SARS coronavirus 2-reactive antibodies in bovine colostrum

Abstract

Objective

To determine if bovine colostrum and sera have antibodies that react with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2).

Animals

Dairy and beef cattle from North America and Europe, sampled before and after the SARS-CoV-2 pandemic.

Procedures

Indirect ELISAs using whole bovine coronavirus (BCoV) and SARS-CoV-2; whole SARS-CoV-2 Spike 1, Spike 2, and nucleocapsid proteins; and SARS-CoV-2-specific nucleocapsid peptide as antigens. Virus neutralization assay for BCoV. Surrogate virus neutralization assay for SARS-CoV-2.

Results

Antibodies reactive to BCoV were highly prevalent in samples collected from cattle before and after the SARS-CoV-2 pandemic. Antibodies reactive with SARS-CoV-2 were present in the same samples, and apparently increased in prevalence after the SARS-CoV-2 pandemic. These antibodies had variable reactivity with the spike and nucleocapsid proteins of SARS-CoV-2 but were apparently not specific for SARS-CoV-2.

Conclusions

Bovine coronavirus continues to be endemic in cattle populations, as indicated by the high prevalence of antibodies to the virus in colostrum and serum samples. Also, the prevalent antibodies to SARS-CoV-2 in bovine samples, before and after the pandemic, are likely the result of responses to epitopes on the spike and nucleocapsid proteins that are shared between the 2 betacoronaviruses. Cross-reactive antibodies in bovine colostrum could be examined for prophylactic or therapeutic effects on SARS-CoV-2 infections in humans.

Introduction

The emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and the disease syndrome causally associated with it, coronavirus infectious disease 2019 (COVID-19), aroused large segments of the human medical community and the general public to the importance of coronaviruses as multi-systemic pathogens. Although it is rarely acknowledged, this awakening occurred nearly 100 y ago in the veterinary medical community, with the seminal studies of the prototypical coronaviruses, mouse hepatitis virus (MHV), and infectious bronchitis virus (IBV) (1).

Due to the status of SARS-CoV-2 as an apparently bat-derived zoonosis, there has been concern since the beginning of the SARS-CoV-2 pandemic that domestic animals, pets, and livestock could be reservoirs for the virus (1). “Species-jumping” is a well-studied phenomenon in veterinary coronavirology (1). Initial in silico studies comparing inter-specific sequence identities among genes coding for angiotensin-converting enzyme 2 (ACE-2), the major receptor for the SARS viruses, revealed variable similarity among the sequences that cannot necessarily be predicted based on the genetic relatedness of the animals carrying the homologous genes (2). Subsequently, numerous attempts have been made to experimentally infect various animal species, including cattle, with SARS-CoV-2. Compared to some other species, there is a “medium” level of sequence identity between genes coding for ACE-2 in humans and cattle (2). Nevertheless, several studies have reported no infection, abortive infection, or minimal prevalence of transient infection in experimentally infected cattle (3–5). However, based on the detection of antibodies to the receptor binding domain (RBD) of the SARS-CoV-2 S1 protein, a recent report suggested that the virus may rarely cause “spillover” infection into cattle (6). Collectively, the available data indicate that cattle are unlikely to be become ill, or even be productively infected, as a result of natural (low-dose) exposure to SARS-CoV-2, and they are unlikely to transmit SARS-CoV-2 — at least the viral strains circulating at the time of this writing — to humans (3–7).

Among the many interesting phenomena that have emerged from the epidemiology of SARS-CoV-2 infections is that one of the reasons for differences in clinical outcomes and severity of COVID-19 is preexisting immunity to infection: a cross-reactive immunity resulting from previous coronavirus infections (8). Since the beginning of the recent SARS-CoV-2 pandemic, few humans have been exposed to SARS-CoV-1 or middle eastern respiratory syndrome (MERS)-CoV, which are the closest relatives of SARS-CoV-2. It is thus unlikely that cross-reactive immunity has resulted from responses to these viruses. Among the viruses most likely to induce SARS-CoV-2 cross-reactive immune responses are other, less closely related, betacoronaviruses, the human cold viruses. It is thought that BCoV is the progenitor of both human and canine cold viruses, HCoV-O43 and canine respiratory coronavirus (CRCoV), respectively; or that all 3 viruses arose from a common ancestor (they have a sequence identity of > 96%) (9,10). Therefore, responses to the animal viruses, without actual cross-infection, could play a role in the complex epidemiology of SARS-CoV-2 infection.

