Increased Plasma Levels of Insulin-Like Growth Factor 2 and Insulin-Like Growth Factor Binding Protein 3 Are Associated with Endometrial Cancer Risk

Endometrial cancer is the most common gynecologic cancer in the United States. It is expected that 39,300 women will develop endometrial cancer in the U.S. in 2003; and of those, 6,600 women will die of their disease (1). Endometrial cancer is associated with excessive estrogen stimulation (2). Risk factors associated with endometrial cancer include obesity, diabetes, and exogenous estrogen use.

A number of epidemiological studies have shown evidence for a relationship between high circulating levels of insulin-like growth factors (IGFs) and the increased risk of several cancers, including premenopausal breast cancer, colon cancer, prostate cancer, and lung cancer (3–7). Additionally, the level of IGF binding protein 3 (IGFBP-3), a major IGFBP that suppresses the mitogenic action of IGF-1, is inversely associated with the risk of these cancers.

The IGF family is a complex system composed of two peptide ligands (IGF-1 and IGF-2), two cell surface receptors (IGF-1R and IGF-2R), 10 specific IGFBPs (IGFBP-1 to IGFBP-10), IGFBP proteases, as well as other IGFBP-interacting molecules (8). IGF-1 and IGF-2 are single-chain polypeptides with 62% homology to proinsulin. In contrast to insulin, they are produced by almost every cell in the body, and circulate in approximately 1000-fold higher concentrations than most other peptide hormones (9). Circulating IGF-1 levels are regulated by many factors, with being the main regulator of its production and secretion from the liver (10). IGF-1, in turn, feedbacks and negatively regulates growth hormone secretion from the pituitary (11).

Circulating IGF-1 levels change substantially with age, increasing slowly from birth to puberty, surging in puberty, and declining with age thereafter (12). Circulating IGF-2 levels are two to three times higher than circulating IGF-1 levels. IGF-2 is relatively stable after puberty and is not regulated by growth hormone (13). Both IGF-1 and IGF-2 are potent mitogenic and anti-apoptotic molecules involved in the regulation of epithelial cell proliferation. IGF-2 plays a fundamental role in embryonic and fetal growth. In the postnatal period, it is traditionally thought to be less important as it is substituted by IGF-1 (14).

IGFBPs are the major determinants of IGF bioavailability. More than 99% of the circulating IGFs are bound to IGFBPs and at least 75% of the bound IGF is carried as a trimeric complex composed of IGFBP-3 and the acid labile subunit (ALS) (15–17). The IGFBPs are found both in extracellular fluid (soluble form) and on the cell surface (membrane-associated form).

It has long been held that unopposed estrogenic stimulation of the endometrium is a risk factor for the development of endometrial hyperplasia and the progression to carcinoma. In vitro studies have demonstrated that estrogen enhances local IGF-1 protein and mRNA expression (18, 19). IGF-1 and IGF-2 are also involved in the regulation of endometrial growth as well as the interactions between the epithelial and stromal compartments of the endometrium (20–22). In the present study, we collected serum from 80 patients with endometrial cancer and 80 age-matched normal individuals with no history of cancer. The purpose of this group-matched hospital-based case-control study was to determine if circulating levels of IGF-1, IGF-2, and IGFBP-3 were associated with endometrial cancer risk.

After receiving IRB approval, a total of 160 subjects (80 cases and 80 controls) was selected for this study. Case subjects were selected from the Gynecologic Oncology Tumor Bank at The University of Texas MD Anderson Cancer Center (MDACC). The Gynecologic Oncology Tumor Bank at MDACC was created in 1997 to preserve tissue, serum, and urine specimens of women diagnosed with gynecologic malignancies for use in future research protocols. The case subjects for this study were patients with a newly histologically confirmed diagnosis of endometrial cancer consecutively enrolled in the tumor bank during 2001 and 2002. All case and control blood samples were taken from a fasting morning sample. They had not received surgery, radiotherapy, or chemotherapy before specimen collection. Control subjects were women with no history of cancer. Controls were matched with the cases by age in a 1:1 ratio. All subjects had previously signed informed consents for the collection and use of blood. Clinical and pathological information for both the case and control groups were obtained by a retrospective chart review.

