Edrv 0190

R E V I E W
Monogenic Diabetes: What It Teaches Us on the

Common Forms of Type 1 and Type 2 Diabetes
Yisheng Yang and Lawrence Chan

Division of Endocrinology (Y.Y.), Department of Medicine, MetroHealth Medical Center, Case Western Reserve
University, Cleveland, Ohio 44109; and Diabetes and Endocrinology Research Center (L.C.), Division of Diabetes,
Endocrinology and Metabolism, Departments of Medicine, Molecular and Cellular Biology, Biochemistry and Molecular
Biology, and Molecular and Human Genetics, Baylor College of Medicine, Houston, Texas 77030
Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

To date, more than 30 genes have been linked to monogenic diabetes. Candidate gene and genome-wide
association studies have identified ⬎ 50 susceptibility loci for common type 1 diabetes (T1D) and approximately
100 susceptibility loci for type 2 diabetes (T2D). About 1–5% of all cases of diabetes result from single-gene
mutations and are called monogenic diabetes. Here, we review the pathophysiological basis of the role of
monogenic diabetes genes that have also been found to be associated with common T1D and/or T2D. Variants
of approximately one-third of monogenic diabetes genes are associated with T2D, but not T1D. Two of the
T2D-associated monogenic diabetes genes—potassium inward-rectifying channel, subfamily J, member 11
(KCNJ11), which controls glucose-stimulated insulin secretion in the ␤-cell; and peroxisome proliferator-acti-
vated receptor ␥ (PPARG), which impacts multiple tissue targets in relation to inflammation and insulin sensi-
tivity— have been developed as major antidiabetic drug targets. Another monogenic diabetes gene, the pre-
proinsulin gene (INS), is unique in that INS mutations can cause hyperinsulinemia, hyperproinsulinemia,
neonatal diabetes mellitus, one type of maturity-onset diabetes of the young (MODY10), and autoantibody-
negative T1D. Dominant heterozygous INS mutations are the second most common cause of permanent neo-
natal diabetes. Moreover, INS gene variants are strongly associated with common T1D (type 1a), but inconsis-
tently with T2D. Variants of the monogenic diabetes gene Gli-similar 3 (GLIS3) are associated with both T1D and
T2D. GLIS3 is a key transcription factor in insulin production and ␤-cell differentiation during embryonic de-
velopment, which perturbation forms the basis of monogenic diabetes as well as its association with T1D. GLIS3
is also required for compensatory ␤-cell proliferation in adults; impairment of this function predisposes to T2D.
Thus, monogenic forms of diabetes are invaluable “human models” that have contributed to our understanding
of the pathophysiological basis of common T1D and T2D. (Endocrine Reviews 37: 190 –222, 2016)
I. Introduction I. Introduction
II. Monogenic Diabetes Genes Associated With T2D, But
ype 1 diabetes (type 1a; hereafter referred to as T1D)
Not T1D
A. KCNJ11 and ABCC8
B. Glucokinase
T and type 2 diabetes (T2D) are of multifactorial eti-
ology, in which genetic predisposition plays a key role.
C. Solute carrier family 2, member 2
Monogenic diabetes is relatively rare. Neonatal diabetes
D. HNF1A and HNF4A
E. Hepatocyte nuclear factor 1 homeobox B mellitus (NDM) and maturity-onset diabetes of the young
F. Pancreatic and duodenal homeobox 1
Abbreviations: ABCC8, ATP-binding cassette transporter subfamily C member 8; CCND2,
G. Paired box 4 cyclin D2; CHI, congenital hyperinsulinism; e, embryonic day; ER, endoplasmic reticulum;
H. NEUROD1/BETA2 FBS, Fanconi-Bickel syndrome; FGF, fibroblast growth factor; GCK, glucokinase; GDM,
I. Wolfram syndrome 1 gestational diabetes mellitus; GKRP, GCK regulatory protein; GLIS3, Gli-similar 3; GLP-1,
J. Peroxisome proliferator-activated receptor ␥ glucagon-like peptide-1; GLUT, glucose transporter; G6PC2, glucose-6-phosphatase, cat-
alytic 2; GSIS, glucose-stimulated insulin secretion; GWAS, genome-wide association stud-
III. Monogenic Diabetes Genes Associated With Both T1D ies; HFD, high-fat diet; HLA, human leukocyte antigen; HLH, helix-loop-helix; HNF1A,
and T2D hepatocyte nuclear factor 1 homeobox A; IGT, impaired glucose tolerance; KATP, ATP-
A. Preproinsulin gene, INS sensitive potassium; KCNJ11, potassium inward-rectifying channel, subfamily J, member
11; KPD, ketosis-prone diabetes; MAF, minor allele frequency; MODY, maturity-onset
B. Gli-similar 3
diabetes of the young; NDH, neonatal diabetes, hypothyroidism; NDM, neonatal diabetes
IV. Concluding Remarks mellitus; PAX4, paired box 4; PDX1, pancreatic and duodenal homeobox 1; PNDM, per-
manent NDM; PPARA, peroxisome proliferator-activated receptor ␣; PPARG, peroxisome
ISSN Print 0163-769X ISSN Online 1945-7189 proliferator-activated receptor ␥; RCAD, renal cysts and diabetes; SGLT2, sodium-glucose
Printed in USA cotransporter 2; SLC2A2, solute carrier family 2, member 2; SUR1, sulfonylurea receptor
Copyright © 2016 by the Endocrine Society 1; T1D, type 1 diabetes; T2D, type 2 diabetes; TH, tyrosine hydroxylase; TNDM, transient
Received October 22, 2015. Accepted March 28, 2016. NDM; TZD, thiazolidinedione; UPR, unfolded protein response; UTR, untranslated region;
First Published Online March 30, 2016 VNTR, variable number of tandem repeats; WFS1, Wolfram syndrome 1.
190 press.endocrine.org/journal/edrv Endocrine Reviews, June 2016, 37(3):190 –222 doi: 10.1210/er.2015-1116
doi: 10.1210/er.2015-1116 press.endocrine.org/journal/edrv 191
Figure 1. 2 (TCF7L2) rs7903146 is one of the

most robust susceptibility variants
for T2D identified to date. Although
T1D and T2D share some clinical
and pathological similarities, GWAS
indicate that their susceptibility loci
are mostly distinct, and only a lim-
ited number of them have been found
to be associated with both T1D and
T2D (12–15). These include: INS,
GLIS3, RAS guanyl nucleotide-

releasing protein 1 (RASGRP1),
cordon-bleu WH2 repeat protein
Figure 1. Monogenic diabetes genes associated with NDM and/or MODY. Over 30 monogenic (COBL), renalase (RNLS), and
diabetes genes have been identified to cause NDM or MODY. Mutations in seven of them, breast cancer antiestrogen resistance
including ABCC8, GCK, HNF1B, INS, KCNJ11, NEUROD1, and PDX1, may lead to both NDM and
MODY. Bolded genes are discussed in this article.
1 (BCAR1). Most of the loci associ-
ated with T1D and T2D (7–10) will
not be discussed further in this article
(MODY) are two major forms of monogenic diabetes, which because they are not known to cause monogenic diabetes.
usually result from mutations of genes for transcription fac- Monogenic diabetes represents unique in vivo models of
tors or other proteins that regulate endocrine pancreas de- disordered developmental and functional biology of pan-
velopment or function. Mutations in a number of monogenic creatic ␤-cells being played out, not in animals, but in
diabetes genes may cause NDM and MODY (Figure 1). people. Importantly, before the development of GWAS,
NDM presents within the first 6 months of life and can persist monogenic diabetes genes provided the genetic landscape
throughout life (called permanent NDM [PNDM]), or it can toward the identification of new candidate genes for the
be transient and disappear during infancy, often reappearing common forms of T1D and T2D. In the GWAS and post-
later in life (called transient NDM [TNDM]) (1, 2). In con- GWAS era, the different types of monogenic diabetes
trast to classic T1D (called common T1D in this article), continue to provide pathophysiological insight for the
NDM is only rarely associated with high-risk human leuko- interpretation of GWAS data. In this context, we will
cyte antigen (HLA) haplotypes or the presence of islet auto- focus on the mechanistic defects that underlie monogenic
immune antibodies (1, 3); eg, mutations of the NDM gene diabetes genes that have also been found to be associated
signal transducer and activator of transcription 3 have re- with common T1D and/or T2D (Table 1 and Figure 2).
cently been reported to be associated with early-onset auto-
immune disorders including T1D in rare cases (4). MODY is
II. Monogenic Diabetes Genes Associated With
an early-onset (presenting usually before the age of 25 years)
T2D, But Not T1D
monogenic diabetes caused by mutations in a number of islet-
related genes and is inherited in an autosomal dominant GWAS revealed that approximately one-third of monogenic
manner. Mutations in hepatocyte nuclear factor 1 homeo- diabetes genes are associated with common T2D, but not
box A (HNF1A) (HNF1A-MODY or -MODY3) and glu- T1D. The 12 monogenic diabetes genes in this category are:
cokinase (GCK-MODY or MODY2) underlie the two most potassium inwardly-rectifying channel, subfamily J, member
common forms of MODY (5, 6). Of note, MODY is a het- 11 (KCNJ11), ATP-binding cassette transporter subfamily C
erogeneous condition. The clinical phenotype can vary sub- member 8 (ABCC8), GCK, solute carrier family 2, member
stantially among patients with the same MODY genotype, 2 (SLC2A2), HNF1A, HNF4A, HNF1B, pancreatic and du-
even among members in the same MODY family. The effect odenal homeobox 1 (PDX1), paired box 4 (PAX4), NEU-
of gene-gene and gene-environment (eg, diet and exercise) ROD1/BETA2, Wolfram syndrome 1 (WFS1), and perox-
interactions may underlie the heterogeneities (5). isome proliferator-activated receptor ␥ (PPARG).
To date, more than 30 genes have been linked to mono-
genic diabetes and related syndromes (1, 2, 5). More than 50 A. KCNJ11 and ABCC8
and 90 susceptibility loci, respectively, are found to be asso- The ATP-sensitive potassium (KATP) channel is composed
ciated with T1D and T2D (7–11) (Figure 2). In particular, the of four pore-forming inwardly-rectifying potassium chan-
HLA class II region of the major histocompatibility complex nel subunits (Kir6.2), encoded by KCNJ11, and four reg-
confers major genetic risk for T1D. Transcription factor 7-like ulatory sulfonylurea receptor 1 (SUR1) subunits, encoded
192 Yang and Chan Monogenic Diabetes Genes and Common Types of Diabetes Endocrine Reviews, June 2016, 37(3):190 –222
Figure 2.

Figure 2. Venn diagram of intersection between the loci/genes associated with T1D or T2D and known monogenic diabetes. Over 50 susceptibility
loci for common T1D and approximately 100 susceptibility loci for T2D have been identified by GWAS and candidate gene association studies. INS,
GLIS3, RASGRP, COBL, RNLS, and BCAR1 are associated with both T1D and T2D. About one-third of the known monogenic diabetes genes are
associated with T2D. INS and GLIS3 are the two known monogenic diabetes genes whose variants are associated with both T1D and T2D.
by ABCC8. An exquisitely regulated KATP channel con- in the regulation of insulin secretion. Loss-of-function mu-
trols insulin secretion by coupling ␤-cell metabolism to tations in the KCNJ11 and ABCC8 genes are the most com-
calcium entry. ATP and MgADP are thought to display mon cause of CHI (17). Most of the disease-causing muta-
opposing roles in closing or opening the KATP channel. tions in the KATP channel genes are recessively inherited,
High blood glucose is taken up and metabolized by the leading to medically unresponsive CHI, which frequently re-
␤-cell that leads to elevated intracellular ATP level and quires near total pancreatectomy to relieve the recurrent se-
closure of the KATP channel. The reduction in K⫹ efflux vere hypoglycemia. Mechanistically, these recessive muta-
gives rise to membrane depolarization, resulting in open- tions adversely affect the biogenesis and turnover of KATP
ing of the voltage-gated Ca2⫹ channel, triggering electrical channels, resulting in defective trafficking of channels to the
activity, Ca2⫹ influx, and insulin secretion (Figure 3A). plasma membrane and altered open-state frequency (18 –
Inactivating mutations in KCNJ11 or ABCC8 result in 20). Increasing numbers of mutations within the nucleotide-
congenital hyperinsulinism (CHI), whereas activating binding domain 2, a hotspot for dominantly acting muta-
mutations in either gene that interferes with the ATP tions in the ABCC8 gene, have been identified in CHI
sensitivity of the KATP channel can lead to NDM (16). patients. These include E1506K, G1479R, R1539Q,
Monogenic mutations in the KATP channel components L1390R, L1431F, Q1459E, A1508P, A1537V, and
have taught us how insulin secretion is regulated and R1420H (21–23). An in-frame heterozygous deletion
how insulin secretagogues such as sulfonylureas work. (I284del) in the KCNJ11 gene (21) and some compound
heterozygous mutations in the ABCC8 gene (24) have also
1. Inactivating mutations in the KCNJ11 and ABCC8 genes been reported in CHI. These inactivating mutations cause
result in CHI hyperinsulinemia predominately through reduced KATP
We will briefly review the molecular basis of CHI because channel activity in pancreatic ␤-cells, which leads to abnor-
it will shed light on the role of the KCNJ11 and ABCC8 genes mal membrane polarization, activation of voltage-gated cal-
Table 1. Monogenic Diabetes Genes Associated With Common T1D and/or T2D
Mutations or Variants
Monogenic Diabetes or Associated With Common
Gene Name Major Function Syndromes T1D and/or T2D Refs.
Monogenic Diabetes Genes Associated With T2D

KCNJ11 Encodes pore-forming inwardly-rectifying PNDM (most common cause) E23K 42– 46
potassium channel subunits (Kir6.2) and TNDM, CHI, MODY
ABCC8 Encodes regulatory SUR1 subunits PNDM and TNDM, CHI, MODY A1369S, 1273AGA, R1420H 46, 47, 52
GCK A key glucose-phosphorylating enzyme; GCK-MODY (MODY2), PNDM, rs1799884 (G/A), rs4607517 (A/G), 75, 78, 79
a glucose sensor CHI 3⬘UTR SNP, chr7:44184184-G/A
SLC2A2 Encodes GLUT2, a high-capacity facilitative FBS SNPs rs5393 (AA) and rs5394 93–100
glucose transporter (CC) in the promoter region
and SNPs rs5400 (T110I) and
rs5404 (T198T)

HNF1A/TCF1 TF; regulator of pancreatic ␤-cell HNF1A-MODY (MODY3), most G319S, c.1522G⬎A (p.E508K) 114, 118, 119
differentiation common cause of MODY,
CHI
HNF4A Key TF for early fetal development HNF4A MODY (MODY1), CHI SNPs rs2144908, rs3818247 and 121–124, 274
rs884614, rs4810424, rs1884613
HNF1B/TCF2 TF; required for the generation of RCAD syndrome, or MODY5; SNP rs757210 A, rs4430796 A, 141, 144
pancreatic and endocrine progenitors TNDM and PNDM (rare) and rs7501939 C
PDX1 TF; required for pancreas development, ␤-cell PNDM, MODY4 C18R, Q59L, D76N, R197H, 163–165, 167
differentiation and the maintenance of G212R, P239Q, InsCCG243,
mature ␤-cell function p.Gly218Alafs*12
PAX4 Islet TF that functions mainly as a MODY9 R121W, R133W, R37W, 180, 181, 187
transcription repressor rs10229583 G
NEUROD1/BETA2 TF; required for the development of the MODY6 and PNDM R111L and 206 ⫹ C; A45T 204 –208
endocrine pancreas; transactivates the variant at rs1801262
INS gene (inconsistent)
WFS1 A transmembrane protein; a negative WFS1, sometimes referred to R456 and H611, SNPs at 223–225
regulator of ER stress as DIDMOAD rs10010131, rs6446482;
variants rs10010131 G,
1801213 G, and
rs734312 A
PPARG TF; master regulator of adipogenesis, energy Monogenic diabetes Pro12Ala variant (rs1801282), 240 –243, 250
balance, lipid biosynthesis, and insulin SNP at rs4684847
sensitivity; cellular target of TZDs
Monogenic Diabetes Genes Associated With Both Common T1D and T2D
INS Predominant glucose-lowering hormone PNDM (2nd most common Class I alleles of INS VNTR 273, 274, 276 –281
cause), TNDM, MODY10 associated with T1D;
Class III alleles of INS
VNTR inconsistently
associated with T2D
GLIS3 TF; regulator of islet development, insulin Neonatal diabetes syndrome rs7020673 G associated 78, 214, 289, 291,
gene transcription, and obesity-induced associated with congenital with T1D; rs7034200 A 292, 295–308
compensatory ␤-cell proliferation hypothyroidism, and and rs7041847 A
polycystic kidneys associated with T2D
Abbreviation: TF, transcription factor.
cium channels, and increased Ca2⫹ concentration causing sive CHI patients (28). The mechanism underlying the in-
insulin hypersecretion (Figure 3B). hibition of insulin secretion by octreotide has not been
Dominant KATP channel mutations causing CHI may fully elucidated. Octreotide inhibits Ca2⫹ entry into pan-
predispose to the development of diabetes in adulthood. creatic ␤-cells via voltage-dependent calcium channels,
Nonpancreatectomized KATP-CHI patients eventually en- leading to suppression of insulin secretion (29, 30). In
ter clinical remission and may progress to diabetes in later addition, octreotide may hyperpolarize ␤-cells by activat-
life (23–27), whereas those who are treated by partial pan- ing KATP channel and thus inhibits insulin secretion (31).
createctomy often develop late-onset diabetes that may
require insulin therapy (17). 2. Activating mutations in the KCNJ11 and ABCC8 genes
Diazoxide is one of the primary medications used to lead to PNDM
treat CHI (28). In pancreatic ␤-cells, diazoxide, a K⫹ chan- Gain-of-function mutations in the KATP channel genes
nel opener, binds to SUR1 subunits to increase KATP chan- impair ATP binding to the channel, leading to KATP chan-
nel activity, promoting K⫹ efflux and cell membrane hy- nel opening, membrane hyperpolarization, impaired in-
perpolarization to block insulin release. Octreotide (a sulin release, and NDM (Figure 3C). Activating mutations
somatostatin analog) can be used for diazoxide-unrespon- of KCNJ11 were first identified over a decade ago in pa-
Figure 3.

Figure 3. Insulin secretion in normal and KATP channel mutant pancreatic ␤-cells. A, Glucose-stimulated insulin secretion in normal ␤-cells. The KATP
channel is composed of four Kir6.2 subunits encoded by KCNJ11 and four SUR1 subunits encoded by ABCC8. High glucose in the circulation leads
to increased glucose uptake into pancreatic ␤-cells. Increased intracellular glucose is metabolized via glycolytic and mitochondrial metabolism,
leading to an increase in ATP production and a fall in MgADP. This results in KATP channel closure, membrane depolarization, opening of
voltage-gated Ca2⫹ channels, Ca2⫹ influx, and insulin release. B, Insulin overproduction in the ␤-cell with KATP channel-inactivating
mutations. Loss-of-function mutations in KATP channel enhances ATP binding to the channel, leading to KATP channel closure, membrane
depolarization, insulin oversecretion and CHI. C, Impaired insulin secretion in the ␤-cell with KATP channel activating mutations. Gain-of-
function mutations in KATP channel impair ATP binding to the channel Kir6.2 mutant (shown in olive green, encoded by KCNJ11), or
enhance the binding of Mg-nucleotide to SUR1 mutant (shown in amber, encoded by ABCC8) leading to KATP channel opening, membrane
hyperpolarization, impaired insulin release, and NDM. D, Oral sulfonylureas stimulate insulin secretion in patients with KATP channel-
activating mutations. Sulfonylureas bind to the ABCC8-encoded SUR1 subunit of the KATP channel to effect channel closure independent of
ATP and enable insulin secretion.
tients with PNDM and account for approximately 30% of a few patients display severe developmental delay, epi-
cases (32). Subsequently, gain-of-function mutations of lepsy, and NDM (called DEND syndrome) (38). In addi-
ABCC8 were reported (33, 34); mutations of these two tion to causing NDM, mutations of KCNJ11 and ABCC8
genes have been identified in ⬎ 40% of PNDM patients have been recently linked to MODY (39, 40).
(35). Mutations in KCNJ11 and ABCC8 may also lead to Subtle differences in the primary structure of the
TNDM. More than 100 different mutations in the ABCC8-encoded SUR1 protein can produce diametrically
KCNJ11 and ABCC8 genes have been found in patients opposite phenotypes, as illustrated by the fact that, de-
with PNDM (36). All but one are missense mutations that pending on the specific amino acid substitution, mutations
cause single amino acid substitutions; the only exception at the same amino acid residue may cause either CHI or
to date is an in-frame 15-bp deletion in the KCNJ11 gene NDM. For example, an E1506K substitution in the nu-
(37). Most of the patients with mutations in the KCNJ11 cleotide-binding domain 2 of SUR1 leads to CHI (23, 27),
and ABCC8 genes present with isolated neonatal diabetes; whereas two other mutations at the same residue
(E1506D, E1506G) were found to cause NDM (41). KATP ant alone did not predict conversion to diabetes (50); in
channels are activated by Mg-nucleotides (via SUR1) and fact, it might even be protective (51). Recently, Baier et al
blocked by ATP (via Kir6.2). All of these mutations reduce (52) reported that ABCC8 R1420H homozygotes even-
channel activation by MgADP. Different sensitivities to tually developed diabetes, despite being hyperinsulinemic
ATP inhibition could be the basis of these opposing clin- during infancy. Furthermore, the R1420H variant was
ical phenotypes. For example, E1506D and E1506G mu- associated with higher birth weights and a 2-fold increased
tant KATP channels display a markedly impaired sensitiv- risk for T2D in 3.3% of an American Indian population.
ity to ATP, resulting in hyperactive channels associated Their increased birth weights are suggestive of insulin
with impaired insulin secretion and PNDM. In contrast, oversecretion due to insulin being a major growth factor
compared to wild-type channels, the E1506K mutant in utero. However, the physiological mechanisms under-
channels are more sensitive to ATP inhibition, produc- lying the reversal from hyperinsulinemia to diabetes re-

