药物毒性试验的设计

Optimising the design of preliminary toxicity studies

for pharmaceutical safety testing in the dog

David Smith a, , , Robert Combes b , Olympe Depelchin c , Soren Dyring Jacobsen d , Ruediger Hack e , Joerg Luft f , Lieve

Lammens g , Friedrich von Landenberg h , Barry Phillips i , Rudolf Pfister j , Yvon Rabemampianina k , Susan Sparrow l , Claudia Stark m and Markus Stephan-Gueldner n

a AstraZeneca, Alderley Park, UK

b FRAME, Nottingham, UK

c Lilly, Mont-Saint Guibert, Belguim

d NovoNordisk, Maaloev, Denmark

e Aventis, Frankfurt, Germany

f AltanaPharma, Hamburg, Germany

g Janssen, Beerse, Belgium

h Merck KGaA, Darmstadt, Germany

i RSPCA, Horsham, UK

j Novartis Pharma, Basel, Switzerland

k Pfizer, Amboise, France

l GlaxoSmithKline, Ware, UK

m Schering AG, Berlin, Germany

n Hoffmann-La Roche AG, Basel, Switzerland

Abstract

A working party, comprising two animal welfare organisations and

some 12 pharmaceutical companies in Europe, was established to minimise the use of the dog in safety testing. As first step, the

participants defined the major objectives of preliminary dose-range finding/MTD toxicity studies in non-rodents, defined the principles and requirements for this study type and agreed on a proposal for an optimised study design, based on collective experience of conducting such studies in industry, involving an evaluation of 100 individual study data sets. The suggested study design is explained and

described, and reflects current best practice in the pharmaceutical industry in Europe. The implementation of such an optimised design is believed to result in a reduction in the overall numbers of animals used for this purpose, without jeopardising the scientific rationale and b b s .y a o z h .c o m

usefulness of the studies for informing the conduct of later regulatory studies.

Keywords: Reduction; Refinement; Alternatives; Animal use; Dogs; Laboratory animal science; Toxicity tests; Regulatory toxicology

1. Introduction

1.1. Background

The pharmaceutical industry recognises the need to implement

strategies for reducing, refining, and replacing the use of animals in toxicity studies (The Three Rs) (Tweats, 2000), particularly regarding the use of companion species in safety testing (Baker and Broadhead, 2000). To this end, several pharmaceutical companies based in Europe have formed a working party with two scientific and animal welfare charities in the UK, the Royal Society for the Prevention of Cruelty to Animals (RSPCA) and the Fund for the Replacement of Animals in Medical Experiments (FRAME). The formation of this collaborative group in 2000 was prompted by a recommendation made at a workshop held to discuss the use of the dog as a second species in regulatory toxicity testing, which in turn was organised as a result of some preliminary research conducted by FRAME and the RSPCA (Broadhead et al., 1998, Broadhead et al., 1999 and Broadhead et al., 2000).

The principal remit of the working party is to propose and, where possible, put into practice scientifically valid and feasible approaches to optimise dog use in the safety evaluation of pharmaceuticals, without compromising human safety or increasing the use of other non-rodent species. 1.2. Initial approach—analysis of the design of

preliminary repeat dose toxicity studies

As one of its first tasks, the working party identified many potential approaches for optimising dog use, and these have been prioritised for further consideration (Smith et al., 2002). One promising

approach related to the design of repeat dose toxicity studies, and included a comprehensive analysis of the different designs currently b b

s .y a o z h .c o m

being used by the member companies of the working party for preliminary escalating dose/maximum tolerated dose (MTD) and dose range finding (DRF) studies, which precede pivotal repeat dose toxicity testing. The objective of this investigation was to define the principles and requirements of MTD and DRF studies so that an optimised study design could be identified. Such a design should deliver appropriate early information on the characteristics of the test compound in non-rodents and allow a reliable prediction of appropriate dose levels for the 14 day/1 month regulatory toxicity study, thus achieving the most effective use of the animals. The information from this optimised study design should also meet single dose (acute) toxicity study requirements.

