Interrogating Post-Transplant Donor HLA-specific Antibody Characteristics and Effector Functions Using Clinical Bead Assays

Abstract

Single antigen bead (SAB) assays are the most common and sensitive method used to detect and monitor post-transplant donor specific HLA antibodies (DSA). However, a direct comparison across traditional and modified SAB assays to improve routine DSA monitoring using pre-treated IgG sera to eliminate interference has not been performed. We performed a technical comparison of 251 post-transplant DSA from n=91 serum samples tested neat (pre-treated, undiluted), at a single 1:16 dilution, in the C1q bead assay, and for IgG subclasses (IgG1, IgG2, IgG3, IgG4) with IgG-enriched sera. We found that DSAs that are detectable by 1:16 dilution and/or C1q are associated with higher IgG MFI values and results could be predicted by testing neat sera. DSA detected at 1:16 dilution correlated with >7000 IgG MFI in neat sera and identified DSA that exceeded the SAB linear range for semiquantitative measurements. C1q positive DSA correlated with >15000 IgG MFI in neat sera. C1q binding correlated most strongly with total IgG MFI (Spearman r = 0.82, p = 0.002) and not specific subclasses, demonstrating that DSA C1q binding capacity in this cohort is driven by HLA-specific IgG concentration. Evaluation of engineered pan-HLA class I-specific human IgG1 and IgG2 subclass monoclonal antibodies by SAB C1q and C3d assays revealed that IgG2 antibodies can bind complement at higher concentrations. The strengths and limitations of modified SAB assays must be considered to optimize efficient testing and accurate clinical interpretation.

Introduction

De novo donor HLA-specific antibody development is associated with antibody-mediated rejection and allograft loss in transplant recipients. Most post-transplant DSA have specificity for class II HLA, in particular HLA-DQ, which is associated with poor clinical outcomes.¹ Luminex single antigen bead (SAB) assays are the most common and sensitive method used to detect and monitor IgG DSA in transplant patients. Unlike high titer humoral immunity generated toward pathogens (e.g. Tetanus Toxoid, Cytomegalovirus, Epstein-Barr Virus), sensitization to HLA alloantigens is generally low titer and clinical testing is performed using undiluted serum. However, testing undiluted serum introduces limitations including interfering proteins that can produce false negative results or suppressed signal and HLA antibody concentrations that exceed the linear range of the assay resulting in the loss of semiquantitative measurements.^2–4 Previous studies have highlighted the importance of removing interfering factors that hinder accurate HLA antibody detection. Treatment of serum with dithiothreitol (DTT), ethylenediaminetetraacetic acid (EDTA), heat (56ºC, 30 minutes), or IgG enrichment columns prior to SAB testing have been shown to effectively reduce interference due to complement and non-IgG antibody isotypes.^5–7

Commonly, DSA results are reported using the IgG mean fluorescence intensity (MFI) from neat (treated, undiluted) sera alone, which is known to be insufficient for quantitation of DSA strength and assessment of injury risk. Despite strong correlations between DSA and adverse outcomes, DSA is often detected in the absence of allograft dysfunction or antibody mediated rejection. This suggests that additional features, such as quantity, specificity, or effector function, might modify the risk profile of DSA. These DSA characteristics can be evaluated using modified SAB assays, including serum dilutions (titer), complement binding capacity, and IgG subclass. Importantly, complement binding and IgG subclass have been shown to correlate with transplant outcomes better than SAB neat MFI alone^8–14, suggesting that utilization of more in-depth DSA characterization may provide valuable information for improved immune monitoring post-transplantation.

