Clinical studies have shown that BR, a cysteine protease obtained from pineapple stems, was effective – presumably through an anti-inflammatory mechanism – in reducing a number of symptoms, including diarrhea, associated with inflammatory bowel disease (IBD).9,24 This is consistent to animal experiments in which it has been shown that BR prevented the intestinal fluid accumulation due to cholera toxin and Escherichia coli,10,11,25 and the damage in the IL-10-deficient murine model of intestinal inflammation.26 Here, we provide evidence that BR affects directly rodents smooth muscle ileum contractility in vitro and reduces intestinal transit in vivo. Furthermore, BR was more potent in pathophysiologic states – such as inflammation or diabetes – than in healthy animals.
In vitro studies
We have shown that BR inhibited intestinal spasms in the isolated ileum. The inhibitory effect of BR was non-selective as this proteolytic compound inhibited with a similar potency the contractions induced by ACh (which acts directly on muscarinic receptors located on smooth muscles), BaCl2 (which enters the intracellular space of the smooth muscle and directly stimulates the contractile mechanism), and by EFS (which acts releasing ACh from myenteric nerves). This non-selective profile of inhibition probably indicates that BR acts directly on smooth muscle contractility (i.e., postjunctional effect) rather than on processes involved in neurotransmitter(s) release. Therefore, to investigate the mechanism of BR-induced contractility inhibition we performed further experiments on the contractions induced by ACh.
To evaluate whether the antispasmodic effect of BR was related to its proteolytic activity, we analyzed the effect of inactivated BR on ACh-induced contractions. Our results showed that the antispasmodic effect of BR requires proteolytic activity, as a treatment with a proteolytically inactive BR did not affect the intestinal motility. Further evidence about the importance of the proteolytic activity in the mode of action of BR comes from the experiments with the proteolytic enzyme inhibitor gabexate, which reduced the inhibitory effect of BR on ACh-induced contractions.
In the gastrointestinal tract there are specific receptors that can be activated by proteases, including the protease-activated receptors (PARs), which comprise a family of four subtypes (PAR-1, PAR-2, PAR-3, and PAR-4) all activated by serine proteases.27,28 These receptors are activated by proteolytic cleavage of a receptor-bound, amino-terminal tethered ligand domain, which is then able to bind to the receptor and to initiate intracellular signalling. PAR-1 and PAR-2 are widely expressed in smooth muscle cells; however their role in motility modulation is very complex, depending upon the animal species and region of the gut studied.29,30 In the isolated mouse small and large intestine, PAR-1 and PAR-2 agonists elicit transient relaxation followed by contraction.31–35 In our experiments the selective PAR-2 antagonist ENMD-1068, but not the selective PAR-1 antagonist SCH 79797, reduced the inhibitory effect of BR. These data provide an indirect evidence about an involvement of PAR-2 on the antispasmodic effect of BR.
In order to better explore BR mode of action on PAR-2, we have synthesized several peptides (I–VII) derived from the fragment Cys22-Lys36 of the extracellular N-terminal PAR-2 sequence. Bromelain has a broad substrate specificity and hydrolyzes a great variety of synthetic and natural substrates extending from synthetic low molecular mass amides and dipeptides up to high molecular substrates such as fibrin, albumin, casein, angiotensin II, bradykinin.5,36. Several studies reported that BR preferentially cleaves glycyl, alanyl, and leucyl bonds; on the other hand, substrate specificity studies performed using combinatorial libraries of small fluorogenic substrates showed a strong preference of BR for Arg at both the P1 and P2 sites.37,38 These data, in conjunction with the observation that SLIGRL-NH2, a well-known PAR-2 agonist, did not exert any action on the BR inhibitory effect (data not shown), prompted us to synthesize peptides I–VII, in order to verify if the BR’s cleavage site on the PAR-2 amino-terminal extracellular domain might be different from that of trypsin. We have shown that longer peptides, such as compounds III (14 aa), IV (13 aa), V (12 aa), and VI (10 aa), did not modify the antispasmodic action of BR. Smaller peptides, such as compounds I (6 aa), II (6 aa), and VII (8 aa) were able to reduce the antispasmodic effect of BR, probably by preventing the binding to the receptor and its following activation. Interestingly, it has been reported that several synthetic peptides induce receptor activation in a different way than the PAR-2 tethered ligand.39–41
A further aim of our work was to investigate the intracellular signalling pathway that follows the activation of PAR-2 by BR. Compared with PAR-1, intracellular signal pathways associated with PAR-2 have been poorly investigated. Nevertheless, it is well known that PAR-2 activation by trypsin leads to activation of the Gq/G11–PLCβ pathway that induce an increase in intracellular calcium concentration.42–44 In order to evaluate the possible contribution of such pathway in BR-induced PAR-2 activation we have performed experiments with a selective inhibitor of PLC. Our results have shown that neomycin (a PLC inhibitor) reduced the inhibitory effect of BR on ACh-induced contractions, thus suggesting the involvement of Gq/G11–PLCβ pathway on BR action.
