An MTCH2 pathway repressing mitochondria metabolism regulates haematopoietic stem cell fate

The metabolic state of stem cells is emerging as an important determinant of their fate. In the bone marrow, haematopoietic stem cell (HSC) entry into cycle, triggered by an increase in intracellular reactive oxygen species (ROS), corresponds to a critical metabolic switch from glycolysis to mitochondrial oxidative phosphorylation (OXPHOS). Here we show that loss of mitochondrial carrier homologue 2 (MTCH2) increases mitochondrial OXPHOS, triggering HSC and progenitor entry into cycle. Elevated OXPHOS is accompanied by an increase in mitochondrial size, increase in ATP and ROS levels, and protection from irradiation-induced apoptosis. In contrast, a phosphorylation-deficient mutant of BID, MTCH2’s ligand, induces a similar increase in OXPHOS, but with higher ROS and reduced ATP levels, and is associated with hypersensitivity to irradiation. Thus, our results demonstrate that MTCH2 is a negative regulator of mitochondrial OXPHOS downstream of BID, indispensible in maintaining HSC homeostasis. Changes in the metabolic state of stem cells can trigger a shift from quiescence into cell cycle entry. Here Maryanovichet al. identify mitochondrial carrier homolog 2 (MCH2) as a negative regulator of mitochondrial oxidative phosphorylation in haematopoietic stem cells, maintaining their homeostasis.

C ellular metabolism plays a critical role in determining the fate of stem cells [1][2][3] . Haematopoietic stem cells (HSCs) are mostly retained in a quiescent non-motile state in their bone marrow niches and shift to a migratory cycling and differentiating state, replenishing the blood with mature leukocytes and red blood cells on demand 4,5 . It is well established that ROS drives HSC entry into cycle, and more recently it has been demonstrated that the transition from stem to progenitor cell corresponds to a metabolic switch from glycolysis to mitochondrial oxidative phosphorylation (OXPHOS) [6][7][8][9][10][11] . Previously, the ataxia-telangiectasia mutated (ATM) kinase was shown to play an important role in regulating the self-renewal and quiescence of HSCs by regulating ROS levels 12,13 . More recently we demonstrated that the BH3-interacting domain death agonist (BID) protein acts as a downstream ATM-effector in this pathway and regulates the quiescence of HSCs by balancing ROS levels produced by mitochondria 14 . In our previous study we demonstrated that BID AA HSCs (expressing the non-phosphorylatable BID protein) escape from quiescence. However, the exact mechanism by which BID regulates mitochondrial ROS and its relation to the switch to mitochondrial OXPHOS described in the transition from stem to progenitor cell remained unknown 15 .
To address this issue, we shifted our focus to mitochondrial carrier homologue 2 (MTCH2), BID's receptor-like protein in the mitochondria that plays a critical role in Fas-induced liver apoptosis 16 . Interestingly, many recent genome-wide association studies associate MTCH2 with metabolic disorders including diabetes and obesity, suggesting that MTCH2 also plays a role in metabolism [17][18][19] . Knockout of mtch2 gene in embryonic stem cells hinders the recruitment of BID to mitochondria 16 , therefore, we hypothesized that MTCH2 serves as a critical regulator of HSC mitochondrial function downstream of ATM and BID.
In the present study we demonstrate that loss of MTCH2 primes mitochondrial OXPHOS and drives haematopoietic stem and progenitor cells (HSPCs) into cycle leading to stem cell exhaustion. Similar findings were obtained with our previously described BID AA mice 14 , indicating that MTCH2 is an essential regulator of haematopoietic homeostasis downstream of ATM and BID.
Loss of MTCH2 primes mitochondrial OXPHOS. To evaluate the effect of MTCH2 knockout on mitochondrial function in HSPCs we initially measured cellular respiration in c-Kit þ progenitor-enriched bone marrow cells and revealed an increase in both basal and maximal oxygen consumption of MTCH2 F/F Vav1-cre þ progenitors (Fig. 2a, left and middle panels). Importantly, the MTCH2 knockout enriched progenitors also possess substantial mitochondrial spare respiratory capacity (SRC; Fig. 2a, right panel), which is the extra capacity available in cells to produce energy in response to increased stress and as such is associated with cellular survival 21 . Similar observations have been recently reported, where a deficiency in the glycolytic enzyme lactate dehydrogenase A (LDHA) induces HSC differentiation and is accompanied by an increase in mitochondrial OXPHOS and SRC in HSPCs 22 .
