Abstract
Objective
Recessive WFS1 mutations are known to cause Wolfram syndrome, a very rare systemic disorder. However, they were also found in non-syndromic diabetes in Han Chinese misdiagnosed with type 1 diabetes (T1D), a molecular cause that appears to be considerably more common than the fully expressed syndrome. We aimed to better define the incidence and clinical features of non-syndromic diabetes due to recessive WFS1 mutation.
Design
We analyzed the genotype and phenotype of 320 consecutive incident Chinese pediatric diabetic patients diagnosed from 2016 to 2019 to search for non-syndromic diabetic cases due to recessive WFS1 mutation.
Methods
A cohort of 105 pancreatic autoantibody-negative patients were recruited for exome sequencing. All patients tested positive for pathogenic diallelic WFS1 mutations were examined for phenotypic features (fundoscopy, audiogram, and urine density).
Results
We found three cases of non-syndromic diabetes due to recessive WFS1 mutations (incidence = 0.94% (95% CI: 0.25–2.7%)). All three cases only had mild diabetes when diagnosed. All patients had well-conserved fasting C-peptide when diagnosed but one of them progressed to T1D-like insulin deficiency. In addition, we found a fourth case with previously undetected features of Wolfram syndrome.
Conclusions
Non-syndromic diabetes due to WFS1 mutation may be common among Chinese pediatric patients with diabetes. It is important to differentiate it from other maturity-onset diabetes in the young subtypes with similar phenotype by molecular diagnosis because of different prognosis and, potentially, therapy.
Introduction
Wolfram syndrome is a rare systemic disorder caused by autosomal recessive WFS1 mutations. The manifestations of the syndrome include juvenile-onset diabetes mellitus, optic nerve atrophy, diabetes insipidus, and deafness (DIDMOAD). These patients usually die of central respiratory failure around the median age of 30 years (1, 2).
However, some cases of recessive WFS1 mutations only have diabetes mellitus without the fully expressed syndrome. Recently, in another study searching for monogenic diabetes cases in a Chinese cohort, clinically diagnosed type 1 diabetes (T1D) but autoantibody negative, we found that cases with diallelic WFS1 variants were the second most common cause of the maturity-onset diabetes in the young (MODY) phenotype next to MODY3, which was the most common one among patients diagnosed because of symptomatic hyperglycemia (3), in the same cohort. These patients were recruited as T1D cases at various ages without any other sign of Wolfram symptoms (4). We recently replicated the same high prevalence of diallelic WFS1 mutations among autoantibody-negative patients of the European ancestry in the T1D Genetics Consortium (5).
In addition to the autosomal recessive Wolfram syndrome, WFS1 mutations may also cause autosomal dominant Wolfram-like syndrome and autosomal dominant deafness (*606201 by Online Mendelian Inheritance in Man). The precise phenotypic spectrum of this condition is not clear so far, and neither is the long-term prognosis. The common phenotype among these cases is isolated diabetes, whose severity ranges from mild to T1D-like manifestations.
Here, we report another three cases of isolated, non-syndromic diabetes caused by recessive WFS1 mutations found in a cohort of 105 consecutive incident autoantibody-negative pediatric cases. One case of fully expressed Wolfram syndrome was also found in this cohort. This case is different in manifestations both at onset and in progression from the non-syndromic recessive WFS1-associated diabetes. This discovery confirms that isolated diabetes mellitus caused by recessive WFS1 mutation is much more common than the fully expressed syndrome and should be considered in the differential diagnosis of T1D and MODY.
Subjects and methods
Subjects
Totally, 320 Chinese pediatric patients were diagnosed with diabetes at the Children’s Hospital, Zhejiang University School of Medicine, from May 2016 to May 2019. Among them, 105 patients from 104 unrelated families were selected with the following criteria: (1) early onset (<18 years) with diabetes mellitus; (2) negative for all pancreatic autoantibodies: GADA, IA-2A, IAA, and ICA; ZnT8A was not available for some of the patients.
Cases with WFS1 mutations were further examined for urine specific gravity, fundus examination, and audiogram. Siblings with the same variants were also examined. Current BMI, HbA1c, and fasting C-peptide were also recorded for probands and affected family members. Pancreatic autoantibodies had been tested on Case 2’s brother who was clinically diagnosed as T1D before.
