Introduction

Top Abstract Introduction Methods Results Discussion References

The existence of distinct Th1 and Th2 phenotypes, defined on the basis of specific cytokine generation, is well established in the murine system; a large body of data have accumulated to support a similar distinction in humans (Mosmann et al., 1986; Romagnani, 1991; Wierenga et al., 1991). While Th2 cells produce IL-4, IL-5 and IL-10, Th1 cells produce IFN- and tumor necrosis factor-. Such phenotypic specificity plays an important role in both protective and pathologic immune responses (Paul and Seder, 1994). While Th1-mediated responses confer protection from intracellular parasites and increase the severity of certain autoimmune diseases, Th2-mediated responses govern both protection from certain extracellular pathogens and the inflammation characteristic of allergic disease. Despite these phenotypic and functional differences, specific regulatory differences between Th1 and Th2 cells have been difficult to establish at a molecular level. Among the candidates for such molecular regulators are CD30 (Del Prete et al., 1995a and b), specific chains of the IFN receptor (Pernis et al., 1995) and intracellular cAMP (Novak and Rothenberg, 1990; Gajewski et al., 1990). Pertaining to cAMP, studies performed in mice have provided evidence for both increased resting levels of intracellular cAMP in Th2-like cells, as well as an enhanced sensitivity of lymphokine production to down-regulation by cAMP-elevating agents in Th1 cells. To date, similar studies in antigen-specific human Th1 and Th2 cells have not been performed.

The steady-state, intracellular level of cAMP is controlled predominantly by PDE. This superfamily of enzymes is comprised of seven distinct families, characterized on the basis of substrate specificity, inhibitor sensitivity and sequence homology (Beavo et al., 1994). Two of these families, PDE3 and PDE4, are primary constituents of lymphocytes (Essayan and Lichtenstein, 1994). Previous studies from our laboratory have documented the ability of PDE4 inhibitors to down-regulate antigen-driven proliferation and cytokine gene expression in peripheral blood mononuclear cells (Essayan et al., 1994, 1995). The potential for differential regulation of PDE4 isoforms in Th1 and Th2 cells has been suggested (Essayan et al., 1995; Engels et al., 1994); however, to date, studies to directly correlate intracellular cAMP, PDE4 isoform expression and functional parameters in antigen-specific Th1 and Th2 clones have not been performed.

In this report, we delineate the cAMP-dependent modulation of antigen-driven proliferation, cytokine gene expression and cytokine protein production in a set of Amb a 1- (a major allergen of short RW, Ambrosia artemisiifolia) specific human T cell clones that have been previously characterized as Th0, Th1 or Th2 (Essayan et al., 1996). Data are included on the modulation of intracellular cAMP by PDE inhibitors and by an adenylyl cyclase activator, as well as assessment of PDE4 isoform composition in the Th1 and Th2 clones.

	Methods

Top Abstract Introduction Methods Results Discussion References

Derivation of antigen-specific T lymphocyte clones. The derivation of the antigen-specific T cell clones used in these experiments has been described (Essayan et al., 1996; Huang et al., 1991). Briefly, PBMCs from an atopic asthmatic subject with epicutaneous skin test reactivity to short RW (A. artemisiifolia) and short RW-specific IgE antibody = 1761 ng/ml by RAST were cultured in the presence of short RW antigen (10 µg/ml). This primary culture was followed by successive, biweekly restimulations of the antigen-specific T cells with the major short RW antigen, Amb a 1, in the presence of autologous, irradiated PBMCs as a source of APCs. The resulting antigen-specific, CD4⁺ T cell line was cloned and subcloned using the limiting dilution technique. Of 30 clones obtained in this manner, 8 were selected for further analysis based on their similar degrees of proliferation to Amb a 1 antigen. Antigen-driven proliferative responses of the clones were limited to Amb a 1, confirming antigen specificity. Cytokine profiles were determined by both RT-PCR and enzyme linked immunosorbant assay for cytokine protein secreted into culture supernatants. Based on these data, phenotypic assignment to Th0, Th1 or Th2 was made.

