Does supplementation with mitochondria improve oocyte competence? A systematic review

Search strategy

The study selection process for this systematic review was based on PRISMA Guidelines (Liberati et al. 2009). A systematic literature review to assess the studies investigating the effect of oocyte mitochondrial supplementation was based on an online search using PubMed (MEDLINE database) and Web of Science until September 2019.

The search terms used to identify the studies were ‘mitochondri* (e.g. mitochondria, mitochondrial and mitochondrion) supplementation’, ‘mitochondri* transfer’, ‘mitochondri* injection’, ‘mitochondri* microinjection’, ooplasm* (e.g. ooplasm and ooplasmic), ‘ooplasm* injection’, ‘cytoplasm* (e.g. cytoplasm and cytoplasmic) transfer’ or ‘cytoplasm* injection’, and ‘oocyte’ or ‘embryo’.

Study selection

Studies were selected by three independent subjects. Studies were first screened for eligibility using article titles and then by reading the study abstracts.

We included clinical (experimental or observational) and animal studies performing mitochondrial transfer, either isolated or within cytoplasm, into fertilized/activated oocytes, designed to improve oocyte competence or explore the safety of this technique. Studies with non-mammalian species were excluded. Narratives, reviews, commentaries, letters to editor and abstract meeting were also excluded, as well as articles with full text in a language other than English, French, Portuguese or Spanish.

Data extraction process

Data on publication date, species, context (study objective), methods (cytoplasm vs mitochondria, source of mitochondria, study outcomes (clinical and biological measurements, molecular parameters) and quality scores were extracted from relevant manuscripts. Data extraction was independently performed by two authors (AFF and MS) and validated by a third author (SAR). If any doubt remained, the full text of the article was discussed. Arbitration by the third author was applied in cases of divergence of opinion.

Quality assessment

The risk of bias of the included studies was evaluated using the SYstematic Review Centre for Laboratory animal Experimentation SYRCLE’s tool (Hooijmans et al. 2014). The SYRCLE’s tool was adapted from the Cochrane risk of bias to assess the methodological quality of animal interventional studies. We have applied this scale to the 36 included studies with slight adaptations for clinical studies.

The score was judged according to the risk of bias: ‘Yes’ indicates a low risk of bias, ‘No’ indicates a high risk of bias, ‘Unclear’ indicates that data supplied was insufficient to assess the risk of bias and ‘NA – not applicable’ indicates inability to apply this item.

In order to attribute a quality score, 10 different criteria were measured: (1) random allocation sequence; (2) baseline characteristics; (3) allocation concealment; (4) random housing; (5) blinded intervention; (6) random outcome assessment; (7) blinded assessment of outcome; (8) incomplete outcome data; (9) selective outcome reporting; (10) other sources of bias. The criteria ‘other source of bias’ was defined by authors, concerning the mitochondrial extraction protocol, to evaluate specifically the description of the protocol and the evaluation of mitochondria activity.

Context (objectives of studies)

Human studies (12 out of 36 studies) were performed in the context of infertile patients with poor embryo quality or advanced maternal age. The aim was to verify if mitochondrial (Fakih 2015, Oktay et al. 2015, Labarta et al. 2019b) or ooplasmic transfer (Cohen et al. 1997, 1998, Huang et al. 1999, Lanzendorf et al. 1999, Dale et al. 2001) could improve oocyte competence and hence the prognosis of the patients with previous in vitro fertilization (IVF) failures. Two studies were designed to evaluate the clinical efficacy of this procedure in women with advanced maternal age (Lanzendorf et al. 1999, Kong et al. 2003) and two others to evaluate the degree of mitochondrial heteroplasmy after ooplasmic transfer (Brenner et al. 2000, Barritt et al. 2001b). One study was conducted with the purpose of evaluating the effect of cytoplasmic transfer on promoting oocyte maturation (Hoseini et al. 2016).

The majority of the experimental studies (14 out of 24) were conducted in animals with good quality oocytes. The effects of mitochondrial transfer were evaluated by comparing IVF/ICSI (intracytoplasmic sperm injection) vs mitochondrial supplementation in combination with ICSI (mICSI) (Li et al. 2017, Tsai & St John 2018) or by comparing somatic cell nuclear transfer (SCNT) in animals of the same species (Lee et al. 2017, Srirattana & St John 2018, Xu et al. 2019) or involving different species (iSCNT) (Sansinena et al. 2011, Hua et al. 2012, Takeda et al. 2012, Gonzalez-Grajales et al. 2016), with or without mitochondrial transfer. Mitochondrial transfer was also compared in the context of parthenogenetic development in bovine species (Takeda et al. 2005). Two studies were designed to evaluate different volumes of injected cytoplasm in mouse species (Nagai et al. 2004, Van Thuan et al. 2006).

Six studies (Petr et al. 1994, El Shourbagy et al. 2006, Hua et al. 2007, Chiaratti et al. 2011, Cagnone et al. 2016, Tsai et al. 2018) were performed in the context of poor oocyte quality while models of aging were explored in three other studies using mouse species (Yi et al. 2007, Igarashi et al. 2016, Sheng et al. 2019). The purpose of these studies was to evaluate oocyte quality and embryonic development after mitochondrial/ooplasmic transfer. Some studies evaluated the effect of this procedure on oocyte/embryo gene expression profiles (Tsai & St John 2018, Tsai et al. 2018) and others have tried to correlate the outcomes with the amount of mtDNA copies (Takeda et al. 2012, Cagnone et al. 2016, Srirattana & St John 2018, Sheng et al. 2019), the ATP content (Van Blerkom et al. 1998, Gonzalez-Grajales et al. 2016, Sheng et al. 2019) and gene expression profile alterations (Hua et al. 2012, Cagnone et al. 2016, Gonzalez-Grajales et al. 2016, Srirattana & St John 2018).

Some of these authors have also explored the safety of the technique (Chiaratti et al. 2011, Hua et al. 2012, Li et al. 2017) whereas another study was designed to evaluate the transgenerational effects of mitochondrial transfer in mouse species (St John et al. 2019).

Mitochondrial vs cytoplasmic transfer and cell source of transplanted mitochondria

The supplementation of isolated mitochondria was reported in 20 papers while cytoplasmic transfer was reported in 16. The material used to supplement oocytes was obtained from different cell types as illustrated in Fig. 3.

