Activation of 5-HT2 Receptors Reduces Inflammation in Vascular Tissue and Cholesterol Levels in High-Fat Diet-Fed Apolipoprotein E Knockout Mice

Coronary artery disease (CAD) is a progressive cardiovascular syndrome characterized by cholesterol-induced focal arterial lesions that impair oxygen delivery to the heart. As both innate and adaptive immune cells play critical roles in the formation and progression of arterial plaques and endothelial cell dysfunction, CAD is commonly viewed as a chronic inflammatory disorder. Our lab has previously discovered that 5-HT2A receptor activation with the 5-HT2 receptor selective agonist (R)-2,5-dimethoxy-4-iodoamphetamine [(R)-DOI] has potent anti-inflammatory activity in both cell culture and whole animal models. Here we have examined the putative therapeutic effects of (R)-DOI in the ApoE−/− high fat model of cardiovascular disease. Subcutaneously implanted osmotic minipumps were used to infuse sustained low rates (0.15 μg / hr) of (R)-DOI∙HCl to mice fed a high-fat “Western” diet. (R)-DOI treated mice had significant reductions in expression levels of mRNA for inflammatory markers like Il6 in vascular tissue, normalized glucose homeostasis, and reduced circulating cholesterol levels. As cardiovascular disease is a leading cause of death both globally and in the Western world, activation of 5-HT2A receptors at sub-behavioral levels may represent a new strategy to treat inflammation-based cardiovascular disease.

Diet and tissue collection. Details of the experimental design are summarized in Fig. 1. Following a two week acclimatization period in which all mice were fed a regular chow diet (Teklad 7012; 5% fat, 19% protein, 5% fat; Harlan Teklad, Madison, WI,USA), animals were divided into 4 groups: ApoE −/− mice fed a regular chow Figure 1. Experimental design and treatment schedule. Following 2 weeks of a normal chow diet, ApoE −/− mice (n = 12) were fed a high-fat (HF; n = 12) diet. 2 weeks post diet change, mice were implanted with an osmotic mini-pump releasing either saline (saline/normal chow, saline/high fat) or (R)-DOI (DOI/normal chow, DOI/high fat). Mice were maintained for an additional 8 week before a second mini-pump was implanted into the opposing flank. 8 weeks post Pump 2 implantation a glucose tolerance test was performed followed by sacrifice and thoracic aorta dissection and blood collection.
diet; mice to receive (R)-DOI fed a normal chow diet; ApoE −/− mice fed a 'Western' high-fat (HF) diet (Teklad 97070; 21% fat by weight, 0.15% cholesterol); and mice to receive (R)-DOI fed a HF diet. Two weeks post HF diet initiation an Alzet mini-osmotic pump (Model 2006; lot no. 10310-13; mean fill volume 237.2 ± 3.0 μl, 0.15 ± 0.01 μl/hr infusion rate; calculated release of 3.6 ± 0.24 μl/24 hours for 65.9 (±0.6) days; DURECT Corp., Cupertino, CA, USA) preloaded with sterile saline (Baxter Healthcare Corp., Deerfield, IL) or (R)-DOI•HCl (2.4 mg/240 μl; 28 mM) dissolved in sterile saline were implanted subcutaneously in ApoE −/− normal chow, (R)-DOI-normal chow, ApoE −/− HF, and (R)-DOI-HF mice to deliver a flow rate k 0 of 1.5 μg/hr. For each surgery, mice were anesthetized with isoflurane (Piramal Healthcare, Andhra Pradesh, India) delivered via nosecone using a vaporizer system (SurgiVet Inc., WI, USA) with isoflurane at 2% in an oxygen flow of 1L/min. Implantation involved making a small incision in the skin on the left flank of each mouse and creating a small subcutaneous pocket using a halsted-mosquito hemostat (FST, Foster City, CA, USA). Pumps were inserted into the pocket of each mouse and the skin incision was closed using 4-0 silk sutures (Oasis Medical, Mettawa, IL, USA). The topical antibiotic Neosporin (Johnson & Johnson Consumer product, Inc., Skillman, NJ, USA) was used post-operatively to prevent infection at the implantation site. Eight weeks post implantation a second osmotic minipump prepared as described above with either sterile saline or drug was implanted subcutaneously into the opposing flank as described above. Animals were maintained an additional eight weeks (16 weeks total drug exposure) before euthanization via a single injection of ketamine/xylazine (60 and 3 mg/kg), thoracic cavity opening, cardiac puncture, exsanguination, and collection of blood. Serum was isolated by centrifuging blood in heparin coated tubes (BD, Billerica, MA, USA), which was flash frozen and stored at −80 °C. Whole aorta samples from the subclavical branch point to the iliac artery were dissected. The aortic arch was then dissected (aortic valve to first branch of aorta) and immediately snap frozen with 2-methylbutane (Sigma, St. Louis, MO, USA, Lot No. SHBJ9704, ≥99%) on dry ice, with storage at −80 °C until further use.
