Microbial Cocoon Helps Cells Penetrate, Shrink Tumors

A tour-de-force in bacterial engineering for cancer therapy.

Niko McCarty
6 min readApr 1, 2022


Subscribe for free at codonmag.com.

Case Study

Sometime in 1913, a patient walked through the doors of New York Hospital, the second oldest hospital in the United States, and met with Dr. William Coley, a bone surgeon.

The patient’s concern was obvious: a large tumor — specifically, a sarcoma, a cancer that originates in bone or connective tissue — had formed in his neck. This particular tumor was so deep, so embedded, that Coley was forced to leave it. A poor prognosis, to be sure.

Shortly after this meeting, though, the patient had two bouts of an infection, called erysipelas, on the skin of his neck. The tumor vanished.

“During these attacks of erysipelas, the tumor of the neck entirely disappeared and the patient left the hospital in good health,” wrote Coley in 1920. “I tracked down the patient and found him alive and well seven years later.”

Bacterial infections had, presumably, caused the tumor to shrink and disappear. And even before Coley’s discovery, in the 1880s, scientists had found bacteria growing inside of tumors. Today, researchers have cataloged bacterial strains and bacterial communities that are associated with various types of tumors.

In the last few years, efforts to harness bacteria as cancer therapies have ramped up, built on the foundation first laid by Coley. Synthetic biology tools, like CRISPR and protein engineering, have helped scientists to build bacteria that can travel through the body, targeting and shrinking tumors with almost pinpoint precision. A new study from researchers at Columbia University has advanced these efforts further still.

A genetic circuit, a biological switch, was created to help E. coli bacteria turn on or off a protective shell. When the shell is switched on, the cells evade the immune system and travel to a tumor. When the shell is off, the cells are vulnerable; white blood cells quickly dispose them. A small molecule, called IPTG, acts as the remote control.

The paper, published in Nature Biotechnology last week, was led by co-first authors, Tetsuhiro Harimoto and Jaeseung Hahn. The paper was a collaboration between the Kam Leong and Tal Danino laboratories.

Transmission electron microscope images show a sugar coat, or cocoon, wrapped around a cell’s outer surface. Deleting a gene called kfiC causes the shell to disappear. Harimoto, T. et al. A programmable encapsulation system improves delivery of therapeutic bacteria in mice. Nat Biotechnol (2022).

Guilty by Association

Microbes in the body probably seek out tumors via chemotaxis. They sense and squirm and inch their way towards cancer cells by following a turbulent stream of molecules. In the body, tumors are also surrounded by a low oxygen environment that many anaerobes (like Clostridium) find ideal.

The Danino lab has a proven track record of engineering microbes to travel to, and colonize, tumors. For a 2020 paper in Science Translational Medicine, his group created an engineered, probiotic strain that homed in on tumors and, once there, released nanobodies targeting cell death-ligand 1 and T lymphocyte-associated protein-4, two proteins that help tumor cells evade the immune system and proliferate.

There are two main challenges with developing a bacterial, cancer-killing system, though: Bacteria need to be potent enough to shrink tumors, but delicate enough so that they die off shortly after completing their mission. That’s because bacteria in the body, left unchecked, can cause sepsis and death.

Clinical trials have been largely unsuccessful, in part, because they miss the mark on one of these two necessities.

For a phase I trial in 24 patients with metastatic melanoma, for instance, a “safer” form of Salmonella typhimurium colonized tumors, but didn’t shrink them. In another trial, patients with advanced, solid tumors were injected with Clostridium novyi-NT, a nontoxic microbe. Out of 24 participants, 10 saw their tumors shrink; toxicities were “significant,” according to the report.

There is a balance, then, between safe and effective, too much and too little. This new study walks the tightrope.

Finding a Target

Bacteria inside the human body evade detection by shielding themselves in a sugar cloak. This has been known, at least by microbiologists, for at least a decade. Sugar shells in some bacteria are structurally similar to sugars found in mammalian cells, and so the human body cannot recognize and destroy them. This observation was a key source of inspiration for the new paper.

Blood samples were first collected from a small group of people (I’m assuming the researchers themselves. When I interviewed for a PhD at Columbia with Tal Danino, a few students mentioned that they had donated blood for a research project, but I can’t recall for certain).

