Loss of IP10 does not attenuate disease in the Ndufs4(−/−) mouse model of LS

To evaluate the role of IP10 in the pathogenesis of LS, we generated and assessed disease onset and progression in Ndufs4(−/−) animals with wildtype, heterozygous knockout or homozygous knockout of IP10. This was accomplished by generating a double mutant IP10 knockout Ndufs4 knockout line (IP10 knockout: Jackson laboratory strain 002287; see Methods). qPCR analysis of IP10 expression in brain tissue revealed no difference in IP10 transcript levels in Ndufs4(−/−)/IP10(+/+) and Ndufs4(−/−)/IP10(+/−) mice. IP10 expression was not detected in Ndufs4(−/−) mice homozygous for deletion of IP10, confirming the effect of the knockout allele (Figure 1A).

Neither heterozygous nor homozygous loss of IP10 had a significant impact on measures of health or neurologic disease in the Ndufs4(−/−) model (Figure 1B–G). IP10 loss did not impact body weight or the onset of weight loss (Figure 1B,C). Loss of IP10 also failed to delay the onset of ataxia and forelimb clasping in the Ndufs4(−/−) model (Figure 1D,E). Consistent with these findings, the progressive decline in rotarod performance that occurs in the Ndufs4(−/−) model was not attenuated by loss of IP10 (Figure 1G). The difference in average latency to fall between Ndufs4(−/−)/IP10(+/+) and Ndufs4(−/−)/IP10(+/−) mice and Ndufs4(−/−)/IP10(+/−) and Ndfus4(−/−)/IP10(−/−) mice at P70 was statistically significant but small.

Survival was not altered by IP10 loss in the Ndufs4(−/−) mouse model (Figure 2A). The primary cause of death in Ndufs4(−/−) mice was euthanasia due to weight loss regardless of IP10 status (Figure 2B). As in LS patients, Ndufs4(−/−) mice develop inflammatory, necrotic lesions in the brainstem, in and around the vestibular nucleus [11]. These lesions are enriched in Iba1+ macrophages. An example from a Ndufs4(−/−)/IP10(+/+) mouse at the time of euthanasia is shown in Figure 2C. Inflammation is also common in the olfactory bulb and cerebellum of Ndufs4(−/−)/IP10(+/+) mice (Figure 2D,E). Ndufs4(−/−)/IP10(−/−) animals also have Iba1+ enriched brainstem lesions at the time of death (Figure 2F) and have areas of inflammation in the olfactory bulb (Figure 2G) and cerebellum (Figure 2H). Like the lesions of Ndufs4(−/−)/IP10(+/+) mice, the lesions of Ndufs4(−/−)/IP10(−/−) mice contain macrophages of peripheral origin (Iba1/Cd45 double-positive cells), consistent with our previous work [14] (Figure S1A,B). Thus, loss of IP10 does not protect against the development of Ndufs4(−/−) disease. Together, the data indicate that IP10 is dispensable in Ndufs4(−/−) disease onset and progression.

Loss of IFNγ modestly impacts disease progression and survival in the Ndufs4(−/−) model

To evaluate the role of IFNγ in the pathogenesis of LS, we generated and assessed disease onset and progression in Ndufs4(−/−) animals with wildtype, heterozygous knockout or homozygous knockout of Ifng by crossing the Ndufs4 model with the well-established Ifng knockout line (Jackson laboratory strain 006087; see Methods). We validated the Ifng knockout by assessing IFNγ expression in CD3/CD28 stimulated T-cells isolated from the spleens of Ndufs4(Ctrl)/Ifng(+/+), Ndufs4(Ctrl)/Ifng(+/−) and Ndufs4(Ctrl)/Ifng(−/−) mice using flow cytometry (see Methods). IFNγ signal was induced in cells isolated from animals carrying one or two wildtype alleles of Ifng but not significantly in Ifng(−/−) CD8 T-cells (Figure 3A,B), as expected.

Ndufs4(−/−)/Ifng(−/−) mice displayed a modest reduction in body size and delay in the onset of weight loss compared to Ndufs4(−/−)/Ifng(+/−) and Ndufs4(−/−)/Ifng(+/+) mice (Figure 3C,D). Reduced body weight has previously been reported as a consequence of Ifng deficiency [19]. Neurological symptom onset (ataxia and forelimb clasping) was modestly attenuated by homozygous loss of Ifng (Figure 3E,F; see Methods), and rotarod performance was marginally improved by Ifng knockout at P50, though benefits were not maintained to later ages of P60 and 70 (Figure 3H).

Ifng loss resulted in a gene dosage-dependent increase in survival: median survival was 72 and 83 days, respectively, in Ndufs4(−/−)/Ifng(+/−) and Ndufs4(−/−)/Ifng(−/−) compared to 58 days in the Ndufs4(−/−)/Ifng(+/+) cohort (Figure 4A). The primary cause of death in Ndufs4(−/−) mice, regardless of Ifng status, was euthanasia due to weight loss (Figure 4B). Like Ndufs4(−/−)/Ifng(+/+) mice, Ndufs4(−/−)/Ifng(−/−) mice develop lesions in the brainstem and neuroinflammation in the olfactory bulb and cerebellum (Figure 4C-H). Also, like Ndufs4(−/−)/Ifng(+/+) lesions, the lesions of Ndufs4(−/−)/Ifng(−/−) animals are enriched in peripheral macrophages (Iba1/Cd45 double positive cells) (Figure S1A,C). Ultimately, Ifng loss modestly extends survival but does not prevent disease development or alter overall disease progression in Ndufs4(−/−) mice.

