NeuroD1, a basic helix-loop-helix transcription factor, not only regulates endocrine pancreatic development, but also activates insulin gene expression by binding to the insulin promoter via heterodimerization with the ubiquitous helix-loop-helix protein E47. Studies have shown that E2A and Pdx1 together with NeuroD1, which appears to play a dominant role in insulin transcription, cooperatively activate the insulin promoter. In addition to the insulin gene, NeuroD1 also activates MODY genes, such as ABCC8/MODY12, GCK/MODY2, and the transcription factor PAX6, thus indirectly regulating the transcription and expression of the insulin gene.
Mice carrying a homozygous targeted disruption of the NeuroD1/BETA2 gene developed severe diabetes and died perinatally, exhibiting defective pancreatic morphogenesis. The newly formed islet cells in NEUROD1/BETA2-defective mice that lived to adulthood failed to form mature islets, which indicates that NEUROD1/BETA2 is essential for the morphogenesis of normal islet structure. Additionally, insulin-expressing cells that differentiated from human embryonic stem cells also required NEUROD1 to fully activate the essential β-cell transcription factor network.
More than 80% of cases of maturity-onset diabetes of the young (MODY) in Chinese patients are genetically unexplained.[4–8] Mutations in at least 13 genes have been identified to cause MODY, wherein NEUROD1/BETA2 is responsible for the MODY6 subtype. To date, NEUROD1 mutations have not been found in classic MODY pedigrees in Chinese, although we have identified one S159P mutation in an early-onset type 2 diabetes family that is a potential MODY family. Homozygous mutations in NEUROD1 resulted in permanent neonatal diabetes, while heterozygous mutations caused MODY.[10,11] Single-nucleotide polymorphisms, such as Ala45Thr in NeuroD1, increase susceptibility to early-onset type 2 diabetes mellitus.NEUROD1/MODY6, in contrast to the most commonly reported forms including GCK/MODY2, HNF1A/MODY3 and HNF1B/MODY5, is a rarer subtype, and only less than 20 NEUROD1 mutations have been reported to be associated with diabetes, but mostly in Caucasians.
In this study, we performed polymerase chain reaction-direct sequencing of the NEUROD1 in 32 MODY probands who fulfilled the classical MODY criteria and were negative for the five known MODY genes, that is, HNF4A, GCK, HNF1A, IPF-1, and HNF1B, and a missense mutation, E59Q, in NEUROD1 was identified. We also investigated the relationship between changes in the E59Q mutant structure and clinical phenotypes using a de novo modeling method.
Participants and methods
Standardized clinical and laboratory evaluation and NeuroD1 sequencing were conducted in 32 unrelated MODY probands who fulfilled the classical MODY criteria and were referred or recruited by the Department of Endocrinology and Metabolism, Shanghai Diabetes Institute, Shanghai Jiao Tong University Affiliated Sixth People's Hospital, China from 2018 to 2019 in this observational study. The screening criteria of these probands were as follows: (1) onset age <25 years; (2) a family history of diabetes for at least three consecutive generations; (3) absence of autoantibodies; (4) no mutations in the following five MODY genes: HNF4A, GCK, HNF1A, IPF-1, and HNF1B. In addition, we enrolled 201 unrelated, non-diabetic control subjects of Han Chinese descent from the Physical Examination Center of Shanghai Jiao Tong University Affiliated Sixth People's Hospital according to the following criteria: age ≥60 years, no diabetic family history and normal glucose tolerance.
All participants completed medical and family history questionnaires, and their information was supplemented with information from the medical records. The American Diabetes Association criteria (2018) were used to diagnose diabetes, impaired fasting glucose and impaired glucose tolerance. Written informed consent was obtained from all participants. This study was approved by the Institutional Review Board of Shanghai Jiao Tong University Affiliated Sixth People's Hospital, China (approval No. YS-2017-83) on March 3, 2017.