Experiments concentrating sera containing weak HLA antibodies or diluting sera with strong HLA antibodies have shown a direct correlation between antibody concentration and positivity in complement binding assays.¹⁵ Evaluation of post-transplant DSA by dilution (titer) and ability to bind complement protein (C1q/C3d) have demonstrated strong correlations with each other and provide better quantitation of DSA that exceeds the linear range of SAB assays using undiluted serum.^4,16–18 These distinct HLA antibody test modifications do provide additional information for assessing DSA strength and injury potential beyond MFI alone yet increase cost and time-to-report for DSA monitoring. A direct comparison across all HLA bead assays to characterize post-transplant IgG DSA has not been performed. Thus, in this study we evaluated 251 post-lung transplant DSA detected using traditional and modified SAB assays to compare results across neat (treated, undiluted) serum, a single 1:16 dilution, the C1q bead assay, and IgG subclasses (IgG1, IgG2, IgG3, IgG4) using sera enriched for IgG using spin columns that remove all protein interference, including complement and non-IgG antibody isotypes. While the removal of non-IgG isotypes from sera was critical to increase sensitivity in the IgG subclass analysis, we also compare DSA results using sera tested undiluted and 1:16 that was pre-treated with EDTA to mitigate complement interference. We report a direct technical comparison across these assays to improve our understanding of DSA IgG characteristics, determine the added value and correlations between the modified assays, and improve efficiency and cost by predicting whether new DSA information is likely to emerge from additional testing.

Methods

Cohort Selection and Approval

Post-transplant HLA class II DSA positive (DSA+) lung transplant recipient sera (n=91) from the Clinical Trials in Organ Transplantation (CTOT)-20 Multicenter Consortium Study (registration number: NCT02631720) were utilized in this study and retested in a centralized laboratory. Pre-transplant DSA was determined absent by the originating HLA laboratory; all DSA were detected de novo, post-transplantation. Use of human samples was approved by the Duke University Institutional Review Board (Protocol 00105330).

HLA class I and class II DSA Single Antigen Bead (SAB) Assay and Analysis

DSA (n=251, detected postoperative day: Range: 7 – 1184. Class I DSA n = 63, mean day: 116, median day: 34. Class II n=188, mean day: 169, median day: 58) were measured by SAB Luminex assays: total IgG (undiluted and 1:16 dilution), C1q, and IgG subclass (IgG1, IgG2, IgG3, and IgG4) (LabScreen and C1qScreen, One Lambda). All sera were pretreated using an IgG enrichment column (Zeba & Melon Column, ThermoFisher) to remove all interfering proteins and acquired on a Luminex 200 instrument (Luminex). Total IgG (undiluted and 1:16 dilution) testing was performed using a modified version of the rapid optimized protocol.¹⁹ Briefly, wells were wet with 1X wash buffer for 10 minutes prior to beginning the assay. Wash buffer was aspirated using a vacuum manifold and 2.5μl of appropriate beads (Class I or Class II) were added to each well. Beads were washed once and 25μl of test serum (undiluted or 1:16 dilution) was added for 15 minutes at room temperature (RT) on a shaker, protected from light. Plates were washed 5x and 20μl of 1:10 PE-conjugated anti-human IgG (One Lambda) was added to each well for 5 minutes at RT on a shaker, protected from light. Plates were washed 5x and samples were resuspended in 60μl of wash buffer prior to acquisition. The C1q assay was also performed using a rapid optimized protocol from the laboratory of Robert Liwski.²⁰ The C1q assay follows a similar protocol, using 2.5μl HLA class I or class II beads with 0.25μl C1q positive control beads per well. Beads were washed and 10μl of serum was added to each well for 20 minutes at RT on a shaker, protected from light. Plates were washed 2x and 10μl of C1q/anti-C1q-PE mixture was added to each well (mixture consists of 2μl working solution of C1q protein (diluted 1:5 per One Lambda package insert), 1.5μl anti-C1q-PE detection antibody, 6.5μl wash buffer). Plates were incubated for 20 minutes, washed 3x, and samples were resuspended in 60μl of wash buffer prior to acquisition on a Luminex 200 instrument (Luminex).

A retrospective comparison was performed using clinical testing data in which sera was pre-treated with EDTA (final concentration 5 mM) and tested undiluted and at a 1:16 dilution. This testing followed a modified manufacturers protocol; briefly, 4μl of appropriate beads (Class I or Class II) and 20μl of test serum (neat undiluted or 1:16 dilution) was added for 30 minutes at room temperature (RT) on a shaker, followed by 4x washes and addition of 100ul of 1:100 PE-conjugated anti-human IgG (One Lambda) to each well for 30 minutes at RT, followed by 2x washes and acquisition on a Luminex 200 instrument (Luminex).

