Impact of probiotic Veillonella atypica FB0054 supplementation on anaerobic capacity and lactate

Summary Seven healthy, physically active men (n = 3) and women (n = 4) (30.7 ± 7.5 years, 172.7 ± 8.7 cm, 70.4 ± 11.6 kg, 23.6 ± 4.1 kg/m2, 49.2 ± 8.4 mL/kg/min) supplemented for 14 days with a placebo (PLA) or 1 × 1010 CFU doses of the probiotic Veillonella atypica FB0054 (FitBiomics, New York, NY). Participants had safety panels, hemodynamics, lactate, and anaerobic capacity assessed. Stool samples were collected to evaluate for metagenomic and metabolomic changes. Exhaustion times were not different between groups, whereas anaerobic capacity tended to shorten with PLA (61.14 ± 72.04 s; 95% CI: −5.49, 127.77 s, p = 0.066) with no change with VA (13.29 ± 100.13 s, 95% CI: −79.32, 105.89 s, p = 0.738). No changes in lactate, hemodynamics, or bacterial community changes were observed, whereas 14 metabolites exhibited differential expression patterns with VA supplementation. In conclusion, VA maintained exercise performance that tended to decline in PLA. Supplementation was well tolerated with no changes in safety markers or reported adverse events.


INTRODUCTION
Our health is determined by a complex interplay between our genetics, diet, exercise, lifestyle, and microbiome, and although it is known that diet, exercise habits, and lifestyle can have significant impacts on prevention of chronic disease, the fact that the microbiome can further influence health continues to be explored. 1For its part, the human microbiota is a metabolic engine, 100 trillion organisms strong, impacting nearly every physiological system and capable of promoting health or morbidity in numerous diseases. 2,35][6] For example, the absence (via germ-free or gnotobiotic methods) or depletion (via antibiotic cocktail) of the microbiome in animal models diminishes muscle mass [7][8][9][10] and results in decreased exercise performance. 7,9,11][14][15] Further, propionate and butyrate, two key metabolites of bacterial fermentation with known health implications, can only be produced by the microbiome and not endogenously by mammalian cells.16 Several human studies have identified correlations between microbiome composition and exercise performance, although causation had not been established.[17][18][19] One bacterial genus of interest is Veillonella, which has the ability to convert lactate in the gastrointestinal tract to the SCFA, propionate. 20 Our rcent work demonstrated that the lactate-utilizing genus Veillonella, which is enriched in the gut microbiome of elite marathoners after running a marathon, significantly boosts exercise performance upon colonization in mice in a manner that can be recapitulated by propionate instillation.14 The working hypothesis is that as glucose is converted to lactate in the muscle, lactate enters the intestinal lumen via blood circulation, creating a lactate reservoir in the gut.There, lactate can act as a carbon source for microbes like Veillonella.The products of lactate utilization, in this case, propionate, are taken up by the host and utilized as an energy source, which can improve endurance performance.This thesis is significant because it directly couples the metabolic stress of exercise to microbiome activity and then back to the observed exercise response.14 Although many commercially available probiotics have been tested for performance improvement with mixed results, Veillonella atypica supplementation has not previously been tested for its influence on exercise performance in humans. He, we conduct a pilot study with the primary aim being to determine the overall clinical safety and influence of V. atypica supplementation on running time to exhaustion in healthy adult volunteers.It was hypothesized the V. atypica (VA) supplementation would be well tolerated and not be responsible for differences in reported adverse events while also promoting increases in exercise performance.

Adverse events
A total of 9 adverse events (nervousness, n = 1; nausea, n = 1; upset stomach, n = 3; stomach cramping, n = 4) were reported throughout the study protocol.Three were reported during PLA (upset stomach, n = 2; stomach cramping, n = 1), and six were reported during VA (nervous, n = 1; nausea, n = 1; upset stomach, n = 1; stomach cramping, n = 3).All reported events were also evaluated for severity where three of the reported adverse events were rated as minor severity and six as a severity of minor-to-moderate.A meta-analysis by Dore et al. 21reported increased relative risks of total reported side effect and those adverse events that specifically centered upon gastrointestinal symptoms and abdominal pain in those people who supplemented with a probiotic versus those who supplemented with a placebo.As such, the reported adverse events after VA supplementation of mild-to-moderate severity align with these findings and results of other studies surrounding reported adverse events with probiotic use.

