Stromal-derived IL-6 alters the balance of myeloerythroid progenitors during Toxoplasma gondii infection

BMSCs are critical regulators of the hematopoietic response to inflammation through secretion of IL-6.

Abstract

Inflammation alters hematopoiesis, often by decreasing erythropoiesis and enhancing myeloid output. The mechanisms behind these changes and how the BM stroma contributes to this process are active areas of research. In this study, we examine these questions in the setting of murine Toxoplasma gondii infection. Our data reveal that infection alters early myeloerythroid differentiation, blocking erythroid development beyond the Pre MegE stage, while expanding the GMP population. IL-6 was found to be a critical mediator of these differences, independent of hepcidin-induced iron restriction. Comparing the BM with the spleen showed that the hematopoietic response was driven by the local microenvironment, and BM chimeras demonstrated that radioresistant cells were the relevant source of IL-6 in vivo. Finally, direct ex vivo sorting revealed that VCAM⁺CD146^lo BM stromal fibroblasts significantly increase IL-6 secretion after infection. These data suggest that BMSCs regulate the hematopoietic changes during inflammation via IL-6.

Introduction

Inflammation can dramatically alter BM hematopoiesis. In many cases, production of erythroid cells is restricted, whereas myelopoiesis is proportionally enhanced. This myeloerythroid shift occurs in multiple models of inflammation, from bacterial infection to sterile immunization [1–4], and is one pathway that leads to the common condition of anemia in the setting of inflammation.

Inflammatory cytokines play a central role in the blockade of red cell development. Relatively recent literature has established a critical function for IL-6, which can act through hepcidin to reduce iron availability and restrict erythropoiesis [5–9]. IL-6 can also block RBC development directly in vitro [10]. Whether IL-6 limits erythropoiesis independently of hepcidin in vivo remains an important question, as therapies specifically neutralizing IL-6 and hepcidin are being developed for use in clinical settings [11, 12]. In addition, past research has indicated important roles for other inflammatory cytokines, including IFN-γ, TNF, and IL-1, in regards to blocking the maturation and proliferation of early erythroid precursors [8]. Their in vivo role relative to the recent elucidation of the IL-6 hepcidin axis is unclear [6, 11].

The significance of the microenvironment to hematopoiesis is widely acknowledged. An important population of reticular marrow fibroblasts, known as BMSCs or mesenchymal stem cells, was first identified in the 1960s. They interact extensively with hematopoietic progenitors and are able to create new bone containing the hematopoietic microenvironment upon in vivo transplantation [13–16]. Their influence on hematopoietic development during inflammation, however, remains largely a black box. In this study, we sought to examine these questions using the T. gondii infection model.

MATERIALS AND METHODS

Mice

C57BL/6 mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA) or Taconic Farms (Rockville, MD, USA). IL-6 KO mice were purchased from The Jackson Laboratory. B6.SJL, IFN-γ KO, iNOS KO, IL-1R KO, IFN-α/βR KO, TNFR superfamily 1a KO, IL-15 KO, and TCR-α KO mice were obtained through Taconic Farms or the NIAID Taconic Exchange. All mice were maintained at Association for Assessment and Accreditation of Laboratory Animal Care-accredited animal facilities, and experiments were performed in accordance with procedures outlined in the Guide for the Care and Use of Laboratory Animals under animal study proposals approved by the NIAID and NIDCR Animal Care and Use Committees. Mice were used between 6 and 12 weeks of age.

Parasite and infection

Infections were performed by orally gavaging brain-derived cysts of a RFP protein-expressing C1 clone of the ME-49 T. gondii strain [17]. Parasite was maintained in C57BL/6 mice through serial passage. Cysts were prepared by homogenizing brains of mice infected 1–3 months prior.

Cell harvest

Murine femoral and tibial BM cells were isolated by flushing the long bones with RPMI. Vertebral BM cells were isolated by crushing two lumbar vertebral bodies in a small petri dish containing RPMI. Spleen cells were isolated by mashing dissected spleens through a 70-μm filter into 50 mL conical tubes.

