Physical fitness influences the composition of human T cell populations
Michael Gustafson, PhD1, Michael Gustafson, PhD1, Courtney Wheatley- Guy, PhD1, Svetlana Bornschlegl2, Allan Dietz, PhD2, Dennis Gastineau, MD1, Emmanuel Katsanis, MD3, Richard Simpson, PhD3, Bruce Johnson, PhD1
1
Mayo Clinic Arizona, Phoenix, AZ, United States;2Mayo Clinic, Rochester, MN, United States;3University of Arizona, Tucson, AZ, United States Correspondence:Michael Gustafson ([email protected])
Journal for ImmunoTherapy of Cancer2019,7(Suppl 1):P744
Background
Despite the remarkable success of cancer immunotherapies, a signifi- cant proportion of patients fail to respond. In order to overcome this, a number of ongoing approaches are being tested to address this need including combinatorial treatment strategies, biomarker identi- fication to tailor optimal responses, and neoantigen discovery. One
Fig. 1 (abstract P742).TIM-3 feedback network
potential avenue that is emerging is the application of exercise to promote overall health. While the mechanisms of exercise-mediated improvements in health are certainly multi-factorial, a substantial amount of data suggests that physical activity can improve outcomes in cancer patients, in part by, improving immunological function. As such, we propose that physical activity may enhance the efficacy of immunotherapy by restoring the quantity and function of leukocyte subsets in cancer patients.
Methods
T cells and other immune cell populations were analyzed by flow cy- tometry from two different cohorts of healthy men[1, 2]. Physical fit- ness parameters including maximum rate of oxygen consumption (VO2max), physical activity, lean body mass, and BMI were assessed or measured and tested for correlations with T cell populations.
Results
Here, we provide evidence demonstrating that the compositions of human T cells circulating in peripheral blood are altered by physical fitness. In one cohort, late differentiated KLRG1+/CD57+/CD28- T cells were inversely associated with a cardiorespiratory fitness param- eter, VO2max. In another we show several T cell populations that are associated with distinct fitness parameters. For example, in baseline peripheral blood samples, PD-1+ T cells were increased in sedentary subjects compared to active ones, numbers of CD8+ T cells and memory CD4+ T cells correlated with lean body mass, and a popula- tion containing stem cell memory CD8+ T cells were correlated with VO2max.
Conclusions
These results suggest that distinct T cell populations may be uniquely sensitive to both negative (age, obesity, and lack of physical activity) and positive (exercise) physiological inputs. These findings provide hypothesis-generating insights into how exercise might be utilized to overcome the immunosuppressive environment in cancer patients potentially improving their responses to immunotherapies.
References
1. Spielmann G, McFarlin BK, O'Connor DP, Smith PJ, Pircher H, Simpson RJ. Aerobic fitness is associated with lower proportions of senescent blood T-cells in man. Brain, behavior, and immunity. 2011;25(8):1521-9. 2. Gustafson MP, DiCostanzo AC, Wheatley CM, Kim CH, Bornschlegl S,
Gastineau DA, et al. A systems biology approach to investigating the influence of exercise and fitness on the composition of leukocytes in peripheral blood. Journal for immunotherapy of cancer. 2017;5:30. Ethics Approval
These studies were approved by the Institutional Review Board by their respective institutions, Mayo Clinic and University of Houston. Consent
Written informed consent was obtained from the patient for publication of this abstract and any accompanying images. A copy of the written consent is available for review by the Editor of this journal.
P745
Tumor cell metabolism as a barrier to immunotherapy in melanoma
Ashley Menk, MS1, Greg Delgoffe, PhD1, Yana Najjar, MD2, Ashley Menk, MS1, John Kirkwood, MD2
1University of Pittsburgh, Pittsburgh, PA, United States;2UPMC Hillman Cancer Center, Pittsburgh, PA, United States
Correspondence:Greg Delgoffe ([email protected])
Journal for ImmunoTherapy of Cancer2019,7(Suppl 1):P745
Background
Immunotherapy has been paradigm-shifting for melanoma, but dur- able responses only occur in a subset of patients. While we have pre-
viously shown that tumor-infiltrating T cells have repressed
metabolic machinery, the environment itself is nutrient poor due to the deregulated metabolism of tumor cells. Recent studies have sug- gested that T cells compete with tumor cells for glucose and oxygen which may be limiting anti-tumor immunity. We hypothesize resist- ance to immunotherapy may be due to deficiencies in the metabolic makeup of the tumor microenvironment, and that we can infer that
environment’s metabolism in patients by metabolically profiling
tumor cells.
