In vitro and in vivo pharmacological role of TLQP-21, a VGF-derived peptide, in the regulation of rat gastric motor functions

Gastrointestinal contraction

Wistar male and female rats (250–350 g; Charles River, Calco, Italy) were killed by inhalation of 75% CO₂ in air. Different portions of the GI tract (oesophagus, stomach, pylorus, jejunum, proximal and distal colon) were removed and washed in fresh Tyrode's solution as previously described (Severini et al., 2000). For the stomach contractility assays, longitudinal strips were obtained from the RLF region and mounted vertically in a 5 mL organ bath in oxygenated (95% O₂ and 5% CO₂) Tyrode's solution at 37°C. Responses were recorded isotonically by a strain gauge transducer (DY 1, Basile, Milan, Italy) and displayed on a recording microdynamometer (Unirecord, Basile. Milano, Italy). Tissues were allowed to equilibrate for about 60 min and the Tyrode buffer was changed every 10 min.

Saturating concentrations of acetylcholine (ACh, 25 µmol·L⁻¹) were added every 20 min, washing the tissue after 1 min contact time, until reproducible contractile responses were obtained. After 30 min equilibration time, TLQP-21 non-cumulative concentration–responses curves were constructed by adding increasing concentrations of the peptide to the tissue chamber, until maximum muscular responses occurred. The tissue was washed as soon as the contraction peak had developed, after a contact time with muscle strips of 1–3 min.

In preliminary experiments, peptide applications were performed at different time intervals (5, 10, 15 and 20 min) to establish the appropriate interval between subsequent administrations. As an interval of 20 min was necessary to obtain a complete recovery, that is, to avoid tachyphylaxis, this interval was used for every successive experiment. Only one concentration–response study was performed on each strip. TLQP-21-induced contractions were expressed as mean ± SEM of the percentage of ACh-induced contractions and EC₅₀ calculated by interpolation from the appropriate concentration–response curve. Maximally effective concentration (E_max) denotes the maximal response achieved by the peptide, expressed as a percentage of that obtained with saturating concentrations of ACh (25 µmol·L⁻¹).

Antagonists were all at a sufficiently high concentration to block the corresponding agonists under the experimental conditions.

To investigate the contribution of the gastric mucosa on the effect of TLQP-21, additional experiments were performed with mucosa-free strips. These preparations were obtained by removing the mucosa from the RLF with a lancet. The tissues were set up in the organ baths as previously mentioned.

Prostaglandin release

Longitudinal RLF strips were prepared as previously described and tested for standard response to ACh. Tissues were weighed and subsequently transferred to an Eppendorf tube and incubated without (control) or with TLQP-21 (400 nmol·L⁻¹, 4 µmol·L⁻¹) for 90 s at 37°C in 0.75 mL of Tyrode's solution. PGE₂ and PGF_2α concentration was measured immediately and in parallel using commercially available EIA kits (Cayman Chemical Company, Ann Arbor, MI, USA) according to the manufacturer's instructions. The reactions were read at 405 nm on a 1420-011 multi-label counter (Wallac Oy, 20101 Turku, Finland).

Immunostaining

Samples of gastric body-forestomach were taken from female adult rats (Wistar, 200–250 g, killed by decapitation: n = 8), pinned flat onto pieces of cork, immersion-fixed in paraformaldehyde (40 g·L⁻¹, in 0.1 mol·L⁻¹ phosphate buffer: 3 h at 0–4°C) and frozen as previously described (Rindi et al., 2007). The rabbit antiserum used was raised against the C-terminal sequence of rat VGF, VGF_609–617, and has been previously characterized (Ferri et al., 1995). Rat gastric sections were treated with Triton X-100 (1 mL·L⁻¹, in H₂O for 1 h), incubated overnight with the anti-VGF antiserum (1:2000–10 000), followed by Cy3-conjugated anti-rabbit antibodies (Jackson Immunoresearch Laboratories, West Grove, PA, USA). Routine controls included substitution of each layer, in turn, with non-immune serum, or phosphate-buffered saline. Pre-absorption of the antiserum with the relevant peptide (approximately 50 nmol·mL⁻¹) virtually abolished staining. Preparations were examined with a Olympus BX60 fluorescence and transmitted light microscope, equipped with a PM30 photographic system.

