Will antibodies be able to protect us from COVID-19?

Researchers try to answer this question by analysing data in a follow-up to the sero-survey conducted in Pune in July-August and conclude that there is some evidence that a high seroprevalence of SARS-CoV-2 in a locality means that the population there may at least in the short term experience community-level immunity. However, vaccination is the only way to ensure long-term immunity.

Community serological surveys are aimed at estimating prevalence of infection in a given region and thereby guiding strategies to prevent its spread. They are based on detecting antibodies (generally the category called immunoglobulin G, or IgG) in the population of the region. The presence of antibodies in an individual is indicative of that person having had infection in the recent past (that is, 10-14 days ago). The estimate of the population share with antibodies (or the seropositivity rate) tells one the level of prevalence of infection in the region. Sero-surveys also tell one about the prevalence of asymptomatic infection in the region.

The immune system mounts a defence against an infection by producing antibodies specifically against the invading pathogen, with the IgGs, in particular, lasting for fairly long (Figure 1). Recent studies on patients who recovered from COVID-19 suggest that IgGs against SARS-CoV-2 may last up to six to eight months. Antibodies develop against different proteins that are part of the virus. Antibodies against one type of viral protein might neutralise the virus by blocking its entry into cells and preventing it from multiplying, while others might not. But do the antibodies picked up during sero-surveys provide protective immunity (that is, prevent infection) by neutralising the virus when it attacks again, at least in the short term?

Long-term immunity is generally conferred only through vaccines. But, in the absence of a vaccine, knowledge of protective immunity would be of importance to evolve public health strategies for infection prevention and control. If it is known that the antibodies detected in the population are of the virus-neutralising kind, then the health authorities can focus their infection control strategies on regions that have low seropositivity rates rather than those that have high seropositivity rates.

But how does one know whether the detected antibodies are virus-neutralising ones? That will depend on the test kits employed in sero-surveys. Most antibody tests cannot distinguish between neutralising and non-neutralising (or binding) antibodies (Figure 2). But if one has specially designed antibody testing kits that pick out neutralising antibodies, then one is closer to answering the question, especially if the test used has a high specificity.

The Pune sero-survey, unlike the others conducted in the country, did precisely that. It was conducted during July 20-August 5 and found an average seropositivity rate of 51.5 per cent in the surveyed five sub-wards, or prabhag s, in high-incidence areas of the city. As reported earlier , the Pune sero-survey employed an indigenous testing kit that the Translational Health Science and Technology Institute (THSTI), Faridabad, Haryana, had developed in-house. Called the THSTI-RBD-ELISA (enzyme-linked immunosorbent assay) kit, it detects IgG antibodies specifically to the receptor-binding domain (RBD) of the spike (S) protein of the causative virus SARS-CoV-2. It has been demonstrated to have 100 per cent specificity and 84.7 per cent sensitivity to the virus.

RBD is the part that enables the virus to bind with the ACE2 receptor molecule in different host cells, enter them and start multiplying into millions of copies that infiltrate various cells of the body that have the ACE2 receptor on their surfaces. So, it stands to reason that if an antibody latches on to the RBD of the virus (an anti-RBD IgG), it would be a virus-neutralising antibody. Thus, an estimate of the presence of such RBD-specific antibodies would be a fairly good correlate for the level of immune protection, both in an individual and at a community level.

Follow-up of sero-survey

In a recent follow-up of the Pune sero-survey, researchers, led by Aurnab Ghose of the Indian Institute of Science Education and Research, Pune, examined the implications of the results for protective immunity in the population surveyed. Accordingly, the researchers used the same randomly selected 1,664 asymptomatic individuals of the sero-survey, though the data of only 1,659 individuals could be used as the remaining five had incomplete demographic information. Of the 1,659 sera, 857 were positive for RBD-IgG.

Noting that current testing guidelines substantially underestimate the spread of SARS-CoV-2 in dense urban populations, the researchers attempted to provide granular estimates of seroprevalence and also determine how well these numbers correlate with virus neutralisation. To establish the correlation of the estimate of seroprevalence with actual neutralising ability, the researchers used large subsets of the positive sera and employed two different assays that measure the extent of neutralisation: one called the Surrogate Virus Neutralisation Test (SVNT) and the other known as the Plaque Reduction Neutralisation Test (PRNT). There is no live virus involved in the SVNT. It only mimics the virus-host interaction by using purified proteins and tests how well the antibody blocks the binding of the spike protein with the ACE2 receptor. On the other hand, the PRNT makes use of live virus grown in a Vero cell line (derived from African green monkeys) to see whether the antibody neutralises the virus. This test, however, is laborious and requires a biosafety level 3 facility, an expensive proposition.

According to the paper reporting the work, which was posted on the e-print repository medRxiv on November 17, this is the first systematic high granularity serosurveillance study along with estimates of correlation with immune protection from a low-to-medium-income country.

“Not all anti-RBD IgGs will be neutralising antibodies, but the bulk of them would be.” Aurnab Ghose pointed out this interesting fact to Frontline in an email correspondence when queried on the need to actually carry out such neutralisation tests with the RBD-IgG positive sera. “Antibodies are generated randomly. Some anti-RBD antibodies, though able to bind to this part of the protein, may have poor binding efficiency and therefore may not efficiently neutralise. Without specifically evaluating those, one cannot comment on individual protection, let alone community-level protection. In the current extension of the earlier report, we have evaluated the presence of antibodies that will block the viral RBD from binding to the ACE2 receptor,” he added. As the work has shown, this correlation, while in the main true, is much more complex than one would imagine.