1.3. Objective of the paper

We intend to share the results of our co-operation and propose an optimised study design, which was formulated as a result of the sharing of data, current working practices, and experiences among members of the working party. Publication of this paper is also intended to promote dialogue with other toxicologists in the industry, and with regulators, who are involved with generating data based on these tests, or who have to assess such information. The work described in this manuscript is part of a wider effort by the working party to optimise and minimise the use of the dog as a second species in the regulatory safety assessment of new pharmaceuticals. Preliminary findings have been published elsewhere (Phillips et al., 2004; Smith et al., 2003). 2. Results of discussions 2.1. The purpose of preliminary studies

The sharing of information between members of the collaborative group soon revealed that, although each organisation used the MTD/DRF study to select dose levels for the repeat dose, regulatory study, the additional uses to which this study were put varied greatly between companies. As a consequence of this difference in custom and practice between each company, different study designs were in use in nearly all cases. b b s .y a o z h .c o m

In consequence, agreement was sought on the primary and

secondary purposes of such studies, together with reaching

consensus on the nature of any optional data sets that could also be generated, before an optimised protocol could be developed. The results of this discussion exercise are summarised in Table 1 and Table 2.

Table 1. Primary and secondary purposes of preliminary DRF/MTD study in the dog Purpose Description

Primary Dose selection (highest dose) for the pivotal repeat dose studies conducted prior to first dose to man

Secondar

y Detection of serious toxicity (often referred to as target organ toxicity) to confirm the selection of candidate drugs

Obtaining toxicokinetic data at a range of doses to allow optimal dose selection for subsequent repeat dose studies in relation to estimated clinical doses

Assessment of suitability of the dog as the second species

Also used for estimating compound requirements for the subsequent repeat dose studies able 2.

Further information obtainable from preliminary MTD/DRF studies in the dog Parameter

Comments No observed effect level (NOEL)

After single or multiple doses

No observed adverse effect level (NOAEL)After single or multiple doses

Systemic exposure

Measured at each dose level in plasma

Overt clinical signs Characterisation of overt toxicity after single or multiple doses b b s .y a o z h .c o m

Parameter Comments

Dose proportionality Saturation of absorption/elimination

Tachyphylaxis Induction after multiple doses

Tolerability/responsiveness To confirm species selection

Delayed onset of toxicity During the repeat dose phase

Recovery from toxic insult After single or multiple doses, e.g., clinical signs

Clinical pathology Within the constraints of the design After considerable discussion concerning the intended uses of maximum tolerated dose (MTD) and dose range finding (DRF) studies, it was agreed that an MTD study should be designed to detect a tolerance limit based on the results of an in vivo test, following the administration of single, escalating doses of test chemical. It was also agreed that DRF studies should be designed to identify a tolerance limit following the repeated administration of the highest feasible dose, based on the outcome of an MTD study. The working party paid particular attention to make suggestions for improving the design of the MTF/DRF preliminary studies that would allow the collection of data compatible with further regulatory

requirements for non-rodent safety testing, so as to avoid

compromising its overall objective of minimising dog usage. To this end, proposals for a new study design were made so that data

generated by the new protocol would meet the requirements to

replace the single dose (acute) toxicity study in non-rodents, since according to the experience of some members of the working party, the strategy of using MTF/DRF data instead of especially generated single dose toxicity data in non-rodents is well accepted by some regulators (namely those in Japan), which still request this type of information collected in non-rodents. Moreover, some regulatory guidelines provide sufficient flexibility to cater for the submission of b b s .y a o z h .c o m

data obtained in this manner (FDA Guidance for Industry, 1996:

Single Dose Acute Toxicity Testing for Pharmaceuticals; Yakuji Nippo, Japan’s and ICH guidelines for new drug registration, 1999). In addition to the primary and secondary purposes of MTD/DRF studies, the working group identified further information that could also be obtained or derived from such studies (Table 2).

The usefulness of including each of the above requirements in any study design is discussed later. It is most important to balance the welfare implications of intensive animal handling, treatment

procedures and sampling methods with the overall number of dogs used when designing studies.