Randomization
Study attrition resulted in five participants being supplemented with PLA first while two participants supplemented with VA first.To identify if an order effect was present, a categorical variable was created on the assigned order (1 = PLA then VA, 2 = VA then PLA), and all ANOVAs were performed with order as an additional fixed effect to identify if any significant interactions were present.No significant interactions were identified (all p > 0.05) for any of the primary or secondary outcomes.

Dietary intake
During each visit, participants were questioned by study investigators about any changes in their exercise training and their compliance with the dietary considerations put forth as part of this research study.All participants reported 100% compliance with completing food records and replicating food and drink intake prior to each study visit.

Hemodynamics
Changes in resting heart rate values indicated a significant main effect for time (p = 0.01), no main effect of condition (p = 0.51), and no significant group 3 time interaction (p = 0.51).The significant main effect was decomposed using paired samples t tests, which revealed that resting heart rate values for PLA did not change (mean difference G SE: 1.3 G 9.1 beats per minute; 95% CI: À7.2, 9.7 beats per minute, p = 0.72) in response to supplementation, whereas heart rate values in VA tended to decrease (mean difference G SE: 5.1 G 5.8 beats per minute; 95% CI: À0.2, 10.5 beats per minute, p = 0.06, partial eta squared = 0.45).Changes in resting systolic blood pressure indicated no main effect for time (p = 0.88) or condition (p = 0.93), whereas the group 3 time interaction tended to be significant (p = 0.07).Forced posthoc comparisons were completed using paired samples t tests to evaluate within-group changes in response to supplementation.Systolic blood pressure values did not change in PLA (mean difference G SE: À5.0 G 9.2 mm Hg; 95% CI: À13.5, 3.5 mm Hg, p = 0.20) or VA in response to supplementation (mean difference G SE: 3.6 G 15.8 mm Hg; 95% CI: À11.0, 18.2 mm Hg, p = 0.57).The group 3 time interaction for diastolic blood pressure was not significant (p = 0.15).No significant main effects for time (p = 0.95) or condition were observed (p = 0.84).All hemodynamic data are presented in Table 1.

Anaerobic capacity (time to exhaustion)
No significant group 3 time interaction was observed for time to exhaustion times (VA-PRE: 346.

