The impact of antibiotics on the presence of the protozoan anaerobeBlastocystisand the surrounding microbiome: a case study

23

Blastocystis, the most prevalent eukaryotic gut microbe in humans, has a global 24 distribution. Studies have linked its presence with distinct gut microbiome and metabolome 25 profiles compared to those where the organism is absent. However, the in vivo effect of 26 antibiotics on Blastocystis and the surrounding gut microbiome remains understudied. This 27 case study aimed to explore how antibiotic consumption influences the presence of 28 Blastocystis and the subsequent changes in the gut microbiome and metabolome of an 29 individual with irritable bowel syndrome (IBS). 30

Stool samples from an IBS patient, collected at various time points, were tested for 31 Blastocystis presence using RT-PCR targeting the SSUrRNA gene, followed by sequencing of 32 positive samples. Illumina sequencing determined the gut microbiome composition, while 33 one-dimensional proton NMR spectroscopy analysed the metabolome composition. Statistical 34 analyses were conducted to identify relationships between antibiotic consumption, bacterial 35 diversity, metabolome composition, and Blastocystis presence. 36

Antibiotics significantly impacted the gut microbiome, with diversity declining early 37 in the antibiotic course, then recovering later and post-course. Blastocystis was detected early, 38 late, and post-course but not mid-course, coinciding with the decline in bacterial diversity. No 39 significant differences were observed between Blastocystis-positive and Blastocystis-negative 40 samples. However, bacterial composition significantly differed between samples collected 41 before, early, and after the antibiotic course compared to those collected mid-course. 42 Metabolite groups, including short-chain fatty acids, amino acids, and succinate, exhibited 43 changes throughout the antibiotic course, indicating that gut metabolite composition is 44 affected by antibiotic consumption. 45 46 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 3 Discussion/Conclusion: While antibiotics did not significantly impact Blastocystis 47 colonisation, they did cause a mid-course decline in microbial diversity and Blastocystis 48 presence. The study also revealed significant alterations in important metabolites such as 49 SCFAs and amino acids throughout the antibiotic course, with an altered metabolome 50 observed post-course. This case study underscores the complex interactions between 51 antibiotics, gut microbiota, and metabolites, highlighting the resilience of Blastocystis in the 52 gut ecosystem. 53 54 55 56 57 58 59 60 61 62 63 64 65 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 4 66

67 The gut microbiome and metabolome are crucial influencers of gastrointestinal (GI) health, 68 and their interactions can be better understood by studying them in tandem [1]. This dual 69 approach is particularly important when investigating GI health differences across various 70 cohorts or when external factors influence the GI tract. For example, a previous 71 metabolomics study using ^1H Nuclear Magnetic Resonance (NMR) spectroscopy revealed 72 significant differences in stool metabolite composition between individuals with diarrhoea 73 and healthy controls and between Blastocystis carriers and non-carriers [2]. Understanding 74 these interactions can provide deeper insights into the complex dynamics of gut health and 75 disease. 76 Blastocystis is a eukaryotic microbe that resides in the GI tract and has a global distribution in 77 a broad range of animal hosts [3], [4]. Epidemiological studies and phylogenetic analysis of 78 the small subunit ribosomal RNA (SSU rRNA) gene have revealed over 44 different subtypes, 79 twelve of which STs 1-9, 12, 16, 23 have been identified in human stool samples [5], [6], [7], 80 [8], [9]. Blastocystis was initially designated as a parasite and linked with IBS and other 81 gastrointestinal disorders [10], [11]; however, more recent studies have indicated a negative 82 correlation between the presence of the organism and gastrointestinal symptoms [8], [9], [12], 83 [13], () muddling its association with disease. Blastocystis’ genetic diversity further 84 complicates interpretations, with many studies finding no relationship between inter and 85 intra-subtype diversity and disease [9], [12]. 86 Nonetheless, specific microbial profiles have been associated with the organism. For 87 instance, Blastocystis presence is more common in the Ruminococcaceae and Prevotella 88 enterotypes, rather than Bacteroides, and associated with higher richness and diversity, which 89 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 5 can be an indicator of good GI health [8], [14], [15]. At the level of subtype, Blastocystis ST3 90 and ST4 have been shown to have an inverse relationship with Akkermansia abundance, an 91 indicator of GI health [8]. Whether Blastocystis is a gut ecosystem engineer, a simple 92 colonizer or just a passenger is still unknown. 