Urologic Oncology

SLIDES & TRANSCRIPTS
Friday, December 5, 2003

Does Prostate Cancer Have an Infectious Etiology?

Elizabeth A. Platz, Sc.D., M.P.H.

Slide 1:

Thanks very much for that introduction. I am going to address a topic that was very popular in the 1970s in the epidemiologic literature. Then it fell by the wayside for quite a number of years and then I guess in the past 5, 7 years the topic has re-emerged, as Dr. Klein mentioned, in part because of some of the genetic research that is being conducted.

So, the question is does prostate cancer have an infectious etiology? We need to set the timer, too, for my talk, so we can keep track.

Slide 2:

So, we know that a number of cancers do have an infectious etiology, stomach cancer, H. pylori for example, liver cancer for hepatitis B and C viruses, also, cervical cancer, anal and penile cancer has human papilloma virus as a very important etiology agent.

The first three cancers that I mentioned are characterized by the inflammatory process because of those infectious agents and we believe that this is the pathway by which carcinogenesis occurs in those organs.

For cervical, anal and penile cancer inflammation is probably not the pathway associated with the infectious agent. Instead the HPV encodes oncogenic sequences, the E6 and E7 proteins that interfere with RB and p53.

Slide 3:

So, here is a very simplified model of infection and inflammation in the etiology of cancer.

So,you can imagine that in the setting of chronic inflammation due to infection there are macrophages, other types of immune cells that move into the environment, begin to release reactive oxygen and nitrogen species. Those reactive agents cause mutations, initiating events increasing the risk of tumorigenesis.

You might, also, imagine that the infectious agent itself or the inflammatory process may cause cell damage. That cell damage may be so extreme as to require replacement hyperproliferation. So, the cells are dead. You need to replace those cells. Therefore, when you look in tumors that have an infectious etiology you would expect to find regenerative lesions.

In the context of replacement hyperproliferation errors can occur. New mutations might occur. They become fixed. These cells may gain some sort of growth advantage and again form a tumor.

So, we would like to consider this model for infectious agents in the etiology of prostate cancer.

Slide 4:

If this model is true for prostate cancer you would expect to find that inflammation is very common in the prostate. You would expect to find regenerative lesions although you would not necessarily expect to find the infectious agent itself. There may be a hit and run style event.

Slide 5:

Well, is there evidence for chronic prostatic inflammation? Indeed there is. Quite a number of studies now have shown that inflammatory infiltrates are common in TURP specimens, biopsy specimens and in prostatectomy specimens.

What we are not sure of yet is how common is inflammation in the prostate in normal men and this is something that we are going to be addressing hopefully in the PCPT.

Slide 6:

So, here is the lesion that Dr. Walsh mentioned. This is proliferative inflammatory atrophy. Focal atrophy has been long recognized in the prostate. It wasn't until more recently when Angelo Demarzo began to really characterize this lesion that people became much more interested in understanding the role of inflammation and atrophy in the etiology of prostate cancer.

So, to give you a little bit of description, and this is all based on Angelo's work, here is the normal prostate. Notice abundant papillary enfoldings. There is a single cell layer of columnar epithelial cells lining the lumen,

In PIA lesions instead there is a loss of papillary enfolding and these cells rather than being tall columnar become more cuboidal in nature and there are a number of other attributes to these lesions, but this is just a couple of characteristics.

Slide 7:

Is this lesion a regenerative lesion? Possibly so, but you notice after staining for Ki67 there has been a number of the cells that are highly proliferative. So, even though the term is atrophy these cells are not atrophic in that sense.

Slide 8:

These lesions exhibit inflammatory infiltrates. Shown here in the lumen of the PIA lesion are numerous macrophages.

Slide 9:

We, also, find PIA lesions with scar tissue surrounding the lesions as well as abundant lymphocytes.

Slide 10:

So, what are the hypothesized causes of PIA lesions? Well, the one that we would like to highlight at the moment is response to infection. You might hypothesize that there are a number of possibilities for causing PIA lesions including things like cell trauma due to oxidant stress, hypoxia-related changes, autoimmunity as well as hormonal changes with age.

Slide 11:

So, where do these lesions fit in the development of prostate cancer? Well, we are not sure. One possible place they might fit in the natural history is as an early step in the pathway, not to imply that all cancers must pass through this step but perhaps some do.

Instead maybe there is some other set of changes that occurs and this lesion may merely represent an intraprostatic milieu that is more favorable to the development of prostate cancer.

