Effects of Smoking on the Gut Microbiota in Individuals with Type 2 Diabetes Mellitus

Smoking affects eating habits; however, few studies on smoking and the gut microbiota have reported the effects of diet in detail. This cross-sectional study aimed to determine the association between smoking and the gut microbiota, considering the impact of smoking on dietary intake. Dietary habits and the composition of the gut microbiota were assessed in 195 men with type 2 diabetes (164 non-current smokers and 31 current smokers) using a brief self-administered diet history questionnaire and 16S ribosomal RNA gene sequencing of fecal samples. The data were compared according to the current smoking status of the participants. Current smokers had high alcohol and sugar/sweetener intake and low fruit intake. The proportion of the Coprococcus genus was higher among current smokers. Multiple regression analysis adjusted for current smoking, age, exercise habits, alcohol intake, sugar and sweetener intake, and fruit intake showed that smoking was associated with the proportion of the Coprococcus genus. Current smoking was associated with both dietary intake and composition of the gut microbiota. Although dietary intake should be considered when investigating the association between smoking and the gut microbiota, the results suggest that the direct effect of smoking is more significant.


Introduction
Cigarette smoking increases the risk of incident diabetes mellitus and exacerbates the microvascular and macrovascular complications of diabetes mellitus [1]. Smoking has also been shown to increase insulin resistance [2]. Moreover, smoking causes systemic inflammation and oxidative stress [3], which not only reduces vascular endothelial function [4] but also directly damages pancreatic β-cell function [5].
Smoking has also been shown to affect the gut microbiota [6]. Gut microbiota affects the regulation of nutritional metabolism [7] and of the immunological and defense mechanisms in the body [8]. Gut dysbiosis participates in the development and progression of many diseases, including type 2 diabetes mellitus [9] and inflammatory bowel disease [10]. Many factors influence the gut microbiota, including diet, lifestyle, smoking, medications, and the genetic background of the host. Weight gain often occurs after smoking cessation [11], and changes in the gut microbiota have been suggested to play a role in this

Study Population
This study was designed to determine the association between diabetes mellitus, the gut microbiota, and various background factors [26,27]. The participants were outpatients at Kyoto Prefectural University of Medicine (KPUM) Hospital and Kameoka Municipal Hospital, Japan. Fecal samples were collected from 522 individuals (17 patients with type 1 diabetes, 383 patients with type 2 diabetes, 8 patients with other types of diabetes, and 114 individuals without diabetes) between November 2016 and October 2018, who had not received antibiotics within the prior 3 months. Individuals without diabetes and those with diabetes other than type 2 diabetes were excluded from the study.

Data Collection
Subject data were collected from medical records on sex, age, blood pressure, height, body weight, body mass index (BMI), duration of diabetes, and available medications for dyslipidemia, hypertension, diabetes, and proton pump inhibitor use. In the questionnaire, participants who reported they were currently smoking even one cigarette per day were categorized as current smokers, whereas those who reported having smoked in the past and who reported never having smoked before were categorized as non-current smokers. The daily number of cigarettes smoked and the years of smoking history were also assessed. Participants who performed some type of sporting activity at least once a week were defined as regular exercisers [28].
Habitual dietary intake data were assessed using a brief self-administered diet history questionnaire (BDHQ). The BDHQ can be used to estimate the dietary intake of 58 food and beverage items in the previous month. The BDHQ has been validated previously [30,31]. In the present study, we excluded patients who did not respond to the BDHQ and those who reported extremely high (>4000 kcal) or low (<600 kcal) energy intake, because of poor reliability [32]. When evaluating energy, protein, fat, and carbohydrate intake, the intake per kilogram of ideal body weight (IBW) was calculated to consider the differences in body size. IBW was calculated as 22× the square of the participant height (m 2 ) [33]. Habitual alcohol intake was defined as alcohol consumption >20 g/day in the BDHQ analysis [34]. Fecal samples were collected and analyzed for gut microbiota composition using previously described methods [35][36][37]. Preservation of the collected fecal samples was performed using a guanidine thiocyanate solution (Feces Collection kit; Techno Suruga Lab, Shizuoka, Japan). Genomic DNA was isolated and purified using a NucleoSpin Microbial DNA kit (Macherey-Nagel, Düren, Germany) and Agencourt AMPure XP beads (Beckman Coulter, Brea, CA, USA).
Sequencing libraries were generated by a two-step polymerase chain reaction (PCR) of the purified DNA samples and the prepared libraries of 250 paired-end sequences were sequenced using the MiSeq Reagent v3 kit and MiSeq (Illumina, San Diego, CA, USA), at Takara Bio's Biomedical Center, as described in a previous study [37].
The DADA2 plugin of Quantitative Insights into Microbial Ecology 2 version 2019.4 was used for table generation for amplicon sequence variants (ASVs) [26] by setting and performing noise removal by DADA2 with the trimming length from the left set at 17 and from the right at 19. Both reads were truncated to <250 base pairs. Each ASV was classified using the sklearn classifier algorithm against Greengenes database version 13_8. Singletons and ASVs assigned to chloroplasts and mitochondria were deleted in this study. Phylogenetic trees were created using SATé-compatible phylogenetic arrangements [38]. Consequently, 6902 ASVs were obtained. Functional profiles from the 16S rRNA dataset were predicted using Phylogenetic Investigation of Communities by Reconstruction of Unobserved States version 2.1.4 [35], as previously described [37].

