BCG infection dose guides dendritic cell migration and T cell priming in the draining lymph node

In contrast to delayed‐type hypersensitivity (DTH) and other hallmark reactions of cell‐mediated immunity that correlate with vaccine‐mediated protection against Mycobacterium tuberculosis, the contribution of vaccine dose on responses that emerge early after infection in the skin with Bacille Calmette–Guérin (BCG) is not well understood. We used a mouse model of BCG skin infection to study the effect of BCG dose on the relocation of skin Dendritic cells (DCs) to draining lymph node (DLN). Mycobacterium antigen 85B‐specific CD4+ P25 T cell‐receptor transgenic (P25 TCRTg) cells were used to probe priming to BCG in DLN. DC migration and T cell priming were studied across BCG inocula that varied up to 100‐fold (104 to 106 Colony‐forming units—CFUs). In line with earlier results in guinea pigs, DTH reaction in our model correlated with BCG dose. Importantly, priming of P25 TCRTg cells in DLN also escalated in a dose‐dependent manner, peaking at day 6 after infection. Similar dose‐escalation effects were seen for DC migration from infected skin and the accompanying transport of BCG to the DLN. BCG‐triggered upregulation of co‐stimulatory molecules on migratory DCs was restricted to the first 24 hour after infection and was independent of BCG dose over a 10‐fold range (105 to 106 CFUs). The dose seemed to be a determinant of the number of total skin DCs that move to the DLN. In summary, our results support the use of higher BCG doses to detect robust DC migration and T cell priming.


| INTRODUCTION
Dendritic cells (DCs) are specialized in capturing, processing, and presenting microbes to naïve T cells for priming.The transport of microbial antigen from the periphery to the draining lymph node (DLN) is a bottleneck in this multistep process which is preceded by the relocation of antigen-laden, migratory DCs from the site of infection in the periphery to DLN. 1 T cell priming remains incompletely understood during infection or vaccination with mycobacteria.In this regard, we have previously shown that when the live tuberculosis (TB) vaccine Mycobacterium bovis Bacille Calmette-Guérin (BCG) is injected in the skin, it mobilizes migratory skin DCs that transport BCG bacilli to DLN for In contrast to delayed-type hypersensitivity (DTH) and other hallmark reactions of cell-mediated immunity that correlate with vaccine-mediated protection against Mycobacterium tuberculosis, the contribution of vaccine dose on responses that emerge early after infection in the skin with Bacille Calmette-Guérin (BCG) is not well understood.We used a mouse model of BCG skin infection to study the effect of BCG dose on the relocation of skin Dendritic cells (DCs) to draining lymph node (DLN).Mycobacterium antigen 85B-specific CD4 + P25 T cell-receptor transgenic (P25 TCRTg) cells were used to probe priming to BCG in DLN.DC migration and T cell priming were studied across BCG inocula that varied up to 100-fold (10 4 to 10 6 Colony-forming units-CFUs).In line with earlier results in guinea pigs, DTH reaction in our model correlated with BCG dose.Importantly, priming of P25 TCRTg cells in DLN also escalated in a dose-dependent manner, peaking at day 6 after infection.Similar dose-escalation effects were seen for DC migration from infected skin and the accompanying transport of BCG to the DLN.BCG-triggered upregulation of co-stimulatory molecules on migratory DCs was restricted to the first 24 hour after infection and was independent of BCG dose over a 10-fold range (10 5 to 10 6 CFUs).The dose seemed to be a determinant of the number of total skin DCs that move to the DLN.In summary, our results support the use of higher BCG doses to detect robust DC migration and T cell priming.
priming of antigen-specific CD4 + T cells. 2 Although a few BCG bacilli can directly drain to LN in our model, this phenomenon does not contribute to T cell priming. 2 With its mix of secreted mycobacterial products, microbial debris and live, infectious bacilli, BCG is immunologically both an interesting and complex antigen preparation that continues to draw attention because of its wide clinical usage and proven safety.BCG was generated by serial passage of virulent M. bovis and has been used against TB since 1921. 3It remains an integral part of many national vaccination programs.The inoculation route has changed over the years.Today, BCG is given in the skin, mostly intradermally but also percutaneously. 4In addition, BCG has been used for more than 40 years as an immunotherapeutic against bladder cancer. 5,6Similar to other mycobacteria, BCG infection causes inflammation and leads to a T helper-cell type 1 (Th1) response.Unlike B cells and antibodies, CD4 + T cells and interferon (IFN)-γ are paramount in protection against mycobacteria largely by activating macrophages to control intracellular M. tuberculosis growth. 7CD4 + T cells and IFN-γ are also involved in BCG-mediated protection against bladder cancer, but the mechanisms at play are not well known compared for TB vaccination.][10] Several explanations have been proposed, but the variation remains incompletely understood.A large body of data supports the efficacy of BCG against disseminated forms of childhood TB, especially TB meningitis, 11 but this discrepancy too, has not been explained.
Inoculation dose is central to vaccination protocols.Not surprisingly, the BCG workhorse has received much consideration on this topic during its 100-year use.Seminal, large-scale studies on guinea pigs performed under the umbrella of the WHO in 1955 showed that BCG dose positively correlates with the size of the vaccine scar, the reaction to tuberculin, and most importantly, protection against M. tuberculosis challenge. 12Similar results were observed in mice vaccinated with the attenuated strain of M. tuberculosis H37Ra. 13The latter study also revealed a threshold after which an escalation in dose did not further increment survival. 13In mouse models, BCG dose can easily range from 5 × 10 5 to 10 7 colony-forming units (CFUs) independent of the route of administration.][16][17] Prior work on the implications of BCG dose on immune responses has mostly focused on protection against M. tuberculosis challenge, while the more immediate, immunological reactions unleashed by escalating BCG dose in vivo are largely uncharacterized.In the current study, we used a mouse model of BCG skin infection to investigate the early events that unfold in the LN draining the BCG injection site in the skin.Specifically, we explored the effect of BCG dose on skin DC migration, the transport of BCG itself, and the priming of antigen-specific CD4 + T cells in DLN.

