Retinoic acid induces homing of protective T and B cells to the gut after subcutaneous immunization in mice

Diarrheal diseases represent a major health burden in developing countries. Parenteral immunization typically does not induce efficient protection against enteropathogens because it does not stimulate migration of immune cells to the gut. Retinoic acid (RA) is critical for gut immunity, inducing upregulation of gut-homing receptors on activated T cells. In this study, we have demonstrated that RA can redirect immune responses elicited by s.c. vaccination of mice from skin-draining inguinal LNs (ingLNs) to the gut. When present during priming, RA induced robust upregulation of gut-homing receptors in ingLNs, imprinting gut-homing capacity on T cells. Concurrently, RA triggered the generation of gut-tropic IgA⁺ plasma cells in ingLNs and raised the levels of antigen-specific IgA in the intestinal lumen and blood. RA applied s.c. in vivo induced autonomous RA production in ingLN DCs, further driving efficient induction of gut-homing molecules on effector cells. Importantly, RA-supplemented s.c. immunization elicited a potent immune response in the small intestine that protected mice from cholera toxin–induced diarrhea and diminished bacterial loads in Peyer patches after oral infection with Salmonella. Thus, the use of RA as a gut-homing navigator represents a powerful tool to induce protective immunity in the intestine after s.c. immunization, offering what we believe to be a novel approach for vaccination against enteropathogens.

Diarrheal diseases are among the major causes of death in developing countries, particularly in young children. Although many parenterally applied vaccines are highly efficient in preventing systemic infectious diseases, this route of vaccination frequently results in rather weak protection against enteropathogens. This is mainly caused by the fact that both effector T cells and antibody-producing plasma cells induced by antigen (applied s.c., intramuscularly, or intravenously) fail to express gut-homing molecules; consequently, a protective shield in the gut mucosa is not installed properly. On the other hand, it is well documented that oral application of appropriate vaccines allows for the induction of protective mucosal immunity in the intestine. Indeed, several studies have shown that such a vaccination strategy generates far more potent mucosal protection than does systemic immunization (1–4). Unfortunately, in developing countries, several oral vaccines have performed rather poorly compared with their effect in industrialized states. Factors responsible for inadequate protection conferred by oral vaccines in developing countries include poor sanitation, digestive and absorptive malfunction, intestinal flora overgrowth, and vitamin A deficiency of the vaccines (5). Moreover, oral vaccination can result in immune tolerance when low-virulent or dead material is used as antigen. Together, these findings underline the urgent need for vaccines that generate effective and long-lasting immunity in the digestive tract.

Effective pathogen clearance requires rapid migration of effector T cells to the site of infection. The homing capacity of effector T cells to distinct nonlymphoid tissues depends on the lymphoid organ in which lymphocyte activation occurs (reviewed in ref. 6). This represents a key feature of the locally restricted immune sub-systems. Migration of antigen-experienced T cells into the small intestine requires expression of α4β7-integrin and the chemokine receptor CCR9. The CCR9/CCL25 interaction also guides plasma cells as well as plasmacytoid DCs into the small intestine (7, 8). In contrast, effector T cells expressing E-selectin ligand (ESL), P-selectin ligand (PSL), CCR4, and very late antigen–4 (VLA-4) preferentially home into the inflamed skin.

Several studies indicated that DCs play a decisive role in the imprinting of tissue tropism of effector T cells. In vitro stimulation by gut-associated lymphoid tissue–derived (GALT-derived) DCs is sufficient to induce expression of α4β7-integrin and CCR9 on T cells (9, 10), whereas DCs isolated from skin-draining LNs confer expression of CCR4, PSL, and ESL (11, 12). LN transplantation studies demonstrated that, aside from GALT-derived DCs, mesenteric LN (mLN) stromal cells are essential for the in vivo generation of gut-homing T cells (13, 14). A key component inducing the expression of α4β7-integrin and CCR9 is all-trans retinoic acid (RA) (15). RA is generated from dietary vitamin A involving retinal dehydrogenases (RALDHs). The ability of GALT-derived DCs and mLN stromal cells to induce gut-tropic T cells critically rests on their selective expression of RALDH (13–15). In addition to imprinting activated T cells, GALT-derived RA confers gut-tropism to B cells and synergizes with IL-6 and IL-5 to induce IgA secretion (16). IgA antibodies represent the predominant Ig class in mucosal secretions and provide specific protection of mucosal surfaces by blocking attachment of pathogens.

Given that the vitamin A metabolite RA confers gut-tropism to effector lymphocytes and acts as cofactor for IgA class switch, we hypothesized that mucosal protection can be induced by s.c. antigen application in the presence of RA. Indeed, we found that such an immunization approach resulted in the induction of gut-homing molecules and allowed for robust homing of effector T cells and plasma cells to the small intestine. Moreover, s.c. immunization combined with RA treatment established a pronounced IgA response in serum as well as intestinal washes. As a result, RA supplementation after s.c. immunization conferred mucosal protection against cholera toxin–induced (CT-induced) diarrhea and reduced bacterial burden in Peyer patches after oral Salmonella infection. Mechanistically, we found that RA applied s.c. induced increased RALDH activity in ingLN DCs in vivo. These DCs subsequently triggered expression of gut-homing molecules on activated T and B cells. Together, these results suggested that RA-supplemented s.c. vaccination efficiently induces intestinal protection by redirecting immune responses initiated in skin-draining LNs toward the small intestine. We believe these findings offer new alternatives for the development of vaccination strategies against enteric pathogens.

