Abstract
The neuropeptide calcitonin gene-related peptide (CGRP) is expressed by one-third of adult rat lumbar dorsal root ganglion (DRG) neurons, many of which mediate pain sensation or cause vasodilation. The factors that regulate the developmental expression of CGRP are poorly understood. Embryonic DRG neurons initially lack CGRP. When these neurons were stimulated in culture by serum or persistent 50 mm KCl application, the same percentage of CGRP-immunoreactive (CGRP-IR) neurons developed in vitroas was seen in the adult DRG in vivo. The addition of the L-type calcium channel blockers, 5 μm nifedipine or 10 μm verapamil, dramatically decreased the proportion of CGRP-IR neurons that developed, although the N-type calcium channel blocker, 2.5 μm ω-conotoxin, was less effective. By contrast, the sodium channel blocker 1 μm tetrodotoxin had no effect on CGRP expression after depolarization. Fura-2 ratiometric imaging demonstrated that mean intracellular free calcium levels increased from 70 to 135 nm with chronic depolarization, and the addition of nifedipine inhibited that increase. Only a subpopulation of neurons had elevated calcium concentrations during chronic depolarization, and they were correlated with CGRP expression. Key signal transduction pathways were tested pharmacologically for their role in CGRP expression after depolarization; the addition of the CaM kinase inhibitor KN-62 reduced the proportion of CGRP-IR neurons to basal levels. By contrast, protein kinase A and protein kinase C were not implicated in the depolarization-induced CGRP increases. These data suggest that depolarization and the subsequent Ca2+-based signal transduction mechanisms play important roles in the de novo expression of CGRP by specific embryonic DRG neurons.
- sensory ganglion
- calcitonin gene-related peptide
- depolarization
- signal transduction
- ion channel
- calcium imaging
Adult sensory ganglia are composed of distinct neuronal populations, but the factors that generate this diversity remain unclear. For example, neuropeptides, including calcitonin gene-related peptide (CGRP), often are found in small-sized neurons that constitute approximately one-third of the dorsal root ganglion (DRG) and that can innervate skin and viscera (Lee et al., 1985; Gibbins et al., 1987; Molander et al., 1987; O’Brien et al., 1989; McCarthy and Lawson, 1990; Noguchi et al., 1990; Kashiba et al., 1991). CGRP is localized in polymodal nociceptors and is a potent vasodilator (Brain et al., 1985; Wallengren and Hakanson, 1987; Holzer, 1988; Scott, 1992). Although CGRP first appears in lumbar DRG in vivo when peripheral target connections are functional on embryonic day 18 (E18) (Narayanan et al., 1971; Saito, 1979; Kudo and Yamada, 1985; Marti et al., 1987; Kucera et al., 1988; Fitzgerald, 1991), cell culture studies with E14 rat DRG indicate that de novo CGRP expression occurs in the absence of those target contacts (Hall et al., 1997). Transcription and steady-state CGRP mRNA levels in cell lines can be upregulated by cAMP, forskolin, phorbol ester, and nerve growth factor or can be inhibited by glucocorticoids, retinoic acid, and vitamin D (deBustros et al., 1985, 1986; Haller-Brem et al., 1988; Naveh-Many and Silver, 1988; Lindsay and Harmar, 1989;Russo et al., 1992; Tverberg and Russo, 1992). Although these studies have identified candidate CGRP regulators, little is known about the initiation of CGRP expression in primary sensory neurons. At least one report suggests caution in assuming that primary sensory neurons and cell lines use equivalent CGRP regulatory mechanisms (Watson et al., 1995). The required combination of factors needed for the generation of CGRP containing neurons in the DRG remains poorly defined.
One important cue that can regulate neuropeptide expression is neuronal activity. In peripheral neurons and neural cell lines, membrane depolarization increases vasoactive intestinal peptide, substance P (Sun et al., 1992; Adler and Fink, 1993), neuropeptide Y (Higuchi et al., 1996), preprotachykinin mRNA (Noguchi et al., 1988), or pituitary adenylate cyclase-activating polypeptide (Brandenburg et al., 1997). In many neurons the elevation of extracellular potassium leads to an increase in intracellular calcium, which can interact with calmodulin and modulate calcium-binding enzymes such as CaM kinases, protein kinases, and adenylyl cyclases. These effectors for second messenger molecules then mediate cellular responses by activating transcription factors that alter immediate early and delayed response gene expression (Misra et al., 1994; Ghosh and Greenberg, 1995; Bito et al., 1997). In most cases, depolarization increases peptide expression in neurons that already synthesize the neuromodulator. By contrast, the early development of the catecholaminergic phenotype in some sensory neurons also can be regulated by neuronal activity (Hertzberg et al., 1995;Brosenitsch et al., 1998). It is not known whether depolarization can affect de novo peptide expression. Embryonic DRG neurons are spontaneously active during the period just preceding and during the formation of functional peripheral target contacts (Fitzgerald, 1987). Thus, early neuronal activity may play a critical role in the neuropeptide differentiation of sensory neurons.
