Metagenomic analysis of the microbiome of the upper reproductive tract: combating ovarian cancer through predictive, preventive, and personalized medicine

General and demographic analysis

Table 1 summarizes the basic information and demographic data for all subjects. As shown, there were 35 patients in the experimental group, with an average age of 52.7 ± 8.9, including 27 serous cell carcinoma, two mucinous carcinoma, three clear cell carcinoma, and three endometrioid carcinoma patients. For 9 cases, the fallopian tube samples were unavailable due to tumor invasion; for 1 case, the cervix was not taken and the cervical and uterine cavity results were not obtained for another subject. There were thirty benign cases (control group) with an average age of 55.1 ± 11.2 in the control group, twenty of which did not undergo ovariectomy. Among the malignant cases, eleven patients were not ligated, and thirteen were ligated. Platinum resistant was defined as relapse 6 months after 1st line platinum-based chemotherapy [25]. Eight cases were platinum-resistant, and nineteen cases were platinum-sensitive.

Table 1 Demographic and clinical characteristics of the subjectsGeneral microbial distribution in the paired samples of both benign and malignant patients

After quality control and de-host analysis, Kraken2 + PlusPF was used for species annotation followed by the seqkit stat used to perform a basic analysis. The average read sequence length of samples from patients in each state was in the range of 61 to 147.3 bp, and the median was in the range of 103 to 144 bp. The base quality of the read sequence of each state sample was acceptable, and the median Q30 ratio was above 96% (SuppFig. 1A). The reads of samples of different parts could be effectively annotated (SuppFig. 1B and C).

Specific composition of microorganism analysis

Bacteria had the highest proportion (median of 99%) in all samples compared to the other microorgansms. The ratio of fungi, viruses, eukaryotes, and archaea was relatively small, less than 1% in most samples. At the phylum level, the top 5 relative abundance was Proteobacteria, Firmicutes, Actinobacteria, Bacteroidetes, and Tenericutes (SuppFig. 1E). The Proteobacteria phylum had the highest abundance in all samples. The proportion of Firmicutes in cervical samples was higher than that of the other three parts. The average abundance of Actinobacteria bacteria from all malignant patients was higher than that in benign patients, especially in the ovarian cases. At the species level, the OUTs that appear in at least 10% of the samples were conserved. The top 5 were Pseudomonas tolaasii, Klebsiella pneumoniae, Salmonella sp., Acinetobacter johnsonii, and Escherichia coli (Fig. 1A). Furthermore, species-level alpha-diversity suggested no significant differences in microbial diversity of the cervix, uterine cavity, and fallopian tube except for ovarian tissues (Fig. 1A, SuppFig. 1D).

Fig. 1figure 1

Signature species for each body site. A Species-level relative abundance community barplot analysis of metagenomic shotgun sequencing in the cervical (CCT), uterine cavity (EMT), fallopian tube (FTT), and ovarian (OCT) from the benign (B) or malignant (M) lesion patients. B Bray-Curtis dissimilarity and Spearman’s correlation coefficient analysis of two adjacent parts at species level. C Species-level beta-diversity analysis (Aitchison distance, PCA). There is taxa diversity in different sites in two types of diseases by PERMAVONA analysis (condition: p = 0.001, sample site: p = 0.001). D Veen plot of microbial species among different sample sites

Migration analysis

To analyze the migration of microorganisms from the lower reproductive tract to the upper reproductive tract, we conducted a similarity analysis of the adjacent positions of samples from the four parts. At the phylum level, the relative abundance of Firmicutes bacteria in both the benign and malignant patients generally decreased from CCT to EMT (SuppFig. 2A and B). The relative abundance of Ascomycota bacteria in malignant patients from EMT to FTT was both increased, and the difference was statistically significant (p < 0.01) (SuppFig. 2C and D). The relative abundance of Uroviricota gate in benign patients from FTT to OCT generally raised, but the difference was not statistically significant (SuppFig. 2E and F). At the level of CCT-EMT, there were apparent differences in the similarity between the two groups (Fig. 1B). There was no significant difference in the parallel of other parts between the two groups; the relative abundance of Brevundimonas sp. DS20 bacteria in group B patients generally decreased from CCT to EMT. The relative abundance of Brevundimonas mediterranea, Brevundimonas sp. scallop, Brevundimonas sp. Bb-A, Brevundimonas sp. DS20, Brevundimonas sp. GW460-12–10-14-LB2, Brevundimonas sp. SGAir0440, and Cutibacterium_acness was reduced from CCT to EMT in the malignant group (SuppFig. 3A and B). Cutibacterium acnes and Komagataella phaffii generally increased in relative abundance from EMT to FTT in patients in malignant group (SuppFig. 3C and D). The relative abundance of Cutibacterium acnes in benign patients from FTT to OCT had an upward trend (p > 0.05). The relative abundance of Brucella intermedia in the experiment group from FTT to OCT decreased, which had a statistical significant.

For the sample diversity, the beta diversity analysis indicated that there were significant differences in the microbial composition of patients in different groups (p <  = 0.001) (Fig. 1C). For presentation purposes, we performed a Venn diagram analysis. The results suggested that the out shared by four parts accounted for the majority, indicating that the lower reproductive tract was the primary source of microorganisms for the upper reproductive tract.