To determine whether the increased respiration observed in the MTCH2-deficient enriched progenitor population correlated with increased respiration in more enriched stem and progenitor cells, we determined mitochondrial NADH levels and redox state in HSPCs (LSK cells), as previously described 23 . As expected, we observed decreased NADH redox due to accelerated electron consumption (Fig. 2b) and an expected increase in mitochondrial NADH content ( Supplementary Fig. 2), correlating with the enhanced mitochondrial respiratory capacity. Further evaluation of mitochondrial parameters in MTCH2-deleted mice showed an increase in the expression of nuclear-encoded subunits of the respiratory complexes in progenitor-enriched population (Fig. 2c). Interestingly, this increase in the expression of respiratory complex subunits was not observed in the LSK cells (not shown), perhaps due to the fact that MTCH2 shows an approximately twofold higher expression level in restricted progenitors (LK population) as compared with HSPCs (LSK population; Fig. 2d). In addition, MTCH2 knockout increased the mitochondria membrane potential (DCm; Fig. 2e), and mildly increased both cellular and mitochondrial ROS levels in HSCs (Fig. 2f). Interestingly, and as would be expected, the increase in respiration/mitochondrial function was accompanied by a decrease in lactate production in enriched progenitor cells (Fig. 2g), indicative of decreased glycolysis. These results identify MTCH2 as a critical suppressor of mitochondrial OXPHOS in HSPCs, and suggest that this suppressive activity is required to prevent excessive differentiation of haematopoietic progenitors leading to HSC exhaustion.
Loss of MTCH2 increases mitochondrial size/volume. To assess whether the mitochondrial functional changes described above in the MTCH2 F/F Vav1-cre þ progenitors were accompanied by changes in mitochondrial structure, we conducted electron microscopy analyses of enriched progenitors. Notably, these analyses revealed that mitochondria of MTCH2 F/F Vav1-cre þ progenitors were significantly enlarged (Fig. 3a, left and middle panels). Interestingly, there was no significant change in the number of mitochondria per cell (Fig. 3a, right panel), which is consistent with the flow cytometric studies showing no change in mitochondria mass ( Supplementary Fig. 3). To obtain a more quantitative analysis of this phenotype, we prepared progenitor-enriched cells by magnetic cell sorting, followed by staining the mitochondria with Mitotracker Red (MitoRed). We then quantified the three-dimensional (3D) volume of mitochondria by computerized segmentation (Fig. 3b, left and middle panels). We found that the MTCH2 F/F Vav1-cre þ cells show an increase in mitochondrial volume as compared with the control MTCH2 F/F cells (Fig. 3b, right panel). To assess whether loss of MTCH2 in a more stem and progenitor-enriched population also results in an increase in mitochondrial volume, enriched cells were stained with anti-Sca-1 antibodies (Fig. 3c, left panel). As expected, Sca-1 þ population, comprising B10% of the enriched progenitor cells, showed an increase in mitochondrial volume in MTCH2 F/F Vav1-cre þ cells as compared with control cells (Fig. 3c, right panel). Taken together, these results indicate that loss of MTCH2 results in an increase in both mitochondrial OXPHOS and mitochondrial size/volume.
Aberrations in the mitochondria fusion/fission machinery 24 , can lead to an increase in mitochondria size and function. Indeed, it was recently demonstrated that improved mitochondrial function correlates with physiological enlargement of mitochondria 25 . Moreover, this mitochondria enlargement was due to less translocation to the mitochondria of the fission protein Drp-1 (ref. 25), which is an essential step in fission of the organelle 26 . Thus we assessed the levels of mitochondrial Drp-1 in MTCH2 F/F and MTCH2 F/F Vav1-cre þ HSPCs using Imaging Flow Cytometry. Bone marrow cells were stained for Drp-1 (green) and MitoRed, gated for LSK, and the co-localization score of the Drp-1 to the mitochondria was calculated using the Bright Detail Similarity (BDS) feature (Fig. 3d, left and middle panels; high BDS represents high Drp-1/MitoRed co-localization, yellow fluorescence, and low BDS represents low Drp-1/MitoRed co-localization, lack of yellow fluorescence). Utilizing Unpaired t-test, a significantly lower Drp-1/MitoRed co-localization   (lower BDS score) was estimated in the MTCH2 F/F Vav1-cre þ cells as compared to the MTCH2 F/F controls (Fig. 3d, right panel). Thus, MTCH2-deficient cells show less mitochondrial localization of the fission protein Drp-1, suggesting that the increase in mitochondria size and possibly function is due to mitochondrial hyperfusion. However, the precise role of MTCH2 in regulating the translocation of Drp-1 to the mitochondria still remains to be determined.

