Interactions between Caveolin-1 polymorphism and Plant-based dietary index on metabolic and inflammatory markers among women with obesity

A series of recent studies have indicated that the Caveolin-1 (CAV-1) gene variant may be associated with metabolic and inflammatory markers and anthropometric measures. Furthermore, it has been shown that a plant-based dietary index (PDI) can elicit a positive impact on these metabolic markers. Therefore, we sought to examine whether PDI intakes may affect the relationship between CAV-1 (rs3807992) and metabolic factors, as well as serum inflammatory markers and anthropometric measures, in women with obesity. This current study consisted of 400 women with overweight and obesity, with a mean (SD) age of 36.67 ± 9.10 years. PDI was calculated by a food frequency questionnaire (FFQ). The anthropometric measurements and serum profiles were measured by standard protocols. Genotyping of the CAV-1(rs3807992) was conducted by the PCR–RFLP method. The following genotypic frequencies were found among the participants: GG (47.8%), AG (22.3%), and AA (2.3%). In comparison to GG homozygotes, risk-allele carriers (AA + AG) with higher PDI intake had lower ALT (P: 0.03), hs-CRP (P: 0.008), insulin (P: 0.01) and MCP-1 (P: 0.04). Furthermore, A-allele carriers were characterized by lower serum ALT (P: 0.04), AST (P: 0.02), insulin (P: 0.03), and TGF-β (P: 0.001) when had the higher following a healthful PDI compared to GG homozygote. Besides, risk-allele carriers who consumed higher unhealthful PDI had higher WC (P: 0.04), TC/HDL (P: 0.04), MCP-1 (P: 0.03), and galactin-3 (P: 0.04). Our study revealed that A-allele carriers might be more sensitive to PDI composition compared to GG homozygotes. Following a healthful PDI in A-allele carriers may be associated with improvements in metabolic and inflammatory markers and anthropometric measures.

www.nature.com/scientificreports/ outside of this range leading to exclusion from the study 38 . All of the participants completed a written informed consent form before taking part in the study.
General, anthropometric and physical activity assessments. Height was measured using Seca 216 to the nearest 0.1 cm and weight of subjects was also assessed using Seca scale to the nearest 0.1 kg, with light clothes and unshod, in a standing position. WC was measured at the point of the smallest girth by a trained expert. BMI was calculated as weight/(height) 2 . In addition, other anthropometric variables were measured using a multi-frequency bioelectrical impedance analyzer InBody 770 scanner (Inbody Co., Seoul, Korea), with participants wearing light clothes without any metal subjects 39 . In addition, physical activity (PA) was assessed via the validated International Physical Activity Questionnaire 40 . All participants' blood pressure was monitored using an automatic sphygmomanometer according to established procedures (OMRON, Germany).

Dietary assessments and plant-based dietary intake.
A validated 147-items semi-quantitative food frequency questionnaire (FFQ) was used by an expert to assess the dietary intake of the subjects. Grams of foods and beverages were entered into NUTRITIONIST IV software (version 7.0; N-Squared Computing, Salem, OR) 41,42 . According to the plant-based dietary intake, which was divided to overall plant-based diet index (PDI), healthy PDI (hPDI), and unhealthy PDI (uPDI), the intakes of the subjects were divided to 18 groups, which were: animal (Dairy, animal fat, egg, meat, fish and seafood, and miscellaneous animal-based foods), healthy (whole grains, vegetables, fruits, legumes, nuts, vegetable oils, tea, and coffee), and unhealthy plant-based foods (fruit juices, sugar-sweetened beverages, refined grains, sweets and desserts, and potatoes) 43,44 . Foods were scored 1-10, after they were transformed to deciles. PDI, hPDI, and uPDI were assessed as scoring 10 and 1 for the highest and lowest deciles of plant food intake, healthy plant foods, and unhealthy plant foods, respectively. Conversely, unhealthy plant foods and animal food items were scored 1 to 10 for the highest and lowest deciles of animal foods and unhealthy plants foods, respectively.
