O C TO B E R 2 0 2 2
Biopharmaceutical
Process Development
Considerations for
Smarter and
Faster Process
Development
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
This custom ebook is sponsored by Cytiva and presented
in partnership with BioPharm International.
Introduction
S
mart process development uses tools and methodologies to improve process
outcomes and shorten time to market. New modalities entering the pipelines often
lack platform process solutions, so working smartly in process development may
become a key differentiator.
For example, an evolution in process development methodology has process developers
moving away from studying one factor at a time toward multivariate data analysis. This
analysis involves statistical models, such as design of experiments (DoE), and, more recently,
mechanistical models based on mathematical equations.
In addition, the use of parallelization and automation for experimental setup, known as high
throughput process development (HTPD), increases the amount of information available and
reduces the sample amount used.
All these tools and methodologies enable accelerated process development and improve
general process understanding, robustness, and performance.
In the following articles, we share some industry drivers to go smarter and faster in process
development as well as insights on how to make both upstream and downstream process
development workflows more efficient.
–Henrik Ihre, Ph.D.
Director, Strategic Downstream Technologies
Cytiva
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
What is driving the need to go smarter
and faster in biopharmaceutical
process development?
T
he biopharmaceutical industry experienced a
decade of tremendous success with more than
1600 commercially available drugs approved
for therapeutic use. At the time of writing in
mid-2022, more than 300 antibodies and recombinant
proteins are in late-stage development, and a great
number are expected to enter late-stage development in
the coming years (1).
Henrik Ihre, Director of
Strategic Downstream A constantly diversifying pipeline
Monoclonal antibodies (mAbs) continue to have a strong
Technologies at
position in the drug pipeline. At the same time, molecule
Cytiva outlines some diversity is constantly increasing thanks to medical
trends in the process innovation. For example, nearly a quarter (24%) of mAbs
development world. and antibody fragments in the pipeline are novel (e.g.,
OCTOBER 2022 | BIOPHARM INTERNATIONAL 3
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
bispecific antibodies, antibody fragments,
and antibody drug conjugates). And the
diversity expands beyond mAbs. The new INTERVIEW
class of mRNA vaccines is just one of Top challenges to overcome
several examples of that diversity, which when developing AAV
processes
also includes plasmids, oligos, cell and
gene therapies, and other novel vaccines
(FIGURE 1).
For these new modalities, one key may not always offer the best solutions.
challenge is the lack of established New products and formats may have to
platforms compared with the now well- be developed to offer novel and flexible
established processing protocols for platform solutions going forward.
traditional mAbs. In addition, these new However, smart process development
modalities have molecular properties combined with existing solutions will
that often differ from conventional enable a good starting point also for these
biomolecules. Therefore, existing tools new modalities.
FIGURE 1: The molecular diversity is constantly increasing, expanding beyond
mAbs, due to medical innovation.
Source: GobalDataTM, 2021
OCTOBER 2022 | BIOPHARM INTERNATIONAL 4
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
The need to reduce development
timelines and costs and secure
process robustness WEBCAST
Historical drug development timelines Navigating the diverse
used to be eight to ten years (FIGURE 2). antibody pipeline with
chromatography innovations
Now however, with accelerated clinical
development, the overall development
timeline can be shortened significantly,
with many companies demanding that However, shortened time to market is not
new molecules be developed in only the only important goal. Equally important
three to five years. Consequently, process is to design, with a quality by design spirit,
development activities have become a manufacturing process that is productive,
critical factors affecting the time to market scalable, and robust. Designing such
with several activities performed in parallel a process is enabled by smart process
with clinical trials. development ways of working.
Reference:
1. GlobalData [Online.] https://www.globaldata.
INTERVIEW com/. Accessed 17 May 2022.
Insights and strategies for
successful development of
scalable AAV processes
FIGURE 2: The typical timelines between investigational new drug (IND) and
biological license application (BLA) have historically taken eight to ten years, but
accelerated processes can shorten it to three to five years.
