Metabolic profiling in personalized medicine: bridging the gap between knowledge and clinical practice in Type 2 diabetes Type 2 diabetes mellitus (DM2) is the most commonly diagnosed metabolic disease and its prevalence is expected to increase. Epidemiological studies clearly show excess mortality associated with DM2, as well as an increased risk of DM2-related complications. Advances in personalized medicine would greatly improve patient care in the field of diabetes and other metabolic diseases. Prediction of the disease in asymptomatic patients as well as its harsh complications in patients already diagnosed is becoming a necessity, with the considerable increase in the cost of the treatment. In the current article, we review the known clinical, molecular metabolic and genetic biomarkers that should be integrated in a future bioinformatic platform to be used at the point-of-care, and discuss the challenges we face in applying this vision of personalized medicine for diabetes into reality.
KEYWORDS: biomarkers n decision support system n diabetes n metabolic profiling
Sagit Zolotov1,
n personalized medicine n prediction
Dafna Ben Yosef1,
Naphtali D Rishe2,

Scope of the problem
medical care. Such a system could be especially Yelena Yesha3
The overflow of scientific information has trans- useful for physicians treating subjects suffering & Eddy Karnieli†1
formed the field of medicine. We are now facing from metabolic diseases such as obesity, Type 2 ¹Institue of Endocrinology, Diabetes & an era in which in addition to the traditional diabetes mellitus (DM2) and hypertension. In Metabolism, Rambam Medical Center medical training, medical information is readily this patient population personalized medicine & Galil Center for Telemedicine, Medical Informatics & Personalized available to professionals and to the public. Easily can assist in identifying risk factors for preven- Medicine, RB Rappaport Faculty of searched databases provide original research arti- tion and prediction of chronic complications Medicine – Technion, 12 Ha'alya St, cles and focused reviews in seconds, and the flow such as cardiovascular disease, as well as tailor- Sami Ofer Tower, #8 Fl, PO Box 9602 Haifa 31096, Israel of new information is constant as databases are ing a treatment plan for an individual patient. 2Florida International University, updated as soon as new information is published. Since these chronic diseases affect a large por- FL, USA 3University of Maryland, Baltimore These databases hold key information that may tion of the population worldwide, require daily County, MD, USA lead to significant progress in clinical medicine. treatment for many years, lead to significant †Author for correspondence: Tel.: +972 4854 1606 This includes clinical and biological markers as deterioration in the quality of life, decrease Fax: +972 4854 2746 well as genetic markers indicating the variabil- life expectancy and increase health costs, such ity between one individual to another. These a system has obvious clinical significance. To advances in medical research may be used to best illustrate how personalized medicine can shed light on the individual's potential medical be utilized to improve diagnoses and patient course, attempting to predict personal risk of outcomes, this article will focus on the future developing a disease state, the course of a known possibilities in employing personalized medi-illness, the response to therapy and potential side cine for improving the management of DM2, effects. The new challenge for the healthcare thus bridging the gap between the advances in system is therefore to be able to translate the research and technology, and medical practice success obtained in molecular, genomic, infor- in the physician's office.
matic research into readily available tools that
can be used in clinical practice for an individual Background
patient, that is, personalized medicine. Indeed, Type 2 diabetes mellitus is the most com-
this trend toward personalized medicine is revo-
monly diagnosed metabolic disease. In 2009, lutionizing the medical world. Understanding approximately 23.7 million people, includ-and integrating genetic information with tra- ing approximately 10.7% of the adult popu- ditional clinical knowledge is the hallmark of lation in the USA, were afflicted with DM2. this transformation. Physicians should be able Unfortunately, its prevalence is expected to to extract the relevant information for an indi- increase to 44.1 million people by 2034 [1]. As vidual patient during an office visit, and use a a result, annual diabetes-related spending dur-decision support system to drive personalized ing this time period is expected to triple from 10.2217/PME.11.36 2011 Future Medicine Ltd Personalized Medicine (2011) 8(4), 445–456
PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karnieli PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karniel Metabolic profiling in personalized medicine PersPective US $113 billion in 2007 to an annual cost of most beneficial for both the physician and the approximately $339 billion. The importance of patient. This approach would allow the physi-DM2 is further emphasized by epidemiological cian to more accurately identify an individual's studies that clearly show excess mortality asso- risk for various complications, more easily iden- ciated with DM2, as well as an increased risk tify patients who require more aggressive medi-of DM2 related complications, including heart cal care and provide the tools a physician can disease, stroke, renal failure, peripheral sensory use to create a personally tailored treatment regi-neuropathy and diabetic retinopathy [101].
men that minimizes the potential for negative According to the American Diabetes side effects. Ideally, this process could be con- Association (ADA), diabetes is diagnosed ducted during a 10–15 min office visit, provid-owing to one of the four following criteria: ing quick, inexpensive, up-to-date and readily Fasting plasma glucose (FPG) above 126 mg% available diagnosis and treatment regimens that (7 mmol/l) fully support clinical guidelines and scientific advancement of the field. The significance of this ƒ Random plasma glucose above 200 mg% challenge clearly demonstrates the critical need (11.1 mmol/l) with classic symptoms for narrowing the gap between knowledge and clinical practice.
ƒ Plasma glucose above 200 mg% (11.1 mmol/l) 2 h after a 75 g oral glucose tolerance test Identifying asymptomatic subjects at
risk of developing DM2
ƒ Glycosylated hemoglobin (HbA1c) above Type 2 diabetes mellitus is a complex poly- genic disorder in which common genetic vari- All diabetic patients share the same treatment ants interact with environmental factors to goals of early diagnosis and tight control of glu- unmask the disease. Early identification of an cose levels. In an attempt to decrease the chance individual's risk for developing DM2 may aid of developing DM2 related-complications, the in the prevention of DM2 and diabetes-related maintenance protocol is aimed at lowering comorbidities [6]. The use of personalized medi-HbA1c below 7%. cine can improve a physician's ability to more Treatment of DM2 is often very problematic. accurately identify a patient's risk for the onset DM2 is diagnosed late in many patients and of DM2. It is well established that certain controlling glucose levels can be extremely dif- populations are at increased risk for developing ficult, which often leads to devastating complica- DM2. The 2011 ADA professional standards tions that cause a significant decrease in quality of medical care in DM2 recommend screening of life. According to the WHO, 50% of people asymptomatic patients for diabetes if they are with DM2 die of cardiovascular disease (primar- overweight or obese, and have additional risk ily heart disease and stroke). Of those patients factors such as physical inactivity, family his-who have lived with DM2 for 15 years or more, tory of DM2, membership in a high-risk eth-approximately 10% develop severe visual impair- nic group (African–American, Hispanic origin, ment and 2% become blind. Between 10–20% Native American, Asian American and Pacific of people with DM2 die of kidney failure. The Islander) and women who have delivered a baby overall risk of dying at least doubles for patients weighing more the 4 kg (9 lbs). Other indi-with DM2, with life expectancy in uncontrolled viduals who should be screened include patients diabetic patients reduced by approximately that have cardiovascular disease or risk factors 8 years [3,4]. On the other hand, some advocate for cardiovascular disease such as hypertension, that in well-controlled DM2 patients, life expec- dyslipidemia or conditions suggesting insulin tancy is similar to that of nondiabetic patients [5]. resistance such as polycystic ovary syndrome and As is mentioned above, in addition to personal acanthosis nigricans [2]. In addition to clinical health impairments, DM2 complications have a manifestations, several laboratory markers have significant economic impact on the heath system been implicated in predicting increased risk for worldwide [102].
