Prevention of respiratory virus transmission by resident memory CD8+ T cells

Abstract

An ideal vaccine both attenuates virus growth and disease in infected individuals and reduces the spread of infections in the population, thereby generating herd immunity. Although this strategy has proved successful by generating humoral immunity to measles, yellow fever and polio, many respiratory viruses evolve to evade pre-existing antibodies¹. One approach for improving the breadth of antiviral immunity against escape variants is through the generation of memory T cells in the respiratory tract, which are positioned to respond rapidly to respiratory virus infections^2-6. However, it is unknown whether memory T cells alone can effectively surveil the respiratory tract to the extent that they eliminate or greatly reduce viral transmission following exposure of an individual to infection. Here we use a mouse model of natural parainfluenza virus transmission to quantify the extent to which memory CD8⁺ T cells resident in the respiratory tract can provide herd immunity by reducing both the susceptibility of acquiring infection and the extent of transmission, even in the absence of virus-specific antibodies. We demonstrate that protection by resident memory CD8⁺ T cells requires the antiviral cytokine interferon-γ (IFNγ) and leads to altered transcriptional programming of epithelial cells within the respiratory tract. These results suggest that tissue-resident CD8⁺ T cells in the respiratory tract can have important roles in protecting the host against viral disease and limiting viral spread throughout the population.

Extended data Fig. 1.. Distribution and characterization of tissue-resident Sendai-specific CD8⁺ T cells limit following intranasal and intraperitoneal infection with recombinant influenza virus.

(a) Schematic of experiment for investigating immunization route on protection from direct infection. (b) Representative plots of tetramer staining at day 35 in immunized mice. (c) Absolute numbers of Sendai NP_324–332/K^b+ CD8 T cells in spleen, BAL, lung, and nasal cavity of immunized mice (n=9 for PR8 SenNP i.p. and PR8 SenNP i.n.). (d) Representative plots of staining for T_RM markers CD69 and CD103 on antigen-specific cells. (e) Absolute numbers of CD69⁺CD103⁺ Sendai NP_324–332/K^b+ CD8⁺ T cells in BAL (n=9 for PR8-SenNP i.n. and n=8 for PR8-SenNP i.p.), lung and nasal cavity (n=10 for PR8-SenNP i.n. and n=9 for PR8-SenNP i.p.). (f) Representative histograms of CD49a, CXCR6, and CXCR3 expression gated on nasal cavity Sendai NP_324–332/K^b+ CD8⁺ T cells following i.n. or i.p. immunization based on T_RM marker expression (CD69⁺CD103⁺, red; CD69⁺CD103⁻, black; CD69⁻CD103⁻, green). (g) Frequency of CD49a, CXCR6, and CXCR3 expression on Sendai NP_324–332/K^b+ CD69⁺CD103⁺ CD8⁺ T cells in the nasal cavity (n=5 per group) following i.n. or i.p. immunization. Data are representative of two individual experiments. Lines represent mean values (c, e, g). Statistical significance was determined using two-sided Mann Whitney test. * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001, ns: non-significant.

Extended data Fig. 2.. Sendai-luciferase bioluminescence strongly correlates with viral titer.

(a) Bioluminescence in Sendai-Luc infected mice combined with viral titer from the total respiratory tract (nasal cavity, trachea, and lungs) on days 2, 4 and 6 post-infection. Black dots represent viral titer, grey curves represent bioluminescence of individual mice, and dotted black line represents background bioluminescence. (b) Correlation of bioluminescence with viral titers. Data shown are from two independent pooled experiments with 10 mice per timepoint.

Extended data Fig. 3.. Contact mice that show no indication of transmission by bioluminescence also fail to develop a Sendai-specific T cell response.