ing underactive channels that result in insulin overse- main elusive. The possible culprits include increased ␤-cell
cretion and CHI (41). apoptosis due to elevated intracellular calcium concentra-
tion and the accompanying insulin resistance secondary to
3. KCNJ11 and ABCC8 gene variants associated with T2D increased body weight (17, 25, 27).
A common variant involving a glutamate (E) to lysine In a subgroup analysis of the UK Prospective Diabetes
(K) substitution at position 23 (E23K) in KCNJ11 has Study (UKPDS), Gloyn et al (53) found that the presence
been found to be associated with increased risk of T2D of KCNJ11 E23K did not predict failure to sulfonylurea
(42– 46). Moreover, a defect in insulin secretion was found treatment at 1 year. Subsequently, Sesti et al (54) reported
to occur in all euglycemic E23K carriers examined (45, that, compared with KCNJ11 E23E homozygotes, a
46). Interestingly, population studies revealed that the higher proportion of E23K carriers failed sulfonylurea-
KCNJ11 E23K variant has a strong allelic association with metformin combination therapy. They further showed
an ABCC8 coding variant S1369A. It is challenging to that islets isolated from E23K carriers displayed a com-
clearly decipher which of the two variants is the respon- promised insulin response to glibenclamide stimulation.
sible gene locus. Functional analyses suggested that a com-
bination of the two variants might underlie the increased 4. Molecular mechanisms and translational significance of
T2D risk (46, 47). Furthermore, impaired glucose toler- mutations or variations in KCNJ11 and ABCC8 and diabetes
ance (IGT) individuals with KCNJ11 E23K are not as well Mechanistically, KCNJ11 activating mutations associ-
protected by 1-year metformin treatment in comparison ated with PNDM decrease ATP sensitivity by reducing
with homozygous E23E IGT individuals who receive the ATP binding to Kir6.2 or prolonging the open state of the
same treatment. A similar relative resistance to met- channel and indirectly influencing ATP sensitivity (55)
formin-mediated protection was also observed in IGT pa- (Figure 3C). Most mutations identified to date lie either in
tients with ABCC8 S1369A. The authors speculated that the predicted ATP-binding site or in the regions associated
the diabetes-protective effect of metformin may interact with channel opening and closing. Activating mutations in
with the impact of these variants on insulin regulation the KCNJ11 gene decrease the sensitivity of the KATP
(47). In vitro experiments indicate that metformin can channel to ATP, thereby compromising the insulin secre-
confer a direct beneficial effect on pancreatic ␤-cells; this tory response. In contrast, mutations in the ABCC8 gene
first-line medication for T2D treatment was shown to re- enhance the binding of Mg-nucleotide to nucleotide-bind-
store the secretory function in isolated islets that has been ing domains of SUR1 or alter the intrinsic gating, leading
impaired by chronic exposure to elevated free fatty acids to membrane hyperpolarization and decreased insulin se-
or glucose (48); it also appears to preserve ␤-cell viability cretion (56).
by inhibiting mitochondrial permeability transition (49). The details of the mechanism that underlies the asso-
Laukkanen et al (50) found that the silent 1273AGA ciation of common variants in the KCNJ11 and ABCC8
variant allele of the ABCC8 gene was associated with in- genes and T2D are still being worked out. KCNJ11 E23K
creased risk of T2D. Importantly, this silent polymor- is thought to exert divergent effects on activatory and in-
phism was in linkage disequilibrium with three promoter hibitory modulators. It facilitates nucleoside diphosphate-
polymorphisms (G-2886A, G-1561A, and A-1273G) of mediated channel opening, whereas it inhibits ATP-trig-
the ABCC8 gene; together they formed a high-risk hap- gered channel closure, synergistically causing overactivity
lotype conferring a significant 2-fold risk of T2D. Indi- of KATP channels in the pancreatic ␤-cell and inhibition of
viduals with this high-risk haplotype and the KCNJ11 insulin release, predisposing to T2D (57). Compared to
E23K allele displayed a 6-fold risk for the conversion to the wild-type allele, E23K also displays significantly en-
diabetes in IGT patients, whereas the KCNJ11 E23K vari- hanced KATP activity in response to long-chain acyl coen-
zyme A, leading to impaired insulin and glucagon-like ferent populations (6). An inactivating mutation in GCK
peptide-1 (GLP-1) secretion and increased glucagon re- increases the glucose threshold needed to trigger insulin
lease (58). A recent report showed that carriers of the secretion, although insulin secretion remains regulated
KCNJ11 variant display reduced Ca2⫹ sensitivity of exo- with an elevated threshold that responds to postprandial
cytosis and decreased insulin release (59). As noted above, glucose excursions (69). Therefore, MODY2 patients, un-
the KCNJ11 E23K variant is strongly associated with like other forms of diabetes, present with mild fasting hy-
ABCC8 S1369A. The K23/A1369 variant was found to perglycemia and can be diagnosed at any age (70). Ho-
exhibit increased KATP channel MgATPase activity in mozygous inactivating GCK mutations lead to PNDM
comparison with the nonrisk E23/S1369 haplotype, pro- presenting at birth. Although over 600 mutations affecting
viding a plausible mechanism by which the E23K/S1369A different parts of the GCK gene have been reported, they
haplotype increases susceptibility to T2D (60). are all relatively rare mutations (66). The prevalence of

The identification of mutations of KCNJ11 and GCK-MODY was found to be 0.5–1% among women
ABCC8 in PNDM patients has important implications for with gestational diabetes mellitus (GDM) in recent studies
therapy because most of these patients (⬎90%) can be of multiethnic cohorts. It was proposed that a combina-
switched from insulin injections to oral sulfonylureas; pa- tion of body mass index ⬍25 kg/m2 and fasting glucose
tients with KATP channel defects respond to this class of ⱖ99 mg/dL differentiates GCK-MODY from common
agents, which enable insulin secretion by binding to the GDM (71, 72). Differentiating the two conditions may
ABCC8-encoded SUR1 subunit of the KATP channel to have therapeutic implications. In contrast to women with
effect channel closure independent of ATP (61) (Figure simple GDM, GCK-MODY mothers and their offspring
3D). An earlier age at initiation of sulfonylurea treatment have a low prevalence of diabetic complications despite
is associated with improved response to the therapy in lifetime hyperglycemia; they rarely require pharmacolog-
KCNJ11-related NDM patients (62). ical treatment outside pregnancy (70, 73). It is important
The therapeutic efficacy of oral sulfonylureas in T2D to note that activating GCK mutations are another cause
patients with KCNJ11 E23K mutation remains contro- of CHI (17, 66), which may presage late-onset diabetes.
versial. KCNJ11 E23K carriers were reported to display a GCK-CHI causes ␤-cell replication early on but subse-
marginally inferior therapeutic response in one study (63), quently leads to extensive ␤-cell apoptosis associated with
whereas another study found that they displayed a better DNA double-strand breaks and activation of the tumor
response to sulfonylureas (64). It is important to point out suppressor p53. The ␤-cell membrane depolarization as-
that the use of sulfonylureas did not increase the risk of sociated with KATP channel inactivation (like inactivating
hypoglycemia in KCNJ11 E23K carriers with T2D in a mutations of the KCNJ11 and ABCC8 genes) is thought to
third study (65). mediate this deleterious effect in GCK-CHI patients (74).
B. Glucokinase 2. Variations in the GCK gene and T2D

The metabolic enzyme GCK is predominantly ex- In a large-scale GWAS and cohort meta-analysis, a
pressed in hepatocytes and pancreatic ␤-cells. GCK cata- common variant rs1799884 G/A of GCK was found to be
lyzes the phosphorylation of glucose to glucose-6-phos- associated with glycated hemoglobin A1C levels, fasting
phate and is involved in the first step of both glycolysis and and 2-hour glucose concentrations, as well as T2D (75).
glycogen synthesis. In pancreatic ␤-cells, GCK is induced The GCK rs1799884 A allele was inconsistently asso-
by glucose and considered as the glucose sensor. Mu- ciated with altered fasting glucose in Chinese popula-
tations in GCK lead to reduced glycolysis and ATP lev- tions (76, 77). In another GWAS meta-analysis, GCK
els and impaired insulin secretion (5, 6). Homozygous rs4607517 A/G was confirmed to be associated with fast-
Gck-deficient mice die perinatally of severe hyperglyce- ing blood glucose levels and T2D (78). In Pima Indians, a
mia, whereas mutant Gck heterozygotes develop mild 3⬘ untranslated region (UTR) single nucleotide polymor-
early-onset diabetes resembling GCK-MODY in humans phism (SNP), chr7:44184184-G/A in the GCK gene, is
(66 – 68). associated with the rate of carbohydrate oxidation post-
absorptively, energy expenditure, and T2D risk, which is
1. GCK mutations underlie MODY2 in agreement with the role of GCK in hepatic glycolysis
Heterozygous inactivating mutations in GCK cause and energy metabolism. However, this variant was not
GCK-MODY (MODY2), which is characterized by mild associated with hepatic glucose production during insulin
fasting hyperglycemia, asymptomatic at birth and often stimulation, despite a previous report demonstrating that
detected later in life during routine screening. Mutations a GCK variant in the promoter region (-258 G/A) was
of GCK account for 30 –50% of cases of MODY in dif- associated with hepatic insulin resistance (79).
3. Molecular mechanisms that underlie the association of of 104 (5%) NDM patients with unknown etiology (ex-
variations in the GCK gene and T2D cluding the common genetic causes) were identified to
The activity of GCK can be regulated by glucose-6- carry homozygous SLC2A2 mutations. Four of these five
phosphatase, catalytic 2 (G6PC2) in the ␤-cell and GCK patients first presented with isolated diabetes and later de-
regulatory protein (GKRP, or GCKR) in the liver. In the veloped other features of FBS, suggesting that GLUT2 may
␤-cell, G6PC2 catalyzes glucose-6-phosphate dephos- play a role in human insulin secretion. Notably, some com-
phorylation, thereby functionally opposing the action of pound heterozygous mutations in the SLC2A2 gene, eg,
GCK. Genetic variants of either gene may affect the bal- c.457_462delCTTATA (p.153_4delLI) and c.1250C⬎G
ance between GCK and G6PC2 activities, resulting in (p.P417R), present an unusually mild clinical course, such as
impaired pulsatile insulin secretion, disrupting proper in- modest glycosuria and tubular proteinuria. Some hallmark
sulin signaling between the pancreas and insulin-sensitive clinical signs of FBS, like hepatomegaly and short stature, are

tissues, and predisposing to T2D development (80). often absent in these patients (91). Therefore, it would be
In the liver, the activity and subcellular localization of prudent to include the SLC2A2 gene for genetic screening in
GCK are allosterically regulated by GKRP. GCK is se- patients with isolated glycosuria.
questered to the nucleus as an inactive complex with
GKRP at low glucose concentrations. It is rapidly disso- 2. SLC2A2 gene mutations or variants and T2D
ciated from the GCK-GKRP complex and translocated to The SNP rs5400 in the SLC2A2 gene, which causes a
the cytoplasm at high glucose concentrations, which trig- GLUT2 T110I substitution, was first reported to be asso-
gers glucose disposal (81, 82). It has been proposed that ciated with T2D in Pima Native Americans (92); subse-
SNP rs780094 at GKRP may perturb the GKRP/GCK quently, other SLC2A2 gene variants and their association
system and thus modulate the effect of SNP rs1799884 at with T2D have been studied extensively, although con-
GCK, resulting in altered glucose metabolism (77). Addi- flicting conclusions remain. To date, variants of the
tional functional studies are needed to prove or disprove SLC2A2 gene are associated with increased risk for T2D
this hypothesis. in some (93, 94), but not most (95–98), studies. In the
Finnish Diabetes Prevention Study, SNPs rs5393 (AA) and
C. Solute carrier family 2, member 2 rs5394 (CC) in the promoter region as well as SNPs rs5400
The SLC2A2 gene encodes glucose transporter (GLUT) (T110I) and rs5404 (GG, T198T) in the exon of the
2, a high-capacity facilitative glucose transporter ex- SLC2A2 gene have been identified to predict the conver-
pressed predominantly in the liver, kidney, intestine, and sion to diabetes in obese subjects with IGT, and rs5393
pancreatic ␤-cells. Ablation of Slc2a2 in mice leads to se- (AA) increased the risk of T2D by 3-fold. Interestingly, the
vere diabetes and lethality usually at 2–3 weeks of age rs3758947 GG of ABCC8 presents an additive effect with
because of an insulin secretion defect (83). It is well es- the rs5393 AA of SLC2A2 to predict the conversion to
tablished that GLUT2 is the predominant glucose trans- diabetes (99), indicating that individuals with risk SNPs in
porter in rodent ␤-cells. A recent study suggested that the network of genes regulating insulin secretion (SLC2A2
GLUT2 may also mediate insulin secretion in human and ABCC8) are particularly at high risk for T2D.
␤-cells (84), although GLUT1, not GLUT2, is thought to Kilpeläinen et al (100) found that moderate-to-vigorous
be the major glucose transporter in human ␤-cells (85). physical activity could lower the risk of conversion to T2D in
the carriers of the common homozygous genotype of rs5393
1. Mutations in SLC2A2 lead to Fanconi-Bickel syndrome (AA), rs5394 (CC), or rs5404 (GG, T198T) of SLC2A2 and
Homozygous patients with inactivating mutations of rs3758947 (GG) of ABCC8, an observation that implies that
SLC2A2 develop Fanconi-Bickel syndrome (FBS; also improvements in insulin sensitivity by lifestyle changes help
known as glycogen storage disease type XI), characterized preserve the insulin secretory function of ␤-cells.
by a renal Fanconi syndrome with glycosuria, galactos-
uria, aminoaciduria, proteinuria and phosphaturia, short 3. Molecular mechanisms that underlie the association of
stature, rickets, poor growth, hepatomegaly, and galac- mutations or variations in the SLC2A2 gene and diabetes
tose intolerance (86). Some FBS patients may present with In vitro experiments conducted by Michau et al (101)
postprandial hyperglycemia and diabetes or IGT. A large revealed that none of four select FBS-associated SLC2A2
number of homozygous mutations (84, 87–90) or com- mutants could transport glucose to a significant level. The
pound heterozygous mutations (91) in the SLC2A2 gene G20D and S242R substitution mutants are associated
have been identified to underlie FBS. with loss of protein expression at the plasma membrane,
Mutations in SLC2A2 were recently noted to be asso- whereas the P417L and W444R mutants lose their trans-
ciated with defective insulin secretion in humans (84). Five port capacity despite adequate membrane targeting, indi-
cating the crucial role of the two amino acids (417 Pro and nephric tubules of the kidney, and the developing pan-
444 Trp) for GLUT2 transport function in humans (101). creas, stomach, and intestine (110). Hnf4a-deficient mice
A recent report showed no difference in insulin secre- display defective gastrulation and die around e9 (111). In
tion during an oral glucose tolerance test in individuals hepatocytes, HNF4A controls the transcription of
with or without the risk allele at SLC2A2 rs5400 (T110I), HNF1A, whereas in pancreatic islets and exocrine cells,
suggesting that neither ␤-cell mass nor insulin secretion is HNF1A controls the transcription of HNF4A through an
significantly affected (102). In agreement with this, this alternate tissue-specific HNF4A promoter (112).
GLUT2 T110I variant displays similar kinetics for glucose
transport as wild-type GLUT2 (101). Additionally, this 1. Mutations in HNF1A and HNF4A lead to MODY3 and
variant is in strong linkage disequilibrium with two other MODY1, respectively
variants (SNPs rs5406 and rs6785803) located in the pro- Mutations in the HNF4A gene lead to the first subtype

moter region of the SLC2A2 gene (103), which raises the of MODY identified (HNF4A-MODY [MODY1]),
possibility that the functional consequence of the rs5400 whereas mutations in the HNF1A gene cause HNF1A-
(T110I) variant could be related to the difference in the MODY (MODY3). These two types of MODY display
transcription of the SLC2A2 gene, rather than the effect of very similar clinical characteristics. The patients develop
the gene product of variant rs5400, ie, GLUT2 T110I it- young-onset diabetes in the second to fourth decade with-
self. Interestingly, the variant rs5400 (T110I) is associated out features of insulin resistance or ␤-cell autoimmunity.
with habitual consumption of sugars (104), which may Pancreatic ␤-cell dysfunction can be observed before the
reflect glucose sensing in the portal vein (105) and/or the onset of diabetes in the carriers with HNF1A mutations.
hypothalamus (106), the two organs that also harbor Glycosuria due to a low renal threshold for glucose is
GLUT2-dependent glucose sensors. Although GLUT2 has observed in HNF1A-MODY patients, and in nondiabetic
been identified in the paraventricular nucleus, lateral hy- HNF1A mutant carriers, but not in HNF4A-MODY pa-
pothalamic area, and arcuate nucleus, it is unclear whether tients (113). Interestingly, inactivating mutations in
the high Km of GLUT2 would allow the putative trans- HNF1A and HNF4A can cause CHI associated with in-
porter to function normally in these regions of the brain creased body weight in early life, which may improve with
with normal blood-brain barrier because of the local low age (114). Both HNF1A and HNF4A MODY patients
glucose concentrations in these brain regions (104). In respond to low-dose sulfonylureas (12.5% or less of the
contrast, GLUT2 is likely operative in parts of the brain maximum licensed dose) (5, 6).
that lack a blood-brain barrier, ie, where the transporter Mutations in HNF1A (MODY3) are the most common
is exposed to physiological levels of circulating blood glu- cause of MODY, which underlie 52% of monogenic di-
cose. Indeed, the circumventricular organ, area postrema, abetes in a large UK series (115). Interestingly, although
is one such brain area that has also been found to harbor patients display inherent ␤-cell defects, they develop dia-
SLC2A2 mRNA (107). These observations suggest that a betes relatively late in life— 63% at age 25 years and 94%
GLUT2-mediated glucose-sensing mechanism could reg- at age 50 (116). The reason that MODY3 presents rela-
ulate food intake in select parts of the brain and thus be tively late in life is caused by the capacity of HNF1A to
associated with the risk of T2D. control the transcription of sodium-glucose cotransporter
2 (SGLT2), a renal glycosuria threshold-controlling gene.
D. HNF1A and HNF4A Therefore, individuals with HNF1A mutations display a
HNF1A, also known as transcription factor 1 (TCF1), drastically decreased expression of SGLT2, thereby reduc-
is first expressed in the yolk sac and in the liver bud. Later ing glucose reabsorption through the proximal tubule
on to adulthood, HNF1A is expressed throughout or- (117). This defective HNF1A-mediated down-regulation
ganogenesis of the liver, kidney and pancreas. Hnf1a-de- of SGLT2 helps maintain relative euglycemia in MODY3
ficient mice develop hepatic dysfunction, phenylketonu- individuals before the appearance of overt hyperglycemia
ria, and renal Fanconi syndrome. A fraction of the mice and diagnosis of diabetes much later in life; ie, HNF1A
lacking Hnf1a die shortly after weaning due to a progres- mutations function like SGLT2 inhibitors, and the belated
sive wasting syndrome (108). Mice with Hnf1a ablation appearance of overt diabetes in affected individuals could
also develop diabetes with a profound defect in glucose or have predicted the development of this new class of hy-
arginine-stimulated insulin secretion (109). poglycemic agent for the treatment of diabetes.
HNF4A is first expressed in the primary endoderm in
mice at embryonic day 4.5 (e4.5) and is restricted to the 2. HNF1A gene variants and T2D
visceral endoderm from e5.5 to e8.5. Thereafter, it is ex- A private mutation (G319S) in HNF1A has been re-
pressed in the liver diverticulum, the hindgut, the meso- ported in Canadian Oji-Cree populations to be associated
with early-onset T2D, lower body mass, and a higher post- a ghrelin receptor antagonist restored glucose homeostasis
challenge plasma glucose (118). Using whole-exome se- in Hnf1␣-null mice (127). Ghrelin suppresses glucose-in-
quencing, MacArthur and colleagues (119) identified a duced insulin secretion, but not basal insulin secretion. It
single low-frequency variant (c.1522G⬎A [p.E508K]) in counteracts cAMP-protein kinase A signaling, a well-es-
HNF1A, which is associated with T2D prevalence in La- tablished pathway of GSIS, by interacting with the GH
tino populations. The variant allele p.E508K is located in secretagogue receptor GHS-R1a on ␤-cells. The suppres-
the HNF1A transactivation domain. Functional studies sion of the cAMP pathway results in activation of delayed
revealed that, despite the preservation of its DNA binding outward K⫹ (Kv) channels and inhibition of Ca2⫹ signal-
activity, the p.E508K variant protein displays compro- ing, causing decreased insulin release (128).
mised transactivation activities on the promoters of In mammals, the HNF4A gene encodes up to nine dis-
SLC2A2 and HNF4A in ␤-cells. Of note, a recent report tinct isoforms, resulting from the use of alternate promot-

showed that HNF1A variant (rs2650000) predisposes ers and alternative splicing. Promoter P1 initiates tran-
to latent autoimmune diabetes in adults and adult-onset scripts that contain exon 1A (isoforms A1–A6) encoding
T1D (120). activation function 1 (AF-1) domain, whereas promoter
P2 initiates transcripts that contain exon 1D (isoforms
3. HNF4A gene variants and T2D A7–A9) lacking AF-1. In pancreatic ␤-cells, the P1 pro-
Multiple SNPs in the alternate upstream promoter (P2) moter exhibits stronger transcriptional activities than the
region of HNF4A have been found to be associated with P2 promoter (129). Pancreatic ␤-cell-specific Hnf4a dele-
increased risk of T2D. Initial studies showed that SNPs at tion leads to hyperinsulinemia in fasted and fed mice but
rs2144908, rs3818247, and rs884614 in the P2 region of paradoxically is also associated with IGT. Hnf4a mutant
the HNF4A gene were associated with T2D risk in Finnish ␤-cells exhibit a defective response to stimulation by glu-
(121) and Ashkenazi (122) populations. Subsequently, cose and sulfonylureas. Experiments in mice showed that
Weedon et al (123) reported that the risk haplotype of the expression of the potassium channel subunit Kir6.2
rs4810424 and rs2144908 in the P2 region of the HNF4A (encoded by KCNJ11), a downstream target of Hnf4a, is
gene was associated with T2D in UK populations in both reduced by 60% in Hnf4a mutant ␤-cells (130).
case-control and family-based studies. Recently, Barroso Studies in rodents further showed that both maternal
et al (124) found that HNF4A P2 SNPs rs1884613 and suboptimal nutrition and aging lead to progressive epige-
rs2144908 were associated with T2D in the Ashkenazi, netic silencing at the enhancer region in the Hnf4a gene,
but not in UK, populations. These findings suggest that P2 disrupting the P2 promoter-enhancer interaction and re-
SNPs in the HNF4A gene confer varying risk effects in sulting in a permanent reduction in Hnf4a expression
different populations. (131). It is unclear whether the variants in HNF4A pre-
dispose to T2D via a similar mechanism in people.
4. Molecular mechanisms and pathophysiology that under-
lie the association of mutations in the HNF1A and HNF4A E. Hepatocyte nuclear factor 1 homeobox B
genes and diabetes HNF1B, also known as transcription factor 2 (TCF2),
To dissect the molecular mechanisms underlying is first expressed in the visceral endoderm at the onset of
MODY3, Bjorkhaug et al (125) investigated the func- gastrulation. Later on to adulthood, it is expressed throughout
tional properties of HNF1A mutations. They reported organogenesis of the liver, kidney and pancreas (132). In
that two mutant HNF1A proteins (R171X and R263C) mice, HNF1B-positive progenitors were shown to be
manifest reduced or absent DNA binding activity. Five equally capable of giving rise to pancreatic duct, acinar,
mutants (R131W, R171X, P379fsdelCT, S445fsdelAG, and endocrine cells from e11.5 to e13.5. Global HNF1␤-
and Q466X) display defective nuclear translocation. Most deficient mice die soon after implantation (132, 133). Pan-
of the MODY3-associated mutants exhibit reduced trans- creatic ␤-cell-specific (RIP-Cre) Hnf1b deletion resulted
activational activity. These loss-of-function HNF1A mu- in IGT with reduced insulin secretion (134). HNF1B is
tants underlie the ␤-cell dysfunction in MODY3 (125). required for the proliferation and survival of pancreatic
Recently, Nowak et al (126) reported that circulating multipotent progenitor cells. Hnf1b deficiency in pancre-
ghrelin levels were higher in HNF1A-MODY patients atic progenitors leads to severe pancreatic hypoplasia and
than those in common T1D or T2D patients. Hnf1a-de- perinatal lethality (135). HNF1B is also a key regulator of
ficient mice develop diabetes caused by a defect in glucose- duct morphogenesis, acinar cell differentiation, and main-
stimulated insulin secretion (GSIS) (109). It is intriguing tenance of acinar cell identity. It is indispensable for the
that a recent study showed that serum ghrelin levels were generation of endocrine precursors, likely by directly reg-
also increased in Hnf1a-deficient mice, and treatment with ulating the pancreatic islet lineage-defining transcription
factor neurogenin 3 (Neurog3, or Ngn3) expression (135). with underdeveloped and disorganized acini in human fe-
The sequential activation of Hnf1b, Hnf6, and Pdx1 is tuses. The HNF1B R112fsdel mutation causes a frame-
believed to direct the differentiation of endodermal cells shift and a truncated protein lacking part of the POU-
into pancreatic progenitors (136). specific domain (POUS), which is essential for DNA
binding specificity, and the entire POU homeodomain
1. Mutations in HNF1B lead to RCAD syndrome, or MODY5 (POUH), as well as the C-terminal transactivation domain,
In addition to causing PNDM or TNDM in very rare whereas the HNF1B P472fsins mutation leads to a pre-
cases (137), heterozygous mutations in HNF1B cause re- mature stop codon and the insertion of 35 novel amino
nal cysts and diabetes (RCAD syndrome, or MODY5). acids at the C-terminus of the transactivation domain.
HNF1B whole-gene deletions account for approximately Functional studies showed that the truncated R112fsdel
50% of MODY5 cases. To date, more than 50 missense protein, retaining the N-terminal dimerization domain,

