Eliminating the Ant1 Isoform Produces a Mouse with CPEO Pathology but Normal Ocular Motility

Abstract

purpose. The adenine nucleotide transporter 1 gene (ANT1) encodes an inner mitochondrial membrane protein that transports ATP into the cell. Mutations within ANT1 produce a syndrome of chronic progressive external ophthalmoplegia (CPEO) in humans. Ant1 knockout (Ant1 ^−/−) mice develop cardiomyopathy and mitochondrial myopathy of limb muscles. Because the extraocular muscles (EOM) are preferentially affected in human CPEO, the objective of this study was to determine whether Ant1 ^−/− mice also exhibit an EOM mitochondrial myopathy.

methods. ANT isoform expression of isolated EOMs, EOM morphology and mitochondrial content, mitochondrial structure and function, ocular motility in intact mice, and contractile performance in isolated muscle preparations were examined.

results. Ant1 ^−/− EOMs had the typical appearance of mitochondrial myopathy, including increase in mitochondrial size, number, and oxidative phosphorylation (OXPHOS) staining. However, there were no measurable ocular motor abnormalities in intact Ant1 ^−/− mice, and their isolated EOMs did not show evidence of increased fatigability. EOMs of wild-type mice exhibited higher levels of Ant2 mRNA compared with hindlimb muscle, which may compensate for the Ant1 loss in mutant mouse EOMs and account for the normal EOM function.

conclusions. The Ant1 ^−/− mice provide a model in which to study CPEO pathology and compensatory mechanisms.

Chronic progressive external ophthalmoplegia (CPEO) is a slowly progressing mitochondrial myopathy that affects primarily the extraocular muscles (EOM). Studies examining the EOMs in patients with CPEO have shown mitochondrial proliferation and morphologic abnormalities such as irregularly shaped inner cristae and increased size. ^{1 2} Mutations in the nuclear-encoded mitochondrial protein adenine nucleotide translocator (ANT)-1 have been found in some patients with autosomal dominant progressive external ophthalmoplegia. ^{3 4 5} ANT is an inner mitochondrial membrane ADP/ATP translocator. Known isoforms are encoded by separate genes and are expressed in a tissue-specific manner. Ant1 is predominantly found in the brain and in heart and skeletal muscles. ⁶ Thus, loss of ANT1 is expected to induce the myopathic abnormalities associated with some forms of CPEO. 

A mouse deficient in Ant1 (Ant1 ^−/−) has been shown to have a mitochondrial myopathy affecting heart and skeletal muscles. ⁶ Hindlimb muscles of Ant1 ^−/− mice exhibit abundant ragged-red fibers, a key pathologic finding in human mitochondrial myopathies. ⁶ Histochemical analyses of succinate dehydrogenase (SDH) and cytochrome c oxidase (COX) in skeletal muscles of Ant1 ^−/− mice show increased mitochondrial activity, which in turn reflects greater mitochondrial content. In addition, Ant1 ^−/− mice are exercise intolerant. ⁶ Since limb muscles of Ant1 ^−/− mice show both histologic and functional abnormalities typical of mitochondrial myopathies, the EOMs would also be expected to have histopathologic ^{1 2} and functional deficiencies ^{7 8} analogous to the EOMs of patients with CPEO. It is important to note that EOMs are often spared in muscular dystrophies. ^{9 10} However, findings of ANT1 mutations in patients with CPEO ^{3 4 5} suggested that the limb and the EOM muscles of the Ant1 ^−/− mouse may be affected. To date, an accurate animal model of CPEO is lacking. To determine whether the Ant1 ^−/− mice provide such a model, we examined mitochondrial morphology, oxidative phosphorylation (OXPHOS) activity of EOMs, and EOM contractile function both in vivo and in vitro. 

Materials and Methods

Animals

All animal procedures were approved by the local animal care and use committee and conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Ant1 ^−/− mice were generated by replacing exons 1 to 3 of Ant1 with a PGKneo cassette in 129/SV mice. ⁶ Both Ant1 ^−/− mice and 129/SV wild-type (WT) controls came from an in-house breeding colony. A total of 22 Ant1 ^−/− mice and 17 WT controls were examined, ranging in age from 1 to 18 months, with a smaller subset used for quantitative analysis. 

