Decreased hepatocyte membrane potential differences and GABAa-β3 expression in human hepatocellular carcinoma

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

To determine whether hepatocyte membrane potential differences (PDs) are depolarized in human HCC and whether depolarization is associated with changes in GABAA receptor expression, hepatocyte PDs and γ-aminobutyric acid (GABA)A receptor messenger RNA (mRNA) and protein expression were documented in HCC tissues via microelectrode impalement, real-time reverse-transcriptase polymerase chain reaction, and Western blot analysis, respectively. HCC tissues were significantly depolarized (−19.8 ± 1.3 versus −25.9 ± 3.2 mV, respectively [P < 0.05]), and GABAA-β3 expression was down-regulated (GABAA-β3 mRNA and protein expression in HCC; 5,693 ± 1,385 and 0.29 ± 0.11 versus 11,046 ± 4,979 copies/100 mg RNA and 0.62 ± 0.16 optical density in adjacent tumor tissues, respectively [P = 0.002 and P < 0.0001, respectively]) when compared with adjacent nontumor tissues. To determine the physiological relevance of the down-regulation, human malignant hepatocytes deficient in GABAA-β3 receptor expression (Huh-7 cells) were transfected with GABAA-β3 complementary DNA (cDNA) or vector alone and injected into nu/nu nude mice (n = 16-17 group). Tumors developed after a mean (± SD) of 51 ± 6 days (range: 41-60 days) in 7/16 (44%) mice injected with vector-transfected cells and 70 ± 12 days (range: 59-86 days) in 4/17 (24%) mice injected with GABAA-β3 cDNA-transfected cells (P < 0.005). Conclusion: The results of this study indicate that (1) human HCC tissues are depolarized compared with adjacent nontumor tissues, (2) hepatic GABAA-β3 receptor expression is down-regulated in human HCC, and (3) restoration of GABAA-β3 receptor expression results in attenuated in vivo tumor growth in nude mice. (HEPATOLOGY 2007;45:735–745.)

Previous studies have demonstrated that tumor tissues are significantly depolarized when compared with adjacent nontumor tissues (reviewed in Marino et al.1). The importance of this finding remains unclear. In recent in vitro studies, we demonstrated that restoration of depolarized malignant hepatocyte cell membrane potential differences (PDs) toward those documented in nonmalignant hepatocytes results in a loss of malignant features, including decreased proliferative activity, absence of colony formation in soft agar, and normalization of phenotypic appearance.2 Whether HCC tissues are depolarized relative to adjacent nontumor tissues and the mechanisms whereby such depolarization might exist have yet to be reported.

γ-Aminobutyric acid (GABA) is the principal inhibitory neurotransmitter in the mammalian brain.3 Of the three major GABA receptor subtypes (GABAA, GABAB, and GABAC), GABAA is most prevalent both in the central nervous system and peripheral tissues.4 Although GABAAreceptors tend to exist as pentameric structures consisting of various combinations of six major subtypes (α, β, γ, δ, ε and π), five of these subtypes (excluding ϵ) have the capacity to form monomeric, ligand-gated receptor complexes.5 The following isoforms have been identified within three of the subtypes: α1-α6, β1-β3, and γ1-γ3. The most ubiquitously expressed isoforms are α1, α2, α3, β2, β3, and γ2. Activation of GABAA receptors results in marked increases in chloride ion influx and thereby hyperpolarization of cell membranes.

In previous studies, we identified GABAA receptor expression (isoforms β3 and ε) in healthy rat and human hepatocytes.6 When activated, these receptors caused hyperpolarization of resting hepatocyte PDs and inhibition of cell proliferative activity.6, 7 Conversely, when GABAA receptors were down-regulated, hepatocyte PDs were depolarized, and proliferative activity was increased.8, 9

Another electrogenic system that may play an important role in regulating hepatocyte PDs is the Na/K adenosine triphosphatase (ATPase) system. Like the GABAergic system, increased Na/K-ATPase activity results in hyperpolarization and inactivation/depolarization of hepatocyte PDs. Also, like the GABAergic system, changes in Na/K ATPase activity and in particular the α1 and perhaps β1 subunits have been described in association with changes in hepatocyte proliferative activity.10–14

The objectives of this study were to (1) document PD levels in HCC and adjacent nontumor tissues, (2) describe the nature and extent of GABA receptor and Na/K ATPase expression in these tissues, and (3) determine what impact enhanced GABAA receptor expression has on malignant hepatocyte growth in vivo.

