Epidemiology and Hematopoiesis and Apoptosis by Martyn Smith, PhD

SLIDES & TRANSCRIPTS
October 30-31, 2000

Epidemiology and Hematopoiesis and Apoptosis

Martyn Smith, PhD
University of California

Slide 1:

DR. SMITH: Thank you very much, Michelle and thank you for inviting me to come here. I should first of all point out that I am not an epidemiologist and that I am going to be talking really about speculative ideas and provocative ideas about what actually causes MDS and what I am going to try to convince you of in the next 10 minutes is that we should really look at the diet as a cause of MDS and possibly of leukemia. What I am saying here is that diet and genetics are possibly the two main factors which are responsible for de novo MDS with alcohol, occupational exposures and smoking also contributing, and these dietary factors I am going to focus on are high protein and phenolic content in the diet, high calories and low vitamin intake. I am also going to touch briefly upon some of the problems we have once we have developed these theories in trying to study them with MDS which are very similar to the studies that you are having problems with in that you have no models. With epidemiology you have a very difficult disease to study because you are having difficulty classifying it. When you want to study diet you have to ask questions like what did you eat in 1992, which is pretty difficult for you to answer. There is no real incidence data on MDS. Classification and diagnosis problems make it very difficult to study from an epidemiological viewpoint.

Slide 2:

Why should we think about the diet? You shouldn't be really too surprised that the diet is an area to look at. This is the causes of cancer death, the famous Doll and Pitot chart, with smoking responsible for about one-third of all cancers and this large green one here is diet, with poor diet being responsible also for about one-third of all cancers. So if you look at what may cause MDS, it is tobacco, alcohol and diet as the main factors with perhaps infection and viral agents also playing some sort of promotional role.

Slide 3:

What do we know about what causes MDS now? We know radiation, alkylating chemotherapy and this compound, benzene, are known causes of MDS. From the epidemiological studies that have been reported there are also increased odds ratios for occupational exposures to halogenated organics, metals, hydrogen peroxide, arc welding, et cetera. But for the majority of cases of MDS these occupational exposures, pesticides, and so on are very unlikely to explain the disease. Also, most of the people who have de novo MDS have not received radiation of any significant levels or chemotherapy, and they may have been exposed to low levels of benzene in the environment. But I have been working for many years on benzene and my studies on benzene really led me to think about what causes leukemia and MDS if benzene is not doing it.

Slide 4:

A little bit about benzene and that is that benzene is metabolized to its two primary metabolites, phenol and hydroquinone and the pathway that appears to be important for benzene toxicity and leukemia induction is the transport of hydroquinone to the bone marrow, its activation to a quinone producing genetic damage and toxicity. Your main protection against this toxicity appears to be an enzyme called NQO1 or NADPH quinone oxidoreductase.

Slide 5:

Together with Nat Rothman of the National Cancer Institute and colleagues in China, we showed that susceptibility to benzene hematotoxicity was related to a polymorphism in this enzyme NADPH quinone oxidoreductase or NQO1 and that people who lacked the activity of this enzyme were twofold to threefold more at risk of getting benzene toxicity than controls or people who had the normal levels of the enzyme. This suggested to us that maybe NQO1 is important in other forms of leukemia and hematotoxicity and

Slide 6:

together with Richard Larson's group, Michelle LeBeau and others in Chicago, we went on to study a group of therapy-related AML patients and found that there was also an association with this polymorphism and NQO1. My former grad student, Joe Wiemels, working with Mel Greaves in London, then showed that infant leukemias with T11 Q23 were also related to this inactivating polymorphism and most recently in a very large study of de novo leukemia we have shown together with collaborators in Leeds that this polymorphism is related to de novo leukemia.

Slide 7:

. I will tell you a little bit about that study, and show a little bit of data. This is a large case-controlled study of leukemia in Northern England with 493 Caucasians diagnosed with acute leukemia and 838 matched controls. Most of them were AMLs and ALLs. We have analyzed various polymorphisms in this population, one of them being this polymorphism in NQO1 and what we found was that the odds ratio for the low activity of NQO1, the heterozygous and homozygous genotypes produced an odds ratio of about a 50 percent increased risk of getting AML and especially in those harboring translocations 821 inversion 16 most notably have an eight-fold increased odds ratio, 5q minus and 7q minus also, an increased odds ratio although not statistically significant. This inversion 16 and these other relationships also held up in another large case series that we looked at. In two large populations of cases we have been able to show this low level, low activity of NQO1 is related to the induction of AML.

