Authors: Masahiro Marshall Nakagawa, Harry Davis, Chozha Vendan Rathinam
Categories: Article
Source: Stem cell research
Authors: Masahiro Marshall Nakagawa, Harry Davis, Chozha Vendan Rathinam
Inflammatory signals have been shown to play a critical role in controlling the maintenance and functions of hematopoietic stem cells (HSCs). While the significance of inflammation in hematopoiesis has begun to unfold, molecular mechanisms and players that govern this mode of HSC regulation remain largely unknown. The E3 ubiquitin ligase A20 has been considered as a central gatekeeper of inflammation. Here, we have specifically depleted A20 in multi-potent progenitors (MPPs) and studied its impact on hematopoiesis. Our data suggest that lack of A20 in Flt3^+^ progenitors causes modest alterations in hematopoietic differentiation. Analysis of hematopoietic stem and progenitor cell (HSPC) pool revealed alterations in HSPC subsets including, HSCs, MPP1, MPP2, MPP3 and MPP4. Interestingly, A20 deficiency in MPPs caused loss of HSC quiescence and compromised long-term hematopoietic reconstitution. Mechanistic studies identified that A20 deficiency caused elevated levels of Interferon-γ signaling and downregulation of p57 in HSCs. In essence, these studies identified A20 as a key regulator of HSC quiescence and cell fate decisions.
Hematopoietic Stem Cells (HSCs) are believed to be largely quiescent under steady state conditions and enter into an actively proliferating state in response to external cues. In the recent years, role of infection and inflammation in the control of hematopoiesis have gained a lot of attention. HSCs have been evolved to sense infections either through direct contact with the pathogens via the Toll Like receptor (TLR) pathways or through actions of inflammatory cytokines produced by the effector cells of the immune system and hematopoietic progenitors (Baldridge et al., 2010; Baldridge et al., 2011; Weiner et al., 2008). A spectrum of pro-inflammatory cytokines and chemokines, that includes IL1, IL6, IL8, TNF, CC-Chemokine ligand 2 (CCL2), IFN-α and IFN-γ, has been identified to regulate HSCs and hematopoiesis. In particular, chronic exposure of HSCs to interferons (both α and γ) results in compromised self-renewal and quiescence of HSCs.
A20 (also known as Tnfaip3) acts as an ubiquitin editing enzyme and has emerged as a key anti-inflammatory molecule of the immune system. A20 contains an amino (N)-terminal cysteine protease/DUB domain (that is necessary for the deubiquitylating functions) and a carboxyl (C)-terminal zinc finger (ZNF) domain (which confers the E3 ubiquitin ligase functions) (Wertz et al., 2004). A20 catalyzes the K48-linked ubiquitylation of target proteins through its caboxy-terminal ZNF domain, therefore it directs its targets for proteasomal degradation. In addition, A20 removes K63–linked ubiquitin chains from its target proteins, which not only inactivates the signaling function of the targets but might also facilitate its K48- linked ubiquitylation and degradation (Wertz et al., 2004). The negative signaling function of A20 involves deconjugation of K63–linked ubiquitin chains from TRAF6 and RIP1, which are central players of the toll like receptor (TLR) and Tumor necrosis factor receptor (TNFR) pathways (Sun, 2008). In addition, A20 also mediates deubiquitylation of RIP2 and thereby negatively regulating the activation of NF-kB and the induction of pro-inflammatory cytokines (Hitotsumatsu et al., 2008; Hymowitz and Wertz, 2010; Sun, 2008; Vereecke et al., 2009). Functions of A20 in many cell types of the immune system have been clearly established, however, its role in hematopoiesis remains largely unknown. We have recently identified that A20 deficiency in HSCs leads to loss of its pool, pathologic hematopoiesis, including auto-inflammatory disease, myeloproliferation and lymphopenia, and postnatal lethality that are dependent on IFNγ (Nakagawa et al., 2015). In the present study, we specifically ablated A20 in (Flt3^+^) multi-potent progenitors (MPPs), but not in HSCs, and our data identified that presence of A20 in HSCs is sufficient and necessary to prevent autoinflammatory disorders. In addition, the current study demonstrates that lack of A20 is sufficient to affect HSC pool and quiescence.
