Durable lymph-node expansion is associated with the efficacy of therapeutic vaccination – Nature.com

Vaccine formulation alters the durability of LN expansion

First, we identified a vaccine formulation eliciting robust and durable LN expansion. Mesoporous silica (MPS) rod-based vaccines, previously found to elicit strong cellular and humoral responses against diverse antigen targets compared with a traditional bolus (liquid) vaccine, were explored18,19,20,21. These high-aspect ratio, silica-based nanoparticles can adsorb vaccine antigens and adjuvants for sustained release, and form a three-dimensional scaffold promoting antigen-presenting cell (APC) recruitment in mouse models. MPS vaccines previously induced potent and long-lived germinal centre responses dependent on sustained antigen release from the vaccine site22,23. Here, MPS rods used in vaccine formulation had an average length of 85.9m and released vaccine components cytosine-guanosine oligodeoxynucleotide (CpG) and granulocyte-macrophage colony-stimulating factor (GM-CSF) in a sustained manner (Extended Data Fig. 1ae). Draining (dLN; ipsilateral to vaccine site) and non-draining (ndLN; contralateral) inguinal LNs of mice immunized with MPS or bolus vaccines were imaged for 100days post-vaccination using HFUS.

Although PBS injection did not affect LN volume, both vaccine formulations induced LN expansion, but with markedly different durability (Fig. 1a,b and Supplementary Fig. 1ac). At the early stage of expansion (within days), both MPS and bolus-vaccinated mouse LNs expanded to a similar extent (Fig. 1c). However, while the bolus vaccine LNs peaked at this time, resulting in a two-fold transient increase in LN volume, the MPS vaccine induced a significantly more substantial (~7) LN expansion over 1week which was maintained for ~3weeks (Fig. 1b,c). Although LN volume in the MPS-vaccinated mice began to decrease ~20days after immunization, it remained elevated out to 100days (Supplementary Fig. 1d). NdLNs did not change in volume with either vaccine, and normalizing the dLN to ndLN volume within each mouse indicated a similar pattern of dynamic LN expansion and contraction (Fig. 1d and Supplementary Fig. 1e). The removal of either CpG or GM-CSF from the vaccine formulation diminished the magnitude of dLN expansion (Extended Data Fig. 2a). An MPS vaccine with log-fold lower doses of ovalbumin (OVA) and CpG also induced long-term LN expansion (Extended Data Fig. 2b). While other published depot-based vaccine formulations including alum, MF59 emulsion and cryogel-based scaffolds also induced LN expansion, expansion was notably lower than with the MPS vaccine (Extended Data Fig. 2c). The MPS vaccine formulation was thus selected as a model of strong vaccination resulting in persistent LN expansion for subsequent investigation.

Mice were immunized with MPS or bolus vaccines delivering GM-CSF, CpG and OVA protein, and compared to PBS-injected controls. Vaccine-draining and non-draining LNs were longitudinally imaged using HFUS. a, Representative HFUS images of vaccine-draining LNs (defined by yellow dashed area) out to 100days after vaccination. Scale bar, 2 mm. b, Quantification of vaccine-draining LN volume over time. Statistical analysis was performed using a two-way analysis of variance (ANOVA) with repeated measures. Significance relative to the PBS group is depicted at each timepoint (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001). Exact P values between MPS and PBS are P=0.04, day 3; P=0.0008, day 5; P=0.006, day 7; P=0.01, day 9; P=0.001, day 11; P=0.004, day 13; P=0.004, day 15; P=0.03, day 17; P=0.01, day 37; P=0.001, day 44; P=0.01, day 62. Exact P values between bolus and PBS are P=0.009, day 1; P=0.02, day 5; P=0.03, day 9; P=0.008, day 11; P=0.04, day 44. c, Plots of LN volume among groups on days 3 (left) and 19 (right). Statistical analysis was performed using ANOVA with Tukeys post hoc test. d, Representative HFUS images of MPS or bolus vaccine-draining or non-draining LNs 15days after vaccination (left) and quantification of dLN/ndLN volume ratio (right). Statistical analysis was performed using ANOVA with Tukeys post hoc test. For ad, n=7 (MPS and bolus) or 8 (PBS) biologically independent animals per group, imaged longitudinally in two cohorts; meanss.d.

