Dr M. R. Karlsson, Department of Clinical Immunology, Guldhedsgatan 10A, 413 46 Göteborg, Sweden.
In the present study we have investigated if transfer of serum from rats fed ovalbumin (OVA) leads to specific tolerance and bystander suppression in recipient animals. Rats that received serum from OVA-fed donors had a lower delayed-type hypersensitivity reaction (DTH) both against OVA and the bystander antigen, human serum albumin (HSA), compared with recipients given serum from control-fed animals. The in vitro proliferation of OVA- and HSA-stimulated spleen cells and the serum immunoglobulin G (IgG) antibody levels against OVA and HSA were also lower in the animals that received serum from OVA-fed animals compared with the controls. There was no reduction of the immune response to HSA if the recipient animals, given serum from OVA-fed donors were immunized with OVA and HSA at separate sites. Depletion of CD25-positive cells from spleen suspensions from rats receiving serum from OVA-fed animals, resulted in a significant increase in proliferation of OVA-stimulated cells in vitro compared with the controls. Tolerogenic activity could be demonstrated, both in a fraction from serum containing structures smaller than 100 000 MW and a fraction with components larger than 100 000 MW, compared with size-related serum fractions obtained from control-fed animals. This implies that the tolerogenic activity could be mediated by more than one serum component. The tolerogenic activity was most prominent in animals receiving the larger size fraction with a more pronounced suppression of the DTH reaction and lower levels of IgG anti-OVA antibodies in serum compared with controls. A novel finding in the present study was that the transfer of serum, collected from rats fed OVA, led to a reduction of the immune response to a bystander antigen in the recipients. This suggests that the induced tolerance is at least partly due to suppression. The suppression could have been mediated by CD25-positive cells since removal of these cells resulted in an increased in vitro proliferation against OVA.
The introduction of protein antigens into the gastrointestinal tract of rodents normally results in reduced antibody production and T-cell responses to the protein. 1–3 This phenomenon is called oral tolerance and is today a well-known concept. At least three main mechanisms are contributing to oral tolerance; anergy, clonal deletion and suppression. Which mechanism will predominate probably depends on factors such as the antigen concentration, the nature of the antigen, the immunization route, the age of the animal, etc. 1
Earlier studies in mice have shown that the processing of a protein in the gut, such as ovalbumin (OVA), generates a tolerogenic component in serum that is transferable and can reduce the delayed-type hypersensitivity reaction (DTH) against OVA in recipient mice. 4–6 It was assumed that the intestinal processing in some way converted the OVA into a tolerogenic form that was released into the circulation. The processing in the gut was proven to be necessary since intravenous injection of the protein did not result in suppression of the immune response. 4
The tolerogenic activity is time dependent and serum taken 5 min after feeding cannot transfer tolerance while serum taken 1 hr after feeding can. 6 Partial characterization of the serum factor has been performed 7 but the immuno-modulating mechanisms by which the factor induces tolerance have not yet been described. If a reduction of the immune response to a bystander antigen can be demonstrated in recipient animals it suggests that suppression by regulatory T cells is one mechanism contributing to the tolerance.
Previously it was suggested that oral tolerance was mainly mediated by CD8-positive cells.8,9 Recent studies have shown that CD4-positive cells are also operative and it is now well established that both CD4- and CD8-positive cell populations are of importance for the induction and maintenance of tolerance. 10–14 Efforts have been made to characterize further the phenotype of suppressor cells. Hence, it seems that T cells expressing CD25 are a unique population of suppressor T cells that can prevent the initiation of autoimmune diseases. 15–21 By which mechanisms these cells regulate the immune response is not known.
In this study we wanted to investigate if a tolerogenic serum factor, similar to the one found in mouse serum, is present in rats and also to study the immunological effects of this factor on recipient animals. Thus, rats were injected with serum prepared from animals fed an OVA-containing diet for 2 h. The development of tolerance in the recipients, i.e. reduction of antibody and inflammatory T-cell responses against OVA and a bystander antigen, human serum albumin (HSA), was tested after parenteral immunization. We show tolerogenic activity both in a serum fraction containing components larger and smaller than 100 000 MW. The tolerogenic serum factor has the potential to down-regulate antibody and T-cell responses not only to the original antigen but also to a bystander antigen in recipient rats. Furthermore, the induced tolerance is at least partly mediated by CD25-positive regulatory cells.