Identification of NeuroD1 mutations
Genomic DNA was extracted from the peripheral leukocytes of the 32 probands, and NeuroD1 mutations in the samples were screened using polymerase chain reaction-direct sequencing and multiplex ligation-dependent probe amplification.[6,7] We used three pairs of primers covering the entire coding sequence of the gene together with flanking sequences of NEUROD1. Sequences were compared with the published sequence (NM_002500.4) using Sequence Navigator (Applied Biosystems Inc., Foster City, CA, USA). The identified mutations were examined for co-segregation with hyperglycemia in other family members using Sanger sequencing, and the genotypes of mutations in the 201 control subjects were tested. The mutations and variants were numbered according to the Human Genome Variation Society (http://www.hgvs.org/). Specific regions of NeuroD1 protein from different mammals were aligned using Clustal O (https://www.ebi.ac.uk/Tools/msa/clustalo/) to evaluate conservation across species.
Three-dimensional model construction for wild-type and the E59Q mutant of NeuroD1
Since there are no crystal structures of NeuroD1 or homologous proteins with high similarity in the protein crystal structure database, a de novo modeling method was used to predict the three-dimensional (3D) structure of wild-type (WT) NeuroD1 and the E59Q mutant.
De novo modeling method
I-TASSER (Iterative Threading the ASSEmbly Refinement) software (http://zhanglab.ccmb.med.umich.edu/I-TASSER) was used to predict the NeuroD1 protein structure. I-TASSER is a method of protein structure and function prediction. Structural templates were first recognized from the PDB library (http://www.rcsb.org/pdb/) using multiple threading alignment approaches, and full-length structure models were then constructed by iterative fragment assembly simulations. Then, the 3D model of the protein was re-optimized using BioLiP (https://zhanglab.ccmb.med.umich.edu/BioLiP/), which integrates the new function library. The I-TASSER program was developed by the Zhang Lab and has been ranked as the top server for protein structure prediction in the Critical Assessment of Structure Prediction project.
Protein structure optimization
To obtain a more reasonable protein structure, the protein model was optimized by the energy optimization method. An amber14 force field was used for energy optimization. The optimization process was divided into two steps: first 5000 steps with the steepest descent followed by 5000 steps with a conjugate gradient were used to further optimize the structure, and the final result was used as a model for subsequent analysis.
The optimized protein model was evaluated by the PROCHECK program (https://www.ebi.ac.uk/thornton-srv/software/PROCHECK/). A Ramachandran plot showed that 97.4% of amino acid residues in the NEUROD1 protein had a reasonable dihedral angle.
Unless otherwise stated, all clinical and laboratory values are expressed as the mean ± standard error of the mean (SEM). These parameters were compared between genotypes using unpaired Student's t tests, in which data with a skewed distribution were logarithmically transformed before analysis and Pearson chi-squared tests (where appropriate). A value of P < 0.05 was considered significant, and SPSS 19.0 (IBM, Armonk, NY, USA) was used for data analysis and processing.
Genetic and clinical phenotype analysis
A heterozygous missense mutation in the NEUROD1 gene, E59Q (c.175 G>C, p.Glu59Gln), was identified in a family with MODY (Fig. 1A and B) but was not detected in other probands or in the 201 non-diabetic control subjects, which suggests that the E59Q mutation is not a simple polymorphism. The E59Q mutation is located at the N-terminal of the NeuroD1 protein (Fig. 1C), and the E59 residue is highly conserved across mammalian species (human, dog, mouse, rat, cattle, horse and monkey) (Fig. 1D), which suggests that E59 may be critical for NeuroD1 protein function.