OCTOBER 2022 | BIOPHARM INTERNATIONAL 5
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
Insights and approaches to
increasing efficiency during
upstream process development
Andreas Castan, Ph.D., the Director of Strategic Upstream
Technologies at Cytiva, recently answered some questions
about process development and scaling.
Q: What does the workflow for process development look like,
and on what important areas should a process developer focus?
AC: Process development creates the basis for a sustainable
manufacturing process. The starting point for process
Andreas Castan, Ph.D.
development is the knowledge of the properties of the product.
Is it stable, unstable, or toxic for cells? This knowledge will
help you decide between two major process modes: fed-batch
development or continuous.
OCTOBER 2022 | BIOPHARM INTERNATIONAL 6
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
Process-development work starts with cell
line development. A good, high-producing
cell line that is stable and produces a product ARTICLE
with the desired quality attributes are key for Optimized process
your process. efficiency in upstream
manufacturing
Let me walk you through a typical workflow
from cell line development through process
development to pilot-scale production of be screened. Next, a relevant number of
material for toxicological studies (tox supply). clones are transferred into a small-scale fed-
FIGURE 1 illustrates what activities are batch development to assess the platform
performed, the analytical focus, how many fit, e.g., the lactate profile, feed compatibility,
clones can be carried into the next phase and sheer tolerance. This stage can be
of process development, and what culture performed in micro-bioreactors. Only a few
volumes are applied. clones are taken further into full process
development where the full product quality
After single-cell cloning, the clones are mainly is assessed in stirred tank bioreactors. The
assessed for titer in micro-titer format. The full process development phase studies
best clones are transferred to the next stage parameters such as pH, feeding strategies,
in deep-well plates. Product quality may also aeration strategies, etc.
FIGURE 1: Workflows in cell line development and upstream process development.
OCTOBER 2022 | BIOPHARM INTERNATIONAL 7
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
You should consider process intensification development activities on the requirements
at an early stage in the development. coming from the target molecule and
Process intensification should focus on the leverage platform technologies, e.g., media
bottlenecks in the process. Examples for platforms for fed-batch and perfusion, that
process intensification are N-1 perfusion, accelerate your development.
high cell density and high-volume cell banks,
and hybrid processes in the production Q: What are the major goals of moving from
bioreactors. Your objective is to always a small-scale culture to larger processes?
improve the volumetric productivity of the And what challenges should be considered?
process, i.e., making more product in smaller AC: When starting a small-scale development
bioreactors in shorter time. activity, you need to keep the end in mind. At
an early stage, you need to have information
The result of the full process development about the pilot, good manufacturing practices
is one fully developed clone. This clone is (GMP), and the commercial plant. This
used to produce the master cell bank (MCB) information is related to the plant equipment
and is scaled up to pilot scale to produce the and the staff. You should also get early
material used for toxicological studies. visibility of the manufacturing strategy, e.g.,
will production be in-house or outsourced
Apart from cell line and process to a contract manufacturing organization
development, the choice of cell culture (CMO)? FIGURE 2 illustrates a product moving
media and feeds is key to a high-producing from early research to process development
process. My advice is to focus your process- to manufacturing.
FIGURE 2: Process scale-up is a given step in all projects going from PD to
commercial manufacturing.
OCTOBER 2022 | BIOPHARM INTERNATIONAL 8
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
When moving from process development
to pilot or clinical manufacturing, there is
“ The strategic choice
a huge difference whether you run a fed- you make in process
batch or a perfusion process. The media development will impact
preparation capability is also an important
factor to consider. Furthermore, you need to manufacturability for a
plan how the upstream process connects to long time.”
the downstream train. For example, you need
to know in what cadence harvest material is
produced. Knowing your projected material
needs helps to determine the scale of clinical Q: What strategies and tools would you
and commercial manufacturing. consider for scaling to bioreactors?