Materials and methods
Eight-week-old inbred male Wistar rats (recipients) and 12-week-old outbred male Wistar rats (donors) were obtained from BK Universal (Stockholm, Sweden) and kept in our animal facilities (Department of Clinical Immunology, Göteborg University, Sweden) under standard conditions.
Serum transfer procedure
The 12-week-old rats were starved over night and then fed an OVA-containing diet (AnalyCen, Lidköping, Sweden) for 2 hr. Control animals were fed a standard diet during the same period. The estimated intake of OVA during this time was 400 mg/rat. Two hours after feeding, blood was collected by heart puncture. The obtained sera were pooled in one OVA-fed pool and one control pool. Recipient rats were injected with 3·5 ml serum intraperitoneally (i.p) on the same day as the serum was collected.
Serum was separated on a filter into two fractions containing structures larger and smaller than 100 000 MW according to the manufacturer's instructions (ultrafree-15, centrifugal filter device, Lot # VS 15358, Millipore, Corporation, Bedford, MA). The larger fraction was resuspended in phosphate-buffered saline (PBS), corresponding to the initial volume of serum. Both preparations, larger and smaller than 100 000 MW, were injected i.p into separate recipient rats.
Determination of OVA concentration in serum after 2 hr of feeding
Round-bottom microtitre plates (Greiner Labortechnik Ltd. Cambridge, UK) were coated with 5 µg/ml of affinity-purified goat immunoglobulin G (IgG) anti-OVA antibodies (Cappel,West Chester, PA, USA). Then 100 µl serum were added to the wells and serially diluted in eight two-fold steps and incubated for 3 hr at room temperature. The plates were washed three times with PBS–Tween and then alkaline phosphatase-conjugated goat IgG anti-OVA antibodies (Cappel) were added overnight. The enzyme activity was visualized by p-nitrophenyl phosphate (1 mg/ml, Sigma Chemical Co., St Louis, MO) dissolved in diethanolamine buffer, pH 9·8. The absorbance values (405 nm, Titertek multiscan, Flow Labs, MacLean, VA) were compared with the absorbance values obtained with an OVA standard. The sensitivity was 3 ng/ml.
Rats were immunized subcutaneously (s.c) in the hind legs 1 week after serum transfer with 100 µl of one of the following antigens mixed in Freund's complete adjuvant (Difco Laboratories, Detroit MI); 100 µg OVA (Grade V, Sigma); 100 µg OVA mixed with 100 µg HSA injected at the same site (Grade V, Sigma); or 100 µg HSA and 100 µg OVA injected at separate sites. A booster immunization was performed, in the same way and with the same doses 5 weeks after the primary immunization with Freund's incomplete adjuvant (Difco Laboratories). One week later the rats were killed by CO2 exposure.
Blood samples were collected from the tip of the tail at different time-points before and after the immunizations.
Two weeks after the primary immunization all rats were intracutaneously challenged with 50 µg OVA in 20 µl PBS in one ear and with 50 µg HSA in 20 µl PBS in the other ear. The ear thickness was measured before and 24 hr after challenge, with a micrometer caliper (Oditest, Kroplin, Hessen, Germany).
Determination of antibody levels
IgG antibodies against OVA and HSA were measured by an enzyme-linked immunosorbent assay (ELISA). Round-bottomed microtitre plates (Greiner Labortechnik) were coated with optimal concentrations of the antigens: OVA (5 µg/ml) and HSA (5 µg/ml, no blocking procedures were necessary). All antigens were diluted in PBS and incubated on the plates overnight at room temperature. The serum samples were diluted in PBS–Tween (0·05%) in eight five-fold steps and incubated overnight at room temperature. The plates were incubated with rabbit anti-rat IgG antibodies, diluted 1/5000 (Zymed Laboratories Inc, San Francisco, CA) for 2–3 hr at room temperature. Alkaline phosphatase-conjugated goat anti-rabbit IgG antibodies, diluted 1/10 000 (Sigma) were then added to the plates and incubated for 1 h. The enzyme activity in the wells was visualized with the substrate p-nitrophenyl phosphate (1 mg/ml, Sigma) dissolved in diethanolamine buffer, pH 9·8. The absorbance was measured in a spectrophotometer at 405 nm (Titertek multiscan). The incubations were performed in a humid atmosphere and between the incubations the plates were washed with 3 × 200 µl PBS–Tween using an ELISA washer (Titertek plus, ICN Biomedicals Inc., Costa Mesa, CA).
