Advertisement

FLVCR is necessary for erythroid maturation, may contribute to platelet maturation, but is dispensable for normal hematopoietic stem cell function

John C. H. Byon, Jing Chen, Raymond T. Doty and Janis L. Abkowitz

Key Points

  • FLVCR deletion causes proerythroblast apoptosis and lethal anemia but leads to increased megakaryocyte ploidy and platelet counts.

  • Loss of FLVCR is dispensable for hematopoietic stem cell function during steady-state and stress conditions.

Abstract

Heme is a pleiotropic molecule that is important for oxygen and oxidative metabolism, most notably as the prosthetic group of hemoglobin and cytochromes. Because excess free intracellular heme is toxic, organisms have developed mechanisms to tightly regulate its concentration. One mechanism is through active heme export by the group C feline leukemia virus receptor (FLVCR). Previously, we have shown that FLVCR is necessary for embryonic and postnatal erythropoiesis. However, FLVCR is also expressed in numerous other tissues, including hematopoietic stem cells (HSCs). To explore a possible role for FLVCR in HSC function, we performed serial, competitive repopulation transplant experiments using FLVCR-deleted and control bone marrow cells, along with wild-type competitor cells. Loss of FLVCR did not impact HSC function under steady-state or myelotoxic stress conditions (such as arsenic or radiation exposure), nor did FLVCR deletion result in alterations in the various progenitor compartments. However, even when 95% of the donor bone marrow cells lacked FLVCR, all red cells in recipient mice were wild type. This is due to the increased apoptosis of FLVCR-deleted proerythroblasts. Also, remarkably, loss of FLVCR increased megakaryocyte ploidy. Together, these findings show FLVCR is redundant in stem cells but has critical and contrasting stage-specific roles in discrete hematopoietic lineages.

Introduction

Hematopoietic stem cells (HSCs) are defined by several properties: quiescence, self-renewal, and multilineage differentiation. Perturbation of one or more of these characteristics leads to abnormal hematopoiesis and, in some cases, malignancy. Our current understanding of HSC quiescence has focused primarily on the role of cell cycle regulatory proteins. Maintenance of either quiescence or cell proliferation requires appropriate matching of the cellular metabolic state to the activation state of the HSCs. Several recent studies have highlighted the importance of metabolic energy balance in regulating HSC function.

The tumor suppressor gene LKB1/STK11 is inactivated in Peutz-Jegher syndrome, a familial cancer predisposition syndrome. LKB1/SKT11 is a serine/threonine kinase that functions upstream of AMP kinase, mammalian target of rapamycin, and FoxO transcription factors. LKB1/STK11 acts to restrict cell growth under energetically unfavorable conditions by regulating cellular metabolism. Mice engineered to lack LKB1/STK11 display loss of HSC quiescence and rapid proliferation of the HSC compartment.1-3 This proliferation is transient, and null animals eventually undergo depletion of HSCs and die of pancytopenia. This effect appears to be independent of AMP kinase, mammalian target of rapamycin, and FoxO signaling. Transcriptome analysis of LKB1/STK11-null mutants revealed downregulation of peroxisome proliferator-activated receptor γ, coactivator (PGC)-1α and PGC-1β, regulators of the peroxisome proliferator-activated receptor γ pathway.1 One key function of peroxisome proliferator-activated receptor γ is to regulate mitochondrial biogenesis. Consistent with this, LKB1/STK11-null animals display defects in mitochondrial number and function. Metabolic profiling of LKB1/STK11 mutants revealed elevated fatty acid synthesis.2 Nakada et al3 discovered that LKB1/STK11-deficient cells underwent aneuploidy and subsequent apoptosis.

Additionally, the majority of the bone marrow is poorly oxygenated, resulting in a hypoxic HSC microenvironment. In response to these hypoxic conditions, HSCs preferentially use glycolysis rather than mitochondrial oxidative phosphorylation for generation of ATP.4 The effect of hypoxia on HSC metabolism is through Hypoxia-inducible factor α, which also promotes quiescence by regulating the cell cycle regulators p16Ink4a and p19Arf.5

Autophagy is a lysosomal degradation pathway that acts to remove damaged or dysfunctional proteins and organelles. It becomes activated in response to cellular damage and nutrient deprivation. Activation of autophagy is also necessary in certain developmental steps, such as reticulocyte maturation.6 Deletion of Atg7, which is important for autophagasome formation, leads to myeloproliferation and death due to eventual loss of HSC function.7 HSCs of Atg7-null mice display accumulation of mitochondria, elevations in reactive oxygen species (ROS), and DNA damage.