To study the role of A20 in hematopoietic differentiation, we crossed A20 floxed mice (Nakagawa et al., 2015) with Flt3^cre/+^ (Benz et al., 2008) transgenic mice to generate A20^F/F^Flt3^cre/+^ mice (henceforth referred to as KO) Flt3 Cre has been shown to induce recombination in all hematopoietic lineage, including myeloid erythroid and lymphoid, cells starting from MPPs (Flt3^+^ LSK) (Boyer et al., 2011). Analysis of hematopoietic organs from KO mice indicated elevated, but statistically insignificant, cellularity of BM and spleen, and relatively normal cell counts in thymus (Fig. 1A). Determination of recombination efficiencies by PCR indicated A20 deletion in majority of BM cells of KO mice (Fig. 1B). Consistently, flow cytometric analysis of A20^F/F^Rosa^RFP^Flt3^cre/+^ mice revealed deletion efficiencies (as inferred by RFP expression) of > 75% in BM, spleen, thymus and peripheral blood of KO mice (Fig. 1C). Our analysis of RFP expression in various hematopoietic progenitor subsets in the BM identified that majority (> 90%) of CD150^+^CD48^−^LSK cells (a subset that is enriched for HSCs and MPP1 (Cabezas-Wallscheid et al., 2014; Wilson et al., 2008)), were RFP^−^ and that all the other downstream hematopoietic subsets, including CD150^+^CD48^+^ LSK (a subset enriched for MPP2) and CD150^+^CD48^+^ LSK (a subset enriched for MPP3 & MPP4), show remarkable expression of RFP (Fig. 1D, E). These data are consistent with earlier studies that Flt3 Cre mouse strain does not induce recombination in LT-HSCs (Boyer et al., 2011). First, we assessed multi-lineage hematopoietic differentiation in KO by flow cytometry and the data indicated increased proportions of CD11b^+^ myeloid lineage cells and CD3e^+^ T lineage cells, reduced proportion of CD19^+^ B lineage cells, and normal frequencies of Ter119^+^ erythroid lineage cells in the BM of KO mice (Fig. 1F). However, determination of absolute cell counts revealed normal numbers of myeloid/erythroid lineage cells, reduction of B cells, and modestly increased numbers of T cells in the BM of KO mice (Fig. 1G). On the other hand, both relative and absolute frequencies of myeloid, erythroid and T cells were normal in the spleen of KO mice (Fig. 1H, I). Whereas the frequencies of B cells in the KO spleen were modestly decreased, their absolute numbers were normal (Fig. 1H, I). Analysis of thymus of KO mice revealed intact thymic differentiation with normal frequencies and numbers of CD4^−^CD8^−^; CD4^+^CD8^+^; CD4^+^CD8^−^; and CD4^−^CD8^+^ cells (Fig. 1J). Finally, analysis of peripheral blood indicated normal frequencies of myeloid (CD11b^+^) cells and T (CD3e^+^) cells, but reduced frequencies of B (CD19^+^) cells (Fig. 1K). Taken together, these data indicated that deletion of A20 at the MPP stage causes only modest alterations in hematopoietic differentiation.
