The interface between the intestinal lumen and the mucosa is the location where the majority of ingested immunogenic particles face the scrutiny of the vast gastrointestinal immune system. Upon regular physiological conditions, the intestinal micro-flora and the epithelial barrier are well prepared to process daily a huge amount of food-derived antigens and non-immunogenic particles. Similarly, they are ready to prevent environmental toxins and microbial antigens to penetrate further and interact with the mucosal-associated immune system. These functions promote the development of proper immune responses and oral tolerance and prevent disease and inflammation. Brain-gut axis structures participate in the processing and execution of response signals to external and internal stimuli. The brain-gut axis integrates local and distant regulatory networks and super-systems that serve key housekeeping physiological functions including the balanced functioning of the intestinal barrier. Disturbance of the brain-gut axis may induce intestinal barrier dysfunction, increasing the risk of uncontrolled immunological reactions, which may indeed trigger transient mucosal inflammation and gut disease. There is a large body of evidence indicating that stress, through the brain-gut axis, may cause intestinal barrier dysfunction, mainly via the systemic and peripheral release of corticotropin-releasing factor. In this review, we describe the role of stress and corticotropin-releasing factor in the regulation of gastrointestinal permeability, and discuss the link to both health and pathological conditions.
The epithelium that lines our digestive tract harbors 100 trillion bacteria1 and mediates our relationship to the world outside. Our intestine is constantly exposed to a wide variety of immunogenic particles and microorganism-derived antigens. Its surface area is large enough to process over 50 tons of food in a lifetime. Intestinal and microbial components are strongly linked to each other and represent the most important barrier to limit luminal antigens from travelling across and reach the mucosal-associated immune system. They also play a critical role in training and modulating our immune system, helping it to distinguish between friend and foe.2,3 The absence of constructive engagement between microbes and intestinal mucosa may render epithelial barrier more permeable, allowing it to be breached. When this happens, resident cells become over-activated, and a cascade of pro-inflammatory signals is initiated.4,5 Altered intestinal permeability has large implications for human health, being involved in the origin and development of many digestive and non-digestive diseases. However, up to date, it is not clear whether intestinal barrier dysfunction is a primary event, an epiphenomenon or a consequence in the pathogenesis of these disorders.
The regulation of intestinal permeability involves terminals from ascending and descending pathways from the autonomic nervous system and the central nervous system, the hypothalamic pituitary-adrenal axis, and the enteric nervous system as well. The vast and versatile array of bidirectional and integrative communications between the brain and the gut allows the brain to respond to internal and external signals and, in return, to modify the autonomic and enteric nervous systems to keep intestinal permeability tight.
Physical and psychological stresses represent convincing evidence of the influence of the brain-gut axis on the intestinal barrier function. In fact, stress has been associated with reactivation of inflammatory and functional gastrointestinal disorders mainly through disruption of the intestinal barrier6 in both human and animal models. The stress response is vehiculated via 2 main systems: the sympatho-adrenomedullary system and the hypothalamic pituitary-adrenal axis. Stress stimulates the parvocellular neurons in the paraventricular nucleus of the hypothalamus to secrete corticotropin-releasing factor (CRF), and arginine vasopressin. CRF promotes the synthesis and release of adenocortico-tropic hormone in the anterior pituitary. The adenocorticotropic hormone, in turn, activates the adrenal cortex to induce a temporary rise in blood levels of cortisol and corticosterone7 and also the release of catecholamines by the adrenal medulla. It has been believed for a long time that central secretion of CRF was the main and unique mediator of the majority of the endocrine, behavioral and gastrointestinal changes induced by stress.8 However, we will show that peripheral release of CRF also plays a key role in the regulation of gastrointestinal permeability, and discuss the link to both health and pathological conditions.
