Regional differences in rat large intestinal crypt function in relation to dehydrating capacity in vivo

Abstract

1. Rat descending colon absorbed fluid against a large hydraulic resistance, imposed by 10% agarose (w/v) gel plugs inserted in the lumen, by raising the tonicity of the absorbate from the gel to 880 ± 54 mosmol kg⁻¹; the tonicity of the absorbate from 2.5% gels was 352 ± 38 mosmol kg⁻¹. The hypertonic absorbate generated an osmotic pressure which created a fluid tension in the crypt lumen. This was monitored as a suction tension in colonic luminal gels of 45.3 ± 3 cmH₂O with 2.5% gels and 725 ± 145 cmH₂O with 10% gels. The caecum was unable to absorb fluid against a significant hydraulic resistance.

2. Fluorescein isothiocyanate-labelled dextran (FITC dextran; molecular mass 10000 Da) accumulated within descending colonic crypt lumens by concentration polarization. Maximal accumulation at a depth of 20–40 μm below the mucosal surface was 5.68 ± 0.2-fold above control levels. Caecal crypts accumulated dextran to a maximum of 1.8 ± 0.17-fold above control levels.

3. The relationship between crypt luminal tension and suction tension of the distal colon was also demonstrated using paraffin, which occluded the crypt lumens with microscopic droplets and completely inhibited fluid absorption from high resistance luminal gels.

4. Reduction in dietary Na⁺ intake raised plasma aldosterone and the capacity of the distal colon to dehydrate against a high luminal hydraulic resistance. The caecum did not respond in this way to varied Na⁺ intake.

The descending colon can exert a very large suction tension (4–5 atm (400–500 kPa)), which dehydrates faeces to give a water content of less than 70 % of the total faecal weight (Van Weerden, 1961; McKie et al. 1990, 1991; Pedley & Naftalin, 1993). The suction tension is produced by hypertonic absorption of NaCl by crypts. The osmotic pressure gradient across the crypt wall produced by the accumulated hypertonic NaCl in the pericryptal fluid induces net fluid outflow across the crypt wall. The fluid tension developed in the crypt lumen sucks fluid out of the adhering faeces, leading to consolidation (Zammit et al. 1994).

Fluid absorption into the crypt lumens has been demonstrated using confocal microscopy (Pedley & Naftalin, 1993; Naftalin & Pedley, 1995) and microperfusion studies in colonic crypts (Singh et al. 1995). Fluorescein isothiocyanate-labelled dextran (FITC dextran; molecular mass 10000 Da) accumulates inside the crypt lumen as a result of concentration polarization, i.e. the dye accumulates in the crypts of rat descending colon by being dragged into the lumens via the absorptive convective stream. The dye is freely diffusible between the crypt lumen and the bathing solution in the colonic lumen. The concentration polarization of FITC dextran in the crypt lumen depends on the relatively low diffusion coefficient of the high molecular weight dextran and the relatively high rate of flow into the crypt as a result of the osmotically induced outflow across the crypt wall.

The distal colon of rat, rabbit and humans responds to aldosterone by generating Na⁺ conductance channels which are inhibited by amiloride (Edmonds & Marriott, 1967; Clauss, 1985; Sellin & Desoignie, 1988). The proximal colon differs from the distal colon in that it does not have amiloride- or benzamil-sensitive Na⁺ conductance channels (Sellin et al. 1988). Instead, electroneutral NaCl absorption is mainly accomplished by dual Na⁺-H⁺ and Cl⁻-bicarbonate exchange (Sellin & Desoigne, 1984; Clauss, 1985; Fromm & Hegel, 1987; Turnamian & Binder, 1990).

It has been demonstrated previously (McKie et al. 1990; Bleakman & Naftalin, 1990) that an alternative explanation for generating the force required to dehydrate faeces, the tension developed by the smooth muscle in the wall of the colon, is insufficient by three orders of magnitude to explain the speed of dehydration from consolidated faeces.

Faecal consolidation is localized only in the distal colon (Hecker & Grovum, 1975; Snipes et al. 1982; Bleakman & Naftalin, 1990; Zammit et al. 1994). However, fluid absorption rates in fluid-perfused proximal and distal colon are approximately the same (Edmonds & Marriott, 1967; Sellin & Desoigne, 1984; Clauss, 1985). Additionally, the electrical conductivity in the proximal colon of ca 10–20 mS cm⁻² decreases only to ca 4–10 mS cm⁻² in the distal colon (Sandle & McGlone, 1987). Furthermore, there are no large obvious differences in histological appearance between the proximal and distal colonic crypts, in the number per unit area, or the size, or the cellularity (Snipes et al. 1982).

