Two Distinct T Cell Subsets, CD4⁺ and CD8⁺CD60⁺, and Their Cytokines Are Required for In Vitro Induction of Human Ragweed-Specific Memory IgE Responses

It is well recognized that CD4⁺ T cells and their cytokines (IL-4 and IL-13) are required for induction of human and murine IgE responses in vivo and in vitro (reviewed in Refs. 1 and 2). However, there is evidence that different cell/cytokine pathways are required for induction/maintenance of memory IgE responses. Herrick et al. (3) were the first to demonstrate that in vitro induction of benzylpenicilloyl (BPO)²-specific memory IgE responses by spleen cells of BPO-keyhole limpet hemocyanin (KLH)-sensitized mice required two distinct T cell subsets: Thy 1⁺ asialo GM1 ganglioside-negative and Thy 1⁺ asialo GM1 ganglioside-positive T cells (3); in her studies, IL-4 was responsible for approximately one-half of the memory IgE response and other cytokines (IFN-α and IFN-γ) for the remainder of the memory IgE response. Studies of Auci et al. (4), in the same BPO-KLH-sensitized mouse model, found that after each booster injection of Ag, Thy l⁺ asialo GM1 ganglioside-positive T cells transiently appeared in spleen and mesenteric lymph node before IgE responses increased in these organs (4, 5). A search for a human counterpart of murine Thy1⁺asialo GM1⁺ T cells proved futile because T cells of allergic patients do not express asialo GM1 ganglioside (6).

Smith-Norowitz (6) and Smith-Norowitz et al. (7) found that there is a statistically significant increase in numbers of the CD8⁺CD60⁺ T cells (CD60 also is a ganglioside) in the blood of serum IgE⁺ ragweed-sensitized (RS) humans at the peak of the ragweed allergy season. Others demonstrated that CD8⁺ T cells, like CD4⁺ T cells, help IgG responses (8, 9) and that CD60⁺ T cells also help these responses (10). However, the role of CD8⁺ or CD60⁺ T cells in IgE responses and T cells that simultaneously express both CD8 and CD60 in humoral responses, including IgE responses, has not been studied. The present studies further characterize CD8⁺CD60⁺ T cells and other lymphocytes in the blood of RS humans obtained at the peak of the ragweed allergy season and investigate the ability of CD8⁺CD60⁺ T cells to regulate human memory IgE responses.

We found that CD8⁺CD60⁺CD45RO⁺ T cell numbers are greatly increased in the blood of serum IgE⁺ RS humans at the peak of the ragweed allergy season. CD8⁺CD60⁺ T cells and their cytokines (IL-2, IL-4, IL-10, IL-12, IFN-α, and IFN-γ) are required for induction of human ragweed-specific memory IgE responses; CD4⁺ T cells also are required for memory IgE responses, as previously shown by others (reviewed in Refs. 1 and 2). CD8⁺CD60⁺ T cells and IFN-α in high numbers/concentration suppressed induction of memory IgE responses.

All subjects (n = 32) were medical student volunteers (males and females, ages 22–40 years). At the time of the study, none of the subjects received allergy therapy and none was being treated with any medication. RS subjects may or may not have exhibited a past history of clinical allergic reactions/symptoms. Approval was obtained from the State University of New York (SUNY) Downstate Institutional Review Board and the procedures followed were in accordance with institutional guidelines involving human subjects.

Peripheral blood (40 ml) was obtained from subjects 2 days after they were defined as either RS, serum IgE⁺ (>100 IU/ml; n = 20), or nonatopic (NA) serum IgE⁻ (<100 IU/ml; n = 12), in August–September, the peak of the ragweed allergy season in New York City. Subjects were skin tested by intradermal injection of pollen Ags (mixed ragweed (tall and short); Center Laboratories), standardized mite Ags (Dermatophagoides farinae; Miles), cat pelt (0.02 ml, 10,000 BAU/ml; Center Laboratories), Histatrol (0.l mg/ml in 0.02 ml, histamine based; Center Laboratories), or sterile diluent for allergenic extract (saline control, 0.02 ml; Center Laboratories) at the Asthma Center of Excellence (SUNY Downstate Medical Center, Brooklyn, NY). All patients defined as RS were skin test positive by a single unequivocally positive test, where the Ag-positive injection site had a mean diameter 1 mm greater than the saline site, and the histamine-positive site had a mean diameter 2 mm greater than the saline site. The RS subjects also were skin test positive to other allergens.

For flow cytometry and cell culture studies, blood was collected into EDTA Monoject tubes (Sherwood Medical) and retained for up to 2 h at room temperature. PBMC used in cell cultures were separated from whole blood on Ficoll-Paque (Pharmacia) gradients (density, 1.077). Viability was >98%, as judged by trypan blue dye exclusion.

For studies of serum IgE, blood was collected into red top Monoject tubes (Sherwood Medical) and allowed to clot for 30 min at room temperature. Sera were collected and stored at −20°C until assayed. All serum IgE determinations were conducted in the Clinical Diagnostic Laboratory at SUNY Downstate Medical Center. Serum IgE levels were determined using the UniCAP Total IgE fluoroenzyme immunoassay (Pharmacia and Upjohn Diagnostics) that was performed according to the manufacturers’ recommendation. To evaluate test results, the response for a subject’s sample was compared directly to the response of the IgE calibrators. Data are expressed as IU/ml, where serum IgE⁺ was >100 IU/ml and serum IgE⁻ was <100 IU/ml.

