γδ T Cells Shape Pre-Immune Peripheral B Cell Populations

Abstract

We previously reported that selective ablation of certain γδ T cell subsets rather than removal of all γδ T cells, strongly affects serum antibody levels in non-immunized mice. This type of manipulation also changed T cells including residual γδ T cells, revealing some interdependence of γδ T cell populations. For example, in mice lacking Vγ4+ and Vγ6+ γδ T cells (B6.TCR-Vγ4^−/−/6^−/−), we observed expanded Vγ1+ cells, which changed in composition and activation and produced more IL-4 upon stimulation in vitro, increased IL-4 production by αβ T cells as well as spontaneous germinal center formation in the spleen, elevated serum Ig and autoantibodies. We therefore examined B cell populations in this and other γδ-deficient mouse strains. Whereas immature bone marrow B cells remained largely unchanged, peripheral B cells underwent several changes. Specifically, transitional and mature B cells in the spleen of B6.TCR-Vγ4^−/−/6^−/− mice and other peripheral B cell populations were diminished, most of all splenic marginal zone (MZ) B cells. However, relative frequencies and absolute numbers of antibody-producing cells, and serum levels of antibodies, IL-4 and BAFF, were increased. Cell transfers confirmed that these changes are directly dependent on the altered γδ T cells in this strain, and their enhanced potential of producing IL-4. Further evidence suggests the possibility of direct interactions between γδ T cells and B cells in the splenic MZ. Together, these data demonstrate the capability of γδ T cells of modulating size and productivity of pre-immune peripheral B cell populations.

Introduction

B cell differentiation from immature precursors to antibody producing plasma cells comprises numerous stages (1, 2). Development begins in the bone marrow with common lymphocyte precursors (CLP), and progresses to immature sIgM^pos B cells, which migrate via the blood stream to the spleen. Here, new arrivals differentiate through transitional stages into mature B cells, including B2 follicular (FO), B2 marginal zone (MZ) and B1 B cells. Particularly in the serous cavities, B1 B cells are further divided into B1a and B1b B cells, which differ from one another by their expression of CD5, developmental requirements and functional roles (3, 4). Some mature B cells recirculate to the bone marrow (5, 6). B cell development is controlled by specific transcription factors (2). Further differentiation also depends on tonic BCR signaling, which is critical for incorporating immature B cells into the peripheral B cell pool (7), as well as on several additional factors and their interplay, including BAFF and NF-kB2 (8, 9). B cells can develop in the absence of IL-4 (10), but when it is present, this cytokine affects B cell development in bone marrow and periphery (11–13), and it enhances CD23 and MHCII expression (11, 14, 15), and suppresses CD5 expression by B cells (16). During B cell differentiation, B cell tolerance is established at several distinct checkpoints, including one in the bone marrow (central tolerance, BCR selection and editing) (17, 18), another during transition (more BCR selection and competition for BAFF) (17, 19), and a third during antigen activation in the germinal center, where B cells undergo somatic mutation as well as positive and negative selection (20, 21).

In contrast to the well-studied role of T cells as B cell helpers during the immune response and the differentiation of mature B cells into specific antibody-producing cells or memory cells (22–29), their role during pre-immune B cell development is unclear. On the other hand, NKT have been implicated in peripheral B cell homeostasis, especially regarding MZ B cells (30), and recent studies of hematopoietic transplantation in humans and humanized mice indicate that αβ T cells play such a role in the setting of transplantation (31, 32). However, studies in mouse strains with impaired TCR signaling suggested that γδ T cells influence antibody production already in non-immunized mice (33–35).

Subsets of murine γδ T cells as defined by their expression of different TCR-Vγ genes develop sequentially in the thymus during ontogeny (36, 37), and segregate to different organs and tissues (38, 39). Vγ1+ and Vγ4+ cells co-localize in the spleen, where they form comparatively large populations, but they are also present in other lymphoid tissues as well as in the lung and the dermal layer of the skin (40, 41). Comparison of these cells in thymus and spleen revealed different gene expression profiles (42, 43), and functional assays showed that they tend to play opposite roles during certain immune responses (44, 45). In particular, some Vγ1+ cells can produce large amounts of IL-4 whereas Vγ4+ cells have the capability of producing IL-17 (39, 46, 47). In addition, studies of the role of γδ T cells in a tumor model and during West Nile virus infection produced an indication of reciprocal regulatory interactions between these two γδ subsets during the immune response (48, 49), and we recently found in untreated mice genetically deficient in two γδ T cell subsets including Vγ4+ cells (B6.TCR-Vγ4^−/−/6^−/−) that the splenic Vγ1+ cell population was substantially altered: In this mouse strain, Vγ1+ cells were expanded, changed in composition, showed signs of activation and produced more IL-4 upon in vitro stimulation (50). Vγ6+ cells are not present in the spleen of untreated mice but they co-localize with Vγ4+ cells in skin and lung (40, 41, 51), and they are also found in tongue and female reproductive tract (38). However, at the present time, we have no indication of interactions between Vγ1+ and Vγ6+ cells.

Mindful of the functional differences between γδ T cell subsets and their ability to cross-regulate each other, we hypothesized that changes in γδ T cell composition might have effects on other immune cells and the immune responses. Our recent study examining mouse strains with genetic deficiencies in distinct γδ T cell subsets (52–54) validates this assumption with regard to serum Ig levels in non-immunized mice (50). Specifically, we found that mice deficient in Vγ1+ cells (B6.TCR-Vγ1^−/−) generally had diminished antibody levels (with the exception of IgE), whereas B6.TCR-Vγ4^−/−/6^−/− mice had increased antibody levels (with the exception of IgG3 and IgA). This mouse strain also developed autoantibodies. The net-effect of γδ T cells assessed in mice deficient in all γδ T cells (B6.TCR-δ^−/−) was neutral (for IgM, IgG3, IgG2c and IgA) or enhancing (for IgG1, IgG2b, and IgE). Several of the effects on the antibodies in γδ-deficient mice could be linked to changes in IL-4 production (50). Furthermore, B6.TCR-Vγ4^−/−/6^−/− mice displayed changes in granulocytes (50) likely to be associated with increased levels of IgE in this mouse strain (55).

Having observed such profound effect of γδ T cell composition on serum antibodies in non-immunized mice, and on IL-4 production (50), we wondered at which stage(s) in B cell development γδ T cells might intervene to effect changes in circulating antibodies. Here we report that γδ T cells begin to shape pre-immune B cell populations during the transitional stage in the spleen, eventually affecting all major populations of mature B cells. Additional data suggest that splenic γδ T cells modulate peripheral B cell populations in part through direct interactions with B cells that migrate through or reside within the MZ.

Materials and Methods

Mice

C57BL/6 mice and γδ T cell-deficient mice of the same genetic background (B6.TCR-δ^−/−) were originally obtained from The Jackson Laboratory and bred at NJH. TCR-Vγ4^−/−/Vγ6^−/− mice were a gift from Dr. K. Ikuta (Kyoto University, Kyoto, Japan), were then backcrossed onto the C57BL/6 genetic background, and re-established after 11 backcross generations. B6.TCR-Vγ1^−/− mice were a gift from Dr. Simon Carding (Norwich Med. Sch., Norwich, UK) and distributed by Dr. C. Wayne Smith (Baylor College of Medicine, Houston, TX). B6.TCR-Vγ1^tg mice were a gift from Dr. Pablo Pereira (Inst. Pasteur, Paris, France. B6.IL-4^−/− mice (C57BL/6-Il4^tm1Nnt/J) were obtained from JAX (Bar Harbor, ME) and were a gift from Dr. P. Marrack at NJH. Double knockout (KO) mice were generated by crossing the corresponding mutant strains and selecting double KO mice in the F2 generation. These mice (TCR-Vγ4^−/−/Vγ6^−/−/IL-4^−/−) were then bred as new homozygous strain. All mice were cared for at National Jewish Health (NJH) (Denver, CO), following guidelines for normal and immune deficient animals, and all experiments were conducted under a protocol approved by the Institutional Animal Care and Use Committee.

