Viral-vectored respiratory mucosal immunization induces Ag-specific CD8 T cells in the lung with distinct gene expression profile

To begin investigating whether respiratory mucosal immunization promotes lung T RM generation, we first set out to characterize the properties of vaccine-induced Ag-specific CD8 T cells in the lung. An adenovirus-vectored tuberculosis vaccine (AdAg85A) was used as a model replication-defective viral vector vaccine and this vaccine, when delivered via respiratory mucosal and parenteral intramuscular routes, induced Ag-specific CD8 T cell responses in the lung21, 28. Interestingly, using intravascular immunostaining it has recently been shown that >95% of T cells in a naive lung are trapped in the pulmonary vasculature and bona fide lung tissue T cells were detected only after respiratory mucosal infection or immunization29, 30. Thus using such intravascular immunostaining we first verified Ag-specific T cell distribution in the lung at 4 weeks following respiratory mucosal and parenteral route of immunization. We found that the vast majority of Ag-specific CD8 T cells induced by respiratory mucosal immunization were bona fide lung tissue T cells. In sharp contrast, most of the Ag-specific CD8 T cells induced by parenteral intramuscular immunization were located in the lung vasculature. To determine the unique properties of respiratory mucosal immunization-induced (i.n.) lung tissue Ag-specific memory CD8 T cells, we compared gene expression of these cells with gene profile in parenteral AdAg85A immunization-induced (i.m.) intravascular Ag-specific CD8 T cells at 4 weeks post-immunization and in naïve CD8 T cells. Such comparisons help identify the genes commonly induced by both routes of immunization and those uniquely expressed in respiratory mucosal immunization-induced lung Ag-specific CD8 T cells. Genes encoding for chemokine receptors, integrin heterodimers, and some activation makers (Supplementary Table 1) implicated in T cell trafficking, maintenance and differentiation8 were profiled in FACS sorting-purified Ag-specific (Ag85A-tetramer-positive) CD8 T cells (CD8+tet+T cells) by using a custom-made PCR array (Fig. 1a). The relative gene expression was determined using data from the real-time cycler and the ΔΔCT method as previously described31. For our first goal, that is, identifying gene profile related to immunization, we focused on comparison of i.n. with control and i.m. with control (Fig. 1b). We found a set of genes commonly induced in Ag-specific CD8 T cells by both i.n. and i.m. immunization compared to naïve CD8 T cells (Fig. 1b). However, the genes encoding proteins CCR1, CCR6, CCR8, and CD103 (itage) were uniquely induced in i.n. immunization-induced CD8 T cells (Fig. 1b). Levels of Ccr1, Ccr6, Ccr8 and itage gene expression by i.n. immunization-induced T cells were at least 30-fold higher than those by i.m. immunization (Fig. 1c). In addition, expression of Cxcr5, Ccr7, CD34, CD44 and itga1 (α1 integrin of VLA-1 or CD49a) genes also increased by more than 2 fold in i.n. immunization-induced memory CD8 T cells (Fig. 1c). Taken together, these data indicate that viral vector mediated respiratory mucosal TB immunization induces lung tissue Ag-specific memory CD8 T cells with a unique set of genes that are implicated in T cell mucosal tissue trafficking and maintenance.

Figure 1 Expression of candidate genes by Ag-specific CD8 T cells induced by replication-defective viral-vectored respiratory mucosal immunization. (a) Experimental schema and flow chart showing the workflow. (b) Venn diagram depicts genes that are commonly expressed on both respiratory mucosal (i.n.) and parenteral intramuscular (i.m.) immunization-induced Ag-specific CD8 T cells, and the genes that are uniquely expressed on i.n.- and i.m.-immunization induced Ag-specific CD8 T cells. (c) Bar graph shows mean ± S.E.M. fold changes of genes expressed by i.n. immunization-induced Ag-specific CD8 T cells compared to i.m. immunization-induced Ag-specific CD8 T cells. Data represent mean fold changes calculated from 3 independent experiments. Full size image

Viral-vectored respiratory mucosal immunization induces Ag-specific CD8 T cells expressing T RM surface markers