the development of DM2. These include meta- Fortunately, not all the patients with DM2 bolic markers, plasma proteins, markers for develop all possible diabetes-related complica- endothelial dysfunction recommendations and tions. As a result, an integrated medical platform genetic predisposition. in which clinical, biological and genetic markers Fasting glucose is frequently used for screening that are most predictive of the development and for diabetes. Current ADA standards of care indi-appropriate treatment of the disease would be cate that impaired fasting glucose measurements Personalized Medicine (2011) 8(4)
future science group PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karnieli PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karniel Metabolic profiling in personalized medicine PersPective between 100 to 125 mg/dl identify individuals mediators contribute to systemic endothelial who are at increased risk for the development of dysfunction. A prospective study was performed DM2 [7,8]. However, even lower levels of impaired to determine whether elevated plasma levels of fasting glucose can be predictive of an increased these biomarkers could predict the develop-risk. For young men who otherwise appear to be ment of DM2 in women who were initially healthy, FPG levels as low as of 87 mg/dl have nondiabetic. It was found that for women who been found to represent an independent risk fac- later developed DM2, baseline median levels tor for DM2. It has been suggested that this crite- of these biomarkers were significantly higher ria could be used in addition to traditional factors than among control subjects. It was con-to identify apparently healthy men at increased cluded that endothelial dysfunction predicts risk for DM2 [9]. An additional biomarker that DM2 in women independent of other known can assist in predicting increased risk for devel- risk factors, including obesity and subclinical oping DM2 in apparently healthy young men i nflammation [15]. involves taking two measurements of triglyceride In addition to biomarkers, multiple genetic levels over a 5-year period of time [10]. Indeed, variants have been associated with the risk of the risk of developing DM2 was higher in men DM2. Recent genetic association studies have who progressed from lower triglyceride levels provided convincing evidence that several novel (30–66 mg/dl) to higher levels (164–299 mg/ loci are associated with increased risk of diabe- dl) as compared with men whose levels remained tes [16]. Two studies have recently tested whether in the lower triglyceride range at both time points the knowledge of these loci provides improved or men whose triglyceride levels decreased from prediction of the risk for development of DM2 the high range in the first measurement and the over clinical predictors such as phenotypic low range in the second measurement.
measurements [17,18]. In these studies, SNPs Sex hormone-binding globulin (SHBG) may at loci associated with DM2 were genotyped also play a role in increased risk of DM2. SHBG and analyzed to evaluate the risk of developing is primarily considered a binding protein of cir- DM2. Researchers looked at the predictive value culating hormones, regulating the bioavailable of these SNPs alone and in conjunction with fraction and sequestering circulating androgens other known DM2-related risk factors. Their and estrogens. Studies since the mid-1990s have findings suggest that a family history of diabe-suggested that SHBG may have biologic func- tes, increased BMI, elevated liver-enzyme levels, tions beyond simply the regulation of the levels current smoking status and reduced measures of of free sex hormones [11]. Research has found insulin secretion and action are predictors for that sex hormone bound to SHBG may directly the development of DM2. Similarly, in a study mediate biological functions such as cell-surface that focused on the Finnish population, clini-signaling and cellular delivery. Recent clinical cal data such as age, BMI, waist circumference, studies have associated low circulating levels of history of antihypertensive drug treatment and SHBG with impaired glucose control, implicat- high-blood glucose, lack of physical activity and ing the globulin in the maintenance of glucose low daily consumption of fruits and vegetables homeostasis [12]. Furthermore, circulating levels were shown as variables predicting diabetes of SHBG may be influenced by genetic variation. risk (FINDRISC) [19]. Independent of clinical In a study aimed at investigating the relationship risk factors, variants in several genes, such as between SHBG plasma levels and SHBG poly- TCF7L2, PPARg, FTO, KCNJ11, NOTCH2, morphisms, it was found that among postmeno- WFS1, CDKAL1, IGF2BP2, SLC30A8, JAZF1, pausal women and men, higher plasma levels of HNF1B, CDK2NA/B and HHEX, were also SHBG were prospectively associated with a lower found to be associated with increased risk of risk of DM2. Importantly, this finding suggests developing diabetes. Based on current research, that SHBG may have a causal role in increased when compared with the predictive value of risk of DM2 [13]. clinical risk factors alone, common genetic vari- Cross-sectional studies have consistently ants that are associated with DM2 have been found that patients with DM2 [14] and individu- found to slightly improve the ability to predict als who are at increased risk for DM2 tend to future onset of DM2 [17,18,20]. Recent findings by have elevated levels of inflammatory mediators, Laakso and colleagues support this concept [21]. such as cellular adhesion molecules (CAMs) When investigating improved identification including intercellular adhesion molecules 1 of previously undiagnosed DM2 individuals, (ICAM-1), E-selectin, and vascular cell adhe- these investigators found that adding measures sion molecules 1 (VCAM-1), and that these of total triglycerides, high-density lipoprotein future science group PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karnieli PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karniel Metabolic profiling in personalized medicine PersPective cholesterol, adiponectin and alanine trans- or genetic based assays. Once identified, physi- aminase (ALT) to the FINDRISC model only cians may use available measures in an attempt slightly increased its predictive value (from 0.727 to prevent the development of these conditions. to 0.772). Adding DM2 risk alleles to the model One example of using metabolic profiling in per-did not further improve its predictive value [21]. sonalized medicine is the prediction of cardio-The 40 known SNP variants only improved vascular complications in DM2 patients. An the predictive value for DM2 risk in patients association between metabolic disorders such younger than 50 years of age. Specifically, as DM2 and cardiovascular disease has been applying our knowledge of SNPs to DM2 risk known since the 1940s, but it wasn't until the appropriately reclassifies patients under the age 1980s that this association became more clearly of 50, but not older patients [22,23]. In addition, defined, and well accepted, and the term meta-it is also strikingly evident that environmental bolic syndrome (also known as syndrome X factors affect gene expression and function. or the dysmetabolic syndrome) was coined. Inducing gene promoter methylation has been Metabolic syndrome describes a cluster of shown to repress gene expression and modulat- metabolic risk factors that, when present in a ing key mitochondrial components results in single individual, increases this individual's risk impairment of essential metabolic function of of developing cardiovascular disease. The main cells. For example, altered metabolite patterns features of metabolic syndrome include abdom-related to lipid pathways may reflect changes in inal adiposity, dyslipidemia (elevated levels of insulin sensitivity [24].