Abbreviations

ATPase, adenosine triphosphatase; GABA, γ-aminobutyric acid; PD, potential differences; RT-PCR, reverse-transcription PCR.

Materials and Methods

HCC and adjacent nontumor tissues were obtained at the time of resection or biopsy, frozen in liquid nitrogen, and stored at −70°C until they were thawed for analyses. The demographic and histological findings of the study population with HCC are provided in Table 1. In the case of five tissue pairs, at the time of surgery small portions (maximum diameter: 2 cm) of tumor and adjacent nontumor tissue were sectioned and placed into chilled transport vials for PD determinations within 10 minutes of resection. The composition of buffer used for transport was as follows: potassium chloride, 5 mM; sodium chloride, 117 mM; sodium bicarbonate, 25 mM; sodium phosphate, 1.5 mM; glucose, 5 mM; magnesium chloride, 0.8 mM; calcium chloride, 1.2 mM; and HEPES, 10 mM. Pilot experiments performed in rats demonstrated that PD values remain stable under these conditions until transit times exceed 20 minutes (Fig. 1).

Table 1. Demographic, Histologic, GABAA-β3, and Na/K ATPase Findings in 12 Subjects with HCC
Patient NumberSexAge (years)Tumor DifferentiationGABAA-β3Na/K ATPase-α1
mRNA (copies/μg RNA)Protein (OD × 10−4)mRNA (copies/μg RNA)Protein (OD × 10−4)
HCCAdjacentHCCAdjacentHCCAdjacentHCCAdjacent
1M32moderate3,43817,3303,9756,8051,03468323,606622
2M41well7,6925,7602,6545,6702,8425516,689491
3M36well7,3779,6711,6562,2102,7192,7848,0611,287
4M15well4,6668,9251,3636,8223,4451,9702531,689
5M22moderate5,21410,7405,3477,0022,0924,5038,218303
6M47well4,1335,8182,6914,0124,2881,250506168
7M29poor to moderate5,5355,7161,9435,7771,81819617,49122,047
8M43moderate4,6439,5144,1146,6534,7202,390367946
9M34moderate5,0665,8773,0697,4302,7722,8712072,983
10M25moderate7,34912,0803,7237,6484,6291,55418823,754
11M38well5,30719,7702,7857,0241,4733,0352,50790
12M31moderate7,29618,6502,1237,8522,5712,56125,105182
Mean 33 5,69311,0462,9966,2462,8672,0297,7664,547
SD 9 1,3854,9791,1041,6311,2071,2449,3118,620
Figure 1.

Representative recording of hepatocyte PDs from a healthy rat liver biopsy. Tissues were bathed in chilled buffer for the period indicated. PDs remained stable for approximately 20 minutes.

This study was approved by the University of Manitoba Conjoint Ethics Committee for Human and Animal Experimentation.

Hepatocyte PD Measurements.

Hepatocyte PDs were measured using intracellular microelectrodes as previously described by Burczynski et al.15 Briefly, single-barreled microelectrodes were drawn on a horizontal puller (Brown-Flaming micropipette puller, model P-87, Sutter Instruments, Novato, CA) using Omega-Dot borosilicate tubing (1.0 mm OD, 0.5 mm ID; Sutter Instruments) filled with 0.5 M potassium chloride and beveled to a 30-degree angle to a tip resistance of approximately 120 MΩ. Electrical signals were conducted via an Ag-AgCl electrode connected to an Axoprobe-1A amplifier (Axon Instruments, Foster City, CA). Data acquisition was controlled by interfacing the amplifier to an 80486 computer via a TL-1 DMA (Axon Instruments) and using the computer program AxoTape (version 2.0; Axon Instruments). Junction potential was accounted for using JPCalc (version 2.2, University of New South Wales, Sydney, Australia). Criteria for acceptable impalements included an abrupt negative deflection on penetration of the cell, a stable intracellular potential for at least 2 minutes, and a return of the electrical potential to within 2 mV of baseline after withdrawal of the microelectrode to the bathing buffer.

Cell microelectrode input resistance was continuously monitored during impalement by passing a current pulse of 0.1 nA, 150-ms duration, every 10 seconds through the microelectrode using a Winston Electronics Timer (Model A65; San Francisco, CA). Following impalement, voltage deflections (V) corresponding to the current pulse (I) were used to estimate cell conductance (gcell = I/V).

GABAA Receptor and Na/K ATPase mRNA Expression: Semiquantitative Reverse-Transcription PCR.

