Abstract
We have investigated cutaneous purified protein derivative-induced delayed-type hypersensitivity (DTH) responses in healthy volunteers to determine features associated with both the generation and resolution of the reaction. The clinical peak of the response occurred at day 3; however, T cell numbers were maximal on day 7. There was a preferential increase of CD4+CD45RO+ T cells on day 7, which was largely due to proliferation, since a mean of 19% was in cycle. The proliferation of this subset was associated with the presence of IL-15, which was expressed as early as 12 h, and IL-2, which showed peak expression at 7 days. By day 14, there was a significant decrease in both the mean T cell number/unit area and IL-2 and IL-15 expression in perivascular infiltrates. Maximal CD95 (Fas/Apo-1) ligand and TNF-α expression were observed at 7 days and were associated with the presence of 1.83% (range 0.81–2.48%) apoptotic T cells. At 14 days, CD95 ligand and TNF-α expression were reduced significantly, and the presence of 2.5% (range 1.5–3.75%) of apoptotic T cells at this time was probably due to cytokine deprivation, associated with decreased Bcl-2 relative to Bax expression. The induction and resolution of the Mantoux reaction may depend on the expression of cytokines, such as IL-2 and IL-15, which regulate both proliferation and apoptosis in T cells. Failure to control either of these phases of the Mantoux reaction may contribute to the chronicity of inflammatory responses in certain cutaneous diseases.
Tcell-mediated inflammatory reactions exhibit an infiltration and expansion of activated CD4 and CD8 T cell populations (1). Resolution is associated with a return to normal in terms of absolute cell numbers and relative proportions of these subsets within the tissue involved (2). It is now recognized that cell numbers are controlled in a wide array of biologic systems by the process of cell suicide or apoptosis (3, 4, 5, 6); however, the contribution of this process to the resolution of cutaneous inflammatory responses has not been studied. Such investigations are of particular importance for understanding persistent inflammatory conditions such as atopic eczema, in which disease chronicity may be perpetuated by the inability to terminate the ongoing response.
The Mantoux reaction is a well-recognized delayed-type hypersensitivity (DTH)3 reaction that peaks clinically at 48 to 72 h and resolves within 14 days (7, 8). This human model of cutaneous inflammation enables the study of the kinetics of the local immune response from onset to resolution. Previous studies have shown that within 12 h of the intradermal injection of PPD, interdigitating dendritic cells appear around dermal blood vessels, and that by 24 to 48 h, large numbers of infiltrating activated macrophages are present (9). T cells begin to accumulate perivascularly within 12 h of challenge and by 48 h, >60% of the mononuclear cell infiltrate is comprised of primed (CD45RO+) T cells (9, 10, 11). The majority of studies of cutaneous DTH reactions in humans have not followed the reaction beyond 96 h, and the process by which activated T cells are removed and inflammation is resolved has not been characterized.
A persistent cutaneous inflammatory response may result from both continued recruitment of leukocytes into the involved tissue, and/or a lack of clearance of the infiltrating cells. Lymphocyte chemotactic factors such as IL-1 and IL-8 have been identified in epidermis overlying cutaneous PPD-induced DTH reactions and may be responsible in part for the infiltration of cells (12, 13, 14). In addition, recent studies have shown that the clearance of activated T lymphocytes by apoptosis can also be prevented by cytokines, in particular those such as IL-2, IL-4, IL-7, and IL-15, which share the IL-2R γ-chain as part of their receptor complexes (15, 16, 17, 18, 19). It is of interest that IL-15 also serves as a chemotactic factor for activated T cells (20, 21, 22). These cytokines were shown to prevent the down-regulation of intracellular regulatory molecules such as Bcl-2 and Bcl-xL, which inhibit apoptosis, but do not alter the expression of Bax and Bcl-xS, which induce death in activated T cells (16, 23, 24, 25, 26, 27). Apoptosis can also be induced by the ligation of CD95 on the surface of activated T cells by its ligand and also by the binding of TNF-α to its receptor (6). The regulation of expression of antiapoptotic cytokines or apoptosis-inducing CD95 ligand and TNF-α during the initiation and resolution of a cutaneous DTH response has not been characterized.
In this study, we have investigated the kinetics of T lymphocyte infiltration, proliferation, and apoptosis during a Mantoux reaction in relation to cytokine and apoptosis regulatory protein expression to characterize features associated with resolution of inflammation.
Materials and Methods
Patients and control samples
Mantoux tests were performed on the volar surface of the nondominant forearm of 20 healthy volunteers previously immunized with Bacille Calmette-Guerin (15 males, age range 23–59, median 30 yr). Testing was first with 0.1 ml of a 1/10,000 solution of tuberculin PPD (Evans Medical, Leatherhead, U.K.), and then, if negative at 48 to 72 h, with 1/1000 strengths. Erythema and induration were measured at 72 h and on the day of the biopsy, as follows. Erythema was scored using a DermaSpectrometer (Cortex Technology, Hadsund, Denmark), a handheld system designed for measuring the erythema index (EI) of the skin by measuring light absorption coefficients (28). Erythema indices obtained were grouped and scored as 1 = EI < 5; 2 = EI 5–10; 3 = EI 10–15; and 4 = EI > 15. Induration was scored as: 1 = none detected; 2 = just palpable; 3 = easily palpable; 4 = marked; and 5 = very marked. The maximum diameter was measured at 72 h in millimeters and scored as 1, 4–9 mm, or 2, >10 mm. The sum of the erythema, induration, and diameter scores was then used to give each subject an overall score, both at the time of biopsy and at 72 h. We examined initiation and resolution of the reaction by biopsying early and late time points after PPD injection. Each volunteer had one 4-mm punch biopsy taken from the intradermal injection site at either 12 h, 72 h, 7 days, or 14 days after the procedure (five subjects per time point). Ethics committee approval and subjects’ informed consent were obtained before performing the biopsies. Normal skin was obtained from surgical specimens in five patients.
