At A Glance
Name: Paul J. Norman
Title: Postdoctoral Researcher, Stanford University, in lab of Peter Parham, professor of Structural Biology. Lymphoma Research Foundation fellow
Background: Worked in South Thames Transplantation Laboratory, at Guy’s hospital in London, before going to Stanford.
Education: PhD, Immunogenetics, from King’s College London
Paul Norman, a second-year postdoc at Stanford University School of Medicine, recently won a $105,000, two-year fellowship from the Lymphoma Research Foundation to study whether genotyping may help physicians better select lymphoma patients for bone marrow transplants (he is also funded by the Leukemia and Lymphoma Society).
Matching bone marrow transplant donors and patients is currently difficult, as human leukocyte-associated surface antigens on marrow cells differ widely from person to person. Plus, recipients will reject donor marrow that is not HLA-matched to theirs. According to the US National Cancer Institute, between 30 and 40 percent of patients have an HLA match with a sibling or parent, and the chances of a match with an unrelated person are much slighter. Even with an HLA match, rejection sometimes occurs.
Though some HLA-testing products or services currently exist on the market — most recently, Orchid Biosciences sold its LifeMatch platform to Tepnel [see 11/13/04 SNPtech Pharmacogenomics Reporter] — Norman hopes that his work will lead to the development of new tests that can use SNPs to match donors to patients. SNPtech Pharmacogenomics Reporter recently spoke with him this week about the research and how he hopes to accomplish this goal.
Can you tell me a little bit about the project for which you received the fellowship?
What we’re doing is we’re focusing on a newly discovered set of genes that are quite important in innate immunity. Essentially these genes are highly polymorphic. The HLA system includes the major transplantation antigens. They are highly polymorphic and they are the reason why people reject or accept bone marrow or other transplants. It has become apparent there is this other set of genes, the products of which interact with HLA. And they can control a person’s innate immunity and other immune reactions.
So they sort of form the first line of defense against invasion — invading organisms, and also against foreign tissue. So that they are potentially quite important molecules in the immune response. This is the reason why we’ve actually focused on these because they could explain why some people that we have seen in the past have rejected bone marrow or kidney or other transplants when we would expect them not to have done because of their tissue type, their HLA type. So there could be a missing link between rejection and non-rejection.
So you are genotyping marrow samples?
We haven’t got that far yet. I will explain what we are doing and why it’s difficult. The problem is, with this gene family is, the proteins are highly polymorphic — more polymorphic than most genes that we know of, but the major obstacle to genotyping is they’ve got polygenic variability. So each person can have between seven and 15 genes, they are called KIR genes … which makes it really difficult to genotype for them. So we don’t know whether one person has one copy, two copies, or no copies of any one of these different KIR [genes], and each one of these different KIR can interact with a different HLA molecule. So the major obstacle is the fact that not every gene is present in every person. So far, there’s been about 25 known haplotypes each with different gene content. So it makes it a challenging area to genotype: It also makes it pretty interesting to genotype.
What were focusing on the moment is, we know that different ethnic populations have got different sort of distribution of these genes, so one group will have a predominance of one particular gene more than another group. So that’s going to make it really difficult for bone marrow transplantation. That means we’re going to have to identify what genes and what alleles are present in all the ethnic groups before we can start looking at the matching. So at the moment we’re collecting samples from ethnic groups. We’re at the sequencing stage. It’s polymorphic enough to ... mean that sequencing is still a valid way of discovering the SNPs. There are other ways of SNP discovery, but sequencing is probably the most efficient because the number of SNPs there are validates using it. So we’re doing that to start with, then were going to look into SNP genotyping methods for a large-scale analysis.
So you are working with the Stanford Genome Center initially, using machines like ABI 3700s or the MegaBace?
I’ve [been] doing a low-scale analysis here to start with using a Beckman 8-channel and an ABI 377. I’m going to scale up the sequencing to using their [machines]. They’ve got a MegaBace and the ABI — they’ve got about seven 96 channel sequencers. So I am going to scale up to that soon and then we’re going to do SNP genotyping using a SNP-specific method such as Pyrosequencing or the Beckman Coulter methods. Those are the two methods I am looking at.
Are you leaning toward one method or the other?
The reason I am choosing those two is because the aim of the SNP-typing will be to use a system that has a moderate number of SNPs to look at but a large number of individuals, and that’s why I have chosen those two methods, because the other ones are [for] large numbers of SNPs and small numbers of individuals. So I am going to be looking at less than 200 SNPs in many thousands of people. …
The attraction of [the Pyrosequencing system] is that we get more information around each SNP. This polymorphism is focused in different areas. So if I get sequence of 30 bases around each SNP, I am going to get more information.
And also I am going to get information as to what haplotypes — whether these SNPs that are in phase or not in phase — from each given read out, which is an attraction. And also if we can look at three or four SNPs that are clustered in one 30-base fragment, then that’s going to mean that Pyrosequencing is the most optimized way of looking for those SNPs. And the attraction for the [Beckman Coulter] method is, I have not actually seen it working yet, but they claim that we can have multiplexing of up to 10 SNPs in a single well. If I can do that, then that will be fantastic.
You haven’t tested that method yet?
I am in the process of doing that. They’re doing it for me. They’re pretty helpful, actually. They’re testing out samples as we speak.
Down the line after you have established the baseline of different polymorphisms in different groups, do you see this turning into some sort of diagnostic test?
Yes. The way I envisage it at the moment … what would be a perfect thing for me really would be to make of what I am thinking of as sort of a maxi chip, which isn’t like a mini array, but say, 150 SNPs particularly focused on this one region, so that we can say from an individual what KIR genes they’ve got and what polymorphisms of those KIR genes they’ve got, and look at that in line with the HLA type of the donor, so we can use this as a diagnostic to say whether a bone marrow transplant is going to succeed or fail. That’s the goal.
How many years out do you see that as being?
The fellowship to design the system and prove it is a two-year fellowship. So, in two years’ time, I would like to start working on the chip, or whatever technology we develop at that time, which might be even better.
Do you see then going into potential partnership with anyone outside of the university?
Yes, I would definitely look towards that. That would be the way forward. I don’t think it is something I could do myself.
But also, what we are doing is were focusing on SNPs that we know or suspect to have a functional consequence, so we’ve also got people that are routinely going through the SNPs and working out what they do by transfecting them into systems. So, every sort of mutation known to man, so far, is evident in this region.
We’ve got SNPs that decide whether the gene is expressed or not; we’ve got SNPs that reduce or SNPs that raise expression levels; we’ve got SNPs that change entirely the molecule that it binds to or change the affinity of the molecule that it binds to. So it’s everything that you can think of that a SNP could do [that] is evident in this region, which makes it interesting.
How will you narrow it down to fewer than 200 SNPs?
We’ve got some redundancy between SNPs. That’s the first line. And that’s part of the purpose of doing the sequencing as well, is to show which SNPs are redundant. So we can knock some out that way. One of the other objectives was also to see which SNPs I got from GenBank were nonexistent. Because previous reports said that 23 percent were not really valid. Unfortunately all the ones I’ve looked at so far have been valid. That hasn’t helped. So this is the functional side really. If we can predict or show what a particular SNP does in a given situation, we’re going to focus on those. And then we’re going to access historical samples from bone marrow transplants and do a retrospective study.
Does Stanford have a bank of these, or how are you going to get the bone marrow samples?
We are talking to the relevant authorities about the prospect of doing that. They seem quite keen to do that.