AC: With bioreactors, you need to find
Other things you should consider suitable operating parameters for the
when designing a process in process larger scale to ensure sufficient oxygen
development include: transfer and CO2 removal. This process is
• The strategic choice you make in called a scale-up strategy. Think about the
process development will impact capabilities and limitations in your large-
manufacturability for a long time. scale bioreactor that you need to consider
• The complexity of the process design during process development. What are the
should be appropriate to the skills of capabilities and limitations in your large-
your staff. Design the process as simple scale bioreactor that you need to consider
as possible. in process development? In other words,
• The scalability and transferability of how can you develop a small-scale process
sensor technology and automation in a design space that overlaps with the
to a GMP environment is important. larger-scale design space? Ensuring an
Select technologies that work in overlap of the small-scale and large-scale
manufacturing. design spaces requires an established
• The factors affecting manufacturability scale-down model (FIGURE 3).
such as the number of feeds, the
process duration, and plant scheduling
considerations, should be taken into
account.
ARTICLE
• The qualities of the cell culture media
system, such as stability and ease Bioreactor scale up and
scale down: applications
of preparation and storage are also and tools
important factors for the ease of
manufacturing.
OCTOBER 2022 | BIOPHARM INTERNATIONAL 9
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
FIGURE 3: Using an established scaling model helps both scale-up and scale-down.
Both the scale-up and scale-down models in mind. Start with the implications of the
need to investigate in a multidimensional type of product you have and what type
design space, which is described by the of process can you run in your plant or
agitation, sparger geometry, and gas-flow at a CMO. The cell line determines the
rates. Tools are needed to ensure sufficient total productivity — invest in it in the
oxygen transfer and CO2 removal across cell line development process and take
scales. Some companies base their scaling several clones with you to final process
strategies on experience. Other companies development. Use platform technologies
use spreadsheet-based tools to guide where available, such as media and feeds,
them. Cytiva™ has developed a bioreactor single-use bioreactors, and automation
scaler that helps you establish sound solutions. Consider process intensification
scaling strategies. options at an early stage where it makes
sense for your process and plant. Develop
Q: What do you recommend to people a process and establish scale-down models
developing new processes? that will facilitate later scale-up. With all
AC: Again, it is important to keep the that in mind, you are in a good position to
end, i.e., the final manufacturing process, develop a successful manufacturing process.
This content is based off an interview with Andreas Castan, Ph.D., that can be viewed here.
OCTOBER 2022 | BIOPHARM INTERNATIONAL 10
HyClone™ cell culture
solutions and services
Media and feeds | Buffers and process liquids
Sera | Microcarriers | Custom services
Better upstream process development
outcomes — improve productivity and
protein quality.
Go smarter. Go faster.
cytiva.com/cell-culture
Cytiva and the Drop logo are trademarks of Life Sciences
IP Holdings Corp. or an affiliate doing business as Cytiva.
HyClone is a trademark of Global Life Sciences Solutions USA
LLC or an affiliate doing business as Cytiva.
©2022 Cytiva
For local office contact information, visit cytiva.com/contact
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
Tools and methodologies for smarter
downstream process development
John Scibetta, Advanced Chromatography Specialist at
Cytiva, explains how to make chromatography process
development smarter and faster.
EVOLUTION OF DOWNSTREAM PROCESS
DEVELOPMENT METHODS: FROM OFAT TO
MECHANISTIC MODELING
If you are experienced in biopharmaceutical process
John Scibetta
development, you know that the technology available to you
has changed dramatically over the years. Initially, process
developers used to depend on one factor at a time (OFAT)
methods, which required many experiments to develop
workable processes. It was time consuming.
OCTOBER 2022 | BIOPHARM INTERNATIONAL 12
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
Design of experiments introduces eventual operating space, or design
randomization space), are tested in a stochastic order.
