Wiskott-Aldrich Syndrome protein is required for homeostasis and function of invariant NKT cells

Abstract

NKT cells comprise a separate T lineage expressing semi-invariant T cell receptors. Canonical iNKT cells specifically recognize lipid antigens presented by CD1d, a MHC Class I-like molecule. iNKT cells function, in part, as initial responders to bacterial infection and play a role in immune surveillance and tumor rejection. The Wiskott-Aldrich Syndrome protein (WASp) serves as a crucial link between cellular stimuli and cytoskeletal rearrangements. While we and others have identified a key role for WASp in homeostasis of T-regulatory and marginal zone B cells, little data exist regarding the role for WASp within the iNKT lineage. Analysis of WASp-expressing cell populations in heterozygous female WASp mice revealed a substantial selective advantage for WASp⁺ vs. WASp⁻ iNKT cells. While adult WASp-deficient (WASp^−/−) mice had normal thymic and BM iNKT numbers, we observed 2–3 fold reduction in the numbers of iNKT cells in the spleen and liver. This peripheral iNKT deficit is manifested, in part, due to defective iNKT homeostasis. WASp^−/− iNKT cells exhibited reduced levels of integrin surface expression and decreased homing and/or retention within peripheral tissues in a competitive repopulation model. In addition, analysis of young mice showed that WASp is important for both maturation and egress of thymic iNKT cells. WASp^−/− iNKT cells also exhibited a marked reduction in antigen-induced proliferation and cytokine production. Our findings highlight the crucial role for WASp in iNKT development, homeostasis and activation, and identify iNKT dysfunction as an additional factor likely to contribute to the clinical features observed in WAS patients.

Introduction

Cross-talk between the adaptive and the innate immune systems is critical for mounting an effective immune response to most infectious challenges. Within the adaptive system, B and T lymphocytes uniquely express a rearranged antigen receptor required for generation of long-term immune memory. Distinct subsets of both B and T cells, however, also exhibit characteristics typically associated with innate cells. Within the B cell lineage, marginal zone (MZ) and B-1 B cells recognize antigen through semi-invariant receptors and rapidly differentiate into antibody secreting cells following pathogen recognition (1). Similarly, NKT cells comprise a unique T cell subset that express an invariant T cell receptor and can rapidly secrete effector cytokines following receptor engagement (2). Both MZ B cells and NKT cells can be positively selected by self-antigen and rely on integrin signals through CD11a for tissue retention (2, 3). We and others have recently shown that the Wiskott-Aldrich Syndrome protein (WASp) is crucial for homeostasis of MZ and B-1 B cells (4–6). This multi-domain adapter protein is essential for linking surface receptor signals to cytoskeletal rearrangement and for efficient integrin signaling (7–9). Despite the functional similarities between NKT and MZ B cells, the potential role for WASp in NKT-associated signals has not been previously examined.

NKT cells are a heterogeneous, thymic-derived population that expresses both NK-cell markers in association with an αβ T cell receptor (2). Classical type I invariant NKT (iNKT) cells specifically utilize the Vα14 chain (in mice, or Vα24 in humans) (10). This population can be activated via glycolipid peptides presented by the MHC class-I type molecule, CD1d and can be identified using CD1d tetramers loaded with the model antigen, α-galactosylceramide (αGalCer) (11). iNKT cells function as immune sentinels, mounting rapid effector responses to infections, participating in immune tolerance, and detecting and eliminating altered host cells (2, 12).

Unlike conventional T-helper cells, activated iNKT cells rapidly produce both interleukin 4 (IL-4) and interferon-γ (IFN-γ), thereby enhancing the adaptive immune responses (13). Absence of iNKT cells is associated with increased rates of infection and autoimmune disease (2). Notably, frequent infection and a high risk of autoimmunity are striking features associated with the Wiskott-Aldrich Syndrome, a rare X-linked immunodeficiency caused by mutations of the WASP gene (14).

WASp deficient (WASp^−/−) mice exhibit partial defects in T lineage development and marked T cell functional deficits thereby providing a useful model for human WAS (15, 16). Contrary to activation defects observed in conventional T cells, WASp-deficient B cells exhibit normal proliferative responses following antigen-receptor stimulation (15, 16). However, WASp-deficient MZ B cells exhibit altered peripheral homeostasis, partly due to defective integrin signaling (4–6). Similarly, WASp-deficient T-regs exhibit defective homing to effector tissues and fail to rescue autoimmune development in the Scurfy mice, which lack regulatory T cells (17–19). In addition to these T and B cell defects, WAS patients exhibit reduced Natural Killer (NK) cell cytotoxicity and impaired migration and antigen presentation by dendritic cells (20–22).

The iNKT lineage shares functional similarities with regulatory T cells and marginal zone B cells (23, 24), suggesting a potential role for WASp in iNKT development and/or function. In the current study, we show that WASp⁻ iNKT cells fail to compete effectively with WASp⁺ iNKT cells in vivo. Further, WASp^−/− mice exhibit defective development and egress of thymic iNKT cells and express lower levels of CD11a integrin. Combined, these defects result in significantly reduced numbers of iNKT cells within the spleen and the liver of adult mice. WASp^−/− iNKT cells also display a marked deficit in antigen-induced proliferation and cytokine secretion. Finally, using adoptive transfer studies we show that WASp^−/− iNKT cells exhibit normal homeostatic proliferation, yet exhibit defects in homing and/or retention within effector tissues. Taken together, these data demonstrate that WASp plays a key role in both iNKT homeostasis and function; and suggest that impaired iNKT function may contribute to the global immune dysregulation observed in WAS patients.

