Antibody-independent protection against heterologous SARS-CoV-2 challenge conferred by prior infection or vaccination

Abstract

Vaccines have reduced severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) morbidity and mortality, yet emerging variants challenge their effectiveness. The prevailing approach to updating vaccines targets the antibody response, operating under the presumption that it is the primary defense mechanism following vaccination or infection. This perspective, however, can overlook the role of T cells, particularly when antibody levels are low or absent. Here we show, through studies in mouse models lacking antibodies but maintaining functional B cells and lymphoid organs, that immunity conferred by prior infection or mRNA vaccination can protect against SARS-CoV-2 challenge independently of antibodies. Our findings, using three distinct models inclusive of a novel human/mouse ACE2 hybrid, highlight that CD8⁺ T cells are essential for combating severe infections, whereas CD4⁺ T cells contribute to managing milder cases, with interferon-γ having an important function in this antibody-independent defense. These findings highlight the importance of T cell responses in vaccine development, urging a broader perspective on protective immunity beyond just antibodies.

Fig. 1. Antibody-independent protection in K18-hACE2 transgenic mice.

a, Experimental setup. Ab⁺ (n = 5) and Ab⁻ (n = 3–7) K18-hACE2 mice were primed with 2 × 10⁵ TCID₅₀ of SARS-CoV-2 D614G and rechallenged with a higher dose (1 × 10⁶ TCID₅₀) of SARS-CoV-2 B.1.1.529 (Omicron). Ab⁺ (n = 4–9) and Ab⁻ (n = 4–7) naïve mice, unexposed to the primary challenge, were infected with 1 × 10⁶ TCID₅₀ of SARS-CoV-2 B.1.1.529. PBS-exposed mice were used as controls. Blood was collected 7, 14 and 21 days after the first infection. Blood, lung, NT and mediastinal lymph node (mLN) were collected 4 days after rechallenge. b, Anti-S1 RBD IgG levels in the plasma after the first challenge. c,d, SARS-CoV-2 RNA in the NT (c) and lung (d). RNA values as copy number per ng of total RNA and the LOD as a dashed line. e, Viral titers in the lung were determined by TCID₅₀. f, Immunohistochemical micrographs of lung sections from PBS-, naïve- and primed-Ab⁺ and Ab⁻ mice. N-SARS-CoV-2-positive cells in brown. Scale bars, 100 μm. g–l, Flow cytometry plots (g and j), frequency (h and k) and absolute number (i and l) of CD8⁺ T cells (g–i) or CD4⁺ T cells (j–l) expressing IFN-γ and TNF in the lungs upon in vitro stimulation with a pool of SARS-CoV-2 peptides. Plots pregated as live⁺/B220⁻/CD19⁻/CD4⁻/CD8⁺ (g–i) or CD8⁻/CD4⁺ (j–l). m, Anti-S1 RBD IgG levels in the plasma 4 days after rechallenge. n,o, Flow cytometry plots (n) and frequency (o) of RBD-specific B cells detected by RBD-tetramers in the lungs (pregated on live⁺/CD4⁻/CD8⁻/B220⁺/CD19⁺). p,q, Flow cytometry histogram (p) and geometric mean fluorescence intensity (gMFI) (q) of surface markers expressed by RBD-specific B cells in the lung of Ab⁺ primed mice. As control, B cells negative for RBD-tetramer staining (gray). gMFI as log₂(fold change) over control B cells. Data are expressed as mean ± s.e.m. and are representative of at least two independent experiments. Data in b–d are pooled from two independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001; Kruskal–Wallis test followed by uncorrected Dunn’s test; each comparison stands alone (c–e, m and o). Two-way ANOVA, Fisher’s LSD test (each comparison stands alone; b, h, i, k and l). LSD, least significant difference. Source data

Fig. 2. A hyACE2 knock-in mouse supports SARS-CoV-2 infection.

