310
volume 29 number 4 april 2011 nature biotechnology
Acceptable changes in quality attributes of glycosylated biopharmaceuticals
To the Editor:
Since the first marketing approvals of recombinant biopharmaceuticals, the question of which changes in quality
attributes, which compri identity, strength and purity, are acceptable in the life cycle of the products without changing the product label has been debated extensively 1. This question is especially important in the context of manufacturing process changes, which happen quite frequently and for
various reasons (e.g., process improvements, scale changes or site transfers). Although companies and health authorities have been managing the quality changes for many years bad on the principle that changes in quality attributes can be accepted only if they do not alter safety and efficacy, the lack of peer-reviewed data in the public domain has limited debates about product qualit
y and variation to a discussion of principles rather than specifics. Here, we prent a study that looks at variation in three major marketed biologics, the purpo of which is to provide more transparency and to anchor the debate about acceptable changes in quality attributes on a firmer factual footing. Identifying such variations in quality attributes could help not only biotech companies in their development efforts but also the medical and scientific communities in understanding the鸡同鸭讲近义词
products. By analyzing the quality profiles of the glycosylated recombinant therapeutic proteins Aranesp (darbepoetin alfa), Rituxan/Mabthera (rituximab) and Enbrel (etanercept) sourced from the market between 2007 and 2010, our data thus provide examples of acceptable variations for products that have remained on the market with unchanged product labels.
Glycosylated proteins are complex molecules and even a well-controlled product may consist of veral hundred or more glycoforms having the same amino acid quence but different glycan composition. When making the products, the manufacturer has to deliver a consistent product quality to guarantee a reproducible
clinical performance. Current analytical methods allow the detection of even small changes in quality attributes and can therefore enable nsitive monitoring of the batch-to-batch consistency and variab
ility of the manufacturing process. Several different factors may account for changes in quality attributes. The first is the inherent batch-to-batch variability in the manufacturing process. Second, process drifts can lead to gradual changes of attributes. Such drifting events are not desired and normally trigger further investigations and corrective actions, or even redevelopment activities to ensure process consistency. Finally, larger and
abrupt changes in quality attributes can occur after implementation of manufacturing process changes, which are all too common in the pharmaceutical industry. Although manufacturers try to prevent associated changes in quality attributes, such changes cannot be avoided in every ca.
Changes in the biologics manufacturing process are tightly regulated by the health authorities. Manufacturers need to
demonstrate that the process change does not alter the clinical safety or efficacy of the biologic product. The evaluation of such changes follows a comparability exerci between the pre- and post-change product, which is focud on the quality level and sometimes, depending on the magnitude of the change and the existing product understanding, also requires comparative data on the preclinical and clinical levels. The principles of the comparability exerci are regulated in guidelines, such as the
International Conference on Harmonisation (ICH) Q5E (/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Quality/Q5E/Step4/Q5E_Guideline.pdf), which acknowledges that “the demonstration of comparability does not necessarily mean that the quality attributes of the pre-change and post-change product are identical, but that they are highly similar and that the existing knowledge is sufficiently predictive to ensure that any differences in quality attributes have no adver impact upon safety or efficacy of the drug product.”
Comparability decisions are difficult to make and require the complete asssment of the existing process and product including the knowledge of structure-function relationships. Others 1 have already acknowledged that a collection of data would be extremely valuable to come to a more informed design and
improved asssment of comparability studies. And yet, very little data on commercialized products can be found in the public domain. In an attempt to at least partially fill this gap, we have analyzed multiple batches of Aranesp, Rituxan/Mabthera and Enbrel to study the variability in the quality attributes of modern therapeutic proteins currently on the market. The data were generated as described in the Supplementary Methods , using the materials listed in Supplementary Tables 1 and 2.The active pharmaceutical ingredient of Aranesp, darbepoetin alfa, is an erythropoiesis-stimulating
protein. It
reprents an engineered analog of human erythropoietin. It differs from endogenous erythropoietin mainly by an alteration of the amino acid quence that introduces two additional N -glycosylation sites, which results in an elongated half-life in vivo . The biological activity and clinical effect of erythropoietins is influenced by the glycosylation profile, which needs to be tightly controlled during production 2.
We have characterized commercial
batches sourced in the European Union (EU) by capillary zone electrophoresis, which parates isoforms with different charges resulting from varying numbers of sialic acids per molecule (Fig. 1 and Supplementary Fig. 1). The in vivo biological activity is known to be dependent on the number of sialic acid units per molecule, which is a result of the available sialylation sites, the antennarity of the N -glycans and the completeness of sialylation 3.
