Lysophosphatidic acid (LPA) is a mediator of multiple cellular responses. LPA mediates its effects predominantly through the G protein-coupled receptors LPA1, LPA2, and LPA3. In the present work, we studied LPA2-mediated signaling using human colon cancer cell lines, which predominantly express LPA2. LPA2 activated Akt and Erk1/2 in response to LPA. LPA mediated Akt activation was inhibited by pertussis toxin (PTX), whereas Erk1/2 activation was completely inhibited by a blocker of phospholipase Cβ, U-73122. LPA also induced interleukin-8 (IL-8) synthesis in the colon cancer cells by primarily activating LPA2 receptor. We also found that LPA2 interacts with Na+/H+ exchanger regulatory factor 2 (NHERF2). Activation of Akt and Erk1/2 was significantly attenuated by silencing of NHERF2 expression by RNA interference, suggesting a pivotal role of NHERF2 in LPA2-mediated signaling. We found that expression of LPA2 was elevated, whereas expression of LPA1 downregulated in several types of cancers, including ovarian and colon cancer. We conclude that LPA2 is the major LPA receptor in colon cancer cells and cellular signals by LPA2 are largely mediated through its ability to interact with NHERF2.
- Na+/H+ exchanger regulatory factor 2
lysophosphatidic acid (LPA), a naturally occurring phospholipid, is a mediator of multiple cellular responses, including proliferation, differentiation, motility, and protection from apoptosis (4, 11, 23). LPA mediates its effects predominantly by interacting with members of the endothelial differentiation gene (EDG) subfamily of G protein-coupled receptors (GPCRs): LPA1/EDG2, LPA2/EDG4, and LPA3/EDG7. Despite the many studies performed on the biological effects of LPA via the LPA receptors, specific functions and cellular signaling mediated by individual LPA receptors are not fully understood. This can partly be attributed to heterogeneous effects observed in different cellular contexts that differ in the receptor expression and cellular environment mediating the biological effects (23). All of the LPA receptors can couple to three distinct families of heterotrimeric GTP binding proteins, including Gαq, Gαi, and Gα12/13 (4, 11, 23). Heterologous expression of LPA receptors in mammalian cells showed that all the receptors mediate inhibition of adenyl cyclase, PLC activation, and mitogen-activated protein kinase (MAPK) activation leading to arachidonic acid release (18). Although the similarity in cellular signaling by these three receptors appears to be redundant, these receptors display distinct genetic and biochemical properties.
LPA1 is expressed in a wide range of tissues and organs, whereas LPA2 and LPA3 show more restricted patterns of expression (1, 2). LPA2 expression is generally low in many tissues, with exceptions of testis, pancreas, prostate, and leukocytes. Genetic deletion of LPA1 in mice results in severe defects, such as craniofacial dysmorphism, defective sucking behaviors rendering semilethality, and a low occurrence of frontal hematoma. In contrast, LPA2-null mice do not show any obvious phenotypic aberrations (5). These studies implicate distinct physiological roles of LPA1 and LPA2 in vivo, and also reveal that normal physiological functions during development do not absolutely require LPA2 activation. On the other hand, LPA2 expression is increased in ovarian and thyroid cancers and inflammation, suggesting that the main function of LPA2 could become evident under pathophysiological conditions (6, 29).
In this study, we report that that LPA2 is the major LPA receptor that mediates mitogenic signals and cytokine induction in colonic epithelial cells. We also report LPA2 receptor-elicited signaling is largely mediated via its ability to interact with NHERF2.
METHODS AND MATERIALS
Cell culture and treatment.
Caco-2 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (FBS) as previously described (38). SW480 cells were grown in RPMI-1640 supplemented with 10% FBS. 1-Oleoyl-2-hydroxy-sn-glycero-3-phosphatidic acid was obtained from Avanti Polar Lipids and prepared in PBS containing 0.1% BSA (vol/vol). Cells were pretreated with 50 ng/ml pertussis toxin (PTX), 50 μM LY-29004, 5 μM U-73122, or vehicle before exposure to LPA. All of the inhibitors were obtained from Calbiochem.
