Drug development is plagued by ineffectiveness or toxicity in significant proportions of the population. This is because blockbuster drug dose is based on the results from cumulative response curves. So the dose is decided by (for example) above 80% of the population respond. However, at this point there will be people that are under-responsive and over-responsive at this dose (figure 1). So they will get no effect or get toxicity at this dose. With certain drugs this is fatal, with 100,000 deaths a year from drug toxicity and 4 million serious adverse reaction events. Variations exist through polymorphisms and drug interactions mostly, the following will discuss some contributions to variability in drug response with some experimental examples.
Absorption, distribution and excretion
There is less variation seen in these processes than metabolism but there are some important variations that can play a big role in drug effectiveness. Absorption is generally similar in most people but disease state can cause variability, for example, 32% of patients requiring irinotecan have diarrhoea, therefore oral administration would create considerable variation with those with diarrhoea absorbing a lower dose than those without. This has to be a consideration in drug development because it suggests the drug should not be developed to be delivered via this route.
Drug distribution and excretion also become variable with disease state, distribution rate is increased if as the disease progresses hypertension and tachycardia increase. In contrast, kidney damage and failure in late stages of disease reduce the body’s ability to excrete the drug, allowing it to be active longer thus increase response.
Metabolism
Metabolism activates and inactivates drugs, so changes to metabolism will cause variation in response because reducing either of these will either lose the effectiveness of the drug or increase the action well over the usual response and be toxic. Many drugs are heavily metabolised by CYP450 enzymes and when an enzyme is metabolising one it cannot metabolise the other. This can cause variability in drug response because one person may be on the drug alone whereas anoth may be taking other drugs. This is exemplified by mibefradil which is safe alone but when administered with propranolol, inhibited CYP450 enzymes, preventing the inactivation of propranolol. A study conducted looked at patients taking both and found systolic blood pressure decreasing from over 140mmHg systolic to 60-70mmHg a fatal level. This was found to be because propranolol was not being inactivated by CYP450s and thus acting excessively on the heart. This drug was soon discontinued.
Other than drug interactions inhibiting or increasing cYP450 metabolism of another drug, CYP enzymes also get polymorphisms, relatively commonly. CYP2C9*2 is the mutant allele in 20% of Caucasians, this is a significant number of people with a loss of function of an important enzyme in drug metabolism. Testing of how much variation the presence of this allele causes showed as much as 30% differences in dose of warfarin required for efficacy. This is a large variation because if a full dose is 30% more effective it is also likely to be toxic at this level. This becomes a problem in drug development because it makes it impossible to give an accurate recommended dose and makes it likely the drug will fail in clinical trials due to toxicity. Evidence that these polymorphisms contribute heavily to variation in drug response between humans, comes from studies adding CYP2C9*2 allele to transgenic mice but using a pharmacogenetics approach of genotyping and phenotyping activity and adjusting the warfarin dose given based on the result. They found that doing this required fewer adjustments to the dose and fewer incidences of toxicity due to too high doses. This suggests that CYP polymorphisms cause lots of variations to metabolism, but pharmacogenetics can be used in drug development to get the correct dose much faster and safer and almost eliminate the variation that is seen.
The future for genetic CYP testing could be in producing transgenic mice with CYP mutations as part of animal models to identify what changes in dose are required for each polymorphism. Evidence it is accurate from measuring CYP2D6 polymorphisms to quinidine and measuring the changes in response to doses and changing dose until variability is within acceptable levels.
CYP polymorphisms are not the only causes of variation in response, other important enzymes also cause variation in response between people including thiopurine methyltransferase (TPMT). TPMT*3A-C polymorphisms cause a total loss of function of this enzyme which allows a build-up of active thiopurines this in turn leads to an increase in efficacy and toxicity which eventually leads to myelosuppression. In this case variation has been fatal, because the toxicity is quickly fatal and a full dose to some heterozygote mutants is 50% higher than the dose should be and 15% of people are at least heterozygous for a TPMT mutation. Again the relationship with drug development is that drug cannot be developed to suit most but not all because it can be fatal. This could lead to a loss of millions at clinical trials or even more if it has to be discontinued. This is more evidence that variation in patients has to be accounted for in drug development, therefore pharmacogenetics may be the approach in rug development to create more individualised medicines from drugs that are technically block busters, it should increase success in clinical trials along with reducing variation.
Non-metabolic genetics
Aside from metabolism genetics comes into other variations between patients. The same cancer can have different responses to the same drug in different people. This is also a type of variation in response. Also taken into account is the stage of cancer because as the cancer progresses mutations are random so the same cancer can look very different by stage 4. A good example of variation to the drug is the drug Herceptin for breast cancer. It acts on the HCN2 receptor and is very effective but only in 25% of cases because men do not have the receptor and the receptor can be down regulated early in the cancer. This shows the variation in drug response because the drug is very effective at treating breast cancer, but the receptor is not present in all people or all breast cancers. This causes a problem in drug development because it is obvious the drug is very effective but getting significant results is difficult because of how variable cancers are.
In summary, variation in drug response is seen more commonly in metabolism but absorption, distribution and excretion also lead to variation between people, mainly due to what state the disease is in. metabolism variation exists mainly through one drug changing the enzymes action on another drug or mutations to important enzymes. However, it is important to note that sometimes the genetics of the disorder are variable as in cancer.
