Introduction

Ghrelin, a ligand for the growth hormone secretagogue receptor, is an orexigenic hormone produced primarily by cells in the oxyntic glands of the stomach [1–4]. It plays a physiological role in the regulation of food intake with plasma ghrelin levels rising shortly before meals and declining rapidly after food is consumed in humans and rodents [3, 5, 6].

Ghrelin is a 28 amino acid gut peptide derived from the 117 amino acid proghrelin. Following cleavage by proprotein convertase 1, it is stored in secretory vesicles of endocrine cells. Further acylation by ghrelin O-acyltransferase, adding an octanoyl group to a serine residue, has been shown to be crucial for its biological function [2]. Ghrelin is a multifunctional molecule, involved in many biological processes ranging from appetite regulation and growth hormone release to gut motility. Though produced throughout the length of the intestine, ghrelin is mainly secreted by the stomach, followed by the small intestine. Negligible quantities are also expressed by peripheral tissues, such as the pancreas and brain [1, 7, 8].

Plasma ghrelin levels rise during food deprivation and decline after food ingestion; this peptide is therefore referred to as a “meal initiator” [5]. Interestingly, fasting plasma ghrelin concentration is negatively correlated with body mass index (BMI) [6].

Sleeve gastrectomy (SG) is rapidly gaining acceptance both as a first stage in high risk patients or standalone procedure for treatment of morbid obesity. It entails a vertical resection of most of the stomach volume, removing part of the antrum, most of the body, and all of the fundus [9]. The proposed mechanism of action of this procedure is a combination of volume restriction, creation of a high-pressure system, and the induction of a favorable hormonal change [10]. Low plasma ghrelin levels were monitored up to 5 years after SG [11].

To better elucidate the “hormonal” mechanism of SG, we examined the distribution of ghrelin mRNA expression by real-time polymerase chain reaction (RT-PCR) and ghrelin-producing cell distribution by immunohistostaining of resected stomach specimens obtained in the course of SG in morbidly obese subjects. Circulating plasma ghrelin levels before and 3 months after surgery were measured as well.

Materials and Methods

Subjects

Twenty morbidly obese patients (3 males and 17 females) scheduled to undergo SG were enrolled in the study. Institutional review board approval was obtained prior to study initiation and all subjects signed an informed consent form.

All subjects met the standard NIH criteria for bariatric surgery (BMI >40 kg/m2 or BMI >35 kg/m2 with obesity-related comorbidities). Exclusion criteria included previous bariatric surgery, chronic liver or renal disease, and heart failure.

Laparoscopic Sleeve Gastrectomy

Surgical technique was identical to our standard practice. A five-port laparoscopic procedure was performed with two 12-mm working ports at the patient’s left upper and right mid-abdomen, one 10.5-mm port for the camera in the supra-umbilical midline, and two 5-mm ports at the subxyphoid and left costal margin, anterior axillary line, for retraction. The greater curvature of the stomach was freed from the omentum using ultrasonic shears from 5 cm proximal to the pylorus to the angle of His. Extra care was taken in dissecting the posterior aspect of the fundus to expose the lesser curvature caudally, all the way to the left crus of the diaphragm, with meticulous division of all short gastric vessels. The gastric sleeve was constructed over a 42F bougie (Pilling; Maloney Bougie Tungsten, Teleflex Medical, High Wycomb, UK). A linear-cutting stapler (Echelon 60 Endopath®, Ethicon Endosurgery, Cincinnati, OH, USA) was used from 5 cm proximal to the pylorus to the angle of His with varying stapler heights, according to tissue thickness, maintaining uniform sleeve diameter throughout its length, with over-suturing of the entire stapler line. The entire fundus was thus completely resected and removed.

Plasma Ghrelin Determination

The levels of plasma total ghrelin were measured by a commercial radioimmunoassay kit (LINCO Cat# GHRT-89HK). Blood samples were taken from patients before and 3 months after surgery after an overnight fast according to the manufacturer’s kit specifications. Radioactivity was determined by the WALLAC gamma counter (1261). Two samples were excluded due to technical error in sample handling.

Immunohistochemistry

Stomach specimens resected during SG were thoroughly washed in saline immediately following extraction to remove debris (blood, loose fat tissue, and mucus). Stomachs were fixed in buffered formalin 10% for 24 h.

After fixation, three tissue samples were obtained from three areas of each stomach: the fundus, body, and pre-antral area. These samples were paraffinized, sectioned, and mounted on slides. Tissue samples were treated by sequential immersion in xylene and graded concentrations of ethanol for removal of paraffin and rehydration. They were then rinsed in distilled water. Endogenous peroxidase activity was blocked by immersion of the slides in 3% hydrogen peroxide diluted in phosphate-buffered saline (PBS) for 10 min followed by rinsing in PBS. Nuclear Decloaker (Biocare Medical # CB911M) with color-coded pH indicator was used for antigen retrieval from the formalin-fixed tissue. Slides were incubated 1 h with GHRL monoclonal antibody (M01) clone 2F4 Abnova (#H00051738-M01). As control, duplicate slides from the same areas were incubated with nonspecific anti-IgG antibodies. After rinsing, slides were incubated for 30 min with Histofine Simple Stain MAX PO (Nichirei Biosciences Inc.) universal immunoperoxidase polymer, anti-mouse, and anti-rabbit and washed in PBS. Chromogen/substrate reagent was added and slides were rinsed and mounted with DAKO liquid DAB (#k3466). Positive cells were counted in 9 of 20 high (×10) magnification fields randomly chosen for each specimen.

