A Mouse Model with a Frameshift Mutation in the Nuclear Factor I/X ( NFIX ) Gene Has Phenotypic Features of Marshall‐Smith Syndrome

Abstract

The nuclear factor I/X (NFIX) gene encodes a ubiquitously expressed transcription factor whose mutations lead to two allelic disorders characterized by developmental, skeletal, and neural abnormalities, namely, Malan syndrome (MAL) and Marshall–Smith syndrome (MSS). NFIX mutations associated with MAL mainly cluster in exon 2 and are cleared by nonsense‐mediated decay (NMD) leading to NFIX haploinsufficiency, whereas NFIX mutations associated with MSS are clustered in exons 6–10 and escape NMD and result in the production of dominant‐negative mutant NFIX proteins. Thus, different NFIX mutations have distinct consequences on NFIX expression. To elucidate the in vivo effects of MSS‐associated NFIX exon 7 mutations, we used CRISPR‐Cas9 to generate mouse models with exon 7 deletions that comprised: a frameshift deletion of two nucleotides (Nfix Del2); in‐frame deletion of 24 nucleotides (Nfix Del24); and deletion of 140 nucleotides (Nfix Del140). Nfix ^+/Del2 , Nfix ^+/Del24 , Nfix ^+/Del140 , Nfix ^Del24/Del24 , and Nfix ^{Del140/Del140} mice were viable, normal, and fertile, with no skeletal abnormalities, but Nfix ^Del2/Del2 mice had significantly reduced viability (p < 0.002) and died at 2–3 weeks of age. Nfix Del2 was not cleared by NMD, and Nfix^Del2/Del2 mice, when compared to Nfix ^+/+ and Nfix ^+/Del2 mice, had: growth retardation; short stature with kyphosis; reduced skull length; marked porosity of the vertebrae with decreased vertebral and femoral bone mineral content; and reduced caudal vertebrae height and femur length. Plasma biochemistry analysis revealed Nfix ^Del2/Del2 mice to have increased total alkaline phosphatase activity but decreased C‐terminal telopeptide and procollagen‐type‐1‐N‐terminal propeptide concentrations compared to Nfix ^+/+ and Nfix ^+/Del2 mice. Nfix ^Del2/Del2 mice were also found to have enlarged cerebral cortices and ventricular areas but smaller dentate gyrus compared to Nfix ^+/+ mice. Thus, Nfix ^Del2/Del2 mice provide a model for studying the in vivo effects of NFIX mutants that escape NMD and result in developmental abnormalities of the skeletal and neural tissues that are associated with MSS. © 2023 The Authors. JBMR Plus published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research.

A mouse model for Marshall–Smith syndrome.

Introduction

The nuclear factor I/X (NFIX) gene (MIM #164005),^{( 1} , ² , ^{3 )} located on chromosome 19p13.2,^{( 4 )} consists of 11 exons (Fig. S1) that encode 14 transcripts, of which 11 are protein coding. NFIX encodes a ubiquitously expressed transcription factor that forms part of the NFI gene family, which in mammals consists of NFIA, NFIB, NFIC, and NFIX. These transcription factors share a highly conserved N‐terminal DNA binding and dimerization domain, which bind as homo‐ or heterodimers to the consensus palindromic sequence 5′‐TTGGC(N5)GCCAA‐3′ present in the promoter regions of viral and cellular genes,^{( 5 )} and a variable C‐terminal transactivation/repression domain, which can potentially provide a range of preferential interactions with other proteins to either activate or suppress transcription.^{( 6} , ⁷ , ⁸ , ^{9 )} NFI transcription factors play important roles in the regulation of stem cell differentiation, quiescence, and differentiation during the development of organs that include lung, kidney, liver, blood, heart, skeleton, and the nervous system.^{( 6} , ^{10 )}

Heterozygous mutations in the NFIX gene can lead to two rare allelic disorders, Malan syndrome (MAL; MIM #614753) and Marshall‐Smith syndrome (MSS; MIM #602535).^{( 1} , ^{9 )} MAL is an overgrowth disorder, characterized by a slender habitus, long hands and advanced bone age, moderate to severe intellectual disability, unusual facial phenotype consisting of a long, triangular face with a prominent forehead, everted lower lip and prominent chin, and behavioral problems, which are usually dominated by anxieties and, less frequently, by aggression^{( 9} , ¹⁰ , ¹¹ , ^{12 )} (Table S1). The missense, nonsense, and frameshift NFIX variants reported in MAL patients predominantly affect exon 2 (Fig. S1), which encodes the highly conserved N‐terminal DNA binding and dimerization domain of the NFIX protein.^{( 1} , ³ , ¹³ , ^{14 )} Entire gene deletions and NFIX mutations observed in MAL patients are predicted to be cleared by nonsense‐mediated mRNA decay (NMD) and lead to NFIX haploinsufficiency.^{( 1} , ² , ³ , ¹¹ , ¹³ , ^{14 )} MSS is characterized by short stature with skeletal abnormalities that may include kyphoscoliosis, abnormal bone maturation, craniofacial defects, and osteopenia and be associated with delays in motor and neural development that lead to moderate to severe mental retardation, limited or absent speech, and postnatal failure to thrive.^{( 15} , ^{16 )} In addition, MSS patients may have distinctive facial features that include a high forehead, proptosis, blue sclera, anteverted nares, small and retracted mandible, gingival hypertrophy, and hypertrichosis (Table S1). MSS patients may also suffer from respiratory difficulties with upper‐airway obstruction and apneas. The de novo frameshift NFIX mutations reported in MSS patients are all clustered in exons 6–10 of the NFIX gene, which encode the variable C‐terminal transactivation/repression domain (Fig. S1). The different mutations result in the production of aberrant transcripts that escape NMD and lead to the production of dysfunctional truncated NFIX proteins, which are predicted to behave in a dominant‐negative manner.^{( 1} , ² , ^{3 )} Thus, mutations that affect different regions of the NFIX gene have distinct consequences on the resulting transcripts and encoded proteins.

To date, only the in vivo consequences of Nfix exon 2 deletion, which encodes the conserved N‐terminal DNA binding and dimerization domain, have been studied in mouse models.^{( 17} , ^{18 )} In one study wherein Nfix exon 2 was replaced with an in‐frame lacZ reporter gene, Nfix ^+/lacZ mice were reported to have normal survival, but reduced body weight, while Nfix ^lacZ/lacZ mice developed skeletal abnormalities due to defects in ossification that resulted in kyphosis and neurological abnormalities such as partial agenesis of the corpus callosum that was associated with hydrocephalus.^{( 17 )} In other studies wherein an Nfix null allele was initially generated via Cre‐recombinase‐mediated excision of Nfix exon 2, the heterozygous Nfix ^+/− mice also had normal survival but with neurological abnormalities,^{( 19 )} and the homozygous Nfix ^−/− mice had neurological defects that included dysgenesis of the corpus callosum but did not have skeletal abnormalities.^{( 18 )} Moreover, Nfix ^−/− mice are reported to have severe delay in intermediate progenitor cells during forebrain development^{( 20 )} and smaller muscle fibers with impairment of muscle regeneration despite the lack of skeletal defects^{( 21 )} (Table S1). These Nfix‐deficient mice with targeted deletions of exon 2 are reported to be representative of MAL. Therefore, to establish potential representative models for MSS, we generated Nfix mouse models with frameshift mutations in exon 7, which is the most commonly mutated exon in MSS patients.^{( 1} , ² , ^{3 )}

Materials and Methods

Study approval

All animal studies were approved by the Medical Research Council Harwell Institute Ethical Review Committee and were licensed under the Animal (Scientific Procedures) Act 1986, issued by the UK Government Home Office Department (PPL30/2433 and PPL30/3271).

