24 th Spring Congress of Korean Diabetes Association Increased risk of cancer with anti-diabetes drugs? : pros Nan Hee Kim, MD, Ph.D. Professor Korea University Medical School
Contents Diabetes and cancer - Hyperinsulinemia - IGF-I Diabetes therapy and cancer - Sulfonylurea & insulin secretagogues - Insulin & insulin analogues - Incretin
Diabetes and risk of colorectal cancer Larsson SC, J Natl Cancer Inst 97: 2005
Diabetes and risk of pancreatic cancer Huxley R, British Journal of Cancer 92: 2005
Diabetes and risk of breast cancer Larsson SC, Int J Cancer 121: 2007
Diabetes and fatal cancer in men Coughlin SS, Am J Epidemiol 159: 2004
Diabetes and fatal cancer in women
Plasma glucose regulation and mortality in Korea : A pooled analysis of three community-based cohort studies Yonchon, Jungeup, and Ansan studies, N=3801 Median f/u: 11yr, N. of all-cause death: 474 The Committee on the Epidemiology of Diabetes, KDA (unpublished)
Plasma glucose regulation and mortality in Korea a) FPG b) 2hPG
Biological plausibility of a link between diabetes, diabetes therapies and cancer Hyperinsulinemia Increased IGF-I activity
Insulin receptor (IR) contents and IR isoform expression in cancer Many cancer cells have an increased IR contents In cancer cells, the IR-A expression is predominant By binding to overexpressed IR-A, insulin may favor cancer progression and facilitate the growth of tumors Vigneri P, Endocrine-Related Cancer 16: 2009
Insulin and IGF-1 receptor signaling Insulin resistance
Hyperinsulinemia, not hyperglycemia or lipidemia acutely enhances colorectal epithelial proliferation in vivo HIHG, COMB HIHG, COMB, H.INS I.INS SAL, LIPID O SAL: saline, HIHG: high insulin, high glucose I.INS: intermediate level of insulin, LIP: lipid H.INS: high level of insulin, COMB: combination of LINS,HG,LIP Tran TT, Endocrinology 147: 2006
Insulin promotion of colon tumors in rats Insulin is a colon tumor promoter in this azoxymethane (carcinogen) injected rat models Group Total no. of rats Rats with tumors (tumor incidence) Total no. of tumors Tumors/rat (Tumor multiplicity) (mean ± SEM) Saline 30 15 (50%) 22 0.73 ± 0.16 a b Insulin 29 23 (79%) 58 2.00 ± 0.33 c Tran TT, Cancer Epidemiol Biomarker Prev 5: 1996
Fasting insulin and outcome in breast cancer In a cohort of 512 women without diabetes, who had early stage breast cancer, survival according to fasting insulin levels Insulin receptor expression was increased in several cancer cells Insulin had mitogenic and tumor promoting effect, and was associated with poor prognosis in cancer patients Pamela J, J Clin Oncol 20: 2001
Signaling of insulin, IGFs in tumor cells IGFBP 1,3
Comparison of highest and lowest IGF-I categories by cancer site Renehan AG, Lancet 353: 2004
More evidences of association of IGF-I and cancer IGF-IR expression is up-regulated in human prostate cancer Tissue over-expression of IGF-I stimulates mammary tumorigenesis Concentration of carcinogen Hellawell GO, Cancer Res, 62: 2002 De Ostrovich KK, Am J Pathol 173: 2008
More evidences of association of IGF-I and cancer Crossing prostate cancer-predisposed TRAMP mice with GH and IGF-I deficient mice Phosphorylated IGF-IR/IR staining in 438 invasive breast cancer tissues TRAMP+GH/IGF-I deficient mice negative positive TRAMP mice negative positive Majeed N, Oncogene 24: 2005 Law JH, Cancer Res 68; 2008
IGF-I contributes to Trasuzumab resistance to breast cancer cells IGF-I activation may cause chemo-resistance, hormonal resistance, resistance to target drug e.g. Trasuzumab (Herceptin ) IGF-I receptor expression was increased in several cancer cells IGF-I overexpression stimulated tumorigenesis, whereas IGF-I deficiency delayed carinogenesis Tissue IGF-I/IR expression was associated with poor prognosis in cancer patients (possibly associated with chemo-resistance) Nahta R, Cancer Res 65: 2005 Lu Y, J Natl Cancer Inst 93: 2001
Do diabetic medications increase the risk of cancer? Sulfonylurea Insulin Insulin analogue
