Absence of TH2 Cytokine Messenger RNA Expression in CD30-NegativePrimary Cutaneous Large T-Cell Lymphomas FREE

CUTANEOUS T-cell lymphomas (CTCLs) are a heterogeneous group of T-celllymphoproliferative disorders that clinically originate in the skin.¹ In most CTCLs, the neoplastic cells have the phenotypeof skin homing CD3⁺, CD4⁺, CD8⁻, andCD45RO⁺ memory T cells. In mice 2 distinct subsets of CD4⁺ T lymphocytes are distinguished on the basis of different patternsof cytokine secretion, ie, T_H1 cells producing interleukin 2 (IL-2),interferon γ (IFN-γ), and tumor necrosis factor α (TNF-α)and T_H2 cells producing IL-4, IL-5, and IL-10.² The cytokines produced by these 2 T-cell subsets have opposing effects. Thus,IL-2 and IFN-γ stimulate the proliferation of T_H1 cells butinhibit the proliferation of T_H2 cells, whereas IL-4 stimulatesthe proliferation of T_H2 cells and both IL-4 and IL-10 inhibitT_H1 cell growth and function.^{3 - 4} Based on these observations in murine CD4⁺ T cells, several groupshave started to evaluate cytokine profiles of the neoplastic CD4⁺T cells in CTCL. Studies in Sézary syndrome (SS), a leukemic form ofCTCL, demonstrated increased IL-4 and IL-5 and low IL-2 and IFN-γ productionby peripheral blood mononuclear cells^{5 - 6} and increased expression of IL-4 and IL-5 messenger RNA (mRNA) within lesionalskin.⁷ Subsequent studies in lesional skinof mycosis fungoides (MF), the most common subtype of CTCL, demonstrated thepresence of IL-4 and IL-5 mRNA in tumor stage MF but not in the early patchstage, whereas IL-2 and IFN-γ were detected in all stages.⁸ These observations resulted in the concept that the neoplastic T cells inMF and SS produce IL-4 and IL-5 (T_H2 cells), whereas the reactiveinflammatory cells are the source of IL-2 and IFN-γ (T_H1cells). This model implies that in the early stages of MF, in which the neoplasticT cells represent only a minority of the cellular infiltrate, the expansionof neoplastic T_H2 cells is inhibited by IL-2– and IFN-γ–producingreactive T_H1 cells. With progression of disease, as the malignantcell population expands, increased production of IL-4 may impair the T_H1 cell–mediated host antitumor response.⁹

This concept of CTCL as neoplasms of IL-4– and IL-5–producingT_H2 cells also gave an explanation for various immune abnormalitiesassociated with the advanced stages of MF and SS and provided a rationalefor treatment with biologic response modifiers as IFN-α, retinoids,IFN-γ, and IL-12, aimed at potentiating a T_H1 antitumor response.^{10 - 13}

From a clinical point of view, it is important to know whether thisnew immunopathogenic concept and, more important, the therapeutic consequencesderived from it also hold true for CTCLs other than MF and SS. In this respect,CD30⁻ primary cutaneous large T-cell lymphomas (PCLTCLs)are the most important group, since these lymphomas do not or insufficientlyrespond to currently available therapies, including multiagent chemotherapy(5-year survival rate of 15%).¹ In the presentstudy, patients with a CD30⁻ PCLTCL were investigated forthe expression of T_H1 and T_H2 cytokines using reversetranscription–polymerase chain reaction (RT-PCR). Skin biopsy specimensfrom patients with plaque or tumor stage MF, SS, psoriasis, lichen planus,and atopic dermatitis were included as controls.

Seven skin biopsy specimens were obtained from 4 patients with a CD30⁻ PCLTCL. At the time of diagnosis, staging procedures did notreveal any evidence of extracutaneous disease. From 2 patients additionalbiopsy specimens obtained during follow-up were available for examination.In all biopsy specimens the histologic findings showed a diffuse infiltrationof medium and large pleomorphic T cells. Detailed immunohistochemical studiesdemonstrated that in all cases the neoplastic T cells made up more than 90%of the infiltrate; CD8⁺-reactive T cells, CD1a⁺ dendriticcells, CD68⁺ macrophages, and CD20⁺ B cells were fewor absent and together never exceeded 10% of the total number of infiltratingcells. Despite multiagent chemotherapy, all patients had diseases that rana progressive clinical course and all patients died of systemic lymphoma 6to 55 months (median, 8 months) after diagnosis (Table 1).