We hypothesized that, based on their repeated exposure to BCoV, a betacoronavirus common in cattle populations, cattle have antibodies that are reactive with SARS-CoV-2. This investigation was undertaken to examine that possibility and to better understand the complex epidemiology of coronavirus infections.

Materials and methods

Colostrum and serum samples

Fresh liquid colostrum samples were obtained by direct collection from individual cows, or by sampling frozen colostrum from single cows that was stored at –20°C in individual 5-gallon pails. Frozen colostrum samples were collected in the years 1999 to 2001 from 1064 dairy cows at 134 farms in Saskatchewan, and in 2017 from 25 cows at a commercial dairy in New York state, USA.

Batches of spray-dried colostrum powder [Saskatoon Colostrum Company Limited (SCCL), Saskatoon, Saskatchewan] were produced from pools of colostrum taken from approximately 8000 individual cows from farms in each of 4 countries: Canada, the USA, Scotland, or the Netherlands. In each country, colostrum from cows at the same farms was collected before the SARS-CoV-2 pandemic, in 2018 or 2019; and after the onset of the pandemic, in 2021 or 2022. Colostrum powder was rehydrated at a concentration of 3 g of powder in 10 mL of sterile water. Before testing, both the liquid and rehydrated colostrum samples were centrifuged at 1500 × g for 15 min. Samples for testing were taken from beneath the resulting upper fat layer. Immunoglobulin G (IgG) concentrations were determined by single radial immunodiffusion (11) and equalized to 0.5 or 1 mg/mL.

In 2021, as a surrogate for obtaining colostrum samples from commercial pastured beef cattle, 36 approximately 1-day-old calves from a herd of multiparous Angus crossbred cows in Saskatchewan that had not received BCoV-containing vaccines for 5 yr were bled by jugular venipuncture after natural suckling.

Two human serum samples were voluntarily obtained from and tested (under biosafety level 2 containment) by the same individual (JE). The samples were collected by a trained phlebotomist both before [in 1999 (12)] and after (in 2022; convalescent sample) the onset of the COVID-19 pandemic. The convalescent sample was obtained approximately 4 wk after a SARS-CoV-2 antigen-confirmed mild acute infection that was contracted after receiving 3 SARS-CoV-2 mRNA vaccinations. These human samples were used as positive controls in all assays.

Antibody assays

ELISA for BCoV-reactive antibodies

An indirect ELISA for BCoV-reactive antibodies was conducted according to standard procedures (13). Briefly, BCoV antigen was sucrose density gradient purified from infected human rectal carcinoma cells. Tissue culture control antigen was prepared in a similar manner from uninfected human rectal carcinoma cells. The optimal concentrations of antigens coated onto 96-well ELISA plates (Immunlon-4HBX; Thermo Fisher Scientific, Waltham, Massachusetts, USA) were determined in a checkerboard analysis using positive and negative control sera. The sera were tested at a dilution of 1/50. Binding of primary antiserum was detected using a 1/5000 dilution of peroxidase-conjugated (Staphylococcus sp.) protein G (Invitrogen, Waltham, Massachusetts, USA), which binds to IgG of numerous species, including cattle and humans (14). This step was followed by reaction with a 2,2′-azino-bis(3-ehtylbenzothiazoline-6-sulfonic acid) (ABTS) substrate (SeraCare Life Sciences, Milford, Massachusetts, USA). Optical densities (OD) were read at 415 nm on an automated ELISA reader (iMark; Bio-Rad Laboratories, Mississauga, Ontario). The units for the ELISA were calculated as follows (13):

