In our study, we definitively diagnosed 12 patients with KS2 based on clinical symptoms and genetic testing and compared the differences in the incidence of various phenotypes between male and female patients with KS2 using a matched case-control study design. A typical dysmorphic feature is an emblematic clinical manifestation of KS. In previous studies, we found that almost all Chinese KS patients demonstrated the characteristic facial features of KS [9]. Makrythanasis et al. established a score list in which half of the scores were derived from facial features. Their study indicated that the score of the KMT2D variant group was higher than that of the non-KMT2D group [6], which was confirmed in our previously reported cases (KS1 4.72 vs KS2 3.33) [9]. In this cohort, Patient 10 (mother of Patient 2) showed eyebrows with lateral sparseness but lacked the other facial features, suggesting that female patients with KS2 may have a lower score than those of the non-KMT2D group. Whether the scoring system mentioned above is applicable to female patients with the KDM6A variant needs further exploration. It may not be appropriate to evaluate female patients with KS2 and male patients with KS2 when exploring the phenotype of KS2. Our statistical findings showed significant differences in the IQ scores, incidence of ID, and CHD between female and male patients with KS2. There is currently no significant difference for other phenotypes, but this may be related to the small number of cases being studied.
There is currently no effective treatment for ID, and this is one of the most confusing parts for families of patients with KS. Our study evaluated the IQ scores of patients with KS2 using the Wechsler Intelligence Test and obtained IQ scores of adult patients with KS2. Cristina et al. tested the IQ of KS1 patients and found an IQ score of 67 ± 24.9 (n = 6), suggesting that a small number of patients with KS have severe ID [10]. When analyzing patients with KS2, we found that the median IQ score of female patients in our case group reached 69 points, which was the highest score for mild ID, while male patients had an IQ score of only 41 points, almost reaching the diagnostic criteria for severe ID. In our study, there was a significant statistical difference in the incidence of moderate-to-severe ID (IQ < 55) between male and female patients. Additionally, two of the four male patients’ IQ scores were less than 40, and the rate of severe ID reached 50%. Although the ID score of KS described in the past indicates mostly mild-to-moderate ID [11], in our study, severe ID was a common finding in male patients with KS2.
The incidence of CHD in patients with KS varied in previous reports from approximately 28–80% [12–14]. Maria et al. summarized the pattern of CHDs in 28 KS patients with KS (27 KS1, 1 KS2); 19 of them had CHD with coarctation of the aorta, aortic valve, atrial septal defect, and ventricular septal defect [12]. In our study, we completed echocardiography of 12 patients with KS2, five of whom had CHD. Patients 1, 2, 3, and 9 had an atrial septal defect (ASD), and Patient 4 had pulmonary stenosis. Previously published studies have suggested that ASD is the most common type of CHD in patients with variants of KDM6A, which is consistent with our findings [3, 12, 15]. Additionally, we further found that the incidence of CHD in male patients with KS2 was significantly higher than that in female patients with KS2. Thus, we assume that CHD also manifests as a sex-specific difference.
In contrast to the above phenotype, it has been reported that the incidence of hypoglycemia was higher in patients with KS2 than in those with KS1. Kai et al. reported ten cases of hyperinsulinemic hypoglycemia among patients with KS, among which five cases were KS2 [16]. In our cohort, two patients were diagnosed with hyperinsulinemic hypoglycemia, among whom Patient 4 required maintenance with glucose transfusion treatment due to persistent hypoglycemia and severe feeding difficulties. When his intravenous glucose level was 2.0 mmol/l, his insulin level was up to 5.5 µIU/ml, and his C-peptide level was 0.46 nmol/l, which indicated hyperinsulinemic hypoglycemia. After obtaining informed consent from the patient’s parents, we administered diazoxide for treatment. During the treatment, the peripheral blood glucose level was maintained at about 4 mmol/L, and there were no adverse reactions. Although the etiology of hyperinsulinemic hypoglycemia in patients with KS is still unclear, its reported onset age in these patients is young, suggesting that KS may be one of the potential genetic factors of hypoglycemia in infants and young children [16–18]. The mechanism of pathogenic variation in the KDM6A gene leading to hyperinsulinemic hypoglycemia is unclear.
The KDM6A gene consists of 29 exons and contains a key domain, Jumonji C (JmjC), whose integrity is essential for maintaining histone demethylase activity. The loss of function of the KDM6A gene causes KS2, and we found nine different variants in 12 patients with KS2. Four of them were frameshift variants located in exons 1, 16, 17, and 19, and one was a nonsense variant located in exon 15. The three missense variants were located in exons 1, 6, and 19 (Fig. 2). Combining our data with previously published cases showed no obvious hot spots or regions [5, 19]. However, all null variant sites in our cases were located in front of the JmjC domain, causing premature termination of gene transcription.
Histone modification is an important part of epigenetics, and different histone modifications have different effects on downstream genes. Methylation of H3K4 plays a role in transcriptional activation, whereas methylation of H3K27 plays a role in transcriptional inhibition [20]. The methylation markers of H3K27 were labeled with Polycomb proteins and removed by KDM6 demethylases, of which KDM6A is an important member [21, 22]. KDM6A is located on the X chromosome, not in the pseudoautosomal region segment, but escapes X chromosome inactivation [23]. Although UTX homologue UTY is found on the male Y chromosome, it was thought to lack demethylation in vitro. As for KS patients with the same variant site, the difference in phenotype between female and male patients may result from the functional difference between UTX and UTY. However, some recent studies have shown that UTY still has an important physiological role [24]. Female homozygous mice could not survive, while hemizygous males could reflect the residual function of UTY [25]. Human KDM6A and UTY transcripts show up to 88% complementary DNA homology and 86% predicted amino acid homology [26]. Both carry the JmjC domain, which is extremely important for demethylation [27]. However, based on our case-control studies and previously reported cases, there are differences in the phenotypic severity of KS2 between the sexes. This phenomenon still requires more case support, and the mechanism still needs to be further explored.
Limitations
Although our study described a KS2 cohort with 12 patients, we also conducted a case-control study that revealed the phenotypic differences between male and female patients with KS. Nevertheless, some limitations should be noted. First, the number of cases in this study was limited. These 12 patients with KS2 were diagnosed within four years at a single center. Considering the rarity of KS2, the diagnosis rate was not low. In future studies, we will collect more cases and conduct multi-center studies to further analyze the phenotype and genotype of patients with KS2. Second, most of the children in our control group did not undergo genetic testing, because genetic testing is not a routine test item for normal children. Therefore, we could only exclude children with other causes of ID based on imaging or laboratory examinations. Third, we did not dynamically monitor the level of patients’ intellectual development and did not repeatedly evaluate patients’ IQ, and this is what we need to further improve in the future. By doing this, more convincing results can be obtained. In addition, for the assessment of CHD, the data we obtained may be biased, because whether the three adult patients had CHD in childhood could only be recalled without objective evidence. In future studies, a greater collection of children’s cases can solve this problem.