To test the role of neuronal activity in the onset of CGRP expression and to elucidate factors that regulate CGRP expression, embryonic rat DRG neurons were cultured in chemically defined medium and depolarized them with 50 mm KCl. Because little is known about the link between calcium entry and the regulation of delayed response genes important for neuronal function, the role of calcium influx and calcium-mediated signaling pathways was examined by ratiometric imaging and by pharmacological analyses. These studies demonstrate that depolarization is sufficient to induce CGRP expression in DRG neurons. Further, molecular components of the signal transduction systems that regulate the onset of CGRP expression were identified.
MATERIALS AND METHODS
Cell culture. E14.5 DRG were dissected from Sprague Dawley rats (Zivic Miller, Zelienople, PA). The dissociation, plating, and survival assays were as described (Hall et al., 1997). Approximately 4000 DRG cells were plated in each 0.32 cm2 well. Growth medium consisted of Neurobasal chemically defined medium (Life Technologies, Grand Island, NY) with B-27 serum-free supplement (Life Technologies), 3 mmglutamine, and nerve growth factor (NGF; 25 ng/ml, Austral Biological, San Ramon, CA). The potassium concentration of Neurobasal growth medium (NB) was 5 mm (NB5). To depolarize cells, we added a stock solution of 4 m KCl to vary the [K+] between 10 and 75 mm. In some cases, 5% heat-inactivated rat serum (RS) was added to L15/CO2 medium with NGF. Iso-osmolar control medium consisted of 1% mannose in NB5 medium, which empirically produced an osmolarity equivalent to Neurobasal medium with 50 mm KCl (NB50).
In some experiments, pharmacological agents were added to cultures in NB50 at the time of plating, and one-half of the medium was exchanged with fresh medium every 1 or 2 d. The agents used were the following: 1 μm tetrodotoxin (TTX; Na+channel blocker; Calbiochem, La Jolla, CA) (Elliott and Elliott, 1993), 10 μm nifedipine or 10 μm verapamil (Calbiochem) (Gault and Siegel, 1997), 2.5 μmω-conotoxin GVIA (ω-Con; Sigma, St. Louis, MO) (McCleskey et al., 1987), 1 μm{1-[N,O-bis(5-isoquinolinesulfonyl)-N-methyl-l-tyrosyl]-4-phenylpiperazine} (KN-62; Calbiochem) (Tokumitsu et al., 1990), 2 μm{N-[2-((p-bromocinnamyl;)amino)ethyl]-5-isoquinolinesulfonamide, HCl} (H-89; Calbiochem) (Findik et al., 1995), and 1 μmbisindolylmaleimide III (BIM III; Calbiochem) (Toullec et al., 1991). For compounds dissolved in dimethyl sulfoxide (DMSO), the control cultures included 0.1% DMSO in the growth medium.
Immunocytochemistry. Cells were fixed in 4% paraformaldehyde and 0.1 m PO4, pH 7.4, for 30 min at room temperature (RT). After being washed, the cells were permeabilized with dilution buffer containing 0.2% Triton X-100/PBS, 20% goat serum, and 0.01% sodium azide for 2 hr and then were incubated in primary antibody (rabbit anti-CGRP, 1:300; Amersham, Arlington Heights, IL) overnight at 4°C. After being rinsed, the secondary antibody (biotin-conjugated goat anti-rabbit IgG, 1:250; Chemicon, Temecula, CA) was applied to the cultures. The staining was visualized with avidin–horseradish peroxidase and a diaminobenzidine (DAB) chromagen. In each case at least 200 neurons were counted per well, and triplicate wells in each of at least two independent experiments were quantified. Data were compared between groups by using an unpaired Student’s t test.
