evidence framework for target validation in dermatologic drug discovery, emphasizing both human relevance and fit-for-purpose use of complementary model systems.

TABLE 1. Practical evidence framework for target validation in dermatologic drug discovery.

2.2 Pathway positioning and disease architecture

Once a target is considered biologically relevant, its position within disease architecture becomes critical. Dermatologic diseases rarely reflect a single pathogenic axis. They arise from layered and interacting processes that include immune activation, barrier dysfunction, microbial interactions, structural remodeling, and neuroimmune signaling. Effective target selection therefore requires a clear understanding of where a target sits within that broader architecture (Beck et al., 2022; Glickman et al., 2025; MacGibeny et al., 2024; Harris-Tryon and Grice, 2022).

Proximal targets that regulate multiple downstream mediators may yield broader efficacy but may also carry greater systemic safety risk. Distal targets may be better tolerated yet may provide insufficient efficacy in the presence of pathway redundancy. This framing aligns with network-based drug discovery principles, which emphasize that target quality depends not only on druggability, assayability, or differential expression, but also on the target’s position and degree of control within a dynamic disease network (Recanatini and Cabrelle, 2020).

These tradeoffs become more meaningful when considered against the intended route of administration, therapeutic index, disease severity, and expected duration of use. Importantly, even two agents directed at the same pathway can differ meaningfully in clinical performance because they disrupt cytokine-receptor complex assembly at distinct structural and energetic steps, as illustrated by comparative structural and thermodynamic analyses of IL-4 and IL-13 axis biologics in atopic dermatitis (Bunick, 2025).

For systemic programs, proximal immune nodes may be appropriate when the anticipated benefit justifies the safety and monitoring burden. For topical or locally delivered programs, distal or compartment-specific targets are often more attractive because meaningful efficacy can be achieved with limited systemic exposure. Target prioritization, modality selection, and delivery strategy are best treated as iterative and interdependent rather than strictly sequential.

2.3 Immune, barrier, and neuroimmune targets

Drug discovery in skin disease has historically focused on immune pathways, particularly in inflammatory dermatoses. That focus remains central, but barrier restoration and neuroimmune modulation have become increasingly important as complementary or alternative strategies.

Barrier-related targets, including those involved in epidermal differentiation, lipid metabolism, and tight-junction integrity, are especially relevant in diseases in which barrier dysfunction amplifies inflammation, antigen penetration, and microbial dysbiosis. Modulating barrier pathways may reduce disease severity indirectly by stabilizing the cutaneous interface, even in the absence of direct immunosuppression (Irvine et al., 2011; Akdis, 2021; Bunick et al., 2015; Goleva et al., 2019; Liu et al., 2025). Barrier biology is also increasingly structurally tractable. Crystallographic definition of the profilaggrin S100 domain and its protein interaction interfaces, for example, provides a concrete molecular basis for barrier-relevant target engagement (Bunick et al., 2015).

Neuroimmune targets have gained prominence due to their role in symptom generation, particularly pruritus and pain. Targeting neural signaling pathways can yield rapid symptomatic relief and may meaningfully improve quality of life, even when objective disease measures change modestly. However, in heterogeneous inflammatory skin diseases, neuroimmune pathways operate within networks involving multiple cytokines and mediators, so focusing on a single pruritogenic cytokine is unlikely to fully address the underlying inflammatory architecture for many patients (Facheris et al., 2023; Speeckaert et al., 2024). Recent mechanistic work illustrates this complexity, showing that IL-31-neuronal signaling can restrain type-2 inflammation in allergic dermatitis models through CGRP release, even as IL-31 drives pruritus (Fassett et al., 2023). Neuroimmune-directed therapies may therefore be best positioned as components of broader strategies that also modulate immune and barrier pathways, or as targeted interventions in carefully phenotyped populations where symptom burden is the dominant unmet need. Emerging therapeutic strategies, including bispecific antibodies and rational combinations, offer a potential solution by simultaneously modulating neuroimmune signaling and upstream inflammatory drivers. Incorporating symptom-driven targets requires careful consideration of endpoint selection, trial duration, and parallel assessment of inflammatory and infection-related outcomes.

2.4 Avoiding non-causal and non-actionable targets

A recurring problem in drug discovery is pursuit of targets that correlate with disease activity but are not causally important to disease initiation or maintenance. Biomarkers of inflammation, stress responses, or tissue injury may track with severity yet fail to yield clinical benefit when modulated directly (Nelson et al., 2015; King et al., 2019).