Histology

Age-matched Ant1 ^−/− and WT mice between 1 month and 1.5 years of age were used for histochemical and ultrastructural analysis of the EOMs. Mice were killed by cervical dislocation while anesthetized [ketamine (80 mg/kg), xylaxine (16 mg/kg)] and decapitated. The left side of each head was fixed in 2% paraformaldehyde/2.5% glutaraldehyde for anatomic analysis of the EOM under transmission electron microscopy (TEM), whereas the right side of the head was dissected further to isolate the eyeball, the attached cone of EOMs, and the Harderian gland. The eyes and the cones were immersed in tissue-freezing medium and flashed frozen with liquid nitrogen for histochemical analysis. The Harderian gland, which surrounds the posterior eye and cone of muscles, was left in place to help support the muscles structurally. 

Both eyes and muscle cones were carefully mounted for sectioning such that cross-sections were obtained starting at the origin of the muscle. Figure 1demonstrates a low-magnification micrograph of a WT mouse, showing the position of the four EOMs surrounding the retractor bulbi, which in turn surrounds the optic nerve. 

Histochemistry of Mitochondrial Enzymes

Cryosections, 7-μm thick, were cut on a cryostat (Leica CM1850; Bannockburn, IL) in the orientation described above. Sections were collected on glass slides and incubated in reaction media for SDH (containing nitroblue tetrazolium) or COX (containing 3,3′-diaminobenzidine tetrahydrochloride), which generate a blue or a brown product, respectively. All sections were processed simultaneously to control for incubation time. Cross-sectional areas of the stained muscles were photographed at ×20 with a Leica DC 300F digital camera and were examined for OXPHOS activity. 

To compare the activity between the Ant1 ^−/− and age-matched WT mice, the intensities of the stains shown in each photograph were digitally analyzed using an image analysis system (Image-Pro Plus 5.0; Media Cybernetics, Silver Spring, MD). For this quantitative analysis four Ant1 ^−/− and four WT mice with confirmed muscle orientation were used. The intensity plot shows the intensity of the stain from the darkest (smaller values) to the lightest (higher values) plotted against the number of pixels at that intensity. This provides a relative measure of intensity to compare the OXPHOS staining between the two groups of mice. To minimize staining variability between samples processed on different days, values were normalized to the largest number of pixels obtained on that day for a particular stain. An ANOVA was performed using SPSS (SPSS, Inc., Chicago, IL) to determine statistical significance between intensity of OXPHOS staining in mutant versus WT mice. 

Electron Microscopy

After 48 hours of fixation, the orbit was further dissected to obtain just the eye and the attached cone of EOMs. The globe and the cone were then dehydrated through a graded series of alcohols followed by propylene oxide before embedding in a resin mixture (Embed 812/Der 736; Electron Microscopy Sciences, Inc., Hartfield, PA) for plastic sectioning. The tissue was carefully embedded and sectioned as described above. Thick sections (0.5 μm), cut with an ultramicrotome (Reichert, Depew, NY) and stained with toluidine blue, were examined to ensure proper orientation of the muscles. Ultrathin silver–gold cross-sections (90–95 nm) were then cut and collected on grids. All sections were stained with uranyl acetate and lead citrate for 30 and 10 minutes, respectively. Using TEM, photographs of mitochondria were taken at varying magnifications (×1000, ×7200, and ×29,000) to examine the structure, the size, and the number of mitochondria. High magnification micrographs (×29,000) were used to calculate the area of each mitochondria within the frame, using an image analysis program (Image-Pro; Media Cybernetics). Micrographs taken at ×7200 magnification were used to count the numbers of mitochondria present within the frame. Four Ant1 ^−/− mice and four WT mice were used for calculating significance, using a two-sample independent t-test. 

