Th1/Th2 cytokine responses following HIV-1 immunization in seronegative volunteers

Abstract

The Th1/Th2 profile that follows human vaccination may profoundly influence the subsequent course of disease after infection. However, the ability to detect IL-4 has been limited outside trials of live vaccination. By using methods in which memory effector cells are allowed to antigenically expand by short term culture, followed by low-dose mitogenic stimulation, we have been able to follow the Th1/Th2 profile in HIV-1⁻volunteers enrolled in two phase I studies of HIV immunogens (a recombinant gp120 and a multivalent, octomeric V3 loop peptide). Antigen-specific interferon-gamma (IFN-γ) could be detected in primary stimulation, but IL-4 was observed only after antigenic expansion and restimulation. In both of these studies the responses after initial immunizations were dominated by IFN-γ, with IL-4 appearing only after multiple rounds of immunization, and IL-4 was temporally related to antibody production. Concomitant with the IL-4 production, the amount of supernatant IFN-γ declined. Antigen-specific IL-10 was not detected in either study. Such techniques, which have been shown to correlate with outcomes in immunotherapy, may prove useful as future surrogates of human vaccine response.

INTRODUCTION

Appropriate immunologic responses induced by vaccination in humans classically have been measured by the development of specific binding or functional antibody that correlates with a protected status as measured by either challenge studies or observational natural history cohorts. In some infections such as HIV-1, the development of appropriate cell-mediated immunity as measured by either lymphocyte proliferation responses, DTH reactions, or cytotoxic T cell activity may be associated with protection upon challenge [1]. More recently, the cellular and humoral responses to vaccinations or to infections have been characterized by the cytokine production of activated peripheral blood lymphocytes or monocytes which follows the infectious challenge [2–4]. Much of the subsequent immune reaction after infection may be controlled by the programmed cytokines produced upon activation by the challenge [5].

This cytokine response to infection in animal models is presently characterized by a Th1/Th2 paradigm in which protection can be associated with either a Th1 response (such as in leprosy, leishmaniasis and other intracellular pathogens) [6–8], or by a Th2 response (such as in many helminthic infections) [9, 10]. The measurement of these responses after vaccination in animal models also has correlated in some studies to protection after subsequent infectious challenge [11, 12]. In general, the Th1 response is characterized by an increase in antigen-specific interferon-gamma (IFN-γ), IL-12, and complement-fixing antibodies, whereas the Th2 phenotype is characterized by production of IL-4, IL-5, IL-10, and an increase in IgE, IgA, and overall IgG antibodies [13, 14]. In addition, progressive HIV infection may be characterized by an increase in Th2-like responses, which may be either a consequence or a cause of the immune deterioration [15–17]. Some individuals who have been exposed, but not infected, appear to have cellular immune responses which may be indicative of a Th1-like mediated response [18, 19].

In human studies to date there is limited direct evidence that the antigen-specific Th1/Th2 phenotype as measured in peripheral blood mononuclear cells (PBMC) existing prior to a pathogenic challenge is predictive of the outcome after subsequent exposure to an infectious agent [20, 21]. The lack of data is mostly due to the difficulty of obtaining cytokine levels immediately before infection. However, it is reasonable to believe that humans with a Th1 phenotype may be protected against some intracellular pathogens, including HIV. Many aspects of human vaccination strategies may affect these measured cytokine responses, including whether the vaccine is live or killed, a recombinant protein or a small peptide, formulated with alum or other adjuvants, and the route by which the vaccine is delivered [22–25]. Many studies performed to date would also predict that initial recombinant protein vaccination would prime with a Th1-like response, which would then be followed by an antibody response dominated by Th2-like signals [26, 27].

Also, the cytokine responses generated in vitro are quite dependent on the laboratory conditions used to generate the results. Specifically, the use of polyclonal activation versus antigenic stimulation may lead to disparate results [28], and the use of antigenically derived clones may be biased by the proliferative capabilities of each subset, which may not reflect the in vivo response [29]. The measurement of Th2-like cytokines in healthy humans is also dependent on the use of in vitro culture techniques which allow for the development of effector populations of memory T cells [30–32]. Even the media used may play a role in the results of a given in vitro experiment [33].

