This analysis was based on a schematic representation by Gaffen SL 31 of the signaling domains of the IL-2R separating pathways associated with the IL-2/15Rβ cytoplasmic tail from those associated with the common γ chain (γc) cytoplasmic tail (Figure 1A and 1B).

The expression of several cytokine and growth factor receptors was altered by rIL-2. Four major self organizing gene clusters were up regulated (a-d, Figure 2A) and two down regulated (e-f, Figure 2A) by rIL-2.

Cluster a contained receptors associated with immune cell survival. The transferrin receptor, CD71 which is vital for continued cell growth, is induced in hematopoietic cells by various cytokines including IL-2 to maintain T cell proliferation 5253 and its pattern of expression coincides with that of CD25 (IL-Rα) in T cells stimulated through CD3 and CD28 triggering 54. The transferrin receptor is necessary for activated T cell survival and its blockage with monoclonal antibodies can arrest proliferation 55. The induction of CD71 expression and its stabilization by IL-2 had been previously reported 56. Surprisingly, the expression of CD71 was tightly linked to that of MET (receptor for hepatocite growth factor and the macrophage-stimulating protein), the urokinase plasminogen activator (PLAU) and its receptor (PLAUR) both required for activation of hepatocite growth factor-mediated cell proliferation 57. This cluster also included CKCL16 which binds to CXCR6 and is involved in trafficking of immune cells to site of inflammation 58 and the interleukin-1 receptor antagonist which inhibits the pro-inflammatory effects of IL-1 59. The induction of expression of these receptors by rIL-2 was likely to occur in cells other than T cells since their expression was not significantly altered in CD4 and CD8 subsets whether stimulated in the presence or absence of bystander PBMC.

Cluster b included the insulin-like growth factor receptor-1 and the colony stimulating factor-2 receptor-β. The former has been associated with T cell activation/proliferation and sensitization to insulin 60. This small cluster included also genes associated with down-regulation of T cell reactivity such as the prostaglandin E receptor 61, the retinoic acid receptor-γ (RARG) 62 and the IL-10Rα. Modulation of the last two receptors was T cell-specific because occurred in CD4 and CD8 T cells stimulated in the presence or absence of bystander PBMC.

Cluster c was characterized by the coordinate up regulation of several receptors for cytokines belonging to the common cytokine-receptor γ-chain (γc) (IL-2Rα, IL-2/IL-15Rβ, L-4Rα and-15Rα) and other immune stimulatory cytokines (IL-12Rβ and the IL18R accessory protein (Figure 2A and 2B). The expression of the IL-21R β-subunit which also participates to the cytokine-receptor common γ-chain complex 63 was not affected by rIL-2 stimulation. This is relevant since IL-21 participates in the modulation of lymphoid proliferation, apoptosis and differentiation. Unfortunately, no cDNA clones representative of the IL-9R (another common cytokine-receptor γc family member) were present in the arrays. Coordinately expressed with these receptors was also the expression of various nuclear receptors, integrin α 4 (CD49D, integral part of the VLA-4 receptor), the TCR δ chain, the platelet-derived growth factor receptor-α and several tumor necrosis factor receptor-associated transcripts). This observation suggests that rIL-2 amplifies its effects not only by co-regulating its own receptors but also by increasing the expression of receptors for other immune regulatory cytokines. Consistent with others' reports 3964., however, the three cDNA clones specific for the IL-7Rα chain (another member of the common cytokine-receptor γc chain family) were strongly and consistently down regulated (Cluster f, Figure 2A and 2B). This is important because this cytokine regulates different phases of immune activation through differential expression of its receptor 656667. Interestingly, IL-7Rα down regulation was not associated with a decrease in the expression of BCL-2 or lung Kruppel-like factor, both involved in IL-7-dependent cell survival 6869. In fact, the expression of both genes was significantly up-regulated by rIL-2 supporting the hypothesis that rIL-2 (and indirectly IL-15) promotes clonal expansion of CD8+ T cells independently of IL-7 in the early phases of activation while it may not be involved in the maintenance of memory T cells which requires both IL-7Rα and IL-15Rα expression. This is also supported by the significant down-regulation of CD62L observed in rIL-2-stimulated PBMC (paired Student t test p₂-value <0.001) compatible with an effector memory T cell phenotype 7071. The significance of the complete shut down of IL-7R transcription warrants more detailed analysis in the future since it was not as clearly seen in T cell subsets stimulated in the absence of PBMC. This suggests that this important modulation is dependent upon the influence of other immune cells present in the PBMC population. Subset analysis also demonstrated that the IL-2Rα chain was preferentially up regulated in CD4 compared with CD8 cells at this rIL-2 concentration. As for several other genes, cyto-fluorimetric analysis demonstrated that the changes in IL-2Rα (Figure 2C) and IL-7Ra (Figure 2D) progressed incrementally in time (3 to 7 days) underlying the lag existing between transcriptional activation and functional effect.