G-CSF mobilization increases mitochondrial OXPHOS and volume.
To test whether the increase in these mitochondrial parameters is physiologically relevant we analysed these parameters following HSC mobilization and entry into active cell cycle 27 . Mice were injected with granulocyte colony stimulating factor (G-CSF), the most commonly used mobilizing agent for stem cell transplantation 28 , and mitochondrial respiration and mitochondrial size/volume were determined. We found that G-CSF administration led to an increase in both basal and maximal oxygen consumption in enriched progenitors (Fig. 4a), and to an increase in mitochondrial volume (Fig. 4b), similar to the phenotypes observed in MTCH2 F/F Vav1-cre þ cells (Figs 2  and 3). Thus, an increase in mitochondrial OXPHOS and mitochondrial size/volume is associated with HSC mobilization and entry into active cell cycle.
Loss of BID phosphorylation primes mitochondrial OXPHOS. We next assessed the involvement of BID, the upstream regulator of MTCH2 (ref. 16), in regulating mitochondrial function. We previously demonstrated that loss of BID phosphorylation (BID AA mice), which results in an increase in mitochondrial BID, leads to HSC exit from quiescence 14 . Since both mouse models, BID AA and MTCH2 F/F Vav1-cre þ , show loss of HSC quiescence, we suspected that both models would also show similar changes in mitochondrial parameters. Indeed, we found an increase in both basal and maximal oxygen consumption in BID AA enriched progenitor cells (Fig. 5a). The increased respiration was accompanied by decreased NADH redox (due to accelerated electron consumption; Fig. 5b), and an increase in DCm (Fig. 5c), similar to the phenotypes observed in the MTCH2 F/F Vav1-cre þ cells (Fig. 2). Moreover, electron microscopy analyses of the BID AA progenitors demonstrated a significant increase in mitochondrial size (Fig. 5d). Thus, an 'activating' mutation in BID and loss of MTCH2 lead to similar phenotypes, arguing that BID acts as a MTCH2 antagonist.

Loss of MTCH2 protects cells from stress-induced death.
It is well established that physiological enlargement of mitochondria, which correlates with improved mitochondrial function/ATP production, protects cells from stress-induced apoptosis 25,29,30 . Since the ATM-BID couple is an important regulator of the DNAdamage response in the bone marrow 12,14,31 we tested the sensitivity of the MTCH2 F/F Vav1-cre þ mice to total body irradiation (TBI). Notably, we found that MTCH2 loss provides protection to HSPCs from irradiation-induced death in vivo (Fig. 6a), which was also evident in vitro (Fig. 6b). In accordance with these findings, the MTCH2 F/F Vav1-cre þ progenitors were resistant to TBI-induced caspase-3 activation/cleavage in vivo (Fig. 6c, Supplementary Fig. 4), and MTCH2 F/F Vav1-cre þ HSCs were considerably less sensitive to TBI-induced apoptosis (Fig. 6d). Finally, we assessed the levels of cellular ATP and as expected MTCH2-deficient progenitors possess considerably higher levels of ATP under basal conditions (Fig. 6e). Surprisingly, however, we found that the BID AA progenitors possess considerably lower basal levels of ATP (Fig. 6f), which seemed consistent with our previous findings that BID AA    irradiation 14 . Interestingly, a recent report 32 argued that an increase in mitochondrial OXPHOS can result in either a significant increase in ATP production, as is the case in the MTCH2 F/F Vav1-cre þ cells, or alternatively a significant increase in ROS production (possibly due to a proton leak), as is the case in the BID AA cells since they show a larger increase in ROS (B3fold 14 ) as compared to a B1.5-fold increase in the MTCH2deficient cells (Fig. 2f).

Discussion
In this study we demonstrate that mitochondrial OXPHOS, regulated by the ATM-BID-MTCH2 pathway, plays a critical role in determining the fate of HSC and progenitor cells (Fig. 6g). Our results demonstrate that loss of MTCH2 primes mitochondrial OXPHOS, which acts as a catalyser to drive HSCs into cycle. Since the expression of MTCH2 is considerably higher in restricted progenitor cells (LK cells) as compared with HSPCs (LSK cells), it is possible that MTCH2 suppresses OXPHOS in restricted progenitors and its loss primarily accelerates progenitor differentiation, which, indirectly, drives HSCs into cycle leading to loss of quiescence and finally exhaustion. These results seem opposite from what occurs in T lymphocytes, where resting cells are long lived and possess considerable SRC, whereas when they become activated and start to divide they lose their SRC 33 . Thus, the metabolic requirements for HSC maintenance seem to differ from the metabolic needs of T lymphocytes. The robust increase in mitochondrial OXPHOS in the MTCH2 F/F Vav1-cre þ cells correlates with an increase in both basal ATP and basal ROS levels that signal the dormant HSCs to cycle, and concurrently equip them with enough energy to survive increased stress (Fig. 6g). On the other hand, the robust increase in mitochondrial OXPHOS in the BID AA cells correlates with a significant increase in basal ROS levels and a decrease in basal ATP levels, which likely contributes to the hypersensitivity of these cells to genotoxic stress (Fig. 6g). It should be emphasized that loss of MTCH2 also results in less mitochondrial targeting of BID 16 , which likely contributes to the decreased sensitivity of the MTCH2-deficient cells to genotoxic stress-induced apoptosis.