Generation of mutant mice and genotyping analysis

Mice were generated using the CRISPR/Cas9 system,^{( 22 )} and genotyping was performed by PCR amplification using genomic DNA and confirmed by RT‐PCR using total RNA extracted, as described in Data S2 Materials and Methods.

Cell lines and in vitro expression assays

Murine embryonic fibroblast (MEF) cells and monkey kidney fibroblast (COS‐7) cells that were used for RNA sequencing analysis or transiently transfected with wild‐type (WT) or mutant murine Nfix cDNA expression constructs and luciferase reporter constructs were utilized for qRT‐PCR, Western blot, and immunofluorescence analyses, as detailed in Data S2 Materials and Methods.

Phenotype analysis

Blood samples were collected and used for plasma biochemical analysis,^{( 22 )} and skeletons and tissues of WT and mutant mice were prepared and used for imaging and histological analyses, as detailed in Data S2 Materials and Methods.^{( 23 )}

Statistical analysis

Data are expressed as mean and standard deviation (SD) or standard errors of mean (SEM). All analyses were performed using Prism (GraphPad), and a value of p < 0.05 was considered significant for all analyses as described in Data S2 Materials and Methods.

Results

Establishment of mutant Nfix mouse models with targeted mutations of exon 7

To derive mouse models with frameshift mutations that affect the variable C‐terminal transactivation or repression domain of the NFIX gene, the CRISPR‐Cas9 system was used to target exon 7 of the murine Nfix gene. Following injection of Cas9 mRNA and Nfix guide RNA into C57BL/6J embryos, founder mice were generated from which three mutant Nfix lines comprising deletions of two nucleotides (Del2), 24 nucleotides (Del24), and 140 nucleotides (Del140) were established. More specifically, Nfix Del2 consists of a frameshift two‐nucleotide deletion from position +49,580 to +49,581 relative to the translation start site (TSS), Nfix Del24 contains an in‐frame 24‐nucleotide deletion (from position +49,561 to +49,584 relative to the TSS), and Nfix Del140 contains a 140‐nucleotide deletion (from position +49,577 to +49,716 relative to the TSS) and comprised 53 nucleotides from exon 7 and 87 nucleotides of intron 7 (Fig. S2A ).

Heterozygous Nfix mice (Nfix ^+/Del2 , Nfix ^+/Del24 , and Nfix ^+/Del140 ) were viable and intercrossed within each line to generate WT (Nfix ^+/+ ), heterozygous (Nfix ^+/Del2 , Nfix ^+/Del24 , and Nfix ^+/Del140 ) and homozygous (Nfix ^Del2/Del2 , Nfix ^Del24/Del24 , and Nfix ^{Del140/Del140} ) mice. Genotypes were confirmed and validated by PCR amplification of exon 7, Sanger DNA sequencing, and, in the case of Nfix Del2 mice, using NlaIII restriction endonuclease digestion analysis (Fig. S2 B–D). Nfix ^+/Del2 , Nfix ^+/Del24 , Nfix ^Del24/Del24 , Nfix ^+/Del140 , and Nfix ^{Del140/Del140} mice were viable, normal, and fertile, but Nfix ^Del2/Del2 mice were subviable by 21 days post term (P21) due to early death around 2–3 weeks of age (p = 0.002; Table 1). Thus, deviation from the normal Mendelian ratio (1:2:1) was not observed in the Nfix ^{Del140/Del140} and Nfix ^Del24/Del24 mice at E18.5, P21, or 12 weeks and in the Nfix ^Del2/Del2 mice at E18.5. However, the Nfix ^Del2/Del2 mice deviated significantly (p = 0.002) from the expected Mendelian ratio at P21 due to early death at 2–3 weeks, indicative of reduced viability of the Nfix ^Del2/Del2 mice. Moreover, the numbers of Nfix ^+/+ , Nfix ^+/Del2 , and Nfix ^Del2/Del2 mouse embryos at day 18.5 (E18.5) did not deviate from the expected 1:2:1 Mendelian ratio (Table 1), thereby suggesting that the life‐limiting mutational effects in Nfix ^Del2/Del2 mice are manifested between E18.5 and P21.

Table 1.

Mendelian Ratios and Binomial Distribution Analysis of Wild‐Type, Nfix Del2, Del24, and Del140 Mice from Intercrosses at E18.5, P21, and 12 Weeks of Age

Stage and mutation	Genotype			Total	p a
	Nfix +/+	Nfix +/−	Nfix −/−
E18.5 b
Del2	19 (13.5) c	26 (27)	9 (13.5)	54	0.101
Del24	14 (16)	31 (32)	19 (16)	64	0.844
Del140	6 (13.25)	27 (26.5)	20 (13.25)	53	0.987
P21 d
Del2	74 (71.5)	162 (143)	50 (71.5)	286	0.002**
Del24	86 (68.25)	122 (136.5)	65 (68.25)	273	0.354
Del140	43 (57.25)	122 (114.5)	64 (57.25)	229	0.865
12 weeks
Del24	86 (67.75)	120 (135.5)	65 (67.75)	271	0.380
Del140	43 (56.75)	120 (113.5)	64 (56.75)	227	0.882

^a p: probability observed number of homozygotes is significantly different from the expected number of homozygotes, which is 25% of the total number of mice obtained from heterozygotes intercrosses, derived by binomial distribution analysis; **p < 0.01.

^b E18.5: embryonic day E18.5.

^c Observed numbers are shown with expected numbers shown in parentheses.

^d P21: postnatal day 21.

Effects of three exon 7 mutations (Nfix Del2, Nfix Del24, and Nfix Del140) on Nfix transcription and translation

