Among the 590 recruited patients, 359 (60.85%) were found to have the disease-causing genetic variations associated with their clinical features. Of these, 10 cases (2.5%, patients #1–10) from unrelated families were identified to have mutations in the MAMLD1 gene. From the literature search, 26 cases (patients #11–36) with MAMLD1-associated 46, XY DSD were identified.
Clinical features
The clinical features of the 10 Chinese patients displaying MAMLD1 gene mutations are summarized in Table 1,and their ages at first visit were all under 3 years; only patient #5 had a family history of hypospadias. Based on the position of the urethral meatus, seven patients (patients #1–3, #6–8, and #10) had severe hypospadias, whereas one patient (patient#5) was classified as having a mild type of hypospadias. In addition to the genital malformation, patient #1 also showed macrocephaly, stunned forehead, and cupped ears.Patient #4 had normal male external genitalia at birth; however, he was admitted to our hospital because of premature secondary sexual characteristics, increased penis size, growing beard, and pubic hair at the age of 1 year. In addition,although the patient showed elevated LH, FSH, and T levels, which supported the precocious puberty, his T level later on decreased, and he displayed hypergonadotropic hypogonadism. According to the ultrasound results, none of the patients had the Müllerian structures, whereas the Wolffian and adrenal structures were normal.
The clinical features, hormone profiles, and molecular results regarding the 26 previously reported cases are listed in Table 2. Hypospadias was the salient phenotype (22/26), and fourpatients manifested with complete external female genitalia.
Hormone level measurements
Detailed endocrine data are shown in Table 1. Serum T,LH, and FSH levels were sufficiently high in patients #3 and #6, who were in the mini-puberty period. In addition, a good response of T to hCG stimulation was observed in patients #1–3, #6, and #8. The LH and FSH levels of patient #4 were elevated at 1 year of age but declined with age; at 2 years and 3 months, the basal LH and FSH levels were decreased compared with the normal range, although the peak values after GnRH stimulation were still significantly high, indicating premature reactivation of the hypothalamic GnRH pulse generator/pituitary gonadotropin-gonadal axis.The AMH and INHB levels in patients #1, #3, #6, #8, and #9 were all within the normal range.
Adrenal function was evaluated in patients #1–9.Patients #2, #3, #6, and #8 showed elevated plasma ACTH at the first visit; however, the ACTH levels were detected to be normal without any medication in the follow-up.
As for the 26 previously reported cases, all the measured baseline serum T levelswere within the normal range, except in patient #23, who manifested with undetectable T in the serum. Notably, patient #31 had hypospadias, short penis, and delayed puberty, although the basal T level was within the normal range for his age. Because the data for the basal LH and FSH levels were not available, we could not confirm hypogonadism by delayed puberty in case #31. Adrenal function was normal in all the nine cases whose data were available.
Molecular analysis of the MAMLD1 gene
In our case series, a total of nine mutations were identified, including six missense variants(p.S662A, p.A421P, p.P542S, p.P334S, p.T992I, and p.A927L) and three nonsense variants (p.A356X, p.G152X, and p.G124X); p.A356X was identified in two patients. All the variants were inherited from the mothers, and exon 3 was the most commonly affected region (6 out of 9 variations, 67%). According to the ACMG clinical practice guidelines, three variants (p.A356X, p.G152X, and p.G124X) in four patients were pathogenic,andvariant p.P334S was likely pathogenic, whereas the missense variants p.S662A, p.A421P, and p.T992I are of uncertain significance,andthe two missense variants p.P542S and p.A927Lare presumably benign.
In the previously reported 26 cases, there were 21MAMLD1 variants (Table 2). Of these, 15 variants were located on exon 3, and 11 variants were considered pathogenic or likely pathogenic according to the ACMG, whereas the remaining 10 variants were considered as polymorphisms.
In silico analysis
The results of the molecular modelling of four selected variants (p.P334S, p.Pro542Ser, p.S662A, and p.A927L) are shown in Figure 1 and are discussed in detail below.In the case of both the WT and mutant p.P334S, there is a hydrogen bond between Pro/Ser334 and Leu362 residues in the middle of the two short-chain α-helix regions; however, the polar atoms forming the hydrogen bond are changed, leading to a small perturbance in the structural balance of the protein (Figure 1B). Compared with the WT, which has two hydrogen bonds formed between Pro542 residue and Ser543 and Leu544 residues, the mutant p.P542S forms a new hydrogen bond between Ser542 and Thr538 residues. Additionally, a part of the nearby coil region is transformed into the short-chain α-helix structure, with the polarity of the protein surface remarkably increased(Figure 1C). The mutant p.S662A has a hydrogen bond between Arg662 and Ser658 residues in addition to the one between Ser662 and Ser658 residues in the WT. This mutantalso has slight changes in the polarity of the protein surface. In addition, the two short-chain β-fold regions in the vicinity are converted into a coil region, with the polarity significantly increased, based on the visualization of the predicted molecular surface (Figure 1D). In mutant p.A927L, three hydrogen bonds that exist between Arg927 and Ala780 residues, and Arg927 and Asp924 residues in the WT are disrupted. Additionally, the polarity of the protein surface is greatly reduced in this mutant (Figure 1E). Based on the molecular modelling of the three selected variants (p.P542S, p.S662A, and p.A927L), it is speculated that the mutations lead to a drastic change in the interaction force of the amino acid chain and flexibility of the spatial structure, and such changes may have an effect on the function of the protein.