A functional comparison of recombinant and native somatostatin sst₂ receptor variants in epithelia

Abstract

Background and purpose.

Somatostatin (SRIF-14) exerts broad spectrum antisecretory effects by activating the somatostatin 2 (sst₂) receptor. The rat (r) sst₂ receptor exists in ‘long' (sst_2a) and ‘short' (sst_2b) forms that differ in their C termini, while a single human (h) sst_2a exists. This study compares the characteristics of recombinant rsst_2a, rsst_2b and hsst_2a activation in human epithelia, and with native sst₂ responses in rat colon.

Experimental approach.

Epithelial layers of each clone or rat colon were placed in Ussing chambers and short-circuit current (I _SC) measured in response to SRIF-14 and chosen analogues. The relative potencies and ability to cause desensitization to SRIF-14 were assessed, and the affinities of the sst₂ antagonist, D-Tyr⁸ CYN154806 for hsst_2a, rsst_2a and native rat colon sst₂ receptors were established.

Key results.

Basolateral SRIF-14 responses were transient in hsst_2a and rsst_2a epithelia, but prolonged in rsst_2b-expressing cells. Activation of rsst_2a resulted in significant desensitization to SRIF-14 and receptor phosphorylation, whereas the rsst_2b receptor did neither. Sst₂-preferred agonists (BIM23190C and BIM23027) reduced I _sc with similar potency and both caused complete desensitization to SRIF-14. CYN154806 antagonized hsst_2a and rsst_2a receptors with pK _B values of 7.9 and 7.8, respectively. In rat colon mucosa, CYN154806 blocked SRIF-14 responses with a pA ₂ value of 8.2, and BIM23190C responses with a pK _B of 8.4.

Conclusions and implications.

SRIF-14 caused rapid rsst_2a receptor phosphorylation and desensitization of epithelial antisecretory responses, neither of which occurred with the rsst_2b receptor. These mechanisms are most likely to be a prerequisite for sensitivity to sst₂-analogues with radiotherapeutic potential.

Introduction

SRIF-14 (somatotrophin release inhibiting factor, or somatostatin-14) is an enteric neuropeptide (Schultzberg et al., 1980; Ekblad et al., 1988) and endocrine cell product (Alumets et al., 1977) with broad spectrum antisecretory effects that include inhibition of gastric acid (Lloyd et al., 1995), intestinal electrolytes (Eklund et al., 1988; Knobloch et al., 1989; Ferrar et al., 1990) and endocrine secretions, for example, growth hormone, glucagon and insulin (for a review, see Weckbecker et al., 2003). These inhibitory actions are mediated by somatostatin 2 (sst₂) receptors, one of five cloned sst receptor types that couple to pertussis toxin-sensitive G_i/o proteins and reduce adenylate cyclase activity (Siehler and Hoyer, 1999b) and can also modulate ion channels, for example, opening neuronal K⁺ channels (Hicks et al., 1998) or closing voltage-dependent Ca²⁺ channels (Kleuss et al., 1991) thereby reducing neuron excitability.

Two sst₂ receptor spice variants are produced in mouse (Vanetti et al., 1993) and rat (r) tissues, alternative splicing occurring via a cryptic splice site within the coding sequence of the sst₂ receptor gene (Schindler et al., 1998). Both rodent species express an sst_2a receptor variant that is 23 amino acids longer than sst_2b, but with different C termini (for a review, see Cole and Schindler, 2000). Activation of the murine (m) short isoform, msst_2b, inhibited adenylate cyclase with greater efficacy and resulted in significantly less agonist-induced desensitization than the msst_2a receptor (Vanetti et al., 1993). In contrast, no differences were observed between rsst_2a and rsst_2b signalling or desensitization after exposure to SRIF-14 (Schindler et al., 1998). The clearest discrimination between these two receptors' signalling capacities was described for the opposing actions of SRIF-14 on CHO cell proliferation; rsst_2a-expressing cells exhibiting an inhibition, and rsst_2b cells, a stimulation of peptide-induced proliferation (Alderton et al., 1998). In contrast with the proven production of two different sst₂ isoforms in rat and mouse, only the human (h) sst_2a has been identified as a functional protein, the hsst_2b orthologue remaining putative (see Cole and Schindler, 2000).

The distribution patterns of the two sst₂ variants differ in mouse and rat nervous systems, sst_2a predominating in the central and peripheral nervous systems (Sarret et al., 1998; Cole and Schindler, 2000; Schulz et al., 2000). In the gastrointestinal tract, differences are also observed. In rat stomach and small intestine, for example, the sst_2a receptor protein is expressed by neuroendocrine cells and enteric nerves (implicating indirect, as well as direct mechanisms of SRIF-14 action), while sst_2b receptor labelling is found in a discrete subpopulation of rat parietal cells (Schindler and Humphrey, 1999). Warhurst et al. (1996) observed both sst₂ receptor transcripts in rat colonic crypt extracts and with equal abundance. In the mouse stomach, parietal and endocrine cells express sst₂ receptors, as do subpopulations of myenteric and submucous neurons, immunostaining for sst_2a being colocalized with nitric oxide synthase immunoreactivity in inhibitory myenteric motor neurons (Allen et al., 2002). In the human gastrointestinal tract, sst_2a is also expressed by endocrine cells (particularly, gastrin-containing cells in the small intestine), as well as myenteric and submucous neurons along the length of the intestine (Gugger et al., 2004) and this pattern is similar to that reported in the rat gastrointestinal tract (Schindler and Humphrey, 1999).

