Functional studies to identify the potential role of a chromosome 3p14-21 gene, protein tyrosine phosphatase receptor type G (PTPRG), were performed. PTPRG was identified as a candidate tumor suppressor gene (TSG) in nasopharyngeal carcinoma (NPC) by differential gene profiling of tumorigenic and nontumorigenic NPC chromosome 3 microcell hybrids (MCH). Down-regulation of this gene was found in tumor segregants when compared with their corresponding tumor-suppressive MCHs, as well as in NPC cell lines and tumor biopsies. Promoter hypermethylation and loss of heterozygosity were found to be important mechanisms contributing to PTPRG silencing. PTPRG overexpression in NPC cell lines induces growth suppression and reduced anchorage-independent growth in vitro. This is the first study to use a tetracycline-responsive vector expression system to study PTPRG stable transfectants. Results indicate its ability to induce significant tumor growth suppression in nude mice under conditions activating transgene expression. These studies now provide functional evidence indicating critical interactions of PTPRG in the extracellular matrix milieu induce cell arrest and changes in cell cycle status. This is associated with inhibition of pRB phosphorylation through down-regulation of cyclin D1. These novel findings enhance our current understanding of how PTPRG may contribute to tumorigenesis. [Cancer Res 2008;68(19):8137–45]
- nasopharyngeal carcinoma
- tumor suppressor gene
- microcell-mediated chromosome transfer
- microarray, pRB, cyclin D1
Previous studies using microcell-mediated chromosome transfer approaches identified chromosome 3p21.3 as a critical region for tumor suppression in nasopharyngeal carcinoma (NPC; ref. 1). A panel of chromosome 3 microcell hybrid (MCH) donor cell lines was transferred into the tumorigenic NPC cell line HONE1; comparison of tumorigenic potential and presence or absence of chromosomal regions defined the critical region for tumor-suppressive activity.
In this study, differential expression of chromosome 3 MCHs and their corresponding tumor segregants (TS) was analyzed by oligonucleotide microarray hybridization. Genes up-regulated in tumor-suppressing MCHs and down-regulated in tumor revertants are presumptive tumor suppressor genes (TSG; ref. 2). One interesting gene, protein tyrosine phosphatase receptor type G (PTPRG), was identified as a candidate gene for this present study.
PTPRG is a member of the protein tyrosine phosphatase family. It maps to chromosome 3p14-21, a region involved in chromosomal translocations and deletions in familial renal cell carcinoma (RCC; refs. 3, 4) and breast cancer ( 5) and tumor suppression in esophageal squamous cell carcinoma ( 6). Allelic loss associated with PTPRG is frequently found in RCC and lung carcinoma cell lines ( 7); PTPRG expression is reduced in gastric cancers ( 8). Functional PTPRG studies show its ability to prolong doubling times and colony sizes of breast cancer cells ( 9). Protein phosphorylation regulators may act as TSGs to play important roles in controlling cell growth and signaling by selective dephosphorylation of key signaling proteins responsive to changes in the microenvironment. However, no direct evidence is available on PTPRG being a candidate TSG associated with NPC.
In this current study, PTPRG expression in chromosome 3 MCHs and their corresponding TSs, as well as in NPC cell lines and biopsies, was verified by quantitative reverse transcription–PCR (qRT-PCR). Tissue microarray (TMA) analysis was used to study PTPRG protein expression directly in NPC tumor tissues. Loss of heterozygosity (LOH) and promoter hypermethylation were investigated to determine the mechanisms of PTPRG gene silencing. Functional studies, including colony formation, soft agar growth, and in vivo tumor growth assays, were performed to examine its tumor-suppressive ability. The interaction between PTPRG and components of the extracellular matrix (ECM) was tested using three-dimensional Matrigel culture methods. Its potential association with cell cycle status and linkage with pRB and cyclin D1 expression were investigated.
Materials and Methods
Cell culture. All MCH, TS, and NPC cell lines, a tetracycline transactivator tTA-producing cell line (HONE1-2), and two immortalized nasopharyngeal epithelial cell lines were cultured, as previously described ( 1, 10– 13).
