Mafia Wars

An increasing amount of data is showing that the cellular battle between pathogens and hosts needs much more than a simple military metaphor to describe it—think undercover infiltration, front organizations, and forced suicide.

Jun 1, 2010
Jef Akst
©Time & Life Pictures / Getty Images / George Silk

When searching for an appropriate description of the mammalian immune system, the vast majority of scientists settle on the metaphor of a war. It’s a battle scene, with the foot soldiers of the immune system (e.g., killer T cells) battling the bacterial or viral particles in an open field (the host’s body). Painting a picture in such strong terms is a good way to attract attention (and funding), and in many ways, it is a good fit—one paper I stumbled on as a graduate student that elegantly modeled the conflict generated nearly identical equations to those used in traditional models of warfare, which predict that military losses are proportional to the size of enemy forces.

Over the last several years, however, scientists have begun to realize that the molecular interactions between a pathogen and its host are quite a bit more complex than simple open field battle, where the power of one’s army is measured by bodies alone. The immune system is a multifaceted defense system, and pathogens have evolved numerous molecular strategies to evade its wrath, including methods that resemble more the devious tactics of organized crime than those of traditional warfare, such as setting up fronts to conduct covert operations, going undercover to infiltrate the opposing gang, and terrifying the enemy into admitting defeat (or committing Seppuku).

In the late 1990s, the discovery of pathogenicity islands—large regions of bacterial and viral genomes unique to pathogenic species—led researchers to recognize that many pathogens were involved in some complex racketeering, says immunologist and microbiologist Igor Brodsky of Yale University. Encoding specialized systems to inject virulence proteins into cells, pathogens are able to manipulate cellular processes in the host for their own benefit, such as initiating immune cell death and blocking a continued immune response.

This prompted the field to start identifying specific virulence factors important for a particular pathogen—either by mining databases for genes that might be behind those factors, knocking them out, and observing the effects on the pathogen’s ability to infect its host, or doing random mutagenesis. Once scientists identified some factors important for virulence, “then the question was what were these genes or proteins doing,” Brodsky says.

Pathogens have evolved molecular strategies to escape the immune system that resemble more the tactics of organized crime than traditional warfare.

In the last 4 or 5 years, Brodsky says, researchers have begun to answer that question thanks to a growing interest in the field, as well as new tools and reagents. “People have made a lot of progress in figuring out what these genes do and how they target host cell biology.” They’ve found, for instance, that bacteria employ a variety of techniques to infiltrate and assassinate undetected, such as innovative camouflages and coaxing hosts to abandon their fight.

“There certainly has been a rapid pace of discovery in the area of the host–pathogen interaction,” agrees cell and molecular biologist John Reed of the Sanford-Burnham Institute for Medical Research in La Jolla, Calif. Like many fields in the life sciences, he adds, the quickly advancing field of “genomics is a big part of that,” as well as the “development of algorithms that allow us to search for hypothetical candidate genes [and] other research technologies [that] keep getting more and more powerful.”

Researchers say that the study of host–pathogen interactions has also benefited from the coming together of two historically distinct disciplines—immunology and microbiology—allowing the two fields to feed off each other in a way that further accelerates the science. “Initially they were sort of kept separate—you had the bacterial pathogenesis folks and then you had the immunologists,” says leukocyte biologist Scott Kobayashi of NIAID’s Rocky Mountain Labs. “I think now really those two areas are starting to merge, [and] more people are focusing in on bridging that gap.”

1. Set up a front organization

©Getty Images / New York Times Co.

Whether it’s “waste management” or a massage parlor, the first step in avoiding capture is to find a way to conduct business undetected. Indeed, while organisms have evolved numerous cellular sensors to identify an infection—which, when bound to a pathogen, elicit a profound immune response that eliminates most invading microorganisms—many pathogens have evolved to escape such detectors.