It is possible that differences in absorptive function of the proximal and distal regions of the colon are only manifest when the colon is subjected to raised luminal hydraulic resistance, as occurs during the normal course of faecal consolidation. Horster & Luckhoff (1981) found that rat distal colon has a much lower hydraulic conductivity (L_p) when exposed to a large osmotic pressure gradient (1000 mosmol kg⁻¹) than when exposed to isotonic conditions. Similarly, Zammit et al. (1994) observed that the L_p of the rat distal colon decreased as the hydraulic resistance of the luminal fluid was raised.

Low Na⁺ diets are known to raise, and high Na⁺ diets to lower, both aldosterone levels and net Na⁺ absorption from distal colon (Edmonds & Marriott, 1967; Sellin & Desoignie, 1987). In this paper we extend earlier investigations to determine whether a low Na⁺ diet increases the capacity of distal colon to absorb fluid against a hydraulic gradient.

METHODS

Confocal microscopy

Rats (Wistar) weighing 150–200 g were killed by cervical dislocation and the colonic mucosa was stripped of its muscle layer and mounted as a 10 mm² sheet in a temperature-controlled perifusion chamber at 35°C. The perifusion chamber was perfused with modified Tyrode solution (mM: NaCl, 136.9; KCl, 4.0; CaCl₂, 1.8; NaHCO₃, 11.8; NaH₂PO₄, 0.9; sodium formate, 4.3; glucose, 5.6). The total measured osmolality of the Tyrode solution was 289 mosmol kg⁻¹. The solution prior to perifusion was gassed with 95 % O₂-5 % CO₂ to maintain a pH of 7.3–7.4.

The tissue was viewed using a Nikon Diaphot inverted microscope with a Nikon Fluor × 20 lens with a numerical aperture of 0.75. The objective lens was attached to an MRC 600 confocal scanning head, equipped with two detection channels and an Ar-Kr mixed gas laser. This allowed simultaneous detection of fluorescein excitation, using the 488 nm line, and the bright field image, using the transmitted light detector in the second channel. Movement along the z-axis was controlled by a software-controlled stepper motor attached to the focusing control of the microscope, which allowed the plane of focus to be changed in 1 μm steps. Fluorescence images can be resolved to a depth of ca 120 μm below the tissue surface. Perifusion media were continuously passed over the tissues via a pre-warming loop and a pair of back-to-back solenoid valves (Lee Products Ltd, Gerrards Cross, UK). This enabled the solutions to be maintained at 35 ± 0.1°C and changed in less than 1 s. Medium was aspirated at the opposite side of the chamber using a suction micropipette, resulting in a near-laminar flow of medium.

Estimation of the average concentration of FITC dextran in crypt lumens

The crypt luminal and pericryptal concentration of fluorescein isothiocyanate-labelled dextran (FITC dextran; molecular mass 10000 Da; Sigma) was estimated here by monitoring the ratio of fluorescence intensity (I) of the dye at any depth x to that in the crypt luminal opening (x= 0). The relative changes in luminal concentrations were quantified using fluorescence after attenuation of the signal with a neutral density filter (3 % transmittance) and subtraction of background. With this high level of neutral density filtration there was no detectable autofluorescence from the tissue. The ratio (I_x - I_bkg)/(I_co- I_bkg) in the crypt lumen is equivalent to the concentration polarization at depths x. I_x is the fluorescence intensity in the crypt lumen at distance x from the crypt luminal opening, I_co is the fluorescence intensity at the crypt luminal opening, and I_bkg is the fluorescence in the absence of dye (I_bkg= 30–50 units per pixel). Estimates of average fluorescence intensity in the crypt lumen were obtained by taking the pixel intensity of a circumscribed area of interest within six to eight crypt lumens, 500–1000 pixels per area at a time. Maximum pixel light intensity = (256 units - background). The intensities of the same regions in the crypt lumen at successive depths were computed.

Ratios of dextran Cy5 and fluorescein acetate within the crypt lumens of caecum and descending colon

A problem with fluorescence microscopy is that the fluorescence signal is progressively quenched on descent through the tissue. The extent of this attenuation varies with the tissue density. Consequently it is difficult to estimate accurately whether the observed decrease in dye concentration with depth is real, or simply due to quenching. One way of resolving this problem is to employ two optically separable fluorescence dyes simultaneously, one of which is both more permeant through the crypt wall and has a higher diffusion coefficient than the other. Dextran accumulation along the axis of the crypt lumen was estimated by this method of ratio imaging.