Total RNA was extracted from tissues via the commercially available Trizol method (Invitrogen, Carlsbad, CA). Twenty microliters of reverse-transcription reactions consisted of the following: 1 μg RNA, 5× reaction buffer (BD Bioscience Clontech, Palo Alto, CA), 0.5 mM deoxyribonucleoside triphosphate, 0.5 U RNase inhibitor, 20 pmol oligo (dT)18 primer, and 20 U Moloney murine leukemia virus reverse transcriptase. Reactions were incubated at 42°C for 60 minutes and terminated at 99°C for 5 minutes. Five microliters of the reactions were used for the PCR reaction. The oligonucleotide primers for PCR were designed against human GABAA-α1-6, -β1-3, -γ1-3, -δ, -ε, and -π, Na/K ATPase-α1 and Na/K ATPase-β1 sequences using an Oligo 5.0 program (NBI, National Biosciences Inc., Plymouth, MN). The sequences of the GABAA receptor oligonucleotide sense and antisense primers are provided in Table 2. The primers were as follows: Na/K ATPase-α1, 5′-TCG CAC TGT GAT GGG AAG AA-3′ (forward) and 5′-GGC ACA TTG GCT ACG ATG A-3′ (reverse); Na/K ATPase-β1, 5′-AAA TGT CCT TCC CGT TCA-3′ (forward) and 5′-TAT GCG AAT TTC AGT GTC CA-3′ (reverse). PCR amplifications were performed in 30 cycles of denaturation (94°C, 45 s), annealing (53°-58°C, 45 s), and elongation (72°C, 2 min) and with an additional 7-minute final extension at 72°C. Finally, 10 μl of the PCR products were run on 2% agarose gels to document gene expression and product lengths. β-Actin RNA was used as a loading control in PCR reactions. The β-actin primers were 5′-GGA CTT GGA GCA AGA GAT GG-3′ (forward) and 5-AGC ACT GTG TTG GCG TAC-3′ (reverse). The product length was 196 bp.

Table 2. Gene-Specific PCR Primers for Human GABAA Receptor Subunits
SubunitPrimerPositionProduct Size (bp)
α1 432–1,334902
 Sense5′-TCGTCACCAGTTTCGGACC-3′  
 Antisense5′-GGTTGCTGTTGGAGCGTAA-3′  
α2 596–1,318722
 Sense5′-TTCACAATGGGAAGAAATCAGTAG-3′  
 Antisense5′-TGCATAAGCGTTGTTCTGTATCA-3′  
α3 779–1,437658
 Sense5′-GGAAGTGGCACAGGATGGTTC-3′  
 Antisense5′-GTTGTAGGTCTTGGTCTCAGTCGG-3′  
α4 577–1,327750
 Sense5′-TGAAATTCGGGAGTTATGCCTATC-3′  
 Antisense5′-GGCTGAATGGGTTTGGACTG-3′  
α5 786–1,612826
 Sense5′-CACCATGCGCTTGACCATCTCT-3′  
 Antisense5′-GCCGAACAAGACTGGGAATA-3′  
α6 46–1,233764
 Sense5′-TGAGGCTTACCATCAATGCTGA-3′  
 Antisense5′-GACAGGTGTTGATTGTAAGATGGG-3′  
β1 476–1,079603
 Sense5′-GTTCTCTATGGACTCCGAATCACA-3′  
 Antisense5′-ATTGGCACTCTGGTCTTGTTTG-3′  
β2 928–1,568640
 Sense5′-AGCTTAAGAGAAACATTGGCTACT-3′  
 Antisense5′-CGATCTATGGCATTCACATCA-3′  
β3 497–1,130633
 Sense5′-AGTGCTGTATGGGCTCAGAATCAC-3′  
 Antisense5′-CCCGGTTGCTTTCGCTCTT-3′  
γ1 1,117–1,379262
 Sense5′-GTGTTTTGCAGCCTTGATGG-3′  
 Antisense5′-TGGCAATGCGTATGTGTATCCT-3′  
γ2 624–1,229605
 Sense5′-AAGTCCTCCGATTGAACAGCAACA-3′  
 Antisense5′-CGCTGTGACATAGGAGACCTT-3′  
γ3 578–1,345767
 Sense5′-ACACTCCTGCCCGCTGATT-3′  
 Antisense5′-TGTCTATGTGAATACGCCCTTTCC-3′  
δ 641–1,295654
 Sense5′-TCACCATCACCAGCTACCACTTCA-3′  
 Antisense5′-GGGCGTAAATGTCAATGGTGTC-3′  
ε 861–1,493632
 Sense5′-GCAGGCGGTTTGGCTATGT-3′  
 Antisense5′-CGAGTAGTTATCCAGGCGGTAG-3′  
π 237–487250
 Sense5′-CGTCGAGGTCGGCAGAAGT-3′  
 Antisense5′-GCGGGCATCCAGAGTGAAG-3′  

Quantitative Real-Time Reverse-Transcription PCR.

