Additive Interaction Between the Renin-Angiotensin System and Lipid Metabolism for Cancer in Type 2 Diabetes

Details on the methodology of the cohort study have been described previously (8). The Hong Kong Diabetes Registry was established in 1995 and enrolls 30–50 ambulatory diabetic patients per week. Patients were referred by general practitioners and internists from community- and hospital-based clinics or were discharged from the Prince of Wales Hospital or other regional hospitals. Less than 10% of these enrolled patients have had hospital admissions within 6–8 weeks before assessment.

The 4-h assessment of complications and risk factors, modified from the European DiabCare protocol (29), was performed on an outpatient basis. Once a diabetic subject had undergone this comprehensive assessment, he or she was considered to have entered this study cohort and would be observed until death. Ethical approval was obtained from the Clinical Research Ethics Committee of the Chinese University of Hong Kong.

Hospital services in Hong Kong were subsidized to a large extent by the government through the Hospital Authority, the governing body for all publicly funded hospitals and outpatient clinics (30). All patients attending Hospital Authority hospital clinics either as outpatients or inpatients are dispensed medications on site. Clinical end points, including discharge diagnoses of hospital admissions and mortality from 1 January 1995 until 30 July 2005 were used for defining the end points. The Hospital Authority Central Computer System was used to retrieve all hospital admissions and drug-dispensing data. These databases were successfully matched by a unique identification number, the Hong Kong Identity Card number, which is compulsory for all residents in Hong Kong and is used by all government departments and major organizations.

From 1995 to 2005, 7,920 diabetic patients were enrolled in this registry. We limited the analysis to 7,387 patients who were enrolled after 1 December 1996 when dispensing data were computerized. The following exclusion criteria were applied before analysis: diagnosis of type 1 diabetes (n = 323) (31), missing data on types of diabetes (n = 5), non-Chinese or unknown nationality (n = 45), cancer or receiving treatment for cancer at enrollment (n = 175), and missing values for variables used in the analysis (n = 736) (see footnotes to Table 2 for the variable list). LDL cholesterol is a major confounding factor for cancer in type 2 diabetes (8). Because the pretreatment LDL cholesterol levels were not documented in the registry and use of statins and RAS inhibitors may modify risk associations between LDL cholesterol and cancer, we excluded 827 patients who were using statins and 1,116 patients who were using RAS inhibitors at enrollment to reduce confounding effects due to treatment at baseline. A total of 4,160 patients were entered into the present analysis.

All hospital discharge principal diagnoses including cancer and non–cancer-related hospital admissions were regularly coded by a team of trained personnel under the Hospital Authority, according to ICD-9. Mortality data from the Hong Kong Death Registry were retrieved, and the causes of death were verified against hospital admission records in the Hong Kong Hospital Authority computer system. ICD-9 codes were used to identify first admissions relating to a diagnosis of cancer. The end point of this study was defined as having a first cancer event during the follow-up period (code 140–208), including fatal and nonfatal cancer.

On the day of enrollment, all patients attended the Diabetes and Endocrine Centre of the Prince of Wales Hospital after at least 8 h of fasting and without taking any medication. A sterile, random spot urine sample was used to measure the albumin-to-creatinine ratio (ACR). Albuminuria was defined as ACR ≥2.5 mg/mmol in men and ≥3.5 mg/mmol in women. Total cholesterol, triglycerides, and HDL cholesterol were measured by enzymatic methods on a Hitachi 911 automated analyzer (Boehringer Mannheim, Mannheim, Germany) using reagent kits supplied by the manufacturer of the analyzer. LDL cholesterol was calculated by the Friedewald equation (32). The precision performance of these assays was within the manufacturer's specifications.

SAS (release 9.10; SAS Institute, Cary, NC) was used to perform all statistical analysis. Cox proportional hazards regression was used to obtain hazard ratios (HRs) with 95% CI. Follow-up time was calculated as the period from enrollment to the date of first admission for cancer, death, or 30 July 2005, whichever came first.

