Transient inhibition of Th1-type cytokine production by CD4⁺ T cells in hepatitis B core antigen immunized mice is mediated by regulatory T cells

Abstract

The non-cytopathic hepatitis B virus (HBV) can induce chronic infections characterized by weak and limited T cell responses against the virus. The factors contributing to the failure to clear HBV and subsequent development of chronic HBV infections are not clearly understood, but a strong interferon-gamma (IFN-γ) response by CD4⁺ T cells against the nucleocapsid hepatitis B core antigen (HBcAg) of the virus appears to be important for viral clearance. The present study documents depressed numbers of CD4⁺ T cells secreting IFN-γ and interleukin-2 (IL-2) in enzyme-linked immunospot assay (ELISPOT) assays restimulated for 24 hr with antigen following both primary and secondary immunizations of mice with recombinant hepatitis B core antigen (rHBcAg). The kinetics of these responses showed that the depression occurred following a peak response and lasted approximately 2 weeks before returning to the previous peak levels. The depression was abrogated by depletion of CD25⁺ cells prior to culture in the ELISPOT assay, suggesting inhibition by regulatory T cells. This inhibition of IFN-γ and IL-2 production was also reversed by in vitro restimulation of the test cells for 48 hr rather than 24 hr in the assay. No such transient, reversible inhibition was detected in the production of IL-5, a Th2-type cytokine. The inhibition in cytokine production did not appear to correlate with the number of antibody-secreting cells or the isotypes produced. This delay by regulatory T cells of Th1-type cytokine production could contribute to viral persistence in chronic HBV infection by interfering with the critical role IFN-γ plays in protection against viral infections.

Introduction

CD4⁺ T cells contribute in major ways to the control and clearance of viral infections, including production by Th1-type cells of the antiviral cytokines interferon-gamma (IFN-γ) and tumour necrosis factor-alpha (TNF-α).^1–5 To focus on the IFN-γ response and to gain insight into how this response is controlled, the response to the nucleocapsid protein [hepatitis B core antigen (HBcAg)] of the hepatitis B virus (HBV) was studied. This viral antigen has been shown to be an efficient immunogen with the ability to elicit potent cellular as well as humoral immune responses.^6–8 HBV is of particular interest because it can induce chronic as well as acute infections in the host and the persistent infections can lead to the development of cirrhosis and liver cancer.⁹ The factors that determine chronicity of HBV infection and the reason for the absence of a strong T cell response in chronic infections remain unclear. Some studies suggest that a strong HBcAg-specific Th1-type response may be essential for eliminating HBV^10–12 because significant CD4⁺ T cell IFN-γ responses to HBcAg have been detected in acute but not chronically infected patients.^13,14 Despite the lack of detectable CD4⁺ T cell responses in patients chronically infected with HBV, these patients produce high titres of anti-HBcAg at levels comparable to those detected in acute patients.^15,16 Understanding the kinetics of the immune response involved in the development of an acute versus chronic response to HBcAg might elucidate the factors responsible for the different disease outcomes.

A kinetic study of epitope-specific CD4⁺ T cell responses using lymphocytic choriomeningitis virus (LCMV) infection demonstrated that following viral infection there is an initial expansion stage, after which the cells enter a contraction phase.¹⁷ The contraction phase for CD4⁺ T cells appears as a steady decrease in the number of specific CD4⁺ T cells,¹⁷ unlike that of CD8⁺ responses, which decline rapidly to a plateau level of memory cells.^17,18 The nature of this phase may be important in determining if an infection, including infection with HBV, will be resolved or persist in the host. More specifically, the persistence or strength of effector functions, such as IFN-γ production, of these CD4⁺ T cells following the peak response could impact the outcome of disease. Unfortunately, little is known about the kinetic patterns of IFN-γ production by CD4⁺ T cells during antiviral immune responses and consequently the mechanisms controlling these patterns remain unknown.

In that regard, previous work in our laboratory evaluating IFN-γ production by CD4⁺ T cells from recombinant HBcAg (rHBcAg) immunized mice in enzyme-linked immunospot (ELISPOT) assays demonstrated that there were differences in response requirements for IFN-γ production early and late after antigen exposure.¹⁹ Cells isolated late after antigen exposure required a longer period of in vitro restimulation to develop a peak IFN-γ response when compared to those isolated earlier in the response. This finding was consistent in both primary and secondary rHBcAg responses. This delay in HBcAg-specific IFN-γ production by CD4⁺ T cells following the peak response provides evidence of unrecognized controls exerted on this response throughout the course of its development. Therefore, we sought to study how the IFN-γ response changed over time with exposure to rHBcAg. The aims of this study were to examine the roles that cytokine secretion and antibody production played in the progression and control of the rHBcAg immune response.

Materials and methods

Immunization

We used (C₃H × CB17)F₁ mice (H-2^{k × d}) in all experiments. These mice were bred and housed in Thomas Jefferson University's Association for Assessment and Accreditation of Laboratory Animal Care (AALAC) accredited animal facility. Mice were immunized at 6–8 weeks of age with 50 µg recombinant hepatitis B core antigen (rHBcAg ayw subtype, kindly provided by Dr Darrell Peterson, Virginia Commonwealth University, Richmond, VA) or 50 µg recombinant hepatitis B surface antigen (rHBsAg adw subtype, kindly provided by Merck Research Laboratories, West Point, PA) subcutaneously in incomplete Freund's adjuvant (IFA; Sigma, St Louis, MO). Secondary responses were tested in mice challenged with 20 µg of rHBcAg or rHBsAg in phosphate-buffered saline (PBS) intraperitoneally (i.p.) 4 weeks after priming. The 50 and 20 µg doses were chosen because they produced the optimal IFN-γ responses as measured by ELISPOT.