2.2. Factors affecting design The following factors were considered by the working party to influence the design of any MTD/DRF study. It should, however, be noted that the following list is unlikely to be exhaustive, but reflects the current approach to drug development. 2.2.1. The overall number of animals required The total number of animals per study depends on the number of dose groups and the group size. Thus, any approach to minimise the total number of animals used has to target on limiting the number of separate dose groups and on minimising the group size needed to achieve a sufficient and relevant data set. If it is accepted that such studies are only designed to detect

significant levels of toxicity, then it is possible to use small group sizes, without control animal groups. Data presented later show that the relevance and predictability of such studies do not correlate to the group sizes used, although the fact that there are limits to reducing group sizes should be recognised. 2.2.2. The amount of compound available for

testing It is, of course, mandatory for such studies to have sufficient chemical, or drug supply to allow inclusion of dose levels high enough to b b

s .y a o z h .c o m

produce significant toxicity. Restriction of dose-escalation due to limitation of compound supply can seriously compromise the validity of the preliminary study, for example, due to a failure to identify target organ toxicity, and in consequence can jeopardise the outcome of a subsequent regulatory 14 day/1 month study. Study designs that entail the use of minimal numbers of animals result in parallel in a minimisation of compound need per dose step and thus allow testing of high doses up to the maximum recommended dose (2 g/kg) and a prolongation of repeat dose treatment.

2.2.

3. The use of previous information

Despite the fact that company practices vary, it is not unusual to have information from in vitro and in vivo low dose pharmacokinetic studies, prior to conducting a MTD/DRF study. However, such data are rarely generated in dogs. In addition, safety pharmacology studies in

telemetered dogs may have been performed. Data from these studies, together with those from rodent studies (in general performed before dog studies), should be used to design the MTD/DRF dog study.

2.2.4. Pharmacological class

The mode of action of a test compound can influence the design of the study, for example cell cycle inhibitors that preclude the re-dosing of animals, due to delayed onset of toxicity.

2.2.5. Compound attributes

The physico-chemical properties of a compound, together with its ADME profile (absorption, distribution, metabolism, and excretion), and its intended route of administration, might be such as to require a non-standard study design to be considered. Compounds with very short or very long elimination half-lives fall into this category. 2.2.6. The timing of a study

The total duration of a study may impact the progression of a

compound during the drug development process. In the very early pre-nomination (research) phase of compound development,

information about toxicity or pharmacological activity might be an essential go/no go criterion for further development. When the test is b b

s .y a o z h .c o m

conducted close to first use in man, the availability of the results of the study might be time-limiting for the commencement of the pivotal 14 day/1 month study. In both scenarios, however, there is pressure to minimise study duration.

2.2.7. The need to replace the acute toxicity study

There was general consensus that standard acute toxicity studies in dogs should be omitted. Some members confirmed that data from escalating, or short-term, repeat dose studies had been accepted as acute toxicity data in non-rodents by Japanese regulatory bodies. The regulatory guidelines for both USA and Japan, which still recommend acute toxicity testing in non-rodents, offer this flexibility (FDA

Guidance for Industry, 1996: Single Dose Acute Toxicity Testing for Pharmaceuticals and Yakuji Nippo Ltd, Japan’s and ICH guidelines for new drug registration, 1999). The essential information required by such regulators for such studies in the non-rodent appear to be: (a) identification of dose levels provoking overt signs of toxicity; (b)

inclusion of appropriate parameters and sufficiently long observation periods proper observations (no more details given); and (c)

conducted according to GLP regulations (see below).

No recommendations are made in the guidelines, however, on the number of animals that should be used, but custom and practice indicates that 2 male and 2 female animals are acceptable.

2.2.8. Conducting studies according to GLP regulations

There is a requirement to conduct the dog MTD/DRF study according to GLP standards if information from this study is to be used for risk assessment in man. When this study is performed early in the developmental programme, however, it is neither possible nor necessary to fulfil all the requirements of GLP. Regulators have accepted such studies provided that: (a) the study has been

performed in a GLP compliant facility; (b) the study plan, conduct, and report are subject to audit; (c) an impurity profile is generated in lieu of a compound specification; (d) non-validated methods for

analysis of drug substance and toxicokinetics are documented in lieu b b s .y a o z h .c o m

of fully validated methods; and (e) all deviations from GLP compliance are noted.