s.c. RA imprints gut-homing specificity on T cells primed in skin-draining LNs. RA has previously been shown to efficiently instruct activated lymphocytes to upregulate the gut-homing molecules CCR9 and α4β7-integrin in vitro (15). We therefore sought to determine whether local abundance of RA, provided by s.c. delivery, allows for induction of gut-homing molecules on activated T cells in the skin-draining inguinal LN (ingLN). To test whether RA provokes conversion of the homing receptor pattern commonly induced in skin-draining LNs, TCR transgenic OT-I or DO11.10 cells (see Methods) were fluorescently labeled with CFSE and adoptively transferred into C57BL/6 or BALB/c wild-type recipients, respectively; 2 hours later, mice received a single s.c. injection of 50 μg ovalbumin (Ova) into the flanks or were fed 50 mg Ova by gavage. We used Ova of greater than 90% purity, which is known to contain immunologically relevant amounts of endotoxins that served as adjuvant in this experimental setup. A group of s.c. immunized mice additionally received s.c. injections of 150 μg all-trans RA on days 0, 1, and 2 after antigen delivery. On day 3, cells were isolated from gut-draining mLNs as well as skin-draining ingLNs and analyzed by flow cytometry. After standard s.c. immunization, CD8⁺ OT-I T cells proliferating in the ingLN acquired a skin-homing phenotype, characterized by upregulation of PSL and ESL and no apparent expression of α4β7-integrin or CCR9 (Figure 1, A and B). In contrast, T cells primed in ingLNs of RA-treated mice readily upregulated α4β7-integrin and CCR9, albeit to a lesser extent than did T cells activated in the mLNs after oral application of Ova. Notably, CD8⁺ OT-I cells of RA-treated mice no longer showed strong and characteristic induction of the skin-homing molecules PSL and ESL, but were comparable to OT-I cells primed in mLNs with regard to the expression of these molecules (Figure 1, A and B). Similarly, CD4⁺ DO11.10 T cells upregulated α4β7-integrin and CCR9 after s.c. antigen plus RA application, even though expression levels were lower than in DO11.10 cells activated in the mLNs after oral antigen delivery (Figure 1, C and D). Similar results were obtained with CD4⁺ OT-II cells (data not shown). Given the role of RA in inducing Tregs, we sought to determine whether addition of RA affects Treg generation. s.c. RA treatment did not enhance the induction of FoxP3⁺ Tregs (Figure 1, E and F), in contrast to a previous report showing generation of FoxP3⁺ T cells in the skin-draining LN after s.c. RA application (17). Of note, the considerably higher amounts of RA used in the prior study might account for the observed differences.

Figure 1

s.c. Ova immunization with RA as additive leads to upregulation of gut-homing molecules on antigen-specific T cells in skin-draining ingLNs. CFSE-labeled OT-I or DO11.10 cells were adoptively transferred into C57BL/6 or BALB/c mice on day 0. 2 hours later, a single dose of Ova was applied orally or injected s.c. A group of s.c. immunized mice additionally received s.c. injections of 150 μg all-trans RA on days 0, 1, and 2. (A, C, and E) Representative flow cytometry plots obtained on day 3 for α4β7-integrin, CCR9, Foxp3, ESL, and PSL expression on OT-I (DAPI^–Ly5.1⁺Vβ5⁺CD8β⁺; A) or DO11.10 (DAPI^–CD4⁺DO11.10⁺; C and E) T cells activated in the mLN after oral antigen application or in the ingLN after s.c. injection of antigen. Numbers indicate percent positive OT-I or DO11.10 cells. (B, D, and F) Percent α4β7-integrin⁺, CCR9⁺, Foxp3⁺, ESL⁺, and PSL⁺ cells among transferred OT-I T cells (B) or DO11.10 T cells (D and F) from the indicated organs. Data were derived from 6–12 mice per group analyzed in 3 independent experiments (OT-I) or 6–10 mice per group analyzed in 2 independent experiments (DO11.10). *P < 0.05; ***P < 0.001.

In contrast to all-trans RA, the isomeric 13-cis RA did not affect the regulation of CCR9 and α4β7-integrin and induced only a mild reduction of ESL and PSL expression (Supplemental Figure 1; supplemental material available online with this article; doi: 10.1172/JCI44262DS1). To determine whether the effect of RA is restricted to the draining LN, we analyzed the pattern of induced homing receptors in popliteal LNs of mice receiving s.c. Ova into the footpad and s.c. RA into the neck region. Of interest, the popliteal LN supported induction of gut-homing receptors and concomitant suppression of skin-homing receptors, albeit to a lesser extent compared with draining axial LNs of mice that received both Ova and RA s.c. into the neck region (Supplemental Figure 2). Thus, the effect was strongest in — but not restricted to — the LN draining the site of RA injection.

Together, these observations show that all-trans RA delivered exogenously via the skin allowed for efficient induction of gut-homing molecules on responding T cells in ingLNs. Our findings are consistent with previous in vitro studies showing that RA-mediated induction of gut-homing molecules goes along with reduced expression of skin-homing molecules (12), which would be the default pathway for T cells activated in skin-draining LNs.

