Skin Tolerance is Supported by the Spleen


Correspondence to: M. Buettner, Funktionelle und Angewandte Anatomie, OE 4120, Medizinische Hochschule Hannover, Carl-Neuberg-Str.1, 30625 Hannover, Germany. E-Mail:


The repeated application of antigens results in the induction of tolerance. Lymph nodes are responsible for this reaction by producing suppressor cells. Using an in vivo transplantation model, we showed recently that stromal cells from different lymph nodes induce different cell populations for suppression, which all produce a tolerogenic phenotype. In this study, we were interested in the role of the spleen in these tolerance reactions. Therefore, tolerance was induced via feeding or injecting ovalbumin several times in control and splenectomized mice. The delayed-type hypersensitivity (DTH) was measured as well as the cell subset composition of the spleen. The spleen of peripherally tolerized mice showed higher proliferation activity and a specific antibody production compared with orally tolerized mice, where regulatory T cells were predominantly found. Tolerance induction after removal of the spleen resulted in a reduced DTH response in antigen fed animals, whereas skin tolerance induction failed. In conclusion, the results illustrate that lymph nodes from different areas employ their individual pathways for similar immune reactions, and the spleen is part of this reaction initiated at the peripheral site.


Induction of tolerance to antigens (Ag) is one of the major functions of the immune system. A decision has to be made whether the Ag are harmless, such as food, or a potentially harmful pathogen. For intestinal tolerance, oral administration of Ag is the classical way to induce Ag-specific systemic immune response suppression. Over the years, different pathways for tolerance formation have been identified. On the one hand, there is deletion or anergy of cells, and on the other hand, the induction of regulatory T cells (Treg) [1-4], both resulting in a suppression of a specific antibody (Ab) production and T cell proliferation [5, 6].

It has been shown that mucosal dendritic cells (DC) migrate via afferent lymphatics into the draining mesenteric lymph node (mLN) [7, 8]. Within the mLN, tolerogenic DC induce Treg via retinoic acid (RA) and IL-10/TGFβ production. Treg proliferate and migrate into the gut where they suppress the induction of effector cells [9, 10]. The importance of the LN was shown in an experiment where the LN was removed and it was attempted to induce tolerance. Without the LN tolerance to the harmless antigen was no longer inducible [11, 12].

Recently, we showed via transplantation of the peripheral lymph node (pLN) into the draining area of the gut that there are differences between pLN and mLN in the way that tolerance is induced [13-15]. The percentage of Treg was diminished within pLN transplants (pLNtx) compared with mLN after oral tolerance induction (OT). However, we observed an induction of B cells which produced Ag-specific immunoglobulin (IgG) in pLNtx. These data were comparable with data of pLN after tolerance induction in the skin tolerance (ST). Furthermore, transfer experiments of these IgG+ B cells isolated from tolerized pLN showed suppression of the delayed-type hypersensitivity (DTH) response [15]. Thus, the induction of tolerance is a LN-specific phenomenon which is not changeable after transplantation and therefore not dependent on the draining area of the LN.

During these experiments, we found indications that the spleen might be involved in tolerance induction. In numerous studies, the spleen, as an organ of systemic response, was shown to be involved in the suppression of T cell proliferation and Ab production after oral tolerance induction [6, 16]. Splenic DC were found to be producers of suppressor cytokines such as TGFβ and later on, suppressor T cells were identified [17, 18].

Therefore, we evaluated the effect of the spleen during intestinal and skin tolerance with regard to DTH response, cell subset composition including the induction of Treg, as well as immunoglobulin production. We found that the spleen is a crucial organ for the induction of skin tolerance, whereas oral tolerance is independent of the spleen. These results indicate that the mechanism of tolerance induction differs between the periphery and the intestinal route after Ag recognition.

Materials and methods


Female C57BL/6 mice were bred at the central animal laboratory of Hannover Medical School and were used at a weight of 18–25 g. All animal experiments were performed in accordance with the institutional guidelines and had been approved by the Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit (No. 33.9-42502-04-09/1667).

Intestinal surgery

Under the combined anaesthesia of Ketamine (Gräub AG, Bern, Switzerland) and Domitor (Pfizer, Karlsruhe, Germany), the abdomen was opened on the lateral side and the spleen was exposed. The arteries were ligated using absorbable fibre (Catgut GmbH, Markneukirchen, Germany), and the spleen was excised. After 4 weeks recovery, tolerance was induced (n = 7–8).