In this family, the 5 carriers of the E59Q mutation included four diabetic and one non-diabetic individual, indicating non-penetrance. The average age of these 4 carriers at the time of diagnosis was 25 (range 12–30, Table 1) years, and at the time of examination, their diabetes was treated with diet, oral hypoglycemic agents or insulin (Fig. 1A). The fasting and 2-hour postprandial serum insulin levels of diabetic individuals with the E59Q mutation were lower than those of unaffected relatives (fasting plasma insulin, P = 0.020; 2-hour postprandial plasma insulin, P = 0.031), indicating a significant decrease in endogenous insulin secretion (Table 1). In addition, the body mass index of the mutation carriers in this family was 20.3–21.2 kg/m2, and the patients were negative for serum antibodies against glutamate decarboxylase and protein tyrosine phosphatase-like protein, which is consistent with the diagnosis of MODY. Furthermore, the proband's mother did not carry the mutation but developed diabetes at the age of 56 years, and because of undetectable glutamate decarboxylase and tyrosine phosphatase-like protein antibodies, she was diagnosed with type 2 diabetes and achieved satisfactory glycemic control with metformin. These diabetic patients with the E59Q mutation did not have any diabetic complications. In addition, no abnormalities in the central nervous system or gastrointestinal tract were observed in these mutation carriers.
Construction of 3D models for WT and the E59Q mutant of NeuroD1
To study the effect of the NEUROD1-E59Q mutation on protein structure and function, we modeled the structure of NeuroD1 as shown in Figure 2A. Glu59 is located between the two helixes of the NeuroD1 protein, and the mutation could affect the structure of the two surrounding helixes.
To further analyze the structural changes of the E59Q mutation, Figure 2B–E shows the interaction patterns between E59 or Q59 and its surrounding amino acids in the WT or mutant, respectively. The negatively charged Glu59 of the WT forms strong salt bridge bonds with the surrounding positively charged Arg54 and Lys88, and the main chains of Glu59 form hydrogen bonds with Gly56 and Asp61, respectively (Fig. 2B and D). In the E59Q mutation, the salt-bridge bonds of the side chains of Glu59 were disrupted, while the mutated Gln59 formed a new hydrogen bond with Arg54 (Fig. 2C and E).
MODY6 is a rare MODY subtype. Since NEUROD1 was reported as MODY6 gene in 1999, E59Q is the first mutation that has been found in a Chinese family with classic MODY. The E59Q mutation has been reported in a potential MODY family in India, where the onset age of the proband was less than 35 years; however, no analysis of the structure and function was conducted. In our study, the four mutation carriers, the proband and her brother, sister and son, developed overt diabetes at the ages of 25, 30, 33, and 12, respectively. Their mutation was inherited from the proband's father, who was non-diabetic according to an oral glucose tolerate test at 63 years of age, indicating incomplete penetrance. This is also the first report of incomplete penetrance inheritance in a Chinese family with MODY.
The first Chinese NEUROD1/MODY6-E59Q mutation showed incomplete manifestation in a family and was also supported by the inheritance of the MODY6 mutation in other races.[17,18] A MODY study in Iceland showed that the onset age of diabetes caused by the NEUROD1-E110K mutation was up to 68 years, while the NEUROD1-H206PfsTer38 mutation in a Japanese study caused diabetes in a patient at the age of 76. These mutation carriers developed diabetes during a wide age range, from childhood to late adulthood, supporting the incomplete penetrance of MODY6. Although the proband's father had not developed diabetes at the age of 63, the possibility of developing diabetes at a later age could not be ruled out. Therefore, “relaxed” criteria for MODY, that is, 2 consecutive generations with diabetes and 1 patient with a diabetes diagnosis before 25 years of age, may be pivotal to identify additional MODY6 patients. Moreover, incomplete penetrance is clearly a clinical feature of MODY6.[17,18] Interestingly, the proband's mother in this study did not carry this mutation and was diagnosed with type 2 diabetes according to her clinical phenotype and biochemical characteristics. This “third-generation family history of diabetes” was also included in our experimental design to recruit probands and family members using the classic MODY diagnostic criteria. Therefore, it is possible to identify more MODY6 patients using the “relaxed” criteria for MODY in the Chinese population.
