Editorial

TGIF 2018: The gut and immune function

We have entered into a rapidly burgeoning renaissance period of more effective immunological therapy for cancer, building upon the initial incremental results from more than a century of work.

Every month, new sets of data are published showing the impact of various treatments targeted on T cells and other components of the immune response, and we now have compelling data regarding substantial anticancer effects against melanoma, renal cell carcinoma, lung cancer and bladder cancer, with several other areas in development.

One of the remaining challenges is the inconsistency of the proportion of responding patients, and the variability of duration of these remissions.

Impact of the microbiome

The thought that the microbiome of the gastrointestinal tract — so heavily influenced by its bacterial composition — might influence the body’s immune response to external and/or internal challenges is not new.

Derek Raghavan, MD, PhD, FACP, FRACP, FASCO
Derek Raghavan

In 2003, Erdman and colleagues, investigating the association between inflammatory bowel disease and colonic malignancy, showed that innate immune dysregulation may provide the link between these two disorders. Specifically, these workers inoculated 129/SvEv Rag2-deficient mice (that were T lymphocyte deficient) and congenic wild-type mice (with normal immune function) with Helicobacter hepaticus or placebo, causing the pathogen-loaded Rag2-deficient mice to develop colitis and, subsequently, colonic cancer. The wild-type mice, with functioning T cells, developed neither condition. To complete their proof, researchers showed that adoptive transfer of regulatory T cells into the Rag2-deficient mice prevented colitis and cancer.

Paulos and colleagues showed that total body irradiation, by lowering levels of inhibitory T cells, increased the activity of adoptively transferred CD8-positive lymphocytes. However, this effect was largely vitiated if antibiotics were used to disrupt the gut microflora, and could be reversed by reintroduction of the bacteria or administration of ultra-pure lipopolysaccharides from bacterial cultures, again suggesting an important interplay between microbiome and immune function.

Although these observations have clearly been of interest in the efforts to explain the connections between inflammatory bowel disease and subsequent gastrointestinal malignancy, they seem to have taken on a new connotation since the recent therapeutic introduction of immune checkpoint blockade into the active therapeutic management of several malignancies.

To remind you, the concept here is that immune checkpoint blockade releases an attack by T lymphocytes by suppressing the interactions between T-cell inhibitory receptors with their ligands on stromal or tumor cells.

One example is the use of monoclonal antibodies raised against the programmed cell death protein 1 (PD-1) or its ligand (PD-L1). To date, it has been thought that the less-than-50% response rate could be due to low antigenicity of tumor cells, low mutational burden, variable expression or distribution of PD-1/PD-L1, aberrations of antigen presentation, or the impact of prior radiotherapy or chemotherapy on the lymphocytes.

Gut response

Now, in addition to these factors, there has been increasing consideration that gut microflora may somehow influence the efficacy of immune checkpoint blockade.

In this year’s first edition of Science, three fascinating papers provide a scientific basis for what our immunotherapy colleagues have suspected for some time — namely, that reduction in the heterogeneity in gut bacteria may impact the efficacy of this treatment against cancer.

Gopalakrishnan and colleagues studied microbiome samples from the oral cavity and gastrointestinal tract of patients with malignant melanoma during the course of their immune checkpoint blockade treatment. Researchers also assessed tumor biopsies and blood samples for parallel genomic alterations and the characteristics of tumor-infiltrating and circulating T cells.

They showed that within-sample diversity of bacterial content in the gut was much higher in patients responding to immune checkpoint blockade, but that there were no response-associated differences in the constituency of oral microflora. Researchers characterized bacterial content, noting that responding patients had a surfeit of Clostridiales and Ruminococcaceae — in particular, Faecalibacterium — whereas nonresponders had greater presence of Bacteroidales.

Further studies with whole-genome sequencing indicated that anabolic functioning pathways predominantly characterized the gut of responding patients, whereas nonresponse to immune checkpoint blockade was predominantly associated with catabolic function. Closing the loop, researchers also showed much higher tumor infiltration by CD8-positive lymphocytes at baseline among responding patients.

This group also completed parallel murine modeling studies, perhaps the most compelling of which were those in which researchers transplanted fecal microbiota from responding patients to germ-free mice treated by immune checkpoint blockade, showing apparent transmission of tumor response. In contrast, mice receiving bacteria from nonresponding patients showed no response.

In a parallel study, Matson and colleagues also studied stool samples from patients treated with immune checkpoint blockade for metastatic melanoma, using three methods of DNA sequence-based bacterial identification. They identified eight bacterial species associated with immune checkpoint blockade response, including Enterococcus faecium, Collinsella aerofaciens, Bifidobacterium adolescentis, Klebsiella pneumoniae and Lactobacillus sp. By contrast, Ruminococcus obeum and Roseburia intestinalis were associated with nonresponse.