Proliferation assays. Proliferation assays were performed as previously described (Essayan et al., 1994, 1996) using conditions optimized for cell number, clone/APC cell ratio, antigen concentration, drug concentration and kinetics. Briefly, 2 × 10⁴ clonal T cells were cultured with 1.5 × 10⁵ APCs in the absence and presence of short RW antigen in 96-well plates. Culture conditions were designated by the presence of various concentrations and combinations of the PDE3 inhibitor (siguazodan), the PDE4 inhibitor (rolipram) and the adenylyl cyclase activator (isoproterenol) for the entire culture period. All selective PDE inhibitors were the kind gift of Drs. Mary Barnette and Ted Torphy, SmithKline Beecham Pharmaceuticals (King of Prussia, PA); isoproterenol was purchased from Sigma Chemical Co. (St. Louis, MO). No exogenous cytokines were used in these cultures. The cells were preincubated with PDE inhibitors for 90 min and/or isoproterenol for 5 min immediately before the addition of antigen, both intervals previously determined as optimal (Henney et al., 1972; data not shown). All culture conditions were performed in triplicate and incubated for 72 hr. The cultures were then pulsed with 1 µCi of [³H]-thymidine for an additional 20 hr, harvested onto glass fiber filter strips in a PHD multichannel harvester (both from Cambridge Technologies Inc., Watertown, MA) and counted in a beta counter (LS 5000 TD, Beckman Instruments Inc., Fullerton, CA).

Gene expression assays. Cytokine gene expression was determined as previously described (Essayan et al., 1995; Huang, et al., 1994) using conditions optimized for cell number, clone/APC cell ratio, antigen concentration, drug concentration and kinetics. Briefly, 2 × 10⁵ clonal T cells were cultured with 3 × 10⁶ APCs in the absence and presence of short RW antigen in slanted 14-ml polypropylene tubes. Culture conditions were designated by the absence or presence of a 10⁵ M concentration of the PDE3 inhibitor (siguazodan), the PDE4 inhibitor (rolipram) or both for the entire culture period. Again, no exogenous cytokine was used. The cells were preincubated with drug for 90 min immediately before the addition of antigen and further culture for 12 hr. Cultures for PDE4 isoform gene expression contained 3 × 10⁶ clonal T cells. To obtain data specific to the designated T cell phenotype, no APCs were used in these cultures; since the lack of APCs obviated the use of antigen, PHA (5 µg/ml) was used to activate the clonal T cells. Cells were incubated for 12 hr in the absence or presence of PHA in slanted 14 ml polypropylene tubes.

View this table: [in this window] [in a new window]   TABLE 1 Cytokine-,  actin- and PDE4-specific RT-PCR primers

Cytokine protein secretion assays. Cytokine protein secretion was assessed by enzyme linked immunosorbant assay (Biosource, Int., Camarillo, CA) according to the manufacturer's instructions. Quantitation was achieved using the WHO standard provided by the company. Briefly, duplicate cultures to those used in the cytokine gene expression experiments were constructed and incubated for 12 hr. Supernatants from these cultures were harvested, and cellular debris was removed by centrifugation. Supernatants were stored at -20°C until assayed. Dilutions of samples, when necessary, were performed in culture medium. All standards and samples were tested in duplicate. Most samples were tested at two different dilutions and compared for internal consistency.

cAMP assays. Intracellular cAMP levels were assessed by EIA (Amersham Corporation, Arlington Heights, IL) according to the manufacturer's instructions. Quantitation was achieved using the nonacetylated standard provided by the company. Again, to obtain data specific to the designated T cell phenotype, no APCs were used in these cultures. Briefly, 5 × 10⁵ clonal T cells were incubated with rolipram (10⁵ M) for 90 min, with or without a terminal 5 min incubation with isoproterenol (10⁶ M). Untreated aliquots of the T cell clones and aliquots treated with siguazodan (10⁶ M) were studied for comparison. On completion of the drug incubation step, the reactions were quenched with ethanol at -20°C. Particulate matter was removed by centrifugation and washed with 65% ethanol; the two ethanol-containing supernatants were combined, dried and resuspended in the assay buffer provided in the EIA kit. The samples were stored at -70°C until assayed.

