PO.PR01.03 · 预防研究
绝经后乳腺癌一种新的潜在危险因素——血浆蛋白LEG1同源物的特征刻画
Characterizing plasma protein LEG1 homolog, a novel potential risk factor for post-menopausal breast cancer
作者与单位 Authors & Affiliations
摘要 Abstract
中文摘要
背景：我们此前在2项队列研究中，在对乳腺癌危险因素进行校正的模型中识别并确认了血浆蛋白LEG1同源物与绝经后乳腺癌风险之间的正相关。这种进化上保守的蛋白尚未被充分研究。我们现在旨在刻画该蛋白随时间的个体内变异、其相关因素以及在校正相关因素后的残余关联。
方法：我们分析了社区动脉粥样硬化风险研究（ARIC）的数据，我们最初在该研究中识别出该蛋白（Syleouni M等，JNCI 2025）。我们纳入了4,403名无癌症的绝经后女性（74%未接受激素替代治疗[HRT]）。血浆蛋白此前在第2、3和5次随访（间隔3年、18年）通过SomaScan® 5K测定。我们使用了通过访谈收集或由经过培训的工作人员测量的人口学、人体测量学、生活方式和生育数据。此前测定了血浆雌二醇和睾酮。我们评估了随时间的个体内变异（ICC），并在log2转换后识别了LEG1的相关因素（线性回归）。我们使用多变量校正的Cox回归，通过分层或校正相关因素重新分析了LEG1-乳腺癌关联。对于高度相关的因素，我们首先将LEG1对每个相关因素进行回归，然后对LEG1残差进行建模。
结果：LEG1在3个时间点上的ICC为0.52。LEG1与催乳素诱导蛋白（r=0.95）和细胞外糖蛋白lacritin（r=0.86）高度相关；这些相关性在不同年龄、种族、HRT使用、BMI和糖尿病状态下几乎相同。其余4,709种蛋白的r < |0.35|，包括与肾功能（白蛋白、ALT、AST）和炎症（CRP）标志物的弱相关。LEG1与雌二醇（间隔3年测定，N=316，r=0.13，p=0.02）和睾酮（间隔3年测定，N=3,134，r=0.05，p=0.02）仅呈弱相关。LEG1随年龄适度下降（p<0.0001）。在校正年龄后，黑人女性的水平显著高于白人女性（p<0.0001）。在同时校正年龄和种族后，BMI较高和患糖尿病的女性LEG1水平较低。就HRT使用和生育因素而言，关联方向似乎因种族而异（仅在白人女性中使用者的水平较高）。LEG1对催乳素诱导蛋白（p=0.0001）或细胞外糖蛋白lacritin（p=0.015）回归后的残差与乳腺癌呈统计学显著的正相关。LEG1的关联在不使用HRT、BMI≥中位数及患糖尿病的女性中更强；我们此前报告该关联在不同种族间无差异。
结论：本研究为血浆蛋白LEG1同源物作为绝经后乳腺癌危险因素提供了持续的支持。应考虑该蛋白在当前风险分层工具中的应用价值。尚需机制研究。资助：NHLBI、NCI、NPCR
查看英文原文 English abstract
Background: We previously identified and confirmed a positive association between plasma protein LEG1 homolog and post-menopausal breast cancer risk in 2 cohort studies in models adjusted for breast cancer risk factors. This evolutionarily conserved protein is understudied. We now aim to characterize the protein's within-person variation over time, its correlates and residual association after accounting for correlates.
Methods: We analyzed data from the Atherosclerosis Risk in Communities study, in which we first identified the protein (Syleouni M et al. JNCI 2025). We included 4,403 post-menopausal women (74% not taking hormone replacement therapy [HRT]) without cancer. Plasma proteins were previously measured by SomaScan® 5K at Visits 2, 3 and 5 (3, 18 years apart). We used demographic, anthropometric, lifestyle, and reproductive data collected by interview or measured by trained staff. Previously, plasma estradiol and testosterone were measured. We assessed within-person variation over time (ICC) and identified correlates of LEG1 after log 2 transformation (linear regression). We re-analyzed the LEG1-breast cancer association stratifying by or adjusting for the correlates using multivariable-adjusted Cox regression. For highly correlated factors, we first regressed LEG1 on each correlate and then modeled LEG1 residuals.
Results: The ICC for LEG1 was 0.52 across the 3 time points. LEG1 was highly correlated with prolactin-inducible protein ( r =0.95) and extracellular glycoprotein lacritin ( r =0.86); these correlations were almost identical by age, race, HRT use, BMI, and diabetes. The remaining 4,709 proteins had r < |0.35|, including weak correlations for kidney function (albumin, ALT, AST) and inflammation (CRP) markers. LEG1 was only weakly correlated with estradiol (measured 3 years apart, N=316, r =0.13, p=0.02) and testosterone (measured 3 years apart, N=3,134, r =0.05, p=0.02). LEG1 declined modestly with age (p<0.0001). Adjusting for age, level was notably higher in Black than White women (p<0.0001). Adjusting for both age and race, LEG1 level was lower in women with higher BMI and diabetes. Direction of the association appeared to differ by race for HRT use (level was higher in users in White women only) and reproductive factors. The residuals of LEG1 after regression on prolactin-inducible protein (p=0.0001) or extracellular glycoprotein lacritin (p=0.015) were statistically significantly positively associated with breast cancer. Associations for LEG1 were stronger in women not using HRT, ≥median BMI, and with diabetes; we previously reported no difference in the association by race.
Conclusions: This study provides continued support for plasma protein LEG1 homolog as a risk factor for post-menopausal breast cancer. This protein should be considered for its utility in current risk stratification tools. Mechanistic studies are needed. Support: NHLBI, NCI, NPCR
利益披露 Disclosure
V. A. Burk, None..
E. A. Platz, None.