Recent experience with RORγt inhibition illustrates this challenge. Despite strong mechanistic rationale as the Th17 master transcription factor, the oral RORγt inhibitor BI 730357 produced only modest efficacy in plaque psoriasis, with a response plateau at higher doses, and development was discontinued. Across the broader RORγt class, programs have been limited by both insufficient efficacy and, for several agents, dose-limiting safety signals, consistent with the difficulty of achieving adequate, durable pathway suppression at a tolerable therapeutic index (Akbarialiabad and Bunick, 2025).

To reduce this risk, target validation should emphasize functional necessity rather than correlation alone. Experimental perturbation in human-relevant systems can help distinguish causal drivers from downstream markers, but no single system is sufficient in isolation. Ex vivo human skin, in vitro co-culture models, fit-for-purpose animal systems, and early human pharmacodynamic readouts can all contribute when matched to the translational question being asked (Copeland, 2026; King et al., 2019).

Actionability matters as much as biological plausibility. Some compelling targets may be difficult to modulate safely or effectively with available modalities. Targets that require sustained or near-complete inhibition to produce benefit may be poorly suited to chronic dermatologic indications. Early assessment of tractability, dosing feasibility, local and systemic tolerability, and therapeutic index is therefore essential (Copeland, 2026).

3 Modality and delivery selection

If Section 2 asks whether a target is worth pursuing, Section 3 asks how that biology can realistically be converted into a therapy. In dermatology, that question cannot be answered by mechanism alone. The same target may lead to very different development programs depending on whether the intended intervention is topical, systemic, local injectable, device-assisted, or soft-drug based.

3.1 Choosing the appropriate therapeutic modality

Once a target has been sufficiently prioritized, the next major decision is therapeutic modality. In this Primer, therapeutic modality refers to the type of intervention used to engage a target, whereas delivery strategy refers to how that intervention is brought to the intended site of action. In dermatology, modality choice is tightly linked to target biology, disease severity, intended route of administration, and desired site of action.

Small molecules are low-molecular-weight chemical entities that are often well suited for intracellular targets. Their ability to penetrate the epidermis and adnexal structures makes them attractive for diseases involving keratinocytes or follicular units. However, systemic exposure may introduce safety considerations, especially for targets involved in immune regulation.

Biologics provide high specificity and have proven highly effective for extracellular or cell-surface targets. Their success in inflammatory skin disease reflects the central role of cytokine signaling in pathogenesis. Limitations include parenteral administration, limited penetration into some skin compartments, and greater manufacturing complexity.

Topical therapies remain foundational in dermatology because they offer the possibility of strong local benefit with limited systemic risk. Advances in formulation science and delivery technology have expanded the range of targets amenable to topical modulation, although achieving sufficient target engagement in deeper skin layers remains a persistent challenge (Singhal et al., 2017; Watkinson et al., 2016).

Modality selection cannot be separated from delivery strategy. Topical, local injectable, device-assisted, and systemic approaches create distinct development programs, each with different constraints on exposure, tolerability, monitoring, setting of care, and proof-of-concept design. Explicit early separation of these tracks reduces late-stage attrition caused by unrealistic assumptions about efficacy, safety, or feasibility. Table 2 compares major modality options across these domains.

TABLE 2. Modality and delivery strategy selection in dermatologic drug discovery: Key considerations.

3.1.1 Identifying a viable starting point

Discovery programs benefit from clear identification of their starting point. In dermatology, five practical pathways commonly initiate successful programs.

First, screening-based approaches, whether target-based or phenotypic, are appropriate for novel targets with limited prior chemical matter. They can identify first-in-class candidates but often require substantial optimization to achieve adequate skin exposure and tolerability.

Second, fast-follower strategies build on established target classes or mechanisms to improve known limitations such as onset of action, selectivity, safety, or suitability for topical delivery.

Third, drug repositioning prioritizes new dermatologic indications for agents with existing human safety, manufacturing, or formulation information. Repositioning can accelerate proof-of-concept, but it still requires careful reassessment of cutaneous exposure, dosing feasibility, indication-specific safety, formulation suitability, and dermatology-specific benefit-risk.

Fourth, reformulation and delivery-led innovation aim to improve performance by changing how an active agent reaches the skin. Examples include strategies that enhance epidermal penetration, improve follicular targeting, or reduce systemic exposure.

Fifth, medicinal-chemistry approaches such as soft-drug design may be particularly attractive in dermatology. Soft drugs are active compounds intentionally designed to undergo predictable inactivation after producing the intended local effect. This strategy is especially useful when high local potency is required but prolonged systemic exposure would narrow the therapeutic window (Thoma et al., 2023; Aprile et al., 2019).

