Profiling the molecular mechanisms driving the fate of human B cells in response to vaccination

Antigen (Ag) encounter activates B cells to proliferate and mature through the formation of germinal centers. Here somatic hypermutation of the variable regions and Immunolgobulin (Ig) isotype switching lead the high affinity Ag-specific clones to two possible differentiation outcomes: antibody (Ab) secreting plasmablasts (PB) or quiescent memory B cells (MBC). The molecular mechanism that drives the fate of a human B cell to differentiate into PB or MBC is poorly understood. Recent studies have provided new insights into the transcriptional program responsible for B cell maturation in mice or human bulk populations. The limited availability of samples and the difficulties in isolating Ag-specific MBCs from peripheral blood make this analysis particularly challenging in humans. We collected samples from human donors that received the seasonal influenza vaccine; those were processed and sorted immediately after the bleed at two different time points: day 8 and day 22 post vaccination, namely the peaks of PBs and MBCs response respectively. The blood samples were used to collect PBs, Ag-specific MBCs and naive B cells (NAIVE) by flow cytometry sorting, exploiting classical surface markers strategies. A new protocol was set up to allow qPCR analysis of multiple genes from sorted single human B cells. This protocol was first used in a pilot study on cells sorted from a first vaccinee, to perform gene expression profiling of 21 relevant genes that allowed us to discriminate the three different B cell populations. Then we up-scaled and optimized the protocol taking advantage of the 96.96 Fluidigm Dynamic Array technology, which enables to perform RT-qPCR for 96 single cells against 96 target genes in one single reaction. This new high-throughput approach was then applied to 240 single cells belonging to Ag-specific MBCs, PBs and NAIVE B cells (80 each) of a second vaccinee, to perform gene expression profiling of 96 genes involved in several pathways of B cell differentiation. By performing unsupervised hierarchical clustering on all the cells, we observed that NAIVE, PBs and MBCs clustered separately and it was possible to identify signatures of gene expression characterizing the three populations. Linear Discriminant Analysis, a dimensionality-reduction analysis, shows that PBs are particularly different from MBCs and NAIVE, that instead share more similarities. By performing statistical analysis we identified the significant differentially expressed genes, which include genes involved in known B cell expression networks and, interestingly, also novel observations (FOXP1, POU2AF1, IRF2). We then compared the gene expression profile of Ag-specific MBCs with MBCs isolated from a healthy donor, to investigate possible differences in the expression patterns of recently activated MBCs and steady-state MBCs. With this analysis we identified 16 genes with a significant differential expression level, denoting a more active profile for the recently activated MBCs isolated from the vaccinee. To further investigate the heterogeneity of Ag-specific MBCs we also recovered immunoglobulin VH sequences from the same cells by sequencing the specific PCR products. Correlation studies showed only weak association between B cell receptor (BCR) maturation (in terms of VH mutation rate) and gene expression data. Conversely, significant association was found between the expression of two genes and the Ig isotype. In particular RORα is associated with IgA, while TBX21 with IgG, in accordance to previous studies performed on mouse bulk B cell populations. The genes identified with this study could be further investigated as they represent potential markers of B cell response to human vaccination