In addition, one of the senior authors (L.W.P.) consented to intradermal injection with 1/10,000 PPD at six different sites on the volar aspect of the forearm, and had 4-mm punch biopsies performed at 6, 12, 24, 72 h, 7 days, and 14 days to ensure that the kinetics of the reaction could be observed in a single individual.
All biopsies were mounted in Cryo-M-Bed (Bright’s instrument Company, Huntingdon, Cambs, U.K.) and snap frozen in isopentane cooled in a bath of liquid nitrogen. Samples were stored in liquid nitrogen until sectioned. Cryostat sections (6 mm) were cut onto poly(l-lysine)-coated slides, air dried for 2 h, and either used immediately or stored wrapped in cling film at −20°C before immunohistologic staining.
Immunohistology
The characteristics of the mAbs/polyclonal antiserum used in this study are documented elsewhere (29) and in Table I. All Abs were mouse anti-human monoclonals, unless otherwise stated.
Ab . | Subclass . | Code . | Source . | Specificity . |
---|---|---|---|---|
Tmix (CD2, CD7, CD8, CD4, CD3) | IgG1 and IgG2a | RFTmix | RFH | T cells |
CD45RO | IgG2a | UCHL-1 | UCHb | Primed T cells |
CD5 | IgM | RFT1 | RFH | T cells and some B cells |
CD8 | IgM | RFT8 | RFH | Class I MHC restricted T cells |
CD4 | IgG1κ | MO716 | Dako (High Wycombe, U.K.) | Class II MHC restricted T cells |
CD4 | IgG2a | MHCD 0400 | Caltag (San Francisco, CA) | Class II MHC restricted T cells |
CD3 | IgG | UCHT1 | UCH | T cells |
Bcl-2 | IgG1 | MO887 | Dako | Anti-apoptotic protein |
Bax | Rabbit anti-human polyclonal IgG | N-20 sc-493 | Santa Cruz Biotechnology (Santa Cruz, CA) | Pro-apoptotic protein |
Ki67 | IgG1κ | MO722 | Dako | Proliferating cells (outside G0) |
Fas-ligand | IgG1κ | NOK-1 | PharMingen (San Diego, CA)c | Membrane bound and soluble Fas-L; proapoptotic TNF family protein that binds to Fas |
Anti-IL-15 | IgG1 | M112 | Genzyme Diagnostics (Cambridge, MA) | T cell chemotaxis and proliferation |
Anti-IL-2 | IgG1 | MCA 745 | Serotec, Oxford, U.K. | T cell activation and proliferation |
Anti-TNF-α | IgG1 | 80-3399-01 | Genzyme Diagnostics | Endothelial cell and phagocyte activation, pro-apoptotic cytokine |
Anti-IL-6 | IgG1 | 1618-01 | Genzyme Diagnostics | Ubiquitous pro-inflammatory cytokine, T cell activation |
Ab . | Subclass . | Code . | Source . | Specificity . |
---|---|---|---|---|
Tmix (CD2, CD7, CD8, CD4, CD3) | IgG1 and IgG2a | RFTmix | RFH | T cells |
CD45RO | IgG2a | UCHL-1 | UCHb | Primed T cells |
CD5 | IgM | RFT1 | RFH | T cells and some B cells |
CD8 | IgM | RFT8 | RFH | Class I MHC restricted T cells |
CD4 | IgG1κ | MO716 | Dako (High Wycombe, U.K.) | Class II MHC restricted T cells |
CD4 | IgG2a | MHCD 0400 | Caltag (San Francisco, CA) | Class II MHC restricted T cells |
CD3 | IgG | UCHT1 | UCH | T cells |
Bcl-2 | IgG1 | MO887 | Dako | Anti-apoptotic protein |
Bax | Rabbit anti-human polyclonal IgG | N-20 sc-493 | Santa Cruz Biotechnology (Santa Cruz, CA) | Pro-apoptotic protein |
Ki67 | IgG1κ | MO722 | Dako | Proliferating cells (outside G0) |
Fas-ligand | IgG1κ | NOK-1 | PharMingen (San Diego, CA)c | Membrane bound and soluble Fas-L; proapoptotic TNF family protein that binds to Fas |
Anti-IL-15 | IgG1 | M112 | Genzyme Diagnostics (Cambridge, MA) | T cell chemotaxis and proliferation |
Anti-IL-2 | IgG1 | MCA 745 | Serotec, Oxford, U.K. | T cell activation and proliferation |
Anti-TNF-α | IgG1 | 80-3399-01 | Genzyme Diagnostics | Endothelial cell and phagocyte activation, pro-apoptotic cytokine |
Anti-IL-6 | IgG1 | 1618-01 | Genzyme Diagnostics | Ubiquitous pro-inflammatory cytokine, T cell activation |
All Abs were mouse anti-human monoclonals unless otherwise stated.
Kindly provided by Prof. P. C. L. Beverley (University College and Middlesex School of Medicine, London, U.K.) and characterized in Reference 29.
Kindly provided by Prof. H. Yagita (Juntendo University School of Medicine, Tokyo, Japan).
The study used indirect immunoperoxidase, immunofluorescence, biotin/streptavidin alkaline phosphatase, and TUNEL methods.
