Degree Type

Dissertation

Date of Award

2019

Degree Name

Doctor of Philosophy

Department

Biochemistry, Biophysics and Molecular Biology

Major

Biochemistry

First Advisor

Adam W. Barb

Second Advisor

Richard . Honzatko

Abstract

N-glycosylation is one of the most abundant post-translational modifications in humans. N-glycans are processed in the ER and Golgi network by enzymes yielding heterogenous populations of N-glycosylation structures in vivo. N-glycosylation is part of the protein quality control and folding pathways. N-glycans also function in antigenicity, serum half-life, proteolysis protection, protein function and much more. N-glycosylation composition differences have been linked to disease states, age, gender, and pregnancy. These functionally relevant composition differences have also been taken advantage of by the pharmaceutical industry.

Monoclonal antibody therapies (mAb) are estimated to exceed $140B in annual sales by 2024. mAb therapies function through direct binding of molecules in inflammation, neutralization or the complement cascade though the majority require Fc gamma R-mediated cell effector functions including antibody dependent cellular phagocytosis (ADCP) and antibody dependent cellular phagocytosis (ADCC). A second generation of glycoengineered monoclonal antibody (mAb) therapies have recently out-performed their counterparts in clinical studies. This efficacy is thought to be directly related to the enhanced binding of antibody for FcRIII (CD16). CD16a is expressed predominantly on NK cells and monocytes and believed to be the main receptor engaged in ADCC. The role of the IgG antibody N-glycan has been well studied by others as well as previous members of the lab. However, the role of Fc gamma R N-glycosylation including CD16a, CD16b and CD32a have not been well studied. Surface plasmon resonance (SPR) experiments designed to study the role of FcR N-glycosylation have shown less processed, oligomannose and hybrid-type N-glycans bind with higher affinity to IgG1 Fc than with complex-type N-glycans. CD16a expressed in a HEK293 Gnt1(-/-) cells give predominant Man5 N-glycosylation and bind 12-50 fold tighter than complex-type while CD16a with hybrid and oligomannose (Man9) bound 2-4 fold tighter than complex-type. Interestingly, when all Fc gamma Rs were expressed with Man5 N-glycans instead of complex-type an enhanced affinity for IgG1 was observed on one Fc gamma R, CD16a, and is specific to one of five N-glycosylations sites, the N162 N-glycan.

CD16b is a nearly identical paralogue to CD16a and differs by only four residues in the extracellular binding domains, thus it is a surprise that CD16b has a much weaker affinity for IgG1 and does not share an enhanced affinity when expressing Man5 as described with CD16a. A better understanding for this could lead to optimal mAb therapies. Interestingly only residue 129, a glycine in CD16a and aspartate in CD16b is the only residue of the four within the binding interface. CD16a G129D and CD16b D129G mutants expressing Man5 or complex-type N-glycans were subjected to SPR binding studies to IgG1 Fc. CD16b D129G bound with a 2-fold higher affinity than CD16a WT and a 90-fold higher affinity than CD16b WT. The reverse mutation in CD16a G129D, bound with a 128-fold decreased affinity to CD16a wt and comparable to CD16b WT. Furthermore, when comparing CD16 Man5 vs. complex-type N-glycans, the ‘lec effect’ was only observed in CD16a WT and CD16b D129G suggesting the interaction is more sensitive to CD16 N-glycosylation when G129 is present because when D129 was present there was a minimal difference between CD16 binding to IgG with a Man5 or complex-type N-glycan. An x-ray crystal structure of glycosylated CD16b in complex with afucosylated IgG1 Fc was determined to 2.2 Å resolution along with a 250 ns all-atom molecular dynamics simulation showed the D129 forces an unfavorable buckle in the CD16b backbone upon binding to IgG1 Fc. Furthermore, the D129 residue forces R155 to make different contacts with the N162 N-glycan. This may force the N162 N-glycan into different conformations which are less sensitive to N-glycan composition on CD16b as compared to CD16a. Overall, these results determine D129 is responsible for the lower affinity of CD16b as compared to CD16a and this finding suggests mAb therapies could design antibodies to accommodate this residue.

CD16a N-glycosylation composition affects binding to IgG1 in vitro. This finding lead our group to question whether immune response is modulated through CD16a N-glycan composition in vivo. However, N-glycosylation composition of CD16a from endogenous tissue has not been elucidated before. Our lab developed methods to isolate CD16a from NK cells and monocytes from apheresis filters after platelet and plasma donations. Immunoprecipitation of CD16a typically yielded 1 to 2 micrograms of CD16a from NK cells and 100 ng to 1 microgram from monocytes. A glycomics approach to CD16a from NK cells found higher oligomannose, less processed, material than what was found in recombinant sources. This study indicated recombinant CD16a glycoforms used by laboratories and the pharmaceutical industry do not recapitulate the material found in humans. A glycoproteomics approach was then used to identify composition of the 5 and 2 N-glycosylation sites of CD16a and CD32a from monocytes. CD16a at N162 was predominantly complex-type biantennerary, and some donors displayed up to 20% of hybrid-types. N38 and N74 displayed highly processed, complex-type biantennerary. N169 displayed predominantly complex-type biantennerary with some highly processed complex-type. CD16a N45 was predominantly under processed with mostly oligomannose and hybrid type. Interestingly, CD16a N45 from monocytes displays more oligomannose than from NK cells. CD16a from NK cells also displayed higher abundance of oligomannose and hybrid-types than CD16a from monocytes indicating higher affinity glycoforms on some NK cell donors. CD32a N64 and N145 from monocytes displayed complex-type biantennerary with two sialic acids unlike any of the CD16a N-glycoyslation sites. Overall, N-glycosylation of CD16a appears to differ by cell-type, donor and site. These methods can now be used to expand donor populations and has the potential to characterize disease-relevant N-glycosylation differences as already has been elucidated in antibodies from populations with various disease states, gender and age.

Copyright Owner

Jacob Timothy Roberts

Language

en

File Format

application/pdf

File Size

150 pages

Available for download on Tuesday, September 29, 2020

Included in

Biochemistry Commons

Share

COinS