Animals

IP10(−/−) and Ifng(−/−) mice are from the Jackson Lab (strains 002287 and 006087). Ndufs4(+/−) mice were originally obtained from the Palmiter laboratory at the University of Washington, Seattle, Washington, United States, but are also available from the Jackson Laboratory (strain 027058). Strain details are described in Kruse et al. [10]. Ndufs4(−/−), Ifng(−/−) and IP10(−/−) lines are all on the C57/BL6 background. Ndufs4(−/−) mice cannot be used for breeding due to their short lifespan and severe disease. Ndufs4(+/−) mice were bred with Ifng(−/−) or IP10(−/−) mice to produce double heterozygous offspring, which were then crossed to produce Ndufs4(−/−)/Ifng(+/−), Ndufs4(−/−)/Ifng(−/−), Ndfus4(−/−)/IP10(+/−) and Ndufs4(−/−)/IP10(−/−) offspring. Genotyping of the Ndufs4, Ifng and IP10 alleles was performed according to the Jackson laboratory methods (strains 027058, 002287 and 006087 respectively).

Mice were weaned at P20–22 days of age. Ndufs4(−/−) animals were housed with control littermates for warmth as Ndufs4(−/−) mice have low body temperature [10]. Mice were weighed and health assessed a minimum of three times a week. Following the onset of Ndufs4(−/−) symptoms, wet food was provided at the bottom of the cage. Animals were euthanized if they lost 20% of maximum body weight for two consecutive days, were immobile or were found moribund [12, 13]. Mice heterozygous for Ndufs4 have no reported phenotype, so controls consisted of both heterozygous and wild-type Ndufs4 animals. We refer to Ndfus4(Ctrl) mice here for clarity. The Ndufs4(Ctrl) and Ndufs4(−/−) mice wild type for Ifng or IP10 used in this study came from crosses of Ndufs4(+/−)/Ifng(+/−) mice, Ndufs4(+/−)/IP10(+/−) mice and our general Ndufs4 colony. Mice were fed PicoLab Diet 5058 and were on a 12-h light–dark cycle. All animal experiments followed Seattle Children's Research Institute (SCRI) guidelines and were approved by the SCRI IACUC.

Clasping and ataxia were assessed by visual scoring and analysed as previously described [12]. During disease progression, Ndufs4(−/−) animals can display intermittent/transient improvement of symptoms, so here we report whether the animal ever displayed the symptoms for two or more consecutive days. The onset of weight loss is indicated by the age of maximum body weight.

A Med Associates ENV-571M single-lane rotarod was used for the rotarod performance test. A mouse was placed on the rod already rotating at 6 rpm, and latency to fall was timed for a maximum of 600 s while rotation speed remained constant. For each mouse, three trials were performed with a minimum of 10 min between each trial. The best of three trials was reported.

Reverse transcriptase quantitative real-time PCR (RT-qPCR)

Brains were collected immediately following euthanasia (criteria defined in the Animals section above), divided along the midline, and each half flash frozen in liquid nitrogen. Half a frozen brain was ground into a fine powder using a mortar and pestle chilled with liquid nitrogen on a bed of dry ice. RNA was extracted from approximately 30 mg of powdered brain using the Qiagen RNeasy Lipid Tissue Mini Kit. RNA was quantified using a NanoDrop 1000 spectrophotometer and reverse transcribed with the SuperScript VILO cDNA Synthesis Kit (Invitrogen) as described in the product manual. PCR was performed with TaqMan Assays Mm00445235_m1 (Cxcl10/IP10) and Mm02619580_g1 (β-Actin) using TaqMan Fast Advanced Master Mix on a StepOnePlus Real-Time PCR System (Applied Biosystems). Samples were plated in triplicate. Quantification was performed using the relative standard curve method, and IP10 levels were normalised to β-actin levels. The median normalised value of each sample triplicate is graphed here.

Brain collection, immune-staining and imaging

Tissue was collected when euthanasia criteria were met. Mice were fixed-perfused with 20 ml of PBS followed by 20 ml of pre-chilled (4°C) 4% PFA. Immediately following perfusion, the brain was removed and placed in 4% PFA for a minimum of 24 h at 4°C. The brain was then moved to cryoprotection solution (30% sucrose, 1% DMSO, 100 μM glycine, 1XPBS) for a minimum of 48 h at 4°C. Cryopreserved tissue was frozen in Tissue-Tek O.C.T. Compound in cryoblocks on dry ice and stored at −80°C until cryosectioning. Cryoblocks were sectioned at 50 μm thickness using a Leica CM30505 cryostat set at −20°C. Slices were immediately placed in 1XPBS with 1 μg/ml DAPI at 4°C for a maximum of 24 h before mounting on slides. Prior to mounting, slices were examined for the presence of lesions using a fluorescent microscope. Slices with lesions were mounted on SuperFrost Plus slides. Four to five slices were mounted on each slide. Sections from the centre of the lesion were chosen for staining and are shown here.

