ToxSci Advance Access originally published online on March 9, 2007
Toxicological Sciences 2007 97(1):4-20; doi:10.1093/toxsci/kfm026
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Preclinical Cardiovascular Risk Assessment in Modern Drug Development
Department of Drug Discovery Support, Boehringer Ingelheim Pharma GmbH and Co. KG, D-88397 Biberach an der Riss, Germany
1 For correspondence via fax: +49 7351-545177. E-mail: brian.guth{at}bc.boehringer-ingelheim.com.
Received December 22, 2006; accepted February 9, 2007
 |
INTRODUCTION |
|---|
TOP
INTRODUCTION
STEPWISE APPROACH TO...The new guidelines including the ICH S7A and S7B (Anonymous, 2001
, 2004) have brought early safety testing of new pharmaceuticals using pharmacological models into the regulatory spotlight in the past few years. Whereas this was motivated to a large extent by a single safety issue (i.e., the proarrhythmic action of drugs that slow myocardial repolarization), the intent was to provide a comprehensive safety assessment based on pharmacodynamic end points, as opposed to histopathological end points used in traditional toxicity studies. These new guidelines were explicitly designed to protect volunteers and patients involved in clinical trials of new drugs from unexpected, life-threatening effects. As such, the emphasis is on acute, functional effects as opposed to chronic histopathological changes. This important distinction was emphasized by Zbinden (1984) some years ago. However, the implications of such testing go well beyond this primary goal and can impact on the later use of a drug in the general population and the wish to avoid any adverse effects, even if not life threatening. The clinical importance of safety pharmacological studies should be obvious, but there is also a substantial economic consequence should adverse effects be found that limit the utility of a given agent and thereby restrict its market potential or even require it to be removed from the market. Thus, safety pharmacology has found a place in modern drug discovery and development.
One should not forget that cardiovascular testing of drug candidates is not a new endeavor. Testing for effects of new agents on arterial blood pressure and heart rate has been conducted routinely for decades, long before the establishment of these new ICH guidelines. The first real attempt to put together a cohesive testing system resulted in the Japanese Guidelines for Nonclinical Studies of Drugs Manual in 1995 (Anonymous, 1995). In this Japanese Guideline, not only there is a requirement for cardiovascular studies but also related effects on the autonomic nervous system and smooth muscle effects are specifically mentioned. Studies on other organ systems were also recommended in the Japanese Guideline making it a very comprehensive package of studies. One might assume that the newer regulatory requirements for safety testing would attempt to further broaden and intensify these safety assessments but this is not the case. In fact, there is a disconcerting trend that the new ICH S7A and S7B guidelines have had just the opposite effect. A strict adherence to the ICH S7A "core battery" studies would mean a substantially less thorough pharmacological profiling than what was recommended in the Japanese Guideline. One can only assume that this was not the intent of the ICH Working Party.
I was reminded recently by a colleague who was in fact involved in the writing of the ICH S7 guideline that one should not forget that guidelines are just that, i.e. guidance and not doctrine. This suggests that a certain amount of flexibility in the testing procedure was actually intended by the authors. Those of us working in the pharmaceutical industry, however, know that such good intentions are difficult to translate into practice, and guidelines are routinely interpreted as law. Nevertheless, it is interesting to observe that approaches to cardiovascular safety testing, particularly in early phases of drug discovery and profiling, are actually quite different from company to company, despite the fact that we are working with the same regulatory guidelines. Some companies continue to support a thorough routine testing of cardiovascular effects early in research, either in the lead optimization phase or at least as a part of the in-depth profiling of selected lead candidates. This may be viewed as a more traditional approach. In contrast, others have decided to undertake only those studies specifically mentioned in the new ICH S7A guideline as core battery studies. This is an alarming trend that was certainly not the intention of the guidance authors. The ICH S7A guideline specifically mentions secondary and follow-up studies that are intended to extend and intensify the study of various organ systems on a project-specific basis. Unfortunately, since they are not included in the core battery, this suggests that they are not essential and are therefore often not performed. A further inadvertent consequence of the ICH S7A guideline relates to its requirement for Good Laboratory Practice (GLP) for the core battery safety pharmacology studies. In order to achieve compliance, some companies have shifted their safety pharmacology studies to a later time than what was typically done in the past. Whereas this has no negative repercussions for drug safety per se, it makes little sense from the perspective of efficient drug development. Specifically, the great utility of safety pharmacology studies (not even restricted to cardiovascular safety) is that these data can be used effectively early in research to select the best possible candidates for further preclinical and clinical development. Delay of these studies until a time in which GLP requirements can be fulfilled means that these data are not available for the candidate selection process. A further inadvertent consequence of the new guidelines is that some companies have chosen to repeat certain studies later in development when GLP compliance is feasible. Again, one can question whether this is a rational approach and may in fact be incompatible with animal use regulations. The pharmaceutical industry should take a broader view of the utility of this type of study and make a differentiation between the fulfillment of regulatory requirements and providing useful in-house data to assist in lead optimization and candidate selection processes.
With this goal in mind, the following will describe an approach to cardiovascular safety testing that positions studies to take advantage of recent advances in testing possibilities and allows early integration into drug research. Early studies are necessarily in vitro, but the subsequent need for in vivo testing of compounds for determining their hemodynamic and electrophysiological effects remains and as yet has not been replaced with alternative approaches. However, refinement of the experimental models and study design can bring about a substantial improvement in data quality (when carefully conducted) and simultaneously result in a dramatic decrease in the number of animals used for such studies. I would also like to take the opportunity to recognize that similar approaches have been recommended by colleagues from other pharmaceutical companies in the past few years and demonstrate that there is a certain consistency between companies that place a high value on early safety testing (Lacroix and Provost, 2000; Redfern et al., 2002) (Table 1).
|
|
STEPWISE APPROACH TO CARDIOVASCULAR RISK ASSESSMENT |
|---|
On the assumption that one chooses to use safety pharmacology studies in a proactive way to support the drug discovery process, there are numerous approaches that one can implement. It is important, however, to not forget some practical constraints inherent with the drug discovery process. An important limitation of "early" safety testing is that the amount of substance available for testing is very limited. Typically, only milligram amounts of compounds are synthesized initially in medicinal chemistry laboratories, and these small amounts are dedicated first to characterization of the effects on the intended pharmacological target or other basic properties needed for a compound to be "drug-like," including chemical and metabolic stability, testing for the inhibition of cytochromes, and pharmacokinetic characterization. Receptor and enzyme activity panels can be conducted with modest amounts of test article and may provide useful data on drug selectivity even at a very early stage. This type of information can contribute substantially to a cardiovascular assessment, but the associated costs usually limit these tests to only selected compounds. A rather new optimization parameter that is also accessible at an early research stage is the activity on human ether-a-go-go-related gene (hERG)–mediated potassium channels. This is the well-known common mechanism for most proarrhythmic compounds that inhibit myocardial repolarization. Methodological approaches have been suggested to address this using high-throughput formats with little test article requirements (see below). At this stage of lead optimization, one should also pay attention to the quality of the test article. In many cases, this may not be ideal and should be considered when deciding which early tests to perform. Finally, many of the early tests require at least modest solubility in aqueous media at physiological pH. Erroneous results can be generated if one does not ensure that compound solubility is adequate for a given test system.
Once test articles are available in amounts from 10 to 100 mg, one can consider further, still mostly in vitro, characterization. Staying with the example of hERG-mediated effects, manual patch-clamping studies can be conducted on not only hERG but also other ion channel targets. In vitro tissue studies using Purkinje fibers, papillary muscles, or isolated intact hearts may also be conducted with limited amounts of test article. Information on hERG-mediated effects together with the evaluation of overall effects on the myocardial action potential in vitro can be used very effectively to anticipate the electrocardiographic effects of drugs in vivo (Guth et al., 2004). Nevertheless, cardiovascular safety testing is still highly dependent upon subsequent testing using in vivo experimental models. There is useful information to be gained from small animal models such as the rat or guinea pig that require modest amounts of test article. Although there has been a trend away from using the rat because of its lack of utility for particularly repolarization-dependent effects and proarrhythmic action, one should not forget that the rat still provides a useful model for evaluating effects on systemic hemodynamics, including arterial blood pressure and heart rate. Miniaturized telemetry technology makes these physiological end points possible while studying the animals in the conscious state. The potential reuse of such instrumented animals can also lead to a reduction in animal usage.
Definitive cardiovascular safety testing is still done in larger, nonrodent species. The most common species remains the dog, both for safety pharmacology studies and for toxicological studies. Increasingly, however, the monkey is being used for toxicological studies and this has necessitated similar cardiovascular models in the monkey. This, in turn, has led to new problems regarding primate availability and health concerns for primates from some sources. This has spurred the search for further alternatives. In this regard, the pig may offer a viable option for toxicological studies and safety pharmacological studies. A limitation is simply that there is less experience with pigs for comparative purposes. Nevertheless, we have recently established a mini-pig model using full-implant telemetry technology for cardiovascular profiling and our initial experience has been very promising.
In vivo cardiovascular studies have been done in the past with anesthetized animals and the Japanese Guideline specifically mentions the anesthetized dog as the model of choice for this study type. Improvements in telemetry technology over the past decade, together with the specification within the ICH S7A guideline calling for the use of conscious animal models have led to widespread application of telemetry systems for not only dogs but also monkeys and pigs. Optimization of these test systems is space- and labor-intensive and start-up costs are comparatively high. Investment in this technology involves not only the substantial start-up price of the system and the needed implants, but requires an adequate infrastructure to obtain the expected high quality data one can achieve with this experimental approach (Klumpp et al., 2006).
Overall, a stepwise approach to cardiovascular safety pharmacology profiling (Fig. 1) appears to be pragmatic and beneficial for the lead optimization process during research. It can be seen that safety pharmacology testing can span a wide range of the drug discovery and development process. The tests increase in complexity and in general move from in vitro to in vivo. The various general study types with suggestions for specific methodologies will be described.
|
|
IN VITRO ELECTROPHYSIOLOGICAL ASSESSMENTS |
|---|
If the ICH S7A and S7B guidelines have had an impact, it is certainly in the area of testing for proarrhythmic potential for drug candidates. In particular, the routine testing for effects on the hERG-mediated potassium has become standard in the pharmaceutical industry over the past few years and it has become a primary optimization parameter in early drug research. Given these screening and profiling activities, it is likely that new drugs with a potential for causing hERG-mediated arrhythmias will become a rarity in the near future. Whereas there appears to be a consensus that this issue should be addressed early in the drug research process, there are different experimental paths one can follow to eliminate this unwanted activity. The gold-standard approach remains the measurement of hERG-mediated current using manual patch-clamp technology, an inherently tedious undertaking. Alternative experimental systems have been devised to allow a higher throughput. These include competitive binding strategies, membrane potential-based marker approaches, and automatization of the more traditional patch-clamp approach. Though all offer higher throughput, they also have disadvantages, as discussed below. In addition, significant advances in the use of in silico approaches for assessing potential interactions with the hERG channel have been reported and can be used to direct the selection of chemical scaffolds and further lead optimization activities (Price et al., 2006
; Sanguinetti and Mitcheson, 2005).