Real-Time Polymerase Chain Reaction

The expression of ghrelin at different regions of the stomach was analyzed. RNA was isolated from the stomach tissue using RecoverAll™ Total Nucleic Acid Isolation (Ambion #1975). RNA was purified using glass fiber filter, with on-filter nuclease treatment, and eluted into water. Single-stranded cDNA from total RNA was synthesized using high capacity cDNA reverse transcription (Applied Biosystems #4374966).

RT-PCR analysis was performed (TaqMan, 7000 sequence detection system, Applied Biosystems). A mixture of 5 μL 1:5 diluted cDNA, 100 nM of primer, 200 nM of probe, and Taqman Universal Master Mix in 20 μL aliquots was prepared. The reaction was initiated by 2 min incubation at 50°C and an initial 10-min denaturation at 94°C. This was followed by 40 cycles consisting of denaturation at 94°C (15 s) and annealing and elongation at 60°C (1 min).

mRNA levels are calculated by comparing the cycle where ghrelin was detected to a control gene. This is the difference in cycle time or ΔCT. The relative quantification (RQ) is the relation between ghrelin and control gene expression and is calculated by the formula RQ = 2−ΔCT. The control gene used was S18, coding for ribosomal RNA, which is avidly transcribed in all tissues.

Statistical Analysis

Data are presented as mean values ± standard deviation. Paired Student’s t test was used for continuous normally distributed variables. Pearson correlation coefficient (r) and the significance for it (P) were calculated between variables.

Analysis of variance using the Duncan multiple comparison option and non-parametric Kruskal–Wallis tests were used to compare measured values from the various regions of the stomach. Statistical significance was considered as a probability of a type 1 error <5%.

Results

Mean total ghrelin plasma levels decreased sharply from 1,676 ± 470 pg/ml before to 1,179 ± 188 pg/ml when measured 3 months after surgery (a 27% drop; P < 0.00001) (Fig. 1). Mean BMI decreased from 46 ± 6 to 38 ± 5 kg/m2 in the same period. However, we found no correlation between the decline in plasma ghrelin levels and BMI in the postoperative measurement (R 2 = 0.0426) (Fig. 2).

Fig. 1
figure 1

Plasma ghrelin levels were determined by radioimmunoassay analysis, measured before and 3 months after sleeve gastrectomy. Data are expressed as mean ± SD. *P < 0.00001, decrease in plasma ghrelin levels was highly significant

Full size image
Fig. 2
figure 2

Correlation between decline in ghrelin levels and BMI. There was no significant correlation between decline in ghrelin levels and BMI. The percent decline in ghrelin levels and percent BMI loss for each subject is shown (R 2 = 0.0425, P > 0.3)

Full size image

Ghrelin mRNA was readily detectable in all examined regions of the stomach with gradually decreasing expression from the fundus distally. The RQ of ghrelin mRNA expression was 0.043 ± 0.016 at the fundus, 0.026 ± 0.01 at the body, and 0.015 ± 0.006 in the pre-antral region. Thus, the fundus expressed ghrelin 1.7- and 2.9-fold more than the body and pre-antral stomach, respectively. The difference between the fundus and pre-antral region was statistically significant (P = 0.05) (Fig. 3). We found that the RT-PCR amplification plot of ghrelin was roughly only four cycles behind that of ribosomal gene S18, a known highly expressed protein, demonstrating the high expression of ghrelin in the stomach of our obese subjects.

Fig. 3
figure 3

Ghrelin gene expression. The relative quantification of ghrelin expression was measured by RT-PCR in the fundus, body, and pre-antral region. There was a trend of decrease in expression when measuring distally from the fundus. *P = 0.05, statistical significance when compared to the fundus

Full size image

The prevalence of ghrelin-producing cells as detected by immunohistochemical staining followed the same pattern, with the greatest percentage of cells staining positive for ghrelin in the fundus, intermediate in the body, and fewest in the pre-antral area. Average ghrelin cell counts were 60 ± 40, 45 ± 20, and 39 ± 13 cells/high power field in the fundus, body, and pre-antral region, respectively. This trend was not statistically significant (P = 0.08) (Figs. 4 and 5).

