INTRODUCTION

The Ca²⁺-mobilizing pathway for G protein-coupled receptors involves multiple steps (1). Agonist binding causes the activation of G protein, resulting in the stimulation of phospholipase C and increased hydrolysis of phosphatidylinositol (4,5)bisphosphate, producing inositol 1,4,5-trisphosphate (IP₃)¹ and diacylglycerol. IP₃ binds to its receptor, a Ca²⁺ channel on the endoplasmic reticulum membrane, releasing Ca²⁺ from the lumen of the endoplasmic reticulum and increasing the intracellular free Ca²⁺ concentration ([Ca²⁺]_i).

Thyrotropin-releasing hormone (TRH) acts on receptors in lactotrophs and thyrotrophs of the anterior pituitary that are coupled to G_q/11 to cause a biphasic increase in [Ca²⁺]_i (2). The initial transient [Ca²⁺]_i spike, which is primarily due to release of intracellular Ca²⁺, is followed by a sustained [Ca²⁺]_i elevation that results from increased influx through L-type Ca²⁺ channels and capacitative Ca²⁺ influx. The initial [Ca²⁺]_i spike is terminated as IP₃ concentrations decline, intracellular Ca²⁺ stores become exhausted, and cytoplasmic Ca²⁺ is resequestered and pumped from the cell.

Desensitization is defined as a decline in the response to an agonist over time or a decline in the response to a subsequent agonist exposure. There are conflicting reports about whether the TRH response undergoes either form of desensitization in pituitary cells or following expression of the receptor in other cell types (3-9). The contradictory findings may be a consequence of differences in the methods used to assess the TRH response. In some cases, IP₃ mass has been measured at intervals after TRH has been given (5, 8, 9), whereas in others (6, 7), the rate of total inositol phosphate accumulation has been measured at different times after TRH has been administered to metabolically labeled cells. In other reports (2-4, 10, 11), only the [Ca²⁺]_i response has been followed. There is little information about the IP₃ response to repetitive applications of TRH, although it is well documented that the [Ca²⁺]_i response to high doses of TRH requires 5-20 min to recover (2). It is unclear whether the refractory period results from desensitization of the receptor or exhaustion of calcium pools.

In this study, we have carried out a detailed analysis of desensitization of the TRH response. To establish the molecular basis for desensitization, we have monitored IP₃ mass, [Ca²⁺]_i, the size of the intracellular Ca²⁺ pool, and the responsiveness to other Ca²⁺-mobilizing hormones in the continued presence of TRH and following the withdrawal and re-administration of TRH. We show that the TRH response undergoes profound desensitization over time and that the receptor is desensitized and unable to stimulate IP₃ production after TRH is withdrawn. The rate-limiting step in the recovery of the [Ca²⁺]_i response to TRH is refilling of the intracellular Ca²⁺ pools.

EXPERIMENTAL PROCEDURES

Hanks' balanced salt solution (HBSS) was purchased from Life Technologies, Inc. Thrombin, carbachol, endothelin, trichlorotrifluoroethane, and trioctylamine were purchased from Sigma. [³H]TRH, [³H](N³-methyl-His²)TRH, and kits for the detection of IP₃ were from NEN Life Science Products. Fura-2/AM and BAPTA were from Molecular Probes, Inc. (Eugene, OR), TRH and ionomycin from Calbiochem, cyclosporin from Sandoz Pharmaceuticals (East Hanover, NJ), and (N³methyl-His²)TRH from Bachem (Philadelphia, PA).

A HEK293 cell line stably expressing the wild-type mouse TRH receptor (301 cells) has been described previously (12). Cells were grown in Dulbecco's modified Eagle's medium supplemented with 5% fetal bovine serum as monolayer cultures at 37 °C in a humidified 95% air and 5% CO₂ environment.

Ca²⁺ imaging was carried out essentially as described by Nelson and Hinkle (13). All Ca²⁺ measurements were performed on cells in HBSS buffered to pH 7.4 with 15 mM HEPES. Cells plated on coverslips were loaded with 4 µM Fura-2/AM, 0.1% bovine serum albumin, and 1 µg/ml cyclosporin A in HBSS at room temperature for 50-60 min. The coverslip was washed and put into a Sykes-Moore chamber on a Nikon inverted microscope on a heated stage at 37 °C. The chamber was perfused with medium at 37 °C, and 340/380 nm fluorescence ratios were acquired every 1200 ms. The length of [Ca²⁺]_i imaging experiments was limited by gradual leaking of the Fura-2; similar difficulties were encountered with Fura-2/PE (data not shown).