Total RNA was prepared from cells or tissues by using TRIzol (Invitrogen). Five micrograms of total RNA were used for the subsequent synthesis of cDNA by using the First Strand Synthesis kit as recommended by the manufacturer (Invitrogen). PCR reactions were performed on 0.1 μl of the cDNA mix with 10 pmol of each primer in a final volume of 50 μl of 1× AmpliTaq Gold PCR Master Mix (Applied Biosystems). The following primer pairs were used: LPA1 (accession no. NM001401): 5′-AACATGGCACCCCTCTACAGTG-3′ and 5′-ATCATTTACTCCTACCG-3′; LPA2 (accession no. NM004720): 5′-AGCCTGGTCAAGACTGTTGTCATC-3′ and 5′-CTAAAGGGTGGAGTCCATC-3′; LPA3 (accession no. NM022983): 5′-GTCAAGAGGAAAACCAACGTC-3′ and 5′-GAGAAGCAGCAGATCATCTTC-3′; IL-8 (accession no. NM000594): 5′-CCAAGCTGGCTGTTGCTCTCTTGG-3′ and 5′-ACCTT(C/T)TG(C/T)ACCCACTTTTCCTTGG-3′; and β-actin (accession no. X00351): 5′-TATGCCAACACAGTGCTGTCTGG-3′ and 5′-TACTCCTGCTTGCTGATCCACAT-3′. The primers sets were designed to cross an exon-intron junction to avoid any contamination from the genomic DNA. Amplification was performed for 30 cycles at 94°C/45 s, 52°C/45 s, and 72°C/30 s.
Cloning of LPA2.
cDNAs for LPA2 were amplified by RT-PCR from total RNA prepared from human embryonic kidney-293 cells. Amplification was performed for 35 cycles at 94°C/45 s, 52°C/45 s, and 72°C/2 min. The following primer pairs were used for the amplification: 5′-TTATGGTCATCATGGGGCAG-3′ and 5′-CTAAAGGGTGGTGGAGTCCATC-3′. The fidelity of the PCR product was confirmed by nucleotide sequencing. The entire LPA2 cDNA was cloned into pVM6 (Roche) to ligate the antibody epitope derived from vesicular stomatitis virus glycoprotein in frame with the NH2 terminus of LPA2, resulting in pVM-LPA2. In addition, the COOH terminal (CT) 4 amino acid residues of LPA2 were truncated by PCR and the PCR product was cloned in pVM6, resulting in pVM-LPA2Δ4.
Hybridization of the cancer profiling array.
cDNA probes corresponding to the entire reading frame of LPA2 or LPA1 were labeled with (α-32P)dATP by nick translation. The cancer profiling array purchased from Clontech was hybridized with the cDNA probes using ExpressHyb solution according to the recommendation by the manufacturer.
In vitro interaction assays.
The CT 44 aa of LPA2 was generated as a glutathione S-transferase (GST) fusion protein in bacteria. Two micrograms of plasmids carrying NHERF1, NHERF2, and CAP70 were labeled with 35S-Met by using the TNT in vitro Transcription-Translation System (Promega). Four micrograms of GST-LPA2 immobilized on glutathione-agarose beads were incubated with compatible amounts of 35S-labeled proteins (20–30 μl of the in vitro products) in a Tris-buffered buffer, as previously described (38). At the end of the incubation period, the complexes were washed three times with the same buffer, resolved by SDS-PAGE, and visualized by autoradiography.
The separate domains of NHERF2 expressed as GST fusion proteins have previously been described (38). These include the full-length NHERF2, the first PSD95-Dlg-ZO1 (PDZ) domain (P1: aa 9–128), the second PDZ domain (P2: aa 92–270), and the CT domain (C: aa 232–337).
Immunoprecipitation and Western immunoblot analysis.
Cells were lysed in lysis buffer (50 mM Tris-Cl, pH 7.4, 150 mM NaCl, 0.1 mM PMSF, 1 mM Na orthovanadate, 10 mM Na fluoride, 10 mM Na pyrophosphate, 25 mM β-glycerophosphate, 1% Triton X-100, and protease inhibitors). Immunoprecipitation and Western blot analyses were performed as previously described (39). The polyclonal antibodies against NHERF1 and NHERF2 have previously been described (39). Anti-LPA2 against the NH2 terminus was obtained from Exalpha Biologicals. All other antibodies were purchased from Cell Signaling.
Small interfering RNA.