Cell preparation

Harvested spleens were teased into single cell suspensions. Red blood cells were lysed using ACK cell lysing buffer [0·15 m NH₄Cl, 10 mm KHCO₃, 0·1 mm Na₂ ethylenediamine tetraacetic acid (EDTA)]. Cells were then washed, passed through a 70 µm cell strainer (BD Falcon, Bedford, MA), and viable cells counted using trypan blue (Gibco, Carlsbad, CA).

Test splenocytes were depleted of CD25⁺ cells by magnetic bead separation (Miltenyi Biotec, Auburn, CA). Splenocytes were incubated with fluorescein isothiocyanate (FITC)-conjugated anti-CD25 (BD Pharmingen, San Diego, CA) for 20 min at 4°. Cells were then washed thoroughly and incubated with anti-FITC MicroBeads (Miltenyi Biotec) for 20 min at 4°. After washing, cell populations were depleted of CD25⁺ cells by passing the labelled splenocytes through a magnetic column (LS column, Miltenyi Biotec) according to the manufacturer's instructions. The depletion of CD25⁺ cells in these test cell populations was > 99% by flow cytometry. The percentage of CD25⁺ cells in the positively selected population ranged between 40 and 60%.

ELISPOT assay

Cytokine ELISPOT assays were performed for IFN-γ, interleukin-5 (IL-5) or IL-2-secreting cells in 96-well MultiScreen-IP plates (Millipore, Bedford, MA). Wells were precoated with 100 µl of either anti-IFN-γ, anti-IL-5 or anti-IL-2 capture antibody (BD Pharmingen) at 5 µg/ml in PBS by incubation overnight at 4°. Plates were then washed with PBS prior to incubation with 200 µl complete culture medium [Iscove's medium (Cellgro, Herndon, VA) supplemented with 5% fetal calf serum (FCS; Gibco), 1% Na pyruvate (100 mm, Cellgro), 1% P/S (100 µg/ml, Cellgro), and 5 × 10⁻⁵ m 2-mercaptoethanol (Gibco)] for 2 hr at 37° to block non-specific binding. After blocking, test splenocytes were then added at 10⁶ cells per well and stimulated 5 µg/ml rHBcAg or rHBsAg in a total volume of 100 µl complete culture medium. Plates were incubated for 24 or 48 hr at 37°, after which they were washed thoroughly with PBS + 0·05% Tween 20 (PBST). Biotinylated anti-IFN-γ, anti-IL-5 or anti-IL-2 (1 µg/ml in PBST + 5% FCS) was then added to each well. After incubation for 2 hr at room temperature, the plates were washed and streptavidin–alkaline phosphatase was added (1:1000, BD Pharmingen) for 1 hr at room temperature. Plates were washed thoroughly with PBST followed by PBS, and 100 µl of one-step NBT/BCIP solution (Pierce, Rockford, IL) was added to each well. The reaction was stopped after 10–15 min by washing the plates thoroughly with tap water. Frequencies of IFN-γ, IL-5 and IL-2 producing cells were determined by counting the number of spots per well using a dissecting microscope. Background cytokine secretion in response to medium alone was subtracted from all values.

Antibody ELISPOT assays for HBcAg-specific antibody-secreting cells were performed in 96-well MultiScreen-HA plates (Millipore), precoating wells with 100 µl of goat anti-mouse IgG (Cappel, MP Biomedicals, Irvine, CA) at 5 µg/ml in PBS by incubation overnight at 4°. Washing and blocking was performed similarly to the cytokine ELISPOT. Test splenocytes were then added at 10⁶, 5 × 10⁵ and 3 × 10⁵ cells per well in a total of 100 µl complete culture medium. After incubating undisturbed for 5 hr at 37°, plates were washed thoroughly with PBST. Biotinylated rHBcAg (1 µg/ml in PBST + 5% FCS) was added to each well. After incubating for 1 hr at room temperature, the plates were washed and streptavidin–alkaline phosphatase was added (1:1000, BD Pharmingen) for 1 hr at room temperature. Plates were developed and antibody forming cells were counted as described for the cytokine ELISPOT assay. The results are expressed as the number of anti-HBcAg spot-forming cell (SFC) per 10⁶ cells plated.

Antibody isotype enzyme-linked immunosorbent assay (ELISA)

ELISA assays for rHBcAg-specific antibody were performed in 96-well Immunlon 4 HBX plates (Thermo Electron Corporation, Milford, MA). Wells were precoated overnight at 4° with 50 µl of rHBcAg at 2·5 µg/ml in phosphate coating buffer (0·1 m Na₂HPO₄ adjusted to pH 9·0 with 0·1 m NaH₂PO₄). Plates were then washed with PBS prior to incubation with 200 µl of PBS + 4% powdered milk for 1 hr at 37° to block non-specific binding. After blocking, wells were washed with PBS + 0·05% Tween 20 (PBST) and test serum samples in titrated twofold dilutions (in PBS + 1% powdered milk) were then added and incubated for 2 hr at 37°. After this incubation, plates were washed thoroughly with PBST and biotinylated anti-IgG1 or anti-IgG2a (1 µg/ml in dilution buffer, BD Pharmingen) was then added to each well. After incubating for 1 hr at room temperature, the plates were washed with PBST and streptavidin–alkaline phosphatase was added (1:2000, BD Pharmingen) for 30 min at room temperature. Plates were washed thoroughly with PBST, and 100 µl of pNPP [1 p-nitrophenyl phosphate disodium hexahydrate tablet (Sigma) dissolved in 1 m diethanolamine pH 9·8 with 0·5 mm MgCl₂] was added to each well. After 5 min, the plates were analysed at 405 nm by an ELISA plate reader (EL_x800, Bio-Tek Instruments Inc., Winooski, VT). The amount of antibody isotype detected in serum was calculated using standard curves determined using murine isotype standards [10 µg/ml IgG2a; Accurate Chemical & Scientific Corporation, Westbury, NY; and 1 µg/ml IgG1; AB1·2 American Type Culture Collection (ATCC), Manassas, VA].

