This report describes a Ca2+-dependent desensitization of the response to 5-HT1A receptor stimulation that occurs when the concentration of Ca2+ in the pipette solution was within the physiological range (~ 100 nM). After the desensitization was induced, the response to 20 nM 5-HT was reduced by 66% and the maximum inhibition could not be achieved even with higher concentrations of 5-HT.
was buffered <1nM, application of 100 nM 5-HT for 10 minutes resulted in a weak, transient and surmountable desensitization to challenge with 20nM 5-HT, but this was not evident when the cell was challenged with a [5–HT] ≥ 100 nM. The desensitization to a 5-HT agonist was heterologous since the effect of baclofen was also reduced (suppl. Fig 3
); this is discussed further below in relation to the mechanism of the effect. Desensitization and percent baseline inhibition of Ca2+
current by 5-HT were unrelated to the degree of Ca2+
current rundown (suppl. Fig 5
Ours results differed from those in cultured embryonic chick dorsal root ganglion (DRG) cells (Diverse-Pierluissi et al., 1996
, Tosetti et al., 2003
). In chick cells the desensitization was rapid and complete within 30-40 seconds, required Ca2+
influx through Ca2+
channels and was prevented when Ba2+
influx was measured in place of Ca2+
. In dr neurons, a fast component was not observed; the desensitization was not dependent on Ca2+
influx through the channels but on the basal level of intracellular Ca2+
. When Ba2+
was used as the charge carrier in place of Ca2+
it was capable of supporting the desensitization.
It is proposed that desensitization occurred in two steps. The first, a non Ca2+
-dependent conformational change in the 5-HT1A
R, as a result of a prolonged exposure to a high concentration of the agonist. In support of this, when internal Ca2+
was buffered to <1 nM, treating dr cells with high concentrations of 5-HT transiently decreased the cell sensitivity to 20nM 5-HT but not 10μM 5-HT. The next step required internal Ca2+
and appeared to involve a phosphorylation event mediated by PKA, leading to a more long-term desensitization (Raymond and Olsen, 1994
). The apparent decrease in 5-HT1A
R signaling strength could be due to i), a decrease in the number of 5-HT1A
receptors on the cell surface, ii), a decrease in the affinity between the 5-HT1A
R and its agonist or iii), a reduction in the coupling between the 5-HT1A
R and the G-protein. In supplemental figure 1
, when the G-protein coupled to the 5-HT1A
receptor, was activated with GTP-γ-S, the inhibition remained stable and did not desensitize over 10 minutes, with or without 5-HT. The desensitization induced by 5-HT could be greatly reduced with external Cd2+
which likely forms “metal bridges” simulating the external disulphide bridges in the receptor that prevent desensitization (Dohlman et al., 1990
, Rosati and Traversa, 1999
, Elliott et al., 2004
). Together these results suggest that the 5-HT1A
receptor itself shows an acute desensitization to the natural agonist. This suggests the following simplest scenario:
A + R ↔ AR ↔AR' (inactivated short-term) ↔AR'-P (inactivated long-term)
Where A= agonist, R= receptor in high affinity (G-protein bound) state, AR' = desensitized receptors having a conformation that is incapable of efficient coupling to the G-protein and P = phosphate.
Desensitization might occur through a well-established pathway of GPCR desensitization involving the binding of arrestin to the phosphorylated receptors, which leads to receptor internalization (Krupnick and Benovic, 1998
). Moreover, it has been shown that in dr neurons, treated with a high concentration of a 5-HT1A
agonist, the 5-HT1A
receptor internalized when observed at a 15 minute time point (Riad et al., 2001
) and this effect was blocked by the 5-HT1A
selective antagonist WAY100635. In the present study the longer lasting desensitization that we observed with 100 nM internal Ca2+
may result in receptor internalization.
The kinase that phosphorylates the 5-HT1A
receptor is likely to be PKA since PKAI5-24
partially antagonized the desensitization, and the adenylyl cyclase activator, forskolin, mimicked it. In contrast, conventional inhibitors of PKC were ineffective even though they blocked the effect of the phorbol ester PMA in dr neurons (Chen and Penington, 1996
). Some of our conclusions, based on kinase inhibitors, are supported by biochemical evidence and the inhibitors of PKA, PKC and phosphatases are considered to be highly selective (Cheng et al., 1986
, House and Kemp, 1987
, Herzig and Neumann, 2000
Inhibition of basal cAMP levels by 5-HT1A
receptors in the dr features in one theory of the action of antidepressant drugs (Honore, 2007
). This suggestion was used to explain an antidepressant-like phenotype in knockout mice lacking a cAMP-sensitive background leakage K+
channel called TREK1 normally found in the dr (Heurteaux et al., 2006
) and which is closed by increasing levels of cAMP. However, the usual negative coupling between the 5-HT1A
receptor and cAMP accumulation may not occur in dr neurons (Clarke et al., 1996
, Johnson et al., 1997
). Liu et. al. (Liu et al., 1999
) reported that in rat pituitary GH4C1 cells, depleted of Gαi1
and transfected with 5-HT1A
receptors, activation of the receptor could lead to activation of AC II or AC IV via the release of Gβγ
subunits (Andrade, 1993
, Tang and Hurley, 1998
). In the present work, physiological and biochemical evidence for PKA stimulation by 5-HT in dr neurons was obtained, although maximally stimulated PKA was inhibited by 5-HT (see below).