The optimal starting point depends on target biology, unmet need, operational capability, and the predefined TPP. In all cases, the TPP should serve as the primary filter for whether an approach is likely to yield a clinically meaningful and commercially viable therapy rather than merely a mechanistically interesting molecule.

3.1.2 Discovery stages with dermatology-specific considerations

Regardless of starting point, dermatologic programs move through the standard discovery stages of hit identification, hit-to-lead, lead optimization, and candidate nomination. At each step, skin-specific criteria should be built in early to avoid later failure caused by inadequate cutaneous exposure, poor local tolerability, or unrealistic operational assumptions.

Hit identification generates initial active compounds through target-based screening, phenotypic assays, or computational methods. In dermatology, this stage should already consider whether the target is accessible by the intended route and whether the initial chemical matter has properties compatible with skin penetration if topical delivery is planned. Phenotypic screens using human keratinocytes, reconstructed epidermis, organotypic skin, or disease-relevant co-cultures can help identify hits with skin-relevant activity from the outset.

Hit-to-lead narrows the chemical series by improving potency, selectivity, and preliminary absorption, distribution, metabolism, and excretion (ADME) properties. For dermatologic programs, this stage should prioritize compounds whose molecular weight, lipophilicity, solubility, and stability fit the intended delivery route. Early permeation testing can eliminate topical candidates that are unlikely to penetrate the stratum corneum, while systemic candidates should be assessed for their ability to achieve adequate skin exposure at tolerable doses.

Lead optimization refines the series through iterative design-make-test-analyze cycles that balance potency, selectivity, drug metabolism and pharmacokinetics (DMPK), safety, manufacturability, and formulation feasibility. In dermatology, formulation development and dermatopharmacokinetic assessment should be integrated during this stage rather than deferred. For topical leads, local tolerability, irritation, and sensitization potential should be evaluated alongside efficacy drivers. For systemic leads, skin-to-plasma ratios and cutaneous target engagement should be characterized in relevant translational systems.

Candidate nomination marks the transition from discovery to development and should be governed by explicit criteria. In dermatologic programs, these criteria should extend beyond standard pharmacology and safety margins to include evidence of skin exposure at pharmacologically relevant concentrations, acceptable local tolerability for the intended duration of use, alignment with the TPP, and a credible path to clinical proof-of-concept using clearly defined endpoints. Before candidates advance, apparent hits should also survive basic quality filters, including reproducibility across primary and orthogonal assays, confirmation of chemical identity and purity, exclusion of obvious assay artifacts such as non-specific activity or flat structure-activity relationships, and sufficient tractability to support early structure-activity refinement (Copeland, 2026).

In this Primer, a decision gate refers to a prespecified checkpoint for advancement. A futility rule refers to a predefined criterion indicating that continued investment is unlikely to be justified without a meaningful change in candidate properties, delivery strategy, or indication. Programs that reach candidate nomination without addressing these dermatology-specific requirements remain vulnerable to late-stage failure due to formulation, exposure, or tolerability problems that could have been identified earlier. Table 3 summarizes practical decision gates and futility rules across discovery stages.

TABLE 3. Decision gates and futility criteria across discovery stages.

3.2 Skin-specific pharmacokinetics and barrier constraints

The skin presents a distinctive pharmacokinetic environment that strongly shapes dermatologic drug discovery. The stratum corneum is an effective barrier to molecular penetration and favors compounds with specific physicochemical properties. As a result, systemic exposure does not reliably predict cutaneous exposure, and pharmacokinetic assumptions borrowed from other therapeutic areas often fail.

Guidelines such as the Rule of 500 remain useful starting points for topical feasibility, but they are not absolute limits (Bos and Meinardi, 2000). Barrier disruption, formulation optimization, prodrug strategies, and device-assisted delivery can expand the range of molecules amenable to topical or localized use. The local-to-systemic exposure ratio is a particularly important optimization parameter when systemic inhibition carries safety liabilities.

Vehicle selection and formulation can markedly alter dermal pharmacokinetics, which is why formulation work should begin early rather than late in discovery. The vehicle is rarely inert. It can affect barrier integrity, hydration, patient acceptability, and local tolerability, each of which may influence apparent efficacy (Ramadon et al., 2022; Kumar et al., 2023).

Even within a given target class, structural determinants can drive meaningful differences in potency, selectivity, and clinical performance, as illustrated by comparative structural analyses of dermatologic PDE4 inhibitors (Wang et al., 2025).