The study has been approved by the Medical Ethics Committees of the Children’s Hospital, Zhejiang University School of Medicine (approval number 2019-IRB-118). Written consent has been obtained from each patient or subject after full explanation of the purpose and nature of all procedures used.
Clinical examinations
The blood glucose was tested by Beckman AU5800 (hexokinase method). The blood C-peptide was tested by Roche 411 (electrochemiluminescence). The HbA1c was tested by Mindray H50 (ion-exchange column chromatography). The pancreatic autoantibodies were tested by YHLO iFlash3000 (immunoblotting).
Whole-exome sequencing and Sanger sequencing
Exome was captured with SureSelect Human All Exon V6 library (Agilent) and sequenced on the NovaSeq-PE150 (Illumina Inc., San Diego, CA, USA) at 100× by Shanghai Origin-gene. Single nucleotide variants (SNVs) and indels in monogenic diabetes genes were called by GATK (RRID:SCR_001876, URL: https://software.broadinstitute.org/gatk/) and SAMtools (RRID:SCR_002105, URL: http://samtools.sourceforge.net/). Protein-altering variants were filtered to retain missense, frameshift, in-frame insertions/deletions, and canonical splicing variants. Common variants were excluded by minor allele frequency (MAF); >0.0001 for missense or >0.001 for truncating variants in dominant genes, according to any of the three public databases (1000 genomes (RRID:SCR_006828, URL: http://www.1000genomes.org/), Exome Aggregation Consortium (ExAc) (RRID:SCR_004068, URL: http://exac.broadinstitute.org/), and Genome Aggregation Database (gnomAD)) (RRID:SCR_014964, URL: http://gnomad.broadinstitute.org/). For the recessive WFS1, the MAF should be lower than 0.005. The 39 monogenic diabetes-associated genes examined included ABCC8, APPL1, BLK, BSCL2, CEL, CISD2, EIF2AK3, FOXP3, GATA4, GATA6, GCK, GLIS3, HNF1A, HNF1B, HNF4A, IER3IP1, IL2RA, INS, INSR, KCNJ11, KLF11, LMNA, MNX1, NEUROD1, NEUROG3, NKX2-2, PAX4, PAX6, PDX1, POLD1, PPARG, PTF1A, RFX6, SLC19A2, SLC2A2, STAT3, TRMT10A, WFS1, and ZFP57. All variants reported here were verified by Sanger sequencing. The variant of MT-TL1 has also been tested.
Variants were then evaluated by five pathogenicity-prediction algorithms – sorting intolerant from tolerant (SIFT) (RRID:SCR_012813, URL: http://sift.bii.a-star.edu.sg/), PolyPhen-2 (RRID:SCR_013189, URL: http://genetics.bwh.harvard.edu/pph2/), likelihood ratio test (LRT), MutationTaster (RRID:SCR_010777, URL: http://www.mutationtaster.org/), and logistic regression (LR)) and rated according to American College of Medical Genetics and Genomics/Association for Molecular Pathology (ACMG/AMP) guidelines. Exome data from 896 unselected non-diabetic Han Chinese subjects were referred as allele frequency control.
Results
Clinical characteristics
Almost 40% of the patients were asymptomatic at diagnosis, and their hyperglycemia was discovered by routine examination. They were further examined for BMI, HbA1c, C-peptide, and pancreatic autoantibodies. Some of the patients had a family history of diabetes. Most of them still preserve endogenous insulin secretion and require none or low dose of insulin injection.
Molecular findings
Among the 105 antibody-negative probands, at least one variant in monogenic diabetes-associated genes was found in 25 patients by whole-exome sequencing (WES). The variants found involve GCK, INS, KCNJ11, NEUROD1, HNF4A, and WFS1. We also found a case of maternally inherited diabetes and deafness due to a mitochondrial mutation. GCK was the most commonly mutated gene (13/25), followed by WFS1 (4/25), KCNJ11 (4/25), NEUROD1 (3/25), INS (1/25), and HNF4A (1/25) mutations (some of the cases have been reported before (6)). Case 1, Case 2, and Case 3 are compound heterozygotes of rare WFS1 variants, confirmed by parental genotyping showing the variants to be in trans in all three cases. These three cases have only developed diabetes but no other symptoms of the Wolfram syndrome since diagnosed. Case 4 is a homozygote for a WFS1 mutation. The patient had only diabetic manifestation at onset, but it had progressed to fully expressed syndrome 3 years later. Positions refer to transcript NM_006005.3.