View Full Size

Illustration of the different cell sources used for oocyte mitochondrial supplementation. EPC, egg precursor cells; ADSC, adipose-derived stem cells; CC, cumulus cells; GC, granulosa cells.

Citation: Reproduction 161, 3; 10.1530/REP-20-0351

• Download Figure
• Download figure as PowerPoint slide

The major source of isolated mitochondria was oocyte (Van Blerkom et al. 1998, El Shourbagy et al. 2006, Cagnone et al. 2016, Srirattana & St John 2018, Tsai & St John 2018, Tsai et al. 2018) and egg precursor cells (EPC) (Fakih 2015, Oktay et al. 2015, Labarta et al. 2019b, St John et al. 2019). Somatic cells were used in six studies, with mitochondria recovered from liver (Yi et al. 2007, Igarashi et al. 2016) or from granulosa/cumulus cells (Kong et al. 2003, Takeda et al. 2005, Hua et al. 2007, 2012) and other cell types were used in four studies, including adipose tissue stem cells (Wang et al. 2017b, Sheng et al. 2019) or fibroblast-derived induced pluripotent stem cells (Takeda et al. 2012, Li et al. 2017).

Regarding cytoplasmic transfer, three types of cell were used as sources: oocytes (Cohen et al. 1997, 1998, Huang et al. 1999, Lanzendorf et al. 1999, Brenner et al. 2000, Barritt et al. 2001b, Dale et al. 2001, Nagai et al. 2004, Chiaratti et al. 2011, Sansinena et al. 2011, Gonzalez-Grajales et al. 2016, Hoseini et al. 2016), granulosa/cumulus cell (Van Thuan et al. 2006), and zygotes/embryos (Huang et al. 1999, Van Thuan et al. 2006). Three papers reported the evaluation of isolated mitochondrial function prior to oocyte supplementation. Mitochondrial membrane potential was determined using the fluorescent probe JC-1 (Hua et al. 2012, Sheng et al. 2019), oxygen consumption rate was determined by high-resolution respirometry (Cagnone et al. 2016) and ATP production was assessed by an ATP bioluminescence assay (Sheng et al. 2019). Mitochondria morphology (assessed by transmission electron microscopy) (Wang et al. 2017b) and mitochondrial content (evaluated by quantification of mtDNA copy number with RT-PCR) (Hua et al. 2012, Cagnone et al. 2016) were also assessed prior to mitochondrial transfer. Additionally, and to confirm that oocytes were indeed injected with isolated mitochondria, several studies labeled mitochondria with MitoTracker Green probe (El Shourbagy et al. 2006, Hua et al. 2007, Takeda et al. 2012, Cagnone et al. 2016, Igarashi et al. 2016) or Red Fluorescent Protein (RFP) (Li et al. 2017).

Outcomes

The outcome measures were categorized in clinical and biological measures, including fertilization and pregnancy rates, embryo development, the existence of offspring and adverse outcomes (summarized in Table 1 – human studies, and Table 2 – animal models), and molecular parameters, including gene expression patterns, mtDNA copy number and ATP content (summarized in Table 3).

Clinical outcomes

There were 11 out of 36 studies assessing clinical outcomes. Autologous mitochondrial transfer, a treatment also referred to as AUGMENT was evaluated at four international fertility centers. Clinical pregnancy rate was not improved in the randomized controlled trial (RCT) conducted in Spain (n = 59) by Labarta et al. (2019b). The cumulative ongoing pregnancy and live birth rate per transferred embryo was 41.2% in the AUGMENT group and 41.7% in the control group. An increased clinical pregnancy rate was demonstrated in the other three non-randomized, experimental studies, in which the control group was the group’s historical IVF outcome. Fakih (2015), reported the clinical experience from two centers: Canada (n = 34) and The United Arab Emirates – UAE (n = 59), reporting a clinical pregnancy rate of 35 and 22%, respectively, in the AUGMENT group, compared to 11 and 4%, respectively, in the previous clinical pregnancy rate. At FAKIH IVF (in UAE), a comparison between treatment group (AUGMENT, n = 171) and control group (ICSI, n = 106) was also performed in a subgroup of 25 women. Embryo quality was higher in the AUGMENT group (embryo transfer performed in 14/25 women in AUGMENT groups vs 2/25 women in ICSI group). Oocytes were not randomly assigned to each group, nor was the criterion for allocation defined (Fakih 2015). Finally, Oktay et al. (2015), reported a clinical pregnancy rate of 40% (4 out of 10 women) in comparison with the 10% baseline of this fertility center in Turkey, among couples with multiple IVF failures.

Other experimental studies with very small samples, again with controls being previous ART cycles, revealed a tendency toward improvement of clinical pregnancy rates after mitochondrial supplementation by ooplasmic transfer. In the study by Cohen et al. (1998), pregnancy was achieved in three out of five patients (18 embryo transfers) with no previous pregnancies (23 prior embryo transfers). In the study of Huang et al. (1999), pregnancy was achieved in four out of nine patients (39 embryo transfers) with no previous pregnancies (85 prior embryo transfers). In the study conducted by Lanzendorf et al. (1999), one out of four patients achieved a pregnancy, although comparisons with previous cycles were not reported. Successful pregnancies after ooplasmic transfer were reported in three publications in the context of aging (Kong et al. 2003) and poor embryo quality (Cohen et al. 1997, Dale et al. 2001).

To evaluate the safety of mitochondrial transfer information regarding genetic screening and newborn/child follow-up was also assessed in the aforementioned studies and other clinical studies. Genetic screening was performed in five studies. Preimplantation genetic screening (PGS) revealed no differences in the euploid rate between the AUGMENT group and control group (43.8 ± 41.7% vs 63.8 ± 44.1%) in one study (Labarta et al. 2019b). The euploid embryo rate was 32% in the AUGMENT group in another study (Oktay et al. 2015). Karyotype (determined after amniocentesis) was normal in all the fetuses of ongoing pregnancies in three other studies (Cohen et al. 1997, 1998, Huang et al. 1999).

