The pituitary adenylate cyclase-activating peptide (PACAP), a member of the secretin/glucagon/VIP family, is a ubiquitous and pleiotropic regulator with a wide range of biological effects (e.g., ontogenesis, pain, circadian rhythm, and behavior). PACAP functions as a hormone/trophic factor and neurotransmitter in the peripheral as well as in the central nervous system. ¹ PACAP can exert its various effects via three different receptors: PAC1, VPAC1, and VPAC2. The VPAC1 and VPAC2 receptors exhibit equal affinities for both PACAP and vasoactive intestinal peptide (VIP), whereas the PAC1 receptor binds PACAP with a 100-fold higher affinity than VIP. ^2–4 In rat, six splice variants of the PAC1 receptor have been identified (Null, Hip, Hop1, Hop2, Hiphop1, and Hiphop2), defined by the insertion of two different cassettes in the third intracellular loop. The Null, Hop1, and Hop2 isoforms trigger both adenylate cyclase (AC) and phospholipase C (PLC) activation, whereas the Hip variant activates AC exclusively. Combinations of both cassettes (Hiphop1 and Hiphop2) result in an intermediate phenotype. ⁵ In addition to the above-mentioned isoforms, two other PAC1 receptor variants (short and very short) have been described as lacking motifs within the extracellular N-terminal domain. ^6–8 First and foremost, PACAP and its receptors have been a focus of neurotoxicity research ^9–12 as a consequence of their strong antiapoptotic effect. However, the expression of both the PACAP gene and its receptors during ontogenesis suggests their involvement in developmental processes. ^13–15 In fact, a number of functional studies have verified that, in the central nervous system, PACAP contributes to cell proliferation, differentiation, apoptosis, and neurite outgrowth. ^16–19 Along with its antiapoptotic effect, PACAP can act as a mitogenic regulator, as well as an antimitogenic regulator. ^19–21 On one hand, it stimulates premature granule cells or PC12 cells to develop processes driving them into differentiation, ^19,21 while on the other hand, endogenous PACAP also exerts a mitogenic function in chick neuroblasts or in 8-day-old cultured cerebellar neurons. ^17,22 Evidently, the versatile and opposing effects correlate with differential expression of PAC1 receptor isoforms. Therefore, to investigate the roles of PACAP in any neural tissue, it is necessary to first identify the PACAP receptor subtypes expressed in the tissue of interest. 

The development of the complex, highly organized retina is an area of active research. The integration of retinal cells into functional circuits results from temporally overlapping processes (i.e., cell genesis, apoptosis, migration, and synaptogenesis) that are regulated by numerous intrinsic and extrinsic environmental factors. ^23–29 Although the presence of PACAP receptors and their isoforms in the mammalian retina has been described, ^15,30 the physiological function of this peptide has not been established, nor has precise quantitative analysis of the PACAP receptor expression in the retina during postnatal development been carried out. 

The present study aimed to assess the expression levels of three PACAP receptors, PAC1, VPAC1, and VPAC2, to identify the PAC1 receptor splice variants, and to analyze their expression patterns during the developmental time course in the rat retina. We describe the presence and expression levels of the three major receptors, PAC1, VPAC1, and VPAC2, and four PAC1 receptor isoforms, Null, Hip, Hop1, and Hiphop1. We demonstrated for the first time that VPAC1 and VPAC2 receptor transcript levels change significantly during postnatal retina development. Our results also reveal unique expression profiles of Null, Hip, Hop1, and Hiphop1 receptor isoforms and their alterations during retina development. The results suggest a complex PAC1 splicing regulation that may contribute to the key mechanisms of retinal development. 

Newborn (postnatal day 0 [P0]) and 1, 3, 5, 10, 15, and 20-day-old (postnatal days P1, P3, P5, P10, P15, and P20) albino Wistar rats were used for this study. Animal handling, housing, and experimental procedures were reviewed and approved by the ethical committee of the University of Pécs (BA02/2000-6/2006), and all animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Animals were anaesthetized by inhalation using isoflurane (Forane; Abbott Laboratories Ltd., Queenborough, UK) or by intraperitoneal injection of thiopental (50 mg/g) and then decapitated. Animals were always sacrificed in the same hour to prevent variations due to the circadian cycle. Eyes were removed, and retinas were dissected in cold phosphate-buffered saline. Upon dissection, tissues were frozen on dry ice and stored at −80 C° until RNA extraction. 

