Pharmacogenetics of testosterone action

view presentation | view abstract | view curriculum vitae | print this page 

Professor Eberhard Nieschlag and M. Zitzmann; Institute Of Reproductive Medicine Of The University, Müenster, Germany

The special topic I am talking about today is the pharmacogenetics of testosterone, which I think is a slowly emerging field, of which we have to take into consideration, probably in the future, when we deal with testosterone treatment. Before I do this I would like to acknowledge my co-workers Dr. Michael Zitzmann, who did most of the clinical work I’m relating to, and Professor Gromoll who is head of our molecular biology laboratory, and in charge of the work in this regard that I am showing. 

When we look at testosterone therapy from an historical point of view, or from what we have available today, you see we have quite a number of preparations, in particular, they came in the last ten years. They show improvement in terms of pharmacokinetics as Dr. Kliesch has shown in the very first talk today. The last one, the buccal testosterone is coming, I think it’s not yet licensed but it is supposed to be on the market this year. So, the preparations on the right side are those with the better kinetics.

As Dr. Kliesch also mentioned already, the principles of clinical use of testosterone still remain the same although these principles were drafted about 12 years ago. First of all to aim at physiologic serum testosterone profiles, and I think within the new preparations, we have been achieving this goal. Secondly to use natural testosterone because testosterone is working on all organs that have androgen receptors. And all organs more or less have androgen receptors, in different densities, but they do.

Testosterone has safety mechanisms, so to speak, although ubiquitously represented in the organism, it is working with another metabolide to make sure to really work as testosterone, and that is DHT, or to work together after the conversion to oestradiol. It’s important to use natural testosterone so that it is converted into these two metabolites.

What we are learning now is that there may be a third element that is important for the principles of the use of testosterone. Let me start out with the observation in Klinefelter patients in whom you see that on the baseline the total testosterone levels are shown and on the Y axis you see the haemoglobin concentrations, and there is some correlation between the circulating testosterone levels and the haemoglobin concentrations. That seems to be not a linear correlation but rather a log correlation.

If we turn, for illustration purposes, to a study from Bhasin, where he gave increasing doses of testosterone, of testosterone enanthate, to normal men who had suppressed Leydig cell function, suppressed by the application of GnRH agonists. You see clearly that there is a dose response in the generation of fat free mass in relation to the testosterone levels, or rather in relation to the testosterone dose given. But if you look carefully on the X axis you see that there is no linear increase in the doses, it is a doubling of the doses, and thereby you get a linear increase in the changes of fat free mass. This indicates again that there is no clear linear dose response relationship between testosterone administered testosterone serum levels, and the biological effects.

It appears that there is some other modifying, mechanism, active, that is modifying the testosterone action. This is what I am concentrating on in my talk here. We, of course, think immediately about the androgen receptor, because the androgen receptor is the mechanism that is transferring testosterone action to the cell, to the nucleus, and finally translating into a biological activity. As you see here, the androgen receptor can be modified by co-activators and co-repressors, which may play a role in the testosterone action, but it could also be the androgen receptor gene as such.

We know for quite some time that there are androgen receptor mutations, which result in dramatic effects of testosterone action. If there is complete androgen insensitivity due to receptor mutations, it results in testicular feminisation. If there are less severe mutations in the androgen receptor, the result is partial androgen insensitivity and we get patients with Reifenstein syndrome, etc.

But I am not talking about these tremendous changes, I am talking about more subtle changes that may modify the testosterone action through the androgen receptor. As you see in this slide, there is the X chromosome, where the androgen receptor gene is located. You see that the androgen receptor consists of 8 exons, and exon 1 is the one we are concentrating on today, because it has a variable length due to the CAG triplets, which may be incorporated into exon1. The CAG triplets in exon 1 of the androgen receptor are those which are responsible for subtle changes in the testosterone action through the androgen receptor. I’m going to show you examples now of how this works, and in the end we will come back to testosterone therapy.

If the CAG repeats - extend beyond the number of 37, the result is Kennedy Syndrome. The spinal bulbar muscular atrophy is a very severe disease, but even when you take the normal distribution below 37 repeats, namely, between 9 and 37, you get differences which can be summarised that the lower the CAG repeat number is, the higher the androgenicity of the receptor of the testosterone action is. That is symbolically shown by this graph here.

