Ann Surg Oncol DOI 10.1245/s10434-015-4488-1
ORIGINAL ARTICLE – TRANSLATIONAL RESEARCH AND BIOMARKERS
Histone Demethylase LSD1 Inhibitors Prevent Cell Growth by Regulating Gene Expression in Esophageal Squamous Cell Carcinoma Cells Isamu Hoshino, MD, PhD1, Yasunori Akutsu, MD, PhD1, Kentaro Murakami, MD, PhD1, Naoki Akanuma, MD, PhD1, Yuka Isozaki, MD, PhD1, Tetsuro Maruyama, MD, PhD1, Takeshi Toyozumi, MD, PhD1, Yasunori Matsumoto, MD1, Hiroshi Suito, MD1, Masahiko Takahashi, MD1, Nobufumi Sekino, MD1, Aki Komatsu, PhD1, Takayoshi Suzuki, PhD2, and Hisahiro Matsubara, MD, PhD1 1
Department of Frontier Surgery, Graduate School of Medicine, Chiba University, Chiba, Japan; 2Graduate School of Medical Science, Kyoto Prefectural University of Medicine, Kyoto, Japan
ABSTRACT Background. The expression of genes can be influenced by the balance of histone acetylation and/or histone demethylation, with an imbalance of these processes possibly observed in many cancers. The histone demethylase LSD1 inhibitor activity is associated with selective transcriptional regulation and alterations in the gene expression. However, the exact mechanisms underlying the antitumor effects of LSD1 inhibitors are not fully understood. Methods. The antitumor effects of NCL1, an LSD1 inhibitor, in esophageal squamous cell cancer (ESCC) cell lines were evaluated. A comprehensive analysis of the changes in the gene expression in ESCC cell lines induced by NCL1 was carried out using a microarray analysis. A loss-of-function assay using a siRNA analysis was performed to examine the oncogenic function of the gene. Results. NCL1 strongly inhibited the cell growth of T.Tn and TE2 ESCC cells and induced apoptosis. According to the microarray analysis, 81 genes in the T.Tn cells and 149 genes in the TE2 cells were up- or down-regulated 2-fold or more by NCL1 exposure. Among these genes, 27 were contained in both cell lines and exhibited similar expression patterns. PHLDB2, one of the genes down-regulated
Electronic supplementary material The online version of this article (doi:10.1245/s10434-015-4488-1) contains supplementary material, which is available to authorized users. Ó Society of Surgical Oncology 2015 First Received: 30 August 2014 I. Hoshino, MD, PhD e-mail: [email protected]
by NCL1, was overexpressed in the ESCC tumor tissues. Moreover, a high expression level of PHLDB2 was found to be significantly correlated with poor prognosis. Conclusions. The present observations of the comprehensive analysis of the gene expression levels provide insight into the mechanisms underlying the antitumor effects of LSD1 inhibitors in ESCC patients. Esophageal cancer consists of two main histological types: esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma. ESCC is the predominant type, accounting for more than 90 % of cases of esophageal cancer in Asian countries, including Japan, Korea, and China,1 and ESCC is recognized to be a malignant tumor with a poor prognosis.2–4 Current chemotherapeutic effectiveness remains low with respect to a radical cure, and the occurrence of adverse events is not negligible.5 Recently, enzymes that regulate the gene expression via epigenetic mechanisms, including DNA methylation and histone modification, have become attractive targets for the treatment of various cancers.6–9 In particular, two histone deacetylase inhibitors have received approval from the U.S. Food and Drug Administration for the treatment of cutaneous T cell lymphoma: vorinostat and depsipeptide.10 Histone acetylation and methylation are two central modifications that function as specific transcription regulators in response to various cellular signals.11 Lysinespecific demethylase 1 (LSD1), a histone demethylase, is an amine oxidase that removes mono- and dimethyl moieties from Lys4 of histone H3 and generates the demethylated H3 tail.12 The overexpression of LSD1 has been reported in various cancers.9,13–15 More recently, histone demethylase inhibitors have been developed, with
I. Hoshino et al.
significant inhibition of tumor growth observed in many cancers, thus indicating their therapeutic potential.16–20 In this study, we investigated the antitumor effects of LSD1 inhibitor in ESCC cell lines and examined the gene expression profiles after LSD1 inhibitor administration in order to clarify the mechanisms underlying the tumor suppressive effects of histone demethylase inhibition. MATERIALS AND METHODS Patients and Samples Tissue samples were obtained from patients who underwent surgery at Chiba University Hospital between 2003 and 2012. Eighty-five pairs of primary ESCC and corresponding normal esophageal epithelia were obtained. Written consent for sample donation for research purposes was obtained from each patient prior to sample collection. The study protocol was approved by the institutional review board of Chiba University.
measured using a xMarkTM microplate absorbance spectrophotometer (Bio-Rad Laboratories, Hercules, CA). Terminal Deoxynucleotidyl Transferase-Mediated Nick-end Labeling Assay Apoptosis was detected based on an analysis of DNA fragmentation using a terminal deoxynucleotidyl transferase-mediated nick end labeling assay (TUNEL) with an In Situ Apoptosis Detection kit (Takara, Tokyo, Japan). The apoptotic index was defined as the percentage of positive cells among 1000 cells in three arbitrary microscopic fields. Cell Motility and Invasion Assay Cell motility was determined using a micropore chamber assay, and the cell invasion assays were carried out using modified Boyden Chambers containing transwell-precoated Matrigel membrane filter inserts with 8 lm pores (BD Biosciences, Bedford, MA), as previously described.21
Cell Culture and Chemicals
In Vivo Animal Model
The human esophageal cell lines T.Tn and TE2 were cultured in Dulbecco modified eagle medium (Life Technologies, Grand Island, NY) supplemented with 10 % fetal calf serum. T.Tn, TE1, -4, -12, and -13 cells were purchased from the Japanese Cancer Research Resources Bank (Tsukuba, Japan). TE2 cells were obtained from Tohoku University (Sendai, Japan). NCL1, an LSD1 inhibitor, was provided by the Graduate School of Medical Science, Kyoto Prefectural University of Medicine (Kyoto, Japan).19,20 In the in vitro studies, the compound was dissolved in dimethyl sulfoxide and diluted with each experimental medium. In the in vivo studies, the compound was dissolved and diluted with 10 % polyoxyethylated (60 mol) hydrogenated castor oil (HCO60) in saline.
TE2 and T.Tn were injected into the backs of BALB/c nu/nu mice. The animals were divided into four groups and treated i.v. four times at 4-day intervals with a dose of 0, 2.5, or 25.0 mg/kg, respectively. The tumor weight was calculated using the following formula: Tumor Weight (mg) = (length 9 width2)/2.
Cytotoxicity Assay The degree of cell growth was determined using the Cell Counting kit-8 (Dojindo, Kumamoto, Japan). The cells (5 9 103 per well) were seeded into 96-well microplates and incubated for 48 h at 37 °C under 5 % CO2. The medium was changed for the test samples and incubated in the presence or absence of the NCL1 for 72 h. Then the Cell Counting kit-8 reagent was added and allowed to react for 3 h. The absorbance at 450 nm was subsequently
Western Blot Analysis The cells were collected and lysed with radioimmunoprecipitation assay buffer. The protein extracts were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes. The membranes were then blocked in phosphate-buffered saline containing 3 % nonfat dried milk and 0.1 % Tween-20. Anti-histone H3 (dimethyl K4) polyclonal antibodies (1:500; Abcam, Inc., Cambridge, UK) and anti-b-actin polyclonal antibodies (1:5000; Abcam, Inc.) were used. After incubation with the primary antibody for 3 h at room temperature, the membranes were incubated for 2 h in secondary antibody in phosphate-buffered saline and visualized using ECL prime western blotting detection reagents (GE Healthcare Life Sciences, Uppsala, Sweden). Densitometry was subsequently performed using the Image J software program (National Institutes of Health, Bethesda, MD).
