Non‐biologic systemic treatments for atopic dermatitis: Current state of the art and future directions

Atopic dermatitis (AD) is a prevalent chronic inflammatory skin condition with an unpredictable clinical course, associated with a significant impact on quality of life. The pathophysiology of AD involves a complex interplay between impaired skin barrier function, immune dysregulation, genetic susceptibility and environmental factors. Advances in understanding of the immunological mechanisms that underpin AD have heralded the recognition of multiple novel therapeutic targets to bolster the systemic treatment armamentarium for patients with severe AD. This review examines current and future directions of non‐biologic systemic treatments for AD, with a focus on their mechanism of action, efficacy and safety, and the key considerations to help inform treatment decisions. We summarize new developments in small molecule systemic therapies which have the potential to further advance our management of AD in this new era of precision medicine.


| PATHOG ENE S IS OF ATOPI C DERMATITIS
AD arises in genetically susceptible individuals who may be exposed to environmental triggers leading to disruption of the epidermal barrier and the initiation of chronic T cell-mediated inflammation. Twin studies have suggested that genetic factors account for the majority, as much as 82%, of individual susceptibility to develop AD, 8  Epidermal barrier dysfunction is considered the primary event in the development of AD, related to intrinsic factors, for example, filaggrin (FLG) mutations, and secondary events such as physical disruption exacerbated by the itch-scratch cycle. 9,10 The skin microbiota also plays an important role and is highly dysregulated in patients with AD. Staphylococcus aureus has long been recognized to play an important role in AD flares and chronicity, with prevalence estimates of S. aureus colonization ranging between 30 and 100%. 11 Furthermore, microbiota shifts, including reduced diversity, are also associated with disease development, flares and disease severity. 12 Epidermal barrier disruptions that have been described in the context of AD include reduced tight junction protein expression, 13 greater transepidermal water loss, 14 higher pH, 15 and reduced antimicrobial peptide expression. 16 The damaged skin barrier permits cross-talk between the skin microbiota, environmental allergens and the immune system, triggering keratinocyte production of alarmins (including IL-1β and IL-33) and cytokines such as IL-25 and thymic stromal lymphopoietin (TSLP) and the engagement of a local innate immune response through the activation of Langerhans and dendritic cells (DC). 17 Activated local immune cells signal the recruitment of circulating lymphocytes to draining lymph nodes and stimulate a class switch towards the production of IgE.
Epidermal T cell infiltration with T-helper 2 (Th2) polarization is an early hallmark of AD and is characterized by secretion of type 2 inflammatory cytokines: IL-4, IL-5, IL-13, IL-31 and IL-33. 18 IL-4 and IL-13 are central mediators of inflammation, epidermal barrier dysfunction and itch. IL-31 mediates itch by stimulating neurons expressing the IL-31 receptor subunitα and is correlated with serum levels of IL-13 and, in turn, disease severity. 19 Although activation of the Th2 immune response is common to all AD patients, there is also varied activation of Th17 and Th22 cells with production of IL-17 and IL-22, respectively, which is thought to contribute to the diversity of AD phenotypes demonstrated amongst different ethnic and paediatric populations. [20][21][22] Analysis of acute and chronic AD lesions has also demonstrated upregulation of Th1 cytokine responses (IFNγ), predominantly present in chronic lesions. 23

| CONVENTIONAL SYS TEMIC THER APIE S IN THE TRE ATMENT OF ATOPI C DERMATITIS
Conventional systemic therapies for moderate-to-severe AD include azathioprine (Aza), ciclosporin (CyA), methotrexate (MTX), and mycophenolate mofetil (MMF). All, except for CyA, are prescribed off-licence for AD and have been used for decades, despite a lack of adequately powered RCTs. However, they continue to play an important role in the treatment of AD in most healthcare settings, partly because regulatory bodies, such as NICE, insist on their use prior to prescribing novel systemic agents. A summary and decision aid for traditional systemic treatments, their efficacy, side effects and other considerations can be found in Table 1.