The extensive inhibitory nature of SRIF-14 actions stimulated early interest in the peptide for therapeutic benefit, for example, as a novel treatment for diabetes, of hormone-secreting tumours and hypersecretory diarrhoea, but the relative instability of plasma SRIF-14 was limiting. One of the first longer-acting cyclic SRIF-14 fragment analogues, octreotide (SMS 201-995, Sandostatin) is used, for example, to treat acromegaly, to prevent complications following pancreatic surgery and relieve symptoms such as chemotherapy-induced diarrhoea (Lamberts et al., 1996). Many other more stable cyclo-octapeptide analogues have since been produced with therapeutic, as well as diagnostic potential (Weckbecker et al., 2003) and some of these have been used in the present study, specifically, BIM23027 (amino acid sequence is as listed in McKeen et al. (1995)), BIM23190C, BIM23014C (Lanreotide) and BIM23268, together with the linear analogues BIM23052 and BIM23056 (amino acid sequences are as listed in Shimon et al. (1997)). Sst receptor-mediated endocytosis has also been utilized clinically to deliver stable radiolabelled SRIF-14 analogues into tumour cells that express sst₂ receptors, thereby allowing metastases to be imaged (Breeman et al., 2001) and treated by receptor-targeted radiotherapy (Weckbecker et al., 2003).

The present study set out to determine whether SRIF-14 activation of recombinant hsst_2a, rsst_2a and rsst_2b expressed in epithelial cells, differed (i) in their response time courses and pharmacology, (ii) in their agonist-induced phosphorylation and desensitization and (iii) whether their pharmacologies differed from that of native sst₂ antisecretory responses in rat colon mucosa. Two SRIF-14 analogues were chosen for their sst₂ affinity, namely, BIM23190C and BIM23027 (McKeen et al., 1995; Shimon et al., 1997; Siehler et al., 1999; Weckbecker et al., 2003). Also included were other SRIF-14 analogues known to stimulate sst₅ receptors (BIM23268 and BIM23052, Shimon et al., 1997; Weckbecker et al., 2003) and the nonselective sst₅/sst₂/sst₃ analogue, BIM23056 (Shimon et al., 1997; Siehler and Hoyer, 1999a, 1999b; Weckbecker et al., 2003) shown also to be an sst₅ antagonist (Wilkinson et al., 1996). The affinity of the selective sst₂ antagonist, D-Tyr⁸-CYN154806 (Ac-(4-NO₂-Phe)-cyc(DCys-Tyr-DTrp-Lys-Thr-Cys)-DTyr-NH₂) (Feniuk et al., 2000; Nunn et al., 2003) for each recombinant sst₂ receptor was also determined and the antagonist used to confirm the predominant involvement of this receptor type in SRIF-14 responses in rat colon mucosa.

Methods

Cell culture and transfection

Colony 1 adenocarcinoma cells (from Dr S Kirkland; Marsh et al., 1993) were incubated in DMEM supplemented with 10% foetal calf serum, 100 μg ml⁻¹ kanamycin and 1.2 μg ml⁻¹ amphotericin B. Cell lines were grown at 37°C in a humidified atmosphere of 95% O₂/5% CO₂ and passaged when confluent by trypsinization (0.25% in versene). Stably transfected clones were generated by calcium phosphate co-precipitation followed by glycerol shock, using cDNA sequences encoding human sst_2a (in pTEJ8, provided by Professor T. Schwartz, Panum Institute, Copenhagen, Denmark), and the rat sst_2a or sst_2b splice variants (N-terminal haemagglutinin (HA) epitope-tagged constructs in pcDNA3.1, from Dr M Schindler; Schindler et al., 1998). Colonies resistant to the antibiotic G418 (1.0 mg ml⁻¹) were isolated directly, expanded and screened for sst₂ receptor expression in short-circuit current (I_SC) studies.

Animals

Male Sprague–Dawley rats (200–250 g, from Banton and Kingman, Hull, UK) were maintained in a 12 h light–dark cycle, with access to standard chow and water ad libitum. The descending colon was removed from rats killed by cervical dislocation, and placed in oxygenated Kreb's Henseleit (KH) buffer (constituents in mM: NaCl 118, KCl 4.7, NaHCO₃ 25, KH₂PO₄ 1.2, MgSO₄ 1.2, CaCl₂ 2.5, glucose 11.1; pH 7.4) until dissection.