The differences in viability between the homozygous Nfix ^Del2/Del2 mutant mice and the homozygous Nfix ^Del24/Del24 and Nfix ^{Del140/Del140} mutant mice suggested that the Nfix allelic variants may have different effects on the expression of this transcription factor. We therefore investigated the effects of these Nfix allele variants on the transcription and translation of Nfix. Murine Nfix contains 11 exons that encode eight transcripts (five of which are protein coding), due to alternative splicing of exons 7 and 9 and the use of different transcription initiation sites (ENSMUSG00000001911.16). Thus, alternative splicing may produce WT Nfix transcripts that retain exon 7 (Nfix long isoform) or shorter conserved isoforms that lack exon 7 (Nfix ΔEx7) (Fig. S3A ). To study the effects of Nfix Del2, Nfix Del24, and Nfix Del140 deletions on splicing of exon 7, RT‐PCR using primers located in exons 6 and 8 and Sanger sequencing were performed on total RNA obtained from MEFs derived from Nfix ^+/+ , Nfix ^+/Del2 , Nfix ^Del2/Del2 , Nfix ^+/Del24 , Nfix ^Del24/Del24 , Nfix ^+/Del140 , and Nfix ^{Del140/Del140} mice. This revealed that Nfix ^+/+ MEFs had the Nfix WT long (317 bp) and short WT ΔEx7 (194 bp) isoforms (Fig. S3 B–D), but the Nfix ^Del2/Del2 and Nfix ^Del24/Del24 MEFs had mutant Nfix long isoforms of 315 and 293 bp, respectively, and the Nfix WT short isoform (ΔEx7 of 194 bp) (Fig. S3 B,C). Nfix ^+/Del2 and Nfix ^+/Del24 MEFs were confirmed to express a WT Nfix long isoform, a mutant Nfix long isoform, and the Nfix WT ΔEx7 short isoform (Fig. S3 B,C). In contrast, Nfix ^{Del140/Del140} MEFs had only the Nfix WT ΔEx7 short isoform, thereby suggesting that the 140‐nucleotide deletion, which comprised 53 nucleotides of the 3′ end of exon 7 along with 87 nucleotides from intron 7 that included the donor splice site, led to exon 7 skipping. Sanger DNA sequence analysis confirmed that the sequence of this Nfix short isoform from the Nfix ^{Del140/Del140} MEFs matched the consensus murine sequence of the Nfix WT ΔEx7 short isoforms (Fig. S3 D). Therefore, the 140‐nucleotide deletion in the Nfix Del140 MEFs led to skipping of exon 7 and alternative splicing of exon 6 to exon 8 due to loss of a donor splice site, resulting in a frameshift and the introduction of a stop codon after 81 amino acids, which corresponded to the WT short NFIX isoforms. However, the two‐nucleotide deletion in Nfix Del2 MEFs resulted in a frameshift and the introduction of a premature stop codon after 65 amino acids, and the 24‐nucleotide in‐frame deletion in Nfix Del24 MEFs predicted the loss of eight amino acids (QGSSPRMA).

To further investigate the effects of the Nfix Del2 and Nfix Del24 mutations on Nfix transcription, translation, and cellular localization, in vitro expression assays in COS‐7 cells transiently transfected with N‐terminal‐FLAG‐tagged WT or mutant (Del2 or Del24) Nfix cDNA constructs that retain exon 7 were undertaken. Analysis by qRT‐PCR showed that there was no significant difference in the amount of Nfix Del2 or Nfix Del24 expression compared to Nfix WT, suggesting that these mutations affecting the C‐terminal part of the Nfix transcripts were not cleared by NMD mechanisms (Fig. 1A ). Furthermore, Western blot analysis demonstrated the production of Nfix Del2 and Nfix Del24 smaller NFIX mutant proteins (<55 kDa), as expected, compared to WT (55 kDa) (Fig. 1B ), thereby confirming that mutations in the C‐terminal part of the Nfix gene produced truncated NFIX proteins. In addition, the expression of the NFIX Del2 protein was significantly decreased (p < 0.05), whereas that of the NFIX Del24 protein was significantly increased (p < 0.01) compared to NFIX WT, thereby revealing differences in the stabilities and likely degradations of the mutant proteins (Fig. 1C ). Immunofluorescence analysis showed that the cellular localization of NFIX Del2 and NFIX Del24 proteins was similar to the predominantly nuclear localization of NFIX WT (Fig. 1D ). The Nfix Del140 mutation, which causes skipping of exon 7 to produce the Nfix WT ΔEx7 isoform, was not investigated in vitro.

Fig. 1.

In vitro expression assays using wild‐type (WT) and mutant (Del2 and Del24) N‐terminal‐FLAG‐tagged Nfix cDNA constructs in COS‐7 cells. COS‐7 cells were transiently transfected with N‐terminal‐FLAG‐tagged WT Nfix cDNA constructs (Nfix WT), mutant Nfix cDNA constructs (Nfix Del2 and Nfix Del24), or an empty expression vector (pCMVTag2C). Untransfected (UT) COS‐7 cells were used as controls. (A) Quantitative real‐time PCR (qRT‐PCR) analysis showed that there was no significant difference in the amount of Nfix Del2 or Nfix Del24 expression compared to Nfix WT. Gapdh and Tbp1 were used as the housekeeping genes against which Nfix expression was normalized. (B) Western blot analysis using anti‐FLAG antibodies revealed Nfix Del2 and Nfix Del24 produced smaller NFIX proteins (< 55 kDa) compared to WT (55 kDa). Antibodies against calnexin (CANX) (75 kDa) were used as loading control. (C) Relative N‐terminal‐FLAG‐tagged NFIX expression, normalized to CANX expression, was quantified by densitometry analysis. (D) Cellular localization of transiently transfected WT and mutant NFIX following immunofluorescence analysis using an anti‐FLAG antibody showed that the nuclear localization of NFIX Del2 and NFIX Del24 was comparable to NFIX WT. (E) In vitro dual luciferase reporter assays, in which the luciferase reporter gene is under the control of the glial fibrillary acidic protein (GFAP) promoter containing three NFIX binding sites, cotransfected with WT or mutant Nfix cDNA constructs. The luciferase construct and varying concentrations of Nfix WT and (F) empty pCMVTag2C vector or (G) Nfix Del 24 or (H) Nfix Del2 cDNA constructs were used to cotransfect COS‐7 cells. The Nfix WT cDNA construct and the Nfix Del 24 mutant construct did not affect the transactivation activity of the NFIX protein at the GFAP locus, while the Nfix Del 2 mutant construct increased NFIX transactivation activity at the GFAP locus in a threshold‐dependent manner. Scale bar = 20 μm. Data are represented as mean ± SEM, n = 4–6, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns = not significant, ^{^} p < 0.01 compared to 500 ng Nfix WT, ^$ p < 0.01 compared to 400 ng Nfix WT, ^& p < 0.01 compared to 300 ng Nfix WT.

To further assess the effects of the Nfix Del2 and Nfix Del24 mutations on NFIX transcription factor function, given that NFIX is reported to activate GFAP expression,^{( 24 )} reporter constructs comprising the luciferase reporter gene downstream of the GFAP promoter were transiently cotransfected with WT or mutant Nfix cDNA constructs into COS‐7 cells. WT NFIX activated the GFAP promoter and caused an approximately ninefold increase (n = 4, p < 0.0001, Fig. 1E ) in luciferase activity in cells with the GFAP promoter cloned in the forward orientation compared to cells with the GFAP promoter cloned in the reverse orientation. Luciferase reporter activity was unaffected by the Nfix Del24 mutation (Fig. 1E,G ) compared to WT NFIX (Fig. 1E,F ). In contrast, the Nfix Del2 mutation caused a significant increase (n = 6, p < 0.05, Fig. 1E ) in luciferase activity, in a threshold‐dependent manner (Fig. 1H ) compared to WT NFIX, suggesting that the Nfix Del2 mutation had aberrant NFIX transactivation activity at the GFAP locus. Overall, these findings suggest that different frameshift mutations or in‐frame deletions affecting the C‐terminal part of the Nfix gene have distinct consequences on the activity of the resulting mutant NFIX proteins. Therefore, the phenotypes of the three mouse models harboring the Nfix allelic variants‐ Del2, Del24, and Del140—were further characterized for features of MSS.

