As a postdoc at the Broad Institute of MIT and Harvard in 2008, cancer biologist Ravid Straussman worked on a collaborative effort to screen normal cells for their influence on cancer cells’ responses to various anticancer therapies. The project turned up hundreds of examples of stromal cells—best known for forming the connective tissues that support organs—somehow conferring drug resistance on their cancerous neighbors in vitro. One of these cases caught Straussman’s attention: human pancreatic and colorectal cancer cell lines could evade the chemotherapeutic drug gemcitabine when cultured with dermal fibroblasts.
In the lab, the cancer cells “continue to proliferate under therapy,” Straussman explains. “This was really fascinating to me because I didn’t understand . . . how stromal cells can protect cancer cells from chemotherapy.”
He spent the next year and a half trying to figure out what was going on. He ruled out several hypotheses, including that the stromal cells were secreting a protective protein or exosome. “We had tons of ideas that turned out to be not true,” Straussman says. Then, he discovered that the fibroblasts he and his colleagues had been using contained bacteria called Mycoplasma, a common contaminant of tissue and cell cultures. “I was devastated at the time,” he recalls—in presentations, he has used a photo of a wrecked sedan to depict his reaction to the discovery. “I spent a year and a half, probably, tracing this, and this is a complete artifact.” He almost trashed the whole project.
“But then I got curious.”
In subsequent experiments, Straussman confirmed that it was the bacteria that were rescuing the cancer cells from chemotherapy: when he treated the fibroblasts with an antibiotic, they no longer protected cancer cells from gemcitabine, and when he infected other stromal cells with Mycoplasma, it endowed them with the ability to shield cancer cells from the drug. “So we know for a fact it wasn’t the stromal cells that rescued the cancer cells from chemotherapy, but actually it was the Mycoplasma that did that.”
The question remained whether such bacteria are naturally present in cancer. Contrary to the common dogma at the time that tumors were sterile, Straussman says he wondered if the cancerous growths could in fact be hosting microbial life, like so many tissues in the human body that modern genomic tools had revealed to harbor their own microbiomes. Gemcitabine was then a first-line treatment for pancreatic cancer, and is still commonly used. “If there are bacteria in pancreatic cancer and bacteria do such a beautiful trick, this can be really interesting.”
When Straussman accepted a faculty position at the Weizmann Institute of Science in Israel in 2013, he decided to pursue this line of research. Along with Weizmann colleagues and collaborators back at the Broad, Straussman sequenced samples from 113 human pancreatic cancers and scoured the results for bacterial DNA. Sure enough, the researchers found genetic signatures of bacteria in 76 percent of them. “We never found Mycoplasma in pancreatic cancer, but we found many other bacteria,” says Straussman, “and we found that [when] we isolated these bacteria from tumors, they can also rescue cancer cells from chemotherapy” in vitro.
Everything that has to do with tumor biology can now be reexamined. There's endless directions in which we can go.—Ravid Straussman, Weizmann Institute of Science
The next piece of the puzzle, then, was how. Further research showed that bacteria lived inside the cancer cells near the nuclear envelope. There, they would take up gemcitabine before it could reach the cell’s genome and they’d cleave an amine group from the drug, effectively deactivating it. The findings were so unexpected that it took one of the group’s own technicians more than a year to be convinced of their validity, notes Straussman, who has since filed for a few patents related to the presence of bacteria in tumors. When the team published its work in Science in September 2017, it was one of the first conclusive demonstrations that a tumor outside the gut had bacteria living within it.
While the thinking at the time was that cancers were generally germ-free tissues, there had long been reports of microbes in tumors. The problem was that contamination could not be ruled out in most cases. But in the less than five years since Straussman and his team revealed those findings, the evidence for the idea has exploded as various groups have carefully documented the presence of cancer type–specific microbial communities. There have been “literally dozens of other papers,” says microbiome researcher Rob Knight of the University of California, San Diego (UCSD), and these have provided “plenty of evidence that tumors do have their own microbiome.” Indeed, this January, Douglas Hanahan of the Swiss Federal Institute of Technology Lausanne added “polymorphic microbiomes” to the famous list of cancer hallmarks, noting that “while the gut microbiome has been the pioneer of this new frontier, multiple tissues and organs have associated microbiomes.”
