Skin biopsy tissues were fixed with 10% neutral buffered formalin for 24 h at room temperature, paraffin-embedded tissue sections (3 µm) from 33 patients at different stages of MF (patch stage MF, n=11; plaque stage MF, n=11; tumor stage MF, n=11) were obtained from the clinical databases of the Department of Dermatology and Venereology, Peking University Third Hospital and the Department of Dermatology and Venereology, Aviation General Hospital, Beijing, China, with approval from the Clinical Ethics Board of Peking University Third Hospital and Aviation General Hospital. All experiments were performed in accordance with the Declaration of Helsinki. The characteristics of the recruited patients are listed in Table I. The histological diagnosis of MF was confirmed by two independent pathologists according to the clinical and pathological diagnoses criteria published by the International Society of Cutaneous Lymphoma. The age range of the patients was 27–82 years, with a mean age of 46.1 years. and there were 18 males and 15 females. The disease durations ranged between 3 months and 30 years, with a mean duration of 15.8 years. Skin biopsies were also obtained from 11 subjects with BIDs (chronic dermatitis n=7, lichen planus n=4) and used as controls.

Paraffin-embedded sections from 33 MF and 11 BID cases were deparaffinized in xylene for 20 min and rehydrated with decreasing concentrations of ethanol solutions for 5 min each, then rinsed three times with PBS for 5 min. Next, the sections were incubated in 3% H₂O₂ for 20 min at 23–26°C. Following microwaving for antigen retrieval (15 min), the sections were washed three times with PBS. They were incubated overnight at 4°C in a 1:50 dilution of polyclonal rabbit anti-human GFI-1 antibody (cat. no. ab21061; Abcam), and then stained using the goat anti-rabbit/mouse antibody detection kit (cat. no. PV-9000D; OriGene Technologies, Inc.). Following diaminobenzidine treatment, the slides were counterstained with Mayer's hematoxylin (0.25%) at room temperature for 10 sec. The tissue slides were scanned using NanoZoomer (NanoZoomer-SQ C13140-01). Positive and negative controls were included with each run of the samples. The percentage of positive cells was assessed under high-power magnification (×400) using an Olympus light microscope, and the mean positive cells of three fields of view was used.

The human CTCL Hut-78 cell line (ATCC no. TIB-161) was cultured in RPMI-1640 medium supplemented with 10% FBS, 100 U/ml penicillin and 0.1 mg/ml streptomycin (Gibco; Thermo Fisher Scientific, Inc.), and incubated at 37°C with 5% CO₂ in a humidified incubator. The peripheral blood cells of 5 patients with benign inflammatory dermatoses (BID) patients were purified by negative selection with monoclonal antibodies directed against granulocytes, B cells by utilizing a Rosette Sep kit (Stemcell Technologies, Inc.). The purity of cells was verified by FACS using fluorescein isothiocyanate-conjugated anti-CD4 antibody (Becton Dickinson Immunocytometry Systems). More than 90% purity of cells was confirmed by immune phenotyping.

A total of four lentiviral shRNA (SH1-4) vectors were constructed by ligating four different oligo nucleotides encoding shRNAs against GFI-1 to an mU6-MCS-Ubi-EGFP vector (GV118; Shanghai GeneChem Co., Ltd.) between the HpaI and XhoI restriction enzyme sites. The oligonucleotides encoding a scrambled shRNA were used as the control (SH0). All constructs were verified by DNA sequencing, Lentiviruses containing shRNA vectors were packaged and produced by Shanghai GeneChem Co., Ltd. The Lentiviruses containing shRNA vectors were diluted at 1:10 in medium containing polybrene (final concentration, 5 mg/ml) in the wells with the cells. Following incubation for 12 h at 37°C, the cultures were replenished with fresh medium and maintained for a further 60 h. Lentiviral production and transduction were performed as previously described (14).

Cell viability analysis was performed using a Cell Viability Colorimetric assay kit, which is an MTS-based cell viability assay (Promega Corporation), according to the manufacturer's protocol. In brief, the Hut-78 cell lines were cultured in growth factor-free RPMI-1640 medium with 10% FBS, and were plated into 6-well plates (2×10⁴ cells/ml per well) in 2 ml medium and incubated at 37°C with 5% CO₂ for 96 h. Every 24 h, 100 µl cell suspension was transferred to a 96-well plate, 20 µl MTS solution was added, and the sample was left for an additional 2 h. The relative cell viability was measured at 490 nm using a spectrophotometer. Each condition was assessed in triplicate, and biological replicates were repeated twice.

