|Home | About | Journals | Submit | Contact Us | Français|
Hematopoietic stem cells give rise to all blood lineages, and retain a remarkable capacity to proliferate in response to insult. While some controls on HSC activation are known, little is understood about how this process is linked to natural signals. We report that the interferon-inducible GTPase Lrg-47 (Irgm1), previously shown to play a critical role in host defense, inhibits baseline HSC proliferation and is required for a normal HSC response to chemical and infectious stimuli. Our results establish a link between the response to infection and HSC activation, and demonstrate a novel function for a member of the p47GTPase family.
Hematopoietic stem cells are self-renewing bone marrow cells which give rise to all blood lineages, and retain a remarkable capacity to proliferate in response to insult. While some controls on HSC activation are known, little is understood about how this process is linked to natural signals. We report that the interferon-inducible GTPase Lrg-47 (Irgm1), previously shown to play a critical role in host defense, inhibits baseline HSC proliferation and is required for a normal HSC response to chemical and infectious stimuli. Overproliferating Lrg-47 −/− HSCs are severely impaired in functional repopulation assays, and, when challenged with hematopoietic ablation by 5-fluorouracil or infection with Mycobacterium avium, Lrg-47 −/− mice fail to achieve the expected expansion response in stem and progenitor cell populations. Our results establish a link between the response to infection and HSC activation, and demonstrate a novel function for a member of the p47GTPase family.
Hematopoietic stem cells (HSCs) are largely quiescent, self-renewing cells that give rise to all adult blood lineages. They have a remarkable capacity to respond to proliferative stimuli from a variety of insults, exiting their quiescent phase and undergoing a period of self-renewal and differentiation in order to restore hematopoietic homeostasis. These cells possess enormous therapeutic potential, which is currently limited by our inability to direct ex vivo what HSCs achieve so easily in vivo – self-renewal events leading to the expansion of HSC number.
In vivo HSC proliferation can be induced by several well-characterized experimental methods (e.g. chemotherapeutic treatment or bone marrow transplantation), however previous attempts to elucidate the genetic and molecular controls on this process have largely ignored the question of how the regulation of HSC function might be linked to natural signals – such as those emanating from the interplay between infectious agents and the host immune system. Recently, this connection has been indicated by results implicating Toll-like receptors as directing stem and progenitor fates in vitro (Nagai et al., 2006), however a genetic link between immune system signaling and in vivo HSC function has remained elusive.
In a previous study of the dynamics of gene expression changes during a cycle of HSC activation (defined as induced expansion, and eventual return to quiescence; Venezia et al., 2004), we identified genes involved in proliferative vs. quiescent HSC functions. Interestingly, one family of genes identified by this approach encode the p47 GTPases (MacMicking, 2004; Taylor et al., 2004), which have been recently redesignated as immunity-related GTPases (IRG). Characterized as interferon (IFN)-responsive effectors of the immune system, these genes are maximally expressed in HSCs within 1 day following exposure to 5-fluorouracil (5FU), and have been identified by other groups in stem cell gene expression profiles (Ivanova et al., 2002; Ramalho-Santos et al., 2002).
In the past, the p47 GTPases have been characterized solely in the context of the immune system. Indeed, one of these family members, Lrg-47 (Sorace et al., 1995; also Ifi1, Irgm1) has been shown to be required for host defense against a broad range of intracellular pathogens, including Listeria, Toxoplasma, Mycobacteria, and Trypanosoma cruzi (Collazo et al., 2001; Feng et al., 2004; MacMicking et al., 2003; Santiago et al., 2005). In previous in vitro studies, the host resistance defects of Lrg-47 deficient mice have been associated with impaired intracellular microbial killing, phagosome maturation and autophagy (Butcher et al., 2005; MacMicking et al., 2003; Martens et al., 2004). Interestingly, recent work shows that the human ortholog, IRGM, serves a similar immune role (Singh et al., 2006), and further, a single nucleotide polymorphism in this gene has been shown to be correlated with Crohn’s disease (Wellcome 2007; Parkes et al., 2007). In characterizing the response of Lrg-47 −/− mice to several of these pathogens, it was also noted that such animals not only fail to control what is normally a non-lethal infection, but also develop a profound pancytopenia prior to death (Feng et al., 2004; Santiago et al., 2005). This unexpected intersection of findings – upregulation of Lrg-47 in HSCs in response to chemotherapeutic stress, coupled with a major hematopoietic defect in Lrg-47 −/− mice confronted with infectious challenge – led us to test a novel hypothesis, that Lrg-47 may regulate HSC function in the face of either chemical or pathogenic stress.
