Abstract
African–American (AA) race/ethnicity, lower body mass index (BMI), and higher IGF1 levels are associated with premenopausal breast cancer risk. This cross-sectional analysis investigated whether BMI or BMI at age 21 years contributes to racial differences in IGF1, IGF2, IGF-binding protein 3 (IGFBP3), or free IGF1. Participants included 816 white and 821 AA women between ages 40 and 79 years across a wide BMI range (18.5–40 kg/m2). Compared with white women, AA women had higher mean IGF1 (146.3 vs 134.4 ng/ml) and free IGF1 (0.145 vs 0.127) levels, and lower IGF2 (1633.0 vs 1769.3 ng/ml) and IGFBP3 (3663.3 vs 3842.5 ng/ml) levels (all P<0.01; adjusted for age, height, BMI, BMI at age 21 years, and menopausal status). Regardless of race, IGF1 and free IGF1 levels rose sharply as BMI increased to 22–24 kg/m2, and then declined thereafter, while IGF2 and IGFBP3 levels tended to rise with BMI. In contrast, BMI at age 21 years was inversely associated with all IGF levels, but only among white women (P-interaction=0.01). With the decline in IGF1 with BMI at age 21 years among whites, racial differences in IGF1 significantly increased among women who were obese in early adulthood. In summary, BMI was associated with IGF1 levels regardless of race/ethnicity, while obesity during childhood or young adulthood may have a greater impact on IGF1 levels among white women. The effects of obesity throughout life on the IGF axis and racial differences in breast cancer risk require study.
Introduction
African–American (AA) women are at a greater risk for premenopausal breast cancer and estrogen and progesterone receptor negative (ER−/PR−) tumors, and have an overall poorer prognosis, compared with white women ( Hausauer et al. 2007, Brinton et al. 2008, Pfeiffer et al. 2008). It is unclear why ER−/PR− tumors are more common among AA women ( Brinton et al. 2008); however, insulin-like growth factor 1 (IGF1) and free (unbound) IGF1 levels are also associated with premenopausal breast cancer risk and therefore may have a role in breast tumor progression in the absence of ER or PR activation ( Renehan et al. 2006).
IGF1 production is stimulated by GH and is a key mediator of mitogenic and anti-apoptotic activity ( Gennigens et al. 2006). IGF-binding protein 3 (IGFBP3) regulates IGF1 bioavailability and also has a functional activity to induce or inhibit apoptosis independently of IGF1 bioavailability. Analyses of data from National Health and Nutrition Examination Survey (NHANES III), Nurses Health Study II (NHS II), and the Multiethnic Cohort Study (MEC) found that AA women had higher IGF1 levels and lower IGFBP3 levels compared with white women ( DeLellis et al. 2004, Pinheiro et al. 2005, Berrigan et al. 2008), consistent with higher IGF1 bioavailability and a potential role for the IGF axis in premenopausal breast cancer risk among AA women.
Aside from race/ethnicity, IGF1 levels may also be associated with obesity or energy availability. Several studies found an inverse association between body mass index (BMI) and IGF1 levels in whites ( Jernstrom et al. 2001b, Gram et al. 2006, Henderson et al. 2006). In contrast, analysis of MEC data found that BMI was not associated with IGF1 levels among AAs ( Henderson et al. 2006). The prevalence of obesity is increasing nationally and is highest among AA women ( Flegal et al. 2002); however, it is unclear whether racial/ethnic differences in the relationship between obesity and the IGF axis may contribute to differences in breast cancer risk between race/ethnicity groups.
Our goal was to investigate the role of obesity upon serum IGF1, IGFBP3, and IGF2 levels in a large sample of white and AA women to help clarify whether obesity contributes to racial differences in the IGF axis.
Materials and methods
Participants
Women in this study were selected from members of the Southern Community Cohort Study (SCCS), a prospective cohort investigation initiated in 2001 designed to identify the determinants of cancer incidence and mortality in a racially diverse population. Detailed methods of the SCCS have been reported ( Signorello et al. 2005), and study information is also available at www.southerncommunitystudy.org. Briefly, adult men and women visiting one of the 71 community health centers located throughout the Southeastern US were approached for recruitment. Potential participants were eligible if they were between 40 and 79 years of age and had not been under treatment for cancer (except nonmelanoma skin cancer) within the past year. All SCCS protocols have been approved by IRBs at Vanderbilt University Medical Center and Meharry Medical College in Nashville, TN, USA.
In 2006, a project was developed within a stratified sample of SCCS female participants to investigate racial/ethnic differences in blood biomarkers potentially relevant to breast cancer risk. This sample included 1000 AA and 1000 white female participants randomly selected from female SCCS cohort members with frozen-stored blood and without a prior breast cancer diagnosis. Furthermore, sampling was stratified by BMI categories (18.5–24.9, 25.0–29.9, and 30.0–45.0), age (in 5-year intervals), and menopausal status. Premenopausal women included women with at least one menstrual cycle within the past 6 months, while postmenopausal women were 6 months or more since their last cycle whether due to natural aging, hysterectomy with or without oophorectomy, or other causes.