For the IgG subclass assay, 12-point standard curves were created by coupling Goat anti-human IgG (Kappa light chain) protein (Invitrogen) to polystyrene Luminex beads via amine coupling with EDC (Sigma) and Sulfo-NHS (Sigma). IgG subclass proteins of known concentration (IgG1, IgG2, IgG3, IgG4) isolated from human myeloma plasma (Sigma) were serially diluted and incubated with the anti-IgG-Kappa-coated beads. Simultaneously, 25ul of patient serum and controls were incubated with 2.5μl/well of HLA-coated beads (One Lambda). All samples were then washed 3x and incubated with IgG subclass-specific biotinylated secondary antibodies (Southern Biotech, 5 ug/ml) for 30 minutes on a shaker at RT. Plates were washed 3x and Streptavidin-PE tertiary detection reagent (BD Biosciences, 5 ug/ml) was added. Plates were incubated at RT for 30 minutes on a shaker, washed 3x, and all samples were resuspended in 60μl of wash buffer prior to acquisition on a Luminex 200 instrument (Luminex). Antibody clones and IgG subclass cross-reactivity data are included in Table 1.

Table 1.

IgG subclass detection reagent cross-reactivity test for IgG subclass proteins isolated from human myeloma plasma. Data is shown for each protein plated at 20 ug/ml. Full titration curves are provided in Supplementary Figure 5.

Monoclonal antibodies		IgG subclass protein
Clone	Target	IgG1	IgG2	IgG3	IgG4
4E3	IgG1 Hinge	18005	555 (3%)	152 (<1%)	73 (<1%)
HP6002	IgG2 Fc	526 (10%)	4829	189 (4%)	51 (1%)
HP6050	IgG3 Hinge	718 (4%)	618 (3%)	18719	44 (<1%)
HP6023	IgG4 pFc’	93 (<1%)	561 (4%)	193 (2%)	15311

Engineered Chimeric Monoclonal Antibody Analyses

Pan-HLA class I-specific human IgG monoclonal antibodies were engineered by cloning murine H+L variable regions from W6/32 onto human IGHG1 and IGHG2 constant regions as we previously reported.²¹ The mAbs were produced in CHO cell lines. Monoclonal antibodies were diluted into normal human AB serum (Complement Technologies, Tyler, TX) that was negative for HLA antibodies by prior single antigen testing. Diluent sera were heat-inactivated (30min at 56°C). Complement binding capacity of each IgG subclass antibody was then evaluated using C1qScreen (OneLambda) and C3d Assay (Immucor, Stamford, CT) as we reported.²² The Luminex assays to detect complement binding induced by chimeric HLA class I hIgG were performed according to manufacturers’ protocols. Briefly, For the C1qScreen Assay (OneLambda), chimeric monoclonals diluted in heat-inactivated serum was incubated with LabScreen single antigen class I beads in the presence of C1q, and after washing, C1q binding was detected with a C1q-PE secondary antibody, and acquired on a Luminex 200 instrument (Luminex, Austin, TX). For the C3d assay, sera with monoclonal antibodies were incubated with LifeCodes LSA class I single antigen beads for 30min at room temperature, followed by addition of 37.5% normal human serum for an addition 30min at room temperature. Then, plates were washed 5 times, then incubated with anti-C3d-PE for 30min at room temperature. After washing twice, samples were resuspended and acquired on Luminex100 instrument (Luminex, Austin, TX).

Data Analysis and Visualization

DSA positivity was determined using full 12-loci (HLA A, B, C, DRB1/3/4/5, DQA1/DQB1, DPA1/DPB1) recipient and donor HLA typing and HLA pattern analysis in HLA Fusion software. Given differences in detection sensitivity across assays, negative detection cutoffs were established for each DSA based on the MFI of the patient’s autologous HLA bead at the corresponding HLA locus. GraphPad Prism (version 9.4.1) and R were used for analysis and figure generation.

