Diabetologia

, 54:2570

Association of the SLC30A8 missense polymorphism R325W with proinsulin levels at baseline and after lifestyle, metformin or troglitazone intervention in the Diabetes Prevention Program

  • A. R. Majithia
  • K. A. Jablonski
  • J. B. McAteer
  • K. J. Mather
  • R. B. Goldberg
  • S. E. Kahn
  • J. C. Florez
  • for the DPP Research Group
Short Communication

DOI: 10.1007/s00125-011-2234-1

Cite this article as:
Majithia, A.R., Jablonski, K.A., McAteer, J.B. et al. Diabetologia (2011) 54: 2570. doi:10.1007/s00125-011-2234-1

Abstract

Aims/hypothesis

Individuals with impaired glucose tolerance have increased proinsulin levels, despite normal glucose or C-peptide levels. In the Diabetes Prevention Program (DPP), increased proinsulin levels predicted type 2 diabetes and proinsulin levels were significantly reduced following treatment with metformin, lifestyle modification or troglitazone compared with placebo. Genetic and physiological studies suggest a role for the zinc transporter gene SLC30A8 in diabetes risk, possibly through effects on insulin-processing in beta cells. We hypothesised that the risk allele at the type 2 diabetes-associated missense polymorphism rs13266634 (R325W) in SLC30A8 would predict proinsulin levels in individuals at risk of type 2 diabetes and may modulate response to preventive interventions.

Methods

We genotyped rs13266634 in 3,007 DPP participants and examined its association with fasting proinsulin and fasting insulin at baseline and at 1 year post-intervention.

Results

We found that increasing dosage of the C risk allele at SLC30A8 rs13266634 was significantly associated with higher proinsulin levels at baseline (p = 0.002) after adjustment for baseline insulin. This supports the hypothesis that risk alleles at SLC30A8 mark individuals with insulin-processing defects. At the 1 year analysis, proinsulin levels decreased significantly in all groups receiving active intervention and were no longer associated with SLC30A8 genotype (p = 0.86) after adjustment for insulin at baseline and 1 year. We found no genotype × treatment interactions at 1 year.

Conclusions/interpretation

In prediabetic individuals, genotype at SLC30A8 predicts baseline proinsulin levels independently of insulin levels, but does not predict proinsulin levels after amelioration of insulin sensitivity at 1 year.

Keywords

Diabetes Prevention ProgramGenetic associationProinsulinSingle nucleotide polymorphismsSLC30A8Zinc transporter

Abbreviations

ANCOVA

Analysis of covariance

DPP

Diabetes Prevention Program

IGT

Impaired glucose tolerance

SNP

Single nucleotide polymorphism

ZnT-8

Zinc transporter 8

Introduction

The first published genome-wide association study for type 2 diabetes showed the single nucleotide polymorphism (SNP) rs13266634 in SLC30A8 (OR 1.26, p = 5.0 × 10−7) to be a diabetes-associated locus [1]. SLC30A8 encodes zinc transporter 8 (ZnT-8), a 369 amino acid transmembrane zinc transporter protein produced at high levels only in the pancreas. Within the pancreatic beta cell, zinc is a necessary component for the formation of insulin granules as it provides the nucleus for crystallisation of insulin proteins within secretory vesicles. The SLC30A8 non-synonymous SNP (C to T) results in a missense R325W substitution. The production pattern and putative function of ZnT-8 suggest that it may increase diabetes risk by impairing insulin secretion. Indeed a defect in first-phase insulin secretion was observed in homozygotes for the high-risk C allele in response to intravenous glucose tolerance test (19% decrease in first-phase insulin release, p = 0.007), but not during an OGTT [2]. However, homology-based structural modelling of ZnT-8 does not indicate an obvious conformational defect in the docking or transporter segments of the protein [3]. In a recently reported beta cell-specific knockout of Slc30a8 in mice, the defect in first-phase insulin secretion was repeated, and glucose intolerance and elevated proinsulin levels were demonstrated [4]. A study of overweight non-diabetic men from a Finnish cohort (30% of whom had impaired glucose tolerance [IGT]) found that individuals with the high-risk genotype at SLC30A8 showed elevated proinsulin/insulin ratios while fasting as well as during OGTT [5].