We tested multiplicative and additive interactions between use of RAS inhibitors and statins for cancer. Multiplicative interaction was tested using a term of the product of two variables in Cox models. There are three measures to test additive interaction (33,34): 1) relative excess risk due to interaction (RERI), 2) attributable proportion due to interaction (AP), and 3) synergy index (S). RERI is the excess risk due to interaction relative to the risk without exposure. AP refers to the attributable proportion of disease that is due to interaction among individuals with both exposures. S is the excess risk from both exposures when there is an additive interaction, relative to the risk from both exposures without interaction. RERI > 0, AP > 0, or S > 1 indicates additive interaction. In Cox models, the RERI is the best choice among the three measures (35). A detailed calculation method of additive interaction including definition of three indicator variables, an SAS program, and a calculator in Excel (available at http://www.epinet.se) was described by Andersson et al. (34). Briefly, the three indicator variables were generated for different combinations of exposure to use of statins and use of RAS inhibitors (1 = Yes/0 = No) (see additive interaction models of Table 2 for details). The SAS program delivered estimates of the required parameters together with the covariance matrix, which are used in calculation of the interaction measures in the Excel calculator.

To control for the confounding effects of drug use, we used Yes/No coding, which was a more robust measure than the duration of drug use with or without adjustment for the period of discontinuation (8). A structured adjustment scheme was used to evaluate the additive interaction of the use of statins and that of RAS inhibitors. First, we adjusted for LDL cholesterol–related risk factors, i.e., LDL cholesterol ≥3.80 mmol/l and LDL cholesterol <2.80 mmol/l plus albuminuria (X.Y., W.Y.S., R.C.W.M., et al., unpublished data), age, sex, BMI, and the use of tobacco and alcohol. Second, we further adjusted for metabolic variables (see footnotes to Table 2 for details) and drug use from enrollment to cancer, death, or censoring dates. To avoid overfitting, a propensity score was used to adjust for the covariates, in which a restricted cubic spline with four knots at the 5th, 35th, 65th, and 95th percentiles was used to adjust for the confounding effects of nonlinear associations of lipids and other continuous covariates as before (8). Stratified Cox models on deciles of the propensity score were used in all of the Cox models to adjust for likelihoods of drug use during follow-up or cancer where appropriate (36) (see Table 2 for details). Proportional hazard assumptions of baseline variables and correlations between pairs of baseline variables were also checked as described previously (8). Two-sided P < 0.05 was considered to be significant.

We developed a UNX rat model characterized by renal carcinogenesis, RAS activation, and dysmetabolism of glucose and lipids to examine disease mechanisms and drug effects. Details of the experimental protocol and phenotypes were described previously (27,28). Male Sprague-Dawley rats (300–350 g) were obtained from the Laboratory Animal Services Centre at the Chinese University of Hong Kong and maintained at our Research Unit at the Prince of Wales Hospital. The animals were caged in pairs, housed at 22–24°C with a 12-h dark/light cycle and free access to water, and fed a standard laboratory rat diet (5001 Rodent Diet; LabDiet, St. Louis, MO). The total duration of the studies was 10 months.

The animals were randomized into three groups: sham operation (n = 8), left UNX (n = 8), and UNX rats treated with the ACE inhibitor (ACEI) lisinopril (n = 8). Lisinopril were dissolved in 3 ml sterile distilled water, with a once-daily dose of 4 mg/kg body weight. All of the sham and UNX rats were also gavaged with distilled water (3 ml) as a placebo control. Ethical approval for the animal study was obtained from the Animal Experimentation Ethics Committee of The Chinese University of Hong Kong and in accordance to the Animals (Control of Experiments) Ordinance of the Department of Health of the Hong Kong SAR Government.

At 3, 6, and 8 months after the operation, 24-h urine samples were collected using metabolic cages (Huang Qiao Yin Xing Animal Cage & Equipments Factory, Suzhou, China). When rats were killed 10 months after the operation, fasting blood samples were taken for the measurement of renal function and lipids including total cholesterol, triglyceride, LDL cholesterol, and HDL cholesterol.