One of the most prevalent classes of tools that hosts use to spot pathogens is the Toll-like receptor (TLR) family. The TLR pathway is “a, if not the, central [pathogen] recognition system we have,” says Thomas Miethke of the Technische Universität München. TLR4, for example, specializes in recognizing a key component of the outer membrane of most Gram-negative bacteria known as lipopolysaccharide (LPS), while TLR5 senses a conserved region of flagellin, the main structural protein in bacterial flagella.

Some bacteria, however, have mastered the art of disguise. Helicobacter pylori, for instance, has evolved ways to be essentially invisible to both these TLR pathways, despite the fact that it is a Gram-

negative bacterium with four to six flagella. To hide from TLR4, H. pylori’s lipid A—the part of the LPS protein that anchors it to the outer part of the bacterial envelope and the domain sensed by TLR4—has taken on a slightly different structure. While TLR4 recognizes lipid A molecules with six acyl chains attached, each 12 to16 carbons in length, the H. pylori lipid A has only five acyl groups, each with up to 6 more carbons. Similarly, amino acid differences in the terminal domain of H. pylori’s flagellin protein allow the bacteria to evade detection by TLR5.1

Other pathogens take a different approach to camouflaging their presence: Rather than disguise the proteins that the host is on the lookout for, some bacteria species have chosen to eliminate them altogether. For example, Anaplasma phagocytophilum—which causes a tick-borne disease known as human granulocytic anaplasmosis—has lost all genes for the biosynthesis of LPS and most genes for the biosynthesis of peptidoglycan, another component of the bacterial cell wall that hosts recognize as a sign of infection. As a result, these pathogens do not trigger an effective innate immune response.

“By removing these LPS and peptidoglycan [proteins], the macrophages and neutrophils cannot recognize these bacteria as a pathogen, so they can enter into [the cells] like stealth pathogens,” says molecular microbiologist Yasuko Rikihisa of the College of Veterinary Medicine at The Ohio State University. “Without being recognized, they can enter, survive, and replicate, and kill the leukocytes eventually.”

The lack of these key cell membrane components, however, is not without consequences, Rikihisa adds. Without LPS and peptidoglycan, “their membrane structure becomes more fragile,” she says. To compensate, the bacteria take up cholesterol from host cells, which stabilizes their membranes. In fact, Rikihisa and her colleagues found that A. phagocytophilum infection rates were 10 times higher in mice fed a high-cholesterol diet than in those on a normal diet.2 “So the cholesterol is helping [the pathogen],” she says.

“Initially, [immunology and microbiology] were sort of kept separate. I think now really those two areas are starting to merge.” —Scott Kobayashi

Further experimentation by Rikihisa and colleagues last year revealed that A. phagocytophilum actually coaxes the host cell to take up more low-density lipoprotein (the so-called “bad”) cholesterol by increasing the expression of the LDL receptor on infected cells, thereby gaining access to the cholesterol it needs to promote infection.3 This represents yet “another subversion mechanism of this bacterium,” Rikihisa says—one that compensates for its initial effort to avoid detection in the first place.

2. Go undercover

©Time & Life Pictures / Getty Images / David Lees

If the host successfully detects the presence of a pathogen, it initiates signaling cascades that kick off a variety of immune responses. Eight years ago, in the course of his studies on the Toll/interleukin-
1 receptor (TIR) domain and its role in TLR immune signaling in response to infection, Miethke started searching the National Center for Biotechnology Information database for additional eukaryotic TIR-containing proteins. What he found gave him a bit of a shock. Instead of finding eukaryotic proteins with TIR homology, “I found, to my surprise, bacterial proteins which also have a TIR domain,” he recalls. “When you see that, fantasy immediately starts up. If they have this TIR domain, maybe they interfere with this main signaling cascade,” he says. Because several components of the host TLR signaling pathway interact at their TIR domains, bacteria encoding similar domains may be able to divert this entire cascade by binding to and blocking those components.