The expiry dates of the tested batches span a range from November 2008 to April 2011. We found a change of the isoform
c o r r e s p o n
d
e n c e
© 2011 N a t u r e A m e r i c a , I n c . A l l r i g h t s r e s e r v e d
.
nature biotechnology volume 29 number 4 april 2011 311
our data reflect this published process change.Rituxan/Mabthera contains a chimeric IgG1 monoclonal antibody against the B-cell surface antigen CD20 (rituximab) as the active ingredient. It is mainly ud in the treatment of B-cell malignancies and rheumatoid arthritis. The glycosylation-dependent Fc effector functions are esntial contributors to its therapeutic mode of action and clinical efficacy 5,6.
We have characterized commercial batches of Rituxan/Mabthera with expiry dates from September 2007 to October 2011 using glycan mapping, cation exchange chromatography (CEX) and antibody-dependent cellular cytotoxicity (ADCC) in vitro bioactivity (Fig. 2). In 2008, an abrupt change in the quality profile became apparent for batches with expiry dates in 2010 or later. The most obvious difference was found in the amount of the C-terminal lysine and N-terminal glutamine variants when analyzed by cation exchange chromatography (Fig. 2a ,b and Supplementary Fig. 2a ,c ). The variants elute as basic variants after the main peak. The post-change batches contain a much smaller amount of the basic variants who relative amounts are reduced from ~30–50% to ~10%. It should be noted that lysine and glutamine heterogeneity is common for monoclonal
groups showed a very high batch-to-batch consistency (Fig. 1a ). The magnitude might indicate that this change is a conquence of a manufacturing process change. Indeed, in 2008 the European Medicines Agency approved a major process change of Aranesp bad on an extensive comparability exerci 4. However, although the time frames are
matching, we have no conclusive evidence that
distribution between two ts of batches (Fig. 1a ,b ). The batches expiring up to April 2010 showed
a higher sialylation rate than the batches expiring after September 2010. The average amount of the more highly sialylated isoform number 5 decread by an average of 10%, whereas the less sialylated isoforms incread 3% (isoform number 4) and 5% (isoform number 3). However, each of the two a
10
电池保养
20
30
40
1
2
3
4
5
6
7
Isoform number
R e l. c o n t e n t (r e l. a r e a %)看的英文
18
22
26
30
34
38
42
46
t (min)
b
5
Figure 1 Comparison of the pre- and post-change Aranesp batches measured by capillary zone electrophoresis. (a ) Relative content of the individual isoforms of the pre-change (n = 18) and the post-change (n = 4) batches. (b ) Reprentative electropherograms; peaks are labeled with the isoform number.
G 0 g l y c a n (r e l . a r e a %)
Expiry date
A D
C C p o t e n c y (% o f r e f e r e n c e )
G0F-GlcNac G0 G0F Man5
GlcNAc Manno Fuco Manno
c
b
a
60
80100120140t (min)
Expiry date Figure 2 Comparison of the different pre- and post-change batches of Rituxan/Mabthera. (a ) Exemplary CEX chromatograms. (b ) Amount of basic variants of the pre-change (n = 12) and post-change (n = 6) batches as measured by CEX. (c ) ADCC potency of the pre-change (n = 11) and post-change (n = 8) batches. (d ) Relative amount of the G0 glycan of the pre-change (n = 13) and post-change (n = 11) batches. (e ) Exemplary glycan mapping chromatograms. (f ) Glycan legend.
C o R R E s p o n
D
E n C E
削弱的反义词© 2011 N a t u r e A m e r i c a , I n c . A l l r i g h t s r e s e r v e d .
312
volume 29 number 4 april 2011 nature biotechnology
COMPETING FINaNCIal INTERESTS
The authors declare competing financial interests: details accompany the full-text HTML version of the paper at /naturebiotechnology/.
Martin Schiestl, Thomas Stangler, Claudia Torella, Tadej Čepeljnik, Hansjörg Toll & Roger Grau
跳跳绳Sandoz Biopharmaceuticals, Kundl, Austria.马桶十大品牌排名
1. Chirino, A.J & Mire-sluis, A. Nat. Biotechnol. 22,
1383–1391 (2004).
2. Egrie, J.C., Dwyer, E., Browne, J.K., Hitz, A. &
Lykos, M.A. Exp. Hematol. 31, 290–299 (2003).3. Egrie, J.C. & Browne, J.K. Br. J. Cancer 84, 3–10
(2001).
4. European public Asssment Report Variation,
Aranesp-H-332-X-42 (2008) <a.europa.eu/ema/index.jsp?curl=pages/medicines/human/medicines/000332/human_med_000651.j s p &m u r l =m e n u s /m e d i c i n e s /m e d i c i n e s.jsp&mid=WC0b01ac058001d124#>
5. Glennie, M.J., French, R.R., Cragg, M.s. & Taylor R.p .
Mol. Immunol. 44, 3823–3837 (2007).
