Biosimilars: How similar is similar enough?

Peter Sebo

sebo@biomed.cas.cz

Biotechnologický ústav AV ČR, v.v.i.

1

What are bilogics and biosimilars ?

Biopharmaceuticals =

Protein drugs produced using live organisms

Biodrugs

Biologicals, Biologics

Biosimilars (EMEA)

Follow – on biologics – FOB (FDA)

2Poděkování: doc. M. Fusek

3

Size matters

Erytropoetin

Atorvastatin

courtessy: doc. M. Fusek

Biologics versus conventional drugs

- production systems – Live organisms used

- molecular wight of the active comnpound

- different analytics used

- different formulation

- potential immunogenicity risk

- Local versus systemic action – selectiveness?

- Legislative obstacles/biosimilars – EMEA guidelines

4

5

Total sales and sales growth trend in the US biotech market for biologic drugs (2003–2008)

6

What's fueling the biotech engine—2008

Saurabh Aggarwal

Nature Biotechnology 27, 987 - 993 (2009)

Top companies that comprise the majority of sales of biologic drugs in 2008.

7

8

Very attractive business:

1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

in 2006 biodrugs represented 33 % of newly approved drugs

Biologicsbiosimilars

follow-on biologics

The issues…

Presentation title Date

11

Main risk =Imune response

• EPREX – Pure Red Cell Aplasia (PRCA) = anemia

• 1 per 5000 patints developed antibodies against EPO

• due to extractables form vial lids in the presence of polysorbate -adjuvans.

• Omnitrope® (somatropin) - Genotropin (Pfizer)57% of patients developed antibodies due to prsence of residual host protein contamination serving as adjuvant

• Purification redeveloped

• Phase III trial repeated

• Saspproved by EMEA, FDA, Australia

Approved Biosimilars - FOB

Name INN original approval

Omnitrope (San) Somatotropin Genotropin (Pf) 12 Duben 2006

Valtropin (Biopartners) Somatotropin Humatrope (EL) 24 Duben 2006

Binocrit (San) Epoetin alfa Eprex (Jan-Cil) 28 Srpen 2007

Epoetin Alfa Hexal (Hex) Epoetin alfa Eprex (Jan-Cil) 28 Srpen 2007

Apseamed (Medice) Epoetin alfa Eprex (Jan-Cil) 28 Srpen 2007

Retacrit (Hospira) Epoetin zeta Eprex (Jan-Cil) 18 Prosinec 2007

Silapo (Stada) Epoetin zeta Eprex (Jan-Cil) 18 Prosinec 2007

Ratigrastim (RatioPharm) Filigrastim Neupogen (Amg) 21 Únor 2008

Biograstim (CT Arzenm) Filigrastim Neupogen (Amg) 21 Únor 2008

Tevagrastim (Teva) Filigrastim Neupogen (Amg) 21 Únor 2008

Filgrastim Ratiopharm Filigrastim Neupogen (Amg) 21 Únor 2008

Future Global player – India• No. 2 producer of pediatric vaccines, No. 4 in drug manufacturing…

• big pharma uses it for outsourcing of R&D and production

• Úpatents prior to 1995 are not respected -1995 patenty =

• 48 biosimilars produced in India in 2008…

• Biosimilars 30-40% cheaper – sales grow

• Biosimilars – 35% - 65% of Indian markert of biologicals

• Qualified workforce - R&D costs 60% cheaper and 35% chepaer than in USA

• Contract research and manufacturing services – 1 mld. USD in 2010

• Indian companies buy EU biopharma firms to penetrate EU markets

• 37% of filed Drug master files in 2007 was from India …

• 130 indian biopharma companies…

• in 2006 sales of $1,5 bil. USD – CAGR 27%...

• 300 milions of Indian can afford expensive treatments - 11th marklet size

• EMEA-certifiied manufacturing of Inmtas Biopharmaceuticals Ltd…13

Will biosimilars really make it?

