Comparative analysis of marketed factor VIII products: reply

Development of factor VIII (FVIII) inhibitors following replacement therapy with FVIII is one of the major challenges faced when treating patients with hemophilia A. Inhibitors develop in 20–32% of previously untreated patients with severe and in 3–13% with moderate or mild hemophilia A 1, 2. The cause of the immunogenicity is not well understood. There is evidence that both genetic and non‐genetic factors influence patients’ susceptibility to develop these antibodies 3, 4. Novel hypothesis and research approaches are required to obtain more clarity on the molecular basis of FVIII immunogenicity and how FVIII triggers unwanted immune responses in some patients but not in others. The aim of our paper ‘Comparative analysis of marketed factor VIII products: recombinant products are not alike vis‐a‐vis soluble protein aggregates and subvisible particles’ 5 was to advance science and help to generate new hypotheses for future research on the molecular basis of FVIII immunogenicity. Our aim was to provide scientific transparency of our rationale and to stay on the level of a scientific debate rather than discussing brands or products. We believe that when it comes to treatment decisions clinical evidence on a robust basis is required. Disclosure of brand names as recently suggested by M Makris and A Farrugia in a letter to the editor of JTH 6 could influence treatment decisions by hypotheses rather than robust clinical evidence.

Data from other protein products suggest that critical quality variables such as soluble protein aggregates (SPAs) and subvisible particles (SVPs) influence the immunogenicity of protein therapeutics 7. We analyzed SPAs and SVPs concentrations in commercially available recombinant FVIII (rFVIII) products to understand if there are differences between these products after reconstitution. Moreover, we wanted to know if and how levels of SPAs and SVPs change upon exposure of rFVIII products to relevant stress conditions such as agitation and sheer stress. Pre‐existing SPAs and SVPs may act as seeds that nucleate further protein aggregation upon exposure to stress 8, 9, 10.

Our data derived from the analysis of three to six different lots of nine rFVIII products revealed the following.

1. SPAs and SVPs were detected in all lots from all products investigated in varying quantities after reconstitution. SPA concentrations ranged from 0.2% to 11.6%; SVP concentrations ranged from 0.7 × 10⁶/1000 IU to 114.0 × 10⁶/1000 IU. There were lot‐to‐lot variations in each product.

2. Upon exposure to relevant stress (agitation and sheer stress) the products formed additional SPAs and SVPs to different degrees. Products with the highest concentrations of SPAs or SVPs after reconstitution showed the highest increase in these variables upon relevant stress, indicating that SPAs and SVPs present in the products after reconstitution might act as seeds that nucleate further SPAs and SVPs upon exposure to stress.

3. The size distribution of SVPs in rFVIII products after exposure to relevant stress was similar to that determined in the products after reconstitution, albeit with a slight increase in larger SVPs (on average 45.8% 0.75–1 μm, 26.4% 1–2 μm, 21.7% 2–4.5 μm and 6.1% 4.5–70 μm in size)

4. The majority (53–99%) of SVPs was protein or contained protein. Individual lots of some products were found to contain increased concentrations of non‐protein particles.

5. The use of any single method for assessment of aggregates is not sufficient to provide a robust measure of protein aggregation.

6. No difference was observed in the initial presence or de novo formation of SPAs or SVPs that could be attributed to the presence or lack of the B‐domain in full‐length rFVIII and B‐domain‐deleted rFVIII products.

The question arises if and how SPAs and SVPs found in the nine recombinant FVIII products influence the immunogenicity of these products. There is experimental evidence that protein aggregates may elicit or enhance immune responses by several mechanisms, including: extensive cross‐linking of B‐cell receptors, causing efficient B‐cell activation 11, 12; enhancing antigen uptake, processing and presentation; and triggering immunostimulatory danger signals 13. However, other critical information that would be essential to directly correlate the concentrations of SPAs and SVPs with protein immunogenicity in patients is still lacking. In particular, the following information would be required.

1. Types and quantities of SPAs and SVPs needed to generate immune responses for any given therapeutic protein product, in our case for rFVIII products. There is evidence that higher‐molecular‐weight aggregates and particles are more potent in eliciting immune responses than lower‐molecular‐weight aggregates 11, 12, 14. However, quantitative information about a correlation between the concentration of these higher‐molecular‐weight aggregates and particles and the immunogenicity of products in patients is missing.

2. The SPAs and SVPs formed and the quantities that efficiently elicit immune responses may differ for different products and in different clinical scenarios. Moreover, the immune system of each patient might have a different sensitivity for immune activation by SPAs and SVPs contained in FVIII products.

In addition to the baseline levels of SPAs and SVPs in each rFVIII product, our data indicate that product mishandling after reconstitution can increase the concentration of SPAs and SVPs. Similar findings for FVIII products were recently published by Tsutomo et al. 15. Thus, it is important to educate end‐users about proper product handling to avoid an amplification of potential adverse effects due to increases in SPAs and SVPs induced by mishandling. 15.

In conclusion, we believe the research community needs to pay more attention to the presence of SPAs and SVPs in FVIII products and how these variables influence the immunogenicity of the products in patients. Our current understanding does not yet allow specific conclusions on how levels of SPAs and SVPs in FVIII products translate into product immunogenicity in patients but we believe that thorough assessment of these variables is important.

Addendum

B. M. Reipert wrote the manuscript. J. Anzengruber and F. Scheiflinger revised the manuscript and all authors approved the final version.

Disclosure of Conflict of Interests

This work was funded by Shire. All authors are employees of Shire.

Reipert BM, Anzengruber J, Scheiflinger F. Comparative analysis of marketed factor VIII products: reply. J Thromb Haemost 2019; 17: 233–5.