The ability to induce expression of a known protein in a model cellular system, such as the cell lines analyzed here, has served as a remarkable, vital tool. However, the GPCR research community that makes use of these cell lines has had only limited data on the complement of GPCRs and related signaling machinery expressed in them. When a cell line is used for heterologous expression, ignorance of the endogenous protein expression profile of a cell line may lead to false conclusions about the nature of receptor function. Indeed, the cellular milieu can have a dramatic impact on GPCR signaling (see . Microarray technology is now sufficiently accessible and mature to determine the mRNA expression for cell lines from species whose genomes have been sequenced. Yet even where microarray studies have been conducted, their data remains only nominally accessible to those with sufficient means, experience and time to extract useful information. The absence of a systematic published study of mRNA expression for signaling proteins of general interest hinders research using these important cell lines. This study provides data examining the presence of several hundred different GPCRs and related signaling proteins in four phenotypically diverse cell lines frequently used in GPCR research.
HEK293 cells are perhaps the most widely used cell line, cited in over four thousand publications and initial work has identified a number of endogenous receptors (see below). We expanded upon this knowledge by demonstrating that HEK293 cells express at least 75 different GPCR mRNAs. Much of what has been reported about GPCR mRNA expression patterns in HEK293 cells was confirmed by our analysis: ADORA2B, CHRM3, F2R, F2RL1, GABABR1, GPCR5A, GPR19, LGR5, LPAR1, LPAR2, P2YR1, PTGER2, S1PR, and SSTR5 [17, 26, 38–46]. HEK293 cells have been reported to possess mRNA for a number of receptors that were not found to be expressed at statistically significant levels in our study: 5HT1D, 5HT6, 5HT7, ADRA2(a-c), ADRAB2, AGTR1, BDKRB2, DRD2, GPR1, GPR161, GRM4, P2YR2 and SSTR2 [17, 26, 46–50]. Our results also indicate that HEK293 cells express mRNA for most G protein subunit isoforms, and numerous isoforms of adenylyl cyclase, protein kinases A and C, and phospholipase C. In addition they are useful for studying regulation of G protein mediated signaling as significant levels of mRNA for beta-arrestin1 and 2, GRK3-5 and a number of RGS subtypes were seen.
Much less is known of the expression patterns of GPCRs in AtT20 cells. We confirmed the presence of Sstr1 and Sstr2 mRNA in AtT20 cells , but did not observe significant levels of any of the other receptors that have previously been reported such as Chrm4 , Vipr1a, Adcyap1r1, or Sstr5 [30, 52, 53]. The lack of significant levels of CRF receptor mRNA was of interest as AtT20 cells have been reported to be sensitive to CRF . mRNA for Hrh3 or Hrh4 histamine receptors were not detectable, despite reports of their presence [55, 56], but a low level of Hrh2 was found. Tacr1  was not detected, nor were Tacr2 and Tacr3. We also did not detect the orphan Cmkor1/Cxcr7 . It is worth noting that there are at least two AtT20 clones available and this may explain some of the differences we found. Generally speaking, AtT20 cells possess a much less diverse complement of GPCR-related signaling and regulatory gene products than the other cell lines tested. Of the approximately 120 GPCR signaling gene products that we studied here, AtT20 cells had 45 significantly expressed signaling related proteins compared to 72 or more in each of the 3 other cell lines. AtT20 cells were the only ones to not have significant levels of beta-arrestin1 (Arrb1) or GRK3-5. Only three RGS gene products were found at significantly high levels: Rgs14, Rgs20, and Snx13.
