Variations in Urinary Microbiota on a Short-Duration Space Mission to the International Space Station

Abstract

Introduction:

Space travel exposes crew to substantial stressors, which may potentially alter their microbiome and be detrimental to their health and safety. We hypothesize that urinary microbiota is altered during a short-duration spaceflight to the International Space Station (ISS).

Methods:

Urine samples were collected from 4 male crew members. One crew member provided samples while on orbit on the ISS using a urine collection kit (UCK) designed for low-gravity environments. This crew member also provided paired mid-stream and UCK samples on Earth prelaunch and on return for direct comparison of collection methods. Three additional crew members provided mid-stream samples prelaunch and on return, and also for follow-up timepoints to serve as ex post facto control samples. Urine was stored at −80℃ and sent for 16s next-generation sequencing (MicroGenDX, Lubbock, TX).

Results:

Bacterial load (P = .52), species richness (P = .64), and differences in microbiota composition (P = .26) did not significantly vary by the collection method (mid-stream vs UCK). Significant differences in species richness, load, and microbial composition among preflight, orbital, and return mission samples were detected (P = .047, F = 1.8092, df = 2.9, R2 = 0.34076). Orbital microbial composition significantly differed from preflight (P = .03) and return (P = .03), whereas preflight and return samples did not differ (P > .05). Comparison with the limited follow-up ex post facto Earth-bound samples serving as controls showed no significant differences among preflight or return mission samples, but preflight (P = .04) and return (P = .05) did significantly differ compositionally from orbital samples. Continuous temporal effects on composition were evaluated through which bacterial load (P = .05) was statistically significant, peaking during the orbital period.

Conclusions:

Results are consistent with an effect of space travel altering the urinary microbiota. Further studies are needed to confirm this result and to better understand whether such changes may be detrimental to the health and safety of space travelers and professional astronauts.

Urine, commonly believed to be sterile, frequently harbors a polymicrobial complement of microorganisms (ie, the microbiome).^1,2 The majority of these microorganisms were previously undetected using traditional microbiology culture techniques. However, rapidly developing sequencing methods, commonly referred to as next-generation sequencing (NGS), in conjunction with bioinformatic techniques are facilitating our ability to understand and characterize the human microbiome.^2,3 NGS affords the potential to characterize microorganisms present in a sample, including species' abundances.^4,5

Characterizing and understanding the microbiome is important because alterations in microbial composition and diversity may affect health.⁶ Several studies have identified that changes in the urinary microbiota may be directly contributory to some of the most frequent and bothersome urological symptoms including kidney stones, lower urinary tract symptoms, urinary tract infections, and even bladder pain syndrome.^3,7,8 Moreover, there is emerging evidence suggesting the development of genitourinary cancers when there are alterations in the urinary microbiome.⁹

Space travel exposes crew to substantial stressors. In the setting of microgravity, urine does not typically collect at the bladder neck, and the surface tension of urine adhering to the bladder wall causing detrusor distention becomes to the main driver to trigger micturition. The resultant urinary stasis, elevated post void residuals, delays in schedule of waste control availability despite use of maximum absorbency garments and the potential changes in urinary microbiota all lends itself as risk factors for UTI development.¹⁰ Two studies have previously demonstrated alterations of crews' gut, skin, nose, tongue, and other skin sites' microbiome during space travel, but did not evaluate urine.^11,12 Specifically, these studies noted a drop in abundance in a few bacterial taxonomies, and clinically, these patients also noted a higher frequency of skin hypersensitivity and rashes due to changes in skin microbiome. Given the perceived importance of the microbiome on urinary health and various symptoms, understanding space travel-induced changes to the urinary microbiome is important for the safety of travelers and professional astronauts.

We aim, for the first time in the literature, to evaluate changes in the urinary microbiota in crew members on a short-duration spaceflight to the International Space Station (ISS) using NGS. The nature of our study is primarily explorative. Yet, we hypothesize that space travel alters the urinary microbiota measured by relative and absolute abundance and diversity of organisms, with specific attention to bacterial load, species richness, and microbial composition.