Commercially available ELISA kits (DSL, Webster, TX) were used to determine the circulating levels of IGF-1, IGF-2, and IGFBP-3. To avoid degradation of IGF-1, IGF-2, and IGFBP-3, none of the samples had previously been defrosted. Case and control subjects were de-linked from any identifying information before assay and the investigators were blinded to whether samples were cases or controls. To separate IGFs from their binding proteins, all blood samples underwent an acid-ethanol extraction procedure before measurement according to the manufacturer's instructions. The sensitivities of the IGF-1, IGF-2, and IGFBP-3 assays are 0.03, 0.25, and 0.04 ng/ml, respectively. The efficiency of extraction had been determined to be between 83% and 101% for IGF-1, 85% and 110% for IGF-2. The coefficient of variation of intra-assay and inter-assay precision are between 4.5–7.1% and 4.8–8.8% for the IGF-1 assay; 1.7–4.2% and 5.2–7.7% for the IGF-2 assay; and 7.3–9.6% and 8.2–11.4% for IGFBP-3, respectively, as reported by the manufacturer (23).

ELISA was performed at room temperature according to the manufacturer's instructions. Briefly, each standard, control, and unknown blood extract were placed into microplate wells in triplicate. Following incubation with the antibody-enzyme conjugate solution, the microplate was washed thoroughly with wash solution. For color development, tetramethylbenzidine (TMD) Chromogen solution was added. After incubation, the reaction was stopped. The absorbance of the solution in the wells was immediately measured by spectrometer at a wavelength of 450 nm. The IGF-1, IGF-2, and IGFBP-3 plasma concentrations were determined from the standard curve by matching their mean absorbance readings with the corresponding IGF-1, IGF-2, and IGFBP-3 concentrations.

Distributions of demographic variables (age and race/ethnicity), clinico-pathological variables [grade, stage, and body mass index (BMI)], and IGF (IGF-1, IGF-2, and IGFBP-3) plasma values were described by percentages, means, and SDs. Differences in the distribution of these variables between case and control subjects were tested using the Pearson's χ² statistic for categorical variables and the Mann-Whitney test for continuous variables. Plasma levels of IGF-1, IGF-2, and IGFBP-3 were further categorized into quartiles based on the distribution of values in the control group for analysis by univariate and multivariate logistic models. Univariate and multivariate logistic regression models were used to assess the association between endometrial cancer and plasma levels of the growth factors of interest and BMI. Taking the lowest quartile level of each of the variables of interest as the reference category, crude and adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were calculated for the other three quartile levels (24). First, crude OR and 95% CI were calculated for each variable of interest (IGF-1 IGF-2, IGFBP-3). Then multivariate models were used to assess the association between each IGF marker and endometrial cancer while controlling for BMI. All P values were two-sided, and considered significant at P < 0.05. Statistical analysis was conducting using SPSS for Windows version 11.5 (SPSS Inc., 2002).

Table 1 summarizes the distribution of cases and controls by demographic and clinico-pathological variables. There were no differences in the distribution of age and race/ethnicity between cases and controls. However, BMI was significantly higher in the case group than in the control group (34.1 versus 27.8, P < 0.001, respectively). A higher proportion of cases than controls (56.3% versus 35%, respectively, P = 0.009) were obese (BMI ≥ 29.00) at the time of diagnosis. The majority of the cases (77.5%) were classified as stage I, whereas 18.8% were stage II or higher, and 3.8% were unstaged. Similarly, 39% of endometrial cancer were grade 1and only 14% were grade 3.

The distribution of IGF plasma levels among cases and controls is shown in Table 2. A statistically significant higher mean plasma level of IGF-2 was observed among cases than among controls (648 ng/ml versus 377 ng/ml, P < 0.001). In contrast, the mean values of IGF-1 and IGFBP-3 were significantly lower among cases than among controls. Mean plasma IGF-1 levels were 155 and 185 ng/ml (P = 0.03) for cases and controls, respectively; IGFBP-3 levels were 1693 and 2160 ng/ml (P = 0.001), respectively.

Crude and adjusted estimates of the association (OR and 95% CI) between IGF markers and endometrial cancer are presented in Table 3. In the univariate logistic regression analysis, IGF-2 levels were found to be positively associated with endometrial cancer. The risk of endometrial cancer was 7.4 times higher among women in the highest quartile of IGF-2 compared to those in the lowest quartile. This association persisted (OR = 9.67; 95% CI 3.29–28.43; P < 0.001) after controlling for BMI in a bivariate logistic model.

In contrast, an inverse relationship between higher plasma levels of IGF-1 and IGFBP-3 and the risk of endometrial cancer was observed. Women in the highest quartile of IGF-1 and IGFBP-3 were at 67% (OR = 0.33) and 75% (OR = 0.25) lower risk of endometrial cancer, respectively, compared with women in the reference category (Table 3). After adjusting for BMI, this inverse association remained statistically significant only for women in the highest quartile of IGFBP-3 (OR = 0.23; 95% CI 0.09–0.60; P = 0.003) but not for higher levels of IGF-1.

