Analysis of STAT4 expression in cutaneous T-cell lymphoma (CTCL) patients and patient-derived cell lines

Abstract

Deregulation of STAT signaling has been implicated in the pathogenesis for a variety of cancers, including CTCL. Recent reports indicate that loss of STAT4 expression is an important prognostic marker for CTCL progression and is associated with the acquisition of T helper 2 cell phenotype by malignant cells. However, little is known about the molecular mechanism behind the downregulation of STAT4 in this cancer. In the current work we test the expression of STAT4 and STAT6 via RT-PCR and/or Western Blot in CTCL lesional skin samples and in immortalized patient-derived cell lines. In these malignant cell lines we correlate the expression of STAT4 and STAT6 with the T helper (Th) phenotype markers and test the effect of Histone Deacetylase (HDAC) inhibitors and siRNA-mediated knock down of miR-155 on STAT4 expression. Our findings demonstrate that STAT4 expression correlates with Th1 phenotype, while STAT6 is associated with the Th2 phenotype. Our results further document that STAT4 and STAT6 genes are inversely regulated in CTCL. Treatment with HDAC inhibitors upregulates STAT4 expression, while at the same time decreases STAT6 expression in MyLa cells. Also, siRNA-mediated knock down of miR-155 leads to upregulation in STAT4 expression in MyLa cells. In summary, our results suggest that loss of STAT4 expression and associated switch to Th2 phenotype during Mycosis Fungoides progression may be driven via aberrant histone acetylation and/or upregulation of oncogenic miR-155 microRNA.

Keywords:

• Cutaneous T-Cell Lymphoma (CTCL)
• Mycosis Fungoides (MF)
• mi-R155
• Sézary Syndrome (SS)
• STAT6 and STAT4

Abbreviations:

CTCL	=	cutaneous T-cell lymphoma
MF	=	mycosis fungoides
SS	=	sézary syndrome
HDAC	=	histone deacetylase
STAT	=	signal transducers and activators of transcription

Introduction

Cutaneous T-Cell Lymphoma (CTCL) is a rare, but potentially devastating malignancy. Mycosis Fungoides (MF) and its leukemic form, Sézary Syndrome (SS), are the most common variants and account for ∼50% of all cutaneous lymphomas.¹ In Caucasians MF/SS primarily affects individuals over 55 years of age, while in African-Americans, Hispanics and Arabic individuals this disease presents at a significantly younger age (i.e. 20s and 30s).^2-4 Multiple epidemiologic studies documented that the incidence of CTCL has increased by ∼3 fold in the last 25–30 years.^4,5

In the early disease stages, which can last for several years, MF presents as flat erythematous skin patches resembling benign inflammatory diseases, whereas in the later stages, MF cells gradually form plaques or tumors and may disseminate to the lymph nodes and internal organs.¹ The early stages of CTCL are often difficult to distinguish clinically and even histologically from other benign entities including chronic eczema, psoriasis and pityriasis rubra pilaris.⁶ In advanced disease, malignant cells may appear in the peripheral blood, leading to the leukemic stage of CTCL. A subset of leukemic CTCL, known as SS, is characterized by a triad of erythroderma, lymphadenopathy and detection of malignant T cells with convoluted/cerebriform nuclei on a peripheral blood smear.¹ The life expectancy of SS patients is usually less than 3 years.¹

The molecular pathogenesis of CTCL remains only partially understood. Recent reports elucidated the nature of cancer initiating cells for MF and SS.⁷ Multiple studies attempted to clarify the genetic multistep carcinogenesis of CTCL.^8-11 Persistent activation of transcription factors of the signal transducers and activators of transcription (STAT) protein family has been implicated in the pathogenesis of a wide variety of human cancers, including CTCL. Expression and function of STAT3, STAT4 and STAT5 have been extensively studied in CTCL and these genes appear to play an important role in the disease pathogenesis and can be used as important prognostic markers.

