83 – Analytical Methods in Process Development (S6E8)

Okay, let's get started. You know that amazing
feeling that spark when a potential new drug
looks really promising in the lab? Oh yeah, that
aha moment. Exactly. But that's really just the
starting line, isn't it? The big challenge comes
next. And making sure you can actually make it
reliably, consistently, you know, batch after
batch when you scale things up. It's a whole
different ball game. It really is. And how we
manage that jump from like a tiny lab flask to
producing medicine for potentially millions of
people. Well, that's where things get interesting.
And that's what we're digging into today. The
analytical tools. The unsung heroes, you could
say. Definitely. They're like the the vigilant
eyes watching the whole process, making sure
the chemistry does what it's supposed to and
that every single pill or vial has the right
stuff in it. Right. So for our listeners, think
of this as your kind of shortcut to understanding
a really key part of drug manufacturing. We're
not going to drown in super technical jargon,
but we want to shed light on how these tools
give us the essential insights we need during
that whole process development stage. And we'll
be drawing on scientific literature, papers from
drug discovery development, including those practical
case studies you often find in journals like
OPRND, Organic Process Research and Development.
So our mission for this deep dive is pretty clear
then. Yeah, I think so We want to unpack how
these core analytical methods the key techniques
are actually used day in day out to monitor what's
happening and Importantly to troubleshoot when
things don't go quite right in developing these
pharmaceutical processes It's all about making
sure that initial aha Reliably becomes a safe
and effective medicine. Okay, so why all this
focus on monitoring? on consistency. I mean,
it seems kind of obvious for medicine, but let's
explore the why a bit more. Well, at its core,
consistent quality isn't just like a nice to
have in pharma. It's absolutely fundamental.
Patient safety depends on it. Right. And there
are strict regulations for very good reason demanding
it. Every single batch has to meet incredibly
tight specifications. It's non -negotiable. Yeah,
you can't have one batch strong in the next one
week or with some impurity. Thinking back to
one of our earlier deep dives, way back we talked
about scale up, right? Going from the lab bench
to giant industrial reactors. Exactly. And that's
where things get tricky. The actual physics can
change, how things mix, how heat gets transferred.
It's not the same in a 10 ,000 liter tank as
it is in a one liter flask. So you need a way
to actually see what's going on inside that big
tank to make sure the reaction is still behaving
as expected despite those changes? Precisely.
And how do we get that visibility? How do we
know reliably that the process is under control?
Well, that's where the data from analytical tools
comes in. They give us the hard evidence, the
numbers, to understand and manage that complexity.
OK. Makes sense. So let's meet some of these
analytical workhorses. First one, HPLC. High
Performance Liquid Chromatography. Sounds pretty
complex. What's the basic idea? Essentially,
HPLC is a technique for separating out all the
different components in a liquid mixture. It
works based on their chemical properties, how
they interact with the materials inside the HPLC
column. Think of it as a very, very sophisticated
filter, sorting molecules. Okay, like molecular
sorting. So how does separating molecules help
us watch a reaction happen? What does it actually
tell us? Well, what you do is take tiny samples
from your reaction mixture at different times,
say every hour. or even more frequently. You
run these samples through the HPLC and the output
basically shows you what's in that sample and
how much of each thing there is. So you can see
the starting materials decreasing over time.
As they get used up. Right. You can see any intermediate
molecules formed along the way appear and then
maybe disappear as they turn into the final product.
And crucially, you see the final product increasing.
So it's like a running tally of all the key players
in the reaction. Exactly. You get a quantitative
picture of the reaction's progress. And this
directly feeds into ensuring that uniformity
in the finished batch from batch to batch that
regulators like in the U .S. CFR Title 21 are
so focused on, you're proving the reaction is
consistent. Okay. That's clear for tracking the
reaction itself. What about making sure the final
drug is pure? No unwanted extras in there. Yeah.
HPLC is absolutely critical for purity testing,
too. Because it separates everything based on
chemistry, it's brilliant at spotting impurities.
