Evaluating the association of biallelic OGDHL variants with significant phenotypic heterogeneity

Abstract

Background: Biallelic variants in OGDHL, encoding part of the α-ketoglutarate dehydrogenase complex, have been associated with highly heterogeneous neurological and neurodevelopmental disorders. However, the validity of this association remains to be confirmed. A second OGDHL patient cohort was recruited to carefully assess the gene-disease relationship.

Methods: Using an unbiased genotype-first approach, we screened large, multiethnic aggregated sequencing datasets worldwide for biallelic OGDHL variants. We used CRISPR/Cas9 to generate zebrafish knockouts of ogdhl, ogdh paralogs, and dhtkd1 to investigate functional relationships and impact during development. Functional complementation with patient variant transcripts was conducted to systematically assess protein functionality as a readout for pathogenicity.

Results: A cohort of 14 individuals from 12 unrelated families exhibited highly variable clinical phenotypes, with the majority of them presenting at least one additional variant, potentially accounting for a blended phenotype and complicating phenotypic understanding. We also uncovered extreme clinical heterogeneity and high allele frequencies, occasionally incompatible with a fully penetrant recessive disorder. Human cDNA of previously described and new variants were tested in an ogdhl zebrafish knockout model, adding functional evidence for variant reclassification. We disclosed evidence of hypomorphic alleles as well as a loss-of-function variant without deleterious effects in zebrafish variant testing also showing discordant familial segregation, challenging the relationship of OGDHL as a conventional Mendelian gene. Going further, we uncovered evidence for a complex compensatory relationship among OGDH, OGDHL, and DHTKD1 isoenzymes that are associated with neurodevelopmental disorders and exhibit complex transcriptional compensation patterns with partial functional redundancy.

Conclusions: Based on the results of genetic, clinical, and functional studies, we formed three hypotheses in which to frame observations: biallelic OGDHL variants lead to a highly variable monogenic disorder, variants in OGDHL are following a complex pattern of inheritance, or they may not be causative at all. Our study further highlights the continuing challenges of assessing the validity of reported disease-gene associations and effects of variants identified in these genes. This is particularly more complicated in making genetic diagnoses based on identification of variants in genes presenting a highly heterogenous phenotype such as "OGDHL-related disorders".

Keywords: 2-oxo acid dehydrogenase; Disease model; Genetic compensation; Mitochondria; Neurodevelopmental disorders; OGDHL; Variant testing; Zebrafish.

Fig. 1

Identification of OGDHL variants in 12 families and mapping of known variants. A Pedigrees and segregation data for the 12 families included in this study. Affected and unaffected individuals are indicated by filled and open squares (males) and circles (females), respectively, and a triangle represents a pregnancy. Affected individuals are indicated by black arrows. Double lines indicate consanguinity. Genetic diagnoses were made in 14 individuals. B The position of changed coding sequence in genomic DNA and the drawing of resulting variants. Seven previously uncharacterized variants described in this study are marked in bold blue. C Seven out of eight missense variants from our studies are highly conserved across species

Fig. 2

Neuroimaging findings of individuals with OGDHL pathogenic variants. Brain MRI findings of individual 1 (A-D), 6 (E–H), and 7 (I-L). Sagittal T2 (A and I) and T1 (E) weighted images showed markedly hypoplastic corpus callosum in individual 1 (A) and 7 (I) and dysplastic corpus callosum in individual 6 (E) with hypoplastic rostrum, genu, and anterior body and absent posterior body and splenium (yellow arrows). Mega cisterna magna was present in all affected individuals (A, E, and I, blue arrows), and individual 6 also had inferior vermian hypoplasia and widening of the foramen of Magendie (E, asterisk). Axial T2-weighted images (B, C, F, G, J, and K) revealed varying degrees of diffuse white matter volume loss, most severe in individual 6 (F and G) and 7 (J and K). Individual 7 also had scattered areas of leukomalacia (J, red arrows) and prominent involvement of brainstem and cerebellar white matter. Individual 6 had ventriculomegaly and colpocephaly (F, orange arrows). Coronal T2 (D and L) and T1 (H) weighted images showed hypoplastic olfactory bulbs in individual 1 (D) and 7 (L) (green arrows) and hypoplastic hippocampi in individual 6 (H, purple arrows)

