MYCOSIS FUNGOIDES AND SEZARY SYNDROME THE JANUS BIFRONS OF THE CUTANEOUS T-CELL LYMPHOMAS: EVALUATION OF SPECIFIC PHENOTYPE AND GENETIC ALTERATIONS.

Mycosis fungoides (MF) and Sézary syndrome (SS) are two variants of cutaneous T-cell lymphomas (CTCL). These entities are considered two distinct diseases, based on phenotypic and genetic analyses. Neoplastic cells in MF seem to derive from T effector memory cells, instead, neoplastic SS cells have a central memory (TCM) profile. Moreover, evidence of a resident memory T cell (TRM) profile in early-stage MF, compared to migratory memory T cells (TMM) profile in advanced-stage MF, seems to explain the clinical behavior of MF during progression. Recent studies have shown more heterogeneity in CTCL phenotypes compared to that theory. Differences between MF and SS also involve the microenvironment, which predominantly expresses Th1 phenotype compared to Th2 phenotype in advanced stages and SS. A series of studies have identified multiple molecular changes in CTCL, showing a heterogeneous landscape of numerous genetic alterations without a strong differential signature between the two diseases. Molecular mutations could be found more frequently in the pathways of epigenetic and/or chromatin regulation, TCR and T-cell/ cytokine signaling, JAK/STAT signaling, and phosphoinositide 3-kinases (PI3K)/protein kinase B (Akt) and NF-kB pathway. The aim of this project is to delineate the differences between MF and SS about the phenotypic and genetic point of view, taking into account all stages of the diseases. In the first phase of the project, 15 patients with a new diagnosis of MF and SS were selected, including 5 early-stage MF (IA-IIA), 3 advanced MF (2 III MF, 1 IIB MF), 5 classical SS, 2 non-erythrodermic SS. Neoplastic T-cell immune phenotypes were evaluated on paraffin-embedded formalin-fixed sections of skin biopsies and on CD4+CD7- sorted T cells of the peripheral blood by flow cytometry (except for two CD7+ cases). About the 3 patients with stage IB MF, two of them revealed a TRM phenotype (CD69+CD103+CCR7-CD62L-). In the third of them, T cells aberrantly showed expression of CD69 and also CD62L. One IIA-MF patient showed expression of CD69 and CD103 (TRM markers) but also a partial expression of CCR7 with the negativity of CD62L, as TMM phenotype. Erythrodermic and tumoral MF patients were characterized by an infiltrate with a TMM phenotype. In SS, 3 patients showed a typical TCM phenotype (CD45RO+CD27+CCR7+CD62L+). Out of 5 CD45RA+ cases, 4 of them evidenced a T naïve phenotype (CD62L+), including an early MF (IIA). One SS showed complete negativity of CD62L and CCR7, arguing a TEMRA phenotype. Immunophenotype of blood samples revealed a better correlation with the skin of SS patients compared to MF patients. Neoplastic T cells in MF mainly had a Th1 phenotype, also in advanced stages, compared to Th2 phenotype of SS patients. This first step of the project confirmed that MF and SS are characterized by heterogeneity of phenotypes with partial correlation to the clinical features. Evidence of the same TCM phenotype in patients with classical and atypical SS suggests that criteria of this disease should be revised, including also non-erythrodermic forms. In the second phase of the project, we performed a gene expression analysis using a 770-genes panel by NanoString technologies (PanCancer Immune Profiling Panel). To obtain more statically significant data, we included 95 FFPE slides of CTCL. Total RNA from 87 samples was extracted (36 SS and 51 MF at any stage). NanoString data were processed and statistical analysis was performed within the statistical environment R. A principal component analysis was performed on all samples together, but we found a homogeneous group without any evidence of clustering between MF and SS. So, we decided to study separately the two entities. About MF, 12 differentially expressed genes (DEGs) between advanced stages and early stages were found (p-value <0.01): 9 of them resulted upregulated (CCR3, PRAME, FPR2, PMCH, AMBP, TRL7, TNFRSF10C, CFI, HAVCR2) and 3 were down-regulated (KLRB1, CD1B, CD5). Gene ontology (GO) enrichment analysis showed that those genes were significantly enriched in the regulation of the immune system, macrophage activation and Toll-like receptor (TLR) signaling. (FDR <0.05) KEGG pathways analysis revealed a not significantly representation in any pathways. Then, MF cohort was divided into two groups based on the median expression for each gene and the effects of high or low expression levels on OS were examined using the Kaplan-Meier (KM) survival curve. A list of 39 genes was identified as significant associated with OS (p-value <0.05). The HRs and p-values of those selected 39 genes were calculated through Cox regression model, revealing a 9-genes signature, which did not match with DEGs but showed a high significance from KM survival analysis. The increased expression of the following genes is significantly associated with poor prognosis: CDK1 (HR=2.06), IL6ST (HR=1.49), CCR4 (HR=1.66), ITK (HR=1.78), NOS2A (HR=1.38), IL2RA (HR=2.06), LRRN3 (HR=1.39), DUSP4 (HR=1.74). Instead, the lower expression of CCL26 (HR=0.53, p-value 0.028) is significantly associated to poor prognosis. A prognostic score was developed based on the incidence of each gene of the signature on overall survival and was defined as the linear combination of logarithmically transformed gene expression levels weighted by average Cox regression co-efficient. Patients with a high score of expression of the signature had a significantly poorer prognosis compared to patients with a low score (p-value < 0.05). Similarly, considering advanced versus early stages, the two groups with high score confirmed to have a statistically significant poorer prognosis compared to low score (p-value <0.05). This result is particularly evident comparing the score in early stages. KEGG pathway analysis showed that the 9-genes signature was significantly enriched in cytokine and JAK-STAT signaling. For SS samples, we did not evaluate DEGs because of the absence of different stages like in MF, but we performed the same analyses based on OS. A 14-gene signature was found, characterized by the association of their high expression and poor prognosis. The genes identified were IL12A (HR=2.12), IL5RA (HR=2.80), IFNL2 (HR=2.21), NT5E (HR=1.97), IL18RAP (HR=1.89), ABCB1 (HR=1.74), CCL16 (HR=1.89), CCL1 (HR=1.95), IL22RA2 (HR=1.67), IFNA17 (HR=1.69), C9 (HR=1.79), CCR9 (HR=3.31), SPANXB1 (HR=1.62), TREM1 (HR=1.76). We applied the prognostic score, showing that patients with a high score had a worse prognosis compared to patients with a low score. GO analysis showed that this 14-gene signature was significantly involved in defense response and Th1 cytokine production (FDR <0.05). KEGG pathway analysis was able to confirm a statistically significant involvement of these genes in JAK-STAT and TLR signaling. In this second phase of the project, we were able to find a gene signature for each disease with a significant prognostic value that could be useful in clinical practice, especially in early stages. Evidence of a strong involvement of JAK-STAT pathway in analysis of both MF and SS is interesting because this pathway is well known to be involved in CTCL pathogenesis and its pharmaceutical inhibition is still studied.