Although not all studies yield similar results, PAR-2 activation may modulate intracellular cAMP levels.45,46 In vascular tissues, the effect of PAR-2 is modulated by PDE4. PDE4, which is involved in smooth muscle contraction, specifically hydrolyses cAMP.47 We have shown that the PDE4 inhibitor rolipram reduced the antispasmodic effect of BR, suggesting an involvement of cAMP in BR action.
Because activation of PAR-2 results in an increase in [Ca2+]i, we performed further experiments in order to evaluate the effect of BR on [Ca2+]i. We found that BR increased [Ca2+]i in Caco-2 cells and this effect was reduced by the selective PAR-2 antagonist ENMD-1068. In addition, preincubation of Caco-2 cells with BR strongly reduce the response to SLIGRL-NH2. Together, these data further support the concept that BR could act by cleaving PAR-2.
In vivo studies
Because intestinal antispasmodic drugs may reduce intestinal motility in vivo and because the PAR-2 is involved in the control of intestinal transit, we investigated the effect of BR and the possible involvement of PAR-2 on BR’s action on intestinal transit in mice. BR, given intraperitoneally at doses ranging from 1 to 10 mg kg−1 reduced intestinal transit in control mice. By contrast, oral administration of BR (upto 500 mg kg−1) was ineffective. This is not surprising as BR is a proteolytic enzyme that is mostly destroyed by gastric acid. Incidentally, in order to prevent it from being destroyed by gastric juice, BR is clinically used in enteric-coated tablets.
As BR has been clinically evaluated in IBD patients, we also investigated its effect on motility in a mouse model of intestinal inflammation. We used the irritant compound CO, which has been extensively studied to induce hypermotility associated with inflammation in mice.12,16 We found that BR, either orally (100–500 mg kg−1) or intraperitoneally administered (1–10 mg kg−1), reduced hypermotility-induced inflammation, being the effect more pronounced in the inflamed gut rather than in the physiologic conditions. Collectively, our in vivo studies suggest that: (i) BR was more active after intraperitoneal than oral administration, and (ii) inflammation increased the potency of BR (see Fig. 6, insert).
The inhibitory effect of BR on intestinal transit, both in physiologic and pathophysiologic states, was reverted by ENMD-1068, thus confirming in vivo the possible involvement of PAR-2. Bromelain was also more active (compared to control mice) in a non-inflammatory model of intestinal dysmotility (i.e., diabetic mice),48 thus excluding the possibility that the higher potency of BR is due to a systemic anti-inflammatory effect. Possible sites of action of BR include epithelial cells, neuronal elements, and myocytes, in which PAR-2 is expressed. The PAR-2-containing cells can be reached by BR via systemic circulation or directly through the gastrointestinal tract (for BR escaping gastric inactivation).
It is also very unlikely that the higher potency of BR is due to PAR-2 hyper-expression because we found – in accordance to other studies49– a down-regulation of such receptor in the inflamed intestine. Further studies are needed to verify if the decreased PAR-2 protein expression is due to proteolysis or to changes in mRNA expression.
It is noteworthy that an increased PAR-2 expression has been observed in other non-intestinal experimental models of inflammation such as joint inflammation and rheumatoid synovium.19,50
We have shown for the first time that a cysteine protease, namely BR, exerts inhibitory effects on intestinal motility. The effect involves PLC and PDE4, and possibly PAR-2. From a clinical point of view, our study provides a further mechanism which may help to explain the efficacy of BR in reducing diarrhea in IBD patients.