Source data

To assess tissue-scale alterations involved in durable LN expansion, LN mechanical properties and extracellular matrix (ECM) distribution were next characterized. Here, MPS-vaccinated mouse dLNs were collected 7days after immunization, beyond the initial expansion phase (when MPS outpaced bolus vaccine LN expansion). At this time, LN collagen architecture was largely maintained, as expected (Fig. 2a)6. Hyaluronic acid (HA) localization was increased in the periphery/follicle, most visibly 7days after immunization, although still notable up to 3weeks later, demonstrating persistent alterations (Fig. 2a,b and Supplementary Fig. 2a). In contrast, the cellular F-actin signal was greater towards the centre of both control and MPS dLNs, with greater polarization between the centre and periphery in the MPS condition (Fig. 2c and Supplementary Fig. 2bd). These changes suggest that LN expansion may be accompanied by changes in tissue mechanical properties, as both HA and F-actin are involved in cellular mechanotransduction and signalling pathways. Through nanoindentation of thick (~500m) LN slices (Fig. 2d), we found that LNs with enduring expansion had reduced stiffness (G) and loss modulus (G) compared with control LNs (Fig. 2e,f). Viscoelasticity, measured by G/G (tan()), was significantly increased in MPS dLNs compared with LNs from control mice, suggesting decreased matrix crosslinking (Fig. 2g).

Mice were treated with MPS vaccines (delivering GM-CSF, CpG, OVA) or PBS, and LNs were collected after 7 and 20days. a, Representative immunohistochemistry (IHC) images depicting LN ECM on day 7. b, Representative IHC image depicting LN ECM on day 20. c, Representative IHC images of LNs stained for F-actin on day 7. For ac, n=3 biologically independent animals per group. d, Schematic depicting nanoindentation of a thick LN slice (above) and experimental timeline (below). eg, Mean G (e), G (f) and tan() (g) across LNs. Statistical analysis was performed using MannWhitney test (e) or two-tailed t-test (f,g). h, Heat maps depicting G across individual LNs. Scale bar, 1mm. i, Mean G of sample points across each LN, separated into those collected at the centre or periphery. n=11 (control, centre), 10 (MPS, centre), 16 (control, periphery) and 15 (MPS, periphery) biologically independent animals per group; results are means.d., combined from three independent experiments. Statistical analysis was performed using MannWhitney test. j, Plot of LN mass versus mean G. For eg, i and j, each data point represents a unique LN per mouse; n=10 (MPS) or 11 (PBS) biologically independent animals per group; means.d., combined from two independent experiments. k, Representative IHC images depicting Hoechst stain within LNs on day 7. Scale bar, 100m. l, Quantification of Hoechst signal across LNs; n=3 (MPS) or 4 (PBS) biologically independent animals per group; means.d.

Source data

Spatial variations in mechanics across LNs were next investigated using nanoindentation (Fig. 2h and Supplementary Fig. 3). Both control and MPS-vaccinated mouse LNs were softer and more viscoelastic in the centre than in the periphery, and this finding was confirmed through intentional sampling at the centre or periphery of nave LNs (Supplementary Fig. 4ae). The LN periphery (~12kPa) was approximately twice as stiff as the centre (~6kPa). Interestingly, after vaccination, LN G and G were only significantly altered at the periphery, while tan() increased only in the LN centre (Fig. 2i and Supplementary Fig. 5ae). LN peripheral stiffness correlated negatively with LN mass, suggesting that the degree of tissue softening relates to the extent of LN enlargement induced by vaccination (Fig. 2j). LN cellular distribution and tissue density remained unaltered, despite expansion (Fig. 2k,l and Supplementary Fig. 6ad). Taken together, these results suggest that LN tissue encompasses a range of mechanical properties, dependent on location within the node, and these parameters change as LNs expand.

