Diabetes and cancer are two heterogeneous, multifactorial, severe, and chronic diseases. Because of their frequency, reciprocal influences – even minor influences – may have a major impact. Epidemiological studies clearly indicate that the risk of several types of cancer (including pancreas, liver, breast, colorectal, urinary tract, and female reproductive organs) is increased in diabetic patients. Mortality is also moderately increased. Several confounding factors, having general or site-specific relevance, make it difficult to accurately assess cancer risk in diabetic patients. These factors include diabetes duration, varying levels of metabolic control, different drugs used for therapy, and the possible presence of chronic complications. Hyperinsulinemia most likely favors cancer in diabetic patients as insulin is a growth factor with pre-eminent metabolic but also mitogenic effects, and its action in malignant cells is favored by mechanisms acting at both the receptor and post-receptor level. Obesity, hyperglycemia, and increased oxidative stress may also contribute to increased cancer risk in diabetes. While anti-diabetic drugs have a minor influence on cancer risk (except perhaps the biguanide metformin that apparently reduces the risk), drugs used to treat cancer may either cause diabetes or worsen a pre-existing diabetes. In addition to the well-known diabetogenic effect of glucocorticoids and anti-androgens, an increasing number of targeted anti-cancer molecules may interfere with glucose metabolism acting at different levels on the signaling substrates shared by IGF-I and insulin receptors. In conclusion, diabetes and cancer have a complex relationship that requires more clinical attention and better-designed studies.
Diabetes mellitus (DM) is a serious and growing health problem worldwide and is associated with severe acute and chronic complications that negatively influence both the quality of life and survival of affected individuals. Today, 250 million people live with diabetes globally, with this figure expected to reach 380 million within 20 years. Therefore, if diabetes is associated even with a small increase in the risk of cancer, this may have important consequences at the population level.
The association between cancer and diabetes has been investigated extensively and most, but not all studies, found that DM is associated with an increased risk of several types of cancer. Most published data, however, requires reinterpretation because DM is not a single disease, but rather a group of metabolic disorders characterized by hyperglycemia. Within this general context, each type of diabetes has additional metabolic and hormonal abnormalities that differently affect diabetic patients. It is therefore inappropriate to consider diabetic patients as a homogeneous cohort. In addition, a series of potential confounders directly related to the disease (obesity, quality of metabolic control, drugs employed for treatment, diet, etc.) and present in diabetic patients may influence the association between diabetes and cancer.
In the present review, we will discuss the available evidence concerning the association between diabetes and cancer, the different aspects of diabetes which may influence this association, and the possible mechanisms involved.
Cancer risk is increased in diabetic patients
A series of recent studies and meta-analyses confirm that the risk for several solid and hematologic malignancies (including liver, pancreas, colorectal, kidney, bladder, endometrial and breast cancers, and non-Hodgkin's lymphoma) is elevated in diabetic patients (Table 1). Evidence for the association of diabetes with other cancers is not available, while for prostate cancer, a reduced incidence has been reported in diabetic patients (Table 1). If we accept that cancer is more frequent in DM, the positive association between diabetes and cancer risk might actually be somewhat underestimated. Diabetes, in fact, is an underdiagnosed disease (3–5% of the adult population has undiagnosed diabetes; Harris et al. 1998) and thus the control population very likely includes individuals with diabetes, which will increase the cancer risk in the ‘normal’ population.
Meta-analyses on the relative risk (RR) of cancer in different organs of diabetic patients
|Cancer||RR (95% CI)|
|Liver (El-Serag et al. 2006)||13 case–control studies||2.50 (1.8–3.5)|
|7 cohort studies||2.51 (1.9–3.2)|
|Pancreas (Huxley et al. 2005)||17 case–control studies||1.94 (1.53–2.46)|
|19 cohort studies||1.73 (1.59–1.88)|
|Kidneya (Lindblad et al. 1999, Washio et al. 2007)||1 cohort study||1.50 (1.30–1.70)|
|1 cohort study||2.22 (1.04–4.70)|
|Endometrium (Friberg et al. 2007)||13 case–control studies||2.22 (1.80–2.74)|
|3 cohort studies||1.62 (1.21–2.16)|
|Colon–rectum (Larsson et al. 2005)||6 case–control studies||1.36 (1.23–1.50)|
|9 cohort studies||1.29 (1.16–1.43)|
|Bladder (Larsson et al. 2006)||7 case–control studies||1.37 (1.04–1.80)|
|3 cohort studies||1.43 (1.18–1.74)|
|Non-Hodgkin's lymphoma (Mitri et al. 2008)||5 cohort studies||1.41 (1.07–1.88)|
|11 case–control studies||1.12 (0.95–1.31)|
|Breast (Larsson et al. 2007)||5 case–control studies||1.18 (1.05–1.32)|
|15 cohort studies||1.20 (1.11–1.30)|
|Prostate (Kasper & Giovannucci 2006)||9 case–control studies||0.89 (0.72–1.11)|
|10 cohort studies||0.81 (0.71–0.92)|
In diabetic patients, cancer may be favored by: i) general mechanisms that promote cancer initiation or progression in any organ because they are due to alterations (i.e. hyperglycemia or hyperinsulinemia or drugs) that affect all tissues; and ii) site-specific mechanisms affecting cancerogenesis of a particular organ.
The incidence of liver and pancreatic cancer is increased in diabetes
Several meta-analyses indicate that the strongest association between DM and increased cancer risk is with pancreatic and liver cancer (Table 1), i.e. two key organs involved in the metabolic derangements typical of diabetes.
Because of the portal circulation, liver cells are exposed to higher insulin concentrations than other tissues, a condition that is exacerbated in insulin-resistant hyperinsulinemic type 2 diabetic individuals, but that is not present in insulin-deficient type 1 diabetic patients treated with exogenous insulin (see Fig. 1). It is unlikely, therefore, that insulin's mitogenic action is specifically involved in the higher incidence of liver cancer in diabetic patients since healthy liver cells are physiologically exposed to higher insulin concentrations than other tissues. Moreover, in diabetic patients injected with exogenous insulin, the liver is exposed to the same insulin levels as the other organs.
Since most epidemiologic studies indicate a two- to threefold increase in hepatocellular carcinomas (HCC) in diabetic patients, other conditions, specific to the liver, must favor liver cancerogenesis in diabetic patients. It has been questioned whether diabetes is a direct risk factor for liver cancer or whether diabetes-related diseases of the liver are also involved. Indeed, steatosis and cirrhosis, both well-known risk factors for HCC, are more frequent in diabetic patients. Likewise, the nonalcoholic fatty liver disease (NAFLD) is very common in both diabetes and obesity and even more frequent in obese-diabetic patients, occurring in over 80% of type 2 diabetic patients. Additional factors that may favor HCC in DM include hepatitis B and C virus (HBV and HCV) infections, both more frequent in diabetic subjects as compared with the nondiabetic population (Davila et al. 2005, Chen et al. 2006).
In conclusion, increased liver cancer incidence in diabetes is well documented and, although the exact mechanisms underlying this association are sill unclear, liver inflammation, hepatocyte damage, and repair are likely to be involved in the higher frequency of HCC among diabetic patients.
