Glucocorticoids Regulate Bone Marrow B Lymphopoiesis After Stroke

Abstract

Rationale:

After stroke, patients frequently experience altered systemic immunity resulting in peripheral immunosuppression and higher susceptibility to infections, which is at least partly attributed to lymphopenia. The mechanisms that profoundly change the systemic leukocyte repertoire after stroke are incompletely understood. Emerging evidence indicates that stroke alters hematopoietic output of the bone marrow.

Objective:

To explore the mechanisms that lead to defects of B lymphopoiesis after ischemic stroke.

Methods and Results:

We here report that ischemic stroke triggers brain-bone marrow communication via hormonal long-range signals that regulate hematopoietic B lineage decisions. Bone marrow FACS analyses and serial intravital microscopy indicate that transient middle cerebral artery (tMCA) occlusion in mice arrests B cell development beginning at the pro-B cell stage. This phenotype was not rescued in Myd88^−/− and TLR4^−/− mice with disrupted Toll like receptor signaling or after blockage of peripheral sympathetic nerves. Mechanistically, we identified stroke-induced glucocorticoid release as the main instigator of B lymphopoiesis defects. B cell lineage-specific deletion of the glucocorticoid receptor in CD19-Cre loxP Nr3c1 mice attenuated lymphocytopenia after tMCA. In twenty patients with acute stroke, increased cortisol levels inversely correlated with blood lymphocyte numbers.

Conclusions:

Our data demonstrate that the hypothalamic–pituitary–adrenal axis mediates B lymphopoiesis defects after ischemic stroke.

INTRODUCTION

Cerebral ischemia triggers a complex neuroinflammatory response involving the early inflammatory activation of local tissue resident cells and infiltration of peripheral innate immune cells¹. Systemic immunity after cerebral ischemia is characterized by a biphasic response: An acute hyper-inflammatory phase promotes immune cell infiltration of the brain, which is rapidly followed by a sustained systemic immunosuppression in the subacute phase^2–5. In mice and patients who survive stroke, this latter phase is marked by a profound alteration of the blood leukocyte repertoire which associates with increased mortality^6,7, poor functional outcome⁸ and high incidence of infections^9–11. Immobilization and neurological deficits, resulting in reduced mucus clearance from the lung, were long thought to be the main cause of post-stroke infections. Lack of a proper immune response, however, may be the actual culprit¹². Indeed, low circulating lymphocyte counts after stroke point to a weakened adaptive immune system as a contributor to poor stroke outcomes^6,13.

Most studies on the adaptive immune system’s contribution to the neuroinflammatory response after stroke focussed on the brain^14,15, delineating protective and harmful functions of distinct lymphocyte subsets, including B cells^16–21. Despite emerging evidence for a pathophysiological role of B cells in stroke, and the dramatic change of the leukocyte repertoire in the peripheral blood, data on cross talk between the injured central nervous and the hematopoietic system are scarce. Given that B cells originate in the bone marrow, we sought to investigate post-stroke mechanisms that may dysregulate bone marrow B lineage decisions. The main hypothesis of our study was that long-distance communication between the ischemic brain and the bone marrow leads to post stroke lymphopenia.

Here we report that hormonal brain-bone marrow communication contributes to the development of systemic lymphopenia after stroke. Transient middle cerebral artery occlusion (tMCAO) in mice, triggers arrested B cell supply at the pro-B cell progenitor level. Ischemic stroke affects lymphopoiesis via endocrine signals, specifically glucocorticoids, resulting in paralyzed adaptive immunity.

METHODS

We confirm that the manuscript adheres to the Transparency and Openness Promotion (TOP) guidelines. The data that support the findings of this study are available from the corresponding author upon reasonable request.

Humans.

Peripheral blood from patients with early cerebral infarction (day 3) was collected in Goyang, South Korea. Patients with stroke were 68.8 ± 3.2 years old (mean ± SEM, n = 20, 65% male). Blood samples from patients diagnosed with dizziness, headaches, and weaknesses were used as controls. The specimens were collected over a 4-month period in a non-consecutive cohort. Sampling for cortisol levels was performed between 7:30 and 8:00 AM. Control patients were 61 ± 2.5 years old (mean ± SEM, n = 14, 50% male). Hematological measurements (Complete Blood Count) were obtained using the XN-9000 analyzer (Sysmex) at the Department of Laboratory Medicine, Dongguk University Ilsan Hospital, Goyang, South Korea. Serum cortisol concentration was measured using a competitive immunoassay (Elecsys cortisol assay) in a Cobas® 8000 Modular Analyzer (Roche Diagnostics). The Institutional Review Board of Dongguk University Ilsan Hospital approved the study. All patients or their legally authorized representatives gave a written informed consent.

Animals and surgical procedures.