Fig. 3

The morphological phenotypes of zebrafish ogdhl F₀ knockouts. A Representative embryo images of uninjected, ogdhl F₀ and ogdhl F₀ co-injected with 200 picogram (pg) human OGDHL mRNA (F₀ + OGDHL) at 3 dpf. Arrow denotes heart edema. B,C Size measurements for head (blue line) and eye (red line) as indicated from genotypes in A were calculated as a percentage difference compared to the mean value of the uninjected embryos. D The presence of heart edema (black arrow in A) was calculated as a percentage of total embryos of each group positive for edema. E Locomotor activities of zebrafish larvae in light and dark conditions at 5 dpf, n = 96 larvae for each group. The larvae were habituated in the dark for 30 min, followed by three cycles of 10-min periods of light and dark. Error bars represent the mean ± SEM. Dark period (D), light period (L). F Average cumulative distance traveled by each larva during three cycles of either light or dark periods. Error bars = mean ± SD. For B, C, and F, each dot represents one embryo, and the mean value of each group is indicated at the bottom of the respective bar in the figure. For B–D, the number of embryos for uninjected = 57, Cas9 protein-injected (Cas9 inj.) = 60, ogdhl F₀ = 63, F₀ + OGDHL (100 pg) = 34, F₀ + OGDHL (150 pg) = 35 and F₀ + OGDHL (200 pg) = 32. Error bars = mean ± SD. Statistical significance was calculated by Brown–Forsthye and Welch’s ANOVA with Dunnett’s T3 multiple comparisons test: not significant (ns) p ≥ 0.05, *p < 0.05, ***p < 0.001, and ****p < 0.0001

Fig. 4

Comparing mRNA rescue levels of OGDHL and OGDH in Dhtkd1 and Ogdh paralog-deficient zebrafish. A Representative image of uninjected, Cas9-injected, single knockout (dhtkd1, ogdha, and ogdhb F₀), double knockout (ogdha;ogdhb F₀), and triple knockout (ogdha;ogdhb;ogdhl F₀) embryos at 3 dpf. B, C Size measurements for head and eye with or without human OGDHL co-injection. The number of embryos for uninjected = 66, Cas9-injected = 60, dhtkd1 F₀ = 36, dhtkd1 F₀ + OGDHL = 36, ogdha F₀ = 66, ogdha F₀ + OGDHL = 36, ogdhb F₀ = 66, ogdhb F₀ + OGDHL = 36, ogdha;ogdhb F₀ = 36, ogdha;ogdhb F₀ + OGDHL = 36, ogdha;ogdhb;ogdhl F₀ = 36 and ogdha;ogdhb;ogdhl F₀ + OGDHL = 36. D, E The size measurements for head and eye with or without human OGDH co-injection. The number of embryos for uninjected = 60, dhtkd1 F₀ = 36, dhtkd1 F₀ + OGDH = 36, ogdha F₀ = 36, ogdha F₀ + OGDH = 36, ogdhb F₀ = 36, ogdhb F₀ + OGDH = 36, ogdhl F₀ = 36, and ogdhl F₀ + OGDH = 36. All values were calculated as a percentage difference compared to the mean value of the uninjected embryos. Each dot represents one embryo, and the mean value of each group is indicated at the bottom of the respective bar in the figure. Error bars = mean ± SD. Statistical significance was calculated by Brown–Forsthye and Welch’s ANOVA with Dunnett’s T3 multiple comparisons test: not significant (ns) p ≥ 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001