$$\begin{array}{l}[\text{net\hspace{0.17em}OD\hspace{0.17em}test\hspace{0.17em}sample\hspace{0.17em}}(\text{OD\hspace{0.17em}antigen-OD\hspace{0.17em}tissue\hspace{0.17em}culture}\\ \text{control})/\text{net\hspace{0.17em}OD\hspace{0.17em}positive\hspace{0.17em}control\hspace{0.17em}serum\hspace{0.17em}}(\text{OD\hspace{0.17em}antigen-OD}\\ \text{tissue\hspace{0.17em}culture\hspace{0.17em}control})]\times 100=\text{units}\end{array}$$

The positive control serum (A/S) from a mature cow with high reactivity in ELISAs for BCoV was considered 100 units.

ELISA for SARS-CoV-2-reactive antibodies

An indirect ELISA for SARS-CoV-2-reactive antibodies was conducted similar to that described for BCoV-reactive antibodies, according to standard procedures (13). Briefly, the OD in wells of ELISA plates coated with sucrose density gradient-purified inactivated SARS-CoV-2 from infected Vero cells (ZeptoMetrix, Buffalo, New York, USA) was compared to the OD in wells coated with tissue culture control antigen prepared in a similar manner from uninfected Vero cells. The human convalescent serum was used as the positive control; fetal bovine serum was used as a negative control.

ELISAs for SARS-CoV-2 S1, S2, and N protein-reactive antibodies

Indirect ELISAs for antibodies reactive with the Spike 1 (S1), Spike 2 (S2), and nucleocapsid (N) recombinant whole proteins of SARS-CoV-2 were conducted according to standard procedures (13). Briefly, 50 μL of reconstituted recombinant N (SARS-CoV-2 nucleocapsid-His; Invitrogen), S1 (Spike S1-His fusion protein; Invitrogen), or S2 (Spike S2 His-tag; Abcam, Cambridge, UK) proteins were coated onto ELISA plates (Immulon 4) at optimal concentrations that were determined in checkerboard analyses using monospecific rabbit antibodies for N (Fisher Scientific, Nepean, Ontario), S1 (Fisher Scientific), and S2 (Abcam) as positive controls. For each reaction, 100 μL of a 1/20 dilution of reconstituted colostrum powder were added, and reactions were visualized as described in the materials and methods herein. Results are reported as OD values compared to SARS-CoV-2 convalescent human serum and blank wells (containing reagents only).

Surrogate virus neutralization test

An ELISA-based surrogate virus neutralization test (cPass SARS-CoV-2 Neutralization Antibody Detection Kit; GenScript, distributed by Fosun Pharma, Princeton, New Jersey, USA) was conducted on selected samples according to the manufacturer’s instructions. This assay allows the detection of SARS-CoV-2-neutralizing antibodies by mimicking the interaction between the viral RBD of the S1 protein and the ACE-2 receptor, and has previously been shown to be highly specific and moderately sensitive using sera from veterinary species, including cattle (6,15).

Capture ELISA for SARS-CoV-2 nuclear protein-specific antibodies

A 1-step capture ELISA (Platelia SARS-CoV-2 Total Ab; Bio-Rad Laboratories, Mississauga, Ontario) was conducted according to the manufacturer’s instructions and using the appropriate dilutions of serum and colostrum to detect total antibodies to recombinant SARS-CoV-2 nuclear protein. According to the manufacturer’s testing, this assay is specific for antibodies to SARS-CoV-2 and does not detect antibodies to other coronaviruses — notably, human coronavirus OC43 — or to a range of other viral and bacterial pathogens.

Virus microneutralization assay

A microneutralization assay to detect BCoV-neutralizing antibodies was conducted using standard procedures (13) and validated, in-house protocols at the Animal Health Laboratory, University of Guelph (Guelph, Ontario).

