IN THE WAR against disease-causing bacteria, our armoury is becoming dangerously depleted. Hospitals, for example, harbour strains of Staphylococcus aureus that can be killed only by vancomycin, one of the oldest antibiotics. When these bugs finally develop resistance to this stalwart of the antibiotic armamentarium, deadly S. aureus infections may become totally untreatable.
Driven by widespread fears that a bacterium will soon appear that is resistant to all existing antibiotics, scientists around the world have been searching for the next generation of antibiotics that will help us keep the upper hand. Mostly, this is a hit-and-miss business that involves screening thousands of new compounds as they emerge from the chemists鈥� labs.
But Peter Nielsen and his fellow scientists at a Danish biotech start-up called Pantheco have a more systematic plan. Their ambitious strategy is to construct molecules that resemble short strands of DNA, which can be used to lock onto the genes of invading bacteria and kill them. Such designer drugs, Nielsen imagines, could also be targeted at specific harmful genes in human DNA, and block them to tackle intractable diseases such as cancer.
Advertisement
The researchers are pinning their hopes on an odd class of molecules known as peptide nucleic acids or PNAs-funky hybrids that were first synthesised by Nielsen and his colleagues Michael Egholm, Rolf Berg and Ole Buchardt in 1991. The molecule鈥檚 secret lies in its similarity to DNA and RNA. Radiating like tiny spokes from a strong but flexible backbone are the same bases that form the genetic code, held in much the same way as they would be in a strand of DNA or RNA. But the difference is in the backbone. Instead of being made of alternating sugar and phosphate units, as in DNA and RNA, the backbone of a PNA is more like that of a protein. It鈥檚 formed from a chain of modified glycine units, the simplest amino acid found in proteins.
The dream is that PNA antibiotics could be tailor-made to disrupt the function of one specific gene of a pathogen, and prevent it doing its job. Such a genetic medicine could offer huge benefits. Being so precisely targeted, it would not have the unwanted side effects common to many of today鈥檚 antimicrobial drugs. Even better, by adjusting the sequence of bases on the PNA, researchers would be able to whip up a new antibiotic whenever a bacterium develops resistance to the existing one.
Like so many discoveries, PNA鈥檚 potential as a revolutionary new medicine came to light while Nielsen and his colleagues were looking for something else entirely. The researchers, then at the University of Copenhagen and the nearby Ris酶 National Laboratory, were trying to build a molecule that would recognise specific short sequences of DNA. Such probes are used to identify particular genes and help figure out what they do.
At first glance, making a gene probe ought to be simplicity itself: just make a molecule that matches the nucleotide sequence along one of DNA鈥檚 two intertwined strands. That is, after all, how DNA itself forms its trademark double helix. The bases on one DNA strand match up with their complements on the other strand-thymine pairing with adenine, and cytosine pairing with guanine. These base pairs then form the rungs that run down the centre of the spiralling ladder of DNA.
But there is a problem. When DNA is not busy directing the production of proteins, its double helix stays zipped up tight, making the bases hard to reach. Only when DNA鈥檚 message needs decoding does the helix unwind, allowing the sequence of bases to be transcribed into messenger RNA (mRNA). The mRNA carries the instructions for generating proteins from the nucleus (where DNA is stored) into the cell cytoplasm, where ribosomes translate the mRNA code into a protein.
To generate a probe that would recognise a specific DNA sequence, Nielsen and his colleagues needed a molecule that could insinuate itself into the tightly coiled double helix and so reach the bases of the gene of interest. In other words, they needed to synthesise a DNA mimic capable of binding to the DNA double helix, forming a molecular triplex.
Fortunately, inspiration was to hand. About ten years earlier, Peter Dervan of the California Institute of Technology and Claude Helene of the National Museum of Natural History in Paris had shown that DNA itself can be encouraged to form such a triple helix-albeit a weak one. The interactions in the DNA triplex are not strong because repulsion between the negative charges carried by the sugar-phosphate backbones of the three strands tends to push the strands apart. To eliminate this electrostatic repulsion, Nielsen and his colleagues decided to design a DNA-like polymer that had a neutral backbone. Thus PNAs were born.
