Research ArticleExperimental Studies

Novel Insights into the Expression of CGB1 & 2 Genes by Epithelial Cancer Cell Lines Secreting Ectopic Free hCGβ

BEATA B. BURCZYNSKA, LATIFA KOBROULY, STEPHEN A. BUTLER, MAHMOUD NAASE and RAY K. ILES
Anticancer Research May 2014, 34 (5) 2239-2248;

Abstract

Background: Ectopic secretion of human chorionic gonadotrophin free beta (hCGβ) by epithelial cancer is associated with aggressive tumors which more readily metastasize, possibly by acting as an autocrine anti-apoptotic agent. hCGβ is encoded by six homologous CGB genes, with poorly-understood variable transcriptionally active expression profiles; CGB1 and CGB2 have always been considered pseudogenes. However, transcripts from CGB1 and -2 can be detected in placental, testicular and pituitary tissues. The expression and function of these genes in cancer is less well-known. Materials and Methods: Expression profiles of CGB genes in epithelial cancer cells by quantitative polymerase chain reaction (qPCR) were explored, along with the consequence of specific siRNA silencing of CGB1 and 2. Immunohistochemical and immunoassay techniques were used to detect the translation and secretion of hCGβ in these cells. Results: CGB1 and -2 gene transcripts were only detected in cells which secreted hCGβ. siRNA-mediated silencing of CGB1 and -2 transcripts significantly reduced secreted protein in concordance with a reduction in cell survival to a greater degree than that of other CGB genes. Conclusion: CGB genes 1 and 2, previously considered as pseudogenes, are notably expressed by epithelial cancer cell lines. The transcription of these genes, but not other CGB genes, correlates with a functionally expressed protein and propensity for cancer growth.

Ectopic expression of human chorionic gonadotropin free beta-subunit (hCGβ) by many types of epithelial cancer is now a recognised phenomenon, overall, in ~32% of carcinomas (9-55% in different cancer subtypes) (1, 2). Such expression is associated with poor prognosis in 87% of cases where prognosis was investigated (1). However, the underlying molecular mechanisms of expression and subsequent influence on oncogenesis are poorly-understood. We reported that hCGβ acts as an autocrine growth factor (3) by inhibiting apoptosis (4). Homodimers of hCGβ have been identified (5) but are not necessarily required for autocrine activity or receptor binding (6). The target receptor for free beta subunit of hCG has not yet been established (7). However, it has been shown that the hormonal luteinizing hormone (LH)/hCG receptor is not implicated (8). It has been inferred that hCGβ-positive cells have a selective survival advantage and potentially undergo further replication and mutation as a result of CGB gene expression.

The beta subunit of hCG is encoded by six CGB genes: CGB (originally CGB3), CGB5, CGB7 and CGB8 (classical hCGβ genes), located on chromosome 19q13.3 in close proximity to CGB1 and CGB2 and the gene for LHβ (CGB4) (9) (Figure 1). While classical CGB genes have been shown to be functional, genes CGB1 and CGB2 were always assumed to be pseudogenes (10).

Unlike the classical CG genes, which each produce a single mRNA transcript, CGB1 and CGB2 potentially have four splice variants (11) (Figure 1). The four variants include a predominant transcript coding a theoretical 132-amino-acid protein (which has never been isolated or characterised to date), and three other splice variants: +47, +166, +176 bp that contain additional DNA sequences (47 bp, 166 bp and 176 bp respectively) taken from within intron 1. An alternative theoretical form of +47 bp variant would result from the re-shift of the open reading frame (ORF) to that of the classical hCGβ-coding transcripts, and would therefore produce an additional short transcript (essentially another exon, which has been termed exon 1b) from the middle of intron 1 (12, 13). Alternative mRNAs +166 bp and +176 bp may hypothetically code for a shorter polypeptide of about 60 amino acids in length (12, 14). Additionally, in testes, a novel rare splice variant form was detected which contains an extra 87-bp exonic sequence (13).

The incorporation of CGB1 and CGB2 within the human genome is believed to have evolved in the cluster as a result of a DNA fragment insertion (736 bp for CGB1 and 724 bp for CGB2) that replaced the 52-bp sequence at the proximal end of the promoter, and also the entire 5’-UTR of the ancestral hCGβ-subunit coding gene fragment, which is still present in classical CGB genes (11, 15). Replacement of the CGB5’-UTR provided a novel putative promoter segment, 5-UTR and exon 1 (which codes for a lead sequence cleaved during synthesis). A single bp frameshift in the ORF for exons 2 and 3 in genes CGB1 and -2 has been proposed to result in an optional earlier stop codon and shorter exon 3 in the transcripts. Thus, expression of CGB1 and -2 genes may serve as a template for a hypothetical protein that is entirely different from the hCG beta subunit of pregnancy, one that lacks similarity to any known hCG protein (11). Previously CGB1 and -2 transcript mixtures have been detected in the placenta (11, 16), testes (17), pituitary gland (18) and other non-cancerous, non-trophoblastic tissues (12). The mRNAs of CGB1 and -2 were observed in brain tissue at a level equal to that found in placenta (19). However, all were detected at much lower levels than that of classical CGB genes expressed by placenta.

In cancer, overexpression of CGB genes does not necessarily correlate with hCGβ protein levels (1). However when CGB genes are silenced, both the secretion and effect of hCGβ are negatively impacted (20, 21). More specifically, the genes CGB, CGB5 and CGB8 are found to be variably expressed in cancer tissue such as of the breast, thyroid, and prostate, but not in non-cancerous tissues of corresponding organs (22), and the genes CGB8, CGB5, CGB and CGB7 (in order of decreasing expression) were found unregulated in invasive breast tumors. CGB7 was also detected in normal breast tissue (23). Additionally, overexpression of CGB, CGB5 and CGB8 in bladder cancer cell lines but lower expression of CGB7 was shown by others (24, 25). In all these studies, neither CGB1 nor CGB2 were specifically addressed within the experimental design, in some cases, perhaps, because they have always been considered pseudogenes and therefore of no significance (10).

In the present study, we re-examined CGB gene expression in non-trophoblastic cancer cells, with a particular focus on the often disregarded genes CGB1 and -2, to explore their potential function in cellular oncogenesis.

