Other studies have shown that caffeine accelerates telencephalic vesicle evagination in early post-implantation mouse embryos. The present study examines the effect of caffeine on gene modulation in post-implantation mouse embryos. Using mRNA differential display, we observed that caffeine increased gene expression of the regulatory subunit (RIα) of cAMP-dependent protein kinase (PKA). RT–PCR analysis confirmed an increase in expression of this gene in caffeine-exposed embryos when compared with saline-treated controls. Using a fluorescent substrate of PKA, we found that PKA activity in the presence of cAMP was lower in caffeine-treated embryos than in controls. Treatment with H89 and PKI(12-24)amide, two inhibitors of PKA activity, mimicked the effects of caffeine on telencephalic vesicle formation. Together these data suggest that in early post-implantation mouse embryos caffeine modulates gene expression of the RIα subunit of PKA and that caffeine-induced inhibition of PKA activity plays a role in early telencephalic evagination.
Morphogenetic development of the mammalian brain depends on a variety of genetic and epigenetic mechanisms (Hatten and Heintz, 1999). However, the molecular basis of epigenetic mechanisms remains largely unexplored. Recently, we published in vitro (Marret et al., 1997) and in vivo (Sahir et al., 2000) studies on the effects of caffeine on brain morphogenesis. These studies found that caffeine accelerates telencephalic vesicle evagination in early post-implantation mouse embryos. We wished to examine the molecular mechanisms involved in these epigenetic morphological events using the murine model developed previously.
The embryonic precursor of the brain is a planar sheet of pseudostratified neuroepithelium, which is referred to as the neural plate. The process of neurulation converts the neural plate into a neural tube through a series of subdivisions. This results in approximate boundaries among the primordia of the brain regions, namely the prosencephalon, the mesencephalon and the rhombencephalon (Vaage, 1969). The prosencephalon develops through three sequential events: prosencephalic formation, prosencephalic cleavage and midline prosencephalic development. Prosencephalic cleavage involves three divisions of the prosencephalon (Puelles et al., 1987; Bulfone et al., 1993; Rubenstein and Puelles, 1994): (i) horizontally, to form the paired optic vesicles, olfactory bulbs and tracts; (ii) transversally, to separate the telencephalon from the diencephalon; (iii) sagitally, to generate the paired cerebral hemispheres from the telencephalon.
Patterning of brain development is controlled by distinct molecular mechanisms that occur along the anterior–posterior and medial–lateral axes. Medial patterning is regulated by the Sonic Hedgehog gene (SHH) and genes induced by SHH (Echlard et al., 1993; Roelink et al., 1994; Hynes et al., 1995; Ericson et al., 1996).
Disorders in formation of the paired cerebral hemispheres can result in holoprosencephaly. This is one of the most common congenital malformations of the human brain (Roach et al., 1975; Matsunaga and Shiota, 1977). Anatomically, this defect is evidenced by an incomplete separation of the mid-line structures into right and left halves within the developing forebrain (Ming et al., 1998; Wallis et al., 1999). Studies of chromosomal rearrangement have identified at least five loci that are associated with holoprosencephaly. The SHH gene is a candidate gene for one of these loci (Roessler et al., 1996).
Many epigenetic factors appear to modify brain morphogenesis (Prassad et al., 2000). Caffeine is one of these factors (Marret et al., 1997). This molecule is widely consumed by pregnant women (Watkinson and Fried, 1985). Caffeine passes freely through the placenta (Kimmel et al., 1984; Abdi et al., 1993) but the enzymes needed for caffeine metabolism are absent in the fetus and neonate (James and Paull, 1985). During the later stages of pregnancy the half-life of caffeine is increased in both rats and humans (Aldrige et al., 1981). This increase puts at risk organs that are still developing in late gestation, such as the brain (Rousseux and Blakley, 1991).
The present study uses our murine model of caffeine-induced early appearance of telencephalic vesicles, which is described elsewhere (Marret et al., 1997; Sahir et al., 2000). Using mRNA differential display, we studied the effects of caffeine on expression of genes that are potentially involved in prosencephalic segmentation.
Materials and Methods
Animals and Drug Administration
Pregnant Swiss mice were used for all experiments. The day of coitus was determined by the presence of a vaginal plug. The starting point of embryonic age (E0) was defined as the midpoint of the dark cycle during which copulation took place. Dams were injected i.p. once a day (at 9 a.m.) between E8.5 and E10.5 with 200 μl of phosphate-buffered saline (PBS) alone or PBS containing one of the following: 25 mg/kg caffeine (50 mg/kg caffeine citrate); 5, 50 or 75 μg myristoylated PKI(14-22)amide (Biomol, Plymouth Meeting, PA); 5, 50 or 75 μg H89 (N-[p-bromocinnamylamino)ethyl]-5-isoquinolinesulfonamide•2HCl) (Biomol). PKI- (14–22)amide and H89 are two inhibitors of cAMP-dependent protein kinase (PKA) (Kimura et al., 1998; Constantinescu et al., 1999). At E10.5 pregnant mice were killed by cervical dislocation. Embryos were either frozen immediately at –80°C (for subsequent RNA extraction and PKA activity measurement), fixed in 4% paraformaldehyde prior to freezing at –80°C in a cryoprotector (for in situ hybridization) or fixed in 4% formalin solution and embedded in paraffin (for histological analysis).
