Sézary syndrome (SS) is the leukemic phase of cutaneous T cell lymphoma characterized by the proliferation of clonally derived CD4+ T cells that release cytokines of the Th2 T cell phenotype (IL-4, IL-5, IL-10), whereas Th1 T cell cytokines (IL-2, IFN-γ) are markedly depressed as is expression of IL-12, a pivotal cytokine for Th1 cell differentiation. Normal Th1 cells express both the β1 and β2 chains of the IL-12 receptor (IL-12R) and tyrosine phosphorylate STAT4 in response to IL-12. Th2 T cells express only the IL-12R β1 and thus do not tyrosine phosphorylate STAT4 in response to IL-12. To determine whether SS cells are Th2-like at the level of IL-12 signal transduction, we analyzed RNA from seven patients for the presence of message for the IL-12R β1 and β2 genes using RNase protection assays and assessed whether IL-12 induced tyrosine-phosphorylation of STAT4 by immunoblotting. In PBL from six of seven SS patients tested, β2 message was expressed at low to undetectable levels and its expression could not be stimulated by either IFN-α or IFN- γ, which stimulated β2 expression in control PBL. The absence of β2 expression is further supportive evidence for the Th2 lineage of SS cells. However, unlike normal Th2 cells, SS cells also showed severely reduced levels of STAT4, suggesting that they have a depressed response to any inducer of the STAT4 signal transduction pathway, including IFN-α. This is the first observation linking STAT4 gene expression with a human disease and suggests that dysregulation of STAT4 expression may be significant to the development and/or progression of SS.
Sézary syndrome (SS)3 is an advanced form of cutaneous T cell lymphoma (CTCL) that is characterized by a clonal expansion of CD4+/CD45RO+ T cells and the appearance of these malignant T cells in the circulation (1, 2). Malignant cells can be identified by their characteristic cerebriform morphology and their aneuploid cytogenetic profile. Our studies and those of other laboratories indicate that SS cells display a Th2-like phenotype characterized by production of high levels of IL-4, IL-5, and IL-10 and low levels of IL-2 and IFN-γ (3, 4, 5, 6). We have also observed that the production of IL-12, a potent inducer of IFN-γ and a pivotal cytokine for Th1 cell differentiation, is depressed in SS (7). Depressed IL-12 levels correlate with an inability to mount an effective Th1 response as reflected in immune abnormalities seen in these patients. Addition of exogenous IL-12 to cell cultures of PBL from SS patients can reverse many of the cytokine and immune abnormalities observed, including decreased IFN-γ production and decreased NK cell activity (8).
Two chains of the IL-12R have been cloned (9, 10, 11, 12), and a recent report suggests the presence of a third chain (13). Studies in both humans and mice suggest that the responsiveness to IL-12 of Th1 cells and the lack of responsiveness of Th2 cells, can be explained by the types of IL-12R chains expressed on the surface of these cells. Th1 cells express both the β1 and β2 IL-12R subunits, whereas Th2 cells express only the β1 subunit (12, 13). Although both Th1 and Th2 cells can bind IL-12, only Th1 cells are capable of signaling in response to IL-12, and a functional signal appears to require binding to a receptor composed of both β1 and β2 subunits (9, 11). IL-4 appears to account, at least in part, for suppressed β2 expression in Th2 cells, although IFN-γ can overcome this suppression under some conditions in the mouse system (14).
Several studies have shown that signaling by IL-12 occurs through the Jak/STAT signal transduction pathway with involvement of two members of the Janus family of tyrosine kinases, Jak2 and Tyk2, and the STAT proteins 1, 3, and 4 (15, 16). Binding of IL-12 to its receptor triggers the recruitment and activation of the Janus kinases, with Jak2 and Tyk2 interacting with the β2 and β1 subunits, respectively, and the subsequent association and tyrosine phosphorylation (pTyr) of the STAT proteins (17, 18, 19). Mice with a targeted disruption of the STAT4 gene have decreased responsiveness to IL-12, an impaired Th1 response, and a Th2 population that demonstrates markedly enhanced development (20, 21), characteristics similar to those associated with SS. Furthermore, only IL-12, the expression of which is depressed in SS patients, and IFN-α (22, 23, 24) have been shown to induce STAT4 phosphorylation in T cells derived from healthy individuals (17). This observation raised the possibility that the inability of SS patients to mount an effective IL-12-directed Th1 response might rest, at least in part, in their failure to phosphorylate and thus activate STAT proteins, in particular STAT4, the phosphorylation of which in response to IL-12 is dependent on the presence of the IL-12R β2 chain. IFN-α not only activates STAT4 but has also been reported to induce expression of the IL-12R β2 chain required for STAT4 phosphorylation in response to IL-12 in normal T cells (17).
