12-month SARS-CoV-2 antibody persistency in a Tyrolean COVID-19 cohort

Coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) leads to an acute immune response transforming into immune memory protecting from infection in previously infected or vaccinated persons [1]. Immunity as a clinical outcome is hard to determine in open-label epidemiological settings. Therefore, surrogates are commonly used mostly by measurement of agent-specific antibodies as a component of immune memory [2]. The short-term antibody response has been shown in COVID-19 cases in several reports [3, 4]; [5‐9], and in few long-term observations up to 12 months [10]. Here we present the 12-month follow-up SARS-CoV‑2 antibody results in a longitudinal prospective Tyrolean cohort which was the primary goal of the study. Additionally, we present pre-vaccine versus post-vaccine SARS-CoV‑2 antibody responses in a subset of 5 study subjects.

Depending on the antigen we found a persistent anti-SARS-COV‑2 antibody response over 12 months after infection with a preference for the RBD and to a lesser degree for the entire S1 subunit of the spike protein. The S1 IgG assay seems somewhat insensitive, as of 7 samples that were formally negative at month 12 only 1 was also negative in the wild type neutralizing assay. The least sensitive antigen was the NCP which turned negative in most patients at month 12 and therefore has only limited use to distinguish between post-infectious and currently available vaccine-induced immunity, the latter exclusively targeting the spike protein. The NAb against SARS-CoV‑2 wild type, which was the predominant type during the first wave when the study participants contracted COVID-19, decreased at month 12 but was still detectable in 92% of convalescents. In contrast, NAb against the variants B.1.1.7 (the predominant variant in Austria at the time of last follow-up, https://​www.​ages.​at/​themen/​krankheitserrege​r/​coronavirus/​sars-cov-2-varianten-in-oesterreich/​) and B.1.351 were less frequently detected at last follow-up. The lower neutralizing capacity against variants is in line with previous observations [12]. The most common spike mutations of the current variants of concern are shown in table S2 which are probably responsible for partial immune evasion; however, there is a wide interindividual variation depending on the individual clonal composition of epitope-specific antibodies [10, 12].

The strong increase of binding and neutralizing antibodies including variants in convalescents after a single dose of vaccine has been described earlier with titres higher than in COVID-19 naïve persons after 2 doses of vaccine [13, 14]. Even post-vaccine, NAb titres against variants were slightly lower compared to the wild type with the limitation of a small sample size in our study; however, this is not consistently reported in the literature with some authors finding significantly higher NAb titres against the alpha variant than against the beta variant [5]. This is an important issue because NAb assays are variably standardized and therefore hardly comparable across different studies. Also, HLA polymorphisms determine the host’s immune response and variations of major histocompatibility complex (MHC) binding affinities depending on SARS-CoV‑2 variants and several HLA alleles have been shown [13]. There are quite a few other factors that need to be considered, such as age of subjects, sex, disease severity, and in cases of vaccine-induced immunization the interval between disease and vaccination or between vaccine doses, all of which are associated with the quantity, the quality, and the duration of the antibody response [4, 13‐16]; however, all these observations underline the importance of the booster vaccinations in previously infected persons by broadening and increasing the immune response probably leading to long-term immunity.

The kinetics of binding antibodies are largely confirmative of previous findings with the most rapid and extensive decline occurring in the NCP IgG assay targeting the least immunogenic antigen as demonstrated by the weak correlation with NAb titres. Even early after disease only 74% of patients had levels above threshold for positivity, possibly indicating a sensitivity issue of this particular test system. This must be considered when using this assay for confirmation of possible infection after vaccination as such a strong decay did not occur using other systems [11].

For S1 IgG the decline of RU/mL values was less pronounced with 79% of patients remaining formally positive at the month 12 follow-up; however, the majority (6 of 7) of S1 IgG negative subjects had a positive result in the wild type NAb assay but only one was NAb positive against the alpha variant and none against the beta variant. On the other hand, of the S1 IgG positives at T4 roughly two thirds had a positive NAb titre against the alpha variant, and roughly one third against the beta variant corroborated by the strong correlation between at least the alpha variant and S1 IgG values.

In contrast to the above antigens, antibodies against RBD showed a stable development over 12 months after infection and correlated strongest with NAb titres. The fact that this assay not only determines IgG but also IgA and IgM might explain this observation apart from RBD seeming to be the strongest driver of humoral immunity [17]. Although IgG is the most long-lasting subclass some people also develop persistent IgM and IgA [10]. Clonal expansion of RBD-specific B‑cells and antibodies has been demonstrated and might also account for the quantitative increase of these antibodies as observed here [10]; however, one must be cautious because IgA and IgM assays might cross-react with non-SARS-CoV‑2 viral and bacterial agents causing respiratory infections [18].

Immunity as a clinical outcome has been described in early recordings of older dynasties of Egypt. It was, however, well-described by the historian Thucydides in his account of the Athenian plague of 430 B.C. [19]: “Yet it was with those who had recovered from the disease that the sick and dying found most compassion. These knew what it was from experience and had now no fear for themselves; for the same man was never attacked twice—never at least fatally.” The NAb tests as surrogates have been reported for COVID-19 finding a strong correlation between NAb titres and clinical vaccine response across seven trials [2], which needs to be interpreted with caution as there were different study designs and equivocal assay systems compared. Moreover, NAb assays only assess limited aspects of clinical immunity in that the interaction between test sera, viruses, and cultured cells susceptible to viral infection are included. There are several details that are not analyzed, such as cellular immunity overall, tissue-based cellular immunity [20], tissue-based humoral immunity with mucosal IgA not correlating well with circulating IgA [21] and of course immune memory [1, 22], which seems to be persisting as of now; however, using antibody tests as surrogates even if not fully validated seems more practical. Assessing long-term clinical immunity required probably 10,000s of unvaccinated convalescents followed prospectively over many years which seems hardly doable at least for the time being. Antibody concentrations at any time point correlated well with future antibody levels independent of the underlying assay. Although correlations are not formally predictive tools one can still assume that antibody persistency depends on the individual quantitative response. In more appropriate regression models others determined the half-life of protection based on NAb decay of 108 days [2]. One would probably rely on NAb titres to determine timing of booster vaccinations until more is known about the duration of clinical protection. In this context, international consensus should be sought, particularly on boosters of previously infected persons as there is broad immune response resulting in an even stronger response after one dose of vaccine [1, 13, 14, 22].

Sex differences with respect to antibody responses have been reported in different contexts [23‐25] with somewhat conflicting results. Often, differential kinetics of titres were not considered. We found higher binding antibody levels in males shortly after infection with a stronger decrease for NCP IgG in females over time possibly related to the higher proportion more severely affected among males. For S1 IgG there was a similar but less pronounced pattern and for RBD pan-Ig antibodies females “caught up” during observation period, starting with lower antibody concentrations shortly after infection and finally slightly higher levels than males. Apparently, the more immune dominant antigens are driving a longer lasting immunity in females as described for influenza [23] and other anti-viral responses, only transiently however [26]. How this evolves long-term for SARS-CoV‑2 remains to be seen.

Finally, although one of the few prospective studies with repetitive testing over 12 months the limitation mainly confers to the relatively low number of participants, which is why we did not perform analytical statistics for some analyses, such as sex differences. Our findings are restricted to a previously healthy population with mild to moderate course of COVID-19 and to the age range between 24 and 64 years. The study was not designed to assess re-infection rates even though none occurred and protective immunity after COVID-19 has been demonstrated in larger populations [2]; however, there seems to be a longer lasting immunity after SARS-CoV‑2 infection than suggested by earlier reports [6] corroborated by a number of more recent publications [1, 4].