MATERIALS AND METHODS

Study Cohort and Questionnaire Data

Our study cohort comprised of 4 healthy male crew members on a short-duration spaceflight to the ISS from April 8, 2022 to April 25, 2022. Each crew members consented to participate in this study approved by the institutional review boards of NASA (STUDY00000411), Sheba Medical Center (8972-21), and Thomas Jefferson University (21D.1072 and iRISID-2023-2592). All crew members completed diet and urinary symptoms questionnaires (Supplementary File, http://links.lww.com/JU9/A178).^13,14

Mission Period and Ex Post Facto Control Urine Sample Collection

Mission period mid-stream urine samples were collected from all 4 crew members before launch (L-13, L-7, L-1) and after return to Earth (R+0, R+1-3; Supplementary Table 1, http://links.lww.com/JU9/A175). Crew member #1 provided urine samples while on orbit on the ISS (O+2, 4, 6, 8) through a urine collection kit (UCK) designed for NASA to facilitate urine collection in a low-gravity environment (Supplementary Figure 1, http://links.lww.com/JU9/A173).¹⁵ Crew member #1 also provided a paired UCK urine sample for each mid-stream urine sample for direct comparison of collection methods prelaunch and upon return to Earth.

Ex post facto control mid-stream urine samples were collected from crew members #2 to 4 during a postmission timeframe (February 21, 2024 to April 25, 2024) to compare with mission period samples. Five ex post facto control samples were collected to match the timeframe of the L-13, L-7, L-1, R+0, and R+3 mission period samples. All urine samples were frozen and stored at −80°C until being shipped on dry ice for NGS testing (MicroGenDx, Lubbock, TX).

Molecular Testing

Total genomic DNA was extracted using a Kingfisher Flex automated platform (Thermo Scientific, Waltham, MA) and ZymoBIOMICS-96 MagBead DNA extraction reagents (Zymo, Irvine, CA). PCR amplifications were selective for the 16S rRNA hypervariable regions V1-V2 using primers 28F (GAGTTTGATCNTGGCTCAG) and 388R (GCTGCCTCCCGTAGGAGT) using Quanta AccuStart II Tough Mix (Quanta bio, Beverly, MA). PCR reactions were conducted on ABI Veriti thermocyclers (Applied Biosystems, Carlsbad, CA) with a thermal profile consisting of 5-minute denaturation step at 95°C, 35 cycles of 94°C for 30 seconds, 52°C for 40 seconds, and 72°C for 60 seconds, and a final extension step of 72°C for 10 minutes. PCR products were grouped equal molar and selected by size in 2 rounds using Agentcourt AMPure XP (Beckman Coulter, Indianapolis, IN) in a 7-10th ratio of AMPure to product. Quantification of each group was performed using a Qubit 2.0 fluorometer (Thermo Fisher, Waltham, MA), and pooled libraries were sequenced on an Illumina MiSeq using 2 × 250 bp chemistry (San Diego, CA).

Bioinformatic processing of sequences was similar as previously described.¹⁶ Denoising of sequence reads, chimera detection, and paired read assembly were conducted using Usearch7, UCHIME, and PEAR, respectively.^17-19 Quality filtered and assembled reads were clustered into operational taxonomic units (OTUs) at 97% sequence similarity threshold using the UPARSE algorithm.²⁰ OTU assignment then used an in-house curated taxonomic reference database. Contaminant screening was achieved using the DECONTAM library55 protocol for prevalence-based filtering (threshold = 0.1) in comparison with DNA extraction and PCR negative controls.²¹ Next, any species that were significantly more abundant in controls compared with crew samples were also removed.²² Scaled with ranked subsampling rarefaction was performed to normalize sequencing effort across samples.²³ Next, Good's Evenness of Coverage was used to estimate how well-sequencing effort characterized microbial profiles present in samples.²⁴ Absolute abundance (bacterial cells per mL of urine) was then calculated for each species for each sample by comparison with ZymoBIOMICS Low Bacterial Load Spike-in Control II (Zymo Research, Irvine, CA), which was added (20 μL) to each urine sample before DNA extraction. Calculation of absolute bacterial cell count per mL of urine was calculated by multiplying the number of Imtechella halotolerans 16S copies spiked-in (3 × 10⁴) by (1/Imtechella halotolerans relative abundance–1), then dividing the product by 3 (assumes an average of 3 16S copies per genome), then dividing by the number of milliliter of urine for each sample.²⁵ From the final estimates of bacterial cells/mL urine for each sample, the absolute cell counts of individual species was obtained by attributing cell counts proportional to each species relative abundance.