Immunoperoxidase staining
An indirect immunoperoxidase technique was used to detect T cells, CD45RO+ and Ki67+ cell numbers, and distribution. Following a 10-min incubation with normal rabbit serum, skin sections were incubated with a pan anti-T cell IgG mAb mix (T mix), CD45RO or Ki67 diluted in PBS at pretitrated optimal concentrations for 45 min at room temperature. The slides were then washed in PBS, and a secondary peroxidase-conjugated goat anti-mouse IgG Ab (P161; IgG, Dako, High Wycombe, Bucks, U.K.) diluted 1/100 in PBS and containing 4% normal human serum was then applied. After an additional 45 min, the slides were again washed in PBS and the reaction was developed using diaminobenzidine. Sections were counterstained in hematoxylin and mounted in dibutyl polystyrene xylene (BDH Laboratory Supplies, Poole, U.K.) Three control preparations were used. Sections of normal human tonsil, in which the distribution and pattern of staining could be tested against tissue architecture, were used as positive controls in each experiment. In addition, control incubations to detect background staining were performed on sections of each skin sample, omitting the primary Ab. Third, isotype specificity was confirmed by comparison with staining with irrelevant mAbs of the same isotype as the mAbs used on tonsil sections.
Immunofluorescence
To determine CD4:CD8 ratios and proportions of T cell subsets expressing CD45RO and Ki67, sections were incubated for 45 min in a moist chamber with appropriate combinations of mAbs diluted in PBS. After rinsing in PBS, Ig isotype-specific FITC- or TRITC-conjugated affinity-purified goat anti-mouse (Southern Biotechnology, Birmingham, AL) second-layer Abs were applied at pretitrated optimal concentrations, and slides were incubated for 40 min. Slides were then rinsed in PBS fixed in 4% paraformaldehyde and mounted in Citifluor (AF1; Citifluor Products, Canterbury, U.K.). Using the above indirect dual immunofluorescence technique, sections were incubated with CD5+CD8 (an IgM Ab mix used to stain T cells) and either Bcl-2 or Bax, and the percentage of T cells expressing Bcl-2 and Bax was determined. Controls were performed as described, but using the fluorochrome-conjugated second layers alone. Sections were fixed as described above.
Quantification of immunohistology
For immunoperoxidase studies, the number and distribution of positive cells were quantified in each section using an image analysis system (Seescan Imaging, Cambridge, U.K.; magnification ×320) per circular frame area centered on the largest dermal perivascular inflammatory cell infiltrates, five times per section. For the purposes of statistical analysis and visual display of the data, results were scaled to a frame area of 1 unit area (UA).
For immunofluorescence studies, the distribution and percentages of T cells were estimated in each section using a Zeiss fluorescence microscope (×400 magnification) in the five largest dermal perivascular inflammatory cell infiltrates present in the sections.
Biotin/Streptavidin
To identify the distribution of IL-2 and IL-15, IL-6, TNF-α, and CD95 ligand, freshly cut cryostat sections were air dried for 2 h, ringed with polysiloxane, and fixed in precooled methanol:acetone 1:1 at −20°C for 10 min. After rinsing in PBS at room temperature, sections were incubated overnight with 100 ml of the appropriately diluted primary Ab in PBS +0.1% BSA. Sections were washed in Tris-buffered saline (TBS) at pH 7.6 and then incubated in a moist covered chamber with 50 ml of affinity-purified horse anti-mouse biotinylated second layer (IgG; Vector Laboratories, Peterborough, U.K.) diluted 1/100 in PBS-BSA for 1 h at room temperature. After rinsing in fresh TBS, sections were then incubated for 1 h with 50 ml of streptavidin-alkaline phosphatase-conjugated third layer (Vector Laboratories) diluted 1/100 in PBS-BSA at room temperature in a moist covered chamber. Sections were again rinsed in fresh TBS, and the reaction was developed by 15-min application of filtered substrate solution (0.005 g naphthol ASBI phosphate, 10 ml Tris-HCl (pH 8.2), 200 μl dimethylformamide, 0.01 g Fast Red (TR), and 10 drops Levamisole added last). Sections were then washed in tap water and counterstained with Mayer’s hematoxylin before mounting in PBS glycerol (9:1). Controls were performed on skin sections as above using the streptavidin/biotin second and third layers alone. Isotype specificity was confirmed by comparison with staining with an irrelevant IgG1 mAb on skin sections. The proportion of perivascular cells with cytoplasmic or membrane staining was estimated using an image analysis system (Seescan Imaging; magnification ×320) in the five largest dermal perivascular inflammatory cell infiltrates in each section.
Identification of apoptotic T cells
The presence of apoptotic T cells within perivascular infiltrates in PPD reactions was confirmed using a combination of indirect immunofluorescence and the TUNEL (terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling) technique (30). Sections were stained as above, then fixed with 4% paraformaldehyde solution for 20 min at room temperature, and washed in PBS for 30 min. Permeabilization was performed by incubating with 0.1% Triton X-100 (Rohm & Haas, Philadelphia, PA), 0.1% sodium citrate for 2 min on ice. After rinsing in PBS, sections were incubated with 50 μl of TUNEL reaction mixture (in situ cell death detection kit, fluorescein; catalogue number 1684795, Boehringer Mannheim, Indianapolis, IN) for 60 min at 37°C in the dark. Sections were rinsed in PBS and mounted in Citifluor. The proportion of TUNEL-positive T cells was estimated in each section using a Zeiss fluorescence microscope in the five largest dermal perivascular inflammatory cell infiltrates. In each experiment, sections of normal human tonsil were used as positive controls, and negative controls were performed using Label solution (without terminal transferase) instead of TUNEL reaction mixture.
