Effect of daily ingestion of Bifidobacterium and dietary fiber on vascular endothelial function: a randomized, double-blind, placebo-controlled, parallel-group comparison study

ABSTRACT

Bifidobacterium animalis subsp. lactis GCL2505 (GCL2505) improves the intestinal microbiota and reduces human visceral fat. This randomized, double-blind, placebo-controlled, parallel-group study was conducted to examine the effects of inulin, a prebiotic dietary fiber, and GCL2505 on vascular endothelial function in healthy subjects (n = 60). The test drink contained 2.0 g/100 g inulin and 1.0 × 10¹⁰ colony-forming units/100 g GCL2505 and was consumed daily for 12 weeks. Flow-mediated dilation was set as the primary endpoint. Subgroup analysis of vascular endothelial function demonstrated a significant increase in the change of flow-mediated dilation (%) from weeks 0 to 12 in the GCL2505 and inulin group (n = 24) compared with the placebo group (n = 23), while an improving trend in low-density lipoprotein cholesterol and plasminogen activator inhibitor-1 were confirmed. Our results indicated that the test drink had a positive effect on vascular endothelial function and related blood parameters.

Graphical Abstract

The suggested mechanism of improving vascular endothelial function via intake of Bifidobacterium animalis subsp. lactis GCL2505 and inulin.

Bifidobacterium animalis subsp. lactis, vascular endothelial function, visceral fat, low-density lipoprotein cholesterol, plasminogen activator inhibitor-1

Cardiovascular disease (CVD) has become a leading cause of death worldwide, with a particularly large impact on low- and middle-income countries (World Health Organization 2021). An estimated 17.9 million people died of CVD in 2019, representing 32% of all deaths worldwide (World Health Organization 2021). Despite current prevention and treatment strategies, mortality from CVD is expected to increase further over the next decade (World Health Organization 2021).

Atherosclerosis is an inflammatory disease of the cardiovascular system characterized by a narrowing of the arterial lumen due to plaque formation. The onset and development of atherosclerosis leads directly to the development of CVD (Weber and Noels 2011; Sakakura et al.2013). Vascular endothelial dysfunction leads to plaque formation and disease progression, resulting in the development of atherosclerosis (Davignon and Ganz 2004). Therefore, improving vascular endothelial function may prevent CVD in at-risk individuals.

Similarly, CVD has been widely reported to be associated with metabolic syndrome (MetS), which is a clinical metabolic condition associated with at least three of the following metabolic risk factors: excess visceral fat (abdominal obesity), insulin resistance, hyperglycemia, hypertension, and dyslipidemia. (Torres et al.2019). These MetS components have been found to impair endothelial function individually (Palomo et al.2009). In addition, MetS patients are known to develop vascular endothelial dysfunction at a high frequency (Benjamin et al.2004).

Regulation of the gut microbiota is an effective approach to address the development and progression of obesity and MetS. Probiotics, which have been defined as “live microorganisms which when administered in adequate amounts confer a health benefit on the host” (Hill et al.2014), are a typical option for regulating the gut microbiota. Bifidobacteria and lactobacilli are representative examples of probiotics. The administration of probiotics contributes to the maintenance of short-chain fatty acid (SCFA)-producing bacteria and the suppression of inflammatory and pathogenic bacteria via regulation of the gut microbiota (Zhang et al.2010; Qin et al.2012), which in turn affects host energy metabolism and the development of MetS, including obesity and insulin resistance (Tremaroli and Bäckhed 2012). A number of reviews have examined the effects of probiotics on type II diabetes (Tao et al.2020) and hypertension (Chi et al.2020). A combined probiotic and prebiotic approach, which is referred to as synbiotics (Gibson and Roberfroid 1995), has also been suggested to increase the number of bifidobacteria in the gut and to regulate the intestinal microbiota more appropriately than with either approach alone. Prebiotics are nonviable food ingredients that provide health benefits to the host in relation to the regulation of the intestinal microbiota (Pineiro et al.2008).

Bifidobacterium animalis subsp. lactis GCL2505 (GCL2505) is a probiotic strain originally isolated from the feces of healthy adults and capable of growing in the gut (Ishizuka et al.2012; Aoki et al.2016). Previously, it was demonstrated that GCL2505 exerts anti-MetS effects such as improving glucose tolerance and suppressing visceral fat accumulation (Aoki et al.2017). Furthermore, GCL2505 affects host metabolic homeostasis (eg enhancing glucose tolerance and suppressing body fat accumulation) in a G protein-coupled receptor 43 (GPR43)-dependent manner by promoting SCFA production in the gut (Horiuchi et al.2020). In a clinical study, it was also found that daily consumption of fermented milk containing GCL2505 reduced abdominal visceral fat mass (Takahashi et al.2016). Our previous findings indicate that GCL2505 exerts its anti-MetS effects by promoting the production of SCFAs in the intestine. In addition, clinical studies have shown that GCL2505, when taken in combination with inulin (Mensink et al.2015), a well-known prebiotic dietary fiber, increases the total number of bifidobacteria in the gut more than GCL2505 alone (Anzawa et al.2019) and improves cognitive function in the elderly (Azuma et al.2023).