Next, we studied the hematopoietic stem and progenitor cell (HSPC) compartments in the BM of KO mice. Analysis of Lin^−^Sca1^+^c-Kit^+^ (LSK) cells of the BM from KO mice indicated; a significant increase in absolute numbers, but normal relative numbers, of CD34^−^Flt3^−^LSK subset; an increase in both relative and absolute numbers of CD34^+^Flt3^−^LSK subset; and a decrease in relative numbers, but an increase in absolute numbers, of CD34^+^Flt3^+^LSK subset (Fig. 2A, C, D). To further characterize the LSK subset, based on CD150 and CD48 expression, we followed a more refined immunophenotyping scheme of HSCs and MPPs (Cabezas-Wallscheid et al., 2014; Wilson et al., 2008). Our analysis identified; a decrease in relative numbers and an increase in absolute numbers of CD150^+^CD48^−^LSK subset (a fraction that is enriched for LT-HSCs and MPP1); an increase in both relative and absolute numbers of CD150^+^CD48^+^ LSK subset (a fraction that is enriched for MPP2); and a decrease in relative numbers, but an increase in absolute numbers, of CD150^−^CD48^+^ LSK subset (a fraction that is enriched for MPP3 & MPP4) (Fig. 2A, E, F). Further enrichment of these subsets into HSCs (CD150^+^CD48^−^CD34^−^Flt3^−^LSK), MPP1(CD150^+^CD48^−^CD34^+^Flt3^−^LSK), MPP2 (CD150^+^CD48^+^CD34^+^Flt3^−^LSK), MPP3 (CD150^−^CD48^+^CD34^+^Flt3^−^LSK) and MPP4 (CD150^−^CD48^+^CD34^+^ Flt3^+^LSK) subsets revealed; a relative decrease, but normal absolute numbers of HSCs; normal relative and absolute numbers of MPP1; normal relative numbers, but increased absolute numbers of MPP2; increased relative and absolute numbers of MPP3; and decreased relative number, but normal absolute numbers of MPP4 (Fig. 2G–I). Overall, these studies specified that A20 deletion in Flt3^+^ progenitors disturbs the maintenance of HSPC pool.
To test if this altered distribution of the HSPC pool results in compromised HSC functions, we first performed a series of bone marrow transplantation (BMT) studies. Transfer of RBC depleted total BM from either control or A20 KO into lethally irradiated congenic (CD45.1) recipients resulted in normal radioprotection (Suppl. Fig. 1A, B) and development of donor derived hematopoiesis at 8 weeks of transplantation (Suppl. Fig. 1C). BMT studies using BM of A20^+/+^ ROSA^RFP^ Flt3^Cre/+^ or A20^F/F^ ROSA^RFP^ Flt3^Cre/+^ mice as donors indicated that majority of the A20 KO derived cells (RFP^+^) underwent recombination (Suppl. Fig. 1D). Multi-lineage analysis indicated comparable frequencies of myeloid cells, B cells and T cells within RFP^+^ cells (Suppl. Fig. 1E).
Consistently, competitive repopulation studies (Suppl. Fig. 1F) indicated comparable levels of donor derived hematopoiesis (Suppl. Fig. 1G). Multi-lineage hematopoiesis in the peripheral blood of recipients that received BM of either control (CD45.2^+^; A20^+/+^ ROSA^RFP^ Flt3^Cre/+^) + competitive (CD45.1) mice or KO (CD45.2^+^; A20^F/F^ ROSA^RFP^ Flt3^Cre/+^ mice) + competitive (CD45.1) mice indicated comparable frequencies of myeloid cells, B cells and T cells (Suppl. Fig. 1H).
To assess the long-term reconstitution abilities of KO HSCs, we performed secondary transplantations (Fig. 3A). Transfer of BM of primary recipients (WT; CD45.1^+^), that received either control or A20 KO BM 12 weeks before, into lethally irradiated secondary recipients (WT; CD45.1^+^) caused a modest, but not statistically significant, loss of radioprotection (Fig. 3B). Analysis of donor derived chimera in the peripheral blood of secondary recipients after 12 weeks of transplantation revealed a modest decrease of KO (CD45.2) derived hematopoiesis (Fig. 3C). Multi-lineage analysis indicated that donor (CD45.1) derived myeloid cells were normal (Fig. 3D). However, donor derived B cells (CD19^+^) and T cells (CD3ε^+^) were reduced in the peripheral blood of secondary recipients that received KO BM (Fig. 3D). Overall these data indicate that A20 deficiency in MPPs affects the long-term, but not reconstitutional abilities of HSCs.