2.2.9. The duration of dosing

The predictiveness of the study is increased with increasing duration of the repeat dose phase. However, as with many of the other factors, the ability to increase the duration of dosing has to be balanced

against compound availability and efficiency (speed). It was apparent that most of the companies in the working group have adopted 5–7 days as the default duration, although 14 days is used by some.

3. Obtaining information on current practice The Steering Group distributed questionnaires to companies that were aimed to: (a) elicit current designs of both MTD and DRF phases of the study and (b) examine the outcome of the most recent studies (at least 10) conducted by each organisation.

3.1. Current designs for the MTD/DRF study

The main reasons for the discrepancies in the various designs for the study by individual companies are: (a) the different emphasis that each company applies to the purpose of the study; (b) variations in the extent of data sets each company wishes to generate; and (c) the confidence that each company requires before conducting the subsequent pivotal 14 day/1 month study.

However, it was apparent that most companies use a two-phase

design involving the administration of an escalating dose, followed by repeat dosing, although some important general and many detailed differences became obvious when the designs were compared (Table 3 and Table 4).

Table 3. Examples of high level design differences in the MTD/DRF study An escalating phase followed by repeat dosing

An escalating phase only

Repeat dosing only

An escalating phase during candidate drug selection, with repeat dosing prior to the pivotal study b b s .y a o z h .c o m

An escalating phase followed by repeat dosing

Repeat dosing of 5–14 days duration Table 4. Examples of detailed design differences in the MTD/DRF study

The re-use of animals between different dosing regimens

The re-use of low-dose animals between studies

The use of parallel groups of dogs during escalation phase (resulting in an overall increase in animals)Repeat dosing during the escalation phase

An escalating phase, with or without a washout period

Some of the above differences are clearly due to natural variation in the way laboratories will design experiments to achieve the same aims. However, further differences in the design are also due to the fact that different custom and practice has been established such that the study can variously be used to establish NOAEL values, provide off-dose recovery data, and meet the needs of the single dose (acute) study. The existence of such diversity in designs for the study is, therefore, not surprising. 3.2. Sharing data to develop a common study design Compound data sets were generated by each participating company from previous studies to reveal: (a) the number of dogs used in the escalating phase of the study; (b) the number of dogs used in the repeat dose phase; (c) the therapeutic indication area of a given project; (d) what route of administration had been used; and (e) the outcome of the subsequent pivotal 14 day/1 month study in terms of whether it was judged a success or a failure.

It was agreed that the criterion for judging whether the MTD/DRF study had been successful was the ability to select an appropriate high dose level for the subsequent pivotal 14 day/1 month study that elicited target organ toxicity or, defined the maximum technically feasible dose. The MTD/DRF study was judged a failure if additional animals or an additional dose group had to be used in the 14 day/1 month study, or if this study had to be repeated due to inappropriate dose selection.

An analysis of the results from 101 pivotal 14 day/1 month studies is illustrated in Fig. 1. The data show that preliminary dose-setting studies that involved the use of up to 4 dogs are as likely to be b b s .y a o z h .c o m

successful in predicting appropriate dose levels for the subsequent 14 day/1 month study, as are those studies that involved substantially more.

Fig. 1. Number of animals vs. success or failure.

For those tests classified as having failed, the cause could not be attributed to drug target or therapy area (Table 5).

Table 5. Study outcome related to therapy area Therapy area Number of compounds

Success

Failure Metabolism 12

2 CNS

28 4

Cardiovascular 16 3 Inflammatory 2 — Oncology 8 — GI tract

6 — Respiratory

8 1 Infection

6 — Dermal

1 — Urogenital

1 — Hormonal

2 — Imaging 1 — It is not possible to assess whether route of exposure affected the outcome of the studies as the oral route was used in 95/101 cases with intravenous, intravenous infusion, and subcutaneous injection each used in 2 studies. All of these six studies were classified as b b

s .y a o z h .c o m

success, but the imbalance in the routes of exposure precludes any conclusions being made.