To estimate the size of the intracellular Ca²⁺ pool, 1.5 mM BAPTA was added to the medium to chelate extracellular Ca²⁺ 30-60 s before 500 nM ionomycin was added. The increase in [Ca²⁺]_i stimulated by ionomycin was used as the measure of pool size. Ionomycin (500 nM) completely eliminated the TRH-induced increase in [Ca²⁺]_i measured 2.5 min later, indicating that the IP₃-releasable pool was completely empty. Conversely, pretreatment of cells with 1 µM TRH for 10 min almost eliminated subsequent ionomycin-induced Ca²⁺ release. Ionomycin did not increase [Ca²⁺]_i in cells that had been pretreated with thapsigargin, whether extracellular Ca²⁺ was present or not, indicating that extracellular Ca²⁺ entry was minimal.

Cells plated in 35-mm dishes were rinsed twice with HBSS and incubated at 37 °C with or without hormone as described below. At the end of the treatment, the medium was aspirated; 0.8 ml of 20% ice-cold trichloroacetic acid was added; and the dish was put on ice immediately. Cells were scraped off the dish, transferred to an Eppendorf tube, and pelleted at 12,000 × g for 1 min at room temperature. The supernatant was extracted with 2 volumes of trichlorotrifluoroethane/trioctylamine (3:1), and the aqueous phase was saved. The radioreceptor assay was performed according to the manufacturer's instructions, except that the reaction mixture was filtered through Whatman GF/C paper from NEN Life Science Products.

Cells plated in 35-mm dishes were washed twice with HBSS and incubated at 37 °C in buffer containing [³H]TRH or [³H](N³-methyl-His²)TRH with or without a 1000-fold molar excess of unlabeled hormone. Cells were then washed three times with HBSS.

RESULTS

The 301 cell line expresses ~200,000 TRH receptors/cell, with an apparent K_d of 10 nM (12). [Ca²⁺]_i was monitored in single 301 cells loaded with Fura-2. TRH stimulation produced a biphasic increase in [Ca²⁺]_i (Fig. 1), with an early transient peak and a later maintained phase. The initial [Ca²⁺]_i spike was abolished by thapsigargin treatment, indicating that the Ca²⁺ came from an intracellular pool, and the sustained [Ca²⁺]_i increase quickly subsided after the addition of the extracellular Ca²⁺ chelator BAPTA, indicating that it depended on the influx of extracellular Ca²⁺. The amplitude of the [Ca²⁺]_i spike depended on TRH concentration, reaching an apparent maximum at 1 nM (Fig. 1 and Table I). The [Ca²⁺]_i response occurred earlier and more synchronously, and the upstroke of the response was steeper at higher doses of TRH (Table I).

[View Larger Version of this Image (12K GIF file)]

Table I. Concentration dependence of TRH-mediated [Ca2+]i responses 301 cells were loaded with Fura-2 and treated with TRH. Single cell [Ca2+]i responses were recorded in experiments like those shown in Fig. 1. Values represent the means ± S.E. of 16-30 cells measured in experiments replicated 2-12 times. The peak increase in [Ca2+]i is defined as the difference between the 340/380 nm fluorescence ratios just before and the peak following TRH addition. The lag is defined as the time between the addition of TRH and the point where the 340/380 nm fluorescence ratio reached half of the peak value. The rate of [Ca2+]i rise is defined as the time between the start of the upstroke in the 340/380 nm fluorescence ratio and the point halfway to the peak. Peak increase in [Ca2+]i (340/380 nm ratio) Lag Rate of [Ca2+]i rise s s TRH 0.1 nM 1.22  ± 0.13 10.4  ± 0.41 0.9  ± 0.3 1 nM 5.39  ± 0.54 5.5  ± 0.13 0.7  ± 0.1 10 nM 4.13  ± 0.32 1.9  ± 0.06 0.2  ± 0.1 1000 nM 5.02  ± 0.37  a   a , Responses complete within time resolution of the system.

