A Uniform Computational Approach Improved on Existing Pipelines to Reveal Microbiome Biomarkers of Nonresponse to Immune Checkpoint Inhibitors

In the paper

“Overall, the majority of these studies used non-parametric testing but did not apply multiple hypothesis testing correction in the initial analysis, with the exception of the Gopalakrishnan cohort (8).”

What the sources say

• Not checked yetcites [8] Gopalakrishnan, 2018, Gut microbiome modulates response to anti-PD-1 immunotherapy in melanoma patients, Science source

In the paper

“Herein, we identify novel bacterial signals associated with clinical response (R) or nonresponse (NR) and develop an Integrated Microbiome Prediction Index.”

In the paper

“Of the 2,132 unique canonically named Bacterial/Archaeal species (i.e. “genus_species”) in the MetaPhlAn2 database, 1,952 (91.56%) are present in the Resphera Insight database.”

In the paper

“Notably, in these datasets, taxa relative abundances that do not correlate across 16S rRNA amplicon and WGS data either represent signals occurring at <0.05% relative abundance or represent two taxa (Ruminococcus bromii and Parabacteroides distasonis) derived from only 2 of the included studies.”

In the paper

“Although F. prausnitzii and B. fragilis showed moderate heterogeneity across studies (I2=54% and 26%, respectively), these differences did not reach statistical significance.”

In the paper

“The NR Integrated Index displayed an average area under the curve (AUC) of 70.3% (± 1.66) compared to 47.8% (± 5.01, p = 0.0022) and 52.1% (±4.51, p = 0.0087) using the Literature and Melanoma Literature NR Indices, respectively (Fig. 4, middle column).”

In the paper

“In the Routy cohort, we also observed significantly higher alpha diversity associated with R among observed species and Chao1 estimators, but not Shannon or Reciprocal Simpson measures”

In the paper

“Overall, no differences were identified in beta diversity since, as expected (19), longitudinal fecal samples were more similar within patients as compared to between patients, despite some drift with time in the Chaput study (Fig. S5A and S5C).”

What the sources say

• Not checked yetcites [19] Costello, 2009, Bacterial community variation in human body habitats across space and time, Science source

In the paper

“Flagella assembly, bacterial motility proteins, and pantothenate/coenzyme A biosysnthesis were enriched in R, along with phagosome and mTOR pathway modifiers. Nucleotide sugar metabolism, nitrogen metabolism, cyanoamino acid metabolism, and glycosphingolipid biosynthesis were all enriched in NR.”

In the paper

“While the majority of species enriched in R corresponded to the previously reported results, for all species level assignments, only a subset of microbial signals retained statistical significance after FDR correction (Table S2).”

In the paper

“In contrast, the AUCs for the Routy NSCLC cohort improved by 7.59–10.91% for Integrated Indices, but the Routy RCC cohort results did not change.”

In the paper

“The Random Effects meta-analysis models were significant individually for R- or NR-associated indices and also as a composite, termed “Integrated Index[Sum]”.”

In the paper

“Conversely, the R signal was consistently, but modestly, enriched in the Integrated and Literature Indices, but not significant in the ICI R Melanoma Literature Index.”

In the paper

“When our Integrated Index, utilizing our newly identified R and NR taxa, was compared to either the Literature or Melanoma Literature Indices, the NR and composite signals outperformed the corresponding signals from either the Literature or Melanoma Literature Indices.”

In the paper

“These newly identified individual members (Fig. 2, above) enabled us to re-evaluate and develop novel indicator indices for R and NR signals, termed Integrated Microbiome Prediction Indices, or “Integrated Index” for short, along with a composite sum of both signals derived as [R index] – [NR index].”

In the paper

“Moreover, while, in general, lower alpha-diversity has been associated with disease states and higher alpha-diversity with healthy states, we found that alpha-diversity was not consistently predictive of response or nonresponse to ICIs, though we confirmed the association of higher alpha diversity in R as reported by Gopalakrishnan et al (8).”

What the sources say

• Not checked yetcites Gopalakrishnan et al (8)

In the paper

“F. prausnitzii and A. muciniphila had been previously identified and we confirmed the enrichment of these species in R though neither species met the initial significance cutoff of p ≤0.10 in the random effects model.”

In the paper

“F. prausnitzii and A. muciniphila had been previously identified and we confirmed the enrichment of these species in R though neither species met the initial significance cutoff of p ≤0.10 in the random effects model.”