Researchers are moving quickly to turn these discoveries into clinically useful tools. The mere presence of microbes in tumors could provide a conceptual basis for new cancer diagnostics, and Knight recently cofounded Micronoma, a company that aims to develop such tests based on bacterial nucleic acid signatures in the blood. Researchers are also exploring the possibility of manipulating the tumor microbiome to treat cancer, and Straussman says he is now in the process of launching a new company, Baccine, to do just that. (He also consults for Biomica, which is focused on gut microbiome–based therapeutics for a range of diseases, including cancer.)
But while some scientists are already looking toward potential clinical applications, many basics of tumor microbiome biology remain unresolved. Some of the most pressing questions pertain to whether and how microbes are influencing the development and progression of cancer, as well as patients’ responses to therapy. “There is a lot of fundamental biology that is yet to be discovered,” says Micronoma cofounder and executive Greg Sepich-Poore, a recent PhD graduate from Knight’s group.
“Describing the presence of these bacteria and viruses, it’s only step one,” agrees virologist Nadim Ajami, director of scientific research for MD Anderson’s Program for Innovative Microbiome and Translational Research (PRIME TR), an effort focused on interrogating microbe-host interactions in cancer. “What we’re really after is [understanding] their role [in the] onset of cancer, progression of disease, response to therapy, with an eye on how can we use all that information to perhaps prevent the development of cancer or increase therapy efficacy.”
See “Microbes Meet Cancer”
Tumors are melting pots
The overturning of the tumors-are-sterile dogma is thanks in large part to the development of next-gen sequencing and the curation of massive data sets of cancer genomes. As MD Anderson’s Jennifer Wargo explains, “Whenever you do whole-genome sequencing, not only do you get human reads, you also get microbial reads. . . . Rather than [filtering] microbes out, we can actually get a lot of clarity about what microbes are there [and] what function do they have.”
“But,” adds Wargo, who started PRIME TR, “there’s a lot of complexity because any time these tissues were collected, it wasn’t necessarily for the sole intent of looking at microbes, and so they may have been collected in a way that really complicates that. There could be a lot of contaminants.”
This is a particular challenge because the microbial communities found in tumors are relatively small—bacteria are far less abundant than in the gut, for instance. Such low-biomass samples, as they’re called, are highly sensitive to contamination, which, according to Straussman, is entirely unavoidable. “A few years ago, we went into a lot of hassle trying to just clean everything completely . . . [and] we found out we just cannot do it,” he says. “There’s always some bacterial DNA contamination.”
So when he, Wargo, and their colleagues decided to survey 1,526 samples of seven different tumor types—including breast, lung, pancreatic, and brain—to rigorously address the question of whether bacteria were commonplace in cancer, they needed to control for this inevitability. They did so by incorporating negative controls in each step of the process, including extracting the DNA, cycling the samples through PCR, and sequencing the resulting nucleic acids. They even took bits of the paraffin block that the tumor tissue was preserved in, to control for contaminants introduced when the samples were first taken. “So at the end of the day,” Straussman says, “we know what the background noise is and what [are] the true sequences that are coming from the tumor.”
After eliminating more than 90 percent of the reads based on these controls, the group was able to identify a distinct bacterial signature for each tumor type, with breast cancer exhibiting a particularly abundant and diverse microbiome. Then, to ascertain if the bacteria were actually living within the tumors, the team cultured tumor slices and treated them with a fluorescently stained enantiomer of the amino acid alanine, which bacteria incorporate into their cell walls. When they saw the glow of the bacteria under the microscope, it was “nice proof of the fact that live bacteria are present in these tumors,” says Straussman.
The results, published in Science in May 2020, took “another step further into the right direction, which is they did not only rely on bioinformatics tools to determine the presence of these bacterial signals; they coupled that with imaging,” says Ajami, who consults or advises for a few biotech companies involved in microbiome research and development. “So they were able to identify intratumoral bacteria—and in some cases, intracellular bacteria—that were present in these tumors.” For good measure, Straussman’s team cultured bacteria from breast tumor samples taken from five women undergoing surgery and was able to cultivate hundreds of colonies, including members of three main phyla, Proteobacteria, Firmicutes, and Actinobacteria (recently renamed Pseudomonadota, Bacillota, and Actinomycetota).
