Mitochondria transfer-based therapies reduce the morbidity and mortality of Leigh Syndrome

Abstract

Mitochondria transfer is a recently described phenomenon in which donor cells deliver mitochondria to recipient cells.^1–3 One possible consequence of mitochondria transfer is energetic support of neighboring cells; for example, exogenous healthy mitochondria can rescue cell-intrinsic defects in mitochondrial metabolism in cultured ρ⁰ cells or Ndufs4^–/– peritoneal macrophages.^4–7 Exposing hematopoietic stem cells (HSC) to purified mitochondria before autologous HSC transplantation allowed for treatment of anemia in patients with large-scale mtDNA mutations^8,9, and mitochondria transplantation was shown to minimize ischemic damage to the heart^10–12, brain^13–15, and limbs¹⁶. However, the therapeutic potential of using mitochondria transfer-based therapies to treat inherited mitochondrial diseases is unclear. Here, we demonstrate improved morbidity and mortality of the Ndufs4^–/– mouse model of Leigh Syndrome (LS) in multiple treatment paradigms associated with mitochondria transfer. Transplantation of bone marrow (BM) from wildtype (WT) mice, which is associated with release of hematopoietic cell-derived extracellular mitochondria into circulation, causes transfer of mitochondria to host cells in multiple organs and ameliorates LS in mice. Furthermore, administering isolated mitochondria from WT mice extends lifespan, improves neurologic function, and increases energy expenditure of Ndufs4^–/– mice, whereas mitochondria from Ndufs4^–/– mice did not improve neurologic function. Finally, we demonstrate that cross-species administration of human mitochondria to Ndufs4^–/– mice also improves LS. These data suggest that mitochondria transfer-related approaches can be harnessed to treat mitochondrial diseases, such as LS.

LS is a primary mitochondrial disease characterized by progressive neurologic deficits, muscle weakness, ataxia, and respiratory failure.¹⁷ Unfortunately, there are no therapies for LS, and the majority of children affected by this disease die by the age of 3-years-old.¹⁸ This disease is characterized by mutations in 113 genes that affect electron transport chain function, with one commonly affected gene being NDUFS4, a nuclear (n)DNA-encoded enzyme that is required for Complex I assembly and function.¹⁹ Mice with a loss-of-function mutation in Ndufs4 exhibit many of the features of LS and die within the first 2–3 months of life.²⁰ We hypothesized that engraftment of a WT hematopoietic system in Ndufs4^–/– mice would provide a systemically distributed, durable supply of healthy cells that could transfer their mitochondria to diseased cells to reduce the morbidity and mortality of LS.

To begin investigating this, we lethally irradiated 3–5-week-old Ndufs4^–/– mice and transplanted them with either WT or Ndufs4^–/– BM (Fig 1a). We found that WT BM transplantation significantly extended the lifespan of Ndufs4^–/– mice (WTKO, n=29, median survival of 74 days) compared to Ndufs4^–/– BM transplantation (KOKO, n=11, median survival of 40 days; Mantel-Cox log-rank: P<0.0001; Fig 1b). Disease severity was assessed 2–4-weeks after transplantation at an age of ~7-weeks, a time-point at which peripheral blood donor chimerism was on average 91.9% ± 2.9% (mean ± standard error of the mean [SEM]). Although there were subtle, non-significant improvements in body weight (Fig 1c) and core body temperature (Fig 1d), we observed significantly improved performance on a neurologic function test where mice are placed on a rotating rod (rotarod). Specifically, KOKO mice could not maintain balance and fell off the rotarod almost immediately, whereas the majority of WTKO mice were able to remain on the rotarod for ~1 min on average (Fig 1e) and tolerate faster rotational speeds (Fig 1f). Four-limb grip strength did not differ between the mice (Fig 1g). Next, mice were singly housed in metabolic cages equipped with an activity monitoring system and indirect calorimetry for ~3 hours after air equilibration. Whole-body energy expenditure was substantially increased in WTKO mice compared to KOKO controls (Fig 1h–i). The distance traveled exploring this new cage environment was notably higher in some WTKO mice, but this comparison did not reach statistical significance (Extended Data Fig. 1a). Food pellets and hydrogel were added to the cage bedding to ensure access to food and water during the experiment, making assessments of food and water intake unreliable, therefore we did not monitor these parameters. However, respiratory exchange ratio (RER), an indicator of energy substrate utilization, did not differ between groups (Extended Data Fig. 1b). Collectively, these data indicate that BM transplantation from healthy, WT donors is associated with improved morbidity and mortality in a mouse model of LS.

Figure 1. Wildtype bone marrow transplantation improves the morbidity and mortality of Leigh Syndrome in NDUFS4-deficient mice.

(a) Experimental design. Ndufs4^—/— (KO) mice were lethally irradiated at 3–5-weeks and transplanted with wildtype (WT) or KO bone marrow cells. (b) Survival expressed as postnatal age in days. Solid black line is WT to KO, dashed blue line is KO to KO. (c) Body weight, (d) rectal core body temperature, (e) time until falling off a rotarod, (f) rotarod rotations per minute (rpm) at the time of falling (starting at rpm=4), and (g) four-limb grip strength at age 7-weeks. (h) Energy expenditure shown per interval. Closed black circles are WT to KO, open blue squares are KO to KO (i) Average energy expenditure. Data are expressed as mean +/− SEM. All data points are unique biological replicates, and all statistical tests are two-sided. For b, n=29 WT to KO, n=11 KO to KO. For c-i, n=5 KO to WT, n=17 WT to WT. For b, Mantel-Cox log-rank test. For e-f, i, Student’s t-tests. For h, two-way ANOVA with repeated measures. *P<0.05, **P<0.01.

To test whether the transplanted hematopoietic cells release mitochondria that are transferred to diseased host cells, we transplanted CD45.1 WT or CD45.1 mtD2^+/– bone marrow into lethally irradiated Ndufs4^–/– mice that naturally have the CD45.2 allele (Fig 2a). mtD2 is a nuclear-encoded Dendra2 fluorescent protein attached to the mitochondria targeting sequence of cytochrome c oxidase subunit VIII²¹ and is a well-established mitochondria reporter system that enables detection of mitochondria transfer by flow cytometric analyses.^4,22 First, we verified successful engraftment of CD45.1 WT and CD45.1 mtD2^+/– immune cells by flow cytometry of peripheral blood 2 weeks post-transplantation (Fig 2b and Extended Data Fig 2a), and found similar degrees of chimerism between the two groups (Extended Data Fig 2b). Next, we used small-particle flow cytometry to quantify extracellular mtD2⁺ mitochondria in blood, as we described previously using submicron size calibration beads and antibodies against CD41, TER119, and CD45 to exclude platelets, red blood cells, and immune cell fragments, respectively⁴. We observed that 3.9% (SEM 1.6%) of the submicron particles were mtD2⁺ in mtD2KO mice compared to 0.0% (SEM 0.0%) in the WTKO controls (Fig 2c and Extended Data Fig 2c). Quantitatively, this corresponded to ~3.5 × 10⁵ extracellular mitochondria per mL of blood (Fig 2d). These results are consistent with prior studies indicating that there are extracellular mitochondria in the blood of mice and humans^23,24 that can be redistributed systemically to distant organs.^4,25

Figure 2. Engrafted hematopoietic cells release mitochondria into the blood and transfer mitochondria to NDUFS4-deficient cells in vivo.

