Clinical and Hematologic Effects of Endotoxin in Warts, Hypogammaglobulinemia, Infections, and Myelokathexis Syndrome Model Mice

Warts, hypogammaglobulinemia, infections, and myelokathexis (WHIM) syndrome is a rare, autosomal-dominant, primary immunodeficiency disease caused by mutations that truncate the C-terminal tail of the chemokine receptor CXCR4 (1–4). WHIM mutations are predominantly CXCR4 protein-terminating nonsense mutations, although frameshifted missense mutations that add neosequence to the truncated natural C-tail also occur. In both cases, critical sites needed for G protein–coupled receptor kinase–mediated phosphorylation are eliminated, and arrestin-dependent desensitization and internalization of the mutant receptor are impaired, resulting in increased cellular responses to the CXCR4 agonist CXCL12 (5).

Many WHIM patients present with panleukopenia, whereas severe noncyclic neutropenia is almost always present. Hence neutropenia pathogenesis has been the most extensively investigated hematologic aspect of the disease (4, 6, 7). Because CXCR4 signaling normally restricts exit of neutrophils from the bone marrow to the blood and, conversely, promotes the return of neutrophils from the blood to the bone marrow, gain-of-function WHIM mutations shift the neutrophil distribution from blood to bone marrow, resulting in myeloid hyperplasia and neutropenia, the key hematopathologic features of myelokathexis, the M in the acronym WHIM. Many bone marrow neutrophils in WHIM patients appear dysmorphic, which may be the result of accelerated senescence of bone marrow neutrophils and/or accelerated homing of senescent apoptotic neutrophils from blood to bone marrow (8). Like neutrophils, lymphocytes appear to develop normally in WHIM syndrome but become abnormally distributed, resulting in peripheral blood lymphopenia (4, 9). B cells are the most commonly and severely deficient lymphocyte subset in the blood, even more severely deficient than neutrophils, and this is associated with mild-to-moderate hypogammaglobulinemia and reports of impaired vaccine responses (4).

Most WHIM patients eventually develop warts, which are refractory to conventional treatments. Several adult patients have been reported to develop squamous cell carcinoma from human papillomavirus strains in the oropharynx and anogenital regions. However, WHIM patients almost always have a history of recurrent acute infections, especially in the otosinopulmonary tract and skin. Yet, invasive infection and sepsis are uncommon, suggesting that patient leukocytes may be functionally normal and that their abnormal distribution is the main factor increasing susceptibility to infection (4). In this regard, during acute localized bacterial infection, leukopenia has been reported to be transiently reversed to normal levels in WHIM patients, consistent with the leukocytosis seen in acutely infected nonimmunocompromised hosts, and this may contribute to the relatively low risk of invasive infection (4).

Treatment of WHIM syndrome is currently directed toward correcting leukopenia. G-CSF rapidly and selectively reverses neutropenia in WHIM patients and is the standard of care for severe congenital neutropenia, although its clinical efficacy has not been specifically tested in WHIM syndrome. Recent clinical trials have begun to test as mechanism-based therapy the specific CXCR4 antagonists plerixafor (also known as AMD3100; brand name, Mozobil; Sanofi-Genzyme) and mavorixafor (X4 Pharma), which rapidly and durably reverse panleukopenia, with preliminary reports of clinical efficacy in small numbers of patients (3, 10). Ig replacement is also commonly used to treat WHIM patients with hypogammaglobulinemia and frequent infections (4, 11).

In this study, we experimentally addressed why sepsis is a relatively uncommon problem in WHIM patients by testing whether WHIM model mice carrying the second most common WHIM mutation at nucleotide 1013, which terminates the mouse CXCR4 protein at Ser³³⁸ (9), might be resistant to bacterial LPS, the master mediator of septic shock. Cxcr4^+/1013 mice have panleukopenia but do not develop spontaneous infections when housed under specific pathogen-free conditions. LPS is a major cell wall component of Gram-negative bacteria, which is recognized by the pattern recognition receptor TLR4 and is widely used to induce endotoxic shock in mice (12). Moreover, although bacterial sepsis and LPS challenge both rapidly induce severe lymphopenia (13, 14), how leukocytes might redistribute during endotoxic shock in WHIM syndrome is not known.