| Mice
C57BL/6NRj mice purchased from Janvier Labs (Janvier Labs, Le Genest-Saint-Isle, France).P25 TCRTg RAG-1 −/− mice expressing EGFP 16 were originally provided by Dr. Ronald Germain, NIAID, NIH.Animals were kept at the Comparative Medicine-Biomedicum (KM-B) facility, Karolinska Institutet.Handling and housing were performed according to the directives and guidelines of the Swedish Board of Agriculture, the Swedish Animal Protection Agency, and Karolinska Institutet.All experiments were approved by the Stockholm Ethics Council.Female mice between 8 and 12 weeks old were used in experiments.

| Inoculation of mice
Mice were inoculated in the hind footpad with 30 μL of BCG at indicated doses.Control animals received 30 μL of PBS.In DC migration experiments, 18 24 hours prior to sacrifice, animals were injected with 20 μL of 0.5 mM 5-and 6-carboxyfluorescein diacetate succinimidyl ester (CFSE) (Sigma) in the same footpad.To assess delayedtype hypersensitivity (DTH), mice were injected with BCG in the hind footpad as above, and 2 weeks later challenged with 10 μg of culture filtrate proteins (CFPs) from M. tuberculosis (BEI Resources, USA) in 30 μL in the contralateral, hind footpad.Footpad thickness was measured using a digital calliper before and 24, 48, and 72 hours after CFP administration.Swelling was determined as the change in footpad thickness after CFP administration.In T cell priming experiments, 10 5 LN cells from naïve P25 TCRTg RAG-1 −/− EGFP mice were injected in the tail vein (200 μL) into C57BL/6 recipients.Recipients were infected 24 hours later with different doses of BCG in the footpad as described above.

| Generation of single-cell suspensions
Single-cell suspensions from popliteal LNs (pLNs) were obtained and prepared for flow cytometry as previously described. 2Briefly, pLNs were aseptically removed, transferred into tubes containing FACS buffer (5 mM EDTA and 2% FBS in PBS), homogenized using a tissue grinder, and cells counted by Trypan Blue exclusion.For CFU analysis, pLNs were put into tubes with PBS containing 0.05% Tween-80 (Sigma), homogenized, and plated onto 7H11 agar.

| Statistical analyses
Graphical representations of data and statistical testing were performed on GraphPad Prism 9.5.1 (GraphPad Software Inc., USA).The significance of differences in data group means was analysed by one-way ANOVA without assuming equal variances, followed by Welch's test for multiple comparisons.P values < .05were considered statistically significant.In some experiments, outliers were excluded from analysis following Grubbs's test for outliers.