Flow cytometric analysis

Cells obtained from single cell suspensions (2×10⁵/well) were stained in 96 well plates (Falcon; BD Biosciences, Franklin Lakes, NJ) for the cell surface markers shown in the figures/tables, using the specific mAbs and derivatized reagents listed in Table 1. CD93^pos cells were detected using mAb AA4.1. Live cells were always gated based on forward and side scatter characteristics (lymphocyte gate), and unless indicated otherwise, forward scatter height and amplitude, and side scatter width and amplitude (to exclude or specifically include cellular conjugates), as well as expression of various B- or T cell markers (Table 1). All samples were analyzed on a LSRII flow cytometer, counting a minimum of 25,000 events per gated region, and the data were processed using FlowJo 9.5.2 software (FlowJo LLC, Ashland, OR).

Table 1.

Anti-mouse antibodies and secondary reagents used in flow cytometry

Antibody	Clone	Dyes/Conjugates	Source
Anti-CD3ɛ	145-2C11	PE-Cy7	Biolegend
Anti-CD5	53–7.3	PE	BD Bioscience
Anti-B220	RA3-6B2	APC-Cy7	Biolegend
Anti-CD19	1D3	APC/ Biotin	eBioscience
Anti-CD22	OX97	FITC	Biolegend
Anti-CD23	B3B4	PE-Cy7	eBioscience
Anti-CD25	PC61	PE	BD Bioscience
Anti-CD32	24G2	FITC	In house
Anti-CD38	90	PE	eBioscience
Anti-CD44	1M7	PE	BD Bioscience
Anti-CD62L	MEL-14	PE	BD Bioscience
Anti-CD69	H1.2F3	PE	BD Bioscience
Anti-CD80	16-10A1	PE	Biolegend
Anti-CD86	GL-1	PE	Biolegend
Anti-CD93	AA4.1	APC	eBioscience
Anti-CD40L	MR1	PE	BD Bioscience
Anti-ICOS	15F9	PE	BD Bioscience
Anti-MHC II	M5/114	PE	eBioscience
Anti-IgKappa	187.1	FITC	Southern Biotech
Anti-IgM	B7.6	FITC/DyLight 488	In house
Anti-NK1.1	PKH136	APC	eBioscience
Anti-TCR-β	H57-597	PE-Cy5	BD Bioscience
Anti-TCR-δ	eBioGL3	APC	eBioscience
Anti-TCR-δ	GL3	FITC/Biotin	In house
Anti-TCR-Vγ1	2.11
Anti-TCR-Vγ4	UC3
PNA		FITC	Vector Laboratories
Streptavidin		FITC	eBioscience
Streptavidin		PE-Cy5
Streptavidin		APC
Streptavidin		eFluor 450

List of staining reagents

Nomenclature

Throughout this article, we use the nomenclature for murine TCR-Vγ genes introduced by Heilig and Tonegawa (56).

T cell purification

Suspensions of splenocytes were prepared by mechanical dispersion, treated with Gey’s solution for lysis of red blood cells, and passed through nylon wool columns to obtain T lymphocyte-enriched cell preparations, as previously described (57). Enriched cells were then incubated with biotinylated anti TCR antibodies (mAb GL3, anti TCR-δ or mAb 2.11, anti TCR-Vγ1) for 15 min at 4°C, washed and incubated with streptavidin-conjugated magnetic beads (Streptavidin Microbeads; Miltenyi Biotec, Bergisch Gladbach, Germany) for 15 min at 4°C, and passed through magnetic columns to purify total γδ T cells, as previously described in detail (58). This produced cell populations containing >85% viable γδ T cells as determined by dye exclusion and staining with specific anti TCR mAbs. These cells were used for cell transfer and co-culture experiments. CD8⁺ and CD8⁻ Vγ1 subpopulations were sorted using a Sony/iCyt Synergy fluorescence activated cell sorter based on their distinctive phenotypes (CD3⁺TCR-β⁻TCR-δ⁺TCR-Vγ1⁺CD8⁺ and CD3⁺TCR-β⁻TCR-δ⁺TCR-Vγ1⁺CD8⁻, respectively).

ELISPOT assay for Ig-producing B cells

High protein-binding microlon ELISA plates were coated with either 3 µg/ml polyclonal goat anti-mouse Ig(H+L) or goat anti-mouse IgG1 (Southern Biotechnology Associates) in 1× PBS overnight at 4°C. Plates were blocked with ICTM for 30 min at room temperature. Splenocytes (5.0 × 10⁵) were added to the first well of a row and titrated in serial 2-fold dilutions in ICTM. After 7 hrs, plates were washed three times with 0.05% Triton X-100 in 1x PBS. Biotinylated goat anti-mouse detecting antibodies were then applied at 0.5 ng/ml in blocking buffer and allowed to incubate overnight at 4°C. Biotinylated anti-IgKappa and anti-IgG were paired with goat anti-mouse Ig(H+L) coating antibody to determine total Ig- and IgG-producing cells, respectively. For detecting IgG1-producing cells, biotinylated anti-IgG was used to pair with anti-IgG1 coating antibody. Plates were washed in 1× PBS, and streptavidin-alkaline phosphatase (Biolegend) was applied at a dilution of 1:2000 in blocking buffer. After washing, plates were developed in 100 mM Tris-HCl (pH 9.5), 100 mM NaCl, and 10 mM MgCl₂ with 1 mg/ml 5-bromo-4-chloro-3-indolyl phosphate (Pierce) for 2 hrs at 37°C. Plates were scanned into TIFF images for blinded counting.

Adoptive transfer of γδ T cells

For in vivo cell transfer, magnetic bead-purified cells were washed in PBS, re-suspended to a concentration of 2.5×10⁷ cells/ml in PBS, and 5×10⁶ cells/mouse were injected in 200 µl PBS via the tail vein of the transfer recipient.

Co-culture of B cells and γδ T cells

For co-culture experiments, MZ B-rich B cells were purified by labeling splenocytes from B6.TCR-Vγ1^−/− mice with anti CD43-conjugated beads, followed by magnetic separation. The flow through was collected and contained >90% viable B220⁺CD43⁻ B cells. These purified B cells at 2×10⁶ per ml in culture medium were incubated with or without the addition of total Vγ1^pos γδ T cells (1×10⁶ cells/ml), or with CD8^pos or CD8^neg fractions of Vγ1^pos cells (0.5×10⁶ cells/ml). Cells were collected after 60 hours of cell culture, stained with the indicated antibodies, and analyzed by flow cytometry.

In vivo labeling of spleen cells

We followed the protocol described by Barral et al. (59), with minor modifications. Briefly, mice were injected via the tail vein with an antibody specific for the pan-lymphocytic marker CD45 (mAb clone 104, anti CD45.2 conjugated with PE or Pacific Blue), at 2 µg antibody mouse in 200 µl PBS, euthanized after 20 min of in vivo incubation, and single cell suspensions of spleen cells were prepared after first perfusing the spleens with PBS to wash out unbound antibodies. Splenocytes were then stained with specific antibodies to identify lymphocyte subsets as indicated in the figures, and analyzed cytofluorimetrically.