For total RNA extraction and purification, homogenizer and spin columns (QIAshredder and RNeasy Plus Mini Kit; Qiagen, Valencia, CA) were used in accordance with the manufacturer's instructions. Total RNA was determined by measuring the optical density at 260 nm with a spectrophotometer (NanoDrop ND 1000; NanoDrop Technologies, Wilmington, DE). Purity was estimated by the absorption ratio at 260/280 nm, which was consistently higher than 1.8. Two micrograms of total RNA was converted into first-strand cDNA using oligo(dT) primer (Fermentas, Glen Burnie, MD) and reverse transcriptase (Maxima II; Fermentas). 

In order to increase specificity and yield, we employed a modification of RT- PCR called touchdown PCR in which the annealing temperature was gradually lowered to a more permissive temperature during the course of cycling, favoring amplification of the desired product. Amplification was carried out in a 25-μL reaction mixture containing 4 μL cDNA, 50 nM forward primer, 50 nM reverse primer, and PCR master mix (Maxima Hot Start PCR Master Mix; Fermentas). The primers used for the PCR reactions are listed in the Table. For VPAC1, VPAC2, and splice-independent PAC1, primers were designed close to the poly(A) tail on the 3′ noncoding area, whereas primers amplifying PAC1 receptor isoforms were designed to bridge the splice variant-specific exons inserted between exons 14 and 15 of the PAC1 receptor gene and their neighboring exon. 

The amplification products were analyzed by 1.5% agarose gel electrophoresis. Then, amplifications of the desired products were confirmed by sequencing. 

Real-time PCR was performed in a 25-μL reaction mixture containing 4 μL cDNA, 5 nM forward primer, 5 nM reverse primer, and 12.5 μL SYBR Green PCR Master Mix (Fermentas). Each sample was analyzed in triplicate. The 2^−ΔΔCt (cycle threshold) method for relative quantification of gene expression was used to calculate mRNA expression levels. Fold changes were adjusted considering the efficiency of each primer pair as determined by the method of serial dilution. The P0 samples were used as a basis for relative quantitation results. To normalize fluorescence signals, RPL13a was utilized as an endogenous control. 

Retinas were harvested from animals on P0, P1, P3, P5, P10, P15, and P20. For protein extraction, tissues were homogenized in 300 μL radioimmunoprecipitation assay (RIPA) buffer (10 mM phosphate buffer pH 7.2, 1% NP-40, 1% Na-deoxycholate, 0.1% SDS, 0.15 M NaCl, 2 mM EDTA, 2 μg/mL aprotinin, 0.5 μg/mL leupeptin, 2mM sodium vanadate, 20 mM sodium fluoride, 0.5 mM dithiothreitol, and 10 mM phenylmethylsulfonyl fluoride) with micropestles on ice for 5 minutes. Thereafter, samples were centrifuged and the supernatants saved. Protein concentration was determined by using a bicinchoninic acid protein assay kit (BCA Protein Assay Kit; Pierce, Rockford, IL). Sample preparation, buffer preparation, gel electrophoresis, and blotting were performed as provided in the NuPAGE Instruction Manual (Invitrogen, Carlsbad, CA). Approximately 25 to 30 μg protein/sample was loaded and run in 10% polyacrylamide gel. For PAC1 receptor detection, membranes were incubated in anti-PACAP receptor type 1 antibody (Pierce) diluted to 1:1000. The antibody recognizes the extracellular N-terminus of PAC1 receptor. For VPAC1 and VPAC2 receptor detection, primary antibodies (Pierce) were used in 1:200 and 1:500 dilution, respectively. The quantitative data were representative of three samples pooled from three animals at each time point. The data were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression using mouse anti-GAPDH antibody (1:10,000) (Sigma-Aldrich, Budapest, Hungary). Anti-mouse and anti-rabbit IgG antibodies conjugated with horseradish peroxidase were diluted to 1:10,000. For signal detection, we used chemiluminescence reagent (Western Lightning Chemiluminescence Plus reagent; PerkinElmer, Waltham, MA). The chemiluminescent signal was captured either on high-sensitivity blue-tinted film (Kodak X-OMAT Blue Film XB; Sigma, Budapest, Hungary) or with an imaging system (FluorChem Q Imaging system; ProteinSimple, Santa Clara, CA). The latter allowed a precise and prompt quantification of the chemiluminescent signals. Quantitation analysis was performed using AlphaView Q software (ProteinSimple, Santa Clara, CA). 

Data are expressed as means ± SD. In all cases, expression values of P0 retinas served as reference groups. Our goal was to identify groups whose means were significantly different from the mean of these reference groups. Therefore, the statistical comparisons were performed by one-way ANOVA analysis followed by post hoc Dunett's test, which is specifically designed for situations where all groups are pitted against one group. A P value of <0.05 was considered statistically significant. 