| DTH reaction correlates with BCG dose
Seminal studies in guinea pigs show that BCG vaccinemediated protection against M. tuberculosis increased with BCG dose. 12In these studies, the protection given by different BCG doses also correlated with DTH to tuberculin prior to M. tuberculosis challenge.To investigate whether a similar correlation between dose and degree of DTH was also evident in mice, we inoculated C57BL/6 mice in the footpad with 4 different doses of BCG (10 4 , 10 5 , 5 × 10 5 , and 10 6 CFUs) and measured 2 weeks later the amount of footpad swelling after injection of M. tuberculosis CFP in the contralateral footpad.In line with results in guinea pigs, 12 DTH in mice was associated with the initial inoculation dose of BCG (Figure 1).DTH responses peaked 2 days after CFP challenge for all doses investigated (Figure 1).

| Priming of P25 TCRTg cells is BCG dose-dependent
Since DTH is mediated by previously primed T cells, we investigated the activation and expansion of naïve CD4 + T cells in response to different doses of BCG.To this end, we took advantage of our previous model reliant on adoptive transfer of naïve Mycobacterium antigen 85B-specific P25 TCRTg cells into C57BL/6 mice. 2 Twenty-four hours after transfer of naïve P25 TCRTg cells, cohorts of C57BL/6 recipients were infected with different doses of BCG as before ( 10 6 and 9 days after infection (Figure 2A,B).For all doses investigated, P25 TCRTg cells were activated 1 day after infection as marked by upregulation of CD69, downregulation of CD62 ligand (CD62L), and upregulation of CD44 (Figure 3).Activation was pronounced with midto-higher doses (10 5 , 5 × 10 5 , and 10 6 CFUs) and delayed with 10 4 CFUs.This trend was also observed at later time points (Figure 3).Interestingly, expression of CD44 on P25 TCRTg cells was higher following infection with lower BCG doses, particularly from day 3 (Figure 3).Gating strategy for P25 TCRTg cells is shown as Figure S1.

| BCG transport and DC migration to DLN are dependent on BCG dose
We have previously shown that priming of P25 TCRTg cells in DLN is dependent on transport of BCG to DLN by migratory DCs from the site of infection in the skin. 2 Therefore, we have examined whether BCG transport to DLN and DC mobilization from skin would follow a similar dose-dependent trend as seen with T cell priming.Consequently, we infected mouse footpads with the same 4 doses of BCG (10 4 , 10 5 , 5 × 10 5 , and 10 6 CFUs) and evaluated mycobacterial CFUs in DLN 3 days after infection.To study DC migration, we employed a CFSE-based migration assay established in our laboratory that enables enumeration of DCs that move to the DLN within a 24-hour interval. 2We found that both BCG and DC transition from skin to DLN exhibited a dose-dependent effect similar to T cell priming (Figure 4A,B).Figure S2 shows gating of CFSE + migratory skin DCs.

| Activation of migratory DCs in DLN peaks 1 day after BCG infection and is independent of dose
Since the accumulation of migratory DCs in DLN was affected by BCG infection dose, we examined whether DC activation status was also influenced by dose.We used the same CFSE-based migration assay to gauge the expression of activation markers CD40 and CD86, and of MHC-II itself, on skin DCs arriving in the DLN across the first 3 days of infection.The intensity of both activation markers and MHC-II on migratory skin DCs reaching the DLN was clearly enhanced 1 day after infection compared to PBS-injected controls, for all three doses investigated (Figure 5).Surprisingly, the expression of these activation markers was similar for the investigated doses and subsided independently of dose after day 1 (Figure 5).These observations suggest that BCG prompts an early but transient activation of migratory skin DCs.