Materials and Methods

Cell culture and controls. The human bladder cancer cell lines RT112, SCaBER and T24, breast cancer cell lines C2235 and C2238, and prostate cancer LN-CAP and PC-3 were obtained from the American Type Culture Collection (Manassas, VA, USA) and from the European Collection of Animal Cell Cultures (Porton Down, Dorset, UK). Cells were cultured in RPMI-1640 medium (Invitrogen, Paisley, Scotland, UK) supplemented with 10% fetal bovine serum (FBS) (Invitrogen) with 5% antibiotics at 37°C in 5% CO2. Positive and negative controls were normal-term placental tissue (placenta obtained after vaginal delivery in the 39th week of pregnancy) and fresh water shrimp, which has no CG-like genomic sequence, respectively.

Immunohistochemical staining. Cancer cells were grown, as above, on cover slips until they reached 80% confluence, and then fixed in 4% paraformaldehyde followed by incubation in 3% H2O2 and washing in phosphate buffered saline (PBS). For permabilization, 0.1% Triton in PBS was added for 8 min, then cells were washed three times in PBS, and then blocked in 50% horse serum for 15 min. The cells were incubated for one hour with 1:100 (diluted in PBS) primary mouse monoclonal anti-hCG beta epitope 7 antibody (INN-hCG-68; Abcam, Cambridge, Cambridgeshire, UK) followed by washing in PBS. This method has been described previously (26). This antibody does not cross-react with intact hCG or hLH/LH free beta-subunit (17, 27). In the control samples, primary antibodies were excluded and substituted with PBS. For detection, we applied the ABC technique and used Vectastain Universal Elite ABC kit (Vector Laboratories, Peterborough, Cambridgeshire, UK) according to the protocol described by the manufacturer. Peroxide/3,3’-Diaminobenzidine (DAB) solution was applied until required stain intensity develops. The cell nuclei were counterstained with haematoxylin for 2 minutes.

Positive control paraffin-embedded sections of term placenta were de-paraffinised and rehydrated. Antigens were retrieved by heating in a microwave in citrate buffer (10 mM, pH=6.0) followed by blocking in horse serum and sequence of incubations as described above for cancer cells.

The immunoreactivity score for hCGβ immunohistochemistry was graded by two researchers independently and then agreed as negative (−ve), uncertain (+/−); weakly positive (+), strongly positive (++) and intensely positive (+++), as previously described (26).

RNA extraction and cDNA synthesis. Total RNA was isolated from all cells using SV Total RNA Isolation System (Promega, Southampton, Hampshire UK). One microgramme of total mRNA was used for cDNA first-strand synthesis by Verso cDNA Kit using oligo-dT primer and random hexamers in a 1:3 ratio (Thermo Scientific, Loughborough, Leicestershire, UK).

Quantitative real-time PCR. Relative quantification of CGB gene expression for the cells was carried out in Quantica thermocycler (Techne, Stone, Staffordshire, UK) using ABsolute™ Blue QPCR SYBR® Green Mix kit (Thermo Scientific) with ROX passive reference dye as recommended by the manufacturer using 2 μl of cDNA sample. The amplification program consisted of 1 cycle of 95°C 10 min, followed by 40 cycles of 95°C for 10 sec, annealing 60°C for 5 sec, and 72°C for 9 sec. After termination of 45 cycles, a dissociation curve analysis was performed.

Figure 1.
Figure 1.

Diagrammatic representation of the hCGβ gene cluster and comparison of different hCGβ genes. A. hCGβ-LHβ gene cluster; diagrammatic representation showing relative positions of the CGB genes, LHB and SNARs. Straight arrows show direction of transcription, and open arrowheads represent SNAR-G genes. Start codons – ATG1 and ATG2 correspond to alternative open reading frame. B: Schematic of CGB1 and -2 genes compared to CGB, CGB5, CGB7, CGB8 gene structures, illustrating exon and the intron regions of the CGB genes and indicating the potential three main splice variants reading through into intron 1 (+47 bp, +166 bp, +167 bp). Translated regions are depicted as black boxes; open reading frames are shown as open boxes. SNAR-Gs are located in the upstream region of exon 1 of genes CGB1 and -2 and are shown as a smaller black box along the gene line.

All primers were designed in NCBI/Primer3 – BLAST tool and purchased from Sigma (Sigma Life Science, Gillingham, Dorset, UK). One set of primers were specific for CGB1 and -2: forward primer (5’-CGTCCAACACCCCTCACTCC-3’) designed to anneal to nucleotides 118-137 in mRNAs of human CGB1 and mRNA of human CGB2 (GenBank accession no. NM_033377 and NM_033378, respectively), and reverse-primer (5’-GGCAGCCCTCCTTCTCCAC-3’) to nucleotides 329-347 of the mRNAs. This set of primers span intron 1 and consequently all major known spliced variants (Figure 1) would be detected. The second set of primers (F 5’-CAGCACCTTTCTCGGGTCAC-3’ and R 5’- CAGGGAGTAGGGTGTAGGAAGG-3’) were specific for classical CGB gene mRNAs and designed to anneal to nucleotides 10-30 and nucleotides 73-95 (GenBank accession no. NM_033183). In order to check the integrity of the reverse transcriptase reaction, the expression of the housekeeping gene encoding glyceraldehyde-3-phosphate dehydrogenase (GAPDH) amplified using a specific primer set (F 5’-CATGGGTGTGAACCATGAGAAG- 3’ and R 5’-GTGCTAAGCA GTTGGTGGTGC-3’) was used as an endogenous control (GenBank accession no. NM_002046) in all reactions. Non-emplate control (NTC) contained complete qPCR master mix, but no cDNA.

Separation and sequencing of PCR products. Real-time PCR products were detected by Microchip Electrophoresis System MCE-202 MultiNA (Shimadzu, Milton Keynes, Buckinghamshire, UK) according to manufacturer instructions for DNA-500 kit (Shimadzu). The precise size estimation length of DNA products was based on separation of 25-bp DNA ladder (Invitrogen) on MultiNA electrochip system. As confirmation, qPCR products were also separated on 1% agarose gel, excised and purified with PureLink™ Quick Gel Extraction Kit (Invitrogen), and sequenced by GATC Biotech (London, UK).