The intracellular concentration of IP₃ was measured in 301 cells at different times after the addition of TRH. The TRH-stimulated increase in IP₃ was also biphasic and strongly concentration-dependent (Fig. 2). The peak concentration of IP₃ occurred within 10 s at TRH concentrations of 10 nM or higher, and IP₃ increased ~15-fold at 1 µM TRH. In contrast to the [Ca²⁺]_i response, the IP₃ response increased with TRH concentrations between 10 and 1000 nM, the highest dose tested. The IP₃ concentration dropped rapidly within 1 min of TRH addition, but remained at more than twice the basal level for at least 10 min. At 1 nM, TRH caused no more than a 70% increase in the overall IP₃ concentration at times from 2 s to 10 min, and this increase was not highly significant (p = 0.07), even though the peak [Ca²⁺]_i response was essentially maximal under these conditions. The concentration dependence of the [Ca²⁺]_i and IP₃ responses is shown in Fig. 3.

[View Larger Version of this Image (20K GIF file)]

[View Larger Version of this Image (18K GIF file)]

Desensitization of the TRH response was measured by exposing cells to TRH, washing to remove free hormone, and then incubating the cells in hormone-free medium for various times before reintroducing TRH and measuring IP₃, [Ca²⁺]_i, and the intracellular Ca²⁺ pool size. Cells were first incubated with 1 µM TRH for 10 s or 10 min or with 1 nM TRH for 10 min. The amplitude of the initial [Ca²⁺]_i spike was the same in all three protocols, but the ability of the cell to respond to a subsequent challenge with TRH depended on both the dose and the duration of the first exposure. When cells were incubated with 1 µM TRH for either 10 s or 10 min, washed, and immediately challenged again with TRH, there was almost no IP₃ response (Fig. 4). Cells gradually recovered the ability to respond to TRH and gave a full IP₃ response by 40 min. The t1/2 for recovery was ~5 min after an initial 10-s incubation with TRH versus ~10 min after an initial 10-min incubation. Since the K_d for TRH is 10 nM, 1 nM TRH occupies only ~10% of receptors at equilibrium. Nonetheless, incubation with 1 nM TRH for 10 min (Fig. 4) significantly desensitized the IP₃ response to a subsequent challenge with 1 µM TRH. Immediately after withdrawal of 1 nM TRH, 1 µM TRH increased IP₃ to only 63% of the IP₃ level reached in naive cells.

[View Larger Version of this Image (14K GIF file)]

The ability of the cell to respond to a second challenge with TRH may be limited by how fast bound TRH dissociates from receptors following the first exposure to agonist. To measure ligand dissociation rates, cells were incubated with 1 µM [³H]TRH for either 10 s or 10 min and washed, and specifically bound [³H]TRH was followed over time (Fig. 5, left panel). The amount of [³H]TRH bound was the same after 10 s or 10 min of incubation, indicating that receptors were essentially saturated in both protocols. However, [³H]TRH dissociated faster from receptors after the 10-s incubation than after 10 min. An additional assay was used to measure the number of unoccupied receptors on the cell surface at various times after cells had been incubated with TRH and then washed (Fig. 5, right panel). Again, TRH receptors became available more rapidly when the initial incubation was brief. These data agree with previous findings in pituitary cells (14) and reflect the fact that the TRH-receptor complex internalizes extensively in 10 min in 301 cells.²

[View Larger Version of this Image (13K GIF file)]

Incubation of 301 cells with 1 µM TRH for 10 min completely abolished the [Ca²⁺]_i response to a second exposure to 1 µM TRH (Fig. 6). The [Ca²⁺]_i response recovered partially after 5 min, but remained at only 25% of the control response at 25 min, even though the IP₃ response was almost fully recovered at this point. When cells were initially exposed to 1 µM TRH for 10 s, the ability of the cells to mount an [Ca²⁺]_i response to 1 µM TRH was half of the control 2 min after washing and nearly 100% after 10 min (Fig. 6). Preincubation with 1 nM TRH for 10 min also prevented an [Ca²⁺]_i response to 1 µM TRH immediately after washing (Fig. 6), even though the IP₃ response was quite large. The [Ca²⁺]_i response recovered to ~70% of the control response by 10 min.