Statistical Analysis
Clinical characteristics, nutrient intake, food group intake, and the proportions of phyla and genera in the gut microbiota were compared between the non-current and current smoker groups using the Mann-Whitney U test. Multiplicity was adjusted by obtaining q values using the Benjamini-Hochberg method to compare the proportions of the gut microbiota.
Furthermore, to assess the effects of smoking on the composition of the gut microbiota, a multiple regression analysis was performed using smoking, smoking-related background factors, and dietary intake as covariates. To investigate multicollinearity, the variance inflation factor (VIF) was checked and all VIFs were confirmed to be less than 2.
Statistical significance was set at p < 0.05 and q < 0.1. JMP version 14.0 (SAS Institute Inc., Cary, NC, USA) was used for the statistical analyses.

Ethics
This study was approved by the Ethics Committee of KPUM (no. ERB-C-534 and no. RBMR-E-466-5) and was performed in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants at the time of enrollment in the study.

Clinical Characteristics
This study analyzed the data of 195 men with type 2 diabetes mellitus ( Figure 1). Women were excluded from the analysis because the number of current smokers was very low (31/195 men and 8/160 women) and to follow the methods of previous studies [6,39].  Table 1 shows the clinical characteristics of participants according to their current smoking status. Of the 195 men, 164 were non-current smokers (84.1%) and 31 were current smokers (15.9%). Comparing the two groups, the current smoker group was younger, had smoked over a longer time period, had fewer exercise habits, and had less sulfonylurea use. There were no differences in BMI, hemoglobin A1c, high-density lipoprotein cholesterol, triglycerides, uric acid, creatinine, and eGFR between the two groups.   Table 1 shows the clinical characteristics of participants according to their current smoking status. Of the 195 men, 164 were non-current smokers (84.1%) and 31 were current smokers (15.9%). Comparing the two groups, the current smoker group was younger, had smoked over a longer time period, had fewer exercise habits, and had less sulfonylurea use. There were no differences in BMI, hemoglobin A1c, high-density lipoprotein cholesterol, triglycerides, uric acid, creatinine, and eGFR between the two groups.  Table 2 shows the differences in nutritional intake between current and non-current smokers. Energy, protein, fat, and carbohydrate intakes did not differ between the two groups in either total intake or intake per IBW. There were also no differences in the intake of dietary fiber, sucrose, salt, and various vitamins according to the current smoking status. In contrast, current smokers consumed significantly more alcohol than did non-current smokers. Values are presented as the mean (standard deviation). IBW, ideal body weight. Differences between the groups were analyzed using the Mann-Whitney U test. Table 3 shows food group intake compared by current smoking status. Of the 15 food groups, two food groups ("sugar/sweeteners" and "fruits") showed differences in intake between current smokers and non-current smokers. Sugar and sweetener intake for coffee or tea and for cooking were significantly higher among the current smokers, whereas the current smokers had lower fruit intake. Values are presented as the mean (standard deviation). Differences between groups were analyzed using the Mann-Whitney U test. Figure 2 shows the proportions of phyla according to the current smoking status. There was no difference between the two groups in the proportions of phyla (Table S1). or tea and for cooking were significantly higher among the current smokers, whereas the current smokers had lower fruit intake. Values are presented as the mean (standard deviation). Differences between groups were analyzed using the Mann-Whitney U test. Figure 2 shows the proportions of phyla according to the current smoking status. There was no difference between the two groups in the proportions of phyla (Table S1).  Of the highest 30 gut microbiota proportions at the genus level, the proportion of Coprococcus was significantly higher in current smokers than in non-current smokers (0.039 versus 0.025, q = 0.030) (Table S2). Figure 3 shows the difference in the proportion of Coprococcus according to the current smoking status.