Statistical analysis

Data are presented as means +/− SD. The unpaired t test was used for two group comparisons, and ANOVA was used for analysis of differences in three or more groups. Statistically significant levels are indicated as follows: NS, not significant, * p < 0.05, ** p < 0.01, *** p < 0.001.

Results

Altered B cells in the peritoneal cavity, lymph nodes and blood; no change of immature B cells in bone marrow

We previously reported that non-immunized mice deficient in individual γδ T cell subsets have changed levels of serum immunoglobulins, and one strain developed autoantibodies (50). The latter strain, which is deficient in Vγ4^pos and Vγ6^pos γδ T cells (B6.TCR-Vγ4^−/−/6^−/−), has much elevated serum levels of IgE and IL-4, and T cells (both αβ T cells and residual γδ T cells) that secrete larger quantities of this cytokine (50). IL-4, originally termed B cell stimulatory factor-1, acts on resting B cells, drives their maturation (60) and, when over-expressed, can promote autoimmunity and allergic-like inflammatory disease (61). These observations, which uncovered an altered regulatory environment for B cells in partially γδ-deficient mice, led us to examine the B cells themselves.

Comparing wt and B6.TCR-Vγ4^−/−6^−/− adult peripheral B cell populations, we found several differences in lymph nodes, blood and peritoneal cavity (Figure 1, Table 2). Total B cells (IgM^posB220^pos) in the lymph nodes of all, and peritoneal cavity of the older mutant mice were dramatically decreased, mainly because of decreases in B2 B cells (IgM^posB220^posCD23^posCD43^neg). This was unexpected given that IL-4 (50) and BAFF (not shown), which are elevated in these mice, promote B cell growth (8, 60). Even more surprising, the removal of IL-4 in B6.TCR-Vγ4^−/−/6^−/−/IL-4^−/− mice restored peripheral B cell numbers (Table 2). Although numbers of B1 B cells (IgM^posB220^posCD23^negCD43^pos) were not substantially changed, their phenotype was altered, in blood and peritoneal cavity (Figure 1, Table 2). B1 B cells have been divided into two subsets based on the expression of the inhibitory receptor CD5 (16). In B6.TCR-Vγ4^−/−/6^−/− mice, cells expressing CD5 at high levels (B1a B cells) were much diminished in numbers and relative frequency whereas cells expressing CD5 at low levels (B1b B cells) were increased. Given that IL-4 inhibits CD5 expression (16), this change was predictable. Moreover, the restored composition of B1 B cells in B6.TCR-Vγ4^−/−/6^−/−/IL-4^−/− mice confirmed the inhibitory role of IL-4 (Table 2). The changes were quite stable and seen in mice between the ages of 4 and 20 weeks (Figure 1F,K). Subsequently, we also examined mice deficient in Vγ1^pos γδ T cells (B6.TCR-Vγ1^pos) and in all γδ T cells (B6.TCR-δ^−/−)(Table 2). In lymph nodes and peritoneal cavity, B6.TCR-δ^−/− mice had normal or somewhat enlarged B cell populations, including both B2 and B1 B cells, with a normal B1a/B1b B cell ratio (peritoneal cavity), and similar results were obtained with B6.TCR-Vγ1^pos mice although peritoneal B1a B cells were increased here, presumably due to the retention of IL-4-suppressive (Vγ4^pos) and absence of IL-4-producing (Vγ1^pos) γδ T cells in these mice (50). In sum, the data show that the particular γδ deficiency in B6.TCR-Vγ4^−/−/6^−/− mice has a large effect on peripheral B cells whereas the absence of Vγ1^pos γδ T cells, or of all γδ T cells, affects peripheral B cell populations more subtly (but see distinct effects with splenic B cells, below).

Figure 1. Influence of γδ T cells on peripheral mature B cell populations.

(A–K) Comparison of B cell populations in female C57BL/6 (B6) and B6.TCR-Vγ4^−/−/Vγ6^−/− (Vγ4^−/−/6^−/−) mice. Total B cells, B1 and B2 B cells, and B1a and B1b B cells were identified using the indicated markers. Panels A, D and G show representative staining profiles of individual 8 wks old mice. (A–C) Lymph node B cells: Panels B and C show total numbers and relative frequencies of inguinal lymph node B cells (both sides pooled), respectively, comparing mice of different ages. The frequency of total B cells was calculated relative to total lymphocytes, and the frequencies of B1 and B2 B cells relative to total B cells. (D–F) Blood B cells: Panels E and F show relative frequencies (total B, B1 and B2) and B1a/B1b ratios, respectively, comparing mice of different ages. (G–K) Peritoneal cavity B cells: Panels H and I show total numbers and relative frequencies, respectively, comparing mice of different ages. Panels J and K show total numbers of B1a and B1b cells and their ratio, respectively, comparing mice of different ages. n ≥ 4 mice per group. For clarity, only comparisons where no significant differences were found are marked (NS), all others are significant at a P value of <0.05 or less.

Table 2.