Statistical analysis
Measurements were taken from five perivascular infiltrates in each subject, and mean values and SDs were calculated. A minimum of three subjects was investigated at each time point. Using the ANOVA method, differences between the values at the five time points were tested for significance, including time and subject as factors in the analysis. For the purpose of visually displaying the data, the mean values for the different subjects were used to calculate the SD for each time point.
Results
Response to PPD of study subjects
All subjects responded to PPD with maximal reactions (as defined by erythema and induration scores) consistently occurring between 48 and 72 h. Of the lesions biopsied at 12 h, two showed a marked response and three a minimal response at that time. When reviewed at 48 to 72 h, all five subjects showed a positive test, even when inflammation induced by the biopsy itself was taken into account. Erythema and induration were scored at 72 h and at the time of biopsy. The maximum diameter of the reaction was scored at 72 h. The sum of these scores for each individual was then used to calculate a mean score and SD for each time point (five subjects per time point) and is presented in Figure 1. There was no significant difference in the mean scores obtained at 72 h in each of the four groups of subjects who received intradermal PPD (ANOVA, p = 0.12), confirming that similar responses were obtained regardless of time of biopsy (Fig. 1,A). The mean scores at the time of biopsy paralleled erythema measurements obtained with the DermaSpectrometer, but also took into account induration, which may better reflect the degree of inflammation occurring in the lesions. These scores were significantly lower at 14 days than at 3 days (ANOVA, p = 0.02), confirming that lesions biopsied at that time point were resolving (Fig. 1 B).
Characteristics of infiltrating T cells
Normal skin contained small numbers of T cells (4.5 ± 1.9/UA). After intradermal PPD, T cells accumulated perivascularly within the dermis and numbers rose significantly by 12 h (15 ± 6.9/UA), and further by 72 h (39.6 ± 7.8/UA); there was also a further increase from this figure up to day 7 (54.2 ± 4.7/UA). T cell numbers then fell significantly by day 14, although they remained in excess of numbers in normal skin (23.7 ± 6.3/UA) (Fig. 2 A). Overall analysis revealed that there were significant differences between T cell numbers at the five time points studied (p < 0.0001). Furthermore, T cell numbers at each time point were significantly different from those at the preceding and subsequent time points. Although a majority of infiltrating cells were present within perivascular areas, smaller numbers of T cells were seen infiltrating the interstitium and epidermis. In these areas, T cell numbers were maximal at 3 days, and subsequently declined up to day 14 after intradermal PPD (data not shown). Thus, although the clinical and overall histologic responses to intradermal PPD showed different kinetics, peaking at 3 and 7 days, respectively, by day 14 both were resolving.
In normal skin, CD4+ cells predominated, but numbers were too small to calculate meaningful ratios. Throughout the course of the Mantoux reaction, the number of CD4+ T cells exceeded the CD8+ cells within the perivascular infiltrates. The ratio of CD4 to CD8 cells was lowest at 12 h (2.23 ± 0.85) and reflected proportions of CD4 and CD8 cells found within the peripheral circulation. Thereafter, the proportion of CD4 cells increased (Fig. 2,B and Fig. 3 A). This suggests that either active recruitment or proliferation of CD4 cells occurred. Dual immunofluorescence was performed on one representative section per time point with CD3 and CD8. No CD3−CD8+ cells were identified in any of the sections examined, indicating that the CD8+ cells identified were T rather than NK cells. In addition, no CD4+CD8+ double-positive T cells were observed.
There were small numbers of CD45RO+ cells in normal skin and at 12 h after intradermal PPD. The number of CD45RO-positive cells within the perivascular infiltrates increased between 12 and 72 h (6 ± 1.8 to 17.1 ± 6.3/UA), and again between day 7 and day 14 (19.8 ± 4 to 27 ± 6.6/UA). Overall analysis revealed that there were significant differences between values at 12 and 72 h, and between those at 7 and 14 days (ANOVA, p < 0.0001).
Double immunofluorescence studies showed that at 12 h, 68.7 ± 4.6% of CD8 cells within perivascular infiltrates were CD45RO+. This figure rose to 83.4 ± 0.6% at 7 days after intradermal PPD (Fig. 4). This suggests that even from early time points during the reaction, CD8 cells recruited into the lesions were already primed.
In contrast, the proportion of primed (CD45RO+) CD4 cells was lowest at 12 h (44.9 ± 22.2%), increased significantly by 72 h (74.4 ± 7.9%), and thereafter continued to rise gradually up to 14 days (86.9 ± 4.2%) (Fig. 4). Overall ANOVA, p < 0.0001.
T cell proliferation in Mantoux reactions
To determine whether in situ proliferation could account for increasing cell numbers within Mantoux reactions, we measured numbers of Ki67+ cells within perivascular infiltrates. Numbers of Ki67+ cells rose significantly from 0.2 ± 0.2/UA at 12 h to 7.3 ± 0.9/UA at 7 days, and then fell to 1.5 ± 1.4 at 14 days (overall ANOVA, p < 0.0001) (Fig. 5,A). To investigate which cells were induced to proliferate, dual immunofluorescence studies were performed (Fig. 5,B). At 12 h, very few T cells expressed Ki67. Seventy-two hours after intradermal PPD, 5.8 ± 3.3% of T cells were Ki67+, and by 7 days, the percentage of proliferating T cells increased significantly to 18.8 ± 3.7%. However, at 14 days, this percentage fell to 4.2 ± 2.3% (overall ANOVA, p < 0.0001). The proliferating cells were CD4+CD45RO+ (Fig. 5, B and C, and Fig. 3 B). At 72 h, 3.9 ± 1.5% of CD4+ cells were actively proliferating; by day 7, this percentage had increased significantly to 19.3 ± 6.6%; and by day 14, it had fallen to 2.63 ± 2.65% (overall ANOVA, p < 0.0001). A similar trend was observed in the CD45RO+ subset. No Ki67+CD8 cells were identified in any of the sections examined. These results suggest that the increase in CD4+CD45RO+ T cells during the course of the Mantoux reaction was most likely to be due to the induction of proliferation within this subset.