Upregulation of STAT5 signaling occurs in the early stages of CTCL.^12,13 A growing body of experimental evidence suggests that this gene is important for expression of anti-apoptotic proteins (bcl-2 and bcl-x), cell cycle genes (Cyclin D and c-myc) and oncogenic miR-155 microRNA,^14-16 which work together to promote cancerogenesis. miR-155 was dubbed as the “bridge between inflammation and cancer”.¹⁷

In the early stages of CTCL, STAT4 appears to be overexpressed, when compared to non-malignant skin samples.¹⁸ The expression of STAT4, which is required for T helper (Th) 1 differentiation, is subsequently lost at the later stages, where the disease acquires predominantly Th2 phenotype.^19-23 In fact, in SS loss of STAT4 expression appears to be a robust and reliable diagnostic marker for this cancer.²¹

Upregulation of STAT3 signaling has been consistently documented in advanced stages of CTCL.^24-28 During late stages autocrine IL-21 stimulation was shown to be responsible for constitutive activation of STAT3.²⁹ Furthermore, others have suggested that malignant T cells may acquire completely cytokine-independent constitutive activation of JAK3 signaling, which can drive a constitutive activation of STAT3.^24,25 Constitutively active STAT3 can increase survival and resistance to apoptosis in malignant T cells, promote sensitization to IL-2 stimulation,^24,29 promote Th2 phenotype in advanced disease and induce expression of miR-21 oncogenic microRNA.³⁰ STAT3 signaling in CTCL was also shown to upregulate VEGF,³¹ the anti-inflammatory cytokine IL-10³² and a suppressor of cytokine signaling-3 (SOCS).³³ The latter was demonstrated to confer resistance to IFN-α in malignant T cells.^32,33

Another important concept that has emerged in recent years is that the phenotype of malignant T cells is not fixed and this cancer exhibits a phenotypic plasticity that is highly sensitive to cytokine milieu in the tumor microenvironment.³² Based on the above described evidence, STAT signaling appears to play a central role in this malignant phenotype switching.

Results

Correlation of STAT 4 and 6 expression with Th1, Th2, Treg, Th9, and Th17 phenotypes

Previous reports documented that many STAT members regulate Th differentiation phenotypes and determine cytokine expression in CTCL malignant cells.^34,35 Hence, we wanted to test the expression of Th1 (IFN-γ, IL-2, T-bet and IL-12A) vs. Th2 (IL-4 and IL-5) vs. Th9 (IL-9) vs. Th17 (IL-17A and IL-17F) vs. Treg (TGF-β1 and FOXP3) markers in these cells (Table S1). As demonstrated in Fig. 1A and Fig. S1, most cell lines spontaneously do not express or only weakly express Th1 markers. However, T cell stimulation with PMA (phorbol 12-myristate 13-acetate) and ionomycin or with CD3/CD28 Dynabeads® in select cell lines (e.g. Hut78, HH, H9, SZ4, Sez4 and Hut102) led to an upregulation of a number of Th1 genes (Fig. 1A; Fig. S1). In these cell lines the expression of Th1 markers correlated with the expression of STAT4 mRNA.

Figure 1. (A) Correlation of STAT4 and STAT6 expression with different T helper phenotype markers in CTCL cell lines under normal control vs. T cell stimulation culturing conditions (e.g., 10 ng/mL of PMA (phorbol 12-myristate 13-acetate) and 1 (M of ionomycin or with CD3/CD28 Dynabeads®). (B) Western blot analysis of STAT4 and STAT6 expression in patient-derived CTCL cell lines under normal control vs. T cell stimulation culturing conditions (i.e., PMA and ionomycin)

On the other hand, most cell lines heterogeneously express to some degree IL-4 and IL-5 Th2 markers and in a number of cases such expression corresponded to STAT6 expression (Fig. 1A; Fig. S1).

Western blot analysis confirmed the expression of STAT4 and STAT6 genes on the protein level (Fig. 1B). Specifically, cell lines that spontaneously express high levels of STAT6 (e.g., MyLa, PB2B, MaC2A, MJ and Hut102) did not express or only weakly expressed STAT4. On the other hand, cell lines that strongly express STAT4 (e.g., Sez4 and SeAx) did not express STAT6 under non-stimulated conditions.