These findings highlight the importance of proper metabolic regulation in maintaining normal HSC function. Additionally, they suggest that dysregulation of other metabolic pathways may influence HSC fate.

The group C feline leukemia virus receptor (FLVCR) is a transporter molecule that exports cytoplasmic heme. Heme is required by all cells, not only erythroid cells, and its intracellular concentrations need to be tightly regulated, as excess free heme is an oxidative stress and toxic. Previously we demonstrated that deletion of FLVCR is embryonically lethal secondary to severe anemia.8 Conditional deletion of FLVCR results in a pure red cell aplasia with a block in erythropoiesis at the proerythroblast stage, likely through excess intracellular heme concentrations, which are toxic.8 Interestingly, FLVCR is also highly expressed in HSCs. We have, for example, demonstrated that group C feline leukemia virus–pseudotyped retroviral particles can efficiently transduce HSCs.9,10 However, it remains to be seen whether FLVCR and regulation of intracellular heme has any functional significance in HSCs.

To assess this we performed serial competitive repopulation studies and also further stressed recipient animals with arsenic or low-dose (200 cGy) irradiation, agents that function in part by increasing intracellular ROS. Our results indicate that FLVCR is dispensable for HSC function, as there likely are redundant pathways regulating oxidative stress in HSC. However, FLVCR is necessary to prevent apoptosis and permit normal erythroid maturation and has a subtle and opposing effect on platelet maturation. Whereas several genes have been identified that direct lineage commitment (erythroid vs megakaryocyte), FLVCR is unique in that it functions after lineage commitment has already occurred. Proerythroblasts lacking FLVCR die, whereas megakaryocytes lacking FLVCR have increased ploidy and robust platelet production. Thus, these two lineages use FLVCR (and presumptively heme export) for opposite consequences.

Materials and methods

Mice

These studies were performed with Institutional Animal Care and Use Committee approval at the University of Washington. All wild-type mice, PeP3b (B6 SJL/Ly5.1), C57BL/6J (B6.Ly5.2), and green flourescent protein (C57BL/6-Tg UBC-GFP), were purchased from The Jackson Laboratory (Bar Harbor, ME), bred, and maintained under specific pathogen-free conditions. Flvcrflox/flox;Mx-Cre+ and Mx-Cre+ mice were generated in our laboratory as previously described,8 and wild-type Cre+ littermates were used as controls.

Transplantation

Eight- to 12-week-old Flvcrflox/flox;Mx-Cre+ or Mx-Cre+ mice were treated with 250 μg poly(I)-poly(C) intraperitoneally every other day for 3 doses8 and were euthanized 10 to 12 days later. Bone marrow cells were harvested and transplanted with wild-type bone marrow cells from Pep3b or GFP mice at a ratio of 50:50 (1:1) or 95:5 (20:1) into lethally irradiated (1100 cGy) C57BL/6 female mice (107cells per animal). Recipient animals were followed with aging for over a year with intermittent blood sampling. Secondary and tertiary transplants were performed for some cohorts (bone marrow cells from 2 to 3 donor mice were pooled and transplanted into lethally irradiated C57BL/6 female mice).

Myelotoxic stress

To induce stress on hematopoiesis, mice received sublethal radiation treatment (200 cGy) at 12 weeks after transplant or arsenic trioxide (Sigma-Aldrich), 8 mg/kg, intraperitoneally, 2 injections given 4 weeks apart, starting at 12 weeks after transplant.

Flow cytometry

For peripheral blood analysis, mice were bled retro-orbitally into EDTA anticoagulated microtainer tubes (Becton-Dickinson). Single cell suspensions were prepared from peripheral blood cells or from freshly isolated bone marrow and were stained with anti–CD45.2-fluorescein isothiocyanate (FITC), anti–CD45.1-PE, anti–Gr-1-APC, anti–Ter119-phycoerythrin (PE)-Cy7, and anti–CD41-PE antibodies (BD Biosciences). Flow cytometry was performed on FACSCanto or LSRII (BD Biosciences), and data were analyzed with FlowJo (TreeStar) software. For GFP analysis of red blood cells and platelets, flow was run on peripheral blood diluted 1:500 with phosphate-buffered saline (Ca+2, Mg+2 free) with forward scatter and side scatter set on a logarithmic scale.