(a) Experimental design. Ndufs4^—/— (KO) mice were lethally irradiated and transplanted with bone marrow cells from CD45.1 WT or CD45.1 mtD2 mitochondria reporter mice. (b) Flow cytometric identification of CD45.1⁺ CD45.2^— donor immune cells and CD45.1^— CD45.2⁺ radioresistant host immune cells. Histograms show mtD2 signal in the host and donor cells within each group. (c) Flow cytometry plots of extracellular mtD2⁺ mitochondria in peripheral blood. Pre-gated on CD41^— CD45^— TER-119^— events less than 2 μm in diameter. (d) Numbers of mtD2⁺ extracellular mitochondria (ex-mito) per mL of peripheral blood. (e) Proportions of singlet, live, CD45.1^— CD45.2⁺ radioresistant host immune cells and (f) CD45.1^— CD45.2^— host non-immune cells that received mtD2 signal from donor cells in the indicated tissues. (g) Proportions of host B cells (B), T cells (T), neutrophils (Neut), monocytes (Mono), epithelial cells (Epi), endothelial cells (Endo) and stromal cells that received mtD2 signal from donor cells in the spleen. For d-g, closed circles are WT, and open squares are mtD2. Data are expressed as mean +/− SEM. All data points are unique biological replicates, and all statistical tests are two-sided. For c-d n=7/group. For e-g, n=3 WT, n=5 mtD2 (except n=4 for spleen). For c-d, Mann-Whitney test. For e-g, two-way ANOVA with Sidak post-hoc test. **P<0.01, ***P<0.001, ****P<0.0001.

Recent studies also indicate that immune cells that infiltrate tissues can also transfer mitochondria to parenchymal cells, including transfer of mitochondria from macrophages to neurons in the dorsal root ganglia.²⁶ To test whether hematopoietic cells transfer mitochondria to host cells in the setting of BM transplantation, we harvested tissues 5-weeks post-transplant and observed that ~15–20% of radioresistant host immune cells were mtD2⁺ in blood, spleen, and liver (Fig 2e). Furthermore, nearly 100% of splenic non-immune cells and ~20% of liver non-immune cells were mtD2+ (Fig 2f). In the spleen, the vast majority of Ndufs4^–/– host-derived B cells, neutrophils, epithelial cells, endothelial cells, and stromal cells were mtD2⁺, with significant but lower mtD2+ frequencies in host T cells and monocytes (Fig 2g and Extended Data Fig 3a–3b). In the liver, the majority of Ndufs4^–/– host-derived B cells and a substantial proportion of epithelial, endothelial, and stromal cells were mtD2⁺, with lower frequencies of T cells, neutrophils, and monocytes that were mtD2⁺ (Extended Data Fig 3c). In the blood, nearly all of the Ndufs4^–/– host-derived B cells and neutrophils were mtD2⁺, with a minor fraction of T cells and monocytes being mtD2⁺ (Extended Data Fig 3d). These data indicate that donor-derived hematopoietic cells provide a source of healthy, cell-free mitochondria, and that BM transplantation is associated with mitochondria transfer to diseased cells in Ndufs4^–/– mice with tissue-specific patterns.

Next, we sought to determine whether inducing mitochondria transfer in the absence of BM transplantation was associated with improved morbidity and mortality in Ndufs4^–/– mice. To accomplish this, we isolated mitochondria from mtD2 mouse livers and confirmed high enrichment of mtD2⁺ mitochondria (94.9 ± 1.9%, Extended Data Fig 4a) and substantially higher yields from liver compared to bone marrow (Extended Data Fig 4b). Based on this higher yield and much faster isolation of mitochondria from liver than bone marrow, we administered 100 μg liver mitochondria or PBS by intraperitoneal injection 1–2 times per week to Ndufs4^–/– mice. Notably, treatment with purified mitochondria increased the survival of Ndufs4^–/– mice (Fig 3a), with median survival of 68 days and 84 days for mice treated with PBS or mitochondria, respectively (Mantel-Cox log-rank: P=0.0052). Body weight (Extended Data Fig 5a) and core body temperature (Extended Data Fig 5b) did not differ between groups, and in many instances the mice treated with purified mitochondria rapidly regrew their fur coats (Fig 3b and Extended Data Video 1). Mice treated with purified mitochondria also exhibited improved balance on a stationary rotarod (Fig 3c) and a ~20% increase in grip strength (Fig 3d). Metabolic cage analyses revealed increased energy expenditure in Ndufs4^–/– mice treated with purified mitochondria (Fig 3e and Extended Data Fig 5c), whereas distance traveled during the 3-hr period did not differ between groups (Extended Data Fig 5d). RER values did not significantly differ in mitochondria-treated mice, suggesting there was no change in energy substrate utilization (Extended Data Fig 5e). These data indicate that mitochondria transplantation improves the morbidity and mortality of mice with LS-like disease.

Figure 3. Administration of exogenous mouse mitochondria reduces the morbidity and mortality of Leigh Syndrome.

(a) Experimental design for a-e. Survival expressed as postnatal age in days of Ndufs4^—/— (KO) mice treated 1–2 times per week with PBS (dashed blue line) or 100μg WT mitochondria (Mito, solid black line). (b) Pictures of mice, (c) time until falling off a stationary rotarod, (d) four-limb grip strength, and (e) energy expenditure at 6–8 weeks of age. (f) Basal oxygen consumption rate (OCR) of isolated KO or WT mitochondria. (g) Time until falling off a stationary rotarod by KO mice were treated once a week with 100 μg KO or WT mitochondria for 4 weeks. (h) Tissue biodistribution of NZB mtDNA in Ndufs4^—/—(KO) mice treated once a week with 100 μg NZB mitochondria for 4 weeks. Background amplification was corrected for with KO mice treated with 100 μg C57BL6/J mitochondria for 4 weeks. Samples were normalized to β-actin as a nuclear DNA control. Data are expressed as mean +/− SEM. All data points are unique biological replicates, and all statistical tests are two-sided.For c, n=15 PBS, n=19 Mito. For d, n=11 PBS, n=12 mito. For e, n=15/group. For d-e, variation in n is due to mortality kinetics. For f, n=6/group. For g, n=9 KO mitochondria, n=8 WT mitochondria. For h, n=6/tissue. For a, Mantel-Cox log-rank test. For c-d, f-g, Student’s t-tests. For e, Mann-Whitney test. *P<0.05.

We previously showed that exogenous mitochondria can rescue cell-intrinsic defects in aerobic respiration in macrophages and BV2 cells, and this response is dependent on expression of NDUFS4 by the donor mitochondria.⁴ To test whether mitochondria transplantation ameliorates LS in an NDUFS4-dependent manner, we isolated WT or Ndufs4^–/– liver mitochondria for administration to Ndufs4^–/– mice. The oxidative consumption rate was significantly lower in isolated mitochondria obtained from Ndufs4^–/– livers compared to WT controls (Fig 3f), demonstrating severe impairment of oxidative phosphorylation. Notably, 6 weeks of weekly intraperitoneal injection with 100 μg WT but not Ndufs4^–/– mitochondria improved balance on a stationary rod (Fig 3g). These data indicate that the therapeutic benefit of mitochondria transplantation in LS is dependent on NDUFS4, suggesting that bioenergetic effects at least partially contribute to the observed improvement in neurologic function.

To determine the tissue biodistribution of the transplanted mitochondria, we used isolated mitochondria from the livers of C57BL6/J or NZB/BINJ (NZB) mice, which have an mtDNA genome that has >76 polymorphisms compared to the C57BL6/J reference mtDNA genome and that can be detected using Amplification Refractory Mutation System (ARMS)-PCR.²⁷ This approach was used instead of mtD2-labeled mitochondria because the mtDNA genome can be maintained in recipient cells, whereas the mtD2 protein cannot. We administered 100 μg of C57BL6/J or NZB mitochondria by intraperitoneal injection once per week to Ndufs4^–/– mice for 5 weeks and then harvested various tissues, with peritoneal exudate cells (PECs) serving as a positive control. We observed NZB mtDNA in the PECs, blood, spleen, and bone marrow, whereas the signal was below the limit of detection in the liver, lung, heart, and brain (Fig 3h). These analyses were normalized to β-actin nDNA content to account for cellularity in the tissue biopsies. These results suggest that the mtDNA from mitochondria transplanted via intraperitoneal injection is well distributed in lymphoid-related tissues.

