Endogenous PYY and NPY mediate tonic Y1- and Y2-mediated absorption in human and mouse colon.

OBJECTIVE: To establish the functional significance of endogenous peptide YY (PYY) and neuropeptide Y (NPY) as mediators of Y(1) and Y(2) absorptive tone in colonic mucosa.
METHODS: Functional studies utilized descending colon from adult mice (wild type [WT] and peptide nulls) and ex vivo human colonic tissue (from patients undergoing bowel resections) measuring changes in basal ion transport. Peak increases in ion transport to Y(1) or Y(2) antagonists (BIBO3304 300 nM; BIIE0246 1 microM) were pooled (mean +/- SEM) and compared using Student's unpaired t test (P <or= 0.05); some tissues received tetrodotoxin (TTX; 100 nM). PYY-positive L-cell numbers and NPY innervation were also compared.
RESULTS: Y(1) and Y(2) tones were present in human and WT mouse colon mucosa and only the latter was TTX sensitive. Y(1) tone was unchanged in NPY(-/-) but was approximately 90% inhibited in PYY(-/-) and abolished in PYYNPY(-/-) colon mucosa. Y(2) tone was reduced approximately 50% in NPY(-/-) and PYY(-/-) tissues and was absent from PYYNPY(-/-) colon. Residual Y(2) and Y(1) tones present in PYY(-/-) mucosa were abolished by TTX. PYY ablation had no apparent effect on NPY innervation and PYY-positive cells were observed at the same frequency in NPY(-/-) (56.7+/-6.8 cells/section) and WT (55.0+/-4.6 cells/section) colons. Double knockouts lacked PYY and NPY expression, but endocrine cells and enteric nerves were present with similar frequencies to those of WT mice.
CONCLUSION: Endogenous PYY mediates Y(1) absorptive tone that is epithelial in origin, whereas Y(2) tone is a combination of PYY and NPY mediation.

Introduction

Peptide YY (PYY) and neuropeptide Y (NPY) are potent antisecretory peptides. Healthy human subjects infused with either peptide exhibit prolonged inhibition of pre-stimulated electrolyte secretion [1,2]. Recent functional studies utilizing selective Y-receptor antagonists and preferred peptide agonists have shown that PYY, NPY, their C-terminal fragments, and pancreatic polypeptide (PP) exert their sustained antisecretory actions via specific Y receptors, (Y1, Y2, and Y4) and that in human and mouse colons the same three Y receptors are involved [3-6]. However, clinical studies identifying the Y-receptor types responsible for these antidiarrheal actions have not been reported to date. Studies with ex vivo human colon have shown the presence of significant levels of Y1- and Y2-mediated absorptive tones [4,5]. In addition, mouse colon (but not other rodent colons) exhibits Y1 and Y2 tone [5-7] and the similarity in pharmacology exhibited by this rodent model and human tissue provides an opportunity to investigate the cellular mechanisms of absorptive tone further. This report is based on an invited lecture to the 9th International NPY meeting (Okinawa, March 2008) and focuses on the local mucosal effects of endogenous PYY and NPY.

Peptides: their locations and functions in the intestine

Peptide YY is expressed predominantly in colorectal endocrine L-cells [8,9] and is colocalized with proglucagon products, glicentin and glucagon-like peptide-1 (GLP-1) and GLP-2 [10]. These peptides are copackaged [10,11] and co-released when food reaches the duodenum. PYY is also found in pancreatic endocrine F-cells, often with PP. In the enteric nervous system, NPY is present in ∼50% of submucous plexus neurons, the majority of which innervate the mucosa and target the lamina propria of most species, including rat, mouse, and human colon [12,13]. Within the submucous plexus, NPY is most frequently present in secretomotor nerves and is colocalized with the inhibitory neurotransmitter, vasoactive intestinal polypeptide [12,14]. PP, in contrast, is present in a sparse population of endocrine cells scattered along the length of the intestine in most species [9,10], its primary source being the pancreatic F-cells.