IL-9 (Th9 marker) was only detected in MyLa cells upon stimulation with CD3/CD28 Dynabeads®, while most CTCL cell lines were IL-9-negative. Consistent with previous reports,^32,36,37 MyLa, PB2B, SZ4, Sez4 Hut102 and SeAx cells express IL-17F, IL-17A or both. Furthermore, these cytokines were further upregulated upon T cell stimulation (Fig. 1; Fig. S1). With respect to Treg phenotype, most malignant cells expressed TGF-β1 to a variable degree, but only MJ, Mac2A, SZ4 and SeAx SS-derived cells expressed detectable FoxP3.

Moreover, Fig. 1 demonstrates that in immortalized malignant T cells the molecular definitions between different T helper subtypes are not strictly followed. Unlike, normal T cells, malignant T cell demonstrate molecular overlap across a number of Th phenotypes.

Effect of Histone Deacetylase (HDAC) inhibitors on STAT expression in MyLa patient-derived malignant cells

Recently, epigenetic changes became a significant focus of basic and clinical research in CTCL. A number of previous studies documented methylation/histone acetylation abnormalities in malignant CTCL cells.^38,39 In fact, 2 of the commonly used medications for advanced stages of this cancer are HDAC inhibitors (Romidepsin and Vorinostat).⁴⁰ Hence, we wanted to test whether histone acetylation mediates the expression of STAT4 and STAT6 thereby potentially affecting the balance between Th1 vs. Th2 phenotype in CTCL. To achieve that we treated MyLa cells for 24 hours with HDAC inhibitors, Romidepsin and Suberoylanilide Hydroxamic Acid (SAHA also known as Vorinostat). MyLa is a commonly used cell line derived from an MF patient skin biopsy.⁴¹ As summarized in Table S2, out of 11 cell lines, this line is the only one derived from a skin biopsy of a typical MF patient. At baseline MyLa cells have very weak STAT4 expression and moderate STAT6 expression, which correlates with the advanced disease phenotype (Fig. 1). As demonstrated by RT-PCR, treatment with Vorinostat and Romidepsin upregulated the expression of STAT4 in these cells (Fig. 2A), while concomitantly downregulating the expression of STAT6 (Fig. 2B). Hence, this finding suggests that there is an inverse relationship in the expression of these genes, which is in part regulated by histone acetylation.

Figure 2. (A and B) Effect SAHA (Vorinostat) and Romidepsin HDAC inhibitors on STAT4 and STAT6 mRNA expression in MyLa cells. (C) Treatment with siRNA targeting miR-155 restores STAT4 expression in MyLa cells.

siRNA-mediated knock down of miR-155 leads to upregulation of STAT4 in MyLa cells

Oncogenic miR-155 microRNA was recently documented to play an important role in CTCL pathogenesis and was found to be upregulated in advanced stages of this cancer.^15,42 Using Bielefeld University Bioinformatic Server database we performed a screening analysis for matches in the STAT4 3'UTR, which identified a putative miR-155 binding site (position 2560 to 2690 of STAT4 mRNA; NM_003151.3). Hence, we wanted to test whether siRNA–mediated knock down of miR-155 expression will lead to the upregulation of STAT4 mRNA. As documented in Fig. 2C, downregulation of miR-155 results in restoration of STAT4 expression in MyLa CTCL malignant cells. This suggests that the observed disease-associated upregulation of miR-155 may be responsible for downregulation of STAT4 and subsequent loss of Th1 phenotype in malignant T cells.