Before Miethke had much time to investigate these bacterial proteins that contain TIR domains (known as bacterial TIR-domain-containing proteins, or Tcps), a paper emerged from Reed’s lab at the Burnham Institute that confirmed Miethke’s suspicion.4 A Tcp dubbed TlpA from Salmonella enteritidis blocked NF-kB—a downstream target of the TLR signaling pathway that triggers the expression of cytokines and recruits white blood cells to the site of infection.

Reed and his collaborators had also scoured the databases, and identified more than 200 bacterial TIR homologs. Choosing one to start with, the team cloned and overexpressed TlpA in mammalian cells and demonstrated that it likely blocks NF-kB activation by somehow interfering with a TIR-containing protein called myeloid differentiation factor 88 (MyD88), used by most TLRs to activate NF-kB.

E. coli's protein TcpC interferes with MyD88, blocking the activation of downstream signaling factors in the TLR pathway, including NF -kB. NF -kB thus fails to enter the nucleus to trigger a pro-inflammatory response, including the expression of cytokines and recruitment of white blood cells to the site of infection.

Like mobsters who send their subordinates undercover to infiltrate enemy gangs, pathogens with such TIR-containing proteins can thus evade the disciplinary actions of the host. Using this “molecular mimicry strategy,” Reed says, the bacterium “is able to go in more stealth mode and hang out in the cell,” allowing it to “replicate intracellularly [and] get a foothold before the immune system” knocks it down.

The finding was initially met with “a fair amount of skepticism,” Reed admits, which is why he was “delighted” when Miethke published the results of his first experiments on Tcps 2 years later.5 Using a similar strategy of cloning the TIR-domain-containing genes, expressing them in host cells, and looking for any biological defects they might cause, Miethke confirmed that two more Tcps—TcpC in Escherichia coli and TcpB in Brucella melitensis—interfered with MyD88. Miethke further showed that TcpC bound to MyD88, suggesting it directly inhibits the host protein—and the entire pathway it activates.

Another growing example of pathogen mimicry comes from viruses, which target the host ubiquitin system. Once thought to be little more than a tag for protein degradation, in the last 5 or 6 years, “it’s become very, very clear that ubiquitylation plays a major role in controlling activity of proteins,” including those involved in immunity, says virologist Adolfo García-Sastre of the Mount Sinai School of Medicine. “Once we know that, then it doesn’t become so surprising that pathogens are intersecting with the ubiquitin and ubiquitin-like pathways for their own benefit.”

One example comes from the “congo” virus, the devastating cause of Crimean-Congo hemorrhagic fever. Searching its sequence for homologous domains, García-Sastre and his colleagues found a couple of interest. One was a viral protein domain that belongs to the family of ovarian tumor (OTU)–like proteases. In mammals, OTU proteases dissociate ubiquitin from its specific target proteins. The reverse process—the addition of ubiquitin or ubiquitin-like molecules, such as interferon-stimulated gene product (ISG15), to target proteins—can initiate antiviral responses, leading García-Sastre to suspect the viral OTU-like proteases may be messing with the normal host immune response.

“Having in mind that ubiquitylation and ISGylation are required for innate immunity, we decided to test whether this domain from ‘congo’ will deconjugate ubiquitin and ubiquitin-like molecules and disarm host innate responses,” García-Sastre explains. Sure enough, expression of the congo virus OTU-like proteases actively detached ISG15 and ubiquitin from their target proteins, and by doing so, prevented two different antiviral pathways.6

This deubiquitylating strategy “most likely represents one of the ways viruses can regulate ubiquitylation inside cells,” García-Sastre says, but it is not the only one. Some viruses encode their own ubiquitin and ubiquitin-like proteins; others encode their own E3 ligases—necessary enzymes for ubiquitin conjugation—and still others encode adaptor proteins that recruit host E3 ligases. Just last year, García-Sastre and his colleagues published one of the first examples of a viral protein that inhibits an E3 ligase to prevent antiviral responses.7 The targeting of the ubiquitin system by viruses to evade the host immune response is turning out “to be a common theme,” García-Sastre says.