Statistical analysis

Data are expressed as mean ± standard deviation. Statistical significance was determined by paired t-tests. Differences between values were considered statistically significant at P ≤ 0·05.

Results

Transient depression in IFN-γ production detected after primary and secondary stimulation with rHBcAg

To identify points of regulation in the CD4⁺ T cell response to rHBcAg, a kinetic study of the IFN-γ response following immunization with rHBcAg in IFA was performed. For analysis of primary responses, mice were immunized with 50 µg rHBcAg subcutaneously (s.c.) in IFA. Splenocytes were isolated from immunized mice on different days after priming and the number of HBcAg-specific IFN-γ-secreting cells was determined in a standard ELISPOT assay following 24 hr in vitro restimulation with rHBcAg (Fig. 1a, solid line). A significant elevation in the number of IFN-γ-secreting cells was detected by day 7 following immunization with rHBcAg. The number of IFN-γ-secreting cells remained constant to day 14; however, it decreased at day 28 when tested in a 24 hr assay. By day 42 after immunization, the number of IFN-γ-secreting cells increased again to levels equal to or greater than those found at 7 and 14 days. Thus, when IFN-γ-secreting cells were assayed after 24 hr of antigen restimulation in vitro, there appeared to be a transient depression occurring 28 days after immunization that was no longer detectable 42 days after immunization.

Figure 1.

Transient depression in interferon-gamma (IFN-γ) production detected after 24 hr antigen restimulation in vitro is alleviated after 48 hr in both primary and secondary responses. Mice were immunized subcutaneously (s.c.) with 50 µg recombinant hepatitis B core antigen (rHBcAg) in incomplete Freund's adjuvant (IFA) with or without a challenge dose of 20 µg rHBcAg in phosphate-buffered saline (PBS) intraperitoneally (i.p.) 4 weeks later. Splenocytes were isolated on indicated days from mice primed only (a) or primed and challenged (b) with rHBcAg. IFN-γ-secreting cells were detected by enzyme-linked immunospot assay (ELISPOT) assays in which 10⁶ splenocytes were restimulated in vitro with 5 µg/ml rHBcAg and incubated for 24 (○) or 48 (•) hr at 37°. Data are represented as the average number of IFN-γ spot-forming cells (SFC)/10⁶ cells of at least six independent experiments ± SEM. Background numbers of SFC detected in medium controls (7 ± 4, in primed cells) (13 ± 8, in primed and challenged cells) were subtracted from all test values. *P ≤ 0·05.

To determine if rHBcAg-specific memory cells were still susceptible to a transient depression in their IFN-γ response following antigen exposure, rHBcAg primed mice were challenged with rHBcAg and tested for IFN-γ-secreting cells at different times thereafter. For this, mice were primed with 50 µg of rHBcAg in IFA and received a 20 µg in vivo challenge 28 days later. The kinetics of the secondary HBcAg-specific IFN-γ response measured in a standard 24 hr ELISPOT assay followed a similar, albeit more rapid, pattern to that seen during the primary response (Fig. 1b, solid line). In splenocytes taken from mice between 7 and 14 days following an in vivo rHBcAg challenge, the number of IFN-γ-secreting cells detected with 24 hr of antigen restimulation in vitro appeared depressed relative to the response detected at day 3 or day 28. This decrease in the numbers of IFN-γ-secreting cells reached statistical significance (P ≤ 0·002) at day 14 after challenge. Thus, when measured after 24 hr of antigen restimulation in vitro, as in the primary response, there was an initial increase in the number of IFN-γ-secreting cells followed by a period of time when the IFN-γ response was decreased, and then elevated again later.

Transient depression in IFN-γ production detected in 24 hr ELISPOT assays was reversed in 48 hr assays

It has been demonstrated previously by this laboratory that CD4⁺ T cell production of IFN-γ in both primary and secondary responses appeared to be delayed at late time-points after rHBcAg exposure relative to that at early time-points.¹⁹ Therefore, we sought to determine if the depressed IFN-γ responses detected at 24 hr of in vitro antigen restimulation were changed by increasing the time of antigen restimulation in the ELISPOT assay. Mice were primed with rHBcAg with or without an in vivo antigen challenge as described above for the 24 hr assays. When test splenocytes from the primary CD4⁺ T cell response were restimulated in vitro with rHBcAg for 48 hr rather than 24 hr, the number of IFN-γ-secreting cells detected in splenocytes from mice 28 days after in vivo challenge were significantly increased, compared to the 24 hr assay, and this response was no longer depressed relative to that detected at 7 or 42 days (Fig. 1a, dashed line). A similar comparison in splenocytes from mice, taken 7 or 42 days after priming, time-points that were not depressed at 24 hr showed no significant difference between the 24 and 48 hr assay results. As in the primary response, when the secondary IFN-γ response was measured after 48 hr in vitro exposure to rHBcAg in the ELISPOT assay, the statistically significant depression in the IFN-γ response detected in the day 14 response in the 24 hr assay was reversed with additional in vitro rHBcAg restimulation (Fig. 1b, dashed line). The number of IFN-γ-secreting cells detected at day 3 and day 28 after in vivo antigen challenge, which were not depressed in 24 hr assays, did not change when the in vitro rHBcAg restimulation was increased to 48 hr. Therefore, it appeared that the 48 hr response was elevated relative to the 24 hr response only when IFN-γ production appeared to be depressed in 24 hr assays. Of note, restimulation in vitro for 72 hr in the ELISPOT assay did not result in a significant increase over the 48 hr response in the number of IFN-γ-secreting cells (data not shown). These data demonstrate that when the primary and secondary rHBcAg-specific IFN-γ responses were measured at 48 hr of antigen restimulation in vitro, the depression in the responses detected in 24 hr assays was reversed.