Peripheral B cells in genetically γδ T cell-deficient mice

Percentage (Absolute cell #)		B6 (n=8)	δ−/− (n=4)	Vγ1−/− (n=5)	Vγ4−/−/6−/− (n=8)	Vγ4−/−/6−/−/IL4−/− (n=7)
Lymph node
B/Lym (%)		23.9±9.8	24.9±0.9	15.0±2.2	8.5±3.0§	27.5±7.9
(B cells (x106))		(1.9±0.7)	(2.9±0.3)*	(1.6±0.5)	(0.4±0.2)¶	(2.8±1.0)
B2 B/B (%)		94.0±1.6	92.9±0.9	62.3±6.2¶	79.9±4.5¶	91.5±2.3
(B2 B cells (x106))		(1.7±0.6)	(2.7±0.3)*	(1.0±0.4)¶	(0.3±0.1)¶	(2.6±0.8)
B1 B/B (%)		2.1±0.6	3.0±0.4	6.3±1.9¶	13.5±4.8¶	3.3±0.8
(B1 B cells (x103))		(54.7±39.6)	(92.9±21.6)	(100.2±47.5)	(43.8±22.5)	(97.2±52.2)
Peritoneal cavity
B/Live (%)		25.1±1.6	21.4±4.0	32.0±7.5*	13.3±5.3¶	26.0±9.5
(B cells (x103))		(394.0±148.6)	(530.2±31.8*)	(1027.5±465.3¶)	(227.2±153.3)	(566.7±243.8)
B2 B/B (%)		18.5±4.1	20.7±7.0	8.2±2.4¶	21.9±7.3	21.8±9.1
(B2 B cells (x103))		(57.4±17.6)	(54.3±4.8)	(84.6±40.2)	(41.0±25.9)	(122.5±87.4)
B1 B/B (%)		67.5±5.6	66.2±7.6	70.8±2.7	60.8±6.8	57.8±17.9
(B1 B cells (x103))		(269.5±112.2)	(346.5±31.2)	(727.4±334.2§)	(143.7±104.84)	(384.1±176.0)
B1a B/B1 B (%)		83.5±2.3	83.8±2.8	88.0±1.8	20.5±7.1¶	74.8±6.1*
(B1a B cells (x103))		(226.7±98.0)	(290.9±35.7)	(642.8±305.9§)	(34.4±32.9¶)	(273.9±137.3)
B1b B/B1 B (%)		16.2±2.3	15.9±2.7	11.5±1.7	79.1±6.8¶	25.0±6.2*
(B1b B cells (x103))		(42.0±14.7)	(54.3±4.8)	(81.5±31.6*)	(109.0±73.8*)	(109.8±57.4*)
B1a/B1b		5.3±0.9	5.4±1.0	7.8±1.4§	0.3±0.1¶	3.2±1.1§
Blood
B/Lym (%)		50.9±10.4	66.9±4.1*	ND	48.1±5.3	68.1±11.0
B2 B/B (%)		83.1±3.2	85.3±1.9	ND	63.1±2.1¶	89.6±2.3
B1 B/B (%)		2.2±0.6	2.0±0.4	ND	3.7±0.2¶	1.4±0.5
B1a B/B1 B (%)		73.3±6.5	70.5±7.1	ND	33.6±7.0¶	67.8±3.5
B1b B/B1 B (%)		26.0±6.6	29.5±7.1	ND	65.4±7.0¶	32.3±3.5
B1a/B1b		3.0±0.9	2.5±0.8	ND	0.5±0.2¶	2.1±0.3
Spleen
Total B cells
B/Lym (%)		58.0±3.0	52.5±8.7	38.3±10.6§	29.3±1.9¶	52.0±2.7*
(B cells)		(45.3±10.2)	(40.1±8.7)	(41.1±9.8)	(11.2±2.4)¶	(45.6±21.3)
Mature B cells
Mature B/Lym (%)		46.8±2.0	43.9±9.7	32.3±8.8§	25.7±1.8¶	46.9±4.5
(Mature B cells)		(36.3±6.7)	(33.6±9.3)	(34.6±8.2)	(9.7±1.8)¶	(41.4±20.8)
MZB/Mature B (%)		4.3±0.8	3.8±1.1	9.5±1.2¶	0.6±0.2¶	4.7±1.8
(MZB)		(1.6±0.4)	(1.3±0.6)	(3.3±0.9)¶	(0.1±0.0) ¶	(1.7±0.2)
FOB/Mature B (%)		84.0±1.4	81.2±3.6	74.9±9.2*	62.7±6.5¶	80.0±2.2
(FOB)		(30.5±5.9)	(27.4±8.1)	(25.9±7.2)	(6.1±1.5)¶	(32.9±15.8)
NewB/Mature B (%)		8.3±1.3	11.2±4.4	10.5±4.0	32.2±8.9¶	11.4±3.6
(NewB)		(3.0±0.6)	(3.6±0.9)	(4.8±3.3)	(3.1±0.9)	(5.3±4.3)
Immature B cells
Immature B/Lym (%)		9.5±1.1	8.6±2.4	6.1±2.2§	3.9±0.4¶	6.1±0.9¶
(Immature B)		(8.1±1.2)	(3.4±0.9)	(2.6±1.3)	(2.1±1.1)¶	(5.7±1.0)*
T1/Imm (%)		27.4±1.2	25.4±1.0	27.6±0.9	48.0±1.2¶	25.8±3.3¶
(T1)		(2.1±0.2)	(1.5±0.3)	(2.0±0.3)	(1.0±0.1)¶	(1.6±0.5)
T2+T3/Imm (%)		59.9±1.5	59.8±1.6	61.1±1.2	33.4±1.1¶	55.9±3.5
(T2+T3)		(5.0±0.8)	(3.3±0.6)§	(4.3±0.5)	(0.7±0.1)¶	(3.0±0.5)
B2 cells
B2/B (%)		76.9±3.0	78.9±0.4	32.5±2.8¶	58.7±5.2¶	80.2±0.5
(B2)		(36.5±10.4)	(31.6±8.0)	(13.2±3.6)¶	(5.9±1.1)¶	(37.8±18.3)
B1 cells
B1/B (%)		2.1±0.3	2.6±0.9	2.8±0.7	10.8±3.3¶	3.2±1.9
(B1)		(0.8±0.2)	(1.2±0.6)	(1.1±0.4)	(0.9±0.3)	(1.3±1.2)

^*p<0.05

^§p<0.01

^¶p<0.001; ND, not determined

Peripheral B cells in various γδ T cell knockout mice. B cells in lymph nodes, peritoneal cavity, blood and spleen were identified based on their surface phenotype as described in Figures 1 and 3, and enumerated in individual mice. Absolute numbers (in parentheses) and relative frequencies are shown. n = 4–8 mice per group, p values, as indicated in the table.

We next examined immature B cells in the bone marrow. Comparing bone marrow B cells from wt, B6.TCR-δ^−/−, B6.TCR-Vγ1^−/− and B6.TCR-Vγ4^−/−6^−/− mice, we hardly found any difference in immature B cell-types (Figure 2A, Table 3), including two developmentally late fractions of immature bone marrow B cells (Fractions E, E’) that give rise to some of the mature B cells in bone marrow as well as bone marrow emigrants (5). Only one fraction (Fraction F), which represents the mature IgD^pos B cell population in bone marrow, was drastically reduced in B6.TCR-Vγ4^−/−6^−/− mice, both in absolute numbers and relative frequency (Figure 2A,B,C and Table 3). However, this fraction consists to a large extent of recirculating peripheral B cells (5, 6). To determine whether halted maturation in bone marrow or diminished recirculation are responsible for the loss of mature bone marrow B cells in B6.TCR-Vγ4^−/−6^−/− mice, we examined bone marrow B cells at several ages, with the older mice having a larger peripheral B cell compartment and increased potential for recirculation (Figure 2B, C). At 4 wks of age, fraction F mature bone marrow B cells in wt and B6.TCR-Vγ4^−/−6^−/− mice were essentially the same, both in absolute numbers and relative frequency, whereas between 8–20 wks of age, mature bone marrow B cells mice increased substantially in wt mice but not in B6.TCR-Vγ4^−/−6^−/− mice. This result is consistent with the interpretation that fewer returning peripheral B cells account for the smaller number of mature B cells in the bone marrow of B6.TCR-Vγ4^−/−6^−/− mice, and that B cell development during the bone marrow stages is unaffected by γδ T cells.

Figure 2. Immature B cells in bone marrow are not affected by γδ T cells.

(A) Comparison of bone marrow B cell populations in 8 wks old female C57BL/6 (B6), B6.TCR-δ^−/− (δ^−/−), B6.TCR-Vγ1^−/− (Vγ1^−/−) and B6.TCR-Vγ4^−/−/Vγ6^−/− (Vγ4^−/−/6^−/−) mice, representative examples. Bone marrow immature and mature B cells (Hardy fractions A-F) were identified using the indicated markers whereby fractions A-C were derived from B220^posCD43^pos and fractions D-F from B220^posCD43^neg cells. Fraction F represents mature B cells within the bone marrow. (B) Comparison of total and mature bone marrow B cell populations (fraction F) in female B6 and Vγ4^−/−/6^−/− mice at ages 4–20 wks in absolute numbers/mouse (left femur plus tibia) and (C) in frequency relative to total live cells or fractions D-F, respectively. For panels B and C, n = 4–8 mice per group. Only significant differences between wt and the γδ-deficient mice are marked. *P<0.05, **P<0.01, ***P<0.001

Table 3.