C57BL/6 (000664), B6.129S7-Rag1^tm1Mom/J (002216), B6.B10ScN-Tlr4^lps-del/JthJ (007227), B6.129P2(SJL)-Myd88^tm1.1Defr/J (009088), B6.Cg-Nr3c1^tm1.1Jda/J (021021) and B6.129P2(C)-Cd19^tm1(cre)Cgn/J (006785), B6.Cg-Gt(ROSA)26Sor^{tm14(CAG-tdTomato)Hze}/J (007914), B6.Cg-Tg(CAG-DsRed*MST)1Nagy/J (006051) were purchased from Jackson Laboratory (Bar Harbor, ME, USA), and 10–14 weeks old mice were used for experiments. All mice were maintained and bred in the pathogen-free environment of the Massachusetts General Hospital animal facility (MGH, Boston, MA, USA). Genotyping for each strain was performed as described on the Jackson Laboratory website. Stroke was induced by transient occlusion of the middle cerebral artery (tMCAO) as described previously²² (14% male mice, 86% female mice). Mice were anesthetized with 1.5–2% isoflurane supplemented with oxygen, and temperature was kept constant using a temperature control system (Harvard Apparatus). Briefly, a midline neck skin incision was performed, the right carotid artery bifurcation was dissected, and the external carotid artery (ECA) was permanently ligated. A clip was then placed on the internal carotid artery, and a temporary ligation was tied onto the common carotid artery. After an arteriotomy was performed on the ECA, a silicon-coated filament (Doccol Corp.) was introduced from the ECA to the internal carotid artery after clamp removal, until block. The filament was left in place for 45 minutes before being pulled out to achieve reperfusion. Success of the occlusion and reperfusion were monitored by laser Doppler flowmetry (PeriFlux System 5000, Perimed). Animals with excessive bleeding during induction of tMCAO, with improper arterial occlusion and reperfusion monitored by laser Doppler, and mice without visible infarct following TTC staining were excluded from the study. Corresponding control groups were sham-operated animals. Here, surgery was performed using similar conditions as in tMCAO, including skin opening, dissection of the carotid bifurcation, and ligation of the right external carotid artery but without inducing brain ischemia. Afterwards, the incision was sutured and the mice allowed to wake up in a clean cage.

Myocardial infarction was induced by permanent ligation of the left anterior descending coronary artery, as described previously²³. Mice were anesthetized with 1.5–2% isoflurane supplemented with oxygen, intubated, and ventilated (Inspira, Harvard Apparatus). After thoracotomy, the heart was exposed; the left coronary artery was identified, and permanently ligated with a monofilament nylon 8–0 suture. The rib cage was then closed with two separate 5–0 sutures, a 20G sheath left between the sutures. Finally, air was removed from the thoracic cavity via the sheath and a 3 ml syringe prior to extubation. Twenty-four hours after this procedure, cardiac magnetic resonance imaging reported a significantly decreased ejection fraction (MI, 41.2% ± 2.5; naive controls, 67.4% ± 1.8; p< 0.001). Ischemia reperfusion injury of the heart was performed by transiently occluding the left anterior descending coronary artery followed by reperfusion 45 min after ischemia onset, as described previously²⁴. Myocardial ischemia was verified by TTC staining, as described below for the brain. In all experiments, mice were randomized to their group assignment and surgical procedures were performed in a blinded fashion. Protocols were approved by the Subcommittee on Animal Research Care at Massachusetts General Hospital.

Treatments.

A specific antagonist for the β3-adrenergic receptor (SR 59230A, Sigma-Aldrich) was used to inhibit β3-adrenergic signaling. Mice were injected at 5 mg/kg body weight i.p. twice per day until sacrifice. To inhibit β2-adrenergic signaling, mice were administered i.p. with ICI-118,551 hydrochloride (Sigma-Aldrich) daily at a dose of 1 mg/kg body weight. We used 6-hydroxydopamine (6-OHDA; Sigma-Aldrich) to deplete the sympathetic nerve fibers. 6-OHDA was dissolved in saline buffer supplemented with 0.01% ascorbic acid and mice were injected with two different doses prior to the surgery, starting with 100 mg/kg body weight five days before and again with 250 mg/kg body weight three days before stroke induction. For HPA axis stimulation, C57BL/6 mice were administered with 100 μg of corticotropin releasing factor (Sigma) for four consecutive days before harvest. Animals injected with PBS served as controls.

Evaluating brain injury.

Brains were harvested and sectioned every millimeter using adult mouse brain matrix slicer (Zivic Instruments). Slices were then incubated in a solution of 2% TTC (2,3,5-Triphenyltetrazolium chloride, Sigma-Aldrich) for 15 minutes while being protected from light. ImageJ software was used to quantitate stroke volume (imagej.nih.gov).

Cell preparation and flow cytometry.

