Quantitatively immunological characterization of mogamulizumab skin disorders in ATL patients

Skin disorders demonstrate highly variable phenotypical and histopathological features. Mogamulizumab, a humanized anti‐CC chemokine receptor 4 monoclonal antibody indicated for the treatment of adult T‐cell leukemia‐lymphoma, has been shown to induce skin‐related adverse events in some patients, including rare cases of Stevens‐Johnson syndrome. Hence, we aimed to elucidate immunological primary reactions in skins of mogamulizumab by quantitatively comparing any patterns of other skin disorders.


| INTRODUC TI ON
Mogamulizumab, a humanized monoclonal antibody (mAb) targeting the CC chemokine receptor 4 (CCR4), has demonstrated efficacy in patients with CCR4-expressing adult T-cell leukemia-lymphoma (ATL). However, in a phase 2 study of mogamulizumab in Japan, approximately 67% of patients with ATL reported skin disorders during or soon after mogamulizumab therapy, 1 including Stevens-Johnson syndrome (SJS) and toxic epidermal necrosis in some patients. 2,3 This could be due to autoimmune reactions induced by the loss of or reduction in regulatory T cells (Treg) known to express CCR4. 4 Indeed, some case reports have indicated reductions in Treg and increases in CD8 + cells with mogamulizumab treatment; albeit, the assessments were qualitative in nature. 2,3,5 Furthermore, mogamulizumabinduced immune response is thought to differ from responses observed in other inflammatory diseases, including autoimmune diseases; however, no reports directly comparing these responses have been published as yet.
In this study, we quantitatively analyzed the expression of

| ME THODS
This analysis was performed retrospectively using 27 archival skin samples. Of the 27 samples, seven were from patients with ATL in a phase 2 study of mogamulizumab. 1 The patients received intravenous infusions of mogamulizumab once per week for at maximum 8 weeks at a dose of 1.0 mg/kg. Twenty samples from 10 patients with psoriasis vulgaris, five patients with atopic dermatitis, and five patients with lichen planus were included for comparison in the study.
The study was compliant with the Ethical Guidelines for Medical and Health Research Involving Human Subjects, and was conducted in accordance with the Declaration of Helsinki. The study protocol was approved by the ethics committee at each participating site, and written informed consent was obtained from each patient or, if obtaining written consent was not possible, an opt-out consent process was used. The results are shown as a box and whisker plot, indicating quartiles, minimum, and maximum values. Steel's test was performed using SAS (release 9.4, SAS Institute) to compare the results between skin disorders caused by mogamulizumab and other skin disorders, with a statistical significance level of P < 0.05.

| Clinical characteristics of ATL patients with mogamulizumab-emergent skin disorders
The quantitative analysis was retrospectively performed using 27 archival skin samples. Of the 27 samples, seven were from ATL patients who experienced grade 2 or 3 rash (n = 5), SJS (n = 1), or nummular dermatitis (n = 1) in a phase 2 study of mogamulizumab. 1 Clinical characteristics of these seven patients are summarized in Table 1 in mogamulizumab-emergent skin disorders. No significant difference between all of the other skin disorders was observed in other immunological markers.

Conclusion:
The low Foxp3 + /CD8 + cell ratio of skins is the underlying reason for mogamulizumab-emergent skin disorders.

K E Y W O R D S
CD8, FoxP3, mogamulizumab

| Immunohistochemical staining of infiltrated Foxp3 + and CD8 + cells in skin
Representative immunohistochemical staining of infiltrated Foxp3+ and CD8+ cells in skin is shown in Figure S2

| Quantitative analysis of infiltrated Foxp3 + , CD8 + , CD4 + , granzyme B + , CD56 + , and MDC + cells in skin
Positively stained cells for each mAb were counted in epidermis, dermis, and basement membrane regions according to the Method. The cell count was summed up from all the selected regions. Mogamulizumab-emergent skin disorder had a lower number of infiltrated Foxp3 + cells compared with psoriasis vulgaris and lichen planus skins, while significant CD8 + cell infiltration was observed in mogamulizumab-emergent skin disorder compared with psoriasis vulgaris and atopic dermatitis skins ( Figure 1A,B).
However, the ratio of Foxp3 + /CD8 + cells in mogamulizumabemergent skin disorder was significantly lower than all of them ( Figure 1C). In inflammatory skins, the more the number of CD8 + cells were infiltrated, the more the number of Foxp3 + cells were prone to be infiltrated, but not in skin disorders induced by mogamulizumab ( Figure 1H).