These studies suggest that SLC30A8 variants link to diabetes risk through alterations in proinsulin to insulin conversion. Previous studies have shown that elevated fasting proinsulin levels are associated with increased risk of developing type 2 diabetes independently of insulin levels [6, 7]. However, elevated proinsulin levels in carriers of the SLC30A8 risk genotype may be an epiphenomenon of increased beta cell stress (from insulin resistance or beta cell death), rather than reflecting a specific deficit in polypeptide processing. Such non-specific mechanisms cannot be ruled out as mediators of the SLC30A8-associated increase in diabetes susceptibility. We reasoned that if SLC30A8-associated proinsulin elevations were largely mediated by a proinsulin-to-insulin conversion defect, interventions that decreased insulin resistance should lower proinsulin in high- and low-risk genotype groups to equal degrees. However, if increased beta cell stress (i.e. increased insulin resistance) was the major determinant of SLC30A8-associated proinsulin levels, interventions that ameliorate insulin resistance should show a selectively beneficial effect in carriers of the SLC30A8 risk genotype as compared with the protective genotype.

In the Diabetes Prevention Program (DPP), we have previously shown that genotype at SLC30A8 rs13266634 is not associated with diabetes incidence, with no significant interaction of genotype with intervention being observed with regard to this outcome [8]. To assess the putative association of SLC30A8 with proinsulin levels and their response to treatments designed to lower insulin resistance, we examined the effect of SLC30A8 genotype on proinsulin levels in the DPP at baseline and 1 year after preventive interventions.

Methods

The DPP (ClinicalTrials.gov number NCT00004992; for a list of investigators see electronic supplementary material [ESM]) enrolled 3,234 US American participants at high risk of developing diabetes (on the basis of overweight, increased fasting glucose and impaired glucose tolerance) and randomised them to placebo, metformin 850 mg twice daily or a lifestyle intervention aimed at ≥7% weight loss and ≥150 min of physical activity per week. A fourth arm of 585 participants who were initially randomised to troglitazone was terminated early because of concerns with hepatotoxicity. Of the above, 3,007 participants who had consented to genetic investigation and for whom valid SLC30A8 genotypes and data at 1 year were available were studied here.

The main endpoint was development of diabetes (as indicated by fasting glucose or 2 h glucose after an OGTT) confirmed by a second measurement. In total, 551 participants developed diabetes; the DPP showed that participants treated with metformin or with a lifestyle intervention were respectively 31% or 58% less likely to develop diabetes after an average of 3.2 years of follow-up [9]. To assess the effects of genotype at SLC30A8 rs13266634 on proinsulin levels in the DPP cohort, we analysed participants at baseline using an analysis of covariance (ANCOVA) model with proinsulin as the dependent variable, and fasting insulin, age, sex, self-reported ethnicity and genotype at SLC30A8 rs13266634 (obtained from our prior study as previously described [8]) as independent variables.

To evaluate the effect of genotype at SLC30A8 rs13266634 on treatment-mediated reductions in proinsulin after 1 year of DPP interventions, we performed a prospective cohort analysis using ANCOVA models with fasting proinsulin at baseline and year 1 as the dependent variables. Participants who developed diabetes at 1 year were excluded from year 1 analyses. Nominal two-sided p values adjusted for multiple comparisons are reported for post-hoc active treatment vs placebo statistical tests. All models included age, sex, self-reported ethnicity and genotype at SLC30A8 rs13266634 as independent variables. Mixed models ANCOVA was used to assess the effect of DPP treatments and genotype at SLC30A8 on proinsulin levels independently of insulin levels; the baseline proinsulin model was additionally adjusted for baseline fasting insulin; the year 1 proinsulin model was adjusted for year 1 fasting insulin. Additional adjustment of the year 1 model for baseline fasting proinsulin and insulin was performed to model change after controlling for the main effect of this SNP on proinsulin. Genotype × treatment interaction tests were performed for all year 1 models with the intention of analysing treatment groups together if there were no significant genotype × treatment interactions.

