ZAP 70

Hi Gale

Dr John Byrd at Ohio State at Columbus is one of the world's foremost CLL experts.

FISH stands for Fluorescent In Situ Hybridization. It is a way of looking for specific chromosome abnormalities that is a lot easier than trying to do the karyotype, which is very difficult in CLL. In CLL the 4 commonest abnormalities are looked for by FISH - three #12 chromosomes (called trisomy 12), and bits broken off the long arms of chromosomes #11 and #13 or the short arm of #17 (known as del 11q23; del 13q14 and del 17p13).

ZAP-70 is an enzyme that is used by lymphocytes to signal messages from the outside of the cell to the inside. It is only there in the unmutated subset (the ones with the worse prognosis) Because the VH gene test is hard to do we have developed the ZAP-70 test which ban be done by flow cytometry and should eventually be available in most labs in the world. There is a paper about ZAP-70 in today's New England Journal of Medicine. Although in most cases the ZAP-70 test gives the same answer as the VH gene test, there are still some teething troubles to sort out.

Trisomy 12 is a marker of intermediate prognosis. It is not an indication for early treatment. Watch and wait is the general prescription. Indications for treatment are clinical - weight loss, fever, rapidly enlarging lymph nodes. or the development of anemia or a low platelet count. Most haematologist would also treat if the white count doubles in 6 months or less - in some parts of the world they start if the white count doubles in less than 12 months.

ZAP-70 results should be interpreted with caution. There is no validation of
the results by a quality assurance scheme, and the only results I really
trust are those done in my own lab. We know that the assay done in San Diego has
much greater discordance with VH gene status than does ours or Montserrat's in
Barcelona. I know that many commercial labs are setting the assay up, but I
don't know whether they are valid or not. However, in our experience of
patients discordant for ZAP-70 and VH genes either one may give a more accurate
assessment of prognosis. Although, there is no evidence as yet that ZAP-70
changes over time, very few sequential studies have been done.

ZAP-70 by flow is much
more variable. The three published papers all used slightly different
techniques, and several well known labs failed to get the assay to work. I am
dubious of everybody's results outside the three labs that have published.
 

In our series we have 185 patients who have had both VH genes and ZAP-70
performed. There are 17 with discordant results. 12 gave low ZAP-70 with
unmutated genes and 5 high ZAP-70 with mutated genes. All these had between 96.9% and
97.9% homology, so they may mot have been discordant at all. There was no
suggestion that the % positivity of ZAP-70 affected the concordance. Four of the
five with mutated VH genes and high ZAP-70 have needed treatment. Eight of
the twelve with unmutated VH genes and low ZAP-70 have required treatment.

You cannot make any assumptions about your ZAP-70 result without having the test done, but the question is should you have it done, and can you rely on the result?

VH gene mutations were the way that we first realised that we could separate CLL into a relatively benign subset with an average survival of about 25 years, and a relatively malignant subset with an average survival of 8 years.

ZAP-70 is a signalling protein used in T-cells, but not normally used in B-cells. However, it does seem to be used in unmutated CLL. Those of us working in the field quickly saw that it might be a way of distinguishing between the mutated and unmutated subset, but because it was an internal protein (not a surface one) and because much more is present in T-cells and there are always T-cell mixed in with CLL cells, it was always going to be technically difficult to measure it.

Ideally we wanted a Flow cytometry assay since these are the cheapest and quickest, and the companies could quickly get hold of them and cell them to labs all over the world. This proved very hard to do and many labs abandoned the task. Eventually, the labs in Barcelona and Bournemouth came up with an assay, quickly followed by a lab in Essen, Germany. Although the methods differed in detail, they were all based on the same antibody, clone 2F3.2, which was used in what is known as an indirect assay. This means that the antibody was applied to the cells first, then detected by an anti-antibody that was fluorescently labelled afterwards.

Now companies are not keen on indirect assays - too many steps, too much to go wrong. They prefer direct assays where the antibody itself is fluorescent labelled. The only trouble is, when we tried directly labelled clone 2F3.2 directly labelled, we didn't get the same results. In fact there were many false positives

Last year a paper was published on ZAP-70 in the New England Journal of Medicine by Rassenti et al on behalf of the CLL Consortium. The results were a bit different to what the Europeans had found. They had found a 94% correlation between ZAP-70 and VH gene mutations; but the American group only found an 77% correlation. I suggested at the time in a leading article in the same journal that this might have been because of a difference in methodology. They used a directly conjugated antibody, but it was a different antibody, clone 1E7.2 coupled to a new fluorescent dye, ALEXA-488.

Other labs around the world have attempted to repeat the work of the CLL Consortium. So far keeping my ear to the ground I hear that difficulties have arisen. with different results using this assay on the same samples that were used for the original European assays.

Clearly, this matter needs resolving and until it is it would be unwise to rely on ZAP-70 results. This is especially so for the conjugated antibodies used by commercial labs. At the moment VH gene mutations seems a more reliable assay even though it is not so readily available.