Short-circuit current studies

Colony 1 epithelial layers were grown to confluence (area 0.2 cm²) on collagen-coated Millipore filters, bathed at 37°C in oxygenated KH and voltage-clamped at 0 mV in Ussing chambers (DVC1000, WPI, Stevenage, UK) as described previously (Holliday et al., 2005). The resulting I_SC was elevated by a maximal concentration of the secretagogue, vasoactive intestinal polypeptide (VIP) (30 nM, 20 min) before basolateral addition of sst₂ ligands. Unless otherwise stated, agonist concentration–response relationships were constructed from single peptide additions. For the determination of CYN154806 IC₅₀ and pK_B values, epithelial layers were pretreated for 10 min with the antagonist before SRIF-14 application. At the end of each experiment, UK14,304 (1 μM) and piretanide (200 μM) were included as inhibitory controls.

Mucosal sheets from rat-descending colon (0.6 cm² area) were voltage-clamped at 0 mV in oxygenated KH, as described previously (Cox et al., 1988). Concentration–response curves to SRIF-14 and other analogues were constructed by cumulative peptide additions to the basolateral reservoir. Changes to I_SC levels were recorded continuously. Tetrodotoxin (TTX; 100 nM) was used to inhibit neuronal activity (as the submucous neuron innervation is intact in these preparations) before addition of the sst₂ antagonist, CYN154806 10 min later. For the determination of CYN154806 pA₂ value in mucosal preparations, the antagonist was added 10 min before the first agonist addition.

Phosphorylation measurements

Colony 1 clones were grown to 80% confluence in six-well plates, loaded with 50 μCi H₃PO₄ in phosphate-free KH buffer for 1 h at 37°C and treated with vehicle or 10 μM SRIF-14 for 5 min. HA-tagged sst₂ receptors were then immunoprecipitated as described previously (Holliday et al., 2005). Briefly, cells were dissolved (2 h at 4°C) in RIPA buffer (50 mM Tris, 100 mM NaCl, 10 mM NaF, 10 mM Na₄P₂O₇, 5 mM EDTA, 1.5% Nonidet P40, 0.5% sodium deoxycholate, 0.2% sodium dodecyl sulphate (SDS), 0.5 mM phenylmethylsulphonyl fluoride, 200 μM activated Na₃VO₄, 100 nM okadaic acid, 10 μg ml⁻¹ leupeptin and aprotinin; pH 8.0) and samples were equalized for protein content (BCA protein assay, Pierce, Cheshire, UK). Immunoprecipitations were carried out overnight at 4°C by addition of anti-HA antibody (rat clone 3f10, Roche Molecular Biochemicals, Lewes, UK) directly conjugated to agarose, and then the washed precipitates were denatured in Laemmli loading buffer (80°C, 3 min). Proteins were resolved by SDS-polyacrylamide gel electrophoresis (PAGE; 10% Tris-HCl Ready Gels; Biorad, Hemel Hempstead, UK) and the dried gels were exposed to pre-flashed Amersham Hyperfilm MP for 72 h at –70°C to detect ³²P labelling. To ensure equivalent receptor loading, immunoprecipitates resolved by SDS-PAGE were also transferred to polyvinyldifluoride membrane and probed overnight (18°C) with the 3f10 anti-HA antibody (100 ng ml⁻¹ in TBST (50 mM Tris, 150 mM NaCl, 0.1% Tween 20; pH 7.5) containing 1% BSA and 0.02% NaN₃). Western blots were developed using a goat anti-rat horseradish peroxidase-conjugated secondary antibody (1:5000 in TBST for 60 min; GE Biosciences, Little Chalfont, UK) and enhanced chemiluminescence detection (ECL plus; GE Biosciences).

Data analysis

Pooled data from I_SC studies present the maximal changes in I_SC as μA cm⁻², mean±1 standard error of the mean (s.e.m.). Agonist pEC₅₀ values and antagonist pIC₅₀ values were obtained by nonlinear iterative fits to the combined data using GraphPad Prism (version 3.03, GraphPad Software Inc., San Diego, CA, USA). Affinity estimates for CYN154806 were obtained from the Gaddum equation (pK_B, for epithelial layers) and Schild analysis (pA₂, for rat colon studies). Statistical comparisons of two data sets were performed by use of Student's t-test, while multiple comparisons were obtained by one-way analysis of variance with Dunnett's post-test.

Materials

Cell culture materials were from the following origin: DMEM and G418 sulphate (Invitrogen, Paisley, UK); foetal calf serum, kanamycin and amphotericin B (ICN Biomedicals, Oxford, UK); trypsin (Lorne Laboratories, Reading, UK). H₃³²PO₄ (10 mCi ml⁻¹) was from Amersham Biosciences (Little Chalfont, Bucks, UK). Agonists BIM23104C, BIM23190C, BIM23052, BIM23056 and BIM23268 (see Shimon et al., 1997 for sequences) were provided by Biomeasure Inc. (Milford, MA 01757, USA); while D-Tyr⁸-CYN154806 and BIM23027 (McKeen et al., 1995) were kind gifts of Dr W Feniuk (Glaxo Institute of Applied Pharmacology, Cambridge, UK). Other peptides were purchased from Bachem (Merseyside, UK); all were stored as single use frozen aliquots of aqueous solution. Piretanide was obtained from Hoechst Marion Roussel (Swindon, UK), and other reagents were from Sigma-Aldrich (Poole, UK) or VWR International (Poole, UK). UK14,304 (5-bromo-N-(4,5-dihydro-1H-imidazol-2-yl)-6-quinoxalinamine) was prepared as a 10 mM solution in dimethylsulphoxide; other chemicals were made up as aqueous stock solutions.