Results

At baseline, the C risk allele at SLC30A8 rs13266634 was significantly associated with higher fasting proinsulin levels in the DPP cohort. This association was in an allele dose-dependent fashion after adjustment for baseline fasting insulin (p = 0.002; Table 1) and remained statistically significant after further adjustment for fasting glucose (p = 0.001; Table 1).
Table 1

Baseline characteristics and association results by rs13266634 genotype

Baseline characteristics

Genotype

p value

CC (n = 1,726)

CT (n = 1,087)

TT (n = 194)

Treatment group, n (%)

 Placebo

522 (59.3)

312 (35.4)

47 (5.3)

0.28a

 Metformin

503 (56.6)

335 (37.7)

51 (5.7)

 

 Lifestyle

513 (56.6)

322 (35.5)

72 (7.9)

 

 Troglitazone

188 (57.0)

118 (35.8)

24 (7.3)

 

Self-reported ethnicity, n (%)

 White

855 (50.0)

713 (41.7)

141 (8.3)

<0.001a

 African-American

475 (80.2)

110 (18.6)

<15d

 

 Hispanic

298 (59.7)

173 (34.7)

28 (5.6)

 

 Asian/Pacific Islanders

48 (37.2)

68 (52.7)

<15d

 

 American Indians

50 (64.1)

23 (29.5)

<15d

 

Demographics

 Age (years)

50.6 ± 10.6

51.1 ± 10.4

53.7 ± 11.2

<0.01b

 Male sex, n (%)

557 (32.3)

389 (35.8)

71 (36.6)

0.11a

Baseline associations

 BMI (kg/m2)

34.0 ± 6.7

33.8 ± 6.5

33.9 ± 6.7

0.83b

 Fasting glucose (mmol/l)

5.9 ± 0.4

5.9 ± 0.4

5.9 ± 0.4

0.62b

 HbA1c (%)

5.9 ± 0.5

5.8 ± 0.5

5.9 ± 0.4

<0.001b

 HbA1c (mmol/mol)

41 ± 5.5

40 ± 5.5

41 ± 4.3

 

 Fasting insulin (pmol/l)e

159.7 (16.6, 1152.8)

159.7 (20.8, 1277.9)

166.7 (42.4, 666.7)

0.68b

 Proinsulin (pmol/l)f

15.81 (15.14, 16.51)

15.56 (14.82, 16.34)

15.29 (13.94, 16.77)

0.38c

 Proinsulin adjustedg (pmol/l)f

12.36 (11.77, 12.97)

11.80 (11.17, 12.47)

11.30 (10.19, 12.53)

0.002c

 Proinsulin adjustedh (pmol/l)f

15.92 (15.41, 16.44)

15.25 (14.70, 15.82)

14.43 (13.46, 15.46)

0.001c

Year 1 association

 Proinsulin adjustedi (pmol/l)f

12.25 (11.72, 12.81)

11.98 (11.39, 12.61)

12.06 (10.94, 13.30)

0.86c

Values are mean ± SD, unless indicated otherwise

All models adjusted for age, sex and ethnicity. The 1 year model was also adjusted for baseline fasting insulin

aχ2; bF test from ANOVA; cANCOVA

dTo preserve confidentiality in accordance with DPP publication policy, absolute numbers are not provided in cells with <15 participants

eFasting insulin values are median (minimum, maximum); fvalues are mean (95% CI)

Adjusted gfor fasting insulin, hfor fasting insulin and fasting glucose, ifor year 1 fasting insulin