Materials and Methods

Mice and tissues

Breeding colonies for WT (C57/Bl6) and C57/Bl6-WASp^−/− (18), and Rag2^−/− mice were maintained within the SPF animal facility of Seattle Children's Research Institute as previously described (4). Animals were analyzed at various ages as noted in results; and all studies were carried out according to the guidelines of Seattle Children's Institutional Animal Care and Use Committee.

Single cell thymus and spleen suspensions were obtained by dissociating tissues with frosted glass slides. For bone marrow isolation, femur and tibia from a single leg were flushed out using a 25G needle and broken up into a single cell suspension with a 22G needle. Erythrocytes were lysed with ammonium chloride potassium phosphate (ACK) buffer. Livers were perfused with 10ml of 1x PBS prior to extraction, dissected into small pieces, and incubated for 40' in 1x HBSS containing 400U/ml collagenase. Cells were washed 2X with staining buffer and resuspended in 2–3ml of 40% Percoll. Cell suspension was underlayed with 2ml of 70% Percoll and spun at 2500 RPM for 20' with minimal acceleration and no brake and interface cells were collected and resuspended in staining buffer.

Flow Cytometry

Intracellular staining for WASp was performed as previously described (4), using a peptide purified rabbit anti-WASp polyclonal antibody (A6503)(25). For surface staining, single cell suspensions were incubated with fluorescently-labeled antibodies for 15 min at 4 °C in staining buffer (PBS w/ 0.5% BSA). Data was collected on a LSR II flow cytometer (BD Biosciences) and analyzed using FlowJo software (Treestar Inc). The following antibodies were used for staining: CD45.1-PeCy7 (clone A20), CD45.2-APC (104), CD3- PeCy7 (145-2C11), cKit-PE (ACK2), Sca1-APC (D7), IL-4-PeCy7 (BVD6-24G2), IFN-γ-APC(XMG1.2) and Streptavidin-PeCy7 from Ebioscience; CD24-APC (M1/69), CD4-PeCy5 (RM-5), from BD Biosciences; CD44-ApcCy7 (IM7) and NK1.1-FITC or PE (PK136) from Biolegend; CD11a-FITC (I21/7) from Southern Biotech; Lin cocktail-Biotin from Miltenyi Biotec, and Alexa568 Goat anti-rabbit 2^nd antibody from Invitrogen. Loaded, PE-labeled CD1d tetramers were generously provided by Mitchell Kronenberg and the NIH tetramer facility. CD1d tetramer staining was performed as previously described (26). For comparison of CD69 MFI values across different experiments, we normalized the CD69 expression to WT levels. For this, the relative CD69 expression was derived through a ratio of CD69 MFI from each sample divided by the CD69 MFI value derived from WT splenic samples.

CFSE proliferation assay

Splenocytes were washed 2x in PBS and resuspended at 5×10⁶ cells/ml with 250nM CFSE (Invitrogen). Cells were incubated for 8–9 minutes at 37 °C, shaking periodically, and washed with complete RPMI media. Cells were washed 3x with complete media and incubated at 1×10⁶ cells/ml in complete media supplemented with αGalCer (1–10ng/ml), IL-15 (100ng/ml), or PMA(10ng/ml)/Ionomycin(500ng/ml). Cells were stained and analyzed at 72 or 96 hours post-stimulation

Cell transfer experiments

Total bone marrow suspensions from WT (CD45.2⁺) and WASp^−/− (CD45.1⁺) mice were mixed at a ratio of 1:9 and transplanted into irradiated (1050 rads) WASp^−/− recipients at a dose of 5×10⁶ cells/recipient. For iNKT transfer, thymocytes were labed with CD8-PE and CD24-PE (BD Biosciences) in MACS buffer and depleted using αPE magnetic beads and LD MACS columns (Miltenyi Biotec). The depletion resulted in ~100-fold enrichment of iNKT cells. The iNKT enriched thymocytes from WT or WASp^−/− mice were mixed at a 1:1 ratio, labeled with CFSE as described above and retro-orbitally injected into unmanipulated Rag2^−/− mice at a dose of 3×10⁶ cells/mouse.

Intracellular cytokine staining

Splenocytes were resuspended at 4×10⁶ cells/ml in 200μl complete RPMI media and stimulated with αGalCer or PMA/Ionomycin. Brefeldin A (final: 1μg/ml) was added 45 minutes after start of incubation. After 5–6 hours, cells were spun down, stained with CD1d tetramer, fixed and permeabilized using cytofix/cytoperm (BD Biosciences) and stained with antibodies against CD4, IL-4 and IFNγ.

Results

WASp⁺ iNKT cells exhibit a selective advantage in vivo

To assess the role for WASp in iNKT development and homeostasis, we initially evaluated selection in WASp^+/− heterozygote females using a flow-cytometry based assay (4). This approach, based upon assessment of inactivation of the X-chromosome that contains the Was gene, permits identification of developmental stage(s) where WASp⁺ cells manifest a selective advantage. As shown in Figure 1A, hematopoietic stem cells (HSC; Lin⁻cKit⁺Sca1⁺) exhibit a random X-inactivation pattern with ~50% of cells expressing WASp. This pattern of WASp expression remained unaltered in neutrophil, monocyte, and dendritic cells derived from either the bone marrow or spleen (data not shown), indicating that WASp⁺ myeloid cells lack a selective advantage despite the documented role for WASp in myeloid migration (14).

Figure 1. WASp⁺ iNKT cells exhibit a marked selective advantage.