Based on their unique gene expression profile and differential localities in the lung, we next selected to determine protein expression levels of CCR1, CCR6, CD103 (itage) and CD49a (itga1or VLA-1) on respiratory mucosal immunization-induced Ag-specific memory CD8 T cells at 4 weeks post-immunization. Although some genes such as Cxcr5 and Ccr7 were also increased in these cells, they were not included in our protein expression analysis as they pertains more to the homing of T cells to secondary lymphoid organs32. Nor was CCR8 protein examined due to limited murine immunoreagents. By flow cytometry only a smaller frequency of CD8+tet+T cells (~20%) expressed CCR1 and CCR6 protein in the lung of i.n. immunized animals (Fig. 2a). In sharp contrast, >80% of Ag-specific CD8 T cells expressed T RM surface markers CD103 and CD49a (VLA-1) (Fig. 2a). In consistent with increased frequencies, we also observed significantly higher numbers of Ag-specific CD8 T cells expressing CD103 or CD49a than those expressing CCR1 or CCR6 in the lung (Fig. 2a). In comparison, very few Ag-specific memory CD8 T cells induced by i.m. immunization expressed T RM surface markers CD103 and CD49a (VLA-1). Together, these data demonstrate that respiratory mucosal TB immunization generates Ag-specific T cells with typical properties of T RM cells in the lung.

Figure 2 Protein expression of T RM surface markers by replication-defective viral-vectored respiratory mucosal immunization-induced Ag-specific CD8 T cells in the lung. Lung mononuclear cells from mice immunized with viral vector vaccine via either respiratory mucosal (i.n.) or parenteral (i.m.) route for four weeks were immunostained for surface markers CCR1, CCR6, CD103 and CD49a and analyzed using flow cytometry. (a) Representative dot plots showing frequencies of tet+CCR1, tet+CCR6+, tet+CD103+, tet+CD49a+CD8 T cells out of total CD8+tet+T cells in the lung of i.n. and i.m. immunized mice. (b) Bar graph showing absolute numbers of tet+CCR1, tet+CCR6+, tet+CD103+, tet+CD49a+CD8 T cells in the lung of i.n. and i.m. immunized mice. Data are presented as mean ± S.E.M. of three mice per group, representative of three independent experiments. *P < 0.05, ***P < 0.001, ****P < 0.0001 compared with i.m. immunization. Full size image

Viral-vectored respiratory mucosal immunization-induced Ag-specific CD8 T cells acquires VLA-1 expression in the draining lymph node

Having established that the majority of respiratory mucosal TB immunization-induced lung tissue Ag-specific CD8 T cells express classical resident memory surface markers, CD103 and CD49a5, we sought to systematically examine the kinetic expression of these T RM surface markers during various phases (effector/expansion, contraction and memory) of T cell responses following respiratory mucosal immunization. To this end we first characterized the phases of T cell responses following viral vector immunization. The CD8+tet+T cells in the lung significantly increased at 10 days and peaked at 14 days post-respiratory mucosal immunization, consistent with the effector/expansion phase of T cell responses (Supplementary Fig. 1). Between 14 and 28 days, the number of CD8+tet+T cells markedly decreased by more than 80% from the peak time indicative of the contraction phase of T cell responses. From 28 days until 45 days, the number of CD8+tet+T cells in the lung remained stable, hence being in the memory phase of T cell responses. In comparison, parenteral intramuscular (i.m.) immunization led to much smaller levels of Ag-specific T cell responses in the lung in various phases (Supplementary Fig. 1).

We next examined CD103 and CD49a expression on Ag-specific CD8 T cells in the lung in different phases of T cell responses. The majority of CD8+tet+T cells in the lung of i.n. immunized animals expressed CD49a upon arrival at the lung and in the expansion/effector phase (d10–d14) and became further enriched for CD49a expression in the contraction (d14–d28) and memory phases (d28–d45) (Fig. 3a/c ). In contrast to CD49a, only a small frequency of CD8+tet+T cells (24% at d10) expressed CD103 in the expansion/effector phase and it progressively increased over the contraction and memory phases (46% at d14 up to 85% at d45) (Fig. 3b/c). The expression profile of CD49a and CD103 on airway luminal (BAL) CD8+tet+T cells was identical to that of the lung cells. Upon closer examination, from d28 onward the majority of CD8+tet+T cells in the lung and airway lumen co-expressed both CD49a and CD103 indicating the acquisition of a bona fide T RM property (Supplementary Fig. 2a/b). In comparison, parenteral intramuscular (i.m.) immunization-induced Ag-specific CD8 T cells only temporarily expressed CD49a and mostly lacked CD103 expression in different phases of T cell responses (Fig. 3a/b). These data suggest that following respiratory mucosal TB immunization the Ag-specific CD8 T cells acquired CD49a (VLA-1) expression before their arrival at the lung whereas they acquired CD103 expression after they entered the lung.