Phenotypic characterization of Nfix Del2, Del24, and Del140 mice

Heterozygous and homozygous Nfix mutant and WT littermates were characterized for features of MSS that included abnormalities of growth, skeleton, central nervous system (CNS), viscera, and plasma biochemistry.

Analysis of growth, length, and body mass

Nfix ^Del2/Del2 mice were characterized by growth retardation and short stature when compared to Nfix ^+/+ , Nfix ^+/Del2 , Nfix ^+/Del24 , Nfix ^Del24/Del24 , Nfix ^+/Del140 , and Nfix ^{Del140/Del140} mice. Thus, at P1, there was no significant difference in the weights of Nfix ^+/+ (1.5 ± 0.1 g), Nfix ^+/Del2 (1.6 ± 0.0 g), and Nfix ^Del2/Del2 (1.5 ± 0.1 g) mice (Fig. S4), irrespective of sex, but by P14 the growth rate, measured as weight gained over time, of the Nfix ^Del2/Del2 mice was significantly reduced compared to the Nfix ^+/+ and Nfix ^+/Del2 mice (0.8‐fold, p < 0.05, Fig. S5 A). In addition, the Nfix ^Del2/Del2 mice were visibly smaller than Nfix ^+/+ and Nfix ^+/Del2 mice at 2 weeks (Fig. S5 A). In contrast, the growth rates of the Nfix ^+/Del24 and Nfix ^Del24/Del24 mice, and the Nfix ^+/Del140 and Nfix ^{Del140/Del140} mice were not significantly different from WT mice between 2 to 12 weeks of age, irrespective of sex (Fig. S5 B,C). Furthermore, visually, the Nfix ^+/Del24 , Nfix ^Del24/Del24 , Nfix ^+/Del140 , and Nfix ^{Del140/Del140} mice were indistinguishable from the Nfix ^+/+ mice at 12 weeks (Fig. S5 B,C).

The tail lengths, indicative of vertebral growth, of the Nfix ^Del2/Del2 mice were also significantly shorter than Nfix ^+/+ and Nfix ^+/Del2 mice (0.9‐fold, p < 0.001; Fig. 2A , Table S2 A). In addition, Echo‐MRI analysis revealed a significant decrease in weight (0.8‐fold, p < 0.0001, Fig. 2B ), lean (0.8‐fold, p < 0.0001, Fig. 2C ) and fat mass content (0.4‐fold, p < 0.0001, Fig. 2E ), and dual‐energy X‐ray absorptiometry (DXA) scan analysis revealed a significant decrease in total tissue mass (TTM, (i.e. sum of total lean and total fat mass), 0.8‐fold, p < 0.0001, Fig. 2G ) of Nfix ^Del2/Del2 mice compared to Nfix ^+/+ and Nfix ^+/Del2 mice at 2–3 weeks, in both males and females (Table S2 A), even when normalized to body weight (Fig. 2D,F,H ). In contrast, the tail lengths, weight, TTM, lean, and fat mass content were not significantly different between Nfix ^+/Del24 and Nfix ^Del24/Del24 mice, or Nfix ^+/Del140 and Nfix ^{Del140/Del140} mice and WT mice at 12 weeks (Table S2 B,C).

Fig. 2.

Length and body mass analyses (by echo‐MRI and dual‐energy X‐ray absorptiometry (DXA scan) of Nfix ^+/+ , Nfix ^+/Del2 , and Nfix ^Del2/Del2 mice at 2–3 weeks of age. (A) Tail length, (B) total weight, (C) lean mass, (D) lean mass normalized to body weight, (E) fat mass, (F) fat mass normalized to body weight, (G) TTM and (H) TTM normalized to body weight was significantly reduced in Nfix ^Del2/Del2 mice compared to Nfix ^+/+ and Nfix ^+/Del2 mice. The number of mice analyzed in each case is indicated in parentheses. Data are represented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, circles represent females, squares represent males.

Skeletal and bone turnover analyses

Micro–computed tomography (MCT) (Fig. 3A ), Alcian Blue, and Alizarin Red staining (Fig. S6) and radiological analyses (Fig. S7 A) of skeletons of 2–3 weeks old mice revealed that >30% of Nfix ^Del2/Del2 mice had kyphosis compared to <10% of Nfix ^+/+ and Nfix ^+/Del2 mice. Craniofacial measurements of the skulls of Nfix ^+/+ , Nfix ^+/Del2 , and Nfix ^Del2/Del2 mice revealed that there was a significant reduction in the skull length of Nfix ^Del2/Del2 mice compared to Nfix ^+/+ and Nfix ^+/Del2 mice (p < 0.05; Fig. S8 A,B). However, there were no significant differences in skull width, nasal bone length, and frontal bone length of Nfix ^Del2/Del2 mice, compared to Nfix ^+/+ and Nfix ^+/Del2 mice (Fig. S8 C–E), although parietal bone length of Nfix ^Del2/Del2 mice was significantly different from that of Nfix ^+/Del2 but not Nfix ^+/+ mice (p < 0.001; Fig. S8 F). In contrast, radiological analyses of Nfix ^+/+ , Nfix ^+/Del24 , and Nfix ^Del24/Del24 mice or Nfix ^+/+ , Nfix ^+/Del140 , and Nfix ^{Del140/Del140} mice revealed no skeletal abnormalities at 12 weeks (Fig. S7 B,C). MCT analysis of the lumbar and thoracic vertebrae also revealed Nfix ^Del2/Del2 mice to have marked porosity at 2–3 weeks (Fig. 3A ), and faxitron digital X‐ray microradiographic analysis of the caudal vertebrae and femora confirmed significantly reduced BMC of vertebrae and femora (p < 0.001; Figs 3B,C ; S9 A,B) and revealed decreases in vertebral height (0.8‐fold; p < 0.05; Fig. 3D ) and femoral length (0.9‐fold; p < 0.01; Fig. 3E ) in Nfix ^Del2/Del2 mice compared to Nfix ^+/+ and Nfix ^+/Del2 mice at 2 weeks. To explore further the basis of these reductions in vertebral and femoral BMC, osteoclast numbers were assessed by histological analysis of the tibia using tartrate‐resistant acid phosphatase (TRAP) staining (Fig. S9C ), which did not reveal significant differences in osteoclast numbers between Nfix ^Del2/Del2 mice compared to Nfix ^+/+ and Nfix ^+/Del2 mice at P21 (Fig. S9 D,E), although the lack of significant difference in the number of osteoclasts could be due to the low number of animals analyzed. To investigate whether the low BMC in Nfix ^Del2/Del2 mice may be a result of abnormal osteoclast activity instead, the plasma concentrations of C‐terminal telopeptide (CTX), procollagen‐type‐1‐N‐terminal propeptide (P1NP), and total alkaline phosphatase (ALP) activity, which are markers of bone resorption, bone formation, and bone mineralization, respectively, were therefore measured. Nfix ^Del2/Del2 mice, when compared to Nfix ^+/+ and Nfix ^+/Del2 mice at 2–3 weeks, had reduced CTX concentrations (0.5‐fold, p < 0.05, Fig. 3F ), due to abnormal osteoclast activity, decreased P1NP concentrations (0.8‐fold, p < 0.001, Fig. 3G ), indicating a reduction in bone formation, and raised plasma ALP activity, possibly implying abnormal bone mineralization (1.5‐fold, p < 0.0001, Fig. S3 H), thereby suggesting an overall abnormal bone turnover phenotype.