Dextran Cy5 was prepared from dextran biotin (molecular mass 10000 Da; Molecular Probes) and streptavidin Cy5 (Pierce Chemicals). Equimolar amounts of the two substances (500 μm) were mixed in a reaction tube at 37°C for 30 min and then diluted to 20 μm in the tissue superfusion solution with 20 μm fluorescein acetate (Sigma). The fluorescence of fluorescein and of Cy5 is detected without significant mutual interference using the 488 and 647 lines of the Ar-Kr laser. Since the dextran Cy5 has a molecular mass in excess of 20 times that of fluorescein acetate it will be ∼5-fold more liable to accumulate within the crypt lumens by concentration polarization than fluorescein acetate. Concentration polarization is an exponential function of the Peclet number (=vx/D; where v is the convective velocity, x the distance moved along the convective stream and D the diffusion coefficient, which is inversely related to √(molecular mass)). Thus, by observing the change in the ratio of dextran Cy5 to fluorescein acetate along the length of the crypts it is possible to obtain an estimate of the relative extents to which concentration polarization affects the accumulation of probe within the crypt. This method, although cumbersome, has the advantage over the method outlined above of being relatively independent of the quenching of the fluorescence signal with depth within the tissue and is not subject to pH-dependent alterations in fluorescence.

Low and high Na⁺ diets

To vary the blood aldosterone levels, rats were fed on low and high Na⁺ diets, which are known to increase and decrease, respectively, circulating aldosterone levels. The rats fed on a low Na⁺ diet received plain flour with 5 % wheat bran and water ad libitum for 10 days. Rats fed a high Na⁺ diet received the flour and bran diet, as above, but isotonic saline was substituted for drinking water. The rats maintained a normal weight gain during this period (Fattah et al. 1977). The effects of the diets on circulating aldosterone were tested to determine whether they conformed to previously demonstrated effects. Blood samples were obtained from portal venous blood immediately after cervical dislocation.

The normal level of circulating aldosterone in Wistar rats is 1500 ± 250 pM. The effect of the high Na⁺ diet was to reduce the aldosterone level to 805 ± 73 pM (n= 3) while the low Na⁺ diet raised the plasma aldosterone level to 3065 ± 573 pM (n= 3), which was significantly different from the level seen with the high Na⁺ diet (P < 0.01, Student's t test). These concentrations are within the range of those previously observed with these dietary regimens (Abayasekara et al. 1993).

Monitoring the changes in fluid and ionic composition of the agarose gels

Wistar rats weighing between 150 and 200 g were anaesthetized by intraperitoneal injection of pentobarbitone sodium BP (120 mg kg⁻¹; May and Baker Ltd, Dagenham, UK). The descending colon was opened by an incision at the level of the pelvic girdle and faeces were flushed from the lumen with Tyrode solution. The gels were inserted into the lumen proximal to the incision. A cotton ligature was placed proximally and distally to the gels to secure them in the loop of colon.

Agarose type II-A (Sigma) was used because of its high strength and biological inertness. The agarose content of the gels in this series was varied from 2.5 g per 100 ml of vehicle solution (140 mM NaCl, 5 mM KCl) to 12.5 g per 100 ml. The agarose solutions were heated under pressure for about 10 min at 110°C and then poured into glass tubes of uniform internal diameter (6 mm).

The solidified cooled gels were expressed from their moulds, cut into 2 cm lengths and weighed prior to insertion into the colonic lumens of anaesthetized rats (Zammit et al. 1994). After incubation, the gels were removed from the lumen and reweighed to determine fluid loss. The macerated gels were extracted overnight in 0.1 M HCl. The amounts of Na⁺ and K⁺ in the HCl extracts were measured using an Instrumentation Laboratory 943 automated flame photometer. The rates of fluid absorption, Na⁺ loss and K⁺ loss or gain were expressed as microlitres (fluid) or micromoles (ions) per square centimetre of surface area of the gel per hour of exposure (μl cm⁻² h⁻¹ or μmol cm⁻² h⁻¹).