Results

Epithelial responses to recombinant sst₂ receptors

We isolated stably transfected Colony 1 cell lines expressing the hsst_2a (hS2a), rsst_2a (rS2a) or rsst_2b (rS2b). These clones exhibited higher basal resistances than non-transfected Colony 1 cells in I_SC studies, but in each case VIP-stimulated sustained elevations in I_SC with similar potencies (pEC₅₀ range: 8.08–8.48, data not shown), reflecting electrogenic chloride secretion. Moreover, VIP responses in Colony 1, hS2a, rS2a and rS2b epithelial layers were all inhibited by subsequent UK14,304 activation of endogenous α₂ receptors (Figure1 and Table 1), or by blockade of basolateral Na⁺/K⁺/2Cl⁻ co-transport using piretanide (Figure 1). However, in contrast to non-transfected host cells, basolateral SRIF-14 decreased VIP-stimulated I_SC in hS2a, rS2a and rS2b clones with equivalent potency (pEC₅₀ range: 7.64–7.86) and similar levels of inhibition, that is, 29–43% of the VIP response after 100 nM SRIF-14 (Figure1 and Table 1). SRIF-14 (100 nM) also inhibited basal I_SC levels, by −13.0±2.8 μA cm⁻² (n=4) in rS2a cells and by −13.0±3.8 μA cm⁻² (n=4) in the rS2b epithelial clone. Apical SRIF-14 addition (100 nM) had little effect on VIP-elevated I_SC in rS2a (−1.1±0.4 μA cm⁻², n=4) or rS2b epithelial layers (−3.5±0.5 μA cm⁻², n=3), indicating preferential targeting of both sst₂ receptor variants to the epithelial basolateral membrane.

Figure 1.

Representative I_SC recordings from Colony 1 hS2a, rS2a and rS2b clones. Confluent epithelial layers were stimulated with 30 nM vasoactive intestinal polypeptide (which produced a sustained rise in I_SC) then by 100 nM SRIF-14 (two additions 20 min apart), 1 μM UK14,304 (UK) and 200 μM piretanide (Piret) as indicated. Initial I_SC levels (in μA) are given to the left of each trace. Note the smaller μA scale for the hS2a trace. SRIF-14, somatotrophin release inhibiting factor; UK14,304, 5-bromo-N-(4,5-dihydro-1H-imidazol-2-yl)-6-quinoxalinamine; VIP, vasoactive intestinal polypeptide.

Table 1.

Electrophysiological parameters and responses to basolateral stimuli in untransfected Colony 1 cells and sst₂ receptor clones

	Basal R	Basal ISC	VIP 30 nM ΔISC	SRIF-14		UK14,304
Cells	(Ω cm2)	(μA cm−2)	(μA cm−2)	pEC50	100 nM ΔISC (μA cm−2)	1 μM ΔISC (μA cm−2)
Colony 1	39.3±0.6 (531)	11.0±0.4 (531)	+29.5±1.2 (200)	—	0.0±0.0 (14)	−10.4±0.9 (4)
hS2a	131.2±4.1 (125)	9.9±4.9 (125)	+22.1±1.0 (103)	7.64±0.07	−9.6±1.1 (8)	−5.8±0.5 (6)
rS2a	100.1±2.5 (237)	50.6±1.7 (237)	+73.2±1.8 (189)	7.67±0.07	−22.4±2.2 (6)	−25.3±3.4 (6)
rS2b	63.4±2.8 (126)	25.2±1.1 (126)	+83.0±3.3 (105)	7.86±0.07	−24.3±4.4 (6)	−19.5±3.6 (6)

Abbreviations: SRIF-14, somatotrophin release inhibiting factor; UK14,304, 5-bromo-N-(4,5-dihydro-1H-imidazol-2-yl)-6-quinoxalinamine; VIP, vasoactive intestinal polypeptide.

pEC₅₀ values for SRIF-14 were calculated from pooled single addition concentration–response relationships for the inhibition of 30 nM VIP-stimulated I_SC (n=3–8).

Values in parentheses indicate the number of observations for basal resistances (R) and I_SC, and for the change in I_SC (ΔI_SC) after each agonist addition at the optimal concentrations shown.

Sst_2a and sst_2b receptor desensitization

Figure 2a illustrates the time-profiles of sst_2a and sst_2b-mediated I_SC responses following basolateral SRIF-14 addition. In both hS2a and rS2a cells, SRIF-14 responses were transient, with increased agonist concentrations leading to both a reduction in the time-to-peak (1–1.5 min after 300 nM SRIF-14), and a more rapid return to I_SC levels before sst_2a stimulation (3–4 min after 300 nM). In contrast to the short-lived nature of rat and human sst_2a responses, the inhibition of VIP-elevated I_SC by SRIF-14 in rS2b epithelial layers was more sustained (Figure 2a). For example, responses at 10 min after 300 nM SRIF-14 had decayed to 47.2±3.3% (n=4) of the peak levels at 2.5–3 min, compared to only 6.4±1.8% (n=5; P<0.001) in rS2a cells.