Despite these drawbacks, the ability to predict whether an appropriate cellular response has occurred resulting in a Th1 phenotype, or whether the system is biased toward a Th2-like response, may have importance. For example, because of the cross-inhibitory effects of IL-4 on IFN-γ production, it may be difficult to drive a Th2 response back toward a Th1 phenotype once such a cytokine profile is established [34, 35]. Likewise, vaccination during a prominent IFN-γ response may not result in appropriately high antibody production due to suppression of a Th2 response [36, 37]. Only through observational and challenge studies of humans in whom these cytokine responses are measured after vaccination will the role of these values, if any, become clear. In the following study, we measured the Th1/Th2 responses after intramuscular immunization with trivalent influenza vaccine, with an HIV-1 V3 loop peptide immunogen, and with a recombinant gp120 envelope immunogen.

MATERIALS AND METHODS

Subjects and vaccinations

Normal healthy blood donors were used to establish many of the culture conditions. These individuals had no history of recent vaccination or were vaccinated with the licensed 1995 trivalent influenza vaccine. Study subjects included volunteers who were enrolled in two phase I trials of the National Institutes of Health (NIH)-sponsored AIDS Vaccine Evaluation Group (AVEG). In the first trial (AVEG 005) individuals received either placebo, 30 μg or 100 μg of a yeast-derived, denatured recombinant HIV-1 SF-2 gp120 vaccine, called Env 2-3 (Biocine Corp., Emeryville, CA) in the adjuvant MF59, along with varying doses of the adjuvant MTP-PE at a 0, 1, 6 and 12–18 month schedule. The results presented are derived from studies on 17 patients from the University of Rochester and University of Washington participating sites. The results of this trial have been previously reported [38]. The second study (AVEG 017) included individuals who received either placebo, low dose (300 μg), or high dose (1200 μg) of an HIV-1 V3 peptide immunogen (United Biomedical Inc., Haupagge, NY) at a 0, 1, 6 and 10 month schedule. This vaccine is a mixture of equal ratios of HIV-1 principal neutralizing domain branched synthetic peptides from 15 HIV-1 strains and formulated in aluminum hydroxide adjuvant (0.4%). Each of the 15 components of the vaccine is derived from one of the major clades of HIV-1 of HIV (A, B, C, D, E), and consists of eight V3-derived homologous peptides attached to a heptalsyl core to form radially branched octomers. Results are reported only for 14 volunteers who had sufficient cells for cytokines and whose cells proliferated to at least one antigen control (Candida or tetanus).

Informed consent was obtained from all subjects under the guidelines of the US Department of Health and Human Services and the Institutional Review Board of the University of Rochester.

Cell culture conditions and reagents

PBMC were isolated by Ficoll–Paque (Pharmacia, Piscataway, NJ) density centrifugation, washed and cryopreserved in liquid nitrogen using a rate controlled freezer. For the optimization of culture conditions, some PBMC from the normal volunteers not receiving HIV vaccines were used immediately without freezing. Lymphocyte proliferation assays (LPA) were carried out in RPMI 1640 (Gibco, Grand Island, NY) plus 10% heat-inactivated human AB (Gemini Bioproducts, Calabasas, CA) sera following the AVEG standard protocol [39]. Cultures were performed in quadruplicate, and the results expressed as a stimulation index (SI), defined as the arithmetic mean ct/min of the antigen-stimulated well divided by the control media stimulus. An SI > 3.0 was considered positive if the Δct/min (difference of mean ct/min of antigen well minus control wells) was > 500.

The PBMC not used in the LPA were placed in 24-well plates at approximately 0.5–1 × 10⁶ cells/ml (0.6–1 ml) in a serum-free, albumin-containing lymphocyte culture media (AIM-V; Gibco). In some experiments the effects of adding heat-inactivated fetal calf serum (FCS; Hyclone, Logan, UT) in increasing dilutions to AIM-V were determined. The cells were stimulated with mitogens and antigens which included media control, phytohaemagglutinin (PHA, 5 μg/ml; Sigma, St Louis, MO), candida antigen (CASTA, 20 μg/ml; Greer Labs, Lenoir, NC), tetanus toxoid (5 LFU/ml; Connaught Labs, Toronto, Canada), and an appropriate vaccine-related antigen. For trial 005 the recombinant gp120 produced in a Chinese hamster ovary cell line (CHO) and CHO control (Chiron, Emeryville, CA) were utilized. The MN strain monomeric V3 loop peptide (10 μg/ml) was used for volunteers from AVEG 017. For some experiments a purified H3N2 haemagglutinin/neuraminidase or an inactivated H1N1 viral stock isolate grown in chick embryo [40] or a purified herpes simplex virus soluble antigen prepared from mink lung cells was utilized [41].