Considering that tissue-level changes may impact or reflect cellular responses, changes in LN cellularity during expansion were next characterized, ranging from early-stage (day 4) to long-term (day 51) changes (Supplementary Figs. 7 and 8a). Cellular expansion was greater and more sustained in MPS-vaccinated mice than in the bolus-vaccinated mice or PBS-injected control; notably, the total cell counts within a LN correlated with its volume (Fig. 3a,b). As early as day 4, monocytes, neutrophils and macrophages were expanded in MPS dLNs, while conventional dendritic cells (DCs), plasmacytoid DCs and T cells peaked at day 7 before declining over time (Fig. 3c,d and Supplementary Fig. 8bi). Monocytes in particular expanded ~80-fold in MPS dLNs compared with PBS controls 4 days after vaccination, relative to ~25 expansion in the bolus group, but this increase was maintained for several weeks in the MPS condition only (Supplementary Fig. 8n). B cells also significantly expanded by day 7 and remained elevated until day 17 (Fig. 3e). A variety of stromal cells expanded following MPS vaccination, typically peaking later (days 1117) than the immune cells, except for natural killer (NK) cells, which also tended to expand later (days 711) (Fig. 3f and Supplementary Fig. 8jm). By comparison, changes in the bolus vaccine group were more modest beyond 4days, and PBS-treated control dLNs and ndLNs from all groups demonstrated minimal changes in cell populations. These results indicate that vaccine-induced LN expansion engages the temporal dynamics of a pathogen-induced immune response, with innate immune cells rapidly responding followed by lymphocytes at later times.

Mice were immunized with MPS or bolus vaccines containing GM-CSF, CpG and OVA protein, euthanized on days 4, 7, 11, 17 and 51 for LN collection and analysis through flow cytometry and compared to PBS-injected controls. a, Total LN cell counts over time. b, Linear regression of LN cell count on a given day versus volume (measured through HFUS). cf, Numbers of dendritic cells (c), T cells (d), B cells (e) and follicular dendritic cells (FDCs; CD45 CD31 CD21/35+) (f) over time. For af, n=4 (MPS dLN days 4, 11 and MPS ndLN day 7) or 5 (all other timepoints and groups) biologically independent animals per group per timepoint; means.d. For a and ce, statistical analysis was performed using ANOVA with Tukeys post hoc test; differences present between one group and all other groups are shown. For f, statistical analysis was performed using KruskalWallis test with Dunns post hoc test; the statistical difference between the MPS and PBS dLN groups is shown. For gj, mice were injected with MPS or bolus vaccines (GM-CSF, CpG, OVA) and dLNs were collected at a late timepoint (days 2021). Nave mice were included as controls. n=5 biologically independent animals per group, barcoded and pooled for sequencing. g, Schematic of processing pipeline for single-cell sequencing. LNs were digested and FACS-sorted to enrich live, CD45+ CD3 CD19 cells for sequencing. h, UMAP of 20,858 cells across conditions coloured by cluster membership. i, UMAP as in h, here coloured by cell density. Red indicates high cell density, blue low density. j, Heat map of relative average expression of marker genes in each cluster from h. Colour bar indicates relative gene expression as z-score. a.u., arbitrary units. k, Frequency of individual cell clusters within each sample. Statistical analysis was performed using ANOVA with Tukeys post hoc test. pDCs, plasmacytoid DCs; Mig., migratory; Infl., inflammatory.

Source data

Because LN expansion is known to be mediated by myeloid interactions with LN stromal cells, we next performed single-cell RNA sequencing (scRNA-seq) on the LN myeloid compartment after vaccination (Fig. 3g and Supplementary Fig. 9a)2. LNs were examined at a late timepoint (days 2021) to consider mediators of durable expansion. After removal of lymphocytes and stromal cells, we identified nine clusters from the remaining 20,858 cells analysed (Fig. 3hj). Clusters were annotated as type-2 conventional DCs (cDC2s; c0, Sirpa, H2-Ab1), plasmacytoid DCs (c1, Siglech, Bst2), migratory DCs (c2, Ccr7, Clu), type-1 conventional DCs (c3, Xcr1, Clec9a), Langerhans cells (c4, Cd207), plasma cells (c5, Ighg2b, Ighg1), inflammatory monocytes (c6, Csf1r, Ly6c2), neutrophils (c7, S100a8, S100a9) and proliferating cDC2s (c8, Top2a, Mki67) (Fig. 3j and Supplementary Fig. 9b). Consistent with the flow cytometry analysis, scRNA-seq identified broad changes in LN cell populations after immunization, with notable differences based on vaccine strength (Fig. 3i,k). Compared with the PBS condition, both bolus and MPS vaccines increased DC2 proportions and decreased frequencies of migratory DCs and DC1s. Maintenance of LN expansion was associated with increased frequencies of inflammatory monocytes and plasma cells and decreased Langerhans cells (Fig. 3k).