Relative quantification of the level of CGB mRNA transcripts. For the calculation of relative quantity of CGB genes, we applied the ΔΔCp method described by Livak and Schmittgen (28). The level of transcription of CGB1 and -2 and classical CGB genes was normalised against housekeeping gene GAPDH levels and calculated to relative gene expression in placental tissue (calibrator). Mean crossing point values (from triplicate reactions) were used in relative quantification. Final results are expressed as N-fold differences in CGB expression relative to calibrator gene expression level equal to 1.0.

siRNA silencing of CGB gene expression. The siRNA duplexes (Sigma-Aldrich, Gillingham, Dorset, UK) were designed primarily to target mRNA arising from the published sequences for CGB2 (GenBank accession no. NM_033378) and CGB8 (GenBank accession no. NM_033183). These duplex sets were distinct in target but CGB2 siRNA duplexes cross-hybridized significantly (100%) with CGB1, and the CGB8 siRNA duplexes significantly cross-hybridized (>95%) with all classical CGB genes (Table I). Bladder cancer cell line SCaBER cells were transfected with siRNA duplexes using CodeBreaker™siRNA Transfection Reagent (Promega, UK) according to the manufacturer's protocol. A total of 5×103 cells per well were plated into 100 μl complete growth medium in 96-well plates (BD Falcon, Oxford, Oxfordshire, UK) and allowed to grow for 24 h (until they were 50% confluent). Cells were then transfected with siRNA at a final concentration of 15 nM, 20 nM and 25 nM of siRNA in sets of six replicates. Plates were incubated for a further 72 h at 37°C and 5% CO2. A non-specific siRNA targeting enhanced green fluorescent protein (EGFP) was included as a transfection response control. Negative controls were cells which had been treated with the transfection reagents (CodeBreaker™siRNA; Promega), but without any siRNA.

Determination of secreted hCG and hCGβ protein by cancer cells. The media from confluent cell cultures grown in 75 cm2 flasks from before and after specific CGB gene silencing were harvested for hCGβ protein quantification. hCGβ concentration was estimated using a two-site FBT-11 immunoenzymetric assay, described previously (27, 29-32). Cell numbers post-culture, were estimated using a haemocytometer and used to normalise protein concentrations (1×106 cells). The assay was calibrated against a standard curve of recombinant hCGβ (Sigma) (range of 50 pg/l to 0.5 pg/l) which had been calibrated against the first international reference preparation of hCGβ (batch 75/551; NIBSC, Potters Bar, UK). Intact hCG concentrations were quantified using the USA hCG Reference Service in-house intact hCG ELISA, utilising the antibody combination of McAb2119 with 4001-POD (33). This was calibrated against the 3rd International Standard preparation of hCG (batch 75/589; NIBSC). The limit of detection (lowest standard) for the intact hCG assay was 0.75 pg/l.

Cell proliferation assay following siRNA transfection. For viability assays, cells were seeded in complete growth medium, as described above. Growth medium was replaced with 20 μl of CellTiter 96® AQueous® One Solution Cell Proliferation Assay reagent (Promega) and cells were incubated at 37°C for 1-4 h until colour was well-developed in a humidified atmosphere with 95% air, 5% CO2. 3-[4,5,dimethylthiazol-2-yl]-5-[3-carboxymethoxy-phenyl]-2-[4-sulfophenyl]-2H-tetrazolium (MTS) is a tetrazolium salt reduction-type assay method - absorbance at 490 nm was measured on a Fluostar OPTIMA (BMG Labtech, Aylesbury, Buckinghamshire, UK), and optical density was measured and expressed as percentage change relative to control cultures in 96-well plates. Data were normalised against the optical density achieved for the untreated controls, and expressed as a percentage change in cell number. Data were then pooled between plates and expressed as mean and standard deviation change in normalised cell populations from quintuplicate experiments. Statistical significance was conducted using Stats Direct™ (Altrincham, Cheshire, UK) software in ANOVA, or Friedman's two-way analysis (data were not normally distributed).

Results

Detection and sequencing of CGB gene transcripts. The CGB genes were detected by two separate sets of primers that differentiated the CGB genes 1/2 and the classical CGB genes 3, 5, 7 and 8. In all cancer cell lines tested (and placenta control), transcripts for classical CGB genes were detected (Figure 2A – product 81 bp) and sequences matched those predicted in the GenBank database.

The transcripts of CGB1 and -2 genes were found only in cell lines SCaBER, RT112, C2235 and placenta (Figure 2B – product 231 bp and 396 bp). All CGB1 and -2 products were sequenced, and aligned with the known CGB 1/2 sequences in the GenBank database. In regard to CGB 1/2 transcripts, sequence analysis identified that the products arose largely from the expression of CGB2 alone (Figure 3).

No product appeared to be derived from genomic DNA, as no high molecular mass sequences (containing introns) were observed.

Relative quantification of the level of CGB gene expression. CGB gene expression in the cancer cell lines was quantified and compared to the expression level of CGB genes in term placenta. Classical CGB genes were expressed at the highest level in the placental tissue and to a much lesser extent by cancer cell lines; in some cases less than 1% of that seen in term placenta. The exceptions were by the bladder cancer cell line T24 at 1.6%, SCaBER at 3.1%, and by the two breast cancer cell lines C2335 at 6.3% and C2238 at 25.0% of that seen in the placental control tissue (Table II).

CGB1 and -2 mRNA was detected in two out of three bladder cancer cell lines (SCaBER and RT112), where the expression level was significantly higher in comparison to term placenta, both 128-fold higher. The breast cancer cell line C2335 demonstrated a 32-fold higher expression of CGB1/2 mRNA compared to the placental control sample (Table II).

Detection of hCGβ protein. Immunostaining for hCGβ scored positively in most of the tested non-trophoblastic cancer cell lines, albeit at different levels. Significantly, bladder cancer cell line SCaBER and breast cancer cell line C2235 exhibited very weakly-positive cellular staining. The remaining cell lines stained more positively, as did the placental positive control (Table II). However, the presence and quantity of free hCGβ protein secreted into the culture media (as measured by specific hCGβ ELISA) was significant in only three of the cancer cell lines: The highest level of hCGβ protein in media was seen in the bladder cancer cell line SCaBER (4.4 ng/106 cells/24 h), breast cancer cell line C2235 (2.3 ng/106 cells/24 h) and bladder cancer cell line RT112 (0.78 ng/106 cells/24 h). In the remaining cancer cell lines (T24, PC-3, LN-CAP, C2238), no hCGβ was detected in the media (<0.5 ng/106 cells/24 h). No intact (αβ heterodimer) hCG of pregnancy was detected in any culture medium (Table II).