Comparison of the E59Q and S159P mutations in NeuroD1
The phenotype of MODY6 exhibits a wide clinical spectrum, from patients with typical characteristics of MODY to those with characteristics of common type 2 diabetes. Different clinical phenotypes in diabetic patients were observed between E59Q and S159P mutation carriers. Diabetic patients with the E59Q mutation developed diabetes at an early age, had a lean body mass index of 20.9 (20.3–21.2)kg/m2, and exhibited impaired insulin secretion, whereas S159P mutation carriers were diagnosed with diabetes at an older age and had a relatively strong ability to secrete insulin (E59Q vs S159P, homeostasis model assessment of β-cell function, 12.2 ± 2.3 vs 56.0 ± 13.9, P = 0.004).
Effects of N-terminal mutation on NeuroD1 function
The human NEUROD1 gene contains two exons and maps to chromosome 2q32, and the encoded protein comprises an N-terminal region, a basic helix-loop-helix domain and a C-terminal transactivation domain.[1,2] Although the basic helix-loop-helix domain is essential for dimerization with E47 and DNA binding, truncated NeuroD1 lacking either the C- or N-terminal region showed a markedly reduced ability to activate the INS promoter. The E59Q mutation identified in this study was located at the N-terminal of the NeuroD1 protein, and the E59 residue is highly conserved across mammalian species, which suggests the functional importance of this residue. In the E59Q mutation, the negative charge of E59 is substituted by a non-charged polar Q59 residue.
We first modeled the 3D structure of the WT and mutant NeuroD1 proteins. In the NEUROD1-E59Q mutant, the loss of negative charge resulted in a weakened interaction between Q59 and the surrounding amino acids, namely the loss of salt bridge bonds (Glu59-Arg54 and Glu59-Lys88) and the formation of a new hydrogen bond (Gln59-Arg54). This change may alter the local conformation of the N-terminal and cause structural instability, thus affecting insulin transcription. Furthermore, this alteration may explain why insulin secretion was significantly decreased in the diabetic individuals carrying the E59Q mutation identified in this study. However, whether and how this mutation affects the transcription of the insulin gene in β-cells still needs to be confirmed. The development of single-cell sequencing technology has allowed the assessment of functional changes of the NeuroD1 protein expressed in different tissues or cells (such as pancreatic β-cells, neuronal cells or gastrointestinal cells) as well as the investigation of the potential regulatory mechanism affected by the E59Q mutation.[1,2,19] Malecki et al indicated that the amino acid residues at the end of the N-terminal comprise the key region of the NeuroD1 protein that combines with Gal4 DNA. Therefore, because E59 is located at the N-terminal, we speculated that the weakened interaction between the E59Q mutant and the surrounding amino acid residues may affect the interaction of the NeuroD1 protein with Gal4 DNA, which in turn leads to weakened or even loss of transcription function. Further experiments will be conducted in either cultured cells in vitro or animal models in vivo to verify the interaction of the NeuroD1 protein and Gal4 DNA and show that insulin transcription is disrupted.
Further studies should add functional tests in vitro, to provide evidence for the pathogenesis of neurod1-E59Q mutations. In addition, to explore the prevalence and more accurate mutation spectrum of NeuroD1 among Chinese patients with MODY, future study should expand the numbers of MODY pedigrees.
We identified the first NEUROD1 mutation, E59Q, in a Chinese family with classic MODY. The NEUROD1-E59Q mutation clearly affected the interaction between the E59 residue and its surrounding amino acids and thus changed the molecular conformation of the N-terminal of NeuroD1. The abnormal conformation of NeuroD1 may decrease binding of the E59Q mutant to the insulin promoter and insulin transcription activity, therefore causing a MODY6 subtype with defective insulin secretion.
We apologize to the many authors whose works could not be cited due to space limitations.
LL designed the study. LL, YL, MJ, and MY revised the manuscript. All the authors contributed to drafting the manuscript and approved the final version of the manuscript.
This work is supported by the National Natural Science Foundation of China (Nos. 81970686, 81770791, 81471012, 81270876; to LL), the Interdisciplinary Program of Shanghai Jiao Tong University, China (No. YG2019ZDA08; to LL), and the Shanghai Leading Talent, China (No. SLJ15055; to LL), and the National Institute of Diabetes and Digestive and Kidney Diseases (No. SC1DK104821; to YL).