Case reports of patients with WFS1 variants
Case 1
A 6-year-old male was found to have hyperglycemia for a week when hospitalized for pneumonia. The random blood glucose (RBG) was between 6.2 and 9.6 mmol/L, and 2-h postprandial blood glucose (2h-PBG) was between 11.8 and 18.8 mmol/L. The patient had no polyphagia, polydipsia, polyuria, weight loss, or other abnormal symptoms.
The BMI of the patient was 17.8 kg/m2, and his HbA1c was 6.3% (45.4 mmol/mol). There was no acanthosis nigricans. After recovery from pneumonia, the fasting plasma glucose (FPG) was 6.5–8.0 mmol/L and the 2h-PG was 8.0–9.5 mmol/L. Oral glucose tolerance test (OGTT) showed that significant pancreatic β-cell function was preserved (Fig. 1). Since the hyperglycemia is mild, no insulin or oral hypoglycemic agents have been administrated. All five pancreatic autoantibodies were negative. The patient had myopia but fundoscopy was negative for optic atrophy (Fig. 1). No other abnormality has been found in laboratory tests or imaging tests.

Clinical characteristics of Case 1. (A) OGTT, patient had been fasted for over 10 h before the test. Glucose was orally taken by 1.75 g/kg body weight within 10 min. The blood glucose, insulin, and C-peptide were tested at the time point of 0, 30, 60, 90, and 120 min. (B) The specific gravity of urine. (C) Image of fundoscopy. (D) Audiogram.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097

Clinical characteristics of Case 1. (A) OGTT, patient had been fasted for over 10 h before the test. Glucose was orally taken by 1.75 g/kg body weight within 10 min. The blood glucose, insulin, and C-peptide were tested at the time point of 0, 30, 60, 90, and 120 min. (B) The specific gravity of urine. (C) Image of fundoscopy. (D) Audiogram.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097
Clinical characteristics of Case 1. (A) OGTT, patient had been fasted for over 10 h before the test. Glucose was orally taken by 1.75 g/kg body weight within 10 min. The blood glucose, insulin, and C-peptide were tested at the time point of 0, 30, 60, 90, and 120 min. (B) The specific gravity of urine. (C) Image of fundoscopy. (D) Audiogram.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097
Since the patient’s only manifestation is isolated and mild hyperglycemia which does not meet the criteria for T1D, it was suspected that he had monogenic diabetes. We found two WFS1 mutations by WES, in trans by parents’ genotype. The two variants both map to WFS1 exon 8. One is a non-frameshift deletion, c.1243_1245del, p.Val415del. This variant is interpreted as likely pathogenic according to the ACMG guidelines and has been reported in Wolfram syndrome cases before (7, 8, 9). The other variant is a missense, c.2534T>G, p.Ile845Ser. It is a variant of uncertain significance (VUS) which has not been previously reported but is predicted deleterious/disease causing by all five computational prediction algorithms (PP3: 5/5) and is estimated to have a Bayesian posterior probability of 80–90% for causing disease (Table 1).
Clinical and biochemical phenotypes and genotypes of the four WFS1 cases. The pathogenicity of the mutations is interpreted based on evidences according to ACMG guidelines. Transcript NM_006005.3.
