In 2000, Kim Lewis, a biologist and director of the Antimicrobial Discovery Center at Northeastern University, and his group noticed that when killing Pseudomonas aeruginosa biofilms with antibiotics, most of the cells were "readily killed at low concentrations…but a small subpopulation appeared invincible," as he wrote in a paper published in Microbe (1). They expected that this subpopulation would include resistant genetic mutants, but it wasn't so. In contrast to cells that are resistant to antibiotics, these persistent cells didn't show genetic mutations.
Then check out our virtual symposium, where Stanford bioengineer Stephen Quake will discuss his latest advances in cell sorting applications, with an eye on microfluidic-based cell sorting devices.
Originally described in the 1940s by Joseph Bigger, persister cells have been largely ignored by scientists until recently. But as scientists began studying biofilms and as more sophisticated methods to study populations of bacterial cells became available, interest has resurged. And now persistence is becoming recognized as a significant factor in chronic, recurrent infections associated with surgical implants, tuberculosis, and cystic fibrosis (CF).
Instead of showing genetics mutations, researchers believe that persister cells are phenotypic variants of normal bacterial cells. Research by Natalie Balaban, a biological physicist at The Hebrew University of Jerusalem, suggests this is generated via a stochastic switching event that happens by chance in individual cells in the population (2).
Persisters survive by going into a slow-growing, dormant state in the presence of antibiotics. When antibiotics are withdrawn, they repopulate the colony. Because antibiotics work on growing cells only, the dormant persisters escape their wrath. "Bactericidal antibiotics kill not simply by inhibiting functions, but by corrupting functions," explains Lewis.
For example, ciprofloxacin kills by inhibiting DNA gyrase, a topoisomerase which is necessary for proper prokaryotic DNA replication and is essential for cell growth. Other antibiotics cause the ribosome to produce dysfunctional peptides that become toxic. But persister cells suppress these normal biological activities, such as DNA replication and protein synthesis, when they go dormant, so the antibiotics become ineffective. And this is why persisters, well, persist.
So, does treatment with antibiotics produce fertile ground for selection of high levels of persisters? In other words, might repeated treatment with antibiotics yield more and more persister cells over time?
"That's what we've found so far," says Lewis, by studying P. aeruginosa isolates from patients with CF. After treating the isolates with high doses of ofloxacin, they saw a dramatic, 100-fold increase in persister levels (3). "It is a strain that evolves to produce a very large number of dormant cells that are tolerant to antibiotics. These cells survive, and as soon as the antibiotic level goes down, they grow. Then you get endless, chronic, relapsing infection."
Madness in the Method
Because they are a tiny percentage of the original bacterial population—typically as few as 1 in 10,000 to 1 in 100,000 cells—scientists must figure out new ways to isolate these persister cells. "We would really like to have single-cell studies," says Lewis. "If you could look at the proteome, the metabolome, on a single-cell level, that is what you would ultimately like to have."
Working toward that goal, Lewis' group has developed a florescence cell-sorting technique to separate Escherichia coli into normal cells and persisters (4). The group used a strain of the bacteria that has a degradable GFP gene downstream of a ribosomal promoter. Normally, these cells express a lot of GFP, glowing bright green. But dormant cells become dim because of their diminished protein synthesis. "This method allowed us to physically isolate persisters," says Lewis. "We isolated enough of them to get their transcriptome, which pointed in the direction of their formation."
That direction turned out to be toxin-antitoxin (TA) modules. These TA modules are pairs of genes in the chromosomes of most bacteria in which one gene acts as a toxin and the other gene cancels out its effect. "The toxin stops important cell functions, primarily targeting protein synthesis through a variety of mechanisms," explains Lewis. Some inhibit translation, while others stop cell wall synthesis or replication, for example. In normally growing cells, the toxin is kept inactive by the antitoxin, which binds tightly to it.
Kenn Gerdes, a cell biologist at Newcastle University in Newcastle, UK, and his group has been studying the toxin-antitoxin modules in E. coli. By deleting all the TA genes from the bacterium, the group found that the persistence level declined as the number of deleted TA genes increased (5).
Because there are so many TA genes, multiple mechanisms through which multidrug tolerance develops are likely. And researchers are hot on the pursuit of several of them already. Lewis reports that 15 TA modules have been identified in E. coli and 80 in Mycobacterium tuberculosis thus far.
He also writes that the stress response is a key part of persister formation. For instance, because ciprofloxacin causes DNA damage, it activates the canonical SOS stress response that leads to increased expression of DNA repair enzymes. At the same time, it activates a TA module called TisAB that overexpresses TisB toxin. TisB disrupts the membrane potential, lowering proton motive force and amount of ATP in the cells. As a result, the cells' systems shut down enough to cause dormancy.
To probe persister cells, microfluidics is becoming an important tool. In this method, bacteria are trapped in tiny spaces just a few microns wide in one layer of a microfluidic chip. On top of this layer is another layer with fluid channels used to flow antibiotics over the bacteria and then wash them away, allowing researchers to observe the reaction of each single bacterium to the antibiotics as they grow and divide through many generations.
