Expansion of Electron Transport Chain Mutants That Cause Anesthetic-Induced Toxicity in Drosophila melanogaster

Abstract

The mitochondrial electron transport chain (mETC) contains molecular targets of volatile general anesthetics (VGAs), which places individuals with mETC mutations at risk for anesthetic complications, as exemplified by patients with Leigh syndrome (LS). The Drosophila melanogaster homozygous mutant for ND-23, which encodes a subunit of mETC Complex I, replicates numerous characteristics of LS, including neurodegeneration, shortened lifespan, behavioral anesthetic hypersensitivity, and toxicity. The anesthetic phenotype of toxicity (lethality) is also observed in flies homozygous for mutations in other Complex I subunits. By contrast, mutations conferring sensitivity have not yet been identified for subunits of Complexes II–V. Furthermore, anesthetic phenotypes are thought to be recessive; that is, risk is not conferred by heterozygous mutations. However, at older ages, exposure of heterozygous mutant ND-23 flies to the VGA isoflurane in 75% oxygen (hyperoxia) results in toxicity. It is also unknown whether combinations of heterozygous mutations in different subunits of the mETC can result in anesthetic toxicity. Here, we show that, following exposure to isoflurane in hyperoxia, flies carrying heterozygous mutations in two Complex I subunits, ND-23 and ND-SGDH (NADH dehydrogenase (ubiquinone) SGDH subunit), had a level of anesthetic toxicity that exceeded the added toxicities of the individual heterozygous mutations. In addition, we show that flies heterozygous for two different alleles of the Complex II gene SdhB were susceptible to isoflurane/hyperoxia-induced anesthetic toxicity. Finally, a mutation in the SdhC subunit of Complex II of Caenorhabditis elegans resulted in isoflurane-induced mortality, supporting the role of Complex II in anesthetic toxicity. These data expand the landscape of mutations in the mETC that increase sensitivity to anesthetic toxicity.

1. Introduction

Volatile general anesthetics (VGAs) in common use today are generally safe. The extent to which the genetic background can increase the risk of adverse effects from exposure to VGAs is only partially understood. Leigh syndrome (LS) is an example of a pharmacogenetic risk factor for the perioperative period [1]. The majority of LS cases are caused by mutations in Complex I of the mitochondrial electron transport chain (mETC). LS results in early severe neurodegeneration and a shortened lifespan, and it is accompanied by hypersensitivity to VGAs [2]. Additionally, individuals with LS are at increased risk of perioperative complications [3], which is attributed, at least partially, to increased susceptibility to the depressant effect of VGAs on oxidative phosphorylation [4]. Unfortunately, no treatment is available for LS. The key to developing effective treatments is an improved understanding of the pathomechanisms of mitochondrial disorders, and to achieve that, animal models are essential.

In Drosophila melanogaster, the ND-23⁶⁰¹¹⁴ allele carries a point mutation in the nuclearly encoded Complex I subunit ND-23 (the Drosophila ortholog of mammalian NDUFS8). Flies homozygous for this allele (ND-23⁶⁰¹¹⁴/ND-23⁶⁰¹¹⁴) reproduce features of LS, including neurodegeneration and a shortened lifespan [5]. Moreover, these animals are hypersensitive to the behavioral effects of the VGAs sevoflurane and isoflurane [6] and, as young adults (i.e., 10–13 days old), they acquire a lethal anesthetic-induced toxicity phenotype, presenting as mortality at 24 h following exposure to the VGA isoflurane. Notably, mortality is increased by hyperoxia and suppressed by hypoxia [7], mimicking oxygen-dependent mortality in rodent models of LS [8]. By contrast, young, heterozygous ND-23⁶⁰¹¹⁴ flies (ND-23⁶⁰¹¹⁴/+) are phenotypically normal and do not show the anesthetic phenotypes. However, at 30–35 days old, they become susceptible to isoflurane-induced toxicity [7]. We also found that flies mutant for other Complex I subunits, including the nuclearly encoded ND-SGDH^G6447 (the Drosophila ortholog of mammalian NDUFB5), are susceptible to isoflurane-induced toxicity at 15–20 days old [9]. However, susceptibility is not conferred by alleles of other Complex I subunits (i.e., ND-13B^EY07792, ND-39^EY06323, ND-15^MI10315, and ND-B22^KG08637) or alleles of Complex II-V subunits (Complex II: SdhB^c00364, Complex III: UQCR-Q^KG08923, Complex IV: levy^EY01834, and Complex V: ATPsynB^EY03793). Thus, the isoflurane-induced toxicity phenotype is specific to alleles of genes encoding particular mETC subunits and depends on the age at the time of isoflurane exposure as well as on the oxygen concentration during isoflurane exposure.