Results

DSA specificity

A total of 251 DSA with specificity for HLA class I (n=63) and class II (n=188) antigens were detected in this cohort. Of class I DSA, 35% (22/63) were directed against HLA-A, 41% (26/63) against HLA-B, and 24% (15/63) against HLA-C (Supplementary Figure 2A). Of Class II DSA, 35% (66/188) were directed against HLA-DR, 57% (107/188) against HLA-DQ, and 8% (15/188) against HLA-DP (Figure 1A). HLA-DQ-specific DSA were the most abundant and had the highest MFI detected by neat serum (HLA-DQ median MFI: 10,792), relative to all other class I (Median MFI for HLA-A: 5471, HLA-B: 4609, HLA-C: 3002; Supplementary Figure 2B) and class II DSA (Median MFI for HLA-DR: 4761, HLA-DP: 4939) detected in this cohort (Figure 1B). The lower number of class I DSA and corresponding lower median MFI limited our statistical analysis capability. Therefore, the remainder of this manuscript focuses on class II DSA and class I DSA analyses are provided as supplementary figures.

Figure 1.

Total IgG post-transplant Class II donor HLA-specific antibody (DSA) (A) distribution of HLA specificities and (B) mean fluorescence intensity (MFI) for HLA-DR, HLA-DQ, and HLA-DP. There are 188 Class II DSA presented from n=77 unique serum samples. A total of n=59 patients are represented. Samples were collected at 1–3 months post-transplant (n=58) and at 6–9 months post-transplant (n=19) for patients with persistent DSA.

HLA class II DSA correlations across SAB assays

The IgG MFI of HLA class II DSA were measured in neat sera and pattern-verified DSA were detected across a range of 427 MFI to 23,718 MFI (Figure 2A). Positive DSA detection at a 1:16 dilution is indicated by the royal blue overlay superimposed onto neat serum MFI (Figure 2B and Figure 2D). DSA detected in 1:16 diluted sera correlated with >7000 neat IgG MFI with 81% sensitivity and 92% specificity. C1q-binding DSA is indicated by the red overlay superimposed onto neat serum MFI (Figure 2C). C1q positive DSA correlated with >15000 neat IgG MFI with 85% sensitivity and 88% specificity, corresponding to sufficient HLA-specific IgG antibody concentration to fix complement. Interestingly, 93% (15/16) of C1q positive DSA were HLA-DQ-specific and one had specificity for HLA-DP. No C1q positive DSA had specificity for HLA-DR. In addition, 93% (15/16) of DSA that were able to bind C1q had higher MFIs than the 15,000 MFI threshold, but 36% (26/72) of DSA with neat MFI above this threshold did not bind C1q. When the neat MFI threshold is raised to >17,000, C1q binding sensitivity increases slightly to 88%, with a specificity of 75%. Further increasing the neat MFI threshold to >20,000 increases C1q binding sensitivity to 96%, while specificity decreases to 50%.

Figure 2.

Total IgG post-transplant Class II donor HLA-specific antibody (DSA) mean fluorescence intensity (MFI) is reported in graphs A-C. Overlay in color indicates positive detection of the DSA by (B) 1:16 dilution assay (dark blue) and (C) C1q binding assay (red). There are 188 Class II DSA presented from n=77 unique serum samples. A total of n=59 patients are represented. Samples were collected at 1–3 months post-transplant (n=58) and at 6–9 months post-transplant (n=19) for patients with persistent DSA. Median indicated as solid line. Sensitivity and specificity for predicting each DSA characteristic was determined by receiver operating characteristic (ROC) analysis and optimal Total IgG MFI. ROC analysis for predicting (D) 1:16 dilution and (E) C1q binding using Total IgG MFI.