Figure 2.

Desensitization of sst_2a and sst_2b receptors. Time courses from hS2a, rS2a and rS2b cells (a) show the reductions in vasoactive intestinal polypeptide-stimulated I_SC following SRIF-14 addition (t=0), at 3 nM (n=3–4), 30 nM (n=3–4) or 300 nM (n=3–5). SRIF-14 concentration–response relationships in rS2a and rS2b epithelial layers (b) were constructed from responses to single agonist concentrations (n=3–8), or to cumulative agonist additions (n=3–4). Sigmoidal fits to the pooled data (excluding the 444 nM data point for rS2a-cumulative responses) yielded the pEC₅₀ values and maximal SRIF-14 responses given in Table 1 and the text. In (c), immunoprecipitated HA-tagged rat sst_2a and sst_2b receptors were resolved by SDS-PAGE and transferred to polyvinyl difluoride membrane for western blotting with anti-HA (left hand photograph). Sst₂ proteins were identified as broad bands of 58–83 kDa (sst_2a) and 53–83 kDa (sst_2b), in addition to bands corresponding to the light (25 kDa, data not shown) and heavy (IgG_H) chains of the immunoprecipitating antibody. To measure phosphorylation, HA-tagged receptors were immunoprecipitated under the same conditions, from rS2a and rS2b cells labelled with ³²P_i and treated with vehicle or 10 μM SRIF-14 for 5 min. The autoradiograph (right, 72 h) of the dried SDS-PAGE gel indicates agonist-induced phosphorylation of rat sst_2a, but not sst_2b receptors. SDS-PAGE, sodium dodecyl sulphate-polyacrylamide gel electrophoresis; SRIF, somatotrophin release inhibiting factor

We next compared rS2a and rS2b concentration–response curves constructed from single agonist additions, with cumulative relationships in which sequential applications of agonist were made at the peak of the previous SRIF-14 response (Figure 2b). The cumulative data for rS2b clones yielded a SRIF-14 pEC₅₀ of 7.83±0.04 and a maximal I_SC response of −21.8±3.4 μA cm⁻² (n=4, 144 nM), similar to values obtained from single peptide additions (Table 1). However, rS2a cumulative SRIF-14 responses were bell-shaped, with a 10-fold higher pEC₅₀ for the inhibitory portion of the curve (8.55±0.08), and a maximal response at 44 nM (−7.0±0.6 μA cm⁻², n=4) that was only 31% of the peak I_SC decrease to 100 nM SRIF-14 (Table 1).

To measure phosphorylation, the HA-tagged rat sst_2a and sst_2b receptors were immunoprecipitated from epithelial clones loaded with ³²P_i, and incorporation of radiolabelled phosphate was determined by gel autoradiography. These experiments were technically challenging, involving the use of a high agonist concentration (10 μM SRIF-14 for 5 min) to overcome the diffusion barrier to the basolateral epithelial domain, and limited by variable background in the autoradiographs from the presence of genomic DNA contamination in the samples. Western blots probed with anti-HA revealed specific broad bands in immunoprecipitates from rS2a (58–83 kDa) and rS2b (53–83 kDa) epithelia (Figure 2c), which were absent in samples from non-transfected Colony 1 cells. Despite the similar loading of receptor proteins, ³²P_i labelling of an equivalent band was observed only in rS2a immunoprecipitates (autoradiograph, Figure 2c), following stimulation by SRIF-14. In contrast, no phosphorylation of sst_2b receptors could be detected.

Effects of sst agonists and antagonist CYN154806 on recombinant sst₂ receptor-expressing epithelia

The sst₂-preferred peptides BIM23027 and BIM23190C were both full agonists in the transfected clones (Figure 3), displaying equivalent potencies and response time-profiles to SRIF-14 in rS2a and rS2b epithelial layers. BIM23027 inhibited VIP-stimulated I_SC with pEC₅₀ values of 7.41±0.26 (rS2a; n=4) and 7.50±0.18 (rS2b; n=3–5), BIM23190C pEC₅₀ values were 7.58±0.14 in rS2a cells (n=3–5) and 7.66±0.18 in the rS2b clone (n=3–5). The sst₅-preferred agonist BIM23268 (100 nM) elicited only small I_SC decreases in each clone, with slight reductions in the subsequent sensitivity to 100 nM SRIF-14 (Figure 3). BIM23056 (a partial agonist at sst₅, sst₃ and sst₂ receptors, 100 nM) did not affect VIP-elevated I_SC, nor did it alter the responses to SRIF-14 added 20 min later (Figure 3).

Figure 3.

