Lipopolysaccharide-binding protein, associated factors, and colorectal cancer

Jessica Citronberg | 2016

Advisor: Johanna W. Lampe

Research Area(s): Cancer Epidemiology, Clinical Epidemiology, Epidemiologic Methods, Pharmaco-epidemiology


BACKGROUND – Recent research has pointed to a possible key role of gut microbial communities (GMC) in understanding the link between obesity, chronic inflammation, and the development of colorectal cancer (CRC) [1]. While some studies have demonstrated a strong association between obesity and circulating levels of lipopolysaccharide (LPS) [2-5], an endotoxin which is found on the outer cell wall of gram negative bacteria, the specific GMC associated with LPS concentrations under non-sepsis conditions remain unknown. LPS increases inflammatory response signaling, and may play a role in the pathogenesis of several adverse outcomes, including diabetes [4, 6], inflammatory bowel disease [7-9], cardiovascular disease [10-12] and cancer [13, 14]. However, no human studies have examined the relationship between lipopolysaccharide-binding protein (LBP; a marker of LPS exposure) levels and CRC risk. Thus, the three studies within this dissertation aim to describe the association between LBP, the gut microbiome, and CRC risk as well as the reliability of LBP as a biomarker [1]. METHODS – We used several different studies to address the dissertation aims. We analyzed the temporal reliability of LBP measured in archived samples from participants in two studies (Aim 1). In Study 1, 60 healthy participants (30 men and 30 women, aged 60-72 years) were recruited to have blood drawn at two time points: baseline and follow-up (either 3, 6, or 9 months) and evenly distributed as much as possible by time interval. In Study 2, we tested 24 individuals (8 men and 16 women, aged 20 to 40 years) with blood drawn 3-4 times over a 7-month period. LBP measured in archived plasma by ELISA was to evaluate within-person reproducibility over time. We evaluated the association of the GMC in stool, serum CRP concentrations, and adiposity with plasma LBP concentrations in 110 premenopausal (ages 40–45 years) women in the United States (Aim 2). Multivariable linear regression analysis was used to assess the relation between LBP concentrations and adiposity. Structural equation modelling (SEM) was used to investigate the indirect effect of circulating LBP concentrations on the association between adiposity and CRP. We examined the association between colorectal cancer (CRC) and LBP in 1,638 participants (819 CRC cases and 819 controls, matched on age, sex, race, location, and time of blood draw, among other factors) from the Multiethnic Cohort study (MEC) (Aim 3). Conditional logistic regression models were used to estimate the multivariable-adjusted odds ratios (OR) and 95% confidence intervals (95% CI). RESULTS – In Aim 1 Plasma LBP concentrations showed low to moderate reliability in both Study 1 (ICC: 0.60, 95% CI: 0.43 to 0.75) and Study 2 (ICC: 0.46, 95% CI: 0.26 to 0.69). Restricting the follow-up period improved reliability. In Study 1, the test-retest reliability of LBP over a three-month period was 0.68 (95% CI: 0.41 to 0.87). In Study 2, the ICC of samples taken ≤7 days apart was 0.61 (95% CI: 0.29 to 0.86). In Aim 2 we found that while alpha diversity did not differ by LBP tertile, the beta-diversity was statistically significantly different between groups using unweighted Unifrac, but not weighted Unifrac. Several taxa, particularly those found in the Clostridia class may be more prevalent in women with low levels of LBP, while Bacteroides may be more prevalent with high levels of LBP. Additionally, findings from the current study suggested that LBP was associated with adiposity, but LBP did not mediate the association between adiposity and CRP. In Aim 3 we did not find a statistically significant association between LBP and CRC. Compared to individuals whose LBP concentrations were in the lowest quartile, the ORs associated with second (range: 24.1-31.7 µg/mL), third (range: 31.8-42.6 µg/mL), and fourth (range: 42.7-107.5 µg/mL) quartiles were 1.23 (95% CI = 0.91-1.67), 1.36 (95% CI = 1.01-1.83), and 1.01 (95% CI = 0.73-1.39), respectively, (Ptrend = 0.66). CONCLUSIONS – Results of Aim 1 suggest that plasma LBP may serve as a reliable marker in short-term studies; however, multiple samples may be needed in longitudinal studies to obtain more stable measures. Results from Aim 2 suggest that there may be differences in distribution of rarer taxa between tertiles of LBP, but overall community structures do not differ between groups. Despite finding an association between inflammation, as measured by CRP, and LBP in Aim 2, results from Aim 3 do not provide clear evidence of an association between plasma LBP concentrations and CRC risk. Additionally, there was no evidence of modification by BMI, dietary fiber intake and saturated fat intake, cancer site, or cancer stage.