We next addressed whether RA-induced upregulation of gut-homing molecules in vivo actually allows for efficient homing of activated CD8⁺ and CD4⁺ T cells to the small intestine. We compared the frequencies of adoptively transferred OT-I or DO11.10 cells in the small intestine of RA-treated and nontreated wild-type recipients on day 5 after s.c. immunization with Ova. As expected, after standard s.c. immunization, very few CD8⁺ or CD4⁺ T cells migrated to the small intestine, as determined by the frequency of Ly5.1⁺Vβ5⁺ cells within the CD8αβ⁺ population and the by frequency of DO11.10 cells (as determined by the clonotypic mAb KJ1-26) within the CD4⁺ population (Figure 2). Compared with mice immunized according to the standard s.c. protocol, both s.c. RA-treated and Ova-gavaged animals showed increased frequencies of OT-I cells recruited to the intestinal epithelial lymphocyte (IEL) and lamina propria lymphocyte (LPL) compartments of the small intestine (Figure 2). It should be noted that only the s.c. Ova plus RA route reached statistical significance in recruiting OT-I cells to the IEL compartment; Ova gavage did not cause a statistically significant increase of OT-I cells in the IEL or LPL compartment (Figure 2B). Given that the intestinal epithelium rather than the lamina propria is the major site where CD8⁺ T cells reside, our findings demonstrated that s.c. application of RA together with antigen is most effective in recruiting antigen-specific CD8⁺ T cells to the intestinal epithelium. Along the same lines, s.c. application of antigen together with RA was efficient in recruiting CD4⁺ T cells to the LPL as well as the IEL compartment, whereas oral application by gavage had no effect on recruiting these cells to either compartment (Figure 2, C and D). Collectively, these data demonstrate that the s.c. application of RA allows not only the induction of gut-homing molecules on activated CD4⁺ and CD8⁺ T cells, but also their efficient homing to the small intestine.

Figure 2

Exogenous RA imprints gut-homing specificity on T cells primed in skin-draining ingLN. CFSE-labeled OT-I or DO11.10 cells were adoptively transferred into C57BL/6 or Balb/c mice on day 0. 2 hours later, a single dose of Ova was applied orally or injected s.c. A group of s.c. immunized mice received additional s.c. RA injections on days 0, 1, and 2. T cell homing to the small intestine was determined on day 5. (A and C) Representative flow cytometry plots for OT-I (A) and DO11.10 (C) T cell migration to the small intestinal IEL and LPL compartments. (B) Percent Ly5.1⁺Vβ5⁺ cells within the gated CD8αβ⁺ population. (D) Percent DO11.10-TCR⁺ cells within the gated CD4⁺ population. Bars represent mean values; symbols represent individual mice pooled from 3 (OT-I) or 2 (DO11.10) independent experiments. *P < 0.05; **P < 0.01.

Skin-draining LNs are the major site for the generation of gut-tropic effector T cells after s.c. antigen and RA application. After s.c. RA injection, gut-homing molecules are induced in the skin-draining LNs. However, it remains possible that a large fraction of the gut-tropic effector cells did not acquire gut-homing capacity in the skin-draining LNs but instead disseminated after priming into distant LNs (18), such as the mLN, to acquire gut-tropism there. To assess whether ingLN imprinting in RA-treated mice is sufficient to generate substantial amounts of gut-tropic effector cells, we compared the generation of gut-homing T cells in mice whose mLN was surgically removed with mice that underwent sham surgery. At 5 days after s.c. RA-supplemented Ova immunization, mLN-explanted OT-I recipients showed the same frequency of antigen-specific T cells in the small intestine as did sham-manipulated OT-I recipients (Figure 3). These data show that the mLN does not contribute to the generation of gut-tropic T cells after s.c. immunization and RA treatment. Instead, local priming and imprinting in skin-draining LN is responsible and sufficient for the induction of gut-homing lymphocytes upon s.c. administration of RA.

Figure 3

mLN does not contribute to the generation of gut-tropic T cells after s.c. immunization and RA treatment. mLN-explanted (mLN X) and mLN-sufficient mice were adoptively transferred with CFSE-labeled OT-I cells. Mice were s.c. immunized with Ova on day 0 and received additional s.c. RA injections on days 0, 1, and 2. T cell migration to the small intestine was determined on day 5. (A) Representative flow cytometry plots for OT-I T cell migration to the small intestinal IEL and LPL compartments. (B) Percent Ly5.1⁺Vβ5⁺ cells within the gated CD8αβ⁺ population. Bars represent mean values; symbols represent individual mice pooled from 2 independent experiments. *P < 0.05; **P < 0.01.

s.c. RA application induces upregulation of gut-homing molecules on plasma cells and class switch to IgA in skin-draining ingLNs. RA not only confers gut-tropism to T and B cells, but also acts as a cofactor to promote IgA secretion (16). The finding that s.c. RA injection modifies the T cell response in skin-draining LN prompted us to investigate whether s.c. RA treatment similarly modifies the outcome of the humoral immune response. Mice were immunized on days 0 and 10 either s.c. with 1 μg CT or orally with 10 μg CT. A group of s.c. immunized mice additionally received 150 μg RA s.c. on days 0, 1, 2, 3, 6, 10, and 13 after antigen delivery. On day 14, the levels of CT-specific Igs in serum and intestinal wash were determined by ELISA. Although it provoked a strong IgG and IgM response in the serum, s.c. immunization with CT failed to elicit detectable anti-CT IgA in serum or intestinal wash (Figure 4A). In contrast, RA-treated mice showed increased levels of antigen-specific IgA levels in serum and intestinal wash, albeit to a lesser extent than orally immunized mice. Notably, antigen-specific IgM and IgG titers of RA-treated mice were as high as those of mice receiving only CT s.c. (Figure 4A). To extend these studies, we assessed the abundance of anti-CT IgA–secreting plasma cells in the small intestinal lamina propria by ELISPOT. Few anti-CT IgA-secreting plasma cells localized to the lamina propria of s.c. immunized mice (Figure 4B). Consistent with the increased antigen-specific IgA titers in the intestinal washes, RA-treated mice also showed increased numbers of anti-CT IgA-secreting plasma cells in the small intestine, and their numbers were comparable to those isolated from orally immunized mice. These observations demonstrated that RA-assisted s.c. immunization results in pronounced localization of antigen-specific plasma cells in the small intestine and a robust IgA response in serum and intestinal wash.