• The segment of the bipological market grows > 200 products on the market, >300 trials ongoing…

• 50% of pharma market by 2010 are biologicals?• Biologicals sales grew 6-fold in a decade …• The older population needs more expensive drugs …• The fastest growinhg market are antibodies (>$75 B)• >25% of drugs in the pipeline are biologicals• Unsaturated market / the producing capacity is not big

enoughnasycen – nedostatečné výrobní kapacity…• 7 patents on 8 key markets expire worth 12 bil. USD• 30% proice reduction and 30-40% market penetration expected• EU could save 20% of costs using biosimilars = 1,6 mld €/year• Much less competition then on generics – complicated

production…14

However, one should not get armed for the past war….Nowadays only a new improved generation of

biosimlars makes sense…

Big pharma biosimilar divisions ready to enter market…

Differ by application route, formulation, glycoprofile, efficay, stability etc.…..

15

Why biosimilars cannot be that much cheaper than original drugs

16

R&D costs and clinical trials plus marketing and market substitution costs high

Problems of biosimilars

• Much higher R&D costs than conventional generics

• Immunofgenicity risk

• Proving identity to original drugs difficult

• Higher production costs + clinical trials needed

• Biocomparability, PK, PD

• High marketing aand market substitution costs

• Much smaller margin in % of „biogenerics“

– Possible „discount“ 10 – 40 % - But how will originator react?

• High societal need – 30 % cheaper mAb is a lot of money

17

Presentation title Slide no 18 Date

The Basics of Biotech Manufacturing Processes and the Starting Point for Biosimilars

DNA Vector Cloning into

DNA Vector

Large-Scale Fermentation

Downstreaming

Formulation

Transfer into

Host Cell,

Expression

e.g., bacterial or mammalian cell

Maybe the same

genetic sequence

(Probably)

a different

DNA vector

Different recombinant cell

system

A different

fermentation

process

A different

purification

protocol

Different

in-process

controls

Maybe a

different

formulation

Presentation title Slide no 19 Date

An Entirely New Manufacturing Process – Innovator or Follow-on Lead to Potential Quality, Efficacy and Safety Issues

DNA Vector Cloning into

DNA Vector

Large-Scale Fermentation

Downstreaming

Formulation

Transfer into

Host Cell,

Expression

e.g., bacterial or mammalian cell

Maybe the same

genetic sequence

(Probably)

a different

DNA vector

Different recombinant cell

system

A different

fermentation

process

A different

purification

protocol

Different

in-process

controls

Maybe a

different

formulation

Clinical Studies are always required for significant risk process changes to demonstrate absence of clinically meaningful differences

Change

filter

supplier

Move

equipment

different part

of facility

Move

manufacturing

to new facility

New cell line

New Process

New Formulation

Scale-up

Manufacturing

Manufacturing Change

• Low risk• Frequent

• Supported by:

Analytical and

Process Data

• Highest risk• Rare

• Extensive Data:

Analytical, Process

and Human data

Follow-on Biologics

New:

•DNA?

•Cell line

•Process

Technology?

•Fermentation

process

•Purification process

•Analytics

•Facilities

•Formulation?

•And - no history

2nd generation innovator processes – and for FOB’sAppropriate human study to demonstrate the equivalence of the efficacy and

safety (including immunogenicity) of the product

Character of Change

Presentation title Slide no 21 Date

Omnitrope® (somatropin)Reference product: Genotropin (Pfizer)

• Early version of the product: 57% of patients developed antibodies against Omnitrope

– Problem was residual host-cell protein

• Re-developed purification process

• Conducted a second phase 3 study

• Antibody levels reduced

• Approved by EMEA, FDA and Australia

Clinical studies evaluating efficacy and safety required to identify and

exclude clinically meaningful difference

Presentation title Slide no 22 Date

Alpheon (alpha interpheron)Reference product: Roferon (Roche)

• Differences in impurity profile observed• Key Clinical Data

– PK (3 studies): supra-bioavilability/comparable/inconclusive

– PD (2 studies): PD equivalence/no PD equivalence

– Safety&Efficacy:• Clinically and statistically significant difference in virological

relapse rate found: more patients on Alpheon had relapse

• Different rate of adverse events and laboratory-related events judged as clinically relevant

Clinical studies evaluating efficacy and safety demonstrated clinically meaningful

differences leading to Rejection by EMEA

Presentation title Slide no 23 Date

Conclusion

• Biosimilars will be similar - not the same as innovator

• Characterization doesn’t tell the whole story – and chemical differences matter!