Results

High prevalence of BCoV-reactive antibodies in colostrum and serum

There was a high prevalence of high titers of BCoV-reactive antibodies among the colostrum samples collected between 1999 and 2001 from the 1064 dairy cattle in Saskatchewan (Figure 1). Similarly, all 25 dairy cattle sampled in 2017 had moderate to high BCoV-reactive antibody concentrations (Table 1).

Figure 1.

Distribution of bovine coronavirus (BCoV)-reactive antibody concentrations in colostrum collected between the years 1999 and 2001 from 1064 dairy cows at 134 farms in Saskatchewan.

Table 1.

ELISA detection of BCoV- and SARS-CoV-2-reactive antibodies in colostrum collected in 2017 from 25 individual cows in New York state, USA.

	BCoV units	SARS-CoV-2 units
	113	216
	65	85
	65	2
	88	30
	72	19
	68	13
	95	22
	68	16
	62	16
	51	24
	70	22
	72	34
	62	23
	45	28
	59	20
	52	40
	83	42
	61	30
	72	15
	49	17
	100	1
	77	0
	64	195
	73	22
	70	13
Mean	70	38
Range	45 to 113	0 to 216

ELISA testing used sucrose-gradient purified whole viruses as antigens.

BCoV — Bovine coronavirus; SARS-CoV-2 — Severe acute respiratory syndrome coronavirus 2.

All 36 post-suckle serum samples obtained from calves in 2021 had variably high concentrations of BCoV-neutralizing antibodies [median (reciprocal) ± standard deviation: 2048 ± 2626; range: 256 to 8192], suggesting high concentrations of BCoV antibodies in the colostrum of beef cows.

Reactivity of bovine antibodies in SARS-CoV-2-reactive ELISAs

Compared to the SARS-CoV-2 convalescent serum, the 25 individual dairy cows had a broad range of concentrations of SARS-CoV-2-reactive antibodies (mean: 35.6 units; range: 0 to 216 units; Table 1). There was no apparent relationship between the units of BCoV-reactive antibodies and SARS-reactive antibodies when comparing units derived from the 2 whole-virus ELISAs (Table 1). All pools of colostrum from cattle in North America and Europe collected before and after the emergence of SARS-CoV-2 contained antibodies that reacted with the virus (Figure 2). After the start of the pandemic, there was an apparent increase in reactivity in each sample of pooled colostrum. Antibodies in each of these pools were variably reactive with epitopes on the S1, S2, and N proteins of the virus (Figure 3).

Figure 2.

Reactivity of pools of rehydrated spray-dried colostrum in ELISAs using sucrose gradient-purified whole inactivated severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) as the antigen. Pools of colostrum were collected before (pre) and after (post) the start of the SARS-CoV-2 pandemic. Units were determined by comparison with a human SARS-CoV-2 convalescent serum (100 units).

Figure 3.

Reactivity of pools of rehydrated spray-dried colostrum in ELISAs using severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) Spike 1 (S1), Spike 2 (S2), and nucleocapsid (N) proteins as the antigens. Pools of colostrum were collected before (pre) and after (post) the start of the SARS-CoV-2 pandemic and were compared with human pre-pandemic and post-SARS-CoV-2 infection convalescent serum. Optical density values in “blank” wells were < 0.100.

Lack of reactivity of bovine antibodies in SARS-specific ELISAs

All of the same rehydrated samples from the same pools of spray-dried colostrum collected from Canada, the USA, Scotland, and the Netherlands before and after the onset of the SARS-CoV-2 pandemic and tested previously, had a similar IgG content when rehydrated (~40 g/L) and were tested at a dilution of 1/20 according to the manufacturer’s recommendations. All samples were negative in both the cPass ELISA test specific for the SARS-CoV-2 RBD of the S1 protein, and the capture ELISA test for a SARS-CoV-2-specific peptide from the N protein.