Biochemical measurements. After 10-12 h night fasting, blood samples were gathered at the Nutrition and Biochemistry laboratory of the School of Nutritional Sciences and Dietetics, TUMS. Triacylglycerol kits (Pars Azmoon Inc, Tehran, Iran) were used to measure serum triglycerides (TG) by colorimetric method tests with Glycerol-3-phosphate oxidase Phenol 4-Aminoantipyrine Peroxidase (GPO-PAP). The cholesterol oxidase Phenol 4-Aminoantipyrine Peroxidase (CHOD-PAP) was used to calculate total cholesterol, and the direct method and immunoinhibition were used to measure low-density lipoprotein (LDL) and high-density lipoprotein (HDL). Pars Azmoon (Pars Azmoon Inc., Tehran, Iran) provided all of the kits. ALT, AST, high-sensitivity C-reactive protein (hs-CRP), MCP-1, interleukin 1 beta (IL-1β), and transforming growth factor-beta (TGFβ) were measured via standard protocols. Plasminogen activator inhibitor-1 (PAI-1) (Human PAI-1*96 T ELISA kit Crystal Company) was measured in triplicate. TC was divided into HDL to discern the ratio of TC/HDL, in addition, LDL was divided by HDL to discern the ratio of LDL/HDL. Serum insulin level was also measured by radioimmune assay. More detailed about biochemical measuring are reported in our previous study 45  Statistical analyses. The Kolmogorov-Smirnov test was used for assessing the normality of the data. The Hardy-Weinberg equilibrium and comparison of categorical variables were assessed with the x 2 test. PDI was transformed to tertiles based on the trends. Quantitative (mean ± SD) and qualitative variables (n (%)) were evaluated via one-way analysis of variance (ANOVA) and chi-square among PDI tertiles, respectively. Genotype groups were considered as a dominant inherent model (AA + AG) versus GG homozygous. Analysis of covariance (ANCOVA) was performed for the adjustment model (adjusting for age, physical activity and energy intake and diastolic blood pressure (DBP)). A generalized linear regression model (GLZM) was used to analyze the interactions between CAV-1 polymorphism (rs3807992) and PDI. Data were analyzed using Statistical Package for Social Sciences (SPSS Inc., Chicago, IL, version 25) and a P-value < 0.05 was considered as significant; but for interactions, P < 0.1 was considered significant.
Ethics approval and consent to participate. All methods were performed in accordance with the relevant guidelines and regulations. The protocol of the study was approved by the ethics committee of TUMS (Ethics number: 97-03-161-41017). All participants completed a written informed consent.

Results
Study population characteristics. The  www.nature.com/scientificreports/ eters according to two genotypes groups (GG and AG + AA) is given in Table 1. The results of this study revealed that DBP (P = 0.007), body fat mass (BFM) (P = 0.01), and fat mass index (FMI) (P = 0.009) were higher in participants carrying the A allele, compared with individuals in the GG genotype. Moreover, A-allele carriers had significantly lower serum HDL (P < 0.001) and LDL (P = 0.01) compared to GG homozygotes; although, the TC/HDL (P = 0.008) ratio was significantly higher. Furthermore, there was no significant association between this polymorphism and other biochemical parameters including TG, inflammatory markers (hs-CRP, MCP-1, TGFβ), and liver enzymes (ALT and AST) in both crude and adjusted models (P > 0.05) ( Table 1).
The difference in means of biochemical parameters and body composition between PDI, hPDI and uPDI. The result of comparison of biochemical parameters and body composition between PDI, hPDI and uPDI are shown in Tables 2, 3 and 4. These results displayed that women with higher PDI were older (P = 0.03) and had higher physical activity (P < 0.001) than those with lower PDI. Women with higher healthful www.nature.com/scientificreports/ PDI had significantly lower LDL/HDL ratios in both crude (P = 0.04) and adjusted models (P = 0.03); although there was no statistically significant difference (P > 0.05) from the low to high unhealthful PDI.