Coronal sections (16 μm) were cut through the entire embryo and stained with hematoxylin and eosin or cresyl violet. The anterior–posterior length of telencephalic vesicles was determined on the serial sections and used as an index of encephalization.
Total RNA was isolated from the telencephalon of 40 caffeine-treated and 40 control embryos using an RNA isolation kit (Bio/RNA-Xcell; Bio/Gene, Kimbolton, UK). To remove DNA contaminants, RNA samples were treated with RNase-free DNase I (Message Clean; GeneHunter Corp., Nashville, TN) for 30 min at 37°C. The integrity of RNA was checked on a 1% electrophoresis agarose gel.
Differential display was performed using the RNA Image kit (GeneHunter Corp.) according to the manufacturer's instructions. Briefly, 0.1 μg total RNA was reverse transcribed with H-T11M (where M may be dA, dC or dG) primer and Moloney murine leukemia virus reverse transcriptase. Polymerase chain reaction (PCR) amplification was performed in the presence of [α-35S]dATP (Amersham, Les Ulis, France) using Taq DNA polymerase (Perkin Elmer, NJ), the H-T11M (downstream) primer and one of the following upstream arbitrary primers: HPA1, HPA2, HPA3 or HPA6. We used the following cycling parameters: 40 cycles at 94°C for 30 s, 40°C for 2 min and 72°C for 30 s, ending with 5 min at 72°C. The PCR-amplified fragments were run in duplicate on a 6% denaturing polyacrylamide gel (Sequagel-6; National Diagnostics, Atlanta, GA). The gel was dried under vacuum and exposed to Kodak XAR film (Biomax, New York, NY). Genes were selected visually from bands that appeared more intense (up-regulated) or less intense (down-regulated) in caffeine- exposed embryos than in controls.
Cloning and Sequencing of cDNA
The cDNA bands of interest were extracted from the dried gels and boiled in 100 μl dH2O for 10 min. cDNA was precipitated from the supernatants using ethanol and glycogen/sodium acetate and redissolved in dH2O. The preparations were used in reamplification by PCR using the corresponding set of primers for 40 cycles. Reamplified cDNA fragments were subcloned into the PCR-TRAP vector (GeneHunter Corp.). Multiple plasmids were analyzed using PCR and 1.5% agarose gel electrophoresis. Inserts were purified on Wizard Columns (Promega, Madison, WI) and sequenced with Appligene (Illkirch, France) using an Applied Biosystems sequencer. Homology analysis with known mouse cDNAs of sequenced bands was performed with Advanced Blast 2.0 (http:// www.ncbi.nlm.nih.gov/Blast/).
Validation of Genes by Radioactive Relative Reverse Transcription (RT)–PCR
The validation of candidate genes was performed using a modified protocol for radioactive relative RT–PCR (Elalouf et al., 1993). A ‘hot-start’ amplification method used AmpliTaq Gold DNA polymerase (Perkin Elmer); initial activation started at 95°C for 10 min. Amplification was performed for 28 cycles. The primers for the mouse RIα subunit of PKA were designed with the Oligo-4S program to obtain a melting temperature of 60°C and to yield minimal self-priming and upper/lower dimer formation. The sequences were as follows: 5′;-ATTGCCCTGCTGATGAATCGTCCT-3′; (sense) and 5′;-TCTGCGTGAGTGAAGCATGGACTG-3′; (antisense) (amplified fragment 216 bp). To control the quantity of RNA in the different samples, 18S rRNA was reverse transcribed and amplified using the sense primer 5′;-TGGTTCCTTTGGTCGCTCGCTCCTC-3′; and the antisense primer 5′;-TTCCTTGGATGTGGTAGCCGTTTCTCA-3′;.
Digoxigenin-labeled RNA probes (sense or antisense) were produced with the DIG RNA Labeling kit (Roche Diagnostics, Meylan, France) according to the manufacturer's instructions. RNA probes were synthesized from the RIα subunit of PKA using T7 or SP-6 polymerase and digoxigenin–UTP.
