Introduction
Since the World Health Organization’s declaration of COVID-19 as a pandemic in March 2020 [
13], the world has seen the emergence of a multiple variations of the original virus, five of which have been declared as variants of concerns (VOCs) and many more considered to be variants of interest. These SARS-CoV-2 variants harbor accumulations of mutations, predominantly within the spike glycoprotein, which are associated with increased viral transmissibility and escape immunity [
9]. The ability for these mutations to permit escape from host immunity raised the question of immune memory durability, which could lead to reinfections and break through infections [
17,
38]. While it is known that antibody levels decline over time following vaccination or infection [
21,
30], an individual’s protective immune responses are also influenced by the strain or combination of strains to which they have been exposed [
42,
58]. In certain cases, cross-reactive immune responses from re-infections may positively impact disease outcomes by promoting beneficial cross-reactive T cell responses [
34,
40], but, in other cases, such as dengue virus and zika virus infections disease severity is potentially exacerbated [
15,
61]. Now more than 3-years into the pandemic, people across the world have very different patterns of immunity to the SARS-CoV-2 virus, based on their exposure and vaccination. Globally, people have been exposed to the original strain (hereon referred to as Ancestral) and/or Alpha, Beta, Gamma, Delta and now Omicron (and all the Omicron subvariants). People may be unvaccinated or have had one or more vaccine doses [
4]. The particular virus strain they are first exposed, whether by natural infection or vaccination, may also shape their subsequent immune patterns, affecting their susceptibility to infection to heterologous strains and disease severity [
2,
31,
35]. The different mutations in the spike protein shape the subsequent antibody and T cell responses [
25], which can result in reduced or enhanced responses to variants further down the line. This has important implications for future proofing vaccine design and dosing strategies.
Although SARS-CoV-2 is considered as pathogenic [
18], COVID-19 shows a diverse range of symptoms from majority of patients reported as asymptomatic or mild disease, to severe cases that include acute respiratory distress syndrome, pneumonia, cardiac arrythmia, encephalopathy, and death [
55]. COVID-19 severity and disease outcomes have been associated with various immunological correlates [
5,
57] that include autoantibodies to type I interferons [
52]; altered myeloid cell populations such as elevated numbers of immature neutrophils and loss of non-classical monocytes [
45,
60]. Increased hyperinflammatory responses and aberrant CD163 + monocytes have also been reported [
19]. Elevated serum cytokine levels are also a strong predictor of severe COVID-19 and adverse disease outcomes [
16]. Similarly, certain T cell responses including unconventional CD16 + T cells, mucosal-associated invariant T (MAIT) cells and γδ T cell may contribute to immunopathology observed in increased COVID-19 disease severity [
36]. Conversely, the development of a coordinated SARS-CoV-2-specific CD4 + and CD8 + T cell response and neutralizing antibodies are associated with reduced COVID-19 disease severity [
44,
46]. Most immunological observations were carried out early during the pandemic when the Ancestral strain was circulating. However, SARS-CoV-2 variance have exhibited variation in disease severity, patterns and waning of immunity, immune evasion and sensitivity to therapeutics [
9,
26].
Our study was initiated during the period of the pandemic when only three VOCs first emerged, so the investigation was limited to a focus on variants Alpha, Beta, and Gamma. We investigated the cross-protection in hamsters previously infected with a VOC and subsequently re-infected with a heterologous variant. We also determined if cross-protection and immunity was dependent on the specific virus to which the hamster was first exposed. We further profiled the host cytokine response induced by each SARS-CoV-2 variant as well as subsequent to re-infection. A comparative analysis of the three VOCs revealed that Alpha variant was the most pathogenic VOC to emerge. We showed that naturally acquired immunity protected hamsters from subsequent re-infection with heterologous SARS-CoV-2 variant, regardless which variant the animal was first exposed to. Our study supports observations that heterologous infection of different SARS-CoV-2 variants do not exacerbate disease in subsequent re-infections. The continual emergence of new SARS-CoV-2 variance mandates a better understanding of cross-protection and immune imprinting in infected individuals. This will contribute to our understanding of the heterogeneity of clinical outcomes in COVID-19 disease and allows us to identify conserved immune epitopes. Such information is essential to guide vaccine strategy and public policy to emerging SARS-CoV-2 VOCs and future novel pandemic coronaviruses.