Comparison of STAT4 expression between CTCL lesional skin, normal skin from healthy volunteers and skin from patients with benign inflammatory dermatoses

Previous reports suggested that expression of certain STAT genes may have diagnostic and prognostic value in CTCL patients.^15,16,21 Hence, we wanted to compare the expression of STAT4 by RT-PCR between CTCL lesional skin, normal skin samples from healthy volunteers and biopsy specimens from patients affected by benign inflammatory dermatoses that often mimic CTCL (e.g., psoriasis, chronic eczema and pityriasis rubra pilaris or PRP). As demonstrated in Fig. 3A, STAT4 was preferentially expressed in CTCL lesional skin, but not in normal skin or skin from patients with benign inflammatory dermatoses. It is of interest that upregulation of STAT4 was previously observed in CTCL lesional skin samples,¹⁸ when compared to non-malignant inflammatory dermatoses.

Figure 3. (A) STAT4 is preferentially expressed in CTCL lesional skin compared to benign inflammatory dermatoses or normal skin from healthy volunteers. (B) Within the CTCL skin samples STAT4 low expression is observed in patients with progressive disease. (C) Kaplan-Meier analysis documented that loss of STAT4 expression is associated with poor disease prognosis.

Further analysis of STAT4 expression in our patient samples documented that this gene was more strongly expressed in patients with stable disease and is downregulated in patients with progressive/advanced disease (Fig. 3B). Using X-Tile software⁴³ we established a cut off <0.02 fold expression for low STAT4 expression and ≥0.02 cut off for high STAT4 expression and documented using Kaplan-Meier analysis (Fig. 3C) that loss of STAT4 expression is associated with progressive disease (Monte Carlo p value = 0.02; Miller-Seigmund p value = 0.0059). This finding is consistent with previous reports demonstrating that loss of STAT4 occurs as malignant cells shift from Th1 toward Th2 phenotype in advanced disease stages.^19-23

Discussion

In the current work we document that the expression of STAT4 correlates with the expression of Th1 markers in immortalized malignant CTCL cells, while STAT6 expression is associated with the Th2 phenotype. As summarized in Suppementary Table 2, most cell lines were derived from CTCL patients with advanced disease. Hence, our findings that the majority of tested cells (e.g., MyLa, PB2B, Mac2a, MJ and Hut102) strongly expressed STAT6 and corresponding Th2 markers, but not STAT4 and Th1 markers, are consistent with this observation. The expression of IL-17 in immortalized cells was addressed in previous experimental studies^36,44 and may correlate with advanced disease in patients.^37,44,45

Our experimental work further suggests that aberrant histone acetylation in CTCL may result in a switch from STAT4-driven Th1 phenotype toward STAT6-associated Th2 phenotype in malignant cells and HDAC inhibitors may be able to restore this balance and upregulate STAT4 expression. More importantly our siRNA knock down experimental data demonstrates that STAT4 is one of the targets of miR-155 and inhibition of this oncogenic microRNA results in upregulation of STAT4. Hence, previously observed upregulation of miR-155 in advanced CTCL stages⁴² may be responsible for downregulation of STAT4 in CTCL lesional skin.

STAT4 was previously shown to be essential for Th1 differentiation and is typically upregulated via IL-12 signaling.⁴⁶ STAT4 binding to IFN-γ promoter/enhancer region is required for IFN-γ transcription.⁴⁷ As a result, STAT4 ^−/- CD4⁺ T cells produce low levels of IFN-γ⁴⁷. Mice with a targeted disruption of the STAT4 gene have decreased responsiveness to IL-12, an impaired Th1 response and skewing toward markedly enhanced Th2 development.⁴⁸ Hence, it is not surprising that in advanced stages of CTCL downregulation of STAT4 appears to play a key role in a subsequent shift from Th1 toward Th2 disease phenotype.

In summary, STAT signaling appears to play a central role in the pathogenesis of this cancer where early in the disease STAT5 upregulation drives the expression of miR-155 oncogene, which targets STAT4 and contributes to a switch from the Th1 to Th2 phenotype, while during later stages activation of STAT3 leads to increased survival and resistance to apoptosis, expression of miR-21 oncogenic microRNA and upregulation of IL-5 and IL-10 signaling, all working in concert to promote carcinogenesis. Further research into the molecular pathogenesis of STAT signaling in this cancer may enable us to develop effective therapies for our patients.