Here, we expand upon our prior studies by investigating the interaction of compound heterozygous mutations in different Complex I subunits in shaping the penetrance of the isoflurane-induced toxicity phenotype. We also examined mutations in Complex II subunits both in Drosophila and in the nematode Caenorhabditis elegans (C. elegans) to test for conservation of the isoflurane-induced toxicity phenotype across phyla.

2. Materials and Methods

2.1. Fly Lines and Culturing

All flies were raised at 25 °C on cornmeal–molasses food, as previously described [6]. ND-SGDH^G6447, SdhB^EY12081, and SdhB^c000364 were obtained from the Bloomington Drosophila Stock Center (Stock IDs 27208, 20343, and 10039, respectively). ND-23⁶⁰¹¹⁴, ND-23^def (Df(3R)Exel8162), and Canton S were provided by Barry Ganetzky (UW-Madison). ND-23⁶⁰¹¹⁴/ND-SGDH^G6447 and ND-23⁶⁰¹¹⁴/+ flies were generated by crossing ND-23⁶⁰¹¹⁴ females with ND-SGDH^G6447 and Canton S males, respectively. ND-SGDH^G6447/+ flies were generated by crossing Canton S females with ND-SGDH^G6447 males. ND-23^def/+ and ND-23^def/ND-SGDH^G6447 flies were generated by crossing ND-23^def/TM6B females with Canton S and ND-SGDH^G6447 males, respectively. SdhB^EY12081/SdhB^c000364 flies were generated by crossing SdhB^EY12081/CyO females with SdhB^C000364/CyO males.

2.2. Anesthetic and O₂ Exposure

Flies were raised at 25 °C for up to 5 days post-eclosion, transferred to vials of 35 mixed-sex flies using CO₂, and incubated at 29 °C until the age of anesthetic exposure. The temperature change from 25 °C to 29 °C served to remain consistent with prior studies [5,6,7]. Vials were placed on their sides to prevent flies from sticking to the food. Flies were transferred to fresh food vials three times per week. On the day of exposure, flies were transferred to 50 mL conical tubes without use of CO₂ and placed onto the serial anesthesia array (SAA) [6]. Each 50 mL conical tube represented a biological replicate. A mixture of O₂ and air (21% O₂, 79% N₂) (Airgas, Madison, WI, USA) was used to generate hyperoxic (75% O₂) conditions using a clinical anesthesia machine (Aestiva/5; Datex-Ohemda) and confirmed by an inline, internal O₂ sensor. Commercial agent-specific vaporizers delivered 2% isoflurane for 2 h (i.e., 4% h (2% × 2 h)) at room temperature (Ohmeda Isotec 5). After exposure, the SAA was flushed with air for 5 min. Flies were transferred to fresh food culture vials and incubated at 29 °C for 24 h. Percent mortality was determined by counting the number of dead flies and dividing by the total number of flies per vial. Mortality is a binary readout and was assessed by an unblinded observer. Each experiment was conducted on at least three separate days. The change in mortality (ΔMortality) was determined by subtracting the average percent mortality from unexposed controls from the percent mortality of each vial exposed to anesthetic. Data were analyzed using GraphPad Prism (version 10.0.2) software. Data were tested for normality using the Shapiro–Wilk test. p-values from comparisons of normally distributed data were generated from Bonferroni’s multiple comparisons test following an ANOVA. p-values from comparisons of non-normally distributed data were generated from Dunn’s multiple comparisons test following a Kruskal–Wallis test.