The intestinal mucosal barrier includes several consecutive layers, from the outermost microbiota, through the external mucus, the epithelium, and down to the innermost lamina propria.9 The epithelial layer is described as a continuous polarized monolayer of columnar cells that separates the intestinal lumen from the internal milieu. Aside from epithelial cells, a variety of different cell types are also intermingled with enterocytes, including goblet cells, Paneth cells, enteroendocrine cells, and M cells. Within the lamina propria we can find blood and lymph vessels, and a plethora of distinct immune cells such as plasma cells, lymphocytes, macrophages, eosinophils, mast cells, dendritic cells, and a significant number of intrinsic and extrinsic nerve terminals.10 All of these components are exquisitely reactive and adaptive and display critical effector and modulatory functions. These functions are relevant for the control of inflammation, absorption and secretion, transport of macromolecules, and metabolic processes.11 Moreover, most of the cellular components have been shown to express receptors for CRF.12,13 But the job of the intestinal barrier now appears to be far more nuanced and complex as it communicates multidirectionally with the immune system and the microbes. Communication is developed through the release of an extensive array of chemical mediators, namely neuropeptides, neurohormones, neurotransmitters, cytokines, chemokines, growth factors, and other regulatory molecules.14,15
Enterocytes are tightly bonded to each other sealing the para-cellular space through the apical junctional complex, composed of tight junctions (TJs), adherens junctions, and desmosomes.16 The transmembrane TJ proteins occludin and claudins form complex protein systems, which interact with zonula occludens proteins that bind to the actin cytoskeleton. When actin contracts, it leads to increased permeability to electrolytes and small molecules.17 The paracellular space is not fully impermeable to molecules and antigens, allowing a controlled amount of small particles (less than 400 daltons) to penetrate across to reach the lamina propria,9 a phenomenon that plays a key role in the induction of immune tolerance. This passage of molecules also takes place through the transcellular pathway via endocytosis or exocytosis.18 Disruption of the intestinal barrier leads to uncontrolled flux of luminal antigens and possibly microbes across the epithelium, which may trigger immune activation and sepsis, and also lead to the development of chronic inflammation in the gut.6,19,20 Therefore, the tight regulation of intestinal permeability emerges as a central mechanism to prevent inflammatory diseases. Numerous pathogens and toxins, hormones and neurotransmitters, and gastrointestinal and non-gastrointestinal diseases have been associated with an augmented intestinal permeability.21 Among them, stress hormones and neurotransmitters have been consistently shown to modulate ion and water secretion, intestinal permeability, mucus secretion, and also intestinal flora.22?25
Stress represents a threat to the internal homeostasis that initiates a systemic coordinated response driven by the autonomic, endocrine, and immune systems to maintain stability. CRF and other members of the CRF signaling family, including urocortin (Ucn) 1, Ucn2, and Ucn3, are the best known and most important neuroendocrine mediators of the stress response.26,27 Central release of CRF and urocortins mediates autonomic, hormonal, and behavioural responses to stress and at the gastrointestinal tract, stimulates the enteric nervous system to modulate gastrointestinal motility and secretion.28?30 In addition, immune cells, regional sensory and sympathetic nerves, enterochromaffin cells, and enteric cells release CRF and urocortins within the gastrointestinal tract28 to modulate mucosal function and gastrointestinal motility.29
Different type of stress, acute or chronic, physical or psychological, have been shown to influence properties of the intestinal barrier function, including ion and water secretion, intestinal permeability, mucus secretion, and also intestinal flora in both human and animal models.6,31?35 Similarly, both central and peripheral administration of exogenous CRF have been shown to mimic the effects of acute stress in the gastrointestinal tract inducing mucin release,36 and increasing ion and water secretion and intestinal permeability.37,38 Figure summarizes the main effects of stress on the intestinal barrier and the role of CRF in gastrointestinal permeability.
The intestinal mucosal surface is covered by a layer of mucus gel containing antibacterial peptides and digestive enzymes.39,40 Mucus protects the epithelial lining from adhesion and invasion by attacking microorganisms, and other antigens present in the intestinal lumen.41 Mucus is produced by goblet cells, which express the CRF receptor 1 (CRF1).42 Stress has been shown to increase mucin release in colonic explants from rats submitted to immobilization, and to enhance rat mast cell protease II and prostaglandin E2 secretion. Intravenous or intracerebral injection of CRF in non-stressed rats reproduced stress-associated changes whereas administration of the CRF antagonist α-helical-CRF9?41 or the mast cell stabilizer lodoxamide inhibited them.36 In addition, those changes were abolished in mast-cell deficient mice, highlighting the key role of the CRF-mast cell axis in stress-mediated mucin release.43 On the contrary, rats submitted to chronic stress display mucus depletion and reduced number of goblet cells in the distal colon in association with increased bacterial adhesion and penetration into enterocytes,44,45 what could represent a step towards intestinal barrier dysfunction.
