Selective IgM digestion improves anti-AAV IgG detection in the immune complex assay format

ABSTRACT

Aim

To confirm the observation of IgM interference in the anti–adeno-associated virus (AAV) IgG immune complex (IC) assay format and to verify that IgM-specific digestion can improve anti-AAV IgG detection in IC assays.

Methods

Treatment-emergent anti-AAV2 and anti-AAV9 IgG signals were measured in IC assays with and without IgM-specific digestion. Anti-AAV2 and anti-AAV9 IgM signals were measured in parallel.

Results

IgM digestion increased anti-AAV2 and anti-AAV9 IgG signals when anti-AAV2 or anti-AAV9 IgM were present in the matrix.

Conclusions

Co-existing anti-AAV IgM cause interference in the anti-AAV IC assay format. Selective IgM digestion improves the detection of anti-AAV IgG in the IC assay.

1. Introduction

The previously published immune complex (IC) assay format for the detection of total anti–adeno-associated virus serotype 2 (AAV2) IgG offers a practical alternative to the existing anti-AAV assays [1]. In the IC format, study samples are preincubated with capsid particles to form anti – AAV-antibody – capsid complexes, which are then captured using an immobilized anti-capsid antibody and detected using antibodies directed against anti-AAV antibodies. This approach contrasts with the direct ELISA format, in which immobilized capsid particles are used as a capture reagent. This fundamental difference of the IC assay format confers several advantages over the direct ELISA as discussed in the previous report [1], most notably low capsid material consumption.

However, when compared with the direct ELISA, the IC assay detected fewer positive samples up to 4 weeks after AAV administration, while showing similar results in samples collected at later timepoints. The authors attributed this difference to lower sensitivity of the IC assay and suggested that it could be study-dependent. Although the overall in-study incidence of anti-AAV2 IgG was equally well determined by both assays, the discordant results at early timepoints indicated limited utility of the IC assay for in-study monitoring of the time course of anti-AAV IgG responses.

Recently, the authors used a panel of routine immunogenicity assays, including anti-AAV2 IC assays, to explore bioanalytical applications of IgM-selective proteolytic digestion [2]. Unexpectedly, the results of this study indicated that co-existing anti-AAV2 IgM could inhibit assay signals in the anti-AAV2 IgG IC assay, with IgM digestion leading to a signal increase in the IgG IC assay. Notably, the signal increases were most prominent in samples collected within the first 4 weeks after viral vector administration. This finding could explain the apparent difference in sensitivity between the IC assay and direct ELISA, which was suggested in the previous report [1], and offer a strategy to improve the performance of anti-AAV IgG IC assays.

Building upon these results, the authors sought to reevaluate the conclusion made in the original publication describing the IC assay format [1]. Specifically, the authors aimed to confirm the observation of IgM interference in the anti-AAV IgG IC assay format using a larger set of preclinical samples, including samples collected in another preclinical study with a different AAV serotype and route of administration, and to verify that IgM-specific digestion can improve anti-AAV IgG detection in IC assays.

2. Materials and methods

2.1. Chemicals, reagents, and equipment

Recombinant AAV vector particles of serotype 9 (rAAV9) were provided by Roche Diagnostics GmbH, Mannheim, Germany. IgM-cleaving protease (IgMBRAZOR™) was provided by Genovis AB, Kävlinge, Sweden. Biotinylated murine monoclonal anti-AAV2 (A20R; Cat No. 610298) and anti-AAV9 antibodies (Clone ADK9; Cat No.615162) was from Progen Biotechnik GmbH, Heidelberg, Germany. Alkaline phosphatase (AP)-conjugated goat anti-human IgM was from Jackson ImmunoResearch Europe Ltd., Ely, UK.

Other chemicals and reagents are described in the previous report [1].

2.2. Collection of preclinical samples

Cynomolgus serum samples were collected as part of routine sample collection in non-Good Laboratory Practice (GLP) preclinical development studies. All study procedures followed current ethical standards and applicable laws and regulations on animal experimentation.