Although seropositivity was extensive, it varied widely in the five prabhag s tested, ranging from 35.8 per cent to 66.4 per cent. As noted in the earlier report on the survey, seropositivity was higher in people living in crowded conditions, such as slums, compared with those living in non-slum areas, and was lower among those above 65 years of age (38.4 per cent). Correspondingly, the infection fatality rate (IFR) was inversely related to seroprevalence; it was highest (0.38 per cent) in the prabhag with the lowest seropositivity rate (Kasbapeth-Somwarpeth) and lowest (0.23 per cent) in the prabagh with highest rate (Lohiyanagar-Kasewadi). And older age groups had higher IFRs compared with younger age groups. While the overall IFR was 0.28 per cent, the higher age groups had IFRs that were significantly above 1 across all the prabhag s.

Notably, the study found considerable heterogeneity in seroprevalence between clusters, even between proximate ones, suggesting that the dynamics of transmission are dependent on the demographic details of the cluster. Significantly, the incidence of COVID-19 after the survey has been found to be lower in areas that had higher seroprevalence. “Declining infection rates in areas with high seropositivity suggest population-level protection [resulting in modification of transmission] and substantial neutralising activity in asymptomatically infected individuals,” notes the research paper.

Aurnab Ghose said in his email: “In terms of community-level protection, what we show is that there is a reduction in the speed of subsequent viral transmission associated with localities showing high levels in infection (seropositivity). This is what is expected if a large number of individuals have had prior infection and have acquired some level of immune protection from reinfection. We are not saying, as some news agencies have chosen to report, that some parts of Pune have attained ‘herd immunity’. We have just seen the above-mentioned trend….

“The virus does not care about administrative boundaries but rather circulates in communities. Hence, with areas barely 2-3 kms apart you see seropositivity varying from 35 to 65 per cent, and within these locations seropositivity varies between crowded hutments and apartments/bungalows within metres of each other. We also cannot tell, without longitudinal follow-up, how long the individual protection lasts and neither can we treat localities as insulated communities with no mixing. The only way to achieve community protection is vaccination.”

Ability to inhibit binding tested

To study this inhibition activity in seropositive individuals, the researchers randomly selected a subset of 697 samples from the 857 RBD-IgG positive sera and, using a specific ELISA-based SVNT, tested for their ability to inhibit the binding of the RBD of the virus to human ACE2 receptor. It was found that 85.1 per cent the 697 RBD-IgG positive sera were also SVNT positive, but not 100 per cent. The researchers also found that there was considerable heterogeneity with age in this association between RBD-IgG positivity and inhibition of RBD binding to the ACE2 receptor.

Interestingly, the correlation was found to increase with age. While the correlation was found to be only 57.2 per cent in the age group 18-30, it was 64 per cent in the age group 31-50 and 75 per cent in people above 50 years. “This is quite unexpected, especially when RBD-IgG levels are not associated with age,” said Aurnab Ghose, “but there are some plausible explanations for this, which need to be investigated further. The biology is quite complex and there are factors that can come into play besides RBD-IgGs to prevent the binding. This cannot, therefore, be taken to imply that you have better protective immunity as you age.”

Confounding observations

Besides this age-related heterogeneity (and complexity), there were other confounding observations with regard to this correlation. It was found that there were samples that had high RBD-IgG levels but low binding inhibition and low RBD-IgG levels but high binding inhibition. Overall, 30 per cent of the positive sera showed departures from a straightforward correlation, with the former group being 6.2 per cent, which suggests that a significant number of seropositive people may not be well protected despite having high titres of RBD-IgG during the survey. According to the paper, a possible explanation for the low RBD-IgG/high inhibition group (which constitutes the remaining about 24 per cent) is the presence of non-IgG antibodies that also bind to the RBD, like the IgAs (immunoglobulin A).

To investigate this aspect, the researchers tested a subset of 639 samples for IgAs against the S1 part of the S-protein, which contains the RBD. It was found that 68.4 per cent of the RBD-IgG positive sera were also S1-IgA positive. S1-IgA was also found to be positively correlated to surrogate virus neutralisation, or SVNT positivity. In particular, the low RBD-IgG/high inhibition category of sera was found to be S1-IgA positive. These observations taken together imply that in the prevention of RBD binding with the ACE2 receptor or, equivalently, giving protective immunity against reinfection, both IgG and IgA antibodies play a role, the latter less but in a significant way. “However, this functional response shows considerable heterogeneity, with patterns of variation suggesting age-associated diversity in the antibody repertoires,” the researchers say in the paper. “More likely,” Aurnab Ghose added, “it is to do with the specific epitopes [sites on the antigen to which antibodies attach] and binding kinetics of the generated antibodies. But biologically it is important to note this heterogeneity in the immune response.”

To validate the SVNT and to verify the correlation of RBD-IgG with neutralising activity, the researchers also evaluated the neutralising activity by running a smaller subset of 192 samples (essentially because of the cost involved) through a PRNT. Both RBD-IgG and SVNT positivity correlated well with high live virus neutralising activity. Thus, in general, elevated RBD-IgG (and surrogate neutralisation) levels are “robust predictors of virus neutralising activity”, notes the paper. In sum, therefore, a sero-survey carried out with an RBD-specific IgG test, such as the THSTI-RBD-ELISA test, is a better proxy for community-level protective immunity than sero-surveys done with other antibody detection kits.

“This study,” notes the paper in conclusion, “demonstrates the heterogeneous but widespread transmission of SARS-CoV-2 in a dense urban population. The high seroprevalence combined with evidence of neutralising antibodies and of declining incidence in settings of high prevalence raises the possibility of population immunity at least in the short term.”