Changes of transcript level during postnatal development were compared to the P0 retina as a reference. First, we used primers specific for all PAC1 isoforms. However, no remarkable overall changes could be detected in PAC1 receptor message level (Fig. 1A). The only statistically significant elevation was measured at P1 (1.38 ± 0.07). In later developmental stages, at P3 (1.08 ± 0.24), P5 (0.89 ± 0.04), P10 (1.12 ± 0.13), P15 (0.83 ± 0.06), and P20 (1.04 ± 0.11), the ratio of PAC1 receptor expression relative to that at P0 was constant. 

In regard to VPAC1 (Fig. 1B), no significant increase was observed as early as P1 (1.06 ± 0.43) and P5 (1.42 ± 0.61), whereas an upregulation was detected at P10 (2.15 ± 0.53). The expression of VPAC1 receptor further increased at P15 (2.61 ± 0.39). Afterwards, a marked but not statistically significant downregulation could be seen at P20 (1.58 ± 0.45). 

The expression pattern of VPAC2 appeared to be biphasic (Fig. 1C). There was a moderate peak at P5 (3.14 ± 0.14) and another more significant increase at P15 (7.16 ± 2.13); VPAC2 was still elevated at P20 (5.12 ± 2.09) as well. 

Using splice variant-specific primers, a sensitive RT-PCR was utilized to reveal the set of PAC1 isoforms expressed in rat retina. Four time points were chosen (i.e., P0, P5, P15, and P20) in order to cover the postnatal development proportionally. At P0 and P5, Null, Hip, Hop1, and Hiphop1 splice variants were detected (Figs. 2A, 2B). At P15, Null, Hip, and Hop1 displayed expression similar to the levels in the earlier stages, whereas Hiphop1 could be seen only as a faint band (Fig. 2C). Expression of Hip seemed to have ceased by P20 (Fig. 2D). Nontemplate controls are shown in Fig. 2E. 

Results obtained from the touchdown PCR indicated that the expression of the individual PAC1 receptor isoforms might display unique profiles. Although RT-PCR is a sensitive method to amplify the desired products, it is not an adequate tool for quantification. Therefore, expression changes were mapped by employing quantitative real-time PCR. 

As a first approach, retinas harvested at P0 were used as reference samples. Consequently, the expression of each PAC1 isoform depicted in Figure 3 was related to its level measured at P0. The Null isoform showed no impressive changes at P1 (1.37 ± 0.06) or P3 (0.9 ± 0.02) but then manifested a decline from P5 (0.64 ± 0.07) to P10 (0.55 ± 0.13), and the message level fell at P15 (0.13 ± 0.02) (Fig. 3A). The Hip isoform had a similar expression pattern (Fig. 3B), with a peak at P1 (1.67 ± 0.29), followed by continuous downregulation by P20 (0.39 ± 0.15). The lowest expression level was detected at P15 (0.15 ± 0.01). The expression levels of the Hop1 splice variant did not change at P1 (1.41 ± 0.07), P3 (1.19 ± 0.01), or P5 (1.46 ± 0.02), but thereafter, the Hop1 message level showed significant increases at P10 (3.33 ± 0.42), P15 (3.35 ± 0.41), and P20 (6.94 ± 1.54) (Fig. 3C). The expression analysis of the Hiphop1 isoform showed one prominent but not statistically significant peak at P10 (5.54 ± 1.06) (Fig. 3D). 

Once the expression profile of each individual PAC1 isoform was determined, we investigated the expression of the isoforms in relation to the Null isoform at each time point to reveal the ratios between the PAC1 isoforms. Since the previous experiment showed that none of the isoforms appeared to change between P3 and P5, we investigated only the following time points: P0, P1, P10, P15, and P20. In early development, the Hip splice variant was expressed to the greatest extent at P0 (10.6 ± 3.41) and P1 (5.00 ± 0.64) (Figs. 4A, 4B). The Hop1 transcript became dominant at P10 (14.75 ± 1.15) (Fig. 4C) and P15 (54.49 ± 2.9) (Fig. 4D) and reached its highest level at P20 (80.6 ± 9.59) (Fig. 4E). Although the Hiphop1 variant displayed considerable elevation at P10 compared to its level at P0, among the PACAP variants, it was present in rat retina in the smallest amount. Looking at any time point from P0 to P20, the Hiphop1 message level was lower than that of any other isoform (Figs. 4A–E). 