It has been demonstrated previously that this IFN-γ production in response to immunization with rHBcAg is a function of CD4⁺ T cells.¹⁹ It should be noted that the percentage of CD4⁺ splenocytes in mice 3, 14 and 28 days after in vivo antigen challenge ranged between 18 and 24%, with no significant difference detected between the different days. Furthermore, the average absolute number of CD4⁺ splenocytes in the different types of mice varied from each other by less than 10%. Thus, the transient depression in IFN-γ production seen in 24 hr ELISPOT assays is not due to changes in the numbers of potential responding cells.

Transient, reversible depression in cytokine-secreting cells detected in IL-2 but not IL-5 ELISPOT assays

To investigate whether a similar transient depression of another Th1-type cytokine occurred in rHBcAg immune mice, IL-2 production during the secondary response was tested. Splenocytes from mice primed and challenged with rHBcAg as above were tested for IL-2 production at different times after challenge by ELISPOT (Fig. 2a). When cells from mice taken 7 and 14 days after an in vivo challenge were restimulated with rHBcAg in vitro for 24 hr in the ELISPOT assay, the number of IL-2-secreting cells was significantly (P ≤ 0·001) decreased compared to IL-2 production by cells from mice 3 days after challenge (Fig. 2a, solid line). This decrease in IL-2 production was transient, as the number of IL-2-secreting cells increased (P ≤ 0·02) by day 28 to levels equivalent to those seen at day 3 after antigen challenge. In contrast, no decrease in the numbers of IL-2-secreting cells was detected when the test cells were restimulated in the ELISPOT assay for 48 instead of 24 hr (Fig. 2a, dashed line). This was seen even in splenocytes taken at day 14 after in vivo challenge, where the depression of the response had been the greatest in 24 hr assays. Splenocytes from mice taken 7 and 14 days after in vivo challenge exhibited a significant increase (P ≤ 0·02) in the numbers of IFN-γ-secreting cells measured 48 hr after in vitro restimulation with rHBcAg relative to that seen at 24 hr. Thus, production of IL-2 appeared to follow a pattern similar to the production of IFN-γ. In both responses, a transient depression in cytokine production was detected in spleen cells from mice 7 and 14 days after antigen challenge in vivo at 24 hr of in vitro restimulation with rHBcAg that was reversed in 48 hr assays.

Figure 2.

Transient depression in interleukin-2 (IL-2), but not IL-5, production was detected in secondary responses. Mice were primed and challenged with recombinant hepatitis B core antigen (rHBcAg) as described in the legend of Fig. 1. Splenocytes were isolated on different days after in vivo rHBcAg challenge and IL-2 (a) and IL-5 (b) secreting cells were detected by enzyme-linked immunospot assay (ELISPOT) assays in which 10⁶ splenocytes were restimulated in vitro with 5 µg/ml rHBcAg and incubated for 24 (○) or 48 (•) hr at 37°. Data are represented as the average number of cytokine spot-forming cells (SFC)/10⁶ cells of at least three independent experiments ± SEM. Background levels of SFC detected in medium controls (134 ± 31, IL-2 SFC; 22 ± 15, IL-5 SFC) and in splenocytes from non-immunized mice (27 ± 11, IL-2 SFC; 6 ± 2, IL-5 SFC) were subtracted from all test values. *P ≤ 0·05.

To determine if the transient depression in IFN-γ and IL-2 production detected in splenocytes following rHBcAg immunization also occurred in the production of a Th2-type cytokine, IL-5-secreting cells were enumerated in mice immunized and challenged with rHBcAg, as described above. IL-5-secreting splenocytes were enumerated at different times after antigen challenge by ELISPOT after 24 and 48 hr of in vitro restimulation with rHBcAg. The number of IL-5-secreting cells remained consistently low throughout day 14 after in vivo challenge when they were tested in the ELISPOT after only 24 hr of antigen restimulation in vitro (Fig. 2b, solid line). The number of IL-5-secreting cells detected at all time-points tested after the challenge dose increased when restimulated in vitro for 48 rather than 24 hr (Fig. 2b, dashed line). This observation is consistent with previous studies, demonstrating that it takes 48 hr to detect a maximal IL-5 response by ELISPOT.²⁰ Similar results were detected when the kinetics of the primary IL-5 response were determined in rHBcAg immune mice (data not shown). These data show that although there is a transient depression in production of the Th1-type cytokines, IFN-γ and IL-2, no such depression was detected after either primary or secondary rHBcAg immunizations when the production of a Th2-type cytokine, IL-5, was analysed.