Responses to SRIF-14 analogues in Colony 1 clones. Epithelial layers from hS2a (a), rS2a (b) or rS2b cells (c) were pre-stimulated with 30 nM vasoactive intestinal polypeptide for 20 min, followed by SRIF-14, BIM agonists, or D-Tyr⁸ CYN154806 (n=3–6). The pooled changes in I_SC to these peptides (mean±1 s.e.m.) are indicated by the left hand of each pair of columns (open for SRIF-14, filled or hatched for analogues). A second addition of 100 nM SRIF-14 was added after 20 min (SRIF-14 or BIM agonists) or 10 min (CYN154806), giving the indicated reductions in I_SC (right hand column of each pair). ^*P<0.05, ^**P<0.01 compared to control 100 nM SRIF-14 responses (open columns). BIM23027 was not tested (NT) in hS2a cells and CYN154806 data for rS2b was n=1. D-Tyr⁸ CYN154806, Ac-(4-NO₂-Phe)-cyc(DCys-Tyr-DTrp-Lys-Thr-Cys)-DTyr-NH₂; SRIF-14, somatotrophin release inhibiting factor.

The D-Tyr⁸ isomer of the peptide sst₂ antagonist CYN154806 (1 μM) produced small increases in I_SC, and attenuated peak SRIF-14 responses added 10 min later in hS2a, rS2a and rS2b clones (Figure 3). CYN154806 pIC₅₀ values for the inhibition of 100 nM SRIF-14 responses were 7.13±0.17 (hS2a) and 7.48±0.04 (rS2a; both n=3–6). SRIF-14 concentration–response relationships were parallel-shifted to the right to the same degree in hS2a and rS2a epithelial layers by 30 nM CYN154806, with no change in the maximum (data not shown). The resulting pEC₅₀ values in the presence of antagonist were 7.17±0.08 (hS2a) and 7.04±0.10 (rS2a; each n=3–8), yielding pK_B values for the hsst_2a and rsst_2a receptor of 7.9 and 7.8, respectively.

Responses to SRIF-14 and analogues in rat descending colon mucosa

Isolated preparations of rat descending colon mucosa exhibited basal resistances of 93.2±4.0 Ω cm² and I_SC levels of 48.7±3.7 μA cm⁻² (n=76). SRIF-14, BIM23190C, BIM23014C, BIM23268 and BIM23052 all inhibited the basal I_SC, with maximal sustained responses of similar magnitude. Concentration–response relationships (Figure 4a and Table 2) revealed the following order of agonist potency: BIM23190C⩾BIM23014C>SRIF-14>BIM23052=BIM23268. BIM23056 was inactive up to the highest concentration tested (444.4 nM; as observed previously in colonic mucosa; McKeen et al., 1995) and a subsequent addition of SRIF-14 (100 nM) still reduced I_SC by 77.0±8.0% (n=6; compared with controls of 71.2±5.2%, n=6).

Figure 4.

Responses to SRIF-14 and BIM agonists in rat descending colon mucosa. In (a), pooled cumulative concentration–response curves for SRIF-14 (n=18), BIM23190C (n=3–5), BIM23052 (n=4), BIM23268 (n=6) or BIM23056 (n=3) produced the pEC₅₀ values presented in Table 2. (b) The effect of CYN154806 pretreatment (CYN; 10 nM–1 μM; n=5) on the SRIF-14 concentration–response curve (antagonist pEC₅₀ values are quoted in the text) in tetrodotoxin (100 nM)-pretreated tissues. The resulting Schild plot (inset) had a slope of 0.85 and a pA₂ value of 8.2. SRIF, somatotrophin release inhibiting factor.

Table 2.

The potencies and peak effect of sst receptor agonists in rat descending colon mucosa

Agonist	Sst receptor preference and order	pEC50	Peak cumulative ΔISC (μA cm−2)
SRIF-14	Nonselective (2, 3, 5 ⩾1, 4)	7.92±0.08 (18)	−67.6±7.5 (444 nM)
BIM23190C	Sst2-preferred(2≫5>3≫4, 1)	8.48±0.06 (5)	−76.7±11.9 (144 nM)
BIM23014C	Sst2-preferred (2≫5>3≫4, 1)	8.21±0.05 (2–5)	−107.8±14.3 (144 nM)
BIM23052	Sst5-preferred (5>3, 2⩾4⩾1)	6.46±0.03 (4)	−83.3±18.6 (4.44 μM)
BIM23268	Sst5-preferred (5≫2, 4, 1>3)	6.84±0.03 (6)	−54.0±8.1 (1.44 μM)
BIM23056	Partial agonist at all (5>2, 3⩾4⩾1)	Inactive (3)	0.0±0.0 (444 nM)

Abbreviation: SRIF, somatotrophin release inhibiting factor.

The described sst receptor preferences of SRIF-14 and BIM analogues are based on binding IC₅₀ values obtained at recombinant human sst_1–5 receptors in stably transfected CHO-K1 cells (Shimon et al. (1997)).

pEC₅₀ values (with n numbers) and peak agonist responses (including the highest cumulative peptide concentration used, in brackets) were calculated from the concentration–response relationships presented in Figure 4a.