RNA interference (RNAi) was achieved by means of small hairpin RNA targeted against NHERF2 using pSilencer-H1 of the pSilencer small interfereing RNA (siRNA) Expression System (Ambion). Synthetic oligonucleotides containing 21 nt of sense and antisense sequences separated by a loop were designed using the siRNA Target Finder program and subcloned into pSilencer-H1. Caco-2 cells seeded at 70% confluence on 10-cm culture dishes were transfected with 10 μg of plasmid DNA using Lipofectamine2000 overnight in the presence of 10% FBS. As a control, scrambled 21-nt sequences were cloned into pSilence-H1. Cells were selected for resistance to 300 μ/ml hygromycin 48 h after transfection. The expression of NHERF2 was determined by Western immunoblot analysis using the polyclonal anti-NHERF2 antibody (39). The expression of NHERF2 was examined every 4–5 passages and its expression was found to be unchanged up to passage 20. For all experiments, the cells were used between passages 6 and 20.
Small interfering RNA duplex (siRNA) directed against LPA2 was obtained from Ambion. Cells were transfected with 50 nM oligonucleotide as described above. Western immunoblot was performed 48 h after transfection using the anti-LPA2 antibody.
Measurement of IL-8 expression.
Overnight serum-starved cells were treated with LPA for 5 h. To determine the amount of IL-8 protein secreted by the cells, conditioned media was collected and centrifuged at 5,000 g for 5 min to remove cell debris. The collected samples were either used for IL-8 measurement by ELISA (Biosource) or frozen at −80°C until needed. For statistic analysis, the LPA-induced IL-8 is defined as the increase in IL-8 production as the result of LPA treatment, i.e., the amount of IL-8 in the presence of LPA minus IL-8 in the absence of LPA treatment.
Statistical significance was assessed using ANOVA and Student's t-test. Results were considered statistically significant when P < 0.05. If not otherwise specified, results are presented as means ± SE.
LPA2 signaling in Caco-2 cells.
We determined expression of LPA receptors in various colonic epithelial cells and colonic tissues by semiquantitative PCR. Figure 1 shows that all three LPA receptors were expressed in colon from human and rabbit. However, in all colon cancer cell lines, including Caco-2, T84, HT29, SW480, RKO, and HCT116, the expression levels of LPA2 were significantly elevated relative to normal colon and noncancerous IEC6 cells when normalized to β-actin levels. On the other hand, the expression levels of LPA1 and LPA3 were low in these cancer cells relative to the controls.
To study LPA2-mediated signaling in colon cancer cell lines, we chose to use Caco-2 cells, which predominantly express LPA2. Differentiated Caco-2 cells were treated with 10 μM LPA for the indicated times, and phosphorylation of Akt and Erk in response to LPA was determined. In all cases, we first determined the amount of phospho-Akt and phospho-Erk1/2. The blots were stripped and blotted for total Akt and Erk1/2. Figure 2A shows that Akt and Erk1/2 were activated in response to LPA in Caco-2 cells. Akt was maximally activated between 30 and 60 min and remained stimulated for at least 2 h. Phosphorylation of Erk1/2 peaked at 10 min but remained activated above the basal level for the following 1 h.
LPA receptors are known to interact with multiple G proteins, including Gαq and Gαi, and the signaling pathways dependent on these G proteins vary among different cell lines and tissues (4, 11, 23). To determine how Gαi and Gαq are involved in LPA2-mediated pathways in Caco-2 cells, we assessed the effects of pharmacological inhibitors on Caco-2 cells. We evaluated the effects of the phosphatidylinositol 3-kinase (PI3-kinase) inhibitor LY-29004, the Gαi inhibitor pertussis toxin (PTX), and the PLC blocker U-73122. Compared with LPA alone, pretreatment of Caco-2 cells with 50 ng/ml PTX significantly blocked LPA-induced activation of Akt, indicating that the PI3-kinase-Akt cascade was mediated by Gαi (Fig. 2B). On the other hand, there was a marginal effect by PTX on Erk1/2 activation by LPA, suggesting that Erk1/2 were activated via a separate pathway.
In contrast to PTX, pretreatment of the cells with U-73122 for 15 min before the addition of LPA had no apparent effect on phosphorylation of Akt (Fig. 2C). Instead, LPA-induced stimulation of Erk1/2 was completely blocked by the inhibitor, indicating that the PLCβ-Erk1/2 pathway was Gαq dependent. Interestingly, the phosphorylation levels of Erk1/2 decreased over time in the presence of U-73122. It seems plausible that there is a basal level of stimulation from PLCβ to the Ras-Erk pathway that maintains the basal phosphorylation of Erk1/2. However, in the presence of the PLCβ blocker, this basal level of stimulation is gradually depleted, and the rate of dephosphorylation predominates.