NPC tissue specimens. Matched normal nasopharyngeal and NPC biopsies from 32 NPC patients were collected from Queen Mary Hospital in 2006, as previously described ( 14). For TMA analysis, 155 primary NPC patient tissues were collected in the Sun Yat-Sen University Cancer Center from 1992 to 2002. Details are as described in Supplementary Data 1. Approval for this study was obtained from Medical Ethics Committee of Sun Yat-Sen University and Hospital Institutional Review Board at the University of Hong Kong.
Oligonucleotide microarray analysis. The 28K oligonucleotide microarray slides were prepared in the Genome Institute of Singapore, as previously described ( 10), and detailed in Supplementary Fig. S2.
Real-time qRT-PCR analysis. One microgram of total RNA was reverse-transcribed with M-MLV Reverse Transcriptase (U.S. Biochemical), as described in the manufacturer's manual. qRT-PCR was performed in a Step-One Plus machine using TaqMan PCR core reagent kits, PTPRG-specific and glyceraldehyde-3-phosphate dehydrogenase (GAPDH)–specific primers, and probes (Applied Biosystems).
Western blot analysis. Western blot analysis of PTPRG was performed as previously described ( 10). Primary antibody incubation was performed with PTPRG B50268 antibody (1:1,000; Stratagene), SC-102 for pRB (1:100; Santa Cruz Biotechnology), SC-753 for cyclin D1 (1:1000; Santa Cruz Biotechnology), and Ab-1 for α-tubulin (1:10,000; Calbiochem). The α-tubulin was used as a normalization control.
NPC TMA and immunohistochemical staining. TMA construction and immunohistochemical staining methodology was as previously reported ( 17) and described in Supplementary Data 1.
Methylation-specific PCR analysis. The 1,319-bp putative PTPRG promoter mapping to position 61521785-61523103 9 (NM_002841) has been described ( 18). Primers for methylated (5′-TTTTATTTTGTAGGGTGTAAGAGGC-3′ and 5′-AAATATCGATCCCAAAATTACGAA-3′) and unmethylated (5′-TTATTTTGTAGGGTGTAAGAGGTGT-3′and 5′-AAAATATCAATCCCAAAATTACAAA-3′) DNA were designed according to the MethPrimer 10 guide ( 13). Genomic DNA bisulfite modification was performed, as previously described ( 10).
5-Aza-2′-deoxycytidine treatment. HONE1 and C666 cell lines were treated with 5 μmol/L 5-aza-2′-deoxycytidine (Sigma) for 5 d, as described ( 19).
Microsatellite typing analysis. Twelve microsatellite markers were used. D3S1239 maps within the PTPRG locus at 62.03 Mb. PCR amplicons were analyzed on an ABI PRISM 3100 Genetic Analyzer using GeneScan and Genotyper software (Applied Biosystems; ref. 6).
Bacterial artificial chromosome fluorescence in situ hybridization. Bacterial artificial chromosome (BAC) fluorescence in situ hybridization (FISH) analysis was performed, as previously described ( 19, 20), and detailed in Supplementary Data 3.
Gene transfection and colony formation assay analysis. The full-length wild-type PTPRG open reading frame was amplified from a nontumorigenic immortalized nasopharyngeal epithelial cell line, NP460, and cloned into pCR3.1 (Invitrogen) and pETE-Bsd ( 13) vectors. All pCR3.1 plasmids were transfected into NPC cell lines with Lipofectamine 2000 Reagent (Invitrogen). Colony formation assays (CFA) were done, as previously described ( 10). For stable inducible PTPRG expression, the pETE-Bsd-PTPRG and vector-alone control were transfected into HONE1-2 cells, as previously described ( 16).
In vivo tumor growth analysis. The tumorigenicity of each cell line was tested by s.c. injection of 107 cells in six sites in three 4-wk-old to 8-wk-old female athymic BALB/c nu/nu mice, as described previously ( 1). Tumor growth was measured weekly. In vivo inhibition of PTPRG expression in PTPRG stable transfectants was achieved by adding 200 μg/mL doxycycline into mouse drinking water ( 16).
Soft agar assay analysis. A total of 2 × 105 cells was mixed with 2 mL of 0.4% agar in DMEM overlaid on 2 mL of 1% plating agar in DMEM and seeded on six-well culture plates. After 2 wk in culture, the colonies were counted. Sizes of 25 colonies in at least five fields were measured using an inverted light microscope (Nikon TMS) at 20× magnification ( 21).