LSD1 Inhibitor Prevents Cell Growth
Immunohistochemistry The sections were mechanically deparaffinized and incubated in Target Retrieval Solution (Dako, Carpinteria, CA). After blocking the endogenous peroxidase activity with methanol containing 3 % hydrogen peroxide (Dako), the tissue sections were incubated with LSD1 (1:500; Cell Signaling Technology, Beverly, MA) and Pleckstrin homology-like domain family B member 2 (PHLDB2) antibodies (1:100; Novus Biologicals, Littleton, CO) at 4 °C overnight. The sections were then incubated in the secondary antibodies at 37 °C for 60 min. Subsequently, the sections were counterstained with hematoxylin for 1 min. Quantitative Reverse Transcription Polymerase Chain Reaction (qRT-PCR) Total RNA was extracted from the frozen surgical specimens and cell lines using Trizol reagent (Invitrogen, Carlsbad, CA) and reverse transcribed into cDNA using a High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City, CA). Real-time RT-PCR was then carried out using the SsoFast EvaGreen Supermix (Bio-Rad, Hercules, CA) with the following primers (PCReady PCR and Sequencing Primers; Operon, Biotechnology, Tokyo, Japan): PHLDB2, 50 -CTCAGACAGGCCTCAGGAAC-30 and 50 -AAGCGACTGTCAGGCTTTGT-30 . ACTA1 was used as an internal control, with sequences of 50 -CCTTCATCGGTA TGGAGTC-30 and 50 -GTTGGCATACAGGTCCTT-30 . Messenger RNA Preparation and cDNA Microarray Analysis The cells were treated with or without an IC80 concentration of LSD1 inhibitor and collected at 24 h. The cells were subsequently processed for RNA extraction with the RNeasy Plus Mini Kit (Qiagen, Chatsworth, CA). A total of 5.5 lg of total RNA derived from cells cultured with NCL1 was compared to 5.5 lg of total RNA derived from cells cultured without NCL1 using an Affymetrix Human Exon 1.0ST Array (Affymetrix, Santa Clara, CA), currently the most dense array designed specifically for profiling the gene expression. All processing was performed according to described protocols.22 Transfection of Small Interfering RNA Small interfering RNA (siRNA) sequences (Stealth RNAi siRNA; Invitrogen) targeting PHLDB2 (siPHLDB21, Cat. No. HSS131884; siPHLDB2-2, Cat. No. HSS131885) and the negative control (Negative Control Duplex med GC Duplex No. 2: Cat. No.12935-112) were
transfected into the cell lines via electroporation using the Neon Transfection System (Invitrogen). Statistical Analysis The relationships between two variables and numerical values obtained via real-time RT-PCR were analyzed by t test. The patients were divided into two groups of high (n = 43) or low PHLDB2 expression (n = 42). A survival analysis was performed according to the Kaplan–Meier method using the logrank test. Clinical features related to the prognosis were selected for multivariate analysis with a Cox proportional hazard regression model. Statistical significance was defined as a P value of \0.05. All statistical analyses were performed by the JMP statistical package (SAS Institute, Cary, NC). RESULTS Immunohistochemical Analysis of LSD1 Expression in ESCC Cells The LSD1 expression was observed in all specimens examined, and the expression in the tumor tissues was higher in all cases compared to that observed in the corresponding normal epithelium (Fig. 1a). NCL1 exhibited an antiproliferative activity by inducing cell apoptosis; it reduced invasion and migration activities in the human esophageal cancer cell lines. We assessed the antitumor activity of NCL1, an LSD1 inhibitor, in the human esophageal cancer cell lines TE2 and T.Tn, and the sensitivity of these cells was evaluated according to the IC50 values of 34.2 and 31.5 lM, respectively (Fig. 1b). In both the TE2 and T.Tn cells, the TUNEL index showed that NCL1 significantly increased the number of apoptotic cells (Fig. 1d). In addition, NCL1 reduced both invasion and migration activities significantly in ESCC cell lines (Supplementary Fig. S1). Effects of NCL1 on T.Tn and TE2 Xenografts As shown in Fig. 1d, 21 days after the first administration of NCL1, the growth of the T.Tn tumors was significantly suppressed compared to that observed in the control group (P = 0.02). On the other hand, in the TE2 cells, although tumor growth inhibition was observed to some extent, there were no significant differences in tumor size (P = 0.11). NCL1 Induced Histone H3 (Lys4) Dimethylation In order to determine whether LSD1 inhibitors induce histone H3 (Lys4) dimethylation, TE2 and T.Tn cells were
I. Hoshino et al. FIG. 1 a LSD1 expression in ESCC tissues. Immunohistochemical c studies identified aberrant expression in tumor tissues. LSD1 expression levels in tumor tissues were higher in all cases compared to those observed in corresponding normal epithelium. b Antiproliferative activity in T.Tn and TE2. Cells were incubated for 72 h in the presence or absence of indicated concentrations of NCL1, and the degree of cell viability was determined using Cell Counting Kit-8, as described in materials and methods. Results are shown as percentage of viability in control specimens. Points indicate mean of eight separate experiments; bars, SD. c In vitro apoptotic effects of NCL1 in T.Tn and TE2 cells. T.Tn and TE2 cells were exposed to NCL1 at IC80 concentrations for 72 h, and apoptotic cells were stained using TUNEL. Number of positive cells increased in a time-dependent manner in both cell lines. Apoptotic index was defined as percentage of positive cells among 1000 cells in three arbitrary microscopic fields. Data are expressed as mean ± SD. Statistical significance was evaluated by unpaired Student’s t test. d NCL1 inhibition of human esophageal cancer cell tumor growth in vivo. Fragments of human solid tumor were implanted s.c. into the backs of BALB/c nu/nu mice. Drugs were given i.v. to mice 4 times at 4-day intervals with NCL1. Tumor weight was calculated from results of five independent experiments. Points indicate mean; bars, SD. Data were analyzed for statistical significance by Dunnett’s test for multiple comparisons (*P \ 0.05 vs control). e Histone H3 (lys4) dimethylation induced by NCL1. T.Tn and TE2 cells were incubated with different amounts of NCL1 for 24 h and then analyzed for dimethylated histone H3 (Lys4) using western blot analysis, as described in materials and methods. Western blot analysis was performed three times; representative data are shown. Data are expressed as mean ± SD in graph. Statistical significance was evaluated by unpaired Student’s t test (*P \ 0.05 vs control)
A
Tumor
B
Normal
120
%of control
100
T.Tn TE2
80 60 40 20 0 0
1
2
3
6
10
20
30
60
100
Concentration (µm)
C
T.Tn
TE2
control
IC50
control
(%)
p=0.023
p=0.028
40
IC50
20 0
IC50
control
In order to determine the target genes of LSD1 in esophageal cancer cell lines, we conducted a cDNA microarray analysis. Twofold or greater alterations in the gene expression, either increasing or decreasing, were considered significant. Eighteen genes were found to be upregulated and nine were found to be down-regulated in both cell lines (Tables 1, 2). Then we determined whether NCL1 can reduce the expression levels of one of the downregulated genes in microarray data. We chose PHLDB2. T.Tn and TE2 cells were exposed to NCL1 at various concentrations for 24 h or various time period at IC80 for 24 h. As expected, NCL1 reduced PHLDB2 expression levels in a dose-dependent and time-dependent manner (Fig. 2a). Effects of Transfection of siRNA Targeting PHLDB2 mRNA The introduction of each of the two siRNA sequences (siPHLDB2-1, siPHLDB2-2) into TE2 and T.Tn cells led to a significant decrease in the level of the target mRNA at
350 300 250 200 150 100 50 0
T.Tn control 2.5mg/kg 25mg/kg
0
3
7
10
14
17
21
1400 1200 1000 800 600 400 200 0
IC50 TE2 control 2.5mg/kg 25mg/kg
0
3
Days after Treatment
E
7
10
14
17
21
Days after Treatment T.Tn
TE2 H3K4 Beta-actin 0
Relative expression levels (normalized to Beta-actin)
NCL1 Regulated Various Gene Expression Levels
control
D Tumor Weight (mg)
exposed to NCL1 at IC80 for 24 h. As expected, NCL1 induced histone H3 (Lys4) dimethylation in a dose-dependent manner (Fig. 1e).
0.1
1.0 10
100
0
0.1 1.0 10 100
4
(µm)
T.Tn
TE2 5
*
*
*
4
3
3 2 2 1 0
1 0
0.1
1.0
10
100
0 0
0.1
1.0
10
100 (mM)
48 h (Fig. 2b). The degree of cell proliferation in the PHLDB2 down-regulated cells was compared to that observed in mock cells, which revealed that cell proliferation was significantly reduced after 96 h in both cell lines (Fig. 2c). TUNEL index showed that siPHLDB2 significantly increased the number of apoptotic cells (Fig. 2d). In both the migration and Matrigel invasion assays, the rate
LSD1 Inhibitor Prevents Cell Growth TABLE 1 List of 18 genes up-regulated ([2-fold changes) in T.Tn and TE2 cells Genbank accession no.
Gene symbol
Description
BC005008
CEACAM6
BC012172 BC008723 L19501 M29540
Maximum fold change T.Tn
TE2
Carcinoembryonic antigen-related cell adhesion molecule 6 (nonspecific cross-reacting antigen)
2.84
3.62
ACSS2
Acl-CoA synthetase short-chain family member 2
1.10
1.24
ASNS CBS
Asparagine synthetase (glutamine hydrolyzing) Cystathionine-beta-synthase
1.92 1.15
1.96 1.18
CEACAM5
Carcinoembryonic antigen-related cell adhesion molecule 5
2.45
2.90 2.11
BC019625
CHAC1
ChaC, cation transport regulator homolog 1 (E. coli)
1.34
U03688
CYP1B1
Cytochrome P4509, family 1, subfamily B, polypeptide 1
1.49
1.67
BC007333
ETV5
Ets variant 5
1.73
1.81
AF110400
FGF19
Fibroblast growth factor 19
1.17
1.26
EF152283
GCNT3
Glucosaminyl (N-acetyl) transferase 3, mucin type
1.17
1.66
AF019770
GDF15
Growth differentiation factor 15
2.01
1.75
BC033089
LCN2
Lipocalin 2
1.62
2.78
BC004863
PSAT1
Phosphoserine aminotransferase 1
1.34
2.87
AF539739
S100P
S11 calcium binding protein P
1.26
2.06
AF097514
SCD
Stearoyl-CoA desaturate (delta-9-desaturase)
1.25
2.38
BC000658
STC2
Stanniocalcin 2
1.95
1.59
BGC011703
TMPRSS4
Transmembrane protease, serine 4
1.03
1.24
AF022375
VEGFA
Vascular endothelial growth factor A
1.62
1.50
TABLE 2 List of genes down-regulated ([2-fold changes) in T.Tn and TE2 cells Genbank accession no.