Straussman is not alone in his mission to document and interrogate microbiomes in tumors. More and more studies present new data and novel approaches to help propel the conversation beyond potential contamination concerns. Margaret Sällberg Chen, a clinical immunologist and cancer microbiome researcher at the Karolinska Institutet in Sweden, worked with physicians there to sample bacteria from precancerous pancreatic cysts—in which they’d previously identified bacterial DNA—in the operating room. “When they lifted out the pancreas, they could take the sample and inoculate it directly to the culture medium,” she explains. The work, published in November 2021, resulted in successful cultures of Enterococcus, Enterobacter, and Klebsiella bacteria, among other groups. Because some of these species die upon exposure to oxygen, time was of the essence when sampling the tumors and preserving their microbial communities, says Sällberg Chen. “By having this very fast culturing method, we succeeded to cultivate some of the pancreas microbiome.”
While much more research is needed to pin down the function of microbes in cancer, researchers who spoke with The Scientist agree that the evidence all points to the idea that tumors are indeed not sterile, but typically harbor a microbiome. “The bottom line is, yes,” says Straussman, “there’s actually live bacteria in tumors.”
MICROBIAL SIGNATURES OF CANCER
In the past few years, researchers have published dozens of studies documenting the presence of bacteria and other microbes in the tumors, gut, and blood of patients with cancer. Multiple groups have uncovered preliminary correlations between microbial signatures in these tissues and a patient’s diagnosis, prognosis, or response to treatment, which could one day help inform clinical care.
Straussman’s 2020 Science paper came on the heels of somewhat similar study, published in Nature by Knight’s group. Like Straussman’s team, Knight and his colleagues had surveyed a range of tumor types and discarded up to nearly 93 percent of the sequence data from their analyses to guard against contamination. And like Straussman’s findings, Knight’s results had showcased microbes as common features of cancer, with microbial makeup varying by cancer type. But the UCSD team had taken an entirely different approach—clean and analyze existing sequence data rather than sequence samples anew—and they had come to the idea for very personal reasons.
Sepich-Poore was a freshman in college when his grandmother was diagnosed with pancreatic cancer the day after Christmas 2012. By the end of January, she had died from the disease. When the doctors were unable to explain why the cancer hadn’t been detected sooner, why it had progressed so quickly, and why it hadn’t responded to therapy, Sepich-Poore decided to go looking for answers himself. He taught himself bioinformatics and machine learning to interrogate how cancers evolve over time. He later began reading the literature on blood-based diagnostics, searching for new strategies in cancer screening and testing. And in 2016, he enrolled in medical school at UCSD, eager to learn about new advances in pancreatic cancer diagnosis and treatment.
Unfortunately, he found that despite the massive genomics revolution that had occurred in recent years, not much progress had been made for patients with pancreatic cancer. The five-year survival rate for the disease has lingered at around 10 or 11 percent for decades. “It made me wonder [if] perhaps there are markers or entities that are not yet being accounted for by those focused on host DNA or host RNA or host proteins,” but which could help catch cancer earlier, notes Sepich-Poore. Then he came across Straussman’s 2017 paper showing that not only did bacteria exist in the tumor microenvironment, they could influence the cancer’s response to therapy. He recalls being stunned.
Sepich-Poore suspected that no tumors were actually sterile, and that the microbes present could provide critical information. He had spent some time in Knight’s lab the summer before starting med school, and he landed back there for his PhD, spending his first six months interrogating whole genome sequences and transcriptomic data that was gathered as part of The Cancer Genome Atlas. His analysis covered the tissue and blood samples of more than 10,000 patients with more than 30 cancer types.
At first blush, the data were disappointing, Knight explains, as most of the variation the researchers found in microbial DNA could be attributed to where the study had been done. But Sepich-Poore applied statistical approaches he’d used to account for batch effects when analyzing human cancer data and “was able to get rid of all that noise and see a signal where the bacteria in a particular sample can give a readout of what kind of tumor you have and what stage,” Knight explains. Then, to scrub the data of potentially contaminating bacteria, Sepich-Poore used a series of existing statistical techniques that take into account, for example, the microbial sequences’ relative abundance across samples of different concentrations.
“What we found through this process, which involved running a supercomputer for six months straight, was really surprising,” Sepich-Poore says. “No cancer type was sterile; we were finding microbial DNA and RNA in every [cancer type]. . . . And on top of that, each cancer type has a unique microbiome.” Blood samples taken from cancer patients also had a distinct microbial signature based on cancer type, raising the possibility of minimally invasive diagnostics based on the cancerous tissue’s microbiome.
