Intratumor microbes, which are present in both tumor and immune cells, have been shown to have multiple effects on human cancer (Cullin et al., 2021). Recent studies have demonstrated that it is possible to characterize the landscape of tumor-specific intratumor microbes in a large number of cancer patients (Nejman et al., 2020), including in samples from multiple cancer types from The Cancer Genome Atlas (TCGA) (Dohlman et al., 2021; Poore et al., 2020). However, the effects of patients’ demographic and clinical factors (e.g., age, gender, body mass index [BMI], and self-reported race) on the intratumor microbiome remain to be elucidated.
Addressing this important open question, we performed a systematic characterization of the tumor microbiome across different cancer types to better understand the effects of these factors on the intratumor microbiome composition. We obtained the microbial abundances in different cancer types from TCGA, including 1,553 genus belonging to 395 families, 175 orders, 76 classes, and 37 phyla (Poore et al., 2020). We capitalized on data by Poore et al., obtaining normalized intratumor microbial abundance data that have been comprehensively filtered for contaminant species (measured in genus units) for 32 tumor types from the online data repository (ftp://ftp.microbio.me/pub/cancer_microbiome_analysis) (Poore et al., 2020). Specifically, we used the intratumor microbial abundance data file “Kraken-TCGA-Voom-SNM-Likely-Contaminants-Removed-Data.csv” with likely contaminants removed. We investigated the associations between the abundance of intratumor microbes, measured in normalized read counts associated with a genus, and gender, age at initial diagnosis, BMI, and race across 27 cancer types having a sample size ≥20. This analysis was performed through a propensity score algorithm (see online methods) (Ye et al., 2019), adjusting for potential confounding factors such as histological type and tumor stage. Our key result is the finding of strong associations between race (European, African, Asian) and different genera abundances in most cancer types (Figure S1A). For example, in breast cancer (BRCA), we identified 551, 701, and 305 significantly differentially abundant genera in European versus Asian, European versus African, and Asian versus African, respectively (while correcting for multiple hypothesis testing with FDR < 0.05). Some microbial genera showed consistent significant differences in race comparisons in numerous cancer types. For example, Brucella, which have been reported to affect immune therapy (Guo et al., 2022), showed significant differences in race comparisons in seven cancer types (data available in our data portal). Some other microbes (n = 172) showed cancer-type-specific associations. For example, Epilithonimonas showed specific significant differences in European versus Asian in uterine cancer (UCEC) only. Reassuringly, the number of significantly differential genera is still high in European versus Asian, European versus African, and Asian versus African, respectively (Figure S1B). Our findings thus highlight the important role that race may play in tumor microbiome composition and the need to consider its effects in future microbiome-based therapeutic strategies.
These findings are perhaps even more notable, as we found that age, gender, and BMI actually have very few associations with genera abundances of intratumor microbes (Figure S1A). For example, we identified only 45 genera in THCA that showed significant differences in their abundance between old (age ≥ 65) versus young (age <65) patients, while most cancer types showed very few such differences, if any (Figure S1A). Similarly, we identified very few associations between gender and genus abundance, as no cancer types showed more than 10 genera whose abundance was significantly different between male and female patients (Figure S1A). Very few differences were also noted between obese patients (BMI ≥30) and nonobese ones (25 > BMI ≥18.5), although it should be noted that BMI information was available in only 10 cancer types. Furthermore, we observed few associations between the abundance of microbiomes and age, gender, and BMI in several other independent datasets (data available in data portal due to space limitations) (Jin et al., 2022; Nejman et al., 2020), further reinforcing these findings.
Our analysis is complemented by a user-friendly data portal, Intratumor Microbiome associated with Clinical Characteristics (IMCC; https://hanlab.tamhsc.edu/IMCC), to allow browsing and searching for the specific associations of the microbiome with gender, age, BMI, and race at the individual genus level in specific cancers of interest (Figure S1C). To facilitate possible future clinical applications, we also characterized associations between microbiomes and overall survival/drug response in cancers.
In summary, we estimated the associations between demographic and clinical factors and microbial genera abundances while controlling for several possible confounding variables by propensity score matching. We observed few associations between the abundance of intratumor genera with age, gender, and BMI but, in marked difference, a large number of strong associations with race. This conclusion remains solid and robust if we performed analysis with other more stringent level data, including “All putative contaminants removed data,” “Plate Center Filtering data,” and “Most stringent filtering data,” or filtering those genera with relatively low abundance (normalized abundance (log2[cpm]) ≤ 1, or count resolved to genus level <10). Of note, a recent study demonstrated that the mammary microbiota may be modulated by dietary patterns (Soto-Pantoja et al., 2021), which could vary due to different races. Race, as a factor related to dietary styles, deserves to be considered in intratumor microbiome studies. Indeed, recent studies highlighted the limited race diversity within cohorts in biomedical studies (Deschasaux et al., 2018), including human microbiome studies (Abdill et al., 2022). Significant efforts have been conducted to eliminate potential contaminant species and the potential confounding effects of low read counts (Nejman et al., 2020; Poore et al., 2020; Sepich-Poore et al., 2021), and further investigation is necessary with high-quality intratumor microbiome data and more complete clinical information (e.g., diet, socioeconomic status, etc.). In particular, our study further puts forward the need to broaden the study of race diversity in future investigations of the tumor microbiome.
