Abundant ghrelin gene expression by monocytes: Putative implications for fat accumulation and obesity

Hormones encoded by the ghrelin gene, GHRL , regulate many body systems. Here, we show that GHRL is abundant in monocytes. Analysis of women subjected to bariatric surgery revealed a signi ﬁ cant post-operative reduction of GHRL in monocytes. We hypothesise that GHRL mediates monocyte-adipocyte cross-talk in health and obesity. the


Introduction
Ghrelin is a pleiotropic hormone with roles in appetite stimulation and energy balance (Gahete et al., 2014a). The ghrelin gene, GHRL, is predominately expressed in the stomach, and transcribed at lower levels in multiple cell types (Gahete et al., 2014a). Our laboratory has a long-standing interest in characterising the structure, expression and function of peptide hormones derived from GHRL (Seim et al., 2013(Seim et al., , 2016a. In this study, we initially assessed the expression of GHRL in 35 cell or tissue types by interrogating publicly available RNA-seq data.

Gene expression data processing
RNA-sequencing data, from 35 human adult somatic tissue or cell types (each with a minimum of three biological replicates), were obtained from the Human Protein Atlas (Uhlen et al., 2015) and ENCODE (Dunham et al., 2012) consortia and interrogated for ghrelin gene (GHRL) expression. For sample details see Seim et al., 2016b. To further investigate GHRL expression in monocytes, we examined an RNA-seq data set from 19 women before and three months after bariatric surgery, each with two technical replicates (NCBI GEO: GSE66306). FASTQ files were aligned to the human genome (hg19) using TopHat (v2.0.9) (Kim et al., 2013). Gene counts were computed by featureCounts v1.4.5-p1 (Liao et al., 2014) and normalised using the quantile method (R package 'pre-processCore'). A paired Student's t-test was used for two-group comparisons, with P 0.05 considered significant.