Given the importance of sustained antigen presentation in maintenance of LN immune responses24,25, we hypothesized that vaccine antigen availability and APC populations may affect LN expansion. Compared with LNs of mice given the full MPS vaccine, LNs of mice given an MPS vaccine without antigen became prominently less enlarged and contracted sooner (Fig. 4a,b and Extended Data Fig. 3a). This indicates that long-term antigen presentation at the vaccine site is important for sustained LN expansion. Indeed, injecting the antigen separately as a bolus (that is, not delivered from the MPS scaffold) similarly reduced the degree and duration of expansion, indicating a critical role of sustained antigen presentation (Extended Data Fig. 3b).

a,b, Mice were immunized on day 0 with a full MPS vaccine (containing GM-CSF, CpG and OVA protein) or an MPS vaccine without antigen (GM-CSF and CpG only). LN volume was tracked using HFUS imaging. n=5 biologically independent animals per group. a, Representative HFUS images of vaccine-draining LNs. Scale bar, 2 mm. b, Quantification of LN volume over time. Statistical analysis was performed using two-tailed t-tests. For a and b, n=5 biologically independent animals per group. cj, Mice were injected with MPS or bolus vaccines (GM-CSF, CpG, OVA) and dLNs were collected at a late timepoint (days 20 and 21). Nave mice were included as controls. n=5 biologically independent animals per group, barcoded and pooled for sequencing. c, Numbers of differentially expressed protein coding genes by cell type between the MPS and bolus conditions. P value calculated using DESeq2. d, Volcano plot displaying differentially expressed cDC2 genes between the MPS and bolus conditions. e, Ltb (lymphotoxin ) expression among DC subtypes in the different conditions. f, Proportion of inflammatory monocytes (cluster 6 from Fig. 3h) in LNs. g, UMAP of 1,468 inflammatory monocytes coloured by cluster membership. h, UMAP as in g, here coloured by cell density. Red indicates high cell density, blue low density. i, Proportion of cluster c0 among inflammatory monocytes. j, Pathway analysis for inflammatory monocyte cluster c0. For c and d, statistical analysis was performed with DESeq2. For f and i, statistical analysis was performed using ANOVA with Tukeys post hoc test.

Source data

To identify potential mediators of this differential response, we next focused the analysis of our scRNA-seq dataset on LN APC populations. Broadly, we identified varying numbers of differentially expressed genes within immune cell clusters between the MPS and bolus vaccine conditions (Fig. 4c). The most dramatic transcriptional changes were in the LN-resident cDC2 and cDC1 compartments, more so than in migratory DCs and Langerhans cells. The cDC2s showed the greatest number of differentially expressed protein coding genes between the two vaccine strengths (Fig. 4c,d), and consensus non-negative matrix factorization (cNMF) analysis26 identified a cDC2-specific programme (CNMF_X14) enriched with MPS vaccination (Supplementary Fig. 10ad). This programme included genes involved in inflammation (Il1r2m, Cd86), immune regulation (Clec4a2, Sirpa, Lst1), cell migration machinery (Rasgef1b, Elmo1) and smooth muscle contraction (Ppp1r14a) (Supplementary Fig. 10d). Furthermore, the gene encoding lymphotoxin beta (Ltb), another member of CNMF_X14, was strongly upregulated in cDC2s with MPS vaccination relative to the bolus condition (Fig. 4d,e and Supplementary Fig. 10d,e). The involvement of mechanosensing genes and Ltb, involved in lymphoid organogenesis, suggests that cDC2s may both respond and contribute to the changing LN microenvironment during expansion. Despite robust LN expansion and immune activation, Cd274 (PD-L1) was not notably upregulated on myeloid cell subsets 20days after immunization (Supplementary Fig. 11a,b). MPS immunization also increased the frequency of CD19 plasma cells and directed gene expression towards more mature immunoglobulin (Ighg1 versus Ighm) expression (Supplementary Fig. 12ad).