(a) Schematic of experiment investigating the cellular immune response in bioluminescence-positive versus bioluminescence-negative contact mice. (b) Representative flow cytometry plots of Sendai NP_324–332K^b+ tetramer staining gated on CD8⁺ T cells in the BAL, lung, nasal cavity, and spleen. (c) Number of total Sendai NP_324–332K^b+ CD8 T cells in the BAL, lung, nasal cavity, and spleen in contact mice 35 after co-housing with PR8-SenNP or WT PR8 immunized index mice. (d) Number of CD69⁺CD103⁺ Sendai NP_324–332K^b+ CD8 T_RM in the BAL, lung, and nasal cavity in contact mice 35 days after co-housing with PR8-SenNP or WT PR8 immunized index mice. Data are representative of two individual experiments with n = 7 mice per group. Lines represent mean values and error bars represent standard deviation (c, d).

Extended data Fig. 4.. Number of SenNP-specific CD8⁺ T_RM in knockout mouse strains following immunization.

(a) Experimental schematic for immunizing WT and knockout mouse strains. (b) Number of CD69⁺ CD103⁺ Sendai NP_324–332/K^b+ CD8 T_RM in the BAL, lung, and nasal cavity 35 days after immunization with LAIV-SenNP for WT (n=15), Prf^−/− (n=12 for BAL, n=13 for lungs and nasal cavity), Ifng^−/− (n=18), and Ifngr^−/− (n=16 for lungs and nasal cavity, n=15 for BAL). Data shown are from 3 independent pooled experiments. (c) Experimental schematic for immunizing WT and knockout mouse strains. (d) Number of NK cells (left graph) and inflammatory monocytes (right graph) in the nasal cavity 35 days after immunization with LAIV WT (n=10) or LAIV-SenNP (n=10 for WT, n=9 for Prf^−/−, n=7 for Ifngr^−/−). Data shown are from 2 pooled independent experiments. Lines represent means (b, d). Statistical significance was determined using two-sided Mann Whitney test. * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001, ns: non-significant.

Extended data Fig. 5.. Immunization does not alter influx of NK cells and monocytes following Sendai virus transmission.

(a) Experimental schematic where immunized contact mice were cohoused with a Sendai-Luc infected index mouse and tissues analyzed for innate immune populations at the time of cohousing (D30) and two days after cohousing (D30+2). (b) Number of NK cells in nasal cavity (D30: n=11 for PR8 WT i.n., n=15 for PR8-SenNP i.n., n=14 for PR8-SenNP i.p.) (D30+2: n= 11 for PR8 WT i.n., n=12 for PR8-SenNP i.n., n=12 for PR8-SenNP i.p.) and BAL (D30: n=10 for PR8 WT i.n., n=14 for PR8-SenNP i.n., n=14 for PR8-SenNP i.p.) (D30+2: n=11 for PR8 WT i.n., n=11 for PR8-SenNP i.n., n=12 for PR8-SenNP i.p.). (c) Number of inflammatory monocytes in nasal cavity and BAL (same n as (b), except n=12 for D30+2 PR8-SenNP i.n. in BAL). Data shown are from 3 independent experiments. Lines represent means and error bars represent 95% confidence interval (b, c). Statistical significance was determined using two-sided Mann Whitney test. * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001, ns: non-significant.

Extended data Fig. 6.. Sendai-specific T_RM numbers and assessment of transmission under different immunization strategies at 1- and 6-months post-immunization.

(a) Experimental schematic where contact mice, immunized as indicated, were cohoused with a Sendai-Luc infected index mouse at 35- or 180-days post-immunization. (b) Number of CD69⁺CD103⁺ Sendai NP_324–332/K^b+ CD8 T_RM in the BAL, lung, and nasal cavity at 1 month or six months post-immunization with x31-SenNP i.n. (n=10 per timepoint except n=9 for 1 month BAL), PR8-SenNP i.p. (n=10 per timepoint except n=9 for 1 month BAL), PR8-SenNP i.n. (1 month n=9 for lung and nasal cavity and n=6 for BAL, 6 month n=10 for lung and nasal cavity and n=9 for BAL), LAIV-SenNP i.n. (1 month n=8, 6 month n=10), and Ad-SenNP i.n. (1 month n=9 for lung and nasal cavity and n=8 for BAL, 6 month n=10 for nasal cavity and BAL, n=9 for lung). (c and d) Bioluminescence curves of immunized contact mice following exposure to an infected index mouse at 1 month (c) or 6 months (d) post-immunization with x31 WT (1 month n=14, 6 month n=16), x31-SenNP (1 month n=16, 6 month n=15), PR8-SenNP i.p. (1 month n=16, 6 month n=16), PR8-SenNP i.n. (1 month n=16, 6 month n=16), LAIV-SenNP i.n. (1 month n=15, 6 month n=16), and Ad-SenNP i.n. (1 month n=15, 6 month n=16). Solid dark lines represent means, solid pale lines represent individual mice, dashed grey line represents background bioluminescence, and dashed red line represents the limit of infection (c, d). Solid lines (b) and error bars (b-d) represent mean with 95% confidence interval. Statistical significance was determined using two-sided Mann Whitney test. Data are combined from two independent experiments. * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001, ns: non-significant.