mutations or deletions have been identified in MODY5 functions as a weak dominant-negative mutant, through
patients. Various phenotypes related to these mutations the formation of nonfunctional heterodimers with wild-
have been described, including urogenital malformations, type HNF1B protein. In contrast, the P472fsins mutant
hyperuricemia, young-onset gout, and pancreatic hyp- exhibits a decreased transactivation capacity, whereas it
oplasia or atrophy. In a literature meta-analysis, renal retains the normal DNA binding activity; however, it does
cysts and diabetes occurred in 89.6 and 45% of the pa- not interfere with the function of the wild-type HNF1B
tients with RCAD syndrome, respectively. However, the protein (146).
concurrence of renal cysts and diabetes was observed in GLUT2, a potential direct target of HNF1B (147), was
only 27.5% of the patients, which could delay the clinical undetectable in islet ␤-cells of human fetuses with
diagnosis and timely genetic testing (138). Patients with R112fsdel or P472fsins mutation (146). Similarly, over-
HNF1B mutations display abnormal ␤-cell development expression of the HNF1B P159L mutant (disrupting its
and reduced insulin section (135, 136) that is usually re- DNA binding domain) led to significantly reduced expres-
sistant to sulfonylureas; most of them require early insulin sion of GLUT2 in Min6 mouse insulinoma cells (148). As
therapy (139, 140). discussed earlier, Slc2a2-deficient mice develop severe di-
abetes and lethality with impaired glucose sensing by pan-
2. HNF1B gene variants and T2D creatic ␤-cells (83). Forced overexpression of a naturally
Winckler et al (141) found that SNP rs757210 A in occurring dominant-negative form of human HNF1A
intron 2 of HNF1B showed a strong association with T2D (P291fsinsC) disrupts the formation of homo- and het-
in multiple populations. Epidemiological studies have erodimers between wild-type HNF1A and HNF1B, lead-
demonstrated an inverse association between T2D and the ing to reduced expression of GLUT2 and E-cadherin
risk of prostate cancer (142, 143). In accordance with this (149). In addition, ␤-catenin and E-cadherin expression
finding, Gudmundsson et al (144) found that a common was drastically reduced in the islet and acinar compart-
variant rs4430796 A in HNF1B was positively associated ments of human fetuses with R112fsdel or P472fsins mu-
with prostate cancer and negatively associated with T2D tation. Wnt/␤-catenin is required for acinar lineage spec-
risk. It has been proposed that with time, metabolic status ification and differentiation (150 –152). These findings
in T2D patients might switch from hyperinsulinemia to suggest that perturbation in the Wnt/␤-catenin signaling
endogenous insulin deficiency, thus alleviating the pro- pathway may underlie the severe pancreas hypoplasia fre-
oncogenic effect of insulin in the prostate. quently observed in HNF1B mutant individuals; early loss
of GLUT2 in ␤-cells may cause early-onset diabetes in
3. Molecular mechanisms and pathophysiology that under- MODY5 patients (146).
lie the association of mutations in the HNF1B gene
and diabetes F. Pancreatic and duodenal homeobox 1
HNF1A and HNF1B share a highly conserved DNA- PDX1 (also known as insulin promoter factor 1 [IPF1])
binding domain and a more divergent C-terminal trans- is a homeodomain transcription factor expressed through-
activation domain. HNF1B may act either as a homodimer out the pancreatic epithelium at e8.5 in mice; from e15.5
or as a heterodimer with HNF1A (145). Diverse mecha- onward, its expression becomes mainly restricted to
nisms have been uncovered that may underlie the devel- ␤-cells and a small subpopulation of ␦ and PP (pancreatic
opment of HNF1B-MODY. For example, frameshift mu- polypeptide) cells (153). PDX1 is essential for pancreas
tations R112fsdel (an 8-bp deletion in exon 2) or development, ␤-cell differentiation, and the maintenance
P472fsins (an 8-bp insertion in the exon 7) in the HNF1B of mature ␤-cell function by regulating the expression of
gene were found to lead to severe pancreatic hypoplasia key islet-specific genes such as Ngn3, insulin, Gck, and
Slc2a2 (153–155). Mice lacking Pdx1 fail to form a pan- variant (p.Gly218Alafs*12) in PDX1 is associated with
creas (156). Pancreatic ␤-cell-specific Pdx1 deficiency in high risk of T2D (167). Furthermore, the same study ob-
mice leads to reduced expression of insulin and Slc2a2, served seven previously reported variants (six missense
causing maturity-onset diabetes (155). A large fraction of variants and one in-frame insertion) in PDX1, but none
Pdx1-deleted ␤-cells rapidly acquires ultrastructural and were associated with diabetes (167). In addition, PDX1
physiological features of ␣-cells, suggesting that Pdx1 variants have also been linked to KPD (see Section II.G for
maintains ␤-cell identity and function via repressing an details).
␣-cell lineage (157). Mutations or variants in the PDX1
gene cause PNDM and MODY4 and are also linked to 3. Molecular mechanisms that underlie the association of
common T2D and ketosis-prone diabetes (KPD). mutations in the PDX1 gene and diabetes
As noted above, PDX1 variants associated with en-

1. PDX1 mutations underlie PNDM and MODY4 hanced T2D risk are rare, despite the fact that PDX1 mu-
Stoffers et al (158) first reported that a homozygous tations cause PNDM and MODY4. Pro63fsdelC point
single cytosine deletion within codon 63 (Pro63fsdelC) of deletion in PDX1 leads to a frame shift, causing a prema-
the human PDX1 gene was responsible for a previously ture truncated protein with 59 aberrant codons that is
described PNDM syndrome that is also associated with lacking transactivation domain at its C terminus. The mu-
pancreatic exocrine insufficiency (159). Two substitution tant protein fails to transactivate the INS gene transcrip-
mutants, E164D and E178K in the human PDX1 gene, tion, leading to PNDM (158). The C-terminal domain of
also individually lead to PNDM and pancreatic exocrine PDX1 is required to maintain all endocrine lineages in the
insufficiency (160), whereas the hypomorphic mutation pancreas. Expression of a single C-terminal truncated
(E178G) in PDX1 is responsible for PNDM but is not form of PDX1 results in mild hyperglycemia at birth and
associated with clinical signs of exocrine insufficiency progresses to overt diabetes by the down-regulation of
(161). In the first family pedigree, Stoffers et al (162) fur- Ngn3 (154). Schwitzgebel et al (160) reported that PDX1
ther identified that the heterozygous Pro63fsdelC muta- E178K displayed reduced transactivation activity on the
tion in PDX1 causes MODY4. Family members devel- INS promoter due to a decrease in PDX1 steady-state pro-
oped diabetes in six generations, with an average age at tein levels resulting from impaired protein stability,
onset of 35 years (range, 17 to 67 years). Six of eight
whereas substitution variant PDX1 E178G displays re-
affected heterozygotes, lacking ketosis or other indica-
duced transactivation activity via an unclear mechanism,
tions of severe insulin deficiency, were treated with diet or
leading to PNDM (161).
oral hypoglycemic agents. Notably, patients afflicted with
A number of PDX1 mutations (eg, C18R, Q59L,
PDX1 mutations (eg, D76N) may concurrently carry
D76N, R197H, and InsCCG243) lead to reduced binding
other MODY gene mutations (eg, S315fsinsA in HNF1␣/
of the mutant protein to the INS gene promoter and
MODY3). Not surprisingly, carriers of mutations in both
decreased INS gene transcription in response to hyper-
genes (PDX1 and HNF1␣) display more severe insulin
glycemia in pancreatic ␤-cells, which may underlie the
deficiency than those who carry the mutation in only one
substantial defect in insulin secretion detected in these pa-
of the two genes (163).
tients (163–165). D76N and Q59L mutations in PDX1
2. Mutations or variations in the PDX1 gene and T2D have also been linked to decreased GSIS in nondiabetic
Shortly after PDX1 mutations were identified in pa- subjects (164).
tients with PNDM and MODY4, other mutations in the In mice, Pdx1 haplodeficiency enhances ␤-cell suscep-
PDX1 gene were also found to be linked to common T2D. tibility to endoplasmic reticulum (ER) stress-associated
Candidate gene association studies showed that PDX1 apoptosis and leads to a failure of ␤-cell compensation for
substitution mutations (C18R, Q59L, D76N, R197H, high-fat diet (HFD)-induced insulin resistance (168).
G212R, and P239Q) and an in-frame proline insertion Mitochondrial dysfunction is an important contributor to
(InsCCG243) predispose to common T2D (163–165). Of ␤-cell failure, insulin resistance, and diabetes (169, 170).
note, approximately 5% of the examined T2D patients Mitochondrial autophagy or mitophagy is required for the
and control nondiabetic individuals carried low-fre- selective elimination of dysfunctional or damaged mito-
quency PDX1 variants that were not associated with di- chondria to maintain cellular respiration. Recently, Pdx1
abetic phenotype, suggesting that PDX1 mutations or has been shown to direct autophagosome-lysosome fusion
variants are a very rare cause of diabetes (166). Consistent during mitophagy and regulate mitochondrial function in
with this observation, whole-genome sequencing of 2630 pancreatic ␤-cells (171). Although it has not been directly
Icelanders showed that a novel rare (0.20%) frameshift examined, it is possible that these mechanisms may also
underlie the association of PDX1 mutations/variants with lated with a positive therapeutic response to rosiglitazone,
diabetes. but not repaglinide. PAX4 variant rs6467136 GA⫹AA
carriers exhibited greater responses to rosiglitazone ther-
G. Paired box 4 apy than GG homozygotes, with greater decrement of
PAX4, a paired-homeodomain transcription factor, 2-hour glucose level and amelioration of insulin resistance
functions mainly as a transcription repressor (172). Pax4 as well as higher cumulative attainment rates of target
mRNA is initially detected in the ventral spinal cord and fasting and 2-hour glucose levels (182).
pancreatic bud at e9.5 in mice; its expression is later re- In addition to the association of PAX4 variants with
stricted to pancreatic ␤- and ␦-cells, reaching a maximal T2D, one candidate gene association study (183) and a
level at e13.5 to e15.5, and thereafter declining to low linkage analysis (184) reported that PAX4 variants in-
levels. PAX4 is also found to be expressed in adult rodent cluding ⫹1,168 C/A of PAX4 (rs712701) were associated

and human islets (173). Pax4-deficient mice develop se- with T1D, but other studies failed to find an association
vere hyperglycemia and die within 2 days of birth. The (185, 186), which suggests that PAX4 variants are un-
␤-cells are replaced by a proportional increase in gluca- likely to play an important role in the development of
gon-positive ␣-cells; somatostatin-expressing ␦-cells are T1D.
also missing in Pax4-deficient mice. Thus, Pax4 deficiency Mauvais-Jarvis et al (187) first reported that PAX4 ho-
is characterized by a favoring of an ␣-cell fate at the ex- mozygous R133W and heterozygous R37W mutations
pense of ␤-/␦-cell lineages (174). The aristaless-related ho- may be associated with KPD. KPD is a heterogeneous
meobox-encoding gene, Arx, and Pax4 are coexpressed in group of atypical diabetic syndromes characterized by re-
the same proendocrine cell during pancreas genesis. Arx- current severe ␤-cell dysfunction and a variable clinical
deficient mice develop early-onset hypoglycemia, dehy- course (188, 189). KPD patients can be classified into four
dration, and die 2 days after birth. In the absence of Arx, subgroups, as defined by: 1) the presence or absence of
endocrine progenitor cells assume ␤-/␦-cell destinies autoantibodies (that are associated with common T1D),
(175). At the molecular level, ARX and PAX4 directly ie, A⫹ or A⫺; and 2) the presence or absence of ␤-cell
bind to the respective promoter or enhancer and drasti- function, ie, ␤⫹ or ␤⫺. The A-␤ subgroup of KPD pa-
cally repress each other’s transcription (176). tients, characterized by ␤-cell failure without evidence of
autoimmunity, often has a strong family history of diabe-
1. Mutations in PAX4 lead to MODY9 tes (188, 189). Balasubramanyam and colleagues (190)
To determine whether PAX4 mutations contribute to systematically sequenced seven genes associated with
MODY in the Thai population, Plengvidhya et al (177) monogenic diabetes in an A-␤ patient subgroup and found
screened PAX4 coding sequences in 46 MODY probands that about 30% of them harbor variants of HNF1A,
who did not have mutations in other known MODY PDX1, or PAX4 genes, suggesting that these variants may
genes. They identified two possible pathogenic mutations underlie the ␤-cell dysfunction in a fraction of patients
of PAX4, R164W and IVS7–1G⬎A (a G to A substitution with A-␤ KPD.
at splice acceptor of intron 7), in patients with MODY
(PAX4-MODY, or MODY9, a very rare type of MODY), 3. Molecular mechanisms and pathophysiology that under-
but not in nondiabetic controls and healthy subjects. Sub- lie the association of mutations in the PAX4 gene
sequently, PAX4 R192H (178) and a 39-base heterozy- and diabetes
gous deletion in exon 3 (C.374 – 412 del 39) in PAX4 (179) Our knowledge of the molecular basis of PAX4 muta-
have also been reported in MODY9 patients. tions causing diabetes remains incomplete. PAX4 pre-
dominantly represses the glucagon promoter activity in
2. PAX4 gene mutations or variants and T2D and KPD ␣-cells and weakly inhibits insulin promoter activity in
In a Japanese population, a missense mutation of PAX4 ␤-cells (172). PAX4 mutations are located either in the
R121W was found to be associated with T2D. The paired domain (R37W), in the homeodomain (R164W),
R121W mutation was located in the C terminus of the or between the paired domain and homeodomain
paired domain and was thought to affect its transcrip- (R133W). These mutations impair its transcriptional re-
tional activity due to a lack of DNA binding (180). In a pressor activity by interfering with its capacity to bind
GWAS meta-analysis, Ma et al (181) reported that the DNA, or the binding specificity or functional stability, or
rs10229583 G risk variant near PAX4 was associated its interaction with other proteins. PAX4 mutants fail to
with elevated fasting glucose, impaired ␤-cell function, repress glucagon transcription, leading to hypergluca-
and early-onset T2D in a Chinese population. In newly gonemia (187). PAX4 mutations may cause insulin defi-
diagnosed T2D patients, PAX4 variant rs6467136 corre- ciency, possibly through the disruption of ␤-cell develop-
ment or impairment of GSIS in ␤-cells. Some PAX4 shift to produce a premature truncated protein lacking the
mutations (eg, R121W, IVS7-1G⬎A) seem to increase the activation domain at the C terminus.
susceptibility of ␤-cells to apoptosis upon cytokine or high
glucose exposure (177, 191). Conversely, forced expres- 2. NEUROD1 gene variants and T2D
sion of wild-type PAX4 has been shown to protect cyto- Malecki et al (204) reported that a missense mutation
kine-induced ␤-cell death in isolated human islets (191). In R111L and a cytosine residue insertion (206 ⫹C) in the
general, most MODY patients are treated with sulfonyl- NEUROD1 gene were associated with the risk of T2D. A
ureas, metformin, or insulin (5). A few recent studies have number of studies showed that a common A45T variant at
also documented the efficacy of GLP-1 receptor agonists rs1801262 in NEUROD1 was inconsistently associated
(192, 193) and dipeptidyl peptidase-IV inhibitors (194, with T2D (204 –208). Recent meta-analyses revealed that
195) in these patients. Although it has not been directly this polymorphism is associated with T2D in European

studied, it is possible that the GLP-1 receptor agonists and descent Caucasians, but not in individuals of Chinese and
dipeptidyl peptidase-IV inhibitors, through their gluca- Japanese descent, or others of East Asian descent (207). Of
note, the A45T variant in NEUROD1 is associated with
gon-inhibiting actions (196), are especially efficacious in
T1D in several small population studies (208 –210), but
the treatment of patients with PAX4 mutations.
not in a larger population study (211).
H. NEUROD1/BETA2
3. Molecular mechanisms that underlie mutations in the
The basic helix-loop-helix (HLH) transcription fac-
NEUROD1 gene and diabetes
tor NEUROD1, also called BETA2, plays a crucial role
NEUROD1 forms a heterodimer with the ubiquitous
in pancreatic ␤-cell maturation and maintenance.
HLH protein E47 to transactivate insulin gene expression
NEUROD1 is detected in the mouse embryo as early as
by binding to a critical E-box motif on the promoter (212),
e9.5 in glucagon-positive cells and is mostly restricted to
synergistically with MAFA, PDX1 (213), and GLIS3
␤-cells at birth. Neurod1-deficient mice fail to develop
(214). The mutation R111L in the NEUROD1 gene is
mature islets due to enhanced apoptosis, leading to NDM
located in the proximal basic portion of the basic HLH
and lethality within the first few days of life (197). Pan- domain, which mediates DNA binding. This mutation
creatic ␤-cell-specific Neurod1-deficient mice exhibit abolishes E-box binding activity of NEUROD1 and dras-
severe glucose intolerance with greatly reduced insulin se- tically compromises insulin gene transcription in pancre-
cretion. Islets lacking Neurod1 respond poorly to glucose atic ␤-cells. The 206 ⫹C insert mutation produces a frame-
and display a glucose metabolic profile similar to imma- shift and synthesis of a nonsense peptide from amino acid
ture ␤-cells, characterized by increased expression of gly- 205 to 242, followed by a stop codon. The mutant peptide
colytic genes, elevated levels of lactate dehydrogenase and has lost its C-terminal transactivation domain, which nor-
basal oxygen consumption, and overexpression of neuro- mally enables interaction with the coactivators cAMP re-
peptide Y. In sum, NEUROD1 is required for the devel- sponse element-binding protein and p300. Thus, these
opment and maintenance of fully functional mature NEUROD1 mutants impair DNA binding and/or the
␤-cells (198). transactivation activity on the INS gene as well as other
target genes in pancreatic ␤-cells leading to T2D (204).
1. Mutations in the NEUROD1 gene lead to MODY6
and PNDM I. Wolfram syndrome 1
Heterozygous loss-of-function mutations in NEUROD1 WFS1 encodes a 100-kDa transmembrane glycopro-
produce a very rare type of MODY (NEUROD1-MODY, tein that is localized in the ER and secretory granules.
or MODY6), with six families reported to date. A missense Unfolded protein response (UPR) is a cellular stress re-
mutation E110K in NEUROD1 was first reported in an sponse induced by the accumulation of unfolded or mis-
Icelandic MODY6 family (199). Subsequently, missense folded proteins within the ER lumen. WFS1 is a compo-
mutations S159P, H241Q, and R103P in NEUROD1 nent of the inositol requiring 1 and PKR-like ER kinase
were identified in Chinese (200), Czech (201), and Polish UPR signaling pathway and is important in the mainte-
(202) MODY6 families, respectively. nance of ER homeostasis. The maintenance of ER integrity
Recently, Rubio-Cabezas et al (203) reported that two by PKR-like ER kinase is crucial to normal ␤-cell function
recessive homozygous mutations in NEUROD1, a single because large amounts of insulin are produced by these
base pair duplication (c.364dupG) and a 2-bp CT deletion cells (215). WFS1 has been implicated in the activation of
(c.427_428del), led to a syndrome of PNDM and neuro- the UPR to mitigate ER stress (216). This ER protein is
logical abnormalities. Both mutations introduced a frame- induced by ER stress and during glucose or potassium
chloride-stimulated insulin secretion in ␤-cells. Wfs1-de- risk in UK and other populations. The frequency at the
ficient ␤-cells display increased susceptibility to ER stress- rs10010131 risk allele is 60%, and it was estimated that
mediated apoptosis (217). the population attributable fraction of rs10010131 is 9%,
explaining 0.3% of the excess familial risk of T2D.
1. Mutations in WFS1 lead to Wolfram syndrome 1 Cheurfa et al (225) examined the association of allelic
Recessive mutations in the WFS1 gene cause pancreatic variations in the WFS1 gene with insulin secretion and risk
␤-cell death, resulting in a monogenic form of diabetes of T2D in the prospective Data from Epidemiological
known as Wolfram syndrome 1 (WFS1), which is some- Study on the Insulin Resistance Syndrome (DESIR) study.
times referred to as DIDMOAD (diabetes insipidus, They found that the three variants rs10010131 G,
diabetes mellitus, optic atrophy, and deafness). It is note- 1801213 G, and rs734312 A were significantly associated
worthy that, in certain populations, eg, a Lebanese pop- with plasma glucose, glycated hemoglobin A1C levels, and