Ex vivo BMSC isolation

BM was digested for 30 min at 37°C in 1 mL RPMI containing 0.5 mg/mL collagenase with 1 mg/mL dispase (Invitrogen, Carlsbad, CA, USA) and 0.5 mg/mL DNase (Sigma-Aldrich, St. Louis, MO, USA) on a rotating platform. Cells were put through a 70-μm filter and stained with CD45 biotin (30-F11) and Ter119 biotin, followed by streptavidin microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany). In some cases, CD45-PE, followed by anti-PE microbeads, was used (Miltenyi Biotec). Depletion was done using LS columns (one column/one to two hindlimbs). Cells were then incubated with streptavidin FITC and antibodies against VCAM, CD146, PDGFRα, ScaI, and CD31. Restaining with CD45 FITC and Ter119 FITC was sometimes used instead of streptavidin FITC. Sorting was performed on a BD FACSAria II by the NIAID Flow Cytometry Core.

Ex vivo BMSC fibroblast colony formation and culture

For colony-forming efficiency assays, cells were cultured in T25 flasks using α-MEM with 20% lot-selected FBS, 2 mM glutamine, 100 U/mL penicillin + 100 μg/mL streptomycin, 0.1 mM β-ME, and 10 million guinea pig BM feeder cells, irradiated at 6000 rads [18]. Colonies were counted by dissecting microscope. For overnight cultures, BMSCs were isolated from two to three mice/group. Cells (30,000–50,000) were plated in one well of a flat-bottom, 96-well plate in 200 μl 10% RPMI. For the endothelial population, all sorted cells were plated, as typical yields were lower than 20,000. ELISA for IL-6 was performed 18 h later and values normalized per 10,000 cells.

Statistical analysis

Statistics were calculated using Prism software (GraphPad Software, La Jolla, CA, USA). Comparisons were evaluated using paired or unpaired Student's t test.

Online Supplemental material

Supplemental Figs. 1–4, as described in the text, and additional Materials and Methods are available online.

RESULTS AND DISCUSSION

T. gondii infection alters hematopoiesis and blocks erythropoietic development beyond the Pre MegE stage

Mice infected orally with T. gondii, a natural rodent pathogen, develop weight loss (Supplemental Fig. 1A) and severe illness, driven by a Th1-biased cytokine storm [19–22]. Previous studies have also demonstrated that infected mice become anemic secondary to IFN-γ-induced hemophagocytosis and that increased myeloid cell production plays a significant, protective role against this pathogen [23–25]. Thus, T. gondii infection allowed us to study how the BM responds to a physiologic and complex inflammatory stimulus.

We found that infected mice showed a >50% loss in total BM cellularity by Day 8 postinfection (Fig. 1A). This was accompanied by a relative increase in myelopoiesis and markedly restricted erythropoiesis (Fig. 1B). Consistent with previous literature [23], analysis of the myeloid changes showed a preferential expansion of the Ly6C^hiLy6G^- “monocytic” population with relatively little change in the Ly6C^loLy6G⁺ “granulocytic” population (Supplemental Fig. 1D) [26–28]. The lymphoid fraction of the BM was also reduced, mostly as a result of decreased B cell production (Supplemental Fig. 1B). As T. gondii spreads systemically, it was possible that these changes were caused by direct infection of BM cells by the parasite. Thus, the BM was assessed for the presence of intracellular T. gondii (made possible, as the strain used expresses RFP). However, only one to two cells/100,000 contained T. gondii (Supplemental Fig. 1C).

Figure 1. T. gondii infection alters hematopoiesis and blocks erythropoietic development beyond the Pre MegE stage.