Histopathological criteria for diagnosis of renal cell carcinoma in the UNX rats included cytological atypia, bizarre nuclei, frequent mitotic figures, and invasive growth. Absence of these morphological characteristics in at least three tissue sections indicated absence of renal cancer. Rats were killed at 10 months after the operation. Kidneys from all of the rats were removed, weighed, and processed for light microscopy. Tissue samples were fixed in 10% neutral formaldehyde and embedded in paraffin. Serial longitudinal sections (4 μm) were spliced parallel to the longest axis of the kidney and stained with periodic acid Schiff. Stained slides were examined with a Zeiss Axioplan 2 imaging microscope (Carl Zeiss, Hamburg, Germany), and representative images were captured using a Spot digital camera.

Tissue total proteins from renal cortex were extracted. The resolved proteins were then transferred onto nitrocellulose membranes. The membranes were blocked for 1 h at room temperature with 5% skimmed milk, incubated with primary antibodies against IGFBP3 (dilution 1:1,000; Santa Cruz Biotechnology, Santa Cruz, CA), Akt1/2/3 (dilution 1:1,000; Cell Signaling Technology, Danvers, MA), and PKCζ (dilution 1:1,000; Santa Cruz Biotechnology) in Tris-buffered saline containing 0.05% Tween 20 (TBS-T) with 5% skimmed milk overnight at 4°C. After washing with TBS-T, membranes were incubated with anti-goat, anti-rabbit, or anti-mouse secondary antibody conjugated to horseradish peroxidase (Upstate Biotechnology, Billerica, MA) with dilution of 1:2,000. Proteins were detected by enhanced chemiluminescence (Amersham, Piscataway, NJ) on Hyperfilm. The major protein bands with ∼42 kDa for IGFBP3, 60 kDa for Akt1/2/3, and 80 kDa for PKCζ were detected. To ensure equal loading of proteins, membranes were incubated and probed with a rabbit anti–β-actin antibody (Abcam, Cambridge, MA) with a dilution of 1:10,000, which recognizes β-actin at ∼43 kDa. Signals were quantitated by densitometry and corrected for the β-actin signal, using a Kodak Digital Image station 440CF and the ID Image Analysis software program. Treatment groups were compared using ANOVA, and P < 0.05 was considered statistically significant.

The median age of the cohort was 55 years (25th–75th percentiles 45–66 years), with a median duration of diabetes of 5 years (1–10 years). During a total of 21,992 person-years of follow-up, 190 patients developed cancer, giving an annual incidence of 8.64 (95% CI 7.42–9.86) per 1,000 person-years. Patients who subsequently developed cancer were older, were more likely to be smokers and alcohol drinkers, and had a longer duration of diabetes and poorer metabolic profile than those without cancer (Table 1). They were also more likely to use antihypertensive drugs other than RAS inhibitors than patients without cancer. Patients who developed cancer were less likely to be treated with statins alone or statins combined with RAS inhibitors during the follow-up period than those who remained free of cancer.

Compared with nonusers of statins and RAS inhibitors, subjects who were exposed to statins and/or RAS inhibitors had a lower risk of cancer after adjustment for drug use indications and demographic and lifestyle covariates. The additive interaction between statins and RAS inhibitors on cancer risk was significant as indicated by the RERI (0.39 [95% CI 0.09–0.69]) and AP values (1.57 [0.21–2.94]). The statistical significances of use of statins only, use of RAS inhibitors only, combined use of statins and RAS inhibitors, and additive interaction measures (RERI and AP) persisted after adjustment for all of the above factors, metabolic covariates, and use of other drugs at enrollment and during follow-up using a propensity score as well as taking nonlinear associations into account (Tables 2 and 3). Combined use of statins and RAS inhibitors was consistently associated with lower risks of cancer even after removal of 682 patients followed for <2.5 years (HR 0.35 [95% CI 0.16–0.75]) and reinclusion of 1,943 patients who used RAS inhibitors or statins at baseline (0.33 [0.21–0.52]). In addition, combined use of RAS inhibitors and statins was also consistently associated with lower risks of a variety of site-specific cancers (supplemental Table, available in an online appendix at http://care.diabetesjournals.org/cgi/content/full/db09-0105/DC1). Figure 1 shows that the cumulative incidence of cancer in patients who used statins or RAS inhibitors alone or both was lower than that in those patients who were not exposed to statins and RAS inhibitors over the follow-up period.