Mouse anti-human mAbs directly conjugated to FITC (IgG1 anti-CD3, TCRα/β, CD45RA, CD23; IgG2b anti-CD14; IgM anti-CD60); PE (IgG1 anti-TCRα/β, TCRγ/Δ, CD1d, CD4, CD8, CD45RA, CD23; IgG2a anti-CD45RO), PE-Texas Red (ECD; IgG2a anti-CD45RO), PE-cyanin 5 (PC5) (IgG2a anti-CD4, CD8). Simultest (FITC/PE-conjugated) reagents (CD3/CD4, CD3/CD8, CD3/CD19, CD3/CD16⁺CD56) and appropriately matched isotype control mAbs (FITC-conjugated IgG1, FITC-conjugated IgG2b, PE-conjugated IgG1 and IgG2a, ECD-conjugated IgG2a, PC5-conjugated IgG2a, Simultest control γ₁/γ_2a, FITC-conjugated IgM). All Abs were purchased from BD Biosciences, except IgM anti-CD60, which was purchased from Ancell, IgG1 anti-CD1d, which was purchased from BD Pharmingen, and CD45RO-ECD, CD4-PC5, CD8-PC5, IgG2a ECD, and IgG2a PC5, which were purchased from Beckman Coulter (Immunotech); all were used according to the manufacturers’ recommendation.

All labeling and sorting studies were conducted within 6 h after blood was obtained. Conjugated Abs (10 μl or 80 μl of titrated anti-CD60) directed against one to three markers were added to the blood (100 μl) in 12 × 75-mm (5-ml) tubes (Fisher Scientific) and incubated for 10 min at room temperature, after which erythrocytes were lysed with whole blood-lysing reagent (Immunoprep; Beckman Coulter), and the cells were counted. Lymphocyte distributions were determined with a Coulter Epics XL/MCL flow cytometer with System II software (Coulter), and CytoComp (Coulter) QC Windows (Flow Cytometry Systems) was used to ensure consistent instrument settings. Forward and side scatter were used to identify the lymphocyte population, with CD45 used to establish an optimal lymphocyte gate. The gain on the photomultiplier tube detecting fluorescence intensity was adjusted so that 99% of cells with background fluorescence staining were scored between 10⁰ and 10¹ on a 4-decade log scale. Specific fluorescence was reported as the percentage of cells with relative fluorescence intensity scored above background; at least 15,000 cells were counted. Data are expressed as total lymphocytes per mm³ and as percent CD8⁺CD60⁺ T cells.

Peripheral blood (3 ml) was added to RPMI 1640 (1:1 ratio) containing l-glutamine (2 mM), penicillin (100 U/ml), and streptomycin (100 μg/ml, l-glutamine-penicillin-streptomycin solution; Sigma-Aldrich). Phytohemagglutinin P (60 μl/ml; Difco Laboratories) and GolgiPlug (1 μl/ml; BD Pharmingen) were added to blood, which was then incubated for 6 h at 37°C. Blood (50 μl) was then incubated with specific mAb (10 μl): CD8-PC5 (IO Test; Beckman Coulter) or CD4-PC5 (IO Test; Beckman Coulter) plus CD45RO-ECD (IO Test; Beckman Coulter) plus CD60-FITC (Ancell) for 10 min at room temperature in the dark, after which Intraprep Reagent 1 (100 μl, Intraprep Kit; Beckman Coulter Immunotech) was added to each tube and the tubes were vortexed and then incubated for 15 min at room temperature in the dark according to the manufacturer’s recommendation. Dulbecco’s PBS (D-PBS, 4 ml; Life Technologies) was added to the tubes, which were then centrifuged at 300 × g for 5 min at room temperature. Supernatants were discarded, cells were resuspended in Intraprep Reagent 2 (100 μl; Beckman Coulter), and the tubes were incubated for 5 min at room temperature. Tubes were then gently agitated, after which Abs to individual cytokines (20 μl; BD Pharmingen) were added: PE-conjugated mouse anti-human IL-2, IL-4, or IL-12; PE-conjugated rat-anti-human IL-6, IL-10, or IL-13; and PE-conjugated mouse anti-human IFN-γ or IFN-α. The tubes were incubated for 10 min at room temperature, after which D-PBS (4 ml) was added, and the tubes were centrifuged at 300 × g for 5 min at room temperature. Supernatants were discarded, pellets were resuspended in D-PBS (400 μl) and D-PBS with 0.1% paraformaldehyde (100 μl, ImmunoPrep C; Beckman Coulter). Cells were then counted with a Coulter Epics XL/MCL Flow Cytometer using System II software (Coulter), with CytoComp (Coulter) and QC Windows (Flow Cytometry Systems) used to ensure consistent instrument settings. Data are expressed as percent total lymphocytes. Some experiments were conducted using PBMC from Ficoll-Paque gradients, with similar results obtained.

