Grand Challenges in Dermatologic Drug Discovery: Four Priorities to Transform Skin Disease Treatment

Ayman A Grada^*

• Department of Dermatology, School of Medicine, Case Western Reserve University, Cleveland, United States

Dermatology is benefiting from targeted therapies that were unimaginable a decade ago. 1 Biologics and JAK inhibitor small molecules now play important roles across multiple conditions, especially in psoriasis and atopic dermatitis, while emerging platforms such as multi-specific antibodies are under active evaluation. [2][3][4][5] Small-molecule inhibition of the Janus kinase-signal transducer and activator of transcription (JAK-STAT) pathway has broadened available options in conditions including atopic dermatitis, alopecia areata, and vitiligo, and has influenced how efficacy endpoints and treatment targets are defined in clinical studies and practice. [6][7][8][9] These advances are cause for optimism, but they have not eliminated the central problem that still defines most of our practice: even our best therapeutics often fail to provide lasting, universal, or curative responses. 10 Many patients never respond. Others lose response with time. For a large share of conditions, including acne (affecting 85% of adolescents), fibrosing disorders such as scleroderma, chronic pigmentary conditions such as melasma, and numerous genodermatoses such as epidermolysis bullosa, our tools remain limited. For other critical areas including non-melanoma skin cancers (among the most commonly diagnosed cancers worldwide and the most common cancer in the United States) 11 , actinic keratoses, the most common diagnosis at U.S. dermatology office visits 12 with pooled global prevalence of about 14% 13 , and chronic cutaneous wounds affecting an estimated 0.2-1% of the population worldwide 14,15 , a burden expected to rise with population aging 16 , current therapies often provide incomplete responses or require repeated invasive procedures. Even for diseases where drugs are highly effective, treatment is usually indefinite and relapse after treatment withdrawal is common across inflammatory dermatoses, and many patients require ongoing therapy to maintain disease control. 17,18 Precision in dermatology begins with acknowledging that diseases such as atopic dermatitis and psoriasis are syndromes rather than single entities. They consist of overlapping phenotypes that share clinical signs but diverge in molecular drivers, comorbidities, trajectories, and responses to therapy. [19][20][21][22] For instance, while currently approved biologic therapies have revolutionized atopic dermatitis treatment, around 40% or more of patients in clinical trials do not reach commonly used response thresholds such as EASI-75 at 16 weeks. [23][24][25][26] Identifying these non-responders before treatment initiation would spare months of ineffective therapy and progression. The current standard of care remains a sequential trial of mechanisms, with each attempt consuming time and exposing patients to fluctuating symptoms and potential risks. The central task is to shift from trial and error to patient-level prediction at baseline, and to maintain that precision longitudinally as the disease and the patient evolve.The field has made progress with diagnostic and prognostic biomarkers, but the most impactful tools are likely to be predictive biomarkers that match an individual patient to a specific disease mechanism and dosing strategy before treatment begins. Real advances will depend on integrating multi-omic measurements (genomics, transcriptomics, proteomics, metabolomics and related layers) with high fidelity clinical phenotyping, imaging, digital biomarkers from wearables, patient reported outcomes, and longitudinal data from routine care. In acne, for example, successive genome wide association studies and meta-analyses now implicate about 50 genetic risk loci, enriching for genes that regulate hair follicle development, pilosebaceous unit structure, lipid metabolism, and stem cell lineage decisions, and pointing toward therapeutic targets that extend beyond traditional anti-inflammatory approaches. 27 Emerging computational approaches such as digital twins, virtual patient models that integrate multiscale biological and clinical data to simulate individual treatment responses, could enable truly personalized therapeutic selection and dosing optimization. 28,29 Predictive signatures will need to be transportable across clinical sites, ancestries, and health systems, and evaluated not only for statistical discrimination, but also for practical decision value such as shorter time to disease control, fewer ineffective treatment cycles, and improved quality of life. Studies that embed pharmacogenomic and pharmacotranscriptomic testing into routine clinical workflows can demonstrate how precision prescribing performs in real-world clinics rather than only in tightly controlled trials. Adaptive, biomarker-anchored trial designs, including enrichment strategies and Nof-1 frameworks, can reduce uncertainty while minimizing prolonged exposure to ineffective options. 