(A) Representative FACS plots showing gating strategies. (B) Combined data showing selection in NK and NKT, but not HSC compartments. Boxes underneath each graph denote statistical significance compared to HSC while bars at top of the graph represent statistical difference between the liver vs. other tissues. Data for each subset includes analysis of multiple (≥6) heterozygous carriers. (C) CD1d tetramers were used to identify iNKT cells in WASp-heterozygous carriers. Left panel shows selection in the thymus, comparing iNKT subsets vs. conventional thymic T cell subsets. Right panel shows selection of CD24⁻CD1d tetr⁺ iNKT cells in the peripheral tissues. Boxes underneath show statistical differences relative to HSC and bars within panel denote statistical differences in thymus vs. peripheral tissues. (D) Representative FACS analysis of one BM chimera. (E) Combined data showing WASp⁺ HSC engraftment and subsequent selection of WASp⁺ NK and NKT subsets. Stars represent statistical differences compared to HSC. Data are representative of 3 independent experiments with 2–4 mice/experiment. * p<0.05, ** p<0.01, *** p<0.001.

We used NK1.1 and CD3 to analyze selective advantage within the NK and the NKT subsets (Fig. 1A). We observed a selective advantage for WASp⁺ NK cells compared to the HSC compartment (Fig. 1A&B), while NK1.1⁺CD3⁺ NKT cells exhibited an even greater selective advantage, especially within the spleen and liver (Fig. 1B). Because NK1.1 is expressed by both classical and non-classical NKT cells, we next used CD1d tetramers loaded with αGalCer to specifically identify invariant NKT cells (11). Within the thymus, tetramer⁺ iNKT cells progress from an immature CD24⁺ to a mature CD24⁻CD44⁺ stage of development (27). Similar to HSCs, ~50–55% of thymic CD24⁺ iNKT expressed WASp⁺. In contrast, a significantly higher proportion of CD24⁻ thymic iNKT cells (~70%) were WASp⁺, suggesting an initial role for WASp during maturation of thymic iNKT cells (Fig. 1C, left). Compared to the thymus, we observed an even more significant selection of CD24⁻ WASp⁺ iNKT cells within the spleen and liver (with an average of 83% WASp⁺ cells in spleen, and 86% in liver). In contrast, we did not observe a difference in selection between the thymus and the bone marrow (Fig.1C, right). These findings suggested that WASp might be involved in late thymic development as well as access and/or retention of iNKT cells in peripheral tissues.

While female heterozygous carriers represent a useful model to study selection of WASp⁺ cells, random X-inactivation might be skewed by lineage extrinsic abnormalities and/or other developmental defects. Therefore, we also used a mixed bone marrow chimera model to evaluate the selection for WASp⁺ cells. We used a relatively low WT:WASp^−/− cell ratio (1:9) to permit detection of any selective advantage. BM from WT and WASp^−/− was transplanted into lethally irradiated WASp^−/− animals and recipient mice were analyzed at 5 months post-transplant. Consistent with the analysis of heterozygous females, we observed no selective advantage for WASp expressing HSC, myeloid or dendritic cells. WASp expression was maintained at the input level of ~5–10% in each of these populations (Fig. 1D and data not shown). In contrast, analysis of NK subsets revealed a progressive increase in the percentage of WASp⁺ cells in the spleen and liver compared to the bone marrow. Even stronger skewing was again observed within the NKT lineage with an increased proportion of WASp⁺CD3⁺NK1.1⁺ NKT cells in all tissues (Fig. 1D&E). Nearly identical data was obtained using congenically marked BM (comprised of a 1:9 mixture of CD45.1⁺ WT and CD45.2 WASp^−/−cells) transplanted into WASp^−/− (CD45.2⁺) recipients (data not shown). Combined, these observations further supported the interpretation that WASp-dependent function(s) impact both late thymic iNKT cell development and peripheral iNKT cells homeostasis.

Reduced peripheral iNKT numbers in WASp^−/− mice

We next analyzed WASp^−/− mice to directly assess the consequences of WASp deficiency on iNKT development and function. We analyzed tissues from 8–14 week old WT and WASp^−/− mice using co-staining with CD1d tetramers, CD3, CD24, NK1.1 and CD44 (Fig. 2A). The absolute number of immature CD24⁺ iNKT cells within the thymus did not differ significantly between WT and WASp^−/− mice (Fig. 2B). While there was a trend for an increase in CD24⁺ iNKT cell numbers in WASp^−/− mice, this difference did not reach statistical significance in adult animals. In addition, we observed normal numbers of mature CD24⁻CD44⁺ iNKT cells within both the thymus and bone marrow, indicating that early iNKT development in adult animals was not significantly perturbed by WASp-deficiency in the absence of competing WT cells (Fig. 2A&C). Consistent with these findings, we also observed similar levels of CD1d expression on DP thymocytes from WT and WASp^−/− mice (data not shown).

Figure 2. WASp^−/− mice generate normal thymic iNKT cell numbers but exhibit a deficit in peripheral iNKT cells.

(A) Representative FACS plot showing the gating strategy used to identify iNKT cell subsets in various tissues in WT vs. WASp^−/− mice. (B) WASp^−/− mice exhibit normal numbers of immature thymic iNKT cells. (C) Reduced mature iNKT numbers in WASp^−/− spleen and the liver. Error bars represent data from 3 independent experiments using 6, 8–12 wk old, animals/experiment. (D) Representative FACS data showing normal expression of CD44 and NK1.1 on splenic and thymic iNKT cells. At least ten animals were examined per genotype. * p<0.05, ** p<0.01, *** p<0.001.