Materials and Methods

Patients and Samples

All patients were enrolled in an IRB-approved study protocol with informed consent.⁹ CTCL patients were recruited from the Cutaneous Lymphoma Clinic at the Dana Farber Cancer Institute (DFCI)/Brigham and Women's Hospital (BWH). All tissue samples were obtained and processed as previously described.⁹ Briefly, punch biopsies from involved skin were collected from 60 CTCL patients between January 26, 2003 and June 1, 2005. Clinical stages of CTCL at the time of diagnosis for each patient are summarized in Supplementary Table 3. The obtained 6 mm biopsies were immediately snap-frozen in liquid nitrogen. Tissue was powdered in liquid nitrogen (Cryo-Press; Microtec Co, http://nition.com/en/products/cryopress.htm), and total RNA was extracted using Trizol (Invitrogen, https://www.lifetechnologies.com/order/catalog/product/15596026) and converted to cDNA using the iScript RT-PCR kit (Bio-Rad, http://www.bio-rad.com/en-us/sku/170-8893-iscript-one-step-rt-pcr-kit-with-sybr-green) according to the manufacturer's instructions. The biopsy samples analyzed in this report are the same samples that were analyzed in previous studies.^8,9,49,50 The diagnosis and clinical staging were established according to the diagnostic criteria of CTCL.⁵¹ Similarly, volunteers with normal healthy skin (N = 6) and benign inflammatory dermatoses (N = 19) were recruited from the outpatient dermatology clinic of the University of British Columbia (Vancouver, Canada) with informed consent. Full-thickness lesional skin punch biopsies were obtained under local anesthesia as previously described.⁵⁰

Cell culture

HH, H9, Hut78, MJ and Hut102 patient-derived CTCL cell lines were previously described^52,53 and were purchased from the American Tissue Culture Collection (ATCC). Detailed information on each cell line is provided in Suppementary Table 2. H9 is a clonal derivative of Hut78 cell line.⁵⁴ MyLa, PB2B, Mac2A, SZ4 and SeAx were a generous gift from professor K. Kaltoft (Copenhagen, Denmark) and were previously described elsewhere.^41,55-58 MJ, Hut78 cells were serially passaged in IMDM media (Invitrogen) containing 10% fetal bovine serum (FBS) (Invitrogen, http://www.lifetechnologies.com/ca/en/home/life-science/cell-culture.html). HH, H9, Hut102, MyLa, Mac2A and SZ4 cells were grown in RPMI media containing 10% FBS. Finally, SeAx cells were grown in RPMI media containing 10% FBS, 5 ng/mL of recombinant human IL-2 and IL-4 (R&D Systems, http://www.rndsystems.com/Products/202-IL). All cells were grown in 5% CO₂, 95% air humidified incubator at 37°C. For stimulation, malignant T cells were treated with 10 ng/mL of PMA (phorbol 12-myristate 13-acetate) and 1 (M of ionomycin (Tocris Bioscience, http://www.tocris.com/dispprod.php?ItemId=65810#.U3AsrODD99A) or with CD3/CD28 Dynabeads® (Invitrogen, http://www.lifetechnologies.com/ca/en/home/references/protocols/proteins-expression-isolation-and-analysis/t-cell-activation-and-expansion/dynabeads-human-t-activator-cd3-cd28.html) according to manufacturer's instructions. To inhibit histone deacetylase (HDAC) activity cells were treated with 2.5-10 (M of Suberoylanilide Hydroxamic Acid (SAHA also known as Vorinostat, Santa Cruz Biotechnology, http://www.scbt.com/datasheet-220139-suberoylanilide-hydroxamic-acid.html) or 250nM-1(M Romidepsin (Adooq Bioscience, http://www.adooq.com/romidepsin-fk228-depsipeptide.html). mRNA from cell lines was isolated using Qiagen RNeasy kit (http://www.qiagen.com/products/catalog/sample-technologies/rna-sample-technologies/total-rna/rneasy-mini-kit) and was converted into cDNA using Bio-Rad iScript cDNA synthesis kit (http://www.bio-rad.com/en-us/product/reverse-transcription-reagents/iscript-cdna-synthesis-kit). Lysates for western blotting were obtained and quantitated as previously described.⁵⁹