B cells in bone marrow from genetically γδ T cell-deficient mice

Percentage % (Absolute cell # x106)	B6 (n=13)	δ−/− (n=6)	Vγ1−/− (n=5)	Vγ4−/−/6−/− (n=4)	Vγ4−/−/6−/−/IL4−/− (n=5)
B cells/live cells	29.0±3.9	29.7±4.0	23.5±3.6*	24.4±1.7§	27.7±3.6
(B cells)	(6.4±2.1)	(6.7±2.2)	(4.6±1.7)	(4.9±0.9)	(6.7±2.1)
Fr’A/(A–C)	51.0±5.1	47.3±6.0	40.5±7.0*	56.1±2.2	53.5±7.0
(Fr’A)	(0.4±0.1)	(0.4±0.1)	(0.3±0.1)§	(0.5±0.2)	(0.3±0.1)
Fr’B/(A–C)	19.3±1.6	20.2±3.0	30.4±9.3	18.2±5.1	18.4±4.0
(Fr’B)	(0.2±0.1)	(0.2±0.0)	(0.2±0.2)	(0.2±0.0)	(0.1±0.0)
Fr’C/(A–C)	26.2±4.3	28.7±3.6	24.2±3.6	25.5±3.5	23.6±5.0
(Fr’C)	(0.2±0.1)	(0.2±0.0)	(0.2±0.1)	(0.3±0.1)	(0.1±0.0)
Fr’D/(D–F)	59.5±4.1	59.7 ±3.8	47.8±5.5§	72.4±9.8*	64.4±4.7
(Fr’D)	(3.3±1.2)	(4.0±1.1)	(1.9±0.9)*	(3.1±0.6)	(3.8±1.2)
Fr’E/(D–F)	15.9±2.4	14.1±2.0	15.5±1.6	17.7±8.0	14.5±2.5
(Fr’E)	(0.9±0.3)	(1.1±0.3)	(0.6±0.3)	(0.7±0.3)	(0.9±0.3)
Fr’E’/(D–F)	4.2±0.6	4.1±0.6	5.1±0.3§	4.1±0.9	3.4±1.1
(Fr’E’)	(0.2±0.1)	(0.2±0.1)	(0.2±0.1)	(0.2±0.0)*	(0.2±0.1)
Fr’F/(D–F)	16.3±4.7	18.6±2.4	20.9±6.4	2.1±1.5¶	14.4±1.8
(Fr’F)	(0.9±0.4)	(1.1±0.3)*	(0.8±0.3)	(0.1±0.1)¶	(0.8±0.2)

^*p<0.05

^§p<0.01

^¶p<0.001

Bone marrow B cells in various γδ T cell knockout mice. B cell types were identified based on their surface phenotype as in Figure 2. Following a scheme first described by Hardy et al. (92), we divided B220^posCD43^pos early stage bone marrow B cells into developmentally consecutive fractions A-C based on their expression of CD24 and BP1. These cells represent pro-B and early pre-B cells. Later stage B220^posCD43^neg B cells, which include those that returned from the periphery, were divided into increasingly mature fractions D (pre-B cells), E and F (B cells), based on their expression of IgM and IgD. Cells contained within these fractions were enumerated in individual mice. Absolute numbers (in parentheses) and relative frequencies are shown. n = 4–13 mice per group, p values, as indicated in the table.

Changed mature and transitional B cells in the spleen

Having found substantial changes among mature peripheral but not immature bone marrow B cells, we proceeded to examine the intermediate stages of B cell development in the spleen (Figure 3, Table 2). We divided B220^pos splenic B cells into mature CD93^neg and immature CD93^pos cells (Figure 3A), and further subdivided the mature B cells into follicular B cells (FOB, CD23^posCD21^int), marginal zone B cells (MZB, CD23^negCD21^hi or CD1d^hiCD21^hi) and “new” B cells (New, CD23^negCD21^neg) (Figure 3G, I, J, H) (62). We also identified B1 B cells in the spleen among B220^posIgM^pos cells based on their CD23^negCD43^pos phenotype (Figure 3D, E, F), and germinal center B cells (GCB) based on their distinctive CD38^negFas^posPNA^hi phenotype (50) and (Figure 3K, L, M). Comparing the same panel of mice as before for these splenic B cell populations, we found that mice lacking all γδ T cells (B6.TCR-δ^−/−) had nearly unaltered B cell populations. In contrast, mature and immature splenic B cell populations in B6.TCR-Vγ4^−/−6^−/− mice were diminished in numbers and relative frequencies (Figure 3A, B, C), and specifically B2 B cells. FO B cells were much diminished but MZ B cells were nearly wiped out (Figure 3G, H, I, J). In contrast, numbers of B1 B (Figure 3B, E, F) were relatively stable, and GC B cells (Figure 3K, L, M) in these mice were relatively increased. Again, removing IL-4 (B6.TCR-Vγ4^−/−/6^−/−/IL-4^−/− mice) reversed all of these changes (Table 2) suggesting that the elevated IL-4 levels in B6.TCR-Vγ4^−/−6^−/− mice are responsible for the changes in their splenic B cells. B6.TCR-Vγ1^−/− mice, on the other hand, showed little changes in numbers of mature and immature splenic B cells, although their MZ and “new” B cells were significantly increased. Apparent decreases in FO B cells in these mice seem to be merely a function of the lower CD23 expression, and a different gating strategy for FO B cells (based on CD21 and IgM expression) revealed normal FO B cell numbers (supplemental Figure 1). Nevertheless, the low CD23 expression in B6.TCR-Vγ1^−/− mice is a distinctive trait (supplemental Figure 4), likely to have functional consequences in the IgE responses. CD23 is positively regulated by IL-4 (14), and was diminished in B6.TCR-Vγ4^−/−/6^−/− mice by ablation of IL-4 (supplemental Figure 4). Furthermore, it was diminished in wt mice by treatment with anti Vγ1 mAbs (supplemental Figure 4), partially restored in cultured splenic B cells of B6.TCR-Vγ1^−/− mice by adding IL-4 in vitro (supplemental Figure 4), and much induced in B cells of B6.TCR-δ^−/− mice following transfer of Vγ1^pos cells from B6.TCR-Vγ4^−/−/6^−/− mice (supplemental Figure 4). Another distinctive trait of B6.TCR-Vγ1^−/− mice is their enlarged population of MZ B cells (Figure 3G,H,I,J, supplemental Figure 1), in direct contrast to the diminished MZ B cells in B6.TCR-Vγ4^−/−6^−/− mice. The mere absence of Vγ1^pos γδ T cells in B6.TCR-Vγ1^−/− mice does not account for these traits because Vγ1^pos cells are also missing in B6.TCR-δ^−/− mice, which have normal CD23 expression and numbers of MZ B cells. Instead, they again probably reflect a changed function of the γδ T cells that remain in B6.TCR-Vγ1^−/− mice mice. In sum, the data revealed that mature B cells in the spleen are sensitive to the influence of γδ T cells, and that much of this influence depends on IL-4.

Figure 3. Influence of γδ T cells on splenic B cell populations.

(A–M) Comparison of B cell populations in 8 wks old female C57BL/6 (B6) (black columns), B6.TCR-δ^−/− (δ^−/−) (open columns), B6.TCR-Vγ1^−/− (Vγ1^−/−) (dark grey columns) and B6.TCR-Vγ4^−/−/Vγ6^−/− (Vγ4^−/−/6^−/−) (light grey columns) mice. Mature and immature B cells, B1 and B2 B cells, marginal zone B cells (MZB), follicular B cells (FOB), “new” B cells (New B) and germinal center B cells (GCB) were identified using the indicated markers. Panels A, D, G, H, and K show representative staining profiles of individual mice. (A–C) Relative frequencies (compared to total B cells) and absolute numbers of mature (m) and immature (imm) B cells. Panels B and C also show numbers and frequencies (compared to total splenic lymphocytes) of total B cells. (D–F) Relative frequencies (compared to total B cells) and absolute numbers of B1 and B2 cells. (G–J) Relative frequencies (compared to total B cells) and absolute numbers of marginal zone B cells (MZB), follicular B cells (FOB) and “new” B cells (New B) B1. Panels G and H show two different ways of identifying MZB, based on expression of CD21 in combination with CD23 or CD1d. The counts of MZB in panels I and J are based on the method shown in panel H. (K–M) Relative frequencies (compared to total B cells) and absolute numbers of germinal center B cells (GCB). n = 5–8 mice per group. For visibility, only significant differences between wt and the γδ-deficient mice are marked. *P<0.05, **P<0.01, ***P<0.001