Testing serum dilutions provides DSA titer information and identifies saturating levels of DSA that exceed the linear range of SAB assays. When comparing the neat IgG MFI and 1:16 dilution MFI, we observed that DSA >15,000 MFI in the neat assay exhibited higher than expected MFI in 1:16 diluted sera, reflecting a potential under assessment of DSA strength (Figure 3). The majority of DSA (107 of 111) detected in the 1:16 assay exhibited <10,000 MFI, indicating that this single 1:16 point dilution was sufficient to retain linearity (Figure 3). In the interest of expanding relevance of this finding, we performed a retrospective analysis of clinical testing data from lung transplant recipients using sera that were treated with EDTA prior to testing. This testing replicated our finding using IgG enriched sera (Zeba & Melon Column), such that DSA >15,000 MFI in the neat assay exhibited higher than expected MFI in 1:16 diluted sera (Supplementary Figure 3).

Figure 3.

Identification of bead saturation and loss of semiquantitative measurement in neat sera. Comparison of post-transplant class I (closed circle, n=27) and class II (open circle, n=84) DSA IgG detection by neat and 1:16 dilution assays. DSA > 15,000 MFI in neat sera display higher than expected MFI in 1:16 assay, indicating a threshold above which DSA can exceed the assay linear range in neat sera.

Analysis of C1q-binding DSAs revealed that 15/16 were detectable in 1:16 diluted sera (Figure 4A) with the majority 68% (11/16) demonstrating an MFI >5,000 (Figure 4B). Examination of the single weakly positive C1q DSA outlier revealed it was specific for HLA-DQ (DQA1*02:01, DQB1*02:01), had a neat IgG MFI of 3025, and fell below the autologous bead MFI in the1:16 dilution assay. These results demonstrate that in the majority of cases, assessment of DSA by 1:16 dilution and C1q-binding SAB assays identify DSA with higher HLA-specific IgG antibody concentrations.

Figure 4.

Comparison of post-transplant Class II DSA detection by 1:16 dilution and C1q binding assays. (A) DSA MFI detected by C1q assay (all), separated by detection in 1:16 dilution assay (blue). (B) DSA MFI detected by 1:16 dilution assay (all), separated by status of C1q binding (red). Median indicated as solid line.

To investigate the IgG subclass composition of the class II post-transplant DSA, we measured IgG1, IgG2, IgG3, and IgG4 subclasses by SAB assay. There were 188 class II DSAs that were detectable by at least one of the IgG subclass assays (Supplementary Table 1). IgG1, IgG2, and IgG3 subclasses were detectable across the full range of IgG MFIs in neat sera, while only 4 DSAs had IgG4 subclass detected (Figure 5A-D). IgG1 and IgG3 subclasses are known to fix complement more efficiently than IgG2 and IgG4 and thus it has been postulated that the C1q assay primarily detects IgG1 and IgG3 DSA.²³ We found that all (16/16, 100%) C1q positive DSA had detectable IgG1 and/or IgG3 subclasses, though IgG2 DSA was also detectable in 14/16 (87%) (Supplementary Figure 4A-D). Interestingly, 8 DSA had only IgG1 and IgG3 detected and, of these, only 1 DSA was positive in the C1q assay. There is a weak positive correlation between IgG1 positive DSA and C1q binding (Figure 6A; Spearman r = 0.64, p = 0.008), but no correlation was found between C1q binding and IgG2 positive DSA (Spearman r = 0.47, p = 0.06) or IgG3 positive DSA (Spearman r = −0.035, p=0.9) (Figure 6B and C). C1q binding correlates most strongly with total IgG MFI (Figure 6D; Spearman r = 0.82, p = 0.002), demonstrating that in this cohort, C1q binding capacity by DSA is driven by HLA-specific total IgG antibody concentration, rather than the presence of specific IgG subclasses.

Figure 5.

Total IgG post-transplant Class II donor HLA-specific antibody (DSA) mean fluorescence intensity (MFI) is reported in graphs A-D. Overlay in color indicates positive detection of the DSA by IgG subclass specific assays: (A) IgG1 (pink), (B) IgG2 (light blue), (C) IgG3 (green), (D) IgG4 (gold). There are 188 Class II DSA presented from n=77 unique serum samples. A total of n=59 patients are represented. Samples were collected at 1–3 months post-transplant (n=58) and at 6–9 months post-transplant (n=19) for patients with persistent DSA. Median indicated as solid line.

Figure 6. IgG Subclass MFI, C1q MFI, and Total IgG MFI were measured by SAB assays and.