93 Individuals with irritable bowel syndrome (IBS) have been known to have distinct gut 94 bacterial compositions/profiles to their non-IBS counterparts, making IBS treatment with 95 antibiotics a potential influencing factor [16], [17], [18], [19]. In vivo and in vitro studies 96 have indicated that the gut microbiomes of individuals colonised with Blastocystis show a 97 decline in abundance of genera such as Bifidobacterium and Lactobacillus [15], [20], [21]. 98 Bifidobacterium has a role in immunomodulation and protection of the GI epithelial cells 99 [20], [22], [23]. Therefore, the results of these studies could link Blastocystis to dysbiosis-100 induced GI symptoms and possibly IBS. 101 Antibiotic administration has been an effective treatment for some IBS cases and other GI 102 conditions, while metronidazole, ciprofloxacin and rifaximin have been effective at 103 decreasing the severity of symptoms in many clinical trials [18], [19], [24], [25], [26]. 104 However, as our understanding of the gut microbiome's role in gastrointestinal health 105 deepens, the impact of antibiotics on gut microbiota modulation is receiving closer scrutiny. 106 Several studies have detected significant changes in the gut microbiome composition during 107 antibiotic treatment and, occasionally, microbiome recovery after treatment [27], [28], [29]. 108 In this case study, we monitored the metabolome and the bacterial gut microbiome 109 composition of a Blastocystis-positive IBS patient during a 14-day course of antibiotics. We 110 also analysed Blastocystis presence over this 14-day period and how it is impacted by 111 antibiotic treatment. The subject’s gut microbiome and metabolome composition were also 112 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 6 analysed following the termination of the antibiotic course to monitor microbial diversity 113 recovery after Blastocystis detection. 114 115

and Method 116 Ethics approval 117 The study was conducted within the guidelines established in IRAS ethics approvals 274985 118 and 286641, following a review by an ethics committee and applying suggested amendments 119 to comply with ethical standards. The UK National Ethic committees of Health Research 120 Authority (HRA) and Health and Care Research Wales (HCRW) under the umbrella of NHS 121 Health Research Authority gave ethical approval of this work. 122 123 Participant recruitment and sample collection 124 The study subject, previously diagnosed with IBS, presented to a hospital in the Kent county 125 (South East England) with gastrointestinal symptoms and was put on a 14-day course of 126 antibiotics. These included 500 mg Amoxicillin (a 3rd generation penicillin antibiotic) x 2 and 127 500 mg Clarithromycin (a 2 nd generation macrolide antibiotic). Daily doses of 30 mg 128 Lansoprazole (proton pump inhibitor) x 2 was also prescribed. The subject was in their 40s 129 (41 to 45) with a BMI of 29 and a mixed diet. The subject was provided with faeces catchers 130 (Zymo Research Cat No R1101-1-10) and two types of collection tubes, one containing 5 ml 131 DNA/RNA shield (Zymo Research Cat No R1100-250) and the other containing 5 ml of 50% 132 methanol. Each faecal sample was distributed in the two tubes. Faecal samples were collected 133 just before the commencement of the treatment course, then on D2 then once daily for the 134 remainder of the first week, then once on D8, D10 and D15 (the day after the completion of 135 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 7 the course) for the following week. Follow-up samples were then collected at D30 and D3M 136 (3 months after course start). The samples were stored in their respective tubes in DNA/RNA 137 shield or methanol at -80ºC. 138 139 DNA extraction 140 200 mg solid stool stored in DNA/RNA shield or 200 µl liquid stool were added to 200 µl 141 PBS (pH 7.4 RNAase free). The samples were then centrifuged for 10 minutes at 10,000 x g 142 at room temperature (RT). The pellet was then resuspended in the supernatant and the DNA 143 was extracted using the QIAamp PowerFecal Pro DNA Kit (Qiagen; Cat. No: 51804) 144 following the manufacturer’s protocol, and 50 µl DNA was eluted. 145 146 qPCR and Blastocystis detection 147 For Blastocystis detection, a 350 bp region of the SSU rRNA gene was targeted using a 148 reaction mixture of 2 µl DNA, 500 nM of primer set PPF1 (fwd) (5’-149 AGTAGTCATACGCTCGTCTCAAA-3’) and R2PP (rvs) (5’-150 TCTTCGTTACCCGTTACTGC-3’) and 5 µl SYBR green making a full reaction volume of 151 10 µl. The qPCR was run on a Quantstudio-3 real-time PCR machine with the following 152 program: initial denaturation 95 ºC for 5 minutes, then 45 cycles of initial denaturation 95ºC 153 for 5 seconds, annealing 68 ºC for 10 seconds, extension 72ºC 10 seconds then a final 154 extension of 72 ºC for15 seconds. 155 156 Sequencing and Subtype annotation 157 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 8 Bi-directional Sanger sequencing using the set of primers for the qPCR reaction was 158 outsourced to and performed by Eurofins (UK). The forward and reverse nucleotide 159 sequences were then assessed and trimmed using SnapGene Viewer Version 6.2.2 160 (https://www.snapgene.com/snapgene-viewer). The final trimmed consensus sequences were 161 then used as queries to check for contamination using the Basic Local Alignment Search Tool 162 (BLAST) from the National Centre for Biotechnology Information (NCBI) 163 (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Once the identity of the sequence was confirmed as 164 Blastocystis, the subtype was assigned using the curated database pubMLST 165 (https://pubmlst.org/organisms/blastocystis-spp). 