To identify physiological alterations that might contribute to the altered distribution of HSPC pool in A20 KO mice, we performed apoptosis and quiescence studies. Maintenance of the HSPC pool has been shown to be tightly regulated through apoptosis and quiescence (Cheung and Rando, 2013). Flow cytometric analysis indicated normal frequencies of apoptosis between control and A20 deficient HSPCs (Fig. 3E). Next, we performed quiescence studies using 3 independent approaches. First, we measured cell cycle through Hoechst and Pyronin Y staining (Shen et al., 2008) and data indicated that the frequencies of quiescent (Hoechst^low^ Pyronin Y^low^) cells were reduced in the HSPC compartment of A20 KO mice (Fig. 3F). Second, we measured cell cycle status using Ki67 and Hoechst (Kim and Sederstrom, 2015) and data revealed that the frequencies of cells in G0 (Quiescent) phase were reduced and in G2/S phase were increased in HSPC compartment of A20 deficient mice (Fig. 3G, H). Similarly, the frequencies of G0 cells were reduced, albeit not significant, and of cells in G2/S phase were remarkably increased in the CD150^+^CD48^−^ LSK (HSC and MPP1 enriched subset) cells of A20 deficient mice (Fig. 3I, J). Third, we performed BrdU incorporation studies to directly measure proliferation of HSPC subsets. Consistent with our Ki67 and Hoechst studies, the frequencies of proliferating cells were more in LSK, CD150^+^CD48^−^LSK, CD34^+^Flt3^−^LSK and CD34^+^Flt3^+^LSK subsets (Fig. 3K, L). These studies have unequivocally demonstrated that A20 deficiency in MPPs causes loss of quiescence in all subsets of HSPCs.
To identify mechanisms through which A20 deficiency affects HSC quiescence, we performed gene expression studies. Based on the previous findings (Baldridge et al., 2010; Baldridge et al., 2011; Essers et al., 2009; King and Goodell, 2011; Mirantes et al., 2014; Sato et al., 2009), including our own (Nakagawa et al., 2015), on the critical role of inflammatory cytokines in HSC quiescence, we have hypothesized that exaggerated expression levels of inflammatory cytokines might be responsible for the loss of HSC quiescence in A20 deficient HSCs. Real-Time PCR analyses on total bone marrow cells suggested normal expression levels of many NF-κB targets, except a modest upregulation of NF-κBIα and a downregulation of BclXL, in the BM of A20 deficient mice (Fig. 4A). While expression levels of IL-6 and TNF-α were normal, a remarkable upregulation of interferon (IFN)γ and a modest down-regulation of IFN-α were noticed in the total BM of A20 deficient mice (Fig. 4A). Consistent with increased IFNγ mRNA expression in the BM, IFNγ levels in the blood serum were elevated in KO mice (Fig. 4B).
Previous studies established that increased IFN-γ signals cause increased surface expression of Seal in HSPCs (King and Goodell, 2011; Snapper et al., 1991). Our analysis of A20 deficient HSPCs revealed a marked upregulation of Sca1 in the surface of Lin^−^c-Kit^+^CD150^+^ cells (Fig. 4C). In agreement with elevated IFN-γ, expression levels of its down targets including Irf1, Irgm1 and Stat1 were elevated in Flt3^+^LSK cells and CD150^+^CD48^−^LSK cells of KO mice (Fig. 4D, E), even though expression levels of many NF-κB targets were normal in Flt3^+^ LSK cells (Fig. 4 D). Amongst other molecular targets of IFN-γ (de Bruin et al., 2013), we focused on the expression levels of p57. Interestingly, expression levels of p57 were reduced in A20 deficient Flt3^+^ LSK cells and CD150^+^CD48^−^LSK cells (Fig. 4 D, E). In keeping with our findings loss of p57 in HSCs results in reduced quiescence and functions of HSCs (Matsumoto et al., 2011). In essence these studies suggest that A20 deficiency in MPP causes elevated IFN-γ expression, and therefore its signaling cascade in HSCs, which eventually may result in the loss of HSC quiescence.
