Decrease of 5-hydroxymethylcytosine and TET1 with nuclear exclusion of TET2 in small intestinal neuroendocrine tumors

All 40 patients included in this study (Additional file 1: Table S1) were diagnosed with SI-NET and operated on at Uppsala University Hospital between 2000 and 2016. The mean age at diagnosis was 62.5 years (range 42–79 years). Specimens were stained with hematoxylin-eosin and those with apparent tumor mass were selected for analyses. In total 87 tumors were analyzed; 40 primary tumors, 37 mesometastases, 7 liver metastases, and 3 lymph node metastases, as well as 4 “normal” small intestine tissue specimens. In addition to being embedded in paraffin, all tissue specimens were snap frozen in liquid nitrogen and kept at − 70 °C. Two human SI-NET cell lines were used in the experiments. CNDT2.5 adhesive cells developed from a liver metastases from a patient diagnosed with primary ileal SI-NET, were kindly provided by Dr. Lee Ellis, MD, Anderson Cancer Center, Houston, TX, USA, and used in this study at cell passages 10–30 [22, 23]. KRJ-I suspension cells established from a multifocal metastatic ileal carcinoid tumor were a kind gift from Dr. Roswitha Pfragner, Medical University of Graz, Austria [24‐26]. Both cell lines expressed the neuroendocrine cell marker synaptophysin (Additional file 2: Figure S1a). In contrast to a recently published study suggesting lymphoblastoid origin of KRJ-I cells [27], the KRJ-I cells used here stained negatively for the lymphoid marker CD45 and the T-cell marker CD3 (Additional file 2: Figure S1b).

Genomic DNA was extracted from 87 frozen primary tumors and the matched metastases from 40 patients using DNeasy Blood and tissue kit (Qiagen GmbH) according to the manufacturer’s instructions. 5hmC DNA standard (Zymo Research Corporation) was used as a control. One microgram DNA was denatured in 0.1 M NaOH at 95 °C for 10 min, then neutralized with 1 M ammonium acetate on ice. Twofold serial dilutions of the DNA samples were spotted onto Hybond-N+ nylon membrane (GE Healthcare) in a Bio-Dot apparatus (Bio-Rad Laboratories, Inc.), which was fixed with UV irradiation (GS Gene Linker UV chamber, Bio-Rad). Subsequently, the membrane was blocked with 5% skimmed milk, and incubated at 4 °C overnight with a rabbit polyclonal anti-5hmC antibody (1:10 K dilution, 39,791; Active Motif). After incubation with an appropriate HRP-conjugated secondary antibody, signals were visualized with the enhanced chemiluminescence system (GE Healthcare). To ensure equal amount of the total DNA on the membrane, the same membrane was stained with 0.02% methylene blue in 0.3 M sodium acetate. The second dot blot signal from the top for each serial dilution was used to quantify the 5hmC relative intensity by the NIH Image-J software according to the program’s instructions using a procedure described previously [18].

Paraffin embedded tumor tissue sections were deparaffinized with xylene and rehydrated through descending alcohol concentrations and distilled water. Background staining was blocked with 3% hydrogen peroxide and heated in EDTA pH 8.0 (Life Technologies Corporation), for 40 min with microwave at 300-W power. Then the sections were incubated with 2 M HCl for 2.5 min and treated with normal goat serum and the rabbit polyclonal anti-5hmC antibody (1:6000 dilution, Active Motif), or the rabbit polyclonal anti-TET1 (HPA019032; Prestige Antibodies, Sigma-Aldrich). For TET2 staining, the sections were heated in citrate buffer pH 6.0 for 10 min with microwave at 800-W power and then for 20 min at 450-W power. After 20 min rest, the sections were incubated with normal goat serum and the rabbit polyclonal anti-TET2 (21207–1-AP, Proteintech Group). For XPO1 staining, the sections were heated in Tris-EDTA buffer pH 9.0 (Bethyl IHC-101 J) and then incubated with normal goat serum and the rabbit monoclonal anti-XPO1 (ab191081, Abcam). The sections were washed three times with PBS, then incubated with a proper secondary antibody and ABC complex. DAB was used for visualization. Consecutive tissue sections of intestinal mucosa were stained with the anti-5hmC antibody (Active Motif), anti-chromogranin A antibody (LK2H10, Thermo Fisher Scientific), anti-TET1 (HPA019032; Prestige Antibodies, Sigma-Aldrich), or anti-TET2 (21207–1-AP, Proteintech). CNDT2.5 and KRJ-I cells first were fixed in formalin and incubated for 20 min with ice-cold 70% ethanol. Then the slides were stained for TET2 as described above. Formalin-fixed CNDT2.5 and KRJ-I cells were first heated in citrate buffer pH 6.0, then incubated with normal goat serum and the rabbit monoclonal anti-synaptophysin (ab32127, Abcam). KRJ-I cells were also stained with the mouse monoclonal anti-CD45 (sc-1178, Santa Cruz Biotechnology), and the rabbit monoclonal anti-CD3 (ab16669, Abcam).