166 167 16S rRNA gene amplicon sequencing 168 Novogene outsourced the high-throughput amplicon sequencing. The protocol used was 169 based on Caporaso et al [30]. [] with some modifications. One ng DNA from extracts was 170 used, fragmented, and then adapted for paired-end sequencing. The DNA was amplified 171 using the primer pair 515F GTGCCAGCMGCCGCGGTAA and 907R 172 CCGTCAATTCCTTTGAGTTT, which amplifies the hypervariable region and then 173 sequenced on the Illumina NovaSeq platform. 174 The raw reads were classified using the Lotus2 software [[31]]. The parameters and tools 175 used are as follows: Chimera checking/removal was performed using Minimap2 [[32]], and 176 Minimap2 was also used to look for off-target hits containing human DNA ‘contaminated’ 177 reads by BLASTing reads against Genome Reference Consortium Human Build 38.p14. V3-178 V4 region trimmed reads were then clustered into ASVs ( ≤ 1 nucleotide dissimilarity) using 179 Divisive Amplicon Denoising Algorithm 2 (DADA2) [[33]], ASVs were taxonomically 180 classified (to species level) using BLAST against the GreenGenes2 (GG2) database [[34]]. 181 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 9 GG2 was chosen for its reliability (GG2 is a unified database suitable for whole genome 182 sequencing (WGS) data and 16s data), as well as replicable results. 183 184 Statistical analysis 185 Statistical analysis and data visualisation were done using the R Studio 4.2.3 package. 186 Relative abundances of each genus were calculated in each sample, and a heatmap was 187 constructed. Diversity index values were calculated using the Phyloseq package. Shannon, 188 Chao1, Simpson and observed taxa values were used. These four values were analysed for 189 statistical differences occurring between the Blastocystis positive and Blastocystis negative 190 samples, as well as differences in diversity score between the ‘antibiotic positive’ time points 191 (days 4-15) and the ‘antibiotic negative’ time points (day 0, 30 days post antibiotics and 3 192 months post antibiotics). First, a Shapiro test was used to determine the data distribution to 193 analyse the statistical differences between the sample groups. Normally distributed data was 194 analysed with ANOVA test followed by Tukey HSD test for pairwise comparison. For 195 samples with a non-normal distribution, the Kruskal-Wallis test was used, followed up by the 196 Dunn Test (Bonferroni P-adjust) for pairwise comparisons. The raw diversity index values 197 were also plotted over time. To visualise microbiome composition, compositional plots 198 showing all taxa making up >1% of the total read counts were produced using the 199 Microbiome package. To look for the presence of ‘bio-markers’ of the presence of 200 Blastocystis, Linear Discriminant Effect Size (LEfSe) analysis was done [35]. LEfSe uses a 201 combination of statistical tests to identify taxa whose high/low abundance or 202 presence/absence allows for the best linear discrimination/explanation of the differences 203 observed (changes in taxa) between the 2 groups of samples ( Blastocystis +ve/-ve). Principle 204 Component Analysis (PCA) was also used to determine differences between the Blastocystis 205 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 10 +ve/-ve groups based on overall taxa presence and distribution. Samples were plotted based 206 on their Dissimilarity matrix values (Euclidian distances), and principal component analysis 207 was performed. Statistical analysis was then done using PERMANOVA [[34]] to determine if 208 the ‘centrons’ of each group (Blastocystis +ve/-ve) differed significantly in location. 209 210 Metabolite extraction 211 200 mg solid stool stored in methanol or 200 µl liquid stool was resuspended in 4 ml 212 methanol, then 200 mg glass beads were added and vortexed for 30 seconds. The samples 213 were then incubated at RT for 3 minutes, then vortexed for a further 30 seconds. The 214 supernatants were then divided in 4 x 1m aliquots then centrifuged at 10,000 x g at 4ºC for 20 215 mins then lyophilized. The lyophilized desiccates were then resolubilised in 375 µl 10% D 2O 216 1 mM non-deuterated DSS and recombined to make 1.5 ml solutions for NMR analysis. 217 The extracts were run on a 600 MHz Avance III NMR spectrometer (Bruker) with QCI-P 218 cryoprobe at a calibrated temperature of 298K to acquire 1D- 1H spectra. For each sample an 219 automated program was set up on the spectrometer using ICON NMR including measurement 220 of water offset, 90º pulse calibration, locking to D 2O, tuning and shimming using an 221 excitation sculpting experiment. A 1D- 1H-NOESY was run with a mixing time of 100 ms, 222 512 scans and 8 dummy scans, a spectral width of 15.98 ppm (9.59 Hz), 32768 data points, 223 an acquisition time of 2.27 s and a relaxation delay of 3 s. 224 The NMR spectra were phased, baseline corrected and had a 1 Hz exponential line 225 broadening window function applied using TOPSPIN 3.6.1 (Bruker) software, then exported 226 into Chenomx 8.4. The water resonance peak between 4.56 pmm and 4.97 ppm was deleted. 