Quantitative reverse-transcription PCR (RT-PCR) was performed in a LightCycler (Roche Diagnostics, Mississauga, Ontario, Canada) using the RNA Master SYBR Green I kit (Roche Diagnostics). A 290-bp fragment of GABAA-β3 was amplified using the primers GABA-β3-645U (5′-AAG GCT GTT ACC GGA GTG GA-3′) and GABA-β3-916L (5′-CGA AGG TGG GTG TTG ATG G-3′). Each 20-μl PCR reaction contained 100 ng of total RNA, 3.5 mM MgCl2, 0.5 μM of each primer, and 2 μl of master mix. Real-time PCR protocol was then employed, beginning with reverse transcription at 55°C for 10 minutes followed by PCR that started at 95°C for 30 seconds and proceeded for 45 cycles in four steps: 95°C for 0 seconds, 57.7°C for 30 seconds, 72°C for 10 seconds, and 72°C for 25 seconds. After amplification, a melting curve was generated by heating the samples to 95°C for 0 seconds, followed by cooling to 65°C for 15 seconds and heating to 95°C at a rate of 0.1°C/s, at which stage SYBR Green I fluorescence was measured continuously. Samples were then cooled to 40°C for 30 seconds. For standards, GABAA-β3 receptor cDNA was obtained from human liver Marathon ready cDNA (Clontech) and cloned into a pCDNA-3.1/his-v5 plasmid. The plasmid containing GABAA-β3 receptor cDNA was linearized via restriction with XbaI and retranscripted to GABAA-β3 receptor RNA using the Ribprobe in vitro Transcription Kit (Promega, Madison, WI). The copy number was estimated by optical density according to the exact molar mass derived from the plasmid and GABAA-β3 receptor cDNA sequences. Different dilutions were made to obtain 102-106 copies in 2 μl, which were used to generate standard curves.

SDS Gel Electrophoresis and Immunoblotting Techniques.

Protein extracts (50 μg) were separated on 12% polyacrylamide–SDS gels and electroblotted to nitrocellulose membranes as described.16 Membranes were blocked with 5% skim milk in Tris-buffered saline [0.02 M Tris-base (pH 7.6)] for 1 hour at room temperature and incubated with rabbit anti-human GABAA-β3 receptor antibody (5.55 μg/ml; kindly provided by Dr. W. Sieghart, University of Vienna, Vienna, Austria) overnight at 4°C. Bands were detected with a horseradish peroxidase–labeled secondary antibody–catalyzed chemiluminescence reaction (Amersham Pharmacia Biotech, Burlington, Ontario, Canada). Controls included rat brain microsomal protein (Upstate Biotechnology Ltd., Lake Placid, NY), and membranes were incubated with secondary antibody but without prior incubation with primary anti–GABA-β3 receptor antibody.

Genomic DNA Isolation and Analysis of GABAA-β3 Promoter Methylation.

Genomic DNA was isolated from HCC and adjacent nontumor tissues using a ZR Genomic DNA II kit (Zymo Research, Orange, CA). Sodium bisulfite modification of the DNA was performed using an EZ DNA Methylation kit (Zymo Research). PCR (Advantage 2 Polymerase Mix, BD Bioscience Clontech) was employed to amplify the human GABAA-β3 gene promoter region (−907 to +98) from the bisulfite modified genomic DNA. The PCR reaction included 32 cycles, 58°C annealing temperature, and the following primers: 5′-CTGCGGTCACATTTTCTGTTCCAA-3′ and 5′-GGTCCCCAGGGTCCAGGAG-3′. This region of the GABAA-β3 promoter contains abundant CpG islands (nucleotides −768 to −644). PCR products were purified with the Wizard SV Gel and PCR Clean-Up system kit from Promega as per the manufacturer's recommendations. Hot-start methylation-specific PCR used two sets of primers designed with MethPrimer software17 to amplify regions of interest (35 cycles, 52°C annealing temperature). One pair recognized sequences in which CpGs are unmethylated, while the other recognized sequences with methylated CpGs. The primers used were unmethylated sense 5′-TTAATTGTATAAATGAAAAATAGGGTTGT-3′, antisense 5′-CCAACCTAACCTACTAAAATCCACT-3′; methylated sense 5′-TTAATTGTATAAATGAAAAATAGGGTCGT-3′, and antisense 5′-GCACTAACCTACTAAAATCCGCT-3′. The methylation-specific PCR products underwent electrophoresis on 2.5% agarose gels and were visualized under ultraviolet illumination after ethidium bromide staining.