B cells newly arrived in the spleen can be divided into discrete transitional stages (T1-T3), distinguished by their surface phenotype and functional capability (17, 63). We identified immature B cells in the spleen of the test panel mice as IgM^posB220^posCD93^pos cells, and further divided these cells into T1 (IgM^hiCD21^neg/lo) and T2 plus T3 (IgM^posCD21^pos) transitional subsets (Figure 4A). Furthermore, we analyzed both subsets for their expression of CD23 and IgD. At 8 wks of age, B6.TCR-δ^−/− mice and B6.TCR-Vγ1^−/− mice had nearly unchanged numbers of transitional B cells (Figure 4B), and relative frequencies of the transitional subsets were normal as well (Figure 4C). In contrast, B6.TCR-Vγ4^−/−6^−/− mice produced significantly fewer T1 and T2 plus T3 B cells (Figure 4B,C). Removal of IL-4 (B6.TCR-Vγ4^−/−/6^−/−/IL-4^−/− mice) partially restored immature B cells, including T1 and T2 plus T3 transitional subsets (Table 2), suggesting that the diminished transitional B cell compartment in B6.TCR-Vγ4^−/−6^−/− mice is a indirect consequence of the deregulated IL-4 production in these animals (see below). CD23 and IgD expression revealed further differences between the transitional B cells of the test panel mice (Figure 4D). Hence, in contrast to immature bone marrow B cells, transitional B cells in the spleen were already altered, and thus might represent the earliest stage in B cell development affected by γδ T cells.

Figure 4. Genetic deficiency in γδ T cells alters transitional B cells in the spleen.

Comparison of transitional B cell populations in the spleens of 8 wks old female C57BL/6 (B6), B6.TCR-δ^−/− (δ^−/−), B6.TCR-Vγ1^−/− (Vγ1^−/−) and B6.TCR-Vγ4^−/−/Vγ6^−/− (Vγ4^−/−/6^−/−) mice. (A) Immature IgM^posB220^posCD93^pos B cells in the spleen were further subdivided into CD21^negIgM^hi (T1) and CD21^posIgM^pos (T2 plus T3) B cells. The test panel mice were also compared for CD23 and IgD expression in the two subsets of transitional B cells. Representative examples are shown. (B) Absolute numbers of transitional B cells/spleen in 8 wks old mice and (C) relative frequencies. (D) Expression of CD23 and IgD in transitional B cells of wt and γδ-deficient mice. Profiles representative of at least three independent staining experiments are shown. For panels b and c, n = 8–14 mice per group. Significant differences between wt and the γδ-deficient mice are marked. ***P<0.001

Increased numbers of antibody-producing B cells in the spleen of B6.TCR-Vγ4^−/−6^−/− mice

Untreated B6.TCR-Vγ4^−/−6^−/− mice have increased levels of circulating antibodies (50) but decreased numbers of mature B cells (this study). Both of these changes are IL-4-dependent. These mice also exhibit spontaneous germinal center formation in the spleen (50). Therefore, we compared numbers of antibody-producing B cells in this partially γδ-deficient strain and untreated wt mice (Figure 5). Total Ig producing cells in the spleen of 8–12 wks old B6.TCR-Vγ4^−/−6^−/− mice were increased > 2fold compared to wt mice, and IgG1-secreting cells > 7fold (Figure 5A). In contrast, there were no significant increased of antibody-producing B cells in bone marrow. IgG1 surface-positive B cells (B220^neg and B220^pos) were substantially increased but not in bone marrow (Figure 5B, D), and IgG1-secreting plasma cells (surface IgM^low, intracellular IgG1^pos) were increased in the spleen as well (Figure 5C, D). Numbers of such cells in bone marrow were low and difficult to quantitate. Taken together, these data document increased numbers of antibody-producing B cells in the spleen but not in the bone marrow of B6.TCR-Vγ4^−/−6^−/− mice, and suggest that the increased levels of circulating antibodies in this mutant strain are a result of this cellular change.

Figure 5. Antibody-producing cells (ASCs) are increased in the spleen of Vγ4^−/−/6^−/− mice.

(A) Total Ig-, IgG-, and IgG1-producing cells determined by ELISPOT assay were enumerated in spleen and bone marrow of female C57BL/6 (wt) and B6.TCR-Vγ4^−/−/Vγ6^−/− (Vγ4^−/−/6^−/−) mice. n = 4 per group. (B–C) Representative FACS plots showing surface IgG1-expressing cells (sIgG1^pos cells, panel B) and intracellular IgG1-expressing cells (iIgG1^pos cells, panel C) in spleen and bone marrow of above mice. Panel D shows relative frequencies and total numbers in spleen and bone marrow of mice indicated in (A–C). n = 4 mice per group. NS, not significant, *P<0.5, **P<0.01, ***P<0.001

γδ T cells themselves shape splenic B cell populations

To address the question of whether γδ T cells themselves modulate the B cell populations, we took advantage of our earlier observation that residual γδ T cells in B6.TCR-Vγ4^−/−6^−/− mice, which mostly belong to the Vγ1^pos subset, are changed (50). Such changes include higher relative frequencies and absolute numbers of IL-4-competent γδ T cells (50), altered TCR-Vδ expression among Vγ1^pos γδ T cells (50) as well as a higher frequency of CD8 expression among these cells when IL-4 is present (see Figure 6E), suggesting IL-4-driven Tc2-like differentiation. We conducted cell transfer experiments with these changed cells, in a manner as previously described (55). Adoptive transfer of B6.TCR-Vγ4^−/−6^−/− -derived splenic γδ T cells into B6.TCR-δ^−/− mice transiently restored splenic Vγ1^pos cells in the cell transfer recipients, albeit only up to about 10% of the population size in wt mice (supplemental Figure 2). Still, the transferred γδ T cells selectively reduced MZ B cells in the transfer recipients (Figure 6A,B), replicating the trend seen in non-manipulated B6.TCR-Vγ4^−/−6^−/− mice (Figure 3). The similar cell transfer experiment shown in supplemental Figure 2 further extends this finding: here the transferred γδ T cells, which again reduced splenic MZB cells, did so despite the presence of recipient γδ T cells (recipient: B6.TCR-Vγ1^−/−), which were unable to prevent this effect. When we examined transitional B cells in the B6.TCR-δ^−/− cell transfer recipients, we found a reduction in CD21 expression (Figure 6C), also replicating the situation in B6.TCR-Vγ4^−/−6^−/− mice (Figure 6D). In all experiments, transferred γδ T cells derived from B6.TCR-Vγ4^−/−/6^−/−/IL-4^−/− mice failed to induce these changes in B cells (Figure 6 and supplemental Figure 2), emphasizing the importance of IL-4 in the functional differentiation of the transferred γδ T cells. The dual effect of the transferred γδ T cells on transitional B cells, which must pass through the MZ (17), and on MZ B cells, which shuttle between follicles and marginal zone (64), is consistent with interactions between γδ T cells and B cells inside the splenic MZ.

Figure 6. Transferred residual γδ T cells alter splenic B cells in vivo and co-cultured residual γδ T cells selectively diminish MZ B cells in vitro.