Methylcellulose medium (MethoCult CFC; Stemcell Technologies, Inc.) was used to analyze the ability of the cells to form colonies in vitro. Methylcellulose medium allows for the formation of colonies from a single cell in situ to form a clonal colony. In the two lentiviral transfection groups (SH1 and SH2), the untransfected cells (Hut-78), and the scramble shRNA sequence group (SH0), 10³ cells were counted under an Olympus confocal microscope (magnification, ×200). Cells were added to the methylcellulose medium and shaken. A syringe with a blunt-tip needle was then used to inoculate the mixed cells into a petri dish, where they were cultured at 37°C with 5% CO₂ for 14 days. Cells were then counted and colony types evaluated using an inverted microscope (×200) and a grid counting dish. The experiment was repeated three times and the mean calculated.

For the analysis of spontaneous apoptosis, the cells in the lentiviral transfection groups (SH1 and SH2), the untransfected group (Hut-78), and the scrambled shRNA sequence group (SH0) were seeded at 4×10⁵/ml in 2 ml RPMI-1640 medium for 24 h, prior to the cells being washed with cold PBS and fixed in cold 70% ethanol overnight at −20°C. Next, the cells were suspended in 100 µl PBS, and then stained with 5 µl 7-aminoactinomycin D and 5 µl phycoerythrin-conjugated Annexin V using the Annexin V-PE Detection kit I (BD Pharmingen) at room temperature for 15 min, and quantified by flow cytometric analysis on a FACScan flow cytometer (BD Biosciences). Cell cycle analysis was performed using PI-mediated flow cytometry. A total of 5×10⁵ transfected cells (SH1 and SH2), as well as control cells (Hut-78 and SH0) were collected, fixed and permeabilized with 100% ethanol for 1 h at 4°C. Following treatment with DNase-free RNase, the cells were stained with 50 µg/ml PI for 1 h at room temperature. Distribution of the cell-cycle phases was determined using a FACScan flow cytometer (BD Pharmingen). For each sample, 10,000 gated events were obtained. Flow cytometry data were analyzed using CellQuest Pro and ModFit v3.3 software (BD Biosciences). The experiments were performed in triplicate and repeated at least three times.

Total RNA was extracted from cells using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.), purified and reverse-transcribed into cDNA using the Reverse Transcription System (Promega Corporation). The reaction steps were as follows: Add 2 µl of total RNA to the RT system to control the total reaction system to 20 µl. The RT reaction conditions were as follows: 42°C for 15 min, 95°C for 5 min and 4°C for 5 min. Hybridization of the cDNA library was performed on Whole Human Genome Oligo microarrays (cat. no. G4112F; Agilent Technologies, Inc.). Agilent microarray slides were scanned using the Agilent DNA Microarray Scanner (Agilent Technologies, Inc.). Hybridization signals were normalized and analyzed using GeneSpring software version 7.3 to identify significantly differentially expressed genes.

Cell lysates were prepared using RIPA lysis and extraction buffer (cat. no. 89900; Thermo Fisher Scientific, Inc.), and the concentrations were quantified using a BCA assay. Protein lysates were then separated on NuPAGE Novex Bis-Tris Gels (Invitrogen; Thermo Fisher Scientific, Inc.). A total of 40 µg protein/lane was separated by 12% SDS-PAGE. The separated proteins were subsequently transferred onto nitrocellulose membranes and blocked with 1×Tris buffered saline containing Tween-20 and 2.5% skimmed milk at 25°C for 1 h. The membranes were incubated with the following primary antibodies: Rabbit anti-GFI-1 antibody (1:500; cat. no. ab21061; Abcam), rabbit anti-Bax antibody (1:1,000; cat. no. Ab32503; Abcam), monoclonal rabbit anti-pro-caspase3 antibody (1:1,000; cat. no. Ab32150; Abcam), monoclonal rabbit anti-cleaved-caspase3 antibody (1:500; cat. no. Ab32042; Abcam), monoclonal rabbit anti-P21 antibody (1:1,000; cat. no. Ab109199; Abcam), monoclonal rabbit anti-CDK2 antibody (1:1,000; cat. no. Ab32147; Abcam), monoclonal rabbit anti-GAPDH antibody (1:2,000; cat. no. Ab8245; Abcam) and monoclonal mouse anti-β-Actin antibody (1:2,000; cat. no. Ab32150; Abcam), overnight at 4°C and rinsed with 1×TBST 3 times for 5 min each. Following the primary incubation, membranes were incubated with 1:10,000 HRP labeled Goat anti-Mouse antibody (cat. no. G-21040; Thermo Fisher Scientific, Inc.), 1:10,000 HRP labeled Goat anti-Rabbit antibody (cat. no. G-21234; Thermo Fisher Scientific, Inc.), for 1 h at 25°C, subsequently washed with TBST and imaged on a LI-COR Odyssey^® Imaging System. Relative protein expression was determined using ImageQuant version 5.2 software (Molecular Dynamics).

