Identifying the Molecular Mechanisms of Early Cachexia Using Whole Transcriptome Sequencing in Muscle and Fat Biopsies from Cancer Patients
Abstract
Cancer cachexia is responsible for one third of cancer-related deaths and contributes to the death of many others. More than 80% of cancer patients are cachectic towards the end of life. Despite intensive research, the mechanisms of cancer cachexia are still poorly understood. It is our hypothesis that identification of early changes in gene expression in cachexia will lead to an improved understanding of the mechanism that trigger this problem in cancer patients.
Thus, to shed light on the mechanisms involved in the major cachexia target tissues, we investigated the entire transcriptome in muscle and fat to identify altered expression of genes in cancer patients with and without cachexia.
Samples of rectus abdominis muscle and visceral fat were collected at surgery from patients exhibiting 5-10% weight loss prior to surgery, compared with stable-weight patients. Analysis of all expressed genes was carried out using next generation sequencing (Illumina HiSeq 2500). Also, selected differentially expressed genes were confirmed using real time RT-PCR.
In muscle, 30 genes showed highly significant changes in expression (25 downregulated and 5 unregulated: P<0.0005 – P<0.00001, FDR 0.2). Analysis of the 25 downregulated genes involved included 7 that are involved with metabolism (5 of which are mitochondrial); 4 with signaling; 4 with ubiquitination; and 3 with intercellular trafficking. There was marked downregulated of multiple genes involved in glycogen metabolism which correlates with the lack of glycogen, muscle weakness, and fatigue; characteristic of cachexia. The 5 upregulated genes include 2 involved with calcium signaling and 2 with cell matrix interactions. Expression of genes previously thought to be important in cachexia, including several inflammatory cytokines, was not significantly different. FBX032, which encodes atrogin-1, upregulated in an in vitro cachexia model, was actually downregulated. No transcripts for the dermicidin gene, which contains the sequence that codes for the backbone peptide of proteolysis-inducing factor, were detected. Expression of myostatin was significantly decreased as was its receptor (ACVR2B), possibly reflecting end organ adaptation to tumor produced myostatin.
In visceral fat, expression of 6 genes were downregulated and 10 upregulated with high statistical significance (P<0.001-0.0002). Several of these encode metabolic enzymes. Of genes in fat previously implicated in cachexia, such as hormone sensitive lipase and adipose tissue triglyceride lipase, were unchanged. In contrast, leptin was significantly downregulated and the zinc- α-2-glycoprotein (lipid mobilizing factor) was significantly upregulated as expected.
These studies confirmed that for a multifactorial condition, genome wide transcriptome analysis is the method of choice to explore the disease complexity. They explain some documented evidence in cachexia pathogenesis, highlight ambiguous data from animal models, and reveal unexpected changes in gene expression that underlie the pathophysiology of the cachectic state in cancer. These results bring reliable, representable, and consistent data from the clinic and back to the bench with more focused insights to be investigated and verified.