Eye Movement Tracking

Eye movement recordings were obtained in four Ant1 ^−/− mice aged 7 to 10 months and four age-matched controls. Recording techniques have been described previously. ^{11 12 13} Briefly, an acrylic fixation pedestal was surgically implanted on each animal’s head several days before the first eye-movement recording session. The pedestal was positioned to maintain the head in a natural-appearing pitch, and in neutral roll and yaw positions. Before each recording session, the recorded eye was pretreated with 0.5% physostigmine ophthalmic drops to limit pupil dilation in darkness. During recordings, the animal’s pedestal was securely bolted to a support armature, and its body was loosely restrained in an acrylic tube attached to the same armature. The assembly was mounted on a turntable and completely enclosed in a drum painted with a high-contrast pattern. Horizontal eye movements were generated by rotating the turntable under manual or computer control. Eye movements were tracked by a commercial video oculography system (ETL200; ISCAN, Burlington, MA) at a sampling rate of 240 samples/s. The eye tracker determined horizontal and vertical linear positions of the pupil and a reference corneal reflection, as well as pupil diameter. These values were subsequently converted to angular eye-in-head position by using a trigonometric algorithm as previously described. ^{11 12 13} The algorithm is based on knowing the distance between the plane of the pupil and the center of corneal curvature (RP), which was determined in each recording session from the relative motion of the pupil and the corneal reflection produced by rotating the video camera around the stationary animal. The algorithm takes into account the fact that RP decreases as the pupil enlarges. The pupil diameter was used to determine instantaneous values for RP, which in turn were used to convert the two separations of the pupil and the corneal reflection centers (expressed in millimeters) to eye-in-head angle. 

Velocity and amplitude characteristics of horizontal fast phases of vestibular nystagmus were determined before and after a “fatigue” stimulus. Fast phases used for quantitative analysis were generated by slowly rotating the animal under manual control in the light. The experimenter monitored the image of the eye and adjusted the rotation speed and the direction to generate a range of fast-phase amplitudes, as well as to assure that eye velocity was low at the moment of fast-phase initiation. The fatigue stimulus consisted of 5 minutes of 0.2 Hz, ±30° amplitude sinusoidal rotation in the light. The parameters of this stimulus were chosen so as to generate large numbers of fast phases in both abducting and adducting directions. Fast phases are more likely to fatigue eye muscles than would slow movements because, due to the mechanical properties of the eyeball and the orbit, fast eye movements require the greatest peak levels of muscular force. The use of fast phases of vestibular nystagmus as a fatiguing stimulus in myopathic mice was motivated by the use of voluntarily generated saccades as a fatigue stimulus in studies of myasthenia gravis in humans. ^{14 15 16}  

Horizontal eye velocity was calculated offline by numerically differentiating the eye position data after smoothing it by convolution with a Blackman window whose cutoff frequency was 80 Hz. Candidate fast phases were identified using a supervised automated program, which searched the records for moments at which eye velocity surpassed 40°/s. The onsets and the offsets of each candidate fast phase were then defined by searching forward and backward from the 40°/s point for the moments at which eye velocity fell below 5°/s. The candidate fast phases and their onsets/offsets were reviewed, and candidates were excluded from analysis if they were disrupted by tracking instabilities or if they were felt to represent slow-phase movements, which occasionally do exceed the velocity threshold of the search program. Linear regression was used to extract the slope of the relation between peak velocity and fast-phase amplitude, with the regression forced to pass through the origin. Abducting and adducting fast phases were treated separately. Most animals were recorded for two sessions, in which case the regression results for the sessions were averaged. 