Although different pathways appear to regulate the PI3-kinase-Akt cascade and the MAPK activation, previous studies showed the presence of “cross-talk” among the signaling cascades initiated by GPCR (28, 34, 37). To test this possibility, Caco-2 cells were pretreated with the PI3-kinase inhibitor LY-294002. As expected, the pretreatment completely blocked activation of Akt (Fig. 2D). In contrast, LY-294002 had a small and statistically insignificant effect on Erk1/2 phosphorylation LPA, suggesting that PI3-kinase does not influence phosphorylation of Erk1/2 in Caco-2 cells.
IL-8 induction by LPA in intestinal epithelial cells.
It has been demonstrated that LPA induces synthesis of IL-6 and IL-8 in ovarian cancer cells and lung epithelia (8, 19). To determine whether LPA can mediate cytokine synthesis in Caco-2 cells in a LPA2-dependent manner, we treated Caco-2 cells with 10 μM LPA for 5 h. The medium was applied to the Human Inflammatory Antibody Array III (RayBiotech, Norcross, GA). Of 40 anti-cytokine antibodies on the array, only IL-8 was significantly increased within the 5-h incubation (Fig. 3A). The stimulation of IL-8 synthesis by LPA was confirmed by RT-PCR. The amount of IL-8 mRNA was increased by 2.3 ± 0.4 fold in Caco-2 treated with LPA (Fig. 3B). Consistently, LPA induced IL-8 induction by similar extent (2.8 ± 0.3 fold increase) in SW480 cells that also overexpress LPA2.
To directly link IL-8 induction to LPA2, we attempted to express GFP-LPA2 in Caco-2 cells. Because of low transfection efficiency (10%) in Caco-2 cells, a ∼15% increase (P < 0.05) in LPA-induced IL-8 induction was attained (data not shown). To confirm this result, we transfected SW480 cells with GFP-LPA2, resulting in a marked increase in IL-8 induction in response to LPA (Fig. 3C). LPA-induced IL-8 synthesis was almost doubled in the cells expressing GFP-LPA2 compared with the control cells. To further corroborate the role of LPA2 in IL-8 induction, we silenced LPA2 expression by siRNA. Transient transfection resulted in 57% LPA2 knocked down as determined by Western immunoblot analysis using the polyclonal anti-LPA2 antibody (Fig. 3D). The knockdown of LPA2 expression decreased the magnitude of IL-8 induction by 50% compared with control (Fig. 3C). The residual IL-8 induction was probably due to the incomplete knockdown of LPA2. These data indicate that the IL-8 production is regulated by LPA2 in these cells.
LPA2 interacts with NHERF2.
PDZ domains preferentially bind the consensus CT sequence-d-S/T-φ-L/V, where φ is a hydrophobic residue (13, 32). Examining the CT sequences of the LPA receptors revealed that LPA2 has the CT sequence of -DSTL, whereas LPA1 and LPA3 have -HSVV and -KSTS, respectively. We predicted that LPA2, but not LPA1 or LPA3 would interact with PDZ domains.
To determine whether LPA2 interacts with PDZ domains, we tested the potential interaction of LPA2 with three PDZ-containing proteins that are known to be present in colonic epithelia: NHERF1, NHERF2, and CAP70. The Na+/H+ exchanger regulatory factor family proteins, NHERF1 and NHERF2, were initially identified as proteins that reconstitute protein kinase A (PKA)-dependent inhibition of the Na+/H+ exchanger NHE3, and contain two tandem PDZ domains (40). CAP70 contains 4 PDZ domains, and interacts with the type IIa Na-phosphate cotransporter and the cystic fibrosis transmembrane conductance regulator (CFTR) (10, 33).
