Pharmacokinetics/Pharmacodynamics of Protein Drugs – Module 2, Session 7

Pharmacokinetics/Pharmacodynamics of Protein Drugs – Module 2, Session 7


>>William Douglas Figg:
I’m excited to introduce today’s lecturer Dr. Jergan Venitz. Jergan is currently Professor in multiple
departments at Virginia Commonwealth University School of Pharmacy. He’s also the Vice Chair of the Department
of Pharmaceutics and the Director of the Pharmacokinetic and the Pharmacodynamics Laboratory at the
school of Pharmacy. In addition, he’s a Fellow in the Center for
Study of Complex Sciences at VCU. In 1981, Jergan received his MD and Ph.D.
from Cologne University in Germany. From 1981 to 1985, he was the Director of
Clinical Research and Development at the Institute for Clinical Pharmacology in Günster, Germany,
then completed post-doctoral fellowship training at the University of Florida. In 1988, Jergan joined the faculty of VCU
School of Pharmacy. He is published extensively and presented
extensively in the area of Quantitative Pharmacology. I’m confident you will enjoy today’s lecture.>>Jergan Venitz:
Welcome, this is Jergan Venitz. I am your entertainer for today as part of
the Principles in Clinical Pharmacology course at NIH. I’m a clinical pharmacologist and a professor
at the Virginia Commonwealth University School of Pharmacy. My topic for today is PK/PD of protein drugs,
so we are going to get into the world of Biotech. So, if you look at the outlines and the objectives
that I put together, we talk about how different those proteins are in terms of their absorption,
distribution, metabolism, and excretion. How different they are from small molecular
drugs that most of us are quite familiar with. We specifically focused then in on PK/PD relationships
of growth-stimulating factors and monochrome antibodies that are commonly used. And then, I’ll wrap up today’s presentation
by getting into something that is very current, and that is the approval of biosimilar drugs
in the United States with are the equivalent of generic small molecular drugs. As you can see, I have a bunch of resources
that are available when you download the class material. My outline is such that we talk about ADME
first, and then we get into PK/PD. We discuss something that is fairly unique
to protein drugs and that is they’re ability to elicit immune response. We’re going to spend a significant amount
of time talking about monochrome antibodies, then go into two examples; one example that
I’m going to lecture on, the second example that I put in as an exercise that you do on
your own and I review the key with you, and then, we get into the biosimilar drugs. All right, so let me just start off by saying
that protein drugs are large molecule drugs are very different form small molecular drugs
that I define as less than 1000 daltons. Now, for those small molecule drugs, the couple
things that we know about in general. That they are subject to passive diffusion
across membranes lipophilic drugs. And, or they may require — if they’re polar
drugs, they may require active drug transports. Okay, and we pretty much know a lot about
the various drug transporters. Again, the depending on the lipophilicity,
they can be plasma bound or tissue bound or not. Depending on the lipophilicity or polarity,
they can also be subject to primary excretion by the kidney or hepatic metabolism or biliary
excretion. And their drug targets the receptors that
we — enzymes that we are trying to inhibit or somehow effect are extravascular. For quite a few drugs, their tissue distribution
organ flow limited so we can draw some fairly general conclusions, for example about liver
clearance and renal clearance. And you can see none of this is true for protein
drugs. So, for small molecular drugs, the main physical
chemical property in general is the lipophilicity polarity. For protein drugs, as you’re going to see
in a minute, is much more molecular weight and their particular structure especially
for glycoproteins. So, here you have a general PK/PD review or
overview that the drug has to get absorbed into the blood. For protein drugs, that’s usually by intravenous
infusion or subcutaneous injection. The drug can interact with plasma proteins,
and you’re going to see for protein drugs, that’s not usually something we worry about. They can distribute into peripheral tissues,
including the tissue that contains the receptors or the targets that we call the effect compartment,
and can interact and cause a pharmacological response. And, the drugs are eliminated by hepatic metabolism
and/or renal excretion, okay. The top part here, anything that relates to
how the dose turns into a concentration, we refer to as kinetics. Anything down here, how the drug bound to
the receptor turns into a pharmacological response, we refer to as pharmacodynamics. Now, if you look at the absorption, that’s
where the first major difference starts to emerge. The protein drugs have large molecular weights,
much larger than the 1000 they use as a cut off, which makes them basically unable to
permeate membranes – GI membranes. So, they would not have any GI permeability
to speak of and sometimes, even more importantly, they are subject to all kinds of chemical
enzymatic degradation. So, there would be a lot of first pass effect
from them from ever reaching systemic levels. Therefore, they are not suitable for all administration,
as I mentioned before. They are usually given intravenously, subcutaneously,
and sometimes intramuscularly, depending on the dose and the volume. Now, even that you have to keep in mind if
you give them by any route other than by the IV route, where they’re being injected or
infused directly into the bloodstream, you give them subcutaneously into a subcutaneous
depot or IM into a muscular depot, you don’t necessarily expect them to be 100% bioavailable,
because what can happen locally is degradation in those tissues. For example, proteolysis in the lymphatic
system. Furthermore, so a reduction in bioavailability,
even if you give them by parental routes. In addition to that the uptake or the absorption
from those tissue compartments for large protein drugs in particular, involves the lymphatic
system rather than the blood capillaries. So, you give a small molecular drug — or
IM the absorption from the side that you setting is bi-capillaries, which are relatively rapidly
profused. On the other had if you give the same — if
you give a different drug, a protein drug by the same route you now have to depend on
the lymphatic system to absorb the drug into the bloodstream, which is convective by nature,
rather than diffusion and blood flow limited. Therefore, it’s much, much more slowly. Distribution, again, due to their large size,
which limits their mobility across biological membranes, most protein drugs are pretty much
restricted to the vascular space for the most part. That means you would expect that volume the
distribution to be, if it’s purely intravascular, three to five liters, for a 75 kilogram per
person. Or if they can at least get to the interstitial
space, it would be around 15 liters per kilogram. So, those are relatively speaking small volumes
of distribution that you expect for protein drugs in general. In addition to that, as I eluded to before,
they’re not expected to be plasma protein bound, like a lot of small molecular especially