| DISCUSSION
Despite its limitations, BCG is still the only available vaccine against TB 100 years after its introduction. 19Thus, understanding BCG function remains relevant to improve the vaccine on a rational basis.In this regard, the early immunological events that unfold in skin and in DLN after BCG infection have been poorly characterized yet stand to innovate BCG and generate basic knowledge on how productive T cell responses to the bacilli are initiated in vivo.Using a mouse model of BCG infection, we show a doseescalating effect of BCG on skin DC migration to DLN and on priming of CD4 + T cells at the same site.BCG-triggered DC activation is transient, peaks within the first 24 hours after injection in the skin, and is independent of dose.In contrast, the number of skin DCs that reach the DLN can be manipulated by increasing the inoculation dose.Given the global shortage of BCG, understanding the contribution of inoculation dose has bearing not only on the generation of productive immune reactions but also on increasing access to the vaccine, if can be used at lower doses.A caveat not limited to the animal models where the effects of dose can be studied is that the definition of a 'low' versus a 'high' dose is subjective.In our BCG infection model, we routinely inject 10 6 CFUs in the footpad skin.This generates a quantifiable BCG-triggered response in DLN.In the past, we have inoculated BCG i.v. at a similar dose but to study chronic responses, including granuloma formation. 16,20The same dose has also been reported in vaccination studies with the novel BCG replacement candidate ΔureChly + rBCG, 21 but this dose is by no means a field standard.Vaccination studies with the MTBVAC candidate used 5 × 10 5 CFUs of BCG as a comparison, 22 while other studies have investigated protective effects of BCG at doses much higher than 10 6 CFUs. 23,24In animal models of bladder cancer, the protective effects of BCG Tokyo were observed at 10 7 CFUs.Similar protective effects were achieved with a lower dose of 10 6 CFUs when given together with GM-CSF and IFN-α. 25otwithstanding differences in patient age and vaccine strain, clinical dose of BCG is brutally higher for bladder cancer than for TB; an equivalent of approximately 1-8 × 10 8 CFUs of BCG TICE is normally instilled in the urethra, 26 while the recommended clinical dose for TB vaccination with BCG SSI 1331 is 1-4 × 10 5 CFUs for neonates and 2-8 × 10 5 CFUs for adults. 4The equivalent of the more commonly administered TB vaccination dose would be approximately 3000 CFUs in mice if based on human infant body weight. 27Interestingly, this dose suffices to protect against M. tuberculosis challenge. 27In our study, the 'low' dose of 10 4 CFUs, which is close to the human neonatal dose, generated muted responses compared to the standard dose of 10 6 CFUs.The same was observed with 10 5 CFUs of BCG.In response to 10 4 CFUs of BCG, skin DCs barely relocated to DLN.In line, detection of BCG and priming of P25 TCRTg cells in DLN were both drastically reduced.This result substantiates the association shown by others and us between T cell priming and the recovery of viable mycobacteria in the DLN. 2,14,28,29ower doses of BCG (40 and 4 × 10 3 CFUs) have been shown to produce a bona fide Th1 response, while a higher dose (4 × 10 7 CFUs) induced Th1 with components of Th2. 30,31That said, no difference in protection against M. tuberculosis challenge was observed using a similar range of doses (2 × 10 3 -2 × 10 7 ). 24Rather, higher doses induce protection against M. tuberculosis faster, while lower doses eventually reach similar protection over time. 24ecause of the dominant role of CD4 + T cells in protection against mycobacteria, we focused on these cells when studying BCG dose effects on adaptive immunity.
In our experiments that tracked transferred, antigen-specific CD4 + T cells, higher doses (5 × 10 5 and 10 6 CFUs) were clearly better compared to lower doses (10 4 and 10 5 CFUs) in expanding P25 TCRTg cells and triggering their activation based on surface expression of activation markers CD69, CD62L, and CD44.Upregulation of CD69 and downregulation of CD62L were delayed when lower doses were used.The lowest dose investigated (10 4 CFUs) did however induce a more robust expression of CD44 on P25 TCRTg cells than the other doses.Although CD44 has been linked to Th1 polarization, 32 the difference observed is not likely relevant for differentiation of P25 TCRTg cells into Th1.The TCR on these Tg cells behaves as high affinity 33 and likely differentiates into Th1 at both low and high Ag doses, as previously demonstrated for another MHC-II-restricted TCR. 34In any case, low doses of BCG were unable to expand P25 TCRTg cells to the equivalent magnitude of higher doses, at least over the time course examined.These data on the early stages of BCG-triggered CD4 + T cell responses provide new mechanistic support for the use of higher doses in protection against M. tuberculosis.That more vaccine should promote more DC migration and T cell priming is not granted in the case of BCG since it retains to some degree the ability exploited by virulent M. tuberculosis to inhibit antigen-presenting cells.
Excessive footpad swelling was recorded in our set-up after injection of 10 7 CFUs, which prevented us from studying the effects of a dose that was one order of magnitude higher than our standard dose of 10 6 CFUs.It is possible that such a dose would stimulate further DC migration and T cell priming in our model, albeit at the expense of adverse inflammation.Regardless of dose and the inflammation that follows, BCG being attenuated is expected to clear over time from immunocompetent mice and humans.Using doses on the higher end of the scale will take longer to clear, allowing for more priming to occur due to persisting bacilli in vivo, in addition to our demonstrated effect of dose on the amplitude of the priming reaction.
4 , 10 5 , 5 × 10 5 , and 10 6 CFUs).Signs of expansion in transferred P25 TCRTg cells were observed already 1 day after BCG infection (Figure 2A,B).The scope of expansion was dose-dependent and particularly clear F I G U R E 1 DTH correlates with BCG inoculation dose.C57BL/6 mice were injected in the footpad with different doses of BCG.Two weeks later, the mice were challenged with CFP in the contralateral footpad and footpad swelling to CFP was measured.Five mice per group were used in experiment.One of 2 independent experiments shown.Bars indicate standard error of the mean (SEM).Asterisk (*) denotes statistical difference between adjacent BCG doses.BCG groups are statistically different from PBS for all doses and time points, with the exception of 10 4 on day 3. F I G U R E 2 T cell priming is dependent on BCG dose.Naïve P25 TCRTg cells were i.v.transferred into C57BL/6 recipients (10 5 cells per mouse), which were 24 hours later infected with different dose of BCG into footpads.One, 3, 6, and 9 days later, DLNs were excised, processed, and analysed by flow cytometry.Total number (A) and percentage (B) of P25 TCRTg cells (EGFP + Vβ11 + CD4 + ) are graphed.(A, B) Left graph shows all time points graphed together, and right part shows data from each time point separately.Four to 6 mice per group.Day 3 graphs show one of two independent experiments; day 1, 6, and 9 experiments have been performed once each.Bars indicate SEM.Asterisk (*) denotes statistical difference between BCG-infected groups.