We first examined the ability of Lrg-47 −/− mice to respond to non-infectious stresses on the hematopoietic system, and we chose two different modalities to address this question. When exposed to a sublethal dose of irradiation, Lrg-47 knockout mice showed impaired recovery of bone marrow and thymic cellularity (Fig. 1a), as well as impaired B lymphopoiesis (Fig. 1b). After 5FU treatment, Lrg-47 −/− mice, although ultimately able to re-establish their baseline blood counts, exhibited substantially delayed hematopoietic recovery (Fig. 1c). Together with the previous findings discussed above, these results suggest a defect at the level of the HSC in mice lacking the Lrg-47 gene.
To examine the functional properties of Lrg-47 −/− HSCs, we performed competitive and non-competitive bone marrow transplantation assays with whole bone marrow obtained from knockout animals (Fig. 2). In competitive repopulation assays, a constant number (250,000) of wild-type competitor cells was admixed with varying amounts of Lrg-47 −/− whole bone marrow. Engraftment analysis by peripheral blood chimerism (CD45.2 vs. CD45.1) at successive 4-week time points post-transplant showed a profound deficit in knockout marrow, as peripheral blood elements derived from Lrg-47 −/− HSCs were progressively lost in favor of wild-type cells. Given a 2:1 advantage over the competitor cells, Lrg-47 −/− marrow achieved on average only 3% engraftment by 4 weeks post-transplant (a 12-fold deficit compared to wild-type), with progressive deterioration of engraftment thereafter (Fig. 2a). More strikingly, a separate experiment showed that even a 25:1 advantage to the knockout marrow failed to produce robust engraftment from Lrg-47 −/− HSCs (Fig. 2b). Interestingly, competitive repopulation assays performed with mice deficient for IGTP (another member of the p47 GTPase family, with a similar immune phenotype to Lrg-47; Taylor et al., 2000) failed to demonstrate an engraftment defect, showing that the requirement for engraftment is not a general property of this gene family (Fig. 2b).
Noncompetitive transplant assays (Fig. 2c) were also performed, with the intent of providing strong selective pressure for Lrg-47-deficient HSCs. Transplantation of 2×105 wild-type bone marrow cells is normally sufficient to rescue and fully repopulate the hematopoietic system of all lethally irradiated recipients, yet this dose of cells from knockout donors failed to rescue any recipients from lethal irradiation. Although a dose of 2×106 knockout cells provided rescue, engraftment was once again poor, with Lrg-47 −/− HSCs ultimately outcompeted by residual host HSCs, indicative of impaired renewal. Importantly, while Lrg-47 −/− HSCs showed poor engraftment, those cells that did engraft satisfy the traditional definition of stem cells – exhibiting multipotentiality (Fig. 2d) and self-renewal (as shown by ability to engraft secondary hosts – Supplementary Fig. S1). In addition to peripheral blood analysis, some recipients from non-competitive experiments were sacrificed, and engraftment analysis in other compartments (bone marrow, spleen, and hematopoietic progenitors) revealed that the peripheral blood engraftment faithfully mirrors engraftment throughout the hematopoietic system (Supplementary Fig. S1). Additional experiments showed that Lrg-47 −/− transplantation defects are not attributable to defects in marrow homing (Supplementary Fig. S2a). Finally, transplantation of wild-type marrow into Lrg-47 −/− recipients demonstrated that this engraftment defect is cell-autonomous, and also showed that knockout recipients were more easily ablated (Supplementary Fig. S2b) – a finding consistent with the above results (Fig. 1a and b).
Despite the striking defects seen in the engraftment ability of Lrg-47 −/− HSCs, at baseline, knockout mice are largely normal, with no conspicuous defects in the absence of immune challenge. We did observe subtle hematopoietic alterations at baseline – including a trend towards reduction in the cellularity of bone marrow and peripheral blood, as well as a shift in the differential counts of peripheral blood leukocytes (Supplementary Fig. S3). In light of this mild hematopoietic phenotype, coupled with the severe functional defect revealed by transplantation challenge, we examined the frequency distribution of HSC, using the side population (SP) method for Hoechst dye efflux (Camargo et al., 2006; Goodell et al., 1996), and found that, while the knockout animals demonstrated a reduction in the relative abundance of long-term HSCs in the bone marrow, this result was not statistically significant in repeated experiments (data not shown); the percentage of the Sca1+, kit+, Lin- (KSL) stem and progenitor cells was comparable to that in wild-type mice. Thus, although Lrg-47 −/− mice exhibit a slight reduction in HSC number, this can not account for the severe functional defects described above.