Addition of paraffin to gels

Prior to insertion of the gels, the lumen of the descending colon was flushed with Tyrode solution. Paraffin oil (0.5 ml, specific gravity 0.830–0.860; Fison's Scientific Apparatus, Loughborough, UK) was then injected into colonic lumen and spread evenly along the segment of colon in which its effect was to be monitored. Either 2.5 or 10 % agarose gel cylinders were then inserted into the lumen. The colonic segment containing the paraffin was sealed with cotton sutures and as controls identical gels without oil, but with saline injections instead, were placed in adjacent colonic segments.

Estimation of the suction tension on the agarose gels and power output of the colon

In the agarose concentration range 2.5–15 %, the L_p of the gels decreases as a monoexponential function of agarose concentration, according to the function y= 1.364× 10⁻⁶exp(-0.547x), where x is the percentage of agarose in the gel. Hence, when the agarose concentration was 2.5 %L_p= 1.12 × 10⁻⁶ cm s⁻¹ cmH₂O⁻¹ and when the concentration was 10 %L_p= 5.75 × 10⁻⁹ cm s⁻¹ cmH₂O⁻¹ (Zammit et al. 1994). The suction tension, ΔP (cmH₂O), across a hydraulic conductance can be characterized by the equation:

where J_v is fluid flux (cm³ cm⁻² s⁻¹). Thus, nearly 200 times more force is required to dehydrate 10 % gels than 2.5 % gels.

Statistics

Estimates of statistical significance were obtained using either Student's t tests for paired data, or one-way analysis of variance for multiple testing (ANOVA) in the case of the effects of high and low Na⁺ diets.

RESULTS

Comparisons of fluid absorption from rat descending colon and caecum

In Fig. 1 the effects of changing gel agarose concentrations on fluid uptake from the descending colon and caecum are shown. Fluid absorption from rat caecum in vivo was not observed from gels with concentrations higher than 2.5 %; whereas the gels inserted into the descending colons of the same animals were dehydrated until the agarose concentration was at least 12.5 % (Zammit et al. 1994). This means that the suction tension exerted by descending colon is at least 100 times greater than that exerted by the caecum. This result confirms the speculation of Horster & Luckhoff (1981) that the passive hydraulic permeability of the caecum is greater than that of the descending colon.

Figure 1. The effects on fluid uptake of changing gel agarose concentrations in the caecal (□) and descending colonic (▪) lumens.

No significant fluid absorption is observed from caeca with gel agarose concentrations >2.5 %. The lines are drawn by a smooth curve fit. The errors shown are ±s.e.m. (n= 5–8).

Effect of high and low Na⁺ diets on fluid and electrolyte absorption from descending colon and caecum

The effects of high, control and low Na⁺ diets (n= 5 animals each) are seen in Figs 2–4. A low Na⁺ diet increased the steady-state concentration of plasma aldosterone, whereas a high Na⁺ diet reduced aldosterone (Abayasekara et al. 1993; see Methods).

Figure 2. Effects of high, control and low Na⁺ diets on fluid absorption from intraluminal 2.5 and 10 % agarose gels by rat descending colon.

As Na⁺ was excluded from the diet, fluid absorption increased from both 2.5 % gels (▪,**P < 0.001, ANOVA) and 10 % gels , *P < 0.02, ANOVA) (n= 5 for each condition, 1 gel of each strength per animal). No significant fluid uptake was observed from 10 % gels in rats fed a high Na⁺ diet (P > 0.1). Fluid absorption from 2.5 % was significantly greater than from 10 % gels under all conditions (†P < 0.001, ANOVA).

Figure 4. Effects of high, control and low Na⁺ diets on the tonicity of absorbate from intraluminal 2.5 and 10 % agarose gels by rat descending colon.

The effect of varying Na⁺ intake did not affect the tonicity of the absorbate significantly from either 2.5 or 10 % gels. Tonicity was estimated from J_Na/J_water, where J_Na is Na⁺ flux (μmol cm⁻² h⁻¹) and J_water is water flux (μl cm⁻² h⁻¹). The mean tonicity of the absorbate was 352 ± 38 mosmol kg⁻¹ from 2.5 % gels and 880 ± 54 mosmol kg⁻¹ from 10 % gels (P < 0.001, ANOVA).

As dietary Na⁺ intake is reduced, so the capacity to absorb fluid against a high hydraulic resistance is increased (P < 0.02). Fluid absorption from 10 % agarose gels by descending colon of rats fed a low Na⁺ diet was significantly above zero, whereas that of rats on a high Na⁺ diet was not (P > 0.1). As shown previously (Zammit et al. 1994), fluid uptake into 10 % gels was significantly lower than into 2.5 % agarose gels (P < 0.001) (Fig. 2).