Statistical analysis. Mean and standard error values, as well as t test comparisons, were derived using StatView (BrainPower, Inc., Calabasas, CA) on a Macintosh PowerBook 145B computer. The t tests were paired, two-tailed. Proliferation data are depicted as percent inhibition for each condition, calculated based on inhibition relative to stimulated, drug-free mean counts, subtracted in every case for background counts with media alone. ELISA and EIA data are presented in standard units. IC₅₀ values represent the concentration of drug at 50% inhibition; EC₅₀ values represent the concentration of drug at 50% efficacy.

	Results

Top Abstract Introduction Methods Results Discussion References

Proliferation assays. Figure 1 depicts the percent inhibition of proliferation achieved with rolipram alone (PDE4 inhibitor), siguazodan alone (PDE3 inhibitor) and rolipram in the presence of 10⁵ M siguazodan from multiple T cell clones assessed on several occasions, segregated by T cell phenotype; solubility of the compounds precluded the use of higher concentrations. Although a modest degree of independent efficacy was evident with the use of siguazodan in the Th1 clones, no independent efficacy was evident with this agent in the Th2 clones. Rolipram showed marked independent efficacy in down-regulating the antigen-driven proliferative response of both Th1 and Th2 clones; however, both IC₅₀ and EC₅₀ values were 4-fold greater in the Th1 clones than the Th2 clones (IC₅₀ = 30 µM vs. 8 µM; EC₅₀ = 10 µM vs. 2 µM; P < .05). Both Th1 and Th2 clones showed a modest but statistically significant additive efficacy of 10⁵ M siguazodan with rolipram at 10⁵ and 10⁴ M (P < .05 for each). These data confirm and extend our previous findings in PBMCs (Essayan et al., 1994). Interestingly, Th0 clones showed a similar pattern of drug efficacy, but their proliferative responses were midway in magnitude between those of the Th1 and Th2 clones (data not shown).

View larger version (K): [in this window] [in a new window]   Fig. 1.   Modulation of clonal T cell proliferative responses by PDE inhibitors. Data are presented for stimulation of two Th1 and two Th2 Amb a 1-specific T cell clones in the presence of varying concentrations of PDE inhibitors. Data are presented as percentage inhibition ± S.E.M. relative to stimulated, drug-free cultures, corrected for background with media alone (63104 ± 7069 and 980 ± 193 cpm, respectively for Th1; 68221 ± 16345 and 1167 ± 325 cpm, respectively for Th2). Each of the four clones was used in three separate replicate experiments; n = 12 individual experiments.

View larger version (K): [in this window] [in a new window]   Fig. 2.   Modulation of clonal T cell proliferative responses by a PDE4 inhibitor and an adenylyl cyclase activator. Data are presented for stimulation of two Th1 and two Th2 Amb a 1-specific T cell clones in the presence of varying concentrations of rolipram and isoproterenol. Data are presented as percentage inhibition ± S.E.M. relative to stimulated, drug-free cultures, corrected for background with media alone (56519 ± 6117 and 851 ± 296 cpm, respectively, for Th1; 57938 ± 7267 and 981 ± 198 cpm, respectively, for Th2). Each clone was used in three separate replicate experiments; n = 12 individual experiments.