Calcium measurement. Intracellular free Ca2+ was monitored with the fluorescent dye fura-2 AM (Molecular Probes, Eugene, OR). Cells were grown on poly-l-lysine/laminin-coated No. 1 glass coverslips affixed to 60 mm Petri plates (Fisher Scientific, Pittsburgh, PA), and cells were loaded with 4 mm fura-2 AM in NB growth medium at 37°C for 20 min. After being washed, the cultures were replaced with fresh growth medium and maintained at 37°C for an additional 20 min to allow for complete hydrolysis of the dye. Coverslips then were placed on the stage of a Zeiss Axiovert 405M microscope (Oberkochen, Germany) prewarmed to 37°C, and intracellular fura-2 was excited at 350 and 380 nm with a xenon lamp. A Hamamatsu SIT camera or Princeton CCD camera was used to collect information. The fluorescence signal at 510 nm was collected, and the ratio of the fluorescence at the two excitation wavelengths was calculated by MetaFluor analysis software (Universal Imaging, West Chester, PA). Areas on the same coverslips without cells were recorded as background images. Intracellular free [Ca2+] was calculated according to Grynkiewicz et al. (1985): [Ca2+] =KD ×Fmin/Fmax × (R −Rmin/Rmax −R), where KD = 225 nm, R is the ratio value,Rmin is the ratio value in a calcium-free solution, Rmax is the ratio value of a saturated calcium solution (1 mm CaCl2),Fmin is the fluorescence intensity of the calcium-free solution at 380 nm, and Fmax is the intensity of a 1 mm calcium solution at 380 nm. Actual values were obtained by using point-to-point correlations obtained from specific free calcium standards between 0 and 40 μm(Molecular Probes). In some cases, cells cultured in NB50 for 4 d on a glass-etched grid (Bellco Glass, Vineland, NJ) were subjected to calcium imaging, followed by immunocytochemistry for CGRP.
RNA isolation and PCR. Total cellular RNA was isolated with RNAzol (Sigma) from each well of a six well plate containing 100,000 DRG cells. Five micrograms of total RNA was DNase I-treated and reverse-transcribed. Primers for CGRP (sense primer 5′-ATGCAGATGAAAGTCAGGGA-3′ and antisense primer 5′-GGGGCTATTATCTGTTCAAG-3′, recommended by Andy Russo, University of Iowa) or elongation factor 1α (sense 5′-TTCACTGCTCAGGTGATTATCC-3′ and antisense 5′-GGCAGCATCACCAGACTTCAAGA-3′) were added in 20 mmTris-HCl, pH 8.4, 50 mm KCl, and 2 mmMgCl2, layered with mineral oil, and subjected to 35 cycles of PCR amplification (denatured at 94°C for 45 sec, annealed at 55°C for 45 sec, and extended at 72°C for 1 min) in a PTC-200 Peltier thermal cycler. PCR products were separated in a 2% agarose gel and detected after ethidium bromide staining and UV illumination.
RESULTS
Embryonic sensory ganglia harvested before CGRP was detectable and before functional connection with targets had occurred in vivo subsequently became CGRP-immunoreactive (CGRP-IR) over time in serum-containing cultures, demonstrating that CGRP expression is an intrinsic property of a subpopulation of DRG neurons (Hall et al., 1997). However, this culture system was not appropriate for identifying factors that regulate CGRP expression because of the uncertainties as to the nature and activity of trophic substances within RS. To understand further the cellular and molecular mechanisms that regulatede novo CGRP expression, we developed a chemically defined culture system and investigated the factors required for CGRP expression.
Depolarization elicited CGRP expression in embryonic sensory neurons
To test the role of depolarization in CGRP expression, we placed E14.5 DRG cells in basal NB5 or depolarizing NB50 medium for 8 d, a period in which CGRP expression develops and stabilizes. The stimulus used, 40–50 mm KCl, is a common tissue culture manipulation used to mimic aspects of neuronal depolarization, and it can alter neuropeptide expression in a variety of neurons (for recent examples, see Rao et al., 1992; Higuchi et al., 1996; MacArthur and Eiden, 1996; Brandenburg et al., 1997; Brosenitsch et al., 1998). CGRP expression was examined with immunocytochemistry (Fig.1A). Few neurons maintained in NB5 medium for 8 d were CGRP-IR (8% ± 0.8). In contrast, CGRP immunoreactivity was detected in approximately one-third of the neurons in depolarizing NB50 (29.3% ± 1.7). The same proportion of CGRP-IR neurons was present in NB50 and in serum-containing media (p = 0.73), and depolarization in the presence of RS did not increase the percentage of CGRP-IR neurons further (p = 0.11). From these data we infer that depolarization and serum affected a single responsive population of DRG neurons. Neuronal survival in all conditions was ∼80% (Fig. 1B), suggesting that the changes in CGRP-IR neurons were attributable to CGRP regulation rather than to selective survival. To rule out the possibility that the increase in the percentage of CGRP-IR neurons in NB50 medium was attributable to an osmolarity change, we added 1% mannose to NB5 medium to produce a control medium with osmolarity empirically similar to NB50; in such mannose-containing medium, CGRP expression remained as low as that in basal NB5 medium (p = 0.98). Neurons cultured in different media maintained their pseudounipolar morphology (Fig. 1C), although their overall size and CGRP distribution varied. In depolarizing medium, neuronal perikarya were approximately the same diameter as in RS but were bigger than those in basal medium. In serum-containing cultures, CGRP immunoreactivity often was detected throughout the whole cell body and processes. However, the majority of neurons cultured in depolarizing NB50 had CGRP immunoreactivity in a perinuclear, Golgi-like distribution.