227 The spectral peaks were then fit into the Chenomx library of metabolites using the profiler 228 tool to match the peaks to their corresponding metabolites and concentrations. 229 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 11 Metabolites at significantly high abundances and of biological importance were divided into 230 four groups; Short chain fatty acids (SCFAs), amino acids, sugars and sugar alcohols and 231 other important metabolites and a time course was plotted to show the change in abundance 232 of each metabolite throughout the antibiotic course. 233 234

and Discussion 235 Composition of gut bacterial communities and Blastocystis colonisation 236 Stool samples from days 0, 2, 3, 4, 5, 6, 7, 8, 10, 15, 30 days and 3 months after the start of 237 the antibiotic course were collected and processed for both Blastocystis screening and 16S 238 gut microbiome analysis. Blastocystis ST1 was present in samples from D2, D3 and 239 consistently after D8 (Table 1). 240 For the microbiome analysis, the most abundant genera (defined here as taxa whose mean 241 abundance across all samples exceeded 1% of the total read count) across all stool samples 242 were plotted as a heatmap and compositional plot ( Figures 1 – 2). Seventeen genera met 243 these criteria, while multiple genera showed patterns of change throughout the antibiotic 244 course, including Phocaeicola_A, Escherichia and Enterococcus_B (Figures 1 – 2). Samples 245 were also grouped by their similarity values (Euclidean distance scores) and ordered in a 246 dendrogram. The bacterial communities of the samples collected 30 days and three months 247 after the start of the antibiotic course were more similar in taxonomic distribution to the 248 samples taken at both ends of the antibiotic course (D0-2 and D10-15) and had the least 249 similarity to samples taken on D3-D7 ( Figure 1). The samples collected on D0 and D2 were 250 strongly similar to each other. Bacteroides and Phocaeicola (both from the phylum 251 Bacteroidota) were the most abundant genera. They became more dominant throughout the 252 first week of the course, with their relative abundances decreasing in the months after the 253 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 12 completion of the course (Figures 1 – 2). Phocaeicola was the dominant genus and showed 254 an increase in relative abundance in the first four days of the course, a slight decrease towards 255 the end of the course and a further decrease in the months following the course (Figure 2 ). 256 Phocaeicola was no longer the dominant genus after the antibiotic course, with Escherichia 257 and Enterococcus being the most abundant in D30 and D3M, respectively ( Figure 2 ). 258 Blastocystis was not detected on D3, D4, D5, D6 and D7 of the course, but was again 259 detected on D8 and recovered at the end. The presence of Blastocystis coincided with the 260 decrease in abundance of Bacteroides and Phocaeicola (Figure 2). The reduced abundance 261 of Bacteroides in the presence of Blastocystis is a consistent finding across studies globally 262 [8], [14], [36], [37], [38], [39]. 263 264 Impact of antibiotic course on alpha diversity 265 To measure changes on alpha diversity, Shannon (factors in both evenness and richness), 266 Chao1 (richness index which factors in potentially relevant singleton and doubleton ‘rare 267 taxa’), Simpson (a measure of ‘dominance’/degree to which a few taxa make up most of the 268 reads) and Observed taxa (true richness) indices were used. These diversity indices were 269 analysed at the antibiotic-negative stage (pooled data of timepoints D0, D30, D3M) and at the 270 antibiotic-positive stage (pooled data of timepoints D2, D3, D4, D5, D6, D7, D8, D10 and 271 D15). All four metrics decreased during antibiotic administration. Antibiotics have been 272 associated with acute gut microbiota perturbations defined by a decrease in taxonomic 273 diversity [40], [41] (). However, these changes when analysed by ANOV A/Kruskal-Wallis 274 were shown to be non-significant (>0.05 P-value) (Supplementary Figure 1 B,D,F,H). 275 The diversity metric scores were also plotted individually over time. The Shannon and 276 Simpson diversity metrics decreased during the antibiotic course however the baseline 277 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 13 composition recovered in the months after the course ended, although a decreasing trend was 278 observed in D3M ( Figure 3 A, D ). This observation aligns with previous studies in both 279 adults and children whereby core microbiome taxa return to their pre-antibiotics abundance 280 [42]. Chao1 and Observed taxa showed sharp reductions at D4, D6 and D15. The reduction 281 in these two indices on D15 contrasts the increase in the Shannon diversity score, indicating 282 that the recovery in Shannon diversity score towards the end/post antibiotics course was 283 characterised by a matching reduction in domination of the microbiome by a handful of taxa. 284 This could be attributed to the elimination of rare taxa by the antibiotics. 