Human PBMC (1 × 10⁵/ml) were washed once with D-PBS, then fixed with 4% paraformaldehyde (Sigma-Aldrich) for 20 min at room temperature, after which they were washed with D-PBS. The cells were then incubated with 0.1% glycine buffer (Sigma-Aldrich) for 10 min at room temperature and then washed twice in D-PBS. Cells were then resuspended in D-PBS (1.0 × 10⁵/ml). For all cell-labeling experiments, normal goat serum (NGS) incubation steps were for 15 min and Ab incubations were for 15 min at room temperature, followed by two washing steps with D-PBS.

To label CD8⁺CD60⁺ TCRα/β T cells, PBMC were incubated in the following sequence: NGS: primary Ab (mouse IgG anti-human CD8 (2–5 μg/μl), clone G42-8; BD Pharmingen) and secondary Ab (Alexa Fluor 647 goat anti-rat IgG, 1/200 dilution in 0.1% BSA; Molecular Probes/Invitrogen); NGS: primary Ab (mouse IgM mAb anti-human CD60 (2–5 μg/μl, clone UM4D4; Ancell) and secondary Ab (Alexa Fluor 488 goat anti-mouse IgM, 1/200 dilution in 0.1% BSA; Molecular Probes); and NGS: primary Ab (mouse IgM anti-human TCRα/β (2–5 μg/μl, clone T10B9.1A-31; BD Pharmingen) and secondary Ab (Alexa Fluor 568 goat anti-mouse IgM, 1/ 200 dilution in 0.1% BSA; Molecular Probes).

To label CD8⁺CD60⁺CD45RO⁺ T cells, PBMC were incubated in the following sequence: NGS: primary Ab (rat IgG2b mAb anti-human CD8 (2–5 μg/μl, clone 5H10-1; Santa Cruz Biotechnology) and secondary Ab (Alexa Fluor 647 goat anti-rat IgG, 1/200 dilution in 0.1% BSA; Molecular Probes) and primary Ab (mouse IgM mAb anti-human CD60 (2–5 μg/μl, clone UM4D4; Ancell) and secondary Ab (Alexa Fluor 488 goat anti-mouse IgM, 1/200 dilution in 0.1% BSA; Molecular Probes) and NGS: primary Ab (mouse IgG anti-human CD45RO (2–5 μg/μl, clone UCHL1; BD Pharmingen) and secondary Ab (Alexa Fluor 568 goat anti-mouse IgG, 1/200 dilution in 0.1% BSA; Molecular Probes).

Cells were then resuspended in 100–200 μl of D-PBS, an aliquot was placed on each slide, then mounted using Prolong Gold (Molecular Probes), and observed under a laser-scanning confocal microscope (Leica Instruments). CD8⁺CD60⁺, CD8⁺CD60⁺CD45RO⁺, and CD8⁺CD60⁺TCRα/β⁺ cells were identified with a confocal laser-scanning microscope (Leica Instruments), with separate settings for excitation and detection of the one, two, or three fluorochromes. No labeled cells were detected when primary Ab was omitted during labeling.

PBMC depleted or not of CD4⁺ or CD8⁺CD60⁺ T cells (2 × 10⁶ in 1 ml of complete medium) were placed in 12 × 75-mm (5 ml) polystyrene tubes (Fisher Scientific) with or without purified ragweed Ag (1–100 μl, tall and short ragweed pollens; Center Laboratories), with or without OVA (1–100 mg), and cultured for 0–12 days at 37°C in a humidified atmosphere of 4% CO₂ in air. Tubes were centrifuged at 700 × g for 10 min at room temperature, supernatants collected, and stored at −20°C until assayed for IgE. The levels of IgE in culture supernatants were determined using solid-phase sandwich ELISA (Total IgE Microplate Test Kits; Kallestead Diagnostics or HOPE) performed according to the manufacturers’ recommendations. Specimens were analyzed in duplicate and a standard curve was derived from known concentrations of IgE. Plates were read using an automated microplate reader (model Elx800; Bio-Tek Instruments), with a 405-nm measurement filter and a 600 reference filter (Bio-Tek Instruments). OD were converted to IU/ml and/or ng/ml (1 IU/ml = 2.4 ng/ml). Complete RPMI 1640 medium (Fisher Scientific) consisted of HEPES (2.5 mM/100 ml; Fisher Scientific), penicillin (100 U/ml; Life Technologies), streptomycin (100 μg/ml; Eli Lilly) and l-glutamine (2 mM; Life Technologies) and was supplemented with 10% FCS (Life Technologies).