Fig. 3.

Skeletal abnormalities in Nfix ^Del2/Del2 mice at 2–3 weeks of age. (A) Representative MCT scans of skeletons (upper panels) and lumbar and thoracic vertebrae (lower panels) of female Nfix ^+/+ , Nfix ^+/Del2 , and Nfix ^Del2/Del2 mice. Nfix ^Del2/Del2 mice had kyphosis of the thoracolumbar spine (white arrow) and marked porosity (red arrows), which was not present in Nfix ^+/+ and Nfix ^+/Del2 mice. (B–E) Faxitron digital X‐ray microradiographic analysis showing cumulative frequency histogram of bone mineral content (BMC) in (B) vertebrae and (C) femora, and (D) vertebral height and (E) femoral length from Nfix ^+/+ , Nfix ^+/Del2 , and Nfix ^Del2/Del2 mice. Nfix ^Del2/Del2 mice had reductions in vertebral and femoral BMC and in length. Grayscale images were pseudocolored according to a 16‐color palette in which low mineral content is black and high mineral content is white (Fig. S8). Plasma biochemistry analysis of bone turnover markers in plasma of Nfix Del2 mice at 2–3 weeks of age revealed that Nfix ^Del2/Del2 mice had (F) decreased C‐terminal telopeptide (CTX) concentration, (G) reduced procollagen‐type‐1‐N‐terminal propeptide (P1NP) concentration, and (H) raised total alkaline phosphatase (ALP) activity compared to Nfix ^+/+ and Nfix ^+/Del2 mice. The number of mice analyzed is indicated in parentheses in each case. Data are represented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, circles represent females, squares represent males.

CNS abnormalities

Histological analysis of brains from 23 week‐old Nfix ^+/+ (Fig. 4A ) and Nfix ^Del2/Del2 (Fig. 4B ) mice revealed that Nfix ^Del2/Del2 mice had enlarged anterior cingulate (p < 0.05; Fig. 4C ), somatosensory (p < 0.05; Fig. 4D ) and retrosplenial (p < 0.01; Fig. 4E ) cortices. Moreover, both the total area of the ventricular zone (at the level of the corpus callosum) and the ventricular area normalized to the total brain section area were significantly larger in Nfix ^Del2/Del2 mice (p < 0.001; Fig. 4J,J ). In contrast, the hippocampal dentate gyrus was significantly smaller (p < 0.05; Figs 4K and S10A ) compared to Nfix ^+/+ mice. Both the superior blade (p < 0.01) and inferior blade areas (p < 0.05) of the dentate gyrus (Fig. 4M,M ) were significantly decreased in Nfix ^Del2/Del2 mice compared to Nfix ^+/+ mice. There was no difference in the size of the corpus callosum, motor cortex, and total brain area in Nfix ^Del2/Del2 mice compared to Nfix ^+/+ mice (Fig. 4H,H ).

Fig. 4.

Analyses of neurological abnormalities in Nfix Del2 mice at 2–3 weeks of age. Hematoxylin stained coronal brain sections of (A) Nfix ^+/+ and (B) Nfix ^Del2/Del2 mice shown at low magnification. Width measurements (black dashed lines shown in (A) and (B); scale bar = 1300 μm) were taken of the (C) cingulate cortex, (D) somatosensory cortex, (F) corpus callosum, and (G) motor cortex using rostral sections. Width measurements (black dashed lines shown in (A′) and (B′); scale bar = 600 μm) were taken of (E) retrosplenial cortex using caudal sections. There was no difference in total brain area (H) in Nfix ^Del2/Del2 mice compared to Nfix ^+/+ . (I) Ventricular area and (J) ventricular area as a proportion of total brain area were significantly larger in Nfix ^Del2/Del2 animals compared to Nfix ^+/+ . Total dentate gyrus area (K) was significantly decreased in Nfix ^Del2/Del2 animals compared to Nfix ^+/+ . Both the superior blade area (L) and inferior blade (M) area (dashed regions in (A') and (B′)) were significantly decreased in Nfix ^Del2/Del2 animals compared to Nfix ^+/+ . V, ventricle; AC, anterior cingulate cortex; M, motor cortex; RS, retrosplenial cortex; SS, somatosensory cortex. Immunofluorescence labeling of (N) Nfix ^+/+ and (O) Nfix ^Del2/Del2 coronal sections with Ki67 (green) and DCX (red). The boxed regions in (N) and (O) are shown at higher magnification in (N′) and (O′), respectively. DCX+ processes extended vertically in the Nfix ^+/+ (arrowheads in (N′)) mice, while DCX+ vertical processes were abnormal in the Nfix ^Del2/Del2 animals (arrowhead in (O′)). Scale bar = 100 μm in (N) and (O), 15 μm in (N′) and (O′). Immunofluorescence labeling of (P, P′) Nfix ^+/+ and (Q, Q') Nfix ^Del2/Del2 coronal sections with glial fibrillary acidic protein (GFAP) (red) and 4′, 6‐diamidino‐2‐phenylindole (DAPI) (white). Radially oriented GFAP+ fibers (arrowheads in (P) and(P′)) were present in the Nfix ^+/+ (Q, Q') mice. GFAP+ radial processes were malformed in Nfix ^Del2/Del2 animals. Scale bar = 15 μm in (P), (P′), (Q), and (Q'). There was no change in (R) the total number of proliferating cells (Ki67+) and (S) the total number of proliferating neuroblasts (Ki67+ DCX+) in Nfix ^Del2/Del2 mice compared to Nfix ^+/+ . (T) There were significantly fewer DCX‐labeled neuroblasts within the dentate gyrus of Nfix ^Del2/Del2 mice compared to Nfix ^+/+ . The number of mice analyzed is indicated in parentheses in each case. Data are represented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, circles represent females, squares represent males.

Mice homozygous for a null exon 2 Nfix allele were previously reported to exhibit similar phenotypes, including expanded ventricles and smaller dentate gyrus.^{( 18 )} These mice have also been shown to have reduced production of new neurons within the postnatal dentate gyrus.^{( 20 )} To determine whether the Nfix ^Del2/Del2 mice exhibited a similar phenotype, we performed co‐immunofluorescence analysis (Fig. 4N,O ) of neuroblasts with the neuroblast marker (doublecortin (DCX)) and proliferating cells (with the marker for proliferation Ki67), as well as with the neural stem cell marker (glial fibrillary acidic protein (GFAP); Fig. 4P,Q ). This analysis revealed no significant difference in either total proliferating cells (Ki67+; Fig. 4R ) or proliferating neuroblasts (Ki67+; DCX+; Fig. 4S ) in Nfix ^Del2/Del2 mice compared to Nfix ^+/+ mice. However, there were significantly fewer neuroblasts in the dentate gyrus of Nfix ^Del2/Del2 mice compared to Nfix ^+/+ (p < 0.01; Fig. 4T ), indicative of a smaller number of differentiating cells and neurons, similar to reports of a global deficit in neuroblasts observed in homozygous Nfix null mice deleted for exon 2.^{( 20 )} The neuroblast vertical processes extended vertically in the granule cell layer in the Nfix ^+/+ mice (Fig. 4N,N') but were abnormal in the neuroblasts of Nfix ^Del2/Del2 mice (Fig. 4O,O'). Similarly, radially oriented GFAP‐labeled fibers corresponding to postnatal neural stem cells (Fig. 4P,P') were present in the Nfix ^+/+ mice, but malformed in the Nfix ^Del2/Del2 mice (Fig. 4Q,Q'). In contrast, histological analysis of the brain of Nfix ^+/+ , Nfix ^+/Del24 , and Nfix ^Del24/Del24 or Nfix ^+/+ , Nfix ^+/Del140 , and Nfix ^{Del140/Del140} mice at 12 weeks revealed no abnormalities (Fig. S10 B,C).