Thus, we identified MTCH2 as a new downstream player in the ATM-BID pathway, at the mitochondria, that via tuning of OXPHOS dictates whether a stem cell will remain quiescent in the bone marrow niche or shift to a migratory cycling and differentiating state replenishing the blood with mature leukocytes.

Methods
Mice. The Weizmann Institute Animal Care and Use Committee approved all animal experiments. MTCH2 F/F mice were generated as described 16 and backcrossed for 10 generations to C57BL6 background. Vav1-Cre male mice (on a C57BL/6 background) were purchased from Jackson Laboratory and bred with MTCH2 F/F females to generate MTCH2 F/F Vav1-Cre þ mice. C57BL/6 CD45.1 congenic mice were from the Weizmann Institute of Science and C57BL/6 CD45.2 congenic mice were purchased from Harlan, Rehovot, Israel. BID AA mice were generated as described 14 . TBI experiments were performed as described 14 . All data presented were repeated in at least three independent experiments with 8-10-weekold age and sex-matched mice. Both males and females were used in this study.
Respiration. Freshly isolated total bone marrow cells were enriched for CD117 þ population using Magnetic cell sorting (MACS; Miltenyi Biotec 130-091-224) and 0.5 Â 10 6 cells were immobilized to XF 24 plate (Seahorse Bioscience, Cat. # 100777-004) pretreated with Cell-Tak (Corning; Cat#354240). Measurement of intact cellular respiration was performed using the Seahorse XF24 analyzer (Seahorse Bioscience Inc.) and the XF Cell Mito Stress Test Kit according to manufacturer's instructions and as described 34 . Respiration was measured under basal conditions, and in response to Oligomycin (ATP synthase inhibitor; 0.5 mM) and the electron transport chain accelerator ionophore FCCP (Trifluorocarbonylcyanide Phenylhydrazone; 0.5 mM). FCCP treatment gives two indexes: the Maximal OCR (Oxygen Consumption Rate) capacity of the cells and their spare respiratory capacity (SRC; indicated by maximum OCR calculated as percentage of baseline OCR). Finally, respiration was stopped by adding the electron transport chain inhibitors Rotenone and Antimycin A (1 mM each).
NADH. Mitochondrial NADH levels were estimated as described 23 . Briefly, mitochondrial NADH was calculated as the difference in arbitrary units between the maximum NADH autofluorescence (in response to 1 mM of potassium cyanide (KCN; Sigma), a potent inhibitor of complex IV resulting in a block in mitochondrial respiration and thus maximum accumulation of mitochondrial NADH); and minimum NADH autofluorescence (in response to 1 mM of the OXPHOS uncoupler FCCP (Seahorse Bioscience Inc.) that induces maximum respiration and thus exhaustion of the mitochondrial NADH pool). NADH redox index was estimated by calculating the initial NADH autofluorescence when the minimum NADH autofluorescence is normalized to 0% and the maximum to 100%. NADH autofluorescence was measured flow cytometrically after excitation with a UV laser with a main emission peak at 470 nm.  Annexin V staining. Mice were either left untreated or subjected to TBI and 0.5 h later mice were killed. Freshly isolated bone marrow was stained for flow cytometric analysis. Annexin V detection kit (BD Biosciences) was used according to manufacturer's instructions, coupled with DAPI (Sigma) staining for dead cells.
Caspase-3 activity. Mice were either left untreated or subjected to TBI and 0.5 h later total bone marrow cells were enriched for CD117 þ population using Magnetic cell sorting (MACS; Miltenyi Biotec 130-091-224). Enriched cells were lysed with Promega cell lysis buffer (E1531) and caspase activity was estimated using Caspase-Glo 3/7 Luminescent Assay kit (Promega) according to manufacturer's instructions.
Statistical analysis. P-values were calculated by using an unpaired Student's t-test.