Ion and water secretions also help to prevent the adhesion of pathogens and noxious substances to the mucosal surface, and dilute and flush out them down the gut to minimize their penetration to the lamina propria. Convergent evidences from studies in animal models indicate that both water and ion secretions increase in the intestinal tract in response to acute and chronic stress24,46 and in response to peripheral administration of CRF47 whereas selective CRF1 receptor agonists prevent this response to stress.48,49 Increased baseline short-circuit current (
Both the acute intraperitoneal administration of CRF and chronic peripheral administration of CRF, the selective CRF1 receptor agonist stressin-1 and the selective CRF2 receptor ligand, Ucn3, increase the basal rat colonic
Apart from ex vivo and in vivo studies, convincing evidence for a direct effect of CRF peptides on intestinal epithelial secretory response comes from in vitro work showing a consistent increase in
Experimental studies in humans reveal that acute physical or psychological stress modulate the intestinal barrier. More than 20 years ago, Barclay and Turnberg et al,62,63 using segmental perfusion techniques in the human jejunum, showed that acute stress reduced net water absorption or increased secretion, supporting the role of the central and the autonomic nervous system in the control of intestinal function. In particular, a reduction in water absorption coupled with net sodium/chloride secretion was observed in healthy subjects subjected to psychological stress induced by dichotomous listening.62,63 Later, Santos et al64 showed that cold pain stress increased jejunal water secretion in healthy subjects and in patients with food allergy. These observations have been confirmed and extended recently: using the cold press-or test. The increase in intestinal water secretion during cold pain stress was higher in healthy female volunteers with lower background stress levels, yet blunted in those with elevated background stress.65 These results suggest that chronic life stress could lead to the loss of regulatory mechanisms in healthy individuals, as previously described in the jejunum of chronic stressed rats after neural stimulation.50 CRF1 receptor is expressed throughout the gastrointestinal tract in healthy subjects, and located in the lamina propria, mostly in macrophages, epithelial cells and enteric neurons, with the highest levels in the ileum and rectum and the lowest level in the colon.66 CRF2 receptors are located in the lamina propria and in the epithelial cells of the distal/sigmoid biopsy samples in healthy subjects.67 In contrast to animal studies, the addition of CRF to human colonic biopsies from healthy individuals mounted in Ussing chambers did not show any effect on
Numerous studies demonstrate the enhancement of intestinal permeability by acute and chronic stressors. In rodents, acute or subchronic stressors, including restraint stress, cold restraint stress, WAS, mild noise stress, and mixed restraint and acoustic stress, increased tissue conductance, and the fluxes of 3H mannitol,51 Cr-labeled EDTA, horseradish peroxidase, in the jejunum, ileum, and colon.32,50,54,69?72 These results indicate the ability of stress to modulate paracellular and transcellular transport of ions, and small and large macromolecules. Multiple mechanisms have been invocated to explain stress-mediated transport abnormalities. The list includes acetylcholine release, histamine, glucagon-like peptide-2, myosin light chain kinase, cytokines such as interferon-γ, interleukin (IL)-4 and IL-13, and mast cell activation, among the most relevant.32,61,71,73?75 Moreover, increased intestinal permeability, measured by lactulose, mannitol and sucralose ratios, was found in rats’ small intestine, following acute swimming stress.23 This increase was glucocorticoid-mediated because adrenalectomy and pharmacologic blockade of glucocorticoid receptors inhibited the response, and dexamethasone increased gastrointestinal permeability in control rats.23
Similarly, rats models of chronic stress (maternal deprivation,55,77?79 WAS,45 and crowding stress52) also show increased ionic permeability, and macromolecular permeability throughout the intestine. The increase was mediated partly by mechanisms including, muscarinic and nicotinic pathways, activation and number of mast cells and the release of nerve growth factor.79 Stress-induced enhancement of colonic permeability was mimicked by exogenous administration of CRF,33 and abolished by pretreatment with the peripheral administration of the non-selective CRF antagonists astressin or α-helical CRF9?41.37,55,80 Likely, the selective CRF1 receptor agonist, cortagine,48 the selective CRF1 receptor antagonist, SSR-125543,80 and the selective CRF2 receptor antagonist, antisauvagine-30,78 reduced the response, supporting the participation of both CRF receptors in the modulation of colonic permeability.