a, Amino acid sequence of human (h)ACE2 and mouse (m)ACE2. In red are the eight residues involved in the interaction with the SARS-CoV-2 spike protein. b, Molecular modeling of the interaction between SARS-CoV-2 spike RBD (orange) and hACE2 or mACE2 (cyan). Crosses indicate the absence of interaction. Electrostatic and hydrophobic interactions in green and yellow dashed lines. c, Experimental setup. The 3T3 cells transduced with lentiviral vectors to express hACE2 (blue symbols), mACE2 (gray symbols) and a hybrid human/mouse (hy)ACE2 (green symbols) were infected with different concentrations of SARS-CoV-2. Nontransduced (WT) 3T3 cells as control. n = 3 biological replicates. d, Dose-dependent viral activity in 3T3 cells infected with SARS-CoV-2 D614G (left), B.1.617.2 (middle) or B.1.1.529 (right). Infection rates as a percentage of the virus-induced cytopathic effect 72 h after infection. Comparison with WT 3T3 cells. n = 3 biological replicates. e, Design of human/mouse hybrid Ace2 allele. f, Experimental setup. K18-hACE2 transgenic mice (n = 4) and hyACE2 knock-in mice (n = 5) were infected with 5 × 10⁵ TCID₅₀ of SARS-CoV-2 B.1.617.2 (Delta). PBS-exposed mice were used as controls (n = 2). Peripheral blood, lung and NT were analyzed 6 days after challenge. g, SARS-CoV-2 RNA in the NT (left) and lung (right). RNA values as copy number per ng of total RNA and the LOD as a dashed line. h, Respiratory frequency (left) and Rpef (right) were assessed by WBP 5 days postinfection (average over a 15-min data collection period). i, Anti-S1 RBD IgG levels in the plasma. j,k, Absolute number of total CD8⁺ T cells (j, left) and CD4⁺ T cells (k, left) and of cytokine-producing CD8⁺ cells (j, right) and CD4⁺ cells (k, right) in the lung on in vitro stimulation with a pool of SARS-CoV-2 peptides. Data are expressed as mean ± s.e.m. and are representative of at least two independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001; Kruskal–Wallis test followed by uncorrected Dunn’s test; each comparison stands alone (g–i, j and k (left)). Two-way ANOVA, Tukey’s multiple comparison (j and k (right)); two-way ANOVA, Fisher’s LSD test (each comparison stands alone; d). Source data

Fig. 3. Antibody-independent protection in hyACE2 mice.

a, Experimental setup. Ab⁺ (n = 5) and Ab⁻ (n = 5) hyACE2 knock-in mice were primed with 5 × 10⁵ TCID₅₀ of SARS-CoV-2 B.1.617.2 (Delta) and rechallenged with a higher dose (1 × 10⁶ TCID₅₀) of SARS-CoV-2 B.1.1.529 (Omicron). Ab⁺ (n = 5) and Ab⁻ (n = 7) naïve mice, unexposed to the primary challenge, were infected only with 1 × 10⁶ TCID₅₀ of SARS-CoV-2 B.1.1.529 (Omicron). As control, PBS-exposed mice. Blood was collected 7, 14 and 21 days after the first infection. Blood, lung, NT and mLN were analyzed 4 days after rechallenge. b, Anti-S1 RBD IgG levels in the plasma after the first challenge. Number (n) of mice as in a. c, Dot plots (left) and frequency (right) of IFN-γ-producing CD8⁺ T cells in the peripheral blood after the first challenge. Number (n) of mice as in a. d,e, SARS-CoV-2 RNA in the NT (d) and lung (e). RNA values as copy number per ng of total RNA and the LOD as a dashed line. f, Viral titers in the lung were determined by TCID₅₀. g,i, Flow cytometry plots (left) and frequency (right) of CD8⁺ T cells (g) or CD4⁺ T cells (i) expressing IFN-γ and TNF in the lungs upon in vitro stimulation with a pool of SARS-CoV-2 peptides. Plots pregated as live⁺/B220⁻/CD19⁻/CD4⁻/CD8⁺ cells (g) or CD8⁻/CD4⁺ (i). h,j, Absolute number of cytokine-producing CD8⁺ T cells (h) or CD4⁺ T cells (j). k, Anti-S1 RBD IgG levels in the plasma 4 days after rechallenge. Data are expressed as mean ± s.e.m. and are representative of at least two independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001; Kruskal–Wallis test followed by uncorrected Dunn’s test; each comparison stands alone (d–f and k). Two-way ANOVA, Fisher’s LSD test (each comparison stands alone; b, c, g–j). Source data