On either the fourth or fifth day of culture, half of the supernatant was removed, frozen in duplicate at −70°C, and the culture media replenished with AIM-V containing 10–20 U IL-2/ml (Genzyme, Cambridge, MA). After an additional 5 days of culture, these PBMC lines were washed in media, recounted, and then restimulated at 0.5–1 × 10⁶ cells/ml with PHA (2.5 μg/ml) plus phorbol myristate acetate (PMA; Sigma; 1 ng/ml). After 24–48 h the supernatants were harvested without disrupting the settled cell pellet and frozen in duplicate at −70°C (referred to hereafter as 10-day samples or effector cell supernatants). The specific conditions are shown for each experiment.

Cytokine measurement

The concentrations of IL-4, IFN-γ and IL-10 were determined with a capture ELISA using antibody pairs obtained from PharMingen (San Diego, CA), and followed the protocol of the manufacturer. In brief, ELISA plates (Corning, Rochester, NY) were coated with mouse anti-human cytokine MoAbs overnight, washed, and blocked with PBS/2% albumin. The samples were then added undiluted (IL-4) or diluted from 1:4 to 1:15 (IFN-γ and IL-10) in duplicate for a 3-h incubation at room temperature. After washing, a biotin-labelled second anti-cytokine MoAb was added, the plates washed, followed by avidin peroxidase, washing, and the substrate ABTS. The plates were read after 20–60 min at 405 nm. Five day and 10 day samples were run on the same plate on almost all occasions. All samples were run in duplicate, and the IL-4 and IFN-γ determinations were run from duplicate frozen samples (never thawed) on two separate occasions. Specific cytokine values were determined by recombinant standards (PharMingen or R&D Systems, Minneapolis, MN) which yielded straight line fits with a simple regression coefficient always ≥ 0.97. The lower limits of detection defined in our assays were 0.5 pg/ml for IL-4, 10 pg/ml for IL-10, and 30 pg/ml for IFN-γ.

Cellular phenotyping

In three individuals PBMC were cultured under identical conditions to those used for the determination of cytokines. These cells were harvested by gentle scraping at 0 and 10 days (prior to restimulation), washed and stained for 15 min with fluorochrome-labelled surface marker antibodies in PBS, then washed and fixed in 1% paraformaldehyde. The following antibodies were obtained from PharMingen, clones listed in parentheses: mouse IgG1-FITC (MOPC-21)/mouse IgG1-PE/CD4-Cy (RPA-T4); mouse IgG1-FITC/mouse IgG1-PE/CD8-CyC (RPA-T8); CD38-FITC (HIT2)/HLA-DR-PE (TU36)/CD4-CyC; CD38-FITC/HLA-DR-PE/CD8-CyC; CD71-FITC (M-A712)/CD25-PE (M-A251)/CD4-CyC; CD71-FITC/CD25-PE/CD8-CyC; CD62L-FITC (Dreg-56)/CD45RA-PE (HI100)/CD4-CyC; CD62L-FITC/CD45RA-PE/CD8-CyC. After washing, 15 000 events were acquired on a FACScan (Becton Dickinson, Mountain View, CA) and were analysed utilizing Cellquest software (Becton Dickinson). The cells were gated on either the CD4 or CD8 FL3 channel, with a tight gate placed on the CD8 cells to exclude the CD8^dim+ population which may represent natural killer (NK) cells. These gates were left invariant for all samples from a given individual.

Statistical analysis

Cytokine values from vaccine recipients were grouped together and compared with placebo recipients. For day 5 supernatants the values are given as absolute numbers, since there was essentially no measurable IL-4, IL-10 or IFN-γ in the control wells. For the restimulation experiments the values are presented as the antigen-specific level minus the control well. Due to the low number of patients involved in these experiments (14 in AVEG 017 and 17 in 005), low-dose and high-dose vaccine recipients were considered as one group (30 μg and 100 μg in AVEG 005 and 300 μg and 1200 μg in AVEG 017). Also due to small numbers, no attempt was made to analyse by the adjuvant received. The values are shown as mean + s.e.m. Comparisons are made using a two-tailed Student's t-test.