Interaction between PDI and CAV-1 rs3807992 genotypes on metabolic and inflammatory markers and anthropometric measures.
There is a gene-diet interaction for PDI and CAV-1 polymor-  (Table 5). Thus, the A-allele carriers who consumed higher PDI had lower ALT, insulin, MCP-1, and hs-CRP, compared to GG homozygotes (Fig. 1).  (Table 6). In particular, A-allele carriers were characterized by lower serum ALT, AST, insulin, and TGF-β, when following a healthful PDI, compared to GG homozygote. (Fig. 2).  Table 6 and Fig. 3.

Discussion
To our knowledge, this is the first study to have investigated the interactions between genetic variants of the CAV-1 gene with PDI on metabolic and inflammatory markers among Iranian women with overweight and obesity. www.nature.com/scientificreports/ Our findings highlight a gene-diet interaction for PDI and CAV-1 polymorphism (rs3807992) on ALT, hs-CRP, insulin, and MCP-1 in both crude and adjusted models; such that the A-allele carriers who had higher PDI, had lower ALT, insulin, MCP-1, and hs-CRP compared to GG homozygotes. Based on our previous findings 46 , CAV-1 (rs3807992) might be related to incremented metabolic disease risk factors in women with overweight and obesity. We also previously posited that among minor allele carriers, insulin pathways may responsible for the relationship between CAV-1 rs3807992 and metabolic factors 46 .
In the present study, DBP, BFM, and FMI were higher in participants carrying the A allele compared to those in the GG genotype. There are some possible functional roles for CAV-1 may be inherited. It has been established by previous studies that CAV-1 has a major regulator function in fat distribution and genetic lipodystrophies in humans 47,48 . Furthermore, experimental studies have reported an association between CAV-1 mRNA expression in adipose tissues in obese women compared with lean subjects 49,50 . CAV-1-deficient mice had a smaller lean body phenotype than their wild-type counterparts 51,52 . Lipodystrophy has been observed in CAV-1 null mice due to a variety of functions attributed to caveolae in adipocytes, including lipid droplet dysfunction, adipocyte differentiation pathway disruption, abnormalities in cholesterol and fatty acid binding, transport, and storage, and an increase in insulin signaling 47,53 .
Moreover, A-allele carriers had significantly lower serum HDL and LDL compared to GG homozygotes. Additionally, there was no significant relationship between this polymorphism and other biochemical parameters containing TG, inflammatory markers (hs-CRP, MCP-1, PAI-1, and IL-1β), and liver enzymes in both crude and adjusted models. Concordant with our study, Khatibi et al. also found a significant association between CAV-1 genotypes with DBP 26 , and observed that participants with prevailing alleles had a lower risk of dyslipidemia, and that risk allele carriers, who adhered more to the Mediterranean-DASH Intervention for Neurodegenerative Delay (MIND) diet, might present with lower dyslipidemia 26 .
In our study, we also observed a significant interaction between hPDI and rs3807992 on ALT, AST, insulin, and TGF-b in both crude and adjusted models. Indeed, A-allele carriers who had higher hPDI scores, had lower serum ALT, AST, insulin, and TGF-b compared to GG homozygote. Further, risk-allele carriers who had higher uPDI scores, had higher WC, TC/HDL, MCP-1, and galactin-3. No significant interactions were observed between CAV-1 rs3807992 variants and uPDI for other metabolic-related traits. A cross-sectional study in 2021 revealed that, among people following antioxidant-rich diets, ' A' allele carriers might be more vulnerable to cardiovascular disorders; the authors also reported no significant relationship between higher adherence to dietary total antioxidant capacity (DTAC) and changes in lipid profile, HOMA-IR, and body composition components 27 .