Despite normal numbers of thymic iNKT cells in adult WASp^−/− animals, analysis of peripheral tissues revealed a significant decrease in total iNKT cells in the spleen and liver (Fig. 2A&C). iNKT cells were reduced by at least 2-fold in the spleen and >3-fold in the liver of WASp^−/− animals. The mature iNKT compartment can be further divided into CD4⁺ and CD4⁻ subsets, which diverge at an early developmental stage and exhibit unique functional activities (28). Both CD4⁺ and CD4⁻ iNKT subsets were decreased in the spleen and liver of WASp^−/− animals, with CD4⁺ iNKT cells showing the greatest numerical decline (>3-fold reduction relative to WT mice, Fig. 2C; bottom panels). The observed reduction in peripheral iNKT numbers did not appear to reflect a maturation defect, as expression of developmental markers including both NK1.1 and CD44 were not altered in WASp^−/− iNKT cells (Fig 2D).

Defective generation of thymic iNKT cells in young WASp^−/− mice

The initial burst of iNKT expansion occurs at ~2–3 weeks post birth, starting in the thymus and culminating in the periphery (29). Immature CD24⁺ iNKT cells develop into mature CD24⁻CD44⁺ cells that subsequently progress from an NK1.1⁻ to an NK1.1⁺ developmental stage within both the thymus and periphery (29). Interestingly, analysis of 15–16 day old mice using antibodies directed against CD44, CD3, CD24, CD4 and NK1.1 revealed significantly reduced numbers of mature CD24⁻ iNKT cells within the thymus of young WASp^−/− animals (Fig. 3A). This change impacted all subsequent stages of mature iNKT development as identified by CD44/NK1.1/CD4 staining (data not shown). In contrast, thymic CD24⁺ immature iNKT cell numbers were significantly increased in WASp^−/− mice suggesting that WASp activity was important for the proliferative expansion of iNKT cells at the CD24⁺ to CD24⁻ transition (Fig. 3A, left). Similar to 6–12 week old animals, iNKT cell numbers were also reduced within the spleen of young animals (Figure 3A). While total iNKT numbers in the liver were also reduced, these data did not reach statistical significance likely reflecting the much smaller number of iNKT cells present in liver at this age and limited numbers of animals analyzed. Similar to older WASp^−/− animals, CD4⁺ iNKT cell numbers were significantly reduced within the liver (data not shown). These findings, in association with our observations in heterozygous female carriers, suggest a role for WASp in the proliferative expansion of immature thymic iNKT cells.

Figure 3. iNKT compartment composition in very young vs. older WASp^−/− mice.

(A) Analysis of iNKT compartment in 15–16 day old WT and WASp^−/− mice. Cells were stained as in Fig. 2. Data represent summary of at least 6 mice per genotype. (B) Analysis of iNKT compartment in 6 month old WT and WASp^−/− mice showing data from 6 mice per genotype. (C) Analysis of CD69 expression in cells obtained from 6-month old WT and WASp^−/− mice. The MFI was normalized as described in materials and methods. Error bars represent standard deviation. * p<0.05, ** p<0.01, *** p<0.001.

Partial correction of peripheral iNKT composition in aged WASp^−/− mice

We next analyzed tissues from older animals to determine whether peripheral iNKT numbers in WASp^−/− mice might normalize over time. Six month old mice had equivalent numbers of iNKT cells within the spleen and the bone marrow, yet exhibited reduced iNKT numbers within the thymus (Fig. 3B). The reduction in the thymic iNKT compartment is consistent with an overall reduction in thymic cellularity observed in aged WASp^−/− animals (data not shown). Similar to 8–12 wk old mice, we observed reduced numbers of iNKT cells within the liver; however this difference did not reach statistical significance perhaps due to small sample size. Similar results were also obtained in one year old mice (data not shown). Combined with findings from 2 week old mice, our observations suggest that a reduced rate of iNKT production in WASp^−/− mice results in lower numbers of splenic iNKT cells that are partially rescued over time.

Additionally, we determined whether WT vs. WASp^−/− mice exhibited any differences in cell surface phenotype within the mature iNKT cell compartment. Mature iNKT cells express a variety of memory-associated cell surface markers, including CD69 and CD122 (IL-2Rβ) (30). While WT and WASp^−/− splenic and liver iNKT cells expressed equivalent levels of CD122 (data not shown), we observed a significant upregulation of CD69 expression on WASp^−/− iNKT cells derived from bone marrow, spleen and liver of 6-month old mice (Fig. 3C). Higher levels of CD69 expression were also observed in liver iNKT cells from 15-day old mice (data not shown). This increase in CD69 expression was cell intrinsic, as it was also observed in WASp^neg iNKT cells in heterozygous WASp carriers (data not shown). Because WASp-deficient iNKT cells expressed WT-like levels of NK1.1, CD44 and CD122, the overall iNKT maturation status appeared to be relatively unperturbed in WASp^−/− mice. Rather, we anticipate that increased CD69 expression in WASp^−/− cells reflect the outcome of reduced signaling through the sphingosine-1-phosphate (S1P) receptor. Binding of S1P to S1P-receptors regulates iNKT thymic egress (31) and CD69 serves as a negative regulator of S1P mediated signaling (32). Higher CD69 expression levels may therefore correlate with reduced S1P responsiveness and inefficient thymic egress of WASp^−/− iNKT cells, as previously shown for WASp-deficient marginal zone B cells (4, 5).

