In vitro differentiation of primary B cells identifies transitionary populations prior to plasma cell terminal differentiation.

(A) Representative plot of the B cell populations in primary tonsil samples with red gates highlighting Naïve B cells (CD27-CD38-), Memory B cells (CD27+CD38-), Germinal Center B cells (CD27intermeidateCD38intermediate), and plasma cells (CD27highCD38high). (B, C) Primary CD19+ tonsil B cells (female) were subjected to single cell paired ATAC and RNA sequencing. Cells were embedded in weighted nearest neighbor UMAP space generated using both ATAC and RNA datasets (B). Cell identities were assigned based on the expression of known lineage-defining genes and cell cycle genes (C). (D) Primary tonsil naïve B cells (female, same donor as in B and C) were cultured with cytokines and mitogens to induce plasma cell differentiation which was confirmed by staining for CD27+CD38+CD138+ cells. Cells harvested over the course of the differentiation were subjected to the same sequencing as in (B) and used to create a new UMAP embedding (E) containing both primary and in vitro differentiating cells. The integrated primary and in vitro datasets were used for pseudotime trajectory analysis with subsequent mapping from the naïve B cell cluster to the plasma cell clusters shown in (F).

Tonsil plasma cells are comprised of at least two distinct populations.

(A) Examples of differentially expressed genes between tonsil plasma cell subsets used for flow cytometry validation. (B) Representative gating strategy for the two identified plasma cell clusters based on CD31, CD38, and CD44v9. (C) CD44v9 protein expression observed in CD27+CD38+ bone marrow plasma cells. (D) Differences in IL-6 receptor expression between plasma cell subsets (n = 8 tonsil donors). (E) Representative RGS13 protein expression in tonsil plasma cell subsets and germinal center B cells. (F) IgH BCR clonal overlap between CD44v9-based tonsil plasma cell subsets, memory B cells, and germinal center B cells (n = 3 tonsil donors). Statistical significance was evaluated using a t-test in figures where only two groups are being compared. In all other instances, one-way analysis of variance (ANOVA) was performed to test for statistical significance. Data are presented as mean ± standard error of the mean (SEM). Statistical significance was defined as *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Plasma cell differentiation proceeds through a transient CD30+ intermediate.

(A) Representative gating scheme for in vitro differentiating cells based on CD27 and CD30 expression. (B) Quantification of cell populations overtime in culture. (C) Representative gating scheme for sorting CD30+ and CD30- populations. (D) Plasma cell yields from day 4 sorted CD30+ or CD30- populations (20,000 each, n=2). (E) Representative gating for sorting CD20+ and CD20- populations. (F) Plasma cell yields from day 5 sorted CD20+ and CD20- populations (20,000 each, n=2). Statistical significance was evaluated using a t-test data are presented as mean ± standard error of the mean (SEM). Statistical significance was defined as *p < 0.05 and **p < 0.01.

Orderly progression of surface markers during in vitro plasma cell differentiation.

(A) Gating scheme used to define CD27, CD31, and CD38 populations during in vitro differentiation of primary naïve B cells. (B) Comparison of CD30 expression between in vitro differentiating germinal center or naïve B cells and its correlation with IRF4 expression.

Identification of formation requirements for CD30+ cells and comparison of naïve and memory B cell differentiation.

(A) Representative gating scheme for sorted naïve (CD23+CD27-), atypical (CD23-CD27-), and canonical (CD23-CD27+) memory B cells used in (B). (B) Changes in CD27 and CD30 populations over time for in vitro differentiating naïve, atypical, or canonical memory B cells (20,000 seeded each, n = 1). (C) Formation of CD30+ cells in response to individual cytokines and mitogens or their combinations (n = 3 separately seeded wells, all from 1 tonsil donor). (D) Comparison of IL-10R and TLR9 between naïve and memory B cells (n = 3 tonsil donors). Statistical significance was evaluated using a t-test in figures where only two groups are being compared. In all other instances, one-way analysis of variance (ANOVA) was performed to test for statistical significance. Data are presented as mean ± standard error of the mean (SEM). Statistical significance was defined as *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Comparison of primary CD30 and CD44v9 populations.

