Changes in inflammatory endometrial cancer risk biomarkers in individuals undergoing surgical weight loss

Abstract

Objective.

Obesity has been strongly linked to endometrial cancer (EC) risk. A number of potential EC risk bio-markers have been proposed, including heightened pro-inflammatory cytokines and adipokines. To evaluate if bariatric surgery can serve as a means for altering levels of such EC risk biomarkers, we investigated changes in these biomarkers after weight loss.

Methods.

Blood samples were collected pre-operatively and 6 months post-operatively in107 female bariatric surgery patients aged 18–72 years. Wilcoxon signed-rank tests were used to compare biomarker levels (measured using xMAP immunoassays) pre- and post-surgery. Normative comparisons were implemented to contrast 6-month post-surgery biomarker levels to levels in a sample of 74 age-matched non-obese women. Linear regression was used to evaluate the relationship between biomarker expression at baseline and 6 months post-surgery and the relationship between race and biomarker levels.

Results.

On average, participants lost 30.15 kg (SD: 12.26) after the bariatric intervention. Levels of C-peptide, insulin, CRP, leptin, IL-1Rα, and IL-6 significantly decreased, while levels of SHBG, IGFBP1, and adiponectin significantly increased with weight loss. Normative comparisons showed the levels of SHBG, C-peptide, insulin, IGFBP1, adiponectin, CRP, and TNFα after bariatric intervention approached the level of markers in comparison group. Multiple regression analyses revealed significant relationships between changes in BMI and changes in biomarker levels. The changes in IL-1Rα were significantly associated with race.

Conclusions.

Our findings demonstrate that normalization of EC risk biomarkers can be achieved with bariatric surgery. Improved understanding of biological mechanisms associated with weight loss may inform preventive strategies for EC.

1. Introduction

Endometrial cancer (EC) is the most common gynecologic malignancy among American women, and has been gradually increasing in incidence in recent years, with approximately 61,380 new diagnoses and 10,920 deaths expected in 2017 [1]. A recent publication from our group estimates a 55% increase in the incidence of EC by 2030 [2]. Although multiple factors are involved, increasing rates of obesity are thought to be the primary driver of increasing EC incidence [3–5]. Prospective studies indicate that EC risk increases 1.6-fold with each additional 5 kg/m² in body mass index (BMI), reaching 9.1-fold higher risk at 42 kg/m² [6]. In a recent publication, we indicated that increasing BMI is also associated with a greater risk of endometrial pathology among women with severe obesity (>35 BMI) [7]. As of 2016, no systemic biomarker (or panel of markers) is available to identify women at high risk of precancerous changes, at a time when preventive interventions such as weight loss or hormone therapy may still be possible.

Accumulating evidence from preclinical research, as well as prospective studies exploring associations between biomarker levels in peripheral blood and the development of EC, strongly implicates three basic biological pathways: pro-inflammatory factors, insulin resistance/metabolic factors, and steroid hormones [8–13]. Obesity is associated with a physiological state of chronic, low-grade inflammation, characterized by elevated systemic levels of circulating inflammatory biomarkers mediating, at least in part, the association between obesity and risk of EC [12,14,15]. Increased adipose tissue mass may contribute to the development of cancer via increased secretion of pro-inflammatory cytokines and chemokines [16,17]. A recent study found CRP, an acute-phase reactant protein that can influence production of inflammatory cytokines, to be positively associated with EC risk [13]. CRP, IL-6, and IL-1Rα have been implicated in EC risk in several prospective investigations [11–13,16]. Circulating adipokines (small protein molecules produced and secreted by white adipose tissue), such as adiponectin, have systemic immunomodulating effects that also play a major role in the development of several cancers [18]. Insulin, IGFBP2, leptin, adiponectin, and C-peptide have been implicated in EC development in prospective studies [11,13,19,20]. While a very limited number of publications have explored the biomarkers associated with endometrial hyperplasia (EC precursor lesion [21 ]), it is likely that the development of hyperplasia and EC are associated with abnormal activity of similar inflammatory, hormonal, and metabolic pathways. One such hormone is leptin, which has been found to be elevated in patients with both endometrial cancer and hyperplasia in comparison with pathology free controls [22].