(A) At the indicated time-points after infection with T. gondii, BM was isolated from one femur and tibia. The percentage of nucleated cells was determined by flow cytometry. Results are representative of two independent experiments with two to three mice/time-point. Error bars represent sds. (B) Nucleated erythroid (Ter119⁺) and myeloid (Ly6C⁺) cells in the BM of naïve versus Day 8- to 10-infected mice; ***P < 0.001. Data points represent individual mice. Results are pooled from greater than three independent experiments. (C) Total BM cells from naïve or infected mice were examined for erythroid lineage precursors. I, Proerythroblasts; II, basophilic + polychromatic erythroblasts; III, orthochromatic erythroblasts and reticulocytes. (D) Percentages of early erythroid precursors among non-GMP lin⁻ cKit⁺ cells and GMPs among lin⁻ cKit⁺ cells from naïve versus infected mice (CFU-E, CFU-erythroid); **P < 0.01; ***P < 0.001. Results are pooled from at least three independent experiments. All error bars represent sds.

We next sought to determine the level at which erythroid development was first impaired. Analysis of erythroblast populations revealed decreases in all stages from proerythroblasts to orthochromatic erythroblasts, suggesting that the block in RBC development occurs earlier (Fig. 1C). Thus, primitive myeloerythroid precursors were analyzed similarly to the methods developed by Pronk et al. [29] (Supplemental Fig. 2A), who characterized the earliest-known erythroid-committed progenitor, the Pre MegE. ScaI positivity was not used for exclusion, as it was up-regulated on most lin⁻/cKit⁺ cells after infection. Our data reveal that erythropoiesis was blocked from progressing beyond the Pre MegE stage, as the fraction of Pre MegE cells stays constant, whereas its progeny is significantly reduced after infection (Fig. 1D). T. gondii also caused a doubling of the percentage of GMPs, consistent with the shift away from erythropoiesis toward myelopoiesis. These results show that infection altered primitive myeloerythroid development and prompted us to look for inflammatory mediators of these changes.

Critical role for IL-6 in controlling the shift of early myeloerythroid progenitors

Many cytokines, previously shown to block erythropoiesis [6, 8], such as IFN-γ, TNF, IL-1, and IL-6, are also prominent players in T. gondii infection [19–23]. We tested the in vivo significance of these and other inflammatory mediators to the observed erythroid block in our model using KO mice. RBC development was similar among all groups before infection (data not shown). Somewhat unexpectedly, blocking most of the tested pathways did not increase erythropoiesis after infection, although analysis of the serum confirmed a systemic Th1 response (Fig. 2A and B). IFN-γ-induced RBC loss is the primary driver of anemia in this model [24, 25], but these data show that IFN-γ is not essential to the reduction in erythropoiesis.

Figure 2. Critical role for IL-6 in controlling the shift of early myeloerythroid progenitors.

Various strains of mice were infected with T. gondii. (A) Eight days after infection, the percentage of Ter119⁺ cells was examined among nucleated BM cells; ***P < 0.001 Data points represent individual mice. Results are pooled from six independent experiments. (B) Serum cytokines and iron levels were determined in naïve and Day 8-infected mice. Results are pooled from two independent experiments. ND, Not detected. (C) Early erythroid precursors and GMPs in WT versus IL-6 KO mice after infection. Results are pooled from three independent experiments (n=8). All error bars represent sds; *P < 0.05; **P < 0.01; ***P < 0.001.

The IL-6 KO was the only tested strain to partially reverse the fall in erythroid precursors after infection (Fig. 2A), although they did not have higher hematocrit levels (Supplemental Fig. 2C). This indicates the RBC loss after T. gondii infection occurs too rapidly to be compensated for by the increased erythropoiesis in IL-6 KOs. However, anemia of inflammation in human disease is often caused by a combination of increased RBC loss and decreased erythropoiesis [8]. This condition is associated with a worse prognosis for patients, and methods to increase RBC production in the setting of inflammatory illnesses are an active area of investigation [8]. Thus, we were interested in the mechanism underlying the erythropoietic difference between WT and IL-6 KO mice.