Fig. 4. Antibody-independent protection against severe disease.

a, Experimental setup. Ab⁺ (n = 5) and Ab⁻ (n = 4) C57BL/6 mice were primed with 5 × 10⁴ TCID₅₀ of rSARS-N501Y_MA30 and rechallenged with a higher dose (3 × 10⁵ TCID50) of rSARS-N501Y_MA30. Ab⁺ (n = 4) and Ab⁻ (n = 4) naïve mice, unexposed to the primary challenge, were infected with 3 × 10⁵ TCID₅₀ of rSARS-N501Y_MA30. PBS-exposed mice were used as controls. Blood was collected 7, 14 and 21 days after the first infection. Blood, lung and NT were analyzed 4 days after rechallenge. b, Anti-S1 RBD IgG levels in the plasma after the first challenge. Number (n) of mice as in a. c,d, Dot plots (c) and frequency (d) of IFN-γ-producing CD8⁺ T cells in the blood after the first challenge. Asterisk indicates Ab⁺ (primed) compared to PBS; hash indicates Ab⁻ (primed) compared to PBS. Number (n) of mice as in a. e, Survival curve after the rechallenge. f, Mouse body weight after the rechallenge as a percentage of weight relative to day 29. Asterisk and hash indicate Ab⁺ and Ab⁻ compared to PBS, respectively. g, Clinical score. Number (n) of mice as in a. h,i, Respiratory frequency (h) and PenH (i) were assessed by WBP 3 days after rechallenge (average over a 15-min data collection period). j,k, SARS-CoV-2 RNA in the NT (j) and lung (k). RNA values as copy number per ng of total RNA and the LOD as a dashed line. l, Viral titers in the lung were determined by TCID₅₀. m, Immunohistochemical micrographs of lung sections from PBS-, naïve- and primed-Ab⁺ and Ab⁻ mice. N-SARS-CoV-2-positive cells in brown. Scale bars, 100 μm. n, Anti-S1 RBD IgG levels in the plasma 4 days after rechallenge. o,r, Flow cytometry histogram and gMFI of surface markers expressed by CD8⁺ T cells (o) and CD4⁺ T cells (r) in the lung. p,s, Frequency of CD11a, CD49d CD8⁺ T cells (p) and CD4⁺ T cells (s) in the lung. q,t, Plots and frequency of CD11a⁺, CD49d⁺ CD8⁺ T cells (q) or CD4⁺ T cells (t) expressing IFN-γ⁺/TNF⁺ upon in vitro stimulation with a pool of SARS-CoV-2 peptides. Plots pregated as live⁺/B220⁻/CD19⁻. Data are expressed as mean ± s.e.m. and are representative of at least two independent experiments. *,^#P < 0.05, **,^##P < 0.01, ***,^###P < 0.001; Kruskal–Wallis test followed by uncorrected Dunn’s test; each comparison stands alone (h–l, n, p and s). Two-way ANOVA, Fisher’s LSD test (each comparison stands alone; b, d, f, g, o, q, r and t). Source data

Fig. 5. Antibody-independent protection conferred by prior mRNA vaccination.