Data and sample collection
All SCCS participants provided written informed consent prior to completing a comprehensive, in-person, baseline interview administered by a trained interviewer. The computer-assisted interview included demographics, weight and height, and a wide range of other potential cancer risk factors. Women reporting 6 or more months since their last menstrual cycle were asked why their periods had stopped. A blood sample was collected from participants at recruitment. Blood samples were refrigerated immediately after collection, and then shipped cold on that day to Vanderbilt University. Blood was processed and stored frozen at −80 °C on average within 1.2 days (range 1–5 days) after collection.
Laboratory assays
Serum concentrations of IGFI, IGF2, and IGFBP3 were determined with the use of commercially available ELISA kits (DSL, Inc., Webster, TX, USA). The calibrators used in the assays ranged between 4.5 and 640 ng/ml for IGF1, 50 and 2000 ng/ml for IGF2, and 2.5 and 100 ng/ml for IGFBP3. For IGFBP3 measurement, plasma samples were diluted at 1:100 in an assay buffer. The intra-assay coefficient of variations (CV) were 4.0, 2.4, and 1.8%, and inter-assay CV were 6.5, 7.4, and 7.1% respectively for IGF1, IGF2, and IGFBP3. Each assay has no cross-reaction with other members of the IGF family. Complete data for blood IGF values were available for 1984 study participants. The molar ratio of IGF1 to IGFBP3 representing an index of free IGF1 was calculated as (IGF1 (ng/ml)×1.30)/(IGFBP3 (ng/ml)×0.36). We did not find an association between IGF markers and transport time, duration of blood storage in freezer, or the time between the last meal and blood collection.
Data analyses
From the initial sample of 1984 participants with IGF data, we excluded women taking hormone replacement therapy (n=202), insulin for the treatment of diabetes (n=111), or missing weight at age 21 data (n=33). We further excluded one participant with an IGF1 value more than 8 s.d.s from the group mean. The final study population included 1637 participants (816 white and 821 AA women). Wilcoxon signed rank and χ2 tests were used to compare median values and categorical levels respectively between AA and white women.
We calculated mean IGF marker levels within white or AA participants after adjusting for age (5-year intervals), BMI (categorized at 2 unit intervals to accommodate nonlinear association), height in meters (continuous), menopausal status at blood collection (premenopausal, postmenopausal), and BMI at age 21 years (<20, 20–24, 25–29, and 30 or more) in a linear regression model. Additional control for time of day of blood collection or year of blood collection (storage time) did not affect our results. Tests for trend were performed by including the categorical variable for the covariate of interest as a continuous variable in the full model. Tests for interaction represented differences with race in the association between each IGF marker and another covariate, and were evaluated by the significance of a race×covariate cross-product term in a model that also included each main effect term and other listed factors. Distributions of all IGF markers approached a normal distribution, with low kurtosis and skewness, such that log transformation of the IGF data did not lead to the normalization of these distributions. A Box–Cox procedure suggested raising values of IGF1 and IGF2 to the power 0.75 or transforming IGFBP3 and free IGF1 values with a square root function. However, study results using either the natural scale for each IGF marker or the transformed variable were almost identical. We therefore report mean IGF levels, standard errors, and P values from statistical tests using IGF values on the natural scale of each IGF marker.
Results
Study population
AA and white women were selected to have a similar age and BMI ( Table 1). AA and white women also had a similar BMI at age 21 years (Wilcoxon P=0.88, χ2 P=0.41) and height (Wilcoxon P=0.07, χ2 P=0.19). There were 714 postmenopausal women, including 383 women who were postmenopausal because of a clinical intervention such as surgery.
Study population characteristics
| White (n=816) | African–American (n=821) | ||||
|---|---|---|---|---|---|
| n | % | n | % | ||
| Age (years) | 40–44 | 301 | 36.9 | 311 | 37.9 |
| 45–49 | 211 | 25.9 | 219 | 26.7 | |
| 50–54 | 107 | 13.1 | 120 | 14.6 | |
| 55–59 | 79 | 9.7 | 69 | 8.4 | |
| 60–64 | 70 | 8.6 | 34 | 4.1 | |
| 65–69 | 25 | 3.1 | 39 | 4.8 | |
| 70–79 | 23 | 2.8 | 29 | 3.5 | |
| BMI | 18–19 | 35 | 4.3 | 19 | 2.3 |
| 20–24 | 169 | 20.7 | 196 | 23.9 | |
| 25–29 | 211 | 25.8 | 205 | 25.0 | |
| 30–34 | 205 | 25.1 | 209 | 25.5 | |
| 35–39 | 114 | 14.0 | 125 | 15.2 | |
| 40–45 | 82 | 10.1 | 67 | 8.2 | |
| BMI at age 21 years | 13–19 | 283 | 34.7 | 268 | 32.6 |
| 20–24 | 366 | 44.9 | 370 | 45.1 | |
| 25–29 | 103 | 12.6 | 125 | 15.2 | |
| 30–51 | 64 | 7.8 | 58 | 7.1 | |
| Height (m) | 1.22–1.49 | 29 | 3.6 | 24 | 2.9 |
| 1.50–1.59 | 196 | 24.0 | 176 | 21.4 | |
| 1.60–1.64 | 222 | 27.2 | 203 | 24.7 | |
| 1.65–1.69 | 191 | 23.4 | 229 | 27.9 | |
| 1.70–1.96 | 178 | 21.8 | 189 | 23.0 | |
| Premenopausal | 458 | 56.1 | 465 | 56.6 | |
| Postmenopausal | 358 | 43.9 | 356 | 43.4 | |
| Natural menopause | 161 | 19.7 | 170 | 20.7 | |
| Surgical or other menopause | 197 | 24.2 | 186 | 22.7 | |
IGFs by menopausal status and race
Premenopausal women had consistently higher IGF levels than postmenopausal women after controlling for age and other factors. These differences in IGF levels associated with menopausal status were similar between AA and white women (P-interaction>0.05 for all tests of a race by menopause interaction) ( Table 2). Racial differences in IGF levels persisted among postmenopausal women regardless of whether they experienced a natural or a surgical menopause (all P-interactions>0.05).