More recently, studies in several cultured epithelial cell lines, including intestinal human HT-29 cells, showed that CRF induced the upregulation of the endotoxin receptor, toll-like receptor 4 (TLR4), and that pretreatment with the CRF2 receptor antagonist, antalarmin, abrogated the response. The concomitant presence of CRF and lipopolysaccharide increased permeability to horseradish peroxidase and decreased transepithelial resistance. This effect was abolished by the CRF2 receptor antagonist, astressin2-B, and was mediated through the upregulation of claudin-2. Similarly, the expression of TLR4 and claudin-2 increased in the small and large intestine of pregnant mice submitted to 10 days of WAS, as was enhanced the permeability to horseradish peroxidase, and the effect abolished by α-helical CRF9?4181 and anti-claudin antibodies. Interestingly, rats submitted to WAS for 10 days, showed higher sensitization to orally-delivered horse-radish peroxidase, as shown by enhanced intestinal permeability upon antigen re-exposure. This effect was prevented by CRF blockade with by α-helical CRF9?41.82
Another line of evidence in animal models highlights the relevance of stress-CRF-mast cell axis in the regulation of intestinal permeability. Elevation of intestinal permeability, in response to restraint stress or intraperitoneal administration of CRF, was reduced or abolished after pretreatment with doxantrazole.59 Similarly, pharmacological inhibition of mast cell activation has been shown to inhibit stress-induced increased intestinal permeability in different animal models.37,58,59,83 Stress/CRF-induced changes in the rat intestinal permeability were also significantly reduced in tissues from mast-cell deficient rats.45,51,59,84 Teitelbaum et al57 also found CRF to cause mast cell hyperplasia and abnormal bacterial attachment and/or penetration into the rat mucosa except in mast-cell deficient rats. The authors revealed that stimulation of CRF1 receptor induced an elevated secretory state, while CRF2 receptor stimulation was associated with permeability dysfunction.57 Although it is clear that mast cell activation disrupts intestinal barrier and increases intestinal permeability, the ultimate mechanisms remain to be elucidated. In this sense, several mast cell mediators have been involved. Tumor necrosis factor-α (TNF-α) has been shown to increase intestinal permeability through nuclear factor-kappa B activation in association with nuclear factor-kappa B-dependent downregulation of zona occludens protein-1 expression and alteration in junctional localization in Caco-2 cells,85 and through myosin light chain kinase phosphorylation.86 More recently, using a porcine ex vivo intestinal model, Overman et al87 found that CRF increased intestinal paracellular permeability via mast cell dependent release of TNF-α and proteases. On the other hand, the tryptase released by mast cells upon activation has been shown to induce TJs disassembly through the activation of proteinase-activated receptor-2 of the epithelial cells.88?90 These receptors can modulate enteric neurotransmission, secretion, motility, epithelial permeability, and visceral sensitivity, and are also known to regulate intestinal inflammation.91 Anatomical contacts between mast cells and enteric nerve fibers have been demonstrated in the human gastrointestinal mucosa and inflammation multiplies these contacts.92,93 Mast cells communicate, bidirectionally, with both the enteric, autonomic and central nervous system through mast cell mediators and neuropeptides.94 In fact, human mast cells synthesize and secrete both CRF and Ucn in response to immunoglobulin E receptor activation.95 Mast cells also express CRF receptors, whose activation leads to the release of cytokines and other pro-inflammatory mediators. CRF enhances the trans-cellular uptake of macromolecules in the human colonic mucosa in vitro, via CRF receptor subtypes CRF1 and CRF2 expressed on subepithelial mast cells.58 CRF1 and CRF2 receptors have been described in human umbilical cord blood-derived mast cells, while only CRF1 receptor has been found in the human leukemic mast cell HMC-1 line. CRF stimulation induces the secretion of vascular endothelial growth factor without tryptase, histamine, IL-6, IL-8, or TNF-α release through CRF1 receptor.96 More recently, acute stress- induced bladder vascular permeability and vascular endothelial growth factor release have been shown to be dependent on CRF2 receptor.97