Most earlier studies investigating the association between diabetes and pancreatic cancer are probably misleading because they do not distinguish between pre-existing diabetes (a condition possibly favoring exocrine pancreatic cancer) and new-onset diabetes (a possible sign of pancreatic functional damage due to a still undiagnosed pancreatic cancer; Noy & Bilezikian 1994). The latter situation is so frequent that hyperglycemia and diabetes, when appearing after the age of 45–50 years, in a lean subject with no family history for diabetes, is considered sufficient to pose an indication for pancreatic cancer screening (Noy & Bilezikian 1994, Chari et al. 2008, Pannala et al. 2009). Similarly, elderly subjects with new-onset diabetes have a 3-year risk of pancreatic cancer nearly eight times higher than a nondiabetic person of similar age and sex (Chari et al. 2005). Laboratory and clinical evidence suggest that diabetes caused by pancreatic cancer is due to cytokines produced by the tumor (Basso et al. 2002) rather than secondary to endocrine pancreatic tissue invasion and damage (Pannala et al. 2009). This conclusion is also supported by the observation that hyperglycemia occurs at an early stage of pancreatic cancer and is independent of tumor size and stage (Chari et al. 2008, Pannala et al. 2008). Epidemiological studies in subjects affected by DM at least 1 year prior to diagnosis or death from pancreatic cancer indicated a relative risk (RR) of 2.1 (95% confidence interval (CI)=1.6–2.8). When the same analysis was carried out including only patients with 5 years of pre-diagnosed diabetes, their RR for pancreatic cancer was similar (RR=2.0; Everhart & Wright 1995). Since all of these data exclude diabetes induced by pancreatic tumors, the reported findings support the possibility that diabetes is indeed a risk factor for pancreatic cancer.
The ‘pre-diabetes’ state should also be considered a risk factor for pancreatic cancer. Studies that evaluated the association between post-load glucose levels and pancreatic tumors in 35 658 individuals reported a higher RR with increasing glucose tolerance impairment. After adjusting for age, race, cigarette smoking, and body mass index (BMI), the risk progressively increased from normal subjects to subjects with slightly altered glycemia (RR=1.65) and then to diabetes (RR=2.15; Gapstur et al. 2000). These results did not change when patients who died of pancreatic cancer during the first 5 years after the assessment of post-load glucose levels were excluded, further suggesting that hyperglycemia and diabetes per se are predisposing factors for pancreatic cancer.
The biological mechanisms underlying the association between diabetes and pancreatic cancer are unclear. Hyperinsulinemia has been indicated as a possible factor because exocrine pancreatic cells, which give rise to most pancreatic cancers, are exposed to very high insulin concentrations because of the common blood supply with the adjacent insulin-secreting islets (Williams & Goldfine 1985). Elevated insulin could act as a tumor growth-promoting factor in many different ways (covered later). This mechanism, however, does not justify the excess of pancreatic cancer in insulin-treated diabetic patients (Green & Jensen 1985) or in type 1 diabetes (Stevens et al. 2007) where pancreatic cells are not exposed to insulin levels higher than those of other tissues. In these studies, however, the analysis is hampered by the insufficient number of cases accrued, due to both the type of diabetes (type 1 diabetes accounts for <10% of all DM patients) and patient age (pancreatic cancer is rare before age 40).
Increased incidence of other cancers in diabetes
An increased frequency of malignancies of other organs has been reported in diabetic patients and has been ascribed to a variety of general and local mechanisms. In these cases, studies are not as numerous as for liver and pancreatic tumors, and the increases in RR are not as statistically significant. However, in many instances, the increased risk is clinically relevant, especially considering the prevalence of the two diseases in the general population.
In diabetic patients, the increased incidence and increased mortality for kidney cancer have been attributed to both general mechanisms (hyperinsulinemia and obesity) and specific factors, mainly hypertension (Yuan et al. 1998, Chow et al. 2000, Zucchetto et al. 2007) and the frequent kidney diseases occurring in diabetic patients (Lindblad & Adami 2002).
Individuals with DM also display a modest increase in the risk of bladder cancer. In this case, in addition to general factors like hyperinsulinemia, the increased frequency of urinary tract infections is also likely to be involved.
The risk of cancers of the female reproductive organs is also increased in DM. Both breast and endometrial cancer risks are increased in diabetic women, and this risk is independent from obesity (a well-established factor promoting breast cancer) as it persists even after correcting epidemiological data for this disease.
Several biological mechanisms may be involved, mostly regarding sex hormone abnormalities. Hyperinsulinemia may increase the levels of bioactive estrogens by decreasing the concentration of circulating sex hormone-binding globulin (Kaaks 1996) and might also stimulate androgen synthesis in the ovarian stroma (Kaaks 1996). Other possible mechanisms include delayed menarche, especially in type 1 diabetic women, who also have a higher incidence of nulliparity, irregular menses, and fertility disorders.
Type 2 diabetes has been associated with an increased risk of colorectal adenomas and carcinomas in most, but not all, studies (Elwing et al. 2006, Limburg et al. 2006). The risk is increased in both women and men for both colon and rectal cancer (Larsson et al. 2005). In addition to hyperinsulinemia, hypothesized mechanisms include slower bowel transit time and the elevated fecal bile acid concentrations often observed in DM (Stadler et al. 1988, Will et al. 1998).
Large prospective cohort studies and case–control studies have shown a moderate increase of non-Hodgkin's lymphoma in diabetic patients, a possible consequence of the immune dysfunction related to impaired neutrophil activity and abnormalities in cellular and humoral immunity in diabetes (Mitri et al. 2008).
Decreased incidence of prostate cancer in diabetes
In contrast to the increased risk of numerous forms of neoplasia, most studies report a reduced risk of prostate cancer in men with diabetes. A recent meta-analysis (Kasper & Giovannucci 2006) including 14 studies carried out in the pre-PSA era (i.e. before the generalised use of prostate specific antigen screening for prostate cancer; Bonovas et al. 2004) and 5 additional studies carried out in the PSA era (and therefore, concerning cancers diagnosed earlier and smaller cancers) has found a significantly reduced risk in diabetic patients (Table 1). The 16% average decreased risk of developing prostate cancer must most likely be attributed to the decreased testosterone levels in diabetic patients (Barrett-Connor 1992, Betancourt-Albrecht & Cunningham 2003). However, other metabolic and hormonal factors, including altered insulin and leptin concentrations, the diffuse use of medications such as statins and metformin, and changes in diet and lifestyle in order to control diabetes, have also been hypothesized as elements potentially contributing to the inverse association between diabetes and prostate cancer (Kasper & Giovannucci 2006).
In conclusion, the epidemiological studies cited above may be partially biased by relevant heterogeneity due to different study design (inclusion criteria), incomplete characterization of DM, failure to consider potential confounders (obesity, diabetes duration, and treatment), and also variably defined control population. However, the overall increased risk for the development of several types of cancer in diabetic patients must be considered well documented. In diabetes, there is a mild to moderate increase in the incidence of pancreas, liver, breast, colorectal, urinary tract, and female reproductive organ cancer and a mild reduction in prostate cancer risk.
