Search

Supplementary Figures S1-S7 and Tables S1 and S2

Skin cancer risk varies substantially across the body, yet how this relates to the mutations found in normal skin is unknown. Here we mapped mutant clones in skin from high- and low-risk sites. The density of mutations varied by location. The prevalence of NOTCH1 and FAT1 mutations in forearm, trunk, and leg skin was similar to that in keratinocyte cancers. Most mutations were caused by ultraviolet light, but mutational signature analysis suggested differences in DNA-repair processes between sites. Eleven mutant genes were under positive selection, with TP53 preferentially selected in the head and FAT1 in the leg. Fine-scale mapping revealed 10% of clones had copy-number alterations. Analysis of hair follicles showed mutations in the upper follicle resembled adjacent skin, but the lower follicle was sparsely mutated. Normal skin is a dense patchwork of mutant clones arising from competitive selection that varies by location.Significance:Mapping mutant clones across the body reveals normal skin is a dense patchwork of mutant cells. The variation in cancer risk between sites substantially exceeds that in mutant clone density. More generally, mutant genes cannot be assigned as cancer drivers until their prevalence in normal tissue is known.See related commentary by De Dominici and DeGregori, p. 227.This article is highlighted in the In This Issue feature, p. 211

Skin cancer risk varies substantially across the body, yet how this relates to the mutations found in normal skin is unknown. Here we mapped mutant clones in skin from high- and low-risk sites. The density of mutations varied by location. The prevalence of NOTCH1 and FAT1 mutations in forearm, trunk, and leg skin was similar to that in keratinocyte cancers. Most mutations were caused by ultraviolet light, but mutational signature analysis suggested differences in DNA-repair processes between sites. Eleven mutant genes were under positive selection, with TP53 preferentially selected in the head and FAT1 in the leg. Fine-scale mapping revealed 10% of clones had copy-number alterations. Analysis of hair follicles showed mutations in the upper follicle resembled adjacent skin, but the lower follicle was sparsely mutated. Normal skin is a dense patchwork of mutant clones arising from competitive selection that varies by location. Significance: Mapping mutant clones across the body reveals normal skin is a dense patchwork of mutant cells. The variation in cancer risk between sites substantially exceeds that in mutant clone density. More generally, mutant genes cannot be assigned as cancer drivers until their prevalence in normal tissue is known. See related commentary by De Dominici and DeGregori, p. 227. This article is highlighted in the In This Issue feature, p. 211

BackgroundImmune checkpoint inhibitors are now standard of care treatment for many cancers. Treatment failure in metastatic melanoma is often due to tumor heterogeneity, which is not easily captured by conventional CT or tumor biopsy. The aim of this prospective study was to investigate early microstructural and functional changes within melanoma metastases following immune checkpoint blockade using multiparametric MRI.MethodsFifteen treatment-naïve metastatic melanoma patients (total 27 measurable target lesions) were imaged at baseline and following 3 and 12 weeks of treatment on immune checkpoint inhibitors using: T2-weighted imaging, diffusion kurtosis imaging, and dynamic contrast-enhanced MRI. Treatment timepoint changes in tumor cellularity, vascularity, and heterogeneity within individual metastases were evaluated and correlated to the clinical outcome in each patient based on Response Evaluation Criteria in Solid Tumors V.1.1 at 1 year.ResultsDifferential tumor growth kinetics in response to immune checkpoint blockade were measured in individual metastases within the same patient, demonstrating significant intertumoral heterogeneity in some patients. Early detection of tumor cell death or cell loss measured by a significant increase in the apparent diffusivity (Dapp) (papp), was consistently higher in the pseudoprogressive and true progressive lesions, compared with the responding lesions throughout the first 12 weeks of treatment. These preceded tumor regression and significant tumor vascularity changes (Ktrans, ve, and vp) detected after 12 weeks of immunotherapy (pConclusionsMultiparametric MRI demonstrated potential for early detection of successful response to immune checkpoint inhibitors in metastatic melanoma.