Extraskeletal abnormalities

Plasma biochemical analysis revealed that Nfix ^Del2/Del2 mice, when compared to Nfix ^+/+ and Nfix ^+/Del2 mice, had raised plasma urea (1.4‐fold, p < 0.0001, Fig. S11 A) and raised total bilirubin (1.5‐fold, p < 0.0001, Fig. S11 B) at 2–3 weeks, in males and females (Table S2 A), consistent with abnormal kidney and liver function, respectively. However, there were no significant differences in the plasma concentrations of sodium, potassium, chloride, total and corrected calcium, inorganic phosphate, aspartate aminotransferase, alanine aminotransferase, albumin, creatinine, and creatine kinase in the Nfix ^Del2/Del2 mice compared to the Nfix ^+/+ and Nfix ^+/Del2 mice (Table S2 A). Moreover, histology of the kidneys and livers from Nfix ^Del2/Del2 mice revealed no abnormalities compared to Nfix ^+/+ and Nfix ^+/Del2 mice (Fig. S10 A). In addition, histological examination for liver inflammation (Fig. S11 C.E.F) and liver fibrosis (Fig. S11 D,G) revealed no hepatic abnormalities in the Nfix ^Del2/Del2 mice compared to the Nfix ^+/+ and Nfix ^+/Del2 mice. The plasma biochemistry of Nfix ^+/+ , Nfix ^+/Del24 , Nfix ^Del24/Del24 , Nfix ^+/Del140 , and Nfix ^{Del140/Del140} mice at 12 weeks was similar (Supplemental Table S2 B,C), suggesting normal kidney and liver function in these mice, and histological analysis of the liver, kidney, lung, and heart of these mice revealed no abnormalities (Fig. S10 B,C).

Effects of the three exon 7 mutations on the expression of Nfix paralogs in Nfix Del2, Nfix Del24, and Nfix Del140 MEFs

Since the variability in the phenotypes in the Nfix mouse models could be due to functional redundancy provided by the other members of the NFI gene family of transcription factors, we pursued RNA sequencing analysis to investigate differences in Nfia, Nfib, and Nfic ^{( 25} , ^{26 )} gene expression in the MEFs derived from the Nfix ^+/+ , Nfix ^+/Del2 , Nfix ^Del2/Del2 , Nfix ^+/Del24 , Nfix ^Del24/Del24 , Nfix ^+/Del140 , and Nfix ^{Del140/Del140} mice. RNA sequencing analysis identified that, compared to the mean of Nfix ^+/+ MEFs, Nfia transcripts were significantly altered (≥2‐fold‐change; p < 0.05; Table 2) in the Nfix ^+/Del140 (2.28‐fold‐change; p = 5.37 × 10⁻¹⁰; Table 2) and Nfix ^{Del140/Del140} MEFs (2.19‐fold‐change; p = 3.84 × 10⁻⁹; Table 2), while Nfib transcripts were significantly altered in the Nfix ^+/Del2 (6.27‐fold‐change; p = 1.11 × 10⁻⁷⁰; Table 2) and Nfix ^{Del140/Del140} MEFs (3.53‐fold‐change; p = 4.30 × 10⁻²⁸; Table 2). However, there was no significant change in Nfic expression between WT and mutant MEFs (Table 2). To validate the RNA sequencing results, we performed qRT‐PCR analysis using total RNA isolated from the MEFs of Nfix ^+/+ , Nfix ^+/Del2 , Nfix ^Del2/Del2 , Nfix ^+/Del24 , Nfix ^Del24/Del24 , Nfix ^+/Del140 , and Nfix ^{Del140/Del140} mice. qRT‐PCR analysis confirmed that, compared to the mean expression in the Nfix ^+/+ MEFs, Nfia expression was significantly increased only in the Nfix ^+/Del140 MEFs (3.18‐fold, p < 0.01, Fig. 5A ), while Nfib expression was significantly increased in the Nfix ^+/Del2 (12.5‐fold‐change, p < 0.0001, Fig. 5B ) and Nfix ^{Del140/Del140} MEFs (4.79‐fold‐change; p < 0.001, Fig. 5B ), consistent with the RNA sequencing data. Moreover, in agreement with the RNA sequencing results, there was no significant change in Nfic expression in the mutant MEFs compared to WT (Fig. 5C ). Therefore, our results suggest that in the Nfix ^+/Del2 , Nfix ^+/Del140 , and Nfix ^{Del140/NfixDel140} mice, Nfia and Nfib but not Nfic change their expression pattern in order to potentially compensate for their respective Nfix Del2 and Del140 frameshift mutations, which could possibly explain the lack of abnormalities in these mice, while Nfia, Nfib, and Nfic expression was unaltered in the unaffected Nfix ^+/Del24 and Nfix ^Del24/Del24 mice, thereby suggesting that the in‐frame Nfix Del2 mutations might potentially be tolerated and is probably not as damaging as a frameshift mutation. This is in contrast to Nfix ^Del2/Del2 mice where unchanged Nfia, Nfib, or Nfic gene expression suggests a lack of functional redundancy from the Nfix paralogs in the homozygous Nfix Del2 mice, which might possibly account for the more severe phenotype observed in the Nfix ^Del2/Del2 mice.

Table 2.

Fold‐Change in Nfia, Nfib, and Nfic Expression in Nfix ^+/Del2 , Nfix ^Del2/Del2 , Nfix ^+/Del24 , Nfix ^Del24/Del24 , Nfix ^+/Del140 , and Nfix ^{Del140/Del140} Murine Embryonic Fibroblasts (MEFs) Compared with Mean of Nfix ^+/+ MEFs as Determined by RNA Sequencing Analysis

Compared to Nfix +/+	Nfia	Nfib	Nfic
Nfix +/Del2	1.97 (2.52 × 10−7) a	6.27 (1.11 × 10−70)	1.11 (0.36)
Nfix Del2/Del2	0.97 (0.86)	1.25 (0.04)	0.97 (0.79)
Nfix +/Del24	1.10 (0.52)	0.57 (5.49 × 10−6)	0.93 (0.49)
Nfix Del24/Del24	0.96 (0.80)	0.58 (1.03 × 10−5)	0.94 (0.59)
Nfix +/Del140	2.28 (5.37 × 10−10)	1.84 (1.41 × 10−7)	0.88 (0.27)
Nfix Del140/Del140	2.19 (3.84 × 10−9)	3.53 (4.30 × 10−28)	0.97 (0.81)

^a Fold‐change with p‐values shown in parentheses.