CGRP expression was induced by serum or depolarization. Dissociated lumbar E14.5 DRGs were plated in basal defined NB5 medium containing 5 mm KCl or in depolarizing NB50 medium containing 50 mm KCl and compared with cultures in rat serum (RS). After 8 d the CGRP-IR neurons were quantified. A, The percentage of CGRP-IR neurons. In either depolarizing medium or in serum a similar proportion of neurons expressed CGRP, and the effects of depolarization and serum were not additive. By contrast, few neurons expressed CGRP in NB5 medium or in the iso-osmolarity control. Asterisksindicate the conditions under which CGRP-IR neurons were equivalent (p > 0.1) and that also differed from basal NB5 values (p < 0.0001). B, Cell survival was good in all conditions, with ∼80% of neurons alive after 8 d. Four independent experiments with each variable in triplicate were performed for A and B.C, CGRP immunocytochemistry in cultures with NB5 (5), NB50 (50), andRS. Few neurons in NB5 were lightly CGRP-immunoreactive (CGRP-IR). In NB50, some neurons had intense staining in the cell body and particularly in the perinuclear region. Intense CGRP staining was observed in the whole cell body of some neurons cultured inRS. D, The proportion of CGRP-IR neurons increased with increased KCl concentration. The concentration of KCl was varied from basal levels (NB5) to medium containing 75 mm KCl (NB75); CGRP-IR neurons were counted after 8 d in culture. Five or six independent experiments with each concentration were performed in triplicate. Approximately 85% of the neurons survived in each condition.
To determine how much KCl was required to induce CGRP immunoreactivity, we varied the concentration of KCl in Neurobasal growth medium. Increases in CGRP-IR neurons were observed only when KCl was applied at a concentration of 25 mm or higher (Fig.1D). The maximal induction of CGRP expression was reached when cells were cultured in NB growth media containing 50 mm KCl. No further increase of the percentage of CGRP-IR neurons was detected at a KCl concentration of 75 mm. Because the maximal induction of CGRP expression in embryonic sensory neurons was reached with 50 mm KCl (Fig.1D), NB50 was used in subsequent studies.
Maintained depolarization was required for increased CGRP expression
To learn whether depolarization-induced increases in CGRP expression required a transient or prolonged signal or if neurons had a critical period in which stimulation must occur, we exposed cells to depolarizing NB50 medium for portions of the culture period and then switched to the basal NB5 medium, or vice versa. After a total of 8 or 10 d in culture, cells were fixed and stained for CGRP (Fig.2). No difference in cell survival was observed in any condition, such that ∼80% neurons present on day 1 were alive after 8 or 10 d in culture. The highest level of CGRP expression was observed in cultures maintained in NB50 for 8 d. At least 4 d in depolarizing medium were required to increase the proportion of CGRP-IR neurons above basal levels. The proportion of CGRP-IR neurons depended on the length of time in depolarizing medium and was not influenced by whether depolarization occurred early or late in the culture period. These data suggest that neurons maintained in basal medium for at least 4 d remained competent to respond to depolarizing signals delivered later in the culture period and that there did not appear to be a critical period during which depolarization had an effect.
Maintained depolarization was required to elicit CGRP expression. Dissociated E14.5 DRG cells were maintained in NB5 or NB50. Growth media were switched at days 2, 4, or 6. After a total of 8 or 10 d the cultures were processed for CGRP immunocytochemistry. The proportion of CGRP-IR neurons was quantified, and the data represent the mean and SEM of eight independent experiments. The highest proportion of CGRP-IR neurons was seen with 8 d of depolarization (chronic or applied after 2 d in NB5 had the same results; p = 0.73). CGRP expression decreased if NB50 was removed during the culture period.
Depolarization in vitro increased the amount of CGRP mRNA in DRG cultures (Fig. 3). A 4 d period was required for sufficient CGRP peptide to be present in the cells so that it could be detected by immunohistochemistry. CGRP mRNA is likely to be present before peptide can be detected. For that reason, to learn if CGRP mRNA was affected by treatments even before peptide changes were apparent, we assayed cultures at earlier times. CGRP mRNA was assayed by RT-PCR from DRG cultures grown in NB5 or NB50 and compared with the expression of the transcription factor elongation factor 1α mRNA. As expected, no CGRP mRNA was present in dissociated DRG from which total RNA was immediately extracted. However, CGRP mRNA was present in both depolarized and nondepolarized neurons after 2 d.