We then explored the role of RAS activation in the development of cancers using the UNX rat model. By 10 months postoperation (Fig. 2), all of the eight untreated UNX rats (100%) developed invasive renal cell carcinoma in the remnant kidney (Fig. 2,B). In comparison, none of the sham rats (Fig. 2,A) or UNX rats treated with the ACEI lisinopril (Fig. 2 C) developed renal cancer. As previously reported (27,28), untreated UNX rats also exhibited a phenotype resembling that of type 2 diabetes, characterized by insulin resistance and pancreatic β-cell deficit, whereas treatment with lisinopril significantly reduced hypertrophy of the remnant kidney (28).

We used the UNX model to examine the longitudinal effect of RAS inhibition on lipid metabolism and renal function. From 3 months onward, untreated UNX rats exhibited progressive chronic renal dysfunction, as indicated by an increased total urine protein-to-creatinine ratio (Fig. 3,A). An elevated LDL cholesterol level was observed in the untreated UNX rats from 6 months after UNX (Fig. 3,B). Interestingly, treatment with ACEI largely attenuated the renal dysfunction (Fig. 3,A) and improved lipid metabolism (Fig. 3,B) in UNX animals (all P < 0.05 vs. untreated UNX rats). Western blot assays of renal tissues revealed a fourfold increase in the protein expression of HMGCR in the untreated UNX rats compared with the sham animals (Fig. 4). Treatment with the ACEI lisinopril normalized the expression of HMGCR (Fig. 4).

We further examined the protein expression of key molecules in the IGF-1 signaling pathway, which may be linked to carcinogenesis. Compared with sham animals, the protein expression of the growth inhibitory factor, IGFBP3, in the remnant kidney was reduced, whereas the protein expression levels of growth-promoting factors, such as Akt/PKB and PKCζ, were increased in the untreated UNX rats (Fig. 5). RAS inhibition by the ACEI nearly normalized the expression of these IGF-1 signal molecules (Fig. 5).

This study provides evidence that the combined use of RAS inhibitors and statins may be associated with greater anticancer effects than use of either class of drugs in isolation. To explain this phenomenon, our animal data indicated that RAS blockade prevented the development of renal cell carcinoma in UNX rats via normalizing the expression of the HMGCR and IGF-1 signaling pathways.

The risk associations of statin use and cancer remain controversial. In epidemiological studies, the use of statins was associated with a large relative risk reduction for cancer (37). Conversely, nearly all meta-analyses of clinical trials (13,38) showed that statins have a neutral effect on incidence of cancer. However, results from meta-analyses are often inconclusive because of heterogeneity of study design, clinical profile of patient cohorts, different definitions for outcome measures, and quality of data. This statement is illustrated by the marked variations in cancer incidence ranging from 0.2 to 6.3% in these reported trials (39). Furthermore, the majority of clinical trials included in these meta-analyses were not conducted in diabetic populations.

The risk association of RAS inhibitors and cancer is also controversial. A meta-analysis of randomized trials of antihypertensive drugs failed to demonstrate a reduced odds of cancer with use of antihypertensive drugs, including RAS inhibitors (40). On the other hand, the Rotterdam Study reported that use of RAS inhibitors was associated with reduced cancer risk in ACE-DD genotype carriers, who are also known to have high levels of ACE (41). Two other studies (11,16) also showed that users of RAS inhibitors had lower risks of cancer than nonusers. Our study further showed that combined use of stains and RAS inhibitors was associated with a larger reduction in cancer risk compared with the added risk reduction associated with the use of either of the two types of drugs in isolation.

