University of Maryland School of Medicine (A.R.S., N.H., T.I.P.), Veterans Administration Geriatrics Research and Education Clinical Center (A.R.S.), Baltimore, Maryland 21201
Address all correspondence and requests for reprints to: Alan R. Shuldiner, M.D., Division of Endocrinology, Diabetes, and Nutrition, University of Maryland School of Medicine, 660 West Redwood Street, Room 494, Baltimore, Maryland 21201. E-mail: ashuldin{at}medicine.umaryland.edu.
The Finnish Diabetes Prevention Study (1), like the U.S. Diabetes Prevention Program Study (2), is a randomized prospective clinical trial of the effects of a lifestyle intervention on the incidence of type 2 diabetes (T2DM). Both studies showed that an intensive lifestyle intervention consisting of diet and modest weight loss, coupled with a regular exercise program in high-risk individuals, resulted in a 58% reduction in the incidence of T2DM compared with the control group in which no intervention was performed. Clearly, in light of the current epidemic of T2DM in virtually all Westernized countries throughout the world, these findings have profound health implications. However, these studies also point out the large interindividual variation in response to the standardized lifestyle intervention: Why did some subjects develop T2DM, whereas others did not despite the same intervention (and similar levels of adherence)?
One such factor that is likely to contribute to this interindividual variation in response is genetic differences among subjects. This month, Todorova et al. (3) present evidence that subjects of the Finnish Diabetes Prevention Study who carried at least one copy of the A allele of the 250 G>A variant in the hepatic lipase gene (LIPC) promoter, which constituted 44% of the group studied, were more responsive to the effect of diet and exercise in preventing or delaying the onset of T2DM than those carrying two G alleles. The authors chose LIPC as a candidate gene because of its function in lipoprotein metabolism and the well-known association of dyslipidemia (most notably, high levels of atherogenic small low-density lipoprotein, low high-density lipoprotein, and high triglycerides) with T2DM, thus implicating the involvement of genes encoding proteins involved in lipid metabolism in T2DM pathogenesis. Furthermore, visceral obesity is associated with high hepatic lipase activity, and weight loss decreases hepatic lipase activity with associated improvements in lipoprotein profile and, as mentioned, prevention of T2DM.
The authors found that the risk of developing T2DM was approximately 2-fold higher in those homozygous for the 250G allele compared with those with at least one 250A allele; the conversion rate to T2DM was 17.8% among subjects homozygous for the 250G allele, compared with only 10.7% among subjects with at least one 250A allele (3). Strikingly, in the lifestyle intervention group, 13.0% of subjects homozygous for the 250G allele converted to diabetes, whereas only 1.0% of subjects with at least one 250A allele converted to diabetes. Although a similar trend was noted in the control group, the more pronounced decrease in the incidence of diabetes in the intervention group suggests that the effectiveness of the lifestyle intervention is influenced by LIPC genotype, i.e. subjects with the A allele are more likely to benefit from lifestyle intervention. This greater benefit was partially attributable to greater weight loss in those with the A allele.
This report differs from most other candidate gene reports, which are cross-sectional and thus can measure only differences in baseline characteristics between genotypes. In this study, a standardized intervention was performed, allowing the authors to directly examine correlation between genotype and response to the intervention. This is particularly important because it emphasizes the idea that genetic and environmental forces (the intervention) are not mutually exclusive. Indeed, for complex diseases like T2DM, gene-environment interactions are presumed to be quite important. The study also exemplifies that some polymorphisms in candidate genes for complex diseases may not have a large causal role in predisposition to disease but rather may act as response elements to environmental influences.
Other investigators have looked at the effect of genotype on diabetes and obesity-related quantitative traits in response to environmental influences. For example, Nicklas et al. (4) examined 70 postmenopausal women with or without the Pro12Ala polymorphism in peroxisome proliferator-activated receptor-2. These women had no differences in baseline traits before intervention. They were then put on hypocaloric diets for 6 months, and it was observed that those with the Ala12 allele showed greater improvement in insulin sensitivity and decreased fat oxidation compared with those with the Pro12 homozygote group. Although both groups lost similar amounts of weight, Ala12 carriers regained the lost weight more rapidly. These findings have clinical implications because they suggest that obese Ala12 carriers who successfully lose weight will have greater difficulty maintaining their reduced weight. Wilund et al. (5) and Halverstadt et al. (6) have reported endothelial lipase and cholesterol ester transfer protein genotype-dependent effects on high-density lipoprotein levels in response to exercise and low-fat diets. These studies along with the current one help to illuminate the idea that gene products do not act in isolation but interact with more established influences on disease. Indeed, many cases exist in which even a gene variant involved in a simple single gene disorder (such as phenylketonuria) requires a certain environment (phenylalanine in the diet) to result in a particular phenotype (mental retardation).