Data and Statistical Analysis

Days of collection were referenced relative to launch date, and time points were also categorized into preflight, ISS orbital samples, and return samples. Differences in species richness, species diversity summarized as Hill₁ numbers, and bacterial load (cells/mL urine) as a function of collection method or time group were tested using ANOVA with crew member #1 (the crew providing samples for all time groups). After nonsignificant effects of collection method, only UCK samples for crew member #1 were retained for subsequent analysis including the other crew members so as not to confound collection method and time group variables. Differences in microbiome compositional profiles among samples were calculated as Bray-Curtis community dissimilarities, and permutational analysis of variance (PERMANOVA) was used to test for the effect of time group, and treatment.^26,27 Here, samples collected around and during the journey to the ISS are referred to as the mission samples, and those collected during the Earth-bound follow-up period are referred to as the ex post facto control samples. The comparison of mission and post facto control sets was accomplished by creating a composite time group × treatment variable (eg, control preflight, experimental preflight, experimental orbital, control return, and experimental return) and pairwise PERMANOVA was used to assess group differences. Polynomial regression was used to test for effects of time on species richness, bacterial load, and the product of load and richness, with the latter being included to explore the possibility that effects of space travel may simultaneously influence both density and number of species in the closed-space urinary microbiome of crew members. Regression models were evaluated both for crew #1 samples only, as well as with the inclusion of all crew member samples. Statistical analyses were conducted in R version 4.1.0 (R Core Team, R Foundation for Statistical Computing).

RESULTS

Questionnaire Data

Four male crew members between the ages of 53 and 72 years participated. All crew members completed the questionnaires prelaunch. None of the members have a history of urinary tract infections, flank or pelvic pain, and pressure, dysuria, gross hematuria, urinary retention, history of catheterization, or diagnosis of benign prostatic hyperplasia. One crew member did have a history of nephrolithiasis within 6-month prelaunch requiring ureteroscopy and laser lithotripsy and took antibiotics during the immediate perioperative period. The baseline AUA Symptom Score (AUASS) ranged from 6 (mild) to 14 (moderate). No crew member reported adverse urinary symptoms or events during the mission. Questionnaire responses are available in Supplementary Table 2, http://links.lww.com/JU9/A176.

Summary of Microbial Detection

Decontamination screening resulted in a total of 114 OTUs and 53 corresponding species identified as likely contaminants and removed from subsequent analysis (Supplementary Table 3, http://links.lww.com/JU9/A177). After decontamination, stratified random subsampling rarefaction resulted in the retention of 836 sequences per sample. This sampling effort was summarized by Good's Evenness of Coverage ranging from 97.73% to 100%, indicating that communities were adequately summarized from sequencing effort. From a total of 23 urine samples, 124 total species were detected, with 99 species occurring at a relative abundance of at least 2% in at least 1 sample. Focusing on the set of 99 more commonly detected species, a mean of 6 species were detected per sample with a minimum and maximum of 1 and 17 species, respectively. From these 99 species, there were 5 phyla detected, with Proteobacteria (54), Firmicutes (38), and Bacteroidetes (5) being the most common phyla detected. Figure 1 illustrates the distribution of most common species across crew members and associated timepoints of urine collection.

Figure 1.

Stacked bar plot of study-wide most common species organized by crew member, timepoint, and control vs mission samples. Control samples were collected from crew members (#2-4) ex post facto to compare with mission samples from crew members (#1-4). Each position on the x-axes corresponds to a sample provided at the corresponding time from each crew member, and colors for each species indicate the relative abundance of each species. L-# refers to days before launch, O+# refers to days after launch, and R+# refers to days after return from orbit.

Comparison of Collection Methods

Crew member #1 provided several paired, mid-stream, and UCK samples, both prelaunch and upon return to Earth to assess whether the collection method noticeably influences microbiome results. First, for these samples, differences in log of bacterial load (cells/mL) based on collection method and time of collection (before or after mission) were tested using ANOVA. For this analysis, the collection method was not significant (P = .41), but time of collection (P = .04) was significant. The same comparisons were also conducted for species richness, but neither collection method (P = .56) nor time of collection (P = .21) was statistically significant. Figure 2 depicts these results. Comparisons for Hill₁ were also nonsignificant for the collection method (P = .97) but significant for time of collection (P = .005). Similarly, for the Pielou evenness, there was no effect of collection method (P = .43), but time was significant (P = .02). Differences in microbiota composition (Bray-Curtis dissimilarities) among samples owing to collection method were similarly tested with PERMANOVA, through which the collection method was not significant (P = .26), but paired time of collection was significant (P = .04, F = 3.13, df = 3.7) explaining 34% of compositional variation. Supplementary Figure 2, http://links.lww.com/JU9/A174 provides a side-by-side summary of the microbiota composition detected in paired samples. To assess whether the nonsignificance of collection method was potentially due to noise introduced by intersample variation among rare species, those with relative abundances less than 2% were eliminated and the ANOVAs were repeated. None of these analyses resulted in significant effects of collection method. Still, the literature has discussed the likelihood of low-level introduction of species through collection methods or sample processing.²⁸ Based on this possibility, only species detected greater than 2% relative abundance were included in subsequent analyses. In addition, because the collection method was consistently a nonsignificant variable for crew member #1, and to not confound collection method and time group variables, only midstream urine samples for crew member #1 were used in subsequent comparisons.