3. Forced suicide

©Popperfoto / Getty Images

Another way to survive is to simply destroy an opponent, which is where Tommy guns and Molotov cocktails come in. Many pathogens do this, of course, killing the host immune cells. But some bacteria, such as the Yersinia family, which includes the causative agent of the plague, have a hard balance to strike—kill enough macrophages to prevent the normal immune response, but not too many that it signals the host to mount an inflammatory immune response. “It’s a little paradoxical in a sense that in some infections, [cell death] is protective, and in some it’s pathogenic,” says molecular biologist Jim Bliska of Stony Brook University in New York. Yersinia’s solution: induce the macrophages to quietly kill themselves.

Yersinia bacteria achieve their mission by secreting factors known as Yersinia outer proteins (Yops) into macrophages to inhibit two key immune signaling pathways—both involved in the suppression of apoptosis following infection. These Yops (YopJ in Y. pestis and Y. pseudotuberculosis; YopP in Y. enterocolitica) block the activation of two kinases (MAPKK and IkB kinase), preventing the action of MAPK and NF-kB and the subsequent expression of genes that suppress cell death, causing the macrophages to undergo apoptosis. Thus, the Yersinia bacterium successfully kills the immune cell without allowing it to tip off the rest of the body that there has been an invasion, preventing the specific immune action of the macrophage (present antigens) as well as a larger immune response.

Yersinia bacteria induce macrophages to undergo apoptosis by inhibiting two key immune signaling pathways, MAPK and NF-kB. By secreting factors known as Yersinia outer proteins (YopJ/P) into the cell, Yersinia binds to and blocks the activation of MAPK kinases and IkB kinase, preventing the action of MAPK and NF-kB and the subsequent expression of genes that suppress apoptosis. With this survival pathway blocked, the cells quietly submit to the infection, and fail to present antigens or initiate a larger immune response.

Although the exact mechanism by which Yops cause apoptosis remains “controversial,” Bliska says, the Yops may alter the host’s proteins, which prevents the activation of antiapoptotic genes. On MAPKK, for example, the Yops may add an acetyl group to the part of the protein normally activated by phosphorylation, thus preventing phosphorylation (and activation). Alternatively, YopJ may induce apoptosis by acting on the ubiquitin system.

By inhibiting these pathways and thus the expression of host antiapoptotic genes, Yersinia effectively stimulates the macrophages to peacefully surrender: They undergo apoptosis without inducing an inflammatory response often caused by other types of cell death. At the same time, inhibiting the function of these immune cells by killing them off essentially “handcuffs the macrophages’ ability to present antigens and also to direct the immune response as well,” Kobayashi says.

Interestingly, while Yersinia also induces low levels of apoptosis in other types of immune cells, such as dendritic cells, this February Kobayashi and his colleagues found that the bacteria have the opposite effect in neutrophils—the most abundant type of white blood cells in mammals.8 Instead of inducing apoptosis, the bacteria postpone normal neutrophil death by inhibiting the production of reactive oxygen species that normally cause rapid cell death following infection. Under normal infection conditions, these apoptotic neutrophils are then digested by macrophages, thereby eliminating the pathogen.

The benefit to Yersinia of preventing neutrophil cell death is still a bit unclear, Kobayashi says, but “the assumption is that anything that survives within these phagocytes is probably helping to promote pathogenesis.” Thus, prolonging the lifespan of the neutrophil “buys the bug time to replicate in the neutrophil and survive,” he says. “There [are] many different ways a cell can die,” Kobayashi adds, “and I think people are starting to recognize that pathogens may differently alter that course of death.”

Tricky Tactic: Target the head honchos or the subordinate hit men

Pathogens have many different effects on and interactions with their hosts. While some, such as Bacillus anthracis, cause acute and often lethal host disease, others, such as Mycobacterium tuberculosis, can maintain long-term persistent infections. As a result, acute pathogens tend to focus on replication and transmission with no regard for their host’s health, while chronic pathogens depend on the survival of their hosts for their own continued existence.