The microbial profile within a subject

To exclude the effects of different microbial populations between/among subjects, we further analyzed the microbial portfolio of four different sites of single subjects. Briefly, we analyzed the relative abundance of microbes at the species level of four sites in 10 patients from control group and 24 subjects from experimental group. The samples of these 34 cases were divided into two clusters according to the salmonella sp. abundance (> 1%), depending on the environmental control and sample classification. We then analyzed the distribution of Salmonella sp. and Pseudomonas Tolaasii in benign and malignant patients with the relative abundance of salmonella sp. more than 1% at any part (CCT, EMT, FTT, or OCT) (Fig. 2A). In the control group, the relative abundance of Pseudomonas tolaasii gradually decreased as migrated upward from the lower reproductive tract to the ovarian tissue. There was a similar trend in the malignant samples. Meanwhile, Salmonella sp. gradually enriched among the 4 sites within the experimental group. The relative abundance in the uterine cavity was significantly higher than that of the lower reproductive tract (cervix) (Fig. 2B). Interestingly, while no difference of Pseudomonas tolaasii abundance was detected in all samples of both groups, however, Salmonella sp. showed an enriched status in both benign and malignant uterine cavities with higher level in the ovary of control group compared to the subjects in the experimental cohort (Fig. 2C). These findings supported the notion that microorganisms in ovarian tissues originate from the reproductive tract. Additionally, these data imply that the migration of the organisms in the upper reproductive tract of the malignant patients was different from those of benign controls with noticeable enriched microorganisms.

Fig. 2figure 2

Signature species within and between individuals. A Species-level relative abundance community barplot analysis at 4 sites per patient. B The relative abundance of Salmonella sp. and P. tolaasii in benign and malignant patients presented in A with the relative abundance of Salmonella sp. more than 1% at any site (CCT, EMT, FTT, or OCT) and C the relative abundance of Salmonella sp. and P. tolaasii in total samples of benign and malignant patients

Malignant ovarian tissue had distinct microbial signatures

To clarify the distribution of microorganisms in malignant ovarian tissues, the ovarian samples from both groups were compared (34 subjects from experimental group and 10 subjects from control group). The two groups showed significant differences in the distribution of β-diversity (p = 0.001, and p = 0.005) at the genus level as well as the species level (Fig. 3A and D). There were 146 genus-level OTUs in all the ovarian samples. The differential bacteria enriched on the surface of malignant ovarian tissues were Salmonella, Asticcacaulis, Arthrobacter, Lactobacillus, Pseudarthrobacter, and Pseudarthrobacter compared to the top 5 different bacteria enriched on the surface of benign ovarian tissue, Brevundimonas, Ralstonia, Pandoraea, Streptococcus, and Corynebacterium (Fig. 3B and C). A total of 329 OTUs at the species level, Salmonella sp., Asticcacaulis excentricus, Acinetobacter sp. NEB 394, Acinetobacter lwoffii, and Arthrobacter sp. FB24, were enriched in malignant tissues. As for control group, Brevundimonas sp. Bb-A, Brevundimonas sp. DS20, Ralstonia pickettii, Pandoraea pnomenusa, Staphylococcus hominis, Komagataella phaffii, Finegoldia magna, Cutibacterium acnes, Prevotella intermedia, and Agrobacterium tumefaciens were enriched (Fig. 3E and F). Notably, Salmonella sp. was the most enriched species in malignant tissues, and its relative abundance was also the highest in the bacteria. While this study was not geared towards the detailed genotyping of every bacterium, however, Salmonella typhimurium is reported as one of the major species in ovarian cancer [4].

Fig. 3figure 3

The distribution of microorganisms on the surface of benign and malignant ovarian tissues. A Genus-level beta-diversity analysis between 10 benign ovarian tissues and 34 malignant ones by PERMAVONA analysis (p = 0.001). B Differential abundant genus between two groups by ANCOM-BC test. C. Relative abundances of the differential abundant genus between two groups. D, E, and F show species’ signature

The influence of tubal ligation on the distribution of microorganisms on the surface of ovarian cancer tissue

A total of 312 OTUs were found between 13 fallopian tube ligated patients and 11 non-ligation patients. Although there was no statistical difference (p = 0.146), however, the β-diversity showed that the two groups share unidentical compositions of microorganisms (Fig. 4A). This might be due to the insufficient sample numbers. Among the different species, the top 5 most enriched bacteria were Arthrobacter sp. J3.40, Arthrobacter sp. UKPF54-2, Arthrobacter sp. KBS0702, Arthrobacter sp. FB24, and Acinetobacter sp. NEB 394 on the surface of the ovarian tissue in the ligated subjects. As for the ovarian tissue of the unligated subjects, the most enriched species were Brevundimonas sp. DS20, Ralstonia mannitolilytica, Brevundimonas mediterranea, Gardnerella vaginalis, and Achromobacter xylosoxidans (Fig. 4B and C).

Fig. 4figure 4

Signature species on the surface of ovarian tissues between ligation and unligation of patients with ovarian cancer. A Species-level beta-diversity analysis between 13 ligation and 11 unligation of patients by PERMAVONA analysis (p = 0.147). B Differential abundant species between two groups by ANCOM-BC test. C Relative abundances of the differential abundant species between two groups

Platinum resistance and the microorganism distribution of the ovarian cancer tissues

We compared the ovarian tissue surface microbial species between 19 postoperative platinum-sensitive and eight platinum-resistant patients. On the species level, the two groups shared 226 species (Fig. 5A). The β-diversity analysis indicated no difference in species between the two groups (p = 0.44). The top 5 enrichment in the platinum-resistant group is Pseudomonas_aeruginosa, Ralstonia mannitolilytica, Achromobacter xylosoxidans, Brevundimonas sp. DS20, and Brevundimonas sp. Bb-A (Fig. 5B and C).

Fig. 5figure 5

Signature species on the surface of ovarian tissues between nineteen postoperative platinum-sensitive and eight platinum-resistant patients. A Species-level beta-diversity analysis indicated no difference in species between the two groups. B Differential abundant species between two groups by ANCOMBC test. C Relative abundances of the differential abundant species between two groups

留言 (0)

沒有登入
gif