key: cord-0292624-de9k5xpf
authors: Bharadwaj, P.; Shrestha, S.; Pongracz, T.; Concetta, C.; Sharma, S.; Lemoine, A.; de Haan, N.; Murakami, N.; Riella, L. V.; Holovska, V.; Wuhrer, M.; Marchant, A.; Ackerman, M. E.
title: Afucosylation of HLA-specific IgG1 as a potential predictor of antibody pathogenicity in kidney transplantation
date: 2022-03-12
journal: nan
DOI: 10.1101/2022.03.09.22272152
sha: 0af8ae1832dd3d3333ffd0e26514e8e3e5e8ed7d
doc_id: 292624
cord_uid: de9k5xpf

Antibody-mediated rejection (AMR) is the leading cause of graft failure. While donor-specific antibodies (DSA) are associated with a higher risk of AMR, not all patients with DSA develop rejection suggesting that the characteristics of alloantibodies that determine their pathogenicity remain undefined. Using human leukocyte antigen (HLA)-A2-specific antibodies as a model, we applied systems serology tools to investigate qualitative features of immunoglobulin G (IgG) alloantibodies including Fc-glycosylation patterns and Fc{gamma}R binding properties. The levels of afucosylation of anti-A2 antibodies were elevated in all seropositive patients and were significantly higher in AMR patients, suggesting potential cytotoxicity via Fc{gamma}RIII-mediated mechanisms. Afucosylation of both glycoengineered monoclonal and naturally glycovariant polyclonal serum IgG specific to HLA-A2 exhibited potentiated binding to, slower dissociation from, and enhanced signaling through Fc{gamma}RIII, a receptor widely expressed on innate effector cells. Collectively, these results suggest that afucosylated DSA may be a biomarker of AMR and could contribute to its pathogenesis.

with 50 μL of sera for 3 hours on a rotational mixer, followed by three washes using PBS-TBN (PBS-156 1X, 0.1% BSA, 0.02% Tween 20, 0.05% sodium azide, pH 7.4). HLA-A2-specific antibodies were LC-MS based Fc glycosylation analysis and data processing quadrupole time-of-flight mass spectrometer (Bruker Daltonics, Billerica, MA), as described (72).

The FcγR binding profiles of polyclonal HLA-A2-specific antibodies were defined using the Fc Array 176 assay described previously (74). Briefly, HLA-A2-coated microspheres were generated as described above for the multiplex assay, but serum antibodies were detected with FcγRIIIa tetramers formed Octet (Forte Bio) instrument. Streptavidin (SA) biosensor tips were first loaded with HLA-A2 antigen, 207 followed by HLA-A2 antibody, and finally dipped into FcγR to study the antibody-receptor kinetics.

Biosensors were equilibrated with 0.05% Tween-PBS for 60s, loaded with 0.25 µg/mL biotinylated 209 HLA-A2 antigen (in 0.05% Tween-PBS) until they reached the threshold of 1nm, and then dipped 210 into 0.05% Tween-PBS to reach baseline for 60s. After that, the biosensors were loaded with 3µg/mL 211 of recombinant HLA-A2 IgG (in 0.05% Tween-PBS) for 600s, dipped into 0.05% Tween-PBS to reach 212 baseline for 60s, and then dipped into 5µM monomeric FcγRIIIa for 180s or 292 nM FcγRI for 300s 213 for association. To improve signal to noise, serum IgG was purified using Melon gel (Thermo) according to the manufacturer's instructions. Polyclonal serum IgG was loaded for 2400s. For FcγR and one as a reference to subtract background signal changes due to dissociation of other 217 components, such as dissociation of HLA-A2-specific IgG from biotinylated HLA. Reference 218 biosensors loaded with IgGs were dipped into buffer (0.05% Tween-PBS) instead of receptors.

