Tumor Biol. DOI 10.1007/s13277-015-4610-9

ORIGINAL ARTICLE

Clinicopathological significance of p38β, p38γ, and p38δ and its biological roles in esophageal squamous cell carcinoma Shutao Zheng 1,2 & Chenchen Yang 1,2 & Tao Liu 1,2 & Qing Liu 1,2 & Fang Dai 1,2 & Ilyar Sheyhidin 2 & Xiaomei Lu 1,2,3

Received: 27 October 2015 / Accepted: 7 December 2015 # International Society of Oncology and BioMarkers (ISOBM) 2015

Abstract P38β, p38γ, and p38δ have been sporadically and scarcely reported to be involved in the carcinogenesis of cancers, compared with p38α isoform. However, little has been known regarding their clinicopathological significance and biological roles in esophageal squamous cell carcinoma (ESCC). Expression status of p38β, p38γ, and p38δ was assayed using immunohistochemistry with ESCC tissue microarray; ensuing clinicopathological significance was statistically analyzed. To define its biological roles on proliferation, migration and invasion of ESCC cell line Eca109 in vitro, MTT, wound healing, and Transwell assays were employed, respectively. As confirmation, athymic nude mice were taken to verify the effect over proliferation in vivo. It was found that both p38β and p38δ expression, other than p38γ, were significantly higher in ESCC tissues compared with paired normal controls. In terms of prognosis, only p38β expression was observed to be significantly associated with overall prognosis. Clinicopathologically, there was significant association

between p38γ expression and clinical stage, lymph nodes metastases, and tumor volume. No significant association was found for p38β and p38δ between its expression and other clinicopathological parameters other than significant difference of expression between ESCC versus normal control. In Eca109, it was observed that p38β, p38γ, and p38δ can promote the cell growth and motility. As verification, overexpression of p38δ can promote, whereas knockdown of p38γ can prevent, the tumorigenesis in nude mice model xenografted with Eca109 cells whose basal level of p38δ was stably over-expressed and p38γ was stably knocked down. Together, our results demonstrate that p38β, p38γ, and p38δ played oncogenic roles in ESCC. Keywords Esophageal squamous cell carcinoma . p38β . p38γ . p38δ . Prognosis . Metastasis

Introduction Electronic supplementary material The online version of this article (doi:10.1007/s13277-015-4610-9) contains supplementary material, which is available to authorized users. * Xiaomei Lu [email protected] 1

Clinical Medical Research Institute, First Affiliated Hospital of Xinjiang Medical University, Xinjiang Uygur Autonomous Region Urumqi, People’s Republic of China

2

State Key Lab Incubation Base of Xinjiang Major Diseases Research, First Affiliated Hospital of Xinjiang Medical University, Xinjiang Uygur Autonomous Region Urumqi, People’s Republic of China

3

Clinical Medical Research Institute, State Key Lab Breeding Base of Xinjiang Major Diseases Research, First Affiliated Hospital of Xinjiang Medical University, Urumqi 830054, Xinjiang Uygur Autonomous Region, People’s Republic of China

p38 mitogen-activated protein kinases (MAPKs) are a class of mitogen-activated protein kinases that are responsive to stress stimuli, such as cytokines, ultraviolet (UV) irradiation, and heat and osmotic shock and are involved in cell differentiation, apoptosis, and autophagy [1]. There are four members of the mammalian p38 MAPK, namely, p38α (MAPK14), p38β (MAPK11), p38γ (MAPK12), and p38δ (MAPK13), with amino acid sequences being identical about 60 % but different in their expression patterns, substrate specificities, and sensitivities to chemical inhibitors, suggesting that different isoforms may mediate different roles [1, 2]. Most of the work published on p38 MAPK pathway involved in carcinogenesis has been focused on studying the role of the p38α isoform, which is widely referred to as p38 in the literature. However, unlike p38α, there have been limited studies about the role of

Tumor Biol.

the other p38 isoforms (β, γ, and δ) in carcinogenesis [2], especially in esophageal squamous cell carcinoma (ESCC). We have previously found p38α to play an anti-oncogenic role in ESCC [3]; however, the roles and implication of the other three p38 isoforms in ESCC have not been defined and investigated so far. Among p38β, p38γ, and p38δ, only p38δ has been mentioned and reported to play anti-oncogenic-like role in ESCC, and loss of p38δ could promote both proliferation, migration, and increases resistance to cisplatin and 5-fluorouracil treatment in ESCC cells ex vivo [4, 5]. However, unlike ESCC that originates squamous epithelial, p38δ has been found to play an oncogenic role in colitis-associated colon adenocarcinoma [6]. Furthermore, p38γ isoform has not been investigated in ESCC till now other than studies carried out in breast cancer [7, 8] and colon adenocarcinoma [6, 9] that mainly defined as oncogenic role. But in terms of whether p38γ could promote metastasis, their results are inconsistent and somewhat controversial even in the same kind of breast cancer [7, 8]. Meanwhile, no study has been available regarding p38β in ESCC either, with the exception of one earlier report that identified p38β as a potential therapeutic target in pancreatic cancer [10] and another report that p38β enhanced the bone metastasis of breast cancer [11]. In consideration of the publications mentioned earlier concerning p38β, p38γ, and p38δ in the involvement of cancers, there is necessity to further clarify their roles in ESCC. In the present study, we have comprehensively investigated the expression of p38β, p38γ, and p38δ in ESCC and paired normal control tissues in vivo. We have also characterized the biological roles using over-expression and knockdown of p38β, p38γ, and p38δ in ESCC cell line Eca109, including proliferation, migration, and invasion ex vivo. To our knowledge, our study comprehensively investigates the role of p38β, p38γ, and p38δ in the setting of ESCC for the first time.

Materials and methods Tissue microarray The study was approved by the Medical Ethics Committee of the First Affiliated Hospital of Xinjiang Medical University, and all animals involved were performed in accordance with approved experimental animal guidelines and regulations of the First Affiliated Hospital of Xinjiang Medical University. Tissue microarrays used for immunohistochemical analysis of p38β, p38γ, and p38δ were commercially from Shanghai Outdo Biotech. Co., Ltd. (Shanghai, China). The microarray was composed of 86 cases of ESCC and 78 paired adjacent normal tissues. Staging and grading were assessed following the World Health Organization classification. None of the patients received chemoradiotherapy before surgical operation. Informed written consents were obtained from all subjects involved.