(A) B6.TCR-δ^−/− mice (δ^−/−) were transferred with purified splenic γδ T cells from B6.TCR-Vγ4^−/−/Vγ6^−/− (Vγ4^−/−/6^−/−) or B6.TCR-Vγ4^−/−/Vγ6^−/−/IL-4^−/− (Vγ4^−/−/6^−/−/IL-4^−/−) mice. 10 days after the cell transfer, B cell populations in the spleen were compared as detailed for Figure 2, using the indicated markers. Data of one representative experiment are shown. (B) Effect of transferred residual γδ T cells on absolute numbers of MZB cells in the spleen of δ^−/− mice, n = 4 mice per group. (C) Effect of transferred residual γδ T cells on CD21 expression in the spleen of δ^−/− mice, n = 4 mice per group. (D) CD21 expression by transitional B cells in untreated wt, Vγ4^−/−/6^−/− and Vγ4^−/−/6^−/−/IL-4^−/−, n = 4 mice per group. *P<0.05, **P<0.01, ***P<0.001 (E) The CD8^pos fraction of Vγ1^pos γδ T cells in the spleen: Relative frequencies of CD8^pos γδ T cells within the splenic Vγ1^pos subset of C57BL/6 (wt), B6.TCR-Vγ4^−/−/6^−/− (Vγ4^−/−/6^−/−) and B6.TCR-Vγ4^−/−/Vγ6^−/−/IL-4^−/− (Vγ4^−/−/6^−/−/IL-4^−/−) mice. n = 5 mice per group, **P<0.01, ***P<0.001 (F) CD43-negative MZ B cell-rich splenic B cells from B6-TCR-Vγ1^−/− (Vγ1^−/−) mice were cultured for 60 hrs alone or in the presence of splenic Vγ1^pos γδ T cells from different mouse strains, and subsequently stained to identify MZ B cells. Only Vγ1^pos cells from B6.TCR-Vγ4^−/−/Vγ6^−/− (Vγ4^−/−/6^−/−) mice substantially diminished MZ B cells. (G) Same B cells as in panel a were cultured alone or in the presence of CD8^pos or CD8^neg fractions of Vγ1^pos cells from Vγ4^−/−/6^−/− mice, and subsequently stained to identify MZ B cells. Only Vγ1^pos cells expressing CD8 diminished MZ B cells. Data panels F and G are representative of several similar experiments.

In addition, we tested for a possible effect of purified Vγ1^pos γδ T cells on MZ B cells in vitro. After a culture period of 60 hrs without any added stimuli or growth factors, enriched CD43^neg splenic B cells from B6.TCR-Vγ1^−/− mice contained ~ 12% MZ B cells (Figure 6F). Co-culturing them with Vγ1^pos γδ T cells from B6.TCR-Vγ4^−/−/6^−/− mice selectively diminished MZ B cells, similarly to the cell transfer experiments in vivo and consistent with the trend in B6.TCR-Vγ4^−/−/6^−/− mice, whereas γδ T cells from wt mice (either Vγ1^pos or Vγ4^pos), or from B6.TCR-Vγ4^−/−/6^−/−/IL-4^−/− mice, failed to mediate this effect. Furthermore, to address the potential significance of the enlarged IL-4-dependent CD8^pos subpopulation among Vγ1^pos cells in B6.TCR-Vγ4^−/−/6^−/− mice, we compared purified CD8^pos and CD8^neg sub-fractions of Vγ1^pos cells in B6.TCR-Vγ4^−/−/6^−/− mice for their effect on MZ B cells in the co-cultures (Figure 6G), and found that only the CD8^pos Vγ1^pos γδ T cells diminished MZ B cells. The effect in vitro - albeit consistent in its selectivity for MZ B cells with the effect in vivo – was comparatively small, which might reflect the importance of co-localization of γδ T cells and MZ B cells in the MZ in vivo. In sum, the combined results of the cell transfer and co-culture experiments suggest that the altered B6.TCR-Vγ4^−/−/6^−/− γδ T cells themselves are responsible for the changes of peripheral B cells in this strain.

γδ T-B cell interactions in the spleen may be facilitated by co-localization

The compartments of the spleen differ in their accessibility to the circulation (65). Recently arrived immature B cells and MZ B cells migrate through or reside within the MZ, a splenic compartment far more accessible to the circulation than the follicles or the peri-arteriolar sheath (PALS) (59, 66). Because of technical difficulties in localizing splenic γδ T cells in wt mice by immunohistochemical methods (67), we instead assessed the exposure of splenic γδ T cells to the circulation. We i.v. injected antibodies specific for the ubiquitous leucocyte marker CD45, and compared labeling levels of the splenic lymphocyte-types at a fixed time point after the injection (Figure 7A). The broad range of labeling intensity in all cell populations examined likely reflects positional differences of individual cells within a given population. However, overall, the percentage of labeled γδ T cells in normal C57BL/6 mice was higher than that of αβ T cells, indicating that splenic γδ T cells tend to be more exposed to the circulation than αβ T cells. The percentage of labeled NK1.1^pos γδ T cells was higher still, and similar to that of NKT cells. This finding in mice is consistent with histological studies in several other species placing γδ T cells in the red pulp and MZ of the spleen (68–72). Among immature B cells, T1 B cells were more exposed than T2 plus T3 B cells, as would be expected (17), and MZ B cells were most exposed among mature B cells, consistent with the literature (59). Although we found some differences in CD45 expression of the splenic cell types examined (based on in vitro antibody staining, see supplemental Figure 3B), the differential labeling in vivo did not correlate with these differences, rather reflecting differential circulation exposure than differential CD45 expression. Thus, circulation exposure assessed here and histological findings of others place γδ T cells, early transitional (T1) B cells and MZ B together inside the MZ. Secondly, we examined γδ T-B cell conjugates in fresh splenocyte preparations. CD93^pos immature B cells (Figure 7B) and MZ B cells (Figure 7C) were enriched in the conjugates, consistent with the notion of encounters and contact between splenic γδ T cells and B cells that pass through (T1 B cells) or reside within the MZ (MZ B cells). In turn, the γδ T cells making these contacts with B cells seemed to be “aware” of them as they expressed the activation markers CD40L and ICOS at higher levels than non-conjugated γδ T cells (Figure 7D).

Figure 7. Localization of γδ T cells and immature B cells in the spleen, propensity to form T-B-conjugates and activation state of γδ T cells in the conjugates.

(A) 8 wks old female C57BL/6 mice were injected i.v. with dye-conjugated antibodies specific for CD45, splenocytes were stained in vitro for subset-specific markers after a 20 min in vivo labeling period, and analyzed cytofluorimetrically. One example representative of six similar experiments is shown. (B) γδ T-B cell conjugates among splenocytes from 8 wks old female C57BL/6 mice were identified based on their simultaneous expression of γδ T cell (TCR-δ) and B cell markers (CD19), and B cells in the conjugates compared with non-conjugated B cells for their expression of CD93, a marker of immature B cells or (C) for their expression of CD23 and CD21 in order to differentiate “new”, MZ and FO B cells. Individual examples representative of four similar experiments are shown. (D) γδ T cells in conjugates with splenic B cells and non-conjugated γδ T cells were compared for their expression of CD40L and ICOS. One experiment representative of four using 8–12 wks old female C57BL/6 mice is shown.

Discussion

Overall, the current study shows that γδ T cells are capable of modulating pre-immune peripheral B cells populations. This work was inspired by our preceding study indicating that γδ T cells strongly affect Ig serum levels and autoantibody development in non-immunized mice (50). Both studies take advantage of three connected observations, namely first that a correlation exists between TCR-Vγ expression by subsets of murine γδ T cells and their function (73); second, that such subsets in isolation tend to have a larger effect on the immune responses than γδ T cells as a whole (74); and third, that the absence of a subset can lead to functional changes in the remaining γδ T cells (50). These features enable the manipulation of γδ T cell function in vivo, using TCR-Vγ knockout as approach. Having employed this approach to reveal the γδ-influence on antibody production and self-tolerance (50, 55), we now demonstrate effects on B cell homeostasis and get a first glimpse at underlying mechanisms.