In some experiments, varying concentrations of Abs to cytokines were included in cultures with unfractionated PBMC and ragweed Ag (10 μl): human anti-IL-2 (1.0–100 μg/ml; 1 μg of Ab will neutralize ∼1.0 biological response modifier program unit of human IL-2); anti- IL-4 (1.0–100 μg/ml; 0.1- to 1.0-μg/ml Ab will neutralize the bioactivity of a 250-pg/ml solution of IL-4); anti-IL-6 (0.2–20 μg/ml; ∼0. 0.2-μg/ml Ab will neutralize >90% of the activity of a 10-pg/ml solution of IL-6); anti- IL-10 (1.0–100 μg/ml; ∼10-μg/ml Ab will neutralize 200 ng of IL-10); and anti-IFN-γ (0.5–50 μg/ml; 1 ng of Ab will neutralize 100% of antiviral activity exhibited by ∼1 U of IFN-γ) Abs, all purchased from Genzyme. Human anti-IL-12 (0.2–20 μg/ml; 50% neutralization dose (ND₅₀) was ∼1.0–2.0 μg/ml in the presence of 1.0 ng/ml IL-12) and anti-IL-13 (0.4–40 μg/ml; ND⁵⁰ was ∼1–4 μg/nl in the presence of 10 ng/ml IL-13) Abs were purchased from R&D Systems. Anti-IFN-α (1.0–100 U/ml; 1 neutralization unit is the amount of antiserum required to neutralize 1 U of IFN-α) Abs were purchased from BioSource International. The ND₅₀ is defined as that concentration of Ab required to yield one-half maximal inhibition of the cytokine activity on a responsive cell line when that cytokine is present at a concentration high enough to elicit a maximum response. All concentrations of anti-cytokine Abs were used within recommended ranges provided by the suppliers. Anti-cytokine Abs were used within 1 mo of receipt.

PBMC were first depleted of CD14⁺ (CD4^+/−) monocytes by incubating PBMC (1 × 10⁷/ml) with anti-human CD14 mAb-coated magnetic beads (25 μl of beads/1 ml of cells) (Dynabeads CD14; Invitrogen/Dynal) in 5-ml tubes (Fisher Scientific) for 20 min at 2–8°C, with gentle tilting and rotation, according to the manufacturer’s recommendation. Tubes were then placed on MPC-1 magnets (Invitrogen/Dynal) for 2 min to allow bead-bound cells to collect on tube walls. Supernatants were removed and stored on ice. Tubes were removed from magnets and bead-bound cells recovered and washed three to four times in buffer 1 (PBS with 0.1% BSA, pH 7.4) by centrifuging at 1000 rpm for 5 min at room temperature. Cells were resuspended in buffer 2 (1 × 10⁷/ml; RPMI 1640/1% FCS) and incubated in complete medium for up to 18 h at 37°C, when most cells had dissociated from beads. Tubes were placed on magnets for 2 min to allow beads to collect on walls, and the cell-containing supernatants were recovered and stored at room temperature. If the numbers of CD14⁺ cells in supernatants were >1%, the separation procedure was repeated. Purity of CD14⁺ cells was >99% (flow cytometry).

CD4⁺CD3⁺ T cells were depleted from PBMC which had been depleted of CD14⁺ cells by incubating cells (1 × 10⁷/ml) with anti-human CD4 mAb-coated magnetic beads (25 μl of beads/1 ml of cells, Dynal CD4-positive Isolation Kit; Invitrogen/Dynal) in 5-ml tubes for 20 min at 2–8°C, with gentle tilting and rotation, according to the manufacturer’s recommendation. Tubes were then placed on MPC-1 magnets (Invitrogen/Dynal) for 2 min to allow bead-bound cells to collect on walls. Supernatants were removed and stored on ice. Tubes were removed from the magnets and bead-bound cells were recovered, washed in buffer 1 by centrifuging at 1000 rpm for 5 min at room temperature, and resuspended in buffer 2 (1 × 10⁷/ml). Bead-bound cells were detached from magnetic beads using CD4 Detachabeads (1 × 10⁷ cells/25 μl of Detachabeads; Dynal), with cells incubated for 45 min at room temperature, with gentle mixing. Tubes were then placed on the magnets for 2 min to allow beads to collect on walls, after which supernatants containing released cells were transferred to a fresh tube and cells were washed twice in buffer 2 by centrifuging at 1000 rpm for 5 min at room temperature. The pellet was resuspended in buffer 2 (1 ml) and stored at room temperature. If >1% CD4⁺CD3⁺ T cells were detected, the separation procedure was repeated. Purity of the CD4⁺ cells recovered from beads was >99% (flow cytometry).

After depletion of CD4⁺CD3⁺ T cells from CD14⁺ cell-depleted PBMC, purified CD14⁺ cells were added back to the depleted PBMC in the same ratio as was present in each subject’s PBMC before fractionation. Purified CD4⁺CD3⁺ T cells, in varying numbers (0.01–10.0 × 10⁴), were then added back to the CD14⁺ cell-reconstituted, CD4⁺CD3⁺ T cell-depleted PBMC (2 × 10⁶/ml), and cells were cultured as described. In some experiments, purified CD4⁺CD3⁺ T cells were added to unfractionated PBMC before culture.

CD8⁺CD3⁺ T cells were depleted from PBMC by incubating cells (1 × 10⁷/ml) with anti-human CD8 mAb-coated magnetic beads (25 μl of beads/1 ml of cells,(Dynal CD8-positive Isolation Kit; Invitrogen/Dynal) in 5-ml tubes for 20 min at 2–8°C, with gentle tilting and rotation, according to the manufacturer’s recommendation. Tubes were then placed on MPC-1 magnets (Invitrogen/Dynal) for 2 min to allow bead-bound cells to collect on walls. Supernatants were removed and stored on ice. Tubes were removed from the magnets and bead-bound cells were recovered, washed in buffer 1 by centrifuging at 1000 rpm for 5 min at room temperature, and resuspended in buffer 2 (1 × 10⁷/ml). CD8⁺ bead-bound cells were detached from the magnetic beads using CD8 Detachabeads (1 × 10⁷ cells/25 μl of Detachabeads; Dynal), with cells incubated for 45 min at room temperature, with gentle mixing. Tubes were then placed on the magnets for 2 min to allow beads to collect on walls, after which supernatants containing released cells were transferred to a fresh tube and cells were washed twice in buffer 2 by centrifuging at 1000 rpm for 5 min at room temperature. The pellet was resuspended in buffer 2 (1 ml) and stored at room temperature. If >1% CD8⁺CD3⁺ T cells were detected in supernatants, the separation procedure was repeated. Purity of CD8⁺ (CD60^+/−) T cells recovered from beads was >99% (flow cytometry).