Resolution of the Mantoux reaction markers of apoptosis
To investigate whether the reduction in T cell numbers and resolution of the DTH reaction occurred as a result of T cell apoptosis, we used a combination of indirect immunofluorescence and TUNEL methodologies. No TUNEL-positive T cells were seen within perivascular areas in normal skin. At 12 h after intradermal PPD, 0.05 ± 0.12% (range 0–0.27%) of perivascular T cells were TUNEL positive. Seventy-two hours after challenge, the percentage of TUNEL-positive perivascular T cells had increased (0.47 ± 0.44%), but not significantly. However, there was a significant increase at day 7 after intradermal PPD (compared with normal skin and the 12- and 72-h time points) to 1.83 ± 0.74% (range 0.81–2.48%), and a further rise to 2.5 ± 0.93% (range 1.5–3.75%) at day 14. This percentage increase at 7 and 14 days remained significant when the variation in T cell numbers between time points was taken into account (overall ANOVA, p < 0.0001). TUNEL-positive T cells were present in all of the 7- and 14-day specimens examined and were located predominantly at the periphery of perivascular infiltrates (Fig. 3 F). Additional investigations revealed that a majority of TUNEL+ cells at these time points were located within macrophages (data not shown), suggesting that the numbers of apoptotic T cells detected represented an underestimate of the total amount of apoptosis occurring.
Previous studies have shown that the propensity for T cells to die by apoptosis due to cytokine deprivation correlates with a reduction in their Bcl-2 expression relative to Bax (16, 23, 25, 26, 27, 31). We thus investigated the percentages of T cells expressing Bcl-2 and Bax within perivascular infiltrates to assess the extent to which the apoptosis observed could be due to lack of cytokines. Dual immunofluorescence studies revealed that the proportions of T cells expressing Bcl-2 perivascularly rose significantly between 12 h after intradermal PPD (43.4 ± 5.8%) and 3 days (76.3 ± 7.5%). This percentage remained relatively stable (70.9 ± 6.1%) to day 7 and then fell significantly by day 14 (23.5 ± 3.3%; overall ANOVA, p < 0.0001) (Fig. 6). In contrast, the proportions of T cells expressing Bax within these areas remained constant at >98% at all of the time points studied, and although T leukocyte numbers were much smaller, high Bax expression was also found in normal skin (data not shown).
The high level of Bcl-2 expression on day 7 suggested that lack of this molecule, and by inference cytokines, was not responsible for the T cell apoptosis observed at this time. We therefore investigated the expression of CD95 ligand in perivascular infiltrates during the Mantoux reaction to determine whether this alternative pathway to apoptosis may contribute to death. Keratinocytes (KC) in normal skin expressed weak to moderate CD95 ligand, while interstitial and perivascular cells showed little or no expression (not shown). At 72 h after intradermal PPD, KC CD95 ligand expression was up-regulated, and 19.9 ± 12.7% of perivascular cells with moderate/strong cytoplasmic as well as membranous CD95 ligand expression was observed (Fig. 6). Cytoplasmic staining was present predominantly in large macrophage-like cells. On day 7, CD95 expression was markedly up-regulated in perivascular areas (84.3 ± 9.4%; Fig. 6). A majority of perivascular cells showed strong cytoplasmic and/or surface CD95 ligand expression (Figs. 6 and 3,G). After 14 days, however, there was a marked reduction in both proportions of cells and the intensity with which they expressed CD95 ligand in perivascular infiltrates (Figs. 6 and 3 H). These results suggest that the apoptosis that was observed on day 7 coincided with peak CD95 ligand expression, suggesting that CD95/CD95 ligand interaction may account for at least a proportion of death observed at this time. However, the significant down-regulation of CD95 ligand expression at day 14 suggests that it is unlikely that CD95/CD95 ligand interactions were the main trigger for apoptosis at this time.
We also investigated expression of TNF-α since this cytokine has also been implicated in the induction of apoptosis (6). We found that peak TNF-α expression occurred between days 3 and 7, and that although there was intersubject variability, at 14 days, proportions of cells in perivascular infiltrates with strong cytoplasmic staining were reduced (summarized in Table II). Thus, although it is possible that this cytokine may contribute to the death observed at the peak of the response on day 7, this cytokine may not have a major role on the apoptosis observed on day 14, when its expression is reduced (Table II).