Extraocular Muscle Function

The contractile properties of EOMs were evaluated in vitro as described previously. ¹⁷ Briefly, whole EOMs were dissected from five control and four Ant1 ^−/− mice aged 7 to 10 months. These included the same eight animals used for the ocular motility arm of the study, plus an additional WT control. Isolated muscles were placed in a small chamber, firmly attached to a force transducer (AE801; SensoNor, Horten, Norway) and a servomotor (Aurora Scientific, Aurora, Canada). The chamber was superfused with a physiologic salt solution: (in mM) 137 NaCl, 5 KCl, 2.0 CaCl₂, 1.0 MgSO₄, 1.0 Na₂HPO₄, 24 NaHCO₃, 11 glucose, and 0.026 d-tubocurarine, bubbled with 95% O₂ to 5% CO₂ to maintain pH at 7.4 at 25°C. Muscles were stretched to the length, giving maximum tetanic force (optimal length, L ₀); force measurements (in newtons) were normalized to muscle cross-sectional area (cm²). The unloaded shortening velocity (V ₀, in L₀ sec⁻¹) was determined with slack tests. After measuring all other functional properties, fatigue was induced by stimulating muscles at a frequency giving approximately one half of maximal tetanic force (50–70 Hz) for 500 ms, followed by a 1.5-second interval between contractions for 10 minutes. 

Quantitative PCR

EOM, heart, and hindlimb muscle samples from 129/SV mice were homogenized in Trizol (GibcoBRL, Rockville, MD), and total RNA was isolated according to the manufacturer’s instructions. Reverse transcription was carried out using Superscript II RNase H⁻ Reverse Transcriptase (Invitrogen, Carlsbad, CA) with random hexamers. cDNA samples (2 μg each) were used for real-time quantitative PCR (qPCR) with the ABI Prism 7700 Sequence Detection System (Applied Biosystems, Foster City, CA). We used the following primer pairs for the Ant isoforms: Ant1 (amplicon size, 75 bp), forward 5′-GAC ACT TGA CTG CTG GAG GAA GA-3′, reverse 5′-TAC ATT GGA CCA AGC ACC TTT G-3′; Ant2 (93 bp), forward 5′-GAC AGA TTC TCT GGG CTT GTC TGT-3′, reverse 5′-AAG AGA AAA CTG GTC AGA TGA ATA TTA TTG-3′. The ABI SYBR Green Master Mix was used according to the manufacturer’s protocol. The relative mRNA abundance in the different tissues was determined with the comparative cycle threshold method. ^{18 19}  

Results

Histochemistry

Staining for the mitochondrial enzymes COX and SDH was much more intense in the EOMs from Ant1 ^−/− mice than in WT mice, demonstrating an increase in mitochondrial content in the mutant mice. Figure 2shows micrographs of COX and SDH staining in EOMs. The WT muscle (Figs. 2C 2D)contained a mix of light- and dark-stained muscle fibers; COX and SDH activities gave the fibers a coarse, granular appearance that corresponded to intracellular mitochondrial distribution. ^{20 21} In contrast, the EOMs from mutant mice show muscle fibers with more intense COX and SDH staining (Figs. 2A and 2B , respectively). There were also areas of darker staining around the fibers, which correspond to subsarcolemmal aggregations of mitochondria. Quantitative analysis of pixel intensity demonstrated significant differences between the Ant1 ^−/− and WT EOM for both the COX and the SDH staining (COX: repeated ANOVA F(92, 552) = 1.45, P = 0.007; SDH: F(98, 588) = 5.25, P < 0.001). 

Interestingly, the retractor bulbi muscle, an accessory EOM, appeared to be even more affected by the absence of Ant1. Figure 3shows micrographs of retractor bulbi stained for COX and SDH in the Ant1 ^−/− and the WT mice. In the WT mice only a few muscle fibers evinced strong COX and SDH reactions. However, in the Ant1 ^−/− mice, nearly all retractor bulbi fibers showed intense COX and SDH activities. This difference would suggest increased mitochondrial content in the Ant1 ^−/− retractor muscle. As in the case of the EOM, quantitative analysis of pixel intensity after COX and SDH staining of retractor bulbi showed significantly greater numbers of dark intensity pixels in the Ant1 ^−/− than in WT mice (COX: ANOVA F(98, 588) = 4.81, P < 0.001; SDH: F(96, 576) = 15.54, P < 0.001). 