CNDT2.5 cells were distributed onto six-well plates (2 × 10⁵) in DMEM-F12 complemented with 10% fetal bovine serum (Sigma Aldrich), 1% nonessential amino acids, and 1% penicillin-streptomycin (PEST), and were cultured at 37 °C in 5% CO2. After overnight incubation, the cells were transfected in triplicate with 4 μg TET1 plasmid expression vector (InvivoGen) or empty vector (pUNO1) using 8 μl Lipofectamin 2000 transfection reagent (Life Technologies), according to the manufacturer’s instructions. Six hours after transfection, a fresh medium was added complemented with 8 μg/ml blasticidin (InvivoGen). KRJ-I cells were seeded onto twelve-well plates (1 × 10⁵) prior to transfection in DMEM-F12 complemented with 10% fetal bovine serum (Sigma Aldrich), 1% nonessential amino acids, and 1% penicillin-streptomycin. The cells were transfected with 0.4 μg TET1 plasmid or empty vector using 1.5 μl Attractene transfection reagent (Qiagen) according to the manufacturer’s instructions. Successful transfection were monitored by RT-PCR and western blotting analysis after 72 h.

CNDT2.5 (2 × 10⁵) and KRJ-I (1 × 10⁵) cells were treated with 5-aza-2′-deoxycytidine (10 μM and 5 μM, respectively) for 72 h. Growth medium with additions were changed every 24 h. Both cell lines were treated with several concentrations with leptomycin B (sc-202,210, Santa Cruz Biotechnology), the nuclear export inhibitor, or KPT-330/selinexor (Selleckchem), an orally bioavailable selective inhibitor of nuclear export for 24 and 72 h, respectively. Two biological replicates of the experiments were performed.

CNDT2.5 cells were seeded onto six-well plates (2 × 10⁵) and transfected in triplicates with TET1 plasmid expression vector or empty vector as described above. Twenty-four hours after transfection, 2000 cells were distributed onto six-well plates, a fresh medium with 8 μg/ml blasticidin was added every 72 h. After 10 days in blasticidin selection, the cells were fixed with 10% acetic acid/10% methanol and stained with 0.4% crystal violet, and the visible colonies were counted. Three biological replicates of the experiment were performed.

To assess cell proliferation, the CyQUANT cell proliferation assay kit (Invitrogen, Thermo Fisher Scientific) was used according to manufacturer’s instructions. The cells were first frozen in the microplate and then lysed and stained with CyQuant GR dye solution. Infinite 200 PRO (TECAN) plate reader was used to measure fluorescence intensity at 480/520 nm. Apoptosis was measured using the Cell Death Detection ELISA kit (Roche Molecular Biochemicals), according to manufacturer’s protocol. As a positive control, CNDT2.5 and KRJ-I cells were incubated with 0.1 μg/ml camptothecin (Sigma-Aldrich) for 48 h.

TET1 promoter methylation status was investigated for 12 CpG sites in 20 primary tumors and metastases, CNDT2.5 and KRJ-I cells. DNA was treated with bisulfite using the EpiTect bisulfite kit (Qiagen) according to the manufacturer’s instructions, and PCR amplified using MyTaq™ HS Mix (Bioline). Touchdown PCR was performed with initial denaturation at 95 °C for 2 min, followed by 10 cycles of 95 °C, 15 s; annealing temperature step downs every cycle of 0.5 °C (from 61 °C to 57 °C); 72 °C, 20 s; the annealing temperature for the final 40 cycles was 57 °C followed by denaturation and extension phases as above. Forward primer was used: 5′-GGGGTTAGGGTTTTGATTGTGTTG-3′, and reverse primer: 5′-biotin-ACCCCCAACTCCAAACCTA-3′. Sequencing primer: 5′-GGTTTTGATTGTGTTGGGAG-3′, and sequenced analyzed: 5′- TYGGTAGGYGTTTTTYGYGATTYGTTYGYGTTTTTYGYGTTYGTYG.

GGGTTTYGGGTTTTAAAGTTGTGGGGAT-3′. The pyrosequencing was carried out using 10 μl PCR product with the PyroMark Q24 system (Qiagen) according to manufacturer’s instructions.