No depression detected in IFN-γ or IL-2 production in splenocytes from rHBsAg immunized mice

To determine if the transient depression in Th1-type (IFN-γ and IL-2) cytokine production in rHBcAg immune mice was unique to rHBcAg immune responses, cytokine production in response to a different antigen was studied. This was tested using rHBsAg, another antigen from the same virus (HBV) that is similar to rHBcAg in its particulate nature. Mice were primed and challenged with rHBsAg using the same protocols described previously for rHBcAg. Cytokine production was then analysed on different days after in vivo rHBsAg challenge by ELISPOT. The number of IFN-γ-secreting cells at 24 hr of in vitro restimulation was elevated (P ≤ 0·04), compared to unchallenged mice, by day 3 following an in vivo challenge and then remained at that level to day 14 (Fig. 3a). The number of IFN-γ-secreting cells did not increase from that found at 24 hr when these cells were restimulated in vitro with rHBsAg for 48 hr. Thus, unlike IFN-γ production stimulated by rHBcAg, no depression in the production of IFN-γ was detected during the first 14 days following in vivo antigen challenge in rHBsAg immune mice. The kinetics of IL-2 production was also determined in spleen cells from rHBsAg immunized mice (Fig. 3b). The number of IL-2-secreting cells at 24 hr of in vitro restimulation remained constant to day 14 after in vivo challenge. Additionally, there was no change in the number of IL-2-secreting cells at any time-point tested in the 14 days following an in vivo challenge when detected at 48 hr, instead of 24 hr, of in vitro rHBsAg restimulation. A similar pattern of cytokine production was detected when the number of IL-5-secreting cells was detected in rHBsAg immune mice following an in vivo antigen challenge (data not shown). Therefore, there was no evidence of transient periods of depression in the production of IFN-γ or IL-2 during rHBsAg immune responses.

Figure 3.

There is no depression in interferon-gamma (IFN-γ) or interleukin-2 (IL-2) production in splenocytes from recombinant hepatitis B surface antigen (rHBsAg) immune mice. Mice were immunized with 50 µg of rHBsAg subcutaneously (s.c.) and challenged 4 weeks later with 20 µg of rHBsAg in phosphate-bufferd saline (PBS) i.p. Splenocytes were isolated from mice on the indicated days after in vivo rHBsAg challenge and IFN-γ(a) or IL-2 (b) secreting cells were detected by enzyme-linked immunospot assay (ELISPOT) assays in which 10⁶ splenocytes were restimulated in vitro with 5 µg/ml of rHBsAg and incubated for 24 (○) or 48 (•) hr. Data are represented as the average number of cytokine spot-forming cells (SFC)/10⁶ cells ± SEM of at least three independent experiments. Background numbers of SFC detected in medium controls (9 ± 6, IFN-γ SFC; 111 ± 32, IL-2 SFC) and in splenocytes from non-immunized mice (9 ± 8, IFN-γ SFC; 29 ± 32, IL-2 SFC) were subtracted from all test values.

The presence of anti-HBcAg antibody-secreting cells in the ELISPOT assay did not contribute to the depression in Th1-type cytokine production

It has been demonstrated previously that HBcAg induces a strong anti-HBcAg antibody response in mice.¹⁵ Therefore, one possible cause of the depression in cytokine production in vitro by cells from rHBcAg immune mice could be the removal of stimulating antigen in culture by specific antibody secreted by HBcAg-specific antibody-secreting cells also present in the wells of the ELISPOT assay. To quantify antibody secretion in the test cultures, splenocytes were isolated from mice on different days after in vivo rHBcAg challenge and assayed for HBcAg-specific antibody-secreting cells using an antibody ELISPOT assay (Fig. 4a). The greatest number of anti-HBcAg-secreting cells was detected at day 3 following challenge. By day 7 after challenge, the number of anti-HBcAg-secreting cells began to decline and returned to the level seen in unchallenged mice by day 14. Thus, the number of cells secreting anti-HBcAg antibody was low 14 days after challenge, a time-point when the IFN-γ response was depressed, and was the largest 3 days after challenge when no depression was detected in IFN-γ production by ELISPOT. These results show no positive correlation between the number of anti-HBcAg antibody-secreting cells and depression of Th1-type cytokine (IFN-γ and IL-2) production, indicating that the presence of anti-HBcAg-secreting cells in the wells of the ELISPOT assay was not the cause of the depression in the IFN-γ and IL-2 responses.

Figure 4.

Transient depression in interferon-gamma (IFN--γ) production after immunization does not correlate with the presence of anti-hepatitis B core antigen (HBcAg) antibody-secreting cells or a switch in antibody isotype. Mice were primed and challenged as described in the legend of Fig. 1. (a) Splenocytes were isolated on the indicated days after in vivo recombinant hepatitis B core antigen (rHBcAg) challenge and assayed for rHBcAg-specific antibody forming cells by enzyme-linked immunospot assay (ELISPOT) after incubating for 5 h at 37°. Data are represented as the average number of rHBcAg-specific spot-forming cells (SFC)/10⁶ cells ± SD of replicate wells. One representative experiment of three independent experiments is shown. (b) The isotype of the HBcAg-specific antibody was determined from serum of HBcAg immune mice obtained at different days after in vivo challenge by enzyme-linked immunosorbent assay (ELISA) comparing the concentration of IgG1 (•) and IgG2a (○) to commercial isotype standards run at the same time. Note that day 0 after challenge is 28 days after the initial priming dose. Data are represented as µg/ml ± SD of HBcAg-specific antibody isotype. One representative experiment of five is shown.