Cancer mortality is increased in diabetic patients
Data on cancer mortality in diabetic patients are less abundant and less homogeneous than data on cancer incidence.
A positive association between breast cancer mortality and diabetes was found in three out of five studies, with a RR from the pooled data of the five studies of 1.24 (95% CI=0.95–1.62; Larsson et al. 2007). In the largest study (cohort size 588 321 with 4346 deaths from breast cancer), after adjusting for age, race, BMI, physical activity, smoking, and alcohol, RR in diabetic women was 1.27 (1.11–1.45) when compared with the nondiabetic female population. In this cohort, as in most others, no stratification was performed for type of diabetes and different treatments. In addition, the menopausal status was not recorded (Coughlin et al. 2004). In a recent study aimed at evaluating whether diabetes could affect breast cancer prognosis, after a 5-year mean follow-up, mortality for breast cancer was significantly higher in women with diabetes (hazard ratio 1.39; 95% CI=1.22–1.59, P<0.0001) suggesting that early survival following breast cancer was reduced in women with diabetes (Lipscombe et al. 2008). This reduced survival might be a consequence of more aggressive breast cancer but also of diabetes-related comorbidities. In fact, in that study, the cause of death was not recorded and diabetic women without breast cancer had an increase in mortality similar to that of diabetic women with breast cancer, suggesting that diabetes, rather than breast cancer, was the major factor contributing to the raised mortality.
Diabetes was also positively associated with colorectal cancer mortality. A statistically significant association was found in three out of six studies (Larsson et al. 2005), and a nonsignificant positive association was reported in a fourth one. Pooled data from the six studies indicated a positive association between diabetes and colorectal cancer mortality (RR=1.26; 95% CI=1.05–1.50), but heterogeneity issues partially compromise the significance of the results. Within these six articles, the two cohort studies that evaluated standardized mortality ratio both indicated a positive association between DM and colorectal cancer death. However, only one study reported a statistically significant increased mortality from colorectal cancer in diabetic patients. A study aimed at evaluating the influence of diabetes on long-term outcome of patients resected for colon cancer (3759 patients, 287 with DM) found that diabetes negatively affected survival in colon cancer patients (Meyerhardt et al. 2003). Data were adjusted for predictors of colon cancer outcome (age, gender, race, clinical status, TNM (tumor, node, metastasis classification) category, Dukes stage, location of primary tumor, and grade of differentiation), and indicated that both disease-free survival (DFS) and overall survival (OS) at 5 years were significantly reduced in diabetic patients (DFS=48 vs 59% in nondiabetics, P<0.0001; OS=57 vs 66% in nondiabetics, P<0.0001). Median survival in diabetic patients was 6.0 years vs 11.3 in nondiabetic subjects. In this study, the role of DM comorbidities (that may negatively affect overall mortality among cancer patients because of adverse health conditions) was probably minor since cancer recurrence was also higher in diabetic patients (recurrence-free survival 56 vs 64% in nondiabetics, P<0.012).
A positive association was also found between diabetes and endometrial cancer mortality in two studies, but it was significant only in one of them (RR=2.38; 95% CI=1.05–5.37; Coughlin et al. 2004, Folsom et al. 2004).
It is interesting to note that, although diabetic patients have a reduced risk of prostate cancer, once an insulin-resistant, overweight man has been diagnosed with prostate cancer, his likelihood of dying from the disease is increased (Ma et al. 2008).
A recent study on the systematic assessment of long-term, all-cause mortality in cancer patients with or without diabetes has evaluated, at 1.41 (95% CI=1.28–1.55), the hazard ratio for death in cancer patients with diabetes compared to cancer patients without diabetes (Barone et al. 2008). Mortality was significantly increased for cancers of the breast, endometrium, colon, and rectum. In this study, the increase in mortality risk was not significantly increased for lung, gastric, liver, pancreatic or prostate cancers. Overall, however, the heterogeneity of the studies analyzed and the length of the observation period (1969–2008, during which treatment for both cancer and diabetes changed markedly) hamper, at least in part, the significance of the data.
Several possible explanations can be put forth to explain the increased risk of cancer death in DM. For instance, it is still unclear whether diabetes, through a number of mechanisms, makes the cancer more aggressive or whether the host organism is less resistant to cancer progression. It is also possible that diabetic patients receive different cancer treatment (i.e. oncologists may employ lower chemotherapy doses in diabetic patients, concerned about their general health and their heart, liver, and kidney function). Of course, it is also possible that diabetic patients may have a worse response to chemotherapy compared with nondiabetic individuals.
In conclusion, epidemiologic studies provide evidence that cancer mortality is moderately increased in diabetic patients. Whether this is a consequence of hyperglycemia and hyperinsulinemia (growth-promoting effect on cancer cells), the impaired health conditions due to diabetes' comorbidities or a combination of the two is still unclear.
Type 1 and type 2 diabetes and cancer risk
DM is a group of metabolic disorders characterized by hyperglycemia. The two most frequent subtypes of DM differ in both metabolic and hormonal characteristics: in type 1 diabetic patients (5–10% of all diabetics), hyperglycemia is associated with an absolute deficiency of endogenous insulin secretion and the absolute requirement for exogenous insulin administration.
In type 2 diabetes, hyperglycemia and hyperinsulinemia coexist for a long time because of insulin resistance in peripheral tissues. Only when β-cell function fails completely will the patient require insulin treatment because of endogenous insulin deficiency.
In spite of these considerable pathogenetic and clinical differences, many studies on the association between diabetes and cancer were carried out without an appropriate distinction between the two forms of diabetes.
For obvious epidemiological reasons, most studies on the association between cancer and diabetes have been carried out in patients with type 2 diabetes (90% of all diabetic patients). As these patients, unlike those with type 1 diabetes, have endogenous hyperinsulinemia and insulin resistance, it is questionable whether these data can be automatically extended to type 1 diabetic patients. This concern is particularly relevant for the older reports in which diabetes assessment was based on self-reported hyperglycemia, with no criteria aimed at distinguishing type 1 from type 2 diabetes. Although more recent studies have been based on medical records, the distinction between type 1 and type 2 diabetes was mostly based on surrogate indicators of diabetes type, like young patient age or insulin treatment (assumed as type 1) versus insulin-independent diabetes (assumed as type 2). This distinction does not take into account many specific conditions, including type 2 diabetic patients that are treated with insulin because oral hypoglycemic agents (OHA) are no longer effective (secondary failure to OHA), type 1 diabetes of the adult (∼5% of adult subjects previously classified as type 2 diabetes; Buzzetti et al. 2007), and other less frequent conditions.
Because of the 10:1 ratio between type 2 and type 1 diabetes, and considering that cancer is mainly a disease of the older population (where type 1 diabetes is less frequent), it is reasonable to assume that the large majority of tumors observed in diabetic patients occurred in type 2 diabetics.
Thus, if cancer association with type 1 diabetes has specific characteristics, these have likely been obscured by the large majority of cancers diagnosed in type 2 diabetic patients.