Thankfully, process development scientists DoE, combined with the use of statistical
developed smarter methods and tools that methods, transformed the framework of
enabled time and cost saving (FIGURE 1). For modern experimental design.
example, they recognized that interactions
between factors could be evaluated with
data mining approaches such as multivariate
DoE, combined with the use of
data analysis using statistical modeling to statistical methods, transformed
help direct, focus, and reduce screening the framework of modern
efforts. So they employed design of
experimental design.
experiments (DoE), which is a systematic
approach to varying several experimental
parameters simultaneously to obtain Another benefit of DoE is that it allows
greater information. you to identify important interactions that
could have been missed if you used only
An important method made possible by OFAT methods. DoE minimizes the number
DoE was the principle of randomization. of experiments in a controlled fashion and
To avoid biased results, combinations of delivers information more quickly — all
running parameters (such as pH, mass load, while increasing process understanding. As
and conductivity) from the total population a result, you need fewer resources and save
they are meant to represent (i.e., the time and cost.
FIGURE 1: The evolution of process development methodologies.
OCTOBER 2022 | BIOPHARM INTERNATIONAL 13
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
High-throughput process development
allows parallel processing “What would take months
The next remarkable advance happened in with multiple scientists could
the late 2000s: high-throughput process
now be done [with HTPD] in
development (HTPD) using multi-well
plates. With HTPD, you can evaluate a wide weeks [by a single scientist].”
range of experimental conditions in parallel.
HTPD shortens the development time while – JOHN SCIBETTA
it increases the amount of information
available during early process development.
You can use HTPD to characterize a design was the practice of quality by design
space and define the process parameters (QbD). QbD was introduced in 2002
that need to be monitored and controlled. by the United States Food and Drug
“What would take months with multiple Administration (FDA) and formally
scientists could now be done in weeks by accepted in the US and Europe in 2009.
a single scientist,” recalls chromatography Other countries implemented it in the
expert John Scibetta. years that followed (1). Regulatory
guidance toward QbD began with the
Quality by design drives the need for recognition that increased testing does not
smarter process development improve product quality. Quality must be
Another great shift in process development built into the product.
FIGURE 2: The QbD principles.
OCTOBER 2022 | BIOPHARM INTERNATIONAL 14
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
QbD is a systematic approach to product of QbD. The question to answer is: How
design, development, and manufacture. does the interplay between critical material
It begins with predefined objectives attributes (CMA) of the resin, such as ligand
and emphasizes product and process density, and critical process parameters
understanding and process control. QbD (CPP) represent a risk of process variation?”
principles lead to better process outcomes This question of resin variability is typically
and, as a result, creates processes that will addressed during process characterization
inherently help you make quality products phase. If a risk is identified, a control
(FIGURE 2). strategy can be implemented to secure
process robustness.
Smart process development tools and
methods are important components of Mechanistic modeling: The future is now
QbD-compliant process development. Years ago, modeling lab work on a
QbD is based on sound science and quality computer may have seemed like something
risk management, which smart process out of a science fiction movie. But it’s now
development strategies elucidate. “QbD is a reality.
driving the FDA, Cytiva, and manufacturers
to develop technologies that support Mechanistic modeling is the most
different new methodologies and ways of recent smart process development
working,” says Scibetta. tool. Mechanistic models give you the
ability to create computer simulations of
Raw material insights support process chromatograms based on physiochemical
robustness phenomena known to occur in
“QbD seeks to develop a process that will chromatography. It is a revolutionary,
be resilient in manufacturing,” he continues. breakthrough methodology that gives
Raw material insights are an important even more process understanding.
component. “The critical material attributes
of the incoming raw materials, such as For example, this methodology uses
chromatography resins, are a key aspect differential equations that describe how
molecules move between resin beads and
inside the bead pores. It also uses adsorption
INTERVIEW
isotherms to quantify how molecules
Insider compete for ligands when binding.
recommendations
for getting started “Right now, we’re able to look at the
with mechanistic
modeling of physiochemical, first-principle interactions
chromatography within a column or within a resin bead
and then make predictions in silico using
OCTOBER 2022 | BIOPHARM INTERNATIONAL 15
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
a modeling software. We can do the TOOLS FOR SMARTER
equivalent of thousands of runs in the DOWNSTREAM PROCESS
software with significantly less wet work in DEVELOPMENT
the lab,” explains Scibetta (FIGURE 3). To benefit from smart process development,
you need the right tools. Cytiva has you
By using computer simulations, covered (FIGURE 4).