Cohen and coworkers (Cohen et al. 1998, Brenner et al. 2000, Barritt et al. 2001b) investigated the pattern of mitochondrial inheritance in the cases of ooplasmic transplantation performed in their laboratory. Their technique involved the transfer of ooplasm from a young donor to a recipient oocyte and therefore, the infusion of a different mtDNA population. mtDNA heteroplasmy was confirmed by mtDNA fingerprinting in samples from embryos (46%), amniocytes (25%), placentas (50%), fetal cord blood (50%) and the blood of two 1-year old children (100%). One of the two babies with mitochondrial heteroplasmy in cord blood was described as healthy at a 6-month follow-up while there was no information on the other. The two 1-year old children were reported as healthy.

A total of 22 healthy infants born after mitochondrial supplementation (14 after allogenic ooplasm) were reported. Despite the presence of heteroplasmy, a normal healthy child reaching appropriate developmental milestones was reported at 6-month pediatric evaluation by Brenner et al. (2000). Sixth month follow-up was also carried out in the study from Labarta et al. (2019b), reporting the absence of congenital malformations, although this study lacks detailed information on how the follow-up had been conducted. Follow-up was not performed after birth in the remaining children.

Some abnormalities could be missed and only diagnosed after the 6-month follow-up period. However, major congenital abnormalities are often diagnosed in prenatal or neonatal period and various pathologies are multifactorial so it would be difficult to associate these abnormalities to the technique.

Fertilization rate and embryo development

Fertilization rate in mitochondrial supplementation group was higher than in control group in 9 out of 25 studies (Cohen et al. 1998, El Shourbagy et al. 2006, Yi et al. 2007, Chiaratti et al. 2011, Hua et al. 2012, Oktay et al. 2015, Cagnone et al. 2016, Wang et al. 2017b, Xu et al. 2019). This rate was not significantly different in 16 studies (Huang et al. 1999, Lanzendorf et al. 1999, Nagai et al. 2004, Takeda et al. 2005, Van Thuan et al. 2006, Hua et al. 2007, Sansinena et al. 2011, Takeda et al. 2012, Gonzalez-Grajales et al. 2016, Igarashi et al. 2016, Lee et al. 2017, Li et al. 2017, Srirattana & St John 2018, Sheng et al. 2019), and was not evaluated in the remaining studies (Petr et al. 1994, Cohen et al. 1997, Van Blerkom et al. 1998, Brenner et al. 2000, Barritt et al. 2001b, Dale et al. 2001, Kong et al. 2003, Hoseini et al. 2016, Tsai & St John 2018, Tsai et al. 2018, St John et al. 2019).

Improved embryo development was reported in 12 out of 27 papers (Cohen et al. 1997, 1998, Dale et al. 2001, El Shourbagy et al. 2006, Hua et al. 2007, Yi et al. 2007, Chiaratti et al. 2011, Hua et al. 2012, Cagnone et al. 2016, Wang et al. 2017b, Xu et al. 2019). Embryo development was not different from controls in 15 studies (Huang et al. 1999, Lanzendorf et al. 1999, Nagai et al. 2004, Takeda et al. 2005, Van Thuan et al. 2006, Sansinena et al. 2011, Takeda et al. 2012, Oktay et al. 2015, Gonzalez-Grajales et al. 2016, Igarashi et al. 2016, Lee et al. 2017, Li et al. 2017, Srirattana & St John 2018, Labarta et al. 2019b, Sheng et al. 2019) and was not evaluated in the remaining (Petr et al. 1994, Van Blerkom et al. 1998, Brenner et al. 2000, Barritt et al. 2001b, Kong et al. 2003, Hoseini et al. 2016, Tsai & St John 2018, Tsai et al. 2018, St John et al. 2019). Although not significant, one study reported a trend toward higher embryo grade on day 3 (Oktay et al. 2015).

One of the studies did not evaluate fertilization or embryo development, and therefore was not included in Table 1. However, it was not excluded from our review since it demonstrates that the microinjection of cytoplasm from fully grown oocytes can induce germinal vesical breakdown in immature oocytes (Petr et al. 1994), a finding that supports the improvement of oocyte competence with mitochondrial supplementation.

Pregnancy and offspring from animal studies

Pregnancy and the offspring were monitored in four of the animal studies, two of them reporting an improvement in pregnancy rates following mitochondria/ooplasmic transfer. Chiaratti et al. demonstrated the efficacy of ooplasmic transfer to rescue bovine oocytes exposed to ethidium bromide, an inhibitor of mtDNA replication and transcription. Blastocyst from the two groups supplemented with ooplasm (ethidium bromide and control) were transferred to recipient cows. The group exposed to ethidium bromide had a gestation rate of 32% compared to 17% of control group (good-quality oocytes not exposed to ethidium bromide). Three calves were born following ooplasmic transfer and were normal (Chiaratti et al. 2011).

Mouse oocytes were injected with autologous mitochondria from adipose tissue-derived stem cells (ADSC) in the other study (Wang et al. 2017b) reporting an improvement in pregnancy rate. The authors compared ICSI with (experimental group) and without mitochondrial supplementation (control group) in mature oocytes from reproductive aged mice (12 months-old) and concluded that the procedure had rescued fertility based on the live-pup rate of 15,7% vs 2%. The study reports the birth of eight mouse pups in the mitochondria transfer group (transfer of 51 embryos to 9 recipient) compared to one pup (transfer of 50 embryos to 7 recipient) in the control group. Data on embryo development and pregnancy rates from young mice after ICSI was not available, neither was a comparison between oocytes from young and aged mice, although the very low pup live birth reported in both experimental and control group is noteworthy.

The other two papers reported no improvement in implantation rate (Li et al. 2017) or birth rates (Van Thuan et al. 2006). In the study from Li et al. (2017), mouse zygotes retrieved from the ampulla of the oviducts were injected with either mitochondria from induced pluripotent stem cell (iPSC) or mitochondrial respiration buffer (control group). They were then transferred into surrogate mice and excised at embryonic day 13.5 to compare placental and fetal weights, which were similar. On the other hand, Van Thuan et al. (2006) described opposite results, obtaining higher placental weights and fetal deaths in the experimental group. However, a very different methodology was used in this study, seeking to determine the effects of cytoplasmic transfer from different mouse strains, different cell sources (cumulus and embryonic cells) and different volumes into mouse oocytes.