Figure 4

Skin-draining ingLNs of RA-treated mice support induction of gut-homing receptor and IgA expression on B cells. Mice were immunized on days 0 and 10 either s.c. with 1 μg CT or orally with 10 μg CT. A group of s.c. immunized mice additionally received s.c. RA injections on days 0, 1, 2, 3, 6, 10, and 13 after antigen delivery. On day 14, mice were sacrificed. (A) Levels of CT-specific Ig in serum and intestinal wash were determined by ELISA. Results are pooled from 4 independent experiments with at least 9 mice total per group. (B) Number of anti-CT IgA-secreting cells in the small intestinal lamina propria (LP), assessed by ELISPOT. Bars denote mean values; symbols denote individual mice pooled from 4 independent experiments. (C–H) Sections from ingLNs of s.c. immunized mice with (D and E) or without (C and F) RA treatment were either analyzed at day 14 by fluorescence microscopy for IgA and Lyve-1 (C and D) or stained for IgA, CCR9, and IgD (E and F). Sections from ingLNs (G) and mLNs (H) from RA-treated s.c. immunized mice were stained for IgA and CT-binding cells. Images are representative for 6 mice analyzed per group in 2 independent experiments. Scale bars: 200 μm (C and D); 50 μm (E and F); 25 μm (G and H). *P < 0.05; **P < 0.01; ***P < 0.001.

After oral challenge, induction of mucosal IgA secretion largely occurs in the organized GALT, such as mLN or Peyer patches. We therefore sought to determine the origin of IgA⁺ plasma cells giving rise to the antigen-specific IgA detected in s.c. immunized RA-treated mice. Immunohistology of ingLN of RA-treated mice revealed the abundant presence of IgA⁺ cells, which were absent in ingLN of s.c. immunized mice without RA treatment (Figure 4, C and D). IgA-expressing cells localized near Lyve-1⁺ lymphatic endothelium, indicative of their imminent egress from the LN. Interestingly, a major fraction of the IgA⁺ plasmablasts also expressed CCR9 in RA-treated mice, but not in nontreated mice (Figure 4, E and F), which explains their increased frequency in the intestinal lamina propria. Importantly, CT-binding plasma cells were detected in ingLN, but not mLN, of RA-treated mice (Figure 4, G and H), although IgA⁺ plasma cells were frequently present in mLN. Since IgA⁺ plasma cells were also found in mLN of nonimmunized mice (data not shown), these cells probably form a constitutively present population. Together, these results identify the skin-draining ingLN of RA-treated mice as a major site for the induction of gut-homing receptors and IgA expression of activated plasmablasts after s.c. RA injection.

RA-supplemented s.c. immunization protects mice from CT-induced diarrhea and diminishes bacterial loads in Peyer patches after oral Salmonella infection. To address whether the intestinal IgA antibodies induced by s.c. application of RA mediate protective immunity, mice were subjected to CT-induced diarrhea (19, 20). When applied orally, CT induces elevated levels of cAMP in intestinal epithelial cells, resulting in secretion of water and chloride ions into the small intestine. A previous study demonstrated that CT-specific secretory IgA in the intestinal lumen was responsible for protective immunity against CT-induced diarrhea (20). Therefore, groups of mice were immunized with CT s.c., orally, or s.c. plus RA treatment, or were left untreated. On day 14, mice were challenged orally by applying 10 μg CT by gavage. At 2 hours after CT challenge, mice were sacrificed, and the weight of the small intestine together with its fluid content was determined. The weight provides information about the degree of fluid influx and gives a rough estimate of antitoxic protection. Standard s.c. immunization did not establish prominent protection, and no significant difference was found in the fluid influx compared with that in nonimmunized animals (Figure 5A). In contrast, the concomitant s.c. RA treatment generated a protective immunity which was comparable in degree to that induced by oral CT immunization. Importantly, the protection provided by s.c. application of the antigen plus RA treatment prevented any detectable fluid influx, since there was no significant increase in weight compared with nonimmunized, nonchallenged mice (Figure 5A). In addition to antitoxic protection, immunity against viable mucosal pathogens is also of high importance; therefore, we assessed protection against Salmonella infection. Mice were s.c. immunized with dead Salmonella Typhimurium SL1344 lacking the aroA gene on days 0 and 2. A group of s.c. immunized mice additionally received s.c. injections of RA on days 0, 1, 2, 3, 4, 7, 8, and 9. On day 14, mice were orally infected with wild-type Salmonella Typhimurium SL1344, and 2 days after infection, bacterial loads were determined in single Peyer patches. RA-treated mice showed significantly less bacteria in Peyer patches than did mice lacking RA treatment (Figure 5B). This difference was more prominent when 7 × 10⁷ bacteria were inoculated compared with high-dose infection with 4 × 10⁹ bacteria. Collectively, these results demonstrated that s.c. application of antigen in RA-treated mice allowed for induction of protective immunity against CT-induced diarrhea and reduced bacterial burden in Peyer patches after oral Salmonella infection.

Figure 5

RA-supplemented s.c. immunization confers mucosal protection against CT-induced diarrhea and diminishes bacterial loads in Peyer patches after oral Salmonella infection. (A) Mice were treated with CT as in Figure 4 or were left untreated (no immunization). At day 14, mice were challenged orally by gavage with 10 μg CT. 2 hours after challenge, mice were sacrificed, and the weight of the small intestine together with its fluid content was determined. Small intestines of mice receiving neither immunization nor challenge served as controls. Bars represent mean values; symbols represent individual mice pooled from 3 independent experiments. Note the discontinuous y axis. (B) Mice were s.c. immunized with 10⁶ PFA-fixed Salmonella Typhimurium SL1344 lacking the aroA gene (SL1344 ′ˆ†aroA) on days 0 and 2. A group of s.c. immunized mice was additionally s.c. injected with RA on days 0, 1, 2, 3, 4, 7, 8, and 9. On day 14, mice were orally inoculated with 7 × 10⁷ or 4 × 10⁹ wild-type Salmonella Typhimurium SL1344. 2 days after infection, bacterial loads were determined in single Peyer patches. Bars represent median values; symbols represent individual Peyer patches from 3 (7 × 10⁷) or 4 (4 × 10⁹) mice per group. Numerals within the graph indicate the number of sterile Peyer patches. *P < 0.05; **P < 0.01.