I am sure you know the story of the drunk who was searching for his lost coin under the lamppost, not because he lost it there, but because the light was better. The ZAP-70 test may be easier to do than VH gene mutations, but until we are sure what it means we would do better to stick to the harder test.

I have to emphasise that ZAP-70 is still an experimental test, and whether it is believable depends on where it was done and how good was the quality assurance. Most patients with stage 0 disease who are CD38 negative can look forward to a long survival even if they are ZAP-70 positive, but immediate treatment is not warranted. The management of your condition should be watch and wait. It is most unlikely that you would have either p53 abnormalities or a deletion of 11q.

I think it would be wise to gauge the pace of your disease over a period, and wait a while before you have further prognostic tests.

The development of the ZAP-70 assay by flow cytometry has been difficult and at least three methods have been published. I have been anxious about this test being adopted generally until we have some harmonization of methods and demonstrated reproducibility.

To this end a meeting took place in Paris last week, where all those in Europe and America who have an interest in this test were able to meet and share technical details. The conclusion was that we should all try and concentrate on a single method and become proficient in its use. The method is basically the one developed at San Diego by Laura Rassenti and Tom Kipps and marketed by Caltag. The advantage of this test is that it uses a directly conjugated antibody and therefore makes the test easier to do in a routine laboratory. Although one French group found that this did not perform well in their hands, both the Bournemouth and Barcelona groups found it to be consistent with their own methodologies (Orchard et al and Crespo et al.).

The UK National Quality Assurance Scheme in Sheffield will oversee sending out samples to participating laboratories in Europe to ensure that results in different laboratories are consistent. Such things as type of anticoagulant, storage and transport will be investigated.

It is clear that this test is not a complete surrogate for VH genes. A positive ZAP-70 correctly predicted VH gene mutational status in just over three quarters of cases (77%) in the Rassenti paper. This is not very much better than CD38 (70%). In our hands the Caltag method does a little better at 84% concordance. For the moment I would still prefer to know the results of VH genes, CD38 and FISH as well as ZAP-70.

In one of the e-mails I sent about ZAP-70 when I had the flu, I gave a false impression and I want to correct it.

Flow cytometric testing for ZAP-70 is not easy. Three labs have developed three different assays. In lots of ways the assay developed in San Diego is the most exciting of the three because it is has fewer steps, and it seems to give better prognostic information than VH gene mutations, in that patients who are VH gene mutated, ZAP-70 positive have a worse prognosis than patients who are VH gene unmutated, ZAP-70 negative.

If I were a commercial lab, this is the one I would latch on to.

BUT so far we have just one paper published using this test and we need to see independent confirmation of the results. This does not mean that I doubt the quality of the work of the San Diego laboratory, far from it. I know that Laura Rassenti and Tom Kipps are most careful workers and highly expert in the field. If I had CLL, Dr Kipps would be one of the first people I would call.

In Germany recently I attended a meeting of ERIC which is the European Union's rather feeble attempt to duplicate the CLL Consortium. We were trying to standardise ZAP-70 testing by Flow. One of the groups present from a well known expert lab had tried to duplicate the results of the New England Journal paper and found slightly different answers. Now that does not mean that San Diego is wrong and Europe is right, nor the other way round. What it does mean is that we are dealing with a complex problem and we should not be relying on a single test to tell the future.

At the moment ZAP-70 testing by FLOW is still under development. If you have had your ZAP-70 tested in San DIego, or in Barcelona or in Bournemouth, you will have had the authenticity of the test validated by RT-PCR and/or Western blotting. The test will mean what it says. If you have had it tested in a commercial laboratory, they may perhaps have validated their result, but I could not guarantee that.

Nobody claims that ZAP-70 is an exact surrogate for VH gene mutations. Different labs find different degrees of separation. Whether this is due to the different assays or to differences in the group of patients studied is not yet clear. More work is needed.

I know patients are impatient with the slow pace of science. If I had the disease, I would be. Sometimes we rage at the FDA because they are slow at licensing new drugs. But in medicine, wrong decisions can kill people.

I believe, and I am sure Dr Kipps does too, that the place for ZAP-70 testing is in clinical trials. The same is true for VH testing and FISH tests. We need to evaluate the performance of these prognostic tests in prospective studies.

People ask me whether I would treat someone with minimal CLL who had unmutated VH genes, was ZAP-70 positive, CD38 positive and had del 11q23 by FISH. I answer, yes: in a clinical trial. The history of medicine abounds with tragedies cause by people who thought they could second guess the future.

The new prognostic markers look like being very useful, but we have no data on what happens if we treat patients on the basis of the markers. That's why we need controlled clinical trials.

27 April 2005

ZAP-70 positivity.

There are three current methods for measuring ZAP-70 by Flow. The two indirect methods Crespo et al and Orchard et al have different upper limits of normal at 20% and 10% respectively. The direct method of Rassenti et al has 20% as the upper limit of normal. The normal range was determined by each lab by comparing positive and negative controls and those were the figures arrived at. The figures given take into account the various artefacts of the methods. There is no 'no-man's-land' Those who are 55% positive are not worse than those with 25%. It should be regarded as an on-off switch.