Figure 3 Expression of T RM surface markers on replication-defective viral-vectored respiratory mucosal immunization-induced Ag-specific CD8 T cells in different phases of T cell responses. Mononuclear cells from lung, BAL, peripheral blood and mediastinal lymph nodes (MLN) obtained at designated time points post-immunization were immunostained for CD49a and CD103 and analyzed using flow cytometry. (a/b) Representative dot plots showing frequencies of tet+ CD49a+ and tet+ CD103+ CD8 T cells out of total CD8+tet+T cells in the lung of respiratory mucosal (i.n.) and parenteral (i.m.) immunized mice in the effector/expansion phase (d10/d14), contraction phase (d14/d28) and memory phase (d28/d45) of T cell responses. (c) Line graph comparing kinetic changes in the expression of CD49a and CD103 on Ag-specific CD8 T cells in the lung induced by viral vector respiratory mucosal immunization. (d) Line graph comparing kinetic changes in the expression of CD49a and CD103 on Ag-specific CD8 T cells in the bronchoalveolar lavage fluid (BAL) induced by respiratory mucosal immunization. (e) Representative histograms showing frequencies of CD8+tet+T cells expressing CD49a in the blood at d14 and d28 post- viral vector respiratory mucosal immunization. (f) Representative dot plot showing the frequency of CD8+tet+T cells out of total T cells in MLN at d14 post- viral vector respiratory mucosal immunization, and the representative histogram showing the frequency of CD8+tet+T cells expressing CD49a. Data are presented as mean ± S.E.M. of three mice per group per time point, representative of three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 compared with i.m. immunization. Full size image

To determine the geographical origin of CD49a acquisition, we examined the CD49a expression on CD8+tet+T cells in the circulation and mediastinal lymph node (MLN), the draining lymph node of the lung. Indeed, a significant number of CD8+tet+T cells in the blood expressed CD49a in the effector phase of T cell responses (d14) (Fig. 3e), consistent with its marked expression on such T cells primed in MLN (d14) (Fig. 3f). Circulating CD8+tet+T cells continued to show high levels of CD49a expression in the memory phase (d28) (Fig. 3e). These data suggest that although respiratory mucosal TB immunization-induced Ag-specific T cells in the lung co-express both CD49a and CD103, these T RM markers are acquired in distinct tissue sites with CD49a (VLA-1) expressed on respiratory mucosal vaccine-induced T cells even before they home to the lung.

VLA-1 is not required for trafficking of viral-vectored respiratory mucosal immunization-induced Ag-specific CD8 T cells to the lung

Having demonstrated that the prominent CD49a expression on Ag-specific CD8 T cells outside and within the lung, we postulated that VLA-1 played a role in the trafficking of Ag-specific CD8 T cells to the lung. To address this question, we blocked CD49a during the initial stage of T cell activation (d6–d12) before Ag-specific CD8 T cells arrived en masse at the lung (Fig. 4a) by using a well-established CD49a functional blocking antibody (CD49a mAb) delivered via intraperitoneal route33. Analysis of CD49a expression on Ag-specific CD8 T cells in the MLN, blood, lung and BAL confirmed complete blockade of CD49a receptor following delivery of CD49a mAb but the isotype control antibody had no effect (Supplementary Fig. 3). Of interest, CD49a blockade did not change the recruitment of CD8+tet+T cells to the lung and airway lumen during the effector/expansion phase (Fig. 4b). These data suggest that VLA-1 does not play a significant role in T cell trafficking to the lung mucosal sites during the initial phase of T cell activation following viral vector mediated respiratory mucosal TB immunization.