(A) Representative gating scheme for comparing CD30 and CD44v9 populations in (B) highlighting the lack of compensation artefacts with fluorescence minus one and minus two controls. (B) Comparison of CD20, CD37, CD31, and CD38 expression between the subsets identified in (A) (n = 5 tonsil donors). Statistical significance was evaluated by one-way analysis of variance (ANOVA). Data are presented as mean ± standard error of the mean (SEM). Statistical significance was defined as *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

MEF2C, BAFF, and APRIL act prior to terminal plasma cell differentiation.

(A) Betweenness centrality for transcriptional regulators determined by CellOracle analysis of the CD30+ Cells and Plasma Cell 01 clusters depicted in Fig 1F. Final plasma cell yields for (B) A366 added on Day 0, 3, 5, or each day (n = 3), (C) BAFF and APRIL given alone or in combination on Day 5 (n = 3), and (D) BAFF and APRIL given alone on Day 5 or Day 7 (n = 6). Statistical significance was evaluated using a t-test in figures where only two groups are being compared. In all other instances, one-way analysis of variance (ANOVA) was performed to test for statistical significance. Data are presented as mean ± standard error of the mean (SEM). Statistical significance was defined as *p < 0.05, **p < 0.01, and ***p < 0.001.

In vitro differentiation scheme.

Primary tonsil naïve B cells were cultured with the indicated cytokines and mitogens to induce plasma cell differentiation.

Gating strategies for identification and comparison of tonsil plasma cell subsets.

(A) Gating scheme for bone marrow plasma cells (CD27+CD38+). (B) Comparison of CD44 and CD44v9 expression between tonsil plasma cell subsets and memory and germinal center B cells (n = 1 tonsil donor).

Changes in IRF4 and CD20 observed during in vitro differentiation.

(A) Comparison of CD20 expression in cells with varying levels of IRF4 expression during Day 3, 5, and 5 of in vitro differentiation staring from primary naïve B cells.

Primary germinal center B cells fail to form the CD30- and IRF4-based populations observed with naïve B cell.

(A) Daily time course of CD30 and IRF4 expression for in vitro differentiating sorted naïve B cells or germinal center B cells showing only those with IRF4 expression above the level of unresponsive cells (best visualized in Fig 4B day 3). (B) Overlaid histograms for the highest IRF4 expressing cells looking at IRF4, CD30, and CD38 expression. (C) Representative CD27 and CD38 profiles at the end of differentiation starting from germinal center or naïve B cells. (D) Representative CD44 and CD44v9 profiles at the end of differentiation starting from germinal center or naïve B cells.

Validation of linked CD30 and CD44v9 expression.

Daily time course of CD30 and CD44v9 expression of in vitro differentiating naïve B cells.

Simulated overexpression or knockout of transcription factors identified by CellOracle analysis.

Overexpression was assessed by a simulated 10 to 20% increase in transcription factor activity and complete knockout for STAT1, XBP1 (as a surrogate for rosiglitazone activity), IRF4, and MEF2C. Green coloration indicates changes predicted to progress towards the plasma cell identity while purple coloration indicates predicted changes away from the plasma cell identity.

Differentially expressed genes in A366 treated cells match with MEF2C direct targets.

(A) Differentially expressed genes between A366 and DMSO treated cells. (B) Log-2-fold change and adjusted p-values for the genes identified in (A) comparing MEF2C knockout cells to wildtype cells reported in Ref. (Ow et al., 2016).

Increased Xbp1, IRF4, and STAT signaling improve plasma cell yields.

Final plasma cell yields for (A) Rosiglitazone added on Day 0 or 3, (B) NX-1607 added on Day 5, or (C) Interferon-γ and 2-NP added on Day 5. Statistical significance was evaluated using a t-test in figures where only two groups are being compared. In all other instances, one-way analysis of variance (ANOVA) was performed to test for statistical significance. Data are presented as mean ± standard error of the mean (SEM). Statistical significance was defined as *p < 0.05.

Regulation of plasma cell differentiation and characteristic changes observed during differentiation.

(A) Identified and functionally validated regulators of different stages of plasma cell differentiation. (B) Changes in protein expression observed during in vitro plasma cell differentiation. Potential fate decisions are indicated with a red star.