In the paper

“In fact, when we were able to exclude SD patients (only the Frankel and Routy cohorts), our overall Integrated Indices improved, with a notable improvement in the Routy NSCLC cohort.”

In the paper

“Further, we were able to validate the Integrated Microbiome Prediction Index in independent cohorts of two tumor types, melanoma and HCC, albeit with small cohorts using publically available data, whereas 16S rRNA sequences derived from NSCLC patients performed less well overall.”

In the paper

“Our analyses enabled identification of novel bacterial signatures associated with clinical R and NR to ICI therapy that were not detected in smaller cohorts, and in particular, an unexpected strong microbiome association with ICI NR.”

In the paper

“Several studies, small in size and using variable analytical methods, suggest the gut microbiome may be a novel, modifiable biomarker for tumor response rates, but the specific bacteria or bacterial communities putatively impacting ICI responses have been inconsistent across the studied populations.”

What the sources say

• Not checked yetcites [5] Matson, 2018, The commensal microbiome is associated with anti-PD-1 efficacy in metastatic melanoma patients, Science source
• Not checked yetcites [6] Routy, 2018, Gut microbiome influences efficacy of PD-1–based immunotherapy against epithelial tumors, Science source
• Not checked yetcites [7] Frankel, 2017, Metagenomic shotgun sequencing and unbiased metabolomic profiling identify specific human gut microbiota and metabolites associated with immune checkpoint therapy efficacy in melanoma patients, Neoplasia source
• Not checked yetcites [8] Gopalakrishnan, 2018, Gut microbiome modulates response to anti-PD-1 immunotherapy in melanoma patients, Science source
• Not checked yetcites [9] Chaput, 2017, Baseline gut microbiota predicts clinical response and colitis in metastatic melanoma patients treated with ipilimumab, Ann Oncol source
• Not checked yetcites [10] Gharaibeh, 2019, Microbiota and cancer immunotherapy: in search of microbial signals, Gut source

In the paper

“Several studies have examined the stool composition of patients on checkpoint blockade with melanoma, non-small cell lung cancer (NSCLC), and renal cell cancer (RCC) (5–10). These studies have implicated increased alpha-diversity (8), along with specific members of each of the major bacterial phyla colonizing humans”

What the sources say

• Not checked yetcites [5] Matson, 2018, The commensal microbiome is associated with anti-PD-1 efficacy in metastatic melanoma patients, Science source
• Not checked yetcites [6] Routy, 2018, Gut microbiome influences efficacy of PD-1–based immunotherapy against epithelial tumors, Science source
• Not checked yetcites [7] Frankel, 2017, Metagenomic shotgun sequencing and unbiased metabolomic profiling identify specific human gut microbiota and metabolites associated with immune checkpoint therapy efficacy in melanoma patients, Neoplasia source
• Not checked yetcites [8] Gopalakrishnan, 2018, Gut microbiome modulates response to anti-PD-1 immunotherapy in melanoma patients, Science source
• Not checked yetcites [9] Chaput, 2017, Baseline gut microbiota predicts clinical response and colitis in metastatic melanoma patients treated with ipilimumab, Ann Oncol source
• Not checked yetcites [10] Gharaibeh, 2019, Microbiota and cancer immunotherapy: in search of microbial signals, Gut source

In the paper

“only 10–30% of treated patients respond and the ability to predict clinical benefit remains elusive”

• Also stated in: § Introduction · plain

  “only ~10–30% of all tumors respond to treatment (3)”

What the sources say

• Not checked yetcites [3] Sears, 2018, The intestinal microbiome influences checkpoint blockade, Nat Med source

In the paper

“Moreover, use of differing pipelines has limited the interpretability of the data presented to date for prediction of clinical anti-tumor responses, and whether microbiota signals predict non-response to ICIs has received less attention.”

In the paper

“In contrast, the gut microbiota of non-responding patients seemed to be less diverse, with the genera Bacteroides (phylum Bacteroidetes), Ruminococci and Roseburia (both Firmicutes) being prominent.”

