In-depth phenotyping reveals common and symptoms in a hemizygous knock-in mouse model (Mut-ko/ki) of mut-type methylmalonic aciduria

Isolated methylmalonic aciduria (MMAuria) is primarily caused by deﬁciency of methylmalonyl-CoA mutase (MMUT or MUT). Biochemically, MUT deficiency results in the accumulation of methylmalonic acid (MMA), propionyl-carnitine (C3) and other metabolites. Patients often exhibit lethargy, failure to thrive and metabolic decompensation leading to coma or even death, with kidney and neurological impairment frequently identified in the long-term. Here, we report a hemizygous mouse model which combines a knock-in (ki) missense allele of Mut with a knock-out (ko) allele ( Mut -ko/ki mice) that was fed a 51%-protein diet from day 12 of life, constituting a bespoke model of MMAuria. Under this diet, mutant mice developed a pronounced metabolic phenotype characterized by drastically increased blood levels of MMA and C3 compared to their littermate controls ( Mut -ki/wt). With this bespoke mouse model, we performed a standardized phenotypic screen to assess the whole-body impairments associated with this strong metabolic condition. We found that Mut -ko/ki mice show common clinical manifestations of MMAuria, including pronounced failure to thrive, indications of mild neurological and kidney dysfunction, and degenerative morphological changes in the liver, along with less well described symptoms such as cardiovascular and hematological abnormalities. The analyses also reveal so far unknown disease characteristics, including low bone mineral density, anxiety-related behaviour and ovarian atrophy. This first phenotypic screening of a MMAuria mouse model confirms its relevance to human disease, reveals new alterations associated with MUT deficiency, and suggests a series of quantifiable readouts that can be used to evaluate potential treatment strategies.

In humans, this reaction represents an important step in propionate catabolism, funneling metabolites from the breakdown of branched-chain amino acids (valine, isoleucine, methionine, and threonine), odd-chain fatty acids, and the side chain of cholesterol into the tricarboxylic acid cycle [2]. Deficient activity of MUT results either from defects of the MUT apo-enzyme or from defects of intracellular synthesis of AdoCbl. Apo-enzyme defects are traditionally divided into mut 0 and mutclasses on the basis of residual in vitro MUT activity, whereby patients classified as mutretain some residual activity and are responsive in vitro to Cbl supplementation while mut 0 are not [3] [4]. This often translates into mutpatients presenting later with milder disease than mut 0 [4] [5]. Biochemically, MMAuria is characterized by the accumulation of metabolites, such as methylmalonic acid (MMA), propionylcarnitine (C3), and 2-methylcitrate in tissues and body fluids [5]. MMAuria patients often present in the newborn period with failure to thrive, lethargy, repeated vomiting and life-threatening metabolic decompensations. Most surviving patients develop chronic kidney damage, and many suffer from neurological deterioration. As a consequence, the risk of severe disability or death is very high [6]. Patients are often supplemented with carnitine and Cbl, and in order to reduce the throughput of the MUT pathway, a dietary restriction involving a low-protein intake is usually implemented [5][7] [8].
To study the pathomechanisms of mut-type MMAuria, mouse models of disease have been utilized. Previously, two Mut knock-out (Mut-ko) mouse models have been generated, resulting in complete null mutants; however, most homozygous pups did not survive past 24 hours [9] [10]. To circumvent this lethality, in separate studies, adeno-associated virus delivery [11] and transgene expression [12] in the liver were introduced. A novel hemizygous mouse model of MMAuria was generated by Forny et al. [13], combining a knock-in (ki) allele J o u r n a l P r e -p r o o f either a high-protein or precursor-enriched diet to increase the throughput of the MUT pathway. These interventions led to more overt signs of the disease, among which were very high levels of accumulating relevant metabolites in blood and tissues [13]. Nevertheless, the dietary-driven increase of the biochemical phenotype could only be shown acutely, as the mice lost weight too rapidly to continue beyond a few days.
To thoroughly investigate in vivo alterations associated with the metabolic impairment in MMAuria, one strategy is to perform a systemic phenotyping study that assesses all clinically relevant organ systems. The full screening approach based on standard protocols may uncover potential mutation-related abnormalities in the whole body in a reliable and reproducible manner, without introducing a researcher's bias as to which systems are expected to be affected [15]. To our knowledge, no mouse model of MMAuria has yet undergone such an unbiased phenotyping screen.
We have developed a new strategy of dietary challenge by feeding our mice with a 51%protein diet from day 12 of life. In this way, Mut-ko/ki mice start accumulating metabolites at a very early age, akin to what is found clinically, but avoid rapid weight loss, enabling long-term investigation. Using this model, we have performed a broad phenotyping screen and found dysfunctions across many organ systems, many of which parallel clinical findings in patients with MMAuria. We have further identified new phenotypic traits that may form the basis of future clinical and molecular investigations.
J o u r n a l P r e -p r o o f J o u r n a l P r e -p r o o f 