mechanistic models decrease the number
of experiments you need to do during Accelerating process development with
process development. It also increases the high-throughput development tools
design space you investigate. You gain the Parallelization and automation are key parts
following advantages: of HTPD, and multiwell plates and robotic
• Cost reduction and time saving: More minicolumns are the smart PD tools that will
data from fewer experiments, which help you do both.
accelerates your process development
• Improved robustness and efficiency to Cytiva’s PreDictor™ 96-well plates can
fulfill regulatory demands on quality be used with a robot or manually with a
by design multichannel pipette. Their versatility allows
you to easily develop results with a minimal
“In the end, what we get are more resilient, amount of investment of energy and capital.
flexible, and enduring processes,” Scibetta “They come in packages of four because
says. “We are reducing the energy and effort you would work it in triplicate to account
required to gain this knowledge, and [we’re] for pipetting errors. So, with four 96-well
achieving better outcomes because we’re plates, you can do 128 experiments in
coming closer to true understanding.” triplicate”, says Scibetta.
FIGURE 3: Mechanistic modeling creates computer-simulated chromatograms.
OCTOBER 2022 | BIOPHARM INTERNATIONAL 16
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
Multiwell plates can be used only in Developing robust processes using
static conditions. So, if you want to look Process Characterization Kits
at dynamic flow characteristics and If you’re a Phase III process developer, you
minimize volume, Scibetta recommends need to perform process characterization
minicolumns. Cytiva’s PreDictor™ and process validation (PC/PV) studies.
RoboColumn™ units can be used as an Cytiva’s Process Characterization Kits can
automated HTPD tool using a robotic help with those studies. Of course, the
liquid handler. earlier in your process development that
you investigate your resin characteristics,
Another HTPD technology that you the greater the compounded benefits you
could use is Cytiva’s recently developed will see.
Fibro™ adsorber. The adsorber is an
electrospun cellulose fiber matrix with Because chromatography resins are
an open pore structure. Mass transfer in always produced within a specification
the matrix is governed by convective flow interval, you need to investigate
rather than diffusive flow observed in whether the inherent variability within a
beads. This structure allows high binding manufacturer’s given specification range
capacities at very short residence times. impacts process outcomes, such as yield
Using Fibro™ rapid cycling chromatography, or productivity. Your process should be
your cycle times will be in minutes instead robust, so that it can handle these normal
of hours. variations.
FIGURE 4: Cytiva tools for smarter downstream process development.
OCTOBER 2022 | BIOPHARM INTERNATIONAL 17
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
“We worked with customers to look at and develop a solid control strategy to
their CPPs and how they interact with obtain a robust and scalable downstream
the resin CMAs. Porosity and bead size chromatography process (FIGURE 5).
can be considered, but ligand density
is most critical,” Scibetta explains. Adding mechanistic modeling to your PD
Quantifying these interactions is important with Cytiva modeling software, columns,
for process understanding. They make a and support
process that is resilient. With mechanistic modeling, you
can move from empirical work to in
Cytiva’s Process Characterization Kits allow silico computational work. GoSilico™
you to study the potential impact that resin Chromatography Modeling Software
ligand density might have on the process democratizes computer simulation and
outcome. For a given chromatography gives you the ability to build your own
resin, the kit consists of three 25 mL mechanistic models. This software creates
bottles with different ligand densities that digital twins of downstream processes. It can
represent low, average, and high values in be used for a wide range of molecules and
the manufacturing envelope. You can obtain applications from early process development
critical insights on resin variation within to late-stage scale-up, troubleshooting, and
the ligand density specification interval chemistry, manufacturing, and controls.
FIGURE 5: Studying how the interplay of process parameters and resin variability
impacts process outcome during process characterization (left). Process parameter
target values can be defined to minimize the impact of resin variability with a control
strategy (e.g., shifting process parameter target value) (right).