For analysis of LinSca-1+kit+ (LSK) and progenitor compartments, bone marrow suspensions were labeled with anti–CD45.1-Alexa 700, anti–CD45.2-V500, anti–CD34-FITC, biotin-conjugated Lin negative cocktail (Miltenyi) and streptavidin APC-Cy7, anti–Sca-1-PE-Cy7, anti–CD16/32-APC, anti–CD127-V450, and anti–CD117-PerCP-Cy5.5. Anti–CD34-FITC was purchased from eBioscience; all other antibodies were purchased from BD Biosciences. LSK cells are defined as LinSca-1+kit+. Common myeloid progenitors are defined as LinSca-1kit+, CD16/32lowCD34low. Granulocyte-megakaryocyte progenitors are defined as LinSca-1lowkit+CD16/32highCD34low. Megakaryocyte-erythroid progenitors are defined as LinSca-1kitlowCD16/32lowCD34low. Common lymphoid progenitors are defined as LinSca-1lowkitlowCD127+.

To identify megakaryocytes, single cell, bone marrow suspensions were stained with anti–CD41-PE and anti–CD45.2-APC.

Intracellular ROS was measured by incubating cultured bone marrow megakaryocytes in 5 μM CM-H2DCFDA (Invitrogen) for 30 minutes in a 37°C water bath, and flow cytometry was performed on a FACS Calibur (BD Biosciences).

Measurement of heme content

Cultured megakaryocytes (100 000 cells per sample) were washed once with phosphate-buffered saline, and the pellet was resuspended in 2 M oxalic acid and heated at 100°C for 30 minutes. This removes heme-bound iron, and the resultant protoporphyrin was measured in a fluorescent plate reader at an emission of 608 nm. Heme content was then calculated by measuring against a standard curve.

Apoptosis

For apoptosis assays, single bone marrow cell suspensions were first stained for anti–CD71-APC and anti–Ter119-PE-Cy7 and then labeled with Annexin V-PE and 7-actinoaminomyci D per the manufacturer’s instructions (Annexin V-PE apoptosis kit; BD Biosciences).

Megakaryocyte ploidy

Total bone marrow was isolated and cultured in StemPro 34 (Invitrogen), 2 mM glutamine, 0.5% Pen-Strep-Fungizone, and 100 ng/mL recombinant human thrombopoietin (PeproTech) for 5 days. Cells were cultured at a density of 5 × 106 cells/mL. On the day of analysis, megakaryocytes were enriched by discontinuous bovine serum albumin gradient (1.5%, 3%) purification. Enriched megakaryocytes were stained with anti–CD41-PE, fixed in 0.1% paraformaldehyde on ice for 1 hour, and permeabilized with 0.1% Triton X-100, and intracellular DNA was stained with 1 μM 4,6 diamidino-2-phenylindole (Accurate Chemicals) on ice for 30 minutes. Flow cytometry was performed on an LSRII (BD Biosciences) and analyzed with FlowJo (TreeStar) software.

Real-time polymerase chain reaction

RNA was isolated from cultured bone marrow megakaryocytes using Trizol reagent (Invitrogen), per the manufacturer’s recommendations. Total RNA was converted to cDNA using the High-Capacity RT kit (Applied Biosystems). Multiplexed real-time polymerase chain reaction (PCR) was performed using FAM-labeled PrimeTime PCR 5′-nuclease assays for FLVCR, heme-oxygenase-1 (HMOX1), and β-actin (purchased from IDT). VIC-labeled TaqMan assays (Applied Biosystems) were used to simultaneously examine expression of TBXAS1 and NQO1. Real-time PCR reactions were set up using TaqMan Fast Advanced Master Mix (Applied Biosystems) per the manufacturer’s instructions. Reactions were run on a Step One Plus (Applied Biosystems) Relative quantitation was performed using the ΔΔCt method and normalized to β-actin expression.