Much controversy has surrounded the use of in vitro models that measure myocardial action potential and can assess effects of drugs on the action potential configuration. Since this is inherently a more physiological test system in comparison to the heterologous cell systems frequently used for measuring activity on hERG channels, there has always been a convincing rationale supporting this approach to obtain a more integrated assessment of a drug effect. However, the ICH S7B guideline has tended to steer one away from conducting these types of studies since they are not included in the core battery of studies. Earlier position papers on this subject, most notably the Committee for Proprietary Medicinal Product "points to consider" paper (Anonymous, 1997) emphasized the great utility of these models. Unfortunately, the fact that they were removed from the recommended core studies and were delegated to "follow-up" study status has been misinterpreted by some to imply that they have little value. As emphasized below, when well done, they can be of tremendous value for the early drug research process. It is perhaps illustrative to observe that companies that have used action potential studies in the past continue to do so and are convinced of their value. It appears to be primarily those without direct experience with the action potential models that tend to choose to not conduct them based on the guidelines.
"High-Throughput" hERG Assays
Recent work has highlighted the K+ channel encoded by the hERG as the molecular target for a wide range of drugs whose administration is associated with an increased risk of an unusual, life-threatening form of arrhythmia known as torsade de pointes (Vandenberg et al., 2001). One could approach this issue within drug development by screening drug libraries to identify and exclude compounds having this potential. This approach has the inherent disadvantage of negatively biasing compounds that may otherwise be good starting points for lead optimization. Given that medicinal chemists are becoming very good at minimizing hERG activity through lead optimization (Jamieson et al., 2006), such a broad screening approach is probably unwarranted. Furthermore, experience has shown that a wide variety of compounds have the potential to block these channels; it is just a question of the test concentration before one sees an effect. Indeed, compounds that show absolutely no effect, even at high concentrations, are more the exception than the rule. Therefore, hERG-blocking activity must always be viewed in conjunction with the activity on the intended target to know if there may be a risk in clinical use (Redfern et al., 2003). Thus, it is more rational to test for this activity in an early secondary screen on hits that emerge from examining effects on the intended therapeutic target. In any case, experimental approaches are needed that allow the testing of large numbers of compounds thereby necessitating a high degree of automation. Several test systems have emerged as potential high-throughput approaches for detecting activity on IKr, the current mediated by the hERG channel. These are relatively new technologies that are still in a development stage. Nevertheless, they are still worth mentioning since they may provide useful approaches for early drug discovery. Such techniques have been reviewed and critically assessed previously (Netzer et al., 2001, 2003).
Competetive Binding Assays
Competitive binding of a test compound in comparison to a radiolabeled compound known to be a high affinity antagonist of the hERG channel has been used. Radiolabeled, potent blockers of hERG such as [3H]-dofetilide (Finlayson et al., 2001a,b) or [35S]MK-499 (Wang et al., 2003) can be used in conjunction with myocardial cells, stably transfected cell lines, or membrane preparations. With this approach, the hERG channels are fully blocked with excess radiolabeled ligand. The test compound is then incubated with the hERG radioligand where it can compete for the binding sites initially occupied by the radioligand as a function of its potentcy.
The test system has relatively low costs and can provide a high throughput. However, compounds that can block hERG-mediated currents but do not compete for the same binding site will not be detected. Due to the heterogeneity of chemical structural classes known to block hERG, it is thought that multiple binding sites exist. This test system detects binding and does not demonstrate any change in the electrophysiological action of the cells used. There may be differences to results of electrophysiological assessments of drug activity. Using a [35S]MK-499–based test system, astemizole and terfenadine produce IC50 results that compare favorably with electrophysiological recordings (Wang et al., 2003). In contrast, the measured value for cisapride using this assay is threefold higher than electrophysiological studies and MK-499 is 30-fold lower than an electrophysiological approach (Wang et al., 2003). Such discrepancies also appear using a [3H]-dofetilide–based system and may even be more pronounced. Thus, the specificity of the binding site for the radiolabeled compound, versus the test compound, may limit the usefulness of this approach. Nevertheless, for specific chemical classes known to compete for a given binding site, this may be a viable approach.
Rubidium Flux Assays
Rubidium exhibits a very high flux through potassium ion channels. Upon depolarization, this flux can be quantified as a measure of channel integrity (Cheng et al., 2002; Tang et al., 2001). Fluorometric assays have been developed that can detect drug-induced effects on hERG channels. The test system utilizes cells expressing hERG channels that are incubated with rubidium that effectively replaces potassium in the cells. The cells are then preincubated with compound and then depolarized with a high concentration of potassium. The supernatant is then removed after several minutes of incubation, and the distribution of rubidium between the supernatant and the cells is determined. Accordingly, a hERG blockade prevents the efflux of rubidium from the cells, whereas the lack of blockade allows rubidium to leave the cells. Radiolabeled rubidium can be measured using scintillation counting (Weir and Weston, 1986) or with atomic-absorption spectroscopy (Terstappen, 1999).
This experimental approach assesses the patency of the hERG potassium channel and therefore provides a functional assessment in contrast to the binding assay. A good correlation between results with this assay system and electrophysiological measurements has been reported (Cheng et al., 2002; Tang et al., 2001) but this may be dependent upon the specific type of blocker being tested. For example, a poorer correlation may be found with voltage-dependent blockers (Cheng et al., 2002). The radioactive waste generated with the use of this technique should also be considered when selecting a test system. The use of atomic-absorption spectroscopy avoids this problem, but has lower throughput.
Fluorescence Ion Channel Assays Using Voltage-Sensitive Dyes
Fluorescent, voltage-sensitive dyes can be employed to investigate the activity of ion channels by reflecting the cellular membrane potential (Epps et al., 1994; Plasek and Sigler, 1996). This approach can be used with different measurement systems including fluorometric imaging plate readers (FLIPR), voltage/ion probe readers (VIPR), or conventional fluorescence readers. The FLIPR technique utilizes an oxonol-derivative dye in a patented test kit (FMP-kit, [Baxter et al., 2002]). Use of this approach for the measurement of IC50 values for known hERG-blocking drugs has been reported (Tang et al., 2001), but there are apparent discrepancies between such values and those obtained using electrophysiological methods. This may, at least in part, be due to the rather slow kinetics of the dye, making steady-state measurements preferable using this system. An approach yielding a higher time resolution uses the fluorescence resonance energy transfer between two dyes. This can be used in conjunction with VIPR configured as a high-throughput system (Gonzalez and Maher, 2002). Conventional fluorescence readers can also be employed to make equilibrium measurements with potential-sensitive oxonol dyes (Netzer et al., 2001, 2003). The reported IC50 values generated with this approach appear to correlate well with electrophysiological measurements (Netzer et al., 2003) with similar absolute values.
The use of membrane potential-sensitive dyes allows a high-throughput approach for detecting drug effects on hERG potassium channels. Independent of the measurement system chosen, common disadvantages are potential fluorescence or quenching artifacts. The membrane potential is also an indirect measurement of a drug effect on the hERG potassium channel. The start-up costs for the necessary instrumentation (e.g., FLIPR of VIPR) may also present a constraint to this approach (Netzer et al., 2001).
Automated Patch-Clamp Systems
Whole-cell patch-clamp techniques, as described below, are still considered to be the gold standard for determining the effects of drugs on the hERG potassium channel. New systems are currently in development that attempt to incorporate the inherent advantages of the patch-clamping approach into an automated and, thereby, higher throughput format. These systems are not truly high throughput but their goal is to speed the study process through automation while preserving the manner in which the study is conducted and the measurements that are made. Thus, cells transfected with hERG and expressing current resembling IKr are used and a whole-cell patch-clamp methodology is applied. Patching of the cells and addition of the test compound is, however, done on an automated basis.
A critical step to the performance of a whole-cell patch-clamp experiment is the formation of a tight, high resistance seal ("gigaohm seal") between the cell membrane and the hole in a planar chip. The formation of this seal remains the limiting factor with this experimental approach with success rates on the order of 50%. An alternative is the use of a so-called "perforated patch" which uses pore-forming agents to form less tight seals that do not need a negative pressure on the cell. The resulting lower resistance seals are reported to be approximately 100 M
(Kiss et al., 2003; Schroeder et al., 2003). While increasing seal success rate, these seals may not be optimal for the electrophysiological recordings. As with standard patch-clamp studies, the stimulation protocol and bath solutions may influence the results.
We have used one such automated patch-clamping approach, as offered through a contract research organization, to generate data at an early time in the lead optimization process. This has provided useful data to guide the synthesis programs in medicinal chemistry. To streamline the process and to reduce costs, single concentrations of the test article are used as opposed to generating IC50 values initially. We noted, however, that as the IC50 approached the test concentration, there was a large variability in the "% inhibition" value achieved. This is not an intrinsic limitation of the automated patch-clamp system, but rather a consequence of the steep relationship between concentration and effect (channel inhibition in this case) near the IC50. We have addressed this by adding a second concentration to the screen to better predict the IC50. A manual patch-clamp study is always performed for interesting compounds and, in general, there is a good correlation between the data generated with the automated patch-clamp system and the in-house manual system. We have noted that certain chemical structural classes have differences in the absolute % inhibition values obtained (3- to 10-fold difference). However, the effect appears to be consistent for a given chemical class and can therefore be accounted for. Limited solubility may also account for these differences and data obtained with poorly soluble compounds should be viewed cautiously and measurement of the actual drug concentration is recommended.
