Weight loss with or without intragastric balloon causes divergent effects on ghrelin cell expression

Abstract Objective The mechanism of action of intragastric balloons in the treatment of obesity is not fully understood. One of the hypotheses is that balloons might have an effect on the fundus, the area of ghrelin production. Methods Participants were randomized to a 13‐week period of sham or balloon treatment followed by a 13‐week period of balloon therapy in everyone. Blood samples for ghrelin levels were taken in the fasting state and after a breakfast at the start, after 13 and 26 weeks. Biopsies for ghrelin cell immunohistochemistry were taken from the fundus at endoscopy. Results Seven participants entered the balloon–balloon (BB) group and 11 the sham–balloon (SB) group. Despite a considerable weight loss, a median −17.9 kg (interquartile ranges −23.8 to −0.5) in the BB group and −18.3 kg (−22.7 to −14.7) in the SB group, fasting ghrelin and meal‐induced ghrelin response did not change. In the SB group, the number of ghrelin cells increased significantly (p 0.001) from 110.6 (83.6–118.9) to 160.2 (128.5–223.0) while on sham treatment and returned to initial levels, 116.3 (91.7–146.9) (p 0.001), when they received their first balloon. No significant changes in ghrelin cell numbers were observed in the BB group. Conclusion In participants without a balloon, weight loss induced an increase in ghrelin cell numbers in the fundus, which was annulled by the subsequent placement of a balloon. The effect of a balloon might be explained by effects on ghrelin cell numbers or ghrelin cell activity.

balloon systems. [4][5][6][7][8] Promising long-term weight maintenance data have been reported with this method. 9,10 In patients with superobesity, preoperative treatment with intragastric balloons considerably reduced liver volume to facilitate surgery 11 and decreased operative time and overall risk of significant adverse outcomes. 12 Meanwhile, two balloons have been approved by the FDA for a 6-month therapy in patients with a BMI 30-40 kg/m 2 with the requirement of supportive therapy for a total of 12 months: that is; (1) the Orbera balloon (formerly the BioEnterics Intragastric Balloon or BIB, Apollo Endosurgery) since 2015 13,14 ; and (2) the Obalon orally ingested intragastric balloon (Obalon Therapeutics) since 2016. 15,16 Pivotal trials with two new and innovative balloon systems have just been finished and await approval by the FDA. [17][18][19] However, the exact mechanism by which intragastric balloons induce weight loss is still not known. Intragastric balloons are hypothesized to mediate satiety peripherally, by being a physical impediment of food intake, by reducing the gastric capacity and by delaying gastric emptying, and centrally, by activating gastric stretch receptors that transmit signals via afferent vagal nerves, the solitary tract and paraventricular nuclei, to the ventromedial and lateral hypothalamus. 4,[20][21][22] Short-term satiety is primarily affected by gastric distension and gastric volume, that is, by the weight and volume of the food rather than its energy content. 22,23 This volumeregulated satiety is thought to result primarily from gastric distension. Indeed, mechanical gastric balloon distension to a volume greater than 400 ml during meals significantly reduced oral intake. 23,24 After the intake of a meal the gastrointestinal tract releases regulatory peptide hormones contributing to the feeling of satiety. 20 The release of postprandial hormones is principally mediated by two mechanisms; (1) by the contact of nutrients, with its different physiochemical properties with receptors in the GI tract; and (2) by the mechanical distension of the stomach leading to increased neural afferent firing in the vagal nerve, affecting feeding-related areas in the brain, including the hypothalamus. 4,21 To date there are only few prospective studies evaluating the effect of chronic gastric distension on neuroendocrine mechanisms in humans. Mathus-Vliegen et al. reported the effects of sham treatment or balloon positioning on gut hormones such as cholecystokinin (CCK) and pancreatic polypeptide (PP). 25 Ghrelin is an orexogenic hormone in contrast to CCK and PP. It is secreted by the fundic glands of the stomach and usually increases food intake. 26 Receptors for ghrelin are expressed on capsaicin sensitive vagal afferent neurons innervating the gut. 27 In animal experiments it has also been shown that gastric distension affects the leptin sensitivity of the nucleus tractus solitarius. 28 Fasting ghrelin levels have been investigated in patients treated with air-filled and fluid-filled balloons. An 800-ml air-filled balloon resulted in increased ghrelin levels after 1 and 4 weeks of balloon insertion and decreased at the time of balloon removal after 16 weeks. 29 Most studies have been performed with a 500-700 ml fluid-filled balloon with divergent results. Fasting ghrelin levels have been reported to remain unchanged 30 or to increase 29-31 during intragastric balloon therapy. Mion et al. reported that chronic gastric dilatation with a balloon reduced fasting plasma ghrelin levels, indirectly indicating that chronic gastric distension might have an inhibitory effect on ghrelin secretion. 32 They also demonstrated a significant decrease in gastric emptying rates.
There are only a few studies that compared balloon treatment with a control group. Konopko-Zubrzycka et al. found a significant increase in fasting ghrelin levels after 1 month and 6 months of balloon therapy compared with controls, but in the former group weight losses were 4 to 5 times as great. 33 Martinez-Brocca et al.
found similar weight losses in balloon-treated and control patients and no differences between the groups in fasting and postprandial ghrelin levels. 34 Mathus-Vliegen et al. performed a patient-and physician-blinded sham-controlled balloon treatment study and found that both balloon-balloon and sham-balloon treatment resulted in significant weight losses after 13 and 26 weeks, but unexpectedly, fasting ghrelin concentrations and the ghrelin response after a meal did not change and certainly did not rise, as is usually seen after energy restriction and weight loss. 35,36 When the balloon was located in the fundus region, ghrelin concentrations were more suppressed. 35 The mode of action of intragastric balloons could thus partly be ascribed to the inhibition of ghrelin secretion, that is, more specifically when the balloon was located in the fundus where ghrelin is secreted. Therefore, we decided to use biopsies taken from the fundus to explore whether weight loss with an intragastric balloon compared with sham treatment was associated with a change in the number of ghrelin cells.