To separate CD8⁺CD60⁻ T cells from CD8⁺CD60⁺ T cells, CD8⁺ T cells (1 × 10⁷/ml) were incubated with mouse IgM anti-CD60 mAb (80 μl of titrated anti-CD60; Ancell) in 5-ml tubes for 20 min at 2–8°C, with gentle tilting and rotation. Cells were washed twice in PBS (2–3 ml) by centrifuging at 1000 rpm for 5 min at room temperature, after which they were resuspended in buffer 1 (1.0 × 10⁷/ml). Rat anti-mouse IgM Dynabeads (25 μl/ml; Invitrogen/Dynal) were added to cells, which were then incubated for 10 min at 2–8°C, with gentle tilting and rotation, according to the manufacturer’s recommendation. Tubes were then placed on magnets for 2 min to allow bead-bound cells to collect on walls, after which supernatants were recovered and stored on ice. Bead-bound cells were recovered, washed three to four times in buffer 1 by centrifuging at 1000 rpm for 5 min at room temperature, resuspended in buffer 2 (1 × 10⁷/ml), and incubated in complete medium for up to 18 h at 37°C, when cells had dissociated from beads. Tubes were placed on magnets for 2 min to allow beads to collect on walls and the cell-containing supernatants were recovered and stored at room temperature. Purity of CD8⁺CD60 (CD3⁺) T cells was >99% (flow cytometry).

After depletion of CD8⁺CD60^+/− (CD3⁺) T cells from PBMC, purified CD8⁺CD60⁻ T cells were added back to PBMC depleted of CD8⁺ T cells in the same ratio as was present in each subject’s PBMC before fractionation. Purified CD8⁺CD60⁺ (CD3⁺) T cells, in varying numbers (0.01–10.0 × 10³), were then added back to the CD8⁺CD60⁺ T cell-depleted PBMC (2 × 10⁶) that had been reconstituted with CD8⁺CD60⁻ T cells and cells were cultured as described. In some experiments, purified CD8⁺CD60⁺ T cells (0.01–10.0 × 10³) were added to unfractionated PBMC (2 × 10⁶) before culture.

Lymphocyte distributions and serum IgE levels of serum IgE⁺ and serum IgE⁻ subjects were compared on each variable. Significance between variables was determined using Student’s t test. A value of p < 0.05 was considered statistically significant for all comparisons. The degree of association between these measures was assessed using Pearson’s correlations. Statistical analyses were performed using SPSS for Windows, version 10.0 software.

All subjects defined as RS (n = 20) were skin test positive to ragweed Ag and to other allergens tested (see Materials and Methods) 2 days before their blood was obtained for the present studies, and all were serum IgE⁺ (107 to >1000 IU/ml; Table I). Subjects defined as NA (n = 12) were skin test negative to ragweed Ag and to the other allergens; all NA subjects were serum IgE⁻ (<100 IU/ml; Table II). Serum IgE levels were significantly increased in RS compared with NA subjects (407 ± 76 and 42 ± 7, respectively; p = 0.001; individual data shown in Tables I and II).

The distributions of blood T cells (CD3⁺, CD4+, CD8⁺, TCRα/β⁺) dramatically increased in RS compared with NA subjects (CD3: 1430/mm³ ± 88 and 956 ± 107, respectively; p = 0.002; CD4: 834/mm³ ± 52 and 618 ± 56, respectively; p = 0.008; CD8: 555/mm³ ± 50 and 352 ± 54, respectively; p = 0.01; TCRα/β: 1334/mm³ ± 81 and 887 ± 109, respectively; p = 0.003; Fig. 1 and Tables I and II). The increase in CD8⁺ T cells in RS, compared with NA subjects, included a dramatic increase in CD8⁺ T cells that simultaneously expressed CD60 (>2-fold) (107/mm³ ± 14 and 39 ± 7.0, respectively; p = 0.001). Similar numbers of CD8⁺CD60⁻ T cells were detected in both groups (p > 0.05). CD4⁺CD60⁺ T cells were not studied. In contrast, the distributions of TCRγ/Δ⁺ T cells, CD19⁺ B cells, and CD16/56⁺ NK precursor cells were similar in the blood of RS and NA subjects (p > 0.05). No CD1d⁺ T cells were detected in the blood of either RS or NA subjects (<1%; data not shown in Fig. 1).

Although the numbers of CD8⁺CD60⁺ T cells and serum IgE levels were significantly increased in RS compared with NA subjects, the correlation between the numbers of CD8⁺CD60⁺ T cells and serum IgE levels was not significant (p > 0.40; see Table I).