| Skin compartment analysis of infiltrated Foxp3 + and CD8 + cells
Infiltrated Foxp3 + and CD8 + cells were separately analyzed from each selected regions of epidermis, dermis, and basement membrane regions. Mogamulizumab-emergent skin disorders had a lower number of Foxp3 + cells compared with those of psoriasis vulgaris, atopic dermatitis, and lichen planus skins. Especially, the decrease was statistically significant in dermis (v.s. psoriasis vulgaris, atopic dermatitis, and lichen planus) and basement membrane (v.s. psoriasis vulgaris and lichen planus) (Figure 2A). On the other hand, the increase in CD8 + cells in mogamulizumab-emergent skin disorders compared with psoriasis vulgaris and atopic dermatitis skins was statistically significant in basement membrane ( Figure 2B).
Accordingly, the ratio of Foxp3 + /CD8 + cells in mogamulizumabemergent skin disorder of dermis and basement membrane was significantly lower than that observed in psoriasis vulgaris, atopic dermatitis, and lichen planus skins ( Figure 2C). Although the Foxp3/CD8 ratio of epidermis was also low, the differences were not statistically significant.

| D ISCUSS I ON
The paucity of Foxp3 + cells and the predominant infiltration of CD8 + cells resulted in a significantly lower Foxp3 + /CD8 + cell ratio Immune dysregulation polyendocrinopathy enteropathy X-linked syndrome, a rare disease involving Treg dysfunction due to germ line mutations in the Foxp3 gene, often presents with eczema as one of the characteristic symptoms. 7 Additionally, in Treg-deficient Scurfy mice injected with α-1,3-fucosyltransferase VII-deficient Treg, which show impaired skin migration, Treg accumulation was reduced in the skin selectively, which in turn resulted in severe cutaneous inflammation. 8 These reports indicate that deficiency or dysfunction of Treg in the skin is closely related to onset of skin disorders. In our study, the absolute number of Treg in inflammatory skin diseases varied widely and was different across diseases; however, it had a positive correlation with that of CD8 + cells, suggesting the possibility that more Treg was infiltrated into the lesion, likely to suppress inflammation involving CD8 + cells. Conversely, Treg depletion in the peripheral blood induced by mogamulizumab was durable and profound in patients with ATL 1 ; the number of Treg in mogamulizumab-emergent skin lesions was consistently very scarce, although abundant CD8 + cell infiltration was observed, which implies a loss of immune suppression in ameliorating an exaggerated immune response.
In human T-cell leukemia virus type 1 (HTLV-1) Tax transgenic mice, a significant increase in CD8 + cells and decrease in Treg in the spleen were observed compared with non-transgenic mice, 9 and a majority of HTLV-1 basic leucine zipper factor transgenic mice were reported to develop skin inflammation due to dysfunction of CD4 + Foxp3 + Treg. 10 Furthermore, HTLV-1 infection has been reported to lead to reductions in Treg and increases in CD8 + cells in skin lesions from patients with HTLV-1-associated infective dermatitis. 11 These findings allude that HTLV-1 infection or ATL disease itself may predispose mogamulizumab-treated patients to an exaggerated immune response in the skin. Indeed, the frequency of skin disorders induced by mogamulizumab in the ATL study was higher than that in the peripheral T-cell lymphoma/cutaneous Tcell lymphoma study: 67% vs. 51%, with grade 3/4 of 22% vs. 11%, respectively. 1,12 In summary, the immune response in mogamulizumab-emergent skin disorders differed from that in other skin disorders. The low Foxp3 + /CD8 + cell ratio, possibly in dermis and/or basement membrane, is the underlying reason for mogamulizumab-emergent skin disorder, suggesting that different management strategies may be needed for skin disorders induced by mogamulizumab compared with other inflammatory skin disorders, including autoimmune skin disorders.
F I G U R E 2 Skin compartment analysis of infiltrated Foxp3 + and CD8 + cells. The number of (A) Foxp3 + , (B) CD8 + , and (C) Foxp3 + /CD8 + cell ratio per 0.005 mm 2 in each skin lesion from patients with mogamulizumab-emergent skin disorders, psoriasis vulgaris, atopic dermatitis, and lichen planus was shown, respectively.