Extramammary Paget disease (EMPD) is a locally aggressive cutaneous malignancy that usually arises in anogenital or axillary skin. Immune checkpoint inhibitors targeting programmed cell death receptor (PD-1) and/or its ligand (PD-L1) are approved for the treatment of several types of cancer, and response to these generally correlates with increased PD-L1 expression by tumor cells. The expression of PD-L1 and composition and density of the tumor-associated immune infiltrate in EMPD have been little studied. To determine whether EMPD might be amenable to immune checkpoint blockade, we analyzed the expression of PD-1 and PD-L1 and the composition and density of the tumor-associated immune infiltrate in EMPD and evaluated associations between biomarker expression and clinicopathologic parameters. Twenty-one EMPD tumors were evaluated for tumor cell PD-L1 expression and for relative expression and distribution of CD3, CD8, PD-1, and PD-L1 in the tumor-associated immune infiltrate by using a combination of visual and image analysis (Aperio ImageScope). In addition, PD-L1 expression was assessed in 10 cases of mammary Paget disease (MPD). In EMPD cases, PD-L1 was expressed by tumor cells (3/21; 14%) and the tumor-associated immune infiltrate (15/21; 71%), and PD-1 was expressed by the tumor-associated immune infiltrate in all cases analyzed (18/18). However, PD-L1 expression by EMPD tumor cells did not correlate with the density of CD3-, CD8-, or PD-1-positive cells in the tumor-associated immune infiltrate or other clinicopathologic parameters. Furthermore, the density of CD3, CD8, PD-1, and PD-L1 in the tumor-associated immune infiltrate did not correlate with any clinicopathologic parameters evaluated with the exception that CD3 positive values were significantly higher in patients who were still alive (median, 1310 cells/mm2; range, 543–2115;) than in those who died (median, 611 cells/mm2; range, 481–908; p = 0.049). In all MPD cases, PD-L1 was absent in tumor cells but present in the tumor-associated immune infiltrate, and PD-L1 expression in lymphocytes was lower in patients with HER2/neu-positive than in those with HER2/neu-negative disease (p = 0.07). Our findings raise the possibility of therapeutic targeting of the PD-1/PD-L1 axis in EMPD.
Cancer immunotherapy has highlighted the clinical relevance of enhancing anti-tumor response of CD8+ T cells in several cancer types. Little is known, however, about the involvement of the immune system in extramammary Paget’s disease (EMPD). We exam- ined the cytotoxicity and the effector functions of CD8+ T cells using paired samples of peripheral blood and tumors by flow cytometry. Expression levels of perforin, granzyme B, IFN-g, TNF-a, and IL-2 in CD8+ tumor-infiltrating lymphocytes (TILs) were significantly lower than those in CD8+ T cells of peripheral blood. Significantly higher expression of PD- 1 was found in CD8+TILs than in CD8+ T cells of peripheral blood. A high number of CD8+ cells was significantly associated with poor overall survival (OS) adjusted with age, sex, and clinical stage (hazard ratio [HR] = 5.03, P = 0.045, 95% confidence interval [CI] 1.03– 24.4). On the other hand, the number of PD-1+ cells was not associated with OS or dis- ease-free survival (DFS). Moreover, we found that tumor cells produced immunosuppres- sive molecule indoleamine 2,3-dyoxygenae (IDO). In conclusion, CD8+ TILs displayed an exhausted phenotype in EMPD. IDO expression seemed more relevant in inducing CD8 exhaustion than PD-1 upregulation or PD-L1 expression by immune cells. Restoring the effector functions of CD8+ TILs could be an effective treatment strategy for advanced EMPD.
Extramammary Paget disease (EMPD) is a rare malignancy of the skin. Because of the scarcity of the cases, genomic alterations in EMPD are poorly characterized. To address this issue, we have interrogated 39 EMPD samples and patients blood with exome sequencing. The mutational load of EMPD was moderately high; the median prevalence of somatic mutations was above 3 mutations per megabase, a number comparable to the one of kidney renal cell carcinoma. Our study identified several putative driver events. ERBB2 mutation, as well as amplification, is frequent in our samples and likely the key driver of EMPD. The mutations are enriched in the tyrosine kinase domain of ERBB2, and are likely to cause functional alteration of the gene product. This observation is in line with previous papers reporting the efficacy of trastuzumab for EMPD. Other cancer genes including ERBB3, KMT2C, MLL4, and COL1A1 are also frequently mutated in EMPD. Driver mutation analysis by OncodriveFM identified potential novel cancer genes that are previously unreported in other cancer types. Copy number analysis identified recurrent somatic copy number aberrations. Frequent deletion peaks included CDKN2A and TSC2, both of which were important tumor suppressor genes. Mutational signature analysis showed that APOBEC3B activation, coupled with aging, was driving the somatic mutations in EMPD. We also identified evidence of APOBEC3B activation including kataegis and strand bias in the EMPD genome. In conclusion, our study provides the comprehensive landscape of somatic mutations in EMPD as well as insights into the mechanisms behind the carcinogenesis of EMPD. We have identified putative driver mutations including ERBB2 and ERBB3, which are readily targeted. We also suggest that EMPD may be treated with cancer immunotherapy, for the moderately high mutational load observed in EMPD is associated with the response to cancer immunotherapy in other cancer types. These insights provide rationale for use of systemic treatments in patients with EMPD.