Mounted sections were baked at 37°C in a white-light LED illuminated incubator for 24 h before storing at −80°C until staining. For immunofluorescent staining, slides were incubated for 24 h in 60°C antigen retrieval buffer (10mM citrate, pH 6.0). Following antigen retrieval, slides were washed for 5 min in 1X PBS on ice before being incubated for 1 h on ice in 1 mg/ml sodium borohydride in PBS. Slides were then washed in 1X PBS with 10 mM glycine for 5 min before being placed in a 0.5 mg/ml Sudan Black in 70% ethanol and incubated overnight with gentle stirring. Slides were then washed for 5 min three times in 1X PBS. Excess moisture was wiped away, and a hybriwell sealing sticker (Grace Bio-Labs, GBL612202) was applied to each slide over the tissue slices. Slides were incubated for 30 min at room temperature in blocking/permeabilization solution (1 mM digitonin, 10% rabbit serum, 0.5% tween in 1X PBS). Following blocking, the hybriwell sealing stickers were removed, and the blocking solution was shaken off the slides. Excess moisture was wiped away, and new hybriwells were placed. The slides were incubated for 24 h at 4°C in antibodies and 1 μg/ml DAPI diluted in the blocking solution. The following fluorescently conjugated primary antibodies were used: Anti-Iba1 (1:150 for Figures 2 and 4 and 1:50 for Figure 1S, Wako 013-26471), anti-Gfap (1:300, clone GA5, Invitrogen 14-9892-82), anti-Cd45 (1:50, clone D3F8Q, Cell Signaling Technologies #19581). After incubation, slides were washed for 5 min four times in 1X PBS. ProLong Gold Antifade was used as mounting media and coverslips sealed with clear nail polish. Slides were stored in an opaque slide box at 4°C.

Imaging in Figures 2 and 4 was performed on a Zeiss LSM710 confocal microscope. Images were collected using a 10X objective with a 0.6X optical zoom. Channels were set to an optical thickness of 15 μm. DAPI was excited with a 405 nm laser, Alexa488 (Gfap, Invitrogen) with a 488 nm laser and fluorochrome 635 (IBA1, Wako) at 633 nm. Images were taken with the same laser intensity and collection filter settings. Gain/brightness/contrast were arbitrarily adjusted for visualisation to make cell morphology clearly visible (intensities not compared). Imaging in Figure S1 was performed on a Zeiss LSM900 confocal microscope. Images were collected using a 20X objective with a 0.45X optical zoom. Channels were set to an optical thickness of 25 μm. DAPI was excited with a 405 nm laser, Alexa555 (Cd45, Cell Signaling Technologies) with a 561 nm laser and fluorochrome 635 (Iba1, Wako) at 640 nm. Images were taken with the same laser intensity and collection filter settings. Gain/brightness/contrast were arbitrarily adjusted for visualisation to make cell morphology clearly visible (intensities not compared).

IFNγ quantification

Spleens were isolated from adult Ndufs4(Ctrl)/Ifng(+/+), Ndufs4(Ctrl)/Ifng(+/−) and Ndufs4(Ctrl)/Ifng(−/−) mice immediately following euthanasia. Spleens were crushed between glass slides in RPMI media with 10% FBS. Cells were pelleted by centrifugation, resuspended and treated with 0.83% ammonium chloride to lyse red blood cells. Cells were passed through a 70-μm cell strainer prior to counting. To stimulate IFNγ production, a six-well plate was coated with 2 μg/ml αCD3 and αCD28 overnight at 4°C. Uncoated plates were used for unstimulated controls. Splenocytes were plated on the coated and uncoated plates at a density of 3 million cells per well for 24 h. Five hours prior to harvesting cells, 1 μg/ml of brefeldein A (BFA) was added to each well to inhibit protein secretion. From each sample, 1.5 × 10⁶ cells were stained for surface antigens (CD44 and CD8), followed by staining for intracellular antigens (IFNγ). All staining was performed on ice and incubated for 45 min protected from light. The Cytofix/Cytoperm kit from BD/Pharmingen was used for intracellular antibody staining as per the manufacturer's instructions. Flow cytometric analysis of stained cells was performed on an LSRII Fortessa (BD Biosciences, San Jose, CA). The following antibodies from BioLegend were used at the indicated dilutions: CD8a BV650 (Clone 53-6.7, 1:100), CD44 APC Cy7 (clone IM7, 1:400) and IFN gamma PE (1:100, clone XMG1.2).

Statistical analysis

All statistical analyses were performed using GraphPad Prism 10.0.0 with statistical tests detailed in figure legends. Error bars represent the standard error of the mean (SEM).

Scientific rigour