Application of fluorescent antigens and cholera toxin

Mice were either fed or injected subcutaneously with 1012 fluorescent latex beads (200 nm diameter; Polyscience, Warrington, PA, USA) or injected subcutaneously in the forefoot with 50 μg ovalbumin conjugated with Alexa Fluor 555 (Invitrogen, Karlsruhe, Germany), both in combination with 10 μg cholera toxin (Sigma-Aldrich, St. Louis, MO, USA). Four hours later, the mice were sacrificed and the LN and the spleen were analysed (n = 4–6).

Induction and measurement of DTH responses

Oral tolerance (OT): mice were fed by gavage with 25 mg OVA (Grade III; Sigma-Aldrich) in 200 μl PBS or PBS only as a control on day 0, 3, 6 and 8. On day 16, mice were immunized by subcutaneous injection of 300 μg OVA (Grade VI; Sigma-Aldrich) in 200 μl PBS emulsified in complete Freud's adjuvant (CFA; Sigma-Aldrich). On day 34, mice were challenged by subcutaneous injection of 50 μg OVA (Grade VI) in 10 μl PBS into the right ear and PBS only into the left ear. Ear swelling was measured before challenge and 48 h later. DTH response was calculated as described previously [11].

Based on the protocol for OT induction, tolerance in the periphery (skin tolerance (ST)) by the skin-draining LN (axillary and brachial LN, pLN) was induced as follows: 4.2 mg OVA (Grade III; Sigma-Aldrich) in 10 μl PBS or PBS only as a control on day 0, 3, 6, and 8 by subcutaneous injection into the forepaw. On day 16, mice were immunized by subcutaneous injection of 300 μg OVA (Grade VI; Sigma-Aldrich) in 200 μl PBS/CFA emulsion. On day 34, mice were challenged by subcutaneous injection of 50 μg OVA (Grade VI) in 10 μl PBS into the right ear and PBS only into the left ear. Ear swelling was measured before challenge and 48 h later, and the DTH response was calculated.


Cryostat sections of spleen from tolerized mice were fixed in acetone/methanol solution. Immunofluorescence histochemistry was performed according to standard protocols. Briefly, sections were rehydrated in TBST (0.1 m Tris pH 7.5, 0.15 m NaCl, 0.1% Tween-20) and preincubated with TBST containing 10% swine serum (Dako, Hamburg, Germany). The sections were stained with biotinylated anti- CD11c (DC, BD Biosciences, Heidelberg, Germany), which was visualized by streptavidine coupled with Cy3 (Invitrogen), fluorescent dye-coupled antibodies against CD3-FITC (eBiosciences, San Diego, CA, USA), with antibodies against B220 (B cells, BD Biosciences), which was visualized by Cy5 (Invitrogen), biotinylated anti-B220 (BD Biosciences) was visualized by streptavidine coupled with FITC (Invitrogen) and fluorescent dye-coupled GL-7-PE (BD Biosciences). Proliferation was visualized using anti-Ki67 (Epitomice, Burlingame, CA, USA), which was detected by a goat anti-rabbit antibody coupled with Cy3 (Dianova, Hamburg, Germany). Sections were mounted with fluorescent mounting medium (Dako). Images were acquired using an Axiovert 200M microscope with Axiovision software (Carl Zeiss, Jena, Germany).

Antibodies for flow cytometry

Cell suspensions from the spleen, mLN or pLN of C57BL/6 mice were made, and 1 × 106 cells were incubated with anti-CD3-FITC, anti-CD19-APC-H7 (BD Biosciences), anti-CD4-APC, anti-CD25-PE, anti-IgG-PE (all acquired from Serotec, Oxford, GB) and anti-FoxP3-FITC (eBiosciences). The FoxP3 staining was performed as following: After labelling the cells with anti-CD4 and anti-CD25 on the surface, cells were incubated for 20 min at room temperature (RT) in the dark with Fix/Perm buffer (eBioscience). Afterwards, cells were washed and incubated for 15 min with Perm buffer (eBiosciences). Without washing, FoxP3 or the isotype antibody was added and incubated for 30 min. All FACS analyses were performed on a FACSCanto (BD Biosciences). Isotype-matched mAb served as controls. Representative FACS illustrations for all stainings are shown in Fig. S1.