Results

Conditional deletion of FLVCR does not impair HSC repopulating activity

To determine whether deletion of FLVCR affects HSC function, serial, competitive repopulation studies were performed. FLVCR-deleted or wild-type Cre+ bone marrow cells and bone marrow from UBC-GFP+ (GFP driven by the ubiquitin promoter) mice were injected into lethally irradiated hosts in equal numbers. Primary transplanted mice were followed for up to 60 weeks. At week 60 after transplant, mice were euthanized, and secondary transplantation was performed. At 36 weeks into the secondary transplant, animals were euthanized, and a tertiary transplant was performed, and transplanted animals were followed for an additional 32 weeks. Granulocyte counts were followed as a marker of reconstitution. There was no difference in the percentage of GFP+ granulocytes after transplantation with FLVCR-deleted or wild-type Cre+ bone marrow cells (Figure 1A). Both FLVCR-deleted and wild-type Cre+ bone marrow cells contributed approximately equally as the GFP+ competitor cells to overall granulocyte reconstitution. This was verified in a second cohort of transplanted mice (J.C. and J.L.A., unpublished data, December 8, 2008). Furthermore, there was no decrease in the contribution of FLVCR-deleted bone marrow cells to the granulocyte lineage even in tertiary transplanted animals. These results indicate that deletion of FLVCR does not impact HSC repopulating activity or granulocyte differentiation.

Figure 1

FLVCR deletion does not impair hematopoietic reconstitution during competitive, serial transplantation. C57Bl/6 mice were transplanted with equal numbers of FLVCR-deleted bone marrow or wild-type Cre+ bone marrow and GFP+ competitor bone marrow cells. The proportion of granulocytes arising from either FLVCR-deleted or wild-type Cre+ cells were distinguished from GFP+ competitor cells by flow cytometry. Black circles represent competitive transplantation with FLVCR-deleted donor cells. Open squares represent competitive transplantation with wild-type Cre+ donor cells. *P ≤ .05.

FLVCR is dispensable for HSC recovery from myelotoxic stress

Although FLVR is not required for HSC repopulating activity, even after serial transplantation, it is possible that FLVCR is required for the HSC response to certain stress stimuli. To test this hypothesis, competitively transplanted animals were exposed to sublethal irradiation (200 cGy) or arsenic trioxide (8 mg/kg, intraperitoneally × 2) 12 weeks after transplantation. Both of these agents act in part by increasing the formation of intracellular ROS,11-14 which may be exacerbated by increased intracellular free heme concentrations. Because UBC-GFP mice can have variable GFP expression in the different lineages of nucleated peripheral blood cells,15 we chose to perform competitive transplants using mice carrying the CD45.1 and CD45.2 allotypes. In mice transplanted with FLVCR-deleted bone marrow cells (CD45.2+), there was no difference in bone marrow recovery either with respect to cell numbers or kinetics between untreated recipients and recipients treated with arsenic trioxide (Figure 2A) or radiation (Figure 2B). Similar findings were seen in mice transplanted with wild-type Cre+ cells (CD45.2+) with and without exposure to arsenic trioxide (Figure 2A) or radiation (Figure 2B). However, differences were seen between FLVCR-deleted recipients and wild-type Cre+ recipients late after hematopoietic recovery. This may be due to some minor variability in stem cells numbers or alloreactivity toward the CD45.1+ competitor cells, as it has been reported that CD45.1+ cells have impaired engraftment when transplanted into CD45.2+ hosts,16 as has been performed here. Thus, FLVCR is also dispensable in HSC recovery from myelotoxic stress.

Figure 2

Loss of FLVCR is dispensable for HSC recovery from oxidative stress. (A) Twelve weeks after competitive transplantation, mice were exposed to 200 cGy of radiation, and hematopoietic recovery was assessed by flow cytometry of peripheral blood granulocytes. (B) A second cohort of transplanted animals was treated with arsenic trioxide (8 mg/kg intraperitoneally) given 12 and 16 weeks after transplant. Hematopoietic recovery was determined by flow cytometry of peripheral blood granulocytes. Open circles represent competitive transplantation with FLVCR-deleted bone marrow and treatment with arsenic trioxide or sublethal radiation. Closed circles represent competitive transplantation with FLVCR-deleted bone marrow. Black squares represent competitive transplantation with wild-type Cre+ donor cells and treatment with arsenic trioxide. Open squares represent competitive transplantation with wild-type Cre+ donor cells.

