Obesity reduces the anticancer effect of AdipoRon against orthotopic pancreatic cancer in diet-induced obese mice

The antidiabetic adiponectin receptor agonist AdipoRon has been shown to suppress the tumour growth of human pancreatic cancer cells. Because obesity and diabetes affect pancreatic cancer progression and chemoresistance, we investigated the effect of AdipoRon on orthotopic tumour growth of Panc02 pancreatic cancer cells in DIO (diet-induced obese) prediabetic mice. Administration of AdipoRon into DIO mice fed high-fat diets, in which prediabetic conditions were alleviated to some extent, did not reduce either body weight or tumour growth. However, when the DIO mice were fed low-fat diets, body weight and the blood leptin level gradually decreased, and importantly, AdipoRon became effective in suppressing tumour growth, which was accompanied by increases in necrotic areas and decreases in Ki67-positive cells and tumour microvessels. AdipoRon inhibited cell growth and induced necrotic cell death of Panc02 cells and suppressed angiogenesis of endothelial MSS31 cells. Insulin and IGF-1 only slightly reversed the AdipoRon-induced suppression of Panc02 cell survival but had no effect on the AdipoRon-induced suppression of MSS31 cell angiogenesis. Leptin significantly ameliorated AdipoRon-induced suppression of angiogenesis through inhibition of ERK1/2 activation. These results suggest that obesity-associated factors weaken the anticancer effect of AdipoRon, which indicates the importance of weight loss in combating pancreatic cancer.

The adiponectin receptor agonist AdipoRon was discovered as the first orally active antidiabetic drug 27 . Its binding to the adiponectin receptors AdipoR1 and AdipoR2 activates the AMPK, p38MAPK and PPARα pathways and ameliorates impaired glucose tolerance and insulin resistance linked to type 2 diabetes 28 . Recently, we and Messaggio et al. reported that AdipoRon induces cell death in either an AdipoR-independent or AdipoRdependent manner and suppresses in vivo tumour growth in human pancreatic cancer cells 29,30 . Because obesity and diabetes influence the sensitivity to chemotherapy of pancreatic cancer 26 , we investigated the anticancer efficacy of AdipoRon in high-fat-diet-induced obese (DIO) prediabetic mice orthotopically implanted with mouse Panc02 pancreatic cancer cells. The results showed that obesity suppressed the anticancer effect of AdipoRon.

Results
Panc02-Luc-ZsGreen orthotopic tumour growth in DIO mice. The body weight of mice fed an LFD (control mice) at 16 weeks of age was 27.71 ± 1.81 g (n = 11), and that of those fed an HFD (DIO mice) was 43.69 ± 3.31 g (n = 11) (Fig. 1A). The blood glucose levels of control mice (n = 5) and DIO mice (n = 5) were 52.0 ± 27.6 mg/dl and 155.6 ± 33.2 mg/dl, respectively (P = 4 × 10 -4 ). The intraperitoneal glucose tolerance test showed that the clearance of an injected glucose load from the body was slower in DIO mice than in control mice (Fig. 1B), indicating impaired glucose tolerance. We injected Panc02-Luc-ZsGreen cells into the pancreas of these mice and monitored tumour growth. DIO mice developed larger orthotopic tumours than control mice on day 22, as demonstrated by bioluminescent imaging and tumour weight (Fig. 1C-F). The incidence of liver metastasis was not different between control mice (3/8 mice) and DIO mice (2/8 mice) (Fig. 1C). There was a strong positive correlation between body weight and tumour burden (r = 0.834, P < 6.04 × 10 -5 ), suggesting that obesity is responsible for the acceleration of Panc02 tumour growth (Fig. 1G).
