Are You Measuring Developability on the Same IgG You Plan to Develop?

The screening-fidelity problem

Changing glycosylation can significantly change the developability result.

Antibody screening is about ranking candidates, not just finding binders. If expression format shifts aggregation, stability, solubility or another developability property differently across sequences, the screen can promote the wrong molecule.

High-throughput discovery can generate more antibody sequences than teams can take through conventional mammalian expression. Early screening is valuable only if the screening molecule is a faithful enough surrogate for the molecule you plan to develop.

For therapeutic IgGs, sequence is not the whole molecule. IgGs are glycoproteins. An expression system can remove N-glycosylation or install materially different glycans.

1. Developability can change even when antigen binding does not

For many antibodies, removing the conserved Fc glycan does not materially change target binding. It can still change developability.

Zheng and colleagues deglycosylated three full-length human IgG1 antibodies with PNGase F. The result: lower CH2 thermal stability, less resistance to guanidine unfolding, greater papain susceptibility and higher aggregation in accelerated stability studies [6].

Kayser and colleagues found the same direction in a full-length human IgG1: the aglycosylated form was less stable and aggregated more readily under temperature stress [7].

Raju and Scallon found faster papain degradation after Fc-glycan removal. Almost all deglycosylated Fc was degraded within four hours; more than 40% of glycosylated Fc remained after 24 hours [8].

Low-pH stress makes the aggregation difference hard to ignore

Under the same acidic aggregation conditions, Latypov and colleagues measured 8% initial monomer loss for glycosylated IgG1 Fc, 31% for partially deglycosylated Fc and 63% for fully deglycosylated Fc [9].

This was isolated Fc, not intact IgG, so the percentages are not universal full-length-antibody values. The result still shows a large glycosylation-dependent shift in stress aggregation.

And sometimes routine assays look fine - until you apply the right stress

Hristodorov and colleagues compared six full-length IgG1s with matched N297A variants. Routine solubility, heterogeneity and antigen affinity were similar, but the aglycosylated antibodies had less stable CH2 domains and significantly greater pH-induced aggregation [10].

2. The exact glycan structure can matter - not just whether a glycan is present

“Glycosylated” is not a single molecular state. Glycan structure can change physical behavior.

More and colleagues compared four defined IgG1 Fc glycoforms. As the N297 glycan shortened from high mannose to Man5 to GlcNAc to aglycosylated N297Q, apparent solubility by PEG precipitation and conformational stability declined. High-mannose and Man5 Fc were most stable; N297Q-Fc was least stable [11].

On the same Fc series, hydrogen-exchange MS showed glycan-dependent backbone flexibility, including two potential aggregation-prone regions [12]. A companion study found glycan-dependent chemical degradation, including Asn315 deamidation and Trp277 oxidation during prolonged thermal stress [13].

Wada and colleagues showed the same principle in full-length IgG1 with homogeneous Fc glycoforms: glycan structure affected thermal stability and aggregation propensity as well as Fc-receptor and complement-related functions [14].

3. Fab glycans can create sequence-dependent developability differences

Variable-region glycans occur in a meaningful fraction of human IgGs: approximately 15–20% of normal polyclonal human IgG molecules have been reported to carry them [1].

These sites can arise through somatic hypermutation [2]. In affinity-maturation or computational libraries, a sequence change can create a potential N-X-S/T glycosylation motif (X ≠ Pro). If occupied in mammalian expression, the mutation changes more than sequence.

van de Bovenkamp and colleagues found that introducing Fab glycans improved stability in some engineered adalimumab variants. Removing naturally acquired Fab glycans lowered melting temperature in three of five antibody clones [15].

Nakamura and colleagues engineered a glycosylated adalimumab Fab that retained binding and thermal stability, resisted protease digestion and showed markedly less pH-shift aggregation/turbidity than wild-type Fab [16].

For a sequence screen, the concern is straightforward: an aglycosylated format can erase a real sequence-dependent developability feature.