J04948
Gene symbol
ALPPL2
Description
Maximum fold change
Alkaline phosphatase, placental-like 2
T.Tn
TE2
-1.01
-1.69
BC098561
EFEMP1
EGF-containing fibulin-like extracellular matrix protein 1
-1.59
-1.76
AB031548
GPR87
G protein-coupled receptor S7
-1.11
-1.06
M19154
TGFB2
Transforming growth factor, beta 2
-1.72
-1.11
BC142678
PHLDB2
Pleckstrin homology-like domain, family B, member 2
-1.14
-1.04
BC146868
COL12A1
Collagen, type XII, alpha 1
-1.13
-1.22
BC017782 AF098807
WISP2 LHFP
WNT1 inducible signaling pathway protein 2 Lipoma HMGIC fusion partner
-1.12 -1.21
-1.31 -1.26
AK123348
C3orf57
Chromosome 3 open reading frame 57
-1.14
-1.72
of penetration of cells through the membrane was significantly decreased after the transfection of each siRNA (Fig. 2e). In addition to T.Tn and TE2 ESCC cell lines, we explored the role of siPHLDB2 in other ESCC cell lines. siPHLDB2 showed an antiproliferation effect on various ESCC cell lines (Fig. S2). PHLDB2 Expression in ESCC Tissues The PHLDB2 expression level in the ESCC tissues was significantly higher than that observed in the corresponding normal epithelium (P = 0.002; Fig. 3a). In addition, we examined the PHLDB2 protein expression in the ESCC
specimens and normal epithelium (Fig. 3b); the PHLDB2 expression in the cytoplasm was observed in all specimens examined. We next evaluated the correlations between the PHLDB2 mRNA expression and various clinicopathologic factors, including sex, age, T factor, N factor, stage, survival, and recurrence (Table 3) after dividing the 85 patients into two groups, a high-expression group (n = 43) and a low-expression group (n = 42). Although none of the factors was found to be significantly related to the PHLDB2 expression (Table 3), a survival analysis using the Kaplan–Meier method and the logrank test showed a significant decrease in overall survival in the highPHLDB2 expression group, with a 5-year survival rate of
I. Hoshino et al.
(%)
140 120
100
*
80
100 80
60 40
**
**
12
24
48 (h)
*
*
*
12
24
48 (h)
40
20
20
0
0 0
0.1
1
10
100 (µm)
0
1
6
TE2 (%)
(%) 160 140 120 100 80 60 40 20 0
140 120 100 80
*
60 40 20 0 0.1
1
10
0
100 (µm)
KDMI concentration
B
Relative expression levels (normalized to ACTA1)
120
100
100
80
80
60
60 40
20
*
*
si1
si2
*
*
si1
si2
20 0
0
NC
T.Tn Relative growth curve
6
TE2
(%)
120
40
C
1
Time after KDMI administration
T.Tn
(%)
NC
TE2
7
7
6
6
5
5
4
4
3
3
*
2
*
1
D
2
*
1
48
72
96
24
48
T.Tn
control
(%)
si1
si2
control
(h)
si1
si2
*
*
si1
si2
*
10
10
5
5
0
0 control
Migration
96
15 *
(%)
72
TE2
15
(%)
NC si1 si2
*
0
0 24
E
**
60
0
DISCUSSION We demonstrated that LSD1 is overexpressed in ESCC specimens and may be a potent therapeutic target in the setting of ESCC. Furthermore, dimethylation of Lys4 of histone H3 in the ESCC cell lines by NCL1 regulated various gene expression levels and is therefore essential for the observed antitumor effect. To date, more than 20 methyl marks on lysine and arginine residues have been identified.23 One of the first methyl marks to be characterized in detail is methylation at Lys4 of histone H3 (H3K4me). According to the histone code hypothesis, single histone modifications or combinations of modifications can be recognized by effector proteins and/or protein complexes that read the marks, converting them into specific functional chromatin states and thereby regulating downstream responses.24,25 In fact, methylation at H3K4 is reported to be strongly associated with transcriptional activation. In contrast, trimethylation at H3K9, H3K27, and H4K20 is commonly considered to be a repressive modification.26 LSD1 removes the methyl groups from mono- and dimethylated Lys4 of histone H3 (H3K4me1/2).27 LSD1 is essential for many biological processes involved in development, growth, differentiation, and morphogenesis.28 LSD1 was reported to interact with the Mi-2/nucleosome remodeling and deacetylase (NuRD) complex and the LSD1/NuRD complexes control several cellular signaling pathways including transforming growth factor b1 signaling pathway.29 Interestingly, the overexpression of LSD1 has been shown to correlate with adverse clinical outcomes in patients with neuroblastoma, and the pharmacological inhibition of LSD1 reduces neuroblastoma growth in xenografted nude mice in vivo.30 In addition, a high expression level of LSD1 in patients with prostate cancer may predict an aggressive tumor biology.28 Meanwhile, one study showed that the overexpression of
T.Tn 120
Invasion
40.0 % (vs 69.1 % in the low-PHLDB2 expression group) (Fig. 3c). On the basis of the results of a multivariate analysis (Table 4), a high expression level of PHLDB2 was found to be the strongest independent risk factor for a poor survival, with a hazard ratio of 2.545 (95 % confidence interval 1.218–5.516; P = 0.013).
A (%) Relative expression levels (normalized to ACTA1)
FIG. 2 a NCL1 reduced PHLDB2 expression levels in a dose-and c time-dependent manner. b Quantification of mRNA of PHLDB2 after transfection of specific siRNA with real-time quantitative PCR. *P \ 0.01. c Silencing of PHLDB2 decreased rate of proliferation of ESCC cells. *P \ 0.05. d TUNEL index showed that siPHLDB2 significantly increased number of apoptotic cells. e Results of Matrigel invasion assay showed that silencing PHLDB2 reduces invasive potential of ESCC cells. NC negative control. *P \ 0.05
si1
si2
control
T.Tn
TE2
140
120
120
100
100
80
80
*
60
*
60
*
*
*
*
si1
si2
40
40 20
20
0
0
140
140
120
120
100
100
80
80
60
*
40
*
60 40
20
20
0
0
NC
si1
si2
NC
LSD1 Inhibitor Prevents Cell Growth * Relative mRNA expression
A
TABLE 3 Correlation between PHLDB2 expression and clinicopathologic features of esophageal squamous cell carcinoma
30.0 25.0 20.0 15.0 10.0 5.0 0.0
Characteristic n Total (%) Tumor
High PHLDB2 expression, n (%)
Low PHLDB2 expression, n (%)
P
85 43 (50.6)
42 (49.4)
70 36 (42.4) 15 7 (8.2)
34 (40.0) 8 (9.4)
0.9600
0.4172
Sex (%)
Normal
Male Female
B
Age (%) \65
32 18 (21.2)
14 (16.5)
C65
53 24 (29.4)
28 (32.9)
Tumor depth (%) Tumor
Overall survival rate
C
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
(×400)
Normal
T1-2
35 15 (17.7)
20 (23.5)
T3-4
50 28 (32.9)
22 (25.9)
0.2329
Lymph node metastasis (%) Low expression (n=42)
High expression (n=43) 5-year OS rate Low: 63.14% High: 40.00% 0
Negative
35 20 (23.5)
15 (17.7)
Positive
50 23 (27.0)
27 (31.8)
I–II
40 18 (21.2)
22 (25.9)
III–IV
45 25 (29.4)
20 (23.5)
0.3119
Stage (%) 0.3313
Patient survival (%)
20
40
60
Months
Alive Dead
52 23 (27.1) 33 20 (23.5)
29 (34.1) 13 (15.3)
0.1411
0.5526
Recurrence (%) FIG. 3 a Expression levels of PHLDB2 in ESCC clinical specimens. Total RNA was isolated from 85 tumor tissues and matched noncancerous tissues. PHLDB2 expression levels were detected using quantitative real-time PCR. *P \ 0.05. b PHLDB2 protein expression in clinical ESCC specimens. Immunohistochemical staining for PHLDB2 showed high expression in cancer tissue (left) and low expression in neighboring normal epithelium (right). c Results of Kaplan-Meier survival analysis of overall survival in high-PHLDB2 expression group versus low-PHLDB2 expression group. Results showed significant decrease in overall survival in high-PHLDB2 expression group (P \ 0.001), with 5-year survival rate of 40.0 % (vs 63.1 % in low-PHLDB2 expression group)
LSD1 is associated with a poor prognosis in patients with non-small cell lung cancer.9 Recently, ORY-1001, a LSD1 inhibitor, is entering phase 1/2 clinical study in acute myeloid leukemia.31 ORY-1001 significantly reduces tumor cells and increases survival time in mouse models of acute lymphoblastic leukemia. Therefore, the inhibition of LSD1 may provide a promising epigenetic target for cancer therapy. Indeed, we showed that LSD1 is overexpressed in ESCC specimens and that NCL1 exerted an antitumor effect in ESCC cell lines. In our experiments, tumor growth was repressed by LSD1 inhibitor injection in the xenograft mice models. LSD1 regulates various gene expression levels. Therefore, LSD1 inhibitors may exhibit efficacy in downregulating oncogenes or up-regulating tumor suppressor genes. Lim et al. reported that the knockdown of LSD1 using an siRNA approach induces the regulation of several
No
54 26 (30.6)
28 (32.9)
Yes
31 17 (20.0)
14 (16.5)
proliferation-associated genes.15 Furthermore, LSD1 knockdown induces changes in the mRNA expression as analyzed according to a microarray analysis.30 In the present study, a microarray analysis of the transcriptome was performed to determine the mRNA expression after LSD1 inhibitor exposure. We then focused on PHLDB2, one of the down-regulated genes, the antitumor effect of NCL1 may be derived from the down-regulation of oncogenes and/or up-regulation of tumor suppressor genes, and the molecule might act as an oncogene according to the latest reports. PHLDB2 was typically down-regulated after LSD1 inhibitor administration in vitro. The expression levels in the ESCC specimens were significantly higher than those observed in the normal esophageal epithelium specimens. PHLDB2 is a 1253 amino acid protein and has been identified to be a microtubule-anchoring factor that attaches EB1/CLIP-associating protein (CLASP)–bound microtubule plus ends to the cell cortex.32 Hotta et al. used the noncancerous human mammary epithelial cell line MCF-10A to investigate the mechanisms determining the cortical distribution of PHLDB2 and showed that PHLDB2 is colocalized with laminin-5 deposition and laminin receptor integrins as well as epithelial tissues and that the localization of PHLDB2 relies on the actions of laminin
I. Hoshino et al. TABLE 4 Results of multivariate analysis for survival Variable
Category
Hazard ratio
95 % confidence interval
P
Tumor depth
T3-4
8.646
1.914–65.937
0.003
Lymph node metastasis
Positive
7.345
1.514–55.123
0.012
Stage
III or IV
0.307
0.037–1.813
0.197
Ly
Positive
1.111
0.497–2.667
0.062
V
Positive
0.902
0.338–2.601
0.842
PHLDB2
High
2.545
1.218–5.516
0.013
Bold values are statistically significant
receptor integrins.33 In addition, Carles et al. found that PHLDB2 was overexpressed in 10/12 head and neck squamous cell carcinoma specimens using RT-PCR.34 These findings support our results regarding PHLDB2 overexpression in ESCC specimens. Moreover, in the siPHLDB2-transfected cells, the inhibition of proliferation, migration, and invasion was observed. We subsequently divided the patients into two groups, a high-PHLDB2 expression group and a low-PHLDB2 expression group, based on the mRNA expression level, and found no significant correlations with any clinicopathologic features in a univariate analysis. According to the multivariate analysis, a high expression level of PHLDB2 was found to be an independent poor prognostic factor. These results suggest that PHLDB2 is a valuable diagnostic and prognostic biomarker for ESCC. In conclusion, LSD1 inhibitors have potent antitumor effects on ESCC cell lines. The antitumor efficacy of LSD1 inhibitors may be derived from their ability to modulate the mRNA expression, while PHLDB2 is an oncogenic factor and promising target in ESCC therapy. DISCLOSURE
The authors declare no conflict of interest.