serum triglycerides and decreased levels of serum Overall, these studies indicate that while high-density lipoproteins), high fasting blood genetic information measured in adulthood glucose levels and hypertension [25]. In addition does not seem to improve the ability to predict to these clinical factors, there are several other DM2 risk, there is support for the supposition factors that influence the prediction of cardio-that heterogeneous metabolite fingerprints vascular complications in individuals diagnosed and phenotypic characteristics better identify with DM2. These include metabolic factors such individuals who are in a prediabetic state and as glucose control, lipoprotein (a; Lp[a]) levels improve the ability to predict the risk of DM2 and microalbuminuria, in addition to genetic onset. The potential use of DNA variants variance-related factors such as variations in the influencing DM2 predisposition and obesity a diponectin gene and its receptor. lies in three main areas: the characterization To evaluate disease management in DM2, of disease mechanisms that provide new tar- physicians rely on fasting glucose levels as well gets for treatment and prevention; improved as HbA1c levels, a marker for average glucose risk prediction and differential diagnosis; and levels over the previous 3 months. Target lev-personalized treatment of DM2 and obesity. els of HbA1c were evaluated in several studies To date, the use of molecular diagnostic tools is in an attempt to determine whether the rate limited to screening of known causal genes for of diabetes-related cardiovascular complica-mutations that are often specific to a given fam- tions could be further decreased when levels of ily enabling a more precise diagnosis [23]. Thus, HbA1C in subjects with diabetes mellitus Type data mining of clinical digital records from 1 (DM1) are kept closer to the levels observed large populations will be helpful in improving in healthy individuals. In several studies, it was the ability to discover more accurate biomarkers found that lowering HbA1c to 6–6.5% slightly of diabetes subtypes and additional risk factors. increased, rather than decreased, cardiac-related These new discoveries, when introduced into mortality [26,27]. Other intervention studies have sophisticated algorithms, would improve our found that the incidence of new cardiovascular decision support system(s), providing physi- events is in fact reduced if postprandial hyper- cians with a greater ability to identify patients glycemia is well managed [28]. However, in with potential risks for developing diabetes and patients with poorly controlled DM2, intensive its complications.
glucose control had no significant effect on the rates of major cardiovascular events, death or Predictors of cardiovascular
microvascular complications [27]. On the other complications in DM2 patients
hand, follow-up studies to the UK Prospective An additional challenge in personalized medi- Diabetes Study (UKPDS) showed a nonsig- cine is identifying individuals prone to suffer nificant trend toward improvement in the rate from DM2-related complications using diagnos- of myocardial infarction in patients who are tic tests, whether looking at biological markers newly diagnosed with DM2 [14]. The Diabetes Personalized Medicine (2011) 8(4)
future science group PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karnieli PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karniel Metabolic profiling in personalized medicine PersPective Control and Complications Trial (DCCT) did on measures of Lp(a) levels when attempting to not show a significant reduction in cardiovas- predict cardiovascular risk in diabetic patients cular events with intensive control in young is still not sufficiently accurate and further patients with DM1, but a follow-up study, the research in this area is needed. Epidemiology of Diabetes Interventions and Microalbuminuria (albumin excretion of Complications (EDIC) trial, found a delayed 20–200 µg/min over a period of 24 h, or albu-benefit for these patients [29]. In a follow-up min/creatinine ration >30 mg/g in first morning study conducted 10 years after both groups midstream sample) has been found to be strongly (intensive or conventional treatment) reached associated with cardiovascular risk in DM2 [37]. similar HbA1c levels, the patients in the origi- Albuminuria reduction may improve cardio- nal intensive-therapy group had significantly vascular outcomes: a 50% decrease in albumin-fewer cardiovascular events than those in the uria has been found to be associated with an standard-therapy group. Similar results were 18 and 27% reduction in cardiovascular and seen in a 10-year follow-up of the UKPDS. heart failure risks, respectively [38]. In patients The Action to Control Cardiovascular Risk diagnosed with DM1, microalbuminurea might in Diabetes (ACCORD) trial [30] reported no serve as an independent predictor for cardiovas-significant decrease in cardio vascular events cular complications after 10 years or more after with intensive glucose control. Furthermore, the diagnosis of DM1 [39]. However, not all stud-the trial ended its intensive therapy early, after ies support the relationship between microalbu-only 3.5 years, because of a significant increase minuria and cardiovascular risk. Some studies in deaths in the intensive-therapy group. The have found that microalbuminuria is not a good Study to Prevent NIDDM (STOP-NIDDM) predictor of cardiovascular risk in nondiabetic found that treating individuals with impaired patients [40]. Thus, suppressing albuminuria glucose tolerance using postprandial hypergly- should be evaluated further as a potential way cemia medication reduces cardiovascular events of achieving optimal cardiovascular p rotection by 49% [31]. These data support the hypoth- in patients with both types of diabetes. esis that glycemic episodes play a role in the Genetic variation in diabetic-related genes is development of cardiovascular complications another area of research being pursued whose in DM2 patients. main focus is on identifying individuals at risk It has also been suggested that measuring for developing cardiovascular complications. Lp(a) levels might aid in predicting the risk of For example, the eNOS gene was reported to cardiovascular events in DM2 patients. Lp(a) be of clinical relevance to cardiovascular disease is a nontraditional biochemical marker that has as well as nephropathy in DM2 [41]. Recently, a been found to be predictive of cardiovascular mutation or polymorphism in eNOS (rs1799983 events [32,33]. Danesh et al. found a clear associa- GG-genotype) was found to be significantly and tion between cardiovascular events and levels inversely associated with cardiovascular disease of Lp(a) in the general population [34]. Normal in DM1 [41]. The mechanism by which such a Lp(a) values are considered any levels below polymorphism affects the phenotype is as yet 30 mg/dl. Higher than normal values of Lp(a) unclear, and further studies are needed to con-have been found to be associated with a high firm these findings. In addition, adiponectin, risk for atherosclerosis, stroke and heart attack. an adipocyte-derived hormone, has been found In addition, a study that followed DM2 patients to be involved in insulin action, while obtain-over a period of 10 years revealed that patients ing anti-inflammatory and antiatherogenic with low levels of Lp(a) (20 mg/dl) were at a effects. However, its plasma level is reduced in higher risk for cardiovascular complications [32]. obesity patients, which is likely contributing, at However, these studies primarily focused on least in part, to impaired insulin