Several studies indicate that uptake of healthy extracellular mitochondria by immune cells such as macrophage and T cells has anti-inflammatory effects.^28–31 Indeed, it was recently reported that inflammation contributes to LS disease pathogenesis.^32–34 Therefore, we hypothesized that mitochondria transplantation might attenuate inflammation in Ndufs4^–/– mice. To assess this possibility, first we compared the peripheral blood immune cell composition in Ndufs4^–/– mice treated with PBS or 100 μg mitochondria weekly for 5 weeks (Extended Data Fig 6a). We found that there were no changes in the numbers per mL of B cells, eosinophils, natural killer (NK) cells, neutrophils, monocytes, CD4 T cells, or CD8 T cells between the two groups (Extended Data Fig 6b). Second, we treated Ndufs4^–/– mice with WT or KO mitochondria for 6 weeks and observed no differences in the numbers of these 7 immune cell populations in the blood (Extended Data Fig 7a), spleen (Extended Data Fig 7b), bone marrow (Extended Data Fig 7c) or liver (Extended Data Fig 7d). Third, we compared the expression of inflammatory cytokines in the spleen (Extended Data Fig 8a), PECS (Extended Data Fig 8b), and liver (Extended Data Fig 8c), and found that the relative expression of Interleukin (IL)-6, Tumor necrosis factor (TNF)-⍶, Interferon (IFN)-γ, IL-18, and IL-4 did not differ in Ndufs4^–/– mice treated with WT or KO mitochondria weekly for 6 weeks. These data suggest that weekly administration of isolated mitochondria to Ndufs4^–/– does not attenuate or enhance inflammation, although we cannot formally exclude a potential effect of mitochondria transplantation on inflammatory responses.

It is known that human mitochondria can fuse with endogenous mitochondria in mouse cells, although human proteins are not maintained over longer periods of time.³⁵ Therefore, in a translational effort, we tested whether xenogenic transplantation of human mitochondria is associated with increased survival in Ndufs4^–/– mice. We obtained HeLa cell-derived human Mitochondria Organelle Complex-Q (MRC-Q, LUCA Science, Inc.), a cell-free product enriched in mitochondria that can be frozen to support distribution and storage and then thawed prior to administration. Compared to human mitochondria isolated from the same cellular source with traditional homogenization and differential centrifugation enrichment methods (H-mito), thawed MRC-Q exhibited higher ATP production that was suppressed after adding the ATP Synthase inhibitor oligomycin (Omy) (Fig 4a) and had substantially better integrity of the outer mitochondrial membrane (OMM, Fig 4b) and inner mitochondrial membrane (IMM, Fig 4c). These data indicate that freeze-thawed MRC-Q is highly enriched in intact mitochondria that retain their ability to respire. We administered 50 μg MRC-Q to 3-week-old Ndufs4^–/– mice intravenously once per week for 4 weeks and observed human mtDNA in various organs including in the blood, spleen, liver, bone marrow, heart, lung, and brain after perfusion with PBS, with the highest signal observed in the lung (Fig 4d). Each sample was normalized to mouse nDNA content to account for cellularity.

Figure 4. Administering a human mitochondria isolate called MRC-Q improves the morbidity and mortality of Leigh Syndrome.

(a) Production of ATP in the presence of DMSO or oligomycin (Omy) by mitochondria isolated from HeLa cells by a commercial kit (H-mito) or mitochondria organelle complex Q (MRC-Q). (b) Outer mitochondrial membrane (OMM) and (c) inner mitochondrial membrane (IMM) integrity. (d) Tissue biodistribution of MRC-Q mtDNA in Ndufs4^—/— (KO) mice treated with 50 μg MRC-Q once per week for 4 weeks. Background amplification was corrected for with KO mice treated with PBS for 4 weeks. Samples were normalized to β-actin as a nuclear DNA control. (e) Survival expressed as postnatal age in days. MRC-Q, solid black line. PBS, dashed blue line. (f) Pictures of mice, (g) time until falling off a hang-wire at age 7-weeks-old, and (h) righting reflex delay when mice were placed on their backs. Data are expressed as mean +/− SEM. All data points are unique biological replicates, and all statistical testing is two-sided. For a and d, n=7/group. For c, n=6/group. For d, n=3/tissue. For g, n=10 PBS, n=14 MRC-Q. For h, n=10/group. For a, two-way ANOVA. For b-c and g-h, Student’s t-tests. For e, Mantel-Cox log-rank test. *P<0.05, ****P<0.0001.

Next, we treated Ndufs4^–/– mice with PBS or MRC-Q intravenously and found that MRC-Q treatment significantly improved the survival of Ndufs4^–/– mice (median survival 80.5 days) compared to PBS-treated controls (median survival 67.5 days, Mantel-Cox log-rank: P=0.0079, Fig 4e). Body weights did not differ between groups (Extended Data Fig 9), however MRC-Q-treated mice rapidly re-grew their fur (Fig 4f), exhibited increased capacity to hang onto a wire (Fig 4g), and had a faster righting reflex (Fig 4h and Extended Data Video 2) compared to PBS-treated controls. MRC-Q-treated mice also had improved stability and posture when suspended by the tail and landed on four limbs when released from low height to the cage bedding (Extended Data Video 3). In contrast, PBS-treated mice gyrated unstably during tail suspension and landed on their backs after release to the cage bedding (Extended Data Video 3). These data suggest that intravenous MRC-Q human mitochondria are distributed to various tissues and ameliorate the morbidity and mortality of Ndufs4^–/– mice.

These studies reveal that mitochondria transfer pathways can be harnessed therapeutically to treat inherited mitochondrial diseases, such as LS. First, transplantation of WT BM to NDUFS4-deficient mice ameliorates LS and is associated with delivery of donor cell-derived mitochondria to diseased host cells. Second, mitochondria transplantation recapitulates many of the beneficial outcomes observed following WT BM transplantation. A comparison of the three therapeutic paradigms is provided in Extended Data Table 1. We reported true survival, therefore our control groups have longer median survival than other studies that utilize euthanasia criteria.^32,36 True survival was determined because some euthanasia criteria are subjective and/or transient (e.g. immobility) or do not reflect future lifespan accurately (e.g., low core body temperature).

The mechanisms by which mitochondria transfer-related therapies ameliorate LS are unclear. Although this study does not provide direct evidence that the uptake of donor mitochondria alters the bioenergetics of acceptor cells in vivo, we speculate that a bioenergetic mechanism is plausible. This possibility is supported by four lines of evidence. First, transplantation of mitochondria from WT but not Ndufs4^–/– mice ameliorated LS in our studies. Second, healthy mitochondria can be taken up and utilized to rescue cell-intrinsic defects in aerobic respiration in Ndufs4^–/– peritoneal macrophages in vivo and in ρ⁰ cells that lack mtDNA in vitro.^4–7 Third, BV2 cells exposed to a brief pulse of rotenone and antimycin A had impaired aerobic respiration that could be rescued by uptake of WT but not Ndufs4^–/– mitochondria.⁴ Fourth, Complex I regenerates NAD⁺ during oxidative phosphorylation and is critical for ATP production. It is possible that delivery of NAD⁺, ATP, or other mitochondrial metabolites ameliorates LS. Indeed, prior studies showed that administering NAD⁺ precursors (e.g., nicotinamide mononucleotide) or expressing the yeast NADH dehydrogenase NDI1 to restore NAD⁺ synthesis decreases LS disease severity in Ndufs4^–/– mice.^37,38