• Only way to exclude clinically meaningful differences in efficacy, safety and immunogenic potential – is from clinical data

• Product class specific guidance will ensure consistency

Presentation title Slide no 24 Date

There are no definitive technologies currently available to predict how the human immune system will respond to a product

• Immune responses to protein products can be extremely serious or life-threatening

• Immunogenicity is a significant characteristic of any protein:

– Must be included in overall comparability exercise

– Immunogenicity can only be determined in human studies…

On the other hand…

• Process can be reverse-engineered to someextent

• Several biosimilars already approved

• EMEA Guidelines available

• Some savings in clinical trials possible… refering to originator product

25

Proving identity: Are cutting edge technologies

getting us there?

dogma or propaganda?

• Comparison is possible only to marketed products…

• No in-process details or batches available from innovators…

• process = product - a dogma used as argument…

• Do laws of chemistry and physcis apply to biosimilars, too?

• Proteins are certainly not rigid and show structural dynamics and modification heterogeneity, but …

• Proteins too are chemical compounds, albeit complex…

27

Process equals product?Case study: Hen egg white lysozyme

28

Overal view Some flexibility/differences

in loops observed

Different crystal forms from different processes• 193L P 43 21 1 1.3Å• 2VB1 P 1 0.65Å• 1AKI P 21 21 21 1.5 Å• 1PS5 C 2 2.0 Å• 1VDP P 21 1.7 Å• Different chicken on differing diet in different countries…

29

Larger protein case:Coagulation factor VIII

• 2R7E P 41 21 2 3.7 Å

• 3CDZ P 41 21 2 3.98 Å

A very large protein crystallized in the same space group by different teams resolution not terrific but no great structure differences seen either. Loops may be poorly defined

The anti-dogma

• X-ray crystallography not the simplest and fastest …

• Cutting-edge Fourier transform mass spectrometry analytical methods and NMR fingerprinting are a breakthrough…

• Exact protein sequence confirmation not an issue

• Assessment of sample modifications not an issue

• Sample similarity (microheterogeineity pattern) assessment possible

• Can we start asking ‘how similar is similar’?

• Proving ‘chemical identity’ or protein and contaminant content (di-)similarity starts being really possible

• Very realistic in 5 years from now

30

The true question?

• The cutting-edge FT MS analytics will allow to optimize the process for biosimilars to yield an ‘identical’ protein product in terms of chemistry – how similar is similar enough?– Combination of ESI, ECD, MALDI approaches allows determining

chemical structure of posttranslational modifications and assessing of sample composition in terms of microheterogeneity

– Structure fingerprinting possible

• Biological activity, pharmacodynamics and equivalence testable

• Massive R&D investment for biosimilars required anyway…

• Is it faster and more cost-effective to make the biosimilarprocess yielding an ‘identical’ protein product or is it faster to go through the full clinical testing and registration procedure?

31

NMR fingerprinting of 3-D structure of unlabeled proteins as QC method…

Anal Chem. 2008 Apr 1;80(7):2623-7

Assessment of the three-dimensional structure of recombinant protein therapeutics by NMR fingerprinting: demonstration on recombinant human granulocyte macrophage-colony stimulation factor.

• Aubin Y, Gingras G, Sauvé S.

• We describe a simple, powerful, and robust NMR-based method that has the potential to greatly impact the characterization of recombinant protein therapeutics. The method

ascertains the bioactive conformational identity of recombinant human granulocyte macrophage-colony stimulation factor (rhGM-CSF) produced in Streptomyces lividans versus Escherichia coli by overlaying their 2D 1H,15N HSQC correlation spectra. An identical match of all resonances implies that

rhGM-CSF from both processes share indistinguishable conformations that correlate with in vitro activity. The result of this method is unique among existing

methods. It can detect and quantify the active ingredient. Moreover it provides a complete assessment of the conformation with high sensitivity to minor structural changes

32

No effect of excipients observed…

Used 700 MHz instrument –more performing instruments are in the pipeline….