Finally, the sensors were dipped into 0.05% Tween-PBS for 180s and 300s for the dissociation of CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) Table 1 ). A2 sensitization was related to previous or current 232 transplantation in 13 patients (41%), pregnancy in 6, blood transfusion in 1 and left ventricular 233 assistance device in 1, whereas the sensitizing event was unknown in 4 patients. AMR was diagnosed 234 in 13/28 (46%) transplanted patients (Supplemental Table 1 ) and in 7/13 (54%) DSA-positive 235 patients (A2 sensitization in a patient transplanted with an HLA-A2 graft) (Supplemental Table 2 ).

Among controls, 9 patients had no detectable anti-HLA antibodies and 9 had detectable antibodies 237 against non-HLA-A2 antigens. Seventeen controls had received an organ transplant and one was on 238 a transplant waiting list.

IgG subclass distribution of HLA-specific antibodies

We first sought to characterize the subclasses of HLA-specific IgGs in HLA-A2 sensitized and control 242 subjects, using research grade single-antigen bead immunoassays. Briefly, single antigens (intact 243 HLA-A02:01 or HLA-A01:01 monomers) were captured on microspheres, and antigen binding of 244 sera samples were measured by multiplex assay. In contrast to the HIV-specific antibody VRC01 245 (negative control), the HLA-A2-specific monoclonal antibody BB7.2 (positive control), which was 246 evaluated in each human IgG subclass, showed specific binding to HLA-A2 ( Figure 1A) , and 247 appropriate staining by IgG subclass detection reagents (Supplemental Figure 1) . In contrast to the 248 negative control mAb, both pooled human serum immunoglobulin (IVIG) and serum from individual 249 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted March 12, 2022. ; https://doi.org/10.1101/2022.03.09.22272152 doi: medRxiv preprint control subjects often showed signal considerably above background. While this profile may be 250 consistent with the widespread prevalence of anti-HLA antibodies among healthy individuals (78, 251 79), and responses among subjects clinically defined as HLA-A2 seropositive were significantly 252 elevated as compared to controls, there was a considerable overlap in distributions. Nonetheless,

there was a distinct difference observed in the subclass distribution between the two groups, with 254 the HLA-A2 seropositive group generally exhibiting a wide range of total HLA-A2 antibodies as 255 compared to the control samples, which tended to show lower levels ( Figure 1A ). The wide 256 distribution of HLA-A2 specific IgG signal was consistent with reported values from clinical testing 257 (Supplemental Figure 2) , with some samples exhibiting moderate to very low HLA-A2 reactivity.

HLA-A2-reactive antibodies predominantly belonged to the IgG1 and IgG2 subclasses, but a few 259 subjects also exhibited higher levels of IgG3 and IgG4; these individuals typically had high levels of 260 total HLA-A2-specific IgG.

A similar analysis of HLA-A1-specific antibody levels was performed. The HLA-A2 seropositive group 263 also showed elevated levels of HLA-A1 antibodies as compared to the control group ( Figure 1B 

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted March 12, 2022. sialyation ( Figure 2D ) between HLA-A2-specific and total serum IgG1. Of note, while HLA-A2-284 specific antibody Fc fucosylation of some individuals was comparable to that of total IgG1, others 285 exhibited fucosylation levels reduced by 5-15%, or even 30% as compared to their total IgG1 286 counterpart ( Figure 2D ). HLA-A2 specific IgG2/3 was characterized by higher bisection, 287 galactosylation and sialylation than total serum IgG2/3, but no afucosylated IgG2/3 glycopeptides 288 were detected ( Figure 2D ). Consistent with multiplex assay data, antigen-specific IgG4 responses 289 were generally too low to be robustly characterized by LC-MS. While IgG2 and IgG3 glycopeptides 290 could not be distinguished by mass spectrometry due to their identical molecular mass, the high 291 levels of antigen-specific IgG2 and low levels of IgG3 observed suggest that the antigen-specific To address the mechanism of improved FcγRIIIa signaling activity, we next sought to characterize 303 the interaction between HLA-A2-specific antibodies and FcγRIIIa using Biolayer Interferometry (BLI) 304 ( Figure 3B , Supplemental Figure 6 ), a label free technique commonly used to study the affinity and 305 kinetics of protein-protein interactions. Antibodies bound to HLA-A2 antigen were allowed to bind 306 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 12, 2022. ; https://doi.org/10.1101/2022.03.09.22272152 doi: medRxiv preprint rates (mean kd = 0.021 s -1 ) as compared to unmodified forms (mean kd = 0.46 s -1 ). This difference in 311 dissociation rate was specific for FcγRIIIa, as fucosylation did not influence FcγRI binding 312 (Supplemental Figure 7) , as expected (63).