Materials and methods
Animals and viruses
Golden Syrian hamsters, 7–8 weeks old (81–90 g) males/females were purchased from the Charles River Laboratories (Saint-Constant, Canada). Animals were maintained at the small animal facility of the National Research Council Canada (NRC) in accordance with the guidelines of the Canadian Council on Animal Care. All procedures performed on animals in this study were in accordance with regulations and guidelines reviewed and approved in animal use protocol 2020.06 by the NRC Human Health Therapeutics Animal Care Committee. Hamsters were anesthetized by injection of Ketamine/Xylazine (90 kg/mg/8 kg/mg) and intranasally challenged with 8.5 × 104 plaque forming unit (PFU) of SARS-CoV-2 (100 µl per animal) or 100 µl of sterile phosphate buffered saline (1X PBS) as control. Hamsters were challenged on 0 dpi (days post-infection; referred to as primary infection) and re-challenged (referred to as secondary infection) with a heterologous SARS-CoV-2 virus (either Ancestral, Alpha, Beta, or Gamma) on 21 dpi. Animals were euthanized by CO2 on days 2, 5, and 7 post primary infection. Selected tissues were collected at necropsy for histology and downstream analysis. All infectious work were conducted under approved containment level-3 (CL-3) conditions at the NRC's CL-3 facility.
The following SARS-CoV-2 isolates were used in this study: hCOV-19/Canada/ON-VIDO-01/2020 (B, Ancestral) (National Microbiology Laboratory, Winnipeg, Canada), hCoV-19/England/204820464/2020 (B.1.1.7, Alpha, NR-54000), hCoV-19/USA/MD-HP01542/2021 (B.1.351, Beta, NR-55282), hCOV-19/Japan/TY7-503/2021 (P.1, Gamma, NR-54982), hCoV-19/South Africa/CERI-KRISP-K040013/2022 (Lineage BA.5; Omicron Variant), hCoV-19/USA/MD-HP40900/2022 (Lineage XBB.1.5; Omicron Variant) were obtained through BEI Resources, NIAID, NIH. Viruses were propagated in Vero E6 cells and quantified in Vero cells. Sanger sequencing of the spike gene was carried out to confirm exact genetic identity to original isolate. Passage 2 or 3 virus stocks were used in all subsequent experiment.
Plaque assay
Virus burden was quantified by plaque assay within the NRC's CL-3 biocontainment facility. Whole left lung was homogenized in 1 ml of 1X phosphate buffered saline (PBS). The plaque assay, in brief, was carried out by diluting the clarified lung homogenate supernatant in a 1:10 serial dilution in infection media (1X DMEM, high glucose media supplemented with 1X non-essential amino acid, 100 U/mL penicillin–streptomycin, 1 mM sodium pyruvate, and 0.1% bovine serum albumin). Vero cells were infected for 1 h at 37 °C before the inoculum was removed and overlay media was added, which consisted of 1X infection media with 0.6% ultrapure, low-melting point agarose). The assay was incubated at 37 °C/5% CO2 for 72 h. After incubation, cells were fixed with 10% formaldehyde and stained with crystal violet. Plaques were enumerated and PFU was determined per gram of lung tissue.
Plaque reduction neutralization test (PRNT)
The PRNT assay was performed in the NRC’s CL-3 facility. Serum samples were inactivated at 56 °C for 30 min and stored on ice. The inactivated serum was serially diluted 1:2 and incubated with equal volume of 100 PFU of SARS-CoV-2 at 37 °C for 1 h, followed by infection of Vero cells. Adsorption of virus were carried out for 1 h at 37 °C. After adsorption, inoculum was removed and cells were overlaid with media as described above. The assay was incubated at 37 °C/5% CO2 for 72 h. Cells were fixed with 10% formaldehyde after incubation and stained with crystal violet. No serum, virus-only back-titer control was included along with naïve animal serum. PRNT50 is defined as the highest dilution of serum that results in 50% reduction of plaque-forming units. The 1:2 dilution of diluted serum to 100 PFU virus was included in the final calculation.