In summary, there are now a variety of techniques available to test compound hERG channel activity at a relatively early stage of drug discovery and lead optimization. Caution should be used, however, in that the absolute values for inhibitory activity may vary substantially from values coming from manual patch-clamp studies. Furthermore, there may be differences in the performance of these test systems depending on the specific chemical class one is working with. Therefore, it is recommended to evaluate the utility of any such test for the specific discovery program and the chemical classes in lead optimization. Automated patch-clamp systems appear to offer a reasonable compromise between a higher throughput while maintaining the physiological integrity of the measurements and thereby their potential relevance (Table 2).
|
|
MANUEL VOLTAGE CLAMP STUDIES ON hERG-MEDIATED POTASSIUM CHANNELS |
|---|
The introduction of the patch-clamp technique (Neher and Sakmann, 1976
) revolutionized the study of cellular physiology by providing a high-resolution method of observing the function of individual ionic channels in a variety of normal and pathological cell types. By varying the basic recording methodology, cellular function and regulation can be studied at a molecular level by observing currents through individual ion channels (Liem et al., 1995). In safety studies during drug development, the single most important channel is the hERG channel. The purpose of this type of study is to examine drug-induced effects on model systems for the ventricular membrane current IKr (i.e., hERG-mediated current) by the use of cells functionally expressing hERG channels. The activity of these voltage-gated channels is best measured using voltage clamp techniques.
Various cells can be used to conduct electrophysiological experiments to determine drug effects on hERG-mediated potassium current. This includes cells that naturally have hERG current including isolated ventricular cardiomyocytes (Jurkiewicz and Sanguinetti, 1993), atrial cell lines (Busch et al., 1996), or neuroblastoma cell lines (Arcangeli et al., 1998). Alternatively, heterologous expression of hERG in other cells was selected to simplify the measurement of the hERG-mediated current. This includes transient expression systems using, for example, Xenopus oocytes, or stably transfected mammalian cells lines, with human embryonal kidney (HEK) 293 and Chinese hamster oocyte (CHO) cells being the most commonly used cell types. Such heterologous expression systems, particularly the mammalian cell-based systems, have become the most widely used and accepted models for hERG channel testing.
Xenopus Oocytes
Xenopus oocytes are a useful heterologous system for measuring hERG current; they are robust cells that express a comparatively large and well-defined hERG current. Their large size also makes the technical aspects of the procedure easier than with small cell types. The disadvantage, however, is that most drugs show lower potency for hERG blockade in Xenopus oocytes in comparison to mammalian cell lines (Po et al., 1999; Weerapura et al., 2002; Witchell et al., 2002). This may be due to absorption of drug in yolk particles within the cells, thereby lowering the intracellular free drug concentration. Whereas the absolute IC50 values may be 3- to 10-fold higher in Xenopus oocytes, their rank order of potency appears to be comparable to that seen in mammalian cells. Nevertheless, there are examples of compounds (e.g., sotalol and erythromycin) shown to block hERG in mammalian cells that do not show a blockade in Xenopus oocytes, suggesting a greater risk of a false negative assessment when using oocytes in comparison to mammalian cells.
Stably Transfected Cells
A preference for a mammalian cell line for measuring drug effects on hERG current is based on the perception that the results may be more representative of what may happen in patients than results from a Xenopus oocyte–based model. The relatively small size of the HEK293 and CHO cells and the lack of a yolk provide a system in which the measured blockade is usually more potent than in Xenopus oocytes. HEK293 cells may have an endogenous transient outward potassium current (Ito) (Snyders and Chaudhary, 1996), whereas CHO cells do not have such a current (Teschemacher et al., 1999). The amount of Ito in HEK293 cells may be variable between batches of cells and could interfere with hERG measurements making CHO cells in this regard superior. HEK293 cells may, however, be more consistent in their level of heterologous expression in comparison to CHO cells (Witchell et al., 2002).
Issues relating to the use of stably versus transient transfection have been reviewed elsewhere (Witchell et al., 2002). Membrane currents can be recorded from HEK293 cells at room temperature (20–22°C) or at physiological temperature (37°C), using the whole-cell patch-clamp technique. Although it is theoretically attractive to conduct such studies at 37°C, there are some practical limitations that suggest conducting the studies at room temperature, at least for routine screening (Witchell et al., 2002). This includes not having to preheat the perfusion media and the observation that cells last substantially longer when studied at room temperature. The measured current has a faster kinetics and is larger at 37°C as compared to room temperature. Also, at 37°C currents measured during strongly depolarizing pulses may diminish during the course of the pulse, a phenomenon not seen at room temperature.
Studies on hERG Channels in Isolated Ventricular Myocytes
Patch-clamp techniques using isolated ventricular myocytes may also be used to measure hERG channel function (Drolet et al., 1999; Kalifa et al., 1999; Salata et al., 1995). For example, ventricular myocytes from guinea pigs can be used (Jurkiewicz and Sanguinetti, 1993). Alternatively, Carlsson et al. (1997) performed voltage clamp studies in isolated ventricular myocytes from rabbits. Whole-cell currents can be recorded in the tight-seal, whole-cell mode of the patch-clamp technique (Sanguinetti and Jurkiewicz, 1990).
Irrespective of the model selected, measurement of hERG-mediated current has become a routine part of safety pharmacological profiling of drug candidates. The selection of the model is secondary to establishing a database using known positive and negative controls with which one can compare new compounds. There will always be differences in the absolute values one obtains for effects on hERG in different systems, even when using what would at first glance appear to be identical test systems. Subtle differences in the test system may be enough to elicit these differences. Nevertheless, using data from reference compounds in relationship to unknown test compounds should provide the basis of comparability between laboratories.
Studies on Myocardial Ion Channels in Addition to hERG
The importance of hERG-blocking activity for drug development is clear when it leads to a prolongation of the myocardial action potential and thereby to a proarrhythmic state. One must recognize, however, that the myocardial action potential is a result of the concerted action of multiple ion channels. Other drug-induced ion channel effects can also lead to changes in the overall action potential and are potentially proarrhythmic such as a blockade of the myocardial sodium channel that is responsible for the depolarization of the heart (Nav1.5). Blockade of these channels leads to a reduced depolarization phase (Vmax) of the action potential and thereby a prolongation of the action potential. In vivo a widening of the QRS complex results as a consequence of the slower depolarization of the left ventricle. Sodium channel–blocking drugs, although developed for their antiarrhythmic action, are well known to be proarrhythmic (Morganroth et al., 1986). A further example is that of the slowly activating repolarziation current IKs, carried by the KvLQT1/minK channel. This current, together with hERG-mediated current, makes up the major repolarization current of the heart (Sanguinetti and Jurkiewicz, 1990). Blockade of these channels can lead to similar action potential prolongation as compared to hERG blockade and it is assumed to present a similar proarrhymic potential.
Alternatively, additional ion channel effects can offset each other, thereby reducing the potential proarrhythmic activity of what might otherwise be interpreted as a proarrhythmic signal. The blockade of myocardial calcium channels has been shown to counteract the effects of drugs that also potently inhibit the hERG channel. A well-known example is verapamil that, despite a potent blockade of hERG channels, is also not associated with depolarization-dependent arrhythmias. One does not see the expected prolongation of the action potential due to hERG blockade, rather one sees a shortening of the action potential due to the calcium channel blockade. Thus, this type of study can assist one in identifying additional proarrhythmic activity of drugs, or can be useful in putting into context an observed activity in blocking hERG. A thorough electrophysiological profile of drug candidates will likely soon include not only the hERG channels but also the major myocardial channels involved in both depolarization and repolarization. Thus, in addition to hERG, one may want to consider studies examining the action of a given agent on Kv1.5 (Ikur), Kv4.3 (Ito), KvLQT/minK (IKs), nav1.5 (Ina), Cav1.2 (Ica-L), or NCX1 Na-Ca exchanger (responsible for the current N/A). This information is invaluable when apparent discrepancies arise between effects seen on hERG channels and the overall effect on the myocardial action potential. Discrepancies suggest that other ion channels may be important and these can be easily measured when one has access to the appropriate test system, something now offered by contract research organizations.
hERG Channel Trafficking
A final aspect deserving mention is that the expression of ion channels in the heart is not static but rather a dynamic process. Ion channels are continuously trafficked to and from the membrane surface where they are active. It has been shown that action potential changes can be elicited by interfering with this process without a direct interaction of a molecule on the channel function, per se (Wible et al., 2005). As such, one should consider also including this type of assessment particularly if effects are seen after longer exposure to a test article but are absent with acute administration.
|
STUDIES ON THE MYOCARDIAL ACTION POTENTIAL |
|---|
Studies to assess the potential of a drug to affect ventricular repolarization, and therefore a prolongation of the QT interval duration, focus on the pivotal role of hERG-mediated potassium current in mediating this effect. Nevertheless, it is well accepted that even a potent blockade of the hERG channel does not necessarily lead to a QT prolongation. For example, the calcium channel blocker verapamil does not lead to QT interval prolongation despite an IC50 on hERG in the range of 100–200nM. The reason for this apparent discrepancy is simply that the myocardial action potential configuration is the net result of the concerted activity of numerous ion channels and effects on a given channel can be masked by the activities of other competing channels; other examples are also apparent (Yuill et al., 2004
). Thus, measuring effects on the myocardial action potential allows one to assess the physiological relevance of any activity on hERG channels that may be present (Anonymous, 1997; Haverkamp et al., 2000). In other words, a compound that shows both potent inhibition of hERG channels and prolongs the myocardial action potential in "low" concentrations will present a higher risk for clinical QT prolongation than the compound that blocks hERG, but shows no effect on the overall action potential. Thus, the performance of hERG channel assays together with action potential assays may provide an early risk assessment with a better correlation to the in vivo and clinical setting than with just one of these assays alone (Guth et al., 2004).
Drug effects on the myocardial action potential are typically measured in vitro in myocardial tissue such as Purkinje fibers (Gintant et al., 2001), papillary muscles, ventricular wedge preparations (albeit less commonly), or the entire isolated heart (Franz, 1991). The focus of all these approaches is to assess the action potential duration, measured as the time required to a given percentage of repolarization, e.g., action potential duration (APD) 90 or the time to 90% repolarization. Additionally, some of these models can respond to hERG-blocking drugs with early after depolarizations (EAD) thought to be a substrate for arrhythmia. The use of the myocardial wedge preparation also addresses the important issue of potential transmural differences in drug-induced effects on repolarization (Antzelevitch et al., 1999), an aspect not addressed by Purkinje fibers or papillary muscles.