Therefore, it was newly thought that the combined intake of GCL2505 and inulin, which was confirmed to have a visceral fat reduction effect in clinical trials (Takahashi et al.2016), may have an effect of reducing arteriosclerosis risk via its anti-MetS effect. However, it has not yet been demonstrated that the anti-MetS effects of probiotics can improve vascular endothelial function to date. Therefore, we conducted a randomized, double-blind, placebo-controlled, parallel-group comparison study to evaluate the combined effect of inulin and GCL2505 on reducing the risk of arteriosclerosis.

Flow-mediated dilation (FMD) was used in this study to assess vascular endothelial function. Flow-mediated dilation is a vascular endothelial function test that utilizes the flow-dependent vasodilatory response that occurs after the release of upper arm ejection. Increased blood flow after anti-myelination causes shear stress on the vascular endothelium, which in turn produces vasodilators such as nitric oxide. These substances act on vascular smooth muscle adjacent to the endothelium, causing the vascular smooth muscle to relax, which in turn triggers a vasodilatory response (Joannides et al.1995; Doshi et al.2001). This response is used for assessing the ratio of maximally dilated vessel diameter to resting vessel diameter, which is defined as FMD (%). Flow-mediated dilation is recognized as a noninvasive, low patient-burden, and useful test for assessing CVD risk (Tomiyama et al.2008; Ito et al.2015). Furthermore, because FMD (%) is reported to correlate with renal function, an improvement in vascular endothelial dysfunction may also reduce the risk of developing kidney disease (Ito et al.2015).

Materials and methods

Participants

The feasibility of clinical trials and the ethical and scientific validity of this study were reviewed and approved by the Bioethics Committee of Hokkaido Information University (approval date: March 15, 2022; approval number: 2021-44). Written informed consent was obtained from all participants prior to enrolment. The trial was conducted in accordance with the guidelines laid down in the Declaration of Helsinki. Participants were Japanese men and women between the ages of 40 and 65 years with a tendency toward reduced vascular endothelial function and who met the selection criteria, did not meet the exclusion criteria, and were deemed eligible by the study investigators. The selection criteria were as follows: (1) fully understand the significance, content, and purpose of the study, and provide written informed consent to participate in the study; (2) Japanese men and women between the ages of 40 and 65 years at the time of screening; (3) have a high pentraxin 3 (PTX3) level at the screening test; and (4) have a high low-density lipoprotein cholesterol (LDL-c) to high-density lipoprotein cholesterol (HDL-c) ratio at the screening test. The exclusion criteria were as follows: (1) have difficulty with FMD measurement on the right arm; (2) receiving medical treatment, medication, or lifestyle guidance from a medical doctor for hypertension, dyslipidemia, or diabetes; (3) a history of bronchial disease such as asthma, tuberculosis, or pleurisy; (4) presence of a pacemaker or defibrillator; (5) serious cerebrovascular disease, heart disease, liver disease, renal disease, gastrointestinal disease, infectious disease requiring notification, and so on; (6) a history of major gastrointestinal surgery such as gastrectomy, gastrointestinal suture, or intestinal resection; (7) significant abnormalities in blood pressure measurement, physical measurements, or blood test; (8) severe anemia; (9) pre- and postmenopausal women with significant physical changes; (10) a history of severe illness such as heart disease, liver disease, kidney disease, digestive disease, or infection that was reported to a hospital; (10) allergic symptoms to drugs or foods (especially milk ingredients); (11) taking or planning to take antibiotics from 12 weeks before the study to the end of the study; (12) regular intake of medications that affect bowel movements (antiflatulents, laxatives, antidiarrheals, and so on); (13) unable to stop consuming yogurt, lactic acid bacteria beverages, and health foods with intestinal-regulation effects (containing ingredients such as lactic acid bacteria, bifidobacteria, oligosaccharides, dietary fiber-enriched foods, and so on) during the test period; (14) smokers, heavy drinkers, or extremely irregular lifestyles; (15) women who donated 400 mL blood within 16 weeks before starting intake, men who donated 400 mL blood within 12 weeks or 200 mL blood within 4 weeks before starting intake, or donated blood components within 2 weeks before starting intake; (16) pregnancy, suspected pregnancy, or breastfeeding; (17) duplicate participation in other clinical trials, observational studies, clinical trials, or home-use tests from 4 weeks before the start of the study period to the end of intake; and (18) any other person deemed ineligible by the study investigator.

Test drinks

The test products were a dairy drink (active drink) containing inulin (Orafti GR; BENEO GmbH, Mannheim, Germany) and GCL2505 or placebo. The active drink was designed to contain 2.0 g/100 g inulin and 1.0 × 10¹⁰ colony-forming units/100 g GCL2505. The placebo was prepared with the same ingredients as the active drink, with the addition of food-grade acetic acid and lactic acid to adjust the flavor and pH. The basic ingredients were skim milk powder, fructose, dextrose, sucrose, yeast extract, acidifier, stabilizer, and flavoring. The nutritional details of the test products are shown in Table 1.

Table 1.

Nutritional details per 100 g of the test drinks

	Active	Placebo
Energy (kcal)	52.0	48.0
Moisture (g)	84.9	86.9
Protein (g)	2.8	2.8
Fat (g)	0.1	0.1
Carbohydrate (g)	11.2	9.1
Ash (g)	1.1	1.1

The active drink contained 2.0 g inulin and 1.0 × 10¹⁰ colony-forming units GCL2505.

