Friendly Bacteria: Allies in the Battle Against Cancer

Based on: Gut microbiome influences efficacy of PD-1-based immunotherapy against epithelial tumors

            Chances are, you’ve heard something in the news about the “gut microbiome” or seen the endless rows of probiotic supplements in the aisles at the grocery store.  Our intestines are filled with multiple different types of bacteria and other microbes, but don’t worry: these little guys are actually here to help us!  Known as commensal microorganisms, these bacteria provide us with lots of different products that have a variety of functions under normal conditions.  A lot of the current hype about the gut microbiome relates to their effects on mental health conditions, and there is some science backing up this association (for more information, check out this Scientific American article!).  Perhaps most importantly, the members of our gut microbiome have been shown to affect our immune systems in multiple ways: they can increase the amounts of anti-inflammatory molecules like interleukin-10 (IL-10), increase the number of regulatory T cells (Tregs, which help control immune responses), and decrease the maturation of Th17 cells that are frequently involved in autoimmune attacks (Omenetti and Pizarro, 2015; Tanaka et al., 2016). 

           

A little bit about cancer immunology

            Currently, immunologists believe that one of the functions of our immune system is to constantly survey for warning signs that a cell may be starting a tumor.  Certain changes occur in a cell when it becomes cancerous, and immune cells are able to recognize these signals and respond by killing the transformed cell before it causes further problems.  Tumor cells express proteins called tumor specific antigens (TSAs) and tumor associated antigens (TAAs) that, along with other danger signals, serve as red flags to tell the B and T cells of the immune system that the cell needs to be destroyed.  The immune response is typically highly inflammatory and is controlled by a particular set of molecules called Th1 cytokines (which will become important for this paper!).

How the immune system detects and destroys cancer cells (Chen and Mellman, 2013)

            Tumor cells use some crafty escape maneuvers to avoid being sensed by the immune system, and one of these tactics is increasing the expression of molecules that inhibit the T cells driving the response.  PDL-1 is usually expressed on the cells that activate T cells in response to damage or pathogen invasion, and it binds to the PD-1 receptor on the T cells themselves.  Activation of PD-1 tells the T cell not to respond to whatever it is being shown.  Normally, PD-1 signaling puts the brakes on the immune response to make sure that T cells aren’t overactivated; however, tumor cells will express high amounts of PDL-1 to trick T cells into ignoring the TSAs and TAAs.  This results in an inability to clear cancerous cells.  A newer form of cancer therapy called checkpoint blockade helps T cells avoid binding to the PDL-1 on tumor cells by covering the PD-1 receptor with an antibody, which then allows the T cells to continue to attack the cancer.  For a quick video explanation of anti-PD1 cancer therapy, follow this link!    

How anti-PD-1 antibody therapy ("antagonistic antibody") unleashes T cells to kill cancer cells (http://bpsbioscience.com/screening-profiling-services/pd-1-screening-services)

A healthy individual’s microbiome can vary due to a variety of factors, including their genetics, where they live, their stress levels, antibiotic use, and the type of diet that they eat (Benson et al., 2010).  When the composition of the microbiome shifts too much, however, the immune system can be severely affected.  This disease-associated change is called dysbiosis, and it has been seen in many autoimmune disorders such as multiple sclerosis, ulcerative colitis, Celiac disease, and lupus (Ott et al., 2008; He et al., 2016; Cekanaviciute et al., 2017; Girbovan et al., 2017).  Since our immune cells also play a role in surveying for cancer, it is possible that differences in the gut microbiome may affect the way that the immune system responds to the threat of tumor cells.

The Experiments

            Because the use of antibiotics typically disrupts the gut microbiome (Blaser, 2016), Routy et al. wanted to determine whether dysbiosis caused by antibiotic use affects the response to checkpoint blockade therapy.  First, they tested their hypothesis in mice with the mouse versions of melanoma (skin cancer) and sarcoma (connective tissue cancer).  They split the mice into four groups: one received antibiotics only, the second received antibiotics and anti-PD1 antibody therapy, the third received saline alone, and the fourth received saline and anti-PD1 antibody therapy.  Tumor size was measured as an index of how well the anti-PD1 antibody worked; a smaller tumor means that the antibody was able to fire up the T cells to attack the tumor.  Mice who were treated with saline and anti-PD1 therapy had the smallest melanoma tumors, but the anti-PD1 treatment did not shrink the tumors as much in mice who also received antibiotics.  These results were also replicated in the sarcoma model, and they suggest that taking antibiotics may reduce the effectiveness of checkpoint blockade.  As another index of anti-PD1 treatment success, the authors also measured survival after the appearance of tumors.  Compared to anti-PD1 therapy alone, mice who were given antibiotics before starting the therapy did not survive either cancer as long.

