1 of 47

Biopharmaceuticals

Challenges in formulation, pro- duction and quality control

Michael Jeltsch, University of Helsinki�& Wihuri Research Institute

29.10.2025

The most recent version of this presentation:�https://mjlab.fi/challenges

MPHARM-002B

PROV-105B

Slide 1 of 46

MPHARM-002B

PROV-105B

& UH

2 of 47

Antibody fusion proteins: Eylea^®, OPT-302

Slide 2 of 47

PROV-105A

& UH

3 of 47

Biologics

Slide 3 of 47

PROV-105A

Biologic(al)s/biological drugs/biopharmaceutical, biologic(al) medical products

A drug that is produced by/from living organisms or contains�components of living organisms.

Is this a good definition?

What about the difference between synthetic and natural vitamins�or the chemical synthesis of small proteins (Merrifield Solid‐Phase�Peptide Synthesis)? Some small-molecule drugs are isolated from�bacteria but are not counted as biologics (e.g. bleomycin, hygromycin).

Old FDA definition

“a virus, therapeutic serum, toxin, antitoxin, vaccine, blood, blood component or derivative, allergenic product, protein, or analogous product, or arsphenamine or derivative of arsphenamine (or any other trivalent organic arsenic compound), applicable to the prevention, treatment, or cure of a disease or condition of human beings”

& UH

4 of 47

Amended definition by the FDA (2020)

Slide 4 of 47

PROV-105A

https://www.federalregister.gov/documents/2020/02/21/2020-03505/definition-of-the-term-biological-product

“Protein means any alpha amino acid polymer with a specific, defined sequence that is greater than 40 amino acids in size.”

Previously, chemically synthesized polypeptides had not been classified as biologics.
Complexity is the important feature (FDA)!
But: In the EU, origin still matters (EMA), example semaglutide (33 amino acids)
For plant-derived “biologicals”, all bets are off (supplement, homeopathic medicine, …)

& UH

5 of 47

Different types of biologics

Slide 5 of 47

PROV-105A

Vaccines

Cell therapy

Gene therapy

Proteins

Transplants

Viruses
DNA
RNA

autologous or allogenic

Stem cells (“regenerative medicinal products”)
Cell-based immunotherapy�(e.g. CAR-T)��

Antibodies:�- polyclonals�- monoclonals�- biosimilars
Protein hormones,�growth factors & cytokines
Enzymes
Protein toxins
Peptides

cell, organ &�tissue trans-�plants (bone marrow, blood, blood products
fecal transplant
xenotransplants
biomaterials

Live virus
Killed virus
Recombinant vaccines
DNA/RNA vaccines
Tolerogens

Advanced Therapy Medicinal Products (ATMPs)

“Traditional”

biological drugs

& UH

6 of 47

Unifying theme of biologics (excluding viruses)?

Slide 6 of 47

PROV-105A

Almost all biologics need to be injected�Biologics are inactivated/degraded in the digestive tract
Most virus-based drugs are also injected�Exceptions: oral polio- & rotavirus vaccination, nasal flu shot, …
Holy Grail of protein drugs: Oral delivery

& UH

7 of 47

Proteolytic cleavage requires water!

Slide 7 of 47

PROV-105A

For chronic diseases, self-administration is highly desirable, but self-administration poses restraints to the formulation

Proteins must be dissolved ↔ Proteins are metastable and degrade relatively quickly in aqueous solutions over time

& UH

8 of 47

Unifying problems of biologics

Slide 8 of 47

PROV-105A

Prefilled syringe/ready dissolved vial: Aggregates are more immunogenic than dissolved material.
Temperature sensitivity → No terminal sterilization but aseptic processing and/or sterile-filtration.
Shearing during filtration (long nucleic acids)
Light-sensitivity (both nucleic acids and proteins are destroyed by UV: you measure them at 260/280 nm)
Lyophilization (freeze-drying) is the preferred method for long-term storage of proteins.

& UH

9 of 47

Biologics are large and complex molecules

Slide 9 of 47

PROV-105A

Aspirin

Typical IgG antibody

& UH

10 of 47

Biologics are large and complex molecules

Slide 10 of 47

PROV-105A

NXT/S (asparagine-�any aa-threonine or serine) = N-linked

glycosylation

Biologics are always chemically somewhat heterogeneous.