Single-cell suspensions were obtained from blood, spleen, and bone marrow. Blood was collected following cardiac puncture in 50 mM EDTA, and erythrocytes were removed using a red blood cell lysis buffer (BioLegend). Bone marrow from the femurs was flushed with FACS buffer (1x PBS supplemented with 0.5% BSA) and gently mashed through 40μm cell strainers (BD Falcon) to obtain cell suspensions. Spleens were triturated and filtered through a 40μm filter to obtain single-cell suspensions, and red blood cells were lysed with 1x red blood cell lysis buffer (BioLegend). After centrifugation (340 g for 5 min at 4℃), splenocytes and bone marrow cells were resuspended in 3ml and 1ml FACS buffer, respectively. Total cell numbers were determined using a hemocytometer and Trypan Blue staining method for cell viability (Cellgro). For flow cytometry, all cells were stained for 30 min at 4°C in 300μl FACS buffer. For B cell progenitor staining, single cell suspensions were first stained with PE-conjugated anti-mouse antibody lineage cocktail including CD3 (100206), CD4 (130310), CD8 (12–0081-85,e-bioscience), CD90.2 (140308), NK1.1 (108707), TER119 (116208), CD11b (101208), Ly6G (127608), CD115 (135506), CD11c (117308). This was followed by a second staining with antibodies for IgM-FITC (406506), CD93-PerCP/Cy5.5 (136512), B220-PE/Cy7 (103222), CD24-BV605 (101827), CD19-BV711 (115555), CD43-APC (143208), IgD-AF700 (405730), all diluted at 1:500, from BioLegend (unless otherwise indicated). For leukocyte staining, cells were stained with Ly6G-FITC (127606), CD19-PE (115508), CD4-PerCP-eFluor710 (46–0042-80, eBioscience), B220-PE/Cy7 (103222), CD11b-APC (101212), CD8-AF700 (100730), CD90.2-APC/Cy7 (105328), CD115-BV421 (135513), Ly6C-BV605 (128036), CD45 BV711 (103147), all 1:500, from BioLegend (unless otherwise indicated). For BrdU incorporation assays, 1 mg of BrdU was injected i.p. either two or 24 hours before harvesting, depending on the experiment. BrdU staining was then performed using the FITC BrdU flow kit (559619, BD Biosciences) according to the manufacturer’s protocol. Apoptotic cell death in bone marrow single cell suspensions was determined by flow cytometry using APC-labeled annexin V kit (550474, BD Biosciences), according to the manufacturer’s instructions. All data were acquired on an LSRII (BD Biosciences) and analyzed with FlowJo software (FlowJo).

Intravital microscopy.

For serial intravital microscopy of the calvarium, bone marrow B cell precursors (Lineage⁻ CD93⁺) from dsRed reporter mice and mature B cells from the spleen (CD19⁺ IgD⁺) were FACS-sorted using a FACSAria II cell sorter (BD). For B cell progenitor imaging experiments, 10⁶ cells were injected i.v. into non-irradiated recipient C57BL/6 mice one day prior to stroke induction. To visualize mature B cell accumulation in the bone marrow, 5×10⁶ cells were injected i.v. one week prior to surgeries. To highlight bone architecture, OsteoSense® 750EX, a fluorescent bisphosphonate imaging agent (Perkin Elmer), was administered i.v. 24 hours prior to imaging (4 nmol/mouse, PerkinElmer). To outline the vasculature, 15μg CD31-Alexa Fluor 647 (102516, BioLegend) 30 min prior to imaging. In vivo imaging was performed using a confocal microscope (IV100 Olympus). Z-stack images for each location were acquired at 2μm steps, and post-processing was performed using Image J software (NIH).

Statistical analyses.

Statistical analyses were conducted with GraphPad Prism software. Animals were randomly assigned to treatment groups. For flow cytometry, power calculations indicated that it was sufficient to use 8 mice/cohort, estimated to give 85% power to detect a 25% difference. Results are depicted as mean ± standard error of mean. Age-matched animals were randomly assigned to control or treatment groups. The data were tested for normality using the D’Agostino-Pearson normality test. For two group comparisons, statistical significance was assessed by the two-tailed unpaired Student’s t test for normally distributed data. If normal distribution assumption was not valid, statistical significance was evaluated using the two-tailed nonparametric Mann-Whitney test. For comparing more than two groups, ANOVA tests followed by Tukey’s multiple comparisons test were applied. Correlations were calculated by using the Pearson correlation coefficient. P values ≤ 0.05 were considered significant.

RESULTS

Stroke triggers B cell development arrest.

To investigate how cerebral ischemia impacts B cell lineage commitment in the bone marrow, we induced transient middle cerebral artery occlusion (tMCAO) in mice (Figure 1A-B). Using multiparameter flow cytometry analyses of the bone marrow, we profiled B lymphoid progenitors on day three of the subacute phase after either tMCAO or sham operation (Figure 1C). We found that tMCAO severely disrupted the B cell lineage downstream of earliest pre-pro B cell precursors, defined as Lineage⁻ B220^int CD93⁺ IgM⁻ CD19⁻ CD43⁺ CD24^{− 25}. Cerebral ischemia halted B cell development at a stage when full B cell lineage commitment occurs, as evidenced by dramatically reduced numbers and frequencies of pro-B cells (Lin⁻ B220^int CD93⁺ IgM⁻ CD19⁺ CD43⁺ CD24⁺), pre-B cells (Lin⁻ B220^int CD93⁺ IgM⁻ CD19⁺ CD43⁻ CD24⁺ cells), and immature IgM⁺ B cells (Lin⁻ B220^int CD93⁺ IgM⁺ CD19⁺) in femoral bone marrow three days after tMCAO (Figure 1C). To confirm these findings, we adoptively transferred dsRed⁺ Lin⁻ B220^int CD93⁺ bone marrow progenitor cells into non-irradiated recipient mice and serially examined calvarium bone marrow cavities using intravital microscopy. One day following their adoptive transfer and one day prior to stroke, dsRed⁺-developing B cells were readily observed in extravascular bone marrow spaces. However, when revisiting the same hematopoietic cavities three days post-stroke, no dsRed⁺ B cell progenitor cells could be detected (Figure 1D). Interestingly, this striking lack of B cell progenitors downstream of the pre-pro B stage did not occur after myocardial ischemia with or without reperfusion (Online Figure I-II), suggesting the involvement of brain-specific pathways.

Figure 1. Stroke impacts bone marrow B cell lineage.