Adjusted insulin-like growth factor (IGF) levels by race and menopausal status
| White | African–American | ||||
|---|---|---|---|---|---|
| Mean | s.e.m. | Mean | s.e.m. | P | |
| Alla | |||||
| IGF1 (ng/ml) | 134.5 | 2.9 | 146.8 | 2.9 | <0.01 |
| IGF2 (ng/ml) | 1773.9 | 20.2 | 1639.7 | 20.4 | <0.01 |
| IGFBP3 (ng/ml) | 3874.6 | 54.1 | 3703.0 | 54.5 | <0.01 |
| Free IGF1 | 0.126 | 0.002 | 0.144 | 0.002 | <0.01 |
| Premenopausalb | |||||
| IGF1 (ng/ml) | 148.4 | 11.9 | 157.0 | 11.7 | 0.02 |
| IGF2 (ng/ml) | 1897.2 | 82.7 | 1745.5 | 82.0 | <0.01 |
| IGFBP3 (ng/ml) | 4183.3 | 212.6 | 3950.3 | 210.6 | <0.01 |
| Free IGF1 | 0.128 | 0.010 | 0.146 | 0.010 | <0.01 |
| Postmenopausalc | |||||
| IGF1 (ng/ml) | 128.2 | 3.9 | 144.7 | 3.9 | <0.01 |
| IGF2 (ng/ml) | 1778.7 | 28.5 | 1666.9 | 28.6 | <0.01 |
| IGFBP3 (ng/ml) | 3872.5 | 79.4 | 3771.4 | 79.6 | 0.25 |
| Free IGF1 | 0.119 | 0.003 | 0.140 | 0.003 | <0.01 |
| Surgical menopaused | |||||
| IGF1 (ng/ml) | 126.7 | 5.2 | 147.4 | 5.4 | <0.01 |
| IGF2 (ng/ml) | 1800.0 | 41.0 | 1723.9 | 42.6 | 0.09 |
| IGFBP3 (ng/ml) | 3934.6 | 111.6 | 3951.6 | 115.9 | 0.89 |
| Free IGF1 | 0.117 | 0.004 | 0.136 | 0.004 | <0.01 |
| Natural menopaused | |||||
| IGF1 (ng/ml) | 132.9 | 6.3 | 145.8 | 6.3 | 0.06 |
| IGF2 (ng/ml) | 1764.9 | 42.3 | 1608.1 | 42.7 | <0.01 |
| IGFBP3 (ng/ml) | 3846.5 | 119.9 | 3615.8 | 121.1 | 0.08 |
| Free IGF1 | 0.124 | 0.005 | 0.145 | 0.005 | <0.01 |
Mean IGF levels adjusted for aage, height, BMI, BMI at age 21 years, and menopausal status; bage, height, BMI, and BMI at age 21 years; cage, height, BMI, BMI at age 21 years, and reason for menopause; dage, height, BMI, and BMI at age 21 years.
AA premenopausal and postmenopausal women had significantly higher IGF1 levels and significantly lower IGF2 levels, compared with similar white women ( Table 2). IGFBP3 levels were higher among whites than AAs, although differences in IGFBP3 levels between white and AA women were statistically significant only among premenopausal women. Free IGF1 levels were significantly higher among premenopausal and postmenopausal AA women compared with similar white women. Indeed, free IGF levels were highest among premenopausal AA women, followed by postmenopausal AA women, premenopausal white women, and postmenopausal white women respectively.