The concentration of OVA-specific IgG in the serum was analysed in the ELISA. Therefore, the plates were coated with 0.5 μg/ml OVA (Grade VI; Sigma-Aldrich) in PBS overnight at 4 °C. After washing, the plates were blocked and samples were added to a concentration of 1:100 to 1:1000 and incubated 120 min at 37 °C. After washing, detection antibodies (biotinylated anti-IgG1, anti-IgG2a, anti-IgG2b and anti-IgG3; BD Biosciences) were added and later detected with horseradish peroxidase (HRP, BD Biosciences), tetramethylbenzidene (TMB, BD Biosciences) and hydrogen peroxide (1:1) as substrate. The reaction was stopped with 2NH2SO4 (Merck, Darmstadt, Germany). The optical density was analysed in an ELISA-reader (Bio-TEK Instruments GmbH, Bad Friedrichshall, Germany).

Data analysis

Calculations, statistical analysis and graphs were performed with Graphpad Prism 4.0 (Graphpad Software Inc., San Diego, CA, USA).


Orally and peripherally applied antigens reach the spleen

It is generally accepted that Ag are transported into the draining LN. In the LN the Ag are presented by DC, macrophages and B cells. To test whether Ag also reach the spleen after Ag administration, fluorescent beads or ovalbumin coupled with PE were injected subcutaneously. After 4 h, the pLN and the spleen were removed and examined for cells carrying fluorescent beads or OVA-PE (Fig. 1A-C). We found fluorescent beads as well as OVA-PE positive cells in the pLN and also in the spleen of these mice (Fig. 1A-C).

Figure 1.

Orally applied or subcutaneously injected Ag are found in the spleen. Fluorophore-coupled beads or OVA were injected or gavaged into control animals (n = 3–6). Four hours later, the peripheral lymph node (pLN), mesenteric lymph node (mLN) and the spleen of these mice were removed and one part was analysed using histology and the second part was analysed using flow cytometry. (A) Representative images of CD11c positive cells in the spleen after Ag administration are shown. CD11c+ cells are found to have incorporated the fluorescent Ag (shown by arrows). (B, C) Using flow cytometry, similar percentages of fluorescent Ag were found in the pLN and the spleen independent of the Ag which was injected in the forefoot of mice. Three independent experiments were performed. Means and standard error are given (n = 3–6). (D, E) Fluorescent beads were injected or gavaged, and fluorescent bead positive cells were detected via flow cytometry. Independent of the region in which the Ag was administered, fluorescent beads were found in the spleen. Three independent experiments were performed. Means and standard error are given (n = 3–6).

Furthermore, we gavaged fluorescent beads to analyse the relevance of different administration routes in Ag transport to the spleen. Comparing the two application routes, we found no difference in the presence of Ag (Fig. 1D, E). Thus, the spleen recognizes all Ag entering the body and could be part of an immune reaction.

Proliferation of B cells is high in the spleen after skin tolerance induction

Reaching the spleen, the Ag are presented by APC and, as known from the LN, an immune response to the foreign Ag will be induced [19]. To evaluate the involvement of the spleen in tolerance induction, it was investigated whether there was an increase in proliferating cells in the spleen after induction of oral or skin tolerance. OT was induced by feeding the mice OVA several times, while ST was induced by subcutaneous injection of OVA into the forefoot of the mice. After tolerance induction, the spleen was analysed. As shown in Fig. 2, there was an increase in proliferating cells, especially B cells which form germinal centres in the spleen of mice with ST, whereas in mice with OT only few proliferating cells and no germinal centres were observed. In addition, GL-7, an antibody against Neu5Ac on germinal centre B cells [20], was also detectable in the spleen of mice with ST. In the spleen of mice with OT, only marginal GL-7 expression was visible (Fig. S2). Thus, the spleen seems to be involved in the induction of ST, but OT is spleen independent.

Figure 2.

Proliferation of B cells in the spleen after induction of skin tolerance. Cryostat sections of spleen from peripherally or orally tolerized mice were made, and immunofluorescent staining for CD3 (green; A and E), B220 (B cells in blue; B and F) and Ki67, used to identify proliferating cells (red; C and G), was performed. Proliferation was visualized to show secondary follicles with developing germinal centres. The merging of (A–C) and (E–G) shows that in the spleen of peripherally tolerized mice, germinal centres are seen, as compared to orally tolerized mice.