Fig. 5.

Nfia, Nfib, and Nfic gene expression in Nfix Del2, Nfix Del24, and Nfix Del140 MEFs. qRT‐PCR analysis in Nfix ^+/+ , Nfix ^+/Del2 , Nfix ^Del2/Del2 , Nfix ^+/Del24 , Nfix ^Del24/Del24 , Nfix ^+/Del140 , and Nfix ^{Del140/Del140} MEFs of (A) Nfia, (B) Nfib, and (C) Nfic expression, with Gapdh and Canx used as the housekeeping genes against which candidate gene expression was normalized. Data are represented as mean ± SEM, n = 4–12, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 compared to mean of Nfix ^+/+ MEFs.

Discussion

Our study reports the phenotypic characterization of three CRISPR‐Cas9 generated mouse models with three allelic variants in Nfix exon 7, which is the most commonly mutated exon in MSS patients. These three allelic mutations, all of which affect the C‐terminal regions of NFIX, have different effects on the phenotypes and on the expression of Nfix transcripts and proteins. Thus, of the Nfix Del2, Nfix Del24, and Nfix Del140 mouse models, only the Nfix ^Del2/Del2 mice developed postnatal skeletal and cranial defects, brain abnormalities, and likely dysfunction of the kidney and liver, which might potentially account for their premature deaths by 2–3 weeks of age, whereas the Nfix ^+/Del2 , Nfix ^+/Del24 , Nfix ^Del24/Del24 , Nfix ^+/Del140 , and Nfix ^{Del140/Del140} mice were viable, normal, and fertile and survived to adulthood. These observations indicate that allelic variation, rather than potential off‐target effects of the CRISPR‐Cas9 system, could be responsible for the differences in phenotypes in the three Nfix mouse models, despite similar allelic mutations, the same environmental conditions, and identical genetic background, which is a common occurrence even on inbred backgrounds.^{( 27} , ^{28 )} Only the two‐nucleotide deletion in Nfix Del2 mice caused a frameshift and the introduction of a premature stop codon, which led to the production of intermediate levels of mutant NFIX Del2 protein with aberrant NFIX protein function that might potentially account for the more severe phenotype observed in the Nfix ^Del2/Del2 mice, while the 24‐nucleotide in‐frame deletion in Nfix Del24 mice caused the loss of eight amino acids, which could potentially be tolerated and is probably not as damaging as a frameshift mutation, whereas the 140‐nucleotide deletion in the Nfix Del140 mice, comprising 53 nucleotides from exon 7 and 87 nucleotides of intron 7 including the splice donor site, caused skipping of exon 7 and alternative splicing of exon 6 to exon 8 to produce WT (normal) Nfix isoforms. This suggests that different frameshift mutations or in‐frame deletions affecting the C‐terminal part of the Nfix gene have different consequences on the transcripts and activity of the resulting proteins, thereby accounting for the different phenotypes in the three mouse models.

Cell autonomous, monoallelic, and stochastic variation in gene expression, as well as functionally redundant paralogs, could also account for phenotypic variability.^{( 25} , ²⁶ , ²⁹ , ³⁰ , ^{31 )} Redundant paralogs that are ubiquitously expressed in a partially overlapping manner and that recognize similar motifs may provide backup for one another in case of mutation by changing their expression pattern and acquiring new regulatory capabilities in order to compensate for the mutation. For example, NFIA, NFIB, NFIC, and NFIX have the same conserved N‐terminal DNA binding and dimerization domain that enables all four related genes to recognize the same consensus sequence present in the promoter region of genes expressed in almost every organ, including the brain, lung, liver, intestine, and skeleton. Nfia ^−/− mice have CNS and kidney abnormalities^{( 32 )} and die perinatally,^{( 33 )} Nfib ^−/− mice have CNS and lung anomalies and die at birth,^{( 34 )} while Nfic ^−/− mice have only a mild phenotype involving abnormal tooth development of the incisors and molars.^{( 35 )} More recently, overlapping patterns of NFIA, NFIB, and NFIX expression have been reported in the brain.^{( 36 )} NFIA, NFIB, NFIC, and NFIX, which were previously shown to interact with each other as well as other cofactors, bind the same regulatory motif of promoters of genes, such as brain fatty acid binding protein (B‐FABP), GFAP, and inscuteable (INSC), and the NFIs or the ratio of the four NFIs have been shown to act either antagonistically or synergistically to regulate transcription in a promoter and context dependent manner.^{( 37} , ³⁸ , ^{39 )} Moreover, knockdown of one NFI member can affect the expression levels of other NFI members, suggesting cross‐talks and possible compensation within the NFI family.^{( 37 )} NFIX was also recently shown to act sequentially after NFIA and NFIB during gliogenesis within the spinal cord, and NFIB was reported to be able to activate Nfix expression in vitro, thereby suggesting autoregulatory mechanisms within the NFI gene family.^{( 38 )} In this study, we have shown that the combination of NFI family expression might potentially influence the phenotypes of the Nfix mouse models. Nfia and Nfib, but not Nfic, change their expression pattern in order to possibly compensate for their respective Nfix Del2 and Del140 frameshift mutations in the unaffected Nfix ^+/Del2 , Nfix ^+/Del140 , and Nfix ^{Del140/NfixDel140} mice, while Nfia, Nfib, and Nfic expression was unaltered in the unaffected Nfix ^+/Del24 and Nfix ^Del24/Del24 mice, thus suggesting that the in‐frame Nfix Del24 mutations might potentially be tolerated and are probably not as damaging as a frameshift mutation. Moreover, the lack of functional redundancy from the Nfix paralogs in the Nfix ^Del2/Del2 mice as well as the presence of intermediate levels of aberrant mutant NFIX Del2 protein might possibly account for the more severe phenotype observed in the Nfix ^Del2/Del2 mice.

Nfix ^Del2/Del2 mice represent a mouse model for MSS in which patients commonly have: reduced growth rate; short stature; craniofacial defects; osteopenia with increased fracture rate and kyphosis that normally worsens in puberty and adolescence and that is possibly aggravated by decreased bone density;^{( 15 )} and anxiety and intellectual disability due to nonspecific rain abnormalities.^{( 1} , ^{2 )} Thus, the Nfix ^Del2/Del2 mice had; short stature; reduced growth and TTM; kyphosis; shortened skull; marked porosity of the vertebrae; reduced BMC; shorter vertebrae height and femur length; reduced plasma CTX and P1NP concentrations but increased total ALP activity, indicative of abnormal bone function; and raised plasma urea and total bilirubin levels, suggestive of renal and hepatic dysfunction, which merits further investigation. Furthermore, Nfix ^Del2/Del2 mice had enlarged anterior cingulate, somatosensory and retrosplenial cortices, and ventricles but reduced dentate gyrus (Table S1). However, other features present in MSS patients, which include intellectual disability, airway obstruction leading to respiratory problem, umbilical hernia, cardiac anomalies, and abnormal bone maturation,^{( 1} , ² , ³ , ¹⁵ , ¹⁶ , ^{40 )} were not assessed in the Nfix ^Del2/Del2 mice in this study. Plasma biochemistry in MSS patients is reported to be usually normal, and our findings of elevated urea and bilirubin concentrations and ALP activity in association with reductions in plasma CTX and P1NP concentrations in the Nfix ^Del2/Del2 mice may represent important differences to MSS patients, or it may be that such abnormalities do occur in MSS patients but have hitherto not been found. This latter notion is a possibility as exemplified by our experience. Thus, following our identification of likely renal dysfunction in the Nfix ^Del2/Del2 mice, ultrasound scan investigations were undertaken in two MSS patients and revealed the occurrence of renal cysts in both patients and nephrocalcinosis in one (Hennekam–personal communication). Moreover, the reduction in plasma CTX concentrations in the Nfix ^Del2/Del2 mice may suggest abnormal osteoclast activity and function, which merits further investigation. Moreover, the paradoxical increased plasma ALP activity, which is a marker of bone turnover, in association with reduced plasma concentrations of CTX and P1NP, which are markers of bone resorption and bone formation respectively, in the Nfix ^Del2/Del2 mice suggests additional extraskeletal origin for the raised ALP activity such as the kidneys or intestine, but not liver as mice, in contrast to humans, express little or no ALP in the liver,^{( 41 )} and a search for additional renal or intestinal abnormalities in MSS may be warranted. Thus, it seems possible that MSS patients may have renal, intestinal, and hepatic dysfunction, and that there may be more similarities with the Nfix ^Del2/Del2 mice.

