Potential conflict of interest: Nothing to report.
MPV17 is a mitochondrial inner membrane protein of unknown function recently recognized as responsible for a mitochondrial DNA depletion syndrome. The aim of this study is to delineate the specific clinical, pathological, biochemical, and molecular features associated with mitochondrial DNA depletion due to MPV17 gene mutations. We report 4 cases from 3 ethnically diverse families with MPV17 mutations. Importantly, 2 of these cases presented with isolated liver failure during infancy without notable neurologic dysfunction. Conclusion: We therefore propose that mutations in the MPV17 gene be considered in the course of evaluating the molecular etiology for isolated, rapidly progressive infantile hepatic failure. (HEPATOLOGY 2007.)
Mitochondrial disorders are established causes of liver failure in early childhood. However, the real incidence is likely to be underestimated because of an underappreciation of this group of disorders, phenotypic heterogeneity, and difficulty in the investigations required to achieve the diagnosis. On the basis of currently available information, most inherited mitochondrial diseases in infants are due to mutations in nuclear genes encoding proteins with specific functions targeted to the mitochondria rather than primary mutations in the mitochondrial genome (mitochondrial DNA [mtDNA]) itself. Mutations in nuclear-encoded genes involved in mtDNA replication and maintenance of mtDNA integrity, including DNA polymerase gamma (POLG), deoxyguanosine kinase (DGUOK), thymidine kinase 2 (TK2), and most recently MPV17, have emerged as being responsible for mtDNA depletion syndromes (low mtDNA copy number). MtDNA depletion typically causes respiratory chain dysfunction with prominent neurological, muscular, and hepatic involvement.1 Mutations in the MPV17 gene that encodes a mitochondrial inner membrane protein were initially identified in 3 families with infantile hepatic mtDNA depletion.2 The phenotype in these patients was characterized by severe liver failure, hypoglycemia, growth retardation, neurological symptoms, and multiple brain lesions during the first year of life.2 Subsequently, a presumed founder MPV17 mutation was also identified in patients with Navajo neurohepatopathy (NNH; MIM 256810), an autosomal recessive disorder prevalent in the Navajo population of the southwestern United States.3, 4 This disorder exhibits severe sensory and motor neuropathy, corneal anesthesia and scarring, cerebral leukoencephalopathy, failure to thrive, recurrent metabolic acidosis with intercurrent illness, and liver disease.3, 5 Decreased respiratory chain activity and mtDNA depletion were demonstrated in liver biopsy samples from 2 NNH patients.6 A homozygous R50Q mutation in the MPV17 gene was identified in all patients with NNH.4
We describe the clinical, biochemical, pathological and molecular features of 4 additional patients from 3 ethnically distinct families with MPV17 mutations. All these cases presented with rapidly progressive liver failure during infancy. Importantly, in 2 patients, neurological dysfunction was not noted.
CS, citrate synthase; ΔCt, difference in the threshold cycle numbers of the nuclear gene and mtDNA; DGUOK, deoxyguanosine kinase; FeCN, ferricyanide; 1H-MRS, proton magnetic resonance spectroscopy; MRI, magnetic resonance imaging; mtDNA, mitochondrial DNA; NA, not available; NADH, nicotinamide adenine dinucleotide; NNH, Navajo neurohepatopathy; ORO, oil red-O; PCR, polymerase chain reaction; POLG, DNA polymerase gamma; SD, standard deviation; SDH, succinate dehydrogenase; TK2, thymidine kinase 2; TM, transmembrane.
Patients and Methods
Samples from the 4 patients included in this study were referred to the Mitochondrial Diagnostics Laboratory at the Baylor College of Medicine for biochemical and molecular studies. The initial workup included biochemical assays of respiratory chain function, determination of the mtDNA copy number, and sequence analysis of the DGUOK and POLG1 genes. Because mutations were not detected in either of these 2 genes despite abnormal electron transport chain activities and reduced mtDNA content, we searched for mutations in the MPV17 gene, which was recently reported to be responsible for a previously unrecognized hepatic mtDNA depletion syndrome.2
The clinical findings of 4 cases caused by MPV17 mutations are summarized in Table 1 and are compared with the findings from patients with documented MPV17 mutations reported in the literature.2, 4
Table 1. Summary of the Clinical and Molecular Findings in Patients with Known MPV17 Mutations
Age of Onset
Failure To Thrive
F indicates female; M, male; and NA, not available.