To examine the interactions between the LPA receptors and PDZ proteins, we generated the carboxy 44 aa of LPA2 as a GST fusion protein. The GST-LPA2 was used to test its interaction against NHERF1, NHERF2, and CAP70 that were labeled with 35S-Met by in vitro transcription and translation. Figure 4A reveals that only NHERF2 showed substantial interaction with LPA2. Figure 4B shows that LPA2, but not LPA1 or LPA3, bound NHERF2, consistent with the presence of PDZ recognition sequences only at the CT of LPA2. LPA1 and LPA3 did not exhibit any interaction with NHERF1 or CAP70 (data not shown). We next examined the structural determinant of the NHERF2 interaction with LPA2. The two PDZ domains (PDZ1 and PDZ2) and the ezrin-radixin-moesin (ERM) binding CT of NHERF2 expressed as GST fusion proteins were used to test their interaction with the LPA2 CT domain. Figure 4C shows that LPA2 preferentially bound the second PDZ domain. An interaction with the first domain was also evident but at much lower efficiency compared with the second PDZ domain. No interaction with the ERM-binding CT of NHERF2 was observed. To determine whether the CT end of LPA2 mediated this interaction, we transfected PS120/NHERF2 fibroblasts with pVM-LPA2 or pVM-LPA2Δ4. NHERF2 was immunoprecipitated using the anti-NHERF2 antibody from the lysates prepared from the transfected cells, followed by Western immunoblot analysis using the monoclonal antibody against the vesicular stomatitis virus glycoprotein tag to detect the presence of coimmunoprecipitated LPA2. In agreement with the in vitro binding study, NHERF2 coimmunoprecipitated LPA2, but not LPA2Δ4, suggesting that the interaction occurred through the CT 4 amino acids of LPA2 (Fig. 4D). To determine whether LPA2 interacted with NHERF2 in a cell line that endogenously expresses all the proteins, NHERF2 was immunoprecipitated from Caco-2 cells. Immunoprecipitation of NHERF2 resulted in coimmunoprecipitation of LPA2, confirming the interaction under conditions where all of the proteins are expressed at endogenous levels (Fig. 4E). In contrast, immunoprecipitation of NHERF1 failed to coprecipitate LPA2, despite the affinity of the anti-NHERF1 is more than several fold higher than the anti-NHERF2.
Knockdown of NHERF2 expression affects LPA2-mediated Akt and MAPK activation.
To determine the role of NHERF2 in LPA2-mediated signaling in Caco-2 cells, we used RNAi approach to suppress NHERF2 expression in Caco-2 cells. The cells stably transfected with the shRNA against NHERF2 were selected by resistance to hygromycin, and these cells were named Caco-2/CL4. Figure 5A shows that the expression level of NHERF2 in Caco-2/CL4 was 18 ± 3% compared with control-transfected Caco-2 (Caco-2/pSiL). As controls, expression levels of NHERF1 and LPA receptors were not affected by the transfection (data not shown). The expression level of NHERF2 in these knockdown cells remained low relative (<25%) to the control cells up to 20 passages in the presence of the antibiotic selection.
NHERF proteins are previously shown to alter GPCR signaling via their association at the CT end of the receptors (14, 21). As in untransfected cells, LPA activated Akt and Erk1/2 in Caco-2/pSiL cells (Fig. 5B). Compared with the control cells, the knockdown of NHERF2 expression in Caco-2/CL4 cells led to almost complete loss of LPA-induced Akt activation in Caco-2/CL4 cells. The effect on Erk1/2 was not as striking as Akt, but yet the knockdown of NHERF2 resulted in statistically significant decreases in phosphorylation of Erk1/2. These results suggest that cellular signaling by LPA2 is largely mediated through its interaction with NHERF2, and that NHERF2 is an essential component of the signal transduction initiated by LPA2.
LPA2-mediated IL-8 induction requires NHERF2.
Our data above showed that the LPA2 receptor induces IL-8 induction in colon cancer cells. To further demonstrate the role of NHERF2 in LPA2-mediated signaling, we determined the effect of NHERF2 knockdown in IL-8 induction in these cells. Treatment of Caco-2/pSiL and Caco-2/CL4 resulted in 2.61 ± 0.13-fold and 1.30 ± 0.09-fold increase, respectively, in IL-8 mRNA in response to 10 μM LPA as determined by semiquantitative RT-PCR (Fig. 6A). These results were confirmed by real-time PCR. When normalized against GAPDH expression levels, LPA showed 7.3 ± 0.3 fold increases in IL-8 mRNA level in control Caco-2/pSiL vs. 1.6 ± 0.1 fold increase in Caco-2/CL4.