Cytokine gene expression. Figure 3 shows the  actin- and cytokine-specific RT-PCR amplification products from a representative study of Th1 and Th2 clones cultured with antigen and APCs in the absence or presence of 10⁵ M rolipram and/or 10⁵ M siguazodan. Doses of drugs were selected based on efficacy in preliminary dose-response studies (10⁷-10⁴ M, data not shown); these concentrations gave consistent results although remaining within the range of selectivity for the individual isozymes. Data from the Th2 clone are shown in the top three rows, whereas data from the Th1 clone are shown in the fourth and fifth rows. Resting clonal cells cultured with APCs in the absence of antigen did not express message for proinflammatory cytokines (data not shown). Adequate normalization of RNA was confirmed by the equality of RT-PCR amplification products at subsaturating cycle number for  actin gene expression (first and fourth rows for Th2 and Th1 clones, respectively). Exposure to 10⁵ M rolipram (second column) caused a clear down-regulation of gene expression for IL-4, IL-5 (Th2 clone, second and third rows) and IFN- (Th1 clone, fifth row) when compared to the drug-free condition (first column). Siguazodan showed no independent efficacy in down-regulating gene expression of proinflammatory cytokines (third column). Finally, in contrast to the proliferative response data, no additional efficacy of siguazodan with rolipram was evident (fourth column). These data confirm and extend our previous findings in PBMCs (Essayan et al., 1995).

View larger version (K): [in this window] [in a new window]   Fig. 3.   Modulation of cytokine gene expression by PDE inhibitors in Th1 and Th2 clones. Representative results of RT-PCR products for  actin, IL-4, IL-5 and IFN- are shown for two different antigen-stimulated T cell clones under four different culture conditions: drug-free, rolipram alone, siguazodan alone and rolipram with siguazodan. The top three rows are from a Th2 clone, whereas the bottom two rows are from a Th1 clone. Normalization by equivalent  actin gene expression at subsaturating cycle number is depicted for each condition in the first and fourth rows. The DNA size marker is shown in the fifth column. Each of four clones was used in two separate replicate experiments.

Cytokine protein secretion. Figure 4 depict the amounts of IL-4 and IFN- secreted from Th0, Th1 and Th2 clones cultured with antigen and APCs in the absence or presence of 10⁵ M rolipram and/or 10⁵ M siguazodan. In both figures, the same individual T cell clones are depicted by the same symbols; Th0 clones are in gray although the Th1 clones are in black and the Th2 clones are in white. The accuracy of the individual values for a clone was confirmed by both duplicate culture experiments as well as replicate ELISA assays at different dilutions of the culture supernatants (data not shown). A clear distinction between Th1 and Th2 clones was evident in the cytokine secretion patterns. Culture with 10⁵ M rolipram, with or without 10⁵ M siguazodan, produced a significant down-regulation of both IL-4 secretion (2048 ± 481 vs. 588 ± 127 or 853 ± 207 pg/ml; P < .01) and IFN- secretion (1938 ± 608 vs. 198 ± 68 or 238 ± 91 pg/ml; P < .01). Siguazodan as a single agent was ineffective in down-regulating IL-4 or IFN- production (2350 ± 591 and 2084 ± 665 pg/ml, respectively); no significant additional efficacy of siguazodan with rolipram was evident. These results are analogous to the results of cytokine gene expression discussed above.

View larger version (K): [in this window] [in a new window]   Fig. 4.   Modulation of IL-4 (top) and IFN- (bottom) secretion by PDE inhibitors in a panel of eight T cell clones. Cytokine protein, measured by ELISA, is depicted for each of eight T cell clones, as indicated by individual symbols, under the same four culture conditions as in figure 3. Gray symbols, Th0; black symbols, Th1 and white symbols, Th2. Mean values in pg/ml are shown by the horizontal bars. Mean stimulated, drug-free IL-4 levels = 2013, 175 and 3690 pg/ml for Th0, Th1 and Th2 clones, respectively. Mean stimulated, drug-free IFN- levels = 2182, 3190 and 198 pg/ml for Th0, Th1 and Th2 clones, respectively. **P < .01.

cAMP assays. Figure 5 depicts the levels of intracellular cAMP in Th1 and Th2 clones cultured in the absence of antigen and APCs, using 10⁵ M rolipram and/or 10⁶ M isoproterenol to modulate cAMP levels; results with 10⁶ M siguazodan are shown for comparison. The resting levels of cAMP were essentially identical in the Th1 and Th2 clones (985 ± 57 and 975 ± 54 fmol/10⁶ cells, respectively). Exposure to rolipram or isoproterenol individually caused significant increases in intracellular cAMP in both Th phenotypes (P < .01); these values are in close agreement with those obtained from purified peripheral CD4⁺ T cells (Giembycz et al., 1996). Although additive efficacy was evident between isoproterenol and rolipram in the Th1 clone, no such effect was evident in the Th2 clone (P < .001 and NS, respectively). Moreover, the Th2 cells were consistently more sensitive to rolipram or isoproterenol as single agents than the Th1 cells, suggesting a mechanistic difference in cAMP regulation between the two lymphocyte subtypes (P < .05 for rolipram or isoproterenol). Siguazodan caused significant and equivalent elevations of intracellular cAMP in both Th1 and Th2 clones.