4. Glycosylation can also change antigen binding - sometimes by a lot

Developability is the central issue, but glycosylation can also significantly alter antigen affinity and binding kinetics.

The anti-dextran antibody TKC3.2.2 carries a heavy-chain variable-region glycan. Removing it produced an apparent association constant about 15-fold lower than the glycosylated antibody [3].

Moving the glycan within the same antibody changed the result again: Asn58 glycosylation increased dextran affinity 10- to 50-fold, Asn60 about threefold, while Asn54 glycosylation blocked detectable binding [4].

The anti-CD33 antibody M195 shows the opposite direction. Removing a heavy-chain variable-region glycosylation site increased CD33 affinity several-fold; restoring the site lowered affinity again [5].

In adalimumab, an anti-TNFα therapeutic antibody, engineered Fab glycans produced position-dependent changes in TNFα binding. Glycosylation-suppression experiments showed that the carbohydrate itself contributed to the effect for multiple variants [2].

5. Tobacco expression can give you a different IgG molecule

Tobacco expression can produce an IgG with a substantially different glycan composition from the intended mammalian product.

For Guy's 13, tobacco expression produced more glycoforms than the corresponding murine IgG1, and about 60% of its N-linked oligosaccharides contained β1,2-xylose and core α1,3-fucose [17].

A separate antibody produced in tobacco BY-2 suspension cells likewise carried major glycan species containing β1,2-xylose and α1,3-fucose [18].

The issue is not simply whether an antibody is “glycosylated.” Same sequence, chemically different glycoprotein.

And production-system differences have shown up in actual antibody readouts

b12 is a broadly neutralizing anti-HIV-1 antibody targeting the gp120 CD4-binding-site region. Rosenberg and colleagues compared plant- and CHO-produced b12 by SPR against soluble BaL gp120. Association overlapped; dissociation differed, with plant b12 showing a more pronounced fast-dissociating component [20].

The same study measured TZM-bl pseudovirus neutralization IC50. Plant-derived b12 was 4- to 10-fold more potent than CHO-derived b12 against several HIV isolates [20].

Within plant-produced b12, the non-KDEL glycoform also showed reproducibly higher neutralization activity than the KDEL form [20].

Not every antibody shifts - which makes prediction harder, not easier

In the same study, plant-derived VRC01 and 2G12 had binding kinetics essentially indistinguishable from their mammalian counterparts, while b12 behaved differently [20].

Teh and colleagues likewise found that tobacco- and HEK-produced VRC01 carried markedly different glycan distributions, including at a variable-light-chain site, yet showed no significant difference in gp120 binding affinity [19].

That is not reassurance. It means expression-system sensitivity is antibody-dependent - and unknown before screening.

6. Fc-dependent readouts are even more glycan-sensitive

For Fc biology, the relationship is direct. Removing the N297 glycan perturbed the Fc C′E loop, whose conformation correlated with FcγRIIIa affinity; the glycan helps preorganize the receptor-binding interface [21].

With defined IgG1 Fc glycoforms, FcγRIIIa binding was strong for high-mannose and Man5 Fc, weak for GlcNAc-Fc and not measurable for aglycosylated N297Q-Fc at the highest concentration tested [22].

Even one glycan feature can matter. Fucose-deficient human IgG1 showed up to a 50-fold increase in FcγRIIIa binding and enhanced ADCC [23].

What this means for developability screening

Suppose you rank hundreds of closely related IgGs. Candidate A wins thermal stability, B apparent solubility, C off-rate, and D carries a mutation that creates an occupied Fab glycosylation site.

In an aglycosylated screen, some sequence-dependent phenotypes disappear. With a substantially different glycan profile, you are measuring a chemically different version of the same sequence.

One benchmark cannot establish fidelity for a library. Preserved binding or stability in one antibody does not prove that unrelated sequences, affinity-maturation variants or glycosylation-site mutants will preserve rank order.

Developability is an ROI decision

Once a lead advances, every downstream dollar compounds the cost of an early ranking error. Years of optimization, manufacturing, toxicology and clinical development can be built around that molecule.