FLVCR deletion does not alter the composition of the HSC or progenitor compartments

To further confirm that FLVCR deletion does not impact the HSC or progenitor compartments, bone marrow from animals competitively transplanted with FLVCR-deleted (CD45.2) or wild-type Cre+ (CD45.2) donor cells and B6.SJL/PeP3b (CD45.1) competitor cells was analyzed by flow cytometry. As shown in Figure 3A-B, there is no difference in LSK or progenitor (common myeloid progenitor, granulocyte-megakaryocyte progenitor, megakaryocyte-erythroid progenitor) cell numbers between FLVCR-deleted and wild-type Cre+ bone marrow cells. These results suggest that FLVCR is not essential for earliest progenitors.

Figure 3

Absence of FLVCR does not impact HSC or progenitor cell numbers. Mice were transplanted with equal numbers of FLVCR-deleted or wild-type Cre+ bone marrow cells and B6SJL/Pep3b competitor cells. (A) HSC and (B) progenitor numbers were determined by flow cytometry (see “Materials and methods”).

The impaired erythroid maturation with FLVCR deletion results from increased apoptosis in proerythroblasts and basophilic erythroblasts.

Previously we showed that constitutive deletion of FLVCR is embryonic lethal and that postnatal deletion of FLVCR leads to rapid morbidity due to severe anemia characterized by an erythroid maturation block at the proerythroblast stage. Therefore, to more fully characterize the effect of FLVCR deletion on erythroid maturation, we used the competitive repopulation model described above.

As expected in FLVCR-deleted:GFP-transplanted animals, by 8 weeks after transplant, all of the red cells come entirely from the GFP+ donor cells. In contrast, wild-type Cre+:GFP-transplanted animals have approximately equal contributions to the erythroid lineage from both cell populations (Figure 4A). Even when FLVCR-deleted cells are transplanted in greater numbers (∼20:1, FLVCR-deleted:GFP+), by 8 weeks after transplant, all of the circulating red cells are of GFP+ donor origin (Figure 3B).

Figure 4

Erythropoiesis is severely impaired in FLVCR-deleted bone marrow with a maturation block at the proerythroblast stage. (A) Competitive repopulation assay using FLVCR-deleted or control bone marrow cells and GFP+ competitor cells. The contribution of each donor population to erythropoiesis was determined by flow cytometry of peripheral blood. (B) Competitive transplantation with 95:5 ratio of FLVCR-deleted:GFP+ bone marrow cells. Hematopoietic recovery is determined by granulocyte recovery by flow cytometry. The contribution of FLVCR-deleted cells to erythropoiesis is determined by flow cytometry. (C) Proerythroblast populations were identified by flow cytometry. (C-D) The extent of apoptosis was determined by staining with 7-AAD and Annexin V. Bone marrow was harvested 24 weeks after transplant.

To further characterize the erythroid maturation arrest at the proerythroblast stage in animals transplanted with FLVCR-deleted bone marrow, we measured apoptosis in this cell population. To address this question, proerythroblasts from FLVCR-deleted:GFP and wild-type Cre+:GFP transplants were examined for apoptosis by flow cytometry using the cell permeable dye, 7-AAD and Annexin V. Live cells exclude 7-AAD, whereas early apoptotic cells will appear 7-AAD dim. Late apoptotic and necrotic cells will appear 7-AAD bright, as they have lost membrane integrity and readily take up the dye. Annexin V is a Ca2+-dependent phospholipid binding protein that binds to phosphatidyl-serine on the cell membrane. Increased cell surface phosphatidyl-serine expression is a marker of compromised membrane integrity and cell death. As shown in Figure 4C-D, erythroblasts (CD71+/Ter119+) derived from wild-type Cre+ and GFP+ competitor HSCs are viable, with low levels of apoptosis. In contrast, erythroblasts derived from FLVCR-deleted donor cells were not only much fewer in number than GFP+ competitor cells but also displayed higher levels of apoptosis (Figure 4C-D).