Effect of AdipoRon on Panc02-Luc-ZsGreen orthotopic tumour growth. We first addressed whether obesity-associated prediabetic conditions influence the suppressive effect of AdipoRon on the orthotopic tumour growth of Panc02-Luc-ZsGreen cells. We chose to administer 5 mg/kg AdipoRon into the mice because the dose is reported to be effective in relieving diabetic symptoms 28 . AdipoRon was intraperitoneally administered once daily into DIO mice for 10 days. Although body weight was not changed ( Fig. 2A), reductions in impaired glucose tolerance, fasting insulin level and insulin resistance represented by the HOMA-IR value were indeed observed in the AdipoRon-administered group compared to the solvent-administered group (Fig. 2B-E), confirming the efficacy of AdipoRon as an antidiabetic drug. Then, the DIO mice that had received AdipoRon for 10 days were orthotopically implanted with Panc02-Luc-ZsGreen cells and continuously received daily administration of solvent alone or AdipoRon. The mice were maintained on an HFD. As a result, it appeared that the average body weight of the DIO mice was not significantly reduced by AdipoRon administration over the course of the experiments ( Fig. 2A). Moreover, AdipoRon did not significantly affect the growth of orthotopic tumours, as evidenced by bioluminescent imaging and tumour weight ( Fig. 2F-H). Notably, there was still a significant positive correlation between body weight and tumour weight at the end of the experiments (Fig. 2I). Thus, administration of 5 mg/kg AdipoRon apparently attenuated the prediabetic conditions to some extent but did not have statistically significant suppressive effects on obesity and tumour growth.
Next, we divided the DIO mice into four groups after orthotopic implantation of Panc02-Luc-ZsGreen cells: HFD/Vehicle group (n = 6): mice fed an HFD continuously and administered vehicle alone intraperitoneally every other day; HFD/AdipoRon group (n = 6): mice fed an HFD continuously and administered AdipoRon (30 mg/kg) intraperitoneally every other day; LFD/Vehicle group (n = 6): mice fed an LFD instead of an HFD and administered vehicle alone intraperitoneally every other day; and LFD/AdipoRon group (n = 6): mice fed an LFD instead of an HFD and administered AdipoRon (30 mg/kg) intraperitoneally every other day. We confirmed body weight loss and a decrease in serum leptin levels in the LFD groups (Fig. 3A,B). Again, as long as the DIO mice were fed an HFD, we did not observe a significant reduction in body weight or tumour growth by AdipoRon (Fig. 3A,C). However, when the diet was changed from an HFD to an LFD, tumour growth tended to be retarded by AdipoRon, although the differences were not statistically significant (Fig. 3A,C). Importantly, Adi-poRon suppressed tumour growth more significantly in the LFD/AdipoRon group than in the HFD/AdipoRon group (Fig. 3C,D). In agreement with these results, haematoxylin and eosin staining demonstrated a noticeable increase in necrotic areas in the tumour tissues of the LFD/AdipoRon group (Fig. 3E). Immunohistochemical analyses showed that although there was a statistically significant difference in the percentage of cleaved caspase-3 (CC3)-positive cells between the LFD/AdipoRon group and the HFD/AdipoRon group, the percentage was too low to explain the difference in the inhibition of tumour growth (Fig. 3F,I). On the other hand, the reduction in the number of Ki67-positive cells (Fig. 3G,J) and decrease in CD31-positive tumour microvessels (Fig. 3H,K) in the LFD/AdipoRon group were more remarkable than those in the HFD/AdipoRon group, suggesting the possibility that suppression of angiogenesis was partially responsible for the growth inhibition and necrosis. These observations suggested that obesity reduced the anticancer effect of AdipoRon, possibly through suppressing the effects of AdipoRon on cell growth and tumour angiogenesis.

Effect of AdipoRon and obesity-related factors on the survival of Panc02-Luc-ZsGreen cells in vitro.
To obtain any clues about the molecular basis for the inhibitory effect of obesity on the antitumour action of AdipoRon, we first examined the effect of AdipoRon and obesity-related factors on the survival of Panc02-Luc-ZsGreen cells. qRT-PCR analyses showed the expression of AdipoR1 and AdipoR2, but not LepRs, in the cells (Fig. 4A). Treatment of the cells with AdipoRon enhanced the phosphorylation of AMPK and p38 MAPK, indicating the functionality of the AdipoRs, whereas as expected, leptin did not increase the phosphorylation of STAT3 (Fig. 4B). AdipoRon caused cell growth inhibition accompanied by cell death in a dose-dependent manner during incubation for 2 days (Fig. 4C,D). The pan-caspase inhibitor z-VAD-fmk did www.nature.com/scientificreports/ not enhance the survival of AdipoRon-treated cells (Fig. 4E), suggesting that AdipoRon did not substantially induce caspase-dependent apoptosis in the cells. As observed in human pancreatic cancer MIAPaCa-2 cells 29 , AdipoRon induced cytoplasmic swelling with large pieces blebbing from the plasma membrane and eventual cytolysis featuring necrotic cell death (Fig. 4D, Supplementary Fig. S1) and transiently increased the phosphorylation of ERK1/2 (Fig. 4B). The MEK inhibitor U0126 slightly recovered the survival of AdipoRon-treated cells (Fig. 4F). Unlike AdipoRon, adiponectin (APN) did not inhibit survival of the cells (Fig. 4G, Supplementary  We observed that HFD serum appeared to have no effect on either cell growth or observable differences in samples treated with Adi-poRon ( Supplementary Fig. S7), although the sera from HFD-fed mice exhibited elevated ratios of leptin/APN compared to the LFD group ( Supplementary Fig. S7). These results, when taken together appeared to indicate that obesity-related factors did not agitate Panc02-Luc-ZsGreen cells and attenuate the effect of AdipoRon in tumour suppression.