We also considered the possibility that the reduced number of iNKT cells in spleen and peripheral tissues might be due to sequestration of iNKT cells within the lymph nodes (LN). As previously reported (14), WASp-deficiency is associated with moderate lymphadenopathy, and this correlated with a modest increase in the number of iNKT cells within WASp^−/− LN. However, LN resident iNKT cells (~5–10×10³) represent only a minor fraction of total iNKT cells within the liver and spleen compartments (2×10⁵ and 5×10⁵, respectively; Fig. 3B) in WASp^−/− mice, suggesting that LN sequestration is unlikely to be a major factor in the observed alterations in iNKT cell numbers and distribution.

Reduced antigen-induced proliferation in WASp^−/− iNKT cells

Mature iNKT cells proliferate and secrete cytokines following antigen receptor engagement. We used the model galactolipid, αGalCer, to determine the role of WASp in the functional responses of iNKT cells. Total splenocytes derived from either WT or WASp-deficient mice were labeled with CFSE, stimulated with αGalCer, and analyzed for both relative percentage of tetramer⁺ cells and CFSE dilution to assess iNKT cell proliferation. We observed a clear expansion of tetramer⁺ cells in αGalCer treated cultures containing WT iNKT cells. This population increased ~6-fold compared to untreated controls with the vast majority of cells undergoing one or more cell divisions (Fig. 4A). In contrast, there was no increase in the proportion of tetramer⁺ cells in αGalCer-treated cultures containing WASp-deficient cells. Furthermore, only ~50% of WASp-deficient tetramer⁺ cells exhibited CFSE dilution compared to ~80% in WT controls (Fig. 4A). These differences in proliferation were evident at 72 hour post-stimulation and especially pronounced at 96 hours post-stimulation. The proliferative defect in WASp-deficient tetramer⁺ cells could not be rescued even with high doses of αGalCer (up to 100ng/ml, data not shown).

Figure 4. WASp^−/− iNKT cells exhibit defects in antigen-induced proliferation.

All FACS plots are representative of the 72hr timepoint. (A) A representative FACS plot of iNKT cells stimulated with 1ng/ml αGalCer. Shaded and open histograms represent control and stimulated cells, respectively. Right panel: Summary of CFSE data showing percentage of tetramer⁺ cells after stimulation (left) and dividing cells within the tetramer⁺ gate (right). (B) CD45.1⁺ (WT) and CD45.2⁺ (WASp^−/−) cells were mixed and co-cultured at 50:50 ratio and stimulated with 1ng/ml αGalCer. (C) Analysis of IL-15 induced proliferation in WT and WASp – deficient splenocytes. Results are representative of 3 independent experiments, 6 mice per experiment. * p<0.05, ** p<0.01, *** p<0.001.

To eliminate the possibility that the reduced proliferative response of WASp-deficient iNKT cells might be secondary to inefficient antigen processing via WASp^−/− DCs, we also carried out studies using mixtures containg an equal number of CFSE labeled CD45.1⁺ WT and CD45.2⁺ WASp^−/− splenocytes. Cells were CFSE-labeled and stimulated with αGalCer as above. Cells were stained with anti-CD45.1 and anti-CD45.2 and proliferation of tetramer⁺ cells was analyzed by flow cytometry (Fig. 4B). As above, αGalCer treatment resulted in 4–7 fold expansion of WT cells, but failed to significantly expand WASp^−/− tetramer⁺ cells (Fig. 4B, right panels). Notably, under these culture conditions a larger percentage of WASp^−/− tetramer⁺ cells exhibited a diluted CFSE signal (~50–70% compared with ~50%; compare Fig 4A to Fig. 4B, right). However, this cycling population still exhibited substantially higher CFSE levels indicating that they sustained fewer cell divisions compared to WT cells (data not shown).

Impaired IL-15 receptor signaling results in significant reduction in thymic and peripheral iNKT numbers (33). Therefore, we also analyzed the relative response of WASp-deficient iNKT cells to IL-15. Total splenocytes from WT or WASp^−/− mice were labeled with CFSE and stimulated with IL-15. Under these conditions, the relative proportion of tetramer⁺ iNKT cells remained essentially identical in WT and WASp^−/− cultures (Fig. 4C). Both populations also exhibited similar levels of IL-15 induced CFSE dilution (Fig. 4C, right). Of note, the IL-15 response was much greater in tetramer⁻ cells and this response was also indistinguishable in cultures containing WT vs. WASp^−/− cells. The more robust response of tetramer⁻ cells is consistent with previous data regarding the IL-15 response of WT primary T and NK cells (34). Notably, both the WT and WASp^−/− iNKT cells also exhibited an equivalent proliferative response to PMA/Ionomycin stimulation (data not shown), consistent with the interpretation that WASp deficiency primarily impacts a proximal signal(s) initiated following antigen-receptor engagement. Together, these data support the conclusion that WASp^−/− iNKT cells exhibit a cell-intrinsic TCR signaling deficit.

WASp^−/− iNKT cells display defective cytokine secretion

The invariant NKT subset is characterized by its ability to rapidly produce large amounts of cytokines upon antigen receptor engagement (2). Previous studies have shown that WASp-deficient T cells and NK cells exhibit marked deficits in receptor-triggered cytokine production (14). Therefore, we also sought to determine whether cytokine production by WASp^−/− iNKT cells was similarly impaired. Splenocytes from WT or WASp^−/− mice were stimulated with either 10 or 100ng/ml αGalCer for 5–6 hours and cytokine production was analyzed using intracellular flow cytometry. As anticipated, a substantial proportion of both CD4⁺ and CD4⁻ WT iNKT cells expressed IL-4 and IFNγ following αGalCer stimulation (2). In contrast, neither CD4⁺ nor CD4⁻ WASp^−/− iNKT cells produced any detectable IL-4 or IFNγ in response to low level antigen (10 ng/ml) stimulation; and this deficit was only partially rescued using higher levels of antigen stimulation (100ng/ml; Fig. 5A&B). In contrast, both WT and WASp^−/− iNKT cells produced similar levels of both cytokines in response to PMA/Ionomycin stimulation (Fig. 5B). As described above, co-cultures containing WT APCs were also used to rule out the possibility that altered antigen presentation might impact these findings. The defect in cytokine production in WASp^−/− iNKT cells was not rescued using this approach (data not shown). Together, these data demonstrate that WASp plays a critical role in promoting antigen-induced, effector cytokine production by iNKT cells.