Routy and colleagues reported that primary resistance to immune checkpoint blockade among epithelial tumors may also be attributed to an abnormal gut microbiome composition — specifically in an elegant series of murine and human paired studies, they have shown that:

What remains unclear is why A. muciniphila is associated with a good response, as it is such a ubiquitously present gut contaminant.

Leaving nothing to chance, these investigators also fed antibiotic-treated mice by gavage a mixture of A. muciniphila and other bacteria, and studied the characteristics of the resulting enhanced T-cell responses, further confirming their hypotheses. In parallel, they studied the results of PD-1/PD-L1 inhibition among patients who had either received or not received antibiotics for a range of clinical indications, noting substantially improved response rates, PFS and OS in those untreated by antibiotics.

Application of translational studies

So, what does this all mean?

These carefully constructed translational studies have all indicated that the gut microbiome matters when considering response to immune checkpoint blockade.

The methods employed were quite different — including different molecular diagnostics, cutoff values from bacterial content and assessment of response — so it is not perturbing that the bacterial species that connote for likely immune checkpoint blockade response differ among these reports.

What is important is that oncologists using immune checkpoint blockade will need to recruit colleagues from the world of microbiology and molecular diagnostics to help to characterize gut microbiota. Additional retrospective and prospective studies — which will soon need to be completed — may define more extensively the mechanistic explanations of the biology of immune checkpoint blockade response and its interaction with normal bacterial constituents of the gastrointestinal tract.

This body of work may ultimately also have relevance to understanding the genesis of human cancer. It is so stimulating to see fine application of the translational interface that is likely to benefit patients directly and improve our treatment algorithms.

References:

Erdman SE, et al. Am J Pathol. 2003;162:691-702.

Gopalakrishnan V, et al. Science. 2018,doi:10.1126/science.aan4236.

Matson V, et al. Science. 2018;doi:10.1126/science.aao3290.

Paulos CM, et al. J Clin Invest. 2007;117:2197-2204.

Routy B, et al. Science. 2018;doi:10.1126/science.aan3706.

For more information:

Derek Raghavan, MD, PhD, FACP, FRACP, FASCO, is HemOnc Today’s Chief Medical Edi­tor for Oncology. He also is president of Levine Cancer Institute at Carolinas HealthCare Sys­tem. He can be reached at derek.raghavan@carolinashealthcare.org.

Disclosure: Raghavan reports no relevant financial disclosures.

We have entered into a rapidly burgeoning renaissance period of more effective immunological therapy for cancer, building upon the initial incremental results from more than a century of work.

Every month, new sets of data are published showing the impact of various treatments targeted on T cells and other components of the immune response, and we now have compelling data regarding substantial anticancer effects against melanoma, renal cell carcinoma, lung cancer and bladder cancer, with several other areas in development.

One of the remaining challenges is the inconsistency of the proportion of responding patients, and the variability of duration of these remissions.

Impact of the microbiome

The thought that the microbiome of the gastrointestinal tract — so heavily influenced by its bacterial composition — might influence the body’s immune response to external and/or internal challenges is not new.

Derek Raghavan, MD, PhD, FACP, FRACP, FASCO
Derek Raghavan

In 2003, Erdman and colleagues, investigating the association between inflammatory bowel disease and colonic malignancy, showed that innate immune dysregulation may provide the link between these two disorders. Specifically, these workers inoculated 129/SvEv Rag2-deficient mice (that were T lymphocyte deficient) and congenic wild-type mice (with normal immune function) with Helicobacter hepaticus or placebo, causing the pathogen-loaded Rag2-deficient mice to develop colitis and, subsequently, colonic cancer. The wild-type mice, with functioning T cells, developed neither condition. To complete their proof, researchers showed that adoptive transfer of regulatory T cells into the Rag2-deficient mice prevented colitis and cancer.

Paulos and colleagues showed that total body irradiation, by lowering levels of inhibitory T cells, increased the activity of adoptively transferred CD8-positive lymphocytes. However, this effect was largely vitiated if antibiotics were used to disrupt the gut microflora, and could be reversed by reintroduction of the bacteria or administration of ultra-pure lipopolysaccharides from bacterial cultures, again suggesting an important interplay between microbiome and immune function.

Although these observations have clearly been of interest in the efforts to explain the connections between inflammatory bowel disease and subsequent gastrointestinal malignancy, they seem to have taken on a new connotation since the recent therapeutic introduction of immune checkpoint blockade into the active therapeutic management of several malignancies.

To remind you, the concept here is that immune checkpoint blockade releases an attack by T lymphocytes by suppressing the interactions between T-cell inhibitory receptors with their ligands on stromal or tumor cells.