a, Experimental setup. Ab⁺ (n = 9) and Ab⁻ (n = 9) hyACE2 knock-in mice were immunized with SARS-CoV-2 full-length spike (S-2P) or firefly luciferase (Luc) mRNA-LNP. Fifteen days after the boost (day 43), mice were exposed to a heterologous challenge of 1 × 10⁶ TCID₅₀ of SARS-CoV-2 B.1.1.529 (Omicron). As control, nonimmunized mice were exposed to PBS. Blood was collected 7, 14, 21 and 35 days after the immunization. Blood, lung, NT, mLN and spleens were analyzed 4 days postinfection. b, Anti-S1 RBD IgG levels in the plasma after the immunization. Number (n) of mice as in a. c, Dot plots (left) and frequency (right) of IFN-γ-producing CD8⁺ T cells in the peripheral blood after the first dose of immunization. Asterisk indicates Ab⁺ (S-2P) compared to Ab⁺ (Luc); hash indicates Ab⁻(S-2P) compared to Ab⁺ (Luc); + symbol indicates Ab⁺ (S-2P) compared to Ab⁻ (Luc); x symbol indicates Ab⁻ (S-2P) compared to Ab⁻(Luc). Number (n) of mice as in a. d,e, SARS-CoV-2 RNA in the (d) NT and (e) lung. RNA values as copy number per ng of total RNA and the LOD as a dashed line. f, Viral titers in the lung were determined by TCID₅₀. g,i Representative flow cytometry plots (left) and frequency (right) of CD8⁺ T cells (g) or CD4⁺ T cells (i) expressing IFN-γ and TNF in the lungs on in vitro stimulation with a pool of SARS-CoV-2 peptides. Plots were pregated as live⁺/B220⁻/CD19⁻/CD4⁻/CD8⁺ cells (g) or CD8⁻/CD4⁺ cells (i). h,j, Absolute number of cytokine-producing CD8⁺ T cells (h) or CD4⁺ T cells (j). k, Anti-S1 RBD IgG levels in the plasma 4 days postinfection. Data are expressed as mean ± s.e.m. and are representative of at least two independent experiments. *,^#,^x,⁺P < 0.05; **,^##,^xx,⁺⁺P < 0.01; ***,^###,^xxx,⁺⁺⁺P < 0.001; Kruskal–Wallis test followed by uncorrected Dunn’s test; each comparison stands alone (d–f and k). Two-way ANOVA, Fisher’s LSD test (each comparison stands alone; b, c, g–j). Source data

Fig. 6. Antibody-independent protection via T cells and IFN-γ.

a, Experimental setup. Ab⁻ hyACE2 knock-in mice were primed with 5 × 10⁵ TCID₅₀ of SARS-CoV-2 B.1.617.2 (Delta) and rechallenged with a higher dose (1 × 10⁶ TCID₅₀) of SARS-CoV-2 B.1.1.529 (Omicron). Ab⁻ (n = 4) naïve mice, unexposed to the primary challenge, were infected with 1 × 10⁶ TCID₅₀ of SARS-CoV-2 B.1.1.529 (Omicron). A group of primed mice was injected intravenously with anti-CD4 (n = 4), or anti-CD8 (n = 4), or the combination of anti-CD4 and anti-CD8 (n = 4) depleting antibodies 2 (day 23) and 1 day (day 25) before re-infection and 3 days later (day 29). A group of primed mice was injected intravenously with anti-IFN-γ (n = 5) blocking antibodies 4 h before and 3 days after (day 29) the re-infection. PBS-exposed mice were used as controls. Blood was collected 21 and 26 days after the first infection. Lung and mLN were analyzed 4 days after rechallenge. b,c, Dot plots (left) and frequency (right) of CD8⁺ T cells (b) and CD4⁺ T cells (c) in the peripheral blood after the first challenge. d, SARS-CoV-2 RNA in the lung. RNA values as copy number per ng of total RNA and the LOD as a dashed line. e, SARS-CoV-2 RNA in the lung. RNA values as fold change of the RNA copy number per ng of total RNA over the LOD. Data are pooled from two independent experiments. f,g, Flow cytometry plots (left) and absolute number (right) of CD8⁺ T cells (f) or CD4⁺ T cells (g) expressing IFN-γ and TNF in the lungs upon in vitro stimulation with a pool of SARS-CoV-2 peptides. Plots pregated as live⁺/B220⁻/CD19⁻/CD4⁻/CD8⁺ cells (f) or CD8⁻/CD4⁺ (g). h, SARS-CoV-2 RNA in the lung. RNA values as copy number per ng of total RNA and the LOD as a dashed line. Data are expressed as mean ± s.e.m. and are representative of two independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001; Kruskal–Wallis test followed by uncorrected Dunn’s test; each comparison stands alone (d, e and h). Two-way ANOVA, Fisher’s LSD test (each comparison stands alone; b, c, f and g). Source data

Extended Data Fig. 1. Antibody-independent protection against heterologous SARS-CoV-2 challenge conferred by prior infection in K18-hACE2 transgenic mice.