BV2 cells are commonly used to study microglial biology. These cells exist in a continuum between two states: a resting state and an activated state. For example, upon interferon gamma treatment (primed state), BV2 cells increase both Cnr2 and Gpr55 mRNA levels, whereas with lipopolysaccharide treatment (activated state) these levels decrease . In our case, BV2 cells were cultured under conditions expected to promote a resting state. We can confirm that as reported in the literature BV2 cells express Adora3 , Cnr2 [33, 58], Mtnr1a , Galr2 , Ccr5, Cxcr3  and Rgs2, 10, 12 [27, 28, 63]. We also detected mRNA for receptors that have been reported to be upregulated in response to stimulation such as Ccrl2 [28, 64], Ptafr, Ptigr  and Rgs7 . Here, too, we identified a few receptors that have been reported present for which we found levels of mRNA below the levels that we defined as statistically significant: Hcrtr1 , Gpr55  and the receptor for Cx3cr1 . We did not detect Prokr1 which is upregulated with activation , nor for Oprm1, P2ry1 and Gpr149 which are downregulated following activation [27, 28]. Several reports have identified the presence of mGluRs in BV2 and primary microglia [21, 68, 69]. Interestingly, Grm1 is generally thought not to be present in BV2 cells, yet it was the only one that we significantly detected. We did find low, non-significant levels of Grm3, 6, and 8. The differences may be due to differences in the activation state of the cell lines. BV2 cells appear have a diverse receptor and signaling gene product profile.
Of the cell lines tested, the least well characterized is the N18 neuroblastoma line. They are known to express Cnr1 , which we confirmed with our analysis. It has been reported by some that they do not express muscarinic receptors , but our analysis shows that they have mRNA for three different types (Chrm2, 3, and 4) consistent with other reports [72, 73]. These studies also found evidence for secretin, prostacyclin, opioid and alpha2-adrenergic receptors In support of this we found significantly high levels of mRNA for Ptgir1, Ptger1, Ptger3, Ptger4, Oprd1, Oprl1 and Adra2b. There were also levels of mRNA for Sctr, Adra2a, Oprk1 and Oprm1 that did not reach statistical significance. Another report describes P2ry2 expression , but for this we only found low non-significant levels. Beyond these reports there seems to be little known of GPCR signaling in this cell line and this analysis substantially extends our knowledge of its complement of GPCR expression, at least at the mRNA level. Like BV2 cells, N18 cells appear to have a large repertoire of GPCRs, effectors, and regulating proteins, being consistently one of the highest expressers in each receptor class analyzed.
Interestingly, even the most prolific cell line expressed only a small portion of the full repertoire of non-chemosensory GPCRs (~360). Of these four cell types, BV2 cells expressed the highest total number of GPCRs (108) followed by N18 (105), AtT20 (79), and HEK293 cells (73). BV2 cells had the greatest total abundance of Class A receptors (88), but had equal numbers of Class A orphan GPCRs as N18 cells (figure 3). On the other hand, HEK293 cells have mRNA for the greater number of Class B and Class C orphans (figure 4). It is interesting that the BV2 and the N18 cell lines, which are more often used as models for specific cell types, express mRNA for the most diverse set of receptors. This presumably results in a greater sensitivity to a broad spectrum of ligands and the large number of signaling proteins for which we measured high mRNA levels in these two cell lines attest to this. It is also of interest that HEK293 and AtT20 cells have the lowest numbers of GPCRs of these four lines. Low expression (both of amount and type) may allow these cell lines to serve as a better-differentiated cell model, and, from an experimental perspective, may make a given cell line more attractive as a 'clean slate' into which gene products of interest can be introduced.
mRNA expression levels do not guarantee that a specific mRNA is properly translated, folded, and trafficked to produce functional protein. Discrepancies between our findings and those that have been reported using PCR-based methods may be due to differences that appear in cell lines with time, passage number and growth conditions. Discrepancies may also be attributed to the PCR-based techniques generally employed to extract and amplify mRNA: in some cases the techniques used in many of the studies cited here, some of which do not allow for relative quantification, may have detected very low levels of receptor mRNA that did not reach our threshold for statistical significance. In light of the potential discrepancies between microarray and PCR-based techniques, we employed qPCR analysis to test the validity of our microarray results (figure 7). The overall consistency between our qPCR results and microarray data lend confidence to the concept that the microarray analyses accurately reflect the presence of mRNA for these GPCRs and their signaling proteins.