Dietary fat, bile salts, carbohydrates, and proteins can stimulate PYY release but to different degrees and with different rates (for review, see Onaga et al. [15]). PYY release is also regulated by, and in turn regulates, vagal nerve activity [15] and the hormone is a major mediator of ileal and colonic brakes, mechanisms that ultimately slow gastric emptying and promote digestive activities to increase nutrient absorption [16-18]. In the circulation ∼40% of released PYY is converted to PYY(3–36) [19,20] and the consequence of this conversion is to amplify Y2-mediated mechanisms at the expense of Y1 (and Y4) responses, because the long fragments have a low affinity for the latter receptors. This hydrolysis occurs via the serine protease, dipeptidyl peptidase-4 (DPP4; EC 3.4.14.5), which cleaves aminoterminal dipeptides from NPY with a higher maximum activity than from PYY [20]. The switching of receptor activities is likely to be important in modulating digestive behavior and in initiating satiety consequent to postprandial increases in plasma PYY and PYY(3–36). DPP4 inhibitors are clinically important because they also prolong the half-life of incretin hormones, GLP-1 and GLP-2 [21], both of which lower blood glucose in a glucose-dependent manner. The inhibitors' potential as new treatments for type 2 diabetes is proven [22] but a proportion of their effects (i.e., promoting satiety and reducing food intake and subsequently body weight) are likely to be mediated by increased half-lives of unrelated peptides, e.g., PYY or NPY, whose prolonged stability (together with other DPP4 substrates [23]) may result in acute intestinal side effects.

Y receptors: their locations and implications for intestinal function

Exogenous PYY, NPY (their fragments), and PP are potent, broad-spectrum inhibitors of electrolyte secretion in human [2,4] and mouse [3,5] intestines and functional studies utilizing genetically modified mice lacking a Y receptor (Y1−/−, Y2−/−, or Y4−/−) combined with selective Y1 [24] or Y2 [25] antagonists and neurotoxin-treatment have shown that Y1 and Y4 are predominantly epithelial, whereas Y2 receptors are neuronal (Y5 (ant)agonists have no effect) [3-5]. Y1−/− mouse tissues are predictably insensitive to Y1 agonists and selectively lack Y1 absorptive tone, whereas Y2 tone and Y2 agonist sensitivity are unaltered [7]. Chemical depolarization of intrinsic submucous neurons (with veratridine) causes sustained epithelial responses that are not altered by Y1 antagonism (BIBO3304) or ablation (i.e., in Y1−/− tissue [7]). Predictably Y2−/− mucosa is Y2 agonist–insensitive and lacking in Y2 absorptive tone but Y1 tone is unchanged [5]. Neurogenic mucosal responses in wild-type (WT) colon are amplified by Y2 antagonist pretreatment and are significantly increased in Y2−/− colon [5]. Irrespective of these differences in cellular localization, Y-receptor activation leads consistently to prolonged antisecretory effects.

Immunohistochemical studies have revealed extensive Y1 labeling on epithelial basolateral membranes and discrete labeling of intrinsic neurons in the lamina propria of pediatric and adult human colon [13,26]. PYY-positive endocrine cells are often surrounded by Y1-immunoreactive epithelia and nerve fibers but are themselves Y1-negative [26]; thus a paracrine rather than an autocrine action for PYY is likely to be significant in human colon. In the rat intestine numerous Y1-positive cell bodies have been observed in myenteric but not in submucous ganglia [27], implicating a neuromodulatory role in smooth muscle activity in this species. In rat jejunum and colon Y2 expression occurs in epithelial and muscle layers [28] but as yet no immunohistochemical Y2-receptor localization in the intestine has been described. Notably, Y2 or Y1 absorptive tone is absent in rat gastrointestinal tissues [29]. In human and mouse colon mucosae, however, there is significant endogenous Y2 tone, which appears to be neurogenic [3,5] and is therefore more likely to be NPY rather than PYY mediated [7].

Despite growing pharmacological evidence, a better understanding of the endogenous mechanisms underpinning tonic absorption is a prerequisite if Y receptors are to be considered potential targets for treating malabsorption. Therefore, we set out to establish whether endogenous PYY was responsible for Y1 absorptive tone and NPY-mediated Y2 tone, making use of single (PYY−/− and NPY−/−) and double (NPYPYY−/−) knockout mice in conjunction with human colonic tissue.