Serum samples from recombinant AAV2 (rAAV2)-treated cynomolgus monkeys were obtained in a 12-week non-GLP study described previously [1].

Serum samples from rAAV9-treated cynomolgus monkeys were obtained in a 12-week non-GLP study. In this study, animals received a single intravenous injection of rAAV9 (target dose 2 × 10¹³ vg/kg). Blood was collected over a period of 4 weeks (predose on Day 1 and on Days 7, 14, 21, 28 postdose). Serum was separated by centrifugation, aliquoted, frozen and stored at − 60°C or lower.

Pre-study testing for neutralizing antibodies against the administered AAV serotypes was performed in both studies and only animals tested negative were included in the treatment phase.

2.3. IgG-specific immune complex assays with IgM-cleaving protease pretreatment

Before analysis, serum samples (10 μL) were incubated overnight at room temperature with the IgM cleaving protease (1 μL, final protease concentration 10 U/μL) or phosphate buffered saline (PBS; 1 μL, “untreated samples”). Anti-AAV2/anti-AAV9 IgG were detected with the IgG-specific immune complex assay described in Wessels et al. [1]. Briefly, the samples were first incubated with capsid particles (rAAV2 or rAAV9) and then with a capture antibody (biotinylated murine monoclonal anti-AAV2 or anti-AAV9 antibody, 0.5 µg/mL). Thereafter, the samples were transferred onto a streptavidin-coated microtiter plate and the captured immune complexes were detected using a detection antibody (horseradish peroxidase [HRP]-labeled goat anti-human IgG).

Anti-AAV2/Anti-AAV9 positivity was determined using the same exploratory in-study cutpoint approach as in the original publication (two times the mean signal value of non-spiked predose samples) [1].

2.4. IgM-Specific immune complex assays

The assays were conducted as described for the IgG-specific IC assays using AP-conjugated goat anti-human IgM (0.5 μg/mL) as a detection antibody and p-nitrophenyl phosphate (pNPP)-based color reaction for signal readout.

3. Results

The first experiment with a larger set of samples from the preclinical study with AAV2-based viral vectors confirmed our initial observation in Pöhler et al. [2]. IgM digestion led to signal increases in the anti-AAV2 IgG IC assay in samples showing high anti-AAV2 IgM signals, with the magnitude of the signal increases correlating with the strength of the anti-AAV2 IgM signals (Figure 1 and Supplementary Table S1). Notably, the effect of IgM digestion was most prominent in animals with early and strong anti-AAV2 IgM responses (signal peak at around 4 weeks postdose; Figure 1(A)), whereas no or small effects were seen in animals with delayed and weak anti-AAV2 IgM responses (signal peak at 8 weeks or later postdose; Figure 1(B)). IgM digestion markedly improved the detection rate of anti-AAV2 IgG in the IC assay, almost eliminating qualitative differences between the IC assay and direct ELISA reported in the previous publication (Table 1).

Figure 1.

The effect of selective IgM digestion on assay signals in the anti–adeno-associated virus serotype 2 (AAV2) IgG immune complex (IC) assay. Signal intensities for anti-AAV2 IgG (with and without IgM protease pretreatment) and anti-AAV2 IgM at different sampling timepoints in representative animals. The results of anti-AAV2 IgG direct ELISA (dashed line) reported previously [1] are shown here for illustrative purposes. (A) Animals with strong anti-AAV2 IgM signals present at early sampling timepoints (before Week 4). (B) Animals with weak anti-AAV2 IgM signals present at late timepoints (after Week 4).

Table 1.

Detection rate of anti-AAV2 IgG in preclinical samples from rAAV2-treated cynomolgus monkeys.