The molecular and cellular mechanisms by which LIPC genotype influenced efficacy of the lifestyle intervention in preventing T2DM require further investigation. Unfortunately, hepatic lipase activity was not measured in this study. In vitro studies of others (7) suggest that another polymorphism in the LIPC promoter, 514C>T, which in Caucasians is in linkage disequilibrium with the 250A allele, has lower transcriptional activity than the 514C allele. This could lead to lower hepatic lipase activity in those with the 250A allele, which would be expected to be beneficial with respect to lipoprotein traits. However, the mechanism whereby decreased hepatic lipase activity leads to decreased rates of conversion to diabetes is not clear.
Interestingly, in another study, the 250A allele was associated with increased insulin resistance (8). An additional study showed an association between the 514T allele, which reportedly reduces transcription of LIPC (7) with reduced glucose tolerance, but only in individuals who also had the APOC3 482C>T variant (9). In light of these previous findings, one might have expected the 250A allele to increase, not decrease, susceptibility to T2DM in the current study. It is possible that these seemingly conflicting results are the consequence of type 1 error. Alternatively, the 250A allele may truly behave in a protective manner in individuals who were selected to have impaired glucose tolerance (and thus, likely insulin resistance) at the onset of the study. An association study by Jansen et al. (10) in coronary artery disease patients provided evidence that the 480 C>T LIPC promoter variant might abolish the effect of insulin on LIPC transcription; such a mechanism might play a role in the putative protective effect of 250C>A or a linked variant in insulin-resistant individuals that is most apparent in the presence of changes in diet and activity levels. Given the possibility of differential effects of LIPC genotype in insulin-resistant subjects, it is unclear whether the results of this study in subjects chosen based on the presence of impaired glucose tolerance and increased risk of T2DM are generalizable to population-based samples, or for that matter, to populations other than Finns.
At this point, many readers are likely asking the question, "How is this finding clinically useful?" Although it does suggest that LIPC genotype helps predict response to exercise and might be a useful means of identifying those who could most benefit from improvements in diet and exercise, it is clear that the lifestyle intervention was beneficial even without considering genotype, and these lifestyle changes prescribed in everyone are unlikely to result in harm. Therefore, from a clinical standpoint, the question becomes, "Why test for the gene polymorphism at all?" Indeed, variation at the LIPC locus overall is a weak predictor of T2DM development, with G allele homozygotes having a sensitivity of 68%, specificity of 46%, and positive predictive value of 18%. If we break down the group into study groups, then the intervention group shows low specificity (45%) and positive predictive value (13%), but impressively high sensitivity (95%). In other words, this study suggests that genetic testing, i.e. identifying those with the A allele, will identify a subgroup of individuals who will be particularly responsive to a diet and exercise intervention. A corollary is that subjects homozygous for the G allele might be more appropriately treated with other interventions (e.g. pharmacotherapy) in addition to diet and exercise.
These studies will need to be confirmed and extended (see below) before LIPC genetic testing is borne out to be clinically useful. This is particularly the case in candidate gene studies in an era in which the entire sequence of the human genome is available, increasing the number of possible candidate genes to test and therefore the potential for type 1 error. A logical testable prediction of the findings of this study is that there should be a higher frequency of the G allele in T2DM, particularly in populations in which high levels of physical activity are the norm. Alternatively, cross-sectional studies in populations with wide variability in the level of physical activity should be able to detect statistical interactions between genotype and habitual physical activity on diabetes prevalence. In the event that enough evidence is accumulated in further research to justify the use of LIPC promoter genotype as a clinical test, it would then have to be validated via several stages as outlined by the American College of Medical Genetics (http://www.acmg.net). The appropriate assay would genotype variants with proven functional involvement in disease, not those in linkage disequilibrium, because linkage disequilibrium varies between populations. In any case, the proposed use of a genetic test such as that for the LIPC variant as a screening test to identify those in need of more aggressive interventions contrasts with the view expressed by Merikangas and Risch (11) that "gene hunting for disorders that appear to be highly amenable to environmental modification ... such as type 2 diabetes" should have low priority. In fact, it is likely at least at present that an identified gene-environment interaction offers more hope for prevention opportunity than a gene variant that is detrimental, regardless of environment, given the limited success achieved in the area of gene manipulation/therapy in humans to date.
Although there are a few forms of diabetes that are due to single gene defects (e.g. maturity onset diabetes of the young, type A syndrome of extreme insulin resistance, congenital generalized lipoatrophic diabetes), the common form of T2DM is likely to be due to the presence of several common gene variants in susceptible individuals. Thus, the combination of specific gene variants, genetic (ethnic) background, and gene-gene interactions will influence in large part not only susceptibility to the disease but also other characteristics, such as age of onset, severity of hyperglycemia, complications, effect of lifestyle on prevention and/or treatment, and response to pharmacological interventions. It stands to reason that genetic tests that capture variation in several genes and an understanding of how these gene variants interact with each other and the environment will have far superior sensitivity and specificity in predicting response to a given intervention than any single gene variant.