ulation, patients with some WFS1 mutations may present insulin secretion at baseline and throughout the study in
with isolated early-onset diabetes, contributing to a sub- individuals with T2D or at risk of developing diabetes, but
stantial proportion of patients who are often mislabeled as not in those who remained euglycemic at the end of the
having “T1D” (218). There is currently no effective treat- follow-up. Carriers of the combined GGA haplotype for
ment for the syndrome, which usually ends in premature these SNPs had a 26% increased risk for developing dia-
death. With respect to the optic atrophy, a recent report betes compared with carriers of the ACG haplotype. These
showed that average retinal thickness is significantly associations between the three variants and T2D were rep-
lower in WFS1 patients, compared with age-matched T1D licated in a large case-control, cross-sectional study. These
patients and healthy individuals. Moreover, a strong neg- findings suggested that variants of the WFS1 gene affect
ative correlation exists between average retinal thickness insulin secretion and are associated with an increased
and both age at examination and duration of diabetes in risk of T2D. However, variants of WFS1 including
WFS1 patients. It has therefore been proposed that retinal rs10010131G were found not to be associated with T2D
thinning is a marker of disease progression in these pa- in a well-designed fine-mapping study, raising the argu-
tients (219). To date, over 180 mutations in the WFS1 ment that low frequency variants in WFS1 may or may
gene, most commonly in exon 8, have been identified not have significant impact on T2D risk in the examined
(220). A WFS1-like syndrome, WFS2, is caused by ho- populations (226).
mozygous mutation in the CDGSH iron sulfur domain 2 Awata et al (227) reported that variant R456H in the
(CISD2) gene. CISD2 encodes a highly conserved zinc- WFS1 gene was associated with an increased risk of com-
finger protein, ER intermembrane small (ERIS), which mon T1D in a small candidate gene association study,
plays a crucial role in calcium homeostasis (221). WFS1 which needs validation in larger populations.
and WFS2 share some common clinical features such as
diabetes mellitus, hearing loss, and optic atrophy or neu- 3. Molecular mechanisms and pathophysiology that
ropathy. They also display distinct characteristics, eg, di- underlie the association of mutations in the WFS1 gene
abetes insipidus is a specific feature for WFS1, whereas and diabetes
defective platelet aggregation resulting in bleeding peptic In WFS1 patients, loss-of-function mutations in the
ulcer appears to be specific for WFS2 (222). WFS1 gene lead to deficiency of wolframin protein and an
inadequate activation of the UPR in response to accumu-
2. WFS1 gene variants and T2D lation of unfolded proteins within the ER lumen, which
Minton et al (223) reported the first evidence that vari- induces ER stress and results in apoptotic cell death. These
ation in the WFS1 gene may influence susceptibility to mechanisms are believed to underlie progressive neurode-
T2D. In a parent-offspring trios study, R456 and H611 in generation and endocrine dysfunction in these patients
WFS1 were found to be overtransmitted to affected off- (216, 228). Interestingly, ER stress-mediated diabetes in
spring from heterozygous parents in diabetic families. In a WFS1 was diagnosed earlier in life than the common type
further cohort study, the H611 allele and the R456-H611 of autoimmune-mediated T1D. WFS1 mutations induce
haplotype were present more frequently in T2D individ- ER stress, which in turn mediates the decline of ␤-cell mass
uals than in control subjects. and diabetes in WFS1 patients. Poor glycemic control-
After genotyping 1536 SNPs in 84 genes regulating induced oxidative stress is thought to further aggravate the
pancreatic ␤-cell development, function, and survival, progression of neurodegeneration in this disease (216).
Sandhu et al (224) found that rs10010131 (minor allele The development of therapeutic measures that alleviate
frequency [MAF] ⫽ 40%) and rs6446482 (MAF ⫽ 41%) ER stress and oxidative stress could benefit WFS1 pa-
in the WFS1 gene were strongly associated with T2D tients. Calpain, a family member of calcium-dependent,
nonlysosomal cysteine proteases, may be a plausible ther- produces strong antidiabetic effects along with bone loss
apeutic target. This enzyme is thought to be a mechanistic (235). In addition, obesity can activate cyclin-dependent
link between the ER and ␤-cell death in WFS1 (229). Loss- kinase 5 in adipose tissue, resulting in the phosphorylation
of-function mutations of WFS1 increase cytoplasmic cal- of PPARG at serine 273. This modification underlies the
cium levels, leading to calpain activation, which causes ER dysregulation of a number of genes associated with obesity
stress-induced ␤-cell death (230). Treatment with the ry- and insulin resistance in adipose tissue. PPARG ligands
anodine receptor inhibitor dantrolene was shown to re- such as TZDs can block this phosphorylation and confer
duce calcium leakage from the ER to cytosol, lowering their antidiabetic therapeutic effects (236, 237).
cytosolic calcium levels and suppressing calpain activa-
tion; it also prevents cell death in pancreatic ␤-cells and 1. Mutations in PPARG lead to severe insulin resistance
neural progenitor cells. The protective effect of dantrolene and diabetes

suggests that modulating calcium levels by suppressing Barroso et al (238) first reported the identification of
calpain activation could be one approach to treat WFS1 two heterozygous mutations, P467L and V290M, in the
and other diseases associated with excess ER stress (229). ligand-binding domain of PPARG in three subjects with
In addition to its role in the UPR, WFS1 may have other severe insulin resistance who presented with diabetes and
␤-cell-specific functions. WFS1 localizes not only to the hypertension at an unusually early age. Although these
ER but also to secretory granules in pancreatic ␤-cells. mutations are uncommon, detected in just three of 85 se-
Compared with wild-type ␤-cells, a 32% reduction in verely insulin-resistant individuals and none in 314 con-
granular acidification was observed in Wfs1-deficient trols, they represent a unique model of monogenic diabe-
␤-cells (231). This aberration may account for the im- tes through decreased insulin sensitivity, without overtly
paired proinsulin processing and an increased circulating defective ␤-cell function.
proinsulin level in Wfs1-null ␤-cells. Electron microscopy
showed that Wfs1 deficiency leads to a reduction in 2. PPARG gene variants and T2D
plasma membrane-attached secretory granules in the Several variants in the PPARG gene have been identi-
␤-cell, indicating that WFS1 may play a role in the intra- fied, with the P12A variant (rs1801282), first identified by
cellular distribution of secretory granules, especially the Yen et al (239), being the most extensively examined in
docking of insulin granules to the plasma membrane. epidemiological studies. Ethnic heterogeneity has been ob-
These perturbations may cause the profoundly impaired served between the PPARG P12A and the magnitude of
GSIS in Wfs1-deficient mice, providing new insights into the association with T2D (240 –242). In a large meta-anal-
the molecular basis of ␤-cell dysfunction in WFS1 pa- ysis that involved 32 849 T2D cases and 47 456 controls,
tients (231). which included a population-based cohort, case-control,
cross-sectional, and GWAS analyses, Gouda et al (243)
J. Peroxisome proliferator-activated receptor ␥ found that the PPARG P12A polymorphism was associ-
The PPARG gene, predominantly expressed in white ated with a reduction in T2D risk, with a combined overall
and brown adipose tissue, is a master regulator of adipo- odds ratio of 0.86.
genesis as well as a potent modulator of whole-body en- Of note, PPARG SNP rs1801282 (P12A C⬎G) was
ergy balance, lipid biosynthesis, and insulin sensitivity found to be associated with T1D in one report (244) but
(232). PPARG exists in two splice isoforms, PPARG1 and not in another (245), suggesting that this T2D locus may
PPARG2, with the latter carrying an additional 30 amino or may not play a role in T1D genetic susceptibility.
acids at its N terminus. PPARG is a nuclear hormone
receptor and a target of thiazolidinediones (TZDs), a com- 3. Molecular mechanisms and pathophysiology that under-
monly used class of antidiabetic medication despite con- lie the association of mutations or variations in the PPARG
cerns on side effects of weight gain, fluid retention, bone gene and diabetes
loss, and heart failure. PPARG activation by TZDs leads The P467L mutation discussed above (238) lies in an
to a marked improvement of whole-body insulin sensitiv- amphipathic ␣-helix at the C terminus of the PPARG li-
ity and glucose metabolism (232). Recently, fibroblast gand-binding domain, a motif that is critical for mediating
growth factor (FGF) 21 and FGF1 are thought to act lo- ligand-dependent transactivation and coactivator recruit-
cally in adipose tissue to mediate the pharmacological ac- ment. The mutant displays a severely impaired binding
tions as well as the side effects of TZDs (233, 234). Fgf21- capacity to its ligand and coactivator. The V290M recep-
deficient mice are resistant to the insulin-sensitizing effects tor mutation is located proximally within the ligand-bind-
of TZDs as well as the associated weight gain and fluid ing domain. Both PPARG mutants inhibit wild-type re-
retention (234). In contrast, FGF21 administration in mice ceptor function in a dominant-negative manner, which
may lead to silencing of basal gene transcription (238). A. Preproinsulin gene, INS
Notably, a recent study highlights that motif-altering INS is unique among monogenic diabetes genes in that
SNPs in target genes of PPARG cause differential PPARG INS mutations can cause hyperinsulinemia, hyperproin-
binding to the DNA, modulating the recruitment of co- sulinemia, NDM, MODY10, as well as autoantibody-
operating factors including CCAAT/enhancer binding negative T1D (type 1b) (Figure 4). Moreover, variants of
protein, and thus determine individual disease risk and the the INS gene are also strongly associated with common
response to TZD treatment (246). T1D (type 1a), but inconsistently with T2D.
Peroxisome proliferator-activated receptor ␣ (PPARA)
1. Mutations in the INS gene lead to hyperinsulinemia
is abundantly expressed in the liver, heart, and muscle,
and hyperproinsulinemia
where it plays important roles in the regulation of fatty
Mutations of the INS gene that produce structurally
acid oxidation, lipoprotein metabolism, and glucose ho-

defective insulins can cause hyperinsulinemia and late-on-
meostasis (247). In the Quebec family study, Bossé et al
set diabetes (Figure 4). These mutated insulins display
(248) examined the interactions and combined effects of
drastically reduced bioactivity and receptor binding ca-
PPARG2 A12 and PPARA V162 on glucose and insulin
pacity, and often prolonged half-life (251). For example,
homeostasis. Carriers of the PPARA V162 allele were
the INS mutations F49L (252), F48S (253, 254), and
found to display higher insulin and C-peptide levels in
V92L (255) have been identified in such patients. It is
response to a glucose challenge. The PPARG2 A12 allele important to note that not all individuals who harbor
reduced plasma insulin or C-peptide levels only on a heterozygous INS gene mutations necessarily develop di-
PPARA V162 genetic background. The PPARG2 A12 al- abetes because affected individuals who produce both
lele appeared to attenuate the deleterious effect of the normal and mutant insulins may compensate for the defect
PPARA V162 allele and improve insulin sensitivity (248), by overproducing the normal insulin. In addition, other
an observation that is consistent with the fact that obese factors such as aging, body weight gain, and insulin resis-
subjects carrying the PPARG2 A12 allele exhibited tance may also play a role.
greater insulin sensitivity as measured by euglycemic-hy- Insulin gene mutations affecting the conversion of pro-
perinsulinemic clamp (249). insulin to insulin cause hyperproinsulinemia with or with-
Recently, Claussnitzer et al (250) identified the T2D out mild diabetes. Mutations at the junction of C-peptide
risk SNP rs4684847 as a cis-regulatory variant located at and ␣-chain (R89H, R89L, R89P; Figure 4) (251) disrupt
6.5 kb upstream of the PPARG2-specific promoter. Func- the cleavage site of the prohormone convertase 2, and the
tional studies revealed that paired related homeobox 1 mutated insulins can be cleaved only at the junction of
(PRRX1) may act as a repressor of PPARG2 via its en- C-peptide and ␤-chain by prohormone convertase 1/3.
hanced binding at the rs4684847 C risk allele. Consistent Another mutation at ␤-chain (H34D) was also shown to
with the role of PPARG2 for insulin sensitivity rather than cause hyperproinsulinemia via an unclear mechanism.
insulin secretion, this enhanced binding activity was ob- H34D proinsulin has a 4- to 5-fold higher receptor binding
served only in insulin target cells (adipocytes, hepatocytes, activity than wild type, and it has been suggested that it
and myocytes), but not in pancreatic ␤-cells or kidney may bind to insulin receptors in the trans-Golgi and be
cells. These findings support the interpretation that the secreted constitutively (256, 257).
rs4684847 C risk allele-mediated enhanced binding of
PRRX1 represses the expression of PPARG2, leading to re- 2. Mutations in the INS gene lead to PNDM, MODY10, and
duced insulin sensitivity and impaired glucose homeostasis. type 1b diabetes
In 2007, Støy et al (258) first reported that mutations
of the INS gene are a novel cause of PNDM. They iden-
III. Monogenic Diabetes Genes Associated tified 10 heterozygous mutations in INS among 16 PNDM
With Both T1D and T2D families. These patients develop diabetes at a median age
of 9 weeks, usually with diabetic ketoacidosis or severe
Among the list of monogenic diabetes genes, variants of hyperglycemia that needs immediate insulin therapy. To
the human preproinsulin gene (INS) are strongly associ- date, heterozygous autosomal dominantly inherited mu-
ated with common T1D, but inconsistently with T2D. In- tations in the INS gene are a common cause of PNDM
terestingly, GWAS and meta-analyses showed that vari- (⬃12%), second only to mutations in KCNJ11 (35). These
ants of GLIS3, a NDM gene, are associated with dominant heterozygous INS mutations lead to preproin-
susceptibility to both common T1D and T2D. sulin misfolding and accumulation in the ER of the ␤-cell,
Figure 4.

Figure 4. Diagram depicting the human preproinsulin molecule with mutations that are currently known to cause diabetes and hyperinsulinemia or
hyperproinsulinemia. The amino acid residues in the signal peptide, the ␤-chain, the C-peptide, and the ␣-chain are shown in different colors. The
dashed circles indicate the basic residues that are the cleavage sites for conversion from proinsulin to insulin. The mutations underlying different
types of diabetes and hyperinsulinemia or hyperproinsulinemia are presented in oval shapes filled in distinct colors as indicated. We note that some
translation initiation codon mutations (eg, 3G⬎T, 3G⬎A), in-frame deletion mutants in the coding region, and mutations in the promoter or
intron regions of the INS gene have been left out of this diagram. In addition, several INS substitution mutants, eg, A23S, A23T, L68M, and G84R,
that were initially reported to occur in patients with diabetes have been left out of this diagram because subsequent studies showed that they are
functional incidental variants. (See Section III.A for details.)
causing ER stress and ␤-cell apoptosis (see Section II.A.5 In addition, this mutation also produces an aberrant tran-
for details). script with an insertion of a 79-nucleotide pseudoexon
Some INS mutations outside the exonic regions can following exon 2 using a native 3⬘ acceptor site. These two
also cause diabetes. For example, mutations in the intron mutant transcripts have been predicted to undergo non-
regions of the INS gene such as heterozygous c.188 – sense and nonstop-mediated mRNA decay, respectively
31G⬎A (259) and homozygous c.187⫹241G⬎A (260) (260). Notably, a number of single amino acid substitu-
cause PNDM. The intronic c.188 –31G⬎A mutation in- tions in insulin, such as A23S, A23T, L68M, and G84R,
troduces an ectopic splice site leading to the insertion of 29 that had been initially reported as possible deleterious
nucleotides from the intronic sequence into the mature pathogenic mutants in subjects with diabetes (35, 262,
mRNA, which results in an aberrant transcript, producing 263) could represent instances of ascertainment bias
misfolded proteins to induce ER stress and ␤-cell death (264); the same “mutated” insulin molecules were subse-
(259). It is interesting that this c.188 –31G⬎A mutation quently shown to be nonpathogenic incidental variants
was also reported to cause MODY in one family (261). (35, 265–267).
The other intronic c.187⫹241G⬎A mutation introduces Molven et al (268) extended the phenotype of the INS
a donor 5⬘ splice site, producing a truncated 482-bp INS gene mutations to include rare cases of MODY
mRNA product that includes a 237-bp intronic sequence. (MODY10) and type 1b diabetes. They reported the INS
mutation c.137G⬎A (R46Q) in a MODY10 family and I alleles of INS VNTR (26 to 63 repeats, frequency ⬃70%
the INS mutation c.163C⬎T (R55C) in a type 1b diabetes in Caucasians) are associated with T1D, whereas the class
family who presented with ketoacidosis and insulin de- III alleles (140 to more than 200 repeats, frequency ⬃30%
pendency. The R46Q mutation disrupts a critical hydro- in Caucasians) seem to confer protection against T1D.
gen bond formation, which impairs the stability of the Although these classes of VNTR alleles have little effect on
insulin molecule. The R55C substitution is thought to per- insulin transcript levels in the pancreas, class I alleles are
turb insulin biosynthesis via the introduction of an un- associated with lower levels of insulin mRNA and protein
paired cysteine, as noted in the Akita mouse harboring a in the thymus, which may compromise negative selection
C96Y mutation in the ins2 gene (269, 270), or it may affect and allow more autoreactive T cells to escape into the
proteolytic processing of proinsulin to insulin (268). It is periphery, increasing susceptibility to T1D, whereas the
significant that the C96Y mutation has also been found in higher insulin levels associated with class III alleles may

rare patients with PNDM (271). A novel heterozygous enhance negative selection for autoreactive T cells, pro-
single nucleotide deletion (c.233delA) leading to a frame- tecting against T1D (274). In support of this hypothesis,
shift mutation (Q78fs) in the INS gene was recently iden- recent GWAS and other studies highlight INS as the sec-
tified in a MODY family. This mutation produces an ab- ond strongest risk locus (right after HLA) for T1D (273),
errant “proinsulin” that is devoid of the native structures insulin and its precursors being major initiating autoan-
of C-peptide and ␣-chain (261). tigens in human T1D (275). It is unclear whether any mis-
Although most of the diabetogenic mutations in the sense mutation in the INS gene underlies common T1D,
INS gene occur in a dominant heterozygous manner, although the INS mutation c.163C⬎T (R55C) has been
Garin et al (272) identified 10 different recessive INS mu- linked to type 1b diabetes (268).
tations that cause PNDM or TNDM by interfering with
the efficiency of insulin biosynthesis. These include four 4. INS gene variants associated with T2D
homozygous mutations involving the coding region: Although the class III alleles of INS VNTR may render
c.184C ⬎ T (p.Q62X), c.3G ⬎ T (p.0?), c.3G ⬎ A (p.0?), protection against T1D, association studies suggested that
and a deletion (c.-370-? 186⫹?del); five homozygous mu- they are inconsistently associated with the risk of T2D,
tations in the promoter or 3⬘ UTR regions: c.-331C ⬎ A, probably due to population admixture (276 –281). In
c.-331C ⬎ G, c.-218A ⬎ C, and c.-366_-343del, and Caucasians, class III INS VNTR appeared to confer a
c.*59A ⬎ G; and one compound heterozygote for two modest increase in risk of T2D (276), but the observation
regulatory region mutations, c.-331C ⬎ G and c.-332C ⬎ could not be replicated in a large family-based association
G. Interestingly, patients with recessive INS mutations study (279). The latter study did, however, report a
were diagnosed earlier (median, 1 week vs 10 weeks) and marked excess transmission of class III INS VNTR from
presented with a lower birth weight (⫺3.2 vs ⫺2.0 SD fathers to diabetic probands, suggesting the potential con-
score) than those with dominant heterozygous INS mu- tribution of imprinted genes in the risk of T2D. In the Pima
tations. A plausible explanation of these observations is population, significant linkage disequilibrium was ob-
that two distinct disease mechanisms are involved in served between INS VNTR (-23Hph1, an A/T SNP within
pathogenesis. In babies with recessive INS mutations, dis- the Hph1 restriction endonuclease site) and birth weight,
rupted insulin synthesis occurs as soon as the ␤-cell starts but not T2D (280).
to produce insulin during early fetal development. In con- Notably, one large-scale meta-analysis showed that ty-
trast, in patients with dominant heterozygous INS muta- rosine hydroxylase (TH)-INS variant (rs10770141), a
tions, the mutant preproinsulin misfolds and accumulates T1D risk allele, is protective for T2D. This variant, dif-
in the ER of the ␤-cell, causing increasingly severe ER ferent from INS VNTR, is located in the promoter region
stress that with time ultimately disrupts insulin production of the TH gene and 11 kb upstream of the INS gene. TH-
by the ␤-cell (267). INS variant regulates the expression of TH, but not INS
(282). The mechanism underlying the association between
3. INS gene variants associated with T1D and this variant of TH and diabetes has not been investigated.
possible mechanisms
Prior to GWAS, the INS gene on chromosome 11p15 5. Molecular mechanisms and pathophysiology that under-
was one of a few loci identified to be associated with sus- lie INS mutations and diabetes
ceptibility to T1D (273). The genetic risk conferred by the Heterozygous dominant mutations of the INS gene
INS locus maps to a DNA region with a variable number have been identified within all regions of the preproinsulin
of tandem repeats (VNTRs) of a 14- to 15-bp sequence molecule including the signal peptide, C-peptide, and
approximately 0.5 kb upstream of the INS gene. The class flanking cleavage sites (Figure 4) (283). The mutations
within the signal peptide region likely impair one or more mutants that cause PNDM (287). Forced expression of
of the earliest events of proinsulin biosynthesis, including mouse proinsulin mutants with either cysteine residue sub-
completion of translocation, excision of the signal pep- stitution (C96Y) or noncysteine residue substitution
tide, initiation of ␤-chain folding, and disulfide bond for- (H29D or L35P), all known to be associated with human
mation (283). For example, R6C and R6H mutations in PNDM (Figure 4), inhibits the production of mature hu-
the INS gene lead to loss of the N-terminal positive charge man insulin in the rat INS-1-derived 832/13 ␤-cell line,
of the 24-amino acid signal peptide, which fails to be fully which coexpresses wild-type human insulin. On the other
translocated across the ER membrane and accumulate in hand, it is noteworthy that the simultaneous introduction
a juxtanuclear cytoplasmic compartment, inducing ER of missense substitutions that replace all six cysteine res-
stress and ␤-cell death (284). The preproinsulin A24D idues of proinsulin fails to repress the secretion of coex-
mutation that affects the last signal peptide residue likely pressed wild-type human insulin. This suggests that path-