The Ndufs4^–/– mouse develops prominent central nervous system (CNS) dysfunction, and we demonstrate using rotarod assays that neurological function is improved following bone marrow or mitochondria transplantation. However, our tissue biodistribution studies revealed no, or at best very little, donor-derived mtDNA in the brains of Ndufs4^–/– mice. This is a puzzling result and raises additional questions about how neurological function and survival could be improved in this context. One possibility is that the exogenous mitochondria might be taken up by cells in the peripheral nervous system, including peripheral neurons, which would not be detected in our flow cytometric assays. This could potentially improve peripheral neuropathy or reflex dysfunction that occur in Leigh Syndrome.³⁹ Indeed, we observed that mitochondria transplantation improved the righting reflex of Ndufs4^–/– mice. A second possibility is that the exogenous mitochondria may induce expression of antioxidant enzymes that quench reactive oxygen species (ROS) and limit accumulation of oxidative stress-associated cellular damage in diseased cells, a process that was shown to occur in the heart.²⁵ Third, recent studies demonstrate that transplanted mitochondria elicit mitophagy¹⁶, raising the possibility that mitochondria transfer-related therapies ameliorate LS by enhancing mitophagy to reduce or delay the accumulation of damaged mitochondria. These potential mechanisms for how mitochondria transfer-based therapies ameliorate LS should be explored in future studies.

There are several limitations of this study. Notably, the tissue biodistribution studies of donor mtDNA were normalized to mouse nDNA content to account for cellularity. We did not determine the percentage of total mtDNA that was donor mitochondria-derived because (1) all cells in tissues have mtDNA but not all of those cells take up donor-derived mitochondria, (2) some cell types take up mitochondria much more efficiently than others, and (3) the mtDNA copy number per cell can vary dramatically across tissues. This precludes us from quantifying how much mtDNA or many mitochondria were taken up in the various tissues, and we have not defined which cell types take up the transplanted mitochondria. Therefore, the exact mechanisms of how transplanted mitochondria improve survival and ameliorate LS severity remain unclear. Future studies are needed to determine the cellular distribution of transplanted mitochondria, the bioenergetic effects of mitochondria transfer on individual cell types in vivo, and the fate and maintenance of the transplanted mitochondria.

Given the complexity of mitochondria transfer pathways and the cell type-specific effects of this process, it is likely that mitochondria transfer-based therapies for LS or other diseases will involve multifaceted mechanisms of action. An exciting future direction will be to determine the safety and efficacy of heterologous hematopoietic stem cell or mitochondria transplants in clinical trials for patients with primary mitochondrial diseases, such as LS.

METHODS

Mice

Experiments performed at Washington University School of Medicine (WUSM) utilized the following mouse strains procured from Jackson Laboratories: wildtype C57BL6/J (strain number 000664), CD45.1 (Ptprca, strain number 002014), PhAM^excised (also known as mtDendra2 or mtD2, strain number 018397), and Ndufs4^+/– (strain number 027058), and New Zealand Black (NZB, strain number 000684). CD45.1 mtD2^+/– mice were generated by crossing mtD2 mice onto a CD45.1 homozygous background. The other strains are on a CD45.2 background. Experiments performed at Osaka University (OU) utilized the following mouse strains procured from Clea Japan unless otherwise noted: wildtype C57BL6/J, Ndufs4^+/– (Jackson Labs via LUCA Science, Inc.), and CD45.1 (Jackson Labs). All mice were on a C57BL6/J background, except for NZB mice. Ndufs4^–/– mice were generated by crossing heterozygous males and females from Jackson Labs stock as duos (1 female and 1 male) or trios (2 females and 1 male per cage), with timed breeding to generate many litters born within a few days of each other. Heterozygotes from that cross were used for subsequent breeding. All mice were housed in a Specific Pathogen Free (SPF) barrier facility at room temperature with a 12h:12h light:dark cycle, with lights on at 6am and lights off at 6pm, and had access to food and water ad libidum. The animal facilities were maintained at room temperature (range of 20–26 °C at WashU, range of 21.5–24.5 °C at Osaka University) with a target humidity setpoint of 30–70% at WashU and 45–65% at Osaka University. Male and female Ndufs4^–/– mice were used interchangeably and in similar proportions with treatment starting at age 3–5-weeks-old. True survival was determined as the actual date of death minus the date of birth without using euthanasia criteria. In order to harvest tissues, separate experiments were performed from survival curve studies, and mice were euthanized using isoflurane asphyxiation before tissue procurement. In some cases, tail vein blood was collected from awake mice. Studies performed at WUSM were approved under Institutional Animal Care and Use Committee (IACUC) protocol 22–0286. Studies performed at OU were approved by the Institutional Animal Care and Use Committee at Osaka University Graduate School of Medicine (Approval number: 30–096-013).

Bone marrow transplantation

Ndufs4^–/– recipient mice aged 3–5-weeks-old were lethally irradiated with 650–800 rads using an X-ray irradiator in an SPF barrier facility. Immediately after irradiation, donor bone marrow was isolated from the femurs and tibias from either CD45.1 Ndufs4^+/+ (WT), CD45.1 Ndufs4^–/–, or CD45.1 mtD2^+/– mice. Bone marrow cells were isolated from the marrow cavity by cutting the bone tips and placing them inside sterile 600 μL polypropylene tubes (Eppendorf) with a hole poked into the bottom with a 16G needle. The 600 μL tube and bones were placed inside sterile 1.5 mL polypropylene tubes (Eppendorf) and centrifuged at 8,000 × g at 4 °C for 15 sec. Bone marrow cell collection was verified by observing a red cell pellet and whitened bone shafts. The bones were discarded, and the cells were resuspended in 1 mL ACK RBC Lysis Buffer (Gibco) for 5 min at room temperature to lyse mature red blood cells. The lysis reaction was quenched with 10 mL sterile Wash Buffer, comprised of high glucose DMEM (Gibco or Corning) supplemented with 5% heat-inactivated fetal bovine serum (FBS), 1X Penicillin/Streptomycin (Gibco) and 2 mM L-glutamine (Gibco). The cells were centrifuged at 500 × g at 4 °C for 5 min in a swinging bucket centrifuge (Eppendorf 5810R), supernatant was aspirated, and the cell pellet was resuspended in 10 mL sterile PBS. The cells were centrifuged again as described above and resuspended in 2–5 mL sterile PBS for cell counting using 1:1 ratios of cells and 0.1% trypan blue in PBS, with live cell identification using either a Countess FL or hemacytometer. Cell counts were performed in triplicate. The cells were pelleted by centrifugation, as described above, the supernatants were aspirated, and the cells were resuspended in PBS at a final concentration of 2 × 10⁸ cells/mL. Male and female cells were processed separately and transplanted to sex-matched recipients under 1–2% isoflurane anesthesia by retro-orbital injection at volumes of 50 μL (10 million cells per recipient). The mice were returned to their home cage over a warming pad until fully recovered. Water bottles containing 3.3% sulfamethoxazole/trimethoprim oral suspension USP (STM, Aurobindo) were provided for 2 weeks, with remixing every 1–2 days and replacement after 7 days. The STM drinking water was made by adding 8 mL STM oral suspension (200 mg sulfamethoxazole and 40 mg trimethoprim per 5 mL) to a 240 mL bottle of autoclaved mouse water. Engraftment was assessed by collecting 5 μL of blood from the tail vein with milking and performing flow cytometric analyses, as described below.