Biologics are defined by the pro- duction process (not by a precisely defined chemical structure).

Example: N-linked glycosylation

Small changes to the process can result in hugely different product

& UH

11 of 47

Heterogenous N-linked glycosylation

Slide 11 of 47

PROV-105A

& UH

12 of 47

Biologics are large and complex molecules

Slide 12 of 47

PROV-105A

Protein destruction

(Chemical modification)

Denaturation

Aggregation

↶↶

& UH

13 of 47

Enemies of physical stability

Slide 13 of 47

PROV-105A

Heat

→ Denaturation/Aggregation

How to protect proteins from becoming denaturated?

& UH

14 of 47

Enemies of physical stability

Slide 14 of 47

PROV-105A

Why does the egg white denaturate despite no pH and temperature change?

https://foto.wuestenigel.com/woman-beats-egg-whites-with-a-mixer-top-view-of-the-process-of-making-a-pie/

& UH

15 of 47

Enemies of physical stability: phase transitions

Slide 15 of 47

PROV-105A

& UH

16 of 47

Bioreactor types and operation modes

Slide 16 of 47

PROV-105A

Different operation

modes:

Batch: Inoculate final culture with cells. Done.
Fed-batch: Inoculate final culture with cells, but keep adding nutrients/medium during production (mostly glucose)
Continuous feed: Inoculate final culture with cells, add continuously medium during production and withdraw continuously medium & cells for purification.

Most protein drugs are secreted proteins.
Turbulent flow is more stressful for cells and proteins then laminar flow

& UH

17 of 47

Enemies of physical stability

Slide 17 of 47

PROV-105A

General rules how to handle proteins

Temperature: Keep cold!
pH: Never change the pH (e.g. by diluting into a buffer with a different pH)!
Phase transitions: Freeze (in the factory) and thaw maximally once (for usage)! Avoid making bubbles!

But:

Proteins are individualists! Some proteins are thermostable, some proteins are phase-transition-resistant, etc.

& UH

18 of 47

From inoculation to harvest

Slide 18 of 47

PROV-105A

Cells don’t like to grow alone!

& UH

19 of 47

Unifying themes in the production of biologics

Slide 19 of 47

PROV-105A

More difficult to produce then chemical synthesis of small molecule drugs (SMDs)!
Even the industry-leaders screw it up sometimes�→ Need for manufacturing process controls and well-defined target specifications
Analytical methods are important (but very challenging)
Biological products can be too complex/heterogenous to be fully characterized using physicochemical testing methods (they are defined via the production process!)

& UH

20 of 47

Unifying themes in the production of biologics

Slide 20 of 47

PROV-105A

Biological activity assays are needed!
Small differences in the production process can lead to big changes in the product.
Possibilities of contamination: viruses, animal-derived components
Fetal calf serum is (despite fully-synthetic alternatives) still a standard component of cell culture medium in the production of biologics (prion proteins, etc.); ethically and environmentally unsustainable
Inertia: it is very expensive to change a once-approved process!
The difficulty of production is reflected in the price! Per-dose-price typically manyfold greater than for SMDs.

& UH

21 of 47

Industrial lyophilization

Slide 21 of 47

PROV-105A

https://commons.wikimedia.org/wiki/File:Lyophilization_of_Pharmaceuticals_in_Vial.JPG

& UH

22 of 47

Contract (& development) manufacturing organizations

Slide 22 of 47

PROV-105A

Clinical trial sponsors do seldomly manufacture the biologics them- selves (Herantis Pharma used Finvector to produce Lymfactin, etc.)
Examples of contract manufacturing organizations: Samsung Biologics, Lonza, MedPharm, Genezen, Finvector (Nadofaragene firadenovec/Adstiladrin)
Supply challenges during Covid-19 pandemic: Many key ingredients of synthetic cell culture media are near-exclusively produced in China.

& UH

23 of 47

Contract (& development) manufacturing organizations

Slide 23 of 47

PROV-105A

If it works in the research lab it does not mean that it can be upscaled for industrial production.
Transfer of the technology research lab to the manufacturing organization.
Pre-clinical technology may be deemed inappropriate for scale-up.
Pre-clinical assets need to be re-implemented under GMP conditions (e.g. cell lines).