IGFs and BMI by race and menopausal status
IGF1 levels were higher among AA compared with white women across the range of BMI categories ( Table 3 and Fig. 1). However, the relationship between IGF1 and BMI was similar between AA and white women (P-interactions>0.05). The highest IGF1 levels were found among participants with a BMI between 22 and 24 kg/m2, with a significant declining trend between BMI and IGF1 thereafter within both race groups (P-trend<0.05). In contrast, IGF2 levels tended to increase somewhat with BMI with AA and white women, while IGFBP3 levels rose slightly (not significantly) as BMI increased beyond 20 kg/m2. Free IGF1 levels across BMI categories followed the pattern established by IGF1 for both AA and white women, with the highest free IGF1 levels among AA women with a BMI between 22 and 24 kg/m2, and then significantly decreasing with a greater BMI. Further investigation of the association between IGF1 or free IGF1 with BMI found a similar pattern within premenopausal or postmenopausal women ( Table 4), although IGF1 levels were not lower among the six AA premenopausal women with a BMI <20 kg/m2.
Relationship between mean insulin-like growth factor (IGF) levels and body mass index (BMI) or BMI at age 21 years by race
| n | IGF1 (ng/ml) | IGF2 (ng/ml) | IGFBP3 (ng/ml) | Free IGF1 | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| AA | W | AA | W | AA | W | AA | W | AA | W | ||
| BMI | 18–19 | 19 | 35 | 133.3 | 114.1 | 1526.9 | 1638.4 | 3448.7 | 3445.4 | 0.139 | 0.119 |
| 20–24 | 196 | 169 | 154.4 | 145.6 | 1616.1 | 1748.1 | 3689.7 | 3821.1 | 0.151 | 0.137 | |
| 25–29 | 205 | 211 | 146.3 | 141.4 | 1641.3 | 1837.6 | 3649.9 | 4058.2 | 0.146 | 0.127 | |
| 30–34 | 209 | 205 | 149.9 | 131.9 | 1676.0 | 1783.7 | 3818.3 | 3897.4 | 0.144 | 0.123 | |
| 35–39 | 125 | 114 | 140.8 | 131.0 | 1736.6 | 1813.5 | 3858.6 | 3937.5 | 0.134 | 0.122 | |
| 40–45 | 67 | 82 | 134.0 | 114.9 | 1709.7 | 1856.2 | 3794.1 | 3918.2 | 0.129 | 0.106 | |
| P-trend | <0.01 | <0.01 | 0.12 | 0.05 | 0.61 | 0.34 | <0.01 | <0.01 | |||
| BMI age 21 | 12–19 | 268 | 283 | 147.0 | 144.5 | 1673.6 | 1881.8 | 3766.2 | 4174.7 | 0.144 | 0.126 |
| 20–24 | 370 | 366 | 147.9 | 133.7 | 1679.6 | 1791.0 | 3780.2 | 3843.5 | 0.143 | 0.126 | |
| 25–29 | 125 | 103 | 146.1 | 126.3 | 1701.3 | 1724.9 | 3759.9 | 3742.0 | 0.141 | 0.121 | |
| 30–45 | 58 | 64 | 149.1 | 119.3 | 1551.2 | 1697.4 | 3609.3 | 3735.4 | 0.148 | 0.115 | |
| P-trend | 0.59 | <0.01 | 0.63 | <0.01 | 0.99 | <0.01 | 0.89 | 0.06 | |||
IGF levels adjusted for age, height, menopausal status, and BMI or BMI at age 21 years. AA, African–American; W, white.
IGF levels by BMI and race. Test for trend (African–American and white; respectively, where BMI>22): IGF1: P<0.01, P<0.01; IGF2: P=0.15, P=0.06; IGFBP3: P=0.66, P=0.40; free IGF1: P<0.01, P<0.01, adjusted for age, height, BMI at age 21 years, and menopausal status.
Citation: Endocrine-Related Cancer 17, 1; 10.1677/ERC-09-0023
Insulin-like growth factor 1 (IGF1) and free-IGF levels and body mass index (BMI) by race and menopausal status
| IGF1 (ng/ml) | Free IGF1 | |||||||
|---|---|---|---|---|---|---|---|---|
| Premenopausal | Postmenopausal | Premenopausal | Postmenopausal | |||||
| AA | W | AA | W | AA | W | AA | W | |
| BMI | ||||||||
| 18–19 | 166.8 | 105.6 | 132.2 | 107.2 | 0.138 | 0.112 | 0.139 | 0.116 |
| 20–24 | 164.7 | 145.5 | 169.2 | 134.1 | 0.160 | 0.136 | 0.148 | 0.127 |
| 25–29 | 162.5 | 139.7 | 146.3 | 133.6 | 0.156 | 0.123 | 0.142 | 0.119 |
| 30–34 | 164.1 | 136.3 | 146.8 | 120.4 | 0.150 | 0.122 | 0.143 | 0.113 |
| 35–39 | 154.7 | 128.9 | 132.7 | 131.9 | 0.142 | 0.109 | 0.128 | 0.127 |
| 40–45 | 139.9 | 116.3 | 134.1 | 110.4 | 0.132 | 0.098 | 0.129 | 0.107 |
| BMI age 21 years | ||||||||
| 12–19 | 155.8 | 140.0 | 143.2 | 142.2 | 0.143 | 0.117 | 0.142 | 0.126 |
| 20–24 | 159.7 | 133.0 | 145.8 | 125.2 | 0.146 | 0.122 | 0.140 | 0.121 |
| 25–29 | 158.9 | 121.8 | 141.3 | 117.9 | 0.147 | 0.114 | 0.130 | 0.117 |
| 30–45 | 160.8 | 120.1 | 143.9 | 106.3 | 0.150 | 0.113 | 0.140 | 0.108 |
Controlling for age, height, and BMI or BMI at age 21 years. Postmenopausal analysis also controlled for the reason for menopause (surgical or natural). AA, African–American; W, white.
