Table 1.
An overview of the observed toxicity-linked clusters in both independent cohorts.
| Cellular phenotype | Cluster number (initial cohort) | Cluster number (validation cohort) | Biological relevance |
|---|---|---|---|
| PD-1hi | 9—PD-1hi CD5hi CD25hi CD27hi CD24lo CD38lo IL-10int |
12—CD5lo 20—CD5hi |
The presence of PD-1hi Bregs has been previously described in hepatocellular carcinoma (HCC) and correlated with disease progression and plasma IL-10 levels25. These exhibited a CD5hi CD24−/+ CD27hi/+ CD38dim phenotype. Incubation with HCC cell-derived supernatants induced B-cell PD-1 expression via TLR-4 ligation which when triggered induced IL-10 production and accelerated tumour progression. PD-1 + B cells from thyroid cancer patients suppressed T-cell proliferation and decreased T-cell viability, effects which could be reversed with PD-L1 blockade41. Patients with oesophageal squamous cell carcinoma had significant elevation of peripheral CD19 + CD24hi CD27 + B10 cells with elevated peripheral B-cell IL-10 production42. B cells could internalise exosomes and treatment with exosomes from these patients could significantly increase the B10 population and stimulate both IL-10 and PD-1 expression, an effect apparently mediated via TLR-4 and MAPK pathway activation. |
| PDL1hi | 2—PDL1hi CD38int CD95int TGF-β-ve IL-10-ve | Not observed | In the 4T1 breast cancer model, PD-L1 + B cells suppressed T-cell proliferation and were increased in the bone marrow, peripheral blood and spleen44. These PD-L1 + CD19 + cells were also upregulated in the peripheral blood of breast cancer patients44. Myeloid-derived suppressor cells (MDSCs) appeared to co-localise with marginal zone B cells, suggesting the potential for crosstalk between these two cell types within the tumour microenvironment. Indeed, co-culture of murine MDSCs with B cells upregulated PD-L1 expression and drove the B cells to become T-cell suppressive44. |
| 18—PDL1hi CD38int CD95lo TGFβhi IL-10lo | 4 | ||
| Plasmablast | 1—Ki67hi CD38hi CD27hi CD95hi IL-10int | 16 | Groups have shown that the main Breg subset supressing myelin oligodendrocyte glycoprotein (MOG)-induced EAE are plasmablasts which regulate the generation of MOG-specific Th1 and Th17 cells by suppressing dendritic cells through high production of IL-1014. It was suggested that the protective effect of splenic B cells was due to plasmablast differentiation in draining lymph nodes: CD1dhi CD5 + B10 cells differentiate into plasmablasts in culture67. |
| 3—Ki67int CD38int CD27int CD95hi IL-10lo | 17 | ||
| B10-type | 12—CD5hi CD24lo CD25lo CD27lo IL-10lo | 7 |
B10 cells are characterised as CD1dhi CD5 + 26. Whilst IL-10 expression is the best marker of human B10 plus B10 progenitor cells27, these cells form a subset of CD24hi CD27 + B cells27 and both of these markers were present on cluster 17, with the expression of CD24 being the highest in all of the strong IL-10 expressing cells in this cluster. B-cell depletion significantly enhanced the local reaction to applied antigen in a contact hypersensitivity model where inflammation is T-cell-dependent26, and this was due to the significant reduction of a splenic IL-10-producing CD1dhi CD5 + B-cell subset—B10 cells. Suppression of T-cell-driven inflammation was IL-10-dependent and in this model, IL-10 production was restricted to this particular B-cell subset. Adoptive transfer of CD1dhi CD5 + B cells into CD19−/− mice inhibited hypersensitivity responses but the transfer of CD1dhi CD5 + B cells from unsensitised mice or from mice sensitised with an unrelated antigen had no effect, suggesting that these B10 cells were likely to be antigen-specific. B-cell depletion prior to EAE induction with MOG significantly exacerbated the resulting disease68. B-cell depletion prior to immunisation led to significant increases in MOG-specific T-effector cells and IFN-γ and IL-17 producing CD4 + cells. Adoptive transfer of modest numbers of B10 cells completely reversed the accelerated disease of B-cell-depleted mice prior to MOG immunisation. B10 function required IL-21R signalling, CD40 interaction and MHC class II expression69. A model was proposed whereby B cells capture autoantigen, thus triggering B-cell receptor (BCR) signalling and driving IL-10-producing capacity. These B cells present antigen via MHC class II in a cognate interaction to CD4 + T cells, which produce local IL-21, which can induce B-cell IL-10 production and subsequent T-cell suppression. |
| 17—CD24hi CD25hi CD27hi IL-10lo | 8 | ||
| Transitional (IgM + IgD+ CD5 + CD21 + CD10 + CD24 + CD38 + ) | 13—CD21lo | 1—CD21lo | Transitional B cells are an intermediate population along the developmental axis, in between immature cells in the bone marrow and mature B cells in the spleen. In humans, they are characterised by high expression of CD10, CD24 and CD3870, but in reality are a heterogeneous population of cells with T1, T2 and T3 subsets having been identified with differences in surface marker expression of IgD, CD10, CD24 and CD3871. Moreover, as these cells differentiate further towards mature cells, there is a gradual loss of CD5, CD10 and IgM and upregulation of CD21 and CD2272. Functionally, the considerable overlap has been described between these cells and CD24hi CD38hi Bregs29,32. A CD27 + transitional B-cell subset has been described in humans which produces IL-10 in vivo and suppresses CD4 + T-cell proliferation and Th1 effector responses32. Cluster 15 was significantly deficient in toxicity patients in our cohort (P = 0.013). |
| 15—CD21hi CD27 + | 2—CD21hi | ||
| 16—CD5 + | 6—CD5 + |