| Study design
The study design has been described in detail. 5,25,35 In brief, participants involved in a trial in the Academic Medical Centre of the University of Amsterdam, investigating the 1-year weight loss with intragastric balloon treatment (Orbera, Apollo Endosurgery) and the weight maintenance in a 1-year balloon-free follow-up period were asked to take part in a study investigating the mechanism of action of intragastric balloons. Before entering the trial, participants had to conform to the criteria of eligibility, that is, age ≥18 years, BMI of at least 32 kg/m 2 without significant changes over the last 3 months, and failure to lose weight within a supervised weight loss program.
Also, gastrointestinal lesions and a large hiatal hernia (assessed at the screening endoscopy) and previous intra-abdominal or bariatric surgery had to be absent.
As shown in Figure 1 and 60 min after the breakfast meal were centrifuged and plasma was stored at À 70°C. Because samples were collected and stored without acidification to stabilize the labile side chain of the acylated ghrelin, total plasma ghrelin was analyzed. Under a wide variety of physical condition the ratio of acylated and total ghrelin is constant and thus total ghrelin is a reasonable surrogate for acylated ghrelin. 37 Total plasma ghrelin was measured using a commercially available RIA kit (Linco Research Inc.). It has a 100% specificity for ghrelin and a limit of sensitivity at 93 pg/ml. The mean intra-assay coefficient of variation was 5.7% (range, 2.4-8.4) at the lower concentration (353-734 pg/ml) and 3.7% (range, 1.0-9.0) at the higher concentration (905-1.879 pg/ml). 35 All samples were measured in duplicate and in one run.

| Participants
The present study comprises all 43 participants from the original trial with complete biopsies and blood ghrelin values from the three timepoints. Another requirement was the absence of Helicobacter pylori because of reported interference with plasma ghrelin levels and ghrelin immunoreactive cells. [38][39][40][41] Participants were withdrawn because of intolerance of the balloon or not achieving the perprotocol required weight loss of 6.5 kg after the first 13 weeks of balloon treatment. This minimum weight loss was required by the FDA for safety and effectiveness reasons in this FDA approved protocol ( Figure 2). A total of nineteen participants were included, seven participants belonged to the balloon-balloon (BB) group and twelve participants to the sham-balloon (SB) group.

| Immunohistochemical procedures
Paraffin embedded biopsies from the 19 participants that participated during the whole study time of 26 weeks were prepared for the immunohistochemical analysis at the Division of Molecular Medicine and Virology, Linköping University, Sweden. Biopsies from T0, T13, and T26 were cut at 5 µm on a microtome. One subject from the SB group was later excluded due to the absence of glandular tissue in the biopsies.
The deparaffinization/rehydration procedure was performed by first heating the slides at 65°C for one hour followed by treatment in

| Statistical analysis
The effects of treatment were tested by Student t test between the groups and by paired t test within the same group for normally distributed variables. For non-normally distributed variables, the Mann Whitney U test and the Wilcoxon signed-rank test were used, respectively. The release of ghrelin during the meal was calculated as the area under the curve (AUC) using the trapezoidal rule (measured 15 min before meal until 60 min after meal). A p value < 0.05 was considered significant. Participant characteristics are given as median (and interquartile [IQR] ranges).