The delayed recovery of Lrg-47 −/− mice from 5FU treatment suggested a possible proliferation defect in knockout HSCs. To test this idea, we assessed the proliferative status of Lrg-47 −/− HSCs using 5-bromodeoxyuridine (BrdU) incorporation. Following a 3-day or 6-day in vivo exposure to BrdU (Hock et al., 2004), HSCs (Sca1+ SP cells) from wild-type and knockout animals were isolated, and BrdU incorporation was analyzed by flow cytometry (Fig. 3a). Surprisingly, after 3 days of labeling, ~20% of the wild-type HSCs had incorporated BrdU, while the knockout HSCs were substantially more proliferative (up to ~65% BrdU positivity). This relationship persisted over a 6 day BrdU exposure (Fig. 3b). In order to examine this phenomenon in the most primitive HSC compartment, further work took advantage of the fact that SPlow cells have the highest level of LT-HSC activity within the SP (Camargo et al., 2006; Goodell et al., 1997); analysis of the fractionated SP confirmed the increased proliferation of the phenotypically-defined knockout long-term HSC compartment (LT-HSC; Fig. 3c)
These findings show that in the absence of Lrg-47, HSCs are hyperproliferative, implicating Lrg-47 as a key regulator of HSC quiescence, even under homeostatic conditions. Previous studies have shown that actively cycling HSCs engraft poorly (Passegue et al., 2005), consistent with the transplant data presented in Figure 2. Yet, even once engrafted, recipients of Lrg-47 −/− HSCs show progressive loss of chimerism over time, indicating a broader functional deficit than engraftment ability alone.
We next hypothesized that, lacking strict control over HSC proliferation, Lrg-47 knockout mice might be impaired in their response to stress at the HSC level. Thus, we next sought to examine the response of Lrg-47 −/− HSCs to proliferative stimuli, using both chemotherapeutic and infectious challenges. Since 5FU consistently induces expansion of the HSC compartment (Harrison and Lerner, 1991; Randall and Weissman, 1997; Van Zant, 1984), causing the stem cells to enter cycle and proliferate – an activity that peaks on day 6 post-treatment, we treated wild-type and knockout mice with this agent and then studied the SP compartment at several timepoints thereafter. As expected, wild-type mice show a marked increase in HSCs at 6 days post-treatment, while Lrg-47 −/− mice showed a drastically impaired expansion of this compartment, on both a relative and absolute basis (Fig. 4a and b). Analysis at later timepoints reveals that Lrg-47 −/− mice are able to muster a subdued LT-HSC expansion, albeit a delayed one. Interestingly, observation at day 1 post-treatment reveals that, in contrast to the slight losses of HSCs sustained by wild-type mice (~35% of absolute HSC number), Lrg-47 −/− knockouts had severe stem cell losses within 24 hours after 5FU administration (~75%) – a result that agrees with their increased proliferative status and consequent enhanced susceptibility to 5FU, a chemotherapeutic agent which kills cycling cells. Analysis of the KSL compartment of knockout animals revealed that caspase activity is increased at roughly twice normal levels following in vitro treatment with 5FU (Supplementary Fig. S4) –suggesting that the vulnerability to 5FU may be mediated at least in part through apoptotic pathways.
Previous observations of pancytopenia in pathogen-infected Lrg-47 −/− mice suggested that this gene might link HSC regulation to the normal immune response –so that infectious stimuli, rather than nucleotide analogs, would provide the impetus for a program of HSC activation. Thus, we investigated the response of wild-type and Lrg-47 −/− stem and progenitor cells 4 weeks following infection with M. avium. Importantly, we detected a striking (15-fold) expansion of the wild-type Sca1+ kit+ Lin- (KSL) compartment in response to this bacterial agent, with a notable lack of expansion in the knockout animals (Fig. 4c) together with reduced marrow cellularity. A similar result was obtained using CD150 (Kiel et al., 2005) as a marker for stem cells (Supplementary Fig. S5). In vitro colony forming assays provide functional confirmation of this result (Fig. 4d), as whole bone marrow progenitor activity increases after M. avium infection in wild-type, but not knockout, animals. These results strongly implicate the stem/progenitor cell populations as playing a role in host resistance– Lrg-47 −/− mice succumb to infection after failure to maintain normal hematopoiesis – and demonstrate at a genetic level that the immune response to infection is a regulator of stem and progenitor function in vivo.