Three-dimensional Matrigel cell culture analysis. A total of 4,000 cells was seeded on top of Matrigel (BD Biosciences) and cultured 10 to 14 d with medium changes at 2-d intervals for spheroid analysis. Colonies were counted and measured using stereomicroscopy at 50× magnification ( 21). For cell cycle and protein analysis, 5 × 105 cells were grown under three conditions: Matrigel for 2 d, normal cell plates for 2 d, and Matrigel for 2 d followed by 2 d reculture on normal cell culture plates. The cells were then harvested for protein, RNA, and cell cycle analysis.
Cell cycle analysis. The cell cycle distribution of each cell line was analyzed by FACScan (Becton Dickinson) flow cytometry ( 16).
Statistical analysis. The χ2 and Fisher's exact tests were used to analyze significant differences of PTPRG gene expression observed in qRT-PCR and TMA analysis. All in vitro assay results represent the arithmetic mean ± SE of triplicate determinations of at least two independent experiments. Student's t test was used to determine the confidence levels in group comparisons. A P value of <0.05 was considered statistically significant.
Microarray analysis of chromosome 3 MCHs, TSs, and HONE1 cell lines. Gene expression profiles of nontumorigenic chromosome 3 MCHs and tumorigenic TSs were analyzed by competitive hybridization on 28K oligonucleotide microarrays. The chromosome 3 MCHs were hybridized against HONE1 and their matched TSs in duplicate dye swap hybridization experiments. A summary of the top 24 up-regulated candidate genes identified from this microarray analysis is shown in Table S1 in Supplementary Data 2. One gene mapping into our critical chromosome 3p region of interest, PTPRG, was differentially expressed and chosen for further in-depth studies. The expression ratios are shown in Table S2 in Supplementary Data 2. An average 1.6-fold up-regulation of PTPRG was observed in four MCHs compared with HONE1 and corresponding TS cell lines.
qRT-PCR analysis of PTPRG in chromosome 3 MCHs and TSs and NPC cell lines and biopsies. The differentially expressed PTPRG levels detected by gene profiling were validated in representative chromosome 3 tumor-suppressive MCHs and corresponding tumorigenic TSs by qRT-PCR. All four MCHs showed up-regulation of PTPRG, after transfer of an exogenous copy of chromosome 3 into HONE1; PTPRG expression was reduced in their corresponding TSs ( Fig. 1A ), showing concordance with microarray results (Table 2 in Supplementary Data 2). The expression levels of PTPRG are significantly reduced in six NPC cell lines, with the exception of CNE2, compared with the immortalized nasopharyngeal epithelial cell line, NP460 ( Fig. 1B).
The clinical significance of PTPRG in NPC was studied by determining its expression levels directly in NPC patient biopsies. Thirteen of 32 (41%) tumors show PTPRG down-regulation compared with their matched normal tissue ( Fig. 1C). There was no significant association of PTPRG down-regulation with pathologic staging, histology, sex, and age of patients (data not shown). In recurrent tumors, however, down-regulation of PTPRG was observed more frequently, but the small sample size precludes meaningful statistical analysis.
PTPRG promoter hypermethylation and LOH analysis in NPC cell lines and biopsies. PTPRG promoter hypermethylation status was analyzed by methylation-specific PCR (MSP). Methylated alleles were observed in HONE1, HK1, and C666 ( Fig. 2A ). Only unmethylated alleles were observed in HNE1, CNE1, CNE2, SUNE1, NP460, and NP69; both alleles are present in HONE1. In seven paired NPC biopsies showing PTPRG down-regulation by qRT-PCR analysis, methylated alleles were exclusively detected in tumor tissues. Unmethylated alleles were detected in both tumor and normal tissues ( Fig. 2B). Treatment with the demethylation agent, 5-aza-2′deoxycytidine, causes restoration of PTPRG expression in HONE1 and C666 cell lines ( Fig. 2C), suggesting promoter hypermethylation is an important mechanism for silencing PTPRG expression.