| Ethics
The study was approved by the Medical Ethical Committee of the Academic Medical Centre, University of Amsterdam, the Netherlands for the first 13 weeks. Thereafter, by amendment, a second period of 13 weeks was allowed.

| Participant characteristics
Participants did not differ as to age, body weight, BMI, fasting gastrin, and insulin levels ( Table 1). In follow-up, weight losses were substantial and did not differ between the two groups at 13 and 26 weeks. However, the second period of 13 weeks (week 13 to week 26) when the SB group received their first balloon while BB group continued with balloon therapy, revealed a significantly greater weight loss for the sham-balloon group, that is, À 9.0 (À 10.8 to À 5.2) versus À 5.0 (À 7.2 to À 3.1) kg (p 0.027) ( Table 1).

| Immunohistochemistry of ghrelin cells
Ghrelin immunoreactive cells were, as expected, seen distributed in the epithelium of the glandular portion in the fundus mucosa ( Figure 3). The number of ghrelin positive cells did not differ between SB and BB group at inclusion or at time point 13 weeks or 26 weeks (Table 1). In contrast, when analyzing data within the groups, participants in the SB group showed a significantly higher number of ghrelin cells after 13 weeks (after sham period) as shown in Figure 4A

| Blood analysis of ghrelin
There was no difference between SB and BB regarding preprandial fasting ghrelin levels or meal-suppressed ghrelin secretion (AUC ghrelin) at baseline or at the follow-up period at week 13 and 26 (Table 1). Similarly no differences were seen over time, analyzing SB and BB group separately (Figures 4B,C). There was also no trend weeks between participants with less severe or severe obesity.

| DISCUSSION
In a previous study Mathus-Vliegen et al. demonstrated that the expected increase in ghrelin after a balloon-induced weight loss of 15% did not occur, 35 which has been noticed after weight losses as small as 5% of body weight. 36 Specifically, when the balloon was located in the fundus region, the place of synthesis of ghrelin, ghrelin concentrations were more suppressed. 35  there was a greater influence on ghrelin secretion when the balloon was in the fundus, we could not find such a correlation as to ghrelin producing cells in the present study. 35 Only 38% of balloons were truly located in the fundus, the remainder was either in the mid stomach or in the antrum.
To our knowledge, this finding of increased ghrelin producing cells in participants losing weight is a new finding. Maksud et al.
compared patients with severe obesity with normal-weight dyspeptic patients and found a higher density of ghrelin producing cells in patients with severe obesity. 42 However, they did not address the fate of these cells after weight loss. Goitein et al. investigated sleeve gastrectomy specimens and discovered a decline in average ghrelin cell counts from fundus to the pre-antral region. 43 In a study, using the articulating circular endoscopic stapler to perform a gastroplication transorally, we found a median excess weight loss of 34.9% (IQR 17.8 -46.6). 44  Most of the studies that tried to explain the mechanism of action of intragastric balloons focused on gastric emptying. Interestingly, in this respect is the finding that the degree of weight loss was highest in participants who had the greatest delay in gastric emptying after balloon positioning. [46][47][48] It would have been interesting to measure fasting and meal-suppressed ghrelin levels in these participants to substantiate our findings.
The strengths and limitations of this study should be acknowledged. The strength of this study is that the study design allowed to study the consecutive effects of sham treatment and balloon posi- levels. 30 Also, we did not measure gastric emptying, but two findings indicated a delayed gastric emptying: the decreased meal-stimulated CCK release 25 and the finding of food obstructed above or adhered to the balloon as could be judged at endoscopy upon balloon exchange. 5

| CONCLUSION
In conclusion, this study showed that weight loss in those without a balloon, that is, the sham-treated group, induced an increase in ghrelin cell numbers in the fundus, in contrast to those who lost weight with having a balloon where no such ghrelin cell increase was seen. To what extent the degree of prolongation of the gastric emptying time and its prognostic value as to weight loss are explained by effects on ghrelin release and ghrelin cell numbers or ghrelin cell activity should be worthwhile to investigate further.