| Azathioprine
Aza is a prodrug that is converted to 6-mercaptopurine in vivo via purine pathway enzymes, including thiopurine methyltransferase (TPMT), to form thioguanine nucleotides which inhibit cell proliferation

Key Messages
• Conventional systemic treatments (e.g. methotrexate) continue to exert disease-modifying effects and provide long-term, cost-effective treatment options.
• Novel small molecule targets such as Janus kinase inhibitors deliver rapid and effective disease control.
• Optimized outcomes in AD depend on personalized treatment strategies with established and novel agents. Additionally, MTX appears to reduce JAK/ STAT phosphorylation and downstream proinflammatory signalling. Mycophenolate mofetil (MMF) is converted to mycophenolic acid which blocks B and T cell proliferation through the inhibition of inosine monophosphate dehydrogenase, preventing de novo nucleotide synthesis. Ciclosporin (CyA) is a calcineurin inhibitor, which blocks the dephosphorylation and nuclear translocation of NFAT, a transcription factor responsible for the upregulation of proinflammatory cytokines including IL-2. Azathioprine (Aza) is converted by TPMT to the active metabolites 6-mercaptopurine and 6-thioguanine, which antagonize purine metabolism and block DNA and RNA synthesis in actively replicating cells. Abbreviations: ATIC, 5-aminoimidazole-4-carboxamide ribonucleotide transformylase/inosine monophosphate cyclohydrolase; IL, interleukin: JAK, Janus kinase; NFAT, nuclear factor of activated T cells; RFC1, reduced folate carrier-1; STAT, signal transducer and activator of transcription; TPMT, thiopurine methyltransferase.

| Ciclosporin
CyA acts to suppress T cell activation and proliferation by blocking the nuclear factor of activated T cell (NFAT)-dependent cytokine production ( Figure 1). It is the only licensed conventional systemic treatment for AD (≥16 years of age) and remains widely prescribed. [28][29][30] CyA is a highly effective treatment for AD and has a rapid onset of action. A systematic review and meta-analysis of 12 studies showed that after just 2 weeks of treatment with 3 mg/ kg CyA, disease severity decreased by 22% (95% CI 8%-36%) and at doses ≥4 mg/kg disease severity decreased by 40% (95% CI 29%-51%). After 6-8 weeks the relative effectiveness was 55% (95% CI 48%-62%). 31 A recent network meta-analysis showed similar improvements in disease severity in those treated with CyA or dupilumab. 32 However, drug survival analyses favour dupilumab (overall probability of dupilumab and CyA survival at 16 months was 82% and 11%, respectively p < .001). 33

| Methotrexate
MTX, one of the most prescribed systemic treatments for AD, is a folic acid antagonist which blocks DNA and RNA synthesis and repair mechanisms and impairs cell division; it is also postulated to act through inhibition of the JAK/STAT pathway ( Figure 1). 34 The efficacy of MTX is similar to that of Aza, supported by an RCT of 43 adults with severe AD. 35 Adults received either MTX (10-22.5 mg/ week) or Aza (1.5-2.5 mg/kg/day); at week 12, the mean relative reduction in SCORing Atopic Dermatitis (SCORAD) score was 42% versus 39% in the Aza group (p = 0.52). 35 In terms of its relative efficacy to CyA, an open-label RCT in children receiving either MTX (7.5 mg/week) or CyA (2.5 mg/kg/day) showed no significant difference in the reduction of SCORAD (Box 1) score between the two groups, but this study was statistically underpowered and used suboptimal treatment doses. 36 MTX has been further assessed in the TREAT trial, an RCT assessing the efficacy and safety of CyA versus MTX in severe AD in children and adolescents. 37 One hundred and two patients aged  Eczema Area and Severity Index, EASI • Provides the average severity of individual clinical signs and body surface area involvement • Scores range from 0 to 72, with higher scores indicating greater severity • EASI-75 response, defined as ≥75% improvement from baseline EASI score, is a commonly utilised end point. • EASI-90 response, defined as ≥90% improvement from baseline EASI score is almost equivalent to IGA 0/1.