LOH at the PTPRG locus was investigated using 12 microsatellite markers. The marker D3S1239, mapping at 62.03 Mb, overlaps with PTPRG and shows a high frequency of nonrandom loss ( Fig. 3A ). All MCHs contain the exogenous copy of the PTPRG region, but the D3S1239 copy is lost in three of four TSs (75%). In NPC biopsies, D3S1239 also shows high LOH. Among seven pairs of tissue having sufficient DNA for analysis, three of seven (43%) show LOH. Representative microsatellite typing results for marker D3S1239 are shown ( Fig. 3B).
Continuous homozygosity in a 23.11-Mb region extending from D3S1300 to D3S1538 in the chromosome 3p12-21.1 region was detected in the HK1 and CNE1 NPC cell lines by microsatellite typing ( Fig. 3A) and suggests loss of one copy of this chromosome in that region. In contrast, the immortalized NP cell line NP69 is heterozygous. This suggests that besides promoter hypermethylation, genomic deletion is another important mechanism silencing PTPRG expression in a portion of NPC tumors.
BAC FISH analysis. The number of copies of the PTPRG region was also examined in MCH4.8 and MCH4.8-1TS cell lines by BAC FISH. MCH4.8 contains three copies of BAC RP11-154D3, whereas MCH4.8-1TS has lost one copy (Fig. S1 in Supplementary Data 3). These findings further confirm the loss in the 3p14 region in these cell lines, which is associated with PTPRG down-regulation in TSs.
Immunohistochemical detection of PTPRG expression in NPC TMA analysis. The clinical relevance of PTPRG protein expression in NPC was further analyzed by immunohistochemistry on “normal” nasopharyngeal tissues and an NPC TMA (Supplementary Data 1). Low expression of PTPRG was observed in 73 of 155 (47%) of the informative NPC cases. However, no significant associations between PTPRG expression and clinical variables were observed (data not shown).
CFA. Four NPC cell lines were transfected with pCR3.1 and pCR3.1-PTPRG. Significant reduction in neomycin-resistant colony numbers was detected in PTPRG down-regulated cell lines, HONE1, HK1, and C666, transfected with wild-type PTPRG, compared with vector-alone controls ( Fig. 4A ). CNE2 was the only NPC cell line not showing PTPRG down-regulation ( Fig. 1B). Transfection of wild-type PTPRG into this cell line, in contrast, did not affect the colony formation ability.
In vivo tumorigenicity assay. pETE-Bsd-PTPRG was transfected into the HONE1-2 cell line. PTPRG-C12 and PTPRG-C15 clones were chosen for further study after Western blot screening to evaluate PTPRG protein expression levels ( Fig. 4B). PTPRG-C12 and PTPRG-C15 both express PTPRG in the absence of doxycycline. PTPRG-C12 (+doxycycline) shows PTPRG protein levels similar to vector-alone levels. Reduced PTPRG levels in PTPRG-C15 (+doxycycline) were still higher than vector-alone controls.
In nude mouse tumorigenicity assays, large tumors formed in all six sites ( Table 1 ) injected with the vector-alone control (± doxycycline) after a 3-week to 4-week latency period. Tumorigenicity can be suppressed by overexpressing PTPRG, as observed with PTPRG-C12 and PTPRG-C15 cell lines ( Fig. 4C and Table 1). The differences in tumor growth kinetics for PTPRG-C12 and PTPRG-C15 (±doxycycline) are statistically significant (P = 0.01 and P = 0.0004, respectively).
Soft agar assay analysis. The anchorage-independent soft agar assay was performed with both tumor-suppressive PTPRG clones and the BSD-C1 control (±doxycycline). BSD-C1 vector-alone controls (±doxycycline) formed larger colonies in soft agar than observed with both PTPRG-expressing C12 and C15 clones ( Fig. 4D). Reduction of the average colony size was statistically significant (P < 0.05), when compared with BSD-C1 controls. PTPRG clones (±doxycycline) did not show statistically significant differences in colony numbers compared with controls. These results suggest that PTPRG expression in HONE1 cells suppresses its colony formation ability in vitro, as evidenced by the observable diminution in colony size.
