Microsoft word - yeald2 - cardiotoxicity.doc

Fair of face, foul of heart - cardiotoxicity and drug
development
Dr. Nick Miller, Beremans Limited

Drug-related cardiotoxicity is difficult to predict in the early stages of drug
development, and now may be the primary cause of drug withdrawals. Here we
examine some of the main issues pertinent to cardiotoxicity and new drug
development.

In a heartbeat

An understanding of cardiotoxicity and of the difficulties in predicting cardiotoxic
potential requires some understanding of the molecular basis of the heartbeat.
Cardiac cells spontaneously generate electrical currents which result in the repetitive
contraction of the heart muscles that keeps us in circulation. This electrical activity
results from the presence of specialised proteins (ion channels) which span the cell
membrane and which can permit the passage of ions (electrically charged
molecules). The movement of these ions can be measured as an electrical current,
which has a characteristic pattern of voltage change known as an action potential.
Action potentials produced in cardiac cells can be measured at the skin surface using
an electrocardiogram (ECG). The ECG measurement is the sum of the action
potentials in all cardiac cells. Synchronised electrical activity in cardiac cells causes
cardiac muscle contraction, and this enables the heart to pump blood.
Change of heart
In normal patients the ECG shows a predictable pattern. Deviations from this pattern
often indicate some form of cardiac insult, and may be predictive of cardiac
problems. Hence, a commonly used measure of the potential for cardiotoxicity of a
drug is the length of the interval between two well-defined points on the ECG trace,
namely the QRS complex and the T wave. This interval is known as the QT interval.

A significantly prolonged QT interval may be associated with perturbations of the
heartbeat, for example the potentially fatal cardiac arrhythmia known as torsades de
pointes (TdP). The ion channels are the most likely molecular targets by which drugs
could prolong the QT interval, by virtue of their causal involvement in cardiac
electrical activity and their exposed (cell surface) location.
A number of cardiac ion channels have been identified. There is now evidence that
the primary target for drug-induced TdP is an ion channel protein known as hERG,
and it seems that all non-cardiac drugs with TdP potential interact with hERG. The
classic example of a TdP-associated drug is terfenadine, an antihistamine which at
one time was widely prescribed for hayfever. Only after it had been extensively used
for several years did it become clear that terfenadine was associated with TdP and
sudden cardiac death. Eventually the drug was withdrawn. The observed side-effects
of the drug are thought to be due to terfenadine interacting with the hERG channel.
Pharmaceutical heartache

The significance of drug-hERG interactions can be measured by the regulatory
consequences and associated expenses. A non-cardiac drug with a TdP incidence of
less than 0.1% can be removed from the market. For example, the pharmaceutical
industry withdrew terfenadine (Seldane) even though it had an incidence of only
1/28,500 prescriptions, and grepafloxicin (Raxar) was withdrawn due to only seven
cardiac-related deaths and three cases of TdP out of 2.7 million prescriptions.
The number of drugs associated with TdP continues to increase. In recent years a
number of non-cardiac blockbuster drugs including cisapride (Propulsid), astemizole,
grepafloxicin and terfenadine have been found to cause TdP and withdrawn from
major markets. Other drugs such as sertindole (Serlect) and and ziprasidone
(Zeldox) have either been withdrawn prior to marketing or required labelling changes
that significantly restricted their use. Lidoflazine had its marketing authorisation
application rejected because of QT effects.
Against this background it is not surprising that the regulatory authorities have
introduced particular guidelines intended to aid detection of QT effects in new drugs.
Currently the QT interval (or its derivation, the ‘corrected’ QT interval, QTc) is used
as a surrogate for TdP, and as a general rule the regulatory authorities appear to
require a comprehensive evaluation of QT effects to be carried out for new drugs.
This has various implications for the design and expense of clinical trials (for
example, it may be necessary to genotype any patients developing QT prolongation
to identify any genetic risk factors such as those associated with long QT syndrome).
Clearly these types of requirement will add to the length and cost of drug
development, and therefore make it yet more imperative that drugs taken into
expensive clinical trials should have the best possible chance of success.
Hence there is an urgent need for high-throughput screening (HTS) methods to
provide detection of TdP prolongation potential early in the drug development
process. Unfortunately, TdP itself is not amenable to most HTS techniques. In
addition, TdP frequency is so low, usually less than 1/100,000, that it is unlikely to be
detected even in clinical trials, which rarely involve more than 2-3,000 patients at
most. For this reason, the TdP potential of drugs may not become manifest until the
drug is being used in large populations, ie in the post-marketing phase. Therefore a
surrogate marker for TdP is needed to identify drugs with TdP potential in the
preclinical or clinical stages of drug development.
Currently QT or QTc prolongation is used as a TdP surrogate. However, it is by no
means ideal; for example, the average QT increase of about 3% shown by
terfenadine is within the daily QT variability often shown by individuals. Also, there
may be incompletely understood drug-specific influences, in that two drugs with the
same QT effect may have very different TdP potentials. Furthermore, while drugs
that cause TdP prolong the QT interval, drugs that prolong the QT interval do not
necessarily cause TdP. Therefore the use of QT as a TdP surrogate risks elimination
of drugs which in fact would be perfectly safe. In addition, it may be that a proportion
of cases of apparent drug-induced TdP have a significant genetic contribution, eg
due to underlying mutations in the hERG gene.
At present, although QT is a poor predictor of TdP, the regulatory sentiment is such
that drugs which show a QT effect in trials are likely to have more approval issues
than drugs which do not. This is in spite of evidence that QT prolongation does not
necessarily mean that a drug is unsafe (of over 200 compounds associated with QT
prolongation, over half are in clinical use in the UK). Hence the need for HTS or in
silico systems that reliably predict drug-associated QT prolongation in patients.
The heart of the matter

The core issue then is that drug development is suffering from the current lack of
HTS procedures capable of distinguishing between drugs which safely interact with
hERG and drugs which interact dangerously, ie which will cause TdP. The main
hurdle is the lack of a convenient TdP surrogate that can easily be measured in HTS
formats. At present the de facto surrogate is QT, but, as we have seen, this is not a
satisfactory solution.
Most HTS systems require tests to be carried out in cell culture. Routine cell-based
screens for cardiotoxic potential have been facilitated by the cloning of the hERG
gene, which can be functionally expressed in stable, cultured cell lines. This allows
the assessment of drug-hERG interaction by monitoring the effect of the drug on the
currents produced by hERG channels in cultured cells. However, this requires a
sophisticated assay technique known as ‘patch clamping’, which isolates regions of
the cell membrane containing hERG channels and measures changes in electrical
potential difference. Use of this method in high throughput requires automation of
patch clamping in array format, which is not widespread, although it appears to be
becoming increasingly possible. If nothing else, such systems will allow the screening
out of those drugs which have a clearly dangerous effect on QT, while allowing the
further, qualified examination of those candidates with a low or intermediate QT
effect.
However, the real prize will go to the enterprise which can develop a new TdP
surrogate that is both meaningful (ie highly predictive of TdP in patients) and also
amenable to high throughput formats. This may not be as simple as measurement or
prediction of interaction with a given ion channel, as it is possible (for example) that
there are unidentified factors in cardiac cells which affect the precise outcome of a
drug-hERG interaction. As always, biology is not simple, but we should take heart
from the exponential rate of technological progress, and we look forward to future
developments in the field of TdP prediction, which shall surely be accelerated by the
commercial rewards that await developers of an effective product in this field.
Comments may be addressed to the author at [email protected]

Source: http://www.beremans.com/pdf/yeald2_cardiotoxicity.pdf