Spearman correlation of Class II DSA C1q binding MFI and (A) IgG1, (B) IgG2, (C) IgG3 subclass, and (D) Total IgG MFI are reported.

To investigate further the contribution of IgG subclass to C1q assay positivity, we engineered pan-HLA Class I-specific human IgG subclass monoclonal antibodies (HLA mAbs) and tested them at different concentrations in C1q (Figure 7) and C3d (Figure 8) assays. The MFI was plotted against the IgG-MFI for each bead, at each dilution. The results show that high concentrations of both hIgG1 and hIgG2 can bind to C1q in solid phase assays. With the exception of outliers among HLA-C locus beads, HLA I IgG1 was strongly C1q positive at high concentrations of antibody with IgG MFI >15,000. Importantly, high concentrations of HLA I IgG2 also caused C1q binding, although the strength of C1q binding reduced more readily when antibody concentration declined, consistent with lower IgG2 affinity for C1q compared with IgG1. Nevertheless, these results demonstrate that IgG2 is capable of causing a positive C1q result if present at high enough concentrations (Figure 7). A similar effect was observed for C3d deposition on single antigen beads, which results from physiological activation of complement in the solid phase assay. Here, both HLA I hIgG1 and hIgG2 could cause a positive C3d result across all beads, although the strength of C3d was lower with hIgG2 than hIgG1, and it became negative at lower concentrations of antibody compared with hIgG1 (Figure 8).

Figure 7.

Titrated HLA class I-specific hIgG1 and hIgG2 antibody binding to C1q on HLA class I single antigen beads.

Figure 8.

Titrated HLA class I-specific hIgG1 and hIgG2 antibody-mediated C3d deposition on HLA class I single antigen beads.

Discussion

This study provides a direct comparison across traditional and modified HLA SAB assays using pre-treated sera and monoclonal human HLA antibodies to assess assay correlations and added value of DSA monitoring using modified SAB assays. We chose to use IgG enriched sera as our primary treatment method to generate test correlations and provide consistent detection of the IgG component across all assays. Traditional clinical histocompatibility SAB assays and flow cytometric crossmatches focus on IgG detection while the C1q assay does not distinguish the contribution of non-IgG isotypes. Given that IgM is the most potent complement binding isotype due to its pentamer structure, if sera is not treated to remove IgM, it is unknown to what extent IgM may contribute to a positive C1q result and its clinical significance. Consistent with other studies, we report that DSAs that are detectable by 1:16 dilution and/or C1q are associated with higher IgG MFI values in neat sera. Earlier studies highlighted benefits of C1q and serum titration analysis to improve HLA antibody detection ^23,24; however, these studies pre-dated current serum treatment strategies such as EDTA, DTT, heat, or IgG enrichment columns that are used to mitigate complement protein interference that can lead to false negative results. Our study found strong correlations between the assays with the ability to predict positivity in both the 1:16 dilution and C1q assay from the neat SAB MFI. Schaub et al used sera pretreated with EDTA from a pre-transplant kidney cohort and found that a neat IgG threshold of 14,154 MFI correlated with C1q positivity ²⁰ While our C1q ROC analysis utilized a limited positive sample size (n=16), we confirmed Schaub’s findings in a post-lung transplant cohort. These SAB MFI correlations can be applied to clinical workflows to improve efficiency and cost by stratifying sera that may benefit from additional modified assays. HLA antibody affinity may also impact detection in diluted serum or ability to bind C1q and may explain the rare outliers of 6 DSA detectable at 1:16 dilution with a neat MFI <7000 and 1 C1q+ DSA with neat MFI <14,000).²⁵

