Protein drugs have the potential to induce immunogenicity, whose consequences may affect the efficacy or even be life-threatening. New models of biotherapy, such as antibody-conjugated (ADC) drugs, fusion proteins, and polyethylene glycolization, all carry unnatural human protein sequences and/or structural motifs, which may increase the risk of immunogenicity. Therefore, comprehensive risk assessment and strategic analysis are needed to monitor and characterize the immunogenicity of ADC and other biotreatments in order to predict potential clinical effects.
ADC is composed of a monoclonal antibody (mAb) covalently linking to cytotoxic agents through a stable linker, which has the specificity of mAb to tumor cell surface targeting antigens and the high efficacy of cytotoxic drug. Although the current ADCs use human or humanized mAbs and small molecular payloads, the hapten-like structure may increase their immunogenicity potentials compared with therapeutic mAbs.
Anti-drug antiby (ADA) can target different domains of ADC, such as mAb epitope, neoepitope, linker, and cytotoxic agent. If large ADC-ADA immune complexes are ingested by non-targeted immune cells, resulting in cell death, then anti-cytotoxic ADAs have safety risk.
It is generally believed that the industry practices and regulatory guidelines used to evaluate the immunogenicity of biotherapeutic drugs can be applied to ADC. The key to the evaluation of ADC immunogenicity is risk assessment, appropriate detection, and additional characteristics of ADA domain specificity. Immunogenicity risk assessment covers a variety of known factors related to patients and products, which may affect the immunogenicity of drugs and the possible consequences of immune responses. The risk for ADC is generally considered higher than that of therapeutic mAbs.
Based on risk assessment, immunogenicity data were analyzed in 11 clinical trials of 8 ADC drugs involving a range of solid and hematological tumor indications.
-
Baseline incidence of ADA
Based on the baseline incidence data obtained from all ADC studies, the baseline incidence of ADAs is between 1.4% and 8.1%, which is within the range reported by other monoclonal antibody-derived biotherapies, including the ADC drug Kadcyla.
In the clinical study of these 8 ADC antibodies, they are unlikely to be related to other components of ADC. In addition, of the 8 ADC evaluated, the background signal in the untreated patients was similar to the mixed serum control signal from healthy volunteers.
-
Characteristics of ADA immune response
For the 8 ADCs, most of the ADAs are against their mAb domains, ranging from 86% to 100%. These results are similar to those of Adcetris (brentuximab-vedotin), which has the same linker and toxic drug as the eight ADCs introduced in this paper.
Overall, these data suggest that the semi-antigen-like structure of these ADCs seems not to increase the risk of immunogenicity in most patients compared with traditional therapeutic mAbs.
As for the ADA level, the total titers of these ADCs range from less than 1.30 to 4.92 per unit, and the average titer exceeds 2.50 per unit.
As for the time of occurrence of treatment-induced ADA, it is shown to be variable in all ADCs, ranging from 3 to 42 weeks. In addition, 6 of the 8 ADCs developed ADAs in more than 60% of patients within 3 to 6 weeks after the start of treatment.
Most of the responses of all ADCs were persistent, ADC H and I had the highest transient responses (43.8% and 34.1%, respectively). However, classifying the ADA responses as persistent or transient in oncology studies may be misleading because the treatment duration is usually short. In fact, a more detailed assessment of the ADA positive response data showed that the three ADCs (ADC F, H, and I) had higher ADA positive rates.
Of the 652 patients, 33 had antibodies at baseline, and 3 of them had increased ADA response. The ADA titers of the three patients varied over time, with baseline titers ranging from 1.89 to 2.71 and post-baseline titers from 2.50 to 4.10. 2 of the 3 patients had peak ADA after the first post-baseline, and the time range of enhanced response was 3 to 9 weeks. The level of ADA in all patients at the last post-baseline was lower than the peak level. The domain specificity of ADA remained unchanged at baseline and post-baseline time points.
-
Effect of ADA on other clinical data
Immunogenicity is an important clinical characteristic of protein drugs, and its data should be evaluated in the context of other factors such as efficacy, competition, and safety. Here we focus on the three ADCs, ADC F, H, and I, which stimulated a high number of ADAs in 19, 16, and 43 post-baseline ADA positive patients, respectively.
The incidence of ADAs in ADC F was 35.8% (19/53), and 18 patients were treatment-induced ADAs and 1 patient was treatment-enhanced ADAs. Overall, no differences were observed in terms of competition, safety, or efficacy outcomes compared with patients who did not produce ADC F antibodies. In addition, the titers of ADA had an effect on the PK of these patients, which had been previously reported in cynomolgus monkey disease model.
As for ADC H, the incidence of ADAs was 25% (16/64), with all 16 patients developed treatment-induced ADAs. Similar to ADC F, the trough antibody level in some patients with higher ADA titer was lower than that in patients with lower ADA titer, but the patients with higher ADA titer did not correspond to those with lower total antibody level.
A total of 43 patients with ADC I had ADAs, with 41 induction and 2 patients having enhancement. It is worth noting that the total antibody levels of 7 patients (including 2 patients with ADAs enhancement) were lower than that at one or more trough points. However, a comprehensive analysis is still needed to determine whether ADA titers have an impact on the PK of these patients. However, it is worth noting that the high titer of ADA (±4.00) had a significant effect on the PK of trastuzumab ADC in cynomolgus monkey models.
Limited data obtained from listed ADC products show that ADA has little impact on clinical outcomes. The incidence of ADA in Adcetris was 37%, and the ADA response frequency of patients with persistent infusion was higher, resulting in the withdrawal of treatment on two patients.
The incidence of ADA of Kadcyla was 5.3%. The development of ADA seems to have no effect on safety, PK, and efficacy. In the first phase of the study, there were some ADA-positive patients treated by Mylotarg, including one with transient shortness of breath associated with ADAs. No additional immunogenicity data for Mylotarg was generated in clinical trials under the recommended dose regimen at that time. In the clinical trial of Besponsa, the incidence of ADAs was 3%, which had no effect on the clearance rate of ADC.
Summary
In a wide range of tumor indications in 11 clinical trials, the incidence of ADAs in 8 vc-MMAE-ADCs ranged from 0 to 35.8%. Most ADA responses target the mAb domain in an ADC, indicating that the hapten-like structure plays a very small role in generating immune responses against these ADCs. These results emphasize that molecular structure is an important but not the only factor causing the immune response to biotherapy.Although the data from the 8 vc-MMAE-ADCs suggest that the risks may be reduced as the projects enter a later stage of clinical development, a conservative approach is recommended for new ADCs that may use different linkers and more powerful cytotoxic drugs. In addition, as predictive immunogenicity tools become more common, they will provide additional information to be considered in risk assessment and immunogenicity strategies for ADC and new treatments.
References:
Immunogenicity of antibody-drug conjugates: observations across 8molecules in 11 clinical trials. Bioanalysis. 2019 Sep;11(17):1555-1568