SCORing of atopic Dermatitis, SCORAD
• Records the extent of the atopic dermatitis (involved body surface according to the rule of nines), the intensity of clinical signs (0-3 scale) and the severity of itch and sleeplessness (evaluated by visual analogue scales). • The 'objective' version, oSCORAD, omits the itch and sleeplessness score.

Symptoms
Peak Pruritus Numerical Rating Scale, PP-NRS • Assesses itch intensity. Patient rates their peak itch sensation during the previous 24 hours, scores range from 0 to 10, with higher scores indicating greater pruritus. • A decrease of at least 4 points, PP-NRS≥4, is considered a clinically relevant end point. disease control seen after treatment cessation with CyA compared to MTX. This suggests that whilst CyA is quicker at reducing disease severity, MTX works better in the longer term with continued clinical improvement often seen for up to a year on therapy, and likely disease-modifying effects, which persist after cessation of treatment. Given its low cost, MTX provides an effective and important treatment, especially in low-resource settings.

| Mycophenolate mofetil
MMF is a prodrug of mycophenolic acid which inhibits the enzyme inosine monophosphate dehydrogenase (Figure 1). This blocks de novo guanine synthesis within T and B lymphocytes thereby inhibiting their proliferation. MMF is the least frequently used conventional immunosuppressant in the treatment of AD and, before the new era of biologics and small molecule inhibitors, was generally reserved for those with few other treatment options, who were poorly responsive or intolerant to other conventional agents. Data on its overall efficacy is limited, although one RCT compares MMF with CyA for long-term treatment of AD. In this study, participants in both arms received a 6-week run-in of 5 mg/kg/day of CyA followed by 1440 mg/day of MMF or 3 mg/kg/day of CyA, respectively, for 30 weeks. Treatment was then stopped in both arms and a 12 week follow-up period ensued. Results showed that those receiving MMF had higher SCORAD scores during the first 10 weeks of treatment.
However, during the maintenance phase (up to 30 weeks), disease activity was comparable in both groups. Furthermore, during the 12 week follow-up phase off treatment, the relapse-free period was greater in the MMF group than CyA. 38  JAKs are crucial to immune function and homeostasis with the complete blockade of JAK signalling leading to severe immunodeficiency. 40 Therefore, unlike biologics which are designed to block specific cytokine signalling pathways, JAK inhibitors exert reversible and partial effects via competitive inhibition to dampen intracellular signal transduction. Furthermore, as opposed to biologics which are highly specific in blocking their target signalling molecule, the new generation of selective JAK inhibitors demonstrate preferential specificity for different JAK isomers, but are likely to become less selective as intracellular drug concentration increases. 41 In AD, the binding of cytokines including IL -4, IL -13, IL -22, IL -31 and TSLP activate JAK1 heterodimeric receptors, mediating Th2 cell differentiation and itch via downstream effects. Three JAK inhibitors (abrocitinib, baricitinib and upadacitinib) have recently been licensed for use in AD. In contrast to biologics, which are large molecules that are administered parenterally, JAK inhibitors are small molecules that are available orally. A decision aid of licensed JAK inhibitors can be found in Table 2.