Multiple studies have shown that C1q positive DSA correlates with worse transplant outcomes compared to SAB DSA MFI values in undiluted sera ^8,12,15,17. However, whether C1q binding provides added information on DSA injury potential or reflects a more valid assessment of DSA concentration has not been robustly studied with clinical outcomes. We and others have shown that many DSA exceed the SAB linear range in undiluted sera and these MFI are no longer semi-quantitative.^4,18 Figure 3 and Supplementary Figure 3 illustrate that many DSA that exceed 15,000 MFI in undiluted sera, whether pre-treated with EDTA or IgG enrichment columns, showed higher MFI values when tested at a 1:16 dilution. Nevertheless, we did observe that one third of DSA with an MFI above the calculated 15,000 neat MFI threshold did not bind C1q and this finding may hold important insight into antibody affinity or polyclonally required to fix complement. If DSA strength assessment is needed for clinical management, increasing the dynamic range of the assay by testing C1q binding or serum titrations should be performed.^4,18,26,27

Previous studies have suggested correlations between DSA of IgG1/IgG3 subclasses and C1q binding, but our analysis using sera enriched for IgG and thresholds specific for each IgG subclass revealed that DSA of IgG1, IgG2, and IgG3 subclasses were all detectable within DSA that bound C1q. Freitas et al investigating acute rejection in a kidney transplant cohort also detected IgG2 in conjunction with IgG1 and IgG3 in all but one patient with late HLA-DQ DSA onset (mean 34 months post-transplant). They reported correlations between acute kidney rejection and HLA-DQ DSA exhibiting IgG3 subclass, C1q binding, and total IgG MFI median of 16,085. ²⁸ Schaub et al interrogated C1q binding HLA antibodies from kidney waiting-list patients and found complement-binding IgG1 and IgG3 subclasses dominated in both frequency and relative strength. Analyses in this study and some others have examined DSA composition categorized within two groups: IgG1 and IgG3 versus IgG2 and IgG4.²⁰ However, our assessment of engineered IgG1 and IgG2 chimeric mAb binding to C1q and C3d complement proteins revealed that IgG2, which is typically considered a “non-complement binding” subclass, can effectively bind C1q at high concentrations, at least in this solid phase assay. Therefore, detection of C1q-binding DSA should not be interpreted to dictate a particular IgG subclass composition. This is also supported by other studies that have demonstrated complement binding by all IgG subclasses under conditions of high complement protein concentration and/or high antigen density.²⁹ In addition to complement activation, it is important to consider that different IgG subclasses may impact clinical outcome by engaging other Fc-mediated effector functions.^30,31 Viglietti et al reported on 110 subjects using multivariate random survival forest analysis to interrogate the additive value of single immunodominant post-transplant DSA characteristics to predict risk of kidney allograft loss.⁹ The addition of C1q binding capacity and IgG3 positivity to the immunodominant DSA neat MFI improved the c statistic for predicting risk of kidney loss. This study and others have identified IgG3 DSA as a significant contributor to allograft loss; yet the potential underestimation of neat DSA MFI in all of these studies, in the absence of serum dilutions, undermines our ability to assess the added value of C1q binding and IgG subclass analysis. Only 4 DSAs had IgG4 subclass detected, likely reflecting the early-post transplant DSA analyzed in this cohort. We detected IgG1, IgG2, and IgG3 subclasses across a broad range of neat SAB MFI, revealing the polyclonality of the DSA, but composition did not correlate with using IgG subclass analysis as a monitoring assay for early (<1 year) DSA. However, the value of IgG subclass composition when assayed at time of acute or chronic AMR in conjunction with quantitative DSA titration experiments are needed to confirm the added value for the diagnostic correlation of IgG3 subclass DSA with AMR and allograft loss.^12,32

The goal of this study was to correlate results from traditional and modified SAB assays to determine the added value for routine clinical decision making. We confirm other studies in finding C1q positive IgG DSA correlates with higher antibody concentration. We highlight the limited linear range in traditional SAB assays above which the MFI value (>15,000 MFI) no longer provides a semiquantitative measurement. We also show the inability to predict IgG subclass composition from C1q or C3d test results. Understanding the strengths and limitations of SAB assays will improve their efficient use and accurate interpretation when used in the clinical management of transplant patients.

Abbreviations:

CHO

Chinese hamster ovary

DTT

dithiothreitol

DSA

donor HLA-specific antibody

EDTA

ethylenediaminetetraacetic acid

MFI

mean fluorescence intensity

PE

phycoerythrin

SAB

single antigen bead

RT

room temperature