Background
The retinoic acid inducible G protein-coupled receptor family C group 5 member A (GPRC5A) is expressed predominantly in normal lung tissue[
1]. Several lines of evidences suggest that it function as a lung-specific tumor suppressor: a)
Gprc5a-/- mice develop spontaneous lung tumors[
2]; b) loss of heterozygosity of chromosome 12p, where GPRC5A gene (12p12.3-p13) resides, is common in NSCLCs[
3,
4]; and c), GPRC5A expression is suppressed in many lung cancer cell lines and tumor tissues compared to adjacent normal lung tissues[
2,
5]. In addition, we showed previously that lung tissues from
Gprc5a
-/-
mouse have increased NF-κB activation compared with that of wild-type mice[
6]. This suggests that Gprc5a negatively regulates a subset of genes involved in inflammation, proliferation, and survival. However, there are several controversy reports showing that elevated GPRC5A expression was found in some tumor cell lines and tumor samples and this expression correlated with increased cell growth and colony formation[
7‐
9]. One possible explanation for these opposite observations is that the tumor-suppressive function of GPRC5A can be inactivated under certain conditions. Understanding the underlying mechanisms of GPRC5A functions will yield new insights into lung tumorigenesis and permit development of novel therapeutic invention for restoring the tumor suppressive functions of GPRC5A in lung cancer.
G protein-coupled receptors (GPCRs) can be modified by glycosylation, phosphorylation and palmitoylation, which alter protein conformation, protein association, subcellular localization, and/or biological functions[
10,
11]. For example, GPCRs are desensitized via phosphorylation following agonist stimulation. This phosphorylation is directed to serine/threonine residues in the cytoplasmic tail and third cytoplasmic loop but rarely on tyrosine residues. The Ser/Thr phosphorylation by GPCR kinases (GRKs) leads to the internalization of GPCRs[
10,
11] and hampers GPCR signaling[
12]. GPCRs can also undergo Tyr-phosphorylation on residues located in the cytoplasmic domain[
13]. It has been suggested that tyrosine phosphorylation of GPCR is required for Src recruitment and subsequent Ser/Thr phosphorylation by GRK. In some GPCRs, a tyrosine containing motif in the cytoplasmic tail has been linked to the internalization of GPCRs. For example, cytokine-induced tyrosine phosphorylation of CXCR4, which reduces the level of functional CXCR4 on cell surface, contributes to GRK3 and β-arrestin2-mediated sequestration of this receptor in the cytoplasm[
14]. It remains elusive whether GPRC5A is subjected to phosphorylation, leading to altered activities in lung cells.
EGFR and its family members are the major groups of receptor tyrosine kinases that are aberrantly activated in many NSCLCs[
15]. EGFR is over-expressed in more than 60% of NSCLC cases[
16]. In addition, oncogenically-activated mutant forms of EGFR and HER2 have been found in lung cancer[
17], and contribute to the development of this disease[
18]. Moreover, EGF and TGF-α, ligands of EGFR, are frequently expressed in NSCLCs, which provides an autocrine mechanism to sustain the hyper-activation of these receptor tyrosine kinases (RTKs)[
19]. In an un-biased whole cell phospho-peptide analysis, GPRC5A was identified as one of the tyrosine phosphorylated protein in HER2-overexpressing HMEC cells after EGF or heregulin (HRG) treatment[
20,
21]. This suggests a potential cross-regulation between EGFR and GPRC5A. In this study, we showed that EGFR interacts with and phosphorylates GPRC5A in two highly conserved double-tyrosine modules (Y317/Y320 and Y347/Y350) at the C-terminal domain of GPRC5A. EGF treatment inhibited GPRC5A-mediated repression of anchorage-independent growth via phosphorylation of these tyrsoine sites since the same treatment failed to do so on GPRC5A-4 F mutant, in which tyrosine residues were replaced with phenylalanine (F). IHC analysis with specific antibody to Y317/Y320-P site showed that GPRC5A in NSCLC tissues is mostly phosphorylated, whereas GPRC5A in adjacent tumor tissues is mostly non-phosphorylated. Thus, EGFR-mediated tyrosine phosphorylation represents a newly identified mechanism by which the tumor suppressive function of GPRC5A is inactivated in lung cancer.