The antibody activity was expressed in arbitrary ELISA units calculated from a standard curve obtained with a hyper-immune serum.
Cell preparations and cell cultures
The spleens were removed and squeezed through a fine wire grid and filtered through a nylon filter (Falcon, Becton, Dickinson, NJ). The cells were then suspended in Hanks' medium and centrifuged on Ficoll (Ficoll–Paque Research Grade, Pharmacia Biotech AB, Uppsala, Sweden) for 30 min at 400 g without brake at room temperature. The mononuclear cell layer was collected and resuspended in Hanks' balanced salt solution. After repeated washes in Hanks' balanced salt solution the cells were then diluted in Iscove's medium supplemented with 10% fetal calf serum, 1% l-glutamine, 1% sodium pyruvate and gentamycin (100 µg/ml Garamycin, Schering-Plough, Madison, NJ).
The cells were finally cultured at a concentration of 4 × 105 viable cells/well in 200 µl Iscove's supplemented medium in flat-bottomed, 96-well plates (Nunc, Roskilde, Denmark). The cultures were stimulated with: concanavalin A, 2 µg/well (Con A, Sigma Chemical Co.); HSA, 80 µg/well (Sigma); or OVA, 80 µg/well (Sigma). All antigens/mitogen were sterile-filtered (0·22 µm Millipore S.A 67 Molheims, France) and diluted in Iscove's supplemented medium. The cells were incubated in a 5% CO2-atmosphere for 72 hr and pulsed with 1 µCi/well [3H]thymidine (Amersham International, Amersham, UK) for 12–16 hr. The cells were then harvested and dried (Inotech, Ninolab AB, Upplands Väsby, Sweden) and the incorporated [3H]thymidine was measured in a β-counter (Matrix96; Canberra Packard, Uppsala, Sweden).
Depletion of CD25-positive cells and fluorescence-activated cell sorter (FACS) analyses
Spleen cells from the group receiving serum from OVA-fed animals and from the controls were separately pooled. The cells were incubated with an optimal dilution of mouse anti-rat CD25 (OX-39) or isotype control antibody. A second incubation was performed with fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse IgG antibodies (Star 70, Serotec, Oxford, UK) and 1% rat serum. A final incubation with Mini Macs anti-FITC beads (Miltenyi, Biotec GmbH, Bergisch Gladbach, Germany) was performed. All incubations were carried out on ice for 20 min and the cells were washed with FACS-buffer (5 m m ethylenediaminetetraacetic acid, 0·25% bovine serum albumin, 0·02% sodium azide, pH 7·2) between the incubations. The cell suspensions were then added to Mini Macs separation columns (Miltenyi, Biotec GmbH). The cells not retained in the columns were seeded out for proliferation (negative fraction).
FACS analyses were performed to determine the frequency of CD25-positive cells. The frequency of CD25-positive cells was 10% before depletion and 2% after.
The Mann–Whitney U test was used for statistical analyses of differences in antibody levels, DTH-responses and T-cell proliferation in vitro. The Student's t-test was used for comparison of the in vitro proliferation after depletion of CD25-positive cells.
Tolerance against OVA in rats receiving serum from OVA-fed donors
Rats were given serum from OVA-fed donor rats or from control rats. The concentration of OVA in the injected serum was 340 ng/ml. The recipients were immunized with OVA s.c. in the hind leg 1 week after serum transfer and then booster immunized in the same manner 5 weeks later. Two weeks after the first immunization the rats were challenged intra-cutaneously in the ear with OVA to measure the DTH response.
Rats receiving serum from OVA-fed donor rats had a significantly reduced DTH response against OVA ( Fig. 1a) compared to animals given serum from control-fed animals.