The practical conclusion

Aglycosylated and alternative-host IgGs can be useful. But if you are choosing which therapeutic IgG has the best developability profile, do not assume glycosylation is irrelevant. The data show it can significantly alter the measurements used to make that decision.

If the intended molecule is a mammalian glycosylated IgG, screening in a mammalian glycosylated format removes a major avoidable uncertainty.

Primary references

All references below are primary research papers. Full-length IgG and isolated-Fc results are distinguished in the text.

1. Mimura Y, Ashton PR, Takahashi N, Harvey DJ, Jefferis R. Contrasting glycosylation profiles between Fab and Fc of a human IgG protein studied by electrospray ionization mass spectrometry. J Immunol Methods. 2007;326(1–2):116–126. PMID 17714731. DOI 10.1016/j.jim.2007.07.014. PubMed
2. van de Bovenkamp FS, Derksen NIL, Ooijevaar-de Heer P, et al. Adaptive antibody diversification through N-linked glycosylation of the immunoglobulin variable region. Proc Natl Acad Sci U S A. 2018;115(8):1901–1906. PMID 29432186. DOI 10.1073/pnas.1711720115. Full text
3. Wallick SC, Kabat EA, Morrison SL. Glycosylation of a VH residue of a monoclonal antibody against alpha(1→6) dextran increases its affinity for antigen. J Exp Med. 1988;168(3):1099–1109. PMID 2459288. DOI 10.1084/jem.168.3.1099. Full text
4. Wright A, Tao MH, Kabat EA, Morrison SL. Antibody variable region glycosylation: position effects on antigen binding and carbohydrate structure. EMBO J. 1991;10(10):2717–2723. PMID 1717254. DOI 10.1002/j.1460-2075.1991.tb07819.x. Full text
5. Co MS, Scheinberg DA, Avdalovic NM, et al. Genetically engineered deglycosylation of the variable domain increases the affinity of an anti-CD33 monoclonal antibody. Mol Immunol. 1993;30(15):1361–1367. PMID 8232322. DOI 10.1016/0161-5890(93)90097-U. PubMed
6. Zheng K, Bantog C, Bayer R. The impact of glycosylation on monoclonal antibody conformation and stability. mAbs. 2011;3(6):568–576. PMID 22123061. DOI 10.4161/mabs.3.6.17922. Full text
7. Kayser V, Chennamsetty N, Voynov V, Forrer K, Helk B, Trout BL. Glycosylation influences on the aggregation propensity of therapeutic monoclonal antibodies. Biotechnol J. 2011;6(1):38–44. PMID 20949542. DOI 10.1002/biot.201000091. PubMed
8. Raju TS, Scallon BJ. Glycosylation in the Fc domain of IgG increases resistance to proteolytic cleavage by papain. Biochem Biophys Res Commun. 2006;341(3):797–803. PMID 16442075. DOI 10.1016/j.bbrc.2006.01.030. PubMed
9. Latypov RF, Hogan S, Lau H, Gadgil H, Liu D. Elucidation of acid-induced unfolding and aggregation of human immunoglobulin IgG1 and IgG2 Fc. J Biol Chem. 2012;287(2):1381–1396. PMID 22084250. DOI 10.1074/jbc.M111.297697. Full text
10. Hristodorov D, Fischer R, Joerissen H, Müller-Tiemann B, Apeler H, Linden L. Generation and comparative characterization of glycosylated and aglycosylated human IgG1 antibodies. Mol Biotechnol. 2013;53(3):326–335. PMID 22427250. DOI 10.1007/s12033-012-9531-x. PubMed
11. More AS, Toprani VM, Okbazghi SZ, et al. Correlating the Impact of Well-Defined Oligosaccharide Structures on Physical Stability Profiles of IgG1-Fc Glycoforms. J Pharm Sci. 2016;105(2):588–601. PMID 26869421. DOI 10.1016/j.xphs.2015.10.014. PubMed
