NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Kvietys PR. The Gastrointestinal Circulation. San Rafael (CA): Morgan & Claypool Life Sciences; 2010.
6.1. ULTRASTRUCTURAL PATHWAYS
The exchange vessels (capillaries) of the gastrointestinal mucosa are fenestrated and relatively permeable to small molecules (e.g., glucose), yet restrict the transcapillary movement of large molecules (e.g., albumin). This allows for the absorption of hydrolytic products of food digestion without compromising the oncotic pressure gradient governing transcapillary fluid movement and edema formation. Ultrastructural studies (e.g., electron microscopy) have identified potential pathways for solute and water exchange across these fenestrated capillaries (Figure 6.1) [28]. Similar pathways for solute and water movement exist in the continuous capillaries of the muscularis, with the obvious exception of fenestrae. A brief description of the enumerated pathways depicted in Figure 6.1 is as follows.
FIGURE 6.1
Schematic of solute and fluid transport pathways in the capillaries of the gastrointestinal tract. Transport pathways: 1, cell membrane; 2, open fenestrae; 3, diaphragmed fenestrae; 4, intercellular junction; 5, pinocytotic vesicles; 6, transendothelial (more...)
6.1.1. Endothelial Cell Membrane
Lipid-soluble substances (O2 and CO2) and very small nonpolar hydrophilic substances (urea) can traverse the fenestrated capillaries by diffusing through the cell membrane proper. The permeability of the endothelial cell has been described in terms of small perforations (4–10 Å radius) in the luminal and abluminal lipid bilayers and assuming negligible resistance to diffusion offered by the cytoplasm. This pathway accounts for less than 10% of the capillary hydraulic conductivity (transcapillary volume flow rate per unit pressure gradient). Thus, the bulk of the transcapillary fluid and solute movement occurs through extracellular pathways [28,42,196].
6.1.2. Fenestrae
Fenestrae are circular openings (200 to 300 Å in radius) in the capillary endothelium. There is an asymmetry in the relative number of fenestrae both along the length of the capillaries as well as with respect to transporting epithelia. The frequency of the fenestrae increases from arterial to venous ends of the capillary, and they are preferentially oriented toward the base of the transporting epithelia. Further, in the small intestine, the capillaries of the villus tips and the crypt region contain more fenestrae per cross-section than the capillaries located at the base of the villus. Tracer molecules ranging in size from 25 to 120 Å can pass through the fenestrae. Approximately half of the fenestrae are closed by a diaphragm, which is believed to offer some restriction to the movement of solutes [12,28,197–202].
6.1.3. Pinocytotic Vesicles
The capillary endothelium also contains plasmalemmal (pinocytotic) vesicles with an internal radius of approximately 250 Å. As is the case with fenestrae, their frequency increases from the arteriolar to venular end. The vesicles are believed to be mobile structures opening on one side of the endothelium (caveolae), acquiring solutes and fluid, moving across the endothelial cytoplasm and discharging their contents on the opposite side. Solutes 25 to 150 Å are readily transported by these vesicles; however, the overall transport of macromolecules by vesicular transport is much slower than their movement through fenestrae. One or more vesicles (caveolae) can fuse and form an open transendothelial channel; however, they have strictures at the points of fusion that reduces their internal radius from 250 Å to between 50 and 200 Å. The effective radius is further reduced in some channels by the presence of diaphragms at the points of fusion [197,201–204].
6.1.4. Interendothelial Cell Junctions
The intercellular junctions can be either open or closed; the frequency of open intercellular junctions increases from the arteriolar to the venular end of the capillaries. The open junctions behave as if they are channels 20 to 60 Å in width, while the closed channels are functionally impermeable to solutes of 20 Å diameter [42,199,202].