Murine mitochondria isolation

Murine mitochondria were isolated using the Mitochondria Isolation Kit, Mouse Tissue (Miltenyibiotec). Briefly, Ndufs4^+/+, Ndufs4^—/—, or NZB mice were euthanized and perfused with 10 mL sterile PBS using a peristaltic pump via the heart’s left ventricle. Perfused livers were harvested, finely minced, and transferred to a 2mL glass dounce homogenizer. A 1 mL aliquot of ice-cold Lysis Buffer (Miltenyibiotec) containing 10 μL of 100X Halt Protease Inhibitor Cocktail (Thermo Fisher) was added, and 8 strokes were performed with twisting downward and straight upward movements. The homogenate was transferred to sterile 15 mL conical tubes, and 9 mL of 1X Separation Buffer was added. The specimen was mixed by gentle inversion, and then 50 μL of anti-TOM20 Microbeads were added. The specimen was wrapped in foil and rocked at 4 °C for 60 min. The sample was pulse-centrifuged at 500 × g at 4 °C for ~5 seconds to pellet large debris and then held on ice as the supernatant was passed over an LS Column attached to a QuadroMACS magnet with stand. The entire supernatant volume was passed through the column and then washed 3 times with 2 mL ice-cold 1X Separation Buffer. After the 3rd wash, the LS Column was removed from the magnet and the bound mitochondria were eluted into a new 15 mL conical tube using 1.5 mL ice-cold 1X Separation Buffer with gentle LS Column plunger force to avoid bubbles. The eluted mitochondria were centrifuged at 13,000 × g at 4 °C for 2 min to pellet the mitochondria. Supernatants were discarded and the mitochondria were resuspended in 1–1.5 mL Storage Buffer. Protein concentration was measured using 10 μL of the mitochondria isolate and 250 μL Coomassie Reagent Plus (Thermo Fisher), with bovine serum albumin (Pierce) serial dilutions as a standard curve. Mitochondria isolates were aliquoted, centrifuged at 13,000 × g for 2 min, and brought to concentrations of 500 μg/mL in ice-cold PBS. Mice were then treated with 200 μL (100 μg) of the mitochondria isolates or PBS by intraperitoneal injection.

Neurologic performance testing

Four-limb grip strength was measured using a Bioseb grip-strength meter with a mesh grid (BIO-GS3). Consistent, smooth force was applied using Bioseb Acquisition Software (BIO-CIS version 1.5.1.0(En), Bioseb) guidance, and the maximum force at the time of release was recorded. Mice were placed on stationary rotarods equipped with 8 infrared lasers per lane to determine the time at which the mouse fell off. The rotation speed ramped up gradually from 4.0 rpm with a rate of increase of 1.2 rpm per min. If one group fell off the stationary rod prior to starting the rotations, the rod was left stationary for the other group. For the hanging wire test, mice were hung from a horizontal rod (1 cm in diameter, at a height of 20 cm) by their forelimbs. After hanging, the time till fall was measured for up to 30 seconds.

Metabolic cages

Mice were weighed, and core body temperature was measured using a RET-3 mouse rectal probe and Fluke 51-II handheld thermometer. The mice were then singly housed for 4 hours in a 16-cage Comprehensive Laboratory Animal Monitoring System (CLAMS, Columbus Instruments, Columbus, OH). Energy expenditure and RER were calculated as previously described.²² The enclosure temperature was set at 22.2 °C with a 12h:12h light:dark cycle with lights on at 6am and lights off at 6pm. Air exchange was equilibrated for one-hour prior to recording. Food pellets and hydrogel were placed on the bedding because the mice could not reach the hanging feeder or water bottle easily, precluding measurement of food or water intake. Distance traveled was determined using an array of infrared lasers and Oxymax software version 5.66 (Columbus Instruments).

Flow cytometric detection of extracellular mitochondria in blood

Peripheral blood was sampled from the tail vein, and 5 μL blood was transferred to 245 μL ACK RBC Lysis Buffer containing 1 mg/mL heparin (Grade 1A, Sigma-Aldrich) with gentle mixing. The cells were pelleted by centrifugation at 500 × g at 4 °C for 5 min. The cell-free supernatant was collected (200 μL, equal to 80% of the total volume) for small-particle flow cytometry to detect extracellular mitochondria in blood, as previously described.⁴ In brief, the cell-free fraction was pelleted at 15,000 × g at 4 °C for 5 min, and the supernatant was removed before resuspending the pellet in 50 μL staining buffer containing the following antibodies: BV421-CD41 (MWReg30, Biolegend, dilution factor 1:300), PB-TER-119 (TER-119, Biolegend, 1:300), PerCP/Cy5.5-CD45 (30-F11, Biolegend, 1:200). The particles were washed by adding 1 mL FACS Buffer (PBS containing 2.5% heat-inactivated FBS and 2.5 mM EDTA followed by filter-sterilization) and centrifuging the samples at 15,000 × g at 4 °C for 5 min. The pelleted particles were resuspended in a final volume of 200 μL FACS buffer, and flow cytometry was performed on a 4-laser Cytek (16V-14B-10YG-8R configuration using SpectroFlo v2.0 software) with small particle settings (FSC = 675, SSC = 235–255, threshold = 6,000). Submicron size calibration beads loaded with FITC (ThermoFisher) were used to identify particles under 2 microns and to set gains to detect the smallest possible particles, with a 300 nM lower limit of detection for this cytometer. We then acquired 100 μL of each specimen. Extracellular mitochondria were defined as CD45^– CD41^– TER-119^– particles <2 microns in diameter and mtD2⁺. Mitochondria percentages were defined as the percentage of all CD45^– CD41^– TER-119^– particles that were mtD2⁺. Cell-free mitochondria counts in 1mL of blood were defined as total mtD2⁺ events × 2 (to account for acquiring half the stained specimen) × 1.25 (to account for processing 80% of the original specimen’s volume) × 200 (to account for collecting 5 μL of whole blood).

Flow cytometric analyses of cells

In some cases, peripheral blood from the tail vein was collected, as described in the section on flow cytometric detection of cell-free mitochondria in blood, however pelleted cells were collected for processing. Mice were perfused with 10 mL PBS, as described above, and the spleens and livers were dissected. Spleen cells were obtained by mashing them through a 100-μm strainer using a sterile syringe plunger and washed with Wash Media. To isolate cells from the bone marrow (BM), the soft tissue was removed from one femur and tibia, an BM cells were isolated using the centrifugation method described above. The pelleted BM was resuspended in Wash Media. Liver non-parenchymal cells were isolated by finely mincing the livers using a razor blade and processed according to established protocols.⁴⁰ Briefly, minced livers were digested in high glucose DMEM containing 0.75 mg/mL collagenase A (Sigma) and 50 μg/mL DNAse I (Sigma) for 30 min in an orbital shaker at 140 rpm at 37 °C. Liver digests were centrifuged at 50 × g at 4 °C for 3 min to pellet hepatocytes. Supernatant containing non-parenchymal (NP) cells was carefully pipetted into a fresh conical tube. Spleen, liver NP, and BM cells were resuspended in 1–2 mL ACK RBC Lysis Buffer and incubated at room temperature for 5 min. The reaction was quenched by adding 10 mL Wash Media. The cells were pelleted and collected in 96-well round-bottom plates for staining. The cells from blood, spleen, liver, and/or bone marrow were spun in a swinging bucket centrifuge at 500 × g at 4 °C for 5 min, and the supernatants were flicked off. Cells were washed once in 200 μL PBS, followed by immediate centrifugation at 500 × g at 4 °C for 3 min. The supernatants were discarded, and the cells were resuspended in 50 μL PBS containing Zombie NIR viability dye (1:1,000 in PBS, BioLegend) and incubated on ice for 5 min. The reaction was quenched by adding 200 μL FACS Buffer. The cells were pelleted at 500 × g at 4 °C for 5 min, and then resuspended in 25 μL Mouse BD Fc Block (anti-CD16/32, BD Pharmigen) at 1:100, diluted in FACS Buffer. After 15 min on ice, a 25 μL aliquot of a 2X antibody cocktail in Brilliant Stain Buffer (BD) was added to the cells and mixed by gently pipetting up and down. The cells were stained with PE/Cy7-CD45.2 (clone 104, Biolegend, dilution factor 1:200) and APC-CD45.1 (clone A20, Biolegend, 1:200) to identify donor and host immune cells. Immune cell populations were further characterized with the following antibodies: PerCP-CD45 (30-F11, Biolegend, 1:200), BV421-SiglecF (E50–2440, BD, 1:400), Spark-NIR685-CD19 (6D5, Biolegend, 1:300), BV750-B220 (RA3–6B2, Biolegend, 1:300), APC/Fire810-NK1.1 (S17016D, Biolegend, 1:300), AF488-TCRβ (H57–597, Biolegend, 1:300), PE/Fire640-CD4 (GK1.5, Biolegend, 1:300), SB550-CD8 (53–6.7, Biolegend, 1:300), PB-CD11b (M1/70, Biolegend, 1:400), BV711-Ly6G (1A8, Biolegend, 1:300), AF700-Ly6C (HK1.4, Biolegend, 1:300). Cells were stained for 30 min on ice and then washed twice with 200 μL FACS Buffer prior to final resuspension in a volume of 200 μL. For each specimen, 100 μL cell suspension was acquired on a Cytek Aurora (SpectroFlo v2.0 software). Analyses of flow cytometry data were conducted with FlowJo (BD) version 10.8 or 10.9 software.

