Retinoic acid catabolizing enzyme CYP26C1 is a genetic modifier in SHOX deficiency

Abstract Mutations in the homeobox gene SHOX cause SHOX deficiency, a condition with clinical manifestations ranging from short stature without dysmorphic signs to severe mesomelic skeletal dysplasia. In rare cases, individuals with SHOX deficiency are asymptomatic. To elucidate the factors that modify disease severity/penetrance, we studied a three‐generation family with SHOX deficiency. The variant p.Phe508Cys of the retinoic acid catabolizing enzyme CYP26C1 co‐segregated with the SHOX variant p.Val161Ala in the affected individuals, while the SHOX mutant alone was present in asymptomatic individuals. Two further cases with SHOX deficiency and damaging CYP26C1 variants were identified in a cohort of 68 individuals with LWD. The identified CYP26C1 variants affected its catabolic activity, leading to an increased level of retinoic acid. High levels of retinoic acid significantly decrease SHOX expression in human primary chondrocytes and zebrafish embryos. Individual morpholino knockdown of either gene shortens the pectoral fins, whereas depletion of both genes leads to a more severe phenotype. Together, our findings describe CYP26C1 as the first genetic modifier for SHOX deficiency.


Appendix Figures Legends
-Treatment of human primary chondrocytes with RA concentrations within the physiological levels.
Relative expression of SHOX mRNA normalized to the reference genes SDHA and HPRT in human primary chondrocytes treated with ATRA 10, 30, and 50 nM for 6h (n = 5).
Data information: The box represents the interquartile range. The whiskers represent Min to Max. n.s., not significant, two-tailed Mann-Whitney non-parametric t test. Variants found in CYP26C1 in control normal height individuals are indicated in green. Cignal-RARE system luciferase assays to test the impact of CYP26C1 variants on its RA degradation activity were performed in U2OS cells treated with 250 nM all-trans retinoic acid (ATRA) for 24 hours (n = 4). Experiments were performed in triplicates. pIRES2-EGFP empty vector was used as control. The residue C459 represents the Iron binding residue (Q6VOL0, UniProtKB). TM, Transmembrane helix; P450, cytochrome p450 domain. RLU, Relative Light Units.
Data information: The box represents the interquartile range. The whiskers represent Min to Max. n.s., not significant, two-tailed Mann-Whitney non-parametric t test.  A. Scheme of cyp26c1 unspliced mRNA. Red bar, region targeted by cyp26c1 MO; arrows, position of the primers used to analyze MO efficacy; f, forward primer; r, reverse primer.
B. RT-PCR analysis of cyp26c1 MO efficacy. The 530 bp band represents the expected wild type product. Samples were obtained from 24 hpf embryos.
C. RT-PCR products were cloned in pSTBlue-1 and sequenced. cyp26c1 MO leads to the exclusion of exon 3. Black vertical segments enclose the out-spliced sequence.

Figure S5 -shox MO1 efficacy analysis.
A. Scheme of shox unspliced mRNA. Red bar, region targeted by shox MO1; arrows, position of the primers used to analyze MO efficacy; f, forward primer; r, reverse primer.
B. RT-PCR analysis of shox MO1 efficacy. The 430 bp band represents the expected wild type product. Samples were obtained from 24 hpf embryos.
C. RT-PCR products were cloned in pSTBlue-1 and sequenced. shox MO1 leads to a spliced product leading to frameshift, p.L142Rfs37*. Black vertical segments enclose the out-spliced sequence. C. RT-PCR products were cloned in pSTBlue-1 and sequenced. MO2 leads to a spliced product leading to frameshift, p.V59Pfs11*. Black vertical segments enclose the out-spliced sequence.

Figure S7 -Functional significance of the SHOX variants identified in families 1 and 2 in zebrafish embryos.
A-C. Wild type embryos injected with sense-capped RNA coding for human SHOX wild type or the variants identified in families 1 and 2, V161A (family 1) and L132V (family 2). Note that the patient in family 3 presented a SHOX deletion. (A) Dorsal views of the embryos at 55 hours post fertilization (hpf). Embryos injected with SHOX wild type RNA displayed longer pectoral fins when compared to control and SHOX variants V161A and L132V (n = 27-46 per condition). These results are in agreement with a preview study reporting that overexpression of Shox in chicken wings increases the length of skeletal elements (Tiecke et al, 2006). Dotted line, pectoral fins. (B, C) Dorsal view and magnification on the pectoral fins of col2a1 expression at 55 hpf. Arrow and dotted line indicate the pectoral fin. Embryos injected with SHOX wild type RNA displayed increased col2a1 expression. D. Pectoral fins area was measured by imageJ (n = 27-46 embryos per condition). Data information: Data are shown as means ± SD. SHOX wild type vs SHOX V161A, * p-value = 0.019. SHOX wild type vs control, ** p-value = 0.006. SHOX wild type vs SHOX L132V, ** p-value = 0.008. One-way ANOVA Bonferroni's multiple comparison test.

Figure S8 -Functional significance of the CYP26C1 variants identified in families 1-3 in zebrafish embryos.
A-C. Wild type embryos injected with sense-capped RNA coding for human CYP26C1 wild type or the variants identified in families 1-3, F508C (family 1), R378H (family 2), Q119P (family 3). (A) Dorsal views of the embryos at 55 hours post fertilization (hpf). Embryos injected with CYP26C1 wild type RNA displayed absent pectoral fins. Ectopic overexpression of CYP26C1 may lead to RA deficiency. Deficiency of RA at early stages of pectoral fin development has been previously shown to lead to similar phenotypes (Begemann et al, 2001). As shown in Fig 2, the CYP26C1 variants identified in families 1-3 disrupt CYP26C1 RA degradation activity and overexpression, therefore, does not alter RA levels within the embryos resulting in normal pectoral fins development (n = 24-32 per condition). Dotted line, pectoral fins. (B, C) Dorsal view and magnification on the pectoral fins of col2a1 expression at 55 hpf. Arrow and dotted line indicate the pectoral fin.