disrupts normal signal peptide cleavage and interferes ways for proinsulin folding/processing are complicated;
with downstream proinsulin folding and trafficking (265), cysteine residues (paired or unpaired) in the INS mutants
leading to PNDM. play a critical role in causing intracellular entrapment of
Proper disulfide pairing is a key event that determines coexpressed wild-type proinsulin, disrupting insulin pro-
whether a proinsulin molecule can achieve its native duction (287). In addition, INS mutants associated with
folded structure. Proinsulin contains six cysteine residues PNDM facilitate the wild-type proinsulin to recruit into
that form three evolutionarily conserved disulfide bonds: non-native disulfide-linked protein complexes in a cell line
B7-A7, B19-A20, and A6-A11 (Figure 4). Many INS mu- and in pancreatic islets of Akita mice (269, 270), which
tations introduce or eliminate a cysteine residue in the leads to disruption of normal folding of wild-type proin-
polypeptide chain, producing unpaired cysteine residues, sulin. Although the C96Y mutation affects only one copy
which may alter or disrupt normal proinsulin disulfide of the Ins2 gene, leaving one normal Ins2 gene and two
pairing (283). For example, the mutations C43G and normal Ins1 genes intact in these animals, ␤-cells from
C96Y in the INS gene disrupt normal disulfide bonds at Akita mice display severe ER stress, accompanied by fail-
B19-A20 and A7-B7, respectively. The mutations R89C ure of GSIS, culminating in diabetes at an early age (269,
and G90C introduce an additional unpaired cysteine res- 270). As noted earlier, heterozygous C96Y mutation was
idue at the ␣-chain-C-peptide cleavage site (258). It is note- also found to cause PNDM in humans (271). Thus, it has
worthy that some mutations, such as G32S, G32R, and been proposed that perturbation of the disulfide-coupled
G47V, do not change proinsulin cysteine content, but may folding pathway of wild-type proinsulin initiates the
still disrupt the normal proinsulin folding, an observation molecular pathogenesis of PNDM associated with INS
that may be related to the fact that most of these mutations mutations (287).
involve evolutionarily highly conserved residues (258). In addition to INS mutations that cause proinsulin mis-
These mutations in the INS gene may lead to proinsulin folding and severe ER stress that cripples the insulin pro-
misfolding and accumulation in the ER of the ␤-cell and duction machinery of the ␤-cell, recessive INS mutations
may cause ER stress, ␤-cell apoptosis, and diabetes. Au- can also cause decreased insulin biosynthesis and diabetes
tophagy is a physiological process that maintains cellular via other mechanisms (272). These include absent or al-
homeostasis via clearance of misfolded proteins and tered translation because of coding-sequence deletions or
damaged organelles (285). Accumulation of excessive mutations, reduced insulin transcripts because of muta-
amounts of misfolded proinsulin molecules per se, eg, in tions in the promoter, or increased insulin mRNA insta-
Akita mice, can stimulate autophagy by inhibiting mam- bility. A few examples of INS mutations that impair in-
malian target of rapamycin complex 1, alleviates ER sulin production directly include: two point mutations
stress, and partially prevents ␤-cell apoptosis. Treatment (c.3G ⬎ A and c.3G ⬎ T) at the translation initiation
of female diabetic Akita mice with rapamycin (mamma- codon that abolish the native translation initiation for the
lian target of rapamycin complex 1 inhibitor) improves preproinsulin protein; a 3⬘ UTR mutation (c.*59A ⬎ G)
diabetes and further down-regulates ␤-cell apoptosis that abolishes the polyadenylation signal and potentially
(286). These findings suggest that augmentation of au- impairs mRNA stability; a 24-bp deletion (c.-366_-
tophagy may be a novel therapeutic approach for this type 343del) in the INS promoter region that disrupts the pro-
of diabetes. The challenge would be to find agents that are moter evolutionary conserved C1 and E1 elements, which
safe and efficacious in alleviating the excess ER stress in are involved in the binding of MAFA and NEUROD1 to
the ␤-cell. the INS promoter, respectively (288); and the CC dinu-
Dominant-negative inhibition of wild-type proinsulin cleotide mutations, c.-331(C ⬎ G, C ⬎ A) and c.-332C ⬎
trafficking appears to be a feature of many of the INS G, that abolish a canonical binding motif for a key INS
transactivator protein GLIS3 (214, 272, 289) (see discus- to that provided by HLA alone. Furthermore, they iden-
sion in the next section), producing up to 90% reduction tified that HLA plus nine SNPs including the one from
in the INS promoter transcriptional activity. GLIS3 could achieve similar prediction accuracy as the
total SNP set for T1D. They proposed a weighted risk
B. Gli-similar 3 model with selected SNPs that could be considered for
GLIS3, a member of the Krüppel-like family of tran- recruitment of at-risk infants into studies of the natural
scription factors, is predominantly expressed in pancreatic history or disease prevention for T1D.
␤-cells, thyroid, and kidney (290, 291). Interestingly,
GLIS3 is a NDM gene that is also associated with the 3. GLIS3 gene variants and T2D
susceptibility of both common T1D and T2D, as identified Shortly after the report that the GLIS3 variant was
by GWAS and large- scale meta-analyses that include mul- strongly associated with common T1D (295), Dupuis et al

tiple populations. (78) identified the association between a variation in
GLIS3 and common T2D, which has been subsequently
1. Mutations in the GLIS3 gene lead to monogenic diabetes- extensively examined in epidemiological studies among
related syndromes different populations.
About a decade ago, Senée et al (291) first reported that In meta-analyses involving 46 186 nondiabetic partic-
mutations in the GLIS3 gene cause a previously described ipants in 21 GWAS and follow-up studies, Dupuis et al
rare syndrome characterized by PNDM, congenital hypo- (78) identified the GLIS3 variant (rs7034200 A) as one of
thyroidism, polycystic kidneys, and in some cases, also nine newly identified loci associated with fasting glucose
with intrauterine growth retardation, facial anomalies, in T2D trait, but not associated with body mass index,
congenital glaucoma, and hepatic fibrosis. Six individuals blood pressure, or lipid profile in participants of European
from three families were affected. In one family (with neo- descent. These findings were replicated in Danes (298),
natal diabetes, hypothyroidism [NDH]), they identified a ho- Chinese (299, 300), South Asians (301), and East Asians
mozygous insertion (2067insC) in GLIS3 leading to a (302). Boesgaard et al (298) found that GLIS3 variant
frameshift and a truncated protein. In two other families rs7034200 A was associated with reduced glucose-stim-
with an incomplete syndrome, affected individuals harbor ulated ␤-cell function in middle-aged Danish people.
homozygous large fragment (426 and 149 Kb) deletions Barker et al (303) reported that GLIS3 variant rs7034200
affecting the 5⬘ UTR of the GLIS3 gene. Recently, addi- was associated with fasting blood glucose levels in healthy
tional patients carrying homozygous partial GLIS3 gene children and adolescents of European origin. GLIS3 vari-
deletions or mutations have been reported, most present- ant rs7034200 carriers showed impaired ␤-cell function,
ing with NDH as well as renal cysts and liver disease (292– as indicated by homeostasis model assessment. Addition-
294). It is noteworthy that Dimitri et al (293) reported a ally, variant rs7041847 A in GLIS3 is associated with
patient with compound heterozygous mutations in GLIS3 susceptibility to T2D in East Asians (302, 304).
(p.Arg589Trp/exons 1–11 del) presenting with NDM, but
not congenital hypothyroidism or renal cysts, who sur- 4. Molecular mechanisms and pathophysiology that
vived to adulthood. underlie the association of GLIS3 mutations or variations
and diabetes
2. GLIS3 gene variants and T1D The gene association studies in humans are tantalizing.
Barrett et al (295) first identified the association be- However, the underlying mechanisms were largely un-
tween a GLIS3 variant and common T1D in a report that known until Glis3 global and ␤-cell-specific knockout
combined a new GWAS and a meta-analysis of two prior mice were generated. Early studies in vitro identified a
studies that involved a total of 7514 T1D patients and GLIS3 response element in the insulin gene promoter and
9045 control individuals. They found that the region of revealed that GLIS3 physically and functionally interacts
strong linkage disequilibrium at chromosome 9p24.2 with with PDX1, MAFA, and NEUROD1 to control insulin
a MAF of 50.2% harbors only a single gene, GLIS3. The gene transcription (214). Moreover, the direct binding of
risk allele rs7020673 G was strongly associated with sus- GLIS3 to the insulin promoter is required for the interac-
ceptibility to T1D in GWAS, case-control, and families tion of PDX1, MAFA, and NEUROD1 with the promoter
analyses. This association was replicated recently in Pak- (305), indicating that GLIS3 plays a pivotal role in form-
istani T1D patients (296). ing a transactivator complex to control insulin transcrip-
In another recent study, Winkler et al (297) found that tion. In agreement with these in vitro findings, Yang et al
the inclusion of 40 non-HLA gene SNPs could signifi- (306) found that tamoxifen-mediated ␤-cell-specific inac-
cantly improve the prediction accuracy of T1D, compared tivation of Glis3 in adult mice down-regulates insulin ex-
Figure 5. cyclin D2 (Ccnd2) mRNA expres-

sion (306). CCND2 is essential for
postnatal pancreatic ␤-cell growth
(309, 310) and compensatory mass
expansion in response to insulin re-
sistance (311). Interestingly, recent
GWAS identified that the variant
rs11063069 G in CCND2 confers
susceptibility risk to T2D (312),
whereas the variant rs76895963 G
in CCND2 reduces the risk of T2D

by half and is correlated with in-
creased CCND2 expression (167).
GLIS3 may also protect pancre-
atic ␤-cells against apoptosis. Proin-
flammatory cytokines, particularly
IL-1␤ and interferon-␥ (313), and
the saturated nonesterified fatty acid
Figure 5. Multifaceted role of GLIS3 in diabetes. In utero, GLIS3 controls islet differentiation by palmitate (314), may contribute to
transactivating Ngn3, synergistically with HNF6 and forkhead box A2 (FOXA2, also known as
HNF3B or transcription factor 3B [TCF-3B]). Loss-of-function mutations or ablation of Glis3 cause
␤-cell loss in T1D and T2D, respec-
impaired islet differentiation and NDM. After birth, GLIS3 predominantly controls insulin gene tively. Nogueira et al (315) reported
transcription, cooperating with MAFA, PDX1, and NEUROD1. GLIS3 is also required for obesity- that Glis3 knockdown in INS-1 cells
induced ␤-cell proliferation and compensatory ␤-cell mass expansion by transactivating Ccnd2. In
increased basal and inflammatory
addition, GLIS3 plays a protective role for ␤-cell survival, possibly by regulating BimS. Therefore,
impairment of GLIS3 function also plays a key role in the development of T1D and T2D. This cytokine (IL-1␤ ⫹ interferon-␥)- or
figure summarizes the pathways by which ablation, loss-of-function mutations, or functional palmitate-induced ␤-cell apoptosis.
impairment of GLIS3 cause NDM, T1D, and T2D. Pathways in the blue rectangle take place in the They suggested that reduced Glis3
developing pancreas in utero and underlie monogenic NDM; pathways in the yellow square and
the red rectangle pertain to the postnatal ␤-cell and contribute to the development of T1D and expression might modulate the alter-
T2D, respectively. native splicing of the proapoptotic
BH3-only protein Bim, promoting
pression, leading to fulminant diabetes and death. Thus, the expression of the prodeath variant BimS and ␤-cell
both in vitro and in vivo data demonstrated that GLIS3 is death.
a potent transactivator of the insulin gene. These data from cell lines and mouse models provide
To gain insight into the role of GLIS3 in monogenic the mechanistic basis for the genetic associations of the
diabetes, three groups independently generated Glis3-de- GLIS3 gene and NDM (291), common T1D (295, 296),
ficient mice (289, 307, 308). Recapitulating the phenotype and T2D (78, 298 –304) (Figure 5). GLIS3 may underlie all
of NDH patients with GLIS3 mutations (291), Glis3-null three forms of diabetes by potently controlling insulin
mice died with severe hyperglycemia and ketoacidosis gene transcription and likely ␤-cell survival. The homozy-
within the first days of life. The mice also developed hy- gous insertion (2067insC) in GLIS3 identified in an NDH
pothyroidism and polycystic kidney, whereas exocrine family (291) leads to a frameshift and a truncated protein
development was unaffected. Mechanistically, GLIS3 was that lacks the C-terminal transactivation domain (290),
shown to control fetal islet differentiation, synergistically which is required to activate both NGN3 (308) and INS
with HNF6 and forkhead box A2 (FOXA2), via direct (214, 305, 306) gene transcription. This serious develop-
transactivation of Ngn3, the endocrine lineage-defining mental deficiency causes severely abnormal ␤-cell devel-
transcription factor (308). opment in utero and marked impairment in insulin pro-
Consistent with an autosomal recessive inheritance duction at birth, culminating in PNDM. In addition to
observed in humans, heterozygous Glis3-mutant mice causing PNDM (289, 291, 307, 308), GLIS3 was also
remained euglycemic while on regular chow diet. How- found by GWAS to be linked to common T1D (295, 296),
ever, when they were challenged with a HFD, adult mice probably through its action on insulin gene transcription
developed diabetes because of impaired compensatory and ␤-cell survival. Finally, GWAS showed that the GLIS3
␤-cell proliferation and mass expansion. It has been locus is also linked to T2D (78, 298 –304). Studies in adult
shown that GLIS3 controls ␤-cell proliferation in re- mouse models suggest that GLIS3 plays a critical role in
sponse to HFD feeding partly by direct regulation of compensatory obesity-induced ␤-cell proliferation via the
regulation of Ccnd2 transcription (306). Of note, the ef- Individual variants may display disrupted islet enhancer
fects of the susceptibility gene variants examined in iso- activity, which may be an important mechanism for the
lation on the risk of T1D or T2D are typically modest, genetic predisposition of T2D (317, 318). It remains a
often making it a challenge to decipher such subtle effects challenge to translate our knowledge from research in the
by current functional assays (10). The precise mechanisms monogenic diabetes genes into effective forms of treat-
underlying the GLIS3 variants and diabetes remain to be ment. Because many of these have been shown to occur in
fully elucidated. T1D or T2D, we are hopeful that future research into these
genes will provide new insights into disease pathogenesis
and treatment target identification, not only for the rare
IV. Concluding Remarks monogenic disorders, but also for the worldwide epidemic
of T1D and T2D.
In the past few decades, an increasing number of genes

have been linked to monogenic diabetes and related syn-
dromes. GWAS have markedly expanded the repertoire of Acknowledgments
susceptibility loci associated with common T1D or T2D.
Although T1D and T2D share overlap in select clinical and Address all correspondence and requests for reprints to: Yisheng Yang,
pathological features, the susceptibility loci identified by Division of Endocrinology, Department of Medicine, MetroHealth
Medical Center, Case Western Reserve University, 2500 Metrohealth
GWAS indicate that the genetic basis for the two common Drive, Cleveland, OH 44109. E-mail: yisheng.yang@case.edu. Or Law-
types of diabetes is mostly distinct (7–15) (Figure 2). In rence Chan, Division of Diabetes, Endocrinology and Metabolism, De-
agreement with this premise, candidate gene-association partments of Medicine and Molecular & Cellular Biology, Baylor Col-
lege of Medicine, One Baylor Plaza, Houston, TX 77030. E-mail:
studies and GWAS showed that, to date, at least 12 mono- lchan@bcm.edu.
genic diabetes genes are associated only with T2D, but not Research in the authors’ laboratories that was discussed in this review
T1D; the INS locus appears to be an exception in that it is was supported by the National Institutes of Health (P30-DK079638) for
a Diabetes Endocrinology Research Center at Baylor College of Medicine
strongly associated with T1D but is inconsistently asso-
(T32-HL066991), Juvenile Diabetes Foundation (Award 46-2010-752),
ciated with T2D. GLIS3 is a developmental transcription the American Diabetes Association (Grant 1-14-MN-01), the T.T. and
factor gene that was shown to underlie a rare kind of W.F. Chao Global Foundation, the Cindy and Frank Liu Family Foun-
monogenic PNDM. It is unique that GWAS showed that dation, the Cunningham Family Foundation, the Betty Rutherford Chair
in Diabetes Research at Baylor St. Luke’s Medical Center (to L.C.), the
variants of GLIS3 are associated with both common T1D American Heart Association (Grant 13SDG17090096; to Y.Y.), and
and T2D in different populations. Not unexpectedly, the MetroHealth Medical Center, Case Western Reserve University start-up
T1D and T2D GLIS3 variants are unique and do not over- funds (to Y.Y.).
Disclosure Summary: The authors have nothing to disclose.
lap, likely because they confer their diabetogenic effects
via different mechanisms.
The monogenic forms of diabetes offer excellent mod-
References
els to unravel the developmental biology of pancreatic
␤-cells and the role of the implicated genes in the regula- 1. Steck AK, Winter WE. Review on monogenic diabetes. Curr
tion of ␤-cell function. Information gained from these Opin Endocrinol Diabetes Obes. 2011;18(4):252–258.
studies also facilitates the interpretation of GWAS data for 2. Greeley SA, Naylor RN, Philipson LH, Bell GI. Neonatal
diabetes: an expanding list of genes allows for improved di-
common T1D and T2D. Because more and more rare vari-
agnosis and treatment. Curr Diab Rep. 2011;11(6):519 –532.
ants are being identified in diabetic patients with advanced 3. Edghill EL, Dix RJ, Flanagan SE, et al. HLA genotyping
sequencing technologies, special caution needs be taken to supports a nonautoimmune etiology in patients diagnosed
avoid ascertainment bias, overinterpretation, and mis- with diabetes under the age of 6 months. Diabetes. 2006;
identification of risk variants for common forms of dia- 55(6):1895–1898.
betes (264). A number of rare mutations of late-onset 4. Flanagan SE, Haapaniemi E, Russell MA, et al. Activating
germline mutations in STAT3 cause early-onset multi-organ
MODY genes (eg, PDX1) have been reported to be vari-
autoimmune disease. Nat Genet. 2014;46(8):812– 814.
ants associated with common T2D, blurring the boundary 5. Fajans SS, Bell GI. MODY: history, genetics, pathophysi-
between the two diabetes syndromes. Differentiating be- ology, and clinical decision making. Diabetes Care. 2011;
tween an incompletely penetrant MODY mutation from a 34(8):1878 –1884.
rare T2D variant is often difficult and can be a matter of 6. Gardner DS, Tai ES. Clinical features and treatment of
maturity onset diabetes of the young (MODY). Diabetes
semantics (264, 316).
Metab Syndr Obes. 2012;5:101–108.
Most of the T2D risk variants in islet transcription fac- 7. Todd JA. Etiology of type 1 diabetes. Immunity. 2010;
tors are enriched in clustered islet enhancers in human islet 32(4):457– 467.
cells where they interact functionally with the epigenome. 8. Grarup N, Sandholt CH, Hansen T, Pedersen O. Genetic
susceptibility to type 2 diabetes and obesity: from genome- served in the long-term follow-up of nonpancreatecto-
wide association studies to rare variants and beyond. Dia- mized patients with persistent hyperinsulinemic hypogly-
betologia. 2014;57(8):1528 –1541. cemia of infancy due to mutations in the ABCC8 gene.
9. Hara K, Shojima N, Hosoe J, Kadowaki T. Genetic archi- Diabetes Care. 2008;31(6):1257–1259.
tecture of type 2 diabetes. Biochem Biophys Res Commun. 25. Abdulhadi-Atwan M, Bushman J, Tornovsky-Babaey S, et al.
2014;452(2):213–220. Novel de novo mutation in sulfonylurea receptor 1 presenting
10. Bonnefond A, Froguel P. Rare and common genetic events as hyperinsulinism in infancy followed by overt diabetes in
in type 2 diabetes: what should biologists know? Cell early adolescence. Diabetes. 2008;57(7):1935–1940.
Metab. 2015;21(3):357–368. 26. Vieira TC, Bergamin CS, Gurgel LC, Moisés RS. Hyper-
11. Prasad RB, Groop L. Genetics of type 2 diabetes-pitfalls insulinemic hypoglycemia evolving to gestational diabetes
and possibilities. Genes. 2015;6(1):87–123. and diabetes mellitus in a family carrying the inactivating
12. Basile KJ, Guy VC, Schwartz S, Grant SF. Overlap of ge- ABCC8 E1506K mutation. Pediatr Diabetes. 2010;11(7):
netic susceptibility to type 1 diabetes, type 2 diabetes, and 505–508.

latent autoimmune diabetes in adults. Curr Diab Rep. 27. Huopio H, Otonkoski T, Vauhkonen I, Reimann F, Ashcroft
2014;14(11):550. FM, Laakso M. A new subtype of autosomal dominant dia-
13. Pociot F, Akolkar B, Concannon P, et al. Genetics of type betes attributable to a mutation in the gene for sulfonylurea
1 diabetes: what’s next? Diabetes. 2010;59(7):1561–1571. receptor 1. Lancet. 2003;361(9354):301–307.
14. DIAbetes Genetics Replication And Meta-analysis 28. Arnoux JB, Verkarre V, Saint-Martin C, et al. Congenital
(DIAGRAM) Consortium, Asian Genetic Epidemiology hyperinsulinism: current trends in diagnosis and therapy.
Network Type 2 Diabetes (AGEN-T2D) Consortium, Orphanet J Rare Dis. 2011;6:63.
South Asian Type 2 Diabetes (SAT2D) Consortium, et al. 29. Hsu WH, Xiang HD, Rajan AS, Kunze DL, Boyd AE 3rd.
Genome-wide trans-ancestry meta-analysis provides in- Somatostatin inhibits insulin secretion by a G-protein-me-
sight into the genetic architecture of type 2 diabetes sus- diated decrease in Ca2⫹ entry through voltage-dependent
ceptibility. Nat Genet. 2014;46(3):234 –244. Ca2⫹ channels in the ␤ cell. J Biol Chem. 1991;266(2):
15. Fortune MD, Guo H, Burren O, et al. Statistical colocal- 837– 843.
ization of genetic risk variants for related autoimmune dis- 30. Schmidt A, Hescheler J, Offermanns S, et al. Involvement
eases in the context of common controls. Nat Genet. 2015; of pertussis toxin-sensitive G-proteins in the hormonal in-
47(7):839 – 846. hibition of dihydropyridine-sensitive Ca2⫹ currents in an
16. McTaggart JS, Clark RH, Ashcroft FM. The role of the insulin-secreting cell line (RINm5F). J Biol Chem. 1991;
KATP channel in glucose homeostasis in health and disease: 266(27):18025–18033.
more than meets the islet. J Physiol. 2010;588:3201–3209. 31. de Weille JR, Schmid-Antomarchi H, Fosset M, Lazdunski
17. Yorifuji T. Congenital hyperinsulinism: current status and M. Regulation of ATP-sensitive K⫹ channels in insuli-
future perspectives. Ann Pediatr Endocrinol Metab. 2014; noma cells: activation by somatostatin and protein kinase
19(2):57– 68. C and the role of cAMP. Proc Natl Acad Sci USA. 1989;
18. Shyng SL, Ferrigni T, Shepard JB, et al. Functional analyses 86(8):2971–2975.
of novel mutations in the sulfonylurea receptor 1 associ- 32. Gloyn AL, Pearson ER, Antcliff JF, et al. Activating mu-
ated with persistent hyperinsulinemic hypoglycemia of in- tations in the gene encoding the ATP-sensitive potassium-
fancy. Diabetes. 1998;47(7):1145–1151. channel subunit Kir6.2 and permanent neonatal diabetes.
19. Cartier EA, Conti LR, Vandenberg CA, Shyng SL. Defec- N Engl J Med. 2004;350(18):1838 –1849.
tive trafficking and function of KATP channels caused by 33. Babenko AP, Polak M, Cavé H, et al. Activating mutations
a sulfonylurea receptor 1 mutation associated with persis- in the ABCC8 gene in neonatal diabetes mellitus. N Engl
tent hyperinsulinemic hypoglycemia of infancy. Proc Natl J Med. 2006;355(5):456 – 466.
Acad Sci USA. 2001;98(5):2882–2887. 34. Proks P, Arnold AL, Bruining J, et al. A heterozygous ac-
20. Taschenberger G, Mougey A, Shen S, Lester LB, LaFranchi tivating mutation in the sulphonylurea receptor SUR1
S, Shyng SL. Identification of a familial hyperinsulinism- (ABCC8) causes neonatal diabetes. Hum Molec Genet.
causing mutation in the sulfonylurea receptor 1 that pre- 2006;15(11):1793–1800.
vents normal trafficking and function of KATP channels. 35. Edghill EL, Flanagan SE, Patch AM, et al. Insulin mutation
J Biol Chem. 2002;277(19):17139 –17146. screening in 1,044 patients with diabetes: mutations in the
21. Kapoor RR, Flanagan SE, James CT, et al. Hyperinsuli- INS gene are a common cause of neonatal diabetes but a
naemic hypoglycaemia and diabetes mellitus due to dom- rare cause of diabetes diagnosed in childhood or adult-
inant ABCC8/KCNJ11 mutations. Diabetologia. 2011; hood. Diabetes. 2008;57(4):1034 –1042.
54(10):2575–2583. 36. Flanagan SE, Clauin S, Bellanné-Chantelot C, et al. Update
22. Pinney SE, MacMullen C, Becker S, et al. Clinical charac- of mutations in the genes encoding the pancreatic ␤-cell
teristics and biochemical mechanisms of congenital hyper- K(ATP) channel subunits Kir6.2 (KCNJ11) and sulfonyl-
insulinism associated with dominant KATP channel mu- urea receptor 1 (ABCC8) in diabetes mellitus and hyper-
tations. J Clin Invest. 2008;118(8):2877–2886. insulinism. Hum Mutat. 2009;30(2):170 –180.
23. Huopio H, Reimann F, Ashfield R, et al. Dominantly inher- 37. Craig TJ, Shimomura K, Holl RW, Flanagan SE, Ellard S,
ited hyperinsulinism caused by a mutation in the sulfonylurea Ashcroft FM. An in-frame deletion in Kir6.2 (KCNJ11)
receptor type 1. J Clin Invest. 2000;106(7):897–906. causing neonatal diabetes reveals a site of interaction be-
24. Gussinyer M, Clemente M, Cebrián R, Yeste D, Albisu M, tween Kir6.2 and SUR1. J Clin Endocrinol Metab. 2009;
Carrascosa A. Glucose intolerance and diabetes are ob- 94(7):2551–2557.
38. Edghill EL, Flanagan SE, Ellard S. Permanent neonatal di- 51. Lyssenko V, Almgren P, Anevski D, et al. Genetic predic-
abetes due to activating mutations in ABCC8 and KCNJ11. tion of future type 2 diabetes. PLoS Med. 2005;2(12):e345.
Rev Endocr Metab Disord. 2010;11(3):193–198. 52. Baier LJ, Muller YL, Remedi MS, et al. ABCC8 R1420H
39. Bonnefond A, Philippe J, Durand E, et al. Whole-exome loss-of-function variant in a Southwest American Indian
sequencing and high throughput genotyping identified community: association with increased birth weight and
KCNJ11 as the thirteenth MODY gene. PLoS One. 2012; doubled risk of type 2 diabetes. Diabetes. 2015;64(12):
7(6):e37423. 4322– 4332.
40. Bowman P, Flanagan SE, Edghill EL, et al. Heterozygous 53. Gloyn AL, Hashim Y, Ashcroft SJ, et al. Association studies
ABCC8 mutations are a cause of MODY. Diabetologia. of variants in promoter and coding regions of ␤-cell ATP-
2012;55(1):123–127. sensitive K-channel genes SUR1 and Kir6.2 with type 2
41. Männikkö R, Flanagan SE, Sim X, et al. Mutations of the diabetes mellitus (UKPDS 53). Diabet Med. 2001;18(3):
same conserved glutamate residue in NBD2 of the sulfo- 206 –212.
nylurea receptor 1 subunit of the KATP channel can result 54. Sesti G, Laratta E, Cardellini M, et al. The E23K variant of