Microneutralization assay
The microneutralization assay was performed in the NRC’s CL-3 facility. Hamster serum samples were inactivated at 56 °C for 30 min and stored on ice. In brief, the inactivated serum was serially diluted 1:5 and incubated with equal volume of 125 PFU of SARS-CoV-2 at 37 °C for 1 h. After incubation, Vero-E6 cells seeded in 96 well plates were infected with sera/virus mixture and adsorption of virus was carried out for 1 h at 37 °C. After adsorption, inoculum was removed and cells were overlaid with media containing serum dilution. The assay was incubated at 37 °C/5% CO2 for 72 h. The plates were examined under a brightfield microscope for cytopathic effect (CPE). The microneutralization titer is determined as the highest dilution factor with no detectable CPE.
Cytokine profiling in hamsters using quantitative real-time PCR
Lung total mRNA was extracted using RNA miniprep kit (Cat # R1058, Zymo Research, Irvine, USA) according to the manufacture’s instruction. 500 ng of RNA was used to synthesize cDNA using reverse transcriptase Super Script III (Cat # 1808-044, Thermofisher, Ottawa, Canada). Cytokine profiling by qRT-PCR was performed in duplicate using SYBR Master mix (Cat # 434446, Applied BioSystems, MA, USA). Fold change gene expression was calculated using ΔΔCt against PBS (mock) infected hamsters as baseline with 18S rRNA as the housekeeping gene. Five animals were analyzed for each experiment. The primer sequences were designed as previously described [
56]
Quantitative real-time PCR of viral genomic RNA
Viral RNA from hamster lung tissues were extracted using Quick-viral RNA kit (Cat #R1035, Zymo Research, Irvine, USA). The extraction was done according to the manufacture’s instruction. Viral RNA was quantified by Luna Universal one-step RT-qPCR kit (Cat #, E3005S, New England Biolabs, MA, USA) with primer/probe sets specific designed for the SARS-CoV-2 E gene (F: ACAGGTACGTTAATAGTTAATAGCGT, R: ATATTGCAGCAGTACGCACACA, Probe: ACACTAGCCATCCTTACTGCGCTTCG [5′]Fam [3′]BHQ-1) Ct values were compared to the SARS-CoV-2 stock (Ancestral) which allowed us to quantify the levels of RNA. Finally, the results were presented as RNA copy numbers/g Lung tissue.
RBD-specific IgG ELISA
Nunc MaxiSorp flat- bottom 96 well plates were coated with recombinant SARS-CoV-2 RBD- His recombinant protein (40595-V80H, Sino Biological, China) and incubated overnight at 4 °C.
After the incubation, plates were washed with PBS containing 0.1% Tween-20 plates were blocked with 3% Bovine Serum Albumin (IgG-Free). Hamster serum was diluted (fivefold serial dilution) from 1:100 up to 1:1562500. Diluted serums were added to the plate and incubated for 1 h at 37 °C. Next, plates were washed with PBS-T and Peroxidase AffiniPure Goat Anti-Syrian Hamster IgG (H + L) (Cat # 107-035-142, Jackson Immuno Research, West Grove, USA) was added to each well and incubated at 37 °C for 1 h. After the last wash with PBS-T, 100 µL of Tetramethylbenzidine (TMB) substrate (Cat# 7004P6, Cell Signaling Technology, MA, USA) was added to each well. After a two-minute incubation at room temperature, 100 µL of Stop solution (Cat# 7002P6, Cell Signaling Technology, MA, USA) was added to terminate the reaction and absorbance was measured at 450 nm. Inhibitory dilution 50 (ID50) was calculated using non-linear regression analysis.
Histopathology and immunohistochemistry
All four lobes of the right lung of infected hamsters were isolated and immersed in 10% neutral buffered formalin. After fixation for 1 week at room temperature, lungs were transferred into 70% ethanol, processed and embedded in paraffin wax. The paraffin blocks were cut into 5 µm sections and placed on Superfrost Plus slides (Fischer Scientific). Sections were dried overnight and duplicate sections were subjected to hematoxylin and Eosin (H&E) or immunohistochemistry (IHC) staining.