REFERENCES 1. Hongo M, Nagasaki Y, Shoji T. Epidemiology of esophageal cancer: orient to occident. Effects of chronology, geography and ethnicity. J Gastroenterol Hepatol. 2009;24:729–35. 2. Ye T, Sun Y, Zhang Y, Zhang Y, Chen H. Three-field or twofield resection for thoracic esophageal cancer: a meta-analysis. Ann Thorac Surg. 2013;96:1933–41. 3. Sugiura T, Uesaka K, Mihara K, et al. Margin status, recurrence pattern, and prognosis after resection of pancreatic cancer. Surgery. 2013;154:1078–86. 4. Zhimin G, Noor H, Jian-Bo Z, Lin W, Jha RK. Advances in diagnosis and treatment of hilar cholangiocarcinoma—a review. Med Sci Monit. 2013;19:648–56. 5. Ando N, Kato H, Igaki H, et al. A randomized trial comparing postoperative adjuvant chemotherapy with cisplatin and 5-fluorouracil versus preoperative chemotherapy for localized advanced squamous cell carcinoma of the thoracic esophagus (JCOG9907). Ann Surg Oncol. 2012;19:68–74. 6. Hoshino I, Matsubara H. Recent advances in histone deacetylase targeted cancer therapy. Surg Today. 2010;40:809–15.
7. Hoshino I, Matsubara H, Hanari N, et al. Histone deacetylase inhibitor FK228 activates tumor suppressor Prdx1 with apoptosis induction in esophageal cancer cells. Clin Cancer Res. 2005;11: 7945–52. 8. Liu J, Liu FY, Tong ZQ, et al. Lysine-specific demethylase 1 in breast cancer cells contributes to the production of endogenous formaldehyde in the metastatic bone cancer pain model of rats. PLoS One. 2013;8:e58957. 9. Lv T, Yuan D, Miao X, et al. Over-expression of LSD1 promotes proliferation, migration and invasion in non–small cell lung cancer. PLoS One. 2012;7:e35065. 10. Ververis K, Hiong A, Karagiannis TC, Licciardi PV. Histone deacetylase inhibitors (HDACIs): multitargeted anticancer agents. Biologics. 2013;7:47–60. 11. An W. Histone acetylation and methylation: combinatorial players for transcriptional regulation. Subcell Biochem. 2007;41:351– 69. 12. Kerenyi MA, Shao Z, Hsu YJ, et al. Histone demethylase Lsd1 represses hematopoietic stem and progenitor cell signatures during blood cell maturation. Elife. 2013;2:e00633. 13. Kashyap V, Ahmad S, Nilsson EM, et al. The lysine specific demethylase-1 (LSD1/KDM1A) regulates VEGF-A expression in prostate cancer. Mol Oncol. 2013;7:555–66. 14. Zhao ZK, Yu HF, Wang DR, et al. Overexpression of lysine specific demethylase 1 predicts worse prognosis in primary hepatocellular carcinoma patients. World J Gastroenterol. 2012; 18:6651–6. 15. Lim S, Janzer A, Becker A, et al. Lysine-specific demethylase 1 (LSD1) is highly expressed in ER-negative breast cancers and a biomarker predicting aggressive biology. Carcinogenesis. 2010;31: 512–20. 16. Murray-Stewart T, Woster PM, Casero RA Jr. The re-expression of the epigenetically silenced e-cadherin gene by a polyamine analogue lysine-specific demethylase-1 (LSD1) inhibitor in human acute myeloid leukemia cell lines. Amino Acids. 2014;46: 585. 17. Wu Y, Steinbergs N, Murray-Stewart T, Marton LJ, Casero RA. Oligoamine analogues in combination with 2-difluoromethylornithine synergistically induce re-expression of aberrantly silenced tumour-suppressor genes. Biochem J. 2012;442:693–701. 18. Willmann D, Lim S, Wetzel S, et al. Impairment of prostate cancer cell growth by a selective and reversible lysine-specific demethylase 1 inhibitor. Int J Cancer. 2012;131:2704–9. 19. Ueda R, Suzuki T, Mino K, et al. Identification of cell-active lysine specific demethylase 1-selective inhibitors. J Am Chem Soc. 2009;131:17536–7. 20. Hamada S, Suzuki T, Mino K, et al. Design, synthesis, enzymeinhibitory activity, and effect on human cancer cells of a novel series of jumonji domain-containing protein 2 histone demethylase inhibitors. J Med Chem. 2010;53:5629–38. 21. Calin GA, Croce CM. MicroRNA signatures in human cancers. Nat Rev Cancer. 2006;6:857–66.
LSD1 Inhibitor Prevents Cell Growth 22. Pitroda SP, Wakim BT, Sood RF, et al. STAT1-dependent expression of energy metabolic pathways links tumour growth and radioresistance to the Warburg effect. BMC Med. 2009;7:68. 23. Mosammaparast N, Shi Y. Reversal of histone methylation: biochemical and molecular mechanisms of histone demethylases. Annu Rev Biochem. 2010;79:155–79. 24. Turner BM. Decoding the nucleosome. Cell. 1993;75:5–8. 25. Strahl BD, Allis CD. The language of covalent histone modifications. Nature. 2000;403(6765):41–5. 26. Barski A, Cuddapah S, Cui K, et al. High-resolution profiling of histone methylations in the human genome. Cell. 2007;129: 823–37. 27. Shi Y, Lan F, Matson C, et al. Histone demethylation mediated by the nuclear amine oxidase homolog LSD1. Cell. 2004;119: 941–53. 28. Metzger E, Wissmann M, Yin N, et al. LSD1 demethylates repressive histone marks to promote androgen-receptor-dependent transcription. Nature. 2005;437(7057):436–9.
29. Wang Y, Zhang H, Chen Y, et al. LSD1 is a subunit of the NuRD complex and targets the metastasis programs in breast cancer. Cell. 2009;138:660–72. 30. Schulte JH, Lim S, Schramm A, et al. Lysine-specific demethylase 1 is strongly expressed in poorly differentiated neuroblastoma: implications for therapy. Cancer Res. 2009;69:2065–71. 31. Ketscher A, Jilg CA, Willmann D, et al. LSD1 controls metastasis of androgen-independent prostate cancer cells through PXN and LPAR6. Oncogenesis. 2014;3:e120. 32. Lansbergen G, Grigoriev I, Mimori-Kiyosue Y, et al. CLASPs attach microtubule plus ends to the cell cortex through a complex with LL5beta. Dev Cell. 2006;11:21–32. 33. Hotta A, Kawakatsu T, Nakatani T, et al. Laminin-based cell adhesion anchors microtubule plus ends to the epithelial cell basal cortex through LL5alpha/beta. J Cell Biol. 2010;189:901–17. 34. Carles A, Millon R, Cromer A, et al. Head and neck squamous cell carcinoma transcriptome analysis by comprehensive validated differential display. Oncogene. 2006;25:1821–31.