Even the few studies specifically addressing cancer incidence in type 1 diabetic patients suffer from poor diabetes type assessment. For example, a recent cohort study evaluating cancer incidence in nearly 30 000 Swedish type 1 diabetic patients diagnosed in the period 1965–1999 has identified 355 cases of cancer (standardized incidence ratio (SIR)=1.2; 95% CI=1.0–1.3, compared with the general Sweden population; Zendehdel et al. 2003). In contrast to type 2 diabetic patients, no increased risk of breast, pancreatic, colorectal, or kidney cancer was found in this cohort. However, type 1 diabetic patients had an increased RR for stomach (SIR=2.3; 95% CI=1.1–4.1), endometrial (SIR=2.7; 95% CI=1.4–4.7), and cervical cancer (1.6; 1.1–2.2). These positive associations have been attributed to the high prevalence of Helicobacter pylori infection or of pernicious anemia (for gastric carcinomas; Oldenburg et al. 1996, De Block et al. 1999) and to the higher incidence of nulliparity, irregular menses, and fertility disorders in type 1 diabetic women (for uterine malignancies). In contrast with this report, a recent meta-analysis including three cohort studies and six case–control studies has found that the RR for pancreatic cancer was doubled in type 1 diabetic patients and young-onset diabetics in comparison with nondiabetics (Stevens et al. 2007).
In conclusion, the large majority of the epidemiological data on cancer incidence and mortality has been obtained in type 2 diabetic patients. Because of the different biology between the two subtypes of diabetes, these findings cannot be acritically extended to type 1 diabetic subjects.
The role of hyperinsulinemia in favoring cancer incidence and progression in diabetic patients
A role for insulin in promoting cancer growth was first recognized by studies in experimental animals. Rats and mice made diabetic with streptozotocin or alloxan (therefore hyperglycemic and insulin deficient) developed less aggressive tumors as they display a longer latency period for cancer development, lower number of tumors, slower cancer progression, and smaller final tumor volume with respect to control animals (Heuson & Legros 1972; Fig. 1). Insulin treatment reversed these effects (Heuson et al. 1972). These results are in concert with the well-known mitogenic effect of insulin that has been extensively documented both in vitro and in vivo.
Most type 1 and type 2 diabetic patients are exposed for decades to increased insulin concentrations, although the physiologic and therapeutic conditions are very different in each individual with diabetes.
Type 1 diabetic patients have an absolute requirement for exogenous insulin because of autoimmune destruction of their pancreatic β-cells, which are therefore unable to produce endogenous insulin. In these patients, insulin administration cannot mimic the physiologic insulin secretion, not only in terms of temporal pattern and hormone serum levels but also in terms of compartment distribution. Indeed, pancreas-secreted insulin is first distributed to the liver (first passage insulin) where a relevant aliquot (up to 80%; Ferrannini & Cobelli 1987) is retained and degraded. The remaining hormone reaches the peripheral tissues through the systemic circulation. The liver/peripheral tissue insulin concentration ratio, therefore, ranges from 3:1 up to 9:1 during insulin secretion bursts. Exogenously administered insulin, in contrast, will arrive to peripheral tissues and to the liver at the same time and at a similar concentration. Peripheral tissue hyperinsulinemia due to exogenous insulin (circulating levels may peak two- to fivefold higher than normal endogenous levels, depending on the dose injected and the type of insulin or analog used) and the ensuing relative liver hypoinsulinemia, therefore, are a common condition in type 1 diabetic patients (Fig. 2).
On the contrary, in most type 2 diabetic patients, hyperglycemia is associated with endogenous hyperinsulinemia, a compensatory state caused by insulin resistance. This condition often persists for many years (decades when including the pre-diabetes period before clinically evident diabetes is diagnosed). Hence, in these patients, the liver/peripheral tissue insulin concentration ratio reflects that of nondiabetic patients, but at a higher level. However, in contrast to normal individuals, in these diabetic patients, increased insulin secretion fails to replete body fuel storages in response to feeding because of insulin resistance. Therefore, in these patients, excess unused substrates (i.e. glucose) are present concomitantly with hyperinsulinemia. This abnormal situation is accompanied by a series of other abnormalities involving other hormones like glucagon, incretins, leptin, etc.
As DM persists for many years, this scenario is often subject to changes, with most type 2 diabetic patients progressively presenting decreased insulin secretion following the failure of β-cells, due to increased apoptosis rates that are not balanced by neogenesis. At this stage, patients with type 2 diabetes may become similar to type 1 diabetic individuals, with endogenous hypoinsulinemia and exogenous insulin requirement.
When studying type 2 diabetic patients, therefore, diabetes duration and insulin requirement may affect tissue exposure to insulin in different ways. If hyperinsulinemia has a role in promoting cancer initiation and/or progression, these aspects should be considered when determining the individual risk of a diabetic patient to develop cancer. Most studies on the diabetes–cancer association overlooked these different biological conditions.
In conclusion, diabetes is generally characterized by hyperglycemia and hyperinsulinemia, often coupled with a reduced metabolic effect of insulin (insulin resistance) in peripheral tissues. Chronic hyperinsulinemia, however, is a possible factor favoring cancer initiation and/or progression in diabetic patients due to the mitogenic effect of insulin. The heterogeneity and complexity of different tissue exposure to hyperinsulinemia in diabetic individuals does not allow the quantification of the role of insulin in promoting cancer risk in the different organs of different diabetic patients.
One example is the potentially increased risk of lung cancer in diabetic patients using the recently introduced inhaled insulin (von Kriegstein & von Kriegstein 2007). The long-term effects of this form of therapy are unknown. Although short-term studies in animals have shown no substantial effect on cell proliferation indices, the high insulin concentration at alveolar and bronchiolar epithelia (due to the fact that only 10–25% of inhaled insulin is absorbed) has raised safety concerns about the possibility that it may promote lung cancer. These concerns have been recently reinforced by the long-time surveillance analysis, indicating that 6 out of 4740 (0.13%) diabetic patients treated with inhaled insulin but only 1 out of 4292 comparator-treated patients (0.02%) developed lung cancer (Mitri & Pittas 2009).
There are multiple and complex mechanisms potentially responsible for the mitogenic effects of insulin.
First, when insulin levels increase (as in the post-prandial surge in insulin-resistant subjects or after insulin injection), insulin may bind and activate the related insulin-like growth factor-I (IGF-I) receptor, which shares ∼80% homology with the insulin receptor (IR), but has a more potent mitogenic and transforming activity. Moreover, insulin decreases IGF-I-binding proteins (IGF-BP1 and, perhaps, IGF-BP2; Kaaks & Lukanova 2001): this will result in increased free IGF-I, the biologically active form of the growth factor.
Secondly, many cancer cells have an increased IR content (Papa et al. 1990; Fig. 3A). The IR may be expressed in two different isoforms, A and B, produced by an alternative splicing of the IR gene transcript (Moller et al. 1989). In malignant cells, the A isoform (IR-A) expression is predominant (Frasca et al. 1999, Sciacca et al. 1999, Kalli et al. 2002; Fig. 3B), and its activation, at variance with the IR-B isoform, elicits more mitogenic than metabolic effects (Frasca et al. 1999). By binding to the overexpressed IR-A, insulin may favor cancer progression and facilitate the growth of tumors that would otherwise have likely remained clinically irrelevant for an undetermined length of time.