Extended data Fig. 7.. High viral burden in index mice does not correlate with increased viral burden in contact mice.

Contact mice were infected intranasally with WT x31 (n=14), x31-SenNP (n=16), PR8-SenNP (n=16), or Ad-SenNP (n=15) and cohoused with a Sendai-Luc infected index mouse at day 35 post-immunization. n= 4 index mice per immunization group are plotted. Total viral burden (AUC) of the co-housed index and contact mice were plotted for each cage. Data were fitted to a generalized linear model with gaussian family for each immunization group to investigate the relationship between AUCs of the index mice and contact mice. Data are combined from two independent experiments.

Extended data Fig. 8.. Pre-existing immunity to related influenza strains limits the efficacy of protective T cell immunity induced by LAIV-SenNP immunization but can be overcome by Ad-SenNP immunization.

(a) Experimental schematic for testing the impact of pre-existing influenza immunity on the ability of LAIV-SenNP to protect against transmission. (b) Bioluminescence curves of A/Cal/09 i.p. & PR8 LAIV WT (n=16), PBS i.p. & PR8 LAIV-SenNP (n=16), A/Cal/09 i.p. & PR8 LAIV-SenNP (n=20), and A/Cal/09 i.p. & Ad-SenNP (n=16) immunized contact mice following exposure to an infected index mouse 30 days after the second immunization. Solid dark lines represent means, solid pale lines represent individual mice, dashed grey line represents background bioluminescence, and dashed red line represents the limit of infection. (c) AUC of bioluminescence in immunized contact mice that become infected following co-housing with an infected index mouse. (d) Probability of infection for immunized contact mice calculated as the proportion of contact mice that became infected. Bars represent 95% binomial confidence intervals (d). Lines represent means (b-d) and error bars represent 95% confidence intervals (b, c). Data are combined from two independent experiments. Statistical significance was determined using two-sided Mann Whitney test. * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001, ns: non-significant.

Extended data Figure 9.. Heterologous influenza prime-boost does not improve the durability of respiratory tract T_RM.

(a) Experimental schematic to assess the durability of respiratory tract T_RM following heterologous PR8-SenNP or Ad-SenNP boosting. (b) Number of CD69⁺CD103⁺ Sendai NP_324–332K^b+ CD8 T_RM in the BAL, lung, and nasal cavity at day 120. Data are combined from two independent experiments with n= 10 mice for PR8 WT i.n., PR8-SenNP i.n., and Ad-SenNP i.n. secondary immunization groups. Lines represent means. Statistical significance was determined using two-sided Mann Whitney test. * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001, ns: non-significant.

Fig. 1.. Respiratory tract CD8 T_RM can limit transmission of respiratory viruses.