CD8⁺CD60⁺ T cells in the blood of RS and NA subjects were virtually all TCRα/β⁺ (>98%; Figs. 2 and 3). The numbers of CD8⁺CD60⁺ T cells expressing the memory marker CD45RO were dramatically increased in the RS, compared with NA, subjects (80–98%, ∼15%, respectively; see also Fig. 3). A subset of the CD8⁺CD60⁺ T cells of RS, but not NA subjects, expressed CD23 (∼20%).

In the present studies, excluding those described in this section, blood of RS and NA subjects always was obtained at the peak of the ragweed allergy season. Fig. 4 shows data for blood from an extensively studied RS subject over a period of 25 mo, including ragweed allergy season and outside ragweed allergy season. CD8⁺CD60⁺ T cells were present in high numbers in August or September, the peak of the ragweed allergy season. However, excluding January 2006, the numbers of CD8⁺CD60⁺ blood T cells were greatly decreased at other times of the year. There was little change in serum IgE during the entire study period, and the numbers of blood CD8⁺CD60⁺ T cells did not correlate with total serum IgE levels.

When PBMC (2 × 10⁶) were cultured for 0–12 days with ragweed Ag, IgE was first detected on day 5, reached peak levels on day 10, and decreased on day 12 (Fig. 5). No IgE responses were induced with non-cross-reacting Ag (OVA) or without Ag. PBMC of NA humans never produced IgE in response to any stimulant. In subsequent experiments, IgE levels were determined on day 10 of culture with ragweed Ag.

Depletion and reconstitution studies (see Materials and Methods) established that both CD8⁺CD60⁺ T cells and CD4⁺ T cells are required for in vitro induction of ragweed-specific memory IgE responses by PBMC of RS humans.

When PBMC of RS humans were depleted of CD8⁺CD60⁺ T cells before culture with ragweed Ag, the IgE responses observed with unfractionated PBMC (Figs. 5 and 6,A) were not induced on days 0–12 of culture (data for day 10 shown in Fig. 6,A). When low numbers of purified CD8⁺CD60⁺ T cells (0.01–1.0 × 10³) were added back to PBMC (2 × 10⁶) depleted of these cells and PBMC cultured for 10 days with ragweed Ag, ragweed-specific IgE responses were restored (Fig. 6,B). In contrast, when higher numbers of purified CD8⁺CD60⁺ T cells (10.0 × 10³) were added back to the depleted PBMC (2 × 10⁶), IgE responses were totally suppressed. Furthermore, when low numbers of purified CD8⁺CD60⁺ T cells (0.01–1.0 × 10³) were added back to unfractionated PBMC (2 × 10⁶) and PBMC cultured for 10 days with ragweed Ag, ragweed-specific IgE responses were induced (Fig. 6 C). Again, when higher numbers of CD8⁺CD60⁺ T cells (10.0 × 10³) were added back to unfractionated PBMC (2 × 10⁶), IgE responses were totally suppressed.

Previous studies by others determined that CD4⁺ T cells are required for in vitro induction of memory IgE responses (1, 2). In the present studies, no ragweed-specific IgE responses were induced when PBMC were depleted of CD4⁺ T cells and cultured for 0–12 days with ragweed Ag. These responses were completely restored when purified CD4⁺ T cells (10.0 × 10⁴⁾ were added back to depleted PBMC (2 × 10⁶; data not shown).

CD8⁺CD60⁺ blood T cells of RS humans produced IL-2, IL-4, IL-10, IL-12, IFN-α, and IFN-γ (23–40%), but not IL-6 or IL-13 (<1%; Fig. 7 A); their CD4⁺ T cells also produced IL-4 (∼30%; data not shown; other cytokines not studied).

When unfractionated PBMC obtained from the RS subjects were cultured for 10 days with ragweed Ag, IgE responses were induced (12–17 ng/ml; Fig. 7 B, see legend). When cultures also contained either anti-IL-2, IL-10, IL-12, or IFN-γ (all concentrations tested) or IL-4 (highest concentration only), in vitro induction of ragweed-specific memory IgE responses was abrogated. Interestingly, anti-IFN-α had a bimodal effect, in that IgE responses decreased with the lowest concentration of anti-IFN-α (1.0 U/ml), whereas an intermediate concentration (10.0 U/ml) potentiated IgE responses and the highest concentration (100 U/ml) had no effect. Anti-IL-6 and anti-IL-13 had no effect on IgE responses at any concentration tested.

The present studies are the first to report that 1) CD8⁺CD60⁺ (CD45RO⁺) T cell numbers are significantly increased in blood of serum IgE⁺ RS humans in ragweed allergy season, a subset of which expressed CD23; 2) CD8⁺CD60⁺ T cells and cytokines made by CD8⁺CD60⁺ T cells (IL-2, IL- 4, IL-10, IL-12, IFN-α, and IFN-γ) are required for induction of human memory IgE responses; 3) CD4⁺ T cells also are required for induction of these responses; the requirement for CD4⁺ T cells was previously reported by others (reviewed in Refs. 1 and 2); and 4) CD8⁺CD60⁺ T cells and IFN-α, depending on their numbers/concentration, can either help or suppress induction of memory IgE responses.