Experimental design

This was a double-blind, placebo-controlled, parallel-group study. For sample size, the final target number of subjects was set at 50, referring to previous reports on probiotic-induced vascular endothelial function (Matsumoto et al.2019; Rezazadeh et al.2019). Considering cases of discontinuation or dropout after the start of intake, the target number of enrollment was set at 60 subjects. The participants were randomized by a stratified permuted block method, with sex, age composition, PTX3 measurement value, and LDL-c/HDL-c ratio as stratification factors, and they were assigned to two groups as subjects. It was reported that LDL-c/HDL-c ratio is a suitable tool for detecting CVD risk and is correlated with FMD (Fernandez and Webb 2008). Pentraxin 3 was also reported to be correlated with FMD (Yilmaz et al.2020). Therefore, PTX3 and L/H ratio were used as screening tests in this study. Subjects in the active drink and placebo groups consumed 100 g of each test drink once daily for 12 weeks. The subjects and observers were blinded to group allocation for the duration of the study. Double blinding was accomplished by displaying only an identification number on the test products. Flow-mediated dilation was used to assess vascular endothelial function, and FMD (%) was used for evaluation (described below in the FMD section). The change in FMD (%) between weeks 0 and 12 was set as the primary outcome. The secondary outcome was the change in FMD (%) from week 0 to 8. Changes in total cholesterol, LDL-c, HDL-c, triglycerides, PAI-1, and TMAO from week 0 to 8 and from week 0 to 12 were also set as secondary outcomes. The tests for the primary and secondary endpoints were performed at weeks 0, 8, and 12. Fecal samples were also collected at weeks 0, 8, and 12. This study was conducted at the Center of Health Information Science, Hokkaido Information University (Ebetsu, Hokkaido, Japan) and was registered in the University Hospital Medical Information Network Clinical Trials Registry (http://www.umin.ac.jp/ctr/index.htm) under the study ID UMIN000047853.

FMD

An UNEX-EF18VG (UNEX Corporation, Aichi, Japan) was used to test the flow-dependent vasodilation response at weeks 0, 8, and 12. The right arm was used for measurements. First, the diameter of the right brachial artery was measured at rest, and then ischemia was induced in the right forearm by inflating a cuff for 5 min. After the pressure of the cuff was released and blood flow restored, the diameter of the maximally diastolic vessel was measured again. The percentage of vascular diameter expansion from rest was expressed as FMD (%). Subjects were restricted from drinking alcohol and overeating the day before FMD measurements, and were prohibited from consuming any food or drink other than water after 9:00 p.m. Subjects were only allowed to drink water (approximately 200 mL) from the time they woke up on the day of the test until 2 h before the meeting time. Data from weeks 8 and 12 were evaluated by calculating the change from week 0, respectively.

Biochemical parameters

Concentrations of biochemical parameters were measured in blood at weeks 0, 8, and 12. Blood samples were taken from each fasting subject who had not consumed anything except water from 21:00 the previous evening. The biochemical parameters measured were general blood parameters (white blood cell count, red blood cell count, hemoglobin, hematocrit, and platelet count), liver function (aspartate transaminase, alanine aminotransaminase, gamma-glutamyl transpeptidase, alkaline phosphatase, and lactate dehydrogenase), kidney function (urea nitrogen, creatinine, and uric acid), blood lipids (total cholesterol, LDL-c, HDL-c, and triglycerides), blood glucose (fasting blood glucose, HbA1c, and insulin), PTX3, PAI-1, and TMAO. The blood draw was performed on the same day as the FMD measurement, but after the FMD test so as not to affect the FMD measurement results. All parameters were examined at Sapporo Clinical Laboratory, Inc. (Sapporo, Japan). Regarding the items of total cholesterol, LDL-c, HDL-c, triglycerides, LDL-c/HDL-c, PAI-1 and TMAO, data from weeks 8 and 12 were evaluated by calculating the change from week 0, respectively.

Fecal collection

Fecal samples were collected by the participants at weeks 0, 8, and 12 at home from 3 days before to the morning of the specified visit, using a fecal container containing RNAlater Stabilization Solution (Invitrogen, Carlsbad, CA). Fecal samples were stored in a refrigerator at home and brought to the laboratory on the day of testing while being cooled with refrigerant.

Fecal DNA extraction

Bacterial DNA was extracted from 10-fold dilutions of fecal samples using an ISOSPIN Fecal DNA Kit (Nippon Gene Co., Ltd., Tokyo, Japan) as described previously (Tourlousse et al.2021). More specifically, the sample (here, 200 µL fecal dilution), 700 µL FE1 buffer, and 10 µL Rnase were added to the provided tubes with beads. Cells were disrupted by bead-beating for 1 min at a speed of 6 m/s using a FastPrep-24 instrument (MP Biomedicals, Irvine, CA). Bead beating was repeated three times, keeping the sample at room temperature for 5 min between rounds. Subsequently, 90 µL FE2 buffer was added and the sample was centrifuged for 15 min at 12 000 × g. The supernatant (up to 500 µL) was recovered and mixed with 0.4 volumes of FB buffer and isopropanol. The sample was finally loaded onto a spin column and cleaned up, according to the manufacturer's instructions. Purified DNA was eluted in 50 µL Tris-EDTA (pH 8.0) buffer.

