CRUK rebuttal

Tamoxifen
(Paragraphs in italics are extracted from CRUK’s letter)

‘Tamoxifen is one of the most successful anti-cancer drugs in current use, and it is estimated to have saved the lives of over 400,000 women worldwide…’

The success of tamoxifen is not questioned. However, its development provides a classic example of researchers choosing to disregard conflicting animal data when they felt it expedient to do so. This kind of selectivity helped the drug find a place in clinical practice, as the following examination makes clear. It also shows animal experiments being conducted to verify or learn information that could have been obtained from humane alternatives.

‘It was first developed as a contraceptive, and works by blocking the action of the female hormone, oestrogen.’ 

This statement omits crucial inter-species variations. The drug was initially developed by ICI (and subsequently shelved) during the 1960s as an oral contraceptive. In rats, tamoxifen can indeed prevent ovulation or terminate pregnancy. In women however, tamoxifen stimulates ovulation and is listed as a treatment for infertility.

Tamoxifen acts in breast cancer therapy by blocking the action of oestrogen in breast tissue. In monkeys, and rats at low doses, tamoxifen also acts as an anti-oestrogen but in mice, dogs, and rats at high doses, the drug has the opposite effect, behaving like an oestrogen. These anomalies prompted the following comment from an eminent Professor of Obstetrics and Gynaecology: ‘…significant species variation has been observed in target tissue response to oestrogens and antioestrogens making it hazardous to predict therapeutic activity in the human by extrapolation of effects in experimental animals…’

‘However, during the 1950s, research involving rats showed that hormones can also play an important role in the development of breast cancer. Following on from this work, scientists discovered that hormones are similarly important in humans.’ 

This is historical revisionism and, at best, a partial truth. It was known since the late nineteenth century that removing the ovaries from patients with breast cancer shrank their tumours. These initial clinical observations were later complemented by animal experiments (much earlier than the 1950s), which showed that early ovariectomy can reduce the incidence of mammary cancer in high-incidence strains of mice. It can persuasively be argued that these animal experiments, and the ones that followed, were cruel and superfluous. Janet Lane-Claypon’s meticulous epidemiological work in the 1920s revealed a great many of the risk factors for breast cancer related to lifetime exposure to oestrogens.

‘As a result of [these human hormone discoveries], tamoxifen was then investigated in breast cancer and is now one of the most successful treatments for this disease.’

This statement again omits crucial information. Rats were used in the initial studies that showed efficacy against breast cancer. Partially as a result of these studies, tamoxifen was licensed and was first used in clinical practice in 1973. At that time, however, there was very little concern about side effects, mainly because other anti-cancer drugs were so toxic by comparison. It was only after extended use in patients with breast cancer, and the proposal to use the drug as a preventative, that further rodent cancer studies were carried out in the late 1980s, and published several years later. These studies found that tamoxifen produced fast-growing liver tumours in rats but not mice.

It is accepted by several figures directly involved in working on tamoxifen that, had these animal studies been carried out in the 1970s, the drug would not have been developed further. John Patterson, then medical director at ICI, declared: ‘If this was a new chemical entity those findings in rats would have caused us to stop its development, but the human experience gave us confidence.’ And V Craig Jordan, a researcher who played a key role in the drug’s evolution, commented in 2003: ‘It is fair to say that if rat liver tumours had been noted in the early 1970s, drug development in this area would have stopped, as there was no successor to tamoxifen.’

The rat data held up a large U.K. cancer-prevention trial involving tamoxifen. On March 12 1992, the Medical Research Council (MRC) withdrew its support for the study pending further toxicological studies. The two remaining sponsors – the Imperial Cancer Research Fund and the Cancer Research Campaign (now combined as Cancer Research UK) – decided to continue the trial on their own, pending Department of Health approval. In other words, the animal data could be selectively ignored when clinical experience in humans was anomalous.