Cell culture The human ESCC cell line Eca109 was purchased from Wuhan University (Hubei; Wuhan), which was cultured in RPMI-1640 medium (Gibco Life Technologies, USA) supplemented with 10 % heat-inactivated fetal bovine serum (FBS, GibcoLife Technologies, USA) at 37 °C with 5 % CO2 in a humidified incubator. Construction and cell transfection Eukaryotic expression vectors containing full-length complementary DNA (cDNA) of p38α, p38β, p38γ, and p38δ fused with Flag Tag in the carboxyl terminal were purchased from Biogot technology (Nanjing, China). The small interference RNA (siRNA) target sequences for p38α, p38β, p38γ, and p38δ were referenced from article by Kukkonen-Macchi A et al. [12]. For a vector-based RNAi approach, a double-stranded short hairpin RNA (shRNA) was cloned into the BamHI-XhoI sites of the pRNA-U6.1/neo-enhanced green fluorescent protein (eGFP) vector (GenScript Corporation). Based on the siRNA sequences, specific shRNA sequence of each p38 isoform was listed in Supplementary Table 1. Italic letters underlined and boldfaced stand for the sites of the restriction enzymes Bam HI and Xho I, hairpin loop and terminal signal, respectively. The construct was confirmed by sequencing. The scramble sequence was inserted into the same plasmid to generate a control vector. Each constructed and purchased vector after verification by sequencing was transfected in ESCC cell line Eca109 cells using Lipofectamine 2000 (Invitrogen Life Technologies, CA, USA) according to the manufacturer’s instruction and eGFP expression was monitored by fluorescent microscopy. After 48 h of transfection, cells were harvested for analysis at 48 h for messenger RNA (mRNA) level by quantitative reversal transcription PCR (qRT-PCR) and 72 h for protein level analysis by western blot. Real-time fluorescent quantitative PCR Total RNA was extracted from cells in each group with TRIzol reagent (Invitrogen Life Technologies, CA, USA) following the manufacturer’s protocol. Both the purity and integrity of RNA extracted was determined by NanoDrop ND1000 equipment. Genomic DNA contamination was removed by adding DNase I according to the manufacturer’s instruction (TaKaRa, Dalian, China), then total RNA were reversely transcribed to cDNA using the prime SCRIPT™ RT-PCR kit (TaKaRa, Dalian, China). The qRT-PCR assay was carried out with the IQ5 system (Bio-Rad, USA) using SYBR Premix Ex Taq (TaKaRa) according to the manufacturer’s instructions. After normalization with β-actin internal control, relative gene expression was determined using a relative standard curve method. The primers for each p38 isoform and β-actin were listed in Supplementary Table 2. All reactions were performed independently in triplicate. The reaction mixture for both p38 isoforms and β-actin were incubated at the following thermal

Tumor Biol.

cycling conditions: 95 °C for 3 min, and 40 cycles at 95 °C for 5 s followed by 56.5 °C for 30 s. Western blotting Seventy-two hours after transfection, cells were harvested in radio-immunoprecipitation assay (RIPA) lysis buffer (Bioteke, Beijing, China) and 50 μg of cellular protein were subjected to 10 % SDS-PAGE separation. Proteins were transferred to polyvinylidene fluoride (PVDF) microporous membrane (Millipore, Boston, MA, USA) and blots were probed with rabbit polyclonal antibody against p 3 8 α ( 1 4 0 6 4 - 1 - A P ) , p 3 8β ( 1 7 3 7 6 - 1 - A P ) , p 3 8 γ (20184-1-AP), and p38δ (10217-1-AP) at dilution 1:800 (Proteintech, China). β-Tubulin (sc-9104, Santa Cruz) was chosen as an internal control and the blots were visualized with western breeze Kit (WB7105, Invitrogen Life Technologies, CA, USA). Immunohistochemistry Immunohistochemical stains were performed using heat-induced epitope retrieval, an avidin-biotin complex method. The rabbit anti- p38α, p38β, p38γ, and p38δ antibody (Proteintech, China) was diluted at 1:100. The sections were evaluated by light microscopic examination, and cellular localization of the protein and immunostaining level in each section was assessed blindly by two separate pathologists. The staining patterns were scored as follows: negative, represented by −; weak (less than 30 % of cells with positive staining), represented by +; moderate (more than 30 % but less than 60 % with positive staining), represented by ++; and strong positive (more than 60 % of cells with positive staining) according to the signal intensity. The negative control for IHC was the substitution of serum IgG at the same protein concentration as the primary antibody, as recommended by Hewitt SM et al. [13], and evaluation of antibody specificity was through antigen preadsorption method as suggested by Burry RW et al. [14]. Cell proliferation assay Methylthiazolyl blue tetrazolium (MTT; Sigma-Aldrich, St. Louis, MO) spectrophotometric dye assay was used to observe and compare cell proliferation ability. ESCC cells were plated in 96-well plates at a density of 4 × 103 cells per well. After transfection experiments, cell proliferation was assessed. Cells were incubated for 4 h in 20 μL MTT at 37 °C. The color was developed by incubating the cells in 150 μL dimethyl sulfoxide (DMSO), and the absorbance was detected at 490 nm wave length. The data were obtained from three independent experiments. Cell migration and invasion assays in vitro Cell migration ability was calculated by the wound healing assay. ESCC cells were plated in a 6-well plate at a concentration of 5 × 105 cells/ well and allowed to form a confluent monolayer for 24 h. After the transfection, the monolayer was scratched with a sterile pipette tip (10 μL), washed with serum-free medium