Seahorse assay

Mitochondria were isolated from the livers of WT or Ndufs4^–/– mice and 50μg of mitochondria were plated onto 8-well Agilent Seahorse XF HS PDL Miniplates (Agilent Technologies). Mitochondria were centrifuged at 2,000 × g for 20 min at 4˚ C and mitochondria seahorse assay media (1mM Pyruvate, 2mM Glutamine, 10mM Glucose, 70mM Sucrose, 18mM EDTA, and 0.2% w/v BSA in Seahorse media) was added to each well. Mitochondrial basal oxygen consumption rate was measured using an Agilent Seahorse XFp HS Mini machine with software version 3.0.0.25.

MRC-Q reagent collection, preparation, and administration

MRC-Q was isolated from HeLa cells (JCRB Cell Bank, cell number JCRB9004) by iMIT (WO2021015298). Briefly, cultured cells with greater than 90% confluency were treated with 30 μM digitonin (WAKO, Cat#043–21376) for 3 min on ice. Cells were washed twice with Tris-sucrose buffer (10mM Tris (WAKO, 011–20095)/250mM sucrose (WAKO, 193–09545)/0.5mM EGTA (DOJINDO, 348–01311), pH7.4) and incubated in Tris-sucrose buffer for 10 min on ice. Cells were detached by gentle pipetting for 3 min and cell suspensions were collected into an ice-cold test tube. Cell debris was removed by two rounds of centrifugation at 500 × g for 10 min at 4 °C. Supernatant containing mitochondria and other organelles was collected and centrifuged at 3,000 × g for 10 min at 4 °C. Pelleted mitochondria were resuspended in Tris-sucrose buffer and quantified by Pierce 660nm Protein Assay Kit (Thermo, 22662). The mitochondria solution (MRC-Q) was diluted further in Tris-sucrose buffer if necessary and stored at −80˚C. Isolation of mitochondria with a traditional homogenized method (H-mito) was conducted with the Mitochondrial Isolation Kit (Thermo, 89874). Cell cultures of over 80% confluency were transferred to a 1 mL dounce homogenizer (DURAN WHEATON KIMBLE, 357538) and homogenized with 60 strokes of the plunger before proceeding to processing and storage as described for MRC-Q. MRC-Q was cryopreserved at −80 °C and then thawed before use. MRC-Q was cryopreserved at a concentration of 100 μg per 100 μL in 10 mM Tris-sucrose (250 mM) plus EGTA (0.5 mM), thawed in a constant temperature bath at 25°C for about 2 minutes and stored on ice. After centrifugation at 3,000 × g for 15 minutes (slow acceleration and slow brake), the supernatant was removed and diluted with PBS to a concentration of 50 μg per 60 μL. Ndufs4^–/– mice were injected with MRC-Q intravenously through the orbital venous plexus under anesthesia. The MRC-Q group received 50 μg of MRC-Q resuspended in 60 μL of PBS per mouse, once a week starting at 3- and ending at 7-weeks old. The control group received 60 μL PBS per mouse in the same manner as mice in the MRC-Q group.

Mitochondrial ATP production assay

The amount of ATP produced by MRC-Q and H-mito was evaluated using the luciferin-luciferase bioluminescence reaction. ATP production reaction was performed for 10 minutes at room temperature in 50 μL reaction volume, in a half-area 96-well plate. The reaction mixture included: 10 mM Malate (WAKO, 199–05642), 5 mM Glutamate (WAKO, 194–02032), 10 mM KH₂PO₄ (WAKO, 164–04245), 0.1 mM ADP (SIGMA, A2754–1G), 70 mM KCl (WAKO, 163–03545) with or without 5 μM Oligomycin A (Selleckchem, S1478). Isolated mitochondria (0.5 μg protein) were added to each well. To quantify the amount of ATP produced, the ATP standard solution was diluted with Tris Buffer (10 mM Tris-HCl, 250 mM Sucrose, 0.5 mM EGTA, pH 7.4) to 0.1, 0.3, 1, 3, 10, or 30 μL and added to each well. The reaction was stopped by 10-minute incubation on ice, after which 50 μL of luciferase reagent (Cell Titer-Glo, Promega) was added. The mixture was incubated at 37 °C for 2 minutes and equilibrated to room temperature for 10 minutes before luminescence was measured with a plate reader (Spark, Tecan). The amount of ATP produced was calculated from the emission intensities of the obtained standard curve and samples.

Measurement of inner membrane integrity and outer membrane integrity of isolated mitochondria

Inner membrane integrity was spectrophotometrically determined by formation of 2-nitro-5-benzoic acid from 5,5’-Dithiobis (2-nitrobenzoic acid) and CoA. Briefly, mitochondria solution (1 mg) was added to the reaction mixture with or without a detergent. Reaction mixture without a detergent contained the following: acetyl coenzyme A (0.3 mM; Sigma, A2056), oxaloacetic acid (5 mM; Sigma, O4126), 5,5’-Dithiobis (2-nitrobenzoic acid) (1 mM; Sigma, D7059). Detergent-added reaction mixture was supplemented with Triton X-100 (0.1%; Sigma, X100). Absorbance at 412 nm was monitored by spectrophotometer for assessment of 2-nitro-5-benzoic acid production. Outer membrane integrity was spectrophotometrically determined at 30˚C using a Cytochrome C Oxidase Assay Kit (Sigma, CYTOCOX1). Briefly, mitochondria solution (1.5 mg) and mitochondria lysate made with 20 mM n-dodecyl beta D-maltoside (Sigma, D4641) were added to a buffer containing reduced-cytochrome C (0.033 mM; Sigma, C2037). The production of oxidated-cytochrome C was determined by measuring absorbance at 550 nm. Membrane integrity was calculated according to the following formula: membrane integrity (%) = (amount of end-product with detergent – amount of end-product without detergent) / amount of end-product with detergent.