Revealingly, an article published in Science News in 1992 quotes a researcher from the Royal Marsden Hospital, who was part of a team that had already been trialling tamoxifen as chemoprophylaxis: ‘I find it remarkable that people are still looking at rat data, which are generally irrelevant anyway. I think the human data now are the critical factor.’

Cutting off a tumour’s blood supply

(Paragraphs in italics are extracted from CRUK’s letter)

‘Research using rodents is helping our scientists to understand how tumours grow and spread, and to develop drugs that block this process.’

This implies that it is only through rodent research that this kind of information can be obtained. In fact, sophisticated human cell culture and computer models are being used to examine the behaviour of tumours. They are also being used in drug development.

‘Only in a living animal is it possible to recreate the growth of blood vessels in a tumour and the interaction between a tumour and neighbouring tissues and organ.’

This statement gives a spurious validity to the concept of using ‘whole’ animals to predict the response of human tumours. In fact, both main types of animal cancer models – xenograft (human tumours transplanted into mice) and genetically engineered mice – suffer from key drawbacks that make them poor surrogates for human disease. A 2011 editorial published in Nature Reviews Clinical Oncology bemoaned the extremely high attrition rates for cancer drugs: ‘Only 5% of agents that have anticancer activity in preclinical development are licensed after demonstrating sufficient efficacy in phase III testing…To compound this issue, many new cancer agents are being withdrawn, suspended or discontinued.’

At least partially to blame for this dire state of affairs are ‘suboptimal’ preclinical strategies – most usually, this means animal experiments. The editorial then goes on to state:

‘A key drawback of animal models is that they do not represent the primary tumors from which they are derived in terms of tumor heterogeneity and the mechanisms of drug resistance. Xenograft models lack the broad molecular transformation events that occur in human tumors. Furthermore, since the stromal component of the tumor [the non-malignant surrounding tissue] is not human the effects of the microenvironment on drug response are often not reflective of the primary tumor. Importantly, the growth rates of human-derived xenografts are considerably more rapid than primary tumors and, as a result, are much more likely to respond to antiproliferative agents. Testing of antiproliferative drugs in animal models might provide a false indication of the potential efficacy of a drug. Also, the immune system in such animal models is compromised, hindering the testing of immunomodulatory agents. Genetically engineered mouse models circumvent some of these limitations as they are immune competent but they still suffer from having rodent-derived stroma.’

This analysis (although followed, bafflingly but true to form, by a plea for better animal models) clearly indicates that these experiments are not succeeding when it comes to delivering new therapies. The intrinsic problem here is that animal researchers are constantly trying to refine a fundamentally flawed paradigm. As Greek (2010) points out: ‘If a modality, be it a research method or weapon or means of transport, is ineffective it should be abandoned regardless of what else is available. No one argues, for example, that we should use Ford trucks to accomplish space travel as the Ford truck, a good product though it may be, is simply not viable for going to the moon or Mars.’

‘Animal research has enabled scientists to develop new drugs that target existing tumour blood vessels or prevent new ones from forming. Some of these are now being used to treat cancer patients. One example is a drug called combretastatin. Testing of this drug in rodents showed that it can effectively cut off the blood supply of tumours. Following on from these studies, Cancer Research UK undertook the initial clinical trials of this drug. It is now in late stage development with a pharmaceutical company.’

These classes of drugs, and combretastatin in particular, are a curious choice if CRUK wishes to highlight a success of its research programme.

Drugs that prevent new blood vessels from forming are known as ‘antiangiogenesis therapies’. Antiangiogenic therapy is based on the theory (now more than 40 years old) that halting new blood vessel formation in tumours will slow their growth. A 2011 review highlights how, for this class of drugs, ‘two decades of positive preclinical studies have yielded only modest incremental changes in the clinic’. It goes on to say ‘the debate has been fuelled by modest clinical benefits, high drug cost, and adverse side effects, in addition to converg¬ing findings published in the past 2 years, which relate to limited drug efficacy in early-stage disease’. The review adds that ‘questions have emerged about the basis of drug resistance and the limitations of predictive preclinical models’.