For topical therapies, successful delivery requires traversal of the stratum corneum while maintaining sufficient stability and activity within the viable epidermis or dermis. Molecular size, lipophilicity, polarity, and formulation all influence penetration efficiency. Even for systemically administered therapies, cutaneous target engagement depends on dermal vascularization, tissue distribution, and local metabolism, which may vary substantially across disease states and anatomical sites.

Pharmacometric exposure-response analyses can determine whether a given pharmacokinetic shift is likely to be clinically meaningful. In acne, for example, modeling indicated that a high-fat meal reduced sarecycline exposure but translated to only an approximately one-lesion difference in clinical outcomes, supporting the conclusion that this pharmacokinetic change was not clinically meaningful, and that weight-based dosing could be applied without food restrictions (Grada et al., 2022). Such analyses prevent both overcorrection and undercorrection of dosing strategies.

Disease-associated barrier disruption can also change penetration over time. Inflamed or damaged skin may allow greater penetration early, followed by reduced exposure as inflammation improves. Programs should therefore assess skin exposure across disease stages and anatomical sites rather than relying on a single condition or time point. For locally acting therapies, demonstration of skin exposure and pharmacodynamic activity should be treated as a primary translational objective.

3.3 Delivery strategies

Delivery strategy determines whether a prioritized target can be engaged at the site of disease at an effective and tolerable exposure. In dermatology, this includes topical and follicular delivery, intralesional or regional administration, device-enabled approaches, systemic delivery, and formulation or medicinal-chemistry strategies designed to maximize local activity while limiting systemic burden.

Follicular delivery is particularly relevant for diseases involving the pilosebaceous unit, including acne, folliculitis, hidradenitis suppurativa, and selected alopecias. Vehicles or particles that preferentially localize to follicular openings may increase exposure in adnexal structures while limiting systemic absorption (Agarwal et al., 2000).

Microneedles, depot formulations, jet injectors, fractional lasers, radiofrequency-assisted delivery, and other barrier-bypassing technologies can expand access to defined skin compartments and enable sustained local exposure. These approaches may be useful when passive topical penetration is insufficient, but they introduce practical considerations related to procedural complexity, cost, patient acceptance, and site-of-care logistics.

Soft-drug design, in which an active compound is designed for predictable inactivation after producing the intended local effect, also belongs within the delivery framework because it can achieve a similar objective through medicinal chemistry: preserving local potency while reducing systemic persistence. This is especially relevant for chronic inflammatory conditions in which cutaneous activity is desired but systemic accumulation could narrow the benefit-risk margin (Thoma et al., 2023).

Delivery strategy should therefore be aligned early with target biology, the TPP, site of action, and intended use setting. Programs that begin with unrealistic delivery assumptions often encounter avoidable formulation, tolerability, or proof-of-concept challenges.

3.4 Operational readiness and development setting

Even a scientifically compelling program may fail if the sponsoring environment cannot execute the required translational work. Academic, biotech, and industry settings each offer distinct strengths and bottlenecks. Academic groups may contribute deep disease expertise and tissue access but may lack medicinal chemistry, formulation capability, regulatory support, or early-phase operational infrastructure. Industry may provide robust cross-functional support but may prioritize only programs that meet predefined portfolio thresholds. Partnership models can therefore be especially important in dermatologic drug discovery.

Operational readiness should be assessed early and explicitly. Key capabilities include medicinal chemistry and formulation support, translational assay development, dermatopharmacokinetic tools, biomarker validation, toxicology planning, regulatory strategy, early clinical operations, biostatistics, and access to appropriate patient populations. This is especially relevant for programs that depend on specialized capabilities such as advanced in vitro or ex vivo skin models, dermatopharmacokinetic platforms, biomarker assays, or early Phase 1 clinical infrastructure, all of which should be assessed at the outset rather than assumed to be available. If a program depends on a capability that is not available in the sponsoring setting, that limitation should be recognized early rather than after candidate nomination or protocol development.

4 Translational models

If target selection defines what biology should be pursued and modality defines how it may be modulated, translational strategy determines whether the program can generate evidence strong enough to move forward. In dermatology, translational models are unusually diverse because the skin can be studied through tissue-based systems, animal models, noninvasive biomarkers, and early human pharmacology. The challenge is not lack of tools, but choosing the right tools for the right question.