Existing literature proposes a central role for hepcidin in blocking RBC development during inflammation [8, 11]. Hepcidin is produced by hepatocytes in response to IL-6, and increased hepcidin during inflammation acts via multiple pathways to limit iron availability for new heme synthesis. Thus, hypoferremia is a defining feature of the IL-6 hepcidin pathway [11]. T. gondii infection did not cause hypoferremia (Fig. 2B), although an increase in total iron-binding capacity modestly decreased transferrin saturations (Supplemental Fig. 2B). This indicated that IL-6 was not acting through hepcidin-mediated iron restriction. In addition, serum IL-6 levels were low compared with previous studies on the IL-6 hepcidin pathway [30], where systemic concentrations measured one order of magnitude greater or more.

IL-6 KOs displayed accelerated erythropoiesis compared with WT mice after infection, evidenced by fewer Pre MegE cells, but increased proportions of their downstream progeny (Fig. 2C). This demonstrates that IL-6 is essential to the early erythropoietic block seen at the Pre MegE stage. It further supports a hepcidin and iron-independent mechanism for IL-6, as the Pre MegE stage occurs before transferrin receptor up-regulation [29]. Along with the differences in erythropoiesis, removing IL-6 significantly blunted the expansion of GMPs (Fig. 2C). Thus, IL-6 is an important regulator of hematopoiesis after T. gondii infection through its action on early myeloerythroid precursors. This conclusion is supported by previous literature showing that multipotent progenitors express IL-6 receptor and can respond to IL-6 via STAT3 phosphorylation [31].

Although the IL-6 hepcidin axis is a physiologically important regulator of RBC production, the hepcidin-independent function of IL-6, seen in our study, is not unprecedented and has been hinted at in human data. For example, neutralizing IL-6 improves hemoglobin levels in patients with multicentric Castleman's disease, even in those who continued to have elevated hepcidin levels [32]. A recent study of Hodgkin's lymphoma patients also concluded that IL-6 was functioning via an alternative mechanism, as anemia correlated with serum IL-6 but not hepcidin [33]. Thus, the use of IL-6 and hepcidin-specific therapies [11, 12] may affect hematopoiesis via different pathways and play nonoverlapping roles in treating anemia of inflammation.

Myeloerythroid response after infection is tissue-specific

Given that IL-6 regulates myeloerythroid development after infection and the significant increase in systemically circulating IL-6 levels, one might expect to observe similar hematopoietic changes outside of the femur and tibia. To test this hypothesis, we compared erythropoiesis in the long bones of the leg versus the vertebral bodies or the spleen. Our data show that erythropoiesis is similarly inhibited whether it occurs in the marrow of the long bones or the vertebrae (Fig. 3A).

Figure 3. Myeloerythroid response after infection is tissue-specific.

(A and B) The percentage of Ter119⁺ nucleated cells was examined in naïve or Day 8-infected mice from the tissues indicated. (A) Marrow of long bones (femur and tibia) versus vertebrae (n=3). (B) Marrow of long bones versus the spleen. Data pooled from two independent experiments (n=4–6). (C) Spleen cells were isolated from naïve or Day 8-infected mice, and early myeloerythroid progenitors were assessed by flow cytometry. GMPs were gated on lin⁻ cKit⁺ cells. Pre MegE, Pre CFU-E, and CFU-E cells were examined among lin⁻ cKit⁺ non-GMP progenitors. Data pooled from at least three independent experiments (n=5–9). All error bars represent sds; **P < 0.01; ***P < 0.001.

The spleen, however, exhibited much lower rates of erythropoiesis at steady-state, which did not decrease after infection (Fig. 3B). In fact, assessment of the early splenic myeloerythroid progenitors showed that T. gondii caused accelerated erythropoiesis and decreased GMP percentages (Fig. 3C). These splenic progenitor changes are opposite compared with those of the femur and tibia. Thus, the hematopoietic response to infection is compartmentalized between different tissues. This implies that the myeloerythroid shift in the BM is a result of the marrow microenvironment and not just systemically circulating cytokine.