Discussion
In this study, we showed that EGFR interacts with and phosphorylates GPRC5A, leading to inactivation of the tumor suppressive function of GPRC5A in lung cancer cells. In this study, EGFR and GPRC5A were found to form a complex together. It remains unclear whether this interaction is directly or indirectly and whether other adaptor molecules, such as Grb-2, are involved in the interaction of EGFR-GPRC5A complex. Further investigation will reveal detailed mechanistic insight for this interaction.
Using systematic site-directed mutagenesis analysis, we identified that Y317, Y320, Y347 and Y350 are involved in the tyrosine phosphorylation of GPRC5A. In a whole-cell proteomic phospho-peptide analysis, Y317, Y320, Y347 and Y350 of GPRC5A were found to be phosphorylated in cells overexpressing EGFR and Src by mass Spectrometry[
22,
23]. These results support our finding that the two double-tyrosine modules in GPRC5A are the major residues responsible for tyrosine phosphorylation. Interestingly, Y317F exhibited much reduced tyrosine phosphorylation than Y320F. One possible explanation is that Y317 is a primary/initial site for tyrosine phosphorylation, whereas Y320 is phosphorylated subsequently. This interesting phenomenon is analogous to GSK-3β-mediated β-catenin phosphorylation, in which CK1 induced-phosphorylation on Ser45 of β-catenin facilitates/primes GSK-3β-mediated sequential phosphorylation on Thr41, Ser37, and Ser33. This reverse order of phosphorylation results in the degradation of phosphorylated β-catenin by ubiquitin–proteasome system[
24]. It is likely that phosphorylation of one residue leads to conformation change of the protein, which in turn results in exposure of other target residue for subsequent phosphorylation. In addition, different phosphorylation motif may have different biochemical or biological changes. For example, we previously demonstrated that two phosphorylation motifs on Snail regulate two different functional properties of Snail, one controls the subcellular localization whereas the other controls proteins ubiquitination and degradation of snail[
25]. In this study, we found that the major phosphorylation sites are in the first double-tyrosine module (Y317 and Y320) in the C-terminal domain of GPRC5A. It is unclear whether other tyrosine kinases are responsible for the phosphorylation of the second double-tyrosine module (Y347 and Y350) of GPRC5A.
Several reports suggest that GPRC5A could be tyrosine phosphorylated in the C-terminal domain, and different stimuli in different cell lines or tissues may induce different tyrosine phosphorylation patterns. For example, HRG-treated 184A1 HMECs cells, which overexpress HER-230-fold, showed 332 tyrosine-phosphorylation sites in 175 proteins. Interestingly, Y347 on GPRC5A (RAI3) was among these peptides[
23]. These discrepancies between theirs and our studies may be due to different tissues or experiment conditions. It is also possible that Y347/Y350 might be preferentially phosphorylated by other tyrosine kinases in different cellular context.
The biological activities of GPRC5A appeared to be regulated differently by tyrosine phosphorylation. First, GPRC5A did not affect cell proliferation, regardless of tyrosine phosphorylation status. Second, GPRC5A-mediated inhibition of cell migration is independent of tyrosine phosphorylation. These results are consistent with Kumar’s finding, in which tyrosine phosphorylation of GPRC5A showed no effect on cell proliferation and migration ability[
26]. And third, tyrosine phosphorylation did decrease or abolish GPRC5A-mediated inhibition on EGF-enhanced anchorage-independent growth of H1792 cells. Taken together, these observations strongly support out hypothesis, that EGFR-mediated tyrosine phosphorylation of GPRC5A inactivates some of the tumor suppressor activities of GPRC5A. Thus, targeting EGFR by RTK inhibitors will restore the tumor suppressor functions of GPRC5A in lung cancer cells.