For H&E staining a fully automated Leica ST5010-CV5030 system was used. Whole slide H&E images were scanned at 20× magnification on a Zeiss Axio Scan.Z1 digital slide scanner capable of brightfield imaging. Histopathology scoring was performed using the criteria described by lien et al. [
28]and the samples were blindly scored by a certified pathologist. Briefly, at scanning magnification the percentage of lung area affected by inflammation was estimated. Subsequently the distribution of lesions, extent of the inflammation as well as type of cell infiltrate was scored as described in Table
1.
Table 1
Lung histopathology scoring criteria
0 | No significant morphological changes |
1 | Minor inflammatory changes with sparse mononuclear cell infiltration, mainly peribronchial/bronchiolar and perivascular |
2 | More apparent interstitial mononuclear inflammatory infiltration, alveolus septa thickening, focal areas of consolidation |
3 | Increased infiltration of inflammatory cells, multiple focal consolidation, diffuse alveolar damage |
4 | Extensive collapse of alveolar spaces, alveolar septa thickening, more inflammatory cell infiltration in alveoli, larger areas of consolidation, diffuse alveolar damage |
5 | Similar findings to 4, but the lung tissue is almost completely consolidated |
For immunohistochemistry, a modified protocol F on the Bond-Max III fully automated staining system (Leica Biosystems, Wetzlar) was employed. All reagents from the Bond Polymer Refine Detection Kit (DC9800) were used. To characterize the immune cell infiltrates in infected lungs, primary antibodies that cross-react with hamster immune cell antigens [
43] were: rabbit polyclonal antibodies against IBA1 (ionized calcium binding adaptor protein, 1), MPO (myeloperoxidase, 1:1000, Dako A0398) and CD3 (1:500, Dako A0452). SARS-CoV-2 was detected using mouse anti-SARS-CoV-2 nucleocapsid monoclonal antibody (1:5000, R&D System MAB10474). Following deparaffinization and rehydration, sections were pre-treated with the Epitope Retrieval Solution 1 (ER1, Citrate buffer, pH 5.0) or Epitope Retrieval Solution 2 (ER2, EDTA buffer, pH 8.8) at 98 °C for 20 min. Epitopes were exposed using ER1 for MPO, IBA1 and SARS-CoV-2 whereas ER2 was used for CD3. After washes, non-specific endogenous peroxidases were quenched using peroxidase block for 5 min. Sections were washed again and then incubated for 15 min at room temperature with primary antibodies. A mouse-on-mouse superblock was applied for 15 min prior to addition of anti-SARS-CoV-2 nucleocaspid antibody (PowerVision IHC/ISH Super Blocking, Leica Biosystems PV6122). After addition of primary and subsequent washes, sections were incubated with polymer refine for 8 min at room temperature and developed with 3, 3′-diaminobenzidine (DAB) chromogen for 10 min. Sections were then washed and counterstained for 6 min with hematoxylin, dehydrated, cleared and mounted. Negative controls included omission of primary antibody and incubation with secondary antibody alone as well as lung tissue from naïve animals.
Image analysis
IHC slides were scanned at 20× magnification using a Zeiss Axio Scan.Z1 digital slide scanner capable of brightfield imaging. QuPath 0.3.2, an open-source software for bioimage analysis
https://qupath.github.io; [
3] was used to detect and count immune-positive cells in whole section brightfield images. Briefly, images were dragged and dropped into the created project folder. Image type was then set to brightfield (H-DAB), which identifies individual cells based on the sum of the hematoxylin and DAB channels. The entire section was carefully selected using the wand tool and immunopositive (stained with hematoxylin and DAB) and negative cells (stained with hematoxylin only) were assigned at this step. The number of immunopositive cells were then calculated using the following steps: Analyze_cell detection_positive cell detection_run. Annotations, as well as markup images of detected cells for visual validation, were then exported into an excel file and the data therein was used for graphing results. It is important to note that the parameters for optimal cell detection and analysis such as threshold value etc. were determined beforehand for each antibody stain, i.e., SARS-CoV-2 nucleocapsid protein, MPO, IBA1 and CD3, and those values were kept constant for all sections. The number of positive cells and area detected were used to calculate the average number of positive cells per mm2.