(a) Experimental setup. Antibody-sufficient (Ab⁺, n = 3-10) and antibody-deficient (Ab⁻, n = 3–28) K18-hACE2 transgenic mice were infected with a target dose of 2 × 10⁵ TCID₅₀ of SARS-CoV-2 D614G through aerosol exposure. Brain was collected and analyzed 9 days postchallenge. (b) Survival curve of Ab⁺ (n = 10, blue line) and Ab⁻ mice upon infection (n = 28, red line). (c) Quantification of SARS-CoV-2 RNA in the brain of the indicated mice. RNA values are expressed as copy number per ng of total RNA and the limit of detection is indicated as a dotted line. n = 3. (d, e) Experimental setup as described in Fig. 1. (d) Representative confocal immunofluorescence staining of lung sections from Ab⁺ (upper panels) and Ab⁻ (lower panels) mice 4 days after re-challenge. Cell nuclei are depicted in blue, B220⁺ cells in green and TCR-β⁺ cells in white. Scale bar, 30 μm. (e) Representative flow cytometry plots (left) and frequency (right) of RBD-specific B cells in the mLN of indicated mice 4 days post re-challenge (pre-gated on live⁺/CD4⁻/ CD8⁻/ B220⁺/CD19⁺ cells). n as indicated in Fig. 1a. Data are expressed as mean ± SEM. Data are representative of at least 2 independent experiments. *p value < 0.05, **p value < 0.01, ***p value < 0.001; log-rank (Mantel-Cox) test (b); Two-tailed unpaired t-test (c); Kruskal-Wallis test followed by uncorrected Dunn’s test, each comparison stands alone (e). Source data

Extended Data Fig. 2. A novel human/mouse hybrid ACE2 knock-in mouse supports SARS-CoV-2 infection.

(a) Amino acid sequence alignment of SARS-CoV-2 D614G, B.1.617.2 and B.1.1.529. The residues involved in the interaction with the human ACE2 are indicated in red. (b) Representative flow cytometry histograms representing the ACE2 expression by 3T3 cells and 3T3 cells transduced with murine ACE2, human ACE2 and hybrid ACE2. (c, d) SARS-CoV-2 titers in transduced 3T3 cells upon infection with D614G (left), B.1.617.2 (middle) or B.1.1.529 (right). In (c) titers were determined as percent of the virus-induced cytopathic effect evaluated 48 hours after infection. In (d) titers were determined by qPCR quantification of SARS-CoV-2 RNA in the supernatant 48 and 72 hours postinfection. n = 3 biological replicates. (e) Human ACE2 expression in the indicated organs from WT (gray), K18-hACE2 (blue) and hyACE2 (green) mice. Values were normalized to the reference gene Gapdh and expressed as fold increase over the limit of detection (LOD). n = 7-8 mice. (f) K18-hACE2 transgenic mice (n = 5) and hyACE2 knock-in mice (n = 5) were infected with a target dose of 2 ×10⁵ or 5 ×10⁵ TCID₅₀ of SARS-CoV-2 B.1.617.2 through aerosol exposure. Lung and nasal turbinates (NT) were collected and analyzed 3 days postchallenge. (g, h) Quantification of SARS-CoV-2 RNA in the NT (g) and in the lung (h) of the indicated mice. n as indicated in (f). RNA values are expressed as copy number per ng of total RNA and the limit of detection is indicated as a dotted line. Data are expressed as mean ± SEM. Data are representative of at least 2 independent experiments. *p value < 0.05, **p value < 0.01, ***p value < 0.001; Two-way ANOVA, Fisher’s LSD test (Each comparison stands alone) (c–e, g, h). Source data

Extended Data Fig. 3. Previously infected, antibody-deficient hyACE2 knock-in mice are protected against heterologous SARS-CoV-2 re-challenge.