BMI at age 21 years and height provide two metrics of body size during growth and development. BMI at age 21 years was inversely associated with IGF1, IGF2, and IGFBP3, but only among white women ( Table 3). Thus, with increasing BMI at age 21 years, racial differences in IGF2 and IGFBP3 diminished, while racial differences in IGF1 significantly increased (P-interaction=0.01). Racial differences in free IGF1 remained fairly constant across the range of BMI at age 21 years ( Fig. 2). Inverse associations between IGF1 and BMI at age 21 years were consistent within premenopausal and postmenopausal white women ( Table 4), and similar trends were observed with IGF2 or IGFBP3 (not shown). Height was not significantly associated with adjusted IGF levels (not shown).
IGF levels by BMI at age 21 years and race. Test for trend (African–American and white; respectively): IGF1: P=0.33, P=0.81; IGF2: P=0.38, P=0.71; IGFBP3: P=0.18, P=0.83; free IGF1: P=0.98, P=0.53, adjusted for age, BMI, menopausal status, and BMI at age 21 years.
Citation: Endocrine-Related Cancer 17, 1; 10.1677/ERC-09-0023
Discussion
AA women are at a greater risk for premenopausal breast cancer and ER−/PR− tumors compared with white women ( Hausauer et al. 2007, Brinton et al. 2008). We found that AA women had significantly higher IGF1 and free IGF1 levels and lower IGF2 and IGFBP3 levels, compared with white women. Any explanation is speculative but may involve racial/ethnic differences in estrogen ( Pinheiro et al. 2005, Setiawan et al. 2006) or vitamin D levels ( Egan et al. 2008) that may in turn affect IGF levels ( Rozen et al. 1997, Janssen et al. 1998, Jorgensen et al. 2004, Lukanova et al. 2004). While beyond the scope of this manuscript, differences in diet, physical activity, or reproductive history need to be considered in future analyses ( Holmes et al. 2002, Probst-Hensch et al. 2003, Gapstur et al. 2004, McGreevy et al. 2007). Furthermore, perhaps one-third of circulating IGF1 and two-thirds of IGF2 and IGFBP3 levels may be attributable to genetic factors ( Harrela et al. 1996), and the differences in the prevalence of these genetic determinants between race/ethnicity groups may contribute to racial/ethnic differences in circulating IGF levels ( Jernström et al. 2001a).
IGF1 levels were sharply lower among women with a BMI <20 kg/m2, and then rose before beginning a significant inverse trend starting with a BMI >22 kg/m2. IGF1 and free IGF1 levels were lower with increasing BMI regardless of menopausal status. This is consistent with several prior analyses ( Jernstrom et al. 2001b, Lukanova et al. 2004, Gram et al. 2006), although studies of premenopausal white women have not always seen lower IGF1 levels with a low (<20 kg/m2) BMI ( Lukanova et al. 2004, Schernhammer et al. 2007). IGF1 levels are strongly affected by energy availability, and lean or fasting individuals in a state of relative energy deprivation may downregulate GH receptor levels or increased resistance to GH signaling, decreasing IGF1 synthesis ( Clemmons & Underwood 1991). This pathway may also affect IGFBP3 and IGF2, as we found the lowest IGFBP3 and IGF2 levels with a BMI<20 kg/m2. The pathophysiology linking obesity to lower GH and IGF1 levels is not well understood, but may involve a negative feedback loop by which lower GH secretion is mediated through decreased IGFBP1 with visceral fat accumulation ( Gram et al. 2006).
The MEC reported that obesity lowers IGF1 levels in white but not in AA participants ( Henderson et al. 2006). However, we found that although AA women had higher IGF1 levels, the associations between BMI and IGF levels were consistent with race/ethnicity. Two studies found that ovulatory function did not mediate the association between BMI and premenopausal breast cancer ( Michels et al. 2006, Palmer et al. 2007), and we might hypothesize that the protective association between BMI and premenopausal breast cancer risk may be attributable to the decrease in IGF1 associated with obesity and that this protective effect could generalize across race/ethnicity. However, the literature regarding BMI and premenopausal breast cancer among AAs is decidedly mixed ( Mayberry 1994, Hall et al. 2000, Zhu et al. 2005, Palmer et al. 2007). For example, Hall et al. (2000) analyzed data from the Carolina Breast Cancer Study and found that a BMI>30 kg/m2 was significantly associated with lower premenopausal breast cancer risk among whites (odds ratio (OR)=0.46, 95% confidence interval (CI) (0.26, 0.80)) but not among AAs (OR=0.89, 95% CI (0.38, 2.07)). In contrast, Zhu et al. (2005) found that BMI was associated with a nonsignificant increase in premenopausal breast cancer risk among AAs (OR=2.49, 95% CI (0.82, 7.59)). Higher overall IGF1 levels among AAs may play a role, and reconciling the possible interactions between IGF levels, obesity, and race becomes more complicated if BMI affects tissue IGF1 receptor expression and potential susceptibility to circulating IGF1 levels ( Suga et al. 2001). Biomarker assays in blood and in target tissue, in addition to body size measures, may be required.