RA treatment facilitates induction of gut-tropism on T and B cells by increasing RALDH activity of skin-draining LN DCs. As described above, s.c. delivery of exogenous RA supported the generation of gut-homing effector T cells and plasma cells in the ingLN. RA has been reported to directly act on activated lymphocytes (15), but there is ample evidence that RA can also affect DC function (21–23). In particular, RA has been shown to induce RALDH expression of bone marrow–derived DCs (24–27). Inversely, mLN DCs of mice raised with a vitamin A–deficient diet show markedly reduced RALDH activity (24–27). Therefore, we assumed that s.c. applied RA might not only act directly on activated lymphocytes in the ingLN, but also exert imprinting activity on resident DC to efficiently synthesize RA. Cells isolated from the ingLN of untreated or s.c. RA-treated mice were incubated with the fluorescent RALDH substrate ALDEFLUOR, and RALDH activity was assessed by flow cytometry. Consistent with a recent study (28), we found that skin-draining LNs of untreated mice contained a small population of RA-producing DCs (Figure 6, A and B). However, ingLN DCs of RA-treated mice showed significantly higher fractions of RA-producing cells than did ingLN DCs of untreated mice, although the proportion of ALDEFLUOR⁺ ingLN DCs derived from RA-treated animals did not reach that of DCs isolated from the mLN. Under all conditions tested, RALDH activity was blocked by the RALDH inhibitor diethylaminobenzaldehyde (DEAB) (Figure 6, A and B), demonstrating the specificity of the assay.

Figure 6

Exogenous RA promotes in vivo the ability of ingLN DCs to induce gut-tropism by increasing their RALDH activity. (A and B) Mice were s.c. immunized with 50 μg Ova; 1 group additionally received RA s.c. at days 0, 1, and 2. At day 3, cells were isolated from mLNs and ingLNs and analyzed for RALDH activity. (A) Flow cytometry plots of gated MHCII⁺CD11c⁺ cells with or without DEAB. (B) Percent ALDEFLUOR⁺ DCs within gated MHCII⁺CD11c⁺ cells. (C and D) CFSE-labeled OT-I cells were cocultured with Ova-loaded ingLN DCs, mLN DCs from untreated mice, or ingLN DCs from RA-treated mice (RA ingLN DC) with or without DEAB. (C) Flow cytometry plots for α4β7-integrin, CCR9, PSL, and ESL expression on OT-I cells on day 5 of coculture. (D) Percent α4β7-integrin⁺, CCR9⁺, ESL⁺, and PSL⁺ cells among OT-I T cells. (E) Percent α4β7-integrin⁺ cells among B cells. CFSE-labeled splenic B cells were activated in the presence of ingLN DCs, mLN DCs, or ingLN DCs from RA-treated mice. (F–H) Langerin⁺ cells were found among CD11c⁺MHCII^hi cells (F, top gate) of skin-draining LNs, whereas MHCII^int DCs (bottom gate) contained few Langerin⁺ cells; Langerin⁺MHCII^hi DC frequency was reduced by RA treatment (G and H). (B, D, E, and H) Bars represent mean values, symbols denote individual mice. Pooled from 2 (E and H) or 3 (D) independent experiments, or (B) similar results were obtained in 2 additional independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001.

Finally, we assessed whether these differences in RALDH activity correlate with the ability of RA-imprinted ingLN DCs to induce gut-homing molecules on responding T cells. CFSE-labeled OT-I cells were cocultured with antigen-loaded ingLN DCs isolated from untreated or RA-treated mice, and their expression of CCR9 and α4β7-integrin was analyzed by flow cytometry. ingLN DCs from RA-treated mice induced significantly more CCR9 and α4β7-integrin on responding T cells than did ingLN DCs from untreated mice (Figure 6, C and D), but were less potent than mLN DCs in upregulating CCR9 (56.4% ± 27.0%) or α4β7-integrin (36.7% ± 28.0%). Upregulation of CCR9 and α4β7-integrin could be blocked by addition of DEAB. These observations indicate that in vitro induction of gut-homing receptors relies on active RA production by ingLN DCs from RA-treated mice, rather than on passive RA carry-over. Apart from T cells, B cells activated in the presence of ingLN DCs from RA-treated mice gained expression of α4β7-integrin (Figure 6E). Interestingly, ingLN DCs of RA-treated mice lost the ability to strongly induce PSL and ESL on activated T cells (Figure 6, C and D). Notably, upregulation of ESL and PSL was also suppressed when RALDH activity was blocked by DEAB. In consideration of a potential link between RA abundance and Langerin-expressing DC generation (29), we next determined Langerin expression by ingLN DCs. Intriguingly, s.c. injection of RA decreased the frequency of Langerin⁺ DCs within the MHCII^hiCD11c⁺ ingLN population (Figure 6, F–H). Together, these results showed that RA modifies DC phenotype and function in terms of RALDH and Langerin expression.

Given the fact that diarrheal diseases are among the major causes of morbidity and mortality among young children in developing countries, there is an urgent need for vaccination regimens that generate effective and long-lasting intestinal immunity. In view of the poor performance of several oral vaccines in developing regions compared with industrialized regions (5), we aimed to develop an immunization strategy based on the targeted homing of immune effector cells to the intestine. Here, we report that mucosal protection was induced by s.c. applied antigens in the presence of RA. RA delivered via s.c. injection induced the generation of gut-homing T cells and gut-tropic IgA-expressing plasma cells in skin-draining LNs. Mechanistically, we showed that s.c. injected RA imprinted ingLN DCs with the ability to generate RA endogenously, which resulted in the capacity to induce gut-homing receptors on activated lymphocytes.