Many angiogenesis inhibitors went to clinical trials but never made it to the marketplace, due to limited efficacy. Two such early candidates, angiostatin and endostatin, showed good results in animals but not in people. These are the rule rather than the exception. A sub-group of antiangiogenesis drugs inhibit a naturally occurring protein called vascular endothelial growth factor (VEGF), which creates new blood vessels. These drugs have some of the highest failure rates of all cancer drugs. However, Avastin (bevacizumab, approved for the treatment of colorectal cancer in humans in 2004) blocks human VEGF but not the mouse version. Had the mouse results been thought to apply to humans, the drug would have been discontinued.

The other class of drugs cited by CRUK, which are supposed to destroy blood vessels already present in cancer tissue, are known as ‘vascular disrupting agents’ (VDAs). Combretastatin is one of these, but it has not been a clinical success, despite many promising animal studies. Developed as Zybrestat by the company Oxigene, Phase I trials have demonstrated several toxic side effects not observed in animals. For example, a potent version of combretastatin could not be given in an effective dose to patients with high blood pressure. CRUK has details on its website about a small 2007 Phase I trial that looked at combretastatin, in combination with immunotherapy, for advanced bowel and pancreatic cancer. The drug combination used had such serious side effects that the researchers could not step up from the lowest possible dose. Of the 12 patients recruited, 10 stopped treatment during their first cycle, and none showed improvements in their tumour size afterwards.

In November 2011, a Phase II lung cancer trial of combretastatin (combined with other chemotherapies) failed to demonstrate any survival benefit. A 2010 Phase II/III trial for a type of aggressive thyroid cancer failed to show a significant improvement in survival for patients taking combretastatin (again as part of a drug cocktail). Finally, a Phase II ovarian cancer trial, despite upbeat spin, could demonstrate only a 13.5 per cent response rate to the drug, as measured by tumour shrinkage of at least a third. There was no progression-free survival or duration of response data reported.

Besides Oxigene, a number of other companies have been testing VDAs on patients, and faring as poorly. Australian drug company Antisoma, in collaboration with Novartis, failed in its 2011 Phase III trials of the drug ASA404 against lung cancer. This was despite the fact that ASA404 resulted in significant long-term response in 80 per cent of tumour-bearing mice. The drug had already failed to deliver benefits in an ovarian cancer trial. AstraZeneca had to halt its Phase II trials of ZD6126, due to severe cardiac side effects. Astoundingly, although AstraZeneca then pulled out of developing the drug further, it still funded invasive and cruel rat studies to try to establish why the drug was so toxic to people.

Understanding the genes that cause cancer

(Paragraphs in italics are extracted from CRUK’s letter)

‘Genes like p53 are therefore very important in cancer development. However, scientists need to understand how these genes work, before they can harness this knowledge to design new ways of preventing and treating cancer.’

The p53 gene was discovered more than 30 years ago. Scientists have been conducting animal experiments to find out how it works for decades, but have not yet managed to ‘harness this knowledge’ for patient benefit. This is in stark contrast to the mass of studies that have shown speculative therapies to be highly effective in animals. According to a breathless article written by key p53 researchers in 2010, ‘sophisticated animal models have shown that activation of the p53 response in even advanced tumors can be curative’. Several candidate drugs have shown activity in xenograft animal models, (usually mice with human tumour tissue grafted under their skin), including the so-called Nutlins and APR-246. A study reported in 2011 apparently showed that ‘activation of p53 by our highly potent and optimized MDM2 inhibitors can achieve complete tumor regression in a mouse model of human cancer’. The study was part-funded by the company that produced the drug, in which the lead researcher owned stock.

