Variant frequencies
Among the 98 CH patient cohort, 24 were diagnosed with TD, 67 presented with normal-sized GIS, and 7 with G; relatedness between the participants was excluded. Using WES, the variant detection rate of all CH patients was 77.55% (76/98), while that of patients with GIS or G was 87.84% (65/74), and patients with TD was only 45.83% (11/24, P = 0.0001; Figure 1A). In total, we identified 149 variants spread across 9 genes, namely TSHR, TOP, TG, FOXE1, GLIS3, DUOXA2, DUOX2, PAX8, and SLC26A4. Among them, the DUOX2 gene had the largest number of detected variants, with a frequency of 74.50% (111/149; Figure 1B).
Characteristics of patients with DUOX2 variants
The DUOX2 gene variant was present in 56 of the 76 CH patients with detected variants. Four patients carried both DUOX2 and other genetic variants, and the remaining patients carried only DUOX2 variants, including 10 patients with single locus mutations and 42 patients with two or more point mutations. Subsequently, 111 variants with 50 types were identified on the DUOX2 gene; most were heterozygous, except for four homozygous DUOX2 mutations (i.e., p.K530* in two cases, and p.R1110Q, p.H678R, and p.R885L in one case each) detected in five patients. All variants were included in the databases or reported in previous studies; hotspot mutations were p.K530* (16.22%, 18/111) in exon 14 and p.R1110Q (17.12%, 19/111) in exon 25 (Table 1, Figure 2A).
In our long-term follow-up of 56 CH patients with DUOX2 variants, 17 GIS patients were diagnosed with TCH and 24 were diagnosed with PCH (21 patients with GIS, two patients with TD, and one patient with G). After analyzing the results, we observed an interesting phenomenon: in 38 GIS patients the proportion of patients with two or more variants that developed into PCH was 46.67% (14/30), while that of patients with single-point variants was 80% (4/5), and the remaining three patients with multiple gene variants all progressed to PCH. In particular, the incidence rate of PCH in GIS patients with the p.R1110Q variant (66.7%, 6/9) was relatively higher than that in patients with the p.K530* variant (50%, 4/8). Localization and distribution of variants in protein functional regions were then analyzed in patients with GIS in the biallelic group. The regions with the highest mutation frequency in TCH and PCH were the peroxidase domain and ferric oxidoreductase region, respectively. We conducted an in-depth analysis of the correlation between different protein domains and the clinical course of 30 GIS patients in the biallelic group. The results showed that when the variant was detected in the peroxidase domain, the probability of PCH occurrence did not differ from that in other regions (P > 0.999). However, when the variant was detected in the ferric oxidoreductase region, the probability of PCH occurrence was significantly higher than that in other regions (P = 0.011; Table 1, Table 2, Figure 2B).
Characteristics of GIS patients in monoallelic and biallelic groups
Our results indicated that most patients carried the DUOX2 gene variant. These patients were divided into three groups according to the pattern of detected variants. Patients with isolated DUOX2 variants were divided into biallelic and monoallelic groups (Figure 3A). The biallelic group contained 42 cases with multiple variants (38 patients with GIS, 3 with G, and 1 with TD), including homozygous or compound heterozygous variants. The monoallelic group included 10 patients with a single variant (6 patients with GIS, 1 with G, and 3 with TD). The oligogenic group included four cases with multiple variants in multiple genes (four patients with GIS). We analyzed the differences in the clinical data between the monoallelic and biallelic groups. We did not observe any differences in sex, birth weight, gestational week, TSH, or FT4 at diagnosis (P > 0.05). However, we observed significant differences in heel blood TSH at screening between the two groups. The monoallelic group displayed significantly higher heel blood TSH levels (176.50 [111.68, 272.50] mIU/L) than the biallelic group (57.50 [15.30, 112.25] mIU/L; P = 0.001). Owing to the small number of oligogenic group cases, we did not compare these cases (Table 3).
L-T4 therapeutic doses administered to GIS patients in monoallelic and biallelic groups
We followed up all patients and collected their L-T4 doses at the initial screening and at 1, 2, and 3 years of age. As shown in Table 1, when participants were enrolled in the study, their ages ranged from 1 to 12 years of age; specifically, some patients had not reached 2 years of age, while others had stopped taking medication at 3 years old (most of these patients were diagnosed with TCH). Thus, 56 patients (48 patients with GIS, 4 with G, and 4 with TD) were receiving treatment at the initial screening and at 1 year of age; 46 patients (40 patients with GIS, 2 with G, and 4 with TD) were receiving treatment at 2 years of age; and only 27 patients (24 patients with GIS, 1 with G, and 2 with TD) were receiving treatment at 3 years of age. We divided these patients into three groups based on the pattern of detected variants (Figure 3A, B and C), and then compared the differences in the L-T4 treatment dose at the initial screening and at 1, 2, and 3 years of age between the monoallelic and biallelic groups. Obviously, the continuous doses of L-T4 in the monoallelic and biallelic groups changed with developmental stage, and the monoallelic group dosage was higher than that of the biallelic group. The L-T4 dose in the biallelic group decreased significantly at 1 year of age, and the dose at 2 and 3 years of age was very similar to that at 1 year of age. The L-T4 dose in the monoallelic group decreased slightly at 1 year of age and then increased gradually. Except for the initial screening and the one performed at 2 years of age, the doses at 1 and 3 years of age were significantly higher in the monoallelic group (36.83 ± 8.23, 39.18 ± 15.71 µg per day) than in the biallelic group (25.87 ± 9.05, 25.38 ± 9.30 µg per day); the differences were statistically significant (P = 0.008, P = 0.030, respectively). Owing to the small number of oligogenic group cases, we did not compare them (Table 4).