30,31 Finally, authoritative reviews that synthesize the fragmented biomarker landscape and propose standard panels, fit-for-purpose thresholds, and minimal reporting standards will be essential to accelerate adoption.We should measure the success of this challenge by prospective validation that includes external cohorts, by prespecified thresholds that clinicians can use at the bedside, and by transparent decision analyses that show how predictions change clinical outcomes. When a model helps a clinician select the first therapy that works and informs when to continue, taper, or switch, the promise of precision becomes real. The second challenge concerns the reliability of the bridge from discovery to proof of concept in patients. Many agents that look compelling in preclinical studies do not survive early clinical testing. 32 For example, RORγt, the lineage-defining transcription factor for Th17 cells, emerged as a highly attractive psoriasis target based on strong genetic and mechanistic data and robust efficacy of IL-17/IL-23 blockade in murine and ex vivo models. Yet systemic RORγt inhibition with the oral small molecule BI 730357 produced only moderate PASI responses in a phase II trial of moderate-to-severe plaque psoriasis, with PASI-75 rates around 30% at the highest dose -clearly below the efficacy typically seen with IL-17 or IL-23 biologics -and its long-term extension was ultimately discontinued because of limited efficacy and non-human carcinogenicity findings. 33,34 These outcomes underscore the limited predictive value of many commonly used models: acute cytokinedriven mouse models and simple two-dimensional cultures capture only fragments of human disease and often fail to reproduce its chronicity, spatial architecture, and multicellular crosstalk. 35 Meanwhile, artificial-intelligence and machine-learning approaches have raised expectations for faster target discovery and design, but their output remains hypothetical until it is rigorously validated in experimental systems and early-phase trials that faithfully recapitulate the relevant human biology. [36][37][38] The immediate need is a clear framework that prioritizes human-relevant models and sets standards for evaluating AI-assisted discovery. Advanced three-dimensional skin organoids, microengineered skin-on-a-chip systems, and humanized mouse models can capture more of the real tissue architecture and multicellular signaling than conventional two-dimensional cultures or short-term inflammatory mouse models. 39,40 Ex vivo human skin and other tissue explants can then act as a practical bridge between in silico predictions and in vivo responses. 41,42 They allow controlled testing of target engagement, downstream pathway modulation, and pharmacodynamic effects in an intact human microenvironment. 43 Model-informed translational pharmacology that links pharmacokinetics, pharmacodynamics, spatial tissue profiling, and imaging can support rational dose selection and increase the chances that early-phase trials test an exposure that truly matches the intended mechanism of action. 44,45 Equally important is a culture of prospective and transparent validation for computational methods. Studies that pit AI-nominated targets and designs against traditional pipelines, that include negative controls, blinding, and preregistered analysis plans, and that report failure modes will separate durable advances from fragile demonstrations. Regulatory science contributions that define fit-for-purpose validation standards, share reference materials, and harmonize assays will make it easier for the community to compare results and build upon them.We will know this challenge is being met when model performance correlates with Phase 2 outcomes across diverse programs, when mechanistic evidence convincingly links target engagement to clinically meaningful biomarkers, and when early clinical studies fail less often because the preclinical package was genuinely predictive. The third challenge is to broaden what we consider "druggable" and how therapies actually reach their cutaneous targets. Small molecules and monoclonal antibodies will remain essential, but on their own they cannot cover the full range of relevant biology. Many key drivers of disease lie in transcriptional programs, intracellular protein-protein interactions, and genetic variants that are hard to reach with conventional modalities. Addressing these targets will require adapting nucleic acid therapeutics, gene-editing systems, protein degraders and molecular glues, bispecific antibodies, cell-mediated delivery systems (including engineered immune and stem cells), and other advanced platforms to the specific barriers and opportunities of human skin. [46][47][48][49][50] Skin is both an obstacle and an ally. Its barrier function protects the host but also makes delivery difficult; at the same time, it allows local administration and direct visualization of response. [51][52][53] Early examples such as topical ruxolitinib cream for vitiligo and atopic dermatitis illustrate how formulation and delivery innovation can turn a