[View Larger Version of this Image (14K GIF file)]

The size of the intracellular Ca²⁺ pool was estimated by adding BAPTA to chelate extracellular Ca²⁺ and then adding 500 nM ionomycin to dump intracellular Ca²⁺ stores. The intracellular Ca²⁺ pool was almost completely depleted by 1 µM TRH within 1 min (Figs. 7 and 8). The pool was also fully depleted by 1 nM TRH, but in this case, only after 10 min (Fig. 7). Ca²⁺ pools refilled slowly, such that they were just 25% replenished after 10-25 min in cells that had been exposed to TRH for 10 min (Fig. 8). Refilling was faster and more complete in cells that had been exposed to 1 µM TRH for only 10 s (Fig. 8). Interestingly, the [Ca²⁺]_i response to TRH recovered before Ca²⁺ pools had refilled, indicating that partially full stores were adequate for a maximal response.

[View Larger Version of this Image (21K GIF file)]

[View Larger Version of this Image (13K GIF file)]

Because dye leakage became a problem in long experiments, we also measured Ca²⁺ pool sizes after loading with Fura-2/AM during the refilling period. In this protocol, Ca²⁺ pools seemed to recover fully within 1 h after cells had been incubated with 1 µM TRH for 10 min and washed. Ca²⁺ pool sizes, determined as the increase in 340/380 nm fluorescence ratios caused by ionomycin, were 2.6 ± 0.2 (n = 26) in control cells and 2.8 ± 0.2 (n = 56) and 3.2 ± 0.7 (n = 24) 1 and 2 h after removal of TRH, respectively.

The results described above imply that following TRH treatment, 301 cells should not be able to increase [Ca²⁺]_i in response to any other agonist until the Ca²⁺ stores are replenished. Thrombin, endothelin, and carbachol all increase [Ca²⁺]_i in 301 cells, although none of these agonists gives a response as large as TRH (data not shown). When combined into an agonist mixture, thrombin, endothelin, and carbachol stimulated large increases in IP₃ and [Ca²⁺]_i (Fig. 9 and Table II). The IP₃ response to the combined agonists occurred as quickly as the IP₃ response to TRH, but the peak was lower (7.5- versus 15-fold), and the IP₃ level fell to base line in 40 s, whereas it remained elevated for at least 10 min in the continued presence of TRH (Fig. 2 and Table II). The peak increase in [Ca²⁺]_i stimulated by the mixture was indistinguishable from that stimulated by TRH (Fig. 9 and Table II).

[View Larger Version of this Image (12K GIF file)]

Table II. Heterologous effects of TRH and an agonist mixture In experiments like those shown in Fig. 9, TRH was added to 301 cells, followed 10 min later by the combined agonist mixture, or vice versa. IP3 was measured immediately before and 10 s after the addition of hormones. Ca2+ pool size was measured immediately before the first hormone addition and 10 min later before the second. Values are the means ± S.E. of three dishes for IP3 or of 20-30 cells for [Ca2+]i and Ca2+ pool size. [Ca2+]i (340/380 nm ratio) IP3 Ca2+ pool, (340/380 nm ratio) Initial Peak Initial Peak pmol/mg cell protein TRH then mixture   0 min, TRH 0.13  ± 0.01 4.83  ± 0.53 6.70  ± 1.79 86.60  ± 0.25 7.99  ± 0.76   10 min, mixture 0.69  ± 0.07 1.00  ± 0.08 20.97  ± 0.04 50.49  ± 5.79 0.25  ± 0.03 Mixture then TRH   0 min, mixture 0.19  ± 0.01 5.58  ± 0.53 6.16  ± 0.19 44.87  ± 0.44 7.99  ± 0.76   10 min, TRH 0.21  ± 0.01 4.66  ± 0.63 8.03  ± 0.61 82.50  ± 5.30 2.14  ± 0.57

As predicted, when cells were first exposed to 1 µM TRH for 10 min and then to the mixture, the combined agonists did not increase [Ca²⁺]_i substantially, although they caused an IP₃ increase that was close to that in untreated cells (Fig. 9 and Table II). These results support the idea that Ca²⁺ pool depletion, not impaired IP₃ generation, prevents [Ca²⁺]_i responses to other hormones. The converse was not true. When cells were preincubated with the mixture for 10 min, they could still respond to TRH with a normal elevation of [Ca²⁺]_i, even though the Ca²⁺ pool was substantially reduced by the agonist mixture (Fig. 9 and Table II). These results again suggest that a full [Ca²⁺]_i response requires only a partly full intracellular Ca²⁺ pool. The IP₃ response to TRH, administered after the combined agonists, was close to that of control cells (Table II).