Metabolite measurements
Thirty microlitres of blood was collected from tail vein and put on a filter card. Whole blood concentration of MMA, acetylcarnitine and propionylcarnitine was determined on dried blood spots, as previously described [13]. Measurements were performed in duplicates and averages were calculated prior to data analysis.
Plasma was prepared by collecting ~50 uL of blood from tail vein in a tube coated with K 2 -EDTA before centrifugation at 1500 g for 10 min. Determination of MMA concentration: Plasma and urine samples (~20 uL and ~70 uL respectively) were analysed by liquid chromatography mass spectrometry (LC-MS/MS) on a Thermo Scientific UltiMate 3000 Rapid Separation LC coupled to an AB Sciex 5500 TripleQuad mass spectrometer using a commercial kit (Recipe ClinMass® advanced). Plasma and urine samples were first diluted with water (factor 2 and 20 for control plasma and urine respectively, and factor 500 and 50'000 for mutant plasma and urine respectively), to which internal standard (d3-MMA) was added. The samples were vortexed, centrifuged, the supernatant transferred to HPLC vials, and 2 µL were injected into the instrument. Creatinine was determined by enzymatic assay on a routine analyser Alinity c from Abbott Laboratories.

Blood gases measurements
Blood was collected from tail vein and analyzed immediately for determination of pH, pCO 2 and HCO 3 on a Radiometer ABL 800 blood gas analyzer.

B12 treatment
Mice were administrated 50 µg/g hydroxocobalamin daily over 50 days, starting at 2 months of age, given intraperitoneally. After one week of treatment, mice were switched to a 51%protein diet.
J o u r n a l P r e -p r o o f

Optimization of dietary challenge
We previously showed that a switch to a high-protein diet (60%-protein) at approximately day 60 of life triggers a significant increase in metabolite levels in Mut-ko/ki mice [13].
Unfortunately, this is accompanied by drastic weight loss requiring the mice to be sacrificed after 4 days. Based on these data, we attempted a new strategy of dietary challenge. We fed 6-week-old Mut-ko/ki females with a protein-content diet of 51%, 42% and 33%-protein chow that were isocaloric to each other and with a reference chow (16%-protein). As might be expected, a protein-content dependent effect was observed in detection of MMA (Figure 1 A) and C3 normalized to C2 (Figure 1 B) detected in dried blood spots, whereby the 51%-protein diet led to the strongest increase in biochemical parameters. However, Mut-ko/ki mice challenged with a 51%-protein diet unfortunately also displayed drastic weight loss (Figure 1 C). We hypothesized that this was due to a sudden reduction in food intake, linked to the switch to a different type of food and pellet texture (unpublished observations and [20]). To circumvent this dietary change, we supplied mice with the 51%-protein pellets starting from day 12 of life, ensuring that from weaning it was the only solid food available. By this method, although Mut-ko/ki mice were smaller than their littermates (Figure 1  were partially amenable to Cbl treatment. In order to determine which further clinically translatable disease aspects are present in these mice, we performed a full standardized phenotyping of our bespoke MMAuria mouse model, completed as part of the mandate of the German Mouse Clinic (GMC) to characterize mouse models of human disease [15][21] [16]; https://www.mouseclinic.de). In accordance with our optimized protocol, for further studies Mut-ko/ki and Mut-ki/wt mice were challenged with a 51%-protein diet from day 12 of life. Cohorts of 15 mice per sex/genotype were used for in-depth phenotyping. Although 60 mice were generated, 4 were excluded from the final analysis -1 Mut-ki/wt female died after the first test, and 3 Mut-ki/wt males had very low body mass combined with extremely high liver enzyme activities, suggesting that they suffered from chronic liver disease, which was confirmed upon autopsy.