It is well accepted that the prevalence of obesity is directly related to dietary patterns 54 . PBD alleviates body fat due to the decreased caloric intake and incremented energy expenditure because of augmented thermogenesis 55 . Reduced use of saturated fats, which are usually derived from animal-based foods, might improve insulin sensitivity 55 ; indeed, plant-based foods are also a main source of phytochemicals 56 . The consumption of phytochemicals, notably polyphenols, which exist in a variety of plant foods (e.g., berries, grapes, onions, apples, Table 5. Interactions between CAV-1 rs3807992 and PDI on metabolic and inflammatory markers and anthropometric measures. Values are represented as β (95%CI). WC waist circumference, TC total cholesterol, TG triglyceride, HDL high density lipoprotein, LDL low density lipoprotein, AST aspartate aminotransferase, ALT alanine aminotransferase, hs-CRP high-sensitivity C-reactive protein, MCP-1 monocyte chemoattractant protein-1, TGFβ transforming growth factor beta. A significant p-values are indicated in bold (significance considered p < 0.05). GLMS was performed to identify significant differences between median of PDI and CAV-1 rs3807992. P* = with unadjusted (crude) model. P** = with adjustments for potential confounding factors including (age, energy intake, physical activity education, job and marriage status). We created an interaction model by using genotype categories and dietary intake groups (PDI, hPDI and UPDI) as fixed variables and confounding markers as covariates in GLMS. GG genotype has 0 risk allele. AG genotype has one and AA genotype have two risk allele. GG genotype is considered as a reference. Low adherence of PDI is considered as a reference. www.nature.com/scientificreports/ cacao, green tea, soy, whole grains, etc.), is related to reduced mortality and chronic disease risk [57][58][59][60] . The food compound of an uPDI might be associated with higher intakes of unpleasant nutrients and lower intakes of micronutrients and antioxidants, which can negatively affect metabolic syndrome and its factors. A high intake of simple carbohydrates from uPDI could also affect glucose control, lipid metabolism, and weight gain 61 , whilst decreased dietary fiber might impact glycemic control, insulin sensitivity, and augment inflammation. These effects could be related to reduced inflammation and oxidative stress 62,63 . Chronic inflammation is often evident in those with poor feeding habits and a sedentary lifestyle, features which are concomitantly associated with obesity progression 64,65 , in addition to various other pathologies, including asthma and Alzheimer disease, and different diseases related to unbalanced metabolisms such as T2DM, atherosclerosis, and cardiovascular diseases [66][67][68][69] . Metabolic inflammation consists of a complex mechanism containing crosstalk between different tissues (like adipose tissue and liver) through the whole body. Generally, this low-grade inflammation emerges when cellular stress is distinguished by the immune system 70 . Another mechanism posited to contribute to the progression of chronic inflammation involves the redundant storage of triglyceride (TG) lipids within adipose tissues. In murine models, exceeding TG storage in white adipose tissue (WAT) induces secretion of pro-inflammatory adipokines, such as IL-1, TNF-α, MCP-1, and IL-6, triggering systemic metabolic inflammation 71 .
Although the exact mechanism of the interaction between CAV-1 and diet has not yet been fully elucidated, some explanations have been proposed. Several studies have expressed that CAV-1 binds to endothelial nitric oxide synthase (eNOS) and HDL receptors in the caveolae and prevents their activity. Some diets, such as the PBD or anti-inflammatory diets, can dislocate CAV-1 from caveolae to the cytoplasm, leading to a reduction in the CAV-1 level, ameliorating the prohibitory effects on HDL and eNOS receptors 72 . Decreased nitric oxide (NO) generation, as a result of increased CAV-1 expression, has been thought to occur from prolonged exposure to high glucose, may play a vital role in inflammatory pathways and expansion of inflammation 73 . Hence, it is not surprising that higher adherence to a uPDI can augment MCP-1 by changing the expression of CAV-1 and other genes.