After 1 year of randomised DPP intervention, fasting proinsulin levels were significantly decreased (p < 0.001) in all active intervention groups compared with placebo, even after adjustment for 1 year fasting insulin, age, sex and ethnicity (Fig. 1). While intervention groups had significantly lower proinsulin levels than placebo group at 1 year, the lifestyle intervention group had the largest decrease and was significantly lower than the metformin (p < 0.001) or troglitazone (p = 0.002) groups. The metformin and troglitazone intervention groups had similar proinsulin levels at 1 year (p = 0.86). There was no genotype × treatment interaction (p = 0.78), suggesting that the dosage of SLC30A8 risk alleles did not modify the effectiveness of any treatment. As there were no significant genotype × treatment interactions, the four groups were also analysed together. At 1 year after DPP interventions, fasting proinsulin levels were no longer associated significantly with genotype at the SLC30A8 locus when adjusted for 1 year fasting insulin, age, sex and ethnicity (p = 0.86; Table 1).
Fig. 1

Fasting proinsulin levels at baseline and 1 year after treatment in placebo (a), metformin (b), lifestyle (c), and troglitazone (d) groups. Proinsulin levels in each treatment panel were stratified by genotype at SLC30A8 rs13266634. Lines coded by genotype connect proinsulin levels at baseline and 1 year for each genotype group (CC continuous lines, CT dotted lines, TT dashed lines) in each treatment arm. Fasting proinsulin decreased significantly in all treatment arms (p < 0.001 by ANCOVA). There was no genotype × treatment interaction (p = 0.78 by ANCOVA). Genotype at SLC30A8 rs13266634 did not predict fasting proinsulin levels at 1 year (p = 0.86 by ANCOVA). All models were adjusted for sex, age and self-reported ethnicity. Baseline values were adjusted for baseline fasting insulin, year 1 values were adjusted for 1 year fasting insulin. Values are mean ± 95% CI

Discussion

Our results demonstrate that genotype at SLC30A8 was associated with fasting proinsulin levels at baseline in the prediabetic DPP cohort; proinsulin levels were positively correlated with dosage of the C diabetes risk allele. These findings are consistent with a previous study demonstrating elevated fasting proinsulin levels in Finnish individuals with IGT [5]. Because fasting proinsulin in our study was adjusted for fasting insulin, this trait represents an elevation of proinsulin that is out of proportion to the hyperinsulinaemia induced by insulin resistance. The persistence of this finding after adjustment for fasting glucose suggests that the effect of genotype on proinsulin is not mediated by ambient glycaemia and its impact on beta cells.

After 1 year of DPP treatments, all of which provided an insulin-sensitising effect, proinsulin levels fell in all treatment groups compared with placebo [10]. This decrease in proinsulin levels remained significant after adjustment for 1 year fasting insulin, suggesting that the decrease in proinsulin levels did not simply reflect a fall in insulin. This indicates that the notable amelioration in insulin sensitivity also signals improved beta cell function.

Genotype at SLC30A8 rs13266634 was no longer associated with proinsulin levels in the DPP cohort after 1 year of treatment. This was true for the placebo as well as the active intervention groups. We also note that genotype at SLC30A8 rs13266634 was no longer associated with proinsulin levels at 1 year even in a model unadjusted for baseline proinsulin levels (data not shown). This lack of association could be due to a loss of power to detect proinsulin differences by genotype when the cohort was divided into intervention groups. But these data are also consistent with the postulate that the various insulin-sensitising interventions leading to lower proinsulin levels could have been potent enough to decrease the main effect of SLC30A8 rs13266634 genotype on proinsulin.

Since elevated proinsulin levels are an independent risk factor for the eventual development of type 2 diabetes, our data suggest that susceptibility to diabetes posed by high-risk alleles at the SLC30A8 locus could be attenuated by preventive treatment prior to the development of diabetes. In the case of the SLC30A8 locus, interventions such as lifestyle could be targeted even before elevated proinsulin levels or IGT manifest in carriers of the risk genotypes.