In particular, we found in the current study that mice deficient in two γδ T cell subsets (B6.TCR-Vγ4^−/−/6^−/−), which have normal numbers of immature bone marrow B cells, nevertheless have much reduced numbers of total peripheral B cells. This occurs in the presence of elevated levels of IL-4 (50) and BAFF (data not shown), both of which would be expected to support B cell growth (8, 60). Indeed, antibody producing B cells are increased, consistent with the increased levels of serum Ig (50). The loss of peripheral B cells seems to take place during development in the spleen because levels of immature B cells in the spleen and derived mature B cell populations were all affected whereas immature B cells in bone marrow were not. Among mature splenic B cell populations, those that reside in or repeatedly shuttle in and out of the MZ - the MZ B cells (64) – were diminished most. Immature B cells, which also have to pass through the MZ on their way to the white pulp (75, 76), were hit as well. Combined, these observations implicate the splenic MZ as a critical site of the γδ-influence on peripheral B cells.

The splenic MZ is a portal for cells in transit from the blood stream to the white pulp (65). It also contains various types of resident cells that depend on each other for their localization and function (77). Besides reticular fibroblasts, these include marginal sinus lining cells as well as several distinct myeloid and lymphoid cell-types. Splenic γδ T cells have been localized in the MZ as well, and within the red pulp, in humans, cattle, sheep, camels and birds (68–72). We previously reported that splenic γδ T cells in mice acquire blood-borne antigen (78), consistent with a similar localization of these γδ T cells. The data of the current study further support this notion: (i) splenic γδ T cells in normal mice were stained well by i.v. injected antibodies, indicating an exposure to the circulation similar to that of early transitional B cells and MZ B cells, (ii) splenic γδ T cells in αβ T cell-deficient mice, where γδ T cells move into the circulation-inaccessible PALS (67), were no longer well-stained by i.v. injected antibodies (data not shown), (iii) immature B cells and MZ B cells were enriched in splenic γδ T – B cell conjugates, and (iv) γδ T cells affected MZ B cells far more than the other mature B cell types. Due to their precursor progeny-relationship, the still substantial but smaller effect of the γδ T cells on other mature B cells likely is a consequence of their interaction with transitional B cells, which must pass through the MZ as well (17).

The cell transfer and co-culture data, and the data co-localizing B cells and γδ T cells in the MZ all suggest that action by splenic γδ T cells themselves, either direct or indirect, is responsible for the diminished peripheral B cells in B6.TCR-Vγ4^−/−/6^−/− mice. Splenic γδ T cells in these mice are mainly Vγ1^pos, and they are also altered in composition and function (50), partly under the influence of IL-4, which drives expansions of both NKT-like (50) and CD8^posTc2-like (79–81) Vγ1^pos cells (this study). Indeed, adoptively transferred γδ T cells from the spleen of B6.TCR-Vγ4^−/−/6^−/− mice selectively diminished MZ B cells in the recipient mice. The transferred γδ T cells also lowered CD21 expression in transitional B cells, consistent with the changed phenotype of these B cells in B6.TCR-Vγ4^−/−/6^−/− mice. Both of these results support the notion of γδ T- B cell interactions in the MZ, but do not rule out such interactions elsewhere. The transfer experiments also underscore the role of IL-4 because γδ T cells obtained from B6.TCR-Vγ4^−/−/6^−/−/IL-4^−/− mice failed to affect the B cells. Normal levels of IL-4 seem to be sufficient for some action of Vγ1^pos γδ T cells in the spleen because wt C57BL/6 mice had significantly fewer MZ B cells than did B6.TCR-Vγ1^−/− mice. However, in B6.TCR-Vγ4^−/−/6^−/− mice with their hyperplastic Vγ1^pos γδ T cell population and increased IL-4 production, the effect is exacerbated, leading to substantial reductions in mature splenic B cells and a near disappearance of MZ B cells. Even co-cultured B6.TCR-Vγ4^−/−/6^−/− -derived Vγ1^pos γδ T cells in vitro reproducibly diminished MZ B cells more than other B cells, although the effect was comparatively small. Both direct and indirect mechanisms might contribute to these γδ-dependent changes in peripheral B cells and serum Ig levels. The elevated IL-4 in B6.TCR-Vγ4^−/−/6^−/− mice potentially could drive premature switching of immature bone marrow B cells (82) but we did not see substantial changes here. The loss of mature splenic B cells in B6.TCR-Vγ4^−/−/6^−/− mice could be the result of accelerated maturation and differentiation into plasma cells, consistent with the increased numbers of antibody producing cells in the spleen of these mice (this study) as well as the elevated levels of circulating antibodies (50). This mechanism seems to be IL-4-dependent because the changes in peripheral B cells are absent when IL-4 is missing. The preferential loss of MZ B cells likely is due to the greater propensity of these cells to form plasma cells (83). In addition, their co-localization with γδ T cells in the spleen might make these B cells more available for the γδ influence. As well, MZ B cells might be more likely to interact with γδ T cells due to their differential expression of ligands for the γδ TCR such as the molecules CD1d and T-22 (84–86). Finally, we cannot exclude a role of changed microbiota (87) in the γδ-deficient mice but this seems less probable under conditions of transient reconstitution.

Although the massive loss of peripheral B cells in B6.TCR-Vγ4^−/−/6^−/− mice was the most noticeable γδ-effect, several phenotypic changes in B cells can be ascribed to the influence of γδ T cells as well, such as the much diminished expression of the IgE receptor CD23 in B6.TCR-Vγ1^−/− mice, the diminished expression of the inhibitory receptor CD5 and FcγRIIB in B6.TCR-Vγ4^−/−/6^−/− mice, or the increased expression of MHCII and IL-4Rα in B6.TCR-Vγ4^−/−/6^−/− mice. All of the listed changes appear to be mediated or at least indirectly connected to γδ-dependent IL-4 (11, 16, 50) and, in the case of B6.TCR-Vγ4^−/−/6^−/− mice, might be exacerbated by the loss of peripheral B cells, which likely contributes to the increase in available serum IL-4 (50) and BAFF (not shown) in these mice.

The findings of this study raise several new questions. For example, γδ T cells in the MZ might be a differentiated and functionally specialized population similarly to the other specialized residents of this site (65), and they might function as part of a cellular network in the MZ. As already mentioned, it appears that MZ γδ T cells also participate in the monitoring of blood-borne antigens (78). Furthermore, our data suggest that the changes in the peripheral B cells of B6.TCR-Vγ4^−/−/6^−/− mice are directly connected to the elevated serum Ig and the development of autoantibodies in these mice (50). By extension, the observations described here predict that changes in γδ T cell populations due to natural causes (26, 33, 88) might similarly affect how peripheral B cells differentiate, how much Ig is produced, and whether or not autoantibodies develop. Consistently, impaired TCR signaling, which causes changes in γδ T cell populations has already been associated with increased IgE production (33, 34, 89). Likewise, changes in size and composition of γδ T cell populations, which have been found in hematopoietic transplantation (90) and in HIV infection (91), might affect T and B cell functions, B cell reconstitution and humoral immune competence.

Abbreviations used in this article

CLP

common lymphocyte precursor

BCR

B cell antigen receptor

TCR

T cell antigen receptor

FOB

follicular B cells

MZB

marginal zone B cells

SLE

systemic lupus erythematosus

IL-4

interleukin 4

BAFF

B cell activating factor