Materials and methods

BIBO3304 and BIIE0246 were gifts from Boehringer-Ingelheim Pharma KG (Biberach an der Riss, Germany) and stocks in 10% dimethyl sulfoxide were stored at −20°C. Peptides were from Bachem Laboratories Inc. (St. Helens, United Kingdom). The DPP4 inhibitor, compound 3, was a gift from Dr. R. Roy (Merck Inc., Rahway, NJ, USA). Anti-PYY (from Dr. E. Ekblad, University of Lund, Lund, Sweden), anti–C-terminal peptide or NPY and anti-NPY antibodies (Affiniti Research Products Ltd., Exeter, United Kingdom), and goat anti-rabbit fluorescein isothiocyanate– or tetramethylrhodamine isothiocyanate–conjugated secondary antibodies (BIOMOL International, Exeter, United Kingdom) were used. All other compounds were of analytical grade (Sigma-Aldrich, Poole, United Kingdom).

Measurement of ion transport and Y1 and Y2 absorptive tone in isolated mucosal preparations

Colon mucosa from clinical specimens obtained from consenting patients undergoing bowel resection surgery (three men and one woman; Table 1) or from WT and knockout mice of different genotypes (NPY−/−, PYY−/− [30] and double knockouts, 16–24 wk old, either gender, with a mixed C57BL/6–129/SvJ background, fed standard chow) were prepared by removing overlying smooth muscle and voltage-clamped at 0 mV in Ussing chambers as described previously [3,4]. Vectorial ion transport (Isc) was measured continuously as microamps per square centimeter and all additions were basolateral. Once stable basal Isc levels were achieved, mucosae were pretreated with vehicle, the DPP4 inhibitor (1 μM of compound 3, [22]) or tetrodotoxin (TTX; 100 nM). Treatment periods were 20–30 min before the addition of the Y1-receptor antagonist BIBO3304 (300 nM [24]), the inactive enantiomer of BIBP3226, BIBP3435 (1 μM, 3435), or the Y2 selective antagonist BIIE0246 (1 μM [25]). The maximum rise in Isc (15–25 min) after each Y antagonist was pooled and compared with controls.

Table 1

Age and gender comparisons with basal resistance and Isc levels and the effect of neuronal blockade by TTX in mucosal preparations from human and murine colon⁎

	Age	No./gender	Resistance (Ω, cm2)	Basal Isc (μA/cm2)	TTX (μA/cm2)
Humans	71.3 ± 7.0 y (4)	3/M, 1/F	75.5 ± 4.9 (21)	72.0 ± 7.6 (21)	−30.5 ± 5.3 (7)
Mice
WT	18.5 ± 1.2 wk (17)	15/M, 2/F	29.3 ± 1.8 (52)	55.3 ± 5.6 (52)	−3.6 ± 1.3 (17)
NPY−/−	24.1 ± 1.4 wk (9)	5/M, 4/F	28.2 ± 2.7 (18)	82.6 ± 8.6 (18)†	−9.4 ± 1.8 (7)†
PYY−/−	18.3 ± 3.3 wk (7)	7/M	37.3 ± 3.0 (20)	73.9 ± 9.8 (20)	−8.6 ± 2.8 (6)
NPYPYY−/−	16.3 ± 0.5 wk (4)	4/F	24.8 ± 3.1 (8)	51.9 ± 15.1 (8)	ND

F, female; Isc, short-circuit current; M, male; ND, not determined; NPY−/−, neuropeptide Y single knockout; NPYPYY−/−, peptide YY/neuropeptide Y double knockout; PYY−/−, peptide YY single knockout; TTX, tetrodotoxin; WT, wild-type.

Values are means ± SEMs (numbers of observations).

P < 0.05 compared with WT controls.

Immunohistochemistry

Lengths (2–3 cm) of mouse descending colon were washed in KH buffer and immersed in paraformaldehyde (4%) for 24 h, washed well in phosphate buffered saline (PBS), cryoprotected in 30% sucrose in PBS for 48 h before being embedded in OCT (VWR International, Lutterworth, UK), and stored at −80°C. Sections (15 μm) were cut, rehydrated in PBS, and blocked in 10% normal goat serum in PBS for 2 h before incubating overnight in polyclonal anti-PYY antibody (1:1000) to visualize PYY-containing endocrine cells or in chromogranin A (1:400) to label all endocrine cells. Longer incubation times (3–4 d) were used to enable anti-NPY labeling (1:400) of NPY-containing neurons or protein gene product (PGP)9.5 (1:400) labeling of all enteric neurons. Primary antibodies were visualized with goat anti-rabbit F(ab′)2 secondary antibodies conjugated to fluorescein isothiocyanate or tetramethylrhodamine isothiocyanate (used at 1:200 for 2 h at room temperature; Chemicon, Harrow, UK). The sections were washed in PBS, mounted in Fluorosave (Calbiochem, Nottingham, UK), and viewed with a Provis microscope fitted with appropriate filters and Axiovision software, and the numbers of fluorescent endocrine cells were counted and innervation compared between genotypes.