The spleen maintains the LN induced tolerance

We further examined the cell pattern of the spleen after tolerance induction in more detail using flow cytometry. Higher percentages of IgG+ B cells were found after ST induction in the spleen (Fig. 3A). In contrast, in the spleen of mice where OT was induced, Treg were found to be increased (Fig. 3B). Furthermore, OVA-specific antibodies were measured in the serum of both tolerance induction routes. OVA-specific IgG2a and IgG 3 were increased in animals with ST compared with mice in which OT was induced (Fig. 3C). Thus, the spleen seems to enforce different immune reactions induced in peripheral and mesenteric sites by additional B cell proliferation and by Treg induction, respectively.

Figure 3.

Increased percentages of IgG+ cells but reduced Treg are found in the spleen after skin tolerance induction. (A, B) The spleen of orally or peripherally tolerized mice was removed, and flow cytometry was performed. Cell suspensions were made, and cells were stained for IgG or CD4, CD25 and Foxp3 to identify Treg. Means and SEM are given (n = 3 mice per group), and significant differences in the unpaired t-test are indicated by * P < 0.05; **P = 0.01. (C) OVA-specific Ig in the serum of orally and peripherally tolerized mice was analysed using Ag-specific ELISA, and the OD was measured (performed in duplicate). Mesenteric lymph node (mLN)-Oral tolerance (OT) values were divided by peripheral lymph node (pLN)-skin tolerance (ST) values, and percentages were calculated. Means and standard error are given (n = 6–8 mice), and significant differences in the unpaired t-test are indicated by *P < 0.05; *** P < 0.001.

After splenectomy oral tolerance but not skin tolerance is induced

The next question was whether tolerance could be induced in splenectomized animals. Four weeks after splenectomy, OT or ST was induced and the DTH reaction was measured. As shown in Fig. 4, repeated OVA administration resulted in a reduced DTH response in OT mice as well as ST mice. However, in splenectomized mice, ST was no longer inducible, whereas OT could be induced without the presence of the spleen. Thus, the spleen is needed for induction of skin tolerance.

Figure 4.

Skin tolerance is not inducible in splenectomized animals. Tolerance was induced in control and splenectomized animals by feeding four times OVA or PBS as a control, and skin tolerance was induced by subcutaneous injection of OVA or PBS. Eight days later, all mice were immunized by subcutaneous injection of OVA in PBS/CFA emulsion and challenged on day 34 by injection of OVA into the right ear and PBS into the left ear. Ear swelling was measured before challenging and 48 h later. Delayed-type hypersensitivity (DTH) responses were calculated (mLN-Oral tolerance (OT) control n = 5–7; pLN-skin tolerance (ST) control n = 5–6; mLN-OT splenectomized n = 8–9; pLN-ST splenectomized n = 7–8). Significant differences in the unpaired t-test are indicated by *P < 0.05; **P < 0.01; ***P < 0.001.

LN react spleen independently

Analysing the mLN or pLN after tolerance induction with or without splenectomy, differences between the LN of different draining areas were observed. Lower percentages of CD3+ and CD4+ T cells and higher percentages of B cells were identified within the pLN (Fig. 5). A more detailed look at the B cell population in control and also in splenectomized mice after ST and OT induction also showed a higher percentage of IgG+ B cells in pLN compared with mLN (control mice: mLN 36.9% ± 0.1, pLN 42.2% ±  2.5 (n = 2); splenectomized animals: mLN 39.8 ± 1.5, pLN 52.2% ± 2.2 (n = 5)). In addition, OVA-specific antibodies were measured in the serum of control and splenectomized animals after ST induction. Decreased IgG1 and IgG2a levels were detected in splenectomized mice compared with control mice (data not shown). However, Tregs were found to be reduced in pLN compared with mLN. Overall, in splenectomized animals, the cell subset composition showed no differences compared with spleen-bearing animals (Fig. 5). Thus, the draining LN started the LN-specific induction of tolerance.

Figure 5.

Mesenteric lymph node (mLN) or peripheral lymph node (pLN) react normally if the spleen is missing. After oral or skin tolerance induction, mLN and pLN of control and splenectomized animals were removed and flow cytometry was performed. Cell suspensions were made, and cells were stained for CD3, CD4 and CD19. Treg were identified via CD4, CD25 and Foxp3 expression. Four independent experiments were performed. Means and standard error are given (n = 8–10 mice per group), and significant differences in the unpaired t-test are indicated by * P < 0.05; ** P < 0.01; *** P < 0.001.