What happened next was unexpected. PNAs did indeed form a triplex with DNA. But instead of sedately aligning itself along the existing double helix, the PNA tore into the helix and formed a stable triplex that contained two strands of PNA and one of DNA. This molecular threesome, it turned out, was much more stable than those achieved with any other synthetic molecules designed to mimic DNA, such as DNAs in which one of the oxygen molecules in the sugar-phosphate backbone is replaced with sulphur. The researchers also found that PNAs formed a stable duplex with single-stranded mRNA. So, in PNAs, Nielsen and his colleagues had what they wanted: a molecular probe-and, they subsequently realised, an agent with potential as a genetic medicine.
Resistance is futile
The ability of PNAs to bind to both double-stranded DNA and single-stranded RNA means that they ought to be able to disrupt the activity of a gene in a number of ways. By binding directly to DNA, a PNA could prevent DNA containing a particular gene being transcribed into mRNA-a tactic called antigene therapy. Alternatively, PNAs that interact with mRNA could block the production of proteins. This strategy is called antisense therapy, because mRNAs are essentially copies of the gene-containing or 鈥渟ense鈥� strand of the DNA double helix. As a third option, PNAs can be targeted to bind to the RNAs that form the ribosome-the cell鈥檚 protein-making factories. Such PNAs could jam the ribosomal machinery, effectively shutting down protein production in the cell. Antisense technology is currently the most advanced of these strategies.
In theory, antisense or antigene therapies could be developed to treat pretty much any kind of disease. In the case of an infection, the drug might target a gene essential for the pathogen鈥檚 reproduction, thus stopping the infection in its tracks. To treat a human disease, the drug would be targeted to shut down the disease-causing gene. For example, tumours might be reduced by shutting down the oncogenes that allow the runaway growth of their cells.
What makes PNAs so attractive as potential genetic medicines is that they bind strongly-and specifically-to their designated targets, whether DNA or RNA. In addition, PNAs are quite stable inside cells. Their peptide-like backbone makes them immune to attack by the enzymes that normally chew up nucleic acids. And the enzymes that digest cellular proteins don鈥檛 recognise PNAs either. Because PNAs consist of chains of modified amino acids that do not occur naturally in proteins, they are immune from attack.
Best of all, it should be easier to rejig PNAs as needed to thwart microbial resistance. Many of today鈥檚 antibiotics work by attacking the proteins that bacteria use to build their cell walls. One way for a bacterium to develop resistance to these drugs is to alter the target protein so that it is no longer recognised by the antibiotic-a trick bacteria accomplish by mutating the appropriate gene sequence. With PNA-based antibiotics on hand, such resistance would be futile: as a bacterium changed its DNA sequence, researchers would be able to follow suit and change the sequence of the antibiotic PNA.
Exciting as the prospects for PNAs sound, many problems remain-some of them related to the very concepts that underlie genetic medicine. In an ideal world, antisense or antigene drugs like PNAs would home in on their target sequences with pinpoint precision. There would be no side effects, because if the sequences did not match exactly, the drug would not bind.
Alas, despite the beguilingly elegant premise, all is not well in the world of genetic medicine. 鈥淭here has been a lot of dubious science surrounding antisense, with only 5 to 10 per cent of papers producing reproducible results,鈥� says Arthur Krieg of the University of Iowa, the editor of the Journal of Antisense and Nucleic Acid Drug Development.
One of the main disputes revolves around the central attraction of antisense therapy, namely that the antisense agent is supposed to bind exclusively and specifically to the portion of mRNA it is targeted against. Many scientists argue that the therapeutic effects seen in clinical trials of other synthetic DNAs are caused by the drug binding to a protein rather than to an mRNA sequence. If this is what is happening, they would in fact be no different from any other drug on the shelf.
But perhaps the most formidable problem facing scientists developing PNA-based antibiotics is finding a way to get them into bacterial cells. In order to target specific sequences, even the smallest PNA-based antibiotic would be a molecular monster-perhaps ten times as big as the antibiotics currently used to combat infections. In addition, because the peptide backbone of a PNA is electrically neutral, it is more difficult to transport across the fortress-like bacterial cell wall, which prefers to admit charged molecules.