2.3. Nematode Culturing and Exposures

Behavior in Anesthetic—Nematode culture and anesthetic exposure was performed as previously described [10]. Briefly, 20–50 adult hermaphrodites were placed on agar plates fed with the bacteria OP50 and allowed to lay eggs for 4 h. The adults were removed, and the eggs were grown at 20 °C overnight. A total of 20–25 L1 animals were then transferred to new agar plates with OP50 as a food source. These plates were grown for 2–3 more days until the animals were first-day, egg-laying adults. The plates were then transferred into a glass chamber and exposed to 6.5% isoflurane for 3 h. The plates were removed, observed after 1 h to ensure all adults had regained mobility, and returned to a 20 °C incubator. Each plate was scored 24 h later for mortality. p-values for single comparisons made between non-normally distributed data were generated from two-tailed Mann–Whitney tests.

3. Results

3.1. Combining Heterozygous Mutations in Different Subunits of Complex I Significantly Increases Isoflurane-Induced Toxicity

To investigate the effects of combined heterozygous mutations in the mETC subunit on anesthetic toxicity, we examined mutations in the Complex I subunits ND-23 (orthologous to mammalian NDUFS8) and ND-SGDH (NADH dehydrogenase (ubiquinone) SGDH subunit). We previously found that exposure of ND-23⁶⁰¹¹⁴/ND-23⁶⁰¹¹⁴ flies at 10–13 [7] or 8–13 [9] days old and ND-SGDH^G6447/ND-SGDH^G6447 flies at 15–20 days old [9] to 4% h isoflurane in 75% O₂, hereafter referred to as hyperoxic isoflurane, caused a significant increase in percent mortality over unexposed flies, and exposure of wild-type Canton S flies at any of these ages did not affect mortality. Moreover, we found that multiple allelic combinations of ND-23 mutants produced a similar phenotype.

To follow up on these findings, we examined heterozygous flies at an older age. Exposure of 30–36 day-old Canton S flies to hyperoxic isoflurane did not increase mortality at 24 h; however, it increased the mortality of ND-23⁶⁰¹¹⁴/+ and ND-SGDH^G6447/+ flies (Figure 1A). Moreover, we found that flies heterozygous for both ND-23⁶⁰¹¹⁴ and ND-SGDH^G6447 had substantially higher mortality than the individual mutants. We then isolated the isoflurane-induced mortality by subtracting the unexposed from the exposed mortality for each genotype (Figure 1A’). We found that the isoflurane-induced mortality of ND-23⁶⁰¹¹⁴/+ and ND-SGDH^G6447/+ flies did not differ significantly (p > 0.99), but the mortality of double heterozygous ND-23⁶⁰¹¹⁴/ND-SGDH^G6447 flies was higher than either mutation alone (p < 0.0001, for both). These results support our prior finding that aging increases the risk of isoflurane-induced toxicity for flies heterozygous for ND-23 or ND-SGDH. Moreover, we conclude that the combination of mutations in distinct subunits of Complex I increases the susceptibility to isoflurane-induced toxicity in older flies.

To test the allelic specificity of the toxicity phenotype, we combined a deficiency allele of ND-23 (ND-23^def) with ND-SGDH^G6447. In this experiment, flies were exposed at a younger age, 24–30 days old, because ND-23^def/ND-SGDH^G6447 flies had low viability at 30–36 days old. Hyperoxic isoflurane did not significantly increase the mortality of Canton S, ND-23^def/+, or ND-23^def/ND-SGDH^G6447 flies, but there was a trend toward an increase for all three genotypes (Figure 1B). Isolating the isoflurane-induced mortality reflected this result (Figure 1B’). These data indicate that introducing a stronger allele of ND-23 may uncover comorbidities that ultimately decrease the resolution of this acute toxicity phenotype.