Isolation and quantification of MRC-Q mtDNA in tissues

Peripheral blood samples were obtained by cardiac puncture. Whole blood was diluted in 1mL BD Pharm Lyse Lysing buffer (BD Biosciences, #555899) and incubated for 2 min at room temperature to lyse mature erythrocytes. Other tissues were removed after mice were transcardially perfused with PBS. Each tissue was chopped into small pieces with scissors and minced with a homogenizer. Frozen specimens were minced by an ultrasonic homogenizer. DNA was extracted using Dneasy Blood & Tissue Kit (QIAGEN USA, #69504) according to the manufacturer’s instructions. Peripheral blood and bone marrow cells were washed twice in PBS and incubated directly with AL buffer and proteinase K at 56°C for 10 min. All other tissues were incubated overnight at 56°C in ATL buffer and proteinase K. The concentration of DNA was determined by a DeNovix spectrophotometer. The isolated DNA samples were stored at −30˚C until further analysis. Quantitative polymerase chain reactions (qPCR) were set using 15 ng of total DNA with TB Green Premix Ex Taq II (Tli Rnase H Plus) (Takara-Bio, # RR820W). Human mtDNA (mt-ND1 and mt-ND5) commercial kit primers (Human Mitochondrial DNA (mtDNA) Monitoring Primer Set- TAKARA Bio; #7246) and mouse β-actin fwd (5’-CTAAGGCCAACCGTGAAAAG-3’) and β-actin rev (5’-ACCAGAGGCATACAGGGACA-3’) primers were used. Relative quantification was performed using QuantStudio 1 or QuantStudio 3 Real-Time PCR System (ThermoFisher Scientific, USA) as follows: an initial denaturation step (95 °C for 1 min) was followed by amplification and quantification steps repeated for 50 cycles (95 °C for 15 s, 60˚C for 1 min). Quadruplicate amplification was carried out for each target gene. Human mtDNA and the mouse nuclear (n)DNA gene β-actin (Ct) were measured in samples obtained from Ndufs4^–/– mice treated with either MRC-Q or PBS. The human mtDNA signal was normalized to mouse β-actin. For each tissue, the average normalized human mtDNA signal from the PBSNdufs4^–/– group (background or non-specific amplification) was subtracted from the normalized human mtDNA signal from each MRC-QNdufs4^–/– sample to obtain a ΔΔCt value. Data are expressed as 2^-ΔΔCt.

Isolation and quantification of NZB mtDNA

Retro-orbital blood was collected and diluted in 250 μL of ACK Lysis Buffer (Thermo Fisher, #A10492). Peritoneal exudate cells (PECs) were obtained with a 7mL PBS lavage. Mice were perfused with 10mL of ice-cold PBS via the left ventricle (right ventricle opened) using a peristaltic pump, and heart, lung, liver, spleen, brain, and bone marrow (BM) were isolated. All samples were placed immediately on dry ice and stored at −80˚C. Frozen tissues were homogenized in 2mL tubes using 1.4mm ceramic beads (Fisher Scientific, #15–340-153) and a beadmill homogenizer (ThermoFisher Scientific, USA). DNA was isolated using the QIAamp DNA mini kit (QIAGEN, #51304) according to manufacturer’s instructions. The concentration of DNA in each sample was measured on a Biotek Synergy H1 Microplate reader using a Take3 multi-volume plate and Gen5 version 3.11 software. DNA samples were stored at – 20˚C until further analysis. Quantitative polymerase chain reactions (qPCR) were set up with 100 ng of DNA and PowerUP SYBR Green Master Mix reagent (Thermo, #A25742). Primers ARMS22 (5’-TTATCCACGCTTCCGTTACGTC-3’) and MT20 (5’-TGGCACTCCCGCTGTAAAAA-3’) were used to amplify NZB mtDNA as previously described²⁷, and mouse β-actin fwd (5’-CTAAGGCCAACCGTGAAAAG-3’) and β-actin rev (5’-ACCAGAGGCATACAGGGACA-3’) primers were used to measure nuclear (n)DNA content. qRT-PCR was performed using QuantStudio 3 (ThermoFisher) with denaturation (95 ˚C for 1 min) followed by amplification and quantification steps repeated for 40 cycles (95 ˚C for 15 s, 62 ˚C for 1 min). Reactions were performed in duplicate and averaged. NZB mtDNA (Ct) and the nuclear (n)DNA gene β-actin (Ct) were measured in samples obtained from Ndufs4^–/– mice treated with either NZB mitochondria or C57BL6/J mitochondria. The NZB mtDNA signal was normalized to β-actin. For each tissue, the average normalized NZB mtDNA signal from the B6Ndufs4^–/– group (background or non-specific amplification) was subtracted from the normalized NZB mtDNA signal from each NZBNdufs4^–/– sample to obtain a ΔΔCt value. Data are expressed as 2^-ΔΔCt.

Isolation of RNA and quantitative real-time (q)RT-PCR

Tissues were homogenized as described above, and RNA was isolated using the Direct-zol RNA MiniPrep Plus kit (Zymo, #R2072) according to manufacturer’s instructions. For PECS, RNA was isolated using the RNeasy Micro Kit (QIAGEN, #74004) according to manufacturer’s instructions. The concentration of RNA in each sample was measured on a Biotek Synergy H1 Microplate reader using a Take3 multi-volume plate and Gen5 3.11 software. RNA samples were converted to cDNA using the SuperScript IV VILO Master Mix (ThermoFisher #11766050) according to manufacturer’s instructions. qRT-PCR was performed using the following primers: Il6 fwd (5’-GACAACTTTGGCATTGTGG-3’), Il6 rev (5’-ATGCAGGGATGATGTTCTG-3’); Il18 fwd (5’-GACTCTTGCGTCAACTTCAAGG-3’), Il18 rev (5’-CAGGCTGTCTTTTGTCAACGA-3’); Tnfa fwd (5′-TATGGCCCAGACCCTCACA-3′), Tnfa rev (5′-GGAGTAGACAAGGTACAACCCATC-3′); Ifng fwd (ATTGCGGGGTTGTATCTGGG), Ifng rev (GGGTCACTGCAGCTCTGAAT); IL-4 fwd (5’-AGATGGATGTGCCAAACGTCCTCA-3’), IL-4 rev (5’-AATATGCCAAGCACCTTGGAAGCC-3’); and β-actin fwd (5’-CTAAGGCCAACCGTGAAAAG-3’), β-actin rev (5’-ACCAGAGGCATACAGGGACA-3’) using a QuantStudio 3 (ThermoFisher) with an initial denaturation step (95 ˚C for 1 min) followed by amplification repeated for 40 cycles (95 ˚C for 15 s, 60 ˚C for 1 min). Reactions were performed in duplicate and averaged. The relative expression was calculated by the ΔΔCt method. β-actin was used for normalization, and data are expressed as a fold change relative to the control group using 2^-ΔΔCt.

Statistics

Each experiment was performed in 2–5 independent cohorts and data were pooled for statistical analyses. All statistical tests were two-sided. Two-group comparisons were made using two-way Student’s t tests if normally distributed, with Welch’s correction applied when the variance between groups differed. The Mann-Whitney test was used for nonparametric comparisons of two groups. Energy expenditure over time was compared using a two-way ANOVA with repeated measures. Immune cell frequencies and numbers were compared using a two-way ANOVA was used with Sidak post hoc test. Survival was compared using the Mantel-Cox log-rank test. Power calculations were performed to estimate the minimum sample sizes needed to achieve at least 80% power with statistical significance set at 0.05 for the primary outcomes (survival or degree of mitochondria transfer). Data are expressed as mean ± standard error of the mean (SEM). The threshold for significance was P<0.05. Data analyses were performed using Prism v9 or v10 (GraphPad) and Excel v16.74 (Microsoft).

Extended Data

Extended Data Fig. 1. Additional metabolic cage parameters from NDUFS4-deficient mice treated with wildtype or NDUFS4-deficient bone marrow transplantation.

(a) Distance travelled and (b) average respiratory exchange ratio (RER) over 3 hr after a 1 hr air equilibration period. Data are expressed as mean +/− SEM. All data points are unique biological replicates. n=5 KO, n=17 WT. Data accompanies Fig. 1.

Extended Data Fig. 2. Engraftment of CD45.1⁺ cells in blood, spleen, and liver in mice transplanted with wildtype or mtD2 bone marrow.