Confirmation that Mut-ko/ki mice show failure to thrive
The first and most obvious phenotypic finding in the Mut-ko/ki mice was that of growth delay in both males and females ( J o u r n a l P r e -p r o o f lumbar spine, we found bone area and bone mineral content (BMC) to be decreased when males and females were combined ( Table 1) or evaluated separately (Error! Reference source not found. and Error! Reference source not found.). As the magnitude of BMC reduction was greater than the decrease in bone area, decreased bone mineral density (BMD) was identified in these mice (Table 1). Therefore, Mut-ko/ki mice have shorter and less dense bones than their littermates. These changes were not due to chronic exposure to metabolic acidosis, since blood pH, pCO 2 and bicarbonate were not found to be changed in selected female mice under the same dietary conditions (Error! Reference source not found.). We further found that Mutko/ki males and females had reduced grip strength using only two paws or with all four paws.
However, regression analysis indicated that this strongly corresponded to their decreased body weight (Error! Reference source not found.). Therefore, muscle function does not appear to be primarily affected by the defect.

Mut-ko/ki mice display hypoactivity and anxiety-related behaviour
We further performed an examination of behavioural and neurological related changes.
Using an open field test, Mut-ko/ki males and females had decreased exploratory behaviour, as they spent less time ( In contrast to these acute tests in a novel environment, when measured over 21 hours in an environment close to home cage conditions, we did not observe any difference in average distance or rearing between Mut-ko/ki and Mut-ki/wt mice (Error! Reference source not found.), although we note the variability between mice in these longer-term tests. In sum, these findings support a hypothesis that Mut-ko/ki mice are anxious and hypoactive, at least when exposed to novel environments.

Further neurological findings
Further neurology and behaviour tests did not identify any major differences between Mutko/ki and Mut-ki/wt mice. Apart from decreased locomotor activity, all observation parameters, part of the SHIRPA protocol, including tremor, transfer arousal, gait, etc were not significantly different between the two groups (Error! Reference source not found.). We also did not identify specific deficiencies of vision, with the caveat that axial length and lens thickness were slightly decreased in some Mut-ko/ki mice (Error! Reference source not found. and Error! Reference source not found.). Finally, we did not find specific hearing deficiencies as determined by acoustic startle (Error! Reference source not found.) and auditory brain response (Error! Reference source not found.) tests.

Evidence of early mild kidney dysfunction
In previous studies, electrolytes were found to be altered in Mut-ko/ki mice, suggestive of kidney dysfunction [13]. Here we found increased plasma levels of sodium, chloride and inorganic phosphate along with decreased total protein and alpha-amylase levels in Mut-ko/ki mice of either sex, while urea levels were increased in females only (

Mut-ko/ki mice show signs of decreased cardiac function and hematological alterations
Awake electrocardiography and echocardiography recordings were performed in Mut-ko/ki and Mut-ki/wt mice to investigate the heart. Echocardiography identified a decrease in cardiac output ( and Error! Reference source not found.). However, electrocardiography revealed a decreased heart rate and a concurrently increased RR length ( identified to be increased in the female mutant cohort, however, following removal of a single outlier, they were found to be unchanged (

J o u r n a l P r e -p r o o f
Hematological investigations further indicated mild macrocytic anemia, as red blood cell counts, hemoglobin and hematocrit were all slightly decreased while mean cell volume and cellular hemoglobin content were increased in mutant mice ( Table 3). These findings were associated with a decrease in plasma iron (Figure 7 A), as well as calculated total iron binding capacity (Figure 7 B) and transferrin saturation (Figure 7 C) in Mut-ko/ki mice. Further, there was evidence of thrombocytopenia as platelet counts and plateletcrit were decreased.
Although platelet distribution width and mean platelet volume were not significantly altered, the proportion of large platelets (PLCR) was slightly reduced, predominantly in females ( Table   3).
Total white blood cell counts, in contrast, were increased in Mut-ko/ki compared to controls (

Alterations in the ovaries of Mut-ko/ki females
Histological investigations in ovaries of Mut-ko/ki females at the age of 20 weeks revealed an ovarian atrophy consisting of predominance of cords of pale stained hyperplastic interstitial tissue with fewer follicles with 100% penetrance (5 out of 5) compared to the control group (0 out of 5) (Figure 8 A and B). Histology of the testes, epididymis, funiculus spermaticus, prostrate, male accessory glands, uterus and vagina did not reveal any significant difference between Mut-ko/ki and their heterozygous littermate controls (data not shown).