in either hyperinsulinism or neonatal diabetes. Diabetes. KCNJ11 encoding the pancreatic ␤-cell adenosine 5⬘-
2011;60(6):1813–1822. triphosphate-sensitive potassium channel subunit Kir6.2 is
42. Hani EH, Boutin P, Durand E, et al. Missense mutations in associated with an increased risk of secondary failure to
the pancreatic islet ␤ cell inwardly rectifying K⫹ channel sulfonylurea in patients with type 2 diabetes. J Clin Endo-
gene (KIR6.2/BIR): a meta-analysis suggests a role in the crinol Metab. 2006;91(6):2334 –2339.
polygenic basis of type II diabetes mellitus in Caucasians. 55. Proks P, Antcliff JF, Lippiat J, Gloyn AL, Hattersley AT,
Diabetologia. 1998;41(12):1511–1515. Ashcroft FM. Molecular basis of Kir6.2 mutations associ-
43. Gloyn AL, Weedon MN, Owen KR, et al. Large-scale as- ated with neonatal diabetes or neonatal diabetes plus neu-
sociation studies of variants in genes encoding the pancre- rological features. Proc Natl Acad Sci USA. 2004;101(50):
atic ␤-cell KATP channel subunits Kir6.2 (KCNJ11) and 17539 –17544.
SUR1 (ABCC8) confirm that the KCNJ11 E23K variant is 56. Proks P, Shimomura K, Craig TJ, Girard CA, Ashcroft FM.
associated with type 2 diabetes. Diabetes. 2003;52(2): Mechanism of action of a sulphonylurea receptor SUR1
568 –572. mutation (F132L) that causes DEND syndrome. Hum Mo-
44. Love-Gregory L, Wasson J, Lin J, Skolnick G, Suarez B, lec Genet. 2007;16(16):2011–2019.
Permutt MA. E23K single nucleotide polymorphism in the 57. Schwanstecher C, Neugebauer B, Schulz M, Schwan-
islet ATP-sensitive potassium channel gene (Kir6.2) con- stecher M. The common single nucleotide polymorphism
tributes as much to the risk of type II diabetes in Caucasians E23K in K(IR)6.2 sensitizes pancreatic ␤-cell ATP-sensitive
as the PPAR␥ Pro12Ala variant. Diabetologia. 2003;46(1): potassium channels toward activation through nucleoside
136 –137. diphosphates. Diabetes. 2002;51(suppl 3):S363–S367.
45. Nielsen EM, Hansen L, Carstensen B, et al. The E23K 58. Riedel MJ, Boora P, Steckley D, de Vries G, Light PE.
variant of Kir6.2 associates with impaired post-OGTT se- Kir6.2 polymorphisms sensitize ␤-cell ATP-sensitive po-
rum insulin response and increased risk of type 2 diabetes. tassium channels to activation by acyl CoAs: a possible
Diabetes. 2003;52(2):573–577. cellular mechanism for increased susceptibility to type 2
46. Florez JC, Burtt N, de Bakker PI, et al. Haplotype structure diabetes? Diabetes. 2003;52(10):2630 –2635.
and genotype-phenotype correlations of the sulfonylurea 59. Rosengren AH, Braun M, Mahdi T, et al. Reduced insulin
receptor and the islet ATP-sensitive potassium channel exocytosis in human pancreatic ␤-cells with gene variants
gene region. Diabetes. 2004;53(5):1360 –1368. linked to type 2 diabetes. Diabetes. 2012;61(7):1726 –1733.
47. Florez JC, Jablonski KA, Kahn SE, et al. Type 2 diabetes- 60. Fatehi M, Raja M, Carter C, Soliman D, Holt A, Light PE.
associated missense polymorphisms KCNJ11 E23K and The ATP-sensitive K(⫹) channel ABCC8 S1369A type 2
ABCC8 A1369S influence progression to diabetes and re- diabetes risk variant increases MgATPase activity. Diabe-
sponse to interventions in the Diabetes Prevention Pro- tes. 2012;61(1):241–249.
gram. Diabetes. 2007;56(2):531–536. 61. Pearson ER, Flechtner I, Njølstad PR, et al. Switching from
48. Patanè G, Piro S, Rabuazzo AM, Anello M, Vigneri R, insulin to oral sulfonylureas in patients with diabetes due
Purrello F. Metformin restores insulin secretion altered by to Kir6.2 mutations. N Engl J Med. 2006;355(5):467– 477.
chronic exposure to free fatty acids or high glucose: a direct 62. Thurber BW, Carmody D, Tadie EC, et al. Age at the time
metformin effect on pancreatic ␤-cells. Diabetes. 2000; of sulfonylurea initiation influences treatment outcomes in
49(5):735–740. KCNJ11-related neonatal diabetes. Diabetologia. 2015;
49. Lablanche S, Cottet-Rousselle C, Lamarche F, et al. Pro- 58(7):1430 –1435.
tection of pancreatic INS-1 ␤-cells from glucose- and fruc- 63. Holstein A, Hahn M, Stumvoll M, Kovacs P. The E23K
tose-induced cell death by inhibiting mitochondrial per- variant of KCNJ11 and the risk for severe sulfonylurea-
meability transition with cyclosporin A or metformin. Cell induced hypoglycemia in patients with type 2 diabetes.
Death Dis. 2011;2:e134. Horm Metab Res. 2009;41(5):387–390.
50. Laukkanen O, Pihlajamäki J, Lindström J, et al. Polymor- 64. Javorsky M, Klimcakova L, Schroner Z, et al. KCNJ11
phisms of the SUR1 (ABCC8) and Kir6.2 (KCNJ11) genes gene E23K variant and therapeutic response to sulfonyl-
predict the conversion from impaired glucose tolerance to ureas. Eur J Intern Med. 2012;23(3):245–249.
type 2 diabetes. The Finnish Diabetes Prevention Study. 65. Ragia G, Tavridou A, Petridis I, Manolopoulos VG. As-
J Clin Endocrinol Metab. 2004;89(12):6286 – 6290. sociation of KCNJ11 E23K gene polymorphism with hy-
poglycemia in sulfonylurea-treated type 2 diabetic pa- 80. Ingelsson E, Langenberg C, Hivert MF, et al. Detailed phys-
tients. Diabetes Res Clin Pract. 2012;98(1):119 –124. iologic characterization reveals diverse mechanisms for
66. Osbak KK, Colclough K, Saint-Martin C, et al. Update on novel genetic Loci regulating glucose and insulin metabo-
mutations in glucokinase (GCK), which cause maturity- lism in humans. Diabetes. 2010;59(5):1266 –1275.
onset diabetes of the young, permanent neonatal diabetes, 81. Brown KS, Kalinowski SS, Megill JR, Durham SK,
and hyperinsulinemic hypoglycemia. Hum Mutat. 2009; Mookhtiar KA. Glucokinase regulatory protein may inter-
30(11):1512–1526. act with glucokinase in the hepatocyte nucleus. Diabetes.
67. Postic C, Shiota M, Niswender KD, et al. Dual roles for 1997;46(2):179 –186.
glucokinase in glucose homeostasis as determined by liver 82. de la Iglesia N, Veiga-da-Cunha M, Van Schaftingen E,
and pancreatic ␤ cell-specific gene knock-outs using Cre Guinovart JJ, Ferrer JC. Glucokinase regulatory protein is
recombinase. J Biol Chem. 1999;274(1):305–315. essential for the proper subcellular localisation of liver glu-
68. Grupe A, Hultgren B, Ryan A, Ma YH, Bauer M, Stewart cokinase. FEBS Lett. 1999;456(2):332–338.
TA. Transgenic knockouts reveal a critical requirement for 83. Guillam MT, Hümmler E, Schaerer E, et al. Early diabetes

pancreatic ␤ cell glucokinase in maintaining glucose ho- and abnormal postnatal pancreatic islet development in
meostasis. Cell. 1995;83(1):69 –78. mice lacking Glut-2. Nat Genet. 1997;17(3):327–330.
69. Byrne MM, Sturis J, Clément K, et al. Insulin secretory 84. Sansbury FH, Flanagan SE, Houghton JA, et al. SLC2A2
abnormalities in subjects with hyperglycemia due to glu- mutations can cause neonatal diabetes, suggesting GLUT2
cokinase mutations. J Clin Invest. 1994;93(3):1120 –1130. may have a role in human insulin secretion. Diabetologia.
70. Steele AM, Shields BM, Wensley KJ, Colclough K, Ellard S, 2012;55(9):2381–2385.
Hattersley AT. Prevalence of vascular complications among 85. De Vos A, Heimberg H, Quartier E, et al. Human and rat
patients with glucokinase mutations and prolonged, mild hy- ␤ cells differ in glucose transporter but not in glucokinase
perglycemia. JAMA. 2014;311(3):279 –286. gene expression. J Clin Invest. 1995;96(5):2489 –2495.
71. Rudland VL, Hinchcliffe M, Pinner J, et al. Identifying 86. Santer R, Schneppenheim R, Dombrowski A, Götze H,
glucokinase monogenic diabetes in a multiethnic gesta- Steinmann B, Schaub J. Mutations in GLUT2, the gene for
tional diabetes mellitus cohort: new pregnancy screening the liver-type glucose transporter, in patients with Fanconi-
criteria and utility of HbA1c. Diabetes Care. 2016;39(1): Bickel syndrome. Nat Genet. 1997;17(3):324 –326.
50 –52. 87. Abbasi F, Azizi F, Javaheri M, Mosallanejad A, Ebrahim-
72. Chakera AJ, Spyer G, Vincent N, Ellard S, Hattersley AT, Habibi A, Ghafouri-Fard S. Segregation of a novel homozy-
Dunne FP. The 0.1% of the population with glucokinase gous 6 nucleotide deletion in GLUT2 gene in a Fanconi-Bickel
monogenic diabetes can be recognized by clinical charac- syndrome family. Gene. 2015;557(1):103–105.
teristics in pregnancy: the Atlantic Diabetes in Pregnancy 88. Mihout F, Devuyst O, Bensman A, et al. Acute metabolic
cohort. Diabetes Care. 2014;37(5):1230 –1236. acidosis in a GLUT2-deficient patient with Fanconi-Bickel
73. Stride A, Shields B, Gill-Carey O, et al. Cross-sectional and syndrome: new pathophysiology insights. Nephrol Dial
longitudinal studies suggest pharmacological treatment Transplant. 2014;29(suppl 4):iv113–iv116.
used in patients with glucokinase mutations does not alter 89. Mannstadt M, Magen D, Segawa H, et al. Fanconi-Bickel
glycaemia. Diabetologia. 2014;57(1):54 –56. syndrome and autosomal recessive proximal tubulopathy
74. Tornovsky-Babeay S, Dadon D, Ziv O, et al. Type 2 dia- with hypercalciuria (ARPTH) are allelic variants caused by
betes and congenital hyperinsulinism cause DNA double- GLUT2 mutations. J Clin Endocrinol Metab. 2012;97(10):
strand breaks and p53 activity in ␤ cells. Cell Metab. 2014; E1978 –E1986.
19(1):109 –121. 90. Al-Haggar M, Sakamoto O, Shaltout A, Al-Hawari A,
75. Soranzo N, Sanna S, Wheeler E, et al. Common variants at Wahba Y, Abdel-Hadi D. Mutation analysis of the GLUT2
10 genomic loci influence hemoglobin A1(C) levels via gly- gene in three unrelated Egyptian families with Fanconi-
cemic and nonglycemic pathways. Diabetes. 2010;59(12): Bickel syndrome: revisited gene atlas for renumbering. Clin
3229 –3239. Exp Nephrol. 2012;16(4):604 – 610.
76. Qi Q, Wu Y, Li H, et al. Association of GCKR rs780094, 91. Grünert SC, Schwab KO, Pohl M, Sass JO, Santer R. Fanconi-
alone or in combination with GCK rs1799884, with type Bickel syndrome: GLUT2 mutations associated with a mild
2 diabetes and related traits in a Han Chinese population. phenotype. Mol Genet Metab. 2012;105(3):433– 437.
Diabetologia. 2009;52(5):834 – 843. 92. Janssen RC, Bogardus C, Takeda J, Knowler WC, Thomp-
77. Tam CH, Ma RC, So WY, et al. Interaction effect of genetic son DB. Linkage analysis of acute insulin secretion with
polymorphisms in glucokinase (GCK) and glucokinase reg- GLUT2 and glucokinase in Pima Indians and the identifi-
ulatory protein (GCKR) on metabolic traits in healthy Chi- cation of a missense mutation in GLUT2. Diabetes. 1994;
nese adults and adolescents. Diabetes. 2009;58(3):765–769. 43(4):558 –563.
78. Dupuis J, Langenberg C, Prokopenko I, et al. New genetic 93. Alcolado JC, Baroni MG, Li SR. Association between a
loci implicated in fasting glucose homeostasis and their restriction fragment length polymorphism at the liver/islet
impact on type 2 diabetes risk. Nat Genet. 2010;42(2): cell (GluT 2) glucose transporter and familial type 2 (non-
105–116. insulin-dependent) diabetes mellitus. Diabetologia. 1991;
79. Muller YL, Piaggi P, Hoffman D, et al. Common genetic vari- 34(10):734 –736.
ation in the glucokinase gene (GCK) is associated with 94. Barroso I, Luan J, Middelberg RP, et al. Candidate gene
type 2 diabetes and rates of carbohydrate oxidation and association study in type 2 diabetes indicates a role for
energy expenditure. Diabetologia. 2014;57(7):1382– genes involved in ␤-cell function as well as insulin action.
1390. PLoS Biol. 2003;1(1):E20.
95. Matsutani A, Koranyi L, Cox N, Permutt MA. Polymor- secretion in hepatocyte nuclear factor 1␣-deficient mice.
phisms of GLUT2 and GLUT4 genes. Use in evaluation of J Clin Invest. 1998;101(10):2215–2222.
genetic susceptibility to NIDDM in blacks. Diabetes. 1990; 110. Cereghini S. Liver-enriched transcription factors and he-
39(12):1534 –1542. patocyte differentiation. FASEB J. 1996;10(2):267–282.
96. Patel P, Bell GI, Cook JT, Turner RC, Wainscoat JS. Mul- 111. Chen WS, Manova K, Weinstein DC, et al. Disruption of
tiple restriction fragment length polymorphisms at the the HNF-4 gene, expressed in visceral endoderm, leads to
GLUT2 locus: GLUT2 haplotypes for genetic analysis of cell death in embryonic ectoderm and impaired gastrula-
type 2 (non-insulin-dependent) diabetes mellitus. Diabe- tion of mouse embryos. Genes Dev. 1994;8(20):2466 –
tologia. 1991;34(11):817– 821. 2477.
97. Tanizawa Y, Riggs AC, Chiu KC, et al. Variability of the 112. Boj SF, Parrizas M, Maestro MA, Ferrer J. A transcription
pancreatic islet ␤ cell/liver (GLUT 2) glucose transporter factor regulatory circuit in differentiated pancreatic cells.
gene in NIDDM patients. Diabetologia. 1994;37(4):420 – Proc Natl Acad Sci USA. 2001;98(25):14481–14486.
427. 113. Stride A, Ellard S, Clark P, et al. ␤-Cell dysfunction, insulin

98. Matsubara A, Tanizawa Y, Matsutani A, Kaneko T, Kaku sensitivity, and glycosuria precede diabetes in hepatocyte
K. Sequence variations of the pancreatic islet/liver glucose nuclear factor-1␣ mutation carriers. Diabetes Care. 2005;
transporter (GLUT2) gene in Japanese subjects with non- 28(7):1751–1756.
insulin dependent diabetes mellitus. J Clin Endocrinol 114. Stanescu DE, Hughes N, Kaplan B, Stanley CA, De León
Metab. 1995;80(11):3131–3135. DD. Novel presentations of congenital hyperinsulinism
99. Laukkanen O, Lindström J, Eriksson J, et al. Polymor- due to mutations in the MODY genes: HNF1A and
phisms in the SLC2A2 (GLUT2) gene are associated with HNF4A. J Clin Endocrinol Metab. 2012;97(10):E2026 –
the conversion from impaired glucose tolerance to type 2 E2030.
diabetes: the Finnish Diabetes Prevention Study. Diabetes. 115. Shields BM, Hicks S, Shepherd MH, Colclough K, Hat-
2005;54(7):2256 –2260. tersley AT, Ellard S. Maturity-onset diabetes of the young
100. Kilpeläinen TO, Lakka TA, Laaksonen DE, et al. Physical (MODY): how many cases are we missing? Diabetologia.
activity modifies the effect of SNPs in the SLC2A2 2010;53(12):2504 –2508.
(GLUT2) and ABCC8 (SUR1) genes on the risk of devel- 116. Ellard S, Colclough K. Mutations in the genes encoding the
oping type 2 diabetes. Physiol Genomics. 2007;31(2): transcription factors hepatocyte nuclear factor 1 ␣
264 –272. (HNF1A) and 4 ␣ (HNF4A) in maturity-onset diabetes of
101. Michau A, Guillemain G, Grosfeld A, et al. Mutations in the young. Hum Mutat. 2006;27(9):854 – 869.
SLC2A2 gene reveal hGLUT2 function in pancreatic ␤ cell 117. Pontoglio M, Prié D, Cheret C, et al. HNF1␣ controls renal
development. J Biol Chem. 2013;288(43):31080 –31092. glucose reabsorption in mouse and man. EMBO Rep.
102. Burgdorf KS, Grarup N, Justesen JM, et al. Studies of the 2000;1(4):359 –365.
association of Arg72Pro of tumor suppressor protein p53 118. Hegele RA, Cao H, Harris SB, Hanley AJ, Zinman B. The
with type 2 diabetes in a combined analysis of 55,521 Eu- hepatic nuclear factor-1␣ G319S variant is associated with
ropeans. PLoS One. 2011;6(1):e15813. early-onset type 2 diabetes in Canadian Oji-Cree. J Clin
103. Willer CJ, Bonnycastle LL, Conneely KN, et al. Screening Endocrinol Metab. 1999;84(3):1077–1082.
of 134 single nucleotide polymorphisms (SNPs) previously 119. SIGMA Type 2 Diabetes Consortium, Estrada K, Aukrust
associated with type 2 diabetes replicates association with I, et al. Association of a low-frequency variant in HNF1A
12 SNPs in nine genes. Diabetes. 2007;56(1):256 –264. with type 2 diabetes in a Latino population. JAMA. 2014;
104. Eny KM, Wolever TM, Fontaine-Bisson B, El-Sohemy A. 311(22):2305–2314.
Genetic variant in the glucose transporter type 2 is asso- 120. Andersen MK, Sterner M, Forsén T, et al. Type 2 diabetes
ciated with higher intakes of sugars in two distinct popu- susceptibility gene variants predispose to adult-onset auto-
lations. Physiol Genomics. 2008;33(3):355–360. immune diabetes. Diabetologia. 2014;57(9):1859 –1868.
105. Burcelin R, Dolci W, Thorens B. Glucose sensing by the 121. Silander K, Mohlke KL, Scott LJ, et al. Genetic variation
hepatoportal sensor is GLUT2-dependent: in vivo analysis near the hepatocyte nuclear factor-4 ␣ gene predicts sus-
in GLUT2-null mice. Diabetes. 2000;49(10):1643–1648. ceptibility to type 2 diabetes. Diabetes. 2004;53(4):1141–
106. Roncero I, Alvarez E, Chowen JA, et al. Expression of 1149.
glucose transporter isoform GLUT-2 and glucokinase 122. Love-Gregory LD, Wasson J, Ma J, et al. A common poly-
genes in human brain. J Neurochem. 2004;88(5):1203– morphism in the upstream promoter region of the hepa-
1210. tocyte nuclear factor-4 ␣ gene on chromosome 20q is as-
107. Li B, Xi X, Roane DS, Ryan DH, Martin RJ. Distribution sociated with type 2 diabetes and appears to contribute to
of glucokinase, glucose transporter GLUT2, sulfonylurea the evidence for linkage in an Ashkenazi Jewish popula-
receptor-1, glucagon-like peptide-1 receptor and neuro- tion. Diabetes. 2004;53(4):1134 –1140.
peptide Y messenger RNAs in rat brain by quantitative real 123. Weedon MN, Owen KR, Shields B, et al. Common variants
time RT-PCR. Brain Res Mol Brain Res. 2003;113:139 – of the hepatocyte nuclear factor-4␣ P2 promoter are asso-
142. ciated with type 2 diabetes in the U.K. population. Diabe-
108. Pontoglio M, Barra J, Hadchouel M, et al. Hepatocyte tes. 2004;53(11):3002–3006.
nuclear factor 1 inactivation results in hepatic dysfunction, 124. Barroso I, Luan J, Wheeler E, et al. Population-specific risk
phenylketonuria, and renal Fanconi syndrome. Cell. 1996; of type 2 diabetes conferred by HNF4A P2 promoter vari-
84(4):575–585. ants: a lesson for replication studies. Diabetes. 2008;
109. Pontoglio M, Sreenan S, Roe M, et al. Defective insulin 57(11):3161–3165.
125. Bjørkhaug L, Sagen JV, Thorsby P, Søvik O, Molven A, onset diabetes but no known renal disease. Diabet Med.
Njølstad PR. Hepatocyte nuclear factor-1 ␣ gene muta- 2013;30(1):114 –117.
tions and diabetes in Norway. J Clin Endocrinol Metab. 140. Haldorsen IS, Vesterhus M, Raeder H, et al. Lack of pan-
2003;88(2):920 –931. creatic body and tail in HNF1B mutation carriers. Diabet
126. Nowak N, Hohendorff J, Solecka I, et al. Circulating ghre- Med. 2008;25(7):782–787.
lin level is higher in HNF1A-MODY and GCK-MODY 141. Winckler W, Weedon MN, Graham RR, et al. Evaluation
than in polygenic forms of diabetes mellitus. Endocrine. of common variants in the six known maturity-onset dia-
2015;50(3):643– 649. betes of the young (MODY) genes for association with type
127. Brial F, Lussier CR, Belleville K, Sarret P, Boudreau F. 2 diabetes. Diabetes. 2007;56(3):685– 693.
Ghrelin inhibition restores glucose homeostasis in hepato- 142. Rodriguez C, Patel AV, Mondul AM, Jacobs EJ, Thun MJ,
cyte nuclear factor-1␣ (MODY3)-deficient mice. Diabetes. Calle EE. Diabetes and risk of prostate cancer in a pro-
2015;64(9):3314 –3320. spective cohort of US men. Am J Epidemiol. 2005;161(2):
128. Dezaki K, Kakei M, Yada T. Ghrelin uses G␣i2 and acti- 147–152.