All animal experiments were performed as defined by an animal study proposal approved by the Animal Care and Use Committee of the National Institute of Allergy and Infectious Diseases, National Institutes of Health (NIH). Wild-type (WT) and WHIM mouse strains of both sexes on a congenic C57BL/6 CD45.1, CD45.2, or heterozygous CD45.1/CD45.2 background were used and have been previously described (9, 15). All WT and WHIM mice were littermates. The mice were maintained in a specific pathogen-free facility at the NIH, with 12 h of day/night cycle, and had access to food and water ad libitum.

EDTA disodium salt solution was purchased from Sigma-Aldrich (St. Louis, MO). FBS was purchased from GeminiBio (West Sacramento, CA), and ACK lysing buffer was purchased from Quality Biological (Gaithersburg, MD). Plerixafor (AMD3100) was provided for investigative use by Sanofi-Genzyme Corp. (Paris, France). LPS was from Escherichia coli strain O128:B12 (Sigma-Aldrich). E. coli strain W3110 was acquired from the Coli Genetic Stock Center at Yale University (New Haven, CT).

LPS from E. coli O128:B12 was freshly dissolved in Dulbecco’s PBS (Thermo Fisher Scientific, Waltham, MA) before performing the injections. The doses of LPS administered i.p. are indicated in the respective figure legends. Control mouse groups were injected with the same volume of PBS (0.2 ml/mouse). Mice were then analyzed as described in the figure legends.

A single colony of E. coli strain W3110 was grown overnight at 37°C in tryptic soy broth. Bacteria were then diluted 1:100 in tryptic soy broth and cultured at 37°C until early log phase was reached (OD₆₀₀ = 0.4–0.6). Bacteria were washed once with PBS and resuspended at 2.5 × 10⁷ CFUs/ml in PBS. Mice were injected i.p. with 200 µl of PBS alone or bacterial suspension (5 × 10⁶ CFUs/mouse). Three hours postinfection, leukocyte populations in the blood were analyzed by flow cytometry.

Blood samples from mice were collected via the maxillary vein using EDTA as anticoagulant (Sarstedt, Nümbrecht, Germany). The spleens, mesenteric lymph nodes (mLNs), and inguinal lymph nodes (iLNs) were harvested, then gently pressed on filter membranes using a syringe piston. RBCs in the blood and spleen samples were lysed using ACK lysis buffer. Bone marrow cells were isolated from the tibia and fibula and were flushed out using FACS buffer (PBS containing 2% FBS and 2 mM EDTA) via a syringe and needle. The peritoneal exudate cells were collected by injecting and collecting 4 ml of RPMI-1640 medium into the peritoneal cavity. Single-cell suspensions of bone marrow cells, splenocytes, and lymph node cells were washed and filtered through a 70-µm filter before cell enumeration and staining for flow cytometric analysis.

Mice were used at ∼6–8 wk of age for competitive transplantation experiments. Total bone marrow cells from CD45 isoform-marked WT and WHIM donor mice were mixed in a ratio of 1:1 and introduced via tail-vein injections into sex-matched CD45 isoform-marked Cxcr4^+/+ recipient mice that had undergone lethal irradiation (9 Gy) 8 h before transplantation. The input donor ratio of mixed bone marrow cells was verified by flow cytometry. Irradiated recipient mice were provided neomycin-supplemented water for 4 wk after transplantation. The mice were challenged with LPS after 8 wk of bone marrow reconstitution. WHIM leukocytes contributed ∼2–4% of total blood leukocytes in PBS-treated host mice after 8 wk of reconstitution.

Cells were resuspended in PBS prior to being mixed in a 1:1 ratio with ViaStain AOPI Staining Solution (Nexcelom Bioscience, Lawrence, MA). The leukocytes were counted using a Cellometer Auto 2000 Cell Viability Counter (Nexcelom Bioscience). The frequencies and absolute cell numbers of leukocytes were established by flow cytometry.

Because older and male mice are more sensitive toward LPS than younger or female mice (16), we used 3- to 6-mo-old male WT and WHIM littermates for the survival experiments. Mice were given a single dose of ∼20 mg/kg LPS i.p., then were monitored every 12 h for morbidity and mortality. Either the mortality was recorded or the mice were euthanized after reaching a pain score of 3 (moribundity), as per NIH guidelines.