Western blot analysis of rat stomach homogenates

Wistar rats (weighing 250–350 g) were killed by inhalation of 75% CO₂ in air. The stomachs were rapidly dissected, minced, boiled for 10 min in 10 mL H₂O·g⁻¹ of tissue and further incubated at room temperature for 15 min. Coarse materials were briefly sedimented and the supernatant cleared at 20 000×g for 45 min at 4°C. This procedure resulted in both protease inactivation and enriched extraction of low molecular weight peptides (Trani et al., 1995). Approximately 0.03 mg of proteins was loaded on gels according to the protocol for NuPAGE system (Invitrogen, Carlsbad, CA, USA). Western blot analysis was performed by using an anti-VGF_609–617 serum (previously mentioned) at 1:5000 dilution for 2 h and goat anti-rabbit-HRP conjugated secondary antibody at 1:2000 dilution for 45 min (Pierce Biotecnology, Rockford, IL, USA). Chemiluminescent detection of immunoblots was achieved by using the Supersignal West Dura kit (Pierce Biotecnology, Rockford, IL, USA) according to the manufacturer's instructions.

Gastric emptying

Adult Wistar rats, weighing 180–200 g (Charles River, Calco, Italy) were used. The animals were housed in single cages with wire net bases to avoid coprophagy. Before the experiments, rats were fasted for 24 h, but had free access to water until the treatment was started. For i.c.v. administration of peptides, a polyethylene cannula (PE10) was implanted into the left lateral ventricle of the brain, 5 days before the experiments as previously described (Netti et al., 1984). At the end of the experiments, dye (0.5% Evans Blue) was injected through the cannula to confirm its position in the ventricle.

The gastric emptying assay, consisting of a phenol red test meal was performed in female and male Wistar rats (Broccardo and Improta, 1990). Briefly, a liquid and acaloric meal (1.5 mL per rat) consisting of a 50 mg phenol red solution in 100 mL aqueous methylcellulose (1.5%) was administered by gavage to conscious rats. TLQP-21 (0.1, 0.4, 2, 8 and 32 nmol per rat) or saline (NaCl 0.9%) were injected i.c.v in a constant volume (5 µL per rat) immediately before meal administration. TLQP-21 (48 or 480 nmol·kg⁻¹) or saline was administered 5 min before meal administration when i.p. injected. Rats were killed by CO₂ (70%) inhalation 10 min after the test meal, when maximal differences between saline and TLQP-21 groups were observed. In each experiment, four non-treated rats were euthanized by CO₂ (70%) inhalation immediately after the administration of the phenol red solution, and this was considered the reference of 0% emptying rate. The stomach and its contents were homogenized with 100 mL of 0.1 N NaOH, protein precipitated with 20% trichloroacetic acid and the supernatant made alkaline with NaOH. The absorbance of the sample was measured at 560 nm (A560) with a spectrophotometer. Gastric emptying (GE) for each rat was calculated according to the following formula: GE (%) = [1 − (A560 sample/mean of A560 reference)] × 100, where A560 sample was the absorption at 560 nm of the gastric content at 10 min and A560 reference was the absorption at 560 nm of the gastric content at zero emptying time. Each value was expressed as percentage change with respect to the percentage gastric emptying in saline-treated rats.

To provide insights into TLQP-21-evoked gastric motor responses, the cyclooxygenase (COX) inhibitor indomethacin was given in different ways to two sets of animals. In the first set, indomethacin was given s.c. (3 mg·kg⁻¹) (Takeuchi et al., 1990), 20 min before i.c.v. TLQP-21. In the second set, indomethacin was given i.c.v. (0.25 mg per rat) (Saperas et al., 1991), 5 min before i.c.v. TLQP-21.

Data analysis

Statistical analysis was performed with a statistic package (GraphPad Prism, GraphPad Software San Diego, CA, USA). All data are represented as the mean ± SEM. Differences between the groups were assessed by one-way analysis of variance (anova) followed by multi-comparison Dunnett's test. The ED₅₀ was calculated by linear regression analysis of the concentration–response data.