The effect of NHERF2 knockdown in IL-8 induction was confirmed by ELISA. LPA stimulated IL-8 protein synthesis in Caco-2/pSiL in a concentration-dependent manner. There was a maximum of fourfold increase in IL-8 protein in response to LPA. On the other hand, the knockdown of NHERF2 attenuated the IL-8 induction at all LPA concentration tested (Fig. 6B).
LPA2 is overexpressed in cancers.
Our study showed that LPA2 was overexpressed in all colon cancer cell lines relative to normal controls. Although it has been demonstrated that LPA2 expression is elevated in ovarian cancer, it is unclear whether a similar overexpression of LPA2 exists in other types of cancer (6, 8). We sought to determine whether LPA2 expression is upregulated in other cancers by using the Cancer Profiling Array. This array contains >100 pairs of cDNAs generated from normal (N) and tumor tissue (C) samples from individual patient, spotted side by side on a nylon membrane. Consistent with the previous observations (6, 8), LPA2 expression was elevated in ovarian cancer (Fig. 7A, lane 2). In addition, we found that several other types of cancers such as colon (lane 3), lung (lane 5), uterus (lane 12), rectum (lane 14), and testis (lane 16) displayed elevated levels of LPA2. Similarly, various cancer cell lines of noncolonic origins (lane 20) showed increased expression of LPA2. One unexpected finding was that LPA1 expression (Fig. 7B) was decreased in some cancers, including ovary, colon, vulva, thyroid, and testis compared with normal tissues. Of these cancers, ovary, colon, rectum, and testis showed concomitant increase LPA2 and decrease LPA1 expression, whereas lung, uterus, vulva, and thyroid cancers were associated with changes in one LPA receptor. Figure 7C shows expression of the housekeeping gene ubiquitin as a control.
In the present work, we showed that signaling mediated by LPA2 receptor induces multiple effects in Caco-2 cells. We utilized Caco-2 cells as a colon cancer cell model to understand LPA2-mediated signaling because LPA2 is the predominant LPA receptor expressed in Caco-2 cell and other colon cancer cell lines. Our assumption that the effects of LPA were mediated by LPA2 in these cells was supported by the significant effects resulted by the knockdown of NHERF2 that interacts only with LPA2. In addition, knockdown of LPA2 expression in SW480 cells significantly inhibited IL-8 synthesis in another colon cancer cell line, SW480, which also overexpresses the LPA2 receptor.
Specific cellular signals mediated by individual receptors vary in different cell types and tissues. In Caco-2 cells, LPA stimulates the PI3-kinase-Akt and PLCβ-Ras-Erk1/2 pathways via GαI and Gαq, respectively. PTX treatment significantly compromised activation of Akt by LPA, suggesting that this activation is mediated by Gαi. On the other hand, blocking PLCβ with U-73122 did not affect the activation of Akt, but completely inhibited Erk1/2 stimulation. Taken together, the Gαi-PI3-kinase-Akt cascade and the Gαq-PLCβ-Erk1/2 cascade appear to operate independently of one another.
Of three LPA receptors, only LPA2 contains the typical PDZ interacting motif, −DSTL, and we anticipated its interaction with PDZ domains. In this study, we demonstrated NHERF2 as a specific interacting partner of LPA2. LPA2 did not interact with NHERF1 or CAP70 that share significant amounts of similarity with NHERF2 (81% vs. NHERF1, 61% vs. CAP70), and in some cases interact with the same set of proteins, such as CFTR and type IIa Na-phosphate cotransporter. As we predicted, NHERF2 did not interact with LPA1 or LPA3. While this work is being pursued, Oh et al. (24) reported a similar finding showing the specific interaction between LPA2 and NHERF2. NHERF proteins mediate assembly of signaling proteins, such as PLCβ, Gαq, and the receptor-activated C kinase-1, in microdomains to facilitate cellular signaling processes (17, 20, 27). Through such an interaction, NHERFs are known to alter not only GPCR-mediated signaling, but also the signaling that regulates downstream targets, such as NHE3 and CFTR (14, 21, 25, 26, 40). In the present study, the knockdown of NHERF2 in Caco-2 cells severely compromised cellular signaling mediated by LPA2. NHERF2 knockdown almost completely abolished LPA-mediated activation of Akt, suggesting that NHERF2 facilitates the interaction between LPA2 and Gαi. The effect on Erk1/2 stimulation was not as striking as Akt. This is in contrast to the recent report by Oh et al. (24) that overexpression of NHERF2 greatly potentiated phosphorylation of Erk in rat fibroblasts. In addition, we were not able to detect a significant effect by NHERF2 knockdown on LPA-induced activation of cyclooxygenase-2 in Caco-2 cells (unpublished data). The reasons for the discrepancy are not clear but there may be cell type differences. An alternative explanation comes from a recent finding that LPA2 interacts with TRIP6 (thyroid receptor interacting protein 6), which associates with focal adhesion kinase (FAK) in response to LPA (36). Because FAK can activate Erk1/2, it is possible that in Caco-2 cells LPA2 partially activates Erk1/2 via the pathway involving TRIP6 and FAK.
