Main

Mitochondrial diseases are the most common metabolic diseases of childhood, with an estimated frequency of 1 in 5000 births.1 These often devastating disorders are clinically characterized by unexplained association of neuromuscular and/or nonneuromuscular symptoms, an often rapidly progressive course, and symptoms involving seemingly unrelated organs or tissues.2–4 The diagnosis of a mitochondrial disorder often relies on the enzymatic analysis of the respiratory chain complexes (RCCs) in muscle homogenates or isolated muscle mitochondria. Muscle tissue assays require a relatively large sample, obtained by a costly, invasive, and potentially dangerous biopsy procedure. The samples are sensitive to temperature changes, prone to spurious results due to metabolite accumulations or mishandling, and may be susceptible to anesthetics used during the biopsy process. In addition, considerable differences exist between clinical laboratories in the character, concentration, and composition of substrates used for RCC assays.5

Proficiency testing programs allow laboratories to regularly evaluate their performance and improve the accuracy of the results they provide to patients. For example, the College of American Pathology provides individual laboratories with unknown specimens for testing, and each participating laboratory receives a report of their performance.6–8 Currently, there is no external quality assurance program for mitochondrial RCC assays or a method for comparison of results between laboratories. It is challenging to identify and distribute adequate, identical human control specimens containing confirmed enzymatic defects to each test center. Frozen tissue from patients confirmed with mitochondrial disease is available only in a very limited amount, insufficient to distribute to many diagnostic centers. Fibroblast lines from patients with known mitochondrial defects are not always available for each complex deficiency, but RCC activities in these fibroblasts are generally low, often not reflective of values obtained from fresh muscle, and assays in fibroblasts are only available in a few diagnostic laboratories.9 Consequently, comparing results from different centers, or even assuring rigorous standardization and controls within a center, have been long-standing hindrances to our ability to adequately diagnose children with mitochondrial disorders.5,10

Caenorhabditis elegans is an aerobic nematode that is 1 mm long and easy to grow inexpensively. The genome of the nematode has been fully sequenced and shares >83% identifiable homology with human genes.11 There is one common wild-type strain with a low forward mutation rate. Animals can be frozen indefinitely, are archived by the Caenorhabditis Genetics Center, and stocks travel easily in the regular mail. Mitochondrial mutants have been identified and well characterized in the nematode.12–16 These mutants manifest clearly defined defects in RCC enzyme activities. They are a readily renewable, inexpensive source of isogenic animals which carry defects within defined steps of electron transport. Available nuclear encoded mutants encoding mitochondrial proteins include gas-1, a mutation in the 49 kD (NDUFS2) subunit of Complex I; mev-1, a mutation in the SDHC subunit of Complex II; and isp-1, a mutation in the Rieske iron sulfur protein subunit of Complex III.12–16 Reproducible knockdown of Complex IV subunits has also been established using interference RNA (RNAi).17 The RCC profile has been characterized for each of these mutations.13,17 Worm mitochondria can be a virtually limitless source of invariable positive controls for mitochondrial RCC enzyme deficiencies, as well as a genetically invariant normal control. Thus, C. elegans is a powerful translational model for human mitochondrial disease. It is clear that nematode mitochondria are not identical to human mitochondria and that the ideal solution to problems in proficiency testing for RCC assays would be a bank of human tissue with defined defects leading to mitochondrial dysfunction. In addition to this goal, we proposed to develop a supply of mitochondria which can be used for quality improvement of mitochondrial RCC enzyme assays using C. elegans mutants that carry invariant, defined defects in electron transport. As a first step, we collected basic information about the assay method of each center. We then compared the results of RCC activities from open-labeled samples, followed by blinded samples designed to reveal potential differences in diagnoses between centers.