Experimental stroke was induced by transient middle cerebral artery occlusion in C57BL/6 mice. Mice were sacrificed after three days and compared to sham-operated animals. A, TTC-stained brain sections and quantification of infarct volumes on day 3 after tMCAO (n= 6 per group). B, Gating for bone marrow FACS of B lymphoid progenitor subsets in mice. C, Bar graphs show the number of Pre-pro, Pro-, Pre-, and IgM⁺ immature B cells per femur as well as their frequency among total BM (n= 8-11 per group). D, Serial intravital microscopy of the calvarium of mice transplanted with lineage⁻ B220^int CD93⁺B cell progenitors isolated from dsRed reporter mice (red). Bone structure is outlined in blue by Osteosense-750; endothelial cells are labeled by injection of Pecam-1 and Sca-1 Abs (green). Scale bar, 50 μm. Higher magnification pictures of the corresponding boxes are shown. Scale bar, 25 μm. Data are means ± SEM with individual values for data distribution. *P < 0.05, **P < 0.01, and ***P < 0.001, ns, not significant, Student’s t test.

The observed post-stroke loss of bone marrow B cell progenitors could be a result of either cell mobilization, a block of progenitor differentiation or induction of apoptosis in these cells. Because the spleen is the main site where bone marrow-generated immature B220^int IgM⁺ B cells mature into IgD⁺ mature B cells in steady state²⁶, and inflammatory signals may enhance B cell precursor mobilization to the spleen²⁷, we looked for evidence of B cell trafficking early after stroke. In the first three days after tMCAO, Lin⁻ B220^int CD93⁺ developing B cells gradually declined as mice entered the subacute phase. However, B cell progenitor cells were barely detectable in the blood and spleen (Figure 2A), suggesting that mobilization into the blood did not cause the cells’ decline in the marrow. Enumerating circulating and splenic CD19⁺ B cells revealed severe peripheral lymphopenia and splenic atrophy, consistent with previous reports^5,12 (Figure 2B-C). In addition, and concomitant with fewer early B cells, we also found a significantly enlarged pool of mature B220^high IgD⁺ B cells in the bone marrow following stroke (Figure 3A). Due to absence of dsRed⁺ B cell precursor-derived cells in the bone marrow after stroke (Figure 1D), we speculated that increased mature IgD⁺ B cell numbers could be the result of their accumulation from the periphery rather than from enhanced B cell maturation directly from upstream progenitors within the bone marrow compartment^28,29. Mature B cells had not incorporated BrdU 24 hours after injection, indicating that their increased number does not arise from upstream progenitors’ proliferation (Figure 3B). To confirm this hypothesis, we adoptively transferred dsRed⁺ B lymphocytes isolated from the spleen into wild-type mice prior to stroke induction. Intravital microscopy of the skull revealed a significant accumulation of dsRed⁺ cells in extravascular bone marrow spaces of animals with stroke when compared to sham-operated controls (Figure 3C-D), which was confirmed by flow cytometry analyses (Figure 3E). We next wondered if the decline of B cell production in the bone marrow was the result of the crosstalk between accumulating peripheral mature B cells and their progenitors as previously reported in the context of aging³⁰. However, we found that bone marrow loss of early B cells still occurred in Rag-deficient mice with stroke (Figure 3F). Therefore, our data indicate that in addition to the B lymphopoietic defects, stroke also alters peripheral mature B cell trafficking by triggering their accumulation in the bone marrow.

Figure 2. B cell progenitor loss in the bone marrow does not result in mobilization.

A, Representative FACS staining of early B cell progenitors (lineage⁻ B220^int CD93⁺ cells) in bone marrow, blood, and spleen at indicated time points after tMCAO. B, FACS-based enumeration of CD19⁺ B-lymphocytes in blood and spleen at indicated days after tMCAO induction. C, Bar graph shows total numbers of cells in the spleen over a three-day period after stroke in mice (n= 6 per group). Mean ± s.e.m; *P < 0.05, **P < 0.01, ***P < 0.001. ns, not significant.

Figure 3. Stroke affects mature B cell trafficking.

A, Quantification of mature lineage⁻ B220^high CD93⁻ CD19⁺ IgD⁺ B cells by flow cytometry in bone marrow from mice with sham surgery versus tMCAO (n = 8-11 per group). B, 5-bromodeoxyuridine (BrdU) incorporation by mature IgD⁺ B cells in the bone marrow after stroke (n = 3 per group). C, C57BL/6 mice were transplanted with 5×10⁶ B cells from the spleen of dsRed donor animals one week prior to stroke or sham (n = 3 per group). Intravital microscopy of the calvarium bone marrow cavities was performed three days after surgery. Osteosense-750 highlights the endosteal surface (blue); blood vessels are visualized following in vivo labeling of endothelial cells (green) and dsRed splenocytes are shown in red. Scale bar indicates 50μm. D, Enumeration of dsRed⁺ cells in calvarium marrow. E, Representative dot plots showing accumulation of transplanted mature B220^high IgD⁺ dsRed⁺ in the marrow. F, Loss of bone marrow lineage⁻ B220^high CD93⁺ in Rag-deficient mice after stroke. Mean ± s.e.m. *P < 0.05, **P < 0.01. ns, not significant, Student’s t test.