CGRP mRNA increased in DRG cultures with depolarization. Dissociated E14.5 DRG cells were placed in tissue culture in basal or depolarizing conditions. At the time indicated, total RNA was extracted and prepared for RT-PCR with primers specific for α-CGRP or the transcription factor elongation factor 1α (EF1α). In control cultures total RNA was extracted immediately (low K+, 0 d), and no CGRP mRNA was detected. After 2 d, CGRP mRNA was present in depolarizing (high K+, 2 d) and basal (low K+, 2 d) conditions. This result was obtained in three independent assays.
L-type calcium channel activity was necessary for depolarization-induced CGRP expression
Because depolarization leads to ionic fluxes across the plasma membrane, we examined the possibility that voltage-activated calcium or sodium channels were required for the CGRP response by adding specific pharmacological agents to NB50 (Table 1). DRG neurons express a variety of calcium and sodium ion channels (Fedulova et al., 1985, 1994; Scroggs and Fox, 1991, 1992; Ogata and Taebayashi, 1992; Elliott and Elliott, 1993). The addition of the L-type calcium channel blockers, 10 μm nifedipine or verapamil, for the entire culture period reduced the proportion of CGRP-IR neurons in depolarizing medium by 60 and 75%, respectively, suggesting that calcium influx is important for CGRP expression. The effect of nifedipine was concentration-dependent, with similar reductions in CGRP expression observed at 20, 10, and 5 μm nifedipine (unpaired t test,p > 0.1), but in control CGRP expression was observed at 1 μm nifedipine (p > 0.4987; from triplicate wells in two experiments). The N-type Ca2+ channel blocker ω-conotoxin reduced CGRP expression, but less effectively than L-type blockers. By contrast, the inhibition of sodium channels with tetrodotoxin had little or no effect on CGRP expression. It is important to note that neuronal survival remained high even in these drugs, indicating no overt effects on survival.
CGRP expression after depolarization required Ca2+ channel activity
Neurons had elevated intracellular calcium in depolarized cultures
To test that calcium was involved in increased CGRP expression under chronic depolarizing conditions, we visualized intracellular free calcium concentrations by ratiometric imaging with fura-2 AM (Fig.4). In general, intracellular free calcium levels were low in unstimulated cultures (NB5). Chronically depolarized neurons (NB50 at 4 d) had higher mean intracellular free calcium levels. The average intracellular calcium concentration in neurons cultured in NB5 for 4 d was 70 nm, whereas depolarized neurons had an average intracellular calcium concentration of 135 nm. The calcium elevation in depolarized cultures was inhibited completely by the L-type calcium channel blocker nifedipine (Fig. 4D).
Calcium imaging revealed that DRG neurons maintained in depolarizing NB50 had elevated free intracellular calcium. A, Mean free intracellular calcium increased with depolarization and was reduced to basal levels after nifedipine treatment. Free intracellular Ca2+ levels were measured in neurons at day 4. DRG cells were grown for 4 d in NB5,NB50, or NB50 with 10 μm nifedipine before they were loaded with fura-2 AM. Free [Ca2+] inside the cell was quantified. At least 300 neurons from three independent experiments were analyzed for each condition.B–D, Ratio histograms. The distributions of free calcium values from neurons cultured in NB5, NB50, or NB50 with 10 μm nifedipine for 4 d are shown in histograms. Most neurons in NB50 had ratios similar to those in NB5, whereas a distinct subpopulation had “higher Ca2+” levels.E, Free calcium levels were higher in a subpopulation of neurons cultured in NB50 for 4 d, shown in the ratio measurement with light blue pseudocolor. Similar observations were obtained in six independent experiments.
Depolarized neurons not only had an increased mean calcium level, but a subpopulation of depolarized DRG neurons at 4 d had fura-2 fluorescence ratios higher than any neurons observed in nondepolarizing media (Fig. 4C,E). The responsive “higher calcium” population with mean values above 0.2 μm was not the only group that could respond to depolarizing signals. When cultures in basal NB5 medium were depolarized with 50 mm KCl for 2 hr before being loaded with fura-2 AM, all neurons responded uniformly, and the mean intracellular free calcium increased to 597 nm. Thus, all neurons in these cultures were capable of responding to short-term depolarization, but only a subpopulation exhibited elevated intracellular calcium levels after chronic stimulation.