Chemicals

TLQP-21 (TLQPPASSRRRHFHHALPPAR) and the other VGF-derived peptides were synthesized by PRIMM (Milan, Italy). Stock solutions of peptides (1–10 mmol·L⁻¹) were made in distilled water, stored at −20°C and then diluted in water before use. COX inhibitors (SC560, NS398) and PG receptor antagonists (SC19220 and PGF_2α dimethylamide; Cayman Chemical Company, Ann Arbor, MI, USA) were dissolved in DMSO. SR140,333 was provided by Sanofi-Synthelabo Ricerche (Montpellier, France) and tetrodotoxin (TTX) was from Alomone (Jerusalem, Israel). The other reagents were from Sigma (St. Louis, MO, USA).

Gastrointestinal contraction

We tested the contractile activity of the natural and synthetic VGF-derived peptides reported in Figure 1 on different regions of the rat GI tracts (oesophagus, pylorus, gastric antrum, gastric forestomach, jejunum, proximal and distal colon). Each of the biologically active VGF-derived peptides purified from brain extracts (TLQP-62, TLQP-21, AQEE-30 and LQEQ-19) failed to evoke contractions in the tested GI preparations (25–50 µmol·L⁻¹, n = 5, data not shown) with the exception of TLQP-21. This peptide elicited a reproducible and concentration-dependent contractile activity (100 nmol·L⁻¹–6 µmol·L⁻¹) of the RLF smooth muscle (Figure 2) and only weak and not concentration-dependent activity on oesophagus, gastric antrum and circular forestomach muscule strips, even at much higher concentrations (25–50 µmol·L⁻¹, data not shown).

In addition, we tested on RLF strips, the contractile action of the synthetic peptides TLQP-11, HFHH-10 and TLQP-30, corresponding to fragments or an extension of the TLQP-21 sequence. In view of the presence of specific cleavage sites in the VGF protein, these peptides could be generated in vivo by endogenous peptidases. Nevertheless, these peptides failed to contract RLF strips (25–50 µmol·L⁻¹, n = 5, data not shown).

TLQP-21 was more potent on contracting RLF strips from female rats than those from male rats. As shown in Figure 2A, TLQP-21 (100 nmol·L⁻¹–10 µmol·L⁻¹) concentration-dependently contracted the RLF strips from male rats with E_max value of 33.4 ± 5.3% and EC₅₀ value of 0.87 µmol·L⁻¹ (n =8 at each concentration), while in female rats E_max was equal to 85.7 ± 7.9% and the EC₅₀ value of 0.47 µmol·L⁻¹ (n =8 at each concentration). In light of these sex differences, we used RLF strips from female rats for the subsequent in vitro pharmacological studies. A representative concentration–response trace is shown in Figure 2B. The effect of TLQP-21 (1 µmol·L⁻¹) was determined against contractions induced by ACh, at a concentration previously determined to be maximally effective (25 µmol·L⁻¹, 1 min contact).

Contractile activity exerted by TLQP-21 showed a clear tachyphylaxis, as a second application of the peptide (1 µmol·L⁻¹) at time intervals shorter than 20 min resulted in a reduced response, while complete reproducibility was obtained on reapplying the peptide after 20 min, as shown in Figure 3A and 3B. In addition, manual removal of the mucosal layer from the RLF strips prevented TLQP-21-induced contractile activity (3 ± 4% of the ACh-induced contraction, n = 5), while ACh-induced contraction was unmodified, when compared with the corresponding contraction of intact RLF preparations, relative to the weight of tissue used (Figure 3C).

Mechanisms of TLQP-21-induced contractions of RLF strips

To characterize the possible mediator(s) of TLQP-21 action, we tested different compounds, capable of inhibiting substances known to stimulate gastric motility (see Table 1). The addition of atropine, SR140,333, a mixture of methysergide and ketanserin, and a mixture of mepyramine and cimetidine did not modify the response elicited by TLQP-21, while it prevented or reduced any contraction evoked by their respective agonists ACh, substance P, 5-HT and histamine. TTX also failed to influence the effect of TLQP-21, suggesting that a neuronal action potential is not involved.

On the other hand, 15 min pretreatment with the non-specific COX inhibitor indomethacin (from 100 nmol·L⁻¹ to10 µmol·L⁻¹), the specific COX-1 inhibitor (SC560, from 10 nmol·L⁻¹ to 10 µmol·L⁻¹) or the specific COX-2 inhibitor (NS398, from 10 nmol·L⁻¹ to 10 µmol·L⁻¹) dose-dependently prevented TLQP-21-induced contractions (1 µmol·L⁻¹), as shown in Table 2.