resistance. The high-risk individuals. Hence, extrapolating association between variants in the adiponectin these results to the entire diabetic population is gene (ADIPOQ) with circulating adiponectin problematic and more research is thus required. levels and cardio vascular risk among women For patients who have been diagnosed with dia- with DM2 has recently been examined [42]. betes for less than 10 years, several studies have The study found that promoter polymorphism found no clear-cut evidence of cardiovascular 11365C3G was significantly associated with complications [35,36]. Previous research has also lower plasma adiponectin levels. Furthermore, found that the ability to predict cardiovascular the patients' homozygous for allele 4034C was risk is less accurate as a patient's age progresses found to be significantly associated with a 60% over 55 years [36]. In conclusion, relying solely increase in cardiovascular risk. Controlling for future science group PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karnieli PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karniel Metabolic profiling in personalized medicine PersPective age, BMI and other covariants did not appre- There are two major known components of ciably change this association. In addition, a drug response: the concentration of active drug common haplotype possessing allele 276T available at its site(s) of action; and the abil-(CAATT) was associated with a significantly ity of the drug to elicit an effect at its site of lower cardio vascular risk than the most com- action. Drugs are often transported and acti- mon haplotype (CAATG). Another polymor- vated (from inactive to active state), or metabo- phism in the adiponectin gene (276G3T) was lized to an inactive form, before excretion from significantly associated with a 45% decrease in the body. Variation of metabolizing enzymes or cardiovascular risk under a recessive mode of transporters might therefore have an impact on inheritance in diabetic patients [42]. Genetic drug response. Furthermore, drug efficiency is variability and presence of polymorphisms, affected mainly by its ability to bind to a receptor, though hypothesized to be a key predictor the function of that receptor, and the function of illness and complications among diabetic of the downstream pathways (direct factors). In patients, still needs further investigation, par- addition factors that are different from the effec- ticularly when trying to determine the poten- tor pathway, such as disease etiology and drug tial mechanisms that could be used as effective combinations, may contribute to the effect of the tools for both d iagnostics and treatment.
drug potency (indirect factors) [46]. Such interac-tion may inhibit the metabolism of the drug or Personalized medicine for
prevent activation of certain m edications, and the treatment of DM2 & its
should be further evaluated [47]. Several studies investigating variation in A variety of treatment options exist for individu- response to DM2 medications have been per- als with DM2. In addition to dietary and physi- formed. For example, studies of the transport cal activity, DM2 patients are currently treated mechanism of metformin, a medication com-pharmacologically with nine major classes of monly administered to diabetes patients, have approved drugs. These medications include recently identified organic cation transporters biguanides, sulfonylureas, thiazolidinediones, that play a role in metformin disposition. Plasma meglitinides, a-glucosidase inhibitors, amylin membrane monoamine transporter (SLC29A4) analogues, incretin-mimetics, dipeptidyl pepti- has been found to be involved in gut absorp- dase 4 (DPP4) inhibitors and insulin and its ana- tion [48], OCT1's (SLC22A1) primarily involve- logues [43]. In addition, patients with DM2 are ment in hepatic uptake, and OCT2 (SLC22A2) often treated with medications for DM2-related in tubular secretion [49,50]. Further studies in complications such as statins to treat hyperlip- humans have demonstrated higher serum met- idemia, angiotensin-converting enzyme inhibi- formin concentrations in individuals carrying the tors, angiotensin receptor blockers , b-adrenergic reduced function polymorphisms of OCT1. This blockers, calcium channel blockers and diuretics indicates that this transporter is important for for hypertension, and antiplatelet agents. Initial metformin's therapeutic action and that genetic therapy for DM2 focuses on lifestyle changes and variation in OCT1 may contribute to variation in administration of metformin (a biguanide). The response to the drug [51]. Metformin acts by acti-rapid addition of other medications and transi- vating adenosine monophosphate-activated pro- tion to new regimens follow when target glycemic tein kinase (AMPK), which leads to suppression goals are not achieved or sustained, and insulin of glucose production via gluconeogenesis and therapy is added for patients who do not meet tar- slightly increased peripheral glucose uptake [52]. get goals [44]. It is recommended that all patients Inhibition of hepatic gluconeogenesis by metfor-achieve near normoglycemia and HbA1C levels min occurs through AMPK-dependent regulation below 7% [45]. Often the response to a specific of the orphan nuclear receptor small heterodimer medication varies between patients, and the partner [53]. Protein-threonine kinase (LKB1), physician may need to make several adjustments which phosphorylates and activates AMPK, is before finding the most appropriate treatment critical for the glucose-lowering effects of metfor-regimen for an individual patient. A primary min in the liver [54]. In a recent study involving focus of personalized medicine is to provide the participants from the DPP study, common varia-physician with tools that aid in selecting the tions in 40 candidate genes previously associated most effective treatment regimen for a patient, with DM2 were analyzed to study their impact while decreasing the possibility for adverse events on diabetes incidence and their interaction with and c omplications related to the recommended response to metformin and lifestyle interven-m edical care. tions [6]. The genes encoding the AMPK kinase Personalized Medicine (2011) 8(4)
future science group PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karnieli PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karniel Metabolic profiling in personalized medicine PersPective STK11 and the AMPK subunit genes PRKAA1, of the ABCC8 gene and rs5210 of the CNJ11 PRKAA2 and PRKAB2 were found to be associ- gene were identified as significantly associated ated with the response to metformin. This prom- with reduction in FPG. Compared with subjects ising finding warrants further investigation [55]. with the Ser/Ser genotype, subjects with the Ala/In addition to these effects, metformin may Ala genotype were found to have a 7.7% greater also exert a direct effect on pancreatic b-cells. decrease in FPG and an 11.9% reduction in 2-h In humans, metformin causes increased insulin plasma glucose. No difference in HbA1c was release in response to glucose [56] and may help found [59].
to preserve b-cell function [57]. In addition, it was In addition to varied responses to DM2 med- recently shown in mice that metformin modu- ications, individual differences are commonly lates multiple components of the incretin axis, found in the levels of risk for DM2-related com-mainly by enhancing the expression of the GLP-1 plications. Individuals with both DM2 and the receptor and related insulinotropic islet receptors haptoglobin (Hp) 2–2 genotype are at increased through a mechanism requiring PPAR-a [58].