Biopsy specimens from 8 patients with CTCL, including plaque stage MF(n = 3), tumor stage MF (n = 2), and SS (n = 3), were included as controls.The diagnoses in these cases were based on clinical, histological, and immunophenotypiccriteria, as described previously.¹ In allCTCL, except for the patients with SS, extensive staging procedures had failedto demonstrate extracutaneous disease at the time of presentation. In addition,skin biopsy specimens from untreated skin lesions of patients with psoriasis(n = 5), lichen planus (n = 2), and atopic dermatitis (n = 3) were includedas benign control groups.

Four-millimeter punch biopsy specimens were snap frozen in liquid nitrogenand stored at −196°C until use. Total RNA was extracted from ten10-µm cryostat sections using RNAzol B (Campro Scientific, Veenendaal,the Netherlands). The integrity and amount of isolated RNA were verified byrunning one fifth of the isolated RNA on a 1.5% Tris-borate-EDTA (TBE) agarosegel stained with SYBR green II RNA gel stain (FMC, Rockland, Me). Next, afterdenaturing for 10 minutes at 70°C, half of the remaining RNA was reversetranscribed by incubating for 60 minutes at 42°C in a cocktail containing200 U of Superscript II reverse transcriptase (Gibco-BRL, Breda, the Netherlands),deoxynucleotide triphosphates (10mM each), 0.5 µg of oligo d(T12-16)(Gibco-BRL), and 10mM dithiothreitol in a total volume of 20 µL. Asa negative control, the other half of the isolated RNA was incubated withan identical cocktail but lacking Superscript II reverse transcriptase. Immediatelyfollowing reverse transcription, samples were diluted to 100 µL withdeionized water. A total of 5 µL of this diluted complementary DNA (cDNA)was amplified using PCR with sense and antisense primers selected with theprimer selection program PC Gene (Table2). To prevent amplification of genomic DNA, only intron-spanningprimers were used. The integrity of cDNA was verified by including controlamplification of cDNA encoding U1A (U1 small nuclear ribonucleoprotein–specificprotein A). Because the U1A protein regulates the production of its own mRNA,an advantage of this control is the equal and low abundant expression in mosttissues.¹⁴ Polymerase chain reaction was carriedout using 0.2 U of SuperTaq (Sphaero Q, Leiden, the Netherlands) in a buffersupplied by the manufacturer with 2mM magnesium chloride, 500µM deoxynucleotidetriphosphate, 10 pmol of each of the sense and antisense primers, and deionizedwater to a total volume of 25 µL. The cDNA was amplified in a thermocycler(Omnigene; Hybaid, Middlesex, England) for 40 cycles, where a single cycleconsisted of 94°C for 40 seconds, 60°C for 50 seconds, and 72°Cfor 60 seconds. Before cycling, samples were denatured for 5 minutes at 94°C,and after cycling, an extra incubation for 5 minutes at 72°C was performed.To avoid carryover contamination, strict physical and procedural precautionswere observed. Furthermore, a negative control water blank was included ineach PCR amplification experiment. Following amplification, a 10-µLsample of the PCR product was size analyzed on a 1.5% agarose TBE gel, stainedwith ethidium bromide, and compared with molecular weight markers. To confirmthe identity of the PCR product, samples were blotted on Qiabrane membranes(Diagen, Dusseldorf, Germany) and hybridized with specific phosphorus 32 end-labeledoligonucleotide probes (Table 2).On each sample of cDNA, all PCRs and hybridization of PCR products were performedin duplicate.