One example is the use of monoclonal antibodies raised against the programmed cell death protein 1 (PD-1) or its ligand (PD-L1). To date, it has been thought that the less-than-50% response rate could be due to low antigenicity of tumor cells, low mutational burden, variable expression or distribution of PD-1/PD-L1, aberrations of antigen presentation, or the impact of prior radiotherapy or chemotherapy on the lymphocytes.

PAGE BREAK

Gut response

Now, in addition to these factors, there has been increasing consideration that gut microflora may somehow influence the efficacy of immune checkpoint blockade.

In this year’s first edition of Science, three fascinating papers provide a scientific basis for what our immunotherapy colleagues have suspected for some time — namely, that reduction in the heterogeneity in gut bacteria may impact the efficacy of this treatment against cancer.

Gopalakrishnan and colleagues studied microbiome samples from the oral cavity and gastrointestinal tract of patients with malignant melanoma during the course of their immune checkpoint blockade treatment. Researchers also assessed tumor biopsies and blood samples for parallel genomic alterations and the characteristics of tumor-infiltrating and circulating T cells.

They showed that within-sample diversity of bacterial content in the gut was much higher in patients responding to immune checkpoint blockade, but that there were no response-associated differences in the constituency of oral microflora. Researchers characterized bacterial content, noting that responding patients had a surfeit of Clostridiales and Ruminococcaceae — in particular, Faecalibacterium — whereas nonresponders had greater presence of Bacteroidales.

Further studies with whole-genome sequencing indicated that anabolic functioning pathways predominantly characterized the gut of responding patients, whereas nonresponse to immune checkpoint blockade was predominantly associated with catabolic function. Closing the loop, researchers also showed much higher tumor infiltration by CD8-positive lymphocytes at baseline among responding patients.

This group also completed parallel murine modeling studies, perhaps the most compelling of which were those in which researchers transplanted fecal microbiota from responding patients to germ-free mice treated by immune checkpoint blockade, showing apparent transmission of tumor response. In contrast, mice receiving bacteria from nonresponding patients showed no response.

In a parallel study, Matson and colleagues also studied stool samples from patients treated with immune checkpoint blockade for metastatic melanoma, using three methods of DNA sequence-based bacterial identification. They identified eight bacterial species associated with immune checkpoint blockade response, including Enterococcus faecium, Collinsella aerofaciens, Bifidobacterium adolescentis, Klebsiella pneumoniae and Lactobacillus sp. By contrast, Ruminococcus obeum and Roseburia intestinalis were associated with nonresponse.

Routy and colleagues reported that primary resistance to immune checkpoint blockade among epithelial tumors may also be attributed to an abnormal gut microbiome composition — specifically in an elegant series of murine and human paired studies, they have shown that:

What remains unclear is why A. muciniphila is associated with a good response, as it is such a ubiquitously present gut contaminant.

Leaving nothing to chance, these investigators also fed antibiotic-treated mice by gavage a mixture of A. muciniphila and other bacteria, and studied the characteristics of the resulting enhanced T-cell responses, further confirming their hypotheses. In parallel, they studied the results of PD-1/PD-L1 inhibition among patients who had either received or not received antibiotics for a range of clinical indications, noting substantially improved response rates, PFS and OS in those untreated by antibiotics.

PAGE BREAK

Application of translational studies

So, what does this all mean?

These carefully constructed translational studies have all indicated that the gut microbiome matters when considering response to immune checkpoint blockade.

The methods employed were quite different — including different molecular diagnostics, cutoff values from bacterial content and assessment of response — so it is not perturbing that the bacterial species that connote for likely immune checkpoint blockade response differ among these reports.

What is important is that oncologists using immune checkpoint blockade will need to recruit colleagues from the world of microbiology and molecular diagnostics to help to characterize gut microbiota. Additional retrospective and prospective studies — which will soon need to be completed — may define more extensively the mechanistic explanations of the biology of immune checkpoint blockade response and its interaction with normal bacterial constituents of the gastrointestinal tract.

This body of work may ultimately also have relevance to understanding the genesis of human cancer. It is so stimulating to see fine application of the translational interface that is likely to benefit patients directly and improve our treatment algorithms.

References:

Erdman SE, et al. Am J Pathol. 2003;162:691-702.

Gopalakrishnan V, et al. Science. 2018,doi:10.1126/science.aan4236.

Matson V, et al. Science. 2018;doi:10.1126/science.aao3290.

Paulos CM, et al. J Clin Invest. 2007;117:2197-2204.

Routy B, et al. Science. 2018;doi:10.1126/science.aan3706.

For more information:

Derek Raghavan, MD, PhD, FACP, FRACP, FASCO, is HemOnc Today’s Chief Medical Edi­tor for Oncology. He also is president of Levine Cancer Institute at Carolinas HealthCare Sys­tem. He can be reached at derek.raghavan@carolinashealthcare.org.

Disclosure: Raghavan reports no relevant financial disclosures.