(a) Survival curve of Ab⁺ (n = 10, blue line) and Ab⁻ hyACE2 upon SARS-CoV-2 infection (n = 10, red line). Mice were exposed to 5 × 10⁵ TCID₅₀ of SARS-CoV-2 B.1.617.2, as described in Fig. 3. (b) Representative flow cytometry plots (left) and frequency (right) of GL-7⁺ FAS⁺ B cells in the mLN of indicated mice 4 days post re-challenge (pre-gated on live⁺/CD4⁻/CD8⁻/B220⁺/CD19⁺ cells). n as indicated in Fig. 3a. Data are expressed as mean ± SEM. Data are representative of at least 2 independent experiments. *p value < 0.05, **p value < 0.01, ***p value < 0.001; Kruskal-Wallis test followed by uncorrected Dunn’s test, each comparison stands alone (b). Source data

Extended Data Fig. 4. Antibody-independent protection against severe disease.

(a, b) Survival curve (a) and body weight loss (b) of Ab⁺ (n = 10, blue symbols) and Ab⁻(n = 10, red symbols) C57BL/6 mice upon SARS-CoV-2 infection. Mice were exposed to 5 × 10⁴ TCID₅₀ of aerosolized rSARS-N501Y_MA30, as described in Fig. 4. Data are expressed as mean ± SEM. Data are representative of at least 2 independent experiments. Source data

Extended Data Fig. 5. Antibody-independent protection against heterologous SARS-CoV-2 infection conferred by prior mRNA vaccination.

(a–d) Frequency of CD8⁺ T cells (left) and CD4⁺ T cells (right) in the mediastinal lymph nodes (a, b) and spleens (c, d) of the indicated mice 4 days postinfection. Experimental setup is described in Fig. 5, and n of mice is indicated in Fig. 5a. Data are expressed as mean ± SEM. Data are representative of at least 2 independent experiments. *p value < 0.05, **p value < 0.01, ***p value < 0.001; Two-way ANOVA, Fisher’s LSD test (Each comparison stands alone) (a–d). Source data

Extended Data Fig. 6. T cells and IFN-γ are key players in the antibody-independent protection against heterologous SARS-CoV-2 challenge.

(a, b) Representative dot plots (left) and frequency (right) of CD4⁺ T cells (upper panel) and CD8⁺ T cells (lower panel) in the lungs (a) and mediastinal lymph nodes (b) of the indicated mice 4 days post re-challenge. Experimental setup is described in Fig. 6, and n of mice is indicated in Fig. 6a. Data are expressed as mean ± SEM. Data are representative of at least 2 independent experiments. *p value < 0.05, **p value < 0.01, ***p value < 0.001; Kruskal-Wallis test followed by uncorrected Dunn’s test, each comparison stands alone (a, b). Source data

Extended Data Fig. 7. Critical role of CD8⁺ T cells in severe disease.

(a) Experimental setup. Ab^- D_HLMP2a mice were infected with a target dose of 5 × 10⁴ TCID₅₀ of rSARS-CoV-2-N501Y_MA30 through aerosol exposure. Thirty days after infection, primed mice were exposed to a heterologous challenge with a target dose of 1x10⁵ TCID₅₀ of rSARS-CoV-2-N501Y_MA30. Ab⁻ (n = 4) naïve mice unexposed to the primary challenge were infected with 1x10⁵ TCID₅₀ of rSARS-CoV-2-N501Y_MA30. A group of primed mice was injected intravenously with anti-CD4 (n = 4), or anti-CD8 (n = 4), or the combination of anti-CD4 and anti-CD8 (n = 5) depleting antibodies three (day 27) and one day (day 29) prior to re-infection and three days later (day 33). Non-infected mice exposed to aerosolized PBS were used as control. Analyses were performed 4 days post re-challenge. (b) Quantification of SARS-CoV-2 RNA in the nasal turbinates of the indicated mice. n as indicated in a. RNA values are expressed as copy number per ng of total RNA and the limit of detection is indicated as a dotted line. (c) Representative immunohistochemical micrographs of lung sections from indicated mice. N-SARS-CoV-2 positive cells are depicted in brown. Scale bars, 100 μm. Data are expressed as mean ± SEM and are representative of two independent experiments. *p value < 0.05, **p value < 0.01; Kruskal-Wallis test followed by uncorrected Dunn’s test, each comparison stands alone (b). Source data