ORIGINAL ARTICLE – TRANSLATIONAL RESEARCH AND BIOMARKERS
Histone Demethylase LSD1 Inhibitors Prevent Cell Growth by Regulating Gene Expression in Esophageal Squamous Cell Carcinoma Cells Isamu Hoshino, MD, PhD1, Yasunori Akutsu, MD, PhD1, Kentaro Murakami, MD, PhD1, Naoki Akanuma, MD, PhD1, Yuka Isozaki, MD, PhD1, Tetsuro Maruyama, MD, PhD1, Takeshi Toyozumi, MD, PhD1, Yasunori Matsumoto, MD1, Hiroshi Suito, MD1, Masahiko Takahashi, MD1, Nobufumi Sekino, MD1, Aki Komatsu, PhD1, Takayoshi Suzuki, PhD2, and Hisahiro Matsubara, MD, PhD1 1
Department of Frontier Surgery, Graduate School of Medicine, Chiba University, Chiba, Japan; 2Graduate School of Medical Science, Kyoto Prefectural University of Medicine, Kyoto, Japan
ABSTRACT Background. The expression of genes can be influenced by the balance of histone acetylation and/or histone demethylation, with an imbalance of these processes possibly observed in many cancers. The histone demethylase LSD1 inhibitor activity is associated with selective transcriptional regulation and alterations in the gene expression. However, the exact mechanisms underlying the antitumor effects of LSD1 inhibitors are not fully understood. Methods. The antitumor effects of NCL1, an LSD1 inhibitor, in esophageal squamous cell cancer (ESCC) cell lines were evaluated. A comprehensive analysis of the changes in the gene expression in ESCC cell lines induced by NCL1 was carried out using a microarray analysis. A loss-of-function assay using a siRNA analysis was performed to examine the oncogenic function of the gene. Results. NCL1 strongly inhibited the cell growth of T.Tn and TE2 ESCC cells and induced apoptosis. According to the microarray analysis, 81 genes in the T.Tn cells and 149 genes in the TE2 cells were up- or down-regulated 2-fold or more by NCL1 exposure. Among these genes, 27 were contained in both cell lines and exhibited similar expression patterns. PHLDB2, one of the genes down-regulated
Electronic supplementary material The online version of this article (doi:10.1245/s10434-015-4488-1) contains supplementary material, which is available to authorized users. Ó Society of Surgical Oncology 2015 First Received: 30 August 2014 I. Hoshino, MD, PhD e-mail: [email protected]
by NCL1, was overexpressed in the ESCC tumor tissues. Moreover, a high expression level of PHLDB2 was found to be significantly correlated with poor prognosis. Conclusions. The present observations of the comprehensive analysis of the gene expression levels provide insight into the mechanisms underlying the antitumor effects of LSD1 inhibitors in ESCC patients. Esophageal cancer consists of two main histological types: esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma. ESCC is the predominant type, accounting for more than 90 % of cases of esophageal cancer in Asian countries, including Japan, Korea, and China,1 and ESCC is recognized to be a malignant tumor with a poor prognosis.2–4 Current chemotherapeutic effectiveness remains low with respect to a radical cure, and the occurrence of adverse events is not negligible.5 Recently, enzymes that regulate the gene expression via epigenetic mechanisms, including DNA methylation and histone modification, have become attractive targets for the treatment of various cancers.6–9 In particular, two histone deacetylase inhibitors have received approval from the U.S. Food and Drug Administration for the treatment of cutaneous T cell lymphoma: vorinostat and depsipeptide.10 Histone acetylation and methylation are two central modifications that function as specific transcription regulators in response to various cellular signals.11 Lysinespecific demethylase 1 (LSD1), a histone demethylase, is an amine oxidase that removes mono- and dimethyl moieties from Lys4 of histone H3 and generates the demethylated H3 tail.12 The overexpression of LSD1 has been reported in various cancers.9,13–15 More recently, histone demethylase inhibitors have been developed, with
I. Hoshino et al.
significant inhibition of tumor growth observed in many cancers, thus indicating their therapeutic potential.16–20 In this study, we investigated the antitumor effects of LSD1 inhibitor in ESCC cell lines and examined the gene expression profiles after LSD1 inhibitor administration in order to clarify the mechanisms underlying the tumor suppressive effects of histone demethylase inhibition. MATERIALS AND METHODS Patients and Samples Tissue samples were obtained from patients who underwent surgery at Chiba University Hospital between 2003 and 2012. Eighty-five pairs of primary ESCC and corresponding normal esophageal epithelia were obtained. Written consent for sample donation for research purposes was obtained from each patient prior to sample collection. The study protocol was approved by the institutional review board of Chiba University.
measured using a xMarkTM microplate absorbance spectrophotometer (Bio-Rad Laboratories, Hercules, CA). Terminal Deoxynucleotidyl Transferase-Mediated Nick-end Labeling Assay Apoptosis was detected based on an analysis of DNA fragmentation using a terminal deoxynucleotidyl transferase-mediated nick end labeling assay (TUNEL) with an In Situ Apoptosis Detection kit (Takara, Tokyo, Japan). The apoptotic index was defined as the percentage of positive cells among 1000 cells in three arbitrary microscopic fields. Cell Motility and Invasion Assay Cell motility was determined using a micropore chamber assay, and the cell invasion assays were carried out using modified Boyden Chambers containing transwell-precoated Matrigel membrane filter inserts with 8 lm pores (BD Biosciences, Bedford, MA), as previously described.21
Cell Culture and Chemicals
In Vivo Animal Model
The human esophageal cell lines T.Tn and TE2 were cultured in Dulbecco modified eagle medium (Life Technologies, Grand Island, NY) supplemented with 10 % fetal calf serum. T.Tn, TE1, -4, -12, and -13 cells were purchased from the Japanese Cancer Research Resources Bank (Tsukuba, Japan). TE2 cells were obtained from Tohoku University (Sendai, Japan). NCL1, an LSD1 inhibitor, was provided by the Graduate School of Medical Science, Kyoto Prefectural University of Medicine (Kyoto, Japan).19,20 In the in vitro studies, the compound was dissolved in dimethyl sulfoxide and diluted with each experimental medium. In the in vivo studies, the compound was dissolved and diluted with 10 % polyoxyethylated (60 mol) hydrogenated castor oil (HCO60) in saline.
TE2 and T.Tn were injected into the backs of BALB/c nu/nu mice. The animals were divided into four groups and treated i.v. four times at 4-day intervals with a dose of 0, 2.5, or 25.0 mg/kg, respectively. The tumor weight was calculated using the following formula: Tumor Weight (mg) = (length 9 width2)/2.
Cytotoxicity Assay The degree of cell growth was determined using the Cell Counting kit-8 (Dojindo, Kumamoto, Japan). The cells (5 9 103 per well) were seeded into 96-well microplates and incubated for 48 h at 37 °C under 5 % CO2. The medium was changed for the test samples and incubated in the presence or absence of the NCL1 for 72 h. Then the Cell Counting kit-8 reagent was added and allowed to react for 3 h. The absorbance at 450 nm was subsequently
Western Blot Analysis The cells were collected and lysed with radioimmunoprecipitation assay buffer. The protein extracts were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes. The membranes were then blocked in phosphate-buffered saline containing 3 % nonfat dried milk and 0.1 % Tween-20. Anti-histone H3 (dimethyl K4) polyclonal antibodies (1:500; Abcam, Inc., Cambridge, UK) and anti-b-actin polyclonal antibodies (1:5000; Abcam, Inc.) were used. After incubation with the primary antibody for 3 h at room temperature, the membranes were incubated for 2 h in secondary antibody in phosphate-buffered saline and visualized using ECL prime western blotting detection reagents (GE Healthcare Life Sciences, Uppsala, Sweden). Densitometry was subsequently performed using the Image J software program (National Institutes of Health, Bethesda, MD).