The Californian company ISIS Pharmaceuticals has been working on developing PNAs as drugs to combat diseases such as cancer, and until recently had been licensed by Nielsen to develop PNAs as antibiotics. But David Eckers, a senior chemist at ISIS, says that he has seen no hint of a successful idea for getting such drugs into bacteria. 鈥淚 don鈥檛 want to come across as too negative, but I think it will take a major breakthrough,鈥� he says.
To get PNAs into bacteria and other target cells, researchers must find a way to disguise their size and shape, and also to make the electrically neutral backbone less unpalatable. One strategy is to attach the molecule to something that the cell needs and actively imports: a nutrient, perhaps, or iron. Or PNAs could be hidden in one of those little fatty packages called liposomes, which can cross the cell wall like Trojan Horses, taking the antisense drugs in with them.
There are some glimmers of hope. In September 1998, Ulo Langel of Stockholm University reported in Nature Biotechnology (vol 16, p 857) that synthetic peptides can carry PNAs into rat cells. Although rat cells are easier to enter than bacteria, Langel says he plans to try to get PNAs into bacteria next. Just how he might do this is something he is coy about discussing, as he says the information might compromise a future patent.
Still, Eckers remains cautious. 鈥淭he ideas currently being tried are all recycled. I鈥檓 sceptical that anyone can come up with a trick that will work,鈥� he says. Which may have something to do with why ISIS last month sold the rights to future PNA antibiotics to a group of venture capitalists led by Pantheco鈥檚 Anker Lundemose, in a deal worth $12.3 million. Now Nielsen and Carsten Schou, director of biology at Pantheco, are working on developing PNAs into antibiotics.
For the time being at least, Nielsen and Liam Good, who is based at the University of Copenhagen, have sidestepped the problem of crossing the bacterial cell wall, and demonstrated that PNAs are indeed capable of mucking up microbial metabolism through an antisense mechanism. The researchers used cells stripped of their walls, as well as a special strain of Escherichia coli with a 鈥渓eaky鈥� wall, to allow the PNAs to get into the bacteria. In a paper published in Proceedings of the National Academy of Sciences in March (vol 95, p 2073), Nielsen and Good demonstrated that PNAs containing a sequence that matched that of E. coli鈥檚 ribosomal RNA inhibited protein production in the bacterium. PNAs with a mismatched sequence had no effect. The sequence-specific PNAs also inhibited cell growth, whereas the mismatched PNAs did not.
Naked and leaky
In a later paper published in Nature Biotechnology (vol 16, p 355), Nielsen and Good demonstrated that PNAs could also be used to inhibit specific bacterial genes. Again, they found that only PNAs with the correct matching sequences could reduce the activity of the genes in their naked or leaky E. coli strains. In addition, they found that the PNAs were so specific that when two alterations were made to the sequence, their ability to inhibit gene activity was diminished. Six changes destroyed the PNA鈥檚 activity completely. 鈥淚 think we have proof of concept and have demonstrated an antisense mechanism,鈥� says Nielsen.
The two papers go a long way to proving that PNAs can act as sequence-specific antibiotics capable of gumming up the works in bacteria. Magdalena Eriksson, a biophysicist at Chalmers University of Technology in Gothenburg, Sweden, agrees that they are on the right track. 鈥淧NAs have the potential to be antibiotics,鈥� she says. But others are unconvinced. 鈥淲e need to see that PNAs can penetrate the bacterial cell wall,鈥� says Susan Freier, director of molecular and structural biology at ISIS. 鈥淚f they can, they could be [used as] antibiotics.鈥�
Time will tell. It will take about ten person-years to sift through the options and to design a PNA against a realistic target like Staphylococcus, Nielsen says. Which is another way of saying that whether PNAs ever make it as antibiotics is anyone鈥檚 guess. Still, PNAs are an example of science in action, and there is nothing more likely to surprise than science in action, particularly when the science is biology.
Nielsen, whose reputation is on the line, remains confident of success. 鈥淲e believe in [PNAs],鈥� he says, 鈥渂ecause they are our baby.鈥�
