The failure of significant clinical responses to immune checkpoint blockade in multiple myeloma in comparison with other lymphoid malignancies (especially HodgkinLs lymphoma) raises considerable questions as to our understanding of human T-cell anti-tumour responses in myeloma. The report by Ray et al. 1 presented in vitro data, which suggested that the inhibition of PD-1/PD-L1 ligation would be a potentially useful therapeutic modality in myeloma. Although this appears to be supported by preclinical murine studies, this data is at variance to the observations being reported in clinical studies in myeloma. 2 For immune checkpoint inhibition to be a successful modality, there must be immune recognition of malignant myeloma cells and the presence of T cells that are capable of being activated. Such activation requires T cells to have an exhausted phenotype rather than an anergic or senescent phenotype as it is the reversal of exhausted T cells rather than the generation of new T cells reacting to the tumour, which is thought to be the key to therapeutic response. The possibility that cytotoxic T cells against common pathogens may be cross reacting with neoantigens from the tumour has been raised in melanoma, 3 and it remains a possibility that cytotoxic T cells can be activated by checkpoint blockade that cross react against myeloma. It has certainly been possible to educate and generate T cells from myeloma patients to malignant plasma cell lines and tumour lysates in vitro, but whether such in vitro recognition and activation reflects the situation in vivo is unclear.

The only definitive T cells that have been shown in clinical studies to be protective and have an impact on survival in myeloma are clonal cytotoxic CD8+ T cells. 4 The presence of these T cells in~ 50% of the patients with myeloma conveys a good prognosis and they are stimulated by thalidomide maintenance therapy. 5 Long-term survivors of myeloma invariably show the presence of such cells. 6 These T cells, however, do not show the exhausted phenotype essential for the reversal of PD-1 blockade. Rather they show a telomere-independent senescent phenotype or senescence-associated secretory phenotype (SASP) 7 and as such would not be expected to be able to be activated by PD-1 blockade. Furthermore, unlike the situation described with solid tumours and tumour infiltrating lymphocytes, clonal bone marrow cytotoxic T cells express low levels of PD-1 expression suggesting that the local immunosuppressive mechanisms involving PD-1/PD-L1 interactions are not active in myeloma (Figure 1). The failure of clinical responses in clinical studies using PD-1 blockade in myeloma support this concept. 2