Supplementary Material
ACKNOWLEDGMENTS
Research reported in this publication was supported by the National Institutes of Health (NIH) (R03AG0 70417, R01HG011633, and R01CA262623 to L.H.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This work was also supported by Cancer Prevention Research Institute of Texas (CPRIT) (RR150085, RP190570) to the CPRIT Scholar in Cancer Research (L.H.). E.R. and E.M.G.’s research was supported in part by the Intramural Research Program of the NIH, National Cancer Institute, and the Center for Cancer Research. Detailed methods are available at https://hanlab.tamhsc.edu/IMCC#!/document, and all other associated data are available in the data portal.
Footnotes
SUPPLEMENTAL INFORMATION
Supplemental information can be found online at https://doi.org/10.1016/j.ccell.2022.08.007.
DECLARATION OF INTERESTS
E.R. is a co-founder of Metabomed Ltd and MedAware and a (divested) co-founder and non-paid scientific consultant for Pangea Biomed.
REFERENCES
- Abdill RJ, Adamowicz EM, and Blekhman R (2022). Public human microbiome data are dominated by highly developed countries. PLoS Biol. 20, e3001536. 10.1371/journal.pbio.3001536. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cullin N, Azevedo Antunes C, Straussman R, Stein-Thoeringer CK, and Elinav E (2021). Microbiome and cancer. Cancer Cell 39, 1317–1341. 10.1016/j.ccell.2021.08.006. [DOI] [PubMed] [Google Scholar]
- Deschasaux M, Bouter KE, Prodan A, Levin E, Groen AK, Herrema H, Tremaroli V, Bakker GJ, Attaye I, Pinto-Sietsma S-J, et al. (2018). Depicting the composition of gut microbiota in a population with varied ethnic origins but shared geography. Nat Med 24, 1526–1531. 10.1038/s41591-018-0160-1. [DOI] [PubMed] [Google Scholar]
- Dohlman AB, Arguijo Mendoza D, Ding S, Gao M, Dressman H, Iliev ID, Lipkin SM, and Shen X (2021). The cancer microbiome atlas: a pan-cancer comparative analysis to distinguish tissue-resident microbiota from contaminants. Cell Host Microbe 29, 281–298.e5. 10.1016/j.chom.2020.12.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Guo F, Das JK, Kobayashi KS, Qin Q-M, A Ficht T, Alaniz RC, Song J, and Figueiredo PD (2022). Live attenuated bacterium limits cancer resistance to CAR-T therapy by remodeling the tumor microenvironment. J Immunother Cancer 10, e003760. 10.1136/jitc-2021-003760. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jin H, Hu G, Sun C, Duan Y, Zhang Z, Liu Z, Zhao X-M, and Chen W-H (2022). mBodyMap: a curated database for microbes across human body and their associations with health and diseases. Nucleic Acids Res. 50, D808–D816. 10.1093/nar/gkab973. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Nejman D, Livyatan I, Fuks G, Gavert N, Zwang Y, Geller LT, Rotter-Maskowitz A, Weiser R, Mallel G, Gigi E, et al. (2020). The human tumor microbiome is composed of tumor type-specific intracellular bacteria. Science 368, 973–980. 10.1126/science.aay9189. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Poore GD, Kopylova E, Zhu Q, Carpenter C, Fraraccio S, Wandro S, Kosciolek T, Janssen S, Metcalf J, Song SJ, et al. (2020). Microbiome analyses of blood and tissues suggest cancer diagnostic approach. Nature 579, 567–574. 10.1038/s41586-020-2095-1. [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
- Sepich-Poore GD, Zitvogel L, Straussman R, Hasty J, Wargo JA, and Knight R (2021). The microbiome and human cancer. Science 371, eabc4552. 10.1126/science.abc4552. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Soto-Pantoja DR, Gaber M, Arnone AA, Bronson SM, Cruz-Diaz N, Wilson AS, Clear KY, Ramirez MU, Kucera GL, Levine EA, et al. (2021). Diet alters entero-mammary signaling to regulate the breast microbiome and tumorigenesis. Cancer Res. 81, 3890–3904. 10.1158/0008-5472.can-20-2983. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ye Y, Hu Q, Chen H, Liang K, Yuan Y, Xiang Y, Ruan H, Zhang Z, Song A, Zhang H, et al. (2019). Characterization of hypoxia-associated molecular features to aid hypoxia-targeted therapy. Nat Metab 1, 431–444. 10.1038/s42255-019-0045-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.