Things like leftover starting materials, byproducts
from side reactions, or even degradation products,
if the molecule isn't stable. So tiny amounts
of things that shouldn't be there will show up
as separate signals? Precisely. They appear as
distinct peaks on the output, the chromatogram.
And we can quantify them, measure exactly how
much of that impurity is present. Method specificity,
making sure the method only measures what it's
supposed to and can distinguish the drug from
potential impurities, is a huge focus in regulatory
guidelines. And I guess that specificity is even
more critical when you have things like... Like
chiral molecules those mirror image one. Oh absolutely
vital Chiral molecules or enantiomers can have
identical chemical formulas But completely different
3d shapes like your left and right hands and
often only one hand has the desired medical effect,
right? The other might be inactive or even harmful
exactly So your HPLC method must be able to separate
those two mirror images and also separate them
from any other impurities that might be structurally
similar. It needs to be highly selective. It's
a real analytical challenge sometimes, but essential.
OK. HPLC separates and quantifies. Got it. Let's
shift gears to another major category, spectroscopy.
That's a pretty broad term, right? It is, yeah.
It covers techniques that look at how matter
interacts with electromagnetic radiation light,
essentially, across different wavelengths. A
really key example in process development is
infrared spectroscopy, or IR. Infrared, like
heat radiation. Sort of, yeah. In IR spectroscopy,
we shine infrared light through a sample. Molecules
absorb specific frequencies of this IR light,
causing their chemical bonds to vibrate, stretch,
bend, wiggle. Wiggle, OK. Yeah, well... vibrate.
And the pattern of which frequencies are absorbed
is unique to that molecule. It's like a molecular
fingerprint. Ah, okay. So unlike HPLC separating
things out, IR identifies a molecule by its unique
vibrational signature when hit with infrared
light. That's a great way to put it. Different
types of bonds like carbon -oxygen bonds or nitrogen
-hydrogen bonds vibrate at characteristic frequencies.
So by looking at the IR spectrum, the plot of
absorbance versus frequency, we can tell what
kinds of bonds or functional groups are present
in the molecule. So you can use it to confirm
you've actually made the thing you intended to
make. Yes, absolutely. You can confirm the formation
of the desired product by seeing its characteristic
peaks appear. You can also monitor the disappearance
of starting material peaks. And importantly,
you can sometimes spot unexpected changes that
might signal a problem, like the formation of
an unwanted structure. And this fingerprint idea...
How does that get used routinely? Well, because
the IR spectrum is so characteristic, it's a
fantastic tool for identity confirmation. You
use it to check incoming raw materials. Is this
really what the label says it is? You compare
its spectrum to a known standard. Same for intermediates
and the final drug. It's a quick and powerful
check. OK. Now I've also heard about X -ray diffraction,
XRD being important, especially for solid drugs
like in tablets. How does that fit in? Is it
related to spectroscopy? It's another way of
probing material structure, but using X -rays
instead of IR light. And it's specifically for
crystalline solids. It's less about the individual
bonds within a molecule and more about how the
molecules pack together in a crystal lattice.
Ah, the crystal structure. Why is that so important?
Because many drug substances can crystallize
in different forms. This is called polymorphous.
Think of it like carbon forming, both graphite
and diamond. Same element, different crystal
structure, very different property. For drugs,
different polymorphs can have different solubilities,
different stabilities. They might dissolve at
different rates in the body, or even process
differently during manufacturing, like when making
cablots. So if you don't control which crystal
form you're making, you could end up with a drug
that doesn't perform consistently. Exactly. One
batch might dissolve quickly, another slowly,
leading to different effects in the patient.
So XRD is used to confirm that you are consistently
producing the correct polymorph, the one with
the desired properties that was used in clinical
trials. It's another critical layer of control,
especially highlighted when you're transferring
technology or ensuring consistency after processes
like milling, which can sometimes induce changes
in crystal form. Okay, so we have HPLC separating
and quantifying, IR giving a molecular fingerprint,
and XRD checking the crystal structure. These
all sound great for monitoring when things are
running smoothly. But what about when they're
not? How do these tools turn into, well, detectives?