Fig. 5

Cell apoptosis was activated by ogdhl loss-of-function. A–C Representative image of uninjected, ogdhl F₀ and ogdhl F₀ co-injected with human OGDHL embryos at 3 dpf after TUNEL staining. Scale bars = 100 μm. A Lateral view of eyes. Anterior to the left and dorsal to the top. B Dorsal view of brains. Anterior to the left. C Lateral view of trunks. Anterior to the left and dorsal to the top. D–G Quantification of the number of TUNEL-positive cells in the eye, midbrain, hindbrain, and spinal cords. The mean value of each group is indicated at the bottom of the respective bar in the figure. TUNEL-positive cells were calculated as a percentage difference compared to the mean value of uninjected embryos. In D, each dot represents one eye. uninjected = 31, ogdhl F₀ = 36 and F₀ + OGDHL = 38 eyes. In E–G, each dot represents one animal. uninjected = 16, ogdhl F₀ = 18 and F₀ + OGDHL = 19 animals. Error bars = mean ± SD. Statistical significance was calculated by Brown–Forsthye and Welch’s ANOVA with Dunnett’s T3 multiple comparisons test: not significant (ns) p ≥ 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001

Fig. 6

Zebrafish ogdhl F₀ mutant displays motor neuron abnormalities. A–L Representative image of the trunk region of uninjected, Cas9-injected, ogdhl F₀ and F₀ + OGDHL in Tg(olig2:DsRed2;mnx1:GFP);Albino embryos at 3 dpf. M The measurement of axon length was started from spinal cord to the end of latest posterior branch (indicated by red arrow) of caudal primary motor neurons (CaP) (indicated by red line). Axon lengths were calculated as a percentage difference compared to the mean value of uninjected embryos. Each dot represents one axon. Uninjected = 24, Cas9-injected = 24, ogdhl F₀ = 30, and F₀ + OGDHL = 26. N The measurements of axon angle (indicated by magenta lines, start from spinal cord). Each dot represents one axon. n = 30 axons for each group. Error bars = mean ± SD. The mean value of each group was indicated at the bottom of the respective bar in the figure. Statistical significance in M and N was calculated by Brown–Forsthye and Welch’s ANOVA with Dunnett’s T3 multiple comparisons test: not significant (ns) p ≥ 0.05, **p < 0.01, and ****p < 0.0001. O Cartoon figures indicate aberrant motor axons observed in ogdhl F₀ embryos such as cross somite segment (blue hash) or missing EGFP signal (magenta asterisk). P Quantification of aberrant motor axon as indicated in O. n = 30 axons for each group

Fig. 7

Functional characterization of human OGDHL variants in zebrafish model. A Experimental approach for characterizing functionality of human OGDHL variants. Human OGDHL variant mRNAs were mixed with synthetic zebrafish-specific ogdhl single-guide RNAs (sgRNAs) and Cas9 protein and microinjected into one-cell stage embryos, followed by phenotypic evaluation at 3 dpf. Blue line indicates head size measurement approach. Red line indicates eye size measurement. Green arrowhead indicates heart edema. B, C Head and eye size measurements were calculated as a percentage difference compared to the mean value of the uninjected embryos. Subsequently, the calculated measurements of uninjected, Cas9-injected, and ogdhl F₀ embryos were compared. Additionally, the ogdhl F₀ co-injected with either WT or mutated OGDHL mRNA were compared to the ogdhl F₀ knockout embryos. Each dot represents one animal and the mean value of each group is indicated at the bottom of the respective bar in the figure. n = the number of embryos. Error bars = mean ± SD. Statistical significance was calculated by Brown–Forsthye and Welch’s ANOVA with Dunnett’s T3 multiple comparisons test: not significant (ns) p ≥ 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Significances for all rescue experiments are in comparison to ogdhl F₀. D The presence of the heart edema phenotype was calculated as a percentage of total embryos of each group. E The schematic of OGDHL truncating variants and the corresponding rescue experiment results