Mark P. Sawicki, MD: Molecular Genetics of Pancreatic Endocrine Tumors (PETs)

I think it's very important for the residents in the group that you start out your research career with a clinical question.

[ Slide 01 ]   Clinical Question: Gastrinoma

I'm going to talk about a rare group of tumors and one of them is shown here. This is a gastrinoma in a patient with sporadic gastrinoma. When I joined the lab as a resident I had no idea what I wanted to study. My mentor, Dr. Pasarro, suggested I try to unravel the molecular biology of these tumors. So we embarked on a very ambitious project to understand the molecular biology of gastrinomas, and it turns out there was nothing known at the time about these tumors in part because they're rare and in part because they're difficult to grow in culture. A whole host of problems. But nonetheless we were surgeons and we were bound to do it.

[ Slide 02 ]   Clinical Questions

So the clinical questions we had was why gastrinoma? Gastrin is not made by the pancreas yet gastrinomas are categorized as pancreatic endocrine tumors. We noted clinically that some of these tumors were curable or so-called benign and some of them were malignant and what was the molecular difference between a benign tumor and malignant tumor? And finally the third question was could some of this technology, this molecular biology, be translated into something clinically useful in treatment of our patients, and in particular the growth of islets and perhaps their applications in diabetes?

[ Slide 03 ]   11q13 Deletion Analysis

Well I started out in the lab and I did what was advised to most people and that is find a basic scientist. I did that and he said "Why don't you throw your gastrinoma DNA blot in with our blots and find out what happens and see if your technology is even working". So I took some tumor and white blood cell DNA and made Southern blots. It happened to be that group that day was using a chromosome 11 probe to look at gene amplification in head and neck cancers and lo and behold, the very first thing we did worked. And we found a deletion of chromosome 11 in about 40 percent of the gastrinomas. Now it did work but I didn't recognize it worked until about 6 months later when Dr. Jim Economou was looking at some slide about 50 feet away and said "Gee that doesn't look right" and he was very insightful and got us on track and we identified that in fact on that Southern blot there was a deletion of that gene probe.

[ Slide 04 ]   Delete Chromosome 11

What we thought was going on is that there was a region on chromosome 11 that was deleted in these pancreatic endocrine tumors. When you compare normal DNA with tumor DNA on a Southern blot with a gene probe from chromosome 11, in some individuals who are heterozygous for this probe with two alleles A and B, in the normal tissue the tumor has only one allele. This change from the heterozygous state in the patient's normal tissue to the homozygous state in the tumor is called loss of heterozygosity. That change at the molecular level is a key hallmark of the presence of a tumor suppressor gene. We looked at DNA probes up and down chromosome 11 and recognized that there was an area consistently deleted in our sample of tumors around chromosome 11q13.

[ Slide 05 ]   INT2 Probe

This nicely illustrates the INT2 probe. This is an oncogene. It happened to be the first probe we used. And it happened to be heterozygous for one of our patients with two alleles, allele 1 here which gave this single band. The second allele from the other chromosome gave 2 bands and the tumor from this patient, this gastrinoma, had loss of these 2 bands on this Southern blot, indicating loss of heterozygosity.

[ Slide 06 ]   11q13 DNA Deletion

What we thought was going on was that there was a stem cell that was going to lead to gastrinoma formation. There was deletion of key critical genetic material at 11q13. At the same time, a group from the Karalinska ... had identified the region of the Multiple Endocrine Neoplasia Type 1 gene at 11q13 and we thought that these sporadic tumors were having deletion of the MEN1 gene.

[ Slide 07 ]   MEN-1 Mapping

We asked everybody who we were collaborating with what we should do. And they said the one thing that you should not do is try and clone this gene. So that's what we did.

We went along and with many collaborations at UCLA, across the country and internationally and pulled out much of the DNA in the area where the MEN1 gene was located at 11q13. As we were sifting through all of this we identified one gene that we thought was the MEN1 gene. This was a GTPase activating protein, very similar to two other known tumor suppressors, so we were ... really confident we had it. As we were going through the mutation analysis, Francis Collins at the NIH did things the old fashioned way, sequenced the entire area and went through 15 genes and eventually stumbled onto the MEN1 gene which was in fact the very same gene we were looking for. It's a relatively small gene with 10 exons and it covers about ten kilobases of DNA.

[ Slide 08 ]   MEN-1 Gene Mutations

This slide summarizes the mutation analysis in some of our tumors. The MEN1 gene, also called menin, maps to chromosome 11q13. We find a high frequency of deletions of this gene in the tumors that we've studied. We studied about 30 tumors. Here's the picture of the gene with 10 exons, 1, 2, 3, all the way up to 10. We find on mutation analysis that 8 of 23 or about a third of benign tumors have mutations. The mutations are scattered throughout the gene. In the malignant tumors three of eight had mutations. So it appears that the MEN1 gene is frequently mutated in both the benign and malignant forms of this tumor.

