Closing in on an understanding of VITT

Early in 2021, rare blood clotting adverse events were linked to the AstraZeneca (AZ) and Johnson & Johnson (J&J) adenovirus vector vaccines (but not the mRNA vaccines). Both vaccines have had excellent safety records, but in the UK where tens of millions of doses of the AZ vaccine had been administered, a small number of unusual blood clotting events associated with the vaccine began to surface. The incidence rate was very low with a frequency of about 1 in 100,000 and a mortality rate of roughly 1 in 1,000,000. In addition, a better understanding of the disorder led to faster recognition and improved treatment (QH). 

Fortuitously, German researchers led by blood clotting expert Andreas Greinacher discovered a link between these deadly events and a rare clotting syndrome called heparin-induced thrombocytopenia (HIT). In HIT, heparin, an anti-coagulant, interacts with Platelet Factor 4 (PF4) to form a complex. This interaction exposes new binding surfaces on PF4, leading to the production of auto-antibodies in certain individuals. These auto-antibodies bind to the PF4-heparin complex, forming immune clusters that activate platelets resulting in excessive clotting and a rapid decrease in platelet levels i.e. thrombocytopenia (QH).

The severe blood clotting events observed in the small number of subjects vaccinated with the AZ vaccine showed similarities to HIT. They exhibited out-of-control clotting in conjunction with a low platelet count, and most tellingly, anti-PF4 antibodies. Dr. Greinacher and others have named the new syndrome vaccine-induced immune thrombotic thrombocytopenia (VITT).

Because of these insights, VITT can now be diagnosed in an expedited manner. It is indicated by high d-dimer levels (a fibrin degradation product arising when a blood clot is degraded by fibrinolysis) and the presence of anti-PF4 antibodies. The diagnosis is confirmed with platelet activation assays. VITT can be treated effectively with therapeutic-dose anticoagulation and high-dose immune globulin.

An outstanding question is the mechanism by which the adenovirus vaccines may give rise to the anti-PF4 auto-antibodies. In a previous post, I explored a speculative hypothesis put forth that cryptic splicing of the expressed spike gene RNA transcript in the cells containing the adenovirus vaccine vector may give rise to abnormal proteins that could induce the auto-antibodies. Cryptic splice sites are not normally used to carry out splicing because they are only weakly recognized by the splicing machinery. However, even if utilized at low frequency, these cryptic sites in a highly expressed gene can produce a significant number of aberrant mRNA molecules (of the wrong sequence) which are translated into aberrant protein molecules.

An alternative hypothesis is that the auto-antibodies may simply be associated with the presence of adenovirus (and its foreign antigens), and hence should be seen (at low frequency) in patients with adenovirus infections. With the former, the solution would be to re-design the vaccine vector to remove cryptic splice sites. In the case of the latter, then any vaccine or gene therapy that uses adenovirus may possibly cause HIT/VITT-like events at very low frequency. Recent data supports the latter.

New work (NEJM) explored this latter possibility by investigating whether VITT-like clinical and laboratory features were found in patients after they had been infected with adenovirus (a relatively common infection that gives rise to mild cold-like symptoms). Researchers identified four adenovirus-infected subjects who exhibited more severe VITT-like clinical symptoms, e.g. blood clotting, high d-dimer levels, platelet activation and most tellingly anti-PF4 auto-antibodies. They conducted proteomic analysis on the anti-PF4 antibodies (i.e. sequenced the antibodies), and quite strikingly found that the antigen-binding sites of these VITT-like anti-PF4 antibodies had highly similar amino acid fingerprints as anti-PF4 antibodies from VITT patients (Figure 1). More specifically the antigen-binding sites in the two sets of antibodies contained the same characteristic amino acids (see Figure 1), which strongly suggests that the antigens that they bind must also be very similar, i.e. the adenovirus proteins.

The conclusion was that that the antibodies induced by adenoviral vector-based Covid-19 vaccination (classic VITT) and the VITT-like antibodies caused by natural adenovirus infection are extremely similar. This surprising resemblance in auto-antibody fingerprints between the two disorders, based on patient-derived antibodies, draws a connection between VITT and the anti-PF4 disorder linked to adenoviral infection as closely-related adverse immune responses induced by adenovirus antigens. These findings suggest that the anti-PF4 disorder initially identified as VITT is essentially the same as the disorder sporadically caused by adenovirus infection.

Figure 1. Similarities between anti-PF4 antibodies in vaccine-induced immune thrombocytopenia and thrombosis (VITT) and VITT-like disorders caused by adenovirus. The amino acid sequences in the antigen-binding site of light chain (left) and heavy chain (right) anti-PF4 antibodies are shown with the characteristic sequences from VITT patients at the top and the four adenovirus-infected patients at the bottom. Quite striking are the conserved D (aspartic acid) and E (glutamic acid) residues in both sets of sequences which help to form a negatively-charged pocket which can bind a positively-charged region of the antigen, i.e. PF4 or adenovirus protein (adapted from Fig. 1 of Wang et al. NEJM, 2024).