Results and discussion
Surprisingly, GHRL expression was relatively abundant in monocytes and lymphoid tissues, such as bone marrow and lymph nodes (Fig. 1A). An assessment of exon coverage in monocytes confirmed expression of canonical preproghrelin coding exons 1 to 4 ( Fig. 1B), as described for peripheral blood mononuclear cells (Dixit et al., 2004;Gahete et al., 2014b).
As the function of GHRL in immune cells and its relationship with disease has hitherto remained enigmatic, we examined transcriptomes of 19 women before and three months after bariatric surgery (Poitou et al., 2015). Our analysis revealed significantly reduced GHRL expression in monocytes after surgery (P ¼ 0.0001) ( Fig. 1C and Table 1), but no difference in adipocytes from the same cohort (data not shown). Serum and enteroendocrine ghrelin production are reduced in most patients following bariatric procedures (Meek et al., 2016), and we now demonstrate that monocyte GHRL expression is also modulated.
Reduced GHRL expression in monocytes postoperatively may result from altered expression or signalling of GHRL-regulating molecules (e.g. leptin and cytokines), or from improved glucose homeostasis. In blood and tissues, the pro-inflammatory cytokine interferon-g (IFN-g) is elevated in obesity and reduced following bariatric surgery (Zhang et al., 2011;Monteiro-Sepulveda et al., 2015), and there is evidence that IFN-g may regulate GHRL expression. GHRL expression was significantly decreased in human Fig. 1. A. A box plot reveals high GHRL expression in monocytes and associated tissues (red dots). From the Human Protein Atlas: adipose tissue, adrenal gland, appendix, bone marrow, cerebral cortex, colon, duodenum, endometrium, oesophagus, gall bladder, kidney, liver, lung, lymph node, ovary, pancreas, placenta, prostate, rectum, salivary gland, skeletal muscle, skin, small intestine, smooth muscle, spleen, stomach, testis, thyroid gland, tonsil, urinary bladder, and uterine tube. From ENCODE: keratinocytes, (CD4 þ ) monocytes, and osteoblasts. B. Integrative Genomics Viewer (IGV) visualisation of the GHRL locus in monocytes validates expression of all canonical preproghrelin-coding exons. C. Monocyte GHRL expression is reduced 3 months after bariatric surgery. Mean ± S.E.M. P value, Student's t-test. D. GHRL expression (qRT-PCR) in human THP-1 leukaemia leukemia cells treated with 50 ng/ml interferon-g (IFN-g) for 24 h (n ¼ 6 per group), compared to untreated controls. Mean ± S.E.M. **P 0.01, ManneWhitney U test. Table 1 Monocyte GHRL expression shifts three months after bariatric surgery. Quantile normalised gene counts are shown (mean ± S.E.M., n ¼ 2). Ydenotes decreased GHRL expression. A dash (À) indicates no change in gene expression. Patients (PM) underwent adjustable gastric banding (AGB) or Roux-en-Y gastric bypass surgery (RYGB THP-1 monocytic leukaemia cells treated with IFN-g ( Fig. 1D; P ¼ 0.0079), suggesting that IFN-g signalling can potently modulate GHRL expression in this monocyte-derived cell line. Indeed, the bariatric surgery cohort examined here showed evidence of a distinct IFN-g signalling pathway in adipose tissue and monocytes postoperatively (Poitou et al., 2015). IFN-g represses gastric GHRL expression in non-obese mice infected with Helicobacter pylori (Strickertsson et al., 2011) and IFN-g regulation has also been associated with dietary interventions in rodents (Lee et al., 2006;Poitou et al., 2015). Taken together, we speculate that weight reduction modulates the expression of GHRL via an altered IFN-g signalling pathway.
What are the consequences of reduced monocyte GHRL expression after bariatric surgery? Obesity induces inflammation in adipose tissue, and metabolic improvements after bariatric surgery are mediated by the reversal of endocrine and immune responses associated with 'pathogenic' adipose tissue. A very recent study demonstrated that inflamed adipose tissue directly stimulates aberrant monocyte production in the bone marrow (monocytosis) to further exacerbate obesity-associated disease processes (Nagareddy et al., 2014). The GHRL derived peptide hormones ghrelin and obestatin regulate adiposity and may promote associated inflammation in adipose tissue (Tsubone et al., 2005;Rodriguez et al., 2009;Gurriaran-Rodriguez et al., 2011). We hypothesise that obesity-associated monocytosis and inflammation, coupled with the action of monocyte-derived paracrine GHRL-derived hormones on adipocytes, regulates metabolism and adipogenesis. Conversely, reduced monocyte GHRL expression following bariatric surgery may improve adipose tissue function.
Monocyte GHRL expression was reduced in~60% of subjects 12 weeks postoperatively, while expression was not altered in the remainder (Table 1). Importantly, the patients were still markedly obese (Poitou et al., 2015). It would be of interest in future studies to examine whether monocyte GHRL expression is reduced for longer than three months after surgery and correlates with successful longer-term bariatric surgery outcomes. In addition, it would be useful to examine the expression of monocyte GHRL in other obesity-associated pathologies, such as Prader-Willi syndrome (where patients exhibit hyperghrelinaemia and hyperphagia from an early age) and metabolic syndrome in general (a risk factor for cancer). Our study demonstrates the power of mining public genome-wide gene expression data and provides further impetus into the study of cross-talk between the endocrine and immune systems and the ghrelin axis.

Funding sources
We acknowledge financial support from the National Health and Medical Research Council Australia (grant no. 1059021; to IS, PLJ, LKC), Cancer Council Queensland (grant no. 1098565; to IS and LKC), a QUT Vice-Chancellor's Senior Research Fellowship (to IS), the Movember Foundation and the Prostate Cancer Foundation of Australia through a Movember Revolutionary Team Award (to LKC), and a travel grant from the European Society of Endocrinology to present this study at the European Congress of Endocrinology, Munich, Germany, May 2016 (to IS).