In support of these clinical observations, experimental studies indicated that RAS activation can influence carcinogenesis and tumor growth by inducing oxidative stress (42) and modulating angiogenesis, cell proliferation, immune responses, and extracellular matrix formation (12). In our experimental studies, the UNX rats developed glucose intolerance and abnormal lipid metabolism and eventually renal cancer, which were all prevented by treatment with ACEIs. This anticancer effect appears to be at least partially mediated through modulation of the HMGCR and IGF-1 signaling pathways, with the latter having complex effects on intermediary metabolism and cellular growth. The tissue activity of IGF-1 is regulated by the levels of its binding proteins as well as by the number and responsiveness of its receptors. In this regard, low levels of IGFBP3 are associated with increased risk of cancer in clinical studies. On the other hand, downstream signals of the IGF-1 pathway such as Akt/PKB and PKCζ can stimulate cell proliferation and promote cell mitosis (43). Thus, given the cancer-enhancing effects of the RAS components (12), interactions between IGF-1 signaling molecules and the tissue RAS components have been shown to stimulate the bcl-2 proto-oncogene–associated cell proliferation and to inhibit the p53 anti-oncogene–mediated cell death (44).

Cholesterol is essential for cell division and growth. In the context, IGF-1 has been shown to regulate the induction and expression of a family of genes involved in cholesterol biosynthesis (45). Thus, it is plausible that overactivation of components of the RAS, IGF-1, and HMGCR pathways may result in dysregulated growth and eventually carcinogenesis, as evidenced by 1) reduced expression of IGFBP3, 2) activation of the IGF-1 signaling pathway (Akt/PKB and PKCζ), and 3) increased HMGCR expression in our UMX rat model. Of note, the interaction between the combined use of statins and RAS inhibitors was observed for multiple cancers in humans, including cancer of genitourinary organs, whereas the mechanistic exploration was made for kidney cancer in animals. In this regard, a strong association between diabetes and kidney cancer has been reported in a large cohort (46). If multiple cancers in type 2 diabetes share some common “pathogenesis,” the IGF-1 signaling pathway is likely to play a role in development of cancer.

Our study has certain limitations. First, the study is not a clinical trial and the findings are only hypothesis generating. Second, we did not perform regular screening for cancer in this cohort because of finite resources. The use of principal discharge diagnosis to identify patients with cancer may lead to the omission of a small number of cancer events. Third, the current method of testing additive interaction does not allow us to quantify the interaction using doses of statins and RAS inhibitors. Fourth, the cohort was mainly clinic based, although the overall clinical profile of patients was comparable to that of many community-based cohorts (47). Last, the dysmetabolism observed in the animal model may not be applicable to humans, although the phenotypes exhibited by the UNX model were highly commensurate with those of type 2 diabetes.

In conclusion, we observed a synergistic effect of the combined use of RAS inhibitors and statins on reducing cancer risk in type 2 diabetes, suggesting that cross-talk of RAS and lipid metabolism may play an important role in the elevated risk of cancer in type 2 diabetes. Diabetes predisposes patients to increased risks of abnormal lipid metabolism, and elevated RAS activity is more frequent in the type 2 diabetic population than in the general population. Presumably, type 2 diabetic individuals with high LDL cholesterol or hypertension would be at risk of cancer via similar mechanisms. Thus, our findings are especially important and relevant to type 2 diabetes. Conversely, given that cross-talk between the RAS and lipid metabolism exists in the general population for the development of atherosclerosis, whether the findings of the present study would apply to the general population warrants further investigations.

This study was partially supported by the Hong Kong Jockey Club Charities Trust [JCICM-P2–05-(CUHK)], Hong Kong Foundation for Research and Development in Diabetes, and Lioa Wun Yuk Diabetes Memorial Research Fund established under the auspices of The Chinese University of Hong Kong.

No potential conflicts of interest relevant to this article were reported.

We thank all the medical and nursing staff of the Diabetes and Endocrine Centre of the Prince of Wales Hospital for recruiting and managing the care of these patients.