Authors: Prof Alan Parker, February 2022.
In late 2019, the world could never have foreseen the impact that a new, emerging zoonotic coronavirus could possibly have across the planet. Fast forward 12 months, hundreds of millions of cases of SARS-CoV-2 later had caused millions of deaths worldwide. Yet amidst the lockdowns, loss of life and long-covid, a newfound optimism was emerging that the pandemic would soon be resolved by the rapid development of new vaccines to protect against SARS-CoV-2. The pandemic would soon be over, or so it seemed. Vaccines based on mRNA, manufactured by Pfizer/BioNTech and Moderna, were demonstrating remarkable protection in clinical trials, but concerns remained over how these vaccines could ever be supplied across swathes of less developed parts of the world, where their necessary cold chain supply simply wasn’t feasible. Step forward the adenoviral vaccines. Easy to mass produce at scale and low cost and highly effective at inducing robust antibody and T-cell immunity. The solution to vaccinating the world, including the hardest to reach and poorest parts of the planet, appeared in hand. The global pandemic required a global technology to deliver on the mantra that “nobody is safe until everybody is safe”, and it appeared adenoviral vaccines were the solution.
As part of the largest global vaccination campaign ever embarked upon, hundreds of millions of doses of AstraZeneca’s ChAdOx1, and Johnson & Johnson’s Ad26 based coronavirus vaccines were rolled out across the planet. For months, euphoria reigned. Then slowly but surely, evidence of a side effect so rare it was almost impossible to detect began to emerge. A handful cases – initially thought to be simply noise in the data – slowly but clearly presented themselves(1,2). Patients presenting typically 5-24 days post vaccination with rare but sometimes devastating blood clots. Each patient characterised by an odd but clear characteristic - high levels of antibodies to a self-protein – PF4. Haematologists had seen this calling card before, but only in patients treated with the anticoagulant, heparin(3). This heparin induced thrombocytopaenia (meaning loss of platelets) “HIT” syndrome had been reported previously, very occasionally, in subsets of heparin treated patients. Its mechanism had been attributed to the binding of a positively charged blood protein, PF4, to negatively charged high molecular weight heparin, resulting in a complex formation, driving a misplaced immune response to PF4. These pathogenic anti-PF4 antibodies bind platelets via a receptor called FcγRIIa, resulting in platelet activation and loss from the blood, increasing the likelihood of blood clotting(4). However, in this new phenomenon, Vaccine Induced Thrombotic Thrombocytopaenia (VITT, also known as Thrombosis with Thrombocytopaenia Syndrome, TTS), no heparin was present to explain the emergence of pathogenic anti-PF4 antibodies. Another complexity – these complications appeared only to evidence themselves following the first dose of adenoviral vaccine, but hardly ever following the second, “booster” dose.
Over the following weeks and months, a flurry of speculative preprints pointed the fingers at potential reasons for these adverse responses. Could it be impurities in the vaccine preps(5)? Potentially, but why would this specifically drive an anti-PF4 response? What about Spike protein splice variants secreting Spike into the blood, binding ACE2 on endothelial cells and activating them, causing blood clotting(6)? Possibly, although the timing of VITT (5-24 days) would not be compatible with this mechanism - since the maximal adenovirus mediated expression of spike protein would be 1-3 days post injection, and greatly reduced by day 7…similarly this would provide no explanation for the very clear anti-PF4 “signature” of VITT. A more likely culprit was the adenovirus vector itself(7).
For more than a decade, the ability of certain adenoviruses to bind proteins involved in blood clotting had been known, investigated, and even abrogated. The best characterised of these interactions is that between the major capsid protein, hexon, of adenovirus type 5 and blood clotting factor, FX – an interaction which drives the rapid accumulation of blood borne Ad5 in hepatocytes(8,9). This interaction, however, does not activate FX, and therefore it would seem unlikely that direct binding of adenovirus to blood clotting factors would underpin VITT, since this would not drive clot formation. Indeed, previous studies had even shown that Ad26, the basis of the J&J vaccine, does not bind FX(10), whilst binding of FX to ChAdOx1 had not previously been evaluated, but the sequence of the hexon protein of ChAdOx indicates it lacks key FX binding amino acids, making any interaction unlikely. An alternative direct interaction proposed to underpin thrombocytopaenia is direct binding of adenovirus to CAR, the primary receptor of Ad5 and ChAdOx1, on platelets. Expression of CAR on platelets has previously been observed(11) but is likely expressed low level and associated with activated platelets(12). The low levels of CAR expression on platelets make it unlikely this could account for such dramatic side effects, and again, the timing of VITT, at 7-20 days post first vaccination, does not seem compatible with such a direct interaction, that would elicit side effects rapidly post vaccination.