285 286 Alpha diversity of Gut Microbial communities and Blastocystis colonisation 287 Blastocystis was detected early and later in the antibiotic course but not between D4 and D7 288 (Figure 3 ). All examined metrics of alpha diversity of Blastocystis-positive stool were 289 increased when compared to the Blastocystis- negative stool ( Figure 3; Supplementary 290 Figure 1). These changes in diversity were analysed for statistical significance and found to 291 all be non-significant (Observed taxa and Chao1 showed very low P-values of 0.058 and 292 0.064, respectively). 293 294 Biomarker analysis of stool samples for Blastocystis colonisation 295 Having shown that (non-significant) increases in diversity occur with the presence of 296 Blastocystis in stool samples, LEfSe was used to look for biomarkers of this change. LEfSE 297 [35] locates taxa whose presence/absence allows for the best identification of a member of 298 the Blastopos group instead of a member of the Blastoneg group [35]. Any LDA score >2 or 299 <-2 is considered to be significant (Figure 4) shows all taxa with a significant LDA score and 300 are coloured to show the indicative group. As shown in the plot, only two taxa were 301 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 14 significantly viable at discriminating against the two groups in favour of Blastocystis negative 302 samples (red, Blastoneg); these taxa were an unclassified member of the family 303 Anaerotignaceae and the genus Lactobacillus. There was a considerably larger amount of 304 taxa that were indicative of the Blastocystis positive group, with five taxa with a <-4 LDA 305 score ( Supplementary Figure 2 ); an unclassified member of the Negativicutes class of 306 bacteria, an unclassified Firmicute bacteria, two different unclassified members of the family 307 Rikenellaceae and an unclassified member of the family Enterobacteriaceae. The data 308 suggests that some bacteria are indicators of the presence of Blastocystis. Still, the presence 309 of so many indicative taxa may imply that a more diverse microbiome is the true indicative 310 factor for Blastocystis. 311 312 Impact of antibiotic course on metabolite composition of the gut and Blastocystis 313 colonisation 314 The metabolite extracts from the stool samples were subjected to 1D 1H NMR. Based on their 315 chemical properties, four groups of metabolites were detected, including short-chain fatty 316 acids (SCFAs) (Figure 5a), amino acids ( Figure 5b), sugars and sugar alcohols ( Figure 5c) 317 and others ( Figure 5d ). Previous metabolome investigations at a single timepoint on 318 Blastocystis positive and negative individuals showed a decreased abundance of certain 319 metabolites in the former [2]. Specifically, Alanine, Glycine, Histidine, Isoleucine, 320 Methionine, Threonine, Tryptophan and Valine all decreased, suggesting an anti-321 inflammatory role of Blastocystis . Significant increases in certain amino acids in the stool 322 have been found in IBD patients [43]. Herein, in a time course metabolome of a single 323 individual, all the amino acids, particularly Alanine and Valine, showed a decrease mid-324 course and recovery towards the end of the course. 325 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 15 Moreover, all amino acids, but glutamate showed a large decrease post-course (Figure 5b ). 326 Regarding SCFAs, the abundance of acetate increased throughout the first week of the 327 antibiotic course but decreased during the second week and then recovered post-course. 328 Butyrate increased after the first four days, decreased on D7, and recovered post-course. 329 Whether these alterations reflect changes in absorption or loss remains an open question. 330 Cellobiose was the most impacted sugar by the antibiotic course and showed a large increase 331 in abundance during the first week, then decline and recovery during the second week 332 (Figure 5c). Malonate steadily declined for the first five days and was undetectable by D6. It 333 then recovered on D7, D8 and D10 but became undetectable on D15 and post-course (Figure 334 5d). Succinate sharply increased from D1 to D2 and stayed high until D5, when it declined 335 again. O-acetylcholine declined for the first five days then recovered on D6 and D7, but 336 disappeared during the second week and didn’t recover post-course. Notably, acetylcholine in 337 the gut plays a role in intestinal homeostasis; hence, its disruption could potentially aggravate 338 inflammation [44]. 339 340