Effect of siRNA duplex silencing. Following siRNA control transfection, secreted hCGβ protein levels ranged from 3-4.6 pg/l for bladder cancer cells. siRNA CGB2 oligo- duplex (25 nM) reduced secretion by 96%, while CGB8 oligo-duplex reduced secretion by 42% of controls. At all concentrations, we observed that siRNA duplexes effectively inhibited CGB expression (Figure 4A). We further examined whether viability of cells in culture could be changed after silencing to determine the impact of hCGβ secretion on survival. At all concentrations of siRNA, CGB2 oligo-duplex reduced cell numbers significantly: 25 nM by 44% (p<0.0001), 20 nM by 24% (p=0.0011), and 15 nM by 19% (p=0.02) compared to controls (Figure 4B). In contrast, the control siRNA directed against the sequence of an unrelated protein (EGFP) failed to produce any significant reduction in relative cell numbers, with reduction of only 13% (p=0.22). When CGB8 gene was targeted by siRNA oligo-duplex, only the highest concentration (25 nM) resulted in a significant reduction of cell growth by 23% compared to control (p=0.0004) (Figure 4B). Non-specific silencing of control EGFP sequence had no effect on cell numbers (1.01%, p=0.42).

Discussion

The overexpression and secretion of ectopic hCGβ by epithelial cancer cells is well-reviewed (1, 34-37) such that vaccines against hCGβ have been developed and discussed as adjuvant treatments in cancer (1, 38-40). However, despite the wealth of evidence in the field, the molecular mechanisms are not well-defined and occasionally result in discordant data. Previous studies have failed to correlate CGB mRNA expression with protein expression as detected by immunohistochemistry and immunoassays (8).

View this table:
Table I.

Detail of siRNA sequences used for silencing CGB genes. siRNA sequences used for silencing CGB2 (and 1) and CGB8 (and CGB, CGB5, -7 and -8); their location on mRNA mapping to base-pairs on exon 1 (according to GenBank accession number NM_033378 for CGB2 and NM_033183 for CGB8, respectively) and the degree of homology between the siRNA duplexes and the different CGB gene sequences.

Several molecular studies have been conducted on placental tissues, choriocarcinoma and epithelial cancer (particularly bladder) to try and establish a model for CGB gene expression. These are not necessarily comparable as the placenta expresses both α and β subunits to form and secrete the pregnancy hormone intact hCG; choriocarcinomas appear to secrete intact hCG and excess free subunits, and epithelial carcinomas express and variably secrete only the free β-subunit hCG.

Due to the high level of homology between the CGB gene cluster sequences, it is difficult to identify base pair sequences that are sufficiently specific enough to differentiate individual CGB genes. In all earlier studies, specific CGB1 and -2 gene expression was excluded from primer design in the belief that these were non-functioning genes (pseudogenes) simply because no functional hCG product has been identified at the time (10). In placental and choriocarcinoma tissue, all CGB gene expression was based solely upon mRNA detection (11) and cancer studies have shown that CGB mRNA is detectable in both normal and carcinoma tissue. Attempts to differentiate CGB expression revealed that only CGB7 (then termed CGB type I) was transcriptionally active in both tumor and normal tissue and the remaining classical genes (termed type II) were indicative of cancer (22, 24, 41). In our study, it was not assumed that CGB1 and -2 genes were transcriptionally silent and we designed specific primers for CGB1 and -2 amplification which spanned intron 1 to allow detection and identification of theoretical spliced variants (11, 16-18).

Critically, our study clearly shows the expression of CGB1 and -2 genes by non-trophoblastic cancer cell lines and at mRNA levels vastly greater than those seen in the placental control tissue. Furthermore, sequencing reveals that this is largely due to the expression of CGB2 rather than that of CGB1. CGB1 and CGB2 only vary by two nucleotides at positions 185 and 223 (accession numbers NM_033377.1 and NM_033378.1, respectively). The predominant nucleotides detected at these sequence positions were C and G, indicative of CGB2 (Figure 3).

The molecular basis of earlier studies to distinguish between CGB gene expression was that type I gene (CGB7) encodes an alanine at position 117, whilst type II encode an aspartic acid (24), and it is this which is associated with malignant transformation of epithelial cells. At first, this appears to be inconsistent with our findings, but it is entirely plausible that the technology employed previously (22, 41) was incomplete. Indeed, the competitive oligonucleotide priming (COP) assay employed (which relies on differential binding on a single base pair mismatch) would have hybridised with CGB1 and -2 mRNA. Examining GenBank CGB gene sequences (Table III), it was found that mRNA arising from CGB1 has the GAC codon (ASP) and CGB2 gene has GCC codon (ALA) at amino acid position 117. Thus, of mRNA discarded, that arising from CGB1 would have registered as type I and that arising from CGB2 as type II. Here we found that abundant mRNA from CGB2 correlated with expressed and secreted free hCGβ protein and confounds the COP assay.

Figure 2.
Figure 2.

Electrophoresis and elecropherogram of qPCR products on Microchip Electrophoresis System MultiNA. Separation of real-time PCR products and analysis by electrophoresis system MultiNA. A: Amplification of classical CGB genes by quantitative PCR. The desired product is seen in appropriate lanes: placenta, SCaBER, RT112, T24, C2235 and C2238. B: PCR products obtained during amplification of genes CGB1 and -2. The product 231 bp corresponds to calculated specific product of CGB1 and -2, whilst the primers are the bands below 25 bp. In addition +166 bp splicing variant was amplified (with calculated by MultiNa software mass 396 bp). The position of upper (UM) and lower (LM) marker dyes of the MultiNA separation are indicated, as is the position of the base pair reference mass markers. The ladder (Invitrogen) shows the bands between 25 bp and 450 bp. Any slight difference in size of product is caused by variations in separation on different microchips. Non-template control (NTC) and negative control (Neg ctrl) are indicated. Representative electropherograms from MultiNA electrochip system of separated PCR products amplified (A) with CGB, CGB5,-7 and - 8 primers with a mass of 81 bp (B) and with CGB1 and -2-specific primers with predominant product with a mass of 231 bp (calculated 229 bp), and less abundant 396 bp. LM and UM were used for calibration according to the manufacturer. The axes of the plot indicate migration index (x) and fluorescence intensity (y). Each peak in the electropherogram represents a DNA fragment of PCR product or primer which migrates according to molecular weight. Migration index is inversely proportional to the length of DNA fragments and is used by software to calculate exact mass.

Figure 3.
Figure 3.

Sequencing analysis of CGB1 and -2 qPCR products. Sections of specific sequencing of CGB1 and -2 PCR products (both 231 bp and 396 bp) were sequenced indicating the single marker nucleotide differences between CGB1 and CGB2 mRNA (at nucleotides 185 and 223 - accession numbers NM_033377.1 and NM_033378.1 respectively). This demonstrates that expression of CGB2 mRNA is the predominant source of CGB1/2 RT-PCR product detected in qPCR with CGB1/2-specific primers.