Methods
Melanoma patient samples were profiled by Seahorse analysis in par- allel to flow cytometric analysis of tumor infiltrating lymphocytes (TIL). Murine melanoma cells were generated including RNAi con- structs to specific metabolic pathways. Responses to PD-1 blockade and adoptive T cell immunotherapy were monitored and lymph node or TIL T cells were tested for effector function, metabolism, and localization by flow cytometry and immunofluorescence.
Results
Analysis of melanoma patient biopsies showed striking heterogeneity in the metabolism of tumor cells. Patients with more metabolically active tumors have more dysfunctional TIL, while tumors that were metabolically quiescent contained T cells with superior effector func- tion. Profiling of patient tumor cells prior to PD-1 blockade therapy revealed that patients with metabolically oxidative tumors had poorer responses while patients with quiescent tumors experienced long-term response. To confirm the role of oxidative metabolism and further investigate the role of tumor cell metabolism, we generated murine melanoma lines in which glucose or oxidative metabolism was inhibited. Tumor cells in which oxidative metabolism showed an increased response to PD-1 blockade immunotherapy while tumor cells with inhibited glycolytic metabolism responded better to adop- tive T cell therapy.
Conclusions
Our data suggest that the metabolic makeup of the tumor micro- environment, driven through deregulated oxidative or glycolytic me- tabolism of the tumor cell, determines whether T cells have a permissive microenvironment for effective immunotherapy, and that inhibiting tumor cell metabolism may be an attractive strategy to im- prove the efficacy of immunotherapy.
P746
Global metabolomics of advanced gastric cancer patients identifies signatures associated with response and clinical benefit from nivolumab
Carrie Brachmann, PhD1, Carrie Brachmann, PhD1, Daniel Cooper2, Kari Wong2, Ethan Grant1, Daniel Catenacci, MD3, David Cunningham4, Jean- Philippe Metges5, Eric Van Cutsem, MD, PhD6, Zev Wainberg, MD7, Dung Thai1, Pankaj Bhargava1, Emon Elboudwarej1, Ondrej Podlaha1, Wadud Khan8, Jennifer Wargo, MD, MMSc8, Manish Shah, MD9
1
Gilead Sciences, Foster City, CA, United States;2Metabolon, Inc, Morrisville, NC, United States;3University of Chicago Medical Center, Chicago, IL,
United States;4Sutton and London Hospital, London, United Kingdom; 5Brest University Hospital;6University Hospitals Leuven & KU Leuven,
Leuven, Belgium;7UCLA School of Medicine, Santa Monica, CA;8UT MD Anderson Cancer Center, Houston, TX, United States;9Weill Cornell
Medicine, NY Pres Hosp, New York, NY
Correspondence:Carrie Brachmann ([email protected])
Journal for ImmunoTherapy of Cancer2019,7(Suppl 1):P746
Background
The gut microbiome has been reported to impact the immune re- sponse to checkpoint blockade [1]. Circulating metabolites relate to both host and microbiome biology and were explored in an ad- vanced gastric cancer population treated with nivolumab.