A higher degree of embryo mitochondrial heteroplasmy was found in two-cell to blastocyst stage embryos and related to embryo arrest (Hua et al. 2012). However, it is important to mention that these results were observed in the context of interspecies somatic cell nuclear transfer (iSCNT). Cloned embryos were produced using bovine oocytes as the recipient cytoplasm and ovine granulosa cells as the donor nuclei, supplemented with bovine (BOBMT embryos) or ovine (BOOMT embryos) mitochondria. The degree of heteroplasmy was calculated as the percentage of bovine mtDNA copy number divided by ovine mtDNA copy number. BOOMT embryos had a higher degree of mitochondrial heteroplasmy and a significantly lower developmental rate in comparison to control group (cloned embryos without supplemented mitochondria). In contrast, the degree of heteroplasmy and developmental rate of BOBMT embryos were not different from controls.

Donor mtDNA was present in mouse placentas and fetuses after supplementation with mitochondria from iPSC. In this study (Li et al. 2017) similar patterns of methylation patterns in the differentially methylated regions (DMRS) of H19 and SNRPN, responsible for embryo development, were observed between embryos from experimental (iPSC) and control groups. Takeda et al. (2012) described no differences in mtDNA content between bovine blastocysts produced after injection of water buffalo mitochondria and controls (injected with buffer or uninjected). Also, parthenogenetic activation to the blastocyst stage was achieved. These findings indicate that intergeneric mitochondria were not destroyed and heteroplasmy did not apparently compromise embryo development. Mitochondrial heteroplasmy was also found in the three calves born after ooplasmic transfer and, as mentioned before, the two gestation had developed to term and the three calves were described to be healthy at birth (Chiaratti et al. 2011).

In addition to previous studies, the safety of mitochondrial transfer was investigated not only in the first but in three generations of offspring by St. John et al. (2019). Autologous egg precursor cell (EPC)-mitochondria were transferred to mouse oocytes and resulted in the birth of seven female founders. When these females reached sexual maturity they were mated, and the offspring formed the first generation. The cycle was repeated until the third generation. Physiological and anatomical evaluation were performed in all three generations. F1 mitochondrial supplemented mice were found to be heavier than controls during the first 3 weeks of life, but no differences were found in subsequent generations. In addition to minor anomalies, the authors observed cardiac valvular pathology in all three of the first-generation mice and in one from the second generation. This cardiac defect was not observed in the third-generation or in control mice.

Gene expression, mitochondrial DNA copy number and ATP content

There were 13/36 studies assessing molecular outcomes in which two used human cells (Hoseini et al. 2016, Labarta et al. 2019b) while the others different animal models such as mouse (Van Blerkom et al. 1998, Sheng et al. 2019, St John et al. 2019), bovine (Hua et al. 2012, Takeda et al. 2012, Gonzalez-Grajales et al. 2016, Srirattana & St John 2018) or porcine (Cagnone et al. 2016, Tsai & St John 2018, Tsai et al. 2018). The expression of genes related to metabolism, oocyte/embryo development and mitochondria evaluated in these studies was analyzed and summarized (Table 3).

Upregulated metabolism-related genes were reported in two studies although regulating different processes. ATP5B and ATP1A1, which are involved in oxidative phosphorylation (OXPHOS), were upregulated in mINT (mitochondrial supplementation and somatic cell nuclear transfer)-derived embryos (Srirattana & St John 2018) while triosephosphate isomerase 1 (TPI1) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), key genes in the glycolysis pathway, were significantly higher in embryos derived from mitochondrial transfer when compared to stage-matched IVF embryos (Tsai & St John 2018). However, another study from the same group showed that blastocyst stage gene expression is altered by supplementation of mtDNA-deficient oocytes. Specifically, metabolism and mitochondrial-related genes, such as COMMD4 (COMM domain containing 4) (Tsai et al. 2018). Furthermore, energy-producing gene ATPase6 was also downregulated (Hoseini et al. 2016).

BCB (Brilliant Cresyl Blue) has been used in mammalian species to assess developmental competence. This dye is reduced to a colorless compound by glucose-6-phosphate dehydrogenase (G6PD). Since G6PD is progressively downregulated during oocyte growth, good-quality oocytes stain blue (BCB⁺) while developmentally incompetent oocytes are colorless (BCB^-). Although mICSI-BCB^- blastocysts had gene expression patterns more similar to ICSI BCB⁺ blastocyst than ICSI BCB^-, which supports a stabilization effect of mitochondrial supplementation, BCB^- blastocysts showed downregulated expression patterns (DEG), particularly in metabolism-related genes. Interestingly, a pathway that regulates mitochondrial biogenesis (involving sirtuins) was activated (Cagnone et al. 2016).

Concerning oocyte and embryo development related genes, no differences in pluripotency genes, such as OCT4 and SOX2, were reported (Hua et al. 2012, Cagnone et al. 2016, Lee et al. 2017, Tsai & St John 2018). Srirattana and St John (2018) found that 188 differentially expressed genes (DEGs) were upregulated and 321 DEGs were downregulated in the miNT blastocysts, being cellular development the most influenced molecular and cellular function, with increased expression. On the contrary, when cytoplasmic transfer from human MII (metaphase II stage) oocytes into GV (germinal vesicle stage) oocytes was performed, genes involved in cytoplasmic maturity (GDF9, BMP15), meiosis (CDC25, AURKC), spindle check point (BUB1, CDC20, MAD2L1), DNA repairing and cell cycle checkpoint enzymes (ATR, ATM, BRCA1) were not expressed (Hoseini et al. 2016).

Mitochondrial transcription factor A (TFAM) was expressed in similar range in mitochondrial supplementation groups, as well as mitochondrial transcription factor B1 (TFB1), Sirtuins (Tsai & St John 2018), POLRMT (Hua et al. 2012) and Mitochondrial Mitofusin-1 (Lee et al. 2017), indicating no differences concerning mitochondrial-related genes expression. However, another study demonstrated that oocytes supplemented with mitochondria modulated methylation of POLG (hypomethylation) at metaphase II oocytes, leading to mtDNA replication prior to implantation (Tsai et al. 2018).