Case ID | Age at | HbA1c (%; mmol/mol) at | FCP (nmol/L) at | CDS change | AA change | Functional effect | Pathogenicity | |||
---|---|---|---|---|---|---|---|---|---|---|
Diagnosis | Study | Diagnosis | Study | Diagnosis | Study | |||||
1 | 5.7 | 8 | 6.3; 45.4 | 8.1; 65.0 | 0.507 | N/A | c.1243_1245del | p.Val415del | Non-frameshift deletion | Likely Pathogenic (1PS+1PM+1PP) |
c.2534T>G | p.Ile845Ser | Non-synonymous SNV | VUS*(2PM+1PP) PP3: DDDDD | |||||||
2 | 3.3 | 6 | 7.2; 55.2 | 7.9; 62.8 | 0.827 | 0.21 | c.1997G>A | p.Trp666Ter | Stop-gain | Pathogenic (1PVS+1PM+1PP) |
c.2643_2646del | p.Phe882Serfs*69 | Frameshift deletion | Pathogenic (1PVS+2PM) | |||||||
3 | 7.0 | 15 | 110.9; 95.6 | 7.1; 54.1 | 0.97 | 0.52 | c.1997G>A | p.Trp666Ter | Stop-gain | Pathogenic (1PVS+1PM+1PP) |
c.2643_2646del | p.Phe882Serfs*69 | Frameshift deletion | Pathogenic (1PVS+2PM) | |||||||
4 | 3.5 | 6 | 12.3; 10.9 | 6.9; 51.9 | 0.189 | 0.042 | c.616C>T | p.Gln206Ter | stop-gain | Likely pathogenic (1PVS+1PM) |
*Classified as ‘hot’ (posterior probability 80–90%) by the 2020 update of the ACGS guidelines on the basis of deleterious prediction (PP3), low frequency (PM2), and being in trans to a pathogenic variant (PM3).
AA, amino acid; CDS, coding DNA sequence; FCP, fasting C-peptide; the results of computational prediction algorithms are indicated in single-letter codes in this order: For SIFT, D, deleterious; T, tolerated. For Polyphen-2, D, probably damaging; P, possibly damaging, B, benign. For LRT, D, deleterious; N, neutral. For MutationTaster, A, disease causing automatic; D, disease causing; N, polymorphism. For LR, D, deleterious; T, tolerated. PVS, very strong evidence of pathogenicity; PS, strong evidence of pathogenicity; PM, moderate evidence of pathogenicity; PP, supporting evidence of pathogenicity, VUS, variants of uncertain significance.
Both parents and 3-year-old sister are healthy with normal FPG. His sister is heterozygous for one of the variants (c.2534T>G, p.Ile845Ser), by Sanger sequencing.
Two years after diagnosis, the patient’s HbA1c was 8.1% (65.0 mmol/mol) with the BMI of 17.4 kg/m2. His fundoscopy and audiogram was normal. The specific gravity of urine was normal. His heterozygous sister has normal HbA1c, urine specific gravity, fundoscopy, and audiogram.
In Case 1, an additional monogenic diabetes gene, PTF1A also had a variant that passed the frequency filter. It was c.229G>A (NM_178161.2), inherited from the normoglycemic mother and rated as VUS according to the ACMG guidelines. This variant is likely benign by prediction algorithms (BP4: only 2/5 of the algorithms predict it as deleterious).
Case 2 and Case 3
A 4-year-old male (Case 2) was hospitalized for polydipsia and polyuria for 2 weeks. The BMI was 16.9 kg/m2, and HbA1c was 7.2% (55.2 mmol/mol). The RBG was 4.0–8.2 mmol/L, and the 2h-PBG was 7.1–13.4 mmol/L. The fasting plasma C-peptide was 0.827 nmol/L (reference range: 0.37–1.47 nmol/L). Pancreatic autoantibodies were all negative. There was no acanthosis nigricans. The patient had myopia, but fundoscopy was negative. No other abnormality has been found in further laboratory tests or imaging. The patient was treated with insulin since diagnosis.
The patient’s parents are healthy and non-consanguineous with normal FPG. However, he has an elder brother (Case 3) who had been diagnosed with T1D at the age of 7 years. His BMI was 17.2 kg/m2 with HbA1c of 10.9% (95.6 mmol/mol) and the plasma C-peptide of 0.97 nmol/L when diagnosed. His RBG was 4.6–7.9 mmol/L, and the 2h-PBG was 5.8–14.0 mmol/L. All pancreatic autoantibodies were negative. There was no acanthosis nigricans, and fundoscopy was normal, as other laboratory and imaging tests. He was clinically diagnosed as T1D and treated with insulin.
Because the possibility of a T1D proband’s sibling to develop T1D before 20 years of age is very low (4%) (10), and the insulin dosage has been low for 5 years since diagnosis, the brothers were suspected to have monogenic diabetes. WES on the younger brother (Case 2) found two WFS1 mutations verified in trans by sequencing his parents. The elder brother shares the same two mutations in WFS1. Both variants are in exon 8, which appears to be a hot spot. One of them is a stop-gain mutation (c.1997G>A, p.Trp666Ter) and the other causes a frameshift deletion (c.2643_2646del, p.Phe882Serfs*69). Both of the variants were interpreted as pathogenic according to the ACMG guidelines and had been reported in Wolfram syndrome cases before (8, 11, 12) (Table 1).