Cytokine . | Days After Intradermal PPD . | . | . | . | . | ||||
---|---|---|---|---|---|---|---|---|---|
. | 0 . | 0.5 . | 3 . | 7 . | 14 . | ||||
IL-15 | |||||||||
% perivascular expressiona | 1.9 | 28.4 | 77.6 | 84.6 | 7.9 | ||||
Range (%) | 0–5 | 15–50 | 70–85 | 75–90 | 5–10 | ||||
Intensity of stainingb | ++ | ++ | +++ | ++ | + | ||||
IL-2 | |||||||||
% perivascular expression | 0% | 0% | 59.5 | 65.7 | 3.5 | ||||
Range (%) | 50–70 | 60–75 | 0–5 | ||||||
Intensity of staining | − | − | ++ | +++ | ++ | ||||
IL-6 | |||||||||
% perivascular expression | 23.3 | 44.7 | 48.5 | 95.1 | 93.7 | ||||
Range (%) | 5–30 | 30–60 | 40–60 | 90–98 | 90–98 | ||||
Intensity of staining | +++ | ++ | ++ | +++ | +++ | ||||
TNF-α | |||||||||
% perivascular expression | 57.6 | 52.9 | 89 | 86.3 | 69.9 | ||||
Range (%) | 50–60 | 50–60 | 80–90 | 75–98 | 40–95 | ||||
Intensity of staining | +++ | ++ | +++ | +++ | + to ++ |
Cytokine . | Days After Intradermal PPD . | . | . | . | . | ||||
---|---|---|---|---|---|---|---|---|---|
. | 0 . | 0.5 . | 3 . | 7 . | 14 . | ||||
IL-15 | |||||||||
% perivascular expressiona | 1.9 | 28.4 | 77.6 | 84.6 | 7.9 | ||||
Range (%) | 0–5 | 15–50 | 70–85 | 75–90 | 5–10 | ||||
Intensity of stainingb | ++ | ++ | +++ | ++ | + | ||||
IL-2 | |||||||||
% perivascular expression | 0% | 0% | 59.5 | 65.7 | 3.5 | ||||
Range (%) | 50–70 | 60–75 | 0–5 | ||||||
Intensity of staining | − | − | ++ | +++ | ++ | ||||
IL-6 | |||||||||
% perivascular expression | 23.3 | 44.7 | 48.5 | 95.1 | 93.7 | ||||
Range (%) | 5–30 | 30–60 | 40–60 | 90–98 | 90–98 | ||||
Intensity of staining | +++ | ++ | ++ | +++ | +++ | ||||
TNF-α | |||||||||
% perivascular expression | 57.6 | 52.9 | 89 | 86.3 | 69.9 | ||||
Range (%) | 50–60 | 50–60 | 80–90 | 75–98 | 40–95 | ||||
Intensity of staining | +++ | ++ | +++ | +++ | + to ++ |
Both cells with membranous and cytoplasmic staining were taken into account when calculating the mean and range percentages of positive perivascular cells (in five perivascular infiltrates per section). A minimum of three subjects were investigated at each time point.
Intensity of staining was graded as follows: −, none; +, weak; ++, moderate; +++, strong.
The results obtained suggested that apoptosis at day 7 was not due to changes in Bcl-2 expression; however, the significant decreased Bcl-2 expression on day 14 was likely to contribute to apoptosis at this time. To further investigate whether apoptosis occurring after the peak of the Mantoux reaction was due to the decrease in Bcl-2 relative to Bax, we examined the kinetics of T cell accumulation together with expression of these molecules in six samples taken at different times from a single individual. This individual had a strongly positive response to PPD (>11 mm induration and >15 EI at 72 h), and showed slightly accelerated kinetics of the reaction. Nevertheless, the trends in T cell numbers and Bcl-2 and Bax expression were similar to the pooled data from different individuals at each time point. In this individual, T cell numbers were maximal 3 days after intradermal PPD, and then progressively declined to day 14. In contrast, Bcl-2 expression appeared maximal at day 1, and had declined significantly by 3 days, before the numbers of T cells were seen to be reduced (Fig. 7). Bax expression remained relatively constant throughout the reaction (data not shown). Thus, the fall in Bcl-2 preceded the fall in T cell numbers, suggesting an association between the down-regulation of this molecule and apoptosis at the later stages of the Mantoux reaction.
Cytokine studies
Withdrawal of cytokines such as IL-2 and IL-15 can induce T cell apoptosis by down-regulating their Bcl-2 expression relative to Bax (16). These cytokines, especially IL-15, are also involved in the induction of the immune response through promotion of T cell chemotaxis and proliferation (20, 21, 22, 32, 33, 34). We therefore investigated whether changes in IL-2 and IL-15 expression occurred throughout the course of the Mantoux reaction in three different subjects per time point. Although some variability occurred between individuals, the overall trends in IL-2 and IL-15 expression at different time points were the same. Normal skin showed no staining with IL-2 (Table II). At 12 h after intradermal PPD, occasional (<5%) dermal interstitial T cells expressed cytoplasmic IL-2, but no positive cells were seen in perivascular infiltrates or the epidermis. By 72 h, there was marked cytoplasmic expression in a majority (>50%) of interstitial T cells, and occasional positive cells were seen penetrating the epidermis (Fig. 3,C). Within perivascular infiltrates at 72 h, 50 to 60% of cells showed membrane staining, while a small percentage expressed cytoplasmic IL-2. Expression of IL-2 appeared maximal at 7 days, when both the numbers of T cells present and the extent of proliferation were at their highest levels. At that time point, >75% of interstitial cells and many perivascular cells expressed cytoplasmic IL-2 (Fig. 3,D), and the majority of the remaining perivascular T cells expressed membrane-bound IL-2. By day 14 after intradermal PPD, only occasional (<5%) interstitial and perivascular T cells expressed cytoplasmic or membrane-bound IL-2 (Fig. 3 E). This decrease in IL-2 at 14 days coincided with the decrease in Bcl-2 expression and cell numbers in the resolving Mantoux reactions.