Initially, 4 cases containing an almost pure population (>90%) of largepleomorphic T cells were investigated (Table 3). Based on the concept that CTCL represents a proliferationof CD4⁺ T_H2 cells, a strong IL-4 mRNA expression wasexpected. However, IL-4 mRNA could not be demonstrated in any of these 4 biopsyspecimens and IL-5 mRNA was demonstrated in only 1 of these (Figure 1). Because of these negative results, 3 additional biopsyspecimens were analyzed from 2 of these patients (patients 2 and 3), eitherat the same time (patient 3) or during progression at 4 and 5 months afterdiagnosis (patient 2). Also, in these additional biopsy specimens, IL-4, IL-5,and IL-10 mRNA were not detected, whereas the presence of U1A confirmed thepresence of intact mRNA. Expression of the T_H1 cytokines IFN-γand IL-2 was detected in 5 of 7 and 4 of 7 biopsy specimens, respectively.

In plaque stage MF, IFN-γ and IL-2 mRNA were detected in 3 of3 and 2 of 3 cases, respectively, whereas IL-4 and IL-10 mRNA were not found(Table 4). In contrast, both casesof tumor stage MF expressed IFN-γ, IL-2, IL-4, and IL-10, whereas IL-5was expressed in 1 of 2 cases. Thus, IL-4 was only found in MF lesions showinga predominance of neoplastic T cells. In the 3 skin biopsy specimens frompatients with SS, IL-4 and IL-5 mRNA were demonstrated in 2 and 3 cases, respectively(Table 4).

In psoriasis, which is widely considered a T_H1-mediated disorder,¹⁵ IFN-γ and IL-2 were detected in all 5 biopsyspecimens, whereas IL-4 and IL-5 were found in 1 of 5 and 1 of 3 cases, respectively(Table 4). Interleukin 10 mRNAwas detected in 4 of 5 psoriatic skin lesions. In atopic dermatitis, IL-2,IFN-γ, and IL-5 were found in 3 of 3 biopsy specimens. Interleukin 4was detected in 2 of 3 biopsy specimens, whereas IL-10 mRNA–derivedcDNA could not be detected (Table 4).

In the present study, we investigated T_H1 and T_H2cytokine profiles in CD30⁻ PCLTCL. Because previous studiesin MF and SS suggested that the neoplastic T cells in CTCL are derived fromT_H2-producing CD4⁺ T cells and all biopsy specimenscontained more than 90% neoplastic T cells, expression of IL-4 and IL-5 mRNAwas expected.

However, in the initial skin biopsy specimens of all 4 patients witha CD30⁻ PCLTCL, IL-4 mRNA was not detected, whereas IL-5mRNA was detected in only 1 of 4 biopsy specimens. These results suggestedthat the neoplastic T cells in these CD30⁻ PCLTCLs do notproduce T_H2 cytokines. Additional evidence for this conclusionis provided by the absence of IL-4 and IL-5 mRNA in 3 additional biopsy specimensfrom 2 of these 4 patients, as well as the presence of IL-4 mRNA in controlbiopsy specimens, run in parallel, from patients with tumor stage MF (2/2),SS (3/3), and atopic dermatitis (2/3), which is consistent with the resultsof recent literature.^{8 ,16} Moreover,all PCR and hybridizations of PCR products were performed in duplicate, andcontrols for the integrity of isolated mRNA and transcribed cDNA were positivein all cases. The expression of IFN-γ and IL-2 found in 5 of 7 and 4of 7 biopsy specimens, respectively, may be attributed to a few scatteredCD8⁺ T cells (always less than 5%), but it cannot be excluded thatthese cytokines are produced by the tumor cells.

In SS and advanced MF, the production of IL-4 and IL-5 has been associatedwith a constellation of immune abnormalities, such as an increased serum IgElevel, decreased T-cell response to antigens, impaired cellular cytotoxicity,and peripheral eosinophilia.^{9 ,17} It is of interest that increased IgE levels and eosinophilia are generallynot observed in CD30⁻ PCLTCL and consistently were not presentin the 4 patients studied. In addition to these immune abnormalities, IL-4and IL-5 have been considered to be responsible for the more aggressive clinicalbehavior of SS and advanced-stage MF by impairing T_H1 cell–mediatedantitumor responses.⁹ On the other hand, theproduction of IFN-γ in early stage MF by the reactive T-cell infiltratecan be an important factor responsible for the indolent course of early patchand plaque stage disease by inhibiting IL-4 production. Thus, in this modelcompeting T_H1 and T_H2 cytokine effects may be importantin disease progression of MF and SS.