LPA is known to induce cytokine production in different cells. Our results show that LPA has a robust effect in IL-8 induction in Caco-2 and SW480 cells. Although we did not determine time course of IL-8 induction by ELISA, induction of IL-8 transcript was observed within 30 min of the addition of LPA (unpublished data). Among 40 different cytokines tested in Caco-2 cells, only IL-8 was induced within the 5 h of LPA exposure. However, these experiments do not rule out the possibility that other cytokines could be induced in these cells upon a longer exposure to LPA. For instance, a prolonged exposure of ovarian cancer cells to LPA induces IL-8 and IL-6 production (8). As further proof that LPA2 is the major LPA receptor in the colon cancer cells, the knockdown of NHERF2 or LPA2 receptor attenuated LPA-induced IL-8 production.
The elevated levels of LPA in the ascites from patients with epithelial ovarian cancer and increased LPA2 expression in ovarian cancer have been demonstrated (7, 12, 22). But whether expression of LPA2 is elevated in other types of cancers is less well known. Using the Cancer Profiling Array, we demonstrated that LPA2 expression is increased in several other types of cancers, including colon, uterus, rectum, testis, and lung, suggesting that the upregulation of LPA2 is a commonly occurring effect in cancer. In accordance with our finding, Shida et al. (31) recently reported a similar increase in LPA2 expression in colorectal cancer patients by real-time PCR. Although the underlying mechanism for overexpression of LPA2 in several types of cancer is not known, it is noteworthy that several sequence variants in the 3′URT of human LPA2 gene that may lead to a loss of normal regulation of LPA2 gene expression have been found in cancerous cells (3). We did not determine LPA3 expression, but Shida et al. (31) did not find significant differences in LPA3 expression between the colorectal cancer patients and healthy subjects. In contrast to the changes in expression of LPA2 and LPA1 in several types of cancers, we did not observe significant difference in the expression of NHERF2 between cancer tissues and the controls (data not shown).
On the basis of the concurrent upregulation of LPA2 and downregulation of LPA1 expression in some cancers, it is plausible that opposing effects are mediated by LPA1. For example, overexpression of LPA1 in ovarian cancer cells induced apoptosis and anoikis, whereas LPA2 promoted proliferation and migration of ovarian cancer cells (9, 15). In addition, it appears that LPA2 is a more efficient in transducing proliferative responses, recruitment of autocrine protein growth factors, such as vascular endothelial growth factor and urokinase-type plasminogen activator (16). On the other hand, other studies have shown both LPA1 and LPA2 promote cell proliferation, suggesting that the specific effects by these receptors may vary depending on cellular context (30).
The findings on Akt activation and induction of IL-8 by LPA may also have physiological and pathophysiological responses in inflammation. Epithelial injury and inflammation release cytokines, including IL-8, to the site of inflammation exerting various effects, including neutrophil infiltration, cell proliferation, and wound healing (35, 41). However, whether the expression of LPA2 receptor or its signaling is altered in response to inflammation or injury is not known.
In summary, we have provided evidence for a physiological importance of the LPA2 receptor in regulation of LPA-induced MAPK activation and IL-8 induction in human colon cancer cells. In addition, our work demonstrated a functional role of NHERF2 in regulation of LPA2-mediated signaling in colon cancer cells that may affect cancer progression. Further understanding of the physiological significance of NHERF2 in LPA2-mediated signaling should help to understand the biological roles of LPA2 in tumorigenesis and inflammation.
This work was supported by the National Institutes of Health Grants DK-61418, DK-64399, GM-60982, and NS-45644, and the Georgia Cancer Coalition.
We are grateful to Dr. Vincent W. Yang for helpful comments throughout the study, and thank Dr. Georgia Chen for a critical reading of the manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society