Total RNA was extracted from aliquots of cells using TRIzol^® Reagent (Invitrogen; Thermo Fisher Scientific, Inc.), according to the manufacturer's protocol. and RT reactions were performed as previously described (15,16). qPCR was performed according to our previous study (17), Reaction conditions were as follows: Denaturation at 95°C for 10 min; then denaturation at 95°C for 15s and annealing at 60°C for 1 min repeat 40 cycles. Real time PCR dissociation curve: 95°C foe 15 sec, 60°C for 1 min and 95°C for 15 sec. Quantification of gene expression was performed using a 7500 Real Time PCR system (Applied Biosystems; Thermo Fisher Scientific, Inc.) with GAPDH as the internal control. The results are presented as copies of the specific genes per 10,000 copies of GAPDH. The sequences of the primers were as follows: GFI-1 forward, 5′-GCCCTACCCCTGTCAGTACTGT-3′ and reverse, 5′-CACCAGTGTGGATGAAAGTGTGT-3′; P21 forward, 5′-GGCAGACCAGCATGACAGATT-3′ and reverse, 5′-TTCCTGTGGGCGGATTAGG-3′; CDK2 forward, 5′-TAAAGTTGTACCTCCCCTGGATGA-3′ and reverse 5′-AAATCCGCTTGTTAGGGTCGTA-3′; Bax forward, 5′-ACCAAGAAGCTGAGCGAATGT-3′ and reverse, 5′-CAGATGGTGAGTGACGCAGTAAG-3′; Caspase-3 forward 5′-GCAAACCTCAGGGAAACATT-3′ and reverse, 5′-TTTTCAGGTCAACAGGTCCA-3′; and GAPDH forward, 5′-AAGATCATCAGCAATGCCTCC-3′ and reverse, 5′-TGGACTGTGGTCATGAGTCCTT-3′.

The wild-type P21 promoter of Hut-78 cell genome (from base pair 1,122 to + 247 relative to the transcription start site) was amplified by PCR. The mutated P21 promoter was generated by introducing point mutations to the key nucleotides in the GFI-1 binding motif using the Fast Mutagenesis system kit (Beijing Transgen Biotech Co., Ltd.), according to the manufacturer's protocol. Wild-type and the mutated promoter were cloned into the pGL3-Basic luciferase vector (Promega Corporation) between the MluI and BglII restriction enzyme sites by MluI, BglII restriction enzyme and T4 DNA Ligase (NEB, Inc.), separately. Hut-78 cells were introduced into cells by electroporation using the dual-luciferase reporter assay system (Promega Corporation), according to the manufacturer's protocol. Entranster™-E (Engreen Biosystem Co, Ltd.) was used to improve electrotransfection efficiency. Cells were cultured for 48 h after transfection for luciferase activity measurement, Relative luciferase activity was quantified by normalization to Renilla luciferase activity, which served as an internal control for transfection efficiency.

Statistical analysis was performed using SPSS version 21.0 (IBM Corp.). A Fisher's exact test was used to compare GFI-1 antibody staining amongst the BID group and the different stages of MF. A non-parametric Mann-Whitney U test was used to compare the difference between GFI-1-knockdown clonal cells and control cells. P<0.05 was considered to indicate a statistically significant difference.

In order to study the expression levels and localization of GFI-1 protein in CTCL, paraffin-embedded sections from 11 cases each of patch, plaque and tumor stage MF were obtained, and immunohistochemical analysis of GFI-1 protein was performed and compared with 11 cases of BID. Representative images of immunohistochemical analysis are presented in Fig. 1. All MF tissues exhibited nuclear staining for GFI-1 protein, with different ratios of infiltrating lymphocytes. The results of the quantitative analysis are presented in Table II; 7 out of 11 (63.6%) patch stage MF specimens, and all plaque and tumor stage MF specimens demonstrated diffuse nuclear staining of GFI-1 protein in >25% of the lymphocytic nuclei, whereas only 1 of 11 (9.1%) BID sections exhibited >25% positive staining. There was a significant difference between all MF tissue specimens and BID specimens (P<0.05; Fisher's exact test), and there was also a significant difference between patch stage MF and BID specimens (P=0.02; Fisher's exact test). Further analysis demonstrated that in 1 of 11 (9.1%) cases with patch stage MF, 7 of 11 (63.6%) cases with plaque stage MF, and in all the cases of tumor stage MF, GFI-1 protein expression was positive in >50% of infiltrating lymphocyte nuclei. There was a significant difference between the patch stage MF and plaque stage MF specimens (P=0.02; Fisher's exact test), and between the plaque stage MF and tumor stage MF specimens (P=0.04; Fisher's exact test), suggesting that the expression of GFI-1 protein gradually increased with the progression of the disease.