Shotgun library construction and sequencing

Unless stated otherwise, construction of metagenome sequencing libraries was performed using a QIAseq FX DNA Library Kit (Qiagen, Hilden, Germany), according to the manufacturer's instructions. Briefly, the enzymatic fragmentation reactions (50 µL) contained 10 × FX buffer, 10 µL FX enzyme mix, and 500 ng DNA template, and the reactions were incubated for 9 min at 32 °C. For adapter ligation, 5 µL adapters was added, along with 20 µL DNA ligase buffer, 10 µL DNA ligase, and 15 µL Rnase-free H₂O, and the reactions were incubated for 15 min at 20 °C. Adapter-ligated fragments were purified and size-selected with the Agencourt AMPure XP PCR purification system (Beckman Coulter, Brea, CA), using sequentially 1 and 0.8 volumes of bead solution, and eluted in 10 m m Tris-HCl buffer.

Quality control of metagenomic reads

Quality control of the metagenomic reads and adapter sequence trimming were performed using fastp (version 0.20.0). Reads less than 50 bp were excluded from further analysis. The remaining reads were mapped to the human (hg38) and phiX bacteriophage genomes, using minimap2 (version 2.17), and mapped reads were excluded. Reads whose pairs were excluded in the filtering steps were also removed.

Construction of a non-redundant gene set and functional annotations

A non-redundant gene set was constructed based on HMP and population-level Japanese metagenomics data. All genes were clustered with a 95% identity threshold using cd-hit (version 4.8.1). Functional annotation of the non-redundant genes was performed using eggNOG-mapper (version 2.1.9), based on the eggNOG orthology database (version 5.0.2). Sequence searches were performed using DIAMOND (version 2.0.15).

Taxonomic and functional profiles of the samples

Taxonomic profiles at the species and genus levels were obtained using the marker-gene-based approach with mOTUs2 (version 3.0.3). For the quantification of gene function, the metagenomic reads were mapped to the non-redundant genes, using minimap2 with a ≥95% identity threshold, and the number of mapped reads for each gene was counted. The number of multiply mapped reads was distributed to the mapped genes based on the ratio of the number of uniquely mapped reads to the genes. The counts were normalized to transcripts per kilobase per million to form a transcripts per million matrix.

Statistical analysis

All measured values are presented as the mean and standard deviation (SD). All statistical analyses were conducted with the open-source software program R (version 4.2.1) (http://cran.r-project.org). A P-value < 0.05 was used as the threshold for determining significant differences. Missing data were treated as missing values and no surrogate values were used. To evaluate the safety of GCL2505 and inulin, test results of week 0 and 12 were compared and paired t-tests were used, and to evaluate the other items, the active and placebo groups were compared and statistical analysis was performed by means of an unpaired t-test (two-tailed test).

Results

Subjects

An overview of the study period is shown in Figure 1. A total of 149 participants were screened for this study. A total of 60participants were eligible as a result of screening: 30 were assigned to the active and placebo groups, respectively. There were no differences in the baseline characteristics of the subject's data between the active and placebo groups at the beginning of the study (Table 2). Each compliance rate of the test drinks was adequate and there was no significant difference in the compliance rate between both groups (Table 2). By the end of the study, three participants had withdrawn for personal reasons (n = 1 from the active group and n = 2 from the placebo group). In addition, one participant was deemed to have discontinued for compliance reasons. After completion of the entire study, one participant was excluded due to an illness unrelated to the study that may have affected the results. Thus, a total of 55 participants—27 in the active group and 28 in the placebo group—were included in the analysis population. Subject selection was designed to screen suitable healthy subjects for the preventive effects of the test drink on vascular endothelial dysfunction. However, there were eight subjects (n = 3 from the active group and n = 5 from the placebo group) who had a pre-intervention FMD (%) of <4%. Because these individuals may have already developed vascular endothelial dysfunction (Tanaka et al. 2018), it was considered necessary to analyze the healthy population without them. Therefore, subgroup analysis was also performed in healthy subjects with FMD (%) ≥4% at week 0 (active group: n = 24; placebo group: n = 23).

Figure 1.

Method followed for subject selection.

Table 2.

Background data of subjects

	Active	Placebo	P-value
Age, years	50.37 (7.47)	50.04 (6.33)	0.859
Number of women	24 of total 27	24 of total 28	–
Systolic blood pressure, mmHg	112.81 (12.52)	114.86 (13.40)	0.562
Diastolic blood pressure, mmHg	77.63 (9.23)	79.21 (8.68)	0.515
Heart rate, bpm	69.22 (11.24)	65.36 (6.75)	0.131
Height, cm	159.66 (5.92)	160.56 (8.03)	0.637
Body weight, kg	54.27 (6.78)	54.24 (9.84)	0.987
Body fat, %	29.11 (6.51)	27.41 (6.37)	0.332
Body mass index, kg/m2	21.29 (2.56)	20.93 (2.59)	0.609
FMD, %	6.03 (2.14)	6.23 (2.72)	0.764
Total cholesterol, mg/dL	219.19 (22.81)	218.18 (27.61)	0.883
LDL-c, mg/dL	122.52 (20.02)	117.71 (21.98)	0.400
HDL-c, mg/dL	77.00 (12.66)	83.14 (17.41)	0.140
Triglyceride, mg/dL	78.81 (30.41)	72.32 (26.19)	0.401
LDL-C/HDL-C	1.64 (0.45)	1.48 (0.38)	0.156
PAI-1, ng/mL	15.44 (10.03)	11.54 (2.40)	0.088
TMAO, µg/mL	0.21 (0.13)	0.26 (0.38)	0.543
Pentraxin 3, ng/mL	1.78 (0.67)	1.79 (0.64)	0.929
Compliance rate of the test sample, %	99.60 (1.08)	99.57 (1.09)	0.749

All data are presented as mean (SD), except for the number of women.