Figure 4 Role of VLA-1 in trafficking of replication-defective viral vector respiratory mucosal immunization-induced Ag-specific CD8 T cells to the lung during the effector phase of T cell responses. (a) Experimental schema depicting the timing of administration of CD49a blocking mAb and isotype control antibody. Blocking/isotype antibodies were administered 6–12 days post- viral vector respiratory mucosal immunization when Ag-specific tet+ CD8 T cells clonally expanded in the draining lymph nodes and in the process of homing to the lung. (b) Representative dot plots showing frequencies of CD8+tet+T cells out of total CD8 T cells in the lung and bronchoalveolar lavage fluid (BAL) of the control and CD49a-blocked mice. Bar graphs comparing absolute numbers of Ag-specific tet+ CD8 T cells in the lung and BAL. Data are presented as mean ± S.E.M. of three mice per group from one experiment. Full size image

VLA-1 negatively regulates viral-vectored vaccine-induced Ag-specific CD8 T cells during the contraction phase in the lung

Integrins such as VLA-1 have previously been implicated in intracellular signalling pathways to regulate cell survival and cell death9, 10 which may be involved in the contraction phase of T cell responses. To determine whether VLA-1 was involved in regulating the contraction of antigen-specific T cells following their effector/expansion responses in the lung, immunized mice were treated with CD49a mAb starting from day 14 post-immunization to block VLA-1 pathway and CD8 + tet+ T cells were examined without Ag re-stimulation at day 18 (Fig. 5a). CD49a blockade led to increased frequencies of CD8+tet+T cells both in the lung and airway (Fig. 5b). It also led to 2–3 times more CD8+tet+T cells in the lung, compared to the isotype control (Fig. 5b). Using an intravascular staining protocol we found that the rise in CD8+tet+T cells in the lung of anti-VLA-1 treated animals occurred only in the lung parenchymal tissue (LPT) while the CD8+tet+T cells in the lung vasculature (LV) remained comparable in numbers to isotype Ab treated animals (Fig. 5c). This finding suggests that VLA-1negatively regulates only the CD8+tet+T cells in the LPT but not those in the LV during the contraction phase.

Figure 5 Role of VLA-1 in regulation of replication-defective viral-vectored respiratory mucosal immunization-induced Ag-specific CD8 T cells in the lung during the contraction phase of T cell responses. (a) Experimental schema depicting the timing of administration of CD49a blocking mAb and isotype control antibody. Blocking and isotype control antibodies were administered at d14 and d16 post- viral vector respiratory mucosal immunization when Ag-specific CD8 T cells in the lung sharply declined. Mice were sacrificed 3 min after i.v. injection of fluorochrome conjugated CD45.2 mAb to differentiate T cells in the lung vasculature from those located in the lung parenchyma. In separate experiments, the mice were treated as above and bromodeoxyuridine (BrdU) was administered i.n. for consecutive four days to assay the in vivo proliferation rate of Ag-specific CD8 T cells. (b) Representative dot plots showing frequencies of CD8+tet+T cells out of total CD8 T cells in the lung and bronchoalveolar lavage fluid (BAL). Bar graphs comparing absolute numbers of Ag-specific tet+ CD8 T cells in the lung and BAL between isotype control antibody and CD49a mAb treated mice. (c) Representative dotplots showing frequencies of CD8+tet+T cells out of total CD8 T cells in the lung parenchymal tissue (LPT) and lung vasculature (LV). Bar graphs comparing absolute numbers of Ag-specific CD8+tet+T cells in the LPT and LV between isotype control antibody and CD49a mAb treated mice. (d) Representative histograms comparing frequencies of BrdU+ proliferating Ag-specific CD8 T cells in lung and BAL between isotype control antiobody and CD49a mAb treated mice. Data are presented as mean ± S.E.M. of three mice per group, representative of two independent experiments. *P < 0.05, compared to isotype control group. Full size image

That VLA-1 blockade increases the number of vaccine-activated CD8+tet+T cell during the contraction phase raised the question whether VLA-1 affects T cell contraction via regulating T cell proliferation. To test this possibility, in separate experiments we examined the proliferation rate of Ag-specific CD8 T cells with and without CD49a blockade by using an in vivo BrdU incorporation assay. T cell BrdU labeling was accomplished following repeated intranasal deliveries of BrdU (Fig. 5a). We found the rates of CD8+tet+T cell proliferation in the lung and BAL of animals with CD49a blockade (CD49a mAb) were comparable to isotype controls (Fig. 5d), suggesting that VLA-1 impacts T cell contraction independent of regulation of T cell proliferation.

