THEY ARE before-and-after shots no women’s magazine would print, but among cancer doctors they have become something of a classic. The first photograph of the woman’s leg is not a pretty sight. What look like large black leathery warts deform the skin around her knee, evidence even to the untrained eye that something is terribly wrong. In the second picture, a few scattered scars are the only visible remnants of what were malignant melanomas.
Slides of these photos are being shown at medical conferences around the world as evidence of the exciting potential of a novel cancer drug called Genasense. It is one of a new class of such agents that can seek out and destroy cancer cells anywhere in the body without the harmful effects of today’s chemotherapy. Clinical trials of Genasense in several types of cancer look promising, and assuming the large final-stage trials give equally impressive results it could go on sale next year.
The day that happens will be hugely significant for a whole biotech field. That’s because Genasense is an “antisense” drug – one designed to switch off the functioning of a specific gene. In the 1980s antisense was touted as the next revolution in drug design, with cures promised for cancer, viral infections and indeed any disease caused by overexpression of a gene. But the field has had a rough ride in the past decade, and it started to look like the antisense bandwagon was being driven more by hyperbole than facts. Genasense looks set to change all that.
Advertisement
But something about the drug does not add up. Several experts think its main mode of action may not be antisense after all: they think Genasense is triggering a different response, a little-known immune reaction called a “CpG effect”, which just happens to shrink tumours. Some experts even think it is possible that most of the other supposed antisense drugs in advanced clinical trials may be working that way too. Genasense’s manufacturer, Genta of New Jersey, vehemently denies these claims. But if the sceptics are right, then a lot of the hype about the leading antisense drugs – and the investment in their success – is based on a fallacy. Patients may not see what the fuss is about – the drug’s efficacy is not being questioned, after all – but repercussions for the biotech industry could be huge. “There definitely is a fear that the investment community would look upon this as a failure of the field,” says Arthur Krieg, chief scientific officer at Coley Pharmaceutical Group, a biotech firm in Wellesley, Massachusetts.
The antisense idea was first demonstrated in 1978 by Paul Zamecnik, then at Massachusetts General Hospital in Boston. The concept is an elegant one that exploits DNA’s basic chemical properties. DNA molecules are long strings of building blocks called nucleotides, which are made of a sugar unit, a phosphate group, and an all-important base, which can be one of four options: A, C, G, or T. The sugar and phosphate groups form the structural backbone of the molecule, while the sequence of bases encodes the useful information held in our genes. Base A will form bonds with T, and C with G, so two strings of DNA with “complementary” sequences of bases will bind stably.
Normally, genes make proteins via an intermediary molecule called messenger RNA (mRNA), which has similar structure to DNA. An antisense drug is a string of nucleotides with bases that are complementary to, and therefore bind to, a short section of a gene’s mRNA. This blocks subsequent protein production, and the mRNA is destroyed by cell enzymes (see Diagram).
And because antisense drugs target specific problem proteins, they should only affect diseased cells, not healthy ones. So they could be given through the desirable “systemic” forms of drug delivery like pills or injection into the bloodstream, without the harsh side-effects of chemotherapy on healthy tissues.
But when scientists tried to translate that concept into useful drugs several problems began to surface. The nucleotide strings, or “oligonucleotides” were rapidly broken down by enzymes in the blood and in cells. To overcome that, oligos with modified sugar-phosphate backbones were tried, with varying degrees of success. Researchers also found that not all genes were useful targets, and not all oligo sequences were effective.
In more than two decades, only one antisense medicine has been approved. This is Vitravene, developed by Isis Pharmaceuticals of Carlsbad, California, to treat a viral eye infection that can affect people with AIDS. But this is only a small market. What’s more, the drug has to be injected into the eye, so failing to live up to antisense’s original promise of easy drug delivery.
That’s why, despite being the second in line, Genasense is seen as the real proof of principle for antisense therapy. It is designed to work against a protein called Bcl-2 that allows cancer cells to ignore signals to commit suicide. When given to patients over several days before standard chemotherapy, Genasense is thought to help the other drugs kill the cancer cells.
Bcl-2 overexpression plays a role in many types of cancer, including breast, prostate, colon and some skin tumours, as well as leukaemia and lymphoma. That’s partly why, two years ago, the pharmaceutical giant Aventis agreed to pay Genta a whopping $477 million to develop Genasense in return for a share in the profits. It remains the second-largest deal for a single compound in the history of biotech, reflecting the excitement over antisense.