FLVCR deletion has subtle effects on megakaryocyte ploidy and platelet numbers

In mice, following postnatal deletion of FLVCR, there is a significant increase in the platelet counts (Figure 5A) compared with wild-type mice. Because conditional deletion of FLVCR causes a rapid-onset, severe anemia, we initially hypothesized that the concurrent thrombocytosis was a reactive event. However, on more detailed analysis, we discovered that there is increased ploidy in FLVCR-deleted megakaryocytes compared with wild-type controls (Figure 5B). This suggested that there might be a role for heme in regulating megakaryocyte physiology. To investigate this hypothesis futher, we examined platelet counts in our competitive transplantation model, where there is no confounding anemia. We detected transient increases in the percentage of FLVCR-deleted platelets when normalized to the number of FLVCR-deleted granulocytes (Figure 5C). In a second cohort of competitively transplanted animals, we examined bone marrow megakaryocyte numbers and ploidy by flow cytometry at a time when the percentage of FLVCR-deleted platelets was higher than the GFP+ competitor-derived platelets. We found that there was no difference in megakaryocyte numbers between FLVCR-deleted and wild-type Cre+ derived megakaryocytes in the simultaneous control experiments (Figure 5D). Analysis of megakaryocyte ploidy, however, revealed a subtle increase in the ploidy (8N fraction) in megakaryocytes derived from FLVCR-deleted cells compared with wild-type Cre+ and GFP+ competitor cells (Figure 5E-F). Beause there is no effect of FLVCR deletion on the numbers of HSCs, progenitors (Figure 3A-B), or megakaryocytes (Figure 5D), we suspect that the increased megakaryocyte ploidy in FLVCR-deleted cells may be responsible for the increase in the percentage of FLVCR-deleted platelets.

Figure 5

FLVCR-deleted bone marrow chimeras display subtle alterations in platelet numbers. (A) FLVCRflox/flox;Mx-Cre mice were injected with polyinosinic:polycytidylic acid, and platelet counts were determined by Hemavet. (B) After red cell lysis, bone marrow mononuclear cells from conditionally deleted and normal mice were cultured in StemPro34 media with 100 ng/mL of recombinant human thrombopoietin (rh-TPO) for 5 days. DNA content was determined by staining for CD41 and propidium iodide. (C) Competitive repopulation assay using FLVCR-deleted or control bone marrow cells and GFP+ competitor cells. The contribution of each donor population to the platelet count was determined by flow cytometry of peripheral blood. (D) Total bone marrow was isolated from competitive repopulation studies, and the proportion of megakaryocytes derived from FLVCR-deleted or control bone marrow cells was determined by staining for CD41. (E-F) After red cell lysis, bone marrow mononuclear cells from mice competitively transplanted with FLVCR-deleted and wild-type Cre+ bone marrow and GFP+ competitor cells were cultured in StemPro34 media with 100 ng/mL of rh-TPO for 5 days. DNA content was determined by staining for CD41 and 4,6 diamidino-2-phenylindole (DAPI).

To uncover the mechanism responsible for the increased ploidy, we measured total heme content in megakaryocytes cultured from FLVCR-deleted or wild-type Cre bone marrow mononuclear cells, but did not detect any difference (Figure 6A). In addition, we measured mRNA expression of HMOX1, which is induced in response to elevations in heme content, and did not see any significant difference between FLVCR-deleted cultured megakaryocytes and controls (Figure 6C), although there was trend toward increased expression in FLVCR-deleted cells. Finally, we examined the expression of TBXAS1, a megakaryocyte gene that is normally repressed due to binding of the heme-regulated transcriptional repressor, Bach1.17 Unexpectedly, TBXAS1 expression was lower in FLVCR-deleted cultured megakaryocytes compared with controls (Figure 6C). As expected, we did see a dramatic decrease in FLVCR mRNA (∼80%) in FLVCR-deleted cultured megakaryocytes.

Figure 6

Effects of FLVCR-deletion on intracellular heme and ROS in megakaryocytes. Bone marrow mononuclear cells from FLVCR-deleted and wild-type Cre mice were isolated and cultured in StemPro34 media with 100 ng/mL of recombinant human thrombopoietin for 5 days. (A) Intracellular heme levels in FLVCR-del and wild-type Cre cultured megakaryocytes. (B) Representative flow cytometry plot of megakaryocyte intracellular ROS. Intracellular ROS was determined by flow cytometry after staining with 5 μM CM-H2DCFDA for 30 minutes. Bold line represents wild-type Cre and the solid line represents FLVCR-deleted cells. (C) Real-time PCR for FLVCR, TBXAS1, NQO1, and HMOX1 gene expression in cultured megakaryocytes derived from FLVCR-deleted or wild-type Cre bone marrow mononuclear cells. Expression was normalized to β-actin. *P < .05.