Upscaling

& UH

24 of 47

Contract (& development) manufacturing organizations

Slide 24 of 47

PROV-105A

If it works in the research lab it does not mean that it can be upscaled for industrial production.
Transfer of the technology research lab to the manufacturing organization.
Pre-clinical technology may be deemed inappropriate for scale-up.

Roller bottles: upscaling without upscaling

In use since 1989 (epoetin alfa, Amgen)?

& UH

25 of 47

Contract (& development) manufacturing organizations

Slide 25 of 47

PROV-105A

CMO have been expanding in both directions: formulation development capabilities & more (→CDMO)
Formulation development, scale-up and GMP manufacture by the same team

& UH

26 of 47

SMD (roxadustat) versus biologic (epoetin)

Slide 26 of 47

PROV-105A

New SMD drug competing with epoetins: Roxadustat (FG-4592), a prolyl hydroxylase (PHD) inhibitor for CKD works better than epoetins and is an oral drug. It was approved by the EMA in 2021 (but is not yet available in Finland).
Risk of immune response against the Epoetins! This wipes out all endogenous EPO resulting in red cell aplasia. Most prevalent in subcutaneous delivery of Eprex between 1998-2003 (epoetin α). Most likely cause is the action of an “adjuvant”, which stimulated the immune response (adjuvant = organic compounds leaching from uncoated rubber stoppers in the prefilled syringes, the rubber stoppers were replaced by teflon stoppers).

https://commons.wikimedia.org/wiki/File:Injection_Syringe.png

& UH

27 of 47

Different expression hosts/systems

Slide 27 of 47

PROV-105A

Bacteria (E. coli)

Yeast

Insect cells

Mammalian cells (CHO)

Cell-free expression/ transcription

Chemical synthesis

Transgenic organisms

Virus vectors

(gene therapy)

Proteins�(incl. antibodies)

Small

peptides

Nucleic

acids

Low yield - High yield

High quality - Low quality

& UH

28 of 47

Protein production in E. coli

Slide 28 of 47

PROV-105A

& UH

29 of 47

Good manufacturing practice (GMP)

Slide 29 of 47

PROV-105A

GMP (Good manufacturing practice) is a system for ensuring that products are consistently produced and controlled according to quality standards.

Raw materials (some raw materials are of variable quality, e.g. FCS)
Standard procedures (low batch-to-batch variability)
Documentation
Equipment & Premise
People

Rules applicable to the pharmaceutical industry�are enforced by FDA (US), EMA (EU), etc.

Quality control is important but cannot replace GMP!

& UH

30 of 47

How to implement GLP/GMP in an academic lab?

Slide 30 of 47

PROV-105A

Atul Gawande

Better: A Surgeon's Notes on Performance
The Checklist Manifesto

How to implement GLP/GMP in an academic lab?

& UH

31 of 47

Cell line generation & banking

Slide 31 of 47

PROV-105A

Cells are alive and change over time!
Production clone selection → characterisation/analysis → cell banking

& UH

32 of 47

Storage, handling, & distribution

Slide 32 of 47

PROV-105A

Can the product be lyophilised and reconstituted without significant change in properties? Freeze-dried protein is often stable for many years! Reconstitution in the pharmacy/hospital pharmacy
Real-Time versus accelerated shelf-life testing (mostly higher temperatures): Accelerated tests are needed, but real-time testing is the gold standard.�Forced degradation tests (freeze-thawing cycles, mechanical stability, etc.)
Is cold storage required? As a rule of thumb for biologics: YES!
Stabilizing agents? … are under development (Vincenzo Cerullo)

Name (Producer)	Vaccine type	Shelf life
Comirnaty (BioNTech/Pfizer)	mRNA	12 months at -60°C to -90°C, 31 days at 2°C to 8°C
Spikevax (Moderna)	mRNA	12 months at -15°C to -50°C, 30 days at 2°C to 8°C
Vaxzevria (AstraZeneca)	Adenovirus	6 months at 2°C to 8°C
Nuvaxovid (Novavax)	Protein	6 months at 2°C to 8°C

& UH

33 of 47

Analytical challenges

Slide 33 of 47

PROV-105A

Characterisation methods are largely determined by the protein’s properties, orthogonal (= different) methods increase the reliability!