It is conceivable that, in contrast to the rather diluted RA arriving via lymph, the RA produced on-site can be delivered in a highly concentrated fashion to those cells directly contacting the DCs. Thus, exogenously applied RA might act not only directly on lymphocytes, but also indirectly via DC-based enhancement of RA signaling. In vivo, this may endow the DC with the capacity to potently manipulate homing, which permits an immune response to be redirected to the intestine even though it originated in the “wrong” compartment, the skin-draining LNs. It is currently unclear whether exogenously applied RA also has some imprinting capacity on stromal cells residing in skin-draining LNs. Since we previously failed to induce gut-homing molecules on T cells in skin-draining LNs transplanted into the mesenterium (13), these results suggest that stromal cells might be less reactive to RA than DCs are, but a role for RA in controlling RALDH expression in stromal cells has also been recently reported (26). Our findings regarding DC imprinting are in line with reports from others suggesting that RA regulates its own synthesis by a positive feedback loop through RA receptor signaling and retinal production (30–32). Furthermore, numerous recent studies provide direct evidence that RA induces RALDH activity in DCs (24–27). In addition to promoting RALDH activity, s.c. applied RA decreased the frequency of Langerin⁺ DCs in skin-draining LNs. Similarly, Chang and colleagues suggest a link between availability of RA and expression of Langerin by showing that vitamin A–deficient mice display increased numbers of Langerin⁺ DCs in GALT (29). It remains to be determined whether the reduction of Langerin⁺ ingLN DCs is responsible for induction of a mucosal immune response after RA-supplemented s.c. immunization. Interestingly, RA has recently been studied as an additive for transcutaneous immunization with inactivated influenza virus (33). This immunization regimen induced a pronounced IgA response in the lung and conferred protection against intranasal influenza virus challenge. The authors attributed this effect primarily to enhanced mobilization of DCs to the skin-draining LNs (33). However, enhanced, RA-driven targeting of lymphocytes to mucosal tissues including the lung might also be a relevant mechanism leading to improved protection. Furthermore, it has been reported that exogenously applied RA leads to increased germinal centers and augmented antibody production (34), which indicates that RA has effects on the immune response other than changing the profile of homing molecules. It also should be mentioned that RA treatment might cause unwanted side-effects, such as allergic responses (35) and weight loss (36).

There is now ample evidence that the induction of skin-homing and gut-homing potential of effector cells of the adaptive immune system is intimately linked to the site of lymphocyte activation. Recent work suggests that DCs and LN stromal cells cooperate in shaping a LN environment that favors induction of the respective homing receptors. Here we report that exogenously applied RA overrides the LN-specific factors driving induction of skin-homing molecules in skin-draining LNs. In doing so, RA redirects the induced plasma cell and T cell response to the intestine. Conversion of the induced type of tissue tropism was triggered by s.c. injections of 150 μg RA, while 20 μg did not show any effect (data not shown). Thus, it appears that distinct threshold levels of RA are required to efficiently affect the microenvironment of the ingLN, investigated in the present study. Importantly, mLN-explanted mice directed T cells to the small intestine as efficiently as did mLN-proficient mice. Furthermore, CT-specific plasma cells were only detectable in ingLN, not mLN, of RA-treated mice. Thus, the mLN does not seem to contribute to the generation of gut-tropic T and B cells after s.c. immunization and RA treatment. Instead, local priming and imprinting in skin-draining LNs is responsible and sufficient for induction of gut-homing lymphocytes upon s.c. administration of RA. In contrast, the mLN was reported to be the site of IgA class switching after topical application of an antigen plus adjuvant to the intact skin surface (37). As additional examples of redirecting immune responses, 2 studies reported that parenteral, adenovirus-based immunizations enhance expansion of gut-homing T cells (38, 39). In mice, the pronounced CD8⁺ T cell response in the gut was attributed to adenovirus-driven induction of RALDH activity in DC (38).

We demonstrated in this study that RA-assisted s.c. immunization generated intestinal immunity and conferred the benefits of mucosal protection otherwise generated only when antigen was applied orally. s.c. RA imprinted gut-homing specificity on T cells and plasma cells activated in skin-draining LNs. Furthermore, RA treatment initiated class switch to IgA in local skin-draining LNs and generated a pronounced IgA response in serum and intestinal lumen. Consequently, RA-assisted s.c. immunization protected mice from CT-induced diarrhea and reduced bacterial loads in Peyer patches after oral Salmonella infection. Notably, RA-supplemented s.c. immunization had the advantage of simultaneously inducing both mucosal and systemic humoral immune responses. As a mode of action of exogenously applied RA in redirecting immune responses, we found that RA treatment imprinted ingLN DCs in vivo with the ability to generate gut-tropic B and T cells by increasing their RA-producing capacities.

Thus, we propose that RA or its precursors, such as vitamin A, might be used as gut-homing navigators for vaccinations given s.c. or intramuscularly in order to generate protective immunity against enteric pathogens. Since vitamin A is already routinely delivered in vaccine clinics to young children in countries in which vitamin A deficiency is a problem, RA-assisted immunization might offer a feasible approach to efficiently combat enteric infections.

View Supplemental data

We thank Siegfried Weiss for providing FGK45 and G. Bernhardt, I. Prinz, and A. Krueger for valuable suggestions on the manuscript. This work was supported by Deutsche Forschungsgemeinschaft grant SFB621-A1 to R. Förster.