Serum HLA-A2 specific antibody fucosylation associates with FcγRIIIa dissociation and ligation

Since afucosylated HLA-A2-specific mAbs had slower dissociation from FcγRIIIa as compared to their 315 unmodified counterparts, we next studied the dissociation profiles of polyclonal HLA-A2 specific 316 antibodies in human serum and found FcγRIIIa dissociation rate to be positively associated with 317 fucosylation ( Figure 4A ). The variable off-rates observed among differentially fucosylated serum-318 derived HLA-A2-specific antibodies suggested that FcγRIIIa ligation and signaling profiles of these 319 HLA-A2 specific antibodies might also vary. To test this possibility, HLA-A2 seropositive subjects 320 were split into tertiles (high, medium, and low) based on their HLA-A2-specific antibody fucosylation 321 levels ( Figure 4B ). A multiplex assay was conducted, in which HLA-A2-specific antibodies were 322 evaluated for their ability to bind FcγRIIIa tetramers (74). Although there was no significant 323 difference in FcγRIIIa binding between high, medium, and low fucose tertiles, HLA-A2 specific 324 antibodies in each of these three categories had higher FcγRIIIa binding as compared to negative 325 controls, as exemplified by the observed negative correlations ( Figure 4B ). In line with observations 326 on variably fucosylated mAbs, these results demonstrate that human serum-derived HLA-A2-specific 327 antibodies with low fucosylation exhibit improved FcγRIIIa binding. 

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) 

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Buffer only blank (cross) and pooled IVIG (diamond) are shown in brown and purple, respectively. Serum samples were tested at a 1:100 dilution. Data shown are representative of two technical replicates. Solid red lines indicate group median. Differences between groups were evaluated using Ordinary Two-way ANOVA adjusted for multiple comparisons using Bonferroni's test (**p<0.01, ***p<0.001, ****p <0.0001, respectively). . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 12, 2022. ; https://doi.org/10.1101/2022.03.09.22272152 doi: medRxiv preprint Figure 2 . Fc glycosylation of HLA-A2-specific antibodies. Representative extracted ion chromatograms and mass spectra (A, B) illustrating the observed variability between HLA-A2specific (A) and total (B) IgG1 Fc glycosylation patterns of the same patient. C. Volcano plot displaying the log10 fold change (x-axis) and -log10 p-value (y-axis) of individual IgG1 (blue circle) and IgG2/3 (maroon square) glycoforms between total and HLA-A2-specific IgG1 and IgG2/3, respectively. D. Violin plots showing the relative prevalence of glycans on bulk and HLA-A2 specific IgG1 (top) and IgG2/3 (bottom), respectively. Statistical analysis was performed using a paired two-tailed Student's t-test (**: p<0.01, ***: p<0.001, ****: p<0.0001, respectively).

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted March 12, 2022. ; https://doi.org/10.1101/2022.03.09.22272152 doi: medRxiv preprint Figure 4 . Associations of serum derived HLA-A2 specific antibody fucosylation with FcγRIIIa binding and signaling. A. Spearman's correlation (RS) between IgG1 fucosylation and FcγRIIIa dissociation rate (n=13). B. FcγRIIIa binding characterization in high (n=8), medium (n=8), low (n=8) fucose samples and negative controls (n=18). Serum samples were tested at a 1:500 serum dilution. Statistical analysis was performed using Ordinary one-way ANOVA adjusted for multiple comparisons using Tukey's test. Solid lines indicate group median. Data shown is representative of two technical replicates. C-D. Spearman's correlations between IgG1 fucosylation (n=24) (C), and FcγRIIIa signaling (n=31) (D) with FcγRIIIa binding (left) and IgG titer (right). Patients with AMR (red), patients without AMR (green), and patients with no AMR information (black) are indicated in color. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted March 12, 2022. ; https://doi.org/10.1101/2022.03.09.22272152 doi: medRxiv preprint Figure 5 . HLA-A2-specific IgG1 afucosylation is associated with AMR. A. Violin plot showing relative prevalence of fucose on HLA-A2-specific IgG1 in individuals with AMR (n=10) and without AMR (n=9). A Mann-Whitney U test was used to compare the two groups. B. Number of subjects in which HLA-A2-specific IgG1 comprise a DSA plotted by fucose content as tertiles (low, medium, high) and AMR status. Statistical significance defined by Fisher's exact test. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted March 12, 2022. ; https://doi.org/10.1101/2022.03.09.22272152 doi: medRxiv preprint