Radioresistant cells are the key source of erythropoiesis-inhibiting IL-6

We next wanted to determine the relevant in vivo sources of IL-6 in the BM. Myeloid cells, for example, can produce large quantities of IL-6 and have been proposed initiators of the IL-6 hepcidin response [5, 8]. To examine the role of hematopoietic versus nonhematopoietic IL-6, we made BM chimeras with WT versus IL-6 KO donors or WT versus IL-6 KO hosts (Fig. 4A). In the initial setup experiments to determine the composition of radioresistant hematopoietic cells using our irradiation protocol, we found that the total percentage of remaining host cells averaged 1.2% in the BM (Supplemental Fig. 2D). The majority of these was TCR-β⁺ lymphocytes (data not shown). Given the expected importance of IL-6 from myeloid cells, we also looked for radioresistant CD11b⁺ cells and found that they made up 1/60 of the host cells or 0.02% of the BM (Supplemental Fig. 2D). As total CD11b⁺ cells in the BM averaged 56% in these mice, donor CD11b⁺ cells outnumbered host CD11b⁺ cells by ∼2800:1. Thus, whereas host myeloid cells do remain in our chimeras, the overwhelming majority of the myeloid compartment is replaced by the donor. Also of note, no differences in engraftment efficiency or lineage reconstitution at steady-state were seen in any of the chimeric mice in our studies (Supplemental Fig. 3A).

Figure 4. Radioresistant cells are the key source of erythropoiesis-inhibiting IL-6.

(A) WT or IL-6 KO mice were lethally irradiated and reconstituted with WT or IL-6 KO BM. Chimerism was examined in the blood after 6 weeks. Mice were then infected with T. gondii, and erythropoiesis was assessed in the BM. A representative erythroblast profile of total BM cells from each group is shown. Data are representative of two independent experiments. (B) Early erythropoietic progenitors were examined in WT and IL-6 KO host chimeric mice after infection. The percentage of Pre MegE, Pre CFU-E, and CFU-E cells in the lin⁻ cKit⁺ non-GMP population was assessed. Data points are pooled from two independent experiments (n=6); **P < 0.01; ***P < 0.001. Error bars represent sds.

To test the importance of hematopoietic IL-6 secretion, we compared WT with IL-6 KO donor chimeras. Somewhat surprisingly, there was no difference in erythropoiesis between mice reconstituted with WT or IL-6 KO hematopoietic cells in response to infection [Fig. 4A, (1) vs. (2), and Supplemental Fig. 3B]. This demonstrated that radiosensitive hematopoietic cells are not the principal source of erythropoiesis inhibiting IL-6.

To examine whether radioresistant cells are the relevant in vivo producers, we compared WT and IL-6 KO host chimeras. IL-6 KO host chimeras contained significantly more nucleated erythroid cells and higher erythroblast percentages than WT host chimeras after infection [Fig. 4A, (1) vs. (3), and Supplemental Fig. 3C]. Examination of early myeloerythroid progenitors revealed that IL-6-deficient host chimeras had decreased GMPs and accelerated maturation of early RBC precursors, mimicking the characteristics of total IL-6 KO mice (Fig. 4B and Supplemental Fig. 3E).

Taken together, the BM chimera data show that IL-6, in radioresistant cells, is sufficient and necessary to produce the myeloerythroid shift after infection. In our IL-6 KO chimeras, radioresistant hematopoietic cells averaged between 1% and 2% in the BM and were mostly composed of T lymphocytes (Supplemental Fig. 2E), consistent with our initial setup experiments. Given that TCR-α-deficient mice did not differ from WT mice in terms of erythropoiesis after infection (Fig. 2A), radioresistant host T cells were unlikely to be functionally important. And whereas we cannot formally rule out a role for the small population of radioresistant host macrophages, we feel that the absence of an effect, despite removing IL-6 in the overwhelming majority of myeloid cells suggests another source.

