New Biologics for the Treatment of Atopic Dermatitis: Analysis of Efficacy, Safety, and Paradoxical Atopic Dermatitis Acceleration

Atopic dermatitis (AD) is a chronic, inflammatory skin disease with an eczematous rash and itching. Due to undesired adverse effects of traditional systemic treatment, there is still an unmet need for safe and effective long-term therapy for refractory AD. As our understanding of the pathogenesis underlying AD grows, novel treatments targeting specific molecules have been developed. Here, we discuss the efficacy and safety profiles of these drugs in recent clinical trials. Among their adverse effects, of particular note is AD acceleration. Although there is still debate about whether certain adverse reactions can be said to be paradoxical adverse events (PAEs), a wide range of PAEs have been reported during biological treatment for chronic immune-mediated diseases. Close surveillance of novel biologics is crucial to detect new undescribed paradoxical reactions and to shed light on the convoluted pathogenesis of AD.


Introduction
Atopic dermatitis (AD) is one of the most common chronic, inflammatory, relapsing skin diseases [1]. Up to 17.1% of adults and 22.6% of children are diagnosed with AD each year [2]. AD is a very distressing disease that is characterized by pruritus and dry skin [3]. For patients with moderate-tosevere AD, systematic treatments are often necessary [4]. The use of traditional systemic treatments (systemic corticosteroids, phototherapy, and immunosuppressants) is limited by safety risks and variable therapeutic benefits [5]. Thus, new systemic therapies have been developed recently.
For approximately 20 years, biological agents (BAs) have been widely used in various autoinflammatory and immune diseases [6]. As new emerging drugs come to market, a tradeoff between efficacy and safety is achieved [7]. Dupilumab, an IL-4 and IL-13 inhibitor, was the first biological drug approved by the FDA for the treatment of AD in adults [8]. It is noteworthy for its acceptable low side effect profile (lower rate of conjunctivitis, injection-site reactions, and infections) and high efficacy (36%-44% of patients achieve clear or almost clear skin) [9][10][11]. Other new biologics that selectively inhibit cytokines involved in the inflammatory component of AD are discussed in our review.
Among the emergent treatment adverse effects, of particular note is AD acceleration. Although there is still a debate about whether certain adverse reactions can be said to be paradoxical adverse events (PAEs), a wide range of PAEs have been reported during biological treatment for chronic immunemediated diseases [12]. PAEs are defined as the occurrence of a pathological condition that usually responds to this class of drug during biological agent therapy [13].
Insufficient data are available concerning the incidence of PAEs [13]. Most paradoxical reactions have been reported to be connected with anti-TNF therapy; however, it is possible that the number of cases will increase as the number of newly introduced biological agents increases [14]. Representative examples of PAEs are palmoplantar pustular reactions, psoriasiform reactions, and hidradenitis suppurativa (HS) in patients under treatment for rheumatoid arthritis (RA) or inflammatory bowel disease (IBD) [15]. A few reviews and case reports have described PAEs: de novo psoriasis in atopic dermatitis patients treated with dupilumab [16][17][18][19][20][21][22][23], paradoxical head and neck erythema in patients with atopic dermatitis treated with dupilumab [24][25][26], mepolizumabinduced alopecia in severe eosinophilic asthma [27], and secukinumab-induced exacerbation of previously diagnosed psoriasis [15]. Regarding atopic dermatitis acceleration, two cases reported the exacerbation of atopic dermatitis symptoms by ustekinumab in psoriatic patients [28].
This review discusses the efficacy, safety, and possible PAEs of novel biological therapies currently in phase II and phase III clinical trials for moderate-to-severe AD.