prolific drugs do. However, they can be taken up or bind to extravascular
tissues and that’s particularly true for monochrome antibodies that target in the blood or in
plasma. And that would increase their vulnerable distribution
beyond the three liters of intravascular space. Just like the absorption from the subcutaneous
or intramuscular depot, any extravascular distribution of protein drugs occurs by convection
and possibly transcellular endocytosis, which are slow processes. So, you can see here, the extra — I’m sorry
the vascular space but along with the flow of fluid, water, and electrolytes, by convection,
you have a flow of protein drugs. Okay, and you can see that the clearance by
the lymphatic system is much, much faster than the clearance by externalization, which
means that they are distributed just like the absorption in general, but extravascular
distribution is very slow. Now, how do protein drugs get eliminated? Again, different from small molecules which
are typically eliminated by the liver and the kidney. They are basically — protein drugs, that
is — are basically subject to the same pathways that dietary or endogenous proteins are subject
to. So, they can form a myriad of metabolites. Ultimately, the terminal metabolites are the
amino acids and whatever carbohydrates were part of the original protein. And those, especially the amino acids are
ultimately recycled into the endogenous pools, so they’re not necessarily excreted into the
bile or the urine. The liver excretion or the renal excretion
of the pan protein, the intact protein itself, is insignificant. Now, the breakdown of those protein drugs,
the proteolysis, typically occurs enzymatically, but can occur almost anywhere in the body,
including the blood. So, for some proteins, they have a very short
half-life in blood because they’re broken down in blood, which is typically not the
case with for very few exceptions in small molecular drugs. What that means is that for some of those,
especially peptides, their total clearance might be so rapid that it actually exceeds
cardiac output. And I’ve already eluded to the fact that they
can be degraded in the gut, which may be relevant in terms of elimination if they are secreted
in the GI tract as well. Now, here we have the first role that the
molecular weight plays, so proteins that are smaller than 30,000 daltons can be filled
up in the renal glomeruli. However — so they’re showing up in the primary
urine. However, after that — subsequently to that
they are subject to renal metabolism. So, what usually shows up in urine is not
the pandrug, but some of the metabolites unless they’re reabsorbed. On the other hand, if you look at the liver,
the role that the liver might play, the only way that those proteins can get into the liver
is by receptor mediated endocytosis or RME. Okay. So, they have to be taken out by particular
transporters, we’re going to look at some in a minute. Before they can metabolize in the lysosomes
by the endopeptidases, as I said before, that are responsible for breaking down endogenous
or dietary proteins as well. So, right here there is a table that I took
from the book chapters that I make reference to that points out the differences. So, you can see the typical small molecules
drugs, 500 to 1000 daltons, elimination side liver, okay. You’re talking about passive diffusion, as
I mentioned before, for some of the bigger ones that may be transporters involved. But the main determinant, physical chemical,
is their lipophilicity and molecular structure. The molecular structure would tell you what
transporters and what enzymes. Lipophilicity is more responsible for crossing
biological membranes. The moment you go above 1000, so let’s look
at small proteins and peptides, the main eliminating organ here would be the kidney. As I mentioned before by microfiltration,
however there’s subsequent renal degradation or metabolism. So, you wouldn’t see any pandrug in urine,
as is the case with the small molecular drugs. So, the molecular weight now becomes a major
determinant. For the larger proteins, so you’re looking
at 50,000 to 200,000, which would include monochrome antibodies, you have this receptor
mediated endocytosis. Not only in the liver but in other organs,
as we’re going to see in a minute as well, and they’re depending very much on the charge
and the sugar entity of those glycol proteins. Because some of those transports target manos
or fucos [spelled phonetically]. The even larger proteins, they actually have
to be opsonized in order to be absorbed. And if they’re really big, this is larger
than 40,000, they need to be taken up into the cell by phagocytosis and particle aggregation
would be the major determinant on how much of this happens. So, let’s look at the renal metabolism so
you can see this is a chlemarolis [spelled phonetically], blood coming in and then blood
flowing along the tubule, so this would be the renal proximal tubule, so the small proteins
up to about 30 to 50,000 get filtered, but then they can be broken down in the primary
urine itself. They can be broken down in the prash [spelled
phonetically] border, or they can be taken up in the tubule cells and then metabolize
or broken down in the lysosomes. Either way, there wouldn’t be any parent protein
that comes out in urine. There’s also a possibility for some of them
at least that the proteins actually get taken up from the peri tubule vasculature into the
tubal cells and being broken down as well. Either way, the only thing that would show
up in the terminal and the ultimate secondary urine would be metabolites not pandrug. And down here we have hepatic uptake, and
it’s not just hepatic uptake, before I eluded to that, it’s also uptake into other tissues,
whether we have endothelia cells, coupled cells. And you can see without going into great detail,
you can see they all involved RME, receptor mediator endocytosis. And you can see those receptors, they basically
target carbohydrates, okay? Manos, fucos and what have you. And the various examples of drugs that are
known to be subject to those uptake systems. So, this is very different from a small molecular
drug. Now, in addition to that, for some proteins,
they actually get broken — and I’m going to show you examples for that — they actually
get metabolized by virtue of interacting with their target. So, they interact with high affinity and low
capacity to surface receptors, and then by virtue of doing that, they’re exerting their
effect, but they’re also being taken up in the cells and then become subject to endocytosis
and lysosome degradation. This is called Target Mediator Drug Disposition,
and we’re going to use monochrome antibody as an example that follows a similar pathway. Now, speaking about immunoglobulin as I point
out here, albumin conjugates and immunoglobulin G’s which are basically monochrome antibodies
or monochrome antibodies are made from the derivatives of IgG. They interact with a very particular receptor
called FCRN or the neo-natal FC receptor. And it’s named neo-natal is because the first
time it was discovered, it was found to be responsible for transferring IgG across the
placenta from the maternal to the fetal blood. It’s also present, for example, outside the
placenta, which obviously is the reason why it plays a role in the proximal small intestine,
and it involves transcytosis. The FCRN as it relates to IgG drugs is present