We report here that expression of the IL-12R β2 gene was low to undetectable in PBL from six of the seven advanced SS patients tested. In vitro culture of SS cells with inducers of β2 expression including IFN-α, IFN-γ, or IL-12 also failed to stimulate β2 expression in these cells. We also show that SS T cells have significantly reduced or undetectable levels of STAT4 proteins, whereas nonmalignant CD8+ T cells from patients express normal levels of STAT4 protein that can be phosphorylated in response to IL-12.
Materials and Methods
IFN-α and IFN-γ were purchased from R&D systems (Minneapolis, MN), PHA was purchased from Sigma Chemicals (St. Louis, MO), and rhIL-12 was a gift from Dr. Stanley Wolfe (Genetics Institute, Amherst, MA).
Venous blood samples were collected from seven SS patients before commencing treatment with IFN-α and during or soon after IFN-α treatment as indicated (Table I). Circulating T cells of patients 1–6 were characterized as 80–99% SS cells as measured by 1-μm section analysis of peripheral blood buffy coats and by FACS analysis. Patient 7 had 35% SS and 15% CD8+ cells before IFN-α and 77% SS cells and 5% CD8+ cells 1 year later during IFN-α treatment.
|Patient .||Date .||IFN-α Treated .||% Malignant .|
|Patient .||Date .||IFN-α Treated .||% Malignant .|
Patient was off IFN-α therapy for several months.
No pre-IFN-α sample available.
Preparation and purification of T cell subpopulations
PBL were prepared as described (21). Briefly, venous blood was drawn from SS patients and normal healthy donors (all with informed consent) and collected into heparinized tubes. Blood was diluted 2-fold with Dulbecco’s PBS, layered over Ficoll-Hypaque (Pharmacia, Piscataway, NJ), and centrifuged for 30 min at 500 × g. The interface containing the mononuclear cell fraction was washed with Dulbecco’s PBS by centrifugation. Monocytes were depleted by adherence to gelatin-coated tissue culture flasks, and the PBL resuspended in complete medium at the indicated cell concentrations. PBL that were skewed to either Th1 or Th2 phenotypes were generated by culturing in RPMI 1640,10% FCS plus 5 μg/ml PHA in the presence of either IL-12 (1 ng/ml), and anti-IL-4 (1/5000 dilution of 4F2 ascites) for the Th1 cells or IL-4 (50 ng/ml) and anti-IL-12 (1/5000 of CB6 ascites) for the Th2 cells (25). Under these conditions, cultures are routinely two-thirds CD4+ and one-third CD8+, with virtually no contaminating monocytes, NK or B cells.
In certain experiments, CD8+ T cells were removed using BioMag immunobeads (PerSeptive Diagnostics, Cambridge, MA) according to the manufacturer’s protocol. Briefly, PBL were incubated with anti-CD8+ mAb (Immunotech, Westbrook, ME) for 30 min at 4°C. Cells were then washed to remove unbound Ab and combined with immunomagnetic beads coated with goat anti-mouse Fc-specific Abs (PerSeptive Diagnostics) at a bead-cell ratio of 25:1. Positively selected CD8+ T cells or clonal malignant T cells were recovered in a magnetic field (with a purity of >98%). The selection process was repeated on the cells not bound to the beads in this first round of separation to maximize yields. The remaining unbound cells typically contained <1% CD8+. Cells were either used immediately after purification or stored in liquid nitrogen in 95% FBS/5% DMSO until use.