To test the hypothesis that these “higher calcium” neurons expressed CGRP, we grew neurons on gridded coverslips in NB50 medium for 4 d. The neurons were subjected to calcium imaging, followed immediately by immunocytochemistry for CGRP. E14 neurons grown for 4 d have only just begun to express CGRP, and the peptide is detectable by immunochemistry in approximately one-half the neurons that will eventually be CGRP-IR at 8 d in vitro (Hall et al., 1997). Nonetheless, CGRP-IR neurons were correlated with neurons that had higher free calcium levels (Fig.5). In a total of 737 neurons from four independent experiments, 63 were CGRP-IR and 181 were “higher calcium” neurons. Twenty-eight CGRP-IR neurons were in the “higher calcium” subpopulation. Thus, higher calcium neurons were twice as likely to be CGRP-IR as average calcium neurons (15.1% ± 2.7 SEM of high Ca2+ neurons were CGRP-IR, whereas 6.8% ± 1.4 SEM of non-high Ca neurons visualized at 4 d were CGRP-IR;p = 0.028).
Higher intracellular calcium levels were correlated with CGRP immunoreactivity. DRG cells were grown for 4 d in depolarizing NB50 on gridded coverslips, and free intracellular calcium was visualized by fura-2 fluorescence ratiometric imaging (A). Although many neurons had basal free calcium levels, some contained higher free calcium and appearedblue in pseudocolor (arrows). The cells were fixed and processed for CGRP immunocytochemistry (B), and the same neurons were located on gridded coverslips. Peptide expression was only just detectable at this early stage and appeared in a perinuclear, Golgi-like location in neurons. CGRP immunoreactivity was more likely to be found in neurons with elevated calcium (arrows).
By contrast, the addition of the calcium ionophore ionomycin did not increase CGRP immunoreactivity. Ionomycin (1 μm) application rapidly increased free calcium levels (mean free calcium concentration ∼360 nm) as compared with control cultures (∼50 nm), but chronic application of ionomycin for 2 or 8 d did not increase CGRP expression or affect survival (data not shown). These data support the notion that the route of calcium entry may be important for neuronal phenotypic differentiation (Ghosh and Greenberg, 1995).
Depolarization-induced CGRP expression was dependent on CaM kinase pathways
To begin to understand which signaling cascades were involved in the depolarization-induced increase of CGRP, we added agents that inhibit CaM kinases, PKA or PKC, to depolarizing NB50 medium. In general, neuronal survival was good in all drugs, although the PKC inhibitor BIM III (Table 2) reduced neuronal survival (p < 0.05). KN-62 (1 μm), which inhibits CaM kinase (Tokumitsu et al., 1990;Enslen et al., 1994), completely blocked the effects of 50 mm KCl on CGRP expression such that the percentage of CGRP-IR neurons dropped to basal levels. In fact, KN-62 concentrations between 0.1 and 2 μm were equally effective in blocking CGRP increases (p < 0.45). By contrast, H-89, a selective PKA inhibitor, did not have inhibitory effects at 2 μm (p > 0.05). This concentration of H-89 is sufficient to inhibit PKA activity in other studies (Findik et al., 1995). Although the PKC inhibitor BIM III reduced CGRP-IR neurons by ∼40% (p < 0.05), this effect may have been the result of differential neuronal survival. The concentrations of KN-62 and H-89 used in these experiments were close to or in excess of their pKi values of 0.9 and 0.048 μm (Chijiwa et al., 1990; Tokumitsu et al., 1990). These studies suggest that depolarization-mediated effects on CGRP induction are mediated via calcium and a calcium/calmodulin kinase signal transduction pathway.
CaM kinases were required for increased CGRP expression after depolarization
Whereas in some cases KN-62 can block L-type calcium channels (Li et al., 1992), the addition of this drug in short-term studies did not alter free intracellular calcium increases after depolarization. To determine whether KN-62 altered free calcium, we examined DRG neurons cultured for 2 d in nondepolarizing medium for free calcium changes after depolarization with NB50 and after depolarization in the presence of nifedipine or KN-62. The addition of KCl to raise the potassium concentration to 50 mm resulted in rapid increases in free calcium (average fluorescence ratio, 0.39 ± 0.17 SEM) in most neurons at 20 min after stimulation. The addition of nifedipine, which blocks L-type channels, reduced intracellular calcium levels (to average fluorescence ratios of 0.22 ± 0.01 SEM), whereas intracellular calcium fluorescent ratios with KN-62 were unaffected (0.35 ± 0.19 SEM; p < 0.0711). Thus, the reduction in CGRP expression with the addition of KN-62 resulted from an inhibition of a CAM kinase pathway rather than from an alteration in free calcium.