The data presented here implicate Lrg-47 as an important regulator of HSC proliferation under both steady state and stress conditions. More broadly, consistent with recent results implicating Toll-like receptors as directing stem and progenitor fates in vitro (Nagai et al., 2006), this work establishes a link between immune system signalling and in vivo HSC function, a relationship that has largely proved elusive. How might Lrg-47 function benefit the HSC compartment? Previous work has shown that, even in steady-state conditions, control of HSC proliferation is tied to functional integrity (Hock et al., 2004; Passegue et al., 2005). Further, we know that in the face of stress, strict control of HSC proliferation is needed, as many stimuli that induce HSC proliferation are themselves often cytotoxic. While HSCs must proliferate and expand when called upon, they must first survive the toxicity associated with these insults. Our previous expression profiling of 5FU-activated HSCs revealed that Lrg-47 is maximally upregulated in the first 24 hours following insult, and as shown here, its function is crucial to adequate HSC protection. Our findings thus point to a dual role for Lrg-47 – first as an inhibitor of HSC proliferation at baseline, and second as an induced player in the response to injury.
Yet, remarkably little is known about how the p47GTPases exert their effects. One possible model derives from the observation that IIGP (a p47GTPase) interacts with hook3 (Kaiser et al., 2004; a microtubule associated protein) – suggesting that p47GTPases may regulate cytoskeletal changes and/or cell motility. This is an attractive hypothesis from a stem cell perspective, since the proliferative status of HSC has been shown to be linked to physical location within the endosteal microenvironment (Wilson et al., 2004). Our data also suggest that Lrg-47 may play a role in pathways regulating cell death. Detailed interactions of Lrg-47 in this system remain an open area of study which may illuminate heretofore unknown mechanisms regulating HSC proliferation and fate decisions.
Interestingly, Lrg-47 has been heretofore known chiefly as an effector of IFNγ signalling – a pathway which has previously been tied to hematopoiesis, as IFNγ contributes to the pathogenesis of Fanconi’s (Haneline et al., 1998; Whitney et al., 1996) and aplastic anemias (Dufour et al., 2001; Nistico and Young, 1994), and has been shown in vitro to exert potent inhibitory effects on hematopoiesis (Selleri et al., 1996). Additionally, when subjected to infectious challenge, IFNγ knockout mice display dysregulated expansion of certain hematopoietic progenitors (Murray et al., 1998). Since Lrg-47 expression is strongly upregulated by IFNγ, it is possible that these effects are mediated through the induction of Lrg-47 or functionally related molecules. In addition, the previous observation that Lrg-47 is upregulated in LPS-stimulated macrophages (Sorace et al., 1995) implicates the TLR signaling pathway as a possible contributor to the Lrg-47-dependent effects of infection on HSC activity. Future studies will examine the requirement for IFN as well as TLR signaling for Lrg-47 function in HSC.
The cellular response to infection requires replenishment of hematopoietically-derived effector populations and consequently provides proliferative stress on bone marrow progenitors. The data presented here demonstrate that a gene previously shown to regulate host resistance to pathogens also operates at the level of the HSC response, and thereby identifies a possible step in its host protective function. More generally, our findings reveal an unexpected intersection point between the fields of stem cell biology and the immune response to pathogens, and suggest that defects in HSC function should be considered as possible determinants of host susceptibility to infection.
Wild-type C57Bl/6 (CD45.2) and C57B/6.SJL (CD45.1) were obtained from Taconic Farms (Germantown, NY) or were bred at the animal care facility at the Baylor College of Medicine (Houston, TX). IGTP−/− and Lrg-47 −/− mice on a mixed C57Bl/6J x 129 background were generated as described (Collazo et al., 2001; Taylor et al., 2000) and were backcrossed 10 times to C57Bl/6 mice. Knockout mice on inbred B6 backgrounds were used for all transplant and infection experiments. Due to poor breeding efficiency of fully backcrossed animals, mice on mixed backgrounds (backcrossed thrice to B6) were used in the 5FU and BrdU experiments. Following the mating of a single mixed background KO male (129 backcrossed twice to C57Bl/6) with a fully inbred WT C57Bl/6 female, F2 progeny of this cross were used to established true breeding WT and KO lines (backcrossed three times to C57Bl/6). From these lines, age and gender matched animals were used for all experiments, and key reported results from mixed background experiments have been verified using inbred B6 mice. All mice were maintained at an AALAC-accredited, specific-pathogen free animal facility at the NIAID, NIH (Bethesda, MD) or the Baylor College of Medicine (Houston, TX). Gender-matched mice of both sexes between 8 and 12 weeks of age were employed.
