By Caroline Ryan
BBC News website health reporter
Animal tests on the kind of drug given to the six men ill in a London hospital may not be the best way of evaluating the effects in people, an expert warns.
Mice may not be the best animals to test all drugs on, experts say
The drug they took stimulates a protein only found in humans.
Dr David Glover, an expert in drug testing, said this meant animal tests of medicines of this sort might give falsely reassuring results.
He said it might be better to look at innovative ways of testing small amounts of such drugs on people.
The drug, TGN 1412, which the six men took belongs to a class called monoclonal antibodies.
It is hoped they could combat a wide range diseases, including cancer.
They are created in the lab by fusing or merging a cell that produces antibodies - the foot soldiers of the immune system - with a cancer cell.
Monoclonal antibodies, such as the breast cancer drug Herceptin, have been increasingly used over the last 25 years.
Most monoclonal antibodies prevent something happening in the body - they are "antagonists".
For example, Herceptin works by blocking the action of the Her2 protein, which can fuel the growth of breast cancers.
But TGN 1412 is an "agonist" - which boosts a particular action.
It boosts the activity of human immune system protein called CD28 which is present on the surface of white blood cells.
There have been concerns it might have been inappropriate to test such drugs on healthy volunteers, whose immune systems are already working effectively, as a further boost might push their systems into over-drive.
At this stage, there is no hard evidence to suggest this is what happened in the study that left six men seriously ill.
The problem might equally be due to a fault with the manufacturing process, or simply a unique reaction to the drug in humans that could not have been predicted.
Dr Glover carried out a large number of studies into monoclonal antibodies in the 10 years he spent as chief medical officer at Cambridge Antibody Technology.
He said that, in this case, animal studies may have given falsely positive safety and effectiveness results prior to the human trial.
He told the BBC: "This is a novel target, an antibody against human CD28 on human white blood cells.
"Therefore, there may be differences in animals, and tests may be very difficult to do or interpret."
He added: "Since this drug was targeted at leukaemia, it may be better to have tested it on leukaemia patients first who could conceivably have got benefit, rather than healthy volunteers."
He said one possible way of testing drugs which were specifically focused on human biological targets was a technique called micro-testing.
This involves creating a blister on a person's arm - which contains cells that a microscopic amount of a drug could be tested on.
"This would mean you could look at the effect on cells before trying it on a whole person."
Another way, he suggested, would be to manipulate the immune system of mice so they contained human immune cells - in order to give a better picture of what might happen when the drug was given to humans.
But he stressed monoclonal antibodies had been a success in treating a range of disorders, and research on them should not be discontinued because of the events at Northwick Park.
Professor David Isenberg, a rheumatologist at University College London, said there were concerns that the current problem could undermine confidence in monoclonal antibodies.
"If it is seen out of context it could indeed give monoclonal antibodies a very bad name.
"But I want to emphasise that there have been wonderful successes in the fields of cancer, and the fields of inflammatory diseases, particularly arthritis.
"People's lives have been saved, and the quality of people's lives has improved dramatically over the last 25 years, thanks to monoclonal antibodies."
The concerns over this drug trial should also not be seen as signalling wholesale problems with animal testing, experts said.
Professor Chris Higgins, director of the Medical Research Council's clinical sciences centre, said that currently, all drugs were tested on at least two species of animals before they can be tested on humans.
"Animals are the best models we have for humans, but we all know they aren't absolutely perfect.
"Very occasionally animals do not pick up potential problems.
"But, of course, there are thousands of potential drugs that were tested on animals, and never got to humans because their potential toxicity effects were recognised."