6.1.5. Glycocalyx
An extra-endothelial barrier to solute movement may be the glycocalyx on the luminal endothelial cell surface [114]. In the fenestrated capillaries of the gastrointestinal mucosa, the glycocalyx not only covers the endothelial cell proper, but also appears to line the luminal portion of the interendothelial junction and covers the surface of fenestral diaphragms [205–207]. Enzymatic removal of the glycocalyx (e.g., pronase) can decrease the hydraulic conductivity of fenestrated capillaries [208].
6.1.6. Basement Membrane
The basement membrane surrounding the endothelium also possesses restrictive properties. Tracers of molecular radius ranging from 25 to 150 Å can readily cross the endothelial lining. Tracers 25–55 Å in radius do not appear to be impeded in their further movement across the basement membrane. Molecules of between 62 and 150 Å are temporarily delayed at the level of the basement membrane, at times accumulating in clusters under permeable fenestrae [28,197,199,209].
6.2. Physiological (Functional) Pathways
Fluid exchange can occur through any of the multiple small ultrastructural “channels,” but exchange of solutes and proteins of different molecular weights will be strictly dependent on channels whose diameters exceed those of the solutes/proteins. In addition to the permeability (porosity) of the capillaries, transcapillary movement of solutes can be influenced by the capillary surface area. Thus, a major physiological approach to this issue has been to assess the transcapillary movement of solutes of different molecular weights (radii), while minimizing the effects of capillary surface area. To this end, a variety of techniques have been used to assess gastrointestinal microvascular permeability to solutes, each with its own set of advantages and limitations [204,210]. Of the approaches used to assess the permeability of the gastrointestinal tract microcirculation, the indicator dilution technique to study small (<37 Å´) solute permeability and the steady-state analysis of lymph protein flux to assess macromolecular (>37 Å) permeability have yielded the most useful information [28,42,197]. Mathematical modeling approaches using experimentally and theoretically derived information have yielded estimates of “equivalent pore radii” to describe the functional permeability characteristics of the gastrointestinal capillaries [196,204,210].
6.2.1. Small Solutes
The multiple indicator dilution technique uses radioactive diffusible tracer molecules to assess their extraction by capillaries during a single pass through the capillary bed [211]. Pairs of diffusible tracers and a nondiffusible (vascular tracer) are simultaneously injected into the arterial supply of isolated organs, and the relative concentrations of the tracers in the venous effluent are assessed. The relative concentration of an isotope in venous blood is obtained by expressing the concentration relative to the original concentration in the injectate. The extraction (E) of a diffusible tracer is calculated by
E = [CV(t) – CD(t)] / CV (t)
where CV(t) equals the relative concentration of the vascular tracer in the venous sample and is equivalent to the relative concentration of the diffusible tracer in arterial blood at time t, and CD(t) equals the relative concentration of the diffusible tracer in venous blood at time t.
When plasma flow (blood volume flow corrected for Hct) to the isolated preparations of the stomach and small intestine are measured, the permeability–surface area product (PS) for the diffusible tracers can be calculated by
PS = Q P ln(1 – E)
where Qp is plasma flow through the preparation.
The use of multiple tracers of different sizes has yielded information relevant to the permselectivity of the gastric [212] and small intestinal [213] capillaries. As mentioned above, the mucosal and muscularis microcirculations are arranged in parallel, and the capillaries of the mucosa have fenestrae, while those of the muscularis do not. Since the mucosal circulation receives approximately 80% of the intramural blood flow, the permeability of the mucosal capillaries will dominate the outcome of the experiments. In these studies, isoproterenol was used to preferentially increase mucosal blood flow, increasing the probability that the experimental outcome reflected primarily the permeability characteristics of the mucosal capillaries. Under these conditions, the ratios of the PS values for simultaneously injected inulin (15 Å) and β-lactoglobulin (28 Å) were greater than the ratios of their free diffusion coefficients. This indicates that the extraction of β-lactoglobulin was more restricted by the capillary wall than that of inulin. Since the diffusible tracers are injected simultaneously, their PS ratios are assumed to be the same as their permeability ratios, given that the surface area is the same for both tracers and cancels out. Thus, based on the relationships among the permeabilities, diffusion coefficients, and radii of these two molecules, [214] an equivalent small pore radius of 53 and 59 Å is predicted for the mucosal capillaries of the stomach [212] and small intestine, [213] respectively.