Inflammatory monocytes demonstrated significant transcriptional changes between the MPS and bolus vaccine groups (Fig. 4c), and the greatest expansion by both total number and relative proportion following MPS vaccination (Fig. 4f and Supplementary Fig. 8n). Therefore, we were interested in how MPS vaccination affected their gene expression profile. Monocytes, similar to DCs, can present antigen to T cells in LNs27, and particular attention was paid to potential T cell interactions. Monocyte-specific clustering identified three subpopulations of inflammatory monocytes (Fig. 4g,h). Of these, c0 formed the predominant monocyte phenotype in LNs with sustained expansion (MPS condition) relative to nave or bolus-vaccinated mice (Fig. 4h,i). Gene set enrichment analysis identified pathways associated with antigen processing and presentation, IFN response and inflammatory signalling that differentiated monocytes in the strong and weak vaccine LNs (Fig. 4j). These gene alterations position inflammatory monocytes as a potential stimulatory, APC type involved in sustained LN expansion.

To confirm the impact of vaccine strength on antigen-presenting, inflammatory monocytes, LNs of mice vaccinated with the MPS vaccine (with or without antigen) were collected and further compared to LNs of mice given bolus or PBS controls (Fig. 5a). Consistent with the scRNA-seq analysis, Ly6Chi inflammatory monocytes27,28 comprised the majority (~6070%) of LN monocytes in the MPS group over time, significantly higher than the PBS and bolus groups (~4050%) by day 20 (Supplementary Fig. 13a,b). Inflammatory monocytes were also significantly expanded in terms of number and proportion in the LNs of MPS-vaccinated mice at day 20 compared with the PBS and bolus, and were visualized in LNs through CCR2 expression (Fig. 5b and Supplementary Fig. 13c,d)29. Inflammatory monocyte responses were abrogated at later timepoints when the MPS vaccine was delivered without antigen, equivalent to the PBS or bolus controls by day 20, suggesting a relationship between long-lived antigen presentation, LN expansion and monocyte responses (Fig. 5b). Consistent with scRNA-seq data and previous investigation on the MPS vaccine system, MPS immunization elicited robust and persistent germinal centre B cell responses, also dependent on the presence of antigen in the vaccine (Extended Data Fig. 4ac).

Mice were treated with MPS or bolus vaccines (containing GM-CSF, CpG, OVA), MPS vaccine without antigen (GM-CSF, CpG only) or PBS, and LNs were collected on days 7, 14 and 20 for cellular analysis. n=5 biologically independent animals per group per timepoint. a, Experimental timeline and conditions. b, Inflammatory monocyte number in LNs over time. Statistical analysis was performed using ANOVA with Tukeys post hoc test. c, Representative flow cytometry histograms depicting MHCII expression on Ly6hi inflammatory monocytes. Median percentage MHCII expression in each group is listed on the right. d, MHCII expression on Ly6hi inflammatory monocytes in the LN at day 20. Statistical analysis was performed using ANOVA with Tukeys post hoc test. For b and d, meanss.d. eh, Mice were administered MPS vaccines (containing GM-CSF, CpG, OVA) or PBS. One group of MPS-vaccinated mice was treated with MC-21 CCR2-depleting mAb daily from days 15 (MC-21 expansion) and one group was treated daily from days 1014 (MC-21 maintenance). Peripheral blood was collected on days 6, 8, 14 and 20 for cellular analysis. n=5 biologically independent animals per group. e, Experimental timeline and conditions. f, Inflammatory monocyte proportion in blood over time. Differences between groups are statistically significant (day 6 MPS versus MPS/MC-21 expansion, P=0.005; day 6 MPS versus PBS, P=0.03; day 8 MPS versus PBS, P=0.001; day 14 MPS versus MPS/MC-21 maintenance, P<0.0001; day 14 MPS versus PBS, P=0.002). Significant differences between the MPS group and other groups are indicated on the figure (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001). g, Proportion of effector CD8+ T cells (CD44+ CD62L) in blood over time. h, Proportion of OVA-tetramer+ of CD8+ T cells in peripheral blood 20days after vaccination. Statistical analysis was performed using ANOVA with Tukeys post hoc test. For fh, meanss.d. For f and g, statistical analysis was performed using KruskalWallis test with Dunns post hoc test (day 6 timepoint) or ANOVA with Tukeys post hoc test (days 8, 14, 20). For b, f and h, only differences between one group and all other groups are shown (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001).