Alterations in the liver of Mut-ko/ki mice
Mutant mice had a hepatocellular hypertrophy compared to control mice (Figure 9). The hepatocellular hypertrophy comprised of enlarged hepatocytes tinctorially distinct and granular cytoplasm. High variation in the size of the cell nucleus (anisonucleosis) was also present as well as a high number of binucleated forms and a high rate of intranuclear inclusions in the liver of Mut-ko/ki mutant mice compared to controls.

Discussion
Journal Pre-proof J o u r n a l P r e -p r o o f In addition to elevated metabolite levels, we also found signs of common clinical manifestations of disease. The most prominent one, growth delay accompanied by decreased feeding and drinking, is akin to failure to thrive in patients. In MMAuria failure to thrive is one of the most prominent clinical findings in the first months to years of life and may be the revealing sign in patients that have not been diagnosed through a neonatal metabolic crisis or by newborn screening programs. In our mice, this failure to thrive is possibly related to the hypoactivity we also identified and may be due to complex multi-organ metabolic interactions, which are worth investigating further. We further observed a mild kidney impairment, characterized by small changes in electrolytes but no histological changes.
These findings may be suggestive of early mild and likely pre-clinical kidney alterations in the Mut-ko/ki mice, with the potential of progression over time. In patients, renal complications are common in the course of the disease, with 47% presenting with chronic kidney disease by 6 years, and 12-14% with end-stage renal disease by adulthood [5]. A similar case is also true for neurological changes. Our model did not display the common neurological signs often described in patients, such as movement disorders and seizures [5]. However, we did identify reduced motor performance, altered thermal sensitivity and changes in anxietyrelated behaviour. As with the kidney dysfunction, it is possible that these changes represent mild neurological changes, which either will get worse over time or illustrate differences in disease manifestation between mice and humans. We do note that this is the first evidence of neural dysfunction identified in a model of MMAuria. Altogether, these data validate the J o u r n a l P r e -p r o o f in the number of organelles such as mitochondria, endoplasmic reticulum or peroxisomes [23]. Liver abnormalities have been reported in MMAuria patients, such as hepatomegaly, histological abnormalities from fibrosis to cirrhosis and enlarged mitochondria accompanied with respiratory chain dysfunction [7][24] [25][26] [27]. More rarely, hepatoblastoma and hepatocellular carcinoma have also been documented in MMAuria [28] [29]. Given the dual roles of the liver in metabolite detoxification and energy metabolism, further studies to delineate its role in MMAuria may be invaluable to understanding disease progression.
By comparison with this model on reference chow [13], our new high-protein diet studies reveal, as might be expected, that most measured biochemical and clinical aspects either remain the same or are worsened in severity for mice on the high-protein diet long-term.
Long-term growth comparison indicates that Mut-ko/ki mice have a much more pronounced failure to thrive when on a high protein diet, especially in the first 50 days of life, while both studies found elevated MMA and C3/C2 levels when mice were on a high-protein diet versus a reference diet. However, both with and without a high-protein diet, we found only small perturbations of electrolyte levels in plasma, indicative of mild early kidney dysfunction, but no histological changes in the kidney. This suggests that even on a high-protein diet Mutko/ki mice represents a mild model of kidney disease.