to remove floating and detached cells, followed by culturing in the setting of serum-free media. Then, photographs were taken (time 0, 24, 48, and 72 h) using inversion fluorescence microscope (Olympus, Takachiho Seisakusho, Japan). Cell culture inserts (24-well, pore size 8 μm; BD Biosciences) were seeded with 5 × 103 cells in 100 μL of medium with 0.1 % fetal bovine serum (FBS). Inserts pre-coated with Matrigel (40 μL, 1 mg/mL; BD Biosciences) were used for invasion assays. Media with 10 % FBS (500 μL) was added to the lower chamber and served as a chemotactic agent. Noninvasive cells were wiped from the upper side of the membrane and cells on the lower side were fixed in cold methanol (−20 °C) and air dried. Cell were stained with 0.1 % crystal violet (dissolved in methanol) and counted using the inverted microscope. Cell cycle and apoptosis analysis Cell cycle and apoptosis were analyzed using flow cytometry (FCM). For cell cycle analysis, cells were plated at a density of 3 × 105 per well in 6-well plates and transfected with plasmid and control vector. Cells were washed twice with cold phosphate buffer saline (PBS) and fixed in 70 % cold ethanol overnight at 4 °C. After incubation with RNase for 1 h at 4 °C, DNA was stained with 2 μL propidium iodide (PI) (500 mg/mL) for 15 min, then analyzed by FCM. For cell cycle analysis, an Annexin V-FITC apoptosis detection kit was used (Invitrogen Life Technologies, CA, USA), and Annexin V-FITC staining was performed according to instructions provided by the manufacturer. Briefly, cells were washed twice with cold PBS and resuspended in 400 μL with 1× binding buffer at a concentration of 1 × 106 cells/mL. Cells were then mixed with 5 μL of the Annexin V-FITC solution and 2 μL of propidium iodide (PI). Cells were incubated for 15 min at 4 °C in the dark, and then analyzed by FCM. Xenografted athymic nude mice model The animal experiment was approved by the Animal Ethnics Committee of the First Affiliated Hospital of Xinjiang Medical University. To evaluate tumorigenic ability of p38γ and p38δ in vivo, Eca109 xenografting mouse model was established. Female BALB/c-nu mice of 3 weeks old were prepared for Eca109 cells implantation. All animals were maintained in a sterile environment on a daily 12-h light/12-h dark cycle, and the 32 mice were grouped into 4, with 8 in each group. Eca109 cells transfected with Lentiviral- shRNA-p38γ as well as Lentiviral-over-p38δ and its control vectors were subjected to sorting using FCM to be purified. Followed by subcutaneous injection (4 × 106/mouse) into the right flank of the nude mice. After 5 weeks, all the 32 mice were euthanatized. Tumor xenografts were harvested and weighted. Tumor volume (TV) was calculated weekly for 5 weeks according to the formula, TV (mm3) = length × width2 × 0.5.

Tumor Biol.

P38γ expression is significantly associated with lymph node metastasis, clinical stage, and tumor volume phenotype in ESCC First of all, given the high similarity in amino acid sequences of p38β, p38γ, and p38δ and to avoid

underlying cross-immunoreactions of the antibodies, we carried out the antigen preadsorption trial before the experiment to test and evaluate the antibody specificity according to the strong suggestion given by Burry RW et al. [14]. The antigen preadsorption trial results showed that the commercially obtained antibodies have adequate specificity of each other (data not shown). To investigate the p38β, p38γ, and p38δ expression status in ESCC, we have first evaluated the expression of p38β, p38γ, and p38δ in human ESCC clinical sample tissue arrays. We found that p38β, p38γ, and p38δ were all expressed in ESCC and paired normal control tissues but just with different immunostaining status for each specific isoform (Fig. 1). Specifically, both p38β and p38δ were significantly expressed higher in ESCC as compared with paired normal control tissues (p = 0.000, respectively); while, there was no significant differential expression of p38γ between ESCC and paired normal control (p = 0.365) (Table 1). In addition, there was also significant correlation between p38β expression and clinical tumor volume (p = 0.044). Namely, the larger the volume of tumor was, the higher the expression of p38β will be. However, no significant association was observed after analysis between p38δ expression and other clinicopathological parameters, including clinical stage, lymph node metastasis,

Fig. 1 Immunohistochemical staining of p38β, p38γ, and p38δ in esophageal squamous cell carcinoma (ESCC) tissue microarrays, which are composed of 86 cases of ESCC and 78 cases of paired normal control

tissue. Representative immunostaining figures of each isoform of p38 MAPK are shown. Inset box shows higher magnification view of each isoform of p38 MAPK positive cells

Statistical analysis For correlations between p38β, p38γ, and p38δ, immunostaining scores, and clinicopathological variables, Fisher’s exact or chi-square tests were used; whereas independent sample Student’s t test was used for continuous data. Differences between groups were calculated with the Student’s t test and one-way ANOVA. In addition, Kaplan-Meier survival curves were carried out to analyze the significant difference between patients for prognosis. All the statistical analysis was carried out with SPSS statistics software package (version 17.0; SPCC Inc., Chicago, IL), and statistical figures were made using the Graphpad Prism (5.0 version; GraphPad Software, La Jolla, CA). Values were expressed as means ± standard error of the mean (SEM). All differences were considered statistically significant for p < 0.05.

Results

Gross classification

Tumor volume (cm3)

4

5 6 43 32 22 27 35

N3 Well Moderate Poor <10 10–20 >20

Differentiation degree

N classification

Eminence type

62 3 50 20 9

T3 T4 N0 N1 N2

46 6 24

47 11 5 12

II III T1 T2

Clinical stage

Ulcerative type Fungating type Medullary type

29 28

≤60 I

T classification

57

>60

78 86 64 22

Number

Age (years)

Male Female

Characteristics

4

40 6 20

4 6 39 27 17 27 30

55 3 46 19 6

44 8 5 10

24 24

52

43 76 56 20

High (++, +++)

0

6 0 4

1 0 4 5 5 0 5

7 0 4 1 3

3 3 0 2

5 4

5

35 10 8 2

Low (+)

p38β expression

1.822

6.295

1.553

6.332

1.390

4.071

1.342

0.185

22.704

χ2

0.609

0.044

0.470

0.109

0.692

0.118

0.294

0.730

0.000

p value

Clinicopathological analysis of expression of p38β, p38γ, and p38δ in ESCC tissue microarray

Adjacent normal ESCC Gender

Table 1

3

36 5 14

2 6 34 20 11 23 27

48 3 41 14 3

33 5 5 6

20 25

43

62 63 47 16

High (++, +++)

1

10 1 10

3 0 9 12 11 4 8

14 0 9 6 6

14 6 0 6

9 3

14

16 23 17 6

Low (+)

p38γ expression

3.547

8.165

4.891

11.578

6.958

8.234

0.411

0.004

0.876

χ2

0.315

0.017

0.093

0.008

0.063

0.017

0.609

1.000

0.365

p value

2

11 2 5

0 2 10 9 4 7 11

18 0 13 6 3

10 1 1 2

6 11

16

1 22 16 6

High (++, +++)

2

35 4 19

5 4 33 23 18 20 24

44 3 37 14 6

37 10 4 10

23 17

41

77 64 48 16

Low (+)

p38δ expression

1.807

1.228

0.412

2.163

2.000

4.791

0.550

0.044

20.030

χ2

0.742

0.575

0.791

0.568

0.646

0.089

0.603

1.000

0.000

p value

Tumor Biol.