There was also a significant increase in fluid uptake against low-hydraulic-resistance 2.5 % agarose gels as dietary Na⁺ was reduced (P < 0.002) (Fig. 2).

The effect of reducing Na⁺ intake in the diet significantly raised the capacity of the colon to absorb Na⁺ against both 2.5 % and 10 % agarose gels (P < 0.01 and n= 5 for each condition). A low Na⁺ diet increased the capacity of the colon to absorb Na⁺ from 10 % gels above that of a high Na⁺ diet from both 2.5 % (P < 0.02) and 10 % gels (P < 0.05). In contrast, a high Na⁺ diet reduced the capacity of the colon to absorb Na⁺ from 2.5 % gels below that of controls (P < 0.02). Raising the gel concentration from 2.5 to 10 % also reduced the rate of Na⁺ uptake (P < 0.01) (Fig. 3).

Figure 3. Effects of high, control and low Na⁺ diets on Na⁺ absorption from intraluminal 2.5 and 10 % agarose gels by rat descending colon.

Reducing Na⁺ intake in the diet significantly raised the capacity of the colon to absorb Na⁺ against both 2.5 % (▪) and 10 % ) agarose gels († P < 0.01 and n= 5 for each condition, 1 gel of each type per rat). A low Na⁺ diet significantly increased the capacity of the colon to absorb Na⁺ from 10 % gels above that of a high Na⁺ diet from 2.5 % (**P < 0.02) and from 10 % (*P < 0.05) gels. A high Na⁺ diet reduced Na⁺ absorption from 2.5 % gels below controls (**P < 0.02).

The effect of varying Na⁺ intake did not significantly affect the tonicity of the absorbate from either 2.5 or 10 % gels. The mean tonicity of the absorbate was 352 ± 38 mosmol kg⁻¹ from 2.5 % gels and 880 ± 54 mosmol kg⁻¹ from 10 % gels (P < 0.001) (Fig. 4). These findings confirm those of Zammit et al. (1994) that the capacity to absorb fluid against a high hydraulic resistance is related to the capacity to generate a hypertonic colonic absorbate. No significant effect of a low Na⁺ diet was observed on fluid absorption by caecum exposed to 2.5 % gels and no significant fluid absorption was observed with either low or high Na⁺ diets in caeca with 10 % gels (not shown).

Effect of paraffin on fluid and electrolyte absorption from agarose gels by rat colon

Since colonic crypts are blind-ended tubules and have a diameter of 1–15 μm, it seems likely that, like renal tubules, they can be occluded by mineral oil droplets, which prevent fluid transit along the lumen (Hellman et al. 1967). By blocking water uptake into the crypt lumen, oil droplets would impair the capacity of the colon to absorb fluid via the lumen, but would not impair absorption via the surface mucosa. Hence paraffin should impair the capacity of the colon to generate hard faeces, i.e. absorb fluid against a large hydraulic gradient, although it should not prevent fluid absorption from the liquid or semi-liquid state (Zammit et al. 1994).

In view of previous findings that crypt fluid absorption has an important role in production of hard faeces, we decided to test whether the aperient laxative action of paraffin oil can be ascribed to blockage of water flow into the crypt lumen.

With the 2.5 % gels in rat distal colon, fluid absorption decreased by 53 % with paraffin (P < 0.001, n= 11 pairs of gels) and with 10 % gels fluid absorption decreased by 93 % (P < 0.001, n= 13 pairs) (Fig. 5). The suction tensions exerted by the distal colon on the 2.5 and 10 % control gels were estimated to be 43 ± 3 and 725 ± 145 cmH₂O, respectively; this also fell, by 53 and 93 %, respectively, with paraffin (P < 0.001) (Fig. 6).

Figure 5. Effect of paraffin on fluid absorption from agarose gels by rat descending colon in vivo.

Fluid absorption from 2.5 % gels is decreased with paraffin (n= 11 pairs; P < 0.001) and is decreased significantly by 93 % from 10 % gels paraffin (n= 13 pairs, P < 0.001). ▪, control; , paraffin.

Figure 6. Effect of paraffin on suction tension exerted by rat descending colon on luminal contents.