Figure 5. WASp^−/− iNKT cells exhibit defects in antigen-triggered cytokine production.

(A) A representative FACS plot showing IL-4 and IFNγ cytokine production in Tetramer⁺CD4⁺ iNKT cells stimulated with 10ng/ml αGalCer for 5 hours. (B) Combined results are shown for CD4⁺ (top right panel) and CD4⁻ (bottom right panel) iNKT subsets. Data shown are representative of one of 2 experiments. * p<0.05, ** p<0.01, *** p<0.001.

WASp promotes iNKT CD11a expression and tissue localization

As shown above, WASp deficiency exerts the greatest peripheral impact on liver resident iNKT cells. Notably, signaling through the CD11a integrin is required for retention of iNKT cells within the liver (35). We therefore assessed CD11a expression in iNKT cells derived from WT vs WASp^−/− animals to determine whether the reduction in tissue resident cells might correlate with altered CD11a expression. Consistent with this idea, we observed a marked reduction in CD11a expression in WASp-deficient iNKT cells (Fig. 6A). This decrease was present in all mature iNKT cells including cells derived from the thymus, spleen, bone marrow and liver (data not shown). It was also present in both CD4⁺ and CD4⁻ mature iNKT cells. Furthermore, decreased CD11a expression was also readily apparent in CD24⁺ immature thymic iNKT cells (data not shown), suggesting that the absence of WASp results in reduced CD11a expression at multiple stages of iNKT development.

Figure 6. WASp^−/− iNKT cells exhibit reduced integrin expression and altered homeostasis in vivo.

(A) Representative CD11a histogram on CD24⁻CD4⁺ splenic iNKT cells. Numbers in the upper left indicate CD11a median fluorescence intensity (MFI). Combined data from splenic iNKT and naïve CD4⁺ T cells are shown in the right panel. Similar data was obtained from thymus, bone marrow and liver (not shown). Results are representative of 2 independent experiments, 6 mice/experiment. (B) Analysis of CD11a expression on WASp⁺ and WASp⁻ splenic iNKT cells in carrier females. Results are representative of two experiments, 3–4 mice/experiment. (C) Analysis of Rag2^−/− mice transplanted with thymic iNKT cells. Upper left panel shows analysis of input cells. Lower left panels show analysis of tissues using CD45.1 (WT) or CD45.2 (WASp^−/−) and homeostatic proliferation was assessed using CFSE dilution (FACS histogram). Data shown are representative of one of 2 experiments. AU: Arbitrary Units. ** p<0.01, *** p<0.001.

We also evaluated heterozygote WASp carriers to determine whether altered CD11a expression levels were likely to represent a cell-intrinsic defect. Analysis using CD11a, iNKT markers and WASp expression again revealed a specific reduction in relative CD11a expression in WASp-deficient iNKT cells within the thymus, bone marrow, spleen and liver (Fig 6B and data not shown). This change was specific to iNKT cells, as similar levels of CD11a were present on conventional, splenic CD4⁺ T cells derived from WT and WASp-deficient animals (Fig 6A, right panel).

Both WASp^−/− regulatory T cells and WASp^−/− mature B cells exhibit impaired homeostasis and decreased fitness in vivo (4, 5, 17–19). To determine whether WASp^−/− iNKT cells might also exhibit altered fitness in vivo we carried out a competitive analysis of WT vs WASp^−/− iNKT cells in lymphopenic mice. This approach allowed us to assess both relative fitness and tissue distribution in the absence of any developmental inputs. We enriched for thymic iNKT cells by magnetically depleting CD8⁺ and CD24⁺ cells, resulting in a 100-fold enrichment of the iNKT population (with iNKT comprising at least 15% of input cells, Fig. 6C, top panel). These enriched cell populations were labeled with CFSE, mixed at 50:50 ratio of WT and WASp^−/− cells, and transferred into unmanipulated Rag2^−/− mice. Recipient animals were evaluated at 7 days post-transplant for the relative proportion of engrafted cells in various lymphoid and tissue compartments. Equivalent numbers of WT and WASp^−/− iNKT cells were present within the bone marrow. In contrast, the relative proportion of WT cells increased substantially within the spleen and, particularly, within the liver (Fig. 6C). This skewing did not appear to be due to decreased proliferative capacity, as both WT and WASp-deficient iNKT cells exhibited similar levels of CFSE dilution (Fig. 6C, histogram). Thus, WASp-deficiency results in impaired iNKT homeostasis by modulating iNKT migration and/or retention within secondary effector tissues.