Discussion

Forty years ago, before the advent of BCoV vaccines, in the first application of ELISA technology to detect antibodies to BCoV in cattle (16), a high prevalence of antibodies was reported in 100 cattle from 12 farms (beef or dairy not indicated) in western Canada. In their report, the authors stated, “It was difficult to find a negative animal.” Subsequently, aside from experimental infections, most investigations of BCoV seroprevalence have been related to seroconversion in the context of respiratory disease: with age and comingling, most cattle develop antibodies to the virus (17). This high prevalence of seroconversion parallels the situations with the other 2 betacoronaviruses in the BCoV cluster, HCoV-O43 in human hosts (18) and CRCoV in canine hosts (10,19), and is consistent with endemic infection. In contrast to a relative abundance of data related to seroconversions resulting from natural exposure to betacoronaviruses, published data concerning coronaviral-specific antibodies in bovine colostrum are scarce, especially as related to the cross-reactive specificities of colostral antibodies (20). The data in this report document a high prevalence of BCoV-reactive antibodies in colostrum samples collected 20 y ago from western Canadian cattle that parallels findings in contemporaneous colostrum samples from unvaccinated cattle in the UK (20). Indicative of the current situation, the high prevalence of BCoV-reactive antibodies in recent samples from both dairy and beef cattle in North America and Europe confirm and extend the seminal observations. Altogether, the data document that BCoV has been, and still is, an endemic infection in cattle populations with a coincident high prevalence of seroconversion. How coronaviral vaccine use may be affecting the prevalence and magnitude of BCoV antibodies in cattle populations is largely unknown. It is estimated that some form of BCoV vaccine (oral or intranasal in calves and injectable in cows) is currently used in 30 to 50% of the dairy herds in Canada and the USA that were the sources of the colostrum pools tested [Richard Hupaelo, Saskatoon Colostrum Company Limited (SCCL), Saskatoon, Saskatchewan; personal communication, 2022]. Limited recent published data from western Canadian cow-calf (beef ) herds indicate that about 50% of herds use at least 1 dose of a BCoV-containing vaccine in breeding animals (21).

Our results confirm and extend the recent report by Oshiro et al (22) that IgG purified from milk-derived whey collected before the SARS-CoV-2 pandemic from dairy cattle in New Zealand contained antibodies that were reactive with SARS-CoV-2-synthesized peptides from the S and N proteins — including peptides overlapping the RBD on the S1 protein. Adding to that, our data suggest that SARS-CoV-2-reactive antibodies in cattle are highly prevalent, not unique to a particular locale, and not the result of unique environmental conditions such as geographically related zoonotic exposure to SARS-CoV-2 “reservoirs” (e.g., Asian bats). The apparent lack of agreement between the ELISAs using equal amounts of sucrose gradient-purified BCoV and SARS-CoV-2 indicate that individual cattle differentially recognize epitopes shared between the 2 viruses. This could be the result of genetic differences in response or differences in immunologic experience with BCoV. Furthermore, our negative results in the surrogate neutralization and 1-step capture ELISAs for antibodies against regions of SARS-CoV-2 S1 RBD and N protein, respectively, that are not conserved (among other coronaviruses) but rather unique to that virus, suggest that the SARS-CoV-2-reactive antibodies detected in the bovine colostrum and serum samples we tested did not result as a consequence of the productive infections of those cattle with SARS-CoV-2. The antibodies are reactive with epitopes that are conserved, at least among betacoronaviruses. Related to this, these results are compatible with available data indicating that cattle are unlikely to be productively infected with SARS-CoV-2. Therefore, SARS-CoV-2-reactive antibodies detected in bovine samples tested are almost certainly cross-reactive antibodies resulting primarily from exposure to endemic BCoV infection — plus, perhaps, some undetermined effect of vaccination against BCoV (20). Nevertheless, when comparing pooled bovine colostrum samples collected from the same herds (although from mostly different cows) before and after the start of the SARS-CoV-2 pandemic, it was interesting that there was an apparent increase in SARS-CoV-2-reactive antibodies in colostrum from cattle in North America and Europe. We cannot exclude the possibility that this increase, if it was real and not simply the result of testing different cows at different times, was due to exposure of the cattle to SARS-CoV-2-infected humans in the close quarters of the dairy operations. In that case, the increase in SARS-CoV-2-reactive antibodies could be the result of “original antigenic sin” — a predominant response to shared epitopes upon exposure to a related virus (23).