Emerging literature suggests that the risk of EC may be particularly responsive to weight loss [3,4,23]. In a large-scale study, Ward et al. recently demonstrated that bariatric surgery is associated with a 71% reduction in risk for uterine malignancy [24]. Similar evidence has been recently published by the Swedish Obese Subject Study, reporting that bariatric surgery lowered the incidence of EC in the bariatric surgery group [25].

The study described in this manuscript was based on the idea that biomarkers in EC risk pathways will be ameliorated by weight loss among women with severe obesity. Our group previously reported that behavioral intervention for weight loss is associated with changes in adiponectin[26]. In this manuscript, we report on the modification of EC risk biomarkers with surgically induced weight loss, an area poorly investigated in existing studies. This study aims to fill an important gap by analyzing biomarkers associated with EC risk in women undergoing weight loss through bariatric surgery and comparing the levels of markers post-bariatric surgery with levels of the same markers in non-obese women.

2. Methods

2.1. Participants and settings

This subanalysis included 107 women aged 18 to 72 (mean age 43.88 years (SD: 11.66 years)) who were participating in the “Effect of weight loss on biomarkers of immunity and inflammation” or Bariatric Marker (BAM) Study at Magee-Womens Hospital which had the goal of determining the extent to which women undergoing bariatric surgery, and its subsequent weight loss, had an overall improvement in inflammatory and endocrine biomarker status. Inclusion criteria for this study included: female, age 18+ years, BMI ≥ 35, approved and scheduled for bariatric surgery (Roux-en-Y gastric bypass, laparoscopic adjustable gastric banding, or sleeve gastrectomy), and life expectancy >3 years. Exclusion criteria included: refusal to sign informed consent, unable to attend study visits, plans to become pregnant within one year after surgery, presence of severe inflammatory disease, previous history of cancer (including gynecologic cancer), recent injury or surgery, and plans to move residence within one year.

Seventy-four female non-obese (BMI < 30) comparison participants aged 21 to 75 (mean age 42.04 years (SD: 13.53 years)) were recruited through a general recruitment campaign at the University of Pittsburgh campus. Comparison participants were matched to bariatric surgery patients on the basis of age. The University of Pittsburgh Human Research Protection Office approved this study. All participants signed informed consent documents. Exclusion criteria for comparisons were inability to sign informed consent, obesity, presence of severe inflammatory disease, previous history of cancer (other than non-melanoma skin cancer), and recent injury or surgery.

At each visit (pre- and 6 months post-operatively), participants in the bariatric surgery group completed a set of validated general health questionnaires, and anthropometric measurements and blood samples were obtained. The same procedure was performed at a single time point with participants in the comparison group. Registered research nurses and staff conducted study procedures in the Clinical and Translational Research Center (CTRC) at Magee-Womens Hospital of UPMC.

2.2. Measures

Anthropometric measurements for participants in both groups were obtained in the CTRC by research staff. Height was measured in centimeters using a wall-mounted stadiometer. Waist and hip circumference was measured in centimeters using a tape measure. Weight (kilograms) and BMI (kg/m²) were obtained from the Tanita body composition analyzer (Model TBF-310, Tanita Corporation of America) with participants wearing light clothing and no footwear.

Reproductive history, menstrual history, and history of hormone use (hormone therapy, birth control, fertility drugs) were obtained from the Reproductive Health Baseline (RHB) and the Screening Questionnaire for General Health History (SQHH). The RHB was used in the Longitudinal Assessment of Bariatric Surgery-2 Study (LABS) to collect information on the status of the reproductive health of women undergoing bariatric surgery [27]. The SQHH has been used in the Paving the road to everlasting food and exercise routines (PREFER) and Self-monitoring and recording using technology (SMART) studies, which tested several methods of behavioral weight management [28,29].

2.3. Statistical analysis

Basic descriptive statistics were used to summarize the characteristics of the bariatric surgery population at baseline and the 6-month post-operative visit. Continuous variables are reported as mean (standard deviation) and categorical variables are reported as N (%). Age, BMI, weight, waist circumference (WC), and waist-to-hip ratio (WHR) were analyzed as continuous variables. Race (dichotomized as European American (EA) or African American (AA)) was analyzed as a categorical variable. The time points were categorized as baseline and 6 months post-surgery.