Lactate changes
Lactate responses were first evaluated using a group [PLA vs. VA] x supplementation status [Pre-Supplementation vs. Post-Supplementation] x Time [Pre-Exercise vs. Immediate Post vs. 5 min Post] and are provided in Table 2.There was no significant main effect for group (p = 0.96), Supplementation status (Supp, p = 0.21), whereas a significant main effect of time (p < 0.001) was observed.No significant two-way interactions (group x supp, p = 0.91; group x time, p = 0.78; Supp x Time) or three-way interactions (group x Supp x Time, p 0.98) were observed.Lactate responses significantly increased under all Group and Supp combinations (See Figure 2 and Table 2, all p < 0.001).Separate 2 3 2 mixed factorial ANOVAs with repeated measures on condition were also evaluated to assess between group differences in lactate values collected before, immediately after, and 5 min after completing the time to exhaustion trial.At the pre-exercise time point, lactate values displayed no significant group [PLA vs. VA] x supplementation status [pre-supplementation vs. post-supplementation] interaction (p = 0.96), main effect for supplementation status [pre-supplementation vs. post-supplementation] (p = 0.46), or main effect for condition [PLA vs. VA] (p = 0.75).When each group was looked at individually, no differences were reported between pre-supplementation and postsupplementation pre-exercise lactate values for PLA (p = 0.56) or VA (p = 0.54).
At the immediate post-exercise time point, lactate values displayed no significant group [PLA vs. VA] x supplementation status [pre-supplementation vs. post-supplementation] interaction (p = 0.89), main effect for supplementation status [pre-supplementation vs. post-supplementation] (p = 0.21), or main effect for condition [PLA vs. VA] (p = 0.78).When each group was looked at individually, no differences were reported between pre-supplementation and post-supplementation lactate levels immediately after exercise for PLA (p = 0.11) or VA (p = 0.53).
At the 5 min post-exercise time point, lactate values displayed no significant group [PLA vs. VA] x supplementation status [pre-supplementation vs. post-supplementation] interaction (p = 0.99), main effect for supplementation status [pre-supplementation vs. post-supplementation] (p = 0.38), or main effect for condition [PLA vs. VA] (p = 0.43).When each group was looked at individually, no differences were reported between pre-supplementation and post-supplementation lactate values 5 min after completing the time trial for PLA (p = 0.59) or VA (p = 0.50).All data associated with the hematological and clinical safety markers are provided in Table 3.A significant group 3 time interaction was observed in the percentage of eosinophils present in the collected samples (p = 0.03).No main effect for time (p = 0.74) or condition (p = 0.26) was observed.When group changes were evaluated separately, the eosinophils percentage significantly decreased in PLA (p = 0.02), whereas no changes were observed in VA (p = 0.19).A significant group 3 time interaction was observed in mean corpuscle hemoglobin content (p = 0.03).A significant main effect for time (p = 0.04) was observed, and the main effect for condition tended to be significant (p = 0.06).When changes in each group were evaluated separately, the mean corpuscle hemoglobin content in PLA (p = 0.21) did not change, whereas VA values significantly changed (p = 0.01).All values for both eosinophils % and mean corpuscle hemoglobin content stayed within clinically accepted normative values for both variables.

Fecal metagenomics and metabolomic analysis
Metagenomics and metabolomics analyses of all collected stool samples were performed to evaluate how Veillonella supplementation might affect the microbiome community.Shannon entropy was used as a measure of alpha diversity (the richness of the microbiota community in terms of the number and abundance of the different bacterial communities that were present).In general, higher alpha diversity is associated with health, and reduced diversity is associated with a variety of chronic and acute health conditions.There were no significant differences in alpha diversity between the various time points (all p > 0.05), suggesting that the community's diversity did not change with the supplementation protocol employed (Figure 3A).To further confirm that there are no changes in overall community structure with VA supplementation, beta diversity (similarity of a bacterial community to another bacterial community) was qualitatively evaluated at different time points compared with baseline (Figure 3B).For four participants, limited to no changes in beta diversity were observed over time, although large increases in beta diversity were observed in VP02, VP07, and VP09 after Veillonella supplementation.Overall, no changes in specific taxa or functions throughout the study were observed after placebo use, the washout, or Veillonella use.Pearson correlations were completed to evaluate any potential relationships between changes in beta diversity and changes in time to exhaustion after VA supplementation (Figure 4).Overall, there was no correlation (r = 0.09) for this relationship.
Longitudinal changes in the microbiome community for each participant were assessed to identify potential non-significant patterns, and although the microbiome did change longitudinally in all participants (some with dramatic changes), no discernable associations with VA supplementation were noted.Metabolomics data provide information on changes in stool metabolites.Variable importance in projection (VIP) scores were used to identify 14 metabolites (Table 4) that were significantly discriminated between Veillonella use and baseline and Veillonella use and placebo but were not changed between placebo and baseline.Many of these metabolites are amino acids, and the majority of the others are metabolites found in foods.Although it is possible that the origin of these molecules is bacterial, and not the diet, the simplest explanation for these results is that these results are due to spurious changes in the participants' diet (even though participants were told to replicate their diet and indicated 100% compliance in doing so), not due to the action of Veillonella.