Methods
Plasma samples (pre-treatment, week 2, week 8 and end of study) from patients enrolled in a randomized phase 2 study in >=2nd line advanced gastric or gastroesophageal junction patients [2] were ex- tracted for analysis on the LC/MS/MS and Polar LC platforms. 141 pa- tients were treated and all received nivolumab. This analysis evaluates all patients as a single population, since andecaliximab failed to provide additional benefit [2-4]. ANCOVA analyses were per- formed for samples stratified by responder/non-responder status or patients that received clinical benefit versus those that did not re- ceive clinical benefit and were adjusted for age and sex. Clinical benefit was defined as patients living >1 year post treatment or had a tumor response. Results with nominal p
Results
Metabolic profiling of human plasma detected 1364 metabolites. The majority of metabolic changes that occurred while the patients were on treatment were detected at end of study (151 and 35 significantly altered metabolites in non-responders and responders, respectively) when compared to the pre-treatment time point. Clinical benefit from checkpoint blockade was associated with higher levels of bacterial-derived indoles, host-derived tryptophan catabolites, lower primary and secondary bile acids, lower levels of lactate, and altered glutamine catabolism. Similar differences were observed when re- sponders and non-responders were compared.
Conclusions
The association of microbial indoles and host-derived tryptophan ca- tabolites with benefit suggests the activation of immunosuppressive mechanisms in non-responders and those who did not experience clinical benefit. Combined with the differences in secondary bile acids, these data suggest a role for the microbiome. Indeed, trypto- phan catabolites can be produced by Clostridiales and enrichment of these bacteria are linked to improved response to anti-PD1 therapy [5]. In addition, differences in circulating lactate, glutamine, and glu- tamate between responders and non-responders suggested differ- ences in tumor metabolic preferences, flexibility, or overall activity.
Acknowledgements
The authors gratefully acknowledge the patients and their families who participated in this study.
Trial Registration
Clinicaltrials.gov NCT02862535
References
1. Routy B, Le Chatelier E, et al. Gut microbiomer influences efficacy of PD- 1-based immunotherapy against epithelial tumors
2. Shah M, Metges J-M, et al. A phase 2, open-label, randomized study to evaluate the efficacy and safety of andecaliximab combined with nivolu- mab versus nivolumab alone in subjects with unresectable or recurrent gastric or gastroesophageal junction adenocarcinoma. ASCO Gastrointes- tinal Cancers Symposium. 2019.
3. Metges J-M, Elboudwarej E, et al. Exploratory evaluation of baseline tumor biomarkers and their association with response and survival in pa- tients with previously treated advanced gastric cancer treated with ande- caliximab combined with nivolumab versus nivolumab. ASCO
Gastrointestinal Cancers Symposium. 2019.
4. Brachmann C, Zhang Y, et al. Evaluation of intratumoral T cells in biopsies from advanced gastric patients treated with andecaliximab and nivolumab. ASCO Gastrointestinal Cancers Symposium. 2019.
5. Gopalakrishnan V, Spencer CN, et al. Gut microbiome modulates response to anti-PD-1 immunotherapy in melanoma patients. Science. 2018.
Ethics Approval
This study was approved by the institutional review board or independent ethics committee appropriate for each site.
P747
Leveraging the Second Genome platform to develop microbiome- derived cancer immunotherapies
Dhwani Haria, Jayamary Ravichandar, Lynn Yamamoto, Cheryl-Emiliane Chow, Bernat Raja, Joanna Dreux, Kareem Graham, Kathryn Iverson, Shoko Iwai, Sunit Jain, Yuliya Katlinskaya, Sabina Lau, Jina Lee, Michelle Lin, Paul Loriaux, Nicole Narayan, Erica Rutherford, Connor Skennerton, Thomas Weinmaier, Michi Wilcoxon, Yonggan Wu, Todd DeSantis, Toshihiko Takeuchi, Karim Dabbagh, Helena Kiefel, PhD , Helena Kiefel, PhD
Second Genome, South San Francisco, CA, United States Correspondence:Helena Kiefel ([email protected])
Journal for ImmunoTherapy of Cancer2019,7(Suppl 1):P747
Background
The gut microbiota has emerged as an important player in cancer pathology and increased evidence supports its influence on clinical response to immune checkpoint inhibitor (ICI) therapy. However, the
specific gut bacteria-derived molecules responsible for the improved response to ICI therapy remain unknown. Second Genome has devel- oped a unique discovery platform to identify, screen, and validate microbiome-derived peptides that promote response to cancer- immunotherapy.