A more recent study, using similar real-time PCR to our study, claimed they were unable to detect any increase in gene CGB1 and -2 expression (23); it should be noted that they exclusively studied normal breast and breast tumour samples. In this study, we also found contradictory patterns of CGB gene expression in the breast cancer cell lines: C2238 did not secrete hCGβ into culture media and did not express CGB1 and -2 genes, but did stain strongly for hCGβ on immunohistochemistry and expressed relatively high levels (25% of that of third trimester pregnancy) of classical CGB genes. C2235 secreted high levels of hCGβ into culture media, expressed high levels of CGB1 and -2 genes, but stained weakly for hCGβ on immunohistochemistry and expressed relatively low levels (6% of that of third trimester pregnancy) of classical CGB genes.

View this table:
Table II.

Relative quantification of CGB gene expression by cancer cell lines and placental tissue compared with detection of hCGβ by immunohistochemistry, and hCGβ and intact hCG secretion into culture media by specific ELISA. Average crossing points (Cp) were normalised to the GAPDH level in order to correct for variation in total cDNA concentration between samples. Normalised expression was compared as relative to that detected in the placental tissue extract which was adopted as an internal calibrator. ND: Not detected, gene expression level was, therefore, not quantified. The immunoreactivity score was graded using a semi-quantitative scoring system which scored both the percentage of positively stained cells and the intensity of immunopositivity as negative (−ve), uncertain (+/−), weakly positive (+), strongly positive (++) and intensely positive (+++). Scoring was based on positivity seen in control placental tissue. The level of hCGβ protein secreted into the culture media by the tumor cell lines (106 cells cultured for 24 h) was determined by specific immunoassay for free hCGβ and intact hCG.

View this table:
Table III.

Alignment of CGB gene mRNA corresponding to exon 3 coding for amino acid 117. Sequence alignment shows that the triplet codon variant GCC (alanine) previously considered unique to hCGβ protein arising from transcription of CGB7 is also found in CGB2. The location of the mutation corresponds to amino acid position 117 usually aspartic acid (GAC). The mRNA GenBank accession number and base pairs position of sequences are indicated in Bold.

Most significantly, in this study, such CGB2 gene expression was only found where hCGβ protein was detected as a secreted (culture media) product. However, the multiple splice variants arising from CGB1 and -2 genes had been presumed by others to translate into a non hCGβ-like protein. This proposed protein was purported to initiate translation from an alternative ATG start site (ATG1) located 47 nucleotides upstream from the conserved classical CGB mRNA start site (Figure 1). Interestingly ATG1 is not present in the other CGB genes due to splicing. Moreover, putative 155- and 163-amino-acid proteins (translated from an ATG2 ORF within genes CGB1 and -2 respectively) are of a comparative size to ‘classical’ CGB gene transcripts and share up to 98% sequence homology. In the present study of CGB1 and -2 gene expression by epithelial cancer, we found that the predominant mRNA spliced variant arising from CGB2 is a fully-spliced 231-bp product (shown in the electropherogram, Figure 2B), which sequenced as predicted, and would therefore have been available for translation to a mature hCGβ-like protein.

Figure 4.
Figure 4.

Effect of silencing of CGB1 and -2 and classical CGB genes on bladder cancer cell line SCaBER. A: Secretion of hCGβ into culture media of SCaBER cancer cells treated with increasing concentrations of siRNA targeting CGB1 (—▪—) and -2 or CGB, CGB5, -7 and -8 (—●—). Values are expressed as percentage of untreated control and each point is the mean of six replicates of the assay ran in duplicate; the bars represent standard deviation about the mean. EGFP silencing represents non-specific effects of introducing a non-functioning siRNA.B: Influence of RNA silencing on SCaBER cell viability – increasing concentration of siRNA (nmol/l) and relative level of viable cells are shown. MTS assay values are expressed as a percentage that of untreated control. Results are the mean from six replicates of the assay. The bars represent standard deviation about the means. EGFP silencing represents non-specific effects of introducing a non- functioning siRNA.

It has been proposed that the expression of CGB type II genes significantly increases in many non-trophoblastic malignant tissues and may correlate with elevated levels of hCGβ in patients' serum. The results obtained by our study showed expression of all classical CGB genes by all non-trophoblastic cell lines. However, this does not correlate with secreted protein levels. Immunohistochemical detection indicates that protein expressed might be held intracellularly for the majority of cancer cell lines, and not secreted. A possible explanation may be that transcription and translation of the CGB gene cluster does not result in secretion per se, and that more specific gene or cellular processing may be required for active autocrine/paracrine expression. The levels of mRNA arising from classical CGB genes in our cell lines are generally less that 1% of that found in third trimester placental tissue. Interestingly, in one breast cancer cell line (C2238), mRNA levels were as high as 25% of that seen in placental tissue, yet no secreted hCGβ protein was detected.

As already stated, transcriptional messages were detected in all cell lines using primers designed to amplify classical CGB genes, including the non-hCGβ protein-secreting bladder cancer cell line T24. However, mRNA arising from CGB1 and -2 gene was only detected in placental tissue, breast cancer cell line C2235, and bladder cancer cell lines RT112 and SCaBER (Figure 2B). Furthermore, these are the only cell lines studied which secrete hCGβ protein into their culture media (Table II). The high level of CGB1 and -2 gene expression in these cancer cell lines may be evidence of their functionality, and this is supported by the interesting correlation with protein secretion. This complex condition of expression may explain ectopic hCGβ secretion by tumors when simple ‘all CGB’ mRNA detection in a tumor fails to correlate with circulating levels found in patients' blood and urine, as reviewed by Butler and Iles (1), and explain why molecular and immunological hCGβ studies often differ. In much the same way as antibody selection has become important in immunoassays detecting the right ‘kind’ of hCG for the management of trophoblastic diseases (42), we might find ourselves having to use the right kind of molecular markers to improve the diagnostic and prognostic power of hCGβ in epithelial cancer.

More recently it has been shown that CGB1 and -2 is expressed in the placenta (14), testes (13, 43) and in ovarian tissue (44). Moreover, in testis, the predominant CGB1 and -2 mRNA transcript is also a fully-spliced mRNA with translation initiation from AUG2. This was also proposed to give rise to an hCGβ protein homologous to the protein coded by classical CGB genes. Similarly, we have now demonstrated that the major transcribed splice variant of CGB2 mRNA is fully-spliced and would result in a recognisable hCGβ peptide in agreement with these findings (13, 43). A +166 splice variant was also detected arising from CGB2 gene expression, but at lower levels in the hCGβ-secreting cancer cell lines. It has been suggested that these are CGB2a and CGB2b proteins (13, 43) and that the splice variant +166 bp encodes a novel peptide using a different open reading frame starting from ATG1. However, the function of this peptide remains unclear.