Ribonuclease protection assay (RPA)
Total RNA was isolated from PBL with TRIzol reagent (Life Technologies, Grand Island, NY), resuspended in diethylpyrocarbonate-treated water, and stored at −20°C. RPA was conducted as described (26). Briefly, antisense RNA probes were prepared with the used of the T7 promoter in Bluescript (Stratagene, La Jolla, CA) or Sp6 promoter in PCRII (Invitrogen, San Diego, CA) and labeled with 400–800 Ci/mmol [32P]UTP or [32P]CTP. Probes were purified by extraction from a 6% acrylamide sequencing gel. Depending on availability, 2–10 μg of total RNA were hybridized with antisense probe (5 × 105 cpm) in 80% formamide buffer at 60°C for 16 h. The IL-12R β1 and β2 probes, generated from human cDNA clones, cover nucleotides 1752–2102 and 2069–2449, respectively (11). A human cyclophilin antisense probe (Ambion, Austin, TX) was used as an internal control for standardization of expression levels between samples. Samples were processed as described (26) and fractionated on a 6% sequencing gel. The dried gels were exposed to a PhosphorImager screen for 1–4 days, and the relative signals were quantitated with Image Quant (Molecular Dynamics, Sunnyvale, CA). Where comparisons were made between the individual messages, the values were standardized for C or U content of the protected fragments.
Determination of STAT protein and phosphorylation levels
Cells depleted of monocytes and/or CD8+ cells were cultured for 3 days in RPMI 1640/10% FBS containing (PHA, 2 μg/ml), washed twice with medium, starved for 20 h in RPMI 1640 containing 2% FBS, and stimulated with IL-12 (22 ng/ml), IFN-α (1000 U/ml), or IFN-γ (1000 U/ml) for 15 min at 37°C. After lysis in ice cold buffer (5 mM Na2EDTA, 50 mM NaCl, 30 mM Na4P2O7, 50 mM NaF, 0.1 mM Na3VO4, 20 mM Tris (pH 7.6), 1% Triton X-100, 0.5 μg/ml leupeptin, 0.7 μg/ml pepstatin A, 50 μg/ml PMSF), cell lysates were cleared by centrifugation and incubated with anti-STAT4 or anti STAT1 (C-20 or E23, respectively, Santa Cruz Biotechnology, Santa Cruz, CA) at 4°C for 60 min. Immune complexes were collected by incubation with protein A-agarose (60 min, 4°C), washed three times with lysis buffer, and heated for 4 min in a boiling water bath in sample buffer for SDS-PAGE. Before immunoprecipitation, aliquots were taken for protein assay (bicinchoninic acid assay, Pierce, Rockford, IL) and for Coomassie blue staining. Immunoprecipitates from extracts containing the same amount of protein were analyzed by Western blotting with Abs against pTyr (4G10 (Upstate Biotechnology, Lake Placid, NY) and PY20 (Transduction Labs, Lexington, KY)), STAT4 (L-18 (Santa Cruz Biotechnology)) or STAT1 (E23 (Santa Cruz Biotechnology)) by chemiluminescence (Renaissance (NEN-DuPont, Boston, MA)) with HRP-conjugated anti-Ig (Figs. 4 and 5). The blots were always analyzed for pTyr first and then stripped by incubation for 30 min at 50°C in stripping medium (2% SDS, 0.05 M Tris (pH 6.8), 0.1 mM β-mercaptoethanol). Stripping was confirmed before reanalysis with anti-STAT4 or anti-STAT1. Western blotting shown in Fig. 6 was conducted with alkaline phosphatase/5-bromo-4-chloro-3-indolyl phosphate-nitroblue tetrazolium for visualization. In this case identical samples were Western blotted and developed with either anti-pTyr or anti-STAT4.