Effect of AdipoRon and obesity-related factors on angiogenesis in vitro.
Based on the suppression of tumour angiogenesis in the LFD/AdipoRon group (Fig. 3H,K) and our previous data that AdipoRon suppressed tumour angiogenesis in a xenograft model of MIAPaCa-2 cells 29 , we examined the effect of AdipoRon on cell growth and tube formation in MSS31 endothelial cells. qRT-PCR analyses showed the expression of AdipoR1 and AdipoR2, and notably LepRb, to a lesser extent, in the cells (Fig. 5A). AdipoRon induced the phosphorylation of AMPK and p38 MAPK, and leptin induced the phosphorylation of STAT3 (Fig. 5B), indicating that both AdipoRs and LepRb elicit signals into the cells. AdipoRon also induced the phosphorylation of ERK1/2 in the cells (Fig. 5B). In contrast to Panc02-Luc-ZsGreen cells, the growth of MSS31 cells was not inhibited by AdipoRon concentrations up to 25 µg/ml (Fig. 5C). A tube formation assay on Matrigel showed that MSS31 cells formed fine tubes in the presence of HGF, and AdipoRon suppressed tube formation (Fig. 5D). Insulin and IGF-1 did not reverse the AdipoRon-induced suppression of tube formation ( Supplementary Fig. S8). However, intriguingly, pretreatment of the cells for 1 h with 100 ng/ml leptin followed by treatment with AdipoRon in the presence of 100 ng/ml leptin alleviated the AdipoRon-induced suppression of tube formation (Fig. 5D). Leptin at 20 ng/ml was also effective ( Supplementary Fig. S8), although it did not enhance the growth of MSS31 cells in a number of culture conditions ( Supplementary Fig. S9). These results suggest that AdipoRon inhibits tube formation and that leptin reverses this effect. To examine the mechanism by which leptin alleviates the suppression of tube formation by AdipoRon, we focused on ERK1/2 activation because we found that the AMPK inhibitor BML-275 and the p38 MAPK inhibitor SB203580 inhibited tube formation, suggesting that both AMPK and p38 MAPK were involved in enhancing MSS31 tube formation, and both inhibitors did not alleviate the suppressive effect of AdipoRon ( Supplementary Fig. S10). On the other hand, the MEK inhibitor U0126 reversed AdipoRoninduced suppression of tube formation (Fig. 5E). We then incubated MSS31 cells that had been pretreated with leptin for 1 h with AdipoRon in the presence of leptin for up to 6 h and examined ERK1/2 activation. As a result, we found that AdipoRon-activated ERK1/2 was suppressed by leptin (Fig. 5F). These results suggest that leptin alleviates the suppressive effect of AdipoRon on tube formation in MSS31 cells, possibly by inhibiting AdipoRon-induced ERK1/2 activation.