Hepatocellular carcinoma; liver transplant; living at 21 years
The patient was the first child of healthy consanguineous parents of Middle Eastern descent. The family history was significant for liver disease in 2 members of the family and for premature deaths in infancy of 2 other individuals (Fig. 1). He presented on day 2 of life with focal seizures, cyanosis, and apnea, which was presumed to be secondary to a documented hypoglycemic episode. A head computed tomography at that time revealed an ischemic area in the distribution of the left middle cerebral artery. At the age of 2 months, the patient was noted to have poor weight gain, irritability, and jaundice. On physical examination, he had hepatosplenomegaly. The neurological examination at that time revealed age-appropriate muscle tone. Laboratory findings included low albumin, high conjugated bilirubin, mild increases in aminotransferases, elevated blood lactate of 6.4 mmol/L (normal: 0.7-2.1), increased plasma methionine and tyrosine (15-fold and 10-fold above the upper limit of the normal range, respectively), and a 2-3–fold increase of glutamine and alanine. There was evidence of a coagulopathy with reduced antithrombin III, protein C, and protein S. The coagulopathy and cholestasis progressed in parallel throughout the course of the disease, with intermittent spontaneous improvements lasting days to weeks, until the final irreversible decline in liver function and death at the age of 5 months. With the exception of the residual changes related to the previously noted infarct, the brain magnetic resonance imaging (MRI) was considered normal, and proton magnetic resonance spectroscopy (1H-MRS) did not show an elevated lactate peak. The brain autopsy showed a single vessel thrombosis without evidence of demyelination.
This infant was the younger brother of case 1. He was found by newborn screening to have increased plasma tyrosine, and there was evidence of liver disease from the newborn period. By 10 weeks of age, he exhibited failure to thrive, jaundice, elevated aminotransferases, and a coagulopathy refractory to treatment. Liver function rapidly deteriorated, with progressive hepatomegaly and cholestasis. He had lactic acidosis that peaked at 9 mmol/L. Notably, up to 5 months of age, psychomotor development and neurological examination were normal, and there was no evidence of other organ system involvement. Brain MRI, 1H-MRS, an electroencephalogram, retinal fundoscopy, and an echocardiogram were all normal. The patient received a liver allograft at the age of 5 months but died 1 month later following multiorgan failure associated with sepsis.
This patient was the daughter of an indigenous language–speaking Hispanic couple from an isolated region in Mexico. She presented at the age of 8 months with failure to thrive and at 14 months with hepatomegaly, moderate elevations in aminotransferases (aspartate aminotransferase, 352 U/L, and alanine aminotransferase, 155 U/L), and hyperbilirubinemia (total 9.1 mg/dL, conjugated 5.6 mg/dL). At 18 months, she developed rapidly progressive liver insufficiency (prothrombin time of 36.2 seconds, with albumin level of 3.0 g/dL). Hypotonia and loss of previously attained skills were noted, as the previously developmentally normal patient became increasingly encephalopathic and minimally interactive. She had lactic acidosis, with blood lactic acid at 9 mmol/L. Five days prior to her death from liver failure at the age of 19 months, brain MRI revealed cytotoxic edema involving the deep and subcortical white matter. The 1H-MRS revealed glutamine and lactate at 3 and 10 times the upper limit of normal.
The patient was the child of healthy, unrelated Caucasian parents. He developed hypoglycemia and lactic acidosis on day 1 of life. At 4 weeks of age, he had persistent lactic acidemia (4.9 mmol/L) and hypoalbuminemia. At 6 weeks of age, he developed recurrent episodes of hypoglycemia requiring large doses of intravenous glucose. An echocardiogram and chest radiograph were reported as normal. Plasma tyrosine was increased. Urinary p-hydroxyphenylacetic, p-hydroxyphenyllactic, lactic, and short-chain dicarboxylic acids were elevated without detectable succinylacetone. Liver failure developed and rapidly progressed, with elevated aminotransferases, extremely high α-fetoprotein (157,000 μg/L), a coagulopathy and hematemesis, jaundice, hypoalbuminemia, ascites, edema, decreased urine output, and severe lactic acidosis, with blood lactate up to 28 mmol/L. The child died at 3 months of age from progressive liver failure. At autopsy, the most striking finding was a very enlarged (285 g versus expected 136 g), yellow, fatty liver.