With the increasing knowledge of complex disease genes and risk factors for complex diseases, geneticsparticularly genetic testingwill be increasingly applied to medicine. Because of this, an important question emerges: Who will administer the tests and provide pretest and posttest counseling regarding their benefits and limitations? In the past, genetic testing has been confined to rare single gene disorders and to women of childbearing age at risk for fetal anomalies; ideally, testing, counseling, and related services have been provided by specialists in the fields of genetic counseling and medical genetics. However, as genetic tests for common disorders become more widespread, the segment of the population eligible for testing will grow dramatically, and primary care providers and other medical specialists along with public health providers will increasingly need to become involved in the process at some level (12, 13, 14). However, limitations in knowledge about and even interest in genetics by primary care physicians have been demonstrated by various studies (15, 16). According to the Association of Professors of Human or Medical Genetics, the average medical student is exposed to only 29 h of coursework in genetics (17). Providing more coursework in genetics may help future physicians both to understand the importance of and to be able to apply the principles of genetics in clinical practice and identify situations in which a genetic specialist consultation might be useful. Progress is being made by organizations such as the American Society of Human Genetics (http://www.ashg.org) and the National Coalition for Health Professional Education in Genetics (http://www.nchpeg.org) in educating primary care providers in genetics, raising awareness, and expanding interest (18). It is equally important that genetic specialists become more educated about general medical and public health practice so that they may be able to assess needs and figure out how to apply their expertise in these areas, whether in direct interactions with patients or education of providers or both.
Another issue with the growth of genetic testing for predisposition to complex disease is the effect that knowing ones susceptibility will have on his/her behavior as well as mental state. It is possible that if a person discovers that he or she carries a predisposition allele for a complex disorder, not only might this cause the person anxiety and stress, but that person might also be discouraged from attempting measures to prevent onset of that disease. In the specific example of the 250G>A LIPC polymorphism, it is not difficult to imagine that those homozygous for the G allele might be discouraged from attempting lifestyle changes. With proper posttest education and counseling, they should realize that lifestyle changes will still reduce their likelihood of developing diabetes. Moreover, risk reduction education and the opportunity to receive medications or other promising interventions, in addition to exercise and dietary changes, may increase their motivation to take action to prevent disease.
Because genetic testing for common forms of diabetes and obesity susceptibility is still far from clinical use, research on peoples true response to such testing is lacking in the literature. However, one small study by Harvey-Berino et al. (19) found that positive gene status for the Trp64Arg variant of the ß3-adrenergic receptor gene (ADRB3), which is associated with obesity, did not adversely affect obese womens confidence in their ability to lose weight or control eating behavior. Further insight might be gleaned from looking at a phenotype such as familial breast cancer, for which mutations in the BRCA1 and BRCA2 genes greatly increase risk for breast cancer. A study of a large Utah kindred found that female BRCA1 mutation carriers increased breast cancer screening behaviors such as mammograms (82% at 2-yr posttesting vs. 35% at baseline) and breast self examination (83% at 2-yr posttesting vs. 43% at baseline) after learning their genotypes. Those found not to carry the mutation were also not deterred from breast cancer screening efforts, with uptake of mammograms and breast self exams nearly as high as that in carriers (20). These findings for breast cancer are obviously not directly applicable to T2DM, both because they involve a different disease and because BRCA1 has a much greater influence on breast cancer than LIPC does on diabetes.
In summary, the study by Todorova et al. (3) provides a glimpse of how clinicians of the future might use genetic testing to target interventions more effectivelyso called individualized medicine. For complex diseases like T2DM, we expect that there are polymorphisms in several/or possibly many genes, each with relatively minor effect on disease susceptibility or on response to a given intervention. This study suggests that LIPC polymorphism may be one such factor. Future studies to confirm these findings, identify polymorphisms in other genes that contribute risk, or determine response to interventions, and determine how these polymorphisms interact with each other will be necessary steps before such genetic tests will have the sensitivity, specificity, and predictive value necessary to be useful to clinicians. Thus, it is possible that in the future, a test for LIPC genotype, particularly in conjunction with family history and/or genetic testing for polymorphisms in other candidate genes, might be useful in guiding prevention and/or therapy. Finally, even if testing for polymorphisms in LIPC or in other diabetes candidate genes does not become standard clinical practice, these research findings are useful for understanding disease pathogenesis, in this case part of the molecular basis of dyslipidemia as a risk factor for diabetes. A better understanding of disease pathogenesis will undoubtedly lead to improved methods of treatment and prevention and more rational use of existing ones.
Acknowledgments
We thank Dr. Braxton Mitchell for the critical reading of this editorial and for his helpful comments.
Footnotes
Abbreviations: LIPC, Hepatic lipase gene; T2DM, type 2 diabetes.
Received March 17, 2004.
Accepted March 17, 2004.
References