To investigate whether stroke induces a block in B cell lymphopoiesis, we isolated femoral bone marrow one and two days after either tMCAO or sham surgery, and performed pre-B colony forming assays. The number of pre-B cell colonies decreased drastically by day two after stroke (Figure 4A), suggesting impaired progenitor cell activity. Consistent with lack of pre-B cell colonies in vitro, on day two after stroke and two hours following a BrdU pulse, we found reduced BrdU incorporation and thus reduced pre-B cell proliferation (Figure 4B-C). Corresponding to blocked B cell lymphopoiesis, apoptotic pre-B cells dramatically increased in the bone marrow on day two after tMCAO, as evidenced by Annexin V staining by flow cytometry (Figure 4D). Altogether, these results demonstrate that cerebral ischemia triggers systemic lymphopenia by acting independently on early B cell survival as well as mature B cell trafficking.

Figure 4. Stroke triggers B cell developmental arrest.

A, Stroke was induced in WT C57BL/6 mice which were sacrificed at day 1 and 2, and compared with sham-operated controls. CFU-pre-B colony formation of whole BM cells was assessed at indicated time points from both groups (n = 3-5 per group). B, Mice were subjected to two hours of 5-bromo-2- deoxyuridine (BrdU) pulse on day 2 after stroke induction and sham operation. Representative BM BrdU staining of pre-B cells is shown. C, Bar graph indicates the frequency of BrdU⁺ pre-B cells in BM (n = 3 per group). D, Representative FACS staining of Annexin V expression in BM pre-B cells on day 2 after stroke (red) and sham controls (blue). Bar graph displays the frequency of Annexin V⁺ cells on day 1 and day 2 after stroke versus sham-operated mice (n= 3-5 per group). Data are means ± SEM with individual values for data distribution. *P < 0.05, **P < 0.01, and ***P < 0.001, ns, not significant, Student’s t test.

Toll-like receptor signaling is not essential for stroke-induced B lymphoid progenitor defects.

To determine the mechanism underlying disrupted B cell lymphopoiesis after stroke, we investigated the possible contributions of Toll-like receptor (TLR) signaling, which can affect hematopoietic progenitor cells’ lineage decisions in response to inflammation³¹. Sources for putative TLR activation after stroke include pathogen-associated molecular patterns (PAMPs) derived from disrupted barriers^12,32 and danger-associated molecular patterns (DAMPs) arising from the ischemic brain³³. For instance, the nuclear protein high-mobility group box-1 (HMGB-1) is released from damaged cells in the ischemic brain, instigating inflammatory responses through TLR4 that contribute to the immunosuppressive stroke phenotype^34,35. We thus induced stroke in mice genetically deficient for either TLR4 or the TLR-adaptor protein Myd88 and subsequently profiled B cell progenitors in their bone marrow. We found that neither lack of TLR4- nor Myd88-mediated signaling rescued B lymphopoiesis (Figure 5A-B). Therefore, stroke-induced B cell progenitor loss is independent of these types of TLR-mediated DAMP- or PAMP signals.

Figure 5. Defects in B cell development after stroke are Myd88/TLR4 independent.

Sham and stroke surgeries were performed in C57BL/6 (n = 3-6 per group), Myd88^−/− (n = 3-5 per group), and TLR4^−/− animals (n = 3-9 per group) and mice were sacrificed three days later. A, Representative BM FACS dot plots showing developing B cells (lineage⁻ B220^int CD93⁺). Here, the percentage of CD93⁺ cells in pre-gated lineage⁻ B220⁺ BM cells is indicated in each dot plot. B, Bar graphs show the number per femur, and the frequency of lineage⁻ B220^int CD93⁺ early B cells among total BM cells. Mean ± s.e.m. *P < 0.05, **P < 0.01.

Stimulating the hypothalamic-pituitary-adrenal axis reduces B lymphopoiesis.

Emerging evidence suggests that post-stroke immune responses may be modulated by both the sympathetic nervous system (SNS) and the hypothalamic-pituitary-adrenal axis (HPA)³⁶. Immune cells express receptors sensing both catecholamines and glucocorticoids³⁷. We have recently shown that SNS is implicated in hematopoietic stem and progenitor cell activation²², and SNS inhibition was reported to reverse peripheral lymphopenia¹². We thus tested the hypothesis that SNS inhibition could restore B lymphopoiesis. For this purpose, mice with stroke were treated with either 6-hydroxydopamine (6-OHDA), which disrupts the peripheral SNS completely, or specific ß2- and ß3-adrenergic receptor antagonists. None of these strategies rescued B cell development as evidenced by flow cytometric quantification of early BM Lin⁻ CD93⁺ B cells on day three after stroke (Figure 6A-C). Hence, the SNS may not impact bone marrow B cell development following stroke.

Figure 6. Inhibiting the sympathetic nervous system has no impact on bone marrow B lymphopoiesis defects after stroke.

C57BL/6 mice were treated with either 6-OHDA or β2- and β3 adrenoreceptor blockers, and flow cytometry was used to quantify developing B cells in the bone marrow after stroke. Mice with stroke treated with PBS and sham-operated animals were used as controls. A, Representative FACS staining of lineage⁻ B220^int CD93⁺ early B cells are displayed for each treatment. B, FACS-based enumeration of B cell progenitor cells per femur. C, Frequency of B cell progenitors among total BM cells compared to sham surgeries are shown in the bar graphs (n = 4-11 per group). Mean ± s.e.m. ***P < 0.001. ns, not significant, one-way analysis of variance (ANOVA).