Aventis may well have backed a winner. Numerous small clinical trials of Genasense have given promising results. Optimism is high in the run-up to the announcement of the first findings from its final-stage trials, due next month. Cy Stein, an antisense expert at Columbia University in New York, who sits on Genta’s scientific advisory board, says: “Genasense has the best chance of any of the oligonucleotides I’ve seen.”
Strange goings-on
So what’s the problem? One of the first to question the way antisense works was Sudhir Agrawal, chief executive of the biotech firm Hybridon in Cambridge, Massachusetts. In 1996, he was testing an antisense oligo against a gene from the human papilloma virus (HPV). When he injected it into infected mice, the characteristic lesions caused by the virus cleared up, as expected. And when he used a different oligo as a control, there was no antiviral activity. “It was a beautiful example of a sequence-specific effect,” says Agrawal. Or so he thought, until he carried out one more control experiment. This time Agrawal used the original antisense against a different strain of HPV. This virus had four different bases in the key gene, so the antisense would not match. But, amazingly, it worked just as well as the first oligo, and the mice’s lesions disappeared. “It was very puzzling,” he recalls.
It took him a few years to work out what was going on. The antiviral effect, it turned out, had been caused by a little-known immune reaction, occurring before the oligos even reached the cancer cells. It had been triggered by the presence of “CpG motifs” – the bases C next to G (“p” refers to the linking phosphate group) in the oligo. In the DNA of vertebrates, CpGs usually have an extra methyl group attached. Any “naked” CpGs knocking around the bloodstream are likely to belong to invading bacteria or viruses. Several types of immune cell have receptors that recognise naked CpGs, and the cells respond by secreting chemical messengers called cytokines that step up the immune response. Over the past few years, several lines of evidence have suggested that this “CpG effect” on immune cells can cause the immune system to kill cancer cells or virus-infected cells.
It is possible, though by no means certain, that Vitravene could be working through the CpG effect, as the motif is present. A 1996 paper published by its manufacturer admitted that some of Vitravene’s antiviral effects could not be due to antisense, although the article also stated CpG effects were unlikely to be responsible. A spokeswoman for the firm says: “The activity observed with Vitravene is consistent with an antisense mechanism. Confirming the mechanism for any drug in humans is difficult, and can take a significant period of time.”
So what about Genasense? The crucial CpG motif occurs twice in the oligo. And at least three groups of researchers have shown through separate cell culture or animal studies that the drug can indeed trigger a CpG response. Coley Pharmaceutical Group, for example, has synthesised ten of the antisense molecules that have reached clinical trials, and compared their ability to cause CpG effects. Arthur Krieg says several had CpG activity, and Genasense had the highest level of all.
That doesn’t mean that Genasense cannot be working through antisense too; the drug could well have a dual mechanism, like plenty of medicines used today. The key question is how much of Genasense’s activity is down to antisense, and how much is down to CpG.
Genta, for one, is adamant that its drug’s main mechanism is antisense. As evidence it cites a study of blood samples from patients treated with Genasense, which revealed no activated immune cells. Sceptics are not convinced, however: they say that such cells wouldn’t necessarily be present in the bloodstream anyway. Genta has also shown that a methylated version of Genasense – which should not trigger the CpG effect – still shrank tumours in mice. But this test involved very high doses of Genasense, so the cancer cells might have been affected even if the drug’s antisense activity was relatively weak, sceptics contend.
Perhaps the firm’s strongest argument is that Genasense causes Bcl-2 levels to fall in some patients’ cancer cells. This appears to confirm that the protein is being selectively blocked – presumably through antisense. Chief executive Raymond Warrell says: “We have very clearly shown that we down-regulate Bcl-2 in patients receiving the drug.”
But a cunning experiment by Gunther Hartmann and colleagues at the University of Munich in Germany suggests that even this could be explained by the CpG effect. They compared the effects of three antisense oligos on cultured human leukaemia cells. The first was Genasense; the second targeted Bcl-2 but bound to a different section of the mRNA and lacked CpGs; and the third targeted a different gene from Bcl-2 and did have CpG motifs.