throughout the entire body, so endothelial cells, monocytes, macrophage, dendritic cells,
all kinds of immune cells, and its road is to protect the IgG, the endogenous as well
as the exogenous drugs, from lysosome degradation. And it provides a salvage pathway and it recycles. So, this is the main reason why the half-life
of IgGs is as long as it is. And we’re going to discuss that in more detail. For right now, let’s just see what the role
plays, the FCRN mediated recycling plays for IgG and albumin. So, this scheme here compares IgG and albumin
relative to their daily formation, so 7.4 percent of the circulating amount gets formed
every day and at steady state 7.4 percent gets removed. For albumin, 11 percent gets formed, 11 percent
gets removed at steady state. But how do we get to this removal or katabatic
route? Well, you can see that if you look at IgG,
I’m sorry — 18 percent would be removed every day, if it were not for about 11 percent being
recycled. So, the net results is that only 7.4 percent
is removed. On the other hand, if you look at albumin,
the difference between the net removal and the actual katabatic efficiency is much less
because the recycling, the FCRN mediated recycling contribution is less. So, you can see that a significant portion
of catabolism for IgGs is protected by recycling the IgG, which as I said before, leads to
a relatively long half-life. Now, let’s look at a couple of just schematically
— a couple of elimination pathways for protein drugs. Three different kinds that I want to show
here, the first one is related to the presence of ADAs, or anti-drug antibodies. So, this is a phenomenon that is particular
to a large molecule protein drug where the body develops its own antibodies towards its
own. So, this would be the body create the antibody
and the antibody is then catalyzed, so you have what we call a turn over model. Now, this antibody, however, has the ability
to react with a protein drug and facilitate its degradation. Okay. So, one elimination would be the unintended
formation, typically unintended, of ADAs that facilitate elimination of protein drugs. The second scheme here looks at the protein
drug that is again different from a small molecular drug, that is eliminated not only
from what we call the central compartment. So, this would be the plasma compartment where
the drug is being given to and it’s being removed from. But it can also, this is the protein drug
now, can also go through tissues and being removed from those peripheral tissue compartments. Okay, so there’s a two-step elimination that
is different most of the time from small molecular drugs. And then, I already used the term Target Mediated
Drug Disposition, which is very relevant for monochrome antibodies, so this is a little
more complicated. So, R would be the target or the receptor
that the monochrome antibodies intended to interact with. It’s being formed continuously, and it’s being
eliminated continuously, so we are at steady state. Now, this receptor that interacts, you can
see with an equilibrium reaction with a protein drug, and you now have a second elimination
pathway for the receptor. So, the receptor is endogenously removed,
but it’s also removed after interaction with monochrome antibody. And obviously, it is that effect that is intended
by giving the protein, the drug, meaning the monochrome antibody, to hit the receptor. Now, the reason why that’s important those
processes can be saturable, meaning those dependent and they can be time-dependent. We can have non-linear PK, which is for small
molecular drugs the exception. Example that I’m showing here, you’re looking
at the plasma concentration on a logarithmic scale versus time and a macrophage colony
stimulating factors, MCSF, is given by IV at three different concentrations. So, you can see this is the low concentration,
the intermediate concentration, and the high concentration. And what you see here is what we like to call
in PK lingo “The Hockey Stick,” which his typical for a non-linear saturable drug. So, initially there is a distribution, there’s
a plateau phase and this is when the concentrations drop below the saturation levels, and we have
a relatively rapid elimination. So, you can see for all those three doses
the most rapid elimination occurs at low concentrations that are below the saturation level and any
higher concentrations lead to this plateau. Now, in addition to this propensity to be
associated with non-linear PK, at least for some of the protein drugs, they are PK/PD
relationship. The relationships between the concentrations
in blood and the target effects can be very, shall we say, indirect. Meaning there are multiple steps between achieving
drug levels and actually observing the affects. So, for example — as I point out here, for
growth stimulating factors, which would include erythropoietin the example I used later on
and the one I just talked about, the MGCFS — the PK/PD typically follows what I call
the indirect effect model, meaning there’s a pronounced lag time between plasma concentrations
and pharmacological effects based on their mechanism of action. So, let me just illustrate that by using erythropoietin,
which, as you know, stimulates the formation of red blood cells. So, what you have here in blue would be the
pharmacokinetic part of the model. The erythropoietin is given to a central compartment,
and it’s being eliminated, and this would be a saturable elimination path. We characterize by Vmax [spelled phonetically],
the maximum elimination rate NKM, which would be the saturation level or the affinity constant
that I eluded to early on. So, this is the kinetic model. Now, how does erythropoietin lead to increased
red blood cell levels in blood? What it does that in the bone marrow by stimulating
precursor cells. So, P1 and P2 are precursor cells all are
the ventricular sides, which are the immediate precursors to red blood cells. So, each of those precursor cells leads to,
by maturation, leads to next level. They all have a finite life span, which is
what the “tall” relates to. Okay, so you can see there is an effect that
slowly occurs because each of those stimulations has to translate into increased concentrations
of the subsequent red blood cell or precursor. In addition to that, as I eluded here in the
scheme, there is a negative feedback system that if the particular sides go up, meaning
a lot of red blood cells are going to be formed down the road, that actually downgrades some
of the precursors. Okay, so what that means is that to get from
the drug level to the effect, there’s a whole chain of events that have to occur, that take
time. And we have nonlinearities built in. And you can see in this particular example,
the stimulation relationships, that’s what the ST stands for, both follow what we call
e-max relationship, so they can be saturable depending on how the drug concentration compares
to the effective concentration, the EC-50 value. To illustrate that, to get away from the scheme
and actually look at data, on the top graph — so, you’re looking at multiple doses of
erythropoietin in humans, and you’re going to look at particular sides on the top and
red blood cells on the bottom and hemoglobin at the very bottom. Okay, and you can see that with each administration
of erythropoietin, the ventricular sides go up and they peak after about 200 hours. And then, they’re being basically maintained