St John et al. (2019), identified 53 differentially expressed genes between the primordial follicles of third-generation mice, indicating that mitochondrial supplementation at the time of fertilization may act on the germline to modify chromosomal gene expression. Indeed, using the Panther Ontology Classification System, the authors found that cellular, metabolic and developmental processes were the most modified biological processes, while binding, catalytic activity and receptor activity were the most modified molecular functions.

After autologous mitochondrial transfer no differences in mtDNA content were evident in AUGMENT euploid embryos (21.8 (IQR 14.6–24.7)) compared to control group (16.9 (IQR 13.8–23.9) (Labarta et al. 2019b). As mentioned before, both groups had similar fertilization and embryo development rates.

In mouse blastocysts mtDNA copy number was significantly higher in aged oocytes injected with mitochondria (240,891 ± 75,642, n = 6) when compared with the buffer-injected group (41,228 ± 5,019, n = 6) (Sheng et al. 2019). Similarly, in bovine, additional mitochondrial injection significantly increased mtDNA copy number in blastocyst stage embryos (Srirattana & St John 2018). However, none of these studies demonstrated better embryo development in association with higher mtDNA copy number.

Cagnone et al. (2016) measured mtDNA copy number in MII oocytes, 2, 4, 6, 8 and 16-cell stage, morula, early blastocysts and expanded blastocysts, to monitor evolution through porcine embryo development. BCB⁺-embryos (derived from developmentally competent oocytes) generated by IVF or ICSI demonstrated a similar profile, showing a decrease from MII oocytes to 2-cell embryos, the same content during preimplantation development and an increase at the expanded blastocyst stage. On the contrary, embryos derived from compromised oocytes (BCB^-) and generated by ICSI had lower mtDNA at the MII stage, a turnover at the 4-cell stage and hereafter, reflecting mitochondrial replication before the expanded blastocyst stage. Embryos derived from compromised oocytes (BCB^-) and generated by mICSI had higher but persistent levels of mtDNA when compared to ICSI-BCB^-. All groups had similar mtDNA copy numbers (>200,000) at expanded blastocyst. This study demonstrated that oocyte supplementation with mitochondria is associated with maintaining a higher mtDNA copy number in preimplantation embryos, higher fertilization rates and better embryo development. Also, low levels of mtDNA copy number in embryos derived from compromised oocytes could activate premature mitochondrial replication (e.g. before expanded blastocyst stage) as a compensatory mechanism.

Finally, only one out of the three studies that determined ATP production after mitochondrial supplementation using a bioluminescence assay demonstrated an increase in net ATP production in recipient oocytes (Van Blerkom et al. 1998), while the others found no differences (Gonzalez-Grajales et al. 2016, Sheng et al. 2019).

On the source of mitochondria for supplementation

The only clinical trials evaluating the effect of mitochondrial transfer were performed with autologous mitochondria isolated from cells - oogonial germline stem cells (OSC) (Fakih 2015, Oktay et al. 2015, Labarta et al. 2019b), the existence of which is still controversial (Gosden 2004, Bristol-Gould et al. 2006, Zhang et al. 2014, Gosden & Johnson 2016). Indeed, the identification of OSC from ovarian cortex biopsies of human and other species have not been replicated by other groups (Liu et al. 2007, Lei & Spradling 2013, Yuan et al. 2013) and detailed information on the protocol for identifying and harvesting these cells is lacking. A rare and mitotically active pool of cells has been purified from the ovarian cortex of reproductive-aged women, considered to be germline stem cells (White et al. 2012). However, the method used to identify these cells, the human Vasa analog – DDX4 antibody, has been shown to label other non-mitotically active cells in postnatal ovaries (Zhang et al. 2012).

Evidence against neo-oogenesis does not exclude the possibility of the survival of a population of stem cells (or stem cell-like cells) of uncertain potency that could be harvested from ovarian cortex. Still, considering the lack of specificity of DDX4 and scarcity of the alleged OSC population in adult ovarian tissue, these cells were surprisingly identified in all except one biopsy (total of 57 biopsies including three fragments of the ovarian cortex) in the study conducted by Labarta et al. (2019b). In addition, the quantity of mtDNA that is actually transferred into oocytes is not described in any of the AUGMENT studies (Fakih 2015, Oktay et al. 2015, Labarta et al. 2019b), but is expectably lower than the content found in most somatic cells, probably ranging from 10-100 copies considering the homology with oogonia and primordial germ cells (Shoubridge & Wai 2007). These facts support the conclusion that injected mitochondria do not modify the dynamics of mtDNA content in blastocysts, inferred by the similar mtDNA content found in blastocysts from AUGMENT and control groups (Labarta et al. 2019b). To better understand if injected mitochondria are destroyed or not replicated it would be important to quantify or evaluate the viability of isolated mitochondria, as well as to label and track mitochondria after injection.

The clinical use of OSC in women with advanced maternal age is also an issue because the intraovarian environment may become inadequate for OSCs differentiation (Wang et al. 2017a). Another aspect questioning the feasibility of the AUGMENT studies is the high cost and invasive method necessary for cell recovery. This method involves a surgery and it is noteworthy to mention the risk of diminishing the ovarian reserve. Thus, the invasiveness of the procedure, the controversial existence of OSC and the uncertainty about the number and quality of isolated mitochondria extracted from OSC are some of the issues pointed out by authors concerning the AUGMENT technology (Gosden & Johnson 2016, Kristensen et al. 2017, Cecchino et al. 2018, Labarta et al. 2019a, Qi et al. 2019).

Regardless, germline or oocyte-differentiated cells are advocated (Schatten et al. 2014, Darbandi et al. 2016) as a privileged source of mitochondria for oocyte transfer because of mitochondrial dependency on the nuclear genome. However, our review shows that other sources of mitochondria such as hepatocytes (Yi et al. 2007) and granulosa cells (Hua et al. 2007) have shown promising results, questioning this preference. Indeed, labeled granulosa cell-mitochondria were identified up to the hatched blastocyst stage in bovine embryos (Hua et al. 2007) and found to improve embryo development and clinical pregnancy. A higher oxygen consumption rate was observed in oocytes from aged mice after injection of mitochondria from liver cells (Igarashi et al. 2016). However, it is unclear how nuclear genome interacts with the extra mtDNA that is supplemented.