Finally, insulin mitogenic activity might be enhanced at the cellular level by post-receptor molecular mechanisms, including insulin (or its synthetic analogs) residence time on the receptor (De Meyts et al. 1995) and the intracellular up-regulation of the insulin mitogenic pathway. Experimental data indicate that this pathway, unlike the insulin metabolic pathway, may not be blunted in the condition of insulin resistance typical of diabetes (Fig. 4). The AMP-activated protein kinase (AMPK), mammalian target of rapamycin (mTOR), and insulin-signaling pathway represent three interrelated components of a complex mechanism controlling cell responses to nutrient availability, and their dysregulation may favor malignant cell proliferation in response to hyperinsulinemia.
In conclusion, strong but circumstantial evidence indicates a role for endogenous hyperinsulinemia and of exogenous insulin or analogs in promoting cancer growth in diabetic patients. However, the clinical relevance of this pro-cancer effect of insulin in diabetic patients is still unclear.
Anti-diabetic drugs that may influence cancer risk in diabetic patients
Most diabetic patients are treated for years or decades with a variety of drugs (Table 2). The potential role of these drugs in favoring cancer is unclear but most likely minor, if any. Data are not conclusive because the large majority of diabetic patients change the drug dosage and/or the type many times during the course of the disease. Moreover, many are treated with more than one drug. Epidemiological studies on this issue, therefore, are difficult to interpret and often inconclusive.
Oral hypoglycemic agents used to treat type 2 diabetes mellitus
|Pharmacological class||Pharmacological compound||Mechanism of action|
|Biguanides||Metformin||Insulin sensitizer (reduces insulin resistance pre-eminently at hepatic level)|
|Thiazolidinediones (glitazones)||Rosiglitazone||Insulin sensitizers (reduce insulin resistance pre-eminently at muscle and fat level)|
|Sulphonylureas||Glipizide||Secretagogues (stimulate insulin secretion)|
|Meglitinides||Repaglinide||Short-term secretagogues (stimulate insulin secretion)|
|α-Glucosidase inhibitors||Acarbose||Reduces carbohydrate absorption|
GLP-1 analogs and gliptines (Dpp-4 inhibitors) have been introduced recently for diabetes treatment and no data are available on their potential influence on the cancer risk in diabetic patients.
The three major oral anti-diabetic drug families (sulphonylureas, biguanides, and thiazolidinediones) have a different mechanism of action. Sulphonylureas stimulate endogenous insulin secretion, while the other two categories of compounds are insulin sensitizers, i.e. they make tissues more responsive to insulin and, therefore, decrease hyperinsulinemia. If hyperinsulinemia plays a role in increasing cancer risk and progression in diabetic patients, it is reasonable to expect that these drugs will have a different effect on the association between diabetes and cancer. The biguanide metformin, widely used for more than 30 years and currently suggested as first-line therapy in type 2 diabetic patients, has been recently reported to reduce cancer risk (odds ratio=0.86) when compared with untreated patients (Evans et al. 2005). In addition to lowering the amount of circulating insulin, another possible mechanism for the anti-cancer effect of metformin is the stimulation of AMPK (an enzyme inducing glucose uptake by muscles) and its upstream regulator LKB1, a well-recognized tumor suppressor protein (Luo et al. 2005). AMPK activators act as anti-proliferative agents because they reduce insulin (and IGF-I)-signaling downstream of the receptor and, therefore, inhibit insulin-stimulated proliferation (McCarty 2004, Ruderman & Prentki 2004). Hence, the anti-cancer effect of metformin can be explained by this dual mechanism.
Recent studies in MCF-7, BT-474, and SKBR-3 human breast cancer cells showed that in vitro metformin inhibited cell proliferation, reduced colony formation, and caused partial cell cycle arrest (Alimova et al. 2009). These effects mainly occurred via MAPK, AKT, and mTOR inhibition and were replicated also in erbB2-overexpressing cells. On the basis of both epidemiological data and in vitro studies, a clinical trial for evaluating metformin activity on breast cancer cell proliferation (Ki67 index) is currently undergoing in 100 breast cancer patients (Cazzaniga et al. 2009).
Data on the other insulin-sensitizing drug (thiazolidinediones) are more controversial. A beneficial (Govindarajan et al. 2007), neutral (Koro et al. 2007), or even deleterious (Ramos-Nino et al. 2007) effect has been reported for different types of cancer. The biological mechanism of these compounds is to activate PPARγ receptors, which, in several in vitro experimental models, has shown a potential anti-cancer effect (Aiello et al. 2006). In addition to lowering hyperinsulinemia, this effect can explain the described anti-cancer effect of glitazones. In any case, the use of these compounds is too recent and too limited to consider the present meager epidemiologic observations reliable.
The third group of drugs (sulphonylureas) are secretagogues, i.e. increase insulin secretion and cause hyperinsulinemia. As expected, therefore, they have been associated with an increased risk of cancer (Bowker et al. 2006). Different sulphonylureas may have different effects, with glyburide being more deleterious than gliclazide (Monami et al. 2007). Although their effect on cancer risk is attributed to the prolonged hyperinsulinemia that they induce in patients, a direct effect on cancer (either positive or negative) cannot be excluded.
In conclusion, some evidence suggests that the biguanide metformin may reduce cancer risk in diabetic patients but, in general, the influence of anti-diabetic drugs on the risk of cancer is not well studied and evidence is weak, indirect, and controversial.
Other factors that may influence the risk of cancer in diabetes
Over 80% of type 2 diabetic patients are obese. Obesity is associated with a higher incidence and a higher mortality in cancer (Adami & Trichopoulos 2003, Vigneri et al. 2006). Moreover, cancer mortality significantly increases with increasing patient BMI (Calle et al. 2003). Fat distribution in the body is also important: central (upper body or android) obesity is more harmful than gynoid obesity in terms of increased risk and worst cancer outcome. Given these observations, it is evident that studies on the diabetes–cancer association are influenced by the high prevalence of obesity in DM patients. Since both DM and obesity are characterized by hyperinsulinemia and higher cancer incidence, it is difficult to identify the contribution of each of the two conditions.
Among the many possible mechanisms involved, hyperinsulinemia (which is typical of central obesity), diet and nutritional factors causing a positive energy balance, and other hormone abnormalities have been indicated as causal factors.