Importantly, IHC analysis showed that GPRC5A in adjacent normal lung tissues is non-tyrosine- phosphorylated, whereas it is tyrosine-phosphorylated in NSCLCs. This observation strongly supports the model that tyrosine phosphorylation inhibited the tumor suppressive activities of GPRC5A. Because GPRC5A in H292G cells was tyrosine-phosphorylated either with or without EGF in vitro, we assume that GPRC5A could be either tyrosine-phosphorylated by EGFR or other receptor tyrosine kinases (RTKs). Thus, we proposed that the tyrosine-phosphorylated GPRC5A could be used as a prognostic marker for tumor progression. Thus, targeting receptor tyrosine kinases will be an effective in preventing and treating lung cancer by restoring the tumor suppressor function of GPRC5A. Our results provide a novel mechanism in supporting this strategy.
Materials and methods
Cell culture and reagents
Human NSCLC cell lines (H460, A549, H226, H157, Calu-1, H226B, H322, H292G, H1792, H358 and Sk-Mes-1) were obtained from Adi Gazdar and John Minna (UT Southwestern, Dallas, TX). Cell culture was performed as described previously[
2]. EGF was obtained from R&D system (Minneapolis, MN, USA). AG1478 was purchased from Merck Millipore (Billerica, MA, USA). EGFR mouse mAb (#2256, IP), EGFR Rabbit mAb (#4267), were from Cell Signaling Technology. Antibody to P-Tyr (PY99) was from Santa Cruz Biotechnology. Antibodies against myc and actin were from Sigma-Aldrich. HRP-conjugated secondary antibodies were from eBioScience (San Diego, CA). Rabbit polyclonal antibody to human GPRC5A was as described previously[
2]. Mouse monoclonal antibody (clone 1 L23) to the C-terminus of GPRC5A (PYKDYEVKKE) was developed and purified in Abmart (Shanghai, China)[
27]. Polyclonal rabbit anti-GPRC5A-Y317/Y320-P antibody was developed against DTLYpAPYpSTH from Abmart (Shanghai, China).
Plasmids and transfection assay
HEK293T cells were transfected using Lipofectamine 2000 transfection reagent (Invitrogen, Grand Island, NY). Transfected plasmids include: pcDNA-EGFR, pcDNA-GPRC5A-myc, and GPRC5A mutants. Cells were starved in serum-free medium for 24 hours, then treated with EGF (100 ng/ml) for various time periods as indicated.
Construction of wild-type and mutant GPRC5A plasmids
Plasmid pcDNA3.1 (+)-GPRC5A-Myc was as described[
2]. GPRC5A mutants were constructed with the following primers: Primer Y317/320 F-F: (5′-GGT TTT GAA GAG ACC GGT GAC ACG CTC TTT GCC CCC TTT TCC ACA CAT TTT C-3′) and primer Y317/320 F-R (5′-GAA AAT GTG TGG AAA AGG GGG CAA AGA GCG TGT CAC CGG TCT CTT CAA AAC C-3′). Primer Y347/350 F-F: (5′-CCA CGC TTG GCC GAG CCC TTT TAA AGA CTT TGA AGT AAA GAA AGA GG-3′) and primer Y347/350 F-R: (5′-CCT CTT TCT TTA CTT CAA AGT CTT TAA AAG GGC TCG GCC AAG CGT GG-3′). 4 F mutant (GPRC5A-Y317/320/347/350 F-myc) of GPRC5A was generated by ligation of GPRC5A-Y317/320 F-myc and GPRC5A-Y347/350 F-myc after digestion with KpnI and EcoRI separately. All sequences were verified by DNA sequencing.
Transient and stable transfection of cells
Transfection of H1792 cells with plasmids was performed by electrophoresis: Briefly, the cells were harvested, and about 3 × 105 cells were mixed with 100 μl of electroporation transfection solution (Solution V, Dharmacon, Lafayette, CO), plus plasmids, and the mixtures were transferred to electroporation cuvettes and subjected to electroporation (Amaxa Biosystems, Cologne, Germany) according to the manufacturer’s pro-grams and instructions.
