Antibodies (mAb) are important in the development of targeted therapy. Different
variations of mAb share similar properties (cytotoxicity/ neutralising
cytokines) but can differ from one another aspects like mechanism of action (CHMP/437/04, 2014). The structure of mAb is
complex, and a single molecule can have multiple functional domains depending
on the isotope. Individual mAbs have unique profiles regarding antigen-binding
regions, FC binding receptors and Fc cytotoxic effector function. Numerous
assays have been developed over the years that allow for in-depth
characterisation of complex proteins based from their physiological and
functional properties. The use of non-clinical and clinical testing is used to
validate the use of the newly produced medicinal product by verifying its
safety, effectiveness and quality. The difficulty increases, however, when the
same testing procedure is applied to determine the significance of minor
changes of physiochemical and biological characterisation when comparing the
mAb to a biosimilar mAb. A biosimilar is a biological medicinal product that
contains the version of the active substance of an already authorised reference
medicinal product. Biosimilars are similar to the reference mAB in terms of
physiochemical and functional properties and any relevant changes observed in
testing must be justified (EMA/CHMP/BWP/247713/2012, 2012).
In order to create an appropriate non-clinical design,
the characteristics of the reference product needs to be clearly understood (EMEA/CHMP/BMWP/42832/2005, 2014). It is not recommended to
conduct repeated dose toxicity studies in non-human primates, as well as conducting
toxicity studies in non-relevant species (EMA/CHMP/BMWP/403543/2010, 2012). By alternating the production
process used by the reference and biosimilar product manufacturers, qualitative
differences of impurities during the process will occur. Process related
impurities should be kept as low as possible to minimize risks associated with
impurities. Quantitative and qualitative differences of variations of products
can affect the biological function of the mAb and need to be assessed using
appropriate in vitro assays. The observed differences in quality could affect
the immunogenic potential and the probability to cause hypersensitivity (EMEA/CHMP/BMWP/42832/2005, 2014). These effects are hard
to determine during the non-clinical phase using animal studies and should be
assessed further in the clinical studies.
The assessment of immunogenicity in animals may not be
used to predict immunogenicity in human. They can be used, however, to
interpret data obtained in in vivo studies in animals. Blood samples are
required for future evaluations and should be taken and stored. Safety
pharmacology and reproductive toxicology studies are not needed during the
non-clinical testing of biosimilars (EMA/CHMP/BMWP/403543/2010, 2012). Studies
on local tolerance are generally not required except when excipients are
introduced where there is no or little information regarding the intended
clinical route. When other in vivo studies are conducted, local tolerance can
be evaluated as part of the design of the study instead of the performance of
separate local tolerance studies.
trials for biosimilars involves comparability testing which is usually a
stepwise procedure that begins with pharmacokinetic (PK) and, if possible,
pharmacodynamic (PD) studies. This is followed by Clinical safety and efficacy
trials or, in some cases, confirmatory PK / PD studies to demonstrate clinical biosimilar
comparability (EMA/CHMP/BMWP/403543/2010, 2012).
the comparison of target-mediated clearance is of major importance in the
biosimilarity exercise, it may not be feasible in patients due to major
variability in target expression (EMEA/CHMP/BMWP/42832/2005, 2014), including variability
over time PK studies are not always feasible in healthy volunteers. In this
case, the PK needs to be studied in patients as part of a multiple dose study,
if a single dose study is not feasible. PK study of intravenous administration
needs to be justified, e.g., in cases when the molecule has an absorption
constant which is much slower than the elimination constant (EMEA/CHMP/BMWP/42832/2005, 2014).
may arise when applying traditional clinical testing methods to biosimilars and
reference mAb under an anticancer setting (EMA/CHMP/BMWP/403543/2010, 2012). Preferred
endpoints under that setting are designed for the benefit to the patient, but
may prove inefficient when trying to establish comparability of biosimilars to
the reference product regarding sensitivity/feasibility. This can be due to
outside factors not related to the differences to the biosimilar and the
reference (e.g. tumours) (EMA/CHMP/BMWP/403543/2010, 2012). It can then be
concluded that the clinical testing is not suited to establish similar efficacy
of the biosimilars and reference mAbs.
tests are designed to show that the biosimilar and the reference medicinal
product are similar to each another regarding safety and efficacy, not in the
patient (EMA/CHMP/BMWP/403543/2010, 2012). The reason behind comparability
testing between the biosimilar and the reference product is to lower remaining
uncertainty following extensive pharmacokinetic, analytical and in-vitro
testing (Kay and Isaacs, 2017). The efficiency of the
reference was approved during the pivotal clinical trials, followed by later
knowledge obtained from clinical practice. Therefore, if bioequivalence between
the reference and the biosimilar can be shown, re-doing the clinical trials
would be redundant. Thus, biosimilar clinical trials can be viewed as bioassays
to show that it produces a therapeutic effect similar to the reference drug,
with a disease that the original drug was approved to treat (Kay and Isaacs,
principles allow for subsequent extension of regulations to other licensed indications
of the reference product, given that therapeutic efficacy relies on a similar
mechanism of action in the extrapolated indications (Kay and Isaacs, 2017).
