Systematic Genetic Study of Youth With Diabetes in a Single Country Reveals the Prevalence of Diabetes Subtypes, Novel Candidate Genes, and Response to Precision Therapy
Article Figures & Tables
Figures
- Figure 1
Outline of the study. All 1,209 participants were tested for GAD65, IA-2, and IAA antibodies (Ab). A total of 153 individuals were selected for genetic testing with suspected MD based on one of the following criteria: antibody negative, IAA+ after introduction of insulin treatment, no insulin requirement, positive family history (one first-degree relative or diabetes in three generations), syndromic diabetes.
Tables
- Table 1
Basic clinical characteristics of the cohort at diabetes diagnosis and comparison according to autoimmunity status
All Ab− group Only IAA+ group Other* P (comparing Ab− and only IAA+ groups) % (n) 153 56.2 (86) 42.5 (65) 4.8 (2) NA Female, % (n) 47.7 (73) 53.5 (46) 40 (26) 50 (1) 0.1 Age at diabetes diagnosis, months 106 ± 68 124 ± 71 82 ± 56 135 ± 16 <0.001 Glycemia, mmol/L 15.8 ± 9.4 14.6 ± 10.9 18.3 ± 6.5 9.4 ± 2.2 0.028 HbA1c, % (mmol/mol) 8.2 ± 2.3 (66 ± 2) 7.9 ± 2.9 (63 ± 9) 9.1 ± 2.9 (76 ± 9) 7.1 ± 0.8 (54 ± 1) 0.98 C-peptide, pmol/L 800 ± 2,100† 1,000 ± 2,600‡ 400 ± 400§ 1,000‖ 0.107 Ketosis, % (n) 51.4 (74 of 144) 44.6 (37 of 83) 62.7 (37 of 59) 0 0.033 OHA, % (n) 3.9 (5 of 129) 6.9 (5 of 72) 0 0 <0.001 Insulin, % (n) 65.1 (84 of 129) 47.2 (34 of 72) 90.9 (50 of 55) 0 <0.001 Insulin dose, units/kg/day 0.76 ± 0.46 0.7 ± 0.4 0.9 ± 0.5 — 0.135 Family history of diabetes, % (n) 47.7 (73 of 153) 55.8 (48 of 86) 36.9 (24 of 65) 50 (1 of 2) 0.021 Variants found in MODY genes, % (n) 27.5 (42 of 153) 40.7 (35 of 86) 9.2 (6 of 65) 50 (1 of 2) <0.001 Data are mean ± SD unless otherwise indicated. The 153 individuals selected for genetic analysis were divided into three groups according to their autoimmunity status (antibody negative, only IAA+, and GAD65+), and clinical characteristics were compared. Ab, antibody; OHA, oral hypoglycemic agents.
* One patient GAD65+ and one patient GAD65+ and IAA+.
†Valid data for 57 of 153 cases.
‡Valid data for 37 of 86 cases.
§Valid data for 20 of 65 cases.
‖Valid data for 1 of 2 cases.
- Table 2
Variants identified in additional genes
Patient no. Gene DNA change Protein effect PolyPhen-2 SIFT Class gnomAD allele frequency Molecular function RNA expression in human β-cells (RPKM)* 43 CASP10 c.104del p.Asp35Valfs*10 — — 4 — Protease 0.2 44 DACH1 c.454_456dup p.Ser152dup — — 4 — DNA binding 6.9 45 GCKR c.1484T>G p.Val495Gly 0.96 0 4 — Carbohydrate binding 0.1 6† HGFAC c.711C>G p.Cys237Trp 1 0 4 0.00026 Protease 0 46 HGFAC c.711C>G p.Cys237Trp 1 0 4 0.00026 Protease 0 18† KCNQ1 c.498C>A p.Phe166Leu 0.967 0.05 4 — Potassium channel 0.3 47 MC4R c.230C>T p.Ser77Leu 0.855 0 4 — G-protein–coupled receptor 0 48 RFX2 c.1894G>A p.Ala632Thr 0.999 0 4 — DNA binding 3.9 49 RREB1 c.3218C>T p.Ser1073Leu 0.988 0.02 4 — DNA binding 9 50 SLC5A1 c.932A>T p.Lys311Met 0.999 0 4 — Glucose:sodium symporter 3.5 51 SLC5A1 c.1415T>C p.Leu472Pro 0.892 0 4 — Glucose:sodium symporter 3.5 52 SLC5A1 c.1415T>C p.Leu472Pro 0.892 0 4 — Glucose:sodium symporter 3.5 53 TMPRSS6 c.2263A>G p.Lys755Glu 0.998 0.01 4 — Peptidase 4.4 54 ZBED3 c.211C>G p.Leu71Val 0.977 0 4 — DNA binding 8.3 Summary of all class 4 (likely pathogenic) variants identified in genes not yet associated with MD. Classification of the variants according to the methodology of Richards et al. (12). PolyPhen-2 score considered pathogenic if >0.47 and SIFT <0.05. gnomAD, Genome Aggregation Database; RPKM, reads per kilobase per million mapped reads.
* Nica et al. (13).
†The variants in the HGFAC and KCNQ1 genes are not likely to be causative for diabetes because the two identified carriers also had a GCK variant and a classical GCK diabetes phenotype. The HGFAC variant may also be considered to be too frequent in the gnomAD database to be an MD variant (26).
- Table 3
Comparison of patients with successful and unsuccessful treatment switch
Patients with successful treatment change (n = 9) Patients with unsuccessful treatment change (n = 7) P** Age at diabetes diagnosis, months 178 (1; 262) 120 (1; 159) 0.203 Duration of diabetes until treatment change, months 54 (16; 267) 221 (30; 336) 0.031 Age at treatment change, months 264 (17; 313) 321 (160; 347) 0.023 HbA1c before treatment change, % [mmol/mol] 6.8 (5.5; 11.6) [51 (37; 103)] 8.7 (7.3; 13.1) [72 (56; 120)] 0.080 HbA1c after treatment change, % [mmol/mol] 6.4 (5.6; 10.2) [46 (38; 88)] 11.6 (7.6; 12.8) [103 (60; 116)] 0.011 C-peptide before treatment change, pmol/L# 498 (220; 1,140) 65 (1; 252) 0.005 Insulin treatment, n (%) 8 (89) 7 (100) 1$ Insulin dose before treatment change, units/kg/day 0.5 (0.2; 1.3) 0.9 (0.6; 1.1) 0.117 Data are median (minimum; maximum) unless otherwise indicated. Summary of the 16 patients who were offered treatment switch after genetic diagnosis. Nine patients (six with HNF1A and three with KCNJ11 diabetes) could successfully be treated with sulfonylureas. Seven patients (three with HNF4A, two with ABCC8, one with HNF1A, and one with KCNJ11 diabetes) had to restart insulin therapy.
*Mann-Whitney U test.
$Fisher exact test.
#One missing value in both groups.
In this Issue