Cell-surface markers were assessed by FACS using mAbs conjugated to fluorochromes. Blood samples or single-cell suspensions of splenocytes, bone marrow cells, and lymph nodes were initially stained with the viability dyes LIVE/DEAD Fixable Aqua (Thermo Fisher Scientific) or Zombie ultraviolet (BioLegend, San Diego, CA), according to the manufacturer’s instructions. Then the cells were stained for 10 min at 4°C with anti-CD16/CD32 mAbs to block Fc receptors. Next, the fluorescently labeled Abs were added to the cells and stained in the dark for 30 min at 4°C. Before acquiring the cell data on the flow cytometer, the cells were washed with FACS buffer. For intracellular Bcl2 staining, the manufacturer’s instructions were followed (PE Hamster Anti-Mouse Bcl-2 Set; BD Biosciences, Franklin Lakes, NJ). All analyses were performed on live, singlet cells except for Annexin V/PI staining (eBioscience Annexin V Apoptosis Detection Kit FITC; catalog number: 88-8005-74), where no viability dyes were incorporated. All data were acquired on the BD LSRFortessa Cell Analyzer flow cytometer using the BD Diva acquisition software (BD Biosciences). The results were analyzed using the commercially available software FlowJo (version 10.6.1; BD Biosciences). The geometric mean fluorescence intensity (MFI) of CD86 or CXCR4 expression in Fig. 5B and 5C on CD45^lo B cells was relative to the geometric MFI on CD45^hi B cells in the same samples. Details of the Abs used in this study are provided in Supplemental Table I.

All mouse experiments were performed using a National Institute of Allergy and Infectious Diseases Animal Care and Use Committee–approved protocol in approved and certified facilities.

The statistical analyses were performed using the commercially available software, GraphPad Prism (GraphPad Software, La Jolla, CA). Data are presented as mean ± SEM.

When Cxcr4^+/+ (WT) and Cxcr4^+/1013 (WHIM) littermates were injected i.p. with a lethal 20 mg/kg dose of LPS from E. coli strain O128:B12, ∼50% of the mice died in both groups during a 14-d observation period postchallenge, most within 48 h after the challenge (Fig. 1). Thus, a WHIM mutation does not appear to alter the sensitivity of mice to fatal endotoxic shock.

LPS is known to induce leukopenia in WT C57BL/6 mice (13, 14). To study whether this also occurs in WHIM mice on this genetic background, we challenged WHIM and WT littermates with a nonlethal 5 mg/kg dose of LPS and 3 h later analyzed blood leukocytes by flow cytometry. This time interval is sufficiently short to obviate a contribution from hematopoiesis to any changes that might be observed. Consistent with previous reports, compared with control PBS-injected WT mice, we observed a marked reduction of the absolute number of CD11b⁺ myeloid cells, NK cells, CD4⁺ and CD8⁺ T cells, and CD19⁺ B cells, but not neutrophils, in LPS-challenged WT mice (Fig. 2A). LPS affected only the distribution of CD8⁺ T cells, which were reduced in frequency by ∼50% compared with control PBS-injected WT mice (Fig. 2B). In marked contrast, the absolute cell numbers were not significantly changed for any of the leukocyte subsets in LPS-challenged WHIM mice relative to control PBS-injected WHIM mice, although there was a trend toward an increase in the absolute B cell number (Fig. 2A). In addition, unlike in WT mice, the percentage of B cells was markedly increased in LPS-challenged WHIM mice compared with PBS-injected WHIM controls, whereas the percentages of CD4⁺ T cells, NK cells, and CD11b⁺ myeloid cells were decreased, and the Ly6G⁺ neutrophil and CD8⁺ T cell frequencies remained unaltered (Fig. 2B). Thus, the leukocyte redistribution dynamics after LPS challenge appeared to be unusual and distinct in WHIM mice for all subsets examined compared with WT mice, with the increase in absolute number and frequency of B cells in endotoxemic WHIM mice being the most anomalous result.

We therefore focused our further investigation on B cells in the model. We found that in endotoxemic WHIM mice, ≥50% of the blood CD19⁺ B cells expressed low levels of cell-surface CD45 (CD45^lo), a tyrosine phosphatase expressed on all leukocytes (Fig. 3A), whereas CD45^lo CD19⁺ B cells were either absent or at most 5% of total B cells in both PBS-injected WT and WHIM controls and were only modestly increased in LPS-challenged WT mice (Fig. 3). Thus, the anomalous increase in B cells in the blood after LPS challenge in WHIM mice appears to be driven mainly by the specific emergence of a CD45^loCD19⁺ subset of B cells.