Composition of Gut Microbiota
Of the highest 30 gut microbiota proportions at the genus level, the proportion of Coprococcus was significantly higher in current smokers than in non-current smokers (0.039 versus 0.025, q = 0.030) (Table S2). Figure 3 shows the difference in the proportion of Coprococcus according to the current smoking status. Multiple regression analysis using current smoking status, age, exercise habits, alcohol intake, sugar and sweetener intake, and fruit intake as covariates showed that smoking was associated with the proportion of the Coprococcus genus (Table 4).

Discussion
This study of individuals with type 2 diabetes mellitus examined the association between smoking and dietary habits and between smoking and the gut microbiota. Current smokers consumed more alcohol, sugar, and sweeteners, and less fruit. We also reported that the gut microbiota of current smokers has a higher proportion of Coprococcus at the genus level. Current smoking was associated with a higher proportion of the Coprococcus genus even after adjusting for background and dietary factors related to current smoking.
Recent studies have suggested that smoking may alter the gut microbiota. Although clinical studies have been reported in healthy people [6,40], patients with inflammatory bowel disease [41,42], and patients with coronary artery disease [39], this is the first study Multiple regression analysis using current smoking status, age, exercise habits, alcohol intake, sugar and sweetener intake, and fruit intake as covariates showed that smoking was associated with the proportion of the Coprococcus genus (Table 4). Smoking status was defined as non-current smoker (=0) or current smoker (=1), and exercise status was defined as non-regular exercise (=0) or regular exercise (=1).