We demonstrate here that tolerance is inducible independent of the regions of Ag application, but that there are differences in the type of induction and performance. In this study, we focused on the role of the spleen in different tolerance strategies.

The spleen as a central organ has contact with Ag entering the body independent of the area. Therefore, it is possible that Ag end up in the blood directly after administration or reaching the LN via the afferent lymphatic. They leave the LN via the efferent lymphatics, which bring the Ag to the thoracic duct. The Ag end up in the blood system and consequently in the spleen [21, 22]. We applied fluorescent-labelled beads or fluorescent-labelled OVA by injection or gavage, and we found positive cells in the draining LN and also in the spleen of these animals. Previous studies showed that Ag were taken up by APC and transported into the mLN on the oral route [23] or into the skin-draining LN after skin painting [24]. Another study showed that after intranasal administration, FITC positive cells were not only found in the draining LN but also in the spleen [12]. Thus, the spleen is exposed to Ag entering the body.

Tolerance is characterized as a systemic unresponsiveness to repeatedly applied and/or large amounts of Ag. Mucosal tolerance is a major reaction induced in the draining LN, which is destroyed after surgical removal of the LN [11, 12]. For oral tolerance, mucosal CD103+ DC have to present the Ag in the LN, where Treg induction has to take place [25]. These Treg migrated into the lamina propria in a CCR9 and α4β7 integrin dependent manner where further proliferation happened. After depletion of Treg, oral tolerance was no longer inducible [26]. In this study, Treg are also found in high percentages in the mLN after tolerance induction. As shown previously, we and others found a tolerogenic microenvironment in the mLN which favours the induction of Treg [15, 25, 27]. In contrast to orally applied Ag, peripheral uptake of Ag drives a different induction of tolerance. We showed recently that Ag-specific B cells were induced in the pLN and that these cells were responsible for skin tolerance [15]. However, tolerance is a systemic reaction, and therefore, we analysed how the spleen is involved.

We determined the proliferation intensity in the spleen and identified increased germinal centre development of mice with skin tolerance compared with oral tolerance. This is in line with previous findings where orally tolerized splenic T cells showed no increase in proliferation after Ag administration [5, 28]. Furthermore, OVA-specific antibodies were found to be absent or greatly diminished in mice which were fed OVA several times [5, 6]. These data were confirmed in our own study for oral tolerance, whereas OVA-specific antibodies were found in animals which induced skin tolerance. These different ways of inducing tolerance seem to be dependent on the LN. After transplantation of a peripheral LN into the mesentery of mice, Ag-specific antibodies were also induced, whereas Treg were not found to be increased [15]. In these pLNtx transplanted mice, we showed that similar pathways were activated. These findings support the idea that the microenvironment of LN regulates the direction of immune reactions.

However, in this study, we were able to show that the spleen is involved in the process of skin tolerance, whereas oral tolerance seems to be spleen independent. Splenectomized orally tolerized mice were able to suppress the immune reaction after Ag administration. These findings indicate that the mLN is the major player in the induction of oral tolerance, while for the induction of skin tolerance, the spleen is needed. In contrast to our own findings, immune suppression in splenectomized animals was detected after Ag injection subcutaneously [18, 29]. One reason for these differences may be the different protocols that were used or the different Ag. But further studies in which tolerance was induced in splenectomized mice detected a reduced number of suppressor cells that support our data. It was concluded, on the one hand that the spleen was involved in the induction of suppressor cells and on the other hand that splenic cells suppressed the proliferation of T cells [6, 16, 18, 29]. Thus, the spleen played an important role in skin tolerance, whereas it is not necessary in oral tolerance induction.

Altogether, our data corroborate the hypothesis that LN from different draining areas showed variable characteristics during tolerance induction. These differences result not only in the induction of different immune cells but also in the involvement of further organs such as the spleen. The spleen, as a systemic organ, seems to receive information not only about all immune reactions in the body, but also whether support is needed.


We wish to thank Sheila Fryk for correction of the English. This work was supported by the Deutsche Forschungsgemeinschaft (SFB621/A10).

Melanie Bornemann performed the experiments. Ulrike Bode designed the research study, analysed the data and wrote the paper. Manuela Buettner designed the research study, performed the experiments, analysed the data and wrote the paper.