Tumor Biol.

tumor volume, and differentiation degree as well as demographic parameters in ESCC tissues. By contrast, there was a significant difference of p38γ expression between clinical stage (p = 0.017), lymph nodes metastases (p = 0.008), and tumor volume (p = 0.017) in ESCC tissues. The results we obtained using ESCC tissue microarray suggest that both p38β and p38δ may be oncogenic property-maintaining genes, whereas p38γ may be a metastasis-associating gene in ESCC.

in ESCC (p = 0.364) (Fig. 2c). However, in the subcohort survival analysis, where no lymph node metastasis occurred, there was also a significant difference of prognosis between p38δ higher expression group and p38δ lower expression group (p = 0.047) (Supplementary Figure 1). Besides, in the subcohort where cases were graded as clinical stage II, there was also a significant difference of prognosis between p38δ higher expression group and lower expression group (p = 0.036) (Supplementary Figure 2).

P38β expression is significantly associated with overall prognosis Next, we sought to determine whether or not the expression of p38β, p38γ, and p38δ correlate with overall survival of patients with ESCC. For one thing, we categorized the cohort into the two subcohorts based on the expression level, high expression (including cases with immunostaining of ++ and +++), and low expression (including cases with immunostaining of +). Then, we analyzed the correlations between p38β, p38γ, and p38δ expression and overall survival in ESCC tissues with Kaplan-Meier curve, according to the definition of subcohort on our own. It is found that there was significant difference of prognosis between patients with high-expression level of p38β versus patients with low level of p38β (p = 0.012). To be specific, the higher the expression of p38β was, the better the prognosis will be (Fig. 2a). Furthermore, in the subcohort where no lymph node metastasized, there was also significant difference of prognosis between p38β higher expression group and p38β lower expression group (p = 0.000) (Supplementary Figure 1). No significant difference of prognosis was observed between patients with higher level of p38γ and patients with lower expression of p38γ (p = 0.667) (Fig. 2b). Likewise, no significant difference of prognosis was found between patients with higher level of p38δ versus patients with lower expression of p38δ

Endogenous level of p38α, p38β, p38γ, and p38δ in ESCC cell line Eca109 To determine the endogenous expression level of p38α, p38β, p38γ, and p38δ in human ESCC cell line Eca109, we have measured the basal level of the four isoforms of p38 MAPK using real-time fluorescent quantitative reversal transcription PCR (qRT-PCR) technique (Fig. 3a). We found that, in terms of mRNA expression level, p38α was the most abundant of all, followed by p38γ, then p38β. The level of p38δ was the lowest of the four p38 MAPK isoforms. Theis result was confirmed with semi-quantitative RT-PCR (Fig. 3b), indicating that in Eca109 cells, p38α and p38γ mRNA were more abundant than that of p38β and p38δ. Next, to explore the role of the four isoforms of p38 MAPK, we have knocked down the four p38MAPK isoforms employing siRNA method (Fig. 3b, c). Knockdown efficiency result showed that all the siRNA against p38α, p38β, p38γ, and p38δ were capable of effectively and significantly decreasing the expression of the four isoforms, respectively, compared with non-transfection control (Fig. 3c). Moreover, we also evaluated the silencing specificity of siRNA against each specific isoform. It was shown that siRNA targeting each isoform of p38 MAPK did not affect the quantity of mRNA encoding the other p38 MAPK isoforms on mRNA (Fig. 3d) and protein level (Fig. 3e),

Fig. 2 Prognostic analysis of expression of p38β, p38γ, and p38δ using Kaplan-Meier survival curves. Kaplan-Meier survival curves for overall survival were compared. a Higher expression of p38β (++, +++) versus lower expression of p38β (−, +); p value was 0.012 after log rank statistics. b Higher expression of p38γ (++, +++) versus lower

expression of p38γ (−, +), p value was 0.667 after log rank statistics. c Higher expression of p38δ (++, +++) versus lower expression of p38δ (−, +); p value was 0.364 after log rank statistics. Higher expression or presence of p38β significantly correlated with better overall prognosis but not p38γ and p38δ

Tumor Biol.

Fig. 3 P38 MAPKs silencing efficacy and siRNA specificity in Eca109 cells. a Endogenous mRNA level of p38α, p38β, p38γ, and p38δ in ESCC cell line Eca109 were measured by qRT-PCR. Each p38 isoform was normalized to β-actin (n = 4). Statistical analysis was performed using Friedman chi-square test and pairwise Wilcoxin signed-rank test found that p38α is the most abundant isoform (p < 0.001).b Quantification of p38 MAPK mRNAs from control (non-transfection)

and isoform-specific siRNA transfected Eca109 cell for 48 h after transfection determined using semi-quantitative RT-PCR. c RNA silencing efficiency was evaluated using qRT-PCR.*p < 0.05, **p < 0.01, compared with control group. d RNA silencing specificity was assayed by qRT-PCR. e RNA silencing was specific, as evidenced by western blot on protein level

indicating that the siRNA sequences we referenced from previous study are specific in reducing the expression of different p38 MAPK isoforms.

figure 4). It can be seen that re-expression of p38α can suppress the proliferation; whereas re-expression of p38β, p38γ, and p38δ were capable of promoting the proliferation (Fig. 4b). Overall, the data we obtained showed that p38α was able to prevent the proliferation whereas p38β, p38γ, and p38δ can promote the proliferation of ESCC cell line Eca109 in vitro.

P38α was able to prevent the proliferation, whereas p38β, p38γ, and p38δ can promote, the proliferation of Eca109 in vitro To analyze the effects of p38α, p38β, p38γ, and p38δ over proliferation, MTT assay was carried out. Construct vectors pRNAT-U6.1-eGFP-p38α, -p38β, -p38γ, -p38δ, and scramble control were all transfected into Eca109. Detection of the enhanced green fluorescent protein (eGFP) was used to evaluate the transfection and expression efficiency. GFP showed that all the vectors transfected into Eca109 cells were highly expressed, as monitored by fluorescence inverted microscope (Supplementary figure 3). We observed that inhibition of p38α expression markedly promoted the cell growth; meanwhile, inhibition of p38β, p38γ, and p38δ significantly suppressed the proliferation in ESCC cell line Eca109 (Fig. 4a). To observe whether the phenotypic variation of proliferation caused by knockdown of each isoform of p38 MAPK could be reversed by over-expression of each isoform, eukaryotic expression vectors were transfected into Eca109 cells, which contained full-length cDNA of Flag-p38α, -p38β, -p38γ, and -p38δ. The four vectors were expressed successfully in Eca109, as evidenced by western blot detection with specific anti-Flag antibody (Supplementary