RESULTS

Culture conditions for cytokine determinations

Initial studies utilizing normal controls stimulated with Candida, tetanus, herpes simplex antigen, or influenza antigen revealed that essentially no IL-4 or IL-10 could be measured in the first 7 days after antigen-specific stimulation. However, by allowing the antigenically stimulated cells to proliferate by adding IL-2 at 5 days, followed by low-dose PHA/PMA mitogenic restimulation at 10 days, large amounts of IL-4 and IL-10 could be detected (Fig. 1). There was little difference in the results if the supernatants were gathered at either 24 h or 48 h after stimulation, or if the restimulation took place any time between 9 and 14 days. The highest antigen-specific values were obtained with the addition of 10–20 U/ml of IL-2 added at either day 4 or day 5. Experiments were also carried out to assess the effect of the density of cells in the wells, the dose of antigen, and the nature of the restimulation. Although very low levels of antigen-specific IL-4 could be measured after restimulation with either specific antigen or soluble CD3 (1–3 pg/ml OKT3; Sigma), these levels were so small as to preclude the measurement of significant differences (data not shown). In contrast, abundant antigen-stimulated IFN-γ could be detected in as little as 3 days. These values range from 30 to 1000 pg/ml, but after restimulation with PHA/PMA often reached 40–50 ng/ml. However, a marked increase in background was also seen in the control wells stimulated with no antigen. Adding FCS at increasing concentrations to the complete, serum-free media resulted in increased background in IFN-γ in non-stimulated control wells measured 24 h after restimulation of 10-day cultures, but gave no relative increase in antigen-stimulated IFN-γ production, and actually caused a decrease in PHA-related responses. Increasing the PHA or PMA concentrations at the time of restimulation induced a similar increase in background with no improvement in antigen-specific responses. Thus, we selected final conditions of 10⁶ cells/ml in AIM-V media alone. Fifty percent of media was replaced, with the addition of a final concentration of 10–20 U/ml of IL-2 at day 4 or day 5, followed by restimulation with PMA (1 ng/ml) and PHA (2.5 μg/ml) after washing and counting the cells at day 10. Supernatants were harvested approximately 24 h later.

Fig. 1.

A representative experiment in a normal volunteer in which peripheral blood mononuclear cells (PBMC) were stimulated with the antigen as described in the text. Supernatant harvest at 4–7 days in the absence of IL-2 failed to detect IL-4 in almost all cases, whereas expansion of the cells with IL-2 for 10 days followed by restimulation led to easily measurable levels of IL-4.

Phenotyping

Lymphocyte phenotyping was undertaken to determine the effects of the 10-day culture system on lymphocyte maturation. Both control wells and antigen-stimulated cells underwent maturation accompanied by an increase in DR and CD38 (Fig. 2). As expected, the use of IL-2 caused almost all cells to express CD25 (data not shown). The increase in these activation markers was greater in antigen-specific stimulated cells than in the media-treated cells alone. Ten day culture in media with or without antigen caused little change in cells which were 45RA⁺, or which expressed the naive phenotype of CD45RA⁺CD62L⁺ as defined by Roederer et al. [42]. However, both CD4⁺ and CD8⁺ cells showed an overall increase in expression of CD62L, consistent with an activated memory state which may correlate to the ability to produce IL-4 [43]. The 10-day culture resulted in an expansion in CD4⁺ cells, and a decrease in CD8⁺ cells, with the CD4/CD8 ratio increasing from 2.46 ± 0.56 to 4.59 ± 0.69. This change was even greater when the cells were antigen stimulated (2.46 ± 0.56 to 8.34 ± 1.39).

Fig. 2.

Peripheral blood mononuclear cells (PBMC) in culture were surface stained as described at day 0 or prior to phytohaemagglutinin (PHA)/phorbol myristate acetate (PMA) restimulation. The populations were gated on the FL3 (CyC) channel by either CD4 (a) or CD8 (b), and the percent of other markers was determined by setting gates using an isotype control. The figure represents PBMC from three individuals stimulated with media alone and two antigens (either Candida, influenza, or tetanus).