for comparison testing it is preferable to have a sensitive population and
endpoint to detect product related differences and minimising risks associated
with patients and disease to improve precision (EMA/CHMP/BMWP/403543/2010,
2012). Clinical testing in a homogenous population measuring activity as a
primary endpoint could be considered. That can be achieved by testing for the
Overall Response Rate (ORR), counting the patients where a Complete Response
(CR) or a Partial Response (PR) was observed (EMA/CHMP/BMWP/403543/2010, 2012).
Clinical Trial Design
are several new approaches to the clinical trial design that can be used to
enhance the drug development process.
and simulation can be used to improve dose selection and study design during
the second clinical phase. It can also be used to determine dose–response and
time–response behaviour of safety and efficacy endpoints (Orloff, J. et al., 2009). When used along with the
Bayesian methods, this approach can deliver a continuous movement of
information through the phases of development (Orloff, J. et al., 2009). Using
modelling during the early development can allow for the use of external
information, (important discussing about sharing placebo data across companies in
the industry) (Orloff, J. et al., 2009), potentially increasing the efficiency
of studies significantly during early development.
Phase three, simulations can be used to show the effect that different study
designs have on the outcome and the probability of success (Orloff, J. et al.,
2009). This can be used to choose the best course of action when carrying out
development. Determining the potential of success can be enabled by combining
numerous sources of data from previous studies done on the drug and external
data which can prove informative when making decisions (Orloff, J. et al.,
2009). Modelling and simulation techniques have the advantage of being used not
just during trial design process but also during the adaptive trial design.
4.2 Bayesian methodology
methodology uses probability models providing information about parameters of
importance, like the therapeutic effect of drug (Menon et al.,
2017). This approach applies the principles
scientific methods combine previous information with data observed resulting in
improved updated information. Using Bayesian methodologies, previous
information obtained about the parameters are summarised in their prior distribution
(Orloff, J. et al., 2009). After that, new data values are obtained through
experimentation and the probability distribution of the data
obtained leads to the probability function (the observed evidence on the parameters).
Bayes’ theorem is used to combine the two sets of data, resulting in the
updated information with the observed evidence (Orloff, J. et al., 2009).
approach used interim data obtained from a trial to improve and modify the
study design, in a pre-planned manner and without adversely affecting validity
or integrity (Mahajan and Gupta, 2010) During clinical phase two,
an adaptive trial can allocate large number of subjects to the treatment group efficient
performance, drop group that are performing inefficiently (Orloff, J. et al.,
2009), and can be used to research a wider range of doses to select doses that
are most likely to succeed in phase three more effectively.
clinical phase three, the approach can enable early recognition of effective
treatments, deciding to drop inefficient trial groups, determining whether the trial
should be stopped for ineffectiveness and ensuring the trial is driven
adequately by making sample-size adjustments at interim time points (Orloff, J.
et al., 2009).
designs combine, the objectives that are tackled in separate trials, in a
single trial. It does this by using data obtained from all stages of
development, such as seamless adaptive Phase II/III trials integrity (Mahajan
and Gupta, 2010).
size re-estimation methods
methods provide flexibility to either increase or decrease the sample size at
an interim point in the trial. This can be useful in the event which there is an
uncertainty about between-subject variance in the response or uncertainty about
the clinically meaningful effect size at which to drive the trial (Orloff, J.
et al., 2009). The approach allows for the study to start with a certain sample
size that can be increased or decreased at an interim point, and allows for an
The current non-clinical and clinical trials are
suitable for testing any newly developed mAb as they are designed for the patients’
needs and benefit. The endpoints being tested in comparability testing,
however, are not in line as its goal is to prove that the biosimilar is like
the mAb it originated from both in physiochemical and biological functionality.
Testing for the safety and efficacy of the biosimilar is redundant as the
biosimilar as the tests were completed in the development when its reference
product was newly developed. This results in several tests used in the trials
being unnecessary, leading to a loss in time and money. Should an appropriate
novel approach to clinical trial design be implemented, one that acknowledges
the information previously obtained can be applied to the biosimilars, then it
would cut money spent on the drug development process significantly as the
amount of time and trials required would be reduced. In conclusion the current
testing procedure in the drug development process needs to be modified to adapt
to the comparability testing of the biosimilar mAb to their reference.