LSD1 Inhibitor Prevents Cell Growth
Immunohistochemistry The sections were mechanically deparaffinized and incubated in Target Retrieval Solution (Dako, Carpinteria, CA). After blocking the endogenous peroxidase activity with methanol containing 3 % hydrogen peroxide (Dako), the tissue sections were incubated with LSD1 (1:500; Cell Signaling Technology, Beverly, MA) and Pleckstrin homology-like domain family B member 2 (PHLDB2) antibodies (1:100; Novus Biologicals, Littleton, CO) at 4 °C overnight. The sections were then incubated in the secondary antibodies at 37 °C for 60 min. Subsequently, the sections were counterstained with hematoxylin for 1 min. Quantitative Reverse Transcription Polymerase Chain Reaction (qRT-PCR) Total RNA was extracted from the frozen surgical specimens and cell lines using Trizol reagent (Invitrogen, Carlsbad, CA) and reverse transcribed into cDNA using a High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City, CA). Real-time RT-PCR was then carried out using the SsoFast EvaGreen Supermix (Bio-Rad, Hercules, CA) with the following primers (PCReady PCR and Sequencing Primers; Operon, Biotechnology, Tokyo, Japan): PHLDB2, 50 -CTCAGACAGGCCTCAGGAAC-30 and 50 -AAGCGACTGTCAGGCTTTGT-30 . ACTA1 was used as an internal control, with sequences of 50 -CCTTCATCGGTA TGGAGTC-30 and 50 -GTTGGCATACAGGTCCTT-30 . Messenger RNA Preparation and cDNA Microarray Analysis The cells were treated with or without an IC80 concentration of LSD1 inhibitor and collected at 24 h. The cells were subsequently processed for RNA extraction with the RNeasy Plus Mini Kit (Qiagen, Chatsworth, CA). A total of 5.5 lg of total RNA derived from cells cultured with NCL1 was compared to 5.5 lg of total RNA derived from cells cultured without NCL1 using an Affymetrix Human Exon 1.0ST Array (Affymetrix, Santa Clara, CA), currently the most dense array designed specifically for profiling the gene expression. All processing was performed according to described protocols.22 Transfection of Small Interfering RNA Small interfering RNA (siRNA) sequences (Stealth RNAi siRNA; Invitrogen) targeting PHLDB2 (siPHLDB21, Cat. No. HSS131884; siPHLDB2-2, Cat. No. HSS131885) and the negative control (Negative Control Duplex med GC Duplex No. 2: Cat. No.12935-112) were
transfected into the cell lines via electroporation using the Neon Transfection System (Invitrogen). Statistical Analysis The relationships between two variables and numerical values obtained via real-time RT-PCR were analyzed by t test. The patients were divided into two groups of high (n = 43) or low PHLDB2 expression (n = 42). A survival analysis was performed according to the Kaplan–Meier method using the logrank test. Clinical features related to the prognosis were selected for multivariate analysis with a Cox proportional hazard regression model. Statistical significance was defined as a P value of \0.05. All statistical analyses were performed by the JMP statistical package (SAS Institute, Cary, NC). RESULTS Immunohistochemical Analysis of LSD1 Expression in ESCC Cells The LSD1 expression was observed in all specimens examined, and the expression in the tumor tissues was higher in all cases compared to that observed in the corresponding normal epithelium (Fig. 1a). NCL1 exhibited an antiproliferative activity by inducing cell apoptosis; it reduced invasion and migration activities in the human esophageal cancer cell lines. We assessed the antitumor activity of NCL1, an LSD1 inhibitor, in the human esophageal cancer cell lines TE2 and T.Tn, and the sensitivity of these cells was evaluated according to the IC50 values of 34.2 and 31.5 lM, respectively (Fig. 1b). In both the TE2 and T.Tn cells, the TUNEL index showed that NCL1 significantly increased the number of apoptotic cells (Fig. 1d). In addition, NCL1 reduced both invasion and migration activities significantly in ESCC cell lines (Supplementary Fig. S1). Effects of NCL1 on T.Tn and TE2 Xenografts As shown in Fig. 1d, 21 days after the first administration of NCL1, the growth of the T.Tn tumors was significantly suppressed compared to that observed in the control group (P = 0.02). On the other hand, in the TE2 cells, although tumor growth inhibition was observed to some extent, there were no significant differences in tumor size (P = 0.11). NCL1 Induced Histone H3 (Lys4) Dimethylation In order to determine whether LSD1 inhibitors induce histone H3 (Lys4) dimethylation, TE2 and T.Tn cells were
I. Hoshino et al. FIG. 1 a LSD1 expression in ESCC tissues. Immunohistochemical c studies identified aberrant expression in tumor tissues. LSD1 expression levels in tumor tissues were higher in all cases compared to those observed in corresponding normal epithelium. b Antiproliferative activity in T.Tn and TE2. Cells were incubated for 72 h in the presence or absence of indicated concentrations of NCL1, and the degree of cell viability was determined using Cell Counting Kit-8, as described in materials and methods. Results are shown as percentage of viability in control specimens. Points indicate mean of eight separate experiments; bars, SD. c In vitro apoptotic effects of NCL1 in T.Tn and TE2 cells. T.Tn and TE2 cells were exposed to NCL1 at IC80 concentrations for 72 h, and apoptotic cells were stained using TUNEL. Number of positive cells increased in a time-dependent manner in both cell lines. Apoptotic index was defined as percentage of positive cells among 1000 cells in three arbitrary microscopic fields. Data are expressed as mean ± SD. Statistical significance was evaluated by unpaired Student’s t test. d NCL1 inhibition of human esophageal cancer cell tumor growth in vivo. Fragments of human solid tumor were implanted s.c. into the backs of BALB/c nu/nu mice. Drugs were given i.v. to mice 4 times at 4-day intervals with NCL1. Tumor weight was calculated from results of five independent experiments. Points indicate mean; bars, SD. Data were analyzed for statistical significance by Dunnett’s test for multiple comparisons (*P \ 0.05 vs control). e Histone H3 (lys4) dimethylation induced by NCL1. T.Tn and TE2 cells were incubated with different amounts of NCL1 for 24 h and then analyzed for dimethylated histone H3 (Lys4) using western blot analysis, as described in materials and methods. Western blot analysis was performed three times; representative data are shown. Data are expressed as mean ± SD in graph. Statistical significance was evaluated by unpaired Student’s t test (*P \ 0.05 vs control)
A
Tumor
B
Normal
120
%of control
100
T.Tn TE2
80 60 40 20 0 0
1
2
3
6
10
20
30
60
100
Concentration (µm)
C
T.Tn
TE2
control
IC50
control
(%)
p=0.023
p=0.028
40
IC50
20 0
IC50
control
In order to determine the target genes of LSD1 in esophageal cancer cell lines, we conducted a cDNA microarray analysis. Twofold or greater alterations in the gene expression, either increasing or decreasing, were considered significant. Eighteen genes were found to be upregulated and nine were found to be down-regulated in both cell lines (Tables 1, 2). Then we determined whether NCL1 can reduce the expression levels of one of the downregulated genes in microarray data. We chose PHLDB2. T.Tn and TE2 cells were exposed to NCL1 at various concentrations for 24 h or various time period at IC80 for 24 h. As expected, NCL1 reduced PHLDB2 expression levels in a dose-dependent and time-dependent manner (Fig. 2a). Effects of Transfection of siRNA Targeting PHLDB2 mRNA The introduction of each of the two siRNA sequences (siPHLDB2-1, siPHLDB2-2) into TE2 and T.Tn cells led to a significant decrease in the level of the target mRNA at
350 300 250 200 150 100 50 0
T.Tn control 2.5mg/kg 25mg/kg
0
3
7
10
14
17
21
1400 1200 1000 800 600 400 200 0
IC50 TE2 control 2.5mg/kg 25mg/kg
0
3
Days after Treatment
E
7
10
14
17
21
Days after Treatment T.Tn
TE2 H3K4 Beta-actin 0
Relative expression levels (normalized to Beta-actin)
NCL1 Regulated Various Gene Expression Levels
control
D Tumor Weight (mg)
exposed to NCL1 at IC80 for 24 h. As expected, NCL1 induced histone H3 (Lys4) dimethylation in a dose-dependent manner (Fig. 1e).
0.1
1.0 10
100
0
0.1 1.0 10 100
4
(µm)
T.Tn
TE2 5
*
*
*
4
3
3 2 2 1 0
1 0
0.1
1.0
10
100
0 0
0.1
1.0
10
100 (mM)
48 h (Fig. 2b). The degree of cell proliferation in the PHLDB2 down-regulated cells was compared to that observed in mock cells, which revealed that cell proliferation was significantly reduced after 96 h in both cell lines (Fig. 2c). TUNEL index showed that siPHLDB2 significantly increased the number of apoptotic cells (Fig. 2d). In both the migration and Matrigel invasion assays, the rate
LSD1 Inhibitor Prevents Cell Growth TABLE 1 List of 18 genes up-regulated ([2-fold changes) in T.Tn and TE2 cells Genbank accession no.
Gene symbol
Description
BC005008
CEACAM6
BC012172 BC008723 L19501 M29540
Maximum fold change T.Tn
TE2
Carcinoembryonic antigen-related cell adhesion molecule 6 (nonspecific cross-reacting antigen)
2.84
3.62
ACSS2
Acl-CoA synthetase short-chain family member 2
1.10
1.24
ASNS CBS
Asparagine synthetase (glutamine hydrolyzing) Cystathionine-beta-synthase
1.92 1.15
1.96 1.18
CEACAM5
Carcinoembryonic antigen-related cell adhesion molecule 5
2.45
2.90 2.11
BC019625
CHAC1
ChaC, cation transport regulator homolog 1 (E. coli)
1.34
U03688
CYP1B1
Cytochrome P4509, family 1, subfamily B, polypeptide 1
1.49
1.67
BC007333
ETV5
Ets variant 5
1.73
1.81
AF110400
FGF19
Fibroblast growth factor 19
1.17
1.26
EF152283
GCNT3
Glucosaminyl (N-acetyl) transferase 3, mucin type
1.17
1.66
AF019770
GDF15
Growth differentiation factor 15
2.01
1.75
BC033089
LCN2
Lipocalin 2
1.62
2.78
BC004863
PSAT1
Phosphoserine aminotransferase 1
1.34
2.87
AF539739
S100P
S11 calcium binding protein P
1.26
2.06
AF097514
SCD
Stearoyl-CoA desaturate (delta-9-desaturase)
1.25
2.38
BC000658
STC2
Stanniocalcin 2
1.95
1.59
BGC011703
TMPRSS4
Transmembrane protease, serine 4
1.03
1.24
AF022375
VEGFA
Vascular endothelial growth factor A
1.62
1.50
TABLE 2 List of genes down-regulated ([2-fold changes) in T.Tn and TE2 cells Genbank accession no.