risk of cardiovascular disease. The antioxidant Sulfonylureas (SU) are metabolized primarily function of the Hp 2–2 protein is often impaired. by the cytochrome P450 2C9 enzyme. Variations As a result, several studies have focused on the in the common allele for CYP2C9 affect the cata- effect of antioxidant vitamin E supplementa-lytic function of the enzyme. These variations, tion on cardiovascular events (stroke, myocar-(Arg144Cys 2C9*2; allele frequency 11% and Ile- dial infarction and cardio vascular death) in Hp 359Leu 2C9*3; allele frequency 7%) may change 2–2 DM individuals. These studies found a the effect of these medications by changing their significant overall reduction in cardiovascular concentration in the serum. Most studies have events [61]. Although in clinical studies vitamin E found that individuals carrying the *2/*3 or supplementation did not provide any mea sure of *3/*3 genotype show reduced drug-metaboliz- cardiovascular protection in unselected popula-ing activities, with a lower dose requirement, as tions with DM2 [62], individuals with Hp 2–2 compared with individuals having the wild-type may benefit from this treatment. In subpopula-Arg144/Ile359 (CYP2C9*1) allele. In healthy tions derived from these trial cohorts, Hp 2–2 volunteers receiving glimepiride, CYP2C9 geno- diabetic patients who were administered vita-type altered the pharmacokinetic profile of the min E appeared to benefit from this treatment by drug significantly, with a much slower elimina- demonstrating decreased rates of cardiovascular tion of glimepiride in individuals carrying the *3 events along with increased life expectancy. This
allele compared with those who carry the *1/*1 finding further emphasizes the importance and
genotype. When drug elimination is delayed, advantages of engaging in personalized medicine
lower doses of the medication can be prescribed over traditional clinical practices.
to patients to achieve the desired effect without
increasing the risk for side effects such as severe Prospective tools for personalizing
hypoglycemia [43]. SUs act by binding to the early diagnosis & treatment
SUR1 moiety of the pancreatic b-cell KATP of diabetes
channel, causing the channel to close and trigger New innovative tools that were originally devel-
insulin secretion. Genetic variation in this path-
oped to study diabetes may also be found to way was identified in subjects with a mutation in be advantageous for future use in personalized the TCF1 gene encoding hepatocyte nuclear fac- medical care. Examples of such tools include tor-1 a (HNF-1a), causing altered SU response. common laboratory markers not currently asso-In a randomized trial involving patients with ciated with diabetes or insulin resistance, new DM2 due to TCF1 mutations, a treatment of SUs technologies for comprehensive ana lysis, such and metformin showed that the decrease in FPG as ‘metabolomics', and employing methods for in response to SU medication such as gliclazide metabolic profiling of discrete small molecule was 3.9-fold greater than their response to met- metabolites using nuclear magnetic resonance formin. In subjects with DM2, no differences in and mass spectrometry. In the past, molecules response to gliclazide or metformin were appar- such as uric acid have been shown to be asso- ent [59,60]. Several studies have been performed ciated with insulin response to a glucose load in an attempt to identify polymorphisms in the in both men and women [63]. Recent studies KATP channel and downstream pathways that of small molecule metabolites, such as amino influence SU response. When 25 SNPs in 11 can- acids and fatty acids, represent the net result of didate genes were examined in a prospective trial genomic, transcriptomic and proteomic vari-of patients treated with gliclazide, Ser1369Ala ability, providing an integrated biological status future science group PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karnieli PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karniel Metabolic profiling in personalized medicine PersPective profile. In addition to studies of the mechanisms the complexity of available data we believe that leading to the development of metabolic diseases, in the near future, during physician–patient these molecules could also be used to study the encounter a specific platform will be applied action mechanisms of commonly used medica- whereby from a single blood test/tissue sample, tions, responses to treatment and adverse event physicians could analyze DNA profile, proteins profiles [64]. Measurements of amino acids, and metabolites, as well as clinical data to find acylcarnitines, free fatty acids and conventional potential therapeutic options suitable for the indi-metabolites such as glucose, lactate, uric acid, vidual patient. Developing the regulatory aspect total ketones, hydroxybutyrate and nonesteri- of such an undertaking is of great importance, fied free fatty acids in sedentary, overweight to particularly as it involves the use of genomic, pro-obese, dyslipidemic individuals have found that teomic and metabolomic information in drug and elevated concentrations of large, neutral amino biomedical device development as well as clinical acids were independently associated with insulin decision-making. Regulation of such requirements resistance, and that large neutral amino acids and must be very strict, should include data protection, fatty acids were related to appropriate pancreatic confidentiality requirements and quality control responses [65]. In a study aimed to elucidate the of genomic or proteomic databases. Furthermore, effect of exercise training on insulin sensitivity, patients should be able to expect to receive the leptin, adipo nectin, D-dimer, paraoxonase activ- right diagnostics and treatment based on tests ity cytokines, inflammatory markers and meta- carried out during a single visit to the hospital, bolic intermediaries were measured at baseline clinic or a physician's office. They should be able and after 6 months of aerobic training. Four fac- to be certain that the diagnostic tests administered tors were found to be independently associated to them will give precise and correct results, and with change in insulin sensitivity: free fatty acids that the treatment decisions made by physicians and by-products of fatty acid oxidation, glycine based on those test results are a result of appropri-and proline, acylcarnitine and C18:1-OH acyl- ate analyses. Regulation should also include more carnitine. Modeling indicated that improvements accurate methods of DNA/protein/drug ana lysis, in insulin sensitivity were retained 15 days after further ensuring that results are correct the first cessation of exercise training and, interestingly, time. The clinical benefits of such an approach will greater sustained levels were seen in men than in increase as new products and approaches are incor-women [65]. Further research into the role of these porated into daily clinical practice. As the field factors in the prevention and treatment of DM2, further develops, it is expected that more efficient with concomitant development of tools for per- clinical trials will be developed based on b etter sonalized medicine, are required before this u nderstanding of the genetic basis of disease.