12. More AS, Toth RT IV, Okbazghi SZ, et al. Impact of Glycosylation on the Local Backbone Flexibility of Well-Defined IgG1-Fc Glycoforms Using Hydrogen Exchange-Mass Spectrometry. J Pharm Sci. 2018;107(9):2315–2324. PMID 29751008. DOI 10.1016/j.xphs.2018.04.026. Full text
13. Mozziconacci O, Okbazghi S, More AS, Volkin DB, Tolbert T, Schöneich C. Comparative Evaluation of the Chemical Stability of 4 Well-Defined Immunoglobulin G1-Fc Glycoforms. J Pharm Sci. 2016;105(2):575–587. PMID 26869420. DOI 10.1016/j.xphs.2015.10.024. Full text
14. Wada R, Matsui M, Kawasaki N. Influence of N-glycosylation on effector functions and thermal stability of glycoengineered IgG1 monoclonal antibody with homogeneous glycoforms. mAbs. 2019;11(2):350–372. PMID 30466347. DOI 10.1080/19420862.2018.1551044. Full text
15. van de Bovenkamp FS, Derksen NIL, van Breemen MJ, et al. Variable Domain N-Linked Glycans Acquired During Antigen-Specific Immune Responses Can Contribute to Immunoglobulin G Antibody Stability. Front Immunol. 2018;9:740. PMID 29706962. DOI 10.3389/fimmu.2018.00740. Full text
16. Nakamura H, Oda-Ueda N, Ueda T, Ohkuri T. Introduction of a glycosylation site in the constant region decreases the aggregation of adalimumab Fab. Biochem Biophys Res Commun. 2018;503(2):752–756. PMID 29909010. DOI 10.1016/j.bbrc.2018.06.071. PubMed
17. Cabanes-Macheteau M, Fitchette-Lainé AC, Loutelier-Bourhis C, et al. N-Glycosylation of a mouse IgG expressed in transgenic tobacco plants. Glycobiology. 1999;9(4):365–372. PMID 10089210. DOI 10.1093/glycob/9.4.365. PubMed
18. Fujiyama K, Misaki R, Katsura A, Tanaka T, Furukawa A, Omasa T, Seki T. N-linked glycan structures of a mouse monoclonal antibody produced from tobacco BY2 suspension-cultured cells. J Biosci Bioeng. 2006;101(3):212–218. PMID 16716920. DOI 10.1263/jbb.101.212. PubMed
19. Teh AYH, Maresch D, Klein K, Ma JKC. Characterization of VRC01, a potent and broadly neutralizing anti-HIV mAb, produced in transiently and stably transformed tobacco. Plant Biotechnol J. 2014;12(3):300–311. PMID 24256218. DOI 10.1111/pbi.12137. Full text
20. Rosenberg Y, Sack M, Montefiori D, et al. Rapid high-level production of functional HIV broadly neutralizing monoclonal antibodies in transient plant expression systems. PLoS One. 2013;8(3):e58724. PMID 23533588. DOI 10.1371/journal.pone.0058724. Full text
21. Subedi GP, Barb AW. The Structural Role of Antibody N-Glycosylation in Receptor Interactions. Structure. 2015;23(9):1573–1583. PMID 26211613. DOI 10.1016/j.str.2015.06.015. Full text
22. Okbazghi SZ, More AS, White DR, et al. Production, Characterization, and Biological Evaluation of Well-Defined IgG1 Fc Glycoforms as a Model System for Biosimilarity Analysis. J Pharm Sci. 2016;105(2):559–574. PMID 26869419. DOI 10.1016/j.xphs.2015.11.003. Full text
23. Shields RL, Lai J, Keck R, et al. Lack of fucose on human IgG1 N-linked oligosaccharide improves binding to human FcγRIII and antibody-dependent cellular toxicity. J Biol Chem. 2002;277(30):26733–26740. PMID 11986321. DOI 10.1074/jbc.M202069200. PubMed