Recessive WFS1 pathogenic variants are thought usually to cause optic atrophy of the fully expressed Wolfram syndrome (DIDMOAD) soon after the onset of diabetes mellitus (13). On a subsequent visit 1 year after the younger brother’s initial diagnosis, the younger brother (Case 2, 7-years old) had a BMI of 16.4 kg/m2 with HbA1c of 7.9% (62.8 mmol/mol) and plasma C-peptide of 0.21 mmol/L. The BMI of the elder brother (Case 3, 14-years old) was 17.5 kg/m2. His HbA1c was 7.1% (54.1 mmol/mol), and the plasma C-peptide was 0.052 nmol/L. The fundoscopy for both brothers was negative, and the audiogram showed mild decrease at high frequency (Figs 2 and 3). The specific gravity of urine was normal for both (Figs 2 and 3). So far, neither of them has developed the typical fully expressed Wolfram syndrome, after, 1 and 5 years of diabetes, respectively. Both of them are still treated with low dose of insulin.

Clinical characteristics of Case 2. (A) The image of the fundoscopy. (B) Audiogram. (C) The specific gravity of urine.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097

Clinical characteristics of Case 2. (A) The image of the fundoscopy. (B) Audiogram. (C) The specific gravity of urine.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097
Clinical characteristics of Case 2. (A) The image of the fundoscopy. (B) Audiogram. (C) The specific gravity of urine.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097

Clinical characteristics of Case 3. (A) The image of the fundoscopy. (B) Audiogram. (C) The specific gravity of urine.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097

Clinical characteristics of Case 3. (A) The image of the fundoscopy. (B) Audiogram. (C) The specific gravity of urine.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097
Clinical characteristics of Case 3. (A) The image of the fundoscopy. (B) Audiogram. (C) The specific gravity of urine.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097
Case 4
A 3-year-old boy was hospitalized for polydipsia, polyuria for a year, and recent aggravation followed by polyphagia and weight loss for 1 month. The BMI was 14.05 kg/m2, and the HbA1c was 12.3% (110.9 mmol/mol). His plasma insulin was 1.2 μIU/mL with the C-peptide of 0.189 nmol/L. He had no acanthosis nigricans, and the pancreatic autoantibodies tests were all negative. All the other physical examinations, laboratory and imaging tests were normal. His parents were both healthy. The patient has been treated with insulin since being hospitalized. A homozygous variant was found in exon 5. It is a stop-gain mutation (c.616C>T, p.Gln206Ter). It was interpreted as likely pathogenic according to the ACMG guidelines. This variant has not been reported before (Table 1).
Three years later, he was found having cataract (the fundus could hardly be assessed for the lenticular opacity) and moderate hearing loss. His HbA1c has decreased (6.9%, 51.9 mmol/mol), but the C-peptide has almost exhausted (0.042 nmol/L). His urine specific gravity was still normal (Fig. 4 and Table 1).

Clinical characteristics of Case 4. (A) The image of the fundoscopy. (B) Audiogram. (C) The specific gravity of urine.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097

Clinical characteristics of Case 4. (A) The image of the fundoscopy. (B) Audiogram. (C) The specific gravity of urine.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097
Clinical characteristics of Case 4. (A) The image of the fundoscopy. (B) Audiogram. (C) The specific gravity of urine.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097
Beyond the study of this cohort of consecutive patients in one center, we also found 2 more WFS1 cases without syndromic manifestations out of the 91 recent clinical cases from November 2019 to August 2020 by routine clinical diagnosis (Fig. 5 and Table 2). The selection criteria for these 91 cases were not defined preventing us from including them in the incidence estimates.

Clinical characteristics of Case S1. (A) The image of the fundoscopy. (B) Audiogram. (C) The specific gravity of urine.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097

Clinical characteristics of Case S1. (A) The image of the fundoscopy. (B) Audiogram. (C) The specific gravity of urine.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097
Clinical characteristics of Case S1. (A) The image of the fundoscopy. (B) Audiogram. (C) The specific gravity of urine.