Fig. 4
figure 4

Average number of ghrelin cells. Immunohistochemical analysis for ghrelin was performed on mucosal specimens from the fundus, body, and pre-antral region. The same pattern of ghrelin cell number decrease from the fundus distally is observed. This trend was not statistically significant (P = 0.08)

Full size image
Fig. 5
figure 5

Distribution of ghrelin immunoreactive cells in the fundic gland. Ghrelin immunoreactive cells were recognized mainly near the lamina muscularis. Arrows indicate ghrelin immunoreactive cells (×40). Immunoperoxidase of monoclonal antibody to human ghrelin on formalin-fixed paraffin-embedded human stomach

Full size image

Discussion

In this work, mapping of ghrelin expression and cell distribution was studied in resected stomach segments of morbidly obese individuals following SG. Ghrelin mRNA expression was threefold higher in the fundus than in the pre-pyloric area. The same pattern was also seen in ghrelin cell distribution.

Plasma ghrelin levels decrease after gastric bypass and lose their meal-associated variance but rise after gastric banding in a way similar to dieting, maintaining this variance [12]. After SG, levels have been shown to decrease up to 5 years after surgery [11, 13]. Substantial weight loss has been well documented for these procedures implying that the specific operative manipulation is responsible for the changes observed in circulating ghrelin levels. We found that the total plasma ghrelin levels significantly decrease 3 months after surgery, and a simultaneous significant weight loss was observed in morbidly obese patients 3 months after surgery. However, there was no distinct correlation between these two parameters. This is consistent with our previous findings implying that ghrelin is but one of the myriad of factors contributing to the success of this procedure [10] and that weight loss per se is not responsible for ghrelin level change.

Ghrelin is produced by oxyntic cells along the entire gastrointestinal tract but its main site of production is the stomach [1]. It is therefore not surprising that ghrelin levels decrease immediately after SG and have been shown to remain low for up to 5 years [11, 14].

The total plasma ghrelin level found in obese patients prior to surgery was quite high even though fasting plasma ghrelin concentration is negatively correlated with body mass index. Nevertheless, our study design included measurement of total plasma ghrelin levels. Active ghrelin was not measured and might have been lower. In addition, higher expression of ghrelin mRNA was found (when compared to the highly expressed ribosomal control gene) in all specimens. High levels of mRNA do not always correlate with final protein levels, and this may be true for ghrelin as well. It has recently been shown that in ob/ob and in diet-induced obese mice, there was a significant increase in preproghrelin mRNA levels while tissue and plasma proghrelin-derived peptide levels were significantly reduced [8]. This discrepancy between ghrelin mRNA expression and circulating ghrelin levels has also been demonstrated in humans. Gravid women with gestational diabetes show no association between circulating ghrelin and the higher ghrelin mRNA expression in their placental tissue [15]. The decrease of ghrelin cell count along the stomach is consistent with similar findings demonstrating higher levels of expression in endoscopically obtained samples from the fundus and antrum of healthy patients [16] and surgically obtained specimens from gastric cancer patients [17].

The efficacy of SG in long-term weight reduction is still under scrutiny. Five- and 6-year excess weight loss results range from 55% to 64%, showing a distinct weight regain after a nadir at 2–3 years [11, 18]. Ghrelin levels seem to increase slowly as well, though still remaining significantly lower than prior to surgery [11]. Some authors have demonstrated the “neo-fundus” in radiological evaluation of patients after SG, attributing weight regain to this stretching and enlargement of the superior portion of the sleeve. Ghrelin levels in this subset of patients were not addressed [18–20].

Obestatin is a peptide originally isolated from the rat stomach, derived from proghrelin by post-transcriptional processing. Though sharing the same precursor with ghrelin, it has been shown to have opposite effects, namely inducing diminished food intake, reducing gut motility, and promoting weight loss, albeit in rodents [21]. Obestatin levels were not obtained in this study but it would be interesting to look at the relationship between ghrelin and obestatin in the future.

The presence of ghrelin-producing cells in the antrum as documented in this study is of special interest. There is an ongoing debate regarding the need of antrum removal and the relationship between its removal and long-term weight loss. Some authors strongly recommend protecting the antrum due to its importance as a pumping mechanism for gastric emptying and also stated that resecting part of the antrum will not significantly change the remaining volume of the stomach [22, 23]. Others argue that leaving an intact antrum leaves room for sleeve dilatation, loss of restriction, and concomitant weight regain and suggest to resect closer to the pylorus about 3–4 cm and not 7 cm [24]. This study proposes that, though to a lesser degree, removal of antral tissue facilitates a more complete removal of ghrelin-producing cells. This too should be taken into account in addition to volume reduction, when discussing the surgical technique of sleeve gastrectomy.

Utilizing large caliber bougies to create the sleeve correlates with poorer weight loss results [25], when compared with thinner bougies, perhaps attributable to both reduced restriction and a reduced hormonal effect with more ghrelin cells remaining. Gut peptides seem to play a crucial role in the weight reduction achieved by bariatric surgery. Ghrelin is the only orexigenic gut peptide described to date. Its levels decrease after SG and may contribute to the impressive outcome of this procedure. To the best of our knowledge, this is the first work looking into specific ghrelin cell distribution and expression in resected stomachs of morbidly obese individuals after SG. Extra care should be taken to resect as much ghrelin-producing tissue during SG, and this study, which helps define the borders of this tissue, should assist in optimizing this procedure to achieve more weight loss.