Measurement of the Concentration of Test Article in In Vitro Systems
The results from studies designed to assess potential drug effects on hERG channels or on the action potential in vitro are used as an early risk assessment for clinically relevant effects on the QT interval duration and even proarrhythmic activity. The use of these data for the estimation of safety margins necessitates an accurate measurement of the drug concentrations present in the test system. The perfusion baths used are typically protein free and one needs to consider potential protein binding of a compound when comparing to the in vivo situation. At early stages of drug research, these data may not be available, however, these can have a substantial impact on the risk assessment since the safety margins are determined with unbound drug fraction (Redfern et al., 2003). Furthermore, compounds may adhere to glass or plastic used in the experimental setups, with the possibility that the intended drug concentrations are not reached, particularly at lower concentrations. Drug solubility in aqueous solutions at pH 7.4 can also limit the concentrations being tested. Multiple aspects of this important issue have been reviewed recently (Herron et al., 2004).
Studies in Isolated Purkinje Fibers
Studies in isolated Purkinje fibers can detect compound-induced effects on the action potential configuration and possible EAD. This type of procedure is a logical adjunct to studies examining effect on hERG-mediated current since it examines the relevance of any hERG-blocking activity on the overall myocardial action potential.
The primary evaluation for detecting hERG-mediated effects involves the determination of the action potential duration. Typically, the time from depolarization to a predefined % repolarization is used to describe the repolarization dynamics of the tissue. That is, time to 10% (APD10), 30% (APD30), 50% (APD50), 70% (APD70) and 90% (APD90) can be used to quantify the repolarization curve. For hERG-dependent effects on the action potential, the APD90 is most commonly reported. Most compounds known to block hERG-mediated potassium current prolong the action potential (e.g., prolongation of the APD90) in this model. The model is capable of detecting both prolongation and shortening of the action potential and it should not be forgotten that other drug-induced effects on electrophysiological function can be detected, including effects on sodium (e.g., effects on the rate of depolarization) and calcium currents (e.g., effects on the plateau phase of the action potential). The Purkinje fiber is a noncontractile tissue, which facilitates electrode positioning and stability. In comparison to the papillary muscle or the monophasic action potential in the intact heart, drug-induced effects on the action potential duration are considerably larger. Whereas this is not necessarily a disadvantage, the model has been viewed as possibly being too sensitive for drug-induced effects on repolarization. A clear disadvantage of the Purkinje fiber model (and the guinea pig papillary model, see below) is that not all hERG-blocking compounds produce the expected prolongation of the action potential. One notable example is terfenadine (Gintant et al., 2001) that in both canine and porcine Purkinje fibers even in supratherapeutic concentrations failed to prolong the action potential. The reason for this is still unclear but the perception of a risk for false negative results has lead to these types of studies receiving secondary status from regulatory authorities (i.e., the Food and Drug Administration). It must be emphasized, however, that most hERG-blocking drugs do indeed lead to an action potential prolongation in this model and that exceptions are rare. Also, potent hERG-blocking compounds that do not prolong the QT interval and are not proarrhythmic (such as verapamil) do not prolong the action potential in this model. Purkinje fibers from the dog or the rabbit have been used typically, while other species, including the pig (Gintant et al., 2001), may also be appropriate models. The use of primate tissue for such studies carries with it ethical and financial considerations, as does the use of dog Purkinje fibers. Studies in isolated rabbit Purkinje fibers have also been used successfully to assess the risk of QT interval prolongation by drugs (Adamantidis et al., 1995, 1998; Cavero et al., 1999; Champeroux et al., 2000; Dumotier et al., 1999).
Studies in Isolated Guinea Pig Papillary Muscles
One of the most commonly used models for assessing the myocardial action potential uses the guinea pig papillary muscle. It has an appropriate size since the papillary muscle from larger species may be compromised by ischemia in the middle of the muscle during the experiment. The use of the guinea pig papillary muscle was also highlighted in the Japanese PRODACT initiative. In these studies, the guinea pig papillary muscle was assessed to be a very sensitive and useful model for predicting effects on the QT interval in vivo (Hayashi et al., 2005). Like studies performed using Purkinje fibers, the primary evaluation involves the determination of the action potential duration as described above.
In contrast to the Purkinje fiber, the papillary muscle is a contractile tissue thus allowing the measurement of contractile force, together with the action potential configuration. Since effects on the inotropic state of the myocardium can also affect the action potential, this is a useful secondary measurement to interpret possible drug-induced effects. However, contraction of the muscle makes the placement of the electrode more difficult and changes in the contractile function can lead to loss of the electrode placement in the course of a study. As with the Purkinje fiber model, some compounds known to potently block hERG channels and to cause QT prolongation in the clinic do not demonstrate the expected action potential prolongation; this includes astemizole, bepridil, pimozide, and terfenadine. In the case of the guinea pig papillary muscle, it was suggested recently in conjunction with the PRODACT evaluation, that the way the data are evaluated may be key for detecting drug-induced effects on the action potential configuration. The PRODACT investigators suggested using the difference between APD90 and APD30 as a novel approach to examine effects on late ventricular repolarization phase that may be affected by hERG-blocking drugs (Hayashi et al., 2005). With this approach, these investigators were able to demonstrate effects of astimizole, bepridil, and pimozide in the isolated guinea pig papillary muscle. Nevertheless, terefenadine did not show an effect on the APD30–90, such that it was suggested that other limitations must prevent terfenadine from demonstrating its expected APD prolongation in this in vitro setting. It should also be noted that with terfenadine it is difficult to show a QT prolongation in vivo, as well.
Arterially Perfused Wedge of Canine Left Ventricle
The M cell is a unique myocardial cell type found in the deeper layers of the ventricular wall (Antzelevitch et al., 1999). These cells respond more sensitively to agents that block hERG channels and, as such, contribute to possible drug-induced transmural heterogeneity of ventricular repolarization and thereby the proarrhythmic potential. The perfused myocardial wedge preparation is designed to allow the study of transmural differences in drug action on the action potential and may therefore provide a better assessment of possible proarrhythmic potential of a test article.
The methodology was first described by Antzelevitch et al. (1996). A transmural electrocardiogram is recorded using extracellular silver chloride electrodes placed near the epicardial and endocardial surfaces of the preparation. Transmembrane action potentials were simultaneously recorded from the epicardial, endocardial, and M regions using three separate intracellular floating microelectrodes filled with KCl and connected to a high input impedance amplifier. Impalements are obtained from the cut surface of the preparation at positions approximating the transmural axis of the electrocardiogram.
The evaluation is based on the measurement of the QT interval of the electrocardiogram, together with the action potential configuration measured using the transmurally located electrodes. The action potential duration is typically measured using the time to either 50 or 90% repolarization. The action potential duration of the M cells (midwall measurement) is expected to be longer than either endocardial or epicardial cells, and drug-induced effects on these cells are also greater. The rate dependency of a given drug effect is also amplified in the M cells and may importantly contribute to the proarrhythmic potential of a drug.
The measurement of a transmural electrocardiogram (ECG) together with local action potentials from across the ventricular wall of the dog provides one of the most sophisticated in vitro approaches for determining drug-induced effects on repolarization, as well as having implications for proarrhythmic potential. Thus, objectively assessed, this model is perhaps the best in vitro model to examine drug-induced effects on repolarization of the heart. Its main disadvantage is the experimental complexity which makes it accessible only in specialized laboratories and requires extensive training to master the technical preparation. A further complicating factor is the use of dogs for studies in which only small portions of the excised heart are utilized, which has both ethical and financial aspects that need to be considered.
Isolated Heart (Langendorff) Preparations
Similar types of studies can be conducted to those described using myocardial tissue in vitro by using the entire isolated heart. One measures an external monophasic action potential, as opposed to an intracellular myocardial action potential (Franz, 1991). Hearts are typically perfused retrogradely from the aorta in variations of the so-called Langendorff preparation.
Eckardt et al. (1998) and Johna et al. (1998) proposed the isolated perfused rabbit heart as a model to study proarrhythmia induced by class III antiarrhythmic drugs. Alternatively, the isolated guinea pig heart can be used effectively in similar preparations for drug evaluation (Hamlin et al., 2004; Stark et al., 1987).
|
THE CASE FOR EARLY IN VIVO TESTING IN RODENTS OR OTHER SMALL ANIMALS |
|---|
There is a clear preference for performing safety pharmacology studies for drug effects on the cardiovascular system in a nonrodent species in order to assess possible effects on ventricular repolarization. Nevertheless, there is still great value in conducting studies to examine drug-induced effects on the cardiovascular system in particularly rats. One reason is that toxicity studies are traditionally conducted in rats and drug effects on cardiovascular parameters in rats are therefore of relevance to such studies. Pharmacokinetic and metabolism data are therefore readily available for the rat, which is not always the case for other species. Furthermore, cardiovascular studies in nonrodents (usually the dog) require substantially more compound than studies in rats. Therefore, as part of the lead optimization, selection and development process, cardiovascular studies in rats can be invaluable. One must simply recognize the limitation of the rat for assessing effects on ventricular repolarization through the blockade of hERG-mediated potassium current. We use a rat model for the assessment of cardiovascular end points prior to testing a given compound in our nonrodent models. This serves to prioritize compounds early and ensures that the compounds that ultimately are tested in the larger animal models are free from major hemodynamic effects.
Small Animal Telemetry System
Cardiovascular parameters such as arterial blood pressure and heart rate can be measured in conscious, free-moving rats using commercially available radiotelemetry systems (Deveney et al., 1998). Such systems include (1) an implantable transmitter with battery capable of sending the pressure signal from a fluid-filled catheter implanted into the abdominal aorta, (2) a receiver unit that detects the transmitted signal and converts it into a digital format, (3) a pressure reference module that adjusts the measured aortic pressure for atmospheric pressure, and (4) data acquisition software for data storage and computation.
The rats are anesthetized and undergo an initial aseptic procedure to implant the aortic catheter distal to the renal arteries. The catheter is secured using cyanoacrylate glue. The transmitter and battery unit are placed in the abdomen and attached to the abdominal musculature using a suture. After closure of the surgical wound, the animals are given sufficient time to recover from the procedure before being used in a study, usually at least 1 week. Using this technology, rats can be studied while still in their home cage. However, since the transmitter frequency used is the same for all animals, they must be held isolated during the study. This means that in most cases the rats are moved from multiple housing cages to single cages for the duration of a study. The cages are placed directly on the receiver units to provide a close contact between the receiver and the animal. The transmitters are activated before starting the study; the transmitter units can be magnetically switched on and off thereby lengthening battery life. A control period of upto 1–2 h allows the animals to acclimate to the environment and to provide a steady baseline prior to administration of a test article. The oral administration of compounds via gavage is most convenient in rats but iv, sc and ip administration is also feasible. Inhalation administration of compounds is also possible by nebulizing the compound into a small exposure box or through the use of an application system bringing the nebulized compound to the nose of the animal. Cardiovascular parameters are affected immediately following the administration of a compound due to the stress inherent with the administration process, however, within 30 min the animals return to a normal hemodynamic steady state.