Serum levels of (R)-DOI. Quantification of drug levels in plasma samples was performed at Pharmoptima LLC (Portage, MI) using optimized methods for (R)-DOI with a LLOQ of 0.1 ng/ml. The LC-MS/MS system used a Thermo Finnigan Quantum Ultra triple-quadrupole mass spectrometer coupled to a Thermo Finnigan Accela HPLC system (San Jose, CA). The HPLC system was equipped with a CTC Analytics HTC PAL autosampler (Zwingen Switzerland). Data acquisition and analysis was performed using Xcalibur V2.0.7, and LCquan V2.5.6 software on 11 samples. The overall mean steady-state trough concentration of plasma (R)-DOI (C ss ) was 0.515 ± 0.20 ng/ml (mean ± SEM) (median of 0.385) with values ranging from <LLOQ to 1.29 ng/ml. Glucose tolerance test. Mice were placed in clean cages with no food but free access to water at ~7:00 p.m.
After an overnight starvation (~12 hr; dark cycle), mice were weighed and a baseline blood sample was obtained by tail bleed. Mice were injected i.p. with glucose (1 g/kg body weight), and additional blood samples (~3 μl) were taken at the indicated time points for measurement of glucose concentration using an Accu-Check Mobile glucometer (Roche, Basel, Switzerland).
Lipoprotein measurement. Serum lipids were determined enzymatically using commercial kits for total cholesterol and high density cholesterol (Total Cholesterol E-test and HDL Cholesterol E-test, respectively; Wako Pure Chemical Industries; Osaka, Japan) according to the manufacturer's instructions.
Quantitative real-time PCR. All nucleic acid was extracted and processed as described previously 24 .
Briefly, total cellular RNA was isolated using TRIZOL reagent (Life Technologies, Carlsbad, CA, USA), per the manufacturer's instruction. First-strand cDNA was generated using the ImProm-II cDNA synthesis kit (Promega, Madison, WI, USA), per the manufacturer's instruction, with 500 ng total RNA per reaction. All primers for quantitative real-time PCR (qPCR) were synthesized by Integrated DNA Technologies, Inc. (Corralville, IA, USA), and designed to be compatible with probes from the Universal ProbeLibrary system using the Universal ProbeLibrary Assay Design Center (Roche Diagnostics, Indianapolis, IN, USA). Primer sequences and probes used in this study are listed in Table 1. Quantification of gene expression using a two-step cycling protocol was performed on a Roche LightCycler 480 Instrument II LC (Roche Diagnostics). Relative gene expression was www.nature.com/scientificreports www.nature.com/scientificreports/ calculated using the 2 −ΔΔCt method, with the levels of all targets from genes of interest normalized to internal Gapdh expression as determined using the Mouse Gapdh Gene Assay (Roche) in multiplex format.
Cytometric bead assay. Serum from treatment groups was collected and processed as described above. A cytometric bead array was used to probe for the pro-inflammatory cytokines and chemokines using the Milliplex cytometric bead array kit (Millipore Sigma, Burlington MA Cat # MCYTOMAG-70K-PMX) according to the manufacturer's instructions. The samples were run on a BioRad Bio-Plex 200 system and data was analyzed relative to a 5-parameter logistical standard curve for each corresponding cytokine/chemokine using Bio-Plex Manager 6.1.1.

Statistical analysis.