View larger version (K): [in this window] [in a new window]   Fig. 5.   Modulation of intracellular cAMP by a PDE4 inhibitor and/or an adenylyl cyclase activator. Values for intracellular cAMP from APC-free Th1 and Th2 cells are depicted for various combinations of exposure to 105 M rolipram and 106 M isoproterenol. Values for 106 M siguazodan (a PDE3 inhibitor) are shown for comparison. Data are presented as the mean value ± S.E.M. *P < .05. Each clone was used in three separate replicate experiments.

PDE gene expression. Figure 6 shows the  actin- and PDE4-specific RT-PCR amplification products from a representative study of Th1 and Th2 clones cultured without APCs in the absence or presence of 5 µg/ml PHA (for cellular activation in the absence of APCs). Adequate normalization of RNA was confirmed by the equality of RT-PCR amplification products at subsaturating cycle number for  actin gene expression (first row). Cellular activation by PHA neither enhanced nor diminished gene expression for any of the PDE4 isoforms in either T cell subset. Although identical levels of gene expression for PDE4A and PDE4B were seen in the Th1 and Th2 clones, a consistently lower level of gene expression for PDE4C and a lack of gene expression for PDE4D was evident in the Th1 cells when compared by RT-PCR to the Th2 cells.

View larger version (K): [in this window] [in a new window]   Fig. 6.   Resting and stimulated PDE4 isoform gene expression in Th1 and Th2 clones. Representative results of RT-PCR products for the four known isoforms of the PDE4 are shown for a Th1 clone and a Th2 clone, with and without stimulation by PHA, all in the absence of APCs. Normalization by equivalent  actin gene expression at subsaturating cycle number is depicted for each condition in the first row. The DNA size marker is shown in the fifth column. Four individual clones were used in two replicate experiments.

	Discussion

Top Abstract Introduction Methods Results Discussion References

We have reported the differential expression of PDE4 isoforms in Th1 and Th2 cells, and provided evidence for the differential efficacy of a PDE4 inhibitor between these T cell phenotypes. These data suggest that Th1 and Th2 clones maintain equivalent steady-state intracellular concentrations of cAMP, but that the Th2 clones exhibit a more marked elevation in cAMP in response to treatment with a PDE4 inhibitor. The functional significance of this finding is that an augmented sensitivity of antigen-driven proliferative responses to treatment with a PDE4 inhibitor exists; the relative excess of PDE4C and specific presence of PDE4D may confer this enhanced sensitivity. Thus, the differential regulation of intracellular cAMP, via the specific use of PDE4 isoforms, may represent an important molecular difference between Th1 and Th2 phenotypes.

A comparison of these data with those of Novak and Rothenberg (1990) provides additional insights into the differential regulation of T cell phenotypes by intracellular cAMP. These investigators, working in the murine system, found higher base-line levels of cAMP in the Th2 clones, D10.G4.1 and CDC-25, than in the Th1 clone, A.E7. However, in these experiments cAMP levels were assessed on day 7 after restimulation, and these data did not differ from those derived at day 4 after restimulation, when cellular activation should still be near maximal. This lack of difference between days 4 and 7 suggests that the cell populations studied by Novak and Rothenberg (1990) were not composed of resting cells. Studies from our laboratory indicate that D10.G4.1 cells may not return fully to a resting state for 10 to 14 days after antigen stimulation (data not shown). Thus the cAMP levels reported in the study of Novak and Rothenberg (1990) are likely to represent stimulated values; under their conditions, higher cAMP levels in the Th2 clones might be expected. However, our data suggest that resting human Th1 and Th2 cells show no differences in baseline intracellular cAMP levels.