We noted that serum levels of IL-6 were reduced when IL-6 was removed from host but not donor cells (Supplemental Fig. 3D and E). This is likely as a result of the fact that large radioresistant organs, such as the liver and intestine, may contribute significant amounts of IL-6 to the systemic circulation (Supplemental Fig. 3F). Thus, whereas our data show that the inhibition of erythropoiesis by IL-6 is a local effect within the BM microenvironment, the marrow is less likely to be a significant, quantitative source of systemically circulating IL-6.

This suggested a couple of explanations. It was possible that IL-6 from other organs in the context of an unspecified characteristic of the marrow microenvironment was driving the myeloerythroid changes. It was also possible that the marrow stroma itself was producing the hematopoiesis-altering IL-6. As the stroma is a relatively minor population, this could explain why no change in total IL-6 mRNA in whole BM was observed after infection (Supplemental Fig. 3F). To better investigate this second hypothesis, we needed a method to effectively isolate BMSCs.

BMSCs increase IL-6 production after T. gondii infection

The BM stroma consists of a variety of radioresistant cells, including vascular endothelium, osteoblasts, adipocytes, and others. However, the reticular fibroblasts, known as BMSCs (also known as mesenchymal stem cells), are unique in their ability to recreate hematopoietic bone after implantation in vivo [13–16]. Initially isolated in the 1960s and distinguished by their ability to form fibroblastic colonies in vitro, they interact extensively with hematopoietic precursors and have been shown to secrete IL-6 after long-term culture [34, 35]. BMSCs are therefore an ideal candidate for producing the IL-6 that regulates early myeloerythroid progenitor differentiation.

Previous methods to study murine BMSCs often relied on weeks of in vitro growth to eliminate contaminating macrophages. We hypothesized that extensive culture time would mask relevant biologic differences between BMSCs from naïve and infected mice. Thus, a method to directly sort these cells was developed (Fig. 5A). Briefly, BM was digested with collagenase/dispase/DNase, and hematopoietic cells were magnetically depleted. Based on previous literature [16, 36], the remaining cells were labeled with stromal markers VCAM and CD146, and cells from the resulting four groups were sorted and plated to determine their fibroblast colony-forming efficiency. The VCAM⁺CD146^lo population was more than one order of magnitude efficient in establishing colonies than the others (Fig. 5B). In fact, >90% of the colonies arose from this fraction (Fig. 5C), indicating that the vast majority of clonogenic stromal cells reside here.

Figure 5. BMSCs increase IL-6 production after T. gondii infection.

(A) Schematic of BMSC isolation. FACS plots show magnetic depletion and sorting profile of CD146 and VCAM stromal markers in remaining nonhematopoietic cells. SS, Side scatter. (B) Fibroblast colony-forming efficiency of the four sorted populations. Data are pooled from two independent experiments. (C) Percentage of total colonies derived from the four sorted populations. Representative of two independent experiments. (D and E) BMSCs were sorted from naïve or infected mice and plated overnight. Supernatants were assessed for IL-6 by ELISA, and cytokine levels were normalized/10,000 cells. Each data point represents an independent experiment; ***P < 0.001. Error bars represent sds.

Surface staining with other markers revealed that the VCAM⁺CD146^lo BMSC population was CD31⁻, whereas the VCAM⁺CD146^hi population was CD31⁺ and thus, composed of endothelial cells (Supplemental Fig. 4A). This differs slightly from human data, where CD146 can be used to sort BMSCs [16]. Two previous studies, including one published while this work was under review, showed that PDGFRα can also be used to sort BMSCs ex vivo [37, 38]. The connection between freshly isolated VCAM⁺ or PDGFRα⁺ BMSCs is not established in the literature, so we costained with both markers and found that all VCAM⁺ BMSCs are also PDGFRα⁺ (Supplemental Fig. 4A). A notable difference between our data and one of the previous studies [37] is that we find most fibroblastic colonies arise from the ScaI⁻ fraction instead of the ScaI⁺ population. The ScaI⁺ cells in our hands expressed high levels of CD146 and were thus, mostly endothelial cells (Supplemental Fig. 4A). However, ScaI expression can be variable, as evidenced by its up-regulation in our VCAM⁺CD146^lo BMSCs after T. gondii infection (Supplemental Fig. 4B). Regardless, we believe that the use of VCAM and PDGFRα will aid future studies of BMSCs.