What the sources say

• Not checked yetcites [5] Matson, 2018, The commensal microbiome is associated with anti-PD-1 efficacy in metastatic melanoma patients, Science source
• Not checked yetcites [6] Routy, 2018, Gut microbiome influences efficacy of PD-1–based immunotherapy against epithelial tumors, Science source
• Not checked yetcites [7] Frankel, 2017, Metagenomic shotgun sequencing and unbiased metabolomic profiling identify specific human gut microbiota and metabolites associated with immune checkpoint therapy efficacy in melanoma patients, Neoplasia source
• Not checked yetcites [8] Gopalakrishnan, 2018, Gut microbiome modulates response to anti-PD-1 immunotherapy in melanoma patients, Science source
• Not checked yetcites [9] Chaput, 2017, Baseline gut microbiota predicts clinical response and colitis in metastatic melanoma patients treated with ipilimumab, Ann Oncol source
• Not checked yetcites [10] Gharaibeh, 2019, Microbiota and cancer immunotherapy: in search of microbial signals, Gut source

In the paper

“While individually significant, a feature of these studies is that each has largely identified different sets of microbes and, overall, there is a lack of consensus regarding the microbial signals associated with clinical response and/or whether putative microbiota signals are tumor-specific or impacted by other clinical metadata.”

What the sources say

• Not checked yetcites [5] Matson, 2018, The commensal microbiome is associated with anti-PD-1 efficacy in metastatic melanoma patients, Science source
• Not checked yetcites [6] Routy, 2018, Gut microbiome influences efficacy of PD-1–based immunotherapy against epithelial tumors, Science source
• Not checked yetcites [7] Frankel, 2017, Metagenomic shotgun sequencing and unbiased metabolomic profiling identify specific human gut microbiota and metabolites associated with immune checkpoint therapy efficacy in melanoma patients, Neoplasia source
• Not checked yetcites [8] Gopalakrishnan, 2018, Gut microbiome modulates response to anti-PD-1 immunotherapy in melanoma patients, Science source
• Not checked yetcites [9] Chaput, 2017, Baseline gut microbiota predicts clinical response and colitis in metastatic melanoma patients treated with ipilimumab, Ann Oncol source
• Not checked yetcites [10] Gharaibeh, 2019, Microbiota and cancer immunotherapy: in search of microbial signals, Gut source

In the paper

“Further Resphera Insight has been validated as largely highly accurate for detection of Bacterioides and Fusobacterium spp. (13) and is also used by the FDA to enhance species identification contaminating the food chain (16,17).”

What the sources say

• Not checked yetcites [13] Drewes, 2017, High-resolution bacterial 16S rRNA gene profile meta-analysis and biofilm status reveal common colorectal cancer consortia, NPJ Biofilms Microbiomes source
• Not checked yetcites [16] Ottesen, 2016, Enrichment dynamics of Listeria monocytogenes and the associated microbiome from naturally contaminated ice cream linked to a listeriosis outbreak, BMC Microbiol source
• Not checked yetcites [17] Grim, 2017, High-resolution microbiome profiling for detection and tracking of Salmonella enterica, Front Microbiol source

In the paper

“Our analysis highlights the development of biomarkers for nonresponse, rather than response, in predicting ICI outcomes and suggests a new approach to identify patients who would benefit from microbiome-based interventions to improve response rates.”

In the paper

“These results suggest stability of the gut microbiome during treatment with checkpoint inhibitors as described in the original reports.”

In the paper

“Butanoate is the main biosynthetic pathway for butyrate, a short-chain fatty acid considered essential for immune and gut homeostasis.”

In the paper

“Butyrate and butyrate-producing bacteria such as F. prausnitzii and R. hominis, have been linked to favorable anti-tumor responses in multiple studies (20).”

What the sources say

• Not checked yetcites [20] Skelly, 2019, Mining the microbiota for microbial and metabolite-based immunotherapies, Nat Rev Immunol source

In the paper

“While these discrepancies could be due to variation(s) among patient cohorts due to regional or other clinical differences, alternatively, this result may argue for functional redundancy among the members of the Bacteroides genus, and in fact, we found three other members of the Bacteroides genus enriched in NR.”

In the paper

“Third, many of these studies utilized ICIs as monotherapy in a heavily pretreated population whereas current standard first-line therapy often includes a combination of ICIs or ICI with chemotherapy for multiple tumor types (35–40).”