Immunoprecipitation (IP) and western blotting
IP was performed with 2 μg of antibodies against myc or EGFR or normal IgG (N IgG) (as a negative control) in 1.0 mg whole cell lysate. Immunoblot[
25,
28] and cell fractionation[
29] was performed as described.
Cell migration assay
The trans-well cell migration system consisted of cell culture inserts with an 8.0 μM pore size in 24-well plate (BD BioCoat #354578, San Jose, CA). Cells were resuspended with fetal bovine serum (FBS)-free DMEM, and seeded onto the insert (4 × 104 cells). DMEM (10% FBS) with or without EGF (100 ng/ml) was loaded into the lower chamber and incubated overnight. Migrated cells, which attached to the lower side of the filter were fixed with 96% ethanol for 30 minutes and stained with 1.5% crystal violet. Migrated cells were counted in five different microscopic fields by 10 × magnifications using a Nikon fluorescence microscope.
Anchorage-independent colony formation in soft agar assay
Soft-agarose assay was performed as described previously[
2]. Aggregates of 50 or more cells were considered to be a colony. Colonies were counted in four different fields under a microscope at 4× magnification and photographed. The means and 95% confidence intervals (CIs) of the number of colonies in four microscopic fields were calculated. Two independent experiments in triplicates were performed.
Tissue samples
We obtained archival, formalin-fixed and paraffin-embedded (FFPE) material from surgically resected lung cancer specimens containing tumor and paired adjacent non-malignant epithelium tissue from the Shanghai Chest Hospital from 2008 to 2013 (Shanghai, CHINA). The adjacent non-malignant epithelia were collected at sites at least 2 cm away from the edge of tumor mass, with best efforts of avoiding contamination by the tumor cells. In total, 150 primary NSCLC patients without prior radiotherapy or chemotherapy were enrolled in this study. All NSCLC samples were confirmed histologically, and tumor samples were rechecked to ensure that tumor tissue was present in more than 80% of the specimens.
Immunohistochemistry
The samples of lung tumor and adjacent normal lung tissues were fixed with formalin buffer and embedded in paraffin. Immunohistochemical (IHC) staining was performed on 3-μm sections of paraffin-embedded specimens with the use of mouse monoclonal antibody (clone 1 L23) to the C-terminus of GPRC5A (PYKDYEVKKE) or rabbit anti-GPRC5A-Y317/Y320-P. These antibodies were developed and purified in Abmart (Shanghai, China). The process of IHC was performed as previously described[
27].
Statistical analyses
All analyses were performed in triplicates, and the significance of differences between groups was calculated using the Student’s t test. P values <0.05 were considered to be statistically significant.
Acknowledgments
We would like to thank Dr. Reuben Lotan (M.D. Anderson, Houston, TX) for his generosity providing NSCLC cell lines and his support. We thank Cathy Anthony for critical reading and editing of this manuscript.
Funding
This work was supported by grants from, Ministry of Science and Technology No. 2011CB504300 (J.Deng) and 2013CB910900 (J.Deng and E.Y.Chin), National Nature Science Foundation of China 91129303 (J.Deng), 81071923 (J.Deng), 81201840 (F.Yao), 81272714 (Q.Li), National Key Basic Research 973 Program of China (2013CB910901) (J.Deng), and Science and Technology Commission of Shanghai (10140902100) (J.Deng). This work was also supported by grants from NIH (2RO1CA125454), Susan G Komen Foundation (KG081310), Mary Kay Ash Foundation (to B.P. Zhou).
Competing interest
The authors declare that they have no competing interests.
Authors’ contributions
XL, SZ, and XY carried out IP-Western, assay of proliferation, migration, invasion and anchorage-independent growth. FY, JZ, and YG collected normal and lung tumor tissue samples. XY participated IHC staining. YL, QL, BS, and YFW participated in development of mutant constructs and stable transfectants. EYC and BH participated in designed the study. BPZ and JD participated in the design of the study and in preparation of the manuscript. All authors read and approved the final manuscript.