Citation: European Journal of Endocrinology 186, 2; 10.1530/EJE-21-0097
Clinical and biochemical phenotypes and genotypes of the two additional WFS1 cases. The pathogenicity of mutations is interpreted based on evidences according to ACMG guidelines. Transcript NM_006005.3.
Case ID | Age at diagnosis | HbA1c at diagnosis | FCP at diagnosis (nmol/L) | CDS change | AA change | Functional effect | Pathogenicity | |
---|---|---|---|---|---|---|---|---|
% | mmol/mol | |||||||
S1 | 11 | 9.5 | 80.3 | 0.672 | c.1997G>A | p.Trp666Ter | Stop-gain | Pathogenic (1PVS+1PM+1PP) |
c.2020G>A | p.Gly674Arg | Non-synonymous SNV | Pathogenic (1PS+2PM+1PP) | |||||
S2 | 4 | 15.6 | 147.0 | 0.782 | c.2054_2055dup | p.Thr686Alafs*25 | Frameshift insertion | Likely pathogenic (1PVS+1PM) |
AA, amino acid; CDS, coding DNA sequence; FCP, fasting C-peptide; the results of computational prediction algorithms are indicated in single-letter codes in this order: For SIFT, D, deleterious; T, tolerated. For Polyphen-2, D, probably damaging; P, possibly damaging, B, benign. For LRT, D, deleterious; N, neutral. For MutationTaster, A, disease causing automatic; D, disease causing; N, polymorphism. For LR, D, deleterious; T, tolerated. PVS, very strong evidence of pathogenicity; PS, strong evidence of pathogenicity; PM, moderate evidence of pathogenicity; PP, supporting evidence of pathogenicity, VUS, variants of uncertain significance. S2 was homozygous.
Statistics against the non-diabetic population
To further validate whether these WFS1 mutations are disease-causing, we compared their results to whole exome data of 896 unselected non-diabetic Han Chinese subjects. Only one had two WFS1 protein-altering variants, both missense predicted benign by all five algorithms and likely benign by ACMG criteria (3/105 vs 0/896, P = 0.0015). None of the WFS1 mutations we reported in the six cases were found in these whole exome data of 896 subjects.
Discussion
Previously, we found that diallelic WFS1 mutations are almost as common as MODY3 in Han Chinese diagnosed with T1D but negative for pancreatic autoantibodies (4). Those cases were recruited as T1D patients for a genome-wide association study (GWAS) and had none of the symptoms of Wolfram syndrome except for diabetes mellitus, when studied. It is possible that some of these patients will progress to other syndromic manifestations; however, what makes us hypothesize that most of these cases will remain non-syndromic is their high incidence, much higher than the extremely rare fully expressed Wolfram syndrome. If all of these cases go on to develop Wolfram syndrome, this will mean a drastic increase in the prevalence of this rare syndrome, worth reporting at this point.
Here, we have confirmed this finding in three cases in an unselected pediatric cohort of only 320, an incidence of almost 1% among all cases of pediatric diabetes or 3% of antibody-negative ones. Only long-term follow-up of these cases, preferably with more sensitive methods such as electroretinography, will give a definite answer.
We found some differences in manifestations between the syndromic and non-syndromic monogenic form of diabetes (MFD) due to WFS1 mutation. Case 4 is a good syndromic control to non-syndromic Case 2 for the same onset age and same follow-up interval since diagnosis. Case 2 had milder diabetes and better conserved C-peptide than Case 4 at onset. Case 4 needed higher dose of insulin (0.74 U/kg) than Case 2 did (0.1 U/kg). Three years after diagnosis, Case 4 had progressed to the syndrome with multiple symptoms and very low C-peptide, while Case 2 still only had isolated diabetes with conserved C-peptide. In general, the tendency has been toward progressive loss of β-cell function.
It is possible that some of these cases we reported would develop other symptoms of Wolfram syndrome in the future. But considering the extreme rarity of the fully expressed Wolfram syndrome, most of these cases should remain non-syndromic.
There may be some similarities between non-syndromic MFD due to WFS1 mutation and other types of MFD in the manifestations at onset. For example, the FPG and HbA1c of Case 1 at initial diagnosis resembled the typical MODY2 (GCK) cases. Specific genetic test of GCK mutation or intensive follow-up might exclude the suspected diagnosis, but the real genetic cause might be ignored. It is necessary to consider WFS1 mutation as well for MODY2-like patients during initial diagnosis.