In the present study we show that deletion of A20 in Flt3^+^ hematopoietic progenitor cells (A20^Flt3-KO^) resulted in altered distribution of HSPC subsets, including HSCs. While loss of HSC quiescence was noticed in A20^Flt3-KO^ mice, this did not impair the short-term repopulation capacities of HSCs and caused only modest reduction in long-term hematopoietic reconstitution functions of HSCs. Interestingly, A20 deficiency in MPPs did not result in overt hematopoietic abnormalities or pathological hematopoiesis. In particular, we did not observe systemic inflammation, lymphopenia and myeloproliferation, even though modest changes in B cell numbers were noticed in hematopoietic organs. These findings were unexpected based on our previous studies on A20 deficient mice (Nakagawa et al., 2015). We previously reported that A20 deficiency in HSCs (A20^Hem-KO^) resulted in a severe and lethal phenotype; including premature death, systemic inflammation, hepatomegaly, splenomegaly, reduced HSC quiescence and loss of HSC pool and functions (Nakagawa et al., 2015). On the other hand, analysis of HSC compartment of A20^Flt3-KO^ mice indicated increased absolute numbers of HSPC subsets, including CD150^+^CD48^−^LSK, CD34^+^Flt3^−^LSK and CD34^+^Flt3^+^LSK fractions. This observation was conflicting with our previous findings that CD150^+^CD48^−^LSK pool and CD34^+^Flt3^+^LSK numbers were severely reduced in A20^Hem-KO^ (Nakagawa et al., 2015). One possible explanation for these discrepancies could be that A20 functions are intact in HSCs of A20^Flt3-KO^ mice, as A20 is deleted in hematopoietic cells after the Flt3^+^ stage (i.e. in MPPs), whereas A20 gene is deleted in all hematopoietic cells, including LT-HSCs, of A20^Hem-KO^ mice. Together these observations suggest that A20 functions in HSCs are critical for the suppression of lethal inflammatory disorders and that A20 is an essential for the maintenance and functions of HSPC pool.
Inflammatory cytokines, including IL1, IL6, TNFα, IFNα, and IFNγ, are shown to regulate hematopoiesis and the physiology of HSCs (Baldridge et al., 2010; Baldridge et al., 2011; Essers et al., 2009; King and Goodell, 2011; Mirantes et al., 2014; Sato et al., 2009). In particular, IFNγ plays a pivotal role in the maintenance of HSC pool and quiescence (Baldridge et al., 2010). Moreover, deregulated IFNγ signaling has been implicated in the pathogenesis of various hematopoietic abnormalities (de Bruin et al., 2014). Our previous studies on A20^Hem-KO^ (Nakagawa et al., 2015) identified a key role for IFNγ in the onset of pathologic hematopoiesis (Nakagawa et al., 2015). Interestingly, complete ablation of IFNγ signals in A20^Hem-KO^ mice resulted in a rescue of the inflammatory phenotype, including premature death, and the HSC phenotype (Nakagawa et al., 2015). Consistent with our data on A20^Hem-KO^ (Nakagawa et al., 2015), in the current study, we noticed marked elevation of IFNγ, very likely due to increased NF-κB binding to IFNγ promoter (Nakagawa et al., 2015), in A20^Flt3-KO^ mice. Increased signaling mediated by IFNγ has been shown to directly affect HSC quiescence by suppressing the expression levels of p57, a key regulator of HSC quiescence (Nakagawa et al., 2015). Our gene expression studies on the HSC subsets identified a striking down regulation of p57 in HSCs. These data suggest that deregulated IFNγ signals might be responsible for the loss of HSC quiescence in A20^Flt3-KO^ mice, however, we cannot rule out the possible involvement of additional pathways.
The study presented here indicated that A20 deficiency in MPPs, is sufficient to alter LT-HSC pool and functions. Because A20 expression and functions are intact in LT-HSCs, we concluded that A20 deficiency affects HSCs in a cell extrinsic manner. Based on the compelling role of IFNγ in the cell extrinsic mode of HSC regulation, we speculate that the HSC phenotype of A20^Flt3-KO^ mice is caused by IFNγ, as we demonstrated earlier (Nakagawa et al., 2015). Even though our studies have not specified the source of IFNγ in these settings, we assume that cells of the lymphoid lineage (as they are differentiated from the Flt3^+^ hematopoietic progenitors), might be predominantly responsible for the elevated IFNγ in A20^Flt3-KO^ mice. Future studies that aim to decipher the significance of IFNγ produced by the cells of the HSC microniche vs. systemic IFNγ produced by immune cells of lymphoid organs would be of great interest to understand the pathophysiological significance of IFNγ. In toto, the study presented here highlights the significance of inflammatory pathways in regulating the HSC pool and identifies A20 as a key regulator of inflammation induced hematopoiesis.