[ Slide 09 ]   MEN-1 Binds AP-1

Well, what's the function of this gene? Some very elegant studies from the NIH have opened some of the clues. This is the gene expressed in fibroblasts. In this case we took the MEN1 gene and attached it to a green fluorescent protein and injected it into the cells and as you can tell it actually goes nuclear. What you can't see on this slide is the outline of the cytoplasm, so this is a nuclear protein. It appears to function as a transcription factor. What it does is bind to the AP-1 transcription complex. Specifically menin in the nucleus binds to JunD, one of the immediate early genes and disrupts the transcription of genes that are regulated by JunD. Some of these are cyclins, myc, the inhibitor of metalloproteinases, collagenase, and now we heard earlier about neurotensin. So the MEN1 gene has some effect on the transcription of key genes and we don't know which ones of these are important, which ones of them regulate cancer or are responsible for the phenotype we're seeing, but we're starting to get some clues about its function.

[ Slide 10 ]   Mutations-Locations

Now there's probably more to the story than that, because as you look at the MEN1 gene, here's where JunD binds. This region up here and this region here and there are mutations that are just amino acid substitutions that are scattered throughout the gene, suggesting that there are other proteins that probably interact with the MEN1 gene and are important for its function. So one of the things we're doing, and I won't show any more slides about this, is trying to find genes that specifically interact with some of these other areas of the protein.

[ Slide 11 ]   MEN-1 Gene Function

In summary, the MEN1 gene ... definitely functions as a tumor suppressor although there's no biological data reported about that yet other than the tumors themselves. It's deleted and mutated in sporadic tumors. Also it is inherited and mutated in MEN1 patients. It functions as a transcription factor by interacting with AP-1 and we're pursuing with finding other functions for the gene.

[ Slide 12 ]   Why Gastrinomas?

So we think to a large part "why gastrinomas?" We think that the MEN1 gene is very critical in the formation of these tumors. So that was the first question, now ten years later, I think we've been able to answer on that list I started out with.

[ Slide 13 ]   Malignant Gastrinoma

Now there were a couple of other questions. How do you go from a benign tumor to a malignant tumor? And that's a common topic today as you listen to various investigators talk about the Vogelstein or Vogelgrams as some people call them, Vogelstein diagrams of progression of colon cancer. We had the same thoughts in mind about gastrinomas, that some of them seem to be benign and curable, some of them are malignant and obviously not curable. This patient unfortunately had an incurable form of this tumor. But what are the molecular mechanisms that lead to that?

[ Slide 14 ]   Delete Chromosome 1

Well another thing we stumbled onto, not by intent, but it worked out in our favor, while we were trying to clone the MEN1 gene, we did cytogenetics of tumors. We looked at them under the microscope, looked at the chromosomes, hoping to find to find a rearrangement with chromosome 11. The only thing we found in one tumor was deletion of the short arm of chromosome 1.

So a fellow in my lab was looking for a project and he decided to look at chromosome 1 probes and found that chromosome 1p36 is deleted in 7 of the 8 tumors that we had with metastatic disease and only 3 of the 21 tumors who did not have mets and the p-value for this data was extremely low, much less than .05 and this work was recently published in Cancer Research. This area is a known tumor suppressor gene region important for many types of cancer such as colon cancer and other endocrine tumors and the area containing this gene is narrowed down to about 800 kilobases completely sequenced and there are about 100 or so genes in that area. The putative tumor suppressor though has not been identified. So we've identified serendipitously, I think, a genetic change that is associated with a poor prognosis in these patients. What we need now is to identify the gene and prove in fact it's related to the progression of these tumors.

[ Slide 15 ]   Menin in Human Islets

And for the final chapter of this series of questions, I want to briefly mention something about the role of the MEN1 or menin gene in human islets and Dr. Craig Smith, a transplant surgeon, and Dr. Yoko Mullen, a transplantation researcher at UCLA, have an interest in developing an islet growth model. I've been trying to apply some of our work to that model. Their model is that they isolate human islets and grow them in culture and form a monolayer of islet-like cells that can be induced to form pseudo islets under certain growth conditions. I wondered whether or not the MEN1 gene played a role in this pathway of development. In fact this is called dedifferentiation or perhaps transdifferentiation.

[ Slide 16 ]   Fresh Islet vs. Proliferating Islets

When I did RtPCR looking at the Menin gene in fresh islets versus these proliferating islets what we found was that as the cells ramp up in proliferation, the MEN1 gene is increased in its expression. And a couple of other transcription factors important for islet development are also regulated. Isl1 is down regulated and Pdx1 is slightly up regulated in this model. And we now have NIH finding to pursue this particular question and understand whether or not these transcription factors play a key role in that islet model.

[ Slide 17 ]   Summary

In summary I think these tumors offer a unique opportunity to study tumor suppressor genes and tumor biology. Menin mutations appear to be an early event involved in both malignant and benign tumors. Deletions of chromosome 1 are a late event and probably associated with malignant progression. We're now looking for candidate oncogenes to try and complete the story of how these tumors occur and we're trying to apply some of the things we're learning in the tumors to more practical application with treatment of diabetes in the understanding islet so growth.

I don't have a slide of my collaborators but I appreciate Dr. Brunicardi's collaboration. He provided us with some of the initial islet material that we started our studies with and also numerous collaborators at UCLA. Thank you very much.