Identity: mass, sequence (mass spec)
Quantity/concentration
Appearance, pH, osmolality
Isoelectric point
Post-translational modifications (disulfide bridges, glycosylation)
Every protein needs one or more dedicated biological activity assays!

Cell-based assays (can be theoretically developed for all proteins)
Whole animal-based assays (e.g. Botox)

& UH

34 of 47

Cell-based assays

Slide 34 of 47

PROV-105A

& UH

35 of 47

Botox bioassay

Slide 35 of 47

PROV-105A

1 unit = mouse LD₅₀ (female Swiss-Webster mice after single intraperitoneal injection)
The mouse assay is the accepted “gold standard”.
Does not scale (compare to the Covid-19-PCR assay)!

& UH

36 of 47

Analytical challenges

Slide 36 of 47

PROV-105A

Characterisation of protein impurities

Analytical methods

SEC: Size exclusion chromatography**
CE-SDS: Capillary electrophoresis-SDS, similar to SDS-PAGE (Polyacrylamide gel electrophoresis) + Coomassie Blue staining
Dynamic Light Scattering (DLS): size distribution

Host cell proteins (HCP)

Protein of Interest, but aggregated or fragmented (inactive)

Microbial contaminants

Protein of Interest

Product-related impurities

SDS-PAGE

CE-SDS

Medium proteins*

& UH

37 of 47

Protein mass spectrometry (MS)

Slide 37 of 47

PROV-105A

Digestion gives “tunnel vision” (what if a protein is protease resistant?)
Automated analysis depends on database quality (what if a protein is not in the database?)
Post-translational modifi- cations are very challenging!
Proteins and long peptides are difficult to measure exactly, short peptides are difficult to assign to proteins (→MS/MS)
De-novo MS sequencing (you get the best mathematical fit, not the only possibility)
Reducing sample complexity before MS by 1/2-D gel & ((U)HP)LC → many samples!

10.3390/proteomes8030014

& UH

38 of 47

Contaminations & safety

Slide 38 of 47

PROV-105A

The multi-step production process is susceptible to microbial contamination

Sources of contamination

Equipment
Personnel
Raw materials

& UH

39 of 47

Viral contamination

Slide 39 of 47

PROV-105A

Human cell lines in the production of biologics have advantages and disadvantages…

+ Host cell protein contamination is much less immunogenic.

− Human cell lines support the replication of human viruses.

Severe drug shortage in 2009 due to viral�contamination: Cerezyme for Gaucher disease)
Viral contamination of a CHO culture of�Boehringer Ingelheim in 2003 (https://doi.org/10.1099/vir.0.19042-0)

Cleaning and disinfecting multiple-use bioreactors�is difficult.

→ Disposable bioreactors for up to 4000 liters

Vesivirus (2GH8)

& UH

40 of 47

Contaminations & safety

Slide 40 of 47

PROV-105A

Typical contamination from bacterial cells: Lipopolysaccharides (LPS, “endotoxins”)

Major component (¾) of the outer membrane of gram-negative bacteria (E. coli)
Is a potent non-specific activator of the�immune system
Removal especially important when bacteria are used as the host organism!

https://www.frontiersin.org/articles/10.3389/fphar.2020.00554/full

& UH

41 of 47

Sourcing biodiversity to improve pharmacokinetics

Slide 41 of 47

PROV-105A

GLP-1 agonists are used in diabetes & obesity treatment.
Technically not biologics (<40 amino acids)!
Human GLP-1 has poor pharmacokinetic properties.
Exendin-4: GLP-1 homolog isolated from the venomous Gila lizard saliva with a longer plasma half life (more resistant to DPP4)
Liraglutide/Semaglutide are sequence-modified and acylated.
Tirzepatide (Mounjaro, made by organic synthesis) outperformed sema- glutide (Ozempic/Wegovy/Rybelsus, made in yeast + chemical modification)
Rybelsus is an oral drug (permeability enhancer salcaprozate sodium/SNAC)

https://commons.wikimedia.org/wiki/File:Gila_monster2.JPG

10.1021/acs.jmedchem.5b00726

& UH

42 of 47

Removing host cell proteins (HCPs)

