Challenges and Opportunities for Consistent Classification of Human B Cell and Plasma Cell Populations

Abstract

The increasingly recognized role of different types of B cells and plasma cells in protective and pathogenic immune responses combined with technological advances have generated a plethora of information regarding the heterogeneity of this human immune compartment. Unfortunately, the lack of a consistent classification of human B cells also creates significant imprecision on the adjudication of different phenotypes to well-defined populations. Additional confusion in the field stems from: the use of non-discriminatory, overlapping markers to define some populations, the extrapolation of mouse concepts to humans, and the assignation of functional significance to populations often defined by insufficient surface markers. In this review, we shall discuss the current understanding of human B cell heterogeneity and define major parental populations and associated subsets while discussing their functional significance. We shall also identify current challenges and opportunities. It stands to reason that a unified approach will not only permit comparison of separate studies but also improve our ability to define deviations from normative values and to create a clean framework for the identification, functional significance, and disease association with new populations.

Keywords: ABC; B cells; Breg; DN2; atypical B cells; memory; naïve; transitional.

Figure 1

Human B cell ontogeny. Illustration of the current understanding of B cell ontogeny in human B cells from late bone marrow (pro- and pre-B cells) through peripheral activation pathways, and into antibody secreting cells (ASC)/long-lived plasma cells compartments (PC). (→) Denotes clear literature supported associations, whereas (- - - >) represent associations which are theoretical but have literature supported plausibility. T1, transitional type 1 B cells; T2, transitional type 2 B cells; T3, transitional type 3 B cells; MZP, marginal zone precursor B cells; DN1, double-negative (IgD– CD27–)−1 B cells; DN2, double-negative (IgD– CD27–)−2 B cells; GC, germinal center; FDC, follicular dendritic cell.

Figure 2

Proposed B cell gating strategies. (A) CD27 vs. IgD gating strategy of CD19+ cells in a normal healthy donor representing four distinct B cell fractions: (I) switched memory (SM) B cells with associated plasmablasts (PB), (II) CD27– IgD– DN (double-negative) B cells, (III) global naïve B cell gate which includes transitional types 1–3 populations (associated FLOCK plots), (IV) Unswitched memory/IgD-only memory B cells. (B) Bm1–Bm5 gating strategy which identifies five B cell fractions which have more difficulty resolving memory from effector B cells, and naïve B cells from memory. (C) CD24 vs. CD38 gating strategy which defines five B cell fractions with good separation of memory from naïve (excluding the fraction II which represents combined activated memory and activate naïve) and naïve from transitional B cell populations. Better population resolution can be attained by subgating from the naïve population within the CD27 vs. IgD strategy followed by this CD24 vs. CD38 strategy. Of note is that transitional type-3 (T3) B cells cannot be resolved in any strategy without the use of MitoTracker® Green (MTG). (D) Representative FACS files showing (CD27– IgD–) DN B cells in a population of rheumatoid arthritis (RA), systemic lupus erythematous (SLE), scleroderma, and both acute and chronic human immunodeficiency virus (HIV) infected patients. Additionally, the DN population can be further defined by markers such as CXCR5. (E) Derived from Jenks et al. (41), Immunity, showing FcRL4 expression in atypical memory B cells from HIV patients but not SLE DN B cells, and the reciprocal expression pattern for FcRL5 (high in SLE but not HIV). This heterogeneity in expression patterns is indicative of multiple DN B cell populations, and can only be resolved by further marker subgating on DN B cells. Additionally, CD11c is highly expressed in SLE B cells and is indicative of DN2 B cells (activated effector cells) and are very different from the proposed atypical memory B cells of HIV which are thought to be anergic. PB, plasmablasts; DN, CD27– IgD– double negative B cells; T1, transitional type 1 B cells; T2, transitional type 2 B cells; T3, transitional type 3 B cells; GC, germinal center-like B cells; SM, switched memory B cells; USM, unswitched memory B cells; aNaive/aNAV, activated naïve B cells; Naïve/rNAV, resting naïve B cells; DN1, double negative 1 B cells; DN2, double negative 2 B cells, HIV, human immunodeficiency virus; SLE, systemic lupus erythematosus.

Figure 3

T-bet, CD21, and CD11c expression ex vivo and after stimulation. (A) The majority of T-bet high B cells are IgD–CD27– DN or IgD+CD27– naïve B cells with a CD11c bright and CXCR5– phenotype characteristic of DN2 and activated naïve B cells, staining of CD19 B cells from a representative SLE patient. (B) Activated naïve and DN2 have the highest levels of intracellular T-bet staining. Gating and histograms are shown for a representative SLE patient is shown on top and quantification of T-bet mean fluorescence intensity for four SLE patients is shown below. Note, while CXCR5– SWM and CD27++ CD38++ PC express some T-bet their MFI is still significantly lower than that of DN2 and activated naïve. (C) Stimulation of HCD naïve B cells with TLR7 agonist R848, cytokines, and interferon gamma but not IL4 results in both plasma cell differentiation and increased T-bet and CD11c expression with concomitant loss of CD21 and CD23 expression. (D) Naive B cells from both HCD and SLE patients gain CD11c and lose CD21 in response to stimulation with interferon gamma, R848, and cytokines. (E) CD21 expression from flow sorted in vitro differentiated B cell populations (starting population indicated above center flow plots). There was a reduction of CD21 expression within all cultures (as compared to residual CD19+ IgD+ undifferentiated resting naïve B cells), independent of starting B cell population, suggesting that the loss of CD21 is indicative of a B cell activation state and recapitulates the in vivo phenotype of DN2 and activated naïve B cells. (B,D) were adapted from Jenks et al. (41).

Figure 4

Phenotypic analysis of antibody secreting cells (ASCs). (A) Representative FACS gating strategy (from an SLE patient) for discerning both CD19+ and CD19– ASCs populations from the blood. Utilizing CD3 and CD14 as “dump” channel markers CD19+ and CD19– (non-T cells, non-monocytes) B cell populations can be extracted from the data. Subgating on the IgD- fraction and utilizing CD138 vs. CD38 allows for the breakdown of ASC populations into five fractions: Pop1, CD19+ CD38mid CD138–; Pop2, CD19+ CD38hi CD138–; Pop3, CD19+ CD38hi CD138+; Pop4, CD19– CD38hi CD138–; Pop5, CD19– CD38hi CD138+. (B) Morphological and further phenotypic alterations in circulating ASC populations in (A) as compared to circulating naïve B cells, revealing plasmablast/plasma cell programming with high BLIMP-1, CD20lo/–, sIglo, and CD27++ expression, with the association of HLA-DR and Ki-67 positivity. (C) Gating strategy of an SLE patient revealing a sizeable presence of CD19+ IgD– CD27– CD38hi CD24– pre-PB B cells. (D) FLOCK analysis of a healthy donor 6 days post vaccination exhibiting CD38mid pre-PB, CD138– plasmablasts (PB), and CD138+ plasma cells (PC) on a CD27 vs. CD38 dot plot. In circulating ASCs Ki-67 was representative of both CD138– PB and CD138+ PC, whereas pre-PB are Ki-67+ without the CD138 expression. Pre-PB show an enrichment for BLIMP-1 expression while simultaneously expressing Pax5, signifying a transitional state toward an ASC fate. (E) CD138 vs. Ki-67 plot in subgated circulating and bone marrow ASCs representing the differences in proliferation status, with bone marrow ASCs exhibiting little Ki-67 expression. This dichotomy likely signifies newly generated vs. resident PC populations.