Three-dimensional Matrigel cell culture analysis. The BSD-C1 control and PTPRG-C12 and PTPRG-C15 transfectants were grown on Matrigel, which provides a more physiologic microenvironment for cell growth. Under these conditions, spheroid numbers formed by PTPRG stable transfectants (−doxycycline), when PTPRG is expressed, are significantly decreased compared with controls ( Fig. 5A ). PTPRG-C12 (+doxycycline) forms similar spheroid numbers as BSD-C1, whereas that of PTPRG-C15 are significantly lower than controls. This was attributed to the leakiness observed in this clone for PTPRG expression ( Fig. 4B). There were no significant differences in spheroid size or morphology between the vector-alone and PTPRG transfectants.
Cell cycle analysis. To investigate possible interactions with the ECM, the PTPRG transfectants (±doxycycline) were grown on Matrigel and then analyzed by DNA flow cytometry. Cell numbers in G0-G1 phase increased in BSD-C1 from 61.7% to 78.65% and 80.1% in PTPRG-C12 and PTPRG-C15 clones in the absence of doxycycline (P < 0.05), respectively. There is a simultaneous significant decrease of cells in S phase from 14.4% to 9.0% and 10.75% in PTPRG-C12 and PTPRG-C15 clones, respectively ( Fig. 5B). There is no significant alteration of cell cycle between the BSD-C1 and PTPRG-C12 (+doxycycline). PTPRG levels still expressed by PTPRG-C15, however, sufficed to induce cell cycle arrest. These same cell lines, when grown on normal cell culture plates, do not show appreciable differences with vector-alone controls. After reculturing cells grown on Matrigel back to normal cell culture plates, no significant differences between the stable PTPRG transfectants and vector-alone control were observed. These results indicate the interaction with ECM components seems to be a prerequisite for subsequent PTPRG-induced G0-G1 phase arrest.
PTPRG activation of pRB and down-regulation of cyclin D1. As PTPRG expression is associated with cell cycle arrest at G0-G1 phase, the status of two key regulators of G1-S phase transition, pRB and cyclin D1, was investigated by Western blotting. When vector-alone and PTPRG transfectants (±doxycycline) were grown on normal cell culture plates, both phosphorylated (115 kDa) and unphosphorylated (110 kDa) forms of pRB are detected ( Fig. 5C). However, when the stable transfectants were grown on Matrigel, the active form (unphosphorylated) of pRB is predominately detected in PTPRG-C12 and PTPRG-C15 cultured in the absence of doxycycline. This is the case also for PTPRG-C15 (+doxycycline) because of continued levels of PTPRG being expressed. The phosphorylated and unphosphorylated forms of pRB are observed in the BSD-C1 control (±doxycycline) and PTPRG-C12 (+doxycycline). Furthermore, when cells grown on Matrigel were then recultured on normal cell culture plates, both forms of pRB were again detected in all clones. These interesting findings suggest an interaction of PTPRG with components of the ECM may inhibit the phosphorylation of pRB and expression levels of cyclin D1 protein. There was no significant difference seen between the clones grown on normal culture plates and those recultured in normal culture conditions after being grown on Matrigel, whereas down-regulation of cyclin D1 protein was observed in PTPRG-C12 and PTPRG-C15, when grown on Matrigel, compared with vector-alone controls ( Fig. 5C). In the presence of doxycycline, cyclin D1 is expressed at comparable levels to the control under all culture conditions. These results are consistent with the hypothesis that PTPRG suppresses tumor growth by inducing G1 arrest through down-regulation of cyclin D1 and, thus, stabilization of pRB.
RT-PCR analysis of E2F1 transcription factor. E2F1 is often involved in regulation of Rb pathways. The expression levels of E2F1 detected by RT-PCR in PTPRG-C12 and PTPRG-C15 clones were similar to BSD-C1 controls (±doxycycline), when the cells were grown on normal culture plates or reculturing on normal culture plates after prior growth on Matrigel ( Fig. 5C). However, when both PTPRG overexpressing transfectant clones (−doxycycline) were grown on Matrigel, E2F1 expression was significantly reduced. PTPRG-C12 (+doxycycline) shows similar E2F1 expression levels with BSD-C1, whereas PTPRG-C15 shows a greater E2F1 expression reduction. The corresponding reduction of E2F1 expression is consistent with observed decreases of pRB phosphorylation.