Source data

Unlike LNs, spleens did not demonstrate superior cellular expansion after MPS vaccination compared with other vaccine groups (Extended Data Fig. 5a). Although total numbers of splenic immune cells including B cells and DCs were largely unaffected by vaccination, transient increases in T cells and macrophages were detected (Extended Data Fig. 5be). Notably, significantly higher numbers and proportions of inflammatory monocytes were found in MPS-vaccinated mouse spleens compared with all other conditions on day 20 (Extended Data Fig. 5f,g). These cells also remained elevated in circulation at the latest timepoint (Extended Data Fig. 5h).

Inflammatory monocytes in the MPS group displayed characteristics of antigen presentation; MHCII expression significantly increased in the MPS vaccine group compared with all others several weeks after vaccination (Fig. 5c,d). Numbers of monocyte-derived DCs (CD11c and MHCII-expressing Ly6Chi monocytes) were also significantly increased in the MPS-vaccinated dLN at this time compared with PBS-treated mice, or any condition in the spleen (Extended Data Fig. 6a). In the spleen, MHCII expression on inflammatory monocytes was unaltered with vaccination (Extended Data Fig. 6b). These results indicate that Ly6Chi monocytes induced by MPS vaccination may engage in antigen presentation, specifically within the LN compartment.

To further discern the impact of inflammatory monocytes on lymph-node expansion and vaccine response, specific depleting reagents were next employed. MPS-vaccinated mice were treated with the CCR2-targeting MC-21 monocolonal antibody (mAb)30,31,32 either early (days 15, LN expansion phase) or later (days 1014, LN maintenance phase) after immunization (Fig. 5e). MC-21 mAb effectively depleted Ly6Chi monocytes in the blood, LN and MPS scaffold during the treatment course, although numbers in the blood rebounded within days (Fig. 5f and Supplementary Fig. 14ac). Early depletion of Ly6Chi monocytes delayed the effector CD8+ T cell response to vaccination, which peaked later, after monocytes had been restored, relative to the MPS vaccine group (Fig. 5g). Furthermore, only the MPS-vaccinated group treated early with MC-21 antibody had significantly elevated tetramer-specific CD8+ T cells by day 20, after the monocyte rebound, compared with the PBS controls (Fig. 5h). Administration of MC-21 mAb in the later phase of the LN response (days 1014) had no discernible impact on the T cell response. These results further suggest a role of inflammatory monocytes in effector CD8+ T cell responses to MPS vaccination, potentially through direct antigen presentation or inflammatory stimulation.

LN expansion kinetics in the absence of inflammatory monocytes or other immune cell subsets were next assessed. MC-21 mAb and/or clodronate liposomes were used to deplete Ly6Chi monocytes and macrophages, respectively (Extended Data Fig. 7a). Lymphocyte (anti-CD4, CD8 and B220) and neutrophil (anti-Ly6G) antibodies were also tested. No differences were observed in the magnitude or kinetics of LN expansion with depletion of any immune cell subset alone (Extended Data Fig. 7bg). However, depleting both inflammatory monocytes and macrophages together restrained the maintenance of LN expansion (Extended Data Fig. 7h). Taken together, these data indicate a stimulatory and antigen-presenting role of inflammatory monocytes, and that these cells in association with macrophages may be required for sustained LN expansion.

Finally, we considered whether durable LN expansion could indicate functional outcomes of vaccination. In a therapeutic model of mouse melanoma, LN expansion after vaccination against a tumour-expressed antigen was not affected by tumour presence (Supplementary Fig. 15ac). The MPS vaccine generated stronger adaptive immune responses than the bolus vaccine, leading to therapeutic benefit (Fig. 6ac and Supplementary Figs. 15dg and 16ac). Importantly, LN expansion associated positively with antibody titres, CD8+ T cell responses and antitumour efficacy of cancer vaccine formulations (Fig. 6df and Supplementary Fig. 17ac). The degree of LN expansion also correlated strongly with effector CD8+ T cell proportions following vaccination across experiments (Supplementary Fig. 17d). In a tumour-free setting, MPS vaccination also enhanced long-term antibody production (day 90) and splenic CD8+ T cell (day 103) responses as compared with the bolus vaccine, and responses associated with earlier degree of LN expansion (Extended Data Fig. 8af). Sustained inflammatory cytokine expression in splenic CD8+ T cells suggested a long-lived adaptive immune response in multiple lymphoid organs.