OCTOBER 2022 | BIOPHARM INTERNATIONAL 18
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
To design a resilient process, you must modeler could become comfortable in
first characterize your chromatography less than a year with support from Cytiva.
column in the lab. Performing column “I think mechanistic modeling is worth
characterization experiments takes trying out. It’s the closest thing you get to
time and requires expert knowledge truly understanding what’s going on in the
to ensure that accurate parameter column,” he says.
values are produced. “Characterization
is exacting work — somewhat art and “With the f(x) column and the software,
a lot of technique. To unburden the I’ve seen one person — a single process
end-user, Cytiva has created a series of developer — create a model with some
precharacterized f(x) columns,” explains support from our team. He never had
Scibetta. With f(x) columns, your parameter any modeling or mathematics experience,
values are instantly available to provide and he was able to create a model for
faster and more reliable mechanistic an adeno-associated virus (AAV)
modeling results. Specific values are purification with a little help. These are
given for column parameters such as total some powerful tools. Users will be amazed
porosity, interstitial porosity, ionic capacity, at how quickly they develop expertise
specific adsorber surface area, and ligand and excellence.”
density. These parameters are plugged into
the GoSilico™ Chromatography Modeling CURRENT AND FUTURE APPLICATIONS
Software to begin the process of creating a FOR MECHANISTIC MODELING
mechanistic model. Scibetta suggests that a key application for
modeling is cation exchange chromatography
for mAb processes. “Cation exchange is a
INTERVIEW very important part of the mAb platform
Using mechanistic process. Although it’s relatively well-
modeling in process characterized and understood, I’ve seen a
development
for advanced lot of customers having a lot of success with
therapeutics mechanistic modeling.”
Other important applications are anion
To help you get started using exchange chromatography and polishing
chromatography modeling software, Cytiva for new therapies, such as AAV or virus-
offers consulting, training, and contract like particles (VLP) which do not have
modeling services to guide you on your established platforms. “The good news
mechanistic modeling journey. “The time is that a virus is relatively easy to model
to learn how to model is time well spent,” because of its characteristic morphology
says Scibetta. He estimates that a nascent and charge regularity,” says Scibetta.
OCTOBER 2022 | BIOPHARM INTERNATIONAL 19
The Need for Smarter Smarter Upstream Smarter Downstream
Process Development Process Development Process Development
“I think the greatest knowledge and “I don’t think it will supplant empirical
understanding of making commercial models. Rather, HTPD, DoE, and
processes for these will be effectuated statistical models will support the outputs
through mechanistic modeling.” from a mechanistic model. I think that
chromatographers are going to enjoy being
in the lab more than ever because of the
“Chromatographers are going ability to do modeling.”
to enjoy being in the lab more
than ever because of the This article was based on a video interview with
Cytiva’s John Scibetta, which can be viewed at
ability to do modeling.” http://www.processdevelopmentforum.com/
articles/smart-process-development-approach/.
– JOHN SCIBETTA
REFERENCE:
“I think mechanistic modeling is the 1. FDA Guidance for Industry, Q8(R2)
future. It’s a 21st century tool for Pharmaceutical Development, Nov 2009 https://
chromatography,” concludes Scibetta. www.ich.org/page/quality-guidelines.
www.cytiva.com
Cytiva and the Drop logo are trademarks of Life Sciences IP Holdings Corp. or an affiliate doing business as Cytiva.
Fibro, GoSilico, and PreDictor are trademarks of Global Life Sciences Solutions USA LLC or an affiliate doing
business as Cytiva. RoboColumn is a trademark of Repligen GmbH Corporation.
© 2022 Cytiva
CY28833-08Oct22-EB
OCTOBER 2022 | BIOPHARM INTERNATIONAL 20
Mechanistic
chromatography
modeling
Software | Columns | Services
Replace lab experiments with computer
simulations — get thousands of results
in a few hours.
Go smarter. Go faster.
cytiva.com/modeling
Cytiva and the Drop logo are trademarks of Life Sciences IP
Holdings Corp. or an affiliate doing business as Cytiva.
©2022 Cytiva
For local office contact information, visit cytiva.com/contact