Diet and plasma-derived nutrients may influence metabolic markers by interacting with caveolae-associated cellular signaling, according to new research 36 . In this regard, Oberleithner et al. suggest that serum sodium and potassium can control eNOS binding to the caveolae membrane and activity 74 . As a result, the beneficial benefits of a healthy PDI might be attributed to components like vegetables and fruits, which have a greater impact on potassium and sodium balance. Table 6. Interactions between CAV-1 rs3807992 and hPDI and uPDI on metabolic and inflammatory markers and anthropometric measures. Values are represented as β (95%CI). WC waist circumference, TC total cholesterol, TG triglyceride, HDL high density lipoprotein, LDL low density lipoprotein, AST aspartate aminotransferase, ALT alanine aminotransferase, hs-CRP high-sensitivity C-reactive protein, MCP-1 monocyte chemoattractant protein-1, TGFβ transforming growth factor beta. A significant p-values are indicated in bold (significance considered p < 0.05). GLMS was performed to identify significant differences between median of PDI and CAV-1 rs3807992. P* = with unadjusted (crude) model. P** = with adjustments for potential confounding factors including (age, energy intake, physical activity education, job and marriage status). We created an interaction model by using genotype categories and dietary intake groups (PDI, hPDI and uPDI) as fixed variables and confounding markers as covariates in GLMS. GG genotype has 0 risk allele. AG genotype has one and AA genotype have two risk allele. GG genotype is considered as a reference. Low adherence of hPDI and uPDI are considered as a reference. www.nature.com/scientificreports/ Beside, the lower liver enzyme in A-allele carriers following an hPDI in the present study can be dependent on the caffeine intake during the hPDI. In this term, we previously found a decreased AST in A-allele carriers who consumed more caffeine in a healthy dietary pattern 45 .Interestingly, a candidate gene for the favorable impact of caffeine consumption is located near the CAV-1 gene 75 , Therefor, we suggested a possible interaction between caffeine and CAV-1, which we hope will be confirmed in future study.
In the current study, we observed a relationship between CAV-1 with body composition, which could suggest a CAV-1 lipogenic pathway interaction. The expression of CAV-1 is found in most normal organs but notably assuming that CAV-1 is greatly expressed in adipose tissue 76 and the interplay between CAV-1 gene and lipogenic genes has been expressed 77 . Visceral fat could also be associated with changes in circulating fatty acid composition 78 . CAV-1 interacts with two popular receptors that augment oxidative stress: the angiotensin II receptor and mineralocorticoid receptor (MR) 79 . MR blockade has been shown to reduce NADPH Oxidase 4 (NOX4) expression and lower reactive oxygen species in the kidney and heart 80,81 .
Several studies have revealed that CAV-1 is associated with oxidative stress, as reactive species could impact the expression, degradation, post-translational changes, and trafficking of Caveolae membranes 82 . Hence, because of the impact of the CAV-1 gene with oxidative stress, it seems that adherence to hPDI could be capable of lowering the metabolic markers, and, as a result, reducing the likelihood of metabolic diseases in women with Although we provide novel findings of gene-diet interactions, some limitations should be considered in the interpretation of the study. First, the results of this cross-sectional study, although nationally illustrative, cannot indicate a causal relationship. Second, even though we applied a validated FFQ, measurement errors are possible. Third, our study only included women, therefore, results are not generalizable to men. Finally, although we considered potential confounders, residual confounding may still exist. We did not have any data related to family history of cancer; however, the expression of CAV-1 can be effected by this cofounding marker.

Conclusion
In conclusion, the results of the current study suggest that diet, gene variants, and their interaction, should be considered in metabolic diseases risk assessment. PDIs may provide advantages in the prohibition of chronic disease. In future studies, investigating samples from a greater geographic area with wide ranges of ages in both sexes and larger sample sizes might prove more important findings. This work clearly highlights the value of considering genotypes in dietary planning and modification. Therefore, these findings from nutrigenetic studies in future may be combine with a patient's genetic history to give more relevant and customized nutritional recommendations for women with obesity in order to avoid or reduce cardiovascular disease.