J04948
Gene symbol
ALPPL2
Description
Maximum fold change
Alkaline phosphatase, placental-like 2
T.Tn
TE2
-1.01
-1.69
BC098561
EFEMP1
EGF-containing fibulin-like extracellular matrix protein 1
-1.59
-1.76
AB031548
GPR87
G protein-coupled receptor S7
-1.11
-1.06
M19154
TGFB2
Transforming growth factor, beta 2
-1.72
-1.11
BC142678
PHLDB2
Pleckstrin homology-like domain, family B, member 2
-1.14
-1.04
BC146868
COL12A1
Collagen, type XII, alpha 1
-1.13
-1.22
BC017782 AF098807
WISP2 LHFP
WNT1 inducible signaling pathway protein 2 Lipoma HMGIC fusion partner
-1.12 -1.21
-1.31 -1.26
AK123348
C3orf57
Chromosome 3 open reading frame 57
-1.14
-1.72
of penetration of cells through the membrane was significantly decreased after the transfection of each siRNA (Fig. 2e). In addition to T.Tn and TE2 ESCC cell lines, we explored the role of siPHLDB2 in other ESCC cell lines. siPHLDB2 showed an antiproliferation effect on various ESCC cell lines (Fig. S2). PHLDB2 Expression in ESCC Tissues The PHLDB2 expression level in the ESCC tissues was significantly higher than that observed in the corresponding normal epithelium (P = 0.002; Fig. 3a). In addition, we examined the PHLDB2 protein expression in the ESCC
specimens and normal epithelium (Fig. 3b); the PHLDB2 expression in the cytoplasm was observed in all specimens examined. We next evaluated the correlations between the PHLDB2 mRNA expression and various clinicopathologic factors, including sex, age, T factor, N factor, stage, survival, and recurrence (Table 3) after dividing the 85 patients into two groups, a high-expression group (n = 43) and a low-expression group (n = 42). Although none of the factors was found to be significantly related to the PHLDB2 expression (Table 3), a survival analysis using the Kaplan–Meier method and the logrank test showed a significant decrease in overall survival in the highPHLDB2 expression group, with a 5-year survival rate of
I. Hoshino et al.
(%)
140 120
100
*
80
100 80
60 40
**
**
12
24
48 (h)
*
*
*
12
24
48 (h)
40
20
20
0
0 0
0.1
1
10
100 (µm)
0
1
6
TE2 (%)
(%) 160 140 120 100 80 60 40 20 0
140 120 100 80
*
60 40 20 0 0.1
1
10
0
100 (µm)
KDMI concentration
B
Relative expression levels (normalized to ACTA1)
120
100
100
80
80
60
60 40
20
*
*
si1
si2
*
*
si1
si2
20 0
0
NC
T.Tn Relative growth curve
6
TE2
(%)
120
40
C
1
Time after KDMI administration
T.Tn
(%)
NC
TE2
7
7
6
6
5
5
4
4
3
3
*
2
*
1
D
2
*
1
48
72
96
24
48
T.Tn
control
(%)
si1
si2
control
(h)
si1
si2
*
*
si1
si2
*
10
10
5
5
0
0 control
Migration
96
15 *
(%)
72
TE2
15
(%)
NC si1 si2
*
0
0 24
E
**
60
0
DISCUSSION We demonstrated that LSD1 is overexpressed in ESCC specimens and may be a potent therapeutic target in the setting of ESCC. Furthermore, dimethylation of Lys4 of histone H3 in the ESCC cell lines by NCL1 regulated various gene expression levels and is therefore essential for the observed antitumor effect. To date, more than 20 methyl marks on lysine and arginine residues have been identified.23 One of the first methyl marks to be characterized in detail is methylation at Lys4 of histone H3 (H3K4me). According to the histone code hypothesis, single histone modifications or combinations of modifications can be recognized by effector proteins and/or protein complexes that read the marks, converting them into specific functional chromatin states and thereby regulating downstream responses.24,25 In fact, methylation at H3K4 is reported to be strongly associated with transcriptional activation. In contrast, trimethylation at H3K9, H3K27, and H4K20 is commonly considered to be a repressive modification.26 LSD1 removes the methyl groups from mono- and dimethylated Lys4 of histone H3 (H3K4me1/2).27 LSD1 is essential for many biological processes involved in development, growth, differentiation, and morphogenesis.28 LSD1 was reported to interact with the Mi-2/nucleosome remodeling and deacetylase (NuRD) complex and the LSD1/NuRD complexes control several cellular signaling pathways including transforming growth factor b1 signaling pathway.29 Interestingly, the overexpression of LSD1 has been shown to correlate with adverse clinical outcomes in patients with neuroblastoma, and the pharmacological inhibition of LSD1 reduces neuroblastoma growth in xenografted nude mice in vivo.30 In addition, a high expression level of LSD1 in patients with prostate cancer may predict an aggressive tumor biology.28 Meanwhile, one study showed that the overexpression of
T.Tn 120
Invasion
40.0 % (vs 69.1 % in the low-PHLDB2 expression group) (Fig. 3c). On the basis of the results of a multivariate analysis (Table 4), a high expression level of PHLDB2 was found to be the strongest independent risk factor for a poor survival, with a hazard ratio of 2.545 (95 % confidence interval 1.218–5.516; P = 0.013).
A (%) Relative expression levels (normalized to ACTA1)
FIG. 2 a NCL1 reduced PHLDB2 expression levels in a dose-and c time-dependent manner. b Quantification of mRNA of PHLDB2 after transfection of specific siRNA with real-time quantitative PCR. *P \ 0.01. c Silencing of PHLDB2 decreased rate of proliferation of ESCC cells. *P \ 0.05. d TUNEL index showed that siPHLDB2 significantly increased number of apoptotic cells. e Results of Matrigel invasion assay showed that silencing PHLDB2 reduces invasive potential of ESCC cells. NC negative control. *P \ 0.05
si1
si2
control
T.Tn
TE2
140
120
120
100
100
80
80
*
60
*
60
*
*
*
*
si1
si2
40
40 20
20
0
0
140
140
120
120
100
100
80
80
60
*
40
*
60 40
20
20
0
0
NC
si1
si2
NC
LSD1 Inhibitor Prevents Cell Growth * Relative mRNA expression
A
TABLE 3 Correlation between PHLDB2 expression and clinicopathologic features of esophageal squamous cell carcinoma
30.0 25.0 20.0 15.0 10.0 5.0 0.0
Characteristic n Total (%) Tumor
High PHLDB2 expression, n (%)
Low PHLDB2 expression, n (%)
P
85 43 (50.6)
42 (49.4)
70 36 (42.4) 15 7 (8.2)
34 (40.0) 8 (9.4)
0.9600
0.4172
Sex (%)
Normal
Male Female
B
Age (%) \65
32 18 (21.2)
14 (16.5)
C65
53 24 (29.4)
28 (32.9)
Tumor depth (%) Tumor
Overall survival rate
C
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
(×400)
Normal
T1-2
35 15 (17.7)
20 (23.5)
T3-4
50 28 (32.9)
22 (25.9)
0.2329
Lymph node metastasis (%) Low expression (n=42)
High expression (n=43) 5-year OS rate Low: 63.14% High: 40.00% 0
Negative
35 20 (23.5)
15 (17.7)
Positive
50 23 (27.0)
27 (31.8)
I–II
40 18 (21.2)
22 (25.9)
III–IV
45 25 (29.4)
20 (23.5)
0.3119
Stage (%) 0.3313
Patient survival (%)
20
40
60
Months
Alive Dead
52 23 (27.1) 33 20 (23.5)
29 (34.1) 13 (15.3)
0.1411
0.5526
Recurrence (%) FIG. 3 a Expression levels of PHLDB2 in ESCC clinical specimens. Total RNA was isolated from 85 tumor tissues and matched noncancerous tissues. PHLDB2 expression levels were detected using quantitative real-time PCR. *P \ 0.05. b PHLDB2 protein expression in clinical ESCC specimens. Immunohistochemical staining for PHLDB2 showed high expression in cancer tissue (left) and low expression in neighboring normal epithelium (right). c Results of Kaplan-Meier survival analysis of overall survival in high-PHLDB2 expression group versus low-PHLDB2 expression group. Results showed significant decrease in overall survival in high-PHLDB2 expression group (P \ 0.001), with 5-year survival rate of 40.0 % (vs 63.1 % in low-PHLDB2 expression group)
LSD1 is associated with a poor prognosis in patients with non-small cell lung cancer.9 Recently, ORY-1001, a LSD1 inhibitor, is entering phase 1/2 clinical study in acute myeloid leukemia.31 ORY-1001 significantly reduces tumor cells and increases survival time in mouse models of acute lymphoblastic leukemia. Therefore, the inhibition of LSD1 may provide a promising epigenetic target for cancer therapy. Indeed, we showed that LSD1 is overexpressed in ESCC specimens and that NCL1 exerted an antitumor effect in ESCC cell lines. In our experiments, tumor growth was repressed by LSD1 inhibitor injection in the xenograft mice models. LSD1 regulates various gene expression levels. Therefore, LSD1 inhibitors may exhibit efficacy in downregulating oncogenes or up-regulating tumor suppressor genes. Lim et al. reported that the knockdown of LSD1 using an siRNA approach induces the regulation of several
No
54 26 (30.6)
28 (32.9)
Yes
31 17 (20.0)
14 (16.5)
proliferation-associated genes.15 Furthermore, LSD1 knockdown induces changes in the mRNA expression as analyzed according to a microarray analysis.30 In the present study, a microarray analysis of the transcriptome was performed to determine the mRNA expression after LSD1 inhibitor exposure. We then focused on PHLDB2, one of the down-regulated genes, the antitumor effect of NCL1 may be derived from the down-regulation of oncogenes and/or up-regulation of tumor suppressor genes, and the molecule might act as an oncogene according to the latest reports. PHLDB2 was typically down-regulated after LSD1 inhibitor administration in vitro. The expression levels in the ESCC specimens were significantly higher than those observed in the normal esophageal epithelium specimens. PHLDB2 is a 1253 amino acid protein and has been identified to be a microtubule-anchoring factor that attaches EB1/CLIP-associating protein (CLASP)–bound microtubule plus ends to the cell cortex.32 Hotta et al. used the noncancerous human mammary epithelial cell line MCF-10A to investigate the mechanisms determining the cortical distribution of PHLDB2 and showed that PHLDB2 is colocalized with laminin-5 deposition and laminin receptor integrins as well as epithelial tissues and that the localization of PHLDB2 relies on the actions of laminin
I. Hoshino et al. TABLE 4 Results of multivariate analysis for survival Variable
Category
Hazard ratio
95 % confidence interval
P
Tumor depth
T3-4
8.646
1.914–65.937
0.003
Lymph node metastasis
Positive
7.345
1.514–55.123
0.012
Stage
III or IV
0.307
0.037–1.813
0.197
Ly
Positive
1.111
0.497–2.667
0.062
V
Positive
0.902
0.338–2.601
0.842
PHLDB2
High
2.545
1.218–5.516
0.013
Bold values are statistically significant
receptor integrins.33 In addition, Carles et al. found that PHLDB2 was overexpressed in 10/12 head and neck squamous cell carcinoma specimens using RT-PCR.34 These findings support our results regarding PHLDB2 overexpression in ESCC specimens. Moreover, in the siPHLDB2-transfected cells, the inhibition of proliferation, migration, and invasion was observed. We subsequently divided the patients into two groups, a high-PHLDB2 expression group and a low-PHLDB2 expression group, based on the mRNA expression level, and found no significant correlations with any clinicopathologic features in a univariate analysis. According to the multivariate analysis, a high expression level of PHLDB2 was found to be an independent poor prognostic factor. These results suggest that PHLDB2 is a valuable diagnostic and prognostic biomarker for ESCC. In conclusion, LSD1 inhibitors have potent antitumor effects on ESCC cell lines. The antitumor efficacy of LSD1 inhibitors may be derived from their ability to modulate the mRNA expression, while PHLDB2 is an oncogenic factor and promising target in ESCC therapy. DISCLOSURE
The authors declare no conflict of interest.