Tnfaip3(A20)^flox/flox^ mice were previously described (Nakagawa et al., 2015). Flt3^Cre/+^ mice were generated and reported earlier (Benz et al., 2008). CD45.1 congenic animals were purchased from the National Cancer Institute. The Institutional Animal Care and Use Committee of Columbia University and University of Maryland School of Medicine approved all mouse experiments.
Cells were analyzed by flow cytometry with FACS Fortessa, LSR II (BD), Attune NXT and FACSDiva software (BD Biosciences) or FlowJo software (Tree Star). The following monoclonal antibodies were anti- CD34 (RAM34), anti-CD45.1 (A20), anti-CD45.2 (104), anti-CD48 (HM48-1), anti-CD117 (2B8), anti-Flt3 (A2F10.1), anti-Sca-1 (D7), anti-B220 (RA3-6B2), anti- CD19 (1D3), anti-CD3 (145-2C11), anti-CD4 (GK1.5), anti-CD8 (53–6.7), anti-CDllb (M1/70), anti– Gr-1 (RB6-8C5) and anti-Ter119 (TER119; from BD Biosciences); anti-CD150 (TC15-12F12.2) from Biolegend; anti-CD16/32 (93) from eBioscience; anti-Mpl (AMM2) from IBL. Cells incubated with biotinylated monoclonal antibodies were incubated with fluorochrome-conjugated streptavidin–peridinin chlorophyll protein–cyanine 5.5 (551419; BD) or streptavidin-allophycocyanin-Cy7 (554063; BD). In all the FACS plots, indicated are the percentages (%) of the gated fraction.
1 × 10^6^ of bone marrow cells were injected into lethally irradiated (10 Gy) congenic (CD45.1^+^) recipient mice. For competitive-repopulation experiments, 5 × 10^5^ BM cells from either control or KO mice were mixed with 5 × 10^5^ of WT (CD45.1^+^) BM cells (to obtain a ratio of 1) and were injected into lethally irradiated congenic WT (CD45.1^+^) recipient mice.
For serial transplantation assays, 1 × 10^6^ of bone marrow cells were injected into lethally irradiated (10 Gy) WT congenic (CD45.1^+^) recipient mice. After 12 weeks of transplantation, 1 × 10^6^ BM cells of primary recipients were injected into lethally irradiated WT congenic secondary recipients.
Apoptotic cells were detected by Annexin V PE Apoptosis Detection Kit according to the manufacturer’s instructions (BD Bioscience).
For bromodeoxyuridine (BrdU) assay, 3.33 mg of BrdU (BD Pharmingen) was injected intraperitoneally and mice were maintained on 0.8 mg/ml BrdU in the drinking water. After 16 h of injection, mice were sacrificed and bone marrow cells were stained for BrdU, following the BrdU Flow Kit manufacturer’s instructions (BD Pharmingen).
Cells were stained with surface markers and fixed and permeabilized with Cytofix/Cytoperm fixation/permeabilization solution kit (BD). These cells were stained with Ki-67 (B56;BD) for 30 min and subsequently stained with 20 μg/ml Hoechst 33342 for 5 min and analyzed by flow cytometry.
Levels of IFNγ in the blood sera were analyzed using IFNγ ELISA Set (BD Biosciences), according to manufacturer’s instructions.
Total RNA was isolated with an RNeasy Mini kit or RNeasy Micro kit (QIAGEN). cDNA was synthesized with Oligo(dT) primer and Maxima reverse transcription (Thermo Fisher Scientific). PCR was performed with T100 thermal cycler (Bio-Rad Laboratories) and TSG Taq (Lamda Biotech). Real-time PCR was performed in duplicates with a CFX-connect real-time PCR system (Bio-Rad Laboratories) and SsoAdvanced SYBR Green Supermix (Bio-Rad Laboratories) according to the manufacturer’s instructions. Relative expression was normalized to the expression levels of the internal control (housekeeping gene) Hprt.
Data represent mean and s.e.m. Two-tailed student’s t-tests were used to assess statistical significance (*P < .05, **P < .01, *** < 0.001). For survival curve analysis, log rank test was used to assess statistical significance (*P < .05, **P < .01, *** < 0.001).