The GFI-1 mRNA expression levels on fresh skin lesions biopsies from 7 MF patients and 10 fresh skin lesions biopsies from patients with BID were analyzed. GFI-1 mRNA levels in MF were significantly higher than that of BID (Fig. 2). There was a significant difference between patch stage MF and BID (P<0.05).

In order to study the role of GFI-1 in the development of MF, lentivirus-mediated RNAi technology was used to inhibit the expression of GFI-1 in the CTCL-derived cell line, Hut-78 (Fig. 3). The expression of GFI-1 in Hut-78 cells transfected with four shRNA sequences (SH1-4) was significantly decreased, and the difference was statistically significant compared with the untransfected group and the scramble transfected shRNA cells (SH0) (P<0.05). An MTS-based cell viability assay was used to assess the proliferation of Hut-78 cells following GFI-1-knockdown. Knockdown of GFI-1 expression resulted in significant inhibition of cell proliferation at 48 and 72 h, compared with the control and SH0 groups (Fig. 2; P<0.05). Colony formation assays were used to determine the in vitro proliferation of cells following GFI-1-knockdown. With the decrease in GFI-1 gene expression, the ability of cells to form colonies in vitro was significantly decreased, compared with the control and SH0 groups (P<0.05).

In order to further analyze the possible mechanisms by which cell proliferation is inhibited following knockdown of GFI-1, the cell cycle distribution was determined using flow cytometry and PI staining. The proportion of cells in the G₀/G₁ phase increased significantly, and the proportion of cells in the G₂/M phase decreased significantly compared with the control group (P<0.05), which demonstrated that knockdown of the GFI-1 gene leads to G₁ cell-cycle arrest. The apoptosis of cells was also studied using flow cytometry with Annexin V-PE/7AAD staining (Fig. 4). The proportion of apoptotic cells was significantly increased following GFI-1-knockdown, compared with the control group (P<0.05). These results indicated that GFI-1 knockdown-induced cell growth inhibition was caused by simultaneous G1 cell-cycle arrest and initiation of apoptosis.

In order to further study the possible molecular mechanisms that lead to cell cycle arrest and apoptosis following GFI-1-knockdown, lentivirus-transfected cells (SH1 and SH2) were used for transcriptome analysis, with non-transfected cells (Hut-78) and scramble shRNA-transfected cells (SH0) as the controls. The results demonstrated that Hut-78 cells with decreased GFI-1 expression exhibited numerous changes in the expression of genes (Fig. 5). Based on the criteria of a fold-change >2 and P<0.05, 39 genes were found to be differentially expressed in the SH1 and SH2 cells, compared with the SH0 and Hut-78 cells. Among these, Bax and Caspase-3 expression levels, both of which are associated with spontaneous apoptosis, were upregulated, while P21 and CDK2 expression levels, both of which are associated with cell cycle inhibition, were upregulated and downregulated, respectively, and this was further confirmed by RT-qPCR and western blotting.

Luciferase reporter assays were used to determine whether GFI-1 directly regulates the transcription of P21 in Hut-78 cells. To begin with, it was predicted that the only GFI-1 binding site on the P21 promoter was located within the region −884 to −860 (Fig. 5A) using two different transcription factor binding analysis databases (MatInspector and JASPAR CORE). The P21 promoter from −1,148 to +247 was cloned and ligated into a luciferase reporter vector, as well as two different promoter mutants with the key GFI-1 binding motif AGAT progressively mutated to CGCT or CACC (Fig. 6A). Dual-luciferase assays in Hut-78 cells revealed a dose-dependent increase in luciferase activity with the mutated promoters, indicating an inhibitory effect of GFI-1 binding on P21 transcription (Fig. 6B). These results demonstrated that GFI-1 inhibits P21 expression by directly binding to a specific locus in the P21 promoter and negatively regulating its transcription.