Targeting of Slc25a21 is associated with orofacial defects and otitis media due to disrupted expression of a neighbouring gene

Abstract

Homozygosity for Slc25a21(tm1a(KOMP)Wtsi) results in mice exhibiting orofacial abnormalities, alterations in carpal and rugae structures, hearing impairment and inflammation in the middle ear. In humans it has been hypothesised that the 2-oxoadipate mitochondrial carrier coded by SLC25A21 may be involved in the disease 2-oxoadipate acidaemia. Unexpectedly, no 2-oxoadipate acidaemia-like symptoms were observed in animals homozygous for Slc25a21(tm1a(KOMP)Wtsi) despite confirmation that this allele reduces Slc25a21 expression by 71.3%. To study the complete knockout, an allelic series was generated using the loxP and FRT sites typical of a Knockout Mouse Project allele. After removal of the critical exon and neomycin selection cassette, Slc25a21 knockout mice homozygous for the Slc25a21(tm1b(KOMP)Wtsi) and Slc25a21(tm1d(KOMP)Wtsi) alleles were phenotypically indistinguishable from wild-type. This led us to explore the genomic environment of Slc25a21 and to discover that expression of Pax9, located 3' of the target gene, was reduced in homozygous Slc25a21(tm1a(KOMP)Wtsi) mice. We hypothesize that the presence of the selection cassette is the cause of the down regulation of Pax9 observed. The phenotypes we observed in homozygous Slc25a21(tm1a(KOMP)Wtsi) mice were broadly consistent with a hypomorphic Pax9 allele with the exception of otitis media and hearing impairment which may be a novel consequence of Pax9 down regulation. We explore the ramifications associated with this particular targeted mutation and emphasise the need to interpret phenotypes taking into consideration all potential underlying genetic mechanisms.

Figure 1. Body weight and fat mass were reduced in homozygous Slc25a21^{tm1a(KOMP)Wtsi} mice.

Homozygous (red symbols) Slc25a21^{tm1a(KOMP)Wtsi} female (A,C) (n = 7) and male (B,D) (n = 7) mice showed decreased body weight and fat mass compared to local wild-type controls run on the same week (green symbols) (female, n = 22; male, n = 15). Mean (± SD) body weight is plotted against age (A,B). Fat mass is presented as a boxplot, with median, 25^th and 75^th percentile (box), and the lowest and highest data point within 1.5× inter quartile range (whiskers) shown (C,D). The median and 95% reference range (2.5–97.5^th percentiles, dotted lines) for all wild-type mice of the same genetic background and sex (n>750) are displayed on the pale green background.

Figure 2. Dental abnormalities observed in homozygous Slc25a21^{tm1a(KOMP)Wtsi} mice.

Dysmorphology analysis revealed underdeveloped, white incisors (A–B) in Slc25a21^{tm1a(KOMP)Wtsi} homozygous mice as compared to wild-type (C). Wholemount (D–F) and subsequent histological sections (G) demonstrated bacterial lacZ activity in cavities within lower jaw molars of Slc25a21^{tm1a(KOMP)Wtsi} homozygous (D,G) mice compared to heterozygous (E) and wild-type (F) mice. Focal fracturing of the surface dentine is indicated with a black arrow and necrosis of the pulp indicated with a red arrow (G). Micro-CT imaging demonstrated that the root aspect of lower jaw incisors was severely under developed (white arrow) and lacked enamel coating in homozygous Slc25a21^{tm1a(KOMP)Wtsi} mice (H–I) compared to wild-type (J).

Figure 3. Rugae and vomer abnormalities observed in homozygous Slc25a21^{tm1a(KOMP)Wtsi} mice.

Abnormalities in rugae (in particular number 3) were observed in homozygous Slc25a21^{tm1a(KOMP)Wtsi} mice (A–C) compared to wild-type (D), indicated with arrows (wt, n = 11; hom, n = 11). Micro-CT analysis revealed that the anterior part of the vomer bone was absent in Slc25a21^{tm1a(KOMP)Wtsi} homozygous mice (E–F) compared to wild-type (G) (wt, n = 2; hom, n = 12). This was confirmed by dissection of the palate of Slc25a21^{tm1a(KOMP)Wtsi} mice (H).

Figure 4. Auditory brainstem response threshold was elevated in Slc25a21^{tm1a(KOMP)Wtsi} mice.

Auditory brainstem response threshold was elevated across all frequencies in homozygous Slc25a21^{tm1a(KOMP)Wtsi} mice aged 1, 2, 3 and 6 months respectively. Mean ABR thresholds (± SD) are shown for wild-type (green symbols), heterozygote (blue symbols) and homozygote (red symbols) Slc25a21^{tm1a(KOMP)Wtsi} mice (A–D). The click-evoked ABR thresholds (mean ± SD) plotted as a function of age (months) demonstrated a progressive deterioration in Slc25a21^{tm1a(KOMP)Wtsi} mice (E); an observation clearly demonstrated by plotting the mean click-evoked ABR threshold elevation of mutants above wild-types as a function of age (months) (F).