To confirm the up- or downregulation of mRNA, immunoblotting with specific antibodies was performed. Unfortunately, there are no commercially available antibodies against PAC1 splice variants. However, due to the presence or absence of the cassettes inserted into the third intracellular loop, isoforms can be separated by polyacrylamide gel electrophoresis. The PAC1 receptor antibody used in our experiments recognizes the extracellular N-terminal sequence; consequently, it binds to all isoforms of the PAC1 receptor that share an identical N-terminal sequence. Since Hop1 and Hip variants have the same molecular mass (56 kDa), they could not be separated, and bands show an additive expression of Hip and Hop1 isoforms (Fig. 5A, tagged as 2). An upregulation of the PAC1 receptor was clearly demonstrated at P15 and P20. Our antibody labeled a protein with the molecular mass of the Null isoform (53 kDa) as well (Fig. 5A, 3), which disappeared by P15 and P20. Evidently, the band around 59 kDa corresponded with the largest PAC1 isoform, namely Hiphop1 (Fig. 5A, 1). The Western blot result showed the presence of the Hiphop1 protein at all time points. 

To validate our quantitative PCR results on the protein level, immunoblotting was carried out. The Western blot results for VPAC1 receptor (47–55 kDa) expression are shown in Figure 5B. The appearance of multiple bands is the result of glycolization of the receptor's N-terminal sequence. Consistent with the gene expression, VPAC1 receptor protein could be detected at all selected time points. However, the amount of the VPAC1 receptor appeared to display no changes during postnatal retinal development. For quantification, the most intense bands that appeared around 50 kDa were scanned. In Figure 5C, changes in the VPAC2 (65 kDa) protein level are depicted. Similar to VPAC1, VPAC2 receptor appears to be expressed in the newborn as well as in the adult retina. VPAC2 expression showed no significant changes from P0 through P20. 

In a wide range of vertebrate species, the expression of the PACAP receptors starts in embryonic stages and continues to be present in the central nervous system through postnatal development. ^13,14,31,32 As a result of alternative splicing, nine isoforms of the PAC1 receptor can be involved in developmental processes. ^6,7,33 In fact, PACAP exerts diverse, occasionally opposing effects on the developmental processes. Depending on the type of the PAC1 receptor isoform, PACAP can induce precursor cells to exit the cell cycle through activation of the Null isoform ³⁴ or can promote proliferation in neuroblasts that express the Hop isoform. ²⁰ Therefore, the expression pattern of PACAP receptors during postnatal development could determine the effect of their ligand. Changes in PAC1 isoform expression have been reported in the cerebellum and suprachiasmatic nucleus, as well as in the cortex. ^30,35,36 However, very little information describing the expression of PACAP receptors in the developing mammalian retina can be found, which provides the rationale for our study. Therefore, we harvested retina tissues on P0, P1, P3, P5, P10, P15, and P20, covering all important developmental stages, to assess the expression changes of the major PACAP receptors. 

According to a recent study, PACAP and PAC1 could be detected as early as E16 in the rat retina and they continued to be expressed throughout postnatal development. ¹⁵ Among the PAC1 receptor isoforms, the presence of Hop1 transcript has been reported in adult rat retina only, ³⁰ and no quantitative data are available from newborn or developing tissue. In the present study, we not only demonstrated the presence of PAC1 receptors in different developmental stages of the postnatal rat retina but we showed that for quantification of PAC1 receptor expression, the existence of PAC1 receptor isoforms must be taken into account. Primers amplifying all isoforms of PAC1 receptor are not adequate to detect expression changes since the fall in message level of one may be compensated by the rise in another. We found that in the newborn rat retina (e.g., P0 and P1), four PAC1 receptor isoforms, namely, Null, Hip, Hop1, and Hiphop1, were expressed. The dominant isoform at this age was the Hip receptor; however, the message level of this isoform appeared to drop at P5 and its expression was downregulated in the mature retina (e.g., P15 and P20). A similar decline was seen for the Null isoform. The seven cell types which later comprise the adult retina can be characterized by an individual proliferation peak from P0 through P6. ^23,28,37 Concurrently, the newborn, postmitotic cells migrate radially, as well as tangentially, to find their destined location. ^38,39 In order to establish the proper cell number, many of the postmitotic cells undergo apoptosis, which sweeps through the developing retina in two waves. ^24,40,41 The enhanced expression of Hip and Null receptor at this period indicates that these splice variants might be involved in regulating the early developmental processes mentioned above. It is noteworthy that the Null receptor engages dual signaling cascades, activating both AC- and PLC-coupled pathways and intracellular Ca²⁺ elevation, whereas Hip is coupled only to the AC pathway. ⁴² Therefore, it is reasonable to assume that these two receptor types might mediate different responses in the developing rat retina. It is important to point out that in spite of their lower expression, a functional role for the other PAC1 isoforms (e.g., Hop1 and Hiphop1) expressed at these time points cannot be excluded. Furthermore, our results showed a splice isoform switching, estimated to occur between P5 and P10, close to the time of eye opening that is scheduled between P10 and P15. In this period, the Hop1 isoform, which is known to have a higher potency for generating cAMP or increase intracellular Ca²⁺ level than Null or Hip receptors, became the dominant isoform. ^42,43 It seems that the period of neurite outgrowth and synapse formation of the inner plexiform layer (IPL) from P3 to P10 ⁴⁴ coincide with an increasing Hop1 expression. Furthermore, the formation of functional neuronal circuitries occurs between P10 and P15. ^25,27,44,45 Our finding that a prevalent expression of Hop1 receptors was maintained from P10 through P20 suggests a late-phase role for this particular isoform in the postnatal retina development. Similar to Hop1, a marked but transient upregulation of the Hiphop1 variant was also found to coincide with eye opening. We note that the expression of this particular isoform varied greatly among the samples harvested at the same time points. Moreover, in spite of a prominent peak that appeared at P10, a variable upregulation was measured in all selected stages. Considering that the literature providing data about the Hiphop1 isoform is relatively sparse, the functional role of this particular PAC1 isoform remains to be defined. However, based on our findings, we speculate that Hiphop1 expression might be coupled to special physiological conditions (i.e., nutritional status or hypoxia) or that it might be determined by the genetic background of the individual animals. In addition, Hiphop1 expression seems to be affected by the PACAP level itself, since intravitreal injection of a potent PAC1 antagonist, PACAP6-38, increases the expression of Hiphop1 receptor (unpublished data). 