We next tested the hypothesis that the HPA axis influences bone marrow B lymphopoiesis, because pro-inflammatory cytokines in the ischemic brain may contribute to HPA axis activation^38,39. In support of this hypothesis, cytokines including TNF-α, IL-1ß and IL-6 mRNA levels and chemokines in the hypothalamus significantly increased (Figure 7A; Online Figure III) and associated with increased corticotrophin releasing hormone (CRH) expression (Figure 7B). Confirming HPA axis activation, adrenal glands were significantly enlarged in mice with stroke (Figure 7C), and corticosterone was strikingly elevated in the blood following cerebral ischemia (Figure 7D). Correspondingly, we detected increased cortisol levels, as well as peripheral lymphopenia, in patients three days after stroke (Figure 7E-F). Increased cortisol inversely correlated with blood lymphocyte numbers in these patients (Figure 7G). There was no correlation between cortisol levels and number of circulating myeloid cells (data not shown).

Figure 7. Stimulating the HPA axis reduces B lymphopoiesis.

A, qPCR analyses for Tnf-α, Il-1β and Il-6 mRNA in the hypothalamus of sham mice and mice with stroke (n = 3-7 per group). B, Corticotropin-releasing hormone (CRH) mRNA from the hypothalamus of sham and stroke-operated animals by qPCR. C, Weight of adrenal glands on day 3 after stroke, normalized to body weight (n = 6-7 per group). D, Corticosterone in serum from mice with or without stroke, and compared to sham operated animals (n = 4-8 per group). E, Cortisol levels by ELISA in the blood of patients three days after stroke (n= 14-20). F, Blood lymphocytes levels from stroke patients (day 3) versus healthy individuals. G, Correlation of lymphocytes with cortisol concentrations in stroke patients. H, CRH was injected i.p. for 4 consecutive days in C57BL/6 mice and marrow B cells were analyzed by flow cytometry. Mice treated with PBS were used as controls. Representative marrow FACS for early B cells is shown for both groups. I, Enumeration of B cell subsets in the femur from mice treated with PBS and CRH (n= 10-12 per group). J, Mice treated with CRH or PBS were given 5-bromo-2- deoxyuridine (BrdU) i.p. (1 mg). BrdU incorporation in pre B cells was analyzed by flow cytometry 24 hours later (n = 7-8 per group). Data are means ± SEM with individual values for data distribution. *P < 0.05, **P < 0.01, and ***P < 0.001, ns, not significant, Student’s t test.

To investigate whether stimulating the HPA axis affects B lymphopoiesis, we treated naive mice without stroke with CRH for four consecutive days. CRH administration significantly increased serum ACTH (Online Figure IVA) and corticosterone levels (Online Figure IVB). In the bone marrow, early Lin⁻ CD93⁺ B220^int B cell progenitors decreased in frequency (Figure 7H). Although not to the same extent, B cell subset analysis revealed that CRH treatment phenocopied the B cell defects we had observed after stroke, including an unaltered pre-pro B stage (Online Figure IVC) and a significant loss of downstream pro- and pre-B cells (Figure 7I; Online Figure IVC). CRH administration, similar to stroke, led to lower B-cell progenitor proliferation in a BrdU incorporation assay (Figure 7J; Online Figure IVD). This mimicry of stroke-induced changes in B lymphopoiesis by CRH indicates that long-range brain-bone marrow communication via the HPA axis disrupts lymphoid cell output.

Targeted disruption of glucocorticoid signaling in CD19⁺ B cells rescues bone marrow lymphopoiesis after stroke.

We next investigated more specifically how HPA axis activation after stroke affects B lymphopoiesis. High glucocorticoid levels are detrimental to multiple immune cell types^40,41. We therefore hypothesized that glucocorticoids could impair B lymphopoiesis by direct signaling to B lymphoid progenitors. To examine the effects of excess glucocorticoid signaling on B cell progenitors after stroke, we crossed mice bearing the loxP-flanked Nr3c1 allele⁴² with mice carrying the Cre recombinase under the transcriptional control of the CD19 promoter⁴³ to specifically delete the glucocorticoid receptor in the B cell lineage (hereafter referred to as CD19-GR^KO mice). Remarkably, disrupting glucocorticoid signaling restored B cell production after stroke, which was readily observed downstream of the pro-B cell stage (Figure 8A-B). Flow cytometry analysis of the bone marrow revealed significantly increased levels of pre-B cell frequencies and numbers in CD19-GR^KO mice compared to CD19-GR^WT on day three after tMCAO (Figure 8A-B). Interestingly, loss of glucocorticoid signaling on CD19⁺ B cells did not affect B lymphopoiesis in sham-operated mice, suggesting stroke makes a specific contribution to arrested development (Figure 8B). Pre-B colony forming units were markedly increased in CD19-GR^KO mice, further indicating B lymphoid progenitors’ restored ability to proliferate and differentiate (Figure 8C). Notably, targeted disruption of glucocorticoid signaling not only prevented the loss of B cell progenitors but also significantly prevented mature B cell accumulation in the bone marrow (Figure 8D), which resulted in higher numbers in the blood of CD19-GR^KO mice (Figure 8E). Although upstream pro-B cells begin expressing CD19, which drives cre expression and glucocorticoid receptor deletion in CD19-GR^KO mice, we did not observe restoration of the pro-B cell pool (Figure 8B). This could be attributed to insufficient glucocorticoid receptor ablation because reporter gene expression was absent from a significant fraction of pro-B cells of CD19^cre:R26-tdTomato mice (Online Figure V). Taken together, our results unveil a previously unknown mechanism by which brain-bone marrow endocrine signals directly impact B lymphocyte production and trafficking via glucocorticoid receptor signaling.