IgG2a was the predominant anti-HBcAg antibody isotype throughout the response

A shift from a dominant Th1-type response to a dominant Th2-type response provides another possible explanation for the decrease in the observed IFN-γ response to rHBcAg. Because a shift in Th1- and Th2-type responses would be reflected in the isotype profile of the antibody response to rHBcAg, we next wanted to see whether this transient depression in Th1-type cytokine production had an effect on the isotype of circulating rHBcAg-specific antibody. Serum was collected from rHBcAg immune mice at different time-points after in vivo rHBcAg challenge and the isotype of the anti-HBcAg antibodies was analysed by ELISA (Fig. 4b). The IgG2a anti-HBcAg peaked between days 7 and 14 after challenge, decreasing slightly by day 28. The dominant anti-HBcAg isotype in the serum detected throughout the response was IgG2a. IgG1 was also detected in the serum of these mice; however, it was present in much smaller concentrations. This IgG1 antibody concentration followed the same kinetic pattern as the IgG2a response. There did not appear to be a change in the ratio of HBcAg-specific IgG1:IgG2a to 14 days after challenge and both antibody isotypes declined in the serum by 28 days, providing no evidence of a switch in the anti-HBcAg isotype from IgG2a to IgG1. The prolonged appearance of secreted rHBcAg-specific antibodies detected in the serum of rHBcAg immune mice differs from the rapid disappearance of HBcAg-specific antibody forming cells in the spleen, which reflects very short-lived stimulation and half-lives of B cells following rHBcAg immunization (Fig. 4a). The lack of a change in IL-5 production (Fig. 2) or in the IgG2a dominance (Fig. 4b) in specific antibody production indicates that the depression seen in the Th1-type cytokine responses to rHBcAg does not involve a switch toward Th2-type responses in rHBcAg immune mice.

Depletion of CD25⁺ cells eliminates inhibition in HBcAg-specific IFN-γ production

An alternative mechanism possibly contributing to the transient depression in IFN-γ production in rHBcAg immune mice could be the activation of a regulatory process. To investigate this possibility, the role of CD4⁺CD25⁺ regulatory T cells in the depression of the rHBcAg-specific IFN-γ response was examined. This was conducted by depletion of CD25⁺ cells from test splenocyte populations prior to in vitro antigen restimulation with rHBcAg in the ELISPOT assay. Mice were primed with rHBcAg in IFA with or without an antigen challenge 28 days later and splenocytes were isolated from mice on days 7 and 14 after the in vivo rHBcAg challenge. CD25⁺ splenocytes were depleted by magnetic bead separation prior to assaying for IFN-γ-secreting cells by ELISPOT. As seen in previous figures, unseparated splenocytes from mice 7 and 14 days after in vivo rHBcAg challenge exhibited in 24 hr ELISPOT assays IFN-γ responses that were depressed relative to those of unchallenged mice (Fig. 5a). When CD25⁺ cells were depleted from these challenged splenocyte populations, the depression of the IFN-γ response seen 7–14 days after antigen challenge was abrogated (Fig. 5a). Removal of CD25⁺ cells had no effect on the IFN-γ response in unchallenged mice at 24 hr. CD25⁺ cell depletion had no effect on the 48 hr IFN-γ response, when IFN-γ production was no longer depressed, in cells from unchallenged or challenged mice (data not shown). When CD25⁺ cells were depleted prior to assaying for IL-2 production, another Th1-type cytokine, in rHBcAg immune mice, an increase in the numbers of IL-2-secreting cells was detected in the 24 hr ELISPOT assay (data not shown), similar to that detected for the IFN-γ response.

Figure 5.

Depletion of CD25⁺ cells prior to in vitro antigen restimulation relieves the depression in interferon-gamma (IFN-γ) production but has no effect on interleukin-5 (IL-5) production. Mice were primed and challenged as described in the legend of Fig. 1. Splenocytes were isolated from mice on the indicated days after in vivo recombinant hepatitis B core antigen (rHBcAg) challenge and CD25⁺ cells were removed by depletion with magnetic bead separation. IFN-γ(a) and IL-5 (b) secreting cells from unseparated (○) or CD25 depleted (•) cell populations were detected by enzyme-linked immunospot assay (ELISPOT) after 24 hr of in vitro restimulation with 5 µg/ml rHBcAg. Data are represented as the average number of cytokine spot-forming cells (SFC)/10⁶ cells ± SD of duplicate wells. Background numbers of spot-forming cells (SFC) detected in medium controls (4 ± 3, IFN-γ SFC; 10 ± 5, IL-5 SFC) and in splenocytes from non-immunized mice (25 ± 14, IFN-γ SFC; 17 ± 7, IL-5 SFC) were subtracted from all test values. *P ≤ 0·05. One representative experiment of three is shown.

In contrast to IFN-γ, IL-5 production in rHBcAg primed mice was not depressed following challenge (Fig. 2). Therefore, the effect of CD25⁺ cell depletion on IL-5 production was studied for comparison with its effects on the depressed production of IFN-γ. As above (Fig. 2b), there was no significant depression in the number of IL-5-secreting cells following in vivo antigen challenge in unseparated spleen cell populations measured in an ELISPOT assay at 24 hr of in vitro rHBcAg restimulation (Fig. 5b). Similar results were seen in 48 hr ELISPOT assays (data not shown). Removal of the CD25⁺ cells did not significantly increase the number of IL-5-secreting cells at either 24 hr (Fig. 5b) or 48 hr (data not shown) of in vitro rHBcAg restimulation. These data, showing that the depressed IFN-γ but not the IL-5 response was enhanced following CD25⁺ cell depletion, support the hypothesis that CD4⁺CD25⁺ regulatory T cells may play a role in actively inhibiting the rHBcAg-specific IFN-γ response.