Since the transcript level might not correlate with protein expression, it was crucial to confirm the existence of PAC1 receptor protein by immunoblot assay. Although specific primary antibodies are not available for PAC1 receptor isoforms, based on their molecular masses, we could identify three bands of PAC1 receptor, suggesting the presence of PAC1 receptor splice variants at the protein level as well. 

PACAP can act via PAC1, as well as VPAC receptors, which have affinity not only for PACAP but also for VIP. In fact, VPAC1 and −2 receptors are exclusively responsible for mediating the effect of the latter peptide, ⁴⁶ and thus, their expression could be related to the presence of VIP. Like PACAP, VIP has been reported to stimulate cell division, differentiation, and growth through VPAC receptors in cultured tissues. ^47–49 With respect to retinal development, the paucity of detailed information on the expression patterns of VPAC1 and −2 receptors is striking. Our analysis revealed statistically significant yet slight increases in transcript levels of VPAC1 during postnatal development. VPAC1 receptor mRNA was first detected at P0, and its level showed modest peaks at P10 and P15. The latter data correlate temporally with the VIP ligand expression, which was first detected by in situ hybridization at P5 and seemed to peak at P15 in retinal amacrine cells. ⁵⁰ However, our protein expression analysis of VPAC1 showed no convincing difference at either time point compared to the newborn retina, instead showing a uniformly enhanced expression of VPAC1 receptors. In the case of the VPAC2 receptor, we found that VPAC2 transcription also became more manifest in the late phase of retina maturation. Likewise, VPAC1 and VPAC2 protein levels did not show positive correlations with the expression of their messenger. The discrepancy between transcript and protein levels indicates either posttranscriptional regulation or fast cellular degradation of the VPAC receptors. Nevertheless, their enhanced expression in the postnatal retina suggests that VPAC1 and −2 receptors could contribute to retinal development. 

In conclusion, in the present paper, we demonstrated that all receptors of the ubiquitous PACAP and VIP have different time courses of expression, and now it remains to be tested how one or more of them influence development. It is important to point out that in order to study PAC1 receptor expression, one must take into account the existence of splice variants because, as indicated in our results, general PAC1 receptor primers are not adequate to analyze expression changes. Since each PAC1 receptor isoform displays a unique expression profile, we suggest that selective splicing and expression of PAC1 receptor may be a pivotal component of developmental processes. Considering that PAC1 isoforms induce different signaling pathways, their combination or changes in their coexpression in relation to evoked responses could be critical. The observations that the receptors displayed enhanced and differential expression in different stages of the postnatal retina ontogenesis suggest a versatile physiological role for both PACAP and VIP throughout retinal development. 

The authors thank Paul Witkovsky (New York University, New York, NY), who improved the language of our manuscript and provided useful critical comments, Csaba Fekete (University of Pécs, Hungary) for technical advice and support, and Peter Kisfali (University of Pécs, Hungary) for performing DNA sequencing.