What the sources say

• Not checked yetcites [35] George, 2019, Emerging role of combination immunotherapy in the first-line treatment of advanced renal cell carcinoma: a review, JAMA Oncol source
• Not checked yetcites [36] Wolchok, 2017, Overall survival with combined nivolumab and ipilimumab in advanced melanoma, N Engl J Med source
• Not checked yetcites [37] Gandhi, 2018, Pembrolizumab plus chemotherapy in metastatic non–small-cell lung cancer, N Engl J Med source
• Not checked yetcites [38] Horn, 2018, First-line atezolizumab plus chemotherapy in extensive-stage small-cell lung cancer, N Engl J Med source
• Not checked yetcites [39] Hellmann, 2019, Nivolumab plus ipilimumab in advanced non-small-cell lung cancer, N Engl J Med source
• Not checked yetcites [40] Motzer, 2018, Nivolumab plus Ipilimumab versus Sunitinib in advanced renal-cell carcinoma, N Engl J Med source

In the paper

“Second, our sample analysis was limited by only having data available from pre-treatment samples and a lack of information to stratify patients based on class of antibiotic exposure, exposure timeline, and other concurrent medication usage.”

In the paper

“Several retrospective studies have shown the detrimental effect of antibiotic use on overall survival in patients undergoing treatment with checkpoint inhibitors (28,29).”

What the sources say

• Not checked yetcites [28] Pinato, 2019, Antibiotic therapy and outcome from immune-checkpoint inhibitors, J Immunother Cancer source
• Not checked yetcites [29] Derosa, 2018, Negative association of antibiotics on clinical activity of immune checkpoint inhibitors in patients with advanced renal cell and non-small-cell lung cancer, Ann Oncol source

In the paper

“In addition. V. parvula has been previously linked to disease states and specifically enriched in the lung microbiome of patients with a primary lung malignancy (32–34).”

What the sources say

• Not checked yetcites [32] Lee, 2016, Characterization of microbiome in bronchoalveolar lavage fluid of patients with lung cancer comparing with benign mass like lesions, Lung Cancer source
• Not checked yetcites [33] Liu, 2018, Difference of lower airway microbiome in bilateral protected specimen brush between lung cancer patients with unilateral lobar masses and control subjects, Int J Cancer source
• Not checked yetcites [34] Tsay, 2018, Airway microbiota is associated with upregulation of the PI3K pathway in lung cancer, Am J Respir Crit Care Med source

In the paper

“Fourth, regional differences in baseline microbiota due to diet or other clinical features could affect the comparison of signals across cohorts.”

In the paper

“Given the consistent enrichment of microbial signals in NR across study cohorts, the NR Integrated Index may be a better predictor of the outcome of ICI therapy and serve to identify patients for whom alteration of the intestinal microbiome might offer clinical benefit.”

In the paper

“This result is not unexpected given the dominant microbiome remodeling induced by antibiotics and reinforces recent observations suggesting that recent antibiotic exposure hinders ICI therapeutic responses (28,29).”

What the sources say

• Not checked yetcites [28] Pinato, 2019, Antibiotic therapy and outcome from immune-checkpoint inhibitors, J Immunother Cancer source
• Not checked yetcites [29] Derosa, 2018, Negative association of antibiotics on clinical activity of immune checkpoint inhibitors in patients with advanced renal cell and non-small-cell lung cancer, Ann Oncol source

In the paper

“For example, the genus Bacteroides in human stools is often more abundant in patients in the USA compared to global levels while the genus Prevotella is more abundant outside the USA (41).”

What the sources say

• Not checked yetcites [41] Gorvitovskaia, 2016, Interpreting prevotella and bacteroides as biomarkers of diet and lifestyle, Microbiome source

In the paper

“Yet recent data identified cross-reactivity between a B. thetaiotaomicron β–galactosidase and myosin peptide associated with development of cardiomyopathy in genetically susceptible individuals (31).”

What the sources say

• Not checked yetcites [31] Gil-Cruz, 2019, Microbiota-derived peptide mimics drive lethal inflammatory cardiomyopathy, Science source

In the paper

“Of note, B. thetaiotaomicron is commonly considered a symbiote (30).”

What the sources say

• Not checked yetcites [30] Xu, 2003, A genomic view of the human-Bacteroides thetaiotaomicron symbiosis, Science source

In the paper

“In this combined dataset, the microbiome signals correlating with nonresponse were dominant, suggesting these microbes provide a potentially powerful biomarker for the selection of patients who might not ordinarily respond to checkpoint blockade but might benefit from microbiome-based interventions that reduce or eliminate specific species.”

In the paper

“In summary, use of a consistent analysis methodology improved on existing pipelines by standardizing critical preprocessing and downstream analysis tools, enabling comprehensive evaluations of taxonomic and functional signals across sequencing datasets that represent patients with melanoma, lung cancer, and renal cell carcinoma.”