Slide 8:

What does this data suggest to us to speculate about a little bit? It suggests that compounds which are substrates for NQO1 all cause oxidative stress and NQO1 defends you against oxidative stress or causative factors in producing leukemia and perhaps MDS. What are substrates for NQO1? They include benzene and its metabolites, phenol, hydroquinone, various other compounds, possibly also flavonoids which are very common in our diet and NQO1 is also going to protect you against oxidative stress generated by things like inflammation which Neal Young was just talking about or low antioxidant intake or certain environmental chemicals which you are exposed to. NQO1 may be working through this oxidative stress mechanism and its defenses.

Slide 9:

We know that benzene causes MDS and leukemia, but it is very unlikely that benzene is really responsible for many of the leukemias or MDS cases in the general population. That is because exposures in the general population to benzene are very low at the level of 1 to 5 parts per billion in air, and we do have, however,

Slide 10:

large amounts of phenol in our urine and also of hydroquinone and this phenol catechol and hydroquinone comes mainly from the diet and not from benzene. This led to me to thinking that maybe the diet and these dietary phenols and quinones are actually what is important, not the benzene that is the common pollutant. These phenol, catechol and hydroquinones are common dietary constituents. There are very widely varying background levels of these in the general population, and these varying levels stem from differences in dietary ingestion, medicines like Pepto Bismol and chlorceptic(?) and most importantly the activity of your gut flora and the make-up of your gut flora. Since excess protein in the diet, things like tyrosine and phenylalanine, these amino acids are converted to phenol by these gut bacteria, so, phenol producers in the gut may be very important.

Slide 11:

We have extremely highly levels of phenol, catechol and hydroquinone in our everyday urine. It is highly variable and in order to get these levels in our urine you would have to be exposed to about 1000 times more benzene than we are typically exposed in every day. It is not coming from benzene.

Slide 12:

It is coming from somewhere else, and it is mainly coming through hydroquinone from something like arbutin, which is present as a glycoside in many things like wheat and pears, and so our main intake of hydroquinone is from these types of foods.

Slide 13:

But phenol which is then converted to hydroquinone is coming mainly from gut bacteria and the breakdown of amino acids from gut bacteria and very interestingly if you have a large number of bacteroides in these phenol-producing bacteria you produce a lot of phenol. You have a high level of phenol in your urine whereas if you have a lot of lactobacillus and clostridium which don't produce phenol then you have very low levels, and interestingly these are associated with meat intake and this is not. This is associated with breast feeding which is protective against leukemia and these are found from meat intake which is also related to leukemia in epidemiological studies. We think this production of phenol by gut bacteria could be very important in producing it.

Slide 14:

The theory goes like this, which will be published in Leukemia in the next month or so, the paper we have in press. I have some copies with me that phenol and hydroquinone in the blood come mainly from arbutin in the diet, a little bit from benzene, a little bit from benzene in cigarette smoke, from medicines such as Pepto Bismol but most importantly from protein and gut flora activity. These travel to the bone marrow where they produce many of the same effects that benzene produces because they are responsible for benzene's effects, oxidative DNA damage, chromosome rearrangements, inhibition of TOPO-2(?) altered hematopoiesis and selection of a leukemic clone. We hypothesize that this could lead to some forms of leukemia and MDS.

Slide 15:

Now, this is very difficult to actually prove or study because of a variety of factors. It is very difficult to look at the diet retrospectively. How do you assess differences in people's gut flora? How do you measure what their gut flora is or what their gut flora was 10 years ago, and there has also been a suggested role of viruses by Azro Raza, the cytomegalovirus. That needs to be looked at and a variety of other things need to be looked at. We have a lot of problems not only looking at the hypothesis but also some epidemiological, methodological problems. There is really no incidence data on MDS. How do you identify patients early enough? It makes this very difficult to do a population-based study, which is what you really need to do. There are problems in diagnosis and classification at different clinical centers. This makes it very difficult to do a study, and as has been mentioned this morning there are also likely to be different types of MDS or subtypes of MDS with differing etiology and in epidemiology this means you are going to be faced with a problem of small numbers.

Slide 16:

What can be done? I would just like to leave you with some of my ideas of what could be done. I think what needs to be done is a very large study of MDS and AML in patients in the major clinical centers which capture the majority of patients in this specific area. By doing this you would get almost a population-based study. Then you would have to decide how you are going to select controls, and it is important to study both MDS and AML together because we may learn a lot about why some MDS progress to AML and others do not, and perhaps I would add to this, since I saw Neal Young's talk a few minutes ago, that we should add anemia and PNH into this also. Then you develop a sophisticated questionnaire and analytical techniques to test these new hypotheses and you perform sophisticated biological sample processing on all these subjects to allow future genomic or genetic analyses, analyses of gene expression and future proteomic analyses, so create a tremendous sample resource for analysis, and this will take a multimillion dollar large systematic effort, but I think it is the only way that we can really get to causation with this particular disease. Thank you. (Applause.)