4.1 In vitro and ex vivo skin models

Translational modeling is central to dermatologic drug discovery, yet model selection is a frequent source of error. Traditional in vitro assays provide scalability and mechanistic insight but may fail to capture the barrier properties and multicellular interactions of human skin.

Ex vivo human skin models offer an important intermediate step by preserving tissue architecture, cellular diversity, and barrier function. These systems allow direct assessment of penetration, target engagement, and local pharmacodynamic effects in human tissue. They are especially useful for topical and locally delivered therapies, for which human-relevant exposure data may be more informative than systemic pharmacology alone.

Their limitations should also be recognized. Ex vivo systems have limited viability, donor variability, restricted capacity to model chronic disease processes, and practical constraints on throughput. They are best used as one component of a complementary translational package rather than as a universal substitute for other models.

4.2 In vivo and humanized models

Animal models continue to play an important role in dermatologic drug discovery, especially for questions related to systemic safety, chronic dosing, comparative pharmacology, formulation tolerability, wound-healing dynamics, and selected delivery problems. Their limitations arise from species differences in skin structure, immune composition, microbiome ecology, and disease mechanisms, all of which can reduce their ability to predict human efficacy.

Humanized mouse models and genetically engineered systems can improve relevance by incorporating human immune components or disease-associated mutations. These models can be useful for hypothesis testing and candidate comparison, but they should not be treated as definitive predictors of clinical efficacy.

Porcine skin models can be particularly informative for questions involving barrier function, penetration, wound repair, device-assisted delivery, and local tolerability because of their structural similarities to human skin. As with all translational systems, their value depends on matching the model to the question being asked (Summerfield et al., 2015; Rosenberg et al., 2021; Lunney et al., 2021).

4.3 Aligning models with translational questions

Effective translational strategy begins with a clear definition of the question at hand. Models may be used to assess target engagement, mechanism of action, relative potency, chronic exposure, safety margins, formulation performance, or delivery feasibility. No single model answers all of these questions well.

The goal is therefore not to rank model systems universally, but to match each system to the translational question. This fit-for-purpose approach is consistent with broader regulatory-science trends toward integrating human biology-based assays, in vitro and ex vivo platforms, computational methods, and animal studies as complementary evidence streams rather than relying on any single model class. Programs that align discovery questions with appropriate models generate more interpretable data and reduce the risk of advancing candidates on the basis of misleading preclinical signals. Supplementary Table S1 provides a practical toolkit of translational methods with their strengths and limitations.

Computational tools, including molecular docking and machine-learning approaches, can support compound prioritization, but they do not adequately capture skin penetration, local pharmacokinetics, or patient tolerability and should not replace empirical validation.

4.4 Biomarkers and non-invasive assessment

The accessibility of skin creates unusual opportunities for biomarker-driven development. Non-invasive and minimally invasive methods, including tape stripping, sebum sampling, suction blister fluid analysis, and high-resolution imaging, permit repeated pharmacodynamic assessment without repeated full-thickness biopsy (Del et al., 2024; Grada and Bunick, 2023; He et al., 2021; Hughes et al., 2021).

These methods can establish target engagement, confirm mechanism of action, support dose selection, and in some settings enable molecular endotyping or patient stratification. Imaging biomarkers such as optical coherence tomography, reflectance confocal microscopy, and multispectral approaches can add objective measures that complement clinical scoring systems and may be particularly useful in early-phase studies with uncertain effect sizes (Florian et al., 2020; Rissmann et al., 2020; Schuck and Sekar, 2025; Tan et al., 2024).

Transcriptomic profiling of lesional and non-lesional skin has demonstrated prognostic and predictive value in inflammatory dermatoses. Gene expression signatures associated with specific immune axes can identify patients most likely to respond to targeted therapies and may enable biomarker-based enrichment strategies in proof-of-concept trials. Circulating biomarkers, while less tissue-specific, offer practical advantages for monitoring in larger trials and may correlate with systemic inflammatory burden in diseases such as psoriasis and atopic dermatitis.

Effective integration of biomarkers into dermatologic development programs requires early planning, assay validation, and alignment with regulatory expectations. Biomarkers are most useful when they are specific to the biological process being measured, technically reliable, and fit for the decision they are intended to inform, whether target engagement, dose selection, patient stratification, or early proof-of-mechanism. Programs that incorporate biomarker strategies from discovery onward are better positioned to generate mechanistic evidence, optimize dosing, and differentiate their candidates in an increasingly competitive landscape (Florian et al., 2020; Rissmann et al., 2020; Schuck and Sekar, 2025; Tan et al., 2024).