p38α promoted, whereas p38β, p38γ, and p38δ prevented, apoptosis of Eca109 in vitro Having discovered that p38β, p38γ, and p38δ can promote proliferation whereas p38α prevents Eca109 proliferation, we therefore hypothesized that p38β, p38γ, and p38δ could inhibit the apoptosis, while p38α could promote apoptosis in Eca109. To test the hypothesis, we assayed using flow cytometry (FCM) both the apoptosis and cell cycle variation after knockdown of p38β, p38γ, and p38δ in Eca109 cells. Apoptosis assay showed that the apoptosis was increased when all p38 subtypes were knocked down; but the increase in apoptosis after p38α was knocked down was pronouncedly lower than that when the other three p38 isoforms were knocked down. Knockdown of p38β, p38γ, and p38δ can markedly increase apoptosis (Fig. 4c). Re-expression of p38β, p38γ, and p38δ could also lead to apoptosis to a certain extent. However, apoptosis rate was markedly higher after re-expression of p38α compared with re-expression of p38β, p38γ, and p38δ (Fig. 4d). Cell

Tumor Biol.

Fig. 4 Effect of silencing and ectopic expression of p38α, p38β, p38γ, and p38δ on cell growth and apoptosis in Eca109. a Proliferation was measured by MTT assay every 24 h (n = 4) after transfection with shRNA-p38α, -p38β, -p38γ, -p38δ, and shRNA scramble into Eca109 cells. b In parallel, proliferation was measured by MTT assay every 24 h (n = 4) after transfection with pcDNA3.1-p38α/β/γ/δ-Flag and pcDNA3.1-Flag, as control vectors, into Eca109 cells. c Apoptosis analysis of Eca109 cells transfected with siRNA. Since eGFP and

Annexin V has the same waver absorbance, to avoid false positive when subjected to FCM, we used chemically synthesized siRNA whose interference target sequence was totally the same as shRNA’s. d Apoptosis analysis of Eca109 cells transfected with over-expression vector of p38α, p38β, p38γ, and p38δ. e Cell cycle analysis of Eca109 cells transfected with siRNA. f Cell cycle analysis of Eca109 cells transfected over-expression vector of p38α, p38β, p38γ, and p38δ. *p < 0.05, **p < 0.01, in comparison with control group

cycle results showed that the S phase significantly increased with the over-expression of p38β, p38γ, and p38δ (Fig. 4e, f). In all, the results from both apoptosis and cell cycle were congruent with results obtained using MTT method, further suggesting that p38β, p38γ, and p38δ isoform have oncogenic-maintaining roles in the growth of Eca109 cells.

Moreover, similar effects on invasion in Eca109 cells was found that knockdown of both p38β, p38γ, and p38δ led to a marked decrease on the invasion that can be prevented by re-expression, while knockdown of p38α significantly increased the invasion that can be rescued by re-expression, too (Figs. 5c–f). Taken together, these data suggest that p38β, p38γ, and p38δ promote the cell motility of Eca109 in vitro.

p38β, p38γ, and p38δ can promote the cell motility of Eca109 in vitro Based on our in vitro observations on proliferation, we hypothesized that p38β, p38γ, and p38δ could promote the cell motility. To test this hypothesis, we transfected shRNA vectors and eukaryotic expression vectors containing full-length cDNA of p38β, p38γ, and p38δ into Eca109. Then, wound healing assay and Transwell assay were used, respectively, to evaluate the effects on cell migration and invasion. Wound healing assay reveals that knockdown of p38β, p38γ, and p38δ were able to markedly inhibit migration, whereas knockdown of p38α could promote migration (Fig. 5a, c), which can be prevented by re-expression of p38α, p38β, p38γ, and p38δ (Fig. 5b, d), suggesting that p38β, p38γ, and p38δ can promote cell migration in Eca109.

Over-expression of p38δ can promote, whereas knockdown of p38γ can suppress the tumorigenesis in athymic nude mice model xenografted with Eca109 cells Given the scarce report of p38γ and p38δ in the setting of ESCC and having found the oncogenic-maintenance properties of p38γ and p38δ, we determined to verify the role of p38γ and p38δ in tumorigenesis using athymic nude mice model. Firstly, we generated the transgenic Eca109 cells whose endogenous p38γ was stably knocked down and p38δ was stably over-expressed using transfection with lentiviral vectors, on the basis of its basal level in Eca109 cells. To make absolute pure, the two stably transgenic Eca109 cells were subjected to being sorted with FCM. Based on which,

Tumor Biol.

Fig. 5 Effect of silencing and ectopic expression of p38α, p38β, p38γ, and p38δ on cell migration and invasion in Eca109. a Wound healing assay of Eca109 cells transfection with shRNA vector of p38α, p38β, p38γ, and p38δ. b Wound-healing assay of Eca109 cells transfection with over-expression vector of p38α, p38β, p38γ, and p38δ. c Quantitative assay of wound healing of Eca109 cells transfection with shRNA vector of p38α, p38β, p38γ, and p38δ. d Quantitative assay of wound-healing

of Eca109 cells transfection with over-expression vector of p38α, p38β, p38γ, and p38δ. e Transwell assay of Eca109 cells after transfection with shRNA and over-expression vector of p38α, p38β, p38γ, and p38δ. f Quantitative assay of Transwell assay of Eca109 cells transfection with shRNA and over-expression vector of p38α, p38β, p38γ, and p38δ. *p < 0.05, **p < 0.01, compared with control group

the sorted transgenic Eca109 cells as well as control cells were subcutaneously injected into athymic nude mice. After 5 weeks breeding, all the mice involved were euthanatized and tumor lesions were dissected and measured both from weight and volume. It was found that over-expression of p38δ can promote, whereas knockdown of p38γ can suppress, the tumorigenesis in athymic nude mice model xenografted with Eca109 cells, further confirming the oncogenic roles of p38γ and p38δ reflected in vitro in

Eca109 cells. In addition, In vivo clinical tissues, p38γ was found to be significantly correlated with tumor volume, which is congruent with tumorigenic observation in vivo in nude mice. As for p38δ, the data we obtained from nude mice also entirely supports its oncogenic role in ESCC that was observed to significantly upregulate in ESCC tissues compared with paired normal controls. However, there is no significantly different effect on the total mice weight (Supplementary Fig. 5) (Fig. 6).