One potential consequence of unchecked elevation in intracellular heme levels is generation of ROS. Furthermore, various studies have suggested that increased ROS enhances platelet formation.18,19 Furthermore, intracellular ROS increases with megakaryocyte maturation and may be due in part to decreased expression of Nrf2, a stress-induced transcription factor that activates transcription of many cytoprotective genes.20 Therefore, we examined several parameters of intracellular ROS in megakaryocytes cultured from the bone marrow of control and FLVCR-deleted mice to determine whether this causally contributed to increased platelet numbers. By flow cytometry, we did not detect any difference in the intracellular ROS concentrations between control and FLVCR-deleted cultured megakaryocytes (Figure 6B). Because it is possible that our assay system is unable to detect subtle changes in intracellular ROS levels that nonetheless have functional consequences, we examined changes in expression of reduced NAD phosphate dehydrogenase, quinone1 (Nqo1), which is induced by increased ROS, through an Nrf2-dependent mechanism. There was no significant difference in Nqo1 expression between FLVCR-deleted megakaryocyte cultures and controls (Figure 6C).

These data show that FLVCR independently regulates the later stages of hematopoietic maturation in multiple lineages. The positive effects of FLVCR deletion in platelets are in stark contrast to its negative effect on red cells, where FLVCR deletion leads to a complete block in erythrocyte maturation. However, whether the regulation of megakaryocyte ploidy is transcriptional (eg, mediated via heme binding to Bach1), translational (eg, mediated via heme binding of heme-regulated eIF2α kinase), mediated via a hemoproteins, or another consequence of FLVCR deletion is unclear.

Discussion

In the present study, we demonstrated that FLVCR is dispensable for normal HSC function, despite high levels of FLVCR expression on these cells. In a competitive repopulation assay, FLVCR-deleted bone marrow cells were able to adequately reconstitute the granulocyte compartment as efficiently as normal, GFP+ competitor bone marrow cells. This held true even after secondary and tertiary transplantation experiments. We did not observe any aging-associated defects. Furthermore, FLVCR deletion did not impair HSC recovery after exposure to sublethal radiation or arsenic trioxide. These data indicate that FLVCR is not required to maintain normal HSC function under steady-state, aging, or stress conditions. This suggests that although FLVCR is present in high quantities on the cell surface of HSCs,9 it may be functionally redundant and that other cell intrinsic mechanisms also regulate intracellular heme. HSCs may have developed unique, specialized pathways to metabolize excess free heme. They may use more heme in the form of hemoproteins, especially those that may act to minimize the levels of intracellular ROS, which are detrimental to the HSC compartment. Alternatively, HSCs may upregulate expression of heme-oxygenases to degrade intracellular heme. Finally, HSCs may express a heretofore, unknown cytosolic heme exporter. This is conceptually similar to the multiple mechanisms used by HSCs to respond to oxidative stress.21-24

In contrast to the normal phenotype in HSCs deficient for FLVCR, normal red cell and likely platelet development appear dependent on appropriate FLVCR expression. In prior work, we have shown that germ-line deletion of FLVCR in mice causes embryonic lethality, whereas conditional deletion leads to a severe anemia and significant morbidity. FLVCR is required for red cell maturation beyond the proerythroblast stage. In this study, our aim was to further characterize the erythroid defect induced by loss of FLVCR. It is clear that FLVCR is essential to erythropoiesis, as nearly all of the circulating red cells in our competitive repopulation experiments derive from the wild-type GFP+ competitor population, even when FLVCR-deleted bone marrow cells are transplanted in significant excess. There is a minor contribution from the FLVCR-deleted donor cells within the first 8 weeks after transplant; however, there is no long-term erythropoiesis derived from cells lacking FLVCR. Further detailed analysis of the proerythroblasts derived from FLVCR-deleted cells indicates increased apoptosis. Although there could simply be a maturation arrest followed by senescence, we favor the possibility that proerythroblasts are especially vulnerable to heme excess and toxicities, as these cells must rapidly synthesize heme yet lack globin.8