4. Discussion

4.1. A proposal for an optimised study design

Having taken into account the primary objective of the preliminary MTD/DRF studies, and after considering the information generated from the review of previous studies, the Steering Group proposed a basic design involving an escalating dose phase involving one male and one female dog dosed to the MTD, followed by repeat dosing (of >4 days duration) in one male and one female na?ve animal at the MTD, with additionally one male and one female non-na?ve animal at the same dose or a lower dose.

On occasions when animals from the escalating dose phase cannot be re-used in the repeat dose phase (if the MTD is exceeded), a further two animals would be required (total six animals).

The above design has been demonstrated to be satisfactory and should be sufficient for almost all circumstances, although, as is the case with the design of all toxicity studies, it is necessary to consider each project on a case-by-case basis. An alternative design, which has also been shown to be satisfactory and could be considered is a repeat dose study involving three dose levels in one male and one female dog.

Using these basic high-level designs, the toxicologist can then

consider the detailed study parameters, which should be included in a study on a case-by-case basis. These parameters are discussed later. One important factor to consider in the escalating dose phase is the need for adequate wash-out (off dose) between administrations during the dose escalation phase of the study. In the absence of pharmacokinetic data, it is usual to fix the interval between

successive dosings at 2 or 3 days. Care is also needed when

evaluating cytotoxic compounds and those that accumulate in the body or that exhibit delayed toxicity. 4.2. Study parameters b b

s .y a o z h .c o m

Several parameters could be included on a case-by-case basis in an MTD/DRF study and these are listed and briefly discussed in Table 6. Table 6. Parameters that could be included in preliminary MTD/DRF studies in dogs End-point Note

Clinical signs Probably one of the most important parameters both in terms of identifying overt

adverse effects and also managing their outcome. Frequent observation is essential as

this parameter is most likely to be the one to establish the MTD

Body weight A very sensitive indicator of toxicity and regular recordings are important

Food

consumption Closely associated with body weight changes

Water

consumption Rarely measured

Haematology/

clinical

chemistry Should be performed at each dose escalation and at the completion of repeat dose phase. After pathology, it is the laboratory investigation most likely to detect toxicity

Urinalysis

Less useful and could be excluded. May be considered in the future for metabonomics

ECG The value of ECG data on less than 4 dogs is limited and is better obtained from Safety Pharmacology using telemetry. However, because of the higher dose levels used in the MTD study it may be useful to analyse data from the highest dose animals. Measurements should include T max

Blood

pressure

Similar consideration as ECG. Effects on BP will not usually prevent a compound going into man and therefore this parameter is not routine

Toxicokinetics Essential component of each phase. Rapid feedback of plasma level data can avoid

unnecessary dosing of animals when saturation of absorption is observed b b s .y a o z h .c o m

End-point Note

Ophthalmolog

y Investigations are non-invasive and should be considered at the end of the repeat dose phase

Gross

pathology Can indicate additional tissues for pathological examination. Necropsy should be performed within two to three days of cessation of dosing (repeat dose phase)

Histopatholog

y Essential to identify target organ toxicity. Major organs/tissues, abnormal tissues and others on a project-specific basis should be examined

Cytochrome

P450 levels Should be considered if findings from previous rat studies indicate induction of metabolism

5. Conclusions

The work of the Steering Group illustrates an initiative by industry and animal welfare to share unpublished data, with the aim of reducing animal usage, thereby improving animal welfare, without compromising scientific validity. In addition to the intrinsic value of reducing animal numbers the project has also been able to provide a mechanism for adding value to the design of animal studies. The worth of sharing unpublished data has been demonstrated by the ability of the group to propose an optimised design for the preliminary toxicity studies in the dog. These preliminary studies are not

undertaken for regulatory purposes, and so it is possible to make progress without change to the regulatory framework. The proposed designs, involving the use of 4 or 6 dogs, will result in the use of less animals. The actual level of reduction in animal numbers is difficult to quantify, but the overall projected impact is considered to be

significant, bearing in mind the welfare cost to each animal in this type of experiment and the number of such experiments undertaken globally in the pharmaceutical industry. It should also be remembered that these studies are performed early within a project life, and many of the compounds subjected to this study do not progress to the market. b b s .y a o z h .c o m

The recommended study design is supported by the data, but it is nevertheless a compromise based on experience. As with all study designs there might be a need for a ‘case-by-case’ approach for

candidate compounds intended for specific therapeutic areas and for individual requirements for each company.