Mitochondria obtained from granulosa or cumulus cells have, in theory, a great disadvantage as a mitochondrial source since these cells are also affected by aging (May-Panloup et al. 2016). Granulosa cells of advanced maternal age women have been shown to contain more mtDNA mutations (Seifer et al. 2002, Tsai et al. 2010) and damaged mitochondria (Tatone & Amicarelli 2013, Liu et al. 2017).

The use of stem cells (e.g. hematopoietic and adipose-derived stem cells) as the source of mitochondria is encouraged by many authors (Schatten et al. 2014, Labarta et al. 2019a) due to the mitochondrial morphological and metabolic similarities with mitochondria in metaphase II oocytes. Indeed, preovulatory secondary oocytes display round-shaped mitochondria and an electron-dense matrix. After fertilization, mitochondrial structure and metabolism are in line with that of somatic cells. Mitochondria acquire an elongated form with many cristae, and gradually changes from glycolysis to OXPHOS (Motta et al. 2000, Eichenlaub-Ritter et al. 2011).

An alternative source, not yet explored, was proposed by Kristensen et al. The authors pointed out that OSC do not pass through the mitochondrial genetic bottleneck (important for selection of subpopulations of mtDNA during oogenesis) and suggested a possible increase in the supply of fully grown oocytes by two methods: in vitro maturation of immature oocytes collected from women undergoing ART or in vitro activation of ovarian cortical biopsies (Kristensen et al. 2017).

The transfer of oosplasm from a young donor has proven efficient in both animal and human studies, resulting in the birth of nearly 30 apparently healthy babies worldwide (Barritt et al. 2001a). Safety is the main disadvantage of this procedure, due to the transfer of foreign populations of mtDNA (addressed ahead), but there is sufficient evidence to suggest that mitochondria are potentially responsible for the improvements of oocyte quality that have been reported. This is based on the demonstrated mitochondrial role in oocyte competence acquisition and embryonic development (Benkhalifa et al. 2014, May-Panloup et al. 2016, Cecchino et al. 2018). Spindle formation is critical for normal chromosome segregation and dependent on mitochondrial ATP supply. Sufficient mitochondria are necessary for the rise in ATP production observed during mitochondria assemble around the spindle in meiosis I and II. Viable mitochondria are also necessary for calcium signaling at the oolemma after sperm entry as it allows the oocyte progression to anaphase and resumption of meiosis II (Eichenlaub-Ritter et al. 2011). Other strategies to improve oocyte competence in the context of aging also rely on increasing mitochondrial biogenesis and activity, without the need for oocyte mitochondrial injection. Resveratrol acts through sirtuin pathway and Coenzyme Q₁₀ contributes to the transport of electrons in the mitochondrial respiratory chain. Both these pharmacological agents were able to enhance embryonic development in animal models of ovarian aging (Schatten et al. 2014, Darbandi et al. 2016, May-Panloup et al. 2016).

In summary, there is still no evidence to support a particular type of cell as the optimal source of mitochondria for oocyte supplementation. What we can conclude is that injected mitochondria must be from the same women (autologous source) and without the same age-related alterations as those found in oocytes (such as mtDNA mutations and mitochondrial dysfunction). Apart from the type of cell source, recover of healthy mitochondria is of utmost importance in the process of mitochondrial isolation. This usually encompasses centrifugation and homogenization steps. Failure to achieve good results after oocyte mitochondrial supplementation might be attributed, for instance, to the presence of cell debris that may cause metabolic overload in poor quality oocytes (Qi et al. 2019).

Furthermore, the viability of isolated mitochondria should be demonstrated prior to oocyte injection. In our review, none of the human studies, and only a few animal studies (Hua et al. 2012, Cagnone et al. 2016, Sheng et al. 2019), evaluated the function of isolated mitochondria prior to supplementation. Likewise, as mentioned before, it would be important to label and track mitochondria during preimplantation development.

On the experimental or clinical context

The studies investigating the efficacy of oocyte mitochondrial supplementation were performed using different species (human, mouse, bovine and porcine), different assisted reproduction techniques (fertilization, SCNT, parthenogenesis) in a set of different contexts, such as IVF failure, aging and poor oocyte or embryo quality. Curiously, this technique was also tested in good quality oocytes by many authors.

Intriguingly, although there is an ongoing debate on using mitochondria as an oocyte rejuvenation tool (Igarashi et al. 2015, Labarta et al. 2019a, Mobarak et al. 2019), none of the recent clinical studies evaluated the effect of mitochondrial transfer in oocytes of advanced maternal age. Aging was only addressed in one case report (Kong et al. 2003) and in a very small sample (n=4) study conducted in 1999, aiming to evaluate the effects of ooplasm transfer (Lanzendorf et al. 1999).

Oocytes from women with advanced maternal age have lower copies of mtDNA (May-Panloup et al. 2005, Duran et al. 2011). Furthermore, the rate of mtDNA mutations is approximately 15 times higher that of the nuclear genome, probably explained by the proximity with the respiratory chain and the lack of histones and repair enzymes (May-Panloup et al. 2016, Reznichenko et al. 2016). Although controversial, a great number of mtDNA abnormalities has been reported in oocytes from women over 35 years of age (Chan et al. 2005), and the higher rate of mutations represents a plausible explanation for this finding. In addition, ovarian aging is associated with mitochondrial dysfunction. Age-related modifications include mitochondrial morphologic changes, different distribution patterns, lower ATP production, lower membrane potential and higher reactive oxygen species (ROS) production (Eichenlaub-Ritter et al. 2011, Qiao et al. 2014, May-Panloup et al. 2016). In addition, genes related to mitochondrial function are downregulated in oocytes from older women (Steuerwald et al. 2007). Therefore, it seems logical to introduce new good-quality mitochondria into oocytes of advanced maternal age women.