Gajewski et al. (1990) have also studied cAMP-mediated effects in antigen-specific murine T cell clones. In Th1 clones, they observed a 30-fold greater sensitivity of IL-2 production to inhibition by 8-Br-cAMP (an analogue of cAMP), compared to its effect on IL-4 production by Th2 clones. However, the converse was observed with IL-2-driven proliferation of Th1 and Th2 clones: The proliferative response of the Th2 clones was 10-fold more sensitive than that of the Th1 clones to inhibition by 8-Br-cAMP (8-bromoadenosine 3,5-cyclic monophosphate). These observations are consistent with our data in human, antigen-driven T cell clones. Regarding proliferative responses, our Th2 clones were 4-fold more sensitive to PDE4 inhibition than were our Th1 clones. Although there was a suggestion in the ELISA studies that rolipram-induced greater suppression of IFN- (in the Th1 clones) than of IL-4 (in the Th2 clones), study design and low sample numbers preclude adequate comparative statistical analysis.

A number of interesting conclusions may be drawn when comparing these current data to our previous studies of PDE4 inhibition in antigen-driven human PBMCs (Essayan et al., 1994, 1995). The current clonal data confirm our earlier finding of a differential response to PDE4 inhibitors between PBMCs stimulated with a Th1-promoting antigen (tetanus toxoid, TT) and a Th2-promoting antigen (RW extract, RW); a relative resistance to PDE4 inhibition was seen in the tetanus toxoid-driven PBMCs compared to the RW-driven PBMCs (Essayan et al., 1994). In that study, the PDE3 inhibitor was ineffective in down-regulating antigen-driven proliferative responses. A subsequent study of cytokine gene expression in RW- and TT-driven PBMCs treated with PDE inhibitors revealed significant down-regulation of IL-5 and IFN- by rolipram (Essayan et al., 1995); a significant contribution of Th0 cells to the antigen-driven response of PBMCs was shown, consistent with the frequency of the various T cell subsets in cloning experiments from our laboratory (Essayan et al., 1996). Valid analysis of cAMP levels in that study was not possible due to the mixed cell design. Evidence for differential expression of PDE4 isoforms in a variety of immune cell lines, including T cells, B cells and basophils, was also presented (Essayan et al., 1995). Specifically, although the HLA-DR2.2-restricted, Amb a 5-specific Th2 cell line, AP.1, expressed mRNA for both PDE4A and PDE4B, the Th1-like Jurkat cell line did not express PDE4A; this last finding has recently been confirmed by others (Engels et al., 1994). However, the current data suggest that the differences in PDE4 gene expression in nontransformed, antigen-specific T cell clones actually occur only in the PDE4C and PDE4D isoforms. Thus, the potential for selective efficacy of PDE4 inhibitors on Th2 cells may positively affect the development of inhibitors selective for the PDE4C and PDE4D isoforms. Finally, the involvement of PDE4 isoforms in the genotype of Jurkat cells, viewed in the context of studies showing modulation of tumor cell growth and differentiation by PDE4 inhibitors, raises the possibility of primary or secondary aberrations of PDE4 isoforms in carcinogenesis (Drees et al., 1993).

The lack of efficacy of siguazodan in these and our previous experiments raises questions about the expression of PDE3 isoforms in T cells and the bioactivity of this particular PDE3 inhibitor. Interestingly, both T cell and PBMC lysates demonstrate a PDE3 peak by diethylaminoethyl-Sepharose column separation that is inhibited by siguazodan (Essayan et al., 1994; Robicsek et al., 1991). However, we have been unable to detect the human cardiac PDE3 by RT-PCR in human T cells (data not shown); other human PDE3 isoforms have not been cloned. Thus, although PDE3 clearly is expressed in human T cells, further molecular analysis cannot be performed at this time. Functionally, treatment of human Th1 or Th2 clones with siguazodan 10⁶ M resulted in a significant elevation of intracellular cAMP, implicating either compartmentalization of PDEs or cGMP-dependence as key modulators of the efficacy of PDE3 inhibitors during T cell activation. Finally, we have used three additional PDE3 inhibitors, each of which is known to bind its intracellular target and to inhibit cAMP hydrolysis in cell lysates. Each of these inhibitors has shown a lack of efficacy in modulating proliferation and cytokine generation from PBMCs (Essayan et al., 1994b). Thus, although PDE3 inhibitors have their predicted biochemical effect in human T cells, their lack of efficacy in modulating proliferation and cytokine generation is a consistent finding.