GABAA-β3 Receptor DNA Sequencing.

Total RNAs were extracted from liquid nitrogen–frozen HCC and adjacent nontumor tissues with Trizo reagent (Invitrogen, Burlington, Ontario, Canada). RNAs were transcripted to cDNAs using the Advantage RT-for-PCR Kit (BD Bioscience Clontech), and the cDNAs were used as templates to perform PCR (Advantage 2 Polymerase Mix, BD Bioscience Clontech). The primers used to amplify the entire length (1,437 bp) of the open reading frame for human GABAA-β3 receptor cDNA were: forward primer 5′-TAGGTACCATGGGGGGCCTTGCGGGAGG-3′, and reverse primer 5′- GCGCGGCCGCTCAGTTAACATAGTACAGCC-3′. PCR products showing a single band on agarose gels were purified from the cut gels with the Wizard SV Gel and PCR Clean-Up system kit (Promega). For the sequencing reactions, 2 μl of 2 ng/μl cDNA were used with the Applied Biosystems 3730 DNA Analyzer; the reactions were performed by Robarts Research (London, Ontario, Canada). Sequences of the GABAA-β3 receptor from HCC and adjacent nontumor tissues were compared with sequences of the human GABAA-β3 receptor from GENEBANK/M82919 (National Center for Biotechnology Information) using GeneJokey Sequence Processor (Biosoft, Cambridge, UK).

GABAA-β3 Receptor Transfections.

Plasmid construction and stable transfections were performed as described previously.2 Briefly, the 1,640-bp cDNA containing the GABAA-β3 receptor coding region was cut with XbaI from the plasmid pRK5-β3 and cloned into the pcDNA3.1/V5-His C vector producing a plasmid known as pcDNA-β3. The cloned cDNA fragment is under the transcriptional control of the immediate early gene of the human cytomegalovirus promoter, and the vector contains polyadenylation signals and ampicillin-resistant and zencin-resistant genes. pcDNA-β3 plasmids were isolated, purified, and sequenced to confirm that the cDNA of the GABAA-β3 receptor was in frame with the cytomegalovirus promoter. Thereafter, 1 × 106 GABAA-β3 isoform–deficient, human malignant hepatocytes (Huh-7 cells) in a 3.5-cm dish were transfected with 1 μg of linearized pcDNA-β3 or pcDNA vector alone using Lipofectamine (GIBCO/BRL, Carlsbad, CA) according to the manufacturer's instructions. Stable transfected cells were established in the presence of G418 (800 μg/ml), and resistant clones were isolated by using cloning cylinders and maintained under G418 selection (200 μg/ml). Clones were analyzed individually via RT-PCR and western blotting for levels of GABAA-β3 receptor mRNA and protein expression, respectively. Restoration of GABAA-β3 receptor mRNA and protein expression after transfection were confirmed via both RT-PCR and western blot analyses but remained undetectable in cells transfected with vector alone (Fig. 2).

Figure 2.

RT-PCR (left) and western blot analysis (right) of GABAA-β3–deficient Huh-7 human malignant hepatocytes following stable transfection with either GABAA-β3 cDNA in a pcDNA 3.1/v5-HisC vector (Huh7-β3) or vector alone (Huh7-V).

Nude Mice Studies.

Male Crl:nu/nu (CD-1) BR nude mice (Charles River Canada, Saint Constant, Quebec, Canada) weighing approximately 20 g were housed in laminar-flow cabinets under specific pathogen-free conditions. At 4-6 weeks of age, the hind flanks of 34 mice were injected with 1 × 107 Huh-7 cells transfected with either GABAA-β3 receptor cDNA (n = 17) or vector alone (n = 17) and assessed daily (by an individual blinded to each mouse's groupings) for tumor development over the subsequent 120 days. Tumors were considered detectable when they reached a minimum diameter of 5 mm as measured via electronic digital calipers (Fisher Scientific Ltd, Nepean, Ontario, Canada). In pilot studies, neither pcDNA5vector plasmid nor pcDNA5-β3 injections were associated with local reactions or toxicity. Mice were cared for according to the National Institutes of Health Guidelines for Care and Use of Laboratory Animals.