Wilcoxon signed-rank tests were used to compare the mean levels of biomarkers of the participants in the bariatric surgery group at the baseline and 6-month post-operative visits. The Wilcoxon rank-sum test was used to compare mean biomarker levels of the participants in the bariatric surgery group at each time point and in the comparison group to determine if there were significant differences in mean levels between the groups. Fisher’s exact test was used to compare categorical variables of participants in the bariatric surgery group at baseline and in the comparison group to determine if there were significant differences between the groups.

Biomarker levels were log transformed to reduce skewness. Normative comparisons are a set of procedures that determine the clinical significance of an intervention used to treat participants compared to the untreated normal participants [30]. If the difference between means of the treated participants and normal participants falls in a specified range of closeness that centers on zero, the two groups are considered clinically equivalent. This range is bounded by a lower limit, negative delta ( − δ), and an upper limit, positive delta ( + δ). For our paper, we chose a range of closeness of one standard deviation of the normative mean for each biomarker level, which is consistent with the previously published article by Kendall et al. [30]. Thus, if the difference is within this range, we conclude that the two groups are clinically equivalent. Next, two one-sided t-tests were done to test the null that the difference of the means is less than the lower limit and greater than the upper limit, respectively. A statistically significant result for both tests indicated that the difference between the means was within the range of closeness and the post-treatment group and the normal group are clinically equivalent. Then, the means of the post-treatment group and the normal group were compared using a traditional two-sided hypothesis test [30] (e.g. two-sided t-test), where the null hypothesis was that the two means are equal. In the final step, the results from the traditional hypothesis test and the clinical equivalency test were compared via a 2 × 2 table adapted from Kendall et al. [30]. The overall goal of this approach was to demonstrate if weight loss resulted in normalization of the biomarker levels of the bariatric surgery participants.

In a secondary analysis, we looked at the relationships between race and biomarker levels at baseline and 6 months post-bariatric surgery. Linear regression was used to evaluate the relationship between biomarker expression from baseline to 6 months and the biomarker levels during the same time period, controlling for baseline BMI, hormone use, diabetes status, and race. An interaction between BMI and race was examined to assess the relationship between biomarker expression, the change in BMI, and race. All statistical analyses were done with SAS version 9.4 (SAS Institute Inc., Cary, NC). α level was set at 0.01 to adjust for multiple comparisons and was two-sided.

3. Results

Table 1 lists the basic characteristics of the bariatric surgery group and the comparison group at baseline. Mean age of the bariatric surgery group participants was 43.88 years (SD: 11.66), mean weight was 123.88 kg (SD: 19.71), and mean BMI was 45.52 kg/m² (SD: 6.19). Bariatric surgery group and control group were similar in terms of racial and age distribution.

Table 1.

Demographic and personal characteristics of bariatric surgery candidates at pre-operative visit and comparison participants.

	Bariatric surgery n = 107	Comparisons n = 74	p-Value
Age (years), mean (SD)	43.88 (11.66)	42.04(13.53)	0.2102*
Weight (kg), mean (SD)	123.88 (19.71)	60.01 (11.36)	<0.0001*
BMI (kg/m2), mean (SD)	45.52 (6.19)	22.61 (2.87)	<0.0001*
Race, N (%)			0.8385**
European American	82 (76.64)	60 (81.08)
African American	25 (23.36)	14 (18.92)
Hormone Use (N = 105)
Yes	17 (16.19%)
No	88 (83.81%)
Diabetes (N = 100)
Yes	15 (15.0%)
No	85 (85.0%)

^*p-Value from Wilcoxon rank sum test.

^**p-Value from Fisher’s exact test.

Bariatric surgery group participants lost an average of 30.15 kg (SD: 12.26) from baseline to the 6-month post-operative visit. BMI decreased an average of 11.03 kg/m² (SD: 4.67), WC decreased by 21.10 cm (SD: 11.37), and the WHR decreased from a mean of 0.88 (SD: 0.07) to a mean of 0.86 (SD: 0.08) in the 6-month post-operative period. The decreases in weight, BMI, WC, and WHR from baseline to 6 months were statistically significant (all p < 0.01).