Statistical analysis was performed using GraphPad Prism v 7.0 software (GraphPad Software, La Jolla, CA, USA). All groups were compared by one-way analysis of variance (ANOVA), followed by Tukey post-test. Results are expressed as mean ± standard error of mean (SEM). P < 0.05 was considered to be significant.

5-HT 2A receptor activation via (R)-DOI does not impact food intake or body weight.
Early studies have indicated that several 5-HT 2A agonists like LSD have an anorexigenic effect 30 . Further, Lorcaserin, a 5-HT 2C preferential agonist with 5-HT 2A activity, has been shown to reduce body weight in obese men and women 31 , a property which is attributed to direct activation of 5-HT 2C receptors in the central nervous system (CNS) 32 . To test if sustained delivery of low-dose (R)-DOI had an appetite suppressing effect in our model, food intake was measured during the final week of experimentation. No significant decrease in food intake was observed between experimental groups given a high fat diet (data not shown). As expected, the mice fed a HF diet demonstrated significantly increased body weight (Fig. 2). HF fed mice implanted with (R)-DOI osmotic minipumps showed a trend for a very slight decrease in weight, but this was not significant (Fig. 2). Taken together these data indicate that sustained delivery of low-dose (R)-DOI results in drug levels that do not influence food intake or body weight.
(R)-DOI increases glucose tolerance in ApoE −/− mice fed a high fat diet. A HF diet in mice has been shown to induce glucose intolerance 15,33 , a major predictor of metabolic and cardiovascular disease in humans 34 . To evaluate the impact of (R)-DOI on glucose sensitivity, we measured blood glucose concentrations in fasting mice after 16 weeks for treatment groups (Fig. 3). (R)-DOI improved glucose tolerance in both HF fed and normal chow fed treatment groups, indicating that 5-HT 2 receptor activation can modulate the impaired glucose tolerance induced by a Western diet.

5-HT 2 receptor activation lowers total and HDL cholesterol levels in Apoe
mice fed a HF diet exhibited total cholesterol levels that were roughly three times higher than ApoE −/− mice fed a normal chow diet (Fig. 4A). The presence of (R)-DOI modestly but significantly reduced total serum cholesterol levels. In a similar manner, HF diet elevated LDL cholesterol levels (Fig. 4B) was also moderately reduced in ApoE −/− mice treated with (R)-DOI.
(R)-DOI reduces the expression of proinflammatory marker genes in the aortic arch.
Inflammation is strongly linked to the development of numerous cardiovascular disorders 35 . We have previously www.nature.com/scientificreports www.nature.com/scientificreports/ shown that (R)-DOI potently blocks TNF-α-induced mRNA expression of pro-inflammatory cell adhesion genes (vcam-1), cytokines (Il-6), and chemokines (mcp-1) in aortic tissue 24 . Here we wanted to examine proinflammatory gene expression in the context of HF diet-induced inflammation, and the effect of (R)-DOI on this expression. Thoracic aorta were dissected and the expression of several inflammation-related genes was evaluated (Fig. 5). mRNA for Il-6, vcam1, and tnf-α were all elevated in the HF fed group, indicating the presence of vascular inflammation. (R)-DOI treatment significantly decreased the expression of all three mRNAs. There was a trend for increased mcp-1 expression in the HF fed animals that was prevented by (R)-DOI, however, these changes were nonsignificant.
HF diet alters CXCL10 expression in the serum. In contrast to tissue-specific upregulation of inflammatory cytokines, circulating serum cytokine levels are known to be only modestly impacted by HF diet feeding [36][37][38] . Cytometric bead array analyses on blood serum collected from the different treatment groups revealed equivalent serum cytokine levels for most markers tested (data not shown). One notable exception was the T cell regulatory chemokine C-X-C motif chemokine 10 (CXCL10) 39 , a biomarker for development of heart failure and adverse cardiac remodeling. CXCL10 was found to be upregulated in HF-diet fed mice, whereas (R)-DOI treated mice exhibited reduced levels that were similar to NC animals (Fig. 6).