Statistical Analyses.

Student t tests were performed for parametric and Wilcoxon rank sum tests for nonparametric data. A P value of less than 0.05 was considered statistically significant. The results are expressed as the mean ± SD.

Results

Figure 3 provides the results of hepatocyte PD determinations in HCC and adjacent nontumor tissues (n = 5). Consistent with previous in vitro findings in malignant and nonmalignant human hepatocyte cell lines,2 HCC tissues were significantly depolarized when compared with adjacent nontumor tissues (−19.8 ± 1.3 vs. −25.9 ± 3.2 mV, respectively [P < 0.05]).

Figure 3.

Hepatocyte membrane potentials in 5 pairs of human HCC and adjacent nontumor tissues. Determinations were performed by documenting cell microelectrode input resistance within 10 minutes of tissue resection. Relative to adjacent nontumor tissues, human HCC tissues were significantly depolarized (P = 0.03).

To determine whether the depolarization of HCC tissues was associated with changes in GABAA receptor and/or Na/K ATPase expression, we documented GABAAand Na/K ATPase mRNA expression in 12 pairs of HCC and adjacent nontumor tissues as well as 5 healthy human liver tissues. The results of semiquantitative RT-PCR on a scale of + (limited expression) to ++++ (highly expressed) are provided in Table 3. As described previously, healthy livers only expressed GABAA-β3 and ε receptor isoforms (data not shown). In tissue adjacent to HCC, GABAA receptor expression was limited to GABAA-β3, ε, and π. In HCC tissues, although GABAA-ε and π expression remained largely unchanged, GABAA-β3 mRNA expression was consistently (11/12) and significantly down-regulated. Also evident in HCC tissues was the appearance and, in one tissue set, slight increase in expression of the GABAA-α3 (7/12) isoform and to a lesser extent GABAA-γ2 (3/12), -γ3 (3/12), -α1 (2/12), -β1 (2/12), and -α4 (1/12) isoforms.

Table 3. GABAA Receptor Subtype/Isoform mRNA Expression in 12 Sets of Human HCC and Adjacent Nontumor Tissues
SubtypeSet
123456789101112
Nontumor            
 GABAA-α1            
 GABAA-α2            
 GABAA-α3            
 GABAA-α4          ++
 GABAA-α5            
 GABAA-α6            
 GABAA-β1            
 GABAA-β2            
 GABAA-β3+++++++++++++++++
 GABAA-γ1            
 GABAA-γ2            
 GABAA-γ3            
 GABAA            
 GABAA+++ +++++++++
 GABAA+++ +++ ++++++
Tumor            
 GABAA-α1+        +  
 GABAA-α2            
 GABAA-α3+++  ++ + +++ +++
 GABAA-α4          ++ 
 GABAA-α5            
 GABAA-α6            
 GABAA-β1   +     ++ 
 GABAA-β2            
 GABAA-β3 +++++  ++ ++
 GABAA-γ1            
 GABAA-γ2+ +      +  
 GABAA-α3+   +    +  
 GABAA        +   
 GABAA++++++++++++++++++++
 GABAA +++ +++++++++

Real-time RT-PCR was performed to more accurately quantitate changes in GABAA3receptor mRNA expression. As shown in Figure 4 and Table 1, expression was decreased by approximately 50% in HCC compared with adjacent nontumor tissues (5,693 ± 1,385 vs. 11,046 ± 4,979 copies/100 mg total RNA, respectively [P = 0.002]). When analyzed individually, relative to adjacent nontumor tissues, GABAA-β3 mRNA expression was decreased in all but 1 of the 12 HCC tissues. Healthy human liver GABAA-β3 receptor mRNA expression was variable but more closely resembled that of adjacent nontumor tissue than HCC tissue (Fig. 4).

Figure 4.

Patient-specific GABAA-β3 mRNA expression via quantitative real-time RT-PCR in 12 pairs of human HCC and adjacent nontumor tissues and 5 healthy human liver tissues. In all but one of the paired tissues (patient 2), GABAA-β3 mRNA expression was decreased in tumor compared with adjacent nontumor tissues. In patient 2, sequencing of the full-length GABAA-β3 open reading frame did not reveal mutations that would be predicted to result in impaired receptor function.