To study the possible effect of any hCGβ protein derived from expression of CGB1 and -2, we inhibited its translation by specifically targeting the mRNA with a mix of siRNA duplexes for different regions within exon 1 of CGB1 and -2, and CGB8 (with high homology to other classical CGB genes – Table I). Silencing of CGB genes with siRNA duplexes resulted in the reduction of the viable cancer cell number (Figure 3B). However, at all concentrations of siRNA, CGB2 oligo-duplex cell numbers were reduced significantly more than if siRNA targeted classical CGB mRNA. This was also reflected in the amount of hCGβ protein secreted by the exposed cells, with CGB2 oligo-duplex siRNA effectively abolishing secretion (Figure 3A).

Non-specific silencing of CGB gene expression was shown to induce increased apoptosis of HeLa cells (20) and inhibit growth of SCaBER cells in vitro (21). The significant reduction of gene expression was demonstrated using a modified U1 snRNA in which the first 10 nucleotides were substituted with a complementary CGB gene sequence present in exon 3 (20). However, in neither of these studies was any distinction made between which of the six CGB genes were silenced. Moreover, silencing of any gene coding beta subunit of human chorionic gonadotropin could lead to reduction not only of free hCGβ, but also of heterodimeric hCG known to be produced by HeLa cells (45-47). Here by distinguishing the target region more specifically’ we were able to examine the respective influence of gene groups on cancer cell viability.

Specific targeting of CGB1 and -2 is much more effective in reducing cancer cell numbers than silencing classical CGB genes, or general silencing of ‘all’ CGB, and although it was once considered to be a pseudogene, CGB2 may now be considered an active gene which is important for cancer growth. Furthermore, CGB2 may interact at the mRNA level with other gene transcripts such as the snRNA, SNAR-G2 which is encoded adjacent to CGB2 (13, 43, 48) (Figure 1). The profile of CGB1 and -2 gene expression during the first trimester correlates with successful pregnancies (14, 16) and particularly elevated levels of CGB1/2 mRNA were reported in ectopic and molar pregnancies (12). It can be postulated that these pregnancy implantation associations, together with data present here, indicate a role for CGB1 and -2 (possibly only CGB2)-derived free hCGβ in the invasive process (elevated free hCGβ is pronounced in early pregnancy).

In conclusion, up-regulation of CGB2 gene expression in common epithelial cancer cells suggests an important function for this presumed pseudogene in ectopic hCGβ secretion and oncogenesis. This finding could play a significant role in cancer diagnostics by differentiating aggressive from non-aggressive tumors.

Acknowledgements

This study was supported by the ELK-Foundation, Peter Luck-Hille and The Lee Smith Foundation, UK.

Footnotes

  • ↵* These Authors contributed equally to this work.

  • This article is freely accessible online.

  • Received March 4, 2014.
  • Revision received March 23, 2014.
  • Accepted March 24, 2014.