Since all three cases of the non-syndromic MFD due to recessive WFS1 mutation preserved endogenic insulin production, the therapeutic alteration from insulin to oral hypoglycemic agent may be helpful to improve the patients’ life quality. WFS1 mutations cause β-cell malfunction by endoplasmic reticulum (ER) stress (14), and one of the main beneficial effects of GLP-1 receptor agonists is relief from ER stress (15). Thus, the GLP-1 receptor agonist could be a potential alternative. In addition, glibenclamide or glyburide should be given consideration, as they have been shown to have retinoprotective properties (16, 17).
It is noticed that MODY3 (HNF1A), which is the most common MFD among individuals diagnosed as T1D, was rarely found in our cohort of 320. It might be due to the relatively small size of the cohort and the much younger age of onset in this pediatric cohort (median age at diagnosis, 8.7 years) than typical MFD cohorts. Our pediatric unit treats patients only up to age 12–15. This is not an ethnicity difference, as MODY3 was found as the most common form in another study we did with patients from Chinese patients diagnosed as T1D including a large proportion of adults group (4).
Declaration of interest
Constantin Polychronakos and Yangxi Li are employees of Maida Inc., a for-profit company that offers genetic testing services. The other authors have nothing to disclose.
Funding
This work was supported by the National Key Research and Development Program of China (No. 2016YFC1305301), the National Natural Science Foundation of China (No. 81570759 and 81270938), and the national multi-center cohort study on registration, precision medicine in pediatric diabetes and on phenotype and genotype in disorders of sex development in Chinese mainland (No. G20A0002).
Acknowledgements
The authors acknowledge the financial support from all the funders and appreciate Dr Yongyong Shi for providing the exome data from 896 unselected non-diabetic Han Chinese subjects. The authors also want to thank all patients and their families who consented to participate in the study.
References
- 1↑
Barrett TG, Bundey SE, Macleod AF. Neurodegeneration and Diabetes: UK nationwide study of Wolfram (DIDMOAD) syndrome. Lancet 1995 346 1458–1463. (https://doi.org/10.1016/s0140-6736(9592473-6)
- 2↑
Barrett TG, Bundey SE. Wolfram (DIDMOAD) syndrome. Journal of Medical Genetics 1997 34 838–841. (https://doi.org/10.1136/jmg.34.10.838)
- 3↑
Carlsson A, Shepherd M, Ellard S, Weedon M, Lernmark Å, Forsander G, Colclough K, Brahimi Q, Valtonen-Andre C & Ivarsson SA et al.Absence of islet autoantibodies and modestly raised glucose values at diabetes diagnosis should lead to testing for MODY: lessons from a 5-year pediatric Swedish national cohort study. Diabetes Care 2020 43 82–89. (https://doi.org/10.2337/dc19-0747)
- 4↑
Li M, Wang S, Xu K, Chen Y, Fu Q, Gu Y, Shi Y, Zhang M, Sun M & Chen H et al.High prevalence of a monogenic cause in Han Chinese diagnosed with type 1 diabetes, partly driven by nonsyndromic recessive WFS1 mutations. Diabetes 2020 69 121–126. (https://doi.org/10.2337/db19-0510)
- 5↑
Marchand L, Li M, Leblicq C, Rafique I, Alarcon-Martinez T, Lange C, Rendon L, Tam E, Courville-Le Bouyonnec A, Polychronakos C. Monogenic causes in the Type 1 diabetes genetics consortium cohort: low genetic risk for autoimmunity in case selection. Journal of Clinical Endocrinology and Metabolism 2021 106 1804–1810. (https://doi.org/10.1210/clinem/dgab056)
- 6↑