Mice were inoculated with B16-OVA melanoma tumours and 3days later treated with MPS or bolus vaccines containing GM-CSF, CpG and OVA protein, and compared to PBS-injected controls. A fourth group of tumour-free mice was treated with MPS vaccines (called MPS, no tumour). Inguinal dLNs were imaged using HFUS at multiple timepoints, and blood was collected 8 and 21days after vaccination to assess T cell responses and serum antibody titres, respectively. n=6 (MPS, B16-OVA) or 5 (all other groups) biologically independent animals per group. a, Serum anti-OVA IgG2a antibody titre 21days after immunization. Statistical analysis was performed using KruskalWallis test with Dunns post hoc test. b,c, T cell analysis in the peripheral blood 8days after immunization. b, Representative flow cytometry plots of OVA-tetramer binding to CD8+ T cells. c, Proportion OVA-tetramer+ of CD8+ T cells in blood. Statistical analysis was performed using ANOVA with Tukeys post hoc test. For a and c, meanss.d. d, LN fold expansion 7days after vaccination versus blood IFN+ CD8+ T cell response to SIINFEKL restimulation 8days after vaccination. e, LN fold expansion 7days after vaccination versus anti-OVA IgG1 titres 21days after vaccination. f, LN fold expansion 7days after vaccination versus tumour area at the latest timepoint with all mice surviving (day 21).

Source data

We next assessed potential indicators of toxicity or T cell dysfunction that could result from sustained LN expansion. In MPS-vaccinated mice, serum HMGB-1 levels, indicative of inflammatory cytokine responses and/or cellular death33,34, were comparable to PBS controls (Extended Data Fig. 8g). Long-term (day 103) PD-1 expression on splenic T cells was also not different between the MPS vaccine group and PBS controls (Extended Data Fig. 8h,i). Mice monitored for 485days after MPS vaccination did not display changes in weight, LN or spleen cell counts, or proportions of immune cell subsets in blood or secondary lymphoid organs, although elevated OVA-specific CD8+ T cells remained detectable in all immune compartments investigated (Extended Data Fig. 9ai). Furthermore, MPS vaccine-generated T cells retained functional, antigen-specific antitumour response when challenged 50 days after immunization (Extended Data Fig. 10ac). Altogether, these results suggest that enduring LN expansion is associated with immune memory and antitumour efficacy, without indications of T cell dysfunction.

To explore whether LN expansion could directly improve vaccine efficacy, the MPS vaccine without antigen (Fig. 4a,b) was employed to jump-start LN expansion before administration of a full, antigen-containing bolus vaccine (Fig. 7a). LNs of mice given the antigen-free MPS jump-start expanded over the first week and continued to increase in size after administration of the bolus vaccine, becoming significantly enlarged compared with all other groups (Fig. 7b). The jump-start plus bolus vaccine broadly improved vaccine responses compared with the traditional bolus vaccine. The proportion of OVA-tetramer+ CD8+ T cells in blood was significantly increased in this condition (Fig. 7c,d). Blood CD8+ T cells restimulated ex vivo with SIINFEKL peptide had superior cytokine production (IFN and TNF) with the jump-start (Supplementary Fig. 18ac), and the jump-start also increased the proportion of effector CD8+ T cells and decreased the blood CD4/CD8 T cell ratio relative to mice given the bolus vaccine alone (Supplementary Fig. 18d,e). The combination treatment improved short- and long-term IgG2a antibody titres, with 10/10 (versus 6/10 with the bolus only) detectable IgG2a titres after 100 days (Fig. 7e and Supplementary Fig. 18f). In these experiments, the jump-start was dosed 7days before the bolus vaccine to match the peak of LN enlargement (Supplementary Fig. 19a). Spacing the jump-start closer to bolus vaccination (4days) tended to increase antigen-specific cytokine expression (IFN and TNF) and OVA-tetramer binding; however, increasing the dose separation (11days) increased granzyme B and reduced PD-1 expression, suggesting that the timing of jump-start and bolus vaccination can alter functional T cell outcomes, and the day 7 timepoint balances both sets of outcomes (Supplementary Fig. 19bh). All additional experiments were conducted with a 7-day spacing. In treating B16-OVA tumour-bearing mice, the jump-start strategy (Supplementary Fig. 20a,b) elicited prolonged tumour regressions compared with the bolus vaccine, which induced only transient tumour regressions, with all mice in this condition eventually succumbing to tumour burden within 50days. In the jump-start plus bolus group, 25% of mice survived at 200days, a significant improvement over all other groups (Fig. 7f,g). In summary, jump-starting LN expansion before vaccine administration improved T cell responses and antitumour efficacy in a model antigen tumour model.