DISCUSSION
Using a randomized, double-blind, placebo-controlled, crossover manner, we examined the impact of a 14-day supplementation regimen of V. atypica FB0054 (VA) on anaerobic capacity (treadmill time to exhaustion), lactate responses to intense exercise, and clinical markers of health in healthy volunteers.The primary findings from this work indicate that a 14-day period of VA supplementation led to no change in anaerobic capacity (treadmill time to exhaustion), whereas PLA supplementation resulted in a tendency for anaerobic capacity to worsen.Supplementation was well tolerated with no clinically meaningful changes in clinical safety biomarkers.This is the first time that a species of live Veillonella has been used in a human population.Traditionally, probiotic supplements, the most common of which include lactic acid bacteria and bifidobacteria, have been isolated from food or animal sources. 22VA was isolated from elite marathon runners and approved for use in humans through the self-GRAS (generally recognized as safe) pathway, which involved whole genome sequencing and annotation to identify potential threats to human health, stringent toxicological testing, 23 and approval by an independent committee.This process was affirmed whereby all reported adverse events were mild to moderate in severity (no serious adverse events were reported) and gastrointestinal in nature.Further, the severity and incidence of reported adverse events were consistent with other forms of probiotic supplementation. 21,24upplementation with VA did not impact anaerobic capacity, whereas anaerobic capacity tended to worsen with PLA (Figure 5).This finding, although statistically non-significant, does align with previous research involving live probiotics (both single and multi-strain) that have investigated the role of probiotic products on various measures of exercise performance. 25For example, Pugh and colleagues 26 examined the effect of a multi-strain (Lactobacillus acidophilus CUL-60, Lactobacillus acidophilus CUL-21, Bifidobacterium bifidum CUL-20, and Bifidobacterium animalis subspecies lactis CUL-34 at a dose 2.5 x 10 10 CFU/day) probiotic (Proven Probiotics, Port Talbot, UK) on substrate metabolism and exercise performance in seven trained cyclists.No changes in markers of gastrointestinal damage and permeability were observed, but changes in carbohydrate and fat oxidation were observed, whereas time-trial performance was unaffected.Lactobacillus plantarum TWK10 is a probiotic strain isolated from Taiwanese pickled cabbage.On multiple occasions, 12,27,28 this strain has demonstrated the ability to improve treadmill run to exhaustion performance and biomarkers associated with fatigue in a dose-dependent manner in a population without a history of regular exercise training.Lactobacillus casei Shirota, which is sold as part of yogurt drinks, has been studied in athletes in the context of improving performance through immunomodulation and infection prevention [29][30][31][32] but has not improved exercise performance.A multi-strain-synbiotic (1 3 10 10 CFU/day each of Lactobacillus acidophilus CUL-21 and Lactobacillus acidophilus CUL-60; 9.5 3 10 9 CFU dose/day of Bifidobacterium bifidum CUL-20; and 5 3 10 8 CFU/day of Bifidobacterium animalis subspecies lactis CUL-34, Bifidobacterium bifidum CUL-20, Bifidobacterium bifidum CUL-20, and Bifidobacterium lactis CUL-34; BioAcidophilus Forte, Biocare Ltd., Birmingham, UK) has been tested for the ability to decrease   endotoxin levels and increase performance in recreational athletes. 33While endotoxin levels did decrease significantly, mean run times between groups exhibited a strong tendency to improve, but ultimately failed to reach statistical significance, in the probiotic group versus placebo.Finally, Strasser et al. 34 reported in trained athletes that supplementation with a multi-strain probiotic (1 x 10 10 total CFU per day of Bifidobacterium bifidum W23, Bifidobacterium lactis W51, Enterococcus faecium W54, Lactobaccilus acidophilus W22, Lactobacillus brevis W63, and Lactococcus lactis W58; Ecologic Performance, Winclove B.V., Amsterdam, The Netherlands) reduced the incidence of upper respiratory tract infections but did not benefit athletic performance.In the context of these mixed results and the pilot nature of our study alongside its small sample size, the results of this study are not unexpected.Many challenges exist when examining the ability of probiotics to influence exercise performance. 24As a starting point, choosing the right endpoint is a significant challenge.Based on previous in vivo work and the anticipated mechanism of action of VA, we chose a run-to-exhaustion model to assess performance.However, in the placebo group, we saw a decrease in time to exhaustion over time, when we would expect to see no change.This unanticipated observation may be due more to the underlying exercise psychology whereby untoward or desirable feelings associated with completion of intense exercise are rewarded by terminating participation.In addition, the study design required four separate experimental visits to the laboratory where the time to exhaustion protocol was included (five including the familiarization trial), thus the mental fatigue surrounding completion of the exercise test may have impacted motivation on subsequent tests.Importantly, we randomized the order of treatment administration to avoid an order effect, which was confirmed in a separate three-way ANOVA performed on the time to exhaustion data.We purposefully chose an exercise intensity (100% VO 2 Peak) that would generate significant production of lactate, but in hindsight, this intensity may have been too high for a few key areas.First, it is well established that cardiac output is redistributed away from the viscera to provide more blood to the working muscle.This physiological response would have led to a significant reduction in blood flow and ultimately limited the ability of VA to metabolize the lactate and offset performance.Thus, we feel that the high exercise intensity may have pushed our participants too quickly and aggressively into anaerobic metabolism, which simply may not have allowed enough time for VA to metabolize lactate.For these reasons, we feel a longer bout of exercise at a challenging (but not maximal) intensity would increase circulating lactate across a longer period and thus would create a richer environment for VA to metabolize to various metabolites.While more work is needed to further verify this consideration, previous research using the TWK10 strain (which has consistently demonstrated ergogenic outcomes) has utilized lower running intensities (60%-70% VO 2 Peak) for longer periods of time. 12,27Along these lines, we have previously posited that it may not be the lactate production and subsequent consumption by metabolic tissues that effects endurance so much as the production of propionate, a short-chain fatty acid energy source for the human host, which could increase endurance.Previous in vitro and in vivo models have linked propionate to mitochondrial (C-F) Although there are differences between the microbiome of the participants, there were no significant patterns in terms of changes in the microbiome of each participant.(C) Four participants (VP01, VP02, VP08, and VP09) had Bacteroides as the most abundant genera in their community throughout the study.(D) One participant (VP03) had Prevotella as the most abundant genus throughout the study.(E) One participant (VP07) had large microbiome changes occur during the study.In this participant, the observed bacterial communities at baseline were dominated by Bacteroides and Alistipes, whereas at post-Veillonella treatment, they were dominated by Prevotella.This participant (VP07) also had the greatest improvement in time to exhaustion.(F) One participant (VP06) had Subdoligranulum as the most abundant genera throughout the study.changes 35,36 and based upon these findings, selecting a longer exercise bout may have allowed more time for propionate to be produced and exert its potential metabolic or ergogenic potential.Certainly, more research is needed to fully explore this possibility.
Beyond the exercise task that was selected, choosing the right population with the appropriate baseline microbiome composition may change how the probiotic affects the host.The composition of the microbiome is determined by host genetics, immune system, diet, lifestyle, and the microbes already resident in the gastrointestinal tract. 2,3The ability for any probiotic strain to exert a positive effect on the host is dependent on all these factors.In this work, we studied a population that was already aerobically fit, lean, and physically active.Thus, it is possible our sample may have already had Veillonella present in the gut, reducing the need for VA supplementation.To this point, analysis of the bacterial communities observed in our collected fecal samples (Figure 3) did not appear to change in response to VA or PLA supplementation.Notably, this may also have been an explanation for the lack of changes observed for exercise performance.In a more sedentary population or a population with low relative abundance of Veillonella, we might expect to see more improvement in endurance and a greater dynamic range of response compared with placebo.
An additional key consideration was our selection and timing of biomarker assessment.We assessed capillary lactate before, immediately after, and 5 min after completing the run to exhaustion.Using this approach, we observed no change in capillary lactate levels between the placebo and VA conditions.In considering this further, one must realize that the majority of lactate is recycled through the liver and resulting plasma concentrations depend upon participant training status and style as well as the intensity and duration of completed exercise and potential differences in the presence and activity of monocarboxylate transporters, which move lactate in and around various tissues. 37These factors alone highlight the difficulty of tracking changes in lactate in response to open-ended exercise tasks as the duration upon which the lactate is being produced has the potential to vary quite a bit between participants and even within participants due to open-ended exercise tasks exhibiting much higher coefficients of variations and lower reliability. 38This was anticipated and explains why participants practiced the time to exhaustion two times prior to their first actual time to exhaustion trial.The key point nonetheless is made that more variation within and between exercise performance will ultimately impact the lactate kinetics during the exercise bouts.Although lactate levels commonly peak approximately 5 min after intense exercise, further consideration of our results against the proposed mechanism highlights the potential need for future investigations to consider lactate measurements for longer periods of time after exercise completion.To this point, evaluation of lactate 15, 30, and 60 min after completion of the exercise bout may have afforded better opportunity to evaluate if VA supplementation was able to metabolize the available lactate.
As mentioned previously, no changes were observed in the bacterial communities assessed in our collected stool samples after performing metagenomic and metabolomic evaluations (Figure 3).Directly related to VA supplementation, widespread changes in the microbiome were not anticipated as Veillonella is a small intestinal bacteria found in relative low abundance in our stool.After supplementation, we would anticipate that VA would engraft in the small intestine (not the large intestine).As a result, large changes in the biome would be surprising after the addition of one species of bacteria.A small number of stool metabolites significantly discriminated between post-Veillonella use and baseline or post-placebo.The majority of these were amino acids (Table 4), and the others were associated with foods.Thus, any associations between these metabolite changes and VA supplementation may be spurious diet-related changes and not specifically microbiome-mediated metabolite changes.
In conclusion, this pilot study of probiotic V. atypica FB0054 in humans shows supplementation did not improve endurance performance but did seem to be able to offset performance decrements observed in the PLA condition without any difference in lactate responses between supplementation.In addition, supplementation of the strain within the confines of our prescribed supplementation regimen (1 3 10 10 CFU daily dose for 14 days) appears to be safe and tolerable.Due to the small pilot nature of our project and the small observed effect on our assessed performance effect, readers are encouraged to view our conclusions with a preliminary mindset.Additional research is needed to explore FB0054, the ergogenic potential of V. atypica, more fully as changes in the population assessed, intensity of exercise employed, and using a closed vs. open-endpoint exercise test could also yield different outcomes.