Slide 12:

So, how has this hypothesis been evaluated so far in the epidemiologic literature? Well, there are a number of studies that have evaluated sexually transmitted infections and there are a number of types of different studies. There have also been studies of clinical prostatitis and then more recently studies of polymorphisms in genes that are involved in the response to infection.

Slide 13:

So, here are a few candidate agents that have been evaluated over the years. Some have been evaluated very systematically. Others have been suggested and maybe looked at just pathologically at this point.

So, this is the agent for syphilis, gonorrhea, and for Chlamydia. These have received the most study so far along with human papilloma virus. You can see that most of these agents though are probably not that common in the general population and would be unlikely to account for anything more than a small percentage of cases. However, some of these viruses are very common and could account possibly for the very common occurrence of prostate cancer.

Slide 14:

This is an ecologic study and it is what I like to show although from an epidemiologist's perspective it is probably not the optimal study design but it is certainly intriguing.

These are data from Denmark showing the occurrence of gonorrhea in the population from 1895 through 1931. Displaced from the other X axis is the occurrence of prostate cancer by year shifted by a number of years and what you can see is that the occurrence of gonorrhea in the population parallels the increase in prostate cancer in that same population, okay, intriguing data although as I said from an epidemiologic perspective this is not the optimal study design.

Slide 15:

There are a number of studies as I mentioned that were conducted in the 1970s looking at STDs and prostate cancer. Those have possibly some limitations. So, instead I am going to show you much more recent studies that were conducted using the population-based case control study design.

At the present there are no prospective studies of self-reported STD history, but there are some that are ongoing. This is a study conducted by Richard Hayes at the National Cancer Institute and what he observed in a very large population-based, case-controlled study which included both black and white men was that men who reported a history of gonorrhea had about a 50 percent higher risk of prostate cancer and as the number of episodes increased so did the apparent risk.

There was also an increased risk of prostate cancer in men who reported that they ever had syphilis.

Slide 16:

There are some very smart epidemiologists out there who like to put all the data together to try to form an overall opinion about our literature. There is a woman who is well known for putting together these kinds of studies called meta analyses and what she did is she took all the epi literature and she reported that STDs are associated with about a 44 percent higher risk of prostate cancer in these studies and you can see the breakdown by syphilis and gonorrhea which as I said are the two most commonly studied self-reported exposures.

Slide 17:

After that study was published there was another population-based case-controlled study that also looked at self-reported history of gonorrhea and syphilis and this study saw approximately the same estimates of effect for prostate cancer, about 1.5, 1.6. It, also noted an increased risk for individuals who reported a history of genital herpes although this was not statistically significant, no association for genital warts and interestingly an inverse association for Chlamydia although not statistically significant.

Slide 18:

So, what are the issues in these epidemiologic studies? They are all retrospective case-controlled studies. While valuable, they do have some limitations including for example if you have prostate cancer you may be more willing to talk about an STD history as you are trying to seek a reason for your diagnosis.

Except for the population-based case-controlled studies most of the early studies had small numbers of cases. The assessment of STD history in the early studies was unblinded. Also, it is unclear whether STDs and I, also, call STDs STI for infection may just merely be acting as a surrogate for the true underlying risk factor, for example, other STDs, androgenicity or life style risk factors for prostate cancer.

Slide 19:

There have, also, been a number of studies that measured circulating antibodies against STDs. Most of these are not prospective. So, the antibodies were measured at the same time that a man had prostate cancer although there is a couple of prospective studies now and as you can see the results are pretty variable.

There are no studies on gonorrhea because individuals do not mount a strong antibody response. There is not a good test for antibodies against gonorrhea. So, some intriguing work that is about to be published, this is an abstract is this virus. This is a herpesvirus that is associated with Kaposi's sarcoma.

So, it is not very common in the general population in the US . It is common elsewhere in the world but they did find an increased risk of prostate cancer for having antibodies against this virus in two different places, here in the US as well as in the Caribbean .

Slide 20:

There have been a number of studies of HPV in prostate cancers. Again, this is antibodies against HPV and again the results are somewhat all over the place. They are either null or suggestive of an increased risk for certain serotypes, perhaps the most commonly associated with cervical cancer, 16 and 18, possibly also 33.