At the protein level, rIL-2 induced a significant release of cytokines belonging to the common cytokine-receptor γc chain family including IL-4, IL-7, IL-9 and IL-15 (Figure 2E). Obviously, IL-2 levels were also significantly increased in culture by it was not possible to distinguish whether endogenous expression of IL-2 by PBMC contributed to this increase. Thus, rIL-2 induces a generalized release of cytokines belonging to the common cytokine-receptor γc chain family but modulates cell responsiveness though the differential expression of cytokine-specific sub-units of the IL-2R. The significant modulation of different subunits of receptors associated to the common γc was not associated with modulation of the γc subunit itself suggesting that cytokine responsiveness within the family relies primarily on the differential expression of private cytokine-specific receptor subunits. Lack of modulation of the γc subunit of the IL-2R contrasts others experimental conditions in which IL-2 rendered T cells susceptible to apoptotic cell death through down-regulation of the γc in mice in vivo 72 underlining the strong dependence of cytokine signaling studies on the experimental condition tested.

The expression of IL-2 and other cytokines is controlled at multiple levels by T cells and it is mutually linked to TCR signaling 7374. Neither the TCR nor its associated co-receptors (CD3, ZAP-70, CD28 and CTLA-4) were affected by rIL-2. However, rIL-2 significantly affected downstream modulators of TCR signaling 14 converging toward cytokine promoter domains. For instance, the phospholipase C (PLC)γ-dependent pathway 7576. activates three major classes of transcription factors: NFAT, NF-kB and AP-1. In turn, rIL-2 can modify the expression of genes associated with these pathways (Figure 3A). The immediate signaling down-stream of linker for the activation of T cells (LAT) was consistently down-regulated (LcK, PLCγ) while the spleen tyrosine kinase (SYK) associated with TCR signaling was not significantly induced by rIL-2 (data not shown). However, downstream activators of cytokine transcription were up regulated. Ras associated GTPases expression was up-regulated in association with several MAP kinases (AMPK6, 7, 8). Tightly co-up regulated in the same cluster with NF-kB were several cDNA clones representing TANK (TRAF family member associated with NF-kB activation), NFAT-5 and Oct-1. Rho-C and associated genes were co-expressed in a separate cluster (Figure 3B). No induction of expression of either Jun or Fos (AP-1 complex) was observed (data not shown). A very interesting cluster included genes most consistently induced by rIL-2 in all donors (between yellow horizontal bars, Figure 3B). This cluster was enriched with genes classically associated with TCR signaling such as PKC-θ, SOS, GRP-1, and VAV. Particularly interesting was the up-regulation of CISH (cytokine inducible SH-2 domain containing protein) (paired t test p₂-value 1 × 10^-23). This was by far the gene most consistently up-regulated by rIL-2. CISH is a member of the SOCS adaptor family associated with TCR-mediated proliferation and survival of T cells 77 and its expression is specifically dependent upon the activation of STAT-5).)45. Increased expression of CISH has been reported in association with MAP kinase activation following TCR triggering leading to proliferative responses and prolonged survival of activated T cells 2877. CISH is a known early responder to IL-2 stimulation 78 and the high consistency of its induction suggests a crucial role of CISH in the interface between TCR-dependent and cytokine-dependent T cell proliferation and survival 28. CISH and SOCS-2 were the only two members of the SOCS family to be significantly modulated by rIL-2 (Figure 4A and 4B). The expression of these two genes correlated loosely in resting PBMC (R² value = 0.23; blue empty circles, Figure 4C) but was tightly correlated following rIL-2 stimulation (R² value = 0.72; red full circles, Figure 4C). The same correlation was described by Li S et al. 77 in response to TCR stimulation suggesting overlapping biology and function of these two members of the SOCS family. Of interest, however, is the distinct behavior of STAT-5 in rIL-2 stimulated PBMC compared with TRC stimulated T cells. While in our study, IL-2 significantly induced the expression of STAT-5A, LI S et al. 77 did not observe induction of expression of STAT-5 mRNA following TCR stimulation although phospho-activation of the protein was observed. Thus, it is possible that rIL-2 enhances the down-stream signaling of TCR stimulation not only by allowing the activation of STAT-5 but also by enhancing its expression. However, since no correlation was noted, in our study, between STAT-5 expression and that of CISH or SOCS-2 (data not shown) it is possible that CISH modulation of TCR and IL-2 signaling is STAT-5 independent in the early phases of the response to rIL-2 as suggested by others 77. CISH expression may be central to TCR-dependent signaling as it may function as an intermediate regulator transmitting signals from the TCR to the MAP kinase pathway through regulation of protein kinase-C (PKC) θ and subsequent activation of c-Jun-N-terminal kinase (JNK, TCR/CD28 dependent pathway) and ERK (TCR-dependent pathway) 7779. It is possible that the IL-2-induced enhancement of the expression of genes associated with the JNK pathway might be at the basis of the reduced dependence of T cells on CD28-mediated co-stimulation in presence of this cytokine. It is also not clear what the biological effects of SOCS-2 could be in response to rIL-2 stimulation. It was recently reported that SOCS-2 may participate in enhancing IL-2 signaling by accelerating the degradation of SOCS-3 80. Thus, rIL-2 induces in the conditions described here only two genes belonging to the SOCS family both of them possibly associated with stimulatory effects on T cell activation. Interestingly, contrary to others' reports 29, we did not identify significant alterations in the expression of SOCS-1. Subsets analysis again suggested that with few exceptions (i.e. CISH and Rho-C) the up-regulation of these signaling pathways was strongly dependent upon bystander PBMC as primarily isolated CD4 and CD8 cultures were not sensitive to this modulation (Figure 3B).