Unlike the inhibitory effects of FLVCR loss on red cell development, deletion of FLVCR may increase platelet production. We explored 2 possible mechanisms. Increased free heme in megakaryocytes may result in increased heme binding to the transcriptional repressor, Bach1, and upregulation of genes involved in platelet formation. When heme is increased, heme binds Bach1 and impairs its ability to bind DNA and repress transcription.25 Similarly, transgenic mice that overexpress Bach1 only in the megakaryocyte-erythroid lineages develop thrombocytopenia secondary to impaired megakaryocyte endomitosis.17 Unexpectedly, we found that expression of the TBXAS1, a BACH1-dependent gene, was lower in FLVCR-deleted cultured megakaryocytes compared with controls. Further studies are required to unravel the molecular mechanisms underlying these observations, including the potential role of FLVCR1b, an FLVCR isoform that exports heme from the mitochondria to the cytoplasm.26

Next, we explored the possibility that elevated free heme may result in increased ROS and platelet formation. Treatment of the MEG-01 cell line and primary human megakaryocytes with the prostaglandin 15d-PGJ(2) causes increased platelet release via induction of ROS.18,19 Furthermore, intracellular ROS increases with megakaryocyte maturation, partially due to decreased expression of Nrf-2, a stress-induced transcription factor that activated transcrption of many cytoprotective genes.20 We detected no difference in intracellular ROS levels in cultured megakaryocytes derived from FLVCR-deleted or control bone marrow mononuclear cells. However, we cannot rule out increased expression of other Nrf2-dependent stress genes that could represent a compensatory response to increased ROS.

These results are difficult to interpret, however, because we anticipate that free heme is only briefly elevated and at a specific stage (perhaps 2N) of early megakaryocyte development. The concept that heme functions as a quick trigger and coordinator of the molecular events needed for endomitosis would be similar to the signaling role of heme in the control of circadian rhythm27 and consistent with our emerging data in early erythroid cells. Since technologies to explore red cell differentiation are better defined, free heme is sufficiently abundant in early erythroid cells to detect and quantitate, and it appears that heme excess triggers macrocytic erythropoiesis, these studies should help inform future studies of megakaryocyte ploidy, an end point that could also reflect impaired cell division.

The contrasting effect of FLVCR deletion on red cells and platelets is similar to several other genes. Overexpression of GATA-228 or miR-15029 drives megakaryocyte differentiation at the expense of erythroid differentiation. In contrast, KLF1 overexpression30 leads to increased erythroid production with a concomitant decrease in megakaryocyte production. Whereas these genes direct lineage commitment (erythroid vs megakaryocyte), FLVCR is unique in that it functions after lineage commitment has already occurred, making our observations mechanistically unique.

In conclusion, we demonstrated that deletion of FLVCR does not affect HSC function under steady state or conditions of oxidative stress. However, loss of FLVCR has significant consequences for red cell and to a lesser extent on platelet maturation. Furthermore, these effects appear after lineage commitment has occurred, as FLVCR deletion does not alter the number of HSCs or progenitors. Red cells require normal FLVCR expression for full maturation, whereas platelet formation may be enhanced by loss of FLVCR. Although we anticipate that both effects are mediated by transient increases in intracellular heme, we cannot fully exclude the possibility that the enhanced platelet formation reflects a different and not-yet-defined role of FLVCR.

Authorship

Contribution: J.C.H.B. designed and performed experiments, analyzed data, and wrote and edited the manuscript; J.C. designed and performed experiments and analyzed data; R.T.D. performed experiments, analyzed data, and edited the manuscript; and J.L.A. designed experiments, reviewed all data, and edited the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Janis L. Abkowitz, University of Washington, Division of Hematology, 1705 NE Pacific St, Seattle, WA 98195; e-mail: janabk{at}uw.edu.

Acknowledgments

The authors thank Zhantao Yang for performing measurements of intracellular heme content.

This work was supported by the National Institutes of Health grants National Institute of Diabetes and Digestive and Kidney Diseases R01 DK085146 and National Heart, Lung, and Blood Institute (NHLBI) R01 HL031823 (to J.L.A.). J.C.H.B. was supported by the NHLBI K12 scholars program K12 HL087165.

Footnotes

  • The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

  • Submitted October 30, 2012.
  • Accepted August 21, 2013.

References

View Abstract