Limitations of the study
The results of this study are limited by the small sample size.That being said, our chosen sample size aligned with other intervention studies involving probiotics using crossover 26 and parallel 27 designs, but the need for a larger small sample size would have helped to further strengthen the observed outcomes.For the current study, each participant supplemented for 14 days with a three-week washout.The duration of the supplementation regimen was selected based on results observed from two recent investigations by our research group involving a similar study design and probiotic supplementation 39,40 as well as a recent literature review on probiotics Mohr and colleagues. 41However, it should be highlighted that the majority of the investigations reported in this review administered probiotics strain(s) for longer periods of time than we did in the present study, thus a longer supplementation period using this strain may have impacted our outcomes.A final factor impacting our outcomes may have been the diverse range of ages in our participants as aging has been shown to impact the composition and quantity of various microbes found in our guts. 42Although a full understanding of how age may impact how the gut responds to exercise remains to be fully established, it seems reasonable to highlight that recruiting a cohort of individuals with a narrower age range may have helped reduce any confounding impact age had on our observed outcomes.Finally, challenging exercise is well established to impact gut permeability, 43 and many studies have demonstrated probiotics ability to improve gut permeability. 25While participants were asked to self-report adverse events and the common side effects connected to poor gut permeability, future research should more closely explore changes in biomarkers associated with gut permeability.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following: All recorded events were systematically categorized using MedDRA system organ class and lowest level terms (LLT) before being graded using Common Terminology Criteria for Adverse Events (CTCAE), v5.0, US Dept Health & Human Services (published: November 27, 2017).