P-value shows the inter-group difference (active group vs. placebo group; Student's t-test).

Safety assessment

Biological parameters at week 12 were compared to those at week 0 to confirm safety in the analyzed population. As a result, significant differences were confirmed in hematocrit of the active and placebo groups, platelet count of the active and placebo groups, alkaline phosphatase of the active and placebo groups, and urea nitrogen of the placebo group. All were within the range of the reference value, so it was determined that there was no problem. For other items, no significant changes were observed during the test period (Table 3). No side effects or medically problematic adverse events were observed. Therefore, the dairy drinks used in the study were considered safe for consumption.

Table 3.

Effect of test drink containing GCL2505 and inulin on biochemical parameters

		Week 0	Week 12	P-value
White blood cell count, ×103/µL	Active	4.95 (1.08)	5.07 (1.29)	0.514
	Placebo	5.19 (1.39)	4.95 (1.18)	0.313
Red blood cell count, ×104/µL	Active	447.85 (33.81)	449.41 (34.50)	0.626
	Placebo	439.36 (37.44)	446.30 (32.07)	0.114
Hemoglobin, g/dL	Active	13.24 (1.00)	13.26 (1.19)	0.790
	Placebo	13.11 (1.22)	13.25 (1.16)	0.151
Hematocrit, %	Active	40.71 (2.86)	41.79 (3.33)	0.003
	Placebo	40.01 (3.58)	41.30 (3.30)	0.001
Platelet count, ×104/µL	Active	25.74 (6.17)	27.81 (6.52)	0.002
	Placebo	24.90 (4.96)	26.55 (5.87)	0.006
Aspartate transaminase, U/L	Active	20.22 (4.48)	20.33 (4.32)	0.900
	Placebo	19.61 (3.27)	19.81 (4.31)	0.694
Alanine aminotransaminase, U/L	Active	17.33 (6.15)	17.33 (7.95)	1.000
	Placebo	15.25 (4.93)	15.22 (5.45)	0.788
Gamma-glutamyl transpeptidase, U/L	Active	31.78 (38.84)	28.85 (31.44)	0.393
	Placebo	17.79 (6.27)	17.78 (7.75)	0.742
Alkaline phosphatase, U/L	Active	66.78 (18.05)	69.52 (19.90)	0.014
	Placebo	58.79 (12.51)	61.11 (12.44)	0.003
Lactate dehydrogenase, U/L	Active	178.63 (36.07)	182.96 (25.87)	0.334
	Placebo	170.86 (21.61)	172.81 (21.80)	0.201
Urea nitrogen, mg/dL	Active	12.21 (2.92)	12.79 (4.24)	0.318
	Placebo	11.24 (2.67)	11.87 (2.93)	0.024
Creatinine, mg/dL	Active	0.71 (0.12)	0.71 (0.14)	0.481
	Placebo	0.71 (0.14)	0.71 (0.13)	1.000
Uric acid, mg/dL	Active	4.49 (0.82)	4.38 (0.70)	0.229
	Placebo	4.23 (0.81)	4.14 (0.91)	0.199
Fasting blood glucose, mg/dL	Active	84.00 (5.10)	85.44 (6.64)	0.161
	Placebo	82.71 (5.47)	84.11 (6.80)	0.053
HbA1c, %	Active	5.44 (0.27)	5.40 (0.26)	0.095
	Placebo	5.32 (0.21)	5.28 (0.24)	0.364
Insulin, µ/IU/mL	Active	4.42 (2.62)	4.95 (3.24)	0.086
	Placebo	3.42 (1.60)	3.58 (1.52)	0.676

All data are presented as mean (SD).

P-value shows the intra-group difference (week 0 vs. 12; paired t-test).

FMD

In the full analysis population, an increase in FMD (%) was observed in the active group at week 12 (1.05% ± 2.21%); the change in FMD (%) from weeks 0 to 12 in the active group showed a statistically increasing trend compared to that in the placebo group (−0.26% ± 2.87%) (P = 0.065, Student's t-test) (Figure 2a). In addition, an increase in FMD (%) was observed in the active group at week 12 (0.88% ± 2.23%) in the population of subjects with an FMD (%) of > 4% at week 0. Furthermore, there was a statistically significant difference between the change in FMD (%) from weeks 0 to 12 in the active group compared with the placebo group (−0.68% ± 2.81%) (P = 0.046, Student's t-test) (Figure 2a). There was no statistically significant difference in FMD (%) change between week 0 and 8 in either the full analysis or the subclass analysis population (Figure 2b).

Figure 2.

Effects of the test drinks on FMD. (a) Change between week 0 and 12, the primary outcome. (b) change between week 0 and 8, a secondary outcome. Values are as means, with error bars asSD. The data show the amount of change from week 0, respectively. P-value shows the inter-group difference (change value of active group vs. placebo group; Student's t-test).