MoFlo, LSRII, and FACS-Scan flow cytometers were used for analysis and sorting. Whole bone marrow was stained with Hoechst 33342 for resolution of the SP population as previously described(Goodell et al., 1996; Goodell et al., 1997), and magnetically enriched for Sca1+ cells (autoMACS; Miltenyi Biotec) when appropriate for sorting. Staining for cell surface markers was performed as previously described(Venezia et al., 2004) – briefly: lineage negative cells were identified by a lineage antibody cocktail (Pharmingen or eBiosciences) consisting of CD4, CD8, B220, GR1, CD11b, and Ter119. For KSL and CD150 analysis, anti-c-kit, anti-Sca1, and anti-CD150 (BioLegend; TC15-12F12.2) antibodies were used in conjunction with the lineage cocktail. For experiments involving 5FU treated animals, CD11b was excluded from the lineage cocktail, as HSCs upregulate CD11b in response to 5FU (Randall and Weissman, 1997).
Non-competitive bone marrow transplants were performed by retrorbital intravenous injection of CD45.2 donor whole bone marrow cells from either 8–12 week old Lrg-47 −/− mice or wild-type controls (C57Bl6) into CD45.1 wild-type C57Bl6 recipients that had been lethally irradiated with a split dose of 10.5 Gy. For competitive transplants, dilutions of CD45.2 donor whole bone marrow cells from wild-type and Lrg-47 −/− mouse strains were admixed with wild-type CD45.1 competitor prior to injection to CD45.1 recipients, as above.
Mice were given intraperitoneal injections of 5FU (American Pharmaceutical Partners) at 150 mg/kg in PBS. For CBCs, 63ul of blood was collected by retro-orbital bleed and diluted 1:2 with CellDyn Diluent prior to analysis on a CellDyn 3500.
Mice were exposed to 600 cGy of whole body irradiation and tissue cellularities examined 4 weeks thereafter. For mycobacterial infection, mice were injected intravenously with 1×106 colony forming units of M. avium (strain SmT 2151) as previously described(Feng et al., 2004), and expansion of KSL and CD150 populations examined 4 weeks later.
Mice received an initial intraperitoneal injection of BrdU (Sigma, 1mg per 6 g of mouse weight), followed by inclusion of BrdU in drinking water (Sigma, 1 mg/ml) for either 3(Hock et al., 2004) or 6 subsequent days. Mice were then sacrificed, and Sca-1+ SP (or Sca-1+ SP low vs. high) cells were sorted into a carrier population of 500,000 B220+ spleen cells. Samples were prepared for analysis of BrdU incorporation using the FITC BrdU Flow Kit (BD PharMingen), and samples were re-analyzed by flow cytometry. Upon re-analysis, HSC’s were readily distinguishable from carrier cells, as Sca-1+ and B220-, and analyzed for BrdU incorporation.
Whole bone marrow from naïve and infected animals was plated in triplicate in MethoCult GF M3434 (StemCell Technologies, Inc.) and incubated at 37°C, 5%CO2. Colony formation was scored on the 12th day after plating.
Student’s T-test was used for statistical analyses, where appropriate. P values are reported in the figure legends, and significance is indicated on the figures using the following convention: * p<.05, ** p<.01, *** p<.001. Error bars in all panels represent the SEM.
The authors would like to thank John Gilbert and Jonathan Berg for critical reading, as well as Wendy Zheng and Sara Hieny for invaluable support in maintaining animal colonies. This work was supported by NIH grants DK58192 (to MAG) and DK63588 (to G. Darlington), as well as the Intramural Research Program of the NIH, NIAID, and an NIAID grant to GAT (AI57831). DCW was supported in part by NIH grant 5 T32 DK64717-02 as well as T32GM008307 from the National Institute of General Medical Sciences. MAG is a Stohlman Scholar of the Leukemia and Lymphoma Society.
The authors declare no competing financial interests
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.