until the last dose is given and they drop off. On the other hand, if you look at the red
blood cells, nothing really changes until maybe 200 hours. So, while those particular stats go up we
don’t see the red blood cells yet because they haven’t matured, or the particular stats
haven’t matured yet. Then they mature, and the red blood cells
continue to go up even as the ventricular sides decline, so it is the life span of those
red blood cells, and the maturation of reticulocytes into red blood cells, that makes the ultimate
response, which is the increased red blood cells or increase in hemoglobin much, much
slower than the primary response, which would be the increased reticulocytes. Similar to that, another growth stimulating
factors, and I’ll be looking a filgrastim. Again, that’s going to be my example at work
in a few minutes. G-CSF. In blue, we have the PK model. So, here we have the drug being given into
a one-compartment body model. The main difference to EPO we now have two
routes of elimination; what’s called a linear route and a neutrophil route. So, this drug is metabolized via binding to
this neutrophil receptors, and I’m going to discuss later on, this is a searchable process. The linear clearance is elimination by the
kidney, because this is a relatively small molecule. Now, how does that relate into pharmacological
effects? Again, we are looking in the bone marrow at
precursor cells to the white blood cells, or the absolute neutrophil count that we can
measure in the peripheral blood. And the stimulation to the precursor cells
takes time, depending on — you can see some of their life spans, 40 hours, 60 hours, 95
hours. So, that’s how long it takes after you stimulate
those precursor cells before they show up in plasma as — or in peripheral blood as
neutrophil count. And here is the corresponding plots. So, on the left-hand side, you’re looking
at the serum or plasma concentration versus time, single doses after subcutaneous injection,
you can see the concentrations increase with higher doses. Their terminal slope seems to be very similar. Okay. On the other hand, if you look at the pharmacological
effects, so this is the absolute neutrophil count versus time. You can see with increasing doses, the peak
effect gets later and later. So, at the highest dose it takes six days
for the neutrophil count to achieve its maximum, even though the peak plasma concentrations
are achieved within two days. What’s the reason? Well, the reason is that the effect has to
work its way through the various progenitor cells, okay? So, this is an indirect effect model, and
it’s the cytokinetic model, meaning the time, the lifespan and the various transfer constants
between those precursor cells drive the time profile in the pharmacological response. Okay, something unique but unrelated to the
PK/PD of protein drugs is their ability, as I mentioned before, to cause immunogenicity. So, they can cause immune responses, anything
from anaphylactic shock to neutralizing antibodies. And I’ve listed here — excuse me. I’ve listed here a few potential consequences. So, you can have loss of efficacy. For example, if you have patients with hemophilia,
and they develop antibodies to Factor X — Factor VIII, they are basically unable, even with
high doses, to use Factor VIII as a way to replace the missing clotting factor. Okay. And you can see some of the monoclonal antibodies. We will talk about that in the exercise. On the other hand, sometimes those immune
responses, meaning those antidrug antibodies, can actually mimic the drug itself. So, growth hormone, the antibodies themselves
also have as a pharmacological response similar to what growth hormones does. They can neutralize endogenous proteins. This is the story about EPO a while ago, when
they changed the formulation of EPO, I think they changed a rubber of the top of the glass
vial that had EPO. It changed the structure of EPO, and EPO became
now immunogenic, and patients that received that erythropoietin actually developed antibody
to erythropoietin that targeted not only the exogenous but also their own erythropoietin,
so they were completely unable to produce red blood cells, because their endogenous
protein was neutralized by those ADA’s. And then I already mentioned allergy anaphylaxis
and serum sickness, if you have lots of circulating ADA complexes with the protein drugs. This is just from a kinetic point of view. The formation of ADA’s can actually typically
increase clearance because you’re now removing protein drug from the circulation, so they
can actually reduce the activity. In rare cases, they can also increase — I’m
sorry, decrease the clearance and increase the activity. But that’s the exception. The main concern is that you have ADA’s that
— the clearance of those protein drugs is enhanced, and the activity is reduced. Where does the immunogenicity come from? Well, before we talked about potential differences
in sequence, and especially when we use non-human amino acid sequences as we do for some of
the monoclonal antibodies. The glycosylation pattern, I think this is
something that is currently investigated very intensely to find out what carbohydrate change
in what position might induce immunogenicity, because if you understand it, you can actually
avoid it. And then, the other thing to keep in mind
is that the proteins themselves can interact and form triominos, tetrominoes, what have
you, and there are impurities and contaminates, as a result of the biological synthesis formulation
technology. All this goes into the potential risk to cause
immunogenicity. For the purpose of this lecture, we are going
to talk about the glycosylation pattern and the sequence variation a little bit later
on. Okay, now, let me switch to then a particular
kind of protein drugs that really have a major — have had a major impact on therapeutics,
and those are monoclonal antibodies. The reason being is they are so highly selective
and have such a high affinity for their intended target that they’re about as close to the
magic bullet that Paul Ehrlich stipulated in the early part of the 20th century. So, they’re about as close to just hitting
the target, rather than having off-target effects. Even as you’re going to see in a minute, they
all have potential for non-target related risks. So, what’s the basic structure of a monoclonal
antibody? As I mentioned before, they are basically
most of — at least the ones that are currently approved are IgG drugs. So, if you look at the molecule here, it looks
like — the only multimedia show for today — they look like a VCU cheerleader or like
a Y. So, you have the FC part of the molecule down
here, and you have the FAB part of the molecule on the two ends of the Y. The top part here, that’s where the antigen
binding region is, composed of a light and a heavy chain. So, this is what makes the specificity of
each of those monoclones. You can see they’re connected, the two chains,
and the various sections of the protein are connected by disulfide bonds. And as I mentioned before, the FC segment,
the stem, if you like, that’s responsible for the effector function, whereas the FAB
section is responsible for the antigen. I would also point out, and it’s highlighted
here in this green circle, that there are lots of carbohydrates in those monoclonal