Mitochondrial Structure in Ant1 ^−/− EOMs

Electron microscopic examination of mitochondrial content and morphology revealed mitochondrial proliferation in the Ant1 ^−/− EOMs. Figure 4presents electron micrographs of the EOMs from Ant1 ^−/− and WT mice. Significantly more mitochondria were found in the muscle fibers of Ant1 ^−/− than in WT mice (125.9 ± 22.4 vs. 58.1 ± 1.7, Student’s t-test: P < 0.001). As shown in Figures 4A and 4B , the increase in mitochondria in the Ant1 ^−/− EOMs was particularly prominent along the inner side of the sarcolemma; these accumulations of mitochondria result in the appearance of “ragged-red fibers” in light microscopy, typifying this and other mitochondrial myopathies. In contrast, in WT EOM (Fig. 4B)the mitochondria were more evenly distributed in the intermyofibrillar and sarcolemmal spaces. The Ant1 ^−/− micrographs also exhibited a number of “onion” membrane whirls, which may be remnants of mitochondria. 

At higher magnification, mitochondrial size and internal structure were rendered visible (Figs. 4C 4D) . Figure 4Bshows mitochondria in normal EOM fibers; the cristae were intact and evenly distributed throughout the mitochondria. In contrast, the mitochondria from Ant1 ^−/− mice contained swollen cristae, tubular inclusions, and areas devoid of matrix (Figs. 4C 4D) . Neighboring cytosol lacked sarcomeric organization and many “onion” membrane whirls could be identified (Fig. 4C) . In addition, mitochondrial size in the Ant1 ^−/− EOM fibers was significantly greater than in the WT EOMs (122.1 ± 49.6 vs. 48.2 ± 17.2 pixels ² , respectively, Student’s t-test: P = 0.03). 

Eye Movement Tracking

Figure 5shows a typical plot of the peak velocity versus amplitude constructed from the 58 fast phases executed by an Ant1 ^−/− mouse during a single recording session. As in a previous study of the C57BL/6 inbred strain, ¹³ the 129/SV controls and 129/SV-derived mutants exhibited a linear velocity–amplitude relationship, and adducting fast phases were more rapid than abducting fast phases. Note that the peak velocities measured in these experiments are not directly comparable to those published in the previous study, owing to differences in the oculography sampling rate and offline smoothing of the eye positions signals. 

Table 1summarizes the parameters extracted from these plots and demonstrates that Ant1 ^−/− mice and controls were essentially identical in performance, including average fast-phase amplitudes, and slopes and correlation coefficients of the velocity–amplitude relations. The similarity of slopes indicates that Ant1 ^−/− mice did not have slower fast phases, as might be expected if the histologically demonstrable myopathy resulted in a functionally significant limitation in peak forces that extraocular muscles could generate. Similarly, the equivalence of the correlation coefficients indicates that Ant1 ^−/− mice did not exhibit a saturation-type nonlinearity, which might have appeared if the myopathy resulted in a force deficiency only for the largest saccades. The similarity in correlation coefficients also indicates that Ant1 ^−/− mice did not exhibit higher degrees of saccade-to-saccade variations in peak velocity, as have been reported for the saccades after botulism intoxication ²² and in myasthenia gravis. ²³  

The fatigue stimulus generated approximately 80 fast phases of nystagmus per minute. Figure 6shows a small sample of the eye and the table position captured during the fatigue stimulus. There was no systematic difference between the number of fast phases generated in the first and the last 50 seconds of the fatigue period. Table 1demonstrates that the parity of mutants and controls was unaffected by the fatigue stimulus. Neither Ant1 ^−/− nor WT mice exhibited any changes in the quantified fast-phase characteristics. 

Extraocular Muscle Function

The in vivo experiments suggest that the bioenergetic effects of the loss of Ant1 do not affect ocular motility at a phenotypic level. However, they do not exclude the possibility that EOM dysfunction is present under contractile conditions more demanding than what could be generated under in vivo conditions. Thus, to determine further if the loss of Ant1 induced changes in EOM function, we evaluated the contractile characteristics of these muscles in vitro. EOMs from Ant1 ^−/− mice generated less force: maximal tetanic force (P₀) decreased by ∼8% (5.1 ± 0.5 vs. 4.7 ± 0.3 N/cm² for control and Ant1 ^−/−, respectively, P < 0.05). However, the velocity of unloaded shortening (V ₀) was not affected: 9.4 ± 0.3 versus 9.6 ± 0.3 L ₀ sec⁻¹ for control and Ant1 ^−/−, respectively. 