6.2.2. Macromolecules
Steady-state analysis of lymphatic protein flux has been used to provide information on the permeability of gastrointestinal capillaries to macromolecules (e.g., endogenous plasma proteins) [210,212,215]. A major assumption of this approach is that the concentration of macromolecules in the lymph draining the gastrointestinal tract is identical to that of interstitial fluid [204]. The validity of this assumption is based on both ultrastructural studies of initial lymphatics indicating that the endothelium is discontinuous and the basement membrane fragmented [198] and physiologic studies indicating that the interstitial and lymph protein concentration is similar [216]. Further, complicating the lymphatic protein flux approach is the anatomical regions drained by the lymphatics. The draining lymphatics in the mesentery contain contributions from both the mucosal and muscularis regions. However, the protein concentration of the rat villus lacteal and the collecting lymphatics are similar [217]. Collectively, these observations support the contention that protein concentration profile of collecting lymphatics provides a reasonable estimate of that in the mucosal interstitial fluid.
Under resting conditions (normal microvascular pressures), transcapillary exchange of macromolecules occurs by both convection and diffusion. Thus, an assessment of the lymph-to-plasma protein concentration ratio (CL/CP) does not allow for an accurate approximation of the sieving characteristics of capillaries. In order for a lymph protein flux analysis to yield useful information regarding the permeability characteristics of the capillaries, the convective flux must be maximized and the diffusive flux minimized. An experimental approach, which increases the convective movement of macromolecules across the capillaries to such an extent that their diffusive exchange becomes negligible, is acute venous hypertension. When capillary filtration (or lymph flow) is increased by graded venous hypertension, CL/CP for total plasma proteins progressively decreases (Figure 6.2). This portion of the relationship is termed “filtration rate-dependent.” At high lymph flows, CL/CP reaches steady-state levels, i.e., becomes “filtration rate-independent.” When CL /CP is filtration rate (or lymph flow)-independent, lymph protein flux analyses can provide information on the true sieving characteristics of the capillary wall. Specifically, the level at which CL /CP becomes filtration rate-independent can provide a reasonable estimate of the osmotic reflection coefficient, σd (1 − CL/CP in Figure 6.3) [210,215].
FIGURE 6.2
Relationship between lymph-to-plasma ratio for total protein concentration (L/P) and lymph flow. Intestinal lymph flow was increased by graded increases in venous pressure in cats (circles) and rats (squares). Used with permission from Handbook of Physiology, (more...)
FIGURE 6.3
Relationship between the ratio of lymph to plasma total protein concentration (L/P) and lymph flow in the small intestine during graded acute venous hypertension. The solid line depicts the control situation (nontransporting small bowel). The experimental (more...)
The relationship between total protein CL/CP and lymph flow rate depicted in Figure 6.2 was derived from studies in the cat and rat small intestine. Similar relationships have been obtained in the feline stomach and canine small intestine and colon. The σd values for total protein and various endogenous proteins of different sizes in the fenestrated capillaries of the gastrointestinal tract [30,212,215,218] are compared to corresponding σd values obtained for the continuous capillaries of the hindpaw [219] in Table 6.1. As shown in Table 6.1, σd increases as the radii of the endogenous proteins increase, indicating permselectivity of gastrointestinal capillaries. Based on the available data, species differences may exist, and caution should be exercised in making comparisons across species. Further, in the dog, σd values for total protein in the small and large intestinal capillaries are similar and not dramatically different from the continuous capillaries of the hindpaw.
TABLE 6.1
Osmotic reflection coefficients for different-sized proteins in the feline and canine gastrointestinal tract.