Solve structure of reference product – compare fingerprints in process…

Comparison of active substancesanalytical scheme for active protein substance – start with basic properties

Sample

Charge

IEX, IEFMW

SEC, SDS-PAGE

(non- and reducing)

Hydrophobicity

RP, HIC Accurate MW

(LC-)MS:

ESI-FTMS

ESI-qToF2D PAGE / HPLC

Poor MW accuracy but provides

information about HMW forms

Highly accurate (sub- / ppm)

directly point on modifications and

isoforms. Still not routine for aggregates or

non-covalent oligomers

Information about number

of isoforms and/or approximate

MW

NMR

fingerprinting

Shape, structuration, oligomerization

centrifugation

gross shape,

conformational

changes

Dynamic light

scattering –

presence of

oligomers/aggrega

tion

Circular dichroism

far UV (190-250nm) –

secondary structure elements,

near UV (250-350nm) – some

aspects of tertiary structure

(follow aromatic amino acids

and SS bonds)

Fluorescence (Trp)

X-ray

full structure

Primary sequence / modifications

Unmodified protein – sequence by accurate mass not an issue anymore

Modified proteins – accurate mass point on the extent of modification

MS/MS on the entire intact protein (top-down)

+ link between selected isoform and its modifications preserved

- limitations – MW, precursor isolation

MS/MS of peptides from enzymatic digestion (bottom-up)

+ MW of the protein no longer limiting

- more difficult to link modification to parent molecule – specific cases where differentially modified isoforms exist

Glycation – Brady LJ – AnalChem 2007

Gadgil HS – AnalBiochem 2006

cutting-edge FT MS

Current limits - to be significantly improved in 3-5 years:

Dynamic range - 2-4 orders of magnitude

Sensitivity – sub-fmol

Resolving power – routinely 200,000 but much higher possible

Accuracy – sub-ppm / tens of ppb

instrument cost 2-3 M Euro

Required great guys…

SoloukiT AnalChem 1997

Selected modifications found on protein pharmaceuticals

• N-/O-glycosylation

• Pyro-Glu, Asn/Gln deamidation, isoAspartate

• Disulfide bridges

• 4-carboxyglutamate

• Sulfation

• Phosphorylation

• Glycation

• Oxidation

• PEGylation

Sample

Example work-flow for IgG characterization

digestion

Reduction/(alkylation)

Deglycosylation

(EndoH, PNGaseF)

LC-MS(/MS)-sequence coverage

-glycopeptides (type

and site of attachment)

LC-MALDI

iontrap (CID, ETD)

FTMS (CID, ECD)

Deglycosylation

(EndoH, PNGaseF)

LC-MS(/MS)

-MW after degly

to asses on

other modifications

ESI-qToF

ESI-FTMS

LC-MS(/MS)-sequence coverage

-confirmation of

glycosylation and

other PTMs

LC-MALDI

iontrap (CID, ETD)

FTMS (CID, ECD)

Sequencing of released glycans

MS(/MS)

HPAEC-PAD + MS

NMR

Glycosidases+MS

LC-MS(/MS)

-disulfide bonding

FTMS

digestion

TopDown

MS, MS/MSFT-MS (CID, EXC,IRMPD,…)

Process optimization case study

tuning the fatty-acylation…

5 15 25 35 Time [min]

Extremely large potential biopharmaceutical

(1706 aa residues)