	Anti-AAV2 IgG Positive Animals (% of total, N = 11)
Timepoint	Immune complex assay	Direct binding assay	Immune complex assay + IgM protease
Predose, Week − 1	1 (9%)	1 (9%)	no data
Predose, Day 1	0	0	0
Postdose, Day 8	1 (9%)	2 (18%)	1 (9%)
Postdose, Day 15	4 (36%)	11 (100%)	9 (82%)
Postdose, Week 4	10 (91%)	11 (100%)	10 (91%)
Postdose, Week 8	11 (100%)	11 (100%)	11 (100%)
Postdose, Week 12	11 (100%)	11 (100%)	11 (100%)

Grey-shaded data reported previously [1] are shown here for illustrative purposes to ease comparison to the data obtained after the treatment with an IgM-selective protease. N: number of analyzed animals; rAAV2: recombinant anti–adeno-associated virus serotype 2.

Similar results were obtained in the second experiment with samples from the preclinical study with an AAV9-based viral vector (Figure 2 and Supplementary Table S2). Prominent anti-AAV9 IgG signal increases upon IgM digestion were only seen in samples with strong anti-AAV9 IgM signals (Figure 2(A)) but not when anti-AAV9 IgM signals were low or declining (Figure 2(B)). The detection rate of anti-AAV9 IgG was improved in the samples pretreated with IgM-cleaving protease, notably, at the first two sampling timepoints (Table 2). Of note, the individual differences in IgM response observed in both studies likely reflect natural biologic variability as no findings that could otherwise explain these differences were reported in the studies (e.g., clinical observations or treatment differences).

Figure 2.

The effect of selective IgM digestion on assay signals in the anti–adeno-associated virus serotype 9 (AAV9) IgG immune complex (IC) assay. Signal intensities for anti-AAV9 IgG (with and without IgM protease pretreatment) and anti-AAV9 IgM at different sampling timepoints in representative animals. (A) Animals with strong anti-AAV9 IgM signals. (B) Animals with weak anti-AAV9 IgM signals.

Table 2.

Detection rate of anti-AAV9 IgG in preclinical samples from rAAV9-treated cynomolgus monkeys.

	Anti-AAV9 IgG Positive Animals (% of total, N = 18)
Timepoint	Immune complex assay	Immune complex assay + IgM protease
Predose, Day 1	0	0
Postdose, Day 7	5 (28%)	7 (39%)
Postdose, Day 14	15 (83%)	16 (89%)
Postdose, Day 21	17 (94%)	17 (94%)
Postdose, Day 28	17 (94%)	17 (94%)

N: number of analyzed animals; rAAV9: recombinant anti–adeno-associated virus serotype 9.

4. Discussion

Intact IgM achieve high avidity despite low binding affinity due to their polymeric (mostly pentameric) structure, which enables polyvalent binding. This structure is especially well suited to binding repetitive epitopes, such as those present on viral capsids [3,4]. In the IC assay, co-existing anti-AAV IgM can thus effectively compete with anti-AAV IgG for binding to capsid particles, preventing the formation and subsequent detection of < anti-AAV IgG – capsid > immune complexes (Figure 3(B)). Interference by co-existing anti-AAV IgM can be eliminated by IgM-specific digestion, which destroys the polymeric structure of IgM and creates low avidity F (ab´)2 fragments unable to effectively compete with anti-AAV IgG (Figure 3(A,C)). Accordingly, IgM digestion resulted in a signal increase in the anti-AAV2 IgG IC assay in our recent exploratory study [2].

Figure 3.

Anti–adeno-associated virus (AAV) IgM interference in the immune complex (IC) assay and its mitigation using IgM-specific digestion. (A) Intact IgM have high avidity due to their polymeric structure. IgM-specific digestion creates F(ab ´)2 fragments with reduced binding strength. (B) Co-existing anti-AAV IgM compete with anti-AAV IgG for capsid antigens and prevent the formation of < anti-AAV IgG – capsid > immune complexes. (C) Low affinity and low avidity F(ab´)2 fragments formed after IgM-specific digestion cannot effectively compete with high affinity anti-AAV IgG.

Our investigation in a larger set of preclinical samples from two different studies with different AAV serotypes confirmed this surprising observation. The apparent lower sensitivity of the IC assay format at earlier timepoints is thus not an inherent limitation of this assay format as suggested in the original publication [1] but the result of interference by coexisting anti-AAV IgM. In line with the competitive mechanism of interference (Figure 3), the effect of coexisting IgM was most prominent in animals with early onset of the anti-AAV IgM response and strong anti-AAV IgM signals. In animals with late IgM response, the increased amount of anti-AAV IgG and their increased affinity appeared to be sufficient to overcome anti-AAV IgM competition.