Methods
Using our multi-technology meta-analysis of published datasets char- acterizing the baseline microbiome from melanoma patients treated with anti-PD-1, we have identified gut microbiota strains differentially abundant in responders versus non-responders that are concordant across multiple cohorts. Next, peptides from the strains associated with responder signatures were predicted from their genome se-
quences. In addition, we identified and selected predicted
bacterially-derived peptides associated with responder signatures from assembled metagenomes. All predicted peptides were screened using phage display technology to select binders to immune cells known to play a role in the tumor microenvironment (TME).
Results
Selected microbiome-derived peptide binders were evaluated for ac- tivity in cell-based assays using isolated human primary T cells, den- dritic cells, and macrophages. We show that several selected peptide binders induce secretion of proinflammatory cytokines and chemo- kines (e.g., IP-10, TNF) from human primary monocyte-derived den- dritic cells, as well as effector cytokine secretion (e.g., IFNγ, IL-2) from human primary T lymphocytes. Additionally, we have identified microbiome-derived peptide binders with capacity to inhibit an M2- like phenotype in macrophages (decreased LPS-induced IL-10 secretion).
Conclusions
Collectively, our results demonstrate the potential of Second Ge- nome’s discovery platform to identify and characterize novel immu- nomodulatory factors produced by the gut microbiota. Second
Genome’s discovery platform offers a unique approach to identify
novel agents with potential for use as therapeutics in cancer immunotherapy.
P748
A rationally-designed consortium of human gut commensals induces IFNg+ CD8+ T cells and enhances tumor immunity
Amanda Prince, PhD1, Johnny Lightcap2, Amit Reddy2, Sachin Deshmukh2, Steve McClellan2, Bruce Roberts, PhD1, Rose Szabady1, Arthur Frankel2
1
Vedanta Biosciences, Inc., Cambridge, MA, United States;2University of South Alabama, Mobile, AL, United States
Correspondence:Bruce Roberts ([email protected]), Arthur Frankel ([email protected])
Journal for ImmunoTherapy of Cancer2019,7(Suppl 1):P748
Background
Efficacy of immune checkpoint inhibitor (ICI) therapy is associated with the composition of the gut microbiome, but the mechanism re- mains to be elucidated. Recent data from Tanoue et al. [1] demon- strated a consortium of 11 bacteria (VE800) induced IFNg production in intestinal CD8+ T cells and enhanced anti-PD-1 cancer activity in tumor models. Vedanta has developed VE800 as a microbiome thera- peutic to enhance the efficacy of ICI therapy. Here, we aimed to fur- ther understand the requirements for VE800 efficacy in pre-clinical models and to understand how immune modulation by VE800 af- fects tumor immunity.
Methods
We used IFNg production in colonic T cells as a readout for VE800 anti-cancer efficacy. Mice were pre-treated with different antibiotics to establish a permissive niche for colonization and given VE800. To determine which bacteria were critical contributors to VE800 func- tion, we created various combinations of individual VE800 bacterial strains and measured induction of IFNg+ CD8+ T cells in the intes- tines of germ-free mice. Consortia unable to induce IFNg+ CD8+ T cells were tested for their anti-tumor efficacy with ICI in the B16 tumor model. In this model, C57BL/6 mice were pre-treated with an- tibiotics to deplete endogenous microbiota, injected with B16, and
treated with ICI on days 4, 8, and 12. VE800 was administered three times a week for three weeks prior to sacrifice to assess levels of tumor infiltrating cells.
Results
Our data suggest that bacterial density drives VE800 efficacy for in- duction of IFNg+ CD8 T cells. We further established that the complete VE800 consortium provides maximal IFNg induction, and that removal of specific strains impairs IFNg induction. Utilizing the B16 tumor model, colonization of VE800 strains was observed in tumor-bearing mice receiving the complete consortia. Additionally, the complete VE800 consortium restored ICI efficacy in antibiotic- treated tumor-bearing mice (Figure 1). However, bacterial consortia incapable of inducing IFNg+ CD8+ T cells do not promote anti-tumor immunity. Pre-treatment with antibiotics abolished the therapeutic effect of ICI, which was accompanied by a reduced level of tumor in- filtrating immune cells indicating the important, supportive role played by gut microbiota.