3.2. Compound Heterozygous Mutations in the SdhB Subunit of Complex II Cause Isoflurane-Induced Toxicity in Older Flies

We previously found that 1–6, 8–13, 15–20, and 22–27-day-old flies homozygous for SdhB^c00364 (Succinate dehydrogenase, subunit B), which encodes a Complex II subunit, showed no toxicity to hyperoxic isoflurane [9]. To determine if toxicity is conferred by other alleles, we tested SdhB^EY12081 as a heterozygote (SdhB^EY12081/CyO) and as a compound heterozygote with SdhB^c00364 (SdhB^EY12081/SdhB^c00364). Exposure of Canton S and SdhB^EY12081/CyO flies to hyperoxic isoflurane at 1–6, 8–13 and 15–20 days old did not increase 24 h mortality over unexposed flies (Figure 2). However, exposure of 29–34 day old SdhB^EY12081/SdhB^c0036 flies significantly increased mortality (p = 0.0007), whereas mortality was not affected for Canton S flies (p = 0.57) (Figure 3). These data indicate that aging sensitizes compound heterozygous SdhB^EY12081/SdhB^c0036 flies but not heterozygous SdhB^c00364 flies to isoflurane-induced toxicity.

3.3. Mutations in the SdhC subunit of Complex II (mev-1) Sensitizes C. elegans to Isoflurane-Induced Toxicity

We investigated the organism specificity of the anesthetic toxicity phenotype using the nematode C. elegans. A missense mutation in mev-1, orthologous to the SdhC subunit of Complex II in humans, has been shown to cause hypersensitivity to oxidative stress and premature aging [11]. We exposed wild-type N2 and mutant mev-1(kn1) worms to 6.5% isoflurane in room air (21% O₂) on the first day of adulthood. Mortality was scored 24 h following exposure. mev-1(kn1) mutants experienced a significant increase in mortality following exposure to isoflurane relative to unexposed (p < 0.0001), whereas N2 worms did not (p > 0.99) (Figure 4). These data indicate that mutations in Complex II can confer sensitivity to isoflurane-induced toxicity across phyla.

4. Discussion

Data presented here show that, following exposure to hyperoxic isoflurane, older flies carrying heterozygous mutations in two Complex I subunits, ND-23 and ND-SGDH, had a level of mortality that was higher than the added mortalities of the individual heterozygous mutations. In addition, the data show that older flies compound heterozygous for two different alleles of the Complex II gene SdhB were susceptible to hyperoxic isoflurane-induced mortality. Thus, in older individuals, sensitization to anesthetic toxicity is not limited to homozygous mutations in Complex I genes.

We are not aware of experimental data directly comparable to ours. The suppressive effect of VGAs on oxidative phosphorylation was discovered in isolated mitochondrial preparations where the Complex I function was found to be particularly sensitive [12,13,14]. Behavioral experiments in Caenorhabditis elegans carrying homozygous mutations in single subunits of Complex I (gas-1 analogous to Nudfs-2) but not in Complex II (mev-1, analogous to SdhC) agreed and assigned Complex I the role of the primary mediator of the behavioral endpoints of the VGAs halothane and isoflurane [4,15]. Furthermore, an observational study in young children also documented hypersensitivity of the central nervous system to the VGA sevoflurane only in carriers of Complex I mutations [2]. To provide context for our findings with data on behavioral sensitivity, it is important to compare the different models. In contrast to other in vivo studies, we used (i) anesthetic-induced mortality (i.e., toxicity) assessed at 24 h after exposure as an endpoint. Experiments testing behavioral sensitivity showed that Complex I mutants were hypersensitive to the VGAs sevoflurane and isoflurane [6] in accordance with behavioral findings in other models; (ii) aged flies as opposed to neonates typically used for experiments addressing toxicity; (iii) isoflurane in 75% O₂ as opposed to air; (iv) heterozygous mutations simultaneously present in different subunits of Complex I; and (v) compound mutations in the SdhB subunit of Complex II. Therefore, our data represent a “proof of principle” that combinations of alleles in different subunits of mitochondrial Complexes I can act more than additively (Figure 1) to result in a phenotype of susceptibility to anesthetic toxicity in aged individuals. Furthermore, the results suggest that gene dose matters for its penetrance (Figure 3). This information is important because mutations in more than one gene coding for a subunit of the mETC are present in approximately half of all patients diagnosed with mitochondrial disease [16], but pharmacogenetic factors are rarely, if at all, considered to contribute to anesthetic complications in adult and aged individuals.