Tissue Histological and Electron Microscopy Studies
All samples were taken after parental informed consent. Liver specimens were obtained through open biopsy in cases 1, 3, and 4 and through explantation in case 2. Patients 3 and 4 died shortly after the biopsy was performed. The liver specimens were fixed in 10% buffered formaldehyde, paraffin-embedded, and stained with hematoxylin and eosin, trichrome, reticulin, and periodic acid–Schiff diastase and for iron. A fresh frozen sample was stained with oil red-O (ORO) for lipids in cases 2 and 4. Case 4 had a muscle biopsy that was subjected to a histochemical study using the following methods: nicotinamide adenine dinucleotide–tetrazolium reductase, ATPase at both pH 9.4 and pH 4.3, cytochrome c oxidase, and ORO. For electron microscopy, biopsy tissues were fixed in 2% or 3% buffered glutaraldehyde, postfixed in 1% osmium tetroxide, dehydrated in graded concentrations of ethyl alcohol followed by propylene oxide, and embedded in Epon 1-μ semithin sections, which were stained with methylene blue and azure II and counterstained with sodium borate and basic fuchsine for light microscopy orientation. Ultrathin sections were stained with uranyl acetate and Sato's lead citrate, mounted on copper grids, and examined with a Phillips CM-12 or Zeiss 912 transmission electron microscope.
Respiratory Chain Enzyme Analysis
Spectrophotometric analysis of the respiratory chain complexes was performed on liver samples from patients 2 (explanted liver), 3, and 4 and on muscle (vastus lateralis of quadriceps femoris) from patients 1 and 4. The liver and muscle samples were immediately frozen in liquid nitrogen after collection, stored at −80°C, and shipped on dry ice. For patient 4, mitochondria were isolated from a fresh muscle sample for polarographic assays of oxidative phosphorylation. Respiratory chain complexes were also assayed from the fresh homogenate.7 Additional muscle and liver specimens were obtained from patient 4 post mortem within 2 hours of death.
The electron transport chain enzymes were assayed at 30°C with a temperature-controlled spectrophotometer (Ultraspec 6300 Pro, Biochrom, Ltd., Cambridge, England). Each assay was performed in duplicate. The activities of nicotinamide adenine dinucleotide (NADH):ferricyanide (FeCN) reductase, succinate dehydrogenase (SDH), rotenone sensitive complex I+III (NADH:cytochrome c reductase), complex II+III (succinate:cytochrome c reductase), and complex IV (cytochrome c oxidase) were measured with appropriate electron acceptors and donors.8, 9 The increase or decrease in the absorbance of cytochrome c at 550 nm was measured for complex I+III, complex II+III, or complex IV. The activity of NADH:FeCN reductase was measured by the oxidation of NADH at 340 nm. For SDH, the reduction of 2,6-dichloroindophenol at 600 nm was measured. Citrate synthase (CS) was used as a marker for the mitochondrial content. Enzyme activities are expressed with respect to both the total protein and CS activity.
Total nuclear DNA and mtDNA were extracted from peripheral blood leukocytes and skin fibroblast cultures with commercially available DNA isolation kits (Gentra Systems Inc., Minneapolis, MN). The mtDNA copy number in various tissues was measured by the real-time quantitative polymerase chain reaction (PCR) method with specific primers for the mitochondrial transfer RNALeu(uur) gene (mtF3212-3231 and mtR3319-3300; see http://www.mitomap.org).10 As a control, a nuclear single-copy gene, β-2-microglobulin, was amplified with primers ntF 589-613 and ntR 674-653 in the 3′UTR (GI:37704390).10 The real-time quantitative PCR was performed twice for each reaction. The 10-μL PCR contained 5X iTaqSYBR Green Supermix with 5-carboxy-X-rhodamine (Bio-Rad, Hercules, CA), 300 nmol/L of each primer, and 0.1-1.0 ng of the total genomic DNA extract. The real-time PCR conditions were 2 minutes at 50°C and 10 minutes at 95°C, followed by 45 cycles of 15 seconds of denaturation at 95°C and 30 seconds of annealing/extension at 63°C. The fluorescent signal intensity of the PCR products was recorded and analyzed on a 7900HT fast real-time PCR system (Applied Biosystems, Foster City, CA) with SDS version 2.2.2 software. The mtDNA content, called the relative mtDNA copy number, was calculated from the difference in the threshold cycle numbers of the nuclear gene and mtDNA (ΔCt). The amount of mtDNA per cell equals 2(2−ΔCt), accounting for the 2 copies of the β-2-microglobulin gene in each cell nucleus. Through a comparison of the mtDNA content of the patient to that of the control, mtDNA depletion could be determined. For muscle, the control value was derived from a pool of 12 samples from age-matched individuals, and for liver, the control value was from a pool of 3 age-matched liver specimens.