Seeing the potential to develop blood-based cancer tests, Knight, Sepich-Poore, and Sandrine Miller-Montgomery, then executive director of UCSD’s Center for Microbiome Innovation, applied for intellectual property to cover the approach and began the process of founding Micronoma. The company incorporated in the summer of 2019 and officially launched in August 2020 (enlisting Straussman and later Wargo to its scientific advisory board and giving them stock options). For its first product, Micronoma is focusing on a test for lung cancer, and collaborations are underway to develop diagnostics for other cancer types, says Miller-Montgomery, now CEO and president of the company. “The goal is to put this in the hands of clinicians and in the hands of patients.” In December, Sepich-Poore defended his dissertation and officially joined Micronoma full-time as its chief analytics officer.
Already, other labs have started to generate findings that support the idea that microbes could aid in cancer diagnosis. Last summer, for example, researchers reported that they’d identified microbial DNA signatures in the blood of melanoma patients that distinguished them from healthy controls. “We see that as a very important confirmation,” says Knight. It’s especially exciting because melanoma was the cancer type for which their diagnostic algorithms performed the worst, he adds—not terrible, but not as good as the rest. “We’re very encouraged that even in that worst case another group was independently able to see highly significant results.”
Microbial signatures may also contain prognostic information, potentially helping to predict treatment response. In Straussman’s 2020 study, for example, melanoma patients in the data set had different microbial signatures in their tumors if they did or did not respond to immunotherapy. “You can find bacteria that are enriched in responders or nonresponders,” Straussman says. And less than a year earlier, MD Anderson physician-scientist Florencia McAllister, along with Wargo and other colleagues, found that the tumor microbiome could indicate whether patients with pancreatic cancer had lived longer than five years following diagnosis with the disease. “The heterogeneity of microbes in the tumor was higher in the patients that live longer,” says McAllister, who has a patent pending based on the work. The million-dollar question now is what effect bacteria and possibly other microbes have on tumor development, progression, and response to treatment. “It’s one thing to descriptively say these bacteria are present in these tumor types,” says Straussman. “We need to learn more and more about how these bacteria affect the biology of the cancer.”
A growing body of literature suggests that bacteria and other microbes living in tumors or in the guts of cancer patients may influence their responses to
A wide-open field
No one really knows what microbes are doing in cancer, but recent results suggest it’s not nothing. In Sällberg Chen’s study of pancreatic cysts, for example, the researchers suspected the bacteria might be triggering inflammation, as they’d noted high levels of cytokines and other inflammatory markers in the cyst fluid. Sure enough, when the researchers put the bacteria they’d cultured from resected tumors into pancreatic cells in vitro, including both healthy cell lines and those with oncogenic mutations, they saw that some of the microbes caused rampant DNA damage and eventually cell death. “Some of them didn’t do so much really, but some of them were kind of vicious,” says Sällberg Chen. “Some of these bacteria that are hiding inside the pancreas are probably not very good. If they are hiding there for a long time, it will probably help [cancer] to develop.”
The million-dollar question now is what effect bacteria and possibly other microbes have on tumor development, progression, and response to treatment.
Researchers are also looking at how the immune system responds to bacteria in tumors. In collaboration with Wargo and others, for example, Straussman and his Weizmann colleague Yardena Samuels found T cells that presented peptides derived from microbes identified in the tumors. “This means that in a way our immune system really sees these bacteria,” says Straussman. Ajami, who was not involved in the study, agrees. “We now know they’re not only there, but they’re being detected and processed by the immune system.” These results point to a potential mechanism by which the microbiome could affect patients’ responses to therapy, Knight adds.
A handful of recent retrospective studies showing that taking antibiotics is associated with responses to both immunotherapies and to chemotherapies such as gemcitabine further support the importance of the microbiome in cancer outcomes, and the nature of the effect appears to vary by cancer type. “Antibiotics seem to have beneficial effects in patients with pancreatic cancer, whereas in melanoma, antibiotics seem to have a deleterious role,” Straussman says. Indeed, researchers studying tumor microbiomes are acutely aware that bacteria are not always bad. “With the microbiome it’s a yin-yang,” says McAllister. “It’s not going to be a one-size-fits-all.”
A limitation of these studies, however, is that the effects may not necessarily have to do with resultant changes in the tumor microbiome, as the gut microbiome will also be affected by antibiotic use. “The role of the tumor microbiome independently of the gut microbiome is a very important question,” says McAllister. “Most of the time when we are modulating the tumor microbiome, we are modulating the gut as well.”