Conclusions
Our data suggests VE800 induces IFNg+ CD8+ T cell induction and provides the necessary microbial component to promote anti-tumor immunity in combination with dual checkpoint inhibitor therapy in the multiple tumor models.
References
1. Tanoue T, Morita S, Plichta DR, Skelly AN, Suda W, Sugiura Y, Narushima S, Vlamakis H, Motoo I, Sugita K, Shiota A, Takeshita K, Yasuma-Mitobe K, Riethmacher D, Kaisho T, Norman JM, Mucida D, Suematsu M, Yaguchi T, Bucci V, Inoue T, Kawakami Y, Olle B, Roberts B, Hattori M, Xavier RJ, Atar- ashi K, Honda K. A defined commensal consortium elicits CD8 T cells and anti-cancer immunity. Nature. 2019 Jan 23.
P749
Investigation of immune-microbiota interaction in lung cancer using genetically engineered mouse models
Chengcheng Jin, PhD, Tyler Jacks, PhD , Georgia Lagoudas
Massachusetts Institute of Technology, Cambridge, MA, United States Correspondence:Tyler Jacks ([email protected])
Journal for ImmunoTherapy of Cancer2019,7(Suppl 1):P749
Background
The development of lung cancer is closely associated with chronic in- flammation, but the source of such inflammation has not been clearly defined and the contribution of specific cellular and molecular component of the immune system is yet to be elucidated. Pulmonary infections are very common in lung cancer patients and affect clinical
outcomes, but these associations are not understood in molecular details. Although increasing evidence has linked commensal micro- biota to oncogenesis and cancer response to therapy, the current lit- erature has been mainly focused on the intestinal microbiota. It remains unclear how the distal intestinal microbiota and local airway microbiota act together to regulate the malignant transformation in the lung.
Methods
We interrogate the immune-microbiota interaction in lung cancer using an autochthonous genetically engineered mouse model of lung adenocarcinoma driven by oncogenic Kras and loss of p53. Lung and intestinal microbiome were characterized by 16S rDNA- based sequencing analysis. Antibiotics treatment and germ-free ani- mals were used to examine the functional importance of microbiota in lung cancer development. Tumor-associated immune cells were profiled by flow cytometry, bulk and single-cell RNA-seq.
Results
We found lung cancer development was associated with increased total bacterial load and reduced bacterial diversity. This dysregulated local microbiota stimulated IL-1βand IL-23 production from myeloid cells, resulting in proliferation and activation of lung-resident Vγ6+Vδ1+γδ T cells. These γδ T cells produced IL-17 to promote neutrophil infiltration and inflammation in the tumor microenviron- ment; they also expressed IL-22, amphiregulin and other effector molecules to directly enhance tumor cell proliferation. Mice rederived to germ-free condition or treated with combined antibiotics were profoundly protected from lung cancer initiation and progression, whereas intratracheal inoculation of bacterial mixture isolated from late-stage lung tumors robustly accelerated tumor growth in naïve animals. Bulk and single-cell RNA-seq based transcriptional profiling further revealed that microbiota induced robust gene expression changes in a number of tumor-associated immune cell populations. In addition, we also performed 16S FISH and multiplex IF to show the direct interaction between bacterial and immune cells in the na- tive lung tumor microenvironment. Therapeutically, we found block- ing theγδT or its downstream IL-17, IL22 signaling all significantly suppressed tumor growth in the lung.
Conclusions
Our study provides strong evidence that local microbiota play a crit- ical role in promoting lung cancer development by provoking tumor- associated inflammation. The key cellular and molecular mediators that we identified in this study may serve as effective targets in lung cancer treatment and prevention.
P750
Predicting response to anti-PD1 therapy from metagenomic sequencing data with machine learning
Jan Majta, MSc, Krzysztof Odrzywołek, Agata Szymanek, Monika Majchrzak-Gorecka, Szymon Wojciechowski, Zuzanna Karwowska, Kaja