Whole-Genome Genomics Correlates of Response to Anti-PD1 Therapy in Relapsed/Refractory Natural Killer/T Cell Lymphoma

The genetic landscape of Natural killer/T-cell nasal-type lymphoma (NKTL) has been recently unraveled by discoveries describing recurring mutations altering the JAK-STAT pathway, epigenetic modifiers, the DDX3X gene and genetic predisposition in the HLA-DPB1 gene but none has employed whole-genome sequencing (WGS).

Whole-genome sequencing was performed for 11 pairs of tumor-blood samples to study the association between somatic mutations and response to pembrolizumab. Interestingly, recurrent PD-L1 SRs were validated in four of the seven complete responders (CR) cases. JAK3-activating (p.A573V) mutations were also validated in another two pembrolizumab-treated patients who have achieved CR. Lastly, we also found a homozygous 3 bp insertion (p.Q131_H132insQ) in the ARID1B gene, a chromatin remodeler gene and a subunit in the SWI/SNF complex in the last remaining CR case. A recent study has also reported PBRM1-deficient and ARID2-deficient tumors correlated with better response to anti-PD1/PD-L1 therapy renal cell carcinoma. There seems to be a relationship between truncating alterations in the subunits of the SWI/SNF complex and response to PD1/PD-L1 therapy. However, the exact mechanisms behind these associations remain to be elucidated for NKTL.

Analysis of the WGS data from the four remaining progressive disease (PD) patients' tumors did not reveal similar alterations in the PD-L1 and JAK3 genes. A careful inspection was also carried out on the genes associated with major histocompatibility complex and interferon gamma pathways, which are known to associate with resistance to immune checkpoint blockade in melanoma, but no further mutation in these groups of genes was found in our cohort. Furthermore, a TP53 (p.W14X) stop-gain mutation, a hallmark tumor suppressor gene, was detected in a patient who had progressive disease after given pembrolizumab.

We went on to check if PD-L1 IHC staining could explain the response of these NKTL patients to pembrolizumab. In this study, tumors were stained and assessed for PD-L1 positivity by the same pathologist. All cases, except two cases, have greater than 20% of tumor cells stained positive for PD-L1. Interesting, both cases are CR. In addition, all four PD cases were strongly stained for PD-L1 with an average of 69% PD-L1 positive cells (range, 50% - 90%) but their outcomes were dismal. This suggests that there could be a companion biomarker that could be added to PD-L1 IHC positivity for better predictive power of response to PD1 blockade therapy.

Here we report retrospectively, for the first time, the genomic mutational profiles of anti-PD1 blockade in 11 relapsed/refractory NKTL patients using WGS data, which provide proof-of-concept data that the response to anti-PD1 is relevant and correlates with recurrent PD-L1 and JAK3 genomic alterations in this malignancy.