Thu, 04/09/2020 - 13:52
Dear colleagues,
Several colleagues asked my attention for the paper (non-peer reviewed) by Wu et et al. on neutralizing antibodies during and after SARS-CoV-2. It is the first published study of this kind and it looks only at non-severe cases, where widely different levels of Neut Ab were found (see Fig 3 A p.15). As could be expected the levels correlated moderately with ELISA Ab against various “spike” antigens (S1, S2 and receptor-binding domain RBD) Fig 2 C p.14. However, although cross-reactivity in ELISA with SARS-CoV-1 has been shown, there was very limited (if any) cross-neutralization of SARS-COV-2 plasma with SARS-CoV-1 (pseudo)virus (Fig 1D p. 13). The SARS-CoV-2 neutralizing titers remained stable 1 or 2 weeks after discharge. Remarkable observations were, however, that the titers rose with age and were positively related with CRP (an inflammatory marker) and also with lymphopenia (indirectly with lymphocyte count). Both of these characteristics have been associated with more severe disease. Unfortunately, no data on viral load are provided.
Longer follow-up of these patients may indicate whether the highly different levels of Neut Ab are somehow predictive of their chances on relapse.
Obviously, at first view, it may seem good news that in vitro neutralizing antibodies are induced in a majority of the patients, but the observation that these antibodies are associated with some “disease severity markers”, such as older age, CRP and lymphopenia may seem strange. Clearly, this illustrates the fact that, in many viral diseases, in vitro neutralizing polyclonal serum antibodies are NOT necessarily a correlate of protection.
We have been confronted with a similar situation in HIV for decades: patients with even broad neutralizing are usually progressors with moderately high viral load.
In some viral infections, where the existence of “naturally induced” protective immunity has been shown (e.g. hepatitis B) convalescent plasma has been successfully used as a prophylactic measure (I’m not sure that a therapeutic effect has been shown?). Giving plasma (polyclonal neutralizing antibodies) to HIV-infected subjects has been tried in the 80s and 90s without success.
Nevertheless, we have now very potent broad HIV-neutralizing monoclonal antibodies, originally derived from HIV-infected progressors, that have shown prophylactic and therapeutic potential in vivo (also in humans). But at least for the extremely variable and rapidly escaping HIV, a combination is needed.
A number of neutralizing monoclonals against SARS-CoV-1 are already available, but it is unlikely that they will directly be useful for SARS-CoV-2, in view of the poor cross-neutralization. Not completely excluded for one or two of them though, but I guess people have tried and failed. Otherwise we would already have heard of it.
Included are some of the papers, mainly with regard to SARS-CoV-1 neutralization and the comparison with SARS-Cov-2. Clearly, while the receptor-binding domain (RBD) is the most obvious target for neutralizing Ab , other epitopes on S1 or S2, involved in the process of binding, fusion, endocytosis and escape to the cytoplasm, may also be important, especially to produce a highly potent “cocktail”.
A last consideration on other potential mechanisms. Besides direct interference with binding and fusion (classical neutralization), Ab binding to the spike or other “surface” proteins (e.g. E) could also be active in viral defense by mechanisms such as “Ab-dependent cellular cytotoxicity” (ADCC) or “Ab-dependent phagocytosis” (ADP), involving NK cells, macrophages etc. Conversely, Ab could enhance cellular infection and worsen disease outcome (Ab dependent Enhancement or ADE). These mechanisms have been shown to be important in other viral infections, but they remain speculative in SARS-CoV-2. Nevertheless ADE has been seen in a model of SARS-CoV-1 vaccination with an MVA vector in ferrets.
Clearly, much remains to be learned about the potential role of Ab in prevention and treatment of COVID-19.
Best wishes,
Guido