We have taken a number of precautions in our methodology to ensure the reliability of our results. First, the time interval for incubation of our proliferation assays has been optimized to maximize the signal/noise ratio. Second, the phenotypic profiles of the T cell clones used in these experiments have remained constant through repeat analyses over the course of 6 mo, precluding the effects of cellular differentiation events on these data. Third, replicate experiments using the same clone over a 3-mo period have yielded nearly identical results. Fourth, we have continued to use a complex, multi-step normalization process with our RT-PCR assay, as previously described, to ensure valid quantitative comparisons between culture conditions; the 12-hr culture interval precludes the effect of cellular proliferation on comparative cytokine gene expression. Fifth, the possibility of cellular senescence as an explanation of our findings is negated by >99% cell viability by trypan blue exclusion at the beginning and end of the culture period. Moreover, cells cultured in the absence or presence of PDE4 inhibitors restimulated with antigen, after drug removal, exhibit responses independent of prior drug treatment. Finally, the meticulous use of subsaturating cycle numbers in the PCR assays assures the validity of the comparative data. The close correlation of gene expression with cytokine protein secretion supports this conclusion.

Although the differential regulation of two PDE4 isoforms between Th1 and Th2 clones is not surprising, a significant difference in the sensitivity of these two cell phenotypes to PDE4 inhibition might not have been predicted. Because rolipram is believed to have no isoform selectivity, a difference due to the specific inhibitor is unlikely. However, a number of other mechanisms could account for this finding. One hypothesis is that PDE4C and PDE4D quantitatively constitute the predominant isoforms in Th2 cells; unfortunately, precise quantitation of PDE4 isoforms in lymphocytes has not been performed. A second hypothesis might invoke functional compartmentalization of PDE4 isoforms, with the C and D isoforms controlling proliferation and possibly cytokine generation. Because the treatment of intact, resting Th2 cells with rolipram induced a more marked elevation of cAMP than did treatment of intact, resting Th1 cells, our data appear to lend more support to the first hypothesis. Clarification of these issues must await improved techniques for PDE4 isoform isolation and quantitation.

Our data also support the differential regulation of intracellular cAMP in T cell subsets. However, the precise mechanisms by which elevations of intracellular cAMP down-regulate immune responses remain unclear. Increases in intracellular cAMP may selectively inhibit Ras/Raf-1 binding through hyperphosphorylation of Raf-1 on serine 43; however, the potential role of PKA isoforms in this model has not been clarified (Wu et al., 1993; Cook and McCormick, 1993). Type I PKA has been shown to interact with the T cell receptor-CD3 complex, but specific consequences of this interaction are unknown (Skalhegg et al., 1994). Finally, elevations of intracellular cAMP, acting through PKA, have been shown to disrupt multiple promoter-enhancer interactions; however, the involvement of specific isoforms of PKA in this model has also not been proven (Chen and Rothenberg, 1994). An improved understanding of the precise pathways involved in cAMP-mediated signal transduction will help to define the potential for differential regulation of downstream effectors between T cell subsets and identify potential new targets for pharmacologic regulation of immune responses.

In conclusion, we provide the first documentation in human Th1 and Th2 clones of the differential regulation of cAMP, associated with differential expression of the PDE4C and PDE4D isoforms. The increased sensitivity of Th2 cells, compared to Th1 cells, to PDE4 inhibition suggests the potential to target treatment to specific T cell phenotypes.