i nformation could be used in c linical settings. In order to more effectively identify, accurately Conclusion & future perspective
diagnose and successfully treat individuals at risk Advances in personalized medicine would for DM2 and related complications, physicians greatly improve patient care in the field of dia-must have access to readily available, reasonably betes and other metabolic diseases. Several chal-priced and reliable tools. Once this is accom- lenges have to be met before this possibility pro- plished, physicians would be able to use these gresses from theory to practice, and the vision tools to tailor individualized treatment regimens of providing individualized medical care for by helping to reveal which patients would benefit metabolic diseases as standard practice advances from specific treatments while, at the same time, towards realization. Physicians and scientists are limiting negative side effects. Moreover, discover- working toward the time when, based on clinical ing which gene mutations or their products could and molecular data, patients suffering from a affect treatment quality would also contribute to metabolic disease such as DM2 can be stratified improved patient health. according to disease risk, and their risk of devel- Several government organizations (e.g., oping disease-related complications. Clinical US FDA and NIH) are currently developing a information obtained during an office visit and regulatory platform for evaluating tools that can molecular data derived from on site ana lysis of connect a range of diagnostic and treatment steps, various domains such as genomics, proteomics, from the identification of a potential therapeutic epigenetics and metabolomics and specific bio-target by academic researchers to the approval of markers, will be combined with information a therapy for clinical use [66]. As direct-to-con- obtained from ana lysis of large population sumer genome-wide profiling is readily available databases and pertinent clinical guidelines and (although not yet approved by FDA), and given integrated into a decision support system that Personalized Medicine (2011) 8(4)
future science group PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karnieli PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karniel Metabolic profiling in personalized medicine PersPective would be used by physicians at the point-of-care ƒ Tools that can provide more robust ana lysis of to provide the best possible outcomes for their the data contained in these large-scale data- patients (Figure 1). Specifically, this increased use
bases are needed. Importantly, the results of of personalized medicine will provide physi- these analyses should include identification of cians with the ability to plan the most effective a set of biomarkers that indicate the expected therapy protocol for their patients, while mini- disease course in an individual patient, as well mizing possible adverse drug events. To achieve as that patient's potential response to various these goals, several advances should be made: treatment options. In the field of biology, biomarkers for disease ƒ Medical devices that could be used by physi- course and management need to be identi- cians in their office during a scheduled visit fied. These include establishing a set of pre- need to be developed. These devices should be dictive biomarkers that would accurately reliable and provide clinically relevant output, identify patients at risk for developing diabe- in a timely manner, and at an acceptable cost. tes and the risk for developing diabetes-related The technology used in these devices should c omplications.
be able to integrate the individual's clinical, ƒ In the field of biomedical informatics, data col- laboratory and genetic information, analyze lection for the identification and validation of known biomarkers and provide guidance for proposed biomarkers needs to be improved. physicians that could be used in an individual This includes the development and accessibility patient's treatment. In addition, the technol- of well organized, large scale, multiuser data- ogy behind these devices should be able to bases that contain clinical, laboratory and assess the response of a particular patient to a molecular information from large samples of specific medical treatment, recommending individual patients. dose adjustments and avoiding medications Point of care
Relevant biomarkers analysis,
Population database single step biochip clinical guidelines Electronic
Figure 1. The decision process of tailoring individualized medical care in complex diseases.
future science group PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karnieli PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karniel Metabolic profiling in personalized medicine PersPective that could potentially lead to negative out- Financial & competing interests disclosure
comes. This technology would need to be This work was supported in part by NSF/NIST grant IIP- made available in a cost-effective manner to 0934364 (Mary Brady, Program Director), by NSF IIS- primary care physicians and s pecialists serving 0837716, CNS-0821345, HRD-0833093, IIP-0829576, the general population. IIP-0931517, CNS-1057661, IIS-1052625, CNS-0959985, by Galil Center, Linda Rose Diamond grant: Accomplishment of these goals would greatly Technion, and by Israel Science Foundation (1235/08), impact the prediction, prevention and treat- University of Maryland Baltimore County, Florida ment of diabetes, as well as provide substantial International University, NOA Inc, and DrRecommend.
advances in personalized medicine. Furthermore, com. The authors have no other relevant affiliations or it would greatly improve the quality of medical financial involvement with any organization or entity with care while, at the same time, reducing costs.
a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. The authors thank Dr Derek Leroith for his comments and No writing assistance was utilized in the production of Dr Debra Davis for editing this manuscript. this manuscript. Executive summary
Scope of the problem
ƒ Currently available scientific information holds the key to progress in clinical medicine.
ƒ The information could be used to predict the individual's potential medical course.
ƒ Tools to extract information and support clinical decisions are not available in clinical practice.
ƒ The focus of this article is on future possibilities in employing personalized medicine for improving the management of Type 2 diabetes mellitus (DM2).
ƒ Diabetes is the most common metabolic disease, expected to affect 44.1 million in the USA by 2034.
ƒ Diabetes complications have a significant economic impact on the health system worldwide.
ƒ There is a need to identify an individual at risk for various diabetes-related complications.
ƒ Tailoring of medical treatment to the individual characteristics of each patient will minimize diabetes complications and the potential for adverse side effects.
Identifying asymptomatic subjects at risk of developing DM2
ƒ Clinical factors such as obesity, physical inactivity, family history can predict the development of DM2.
ƒ Laboratory factors such as fasting glucose, sex hormone-binding globulin, inflammatory mediators are associated with increased risk for developing DM2.
ƒ Common genetic variants associated with DM2 have slightly improved the ability to predict future onset of DM2.
Predictors of cardiovascular complications in DM2 patients
ƒ Potential predictors of cardiovascular complications in DM2 patients are glucose control, lipoprotein (a) levels, microalbuminuria and genetic variance.
ƒ Subject with haptoglobin 2–2 genotype are at increased risk of cardiovascular disease and may benefit from treatment with vitamin E.
Prospective tools for personalizing early diagnosis & treatment of diabetes
ƒ Metabolic profiling of discrete small molecule metabolites using nuclear magnetic resonance and mass spectrometry such as amino acids and fatty acids.
ƒ Development of a regulatory platform for evaluating tools that can connect a range of diagnostic and treatment steps.
Most updated guidelines of the
Lutgers HL, Gerrits EG, Sluiter WJ et al. Life American Diabetes Association's (ADA)
expectancy in a large cohort of Type 2 diabetes Papers of special note have been highlighted as: recommendations for treating diabetes.
patients treated in primary care (ZODIAC-10). Franco OH, Steyerberg EW, Hu FB, PLoS ONE 4(8), e6817 (2009).
nn of considerable interest Mackenbach J, Nusselder W. Associations of Knowler WC, Barrett–Connor E, Fowler SE Huang ES, Basu A, O'Grady M, diabetes mellitus with total life expectancy et al. Reduction in the incidence of Type 2 Capretta JC. Projecting the future and life expectancy with and without diabetes with lifestyle intervention or diabetes population size and related costs for cardiovascular disease. Arch. Intern. Med. 167, metformin. N. Engl. J. Med. 346, 393–403 the US. Diabetes Care 32, 2225–2229 1145–1151 (2007).
Gu K, Cowie CC, Harris MI. Mortality in Tai ES, Goh SY, Lee JJ et al. Lowering the Update of ADA's major position statement, adults with and without diabetes in a criterion for impaired fasting glucose: impact ‘standards of medical care in diabetes'. national cohort of the US population, on disease prevalence and associated risk of Introduction. Diabetes Care 34(Suppl. 1), 1971–1993. Diabetes Care 21, 1138–1145 diabetes and ischemic heart disease. Diabetes S1–S2 (2011).