Discussion
The standard, 'Western diet' overloads blood with cholesterol, triglycerides, and glucose. This diet can result in altered glucose homeostasis, metabolic disease, and cardiovascular dysfunction. Traditional 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors, or statins, represent the gold-standard in lipid-lowering therapy due to their efficacy at reducing atherosclerotic plaques. However, muscular symptoms and/or elevation of hepatic aminotransferases during statin therapy are serious side effects that contribute to statin intolerance in a sizable population of patients 40 . For statin-intolerant patients, alternative  www.nature.com/scientificreports www.nature.com/scientificreports/ therapies first start with dosage lowering or alternative dosing regimens. While these treatments can result in reduced myopathies and liver toxicity, their ability to provide a long-term reduction in cardiac risk has not been established. Furthermore, the use of non-statin lipid-lowering drugs constitutes an underutilized approach to hyperlipidemia. Preliminary results into the LDL-C reducing potential of various nutraceuticals and supplements is promising, but further study is necessary to confirm their effectiveness and safety 41 .
A recent study has found that cytokine-based/anti-inflammatory therapy may be more efficacious in preventing cardiovascular disease than cholesterol-lowering compounds 42 . Results from the CANTOS (Canakinumab Anti-Inflammatory Thrombosis Outcome Study) trial show that the IL-1β inhibitor canakunumab, independent of lipid-lowering, significantly reduces the risk of cardiovascular disease and recurrent cardiovascular events in patients with previously diagnosed myocardial infarction 43 . Statins themselves have been shown to possess a variety of anti-inflammatory properties, which has led to speculation that the inflammatory inhibiting effects of statins may be equally if not more important for their cardioprotective effects than simple cholesterol lowering 44 . Thus, when designing a drug to be an alternative to statins, any compound that lowers cholesterol levels while concurrently having anti-inflammatory properties is highly desirable.  www.nature.com/scientificreports www.nature.com/scientificreports/ We have previously shown that (R)-DOI has potent anti-inflammatory properties in primary vascular cells, against TNF-α-mediated inflammation in the aortic arch, and in a model of allergic asthma 45,24,46 . The levels of (R)-DOI necessary to produce these anti-inflammatory effects are low, and in some instances far below those necessary to produce behavioral effects. Frequent consumption of a "Western" diet is known to promote systemic inflammation and immune cell dysregulation 47 . HF-diet feeding has been shown by others to induce the expression of a number of proinflammatory factors that we have previously found to be elevated in our TNF-α and/or ovalbumin models (i.e. IL-5, IL-6, IL-13, MCP-1, GM-CSF) 48 ; therefore, we examined the ability of (R)-DOI to protect against the effects of a high fat diet in the ApoE −/− mouse model. We chose the ApoE −/− model because this mouse robustly develops metabolic and vascular disease in response to being fed a "Western" style diet. As expected, we found that expression of proinflammatory genes, including the cell adhesion molecule vcam-1 and the cytokine Il-6, were significantly elevated in aortic tissues following 16 weeks of high-fat diet feeding (Fig. 5). Significantly, (R)-DOI prevented high fat diet-induced expression of several inflammatory markers, suggesting that 5-HT 2 receptor activation is a viable strategy to combat diet-induced vascular inflammation. An interesting exception was for tnf-α, where the HF diet did not induce mRNA expression, but the combination of the HF diet and (R)-DOI resulted in significantly lowered tnf-α expression below that of even control. Another interesting result is the upregulation of the T-cell chemokine and cardiovascular disease biomarker CXCL10 in our HF-diet fed animals (Fig. 6). While the standard Western diet increased the overall expression of inflammatory markers in vascular tissues, its impact on circulating cytokines is minimal 36 . However, higher concentrations of CXCL10 (IP-10) have been found in the plasma of patients with coronary artery disease 49 , which has been theorized to modulate the balance of effector and regulatory T cells in atherogenesis 39 .