To study the effect of LPS on CD45^lo B cells in more detail, we next performed LPS time-course and dose-response studies. In WT mice, the proportion and absolute numbers of total B cells reduced drastically after 3 h of LPS challenge, while in WHIM mice, there was a transient increase of B cell proportions and numbers at 3 h, only to gradually wane by 24 h (Fig. 4A, 4B). The proportions of CD45^lo B cells in WT mice peaked at ∼15% 6 h after the challenge and returned to very low baseline levels, <5%, 24 h postchallenge (Fig. 4C). In contrast, in WHIM mice, CD45^lo B cells peaked at ∼60% of total B cells 3 h after the challenge and remained at that level through 12 h postchallenge until declining to ∼40% by 24 h postchallenge (Fig. 4C). In terms of absolute cell numbers, the CD45^lo B cells in WHIM mice rapidly increased at 3 h; however, by 14 and 24 h, their numbers were negligible and comparable with those in WT mice (Fig 4D). When we treated WHIM mice with increasing doses of LPS, we found a dose-dependent increase in the proportion and numbers of CD19⁺ and CD45^lo B cells at 3 h postchallenge, with a maximal response at 5 mg/kg (Fig. 4E–(H). Therefore, the phenomenon of enhanced emergence of CD45^lo B cells in the blood of WHIM mice on LPS exposure is dose dependent and rapid, which indicates that the cells are most likely being mobilized from the bone marrow by endotoxemia.

To understand the origin of CD45^lo B cells that emerged after LPS challenge, we studied the expression of a panel of B cell developmental markers (17, 18). Because the frequencies of CD45^lo B cells were very low in LPS-challenged WT mice, we limited the analysis to WHIM mice 3 h after LPS challenge. We found that CD45^lo B cells were negative for surface expression of CD43, which is a marker of Hardy fractions A–C of developing B cells (17, 18), and expressed lower levels of the pan-B cell marker B220 than CD45^hi B cells (Fig. 5A). Moreover, whereas CD45^hi B cells were heterogenous for IgD and IgM expression, all CD45^lo B cells were IgD⁻IgM⁻ (Fig. 5A). Differential expression of surface IgD and IgM enables identification of Hardy fractions D–F of developing B cells. Based on these maturation markers, the surface phenotype of CD45^lo B cells best fits the developing Hardy fraction D of CD43⁻B220⁺IgD⁻IgM⁻ late pre-B cells (17, 18). Furthermore, we also analyzed the recombination activating gene (RAG) expression in CD45^lo and CD45^hi B cells in LPS-treated WHIM mice expressing RAG-GFP (19). Consistent with the phenotype of late pre-B cells (18), all CD45^lo B cells in blood of WHIM mice were RAG⁺ (Supplemental Fig. 1A). Importantly, the MFI of RAG-GFP on CD45^lo B cells in blood and on late pre-B cells in the bone marrow from the same mice were comparable (Supplemental Fig. 1B). We also studied the cell-surface expression of the molecules, CD127 (IL-7Rα), CD23, and CD21. Late pre-B cells express high amounts of CD127, whereas peripheral B cells express CD23 and CD21 (18). Compared with the CD45^hi B cells, the CD45^lo B cells in the blood of LPS-treated WHIM mice expressed high amounts of CD127 and were CD23⁻CD21⁻ (Supplemental Fig. 1C, 1D). Therefore, most likely, LPS induces rapid and selective mobilization of late pre-B cells into the peripheral blood of WHIM mice, opposite to the expected CXCR4-dependent bone marrow retention mechanism. Consistent with this, the LPS-inducible B cell activation marker CD86 (20) was ∼50% lower on CD45^lo B cells compared with CD45^hi B cells in LPS-challenged WHIM mice (Fig. 5B). Furthermore, compared with CD45^hi B cells postchallenge, CD45^lo B cells expressed about three times higher levels of CXCR4 (Fig. 5C), which is known to be most highly expressed on B cells during development in the bone marrow (21, 22). Therefore, we defined the mobilized CD45^lo B cells in LPS-treated WHIM mice immunophenotypically as CD43⁻B220^loIgD⁻IgM⁻CXCR4^hiCD86^lo. Regarding viability, ∼20% of CD45^lo B cells stained with Annexin V, an ∼4-fold higher proportion than for CD45^hi B cells postchallenge (Fig. 5D). Consistent with this, the antiapoptotic molecule Bcl-2 was not detected in CD45^lo B cells but was expressed in ∼38% of CD45^hi B cells postchallenge (Fig. 5D).