(a) Representative gating for identification of CD45.1⁺ CD45.2^— donor immune cells and CD45.1^— CD45.2⁺ radioresistant host cells. (b) Chimerism expressed as a percentage of the ratio of CD45.1⁺ cells to the sum of CD45.1⁺ and CD45.2⁺ cells. (c) Representative gating in the blood to exclude red blood cells (RBCs) and platelets from particles less than 2 μm in diameter for identification of extracellular mtD2⁺ mitochondria. Data are expressed as mean +/− SEM. All data points are unique biological replicates. For b, n=3 WT, n=5 mtD2 (except n=4 for spleen). Data accompanies Fig. 2.

Extended Data Fig. 3. Bone marrow transplantation leads to transfer of mtD2+ mitochondria to host cells in the blood, spleen, and liver.

(a) Representative gating for identification of CD45.1^— CD45.2⁺ radioresistant host B cells, T cells, neutrophils, and monocytes as well as CD45.1^— CD45.2^— host epithelial, endothelial, and stromal cells in the spleen. (b) Representative gating to identify the proportions of host B cells, T cells, neutrophils, monocytes, epithelial, endothelial, and stromal cells that received mtD2⁺ mitochondria in the spleen. (c) Proportions of B cells (B), T cells (T), neutrophils (neut), monocytes (mono), epithelial cells (epi), endothelial cells (endo), and stromal cells that received mtD2+ mitochondria in the liver and (d) peripheral blood. For c-d, closed circles are WT and open squares are mtD2. Data are expressed as mean +/− SEM. All data points are unique biological replicates. For b-d, n=3 WT, n=5 mtD2 (except n=4 for spleen). For b-d, 2-way ANOVA with Sidak post-hoc test. ***P<0.001, ****P<0.0001. Data accompanies Fig. 2.

Extended Data Fig. 4. Mitochondria isolates from the liver are enriched in mtD2⁺ events and produce more yield than bone marrow mitochondria isolates.

(a) Flow cytometric identification of the proportion of mtD2⁺ mitochondria. Pre-gated on CD41^— CD45^— TER-119^— events less than 2 μm in diameter. (b) Mitochondrial yield from mouse liver or bone marrow (BM). Data are expressed as mean +/− SEM. All data points are unique biological replicates. For a-b, n=4 biological replicates/group. For b, Student’s t-test (two-sided). *P<0.05 Data accompanies Fig. 3.

Extended Data Fig. 5. Additional metabolic cage parameters from NDUFS4-deficient mice treated with mitochondria or PBS.

(a) Body weight, and (b) rectal core body temperature of 7-week-old KO mice treated with PBS or 100μg mitochondria 1–2 times per week. (c) Average energy expenditure (d) distance travelled, and (e) average respiratory exchange ratio (RER) over 3 hr after a 1 hr air equilibration period. Data are expressed as mean +/− SEM. All data points are unique biological replicates. For a, n=11 PBS, n=12 WT. For b, n=11/group. For c-e, n=15/group. Variation in n is due to mouse mortality. For c, Student’s t-test (two-sided). *P<0.05. Data accompanies Fig. 3.

Extended Data Fig. 6. Administration of wildtype mitochondria to NDUFS4-deficient mice does not alter immune cell composition in the blood.

(a) Representative gating to identify immune cells in the peripheral blood of Ndufs4^—/— (KO) mice. (b) Overall number of B cells (B), eosinophils (eos), natural killer cells (NK), neutrophils (neut), monocytes (mono), CD4 T cells, and CD8 T cells per mL of peripheral blood from KO mice treated weekly with PBS (closed circle) or 100μg mitochondria (open square) for 5 weeks. Data are expressed as mean +/− SEM. All data points are unique biological replicates. For b, n=14 PBS, n=18 Mito. For b, 2-way ANOVA with Sidak post-hoc test. Data accompanies Fig. 3.

Extended Data Fig. 7. Administration of mitochondria to NDUFS4-deficient mice does not alter immune cell composition in the blood or tissues.

(a) Overall number of immune cells in peripheral blood, (b) spleen, (c) bone marrow, and (d) liver of Ndufs4^—/— mice treated weekly with 100μg KO or WT mitochondria for 7 weeks. For a-d, closed circles are KO mito and open squares are WT mito. Data are expressed as mean +/− SEM. All data points are unique biological replicates. For a, n=3 KO Mito, n=4 WT Mito. For b and d, n=4/group. For c, n=4 KO Mito, n=3 WT Mito. For a-d, 2-way ANOVA with Sidak post-hoc test. Data accompanies Fig. 3.

Extended Data Fig. 8. Administration of mitochondria to NDUFS4-deficient mice does not alter expression of inflammatory cytokines.

(a) Inflammatory cytokine expression in the spleen (b) peritoneal exudate cells (PECS), and (c) liver of Ndufs4^—/— mice treated weekly with 100μg KO or WT mitochondria for 7 weeks. For a-c, closed circles are KO mito and open squares are WT mito. Data are expressed as mean +/− SEM. All data points are unique biological replicates. For a-c, n=4 biological replicates/group. For a-c, 2-way ANOVA with Sidak post-hoc test. Data accompanies Fig. 3.

Extended Data Fig. 9. Administration of MRC-Q does not alter body weight in NDUFS4-deficient mice.

Body weight of Ndufs4^—/— mice treated with PBS or 50 μg MRC-Q once per week from 3- to 7-weeks old. Data are expressed as mean +/− SEM. All data points are unique biological replicates. n=12 PBS, n=14 MRC-Q. Data accompanies Fig. 4.

Extended Data Table 1.

Comparison of the three treatment paradigms

		Figure 1	Figure 3	Figure 4
Experimental parameters:	Treatment paradigm	BMT	Mitochondria transplant
	Irradiation of treated mice	Yes	No	No
	Bone marrow cells	Yes	No	No
	Mitochondria transplant	No	Yes	Yes
	Donor species	Mouse	Mouse	Human
	Mitochondria dose	n/a	PBS: 0 Mito: 100 μg	PBS: 0 MRC-Q: 50 μg
	Performance sites	WUSM, OU	WUSM	OU
Survival outcomes:	Median survival (days)	KO to KO: 40 WT to KO: 74	PBS: 68 Mito: 84	PBS: 67.5 MRCQ: 80.5
	Sample size per group (n)	KO to KO: 11 WT to KO: 29	PBS: 27 Mito: 23	PBS: 14 MRC-Q: 14
	P-value, Mantel-Cox logrank test of survival curves	<0.0001	0.0052	0.0079

Abbreviations: BMT, bone marrow transplant; n/a, not applicable; PBS, phosphate buffered saline; Mito, mitochondria; WUSM, Washington University School of Medicine; OU, Osaka University; WT, wildtype; KO, knockout (of Ndufs4)

Supplementary Material

Video 1

Supplementary Video 1. Treatment of Ndufs4^–/– mice with exogenous mitochondria leads to improved hair regrowth and improved neurobehavioral characteristics. Mice were treated once per week with PBS (left cage) or 100 μg Mito (right cage).

Video 2

Supplementary Video 2. Administration of MRC-Q improves righting reflex in Ndufs4^–/–mice. Mice were treated once a week with PBS (right cage) or 50 μg MRC-Q (left cage).

Video 3

Supplementary Video 3. Administration of MRC-Q improves tail suspension stability and landing recovery in Ndufs4^–/– mice. Mice were treated once a week with PBS (right cage) or 50 μg MRC-Q (left cage).

Data Availability

All data supporting the findings of this study are available within the paper and its Supplementary Information or are available by request to the corresponding authors. MRC-Q is a proprietary, unique biological material and its use is restricted by a material transfer agreement. MRC-Q is available upon request from LUCA Science, Inc.