Fig. 6 Effect of stably silencing p38γ and ectopic expression of p38δ in tumorigenesis of athyhmic nude mice xenografted with Eca109. a FCM sorted Eca109 cells whose basal level of p38γ was stably knocked down using lentiviral recombinant vector produced smaller tumor lesions, whereas FCM sorted Eca109 cells whose basal level of p38δ was stably

over-expressed using lentiviral recombinant vector gave birth to larger tumor lesions in comparison with Eca109 cells transfected with negative control lentiviral vectors. b Quantitation of tumorigenesis in athymic nude mice. NS non significant, OE over-expression, NC negative control, KD knock down

Tumor Biol.

Discussion In the present study, we found that among p38β, p38γ, and p38δ, only p38β expression was significantly associated with overall prognosis, while p38γ expression was significantly associated with lymph node metastasis, clinical stage, and tumor volume in ESCC in vivo; and that both p38β and p38δ were able to promote proliferation and motility of ESCC cell line Eca109 whereas p38γ can significantly promote motility but not proliferation in vitro, suggesting that unlike the oncogenic properties of p38β, p38δ, and p38γ may be a metastasis-associating gene in ESCC cells. The importance of p38 MAPK signal pathway involved in cancer has been increasingly acknowledged [15, 16]; but diverse, sometimes even contradictory roles have been reviewed for p38 MAPK in cancers [1, 17, 18]. P38 MAPK has been reported to induce apoptosis in some cells, but prevent apoptosis in others [19]. P38 MAPK has also been shown to be implicated in cell proliferation and tumorigenesis [20]. Conversely, a number of studies implicated p38 MAPK was a negative regulator of cell proliferation [21, 22]. In agreement with other recent findings [5, 7, 9], we showed that at least some of the discrepancy in p38 function may be attributed to the distinct contributions of p38 isoforms. given that p38 MAPK pathway is ubiquitously utilized in stress-response setting, it logically follows that metastasizing cancer cells, which encounter an ever-changing milieu of cellular stresses, may gain survival advantages under stressed conditions upon modulation of the appropriate p38 MAPK isoforms [19]; Therefore, specifying the contributions of each specific p38 isoform would allow more precise targeting of specific subsets of cancer. In our study, despite that expression of p38α on mRNA level was most abundant compared with p38β, p38γ, and p38δ in ESCC cell line Eca109, p38α has been shown to play a tumor-suppressing role in ESCC in our previous study [3], which was also supported by and consistent with the study conducted in cervical squamous cell carcinoma cell line HeLa cell [12]. However, this is inconsistent, even contradictory with the studies performed in adenocarcinomas, such as breast cancer [19] and colorectal cancer [23], suggesting that the dual role of p38α both as tumor promoter and suppressor in the tumorigenesis of cancers, especially inflammation associated and caused cancers [21, 24]. To date, no conflicting report has been emerged in the setting of squamous cell carcinoma where p38α was tentatively found to play a tumor-suppressing role. p38β has been reported to be a metastasis-associating gene in bone metastasis of breast cancer [11] and to be mediated in bladder cancer cell migration in complex with integrin-linked kinase [25]. In addition, p38β has been proposed to be a therapeutic target in pancreatic cancer [10]. However, the evidence regarding clinicopathological significance in cancers has been scarce. In our study, p38β was observed to be

significantly upregulated in ESCC tissues as compared with paired adjacent normal control tissue. Moreover, there is significant correlation between p38β expression and clinical tumor volume. Specifically, the higher the expression of p38β is, the larger volume of the tumor tends to be. The only two significant clinical associations observed earlier tentatively indicate that p38β seems to be an oncogenic property. Survival analysis showed that there is statistically significant difference of overall prognosis between patients with high level of p38β versus patients with low expression of p38β. That is, the higher the expression of p38β is, the better the prognosis will be. In addition, in the subcohort where no lymph node metastasized, there was also significant difference of prognosis between p38β high-expression group and p38β low-expression group. Likewise, the higher the level of p38β is, the more superior the prognosis will be. Although there was significant difference of prognosis for p38β, the trend reflected by p38β in prognosis is somewhat inconsistent with its expression status in ESCC tissues. In terms of the property suggested by prognosis of p38β, which seems to be somewhat at odds with the evidence that p38β was significantly upregulated in ESCC tissues, as observed. In light of evidence concerning the clinicopathological role of p38β has been rather limited and scarce, both prognosis and clinicopathological significance of p38β expression in our study needs to be further warranted and confirmed in different cohorts of ESCC and other types of cancers. In vitro, we found that p38β was able to promote the proliferation and motility of Eca109 cells, suggesting that p38β could have oncogenic roles, which can entirely account for its expression status on clinical tissue and which are totally in agreement with Yu L et al.’s findings in bladder cancer cell line TSU-Pr1 [25]. Unlike p38α and p38β isoforms on which the majority of studies to date have focused, p38γ and p38δ are comparatively less researched and easily neglected [2, 26] owing to its roles that have been largely unknown and left to be investigated. In the case of involvement of p38γ in the progression of cancers, it has been suggested that p38γ is an oncogenic property-maintaining gene [27], and that p38γ was required in the inflammation-associated colon tumorigenesis [6, 9]. As for whether or not the role of p38γ was associated with metastasis, it is somewhat controversial and inconsistent even in breast cancer [7, 8]. However, there was an agreement supporting that p38γ expression was significantly associated with overall survival of patients with breast cancer [7, 8]. In our study, it was shown that among p38β, p38γ, and p38δ, only expression of p38γ was significantly associated with lymph nodes metastases, clinical stage, and tumor volume in ESCC, which is partially consistent with previous studies finding that p38γ was metastasis-promoting in breast cancer [8]. In terms of prognosis, no significant difference was observed between patients with different p38γ expression status, which is in disagreement with the consensus reached by two

Tumor Biol.