DISCUSSION
The same proportion of CGRP-IR neurons that develop in vivo (Hall et al., 1997) was elicited in vitro by stimulating embryonic sensory neurons with high potassium, suggesting that membrane depolarization and perhaps neuronal activity play critical roles in the initial development of this sensory neuron phenotype. This observation not only confirmed our previous finding that initial CGRP expression is independent of target tissues but also provides an important model system for investigating the cellular and molecular mechanisms that regulate de novo CGRP expression. Our analysis with pharmacological agents and calcium ratiometric imaging suggested that intracellular calcium changes and subsequent calcium-based signal transduction pathways were required for activity-induced CGRP expression in embryonic sensory neurons. An important aspect of this study is that it links depolarization with neuropeptide regulation via a CaM kinase pathway.
These experiments make no assumption about the order of action of particular signaling components and are intended to begin to identify “players” in the signaling cascade. Given the lag time between depolarization and the observed increases in CGRP immunoreactivity, a complex cascade of intervening events involving changes in second and third messengers may be very time-dependent.
Depolarization of developing sensory neurons in high potassium in vitro may mimic physiological processes. Sensory afferents of DRG neurons have a high level of activity during embryogenesis, but not at birth or in the adult (Fitzgerald, 1987). In vivo, lumbar DRG neurons begin to fire on E16, and the firing peaks at E18/19, just the time period in which the functional contacts have been established and CGRP first can be detected. The firing may be triggered by chemical stimuli from peripheral targets or membrane properties of the DRG neurons themselves (Scott, 1992).
In our study, Ca2+ entry via L-type calcium channels was important for depolarization-induced CGRP expression in embryonic sensory neurons. Voltage-dependent calcium channels, especially L-type channels, have a long open time and are implicated in increasing the expression of c-fos (Greenberg et al., 1986; Bading et al., 1993; Misra et al., 1994), tyrosine hydroxylase (Brosenitsch et al., 1998), nicotinic ACh receptor subunit genes (DeKoninck and Cooper, 1995), and GABA receptor genes (Gault and Siegel, 1997). Because L-type calcium channels require strong depolarization for activation, they usually are triggered experimentally in cultured cells by increasing extracellular potassium to 50 mm. In the present study, nifedipine and verapamil dramatically decreased the proportion of CGRP-IR neurons. The blockade of Ca2+ entry by nifedipine was confirmed by direct visualization of free intracellular Ca2+ with calcium imaging on cultures at day 4. Although their effects were dramatic, none of these calcium channel blockers showed complete inhibition of the depolarization-induced CGRP expression. One explanation for the partial effect by channel blockers is that the activation of CGRP expression by depolarization was not linked to a single type of voltage-activated Ca2+channel. Another possibility is that calcium-induced calcium release from intracellular stores contributes to the intracellular free Ca2+ level (Clapham, 1995). This notion may be unlikely, because ionomycin treatment did not increase CGRP. A third possibility is that a Ca2+-independent mechanism also contributes to CGRP induction after depolarization. Future studies using combinations of drugs or agents that deplete intracellular Ca2+ stores will differentiate among these possibilities. Although our data implicate calcium entry via L-type calcium channels in CGRP regulation, the activity of other calcium channels cannot be ruled out. Chronic depolarization by high potassium is likely to inactivate other channels, including the N-type channels. By contrast, L-type calcium channels have been implicated in the regulation of a number of neuronal genes (Higuchi et al., 1996; Gault and Siegel, 1997; Brosenitsch et al., 1998).
The elevation of intracellular calcium in DRG neurons maintained in depolarizing conditions may reflect calcium entry through voltage-sensitive or ligand-gated channels as well as calcium release from intracellular stores. DRG neurons also express a variety of receptor-mediated Ca2+ channels, such as NMDA receptors, that can be associated with depolarization-induced gene expression (Bading et al., 1993). However, the addition of 100 μm AP-V, a potent and selective NMDA receptor blocker, failed to inhibit CGRP expression (data not shown), suggesting that NMDA receptor was not involved in depolarization-induced CGRP expression in developing sensory neurons.