To measure epithelial responses to these agonists in the absence of potentially confounding neuronal sst receptor-mediated effects, we added the neurotoxin, TTX (100 nM) to mucosal preparations. TTX reduced the basal I_SC slightly (−5.6±0.9 μA cm⁻²; n=40, as shown previously, Ferrar et al., 1990), but had no significant effect on the subsequent pEC₅₀ values for BIM23190C (8.61±0.33, n=5) or BIM23268 (6.84±0.30, n=4) compared with controls (Table 2). Pre-incubation of TTX-treated tissues with CYN154806 (10 min, produced no change in I_SC) elicited rightward shifts in the SRIF-14 concentration–response curve without a change in the maximum inhibitory response (Figure 4b), yielding pEC₅₀ values of 7.38±0.01 (10 nM antagonist), 6.90±0.02 (100 nM) and 6.01±0.03 (1 μM, each n=5). The linearity and slope (0.85±0.13) of the resulting Schild analysis was consistent with reversible, competitive antagonism, and gave a pA₂ value of 8.2. The BIM23190C concentration–response curve was also right-shifted (pEC₅₀: 7.20±0.05, compared to 8.61±0.05 in control tissues; n=5) in the presence of 100 nM CYN154806 and the resulting pK_B value (8.4) was similar to the pA₂ obtained with SRIF-14 as the agonist. The effect of CYN154806 on sst₅ preferred agonist, BIM23268-stimulated responses was less pronounced, with a small decrease in agonist potency after 100 nM antagonist (pEC₅₀: 7.21±0.07 compared to control value of 7.49±0.18, n=3–5), resulting in a 10-fold lower pK_B estimate of 7.0.

Discussion

SRIF-14 antisecretory responses, desensitization and phosphorylation of recombinant sst₂-expressing epithelia.

We characterized the splice variants of the human and rat sst₂ receptors in stably transfected epithelial cells, to assess how differences between their respective C termini altered functional responses. Sst_2a and sst_2b receptors were both targeted to the basolateral epithelial membrane, indicating that the shortened C-tail of the sst_2b receptor did not cause its apical misdirection after synthesis (Beau et al., 2004). In each cell line (hS2a, rS2a and rS2b), basolateral SRIF-14 inhibited VIP-stimulated I_sc levels (pEC₅₀ 7.64–7.86), as expected from the preference of the former receptors for G_i/o protein coupling (Schindler et al., 1998; Siehler and Hoyer, 1999b). However, SRIF time-courses were strikingly more transient in hS2a and rS2a cells than observed for rS2b responses. Together with the comparison between SRIF cumulative and single addition concentration–response relationships (Figure 2b), these data provide strong evidence that both sst_2a receptors undergo much more rapid desensitization than the rsst_2b variant.

As we also observed in rS2a epithelial cells, native and recombinant sst_2a receptors are phosphorylated within minutes (Liu et al., 2003, 2005; Tulipano et al., 2004), either by a G-protein-coupled receptor kinase, GRK2, after homologous SRIF-14 stimulation (Elberg et al., 2002; Tulipano et al., 2004), or after heterologous activation of protein kinase C by other G-protein-coupled receptors (Hipkin et al., 2000; Elberg et al., 2002). Phosphorylated sst_2a receptors bound to SRIF-14 recruit β-arrestin adaptor proteins, preventing G-protein coupling (desensitization) and also initiating clathrin-mediated internalization (Tulipano et al. 2004). SRIF-14 and its peptide analogues (but not small molecule agonists) stimulate similar patterns of sst_2a β-arrestin2 recruitment and internalization (Liu et al., 2005), consistent with the identical time-course profiles of SRIF-14, BIM23027 and BIM23190C in hS2a and rS2a cells. The reduced desensitization observed here for the sst_2b variant is supported by an earlier study on mouse sst_2a and sst_2b isoforms (Vanetti et al., 1993), but differs from the comparison of rat sst_2a and sst_2b receptors made by Schindler et al. (1998). However, these authors used a much longer conditioning SRIF-14 treatment (60 min) to establish sst_2a and sst_2b desensitization, which contrasts with our examination of the real-time effects of a single agonist concentration over a shorter period (20 min). Our demonstration for the first time that rsst_2b receptors (containing 3 C-terminal Ser/Thr residues) are not phosphorylated as efficiently as the rsst_2a isoform (10 C-tail sites) suggests the underlying mechanism. In particular, alanine mutation of a C-terminal Thr cluster surrounded by acidic residues (E₃₅₂TTET) inhibits sst_2a phosphorylation and β-arrestin2 recruitment (Tulipano et al., 2004), and the absence of this key sequence in the sst_2b receptor may be sufficient to explain its resistance to phosphorylation and functional desensitization.