Data analyses

Maximal changes in Isc at 15 or 25 min are expressed throughout as mean ± SEM from a minimum of three experiments. Single comparisons between data groups were performed using Student's unpaired t test, whereas multiple comparisons used one-way analysis of variance with Dunnett's post-test with P ≤ 0.05 considered statistically significantly different.

Results

Table 1 presents the basal resistances and Isc levels for human and murine colon mucosae. Values were similar to those published previously for human and WT mouse mucosae [5,6] and basal levels of Isc and TTX-sensitive Isc in NPY−/− colon specifically were significantly higher than those of WT tissue. The competitive Y1 antagonist, BIBO3304, caused sustained elevations in Isc that were maximal at 15 min in WT mouse and human colon mucosa and neither of these effects was sensitive to TTX pretreatment (Fig. 1A,C). The inactive Y1 antagonist enantiomer, BIBP3435, had no effect per se (P ≤ 0.01 in both tissues). Blockade of Y2-mediated absorption (with Y2 antagonist BIIE0246) also increased basal Isc levels that were virtually abolished by the neurotoxin TTX (Fig. 1B,D). This indicates that Y2 tone is predominantly neuronal in contrast to Y1 absorptive tone that is non-neuronal in both colonic tissues.

Fig. 1

Y1 (3304) and Y2 (0246) antagonists reveal absorptive tone but 3435 (an inactive Y1 isomer) was ineffective. Y1 and Y2 antagonism raised Isc in human (A, B) and wild-type (C, D) mouse colon mucosa, respectively. Y1 tone in both tissues was insensitive to TTX (+TTX, 100 nM; A, C), whereas Y2 tone was significantly reduced by TTX pretreatment of both mucosae (B, D). Asterisks indicate statistical differences between control and experimental data groups (*P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001) and bars represent mean ± SEM from 3–10 observations. ΔIsc, change in short-circuit current; TTX, tetrodotoxin; 0246, BIIE0246; 3304, BIBO3304; 3435, BIBP3435.

Because NPY is a better substrate for DPP4, Y2 tone was predicted to be amplified by a selective DPP4 inhibitor. Whereas Y1 tone was unaffected in mouse or human mucosa (data not shown), the same pretreatment with compound 3 significantly augmented Y2 tone at 25 min in human mucosa (control [n = 4] 9.6 ± 4.7 μA/cm2 versus compound 3 pretreatment [n = 4] 29.5 ± 5.9 μA/cm2, P ≤ 0.05) and at 15 min after BIIE0246 addition to mouse mucosa (controls [n = 8] 8.7 ± 2.3 μA/cm2 versus pretreatment [n = 8] 17.1 ± 2.6 μA/cm2, P ≤ 0.05).

To establish the relative contributions of endogenous NPY and PYY toward each Y-receptor–mediated tone, we monitored the effects of Y1 or Y2 antagonists in colon mucosa from single knockout mice (NPY−/− or PYY−/−) and the double-null mice (NPYPYY−/−). Comparison of the maximal increases in Isc 15 min after antagonist additions in WT versus null mucosa showed that NPY−/− colon exhibited normal Y1 tone (that was not TTX sensitive; Fig. 2A), whereas Y2 tone was partially (but not significantly) reduced by NPY ablation (Fig. 2B). TTX pretreatment of NPY−/− tissue did not further inhibit Y1 or Y2 residual tone (Fig. 2A,B). In marked contrast, PYY−/− mucosa was markedly less sensitive than WT colon to Y1 (Fig. 2C) and Y2 (Fig. 2D) antagonism and blocking neuronal activity in this tissue abolished these residual increases in Isc. NPYPYY−/− tissues were insensitive to both Y antagonists (Fig. 2A,B).

