Does the Trojan Horse Have an Achilles' Heel?

A golden era followed Fleming's discovery of penicillin, during which many novel classes of antibacterial drugs were developed. Most of these antibiotics, such as streptomycin and tetracycline, are derivatives of natural compounds released by microbes (often fungi or soil bacteria) to kill competitors found nearby in the environment. Initially, candidate antibiotics were identified in simple inhibition or killing screens with the use of sensitive target bacteria grown on laboratory culture media. (These are referred to as in vitro–grown bacteria.)

After several decades of success, the yield of natural antibiotics began to dry up, and pharmaceutical companies turned to new, more sophisticated methods to improve their screens. These approaches required the use of ever more complex chemical libraries containing thousands of compounds in high throughput, sometimes automated, screens. Screening was often conducted against in vitro–grown bacteria, although in recent years, purified protein targets have also been used. However, even with these extremely sophisticated approaches, novel classes of antimicrobial drugs have proved largely elusive.

One limitation of such a screening process is that it targets bacteria or their subcellular components when they are growing in vitro. However, genetic analysis has shown that bacteria express only a portion of the genome when they are grown in the laboratory. Indeed, many of the genes that are essential for virulence in humans are expressed only in vivo during infection. Thus, screening that targeted proteins critical for bacteria grown in vitro could not snare proteins expressed only in vivo. The concept that in vivo–grown bacteria are different from those of the same isolate grown in vitro was developed by scientists such as Henry Smith in the 1950s.1 The idea of targeting genes and their products that contribute to infection but are not needed for life in vitro has been emerging for some years.

Rasko and colleagues2 have recently taken this idea through an experimental proof of concept. Many bacteria use protein-based signaling systems to interrogate their environment, and these systems can sense whether the bacteria have entered a potential host. Some of the sensors bind to specific small molecules in the environment and, when bound, send signals that alter bacterial gene expression. The molecular pathways associated with these sensors can activate virulence genes in vivo.

An example is the Qse signaling system found in many bacterial pathogens, including enterohemorrhagic Escherichia coli (EHEC). This system is driven by a membrane-associated histidine sensor kinase that responds to the presence of both adrenergic signals (for example, epinephrine or norepinephrine) from the host or signal molecules from other bacteria. Specific binding to the Qse protein causes autophosphorylation and transmission of signals through a second protein called QseB, which subsequently triggers increases in the levels of expression of key virulence genes.

Since it is difficult to get hold of large amounts of bacteria grown in vivo, the authors designed a clever screen that used gene reporters hooked up to virulence genes activated by the quorum-sensing E. coli regulator (Qse) system (Figure 1Figure 1Targeting Bacterial Virulence.). They then screened a chemical library and identified a compound, named LED209, which prevented the activation of the reporter and could inhibit the binding of norepinephrine to a receptor protein, called QseC. They were then able to assess the ability of LED209 to inhibit the productive attachment of EHEC to host cells and to prevent infection of rabbits by EHEC bacteria. At nanomolar concentrations, LED209 was effective in preventing productive cell binding but was not so efficient in preventing infection in rabbits. However, LED209 showed efficacy in two other in vivo models involving the infection of mice with either Salmonella typhimurium or Francisella tularensis. LED209 did not appear to interfere with adrenergic signaling in the host.

Clearly, these experiments are at the beginning of a long road to antibiotic development, but they do set several important precedents that are worth further consideration. Perhaps the most remarkable property of LED209 is that although it can attenuate the virulence of bacteria in a host, it does not inhibit the growth of bacteria on normal laboratory medium. This is an exciting property with many implications, and yet it poses challenges for industry. How will screens be set up for the routine testing of the antimicrobial activity of this molecule in bacteria for which genetic manipulation or even culture is a challenge? How will screens be set up for the emergence of resistance? Is there a danger that mutants will arise in situations in which virulence genes are actually turned on rather than off, and what would be the consequences of such an effect, if any?

No potential conflict of interest relevant to this article was reported.
Source Information

From the Wellcome Trust Sanger Institute, Genome Campus, Hinxton, Cambridge, United Kingdom.

DOWNLOAD PDF FULL

SHARECASH (U.S recommended)
ZIDDU (worldwide)

Mirror 1
Mirror 2