Discoveries of cancer-causing genes by powerful new unbiased genomic and transcriptome sequencing have identified candidate genes in a number of tumors, including breast and colon ( 22) and pleural mesotheliomas ( 23). In this study, we took a more targeted and functional approach using gene profiling of nontumorigenic chromosome 3 MCH and tumorigenic TS cell lines to identify interesting candidate genes having possible functional significance in NPC tumorigenesis (Table S1 in Supplementary Data 2). In this particular study, however, our focus was on one candidate, the differentially expressed gene PTPRG, which maps to the critical chromosome 3p14-21 region associated with several human tumors and in which our laboratory has had a long-term interest ( 1). All tumor-suppressive MCHs show PTPRG up-regulation when compared with parental HONE1 cells. All tumorigenic TSs showed dramatically decreased PTPRG expression. The genotyping results of MCHs and TSs show nonrandom high-frequency loss of the D3S1239 microsatellite marker mapping within the PTPRG gene, which is independently confirmed by BAC FISH. This provides an explanation for the down-regulation of PTPRG in TSs. Genomic loss in this region is associated with recurrence of tumorigenicity and a simultaneous down-regulation of PTPRG expression in chromosome 3 TS cell lines.
Recent studies report variable levels of expression of PTPRG observed by immunohistochemical staining of tissue sections in different cancers ( 24). In particular, immunohistochemical staining of PTPRG shows loss of expression in lung, breast, and ovarian cancers, whereas high-grade lymphomas and astrocytomas show PTPRG overexpression. Thus, the level of PTPRG expression is variable in different cancers with different states of differentiation. In this study, both low (47%) and high (53%) expression of PTPRG was observed in NPC tumor sections by immunohistochemical staining. More extensive studies are required to elucidate the complex role of PTPRG in tumor progression.
Loss of PTPRG expression was observed in a portion of NPC cell lines (six of seven, 86%) and tumor specimens (13 of 32, 41%). Its mechanism of inactivation is attributed to both promoter hypermethylation and LOH. Promoter hypermethylation was detected in NPC cell lines and demethylation restored PTPRG gene expression. PTPRG promoter hypermethylation was reported for T-cell lymphomas ( 18), gastric cancer ( 25), and melanoma cell lines ( 26), suggesting promoter hypermethylation is an important mechanism to silence PTPRG expression. The homozygous pattern of the chromosome 3p microsatellite markers in NPC cell lines suggests that besides promoter hypermethylation, allelic loss is another important mechanism for PTPRG inactivation. Analysis of clinical specimens further confirms the role of these two mechanisms in PTPRG down-regulation. Loss of the chromosome 3p14-21 region is commonly found in NPC ( 27– 30), renal and lung cancer cell lines ( 7), and cervical cancer ( 31). This study shows only 41% of NPC tumors having down-regulation of PTPRG by qRT-PCR, and 47% show lower PTPRG expression by immunohistochemical staining, with no statistically significant associations observed between PTPRG expression and clinicopathologic variables.
Phosphorylation and dephosphorylation of proteins provide important means for orchestrating cellular signals affecting cell growth, differentiation, cell cycle control, and apoptosis. PTPRG is a member of the protein tyrosine phosphatase family. It is down-regulated in a number of cancers, including gastric ( 8), lung, and ovarian ( 32) cancers. Mutations of PTPRG have been reported in colorectal cancer ( 33). Clarifying the functional role of PTPRG in cancer is, therefore, of importance.
Our current functional studies show that overexpression of PTPRG is sufficient to inhibit cell growth in in vitro CFAs and induce tumor growth suppression in vivo. When the gene is shut off with doxycycline and PTPRG protein expression is restored to low levels, cells revert back to their tumorigenic phenotype, as observed for PTPRG-C12. The continued growth inhibitory effect observed with PTPRG-C15 (+doxycycline) was attributed to leakiness of PTPRG expression in this clone. Importantly, these results show that even weak expression of PTPRG is capable of inducing a strong tumor-suppressive effect. Previous studies using this tet-off vector system have reported clonal variation in leakiness for TSLC1 (now named CADM1; ref. 16) and RBSP3/HYA22 ( 34). The PTPRG-C15 clone expresses higher PTPRG levels similar to that of NP460, the immortalized NPC cell line, and shows greater tumor-suppressive ability than PTPRG-C12. Thus, the in vivo nude mouse assay provides important functional evidence that expression of wild-type PTPRG can suppress the tumorigenic ability of HONE1 cells.