PTMs consisting of palmitoylation

HPLC of trypsin digest

Process optimization case study

tuning the fatty-acylation…

Sequence coverage from a single LC/FT MS analytical run 94%

A mutation detected V-->L

Disulphides unambigously assigned

Acylation determined modified lysine 1 = 58% palmitoleate, 42% palmitate

modified lysine 2 = 37% palmitoleate, 43% palmitate

Tuning protein acylation in process:Extracted ion chromatograms from LC/FT MS

K1069

K946

unmodified C16:1

C16:1

C16:0

C16:0

unmodified

C16:1

C16:1

C16:0

C16:0

K1069

K946

unmodified

unmodified

Inappropriate process

1ASNSLQYVNVQVKDIEADLQHGVDESYTLDVEEDSDTITINAETVWGALHAFTTLQQLV

60ISDGHGGLIIEEPVNIKDSPLYPYRGIMLDTGRNFVSLPKIFEQLEGMSLSKLNVLHWH

119IDDAQSWPIWVDVYPEMVKDAYSPHEIYSRNDVRNIVNYARARGIRVIPEIDMPSHS

176SSGWKQVDPEMVTCTDSWWSNDDWPLHTAVEPNPGQLDIIYNKTYEVVGNVYKEL

231SDIFPDHWFHVGGDEIQPNCFNFSTHVTKWFAEDPSRTYHDLAQYWVDHAVPIFQ

286NYSQERRLVMWEDIALSADNAHDVPKNIVMQSWNNGLEYISNLTARGYDVIVSSSD

342FLYLDCGHGGFVTNDPRYNVMANPDANTPNFNYGGNGGSWCAPYKTWQRIYDYD

396FTLNLTETQAKHIIGATAPLWGEQVDDINVSSMFWPRAAALAELVWSGNRDANGNK

452RTTEMTQRILNFREYLVANGVQAQALVPKYCLQHPHACDLYRNQAAIQ

Disulfide Bridges:b-N-Acetylhexosaminidase

Six cysteine residues at positions: 189, 250, 347, 382, 482 and 489.

Used protease digestion sites underlined and in colour

Strategy

• Non-reducing SDS-PAGE

• In-gel deglycosylation (PNGaseF) and digestion(Trypsin and Asp-N).

• Capillary LC-FTMS (C18, 0.1x150 mm, 3mm, 200A,APEX-Q 9.4T).

• Data reduction (Averagine, THRASH and Z-score algorithms).

• Data interpretation (Links).

Deconvoluted spectra of cystine peptides from β-N-Acetylhexosaminidase (A. oryzae). (a) Asp-N cystine peptide Cys189/Cys250

at m/z 1857.7348. (b) Tryptic cystine peptide Cys347/Cys382 at m/z 5952.5998 and its oxidized form. (c) intrachain tryptic

cystine peptide Cys482/Cys489 at m/z 1616.7095. All peptides were measured with the error below 1.0 ppm. (d) The contribution

of 34S visible to the third isotope – I.) experimental data - cystine peptide Cys482/Cys489 at m/z 540.2401 (z = 3); II.) theoretical

spectrum of the same peptide; III.) theoretical spectrum of a peptide (SSDFLYLDCGHGGFV) which fits into the window (3ppm) of

accurate mass search, but has only one sulfur.