The nature of these different approaches to life came up in a meeting of the Ruslan Medzhitov lab at Yale University. “Very acute severe pathogens like anthrax don’t have a requirement for long-term interactions with the host, and the kinds of things they target are going to be extremely disruptive to host physiology,” says immunologist and microbiologist Igor Brodsky, a postdoc in the lab. “In contrast, persistent pathogens can’t afford to do that because they would eliminate their own biological niche. The nature of the requirements for an acute infection differ fundamentally [from] chronic or persistent infection.”

This discussion led the group to an idea: Perhaps acute pathogens target more central players in the host immune system, often with devastating consequences to the host, while chronic or persistent pathogens merely pick away at the peripheral components to avoid being eliminated from the host while causing minimal disease.

“The idea of nodes and hubs in signaling networks and biological networks is not a new idea,” says Brodsky, who wrote a review with Medzhitov on the topic last year,9 but for how pathogens interact with their hosts, “it’s a relatively new concept.”

The concept makes sense, researchers agree, but the evidence for it is still limited. “The idea is certainly interesting and a lot of the aspects are true,” says molecular biologist Jim Bliska of Stony Brook University in New York. “Obviously more work has to be done to validate this model, and there’s probably going to be examples of pathogens that cause chronic infections that do target central hubs, and certainly pathogens that cause acute infections will also affect the peripheral networks.”

In addition, many chronic pathogens can alternate between a latent stage, with little or no effect on host health, to an active infection, says cell and molecular biologist John Reed of the Burnham Institute for Medical Research in La Jolla, Calif., in which case the same pathogen might target different aspects of the host immune system at different points in its life cycle. “We don’t know a lot yet about the circumstances that would tell bacteria to turn on or off certain genes,” Reed says. The host–pathogens interactions “might have a big role for when they are expressed and how that manifests in the context of clinical disease.”


1. E. Andersen-Nissen et al., “Evasion of Toll-like receptor 5 by flagellated bacteria,” Proc Natl Acad Sci, 102:9247–52, 2005.
2. Q. Xiong et al., “High-cholesterol diet facilitates Anaplasma phagocytophilum infection and up-regulates macrophage inflammatory protein-2 and CXCR2 expression in apolipoprotein E-deficient mice,” J Infect Dis, 195:1497–1503, 2007.
3. Q. Xiong et al., “Cholesterol-dependent Anaplasma phagocytophilum exploits the low-density lipoprotein uptake pathway,” PLoS Pathog 5:e1000329, 2009.
4. R.M. Newman et al., “Identification and characterization of a novel bacterial virulence factor that shares homology with mammalian Toll/interleukin-1 receptor family proteins,” Infect and Immun, 74:594–601, 2006.
5. C. Cirl et al., “Subversion of Toll-like receptor signaling by a unique family of bacterial Toll/interleukin-1 receptor domain-containing proteins,” Nat Med, 14:399–406, 2008.
6. N. Frias-Staheli et al., “Ovarian tumor domain-containing viral proteases evade ubiquitin- and ISG15-dependent innate immune responses,” Cell Host & Microbe, 2:404–16, 2007.
7. M.U. Gack et al., “Influenza A virus NS1 targets the ubiquitin ligase TRIM25 to evade recognition by the host viral RNA sensor RIG-I,” Cell Host & Microbe, 5:439–49, 2009.
8. J.L. Spinner et al., “Neutrophils are resistant to Yersinia YopJ/P-induced apoptosis and are protected from ROS-mediated cell death by the type III secretion system,” PLoS One, 5:e9279, 2010.
9. I.E. Brodsky, R. Medzhitov, “Targeting of immune signalling networks by bacterial pathogens,” Nat Cell Bio, 11:521–26, 2009.