5 Early clinical development and proof-of-concept

By the time a candidate reaches early clinical development, the central questions should be more focused. Is the intended population appropriate? Are the endpoints sufficiently sensitive and mechanistically aligned? Does the study design allow meaningful interpretation of both signal and failure? In dermatology, the visibility of disease and accessibility of tissue can be major advantages, but only if trial design is matched to disease biology and treatment intent.

5.1 Patient selection and enrichment strategies

Early clinical development in dermatology offers several advantages, including visible disease manifestations, accessible tissue, and sensitive clinical endpoints. These advantages can be lost, however, when heterogeneity in disease severity, prior treatment exposure, symptom burden, anatomical distribution, or underlying biology is not addressed in eligibility criteria, stratification, and analysis plans (Rissmann et al., 2020). This is a recognized driver of phase 2 failure in immunologic diseases, where lack of control for interpatient heterogeneity and absence of predictive biomarkers are repeatedly cited as major causes of unsuccessful phase 2 programs and poor phase-transition efficiency (Patel et al., 2017).

Careful patient selection and enrichment can improve signal detection in proof-of-concept studies. Enrichment based on severity, biomarker profile, treatment history, or dominant symptom phenotype may reduce variability and increase the likelihood of demonstrating meaningful benefit within a trial duration aligned with the TPP. Inclusion and exclusion criteria do more than define logistics. They help define the population for whom the therapy is actually being developed.

Failure to adequately define the target population at this stage may result in underpowered studies, ambiguous outcomes, or false-negative results that prematurely terminate otherwise viable programs.

5.2 Endpoint selection and trial duration

Endpoint selection is a defining feature of dermatologic proof-of-concept studies. Dermatology benefits from validated clinician-reported and patient-reported outcomes sensitive to change over relatively short timeframes. However, not all endpoints are equally suited to early development (Rissmann et al., 2020). When feasible, aligning phase 2 outcomes with intended phase 3 or registration endpoints and standardizing endpoint implementation reduces interpretability risk and improves the value of phase 2 signals for downstream decisions (Patel et al., 2017).

In this framework, proof-of-mechanism (PoM) refers to demonstrating target engagement and expected pathway modulation in humans, whereas proof-of-concept (PoC) refers to demonstrating a clinically meaningful effect on prespecified endpoints in the intended patient population.

Endpoints should be chosen according to the drug’s mechanism of action, expected time course of effect, and the minimum difference that would matter to patients and clinicians. Symptom-focused endpoints, such as pruritus or pain scores, may show early benefit even when objective signs change more slowly. Objective clinical-sign endpoints may provide earlier evidence of anti-inflammatory activity but can still lag behind symptom relief depending on disease biology and endpoint construction. Structural or functional endpoints, such as scarring, hair regrowth, repigmentation, or barrier restoration, often require longer studies and larger sample sizes (Silverberg et al., 2024; Armstrong et al., 2017; Armstrong et al., 2025).

An important distinction is whether the therapeutic goal is symptomatic relief, disease modification, or both. Some targets, particularly those in neuroimmune pathways, produce rapid improvement in patient-reported symptoms such as pruritus without necessarily altering underlying disease architecture; itch reduction may precede or exceed changes in objective inflammatory measures. Targets that modulate central immune signaling or restore barrier integrity may demonstrate disease-modifying effects on tissue inflammation, epidermal structure, and long-term disease control. The emerging concept of deep remission, defined as sustained clearance with histologic and molecular normalization (Bieber, 2023), reflects a higher bar for disease modification with implications for treatment duration and discontinuation strategies. Programs should explicitly define whether they are targeting symptom control, disease modification, or both, as this distinction shapes endpoint hierarchies, trial duration, and the evidentiary package required to demonstrate clinically meaningful benefit.

Overly ambitious endpoint selection can obscure meaningful biological effects, while narrowly focused endpoints may fail to capture clinically relevant benefit. Balanced endpoint hierarchies incorporating both objective measures and patient-reported outcomes are often most informative. Supplementary Table S2 offers an operational checklist for proof-of-concept trial design.

5.3 Interpreting early efficacy signals

Early efficacy signals should be interpreted alongside exposure, target engagement, and safety. Apparent clinical benefit without mechanistic confirmation should be viewed cautiously. Conversely, modest clinical change may still represent meaningful proof-of-mechanism if supported by clear pharmacodynamic evidence and if it remains consistent with the TPP (Rissmann et al., 2020; Go et al., 2024).