Studies can be conducted with a group comparison or a crossover design. Individual parameters are recorded continuously for 8–24 h after drug administration to ensure all drug-induced effects are detected. Evaluation of the data can be performed by calculating areas under the parameter-time curve and testing for drug-induced effects. In this way, the analysis is not restricted to a given time point. With crossover design an appropriately long washout phase is needed between treatments. A treatment arm with vehicle is also useful to demonstrate the stability of the system. Water is usually provided ad libitum and food can be given should the measurements be performed over an extended time period.
Instrumented animals may be reused after appropriate washout periods on the assumption that the compounds in the doses tested cause no irreversible damage. Since the cardiovascular parameters measured are sensitive to the overall well being of the animals, changes in baseline conditions should be examined for detecting possible compound-related toxicities. Assuming such toxicities are not observed, animals can be maintained for over 1 year with appropriate care.
Due to the fact that whole-body plethysmographs for rats or guinea pigs are constructed out of plexiglass, they can be used in conjunction with telemetry systems without interfering with the transmitted signals. Therefore, this offers the opportunity to combine studies for determining cardiovascular function using telemetry, with whole-body plethysmograph studies for simultaneously measuring respiratory function (Schierok et al., 2000).
As mentioned above, the most important limitation of the rat as cardiovascular model for safety pharmacology studies is its "lack" of utility for detecting ventricular repolarization-related drug effects. Whereas this is a substantial limitation, there are nevertheless other drug-induced effects that can be observed in the rat model including effects on arterial blood pressure, heart rate, and contractile ventricular function. The rat, due to its rather small size, requires substantially less compound to perform a study in comparison to that what is needed for larger animals. Thus, this model can be used at an earlier time during drug optimization and selection when amounts of a given test compound are limited. One should keep in mind, however, that a rat instrumented for telemetric collection of cardiovascular data, if kept for longer periods of time, grows considerably in comparison to the "typical" laboratory rat and can grow to over 600 g.
An interesting alternative to the rat for this type of study is use of the guinea pig with chronically implanted instrumentation for cardiovascular measurements. We have used this particularly for obtaining ECG data but have had difficulties with the chronic implantation of arterial catheters. A recent study from Provan et al. (2005) may show a way to overcome this technical limitation. The guinea pig has been shown to demonstrate drug-induced effects on the QT interval duration (Hamlin et al., 2003) and is suitable for use with telemetry-based systems. Hey et al. (1996) analyzed the ECG wave in anesthetized guinea pigs to determine QT interval, QTc interval, PR interval, QRS interval, and heart rate after the administration of the second-generation antihistamines ebastine and terfenadine. More recently, the anesthetized guinea pig has been shown to be useful for safety pharmacological evaluations (Hamlin et al., 2003; Hauser et al., 2005).
|
IN VIVO APPROACHES TO CARDIOVASCULAR PROFILING AND ELECTROCARDIOGRAPHIC ASSESSMENTS |
|---|
Cardiovascular Assessments Using Anesthetized Animal Models
The basic parameters needed for a safety pharmacology evaluation of possible cardiovascular effects of drugs includes heart rate, arterial blood pressure, and the electrocardiogram. These were often measured in anesthetized animals, however, the potential influence of anesthesia on the parameters measured needed to be taken into account. Hence, the cited preference in the ICH S7A guideline for the use of conscious animals for such assessments. There are, however, advantages to anesthetized preparations. More invasive techniques can be applied, thereby giving an in-depth evaluation of possible drug effects on heart and vascular function in a variety of perfusion beds. A well-performed study using an anesthetized animal model is characterized by a highly stable hemodynamic state with very low variability of the parameters measured. This results in a highly sensitive model for detecting possible drug-induced effects. The oral route is, however, not accessible in anesthetized animals although intraduodenal administration may provide a reasonable alternative if the iv route is not feasible or not desirable. A "well-performed" study refers to the manner in which the anesthesia is conducted, maintained, and monitored. The goal is to produce a profound (i.e., surgical level without pain responses) and stable level of anesthesia that does not vary appreciably over time. This is feasible using either inhalation anesthetics or through steady-state infusion protocols. Bolus iv administration of anesthetic agents having hemodynamic effects (e.g., pentobarbital) leads to marked changes and should be avoided.
Measurements possible in this model include end-diastolic and systolic pressure of the left ventricle, contractility of the heart (usually using peak positive LV dP/dt or LV dP/dt at a developed pressure of 40 mmHg), heart rate, cardiac output, and arterial blood flow in various local perfusion beds, as may be needed. Test compounds can be classified in terms of a variety of pharmacological actions:
- Positive and negative inotropic effects
- Arrhythmogenic effects
- Hyper- or hypotensive effects
- Tachycardic or bradycardic effects
Until recently, this experimental approach was the most common for studying drug-induced effects on the cardiovascular system and was specifically mentioned in the Japanese Guidelines for general pharmacology as the standard test (Anonymous, 1995). As such, there is a vast amount of experience with this type of study and a large amount of comparable data available. The main reason why anesthetized animal models lost their primary role in pharmacological studies was due to the recognition of the possible effects of anesthesia (particularly pentobarbital) on ventricular repolarization and therefore on drug-induced effects on QT interval duration (Bachmann et al., 2002; Weissenburger et al., 2000). However, a well-performed study using anesthetized animals, as supported recently by the Japanese PRODACT investigators, can provide a useful and sensitive model of detecting drug-induced effects on the QT interval duration (Tashibu et al., 2005). Furthermore, the utility of this model for sensitively detecting effects on arterial blood pressure, heart rate, and ventricular contractility is well established. A possible limitation of this approach is the use of iv administration for drugs intended for the oral route. Clearly, an adequate compound solubility must be present or one needs to resort to vehicles, many of which have independent pharmacodynamic effects (Pestel et al., 2006). Anesthetized preparations also lend themselves to frequent blood sampling so that drug concentrations can easily be measured. We typically use an infusion administration scheme as opposed to using iv boluses to allow for a better correlation of effects to "steady-state" drug plasma levels.
Pigs may also be used for this type of study and are an attractive alternative if the dog is not suitable. Adult domestic pigs are difficult to use due to their size, such that if adult animals are preferred, one of several breeds of mini- or micropigs may be used. The induction of anesthesia is different than with the dog and typically an intramuscular sedative is administered first (e.g., ketamine) followed by the anesthesia used for the remainder of the study. Halothane anesthesia should not be used in pigs due to a high incidence of malignant hyperthermic reactions.
|
CARDIOVASCULAR SAFETY STUDIES IN CONSCIOUS DOGS AND OTHER NONRODENT SPECIES |
|---|
The preferred model for performing safety pharmacology studies on the cardiovascular system according to the ICH S7A guideline is the conscious animal under nonstressed, physiological conditions (Anonymous, 2001
). This includes the evaluation of drug-induced effects on systemic arterial hemodynamics and the electrocardiogram. Important cardiovascular parameters are, however, best measured using invasive measurement techniques but these can interfere with the physiological status of the animal subject. Relatively new telemetry-based systems (Mellor and Pettinger, 1986) allow one to monitor hemodynamic parameters including heart rate, arterial blood pressure, and ECG (and, with some systems, left ventricular blood pressure and left ventricular dP/dt) with the animals in their home cage, thereby serving to reduce the stress associated with the collection of data. It should be noted, however, that the use of telemetry is not essential for the performance of this type of study. For example, animals can be trained to the laboratory environment such that they can be examined using light restraint (e.g., sling or chair) under physiological conditions. This approach is suitable if the main objective of the study is to obtain electrophysiological data using external limb leads. Also, noninvasive measurements of arterial blood pressure are possible using well-trained conscious dogs, albeit with somewhat less sensitivity in comparison to invasive measurement. Other useful cardiovascular end points, such as left ventricular dP/dt require more invasive approaches. Externalization of implanted transducers and "hard-wiring" animals has been used successfully in the past, but this has the inherent disadvantage that such wires provide sites for infection. Even with meticulous care, such infections are unavoidable. In general, the use of fully implantable telemetry technology has replaced these approaches in the pharmaceutical industry in recent years.
Full-implant telemetry systems typically measure arterial blood pressure and a single electrocardiogram, but they can also be configured for the measurement of left ventricular pressure or multiple ECG leads. However, a single ECG lead with clearly definable waveforms is usually sufficient for detecting drug-induced effects. Beagles are frequently used for these studies but other larger breeds have the advantage of a simplified implantation procedure (since they are larger) and, in general, lower heart rates at the time of the subsequent studies. We use Labrador dogs with success and have never seen data with Beagles that achieve similarly low heart rates during studies. There is some evidence that drug-induced effects on the QT interval duration are more pronounced in female animals. For the Labrador dog, we find absolutely no difference between males and females such that both sexes can be included in these studies. Of utmost importance is that all animals are adequately trained for adaptation to their home cage and laboratory environment prior to being instrumented. Furthermore, the training process should continue indefinitely even after instrumentation. Instrumented animals should receive regular contact with the people involved in the studies. We have investigated the impact of housing conditions on the parameters measured and have found that this is an important consideration (Klumpp et al., 2006). Dogs should be paired with their usual housing mates (the dogs are routinely group housed) during the measurements and visual contact with the other dogs used in the study should be maintained. This approach requires that dogs have independent carrier frequencies for the transmitted data, a feature not available in all telemetry systems.
The transducers of the telemetry implant are calibrated prior to implantation and the unit is sterilized using a low-pressure ethylene oxide process. For the one-time surgery, dogs are anaesthetized and all procedures are performed under aseptic conditions using sterilized equipment. Details of the surgical procedure have been published in the past (Markert et al., 2004). Studies can be conducted with the animals in their home cages if they are equipped with antennae to pick up the transmitted signals. Drugs can be applied orally, iv, sc, or through inhalation although the latter requires some additional training (Markert et al., 2004). After treatment, dogs are returned to their home cages for the duration of the study that can last up to 24 h, if needed. This is usually based on the duration of the expected drug exposure or possible pharmacological activity of the test article. Longer studies should allow for feeding of the animal and water should be available ad libitum.