Phenotyping screen reveals rare and potentially new clinical manifestations of disease
In addition to confirming at least mild versions of common clinical disease manifestations, we further identified less common and novel disease symptoms in our mice. This includes decreased iron levels, red blood cells, platelet and plateletcrit, and an increase in mean corpuscular volume. MMAuria patients usually have normal hematology values but neutropenia and pancytopenia were already noted as acute and chronic presentations of the disease [5]. Leukopenia, thrombocytopenia and anemia have also been described in patients [5][30][7] [31], whereby anemia has been potentially linked to chronic kidney failure due to an insufficient erythropoietin production. Our results on one hand reflect the thrombocytopenia described in patients with isolated MMAuria and on the other indicate impaired erythropoiesis and macrocytic anemia, which is a typical consequence of complete cobalamin deficiency [32] but not otherwise known in isolated MMAuria.
In addition, Mut-ko/ki animals showed alterations in electrocardiography recordings, namely elevated RR and QTc lengths. A prolonged QTc interval causes premature action potentials during the late phases of depolarization and can increase the risk of ventricular arrhythmia or abnormal electrical conduction in the heart. Increased QTc interval, along with J o u r n a l P r e -p r o o f cardiomyopathy, is a common and feared manifestation of the related disorder propionic aciduria -a disease due to a defect of propionyl-CoA carboxylase, two enzymatic steps upstream of MUT -and has also been described in MMAuria [5]. Since many of the other disease manifestations are shared between propionic aciduria and MMAuria, it has been hypothesized that alterations in the heart may be secondary due to the accumulation of propionic acid and other metabolites, which are known to be drastically elevated in propionic aciduria and to a lesser degree in MMAuria. Alternatively, these cardiac observations could be secondary to impairments in other organ systems. Further investigation is therefore warranted to confirm the nature of the cardiac defects in this model. Nevertheless, these data highlight the utility of this murine model to bridge the gap from bench to bedside.
In terms of novel findings, we identified decreased bone mineral content and density in Mutko/ki mice, which could be correlated with the risk of osteopenia and osteoporosis previously described in MMAuria patients [5] [8]. It is often discussed whether this risk is iatrogenic in patients receiving a metabolic diet containing low protein, calcium and vitamin D levels [5].
Interestingly, our findings, provided by mutant and control mice given the same high-protein chow, tend to refute this hypothesis. To our knowledge, decreased bone density has not been thoroughly explored in patients so far. However, with the currently increasing life span patients have experienced over the last decades, this may become an important secondary effect later in life and should be monitored. We also found ovarian atrophy, which could be associated with an ovarian follicular insufficiency or impaired regulation of ovarian folliculogenesis. In mice, ovarian atrophy typically represents an age-related change (e.g. > 1 year old) but when it occurs prematurely (20 weeks old) it represents an aberrant state [33].
Although MUT is known to be moderately expressed at the mRNA and protein levels in ovarian tissue in humans (The Human Protein Atlas, https://www.proteinatlas.org/) and mice [34], this is to our knowledge the first time that ovaries have been investigated in MMAuria.
This suggests that the observed alterations might be caused by an impaired follicular reserve in Mut-ko/ki females. This may be consistent with the premature menopause reported in a 36-year-old female patient with MMAuria and a 45-year-old female patient suffering from propionic aciduria [35] [36]. Again, in light of the better clinical management in current times, having children may be an option for some patients in the near future. In this case, it will be interesting to identify whether their ability to do so is affected, and if so, how they can best be supported.

Conclusions
Journal Pre-proof J o u r n a l P r e -p r o o f In sum, we optimized a bespoke mouse model of MMAuria, and by performing a broad phenotypic screen not only identified common clinical manifestations of the disease, but also novel presentations. It will be important to determine whether these latter are truly applicable to humans or possibly due to the divergence between human and mouse biology. We suggest attention should be paid to decreased cardiac function, bone mineral density especially in aging patients, for signs of liver damage in both sexes and reproductive capacity in females.
So far, the treatments that have been attempted in MMAuria mouse models were assessed by investigating individual biomarkers only, thus limiting their clinical assessment. In this study, we identified quantifiable readouts, such as body weight, bone mineral density and content, hypoactivity in the acute setting, plasma electrolyte levels, QTc, and number of follicles. Once confirmed and studied in patients, these readouts would enable a comprehensive clinical output that could be used to assess the efficacy of potential novel therapeutic avenues, such as gene or mRNA therapy. Finally, many of the symptoms identified in these mice are likely related to metabolic alterations affecting energy. A more indepth study of these changes may provide an exciting opportunity to elucidate further pathophysiological mechanisms of the disease.
J o u r n a l P r e -p r o o f Table 1: DEXA analysis of whole mouse and region-specific bone mineral density and content. J o u r n a l P r e -p r o o f