To remind you what happened around our Lancet article.
According to my contract with the Rowett, my former employer, I have to
send my papers to the Director, Dr Peter Morgan, for "scrutiny" as
Queen's Council advised me.  What this means is that because of Pillip
James' carelessly drafted letters of 16 & 19 February 99, I am no longer
required to submit my drafts of scientific papers for his approval
before I can submit them to any scientific journal for publication.
However, I am still under a legal obligation to show the Director any
paper for his "examination" but not for editing and changing them.  This
I have to do sometimes between submission and actual publication.  This
was precisely what happened with the Lancet paper whose galley proofs
were sent to Dr Morgan at the end of September when we thought that the
paper would in fact be published by 1 October.  For a number of
technical reasons the paper was in fact published on 15 October that
gave two weeks for the "rubbishing brigade" to do their job.  Dr Morgan
advised me in his letter that 1/. I was still in their view under an
obligation to send my drafts to him before submission to a journal and
that the Rowett had the right to change anything in it even though they
could not possibly do this with these galley proofs and 2/. he sent the
galley proofs to a number of people, the Scottish Office, scientists in
the pay of the Scottish Office who oppose our work and, obviously, to
the Royal Society (RS).  The results, the threats to the Lancet Editor
and our villification by the RS and other top establishment figures in
the newspapers, BBC and in letters to the Editor in the Lancet are now
well-known and I am not going to deal with them here.

Although most of the attacks on us were very personal and the language
sometimes touching the level of the gutter, I am not willing to join in
such a debate debasing science.  However, I shall want to summarize the
scientific, quasi-scientific and misinformed criticisms on our paper,
which appeared mainly in letters to Editor in the Lancet and also some
magazine/newspaper articles.  I shall also try to summarize our
responses to these which appeared mainly in the 13 November issue of the
Lancet.

Even before our article appeared in the Lancet on 15 October, Prof.
Pickett, one of the six referees of our paper who apparently opposed its
publication, broke the over 200 year gentleman-convention of refereeing
and gave interviews to the press in which he stated his opposition to
the paper in a few  not-so-gentlemanly sentences.  However, his main
"scientific message" was that our results were uninterpretable because
we "changed horses in mid-stream", meaning that we changed the diets of
the rats from GM-potatoes to controls diets.  We have but two
interpretations of this criticism: either Prof. Pickett did not read the
paper he was supposed to have refereed or he deliberately misinformed
the reporter.   In either case, he was not the right person to referee
our paper.

Sir Aaron Klug, the President of the RS criticised our paper and said
that it would not have been accepted by the RS for publication because
we did not have low-protein control diets in our experiment.  Apart from
the fact that the RS does not publish peer-reviewed scientific papers,
his comments suggested that he may not have read our paper in which it
is clearly said that all diets (including control diets) contained the
same amount of protein and were iso-energetic.

In the Lancet letters mainly two peoples' comments were scientifically
worthy of reply.

In the invited commentary letter Kuiper et al (13 October) addressed two
main issues:  The first one concerned the specific criticisms of our
paper and secondly they dealt with the issue whether there are adequate
methods for the testing of the safety of GM-foods and whether this is
necessary at all.
Concerning the specific criticisms of our paper and our replies could be
summarized as follows:
1.  Although we suggested that the changes in gut structure and
metabolism reported in the paper were due to compositional changes in
the potato on genetic modification, we did not report these changes.
-    Clearly, in a short research letter this was not permitted by the
journal.  However, Kuiper et al have a copy of the two Rowett internal
reports in the public domain which were released by the Rowett on the
internet against my expressed wishes (and not by me as Kuiper et al say)
with the results of extensive analyses.   He could have checked up on
these showing that the two GM-potato lines were not substantially
equivalent in composition with the parent line or with each other.  All
the same, and most importantly, the GM-line 71 used for the Lancet
experiment had the same protein content as the parent line and therefore
all diets (including the controls) were iso-proteinic and iso-energetic.

2.  The diets were protein-deficient, contained only 6% protein and
therefore the rats were "starved".  -  It is true that 6% protein is
just over half of the protein content of the optimal rat diet.  However,
in a short-term (10 day) experiment the rats can cope with that and in
any case the control diets had the same low protein content and the
results were therefore comparable.  In contrast, Kuiper et al say that
low protein diets impair rat metabolism and therefore invalidate our
results.  Unfortunately, as their reference shows this conclusion is
more likely to be valid only for rats kept on protein-free diet and food
restriction.  Starvation is an emotive and meaningless term in respect
of our experiment because despite the low protein content of the diet
the rats were in fact growing, albeit less slowly than on optimal diets.