On LPS exposure, bone marrow myelopoiesis increases, whereas lymphopoiesis decreases (13, 23, 24). Consequently, inflammation-associated reduction of immature B cells in the bone marrow can be associated with extramedullary B lymphopoiesis in the spleen (25–27). Therefore, we next investigated whether CD45^lo B cells were observed in other sites besides the blood 3 h after LPS treatment of WHIM mice (Fig. 6). Only CD45^hi B cells were detected in the mLNs and iLNs, whereas both CD45^lo and CD45^hi B cells were detected in the spleen (Fig. 6A). Nevertheless, CD45^hi B cells were the predominant population in spleen with CD45^lo B cells accounting for ∼5% of total B cells (Fig. 6B). We also quantified the CD45^lo and CD45^hi B cells at the site of injection, i.e., the peritoneal cavity. Similar to the lymph nodes, the B cells were exclusively CD45^hi in the peritoneum (Supplemental Fig. 1E). Given the marked disparity in the distribution of B cells stratified by CD45 expression in blood and spleen, we stimulated splenocytes from naive WT and WHIM mice in vitro with varying doses of LPS in the presence or absence of mouse CXCL12. However, consistent with a previous report (25), we failed to see the appearance of CD45^lo B cells after 3 h of stimulation (data not shown). Therefore, in WHIM mice, during endotoxic shock, CD45^lo B cells are most likely being mobilized from the bone marrow to the blood from which they may seed the spleen where extramedullary B lymphopoiesis might then occur.

The enhanced mobilization of CD45^lo B cells to the blood after LPS challenge in WHIM mice compared with WT mice provides genetic evidence that the phenomenon is mediated by CXCR4 signaling. Because pre-B cells from WT bone marrow are known to express very high levels of CXCR4 (22), as do CD45^lo B cells in the blood of WHIM mice challenged with LPS, the mobilization of CD45^lo B cells by LPS may be dependent on not just CXCR4 signaling but also on expression of the WHIM receptor on these cells. However, because we tested only one line of WHIM mice in discovering this phenomenon, it remains possible that it results from a CXCR4-independent genetic artifact intrinsic to this line of mice. We therefore tested the CXCR4 signaling dependence hypothesis pharmacologically by treating the mice with 10 mg/kg of the specific CXCR4 antagonist AMD3100 before LPS challenge.

AMD3100 is known to mobilize most subsets of leukocytes into the bloodstream of WT mice, with a peak at ∼2.5 h after injection (28), whereas the peak occurs at ∼6 h for total leukocytes in healthy human subjects and WHIM patients with individual subset variation (29, 30). For our experiments, we treated naive WT and WHIM mice with AMD3100, and 2.5 h later either collected blood samples for immediate FACS analysis or injected LPS and analyzed blood leukocyte populations 3 h after endotoxin challenge. Consistent with previous reports (9, 28), AMD3100 treatment resulted in a surge of the absolute number of all subsets of leukocytes measured in the blood of WT mice, and we observed a similar response in WHIM mice (Supplemental Figs. 2, 3). Thus, panleukopenia observed in WHIM mice was corrected by injecting a high-dose bolus of AMD3100.