DISCUSSION

In this report, we have demonstrated that the TRH response undergoes profound desensitization. Restoration of the [Ca²⁺]_i response to TRH involves multiple steps, including ligand dissociation, recovery of receptor 1G protein 1 phospholipase C coupling, and intracellular Ca²⁺ pool replenishment. The characteristics of the IP₃ response measured in 301 cells expressing the TRH receptor closely resemble those reported previously for cell lines expressing an endogenous TRH receptor (5, 8, 9, 15). TRH stimulates a rapid and transient increase in IP₃ mass, which reaches a peak between 5 and 15 s (Fig. 2). In 301 cells, the dose dependence of the IP₃ and [Ca²⁺]_i responses to TRH was very different. A low dose of TRH (1 nM) produced an essentially maximal peak of [Ca²⁺]_i, but only a small increase in IP₃. A similar although less pronounced discrepancy in the dose-response curves for IP₃ and [Ca²⁺]_i has been reported previously for pituitary cells (3, 16, 17), and the activity of phospholipase C has been reported to increase with TRH doses up to 1-10 µM in isolated membranes (18). Rapid kinetic studies have shown that the [Ca²⁺]_i peak precedes the IP₃ peak (19). These findings all indicate that a very small increase in the average IP₃ concentration is sufficient to trigger a large increase in [Ca²⁺]_i. There are a number of possible explanations. Single cell Ca²⁺ imaging has shown that there is a highly variable delay between the addition of a low dose of TRH and the onset of the [Ca²⁺]_i rise (Table I) (2, 10). If there is a similar asynchrony in the TRH-mediated increase in IP₃, then the average IP₃ value will seriously underestimate the amplitude of a short-lived increase in IP₃ in individual cells. High concentrations of TRH stimulate a rapid and highly synchronous increase in [Ca²⁺]_i (Table I) (2, 13). Imaging studies have provided strong evidence for spatial heterogeneity in intracellular Ca²⁺ release (20), and biochemical evidence supports the existence of heterogeneous Ca²⁺ stores in pituitary cells responsive to TRH (21). The small amount of IP₃ generated by low concentrations of TRH may release Ca²⁺ from stores near the plasma membrane, and the resultant increase in [Ca²⁺]_i may sensitize IP₃ receptors (1, 20) or otherwise contribute to Ca²⁺ release as it spreads through the cell.

At the other extreme of the dose-response relationship, very high concentrations of TRH produced much higher levels of IP₃ than necessary for a maximal [Ca²⁺]_i spike. Very high levels of IP₃ are likely to alter the kinetics of pool refilling, even though they do not increase the size of the initial [Ca²⁺]_i spike. In addition, the peak Ca²⁺ concentration in the vicinity of the secretory granules may increase as the agonist dose is raised, even though the average Ca²⁺ concentration reported by fluorescent indicators does not.

Desensitization is defined as a diminishing response in the continued presence of an agonist or a diminished response to a subsequent exposure to an agonist. In this study, we documented both of these forms of desensitization. In the continued presence of TRH, IP₃ increased 15-fold and then fell rapidly to a plateau 1.5-3 times the basal level. This IP₃ response is typical of many receptors coupled to G_q (22-25). Gershengorn and co-workers found that the rate of total inositol phosphate production, measured over 30 min, decreases with time of exposure to TRH in pituitary cells (7) and in several other cell types (6), but not in HEK293 cells (6). The reason for the discrepancy is not known, but it may be because the transient IP₃ response to TRH was obscured in measurements done over a 30-min period or because the density of receptors was much greater when they were introduced by adenovirus-mediated gene transfer (6) rather than stable transfection, as in our work.

In principle, IP₃ concentrations could rise and then fall for several reasons. 1) Metabolism of IP₃ may be accelerated. 2) Substrate (phosphatidylinositol (4,5)bisphosphate) may be depleted. 3) Downstream kinases may turn off phospholipase C. 4) The receptor/G protein/phospholipase signal pathway may become uncoupled. Since TRH causes a biphasic increase in diacylglycerol as well as IP₃ and a burst followed by a gradual increase in total inositol phosphates (16), the activity of phospholipase C must change over time, not the metabolism of IP₃. The concentration of phosphatidylinositol (4,5)bisphosphate declines after TRH is added, but quickly recovers (26), making substrate exhaustion unlikely. Downstream kinases do not turn off phospholipase C generally because TRH caused little reduction in the IP₃ response to other G_q-coupled receptors. Activation of protein kinase C with phorbol esters does decrease the subsequent phospholipase C response to TRH (6, 7, 16), but this level of regulation is probably exerted later. The simplest explanation is that the TRH receptor became uncoupled.