QUANTIFICATION AND STATISTICAL ANALYSIS
All statistical analysis was completed in a blinded fashion.Primary endpoints for this study were treadmill time to exhaustion and lactate responses to the time to exhaustion trial.Secondary endpoints included hemodynamic responses and tertiary outcomes were changes in hematological, clinical safety, and adverse event outcomes.All data is presented as means G standard deviations and analyzed using IBM SPSS 27 (Armonk, NY USA) and graphs were generated using GraphPad (La Jolla, CA) or custom python scripts (Python 3.9.13,numpy 1.21.5, pandas 1.4.4., seaborn 0.11.2).A p value of 0.05 was used to make all statistical determination and any p value between p = 0.051-0.10was considered a statistical trend.Differences in Shannon diversity were determined by the Kruskal-Wallis test.Normality was assessed using Shapiro-wilk normality tests.Within-within (condition x time) factorial ANOVAs with repeated measures on condition and time with Least Significant Difference adjustments for multiple comparisons were used to assess differences between conditions at each individual time point for time to exhaustion, hemodynamics, all clinical safety parameters, and lactate changes.If the assumption of heteroscedasticity for repeated measures were violated, a Greenhouse-Geisser correction factor was applied.Paired samples t-tests were used to compare differences before and after supplementation for each condition.Grubbs outlier tests were completed on all data to evaluate the presence of any statistical outliers that may otherwise impact the reported outcomes.Pearson correlations were completed to evaluate the strength of any relationships.