The σd values for different-sized macromolecules may be used to quantitatively describe the permeability characteristics of the gastrointestinal microvasculature in terms of “equivalent pores” using a graphic analysis [204,210,220]. An example of such an analysis for the small intestine [215] is shown in Figure 6.4, where 1 − σd is plotted as a function of solute radius. The analysis involves fitting the data with two sets of equivalent pores: small and large. First, a theoretical large-pore curve is fitted to the data points representing the large molecules. Subsequently, by a “curve-peeling” process, the resulting values for 1 − σd for small solutes are fitted with a smaller theoretical pore curve. The ordinate intercept predicts the percentage of the total hydraulic conductance occurring through each set of pores. The relative areas of the small and large-pore populations can be assessed using
FIGURE 6.4
Application of pore-stripping analysis to sd values for plasma proteins of different sizes. The data are fitted using two sets of equivalent pores. Dashed lines depict analysis of data from nontransporting small intestine, while solid lines depict analysis (more...)
As /Al = (Fs /Fl) × [(rl )2 / (rs )2]
where A equals pore area, F equals the fraction of hydraulic flow through the pores, r equals pore radius, and the subscripts s and l refer to small and large pores, respectively. The relative frequency of each type of pore can be calculated by
Ns /N1 = (As /Al) × [(rl )2/(rs )2]
where Ns /Nl is the ratio of small-to-large pore number.
This type of mathematical analysis has also been applied to data obtained for the stomach [212] and colon [218] and cumulatively presented in Table 6.2. Several general features of the permselectivity of the gastrointestinal capillaries can gleaned from this information. The effective radius of the small and large pores do not vary dramatically between various regions of the gastrointestinal tract (15–30%), and the radius of the small pores derived using lymph protein flux data is similar to that predicted by the multiple indicator dilution technique, i.e., 53–59 Å. Further, the relative density (number and area occupied) by the small-pore component is much greater than the large-pore component, the disparity being the greatest in the capillaries of the small intestine. Also of interest is that the dimensions and relative frequencies of equivalent pores in the fenestrated capillaries of the gastrointestinal tract are similar to the continuous capillaries of the hind-paw.
TABLE 6.2
Dimensions and relative frequencies of equivalent pores in the fenestrated capillaries of the gastrointestinal tract compared to continuous capillaries of the hind-paw.
In general, agents or conditions that increase intestinal capillary permeability do so by increasing the dimensions of the large pores [210]. One condition that appears to selectively influence the small-pore population is severe arterial hypoxemia ( pO2 of 35 mmHg) [221]. In this situation, the small pore size was increased (from 59 to 67 Å) as assessed by the multiple indicator dilution technique. However, lymph protein flux did not change, indicating that the dimensions of the large pores were not affected.
The capillaries of the gastrointestinal tract may also exhibit charge selectivity. Studies in the small intestine indicate endogenous lactate dehydrogenase (LDH; 42 Å radius) isoenzymes are selectively restricted by the capillaries based on their isoelectric points [222]. The σd for LDH isoenzymes decreased from 0.95 for the most positive isoenzyme (isoelectric point, 8.3) to 0.71 for the most negative isoenzyme (isoelectric point, 5.2). Similar results were noted using exogenously administered dextran molecules (35–36 Å radius), i.e., the steady-state CL/CP of the neutral dextran was twice that of the positive dextran. Other studies assessing binding of cationized ferritin to intestinal mucosal capillaries demonstrated that very little binding of this cationic molecule occurs, indicating charge repulsion by the endothelium [223]. Collectively, these observations indicate that the small intestinal microvessels behave as a positively charged barrier. This is in stark contrast to the renal glomerular microcirculation, which behaves as a negatively charged barrier [224]. It has been suggested that the cationic nature of the small intestinal capillaries promotes transcapillary movement of negatively charged proteins, such as albumin, which is ultimately returned to the circulation via lymphatics (and capillaries) [42]. This would favor assimilation of absorbed substances, which bind to albumin (e.g., fatty acids). By contrast, the anionic nature of the glomerular capillaries would prevent the urinary loss of negatively charged proteins [224].