Slide 42 of 47

PROV-105A

Host cell proteins (HCP) are a typical impurity (currently <100 ppm is allowed for protein products, 0.01%)
Even if the biological drug molecule itself is not immunogenicity, HCPs may be highly immunogenic
It can be difficult (impossible) to get rid of all HCPs by protein purification
The starting material contains thousands of different proteins. The POI is sometimes a minor component of�the starting material. Reasons for co-purification:

Chemical-physical similarities with the product molecule.
Natural interactions between the POI and other proteins are often functionally essential and can�be very strong (BSA)

Many "household" proteins are known, which are�difficult to separate chromatographically,�e.g. alpha-enolase, aldolase, ribonucleoproteins
Room for improvement: e.g. secretion deficient cell-lines

& UH

43 of 47

Challenges: Viral therapies

Slide 43 of 47

PROV-105A

Technical & safety aspects

Viral vs non-viral gene therapy
In-vivo vs ex-vivo gene therapy
Replicating vs non-replicating virus
Traditional gene therapy → non-replicating
Oncolytic viruses → conditionally replicable�(e.g. only replicate in cancer cells)
Inactivation by immune response
Permanent changes in target cells (AAV)

Commercial & ethical aspects

How to cover the costs of drug development�if the product is used once per patient?
Can s single dose of a drug cost $2 Mio.? https://innovativegenomics.org/news/paying-for-crispr-cures/

& UH

44 of 47

Challenges: Biosimilars & Interchangeable biosimilars

Slide 44 of 47

PROV-105A

The biosimilars market is exploding because patents are expiring.
Biosimilars require ~250 test methods for QC (SMD: ~50 test methods)
Comparable with the originator product

Physicochemical characterisation
Biological activity
Safety, immunogenicity
Preclinical in-vivo reproducibility
Clinical comparability

Interchangeable biosimilar (US market)�Can be exchanged without consulting with the prescriber (i.e. in the pharmacy). The EMA regards all biosimilars as interchangeable.
Biosimilars still relatively new on the pharmaceutical market�→ Regulatory processes are evolving and requirements are changing

& UH

45 of 47

Challenges: Intellectual Property (IP)

Slide 45 of 47

PROV-105A

~70% of all new drug patents are for biologicals.
Only the US and EU are considered significant patent-protected markets.
Nearly impossible to develop a drug without IP protection! No “trade secrets” in the pharma industry!
Nucleotide and protein sequences can still be patented both in the US and EU. Limitations:

Sequence alone is not sufficient, there must be an application or other “inventive” act.
Must be sufficiently isolated from “nature” = cDNA does not exist in nature.

Greedy patents: “sequence and all related sequences with >90% homology” → human patent covers monkey, but not dog or mouse sequences.
Bottom line: IP is complicated. You need a patent lawyer, even if it costs you $800/h.

& UH

46 of 47

Challenges versus advantages of biologics

Slide 46 of 47

PROV-105A

	Advantages	Challenges
Properties	Nearly identical to endogenous molecules, strong and specific effects.	Complex ADME (absorption, distribution, metabolism, and excretion). Drug is often a heterogenous mixture of biologically active variants. Irreversibility of some treatments (gene therapy).
Production	Recombinant DNA technology allows for the production of virtually any protein drug. Use of autologous, patient-derived (stem) cells for therapy.	Production is difficult. Analysis is difficult.
Safety	Generally very safe and non-toxic (since often identical to the body’s own proteins/cells). Typically less side effects.	Can provoke an immune response (“anti-drug antibodies”).
Usage	Future potential: single-dose curative treatment drugs (gene therapy).	Lower stability (cold storage required). Must be injected for the most part.

& UH

47 of 47

That’s all, folks!

End

Slide 47 of 47

PROV-105A

This presentation https://mjlab.fi/uef

Lab http://mjlab.fi

Private rumblings https://jeltsch.org(/science)

More questions? Ask via email: michael@jeltsch.org

@JELTSCH

& UH