That's often where they really earn their keep,
honestly. When a process goes sideways, maybe
the yield drops or an impurity pops up or the
reaction just stalls, these techniques are indispensable
for figuring out why. Okay, let's take an example.
Say your reaction yield is suddenly much lower
than usual. How would you use, say, HPLC to investigate?
Right. So you take a sample of that final reaction
mixture and run a detailed HPLC analysis. What
are you looking for? Well, maybe you see a huge
peak for your starting material still remaining.
Meaning the reaction just didn't finish? Exactly.
Or perhaps you see very little starting material,
but also less product than expected, and a bunch
of new unexpected peaks. Suggesting side reactions
chewed up your material. Precisely. Unwanted
byproducts. Or maybe you see a large peak for
an intermediate compound that's supposed to react
further but hasn't. That points to a bottleneck
later in the reaction sequence. The HPLC quantifies
all these things, so you're not just guessing,
you're seeing the molecular evidence of what
went wrong. It really is like molecular accounting,
tracing where everything went. What about the
other scenario? A finished batch fails its purity
test and unexpited impurity shows up. Yeah, similar
approach. HPLC is usually the first line of attack.
That unexpected peak tells you something extra
is there. The next step is often to identify
it. Sometimes you can couple the HPLC directly
to a mass spectrometer, HPLC -MS. Ah, which gives
you the mass of the molecule. Exactly, which
is a huge clue to its identity. Or you might
try to isolate that impurity peak from the HPLC
and then analyze it using spectroscopy IR, or
perhaps NMR, nuclear magnetic resonance spectroscopy,
which gives even more detailed structural information.
The goal is to figure out the chemical structure
of that contaminant so you can then trace back
how it might have formed. It's pretty amazing
the level of detail you can get. It really highlights
that making medicines isn't just cookbook chemistry.
It's incredibly precise science and engineering.
Absolutely. And all this ties into the broader
concept of quality by design or QBD in pharma.
The idea isn't just to test quality at the end,
but to build it into the process by understanding
how different factors affect the outcome. And
these analytical tools provide the data to build
that understanding. Exactly. They let you monitor
the critical parameters in real time or near
real time, understand the process deeply, and
control it to ensure consistent quality batch
after batch. It's proactive. not just reactive.
So let's sort of wrap this part up. We've seen
how these analytical tools, things like HPLC
and spectroscopy IR and XRD included, are just
fundamental. They're the eyes and ears during
pharmaceutical process development. They watch
the reactions. They check the quality. They ensure
consistency. And they play that crucial detective
role when troubleshooting is needed. They let
scientists understand what happened at a molecular
level and figure out how to fix it or prevent
it from happening again. You see countless examples
of this detailed analytical work underpinning
process improvements in journals like OPRND optimizing
conditions, controlling tricky intermediates,
solving scale -up impurity issues. That's where
the practical application really shines. Right.
So it's this detailed monitoring and troubleshooting
underpinned by sophisticated analysis that ultimately
ensures the medicines people rely on are both
safe and effective every single time. It really
is the bedrock of quality, which maybe leads
to a final thought to chew on. Go for it. Well,
we've talked about these established techniques,
but analytical technology is always advancing,
right? Things are getting faster, more sensitive,
more automated. And now we're seeing things like
AI and machine learning starting to be applied
to analyze the massive amounts of data these
instruments generate. So the question is, how
might these future analytical capabilities, maybe
even smarter tools combined with AI, how might
they further revolutionize our ability to develop
complex medicines and guarantee their quality
even more robustly? What new levels of understanding
and control might become possible? That's a fascinating
question. Moving beyond just monitoring to maybe
even predicting and autocorrecting processes
in real time using that data. Definitely something
to think about. Yeah, it's an exciting space
to watch.