It is not clear how the TRH receptor/G protein/phospholipase C cascade becomes uncoupled. Phospholipase C has GAP (GTPase-activating protein) activity (27) that may account for the rapid turnoff of G protein activation. However, since other agonists activated phospholipase C normally when the TRH response was uncoupled, it is necessary to postulate that the G proteins and phospholipase C coupled to the TRH receptor are spatially restricted if the GAP activity of the effector is responsible. An alternative, or additional, mechanism may involve uncoupling of the receptor from the G protein, possibly as a result of phosphorylation by a G protein-coupled receptor kinase or other downstream kinase (28). There is no evidence about whether the TRH receptor is phosphorylated by G protein-coupled receptor kinases or other kinases (28), but we have shown that a truncated form of the TRH receptor, which lacks probable G protein-coupled receptor kinase phosphorylation sites, does not undergo this form of desensitization effectively.²

When TRH was withdrawn from the medium, a second form of desensitization was observed, whereby the cell was refractory to re-addition of TRH for up to 30 min. The pattern of recovery is shown schematically in Fig. 10. Cells first recovered the ability to increase IP₃, but elevated IP₃ did not increase [Ca²⁺]_i because intracellular Ca²⁺ pools were almost completely depleted. The [Ca²⁺]_i response recovered ~10 min behind the IP₃ response and was maximal when Ca²⁺ pools were partially full. The Ca²⁺ pool eventually recovered fully. The consequence of depleted Ca²⁺ pools was heterologous desensitization of the [Ca²⁺]_i response. Although other agonists were able to evoke a nearly normal increase in IP₃ after TRH, they could not increase [Ca²⁺]_i because the intracellular Ca²⁺ pool was empty.

[View Larger Version of this Image (30K GIF file)]

Hormone dissociation is a necessary first step before a previously occupied receptor can be activated again. Ligand dissociation is not likely to be the only factor in desensitization of the IP₃ response, however. There was a sizable plasma membrane receptor pool available for activation shortly after a 10-s exposure to 1 µM TRH, but TRH could not stimulate phospholipase C activity effectively. A cell-surface receptor pool of similar size was present 25 min after exposure to 1 µM TRH for 10 min, and in this case, TRH could generate a nearly maximal IP₃ response.

Desensitization of the [Ca²⁺]_i response to agonists has been documented in many other studies (22-25). Anderson et al. (4) expressed rat and human TRH receptors in HEK293 cells and found that the [Ca²⁺]_i response mediated by the receptors was desensitized by high concentrations of TRH and did not recover at all, even in the face of normal Ca²⁺ pool refilling. The difference between previous results and ours may be due to the much higher concentration of TRH receptors in the transfected HEK293 cells used by Anderson et al. (4). In GH-cells and primary pituitary cell cultures, pool refilling does not seem to occur until TRH is withdrawn and then requires as long as 20 min (2).

Our study highlights the importance of regulation of the Ca²⁺ pool as a means of desensitizing receptor-mediated [Ca²⁺]_i responses. Pool replenishment was the rate-limiting step in recovery of the TRH response in 301 cells. The other agonists tested here did not deplete Ca²⁺ pools as thoroughly as TRH, even though the IP₃ responses were large. One of the reasons that TRH depletes the Ca²⁺ pool so thoroughly is that it stimulates efflux of Ca²⁺ from the cytoplasm, apparently by activating a plasma membrane Ca²⁺ pump (29). Nonetheless, all methods used to quantitate Ca²⁺ pool sizes are indirect. If TRH treatment changes cytoplasmic Ca²⁺ buffering capacity or Ca²⁺ reuptake, then the increase in [Ca²⁺]_i caused by ionomycin, used here to estimate pool size, might be altered (30), as might any localized release of Ca²⁺ from specialized domains on the endoplasmic reticulum membrane (31, 32).

In summary, we have shown that the signal transduction pathway to TRH becomes profoundly desensitized at multiple levels. In 301 cells, the IP₃ response to TRH undergoes homologous desensitization, but the [Ca²⁺]_i response undergoes heterologous desensitization because the TRH-treated cell cannot respond to any Ca²⁺-mobilizing agonist until pool refilling has occurred. Additional study is needed to identify the molecular mechanisms responsible for these complex levels of desensitization.