As a result of our collaborative study, it is recommended that

companies review their study designs with the aim of adopting our proposed new study design for reduced animal numbers. A number of Steering Group members have already done so but not without a change of mind set.

References Baker and Broadhead, 2000 R. Baker and C. Broadhead, Should cats and dogs be used as laboratory animals In: M. Balls, A.-M. vanZeller and M. Halder, Editors, Developments in Animal and Veterinary Sciences, 31B: Progress in the Reduction, Refinement and Replacement of Animal Experimentation , Elsevier, Oxford (2000), pp. 1725–1730. Broadhead et al., 1998 Broadhead, C.L., Jennings, M., Heywood, R., Combes, R.D., 1998. The use of dogs in regulatory toxicology. In: O’Donoghue, P.N. (Ed.), The Ethics of Animal Experimentation: Proceedings of the European Congress held 17–18 December, 1996, at the Palais des Congres, Brussels. EBRA, London, p. 261. Broadhead et al., 1999 Broadhead, C.L., Jennings, M., Combes, R.D., 1999. A Critical Evaluation of the Use of Dogs in the Regulatory Toxicity Testing of Pharmaceuticals. FRAME, Nottingham, UK (ISBN 0 9501 700 5 4), 106 pp. Broadhead et al., 2000 Broadhead, C.L., Betton, G., Combes, R., Damment, S., Everett, D., Garner, C., Godsafe, Z., Healing, G., Heywood, R., Jennings, M., Lumley,

C., Oliver, G., Smith,

D., Straughan, D., Topham, J., Wallis, R., Wilson, S., Buckley, P. (observer) 2000. Prospects for reducing and refining the use of dogs in regulatory toxicity testing of pharmaceuticals. Human and Experimental Toxicology 19, 440–447. Phillips et al., 2004 B. Phillips, D. Smith, R. Combes, G. Descotes, S. Dyring Jacobsen, R. Hack, J. Kemkowski, K. Krauser, R. Pfister, Y. Rabemampianina, S. Sparrow, M. Stephan-Gueldner and F. von Landenberg, An approach to minimise dog use in regulatory toxicology: production of a best practice guide to study design, Alternatives to Laboratory Animals 32S1 (2004), pp. 447–451. Abstract-EMBASE | $Order Document b b s .y a o z h .c o m

Smith et al., 2002 D. Smith, C. Broadhead, G. Descsotes, R. Fosse, R. Hack, K. Krauser, R. Pfister, B. Phillips, Y. Rabemampianina, J. Sanders, S. Sparrow, M. Stephan-Gueldner and S.D. Jacobsen, Preclinical safety evaluation using nonrodent species: an industry/welfare project to minimize dog use, ILAR Journal, Suppl. (2002), pp. 539–542. Smith et al., 2003 D. Smith, R. Combes, G. Descotes, S. Dyring Jacobsen, R. Hack, J. Kemkowski, K. Krauser, L. Lammens, R. Pfister, B. Phillips, Y. Rabemampianina, S. Sparrow, M. Stephan-Gueldner and F. von Landenberg, Approaches to the minimisation of dog use in the safety assessment of pharmaceuticals: an industry/animal welfare initiative, Toxicology Letters 144 (2003) (Suppl. 1), p. 48. Abstract-Compendex | Abstract-INSPEC | $Order Document Tweats, 2000 D. Tweats, A review of the reduction and refinement of regulatory toxicity studies for pharmaceuticals In: M. Balls, A.-M. van Zeller and M. Halder, Editors, Developments in Animal and Veterinary Sciences, 31B: Progress in the Reduction, Refinement and Replacement of Animal Experimentation , Elsevier, Oxford (2000), pp. 783–791.

b b s .y a o z h .

c o m