Fig. 1 shows the mean levels of the biomarkers for the 6-month visit for the bariatric surgery group and the comparison group. SHBG, adiponectin, IGFBP1, and IGFBP2 significantly increased (p-values < 0.01) from the baseline to the 6-month visit, and C-peptide, insulin, CRP, leptin, IL-1Rα, and IL-6 significantly decreased from the baseline to the 6-month visit (p-values < 0.01). When compared to biomarker levels in the comparison group, SHBG, IGFBP1, and adiponectin were significantly lower than comparison participants at both the baseline and 6-month visits, and C-peptide, insulin, CRP, resistin, leptin, IL-1Rα, IL-6, and TNFα were significantly higher in the bariatric surgery group than the levels of the comparison group at both visits. IGFBP2 was lower at baseline in the bariatric surgery group, but the levels increased significantly between baseline and the 6-month postoperative visit to levels significantly higher than was observed among comparison group participants.

Fig. 1.

Biomarker levels over time in bariatric surgery patients and in non-obese comparison participants (log transformed values).

From the normative comparisons, SHBG, C-peptide, IGFBP1, adiponectin, CRP, and TNFα levels in the bariatric surgery group were statistically different, but clinically equivalent to the comparison group at the 6-month post-operative visit (data not shown). IGFBP2, resistin, leptin, IL-1Rα, and IL-6 were clinically different from the comparison group at the 6-month post-operative visit. Insulin was clinically equivalent to the comparison group at the 6-month post-operative visit.

In the analysis by race for the surgery group, at baseline (data not shown), AA women had lower levels of potential EC risk biomarkers than EA women; however, only IGFBP1 and adiponectin were significantly lower in AA women (p < 0.01). At 6 months (data not shown), IGFBP1,IGFBP2, and IL-1 Rα were significantly lower in AA women (p < 0.01). From the linear regression models, the change in IL-1Rα was significantly associated with race (p-value = 0.0009). Hormone use and diabetes status were not significantly associated with any of the biomarkers. Table 2 details the results of the linear regression.

Table 2.

The linear regression equation: Δbiomarker level = age + race + baseline biomarker + baseline BMI + ΔBMI.

Change (Δ)in biomarker levels	Age	Race	Baseline biomarker level	Baseline BMI	ΔBMI
ΔSHBG	–0.006	– 0.297	0.374	0.017	−0.041
	(0.26)*	(0.06)	(0.002)	(0.16)	(0.01)
ΔC-peptide	–0.002	0.114	0.788	– 0.006	0.025
	(0.57)	(0.30)	(<0.0001)	(0.49)	(0.03)
ΔInsulin	0.001	– 0.043	0.794	–0.010	0.041
	(0.85)	(0.81)	(<0.0001)	(0.43)	(0.02)
ΔIGFBP1	–0.010	0.047	0.620	0.036	− 0.078
	(0.14)	(0.79)	(<0.0001)	(0.006)	(<0.0001)
ΔIGFBP2	–0.014	0.228	– 0.401	0.044	− 0.132
	(0.03)	(0.19)	(0.22)	(0.001)	(<0.0001)
ΔAdiponectin	– 0.004	– 0.039	0.227	0.014	− 0.035
	(0.16)	(0.59)	(0.01)	(0.008)	(<0.0001)
ΔCRP	– 0.005	0.446	0.226	– 0.056	0.061
	(0.48)	(0.02)	(0.18)	(0.0001)	(0.002)
ΔResistin	0.0004	0.023	0.294	–0.013	0.0003
	(0.84)	(0.65)	(<0.0001)	(0.0006)	(0.95)
ΔLeptin	0.001	–0.121	0.518	– 0.059	0.124
	(0.73)	(0.17)	(<0.0001)	(<0.0001 )	(< 0.0001)
ΔIL-1Rα	0.004	0.360	0.458	–0.016	0.023
	(0.27)	(0.0009)	(< 0.0001)	(0.04)	(0.03)
ΔIL-6	–0.0003	– 0.093	0.342	– 0.007	0.020
	(0.94)	(0.34)	(<0.0001)	(0.36)	(0.04)
ΔTNFα	– 0.0005	0.141	0.315	– 0.006	0.002
	(0.88)	(0.11)	(<0.0001)	(0.39)	(0.86)

^*p-Values are in parentheses.