Exercise intolerance is a hallmark of mitochondrial myopathies. We studied the response of control and Ant1 ^−/− EOMs to increased contractile activity using an in vitro fatigue protocol. The typical response of submaximally stimulated muscles (not fully fused tetani) includes a variable force increase early during the fatigue run (potentiation), followed by a monotonic decrease in force (Control in Fig. 7 ). Surprisingly, Ant1 ^−/− EOMs had the same qualitative response; the percent of initial force remaining at the end of the protocol was essentially the same for both groups: 49.1 ± 2.1% versus 53.4 ± 2.6%, control and Ant1 ^−/−, respectively. These data indicate that the mutation did not render these muscles more fatigable, at least under these conditions. 

Expression of ANT Isoforms

The paucity of eye movement abnormalities in vivo and EOM dysfunction in vitro suggest that EOMs may compensate for the absence of Ant1. To examine this issue further, we measured the relative abundance of mRNAs coding for two Ant isoforms, Ant1 and Ant2, in WT EOMs by qPCR. Data are means from 3 samples and are normalized to levels measured in hindlimb skeletal muscle from control mice. Ant1 mRNA abundance in EOM and heart was not significantly different from hindlimb muscle: 0.97 ± 0.12- and 0.81 ± 0.15-fold of skeletal muscle level (mean ± SD). Given that Ant2 is highly expressed in all mouse tissue, except skeletal muscle, ²⁴ we expected Ant2 message levels to be low in EOM following the pattern set by Ant1 expression. Ant2 expression was significantly higher in heart (8.17 ± 3.8-fold versus hindlimb muscle, P < 0.05) and in the extraocular muscle (7.85 ± 2.8-fold greater than in hindlimb muscle, P < 0.05). 

Discussion

Without a properly functioning Ant1 protein to mediate ATP/ADP exchange between the mitochondria and the cytosol, the skeletal muscles of Ant1 ^−/− mice develop a mitochondrial myopathy of limb muscles. ⁶ Our histochemical and morphologic findings indicate that the EOMs of Ant1 ^−/− mice develop similar myopathic changes. Most of these changes are presumably compensatory in nature. Ineffective oxidative phosphorylation triggers mitochondrial biogenesis, as demonstrated by increased SDH and COX activity, increased mitochondrial content and size, and subsarcolemmal accumulation of mitochondria. Other findings are degenerative, such as onion-ring-like cytosolic inclusions (putative mitochondrial remnants). Still other changes reflect derangement of organizational cues by loss of abnormal intramitochondrial structure. These findings are all consistent with histopathologic changes previously described in patients with CPEO. ^{1 2}  

Given the histochemical and structural abnormalities of Ant1 ^−/− EOMs, the previous report of severe exercise intolerance in these animals, ⁶ and the knowledge that saccades in human patients with CPEO have reduced velocities, ^{7 8} we expected our mutant mice to have abnormal eye movements. However, this was not the case. Although our study of ocular motility was limited to assessing peak velocity of fast phases, it is unlikely, given the absence of any slowing of fast phases, that a more extensive assessment would detect any other functional abnormality attributable to insufficiency of force production. This follows from the principles that motion of the eye is opposed by the viscous properties of the eyeball and orbital tissues, and, consequently, EOMs are called on to produce their highest levels of force during rapid movements, such as saccades (or equivalently, fast phases of nystagmus). Disorders that mildly impair EOM force production may manifest as reductions in saccade peak velocities when other aspects of ocular motility (such as eye-movement range) appear unimpaired. ²⁵ While the functional findings were surprising, they are supported by the in vitro data, which detected only a small decrease of EOM P ₀, normal velocity of shortening, and no evidence of greater fatigability. The lack of functional effects despite the prominence of the histologic changes may be attributed to three possibilities. First, limitations in the amount of force generated by single muscle fibers can be compensated by alterations in the firing patterns of the motor neurons. ²⁶ Second, the small force deficit detected in vitro may only be significant when muscles are generating force levels greater than those required to accomplish the largest fast phases we were able to elicit (approximately 12°). Third, the structural alterations we observed (the “myopathic” changes) may actually succeed in compensating for the absence of Ant1. Note that none of these explanations are mutually exclusive. 