Fecal microbiota

The fecal microbiota was examined by shotgun metagenomic analysis in the subgroup analysis population at weeks 0 and 12. Alpha-diversity (Chao1) and beta-diversity analyses were performed to compare the fecal microbiota of both groups. Interestingly, no differences were observed between or within both groups at weeks 0 and 12 (Figure 3a and b). The relative abundance of B. animalis in the intestine of the active group was 0.061% ± 0.002% at week 0 and increased to 1.503% ± 0.018% at week 12; its abundance was 0.085% ± 0.004% at week 0 and 0.014% ± 0.001% at week 12 in the placebo group. Differences in abundance between both groups were assessed using the Linear Model for Differential Abundance (LinDA) method for compositional data, with P-values corrected using the Benjamini-Hochberg procedure to minimize the false discovery rate (Zhou et al.2022). At week 12, there was a statistically significant difference between the relative abundance of B. animalis in the active group compared with the placebo group (P = 5.62 × 10⁻⁹) (Figure 3c).

Figure 3.

Effect of consuming the test drink containing GCL2505 and inulin on the fecal microbiota in a population of subgroup analysis (active group: n = 22; placebo group: n = 21). Boxplots represent the 5th percentile, 95th percentile, interquartile range (25%-75%), and median. (a) Alpha-diversity (Chao1). (b) Beta-diversity (principal components analysis of genus-level Bray-Curtis distance). (c) Relative abundance of B. animalis at week 12. Data were analyzed by LinDA.

Physical parameters

In the subgroup analysis population, blood PAI-1 levels decreased in the active group (week 8: −1.83 ± 5.25 ng/mL, week 12: −1.25 ± 4.66 ng/mL). The change in blood PAI-1 levels in the active group from weeks 0 to 8 showed a statistically increasing trend compared to that in the placebo group (0.52 ± 2.61 ng/mL) (P = 0.058, Student's t-test), and the change in blood PAI-1 levels in the active group from weeks 0 to 12 was also showed a statistically increasing trend compared to that in the placebo group (0.86 ± 3.12 ng/mL) (P = 0.076, Student's t-test) (Table 4). In the same population, blood total cholesterol levels decreased in the active group (week 8: −2.75 ± 11.51 mg/dL, week 12: −3.75 ± 21.09 mg/dL), and the change in the active group from weeks 0 to 8 showed a statistically increasing trend compared to that in the placebo group (3.87 ± 13.74 mg/dL) (P = 0.081, Student's t-test). Blood LDL-c levels also decreased in the active group (week 8: −4.17 ± 9.18 mg/dL; week 12: −4.29 ± 15.14 mg/dL). There was a statistically significant difference between the change in blood LDL-c levels in the active group from weeks 0 to 8 compared with the placebo group (3.57 ± 10.93 mg/dL) (P = 0.012, Student's t-test). The change in blood LDL-c levels in the active group from weeks 0 to 12 showed a statistically increasing trend compared to that in the placebo group (2.14 ± 9.73 mg/dL) (P = 0.092, Student's t-test) (Table 4).

Table 4.

Effects of the test drinks on blood parameters in a population of subgroup analysis (active group: n = 24; placebo group: n = 23)

		Week 8		Week 12
		Change	P-value	Change	P-value
Total cholesterol, mg/dL	Active	−2.75 (11.51)		−3.75 (21.09)
	Placebo	3.87 (13.74)	0.081	3.00 (13.55)	0.200
LDL-c, mg/dL	Active	−4.17 (9.18)		−4.29 (15.14)
	Placebo	3.57 (10.93)	0.012	2.14 (9.73)	0.092
HDL-c, mg/dL	Active	2.08 (6.50)		2.04 (7.39)
	Placebo	1.74 (5.79)	0.849	2.00 (7.98)	0.985
Triglycerides, mg/dL	Active	−6.08 (29.03)		−4.42 (25.02)
	Placebo	−6.52 (23.71)	0.955	−2.41 (18.43)	0.757
LDL-c/HDL-c	Active	−0.10 (0.21)		−0.10 (0.24)
	Placebo	0.01 (0.14)	0.041	−0.02 (0.14)	0.150
PAI-1, ng/mL	Active	−1.83 (5.25)		−1.25 (4.66)
	Placebo	0.52 (2.61)	0.058	0.86 (3.12)	0.076
TMAO, µg/mL	Active	0.12 (0.41)		0.34 (0.86)
	Placebo	−0.02 (0.37)	0.228	0.06 (0.59)	0.197

All data are presented as mean (SD).

The data for weeks 8 and 12 show the amount of change from week 0, respectively.

P-value shows the inter-group difference (change value of active group vs. placebo group; Student's t-test).

Discussion

We investigated the effects of consuming a dairy drink containing Bifidobacterium animalis subsp. lactis GCL2505 (GCL2505) and inulin on vascular endothelial function in healthy adults. Our results indicated that the test drink had a positive effect on vascular endothelial function and related blood parameters.

Flow-mediated dilation was used in this study, and the primary endpoint was the change in FMD (%), the ratio of the maximally dilated vessel diameter to the resting vessel diameter, during the test period. The Japan Society for Vascular Failure and the Japanese Circulation Society have jointly proposed standard values for FMD (%). In their report, an FMD of <4% is considered abnormal with a high suspicion of vascular endothelial dysfunction (Tanaka et al.2018). While a pharmaceutical approach is appropriate for patients, it is always desirable to find ways to prevent disease before it occurs. For prevention, it is preferable to use foods that are commonly consumed in daily life. Therefore, in the present study, in which a dairy drink was used as an intervention, subgroup analysis was conducted in healthy subjects with an FMD (%) of ≥ 4% before intervention.