It is important to note here the ability of eosinophils to alter intestinal permeability and the mechanisms and mediators involved.98 Barrier function has been shown to be affected by eosinophil-derived major basic protein (MBP) through downregulation of occludin in a mast cell independent fashion.98 The neuropeptide substance P (SP) has been found to induce the release of vasoactive mediators from mast cells, macrophages, and T cells, contributing to chloride secretion, enhanced intestinal permeability, and vascular leakiness.99,100 SP increases the expression of CRF1 receptor in mast cells, and in turn, CRF induces the expression of the specific receptor neurokinin (NK)-1.101 More recently, psychological stress, through SP release, mediated stress-induced CRF expression in mice eosinophils and eosinophil-derived CRF was responsible for mast cell activation and epithelial barrier dysfunction. In this work, a cell line of eosinophils was treated with a number of stress mediators, but only SP induced CRF release via NK-1 and NK-2. Moreover, priming of eosinophils with SP resulted in mast cell activation through eosinophil-derived CRF that in turn induced intestinal barrier dysfunction.102 Wallon et al83 examined non-inflamed colonic mucosal biopses from patients with ulcerative colitis and found that eosinophils displayed immunoreactivity to CRF. In addition, in co-culture studies, carbachol activation of eosinophils induced CRF release and subsequent activation of mast cells, which increased permeability of epithelial cells to macromolecules.83 Similar to mast cells, eosinophils have also been shown to localize close to airway nerve terminals in patients with asthma as well as in animal models of bronchial hyperreactivity.103,104 This proximity to nerve terminals is described in certain digestive disorders,105,106 providing the anatomical substrate to understand the relevance of the mast cell-eosinophil-stress/CRF axis in the regulation of epithelial permeability and the initiation of immune and inflammatory diseases.
Stress-induced increase in gut paracellular permeability has also been shown to depend on CRF1 receptor-mediated mast cell release of nerve growth factor (NGF). Maternal deprivation has been shown to enhance colon permeability in association with elevated NGF expression.76 A subsequent study from the same group showed that CRF, acting through its receptor CRF1 receptor, stimulated NGF release from mast cells, which in turn increased gut paracellular permeability.80 Dendritic cells are relevant for the regulation of intestinal immune function and permeability through CRF production, a process augmented by commensal bacteria.107 Similarly, CRF1 and CRF2 receptor agonists exert a biphasic effect on macrophages. During the early stages of the inflammatory response, they suppress TNF-α release via induction of cyclooxygenase-2/prostaglandin E2 while later on they induce TNF-α transcription.108 Unfortunately, the role of CRF-mediated activation of macrophages and dendritic cells in stress-related intestinal barrier dysfunction is largely ignored.
Vanuytsel et al38 have recently shown that, both psychological stress (public speech) and a single intravenous bolus of CRF (100 μg) increased small intestinal permeability in healthy humans, measured by the lactulose/mannitol ratio. Two weeks pretreatment with 800 mg/day of the mast cell stabilizer, disodium cromoglycate, blocked the effect of both stress and CRF, invoking the participation of the CRF-mast cell axis in this response.38 We have also observed in healthy volunteers that cold pain stress enhanced both the blood-to-lumen albumin65 ratio and the blood-to-lumen mannitol and xylose permeability.109 This response was mainly observed in females with higher background stress levels, suggesting an impaired epithelial response to incoming stressful stimuli in this group. In addition, acute cold stress was shown to induce a significant release of α-defensin in the jejunum in this study, supporting the possibility that stress might affect this protective pathway in the gut, as have been shown in the skin.110 Furthermore, we have also found that cold stress evoked a differential gender-determined increase in human intestinal macromolecular permeability.111 This enhanced permeability could lead to excessive uptake of luminal antigens and bacterial products that may initiate an inflammatory response in the mucosa.112 Other studies showing the effect of stress and CRF on intestinal and extra-intestinal permeability are shown in Tables 1 and 2, respectively.
The growing acknowledgment of the scientific community to the role of intestinal permeability in keeping health and well-being and its relation to the origin of digestive and extradigestive disorders is becoming more and more universal for clinicians. Increased permeability and breakdown of intestinal barrier have been implicated in the origin of gastrointestinal and liver disorders, including celiac disease,113 inflammatory bowel disease,114?118 food allergy,119 acute pancreatitis,120 irritable bowel syndrome (IBS),121,122 functional dyspepsia,123 infectious diarrheal syndromes,124 primary biliary cirrhosis, and primary sclerosing cholangitis,125 liver cirrhosis,126 alcoholic liver disease,127 liver encephalopathy,128 and gastroesophageal reflux disease,129?131 among others. However, whether enhanced gut permeability is an early manifestation of disease, a central step in disease pathogenesis, or a simple epiphenomenon, and its relationship with life stress in the clinical setting, remains to be elucidated. A few examples can illustrate this controversy.