The phenotype of the Ant1 ^−/− mouse includes exercise intolerance, which may reflect an intrinsic increase in skeletal-muscle fatigability, decreased cardiac output due to cardiomyopathy, ⁶ or both. In the present study, we tested large numbers of fast phases of vestibular nystagmus in an attempt to fatigue the EOMs. Although fast saccadic rates have been reported to engender ocular motor fatigue in myasthenia gravis ¹⁵ (a disorder of neuromuscular transmission), the ability of saccades to generate fatigue has never been assessed in ocular motor myopathies, i.e., in disorders more analogous to the defect in the Ant1 ^−/− mouse. The Ant1 mutation did not affect the saccadic rate generated by this protocol. Because absence of fatigue under these in vivo conditions could reflect adaptive changes in motor unit recruitment, we also determined the fatigue resistance of isolated EOMs. Surprisingly, Ant1 ^−/− EOMs fatigued at the same rate as control muscles. This finding indicates that absence of Ant1 did not cause the same metabolic dysfunction in EOMs as occurred in skeletal muscles. However, it should be noted that extraocular muscle has an inherent resistance to fatigue in comparison to skeletal muscle (see Ref. ⁹ for review), which may explain the absence of EOM fatigue in the Ant1 ^−/− mice. Potentially, the increased expression in EOMs of Ant2, an isoform not typically found in skeletal muscle, is sufficient to functionally compensate for Ant1 loss. Nevertheless, the Ant1 ^−/− EOMs show the morphologic hallmarks of a mitochondrial myopathy, suggesting that upregulation of Ant2 in EOMs is only partially protective and does not fully suppress abnormal mitochondrial biogenesis. 

Humans with CPEO typically also have ptosis due to dysfunction of the levator palpebralis superioris (LPS). Techniques for assessing LPS function quantitatively have not been developed for the mouse, although our experience in the present study suggests that if there is LPS dysfunction in Ant1 ^−/− mice, it is subtle; gross degrees of ptosis would have interfered with the video oculography, and no such difficulties were encountered. Histologic assessment did demonstrate, however, increased COX and SDH activity in both LPS (data not shown) and the embryologically related retractor bulbi. Thus, it is likely that the Ant1 ^−/− strain could serve as a histologic, but not functional, model of ptosis in CPEO, just as their EOMs can serve as a histologic model of the EOMs in patients with CPEO. It is of interest, however, that other mouse models of human disorders affecting the EOMs also show sparing of the EOMs, whereas the retractor bulbi and LPS muscle are morphologically affected. ^{27 28}  

Conclusions

The histopathologic changes in Ant1 ^−/− EOMs are consistent with findings in patients with CPEO. ^{1 2} Unlike the case in the human patients, however, the histologic findings in the Ant ^−/− mice were not accompanied by significant deficits in ocular motility in vivo or EOM function in vitro. We hypothesize that EOM function in the mice may have been “rescued” by a compensatory upregulation of Ant2 expression, which substituted for the missing ANT1 isoform, a process perhaps not possible in hindlimb skeletal muscle where Ant2 is not typically found. Indeed, the results of our present study suggest it would be of interest to assess levels of Ant2 expression in the EOMs of human CPEO patients. In sum, while Ant1 ^−/− mice cannot serve as a model of the ocular motility deficits in CPEO, they can serve as a model of the histochemical and structural abnormalities, as well as a context in which to investigate the cellular mechanisms involved in compensating for the absence of this protein. 

 Supported by Grants EY-13370 (JSS), EY-014837 (JSS), EY-13238 (JSS), EY12998 (FHA) EY13724 (FHA), NS-41,850 (DCW), NS-21,328 (DCW), P30-EY06360 (Emory); and Department of Veterans Affairs, Research to Prevent Blindness. NJN is a recipient of a Research to Prevent Blindness Lew R. Wasserman Merit Award.