BMI at age 21 years was inversely associated with all IGF markers levels in later life, but only among white women. The decline in IGF1 among whites and the constant levels among AAs with BMI at age 21 years resulted in considerably lower IGF1 levels for whites than AAs for those who were obese in early adulthood. The reason is not clear but appears consistent with an analysis from the NHS II showing that BMI at age 18 years was associated with reduced IGF1 levels among white premenopausal women ( Schernhammer et al. 2007). Our findings also raise the possibility that lower IGF1 status may be implicated in a premenopausal breast cancer risk reduction with obesity among whites but not among AAs. The investigation of early-life obesity and breast cancer among AAs is limited, with Zhu et al. (2005) reporting no association between BMI at age 18 years and premenopausal breast cancer, while Palmer et al. (2007) reporting that early-life BMI was associated with a reduced risk among AAs. BMI at age 21 years also appeared to decrease the racial differences in IGF2 and IGFBP3 associated with steroid hormone activity and breast cancer risk ( Figueroa et al. 1993, Gronbaek et al. 2004, Allen et al. 2005). Thus, the relevance of early-life obesity on racial/ethnic differences in IGF1 in the context of decreasing racial/ethnic differences in IGFBP3 or IGF2 remains to be investigated.
Strengths of this analysis include evaluation of a large number of AA and white women after excluding participants taking hormone replacement therapy or insulin. Furthermore, the range of BMI available for analyses extended beyond most studies. Rather than using only correlation coefficients to summarize IGF and BMI associations, we evaluated nonlinear trends graphically over BMI values and controlled for potential confounders such as menopausal status and the reason for menopause. We excluded women with any cancer treatment within the past year, and lifetime cancer history was not significantly associated with any IGF level (all P>0.56). The women evaluated were all participants in the SCCS and were generally of a similar socioeconomic status regardless of race. Hence, not only is the underlying study base population well defined, but also the differences in education, income, or other attributes, which sometimes exist in comparisons of AAs and whites were minimized by the study design.
Our study has some limitations. Although temporal inference from BMI at age 21 years to current IGF levels may be reasonable, the cross-sectional approach did not allow us to determine the temporal relationship between IGF markers and current BMI. BMI is based on self-reported data, although we have reported that self-reported BMI is correlated with blood leptin levels ( Fowke et al. 2008). IGF measurements were based on a single serum sample stored at −80 °C for up to 4 years, although IGF marker assays appear reliable when biospecimens are properly stored ( Berrigan et al. 2007). Our IGF values are similar to those reported by MEC, NHS, NHANES, and other research groups. Oral contraceptive use is associated with lower IGF1 levels ( Jernström et al. 2001a); however, we did not have data to exclude current oral contraceptive users from our analysis. If oral contraceptive use lowers IGF1 among premenopausal women ( Jernström et al. 2001a), the differences between premenopausal and postmenopausal women may be conservative. Generalization of our results to AA and white women will require confirmatory analyses in other studies. Our study also did not allow us to determine the combined effect of past and present obesity with the IGF axis on the risk of ER−/PR− breast cancer among AA women. Finding such relationship(s) requires further follow-up investigations in our cohort.
Conclusion
BMI was associated with IGF1 levels regardless of race/ethnicity, while obesity during childhood or young adulthood may have a greater impact on IGF1 levels among white women. The separate effects of obesity throughout life and the IGF axis on the risk of ER−/PR− breast cancer among AA women require investigation.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This project was supported by grant OP05-0927-DR1 from Susan G Komen for the Cure. The Southern Community Cohort Study is supported by grant R01 CA92447 from the National Cancer Institute.
Author contributions statement
J H Fowke, C E Matthews, M S Buchowski, W Zheng, and W J Blot developed the hypotheses and research approach. J H Fowke conducted the analyses and was the primary author. H Yu supervised all IGF assays. Q Cai and S Cohen managed biospecimen and data repositories. All authors provided comments on an early draft.
References
Allen NE, Roddam AW, Allen DS, Fentiman IS, Dos Santos Silva I, Peto J, Holly JM & Key TJ 2005 A prospective study of serum insulin-like growth factor-I (IGF-I), IGF-II, IGF-binding protein-3 and breast cancer risk. British Journal of Cancer 92 1283– 1287.