OVA given as an i.p. dose corresponding to the amount in the transferred volume of serum from OVA-fed animals did not induce tolerance ( Fig. 1a). OVA-stimulated spleen cells from rats receiving tolerogenic serum showed a significantly lower in vitro proliferation compared to spleen cells from control rats 1 week after the booster immunization ( Fig. 1b). The counts per minute (c.p.m.) value of Con A-stimulated spleen cells from rats receiving tolerogenic serum was 25 214 c.p.m. (mean, n = 8) and in control cultures it was 25 864 c.p.m. (mean, n = 8). The background c.p.m. values in cultures of spleen cells without antigen were less than 100 c.p.m. in both groups. The background is substracted from the values shown in Fig. 1(b).
There were no differences in the IgG anti-OVA antibody production at the beginning of the experiment, but 6 weeks after the serum transfer the group receiving tolerogenic serum had lower IgG anti-OVA antibody levels than the control group ( Fig. 1c).
Distinct molecular size serum fractions, obtained from animals fed OVA, give rise to different tolerogenic activities in recipients.
Serum was separated into one fraction containing components smaller than 100 000 MW and one fraction with structures larger than 100 000 MW. Both fractions, prepared from serum collected from animals fed OVA, induced tolerance, i.e. lower DTH reaction against OVA in the recipients compared to the corresponding control group ( Fig. 2a). However, the tolerogenic activity was more prominent in the larger fraction compared to the smaller fraction ( Fig. 2a). The rats receiving the fraction containing larger components had, 3 weeks after the serum transfer, lower levels of IgG anti-OVA antibodies in serum compared to the control group ( Fig. 2b). This decrease in IgG antibody levels against OVA was not observed in recipients given components smaller than 100 000 MW.
CD25-positive cells are important for the suppression of the in vitro proliferation
Depletion of CD25-positive cells from suspensions of spleen cells taken 1 week after booster immunization from the group that received tolerogenic serum, resulted in an almost four-fold increase in the in vitro proliferation against OVA ( Fig. 3). Such an increase was not detected when CD25-positive cells were depleted from the spleen suspensions from the control group.
Bystander suppression in recipient rats
Rats were given serum from OVA-fed donor rats, or from control rats and were then immunized 1 week later with a mixture of OVA and HSA s.c. in the hind leg and were booster-immunized in the same manner 5 weeks later. The DTH response was measured 2 weeks after the primary immunization by intracutaneous challenge in the ear with HSA.
Twenty-four hours after the challenge with HSA the group that received serum from OVA-fed animals had a significantly lower DTH response against HSA compared to the controls given serum from control-fed animals ( Fig. 4a). The in vitro proliferation of HSA-stimulated spleen cells from rats receiving OVA tolerogenic serum was lower than the proliferation of HSA-stimulated spleen cells from control animals 1 week after booster immunization ( Fig. 4b). Significantly lower IgG anti-HSA antibody levels were detected 3 weeks after serum transfer in the group receiving the tolerogenic serum ( Fig. 4c). The IgG anti-HSA antibody levels continued to be low throughout the experiment. The corresponding values for the IgG anti-OVA antibody activity are shown in Fig. 1(c).
Reduced responses to HSA in animals receiving serum from OVA-fed donors were not seen if the rats were immunized with OVA and HSA at separate sites (not shown).
In the present study we have demonstrated that serum from animals fed OVA caused reduced T-cell responses and antibody production against OVA in recipients. The transferable tolerogenic activity could be recovered both in serum fractions containing components smaller or larger than 100 000 MW. CD25-positive cells appeared to be of importance for the observed tolerance since removal of these cells led to enhanced proliferation of OVA-stimulated spleen cells in vitro. Furhermore, suppression seemed to be one mechanism contributing to the observed tolerance since a reduction of the immune response to a bystander antigen, HSA, was demonstrated.