 Submitted for publication June 3, 2005; revised July 1, 2005; accepted September 7, 2005.

 Disclosure: H. Yin, None; J.S. Stahl, None; F.H. Andrade, None; C.A. McMullen, None; S. Webb-Wood, None; N.J. Newman, None; V. Biousse, None; D.C. Wallace, None; M.T. Pardue, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Cross-sectioned cone of muscles stained with COX. The retractor bulbi is outlined with a thin black line while the EOMs are outlined with a thick black line. Optic nerve (ON) is located in the center of the muscle cone while the Harderian gland (HG) wraps around and encases the muscle cone.

Figure 2.

 

Histochemical analyses of key mitochondrial enzymes in Ant1 ^−/− and WT EOMs. (A) COX activity in Ant1 ^−/− EOMs. (B) SDH activity in Ant1 ^−/− EOMs. (C) COX activity in WT EOMs. (D) SDH activity in WT EOMs.

Figure 3.

 

Histochemical analyses of key mitochondrial enzymes in Ant1 ^−/− and WT retractor bulbi muscle. (A) COX activity in Ant1 ^−/− retractor bulbi. (B) SDH activity in Ant1 ^−/− retractor bulbi. (C) COX activity in WT retractor bulbi. (D) SDH activity in WT retractor bulbi.

Figure 4.

 

Electron micrographs of EOMs from Ant1 ^−/− (A, C, D) and WT (B) mice. (A, B) Mitochondria are shown in the subsarcolemmal space of muscle fibers where the mitochondria are more numerous in the Ant1 ^−/− EOMs. Magnification, ×7200. (C, D) High magnification (×25,000) electron micrographs of the mitochondria within the EOMs of Ant1 ^−/− mice. The cristae in the mitochondria of Ant1 ^−/− EOMs are deteriorated or swollen. In addition, numerous onion-like cytosolic inclusions were observed in the mutant EOMs (C).

Figure 5.

 

Relationship of peak velocity to amplitude of 58 fast phases of nystagmus, obtained in an Ant1 ^−/− mouse. Linear regression fits have been superimposed.

Table 1.

 

Average Fast-Phase Amplitudes, Regression Slopes and Correlation Coefficients for the Velocity–Amplitude Relations

		Prefatigue			Postfatigue
				Ant1 −/−	Control	Ant1 −/−	Control
Abduct	Amplitude	6.7 ± 1.2°	7.4 ± 0.8°	7.9 ± 1.7°	7.2 ± 0.9°
	Slope	38.9 ± 6.4 s−1	38.2 ± 3.8 s−1	36.5 ± 6.9 s−1	39.0 ± 3.9 s−1
	r 2	0.97 ± 0.01	0.98 ± 0.01	0.98 ± 0.02	0.98 ± 0.01
Adduct	Amplitude	7.8 ± 1.5°	7.8 ± 1.0°	8.2 ± 1.5°	8.4 ± 1.0°
	Slope	48.6 ± 0.7 s−1	45.2 ± 3.0 s−1	48.0 ± 2.5 s−1	45.0 ± 3.9 s−1
	r 2	0.97 ± 0.02	0.97 ± 0.01	0.97 ± 0.02	0.98 ± 0.01

 Data reported separately for abducting and adducting fast phases of the recorded eye, and before and after the fatigue stimulus. All values are mean ± SD. N = 4 for both Ant1 ^−/− and control groups.

Figure 6.

 

Five-second sample of horizontal eye, vertical eye, and head position during the fatigue stimulus. Horizontal and vertical eye position traces share the same vertical calibration. Arrow indicates one of the many fast phases of nystagmus.

Figure 7.

 

Changes in force during an in vitro fatigue protocol for WT (Control, n = 5) and Ant1 ^−/− (n = 4) EOMs. Data are mean peak force ± SEM determined every 30 seconds for the duration of the fatigue protocol.

Robert James (Dept. of Neurology, Case Western Reserve University, Cleveland, OH) acquired and processed the ocular motility data.