Significant differences in colony sizes between vector-alone controls and PTPRG stable transfectants in soft agar assays suggest PTPRG expression can reduce the transforming ability of HONE1 cells. Anchorage-independent growth is an important requirement for tumors undergoing metastasis. Similarly, when PTPRG was transfected into MCF-7 breast cancer cells, smaller and fewer colonies were formed compared with vector-alone controls ( 9).
As PTPRG is a surface protein ( 35), its interactions within an ECM environment were tested by growing cells on Matrigel to mimic the in vivo environment ( 36, 37). Under such growth conditions, the PTPRG-expressing stable transfectants form fewer spheroids than vector-alone controls, suggesting PTPRG expression within the ECM microenvironment can inhibit spheroid formation. Furthermore, PTPRG interactions with ECM components seem to induce cell cycle arrest. Significant cell cycle alteration between PTPRG-expressing clones and controls are specifically observed after growth on Matrigel, but this effect is reversible when the cells are recultured after Matrigel growth back to normal cell culture plates. Western blot analysis confirms interactions between PTPRG and ECM components inhibit pRB phosphorylation, causing cell cycle arrest. In NPC, mutation or alteration of RB expression is rare ( 38), but overexpression of cyclin D1, which can control the phosphorylation status of pRB, is common ( 39). The decrease of cyclin D1 associated with PTPRG expression provides a mechanism for loss of pRB function in NPC being associated with its phosphorylation status rather than to mutations.
Phosphorylation of pRB is a regulator of cell cycle events. The inhibition of pRB phosphorylation was further confirmed by showing a concomitant reduction of E2F1 expression in Matrigel-grown stable PTPRG-expressing clones. Unphosphorylated pRB can control the G1-S transition by interacting with E2F family proteins through direct protein-protein interaction to prevent activation of downstream targets for G1-S transition ( 40). E2F1 is a downstream target of pRB. Unphosphorylated pRB tightly binds E2F1 protein to inhibit its transcription factor function. During the G1-S transition, pRB becomes phosphorylated and releases E2F1, allowing activation of its downstream targets. As E2F1 can activate its own transcription ( 15, 41, 42), therefore, the change of E2F1 transcription levels provides further evidence for a change in phosphorylation status of pRB. The down-regulation of E2F1 observed in the PTPRG-expressing clones grown on Matrigel correlates with the finding that these clones exhibit higher levels of unphosphorylated pRB than vector-alone controls. This phenomenon seems related to PTPRG interactions occurring specifically within the ECM milieu. Cyclin D1 is involved in pRB phosphorylation during the G1-S transition. It is overexpressed and is a target oncogene in NPC ( 43). Down-regulation of the cyclin D1 oncogene suggests the tumor-suppressive ability of PTPRG is at least, in part, related to reduction of cyclin D1 levels. These results provide strong evidence, for the first time, that, in the ECM microenvironment, PTPRG exerts its functional effect to inhibit the phosphorylation of pRB to cause G0-G1 arrest.
These studies are the first to provide evidence for the importance of PTPRG in NPC. Frequent PTPRG down-regulation in NPC cell lines and in 41% of tumor biopsies is observed. Both allelic loss and promoter hypermethylation mechanisms silence PTPRG expression, although hypermethylation seems involved more often. This is the first report to identify PTPRG association with suppression of NPC cell growth and tumorigenicity. It provides clear functional evidence that PTPRG expression suppresses cell growth and transformation and induces cell cycle arrest. These novel findings show the cellular microenvironment and interactions between PTPRG and ECM elements are important in determining the cellular responsiveness to growth suppression pathways. Current findings indicate PTPRG suppression of NPC growth involves deregulation of the cyclin D/RB pathway.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Grant support: Research Grants Council of the Hong Kong Special Administrative Region, People's Republic of China grant HKUST6112/04M (M.L. Lung), and Swedish Cancer Society, Swedish Research Council, Swedish Institute, Royal Swedish Academy of Sciences, and Karolinska Institute (E.R. Zabarovsky).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
- Received March 7, 2008.
- Revision received July 8, 2008.
- Accepted July 11, 2008.
- ©2008 American Association for Cancer Research.