Lower copies of mtDNA were also found in oocytes from women with poor-quality oocyte (not age-related), and this was associated with lower fertilization rates (Reynier et al. 2001, Santos et al. 2006). Therefore, it also seems logical to test the efficacy of mitochondrial supplementation in couples with multiple IVF failures. The AUGMENT treatment had apparent good results in the first clinical trials performed (Fakih 2015, Oktay et al. 2015) but concerns were raised about the methodology of these studies, such as the risk of bias introduced by IVF intercycle variability (Cozzolino et al. 2019, Labarta et al. 2019b). These flaws were clarified with the RCT from Labarta et al. (2019b) demonstrating no benefit of using mitochondrial oocyte supplementation from OSC. While the group from IVI-RMA in Spain disproved the efficiency of the AUGMENT treatment (Labarta et al. 2019a), there is still one clinical trial in progress (ClinicalTrials.gov) (ClinicalTrials.gov known as a database of general and private clinical studies that have been conducted around the world and a clinical studies source founded by the United States NLM (National Library of Medicine).) employing this treatment.

In terms of animal models, most of the studies have compared mitochondrial transfer with IVF and ICSI (both as control groups) in oocytes with presumably normal mitochondrial function and content. Testing the efficacy of mitochondrial supplementation in good-quality oocytes may be problematic for two reasons. First, as already mentioned, the rationale to supplement oocytes with mitochondria is based on a strong body of the evidence that low quality oocytes and/or poor embryo development have fewer mtDNA copies (Reynier et al. 2001, Chan et al. 2005, Santos et al. 2006, Murakoshi et al. 2013, Fragouli et al. 2015) as well as reduced metabolic activity, low ATP content and mitochondrial dysfunction (Eichenlaub-Ritter et al. 2011, Benkhalifa et al. 2014, Schatten et al. 2014, May-Panloup et al. 2016). Second, it has been stated that IVF is a more efficient reproductive technique compared to ICSI concerning in vitro production of embryos in farm animals (Betteridge 2006, Knox 2014) since it requires less manipulation and causes less mechanical damage in oocytes (Schultz & Williams 2002). Additionally, studies in animals showed that fertilization by ICSI cause alterations in the transcriptome of the developing embryo (Giritharan et al. 2010) and may induce a delay in preimplantation embryo cleavage (Dumoulin et al. 2000). Thus, the benefits of mitochondrial transfer should be tested in conditions of low mtDNA and/or mitochondrial dysfunction and the most suited control to mICSI injection should be ICSI and not IVF.

The purported aim of mitochondrial supplementation at the time of fertilization is to increase the amount of mtDNA available to support embryo development until replication resumes. However, two of the analyzed studies have failed to demonstrate a better embryo development despite the higher mtDNA copies at blastocyst stage (Srirattana & St John 2018, Sheng et al. 2019).

Although an increase in mitochondrial content was demonstrated, the viability of mitochondria after transfer was not evaluated and the energy supplied may have been insufficient for the energy demand, which is in agreement with the similar ATP content observed in blastocysts from mitochondria-injection group and control group (Sheng et al. 2019). Accordingly, additional mtDNA copies lead to an increase in the expression of genes involved in the glycolytic metabolism and embryonic development, indicating a state of readiness to promote cellular proliferation, which shows that the mitochondrial and nuclear genomes interact to promote embryonic development (Srirattana & St John 2018).

Oocyte mitochondrial content is distributed to daughter cells during early embryogenesis. Blastomeres will depend on these mitochondria to cell function until resumption of mitochondrial replication, that is thought to occur around the time of implantation, at the expanded blastocyst stage in human (Hashimoto et al. 2017), bovine (Hendriks et al. 2019) and porcine (Cagnone et al. 2016) and E6 day in mouse (Facucho-Oliveira et al. 2007). In line with oocyte, embryo mtDNA copies would be expected to reflect embryo quality. Indeed, low levels of mtDNA were found in uncleaved embryos (Murakoshi et al. 2013). However, the role of mtDNA as a biomarker of viability seems to be more complex.

Klimczac et al. could not demonstrate a relationship between mtDNA content and blastocyst morphology and implantation potential after evaluation of human euploid embryos (Klimczak et al. 2018). On the other hand, higher mtDNA copies have been associated with poor implantation potential in human embryos. This association was perceived in both day 3 and blastocyst embryos (Diez-Juan et al. 2015).These findings have been reproduced by another group (Fragouli et al. 2015) with some disparities. Fragouli et al. also reported a low implantation potential in blastocysts with unusually high mtDNA levels (above a threshold). The authors evaluated mtDNA copies according to women’s age and found high levels of mtDNA in blastocysts from older women (38–42 years old) when compared to younger women (29–37 years old). However, opposite results were found in cleavage stage embryos, with higher mtDNA levels in young women and lower mtDNA levels in older women. A suggestion to explain this paradox was proposed by the authors, in which mitochondrial replication would occur before the blastocyst stage in older women (Fragouli et al. 2015). In addition, it has been proposed that mitochondrial proliferation can be a compensatory mechanism for mitochondrial insufficiency, leading to oxidative damage and dysregulation of pathways involving mitochondria (May-Panloup et al. 2016).

Therefore, higher levels of mtDNA in preimplantation embryos, as seen in ICSI BCB- (derived from developmentally incompetent oocytes) porcine embryos (Cagnone et al. 2016) and aged mouse (mitochondria-injected) embryos (Sheng et al. 2019), compared to matched good-quality embryos, may indicate premature mitochondrial replication.

Although premature mitochondrial replication could potentially indicate poor embryo quality, the clinical value of using quantitative mtDNA for assessing blastocyst viability still needs to be explored.

Apart, these results support the importance of a sufficient oocyte mitochondrial content before fertilization to allow normal embryonic development. Also, equal expression of mitochondrial-related genes (such as TFAM) in mICSI-BCB- compared to ICSI BCB+ (derived from competent oocytes) blastocysts also supports a stabilization effect of mitochondrial transfer, without compensatory mitochondrial proliferation activation.

On treatment safety

None of the studies included in this review reported cases of congenital malformations or chromosomal abnormalities, neither in ooplasmic nor mitochondrial transfer. Though, it is not clear from studies if miscarriages and stillbirths were systematically screened for congenital malformations that could have potentially been induced by epigenetic modifications.