Indeed, when she and her colleagues took bacteria from pancreatic cancer patients and gave them orally to antibiotic-treated mice, they changed the composition of the gut microbiome first and foremost. That said, the researchers did see notable shifts in the tumor microbiome as well, including increases in bacteria that were enriched in the patients’ tumors. “This data suggests that the gut microbiome can modulate the tumor microbiome, in minor part by direct translocation into the tumors, but more significantly, by altering the microbial landscape,” the authors write in their paper. (See “Where Does the Tumor Microbiome Come From?” below.) And the outcome was significant: animals who’d received microbiome transplants from long-term survivors had slower-growing tumors than did mice who’d gotten transplants from short-term survivors, suggesting a potential protective effect of the bacteria.
Two studies by separate groups published in the same issue of Science last year tested the concept of fecal transplants for modulating response to cancer therapy in humans. These trials showed that metastatic melanoma patients treated with fecal microbiota transplants from a treatment-responding donor had improved responses to the same checkpoint inhibitor immunotherapy the donor had received. Using fecal transplant to turn a nonresponder into a responder is “an amazing event,” says Knight. Although in this case the research focused on manipulating the gut microbiome, he adds, “it might be possible to modify what’s in the tumor by modifying what’s in the gut.”
Another possibility researchers are pursuing is the use of bacteria themselves as cancer therapeutics. Indeed, scientists have for years been working to engineer bacteria to seek and destroy cancer. “Now, the more we know about the bacteria [that] are found in tumors, probably we can use this as a new modality of treating cancer,” says Straussman.
To be able to fully harness the power of bacteria and other microbes to treat cancer, researchers need to better understand the relationship between the two. The list of outstanding questions goes on and on: How stable is the tumor microbiome? How does it interact with the gut microbiome? How have the bacteria in tumors evolved in the unique microenvironment? And that’s just the microbiology side; then there’s the cancer side. What role are the microbes playing in clonal evolution, metastasis, response to therapy, the immune landscape, and so on?
“And this is a short list. . . . Everything that has to do with tumor biology can now be reexamined,” says Straussman, who says he appreciates the collaborative spirit of the small but growing field. “There’s endless directions in which we can go.”
WHERE DOES THE TUMOR MICROBIOME COME FROM?
Among the many questions that remain about the cancer microbiome is where tumor-bound bacteria come from. One likely source is the abundant microbiome of the gastrointestinal tract. When MD Anderson physician-scientist Florencia McAllister and colleagues sequenced samples of patient tumors, they found that “surprisingly, around twenty percent of microbes in tumors was coming from the gut,” McAllister says. And the fact that they were absent in adjacent tissues suggested that “they were uniquely translocated to the tumors,” she adds. Indeed, when the team treated mice orally with fecal microbes from pancreatic cancer patients and then implanted mouse pancreatic cancer cells, some of those bacteria showed up in the animals’ subsequent tumors. McAllister and her colleagues also found plenty of microbes in the mice’s tumors that were in neither the gut nor the healthy tissue, however. McAllister says she suspects these are coming from the oral microbiome.
Karolinska Institutet clinical immunologist Margaret Sällberg Chen and colleagues have found evidence of just that in precancerous pancreatic cysts. “We were just surprised” to initially see bacteria known to be members of the oral microbiome, Sällberg Chen says, but the results held up in a larger cohort. Suspecting they had traveled through the blood to get there, the team looked in plasma from patients who had developed pancreatic cancer and found increased antibody levels against some of the oral microbes, including Fusobacterium, which had previously been associated with other types of cancer. “So that was seemed like confirmation that oral microbes might have some role,” she says.
One type of cancer Fusobacterium has been strongly linked to is colorectal, which is perhaps not surprising as the bacteria are common members of the gut microbiome in addition to the oral microbiome. But the Fusobacterium in colorectal tumors may not come from the nearby gut but rather from the bloodstream, having entered the circulation from the oral cavity. “Even though it’s close to the lumen of the colon, many of the bacteria may end up in the tumor from the bloodstream and not from the colon,” says Ravid Straussman of the Weizmann Institute of Science in Israel who was not involved in the study.
Regardless of where they originate, the bacteria found in tumors may then travel to distant sites in the body during metastasis. In December 2017, researchers published evidence that Fusobacterium and a suite of bacteria that it often co-occurs with are found in primary colorectal tumors and in metastases in the liver and other organs. “[W]e hypothesize that Fusobacterium travels with the primary tumor cells to distant sites, as part of metastatic tissue colonization,” the authors write.
adapted from Science, 371:eabc4552, 2021; Illustration by:© Natasha Mutch, SayoStudio