MATERIALS AND METHODS

C. elegans

All C. elegans strains were obtained from the Caenorhabditis Genetics Center (Minneapolis, MN). Wild type (N2) and mutants, including a Complex I mutant (gas-1(fc21)), a Complex II mutant (mev-1(kn1)), and a Complex III mutant (isp-1(qm150)), were used in this study. All are canonical missense alleles in nuclear DNA. Worms were grown according to standard protocols, and mitochondrial fractions were prepared from wild-type and mutant animals using established techniques.13

RCC assays

The RCC profile is a collection of spectrophotometric enzyme assays measuring rotenone-sensitive NADH CoQ reductase (Complex I), NADH ferricyanide reductase (first four subunits of Complex I), succinate dehydrogenase (part of Complex II), thenoyltriflouroacetone-inhibited rate of succinate-dichlorophenolindophenol reductase with and without duroquinone, antimycin-sensitive decylubiquinol cytochrome c reductase (Complex III), rotenone-sensitive NADH-cytochrome c reductase (Complexes I + III), antimycin A-inhibited succinate cytochrome c reductase (Complexes II + III), cytochrome c oxidase (Complex IV), and citrate synthase as a mitochondrial internal marker enzyme.18–24 Each laboratory used their usual muscle tissue protocol for patient diagnosis. Five laboratories in the United States and one in Australia received worm mitochondria samples. The results of the Hahn laboratory are designated as A in all the figures comparing results between centers.

Determination of optimal conditions for large-scale preparation, storage, and distribution of mitochondria from C. elegans

The stability of RCC enzyme activities, including Complexes I, II, III, IV, I + III, II + III, and citrate synthase, was studied using mitochondria from wild-type C. elegans, N2. This study was undertaken only by the Hahn laboratory. The original sample is from the same preparation as that which provided laboratory A comparison data of RCC enzyme activities between laboratories. The effects of storage at −80°C for up to 3 months and four freeze-thaws on mitochondrial RCC enzyme activities were also investigated. The study design is illustrated in Figure 1. Briefly, the mitochondria isolated from N2 were aliquoted and stored at −80°C. After a week, aliquots of mitochondria were taken out of the freezer and thawed on ice. Part of the thawed aliquot was used for RCC testing, and this sample was labeled Sample A. The rest of the aliquot was returned to the −80°C freezer and was labeled Sample B. After another month, Sample B was thawed, and part of the aliquot was used for testing RCC enzyme activities. The rest of the aliquot was returned to the freezer and labeled as Sample D. In a similar fashion, Samples C and E–K were generated for up to 3 months storage and four freeze-thaws. The RCC profile was tested at each step.

Fig. 1
figure 1

Study design for C. elegans mitochondrial sample stability study. The samples were set up in the fashion described in “Materials and Methods.” The storage time in −80°C and freeze-thaw conditions are summarized in the right panel. RCC, respiratory chain complex.

Full size image

Mitochondria sample preparation and distribution

Nematode mitochondria were isolated as previously published13 at the Seattle Children's Research Institute, which also processed and mailed samples to each center. Mitochondria extracted from multiple cultures of each genotype were pooled and realiquoted at 200 μg protein per tube. The protein concentration was determined using the Bradford assay. Mitochondrial aliquots were stored at −80°C until shipped to participating centers on dry ice. In the first survey, we provided open-labeled samples to establish the range of enzymatic activities for control and mutants at each test center. Each test center received three strains, wild type (N2), Complex III mutant (isp-1), and Complex I mutant (gas-1). The ratios of the RCC activities of defective mitochondria over the normal control were used for comparison. In the second survey, each test center received three strains, N2, the wild-type control, and two blinded mutants, a Complex III mutant (isp-1) and a Complex II mutant (mev-1). For wild-type normal controls, 800 μg of mitochondrial protein was provided in four tubes. For each of the mutants, 600 μg of mitochondrial protein was provided in three tubes. Samples were shipped on dry ice and arrived frozen. All samples were kept at −80°C until the assay was done. The guidelines for sample preparation were provided to each center with instructions to send results back to the Seattle Children's Hospital Research Institute within 3 months.

Survey of RCC protocols

Each test center completed a survey to compare the methods used for clinical testing. The survey included sample type, sample preparation, methods of calculations, units of enzymatic activity, and methods for each enzymatic activity assay.