Berrigan D, Potischman N, Dodd KW, Nicar M, McQuillan G, Lavigne JA, Barrett JC & Ballard-Barbash R 2007 Serum levels of insulin-like growth factor-I and insulin-like growth factor-I binding protein-3: quality control for studies of stored serum. Cancer Epidemiology, Biomarkers & Prevention 16 1017– 1022.
Berrigan D, Potischman N, Dodd KW, Hursting SD, Lavigne J, Barrett JC & Ballard-Barbash R 2008 Race/ethnic variation in serum levels of IGF-I and IGFBP-3 in US adults. Growth Hormone & IGF Research 19 146– 155.
Brinton LA, Sherman ME, Carreon JD & Anderson WF 2008 Recent trends in breast cancer among younger women in the United States. Journal of the National Cancer Institute 100 1643– 1648.
Clemmons DR & Underwood LE 1991 Nutritional regulation of IGF-I and IGF binding proteins. Annual Review of Nutrition 11 393– 412.
DeLellis K, Rinaldi S, Kaaks RJ, Kolonel LN, Henderson B & Le Marchand L 2004 Dietary and lifestyle correlates of plasma insulin-like growth factor-I (IGF-I) and IGF binding protein-3 (IGFBP-3): the Multiethnic Cohort. Cancer Epidemiology, Biomarkers & Prevention 13 1444– 1451.
Egan KM, Signorello LB, Munro HM, Hargreaves MK, Hollis BW & Blot WJ 2008 Vitamin D insufficiency among African–Americans in the Southeastern United States: implications for cancer disparities (United States). Cancer Causes & Control 19 527– 535.
Figueroa JA, Jackson JG, McGuire WL, Krywicki RF & Yee D 1993 Expression of insulin-like growth factor binding proteins in human breast cancer correlates with estrogen receptor status. Journal of Cellular Biochemistry 52 196– 205.
Flegal KM, Carroll MD, Ogden CL & Johnson CL 2002 Prevalence and trends in obesity among US adults, 1999–2000. Journal of the American Medical Association 288 1723– 1727.
Fowke JH, Matthews CM, Buchowski MS, Signorello LB, Chang SS, Cookson MS & Blot WJ 2008 Association between prostate-specific antigen and leptin, adiponectin, HbA1c or C-peptide among African–American and Caucasian men. Prostate Cancer and Prostatic Diseases 11 264– 269.
Gapstur SM, Kopp P, Chiu BCH, Gann PH, Colangelo LA & Liu K 2004 Longitudinal associations of age, anthropometric and lifestyle factors with serum total insulin-like growth factor-I and IGF binding protein-3 levels in black and white men: the CARDIA Male Hormone Study. Cancer Epidemiology, Biomarkers & Prevention 13 2208– 2216.
Gennigens C, Menetrier-Caux C & Droz JP 2006 Insulin-like growth factor (IGF) family and prostate cancer. Critical Reviews in Oncology/Hematology 58 124– 145.
Gram IT, Norat T, Rinaldi S, Dossus L, Lukanova A, Tehard B, Clavel-Chapelon F, van Gils CH, van Noord PA & Peeters PH et al. 2006 Body mass index, waist circumference and waist–hip ratio and serum levels of IGF-I and IGFBP-3 in European women. International Journal of Obesity 30 1623– 1631.
Gronbaek H, Flyvbjerg A, Mellemkjaer L, Tjonneland A, Christensen J, Sorensen HT & Overvad K 2004 Serum insulin-like growth factors, insulin-like growth factor binding proteins, and breast cancer risk in postmenopausal women. Cancer Epidemiology, Biomarkers & Prevention 13 1759– 1764.
Hall IJ, Newman B, Millikan RC & Moorman PG 2000 Body size and breast cancer risk in black women and white women: the Carolina Breast Cancer Study. American Journal of Epidemiology 151 754– 764.
Harrela M, Koistinen H, Kaprio J, Lehtovirta M, Tuomilehto J, Eriksson J, Toivanen L, Koskenvuo M, Leinonen P & Koistinen R et al. 1996 Genetic and environmental components of interindividual variation in circulating levels of IGF-I, IGF-II, IGFBP-1, and IGFBP-3. Journal of Clinical Investigation 98 2612– 2615.
Hausauer AK, Keegan TH, Chang ET & Clarke CA 2007 Recent breast cancer trends among Asian/Pacific Islander, Hispanic, and African–American women in the US: changes by tumor subtype. Breast Cancer Research 9 R90.
Henderson KD, Goran MI, Kolonel LN, Henderson BE & Le ML 2006 Ethnic disparity in the relationship between obesity and plasma insulin-like growth factors: the Multiethnic Cohort. Cancer Epidemiology, Biomarkers & Prevention 15 2298– 2302.
Holmes MD, Pollak MN, Willett WC & Hankinson SE 2002 Dietary correlates of plasma insulin-like growth factor I and insulin-like growth factor binding protein 3 concentrations. Cancer Epidemiology, Biomarkers & Prevention 11 852– 861.