Comparison of hsst_2a and rsst_2a receptor pharmacology and desensitization

There were no apparent differences in the sensitivity of epithelial layers expressing either hsst_2a or rsst_2a to SRIF-14, the sst₂ antagonist CYN154806 or BIM compounds; sst₂ agonist BIM23190C or the sst₅-preferred BIM23268. Both epithelial clones were insensitive to the nonselective sst₅/sst₂/sst₃ agonist, BIM23056. CYN154806 was a competitive antagonist, blocking antisecretory responses with the same potency (pK_B values of 7.9 and 7.8) for hsst_2a and rsst_2a and this was similar to the potencies obtained for the cyclic octapeptide in other functional assays (Feniuk et al., 2000; they also used the D-Tyr⁸ isomer of CYN154806).

The sst₅-preferred agonists, BIM23268 (Figure 3) and BIM23052 (data not shown) were partial activators at the highest concentration tested (100 nM) and they only caused partial subsequent desensitization of SRIF-14 responses. BIM23056 had no effect alone, or on subsequent SRIF-14 responses (Figure 3) and we conclude it has a very low affinity for sst₂ receptors in epithelia, or in rat mucosal preparations (see below). Our observations are in keeping with reports that BIM23056 has a higher affinity for sst₅ (Shimon et al., 1997; Siehler et al., 1998) and is actually a potent sst₅ antagonist (Wilkinson et al., 1996). As our epithelia only express sst₂ receptors, we would not expect to see any sst₅-mediated effects to this, or other less potent sst₅ analogues. We have ascertained that BIM23190C and BIM23037 are full sst₂ agonists, while BIM23268 and BIM23052 were less potent in all preparations. It is notable that in another colonic adenocarcinoma cell line (Col-24 cells, that constitutively express other G_i-coupled receptors; Cox et al., 2001) BIM23268 was a potent antisecretory agonist (exhibiting an EC₅₀ value of 8.7 nM) and thus these cells appear to express sst₅ receptors constitutively (Cox et al., unpublished observations).

Sst₂ receptor pharmacology in rat-isolated colon mucosa

Reverse transcription-PCR analysis of crypt epithelia isolated from the rat descending colon, show expression of sst_2a and sst_2b receptors together with sst₁ transcripts and low levels of sst₅ RNA (but no sst₃ or sst₄ products, Warhurst et al., 1996). Given this combination of sst receptor expression, we might predict a mixed pharmacology. However, the sst₂-preferring agonists BIM23190C, BIM23014C (Table 2) and SRIF-14 were potent antisecretory agonists with similar efficacy, producing long-lasting reductions in I_SC (as shown previously for SRIF-14; Ferrar et al., 1990; McKeen et al., 1995). These inhibitory responses were unaffected by the neurotoxin, TTX and are therefore epithelial in origin. However, they were abolished by CYN154806, while BIM23056 (with purported sst₅/sst₂/sst₃ activity, Weckbecker et al., 2003; and sst₅ antagonism; Wilkinson et al., 1996) had no significant sst₂ activity in our epithelial models. Colonic SRIF-14 and BIM23190C responses were both blocked by CYN154806 (with a pA₂ value of 8.2 and pK_B of 8.4, respectively) and the cyclic octapeptide was an order of magnitude less potent at inhibiting the antisecretory responses of sst₅-preferred (but also sst₂-activating) BIM23268 (pK_B of 7.0). CYN154806 has in addition been found to have a low affinity for sst₅ receptors (two orders of magnitude less than its reported sst₂ affinity; Feniuk et al., 2000; Nunn et al., 2003), so inhibition, albeit at high nanomolar antagonist concentrations might be expected in colonic mucosa.

In conclusion, the rapid agonist-induced sst_2a receptor phosphorylation and coincident desensitization that we have observed is consistent with the pronounced sst_2a receptor internalization that Cole and Schindler (2000) and others (Hipkin et al., 2000) described. This acute process has already been harnessed to deliver stable radiolabelled SRIF-14 analogues into sst₂-expressing neuroendocrine tumours to allow either imaging (Breeman et al., 2001) or targeted chemotherapy and radiotherapy treatment (for a review, see Hofland and Lamberts, 2003; Weckbecker et al., 2003). However, certain patients with islet or carcinoid tumours can become insensitive to treatment with long-term SRIF-agonists such as lanreotide or octreotide (Hofland and Lamberts, 2003; Zomerhuis et al., 2005). It remains to be seen whether sst_2a receptor desensitization contributes in any way to this chronic reduction in efficacy.

Abbreviations

D-Tyr⁸ CYN154806

Ac-(4-NO₂-Phe)-cyc(DCys-Tyr-DTrp-Lys-Thr-Cys)-DTyr-NH₂

DMEM

Dulbecco's modified Eagle's medium

HA

haemagglutinin

hS2a

human sst_2a clone

rS2a

rat sst_2a clone

rS2b

rat sst_2b clone

I_SC

short-circuit current

KH

Kreb's Henseleit

SRIF-14

somatotrophin release inhibiting factor (somatostatin)14–28

TTX

tetrodotoxin

VIP

vasoactive intestinal polypeptide

UK14,304

5-bromo-N-(4,5-dihydro-1H-imidazol-2-yl)-6-quinoxalinamine

Conflict of interest

The authors state no conflict of interest.