systemic mechanism into a targeted cutaneous therapy. Further progress will depend on understanding the structure-function relationships that govern penetration, cellular targeting, intracellular trafficking, and expression within keratinocytes, fibroblasts, melanocytes, endothelial and immune cells, and adnexal structures. Dissolving microneedles, optimized lipid nanoparticles, polymeric or extracellular vesicle-based carriers, follicular targeting systems, and intradermal or regional delivery approaches offer ways to concentrate drugs in specific skin compartments while limiting systemic exposure. [54][55][56][57] The same delivery strategies can also be used to repurpose known agents by reshaping local exposure and improving benefit-risk profiles without altering the pharmacology of the active molecule.Studies that quantify delivery to defined skin compartments, show on-target activity in the intended cell types, and clearly describe local tolerability will be the most informative for moving new platforms forward. [58][59][60] In parallel, Chemistry, Manufacturing, and Controls (CMC) and manufacturability descriptions for cutaneous advanced therapies -including stability data, scalable processes, and well-defined critical quality attributes -can accelerate translation by resolving practical barriers that often stall promising concepts. Reviews that organize structure-function principles for skin-directed delivery into clear design rules would serve as key reference points for both academic groups and industry teams.The field will know it is making real progress when platforms reliably achieve therapeutic concentrations in targeted cutaneous cells, when pharmacodynamic evidence shows specific pathway modulation with acceptable local and systemic safety, and when integrated translational packages (pharmacology, toxicology, and CMC) are routinely mature enough to support first-in-human studies for modalities that have not traditionally been used in dermatology. 61,62 The fourth challenge is the most ambitious. Many inflammatory and pigmentary diseases relapse after treatment withdrawal because tissues retain a "memory" of disease. 63,64 This memory likely includes pathogenic tissue-resident T-cell populations that persist in previously affected sites, alongside durable changes in keratinocytes, fibroblasts, endothelial cells, and other structural cells that keep the tissue in a primed state. Neuroimmune circuits, vascular and stromal remodeling, and epigenetic marks may all reinforce this susceptibility. Current drugs generally suppress inflammatory cascades without fully erasing these tissue programs. 10,65,66 In this Perspective, we define sustained drug-free remission as maintenance of clinically meaningful disease control for at least 12 months after complete withdrawal of active therapy. Achieving this consistently will require strategies that disrupt the maintenance and reactivation of disease memory and that actively restore healthy tissue states rather than simply quieting inflammation. 67 Beyond chronic inflammatory dermatoses, the curative challenge extends to skin cancers, precancerous lesions, and chronic wounds. In actinic keratoses and field cancerization, most lesions never progress, yet the surrounding photodamaged field harbors clonally altered keratinocytes and represents a reservoir of subclinical disease. 68,69 Therapies that clear both visible and subclinical lesions while normalizing keratinocyte differentiation programs could reduce progression to cutaneous squamous cell carcinoma and decrease the need for repeated destructive procedures. 70,71 For high-risk or advanced squamous cell carcinoma, immune checkpoint inhibitors have already transformed outcomes 72 , while in the field-cancerization setting, field-directed regimens -including topical immunotherapies, photodynamic therapy, and combination strategies -aim for comprehensive field clearance rather than lesion-by-lesion ablation.In chronic wounds, failed regeneration, persistent inflammation, and cellular senescence trap tissues in a non-healing state. 73 Strategies that reactivate dormant regenerative programs, modulate senescent and dysfunctional cell populations, and restore appropriate growth factor and matrix signaling could transform outcomes for patients with diabetic foot ulcers, venous leg ulcers, pressure injuries, and other non-healing wounds. Emerging regenerative approaches -including bioengineered skin substitutes, stem-cellbased therapies, extracellular-vesicle and scaffold platforms, and pro-resolving mediators -provide concrete routes to rebuild architecture and re-establish tissue homeostasis. [74][75][76] Progress on this challenge will depend on better human mapping of disease trajectories. Longitudinal and spatial analyses that track how resident immune cells and stromal compartments are established, maintained, and reactivated in human skin can identify which cell states must be depleted, displaced, or reprogrammed. 