            Since the mouse models had supported the hypothesis that dysbiosis may affect how well anti-PD1 works to shrink tumors, the authors also studied a group of human patients with non-small cell lung cancer and kidney cancer who had been treated with anti-PD1.  They asked whether the patients had been prescribed antibiotics (for typical infections) before receiving anti-PD1 antibodies and also recorded how long they survived without their cancer getting worse.   Patients who had not taken antibiotics lived for a significantly longer time without their symptoms getting worse and for a longer time overall, which agrees with what Routy et al. saw in the mouse models.  Both support the idea that disturbances in the gut microbiome might affect the way that the immune system responds to tumor cells after being given a boost by anti-PD1 antibodies.

Human cancer patients survived longer after anti-PD1 therapy if they had not been prescribed antibiotics (black line)

            To evaluate what kinds of bacteria may be involved in altering the effectiveness of anti-PD1 checkpoint blockade, Routy et al. used a standardized set of criteria to designate some cancer patients as “responders” to the therapy (they improved after treatment) and some as “non-responders” (whose status did not change or got worse).  They then analyzed stool samples from each patient for various types of bacteria.  Some types of bacteria were enriched in responders, while others were found more commonly in non-responders.  One bacterium that was found more commonly both in those who partially responded to anti-PD1 treatment and those who survived without progression for more than three months was Akkermansia muciniphila. 

Because A. muciniphila yielded the most significant results, the authors decided to study it further.  They then classified the lung and kidney cancer patients into three groups: partial responders, those with stable cancer, and those whose cancer got worse (progressive disease).  The partial responder category had the highest percentage of patients with A. muciniphila in their stool at diagnosis, and the progressive disease had the lowest percentage of patients with A. muciniphila.  To determine the effects of the bacterium on systemic T cell responses, Routy et al. grew T cells from all categories of patients in a plate with A. muciniphila.  This experiment would allow them to see whether the patients had T cells that recognized the bacterium and what signaling molecules are produced in response.  They found that A. muciniphila caused T cells to produce a lot of interferon-γ (IFN-γ), which drives the highly inflammatory Th1 response that helps kill tumor cells.  They then correlated the amount of IFN-γ made by a patient’s T cells in response to A. muciniphila with how long the patient survived without progression; patients whose cells made high levels of IFN-γ survived longer without their cancer getting worse.  Another bacterium, Enterococcus hirae, showed similar effects.

Patients who had a partial response (PR) had the most Akkermansia muciniphila in their stool, while levels decreased for patients with stable disease (SD) and those with progressive disease (PD)

To provide stronger evidence that microbiome alterations cause failure to respond to anti-PD1 therapy, Routy et al. used fecal microbiota transfers to manipulate the microbiomes of mice.  They treated mice designed to develop sarcoma with high doses of antibiotics to wipe out their natural microbiomes before recolonizing their guts with fecal matter from human responders or non-responders.  The mice were then treated with either control antibodies or antibodies directed against PD-1.  When the mice received microbiomes from non-responders, they too failed to see a reduction in tumor size after anti-PD1 therapy.  Mice recolonized with stool from responders, however, had significantly smaller tumors after anti-PD1 treatment.  The authors then measured the amount of T cells expressing a receptor called CXCR3 (which allows the cells to move towards sites of inflammation) that were able to move into the tumor after anti-PD1 treatment.  Significantly more T cells expressing CXCR3 made it into the tumor when mice had microbiomes from responders than when they received fecal transfers from non-responders.

Schematic of experiment (L) and mice who received stool from responders had a bigger reduction in tumor size after anti-PD1 therapy than mice who received microbiomes from non-responders

 To verify their results in the sarcoma model, Routy et al. then used another mouse cancer model in which kidney cancer cells that express a fluorescent chemical are transferred into the mice instead of sarcoma cells.  The mice were then treated with antibiotics and recolonized with stool from responders or non-responders.  Though this type of kidney cancer usually responds well to treatment with anti-PD1 antibodies, the tumors were even smaller in mice who received stool from patients who also responded to anti-PD1 therapy.

Next, Routy et al. investigated the direct effects of reconstituting antibiotic-treated mice’s microbiomes with A. muciniphila.  The mice were treated with anti-PD1 antibodies and A. muciniphila simultaneously, and melanoma tumors were significantly smaller than mice who received control antibodies with no bacteria.  Treatment with anti-PD1 therapy alone, however, did not produce this shrinking effect.  The authors then used yet another mouse model of lung cancer; simultaneous treatment with A. muciniphila and anti-PD1 antibodies after radiation therapy was more effective than only radiation and anti-PD1 therapies.  In the sarcoma model, adding A. muciniphila along with stool from non-responders prevented tumor growth better after anti-PD1 treatment than simply reconstituting with non-responder stool alone, indicating that the bacterium is the component responsible for making the mice respond to anti-PD1 therapy.