References

  1. ↵
    1. Cole L,
    2. Butler SA
    1. Butler SA,
    2. Iles RK
    : Biological function of the free beta subunit: Expression and treatment target in cancer. In: Human Chorionic Gonadotropin (hCG). Cole L, Butler SA (eds.). London, Elsevier Health Sciences: pp. 153-171, 2010.
  2. ↵
    1. Iles RK,
    2. Delves PJ,
    3. Butler SA
    : Does hCG or hCGβ play a role in cancer cell biology? Mol Cell Endocrinol 329: 62-70, 2010.
    OpenUrlCrossRefPubMed
  3. ↵
    1. Gillott DJ,
    2. Iles RK,
    3. Chard T
    : The effects of human chorionic gonadotropin on the in vitro growth of bladder cancer cell lines. Br J Cancer 73: 323-326, 1996.
    OpenUrlCrossRefPubMed
  4. ↵
    1. Butler SA,
    2. Ikram MS,
    3. Mathieu S,
    4. Iles RK
    : The increase in bladder carcinoma cell population induced by the free beta subunit of human chorionic gonadotrophin is a result of an anti-apoptosis effect and not cell proliferation. Br J Cancer 82: 1553-1556, 2000.
    OpenUrlCrossRefPubMed
  5. ↵
    1. Butler SA,
    2. Laidler P,
    3. Porter JR,
    4. Kicman AT,
    5. Chard T,
    6. Cowan DA,
    7. Iles RK
    : The beta subunit of human chorionic gonadotrophin exists as a homodimer. Journal of Molecular Endocrinology 22: 185-192, 1999.
    OpenUrlAbstract
  6. ↵
    1. Butler SA,
    2. Iles RK
    : Monomeric human chorionic gonadotrophin beta is responsible for autocrine growth effects. Tumor Biology 25: 18-23, 2004.
    OpenUrlCrossRefPubMed
  7. ↵
    1. Iles RK
    : Ectopic hCGbeta expression by epithelial cancer: malignant behaviour, metastasis and inhibition of tumor cell apoptosis. Mol Cell Endocrinol 260-262: 264-270, 2007.
    OpenUrlCrossRefPubMed
  8. ↵
    1. Davies S
    : Ectopic Expression of Glycoprotein Hormones and Their Receptors by Urogenital Cancers. Ph.D. thesis, University of London. 2001.
  9. ↵
    1. Talmadge K,
    2. Boorstein WR,
    3. Vamvakopoulos NC,
    4. Gething MJ,
    5. Fiddes JC
    : Only three of the seven human chorionic gonadotropin beta subunit genes can be expressed in the placenta. Nucleic Acids Res 12: 8415-8436, 1984.
    OpenUrlCrossRefPubMed
  10. ↵
    1. Boorstein WR,
    2. Vamvakopoulos NC,
    3. Fiddes JC
    : Human chorionic gonadotropin beta subunit is encoded by at least eight genes arranged in tandem and inverted pairs. Nature 300: 419-422, 1982.
    OpenUrlCrossRefPubMed
  11. ↵
    1. Bo M,
    2. Boime I
    : Identification of the transcriptionally active genes of the chorionic gonadotropin beta gene cluster in vivo. J Biol Chem 267: 3179-3184, 1992.
    OpenUrlAbstract/FREE Full Text
  12. ↵
    1. Rull K,
    2. Hallast P,
    3. Uusku L,
    4. Jackson J,
    5. Punab M,
    6. Salumets A,
    7. Campbell RK,
    8. Laan M
    : Fine-scale quantification of HCG beta gene transcription in human trophoblastic and non-malignant non-trophoblastic tissues. MHR-Basic Science of Reproductive Medicine 14: 23-31, 2008.
    OpenUrl
  13. ↵
    1. Parrott AM,
    2. Sriram G,
    3. Liu Y,
    4. Mathews MB
    : Expression of type II Chorionic Gonadotropin genes supports a role in the male reproductive system. Molecular and Cellular Biology 31: 287-299, 2011.
    OpenUrlAbstract/FREE Full Text
  14. ↵
    1. Hallast P,
    2. Rull K,
    3. Laan M
    : The evolution and genomic landscape of CGB1 and CGB2 genes. Mol Cell Endocrinol 260,262: 2-11, 2007.
    OpenUrl
  15. ↵
    1. Hollenberg AN,
    2. Pestell RG,
    3. Albanese C,
    4. Boers ME,
    5. Jameson JL
    : Multiple promoter elements in the human chorionic gonadotropin beta subunit genes distinguish their expression from the luteinizing hormone beta gene. Mol Cell Endocrinol 106: 111-119, 1994.
    OpenUrlCrossRefPubMed
  16. ↵
    1. Rull K,
    2. Laan M
    : Expression of beta-subunit of HCG genes during normal and failed pregnancy. Hum Reprod 20: 3360-3368, 2005.
    OpenUrlCrossRefPubMed
  17. ↵
    1. Berger P,
    2. Kranewitter W,
    3. Madersbacher S,
    4. Gerth R,
    5. Geley S,
    6. Dirnhofer S
    : Eutopic production of human chorionic gonadotropin β (hCGβ) and luteinizing hormone β (hLHβ) in the human testis. FEBS Lett 343: 229-233, 1994.
    OpenUrlCrossRefPubMed
  18. ↵
    1. Dirnhofer S,
    2. Hermann M,
    3. Hittmair R,
    4. Hoermann K,
    5. Kapelari K,
    6. Berger P
    : Expression of the human chorionic gonadotropin-beta gene cluster in human pituitaries and alternate use of exon 1. J Clin Endocrinol Metab 81: 4212-4217, 1996.
    OpenUrlCrossRefPubMed
  19. ↵
    1. Strauss BL,
    2. Pittman R,
    3. Pixley M,
    4. Nilson JH,
    5. Boime HI
    : Expression of the beta subunit of chorionic gonadotropin in transgenic mice. J Biol Chem 269: 4968-4973, 1993.
    OpenUrl
  20. ↵
    1. Jankowska A,
    2. Gunderson SI,
    3. Andrusiewicz M,
    4. Burczynska B,
    5. Szczerba A,
    6. Jarmolowski A,
    7. Nowak-Markwitz E,
    8. Warchol JB
    : Reduction of human chorionic gonadotropin beta subunit expression by modified U1 snRNA caused apoptosis in cervical cancer cells. Mol Cancer 7: 26, 2008.
    OpenUrlCrossRefPubMed
  21. ↵
    1. Burczynska B,
    2. Booth MJ,
    3. Iles RK,
    4. Shah A,
    5. Shiled A,
    6. Butler SA
    : Stable knockdown of hCGβ mRNA expression in bladder cancer cells results in significant growth inhibition. Anticancer Res 33: 3611-3614, 2013.
    OpenUrlAbstract/FREE Full Text
  22. ↵
    1. Bellet D,
    2. Lazar V,
    3. Bièche I,
    4. Paradis V,
    5. Giovangrandi Y,
    6. Paterlini Y,
    7. Lidereau R,
    8. Bedossa,
    9. Bidart J,
    10. Vidaud M
    : Malignant transformation of nontrophoblastic cells is associated with the expression of chorionic gonadotropin β genes normally transcribed in trophoblastic cells. Cancer Research 57: 516-523, 1997.
    OpenUrlAbstract/FREE Full Text
  23. ↵
    1. Giovangrandi Y,
    2. Parfait B,
    3. Asheuer M,
    4. Olivi M,
    5. Lidereau R,
    6. Vidaud M,
    7. Bièche I
    : Analysis of the human CGB/LHB gene cluster in breast tumors by real-time quantitative RT-PCR assays. Cancer Lett 168: 93-100, 2001.
    OpenUrlCrossRefPubMed
  24. ↵
    1. Lazar V,
    2. Diez SG,
    3. Laurent A,
    4. Giovandgrandi Y,
    5. Radvanyi F,
    6. Chopin D,
    7. Bidart JM,
    8. Bellet D,
    9. Vidaud M
    : Expression of human chorionic gonadotropin beta subunit genes in superficial and invasive bladder carcinomas. Cancer Research 55: 3735-3738, 1995.
    OpenUrlAbstract/FREE Full Text
  25. ↵
    1. Hotakainen K,