As expected, AMD3100 alone also increased the absolute number of CD45^lo B cells in the blood of WT and WHIM mice; however, unlike the far greater increase in WHIM mice compared with WT mice after LPS stimulation, the increase in absolute cell number induced by AMD3100 alone did not differ between the two mouse strains (Fig. 7C). Moreover, the proportion of total B cells in the blood represented by CD45^lo B cells after AMD3100 injection did not differ between WT and WHIM mice and was very low, <5% (Fig. 7B). Together, the results are consistent with a CXCR4-dependent but WHIM allele-independent response of CD45^lo B cells to AMD3100 alone and differ fundamentally from the LPS-stimulated mobilization of CD45^lo B cells, which appeared to be CXCR4 dependent and enhanced by ∼4-fold by the WHIM allele (Fig. 7B). Consistent with this interpretation, pretreatment with AMD3100 completely blocked LPS-enhanced CD45^lo B cell mobilization in WHIM mice, whereas it had no effect on the low-level LPS mobilization of CD45^lo B cells in WT mice, as determined by the percentage of total B cells in the blood (Fig. 7B). The selective inhibitory effect of AMD3100 on LPS-enhanced CD45^lo B cell mobilization in WHIM mice was also reflected in the abolition of the difference in absolute numbers of CD45^lo B cells in the blood of LPS-challenged WHIM and WT mice when pretreated with AMD3100 (Fig. 7C). Therefore, the combined genetic and pharmacologic data support the notion that the anomalous emergence of CD45^lo B cells after LPS challenge depends on the variant CXCR4 WHIM receptor and not on WT CXCR4. The pattern was different for all other blood leukocyte and bone marrow progenitor cells tested. In particular, AMD3100 consistently increased the absolute number of each cell type in the blood to a similar level in WT and WHIM mice; LPS alone did not increase the absolute number or frequency of other cell types, and LPS contracted the rise in the absolute number of other cell types mobilized by AMD3100 without affecting the subset frequency. The frequencies were also not different for all other subsets in WT versus WHIM mice under the various experimental conditions (Supplemental Figs. 2, 3).

Because the anomalous LPS-enhanced emergence of CD45^lo B cells in WHIM mice is most likely driven by aberrant CXCR4 signaling because of the WHIM mutation, we next tested whether the phenomenon was cell intrinsic. For this, we conducted competitive transplantation experiments using mixed bone marrow cells from WT and WHIM mice as donors in irradiated WT recipients. After 6–8 wk of reconstitution, the mice were challenged with LPS, and their blood samples were collected for leukocyte analysis 3 h later (Fig. 8A). An increase in the proportion of both WT and WHIM bone marrow–derived CD45^lo B cells was observed in LPS-treated mice. However, WHIM CD45^lo B cells were significantly higher compared with those of WT origin (Fig. 8B, 8C). Therefore, anomalous LPS-enhanced mobilization of CD45^lo WHIM B cells is cell intrinsic.

Finally, we tested whether infection with E. coli would also induce mobilization of CD45^lo B cells in WHIM mice. For this, we injected naive WT and WHIM mice i.p. with 5 × 10⁶ CFUs of E. coli, and blood leukocyte populations were analyzed by FACS 3 h later. LPS-treated WT and WHIM mice were included as positive controls. As shown in (Fig. 9, the frequency of CD45^lo B cells in E. coli–treated WHIM mice was markedly increased compared with that of WT mice. Therefore, E. coli infection recapitulated the phenomenon of selective LPS-enhanced CD45^lo B cell mobilization in WHIM mice.

In this study, we have shown in a mouse model of WHIM syndrome immunodeficiency that the S338X WHIM variant of CXCR4 mediates selective and anomalous egress of Hardy fraction D of late pre-B cells from the bone marrow to the blood in response to LPS and E. coli challenge, amounting to a majority of total B cells in the blood and accounting for all of the increase in total B cells in the blood in response to the challenge. Cells at this stage of B cell development are known to express high levels of CXCR4 (21, 22) and to downregulate Bcl-2 (31), which we confirmed. The mobilized cells have the immunophenotype CD43⁻B220^loIgD⁻IgM⁻CXCR4^hiCD86^lo and can be found in the blood in a CXCR4^WT-independent and CXCR4^WHIM-dependent manner within 3 h of LPS challenge. No other lymphocyte subsets, including mature B cells, were mobilized to the blood of WHIM mice by LPS, and the selective mobilization of late pre-B cells is opposite to the canonical ability of WHIM mutations to cause constitutive panleukopenia, including severe B lymphocytopenia, in part by promoting mature leukocyte retention in the bone marrow and leukocyte homing to the bone marrow from blood. Our results suggest a qualitatively distinct cell-intrinsic signaling capacity of CXCR4^WHIM compared with CXCR4^WT to mediate late pre-B cell egress that is revealed in the context of LPS and E. coli challenge.