RESULTS

The sample stability

All RCC enzyme activities were stable over 3 months, even after four freeze-thaw cycles. The data are shown in Figure 2. Complex IV activity marginally declined over time with successive cycles of freeze-thaw but not at a magnitude that would affect interpretation when compared with mutants. Complex IV activity was missing in Sample J due to a technical error.

Fig. 2
figure 2

C. elegans mitochondrial stability study. The RCC was measured by the Hahn laboratory after multiple cycles of freeze-thaw over time. The panel on the right summarizes the number of cycles of freeze-thaws and length of storage. Complex IV activity was missing in Sample J due to spectrophotometer technical issue. Complex IV activity marginally declined over time with successive cycles of freeze-thaw but not at a magnitude that would affect interpretation, when compared with mutants. Overall, the entire RCC activities remained stable over 3 months after up to four freeze and thaws. CS, citrate synthase; I + III, Complexes I + III; I, Complex I; II + III, Complexes II + III; II, Complex II; III, Complex III; IV, Complex IV.

Full size image

The first survey results

The first survey samples were kept at −80°C before the analysis at each center for a period ranging from 2 weeks to 3 months. The RCC data from each center are summarized in Figure 3. It is known that the Complex III deficiency mutant has significantly decreased Complex III enzyme activity and also reduced Complex I, Complexes I + III, and Complexes II + III activity.25 Only two centers reported Complex III activity, with the average ratio to normal control of 0.02. All test centers were able to detect the reduced enzyme activities of Complex I and Complexes I + III in the Complex-I-deficient mutant. The average ratio of Complex I enzyme activity was 0.56 for NADH dehydrogenase and 0.26 for NADH CoQ reductase. Complexes I + III enzyme activity measured in three centers showed an average ratio to normal control of 0.39, an expected range from previous study.13

Fig. 3
figure 3

Summary of the RCC data on Complexes I and III deficiency mutant mitochondria in the first survey. The ratios of RCC enzyme activities in Complex I deficiency mutant (top panel) and Complex III deficiency mutant (bottom panel) to normal control (N2) were compared. Each column group represents an RCC complex: Complexes I, I + III, II, II + III, III, IV, and CS. Each color represents a different test center. The first column of each group represents the mean of different test centers, and standard deviations are shown as bar. The column labeled as “A” represents the result from Seattle Children's Research Institute. *I denotes NADH dehydrogenase activity. **I denotes measurements of NADH CoQ Reductase.

Full size image

The second survey results

The second survey contained three samples, including one wild-type control (N2), one Complex-III-deficiency mutant (isp-1), and one Complex-II-deficiency mutant (mev-1). The samples were stored by different centers for time periods varying from 10 days to 2.5 months. The RCC data from each diagnostic center are summarized in Figure 4. RCC enzyme activities in mutants were normalized to normal controls, and the ratios are presented. Two centers reported Complex III activity with a mean of 0.08. The Complex II mutant had reduced Complex II and Complexes II + III activities. All six test centers were able to show reduced Complex II and Complexes II + III activities. Five test centers were able to detect the significantly reduced Complex II activity, whereas one test center reported the ratio to normal control as 0.86 (the average ratio to normal control was 0.24). All test centers were able to find significantly reduced Complexes II + III activity (average ratio to normal control is 0.07). However, as the activities were scattered for various enzymes, the exact interpretation was significantly different among centers (Table 1).

Fig. 4
figure 4

Summary of the RCC data on Complexes II and III deficiency mutant mitochondria in the second survey. The ratios of RCC enzyme activities in Complex II deficiency mutant (top panel) and Complex III deficiency mutant (bottom panel) to normal control (N2) were compared. Each column group represents an RCC complex: Complexes I, I + III, II, II + III, III, IV, and CS. Each color represents a different test center. The first column of each group represents the mean of different test centers results, and standard deviations are shown. The column labeled as “A” represents the result from Seattle Children's Research Institute. *I was measured as NADH dehydrogenase. **I was measured as NADH CoQ reductase.