The active group had a 0.85% increase in FMD (%) after consuming the test drink containing GCL2505 and inulin during the 12-week study period. Furthermore, the change in FMD (%) from weeks 0 to 12 was significantly greater in the active group than in the placebo group. Because meta-analyses have shown that a 1% increase in FMD (%) results in an approximately 13% reduction in the risk of developing CVD (Inaba, Chen and Bergmann 2010; Matsuzawa et al. 2015), the findings of the present study are considered clinically meaningful.

Analysis of the fecal microbiota of the subjects showed no effect of GCL2505 and inulin in alpha-diversity and beta-diversity analyses. It has been reported that dietary fiber intervention in healthy humans increases the abundance of certain bacteria but does not affect the diversity of the intestinal microbiota (So et al.2018). The lack of a change in the fecal microbiota may be because the subjects in this study were healthy and did not have dysbiosis. The relative abundance of B. animalis increased in the intestines of subjects in the active group only and did not increase in the placebo group. In a clinical trial, Anzawa et al. confirmed the number of bifidobacteria in the feces of a group that intook only GCL2505 and a group that intook GCL2505 and inulin. As a result, there was no difference in the number of B. animalis subsp. lactis, a subspecies of B. animalis, between the two groups (Anzawa et al. 2019). The growth-enhancing effect of inulin on endogenous B. animalis has not been confirmed. Therefore, it was determined that increased B. animalis in this study means GCL2525, and the ingested GCL2505 was able to reach the intestine alive. The increased relative abundance of B. animalis in the intestine may have increased the production of acetic acid, a major SCFA.

In this study, we found an effect of the test drink on blood LDL-c levels. High plasma cholesterol concentrations, especially LDL-c concentration, are known to be one of the major risk factors for atherosclerosis (Ross 1999). Low-density lipoprotein can be modified by oxidation, glycation (in diabetes), aggregation, binding to proteoglycans, or incorporation into immune complexes, and are a major cause of damage to the endothelium and underlying smooth muscle. When LDL particles become trapped in an artery, they can undergo progressive oxidation and be internalized by macrophages by means of the scavenger receptors on the surfaces of these cells. Low-density lipoprotein cholesterol deposited in the vascular wall becomes oxidized LDL, which contributes to the dysfunction of vascular endothelial cells and the increased expression of adhesion factors (Chinetti-Gbaguidi, Colin and Staels 2015). Oxidized LDL is taken up by macrophages and is associated with foam cell formation, which promotes local and systemic plaque inflammation, thereby growing atherosclerotic lesions (Barrett 2020). Oxidized LDL is one of the major causes of atherosclerosis in humans (Ross 1999). Several clinical trials have been conducted to investigate the association between LDL-c and vascular events and a direct involvement of oxidized LDL in CVD, which is produced by oxidation of LDL-c, is has been reported (Baigent et al. 2010; Cannon et al.2015). In this study, a decrease in blood LDL-c level was confirmed along with an improvement in vascular endothelial function, but there was no statistically significant correlation between the amount of change in FMD (%) and in LDL-c level (data not shown). In order to discuss the relationship between these in more depth, it is desirable to measure the amount of oxidized LDL, which is directly related to CVD. In addition, blood PAI-1 levels were lower in the active group than in the placebo group. PAI-1 is a cytokine that strongly inhibits thrombolysis by blocking tissue plasminogen activator, a known thrombolytic agent (Vaughan 2005). Elevated PAI-1 levels correlate with atherothrombosis (Sobel, Taatjes and Schneider 2003), and PAI-1 is increased in obese MetS patients and type II diabetics (Juhan-Vague et al.2003). Therefore, the intake of inulin and GCL2505, which has anti-MetS effects, reduced blood LDL-c and PAI-1 levels. Although there was no statistically significant difference between the two groups in the change in LDL-c at week 12, a significant trend was observed. The number of subjects was small due to subclass analysis, which may have increased the variance. And since no statistically significant correlation was found between LDL-c and FMD (%), further studies are needed to determine a causal relationship between them.