a, Experimental timeline for be; mice were injected with PBS or a bolus vaccine on day 0, or injected with an MPS no-antigen jump-start on day 7 followed by PBS or a bolus vaccine (GM-CSF, CpG and OVA protein) on day 0. Mice were bled after 8 and 21days for T cell analysis and serum antibody titres, respectively. b, LN expansion measured by HFUS imaging. Values are normalized to the baseline volume for each individual LN. n=5 biologically independent animals per group; only differences between one group and all other groups are shown. c, Representative flow cytometry plots depicting CD8+ T cell OVA-tetramer binding in cells derived from blood on day 8. d, OVA-tetramer+ proportion of CD8+ T cells. e, Anti-OVA IgG2a titre on day 21. Statistical analysis was performed using KruskalWallis test with Dunns post hoc test. f,g, Mice bearing B16-OVA tumours (inoculated on day 8) were treated starting at day 7 as per studies in ae with an MPS jump-start (MPS material, GM-CSF and CpG without antigen) or left untreated, and then injected with a bolus vaccine (GM-CSF, CpG and OVA) or left untreated at day 0. Tumour growth and survival were tracked. f, Tumour growth curves. n=10 biologically independent animals per group. g, KaplanMeier curves depicting survival. n indicates naive/non-vaccinated (n+bolus, naive + bolus vaccine; jump-start+n, jump-start + naive; n+n, naive + non-vaccinated). Statistical analysis was performed using log-rank (MantelCox) test, correcting for multiple comparisons. n=17 (n/n) or 16 (all other groups) biologically independent animals per group; results are combined from two independent experiments, the second performed in a blinded manner. For b and d, statistical analysis was performed using ANOVA with Tukeys post hoc test. For ce, n=9 (PBS) or 15 (all other groups) biologically independent animals per group; results are combined from two independent experiments. For b, d and e, meanss.d.

Source data

Finally, we assessed the impact of a booster vaccine format on LN expansion kinetics and adaptive immune responses. Following the MPS prime vaccine, dLN volume increased over the following 12weeks and declined by day 42 (Supplementary Fig. 21a,b). On day 43, a booster MPS vaccine was delivered, and this led to more immediate LN expansion, reaching peak volumes within 4days, compared with day 7 with the initial vaccine. Seven days after the booster vaccine, peripheral blood was collected and compared to mice that had received only prime vaccination at the same timepoint as the boost in the prime-boost group. No differences in the IFN+ proportion of CD8+ T cells after OVA peptide restimulation were detected; however, the IFN+ proportion of CD4+ T cells was significantly increased relative to both nave control mice and mice that had received only prime vaccination (Supplementary Fig. 21c,d). The proportion of effector-phenotype (CD44+ CD62L) CD8+ T cells was elevated with the MPS prime and further increased after the booster (Supplementary Fig. 21e). Both IgG1 and IgG2a titres against OVA were increased after the booster dose compared with either the same mice on day 21 (pre-boost) or the prime-only mice at the same timepoint (Supplementary Fig. 21f,g). These results indicate that a booster vaccine may elicit more rapid LN expansion along with a stronger adaptive immune response.

See the rest here:

Durable lymph-node expansion is associated with the efficacy of therapeutic vaccination - Nature.com