in the pipette solution, the desensitization to prolonged application of 5-HT was weak, transient and surmountable when the cell was challenged with 100nM 5-HT, but after inhibition of PP1/2A with microcystin the application of 5-HT resulted in desensitization to the same extent as when [Ca2+
was ~100 nM. Thus in the relative absence of [Ca2+
the 5-HT-induced desensitization appears to be limited by phosphatase activity. Our results reveal that 5-HT may both stimulate the kinase (not Ca2+
-dependent) and inhibit the phosphatase in a Ca2+
-dependent manner. A mechanism for the second possibility has been described in which the activity of PP1 can also be modulated by PKA (Walaas et al., 1983
, Winder and Sweatt, 2001
). Thus 5-HT-stimulated PKA activity could contribute to desensitization of the 5-HT1A
receptor by inhibiting phosphatase activity.
Activation of adenylyl cyclase by forskolin reduced the effect of 5-HT, but the effect of forskolin was Ca2+-dependent. Further, addition of microcystin to the pipette did not allow forskolin to decrease the effects of 5-HT in low Ca2+ pipette conditions (as it did with 5–HT) so the action of forskolin was not the result of indirect phosphatase inhibition. The agonist-induced desensitization was capable of inducing the first step of desensitization in the absence of Ca2+, and the full long-term effect was then revealed by phosphatase inhibition. However, in the absence of internal Ca2+, forskolin could not induce the first stage, and phosphatase inhibition was ineffective. Thus, different adenylyl cyclase isoforms are potentially involved in the effects of forskolin and 5-HT.
The inhibition by an agonist of maximal cAMP-induced PKA activity is not a “physiological” measure, and the stimulation by an agonist of basal PKA activity is perhaps more meaningful. It was expected that stimulation of the basal level of PKA activity would occur after treatment with DP-5-CT with no change in the total activity (activity ratio increases). In the present study, there was a significant increase in the AR of PKA by 80 % caused by the 5-HT1A agonist; however this was brought about by a significant decrease in the total activity and an average stimulation in basal levels of PKA activity with DP-5-CT. These effects were mediated by stimulation of the 5-HT1A receptor since it was completely antagonized by incubation with WAY100635.
The decrease in the “total” PKA activity after treatment with the 5-HT agonist suggests that a saturating concentration of cAMP was not capable of producing as many free cPKA subunits. However, the increase in AR with the 5-HT agonist suggests that the percentage of total PKA in the slice that was active under basal conditions was greater than control after treatment with the 5-HT agonist. Since there appears also to be a stimulation of basal PKA activity there could be a differential effect of 5-HT receptor stimulation on the two main types of PKA (I and II isoforms).
It should be noted that the biochemical measurement of stimulation of PKA activity and inhibition of phosphatase activity observed in this study was a population response from the whole dr nucleus which contains several different types of cells. However a similar effect was also seen when recording from individual serotonergic cells isolated from the nucleus suggesting that the population response is a signal from the serotonergic neurons in the dr nucleus. Application of serotonin also desensitized the response to baclofen; this means that the desensitization was heterologous. This type of desensitization, rather than one specific for the 5-HT1A receptor, would be expected given the mechanism of the desensitization, since activation of a freely diffusible messenger such as cAMP would be expected to influence the response to several transmitters and not just the 5-HT1AR. It would be interesting to determine whether the application of baclofen would also desensitize the response to 5-HT although whether every G-protein coupled transmitter would activate this mechanism would presumably depend on the Ca2+-sensitivity of the individual transduction mechanism.
The effect that we have described should have physiological consequences. A long burst of spikes in the dr could expose the raphe cells to at least 100 nM 5-HT, perhaps remaining in the synaptic space long enough to desensitize the 5-HT receptors. The response of the neurons to small amounts of 5-HT, producing IPSPS (Aghajanian and Lakoski, 1984
, Williams et al., 1988
) and inhibiting Ca2+
influx during the action potential (Penington et al., 1992
), would be depressed for some time. This effect may be surmountable only if [Ca2+
is low enough, and a large amount of 5-HT was released, so that strong stimulation would continue to produce profound inhibition of the cell bodies and prevent action potential generation. The relationship between the acute effects that we have described and the effect of chronically flooding raphe neurons with 5-HT, such as occurs after prolonged application of re-uptake blockers, is a subject worthy of further investigation.