6.3. FACTORS INFLUENCING VASCULAR PERMEABILITY
Several experimental interventions to mimic physiological and pathological situations have been shown to alter the permeability of small intestinal capillaries to endogenous plasma proteins (Table 6.3) [42,197].
TABLE 6.3
Effects of physiologic and pathologic interventions on the osmotic reflection coefficient (σd) for total plasma proteins in the intestinal microcirculation.
Simulation of the postprandial state by intraluminal placement of micellar fatty acids or glucose and electrolytes is associated with an increase in transcapillary protein movement as reflected by an increase in lymphatic protein flux [225,226]. This increased protein flux may be the result of an increase in capillary surface area and/or an increase in capillary permeability to proteins. As shown in Table 6.3, the σd for total protein is not affected by glucose or electrolyte absorption, indicating that an increase in capillary surface area, rather than an increase in capillary permeability, is responsible for the increased lymphatic protein flux. This contention is supported by the observations that Kf,c [227] and the PS for rubidium [71] is increased during glucose/electrolyte absorption. By contrast, absorption of micellar oleic acid is associated with a decrease in σd (Figure 6.3) [225]. The σd for proteins of different molecular sizes has been used to calculate equivalent pore sizes in the intestinal capillaries during lipid absorption (Figure 6.4). There was no change in the size of the small pores during lipid absorption (46 vs 51 Å). However, the dimensions of the large pores increased from 200 to 300 Å. Both cholecystokinin [156,228] and neurotensin [179,229] have been shown to be released by the upper small intestine after intraluminal administration of lipids. Intravascular administration of neurotensin to achieve circulating postprandial concentrations decreased, σd [230] while cholecystokinin did not [225]. The neurotensin-induced increase in permeability was a result of a selective increase in the dimensions of the large pores (from 47 to 330 Å). These findings suggest that neurotensin may mediate the vascular permeability changes produced by lipid absorption. However, conclusive evidence (e.g., pharmacologic or genetic blockade) to support this contention is lacking.
Several pathologic interventions have been associated with changes in σd for total plasma proteins. I/R of the small intestine is associated with a decrease in σd for total proteins [210]. Mast cell degranulation decreases σd, an effect attributed to the action of histamine on H2 receptors [231]. Experimental arterial hypertension also decreases σd [232]. A role for angiotensin II in this phenomenon is not likely, since angiotensin II actually increases, rather than decreases σd [30]. The mechanism by which angiotensin II decreases intestinal vascular permeability to plasma proteins has not been addressed, but angiotensin II decreases the movement of albumin across brain endothelial cell monolayers by modifying the functional activity of occludin, a tight junction protein [233]. Since angiotensin II may be one of the few endogenous vasoactive agents that actually decrease vascular permeability, further studies are warranted to more clearly define the mechanism by which angiotensin II strengthens endothelial barrier function.
6.4. ULTRASTRUCTURAL CORRELATES FOR THE FUNCTIONAL PATHWAYS
The specific capillary ultrastructural pathways used by solutes and water are a matter of debate [196,202,204,207,234–236]. However, it is generally agreed that two different pathways must exist to explain physiological data: a large-pore pathway (175–250 Å radius) and a small-pore pathway (45–55 Å radius) [202,204,210].
The structural equivalents to the large-pore pathway are either the open fenestrae (200–300 Å radius) or the plasmalemmal vesicles/caveolae (250 Å radius) or a combination of both [199,202]. Obviously, the plasmalemmal vesicles assume a dominant role as the large-pore pathway in continuous capillaries (lacking fenestrae). Although the transcellular movement of the plasmalemmal vesicles is too slow to account for transendothelial movement of macromolecules [196,199], the transendothelial channels formed from multiple vesicles/caveolae would allow for more rapid transendothelial protein flux.