Fig. 2

Effects of individual NPY or PYY ablation (NPY−/−, PYY−/−) and double knockout (NPYPYY−/−) on Y1 and Y2 tone in mouse colon mucosa. (A) WT Y1 tone was similar to that of NPY−/− with or without TTX but was absent from NPYPYY−/− tissues (**P ≤ 0.01). (B) Y2 tone was partially reduced in NPY−/− tissue with or without TTX (not significantly) and abolished in NPYPYY−/− preparations (*P ≤ 0.05). (C) Y1 tone was reduced in PYY−/− colon (***P ≤ 0.001) and residual Y1 tone was abolished by neuronal blockade with TTX. (D) Y2 tone was partially inhibited in PYY−/− mucosae (*P ≤ 0.05) and abolished by TTX. Bars represent mean ± SEM from 3–7 observations. ΔIsc, change in short-circuit current; NPY−/−, neuropeptide Y single knockout; NPYPYY−/−, peptide YY/neuropeptide Y double knockout; PYY−/−, peptide YY single knockout; TTX, tetrodotoxin; WT, wild-type.

The NPYPYY−/− colon lacked PYY endocrine cells and NPY-containing neurons in both intramural plexi, although chromogranin A–stained endocrine cells and PGP9.5-positive neurons were present (data not shown). The PYY−/− colon exhibited no obvious alteration in the pattern of intramural NPY innervation, but PYY-positive cells were absent. The NPY−/− colon exhibited similar numbers of PYY-labeled endocrine cells (56.7 ± 6.8 cells/section) compared with WT mucosa (55.0 ± 4.6 cells/section) and the morphology of WT and null tissues was similar.

Discussion

Y1 tone is clearly stereospecific and not mediated by TTX-sensitive enteric neurons in human and mouse colon (Fig. 1A,C). In contrast, Y2 tone is predominantly neurogenic (Fig. 1B,D), as described previously in normal human colon and WT mouse mucosa [4,5]. The link between NPY and Y2 tone was initially revealed in ex vivo studies showing that NPY−/− and Y2−/− tissues lost Y2 tone and were indistinguishable in this regard [7]. The present study included NPY−/− mice from a source different from that used previously [7] but with a similar mixed genetic background (C57BL/6–129/SvJ). Apart from an elevation in basal Isc and TTX sensitivity in NPY−/− compared with WT values (Table 1), there were no differences in electrophysiological parameters between genotypes. Loss of inhibitory NPY could be partly responsible for the increase in basal Isc levels if tonic activation of intrinsic submucous nerves is contributing to basal Isc. The increased effect of TTX on basal Isc in NPY−/− compared with WT values would also confirm this is the case.

The selective amplification of Y2 tone by DPP4 inhibition also indicates a significant tonic inhibition for endogenous NPY in human and mouse colon. Prolonging NPY's half-life would result in potentiation of neurogenic responses and, because Y2 effects are primarily neuronal, they were increased, whereas the antisecretory actions of PYY at epithelial Y1 receptors were not significantly altered. Peptide levels before and after DPP4 inhibition will be measured to confirm the relative stability of NPY over PYY, particularly at times when Y2 (but not Y1) tone is significantly amplified.

Knockout of individual peptides revealed clear differences in functional losses in the absence of any changes in peptide distribution. The NPY−/− colon exhibited normal levels of Y1 tone that were not significantly altered by neuronal blockade, indicating a direct epithelial mechanism of PYY action (Fig. 2A,B). This agrees with previous studies of antisecretory responses activated by exogenous analogs [5,6] (Cox et al., unpublished observations). PYY ablation conversely reduced Y1 tone by ∼90% (Fig. 2C), whereas Y2 tone was partially inhibited (Fig. 2D) and the remaining residual responses were abolished by TTX and by inference must be NPY mediated. As predicted, double knockouts lacked Y1 and Y2 tonic absorption (Fig. 2A,B). We conclude that a combination of the two peptides are responsible for Y2 tonic absorption and that neuronal and epithelial mechanisms underpin this tonic effect, as indicated by the neurotoxin TTX's ability to partially inhibit antagonist responses in NPY−/− (Fig. 2B) and abolish remaining Y2 tone in the PYY−/− colon (Fig. 2D). In contrast, NPY has a minor role as a mediator of Y1 tone (Fig. 2C). The location of Y1 receptors in basolateral poles of epithelia surrounding PYY-positive endocrine cells in human colon [26] fits well with the functional PYY–Y1 interactions observed in this tissue and in mouse colon (as shown in Fig. 3). In these tissues we have no evidence for submucous plexus or mucosal neuron Y1 expression and this agrees with the more limited distribution of Y1-receptor immunoreactivity noted by Matsuda et al. [27]. They observed only Y1 labeling of myenteric nerves (removed with smooth muscle from our preparations), a few endocrine-like cells, and specific larger blood vessels.