A tight correlation has been observed between obesity, circulating estrogen levels, and increased breast cancer risk (Key et al. 2003, Cleary & Grossmann 2009) especially in post-menopausal women. Obese post-menopausal women usually present an increase in both estrone and estradiol, a likely consequence of the increased aromatase activity of the adipose tissue (Reed & Purohit 2001). Considering the growing prevalence of obesity and diabetes in both developed and developing countries, these data might explain the reported rise in estrogen receptor-positive breast cancers (Glass et al. 2007). Several other molecular alterations associated with obesity might also be responsible for the higher incidence of breast cancer found in obese (and obese-diabetic) pre- and post-menopausal women. Preclinical evidence has suggested that leptin, an adipocyte-derived cytokine, highly expressed in obese subjects, promotes breast cancer cell proliferation (Hu et al. 2002), an observation that has not yet been confirmed in the clinical setting since an association between leptin levels and breast cancer outcome has not been demonstrated (Goodwin et al. 2005). Another adipokine produced by the adipose tissue, adiponectin, which is inversely correlated with body fat, might exert a protective effect on breast epithelial cells since its addition to different breast cancer cell lines inhibited proliferation and enhanced apoptosis (Cleary et al. 2009).
Most diabetic patients present both hyperglycemia and hyperinsulinemia. Thus, it is difficult to distinguish the specific role of each abnormality in increasing cancer risk.
It is known that a high intake of sugar and refined carbohydrates and elevated blood glucose levels are strongly associated with the risk of cancer (Krone & Ely 2005). It is also known that impaired glucose tolerance without diabetes is associated with increased cancer risk (Dankner et al. 2007). Both these conditions, however, are also characterized by hyperinsulinemia. Although much convincing evidence demonstrates an association between hyperglycemia and cancer, it has yet to be demonstrated that hyperglycemia per se is an independent risk factor.
Possible mechanisms implicated include the role of an abnormal energy balance and the effect of hyperglycemia in impairing the effect of ascorbic acid on the intracellular metabolism and reducing the effectiveness of the immune system. Further evidence suggests a role for the oxidative stress-responsive genes (like thioreodoxin-interacting protein) that are sensitive to hyperglycemia and regulate the level of reactive oxygen species (ROS; Turturro et al. 2007).
Free fatty acids
Deregulation of fatty acid synthase (FASN) activity, which catalyzes de novo fatty acids biogenesis (Hillgartner et al. 1995, Semenkovich et al. 1995), could also play a role in the pathogenesis of insulin resistance, diabetes, and cancer. FASN activity is important for de novo fatty acid synthesis in the liver and is stimulated by a low-fat/high-carbohydrate diet (Hudgins 2000, Hudgins et al. 2000). Interestingly, FASN expression is increased in insulin-resistant/hyperinsulinemic patients (Moustaid et al. 1996, Claycombe et al. 1998), and its increased activity further worsens insulin resistance and may result in NAFLD (Postic & Girard 2008), which is associated with an increased risk of hepatocarcinoma (Caldwell & Lazo 2009). FASN activity is also increased in cancer cells, where de novo fatty acid synthesis is crucial for membrane remodeling during cell migration and proliferation, as well as for lipid-based post-translational modifications of intracellular proteins in highly proliferating cell populations (i.e. myristylation of RAS). The concept that FASN is directly involved in affecting tumor progression derives also from studies with the FASN blocker cerulenin (Lupu & Menendez 2006a,b). Indeed, cell exposure to this inhibitor results in cytostatic, cytotoxic, and apoptotic effects in vitro and retards the growth of tumor in xenograft models (Menendez et al. 2009).
Therefore, FASN activity and fatty acid production are another possible link between diabetes and cancer as indicated by the hypothesis that insulin-resistant conditions such as obesity, type 2 diabetes, and cancer are favored by common FASN-driven ‘lipogenic state’ (Menendez et al. 2009).
Chronic inflammation and oxidative stress
The metabolic abnormalities that characterize diabetes, especially under conditions of poor metabolic control, increase oxidative stress and cause a permanent pro-inflammatory condition. This chronic pro-inflammatory state (which persists for years or decades) reduces intracellular anti-oxidant capacity, predisposing susceptible cells to malignant transformation. In fact, high concentrations of diverse free radicals and oxidants generate a potent ROS that can damage cell DNA by direct oxidation or by interfering with the mechanisms of DNA repair (Federico et al. 2007). ROS may also react with proteins and lipids, forming derivative products that may alter intracellular homeostasis favoring the accumulation of mutations that, in turn, contribute to the multistage carcinogenesis process (Ohshima et al. 2003).
A possible additional mechanism is related to mitochondrial dysfunction, a well-recognized abnormality in diabetes. DNA repair is a high energy consuming process that requires increased mitochondrial activity: stimulating malfunctioning mitochondria will not only provide low, insufficient energy supply, but also increase ROS production (Cebioglu et al. 2008).
Moreover, an additional factor correlated with insulin resistance is the pro-inflammatory cytokine tumor necrosis factor α (TNFα) produced by the adipose tissue (Kern et al. 2001). TNFα induces development and progression of many tumors (Szlosarek et al. 2006) by strongly activating nuclear factor-kappa B (NF-κB), which mediates many of the pro-tumoral effects of TNFα.
In conclusion, DM, by mechanisms both specific to diabetes and common with other chronic degenerative diseases, might accelerate the aging biological processes that favor cancerogenesis.
Drugs used to treat cancer may favor diabetes
A recently emerging issue is the possible adverse effect on glucose metabolism of anti-cancer therapies. Cancer patients can exhibit temporary hyperglycemic states or full-blown diabetes following steroid-based medication (administered before and during chemotherapy), or because of the specific mechanism of action of an anti-cancer drugs. Glucocorticoids are frequently used at a high dosage both to prevent and/or cure allergic reactions, inflammatory states caused by anti-cancer treatment, for their anti-edema effect and to alleviate fatigue. Glucocorticoids, however, have a potent diabetogenic effect because at high doses they cause severe insulin resistance, which can be compensated by hyperinsulinemia only when the patient's pancreas is functioning well. Otherwise, glucocorticoid administration may result in the worsening of a condition of pre-diabetes or undiagnosed diabetes and may transform mild diabetes into a clinically severe illness, possibly leading to a deadly hyperosmolar coma. Owing to the high prevalence of diabetes and pre-diabetes (over 15–20% in the aged population, which is more prone to cancer), this is a real health risk.
Apart from corticosteroids, anti-androgens may also adversely affect glucose metabolism. Androgen deprivation therapy is the fundamental treatment of prostate cancer. This therapy causes a variety of metabolic abnormalities that include decreased insulin sensitivity and altered lipid profile. It therefore, increases risk of diabetes and cardiovascular disease (Saylor & Smith 2009).
Androgens are important determinants of body composition: their inhibition increases fat mass and decreases lean body mass. In patients treated with GnRH agonists and/or nonsteroidal anti-androgens, such as flutamide, bicalutamide, and nilutamide, or with the steroidal anti-androgen cyproterone acetate, ‘sarcopenic obesity’ is favored, a combination of excess body weight and reduced muscle mass. Fat accumulation is primarily subcutaneous and is often associated with increased total cholesterol, triglycerides, and high-density lipoprotein (HDL). These changes result in insulin resistance and, sometimes, diabetes. In a recent study in over 70 000 subjects with locoregional prostate cancer, those who were treated with GnRH had a 44% increased risk of developing diabetes (Keating et al. 2006). Diet and lifestyle interventions with a 5–10% weight loss and statin drugs are the main strategies for preventing or treating the metabolic complications of androgen deprivation therapy in prostate cancer patients.