Discussion

The data presented here demonstrate that following immunization with rHBcAg there is a period of time when CD4⁺ T cell production of Th1-type cytokines, IFN-γ and IL-2 is depressed, requiring longer in vitro restimulation in the ELISPOT assay to develop. This transient depression was detected as decreased numbers of IFN-γ and IL-2-secreting cells after 24 hr of in vitro restimulation that was reversed by 48 hr in the presence of antigen. The depression was observed following strong antigen-specific IFN-γ and IL-2 responses by splenocytes from mice immunized with rHBcAg, but not rHBsAg, in both primary and secondary immunization protocols. Production of the Th2-type cytokine IL-5 was not inhibited similarly, nor did it increase at times when Th1-type cytokine production was inhibited. This lack of increase in IL-5 production is consistent with the observed absence of a switch in the isotype of the anti-HBcAg response from IgG2a to IgG1. The depressed IFN-γ response was restored upon depletion of CD25⁺ cells prior to rHBcAg restimulation in vitro, demonstrating that regulatory T cells mediated an active inhibition of this response.

The general pattern of this regulatory T cell-mediated transient inhibition in IFN-γ production was similar in both the primary and secondary rHBcAg-specific responses, with the major difference being the time of onset of the inhibition. Despite the earlier appearance of inhibition in the secondary response, it took the same number of days (14 days) for the depressed IFN-γ production in both the primary and secondary responses to recover. Another difference between these two responses was that the number of IFN-γ-secreting cells detected throughout the tested time-points remained higher, at an average of 44%, in the secondary response. However, the degree of inhibition in IFN-γ production was similar for both primary and secondary responses. The more rapid inhibition in IFN-γ production in secondary compared to primary rHBcAg responses seems likely to be induced by the more rapid, stronger secondary response.

When the production of IL-2 was assayed in the same secondary spleen cell populations a similar pattern of transient inhibition was detected, in that the number of days it took for the inhibition of IL-2 and IFN-γ production to appear and disappear (11 and 14 days, respectively) was the same with both cytokine responses. In general, the numbers of IL-2-secreting cells were greater than the numbers of IFN-γ-secreting cells following rHBcAg stimulation and IL-2 production was decreased to a much greater extent (an average of 90% relative to 69%) than IFN-γ production. The increased percentage of inhibition in the IL-2 response combined with the greater number of cells secreting the cytokine suggests that there may be a greater suppressive effect on IL-2 compared to IFN-γ production. Additionally, the fact that the rHBcAg-specific production of both of these cytokines follows the same kinetic pattern suggests that IFN-γ production is dependent on the presence of IL-2. This finding, that the number of cells producing IL-2 declined in parallel with the loss of IFN-γ production, has been described in other models of chronic antigen exposure including mice chronically infected with Mycobacterium avium²¹ and murine acquired immunodeficiency syndrome.²²

The data presented here also demonstrate that, unlike rHBcAg, Th1-type responses to rHBsAg another particulate HBV antigen, were not subject to inhibition by the same mechanism of regulation under the same experimental conditions. When mice were immunized with rHBsAg instead of rHBcAg, there was no apparent inhibition in Th1-type cytokine production. It is possible that the induction of rHBsAg-specific regulatory mechanism may take more time to develop than was assayed in this study. It is also possible that the activation of regulatory T cells and their inhibition of Th1-type cytokine production in response to rHBcAg is due to the strong immunogenicity of this antigen. Further studies will be needed to determine if this activation of regulatory mediated inhibition described in this study is a function of a certain type of antigen, such as nucleocapsid antigens, or indeed is unique to rHBcAg.

The transient inhibition in cytokine production described here appears to be restricted to predominantly Th1-type cytokine responses. This is supported by data demonstrating that rHBcAg-specific IL-5, a Th2-type cytokine, production remained constant throughout the response. CD8⁺ T cells do not contribute significantly to the IFN-γ response stimulated by s.c. immunization with rHBcAg in IFA,¹⁹ so the present data do not address the question of whether CD8⁺ T cells exhibit a similar delay in their IFN-γ production. However, a recent study demonstrated that adoptively transferred HBcAg epitope-specific CD8⁺ T cells isolated from the liver of HBV transgenic mice displayed ‘oscillating effector functions’ characterized by a transient decrease in IFN-γ production that coincided with an increase in cytolytic activity.²³ It is as yet unclear if the CD4⁺ T cell responses to rHBcAg also show oscillating effector functions.

Some of the potential mechanisms for the inhibition seen in IL-2 and IFN-γ production during the CD4⁺ T cell response to rHBcAg can be ruled out with the findings presented here. Removal of available rHBcAg in vitro within the ELISPOT assay by the presence of rHBcAg-specific antibody is unlikely as a cause of the inhibition, because antibody ELISPOT data demonstrated that the number of antibody-secreting cells was low when the IFN-γ and IL-2 responses were depressed and high when production of these cytokines were no longer inhibited. The fact that the inhibited cytokine responses seen in 24 hr ELISPOT assays were reversible by 48 hr assays and by depletion of CD25⁺ cells indicates that the 24 hr depression is not due to activation-induced cell death or anergy in the effector cells. Further evidence against a role of activation induced cell death or anergy in the observed inhibition in cytokine production is the transient nature of the inhibition following in vivo challenge. Additionally, the transient inhibition in IFN-γ and IL-2 production does not appear to be due to a switch from a Th1- to a Th2-type rHBcAg-specific response. The number of IL-5-secreting cells is not increased at time-points when IFN-γ- and IL-2-secreting cells are inhibited. Further evidence against a switch to Th2-type responses come from the observation that the dominant isotype of the anti-HBcAg response remained IgG2a, the isotype promoted in the mouse mainly by IFN-γ^24–26 throughout the response, and did not show an increase in IgG1 anti-HBcAg antibody, the isotype promoted predominantly by Th2-type cytokines,^24,27,28 even at times when IFN-γ production was depressed.