Our Nfix ^Del2/Del2 mice have similarities and differences when compared to two previous homozygous Nfix‐deficient mouse models that had targeted deletions of exon 2^{( 17} , ^{18 )} (Table S1). Thus, homozygous Nfix‐deficient mice (Nfix ^lacZ/lacZ ) were viable and had: growth retardation; an inability to fully open eyes; ataxic gait; feet‐clasping posture when lifted by their tail indicating neurological abnormalities; gastrointestinal defects; brain malformations consisting of hydrocephalus and partial agenesis of the corpus callosum; defects in endochondral ossification, reduction in trabecular bone formation and calcification; thinning of cranial bones; kyphotic deformation of the spine; and early postnatal death between 3 and 4 weeks of age.^{( 17 )} The other homozygous Nfix ^−/− mice showed; failure to thrive and grow when on a standard lab chow diet; delayed eye and ear opening; leg‐clasping phenotypes indicating neuroanatomical defects; increased brain weight due to expansion of the cortex and entire brain along the dorsal ventral axis; aberrant neocortex, cerebellum, hippocampus, and spinal cord formation; and an abnormal ventricular cell population due to excessive generation of Pax6‐expressing ventricular cells with hydrocephalus.^{( 18} , ²⁰ , ³⁸ , ³⁹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ , ^{46 )} Liver and kidney phenotypes were not assessed in these two previously reported Nfix‐deficient mouse models, although it is important to note that Nfix ^lacZ/lacZ mice had gastrointestinal defects.^{( 17 )} Importantly, the Nfix ^Del2/Del2 mice are not Nfix‐deficient but instead have aberrant Nfix transcripts that escape NMD and lead to the production of mutant truncated NFIX protein, which is representative of MSS. Interestingly, MSS patients are heterozygous for NFIX mutations, and this contrasts with Nfix ^+/Del2 mice, which are normal, while developmental, skeletal, cranial, neural, hepatic, and renal abnormalities are observed in Nfix ^Del2/Del2 mice, which could account for their reduced viability. However, phenotypic differences between organisms are not uncommon and can be attributed to allelic variation, modifier genes, genetic variations, genetic background, environmental conditions, and reduced sensitivity of assays, such as behavioral assays, in animal models versus in patients.^{( 29} , ³⁰ , ^{31 )} For example, the autosomal dominant disorder spondyloepimetaphyseal dysplasia, Missouri type (SEMD_MO) in humans, is due to a heterozygous matrix metalloproteinase 13 (MMP13) missense F56S mutation, whereas heterozygous Mmp13 ^+/− mice deleted for exons 3, 4, and 5 have normal growth plates, but the homozygous Mmp13 ^−/− mice have defects in growth plate cartilage and delayed endochondral ossification.^{( 47 )}

In summary, in this study we report three Nfix mouse models with three different targeted mutations in exon 7 of the Nfix gene, which are representative of the most frequent NFIX mutations observed in MSS patients. The three mouse models, although being on the same genetic background, have differing phenotypes and viability. While the Nfix ^Del2/Del2 mice have some similarities to previously reported Nfix deficient mouse models, they also have a number of other phenotypes that are consistent with MSS. Further studies of the Nfix ^Del2/Del2 mice will help better understand the role of NFIX mutations that result in dominant‐negative NFIX proteins and give rise to MSS, as well as provide useful resources for testing potential future treatments.

Author Contributions

Kreepa G. Kooblall: Conceptualization; data curation; formal analysis; investigation; methodology; validation; writing – original draft; writing – review and editing. Mark Stevenson: Supervision; writing – original draft; writing – review and editing. Michelle Stewart: Data curation; methodology; project administration; writing – review and editing. Lachlan Harris: Conceptualization; data curation; formal analysis; investigation; methodology; validation; writing – review and editing. Oressia Zalucki: Conceptualization; data curation; formal analysis; investigation; methodology; validation; writing – review and editing. Hannah Dewhurst: Data curation; formal analysis; investigation; methodology; validation; writing – review and editing. Natalie Butterfield: Data curation; formal analysis; investigation; methodology; validation; writing – review and editing. Houfu Leng: Data curation; formal analysis; investigation; methodology; validation; writing – review and editing. Tertius A. Hough: Data curation; investigation; methodology; validation; writing – review and editing. Da Ma: Data curation; formal analysis; investigation; methodology; software; validation; writing – review and editing. Bernard Siow: Conceptualization; data curation; formal analysis; investigation; methodology; software; validation; writing – review and editing. Paul Potter: Writing – review and editing. Roger D. Cox: Writing – review and editing. Stephen D.M. Brown: Writing – review and editing. Nicole Horwood: Supervision; writing – review and editing. Benjamin Wright: Data curation; formal analysis; investigation; methodology; software; validation; writing – review and editing. Helen Lockstone: Conceptualization; data curation; formal analysis; investigation; methodology; software; validation; writing – review and editing. David Buck: Software; supervision; writing – review and editing. Tonia Vincent: Supervision; writing – review and editing. Fadil M. Hannan: Conceptualization; writing – review and editing. J.H. Duncan Bassett: Conceptualization; funding acquisition; supervision; writing – review and editing. Graham R. Williams: Conceptualization; funding acquisition; supervision; writing – review and editing. Kate E. Lines: Supervision; writing – original draft; writing – review and editing. Michael Piper: Conceptualization; funding acquisition; supervision; writing – review and editing. Sara Wells: Resources; writing – review and editing. Lydia Teboul: Conceptualization; data curation; formal analysis; investigation; methodology; resources; supervision; validation; writing – review and editing. Raoul C. Hennekam: Conceptualization; funding acquisition; supervision; writing – review and editing. Rajesh V. Thakker: Conceptualization; funding acquisition; resources; supervision; writing – original draft; writing – review and editing.

Peer Review

The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1002/jbm4.10739.

Supporting information

Data S1. Supporting Information.

Data S2. Supporting Information.