Background The gut microbiome of cancer patients appears to be associated with response to Immune Checkpoint Inhibitor (ICIs) treatment.1–4 However, the bacteria linked to response differ between published studies. Methods Longitudinal stool samples were collected from 69 patients with advanced melanoma receiving approved ICIs in the Cambridge (UK) MELRESIST study. Pretreatment samples were analysed by Microbiotica, using shotgun metagenomic sequencing. Microbiotica’s sequencing platform comprises the world’s leading Reference Genome Database and advanced Microbiome Bioinformatics to give the most comprehensive and precise mapping of the gut microbiome. This has enabled us to identify gut bacteria associated with ICI response missed using public reference genomes. Published microbiome studies in advanced melanoma,1–3renal cell carcinoma (RCC) and non-small cell lung cancer (NSCLC)4 were reanalysed with the same platform. Results Analysis of the MELRESIST samples showed an overall change in the microbiome composition between advanced melanoma patients and a panel of healthy donor samples, but not between patients who subsequently responded or did not respond to ICIs. However, we did identify a discrete microbiome signature which correlated with response. This signature predicted response with an accuracy of 93% in the MELRESIST cohort, but was less predictive in the published melanoma cohorts.1–3 Therefore, we developed a bioinformatic analytical model, incorporating an interactive random forest model and the MELRESIST dataset, to identify a microbiome signature which was consistent across all published melanoma studies. This model was validated three times by accurately predicting the outcome of an independent cohort. A final microbiome signature was defined using the validated model on MELRESIST and the three published melanoma cohorts. This was very accurate at predicting response in all four studies combined (91%), or individually (82–100%). This signature was also predictive of response in a NSCLC study and to a lesser extent in RCC. The core of this signature is nine bacteria significantly increased in abundance in responders. Conclusions Analysis of the MELRESIST study samples, precision microbiome profiling by the Microbiotica Platform and a validated bioinformatic analysis, have enabled us to identify a unique microbiome signature predictive of response to ICI therapy in four independent melanoma studies. This removes the challenge to the field of different bacteria apparently being associated with response in different studies, and could represent a new microbiome biomarker with clinical application. Nine core bacteria may be driving response and hold potential for co-therapy with ICIs. Ethics Approval The study was approved by Newcastle & North Tyneside 2 Research Ethics Committee, approval number 11/NE/0312. References Matson V, Fessler J, Bao R, et al. The commensal microbiome is associated with anti-PD-1 efficacy in metastatic melanoma patients. Science 2018;359(6371):104–108. Gopalakrishnan V, Spencer CN, Nezi L, et al. Gut microbiome modulates response to anti-PD-1 immunotherapy in melanoma patients. Science 2018;359(6371):97–103. Frankel AE, Coughlin LA, Kim J, et al. Metagenomic shotgun sequencing and unbiased metabolomic profiling identify specific human gut microbiota and metabolites associated with immune checkpoint therapy efficacy in melanoma patients. Neoplasia 2017;19(10):848–855. Routy B, Le Chatelier E, Derosa L, et al. Gut microbiome influences efficacy of PD-1-based immunotherapy against epithelial tumors. Science 2018;359(6371):91–97.