Statistics
Data were analyzed using GraphPad Prism® version 9 (GraphPad Software). Statistical significance of the difference between groups was calculated by one-way (ANOVA) followed by post-hoc analysis using Tukey’s (comparison across all groups) multiple comparison test. Data was log transformed (except for % neutralization and % body weight loss) prior to statistical analysis. For all analyses, differences were considered to be nonsignificant with p > 0.05. Significance was indicated in the graphs as follows: *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001.
Discussion
COVID-19 pandemic saw the emergence of multiple waves of different SARS-CoV-2 variants with accumulating mutations that proffered the virus with varying levels of fitness. As such, individuals have become infected multiple times with different variants but very few studies have been carried out to understand the effect of multiple heterologous infections within a host. While the WHO had declared the pandemic as officially over, the continued infection and emergence of new variants remain a risk to global health. In this study we characterized the different pathogenicity and immune profile of three different VOCs in Golden Syrian hamsters. We further investigated the effect of multiple infections within a single host, and whether the initial infection affects the immune response to subsequent infection with a heterologous variant. Our data demonstrated that Alpha variant was the most pathogenic variant compared to Ancestral, Beta, and Gamma. Infection with Alpha variant resulted in significant weight loss with some hamsters showing lethality (a survival rate of 77%), delayed viral clearance, and pronounced cytokine gene expressions. Thus far, only Alpha variant infections have shown lethality in hamsters compared to the three other SARS-CoV-2 viruses we have tested (Ancestral, Beta, and Gamma). Our findings are in agreement with the literature and support the evidence that the Alpha variant might be the most virulent variant to have emerged thus far as demonstrated in hamsters and mice [
22,
50,
51].
We demonstrated that induction of RBD-specific IgG titer was similar for all tested VOCs and Ancestral virus, unrelated to the pathogenicity of the virus. While we did not observe significant reduction in neutralization activity in most of the convalescent sera of animals infected with the Ancestral or any of the VOCs to each other, possibly due to the low number of animal samples used in the study. However, we do see a trend that supports prior studies suggesting a slight reduction in neutralizing capability of certain convalescent sera to specific isolates tested [
14,
33,
59]. Nevertheless, our observation provides evidence that cross-neutralization is retained for the most part for all heterologous strains.
When this study was conducted, Delta and Omicron variants had not emerged; therefore, we were not able to include these more recently emerged VOCs in our in vivo challenges. However, we did evaluate if the convalescent sera from these infected hamsters were able to neutralize the more recent Omicron sublineages BA.5 and XBB1.5. In line with reports that infection or vaccination with previous SARS-CoV-2 strains resulted in reduced neutralizing activity to these recent Omicron subvariants [
37,
53], we observed substantial decline in neutralizing activity to BA.5 and especially to XBB1.5. Only 1 out of 5 animals showed any detectable levels of neutralization to XBB1.5 for Ancestral and Alpha strains. Interestingly, sera from Beta-infected animals showed at least 2-log neutralization titer that is unchanged for both BA.5 and XBB1.5 with 3 or 4 animals out of 5 showing detectable levels of neutralization. Of the initial VOCs that emerged, Beta was reported to have the highest resistance to neutralization by Ancestral-elicited sera, suggesting antigenic differences [
10,
29]. Moreover, studies have determined that Omicron shared mutations K417N and N501Y with Beta [
8]. A study that analysed different antibody subsets elicited by Beta infection revealed that a certain population of those neutralizing antibodies retained the ability to also neutralize Omicron, targeting key mutations that are shared between the two variants [
41]. This supports our observation that Beta-elicited sera retains low levels of neutralizing activity to both BA.5 and XBB1.5.