Programs should prespecify success and failure criteria before study initiation. Clear go or no-go thresholds reduce interpretive bias and support disciplined decision-making when tied to primary and key secondary endpoints. Early studies should also be interpreted in light of whether their duration was appropriate for the mechanism and disease biology under study. A negative short study may not invalidate a target intended to influence slower structural endpoints, just as a symptom-only signal may be insufficient for a program seeking a disease-modifying claim. Quantitative systems pharmacology and model-based approaches can help optimize design and stratification (Go et al., 2024).

6 Safety and benefit-risk fundamentals

The final major theme in this framework is that safety cannot be treated as a late-stage overlay placed on top of otherwise successful development. In dermatology, where many indications require chronic intermittent or continuous treatment and where tolerability strongly influences adherence and uptake, safety and benefit-risk context must be built into development from the beginning. Supplementary Table S3 provides a practical framework for benefit-risk communication across the development lifecycle.

6.1 Anticipating safety liabilities early

Safety should be integrated into dermatologic drug discovery from the beginning rather than deferred to late-stage development. Many dermatologic indications require chronic intermittent or continuous therapy, which places high value on tolerability, monitoring burden, and patient acceptance. Targets involved in immune regulation, cell proliferation, fibrosis, or systemic signaling warrant particular scrutiny.

Early safety assessment should address both on-target and off-target effects. On-target risks may be predictable from pathway biology, whereas off-target effects may arise from modality-specific properties, formulation effects, or distribution patterns. Mechanistic toxicology, dose-response modeling, tissue-distribution studies, and local tolerability assessment can help define a realistic therapeutic window.

6.2 Baseline disease risk and contextualizing adverse events

Safety signals in dermatology must be interpreted in the context of baseline disease-associated risk. Many inflammatory skin diseases are associated with systemic comorbidities, infection risk, psychosocial burden, and concomitant therapies. Without this context, adverse events may be misattributed and development decisions may become overly conservative.

Benefit-risk assessment should therefore incorporate disease severity, unmet need, available alternatives, expected treatment duration, and feasible monitoring. In severe or refractory disease, higher monitored risk may be acceptable if balanced by benefit that cannot be achieved with existing options.

6.3 Risk communication as a development competency

Risk communication is not only a post-approval issue. Early programs should anticipate how safety information will be interpreted by regulators, clinicians, patients, and payers. Clear differentiation between class effects, target-specific risks, route-related tolerability issues, and coincidental findings improves transparency and supports better decisions throughout development.

Programs that define safety narratives early, grounded in biology, epidemiology, and intended use context, are better positioned to navigate later scrutiny and reduce misperceptions that can limit adoption even when the overall benefit-risk profile is favorable.

7 Illustrative examples

The principles outlined above become more intuitive when viewed through common dermatologic development contexts.

7.1 Inflammatory skin disease

Drug discovery in inflammatory dermatologic conditions illustrates the importance of aligning target biology with clinical endpoints. Programs targeting central immune signaling nodes have demonstrated robust efficacy when paired with sensitive and validated outcome measures. The validation of the IL-17/IL-23 axis in psoriasis, supported by human genetics (Nair et al., 2009; Dand et al., 2025) and confirmed through rapid clinical proof-of-concept with targeted biologics, exemplifies this alignment. Conversely, agents directed at peripheral or redundant pathways often show limited benefit unless delivered locally or combined with complementary mechanisms. Key lessons include the value of early human tissue validation, careful dose selection to balance efficacy and safety, and use of rapid-onset endpoints to demonstrate proof-of-concept.

7.2 Follicular and adnexal disease

Follicular disorders highlight the importance of delivery strategy and compartment-specific targeting. As outlined in Section 3.3, diseases such as acne, folliculitis, hidradenitis suppurativa, and certain forms of alopecia are driven by biology centered in the pilosebaceous unit and adnexal structures, where systemic exposure alone may be inefficient or unnecessarily risky (Agarwal et al., 2000).

In acne specifically, optimization may also include spectrum and microbiome considerations, as well as structural determinants of target engagement, where rational design approaches that narrow antibacterial activity can support stewardship while preserving efficacy against relevant Gram-positive organisms (for example, sarecycline as a narrow-spectrum tetracycline) (Bunick et al., 2021). In parallel, formulation strategies that increase follicular exposure relative to systemic circulation are most likely to deliver favorable benefit-risk profiles. Acne also illustrates how structure-informed target engagement can support rational antibiotic discovery, as cryo-EM analysis of the Cutibacterium acnes 70S ribosome demonstrates that sarecycline inhibits translation via a two-site binding mechanism (Lomakin et al., 2023). More recent comparative cryo-EM analyses of doxycycline, minocycline, and sarecycline bound to Cutibacterium acnes and Escherichia coli 70S ribosomes show that tetracyclines can target both the 30S mRNA decoding center and the 50S nascent peptide exit tunnel/peptidyl-transferase center region, with species-specific ribosomal differences and tetracycline-specific chemical modifications influencing binding pose, occupancy, and translational inhibition (Devarkar et al., 2026). Target biology, medicinal chemistry, structural biology, and clinical positioning can converge to support a differentiated dermatologic therapeutic.