In Situ Hybridization
Coronal cryostat sections (10 μm) were cut from caffeine-treated (n = 10) or control (n = 10) embryos. Sections were hybridized overnight at 65°C with antisense RNA probes (1 ng/μl). Slides were washed twice with 50% formamide, 1× SSC, 0.1% Tween-20 at 65°C for 30 min each. Two final washes were done with a solution containing 100 mM maleic acid, 50 mM NaCl and 0.1% Tween-20 at room temperature for 30 min. Slides were incubated overnight at room temperature with alkaline phosphatase- coupled anti-digoxigenin antibody (1/2000) (Roche Diagnostics). BM purple AP substrate (Roche Diagnostics) was used as the chromogenic substrate for alkaline phosphatase. Sense probes were used as controls.
Measurement of PKA Activity
PKA activity was measured using fluorescent-labeled kemptide (Pep Tag assay; Promega). In this assay enzyme-induced phosphorylation is assessed following electrophoretic separation on agarose gels of the phosphorylated peptide from the non-phosphorylated peptide. Both forms of peptide were visualized with UV light. Caffeine-treated (n = 10) and control (n = 10) embryos were homogenized in buffer (20 mM Tris, 10 μM leupeptin, 25 μg/ml aprotinin, 1 mM EGTA, final pH 7.4) (Macala et al., 1998). Cells were lysed by probe sonication for 10 s. The samples (12 μg total protein) were then assayed immediately according to the manufacturer's instructions. Measurement of PKA activity was repeated three times.
RT–PCR Analysis of SHH mRNA
RT–PCR analysis of SHH mRNA was performed as described above on control and caffeine-exposed embryos. The sequences were as follows: 5′;-TCTGTGATGAACCAGTGGCC-3′; (sense) and 5′;-GCCACGGAGTTCTCTGCTTT-3′; (antisense) (amplified fragment 222 bp).
Identification of Expressed Genes
We looked at mRNA to see if telencephalic transcripts were expressed differentially in control versus caffeine-treated embryos. Samples from control and treated embryos were run in duplicate. Consistent with other findings (Sahir et al., 2000), embryos treated with caffeine (25 mg/kg) showed significant acceleration of telencephalic vesicle formation when compared with control embryos (Fig. 1A–C). To explore molecular mechanisms associated with this phenomena, we used a mRNA differential display approach. Using four of the eight pairs of primers we found that 25 cDNA fragments have a different level of intensity for caffeine-exposed samples compared with controls. Twenty-four bands were less intense in the telencephalons of caffeine-treated embryo samples, while one band was more intense (Fig. 1D). In the present study we focus on this band (430 bp), corresponding to a gene up-regulated in caffeine-exposed telencephalons.
This 430 bp fragment was cloned and sequenced. Blast analysis indicated that this sequence has a 230 bp open reading frame (ORF) and 200 bp of 3′;-UTR with a polyadenylation signal at 408 bp. The 230 bp ORF translation product is identical to the C-terminus of the mouse RIα subunit of PKA (GenBank accession no. CAB94718.1) (Fig. 2).
To validate the finding of increased expression of the RIα subunit of PKA in caffeine-exposed telencephalons we used relative RT–PCR, as described in Materials and Methods above. Relative RT–PCR experiments using 20 and 60 ng total RNA found increased expression of the RIα subunit of PKA in caffeine-treated embryos. Compared with controls, there was a 58% increase in the RIα subunit of PKA in caffeine-exposed telencephalons based on PhosphorImager quantification (Fig. 3A). In contrast, 18S rRNA, which was used as an internal control, was not modified (Fig. 3B). In situ hybridization of the RIα subunit of PKA mRNA showed that expression of the RIα subunit of PKA was ubiquitous and that the level of expression of this gene was similar in all embryonic tissues. The pattern of expression of the RIα subunit of PKA was identical in caffeine- treated and control embryos (Fig. 4). However, expression of the RIα subunit of PKA was significantly increased in all tissues of caffeine-exposed embryos (Fig. 4).
PKA Activity Measurement
The PKA holoenzyme contains a regulatory subunit dimer and two catalytic subunits. The enzyme is activated when the regulatory domain binds four molecules of cAMP and the catalytic subunits are released. In this study we found that caffeine increased expression of the regulatory subunit of PKA. Therefore, it also held the potential to decrease PKA activity. To test this hypothesis we used a fluorescent-labeled kemptide assay to measure PKA activity (Macala et al., 1998). Compared with PBS controls, caffeine produced a minimal decrease (3% reduction) in basal PKA activity (without cAMP addition) and a marked decrease (44% reduction) in PKA activity in samples where cAMP was added to obtain maximal PKA activity (Fig. 5).
Inhibition of PKA Activity
Mouse embryos were treated in vivo with inhibitors of PKA [H89 and PKI(12-24)amide] to assess the effects of caffeine on (i) RIα subunit of PKA expression and (ii) PKA activity and telencephalic vesicle formation. Measurement of the rostrocaudal length of telencephalic vesicles showed that both inhibitors accelerated the evagination of telencephalic vesicles in a dose-dependent manner (Fig. 6).