Studies may be conducted as group comparisons if sufficient numbers of animals are available. Alternatively, the use of a Latin square crossover experimental design allows for studies with fewer animals (e.g., n = 4). The experiment starts after an equilibrium period of 60–120 min to allow the dogs to acclimate to the measurement pens. The administration of the test compound is started after a 30- to 45-min control period. Experiments involving iv infusions or frequent blood sampling are performed in smaller cages to allow close monitoring of indwelling catheters. Continuous measurements should not, however, exceed 6–8 h without allowing for a short pause for exercise. Continuous measurements for up to 24 h are acceptable if the animals are kept in larger cages.
The hemodynamic and ECG parameters can include systolic, diastolic, and mean aortic pressure, peak systolic and end-diastolic left ventricular pressure, LV dP/dt max and dP/dt min, heart rate, PQ, QRS and QT intervals. As mentioned, not all systems include a pressure transducer for left ventricular pressure. We have found this to be an invaluable parameter, particularly the derivative (dP/dt) that serves as an index of myocardial contractility. Drug-induced effects on left ventricular dP/dt are, at least based on our experience, more common than one might think. We have detected on three occasions that compounds intended for entering development were found to have a negative inotropic effect. Based on these observations, the compounds were deprioritized and alternatives were found. Various commercially available software programs can be used for data acquisition and analysis. Whereas all physiological parameters are routinely averaged over predefined time intervals, it has been proposed that for the ECG only a few beats are required to detect effects (Hamlin et al., 2004). Of particular interest are time points corresponding to the time of Cmax of the test agent. One can include blood sampling within a test protocol for the measurement of plasma drug concentration. Even in well-trained animals, this leads to a transient stress that interferes with the cardiovascular data. Therefore, one should keep such sampling to a minimum. It is preferable to have a well-documented pharmacokinetic profile of a given test article prior to conducting such a study.
Correction of QT Interval for Heart Rate
The QT interval duration is heart rate dependent. The use of correction formulae derived from clinical data is not appropriate for use with dog ECGs. Algorithms designed specifically for the dog are required and historical data from the dogs actually used in a given study is the preferred way to derive any QT correction (Meyners and Markert, 2004). Still better, if no drug-induced effect on heart rate is observed, no correction of the QT interval should be undertaken.
Animal Reuse
Animals instrumented for this type of study may be reused. Since the instrumentation is fully implanted without externalization of wires or catheters, there is little risk of developing sepsis after successful implantation. There may be problems that arise with erosion of either the battery or the transmitter unit, but fortunately these are rare. The batteries are designed to provide long life well in excess of a year. After the completion of a given study and an appropriate washout time, further studies can be conducted. This is facilitated by the fact that studies are typically single administration studies using doses that are not intended to cause irreversible effects. Nevertheless, there is a need to explore supratherapeutic doses in order to get an impression of a safety window with regard to cardiovascular effects. The qualification of animals for subsequent use should be based on the health status of the animals. Given the ease of obtaining hemodynamic data from these animals, one can maintain historical data on heart rate and blood pressure, parameters that are sensitive to the overall well being of the animal. Additionally, clinical chemistry parameters can be monitored to detect possible effects on kidney and liver function.
Similar telemetry-based cardiovascular safety pharmacology studies can be conducted in pigs and monkeys (normally Macaca mulatta [rhesus] or Macaca fascicularis [cynomulgus]). These two alternative species are of interest due to their use in toxicology studies where the dog is deemed inappropriate.
Incorporating Safety Pharmacology End Points in Toxicology Studies
A common approach, also mentioned in the ICH S7A guideline, is the incorporation of safety pharmacology cardiovascular end points into toxicological studies. This is an attractive approach in that it theoretically leads to a reduction in the number of animals used for the safety evaluation of a new compound. Since toxicological studies are routinely performed under GLP conditions, the new requirement for the core battery safety pharmacology can also be accommodated without additional effort to qualify test systems. There are several arguments against this approach, however. One consideration is the delay in the availability of the resultant data. The amount of delay may vary between companies based on when GLP toxicology studies are performed, but is certainly measured in months. We use the cardiovascular safety pharmacology data as an important qualifier for a compound to enter into development and this decision is made well before the first GLP toxicology studies are performed. Another major consideration, particularly for cardiovascular studies, has to do with the quality of the data one can expect from a toxicology study. Simply stated, the data are highly variable and difficult to standardize. As an example, I was recently asked to review the results of a 2-week toxicology study in which arterial blood pressure, heart rate, and ECG were measured in Beagles. The heart rates reported under control conditions ranged from a low of 90 beats/min to a high of 174 beats/min (our Labrador dogs have heart rates under 60 beats/min). Heart rates following treatment in this case study increased to over 200 beats/min. Paper tracings were made (50 mm/s) using temporarily placed limb leads. Given the huge variability in this set of data, it is impossible to make a detailed ECG assessment. Finally, toxicological studies include doses that will cause some adverse effect that may be associated with discomfort or even overt side effects. These may elicit secondarily an activation of the sympathetic nervous system. Sympathetic stimulation activates mechanisms that indirectly affect both IKr and IKs thereby leading to a prolongation of the QT interval (Fossa et al., 2005). Thus, QT prolongation should be a common observation in such studies at doses associated with other side effects. Measuring heart rate, blood pressure, and ECG during a toxicological study may provide some important insights into the toxicological results since they will detect major changes in hemodynamic status and can detect arrhythmias. However, a detailed analysis of the ECG, including subtle morphological changes or measurement of intervals is not feasible in the context of a normal toxicity study and should not be attempted.
|
DESIGNING A COMPREHENSIVE CARDIOVASCULAR ASSESSMENT APPROACH |
|---|
Integrating a safety pharmacological assessment into drug discovery and development has become essential to fulfill new regulatory requirements but more importantly this assessment provides important data for the selection of drug candidates. As mentioned from the onset, only doing GLP safety pharmacology studies on drugs prior to phase I clinical trials will mean that data are available too late for lead optimization and the drug candidate selection process. Furthermore, a strict adherence to the ICH S7 guidelines will result in a narrow cardiovascular assessment that does not address all potentially important aspects of cardiovascular function. Thus, a thorough cardiovascular safety pharmacological assessment should consider additional aspects and may include a variety of experimental models not explicitly listed in the guidelines. There is clearly no optimal approach that fits all needs since pharmacological profiling must be tailored to the needs of individual projects. The approach chosen should reflect a company's expertise and project-specific potential liabilities. Nevertheless, some general trends have emerged in the past few years. The general approach we use was summarized graphically in Figure 1.
In vitro electrophysiological testing is a rather new element of safety pharmacology profiling for many companies, but is here to stay. In particular, screening for hERG channel activity can and should be done rather early in the lead optimization process. Methodologies are available for reasonable throughput and the potential for interactions with hERG should be identified early enough to allow medicinal chemists the time needed to reduce this unwanted activity. It should be noted that this can be done successfully with current knowledge of the hERG channel and potential drug interactions (Jamieson et al., 2006). Effective tools are now available and include in silico approaches (Aronov, 2006; Sanguinetti and Mitcheson, 2005), high-throughput approaches, as well as traditional manual electrophysiological studies in well-profiled test systems. Our experience has been that testing for hERG activity as projects make a transition from the lead identification phase (i.e., identification of chemical scaffolds with target activity and the possibility to optimize) to lead optimization activity is a reasonable time to perform these studies. There is typically enough time remaining and structural diversity available to reduce hERG-blocking activity in the ultimate drug candidates.
A second in vitro step that adds value to the overall electrophysiological assessment is to evaluate effects on myocardial action potential using systems mentioned above. Our initial focus on hERG is necessary from a pragmatic point of view, based on the regulatory fixation on this single channel activity. However, there is a strong argument that ultimately only effects that result in changes on the myocardial action potential are of relevance. The models needed for assessing effects on the myocardial action potential are a bit more complex and time consuming that those used for measuring hERG activity, but they require little compound and give additional value to the overall risk assessment. Paired with results from a hERG assay, one achieves rather early on, a predictive assessment for cardiovascular liability (Guth et al., 2004). The myocardial action potential assay can also be included during early lead optimization.
If one focuses on the potential for compounds to affect ventricular repolarization, one might argue that the next essential step entails an evaluation of the electrocardiogram in a nonrodent species to complement the in vitro evaluation. Indeed, this is what is found in the ICH S7B guideline (Anonymous, 2004). However, there is a major step in terms of compound requirements from low milligram quantities to hundred milligram or even gram amounts of compound to move into such in vivo models. One should recognize that potential hemodynamic activity (changes in arterial blood pressure, heart rate) can be addressed easily in small animal models including the rat and guinea pig. The guinea pig, in contrast to the rat, offers the additional advantage of being useful for assessing effects on ventricular repolarization. One note of caution, however, is that the guinea pig proves to be more challenging than the rat in terms of implantation of particularly arterial catheters, in comparison to the rat (Provan et al., 2005). Such studies may be done later in lead optimization when candidate selection and prioritization is going on. These studies form a valuable bridge to later studies in larger animals that are only feasible for a few drug candidates that are thought to have a chance for further development. Unexpected effects on blood pressure and heart rate can be just as devastating to a lead optimization program, as are effects on the QT interval.
Finally, a thorough in vivo assessment of both systemic hemodynamics and the electrocardiogram is indispensable. The dog has been most commonly used species for these assessments and full implant technology has become the gold-standard experimental approach to allow for studies in conscious animals. Our experience with the use of Labrador dogs so instrumented and extensively trained is that we can achieve optimized physiological conditions during such studies with very low variability of the parameters measured and therefore a high sensitivity for drug-induced effects (Meyners and Markert, 2004). Heart rates ranging between 50 and 60 beats/min throughout a 7- to 8-h observation period after dosing are routinely measured with our model. The laboratory environment is critical to achieving such results with attention to noise or other disturbances that stimulate the dogs. Routine training and a well thought out housing strategy contribute importantly to the optimization of this type of system (Klumpp et al., 2006). Although we have less experience with the mini-pig and monkey, both species can also be successfully used for generating high quality cardiovascular data.
My intention from the onset was not to be overly critical of the new guidelines for safety pharmacology. It is, however, important to keep them in context as guidance outlining a general approach for profiling drug candidates. It would be a mistake to interpret them as a cookbook of core battery studies, perform only these, tick the boxes, and then think one has done an adequate profiling. A guideline such as the ICH S7A must be very general in nature and cannot specify exactly what is reasonable to do for every development compound. Good pharmacological science should be the driving force behind the selection of studies performed and the responsibility for this selection still lies with the sponsor.
|
REFERENCES |
|---|
Adamantidis MM, Dumotier BM, Caron JF. Sparfloxacin but not levofloxacin or ofloxacin prolongs cardiac repolarization in rabbit Purkinje fibers. Fundam. Clin. Pharmacol (1998) 12:70–76.[Web of Science][Medline]
Adamantidis MM, Lacroix DL, Caron JF, Dupuis BA. Electrophysiological and arrhythmogenic effects of the histamine type 1-receptor antagonist astemizole on rabbit Purkinje fibers: clinical relevance. J. Cardiovasc. Pharmacol (1995) 26:319–327.[Web of Science][Medline]
Anonymous. (1995). Japanese Guidelines for Nonclinical Studies of Drugs Manual 1995. Yakuji Nippo Limited, Tokyo, Japan.