IRF-4 expression was strongly up regulated by rIL-2 (Figure 1C) confirming previous observations 39. Expression of this transcription factor is specific to lymphoid and myeloid cells and combines adjacent to NFAT binding motives on cytokine promoter regions interacting with NFATc1 to enhance the production of IL-2, IL-4, IL-10 and IL-13 81. It has been shown that IRF-4 expression is driven through activation of NF-kB and NFAT pathways 82 and it is dependent upon antigen-mimetic stimuli such as TCR cross-linking or treatment with phorbol ester 83. Interestingly, while IL-4 (Figure 2D) and IL-13 (Figure 5) were up-regulated by rIL-2 no change in the release of IL-10 was noted in this study either at mRNA or protein level.

Immune effector functions including lymphokines, lytic granule-associated molecules and TNF/TNFR family members were extensively modulated by rIL-2 (Figures 5 and 6). Four major clusters were identified. Cluster a included genes down-regulated such as IL-8 and TGF-β also reduced at the protein level in culture supernatants (data not shown). This is in contrast with our previous in vivo observation of consistently increased serum levels following systemic rIL-2 administration suggesting that the release of these cytokines is dependent upon interactions with cells other than PBMC 1124. Down regulation of these cytokines was specific to CD4 and CD8 cells independent of the presence of bystander PBMC. Consistent with others' reports 2939., several receptors associated with cell survival were also down-regulated including CD-27, the nerve growth factor receptor and TOSO. TIA-1 (a lytic granule associated protein whose over expression before rIL-2 therapy but not after has been linked to the immune responsiveness of melanoma metastases) was down-regulated in accordance to others' in vitro experimental findings 29. Interestingly, down regulation of TIA-1 occurred only in CD4 and CD8 subsets exposed to rIL-2 in the absence of bystander PBMC suggesting that the regulation of this biomarker of responsiveness to rIL-2 is strongly dependent upon the immunological microenvironment 26. Cluster b) included genes associated with lytic function such as cathepsins, granzyme-B, perforin 1, NK-4/IL32 and NKG7 which we have previously reported over expressed by in vitro-activated immunization-induced T cells 25. Several of these genes were also found to be over expressed in a melanoma metastasis regressing in response to rIL-2 therapy 24. Interestingly, this cluster also included IL-15, lymphotoxin-α and β, IFN-γ, FLT-3, caspase-2 and -8 and BCL-2. The expression of the genes included in this cluster was strictly dependent on rIL-2 stimulation as practically none of the non-stimulated samples expressed them at any level. CD4 and CD8 subsets stimulated in the presence of bystander PBMC behaved similarly with strong up regulation of all the genes included in cluster b. However, the induction was subset-specific in CD4 and CD8 cells exposed to rIL-2 in the absence of bystander cells. In particular, CD4 T cells over expressed caspase-8, lymphotoxin-α and β, FLT-3, and NK-4/IL-32 while the over expression of granzyme-B, perforin and NKG-7 became restricted to CD8 T cells (see also later). This information may be extremely important for future interpretation of in vivo data. For instance, as later discussed, NK-4/IL-32 is becoming increasingly recognized as a biomarker central to immune effector function. Its over expression in acutely inflamed organs may depend upon modulation by bystander cells.