An interesting possibility that the large-pore system may be a dynamic ultrastructural entity encompassing both fenestrae and plasmalemmal vesicles has been proposed [202]. In this scheme (Figure 6.5), a plasmalemmal vesicle/caveolae (or vesicles) fuses with the endothelial cell membrane at both the blood and tissue fronts forming a transendothelial channel with two diaphragms, which eventually becomes a fenestra by a collapse to minimal length. This possibility is intriguing in that it may explain why the macromolecular permeability characteristics of the fenestrated capillaries of the gastrointestinal tract and continuous (containing only vesicles) capillaries of the hind-paw are similar (Tables 6.1 and 6.2). Structures identical to those schematically depicted in Figure 6.5C have been identified by electron microscopy and referred to as “barrel-shaped fenestrae [206].” Further, biochemical support for this possibility stems from the observation that a caveolar protein, PV-1, is also found on diaphragms of fenestrae and transendothelial channels [237]. Since current ultrastructural studies involve static approaches (a snap-shot in time) rather than continuous monitoring of structural changes, support for this contention awaits further refinement in the available technology. In vitro approaches may also be used to address this issue, particularly, since fenestrae have been induced in endothelial cells in culture [238,239].
FIGURE 6.5
Proposed development of fenestrae from plasmalemmal vesicles. (A) plasmalemmal vesicle; (B) fusion of vesicle with both aspects of endothelial cell membrane; (C) formation of transendothelial channel with two diaphragms; (D and E) development of typical (more...)
The structural equivalents to the small-pore pathway in gastrointestinal capillaries are generally believed to be the diaphragmed fenestrae [42] and, perhaps, the endothelial cell junctions [205]. The junctional pathway may be more important in the continuous capillaries, such as those of the muscularis. However, the relative number of small pores predicted from functional studies (small / large pore numbers / areas >500:1; Table 6.2) do not coincide with the ratio of diaphragmed / fenestrae noted in ultrastructural studies (~50:50). The presence of a glycocalyx-like structure covering the fenestral diaphragms has been proposed to provide the necessary sieving characteristics of small pores predicted by functional studies [114,206,208].
With respect to the charge selectivity of intestinal capillaries, physiological assessments indicate that the microcirculation behaves as a positively charged barrier, thereby facilitating the transcapillary movement of negatively charged macromolecules (albumin) [222]. The presence of a glycocalyx-like structure covering the endothelial lining not only may contribute to the sieving characteristics of capillaries, but may also be an important contributor to the charge-selectivity of capillaries [114,206,208]. However, the polyanionic nature of its constituents (heparin sulfate, hyaluronic acid, etc.) imparts a negative charge to the glycocalyx. Thus, it is difficult to reconcile the positively charged barrier noted in physiological studies with the macromolecular composition of the glycocalyx. One possibility is that the composition of the glycocalyx of intestinal capillaries differs from that of glomerular capillaries. This possibility is tenable, since endothelial cells can readily alter the content and physiochemical properties of the glycocalyx [114]. However, experimental verification of this possibility is lacking.
As mentioned above, increases in microvascular permeability usually involve increases in the size of the large-pore system or the formation of capillary “gaps.” Although it is generally agreed that the gaps form at the venular end of the microcirculation, the ultrastructural equivalent of these gaps is under considerable debate [240]. The formation of gaps has been attributed to contraction of endothelial cells leading to their separation [241,242]. Alternatively, it has been proposed that the plasmalemmal vesicles/caveolae (and transendothelial channels formed by their fusion) are opened wider by the endothelial cell contractile machinery [203]. Interestingly, a similar debate is ongoing regarding the ultrastructural pathway (paracellular vs vesicular) used by neutrophils to cross the venular capillaries during inflammation [243].
- Transcapillary Solute Exchange - The Gastrointestinal CirculationTranscapillary Solute Exchange - The Gastrointestinal Circulation
Your browsing activity is empty.
Activity recording is turned off.
See more...