Discussion

As part of this work and previous studies, we have completed extensive analyses of the relative selective advantage for WASp⁺ cells in distinct myeloid and lymphoid cell subsets. Our current and previous data (4, 19) fail to demonstrate any selective advantage WT vs. WASp⁻ multipotent hematopoietic stem cells (HSC) in mice. In addition, we have observed no significant differences among various myeloid-derived subsets (data not shown). In contrast, there is a pronounced enrichment for WASp⁺ cells within lymphoid populations, including various T and B cell subsets, as well as in the NK lineage. Our findings are consistent with a recently published report that used a similar methodology to examine multiple hematopoietic cell populations (5). The authors of that study also failed to identify any selective advantage for WT HSC or myeloid cells but demonstrated a pronounced selective advantage of WASp⁺ cells within lymphoid subsets, including peripheral T cell populations and MZ B cells. Surprisingly, while these authors were the first to report a selective advantage for thymic and splenic iNKT cells, they did not detect a significant advantage for WASp⁺ splenic NK cells. This latter difference may reflect the smaller number of animals examined in the previous study Consistent with our findings, we and others have previously observed selective outgrowth of WASp⁺ T-regs and NK cells in patients with revertant mutations within the WAS gene (19, 36, 37). Thus, our current findings and other previously published data (5, 38–40) clearly demonstrate the significant selective advantage for WASp⁺ cells within nearly all mature lymphoid subsets in human WAS patients and murine models of WAS.

In the current work, we focused on the marked selective advantage observed for mature WASp⁺ iNKT cells and also carried out studies to assess the functional impact of WASp deficiency within this subset. Using heterozygous carrier female mice, mixed bone marrow chimera studies, and adoptive cell transfer experiments we demonstrate that WASp⁺ iNKT cells exhibit a strong selective advantage in comparison with WASp⁻ iNKT cells in the thymus, spleen and liver. Our combined data also indicate that WASp is partially required for late-stage thymic iNKT development and egress; is essential for antigen-dependent iNKT function; and is important for peripheral iNKT homeostasis.

WASp⁻ iNKT cells exhibit a modest selective deficit during the final stage of mature thymic iNKT (CD24⁺ to CD24⁻CD44⁺) development. This deficit is only apparent, however, in two week old animals or within the context of competing WASp⁺ iNKT cells. In contrast, adult WASp^−/− animals generate normal numbers of mature thymic iNKT cells. Our findings suggest that WASp^−/− iNKT cells transition normally through positive selection, but exhibit a reduced proliferative burst during the transition from the CD24⁺ to CD24^neg developmental stage. This decreased maturation rate is readily detected in very young animals and in heterozygous WASp carriers where WASp^neg iNKT cells are partially out-competed in the presence of WASp⁺ mature thymic iNKT cells. Transition into the mature (CD24⁻) iNKT stage is mediated via TCR and SLAM-family receptor signaling (41–43) and WASp has previously been shown to impact both of these signaling cascades in primary T cells (7, 44). In addition, our data clearly show that WASp deficiency impacts TCR signaling in mature iNKT cells. Based on our data, we anticipate that the more limited impact of WASp-deficiency on thymic iNKT numbers in adult WASp^−/− mice reflects inefficient egress and resultant accumulation of thymic iNKT cells. Emigration of thymic iNKT cells is regulated by S1P signaling (31). CD69 acts as a negative regulator of these events; and S1P-receptor engagement also promotes down-regulation of CD69 surface expression (32). We and others have shown that WASp mediates S1P signaling in marginal zone B cells (4, 5), suggesting that increased CD69 expression on WASp^−/− peripheral iNKT cells may represent inefficient S1P signaling in WASp^−/− iNKT cells. Together, these findings suggest that WASp plays a role in both the initial development of CD24⁻ iNKT cells and in the egress of mature iNKT cells from the thymus to the periphery.

The most significant changes observed in our study were the alterations in peripheral iNKT compartment in WASp^−/− mice. WASp deficiency specifically impacted cells within the liver and spleen but not the bone marrow. Interestingly, the numbers of splenic, but not liver, iNKT cells reached WT levels in older animals. This tissue and age-specific phenotype suggests several roles for WASp in peripheral iNKT homeostasis. First, reduced egress of thymic iNKT cells likely leads to a diminished pool of peripheral iNKT cells and slower seeding of peripheral tissues. Secondly, WASp−/− iNKT exhibit a peripheral homeostatic defect most likely due to altered tissue-specific migration and/or retention. We observed a marked selective advantage for WASp⁺ iNKT cells within the spleen and the liver (compared to the BM) in both WASp heterozygous carriers and in Rag2-KO mice transplanted with iNKT-enriched thymocytes. The migration and retention of iNKT cells within peripheral tissues is mediated by a complex interplay between integrins (CD11a(35)), chemokines (CXCL16(45)) and co-stimulatory factors (ICOS(46)). We observed significantly reduced levels of CD11a integrin expression on WASp-deficient iNKT cells derived from all tissues; and this defect was cell-intrinsic, as analysis of WASp heterozygous mice also revealed reduced CD11a levels specifically in WASp⁻ iNKT cells. Signaling through CD11a is crucial for iNKT migration to the liver (35) and we have recently demonstrated that WASp-deficient marginal zone B cells are impaired in their ability to aggregate CD11a and form a functional immune synapse (4). Notably, we did not observe lower CD11a levels on WASp^−/− MZ B cells (4) or naïve CD4⁺ T cells, suggesting differential regulation of CD11a expression in iNKT vs. conventional T cells or MZ B cells. Our combined observations suggest that WASp-mediated signals might be required for sustained CD11a expression in vivo; and that altered CD11a signals are likely to impact iNKT homeostasis in the liver. Altered signaling through the chemokine receptor, CXCR6, may also impact this process since CXCR6 signaling, via interaction with CXCL16, is essential for iNKT migration to the liver (45). Further experiments will be required to determine whether WASp-deficient iNKT cells exhibit altered CXCR6 expression and/or reduced migration to CXCL16.