Slide 21:

Dr. Hayes along with his study of self-reported STDs looked at circulating antibodies against syphilis and HPV 16 and what he found again was a suggestion of an increased risk for having been exposed to syphilis and also possibly an increased risk for HPV 16 but this was not statistically significant.

Slide 22:

This is a prospective study that was conducted in Europe . It is not very large but again it is prospective and this study noted increased risk for prostate cancer for HPV 18 and 16 but not 33 or 11 nor did they find an association for Chlamydia.

Slide 23:

So, what are the issues in the serology studies? Most of the populations that have been studied to date with the exception of the Hayes study were low STD populations. So, it is much more difficult to detect an association when the exposure has a low prevalence. There is also the issue of waning antibodies over time. These don't necessarily persist and if you expect that the exposure occurred early in life but you are measuring in middle age it is possible that antibody levels have gone down.

Another question is whether circulating antibodies correspond to having had the infection within the prostate. The same issue that arises for the self-reported studies is it is unclear whether measurement of antibodies merely corresponds to some other risk factor that co-occurs and finally some of the studies use men with BPH as the control group and it is unclear whether BPH has an infectious etiology.

Slide 24:

There have been a number of studies done largely by pathologists rather than epidemiologists looking at whether these agents can be detected in tissue either in normal tissue or within prostate cancer.

Also, we do know that several STDs can cause prostatitis. So, these are clues but like I said the agents for the most part in tissue itself have been inconsistently found although one point that I think is kind of interesting is that some studies now have shown that perhaps there is bacterial infection but the nature of the bacteria cannot be identified, and interestingly the detection of evidence of bacteria appears to be correlated with the extent of inflammation within the prostate.

Slide 25:

So, what are the issues in the tissue studies? Well, typically these have been small studies. The populations studied have low STD prevalences. There may be contamination from the urethra or rectum depending on how the tissue is procured. There is also the issue still of persistence versus clearance. You don't necessarily expect to see the agent there that many decades after the infection. Also, it is important for some of these agents to make sure that a very sensitive method of detection of the agent is used and also at this point we really do need to have tissue from men without prostate cancer to see which agents truly do infect the prostate ordinarily.

Slide 26:

Moving on to prostatitis the same person who put together the other meta analysis performed this one and what she found was a summary estimate of about 1.6 for clinical prostatitis.

Slide 27:

Again, there are several possible issues, the issue of recall bias, the mixed unknown etiology of prostatitis; much of it doesn't even have an inflammatory component to it.

Ascertainment is difficult in epidemiologic studies, ascertainment of prostatitis is quite difficult and certainly we won't be able to ascertain asymptomatic inflammatory prostatitis in men in epidemiologic studies.

The other issue is being able to capture duration and treatment for this condition. Plus there is the issue of prostate cancer detection bias and men who seek treatment for prostatitis may be much more willing to go back to their urologist for screening for prostate cancer.

Slide 28:

So, I am not going to discuss the next two items. I am going to leave it up to Dr. Zhou and Dr. Klein tomorrow to discuss genes involved in the response to infection.

I just want to throw this out there that the idea of infection is now being further supported by the identification of these genes and changes in these genes that render these proteins less likely to respond to infection I think really is an important finding.

Slide 29:

[No text is associated with this slide]

Slide 30:

[No text is associated with this slide]

Slide 31:

[No text is associated with this slide]

Slide 32:

So, in summary the epidemiologic literature hints that infection and associated inflammation are linked to prostate cancer, but we certainly need much more work. We need prospective designs with good exposure characterization.

We certainly need appropriate control groups for the genetic studies. We need to ensure that the underlying allele frequencies and the controls really correspond to the population that gave rise few the cases.

We, also, need to make sure that the control group has an equal opportunity for detection of prostate cancer.

Finally we do also need prostate tissue from men without prostate cancer to be able to identify what those agents are that might increase the risk.

Slide 33:

So, a number of things need to be considered, persistence versus clearance leading to inflammation, age at infection, extent of immune response. Prevalence of infection must be high in order to have much of prostate cancer to be accounted for by these agents. We need to account for things like what is the explanation for racial variation. Could it be differences in infectious agents prevalence and also we need to determine whether infection is at all involved in the etiology of PIA lesions?

Slide 34:

So, I would like to acknowledge my collaborators on a planned study and an ongoing study of STDs and other infections in prostate cancer and I would also like to thank Dr. Walsh who has really created the environment for epidemiologists to work alongside with clinicians and laboratory scientists on these types of studies.

Thanks.