7.3 Local exposure as a differentiator

A recurring lesson across dermatologic programs is that the active moiety alone does not determine success. Formulation, cutaneous distribution, local tolerability, and systemic burden materially alter clinical performance even within the same target class, and should be treated as primary design variables rather than downstream implementation details.

PDE4 inhibitors illustrate this principle within a single target class. Comparative chemical, biochemical, and structural analyses of dermatologic cAMP PDE4 inhibitors show that structural and physicochemical differences among agents translate into meaningful differences in potency, selectivity, formulation behavior, and skin-directed performance (Wang et al., 2025). These within-class differences help explain why several topical PDE4 inhibitors have reached clinical use for different dermatologic indications despite sharing a common mechanism, and they reinforce the point that compound design and formulation are inseparable from delivery strategy.

8 Future directions

Emerging technologies are likely to accelerate dermatologic drug discovery by refining disease biology, expanding the range of tractable targets, and improving development efficiency. High-resolution human datasets, including single-cell and spatial transcriptomics, multi-omics integration, and large-scale human genetic studies, are increasingly able to resolve skin disease into distinct endotypes and progression trajectories (Castillo et al., 2023; Cho et al., 2025; Wu et al., 2021; Xia et al., 2022).

Artificial-intelligence systems that integrate these data with chemical, imaging, and clinical information may support target identification, repurposing, lead optimization, and trial design. Their value, however, depends on high-quality dermatology-specific data, transparent evaluation, expert interpretation, and close linkage to human-relevant experimental validation rather than stand-alone algorithmic decision-making (Utti et al., 2025; Bordukova et al., 2024). Network-based approaches may help prioritize disease modules, shared mechanisms, and repurposing candidates by integrating genes, drugs, targets, disease annotations, and protein-interaction data; however, such analyses should be treated as hypothesis-generating and supplemented with disease-specific molecular data from well-characterized patient cohorts (Sadegh et al., 2021; Sadegh et al., 2023).

New therapeutic modalities and delivery platforms are also reshaping what is feasible in skin disease. RNA-based agents, gene-editing approaches, cell and microbiome-directed therapies, advanced topical systems, and organoid or organ-on-chip platforms may expand the biology that can be modulated locally or systemically. Their success in dermatology will still depend on the same fundamentals outlined in this Primer: a clear TPP, early attention to human-relevant pharmacokinetics and pharmacodynamics, realistic alignment between mechanism and delivery strategy, and disciplined evaluation of benefit-risk.

9 Conclusion

Dermatologic drug discovery benefits from direct access to human tissue, sensitive clinical endpoints, a growing range of therapeutic modalities, and increasing ability to study cutaneous biology in detail. At the same time, it poses distinct challenges related to barrier function, compartment-specific delivery, immune complexity, operational execution, and long-term safety expectations.

This Primer outlines core principles that support successful dermatologic drug discovery: human relevance, clear target product profiles, iterative alignment among target, modality, and delivery strategy, careful use of translational models, operational readiness, and disciplined early clinical development with clear definitions of proof-of-mechanism and proof-of-concept. Box 1 distills these concepts into a practical checklist for discovery teams. Applying these fundamentals can improve translational efficiency and increase the likelihood that mechanistic insight leads to durable clinical benefit for patients with skin disease while preserving benefit-risk profiles that support real-world use.

Ten principles for dermatologic drug discovery.

The following principles distill key decision points that distinguish successful dermatologic programs from those that encounter preventable late-stage failures.

Table 4.

Table 5.

Correction note

A correction has been made to this article. Details can be found at: 10.3389/fddsv.2026.1939675.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fddsv.2026.1777138/full#supplementary-material

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Footnotes

Edited by: Brian Poligone, Rochester Skin Lymphoma Medical Group, PLLC, United States

Reviewed by: Carolina Victoria Alexander-Savino, University of North Carolina at Chapel Hill, United States

Marta Serafini, University of Oxford, United Kingdom