Pathogenesis
Atopic dermatitis is characterized by T cell-mediated skin inflammation and an impaired skin barrier. The acute phase of AD is characterized by a strong modulation of Th2 and Th22 immune responses, along with effects on the Th17/IL-17 and IL-23 pathways [29,30]. Barrier-disrupted keratinocytes are potent producers of immunoregulatory cytokines such as thymic stromal lymphopoietin (TSLP), IL-25, and IL-33 [31]. TSLP plays a critical role in activating the Th2 cascade [29]. TSLP and IL-25 activate dendritic cells (DCs) to express OX40 L. OX40 L/OX40 initiates type 2 immune differentiation of T cells. TSLP also induces IL-23 production by human DCs [32]. IL-33 can positively regulate the TSLPdendritic cell-OX40 L axis, participating in the induction and maintenance of the Th2 response [33,34]. The Th22 pathway is consistently activated by the Th2 pathway in AD, and both are considered key immune drivers of AD [35,36]. While acute AD pathogenesis is polarized towards Th2 and Th22 immune responses, chronic AD lesions additionally exhibit a substantial Th1 component. Th2 cells release IL-4, IL-13, IL-31, and IL-5 [31]. IL-4 and IL-13 disrupt barrier function by downregulating filaggrin (FLG) expression. Furthermore, IL-4 and IL-13 prompt inflammation through the stimulation of IgE production from plasma cells and B cell and plasma cell differentiation [37]. IL-4 and IL-13 also amplify IL-31-induced and histamine-induced pruritus [31]. IL-4 and IL-13 augment the production of CCL17, CCL22, and CCL26. These chemokines, along with IL-5, recruit Th2 cells and eosinophils [31,38]. IL-31 stimulates sensory nerves and induces pruritus [39], the itching evokes scratching, and the itch-scratch cycle aggravates barrier disruption [31].
IL-22, the leading Th22 cytokine, was suggested to have a major pathogenic role in epidermal pathology, induce keratinocyte proliferation, and downregulate FLG expression, resulting in barrier dysfunction and epidermal hyperplasia [40]. Th17/IL-17 and IL-23 pathway-associated cytokines (IL-17 and IL-12/23p40) are increased in several AD subtypes, including intrinsic [41], Asian [42], and paediatric AD [43]. IL-23 is composed of a p19 subunit in addition to a p40 subunit, which is also a component of IL-12. IL-23 also has one receptor subunit in common with IL-12 and IL-12Rβ1. IL-23 is a crucial player in the expansion and survival of Th17 T cells [44]. IL-17A and IL-17F secreted by Th17 cells can promote eosinophil production [45]. IL-12 secreted by eosinophils, dermal dendritic cells (DDCs), and inflammatory epidermal dendritic cells (IDECs) [45] induces the production of IFN-γ [46], which results in the Th2 acute phase-to-Th1 chronic phase switch in AD [45] (Figure 1).

Targeted to IL-13
3.3.1. Tralokinumab. Tralokinumab is a fully human monoclonal antibody that potently binds to and neutralizes the effects of IL-13 [50]. In a phase 2b study (NCT02347176), 204 adults were randomized 1 : 1 : 1 : 1 to receive 45, 150, and 300 mg of subcutaneous tralokinumab or placebo every 2 weeks for 12 weeks with concomitant topical glucocorticoids. At week 12, the adjusted mean difference from baseline in EASI score (primary endpoint) was significantly different than that in the placebo group, 150 mg group -4.36 (P = 0:03) and 300 mg group -4.94 (P = 0:01), while there was no significant difference in the percentage of participants with an IGA response (coprimary endpoint) at week 12 (23.0% vs. 11.8%, P = 0:10). More responses were found in participants with greater concentrations of biomarkers (DPP-4 and periostin) [50]. Tralokinumab has an acceptable safety profile with TEAEs relative to placebo (60.8%) and pooled tralokinumab (66%). The most common adverse events were upper respiratory tract infections and headache [50].
In a phase 2b, double-blind, placebo-controlled, randomized clinical trial (NCT03443024), a total of 280 patients were randomized 3 : 3 : 3 : 2 to subcutaneous injections of lebrikizumab at doses of 125 mg every 4 weeks (250 mg loading dose (LD)), 250 mg every 4 weeks (500 mg LD), or 250 mg every 2 weeks (500 mg LD at baseline and week 2) or to placebo every 2 weeks or for 16 weeks [55]. At week 16, the lebrikizumab groups showed significant dose-dependent improvements in the percentage change in EASI score ( [55].