77 Mechanistic studies should test whether targeting pathogenic tissue-resident populations, or their sustaining cytokine networks, can reset the local milieu safely. Interventions might combine short, intensive therapy to extinguish active pathways with biomarker-guided tapering and structured withdrawal. 78 Trials should prespecify remission definitions, relapse biomarkers, and stopping rules that are clinically meaningful and reproducible.In fibrosing conditions and scarring genodermatoses, regenerative strategies that rebuild architecture and barrier function will need to work alongside immunomodulation. Evidence that this challenge is being met will include biomarkers that prospectively identify readiness for withdrawal, sustained control after planned discontinuation, and molecular signatures showing a shift toward healthy tissue states. For cancer prevention, success means durable field clearance with reduced progression to invasive disease. For wound healing, it means complete and stable closure with restoration of functional skin. Authoritative reviews that integrate emerging frameworks for disease modification, cancer prevention, and tissue regeneration in skin will help knit together insights from inflammation biology, oncology, and regenerative medicine. Each of the four challenges depends on shared technical and societal foundations. At the core are standardized outcomes and biomarker platforms: harmonized clinical measures, imaging, digital readouts, and patient-reported outcomes, together with validated pharmacodynamic and predictive biomarkers, are essential for comparing interventions across studies and for meta-analyses that can inform practice and policy. 79,80 Patient engagement and co-design align research priorities with lived experience and ensure outcomes measure what matters to patients. 80 Data and model sharing, clear reporting, and where feasible open code reduce duplication and improve reproducibility, while AI/ML and emerging digital-twin approaches require well-annotated, diverse datasets to be trustworthy.Diversity and equity must be built in from the outset, with enrollment and validation across skin tones, ancestries, and geographies so that discoveries are generalizable. Global health perspectives are critical because skin disease burden is high in resource-limited settings where access to advanced diagnostics and therapeutics is limited. Climate change is emerging as a modifier of skin disease patterns: rising temperatures, air pollution, ultraviolet exposure, and extreme weather events are all associated with shifts in the prevalence, severity, and triggers of many dermatoses. 81,82 Even more extreme environments, such as spaceflight, reveal how skin adapts under combined radiation, microgravity, and confined-habitat stressors. 83,84 Astronauts frequently develop erythema, xerosis, pruritus, delayed wound healing, and features of accelerated skin aging, providing unique models for barrier disruption, impaired repair, and microbiome change that can inform terrestrial therapies. 83,85,86 Public-private partnerships can accelerate translation by pairing academic innovation with industrial scale and development capabilities. Real-world learning systems, including prospective registries, pragmatic and decentralized trials, and systematic durability and safety surveillance, can embed biomarker sampling, switching rules, and structured withdrawal into routine care. 87,88 Regulatory and policy science can provide fit-for-purpose pathways for biomarker and model qualification and for novel modalities and combinations. An emerging frontier that intersects all four challenges is the cutaneous microbiome. Dysbiosis contributes to multiple inflammatory, infectious, and neoplastic skin conditions, and microbiome-targeted interventions -including prebiotics, probiotics, postbiotics, bacteriophages, and engineered commensals -are beginning to show clinical promise as adjuncts or alternatives to traditional therapies. [89][90][91] The Dermatologic Drugs section invites: Precision, Translational, Modality, and Curative challenges are interdependent. Better models will increase the yield of AI-assisted discovery. New delivery systems will expand the range of targets we can engage safely in human skin. Predictive biomarkers will allow focused, efficient trials that test whether we are modifying disease rather than only suppressing it. Together, these efforts can move dermatology from symptomatic control to treatment-free disease control and potential cure.The tools to meet these challenges already exist. We invite colleagues across academia, industry, and policy to bring forward work that is ambitious, testable, and centered on patient benefit. The Dermatologic Drugs section stands ready to disseminate studies that advance SDG 3 and to help shape a future in which long-term disease quiescence is a standard expectation and cure is a realistic objective. We invite submissions that challenge conventional approaches, embrace innovative methodologies, and prioritize patientcentered outcomes. Together, these efforts can move dermatology from symptomatic control to treatment-free disease control and, ultimately, to curative strategies for selected conditions.