Spatial proteomics of tertiary lymphoid structures reveals correlation with immunotherapy response in Merkel cell carcinoma

BACKGROUND: Tertiary lymphoid structures (TLS) are ectopic lymphoid formations within or around tumours. They are emerging as important predictors of prognosis and response to immune checkpoint inhibitors (ICIs) in various cancers. However, in Merkel cell carcinoma (MCC), the correlation between TLS and ICI response has not yet been evaluated. Furthermore, the criteria for assessing TLS maturity vary by report, and there is no unified consensus. OBJECTIVES: To investigate the relationship between TLS and ICI responsiveness in MCC using spatial proteomics at the single-cell level. By comparing the spatial characteristics of TLS in people whose disease responds to ICI and those whose disease does not, we also aimed to identify features related to functionality of TLS, distinct from the previously reported maturity of TLS. METHODS: We analysed 24 TLS from 13 patients with MCC treated with the anti-programmed cell death ligand 1 antibody avelumab, using the PhenoCyclerTM platform with 16 oligonucleotide-labelled antibodies. Infiltrating immune cells and tumour cells were phenotyped based on machine learning, and spatial single-cell analysis was performed, including cellular neighbourhood analysis and cell-cell correlation analysis. RESULTS: Our findings revealed that the number of TLS positively correlated with survival and ICI response (P = 0.03). The number of TLS observed in tumours was significantly higher in those whose disease responded to the ICI compared with those whose disease did not (P = 0.004). Cellular neighbourhood analysis and cell-cell correlation analysis demonstrated that in functional TLS in those whose disease responded to treatment, T cells infiltrated densely within B-cell clusters, extending close to follicular dendritic cells, and that CD4+ T cells adjacent to CD8+ T cells express programmed cell death protein 1 (PD-1). In contrast, in TLS in those whose disease did not respond to treatment, although the B-cell density was high, T cells did not infiltrate B-cell cluster areas but instead resided in the peripheral areas. Furthermore, it was found that CD8+ T cells and CD4+ T cells approaching B cells expressed PD-1. CONCLUSIONS: We found a significant positive correlation between the number of TLS and ICI response in MCC. Additionally, our results suggest that the immune composition and spatial configuration within TLS may be important determinants of ICI efficacy.