The other most currently employed targeted anti-cancer molecules do not significantly affect glucose homeostasis. However, an increasing number of compounds are being tested for therapeutic use which alter the IGF-I system and its intracellular pathways. The increasing use of these compounds may amplify the frequency of anti-cancer drug-related diabetes. Since IGF-I signaling plays a key role in both tumor progression and glucose homeostasis, therapies targeting the IGF system for its pro-cancer effect may at the same time cause hyperglycemia. In this paragraph, we will examine drugs and mechanisms responsible for hyperglycemia induced by novel anti-cancer therapies that may alter the insulin–glucose balance.
IGF-I system targeting anti-cancer treatments
IGF-I and insulin, their receptors and their intracellular signaling pathways share large similarities. Likewise, the biological (metabolic and mitogenic) effects of the two hormones partially overlap. Because of the well-known role of IGF-I as a cancer-promoting factor, many efforts have been made to block its function in cancer patients. However, these efforts may have a detrimental effect on glucose metabolism through three different mechanisms: i) the inhibition of the IGF-I insulin-mimetic effect (Kuzuya et al. 1993, Fernandez et al. 2001, Pennisi et al. 2006); ii) the increase in circulating GH levels due to the lack of IGF-I feedback (GH is a potent diabetogenic hormone; Yakar et al. 2004); and iii) the possibility that agents that block IGF-I signaling might also cross inhibit the insulin-signaling pathway.
Currently, anti-cancer strategies inhibiting the IGF system include both direct targeting of the IGF-I receptor (IGF-IR) with both monoclonal antibodies and suppression of the IGF-IR-signaling pathway by protein kinase inhibitors (Fig. 5).
Several antibodies targeting the IGF-I peptide or the IGF-IR have been tested, but only the latter are currently undergoing preclinical testing or are in phase I–II trials for the treatment of both hematological (multiple myeloma, and leukemia) and solid (sarcomas, carcinomas of the lung, breast, colon, and prostate) tumors (Haluska et al. 2007, Lacy et al. 2008).
Hyperglycemia has been observed in a few patients enrolled in studies with the anti-IGF-IR antibody (Haluska et al. 2007, Lacy et al. 2008). This is likely to be a consequence of a compensatory increase in the circulating concentration of GH after IGF-I blockade, with the consequent insulin resistance (del Rincon et al. 2007) that may cause or worsen diabetes.
A second approach to IGF-I inhibition is to block IGF-IR signaling at the enzymatic level. Since IGF-IR is a transmembrane tyrosine kinase (TK) receptor, several TK inhibitors targeting IGF-IR have been developed and found to be active in preclinical models and in phase I clinical trials (Hofmann & Garcia-Echeverria 2005, Gable et al. 2006, Haluska et al. 2006, Ji et al. 2007, Mulvihill et al. 2008, Vasilcanu et al. 2008, Zimmermann et al. 2008). These small molecules may cause more serious toxicity than that observed with the IGF-IR-specific antibodies, as they cross the blood–brain barrier with the possibility of neurotoxicity for the inhibition of the neuroprotective effect of IGF-I. Unexpectedly, these TK inhibitors are associated with less hyperglycemia than IGF-IR-blocking antibodies. One possible explanation for this difference is that the TK inhibitors do not accumulate in muscle, leaving unaffected IR function on the metabolic process of this tissue (Pollak 2008). More research is needed to clarify this point.
Downstream of the receptor, IGF-I signaling occurs via the activation of enzymes and substrates like phosphatidylinositol 3-kinase (PI3K), AKT, and mTOR. When activated via IGF-IR, these substrates play a role in tumor cell proliferation and survival, but they are also activated via the IR and heavily contribute to glucose homeostasis. Several compounds targeting different signaling molecules downstream of the IGF-IR have been tested, as anti-cancer therapies are able to inhibit the mitogenic and anti-apoptotic effects of IGF-I in cancer cells (Fig. 5).
Targeting PI3K, the most proximal pathway component, has the advantage of providing a broader inhibition of downstream signaling compared with distal component inhibition (such as AKT and mTOR). Inhibitors like LY294002 and wortmannin effectively inhibit PI3K, but poor solubility and high toxicity have prevented their clinical application. New compounds (like PX-866) are now being tested in xenograft models and in phase I clinical trials (Ihle et al. 2004, 2009a,b, LoPiccolo et al. 2008). In xenograft models, PX-866 increases glucose and insulin levels as well as glucose intolerance. While metformin is not effective in counteracting this effect and lowering glucose levels, glitazones (e.g. pioglitazone) ameliorate glucose balance in these patients, without affecting the anti-tumor activity of the compound (Ihle et al. 2004, 2009a,b, LoPiccolo et al. 2008).
A variety of AKT inhibitors have been developed (including perifosine, phosphatidylinositol ether lipid analogs PIAs, and triciribine phosphate; Ihle et al. 2004, 2009a,b, LoPiccolo et al. 2008; Fig. 5). Clinical data concerning the anti-tumor activity of AKT inhibitors as well as their effect on glucose homeostasis are insufficient. Recent preliminary data obtained in a xenograft model with GSK690693, a novel ATP-competitive/pan-AKT kinase inhibitor, indicate that abrogating AKT activity results in increased glucose and insulin levels. Interestingly, the diabetogenic effect of GSK690693 is not reverted by either metformin or pioglitazone or GLP-I agonist exendin-4, but only by a low-carbohydrate diet (Rhodes et al. 2008, Crouthamel et al. 2009).
A further class of targeted drugs that may interfere with blood glucose levels is the inhibitors of the mTOR kinase. This mTOR serine/threonine kinase and the mTOR–raptor complex (TORC1) regulate cell cycle progression (i.e. G1–S phase transition) and increase the expression of angiogenic factors. When dysregulated, mTOR plays a key role in cell proliferation and neoplastic transformation favoring the development of resistance to several types of cancer therapy (Bjornsti & Houghton 2004, Panwalkar et al. 2004).
Several mTOR inhibitors have been developed in vitro (Fig. 5). Some of them have been used in clinical trials. The most important, everolimus (RAD001; 40-O-2-hydroxyethyl-rapamycil), an orally available ester derivative of the anti-fungal antibiotic sirolimus (rapamycin), is currently used as an immunosuppressive agent to prevent rejection in transplant recipients (Eisen et al. 2003, Lorber et al. 2005). Immunosuppression maintenance with everolimus has been associated with a significantly reduced risk of developing de novo malignancies after renal transplant (Kauffman et al. 2005). Everolimus forms a complex with the immunophilin FKBP-12, which then binds to and disrupts TORC1, leading to mTOR inhibition and G1 phase cell cycle arrest, apoptosis (Aguirre et al. 2004, Majumder et al. 2004), and angiogenesis suppression (Majumder et al. 2004). Temsirolimus is a further novel mTOR inhibitor of the same family, recently approved for the treatment of renal cell carcinoma with unfavorable clinical characteristics. As expected from mTOR inhibition, hyperglycemia, hypertriglyceridemia, and hypercholesterolemia have been observed in ∼20% of patients treated with these inhibitors. In particular, recent data have reported increased blood glucose levels in 26% of temsirolimus-treated patients with 11% displaying G3/G4 hyperglycemia (Bellmunt et al. 2008, Malizzia & Hsu 2008). Most diabetic patients treated with temsirolimus required an increase in their hypoglycemic treatment, and roughly 30% of nondiabetic patients had to begin a specific therapy to lower their blood glucose (Bellmunt et al. 2008, Malizzia & Hsu 2008). Treatment of mTOR inhibition-related hyperglycemia has not yet been studied.