Low capsid material consumption in the IC assay, which was the primary driver of the assay development, apparently makes it susceptible to IgM interference. In the direct ELISA, large quantities of capsid particles used for coating provide sufficient number of capsid epitopes to avoid saturation and masking by coexisting anti-AAV IgM. In contrast, competitive signal suppression in anti-AAV IgG IC assays is enabled by much lower number of capsid particles and different binding kinetics in solution as compared with immobilized capsids in the direct ELISA. This effect could be mitigated by selective IgM digestion performed before sample analysis, which increased the signal strength in the anti-AAV IgG IC assay, improved the detection rate of anti-AAV IgG and thus eliminated the observed performance differences to direct ELISA.

The sensitivity of the IgG IC assay to co-existing IgM is specifically relevant for gene therapy as it appears to be driven by the nature of capsid epitopes favoring polyvalent binding of anti-capsid IgM. No such effect is expected in the related anti-drug antibody (ADA) IC assay for antibody-based drugs [5] because drug antibodies lack sufficient number of epitopes to enable high avidity binding of IgM and thus a noticeable competition between IgG and IgM ADAs. Moreover, anti-AAV IgM often co-exist with anti-AAV IgG [6], making anti-AAV IgM interference practically relevant for monitoring of the immune response to gene therapy. Mitigation of IgM interference is thus key for successful implementation of anti-AAV IC assays in this field. Since IgM-protease pretreatment can be easily added to the established IC assays without a negative effect on assay reliability [2], we recommend using this approach in situations when coexisting IgM may be present in the matrix, such as for in-study monitoring with sampling shortly after viral vector administration.

5. Conclusion

Co-existing anti-AAV IgM can cause interference in the anti-AAV IgG IC assay format, suppressing assay signal and potentially leading to false negative results. After disruption of the pentameric structure of IgM, the sensitivity of the anti-AAV IgG IC assay appears to be similar to that of the direct ELISA. Easy-to-implement selective IgM digestion, which can mitigate IgM interference, is recommended when the IC assay is used to detect anti-AAV IgG in samples with high likelihood of co-existing anti-AAV IgM, such as samples collected within first weeks postdose during in-study monitoring.

Funding Statement

This manuscript was funded by Roche Diagnostics GmbH, Penzberg, Germany.

Article highlights

• Anti-AAV IgG IC assay format is susceptible to interference by co-existing anti-AAV IgM.

• The IgM interference was consistently observed across two different gene therapies and routes of administration (intravitreally administered AAV2 and intravenously administered AAV9-based viral vectors). It was especially prominent in animals with early and strong anti-AAV IgM response.

• Sample pretreatment with an IgM-specific protease mitigated anti-AAV IgM interference in the IC assay.

• Selective IgM digestion is an easy-to-implement approach to improve the performance of anti-AAV IgG IC assays. It is recommended for analysis of samples with high likelihood of co-existing anti-AAV IgM.

Disclosure statement

All authors are employees of Roche Diagnostics GmbH and owners of Roche shares. Uwe Wessels, Alexander Pöhler and Kay-Gunnar Stubenrauch have a pending patent application related to the subject matter presented. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Writing assistance disclosure

Medical writing support was provided by Alexander Nürnberg and was funded by Roche Diagnostics GmbH.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Author contributions

Uwe Wessels: Conceptualization, Investigation, Methodology, Validation, Visualization, Writing – original draft; Alexander Pöhler: Conceptualization, Methodology, Writing – review & editing; Francesca Ros: Conceptualization, Methodology, Writing – review & editing; KayGunnar Stubenrauch: Conceptualization, Methodology, Supervision; Writing – review & editing.

Ethical conduct of research

All animal study procedures followed current ethical standards and applicable laws and regulations on animal experimentation.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/17576180.2025.2548191