Acknowledgements

The investigators gratefully acknowledge the commitment and dedication of the participants of the DPP. The National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) of the National Institutes of Health provided funding to the clinical centres and the Coordinating Center for the design and conduct of the study, and for collection, management, analysis and interpretation of the data. The Southwestern American Indian Centers were supported directly by the NIDDK and the Indian Health Service. The General Clinical Research Center Program, National Center for Research Resources and the Department of Veterans Affairs supported data collection at many of the clinical centres. Funding for data collection and participant support was also provided by the Office of Research on Minority Health, the National Institute of Child Health and Human Development, the National Institute on Aging, the Office of Research on Women’s Health, the Centers for Disease Control and Prevention and the American Diabetes Association. Bristol-Myers Squibb and Parke-Davis provided medication. This research was also supported, in part, by the intramural research programme of the NIDDK. LifeScan, Health O Meter, Hoechst Marion Roussel, Merck-Medco Managed Care, Merck, Nike Sports Marketing, Slim Fast Foods and Quaker Oats donated materials, equipment or medicines for concomitant conditions. McKesson BioServices, Matthews Media Group and the Henry M. Jackson Foundation provided support services under subcontract with the Coordinating Center. The opinions expressed are those of the investigators and do not necessarily reflect the views of the Indian Health Service or other funding agencies. A complete list of centres, investigators and staff can be found in the ESM. This work was supported in part by R01 DK072041 to J. C. Florez and K. A. Jablonski. S. E. Kahn is supported in part by the Department of Veterans Affairs. J. C. Florez is supported by a Physician Scientist Development Award by the Massachusetts General Hospital and a Clinical Scientist Development Award from the Doris Duke Charitable Foundation. We also thank the late A. F. Moore for his intellectual contribution to the genesis of this project.

Contribution statement

J.B.M. directed the genotyping with supervision from J.C.F. Recruitment and phenotyping were performed previously by the Diabetes Prevention Program Research Group. K.A.J. conducted statistical analyses with input from A.R.M. and J.C.F. A.R.M. wrote the manuscript with supervision from J.C.F. All authors were involved in the analysis and interpretation of results, contributed to the discussion, and reviewed and edited the manuscript. All the authors approved the final version of the manuscript.

Duality of interest

J. C. Florez has received consulting honoraria from Daiichi-Sankyo and AstraZeneca. All other authors declare that there is no duality of interest associated with this manuscript.

Supplementary material

125_2011_2234_MOESM1_ESM.pdf (154 kb)
ESMList of Diabetes Prevention Program Research Group investigators (PDF 154 kb)

Copyright information

© Springer-Verlag 2011

Authors and Affiliations

  • A. R. Majithia
    • 1
    • 2
    • 3
  • K. A. Jablonski
    • 4
  • J. B. McAteer
    • 1
    • 2
  • K. J. Mather
    • 5
  • R. B. Goldberg
    • 6
  • S. E. Kahn
    • 7
  • J. C. Florez
    • 1
    • 2
    • 3
    • 8
  • for the DPP Research Group
  1. 1.Center for Human Genetic Research and Diabetes Research Center (Diabetes Unit)Massachusetts General HospitalBostonUSA
  2. 2.Program in Medical and Population GeneticsBroad InstituteCambridgeUSA
  3. 3.Department of MedicineHarvard Medical SchoolBostonUSA
  4. 4.The George Washington University Biostatistsics CenterRockvilleUSA
  5. 5.Division of EndocrinologyIndiana University School of MedicineIndianapolisUSA
  6. 6.Division of Endocrinology, Diabetes, and Metabolism, Leonard M. Miller School of MedicineUniversity of MiamiMiamiUSA
  7. 7.Division of Metabolism, Endocrinology and NutritionVA Puget Sound Health Care System and University of WashingtonSeattleUSA
  8. 8.c/o Diabetes Prevention Program Coordinating CenterThe George Washington University Biostatistics CenterRockvilleUSA