The finding that animals receiving serum from OVA-fed rats had a lower DTH reaction against OVA after immunization is consistent with studies in mice.4,5,22 In addition to the suppressed DTH reaction against OVA we also observed decreased antibody levels against OVA in the rats receiving serum from OVA-fed animals. Reduction of the antibody production has been reported in mice receiving tolerogenic serum but was assumed to be fortuitous in that study. 5 It is usually considered that antibody responses are more difficult to suppress than T-cell responses. 12 We determined the antibody activity in serum from the recipients on several occasions during the experiment. The difference between the groups in our study was first evident 5 weeks after the serum transfer. In the mouse studies, antibody analyses were only performed 3 weeks after serum transfer. Another difference between the current study and the studies on mice is the feeding regimen. In the present study the animals were fed an OVA-containing diet ad libitum for 2 h in contrast to the mouse studies where the animals were tube-fed with OVA. This may lead to a different processing of the protein that might affect the nature of the tolerogenic serum factor. It should be noted, though, that the concentrations of OVA in the donor sera were comparable to the concentration in donor mice sera in previously published studies.5,23 The serum concentration of OVA does not, however, correlate with the concentration of the serum factor. 6 Instead, gut processing of the antigen has been proven necessary since i.p. injection of OVA corresponding to the amount of OVA in the transferred serum from OVA-fed animals did not induce tolerance. Hence, the tolerance-inducing activity found in serum is not just native protein. In accordance, we showed that the tolerogenic activity could be found in a fraction of serum containing components larger than 100 000 MW. This high molecular weight fraction from serum induced a decreased DTH reaction against OVA and lower levels of IgG anti-OVA antibodies in recipients. Thus, the tolerogenic activity, or factor, could be recovered in a fraction containing molecules larger than monomeric OVA which has a molecular weight of 43 000. It is not likely that the activity in the larger fraction is aggregated OVA, since aggregated proteins prime animals.24,25 The fraction of serum containing structures smaller than 100 000 MW also induced a decreased DTH reaction against OVA in the recipients but to a lesser extent. This fraction did not suppress the levels of IgG anti-OVA antibodies in the recipients. The finding that a low molecular weight fraction transferred tolerance is consistent with a study done by Furrie et al. 7 One can speculate that the tolerance induced by the smaller fraction could partly be due to monomeric OVA since it has been shown that monomeric proteins are tolerogenic. 26
A novel finding in this study was that depletion of CD25-positive cells in vitro increased the proliferation of OVA-stimulated spleen cells from rats receiving tolerogenic serum, suggesting that these cells had immunomodulating functions at least in vitro. Regulatory effects have previously been associated with CD25-positive cells.15,17, 18,20,27 Thornton et al. recently showed that co-culturing of CD25-positive/CD4-positive cells with CD25-negative/CD4-positive cells markedly suppressed the proliferation of the CD25-negative cells after polyclonal activation. 17 It has been suggested that CD25-positive cells act as scavenger cells and consume interleukin-2, resulting in a lack of this growth factor for other T cells. 20
Another new finding in this study was that the tolerogenic serum suppressed the DTH and antibody responses to an unrelated antigen, i.e. bystander suppression or active tolerance. Bystander suppression is presumably driven by antigen-specific cells that secrete suppressive cytokines that act unspecifically. 28–32 Recently it was also shown by Taams et al. that antigen-specific cells can down-regulate the polyclonal response of other T cells through contact. 33 They called this phenomenon linked suppression.
The reduction of the IgG anti-HSA antibody production seemed to occur more rapidly than the down-regulation of the IgG antibody production against OVA. This could be due to the tolerant animals having had contact with OVA prior to immunization, in conjunction with the transfer of tolerogenic serum. This finding was consistent with a study from our laboratory where an initial increase of specific antibodies to the fed antigen was observed in the tolerant group after immunization. 34 In conclusion, we have demonstrated that rats fed OVA develop an OVA-specific tolerogen activity or factor in serum. The activity could be found both in a fraction containing components larger than 100 000 MW and in a fraction with structures smaller than 100 000 MW. Thus, indicating that there could be more than one factor operating. The serum factor(s), presumably induces regulatory suppressor cells in the recipients since we noted impaired antibody and T-cell responses to a bystander antigen both in vitro and in vivo. The tolerance was at least partly mediated by CD25-positive cells in the recipients since depletion of these cells abolished the suppression in vitro.
We are grateful for the invaluable work of the staff in the animal house at Sahlgrenska University Hospital and to Inger Pettersson for excellent help with the FACS analyses. This research was supported by grants from Swedish Medical Research Council (No. 215), the Vårdal Foundation, the Ellen, Walter and Lennart Hesselmann Foundation and the Konsul Th C Berghs Foundation.