(a) Schematic of experiment to assess Sendai viral titers in the nasal cavity of immunized mice and investigate immunization route on protection from direct infection. (b) Sendai virus titers in the nasal cavity of immunized mice (n=10 for PR8 WT n=10 for PR8-SenNP i.p., n=9 for PR8-SenNP i.n.). c) Representative in vivo imaging (IVIS) images of immunized mice following direct infection with Sendai-Luc. (d) Bioluminescence curves over the course of infection (n=10 per group). (e) Area under curve (AUC) of bioluminescence signal over the course of infection (n=10 for each experimental group). (f) Schematic of experiment to investigate transmission from immunized index mice to naïve contact mice at indicated time points. (g) Example images of contact mice where index mice were added 3 days post infection. (h) Bioluminescence curves of contact mice following co-housing on day 1 (n=12 per group), day 3 (n=12 per group), day 5 (n=8 per group), and day 7 (n=8 per group) post infection of index mice. (i) AUC of bioluminescence of contact mice that become infected following co-housing with an infected index mouse. Grey circles represent uninfected mice. (j) Percent infected contact mice following co-housing with an infected index mouse and the probability of transmission (inset). Dots represent proportion of contact mice that become infected and error bars represent 95% binomial confidence intervals. All data are representative of 2 combined experiments with similar results. For bioluminescence curves (d, h), solid dark lines represent means, solid pale lines represent individual mice, dashed grey line represents background bioluminescence, and dashed red line represents the limit of infection. Solid lines (b, e, i, j) and error bars (d, e, h, i) represent mean with 95% confidence interval. Statistical significance was determined using a two-sided Mann Whitney test. * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001, ns: non-significant.

Fig. 2.. IFNγ signaling plays an essential role in preventing transmission of respiratory viruses.

(a) PCA plot of sorted virus-specific CD8 T_RM and T_EM from BAL, lung and spleen of influenza primed mice at resting memory (d30, n=3 per group) and 3 days post challenge (d30+3, n=3 per group), pooled over 3 independent experiments. (b) Heat map of RNA transcript expression for the listed differentially expressed genes. (c) Schematic of experiment to investigate the importance of IFNγR or perforin for limiting transmission to naïve contact mice. (d) Bioluminescence of WT (n=11), Prf^−/− (n=10), and Ifngr^−/− (n=6) index mice following direct infection with Sendai-Luc. (e) AUC of bioluminescence in index mice. (f) Bioluminescence in B6 contact mice (WT, n=40 contacts; Prf^−/−, n=32 contacts; Ifngr^−/−, n=22 contacts) following co-housing with infected index mice on day 3 post infection. (g) AUC of bioluminescence in contact mice that become infected following co-housing with an infected index mouse. (h) Probability of transmission to naïve contact mice for each genotype of immunized index mouse, calculated as the proportion of contact mice that became infected. Error bars represent 95% binomial confidence interval. Data are combined from 3 independent experiments (d-h). For bioluminescence curves (d, f), solid dark lines represent means, solid pale lines represent individual mice, dashed grey line represents background bioluminescence, and dashed red line represents the limit of infection. Solid lines (e, g, h) and error bars (d-g) represent mean with 95% confidence interval. Statistical significance was determined using two-sided Mann Whitney test. * p<0.05, *** p<0.001, **** p<0.0001, ns: non-significant.

Fig. 3.. Respiratory tract CD8 T_RM protect against viral propagation following transmission through IFNγ.

(a) Schematic of experimental setup where immunized contact mice were cohoused with a Sendai-Luc infected index mouse. (b) Bioluminescence curves of contact mice immunized with PR8 WT i.n. (n=13), PR8 SenNP i.p. (n=16), or PR8-SenNP i.n. (n=15). (c) AUC of bioluminescence in immunized contact mice that become infected following co-housing with an infected index mouse. (d) Probability of infection for immunized contact mice, calculated as the proportion of contact mice that became infected. (e) Schematic of experiment to investigate protection from transmission in immunized WT, Prf^−/− Ifng^−/−, and Ifngr^−/− contact mice and WT mice treated with anti-FasL antibody. (f) Bioluminescence curves of immunized WT (n=16 for LAIV WT and LAIV-SenNP immunization), Prf^−/− (n=16), Ifng^−/−, (n=13), Ifngr^−/− (n=8), and WT+anti-FasL (n=14) contact mice after co-housing with an infected index mouse. (g) AUC of bioluminescence in contact mice that become infected following co-housing with an infected index mouse. (h) Probability of infection for immunized contact mice. Error bars represent 95% binomial confidence intervals (d, h). All data are combined from 2 independent replicates. For bioluminescence curves (b, f), solid dark lines represent means, solid pale lines represent individual mice, dashed grey line represents background bioluminescence, and dashed red line represents the limit of infection. Solid lines (c, d, g, h) and error bars (b, c, f, g) represent mean with 95% confidence interval. Statistical significance was determined using two-sided Mann Whitney test. * p<0.05, *** p<0.001, **** p<0.0001, ns: non-significant.