vates voltage-dependent K⫹ channels to attenuate glucose- 143. Giovannucci E, Harlan DM, Archer MC, et al. Diabetes
induced Ca2⫹ signaling and insulin release in islet ␤-cells: and cancer: a consensus report. Diabetes Care. 2010;
novel signal transduction of ghrelin. Diabetes. 2007;56(9): 33(7):1674 –1685.
2319 –2327. 144. Gudmundsson J, Sulem P, Steinthorsdottir V, et al. Two
129. Eeckhoute J, Moerman E, Bouckenooghe T, et al. Hepa- variants on chromosome 17 confer prostate cancer risk,
tocyte nuclear factor 4 ␣ isoforms originated from the P1 and the one in TCF2 protects against type 2 diabetes. Nat
promoter are expressed in human pancreatic ␤-cells and Genet. 2007;39(8):977–983.
exhibit stronger transcriptional potentials than P2 promot- 145. Barbacci E, Chalkiadaki A, Masdeu C, et al. HNF1␤/TCF2
er-driven isoforms. Endocrinology. 2003;144(5):1686 – mutations impair transactivation potential through altered
1694. co-regulator recruitment. Hum Molec Genet. 2004;
130. Gupta RK, Vatamaniuk MZ, Lee CS, et al. The MODY1 13(24):3139 –3149.
146. Haumaitre C, Fabre M, Cormier S, Baumann C, Delezoide
gene HNF-4␣ regulates selected genes involved in insulin
AL, Cereghini S. Severe pancreas hypoplasia and multi-
secretion. J Clin Invest. 2005;115(4):1006 –1015.
cystic renal dysplasia in two human fetuses carrying novel
131. Sandovici I, Smith NH, Nitert MD, et al. Maternal diet and
HNF1␤/MODY5 mutations. Hum Molec Genet. 2006;
aging alter the epigenetic control of a promoter-enhancer
15(15):2363–2375.
interaction at the Hnf4a gene in rat pancreatic islets. Proc
147. Cha JY, Kim H, Kim KS, Hur MW, Ahn Y. Identification
Natl Acad Sci USA. 2011;108(13):5449 –5454.
of transacting factors responsible for the tissue-specific ex-
132. Coffinier C, Thépot D, Babinet C, Yaniv M, Barra J.
pression of human glucose transporter type 2 isoform gene.
Essential role for the homeoprotein vHNF1/HNF1␤ in vis-
Cooperative role of hepatocyte nuclear factors 1␣ and 3␤.
ceral endoderm differentiation. Development. 1999;
J Biol Chem. 2000;275(24):18358 –18365.
126(21):4785– 4794.
148. Kim EK, Lee JS, Cheong HI, Chung SS, Kwak SH, Park KS.
133. Barbacci E, Reber M, Ott MO, Breillat C, Huetz F, Ce-
Identification and functional characterization of P159L
reghini S. Variant hepatocyte nuclear factor 1 is required mutation in HNF1B in a family with maturity-onset dia-
for visceral endoderm specification. Development. 1999; betes of the young 5 (MODY5). Genomics Inform. 2014;
126(21):4795– 4805. 12(4):240 –246.
134. Wang L, Coffinier C, Thomas MK, et al. Selective deletion 149. Yamagata K, Nammo T, Moriwaki M, et al. Overexpres-
of the Hnf1␤ (MODY5) gene in ␤-cells leads to altered gene sion of dominant-negative mutant hepatocyte nuclear fc-
expression and defective insulin release. Endocrinology. tor-1 ␣ in pancreatic ␤-cells causes abnormal islet archi-
2004;145(8):3941–3949. tecture with decreased expression of E-cadherin, reduced
135. De Vas MG, Kopp JL, Heliot C, Sander M, Cereghini S, ␤-cell proliferation, and diabetes. Diabetes. 2002;51(1):
Haumaitre C. Hnf1b controls pancreas morphogenesis 114 –123.
and the generation of Ngn3⫹ endocrine progenitors. De- 150. Dessimoz J, Bonnard C, Huelsken J, Grapin-Botton A.
velopment. 2015;142(5):871– 882. Pancreas-specific deletion of ␤-catenin reveals Wnt-depen-
136. Poll AV, Pierreux CE, Lokmane L, et al. A vHNF1/TCF2- dent and Wnt-independent functions during development.
HNF6 cascade regulates the transcription factor network Curr Biol. 2005;15(18):1677–1683.
that controls generation of pancreatic precursor cells. Di- 151. Papadopoulou S, Edlund H. Attenuated Wnt signaling per-
abetes. 2006;55(1):61– 69. turbs pancreatic growth but not pancreatic function. Di-
137. Murphy R, Ellard S, Hattersley AT. Clinical implications abetes. 2005;54(10):2844 –2851.
of a molecular genetic classification of monogenic ␤-cell 152. Murtaugh LC, Law AC, Dor Y, Melton DA. ␤-Catenin is
diabetes. Nat Clin Pract Endocrinol Metab. 2008;4(4): essential for pancreatic acinar but not islet development.
200 –213. Development. 2005;132(21):4663– 4674.
138. Chen YZ, Gao Q, Zhao XZ, et al. Systematic review of 153. Oliver-Krasinski JM, Stoffers DA. On the origin of the ␤
TCF2 anomalies in renal cysts and diabetes syndrome/ma- cell. Genes Dev. 2008;22(15):1998 –2021.
turity onset diabetes of the young type 5. Chin Med J. 154. Oliver-Krasinski JM, Kasner MT, Yang J, et al. The dia-
2010;123(22):3326 –3333. betes gene Pdx1 regulates the transcriptional network of
139. Edghill EL, Stals K, Oram RA, Shepherd MH, Hattersley pancreatic endocrine progenitor cells in mice. J Clin Invest.
AT, Ellard S. HNF1B deletions in patients with young- 2009;119(7):1888 –1898.
155. Ahlgren U, Jonsson J, Jonsson L, Simu K, Edlund H. susceptibility genes Pdx1 and Clec16a function in a path-
␤-Cell-specific inactivation of the mouse Ipf1/Pdx1 gene way regulating mitophagy in ␤-cells. Diabetes. 2015;
results in loss of the ␤-cell phenotype and maturity onset 64(10):3475–3484.
diabetes. Genes Dev. 1998;12(12):1763–1768. 172. Smith SB, Ee HC, Conners JR, German MS. Paired-home-
156. Jonsson J, Carlsson L, Edlund T, Edlund H. Insulin-pro- odomain transcription factor PAX4 acts as a transcrip-
moter-factor 1 is required for pancreas development in tional repressor in early pancreatic development. Molec
mice. Nature. 1994;371(6498):606 – 609. Cell Biol. 1999;19(12):8272– 8280.
157. Gao T, McKenna B, Li C, et al. Pdx1 maintains ␤ cell 173. Brun T, Gauthier BR. A focus on the role of Pax4 in mature
identity and function by repressing an ␣ cell program. Cell pancreatic islet ␤-cell expansion and survival in health and
Metab. 2014;19(2):259 –271. disease. J Mol Endocrinol. 2008;40(2):37– 45.
158. Stoffers DA, Zinkin NT, Stanojevic V, Clarke WL, Ha- 174. Sosa-Pineda B, Chowdhury K, Torres M, Oliver G, Gruss
bener JF. Pancreatic agenesis attributable to a single nu- P. The Pax4 gene is essential for differentiation of insulin-
cleotide deletion in the human IPF1 gene coding sequence. producing ␤ cells in the mammalian pancreas. Nature.

Nat Genet. 1997;15(1):106 –110. 1997;386(6623):399 – 402.
159. Wright NM, Metzger DL, Borowitz SM, Clarke WL. Per- 175. Collombat P, Mansouri A, Hecksher-Sorensen J, et al. Op-
manent neonatal diabetes mellitus and pancreatic exocrine posing actions of Arx and Pax4 in endocrine pancreas de-
insufficiency resulting from congenital pancreatic agene- velopment. Genes Dev. 2003;17(20):2591–2603.
sis. Am J Dis Child. 1993;147(6):607– 609. 176. Collombat P, Hecksher-Sørensen J, Broccoli V, et al. The
160. Schwitzgebel VM, Mamin A, Brun T, et al. Agenesis of simultaneous loss of Arx and Pax4 genes promotes a so-
human pancreas due to decreased half-life of insulin pro- matostatin-producing cell fate specification at the expense
moter factor 1. J Clin Endocrinol Metab. 2003;88(9): of the ␣- and ␤-cell lineages in the mouse endocrine pan-
4398 – 4406. creas. Development. 2005;132(13):2969 –2980.
161. Nicolino M, Claiborn KC, Senée V, Boland A, Stoffers DA, 177. Plengvidhya N, Kooptiwut S, Songtawee N, et al. PAX4
Julier C. A novel hypomorphic PDX1 mutation responsi- mutations in Thais with maturity onset diabetes of the
ble for permanent neonatal diabetes with subclinical exo- young. J Clin Endocrinol Metab. 2007;92(7):2821–2826.
crine deficiency. Diabetes. 2010;59(3):733–740. 178. Kooptiwut S, Plengvidhya N, Chukijrungroat T, et al. De-
162. Stoffers DA, Ferrer J, Clarke WL, Habener JF. Early-onset fective PAX4 R192H transcriptional repressor activities
type-II diabetes mellitus (MODY4) linked to IPF1. Nat associated with maturity onset diabetes of the young and
Genet. 1997;17(2):138 –139. early onset-age of type 2 diabetes. J Diabetes Complica-
163. Weng J, Macfarlane WM, Lehto M, et al. Functional con- tions. 2012;26(4):343–347.
sequences of mutations in the MODY4 gene (IPF1) and 179. Jo W, Endo M, Ishizu K, Nakamura A, Tajima T. A novel
coexistence with MODY3 mutations. Diabetologia. 2001; PAX4 mutation in a Japanese patient with maturity-onset
44(2):249 –258. diabetes of the young. Tohoku J Exp Med. 2011;223(2):
164. Hani EH, Stoffers DA, Chèvre JC, et al. Defective muta- 113–118.
tions in the insulin promoter factor-1 (IPF-1) gene in late- 180. Shimajiri Y, Sanke T, Furuta H, et al. A missense mutation
onset type 2 diabetes mellitus. J Clin Invest. 1999;104(9): of Pax4 gene (R121W) is associated with type 2 diabetes in
R41–R48. Japanese. Diabetes. 2001;50(12):2864 –2869.
165. Macfarlane WM, Frayling TM, Ellard S, et al. Missense 181. Ma RC, Hu C, Tam CH, et al. Genome-wide association
mutations in the insulin promoter factor-1 gene predispose study in a Chinese population identifies a susceptibility
to type 2 diabetes. J Clin Invest. 1999;104(9):R33–R39. locus for type 2 diabetes at 7q32 near PAX4. Diabetologia.
166. Edghill EL, Khamis A, Weedon MN, et al. Sequencing 2013;56(6):1291–1305.
PDX1 (insulin promoter factor 1) in 1788 UK individuals 182. Chen M, Hu C, Zhang R, et al. Association of PAX4 ge-
found 5% had a low frequency coding variant, but these netic variants with oral antidiabetic drugs efficacy in Chi-
variants are not associated with Type 2 diabetes. Diabet nese type 2 diabetes patients. Pharmacogenomics J. 2014;
Med. 2011;28(6):681– 684. 14(5):488 – 492.
167. Steinthorsdottir V, Thorleifsson G, Sulem P, et al. Identi- 183. Biason-Lauber A, Boehm B, Lang-Muritano M, et al. As-
fication of low-frequency and rare sequence variants association of childhood type 1 diabetes mellitus with a vari-
sociated with elevated or reduced risk of type 2 diabetes. ant of PAX4: possible link to ␤ cell regenerative capacity.
Nat Genet. 2014;46(3):294 –298. Diabetologia. 2005;48(5):900 –905.
168. Sachdeva MM, Claiborn KC, Khoo C, et al. Pdx1 184. Holm P, Rydlander B, Luthman H, Kockum I. Interaction
(MODY4) regulates pancreatic ␤ cell susceptibility to ER and association analysis of a type 1 diabetes susceptibility
stress. Proc Natl Acad Sci USA. 2009;106(45):19090 – locus on chromosome 5q11– q13 and the 7q32 chromo-
19095. somal region in Scandinavian families. Diabetes. 2004;
169. Szendroedi J, Phielix E, Roden M. The role of mitochon- 53(6):1584 –1591.
dria in insulin resistance and type 2 diabetes mellitus. Nat 185. Hermann R, Mantere J, Lipponen K, et al. Lack of asso-
Rev Endocrinol. 2012;8(2):92–103. ciation of PAX4 gene with type 1 diabetes in the Finnish
170. Sivitz WI, Yorek MA. Mitochondrial dysfunction in dia- and Hungarian populations. Diabetes. 2005;54(9):2816 –
betes: from molecular mechanisms to functional signifi- 2819.
cance and therapeutic opportunities. Antioxid Redox Sig- 186. Zhang Y, Xiao X, Liu Y, et al. The association of the PAX4
nal. 2010;12(4):537–577. gene with type 1 diabetes in Han Chinese. Diabetes Res
171. Soleimanpour SA, Ferrari AM, Raum JC, et al. Diabetes Clin Pract. 2008;81(3):365–369.
187. Mauvais-Jarvis F, Smith SB, Le May C, et al. PAX4 gene ROD1 are responsible for a novel syndrome of permanent
variations predispose to ketosis-prone diabetes. Hum Mo- neonatal diabetes and neurological abnormalities. Diabe-
lec Genet. 2004;13(24):3151–3159. tes. 2010;59(9):2326 –2331.
188. Balasubramanyam A, Nalini R, Hampe CS, Maldonado 204. Malecki MT, Jhala US, Antonellis A, et al. Mutations in
M. Syndromes of ketosis-prone diabetes mellitus. Endocr NEUROD1 are associated with the development of type 2
Rev. 2008;29(3):292–302. diabetes mellitus. Nat Genet. 1999;23(3):323–328.
189. Patel SG, Hsu JW, Jahoor F, et al. Pathogenesis of A⫺␤⫹ 205. Liu L, Jia W, Zheng T, Li M, Lu H, Xiang K. Ala45Thr
ketosis-prone diabetes. Diabetes. 2013;62(3):912–922. variation in neuroD1 gene is associated with early-onset
190. Haaland WC, Scaduto DI, Maldonado MR, et al. A-␤- type 2 diabetes with or without diabetic pedigree in Chi-
subtype of ketosis-prone diabetes is not predominantly a nese. Mol Cell Biochem. 2006;290:199 –204.
monogenic diabetic syndrome. Diabetes Care. 2009;32(5): 206. Malecki MT, Cyganek K, Klupa T, Sieradzki J. The
873– 877. Ala45Thr polymorphism of BETA2/NeuroD1 gene and
191. Brun T, Franklin I, St-Onge L, et al. The diabetes-linked susceptibility to type 2 diabetes mellitus in a Polish popu-

transcription factor PAX4 promotes ␤-cell proliferation lation. Acta Diabetol. 2003;40(2):109 –111.
and survival in rat and human islets. J Cell Biol. 2004; 207. Han X, Xiao J, Ren Q, Tang Y, Yang W, Ji L. Evaluation of
167(6):1123–1135. variant A45T in NEUROD1/BETA2 for its association with
192. Docena MK, Faiman C, Stanley CM, Pantalone KM. type 2 diabetes mellitus. Endocrine. 2013;44(1):99 –106.
Mody-3: novel HNF1A mutation and the utility of gluca- 208. Hansen L, Jensen JN, Urioste S, et al. NeuroD/BETA2 gene
gon-like peptide (GLP)-1 receptor agonist therapy. Endocr variability and diabetes: no associations to late-onset type
Pract. 2014;20(2):107–111. 2 diabetes but an A45 allele may represent a susceptibility
193. Østoft SH, Bagger JI, Hansen T, et al. Glucose-lowering marker for type 1 diabetes among Danes. Danish Study
effects and low risk of hypoglycemia in patients with ma- Group of Diabetes in Childhood, and the Danish IDDM
turity-onset diabetes of the young when treated with a Epidemiology and Genetics Group. Diabetes. 2000;49(5):
GLP-1 receptor agonist: a double-blind, randomized, 876 – 878.
crossover trial. Diabetes Care. 2014;37(7):1797–1805. 209. Malecki MT, Klupa T, Moczulski DK, Rogus JJ. The
194. Lumb AN, Gallen IW. Treatment of HNF1-␣ MODY with Ala45Thr polymorphism of BETA2/NeuroD1 gene and
the DPP-4 inhibitor sitagliptin(1). Diabet Med. 2009; susceptibility to type 1 diabetes mellitus in Caucasians.
26(2):189 –190. Exp Clin Endocrinol Diabetes. 2003;111(5):251–254.
195. Katra B, Klupa T, Skupien J, et al. Dipeptidyl peptidase-IV 210. Oliveira CS, Hauache OM, Vieira JG, et al. The Ala45Thr
inhibitors are efficient adjunct therapy in HNF1A matu- polymorphism of NEUROD1 is associated with type 1 di-
rity-onset diabetes of the young patients–report of two abetes in Brazilian women. Diabetes Metab. 2005;31(6):
cases. Diabetes Technol Ther. 2010;12(4):313–316. 599 – 602.
196. Lefèbvre PJ, Paquot N, Scheen AJ. Inhibiting or antago- 211. Vella A, Howson JM, Barratt BJ, et al. Lack of association
nizing glucagon: making progress in diabetes care. Diabe- of the Ala(45)Thr polymorphism and other common vari-
tes Obes Metab. 2015;17(8):720 –725. ants of the NeuroD gene with type 1 diabetes. Diabetes.
197. Naya FJ, Huang HP, Qiu Y, et al. Diabetes, defective pan- 2004;53(4):1158 –1161.
creatic morphogenesis, and abnormal enteroendocrine dif- 212. Naya FJ, Stellrecht CM, Tsai MJ. Tissue-specific regula-
ferentiation in BETA2/neuroD-deficient mice. Genes Dev. tion of the insulin gene by a novel basic helix-loop-helix
1997;11(18):2323–2334. transcription factor. Genes Dev. 1995;9(8):1009 –1019.
198. Gu C, Stein GH, Pan N, et al. Pancreatic ␤ cells require 213. Zhao L, Guo M, Matsuoka TA, et al. The islet ␤ cell-
NeuroD to achieve and maintain functional maturity. Cell enriched MafA activator is a key regulator of insulin gene
Metab. 2010;11(4):298 –310. transcription. J Biol Chem. 2005;280(12):11887–11894.
199. Kristinsson SY, Thorolfsdottir ET, Talseth B, et al. 214. Yang Y, Chang BH, Samson SL, Li MV, Chan L. The Krüp-
MODY in Iceland is associated with mutations in HNF-1␣ pel-like zinc finger protein Glis3 directly and indirectly
and a novel mutation in NeuroD1. Diabetologia. 2001; activates insulin gene transcription. Nucleic Acids Res.
44(11):2098 –2103. 2009;37(8):2529 –2538.
200. Liu L, Furuta H, Minami A, et al. A novel mutation, 215. Fonseca SG, Fukuma M, Lipson KL, et al. WFS1 is a novel
Ser159Pro in the NeuroD1/BETA2 gene contributes to the component of the unfolded protein response and maintains
development of diabetes in a Chinese potential MODY homeostasis of the endoplasmic reticulum in pancreatic
family. Mol Cell Biochem. 2007;303:115–120. ␤-cells. J Biol Chem. 2005;280(47):39609 –39615.
201. Gonsorcíková L, Průhová S, Cinek O, et al. Autosomal 216. Rohayem J, Ehlers C, Wiedemann B, et al. Diabetes and
inheritance of diabetes in two families characterized by neurodegeneration in Wolfram syndrome: a multicenter
obesity and a novel H241Q mutation in NEUROD1. Pe- study of phenotype and genotype. Diabetes Care. 2011;
diatr Diabetes. 2008;9:367–372. 34(7):1503–1510.
202. Szopa M, Ludwig-Galezowska AH, Radkowski P, et al. A 217. Fonseca SG, Ishigaki S, Oslowski CM, et al. Wolfram syn-
family with the Arg103Pro mutation in the NEUROD1 drome 1 gene negatively regulates ER stress signaling in ro-
gene detected by next-generation sequencing-clinical char- dent and human cells. J Clin Invest. 2010;120(3):744 –755.
acteristics of mutation carriers. Eur J Med Genet. 2016; 218. Zalloua PA, Azar ST, Delépine M, et al. WFS1 mutations
59(2):75–79. are frequent monogenic causes of juvenile-onset diabetes
203. Rubio-Cabezas O, Minton JA, Kantor I, Williams D, El- mellitus in Lebanon. Hum Molec Genet. 2008;17(24):
lard S, Hattersley AT. Homozygous mutations in NEU- 4012– 4021.
219. Zmyslowska A, Fendler W, Niwald A, et al. Retinal thin- 237. Banks AS, McAllister FE, Camporez JP, et al. An ERK/
ning as a marker of disease progression in patients with Cdk5 axis controls the diabetogenic actions of PPAR␥.
Wolfram syndrome. Diabetes Care. 2015;38(3):e36 – e37. Nature. 2015;517(7534):391–395.
220. Rigoli L, Lombardo F, Di Bella C. Wolfram syndrome and 238. Barroso I, Gurnell M, Crowley VE, et al. Dominant neg-
WFS1 gene. Clin Genet. 2011;79(2):103–117. ative mutations in human PPAR␥ associated with severe
221. Amr S, Heisey C, Zhang M, et al. A homozygous mutation insulin resistance, diabetes mellitus and hypertension. Na-
in a novel zinc-finger protein, ERIS, is responsible for Wol- ture. 1999;402(6764):880 – 883.
fram syndrome 2. Am J Hum Genet. 2007;81(4):673– 683. 239. Yen CJ, Beamer BA, Negri C, et al. Molecular scanning of
222. Rigoli L, Di Bella C. Wolfram syndrome 1 and Wolfram the human peroxisome proliferator activated receptor ␥
syndrome 2. Curr Opin Pediatr. 2012;24(4):512–517. (hPPAR ␥) gene in diabetic Caucasians: identification of a
223. Minton JA, Hattersley AT, Owen K, et al. Association studies Pro12Ala PPAR ␥ 2 missense mutation. Biochem Biophys
of genetic variation in the WFS1 gene and type 2 diabetes in Res Commun. 1997;241(2):270 –274.
U.K. populations. Diabetes. 2002;51(4):1287–1290. 240. Altshuler D, Hirschhorn JN, Klannemark M, et al. The