We hypothesize that the improvement of vascular endothelial function by GCL2505 and inulin is achieved by a mechanism of action consisting of the following two steps. Step 1: the ingestion of GCL2505 and inulin exerts a visceral fat-reducing effect via increasing intestinal SCFA. Clinical studies have shown that the daily consumption of yogurt containing GCL2505 reduces abdominal visceral fat mass (Takahashi et al.2016). In addition, Horiuchi et al. demonstrated that the administration of GCL2505 increases intestinal B. lactis counts and plasma acetate concentrations and is followed by GPR43-dependent regulation of host energy metabolism (eg suppressed body fat accumulation, improved insulin sensitivity, and enhanced whole body fatty acid metabolism) in GPR43 knockout mice (Horiuchi et al.2020). G protein-coupled receptor 43 a member of the G protein-coupled receptor family, which are SCFA receptors in the host, and it has been reported that acetic acid inhibits fat accumulation by activating GPR43 in adipocytes and modulating insulin signaling in adipocytes (Kimura et al.2013). Clinical studies have shown that taking GCL2505 in combination with inulin increases the total number of bifidobacteria in the intestine (Anzawa et al.2019), so inulin may enhance the visceral fat reduction effect of GCL2505. Step 2: vascular endothelial dysfunction is prevented through the following sub-steps triggered by the reduction of visceral fat: (1) reduction of the blood LDL-c concentration by decreasing the blood lipid concentration through the reduction of visceral fat; LDL-c is a causative agent of plaque formation, and its synthesis is related to free fatty acid levels; and (2) suppression of PAI-1 expression via the reduction of visceral fat; PAI-1 is an adipocytokine produced by adipocytes and regulates fibrinolysis, which is strongly associated with atherosclerosis (Alessi and Juhan-Vague 2006). A possible reason why GCL2505 and inulin improved vascular endothelial function is that blood PAI-1 levels were reduced by the reduction of visceral fat, which activated fibrinolysis. Unfortunately, changes in the visceral fat area of the subjects were not measured in this study. However, because a positive correlation between visceral fat and blood LDL-c concentration has been reported (Luo et al.2014), it is possible that visceral fat mass in the active group in this study, which had reduced blood LDL-c levels, may have been affected. Therefore, the present study indicates that GCL2505 and inulin may contribute to vascular endothelial function via inhibitory effects on fat accumulation. The proportion of women in the study was high, 24 out of a total of 27 in the active group and 24 out of a total of 28 in the placebo group. However, there is no difference in the mechanism of CVD between men and women. And because of the protective effect of the known female hormone estrogen on cardiovascular diameter, the incidence of CVD in women was reported to be lower than in men (Stampfer et al.1991; Grodstein et al.1996). In addition, the incidence of CVD in postmenopausal women is increased compared to premenopausal women and is comparable to that in men (Kannel et al.1976). The average age of menopause among Japanese women is 49.47 years, which was found to be close to the average age of the subjects in this study (the active group : 50.37 ± 7.47 years, the placebo group: 50.64 ± 6.33 years). So it is likely that there was no significant difference in incidence between the sexes of the subjects in this study. It is also possible that if the proportion of male subjects in this study had been further increased, the effects of GCL2505 and inulin on vascular function would have been more pronounced. Based on these considerations, the efficacy assessment in this study was considered acceptable.

Species-specific PCR assays have shown that synbiotics containing GCL2505 and inulin are highly effective in increasing the number of bifidobacteria in the intestine (Anzawa et al.2019). However, in the present study, the number of intestinal bifidobacteria increased in the active group, but there was no statistically significant difference from that in the placebo group (P = 0.899 by LinDA, data not shown). There are two possible reasons for this result. One is that shotgun metagenomic sequencing was used as a tool, which is less accurate for quantification than the species-specific PCR method. In this study, the intestinal microbiota was examined by shotgun metagenomic sequencing in order to comprehensively analyze the effects of the test drink on the intestinal microbiota. The other reason is that the number of subjects used to analyze the fecal microbiota may be too small. Shotgun metagenomes have been used in human intervention studies on healthy subjects, and the number of subjects in these studies has been set at approximately 100 or more (Hibberd et al.2019; Martoni et al.2019; Mokkala et al.2021; Wang et al.2021). Alpha-diversity and beta-diversity analyses indicated that GCL2505 and inulin did not alter the intestinal microbiota, but future trials of shotgun metagenomics with an adequate sample size may reveal the impact of the intervention.

In addition, there was no change in blood TMAO levels in this study. TMAO is a metabolite of great interest as a factor in CVD among the various physiological kinetics of bacterial metabolism (Sanchez-Rodriguez et al.2020; Alhajri et al.2021). The intake of animal foods rich in trimethylamine sites such as l-carnitine and choline (eg milk, red meat, eggs) promotes TMAO production (Alhajri et al.2021). It has been reported that elevated TMAO levels increase the risk and prevalence of CVD, although the exact mechanism is not known (Ge et al.2020). Because TMAO levels did not change in the present study, it is considered that GCL2505 and inulin have no effect on TMAO. Dietary components can affect trimethylamine production by intestinal bacteria, but regrettably, the subjects’ diets were uncontrolled in this study. Therefore, further studies are needed to clarify the exact effects of GCL2505 and inulin on TMAO.

In summary, it is suggested that the ingestion of a drink containing GCL2505 and inulin improved vascular endothelial function possibly via the increasing intestinal SCFA and subsequent improving of lipid metabolism. These findings are considered to be clinically meaningful for reducing the risk of CVD. Therefore, this combination of GCL2505 and inulin may contribute significantly to improving health by reducing the risk for multiple diseases beyond CVD alone.

Data Availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Author contribution

Conceptualization, N.A., Y.S., and T.N.; methodology, N.A., Y.S., and T.N.; validation, N.A. and Y.S.; formal analysis, N.A. and R.A.; investigation, N.A.; writing-original draft preparation, N.A. and Y.S.; writing—review and editing, Y.S. and T.N.; visualization, N.A.; supervision, J.N.; and funding acquisition, Y.S.

Funding

This study was funded by Ezaki Glico Co., Ltd., Japan and was supported by the cross-ministerial Strategic Innovation Promotion Program (SIP) “Technologies for Smart Bio-industry and Agriculture” (funding agency: Bio-oriented Technology Research Advancement Institution, NARO, Japan; grant number: SIP2B).

Disclosure Statement

N.A., Y.S., T.N., and R.A. are employees of Ezaki Glico Co., Ltd. The other author, J.N., reports no conflict of interest in this work.

Acknowledgments

The authors would like to thank the staff at the Center of Health Information Science, Hokkaido Information University for their technical assistance with the clinical trial.