antibodies. So, there are large molecules and there are
glycoproteins. This is to give you some idea how they compare
across the various classes, so I’m only going to focus in on IgG, because as far as I’m
aware, those are the only ones that are therapeutically available. But I want you to realize that endogenously,
we have IgA. Those are typically on epithelial surfaces. IgM, those are the macroglobulins. IgD and IgE, those are related to histamine
release on mass cells. But the monoclonal antibodies are typically
from the IgG class. So, let’s see what they have in common. You can see their molecular weight, as I said
before, relatively high, 150- to 160,000. You can see their half-life is to the tune
of three to four weeks, with the exception of Ig3. Okay. Now, what do they do in terms of the effector
function? They all activate to some extent the complement
pathway I’m going to talk in a minute about. They do not activate the alternate — alternative
complement pathway. They are not present on mature B cells. Probably the most important thing to keep
in mind is their ability to bind or not to bind to the FC receptors. Remember, those are the receptors that contribute
to the recycling and the prolonged half-life. Okay. So, IgG1 is the most commonly used backbone
for drugs that you want to target and knock out, because they allow you to bind to the
target and induce a factor to remove it. Now, in terms of the convention to be used
to name them, again, this is different from small molecular weight drugs that all have
their pretty generic name. We have murine antibodies. They would be MoMap. I’m not going to talk about them because they
all have an unacceptable risk of causing immune responses. However, we do have chimeric antibodies. So, you can see those are antibodies where
the FAB, the antigen binding part of the molecule is murine, meaning based on mouse sequence. The rest of the molecule is human. We have humanized antibodies, the zoo [spelled
phonetically] maps, that have basically small sequences of murine, but they are — otherwise,
they are fully humanized. And then we have the gold standard nowadays,
which would be a fully human antibodies, where both the FAB and FC sequence are all human. So, you can see at adalimumab, which is one
of the TNF antagonists, is human antibody. And you can see that infliximab, which is
another TNF, that’s a chimeric antibody. Now, what are the Factor processes? In other words, once those monoclonal antibodies
hit their target, how do they remove it? Either by complement-dependent cytotoxicity
or by antibody-dependent cellular toxicity, CDCC or ADCC. Let’s look at CDC first. So, here you can see that’s the monoclonal
antibody, the little Y that binds to a target. In this case, it would be a surface target. This would be a CD20, which is a target for
rituximab. By binding the FC receptors, so the bottom
part, if you like, the stem of the molecule activates complement. Complement proteins gather, and they form
what’s called a membrane attack complex — a membrane. I’m sorry, a protein complex that basically
opens up a pore and through that pore, basically, you have the extracellular and the intracellular
fluid being in contact, and the cell is going to be lysed. Okay. So, the binding of the monoclonal antibody
to the target leads to FC-mediated complement activation, which kills the cell. That would be CDC complement. On the other hand, if you look at the ADCC,
you now have again a target that the monoclonal antibody has bound to, and now the FC, the
bottom part of that molecule, interacts with FC gamma receptors that are present on other
immune cells. In this case, it would be macrophage or monocyte,
or natural killer cell. Okay. So, here the monoclonal antibody hits the
target and then it activates other immune cells to destroy the target and the cell associated
with it. So, this requires complement. This requires other immune cells. The last thing that I want to review with
you in more detail is this FcRn-mediated recycling. So, regardless of what target IgG or monoclonal
antibodies have, they are subject to FcRn or the neonatal FcR receptor recycling. So, how does that work? Well, this would be the IgG. It could be the endogenous IgG, or it could
be the monoclonal antibody. It gets — it binds to the FcRn on the cell
surface. First, it forms an endosome, and the endosome
gets internalized. Okay, so you now have inside the endosome,
you have the IgG, both free and the IgG that is bound to the FcRn. The IgG that is free is then metabolized,
broken down in the lysosome. So, that’s gone. On the other hand, the FcRn-bound monoclonal
antibody is protected from this degradation, and it gets recycled to the cell surface where
the antigen — I mean, the antibody is then released again. So, you have recycling of a portion and you
have breakdown of some of it. And remember, this is not related to the FAB. It’s not related to the antibody targeting
part of the molecule. This is all related to the FC portion of it. So, if you put this together, monoclonal antibodies
are subject to two parallel elimination pathways. The first one that I’ve already alluded to
is TMDD or target mediated drug disposition. So, here the monoclonal antibody, by virtue
of its FAB forms an equilibrium with a drug target, it tags it, and then you have subsequent
elimination via FC-mediated processes. So, this could be CDC or ADCC. This is a pathway that is highly selective. It only occurs if the drug binds to the target,
and it depends on how much target and how much drug there is. In other words, it is searchable and depends
on the concentration of both of the contributors to this complex. But in addition to that, any monoclonal antibodies
are also subject to this relatively nonselective FcRn-mediated. Okay. So, here with the Fc part of the molecule,
you have endocytosis and subsequent degradation. However, any of the monoclonal that is bound
to the FcRn is protected and gets recycled. So, this is a pathway that does not depend
on target concentration because the target is not involved. And it is usually not searchable. However, something that I didn’t point out
here that I want to point out verbally is that this process happens not only for monoclonals,
but also for endogenous IgG. So, you have endogenous IgG, the patient’s
IgG compete with — potentially at least, compete with monoclonal antibodies for those
binding sites which then could prevent them from being recycled. The reason why this is important from a kinetic
point of view, the combination of those pathways, along with the fact that the target itself
changes over time, means that there is a potential for dose and time dependent pharmacokinetics. So, let’s look at an example where we work
out those various relationships. So, this is an antibody that targets a surface
receptor CD11a on T cells. Okay. So, you have the free antibody that interacts
with its target. It forms a complex, and this is a reversible
interaction. However, this complex itself is now subject
to this, as I mentioned before, TMVD. So, specific searchable receptor mediated
clearance. And you can see there are various FC receptors
that lead to the removal of the target that is bound to the monoclonal. In addition to that, the monoclonal by itself
can also be removed by these non-searchable FcRn pathways. Okay. So, this is non-searchable. This is searchable and depends on how much