Anonymous. (1997). Committee for Proprietary Medicinal Product. Points to Consider. The Assessment of QT Interval Prolongation by Non-Cardiovascular Medicinal Products. CPMP/986/96.
Anonymous. (2001). ICH Harmonized Tripartite Guideline, S7A Safety Pharmacology Studies for Human Pharmaceuticals. U.S. Department of Health and Human Services, FDA. Available at: http://www.fda.gov/cder/guidance/P266_27807. Accessed July 2001.
Anonymous (2004). ICH S7B Guideline Step 2 Revision. The Non-Clinical Evaluation of the Potential for Delayed Ventricular Repolarization (QT Interval Prolongation) by Human Pharmaceuticals. U.S. Department of Health and Human Services, FDA. Available at: http://www.fda.gov/cder/guidance/5533drf.htm. Accessed 10 June 2004.
Antzelevitch C, Sun Z-Q, Zhan Z-Q, Yan G-X. Cellular and ionic mechanisms underlying erythromycin-induced long QT intervals and torsade de Pointes. J. Am. Coll. Cardiol (1996) 28:1836–1848.[Abstract]
Antzelevitch C, Shimizu W, Yan G-X, Sicouri S, Weissenburger J, Nesterenko VV, Burasnikov A, Diego JD, Saffitz J, Thomas GP. The M cell: its contribution to the ECG and to normal and abnormal electrical function of the heart. J. Cardiovasc. Electrophysiol (1999) 10:1124–1152.[Web of Science][Medline]
Arcangeli A, Rosati B, Cherubini A, Crociani O, Fontana L, Passani B, Wanke E, Olivotto M. Long term exposure to retinoic acid induced the expression of IRK1 channels in HERG channel-endowed neuroblastoma cells. Biochem. Biophys. Res. Commun (1998) 244:706–711.[CrossRef][Web of Science][Medline]
Aronov AM. Common pharmacophores for uncharged human ether-a-go-go-related gene (hERG) blockers. J. Med. Chem (2006) 49:6917–6921.[CrossRef][Web of Science][Medline]
Bachmann A, Mueller S, Kopp K, Brueggemann A, Suessbrich H, Gerlach U, Busch AE. Inhibition of cardiac potassium currents by pentobarbital. Naunyn Schmiedeberg Arch. Pharmacol (2002) 365:29–37.[CrossRef][Web of Science][Medline]
Baxter DF, Kirk M, Garcia AF, Raimondi A, Holmqvist MH, Flint KK, Bojanic D, Distefano PS, Curtis R, Xie Y. A novel membrane potential-sensitive fluorescent dye improves cell-based assays for ion channels. J. Biomol. Screen (2002) 7:79–85.
Busch AE, Suessbrich H, Waldegger S, Sailer E, Greger R, Lang HJ, Lang F, Gibson KJ, Maylie JG. Inhibition of IKs in guinea pig cardiac myocytes and guinea pig IKs channels by the chromanol 293B. Pflügers Arch (1996) 432:1094–1096.[CrossRef][Web of Science][Medline]
Carlsson L, Amis GJ, Andersson B, Drews L, Duker G, Wadstedt G. Electrophysiological characterization of the prokinetic agents cisapride and mosapride in vivo and in vitro: implications for proarrhythmic potential? J. Pharmacol. Exp. Ther (1997) 282:220–227.
Cavero I, Mestre M, Guillon JM, Heuilett E, Roach AG. Preclinical in vitro cardiac electrophysiology: a method of predicting arrhythmogenic potential of antihistamines in humans? Drug Saf. (1999) 21(Suppl 1):19–31.[CrossRef][Web of Science][Medline]
Champeroux P, Martel E, Vannier C, Blanc V, Leguennec JY, Fowler J, Richard S. The preclinical assessment of the risk for QT interval prolongation. Thérapie (2000) 55:101–109.[Web of Science][Medline]
Cheng CS, Alderman D, Kwash J, Dessaint J, Patel R, Lescoe MD, Kinrade MB, Yu W. A high-throughput HERG potassium channel function assay: an old assay with a new look. Drug Dev. Ind. Pharm (2002) 28:177–191.[CrossRef][Web of Science][Medline]
Deveney AM, Kjellström A, Forsberg T, Jackson DM. A pharmacological validation of radiotelemetry in conscious, freely moving rats. J. Pharmacol. Toxicol. Methods (1998) 40:71–79.[CrossRef][Web of Science][Medline]
Drolet B, Vincent F, Rail J, Chahine M, Deschenes D, Nadeau S, Kalifa M, Hamelin BA, Turgeon J. Thioridazine lengthens repolarization of cardiac ventricular myocytes by blocking the delayed rectifier potassium current. J. Pharmacol. Exp. Ther (1999) 288:1261–1268.
Dumotier BM, Adamantidis MM, Puisieux FL, Bastide MM, Dupuis BA. Repercussions of pharmacologic reduction in ionic currents on action potential configuration in rabbit Purkinje fibers: are they indicative of proarrhythmic potential? Drug Dev. Res (1999) 47:83–76.
Eckardt L, Haverkamp W, Mertens H, Johna R, Clague JR, Borggrefe M, Breithart G. Drug-related torsades de pointes in the isolated rabbit heart: Comparison of clofilium, d,l-sotalol, and erythromycin. J. Cardiovasc. Pharmacol (1998) 32:425–434.[CrossRef][Web of Science][Medline]
Epps DE, Wolfe ML, Groppi V. Characterization of the steady-state and dynamic fluorescence properties of the potential-sensitive dye bis-(1,3-dibutylbarbituric acid)trimethine oxonol (BiBAC4(3)) in model systems and cells. Chem. Phys. Lipids (1994) 69:137–150.[CrossRef][Web of Science][Medline]
Finlayson K, Pennington AJ, Kelly JS. [3H]Dofetilide binding in SHSY5Y and HEK 293 cells expressing a HERG-like K+ channel? Eur. J. Pharmacol (2001b) 412:202–212.
Finlayson K, Turnbull L, January CT, Sharkey J, Kelly JS. [3H]Dofetilide binding to HERG transfected membranes: a potential high throughput preclinical screen. Eur. J. Pharmacol (2001a) 430:147–148.[CrossRef][Web of Science][Medline]
Fossa AA, Wisialowski T, Magnano A, Eric W, Winslow R, Gorczyca W, Crimin K, Raunig DL. Dynamic beat-to-beat modeling of the QT-RR interval relationship: analysis fof QT prolongation during alterations of autonomic state versus human ether a-go-go-related gene inhibition. J. Pharmacol. Exp. Ther (2005) 312:1–11.
Franz MR. Method and theory of monophasic action potential recording. Prog. Cardiovasc. Dis (1991) 6:347–368.
Gintant GA, Limberis JT, McDermott JS, Wegner CD, Cox BF. The canine Purkinje fiber: an in vitro model system for acquired long QT syndrome and drug-induced arrhythmogenesis. J. Cardiovasc. Pharmacol (2001) 37:607–618.[CrossRef][Web of Science][Medline]
Gonzalez JE, Maher MP. Cellular fluorescent indicators and voltage/ion probe reader (VIPRTM): tools for ion channel and receptor drug discovery. Recept. Channels (2002) 8:283–295.[CrossRef][Web of Science][Medline]
Guth BD, Germeyer S, Kolb W, Markert M. Developing a strategy for the nonclinical assessment of proarrhythmic risk of pharmaceuticals due to prolonged ventricular repolarization. J. Pharmacol. Toxicol. Methods (2004) 49:159–169.[CrossRef][Medline]
Hamlin RL, Kijtawornrat A, Keene BW. How many cardiac cycles must be measured to permit accurate RR, QT, and QTc estimates in conscious dogs? J. Pharmacol. Toxicol. Methods (2004) 50:103–108.[CrossRef][Medline]
Hamlin RL, Kijtawornrat A, Keene BW, Hamlin DM. QT and RR intervals in conscious and anesthetized guinea pigs with highly varying RR intervals and given QTc-lengthening test articles. Toxicol. Sci (2003) 76:437–442.
Hauser D, Stade M, Schmidt A, Hanauer G. Cardiovascular parameters in anaesthetized guinea pigs: a safety pharmacology screening model. J. Pharmacol. Toxicol. Methods (2005) 52:106–114.[CrossRef][Medline]
Haverkamp W, Breithardt G, Camm AJ, Janse MJ, Rosen MR, Antzelevitch C, Escande D, Franz M, Malik M, Moss A, et al. The potential for QT prolongation and pro-arrhythmia by non-anti-arrhythmic drugs: clinical and regulatory implications report on a policy conference of the European society of cardiology. Cardiovasc. Res (2000) 47:219–233.
Hayashi S, Yoshihide K, Tabo M, Fukuda H, Itoh T, Shimosato T, Amano H, Saito M, Morimoto H, Yamada K, et al. QT PRODACT: a multi-site study of in vitro action potential assays on 21 compounds in isolated guinea-pig papillary muscles. J. Pharmacol. Sci (2005) 99:423–437.[CrossRef][Web of Science][Medline]
Herron W, Towers C, Templeton A. Experiences in method development for the analysis of in vitro study solutions for content. J. Pharmacol. Toxicol. Methods (2004) 49:211–216.[CrossRef][Medline]
Hey JA, del Prado M, Kreutner W, Egan RW. Cardiotoxic and drug interaction profile of the second generation antihistamines ebastine and terfenadine in an experimental model of torsade de pointes. Arzneimforsch/Drug Res (1996) 46:159–163.