The results were revealing. Genasense and oligo number three strongly activated several genes involved in the immune response, but oligo number two didn’t touch them. Genasense and number three had changeable effects on Bcl-2 and several other proteins, sometimes lowering and sometimes raising them, and also had variable effects on the rate of cell death. The researchers concluded that it is probably the CpG response rather than antisense that causes Genasense’s anti-cancer activity (Journal of Leukocyte Biology, vol 72, p 83).
Finally, a preliminary study carried out by Cy Stein challenges the premise that shutting down Bcl-2 is the main reason for the therapeutic effects of Genasense, at least in a prostate tumour model. He used a technique called RNA interference, which can block the effects of specific genes, to block the Bcl-2 protein in cultured prostate cancer cells. The cells didn’t die, suggesting that turning off Bcl-2 via antisense wouldn’t harm them either.
When asked to explain these results, Genta chief executive Raymond Warrell says: “Both studies are from in vitro experiments – their reference to in vivo human biology remains to be proven. There is no evidence whatsoever for any immunomodulating effect from [Genasense] in human subjects.”
And it is unlikely that human trials will be carried out to resolve the issue, as it would be unethical to give some patients Genasense and others controls like the ones Hartmann used. Figuring out the mechanism is low priority for Genta. “It’s not essential for us to prove it is X and not Y,” says Warrell. “I truly don’t care. I don’t think it’s a meaningful question.”
That sounds dismissive, but from a regulatory standpoint Warrell is right. Drug regulatory agencies don’t require firms to prove their medicine works a certain way before approving it. If that were the case, the vast majority of today’s medicines wouldn’t be here. Nor would Genasense be the first medicine people thought worked in a particular way, only to be proven wrong. Take statins, the drugs given to lower cholesterol levels in people who have had heart attacks. Originally, statins were thought to work by inhibiting an enzyme involved in cholesterol synthesis. But it turns out they also have an anti-inflammatory effect that may be just as crucial as their cholesterol-lowering abilities in stopping blood-vessel plaques from bursting (żěè¶ĚĘÓƵ, 11 January, p 36). Aspirin is another example, and there are many more. It’s not that biochemists keep imagining things. Drugs can affect the body’s chemistry in many ways, and the effect that is discovered first may not be the most significant.
In the case of Genasense, however, its mechanism of action is being questioned before it is approved. So a different issue arises: should Genta be allowed to go ahead with its plans to brand and advertise its product as an antisense drug even if that effect were only to play a minor role in its activity?
And the question may not only apply to Genasense. Most of the antisense drugs in advanced human tests have CpG motifs.
Of course, patients care more about whether they will be cured than exactly how that cure is achieved. As James Mannion, chief executive of Epigenesis, another New Jersey antisense company, says: “At the end of the day, this is about improving patients’ health. I don’t think the patients care one bit.”
Reputations at risk
But investors are another matter. They do care if the science they are bankrolling isn’t real. If the poster child for a new technology turns out to be an impostor, the whole antisense field could suffer. In many people’s minds, Genasense and antisense are inextricably linked; the drug’s moniker is no coincidence, after all. “If it is approved, and it does not work that way, it’s bad for the field,” says Hartmann. Biotech analyst Cory Kasimov of the investment bank Ryan, Beck and Co in New York says that if Genasense does turn out to work mainly through the CpG effect, it may not affect sales, but it could hurt other antisense firms.
Ironically, the next generation of antisense drugs coming through early development cannot be working through CpG as they have a different sugar-phosphate backbone. The prospects for this new breed of antisense agents should look bright. They are potent molecules with a clear mode of action, says Agrawal, and they are so stable that they can be given as pills, something researchers could not have dreamed of 10 years ago. But if the antisense field suffers yet another embarrassing setback, the reputations of these drugs could suffer too. How they will weather the oncoming storm remains to be seen.
The cloud hanging over antisense does have a silver lining, however. When scientists began to suspect that Genasense and its like had non-antisense effects, they started investigating this mysterious immune response. Now several firms, including Coley and Hybridon, are developing oligos that maximise what was once seen as an undesirable side effect, to create new treatments for asthma and allergy, as well as cancer. The first candidates have reached clinical trials. “We found a problem with antisense, but we exploited it,” says Agrawal.
But that’s another story.