Fig. 4.. IFNγ signaling induces antiviral gene expression and increases antigen presentation in nasal cavity epithelial cells.

(a) Schematic of experiment to evaluate the impact of T_RM-derived IFNγ on nasal cavity epithelial cells. (b) PCA plot of sorted nasal cavity epithelial cells from LAIV-SenNP immunized WT and Ifng^−/− mice at resting memory (d30) (WT n=10, Ifng^−/− n=9) and 3 days post peptide administration (d30+3) (WT n=10, Ifng^−/− n=8). (c) Hierarchical clustering of RNA transcript expression for the listed differentially expressed genes. (d) GSEA plots of listed immune pathways. (e) Representative histograms of MHC-I and MHC-II expression on nasal cavity epithelial cells. (f) Frequency of MHC-I^hi and MHC-II⁺ nasal cavity epithelial cells (n=10 per group). Bars represent mean and standard deviation. (g) Cytokine and chemokine concentrations in the nasal cavity of immunized WT (n=10 per timepoint) and Ifng^−/− (n=10 per timepoint) mice at resting memory and 3 days following peptide administration. Bars represent median. The data shown are combined from 2 independent experiments. Statistical significance was determined using two-sided Mann Whitney test. * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001, ns: non-significant.

Fig. 5.. The number of respiratory tract CD8 T_RM is strongly linked to protection from transmission.

(a) Schematic of experiment to investigate the durability of T_RM-mediated protection and ability to limit viral transmission. (b) Bioluminescence curves of contact mice immunized with x31 WT (n=16), PR8-SenNP (n=16), x31-SenNP (n=15), and Ad-SenNP (n=16) after co-housing with an infected index mouse. (c) AUC of bioluminescence in contact mice that become infected following co-housing with an infected index mouse. (d) Probability of infection for immunized contact mice, calculated as proportion of contact mice that become infected. (e) Number of CD69⁺CD103⁺ Sendai NP_324–332/K^b+ CD8 T_RM in the BAL (n=9 for PR8-SenNP, n=10 for x31-SenNP and Ad-SenNP), lung (n=9 for Ad-SenNP, n=10 for PR8-SenNP and x31-SenNP), and nasal cavity (n=10 per group), and Sendai NP_324–332/K^b+ CD8 T_EM in the spleen (n=10 per group), on D180 post-immunization. (f) Correlation between AUC of bioluminescence and the number of CD69⁺CD103⁺ Sendai NP_324–332/Kb+ T_RM in the BAL, lung, and nasal cavity. (g) Correlation between the probability of infection and number of CD69⁺CD103⁺ Sendai NP_324–332/Kb+ T_RM in the BAL, lung, and nasal cavity. (h) Schematic of experiment to investigate the impact of multiple immunizations on the durability of T_RM-mediated protection against Sendai virus transmission. (i) Bioluminescence curves of immunized contact mice after co-housing with an infected index mouse (n=15 for PR8 WT boosted contacts, n=19 for PR8-SenNP boosted contacts, and n=16 for Ad-SenNP boosted contacts). (j) AUC of bioluminescence in primed and boosted contact mice that become infected following co-housing with an infected index mouse. (k) Probability of infection for immunized contact mice. Error bars represent 95% binomial proportion confidence intervals (d, k). All data are combined from 2 independent experiments. For bioluminescence curves (b, i), solid dark lines represent means, solid pale lines represent individual mice, dashed grey line represents background bioluminescence, and dashed red line represents the limit of infection. Solid lines (c, d, e, j, k) and error bars (b, c, i, j) represent mean with 95% confidence interval. Statistical significance was determined using two-sided Mann Whitney test * p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001, ns: non-significant.