Apart from tetramer specificity and proliferation of CD8+tet+T cell following CD49a blockade, we further examined other functional properties of CD8 T cells including IFN-γ production and degranulation as indicated by CD107a expression upon ex vivo Ag re-stimulation. Indeed, compared to their control counterparts, in vivo CD49a blockade led to significantly increased frequencies and numbers of Ag-specific CD8 T cells capable of IFN-γ production (Fig. 6a) and degranulation (Fig. 6b) upon ex vivo Ag re-stimulation. Furthermore, we found that in vivo CD49a blockade also led to significantly increased production of IFN-γ and degranulation marker CD107a per cell basis measured by mean fluorescent intensity (MFI) of signals (Fig. 6a/b).

Figure 6 Role of VLA-1 in regulation of effector functions of replication-defective viral-vectored respiratory mucosal immunization-induced Ag-specific CD8 T cells in the lung during the contraction phase of T cell responses. Experimental conditions were described in Fig. 5a except that the cells were ex vivo re-stimulated with Ag85A antigens. (a) Representative dot plots showing frequencies of CD8+ IFN-γ+ T cells out of total CD8 T cells in the lung and BAL. Bar graphs comparing absolute numbers and IFN-γ mean fluorescence intensity (MFI) of CD8+ IFN-γ+ T cells in the lung and bronchoalveolar lavage fluid (BAL) between isotype control antibody and CD49a mAb treated mice. (b) Representative dot plots showing frequencies of CD8+ CD107a+ degranulating T cells out of total CD8 T cells in the lung and BAL. Bar graphs comparing absolute numbers and CD107a MFI of CD8+ CD107a+ T cells in the lung and BAL between isotype control antibody and CD49a mAb treated mice. (c) Western blot depicts levels of Ag85A protein and β-Tubulin in the total lung of 3 mice (1–3) treated with anti-CD49a mAb, 3 mice (1–3) treated with isotype antibody, a mouse immunized for 3 days (3dpi) and an unimmunized mouse (Naïve). Bar graph shows relative levels of Ag85A calculated in respect to β-Tubulin using Image Studio Lite. Data are presented as mean ± S.E.M. of three mice per group, representative of two independent experiments except western plot data, which is representation of one experiment. *P < 0.05, **P < 0.01 compared to isotype control group. Full size image

That VLA-1 blockade increased the number of vaccine-activated CD8+tet+T cell during the contraction phase also raised the question whether VLA-1 affects T cell contraction via altering the viral clearance as such affecting the level of Ag85A protein in the lung. Thus, in a separate experiment we quantified the antigenic (Ag85A) load in the lung following immunization in isotype and CD49a blocking antibody treated animals (Fig. 5a). Mouse either left without immunization (naïve) or immunized and sacrificed 3 days post-immunization was used as negative and positive controls, respectively. Western blot analysis of the total lung protein showed comparable levels of Ag85A protein in isotype and CD49a blocking antibody treated animals, suggesting that the differential contraction level of CD8+tet+T cell in CD49a blocking antibody treated animals is independent of antigenic load in the lung (Fig. 6c).

Taken together these data suggest that VLA-1 pathway plays a critical role in negatively regulating Ag-specific CD8 T cell responses during the contraction phase following viral vector based respiratory mucosal TB immunization.

VLA-1 is not required for maintenance of viral-vectored respiratory mucosal immunization-induced Ag-specific T RM during the memory phase in the lung

VLA-1 was previously shown to contribute to the retention of Ag-specific CD8 T cells in the lung tissue after influenza infection12. We have shown that immune protective Ag-specific CD8 T cells induced by viral vector respiratory mucosal immunization persist in the lung for a long time34, 35 and we have here found these cells to be of T RM phenotype (Fig. 2/3). We thus next determined whether VLA-1 also played a role in maintenance of viral vector vaccine-induced T RM cells in the lung. To test this, CD49a receptor was functionally blocked for a total of four days beginning from day 28 of the memory phase following respiratory mucosal immunization and CD8+tet+T cells were examined at day 32 (Fig. 7a). We found that CD49a blockade during the memory phase had no effect on both the frequencies and numbers of CD8+tet+T cells in the lung and airway, compared to the isotype control animals (Fig. 7b). These data suggest that VLA-1 is not required for the retention of Ag-specific tissue resident memory CD8 T cells in the lung following replication-defective viral vector respiratory mucosal TB immunization.