Independent groups have demonstrated that the pre-treatment gut microbiome of cancer patients impacts the subsequent response to Immune Checkpoint Inhibitor (ICIs) therapy [1-4]. However, each study identified different sets of bacteria linked to outcome, which has limited the development of drug response biomarkers and clinic-first design of novel microbiome-based therapeutics. The Cambridge (UK) MELRESIST study includes a cohort of advanced melanoma patients receiving approved ICIs. Pre-treatment stool samples from MELRESIST were analysed by Microbiotica using shotgun metagenomic sequencing. Microbiotica's platform comprises the leading Reference Genome Database to give the most comprehensive and precise mapping of the gut microbiome. A bioinformatic analysis identify a small discrete microbiome signature that was different between responders and non-responders. We extended this signature by reanalysing three published melanoma cohorts [1-3] using the Microbiotica platform. The resultant bacterial signature predicted whether or not a patient responded to anti-PD1-based therapy with an accuracy of 91% in all four studies combined and was also an effective biomarker for each cohort individually. We validated the signature using a NSCLC study [4] indicating that it has great potential as a clinical biomarker for a number of indications. The signature was strongly skewed towards species raised in abundance in responding patients, suggesting that the microbiome influences ICI treatment primarily through bacteria that enhance the efficacy of the drugs. At the core of the signature was nine species strongly associated a positive outcome, which we hypothesized to be a central driver of drug response. MB097 is a consortium comprised of all nine bacteria. In a syngeneic mouse model of cancer, MB097 was able inhibit tumor growth, but most strikingly was potently synergistic when dose with anti-PD1. To understand the mechanisms by which these bacteria drive an anti-tumor response, we have profiled the bacteria individually and as a consortium in multiple assays with primary human immune cells. The bacteria strongly activate dendritic cells with a number inducing high levels of IL-12 relative to IL-10. These bacteria-stimulated dendritic cells went on to trigger Cytotoxic T Lymphocytes (CTLs) to upregulate Granzyme B, Perforin and IFNg. Further, we have demonstrated that these primed CTLs are very effective at tumor cell killing in vitro. In summary, Microbiotica's precision microbiome profiling and the MELRESIST study has allowed us to identify a consortium of bacteria, MB097, strongly linked to response in multiple melanoma cohorts and a NSCLC study. The consortium drives immune-mediated tumor killing in vivo and in vitro. MB097 is being scaled up for manufacture as a novel co-therapy with ICIs. References 1 Matson V et al Science (2018) 359:104 2 Gopalakrishnan V Science (2018) 359:97 3 Frankel AE et al Neoplasia (2017) 19:848 4 Routy B et al Science (2018) 359:91 Citation Format: Matthew J. Robinson, Kevin Vervier, Simon Harris, Amy Popple, Dominika Klisko, Robyne Hudson, Ghaith Bakdash, Laure Castan, Clelia Villemin, David J. Adams, Doreen Milne, Catherine Booth, Christine Parkinson, Roy Rabbie, Sarah J. Welsh, Emily Barker, Katie Dalchau, Pippa Corrie, Trevor Lawley. MB097: A therapeutic consortium of bacteria clinically-defined by precision microbiome profiling of immune checkpoint inhibitor patients with potent anti-tumor efficacy in vitro and in vivo [abstract]. In: Abstracts: AACR Virtual Special Conference: Tumor Immunology and Immunotherapy; 2021 Oct 5-6. Philadelphia (PA): AACR; Cancer Immunol Res 2022;10(1 Suppl):Abstract nr P074.