Another potential mechanism might involve the chronic exposure to the rHBcAg allowed by immunization with IFA, which has been shown to release antigen slowly over time^29–31 and prolong antigen presentation for months.³² Other studies have reported decreases in CD4⁺ T cell cytokine production under conditions of repeated antigen stimulation.^33,34 These decreases were transient in nature and were detectable in T cell lines and clones after in vitro^33,34 and in vivo³⁴ antigen restimulation. Furthermore, when effector cells were stimulated repeatedly with antigen, they showed diminished ability to provide protection against influenza infection.³⁴ The decreased effector functions that were described in these studies had similar effects on Th1- and Th2-type cytokine production. In the present study, similar transient inhibitory effects were seen in CD4⁺ T cells from mice chronically exposed to rHBcAg antigen. In contrast, the inhibition described in our study seemed restricted to Th1-type cytokine production and had no effect on Th2-type rHBcAg responses.

The most likely mechanism for the observed depression in IFN-γ and IL-2 production during the rHBcAg response supported by the current study is via the activation of an active regulatory process. More specifically, the presence of a regulatory cell appears to be required for the inhibition in Th1-type (IFN-γ and IL-2) cytokine production described in the present study. It has been demonstrated that CD4⁺CD25⁺ regulatory T cells possess the ability to actively regulate immune responses against foreign antigens and viruses^35–38 including HBV.^39,40 Data from the present study, demonstrating that when CD25⁺ cells were depleted prior to restimulation in vitro with rHBcAg CD25^– cells were no longer depressed relative to unseparated populations in their ability to produce IFN-γ, support this hypothesis. This finding is consistent with other in vivo studies demonstrating that suppression by regulatory T cells is reversible when suppressor cells are depleted from suppressed effector cells, allowing these cells to regain effector function.^41,42 It has been demonstrated that CD4⁺CD25⁺ regulatory T cells can suppress T cell activation in vitro by inhibiting IL-2 gene transcription in the CD4⁺CD25^– responder T cells⁴³. Data demonstrating that IL-2 production in addition to IFN-γ production was inhibited in our model supports this type of suppressive mechanism by regulatory T cells. The reversal of the inhibition in IFN-γ production by 48 hr in culture with antigen may be due to inactivation of regulatory T cells during the culture. Alternatively, it may be more likely that an increased time of antigen restimulation allows accumulation of enough IL-2 in the suppressed cultures to overcome the inhibition of IFN-γ. Further studies are needed to explore the reasons for the apparent transient nature of the CD4⁺CD25⁺ regulatory T cell-mediated suppression of CD4⁺ T cell IFN-γ and IL-2 production in rHBcAg immune mice.

Although there is no direct evidence in the present study for inhibition of HBcAg-specific Th1-type cytokine production occurring in vivo, it seems likely that this inhibition is controlled by in vivo events, given its transient appearance in spleen cells isolated at defined times after both primary and secondary immunizations with rHBcAg. The recovery of cytokine secretion in the suppressed cell populations detected after 48 hr of in vitro antigen restimulation, however, may be detectable in vitro only when the architecture of the spleen that limits cell interactions in vivo is broken down in the single cell suspensions of the ELISPOT assay. The recovery of this response may depend upon a small number of unsuppressed T cells in culture that produce enough IL-2 to overcome the CD25⁺ cell-mediated suppression of CD4⁺ T cells. In fact, without this reversal of suppression, the inhibition in vivo may be greater than the in vitro suppression described in this study.

Our study, along with others,^23,33,34 suggests that under conditions of persistent or long-term antigen exposure, as in chronic viral infections and exposure to antigen emulsified in Freund's adjuvant, T cells with impaired effector functions may develop. Decreased production of the Th1-type cytokines IFN-γ and IL-2, which is involved principally in cell-mediated immunity, could contribute to an inefficient immune response resulting in chronic infection. In the context of HBV, a HBcAg-specific IFN-γ CD4⁺ T cell response has been demonstrated to be important for viral elimination.^44–47 This current study points to the presence of a regulatory mechanism targeted towards HBcAg-specific Th1-type cytokine responses during antigen exposure. An insufficient Th1-type immune response against HBV, specifically the response directed against HBcAg, at a critical time in infection could facilitate the establishment of the chronic disease due to viral persistence. Further studies are necessary to explore the mechanism of action of the regulatory cells activated in the HBcAg-specific CD4⁺ T cell response.

Abbreviations

ELISPOT

enzyme-linked immunospot assay

ELISA

enzyme-linked immunosorbent assay

FCS

fetal calf serum

HBcAg

hepatitis B core antigen

HBV

hepatitis B virus

rHBcAg

recombinant hepatitis B core antigen

rHBsAg

recombinant hepatitis B surface antigen

IFA

incomplete Freund's adjuvant

IFN-γ

interferon gamma

IL

interleukin

SFC

spot-forming cell

Th

helper T cell