Figure 2.

Crew member #1 provided paired, mid-stream, and UCK samples, both prelaunch and upon return to Earth to assess whether collection method noticeably influences microbiome results. Distribution of (A) bacterial load and (B) species richness across samples for method and time of collection, including species detected > 2% in any sample. C, Axes 1 and 2 of PCoA based on microbial composition (Bray-Curtis dissimilarities) across samples for method and time of collection, including species detected > 2% in any sample. L-# refers to days before launch, and R+# refers to days after return from orbit.

Temporal Trends

Focusing on crew member #1, mission time (preflight, ISS orbital samples, and return samples) significantly influenced log of bacterial load (P = .02, F = 7.07, df = 2.7, R2 = 0.67) and had notable effects on richness (P = .12) and Hill₁ (P = .13), but was not significant for the Pielou evenness (P = 3). In addition, the interaction of log of bacterial load and richness was significantly explained by time (P = .03, F = 6.43, df = 2.7, R2 = 0.66). In addition, microbial composition summarized as Bray-Curtis dissimilarities significantly differed among groups (P = .047, F = 1.8092, df = 2.9, R2 = 0.34076). Pairwise testing between time groups for compositional differences revealed that preflight and orbital samples significantly differed (P = .03), orbital and return samples significantly differed (P = .03), whereas preflight and return samples did not significantly differ (P = .9). Using principal coordinate analysis (PCoA), Figure 3 illustrates the compositional similarity across time for urine samples from all crew members. Orbital samples were the most compositionally clustered set of samples if comparison with the other time points; however, orbital samples were all from crew member #1. Comparison of experimental and control samples by time and group revealed that both control preflight (P = .04) and control return (P = .05) significantly differed from experimental orbital, but no comparisons between control and experimental preflight or return groups significantly differed (P > .05 in all 6 comparisons). We assessed the association of temporal effects (days of mission duration starting at L-13) on urinary microbial composition (bacterial load, species richness, product of load, and richness) using polynomial regressions including linear, quadratic, and cubic terms. These findings are depicted in Figure 4. Analysis of mission period samples from all crew members resulted in bacterial load being significantly explained by time in a second order polynomial term (P = .05).

Figure 3.

Plots of Axes 1 and 2 of PCoA based on microbial composition (Bray-Curtis dissimilarities) for samples from all crew members across time of urine sample collection and treatment. Samples are color-coded according to period of sample collection, with shapes corresponding to each crew member. Percentage of total correlated variation explained on each axis is provided.

Figure 4.

Measures of species richness, log transformed bacterial load (cells/mL), and their product in relation to timepoint, color coded by preflight, orbital, and return periods for each crew member separated by control vs mission samples is shown. Control samples were collected from crew members (#2-4) ex post facto to compare with mission samples from crew members (#1-4). The significant second order polynomial regression line for bacterial load is shown (P = .05).

DISCUSSION

The risk of urinary problems during spaceflight is real. A recent report evaluating urinary symptoms in space during the Mercury, Gemini, Apollo, Mir, Shuttle, and ISS expeditions 1 to 38 identified NASA astronauts experienced 16 cases of urinary retention and 9 reported urinary tract infections during spaceflight.²⁹ Furthermore, an astronaut with urinary retention is 25 times more likely to have a UTI with a 17% infection rate per mission.²⁹ Astronauts are also at risk for kidney stones with 7 cases reported in the 12 months after flight.²⁹ Another study evaluated urinary composition during spaceflight and identified the presence of a urinary environment that favored the supersaturation of stone-forming salts which may have increased the risk for kidney stones.¹⁵ The use of specialized suited body waste management systems which allows for additional moisture wicking capabilities may also play a role in the changes of the urinary microbiome.¹⁰ On this mission to the ISS, no crew members experienced adverse urinary events. Still, this study reported higher bacterial loads and shifted urinary composition during orbit, suggesting space travel disturbs the normal urinary microbiota and could contribute to increased incidence of UTI. These changes coupled together with the altered gravity vector, increased post void residuals, and several hygiene factors such as delayed access to voiding and different urine collection devices may play a role in documented cases of UTIs in previous expeditions.³⁰