Sequence-specific oligonucleotide primers linked to M13 universal primers were designed to amplify all coding exons of the MPV17 gene and at least 50 nucleotides of the flanking introns. PCR products generated with Fast Start DNA polymerase (Roche, Indianapolis, IN) were purified on ExcelaPure 96-well ultrafiltration PCR purification plates (Edge BioSystems, Gaithersburg, MD). Sequencing reactions were performed with the BigDye Terminator cycle sequencing kit (version 3.1), purified with Performa DTR 96-well V3 short plates (Edge BioSystems), and analyzed on an ABI3730XL automated DNA sequencer with Sequencing Analysis Software version 5.1 (Applied Biosystems). DNA sequences were analyzed with Mutation Surveyor version 2.6.1, and the MPV17 Genebank sequence (NT_022184 gene ID 4358) was used as the reference.
The liver samples for histology (4 cases) and electron microscopy (3 cases) were reviewed. Light microscopic features at the time of tissue collection were similar in all samples and were typical for liver failure due to a mitochondriopathy (Fig. 2). In all cases, hepatocyte cytoplasm was severely expanded and contained coarse, evenly dispersed eosinophilic granules, numerous small lipid vacuoles, or both (Fig. 2A,F,I,J). The largest granules measured up to 4-5 μm in diameter and most likely represented abnormal mitochondria (Fig. 2A,K). Distinctive bright eosinophilic granules not associated with lipid vacuoles were seen in a minority of the hepatocytes in 3 cases and were very prevalent in case 3 (Fig. 2A,J). The ORO staining showed most hepatocytes contained lipid droplets of various sizes, from extremely fine to medium-diameter vacuoles (data not shown). The number of hepatocytes containing lipid exceeded expectations based on the appearance in paraffin sections. Canalicular and cytoplasmic bile stasis was common. The prevalence of hepatocyte nuclear immunoreactivity for MIB-1 (Ki67 monoclonal antibody) supported the evidence of regenerative activity, such as scattered oval-shaped regenerative hepatocytes (Fig. 2K). Isolated hepatocyte necrosis was a minor feature in all but case 3. Degenerating and necrotic hepatocytes lacking obvious lipid vacuoles exhibited unusual coarse eosinophilic cytoplasmic granularity. In general, liver inflammation was minimal to absent, and the liver architecture was characterized by delicate periportal and pericellular fibrosis coexisting with areas of intralobular collapse (Fig. 2A,D,F,G).
Ultrastructural studies revealed nonspecific changes related to cholestasis. In most hepatocytes, increased residual bodies, decreased glycogen, increased randomly scattered rough endoplasmic reticulum (often dissociated from mitochondria), and abundant non–membrane-bound minute lipid droplets were observed. Mitochondrial abnormalities were easily detected in almost all hepatocytes and included increased numbers per cell, abnormal variations in the size and shape, distorted tubular dilatation of the cristae, and absence of dense matrix granules (Fig. 2B,C,E,H). The accumulation of an abnormally dense matrix with the displacement of the cristae was a prominent focal feature (Fig. 2C,H). Megamitochondria due to a massive accumulation of the matrix were found only in case 3 (Fig. 2H). There were no paracrystalline inclusions in the matrix or in the intracristal space. Interestingly, the mitochondria in other cell types within the liver and in the muscle (cases 1 and 4) had a normal structure (data not shown). The brain was examined at autopsy in 2 patients (cases 1 and 4). Both had Alzheimer type II astrocytosis, which is consistent with hepatic encephalopathy (data not shown). Patient 1 also had an old ischemic infarct in the left temporal lobe. The size, character, and advanced age of this lesion were consistent with a major arterial occlusion during the perinatal period. There were no features of mitochondrial encephalopathy.