Ming-Qiang Z, Yang-Li D, Ke H, Wei W, Jun-Fen F, Chao-Chun Z, Guan-Ping D. Maturity onset diabetes of the young (MODY) in Chinese children: genes and clinical phenotypes. Journal of Pediatric Endocrinology and Metabolism 2019 32 759–765. (https://doi.org/10.1515/jpem-2018-0446)
- 7↑
Hansen L, Eiberg H, Barrett T, Bek T, Kjaersgaard P, Tranebjaerg L, Rosenberg T. Mutation analysis of the WFS1 gene in seven Danish Wolfram syndrome families; four new mutations identified. European Journal of Human Genetics 2005 13 1275–1284. (https://doi.org/10.1038/sj.ejhg.5201491)
- 8↑
Gasparin MR, Crispim F, Paula SL, Freire MB, Dalbosco IS, Manna TD, Salles JE, Gasparin F, Guedes A & Marcantonio JM et al.Identification of novel mutations of the WFS1 gene in Brazilian patients with Wolfram syndrome. European Journal of Endocrinology 2009 160 309–316. (https://doi.org/10.1530/EJE-08-0698)
- 9↑
Bodoor K, Batiha O, Abu-Awad A, Al-Sarihin K, Ziad H, Jarun Y, Abu-Sheikha A, Abu Jalboush S, Alibrahim KS. Identification of a novel WFS1 homozygous nonsense mutation in Jordanian children with Wolfram syndrome. Meta Gene 2016 9 219–224. (https://doi.org/10.1016/j.mgene.2016.07.003)
- 10↑
Mayer-Davis EJ, Kahkoska AR, Jefferies C, Dabelea D, Balde N, Gong CX, Aschner P, Craig ME. ISPAD Clinical Practice Consensus Guidelines 2018: definition, epidemiology, and classification of diabetes in children and adolescents. Pediatric Diabetes 2018 19 (Supplement 27) 7–19. (https://doi.org/10.1111/pedi.12773)
- 11↑
Hong J, Zhang YW, Zhang HJ, Jia HY, Zhang Y, Ding XY, Zhou DY, Chen HP, Jiang XH & Cui B et al.The novel compound heterozygous mutations, V434del and W666X, in WFS1 gene causing the Wolfram syndrome in a Chinese family. Endocrine 2009 35 151–157. (https://doi.org/10.1007/s12020-009-9145-7)
- 12↑
Blanco-Aguirre ME, la Parra DR, Tapia-Garcia H, Gonzalez-Rodriguez J, Welschen D, Arroyo-Yllanes ME, Escudero I, Nunez-Hernandez JA, Medina-Bravo P, Zenteno JC. Identification of unsuspected Wolfram syndrome cases through clinical assessment and WFS1 gene screening in type 1 diabetes mellitus patients. Gene 2015 566 63–67. (https://doi.org/10.1016/j.gene.2015.04.040)
- 13↑
Urano F Wolfram syndrome: diagnosis, management, and treatment. Current Diabetes Reports 2016 16 6. (https://doi.org/10.1007/s11892-015-0702-6)
- 14↑
Yamada T, Ishihara H, Tamura A, Takahashi R, Yamaguchi S, Takei D, Tokita A, Satake C, Tashiro F & Katagiri H et al.WFS1-deficiency increases endoplasmic reticulum stress, impairs cell cycle progression and triggers the apoptotic pathway specifically in pancreatic beta-cells. Human Molecular Genetics 2006 15 1600–1609. (https://doi.org/10.1093/hmg/ddl081)
- 15↑
Yusta B, Baggio LL, Estall JL, Koehler JA, Holland DP, Li H, Pipeleers D, Ling Z, Drucker DJ. GLP-1 receptor activation improves beta cell function and survival following induction of endoplasmic reticulum stress. Cell Metabolism 2006 4 391–406. (https://doi.org/10.1016/j.cmet.2006.10.001)
- 16↑
Berdugo M, Delaunay K, Lebon C, Naud MC, Radet L, Zennaro L, Picard E, Daruich A, Beltrand J & Kermorvant-Duchemin E et al.Long-term oral treatment with non-hypoglycemic dose of glibenclamide reduces diabetic retinopathy damage in the goto-KakizakiRat model. Pharmaceutics 2021 13 1095. (https://doi.org/10.3390/pharmaceutics13071095)
- 17↑
Berdugo M, Delaunay K, Naud MC, Guegan J, Moulin A, Savoldelli M, Picard E, Radet L, Jonet L & Djerada Z et al.The antidiabetic drug glibenclamide exerts direct retinal neuroprotection. Translational Research 2021 229 83–99. (https://doi.org/10.1016/j.trsl.2020.10.003)