Full size image
Table 1 Summary of interpretations for Complex-III-deficient blind sample
Full size table

RCC protocol survey results

We sent out a survey to collect basic information about methods and sample requirements used for clinical assay at each center. The results are summarized in Table 2. It was notable that some centers use crude homogenates, whereas others use isolated mitochondria for sample preparation. There are variations among protocols, in particular for Complexes I, II + III, III, and IV. For Complexes III and IV, some test centers use initial rates, whereas others use a first-order rate constant. The units for reporting are also variable among centers, from grams of wet tissue to milligrams of protein for normalization of enzymatic activity. To assay Complex I activity, four test centers measure NADH CoQ Reductase activity, while three test centers defined NADH Dehydrogenase as representing Complex I.

Table 2 Survey summary of the RCC protocol used for clinical diagnosis in different test centers
Full size table

DISCUSSION

Enzyme assays of muscle or other tissues are a critical component for diagnosing patients with suspected mitochondrial disorders. The clinician caring for the mystifying patient must process potentially very complicated RCC data with other laboratory studies, as well as the clinical picture. It is unfortunate that there is no external proficiency testing program for these clinical assays given that muscle biopsy is a relatively frequently performed procedure. Multiple issues exist concerning assays of mitochondrial function that have led to this situation. There is no standard assay that is agreed on by clinical laboratories for individual electron transport chain enzymes as the correct assay. Even if the same assay is used, treatment of the specimen before assay may be different, as is the language used in patient reports. The identification of normal controls is also extremely problematic. In addition, it is easily conceivable that regardless of the choice of a standard, the assay would have intrinsic limitations. For example, it is doubtful that any one enzymatic assay could interrogate all possible functions of complicated entity similar to a 45-subunit protein complex. Given all these variables, it is perhaps not surprising that, to date, there has not been any comparison study for the assays that are currently performed in the United States.

Our supply of purified nematode mitochondria of course bypassed many issues that will need to be addressed by true proficiency testing of RCC assays. Our nematode mitochondria were harvested from the whole organism; all laboratories received a purified sample initially processed by a central laboratory. Not all centers isolate mitochondria for their RCC assays. In addition, procurement of human tissue is a very large concern and a very different clinical situation. Tissue harvesting, transport, and treatment on arrival in the laboratory are just a few of the steps that we have completely sidestepped in this study. Also, mitochondria from liver may need to be assayed differently than mitochondria from muscle or fibroblasts. This is not a concern in worms. In addition, although we have been very impressed with the similarity between human and nematode mitochondria, differences do exist. On the one hand, we can note that virtually every mitochondrial protein listed in MitoCarta can be found within the worm genome. Moreover, our laboratory did not find it necessary to modify any protocols originally developed for mammalian mitochondria in their RCC assays (Morgan and Sedensky, Seattle Children's Hospital, personal communication, 2009).

Nevertheless, our results from two surveys of RCC activities are quite encouraging, in that most centers were able to detect electron transport chain defects in mutant samples. For example, all centers that report Complex III activity detected the severely reduced activity in isp-1, and Complex I activity was reported as reduced by all centers for gas-1. This was precisely the result we had envisioned. Nevertheless, there was significant variability between centers for absolute amounts of activity for certain enzymatic assays, particularly I + III, II, III, and IV. Not all centers performed the same RCC assays. For example, some infer Complex III activity by analyzing overlapping Complexes I + III and II + III. Ultimately, this difference in enzyme assays led to significantly different interpretations of identical samples between centers. This is seen in Table 1, where it appears that four centers arrived at very different diagnoses of isp-1. This variation does not reflect widely varying abilities to measure Complex III activity but rather reveals the complexity of interpreting intertwined data, albeit data from a simple animal, that ultimately leads to a diagnosis. This difference in interpretation is probably not surprising, as significant variations of RCC test has been previously reported by different groups.5,10 However, it is of concern as it could potentially affect patient care. The reasons for the discrepant results in specific enzyme activities are probably due to the fact that the diagnostic centers use different protocols in the tissue preparation, kinetic determination, assay buffers, temperatures, and substrate concentrations.5 For Complex I, different subsets of the complex's enzymatic activities can be reported simply as Complex I. We confirmed that the assay methods including sample preparation are variable among participating centers. It is conceivable that these many differences can each potentially contribute to the different RCC assay results.