The activities of mitochondrial enzymes, including respiratory chain complexes, are summarized in Table 2. In liver, the CS, NADH:FeCN reductase, and SDH activities, which are not encoded by mtDNA, were found to be increased in all 3 cases studied, in agreement with mitochondrial proliferation, whereas complexes I/III and IV, which include subunits encoded by mtDNA, were found to be deficient in all 3 patients after correction for increased CS activity (Table 2). From the muscle biopsy of case 1 and the postmortem muscle sample of case 4, all enzyme activities, including CS, were reduced, indicating defects in mitochondrial biogenesis or a generalized tissue deterioration, although after correction to reduced CS, complex IV remained deficient (Table 2). However, in fresh muscle obtained from case 4 a few hours before death, the activities of respiratory chain complexes I (rotenone sensitive), II, III, and IV and CS, NADH:FeCN reductase, SDH, carnitine palmitoyl transferases I and II, and enzymes of fatty acid oxidation were all within the normal control range. Mitochondria isolated from the same specimen showed reduced oxidation of tetramethylenephenyldiamine plus ascorbate (52% of the control mean), which suggested reduced function of at least complex IV in the intact muscle mitochondria.
Table 2. Mitochondrial Enzyme Activities in Patients with MPV17 Mutations
NA indicates not available, and SD, standard deviation.
The activities are given as micromoles per minute per gram of wet weight for muscle and as nanomoles per minute per milligram of protein for liver. The first number in the parentheses is the percentage of the control mean, and the second number in the parentheses is the percentage of the control mean corrected to CS activity. Case 4 muscle was postmortem tissue. The other tissues were all biopsies.
381 (154; 63)
19.2 (19; 8)
31.5 (179; 73)
11 (93; 38)
7.7 (29; 12)
765 (310; 53)
48.3 (47; 8)
87.6 (497; 85)
18.8 (71; 12)
760 (308; 100)
61.8 (60; 19)
35.6 (202; 66)
17.6 (149; 48)
25.5 (97; 31)
Controls (n = 10)
Mean ± SD
62.7 ± 25.8
247 ± 127
188 ± 58.7
17.6 ± 5.3
11.8 ± 3.2
26.4 ± 10.8
3.39 (16; 45)
0.27 (11; 29)
0.09 (7; 20)
0.51 (43; 118)
0.31 (9; 25)
0.75 (4; 15)
0.54 (21; 78)
0.27 (22; 81)
0.21 (18; 67)
0.16 (5; 19)
Controls (n = 10)
Mean ± SD
9.6 ± 4.8
20.7 ± 8.7
2.5 ± 1.3
1.2 ± 0.3
1.2 ± 0.5
3.4 ± 1.4
The mtDNA content was measured in all available tissues (Table 3). The liver mtDNA content in patients 2, 3, and 4 was severely reduced to 3.9%, 4.5%, and 3% of the mean of the age and tissue-matched controls, respectively. The mtDNA content in muscle was also reduced, but not as severely as in liver: 12.6% and 8.4% of the matched controls for patients 1 and 4, respectively (Table 3). Although only measured in 1 patient (case 2), the reduction of the mtDNA content in blood was found to be mild (50% of the age-matched mean).
Table 3. Quantification of mtDNA in Patients with MPV17 Mutations
The results are expressed as the copy number per haploid nuclear gene. The percentage of the control is shown in parentheses. The control value of the liver mtDNA copy number was obtained through the pooling of 3 age-matched liver specimens. The control value of the muscle mtDNA copy number was obtained through the pooling of 12 age-matched muscle samples. NA indicates not available.
Deleterious mutations in the MPV17 gene were detected in all 4 cases (Fig. 1 and Table 1). Three novel mutations were identified in 3 patients (cases 1, 2, and 4). Case 3 was found to harbor a homozygous R50W mutation that has previously been reported.2 The R50W mutation changes the basic amino acid arginine to the hydrophobic aromatic amino acid tryptophan. This mutation is predicted to be more drastic than the arginine-to-glutamine (R50Q) change observed in NNH,4 as supported by the observation that in a yeast model devoid of the MPV17 ortholog, partial phenotypic correction was noted upon transfection with a construct containing R50Q, but not with R50W.2 The previously unreported W69X mutation was observed in patients 1 and 2 and is expected to be severe because the truncated mutant protein, if stably synthesized, is missing two-thirds of the C-terminal portion (Fig. 3). Patient 4 was found to be a compound heterozygote for 2 in-frame deletions (c.263_265del3 and c.234_242del9) of 3 and 9 nucleotides.