224. Sandhu MS, Weedon MN, Fawcett KA, et al. Common common PPAR␥ Pro12Ala polymorphism is associated
variants in WFS1 confer risk of type 2 diabetes. Nat Genet. with decreased risk of type 2 diabetes. Nat Genet. 2000;
2007;39(8):951–953. 26(1):76 – 80.
225. Cheurfa N, Brenner GM, Reis AF, et al. Decreased insulin 241. Soriguer F, Morcillo S, Cardona F, et al. Pro12Ala poly-
secretion and increased risk of type 2 diabetes associated morphism of the PPARG2 gene is associated with type 2
with allelic variations of the WFS1 gene: the Data from diabetes mellitus and peripheral insulin sensitivity in a pop-
Epidemiological Study on the Insulin Resistance Syndrome ulation with a high intake of oleic acid. J Nutr. 2006;
(DESIR) prospective study. Diabetologia. 2011;54(3): 136(9):2325–2330.
554 –562. 242. Stumvoll M, Häring H. The peroxisome proliferator-acti-
226. Fawcett KA, Wheeler E, Morris AP, et al. Detailed investiga- vated receptor-␥2 Pro12Ala polymorphism. Diabetes.
tion of the role of common and low-frequency WFS1 variants 2002;51(8):2341–2347.
in type 2 diabetes risk. Diabetes. 2010;59(3):741–746. 243. Gouda HN, Sagoo GS, Harding AH, Yates J, Sandhu MS,
227. Awata T, Inoue K, Kurihara S, et al. Missense variations of Higgins JP. The association between the peroxisome pro-
the gene responsible for Wolfram syndrome (WFS1/wol- liferator-activated receptor-␥2 (PPARG2) Pro12Ala gene
framin) in Japanese: possible contribution of the variant and type 2 diabetes mellitus: a HuGE review and
Arg456His mutation to type 1 diabetes as a nonautoim- meta-analysis. Am J Epidemiol. 2010;171(6):645– 655.
mune genetic basis. Biochem Biophys Res Commun. 2000; 244. Raj SM, Howson JM, Walker NM, et al. No association of
268(2):612– 616. multiple type 2 diabetes loci with type 1 diabetes. Diabe-
228. Lemaire K, Schuit F. Integrating insulin secretion and ER tologia. 2009;52(10):2109 –2116.
stress in pancreatic ␤-cells. Nat Cell Biol. 2012;14(10): 245. Qu HQ, Grant SF, Bradfield JP, et al. Association analysis
979 –981. of type 2 diabetes loci in type 1 diabetes. Diabetes. 2008;
229. Lu S, Kanekura K, Hara T, et al. A calcium-dependent pro- 57(7):1983–1986.
tease as a potential therapeutic target for Wolfram syndrome. 246. Soccio RE, Chen ER, Rajapurkar SR, et al. Genetic vari-
Proc Natl Acad Sci USA. 2014;111(49):E5292–E5301. ation determines PPAR␥ function and anti-diabetic drug
230. Huang CJ, Gurlo T, Haataja L, et al. Calcium-activated response in vivo. Cell. 2015;162(1):33– 44.
calpain-2 is a mediator of ␤ cell dysfunction and apoptosis 247. Grygiel-Górniak B. Peroxisome proliferator-activated re-
in type 2 diabetes. J Biol Chem. 2010;285(1):339 –348. ceptors and their ligands: nutritional and clinical implica-
231. Hatanaka M, Tanabe K, Yanai A, et al. Wolfram syn- tions–a review. Nutr J. 2014;13:17.
drome 1 gene (WFS1) product localizes to secretory gran- 248. Bossé Y, Weisnagel SJ, Bouchard C, Després JP, Pérusse L,
ules and determines granule acidification in pancreatic Vohl MC. Combined effects of PPAR␥2 P12A and PPAR␣
␤-cells. Hum Molec Genet. 2011;20(7):1274 –1284. L162V polymorphisms on glucose and insulin homeosta-
232. Ahmadian M, Suh JM, Hah N, et al. PPAR␥ signaling and sis: the Québec Family Study. J Hum Genet. 2003;48(12):
metabolism: the good, the bad and the future. Nat Med. 614 – 621.
2013;19(5):557–566. 249. Koch M, Rett K, Maerker E, et al. The PPAR␥2 amino acid
233. Jonker JW, Suh JM, Atkins AR, et al. A PPAR␥-FGF1 axis polymorphism Pro 12 Ala is prevalent in offspring of type
is required for adaptive adipose remodelling and metabolic II diabetic patients and is associated to increased insulin
homeostasis. Nature. 2012;485(7398):391–394. sensitivity in a subgroup of obese subjects. Diabetologia.
234. Dutchak PA, Katafuchi T, Bookout AL, et al. Fibroblast 1999;42(6):758 –762.
growth factor-21 regulates PPAR␥ activity and the antidi- 250. Claussnitzer M, Dankel SN, Klocke B, et al. Leveraging
abetic actions of thiazolidinediones. Cell. 2012;148(3): cross-species transcription factor binding site patterns:
556 –567. from diabetes risk loci to disease mechanisms. Cell. 2014;
235. Wei W, Dutchak PA, Wang X, et al. Fibroblast growth 156:343–358.
factor 21 promotes bone loss by potentiating the effects of 251. Steiner DF, Tager HS, Chan SJ, Nanjo K, Sanke T, Ruben-
peroxisome proliferator-activated receptor ␥. Proc Natl stein AH. Lessons learned from molecular biology of insulin-
Acad Sci USA. 2012;109(8):3143–3148. gene mutations. Diabetes Care. 1990;13(6):600 – 609.
236. Choi JH, Banks AS, Estall JL, et al. Anti-diabetic drugs 252. Tager H, Given B, Baldwin D, et al. A structurally abnor-
inhibit obesity-linked phosphorylation of PPAR␥ by mal insulin causing human diabetes. Nature. 1979;
Cdk5. Nature. 2010;466(7305):451– 456. 281(5727):122–125.
253. Haneda M, Chan SJ, Kwok SC, Rubenstein AH, Steiner 269. Wang J, Takeuchi T, Tanaka S, et al. A mutation in the
DF. Studies on mutant human insulin genes: identification insulin 2 gene induces diabetes with severe pancreatic
and sequence analysis of a gene encoding [SerB24]insulin. ␤-cell dysfunction in the Mody mouse. J Clin Invest. 1999;
Proc Natl Acad Sci USA. 1983;80(20):6366 – 6370. 103(1):27–37.
254. Shoelson S, Fickova M, Haneda M, et al. Identification of 270. Liu M, Hodish I, Rhodes CJ, Arvan P. Proinsulin matura-
a mutant human insulin predicted to contain a serine-for- tion, misfolding, and proteotoxicity. Proc Natl Acad Sci
phenylalanine substitution. Proc Natl Acad Sci USA. 1983; USA. 2007;104(40):15841–15846.
80(24):7390 –7394. 271. Ahamed A, Unnikrishnan AG, Pendsey SS, et al. Perma-
255. Nanjo K, Sanke T, Miyano M, et al. Diabetes due to se- nent neonatal diabetes mellitus due to a C96Y heterozy-
cretion of a structurally abnormal insulin (insulin Wa- gous mutation in the insulin gene. A case report. JOP.
kayama). Clinical and functional characteristics of 2008;9(6):715–718.
[LeuA3] insulin. J Clin Invest. 1986;77(2):514 –519. 272. Garin I, Edghill EL, Akerman I, et al. Recessive mutations
256. Carroll RJ, Hammer RE, Chan SJ, Swift HH, Rubenstein in the INS gene result in neonatal diabetes through reduced

AH, Steiner DF. A mutant human proinsulin is secreted insulin biosynthesis. Proc Natl Acad Sci USA. 2010;
from islets of Langerhans in increased amounts via an un- 107(7):3105–3110.
regulated pathway. Proc Natl Acad Sci USA. 1988;85(23): 273. Bakay M, Pandey R, Hakonarson H. Genes involved in
8943– 8947. type 1 diabetes: an update. Genes. 2013;4(3):499 –521.
257. Schwartz GP, Burke GT, Katsoyannis PG. A superactive 274. Vafiadis P, Bennett ST, Todd JA, et al. Insulin expression
insulin: [B10-aspartic acid]insulin(human). Proc Natl in human thymus is modulated by INS VNTR alleles at the
Acad Sci USA. 1987;84(18):6408 – 6411. IDDM2 locus. Nat Genet. 1997;15(3):289 –292.
258. Støy J, Edghill EL, Flanagan SE, et al. Insulin gene muta- 275. Zhang L, Nakayama M, Eisenbarth GS. Insulin as an au-
tions as a cause of permanent neonatal diabetes. Proc Natl toantigen in NOD/human diabetes. Curr Opin Immunol.
Acad Sci USA. 2007;104(38):15040 –15044. 2008;20(1):111–118.
259. Garin I, Perez de Nanclares G, Gastaldo E, Harries LW, 276. Bennett ST, Todd JA. Human type 1 diabetes and the in-
Rubio-Cabezas O, Castano L. Permanent neonatal diabe- sulin gene: principles of mapping polygenes. Ann Rev
tes caused by creation of an ectopic splice site within the Genet. 1996;30:343–370.
INS gene. PLoS One. 2012;7(1):e29205. 277. Meigs JB, Dupuis J, Herbert AG, Liu C, Wilson PW,
260. Carmody D, Park SY, Ye H, et al. Continued lessons from Cupples LA. The insulin gene variable number tandem re-
the INS gene: an intronic mutation causing diabetes peat and risk of type 2 diabetes in a population-based sam-
through a novel mechanism. J Med Genet. 2015;52(9): ple of families and unrelated men and women. J Clin En-
612– 616. docrinol Metab. 2005;90(2):1137–1143.
261. Dusatkova L, Dusatkova P, Vosahlo J, et al. Frameshift 278. Hansen SK, Gjesing AP, Rasmussen SK, et al. Large-scale
mutations in the insulin gene leading to prolonged mole- studies of the HphI insulin gene variable-number-of-tan-
cule of insulin in two families with maturity-onset diabetes dem-repeats polymorphism in relation to type 2 diabetes
of the young. Eur J Med Genet. 2015;58(4):230 –234. mellitus and insulin release. Diabetologia. 2004;47(6):
262. Bonfanti R, Colombo C, Nocerino V, et al. Insulin gene 1079 –1087.
mutations as cause of diabetes in children negative for five 279. Huxtable SJ, Saker PJ, Haddad L, et al. Analysis of parent-
type 1 diabetes autoantibodies. Diabetes Care. 2009; offspring trios provides evidence for linkage and associa-
32(1):123–125. tion between the insulin gene and type 2 diabetes mediated
263. Oda N, Nakai A, Fujiwara K, et al. Polymorphisms of the exclusively through paternally transmitted class III vari-
insulin gene among Japanese subjects. Metabolism. 2001; able number tandem repeat alleles. Diabetes. 2000;49(1):
50(6):631– 634. 126 –130.
264. Flannick J, Beer NL, Bick AG, et al. Assessing the pheno- 280. Lindsay RS, Hanson RL, Wiedrich C, Knowler WC, Ben-
typic effects in the general population of rare variants in nett PH, Baier LJ. The insulin gene variable number tan-
genes for a dominant Mendelian form of diabetes. Nat dem repeat class I/III polymorphism is in linkage disequi-
Genet. 2013;45(11):1380 –1385. librium with birth weight but not type 2 diabetes in the
265. Park SY, Ye H, Steiner DF, Bell GI. Mutant proinsulin Pima population. Diabetes. 2003;52(1):187–193.
proteins associated with neonatal diabetes are retained in 281. Brookes KJ. The VNTR in complex disorders: the forgot-
the endoplasmic reticulum and not efficiently secreted. ten polymorphisms? A functional way forward? Genom-
Biochem Biophys Res Commun. 2010;391(3):1449 – ics. 2013;101(5):273–281.
1454. 282. Saxena R, Elbers CC, Guo Y, et al. Large-scale gene-centric
266. Sunyaev S, Ramensky V, Bork P. Towards a structural meta-analysis across 39 studies identifies type 2 diabetes
basis of human non-synonymous single nucleotide poly- loci. Am J Hum Genet. 2012;90(3):410 – 425.
morphisms. Trends Genet. 2000;16(5):198 –200. 283. Liu M, Hodish I, Haataja L, et al. Proinsulin misfolding
267. Støy J, Steiner DF, Park SY, Ye H, Philipson LH, Bell GI. and diabetes: mutant INS gene-induced diabetes of youth.
Clinical and molecular genetics of neonatal diabetes due to Trends Endocrinol Metab. 2010;21(11):652– 659.
mutations in the insulin gene. Rev Endocr Metab Disord. 284. Guo H, Xiong Y, Witkowski P, et al. Inefficient translo-
2010;11(3):205–215. cation of preproinsulin contributes to pancreatic ␤ cell fail-
268. Molven A, Ringdal M, Nordbø AM, et al. Mutations in the ure and late-onset diabetes. J Biol Chem. 2014;289(23):
insulin gene can cause MODY and autoantibody-negative 16290 –16302.
type 1 diabetes. Diabetes. 2008;57(4):1131–1135. 285. Fernández ÁF, López-Otín C. The functional and patho-
logic relevance of autophagy proteases. J Clin Invest. 2015; 302. Cho YS, Chen CH, Hu C, et al. Meta-analysis of genome-
125(1):33– 41. wide association studies identifies eight new loci for type 2
286. Bachar-Wikstrom E, Wikstrom JD, Ariav Y, et al. Stimula- diabetes in east Asians. Nat Genet. 2012;44(1):67–72.
tion of autophagy improves endoplasmic reticulum stress- 303. Barker A, Sharp SJ, Timpson NJ, et al. Association of ge-
induced diabetes. Diabetes. 2013;62(4):1227–1237. netic loci with glucose levels in childhood and adolescence:
287. Liu M, Haataja L, Wright J, et al. Mutant INS-gene in- a meta-analysis of over 6,000 children. Diabetes. 2011;
duced diabetes of youth: proinsulin cysteine residues im- 60(6):1805–1812.
pose dominant-negative inhibition on wild-type proinsulin 304. Sakai K, Imamura M, Tanaka Y, et al. Replication study
transport. PLoS One. 2010;5(10):e13333. for the association of 9 East Asian GWAS-derived loci with
288. Hay CW, Docherty K. Comparative analysis of insulin susceptibility to type 2 diabetes in a Japanese population.
gene promoters: implications for diabetes research. Dia- PLoS One. 2013;8(9):e76317.
betes. 2006;55(12):3201–3213. 305. ZeRuth GT, Takeda Y, Jetten AM. The Krüppel-like pro-
289. Kang HS, Kim YS, ZeRuth G, et al. Transcription factor tein Gli-similar 3 (Glis3) functions as a key regulator of

Glis3, a novel critical player in the regulation of pancreatic insulin transcription. Mol Endocrinol. 2013;27(10):
␤-cell development and insulin gene expression. Mol Cell 1692–1705.
Biol. 2009;29(24):6366 – 6379. 306. Yang Y, Chang BH, Chan L. Sustained expression of the
290. Kim YS, Nakanishi G, Lewandoski M, Jetten AM. GLIS3, transcription factor GLIS3 is required for normal ␤ cell
a novel member of the GLIS subfamily of Krüppel-like zinc function in adults. EMBO Mol Med. 2013;5(1):92–104.
finger proteins with repressor and activation functions. 307. Watanabe N, Hiramatsu K, Miyamoto R, et al. A murine
Nucleic Acids Res. 2003;31(19):5513–5525. model of neonatal diabetes mellitus in Glis3-deficient mice.
291. Senée V, Chelala C, Duchatelet S, et al. Mutations in GLIS3 FEBS Lett. 2009;583(12):2108 –2113.
are responsible for a rare syndrome with neonatal diabetes 308. Yang Y, Chang BH, Yechoor V, et al. The Krüppel-like zinc
mellitus and congenital hypothyroidism. Nat Genet. 2006; finger protein GLIS3 transactivates neurogenin 3 for
proper fetal pancreatic islet differentiation in mice. Dia-
38(6):682– 687.
betologia. 2011;54(10):2595–2605.
292. Dimitri P, Warner JT, Minton JA, et al. Novel GLIS3 mu-
309. Georgia S, Bhushan A. ␤-Cell replication is the primary
tations demonstrate an extended multisystem phenotype.
mechanism for maintaining postnatal ␤ cell mass. J Clin
Eur J Endocrinol. 2011;164(3):437– 443.
Invest. 2004;114(7):963–968.
293. Dimitri P, Habeb AM, Garbuz F, et al. Expanding the clin-
310. Kushner JA, Ciemerych MA, Sicinska E, et al. Cyclins D2
ical spectrum associated with GLIS3 mutations. J Clin En-
and D1 are essential for postnatal pancreatic ␤-cell growth.
docrinol Metab. 2015;100(10):E1362–E1369.
Mol Cell Biol. 2005;25(9):3752–3762.
294. Cao B, Gong C, Wu D, et al. Genetic analysis and fol-
311. Georgia S, Hinault C, Kawamori D, et al. Cyclin D2 is
low-up of 25 neonatal diabetes mellitus patients in China.
essential for the compensatory ␤-cell hyperplastic response
J Diabetes Res. 2016;2016:6314368.
to insulin resistance in rodents. Diabetes. 2010;59(4):987–
295. Barrett JC, Clayton DG, Concannon P, et al. Genome-wide
996.
association study and meta-analysis find that over 40 loci 312. Morris AP, Voight BF, Teslovich TM, et al. Large-scale
affect risk of type 1 diabetes. Nat Genet. 2009;41(6):703– association analysis provides insights into the genetic ar-
707. chitecture and pathophysiology of type 2 diabetes. Nat
296. Kiani AK, John P, Bhatti A, et al. Association of 32 type 1 Genet. 2012;44(9):981–990.
diabetes risk loci in Pakistani patients. Diabetes Res Clin 313. Eizirik DL, Colli ML, Ortis F. The role of inflammation in
Pract. 2015;108(1):137–142. insulitis and ␤-cell loss in type 1 diabetes. Nat Rev Endo-
297. Winkler C, Krumsiek J, Buettner F, et al. Feature ranking crinol. 2009;5(4):219 –226.
of type 1 diabetes susceptibility genes improves prediction 314. Cnop M, Ladrière L, Igoillo-Esteve M, Moura RF, Cunha
of type 1 diabetes. Diabetologia. 2014;57(12):2521–2529. DA. Causes and cures for endoplasmic reticulum stress in
298. Boesgaard TW, Grarup N, Jørgensen T, et al. Variants at lipotoxic ␤-cell dysfunction. Diabetes Obes Metab. 2010;
DGKB/TMEM195, ADRA2A, GLIS3 and C2CD4B loci 12(suppl 2):76 – 82.
are associated with reduced glucose-stimulated ␤ cell func- 315. Nogueira TC, Paula FM, Villate O, et al. GLIS3, a suscep-
tion in middle-aged Danish people. Diabetologia. 2010; tibility gene for type 1 and type 2 diabetes, modulates pan-
53(8):1647–1655. creatic ␤ cell apoptosis via regulation of a splice variant of
299. Liu C, Li H, Qi L, et al. Variants in GLIS3 and CRY2 are the BH3-only protein Bim. PLoS Genet. 2013;9(5):
associated with type 2 diabetes and impaired fasting glu- e1003532.
cose in Chinese Hans. PLoS One. 2011;6(6):e21464. 316. Thomsen SK, Gloyn AL. The pancreatic ␤ cell: recent in-
300. Hu C, Zhang R, Wang C, et al. Variants from GIPR, sights from human genetics. Trends Endocrinol Metab.
TCF7L2, DGKB, MADD, CRY2, GLIS3, PROX1, 2014;25(8):425– 434.
SLC30A8 and IGF1 are associated with glucose metabo- 317. Gaulton KJ, Nammo T, Pasquali L, et al. A map of open
lism in the Chinese. PLoS One. 2010;5(11):e15542. chromatin in human pancreatic islets. Nat Genet. 2010;
301. Rees SD, Hydrie MZ, O’Hare JP, et al. Effects of 16 genetic 42(3):255–259.
variants on fasting glucose and type 2 diabetes in South 318. Pasquali L, Gaulton KJ, Rodríguez-Seguí SA, et al. Pan-
Asians: ADCY5 and GLIS3 variants may predispose to creatic islet enhancer clusters enriched in type 2 diabetes
type 2 diabetes. PLoS One. 2011;6(9):e24710. risk-associated variants. Nat Genet. 2014;46(2):136 –143.

Edrv 0190

Uploaded by

Copyright:

Available Formats

Edrv 0190

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Edrv 0190

Uploaded by

Copyright:

Available Formats

R E V I E W

Monogenic Diabetes: What It Teaches Us on the

Yisheng Yang and Lawrence Chan

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Figure 1. 2 (TCF7L2) rs7903146 is one of the

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Monogenic Diabetes Genes Associated With T2D

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Abbreviation: TF, transcription factor.

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

B. Glucokinase 2. Variations in the GCK gene and T2D

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Figure 5. cyclin D2 (Ccnd2) mRNA expres-

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

Downloaded from https://academic.oup.com/edrv/article/37/3/190/2354693 by guest on 11 November 2022

You might also like