To place our finding in a broader context, before the identification of CXCL12 as the primary ligand of CXCR4, it was characterized as a pre-B cell growth-stimulating factor (32). Mice deficient for CXCL12 die perinatally, and the mutant embryos display severely reduced B cell progenitors in fetal liver and bone marrow, highlighting its indispensable role in B lymphopoiesis (33). In adult vertebrates, B lymphopoiesis occurs in the bone marrow. Hematopoietic stem cells undergo a series of intermediate stages to become more committed to the B cell lineage. The B cell progenitors at the pro-B cell stage initiate rearrangement of IgH, followed by numerous rounds of clonal expansion at the pre-B cell stage, and then rearrange the L chain gene to express surface IgM on B cells. CXCL12/CXCR4 signaling is required for the development of small pre-B cells (B220⁺CD19⁺CD43⁻IgM⁻Forward scatter^lo) and subsequently immature B cells (B220⁺CD19⁺CD43⁻IgM⁺) from large pre-B cells (B220⁺CD19⁺CD43⁻IgM⁻Forward scatter^hi) because of its requirement for Ig gene L chain recombination (22). The semimature B cells expressing IgM and IgD on their surface exit the bone marrow to colonize the spleen, where they undergo maturation to become naive B cells (34). A small group of reticular stromal cells in the bone marrow designated CXCL12-associated reticular cells express high levels of CXCL12 and mediate development and homing of immature and mature B cells inside the niche environment (35). Thus, CXCL12/CXCR4 signaling is critical for B lymphopoiesis and accounts for severe B lymphocytopenia in WHIM syndrome, yet LPS appears to anomalously abrogate this signal in a variant WHIM receptor-specific and pre-B cell–specific manner. Inflammation is known to downregulate CXCL12 in the bone marrow, but this should have a general effect of reducing retention of cells in the bone marrow (14, 24, 36). In addition to directly binding to and blocking CXCR4, AMD3100 has been reported to induce downregulation of CXCL12 in the bone marrow (37, 38); however, there are no studies addressing the effect of LPS on this. Altogether, AMD3100 has a complex effect on the mobilization of the immature CD45^lo B cells in WHIM mice in that downregulation of bone marrow CXCL12 potentially occurs by both LPS and AMD3100 treatment, and both stimuli given alone are capable of successfully mobilizing CD45^lo B cells in WHIM mice, but they act antagonistically when given sequentially.

We also found that LPS-induced T lymphocytopenia, which is a well-known phenomenon, is neither CXCR4^WHIM nor CXCR4^WT dependent. Instead, the low levels of circulating T cells found in WHIM mice do not decline further after LPS challenge, and conversely, LPS challenge causes rapid and complete contraction of T cells mobilized to the blood by pretreatment with the CXCR4 antagonist AMD3100 (28, 39). The mechanism of LPS-induced lymphopenia is not well understood. Mice deficient for the LPS-sensing molecules, CD14 and TLR4, and IFN-γR^−/− mice show resistance toward the pathophysiologic effects of LPS. In all these mouse strains, acute leukopenia either does not occur or is transient on LPS treatment (40, 41). During inflammation, concomitant with lymphopenia, T cells rapidly infiltrate organs such as lung and liver in mice. In a mouse model of acute liver injury induced by LPS and Mycobacterium bovis bacillus Calmette–Guérin, the intrahepatic infiltrating T cells were significantly reduced by administration of anti-CXCL16 Ab, and the chemotactic potency of CXCL16 toward activated T cells has been verified in vitro (42). Also, IL-16 was shown to regulate T cell trafficking in lungs of mice with acute lung injury induced by hemorrhagic shock and cecal ligation and puncture (43). In contrast, the recruitment of adoptively transferred Ag-specific Th1 cells in lungs of intranasally LPS-treated STAT1^−/− mice was independent of expression of STAT-1 and its inducible CXCR3 ligands CXCL9, CXCL10, and CXCL11 (44). Moreover, in ICAM-1–deficient mice, T cell lymphopenia induced by high doses of LPS occurs to a similar extent as WT mice, unlike neutrophils, which rebound in the mutant mice (45).