Figure 5. Signs of otitis media in Slc25a21^{tm1a(KOMP)Wtsi} mice.

The tympanic membrane viewed from the external ear canal showed an opaque fluid present in Slc25a21^{tm1a(KOMP)Wtsi} homozygous mice (B) that was not present in wild-type controls (A). The opaque exudate was clearly visible on haematoxylin and eosin stained histology sections through the middle ear of Slc25a21^{tm1a(KOMP)Wtsi} homozygous mice (D) compared to wild-types (C). Examination at higher magnification revealed that the exudate containing inflammatory cells including foamy macrophages (solid arrow head), and neutrophils (short arrow) in the middle ear cavity (MEC), and a thickened mucosa (M, double headed arrow) in Slc25a21^{tm1a(KOMP)Wtsi} homozygous mice due to epithelial hyperplasia, oedema and congestion (F). Wild-type mice appeared normal (E).

Figure 6. Schematic of the Slc25a21^{tm1(KOMP)Wtsi} allelic series and the associated RNA expression levels.

Quantitative PCR on RNA from E13.5 embryo heads demonstrated that Slc25a21 RNA expression was significantly reduced in Slc25a21^{tm1a(KOMP)Wtsi}, Slc25a21^{tm1b(KOMP)Wtsi} and Slc25a21^{tm1d(KOMP)Wtsi} homozygotes (* p<0.05; *** p<0.001), but equivalent to wild-type levels in Slc25a21^{tm1c(KOMP)Wtsi} homozygotes (A). A schematic of the Slc25a21^{tm1(KOMP)Wtsi} allelic series is presented, consisting of: (B) Slc25a21^{tm1a(KOMP)Wtsi}, the knockout first reporter (lacZ) tagged insertion allele; (C) Slc25a21^{tm1b(KOMP)Wts i}, the lacZ tagged deletion allele generated by breeding with mice ubiquitously expressing Cre recombinase; (D) Slc25a21^{tm1c(KOMP)Wtsi}, the conditional allele generated by breeding mice carrying the Slc25a21^{tm1a(KOMP)Wtsi} allele with mice ubiquitously expressing Flp recombinase; and (E) Slc25a21^{tm1d(KOMP)Wtsi}, the deletion allele arising after breeding mice carrying the Slc25a21^{tm1c(KOMP)Wtsi} allele with mice ubiquitously expressing Cre recombinase (E).

Figure 7. Normal phenotype observed in Slc25a21^{tm1b(KOMP)Wtsi}, Slc25a21^{tm1c(KOMP)Wtsi} and Slc25a21^{tm1d(KOMP)Wtsi} homozygous mice.

Body weight and fat mass were normal in females homozygous (red symbols) for Slc25a21^{tm1b(KOMP)Wtsi} (A,D) (n = 7), Slc25a21^{tm1c(KOMP)Wtsi} (B,E) (n = 7), and Slc25a21^{tm1d(KOMP)Wtsi} mice (C,F) (body weight n = 7, fat mass n = 5) compared to local wild-type controls run on the same week (green symbols) (n≥9). Male homozygotes for each of these three alleles and two parameters were also phenotypically normal (data not shown). Mean (± SD) body weight is plotted against age (A–C). Fat mass is presented as a boxplot, with median, 25^th and 75^th percentile (box), and the lowest and highest data point within 1.5× inter quartile range (whiskers) shown (D–F). The median and 95% reference range (2.5–97.5^th percentiles, dotted lines) for all wild-type mice of the same genetic background and sex (n>80) are displayed on the pale green background. ABR thresholds were normal in homozygous Slc25a21^{tm1b(KOMP)Wtsi} (G), Slc25a21^{tm1c(KOMP)Wtsi} (H) and Slc25a21^{tm1d(KOMP)Wtsi} mice (I) (red symbols) compared to wild-type controls (green symbols).

Figure 8. RNA expression of Pax9 was reduced in Slc25a21^{tm1a(KOMP)Wtsi} mice.

Quantitative PCR on RNA from E13.5 embryo heads demonstrated that Pax9 RNA expression was significantly reduced in Slc25a21^{tm1a(KOMP)Wtsi} (*** p = 1.1E-06) homozygotes, but equivalent to wild-type levels in Slc25a21^{tm1b(KOMP)Wtsi}, Slc25a21^{tm1c(KOMP)Wtsi} and Slc25a21^{tm1d(KOMP)Wtsi} homozygotes.