Four independent international groups have demonstrated that the pre-treatment gut microbiome of cancer patients is associated with the subsequent response to treatment with Immune Checkpoint Inhibitors (ICI) [1-4]. However, each study identified different bacteria as being linked to outcome, which has limited the development of drug response biomarkers and novel microbiome-based therapeutics. Here we describe the identification of a microbial signature predictive of response to ICI across multiple melanoma studies, and a derived Live Bacterial Therapeutic with potent anti-tumour activity. MELRESIST is a single centre, prospective melanoma patient data and biosample collection research study. We collected longitudinal stool samples from 69 patients with advanced melanoma who received standard anti-PD-1+/- anti-CTLA-4 antibodies. Shotgun metagenomic sequencing analysis of the baseline stool microbiome was done using Microbiotica's platform, which comprises the world's leading Reference Genome Database to give the most comprehensive and precise mapping of gut microbiomes. Using 6 months progression-free survival as our cut-off for response, the analysis revealed a small but discrete microbiome signature that differentiated responders and non-responders with an accuracy of 93%. We extended this signature by reanalysing another 3 melanoma patient stool sample sequence datasets [1-3] using the Microbiotica platform, and a machine learning-based bioinformatic model. The resultant bacterial signature accurately predicted response when all 4 studies when combined (91%), as well as when the cohorts were analysed individually (82-100%). We validated the model using independent cohorts and the signature using NSCLC and Renal Cell Carcinoma (RCC) datasets [4]. The latter indicated the bacteria associated with response may differ slightly between indications. At the core of the signature was 9 bacteria that were all overrepresented in patients that responded to ICI treatment. Notably as a consortium, these 9 bacteria demonstrated tumor growth inhibition when dosed in a syngeneic mouse model. These strains also stimulate primary immune cells in vitro leading to tumor cell killing. In summary, we have identified a microbiome biomarker that is predictive of response to ICI treatment in multiple clinical studies from different countries. In addition, a unique set of bacteria derived from the signature has great therapeutic potential in combination with ICIs. References 1 Matson V et al Science (2018) 359:104 2 Gopalakrishnan V Science (2018) 359:97 3 Frankel AE et al Neoplasia (2017) 19:848 4 Routy B et al Science (2018) 359:91 Citation Format: Matthew J. Robinson, Kevin Vervier, Simon Harris, Roy Rabbie, Doreen Milne, Catherine Booth, Christine Parkinson, Sarah J. Welsh, David Bruce, Emily Barker, David Adams, Pippa Corrie, Trevor D. Lawley. Precision microbiome profiling identifies a novel biomarker predictive of Immune Checkpoint Inhibitor response in multiple cohorts and a potent therapeutic consortium of bacteria [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 1783.

e21546 Background: Four independent groups have demonstrated that the pre-treatment gut microbiome of cancer patients impacts the subsequent response to Immune Checkpoint Inhibitor (ICIs) therapy [1-4]. However, the patient’s outcome was linked to different bacteria in each study, which has limited the development of drug response biomarkers and clinic-first design of novel microbiome-based therapeutics. Methods: The Cambridge (UK) MELRESIST study includes a cohort of advanced melanoma patients receiving approved ICIs. Pretreatment stool samples from MELRESIST were analysed by Microbiotica using shotgun metagenomic sequencing. Microbiotica’s platform comprises the world’s leading Reference Genome Database to give the most comprehensive and precise mapping of the gut microbiome. Results: MELRESIST samples showed an overall difference in the microbiome composition between advanced melanoma patients and healthy donors, but not between patients who did or did not respond to ICIs. However, we did identify a discrete microbiome signature that differentiated responders and non-responders with an accuracy of 93%. We extended this signature by reanalysing three published melanoma cohorts [1-3] using the Microbiotica platform, and a propriety bioinformatic model. The resultant bacterial signature was very accurate at predicting response in all 4 published studies combined (91%), and each cohort individually (82-100%). We validated the model using independent validation cohorts and the signature using lung and renal cancer studies [4]. At the core of our microbiome signature was 9 bacteria most significantly associated with ICI efficacy. All 9 were overrepresented in patients who responded to immunotherapy suggesting high abundance of these bacteria is a central driver of ICI response. A consortium comprised of all 9 strains had very potent anti-tumor efficacy in a cancer syngeneic mouse model. The bacteria also demonstrate multiple interactions with primary human immune cells in vitro leading to dendritic cells activation, Cytotoxic T lymphocyte activation and tumor cell killing. These validate the potential of this consortium as a novel therapy for use in combination with ICIs. Conclusions: We have identified a unique microbiome signature predictive of ICI response in 4 independent melanoma cancer cohorts. This removes a major challenge to the field, and could represent a new highly accurate biomarker with clinical application. Nine core bacteria appear to be driving response, and demonstrate anti-tumor activity in vivo and in vitro. This consortium holds great potential as a co-therapy with ICIs. References:1 Matson V et al, Science (2018) 359:104; 2 Gopalakrishnan V et al, Science (2018) 359:97; 3 Frankel AE et al, Neoplasia (2017) 19:848; 4 Routy B et al, Science (2018) 359:91.