More recently, a more likely explanation has been proposed. By studying the ChAdOx1 capsid structure and its interactions at near atomic level detail, an international team of researchers were able to show structurally and biochemically, how the adenovirus itself may form a complex with PF4(13). Hypothetically, this may occur following microvascular damage during intramuscular vaccination, allowing trace amounts of virus to enter the blood, where PF4, present in extremely low concentrations, may bind the capsid, forming a complex. This complex may then be phagocytosed, presented in the lymph nodes, and in rare cases induce either de-novo production of misplaced, high affinity, pathogenic anti-PF4 antibodies, or, perhaps more likely, restimulate memory B cells to secrete high affinity anti-PF4 antibodies. Such antibody secretion and resultant antibody/PF4 immune complexes could then bind and activate platelets via FcγRIIa, with consequential thrombocytopenia and thrombosis. Such a model appears valid given the timing and frequency of VITT.
This model may also provide insight as to why VITT occurs only following the first dose of vaccine. Both Ad26 and ChAdOx1 were developed based on their low seroprevalence rates in the population. This approach is doubtless beneficial from the point of view of vaccine efficacy. However, the lack of antibodies against the capsid, and specifically the immunodominant hexon protein(14) where PF4 was shown to bind, may allow the complex to form between PF4 and blood borne adenovirus. Following the first dose of vaccine, the vaccine recipient will develop antibodies both against the spike antigen, to protect against SARS-CoV-2, but also against the adenovirus vector itself (ChAdOx1 or Ad26), which will be high affinity, and likely to prevent PF4 binding following any “booster” dose. This may also explain why Ad5 based vaccines, which have significantly higher seroprevalence rates in the population, may not induce VITT, since binding of PF4 might be prevented even after the first dose by either neutralising antibodies, or potentially FX binding with high affinity to the capsid, out competing low affinity PF4 binding to the adenoviral vector.
So, where does this leave adenoviral SARS-CoV-2 vaccines? It should be emphasised again that these vaccines are widely regarded as safe and highly effective. Adverse events, where they occur, so with a vanishingly low frequency, and the protective effects they provide against COVID-19 hugely outweigh any potential risk from the vaccine itself. It also remains clear that the world still requires vaccine doses to be administered across the planet to provide sufficient immunity to control the coronavirus pandemic. Therefore, vaccines which do not require cold chain supply, such as the adenovirus, remain central to any plan to control this, or future pandemics.
So, how can adenoviral vectors safely be deployed, given what we have learned from VITT? Several possibilities exist. The option that many canvass for is to ensure that vaccines are appropriately aspirated following intramuscular injection. However, it has not been suggested that VITT occurs following inadvertent injection of viral vaccine directly into the bloodstream. Rather, that microvascular damage – an inevitable consequence of any injection – causes trace amounts to enter the blood. Therefore, aspiration as part of intramuscular vaccination will likely have limited impact. Another, elegant solution might be to fully define how PF4 binds to the adenovirus capsid, then use this information to engineer PF4-binding resistant capsids. However, there is no way of telling if such changes would be compatible with virus formation, or what new interactions my inadvertently be introduced because of the alterations to the viral capsid. It would be highly unlikely that such an altered vaccine could immediately be introduced into widespread clinical use and would require rigorous trialling first. A much simpler solution could be simply changing the route of administration of the vaccine. Those of us with school age children will know that whilst adults receive annual flu vaccines by intramuscular injection, our children are given the vaccine in an aerosolised formulation via the nasal route of delivery. This intranasal approach, whilst originally taken to make vaccine administration more palatable for children, has a significant advantage of providing enhanced mucosal immunity where it is needed to cope with a respiratory pathogen – i.e., in the lungs. From the perspective of avoiding VITT, simultaneously preventing contact with the blood by giving the vaccine intranasally would appear to be an ideal solution. Trials of this approach are underway, and if successful, it may well be that adenovirus-based SARS-CoV-2 vaccines may still save the day by providing a cheap, effective and cold chain free means to deliver immunity across the planet.
Schultz NH, Sørvoll IH, Michelsen AE, Munthe LA, Lund-Johansen F, Ahlen MT, Wiedmann M, Aamodt AH, Skattør TH, Tjønnfjord GE, Holme PA. Thrombosis and Thrombocytopenia after ChAdOx1 nCoV-19 Vaccination. N Engl J Med. 2021 Jun 3;384(22):2124-2130.