"In those same cells, we can have reporter genes, and we can then see if they're turned on or off and determine if this gene expression coincides with persistence," says Gerdes.
Balaban’s group has developed an automated system—dubbed ScanLag—that uses standard office scanners arranged in an array to image the bacterial colonies growing on Petri dishes. The images are relayed to a software program, which analyzes the images and pulls out information related to the colonies' growth rate. Compared to observing single cells under the microscope and monitoring the length of time until they begin dividing, ScanLag allows for a far greater number of single cells to be monitored.
"This gives us a very nice measurement of how long single bacteria stay dormant. We were able to really characterize this dormancy and fit it into a more elaborate mathematical model of persistence," says Balaban.
They've also used ScanLag to map out how a toxin-antitoxin module can lead to partial dormancy in E. coli, because though all the bacterial cells have the same TA genes, only some of them end up going dormant (6).
With the goal of helping those who suffer from chronic infections, James Collins, a biomedical engineer at Boston University and a Howard Hughes Medical Institute investigator, is searching for ways to eradicate the responsible persister cells. So, he wondered if simple sugars like mannitol, fructose, and glucose could wake up the dormant persisters, nudging them into a growth cycle. Although the sugars had no effect on the killing efficacy of some commonly used antibiotics, in combination with gentamicin—a generic aminoglycoside antibiotic—they decreased the number of persistent bacteria by more than 1000-fold.
Using metabolic network analysis, Collins' group found that the sugars "were stimulating proton motive force, which is known to be needed for the uptake of aminoglycosides," he says. "We were able to validate that to show that persisters were not normally taking up aminoglycosides, but that with the sugars, they were." The approach was effective in both Gram-negative and Gram-positive bacteria. Taking their work a step further, they used a mouse model of E. coli urinary tract infection with a catheter implanted to show that co-treatment with gentamicin and mannitol eradicated the infection and reduced its spread to the bladder and kidneys (7).
Intrigued, Collins wanted to know if there were community aspects to persister cell formation. If so, this would be a different mechanism than the stochastic switching event that was suggested by Balaban’s group. Using microfluidic techniques, his team looked at the effects of indole—an endogenous signaling molecule—on persister cells. As a result, they found that the bacteria that took up more indole were more likely to become persisters (8). "There are bet-hedging strategies that appear to be functioning on a population level, where the bugs can communicate to each other via something like indole, and then depending on the amount of indole that a microbe sees, that will dictate whether it becomes a persister or not," he says.
Meanwhile, Lewis' group has been taking another approach to treating persister cells—using a prodrug. Prodrugs are molecules that kill a pathogen only once it gets inside the cell. "We realize that with persisters, you don't have a single target. So, you want to create a highly reactive molecule inside the bacterial cell—something that won't kill the host cell, of course." A prodrug will get into the bacterial cell, and is converted by bacteria-specific enzymes into a highly reactive molecule that then kills dormant cells just as easily as growing ones.
In the end, developing techniques to study the workings of a single cell will generate new information that hopefully will translate to the treatment of persistent chronic infections. “I think they'll become increasingly important in the study of persister cells, given that the phenomenon is still a very rare one,” says Collins. “But the ability to then study the responses of single cells will help us understand what's happening inside those cells, to either make them persisters or enable them to come out of the persister state."
1. Lewis K. 2010. Persister cells and the paradox of chronic infections.
Microbe 5: 429-437.
2. Balaban NQ, Merrin J, Chait R, Kowalik L, Leibler S. 2004. Bacterial persistence as a phenotypic switch. Science 305: 1622-1665.
3. Mulcahy LR, Burns JL, Lory S, Lewis K. 2010. Emergence of Pseudomonas aeruginosa strains producing high levels of persister cells in patients with cystic fibrosis. Journal of Bacteriology 192: 6191-6199.
4. Shah D, Zhang Z, Khodursky AB, Kaldalu N, Kurg K, Lewis K. Persisters: a distinct physiological state of E. coli. 2006. BMC Microbiology 6: 53.
5. Maisonneuve E, Shakespeare LJ, Jørgensen MG, Gerdes K. 2011. Bacterial persistence by RNA endonucleases. PNAS 108:13206–11.
6. Rotem E, Loinger A, Ronin I, Levin-Reisman I, Gabay C, Shoresh N, Biham O, Balaban N. 2010. Regulation of phenotypic variability by a threshold-based mechanism underlies bacterial persistence. PNAS 107: 12541-12546.
7. Allison KR, Brynildsen MP, Collins JJ. 2011. Metabolite-enabled eradication of bacterial persisters by aminoglycosides. Nature 473: 216-220.
8. Vega NM, Allison KR, Khalil AS, Collins JJ. 2012. Signaling-mediated bacterial persister formation. Nature Chemical Biology 8: 431-433.