This concept may explain at least in part the beneficial effects ofIFN-α and retinoids, both of which have T_H1 response–augmentingactivities, in the treatment of MF and SS.^{10 ,18 - 20} In addition, it provides a rationale for treatment with still experimentalbiologic response modifiers, such as IFN-γ and IL-12, the mean inducerof IFN-γ.^{11 - 12} In vitrostudies have already demonstrated that the excess IL-4 production in peripheralblood mononuclear cells from patients with SS can be inhibited by IL-12, IFN-γ,and IFN-α.^{5 ,12} In phase2 studies with IFN-γ, partial responses were observed in approximately30% of patients with CTCL.¹¹ In a phase 1 studyin MF and SS, subcutaneous IL-12 resulted in complete or partial responsesin 4 of 5 MF plaques, 2 MF tumors, and 1 of 2 patients with SS.¹³ In addition to an inhibitory effect on T_H2 cells, IFN-γ mayinduce the production of the CXCR3-targeting chemokines IFN-inducible protein10 (IP-10), monokine induced by IFN-γ (MIG), and IFN-inducible protein9/IFN-inducible T-cell α-chemoattractant.^{21 - 23} These chemokines specifically attract CXCR3-bearing activated T cells^{24 - 25} and are considered to play an importantrole in the antitumor responses.²⁶ In accordance,both biopsy specimens in which IFN-γ was not detectable by RT-PCR werealso negative for IP-10 as determined by in situ hybridization.²³

Because recent studies demonstrated that p53 protein can downmodulateIL-4 gene expression²⁷ and overexpression ofp53 protein was found on neoplastic cells in 4 of 8 CD30⁻PCLTCLs,²⁸ the expression of p53 could be afactor in the absence of IL-4 in these lymphomas.

Expression of IL-10 in CTCL is of interest because in MF an increasedexpression of IL-10 mRNA is associated with tumor progression.²⁹ However, in this study IL-10 mRNA was detected in only 1 of 7 CD30⁻ PCLTCL biopsy specimens, making it unlikely that production of IL-10is an important mechanism in the pathogenesis of CD30⁻ PCLTCL.

Previous studies established that in the group of PCLTCLs, expressionof the CD30 antigen on most tumor cells is the most important prognostic parameter.Thus, whereas CD30⁺ PCLTCLs have an excellent prognosis (5-yearsurvival rate of >90%), the prognosis of CD30⁻ PCLTCL ispoor (5-year survival rate of <15%). The molecular and genetic mechanismsunderlying these differences in clinical behavior are as yet unexplained.A study on a small number of CD30⁺ PCLTCLs suggested that thistype of CTCL is characterized by production of IL-4 and IL-10.³⁰ Our observations demonstrating lack of T_H2 cytokine mRNA expressionin CD30⁻ PCLTCL provide another biological difference withCD30⁺ PCLTCL and suggest differences in the regulation of tumorcell proliferation by the cytokine network.

Cytogenetic studies in CTCL have identified structural chromosomal abnormalitiesin MF, including 1p, 2p, 6q, 10q, and gene alterations in p16^INK4a.^{31 - 36} Multiple chromosomal abnormalities and alterations of p16^INK4awere found to be associated with tumor progression and a poorer prognosis.In contrast, no data are available on the genetic alterations in CD30⁻ PCLTCL, and investigations in this field are clearly warranted.

In conclusion, the results of the present study suggest that, in contrastto SS, MF, and CD30⁺ PCLTCL, the neoplastic T cells of CD30⁻ PCLTCL do not or rarely express IL-4 and IL-5 mRNA and thusdo not display a T_H2 cytokine profile. This observation contrastswith the current concept of CTCLs as lymphomas producing IL-4 and IL-5 anddemonstrates heterogeneity of the cytokine profile in CTCL.

Accepted for publication February 6, 2001.

We thank K. Thestrup-Pedersen, MD, PhD, for critical reading of themanuscript and helpful discussion.

Corresponding author: Maarten H. Vermeer, MD, Department of Dermatology,LUMC, Albinusdreef 2, 2300 RC Leiden, the Netherlands (e-mail: m.h.vermeer@lumc.nl).