RT–PCR Analysis of SHH mRNA
Relative RT–PCR experiments using 20 and 60 ng total RNA found increased expression of SHH in caffeine-treated embryos. Compared with controls, there was a 38% increase in SHH in caffeine-exposed telencephalons based on PhosphorImager quantification (Fig. 7) while 18S rRNA, which was used as an internal control, was not modified.
This study sought to identify genes that are modulated by caffeine treatment and to assess their potential role in caffeine- induced acceleration of telencephalic vesicle evagination (Marret et al., 1997; Sahir et al., 2000). The findings of studies using different methods (i.e. mRNA differential display strategy, relative RT–PCR and in situ hybridization) support the notion that caffeine can increase expression of the RIα subunit of PKA. However, less can be said about the molecular mechanisms leading to this increase.
Encephalic vesicle formation involves a combination of cell proliferation, cell death and cell migration (Zaki and Van der Loos, 1980; Haydar et al., 1999). Although these parameters have not been directly studied in the present in vivo study, caffeine was previously shown to significantly increase both the mitotic and pycnotic indices in the prosencephalon of in toto cultured embryos (Marret et al., 1997). On the other hand, the previously described normalization of brain anatomy and histology in our caffeine-exposed embryos within a few days following caffeine discontinuation (Sahir et al., 2000) is in agreement with an acceleration of telencephalic vesicle formation under caffeine exposure rather than a permanent caffeine-induced modification of vesicle size.
Measurement of the phosphorylase activity of PKA revealed that caffeine exposure resulted in increased expression of the RIα subunit of PKA as well as a decrease in PKA activity in the presence of cAMP. Treatment of embryos with two different PKA inhibitors (H89 and PKI) produced an effect on telencephalic vesicle evagination that mimicked the effect seen with caffeine treatment. Taken together, these data suggest that caffeine-induced inhibition of PKA activity contributes to accelerated encephalization.
Is there a relationship between increased expression of the RIα subunit of PKA and decreased activity of PKA following exposure to caffeine? The present study does not address this question, but published data raise some interesting hypotheses. PKA is a tetrameric protein composed of two regulatory subunits (RIF128a, RIβ, RIIα or RIIβ homodimers) and two catalytic subunits (Reiman et al., 1971; Krebs and Beavo, 1979). Fixation of four cAMP molecules on the regulatory homodimer will induce release of the catalytic subunits, permitting phosphorylase activity of PKA. The level of PKA activity can be modulated by its subunit composition (Sarkar et al., 1984; Leiser et al., 1986; Bregman et al., 1989; Cummimgs et al., 1996) and by the level of intracellular cAMP (Schwartz and Rubin, 1983; Hedin et al., 1987; Liu and Lin, 1988). Therefore, in the present study PKA activity could be modified by a change in subunit composition in caffeine-treated embryos. Furthermore, if we assume that caffeine increases gene expression of the RIα subunit without interfering with the catalytic subunits and that caffeine does not modify the equilibrium constant, we would expect to see reduced activity of PKA by displacement of the chemical equilibrium (Houge et al., 1990).
Interestingly, transgenic mouse embryos expressing a dominant negative form of PKA that inhibits PKA activity display increased activation of the SHH pathway (Epstein et al., 1996; Hammerschmidt et al., 1996). As mentioned previously, SHH gene inactivation has been implicated in the pathophysiology of human and murine holoprosencephaly (Belloni et al., 1996). This suggests that early exposure to caffeine can produce an inverted picture of holoprosencephaly by inhibiting PKA activity, which in turn could favor the SHH pathway. Further- more, using relative RT–PCR, we found that caffeine enhances SHH gene expression.
In situ hybridization experiments found that caffeine increases RIα subunit expression in all embryonic tissues. As our study focused on the prosencephalon, we cannot exclude the possibility that caffeine interferes with the early development of other embryonic tissues. Furthermore, previous reports have demonstrated that PKA activation is implicated in long-term potentiation in the hippocampus (Taylor et al., 1999). Additional studies to assess the potential effects of caffeine on PKA gene expression at later stages of brain development would be of interest.
In conclusion, the present study shows the involvement of PKA activity in caffeine-induced acceleration of encephalization. Further studies are needed to understand the precise mechanisms by which caffeine modulates expression of the RIα subunit of PKA as well as the potential role of other, unidentified genes differentially expressed in caffeine-exposed embryos.
We are grateful to Heather G. Miller for critical reading of the manuscript and we particularly appreciate the assistance of Jean-Marie Moalic and Jean-François Benoist. This work was supported by INSERM and the Fondation pour la Recherche Médicale.
Address correspondence to Pierre Gressens, INSERM E 9935, Hôpital Robert-Debré, 48 Boulevard Sérurier, F-75019 Paris, France.