Post-transplant development of C1q-positive HLA 672 antibodies and kidney graft survival

Functional diversification of IgGs through Fc glycosylation

Fc-galactosylation modulates 676 antibody-dependent cellular cytotoxicity of therapeutic antibodies

Impact of Glycosylation on Effector Functions of Therapeutic IgG †. 679

Terminal sugars of Fc glycans influence antibody effector functions of IgGs. Current 681 Opinion in

The Absence 683 of Fucose but Not the Presence of Galactose or Bisecting N -Acetylglucosamine of Human IgG1

Complex-type Oligosaccharides Shows the Critical Role of Enhancing Antibody-dependent 685

Decoding the FcγR Binding and ADCC Activity of Human IgG Allotypes

Multi-level glyco-699 engineering techniques to generate IgG with defined Fc-glycans

Microscale purification 702 of antigen-specific antibodies

multiplexed IgG subclassing of antigen-specific antibodies from clinical samples

RhD IgG-Fc-fucosylation in pregnancy: a new variable predicting severity in haemolytic disease 708 of the fetus and newborn

High-Throughput Analysis of IgG Fc Glycopeptides 710 by LC-MS

The N-Glycosylation 712 of Mouse Immunoglobulin G (IgG)-Fragment Crystallizable Differs Between IgG Subclasses and 713 Strains

Immunoglobulin 715 G1 Fc glycosylation as an early hallmark of severe COVID-19

Multiplexed Fc array 717 for evaluation of antigen-specific antibody effector profiles

Highly parallel 720 characterization of IgG Fc binding interactions

IgG Fucosylation Predicts Susceptibility to Dengue Disease in Infants

Distinct Features and 725 Functions of Systemic and Mucosal Humoral Immunity Among SARS-CoV-2 Convalescent 726 Individuals

Naturally occuring" anti-728 HLA antibodies

Natural" 730 human leukocyte antigen antibodies found in nonalloimmunized healthy males

Detecting and monitoring human leukocyte antigen-specific antibodies

Sialylation 743 of antibodies in kidney recipients with de novo donor specific antibody, with or without antibody 744 mediated rejection

Donor-specific HLA 746 antibodies: evaluating the risk for graft loss in renal transplant recipients with isotype switch from 747 complement fixing IgG1/IgG3 to noncomplement fixing IgG2/IgG4 anti-HLA alloantibodies

The Immunoregulatory Roles of Antibody Glycosylation

Modulating IgG effector function 752 by Fc glycan engineering

IgG subclasses and allotypes: from structure to effector 754 functions

Natural variation in Fc 756 glycosylation of HIV-specific antibodies impacts antiviral activity

Antibody 759 fucosylation predicts disease severity in secondary dengue infection

Afucosylated IgG 762 characterizes enveloped viral responses and correlates with COVID-19 severity

Immunoglobulin G glycosylation in aging and diseases

High titers and 767 low fucosylation of early human anti-SARS-CoV-2 IgG promote inflammation by alveolar 768 macrophages

A prominent 770 lack of IgG1-Fc fucosylation of platelet alloantibodies in pregnancy

IgG antibodies 772 to dengue enhanced for FcgammaRIIIA binding determine disease severity

Systems serology for evaluation of HIV vaccine trials

Potential influence of HLA-A2-specific IgG1 afucosylation, FcγRIIIa binding and activation on ADCC and graft rejection