The analysis of cytokine profiling and gene expression in lung tissue of infected hamsters indicate that each variant induced distinct cytokine and chemokine profiles. Infection by Alpha resulted in significant increase of several inflammatory cytokines such as CCL2, CXCL10 and IL-6 over a longer period of time compared to Ancestral strain. Gamma variant also showed sustained elevated CXCL10 and CCL2 chemokines compared to Ancestral infection. Elevated levels of CXCL10 and IL-6 has been shown to correlate with severe COVID-19 [
23]. CCL2 has a critical role in monocyte infiltration and furthering lung tissue damage; increased levels of CCL2 by Alpha infection is consistent with tissue damage observed in lung pathology of infected hamsters [
39]. High pulmonary expression of type III IFNs is detected in critical COVID-19 cases. IFN-λ is the dominant IFN produced in respiratory tissues against viral infection to suppress the viral spread [
6]. We observed equally high levels of IFN-λ expressed in the respiratory tissues infected with all tested variants. The levels of the IFN-λ peaked at day 2 post infection and reduced by day 7 post infection; however, even at day 7 post infection, the levels of IFN-λ remain the highest compared to other IFNs and cytokines.
We profiled immune infiltration in the respiratory tissues by IHC. Due to a lack of reagents to carry out in-depth immune characterization in hamsters, our investigation was limited to a few immune markers for activated macrophages, neutrophils, and T lymphocytes. Our data indicated a stronger neutrophil response upon Alpha infection. Neutrophils are one of the earliest immune cells to be recruited and activated during viral infection. Previous study has shown an increase in the number of circulating neutrophils in lung tissue of COVID-19 patients; this increase in number of neutrophils was correlated to severity of COVID-19 [
27]. Therefore, an increased number of neutrophils in lung tissue by Alpha infection may indicate a higher pathogenicity of Alpha variant compared to other variants. T cell response plays a key role in protection against severe COVID-19 disease; moreover, with the emergence of SARS-CoV-2 variants of concern and their capability to escape neutralizing antibodies; T cell immune response is critical in containing the infection [
54]. Our IHC data shows that there was a notable increase in CD3
+ T lymphocytes infiltration in the lung tissue at 5 dpi but little was detected earlier. Importantly, we observed different levels of T lymphocyte infiltration from different variants. Beta and Gamma induced stronger CD3
+ T lymphocyte response compared to Alpha and Ancestral infection. Prior study has shown that T cell immunity is not disrupted by the mutations in variants [
49]; the differences in T cell response detected in our IHC study might reflect the longer time needed for alpha and ancestral infection to produce T cell immunity. Further investigations are needed to study the molecular mechanism underlying these immunological differences triggered by different SARS-CoV-2 variants of concern and to better understand the host-virus interaction. These findings provide insights into immunomodulatory interventions that can be used as therapeutics to treat COVID-19.
Previous studies have shown that primary infection in hamsters protects the animal from re-infection [
7,
11,
20]. Our re-infection data is in line with prior observations that reinfection results in lower viral loads and limited viral replication in hamsters. Re-infected hamsters did not lose any noticeable weight compared to primary infection where animals lost up to 18% of their weight by 5 dpi. SARS-CoV-2 viral load was detected in lung tissue of re-infected hamsters; however, the quantified viral load was substantially lower compared to primary infection, which suggests that initial infection with SARS-CoV-2 provides a certain degree of immunity to secondary infection with another SARS-CoV-2 variant. Significantly, we observed that immunity acquired from the initial infection greatly reduced the infiltration of immune cells into the respiratory tissue and also promoted strong CD3
+ T lymphocyte recruitment to the area of infection, regardless of which variant caused the initial infection.
Our data with Ancestral, Alpha, Beta and Gamma, strongly suggests that naturally-acquired immunity provides protection from subsequent infection with a heterologous SARS-CoV-2 strain. This would minimizes disease severity, viral burden, as well as reduces aberrant immune response that lead to severe pathology. Our study provides evidence that infection with a heterologous strain confers protection with no observable enhancement of disease. Moreover, our observations underscore the need to better understand the pathogenicity and host immune response to each variance in order to predict the nature of new emerging SARS-CoV-2 variants, which would permit more informed decisions about future vaccine design, strategy and intervention.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.