As mentioned above the 5-HT 2A receptor is present in cardiovascular tissue critical to autonomic functioning (vascular smooth muscle, endothelial cells, cardiomyoctyes) and the immune cell populations that resides in cardiac tissue (mononuclear phagocytes, neutrophils, B and T cells, macrophages) 13,50 . Accordingly, the receptor for CXCL10, CXCR3, is also expressed both in non-immune (endothelial and smooth muscle cells) and immune (T lymphoctyes, natural killer cells, monocytes) cardiovascular tissue 51 , with CXCL10 binding to CXCR3 mediating a plethora of cell functions, including chemotaxis, proliferation, migration and survival. As both 5-HT 2A and CXCR3 receptors reside on both these immune and non-immune cell populations, it's possible a dynamic interplay exists between 5-HT 2A and CXCR3 receptor activation. Therefore, it is conceivable that the primary targets of (R)-DOI in this model are immune cells. The finding that elevated levels of CXCL10 induced by high fat feeding are reduced to control levels in mice administered (R)-DOI supports the notion that activation of 5-HT 2A receptors on discreet subsets of lymphocytes modulates vascular inflammation in a therapeutic manner. Further studies are necessary, however, to determine the role of (R)-DOI-mediated 5-HT 2 receptor activation on these different cellular populations.
Impaired glucose tolerance is a major risk factor for adverse cardiovascular events 52,53 . In particular, population studies have revealed that impaired glucose metabolism combined with dyslipidemia are major risk factors for cardiovascular disease 54 . Here we demonstrated that (R)-DOI improved glucose tolerance for both HF and normal chow fed mice. We cannot yet say what the underlying mechanism of this effect is, but previous work by others has shown that beta-islet cells in the pancreas express 5-HT 2A receptors and that agonism of these receptors can increase insulin secretion 55 .
Interestingly, (R)-DOI decreased both total cholesterol and LDL cholesterol in HF-diet fed mice. The reduction in total and LDL cholesterol suggests a possible protective therapeutic effect for (R)-DOI against hypercholesterolemia; however, further experiments are necessary to fully understand the effects of this drug on cholesterol and how this may impact therapeutic strategies. Interestingly, cholesterol is known to directly interact with the 5-HT 2A receptor to induce increased receptor conformational variability 56 . One may speculate that conformational changes induced by cholesterol contribute to its anti-inflammatory effect by either enhancing interactions with anti-inflammatory signaling pathways or preventing interactions with downstream pro-inflammatory signaling cascades. This, however, will also require further experimentation. Polymorphisms at the HTR2A locus have been shown to be associated with cholesterol levels 57 . Therefore, 5-HT 2A receptor function in general may modulate other aspects of lipid homeostasis and that activation with (R)-DOI is affecting these processes.
Whereas (R)-DOI is an agonist of 5-HT 2 receptors, previous work by others has demonstrated that antagonists for these receptors can protect against vascular inflammation. For example, the 5-HT 2 receptor antagonist sarpogrelate retards the progression of atherosclerosis in rabbits 58 . We speculate that while the effects of (R)-DOI are active mediation of anti-inflammatory processes, 5-HT 2 receptor antagonists may merely be blocking the well-established proinflammatory effects of serotonin. For example, 5-HT is known to have proliferative effects on vascular smooth muscle cells and macrophages. Sarpogrelate may simply be blocking the effects of 5-HT on these cells and preventing inflammation induced proliferation, resulting in protection against high fat diet-induced atherosclerosis and vascular inflammation. Based on our previous studies on the ability of (R)-DOI to prevent vascular-related cell and tissue inflammation induced by TNF-α, which is a key pro-inflammatory cytokine in atherosclerosis and vascular inflammation, via 5-HT 2A receptor activation we propose the following model. Sub-behavioral levels of systemic circulating (R)-DOI activate 5-HT 2A receptors to induce anti-inflammatory pathways that include blocking the expression of IL-6, VCAM-1, CXCL10, and TNF-α from vascular endothelial and smooth muscle cells as well as macrophages that ultimately limit high fat-induced vascular inflammation and recruitment of macrophages to the aorta that would otherwise differentiate to foam cells producing tissue damage and more inflammation. This reduced vascular inflammation may also include a component resulting from the observed decrease in total plasma and LDL cholesterol by drug treatment. Consistent with our proposed model, a recent study found that the antipsychotic drug olanzapine, a 5-HT 2A receptor inverse agonist, increases serum levels of total cholesterol, non-HDL, HDL-c, and triglycerides, deregulates hepatic lipid metabolism, and increases aortic proinflammatory protein expression (VCAM-1, TNF-α, and IL-6) in apoE −/− mice 59 . Along a similar vein, the 5-HT 2C selective agonist lorcaserin effectively reduces appetite to induce weight loss, a result we