ADDITIONAL RESOURCES
This study protocol was retrospectively registered on ClinicalTrials.orgas NCT05816291 on April 13, 2023.

Figure 1 .
Figure 1.Individual responses for the observed changes in treadmill time to exhaustion for each supplemental condition Control = Placebo (PLA); Veillonella = Veillonella atypica (VA).Each respective participant has the same color symbol for each condition.The large horizontal bar is the average change in treadmill time to exhaustion (post-to pre-supplementation) for each respective condition.

Figure 2 .
Figure 2. Capillary lactate concentrations in response to time to exhaustion trial at 100%VO2Peak Solid and dotted red lines are associated with the mean changes in control = Placebo (PLA).Solid and dotted blue lines are associated with the mean changes in Veillonella atypica (VA).Transparent shading underlaying the mean responses illustrate individual patterns of change for each participant.Pre-Ex = pre-exercise lactate measurement; Immed Post-Ex = immediate post-exercise lactate measurement; 5 Min Post-Ex = 5 minutes post-exercise lactate.

Figure 3 .
Figure 3. Individual participant microbiome data after baseline, placebo, and Veillonella atypica FB0054 supplementation (A) Using independent t tests, no significant changes (all comparisons p > 0.05) in alpha diversity were identified for participants throughout the study.(B) Beta diversity (similarity of samples to other baseline samples) did not change throughout the study, although there were striking increases in beta diversity post-Veillonella in a subset of participants.(C-F)Although there are differences between the microbiome of the participants, there were no significant patterns in terms of changes in the microbiome of each participant.(C) Four participants (VP01, VP02, VP08, and VP09) had Bacteroides as the most abundant genera in their community throughout the study.(D) One participant (VP03) had Prevotella as the most abundant genus throughout the study.(E) One participant (VP07) had large microbiome changes occur during the study.In this participant, the observed bacterial communities at baseline were dominated by Bacteroides and Alistipes, whereas at post-Veillonella treatment, they were dominated by Prevotella.This participant (VP07) also had the greatest improvement in time to exhaustion.(F) One participant (VP06) had Subdoligranulum as the most abundant genera throughout the study.

Figure 4 .
Figure 4. Scatterplot demonstrating beta diversity changes in response to Veillonella atypica FB0054 supplementation against changes in time to exhaustion (TTE).Pearson correlations (r) were computed between the change in time to exhaustion and the Bray-Curtis distance between pre-and post-Veillonella.
Recorded diet records indicated participants consumed 2190 G 913 kcal per day, 212.5 G 135.2 g of carbohydrate per day, 113.3 G 135.2 g of protein per day, and 100.8G 49.3 g of fat per day.

Table 1 .
Research activities by study visit

Table 3 .
Hemodynamic, hematological, and clinical chemistry parameters (Continued on next page)

Table 3 .
Continued (Continued on next page)

Table 4 .
Metabolites that discriminate between Veillonella treatment and baseline and/or placebo B Data and code availability d EXPERIMENTAL MODEL B Study participant details d METHOD DETAILS B Procedures d QUANTIFICATION AND STATISTICAL ANALYSIS d ADDITIONAL RESOURCES