independent studies by Rosenthal DT [8] et al. and Meng F et al. [7] in breast cancer. Besides, no more evidence is available regarding the relationship between p38γ expression and prognosis. In vitro, we found that p38γ was capable of promoting motility but not proliferation in Eca109 cells, which is fully in line with earlier reports [7, 27] finding that p38γ made much difference on metastasis but little difference on proliferation; but our findings on the role of p38γ in vitro is partially inconsistent with publications [8, 28] and reported that p38γ not only can promote metastasis but also can facilitate cell growth in cancer cells. Like p38α, p38δ would also appear to have both pro-and anti-oncogenic roles, depending on the cell type studies [26]. Interest in p38δ as a potential tumor promoter was based on the evidence that p38δ expression and activation are significantly increased in a variety of carcinoma cell lines, such as head and neck squamous carcinoma cells and tumors [29] and cholangiocarcinoma [30]. In contrast to the relatively well-characterized role of p38δ as a tumor promoter, an increasing number of reports outlined its property as tumor suppressor [4, 5, 26]. The indication of a tumor suppressive role for p38δ comes from O’Callaghan C et al.’s two serial studies about p38δ in ESCC [4, 5] where they observed that re-expression of p38δ into ESCC cells lacking endogenous expression markedly impaired cell proliferation, migration, and invasion. In addition, another report in triple-negative breast cancer found that abolition of p38δ expression induced cell growth, whereas over-expression of p38δ reduced growth rate in brain metastases [31]. In our study, we found that p38δ was significantly upregulated in ESCC in comparison with paired normal control tissues, which is suggestive of its oncogenic role in vivo. P38δ expression can promote both proliferation and motility in Eca109 cells, which is in disagreement with O’Callaghan C and et al.’s findings in ESCC cell lines in vitro [5]. No significant correlation was found between p38δ expression and clinicopathological parameters other than significant difference between ESCC and paired normal control, nor is there significant difference in terms of overall prognosis. However, in the subcohorts where no lymph node metastasized and all cases graded as clinical stage II, there is significant difference of prognosis between patients with high expression of p38δ and patients with low p38δ. To be specific, the higher the expression of p38δ is, the better the prognosis tends to be, which is in somewhat conflict with its expression status that was markedly upregulated in ESCC tissues compared with paired normal control. It has been reported that p38δ gene promoter region was discovered to be hypermethylated [32]. This methylation in promoter region leads to downregulation of p38δ mRNA and protein, which may account for the reason why p38δ mRNA was most significantly downregulated both in Eca109 cells and ESCC tissues compared with p38α, p38β, and p38γ.

Despite our study being the first time to comprehensively provide evidences concerning p38β, p38γ, and p38δ in the setting of ESCC, there were still several limitations that have to be acknowledged that we cannot evade. Firstly, the clinical sample size of ESCC tissues we enrolled and the number of ESCC cell line was rather limited [32, 33] and therefore may lead to the potentially biased or insufficient conclusions. Consequently, the observations we presented here may need to be further warranted; secondly, considering that the specificity of primary antibodies used could lead to the irreproducibility of biomedical research [34–36], all the primary antibodies should have been tested and evaluated before being used. However, we failed to test before it being used; thirdly, in our study, immunophenotype was examined using tissue microarray [37], which may not reflect the heterogeneity of protein expression within individual tumors. Another issue arising from our work that will be worthwhile to address in future studies in particular is that, in terms of the prognostic significance of p38β and p38δ, no matter what in cohort or subcohort, was somewhat in conflict with their property suggested by their expression status in ESCC, which remains unknown and which is hard to interpret in the study. Despite the relatively small cases included in our study, our data may have important implications for understanding the clinicopathological significance as well as its biological role of p38β, p38γ, and p38δ in ESCC. In conclusion, among p38β, p38γ, and p38δ, only p38β was found to be significantly associated with overall prognosis. In the subcohort where no lymph node metastasized, both p38β and p38δ were shown to be significantly associated with overall prognosis. Clinicopathologically, p38γ expression was significantly associated with lymph nodes metastases, clinical stage, and tumor volume. In vitro in ESCC cell line Eca109, p38β, p38γ, and p38δ were able to promote proliferation and motility of Eca109, suggesting their oncogenic-maintaining roles. Our study highlights once again the distinct and often opposing functions of the individual p38 MAPK isoform in cancer. Acknowledgments The study was supported by National Science Foundation of China (no. 81360357, 81160303, 81260359, 81201891, U1303321), from Major Science and Technology Projects of the Xinjiang Uygur Autonomous Region (no. 201430123-1) and Opening Project of Xinjiang Medical Animal Model Research Key Laboratory (XJDX11032012-05). We are especially appreciative of Professor Ana Cuenda in the Department of Immunology and Oncology, Centro National de Biotecnología/CSIC, Madrid, Spain, for kindly proof reading and giving constructive comments in the manuscript. Compliance with ethical standards Conflicts of interest None

Tumor Biol.

References 20. 1.

2. 3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14. 15. 16.

17.

18. 19.

Cuenda A, Rousseau S. p38 MAP-kinases pathway regulation, function and role in human diseases. Biochim Biophys Acta. 2007;1773(8):1358–75. Risco A, Cuenda A. New insights into the p38gamma and p38delta MAPK Pathways. J Signal Transduct. 2012;2012:520289. Zheng ST, Liu T, Liu Q, Lu M, Gao XP, Sheyhidin I, et al. Investigation on role of p38alpha mitogen-activated protein kinases in human esophageal squamous cell carcinoma cell line Eca109. Zhonghua yu fang yi xue za zhi Chin J Prev Med. 2013;47(8):757–61. O’Callaghan C, Fanning LJ, Barry OP. p38delta MAPK phenotype: an indicator of chemotherapeutic response in oesophageal squamous cell carcinoma. Anti-Cancer Drugs. 2015;26(1):46–55. O’Callaghan C, Fanning LJ, Houston A, Barry OP. Loss of p38delta mitogen-activated protein kinase expression promotes oesophageal squamous cell carcinoma proliferation, migration and anchorage-independent growth. Int J Oncol. 2013;43(2):405–15. Del Reino P, Alsina-Beauchamp D, Escos A, Cerezo-Guisado MI, Risco A, Aparicio N, et al. Pro-oncogenic role of alternative p38 mitogen-activated protein kinases p38gamma and p38delta, linking inflammation and cancer in colitis-associated colon cancer. Cancer Res. 2014;74(21):6150–60. Meng F, Zhang H, Liu G, Kreike B, Chen W, Sethi S, et al. p38gamma mitogen-activated protein kinase contributes to oncogenic properties maintenance and resistance to poly (ADP-ribose)-polymerase-1 inhibition in breast cancer. Neoplasia. 2011;13(5):472–82. Rosenthal DT, Iyer H, Escudero S, Bao L, Wu Z, Ventura AC, et al. p38gamma promotes breast cancer cell motility and metastasis through regulation of RhoC GTPase, cytoskeletal architecture, and a novel leading edge behavior. Cancer Res. 2011;71(20):6338–49. Yin N, Qi X, Tsai S, Lu Y, Basir Z, Oshima K, Thomas JP, Myers CR, Stoner G, Chen G. p38gamma MAPK is required for inflammation-associated colon tumorigenesis. Oncogene 2015. Singh AK, Pandey R, Gill K, Singh R, Saraya A, Chauhan SS, et al. p38beta MAP kinase as a therapeutic target for pancreatic cancer. Chem Biol Drug Des. 2012;80(2):266–73. He Z, He J, Liu Z, Xu J, Yi SF, Liu H, et al. MAPK11 in breast cancer cells enhances osteoclastogenesis and bone resorption. Biochimie. 2014;106:24–32. Kukkonen-Macchi A, Sicora O, Kaczynska K, Oetken-Lindholm C, Pouwels J, Laine L, et al. Loss of p38gamma MAPK induces pleiotropic mitotic defects and massive cell death. J Cell Sci. 2011;124(Pt 2):216–27. Hewitt SM, Baskin DG, Frevert CW, Stahl WL, Rosa-Molinar E. Controls for immunohistochemistry: the Histochemical Society’s standards of practice for validation of immunohistochemical assays. J Histochem Cytochem Off J Histochem Soc. 2014;62(10):693–7. Burry RW. Controls for immunocytochemistry: an update. J Histochem Cytochem Off J Histochem Soc. 2011;59(1):6–12. Liu H, He J, Yang J. Tumor cell p38 MAPK: a trigger of cancer bone osteolysis. Cancer Cell Microenviron. 2015;2(1). Gonzalez-Villasana V, Fuentes-Mattei E, Ivan C, Dalton HJ, Rodriguez-Aguayo C, Fernandez-de Thomas RJ, et al. Rac1/Pak1/p38/MMP-2 axis regulates angiogenesis in ovarian cancer. Clin Cancer Res Off J Am Assoc Cancer Res. 2015;21(9): 2127–37. Rennefahrt U, Janakiraman M, Ollinger R, Troppmair J. Stress kinase signaling in cancer: fact or fiction? Cancer Lett. 2005;217(1):1–9. Loesch M, Chen G. The p38 MAPK stress pathway as a tumor suppressor or more? Front Biosci J Virtual Libr. 2008;13:3581–93. Chen L, Mayer JA, Krisko TI, Speers CW, Wang T, Hilsenbeck SG, et al. Inhibition of the p38 kinase suppresses the proliferation of