Cytobuster Protein Extract Reagent (Merck Millipore, Billerica, MA, USA) complemented with protease inhibitor cocktail (Roche Diagnostics Scandinavia AB, Bromma, Sweden) used to prepare protein extracts. Rabbit polyclonal anti-TET1 (GTX124207, GeneTex Inc) and goat polyclonal anti-Actin (sc-1616, Santa Cruz Biotechnology) were used. Separate cytoplasmic and nuclear protein extracts were prepared using NE-PER Nuclear and Cytoplasmic Extraction Reagents kit (Thermo Scientific) with Halt Protease Inhibitor Single-Use Cocktail EDTA-free (Thermo Scientific) according to manufacturer’s instructions. Rabbit polyclonal anti-TET2 (21207–1-AP, Proteintech Group), mouse monoclonal anti-Lamin A/C (sc-376,248, Santa Cruz Biotechnology), and rabbit polyclonal anti-β tubulin (sc-9104, Santa Cruz Biotechnology) were used. After incubation with the proper secondary antibody, except for anti-Lamin A/C-HRP conjugated, bands were visualized using the enhanced chemiluminescence system (GE Healthcare).

DNA-free total RNA was extracted using RNeasy Plus Mini kit (Qiagen) according to manufacturer’s instructions and treated with DNase I using TURBO DNA-free™ kit (Life Technologies Corporation, Thermo Fisher Scientific), and PCR used to establish successful treatment of all RNA preparations. cDNA synthesis was performed with random hexamer primers using the First-Strand cDNA Synthesis kit (Fermentas, Thermo Fisher Scientific) according to the manufacturer’s instructions. qRT-PCR was performed on StepOnePlus RealTime PCR systems (Life Technologies Corporation) using TaqMan gene expression Master Mix and assays for TET1 (Hs00286756_m1), TET2 (Hs00325999_m1), XPO1 (Hs00185645_m1), 18S rRNA (Hs03928990_g1), and GAPDH (Hs02758991_g1) transcripts. Each cDNA sample was analyzed in triplicates.

Paired and unpaired t test were used for statistical analysis, and p < 0.05 was considered significant. All data are presented as mean ± S.D.

The relative level of 5hmC in 87 primary tumors (PT) and corresponding metastases (Met) from 40 patients was determined by a semi-quantitative DNA immune-dot blot assay using a specific rabbit polyclonal antibody against 5hmC [18]. A commercial 5hmC DNA ELISA kit was also applied, but showed too low sensitivity (not shown). The obtained data separated the 40 patients into 3 groups (Fig. 1a - c). Seventeen patients showed a relative low level of 5hmC in the Mets compared to the PTs, 15 patients showed no difference, and 8 patients showed a relative high level of 5hmC in the Mets compared to the PTs. Statistical analysis did not reveal any relation to clinical data (Additional file 1: Table S1) for the 3 patient groups (not shown).

Next, immunohistochemical (IHC) analysis of 5hmC was performed in 32 paired PTs and Mets from 15 patients, represented in the 3 identified patient groups above (Fig. 1). All analyzed SI-NETs showed a mosaic pattern of staining, i.e. a mixture of positive and negative cell nuclei, regardless of cell number, staining strength or patient group (Fig. 2a, b). Some tumors (9/32) also displayed mosaic pattern together with negative 5hmC staining in larger areas of the section (Fig. 2a). 11 out of 15 patients displayed similar pattern of staining for both primary tumors and metastases and 4 patients displayed mosaic pattern together with negative 5hmC staining in larger areas of the section only in metastases. No staining was observed in the absence of the primary antibody (Fig. 2c), and chromogranin A-positive cells in the normal small intestine stained positive for 5hmC (Additional file 3: Figure S2). These cells likely represent the enterochromaffin cell of origin of SI-NETs.

IHC staining of TET1 and TET2 in the 32 paired PTs and Mets from 15 patients closely resembled those observed for 5hmC, with cell nuclear staining and mosaic pattern, i.e. a mixture of positive and negative cell nuclei, regardless of cell number, staining strength or patient group (Fig. 2a-c). Similar to the observed aberrant staining patterns of 5hmC, tumors displayed mosaic regions together with areas of negative staining (15 out of 32 tumors for TET1, and 2 out of 32 tumors for TET2). For one additional tumor almost all cell nuclei stained positive for TET1. Prominent overall staining of TET1 and TET2 was observed in normal small intestine (Additional file 3: Figure S2).

TET2 also showed cytoplasmic staining in addition to the mosaic nuclear staining pattern (Fig. 2a - c). Fifteen out of the analyzed 32 SI-NETs showed cells with TET2 staining of the cytoplasm and of the nucleus. Seventeen tumors showed the presence of cells with detectable TET2 staining of the cytoplasm and not in the nucleus (Fig. 2c). Different patterns of staining in primary tumors compared to paired metastases were observed in 7 patients out of 15 for TET1, and in 2 patients for TET2. The observed staining patterns of TET1 resembled those of 5hmC in 19 tumors out of 32 and for 3 tumors staining of consecutive tissue sections revealed the same negative areas for both 5hmC and TET1. A correlation between the TET1 and TET2 stainings and the Ki67 index (Additional file 1: Table S1) was not observed. No significant difference in mRNA expression level of TET1 and TET2 was observed when PTs and Mets were compared (Fig. 2d).

Treatment of SI-NET cell lines CNDT2.5 (adherent cells) and KRJ-I (suspension cells) with the DNA methylation inhibitor 5-aza-2′-deoxycytidine caused clear induction of TET1 mRNA expression in the CNDT2.5 cells. No effects on TET2 expression was observed (Additional file 4: Figure S3a). The TET1 promoter and exon 1 CpG island has been reported to be methylated in multiple cancers [28]. Therefore, DNA methylation levels of 12 CpG residues in this region was determined for the 2 cell lines by quantitative bisulfite pyrosequencing analysis. CNDT2.5 cells showed high level of methylation (90%), while KRJ-I cells were methylated at low level (20%) (Additional file 4: Figure S3b). This was consistent with the observed activation of TET1 expression in Aza-treated CNDT2.5 cells. Next, methylation levels were determined in 20 PTs and Mets, and only very low levels of methylation (10%) were observed (Additional file 4: Figure S3b). Thus, it is not likely that the observed reduced expression of TET1 in SI-NETs involves DNA hypermethylation of the promoter region.

In order to investigate whether TET1 could play a cell growth regulatory role in SI-NET cells, a colony forming assay was used. Stable overexpression of TET1 for 10 days resulted in a significant reduced number of CNDT2.5 cell colonies, compared to transfection of empty expression vector (Fig. 3a). Transient overexpression of TET1 for 72 h in CNDT2.5 and KRJ-I cells resulted in induction of apoptosis, as determined by quantifying cytoplasmic histone-associated-DNA-fragments (Fig. 3b). Successful increased expression of TET1 after transfection was monitored by real-time quantitative PCR and Western blotting analysis (Fig. 3c). Neither transfection with TET2 CRISPR double nickase plasmids [29] nor TET2 shRNAs resulted in efficient knockout of TET2 expression (data not shown).

Taken together, these results strongly suggest a cell growth regulatory role of TET1 in neuroendocrine cells of the small intestine and function as candidate tumor suppressor gene.

In addition to the mosaic nuclear staining pattern, the IHC analysis described above displayed variable cytoplasmic expression of TET2 in all analyzed PTs and Mets (n = 32), and presence of cytoplasmic positive cells with absent TET2 nuclear localization in 17 of these tumors. To investigate whether the nuclear exclusion of TET2 in SI-NETs involved the nuclear export machinery and the nuclear export protein exportin-1 (XPO1/CRM1), CNDT2.5 and KRJ-I cells were treated with the nuclear export inhibitor leptomycin B for 24 h, and analyzed by IHC and Western blotting. Figure 4a shows accumulation of TET2 in the nuclei of leptomycin B treated cells from both SI-NET cell lines. Partitioning of cellular protein extracts into cytoplasmic and nuclear fractions followed by Western blotting analysis showed induced absence in the cytoplasm and nuclear retention of the slowest migrating polypeptide of TET2 after leptomycin B treatment (Fig. 4b). These results together strongly suggested aberrant transport of TET2 from the nucleus to the cytoplasm by the exportin-1 nuclear export machinery. Overexpression of XPO1 is common in other tumor types and real-time quantitative RT-PCR analysis revealed highly variable mRNA expression levels (Fig. 4c), suggesting a possibility of XPO1 overactivity also in SI-NETs. IHC staining of XPO1 in the 31 PTs and Mets from 15 patients showed overall positive staining regardless of strength in 13 tumors, and mosaic pattern of staining or positive staining together with negative areas was observed in 18 tumors (data not shown). Of the 13 tumors that stained overall positively for XPO1, 12 tumors showed detectable TET2 staining in the cytoplasm and not in the nucleus. Furthermore, treatment with leptomycin B or KPT-330 (selinexor), a selective oral reversible inhibitor of XPO1 and nuclear export [30], resulted in reduced cell proliferation and induction of apoptosis of CNDT2.5 and KRJ-I cells (Fig. 5a, b).