The first study which applied this to a clinical situation fits well to the topic we are discussing today, and this is the number of CAG repeats in relation to prostate cancer. This is a study published in 1997 by Giovanucci, and you see here that in a large number of patients, in particular, those who developed an advanced stage, they have a higher odds ratio for prostate cancer, for more aggressive cancer, in the metastatic cases 2.44 if they have shorter CAG repeats. So, obviously in the severity of the prostate cancer, the CAG repeats play a role. There are studies which negate this correlation, but about 3,000 cases have been investigated today and the majority shows that this correlation may exist, and that the shorter CAG repeats go along with a higher incidence of prostate carcinoma, which of course may be important for the treatment of androgen deficient men.

Now, there are other areas where the CAG repeats play an important role. We have investigated fertile men, and I emphasize fertile men, most of them proven fathers, not patients attending the infertility clinic, they were proven fathers, and you see here that in these normal men where we all know there is wide variation of normal sperm counts all above 20 – 40 million, that the CAG repeats correlate quite nicely with the sperm concentrations. You see those on the left with the smaller number of CAG repeats have the highest sperm concentration, whereas those on the right with the longer CAG repeats, they have the lowest sperm concentrations. There was some controversy about CAG repeats and male infertility. I think the picture there is confused because various causes for infertility are added to these normal values and then the picture gets blurred. But if you take really normal men as we have done here, you find this nice correlation. You may say thereby that those with shorter CAG repeats are more masculine, if more sperm means more masculinity. Whereas those with lower CAG repeats, longer CAG repeats have lower androgenicity or masculinity.

Let’s go to another topic which is quite dear to some of us, many of us, and that is balding. You see here a study from Ellis, published two or three years ago, where he looked at the CAG repeats shorter than 21 in those men on the left, with the Hamilton baldness type 1, which is the complete male hair pattern, or going to type 4, where you have the beginning of balding and the more advanced male hair pattern. You see no clear difference between the two groups, but, the group with the dense hair has only 29% of men with CAG repeats shorter than 21, whereas those with the beginning balding they have a much higher percentage; half of them, who show shorter CAG repeats. So, again, maleness related to the number of CAG repeats.

If you go through the ethnic groups it all of a sudden becomes very interesting that we know the different hair pattern between Asians, East Asians, Caucasians and Africans, and these differences may reside in CAG repeats. You see on this picture here, which is combined from various studies, (including some of us, some Amerindians), that the Chinese have the longest and the Africans have the shortest CAG repeats. This may well explain the hair pattern. It may also explain coming back to the prostate carcinoma that the Africans have the highest incidence of prostate carcinoma. So, the CAG repeats play a role everywhere.

They also play a role in the HDL cholesterol in the blood lipids. Here is a study from Michael Zitzmann where he showed that the shorter the CAG repeats are, the lower the HDL cholesterol, which as you know has a protective effect. Those with longer CAG repeats have higher HDL cholesterol values, so they are better protected. They are not as "male" as those on the left side.

Now, bones again are an issue which we could discuss and I’m first showing you some data on the bone density as it develops in normal men, and in eugonadal men, and hypogonadal men. You see here a three dimensional picture where the Y axis is the bone density, and on the X axis you see age and testosterone. You see quite clearly that the threshold to hypogonadism, also in terms of testosterone levels, is also connected with the decrease in bone density, which may even go into a range that is below the fracture threshold. So it shows that age and testosterone are important determinants of bone density.

Now, if we look at the continuation of the study on the CAG repeats, the age as such, and the oestradiol are important for bone density, but the strongest correlation here could be found with the CAG repeats. I’m not saying that CAG repeats are the whole message, but in a multi-factorial area of factors, the CAG repeats play an important role, even in bones. 

Finally, if we look at the 77 Klinefelter patients, whom I have shown you at the beginning, we find some very interesting phenotype genotype correlations. You know that Klinefelter patients, excluding those with a mosaic are a very heterogeneous group in terms of clinical phenotype. And here we may have a clue to this great variation. As we are working in a centre where we get endocrine patients as well as infertile patients, we get a broad spectrum of Klinefelter, which maybe in a centre working with only one or the other aspect, do not turn up.

If you look at the upper left at the gynaecomastia you see that those patients with long CAG repeats have a higher incidence of gynaecomastia than those with short CAG repeats. If you go to the upper right panel you see that those patients with short CAG repeats have a higher chance to live in a partnership than those with long CAG repeats. If you go to the left lower panel you see that the reason for referral correlates with the CAG repeats. Those with the short CAG repeats and a partner they come for infertility, makes sense; whereas those with long CAG repeats, no partner, they have more endocrine reasons to come. They come for features of hypogonadism. Also the range of professions in those patients is correlated somehow with the CAG repeat lengths. You see those with a university degree are mainly those with a short CAG repeat, and those with the longer CAG repeats tend to be craftsmen or no learned profession. So CAG repeats all over.