All 3 disulphides readily identified

The accuracy of FT MS is the key…

S-S Peptide(s) Exp. Theor. Error

1-2 183-190,244-251 1857.7348 1857.7344 0.22 ppm

183-190,244-251 ox. M 1873.7300 1873.7292 0.44 ppm

170-190,244-251 3284.3680 3284.3645 1.07 ppm

170-190,244-251,ox. M 3300.3573 3300.3594 0.67 ppm

3-4 332-358,359-386 5952.6206 5952.6069 2.30 ppm

332-358,359-386 ox. M 5968.6111 5968.6016 1.59 ppm

5-6 481-493+1S_S 1616.7095 1616.7102 0.45 ppm

S-S Linkages

1ASNSLQYVNVQVKDIEADLQHGVDESYTLDVEEDSDTITINAETVWGALHAFTTLQQLV

60ISDGHGGLIIEEPVNIKDSPLYPYRGIMLDTGRNFVSLPKIFEQLEGMSLSKLNVLHWH

119IDDAQSWPIWVDVYPEMVKDAYSPHEIYSRNDVRNIVNYARARGIRVIPEIDMPSHS

176SSGWKQVDPEMVTCTDSWWSNDDWPLHTAVEPNPGQLDIIYNKTYEVVGNVYKEL

231SDIFPDHWFHVGGDEIQPNCFNFSTHVTKWFAEDPSRTYHDLAQYWVDHAVPIFQ

286NYSQERRLVMWEDIALSADNAHDVPKNIVMQSWNNGLEYISNLTARGYDVIVSSSD

342FLYLDCGHGGFVTNDPRYNVMANPDANTPNFNYGGNGGSWCAPYKTWQRIYDYD

396FTLNLTETQAKHIIGATAPLWGEQVDDINVSSMFWPRAAALAELVWSGNRDANGNK

452RTTEMTQRILNFREYLVANGVQAQALVPKYCLQHPHACDLYRNQAAIQ

AAKAPFKIDFEVRRGESKDDLSPEDDSN

PRFVKRDGSLDMTLTNKQTFYMATLKIG

SNEDENRVLVDTGSSDLWVMSHDLKCV

SAPNSKRNERSFGHGTGVKLNERELMQ

KRKNLYQPSRTIETDEEKEASEKIHNKL

FGFGSIYSTVYITEGPGAYSTFSPFVGTE

GGSGGSGGSNTCTSYGSFNTENSDTFK

KNNTNDFEIQYADDTSAIGIWGYDDVTIS

NVTVKDLSFAIANETSSDVGVLGIGLPGL

EVTTQYGYTYQNLPLKLKADGIIAKSLYS

LYLNTADAKAGSILFGAIDHAKYQGDLV

TVKMMRTYSQISYPVRIQVPVSKIDVES

SSGSTTNILSSTTGVVLDTGSTLSYVFSD

TLQSLGKALNGQYSNSVGAYVVNCNLA

DSSRTVDIEFGGNKTIKVPISDLVLQASK

STCILGVMQQSSSSSYMLFGDNILRSAYI

VYDLDDYEVSLAQVSYTNKESIEVIGA

IKLDFNKVSTPSKYTKRDALPMPLIND

KILYTTELEIGSNKDKVSVSIDTGSYDL

WVMSNDAVCYKVSEFQTEGAPQLPDI

FNDIDQDYSCTFNGTYNSKSSKTFKN

TSEDFSIGYVDGSAAQGVWGYDSVQF

GQYGVTGLKIGIANRSSVSDGILGIGIA

NGYDNFPVLLQKQGLINKIAYSVYLNS

SNSTTGTILFGAIDHAKYKGALSTVPV

DSKSQLSVNVTNLKTKNGNVASGGHS

ILLDTGSTFSIFPDEWIDALGHSLDATY

DEDESVYEIECDGYDEHFFGFSIGDSD

FSVPIQDLKTEKDGQCYLAIMSNSVIG

GGGILFGDDILRQIYLVYDLQDMTISVA

PVVYTEDEDIEEIL

SAP 9 SAP 10

Rapid characterization of protein glycosylation at multiple sites

Strategy

• Protein reduction (TCEP) and alkylation (IAA)

• Enzymatic proteolysis (Trypsin and Asp-N)

• Capillary LC-FTMS (C18, 0.1x150 mm, 3mm, 200A, APEX-Q 9.4T)

• Data reduction (Averagine, THRASH and Z-score algorithms) and interpretation (GlycoS)

LC-FTMS 2D Plot of Sap10 Trypsin Digest and Spectra of Selected Glycopeptides.

GlycoS

SAP10 glycosylation – LC-MS/MS

Sample

microheterogeneity

Assessment

How similar

is similar enough…

Single hexose peak m/z spacing

SAP10 glycosylation – LC-MS/MS

Single hexose peak m/z spacing

Results

Generic therapeutic antibodies

Targeted biological - antibody drugs against

cancer, pain, arthritis, osteoporosis

50% of current biologicals market?

Robust

nowadays rather easy to produce

minimal postrranslational modification (PTM)

Straightforward analytics of PTM

probably the easiest to copy…

Can be produced in glycooptimized cells

established structure characterization for Fab

and Fc fragments

Take home message:

• Tremendous improvement and steady advance of analytical instrumentation makes in-process comparisons of chemical identity (similarity) of biosimilars and original products realistic

• The nebula of isoforms of original and similar molecules compared…

• Matter of time and cost-effcetiveness

• Biosimilars require substantial R&D effort anyway…

• Full clinical testing of biosimilars, can be lengthy (not only costly) due to limited numbers of patients

• Advanced analytics require costly equipments and great guys to operate them… - partner with academia!

54

Petr Man Petr Novákhttp://ms.biomed.cas.cz/equipment.php

Institute of Microbiology AS CR, v.v.i.

Prague, CR

Acknowledgements: