This article is mentioned in:

Introduction

Hepatocellular carcinoma (HCC) is the fifth most common cancer and third leading cause of cancer-related deaths worldwide and its prognosis is poor despite advances in diagnostic and therapeutic modalities (1,2). A minority of patients with HCC are diagnosed during early stages when the disease is amenable to potentially curative treatments such as surgery and topical therapy (for example, radiofrequency ablation) (3). The prognosis is poor for patients diagnosed at an advanced stage or with recurrent lesions after surgery or topical therapies, owing to underlying liver damage and lack of effective treatment options (1,4). Therefore, investigations that clarify the mechanisms of carcinogenesis and tumor progression in HCC are urgently needed to discover targets for therapy and prognostic markers for recurrence and progression of HCC.

Evidence indicates that the pathogenesis of HCC occurs in a stepwise process driven by oncogene activation and inactivation of tumor suppressor genes (TSGs) (5,6). Epigenetic alterations such as promoter hypermethylation can lead to the transcriptional silencing of TSGs (7,8). We identified several tumor-related genes frequently silenced through promoter hypermethylation in HCC, suggesting that expression status and promoter hypermethylation of these genes can serve as a biomarker for early detection of HCC (9–11).

The differentially expressed in normal and neoplastic cells (DENN) domain proteins regulate Rab GTPases and represent a newly recognized class of membrane trafficking proteins (12–16). The Rab family comprises 70 members in humans and represents the largest family of small GTPases (17). Rab proteins cycle between inactive GDP-bound and active GTP-bound states. In the active state, they recruit effectors that control multiple aspects of membrane trafficking (13,18,19). The DENN domain present in members of the connecdenn family of proteins, interacts directly with Rab35 and functions as a guanine nucleotide exchange factor (GEF) for this GTPase (20). GEFs activate Rabs by mediating the exchange of GDP for GTP. The human genome encodes eight DENND (DENN domain) proteins that form eight families based on homology and domain structure as follows: DENND1A-1C, DENND2A-2D, DENND3, DENND4A-4C, DENND5A/5B, DENND6A/6B, MTMR5/13 and DENN/MADD. The DENN domain is located toward the N-terminus, except for the DENND2 family, where it is located toward the C-terminus (21–24). There is no significant sequence similarity to other proteins outside of the DENN domain.

Little is known about the functions and expression patterns of DENND family proteins in malignant tumors, although they play important roles in intracellular signaling pathways by integrating the activity of Rab pathways. DENND2D, which is located on chromosome 1p13.3, encodes a 53-kDa protein, which suppresses the proliferation and tumorigenicity of non-small cell lung cancer cells (23,25); however, the role of DENND family proteins in gastroenterological cancers has not been reported. Accordingly, we focused on DENND2D and investigated the regulation of its expression in an attempt to identify a TSG regulated by silencing through promoter hypermethylation. Our present study also focused on developing novel epigenetic biomarkers for HCC.

Materials and methods

Sample collection

Nine HCC cell lines (Hep3B, HepG2, HLE, HLF, HuH1, HuH2, HuH7, PLC/PRF/5 HepG2 and SK-Hep1) were obtained from the American Type Culture Collection (Manassas, VA, USA) and cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum at 37°C in an atmosphere containing 5% CO2. Primary HCC tissues and corresponding non-cancer tissues were collected from 92 patients (75 men, 17 women; mean age, 63.2±9.9 years; range, 34–84 years) who underwent liver resection for HCC at Nagoya University Hospital between January, 1998 and July, 2008. Mean duration of patient follow-up was 48.3±36.9 months (range, 0.8–147 months). Specimens were classified histologically using the 7th edition of the Union for International Cancer Control (UICC) classification (26). Written informed consent for surgery and use of clinical data, as required by the institutional review board, was obtained from all patients. Tissue samples were immediately flash frozen in liquid nitrogen and stored at −80°C. Microscopic evaluations were performed to ensure that tumor samples contained >80% tumor cells, whereas non-cancer liver tissue samples did not contain regenerative or dysplastic nodules.

Reverse transcription-polymerase chain reaction (RT-PCR) and quantitative real-time RT-PCR (qPCR)

The levels of DENND2D mRNA expression were analyzed using RT-PCR and qRT-PCR. Total RNAs (10 μg) isolated from each of the HCC cell lines listed above, 92 primary HCC tissues and corresponding non-cancer tissues were used as templates to generate complementary DNAs (cDNAs). PCR primers for DENND2D were as follows: sense (S) (5′-CACTGCTCTACCCCTTCAGC-3′ in exon 7) and anti-sense (AS) (5′-TTTTTCATCACCAACCGACA-3′ in exons 9 and 10), which amplify a 204-base-pair (bp) product. RT-PCR amplification was performed as follows: 40 cycles at 94°C for 30 sec, 60°C for 30 sec and 72°C for 30 sec after an initial denaturation step at 94°C for 5 min. To confirm that equal amounts of cDNA were used as templates, RT-PCR of β-actin was performed. Each RT-PCR product was loaded directly onto 2% agarose gels, electrophoresed, stained with ethidium bromide and visualized with ultraviolet light. The qPCR reactions were performed using a SYBR® Green PCR Core Reagents kit (Life Technologies, Carlsbad, CA, USA) under the following conditions: 1 cycle at 50°C for 2 min, 1 cycle at 95°C for 10 min and 45 cycles at 95°C for 15 sec and 60°C for 30 sec. Real-time detection of the SYBR-Green emission was conducted with an ABI PRISM® 7000 Sequence Detection System (Life Technologies). The primers were those described above. For standardization, GAPDH (TaqMan®, GAPDH Control Reagents, Life Technologies) was amplified in each sample. Nine HCC cell lines and 92 pairs of clinical samples and negative control reactions without templates were analyzed. Reactions were performed in triplicate. The amount of amplified DENND2D DNA in each sample was normalized to that of GAPDH. The expression of DENND2D mRNA was defined as downregulated in tumor tissues when its level was less than one-third that of the corresponding non-cancer tissues.

Analysis of the promoter region of DENND2D

The nucleotide sequence of the DENND2D promoter region was analyzed to determine the presence or absence of CpG islands defined as: ≥200-bp region of DNA with a high GC content (>50%) and an Observed CpG/Expected CpG ratio ≥0.6 (27). We used CpG Island Searcher software (http://cpgislands.usc.edu/) to determine the locations of CpG islands (28).

Methylation-specific PCR (MSP)

DNA samples from HCC cell lines, HCC tissues and corresponding non-cancer tissues were treated with bisulfite. Briefly, 2 μg of DNA was denatured with NaOH, reacted with sodium bisulfite and purified using the Wizard® PCR Preps DNA Purification System resin (Promega, Madison, WI, USA), treated again with NaOH, precipitated with ethanol and resuspended in water. The sequences of the unmethylated primer pairs that amplify a 102-bp product were derived from the DENND2D promoter region upstream of exon 1 are: S (5′-GATATGTGTTTTTGTGGATT-3′) and AS (5′-ACACATCCAAAACTAAAC-3′). Primer sequences derived from the DENND2D promoter region used to detect methylated DNA amplify a 193-bp product and are: S (5′-AGGTGGCGTCGTTTAGTTTC-3′) and AS (5′-GCGAATCCGACACTTTCACT-3′). DNA was amplified as follows: 45 cycles of 94°C for 30 sec 58°C for 30 sec and 72°C for 30 sec after an initial denaturation step at 94°C for 5 min. Each PCR product was loaded directly on 2% agarose gels, electrophoresed, stained with ethidium bromide and visualized with ultraviolet light.

Bisulfite sequence analysis

Genomic bisulfite-treated DNAs from HCC cell lines were sequenced to verify the MSP results. The sequences of the primer pair used to generate a fragment for sequencing was derived from the DENND2D promoter region: S (5′-GGAGGTTAAGGATAGGGG-3′) and AS (5′-ACACTAACCCCCATAACC-3′), which amplify a 133-bp product. DNA was amplified as follows: 50 cycles at 94°C for 30 sec, 63°C for 30 sec and 72°C for 30 sec following an initial denaturation step at 94°C for 5 min. PCR products were purified directly using the QIAquick PCR Purification kit (Qiagen, Hilden, Germany). Purified DNA fragments were subcloned into the TA cloning vector (Life Technologies). Each DNA sample was mixed with 3 μl of specific primer (M13) and 4 μl of Cycle Sequence Mix (BigDye® Terminator v1. 1 Cycle Sequencing kit, Life Technologies, Grand Island, NY, USA). Sequences were analyzed using an Applied Biosystems ABI PRISM 310 DNA Analyzer and sequence electropherograms were generated using ABI Sequence Analysis 3.0 software (Life Technologies).

5-Aza-2′-deoxycytidine (5-aza-dC) treatment

To assess the relation of promoter hypermethylation to DENND2D expression, HCC cells (1.5×106) were treated with 5-aza-dC (Sigma-Aldrich, St. Louis, MO, USA) to inhibit DNA methylation and were cultured for 6 days with medium changes on days 1, 3 and 5. RNA was extracted and RT-PCR was performed as described above.

Immunohistochemistry (IHC)

We used IHC to investigate DENND2D localization in 30 representative sections of well preserved HCC tissues. Formalin-fixed, paraffin-embedded tissues were treated with 3% H2O2 to inhibit endogenous peroxidase, followed by epitope retrieval using five incubations in 10 mM citrate buffer at 95°C, 5 min each. The samples were incubated with Histofine SAB-PO(R) (Nichirei, Tokyo, Japan) for 5 min to limit non-specific reactivity and were then incubated for 1 h at room temperature with a rabbit antibody against DENND2D (PA5-24032, Thermo Fisher Scientific Inc., Rockford, IL, USA) diluted 1:100 in Antibody Diluent (Dako). Sections were developed for 2 min using liquid 3, 3′-diaminobenzidine as the substrate (Nichirei). Staining properties were determined using surrounding hepatic veins as internal controls and staining patterns were compared between HCCs and the corresponding non-cancer tissues. To avoid subjectivity, specimens were randomized and coded before analysis by two independent observers blinded to the status of the samples. Each observer evaluated all specimens at least twice within a given time interval to minimize intraobserver variation.

Statistical analysis

The relative mRNA expression levels (DENND2D/GAPDH) between HCC and non-cancer tissues were analyzed using the Mann-Whitney U test. The χ2 test was used to analyze the association between the expression and methylation status of DENND2D and clinicopathological parameters. Overall and disease-free survival rates were calculated using the Kaplan-Meier method and the difference in survival curves was analyzed using the log-rank test. We performed multivariable regression analysis to detect prognostic factors using the Cox proportional hazards model and variables with a P<0.05 were entered into the final model. All statistical analysis was performed using JMP® 10 software (SAS Institute Inc, Cary, NC, USA). P<0.05 was considered statistically significant.

Results

DENND2D mRNA expression in HCC cell lines and tumor tissues

The levels of DENND2D mRNA detected using qPCR were reduced significantly relative to median value of normal liver controls in all HCC cell lines except HuH1 (Fig. 1A). In particular, the DENND2D mRNA levels were 20-times lower in Hep3B, HLE, HuH2 and PLC-PRF5 cells. When we compared the levels of DENND2D mRNA in non-cancer tissues of HCC patients with (n=35) or without (n=57) cirrhosis, no significant difference was observed, suggesting that the expression of DENND2D mRNA in non-cancer liver tissue was not affected by liver fibrosis (Fig. 1B). The expression level of DENND2D mRNA in 72 (78%) of 92 patients was lower in HCC tissues than in the corresponding normal tissues. Further, the mean expression of DENND2D mRNA was 2-fold lower in HCC tissues than in corresponding normal tissues (P<0.001, Fig. 1C).

Figure 1. (A) DENND2D mRNA expression in HCC cell lines and median value of normal liver tissues (without chronic hepatitis and cirrhosis) using qPCR. Reduced expression was detected in all HCC cell lines compared with controls, except for HuH1. (B) There were no significant differences in DENND2D mRNA expression level in livers of patients with and without cirrhosis. (C) The mean expression level of DENND2D mRNA was significantly higher in HCC tissues than in corresponding normal tissues. NS, not significant.

Identification of a CpG island in the DENND2D promoter

A CpG island was identified at the DENND2D promoter region using the CpG Island Searcher. The properties of the CpG island are: 1214 bp, 68.7% GC and 0.68 Observed CpG/Expected CpG ratio (Fig. 2A). Therefore, we hypothesized that hypermethylation of the CpG islands regulates the expression of DENND2D in HCC tissue.

Figure 2. (A) The CpG island indicated by the blue line was centered around the DENND2D transcription initiation site extending upstream into the promoter region. (B) MSP analysis. The DENND2D promoter was completely methylated in Hep3B, HepG2, HuH2, HuH7 and SK-Hep1 cells, in contrast, methylation was partial in HLE, HLF and PLC/PRF/5 cells and was not detected in HuH1 cells. (C) RT-PCR analysis before and after 5-aza-dC treatment. Reactivation or an increase if DENND2D expression was detected in all HCC cell lines except HuH1.

MSP analysis of HCC cell lines

MSP was conducted to verify the hypothesis described above. We first determined the methylation status of DENND2D in nine HCC cell lines. Bands consistent with methylated DNA were detected in all HCC cell lines except HuH1. PCR using unmethylated primers, amplified bands from HLE, HLF, HuH1 and PLC/PRF/5 DENND2D DNA (Fig. 2B). We conclude that methylation of the DENND2D promoter was complete in Hep3B, HepG2, HuH2, HuH7 and SK-Hep1; partial in HLE, HLF and PLC/PRF/5 cells and undetectable in HuH1 cells.

Transcription of DENND2D in cells treated with 5-aza-dC

To determine whether promoter hypermethylation leads to the suppression of DENND2D transcription, we analyzed HCC cell lines before and after treatment with the DNA methylation inhibitor, 5-aza-dC. Using semi-quantitative RT-PCR, reactivation or an increase in DENND2D expression was detected in all HCC cell lines except HuH1, consistent with the results of the MSP analysis (Fig. 2C).

Bisulfite sequence analysis

To confirm the results of the MSP experiments, we directly sequenced the DENND2D promoter in HepG2 (complete methylation) and HuH1 (undetectable methylation) cells and found that all CpGs in the HepG2 fragment were CG, while TG was present at the corresponding position in HuH1 cells (Fig. 3), thus confirming the MSP data.

Figure 3. Bisulfite sequence analysis of genomic DNA isolated from HepG2 and HuH1 cells. All CpGs in HepG2 were CG and TG was present at the same position in HuH1.

Prognostic value of DENND2D expression in 92 patients with HCC

Downregulation of DENND2D mRNA was detected in tumor samples from 34 of 92 (37.0%) patients with HCC. Median disease specific survival (32.6 versus 112 months, P=0.020, Fig. 4A) and relapse-free survival (10.3 versus 23.0 months, P<0.001, Fig. 4B) were significantly shorter in patients with downregulation of DENND2D mRNA. Tumor size ≥3.0 cm and vascular invasion, but not DENND2D expression, were identified as independent prognostic factors for disease specific survival using multivariate analysis (Table I). In contrast, univariate analysis for recurrence-free survival showed that serosal infiltration, vascular invasion and downregulation of DENND2D mRNA were significantly prognostic of adverse outcomes. Multivariate analysis identified downregulation of DENND2D mRNA as an independent prognostic factor for recurrence-free survival (hazard ratio 2.86, P=0.002, Table II). Downregulation of DENND2D mRNA was not significantly associated with other clinicopathological parameters.

Figure 4. Kaplan-Meier survival curves for 94 patients categorized according to downregulation of DENND2D mRNA in HCC tissues. The overall (A) and recurrence-free (B) survivals were significantly shorter in patients with downregulation of DENND2D.

Table I. Prognostic factors for disease-specific survival in 92 patients with hepatocellular carcinoma.

Table I.

Prognostic factors for disease-specific survival in 92 patients with hepatocellular carcinoma.

Variable	n	Univariate			Multivariate
		Hazard ratio	95% CI	P-value	Hazard ratio	95% CI	P-value
Age (≥65)	47	1.61	0.87–3.03	0.130
Gender (male)	77	1.18	0.55–2.92	0.695
Background liver (cirrhosis)	36	1.44	0.77–2.67	0.250
Pugh-Child’s classification (B)	7	1.28	0.31–3.57	0.692
AFP (>20 ng/ml)	47	1.90	1.03–3.57	0.042a	1.70	0.87–3.36	0.121
PIVKA II (>40 mAU/ml)	54	2.25	1.18–4.51	0.013a	1.36	0.65–2.99	0.415
Tumor multiplicity (multiple)	24	1.97	1.00–3.69	0.049a	2.26	1.08–4.62	0.032a
Tumor size (≥3.0 cm)	67	2.41	1.16–5.67	0.017a	1.18	0.49–3.07	0.712
Tumor differentiation (well)	28	0.51	0.23–1.03	0.060
Growth type (invasive growth)	17	1.16	0.51–2.35	0.705
Serosal infiltration	25	2.33	1.16–4.49	0.018a	1.05	0.46–2.34	0.898
Formation of capsule	68	0.98	0.51–2.00	0.952
Infiltration to capsule	54	1.07	0.58–2.03	0.822
Septum formation	34	0.91	0.49–1.75	0.776
Vascular invasion	23	3.33	1.73–6.23	<0.001a	2.35	1.08–5.11	0.032a
Margin status (positive)	24	2.23	1.17–4.15	0.016a	1.89	0.96–3.60	0.064
Hypermethylation of DENND2D	69	1.22	0.61–2.73	0.588
Downregulation of DENND2D mRNA	34	2.08	1.10–3.89	0.025a	1.79	0.89–3.59	0.101

a Statistically significant (P<0.05). CI, confidence interval; AFP, α-fetoprotein; PIVKA, protein induced by vitamin K antagonists. Univariate and multivariate analyses were performed using the log-rank test and the Cox proportional hazards model, respectively.

Table II. Prognostic factors for recurrence-free survival in 92 patients with hepatocellular carcinoma.

Table II.

Prognostic factors for recurrence-free survival in 92 patients with hepatocellular carcinoma.

Variable	n	Univariate			Multivariate
		Hazard ratio	95% CI	P-value	Hazard ratio	95% CI	P-value
Age (≥65)	47	1.72	0.96–3.08	0.068
Gender (male)	77	1.41	0.64–3.73	0.418
Background liver (cirrhosis)	36	1.68	0.86–3.10	0.124
Pugh-Child’s classification (B)	7	0.93	0.28–2.32	0.889
AFP (>20 ng/ml)	47	1.58	0.89–2.82	0.116
PIVKA II (>40 mAU/ml)	54	1.10	0.61–2.03	0.759
Tumor multiplicity (multiple)	24	1.06	0.54–1.96	0.861
Tumor size (≥3.0 cm)	67	1.93	0.95–4.47	0.073
Tumor differentiation (well)	28	1.05	0.51–2.02	0.879
Growth type (invasive growth)	17	1.33	0.57–2.75	0.483
Serosal infiltration	25	2.62	1.35–5.00	0.005a	2.06	1.04–4.02	0.039a
Formation of capsule	68	0.57	0.29–1.22	0.139
Infiltration to capsule	54	0.84	0.47–1.55	0.577
Septum formation	34	0.68	0.38–1.26	0.214
Vascular invasion	23	2.41	1.27–4.39	0.008a	2.12	1.08–4.01	0.029a
Margin status (positive)	24	1.79	0.97–3.23	0.064
Hypermethylation of DENND2D	69	1.11	0.59–2.24	0.749
Downregulation of DENND2D mRNA	34	2.86	1.52–5.35	0.001a	2.86	1.49–5.44	0.002a

a Statistically significant (P<0.05). CI, confidence interval; AFP, α-fetoprotein; PIVKA, protein induced by vitamin K antagonists. Univariate and multivariate analyses were performed using the log-rank test and the Cox proportional hazards model, respectively.

Methylation status of DENND2D in 92 clinical HCC samples

MSP analysis revealed that 69 (75.0%) out of 92 HCC tissue samples and only 4 (4.3%) of 92 corresponding non-cancer tissues showed hypermethylation of the DENND2D promoter. There was no significant association between hypermethylation of DENND2D in HCC tissues and overall or recurrence-free survival (Tables I and II). Analysis of the associations between the methylation status of DENND2D and clinicopathological factors, including demographics, background liver status, pathological findings and expression level of DENND2D mRNA showed that promoter hypermethylation of DENND2D in HCCs was significantly associated with tumor size ≥3 cm, serosal infiltration and downregulation of DENND2D (P=0.048, 0.049 and 0.004, respectively, Table III).

Table III. Association between methylation status of DENND2D and clinicopathological parameters in 92 HCC patients.

Table III.

Association between methylation status of DENND2D and clinicopathological parameters in 92 HCC patients.

Clinicopathological parameters	Methylation positive in tumor tissue (n)	Methylation negative in tumor tissue (n)	P-value
Age
<65 year	31	15	0.090
≥65 year	38	8
Gender
Male	59	16	0.101
Female	10	7
Background liver
Normal liver	4	5	0.080
Chronic hepatitis	36	12
Cirrhosis	29	6
Pugh-Child’s classification
A	64	21	0.823
B	5	2
Hepatitis virus
Absent	9	5	0.617
HBV	19	6
HCV	41	12
AFP (ng/ml)
≤20	34	13	0.547
>20	35	10
PIVKA II (mAU/ml)
≤40	32	7	0.175
>40	37	16
Tumor multiplicity
Solitary	53	16	0.493
Multiple	16	7
Tumor size
<3.0 cm	23	3	0.048a
≥3.0 cm	46	20
Differentiation
Well	21	6	0.689
Moderate to poor	48	17
Growth type
Expansive growth	58	18	0.533
Invasive growth	11	5
Serosal infiltration
Absent	54	13	0.049a
Present	15	10
Formation of capsule
Absent	19	7	0.790
Present	50	16
Infiltration to capsule
Absent	29	11	0.628
Present	40	12
Septum formation
Absent	26	8	0.803
Present	43	15
Vascular invasion
Absent	53	17	0.779
Present	16	6
Margin status
Negative	51	17	1.000
Positive	18	6
UICC pathological stage
I, II	44	10	0.089
III, IV	25	13
Downregulation of DENND2D
Absent	38	20	0.004a
Present	31	3

{ label (or @symbol) needed for fn[@id='tfn3-ijo-44-01-0044'] } UICC, Union for International Cancer Control; HBV, hepatitis B virus; HCV, hepatitis C virus; AFP, α-fetoprotein; PIVKA, protein induced by vitamin K antagonists; S, serosal infiltration; Fc, formation of capsule; Fc-inf, infiltration to capsule.

a Statistically significant difference (P<0.05).

IHC

The expression of DENND2D was determined using IHC in 30 cases showing relative overexpression, underexpression, or equivalent DENND2D mRNA expression in HCC tissues compared with the corresponding non-cancer tissues. Two representative cases with the lowest expression level of DENND2D mRNA in HCC tissues showed downregulation of DENND2D in the membrane and cytoplasm of tumor cells compared with adjacent non-cancer tissues (Fig. 5A and B). In contrast, equivalent expression of DENND2D protein in tumor and normal cells was detected in the case without reduced DENND2D mRNA expression in HCC tissues (Fig. 5C). The MSP analysis of these cases is shown in Fig. 5D.

Figure 5. IHC analysis of DENND2D expression in representative patients with HCC. (A) Case 1; (B) Case 2, cancer tissue cells showed downregulation levels of DENND2D compared with adjacent non-cancer tissue cells (×100). (C) Case 3, equivalent expression of DENND2D was detected in cancer and non-cancer tissue cells in patients without downregulation of DENND2D mRNA expression (×40). (D) MSP results for cases 1–3 whose IHC data are shown in (A–C). N, non-cancer tissue cells; T, tumor tissue cells.

Discussion

Research conducted over the past decade has demonstrated that certain TSGs are epigenetically inactivated in HCC, indicating that this is one of the major molecular alterations that occurs during hepatocarcinogenesis (7,29). Moreover, there is growing evidence that a major mechanism of epigenetic silencing in cancers involves hypermethylation of the TSG promoter (10,30,31). Promoter hypermethylation can be used as a sensitive marker for TSG identification, cancer diagnosis and predicting prognosis. In the present study, DENND2D was identified as another candidate TSG that is epigenetically inactivated in HCC.

We show here that the expression level of DENND2D mRNA was independent of fibrosis of the liver, because there was no significant difference in DENND2D mRNA levels between patients with and without cirrhosis. In contrast, the level of DENND2D mRNA was reduced in 8/9 HCC cell lines and in 72/92 surgically resected specimens and the mean expression level was significantly lower in cancer tissues than in corresponding normal tissues. This result indicates that DENND2D plays an important role in hepatocarcinogenesis but not in the fibrous response of the liver. Further, the expression pattern of DENND2D was consistent with that of its mRNA according to the results of qPCR and IHC analyses. A striking discovery was that a significant reduction in DENND2D mRNA level in HCC tissues correlated with remarkably earlier recurrence (hazard ratio 2.86, P=0.002, Table I) and subsequent adverse prognosis. These findings suggest that DENND2D acts as a TSG and are consistent with the results of a study on DENND2D expression in lung cancer (25). HCC frequently relapses after hepatectomy, becoming multicentric metastatic within the liver. Downregulation of DENND2D expression in HCC tissues could therefore be a biomarker of HCC recurrence.

To understand the regulation of DENND2D transcription, we performed methylation analysis of DENND2D after we identified a CpG island within its promoter. The DENND2D promoter was hypermethylated in 8/9 HCC cell lines and DENND2D transcription could be reactivated in cells treated with an inhibitor of methylation. Hypermethylation was frequently (75%) detected in HCC tissues and significantly associated with substantial (≥3-times) reduction of DENND2D mRNA levels. Therefore, we consider promoter hypermethylation as a potent regulatory factor of DENND2D transcription in HCC. Further, hypermethylation of the DENND2D promoter in HCC tissues significantly associated with tumor size and serosal infiltration, indicating the importance of DENND2D in tumor-cell growth and migration. Therefore, hypermethylation of DENND2D should provide an important new biomarker of HCC progression.

The DENND2 family is the only example where the DENN domain is located within the C-terminal region (23,32). Each DENND2 protein acts as a GEF for Rab9a/b. DENND2D is composed of only a DENN domain (33); therefore, the DENN domain mediates GEF activity. Rab9 functions in retrograde trafficking of the mannose phosphate receptor from late endosomes to the trans-Golgi network and depletion of DENND2A, but not other DENND2 family members, disrupting trafficking (34,35). This suggests that DENND2 proteins activate other Rab9 functions, such as biogenesis of lysosome-related organelles (36).

This study is limited by its lack of sufficient functional analysis of DENND2D, which tempers the conclusion that it acts as a tumor suppressor for HCC. Further studies will be required to clarify the molecular mechanisms underlying the biological activities of DENND2D in HCC.

In conclusion, we propose DENND2D as a candidate tumor suppressor gene that is inactivated by promoter hypermethylation in HCC and shows promise as a novel biomarker of early recurrence of HCC.

References