Mice who were treated with A. muciniphila, radiotherapy, and anti-PD1 treatment (triangles) had smaller lung tumors than mice who got radiotherapy and anti-PD1 therapy without bacteria (x shapes).

Since the goal of anti-PD1 therapy is to ramp up T cell activation and Tregs suppress the response, the authors also wanted to investigate whether A. muciniphila increased the effectiveness of anti-PD1 treatment by reducing Treg numbers.  They found that mice who received A. muciniphila had fewer Treg cells in their tumors, and that having fewer Treg cells was correlated with smaller tumor size.  Additionally, Routy et al. wanted to confirm that A. muciniphila promotes a pro-inflammatory Th1 response.  They used antibodies that cause an important Th1 signaling molecule, interleukin 12 (IL-12), to stop functioning.  When mice were treated with a combination of the anti-IL-12 antibodies, anti-PD1 antibodies, and A. muciniphila, the enhanced tumor-shrinking effects were reversed.  These results indicate that the capability of the bacterium to promote responsiveness to anti-PD1 cancer therapy relies on IL-12 signaling and decreases the amount of inhibitory Tregs to promote tumor-suppressive inflammatory effects.

So what?

Routy et al. demonstrated that the presence of Akkermansia muciniphila affects the way that mice with cancer respond to anti-PD1 antibody therapy.  They also found that human cancer patients who did improve with anti-PD1 therapy had more A. muciniphila than patients who did not respond to anti-PD1 checkpoint blockade.  The bacterium likely works by making it easier for T cells to invade the tumor and depends on pro-inflammatory molecules.  Though the friendly A. muciniphila is usually not attacked by the immune system, unleashing the T cell response with anti-PD1 antibodies allows the bacterium to activate Th1 cells that can then attack cancer in other places.  In the near future, supplementing the gut microbiomes of cancer patients might help checkpoint blockade immunotherapy work better for a wider group of patients!

References

Benson AK, Kelly SA, Legge R, Ma F, Low SJ, Kim J, Zhang M, Oh L, Nehrenberg D, Hua K, Kachman SD, Moriyama EN, Walter J, Peterson DA, Pomp D (2010) Individuality in gut microbiota composition is a complex polygenic trait shaped by multiple environmental and host genetic factors. Available at: www.pnas.org/cgi/doi/10.1073/pnas.1007028107.

Blaser MJ (2016) Antibiotic use and its consequences for the normal microbiome. Science 352:544–545 Available at: http://www.ncbi.nlm.nih.gov/pubmed/27126037.

Cekanaviciute E, Yoo BB, Runia TF, Debelius JW, Singh S, Nelson CA, Kanner R, Bencosme Y, Lee YK, Hauser SL, Crabtree-Hartman E, Sand IK, Gacias M, Zhu Y, Casaccia P, Cree BAC, Knight R, Mazmanian SK, Baranzini SE (2017) Gut bacteria from multiple sclerosis patients modulate human T cells and exacerbate symptoms in mouse models. Proc Natl Acad Sci U S A 114:10713–10718.

Chen DS, Mellman I (2013) Oncology Meets Immunology: The Cancer-Immunity Cycle. Immunity 39:1–10 Available at: https://www.sciencedirect.com/science/article/pii/S1074761313002963?via%3Dihub#fig1.

Girbovan A, Sur G, Samasca G, Lupan I (2017) Dysbiosis a risk factor for celiac disease. Med Microbiol Immunol 206:83–91 Available at: http://link.springer.com/10.1007/s00430-017-0496-z.

He Z, Shao T, Li H, Xie Z, Wen C (2016) Alterations of the gut microbiome in Chinese patients with systemic lupus erythematosus. Gut Pathog 8:64 Available at: http://gutpathogens.biomedcentral.com/articles/10.1186/s13099-016-0146-9.

Omenetti S, Pizarro TT (2015) The Treg/Th17 Axis: A Dynamic Balance Regulated by the Gut Microbiome. Front Immunol 6:639 Available at: http://journal.frontiersin.org/Article/10.3389/fimmu.2015.00639/abstract.

Ott SJ, Plamondon S, Hart A, Begun A, Rehman A, Kamm MA, Schreiber S (2008) Dynamics of the Mucosa-Associated Flora in Ulcerative Colitis Patients during Remission and Clinical Relapse. J Clin Microbiol 46:3510–3513 Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2566070/pdf/1512-08.pdf.

Tanaka S, Yamamoto K, Yamada K, Furuya K, Uyeno Y (2016) Relationship of Enhanced Butyrate Production by Colonic Butyrate-Producing Bacteria to Immunomodulatory Effects in Normal Mice Fed an Insoluble Fraction of Brassica rapa L. Appl Environ Microbiol 82:2693–2699 Available at: http://www.ncbi.nlm.nih.gov/pubmed/26921420.