341 This case study provides valuable insights into the impact of antibiotic treatment on the 342 presence of Blastocystis and the overall gut microbiome and metabolome. The 14-day course 343 of Lansoprazole, Amoxicillin, and Clarithromycin significantly altered the gut microbial 344 composition, causing a notable decline in microbial diversity mid-course and leading to a 345 temporary absence of Blastocystis. Despite the antibiotic-induced perturbations, Blastocystis 346 demonstrated resilience, re-emerging post-treatment. 347 Additionally, the study highlighted significant fluctuations in metabolite levels, including 348 short-chain fatty acids and amino acids, which are critical for gut health. These changes 349 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 16 underscore the profound influence antibiotics have not only on microbial populations but also 350 on the metabolic environment of the gut. While antibiotics did not have a lasting effect on 351 Blastocystis colonisation, their temporary impact on microbial diversity and metabolite 352 composition points to gut ecology's intricate and dynamic nature. The findings emphasise the 353 need for careful consideration of antibiotic use, especially in conditions like IBS, where 354 maintaining a balanced gut microbiome is crucial. 355 Future research should expand on these findings by exploring the long-term effects of 356 antibiotics on gut microbiota and metabolites in larger cohorts. Understanding these 357 interactions will be essential for developing targeted therapies that mitigate adverse impacts 358 on the gut ecosystem while effectively treating gastrointestinal conditions. 359 360

361 We would like to thank the volunteer for participating in this study. Many thanks to the 362 Tsaousis Lab members (2020 – 2021) for their help and support in sample collection and 363 managing this project. J.M.N. was supported by a Kent Health studentship and W.J.S.E. by a 364 SoCoBio DTP studentship. 365 366 Contributions 367 Conceptualisation, A.D.T.; methodology, J.M.N., W.J.S.E. and G.S.T.; software, G.S.T.; 368 validation, G.S.T. and W.G. and W.J.S.E..; formal analysis, J.M.N. and W.J.S.E.; 369 investigation, J.M.N. and W.J.S.E.; resources, A.D.T.; data curation, J.M.N. and E.G.; 370 writing—original draft preparation, J.M.N.; writing—review and editing, A.D.T., E.G.; 371 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 17 supervision, A.D.T.; project administration, A.D.T.; funding acquisition, A.D.T. All authors 372 have read and agreed to the published version of the manuscript. 373 374 375 376 377

378 379 [1] A. Visconti et al., “Interplay between the human gut microbiome and host metabolism,” Nat 380 Commun, vol. 10, no. 1, Dec. 2019, doi: 10.1038/s41467-019-12476-z. 381 [2] E. L. Betts et al., “Metabolic fluctuations in the human stool obtained from blastocystis 382 carriers and non-carriers,” Metabolites, vol. 11, no. 12, Dec. 2021, doi: 383 10.3390/metabo11120883. 384 [3] K. S. W. Tan, “Blastocystis in humans and animals: New insights using modern 385 methodologies,” Vet Parasitol, vol. 126, no. 1-2 SPEC.ISS., pp. 121–144, 2004, doi: 386 10.1016/j.vetpar.2004.09.017. 387 [4] K. S. W. Tan, “New insights on classification, identification, and clinical relevance of 388 Blastocystis spp.,” Clin Microbiol Rev, vol. 21, no. 4, pp. 639–665, 2008, doi: 389 10.1128/CMR.00022-08. 390 [5] B. Skotarczak, “Genetic diversity and pathogenicity of blastocystis,” Annals of Agricultural 391 and Environmental Medicine, vol. 25, no. 3, pp. 411–416, 2018, doi: 10.26444/aaem/81315. 392 [6] S. K. Kamaruddin, A. Mat Yusof, and M. Mohammad, “Prevalence and subtype distribution of 393 Blastocystis sp. in cattle from Pahang, Malaysia,” Trop Biomed, vol. 37, no. 1, pp. 127–141, 394 2020. 395 [7] J. G. Maloney et al., “Identification and Molecular Characterization of Four New Blastocystis 396 Subtypes Designated ST35-ST38,” Microorganisms, vol. 11, no. 1, Jan. 2023, doi: 397 10.3390/microorganisms11010046. 398 [8] R. Y . Tito et al., “Population-level analysis of Blastocystis subtype prevalence and variation in 399 the human gut microbiota,” Gut, vol. 68, no. 7, pp. 1180–1189, Jul. 2019, doi: 10.1136/gutjnl-400 2018-316106. 401 [9] M. J. Kim, E. J. Won, S. H. Kim, J. H. Shin, and J. Y . Chai, “Molecular detection and 402 subtyping of human blastocystis and the clinical implications: Comparisons between diarrheal 403 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 18 and non-diarrheal groups in korean populations,” Korean Journal of Parasitology, vol. 58, no. 404 3, pp. 321–326, 2020, doi: 10.3347/kjp.2020.58.3.321. 405 [10] C. R. Stensvold et al., “Blastocystis: Unravelling potential risk factors and clinical significance 406 of a common but neglected parasite,” Epidemiol Infect, vol. 137, no. 11, pp. 1655–1663, 2009, 407 doi: 10.1017/S0950268809002672. 408 [11] A. M. Abdulsalam et al., “Prevalence, predictors and clinical significance of Blastocystis sp. in 409 Sebha, Libya,” Parasit Vectors, vol. 6, no. 86, pp. 1–8, 2013, doi: 10.1186/1756-3305-6-86. 410 [12] A. Alinaghizade, H. Mirjalali, M. Mohebali, C. R. Stensvold, and M. Rezaeian, “Inter- and 411 intra-subtype variation of Blastocystis subtypes isolated from diarrheic and non-diarrheic 412 patients in Iran,” Infection, Genetics and Evolution, vol. 50, pp. 77–82, 2017, doi: 413 10.1016/j.meegid.2017.02.016. 414 [13] E. Piperni et al., “Intestinal Blastocystis is linked to healthier diets and more favorable 415 cardiometabolic outcomes in 56,989 individuals from 32 countries,” Cell, 2024, doi: 416 10.1016/j.cell.2024.06.018. 