Cluster c) included predominantly lymphokines (CSF-1, TNF-α, IL-15, SCYE, MCP-1, MIP-1α and MIP-1β) whose expression was coordinately associated with that of NF-kB and TANK 84. This cluster was characterized by expression of genes preferentially expressed by CD4 and CD8 T cells in the absence of by-stander PBMC even in the absence of rIL-2 stimulation. Finally, cluster d) contained a functionally heterogeneous array of genes including IP-10 and IL-12A also found to be over-expressed at the protein level in culture supernatants (paired t test p₂-value < 0.01 and = 0.005 respectively) as well as survival/apoptosis associated factors such as calpain-1, Enolase-3 and SMAD-7 8586.

The transcriptional alterations induced by rIL-2 in PBMC were largely consistent with previous reports 2939. However, it remained to be elucidated to what degree rIL-2 acts directly on T cells rather than through bystander interactions. Therefore, we compared the transcriptional profile of CD4 and CD8-expressing T cell obtained from 6 donors (3 Chinese: D12, 15 and 36 and 3 Caucasian: D38, 43 and 45). T cell subsets were sorted by negative bead separation before (referred thereafter as CD4 or CD8 T cells) or after (referred here as PBMC/CD4 or PBMC/CD8) exposure to 300 IU/ml rIL-2 for 4 hours. Purity after separation was similar in all conditions and averaged 89.3% ± 0.9% (range 80% to 97%). This approach could segregate transcriptional alterations induced in CD4 and CD8 T cells by rIL-2 directly (CD4 and CD8 subsets) from those requiring the presence of PBMC (PBMC/CD4 or PBMC/CD8 subsets). The complete data set is available as additional file 3. All subsequent analyses were performed by comparing samples with a two-tailed paired or unpaired Student t test as appropriate. For the subset analyses the cut-off p₂-value of significance was set at < 0.05 due to the smaller sample size as done by others 2939.

PBMC from leukocyte aphaeresis of 6 donors were thawed and plated in complete Iscove medium (Life Technologies, Grand Island, NY) supplemented with 10% heat inactivated human AB serum, 10 mM HEPES buffer, 100 U/ml penicillin-streptomycin, 0.5 mg/ml amphotericin B and 0.03% glutamine, at the density of 10^6 cells/well in 48 multiwell plate. After resting overnight half of the wells were treated with human recombinant IL-2 300 U/ml (rHuIL-2, Chiron Co, Emeryville, CA). IL-2 was added every two days. Treated and untreated cells were harvested after 1–3–7 days. The modulation of the expression of CD25 (IL-2 receptor alpha chain) and CD127 (IL-7 receptor) were tested by staining the cells with the following antibodies: CD4-APC, CD8-PE or CD8-FITC, CD3-PerCP and CD25-FITC (all from BD Biosciences Pharmingen, San Diego, CA) or CD127-PE (Beckman Coulter, Fullerton, CA). As negative control cells were stained with IgG FITC or PE conjugated, according with the antibody's isotype. Cells were analyzed with FACS sort (BD Bioscience) gathering them on living lymphocytes CD3 positive.

Total RNAs were extracted from each donor's PBMC as previous described and converted to cDNA by reverse transcription. PCR was performed by using iProof High-Fidelity DNA polymerase (Bio-Rad) according to the manufacturer protocol.

The following primers were used in the PCR reaction: forward: 5'-TGCCACCGCCCCATGTCTCA-3' reverse: 5'-CACAAAGATGGTACACACGGATCATT-3' using the following conditions:

98° 30 seconds, 98° 10 seconds, 62° 30 seconds, 72° 2 minutes for 30 cycles, 72° 10 minutes for final extension. PCR products were purified by adding 3 μl ExoSAP-IT (USB) per 20 μl of PCR product, at 37° for 15 minutes followed by incubation at 80° 15 minutes to inactivate the enzyme. The sequencing reaction was set up using the BigDye Terminator v3.1 Cycle Sequencing kit and ABI Prism 3700 DNA Analyzer.

The following primers were used for sequencing:

F1: 5'-TGCCACCGCCCCATGTCTCA-3'

F2: 5'-CCATCCAGGACTTCAAGCC -3'

F3: 5'-AGGACAAGGTGCCTGAGC-3'

F4:5'-TGGTGCTGCGAGAGGCTG-3'

F5: 5'-CAGCCTGAGCGTGCTTTC-3'

F6: 5'CCTGCTGCATCTTCCCACA-3'

F7: 5'TCTGACCAGCAGCCTATGAG-3'

R1: 5'-CGAACTCCAGGTGTCTTTCAA-3'

R2: 5'-CTCTATCTCCAAGGCATCCG-3'

R3: 5'-AACAGGGTCCTTCTGAGGCT-3'

R4: 5'-GGAATAGCATGTGCAACAGAG-3'

R5: 5'-GTCAGAGTTAGCTGGGACTGG-3'

R6: 5'-GGATAAGGAGACCGACTTGC-3'