F I G U R E 3 T
cell activation depends on BCG inoculation dose.Naïve P25 TCRTg cells were transferred into C57BL/6 recipients as in Figure 2 (10 5 cells per mouse).Twenty-four hours later, mice were infected with different dose of BCG into footpads.One, 3, 6, and 9 days later, DLNs were excised, processed, and analysed by flow cytometry.Graphs show mean fluorescent intensity (MFI) for CD69 (top), CD62L (center), and CD44 (bottom) of transferred P25 TCRTg cells (EGFP + Vβ11 + CD4 + ).Four to 6 mice per group.Day 3 graphs show one of two independent experiments; day 1, 6, and 9 experiments have been performed once each.Bars indicate SEM.Asterisk (*) denotes statistical difference between adjacent BCG-infected groups.Octothorpe (#) shows statistical difference of BCG-infected groups over PBS.

F I G U R E 4
BCG transport and BCG-triggered DC migration correlate with the infection dose.C57BL/6 mice were inoculated with different doses of BCG (A, B). (A) BCG CFUs recovered from pLNs of infected animals.Ten mice per group.(B) The CFSE-based migration assay was used to assess skin DC migration to DLN.Total numbers of CFSE + skin DCs (MHC-II high CD11c int cells) in BCG-draining pLN were determined 3 days after infection by flow cytometry.Dashed line represents average of CFSE + migratory skin DCs in PBS controls.The DC migration in all BCG-infected groups is significantly increased over PBS controls (not indicated).Four to six animals per group.Both (A) and (B) show 1 of at least 2 independent experiments for dose 10 5 CFU BCG and higher.Bars indicate SEM.Asterisk (*) denotes statistical difference between indicated groups (A, B).F I G U R E 5 Activation markers on migratory DC in DLN are upregulated 24 hours after BCG infection in the skin.C57BL/6 mice were inoculated in the footpad with different doses of BCG and the CFSE-based migration assay performed at different time points.MFI is shown for CD40, CD86, and MHC-II on CFSE + MHC-II high CD11c int migratory skin DCs.Five mice per group.One out of 2 independent experiments shown.Bars indicate SEM.Asterisk (*) denotes statistical difference between adjacent BCG-infected groups.Octothorpe (#) shows statistical difference of BCG-infected groups over PBS.