    2. Lintula S,
    3. Jarvinen R,
    4. Paju A,
    5. Stenman J,
    6. Rintala E,
    7. μlf-Håkan,
    8. Stenman UH
    : Overexpression of human chorionic gonadotropin beta genes 3, 5 and 8 in tumor tissue and urinary cells of bladder cancer patients. Tumor Biology 28: 52-56, 2006.
    OpenUrlPubMed
  26. ↵
    1. Li D,
    2. Wen X,
    3. Ghali L,
    4. Al-Shalabi FM,
    5. Docherty SM,
    6. Purkis P,
    7. Iles RK
    : hCG beta expression by cervical squamous carcinoma–in vivo histological association with tumour invasion and apoptosis. Histopathology 53: 147-155, 2008.
    OpenUrlPubMed
  27. ↵
    1. Berger P,
    2. Sturgeon C,
    3. Bidart JM,
    4. Paus E,
    5. Gerth R,
    6. Niang M,
    7. Bristow,
    8. Birken S,
    9. Stenman UH
    : The ISOBM TD-7 Workshop on hCG and Related Molecules. Tumor Biology 23: 1-38, 2002.
    OpenUrlCrossRefPubMed
  28. ↵
    1. Livak KJ,
    2. Schmittgen TD
    : Analysis of relative gene expression data using real time quantitative PCR and the 2 ΔΔC(T) Method. Methods 25: 402-408, 2001.
    OpenUrlCrossRefPubMed
  29. ↵
    1. Bidart JM,
    2. Troalen F,
    3. Lazar V,
    4. Berger P,
    5. Marcillac I,
    6. Lhomme C,
    7. Droz JP,
    8. Bellet D
    : Monoclonal antibodies to the free beta-subunit of human chorionic gonadotropin define three distinct antigenic domains and distinguish between intact and nicked molecules. Endocrinology 131: 1832-1840, 1992.
    OpenUrlCrossRefPubMed
    1. Kardana A,
    2. Cole LA
    : Polypeptide nicks cause erroneous results in assays of human chorionic gonadotropin free β-Subunit. Clin Chem 38: 26-33, 1992.
    OpenUrlAbstract/FREE Full Text
    1. Cole LA,
    2. Kardana A
    : Discordant results in human chorionic gonadotropin assays. Clin Chem 38: 263-270, 1992.
    OpenUrlAbstract/FREE Full Text
  32. ↵
    1. Cole LA,
    2. Kardana A,
    3. Seifer DB,
    4. Bohler HC Jr..
    : Urine hCG beta-subunit core fragment, a sensitive test for ectopic pregnancy. J Clin Endocrinol Metab 78: 497-479, 1994.
    OpenUrlCrossRefPubMed
  33. ↵
    1. Cole LA,
    2. Seifer DB,
    3. Kardana A,
    4. Braunstein GD
    : Selecting human chorionic gonadotropin immunoassays: Consideration of cross-reacting molecules in first-trimester pregnancy serum and urine. Am J Obstet Gynecol 168: 1580-1586, 1993.
    OpenUrlPubMed
  34. ↵
    1. Iles RK,
    2. Butler SA
    : Human urothelial carcinomas–a typical disease of the aged: the clinical utility of chorionic gonadotrophin in patient management and future therapy. Exp Gerontol 33: 379-391, 1998.
    OpenUrlCrossRefPubMed
    1. Butler SA,
    2. Iles RK
    : Review - ectopic hCGβ secretion by epithelial tumors and hCGβ induced apoptosis in Kaposi's sarcoma – is there a connection? Clin Cancer Res 9: 4666-4673, 2003.
    OpenUrlAbstract/FREE Full Text
    1. Stenman UH,
    2. Alfthanm H,
    3. Hotakainen K
    : Human chorionic gonadotropin in cancer. Clin Biochem 37: 549-561, 2004.
    OpenUrlCrossRefPubMed
  37. ↵
    1. Lempiäinen A,
    2. Sankila A,
    3. Hotakainen K,
    4. Haglund C,
    5. Blomqvist C,
    6. Stenman UH
    : Expression of human chorionic gonadotropin in testicular germ cell tumors. Urol Oncol pii: S1078-1439, 2014.
  38. ↵
    1. Butler SA,
    2. Staite EM,
    3. Iles RK
    : Reduction of bladder cancer growth in response to hCGβ CTP37 vaccinated mouse serum Oncology Research 14: 93-100, 2003.
    OpenUrlPubMed
    1. Delves PJ,
    2. Iles RK,
    3. Roitt IM,
    4. Lund T
    : Designing a new generation of anti-hCG vaccines for cancer therapy. Mol Cell Endocrinol 260-262: 276-281, 2007.
    OpenUrl
  40. ↵
    1. Morse MA,
    2. Chapman R,
    3. Powderly J,
    4. Blackwell K,
    5. Keler T,
    6. Green J,
    7. Riggs R,
    8. He LZ,
    9. Ramakrishna V,
    10. Vitale L,
    11. Zhao B,
    12. Butler SA,
    13. Hobeika A,
    14. Osada T,
    15. Davis T,
    16. Clay T,
    17. Lyerly HK
    : Phase I study utilizing a novel antigen-presenting cell-targeted vaccine with toll-like receptor stimulation to induce immunity to self-antigens in cancer patients. Clinical Cancer Res 17: 4844-4853, 2011.
    OpenUrlAbstract/FREE Full Text
  41. ↵
    1. Bièche I,
    2. Lazar V,
    3. Noguès C,
    4. Poynard T,
    5. Giovangrandi Y,
    6. Bellet D,
    7. Lidereau R
    : Prognostic value of chorionic gonadotropin beta gene transcripts in human breast carcinoma. Clin Cancer Res 4: 671-676, 1998.
    OpenUrlAbstract
  42. ↵
    1. Cole LA,
    2. Shahabi S,
    3. Butler SA,
    4. Mitchell H,
    5. Newlands ES,
    6. Behrman HR,
    7. Verrill HL
    : Utility of commonly used commercial hCG immunoassays in the diagnosis and management of trophoblastic diseases. Clinical Chemistry 47: 308-315, 2001.
    OpenUrlAbstract/FREE Full Text
  43. ↵
    1. Parrott AM,
    2. Tsai M,
    3. Batchu P,
    4. Ryan K,
    5. Ozer HL,
    6. Tian B,
    7. Mathews MB
    : The evolution and expression of the snaR family of small non-coding RNAs. Nucleic Acid Res 39: 1485-1500, 2011.
    OpenUrlCrossRefPubMed
  44. ↵
    1. Kubiczak M,
    2. Walkowiak GP,
    3. Nowak-Markwitz E,
    4. Jankowska A
    : Human chorionic gonadotropin beta subunit genes CGB1 and CGB2 are transcriptionally active in ovarian cancer. Int J Mol Sci 14: 12650-12660, 2013.
    OpenUrlPubMed
  45. ↵
    1. Ghosh NK,
    2. Cox RP
    : Production of human chorionic gonadotropin in HeLa cell cultures. Nature 259: 416-417, 1976.
    OpenUrlCrossRefPubMed
    1. Lieblich JM,
    2. Weintraub BD,
    3. Rosen SW,
    4. Chou JY,
    5. Robinson JC
    : HeLa cells secrete the a subunit of glycoprotein tropic hormones. Nature 260: 530-532, 1976.
    OpenUrlCrossRefPubMed
  47. ↵
    1. Cox GS
    : Glycosylation of the chorionic gonadotropin a subunit synthesized by HeLa cells. Cancer Research 41: 3087-3094, 1981.
    OpenUrlAbstract/FREE Full Text
  48. ↵
    1. Parrott AM,
    2. Mathews MB
    : Novel rapidly evolving hominid RNAs bind nuclear factor 90 and display tissue-restricted distribution. Nucleic Acid Res 35: 6249-6258, 2007.
    OpenUrlCrossRefPubMed
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools
Reprints and Permissions
Novel Insights into the Expression of CGB1 & 2 Genes by Epithelial Cancer Cell Lines Secreting Ectopic Free hCGβ
BEATA B. BURCZYNSKA, LATIFA KOBROULY, STEPHEN A. BUTLER, MAHMOUD NAASE, RAY K. ILES
Anticancer Research May 2014, 34 (5) 2239-2248;
Twitter logo Facebook logo Mendeley logo

Jump to section

Keywords