The results of GABAA-β3 protein determinations supported the mRNA findings (Fig. 5 and Table 1). Specifically, in HCC tissues, GABAA-β3 protein levels were only 0.29 ± 0.11 versus 0.62 ± 0.16 × 10−4 in adjacent nontumor tissues (P < 0.0001), a decrease of approximately 50%. Of note, GABAA-β3 protein levels were decreased in all 12 HCC tissues, including the one tissue set in which transcript levels were slightly higher in HCC than in adjacent nontumor tissue. Sequencing of the open reading frame for the human GABAA-β3 receptor gene from this HCC tissue did not reveal mutations that would explain the decreased protein level relative to the adjacent nontumor tissue (data not shown). Once again, GABAA-β3 receptor protein expression in healthy human livers was variable but more closely resembled that of adjacent nontumor tissue than HCC tissue (Fig. 5A).

Figure 5.

(A) GABAA-β3 receptor protein levels via western blot analyses in 12 pairs of human HCC and adjacent nontumor tissues and 5 healthy human liver tissues. As observed at the transcript level, compared with healthy livers and tissue adjacent to tumor tissues, GABAA-β3 receptor protein expression was significantly down-regulated in human HCC tissues (P < 0.0001). (B) Results in individual pairs (odd-numbered lanes represent adjacent nontumor tissues; even-numbered lanes represent HCC tissues). Depending on the degree of glycosylation and/or alternative splicing, the weight of GABAA-β3 receptors is approximately 53-59 kDa and often appears as 2 distinct bands. β-Actin expression was used as a loading control.

To elucidate the mechanism whereby GABAA-β3 gene expression is down-regulated in HCC tissues, we documented the methylation status of the upstream promoter (−907 to +98) in 3 randomly selected HCC tissues. This region (nucleotides −768 to −644) contains abundant CpG islands, the hypermethylation of which would be predicted to result in attenuated downstream GABAA-β3 gene expression. The GABAA-β3 promoter gene was methylated in both HCC and healthy control livers (Fig. 6). An unmethylated form of the promoter gene was also present in 2 of the 3 HCC tissues, a finding not seen in control livers.

Figure 6.

Methylation status of the upstream GABAA-β3 receptor promoter gene in 3 human HCC and 2 healthy control livers as determined via RT-PCR of bisulfite-modified genomic DNA using computer-generated primer sequences. The promoter region derived from HCC tissues was not hypermethylated relative to control livers.

Figures 7 and 8 and Table 1 provide the results of Na/K ATPase-α1 mRNA and protein determinations. Here, expression was similar in HCC compared with adjacent nontumor tissues. The results of Na/K ATPase-β1 mRNA and protein expression were similar to those of α1 mRNA and protein expression (data not shown).

Figure 7.

Na/K ATPase-α1 receptor mRNA expression via real-time RT-PCR in 12 pairs of human HCC and adjacent nontumor tissues. The increase in expression in tumor versus adjacent nontumor tissues was not significant (P = 0.06).

Figure 8.

Na/K ATPase-α1 receptor protein expression via Western blot analysis in 12 pairs of human HCC and adjacent nontumor tissues. Differences in expression between tumor and adjacent nontumor tissues were not significant.

Finally, to determine whether the down-regulation of GABAA-β3 receptor mRNA expression in HCC influences malignant hepatocyte growth in vivo, GABAA-β3–deficient Huh-7 cells transfected with either GABAA-β3 cDNA or vector alone were injected (1 × 107 cells) into the hind flanks of nude mice (n = 17/group). One control mouse died of unknown causes within 24 hours of injection. In the remaining mice, tumors developed in 7/16 (44%) mice injected with vector-transfected cells after a mean of 51 ± 6 days (range: 41-60 days), whereas 4/17 (24%) mice (P = 0.28) injected with GABAA-β3–transfected cells developed tumors after a mean of 70 ± 12 days (range: 59-86 days [P < 0.005]).

Discussion

To date, cellular PDs have been documented in tumor and adjacent nontumor tissues from 4 sites in humans: the breast, cervix, brain, and stomach.1, 18–21 In each case, tumor PDs were significantly depolarized (breast, 18%; cervix, 33%; brain, 54%; stomach, 34%). Thus, our finding of a 24% decrease in HCC tissue PD is not only consistent with these reports but also reports describing depolarization in various rat and human malignant cell lines.2, 22 Although the mechanism whereby such depolarization might result in enhanced cell proliferation and tumor growth remains to be determined, electrogenic translocation of polyvalent cationic growth promoters such as polyamines from the relatively positive-charged environment of the cytoplasm to the negative-charged DNA of the nucleus has been implicated.23