Multivalent V3 peptide vaccine trial

Responses to the multivalent HIV-1 peptide immunogen consisting of octomeric units from multiple clades were made at baseline, 2 weeks after the second immunization and 2 weeks after the 6-month immunization. In placebo recipients (n = 3) no IFN-γ, IL-10 or IL-4 were seen after MNV3 peptide stimulation at day 4 or after 10-day restimulation, although responses to the control Candida antigen were equivalent to the values in vaccine recipients. Similarly, in vaccine recipient prevaccination, no antigen-specific cytokines were produced after stimulation with MNV3 loop peptide (Fig. 3a). However, after two immunizations a non-significant trend toward an increase for IFN-γ was seen at day 4, and significant amounts of V3-specific IFN-γ (P = 0.05) could be measured after restimulation when compared with control wells (Fig. 3b). The significant IL-4 response was confined to the Candida antigen stimulation, although two individuals receiving the immunogen showed an increase in IL-4 to V3 loop. After three immunizations, the 10-day IL-4 response to the V3 loop was significant (P = 0.039), with a concomitant decrease in the IFN-γ response (Fig. 3c). Of interest, all seven immunized volunteers studied produced detectable quantities of IL-4 at this time, compared with very low levels in only two of nine and three of eight subjects at the earlier time points. Despite high levels of measurable IL-10 in the supernatants at 10 days, no significant antigen-specific IL-10 response was detected in any of these experiments to either Candida, the V3 loop peptide, the core control protein, or other mitogens (Fig. 4). Despite a trend toward increased proliferation and increased IFN-γ, no statistically significant correlation could be found between lymphocyte proliferation SIs and either IFN-γ, IL-4, or IL-10 (data not shown). This held true for both 4-day and 10-day restimulation supernatants.

Fig. 3.

The cytokine values were measured as described in supernatants harvested at 4 days (▪) or at the end of the 10-day restimulation period (□). IFN-γ is shown as ng/ml, whereas IL-4 is shown as pg/ml. When cell numbers were limited, antigens were used in the following order: V3 loop, media control, Candida, phytohaemagglutinin (PHA). The error bar represents the s.e.m. (a) Cells from vaccine recipients preimmunization stimulated with media alone (n = 8), the V3 loop (n = 9), and Candida antigen (n = 5). (b) The cytokine values 2 weeks after the second immunization (n = 9, 10 and 3, respectively). (c) The cytokine values 2 weeks after the third immunization (n = 6, 7, and 4, respectively).

Fig. 4.

Supernatant amounts of IL-10 (± s.e.m.) produced by effector cells after restimulation with phytohaemagglutinin (PHA) and phorbol myristate acetate (PMA) at 10 days in five (preimmunization), six (post-second immunization), and five (post-third immunization) volunteers who received the V3 peptide immunogen. No IL-10 was detectable in the primary culture supernatant harvested at day 4.

SF-2 HIV-1 env 2-3 trial

Similar to the V3 loop trial, lymphocyte proliferation could be measured 2 weeks after the second immunization (Fig. 5). These proliferative responses actually increased only slightly after the third and fourth vaccinations. IFN-γ production was seen at 5 days after stimulation with PHA (20–60 ng/ml), and was seen in three vaccine recipients after the third immunization (which was not seen in the placebo group receiving adjuvant alone, Fig. 6a). The day 10 IFN-γ values were similar to the day 5 results with no statistical difference. As in all of our other experiments, essentially no antigen-specific IL-4 could be found in 5-day cultures, whether measured at baseline, 2 weeks after the third immunization, or 2 weeks after the fourth immunization (≤ 5 pg/ml in 2/30 vaccinated volunteers studied). Once again, similar to AVEG protocol 017, antigen-specific IL-4 was measured in vaccinated individuals (and not placebo recipients) after three immunizations (3/9) or four immunizations (3/6). However, due to the large intersubject variations, significance was not achieved (Fig. 6b).

Fig. 5.

The logarithmic scale lymphocyte proliferative indices (± s.d.) are shown to Candida, tetanus, and gp120 (10 μg/ml) in four placebo recipients (a) and a variable number of vaccinees (b) (depending on cell viability). The graph shows that excellent proliferation was achieved 2 weeks after the second immunization (visit 6), and increased thereafter at the post-third immunization at 6 months (visit 11) and the 2 weeks after post-fourth immunization (visit 19) at 12–18 months. CHO, Chinese hamster ovary.

Fig. 6.

IFN-γ (a) and IL-4 levels (b) in supernatants of cultures (± s.e.m.) stimulated with Candida, the gp120 protein, or phytohaemagglutinin (PHA) harvested at 5 days in primary culture, or 24 h after restimulation at 10 days.

DISCUSSION

These studies were designed to measure antigen-specific IFN-γ and IL-4 responses after HIV immunization to better characterize Thl/Th2 responses induced by these immunogens. By using recombinant or synthetic peptide HIV immunogens which were given in conjunction with either novel adjuvants or alum alone, the antigen-specific IFN-γ response appeared to occur early and was then followed by IL-4 responses to the antigens. This timing of the IL-4 response was concomitant with the overall antibody responses in these trials. This IL-4 response was accompanied by a decrease in the antigen-specific IFN-γ response. In these experiments IL-4, which is the best single marker of a Th2 phenotype, could only be measured by using a protocol which allowed for the development of effector cells, known as eTh1 and eTh2 in the nomenclature of Swain & Paul [26, 44].