Figure 8. Deleting glucocorticoid receptor in CD19⁺ B cells prevents stroke-induced lymphopenia.

Stroke was induced in CD19-GR^WT (n = 7; blue bars) and CD19-GR^KO (n = 7 mice; red). Mice were sacrificed on day 3 following stroke. Sham control surgeries were also performed in CD19-GR^WT (n = 5) and CD19-GR^KO mice (n = 4). A, TTC staining and FACS staining of B lymphoid progenitor subsets are shown from mice with tMCAO. Percentages within dot plots indicate frequency among total BM cells. B, Bar graphs show the number per femur (left) and frequency among total BM cells (right) for each subset investigated. C, Proliferation was performed in a CFU-pre-B cell colony forming unit assay in vitro for each cohort of mice. D, Flow cytometry quantification of mature B cells per femur. E, Enumeration of mature B cells circulating in blood by flow cytometry. Data are mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001, ns, not significant, one-way analysis of variance (ANOVA).

DISCUSSION

Cerebral ischemia dramatically alters the hematopoietic tree, resulting in increased myeloid cells and substantially fewer lymphocytes. The precise mechanisms that govern this imbalance after ischemic injury of the brain, however, are not fully understood. Here we describe that B cell development in the bone marrow is severely impaired at the pro-B cell stage three days after transient MCA occlusion and that this impairment contributes to peripheral lymphopenia. This rapid decline of early B cells in the bone marrow is not due to cells mobilizing or relocating to other peripheral organs such as the spleen. Rather, B cell loss results from strikingly reduced B cell progenitor proliferation and differentiation, as well as enhanced apoptosis. These observations in adaptive immune cells clearly contrast to the post-stroke effects on innate immune cell production. Increased sympathetic innervation of bone marrow stromal cells results in exaggerated myeloid cell progenitor proliferation and differentiation²². This direct sympathetic cross talk between the brain and bone marrow occurs in concert with brain-derived signals that govern the severe impairment of lymphopoiesis. Indeed, several stimuli, including long-range signals and microenvironmental cues from the bone marrow stem cell niche, may cooperate to differentially affect hematopoietic stem and progenitor cell maintenance and differentiation into different lineages. To test our hypothesis that brain-derived signals could account for the observed lymphocyte phenotype, we investigated the brain-bone marrow axis and long-range signals involved in regulating B cell production in the bone marrow after stroke.

We first explored whether sympathetic activation and innervation of bone marrow stromal cells could also be responsible for preventing B cell development and, thus, would directly account for imbalanced myelopoiesis and lymphopoiesis post-stroke. The sympathetic nervous system is a major pathophysiological pathway contributing to post-stroke alteration of the leukocyte repertoire and immunosuppression^2,12,44,45. Further, sympathetic regulation of bone marrow stem cell niches has been linked to hematopoietic stem and progenitor (HSPC) cell quiescence and mobilization, as well as lineage decisions favoring myelopoiesis^22,23,46. We previously reported that post-stroke sympathetic nervous signaling increases HSPC activity, instigates a bias towards the myeloid lineage and consequently increases monocyte and neutrophil production²². However, we did not observe any catecholaminergic effect on lymphopoiesis in the experiments described here. This suggests that distinct mechanisms drive lymphoid versus myeloid lineage fate decisions after stroke.

Several brain-derived signals may directly affect hematopoietic cells and subsequently alter leukocyte output after stroke. For instance, ischemic injuries cause dying cells to release DAMPs, such as high mobility box 1 (HMGB1) and heat shock proteins, which instigate inflammatory responses via TLR signals. Indeed, the TLR4 ligand HMGB1, which increases HSPC activity³³, has been detected in the blood after brain injury⁴⁷. TLR4 is highly expressed by HSPCs and increases their proliferation and expansion toward the myeloid lineage³¹. We therefore used TLR4^−/− and Myd88^−/− mice to examine if brain insult-derived alarmins transported through the blood stream could disrupt B cell development. Neither of these gene deletions rescued B lymphopoiesis, suggesting that DAMP-TLR-mediated immune signaling after cerebral ischemia does not influence bone marrow B cell progenitors. This suggests that post-stroke barrier function loss^12,48 does not directly compromise adaptive immunity.

Stroke results in significant stress, which activates both the SNS and the HPA axis, resulting in high systemic levels of the stress hormone glucocorticoid. Although glucocorticoids’ immunomodulatory effects on lymphocytes have been reported previously, including the glucocorticoid receptor dependent apoptosis induction in thymic T cells^40,49, it was unclear whether stroke-induced stimulation of the HPA axis impacts hematopoietic progenitor fate in the bone marrow. To determine if HPA axis activation could cause post-stroke observed B cell losses, we administered CRH to wild-type mice. This pituitary gland stimulation efficiently triggered the production of ACTH and, in turn, glucocorticoids. Strikingly, this phenocopied the bone marrow hematopoietic defect observed after stroke. While prolonged high systemic glucocorticoid levels’ detrimental consequences on immune cells are well documented⁴⁰, the direct effects of glucocorticoid signaling on specific immune cell lineages remain to be explored in vivo. For example, lymphocytes and myeloid cells could be differentially affected by corticosteroids after cerebral ischemia. We therefore addressed, for the first time, whether glucocorticoid signaling could directly influence B cells and their development by generating mice that specifically lack the glucocorticoid receptor in CD19-expressing B cells. Targeted disruption of glucocorticoid signaling in B cells rescued bone marrow B cell development and peripheral B cell numbers in mice after stroke. We thus show that glucocorticoids directly induce apoptosis in B cell progenitors, uncovering the endocrine system’s importance in suppressing lymphopoiesis post-stroke. The intracellular glucocorticoid receptor, upon binding its ligand, translocates to the nucleus where it can directly bind to the DNA, specifically to palindromic sequences that are called glucorticoid response elements⁴⁰. Once bound to DNA, the complex acts as a transcription factor leading to a change in gene expression that can be cell specific and affects both pro- and anti-apoptotic genes, resulting in apoptosis or cell survival depending on the cell type⁴⁹. The precise molecular circuitry downstream of the glucocorticoid receptor that leads to the post-stroke effects on the B-cell lineage remains to be elaborated.