Finally, inhibitors of the ABL TK may also affect glucose homeostasis. In vitro results indicate that ABL is involved in IR signaling and upon insulin stimulation enhances IR-dependent metabolic effects while attenuating the nonmetabolic ones (Frasca et al. 2007, Genua et al. 2009). Therefore, treatment with an ABL inhibitor was expected to impair glucose homeostasis (Fig. 6). However, adult patients with chronic and accelerated phase chronic myelogenous leukemia (CML), treated with the ABL inhibitor imatinib mesylate, have actually shown a consistent reduction in their blood glucose levels (Veneri et al. 2005). Interestingly, a recent report has described hyperglycemia in ∼10% of CML patients treated with nilotinib, a second-generation ABL kinase inhibitor currently used for individuals resistant or intolerant to imatinib (Kantarjian et al. 2006). The increase in fasting glucose registered after nilotinib therapy is predictive of drug response and apparently does not require administration of hypoglycemic drugs (Deremer et al. 2008). However, the follow-up of the study is too short to yield conclusive evidence, especially considering that patients responding to nilotinib will have to continue drug treatment indefinitely until disease progression.
In conclusion, in addition to glucocorticoid- and anti-androgen-dependent hyperglycemia, the use of molecular inhibitors of IGF-I, pro-mitogenic and anti-apoptotic signaling will likely become more diffuse in cancer patients, possibly causing hyperglycemia. Since diabetes and pre-diabetes have a high prevalence in the general population and patients treated with these novel anti-cancer compounds often have a considerable life expectancy, careful monitoring of glycemia is a requirement in all patients treated with agents that may interfere, at different levels, with glucose metabolism.
The complexity of the various diabetic conditions, the diversities in the biology of different forms of cancer, and the multiplicity of the possible mechanisms involved prevent a comprehensive and definite answer to many questions regarding the association of diabetes with an increased risk of cancer initiation and progression. Most epidemiologic studies have not carefully considered a series of confounding factors, and diabetic patients have not been adequately characterized for the type of diabetes, the duration of the disease, the drugs used for therapy, the quality of the metabolic control, or the presence of comorbidities.
Because of the intrinsic heterogeneity of both diabetes and cancer, studies on the association of the two diseases are not easy to carry out. Indeed, considering the wide array of possible mechanisms causing increased cancer incidence and mortality in diabetic patients, it is difficult to accurately define the aims, the recruitment criteria, and the appropriate design for such studies.
The available evidence indicates that the level of cancer risk related to diabetes will probably differ for each diabetic patient, on the basis of the cancer type and many other diabetes-related factors. Our present knowledge provides good evidence for a mild increase of cancer risk (and cancer mortality) in diabetic patients, more evident for some site-specific cancers. Present evidence, however, does not allow us to accurately define the general and the specific organ cancer risks in the individual diabetic patient. Because of the growing worldwide frequency of diabetes, this question needs to be properly addressed, in order to acquire a more rational approach to cancer prevention and treatment in diabetic patients.
Declaration of interest
There is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
The studies quoted in this work were partially supported by Associazione Italiana Ricerca sul Cancro (AIRC) to R Vigneri.
The authors are grateful to A Belfiore (University of Catanzaro) for useful comments and discussions.
De BlockCEDe LeeuwIHVan GaalLF1999High prevalence of manifestations of gastric autoimmunity in parietal cell antibody-positive type 1 (insulin-dependent) diabetic patients. The Belgian Diabetes Registry. Journal of Clinical Endocrinology and Metabolism844062–4067.
HaluskaPShawHMBatzelGNYinDMolinaJRMolifeLRYapTARobertsMLSharmaAGualbertoA2007Phase I dose escalation study of the anti insulin-like growth factor-I receptor monoclonal antibody CP-751,871 in patients with refractory solid tumors. Clinical Cancer Research135834–5840.
HarrisMIFlegalKMCowieCCEberhardtMSGoldsteinDELittleRRWiedmeyerHMByrd-HoltDD1998Prevalence of diabetes, impaired fasting glucose, and impaired glucose tolerance in U.S. adults. The Third National Health and Nutrition Examination Survey, 1988–1994. Diabetes Care21518–524.
IhleNTLemosRJrWipfPYacoubAMitchellCSiwakDMillsGBDentPKirkpatrickDLPowisG2009aMutations in the phosphatidylinositol-3-kinase pathway predict for antitumor activity of the inhibitor PX-866 whereas oncogenic Ras is a dominant predictor for resistance. Cancer Research69143–150.
IhleNTLemosRSchwartzDOhJHalterRJWipfPKirkpatrickLPowisG2009bPeroxisome proliferator-activated receptor gamma agonist pioglitazone prevents the hyperglycemia caused by phosphatidylinositol 3-kinase pathway inhibition by PX-866 without affecting antitumor activity. Molecular Cancer Therapeutics894–100.
JiQSMulvihillMJRosenfeld-FranklinMCookeAFengLMakGO'ConnorMYaoYPirrittCBuckE2007A novel, potent, and selective insulin-like growth factor-I receptor kinase inhibitor blocks insulin-like growth factor-I receptor signaling in vitro and inhibits insulin-like growth factor-I receptor dependent tumor growth in vivo. Molecular Cancer Therapeutics62158–2167.
LacyMQAlsinaMFonsecaRPaccagnellaMLMelvinCLYinDSharmaAEnriquez SaranoMPollakMJagannathS2008Phase I, pharmacokinetic and pharmacodynamic study of the anti-insulinlike growth factor type 1 receptor monoclonal antibody CP-751,871 in patients with multiple myeloma. Journal of Clinical Oncology263196–3203.
LimburgPJVierkantRAFredericksenZSLeibsonCLRizzaRAGuptaAKAhlquistDAMeltonLJIIISellersTACerhanJR2006Clinically confirmed type 2 diabetes mellitus and colorectal cancer risk: a population-based, retrospective cohort study. American Journal of Gastroenterology1011872–1879.
LorberMIMulgaonkarSButtKMElkhammasEMendezRRajagopalanPRKahanBSollingerHLiYCretinN2005Everolimus versus mycophenolate mofetil in the prevention of rejection in de novo renal transplant recipients: a 3-year randomized, multicenter, phase III study. Transplantation80244–252.
del RinconJPIidaKGaylinnBDMcCurdyCELeitnerJWBarbourLAKopchickJJFriedmanJEDrazninBThornerMO2007Growth hormone regulation of p85alpha expression and phosphoinositide 3-kinase activity in adipose tissue: mechanism for growth hormone-mediated insulin resistance. Diabetes561638–1646.