The methods used here were modified from procedures of others to maximize the likelihood of detecting a Th2-like response after immunization. Serum-free media were chosen because of reports that human sera or fetal calf sera containing platelet-derived growth factor inhibited the production of IL-4 [33]. Initial attempts by our group and others have failed to measure IL-4 after brief antigenic stimulation, even in situations in which theory and animal models would predict large amounts, such as in human visceral leishmaniasis [45]. In short term mitogenic in vitro stimulation, we have only been able to measure IL-4 in the first few days after stimulation by either using large quantities of mitogens, such as PHA or PMA plus calcium ionophore, or by measuring the responses in a very dysregulated patient subset, such as patients with advanced HIV infection. However, the use of antigenic expansion for 10 days followed by a submaximal mitogenic stimulation was successful in eliciting antigenic IL-4 responses 2 weeks after multiple immunizations. Although this methodology is clearly an artificial in vitro system, such methods have been used successfully by other investigators to measure allergen-specific IL-4 responses following immunotherapy [28, 31, 46].

Transgenic animal models have shown that in vitro priming with agents that promote a specific Th1/Th2 response can modify the subsequent recall response [27, 35]. However, the murine model has the advantage of the use of lymphoid or splenic tissue, whereas large scale human studies are confined to the use of PBMC. In many of these animal models, an initial Th1 response to vaccination is either accompanied by or is followed by a Th2 response, similar to our observations in human volunteers. Such sequential production of Th1 followed by Th2 cytokines has also been shown by intracellular cytokine analysis at the single-cell level [3]. The need for multiple immunizations to elicit antibody may be viewed in part by this Th1/Th2 paradigm. In addition, an initial antigen-specific response, which may at first elicit a Th1 response, may over time develop into a Th2 phenotype, as noted in the animal experiments of Bordetella pertussis vaccines [27] or in the Leishmania major model of Menon & Bretscher [47]. Whether the results of late IL-4 development seen with these recombinant and peptide vaccines are due to a time lag or to multiple immunizations will require further study. Of interest, it has often been found that boosting doses of recombinant immunogens is more effective for antibody production if a greater time lag is allowed; whether this may relate to immunization during a more ‘adaptive’ or ‘friendly’ cytokine environment characterized by a higher Th2/Th1 ratio deserves further investigation.

The restimulation of antigenically proliferating cells has been shown to result in programmed cell death, and Th1 cells may be preferentially eliminated compared with Th2 clones or lines [48–50]. It is possible that this could have led to an underestimation of the Th1 magnitude in our studies. However, we collected supernatants 4, 24 and 48 h after stimulation, and noticed minimal change in the IL-4/IFN-γ ratios. Whether this programmed death could account for the lag time required for antibody boosting as discussed above is clearly worthy of study.

Antigen-specific IL-10 was not detected with any of our methods in these studies. This finding is similar to the results of Imada et al., who compared antigen-specific activation with polyclonal stimulation in allergic human subjects [28]. Although differences could be detected in IL-4 production between normal and allergic subjects using short-term culture with antigen stimulation (no differences were seen with polyclonal stimulation), no differences were seen in antigen-specific IL-10. We were also unable to detect IL-12 in our supernatants at either 4, 5, or 10 days (data not shown).

The development of effector cells by antigen-specific expansion was associated with a number of phenotypic changes. In general, the number of cells in the control wells decreased by 10–30% over the 10 days, whereas the antigen-stimulated cells underwent limited (10–20%) to marked (400%) expansion. However, there was not always a correlation between the level of proliferation and cytokines produced. The cells expanded were generally CD4, and increased their expression of the activation markers DR, CD38 and CD25. The expansion was also accompanied by an increase of CD62L (selectin), but with little change in the naive CD45RA⁺CD62L⁺ component. As the memory phenotype has been associated with a greater production of IL-4, it will be important to measure the relative contributions of these lymphocyte subsets to the Th1/Th2 phenotype.

The relative utility of measuring Th1/Th2 markers after human immunization remains unproven. These techniques which allow the development of effector cells which produce antigen-specific IL-4 may allow for further correlation of these measurements with both total and isotypic antibody responses and other cellular markers. However, only challenge or observational natural history studies will prove the true worth of these markers and define their roles as potential correlates of immunity.