Sulfonylurea and cancer: case-control study Case: 195 diabetic patients with incident malignancy Control: 195 diabetes without cancer, matched for age, sex, duration of DM, BMI, HbA1c, comorbidity, alcohol, smoking Exposure to hypoglycemic drugs during 10 years preceding the event Age 69.0±9.2 years, 59% men Monami M, Acta Diabetol 46: 2009
Sulfonylurea, insulin and mortality from cancer Population-based retrospective cohort study in Canada New users of metformin or sulfonylureas (N=10309), f/u 5.4±1.9 years Age 63.4±13.3 years, 55% men, 474 cancer deaths occurred Variable Oral glucose-lowering agents Time at risk (person-days) Cancer deaths N, (%) Sulfonylurea monotherapy 16,737 162 (4.9) 1.0 Adjusted HR (95% CK)ª Metformin±SU 38,973 245 (3.5) 0.80 (0.65-0.98) Insulin use (dispensations/year) No insulin ever 52,131 323 (3.6) 1.0 <3 400 6 (3.8) 2.22 (0.99-5.00) 3 to 11 1,543 29 (5.0) 3.33 (2.26-4.89) 12 1,363 49 (7.0) 6.40 (4.69-8.73) Adjusted for age, sex and comorbidity Bowker SL, Diabetologia 53: 2010
Insulin secretagogues and mortality from cancer Hospital based retrospective observational study in 3002 diabetics During f/u 5.4±1.9 years, 87 cancer-related deaths There were limited evidence of the association between SU and cancer Insulin sensitizers (N=444, HR=0.77) No treatment (N=727, HR=1) In a case-control study, glibenclamide Insulin (N=421, HR=2.26) had higher OR for cancer, whereas gliclazide had lower OR Mortality from cancer was Scretagogues higher in SU (N=372, group HR=2.29) than those with metformin or no treatment group Adjusted for age, sex, BMI, comorbidity Monami M, Exp Clin Endocrinol Diabetes 116: 2008
Insulin therapy and colorectal cancer Retrospective case-control study of the General Practice Research Database from UK Insulin users (at least > 1yr, N=3160) vs. noninsulin-using T2DM patients (N=24918) for occurrence of colorectal cancer Age 74.9±8.4 yrs, f/u 5.6±1.8 yrs insulin users noninsulin users Yang Y, Gastroenterology 127: 2004
Insulin therapy and colorectal cancer Nested case-control analysis (matched for age, duration of DM) Unadjusted OR (95% CI) Adjusted OR (95% CI) No insulin therapy (reference) 1.0 1.0 1 year and <3 years of insulin use 1.2 (0.6-2.6) 1.4 (0.6-2.9) 3 years and <5 years of insulin use 2.2 (0.9-5.5) 2.9 (1.1-7.7) 5 years of insulin use 2.8 (0.9-8.5) 4.7 (1.3-16.7) Each incremental year of insulin therapy 1.13 (0.99-1.3) 1.2 (1.03-1.4) Adjusted for sex, history of cholecystectomy, smoking, duration of diabetes, BMI, use of metformin, sulfonylurea, or NSAID Yang Y, Gastroenterology 127: 2004
Anti-diabetes therapies and pancreatic cancer Hospital based, case-control study 973 pancreatic cancer (259 DM) vs. age, sex, race matched 863 controls (109 DM) Proportion of DM in case and control Type of Therapy AII Diabetics Duration of DM>2 yrs DM Cases (%) Control (%) No 73 87 1 AOR (95% CI) Yes 27 13 2.4 (1.8-3.1) Dur<2 14 3 4.5 (3.0-6.9) 3-5 yr 4 3 1.7 (1.0-2.9) >5 yr 9 6 1.5 (1.1-2.2) Insulin Never Ever Insulin secretagogues Never Ever Metformin Never Ever TZDs Never Ever AOR (95% CI) 1.0 5.0 (2.6-9.6) 1.0 2.3 (1.3-4.8) 1.0 0.4 (0.2-0.7) 1.0 1.6 (0.8-3.1) AOR (95% CI) 1.0 5.0 (2.4-10.7) 1.0 1.7 (0.8-3.8) 1.0 0.4 (0.2-0.9) 1.0 1.7 (0.7-3.9) Adjusted for age, sex, race, BMI, smoking, alcohol, family history of cancer, DM duration, insulin use Li D, Gastroenterology 137: 2009
UK GP diabetes study Retrospective cohort study of people treated in UK general practices Diabetic patients >40 yrs of age, and started treatment with oral agents or insulin after 2000 (N=62809) HR (95% CI) P-value SU 1.36 (1.19-1.54) <0.001 Metformin + SU 1.08 (0.96-1.21) 0.21 insulin 1.42 (1.27-1.60) <0.001 Adjusted for age, sex, smoking, previous cancer HR are relative to metformin monotherapy Metformin Metformin+SU No treatment SU Insulin Currie CJ, Diabetologia 52; 2009