Janssen JA, Stolk RP, Pols HA, Grobbee DE, de Jong FH & Lamberts SW 1998 Serum free IGF-I, total IGF-I, IGFBP-1 and IGFBP-3 levels in an elderly population: relation to age and sex steroid levels. Clinical Endocrinology 48 471– 478.
Jernström H, Chu W, Vesprini D, Tao Y, Majeed N, Deal C, Pollak M & Narod SA 2001a Genetic factors related to racial variation in plasma levels of insulin-like growth factor-1: implications for premenopausal breast cancer risk. Molecular Genetics and Metabolism 72 144– 154.
Jernstrom H, Deal C, Wilkin F, Chu W, Tao Y, Majeed N, Hudson T, Narod SA & Pollak M 2001b Genetic and nongenetic factors associated with variation of plasma levels of insulin-like growth factor-I and insulin-like growth factor-binding protein-3 in healthy premenopausal women. Cancer Epidemiology, Biomarkers & Prevention 10 377– 384.
Jorgensen JO, Christensen JJ, Krag M, Fisker S, Ovesen P & Christiansen JS 2004 Serum insulin-like growth factor I levels in growth hormone-deficient adults: influence of sex steroids. Hormone Research 62 73– 76.
Lukanova A, Lundin E, Zeleniuch-Jacquotte A, Muti P, Mure A, Rinaldi S, Dossus L, Micheli A, Arslan A & Lenner P et al. 2004 Body mass index, circulating levels of sex-steroid hormones, IGF-I and IGF-binding protein-3: a cross-sectional study in healthy women. European Journal of Endocrinology 150 161– 171.
Mayberry RM 1994 Age-specific patterns of association between breast cancer and risk factors in black women, ages 20 to 39 and 40 to 54. Annals of Epidemiology 4 205– 213.
McGreevy K, Hoel B, Lipsitz S & Hoel D 2007 Impact of nutreints on insulin-like growth factor-I, insulin-like growth factor binding protein-3, and their ratio in African American and white males. Public Health Nutrition 10 97– 105.
Michels KB, Terry KL & Willett WC2006 2006 Longitudinal study on the role of body size in premenopausal breast cancer. Archives of Internal Medicine 166 2395– 2402.
Palmer JR, dams-Campbell LL, Boggs DA, Wise LA & Rosenberg L 2007 A prospective study of body size and breast cancer in black women. Cancer Epidemiology, Biomarkers & Prevention 16 1795– 1802.
Pfeiffer RM, Mitani A, Matsuno RK & Anderson WF 2008 Racial differences in breast cancer trends in the United States (2000–2004). Journal of the National Cancer Institute 100 751– 752.
Pinheiro SP, Holmes MD, Pollak MN, Barbieri RL & Hankinson SE 2005 Racial differences in premenopausal endogenous hormones. Cancer Epidemiology, Biomarkers & Prevention 14 2147– 2153.
Probst-Hensch NM, Wang H, Goh VHH, Seow A, Lee HP & Yu MC 2003 Determinants of circulating insulin-like growth factor I and insulin-like growth factor binding protein 3 concentrations in a cohort of Singapore men and women. Cancer Epidemiology, Biomarkers & Prevention 12 739– 746.
Renehan AG, Harvie M & Howell A 2006 Insulin-like growth factor (IGF)-I, IGF binding protein-3, and breast cancer risk: eight years on. Endocrine-Related Cancer 13 273– 278.
Rozen F, Yang XF, Huynh H & Pollak M 1997 Antiproliferative action of vitamin D-related compounds and insulin-like growth factor-binding protein 5 accumulation. Journal of the National Cancer Institute 89 652– 656.
Schernhammer ES, Tworoger SS, Eliassen AH, Missmer SA, Holly JM, Pollak MN & Hankinson SE 2007 Body shape throughout life and correlations with IGFs and GH. Endocrine-Related Cancer 14 721– 732.
Setiawan VW, Haiman CA, Stanczyk FZ, Le Marchand L & Henderson BE 2006 Racial/ethnic differences in postmenopausal endogenous hormones: the Multiethnic Cohort Study. Cancer Epidemiology, Biomarkers & Prevention 15 1849– 1855.
Signorello LB, Hargreaves MK, Steinwandel MD, Zheng W, Cai Q, Schlundt DG, Buchowski MS, Arnold CW, McLaughlin JK & Blot WJ 2005 Southern community cohort study: establishing a cohort to investigate health disparities. Journal of the National Medical Association 97 972– 979.
Suga K, Imai K, Eguchi H, Hayashi S, Higashi Y & Nakachi K 2001 Molecular significance of excess body weight in postmenopausal breast cancer patients, in relation to expression of insulin-like growth factor I receptor and insulin-like growth factor II genes. Japanese Journal of Cancer Research 92 127– 134.
Zhu K, Caulfield J, Hunter S, Roland CL, Payne-Wilks K & Texter L 2005 Body mass index and breast cancer risk in African American women. Annals of Epidemiology 15 123– 128.