Consistent with other studies, higher levels of BMI were found to be positively associated with endometrial cancer and this association remained after controlling for IGF markers, individually and simultaneously. Women in the highest BMI quartile were 3.45 times higher risk of endometrial cancer (95% CI 1.39–8.61, P = 0.008) compared to women in the lowest BMI quartile after controlling for IGF markers (Table 3).

It has been hypothesized that increased cell proliferation and cell division stimulated by external or internal factors are associated with the development of many human cancers (25). IGFs play an important role in regulating epithelial cell proliferation, differentiation, apoptosis, and transformation (26). Case-control and prospective studies have reported that higher circulating levels of IGFs are associated with an increased risk for the development of a number of human cancers including breast, prostate, and lung cancers (3–5, 27).

To date, most studies have focused on the plasma levels of IGF-1 and IGFBP-3, or combinations thereof, in association with cancer risk (3, 4, 27, 28). Consistent results have shown an increased risk of solid tumors in association with high levels of IGF-1 and reduced risk of solid tumors in association with high levels of IGFBP-3. However, the role of IGF-2 in cancer development generally has been less well recognized. This may be largely due to early evidence suggesting that IGF-1 is a more important growth factor in postnatal development with IGF-2 playing a larger role in embryologic development (29).

On the basis of the previous investigations, we hypothesized that elevated plasma levels of IGF-1 and IGF-2 and decreased plasma levels of IGFBP-3 may be associated with an increased risk of endometrial cancer development. Our results demonstrate a significantly higher level of plasma IGF-2 in the case subjects when compared to the controls. This relationship persisted even after controlling for BMI. In addition, a statistically significant dose-response relationship between the level of IGF-2 and the risk for developing endometrial cancer was seen. Plasma levels of IGFBP-3 were significantly lower in women with endometrial cancer than in control subjects, demonstrating an inverse relationship between the plasma level of IGFBP-3 and the risk of developing endometrial cancer. This finding is similar to what investigators have reported for other cancer sites (28). In contrast to other solid tumor sites, we found lower levels of IGF-1 in the endometrial cancer cases as compared to the controls.

Our findings are similar to what has been reported by Petridou et al. (30. In their recently published study, they report a positive association between plasma levels of IGF-2 and the risk of developing endometrial cancer and an inverse relationship between IGF-1 and endometrial cancer risk. IGFBP-3 was not found to be significantly associated with endometrial cancer risk in their study. In contrast, a previous study by Ayabe et al. ([31]) has shown that increased circulating levels of IGF-1 and decreased serum levels of IGFBP-1 are associated with increased endometrial cancer risk in postmenopausal women. However, the study by Ayabe et al. only included 23 patients with endometrial cancer, and accordingly was limited by the relatively small number of patients.

Obese individuals are presumed to be at higher risk for the development of endometrial cancer due to the increased bioavailability of estrogen. In addition to its mitogenic activity, studies have demonstrated that elevated IGF-2 is directly associated with obesity (32–35). Molecular genetic analysis found three genotypes of IGF-2 categorized as G/G, G/A, and A/A. It has been shown that the homozygous A/A IGF-2 genotype is associated with obesity in Caucasians (32, 33). Furthermore, free and total plasma levels of IGF-2 were found increased in both non-diabetic obese subjects and type 2 diabetics, whereas IGF-1 remained unchanged (34). Our own data failed to show a statistically significant relationship between BMI and levels of IGF-1, IGF-2, or IGFBP-3; however, BMI was found to be positively associated with endometrial cancer. Given the well-established association between obesity and IGF-2, it is possible that direct link between BMI, IGF-2 levels, and the risk for developing endometrial cancer exists. A larger series may be necessary to identify this relationship.

Although this study did not find obesity and IGF-2 to be associated, we did find the two factors to be independently associated with endometrial cancer. As this was a retrospective case-control study and examined a relatively small number of patients, no definitive conclusions can be made as to whether IGF-2 actively promotes the development of endometrial cancer or is simply a biomarker for the risk of developing endometrial cancer.

In summary, we found that elevated plasma IGF-2 levels and lower IGFBP-3 levels were positively associated with an increased risk for developing endometrial cancer. The results warrant further prospective studies to elucidate the possible mechanisms of IGF-2 in the development of endometrial cancer and to determine if plasma levels of IGF-2 and IGFBP-3 could serve as a useful biomarker for the development of endometrial cancer.