6. Reslan, L., Dalle, s. & Dumontet, C. MAbs 1, 222–229
(2009).
7. Dick, L.W., Kim, C., Qiu, D. & Cheng, K.-C. Biotechnol.
Bioeng. 97, 544–553 (2007).
8. Antes, B. et al. J. Chromatogr. B 852, 250–256
(2007).
9. shields, R.L. et al. J. Biol. Chem. 277, 26733–26740
(2002).
10. Tracey, D. et al. Pharmacol. Ther. 117, 244–279
(2008).
C-terminal heterogeneity, and the former also showed variation in ADCC activity among batches. Becau of the abruptness and the magnitude of the obrved alterations, they are most probably caud by changes in the manufacturing process. As the glycosylation profile is defined by the production cell line, growth conditions and the purification quence, the findings may reflect changes in one or more of the components. The data indicate the magnitude of changes in quality attributes of marketed products. All tested products remained on the market with unaltered labels in the tested time
frame, indicating the obrved changes were predicted to not result in an altered clinical profile and are therefore acceptable by the health authorities.
Note: Supplementary information is available on the Nature Biotechnology website.
aCkNOwlEdGEMENTS
NK3.3 cell line was generated by J. Kornbluth, St. Louis University School of Medicine, and the cell line was obtained by Novartis from St. Louis University, St. Louis. We would like to thank J. Windisch,
M. McCamish, C. Sonderegger, M. Lang-Salchner and A. Seidl from Sandoz Biopharmaceuticals for their thorough review and support.
antibodies 7,8, which most likely has no
significant impact on the biological properties of the molecule. However, our finding is an indication of an alternation in the product, which might be related to a change in the manufacturing process.
Another physicochemical difference was detected in the glycan map for unfucosylated G0 glycans (F
ig. 2d,e and Supplementary Fig. 2b ,d ). The abundance of this structure is only ~1%, but it has a substantial effect on ADCC potency, which is reflective of an esntial part of the clinical mode of action. MAbs having only unfucosylated glycans are known to exert much higher ADCC potency than the fucosylated ones 9. The abundance of unfucosylated G0 glycan in Rituxan/Mabthera incread approximately by a factor of three (Fig. 2d ) and the measured ADCC potency also showed an increa (Fig. 2c ), although by a factor less than 3, indicating that ADCC may depend on structural features in addition to the level of unfucosylated G0.
A third product, Enbrel, contains a dimeric fusion protein that binds tumor necrosis factor (TNF) alpha as the active ingredient. It compris the extracellular ligand-binding domain of the human 75 kDa (P75) tumor necrosis factor receptor (TNFR2/p75) and the Fc part of a human IgG1 antibody. The protein is glycosylated containing the IgG1-specific N -glycosylation sites in the Fc part and multiple O -glycans in the receptor
part 10. We have analyzed commercial batches sourced in the European Union and the United States using glycan mapping and CEX. The data revealed a highly consistent quality profile for batches having expiry dates until the end of 2009. After this time period, batches with a cond and changed quality profile appeared on the market in parallel (Fig. 3). Major differences were found in th
颊边痣e glycosylation profile. The amount of variants containing the N -glycan G2F decread from ~50% in the pre-change to ~30% in the post-change material (Fig. 3b ,d and
Supplementary Fig. 3b ,d ). The CEX analysis showed a change of the amount of the basic variants, which corresponds primarily to C-terminal lysine variants from 15–30% in the pre-change to 40–60% in the post-change material (Fig. 3a,c and Supplementary
Fig. 3a ,c ). As for the other products, Aranesp and Rituxan/Mabthera, the pre- and the post-change versions of Enbrel were also marketed under the same label.
In conclusion, the data we prent here reveal substantial alterations of the glycosylation profile for all tested products. Different lots of Rituxan/Mabthera and Enbrel also showed changes of the N- and
b
12.0
B a s i c v a r i a n t s (r e l . a r e a %)
爱就在我身边G 2F g l y c a n s (r e l . a r e a %)
a
t (min)Expiry date Expiry date
t (min)
8.010.018.014.016.0Figure 3 Comparison of the different pre- and post-change batches of Enbrel. (a ) Relative amounts of basic variants of the pre-change (n = 6) and the post-change (n = 6) batches as measured by CEX.
(b ) Relative amount of the G2F glycan of the pre-change (n = 25) and the post-change (n = 9) batches. (c ) Exemplary CEX chromatograms. (d ) Exemplary glycan mapping chromatograms.
C o R R E s p o n
D
E n C E
© 2011 N a t u r e A m e r i c a , I n c . A l l r i g h t s r e s e r v e d .