target there is. So, if you look at again the concentration
of plasma versus time used for the different doses from 1.2 to 10 mg, so this is a 100-fold
range in doses, you see again what I like to call the hockey stick, that with higher
doses you get a plateau, but the terminal rapid elimination is the same across doses,
because now the concentrations are below the saturation level. What is — and I didn’t put the numbers here,
but if you look at the clearance values across those doses, you can see that from the lowest
to the highest dose, the clearance is reduced 60-fold. So, this is strongly nonlinear. You also see an increase in — I’m sorry,
decrease in volume of distribution only about twofold. That means with increasing doses, you have
saturated now the binding sites, not only the removal sites. Okay. So, this is typical for TMDD that both clearance
and volume go down. Now, if you look at the PK/PD relationship
— so, this is now not after IV but after subcutaneous administration of one of those
doses. And we are looking at the left-hand side at
the plasma concentration, and here we are looking at the concentration of the target
on the right-hand side versus time. So, you can see after subcutaneous administration,
the monoclonal antibody concentrations in plasma go up, they peak, and they decline. Consistent with that, you can see that the
target concentrations go down as well. Okay. So, the drug hits the target, that binding
almost instantaneously, and the target gets removed. And then as the monoclonal antibody concentrations
decline, meaning as that monoclonal antibody complex is being removed, you have recovery
of the target. So, the target comes back because the removal
now gets less and less, and the target gets regenerated. Okay. And you can see that is true whether you look
at the expression or you’re looking at the binding sites. Okay. So, here you can see that the target concentrations
have a direct — or the target levels, I should say, have a direct correlation with the concentration
of drug. But they also depend on the turnover of the
target. Underneath, you have now repeat those studies
in patients with psoriasis. The idea being is that by hitting the target,
by knocking out those T cells that the psoriasis and autoimmune disease should be benefitting
from that. So, what you’re looking at here is again over
time. Now, look at the different time scale because
you’re now looking at repeat doses once a week. On the left-hand side, we are looking at the
drug concentrations. So, you can see weekly doses, the drug concentrations,
they go up and then we are at steady state, and at the last dose, the drug concentrations
decline. In yellow, we are looking now at the target,
and the target is on the right-hand scale. And you can see that with the first dose,
the target is knocked out and it remains knocked out basically until the last dose is given
and then the target gets regenerated. It comes back. Okay. And you can see that the target and the three
binding sites, just like in this — well, here, they basically parallel. If you now relate that to the clinical effect
— that’s what the clinicians care about, they have what’s called a PASI score, psoriasis
area and severity index. And that’s in purple. And you can see that that score, even though
the target gets knocked out right away, that score changes very slowly. And even after the last dose, that score only
slowly returns back. So, there’s a further delay from translating
the knockdown of the target, the CD11a to relating that to the clinical symptom, which
tells us downstream of this target, there is pathophysiology that basically explains
for this lag as well. Okay. Let me go through one example. And the second one is, as I said, an exercise. So, the first example is a growth stimulating
factor, filgrastim (G-CSF). And I want to start off by showing you how
different this protein is from a monoclonal. Okay, you can see it’s much smaller. Okay. It’s about 20,000 daltons. Monoclonal is about 150,000. It is non-glycosylated, so we don’t have to
worry about various carbohydrates. It is a single chain protein as opposed to
the two light and the two heavy chains that are monoclonal. Okay. So, filgrastim should be much easier to deal
with in terms of the manufacturing than a monoclonal. What does it do? Well, as I mentioned before, it binds to a
specific receptor on the surface of white blood cells and their precursors. And by binding to the cell surface receptor,
G-CSF receptor, that gets translated in cell proliferation. So, it leads to increased maturation of white
blood cells. However, and the reason why I use it as an
example, it has very complex kinetic properties, because it is subject to nonsearchable renal
elimination, because it’s a relatively small protein. And then it has a second pathway, and that’s
the one that causes trouble. Because that pathway depends on how much receptor
there is. So, this binding to the receptor not only
leads to the intended pharmacological response, increased white blood cells, but it also leads
to the breakdown of the drug in the first place. So, it’s limited. This elimination is limited by how many white
blood cells there are, and therefore, it becomes searchable and time dependent. So, let’s look at a couple of plots to illustrate
that, and they are part of one of the handouts that I’ve included with the class material. So, let’s look at after IV administration. So, here filgrastim is given intravenously
as an infusion, I think over 30 minutes. And you can see after a single dose, it looks
— doesn’t look like hockey sticks; it looks like it might follow linear PK. If you look underneath, the pharmacological
response, the absolute neutrophil count versus dose — I’m sorry, versus time — you can
see that the absolute neutrophil count goes up, peaks, and then it goes down. The main thing, as I alluded to before, the
plasma concentrations peak right at the end of the infusion, after 30 minutes. However, the white blood cells don’t peak
until after about 20 hours. So, again we have this disconnect, this lag
time that is caused by the maturation, the cytokinetic model. On the right-hand side, we are now changing
the administration route and we are giving filgrastim subcutaneously, so we now can see
concentrations in plasma go up, they peak and then they go down, maybe with a longer
half-life. So, maybe there’s a little flip flop. Underneath, we can see, again, the absolute
neutrophil count goes up and it peaks, and you can see the change here is less than it
is after IV, because we don’t have complete bioavailability. You can see after subcutaneous administration,
the peak level is about 10. After IV, they are more than 100. More importantly, when we look at repeat dose
administration — so, now we’re giving the drug subcutaneously repeatedly at different
doses. So, at the very top, we are looking at the
plasma concentration of the filgrastim versus time. So, this is the first dose, second dose and
so on. Let’s just compare for the low dose — that’s
2.5 mcg/kg. The first dose and the last dose. And you can see that at the first dose, the
drug levels are higher than after the last dose. Let’s look at the higher dose, 5 mcg/kg. The first dose has higher doses than the last
dose. And the same is true for the highest dose. So, over time, the drug concentrations behave
very different from what you would expect. There is no accumulation. There is the opposite. There is a loss of concentrations over time. All right. Let’s look underneath, what happens to the
pharmacological response. So, here you’re looking at the absolute neutrophil
count versus time, and you can see the absolute neutrophil count goes up, goes up, and the
highest level are achieved as you would have expected after the last dose. Okay. So, you can note that at the first dose when
we have the highest drug level, we also have the lowest white blood cell count. At the last dose, when we have reduced drug
levels, we also have the highest white blood cell count. And obviously, this is the explanation. As I mentioned before, the elimination, at
least part of it, a major part of it, is — of filgrastim, it depends on how many neutrophils
we have. So, as we have more neutrophils, the clearance
is increased, and we have lowered or reduced levels. Okay. So, when you look over time, so this is based
on a model that they used in this particular paper, if you look at the amount of receptor
that is available for the drug to bind to exert its effect but also to be eliminated,
that receptor goes up. Or, if you look at the better way to look
at that — if you look at the clearance, the instantaneous clearance over time, you can
see within each dose, the clearance, because it’s a searchable clearance, goes down. And then, it recovers prior to the next dose. Okay. But in addition to that, over time, because
you now have more white blood cells, you also have an increase in residual clearance. So, at the last dose, the clearance is much
higher than after the first dose. Now, so that’s complex — time and consideration
dependent. Now, one of the advantages — one of the disadvantages
of filgrastim is it has to be given fairly frequently. So, attempts were made to improve that, and
one way to do that is by PEGylating it. So, this is a [unintelligible] binding of
filgrastim to polyethylene glycol that increased the molecular weight to above 30,000, which
basically shuts down the renal excretion. And if you go back, we are shutting down this
pathway, and the PEGylated filgrastim now has only one way to leave the body. And that is why this receptor-mediated degradation. So, again, let’s look at a study. So, this is a study, and top we’re looking
at the concentration of the PEGylated filgrastim, and underneath we are looking at the neutrophil
count. And I’m going to explain to you in a minute
what happens to those neutrophils. So, we can see first, that the concentrations
increase dose-dependently and stay around longer than after the non-PEGylated. So, the non-PEGylated are the triangles, and
you can see all those half-lives or terminal decline is more slowly. So, we can give it less frequently. That’s obviously the intent. Now, if you look at the ANC, we now have to
appreciate that those studies were done in patients that underwent chemotherapy. So, they got a test dose, if you like, or
a pre-chemotherapy dose. And they saw a time-dependent increase in
neutrophil count. On day 15, they received their chemotherapy,
which was highly myelosuppressive. So, you would expect the white blood cells
to go down. A day later, they got their second dose of
filgrastim, and you can see that even though there was a drop in white blood cells because
of the chemotherapy, that drop was actually less with the PEGylated than it was with the
non-PEGylated, okay, which again suggests that the difference in half-life — that this
non-PEGylate has such a short half-life that it does not cover all the post-chemotherapy
myelosuppression. If you look at the clearance values — again
to illustrate how non-linear the kinetics has become — if you look at those, those
were the patients that I just reviewed for you with non-small cell lung cancer before
chemotherapy. And let’s just look at the clearance. With increasing doses, the clearance declines
from 64 to seven. So, this is a sign of the concentration or
dose-dependent kinetics. We are saturating the pathways via the white
blood cells. After chemotherapy — so this is when the
chemotherapy now has reduced the white blood cells — the clearance across all doses is
further reduced, because we basically have reduced their elimination pathway. We still have dose-dependence. And what I’ve done here is I’ve actually plotted
the area under the curve. This is for the drug in plasma versus dose. And I’ve plotted the area under the effect
curve, which is the ANC count versus dose as well. So, let’s look at the kinetics in blue. And you can see that this is not a straight
line, but it is a super-proportional relationship, meaning with increasing doses, the area increases
more than proportion to dose. This is true for the kinetics, and the reason
for that is that we have saturation. However, if we do the same plot for the pharmacodynamic
response, you can see that the pharmacodynamic response is actually infra-proportional. So, increasing the dose, we get a less than
proportion increase in the area of the ANC count. That has to do with the fact that those doses
are already saturating the maximum possible response. So, for the purposes of dose selection — as
the paper that I have included with my handouts points out — they looked at the relationship
based on a model between the maximum effect, okay, as a function of the concentration of
filgrastim. They defined it — I’m sorry — by an EC50
value of about eight nanograms per mil, and you can see the various doses. The top two doses basically fall into the
plateau portion, and you can see those original doses in the studies that I just reviewed
for you were body weight corrected. And they actually then picked a non-body weight
corrected dose that allowed them across a fairly large range of body weights to fall
into the plateau. All right. So, filgrastim has a reduced total clearance
relative to filgrastim because we’re taking out the renal elimination pathway, and it
depends exclusively on the neutrophil-mediated elimination pathway. As a result, the pharmacokinetics becomes
strongly non-linear and saturable after single doses. However, as far as the dynamics is concerned,
we can select doses that fall into the plateau. And as a result, we get those effects that
we’re talking about here in preventing the chemotherapy-induced myelosuppression. The TNF blockers, we’ll talk about that after
the end of my lecture. So, let me now wrap up my presentation by
talking about biosimilars. So, biosimilars are the equivalent in the
biotech world of generic drugs. The idea basically is if we can classify a
biosimilar to a reference protein drug, they are therapeutically substitutable. Okay. Another term that you find in the literature
further that are follow-on biologics. Now, just to give you a reference — no pun
intended. Generic drug products for small molecular
drugs basically depend on a comparison of the area under the curve in bioequivalence
assessment, okay, in order to conclude that they are therapeutically substitutable, unless
they are locally-acting drugs. However, for biosimilars, because of the complexity
of the molecule itself, that’s not sufficient. So, the approach is different. And you can see this is the pyramid that you
find lots of times that’s supposed to help understand how we conclude that biosimilar
is biosimilar or highly similar to the reference product. Most of this pyramid refers to analytical