Let’s turn to the treatment. We have heard about the prostate and prostate growth, and I’m showing you a study here, which Hermann Behre did some years ago while he was still in our Institute, where we could show, and these are the black dots, that patients with hypogonadism have no testosterone values, but when you treat them with testosterone, these are the red triangles, their prostate volume increases. It comes into the same range as in normal age matched controls. That is fine but if you look, there is still a big scatter among the hypogonadal men in the prostate volume finally achieved.

We looked at patients with testosterone substitution and also analysed their CAG repeat polymorphism and the androgen receptor, and asked the question whether this may play a role.

So the method was retrospective 131 hypogonadal men treated with different testosterone preparations, but you see here that the mean treatment lasted for 18 to 69 months.

The prostate measurement was done by transrectal ultrasonography.

Let’s look at the results. You see here the prostate volume baseline compared to the CAG repeats. There is no correlation. It’s a relatively low volume in most of these patients but there is no significant correlation between the CAG repeats and the prostate volume. Remember these are all patients who have low testosterone and have never seen exogenous testosterone.

When we start treating these, the picture changes. You see the red dots are those under therapy, and here you see a quite clear correlation to the CAG repeats. You see clearly that those on the left side with the short repeats attain a higher volume than those on the right side.

I think this is a very important finding which also holds true for the absolute growth, if you calculate this during the time of treatment as growth per year. You see here clearly again that those on the left side have a higher growth of the prostate than those on the right side with the longer CAG repeats.

I do think this is an important finding. If you calculate an odds ratio for this development of the prostate sizes under testosterone substitution, you will find that those with CAG repeats below 19 have a 10.4 higher odds ratio for a larger prostate, compared to those with a higher number of CAG repeats.

If we put this in a score, then of course we know that CAG repeats are not the only parameter. There are multi-factorial reasons as to why the prostate grows. First of all age plays an important role, in particular for the baseline growth, and for the prostate growth per year. The baseline testosterone is of medium importance. The initial prostate size is of course quite important for the volume to be attained. There is a negative correlation between the size and the growth, meaning that a large prostate grows less than a small prostate. But when you look at the length of the CAG repeats there is clearly a strong correlation with the prostate volume under treatment and the growth per year. Duration of treatment doesn’t count so much if it’s over one year, and that absolute testosterone levels under treatment are also not so important. The receptor seems to play a more important role here.

This is the last example I’m going to show you, where the pharmacogenetics of testosterone treatment may be very important, is hormonal male contraception. Where we give testosterone alone or in combination with mainly gestagens to suppress the pituitary and to suppress spermatogenesis so that the volunteers may get azoospermic. All those working in the field know the big problem of responders and non-responders, whereby we mean by responders those who react to the treatment with azoospermia and those who don’t respond suppress their sperm counts to a certain level, which may not really be compatible with contraception. If you look at these volunteers, and we pooled several studies that we have done in the past, where we could get DNA samples from the volunteers, and a strong predictor of whether someone will respond with azoospermia or not, are the LH levels. You see here the three groups we formed of those who have bad suppression, those who have medium suppression, and those who have very high suppression. You see particularly in the group with the low suppression of LH the length of the CAG plays an important role. This is a good example how the pharmacogenetics of testosterone may affect our day-to-day work, right into the future when male contraception as a hormonal entity may become available.

In conclusion we can say that the CAG repeats polymorphism of the androgen receptor gene exerts pharmacogenetic influence on the effects of testosterone at least in prostate growth and suppression of spermatogenesis. I guess we will see more effects in the future.

If we now want to develop a hypothetical model for the testosterone action, then we know that the action of testosterone from hypogonadism to true normal eugonadal values, is not a linear one, it is curved as shown here. But there may be a difference between those with short and those with long CAG repeats. The red curve shows you those men who have shorter CAG repeats who are more sensitive to testosterone, and react more with androgenicity, so to speak. Those with the lower are those who respond less to the same testosterone dose.

If we want to show this graphically I would like to conclude with this cartoon where the scale for androgenicity is mainly determined, as we all know, by total testosterone, but we have learned that the SHBG on the left side is of great importance, modifying testosterone action in circulation. Now we know that the length of the CAG repeats in the androgen receptor on the cellular level is also of importance, modifying testosterone action in the creation of androgenicity.