417 [14] J. Forsell, J. Bengtsson-Palme, M. Angelin, A. Johansson, B. Evengård, and M. Granlund, 418 “The relation between Blastocystis and the intestinal microbiota in Swedish travellers,” BMC 419 Microbiol, vol. 17, no. 231, pp. 1–9, 2017, doi: 10.1186/s12866-017-1139-7. 420 [15] A. Kodio et al., “Blastocystis colonization is associated with increased diversity and altered 421 gut bacterial communities in healthy malian children,” Microorganisms, vol. 7, no. 649, pp. 1–422 12, 2019, doi: 10.3390/microorganisms7120649. 423 [16] K.L Glassner, B.P Abraham, E.M.M. Quigley “The microbiome and inflammatory bowel 424 disease,” J Allergy Clin Immunol, vol. 145, no. 1, pp. 16–27, Jan. 2020. 425 [17] C. Milani et al., “The First Microbial Colonizers of the Human Gut: Composition, Activities, 426 and Health Implications of the Infant Gut Microbiota,” Microbiology and Molecular Biology 427 Reviews, vol. 81, no. 4, 2017, doi: 10.1128/mmbr.00036-17. 428 [18] P. Rutgeerts et al., “Controlled trial of metronidazole treatment for prevention of crohn’s 429 recurrence after ileal resection,” Gastroenterology, vol. 108, no. 6, pp. 1617–1621, 1995, doi: 430 10.1016/0016-5085(95)90121-3. 431 [19] L. R. Glick et al., “Low-Dose Metronidazole is Associated with a Decreased Rate of 432 Endoscopic Recurrence of Crohn’s Disease after Ileal Resection: A Retrospective Cohort 433 Study,” J Crohns Colitis, vol. 13, no. 9, pp. 1158–1162, 2019, doi: 10.1093/ecco-jcc/jjz047. 434 [20] J. A. Yason, Y . R. Liang, C. W. Png, Y. Zhang, and K. S. W. Tan, “Interactions between a 435 pathogenic Blastocystis subtype and gut microbiota: In vitro and in vivo studies,” Microbiome, 436 vol. 7, no. 30, pp. 1–13, 2019, doi: 10.1186/s40168-019-0644-3. 437 [21] C. Audebert et al., “Colonization with the enteric protozoa Blastocystis is associated with 438 increased diversity of human gut bacterial microbiota,” Sci Rep, vol. 6, no. April, pp. 1–11, 439 2016, doi: 10.1038/srep25255. 440 [22] H. Li et al., “Bifidobacterium spp. and their metabolite lactate protect against acute 441 pancreatitis via inhibition of pancreatic and systemic inflammatory responses,” Gut Microbes, 442 vol. 14, no. 1, Dec. 2022, doi: 10.1080/19490976.2022.2127456. 443 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 19 [23] A. Sivan et al., “Commensal Bifidobacterium promotes antitumor immunity and facilitates 444 anti–PD-L1 efficacy,” Science (1979), vol. 350, no. 6264, pp. 1084–1089, Nov. 2015, doi: 445 10.1126/science.aac4255. 446 [24] K.T. Thia et al., “Ciprofloxacin or metronidazole for the treatment of perianal fistulas in 447 patients with Crohn’s disease: a randomized, double-blind, placebo-controlled pilot study,” 448 Inflamm Bowel Dis, vol. 15, no. 1, pp. 17–24, 2009, doi: 10.1002/ibd.20608. 449 [25] B. Shen et al., “A randomized clinical trial of ciprofloxacin and metronidazole to treat acute 450 pouchitis,” Inflamm Bowel Dis, vol. 7, no. 4, pp. 301–305, 2001, doi: 10.1097/00054725-451 200111000-00004. 452 [26] B. E. Lacy, L. Chang, S. S. C. Rao, Z. Heimanson, and G. S. Sayuk, “Rifaximin Treatment for 453 Individual and Multiple Symptoms of Irritable Bowel Syndrome With Diarrhea: An Analysis 454 Using New End Points,” Clin Ther, vol. 45, no. 3, pp. 198–209, Mar. 2023, doi: 455 10.1016/j.clinthera.2023.01.010. 456 [27] M. Willmann et al., “Distinct impact of antibiotics on the gut microbiome and resistome: A 457 longitudinal multicenter cohort study,” BMC Biol, vol. 17, no. 1, Sep. 2019, doi: 458 10.1186/s12915-019-0692-y. 459 [28] M. Reyman et al., “Effects of early-life antibiotics on the developing infant gut microbiome 460 and resistome: a randomized trial,” Nat Commun, vol. 13, no. 1, Dec. 2022, doi: 461 10.1038/s41467-022-28525-z. 462 [29] T. Nogueira, P. H. C. David, and J. Pothier, “Antibiotics as both friends and foes of the human 463 gut microbiome: The microbial community approach,” Feb. 01, 2019, Wiley-Liss Inc. doi: 464 10.1002/ddr.21466. 465 [30] J. G. Caporaso et al., “Global patterns of 16S rRNA diversity at a depth of millions of 466 sequences per sample,” Proc Natl Acad Sci U S A, vol. 108, no. SUPPL. 1, pp. 4516–4522, 467 Mar. 2011, doi: 10.1073/pnas.1000080107. 468 [31] E. Özkurt et al., “LotuS2: an ultrafast and highly accurate tool for amplicon sequencing 469 analysis,” Microbiome, vol. 10, no. 1, Dec. 2022, doi: 10.1186/s40168-022-01365-1. 470 [32] H. Li, “Minimap2: Pairwise alignment for nucleotide sequences,” Bioinformatics, vol. 34, no. 471 18, pp. 3094–3100, Sep. 2018, doi: 10.1093/bioinformatics/bty191. 472 [33] B. J. Callahan, P . J. McMurdie, M. J. Rosen, A. W. Han, A. J. A. Johnson, and S. P. Holmes, 473 “DADA2: High-resolution sample inference from Illumina amplicon data,” Nat Methods, vol. 474 13, no. 7, pp. 581–583, Jul. 2016, doi: 10.1038/nmeth.3869. 475 [34] T. Z. DeSantis et al., “Greengenes, a chimera-checked 16S rRNA gene database and 476 workbench compatible with ARB,” Appl Environ Microbiol, vol. 72, no. 7, pp. 5069–5072, Jul. 477 2006, doi: 10.1128/AEM.03006-05. 478 [35] N. Segata et al., “Metagenomic biomarker discovery and explanation,” Genome Biol, vol. 12, 479 no. 6, Jun. 2011, doi: 10.1186/gb-2011-12-6-r60. 