Greinacher A, Thiele T, Warkentin TE, Weisser K, Kyrle PA, Eichinger S. Thrombotic Thrombocytopenia after ChAdOx1 nCov-19 Vaccination. N Engl J Med. 2021 Jun 3;384(22):2092-2101.
Visentin GP, Ford SE, Scott JP, Aster RH. Antibodies from patients with heparin-induced thrombocytopenia/thrombosis are specific for platelet factor 4 complexed with heparin or bound to endothelial cells. J Clin Invest. 1994 Jan;93(1):81-8.
Denomme GA, Warkentin TE, Horsewood P, Sheppard JA, Warner MN, Kelton JG. Activation of platelets by sera containing IgG1 heparin-dependent antibodies: an explanation for the predominance of the Fc gammaRIIa "low responder" (his131) gene in patients with heparin-induced thrombocytopenia. J Lab Clin Med. 1997 Sep;130(3):278-84.
Michalik S, Siegerist F, Palankar R, Franzke K, Schindler M, Reder A, Seifert U, Cammann C, Wesche J, Steil L, Hentschker C, Gesell-Salazar M, Reisinger E, Beer M, Endlich N, Greinacher A, Völker U. Comparative analysis of ChAdOx1 nCoV-19 and Ad26.COV2.S SARS-CoV-2 vector vaccines. Haematologica. 2022 Jan 20.
Kowarz E, Krutzke L, Külp M, Streb P, Larghero P, Reis J, Bracharz S, Engler T, Kochanek S, Marschalek R. Vaccine-induced COVID-19 mimicry syndrome. Elife. 2022 Jan 27;11:e74974.
Othman M, Baker AT, Gupalo E, Elsebaie A, Bliss CM, Rondina MT, Lillicrap D, Parker AL. To clot or not to clot? Ad is the question-Insights on mechanisms related to vaccine-induced thrombotic thrombocytopenia. J Thromb Haemost. 2021 Nov;19(11):2845-2856.
Waddington SN, McVey JH, Bhella D, Parker AL, Barker K, Atoda H, Pink R, Buckley SM, Greig JA, Denby L, Custers J, Morita T, Francischetti IM, Monteiro RQ, Barouch DH, van Rooijen N, Napoli C, Havenga MJ, Nicklin SA, Baker AH. Adenovirus serotype 5 hexon mediates liver gene transfer. Cell. 2008 Feb 8;132(3):397-409.
Parker AL, Waddington SN, Nicol CG, Shayakhmetov DM, Buckley SM, Denby L, Kemball-Cook G, Ni S, Lieber A, McVey JH, Nicklin SA, Baker AH. Multiple vitamin K-dependent coagulation zymogens promote adenovirus-mediated gene delivery to hepatocytes. Blood. 2006 Oct 15;108(8):2554-61.
Alba R, Bradshaw AC, Parker AL, Bhella D, Waddington SN, Nicklin SA, van Rooijen N, Custers J, Goudsmit J, Barouch DH, McVey JH, Baker AH. Identification of coagulation factor (F)X binding sites on the adenovirus serotype 5 hexon: effect of mutagenesis on FX interactions and gene transfer. Blood. 2009 Jul 30;114(5):965-71.
Othman M, Labelle A, Mazzetti I, Elbatarny HS, Lillicrap D. Adenovirus-induced thrombocytopenia: the role of von Willebrand factor and P-selectin in mediating accelerated platelet clearance. Blood. 2007 Apr 1;109(7):2832-9.
Gupalo E, Buriachkovskaia L, Othman M. Human platelets express CAR with localization at the sites of intercellular interaction. Virol J. 2011 Sep 30;8:456.
Baker AT, Boyd RJ, Sarkar D, Teijeira-Crespo A, Chan CK, Bates E, Waraich K, Vant J, Wilson E, Truong CD, Lipka-Lloyd M, Fromme P, Vermaas J, Williams D, Machiesky L, Heurich M, Nagalo BM, Coughlan L, Umlauf S, Chiu PL, Rizkallah PJ, Cohen TS, Parker AL, Singharoy A, Borad MJ. ChAdOx1 interacts with CAR and PF4 with implications for thrombosis with thrombocytopenia syndrome. Sci Adv. 2021 Dec 3;7(49):eabl8213.
Sumida SM, Truitt DM, Lemckert AA, Vogels R, Custers JH, Addo MM, Lockman S, Peter T, Peyerl FW, Kishko MG, Jackson SS, Gorgone DA, Lifton MA, Essex M, Walker BD, Goudsmit J, Havenga MJ, Barouch DH. Neutralizing antibodies to adenovirus serotype 5 vaccine vectors are directed primarily against the adenovirus hexon protein. J Immunol. 2005 Jun 1;174(11):7179-85.