In the context of ooplasmic transplantation, Cohen and coworkers reported two cases of abnormal 45,X fetuses (Barritt et al. 2001a). One was a spontaneous first-trimester miscarriage and the other a termination of pregnancy at 16 weeks. A pervasive developmental syndrome in a 5-year-old child has also been reported (Labarta et al. 2019b). The ooplasm is highly structured, thus subtle problems might well arise if the maturation of the donating oocyte is not in perfect synchrony with the recipient oocyte. Although the number of successful pregnancies achieved following ooplasmic donation is too small to allow any meaningful interpretation, the genetic abnormalities reported raised safety concerns regarding mitochondrial heteroplasmy. Mixing two entities (ooplasm from donor and recipient) could influence the balance of pro-apoptotic and anti-apoptotic factors, affect proteins closely linked to microtubule dynamics, such as mitogen-associated protein (MAPK) and M-phase promoting factor (MPF), modify DNA methylation and disturb mRNA translation (Darbandi et al. 2017). The major concern relates to the possible incompatibility between nuclear and mitochondrial genomes. Indeed, Hua et al. have demonstrated significant alterations in the expression of pluripotency and mitochondrial-related genes in association with mitochondrial heteroplasmy (Hua et al. 2012). Conversely, mtDNA heteroplasmy seems to be common in the general population (May-Panloup et al. 2016) and several murine studies failed to demonstrate mitonuclear incompatibility (Reznichenko et al. 2016).

The studies analyzed in our review confirmed the presence of donor mtDNA in the offspring with no apparent adverse effects on their development, although few tissues have been analyzed. The degree of heteroplasmy may be more relevant in organs/systems that are more dependent on mitochondrial activity and ATP production by oxidative phosphorylation, and that may therefore be more affected by any changes at this level (such as the heart muscle). This issue remains to be investigated in greater detail.

The euploid rate and the absence of reported congenital malformation from the clinical studies are reassuring. Animal research also favors the safety of the technique in that offspring was normal in all except one of the animal studies (Van Thuan et al. 2006, Chiaratti et al. 2011, Li et al. 2017, Wang et al. 2017b) and, as mentioned before, the fetal deaths reported derived from a study using different strains, different sources (cumulus and embryonic cells) and different volumes into mouse oocytes (Van Thuan et al. 2006).

Although unrelated to mitochondrial heteroplasmy, the safety of mitochondrial transfer from an autologous source is still a concern, considering the specificities of the morphologic and bioenergetic profile of mitochondria between oocytes and donor cells.

Oocyte maturation is characterized by a notorious increase in mitochondrial number accompanied also by a shift in mitochondria structure, suggesting functional differentiation. Oocytes from preantral follicles contain about 10,000 mitochondria with many cristae and light matrix, while oocytes from preovulatory follicles have about 100,000 round-shaped mitochondria with few cristae and dense matrix. In addition, the distribution of mitochondria also changes during oocyte maturation, from a homogeneous distribution to assembly close to the spindle and the oolemma, presumably for privileged energy supply. After fertilization, mitochondria acquire an elongated form with many cristae, and gradually changes from glycolysis to OXPHOS, although mitochondrial structure and function seem to be different in the two cell types (trophectoderm and inner cell mass) of preimplantation embryo. This shift from mainly glycolytic to OXPHOS is perceived in the trophectoderm, whereas the inner cell mass cells retains round-shaped mitochondria with few cristae (Motta et al. 2000, Eichenlaub-Ritter et al. 2011).

The dynamic balance among metabolic pathways (glycolysis and OXPHOS) has not only implications in energy supply but has also been implicated in cell regulation and apoptotic processes (Zhang et al. 2018, Rastogi et al. 2019). Therefore, the introduction of morphological and likely functionally different mitochondria into mature oocytes could induce mitochondrial reprogramming and adversely affect embryonic development. Mitochondria from somatic cells, such as GC/cumulus (Eichenlaub-Ritter et al. 2011) have been shown to have many cristae and a light matrix, features that are different from that of the mature oocyte. On the contrary, spherical mitochondria from stem cells are comparable to mature oocytes. Because of this morphological and metabolic (glycolytic) similarity, the use of mitochondria from stem cells have been encouraged by some authors (Schatten et al. 2014, Labarta et al. 2019a), as mentioned before.

Studies analyzed in our review have not explored metabolism-related gene expression in embryos derived from somatic cell mitochondrial supplementation. Gene expression from oocyte-mitochondria source have shown contradictory results, with increased expression of OXPHOS (Srirattana & St John 2018) and glycolytic (Tsai & St John 2018) pathways.

As mentioned, the coordinated expression of mitochondrial and nuclear-encoded genes is important to maintain the normal cellular function. Particularly, TFAM, which is involved in mtDNA stability, transcription and replication, seems to have an essential role in postimplantation embryos, leading to lethality in knockout TFAM mice (Ekstrand et al. 2004). Our review demonstrates a similar range of expression of TFAM in mitochondrial supplementation groups (Cagnone et al. 2016, Lee et al. 2017, Tsai & St John 2018) even when granulosa cells were used as mitochondrial source (Hua et al. 2012). Also, no differences in pluripotency genes, such as OCT4 and SOX2, were reported in this study using somatic cell-mitochondria, although these results were observed in the context of interspecies somatic cell nuclear transfer (iSCNT). Likewise, our review has shown promising results in studies using hepatocytes (Yi et al. 2007) and granulosa cells (Hua et al. 2007) as somatic sources of mitochondria.

The possibility and consequences of mitochondrial reprogramming after oocyte supplementation with other (than oocyte) sources of mitochondria are still unclear and should be explored in further studies.

The possible association between mitochondrial transfer (from OSC) and cardiac defects raised by St. John et al. (2019) does not seem to be well-founded. The authors have suggested a link between mtDNA variants and heart defects, but the non-classified (insertion/deletion at nt9821) and probably damaging (single nucleotide nt12581) variants were found in primordial follicles of the third-generation offspring, in which no heart defect was found. Moreover, the selection of mice for histopathological analysis and mtDNA sequencing was not clear, since out of the total mice from each generation only three were selected without randomization or a specific justification.

When considering epigenetic modifications, hypomethylation of the H19 and SNRPN differentially methylated regions (DMRS), have been associated with Silver-Russel syndrome and SNRPN gene may be implicated in Prader-Willi and Angelman syndromes. Li et al. have explored the methylation patterns of these genes, involved in embryo development, and found similar patterns, adding support to the safety of the procedure (Li et al. 2017), although more experimental data are needed.