tools that allow us to characterize the molecular structure, the physical-chemical properties,
the in vitro properties, in terms of the FC and the FE part of monoclonal antibodies,
for example. So, this is all based on analytical work. Then there are nonclinical studies to show
that the PK and if possible the PD in animals are comparable. What I’m going to focus are the top two parts
of the pyramid. We want to show, also in order for a biosimilar
to demonstrate biosimilarity, the reference and the potential biosimilar have to show
that they are bioequivalent kinetically, bioequivalent dynamically, and that they are clinically
not distinguishable. So, as we move up, we are reducing uncertainty. Uncertainty meaning we are convincing ourselves
more and more that those two products are truly therapeutically substitutable. And as I point out here, PK/PD including immunogenicity
studies, they are intended — so, that’s the top part here — they are intended to show
that any differences between the two products are not clinically significant between the
proposed biosimilar and the reference product. And that’s all part of the totality of evidence,
which would be the entire pyramid. All right. Now, for the PK/PD part, typically we have
clinical pharmacology studies in healthy volunteers. So, those are not patients but study subjects
in order to demonstrate PK and PD, if possible, bioequivalence. In addition to that, this is required in the
United States per guidance. In a phase three study with patients, we have
to demonstrate, or it has to be demonstrated, that there is no clinical difference between
outcomes, or for outcomes, between test and reference products. And, as part of those studies, there’s usually
a sub-study where the PK in patients will be assessed. And lastly, clinical immunogenicity should
be comparable. So, what I’ve done on the next few pages,
I have summarized the four biosimilars that as of right now, are approved in the U.S. And you can see one of them is a GCFF that
we’ve already talked about, and three of them are monoclonals. Okay. So, we have filgrastim. We just talked about that. There is one biosimilar approved. And then, we have three TNF blockers that
are part of the exercise that also have been approved. And you can see how were they studied? Well, they all underwent pharmacokinetic study
in healthy volunteers, either as a crossover or parallel group. So, there was washout involved for the crossover
studies to make sure that the two treatments didn’t carry over. You can see that in all across the board,
both U.S.- and EU-sourced reference product was used. So, those are all multinational companies. They wanted to target approval in both the
EU and the United States. You can see fairly large sample sizes to demonstrate
bioequivalence. On the page, I have listed according to the
label whether the drugs follow linear or non-linear PK. And as we talked about for filgrastim, it
is subject to a strongly super-proportional PK. All the other — at least two of them are
known to be linear in the therapeutic range. The third one is not stated in the label. As a result, you can see that for the filgrastim
biosimilar, different doses and different routes were studied to deal with the fact
that there are super-proportional PK. On the other hand, if you look at the monoclonals,
pretty much straight, single doses and a fixed dose, a therapeutic dose, was studied. All those studies used bioequivalents. So, the area in the Cmax/AUC ratios between
biosimilar and reference had to be within 80 to 125 percent. In addition to that — and I think this is
kind of tough to see — for all of those four biosimilars that are approved, phase three
studies were done in patients. The dosing regimen that was used was obviously
a therapeutic dosing regimen, and in sub-groups of patients, at least informally, the pharmacokinetics
was assessed using either areas under the curve during the dosing interval or trough
levels. Other than the ANC count for filgrastim for
any of the TNF blockers, no PD metrics was assessed. As far as immunogenicity is concerned, you
can see that of the four approved, the immunogenicity rates were “similar,” meaning in the eyes
of the FDA and the Swanson [spelled phonetically] are different between the biosimilar and the
reference for one of the two — for one of the four, a post-marketing study actually
was required because there were differences that could have been clinically significant. So, let me walk you through an example. And I picked the Zarxio, which was the first
U.S.-approved biosimilar. I picked that as my example. Okay. So, the biosimilar needed to demonstrate biosimilarity
for the U.S.-approved reference product, Neupogen, included PK/PD studies. So, let’s walk through them. One of the phase one studies — one of the
PK bioequivalent studies that I alluded to in my summary sheet, now looks at the concentrations
of filgrastim versus time for this Zarxio which is now the biosimilar, and the reference
product the Neupogen. So, here you can see that the profiles seem
to overlap nicely. The formal bioequivalence assessment tells
us that both the area under the curve and the Cmax/AUC, their 90 percent confidence
level falls within 80 to 125 percent. So, they passed the pharmacokinetic bioequivalence. In addition to that, because we have a marker
— a pharmacodynamic marker — as part of this study, they also measured the pharmacodynamic
bioequivalence. And in order to do that, they — for the same
study now — they measured over time the ANC, the Absent Neutrophil Count, versus time. Again, you can see the profiles are virtually
superimposable. Using the bioequivalence criteria for both
the area under the effect curve and the peak effect fall within the bioequivalence [unintelligible]
check. Then, I mentioned before given the fact that
there is a relatively complicated dose dependence, they also looked at two and a half, five,
and 10 micrograms per kilogram sub-cu. Again, you can see without any formal testing
that the dose dependence for both the Zarxio biosimilar and the Neupogen reference are
basically superimposable. So, in all the clinical — the phase one PK/PD
bioequivalence was achieved. And then, to confirm that this translates
in patients in any or in no clinically-significant difference between the two, so they did a
study in patients with breast cancer that underwent chemotherapy. And they received two doses of filgrastim. So, just like we talked about before, this
is the first does of filgrastim. And somewhere here they received chemotherapy. You can see the white blood cell count declines,
and then they get the second dose. And you can see, again, if you compare the
Zarxio the biosimilar and the Neupogen the reference, the profiles are virtually superimposable. And the clinically-relevant outcome that was
used to assess this study was the duration of severe neutropenia. So, if you look at for the both treatments,
the mean and the 90 percent confidence interval, they basically overlap. And they could conclude that the two treatments
are bioequivalent. So, based on the totality of the evidence,
they were able to show that Zarxio and Neupogen were biosimilar, and Zarxio in 2005 became
the first biosimilar ever. Now, that concludes my formal presentation.

Leave a Reply

Your email address will not be published. Required fields are marked *