With blood in hand, the researchers mixed in different bacterial strains, including E. coli and Salmonella typhimurium, a human pathogen that can cause gastroenteritis and trigger diarrhea, cramps and vomiting. A type of E. coli called Nissle 1917 survived the longest in blood, they found. When injected into mice, Nissle also caused a minimal immune response.

To figure out which genes coordinate sugar shell production in Nissle, small RNAs were used to knock down the expression of different genes. Knocking out a gene called kfiC caused bacterial cells, in blood, to be swiftly destroyed by white blood cells. That gene, then, appears to play a defensive role, helping cells to evade the host’s immune system.

Controlled Cocoon

A genetic circuit was next constructed that could reversibly turn kfiC on or off. Switching off kfiC caused bacteria to drop their protective sugar shells, while switching it on caused them to don their ‘invisibility cloak.’

The circuit is simple: In the cells’ natural state (no shell), the production of kfiC is blocked by the transcriptional repressor, lacI.

But when IPTG is added to the cells, this small molecule passes through their membranes and binds to the lac repressor, blocking its function. The kfiC gene switches on and makes the sugar shell (called capsular polysaccharide, or CAP, for short).

a, Circuit diagram of the “sugar coat” on-off switch. A small molecule, IPTG, blocks lacI expression, which turns on a gene that helps produce CAP. b, Turning on the ‘cloak’ with more IPTG results in a thicker sugar shell. c, Circuit dynamics for ‘on’ and ‘off.’ The cells are fully surrounded by sugar after about six hours, and the shell disappears in a similar amount of time. Harimoto, T. et al. Nat Biotechnol (2022).

It takes about six hours, after introducing IPTG, for the cells to fully enmesh themselves in the cocoon. Removing IPTG causes the sugar shell to disappear at a similar pace.

Bacterial cells with this genetic circuit were, once again, mixed with human blood. Adding IPTG, and switching on the protective coat, increased their survival by about 100,000-fold.

Wild-type Nissle, without the protective shell, can survive in human blood for more than 6 hours. A microbe with depleted kfiC disappears within 30 minutes. Bacteria with the sugar cocoon were destroyed by white blood cells about 10-fold less than cells without that protection.[1]

Cancer Payloads

The goal, in this study, was two-fold: Reduce the immune response caused by injecting bacteria into mice (thus increasing the amount of therapy that can be delivered to tumors) and then, somehow, get rid of the bacteria.

(Left) Bacteria with the sugar shell have reduced immunogenicity in cancer mouse models. The sugar shells reduce initial inflammation while effectively clearing bacteria over time. (Right) Turning on the CAP system causes cells at one tumor to translocate to another tumor. Harimoto, T. et al. Nat Biotechnol (2022).

In creating their genetic circuit, the researchers increased the maximum tolerable dose — the total number of bacteria that can be delivered to a tumor — more than 10-fold, relative to other strains, in a cancer mouse model.

For one experiment, the researchers expanded upon their genetic circuit so that bacteria would release an antitumor toxin, called theta, only after reaching a target tumor. The researchers stimulated the microbes with IPTG, and injected them into two mouse models for both colorectal and breast cancers. In both cases, the tumors shrank, without triggering a strong immune response. And once inside the mouse — in the absence of IPTG — the cells soon lost their protective coats, and were eradicated by white blood cells.

This genetic circuit is akin to a reversible invisibility cloak. Cells can switch this cloak on and off, using IPTG as a remote. The result is a bacterial cancer therapy that, at least in mice, has improved safety and efficacy.

Until next time,

— Niko // @NikoMcCarty

  1. This experiment, in my opinion, suggests a secondary application for bacterial cocoons. Under normal circumstances, bacteria are swiftly killed by white blood cells. With a protective coat, bacteria can survive for several hours. Could protected bacteria, then, be used to metabolize and remove contaminants in human blood samples, say, before a transfusion? Staphylococcus bacteria are the most common contaminants in blood products, according to the CDC. Perhaps bacteria could be engineered to don their cocoon, destroy the Staphylococcus, and then remove their shell. This might help to keep valuable blood donations from going to waste.



Niko McCarty
Editor for

Science journalism at NYU. Previously Caltech, Imperial College. #SynBio newsletter: https://synbio.substack.com Web: https://nikomccarty.com