A study evaluating the changes of astronaut's gastrointestinal microbiome demonstrated changes while on the ISS. They found that alpha diversity increased significantly during the 6-month period on the ISS, while compositional changes included relative abundance shifts and the loss and acquisition of microbial species in space. Their data also revealed that orbital changes to the microbiome occurred relatively early during space travel, and in most astronauts, this was evident by flight day 7.¹¹ A second study evaluating the microbiome during a short-duration spaceflight demonstrated alterations across several body sites, which included cutaneous, nasal and oral microbiome. Specific to the oral microbiome, the authors observed persistent, long-term shifts in several bacterial species which was associated with plaque formation.¹²

In our study, when comparing trends in bacterial composition and load among preflight, orbital and postflight samples, we also found that the orbital samples significantly differed, while there were no significant differences between preflight and postflight urine samples. We were able to demonstrate microbiome changes in the urine as early as 2 days in space. It is also notable that crew member #1's first postflight sample, especially bacterial load (sample collection point 7, Figure 4), was more like other orbital samples than postflight samples, a result consistent with the expectation that orbital effects on the urinary microbiome should still be measurable shortly after returning to Earth.

In addition to significant compositional differences among preflight, orbit, and return urine samples, we also found significant differences in bacterial load as well as the interaction of load and species richness over time peaking during the orbital period. Whereas microbiome studies typically do not consider absolute abundance, measuring load appears consequential to detecting effects of short-duration space travel. Infrequent urination may provide more opportunities for urinary tract bacterial accumulation, and because species interact and vary in replication rate, this may also influence community composition. Previous ecology literature has not considered the theoretical relationship between biomass (bacterial load) and species richness, but a few studies have measured the relationship between the 2 variables. For example, Tipton et al previously found a negative relationship between load and richness in a chronic wound setting, whereas Lowman et al graphically reported a positive relationship in the setting of equine sinusitis.^31,32 Although there is not a precedent for the approach of using the product (interaction) of load and richness as the dependent variable, it seems reasonable that these 2 variables may interact, especially in closed-space body sites, and their relationship could vary depending on synergism between colonizing microbes. Larger sample sizes in future studies will also allow explicit investigation of the interaction effects of biomass and species richness.

The nature of our study is exploratory and descriptive. There are inherent biases when sample sizes are small with limited availability of controls and all participants were male which does not allow for evaluation of sex on the urinary microbiome. Moreover, NGS testing is a sensitive technology, and the clinical significance the findings in this setting are unknown. However, it is important to consider the rare opportunity of obtaining urine samples from the few people who currently travel to space. Being private space travelers, and not professional astronauts, crew members from this study were able to elect whether to participate and select and exclude portions of the study. Nonetheless, this is the first report that describes the changes of the human urinary microbiota during space travel, which previous spaceflight microbiome studies did not. Although it may be challenging to replicate this study with a large sample size, future work could potentially conduct studies from astronauts being trained in simulators, as well as specific chemical changes in the urine that may lead to alterations in urinary microbiota. Future studies with larger cohorts will also allow for the evaluation of questionnaire data which could not be analyzed statistically in this study because of the small sample size. This pilot study is hypothesis generating and will help to create the foundation and pave the way for further research to come with the goal of protecting the urinary health of astronauts and future space travelers.

CONCLUSION

Space travel appears to alter the urinary microbiota. To validate our findings, and to understand consequences for long-duration space travel on the urine microbiota, follow-up studies based on future missions including Earth-bound controls in a balanced study design are needed.

FUNDING

Thomas Jefferson University, Sheba Medical Center, MicroGenDx, NASA, and Axiom Space provided funding in support of the research.

CONFLICT OF INTEREST DISCLOSURES

The authors have the following disclosures: JCN and JP scientific study/trial/consultant with MicroGenDx, PHC scientific study/trial with MicroGenDx.

ETHICS STATEMENT

Our study has been approved by the institutional review boards of NASA, Sheba Medical Center, and Thomas Jefferson University.

DATA AVAILABILITY STATEMENT

The data sets generated during and/or analyzed during this study are available from the corresponding author on reasonable request.

AUTHOR CONTRIBUTIONS

Conceptualization: Chung, Boursi, Baris, Nickel, Phillips. Data curation: Chung, Baris, Phillips. Formal analysis: Chung, Leong, Tipton, White, Ancira, Phillips. Methodology: Chung, Boursi, Parvizi, Phillips. Project administration: Chung. Writing–original draft: Chung, Leong. Writing–review and editing: Chung, Boursi, Baris, Leong, Nickel, Parvizi, Tipton, White, Ancira, Hochberg, Rifkin, Phillips. Supervision: Baris. Software: Phillips.