Complex II mutations are rare, even though, ironically, the SdhA subunit was the site of the first reported mutation causing a defect in the oxidative phosphorylation [17]. SdhA mutations preferentially present with an LS-like phenotype, while SdhB and SdhC mutations present with malignancies in adulthood, necessitating surgical intervention and, hence, exposure to anesthesia. To date, only 61 patients have been reported in the literature, harboring a total of 32 pathogenic variants in the Complex II genes [18]. This rarity could be due to the non-viability of most homozygous carriers of mutations in Complex II, possibly because of the unique role of Complex II as the only direct link between the Krebs cycle (located in the mitochondrial matrix) and oxidative phosphorylation (located at the inner mitochondrial membrane). Therefore, substantial functional disruption of SdhB via homozygosity of deleterious mutations will likely result in non-viability. In contrast, as all subunits of Complex II are nuclearly encoded, minor alleles and compound mutations in the general population, when present, are likely to be asymptomatic.

We examined compound mutants of the SdhB subunit. Our findings that SdhB^EY12081/SdhB^c00364 flies showed isoflurane-induced toxicity, but SdhB^c00364/SdhB^c00364 did not, indicates that SdhB^c00364 is a functionally weaker allele than SdhB^EY10281. This conclusion is supported by the fact that SdhB^c00364/SdhB^c00364 flies are homozygous viable, but SdhB^EY12081/SdhB ^EY12081 flies are homozygous lethal. Furthermore, the fact that SdhB^EY12081 and SdhB^c00364 have similar mutations (i.e., transposons inserted near the 5′ end of the gene, located 40 base pairs apart in opposite orientations) suggests that mutations compatible with survival into adulthood rarely meet the criteria of impairing Complex II function sufficiently to present an anesthetic toxicity phenotype. It is also compatible with the lack of a behavioral anesthetic phenotype in the SdhC subunit mutant mev-1 [15]. Notably, we found that mev-1 carries a distinct anesthetic toxicity phenotype even under normoxic conditions (mev-1 was isolated as a strain hypersensitive to O₂ [19]) when assessed in the 24 h mortality assay (Figure 4), thereby mimicking the dissociation between behavioral sensitivity to and toxicity of the VGAs sevoflurane and isoflurane described previously [6,7].

It is tempting to speculate whether the simultaneous presence of two minor alleles—one in Complex I and the other in Complex II—would present with an anesthetic toxicity phenotype. We are not aware of experimental data addressing this issue, but clinical data can provide a basis for speculation. Generally, compound heterozygosity may lower the penetrance compared to homozygous mutants of one of the alleles. This is the likely cause for the viability of and lack of anesthetic-induced mortality in one of the SdhB alleles tested. Some compound heterozygous combinations may provide an evolutionary advantage under certain environmental conditions, as exemplified by beta-hemoglobinopathies [20]. However, at least one reported case of the Complex II compound heterozygosity affecting the SdhA subunit presented clinically with LS [21], which is the common clinical presentation of homozygous SdhA mutants. It is noteworthy that the SdhA subunit is the only Complex II subunit in which dominant pathogenic variants have been identified [18]. We have not yet tested SdhA subunits. However, our established experimental system would be a convenient platform to this end.

5. Limitations

SdhB heterozygous and compound heterozygous mutant flies were tested at different ages. Therefore, toxicity could be due to age or compound heterozygosity for SdhB mutations. Hyperoxia and isoflurane were not independently tested, so toxicity may be due to hyperoxia, isoflurane, or a combination of the two. Flies heterozygous for Complex I and II mutations simultaneously were not tested; thus, it is not known if toxicity is limited only to mutations in two subunits of the same complex. Only one allelic combination each for Complex I and Complex II exhibited a significant increase in anesthetic-induced mortality, so sensitization to toxic effects due to these combinations may be allele-specific.