Address correspondence to: Reinhold Förster, Institute of Immunology, Hannover Medical School, Carl-Neuberg Strasse 1, 30625 Hannover, Germany. Phone: 49.511.5329733; Fax: 49.511.5329722; E-mail: Foerster.Reinhold@MH-Hannover.de.

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J Clin Invest. 2011;121(8):3051–3061. doi:10.1172/JCI44262.

1. Belyakov IM, Isakov D, Zhu Q, Dzutsev A, Berzofsky JA. A novel functional CTL avidity/activity compartmentalization to the site of mucosal immunization contributes to protection of macaques against simian/human immunodeficiency viral depletion of mucosal CD4+ T cells. J Immunol. 2007;178(11):7211–7221.
   View this article via: PubMed Google Scholar
2. Belyakov IM, Ahlers JD, Nabel GJ, Moss B, Berzofsky JA. Generation of functionally active HIV-1 specific CD8+ CTL in intestinal mucosa following mucosal, systemic or mixed prime-boost immunization. Virology. 2008;381(1):106–115.
   View this article via: PubMed CrossRef Google Scholar
3. Gallichan WS, Rosenthal KL. Long-lived cytotoxic T lymphocyte memory in mucosal tissues after mucosal but not systemic immunization. J Exp Med. 1996;184(5):1879–1890.
   View this article via: PubMed CrossRef Google Scholar
4. Santosuosso M, Zhang X, McCormick S, Wang J, Hitt M, Xing Z. Mechanisms of mucosal and parenteral tuberculosis vaccinations: adenoviral-based mucosal immunization preferentially elicits sustained accumulation of immune protective CD4 and CD8 T cells within the airway lumen. J Immunol. 2005;174(12):7986–7994.
   View this article via: PubMed Google Scholar
5. Czerkinsky C, Holmgren J. Enteric vaccines for the developing world: a challenge for mucosal immunology. Mucosal Immunol. 2009;2(4):284–287.
   View this article via: PubMed CrossRef Google Scholar
6. Agace WW. Tissue-tropic effector T cells: generation and targeting opportunities. Nat Rev Immunol. 2006;6(9):682–692.
   View this article via: PubMed CrossRef Google Scholar
7. Pabst O, et al. Chemokine receptor CCR9 contributes to the localization of plasma cells to the small intestine. J Exp Med. 2004;199(3):411–416.
   View this article via: PubMed CrossRef Google Scholar
8. Wendland M, et al. CCR9 is a homing receptor for plasmacytoid dendritic cells to the small intestine. Proc Natl Acad Sci U S A. 2007;104(15):6347–6352.
   View this article via: PubMed CrossRef Google Scholar
9. Mora JR, et al. Selective imprinting of gut-homing T cells by Peyer’s patch dendritic cells. Nature. 2003;424(6944):88–93.
   View this article via: PubMed CrossRef Google Scholar
10. Johansson-Lindbom B, Svensson M, Wurbel MA, Malissen B, Marquez G, Agace W. Selective generation of gut tropic T cells in gut-associated lymphoid tissue (GALT): requirement for GALT dendritic cells and adjuvant. J Exp Med. 2003;198(6):963–969.
    View this article via: PubMed CrossRef Google Scholar
11. Dudda JC, Simon JC, Martin S. Dendritic cell immunization route determines CD8+ T cell trafficking to inflamed skin: role for tissue microenvironment and dendritic cells in establishment of T cell-homing subsets. J Immunol. 2004;172(2):857–863.
    View this article via: PubMed Google Scholar
12. Mora JR, Cheng G, Picarella D, Briskin M, Buchanan N, von Andrian UH. Reciprocal and dynamic control of CD8 T cell homing by dendritic cells from skin- and gut-associated lymphoid tissues. J Exp Med. 2005;201(2):303–316.
    View this article via: PubMed CrossRef Google Scholar
13. Hammerschmidt SI, et al. Stromal mesenteric lymph node cells are essential for the generation of gut-homing T cells in vivo. J Exp Med. 2008;205(11):2483–2490.
    View this article via: PubMed CrossRef Google Scholar
14. Molenaar R, et al. Lymph node stromal cells support dendritic cell-induced gut-homing of T cells. J Immunol. 2009;183(10):6395–6402.
    View this article via: PubMed CrossRef Google Scholar
15. Iwata M, Hirakiyama A, Eshima Y, Kagechika H, Kato C, Song SY. Retinoic acid imprints gut-homing specificity on T cells. Immunity. 2004;21(4):527–538.
    View this article via: PubMed CrossRef Google Scholar
16. Mora JR, et al. Generation of gut-homing IgA-secreting B cells by intestinal dendritic cells. Science. 2006;314(5802):1157–1160.
    View this article via: PubMed CrossRef Google Scholar
17. Kang SG, Lim HW, Andrisani OM, Broxmeyer HE, Kim CH. Vitamin A metabolites induce gut-homing FoxP3+ regulatory T cells. J Immunol. 2007;179(6):3724–3733.
    View this article via: PubMed Google Scholar
18. Liu L, Fuhlbrigge RC, Karibian K, Tian T, Kupper TS. Dynamic programming of CD8+ T cell trafficking after live viral immunization. Immunity. 2006;25(3):511–520.
    View this article via: PubMed CrossRef Google Scholar
19. Nochi T, et al. Rice-based mucosal vaccine as a global strategy for cold-chain- and needle-free vaccination. Proc Natl Acad Sci U S A. 2007;104(26):10986–10991.
    View this article via: PubMed CrossRef Google Scholar
20. Tokuhara D, et al. Secretory IgA-mediated protection against V. cholerae and heat-labile enterotoxin-producing enterotoxigenic Escherichia coli by rice-based vaccine. Proc Natl Acad Sci U S A. 2010;107(19):8794–8799.