CGRP expression in depolarized sensory neurons was detected when the mean free intracellular Ca2+ level was elevated in the population. Indeed, a distinct subset of “higher Ca2+” neurons in depolarized cultures at day 4 was correlated with the CGRP-IR neurons just detectable in the population. Because the calcium measurements we have done revealed the intracellular Ca2+ level at the specific time these neurons were imaged, it is not clear whether this subset of “higher Ca2+” neurons differed uniquely from others by their electrical properties so that they were capable of maintaining high calcium levels or if they represented a changing subpopulation of DRG neurons that happened to have higher free calcium levels at that instant. It will be important to observe calcium changes over time in these cultures to learn whether neurons sustain increased calcium for long periods. Interestingly, no such “higher Ca2+” subsets were observed in depolarized cultures at day 2 (data not shown), further suggesting that calcium homeostasis changes over time in culture.
A common mechanism by which elevated intracellular Ca2+ regulates gene expression is by the activation of calmodulin. The calcium/calmodulin complex then binds and modulates multiple regulatory proteins, such as CaM kinases and calcineurin (Enslen and Soderling, 1994; Bito et al., 1996, 1997). KN-62, which abolished the increase in CGRP caused by depolarization, blocks CaM kinase IV and V as well as CaM kinase II (Tokumitsu et al., 1990;Enslen et al., 1994), a finding that suggests that membrane depolarization acts via CaM kinase pathways to initiate CGRP expression. One CaM kinase, CaM kinase IV, is localized in many DRG neurons (Sakagami et al., 1994) and is implicated in regulating gene expression (Sun et al., 1996). Previous studies on CaM kinase IV in embryonic and adult DRG indicated that it is detectable in rat DRG as early as E15 and is expressed preferentially in small neurons in sensory ganglion (Sakagami et al., 1994; Ji et al., 1996). However, CaM kinase IV expression is not well correlated with CGRP in adult DRG neurons and was not confined to CGRP-IR neurons in our RS-containing cultures (data not shown). It is important to recognize that pharmacological probes, but not direct biochemical enzyme assays, implicate these messengers in effecting CGRP expression in the developing DRG. Alternatively, changes in neuronal gene expression can be regulated by neuronal activity via the phosphorylation of cAMP response element-binding protein (CREB) by various Ca2+-dependent protein kinases, including CaM kinases (Misra et al., 1994; Bito et al., 1996). However, our preliminary immunocytochemical study with antibodies against pCREB did not reveal a correlation of pCREB with CGRP at 4 d (data not shown), suggesting either that the depolarization-induced CGRP expression is not mainly attributable to the activation of CREB or that pCREB-mediated events occurred at earlier times.
Stimulation by depolarization or serum resulted in the same proportion of CGRP-IR neurons with a similar onset of expression, suggesting that these stimuli are eliciting the phenotype in a competent population rather than inducing plastic populations of neuronal precursors to express CGRP. For this reason, to understand how that subpopulation becomes defined during development, we must address which factors restrict CGRP to one-third of DRG neurons. It is not clear which active agent in RS results in CGRP expression.
Although this study demonstrates changes in CGRP immunoreactivity that follow specific stimuli, it reflects steady-state CGRP amounts and does not identify which cellular processes limit the CGRP changes. In particular, it is not clear if depolarization directly increases transcription or if increases in CGRP immunoreactivity reflect changes in peptide processing. The amount of CGRP mRNA present after 2 d of stimulation is increased with high KCl, suggesting but not proving that transcriptional events are regulated by depolarization. It is unlikely that depolarization increases peptide storage, however, because depolarization generally causes peptide release. CGRP gene expression is regulated in part at the transcriptional level, as indicated by promoter analysis in thyroid C cell lines (deBustros et al., 1985, 1986; Haller-Brem et al., 1988; Naveh-Many and Silver, 1988;Lindsay and Harmar, 1989; Russo et al., 1992; Tverberg and Russo, 1992). The 5′ flanking sequence of the calcitonin/CGRP gene contains a cAMP-responsive element (CRE; Watson and Latchman, 1995) and an E-box/helix–loop–helix (HLH) enhancer region (Peleg et al., 1990;Tverberg and Russo, 1993). It is interesting to speculate that the E-box/HLH imparts depolarization sensitivity of CGRP expression, because this motif acts as a depolarization response element in other neuronal genes (Su et al., 1995; Higuchi et al., 1996; Walke et al., 1996).
In combination, these data demonstrate that CGRP expression during embryonic development of the DRG is sensitive to depolarization and raise the possibility that neuronal activity plays an important role in the differentiation of DRG neuronal phenotypes.
Footnotes
This study was supported by National Institutes of Health (NS-30842 and NS-23678) and the March of Dimes (FY98-0506). We thank Drs. Richard Zigmond and Ben Strowbridge for helpful critiques of this manuscript.
Correspondence should be addressed to Dr. Alison K. Hall, Department of Neurosciences, Case Western Reserve University, School of Medicine, Cleveland, OH 44106-4975.