The CD8⁺CD60⁺ T cells found in high numbers in blood of RS humans during peak ragweed allergy season were comprised of cells that coexpressed the memory marker CD45RO (80–98%) compared with ∼15% in NA humans. From experiments in which purified CD8⁺CD60⁺ T cells were used to restore memory IgE responses by PBMC depleted of CD8⁺CD60⁺ T cells, it is known that CD8⁺CD60⁺ T cells are required for memory IgE responses. The organ(s) in which blood CD8⁺CD60⁺ (CD45RO⁺) T memory cells were generated is unknown. The number of CD8⁺CD60⁺ T cells with specificity for ragweed Ag also is unknown because the RS donors of these cells also were sensitized to other allergens (skin testing; see Materials and Methods). However, the CD8⁺CD60⁺ blood T cells were 1) obtained in peak ragweed allergy season and 2) required for induction of ragweed-specific memory IgE responses. Therefore, many, if not most, of the CD8⁺CD60⁺CD45RO⁺ T cells probably are associated with specificity to ragweed Ag. Nevertheless, the exact numbers of CD8⁺CD60⁺CD45RO⁺ T cells that were ragweed specific, rather than T cells recruited into the ragweed IgE response, is unknown. It seems reasonable to infer that during peak ragweed allergy season, ragweed-specific memory CD8⁺CD60⁺ T cells divided in response to this allergen in lymphoid organs (MALT?) and emigrated and entered the circulation, whereas after ragweed allergy season, when allergen-induced cloning decreased, the numbers of ragweed-specific blood CD8⁺CD60⁺ T cells decreased. This idea is supported by the fact that high numbers of CD8⁺CD60⁺CD45RO⁺ T cells were present in blood of one extensively studied RS human at the peak of two ragweed allergy seasons (Fig. 4) and dramatically decreased thereafter to levels observed in NA humans (Fig. 1 and Table II). Nevertheless, because all RS humans in our study also were sensitized to other allergens, it would follow that CD8⁺CD60⁺ (CD45RO⁺) T memory cells related to these allergens might also be present in blood.

In the present studies, the highest numbers of CD8⁺CD60⁺ T cells were identified in blood during ragweed allergy season, after which their numbers dramatically decreased, but there were no significant decreases in total serum IgE levels. The lack of correlation between numbers of CD8⁺CD60⁺ T cells and serum IgE levels might be explained by the fact that all RS humans also were skin test positive to other allergens. Their blood was obtained at the peak of the ragweed allergy season; therefore, many, if not most, of the CD8⁺CD60⁺ (CD45RO⁺) T cells were likely to be ragweed-specific memory cells, whereas serum IgE probably was directed against several allergens, including ragweed, so that no correlation would be expected because of the multiple specificities. Nevertheless, ragweed-specific IgE remains to be quantified in and out of ragweed allergy season, and it is possible that changes in CD8⁺CD60⁺ T cell numbers will directly correlate with changes in ragweed-specific IgE.

The present studies have shown that a subset of CD8⁺CD60⁺ T cells (∼20%) simultaneously expressed CD23, in contrast to <1% in serum IgE⁻ NA humans. Therefore, it is important to better define the identities of the CD8⁺CD60⁺ T cell subset (CD23^+/−) that help and/or suppress memory IgE responses in order that cells/cytokine pathways might be accurately determined. Because CD23 is shed from cell surfaces (11, 12, 13), it could be that in vivo, even more CD8⁺CD60⁺CD45RO⁺ T cells of RS subjects coexpressed CD23. There is substantial literature implicating CD23⁺ cells, T and B, and soluble CD23 as regulators of IgE responses (reviewed in Refs. 14, 15, 16, 17). If so, CD23⁺ T cells that coexpress CD8 and CD60 may distinguish the ε-specific helper or suppressor cells from CD8⁺ T cells that are known to help IgG (8, 9) and CD60⁺ T cells that are known to help IgG and IgA (10) responses and permit their selective deletion. However, the present experiments lay the groundwork for isolation and functional studies of this potentially important T cell subset.

The present studies confirm the well-known requirement for CD4⁺ T cells and IL-4 for induction of human IgE responses (reviewed in Refs 1 and 2). However, our studies also demonstrated that two distinct T cell subsets, CD4⁺ and CD8⁺CD60⁺, are required for induction of memory IgE responses, extending the network(s) leading to memory IgE responses to include CD4⁺ (18, 19) and CD8⁺CD60⁺ T cells and cytokines produced by CD8⁺CD60⁺ T cells (IL-2, IL-4, IL-10, IL-12, IFN-α, IFN-γ; IL-4 also was produced by CD4⁺ T cells; other cytokines were not studied). Interestingly, in the present studies, IL-13 was not required for induction of memory IgE responses. The foregoing notwithstanding, it is possible that these cytokines are also produced by other PBMC.

Although the present studies are the first to demonstrate the strict requirement for two distinct T cell subsets, CD4⁺ and CD8⁺CD60⁺ for induction of human memory IgE responses, earlier studies of Herrick et al. (3) were the first to demonstrate the requirement for two T cell subsets: Thy1⁺ asialo GM1 ganglioside-negative and Thy1⁺ asialo GM1 ganglioside-positive for induction of murine hapten-specific memory IgE responses. Furthermore, the observations with cytokines in the present human studies directly parallel those of Herrick et al. (3). Both studies demonstrated that IL- 4 was responsible for ∼50% of the in vitro memory IgE response and that IFN-γ and IFN-α also were required, with IFN-α either helping or suppressing these IgE responses, depending on its concentration. In the present studies, human memory IgE responses additionally require IL- 2, IL- 10, and IL- 12, but not IL- 13 (or IL- 6).