From a comparative, One Health-focused medical standpoint, our results are important for at least 2 reasons. First, the detection of SARS-CoV-2-reactive antibodies in bovine samples parallels the situation in humans, including the positive results with the pre-pandemic human sample tested herein, as a probable result of exposure to the endemic betacoronavirus HCoV-O43 (18). Moreover, although not examined here, there were at least antecedent cross-reactive CD4+ T-cell responses necessary to produce SARS-CoV-2 antibodies in cattle — and perhaps also CD8+ T-cell responses. The application of the indirect ELISAs using SARS-CoV-2 proteins indicates that among the epitopes recognized are those comprising at least the S1, S2, and N proteins of BCoV. By extension, human exposure to the betacoronaviruses that are closely related to HCoV-O43, BCoV and CRCoV, may induce similar cross-reactive responses (24). In other words, these data from cattle further support the concept that both B-cell and T-cell responses cross-reactive with SARS-CoV-2 can be induced by exposure to other betacoronaviruses. The fine specificities of these responses are currently under investigation in humans (24). Furthermore, consistent with our data, recent in silico studies have predicted several shared peptides between the BCoV and SARS-CoV-2 S, N, M, and ORF1-encoded (nonstructural) proteins that could serve as T-cell (major histocompatibility complex Class 2) and B-cell epitopes, and suggest the possibility that previous exposure to cattle infected with BCoV could induce cross-protective responses (25).

Second, these results provide insight into the possible utility of bovine colostrum as an adjunct preventative or therapy for COVID-19 in humans. Indeed, clinical trials using bovine colostrum as a therapeutic intervention in COVID-19 patients are currently underway (26). Any specific anti-SARS-CoV-2 effects of SARS-CoV-2-reactive antibodies in cattle are unresolved. It is interesting that, in one study of cattle experimentally infected with SARS-CoV-2, the authors concluded that the observed mild clinical infections occurred despite high concentrations of antibodies to BCoV (3). Another interpretation of those data could be that cross-reactive antibodies may have at least partially affected the clinical outcome without preventing infection. The negative results in the RBD-binding surrogate neutralization ELISA suggest that antibodies to BCoV are unlikely to neutralize SARS-CoV-2 infectivity by “blocking” the interaction between the S1 subunit of the viral spike protein with the ACE-2 receptor. There are, however, mechanisms of virus “neutralization” other than blocking of attachment to the receptor — including, notably, blocking of fusion (27), which could be mediated by antibodies to the S2 subunit of the SARS-CoV-2 spike protein (28). Indeed, the possibility of using S2 as an immunogen to induce cross-reactive responses is currently under investigation (27).

Before the emergence of COVID-19, numerous studies in a variety of species suggested that bovine colostrum could have ameliorative effects on respiratory disease and enteric diseases similar to those comprising the clinical syndrome of COVID-19 (29). From a mechanistic standpoint, in the respiratory tract, most of the benefits of ingesting bovine colostrum in species other than cattle are attributed to the non-specific antimicrobial activity of non-immunoglobulin components of colostrum, notably lactoferrin; whereas in the gut, in addition to antimicrobial activities, colostral growth factors are thought to aid in the repair of tissue damage (29,30). Our data suggest that, in addition to these components, cross-reactive antibodies — notably, those against the S2 subunit of the spike protein — could have disease-sparing effects, especially if they are present in the oropharynx at the time of SARS-CoV-2 infection (25,28). This possibility warrants further investigation.

In conclusion, our results, which focus on colostral antibodies, confirm and extend the data related to the endemicity of BCoV in cattle, and provide further insight into the cross-reactive responses generated during the course of betacoronaviral infections in their target species.