Fig. 3

Schematic diagram showing the intramural sites of action of endogenous PYY and NPY in normal mouse and human colon mucosa. Direct activation of epithelial Y1 receptors by PYY (and NPY) will inhibit epithelial anion (Cl−) secretion. NPY released from submucosal secretomotor neurons can auto-inhibit its release (lefthand side, a Y2-mediated effect) and also, when released from interneurons, can inhibit (again via Y2 receptors) secretomotor (e.g., VIP-ergic) neurons. Endocrine PYY can co-activate neuronal Y2 receptors and predominant epithelial Y1 receptors, and both mechanisms result in sustained inhibition of epithelial Cl− secretion. The NANC neurotransmitter in the final secretomotor neuron (righthand side) has yet to be positively identified but is likely to be VIP, which in turn stimulates prolonged epithelial cyclic adenosine monophosphate–dependent Cl− secretion that can be inhibited by Y1 (or Y4) receptor activation. NPY, neuropeptide Y; PYY, peptide YY; VIP, vasoactive intestinal polypeptide.

The selective inhibition of Y2 responses by TTX indicate the absence of Y2 receptors in human and mouse colonic epithelia, in contrast to previous studies of rat jejunum, where functional studies [29] and polymerase chain reaction–based detection of Y2 mRNA [28] showed epithelial Y2 expression. We propose that activation of presynaptic Y2 receptors by NPY and PYY inhibits secretomotor nerves (most probably vasoactive intestinal polypeptide–ergic) and provides a mechanism by which Y2 antagonism elevates basal Isc and causes hypersecretion (Fig. 3) in human and mouse colon. NPY released from other non-adrenergic, non-cholinergic secretomotor nerves could auto-inhibit NPY release by Y2 receptor activation and local NPY, NPY(3–36), PYY, or PYY(3–36) could act on these neuronal Y2 receptors to modulate ongoing mucosal electrolyte secretion.

Ultimately, whatever the luminal or neural (vagal) stimulus, released PYY will rapidly activate several local (paracrine) targets, primary among them, the epithelium. In addition to its hormonal effects that occur within minutes to hours, PYY can also exert longer-term responses to alter epithelial adhesion, differentiation, and increasing cell migration [31], potentially protecting against stimuli that cause mucosal erosion. The same PYY−/− mice that we have used have been shown to exhibit a more prominent female development of mild late-onset obesity with a high-fat diet [30] and this outcome was not dissimilar to the age-related weight gained by both genders of another PYY−/− recently described [32]. Whether hyperinsulinemia contributes to this phenotype is not agreed but tonic inhibition afforded by PYY acting on Y1 receptors in pancreatic islets combined with Y2 inhibition of vagal output have been proposed [30] as coincident contributing factors. It is interesting to note that PYY and NPY are responsible for Y1- and Y2-mediated tonic inhibition in central and peripheral targets including the intestinal tract, where I have uncovered significant tonic activity using selective Y1 and Y2 antagonists.

In conclusion, PYY and NPY exert common final antisecretory actions but mediated by different pathways involving two Y receptors (Y1 and Y2) with distinct distributions. The PYY-synthesizing L-cells of the large bowel play a pivotal role, not only in the regulation of satiety and ileal brake but also in response to different luminal cues, e.g., fatty acid chain length [33], resulting in predominant Y1 absorptive responses. NPY exerts its inhibitory effects predominantly through submucous neuronal innervation of the mucosa to bring about indirect Y2-mediated absorption and both tonically active mechanisms are present in human and mouse colon mucosa.