Discussion
We demonstrated here that the orthotopic tumour growth of Panc02-Luc-ZsGreen cells was stimulated by obesity, in good agreement with a previous study 22 . We then investigated the anticancer efficacy of AdipoRon in DIO mice implanted with such cells. The results showed that although AdipoRon at the given dose effectively ameliorated the prediabetic condition in terms of reductions in blood glucose and insulin, the treatment failed to suppress tumour growth. This could be due to the rapid tumour growth of Panc02-Luc-ZsGreen cells in the pancreas, which did not give enough time for the tumour growth suppression by AdipoRon to manifest. However, we realized that glucose, insulin and IGF-1 did not affect the growth of Panc02-Luc-ZsGreen cells even at high concentrations, although these have been reported to influence the proliferation of human pancreatic cancer cells [9][10][11][12][13][14][15][16][17][18] . Thus, we concluded that amelioration of the prediabetic condition, at least in terms of blood glucose and insulin levels, by a high dose of AdipoRon was unable to suppress orthotopic tumour growth of the cells in the present experimental setting. However, it should be noted that this does not necessarily exclude the involvement of glucose, insulin, IGF-I and other diabetes-related factors, such as circulating lipids and cytokines, and changes in the intestinal microbiome observed in diabetic patients 19,20 in the growth of human pancreatic cancer cells. Consistent control of these factors by AdipoRon administration might be able to suppress the growth of pancreatic cancer associated with prediabetes and obesity in humans. A high dose of AdipoRon was also ineffective in inhibiting Panc02 tumour growth in DIO mice as long as they were fed an HFD. However, it significantly suppressed tumour growth when the diet was changed to an LFD, which resulted in the mice losing weight. Histochemical studies indicated an increase in necrotic areas in the tumours of AdipoRon-administered mice fed an LFD. z-VAD-fmk only slightly reversed the AdipoRoninduced growth inhibition, and only a small fraction of tumour cells was positive for CC3 in both the mouse group fed an HFD and the mouse group fed an LFD. These results excluded apoptosis as a main cause of tumour growth inhibition, although caspase-independent apoptosis could not be completely denied. Because AdipoRon inhibited cell proliferation and concomitantly caused necrotic cell death in Panc02 cells in vitro, the increase in necrotic areas could be a direct effect of AdipoRon on Panc02 cells. In addition, because we previously observed the suppression of tumour angiogenesis in human pancreatic cancer tissues of a xenograft mouse model administered AdipoRon 29 , it is also possible that AdipoRon acted on endothelial cells. Impaired tumour angiogenesis should cause a shortage of nutrients and oxygen supply, which brings about necrotic cell death in tumours and a decrease in cell proliferation. Our results showed a significant decrease in the number of microvessels in AdipoRon-administered mice fed an LFD. Coinciding with this, we found that AdipoRon inhibited tube formation of MSS31 cells at a concentration that did not suppress cell proliferation.
To address what kind of obesity-associated factors reduced the anticancer efficacy of AdipoRon, we examined the effect of glucose, insulin, IGF-1 and leptin on the survival of AdipoRon-treated Panc02-Luc-ZsGreen cells in vitro. The results showed that insulin and IGF-1 but not glucose slightly alleviated the suppressive effect of AdipoRon, which indicates that these factors can reduce the effect of AdipoRon in vivo. However, we believe that this is not the main cause because the effect was small, and amelioration of diabetic conditions by AdipoRon marginally affected tumour growth. Leptin neither enhanced the growth of the cells nor interfered with AdipoRon, a finding consistent with the absence LepRb; additionally we excluded palmitic acid as a tumour growth-enhancing factor since it promoted Panc02-Luc-ZsGreen cell death. Of note, as we found sera from HFD-fed mice not to affect the survival of AdipoRon-treated cells, we hypothesized that obesity-related factors may exert their effect through altering the homeostasis of the host rather than cancer cells.
Our earlier 29 and aforementioned findings then promoted us to investigate the role of tumour angiogenesis and investigated the effect of insulin, IGF-1 and leptin on tube formation of MSS31 cells that expressed both AdipoRs and LepRb, as do HUVECs 31 . Leptin is said to be pro-angiogenic in the context of wound healing [32][33][34] and cancer progression 25,32,37 , and serum leptin levels obviously decreased after a change in diet in accordance with previous findings that the leptin blood level increases as body weight increases and is decreased by initial weight loss 24 . This echoed the possibility that leptin affected the antiangiogenic effect of AdipoRon. Indeed, we found that AdipoRon suppressed tube formation of MSS31 cells on Matrigel and leptin alleviated the suppression of tube formation.