Institutional review board statement
This study was approved by the Institutional Review Board of Shanghai Jiao Tong University Affiliated Sixth People's Hospital, China (approval No. YS-2017-83) on March 3, 2017.
Declaration of participant consent
The authors certify that they have obtained the participant consent forms. In the forms, participants have given their consent for their images and other clinical information to be reported in the journal. The participants understand that their names and initials will not be published and due efforts will be made to conceal their identity.
Conflicts of interest
The authors declare that they have no conflicts of interest.
1. Naya FJ, Huang HP, Qiu Y, et al. Diabetes, defective pancreatic morphogenesis, and abnormal enteroendocrine differentiation in BETA2/neuroD-deficient mice. Genes Dev
2. Horikawa Y, Enya M. Genetic dissection and clinical features of MODY6
(NEUROD1-MODY). Curr Diab Rep
3. Romer AI, Singer RA, Sui L, et al. Murine perinatal β-cell proliferation and the differentiation of human stem cell-derived insulin-expressing cells require NEUROD1. Diabetes
4. Xu JY, Dan QH, Chan V, et al. Genetic and clinical characteristics of maturity-onset diabetes of the young in Chinese
patients. Eur J Hum Genet
5. Liu L, Furuta H, Minami A, et al. A novel mutation
, Ser159Pro in the NeuroD1/BETA2 gene contributes to the development of diabetes in a Chinese
potential MODY family. Mol Cell Biochem
6. Liu L, Nagashima K, Yasuda T, et al. Mutations in KCNJ11 are associated with the development of autosomal dominant, early-onset type 2 diabetes. Diabetologia
7. Liu L, Liu Y, Ge X, et al. Insights into pathogenesis of five novel GCK mutations identified in Chinese
MODY patients. Metabolism
8. Zhang J, Liu Y, Li M, et al. Identification of Ala2Thr mutation
in insulin gene from a Chinese
MODY10 family. Mol Cell Biochem
9. American Diabetes Association. Classification and diagnosis of diabetes: Standards of medical care in diabetes-2020. Diabetes Care
10. Demirbilek H, Hatipoglu N, Gul U, et al. Permanent neonatal diabetes mellitus and neurological abnormalities due to a novel homozygous missense mutation
in NEUROD1. Pediatr Diabetes
11. Bouillet B, Crevisy E, Baillot-Rudoni S, et al. Whole-exome sequencing identifies the first French MODY 6 family with a new mutation
in the NEUROD1 gene
. Diabetes Metab
2020; doi: 10.1016/j.diabet.2020.03.001.
12. Liu L, Jia W, Zheng T, et al. Ala45Thr variation in neuroD1 gene
is associated with early-onset type 2 diabetes with or without diabetic pedigree in Chinese
. Mol Cell Biochem
13. Fajans SS, Bell GI, Polonsky KS. Molecular mechanisms and clinical pathophysiology of maturity-onset diabetes of the young. N Engl J Med
14. American Diabetes Association. 2. Classification and diagnosis of diabetes: Standards of medical care in diabetes-2018. Diabetes Care
15. Yang J, Zhang Y. I-TASSER server: new development for protein structure and function predictions. Nucleic Acids Res
16. Chapla A, Mruthyunjaya MD, Asha HS, et al. Maturity onset diabetes of the young in India - a distinctive mutation
pattern identified through targeted next-generation sequencing. Clin Endocrinol (Oxf)
17. Malecki MT, Jhala US, Antonellis A, et al. Mutations in NEUROD1 are associated with the development of type 2 diabetes mellitus. Nat Genet
18. Horikawa Y, Enya M, Mabe H, et al. NEUROD1-deficient diabetes (MODY6
): Identification of the first cases in Japanese and the clinical features. Pediatr Diabetes
19. Zhang H, Cui N, Cai Y, et al. Single-cell sequencing leads a new era of profiling transcriptomic landscape. J Bio-X Res