4. Discussion

In this longitudinal analysis of biomarker changes occurring with weight loss, we found that the mean levels of the following biomarkers in the bariatric surgery group after 6 months of weight loss approach normative levels in a comparison group: SHBG, C-peptide, insulin, IGFBP1, adiponectin, CRP, and TNFα. The increase in adiponectin reported with weight loss is consistent with findings reported by Arita [31]. Findings are also consistent with previously published literature on biomarker changes occurring with weight loss [26,32].

This is one of the first studies to evaluate changes in EC risk biomarkers in a group of women undergoing bariatric surgery and to demonstrate that these biomarkers trend toward normalization with weight loss. The idea of preventing EC by weight loss is an attractive idea, as weight loss would also improve cardiovascular fitness, reduce Type 2 diabetes, and possibly reduce the risk of other obesity-related cancers [3]. The possibility of preventing EC through weight loss is particularly important to investigate for AA women, who bear a disproportionate burden of EC mortality [33]. Our findings that race is associated with changes in several EC risk biomarkers is interesting in the context of our previous publication suggestion differential expression of EC-associated biomarkers in AA in comparison to EA women [34]. Bariatric surgery patients would be an ideal group in which to investigate EC prevention through weight loss because they exhibit a very substantial and rapid weight loss.

Although women can lose weight through other means, such as diet and exercise, the literature suggests that diet and exercise programs for individuals with severe obesity result in sustainable weight loss in <20% of program participants [35], while bariatric surgery results in sustainable weight loss in over 80% of patients [36]. In a prospective Swedish study of a cohort of bariatric surgery patients and matched comparisons, sustained weight loss was associated with a 38% reduction of cancer incidence in women [37]. In another prospective study, a 24% reduction in incident cancers was seen among 6596 women who underwent bariatric surgery, in comparison with controls, over a 12.5-year follow-up period [37]. The most impressive reduction in cancer risk was seen for EC with a seven-fold risk reduction [37]. Taken together, these studies strongly suggest that EC is preventable through bariatric surgery-induced weight loss [3]. Our results raise the possibility that these effects may be mediated by changes in EC risk biomarker levels.

Key limitations of this study include a relatively small sample size and a short follow-up period; however, the results of this initial study will help to form future research questions and research studies in the field of EC prevention. Additionally, future investigations will include more than one follow-up point and be expanded to multiple ethnic and racial groups. While this study focused on EC risk biomarkers, our future studies will also evaluate similar biomarkers in the context of confirmed cancer, as similar biomarkers have are associated with EC development, progression, and prognosis [38]. Furthermore, our future research will focus on exploring endometrial tissues of women at high risk, as our previous studies indicated that bariatric patients may have unrecognized endometrial pathologies that may resolve following bariatric intervention and may be associated with normalization of endometrial tissue biomarkers [7,39–41].

EC risk has been associated with a number of factors in addition to obesity, including diabetes, hormone use, and smoking history [42]. These variables, as well as depressive symptoms and quality of life, have been implicated in altered levels of several cytokines [43] and have been positively associated with inflammatory markers such as CRP and IL-6, with BMI implicated as a mediating/moderating factor [44]. Thus, additional contributions of these factors will need to be assessed in conjunction with further investigation of inflammatory cytokines, metabolic, and hormonal factors in future investigations. Results of this study will have important implications for weight loss-based EC prevention programs and alternative treatments for women with high EC risk due to obesity. This concept is consistent with EC prevention targets identified by Hursting [45], which include host factors such as inflammatory-related molecules, insulin, leptin, adiponectin, and other cytokines evaluated in conjunction with systems level look at EC prevention targets.

HIGHLIGHTS.

• Endometrial cancer associated biomarker levels decreased after bariatric surgery.

• The change in CRP and IL-1Rα, and SHBG were associated with race.

• After surgery biomarker levels in surgery group were equivalent to healthy controls.