Mitochondrial respiratory chain disorders are often recognized on the basis of multisystemic involvement that typically includes neurological findings. Liver complications are typically considered a late feature of a multisystem mitochondrial disorder in which neuromuscular disease is prominent. In contrast, the clinical presentation of our 4 cases of mitochondrial disease due to MPV17 mutations was dominated by rapidly progressive liver failure in infancy, and neuromuscular disease was not appreciated (cases 2 and 4), was attributed to a neonatal stroke (case 1), or was of lesser severity than the hepatic disease (case 3).
In the past few years, there has been a substantial increase in the understanding of the molecular basis of mtDNA depletion syndromes. There are at least 3 nuclear genes currently known to be responsible for liver mtDNA depletion: POLG,11DGUOK,12 and the recently discovered MPV17.2 Given the paucity of cases, there are still unanswered questions regarding MPV17 mitochondrial hepatopathy, including specific clinical, pathological, and biochemical features and appropriate testing strategies.
From a clinical perspective, similarly to the cases reported by Spinazzola et al.,2 the onset of liver disease in our patients occurred early, primarily in the first year of life. The presence of lactic acidosis and hypoglycemia are useful diagnostic features, as is the detection of elevated tyrosine by expanded newborn screening. However, these findings are very nonspecific.
With respect to the pathologic features of MPV17-associated liver disease, mixed macrovesicular and microvesicular steatosis and cholestasis have been reported in the livers of patients with NNH and liver failure prior to the discovery of the causative role of the MPV17 gene.6 At the histological level, all 4 of our cases exhibited similar features, including swollen granular hepatocytes, microvesicular steatosis, and focal pericellular and periportal fibrosis. Taken together, these findings are not specific and have been previously observed in other forms of mtDNA depletion due to molecular defects in other nuclear genes such as DGUOK13 and POLG (Alpers syndrome).14 The ultrastructure of MPV17-deficient mitochondria is characterized by dilated and distorted cristae and by central accumulation of the matrix, which displaces the cristae toward the periphery of the organelle and causes enlargement in some instances to an extreme degree. There is also diminished prominence of matrix granules along with a disturbance of the normally intimate relationship of a rough endoplasmic reticulum with mitochondria. In aggregate, these changes also are not unique for MPV17-deficient hepatocytes but constitute a pattern that is highly suggestive of a respiratory chain disorder.13, 15 Thus, our detailed pathological examination does not support the existence of a specific pathological hallmark for MPV17 hepatopathy.
As shown in cases 1 and 4, histochemical, ultrastructural, and respiratory chain studies of skeletal muscle can be uninformative, whereas similar assays on liver samples appear more helpful (Table 2). The observation that mtDNA is depleted in both muscle and liver tissue is diagnostically useful and fairly specific, yet why both tissues exhibit mtDNA depletion but not similar enzymologic deficits remains a puzzle.
To date, only 4 mutations in the MPV17 gene have been reported (Table 1 and Fig. 3). The R50Q mutation has been observed in NNH4 and in 1 Caucasian family.2 A previously reported mutation at the same amino acid position, the R50W mutation,2 was seen also in 1 of our cases, and this suggests that this codon, which includes a CpG dinucleotide, may be a hot spot for mutations. The R50Q mutation appears to be associated with longer survival, and the onset of the liver failure may be preceded by neurological involvement (Table 1). In contrast, all the other MPV17 mutations are associated with a more severe and rapidly progressive disease. We have also detected 3 previously unreported mutations: W69X and 2 deletions (c.263_265del3 and c.234_242del9; Fig. 3). R50 and W69 are 2 of the 6 invariant amino acids in MPV17 that are conserved in yeast, fish, flies, frogs, mice, and humans. Molecular modeling of the MPV17 protein has revealed that this 176 amino acid protein is predicted to contain 4 transmembrane (TM) spans: TM1 from amino acid 18–38, TM2 from 53–73, TM3 from 94–114, and TM4 from 131–151, with short flanking hydrophilic intermembrane and matrix regions (Fig. 3). The R50 residue is located within the intermitochondrial membrane space. The W69 residue is located within the second TM span (Fig. 3). The W69X mutation is expected to be severe because the truncated mutant protein is missing two-thirds of the C-terminal portion of the MPV17 protein, and this allele may well undergo nonsense-mediated decay of the messenger RNA. The in-frame deletion c.234_242del9 (p.79_81del3aa) in patient 4 occurred in the region between TM2 and TM3 and led to the deletion of a putative protein kinase C phosphorylation site (Fig. 3).