Most TEAEs were mild to moderate. TEAEs were reported in 46.2% of placebo patients, 57.5% of 125 mg Q4W patients, 48.8% of 250 mg Q4W patients, and 61.3% of 250 mg Q2W lebrikizumab-treated patients. Among the most frequently reported AEs, upper respiratory tract infections, nasopharyngitis, injection-site pain, and fatigue occurred more frequently with pooled lebrikizumab than with placebo, and headache occurred more frequently with placebo [55].

Targeted to IL-31.
Nemolizumab is a humanized monoclonal antibody against the interleukin-31 receptor. In a recently published 16-week, double-blind, phase 3 clinical trial, a total of 215 Japanese patients (≥13 years old) with atopic moderate-to-severe dermatitis were randomized a 2 : 1 ratio to receive subcutaneous nemolizumab 60 mg or placebo every 4 weeks until week 16, with concomitant topical glucocorticoids [56]. At week 16, the mean percent change in the visual analogue scale (VAS) score for pruritus was −42.8% in the nemolizumab group and −21.4% in the placebo group (P < 0:001). The mean percent change in the EASI score (one of the secondary outcomes) was −45.9% with nemolizumab and −33.2% with placebo [56].
Nemolizumab was generally well tolerated, with an equal TEAE distribution (nemolizumab, 71%; placebo, 71%). The most commonly reported adverse event of special interest was worsening atopic dermatitis, occurring in 24% of the nemolizumab group and 21% of the placebo group; one patient discontinued nemolizumab as a result. The incidence of injection-related reactions was 8% in the nemolizumabtreated and 3% in the placebo-treated patients [56].
In a randomized, double-blind, phase 2b study (NCT03100344), a total of 226 adult patients with moderate-to-severe AD were randomized 1 : 1 : 1 : 1 to receive subcutaneous injections of nemolizumab 10, 30, and 90 mg every 4 weeks or placebo with topical agents [57]. At week 24, among the three nemolizumab-treated groups, the 30 mg dose had the best response rates. The percentage change in EASI score (the primary endpoint) was statistically significant at the 30 mg nemolizumab dose compared with the placebo (−68.8% vs. −52.1%, P = 0:016) and borderline statistically significant at the 10 mg dose (P = 0:051). With respect to the secondary endpoints, 36.8% of subjects in the 30 mg nemolizumab group achieved IGA 0/1 versus 21.1% in the placebo group (P = 0:06). Compared with the placebo, the 30 mg nemolizumab arm achieved the most distinct improvement in peak pruritus NRS (PP-NRS) scores (−68.6% vs. −34.3%, P < 0:0001) [57]. The rate of TEAEs was slightly higher in the nemolizumab groups than in the placebo group. Nonskin infections, including nasopharyngitis, upper respiratory tract infections, and gastroenteritis, occurred more frequently with nemolizumab than with placebo [57].
In a phase 2 (part A) randomized, double-blind, placebocontrolled study (NCT01986933), a total of 264 adult patients with moderate-to-severe AD were randomized 1 : 1 : 1 : 1 to receive subcutaneous nemolizumab (at a dose of 0.1 mg, 0.5 mg, or 2.0 mg per kilogram of body weight) or placebo every 4 weeks or an exploratory dose of 2.0 mg of nemolizumab per kilogram every 8 weeks [58]. At week 12, the percentage changes in the pruritus visual analogue scale (P-VAS) score (the primary endpoint) were −43.7% in the 0.1 mg Q4W group, −59.8% in the 0.5 mg Q4W group, and −63.1% in the 2.0 mg Q4W group versus −20.9% in the placebo group (P < 0:01 for all comparisons). For the secondary endpoints, changes in the EASI were −23.0%, −42.3%, and −40.9%, respectively, in the nemolizumab groups versus −26.6% in the placebo group. Changes in BSA were −7.5%, −20.0%, and −19.4%, respectively, in the nemolizumab arms versus −15.7% in the placebo [58]. The incidence of TEAEs was similar among all groups. The most frequent adverse events included exacerbation of AD, nasopharyngitis, upper respiratory tract infection, peripheral oedema, and increased creatine kinase levels. Exacerbation of AD and peripheral oedema were more common in the nemolizumab groups than in the placebo group [58].