Care 27, 1728–1734 (2004).
Personalized Medicine (2011) 8(4)
future science group PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karnieli PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karniel Metabolic profiling in personalized medicine PersPective Schriger DL, Lorber B. Lowering the cut Identifying undiagnosed and
34 Danesh J, Collins R, Peto R. Lipoprotein(a) point for impaired fasting glucose: where is and coronary heart disease. Meta-ana lysis of the evidence? Where is the logic? Diabetes 22 De Miguel-Yanes J, Shrader P, Pencina M prospective studies. Circulation 102, Care 27, 592–601 (2004).
et al. Genetic risk reclassification for Type 2 1082–1085 (2000).
Tirosh A, Shai I, Tekes-Manova D, Israeli E diabetes by age below or above 50 years using 35 Abu-Lebdeh HS, Hodge DO, Nguyen TT. et al. Normal fasting plasma glucose levels 40 Type 2 diabetes risk single nucleotide Predictors of macrovascular disease in patients and Type 2 diabetes in young men. N. Engl. polymorphisms. Diabetes Care 34, 121–125 with Type 2 diabetes mellitus. Mayo Clin. J. Med. 353, 1454–1462 (2005).
Proc. 76, 707–712 (2001).
10 Tirosh A, Shai I, Bitzur R et al. Changes in 23 Lucio M, Fekete A, Weigert C et al. Insulin 36 Sunayama S, Daida H, Mokuno H et al. Lack triglyceride levels over time and risk of Type 2 sensitivity is reflected by characteristic of increased coronary atherosclerotic risk due diabetes in young men. Diabetes Care 31, metabolic fingerprints--a Fourier transform to elevated lipoprotein(a) in women > or = 2032–2037 (2008).
mass spectrometric non-targeted 55 years of age. Circulation 94, 1263–1268 11 Polderman KH, Gooren LJ, Asscheman H, metabolomics approach. PLoS ONE 5, e13317 Bakker A, Heine RJ. Induction of insulin 37 Mogensen CE. Microalbuminuria predicts resistance by androgens and estrogens. J. Clin. Identifying undiagnosed and
clinical proteinuria and early mortality in Endocrinol. Metab. 79, 265–271 (1994).
maturity-onset diabetes. N. Engl. J. Med. 310, 12 Golden SH, Dobs AS, Vaidya D et al. 356–360 (1984).
24 Feero WG, Guttmacher AE, McCarthy MI. Endogenous sex hormones and glucose tolerance Genomics, Type 2 diabetes, and obesity. 38 de Zeeuw D. Albuminuria, not only a status in postmenopausal women. J. Clin. N. Engl. J. Med. 363, 2339–2350 (2010).
cardiovascular/renal risk marker, but also a Endocrinol. Metab. 92, 1289–1295 (2007).
target for treatment? Kidney Int. Suppl. 92, 25 Gallagher EJ, Leroith D, Karnieli E. Insulin 13 Ding EL, Song Y, Manson JE et al. Sex S2–S6 (2004).
resistance in obesity as the underlying cause hormone-binding globulin and risk of Type 2 for the metabolic syndrome. Mt Sinai J. Med. 39 Messent JW, Elliott TG, Hill RD, Jarrett RJ, diabetes in women and men. N. Engl. J. Med. 77, 511–523 (2010).
Keen H, Viberti GC. Prognostic significance 361, 1152–1163 (2009).
of microalbuminuria in insulin-dependent 26 Cheng AY, Leiter LA. Cardiovascular risk and 14 Albertini JP, Valensi P, Lormeau B et al. diabetes mellitus: a twenty-three year glycemic control. CMAJ 180, 907–908 Elevated concentrations of soluble E-selectin follow-up study. Kidney Int. 41, 836–839 and vascular cell adhesion molecule-1 in 27 Duckworth W, Abraira C, Moritz T et al. NIDDM. Effect of intensive insulin treatment. 40 Agewall S, Wikstrand J, Ljungman S, Glucose control and vascular complications in Diabetes Care 21, 1008–1013 (1998).
Fagerberg B. Usefulness of microalbuminuria veterans with Type 2 diabetes. N. Engl. 15 Meigs JB, Hu FB, Rifai N, Manson JE. in predicting cardiovascular mortality in J. Med. 360, 129–139 (2009).
Biomarkers of endothelial dysfunction and treated hypertensive men with and without 28 Ceriello A, Hanefeld M, Leiter L et al. risk of Type 2 diabetes mellitus. JAMA 291, diabetes mellitus. Risk factor intervention study Postprandial glucose regulation and diabetic 1978–1986 (2004).
group. Am. J. Cardiol. 80, 164–169 (1997).
complications. Arch. Intern. Med. 164, 16 Sladek R, Rocheleau G, Rung J et al. A 41 Mollsten A, Lajer M, Jorsal A, Tarnow L. The 2090–2095 (2004).
genome-wide association study identifies endothelial nitric oxide synthase gene and risk 29 Nathan DM, Cleary PA, Backlund JY et al. novel risk loci for Type 2 diabetes. Nature of diabetic nephropathy and development of Intensive diabetes treatment and 445, 881–885 (2007).
cardiovascular disease in Type 1 diabetes. cardiovascular disease in patients with Type 1 Mol. Genet. Metab. 97, 80–84 (2009).
Novel risk loci for Type 2 diabetes.
diabetes. N. Engl. J. Med. 353, 2643–2653 42 Qi L, Doria A, Manson JE et al. Adiponectin 17 Meigs JB, Shrader P, Sullivan LM et al. genetic variability, plasma adiponectin, and Genotype score in addition to common risk 30 Gerstein HC, Miller ME, Byington RP et al. cardiovascular risk in patients with Type 2 factors for prediction of Type 2 diabetes. Effects of intensive glucose lowering in Type 2 diabetes. Diabetes 55, 1512–1516 (2006).
N. Engl. J. Med. 359, 2208–2219 (2008).
diabetes. N. Engl. J. Med. 358, 2545–2559 43 Distefano JK, Watanabe RM. 18 Lyssenko V, Jonsson A, Almgren P et al. Pharmacogenetics of anti-diabetes drugs. Clinical risk factors, DNA variants, and the 31 Chiasson JL, Josse RG, Gomis R, Hanefeld Pharmaceuticals (Basel) 3, 2610–2646 (2010).
development of Type 2 diabetes. N. Engl. M, Karasik A, Laakso M. Acarbose treatment J. Med. 359, 2220–2232 (2008).
and the risk of cardiovascular disease and Describes polypharmacy and drug interaction.