Jamieson C, Moir EM, Rankovic Z, Wishart G. Medicinal chemistry of hERG optimizations: highlights and hang-ups. J. Med. Chem (2006) 49:5029–5046.[CrossRef][Web of Science][Medline]
Johna R, Mertens H, Haverkamp W, Eckardt L, Niederbröcker T, Borggrefe M, Breithardt G. Clofilium in the isolated perfused rabbit heart: a new model to study proarrhythmia by class III antiarrhythmic drugs. Basic Res. Cardiol (1998) 93:127–135.[CrossRef][Web of Science][Medline]
Jurkiewicz NK, Sanguinetti MC. Rate dependent prolongation of cardiac action potentials by a methanosulfonanilide class III antiarrhythmic agent: specific block of a rapidly activating delayed rectifier K+ current by dofetilide. Circ. Res (1993) 72:75–83.
Kalifa M, Drolet B, Daleau P, Lefez C, Gilbert M, Plante S, O'Hara GE, Gleeton O, Hamelin BA, Turgeon J. Block of potassium currents in guinea pig ventricular myocytes and lengthening of cardiac repolarization in man by the histamine H1 antagonist diphenhydramine. J. Pharmacol. Exp. Ther (1999) 288:858–865.
Kiss L, Bennett PB, Uebele VN, Koblan KS, Kane SA, Neagle B, Schroder K. High throughput ion-channel pharmacology: planar-array-based voltage clamp. Assay Drug Dev. Technol (2003) 1:127–135.[CrossRef][Web of Science][Medline]
Klumpp A, Trautmann T, Markert M, Guth B. Optimizing the experimental environment for dog telemetry studies. J. Pharmacol. Toxicol. Methods (2006) 54:141–149.[CrossRef][Medline]
Lacroix P, Provost D. Basic safety pharmacology: the cardiovascular system. Thérapie (2000) 55:63–69.[Web of Science][Medline]
Liem LK, Simard JM, Song Y, Tewari K. The patch clamp technique. Neurosurgery (1995) 36:382–392.[Web of Science][Medline]
Markert M, Klumpp A, Trautmann T, Guth BD. A novel propellant-free inhalation drug delivery system for cardiovascular safety pharmacology evaluations in dogs. J Pharmacol. Toxicol. Methods (2004) 50:109–199.[CrossRef][Medline]
Mellor PM, Pettinger SJ. Application of radio telemetry to cardiovascular monitoring in unrestrained animals. J. Pharmacol. Methods (1986) 16:181–184.[CrossRef][Web of Science][Medline]
Meyners M, Markert M. Correcting the QT interval for changes in HR in pre-clinical durg development. J. Pharmacol. Toxicol. Methods (2004) 50:109–119.[CrossRef][Medline]
Morganroth J, Anderson JL, Gentzkow GD. Classification by type of ventricular arrhythmias predicts frequency of adverse cardiac events from flecainide. J. Am. Coll. Cardiol (1986) 8:607–615.[Abstract]
Neher E, Sakmann B. Single-channel currents recorded from membranes of denervated frog muscle fibres. Nature (1976) 260:799–802.[CrossRef][Medline]
Netzer R, Bischoff U, Ebneth A. HTS techniques to investigate the potential effects of compounds on cardiac ion channels at early-stages of drug discovery. Curr. Opin. Drug Discov. Dev (2003) 6:462–469.[Web of Science][Medline]
Netzer R, Ebneth A, Bischoff U, Pongs O. Screening lead compounds for QT interval prolongation. Drug Discov. Today (2001) 6:78–84.[CrossRef][Web of Science][Medline]
Pestel S, Martin H-J, Maier G-M, Guth B. Effect of commonly used vehicles on gastrointestinal, renal, and liver function in rats. J Pharmacol. Toxicol. Methods (2006) 54:200–214.[CrossRef][Medline]
Plasek J, Sigler K. Slow fluorescent indicators of membrane potential: a survey of different approaches to probe response analysis. J. Photochem. Photobiol. B Biol (1996) 33:101–124.[CrossRef][Medline]
Po SS, Wang DW, Zang ICH, Johnson JP, Nie L, Bennett PB. Modulation of hERG potassium channels by extracellular magnesium and quinidine. J. Cardiovasc. Pharmacol (1999) 33:181–185.[CrossRef][Web of Science][Medline]
Price DA, Armour D, De Groot M, Lieshman D, Napier C, Perros M, Stammen BL, Wood A. Overcoming HERG affinity in the discovery of the CCR5 antagonist maraviroc. Bioorg. Med. Chem. Lett (2006) 16:4633–4637.[CrossRef][Medline]
Provan G, Stanton A, Sutton A, Rankin-Burkart A, Laycock SK. Development of a surgical approach for telemetering guinea pigs as a model for screening QT interval effects. J. Pharmacol. Toxicol. Methods (2005) 52:223–228.[CrossRef][Medline]
Redfern WS, Carlsson L, Davis AS, Lynch WG, MacKenzie I, Palethorpe S, Siegl PKS, Strang I, Sullivan AT, Wallis R, Camm AJ, et al. Relationships between preclinical cardiac electrophysiology, clinical QT interval prolongation and torsade de pointes for a broad range of drugs: evidence for a provisional safety margin in drug development. Cardiovasc. Res (2003) 58:32–45.
Redfern WS, Wakefield ID, Prior H, Pollard CE, Hammond TG, Valentin J-P. Safety pharmacology—a progressive approach. Fundam. Clin. Pharmacol (2002) 16:161–173.[CrossRef][Web of Science][Medline]
Salata JJ, Jurkiewicz NK, Wallace AA, Stupineski RF III, Guinosso PJ Jr, Lynch JJ Jr. Cardiac electrophysiological actions of the histamine H1-receptor antagonists astemizole and terfenadine compared with chlorpheniramine and pyrilamine. Circ. Res. (1995) 76:110–119.
Sanguinetti MC, Jurkiewicz NK. Two components of cardiac delayed rectifier K+ currents: differential sensitivity to block by class III antiarrhythmic agents. J. Gen. Physiol (1990) 96:195–215.
Sanguinetti MC, Mitcheson JS. Predicting drug-hERG channel interactions that cause acquired long QT syndrome. Trends Pharmacol. Sci (2005) 26:119–124.[CrossRef][Medline]
Schierok H, Markert M, Pairet M, Guth B. Continuous assessment of multiple vital physiological functions in conscious freely moving rats using telemetry and a plethysmography system. J. Pharmacol. Toxicol. Methods (2000) 43:211–217.[CrossRef][Web of Science][Medline]
Schroeder K, Neagle B, Trezise DJ, Worley J. IonWorksTM HT: a new high-throughput electrophysiology measurement platform. J. Biomol. Screen (2003) 8:50–64.
Snyders DJ, Chaudhary A. High affinity open channels blockade by dofetilide of HERG expressed in a human cell line. Mol. Pharmacol (1996) 49:949–955.[Abstract]
Stark G, Stark U, Tritthart HA. Continuous ECG measurements of intracardiac activity from the surface of Landorff-perfused guinea pig hearts. Basic Res. Cardiol (1987) 83:202–212.[CrossRef][Web of Science]
Tang W, Kang J, Wu X, Rampe D, Wang L, Shen H, Li Z, Dunnington D, Garyantes T. Development and evaluation of high throughput functional assay method for HERG potassium channel. J. Biomol. Screen (2001) 6:325–331.
Tashibu H, Myazaki H, Aoki K, Akie Y, Yamamoto K. QT PRODACT: in vivo QT assay in anesthetized dog for detecting the potential for QT interval prolongation by human pharmaceuticals. J. Pharm. Sci (2005) 99:473–486.[CrossRef][Web of Science]
Terstappen GC. Functional analysis of native and recombinant ion channels using a high-capacity nonradioactive rubidium efflux assay. Anal. Biochem (1999) 272:149–155.[CrossRef][Web of Science][Medline]
Teschemacher AG, Seward EP, Hancox JC, Witchel HJ. Inhibition of the current of heterologously expressed HERG potassium channels by imipramine and amitriptyline. Br. J. Pharmacol (1999) 128:479–485.[CrossRef][Web of Science][Medline]
Vandenberg JI, Walker BD, Campbell TJ. HERG K+ channels: friend and foe. Trends Pharmacol. Sci (2001) 22:240–246.[CrossRef][Medline]
Wang J, Della Penna K, Wang H, Karczewski J, Connolly TM, Koblan KS, Bennett PB, Salata JJ. Functional and pharmacological properties of canine ERG potassium channels. Am. J. Physiol. Heart Circ. Physiol. (2003) 284:H256–H267.
Weerapura M, Nattel S, Chartier D, Caballero R, Hebert TE. A comparison of current carried by hERG, with and without coexpression of MiRP1, and the native rapid delayed rectifier current. Is MiRP1 the missing link? J. Physiol (2002) 540:15–27.
Weir SW, Weston AH. The effects of BRL 34915 and nicorandil on electrical and mechanical activity and on 86Rb efflux in rat blood vessels. Br. J. Pharmacol (1986) 88:121–128.[Web of Science][Medline]
Weissenburger J, Nesterenko VV, Antzelevitch C. Transmural heterogeneity of ventricular repolarization under baseline and long QT conditions in the canine heart in vivo: torsades de pointes develops with halothane but not pentobarbital anesthesia. J. Cardiovasc. Electrophysiol (2000) 11:290–304.[Web of Science][Medline]
Wible BA, Hawryluk P, Ficker E, Kuryshev YA, Kirsch G, Brown AM. HERG-Lite®: a novel comprehensive high-throughput screen for drug-induced HERG risk. J. Pharmacol. Toxicol. Methods (2005) 52:136–145.[CrossRef][Medline]
Witchell HJ, Milnes JT, Mitcheson JS, Hancox JC. Troubleshooting problems with in vitro screening of drugs for QT interval prolongation using HERG K+ channels expressed in mammalian cell lines and Xenopus oocytes. J. Pharmacol. Toxicol. Methods (2002) 48:65–80.[CrossRef][Web of Science][Medline]
Yuill KH, Borg JJ, Ridley JM, Milnes JT, Witchel HJ, Paul AA, Kozlowski RZ, Hancox JC. Potent inhibition of human cardiac potassium (HERG) channels by the anti-estrogen agent clomiphene- without QT interval prolongation. Biochem. Biophys. Res. Commun (2004) 318:556–561.[CrossRef][Web of Science][Medline]
Zbinden G. Neglect of function and obsession with structure in toxicity testing. In: Proceedings, 9th International Congress of Pharmacology (1984) Vo1. 43–49. New York: Macmillan.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. J. Perrin, P. W. Kuchel, T. J. Campbell, and J. I. Vandenberg Drug Binding to the Inactivated State Is Necessary but Not Sufficient for High-Affinity Binding to Human Ether-a-go-go-Related Gene Channels Mol. Pharmacol., November 1, 2008; 74(5): 1443 - 1452. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||