The French network of mitochondrial disease diagnostic centers undertook a monumental task to standardize the RCC protocols among French test centers.26 The results are not only encouraging but also limited because of a lack of proper quality control. Without proper quality controls, it is difficult to evaluate whether the present French standardization is sufficient to obtain comparable data in all centers.

The muscle samples from true mitochondrial patients would be ideal for proficiency testing, but it is challenging to obtain enough tissue to distribute the same sample to many testing centers at one time. A set of cell lines with diverse enzymatic defect can be another possible option to serve as blinded test samples. However, culturing cell lines with defects in electron transport chains in large quantities is challenging due to very slow growth. Many diagnostic centers do not have a protocol for skin fibroblast samples. Although imperfect, nematode mitochondria may be useful as a quality improvement scheme until true proficiency testing via a consortium of banked patient tissues is available.

Our exploratory study involved only two test panels in 1 year, so it is premature to make any definitive conclusions concerning the assay performance in participating centers. Multiple assays are required to establish the level of discordance in repeated samples and error rates for specific enzymes. Nevertheless, we believe that optimization of the assay using C. elegans would be a feasible project. Laboratories could certainly turn to nematode mitochondria if they encounter problems in an established test, one that might require repeated efforts to ensure that the assay can detect a specific defect. Any doubts about their test's ability to discover a certain enzymatic defect might be relieved by using the worm mitochondria to provide a large supply of genetically identical, defective, mitochondria. In addition, if a laboratory would like to develop a new assay, it might be beneficial to use nematode mitochondria with a defined defect in the development of that test before potentially squandering precious human tissue from a muscle biopsy. In addition, worms with customized defects could also be generated by RNAi knockdown of specific subunits of the RCC. However, this would only be true for nuclear encoded genes. We could not, for example, knock down function of ND1 or ND3 by RNAi. As such they cannot be used to qualify if tests would recognize defects in coenzyme Q binding, which depend on the function of ND1 and ND3.

By establishing a standard protocol for mitochondrial RCC enzyme assays and a quality improvement network it would be possible to exchange and compare data between diagnostic centers. In addition, it would be feasible to collectively survey viable human tissues to help determine the reference range for normal controls and disease states. Ultimately, standardization of RCC enzyme assays could contribute to the highest quality of patient care in the future.

In summary, we were able to supply clinical laboratories with a virtually limitless supply of isogenic animals in which 100% of all tissues share identical mitochondria. Our C. elegans models not only provide us virtually unlimited “normal” mitochondria but also positive controls for mitochondrial defects. The mutants have well-characterized deficiencies in specific subunits of the electron transport chain, but are viable, fertile, and cheap to grow and maintain. Each laboratory received a sample that was identical to those received by all the other centers. This allowed comparison of centers, something that has been quite elusive to date. It was very encouraging to discover that virtually all laboratories easily detected a primary defect in Complex I, II, or III. We envision that clinical laboratories could eventually find that worm mitochondria, which tolerate freezing well, might be a very easy way to check assays when technical issues arise. These strengths, however, coexist with limitations to the study. Our survey can never test very important, concrete features of the clinical laboratories, such as sample preparation, tissue handling, or checking the reference range, issues that are essential in proficiency testing. Nevertheless, our data strongly support future development of standard proficiency testing for mitochondrial RCC enzyme assays. This is critical for quality assurance in the laboratory and for patient care. It seems very promising that mitochondria from C. elegans with specific defects in the electron transport chain can serve as an excellent tool for quality improvement of RCC assays.