To compare the TLQP-21 contractile activity on RLF strips with that of PGF_2α and PGE_2,non-cumulative concentration–responses curves of these PGs were made. The EC₅₀ value for PGF_2α was 11.5 nmol·L⁻¹ (n =4 at each concentration), and the EC₅₀ value for PGE₂ was 5.5 nmol·L⁻¹ (n =4 at each concentration), as compared with the TLQP-21 EC₅₀ value of 0.87 µmol·L⁻¹.

In the light of these findings, we investigated the identity of the PGs released in response to TLQP-21 using antagonists for the FP and the EP receptors (prostaglandin E₂ and F_2α receptors, nomenclature follows Alexander et al., 2008). Either PGF_2a dimethylamide (antagonist for the FP receptor, 10–100 µmol·L⁻¹) or SC19220 (antagonist for the EP receptor, 1–100 µmol·L⁻¹) significantly reduced the TLQP-21-induced contraction as well as the activity of the corresponding agonists PGF_2α and PGE₂ used at EC₅₀ concentrations (Table 2). To confirm the release of these PGs from RLF strips, these strips were incubated with different concentrations of TLQP-21 (400 nmol·L⁻¹, 4 µmol·L⁻¹) and the PG content of the incubation medium assayed by EIA. As shown in Figure 4, a dose-dependent increase in the concentration of PGE₂ and PGF_2α was detected in the medium following incubation with TLQP-21, when compared with the corresponding basal levels. On the contrary, no increase in PG production was detected by using the same tissues after mechanical removal of the mucosa (data not shown).

VGF expression and processing in the rat stomach

We studied the presence and distribution of VGF-containing nerve structures in the stomach by means of immunocytochemistry using antibodies raised against the C-terminal region of VGF (Figure 5) (Ferri et al., 1995). As shown in Figure 5A, a number of strongly labelled, VGF C-terminus-immunoreactive axons were shown in the different layers of the forestomach (Figure 5A, panels a-c). Scattered bundles of beaded axons were seen in the mucosa, running in the lamina propria between the gastric glands (Figure 5A, panel a). The VGF-containing innervation of the muscularis externa was especially rich, with very numerous varicose axons running in both the circular and the longitudinal layers (Figure 5A, panel a). Immunoreactive terminals were also shown in myenteric ganglia, largely depicting non-labelled cell bodies (Figure 5, panels a and c). Weak but distinct labelling was seen in a few perikarya (Figure 5A, panel b).

To characterize which of VGF-derived peptides, previously purified by brain extracts (Trani et al., 2002) were present, Western blot analysis was performed on extracts from rat stomach, comparing bands with those from primary cultures of rat cerebellar granule cells.

As shown in Figure 5B, in addition to high-form precursors (double arrows), two bands detected in stomach samples could be identified as NAPP-129 and TLQP-62 (single arrows) on the basis of their electrophoretic point. The presence of other peptides of lower molecular weight is difficult to detect by Western blot analysis.

Gastric emptying

I.c.v. injections of TLQP-21 (from 0.1 to 32 nmol per rat) delayed gastric emptying of a phenol red meal, measured both in male and female rats, as reported in Figure 6A. TLQP-21 decreased the gastric emptying significantly, but not in a dose-related manner (2, 8 and 32 nmol per rat) in comparison with gastric emptying values in rats given i.c.v. saline (value marked 0). The maximal gastric emptying inhibitory response was about 40% both in male (Figure 6A) and female rats (data not shown). The effect of i.c.v. administration of TLQP-21 on gastric emptying was evaluated after i.p. and i.c.v. injection of the COX inhibitor indomethacin at dose levels, which had no effect on gastric emptying. Indomethacin (3 mg·kg⁻¹), when given i.p., did not modify the delay of gastric emptying induced by i.c.v. administration of TLQP-21 (8 nmol per rat). However, when indomethacin was given i.c.v. (0.25 mg per rat), before i.c.v. TLQP-21, it totally blocked the inhibitory effect of the peptide (Figure 6B). TLQP-21, when administered peripherally (48 and 480 nmol·kg⁻¹ i.p.), did not affect gastric emptying of a phenol red meal measured in male (Figure 6C) and female rats (data not shown).