Our study also shows that a WHIM mutation does not affect clinical sensitivity to fatal outcome after LPS challenge in mice, which aligns with the low risk for sepsis in WHIM patients despite their high risk for recurrent noninvasive bacterial infection (4). In WHIM patients, neutropenia and B lymphocytopenia are the most common and severe hematologic abnormalities, and both occur in the WHIM mouse. However, when housed in specific pathogen-free facilities, there is a much lower proportion of neutrophils in C57BL/6 mice compared with healthy humans, and their number is reduced less severely in WHIM mice than in WHIM patients. LPS did not affect the percentage or absolute number of neutrophils in WT or WHIM mice; however, it dramatically reversed the neutrophilia induced by AMD3100 in both WT and WHIM mice. Our study did not address the role of CXCR4 in neutrophil dynamics and host defense during bacterial infection. In previous studies, neutrophil mobilization during cecal ligation and puncture in mice has been shown to be dependent on CXCL12/CXCR4 signaling, which in turn impacts survival. Moreover, anti-CXCL12 antisera has been reported to reduce neutrophil mobilization from the bone marrow to blood and peritoneum, which results in higher bacterial load in the peritoneal cavity and reduced survival (36).

Mobilization of developing B cells from bone marrow to the periphery during inflammation is not unique to LPS challenge. Immature B220^loRAG⁺ B cells have been reported to be mobilized from the bone marrow to the spleen after immunization and during infection with the parasite Plasmodium yoelii (26). Similarly, during cecal ligation and puncture–induced polymicrobial sepsis in mice, CXCL12 expression in bone marrow is downregulated, resulting in the mobilization of immature B cell precursors. Furthermore, CXCL12 antisera reduces the inflammation-driven mobilization of CD19⁺AA4.1^hiB220^intermediate immature B cells (36). However, our study provides the first example of mobilization of late pre-B cells in a manner that is promoted selectively by LPS and involving a CXCR4 WHIM mutation. CXCR4 is known to be a part of the LPS-sensing apparatus, contributing to the activation of NF-κB and MAPKs (46); however, coupling of this pathway selectively to a variant WHIM receptor has not been investigated.

Inflammation promotes myelopoiesis (emergency granulopoiesis) while simultaneously downregulating B lymphopoiesis (13). It has been previously shown that in endotoxemic mice, LPS induces an increase in granulocyte-macrophage progenitors by upregulating G-CSF, IL-6, and GM-CSF in the bone marrow and spleen, whereas it reduces B cell progenitors in the bone marrow by downregulating CXCL12, stem cell factor, and IL-7 (23, 24). Similarly, infection with Trypanosoma brucei or Plasmodium yoelii in mice is known to favor granulopoiesis over lymphopoiesis and to induce transient extramedullary B lymphopoiesis in the spleen (26, 47). Consistent with this, we found that the frequency of CD45^lo B cells in the spleen of LPS-treated WHIM mice was increased, which suggests the potential for extramedullary B lymphopoiesis in LPS-challenged WHIM mice.

The biological consequences of LPS-driven mobilization of late pre-B cells remain to be explored. One possibility is that the cells might escape tolerance, which may lead to the development of B cell lymphomas, which have been reported in several WHIM patients (4). Of note, Waldenström macroglobulinemia (WM) is an incurable, indolent, IgM-secreting lymphoplasmacytic lymphoma (LPL), involving somatic mutations in the gene encoding the adaptor protein MYD88 (point mutations) in >90% of cases and mutations in the CXCR4 (frameshift or non-sense mutations, including known WHIM mutations) in ∼30% of patients. The disease is characterized by infiltration of LPL in the bone marrow, as well as IgM monoclonal gammopathy (48). A risk factor for developing WM and LPL-WM is septicemia with odds ratios of 2.6 and 2.4, respectively. Association between genetic mutations and systemic inflammation raises the possibility of immune disruption by the latter to augment underlying congenital or acquired genetic risk factors (49).

In conclusion, we report in this article the novel phenomenon of late pre-B cell egress during inflammation in the presence of a CXCR4 WHIM mutation. Further studies will be necessary to delineate the mechanism and biological implications of the mobilization of these B cells in the context of sepsis and aberrant signaling through a WHIM receptor.

We thank Dr. Alfred Singer for kindly providing the RAG-GFP mice. We acknowledge Dr. David H. McDermott and Dr. Mukta Deobagkar-Lele for critically reading the manuscript and providing helpful suggestions. (Fig. 8A was created with BioRender.com.

The authors have no financial conflicts of interest.