In normal skin, epidermal KC showed moderate cytoplasmic staining with IL-15 (Fig. 8,A, and Table II). Within the dermis, only occasional cells with cytoplasmic staining were seen. Twelve hours after intradermal PPD, staining intensity was greater in epidermal KC (Fig. 8,B). Occasional strongly positive cells, with a dendritic morphology resembling Langerhans cells, were also present in this area. In the papillary dermis, intracytoplasmic IL-15 was present in numerous large, macrophage-like cells and occasional dendritic cells (DC), both within perivascular infiltrates in close proximity to lymphocytes, and in the interstitium (Fig. 8,B). In addition, up to 50% of perivascular T cells showed membrane staining with IL-15. At 72 h, although fewer IL-15-positive KC were present, greater numbers of strongly positive epidermal DC were seen. In the papillary and upper reticular dermis, the distribution of IL-15 was the same as 12 h, but the staining intensity and number of positive cells were greater (Fig. 8,C). In particular, >75% of perivascular lymphocytes expressed membrane-bound IL-15. At 7 days, when T cell proliferation and cell numbers were maximal, fewer dermal cells with intracytoplasmic staining were seen, but a majority (75–90%) of perivascular cells still expressed membrane-bound IL-15 (Fig. 8,D). Epidermal KC IL-15, however, was markedly reduced by day 7 and undetectable at day 14, although occasional positive DC remained. By day 14, only very occasional perivascular macrophage-like cells were seen and minimal or no membranous IL-15 was present on perivascular lymphocytes (Fig. 8 E). These results suggest that although IL-15 was present earlier than IL-2 during a Mantoux reaction, high levels of expression of both cytokines at day 7 coincided with increased cell numbers and proliferation. Conversely, the decrease in cell numbers on day 14 was associated with a marked reduction of expression of both IL-15 and IL-2. Thus, high levels of expression of these IL-2R γ-chain signaling cytokines were present at the height of the DTH response, whereas resolution was associated with markedly reduced levels, a situation favoring T cell apoptosis due to cytokine deprivation.
To determine whether this reduction in cytokine expression was specific to IL-2 and IL-15, we also investigated expression of IL-6 (Table II). This cytokine has previously been identified in blister fluid from PPD-induced DTH reactions (11), and in addition to its T cell costimulatory effects (35), may be involved in up-regulation of the cutaneous lymphocyte Ag on CD4+CD45RO+ responsible for this subset’s ability to home to skin (36). Seven days after intradermal PPD, IL-6 expression was markedly up-regulated in epidermal KC, dermal DC, and perivascular areas (in which up to 98% of lymphocytes expressed cytoplasmic and/or membranous IL-6) compared with the 12- and 72-h time points (data not shown). However, in contrast to the marked reduction in IL-15 and IL-2 expression that occurred at day 14, membranous and cytoplasmic IL-6 expression remained high in perivascular lymphocytes at this time point (Fig. 8 F). Collectively, these observations suggest that changes in expression of IL-2 and IL-15 may play a role in the generation and resolution of the Mantoux reaction.
Discussion
The generation and resolution of a localized immune response are governed by the migration of leukocytes into the site of injury, the proliferation of cells in situ, and the removal of these cells after antigenic clearance (37, 38, 39). Abnormal regulation of any of these phenomena may lead to chronic inflammation. We have investigated the kinetics of a Mantoux reaction to clarify factors that influence the generation and resolution of the response to better understand why chronic inflammation in cutaneous diseases such as atopic eczema persists (40).
Following the intradermal injection of PPD, T cells begin to accumulate perivascularly within the dermis by 12 h, reflecting increased transendothelial migration rather than proliferation, since no Ki67+ T cells are present at this time. This increased migration has been attributed previously to the release of chemotactic factors such as IL-8 (12, 13, 14); however, other factors may also be involved. For example, naive and memory T cells show different recirculation pathways (41), and migration of skin-homing T cells is dependent on interactions between the lymphocyte homing receptors cutaneous lymphocyte Ag, VLA-4, and LFA-1, and their endothelial cell counter-receptors E-selectin, VCAM-1, and ICAM-1 (36, 42). Furthermore, highly differentiated CD45RO+CD45RBlowCD4+ T cells migrate preferentially (43, 44), and their rate of migration can be increased by endothelial cell activation, IL-15, and chemokines such as RANTES (44, 45). We have demonstrated that there was up-regulation of IL-15 as early as 12 h after PPD challenge. In addition to its effects on the rates of T cell transmigration, this cytokine has been shown to be an important chemoattractant for T cells (20, 21, 22). IL-15 may therefore also play a role in the early accumulation of T cells after PPD challenge.
There was an initial recruitment of both CD4 and CD8 cells into the lesions, followed later by a selective increase in CD4+ T cell numbers. The increase of CD4+CD45RO+ T cells at the later time points is probably due to proliferation rather than migration, since 19% of this subset expressed Ki67 reactivity at 7 days after initiation of the reaction, coinciding with the peak in T cell numbers. Furthermore, this expansion of the CD4+ T cell subset was selective, since no Ki67 reactivity was found in CD8+ T cells in any of the samples tested. The proportion of Ag-specific T cells accumulating within DTH lesions is uncertain, as these have been variably reported to comprise either a majority or <1% of infiltrating T cells (46, 47), and T cell proliferation may therefore reflect a bystander (non-Ag-specific) response that may be driven by cytokines alone (48, 49). However, a recent observation in mice suggests that although Ag specificity does not influence migration into inflamed tissue, only Ag-specific cells are retained (50).
After antigenic stimulation, the induction of cell cycling is driven by cytokines such as IL-2 (51). In addition, it has recently been shown that IL-15 also triggers proliferation in activated T cells (22, 32, 33, 34). We showed that the substantial proliferative activity in T cells at 7 days after PPD challenge was associated with the presence of both of these cytokines, and confirmed that epidermal KC and DC, and dermal DC expressed IL-15 (52, 53). There was a marked reduction in staining intensity for IL-15 in both KC and monocytes/macrophages on day 7, when T cell numbers and T cell proliferation were maximal. This suggests that this cytokine may only contribute to the induction of the T cell proliferation during the early phases of the DTH response, and that other cytokines may then take over this role at later stages after PPD challenge. We found that in contrast to IL-15, IL-2 expression was low 12 h after intradermal PPD, and appreciable amounts were only observed at 72 h. At 7 days after challenge, when maximal proliferation and T cell numbers were evident, maximal IL-2 expression was observed. These data, although circumstantial, are compatible with the possibility that during the Mantoux reaction, IL-15, a non-T cell-derived cytokine, may promote the initial proliferative drive until T cells themselves synthesize IL-2, which maintains the proliferative activity.