SS T cells express little or no message for the IL-12R β2 subunit
Levels of IL-12R β1 and β2 subunit mRNA in PBL from 7 SS patients and four normal controls were determined by RPA. Samples for RNA were obtained either before or during IFN-α therapy, which involved three weekly injections for extended periods of time (Table I). Six of the seven patients had little or no detectable β2 message (Fig. 1). β2 message levels were markedly lower than in RNA extracted from freshly isolated control PBL which typically express low levels of β2 and were even lower than in the Th2-skewed normal control PBL. By contrast, β1 message levels were 3- to 5-fold higher in the SS samples than in normal PBL but were still significantly lower than in the Th1-skewed normal controls. Unlike β2 message levels which remained essentially unchanged before or during IFN-α therapy, β1 message levels increased from 30 to 50% in samples taken from several patients during IFN-α treatment, suggesting that some patient cells might be responding to IFN-α at the level of receptor gene expression (Fig. 1).
Both IFN-α and IFN-γ induce expression of IL-12R β2 in normal PBL but not in SS cells
To further compare the effects of biological response modifiers on IL-12R gene expression in normal controls and SS patients, freshly isolated PBL were cultured in the presence of either IFN-α, which has been reported to induce human β2 expression (27), or IFN-γ, which has been reported to induce β1 and β2 expression in humans (28) and mice (29). Freshly isolated normal control PBL had low but relatively similar levels of β1 and β2 mRNA, whereas no β2 message was detected in the RNA isolated from SS patient 1 (Fig. 2). Both β1 and β2 message levels increased 2- to 3-fold in control cells cultured in the presence of IFN-α for 24 h, but there was no significant increase in either β1 or β2 message levels in the SS cells even after 56 h.
The effect of IFN-γ on β1 and β2 message levels in control PBL was even greater than that of IFN-α, with 5- and 7-fold increases respectively (Fig. 2). These results are consistent with previous reports in mice and humans that IFN-γ enhances IL-12 responsiveness (29, 30). In contrast, β2 expression in SS cells was unaffected by IFN-γ and remained undetectable, although β1 message levels increased more than 4-fold by 18 h and remained elevated 56 h after addition of IFN-γ. Because the increases in the levels of β1 message were similar in both the control and patient RNA, β1 expression was likely induced in the patient’s SS cells in addition to the normal cells because 99% of the T cells of patient 1 were identified as SS cells by FACS analysis of clonal Vβ expression, and NK cell numbers were also very low. The inability to detect any increase in β2 message suggests that β2 transcription may be suppressed in the SS cells, although we cannot eliminate the possibility that increased message turnover is also involved. The extremely low number of normal T cells in this patient precluded determination of induced β2 expression in the nonmalignant T cells. The finding that IFN-α and IFN-γ induce both receptor genes in normal PBL supports the previous report that IFN-γ up-regulates IL-12 receptors and IL-12 responsiveness in freshly isolated human PBL (29), although no effect of IFN-γ on either β1 or β2 expression was detected in similar studies using human Th cell lines (27).
Neutralization of IL-4 in SS cell cultures fails to result in the induction of IL-12R β2 expression by either IFN-α or IL-12
SS is characterized by production of Th2 cytokines including IL-4, which has been shown to suppress β2 expression (14, 30, 31). However, neutralization of IL-4 activity in SS cultures with anti-IL-4 Abs (anti-IL-4) did not result in any detectable β2 message in the presence or absence of IFN-α (Fig. 3,A). By contrast, the induction of both β1 and β2 expression by IFN-α in normal control PBL was enhanced by the presence of anti-IL-4, with the greatest effect on β2 expression (Fig. 3 B).
Several recent findings raised the possibility that IL-12 could have some effect on the SS cells that was independent of β2. Early studies reported that IL-12 can up-regulate β2 expression on normal PBL by some unknown mechanism (27). In addition, a third prospective chain of the IL-12 receptor has been reported to be expressed in Jurkat cells. This protein appears to be able to interact with β1 expressed from a transfected cDNA to produce a high affinity IL-12 receptor. IL-12 appears to induce tyrosine phosphorylation of this chain even in the absence of the β2 chain, which is not expressed in Jurkat cells (13). Finally, cells transfected with only the β1 cDNA, which lacks a kinase activity, have been shown to phosphorylate Tyk-2. Therefore, the SS cells were also incubated with anti-IL-4 and rIL-12 (Fig. 3 A). However, anti-IL-4 had no detectable effect on β2 expression in the SS cells even when IL-12 was added.
