Single-cell trajectories in Sézary syndrome

In this issue of Blood, Herrera et al¹ identified a restricted phylogeny together with a proliferation and T-cell activation signature in the skin compartment of leukemic cutaneous T-cell lymphoma (CTCL) by utilizing a multimodal single-cell analysis of paired blood and skin samples.

In recent years, both phenotypic analysis and next-generation sequencing have revealed significant interpatient diversity of Sézary syndrome (SS).^2-4 Assessing blood tumor burden has always been a challenge in CTCL, requiring international efforts to define malignant cells because of their aberrant phenotype and/or T-cell receptor gene clonotype. Herrera et al employed a CRISPR-compatible cellular indexing of transcriptomes and epitopes by sequencing (ECCITE-seq), combining cell hashing plus phenotypic analysis by antibody-derived tags (ADT) with single-guide RNA capture of individual cells identified by molecule barcoding (see figure). This multimodal approach permits the detection of transcriptome, T-cell receptor α/β (TCRα/β) and TCRγ/δ clonotype, and surface proteins expression at the single-cell level.⁵ 

Thanks to integrated bioinformatics tools, Herrera et al established the transcriptional profile of individual skin and blood clonal T cells in 4 patients with SS and 1 patient with leukemic mycosis fungoides (MF), thus defining clusters in both compartments. Interestingly, malignant skin T cells mainly clustered together, whereas blood contained several clusters revealing a greater transcriptional heterogeneity. This approach also demonstrated inter- and intraindividual heterogeneity according to the transcriptional and phenotypic profile of the malignant T cells. Although the expression of transcription factors resulted in a relatively homogeneous T helper 2 cell (Th2) or Th17 signature, the study confirmed the heterogeneity of SS cells according to their naive, central, or effector memory phenotypes as previously shown using comparative phenotypic and molecular profiling of skin and blood CD4⁺ cells.² This suggested that skin SS cells exhibit a more advanced maturation pattern than do their circulating counterparts, but the present data support that SS cells may be not fixed at a specific cell of origin or maturation stage, as recently observed in SS preclinical models.⁶ 

In this study, Herrera et al assessed single-cell copy number variation (CNV) by evaluating gene expression data. The cytogenetic profile inferred from single-cell transcriptome with gains in 8q and 17q and losses in 10q and 17q is highly concordant with data obtained by bulk comparative genomic hybridization of SS samples.⁷ Among CTCL, the characteristic cytogenetic profile of SS contrasts with the diversity seen in its mutational landscape except for some notable recurrent alterations, such as those of the TP53 gene.^3,4,8 The study did not investigate correlations with these specific gene mutations. Interestingly, CNV at the TP53 locus was absent in their leukemic MF case and in 1 SS sample. As discussed, a limitation of the CNV inference method may be the detection of small-scale alterations, especially focal deletions that are highly recurrent in SS.⁸ However, phylogenetic trees based on CNV analysis have shown parallel and shared branches of clonal evolution generating subclones supporting a continuous or stepwise migration between skin and blood. This approach did not show a clear monodirectional relationship to support a tissue of origin of the disease.

However, pseudo-time trajectory analysis of matched blood and malignant skin T cells revealed a tissue-dependent transcriptional signature masking subclonal differences among malignant skin T cells characterized by a strong and consistent upregulation of T-cell activation and proliferation signatures. This implies that the skin microenvironment determines the transcriptional program of malignant T cells and may be critical for disease initiation, as suggested by the presence of a UV signature in the mutational profile of SS cells.^3,4,8 Both skin pathobionts and the cutaneous niche may support malignant T-cell activation and expansion. The data are in accordance with differences in the proliferation index of SS between blood and skin compartments.⁹ The quiescent status of blood SS cells and the proliferative status of cutaneous SS cells were associated with the expression of CD-62L or PD-1, respectively. These findings also support the need for a multiagent therapeutic approach rather than the sequential single-agent therapies commonly used for refractory advanced-stage CTCL. For example, the response rate to mogamulizumab, an anti-CCR4 monoclonal antibody, is lower in the skin than in the blood.¹⁰ Indeed, the ECCITE-seq analysis should be applied to pre- and posttherapeutic samples to further monitor the therapeutic response in the blood and skin compartments, especially in early or limited CTCL stages. A sequential trajectory analysis would identify the transcriptional and genetic features associated with either therapeutic response or primary or secondary resistance to multimodal or single-agent therapies.