With the use of the above protocol, VCAM⁺CD146^lo stromal cells were sorted from naïve and T. gondii-infected mice. Total sorted numbers were not different after infection (Fig. 5D), suggesting that BMSCs do not undergo significant hypoplasia, unlike the BM hematopoietic compartment. The freshly isolated BMSCs were plated overnight, and analysis of the supernatant revealed that they produce IL-6 directly ex vivo. Importantly, they secrete more than twice the amount when taken from infected mice (Fig. 5E). This demonstrates, for the first time, that IL-6 up-regulation is an endogenous response of the fibroblastic marrow microenvironment after infection.

The combination of data, showing a localized effect in the marrow, chimera experiments pointing to a radioresistant population, and functional demonstration of increased IL-6 production by BMSCs, suggests that marrow stromal IL-6 regulates hematopoiesis during infection. However, to the authors' knowledge, BMSC-specific IL-6 KO mice do not exist so the definitive experiment cannot be performed. It remains possible that increased systemic IL-6 in the context of some other aspect of the marrow microenvironment produces the hematopoietic alterations.

In these experiments, VCAM⁺CD146^hi endothelial cells were also included for comparison as a separate nonhematopoietic control population. VCAM⁺CD146^hi cells produced significantly less IL-6 and did not increase their production after infection (Supplemental Fig. 4C), highlighting the distinctive biology of clonogenic BMSCs. Data comparing BMSCs from naïve and infected mice also showed that IL-6 secretion tends to equalize over time in culture (Supplemental Fig. 4D). This validated our initial concerns and highlights the need to avoid extensive in vitro growth when studying the endogenous behavior of these cells.

It is important to mention two notable and recently published papers were the first to address the in vivo function of murine BMSCs, taking advantage of nestin and monocyte chemotactic protein-1 fluorescent reporter mice [15, 39]. These studies provided significant in vivo insight into BMSC biology, demonstrating that they play a critical role in proper homing and engraftment of hematopoietic stem cells and also regulate monocyte trafficking in response to circulating TLR ligands, which after T. gondii infection, may trigger the increased IL-6 production seen in our experiments, but this remains an avenue for future investigation. BMSCs might also respond to other molecules produced in an inflammatory setting [40], including neurotrophins and cytokines not tested here.

How inflammation alters hematopoiesis and the role of the marrow microenvironment in this process remain classic questions in hematology. Our study sheds light on one pathway, where stromal IL-6, potentially from BMSCs, regulates early myeloerythroid differentiation by limiting RBC development beyond the Pre MegE stage and expanding GMPs. As the field progresses, increasingly sophisticated methods to study BMSC function, including the use of culture-independent isolation methods, are likely to open new avenues of investigation. Given the prevalence of hematopoietic changes in human disease, further research into the interplay between BMSCs and hematopoietic progenitors may benefit patients with illnesses ranging from cancer to chronic infection.

AUTHORSHIP

D.B.C. and B.S. designed the research, performed experiments, analyzed results, and prepared the manuscript. N.B. and A.M.U. helped perform experiments and prepare the manuscript. M.G. helped design experiments and prepare the manuscript. C.N.R. helped design and perform experiments and prepare the manuscript. P.G.R. and Y.B. designed and oversaw the research and helped prepare the manuscript.

DISCLOSURES

The authors declare no competing financial interests.