The clearance of T cells during the resolution of the DTH response may be due to both the efflux of cells or to the death of cells in situ. We found that the numbers of T cells were reduced significantly at 14 days after the initiation of the PPD challenge. At this time, significant numbers of apoptotic T cells were detected. These apoptotic cells could also be detected inside macrophages (data not shown). It is likely that the numbers of apoptotic cells detected represent a substantial underestimate of the total extent of apoptosis taking place (54). It is well recognized that activated T cells require the continued presence of certain cytokines, such as those that signal via the IL-2R common γ-chain, to prevent apoptosis (3, 4, 5, 6, 16). It has been shown that these cytokines may prevent apoptosis by up-regulating Bcl-2 relative to Bax in T cells (15, 16, 18). For mature activated T cells, IL-2 and IL-15 are the most efficient at preventing death (15, 16, 17, 18, 19). It is of interest, therefore, that on day 14, when most of the apoptosis was detected, both IL-2 and IL-15 were decreased significantly as compared with day 7, when maximal proliferation and T cell numbers were detected. The strikingly decreased Bcl-2 that was observed at this time is compatible with previous observations that these cytokines regulate apoptosis via the induction of this molecule. This suggests that when maximal levels of IL-2 and IL-15 are present, T cell proliferation may occur. Conversely, when levels of these cytokines are limiting, T cells undergo apoptosis due to cytokine withdrawal. We also investigated T cell expression of Bcl-2 and Bax in a single subject, in which the PPD response was investigated at multiple time points. The reduction of Bcl-2 after the peak of the response preceded the fall in T cell numbers, further suggesting that cytokine deprivation was involved in the resolution of the response. However, apoptosis was also observed at the peak of the response on day 7, when high levels of proliferation, IL-2, IL-15, and Bcl-2 were observed, suggesting that other mechanisms were responsible for the induction of apoptosis at this time.
The apoptosis of mature activated T cells may occur as consequence of religation of the TCR in cells that are already in cycle (55, 56, 57). This activation-induced cell death (AICD) is mediated by interaction of CD95 (Fas/Apo-1) with its ligand, which is transiently expressed on activated T cells (6, 57). The phenomenon of AICD is thought to operate in situations in which there is an excess of Ag, and may be a mechanism that prevents immunopathology resulting from overactivation of the immune system (58). We therefore investigated the expression of CD95 ligand during the Mantoux reaction. The kinetics of CD95 expression indicated that maximal expression was found in the perivascular infiltrates at the peak of the response, but was reduced significantly at 14 days. Furthermore, the expression of TNF-α, which can also induce T cell apoptosis as a consequence of binding to its receptor, was also maximal at 7 days, but was reduced at 14 days. These data suggest that at the peak of the Mantoux reaction, both CD95- and TNF-α-mediated death are likely to be involved with the apoptosis observed. However, the relative contribution of each of these pathways to the overall death observed at this time is not clear. In contrast, at 14 days, when there has been clinical resolution of the response, presumably as a result of antigenic clearance, and when T cell proliferation, CD95 ligand, and TNF-α expression are substantially reduced, it is unlikely that AICD plays a major role.
These results collectively suggest that, while apoptosis occurring during the induction phase of the PPD response may involve the interactions of either CD95 or TNFR with their ligands, apoptosis during the resolution phase may be controlled by the regulation of Bcl-2/Bax levels by cytokines (6, 57, 58). Apart from the γ-chain cytokines, stromal cell factors have also been shown in vivo to rescue activated T cells from apoptosis via a mechanism involving up-regulation of Bcl-xL independently of Bcl-2 (2, 59, 60). It would therefore be of interest to determine whether stromal cell-mediated mechanisms are also involved, both in the induction and resolution of the PPD response.
In summary, we have shown that in the Mantoux reaction, the generation of the response involves not only recruitment, but also T cell proliferation, while resolution occurs in part by induction of apoptosis in infiltrating T cells. We hypothesize that the proliferative phase and the resolution of the response appear to be controlled by different levels of the same group of cytokines, the presence of which promotes proliferation, while the absence of these mediators leads to apoptosis. Our recent observations suggest that there are high levels of Bcl-2, low levels of CD95 ligand expression, and only low levels of T cell apoptosis in cutaneous lesions of atopic eczema patients (C. H. Orteu, A. N. Akbar, L. W. Poulter, and M. H. A. Rustin, in preparation). This suggests that dysregulation of T cell apoptosis may contribute to chronicity of inflammation in cutaneous disease. Studies into the regulation of apoptosis in these lesions are clearly pertinent to their future management.
Acknowledgements
We thank Huda Al-Doujaily, Shelley Horne, Sharon Bernard, Melanie Saunders, Nicola J. Borthwick, Jonathan Crowston, and Maria Soares for excellent technical assistance. We also thank Glaxo-Wellcome for their BAD Travelling Fellowship and all of the volunteers who participated and without whom this study would not have been possible.
Footnotes
This work was supported by a grant from Sir Jules Thorn Charitable Trust (number 95/04A).
Abbreviations used in this paper: DTH, delayed-type hypersensitivity; AICD, activation-induced cell death; DC, dendritic cells; EI, erythema index; KC, keratinocytes; PPD, purified protein derivative; TBS, Tris-buffered saline; TRITC, tetramethyl rhodamine isothiocyanate; TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling; UA, unit area.