From the small number of patients described so far, it is clear that MPV17 mutations lead to mtDNA depletion. However, the specific function of the MPV17 protein remains unknown. The mouse model of MPV17 deficiency generated more than 10 years ago has not been instructive in elucidating the role of this gene in human disease.16 In contrast to humans, the absence of Mpv17 in the mouse is compatible with survival to adulthood and, with the exception of an age-dependent hearing loss, is without significant clinical problems, at least in a controlled laboratory environment.17
The biosynthetic and detoxifying properties of the liver are highly dependent on adenosine triphosphate, and it is possible that liver failure may be triggered by environmental agents that MPV17-deficient hepatocytes are not able to detoxify. Thus, Mpv17−/− mice may not exhibit liver damage because they are not exposed to these environmental factors in a laboratory setting.
Given its localization within the mitochondrial inner membrane and the similarity with the hepatic phenotype caused by DGUOK and TK2 mutations, it is conceivable that MPV17 participates in maintaining the deoxyribonucleotide pool necessary for mtDNA synthesis. In fact, it has been shown that mtDNA integrity is severely affected in patients with a deficiency of succinyl-CoA synthetase, which is tightly associated in a complex with nucleoside diphosphate kinase, an enzyme that is crucial in maintaining the homeostasis of ribonucleotides and deoxyribonucleotides.18, 19
Our cases illustrate that liver failure may be the only presenting feature at the time of disease onset, as reflected by the observation that 2 of our 4 cases initially exhibited normal neurological examinations. Case 1 had a neonatal stroke that, on the basis of autopsy findings, was inconsistent with mitochondrial dysfunction. It is plausible that the stroke may have been the result of an altered anticoagulation cascade due to liver disease because, similar to that of his sibling, his newborn screen was clearly abnormal and suggestive of a perinatal onset of liver dysfunction. Although neurological deterioration would likely have occurred in our patients had they survived through infancy, our experience highlights the importance of suspecting an mtDNA depletion syndrome even in the absence of involvement of other systems.
In an era when liver transplantation is available for liver failure, mitochondrial cytopathies represent a special challenge. With scarce organ availability, poor outcomes in those patients undergoing transplantation, and a lack of effective medical treatment for the underlying mitochondrial disorder, transplantation is usually contraindicated in the presence of multiorgan involvement from a mitochondrial disorder. Although there has been no systematic evaluation of liver transplantation in such patients, it appears that transplantation does not improve survival and may in fact precipitate multiorgan deterioration because of the added stress associated with the procedure. However, this may not be true in the case of mitochondrial disorders with predominant liver involvement. There are reports of at least 4 cases of MPV17 hepatopathy who received liver transplantation with relatively acceptable outcomes.2, 4 The inability to predict the timing and the severity of extrahepatic organ involvement remain major issues in the management of these diseases.
The differing time of onset and severity of neurological deterioration in reported cases of MPV17 mitochondrial cytopathy merit frank discussion in the course of evaluating a child for transplantation candidacy. As we experienced in a subset of our patients, neurologic involvement can clearly be absent at the initial presentation of liver failure. When one is faced with a mitochondrial disorder, clinical deterioration due to the stress of transplantation, infections, and bleeding should also be emphasized and anticipated if transplantation is considered. Unfortunately, therapeutic support of mitochondrial functions in those instances is still not available. A better understanding of the role of MPV17 in mitochondrial functions could shed light on the pathogenesis of the disorder and potentially inform new treatments.
Finally, our first 2 patients showed evidence of liver disease in the expanded newborn screen. As we develop more knowledge about this disorder, early detection and possible early treatment might delay or prevent entirely the development of severe liver disease and improve outcomes. With the increasing availability of appropriate DNA diagnostics, prenatal testing for the causative gene can be offered to affected families.
In summary, we report the occurrence of novel MPV17 mutations in 3 families from different ethnic groups. Liver failure occurred in 2 of the 4 patients in the absence of neurological findings, and therefore we recommend that testing for mtDNA depletion be considered in infants and children with isolated liver failure. Quantitative testing of the mtDNA content provides an attractive approach to establishing a diagnosis, but its sensitivity and specificity remain unknown. The availability of DGUOK, POLG, and MPV17 molecular analyses can help establish the etiology of liver disease in a timely fashion in acutely ill patients.
We thank Dr. Mark Cohen and Dr. Gretta Jacobs, Case Western Reserve University, for their assistance with the pathological descriptions. We appreciate the cooperation of the families of these patients in supporting these investigations.