Interestingly, life events may favor the clinical appearance of celiac disease,132 and cellular stress, through MHC class I chain related genes A and B and endoplasmic reticulum stress pathways, is linked to the disregulation of mucosal homeostasis.133,134 Early in the 80s, celiac patients were shown to display increased intestinal permeability that normalized after several months on a gluten-free diet.113 More precise in vitro investigations revealed that although strict gluten withdrawal restored intestinal histology, a subjacent defect in mucosal permeability, measured by cellobiose/mannitol ratio, was transiently induced by short exposure to gluten, suggesting that increased intestinal permeability in celiac disease could be a primary defect.135 Recently, this defect in intestinal permeability has been related to altered expression of TJ genes related to permeability, polarity, and cell proliferation in active celiac disease,136 partly through the activation of the zonulin pathway in a MyD88-dependent fashion.137,138 Again, the majority of genes returned to normal after 2 years of gluten eviction with the exception of PPP2R3A, possibly indicating a constitutive defect in these patients.136 Moreover, unlike celiac disease, gluten sensitivity is not associated with increased intestinal permeability.139
Increased intestinal permeability has also been described in patients with inflammatory bowel disease. Enhanced intestinal permeability in this population is considered an initial event because it is increased in first-degree relatives of Crohn’s disease patients.140,141 Several observations link clinical relapses to the increase in intestinal permeability142,143 and to life stress,144,145 partly through mast cell activation146 and the release of CRF from eosinophils83 and neighboring nerves.68 Moreover, inflammatory bowel disease patients show associations with genes involved in the regulation of intestinal barrier and various alterations of the transmembrane and intracytoplasmic proteins.147,148 In addition, increased CRF1 receptor66 and Ucn expression, but decreased CRF2 receptor67 expression have been shown in the colonic mucosa of active ulcerative colitis patients.149,150
Accumulating evidence also indicates that IBS is linked to abnormal intestinal permeability,151?153 CRF release and life stress154?157 in close association with low-grade mucosal inflammation and immune activation.158 Several groups have provided preliminary evidence linking clinical manifestations of IBS to structural abnormalities of the apical junctional complex in both the small159,160 and large bowel mucosa.81,161?163 Furthermore, unpublished observations from our group indicate that a single intravenous bolus of CRF (100 μg) increased intestinal permeability, measured as the blood-to-lumen albumin ratio, in healthy subjects and in IBS patients through mast cell activation.164
In the same line, increased intestinal permeability and stress/CRF axis have also been involved in the pathophysiology of food allergy, as these patients display an enhancement of intestinal permeability even in the absence of food allergens165 and stress facilitates sensitization to luminal antigens.82 In fact, patients following immunosuppressive treatment have been shown to develop new-onset food allergies that may be related to the increase in intestinal permeability induced by treatment.166
Many other stress-related conditions such as severe burn,167,168 hemorragic shock,169 chronic kidney disease,170 type 1 diabetes and the metabolic syndrome,171,172 neuropsychiatric disorders,173 autism,174 autoimmune thyroiditis,175 IgA nephropathy,176 patients with primary immunodeficiencies,177 and sepsis178 have been associated with increased intestinal permeability.
The ability of stress and peripheral CRF to affect intestinal epithelial function and, particularly, intestinal permeability, is well documented in human and several animal species. Interestingly, apart from the direct effects on the enterocytes, the stress-induced increase of intestinal permeability is mediated via recruitment and activation of mast cells, eosinophils, macrophages and other mononuclear cells, and implicates both CRFR1 and CRF2 receptors.28
There is abundant literature supporting the link between life stress and the origin and clinical course of several human disorders,179 however, the ultimate clinical relevance of stress-CRF-mediated dysregulation of gastrointestinal permeability, remains at present mostly intuitive, with few exceptions. More work in this area is needed to confirm the findings. In this sense, studies with CRF and related peptides and the development of new antagonists for human use, will offer the opportunity to test this hypoyhesis.