These kinds of wondrous results in mice have not to date been replicated in humans. More than 150 trials exploiting the p53 pathway have been conducted, but only one drug has been licensed so far – a highly controversial gene therapy, which is available only in China (the FDA in the US refuses to back it). Researchers point out that it is too early to say if the main candidate drugs will enter the clinic, but in the meantime ‘these compounds have become popular research tools’. In other words, no patient benefits may arise, but the substances are still useful for further animal testing.

Hopes that the p53 gene would help to stratify cancer patients into different groups for more tailored therapy are also not being realised. One of the co-discoverers of p53 wrote in 2009: ‘The discovery that p53 has a pivotal role in cell killing by DNA-damaging agents, many of which are in routine use for cancer chemotherapy, gave rise to the expectation that TP53 mutation status would prove to be a reliable predictor of therapy response and patient prognosis. However, these expectations have not yet been fulfilled, probably reflecting the complex genetic nature and extensive diversity of individual tumours.’

This is an argument that Animal Aid has made before. Simplistic animal models, which involve turning off biological pathways or key chemical reactions (by ‘knocking out’ certain genes), do not reflect the far more complex behaviour and genetics of human cancers. This is partly why these models are so poorly predictive. In addition, p53 is not just a cancer-suppressing gene. Its presence in the human genome, thousands of years before cancer became a disease of such significance for our own species, points to a much wider role. Studies have demonstrated that the gene could play an important role in female fertility, development, stem cell division and ageing. These aspects are simply not studied in single outcome, short-lived animal cancer victims.

Despite the failure to translate animal research into clinical progress, the p53 gene is still the subject of countless hours of research work (in 2009, there were ‘nearly 50,000 PubMed-listed publications so far and a steady flow of new ones entering cyberspace every week’). A fortune has undoubtedly been spent funding such programmes. The real reason for its popularity is probably not its hypothetical, endlessly postponed clinical benefits, but its trendiness – ‘in the early part of the last decade of the twentieth century, p53 was recognized as a major tumour suppressor and became a most fashionable gene and protein to study – a fashion that remains today’. Researchers in the field remain shameless in their hyperbole, despite decades of experimentation yielding, at best, paltry benefits to cancer patients.

‘Research with mice has been essential for providing information about p53. When scientists bred mice with a defective version of this gene, they found that they were very susceptible to cancer, proving that p53 is important in the development of the disease.’

This is again a highly contestable assertion – leaving aside the issue of clinical benefit. The initial discoveries about p53, although made during animal experiments, suggested that it was an oncogene (a cancer promoter), when in fact the opposite was true. Later, several mouse leukaemia experiments purported to show support for this hypothesis, although with hindsight this was far from being the case. The anomalous findings were disregarded as ‘exceptions to particular models’. Animal models of disease seem to facilitate this kind of selective interpretation, in line with the prevailing medico-scientific orthodoxy. Unexpected or inconvenient data are easier to disregard as the subjects are only ‘models’, and therefore prone to idiosyncrasies.

Furthermore, it is clear that human cell cultures and tissues have been used throughout the research history of p53, running parallel to the mouse experiments. What is to be done when cell cultures suggest a different conclusion from genetically engineered mice? How essential is the mouse data then to be regarded? A lead researcher commented in 2009:

‘Although studies in cultured cells have provided a rich body of seemingly impeccable evidence for the pivotal importance of p53 post-translational modifications, many of these conclusions are challenged by data from mutant Trp53 knock-in mice. It is these confusing results that will need to be reconciled over the next 10 years.’

Lastly, it was discovered that the p53 gene is lost or mutated in human colorectal tumours, and other human-derived cancer cells, before the knockout mice experiments that ‘proved that p53 is important in the development of the disease’. Similarly, the discovery that p53 mutations cause the human Li-Fraumeni syndrome (a very rare hereditary condition, characterised by early-onset of a wide range of malignancies) predated the mouse experiments by two years.

(Full references for quoted passages are available on request.)