Discussion
This study of individuals with type 2 diabetes mellitus examined the association between smoking and dietary habits and between smoking and the gut microbiota. Current smokers consumed more alcohol, sugar, and sweeteners, and less fruit. We also reported that the gut microbiota of current smokers has a higher proportion of Coprococcus at the genus level. Current smoking was associated with a higher proportion of the Coprococcus genus even after adjusting for background and dietary factors related to current smoking.
Recent studies have suggested that smoking may alter the gut microbiota. Although clinical studies have been reported in healthy people [6,40], patients with inflammatory bowel disease [41,42], and patients with coronary artery disease [39], this is the first study to evaluate the association between smoking and the gut microbiota in a population of people with type 2 diabetes mellitus. Furthermore, despite many studies reporting that smoking affects eating habits [43,44], previous studies on smoking and the gut microbiota found no difference in dietary intake according to smoking status or did not assess dietary intake. A significant finding of this study is the association between current smoking and dietary intake. After considering the effects of these factors, we identified the relationship between current smoking and the gut microbiota. A previous large cross-sectional study found no significant differences in the composition of the gut microbiota between those who had never smoked and former smokers, indicating that recovery of the gut microbiota composition to its pre-smoking status is likely to occur if smokers quit smoking [6]. Another longitudinal study on changes in the gut microbiota before and after smoking cessation showed that the effects of smoking cessation on the gut microbiota occurred quickly [12]. Similarly, the change in eating habits due to smoking cessation is also expected to occur relatively rapidly, as taste recovery was observed after two weeks [45] and weight gain appeared within three months [46] after smoking cessation. Therefore, in this study, the gut microbiota and dietary intake were compared between two groups depending on current smoking status.
The dietary habits of smokers are characterized by a low intake of vegetables and fruits, high intake of meat and alcoholic beverages, and low intake of antioxidant β-carotene and vitamin C [43,44]. The current smokers in this study also had high alcohol and low fruit intake, similar to previous findings. In contrast, the intake of vegetables and antioxidant nutrients did not differ between current and non-current smokers. This may be because of the influence of daily dietary advice and encouragement of vegetable consumption as part of the treatment of diabetes mellitus. It was also noted that the validity of the BDHQ used in this study may have been lower if only men responded without the advice from their wives [30]. Remarkably, the amounts of sugar and sweeteners used in coffee, tea, and cooking were higher among smokers in this study. Excessive intake of sucrose has been suggested to cause dysbiosis [47], and it is important to evaluate the intake of sucrose. However, the BDHQ is unreliable for measuring sucrose consumption and it is difficult to estimate the amount of sucrose consumed.
The relationship between smoking and the gut microbiota has been investigated in terms of diversity, composition, and function. Many studies have reported a reduction in the diversity of bacterial species in fecal samples from smokers [48]. Several studies have shown that smokers have a higher proportion of the Bacteroidetes phylum than nonsmokers [6,41,42]. Furthermore, there have been several reports of higher proportions of Prevotella in smokers [49][50][51]. The present study did not find results similar to those previously reported; however, new findings were obtained for the genus Coprococcus. The genus Coprococcus plays an important role as a protective factor against Clostridium difficile infection, and its proportion has been reported to decrease with strong inhibition of gastric acid secretion [52]. Smoking stimulates gastric acid secretion [53], which may increase the proportion of Coprococcus.
The mechanisms by which smoking affects the gut microbiota include immunosuppression, increased oxidative stress, altered gut barrier function, and changes in acid-base equilibrium [54]. Cigarette smoke contains nicotine, aldehydes, polycyclic aromatic hydrocarbons, heavy metals, toxic gases, and volatile organic compounds. The effects of each of these substances on the gut microbiota have been studied; however, the co-impact of multiple toxic substances in cigarette smoke on the gut microbiota remains unknown [55]. In addition, alcohol consumption, which was positively correlated with current smoking in this study, was reported to be associated with gut dysbiosis [56]. High fat, high sugar, and low fiber diets are also thought to cause gut dysbiosis in smokers [57]. Therefore, smoking could affect the gut microbiota not only directly through toxic substances but also indirectly through changes in dietary intake. Moreover, animal studies have indicated that a fiber-free diet further contributes to the reduction in antioxidant capacity caused by smoking [58], and smoking is expected to have a synergistic direct and indirect negative effect.
This study had several limitations. The study population of individuals with type 2 diabetes mellitus was originally prone to dysbiosis, and changes in the gut microbiota due to smoking may not have been adequately assessed. Moreover, if the direct effects of smoking on the gut microbiota are offset by indirect effects via dietary changes, they may not be detected as differences in the gut microbiota of current and non-current smokers. The quantitative impact of the number of cigarettes smoked or years of smoking was not evaluated in this study. Additionally, since this was a cross-sectional study, the causal relationship between changes in the gut microbiota, smoking, and diet is unknown.
Smoking cessation is important for people with diabetes because smoking has a significant negative influence on the pathogenesis of diabetes and its complications. Unhealthy diets consumed by patients with diabetes may have been influenced by smoking. However, many people are reluctant to quit smoking because of weight gain concerns due to smoking cessation. Weight gain associated with smoking cessation has been shown to not reduce the long-term benefit of smoking cessation in reducing cardiovascular and all-cause mortality [11,59]. More research is needed to clarify the relationship between smoking cessation and the gut microbiota, diet, and body weight because if means can be found to control weight gain with smoking cessation, more people are likely to quit smoking, and the benefits of smoking cessation would increase.

Conclusions
This study revealed that both dietary intake and the gut microbiota are associated with current smoking status in men with type 2 diabetes mellitus. However, even after adjusting for the effects of smoking on dietary intake, associations between current smoking and the gut microbiota were still observed. The results of this study suggest that although dietary intake should be considered when examining gut microbiota associations, the non-dietary effects of smoking are more significant.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/nu14224800/s1, Table S1: The proportions of phyla by current smoking status; Table S2: The proportions of genera by current smoking status. Funding: This research was partly supported by MAFF Commissioned project study (Grant Number JPJ009842) on "Project for the realization of foods and dietary habits to extend healthy life expectancy" to Y.N. No funding bodies played a role in this study.

Institutional Review Board Statement:
The study was conducted in accordance with the guidelines of the Declaration of Helsinki and was approved by the Research Ethics Committee of KPUM (no. ERB-C-534 and no. RBMR-E-466-5).
Informed Consent Statement: Informed consent was obtained from all the participants involved in the study.