In the present study, help (or potentiation) or suppression of memory IgE responses was observed when purified CD8⁺CD60⁺ T cells, which made IL-2, IL-4, IL-10, IL-12, IFN-α, and IFN-γ, interacted in different proportions with PBMC of RS donors, which had been depleted of these cells, or with unfractionated PBMC. The mechanism(s) by which addition of low numbers of CD8⁺CD60⁺ T cells back to the depleted PBMC restored the ability of ragweed Ag to induce memory IgE responses, whereas addition of higher numbers totally suppressed these responses, are unknown. Because low numbers of CD8⁺CD60⁺ T cells and IFN-α helped memory IgE responses, whereas higher numbers and IFN-α suppressed them, an attractive possibility is that CD8⁺CD60⁺ (CD45RO⁺; CD23+?) IFN-α⁺ T cells secrete IFN-α, which directly or indirectly helps and/or suppresses memory IgE responses. This hypothesis is made more attractive by the fact that two T cell subsets and IFN-α performed similar helper and suppressor functions in mice (3).

The present findings with CD8⁺CD60⁺ T cells resemble earlier findings of Haskill and Axelrad (20) and Haskill and Harbrook (21), who showed that addition of low numbers of large spleen cells from mice primed with sheep erythrocytes suppressed plaque-forming cell responses of small lymphocytes, whereas higher numbers failed to do so. Kontiainen and Feldmann (22) found carrier-specific suppression with low numbers of KLH-primed cells, as low as 1 × 10³, in a system which tested helper activity as the target of suppression. Durkin et al. (23) demonstrated that low numbers of OVA-sensitized T memory cells added to higher numbers of lymph node cells from the same OVA-sensitized rats sometimes helped or potentiated and sometimes totally suppressed proliferative responses to either OVA or phytohemagglutinin. In the present studies, as was true of the earlier studies by others (20, 21, 22, 23), the mechanisms involved in the network(s) leading to help/potentiation or suppression of memory responses obviously are complex, involving other cells and mediators released by cells participating in the memory response and feedback loops (24, 25).

The mechanism(s) by which CD8⁺CD60⁺ T cells helped IgE responses in the present studies is unknown, including the exact cell-cell and cell-cytokine interactions involved. Unraveling these mechanisms will require studies such as use of chamber experiments (26, 27). These studies would determine whether, for example, CD4⁺ T cells interacted with CD8⁺CD60⁺ T cells, or vice versa, or with intermediary cells to induce memory IgE responses. The requirement for six cytokines (IL- 2, IL- 4, IL- 10, IL- 12, IFN-α, and IFN-γ), each of which was produced by ∼25% of CD8⁺CD60⁺ T cells, with IL-4 also produced by CD4⁺ T cells, and the probability that some these cytokines also might be produced by other PBMC, points to the fact that CD8⁺CD60⁺ T cells probably acted indirectly to induce memory IgE responses.

There is substantial literature related to CD8⁺ T cell- mediated killing (28, 29, 30, 31). In the present studies, it remains to be determined whether suppression of memory IgE responses observed with higher numbers of CD8⁺CD60⁺ T cells than those required for helper activity involved killing mechanisms. However, if CD8⁺CD60⁺ T cells prove to kill other PBMC to suppress memory IgE responses, this could be a mechanism by which IgE responses, in general, are selectively suppressed.

With respect to requirements for two distinct T cell populations for in vitro induction of specific memory IgE responses observed by Herrick et al. (3), Thy1⁺ asialo GM1 ganglioside negative and Thy1⁺ asialo GM1 ganglioside positive, and, in the present studies CD4⁺ and CD8⁺CD60⁺, it should be mentioned that CD60 also is a ganglioside (disialoganglioside II³(Neu5Ac)₂-LacCer(9-O-acetyl G_D3) (10), albeit a different ganglioside from asialo GM1 ganglioside. This, of course, raises the question of whether or not it is simply happenstance that one of the two T cells subsets required for memory IgE responses in the murine and human systems coexpress gangliosides or if the gangliosides play an important role in generation of memory IgE responses, and this is the subject of our ongoing studies.

Lastly, it should be mentioned that in studies in our laboratory, when PBMC obtained from serum IgE⁺ RS donors outside the allergy season were cultured with ragweed Ag, IgE frequently was not induced (>60% of experiments). When blood was obtained in season, IgE was not produced in ∼25% of experiments. The reason(s) for this is unknown. One explanation for the successful induction of ragweed-specific memory IgE responses reported in the present studies is the fact that a requirement for induction of ragweed-specific memory IgE responses was met because blood was obtained in ragweed allergy season and contained the requisite numbers of CD8⁺CD60⁺ T cells required for induction of these responses. This, however, does not rule out that there may be additional unknown requirements for induction of memory IgE responses, which have been met in the present studies by obtaining PBMC during peak ragweed allergy season.

The authors have no financial conflict of interest.