STAT4 protein levels are undetectable in PBL from SS patients with high levels of SS cells
The biological effects of the IFNs and IL-12 on T cells are mediated through the Jak/STAT signal transduction pathways. In particular, signal transduction by both IFN-α and IL-12 proceeds in part, through the activation of STAT4 (17, 18, 32). Because activation of STAT4 is important for development of an effective Th1 response, we compared the ability of SS cells and PBL from normal healthy individuals to activate the STAT4 signal transduction pathway, as measured by tyrosine phosphorylation of STAT4, in response to IL-12 or IFN-α (Fig. 4).
Extracts for immunoprecipitations were prepared from cells of SS patients in whom >85% of the lymphocytes represented a single clonotypic TCR Vβ rearrangement and were thus derived from the identical transformed clone. STAT4 protein was essentially undetectable in the SS cell extracts from patient 2 (Fig. 4) or patient 1 (data not shown). In some cases, low levels of STAT4 protein could be detected if a 10-fold greater number of SS cells (2 × 108) were used to prepare the extracts (data not shown). We attribute the detection of STAT4 protein under these conditions to the small proportion of normal T cells present.
To determine whether the expression of other STAT proteins was also repressed in SS cells we examined protein extracts prepared from PBL from patients 1 and 2 for the expression of STAT1. STAT1 is required for signaling by both IFN-α and IFN-γ (33) and has also been reported to be activated by IL-12 in human T cells (28). Fig. 5 shows the results of the immunoblotting experiments, which clearly detect STAT1 protein and pTyr phosphorylation at similar levels in both patients 1 and 2 and the normal control.
The normal CD8+ T cells from SS patients with moderate levels of SS cells express STAT4 and phosphorylate it in response to IL-12
To determine whether the patient’s normal T cells also have reduced levels of STAT4, we prepared protein extracts from PBL of patient 7, in whom ∼15% of the T cells were normal CD8+ cells and ∼35% of the T cells were SS cells at the time of sampling. Immunoblotting revealed reduced levels of STAT4 protein as compared with the normal controls and greatly decreased, but none the less detectable levels of tyrosine phosphorylation of STAT4 in response to IL-12 (Fig. 5,A). To determine whether the reduced levels of STAT4 phosphorylation correlated with the lack of β2 expression on the SS cells, the analysis was repeated using extracts prepared from the patient PBL depleted of the normal CD8+ T cells. Indeed, STAT4 protein levels were greatly reduced by removal of the CD8+ T cells and tyrosine phosphorylation of STAT4 was no longer detectable under these conditions (Fig. 5,B), suggesting that most of the STAT4 protein and all of the STAT4 phosphorylation detected in these experiments could be attributed to the CD8+ T cells. To ensure that equivalent amounts of the normal and CD8+ depleted SS immunoprecipitates were Western blotted, the precipitated extracts were analyzed on an SDS-acrylamide gel by Coomassie blue staining (Fig. 5 C) to confirm the protein concentrations determined with bicinchoninic acid reagent (data not shown).
The present analysis of SS patients with similar levels of disease for 1) the expression of the IL-12R β1 and β2 genes and 2) the ability of SS cells to phosphorylate STAT4 in response to IL-12 represents an effort to understand the phenotypic characterization of these cells as compared with normal T cells. It also furthers our understanding of the mechanism of action of biological response modifiers, in particular IL-12 and IFN-α, that have potential therapeutic utility in the treatment of CTCL.