In a 52-week double-blind extension phase II (part B) trial (NCT01986933), long-term efficacy and safety were assessed in patients who completed part A of the study. Previous placebo patients in part A were rerandomized 1 : 1 : 1 to receive subcutaneous nemolizumab (0.1, 0.5, or 2.0 mg/kg Q4W) in part B [59]. The improvement from baseline on the pruritus visual analogue scale (VAS) score was maintained or increased from weeks 12 to 64. The greatest improvement was observed in the 0.5 mg/kg nemolizumab group [59].
Among the abovementioned trials, exacerbation of AD occurred more frequently in the placebo group in the phase 2B study [57], whereas it occurred more often in the nemolizumab groups in the phase 3 study [56] and phase 2 part A [58]+part B study [59].
Among these aforementioned agents, some drugs showed more frequent atopic dermatitis TEAEs than the placebo, including nemolizumab and GBR 830 (Table 3). This atopic dermatitis as a safety outcome was not associated with the efficacy outcome, opening the question of whether some of these cases are paradoxical adverse events. According to the data we showed above and the definition of PAE, some of PAEs may have occurred with these drugs, especially nemolizumab and maybe GBR 830 and the others.
Mechanisms involved in PAEs are complicated. Limited hypotheses have been proposed based on TNF inhibitor investigations. First, an imbalance in the cytokine milieu is advanced during most PAEs [13]. The prototypical example is biological agent-induced psoriasis, due to a TNF-α/type-1 IFN cytokine imbalance: TNF-inhibitors (TNFi) block TNF-α, which results in uncontrolled activation of plasmacytoid dendritic cells (pDCs), with surplus production of IFN-α. IFN-α drives paradoxical skin inflammation [67]. Atopic dermatitis is also an autoimmune disease that involves several cytokines. These cytokines are interwoven in the pathogenesis of AD, and targeting one of these cytokines may have effects on the others. Second, individual genetic susceptibility might play a role [68]. The relationship between AD and single nucleotide polymorphisms of some genes has been investigated, such as SNPs of the interleu-kin-4/interleukin-13 receptor gene and the β-defensin 1 gene [69,70]. These polymorphisms influence the genes involved in cytokine production, and it is probable that paradoxical reactions occur in patients with an underlying genetic predisposition [14]. This might be one reason why biological agents have been used successfully in some patients with atopic dermatitis, while paradoxically these same types of atopic dermatitis are triggered by the same biological agents. Third, there is a shift in the cutaneous immune response pattern, for example, psoriatic morphology changes (plaque to pustular) [67]. Fourth, a spatial shift of immune cells to the skin [67], for example, a spatial shift of lymphocytes from the gastrointestinal system to the skin, gives rise to the development of psoriasis-like skin inflammation in patients treated for IBD [71]. Fifth, imbalance or dysfunction of regulatory T cells [67], paradoxical cutaneous sarcoidosis and granulomatous disease are prime examples. Drugs targeting TNF result in TNF-α/IL-10 cytokine imbalance and a decrease in TNFR2, followed by dysfunctional Treg increases [72].
Although a few AD cases of paradoxical reactions to biological therapy have been reported, with their increasing use for AD, an increasing number of reports of paradoxical adverse events of AD might be seen. Recently, a systematic review of paradoxical eruptions in response to targeted therapies in dermatology was published, and they identified that TNF-α inhibitors resulted in 91.2% (1869/2049) of all cases, followed by IL-17/17R (3.5%), IL-4Rα (2.7%), IL 12/23 (2.4%), and IL-23 (0.01%) inhibitors in 2049 cases of paradoxical reactions. Psoriasiform and eczematous eruptions were the most commonly reported [73].
Biological therapies associated with PAE onset are a challenging issue. Therefore, careful clinical and immunological evaluation should accompany the initiation of biological therapies. In addition, closely monitoring patients receiving biological treatment to detect such reactions is also recommended. These countermeasures will extend our clinical knowledge and shed light on our understanding of the complex immune mechanisms underlying PAEs. The understanding of these new types of adverse reactions will help us to optimize our choices for atopic dermatitis treatment.

Conflicts of Interest
The authors declare that they have no conflicts of interest.