Quantitative real-time reverse transcription-PCR gene expression analysis

STAT gene expression was tested in CTCL patients lesional skin, normal skin form healthy volunteers, lesional skin from patients with benign inflammatory dermatoses, and in patient-derived CTCL cell lines as previously described.^8,11,50 Primers for candidate human genes were designed using Primer 3 web software⁶⁰ and were purchased from Invitrogen Inc. Primer pair sequences for tested cytokines, STAT genes and control housekeeping genes are listed in Table S1. RT-PCR was performed utilizing the obtained cDNA from patients and iScript RT-PCR mix (Bio-Rad, http://www.bio-rad.com/en-us/sku/170-8893-iscript-one-step-rt-pcr-kit-with-sybr-green) on Bio-Rad iCycler as previously described.^8,49,50 The expression was standardized using genorm method⁶¹ utilizing ACTB, SDHA and GAPDH housekeeping genes. The obtained data was analyzed using XLSTAT 2009 software to obtain Kaplan-Meier curves.⁶² We determined the optimal cut off point of STAT4 mRNA expression as 0.02 using the X-Tile software.⁴³ In our analysis, STAT4 high and STAT4 low groups were defined by the expression levels higher or lower than 0.02, respectively. p values < 0.05 were considered to be statistically significant.

Western blotting

Western blotting was performed as described previously.^11,49 STAT antibodies were purchased from Cell Signaling (http://www.cellsignal.com/products/primary-antibodies/9939?Ntt=%239939&fromPage=search) as part of Stat Antibody Sampler kit (Catalog #9939). STAT4 antibody was purchased separately (http://www.cellsignal.com/products/primary-antibodies/2653?Ntt=stat4&fromPage=search). β-Actin antibody was purchased from Santa Cruz Biotechnology (http://www.scbt.com/datasheet-130656-β-actin-n-21-antibody.html). Chemiluminescent detection reagents (ECL) were purchased from Amersham Biosciences (http://www.gelifesciences.com/webapp/wcs/stores/servlet/productById/en/GELifeSciences/28980926).

Bioinformatics sequence analysis for putative targets of miR-155

Following our report on the key function of miR-155 in CTCL,¹⁵ we observed a reverse correlation between miR-155 and STAT4 expression in malignant T cells, which prompted us to use RNAhybrid analysis from Bielefeld University Bioinformatic Server database (http://bibiserv.techfak.uni-bielefeld.de/rnahybrid) to test our hypothesis that miR-155 targets STAT4.⁶³ Screening for matches in the STAT4 3'UTR, identified a putative miR-155 binding site (position 2560 to 2690 of STAT4 mRNA NM_003151.3).

Transient transfection with miRNA inhibitors (antagomiRs)

Cells were transfected with miR-155 antagomiR and negative control (mirVana^TM miRNA inhibitors, Ambion, http://www.lifetechnologies.com/ca/en/home/life-science/epigenetics-noncoding-rna-research/mirna-profiling-/ambion-anti-mir-inhibitors.html) using 0.625 nmol of the respective miR/control on 5.0 × 10⁶ cells with an Amaxa Nucleofector as descriped previously.¹⁵

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank Mr. Gregory Cormack for his technical assistance in performing molecular experiments.

Supplemental Materials

Supplemental data for this article can be accessed on the publisher's website.

Additional information

Funding

This work was supported by the Canadian Dermatology Foundation research grants to Dr. Sasseville, Dr. Litvinov and Dr. Zhou, the Fonds de la recherche en santé du Québec (FRSQ) research grant to Dr. Sasseville, Canadian Institutes of Health Research to Dr. Zhou, and the National Institutes of Health SPORE program (P50CA093683) to Dr. Kupper.