REFERENCES 1. Hongo M, Nagasaki Y, Shoji T. Epidemiology of esophageal cancer: orient to occident. Effects of chronology, geography and ethnicity. J Gastroenterol Hepatol. 2009;24:729–35. 2. Ye T, Sun Y, Zhang Y, Zhang Y, Chen H. Three-field or twofield resection for thoracic esophageal cancer: a meta-analysis. Ann Thorac Surg. 2013;96:1933–41. 3. Sugiura T, Uesaka K, Mihara K, et al. Margin status, recurrence pattern, and prognosis after resection of pancreatic cancer. Surgery. 2013;154:1078–86. 4. Zhimin G, Noor H, Jian-Bo Z, Lin W, Jha RK. Advances in diagnosis and treatment of hilar cholangiocarcinoma—a review. Med Sci Monit. 2013;19:648–56. 5. Ando N, Kato H, Igaki H, et al. A randomized trial comparing postoperative adjuvant chemotherapy with cisplatin and 5-fluorouracil versus preoperative chemotherapy for localized advanced squamous cell carcinoma of the thoracic esophagus (JCOG9907). Ann Surg Oncol. 2012;19:68–74. 6. Hoshino I, Matsubara H. Recent advances in histone deacetylase targeted cancer therapy. Surg Today. 2010;40:809–15.
7. Hoshino I, Matsubara H, Hanari N, et al. Histone deacetylase inhibitor FK228 activates tumor suppressor Prdx1 with apoptosis induction in esophageal cancer cells. Clin Cancer Res. 2005;11: 7945–52. 8. Liu J, Liu FY, Tong ZQ, et al. Lysine-specific demethylase 1 in breast cancer cells contributes to the production of endogenous formaldehyde in the metastatic bone cancer pain model of rats. PLoS One. 2013;8:e58957. 9. Lv T, Yuan D, Miao X, et al. Over-expression of LSD1 promotes proliferation, migration and invasion in non–small cell lung cancer. PLoS One. 2012;7:e35065. 10. Ververis K, Hiong A, Karagiannis TC, Licciardi PV. Histone deacetylase inhibitors (HDACIs): multitargeted anticancer agents. Biologics. 2013;7:47–60. 11. An W. Histone acetylation and methylation: combinatorial players for transcriptional regulation. Subcell Biochem. 2007;41:351– 69. 12. Kerenyi MA, Shao Z, Hsu YJ, et al. Histone demethylase Lsd1 represses hematopoietic stem and progenitor cell signatures during blood cell maturation. Elife. 2013;2:e00633. 13. Kashyap V, Ahmad S, Nilsson EM, et al. The lysine specific demethylase-1 (LSD1/KDM1A) regulates VEGF-A expression in prostate cancer. Mol Oncol. 2013;7:555–66. 14. Zhao ZK, Yu HF, Wang DR, et al. Overexpression of lysine specific demethylase 1 predicts worse prognosis in primary hepatocellular carcinoma patients. World J Gastroenterol. 2012; 18:6651–6. 15. Lim S, Janzer A, Becker A, et al. Lysine-specific demethylase 1 (LSD1) is highly expressed in ER-negative breast cancers and a biomarker predicting aggressive biology. Carcinogenesis. 2010;31: 512–20. 16. Murray-Stewart T, Woster PM, Casero RA Jr. The re-expression of the epigenetically silenced e-cadherin gene by a polyamine analogue lysine-specific demethylase-1 (LSD1) inhibitor in human acute myeloid leukemia cell lines. Amino Acids. 2014;46: 585. 17. Wu Y, Steinbergs N, Murray-Stewart T, Marton LJ, Casero RA. Oligoamine analogues in combination with 2-difluoromethylornithine synergistically induce re-expression of aberrantly silenced tumour-suppressor genes. Biochem J. 2012;442:693–701. 18. Willmann D, Lim S, Wetzel S, et al. Impairment of prostate cancer cell growth by a selective and reversible lysine-specific demethylase 1 inhibitor. Int J Cancer. 2012;131:2704–9. 19. Ueda R, Suzuki T, Mino K, et al. Identification of cell-active lysine specific demethylase 1-selective inhibitors. J Am Chem Soc. 2009;131:17536–7. 20. Hamada S, Suzuki T, Mino K, et al. Design, synthesis, enzymeinhibitory activity, and effect on human cancer cells of a novel series of jumonji domain-containing protein 2 histone demethylase inhibitors. J Med Chem. 2010;53:5629–38. 21. Calin GA, Croce CM. MicroRNA signatures in human cancers. Nat Rev Cancer. 2006;6:857–66.
LSD1 Inhibitor Prevents Cell Growth 22. Pitroda SP, Wakim BT, Sood RF, et al. STAT1-dependent expression of energy metabolic pathways links tumour growth and radioresistance to the Warburg effect. BMC Med. 2009;7:68. 23. Mosammaparast N, Shi Y. Reversal of histone methylation: biochemical and molecular mechanisms of histone demethylases. Annu Rev Biochem. 2010;79:155–79. 24. Turner BM. Decoding the nucleosome. Cell. 1993;75:5–8. 25. Strahl BD, Allis CD. The language of covalent histone modifications. Nature. 2000;403(6765):41–5. 26. Barski A, Cuddapah S, Cui K, et al. High-resolution profiling of histone methylations in the human genome. Cell. 2007;129: 823–37. 27. Shi Y, Lan F, Matson C, et al. Histone demethylation mediated by the nuclear amine oxidase homolog LSD1. Cell. 2004;119: 941–53. 28. Metzger E, Wissmann M, Yin N, et al. LSD1 demethylates repressive histone marks to promote androgen-receptor-dependent transcription. Nature. 2005;437(7057):436–9.
29. Wang Y, Zhang H, Chen Y, et al. LSD1 is a subunit of the NuRD complex and targets the metastasis programs in breast cancer. Cell. 2009;138:660–72. 30. Schulte JH, Lim S, Schramm A, et al. Lysine-specific demethylase 1 is strongly expressed in poorly differentiated neuroblastoma: implications for therapy. Cancer Res. 2009;69:2065–71. 31. Ketscher A, Jilg CA, Willmann D, et al. LSD1 controls metastasis of androgen-independent prostate cancer cells through PXN and LPAR6. Oncogenesis. 2014;3:e120. 32. Lansbergen G, Grigoriev I, Mimori-Kiyosue Y, et al. CLASPs attach microtubule plus ends to the cell cortex through a complex with LL5beta. Dev Cell. 2006;11:21–32. 33. Hotta A, Kawakatsu T, Nakatani T, et al. Laminin-based cell adhesion anchors microtubule plus ends to the epithelial cell basal cortex through LL5alpha/beta. J Cell Biol. 2010;189:901–17. 34. Carles A, Millon R, Cromer A, et al. Head and neck squamous cell carcinoma transcriptome analysis by comprehensive validated differential display. Oncogene. 2006;25:1821–31.