Bone marrow stromal cells may also respond to ischemic brain injury and thereby affect hematopoiesis⁵⁰. They support both myeloid and lymphoid maintenance, differentiation, and mobilization by providing local signals through cell-cell contacts and secreted proteins, including cytokines, chemokines, and growth factors. In light of the bone marrow stem cell niche heterogeneity, our findings call for further investigations into the cellular components that may dysregulate HSPC cell activity after stroke. A number of distinct niche cells residing in the bone marrow participate in regulating hematopoiesis and leukocyte production. These cells, including mesenchymal stromal cells, osteolineage niche cells, endothelial cells and bone marrow macrophages regulate HSPC activity in steady state^51,52. Future studies should investigate how bone marrow stromal cells modulate and mobilize bone marrow hematopoietic cells and their contributions to the post-stroke inflammatory state.

Our data show that the HPA axis is a major contributor to stroke-induced lymphopoiesis arrest. This clearly links the neuroendocrine system to regulation of emergency hematopoiesis after stroke. We show that lymphopenia and sustained elevated cortisol also occur in stroke patients and strongly correlate with each other, which suggests that similar mechanisms lead to immunosuppression in humans. Further, this work indicates that post-stroke humoral immunosuppression can be attenuated by inhibiting the neuroendocrine stress response after stroke. In patients, the neutrophil to lymphocyte ratio is associated with poor long-term outcome⁶. It is thus tempting to speculate that restoring cellular and humoral responses could prevent post-stroke complications and mitigate risk of infections. In addition, studies in mice suggest B cells play a neuroprotective role after stroke, as genetically depleting B cells exacerbates stroke outcome while adoptively transferring B cells limits CNS inflammation, mortality, and neurological deficits¹⁶. Yet, whether or not B cells exert beneficial or detrimental effects in post-stroke recovery is still a matter of debate. Studies in murine models reported minimal improvements or even lack of post-stroke neuroprotection⁵³. Perhaps a lymphopoiesis shutdown is even required to prevent autoreactive B cells triggered autoimmunity to brain-specific antigens. Indeed, such autoreactive B cells have been detected in mice and patients with stroke⁵⁴. Future studies are thus warranted to further delineate the pathophysiology of stroke, which likely involves and affects multiple organ systems and cell lineages. Post-stroke immune system modulations are both complex and dynamic; in this study, we contribute novel insight into how lymphopoiesis is regulated, which is important for a better pathophysiological understanding of cerebral ischemia and could enable targeted treatment strategies for patients with stroke.

NOVELTY AND SIGNIFICANCE.

What Is Known?

• Patients with ischemic stroke often suffer from secondary complications such as pneumonia.

• Ischemic stroke profoundly changes the number and phenotype of circulating white blood cells.

• Most circulating blood cells are made in the bone marrow.

What New Information Does This Article Contribute?

• A map of changes in B lymphopoiesis after ischemic stroke in mice showed that the production of B-lymphocytes is disrupted at the upstream progenitor level.

• Increased glucocorticoid signaling after brain ischemia leads to a B-cell development arrest.

The immune system plays a critical role in the genesis of atherosclerosis, which is the underlying cause of myocardial infarction and ischemic stroke. During stroke, ischemic injury to the brain rapidly activates production of myeloid cells while adaptive immune cells become rare. This decline in lymphocytes compromises resistance to post-stroke infections, such as pneumonia and may also influence the recovery of the brain from ischemia. Here we describe that post-stroke lymphopenia arises across the entire B-cell lineage and includes B-cell progenitors in the bone marrow. Our results suggest that the hypothalamic–pituitary–adrenal axis mediates B lymphopoiesis defects after ischemic stroke via the glucocorticoid receptor, which is widely expressed in hematopoietic cells. Such increased understanding of the hematopoietic response to brain ischemia may provide new avenues to augment stroke recovery.

Nonstandard Abbreviations and Acronyms:

t(tMCA)

ransient middle cerebral artery

(CFU)

colony-forming unit cell

(TLR)

Toll-like receptor

(HMGB-1)

high-mobility group box-1

(6-OHDA)

6-hydroxydopamine

(HPA)

hypothalamic–pituitary–adrenal

(CRH)

Corticotropin releasing hormone

(ACTH)

Adrenocorticotropic hormone

(PAMPs)

pathogen-associated molecular patterns

(DAMPs)

danger-associated molecular patterns

(GR)

glucocorticoid receptor

(SNS)

sympathetic nervous system

(HSPC)

hematopoietic stem and progenitor cells