In six of our patients, the SS cells can be classified as Th2 T cells based on the relative absence of message for the IL-12R β2 chain, supporting previous studies that classify SS cells as Th2 cells based on cytokine secretion profiles (3, 4, 34, 35). β2 expression appears to be strongly suppressed in these cells because in vitro exposure to IFN-α and IFN-γ both of which induced β2 message in control PBL, was ineffective on SS cells. SS cells do express the GATA-3 protein (data not shown), which has been suggested to suppress Th1 differentiation (36, 37). However, IFN-γ, which was previously reported to partially overcome the suppression of β2 expression in normal mouse Th2 cells (14), did induce significant levels of β1 message in the SS cells but not β2, suggesting that factors in addition to GATA-3 may be involved. The induction of β1 in the SS cells by IFN- γ further suggests that signal transduction pathways for induction of IL-12R β1 and β2 expression by IFN- γ are independent at some level. Indeed, the synergy demonstrated between IFN-γ and IL-12 in activating cytotoxic activity of the normal CD8+ T cells against autologous SS cells demonstrated by Seo et al. (38) might reflect, in part, the induction of IL-12R expression in the CD8+ T cells by IFN-γ, consequently enhancing their responsiveness to IL-12.
The apparent lack of β2 chain expression by most SS cells is consistent with their characterization as Th2 cells. However, normal Th1 and Th2 cells, in both humans (27) and murine models (19, 31) have similar levels of STAT4 protein, although we find that STAT4 levels in the Th2-like SS cells are essentially undetectable as compared with their normal counterparts. By contrast, STAT1 protein levels appear to be normal in SS cells; therefore, the lack of STAT4 that we detect is not symptomatic of a more general effect on all STAT proteins. These results suggest that SS cells may have the capacity to signal through the unaffected STAT pathways and raise the question as to whether some SS cells could be susceptible to the growth-inhibiting effects of the IFNs. Understanding at what point in the progression of SS the STAT4 expression is lost, whether a similar loss of STAT4 is associated with other forms of CTCL, and how the loss of STAT4 correlates with the loss of β2 expression may be an important element in understanding the development and progression of this disease and in designing treatments for its control.
The studies described here also clearly show that the normal patient CD8+ cells retain the capacity to respond to IL-12 by phosphorylating STAT4 and thus express functional IL-12 receptors. The phosphorylation of STAT4 detected in extracts of SS cultures with significant numbers of normal CD8+ T cells supports our previous report that IL-12 can partially reverse the Th2 cytokine profile of PBL isolated from CTCL patients. The increases in IFN-γ production and NK cell activity detected in these studies (7) was likely due to the effects of IL-12 on the normal T cells present in these preparations. These hypotheses are further supported by our in vitro data (39) as well as those of others (40) derived from clinical trials utilizing recombinant IL-12 which suggest that CD8+ cytotoxic T cells are activated in patients receiving IL-12 therapy. Thus, IL-12 can mediate immune-augmenting effects on the nonmalignant cells of SS patients, and this may lead to a profound clinical benefit.
The cells of patient 5 expressed significantly higher levels of β2 message than those of the other six patients, suggesting that manifestation of the Th2 phenotype may not absolutely depend on silenced β2 expression. On the other hand, the ratio of β1 to β2 message in this patient was 6:1, comparable to with ratios we find in PBL cultures skewed to the Th2 phenotype by growth in the presence of IL-4 and anti-IL-12, and unlike the 1:1 to 2:1 ratios in cultures skewed to the Th1 phenotype (see Fig. 1). Using enhanced chemiluminescence, we could detect some STAT4 expression in this patient but not phosphorylation of STAT4 (data not shown), suggesting that signal transduction by IL-12 through STAT4 is blocked. Perhaps the levels of β2 expression as compared with that of β1 are too low to assemble significant numbers of functional receptors. Whether β2 expression is a unique characteristic of this particular patient or whether it is a marker that may be significant for classification of a SS subtype will require the identification and characterization of additional β2-expressing SS patients.
We thank Drs. A. Caton and G. Rovera for critically reading the manuscript, Ms. Marina Hoffman and the Wistar Editorial Department for editing and preparing the manuscript, Dr. D. Peritt for Th1 and Th2 RNA controls, and Drs. M. Kobayashi and S. Wolfe for rIL-12.
This work was supported by the Leukemia Society of America, the Elsa Pardee Foundation, and the Ruth Estrin Goldberg Memorial Foundation. K.A. was supported by the Federal Work Study Program.
Abbreviations used in this paper: SS, Sézary syndrome; CTCL, cutaneous T cell lymphoma; RPA, RNase protection assay; pTyr, phosphotyrosine.