NonClinical Toxicology of AntiCancer Drugs

1
Non-Clinical Toxicology of Anti-Cancer Drugs

• Andrew Makin, LAB Scantox
• 27 September 2006

2
Points to cover in this presentation

• Why do we need animal testing?
• Which animal studies must be performed before
  starting Phase I and II studies in cancer
  patients in the EU?
• How should animal studies be designed?
• How does animal toxicology correlate with
  toxicity in patients?

3
Key Legislation

• Testing of medicinal products is governed by EU
  Directive 75/318
• Guidance on the requirements of the Directive is
  given in CPMP/ICH/286/95
• Pre-clinical testing of anti-cancer drug products
  is covered by a further guidance note from the
  European Medicines Evaluation Agency (EMEA)
  CPMP/SWP/997/96

4
Special populations

• Normally the first human patients to receive a
  new medicinal product are healthy (males)
• Patients suffering from cancers are a special
  population
• Therapies are aggressive with many side effects
• They are not suitable to be given to healthy
  volunteers
• In the following series of slides I will try to
  show what tests are normally required before
  going into healthy humans withstandard
  medicines.
• I will also show how the requirements are
  modified for the special population of cancer
  patients

5
Special populations

• Remember that in any evaluation of new medicines,
  doctors in particular have to make a cost-benefit
  analysis where the risks associated with the
  treatment are weighed against the possible
  positive outcomes
• Cancer is a life-threatening disease and
  therefore a greater risk of potential therapies
  is acceptable in comparison with lesser, non
  life-threatening conditions

6
Preclinical Testing Requirements

• Note that the testing requirements discussed
  relate to single drugs. Combination therapies
  will require special consideration e.g.
  pharmacokinetics and interactions
• In addition, I am for the most interested in what
  is required for a first into man programme of
  studies Based on EMEA requirements (Note for
  Guidance on the pre-clinical Evaluation of
  Anticancer Medicinal Products CPMP/SWP/997/96)
• Testing falls into 6 broad categories
• Acute studies
• Genetic Toxicity
• Repeat Dose Toxicity
• Safety Pharmacology
• Reproductive Toxicology
• Local Tolerance

7
Preclinical Testing Requirements - 2
8
Preclinical Testing Requirements - 3
Tests generally not required because the clinical
conditions are assumed to have no available
alternative treatment
9
Preclinical Testing Requirements - 4
10
Preclinical Testing Requirements - 5
11
Preclinical Testing Requirements - 6
12
Preclinical Testing Requirements - 7
13
Regional variations

• Note that there have been regional variations in
  pre-Phase I requirements
• US FDA has required pre-clinical studies in
  rodent (rat or mouse) and non-rodent (normally
  dog) prior to Phase I (rodent study identifies
  potential life-threatening toxicities and dog
  study shows whether doses tolerated by rodents
  are life-threatening in a non-rodent species)
• US FDA has a stronger requrement for embryofetal
  studies
• European Phase I studies can be supported by
  rodent data alone
• Note, however, in summary, for FDA there are
  basically only 2 non-clinical studies that are
  absolutely required

14
Non-clinical study designs vs clinical

• Often human patients are dosed on a cyclical
  basis (e.g. x days dosing followed by y days
  recovery)
• This is because more aggressive therapeutic
  regimes cause unwanted side-effects due to the
  toxic nature of the drugs in question (e.g. bone
  marrow effects, GI disturbance etc).
• The same happens with animals, but the recovery
  period could differ due to species differences in
  responsiveness
• Therefore animals will be dosed for the same
  duration (i.e. x days), but the recovery period
  could be longer or shorter i.e. y days /- z days

15
Non-clinical study designs vs clinical

• If the human exposure is not followed in the
  animal species, then the predictivity of the
  animal studies is impaired or rendered irrelevant
• E.g. if animals are dosed continually and humans
  will be dosed intermittently, this could result
  in an underestimate of the human starting dose

16
Non-clinical study designs vs clinical

• A key point in the animal studies is that
  toxicity of the drug is of the greatest
  importance.
• For a normal NCE, human doses and inter-species
  comparisons of dose and effect are based on
  pharmacokinetics (exposure, AUC etc)
• For anti-cancer drugs, PK is not an essential
  part of the non-clinical toxicity studies, and
  doses are based almost entirely on toxicity
• This comes back to the earlier point that the
  cost-benefit analysis for anti-cancer drugs is
  different than for other therapies

17
How do the non-clinical studies predict for human
studies?

• The body surface area concept in dose
  normalisation
• In order to get a direct inter-species comparison
  of doses, the actual doses given (in mg drug/kg
  bodyweight) are converted to mg drug/m2.
• In simple terms, the mouse dose in mg/kg is
  multiplied by 3the rat dose is multiplied by
  6and the dog dose is multiplied by 20to give
  the equivalent human dose
• For practical reasons, doses for individual
  animals are measured on a kg bodyweight basis

18
How do the non-clinical studies predict for human
studies?

• Selection of a starting dose in humans
• One tenth of the rodent LD10 or STD10 (Severely
  Toxic Dose)
• or
• One sixth of the dog TDH (Toxic Dose High the
  highest dose tested that does not cause severe,
  irreversible toxicity in dogs)

19
How do the non-clinical studies predict for human
studies?

• Using Platinumcontaining compounds as an example
  
• FDA has published a review of the predictive
  value of preclinical toxicology studies for
  platinum anticancer drugs (Clinical Cancer
  Research, 5, 1161-1167, May 1999)

20
How do the non-clinical studies predict for human
studies?

• 12 Platinum analogues for which data had been
  submitted to FDA were selected (all studies had
  preclinical and clinical data from matching drug
  administration schedules)
• Doses were correlated on an mg/m2 basis and
  showed
• One third rodent LD10 or one third dog TDH would
  have predicted starting human dose and first
  escalation without exceeding clinical MTD
• Human dose limiting toxicities were well
  predicted from non-clinical studies
• Human starting dose and dose limiting toxicities
  could have been predicted from rodent data alone

21
How do the non-clinical studies predict for human
studies?

• Recurring themes in the platinum compound review
  and in other reviews comparing preclinical with
  clinical data are
• Rodent data alone are probably sufficient to
  predict both a safe starting dose for humans plus
  likely human toxicities
• Use of current methodologies (e.g. one tenth of
  rodent LD10) to predict human starting dose
  actually give a human dose that is generally too
  low, resulting in more escalation steps than are
  necessary in human studies.

22
A final health warning

• The preceeding comments that I have made relate
  principally to new anti-cancer drugs whose mode
  of action is principally cytotoxic.
• There are many cases where special considerations
  may apply e.g.
• Combination products
• Specialised delivery systems e.g. liposomal
  encapsulation, conjugation to antibodies
• Light activated drugs

23

• Thank you for your attention