Perhaps the most striking finding of the present study was the consistent and significant decrease in GABAA-β3 receptor expression in HCC compared with adjacent nontumor tissues. This finding supports previous descriptions of down-regulated or absent expression of this receptor isoform in human malignant hepatocyte cell lines and regenerating livers following partial hepatectomy.2, 6 Although the correlation of decreased receptor expression with increased proliferative activity may represent an association rather than effect, previous in vitro studies have shown that restoration and/or activation of GABAA-β3 receptors results in attenuation of proliferative activity in malignant hepatocytes.2, 24, 25 Moreover, reports of chromosomal deletions in the region encoding the GABAA-β3 receptor gene being associated with hepatic tumor development in humans also argues in favor of a causative role.26 Finally, the results of the nude mice experiments described in this study in which restoration of GABAA-β3 receptor expression in GABAA-β3–deficient Huh-7 cells resulted in attenuated tumor growth compared with mice injected with Huh-7 cells transfected with vector alone also supports a causative role.

The mechanism(s) whereby GABAA-β3 receptor expression is down-regulated in HCC remains unclear. Because the GABAA-β3 gene has been localized to a region of chromosome 15 that is associated with genetic imprinting,27, 28 hypermethylation of the CpG islands present on the upstream promoter could explain this finding.29 However, we could not detect such hypermethylation using computer-generated primers. Thus, additional studies are required to elucidate the mechanism(s) involved.

Regarding the expression of other GABAA receptor subtypes/isoforms, that GABAA-α3 expression became detectable in the majority of HCC tissues has not been previously described, and the significance of this finding has yet to be elucidated. Although GABAA-α1, -α4, -β1, -γ2, and -γ3 expression either became detectable or, as in one case (α4 isoform), increased in expression, this activity occurred in only a minority of tumors and is therefore of questionable significance. Moreover, these findings were not unexpected, because previous reports have documented that the spectrum of GABAA receptor isoform expression correlates with the state of cell differentiation.30, 31 It should be noted, however, that because HCC tissues are significantly depolarized compared with adjacent nontumor tissues, it is unlikely the limited expression of these isoforms results in the formation of functional GABA-gated receptors.

Unlike the results reported by Johnson and Haun32 in pancreatic tumors, we did not observe a down-regulation of GABAA-π mRNA expression in tumor versus nontumor tissues. The reasons for this discrepancy are unclear, because the oligonucleotide primers employed for GABAA-π mRNA detection and the number of tumors analyzed (n = 12) were identical in both studies.

Despite years of research, the actual contribution of Na/K ATPase activity to total hepatocyte PD remains unclear; some investigators suggest it is a major determinant, while others argue that its contribution is minimal.33, 34 Some of this controversy relates to the uncertainty regarding which subtypes are expressed in the human liver. Although there is general agreement on the presence of the α1 subunit, the status of the β1 subunit is less clear.13, 35, 36 We were able to detect both α1 and β1 expression via RT-PCR. However, the pattern of each subunit expression was similar (and certainly not decreased, which would be required to explain HCC depolarization relative to adjacent nontumor tissues). It is tempting to speculate that the trend (P = 0.06) toward Na/K ATPase-α1 up-regulation in HCC represents a compensatory response of the malignant cell to a depolarized state. While determinations of actual Na/K ATPase enzyme activity would have been ideal (enzymatic activity has been reported to be increased despite no change in transcript expression), tissue samples were inadequate to perform such an analysis.37

Although attenuated, the presence of GABAA3receptor expression in human HCC has potential clinical relevance. For example, GABA and GABAA receptor agonists have been demonstrated to inhibit tumor growth in other tissues.38 Conversely, GABAA receptor antagonists such as fluoroquinolone antibiotics and calcineurin inhibitors could theoretically enhance tumor growth and might therefore be avoided in patients with HCC.39–41 The discrepancy between HCC and adjacent nontumor tissue PDs could also be exploited in terms of developing charged compounds that would serve as the basis for new imaging modalities when the diagnosis of HCC is being considered.42

In conclusion, relative to adjacent nontumor tissues, HCC tissues are depolarized and have decreased GABAA3 receptor expression. Moreover, restoration of GABAA3 receptor expression in receptor deficient malignant hepatocytes results in attenuated in vivo tumor growth in nude mice. These findings highlight the importance of elucidating the role hepatocyte PDs and/or GABAergic activity play in the pathogenesis of HCC. They also raise the potential for new therapeutic and diagnostic approaches to HCC in humans.

Acknowledgements

The authors thank S. Zdanuk for her prompt and accurate typing of the manuscript.

Ancillary