3.  Our results are probably due to the presence of poorly digestible
carbohydrates in our GM-potato diets.  -  All our diets, including
control diets contained the same poorly digestible raw potato starch.
Moreover, some of the changes in gut parameters also occured in rats fed
diets containing boiled GM-potatoes when compared to non-GM boiled
potato diets and these contained no poorly digestible raw potato
starches.

4.  Our experiments were incomplete, contained too few animals per group
and there was no optimal control group, necessary to discern any
toxicological significance.  -  Our study was not a toxicological but a
nutritional study.  Kuiper et al confuse these two different types of
evaluation experiments.  In our experiments there were six rats per
group.   This was more than adequate in comparison with the four or five
rats per group usually used in nutritional experiments and described in
top nutritional journals.  I have published at least 40 such papers of
nutritional evaluation of rat diets in the past 20 or so years.  In any
case, toxicological LD50-type experiments are only permitted in
exceptional cases in the UK.

5.   We had no "empty vector" (no GNA gene but all other genes in the
vector) containing potatoes as controls.  -  True,  this potato was
created at Durham at my request but our work was stopped before we could
test these.

The general comments of Kuiper et al which refer to the adequacy of the
testing of GM-food:

1.   Investigation of the gene product's role should not be confined to
only its effect on the gastrointestinal tract. -  Although they hold up
as an example of some work carried out in Kuiper's lab that was
published as a book Chapter (and not peer-reviewed!) this was somewhat
disingenious on their part because the work (Bt toxin from GM-tomatoes)
was not carried out with the gene product isolated from the GM-plant as
it should have been but was done with an E. coli recombinant.  As these
two are quite different, conclusions drawn from recombinant work cannot
be accepted as valid.  Incidentally, all biotech companies do the same
according to their submitted evidence to the regulatory authorities.
All the same, we wholeheartedly endorse Kuiper's views that ALL
GM-PRODUCTS SHOULD BE FULLY TESTED FOR THEIR BIOLOGICAL EFFECTS ON
ANIMALS AND THEN HUMAN VOLUNTEERS!  However, such studies have so far
only been carried out by us at the Rowett on GM-potatoes and GM-peas and
only with rats but their results have been published!

2.  All presently accepted GM-crops, GM-soya, GM-maize, etc have been
adequately tested.  -  This is a really strange statement because no
such studies, except one on round up-ready-soya in 1996, can be found in
the scientific literature.  In fact, what is in the literature (even in
the above Monsanto paper) shows that GM-soya is compositionally
different from non-GM-soya.  I am afraid, as scientists we need more
published evidence before we can accept the GM-foods already on the
market are adequately tested.

3.  Safety testing will have to be adjusted for the "second generation"
of GM-food plants.  -  This is a baffling statement.  What is the
difference between the need to test the first and second generation of
GM-food?  Why is it that it is only the second generation of GM-food
that needs to undergo "extensive toxicological and nutritional (for the
first time that this is demanded by someone high up in the EU
authorities) testing"?   We do not argue against this because it is a
great advance on Kuiper et al's previous stance but we need to stress
that this statement should equally apply to the first generation!  We
have no objections to the incorporation into the testing procedures his
proposed more sophisticated analytical testing methods (microarray
technology, mRNA fingerprinting, proteomics, etc) but find their logic
somewhat inconsequentional when they say that "depending on these
studies, further toxicological and nutritional studies may be needed".
This means in plain language that they first propose to carry out the
analytical testing for some toxicant or allergen in the GM-crop but
without knowing if they are there at all.  How can you find something
that you do not know what to look for and whether it is there.  Surely,
one does not want to waste time and precious resources by performing a
lot of sophisticated analytical tests when in fact the GM-food may not
show any signs of harming animals or humans.  It is more logical that
after some crude analyses (protein, starch oil) needed for formulating
diets for animal testing, nutritional and toxicological evaluations are
carried out first and if these indicate problems with the GM-food, we
ought to find what is the cause(s) for them.  However, considering
everything Kuiper et al's commentary is a good step forward in the
safety testing of GM-food.  Now we only want to establish who is doing
the testing, where is it going to be done and who is going to pay for
it?  It is gratifying to know that someone like Kuiper to whom funds for
such testing have already been allocated by the EU is a passionate
advocate for the biological testing of at least the second generation of
GM-crops.  Hopefully, and particularly if the first generation is also
included in his testing programme, the EU consumers will reap the
benefits, and if all these studies will be as transparently presented to
the scientific community (and consumers!) as promised by him in the last
sentence of his Lancet commentary.