19 Lindstrom J, Tuomilehto J. The diabetes risk hypertension in patients with impaired 44 Nathan DM, Buse JB, Davidson MB et al. score: a practical tool to predict Type 2 diabetes glucose tolerance: the STOP-NIDDM trial. Medical management of hyperglycaemia in risk. Diabetes Care 26, 725–731 (2003).
JAMA 290, 486–494 (2003).
Type 2 diabetes mellitus: a consensus algorithm for the initiation and adjustment of 20 Grant RW, Moore AF, Florez JC. Genetic 32 Hernandez C, Francisco G, Chacon P, Simo therapy: a consensus statement from the architecture of Type 2 diabetes: recent R. Lipoprotein(a) as a risk factor for American diabetes association and the progress and clinical implications. Diabetes cardiovascular mortality in Type 2 diabetic European association for the study of Care 32, 1107–1114 (2009).
patients: a 10-year follow-up study. Diabetes Care 28, 931–933 (2005).
diabetes. Diabetologia 52, 17–30 (2009).
21 Wang J, Stancakova A, Kuusisto J, Laakso M. 45 Standards of medical care in diabetes – 2010. Identification of undiagnosed Type 2 diabetic 33 Kollerits B, Auinger M, Reisig V et al. Diabetes Care 33(Suppl. 1), S11–S61 (2010).
individuals by the Finnish diabetes risk score Lipoprotein(a) as a predictor of cardiovascular and biochemical and genetic markers: a disease in a prospectively followed cohort of 46 Pearson ER. Pharmacogenetics in diabetes. population-based study of 7232 Finnish men. patients with Type 1 diabetes. Diabetes Care Curr. Diab. Rep. 9, 172–181 (2009).
J. Clin. Endocrinol. Metab. 95, 3858–3862 29, 1661–1663 (2006).
47 Hutson S. Researchers sound alarm on ‘silent' drug interactions. Nat. Med. 17, 6 (2011).
future science group PersPective Zolotov, Ben Yosef, Rishe, Yesha & Karnieli Describes polypharmacy and
54 Shaw RJ, Lamia KA, Vasquez D et al. The 61 Feng Y, Mao G, Ren X et al. Ser1369Ala kinase LKB1 mediates glucose homeostasis in variant in sulfonylurea receptor gene ABCC8 48 Zhou M, Xia L, Wang J. Metformin transport liver and therapeutic effects of metformin. is associated with antidiabetic efficacy of by a newly cloned proton-stimulated organic Science 310, 1642–1646 (2005).
gliclazide in Chinese Type 2 diabetic patients. cation transporter (plasma membrane 55 Jablonski KA, McAteer JB, de Bakker PI et al. Diabetes Care 31, 1939–1944 (2008).
monoamine transporter) expressed in human Common variants in 40 genes assessed for 62 Yusuf S, Dagenais G, Pogue J, Bosch J, intestine. Drug. Metab. Dispos. 35, 1956–1962 diabetes incidence and response to metformin Sleight P. Vitamin E supplementation and and lifestyle intervention in the diabetes cardiovascular events in high-risk patients. The 49 Kimura N, Masuda S, Tanihara Y et al. prevention program. Diabetes 59, 2672–2681 heart outcomes prevention evaluation study Metformin is a superior substrate for renal investigators. N. Engl. J. Med. 342, 154–160 organic cation transporter OCT2 rather than 56 Marchetti P, Scharp DW, Giannarelli R et al. hepatic OCT1. Drug Metab. Pharmacokinet. Metformin potentiates glucose-stimulated 63 Bain JR, Stevens RD, Wenner BR et al. 20, 379–386 (2005).
insulin secretion. Diabetes Care 19, 781–782 Metabolomics applied to diabetes research: 50 Wang DS, Jonker JW, Kato Y, Kusuhara H, moving from information to knowledge. Schinkel AH, Sugiyama Y. Involvement of 57 Marchetti P, Lupi R, Del Guerra S et al. Goals Diabetes 58, 2429–2443 (2009).
organic cation transporter 1 in hepatic and of treatment for Type 2 diabetes: b-cell 64 Modan M, Halkin H, Karasik A, Lusky A. intestinal distribution of metformin. preservation for glycemic control. Diabetes Elevated serum uric acid – a facet of J. Pharmacol. Exp. Ther. 302, 510–515 Care 32(Suppl. 2), S178–S183 (2009).
hyperinsulinaemia. Diabetologia 30, 713–718 58 Maida A, Lamont BJ, Cao X, 51 Shu Y, Sheardown SA, Brown C et al. Effect Drucker DJ. Metformin regulates the 65 Huffman KM, Shah SH, Stevens RD et al. of genetic variation in the organic cation incretin receptor axis via a pathway Relationships between circulating metabolic transporter 1 (OCT1) on metformin action. dependent on peroxisome proliferator- intermediates and insulin action in J. Clin. Invest. 117, 1422–1431 (2007).
activated receptor-a in mice. Diabetologia overweight to obese, inactive men and 52 Zhou G, Myers R, Li Y et al. Role of 54(2), 339–349 (2011).
women. Diabetes Care 32, 1678–1683 (2009).
AMP-activated protein kinase in mechanism 59 Pearson ER, Starkey BJ, Powell RJ, Gribble 66 Hamburg MA, Collins FS. The path to of metformin action. J. Clin. Invest. 108, FM, Clark PM, Hattersley AT. Genetic personalized medicine. N. Engl. J. Med. 363, 1167–1174 (2001).
cause of hyperglycaemia and response to 301–304 (2010).
53 Kim YD, Park KG, Lee YS et al. treatment in diabetes. Lancet 362, Metformin inhibits hepatic gluconeogenesis 1275–1281 (2003).
through AMP-activated protein 60 Blum S, Vardi M, Brown JB et al. Vitamin E kinase-dependent regulation of the orphan reduces cardiovascular disease in individuals 101 Diabetes Public Health Resource nuclear receptor SHP. Diabetes 57, 306–314 with diabetes mellitus and the haptoglobin 2–2 genotype. Pharmacogenomics 11, 675–684 102 WHO: Diabetes fact sheet 312 Personalized Medicine (2011) 8(4)
future science group




Assessment of the U.S. Fibromyalgia Market in Support of S1 Filing (Phase 2) Presented to TONIX Pharmaceuticals Frost & Sullivan Healthcare and Life Sciences Consulting Group CONFIDENTIAL This report is solely for the use of client personnel. No part of it may be circulated, quoted, or reproduced for distribution outside the client organization without prior written approval from Frost & Sullivan. This material was used by Frost & Sullivan during an oral presentation; it is not a complete record of the discussion.