Genomic and Behavioral Signatures of Selection for Ethanol Preference from the Heterogeneous Stock Collaborative Cross Mice – The Central Nucleus of the Amygdala II

Alcohol use disorder (AUD) is a complex disease with heritability of ~0.5, indicating genetic and non-genetic factors contribute to risk. Identifying gene expression networks contributing to risk using post-mortem human brain tissue has the limitation of conflating risk for AUD with consequences of alcohol use. We leveraged mice selectively bred for differential ethanol preference from a highly genetically diverse population to overcome this limitation. Ethanol intake was highly correlated with preference, high-preferring (HP) mice consumed more sweet- but not bitter-tasting solutions compared to low-preferring (LP) mice, and the lines did not differ in rate of ethanol elimination. Adult, ethanol-naïve HP and LP mice contributed tissue from the central nucleus of the amygdala (CeA), a region critical to ethanol preference and intake. Single-nuclei and bulk RNA sequencing data were used to identify cell types and transcriptome changes related to selective breeding for differential risk for ethanol preference. Single nuclei analysis identified populations of inhibitory (~48% of cells) and excitatory (~23%) neurons, and non-neuronal (~29%) cells, but no differences in cell-type composition or gene expression were identified between the lines. Bulk CeA analysis identified differences between the lines for: (1) gene expression (2996 genes), (2) expression variability (426 genes), and (3) wiring (407 significant gene-gene correlations). Overall, lower variance was found in the HP line. Reduced gene-gene correlation, also found in HP mice, suggested that selection for high preference induced changes in transcriptional regulation resulting in reduced connectivity, specific to gene networks enriched in markers for inhibitory neurons expressing Isl1 and Tac1. Mice underwent bidirectional selective breeding for ethanol intake (10% ethanol v/v in water, two bottle choice) for 5 generations starting from the Heteroengeous Stock - Collaborative Cross, a strain which emcompasses ~90% of all genetic diversity known in Mus Musculus. After 5 generations, high and low preference mouse had ~10fold different prefernce for ethanol and consumed 7fold differently. Ethanol naive mice from generation 5 were sacrificied at adulthood and brain tissue was collected for the central nucleus of the amygdala, a region important for addiction. bulk RNA sequencing (paired-end) was run on 200 samples (n=50/sex/strain) to a mean depth of 20M, and aligned to the GRCm39 using STAR. Raw counts were normalized using the limma+voom method, and differential expression and weighted gene co-expression analysis were carried out to identify risk vs protection markers for ethanol drinking.

酒精使用障碍（Alcohol use disorder, AUD）是一种复杂疾病，其遗传力约为0.5，表明遗传与非遗传因素共同参与了疾病风险的形成。利用死后人类脑组织鉴定与疾病风险相关的基因表达网络存在局限性，因为无法区分AUD的患病风险与酒精使用带来的生理后果。为克服这一局限，本研究利用从高度遗传多样性群体中选择性繁育出的、具有不同乙醇偏好性的小鼠模型。乙醇摄入量与偏好性高度相关；与低偏好（low-preferring, LP）小鼠相比，高偏好（high-preferring, HP）小鼠摄入更多甜味溶液，但对苦味溶液的摄入无显著差异，且两个品系的乙醇清除速率无明显区别。成年且未接触过乙醇的HP和LP小鼠的杏仁核中央核（central nucleus of the amygdala, CeA）组织被采集——该区域对乙醇偏好与摄入行为至关重要。本研究采用单细胞核测序与批量RNA测序技术，鉴定与乙醇偏好差异风险的选择性繁育相关的细胞类型及转录组变化。单细胞核测序分析鉴定出抑制性神经元（约占细胞总数的48%）、兴奋性神经元（约占23%）以及非神经元细胞（约占29%），但两个品系间未发现细胞类型组成或基因表达的显著差异。批量CeA组织测序分析则发现两个品系存在三方面差异：（1）基因表达水平差异（共2996个基因）；（2）表达变异性差异（共426个基因）；（3）调控网络差异（共407个显著基因-基因相关关系）。总体而言，HP品系的基因表达方差更低。HP小鼠中还观察到基因-基因相关性降低，这提示针对高偏好性的选育改变了转录调控模式，导致特定基因网络的连接性减弱——这类基因网络在表达Isl1和Tac1的抑制性神经元的标志物中显著富集。本研究以远交系-协作杂交（Heterogeneous Stock - Collaborative Cross）小鼠为起始种群——该品系涵盖了小家鼠（Mus Musculus）已知全部遗传多样性的约90%——开展了5代的双向选择性繁育，选育标准为乙醇摄入量（采用10%体积分数乙醇水溶液的双瓶选择实验）。经过5代繁育后，高、低偏好品系小鼠的乙醇偏好性差异达到约10倍，摄入量差异达到7倍。第5代未接触过乙醇的小鼠在成年后被处死，采集其杏仁核中央核组织——该区域与成瘾行为密切相关。随后对200个样本（每个品系/性别各50个样本）进行了双端批量RNA测序，平均测序深度达20M，采用STAR工具将测序reads比对至小鼠参考基因组GRCm39。原始计数采用limma+voom方法进行标准化，并开展差异表达分析与加权基因共表达分析，以鉴定与乙醇饮酒行为相关的风险与保护标志物。