Allan Mowat in his letter to the Lancet (15 November) also makes a
number of points which, together with our replies, are dealt with in the
following:

1.  Mowat raises a technical point of doubts about the role of tissue
fixation and its potential to create problems and urges extreme caution
in the interpretation of our results of gut histology, particularly as
the crypt depths reported in our paper are below (about 60 um) what is
usually found by others. -  He is right, the values found with these
potato diets are below of that found in well-fed normal rats.  However,
Mowat does not appreciate that as the potato lectin is antimitogenic it
causes a reduction in crypt size and this is re-inforced by the low
protein content of all the diets.  It is the more remarkable therefore
that even in the presence of such factors the GM-potato diet
dramatically and significantly increases crypt depth to about 90 um.  In
any case, we have something like well over ten years of experience in
carrying out these measurements on our rats kept on a great variety of
different diets and therefore we are in a unique situation to make
authorative statements on comparative and absolute values of crypt
depth.

2.  Mowat raises another technical point concerning IEL (intraepithelial
lymphocyte infiltration) counts in our jejunal sections.  According to
him, increased IEL is a sure sign of immunologically-mediated damage to
the gut induced by the use of lectins and not due to exposure to
GM-potatoes.  In any case, GNA as lectin could cause this.  -  We did
not give IEL counts in the gut of rats fed on GNA-spiked potato diets
for the simple reason that in our previous studies no lymphocyte
infiltration was observed even with a 1000-fold excess of GNA in the
diet.  Moreover, as published by us even a 1000-fold excess of PHA,
kidney bean lectin, caused no significant IEL increase in the gut.

3.  Our data do not confirm that the crypt hyperplasia observed by us in
GM-potato diets was due to genetic modification in the absence of data
on the increase in mitotic counts.  It is possible that the changes are
secondary to lectin-induced villus damage.  -  First, there is no villus
damage!  Second, it is true that mitotic counts are not given in the
paper but these will be included in a second paper which is in
preparation.  However, we can confirm that only the GM-potato diets
increased mitotic counts and no such increase was observed with the
parent potato diets, with or without spiking with GNA.
  There was a similar comment made by FitzGerald et al about the need
for the measurement of mitotic counts in the same issue of the Lancet.
Our answer is of course the same as above.

Peter Lachmann's criticisms in the same issue of the Lancet were hostile
but as they touched on scientific issues we decided to deal with them:

Points no. 1 & 2 have already been answered above.

3.  Were the assays done blind on coded samples? -  Yes, they were.

4.  Lachmann states that as the differences in gut structure between
rats fed GM and non-GM diets may not be pathological, they can be
dismissed.  -  This just shows that Lachmann does not understand the
issue.  All GM-foodstuffs so far accepted have been passed on the basis
that they were substantially equivalent to their non-GM counterparts.
Thus, we do NOT have to prove pathological differences but only that the
differences are significant.  It is up to the companies and the
regulatory authorities to show whether the changes are pathological.  In
any case, with Lachmann's background in medicine he ought to have
appreciated that stimulation of gut growth in the young is at the
expense of body growth (which is potentially harmful) as intestinal
turnover uses up somewhere between 30-40% of the dietary intake of
proteins.  In the colon it is potentially even more serious because
crypt cell proliferation in the colon may aid the development of
tumours, and you cannot have anything more pathological than that.

5.  His final point is that we have made no correction for "data
dredging" in our statistical analysis.  -  We did not have to correct
for data dredging  because we did do no such thing.  The statistical
methods of multivariate analysis of significance, one-way ANOVA, paired
t-tests and Tukey's tests are all described in our Table 1.  These
analyses have been seen and approved by independent statisticians,
including one of the referees of our paper.  We are more likely to
accept the advice of these independent people than that of Lachmann with
his partizan views and of rather unproven statistical expertise.

In our final analysis, most of the comments by Kuiper et al and Mowat
had legitimate scientific basis and were not personal.  This is
something that cannot be said of most of the others.  We think that
overall the Lancet initiative was helpful for both science and consumers
by moving away from personal attacks and misinformation to a valid
scientific debate on the safety testing of GM-foodstuffs.  Let us hope
that some of the conclusions will pass into practice and be accepted by
the EU and all the regulatory authorities whose job was in the first
place to start up this process but did not do so.

Links
 

Pusztai and Ewen write in the Lancet - Go to BBC - Hint: search "Pusztai" or "GM food"
The paper: "Effect of diets containing genetically modified potatoes expressing Galanthus nivalis lectin on rat small intestine."  Lancet. 1999 Oct 16;354(9187):1353-4.
Read the storry behind the storry: Journal editor threatened
 

A commentary by Thorkild C. Bøg-Hansen, Senior Associate Professor, University of Copenhagen, Denmark. January 1, 1999 - updated with  links on 26 February, 1999. Write to T.C.Bøg-Hansen,