The precise mechanisms by which AdipoRon inhibits tube formation and leptin reduces this inhibition remain to be investigated. AdipoRon activated AMPK, p38 MAPK and ERK1/2 in MSS31 cells. We thought that activation of AMPK and p38MAPK enhanced tube formation because activation of AdipoR signalling acts as a survival signal in MIAPaCa-2 cells 29 . In reality, BML-275 and SB203580 inhibited tube formation. On the other hand, U0126 ameliorated AdipoRon-induced suppression of tube formation, indicating that ERK1/2 activation is a cause of the suppression. ERK1/2 activation may overcome AMPK and p38 MAPK activation, leading to the suppression of tube formation. Intriguingly, leptin suppressed AdipoRon-enhanced ERK1/2 activation, in good agreement with the effect of leptin on AdipoRon-induced suppression of tube formation. Leptin has also been demonstrated to activate Notch signalling, PI-3K/Akt signalling and NF-kB signalling 25 . There might be connections between these signalling pathways and AdipoRon signalling pathways. Further studies are necessary to elucidate the mechanisms.
In conclusion, AdipoRon suppressed orthotopic Panc02 tumour growth through inhibition of tumour cell survival and tumour angiogenesis in part via ERK1/2 activation, leading to induction of necrotic cell death in tumours, and this effect was ameliorated by obesity-associated factors (Fig. 6). Among these factors, leptin may play an important role in reducing the anticancer activity of AdipoRon. Obesity and leptin signalling have also been reported to enhance chemoresistance to anticancer drugs in human pancreatic cancer 25 . Therefore, a further understanding of why obesity is associated with chemoresistance might lead to novel treatments and enhance the outcome of current therapies against pancreatic cancer.

Methods
Cells and cell culture. Mouse pancreatic cancer Panc02-Luc-ZsGreen cells expressing firefly luciferase and ZsGreen were established as described previously 38 . The characteristics of mouse endothelial MSS31 cells were described previously [39][40][41] . The cells were usually cultured in high-glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% foetal bovine serum and 40 μg/ml gentamycin in a humidified atmosphere with 21% O 2 /5% CO 2 .   Assessment of insulin resistance (HOMA-IR). Peripheral blood was collected from the tail veins of overnight-fasted mice. After centrifugation of the blood at 1000×g for 15 min at 4 °C, the plasma fraction collected from the supernatant was used to estimate blood insulin levels with an Ultra Sensitive Mouse Insulin ELISA Kit (Morinaga Institute of Biological Science, Inc., Yokohama, Japan). The HOMA-IR value was calculated based on the blood glucose concentration and insulin levels 43 .

Reagent.
Orthotopic tumour injection. For orthotopic implantation, Panc02-Luc-ZsGreen cells (2 × 10 5 cells) were implanted with 50% Matrigel into the pancreas of anaesthetized mice with medetomidine (0.3 mg/kg)/midazolam (4.0 mg/kg)/butorphanol (5.0 mg/kg). The pancreas together with the spleen was exteriorized through a laparotomy, and the cells were injected using a 30-gauge needle attached to an insulin syringe. The pancreas and spleen were returned to the peritoneal cavity, and the incision was closed with surgical staples. The mice and surgical wounds were observed and evaluated once a day until the mice returned to normal behaviour. The staples were removed 10 days after surgery. The mice were further examined daily for their health, including checks of infection, wound dehiscence and excessive weight loss, and, if necessary, received subcutaneous administration of analgesics to minimize pain and distress. Mice were continued on their diet regimen while tumour growth was monitored. Mice were normally euthanized by CO 2 inhalation at the end of a study.
Mice were intraperitoneally administered 5 mg/kg AdipoRon with single daily dosing for 10 days before the glucose tolerance test. Except for the day of Panc02 cell implantation, administration of the same dose of AdipoRon was continued until the end of the experiments. In other experiments, mice were intraperitoneally administered 30 mg/kg AdipoRon every other day after Panc02 cell implantation.
Bioluminescent imaging. In vivo bioluminescent imaging was performed using the IVIS imaging system (Summit Pharmaceuticals International Corp., Tokyo, Japan). All mice were injected intraperitoneally with 150 mg/kg D-luciferin (PerkinElmer, Waltham, MA, USA) and anaesthetized with 2.5% isoflurane. Ten minutes later, photons from animals' whole bodies were imaged using the IVIS imaging system according to the manufacturer's instructions. Data were analysed by living image 2.50 software.