| Abrocitinib
Abrocitinib is a selective oral JAK1 inhibitor. Two phase III, multicen-  With regards to safety data, the incidence of adverse events was highest in the abrocitinib group (74% vs. 65%), however serious adverse events that led to discontinuation were equal between the two treatment groups (2%, n = 6 in each group). Mild-moderate acne was reported more frequently with abrocitinib, whilst conjunctivitis was reported more frequently with dupilumab (11% vs. 3%).
The safety of abrocitinib has been examined in a long-term analysis from phase II and III trials and one long-term extension study of the drug. 46 Total exposure in the all-abrocitinib cohort (n = 2856) was 1614 patient-years (exposure was ≥24 weeks in 1248 patients and ≥48 weeks in 606). In the placebo-controlled cohort, dose-related adverse events (200 mg, 100 mg, placebo) included nausea (14.6%, 6.1%, 2.0%), headache (7.8%, 5.9%, 3.5%), and acne (4.7%, 1.6%, 0%). Platelet count was reduced transiently in a dose-dependent manner, with 2/2718 patients  48 The most common side effects reported with F I G U R E 2 Cytokine signalling via JAKs and their inhibitors. Different cytokines bind to receptors associated with different pairs of Janus kinases. Ligand binding of the receptors leads to phosphorylation of JAK isomer pairs (either homodimers or heterodimers) which allows recruitment and phosphorylation of STAT proteins which translocate to the nucleus and regulate gene transcription. JAK inhibitors are differentially selective in their inhibition of JAKs and reduce the activity of one or more isoforms. Abbreviations: JAK, Janus kinase; STAT, signal transducer and activator of transcription.

| Upadacitinib
Upadacitinib is an oral JAK inhibitor with predominant inhibitory effects on JAK1. Like, abrocitinib, it was engineered with the objective of providing more selective inhibition to improve its risk-benefit profile. Upadacitinib is licenced for moderate-to-severe AD in adolescents and adults (12 years and above) and was the first JAK inhibitor ap-

| The safety of JAK inhibitors
The safety of JAK inhibitors has largely been established from their use in rheumatoid arthritis (RA). Tofacitinib, which reduces the

| FURTHER S MALL MOLECULE TARG E TS IN DE VELOPMENT
There are numerous other small molecule systemics in early-phase development which act on promising novel potential targets but are yet to demonstrate significant therapeutic effects with an acceptable safety profile.  with an unfavourable adverse event profile (including nausea, diarrhoea, headache, nasopharyngitis and cellulitis), leading to the discontinuation of the study. 64 Orismilast is a next-generation oral PDE-4 inhibitor targeting specific PDE-4 subtypes linked to cutaneous inflammation that is being evaluated in a Phase IIb study (NCT05469464) and has secured FDA fast track approval. BEACON, a UK-wide RCT, will provide the first comprehensive head-to-head evidence on the comparative effectiveness, tolerability and cost-effectiveness of CyA, subcutaneous MTX, and dupilumab in the treatment of moderate-to-severe adult AD. 66 It will also act as a future platform for comparisons across the large pipeline of emerging novel therapies by permitting the inclusion of additional treatment arms in a future multi-arm multi-stage platform trial which will provide critical information to patients and other stakeholders as the choice of treatments expands.

| CON CLUS ION
The fluctuant and unpredictable course of AD and the relatively limited number of approved drugs with an acceptable risk-benefit ratio continue to present major challenges in managing the disorder. As our understanding of AD develops, with increasing data on phenotypes, biomarkers and aetiopathogenesis we aim to incorporate this knowledge to provide personalized care with existing and novel therapies. JAK inhibitors, such as abrocitinib, baricitinib and upadacitinib, represent novel, orally administered, highly effective, rapid onset and non-immunogenic therapeutic options. There is huge untapped potential in further small molecular targets that will need to prove reproducible clinical efficacy and be balanced against the potential for off-target side effects and their cost, compared to conventional systemic treatments, such as methotrexate.

AUTH O R S CO NTR I B UTI O N S
The manuscript was written and prepared by AP under guidance from CF. Figures and tables were prepared by AP.
All authors have reviewed and approved the final submitted manuscript.

CO N FLI C T O F I NTE R E S T S TATE M E NT
CF is the Chief Investigator of the UK National Institute for Health He also leads the EU Trans-Foods consortium. His department has received investigator-led funding from Sanofi-Genzyme and Pfizer for microbiome work. All other authors declare no conflict of interest.