480 [36] F. Beghini, E. Pasolli, T. D. Truong, L. Putignani, S. M. Cacciò, and N. Segata, “Large-scale 481 comparative metagenomics of Blastocystis, a common member of the human gut 482 microbiome,” ISME Journal, vol. 11, no. 12, pp. 2848–2863, Dec. 2017, doi: 483 10.1038/ismej.2017.139. 484 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint 20 [37] S. Gabrielli, F. Furzi, L. Fontanelli Sulekova, G. Taliani, and S. Mattiucci, “Occurrence of 485 Blastocystis-subtypes in patients from Italy revealed association of ST3 with a healthy gut 486 microbiota,” Parasite Epidemiol Control, vol. 9, May 2020, doi: 487 10.1016/j.parepi.2020.e00134. 488 [38] L. O’Brien Andersen et al., “Associations between common intestinal parasites and bacteria in 489 humans as revealed by qPCR,” European Journal of Clinical Microbiology & Infectious 490 Diseases, vol. 35, no. 9, pp. 1427–1431, Sep. 2016, doi: 10.1007/s10096-016-2680-2. 491 [39] A. Rajamanikam, M. N. M. Isa, C. Samudi, S. Devaraj, and S. K. Govind, “Gut bacteria 492 influence Blastocystis sp. phenotypes and may trigger pathogenicity,” PLoS Negl Trop Dis, 493 vol. 17, no. 3, 2023, doi: 10.1371/journal.pntd.0011170. 494 [40] L. McDonnell et al., “Association between antibiotics and gut microbiome dysbiosis in 495 children: systematic review and meta-analysis,” Gut Microbes, vol. 13, no. 1, pp. 1–18, 2021, 496 doi: 10.1080/19490976.2020.1870402. 497 [41] W. E. Anthony et al., “Acute and persistent effects of commonly used antibiotics on the gut 498 microbiome and resistome in healthy adults,” Cell Rep, vol. 39, no. 2, Apr. 2022, doi: 499 10.1016/j.celrep.2022.110649. 500 [42] A. Palleja et al., “Recovery of gut microbiota of healthy adults following antibiotic exposure,” 501 Nat Microbiol, vol. 3, no. 11, pp. 1255–1265, Oct. 2018, doi: 10.1038/s41564-018-0257-9. 502 [43] S. Bosch et al., “Fecal Amino Acid Analysis Can Discriminate De Novo Treatment/i1Naïve 503 Pediatric Inflammatory Bowel Disease From Controls,” J Pediatr Gastroenterol Nutr, vol. 66, 504 no. 5, pp. 773–778, May 2018, doi: 10.1097/MPG.0000000000001812. 505 [44] J. Uwada, H. Nakazawa, I. Muramatsu, T. Masuoka, and T. Yazawa, “Role of Muscarinic 506 Acetylcholine Receptors in Intestinal Epithelial Homeostasis: Insights for the Treatment of 507 Inflammatory Bowel Disease,” Apr. 01, 2023, Multidisciplinary Digital Publishing Institute 508 (MDPI). doi: 10.3390/ijms24076508. 509 510 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint Table 1 Blastocystis colonisation of stool samples was collected on different dates of the antibiotic course, 30 days, and three months after the completion of the course. + indicates the sample is Blastocystis + and – indicates the sample is Blastocystis –. Date of antibiotic course Blastocystis +/- Day 0 - Day 2 + Day 3 + Day 4 - Day 5 - Day 6 - Day 7 - Day 8 + Day 10 + Day 15 + 30 days after course + 3 months after course + All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint Figure 1. Heatmaps showing the relative abundance of taxa within the faecal samples, shown are all taxa which made up >1% of the total reads, all samples are indicated to be Blastocystis +ve (blue) or - ve (red) by a colour-coordinated legend. Dendrograms cluster both taxa (A-B) and the samples (B) using Euclidean distances and clustered using optimal leaf ordering. A). Heatmap showing the samples ordered by their collection timepoint during the antibiotics course. B). The heat map shows the samples ordered based on their similarity. Figure 2. (A-C). Compositional plots showing the bacterial composition of the gut taxa aggregated to varying taxonomic levels. A). Phylum level B). Genus level C). species level. Taxa included made up >1% of total read counts respectively. Taxa abundances are shown as % relative abundance. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint Figure 3. Statistical Diversity analysis of samples taken throughout the antibiotics course. A,B,C,D). Shannon, Chao1, Observed (richness), Simpson scores over time. Indicated if the sample was +ve (blue) or -ve (red) for Blastocystis.). Kruskal-Wallis H-test and Dunn’s test (Bonferroni p-adjust method) or ANOVA and Tukey-HSD test were used for statistical analysis (this was based on normality of the data, determined using the Shapiro test). Kruskal Wallis/ANOVA scores were all shown to be >0.05 indicating no-significance between the samples. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint Figure 4. Linear Discriminant Analysis (LDA) Effect Size (LEfSe) plot. LDA scores indicate the presence/increased abundance of each taxa to discriminate between two conditions, Blastocystis +ve (blue) and Blastocystis -ve (red). Taxa with LDA scores between -2 and 2 are considered insignificant ‘biomarkers’ and are not included in the plot. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint Figure 5 Time course of metabolite abundances of four different groups of metabolites throughout the antibiotic course as well as Blastocystis Colonisation. a. SCFAs b. Amino acids. c. Sugars and Sugar alcohols d. Other important metabolites All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted July 26, 2024. ; https://doi.org/10.1101/2024.07.25.24310942doi: medRxiv preprint