Notably, previous studies have shown that CD11a-deficient or CXCR6-deficient mice do not exhibit deficits in splenic iNKT numbers (35, 45). Thus, while we observed reduced CD11a on iNKT cells derived from all the tissues, this defect cannot explain the altered cell numbers within the spleen in WASp^−/− animals. In addition to chemokines and integrins, peripheral iNKT homeostasis is dependent on a steady provision of costimulatory signals. While thymic iNKT development depends on CD28-B7 interactions (47), peripheral survival relies on co-stimulatory signals delivered through ICOS/ICOS-L interaction (46). Although WASp is involved in CD28-mediated signaling (48), its role in ICOS signaling has not yet been established.

Notably, we while observed a marked selective advantage for WASp⁺ iNKT cells within the spleen and the liver, we observed similar levels of CFSE dilution in WASp−/− vs. WT iNKT cells in transplanted Rag2^−/− mice. This observation strongly suggests that, unlike the positioning signals described above, the events mediating iNKT homeostatic proliferation are intact in the absence of WASp. Consistent with this view, we also observed no difference in the proliferative response of WT vs. WASp^−/− iNKT cells to IL-15, the primary mediator of iNKT homeostatic proliferation in vivo (33).

Taken together, our findings suggest a dual role for WASp in peripheral iNKT homeostasis: WASp is important for iNKT migration to the peripheral tissues and is involved in promoting cellular retention within the tissues, in part by regulating integrin surface expression. While WASp is required for the homeostasis of both iNKT and regulatory T cells, the mechanism(s) involved appear to be distinct for each subset. Unlike iNKT cells, WASp^−/− mice generate normal numbers of splenic T-regs (18, 19) . However, in contrast to WASp^−/− iNKT cells, WASp^−/− T-regs exhibit a markedly reduced rate of homeostatic proliferation and this correlates with progressive loss of this population when placed in direct competition with WT T-regs (19). These contrasting phenotypes suggest separate and non-overlapping functions for WASp in promoting homeostasis of specific T cell subsets: WASp is involved in both antigen-driven homeostatic proliferation and in chemokine and integrin-mediated tissue migration and retention. Maintenance of regulatory T cells and other antigen-dependent T cell lineages requires WASp activity for persistent signaling through the antigen receptor. In non-canonical subsets such as iNKT cells, homeostatic proliferation is driven by cytokines rather than antigen, and our data suggest that this subset requires WASp for entry and retention in secondary lymphoid tissues and, possibly, also for key co-stimulatory signals.

Finally, our data demonstrate, not surprisingly, that WASp is essential for antigen-induced proliferation and cytokine production in iNKT cells. As previously reported for mutant CD4 T cells (14–16), our findings imply that WASp-deficient iNKT cells manifest a proximal, cell-intrinsic deficit in antigen-receptor signaling. Consistent with this view, WT and WASp^−/− iNKT cells expand equivalently in response to signals (including PMA/Ionomycin stimulation) that bypass these proximal events. Notably, the proliferation defect in WASp-deficient CD4⁺ T cells correlates with markedly reduced IL-2 production; and can be rescued by provision of exogenous IL-2 (15, 16). In contrast, provision of WT iNKT in trans failed to rescue the sustained cycling defect in WASp^−/− iNKT cells. This observation suggests that, unlike CD4 T cells, antigen-driven iNKT activation is less dependent upon autocrine signals. The cytokine production defect observed in antigen-triggered WASp-deficient iNKT cells is also consistent with previous work in both T and NK cells (15, 49). As with proliferation, the defect in cytokine production appears to reflect inefficient proximal signaling, as PMA/Ionomycin stimulation resulted in equivalent cytokine production by WT and WASp-deficient cells. In peripheral tissues, iNKT cells serve as immune sentinels, proliferating and producing large amounts of cytokines following triggering of the antigen receptor (2). WASp-deficient iNKT cells are unable to properly perform both of these crucial tasks. These findings imply that WASp^−/− animals should manifest compromised iNKT response in vivo and be at increased risk for specific viral and intracellular infections. Further work using lineage specific targeting will be required to directly test this idea.

Of note, a report published during the review of this manuscript also highlights the crucial role for WASp in antigen-mediated proliferation and cytokine production of iNKT cells (50). Consistent with our in vitro data, in vivo antigen–challenged WASp^−/− iNKT cells exhibit defective cytokine production and proliferation. In contrast to our data, however, the authors imply a crucial role for WASp in the transition from the NK1.1⁻ to NK1.1⁺ mature iNKT cell stage. These authors reported a near complete absence of CD44⁺ NK1.1⁺ cells in both the thymus and periphery of WASp^−/− animals (50). In contrast, as described herein, while we observed decreased peripheral iNKT numbers, these cells exhibited WT levels of CD44, NK1.1 and all other maturation markers (with the exception of CD69). Notably, the WASp^−/− mice used in our study were generated independently (15) and have been backcrossed unto the B6 background for more than 10 generations. Since NK1.1 is expressed on the B6 but not the 129 background (51), is seems most likely that background strain differences are responsible for the aberrant NK1.1 expression observed by Locci and colleagues (50). The iNKT maturation deficit implicated by these authors might be clarified by analysis based upon additional maturation markers.

Combined, our findings suggest a model in which WASp is important for both homeostasis and function of iNKT cells. Our data would predict that while WAS patients are likely to exhibit normal or only mildly reduced iNKT numbers in the peripheral blood, that such cells are likely to show significant defects in antigen-induced proliferation, cytokine secretion and tissue targeting. Combined with the overall immunodeficiency and immune dysfunction observed in WAS patients, reduced iNKT functionality may contribute to both the elevated infection rate and increased autoimmune incidence observed in this complex disorder.