Hazard ratios for progression to several cancers Insulin therapy was associated with 20% increased risk for colorectal cancer per each year Although insulin therapy had positive association with the presence of pancreatic cancer, reversecausality may overestimate the actual risk Insulin and SU were associated with the development of colorectal and pancreatic cancers in the large UK retrospective cohort study Currie CJ, Diabetologia 52; 2009
Insulin glargine and cancer: a German study German statutory health insurance fund Diabetic patients without known malignancies and had received first-time therapy exclusively with human insulin, aspart, lispro, glargine N=127031, f/u 1.63 yrs HR (95% CI, reference group: human insulin) Covariates Aspart Lispro Glargine None 0.86 0.85 0.85 (0.79-0.93) Age, sex 0.95 0.90 0.86 (0.79-0.94) Age, sex, dose 1.01 0.99 1.14 (1.05-1.24) Final model 10 IU 1.00 0.99 1.09 (1.00-1.19) 30 IU 1.02 0.98 1.19 (1.10-1.30) 50 IU 1.04 0.98 1.31 (1.20-1.42) Hemkens LG, Diabetologia 52; 2009
Insulin doses and cancer occurrence Within a cohort of insulin-treated T2DM, those with incidence cancer during a longitudinal f/u were identified as controls (112 cases vs. 370 controls) Controls are matched for age, sex, BMI, follow-up time, Median f/u: 75.9 months Comparison of insulin doses in case/control Risk of cancer associated with doses of each insulin type 0.3 IU/kg/day Mannucci E, Diabetes Care 33: 2010 Adjusted for comorbidity, metformin, other types of insulin
Insulin analogues
Insulin analogues and mitogenicity IGF-I receptor affinity is highly correlated with the potency in stimulating growth of human mammary epithelial cells and various cancer cell lines Insulin receptor occupancy time is also correlated with mitogenic potential Analog Insulin receptor affinity (%) Insulin receptor off-rate (%) IGF-I receptor affinity (%) Mitogenic potency (%) Human insulin 100 100 100 100 B10Asp 205 ± 20 14 ± 1 587 ± 50 975 ± 173 Aspart 92 ± 6 81 ± 8 81 ± 9 58 ± 22 Lispro 84 ± 6 100 ± 11 156 ± 16 66 ± 10 Glargine 86 ± 3 152 ± 13 641 ± 51 783 ± 132 A21Gly 78 ± 10 162 ± 11 42 ± 11 34 ± 12 B31B32diArg 120 ± 4 75 ± 8 2,049 ± 202 2,180 ± 390 Detemir 46 ± 5/18 ± 2 204 ± 9 16 ± 1 11 Kurtzhals P, Diabetes 29: 2000
Insulin analogues and carcinogensis B10Asp: increased tumorigenic potential in SD rats Glargine: no increased cancer incidence in SD rats Does difference of two studies B10Asp: mammary tumor at high doses (12.5-200 U/kg) Glargine: no tumor at low doses (2-12.5 U/kg)
Comparison of mitogenic potential in breast ca cells Strong mitogenicity of insulin glargine was observed in several cell lines with higher IGF-I receptor/ insulin receptor ratios (human osteosarcoma cells, breast cancer cells, normal fibroblasts and smooth muscle cells) MCF-7 cells, a human breast cancer cell line, have IGF-IR/IR ratio of 7:1 Shukla A, Endocrine-Related Cancer 16: 2009
Activation of PI3K and MAPK pathways by insulin glargine Unlike regular insulin and other insulin analogs, insulin glargine strongly activates Insulin the IGF-IR glargine and was MAPK associated pathway in malignant with increased mammary risk epithelial for cells, which express high IGF-IR/IR ratio cancer when the dosage was taken into account in clinical studies MAPK pathway Effect of IGF-IR knockdown on cellular proliferation Insulin glargine had 8 fold higher mitogenic potency compared to human insulin, especially in cells with high IGF-IR/IR ratio In in vitro study, insulin glargine had strong mitogenic activity via activation of IGF-IR Shukla A, Endocrine-Related Cancer 16: 2009
Summary Several methodological problems in clinical studies Incomplete evidence of risk evidence of no risk Treatment induced hyperinsulinemia, increased IGF-I bioactivity may accelerate the progression of pre-clinical cancer and be a mechanism of chemo-resistance especially in tissues with high IGF-IR/IR expression Future studies
How can we treat diabetic patients? Risk and benefit Careful consideration of treatment choices Optimal doses of medication to achieve normoglycemia Amelioration of insulin resistance - diet, exercise, metformin
Thank you for your attention!