    View this article via: PubMed CrossRef Google Scholar
21. Geissmann F, et al. Retinoids regulate survival and antigen presentation by immature dendritic cells. J Exp Med. 2003;198(4):623–634.
    View this article via: PubMed CrossRef Google Scholar
22. Wada Y, Hisamatsu T, Kamada N, Okamoto S, Hibi T. Retinoic acid contributes to the induction of IL-12-hypoproducing dendritic cells. Inflamm Bowel Dis. 2009;15(10):1548–1556.
    View this article via: PubMed CrossRef Google Scholar
23. Manicassamy S, et al. Toll-like receptor 2-dependent induction of vitamin A-metabolizing enzymes in dendritic cells promotes T regulatory responses and inhibits autoimmunity. Nat Med. 2009;15(4):401–409.
    View this article via: PubMed CrossRef Google Scholar
24. Yokota A, et al. GM-CSF and IL-4 synergistically trigger dendritic cells to acquire retinoic acid-producing capacity. Int Immunol. 2009;21(4):361–377.
    View this article via: PubMed CrossRef Google Scholar
25. Feng T, Cong Y, Qin H, Benveniste EN, Elson CO. Generation of mucosal dendritic cells from bone marrow reveals a critical role of retinoic acid. J Immunol. 2010;185(10):5915–5925.
    View this article via: PubMed CrossRef Google Scholar
26. Molenaar R, et al. Expression of retinaldehyde dehydrogenase enzymes in mucosal dendritic cells and gut-draining lymph node stromal cells is controlled by dietary vitamin A. J Immunol. 2011;186(4):1934–1942.
    View this article via: PubMed CrossRef Google Scholar
27. Jaensson-Gyllenback E, et al. Bile retinoids imprint intestinal CD103(+) dendritic cells with the ability to generate gut-tropic T cells [published online ahead of print February 2, 2011]. Mucosal Immunol . doi:10.1038/mi.2010.91 .
28. Guilliams M, et al. Skin-draining lymph nodes contain dermis-derived CD103(–) dendritic cells that constitutively produce retinoic acid and induce Foxp3(+) regulatory T cells. Blood. 2010;115(10):1958–1968.
    View this article via: PubMed CrossRef Google Scholar
29. Chang SY, et al. Lack of retinoic acid leads to increased langerin–expressing dendritic cells in gut–associated lymphoid tissues. Gastroenterology. 2010;138(4):1468–1478.
    View this article via: PubMed CrossRef Google Scholar
30. de The H, Vivanco-Ruiz MM, Tiollais P, Stunnenberg H, Dejean A. Identification of a retinoic acid responsive element in the retinoic acid receptor beta gene. Nature. 1990;343(6254):177–180.
    View this article via: PubMed CrossRef Google Scholar
31. Leroy P, Nakshatri H, Chambon P. Mouse retinoic acid receptor alpha 2 isoform is transcribed from a promoter that contains a retinoic acid response element. Proc Natl Acad Sci U S A. 1991;88(22):10138–10142.
    View this article via: PubMed Google Scholar
32. Duester G, Shean ML, McBride MS, Stewart MJ. Retinoic acid response element in the human alcohol dehydrogenase gene ADH3: implications for regulation of retinoic acid synthesis. Mol Cell Biol. 1991;11(3):1638–1646.
    View this article via: PubMed Google Scholar
33. Martin Mdel P, Seth S, Koutsonanos DG, Jacob J, Compans RW, Skountzou I. Adjuvanted influenza vaccine administered intradermally elicits robust long-term immune responses that confer protection from lethal challenge. PLoS One. 2010;5(5):e10897.
    View this article via: PubMed CrossRef Google Scholar
34. Ma Y, Ross AC. Toll-like receptor 3 ligand and retinoic acid enhance germinal center formation and increase the tetanus toxoid vaccine response. Clin Vaccine Immunol. 2009;16(10):1476–1484.
    View this article via: PubMed CrossRef Google Scholar
35. Matheu V, et al. Impact on allergic immune response after treatment with vitamin A. Nutr Metab (Lond). 2009;6:44.
36. Amengual J, Ribot J, Bonet ML, Palou A. Retinoic acid treatment increases lipid oxidation capacity in skeletal muscle of mice. Obesity (Silver Spring). 2008;16(3):585–591.
    View this article via: CrossRef Google Scholar
37. Chang SY, et al. Cutting edge: Langerin+ dendritic cells in the mesenteric lymph node set the stage for skin and gut immune system cross-talk. J Immunol. 2008;180(7):4361–4365.
    View this article via: PubMed Google Scholar
38. Ganguly S, Manicassamy S, Blackwell J, Pulendran B, Amara RR. Adenovirus type 5 induces vitamin A-metabolizing enzymes in dendritic cells and enhances priming of gut-homing CD8 T cells [published online ahead of print February 2, 2011]. Mucosal Immunol . doi:10.1038/mi.2011.1.
39. Benlahrech A, et al. Adenovirus vector vaccination induces expansion of memory CD4 T cells with a mucosal homing phenotype that are readily susceptible to HIV-1. Proc Natl Acad Sci U S A. 2009;106(47):19940–19945.
    View this article via: PubMed Google Scholar
40. Voedisch S, et al. Mesenteric lymph nodes confine dendritic cell-mediated dissemination of Salmonella enterica serovar Typhimurium and limit systemic disease in mice. Infect Immun. 2009;77(8):3170–3180.
    View this article via: PubMed Google Scholar
41. Rolink A, Melchers F, Andersson J. The SCID but not the RAG-2 gene product is required for S mu-S epsilon heavy chain class switching. Immunity. 1996;5(4):319–330.
    View this article via: PubMed CrossRef Google Scholar