21.

22.

23.

24.

25.

26. 27.

28.

29.

30.

31.

32.

33.

34. 35.

36. 37.

human ER-negative breast cancer cells. Cancer Res. 2009;69(23): 8853–61. Radziwon-Balicka A, Santos-Martinez MJ, Corbalan JJ, O’Sullivan S, Treumann A, Gilmer JF, et al. Mechanisms of platelet-stimulated colon cancer invasion: role of clusterin and thrombospondin 1 in regulation of the P38MAPK-MMP-9 pathway. Carcinogenesis. 2014;35(2):324–32. Grossi V, Peserico A, Tezil T, Simone C. p38alpha MAPK pathway: a key factor in colorectal cancer therapy and chemoresistance. World J Gastroenterol WJG. 2014;20(29):9744–58. Tai TW, Su FC, Chen CY, Jou IM, Lin CF. Activation of p38 MAPK-regulated Bcl-xL signaling increases survival against zoledronic acid-induced apoptosis in osteoclast precursors. Bone. 2014;67:166–74. Germani A, Matrone A, Grossi V, Peserico A, Sanese P, Liuzzi M, et al. Targeted therapy against chemoresistant colorectal cancers: inhibition of p38alpha modulates the effect of cisplatin in vitro and in vivo through the tumor suppressor FoxO3A. Cancer Lett. 2014;344(1):110–8. Gupta J, del Barco BI, Igea A, Sakellariou S, Pateras IS, Gorgoulis VG, et al. Dual function of p38alpha MAPK in colon cancer: suppression of colitis-associated tumor initiation but requirement for cancer cell survival. Cancer Cell. 2014;25(4):484–500. Yu L, Yuan X, Wang D, Barakat B, Williams ED, Hannigan GE. Selective regulation of p38beta protein and signaling by integrinlinked kinase mediates bladder cancer cell migration. Oncogene. 2014;33(6):690–701. O’Callaghan C, Fanning LJ, Barry OP. p38delta MAPK: emerging roles of a neglected isoform. Int J Cell Biol. 2014;2014:272689. Meng F, Wu G. Is p38gamma MAPK a metastasis-promoting gene or an oncogenic property-maintaining gene? Cell Cycle. 2013;12(14):2329–30. Yang K, Liu Y, Liu Z, Liu J, Liu X, Chen X, et al. p38gamma overexpression in gliomas and its role in proliferation and apoptosis. Sci Rep. 2013;3:2089. Junttila MR, Ala-Aho R, Jokilehto T, Peltonen J, Kallajoki M, Grenman R, et al. p38alpha and p38delta mitogen-activated protein kinase isoforms regulate invasion and growth of head and neck squamous carcinoma cells. Oncogene. 2007;26(36):5267–79. Tan FL, Ooi A, Huang D, Wong JC, Qian CN, Chao C, et al. p38delta/MAPK13 as a diagnostic marker for cholangiocarcinoma and its involvement in cell motility and invasion. Int J Cancer J Int Cancer. 2010;126(10):2353–61. Choi YK, Woo SM, Cho SG, Moon HE, Yun YJ, Kim JW, et al. Brain-metastatic triple-negative breast cancer cells regain growth ability by altering gene expression patterns. Cancer Genomics Proteomics. 2013;10(6):265–75. Kaplan RM, Chambers DA, Glasgow RE. Big data and large sample size: a cautionary note on the potential for bias. Clin Transl Sci. 2014;7(4):342–6. Meier DT, Entrup L, Templin AT, Hogan MF, Samarasekera T, Zraika S, et al. Determination of optimal sample size for quantification of beta-cell area, amyloid area and beta-cell apoptosis in isolated islets. J Histochem Cytochem Off J Histochem Soc. 2015;63(8):663–73. Baker M. Reproducibility crisis: blame it on the antibodies. Nature. 2015;521(7552):274–6. Helsby MA, Fenn JR, Chalmers AD. Reporting research antibody use: how to increase experimental reproducibility. F1000Res. 2013;2:153. Helsby MA, Leung MY, Chalmers AD. The F1000 research antibody validation article collection. F1000Res. 2014;3:241. Linderoth J, Ehinger M, Akerman M, Cavallin-Stahl E, Enblad G, Erlanson M, et al. Tissue microarray is inappropriate for analysis of BCL6 expression in diffuse large B-cell lymphoma. Eur J Haematol. 2007;79(2):146–9.