The suction tension exerted on control 10 % gels was significantly higher than that exerted on 2.5 % gels (P < 0.001). Paraffin significantly reduced the suction tension exerted by the distal colon on the 2.5 % and 10 % gels by 53 and 93 %, respectively (P < 0.001). ▪, control; , paraffin.

Paraffin reduced net Na⁺ absorption from 2.5 % gels from 8.74 ± 0.75 to 6.75 ± 0.53 μmol cm⁻² h⁻¹ (P < 0.025, n= 11) and from 10 % gels from 3.99 ± 0.52 to 3.59 ± 0.57 μmol cm⁻² h⁻¹ (not significant, n= 13). Thus Na⁺ absorption was reduced by paraffin from 2.5 % gels by 23 ± 6 %. This paraffin-dependent reduction in Na⁺ absorption was significantly less than the 53 ± 2 % reduction in water (t test of difference, P < 0.025).

Thus whilst paraffin was ineffective in reducing Na⁺ absorption from the 10 % gels, it reduced water absorption from 10 % gels to zero (t test of difference between Na⁺ and water absorption from 10 % gels; P < 0.001).

Direct observation of the effect of paraffin oil on FITC dextran accumulation into rat descending colonic crypts

Agarose gels were inserted into the rat descending colon with either 20 μm FITC dextran or 20 μm FITC dextran and paraffin (0.5 ml per 2 cm length of sac). Immediately following incubation, the sacs were removed and the colonic mucosae were stripped of external muscle layers in Tyrode solution containing 20 μm FITC dextran without paraffin and mounted in the perifusion chamber for viewing by confocal microscopy at 35.0 ± 0.1°C. Figure 7 shows the confocal fluorescence views. The plane of focus of the confocal images is between 30 and 40 μm below the mucosal surface. All of the crypt lumens in tissue exposed to paraffin oil are occupied with refractile oil droplets, whereas no intraluminal droplets are evident in controls. These inclusions within the crypt lumen are most evident when shown up by negative contrast in the fluorescence images. FITC dextran is not miscible with the oil so the paraffin droplet within the crypt lumen appears as an opaque black inclusion. The mean diameter of the droplets was 6 ± 2 μm. Surrounding each droplet is a thin layer of fluorescent dextran which merges with the crypt wall. The fluorescence intensity of FITC dextran within the crypts of tissue exposed to paraffin was only 52 ± 3 % (n= 14) of that found in controls. FITC dextran accumulation at a depth of 40 μm was 5.05 ± 0.02-fold above background levels within the lumens of control crypts and 2.62 ± 0.03-fold above background in paraffin-treated tissue (significant difference between control and paraffin, P < 0.001, n= 11).

Figure 7. Confocal microscopic view of rat distal colonic crypts filled with FITC dextran showing effects of paraffin droplets in the lumen.

The confocal image is between 30 and 40 μm below the mucosal surface. All of the crypt lumens in tissue exposed to paraffin oil are occupied with refractile oil droplets (B), whereas no intraluminal droplets are evident in controls (A). Width of both panels, 210 μm.

Comparison of FITC dextran accumulation in the descending colonic and caecal crypt lumens

Uptake of FITC dextran was compared in descending colon and caecal crypts (Fig. 8) using the ratios of fluorescence images of Cy5 dextran and fluorescein acetate and normalized to the fluorescein acetate fluorescence at the crypt luminal opening (Fig. 9). In descending colonic crypts Cy5 dextran accumulated to a maximum at a depth of 20–40 μm. This was 5.7 ± 0.2-fold above the fluorescence intensity observed at the luminal opening, i.e. z= 0μm. Cy5 dextran also accumulated to a higher concentration in the pericryptal space than at the crypt opening. As found previously with FITC dextran, the rate of accumulation of Cy5 dextran in the pericryptal space was slower than in the crypt lumen (Zammit et al. 1994). However, fluorescein acetate did not accumulate to a higher concentration than at the crypt opening either in the depths of the crypt lumen or in the pericryptal space. These findings are consistent with the view that Cy5 dextran accumulates in the crypt lumen because it has a much lower diffusion coefficient than fluorescein acetate and because it permeates relatively slowly through the crypt wall. Conversely the absence of accumulation of fluorescein acetate is consistent with its high diffusion coefficient and high permeability (see Discussion sections in this and the following paper (Naftalin & Pedley, 1999)).

Figure 8. Confocal images at varying depths within rat caecal and descending colonic mucosa showing steady-state accumulation of FITC dextran.

The upper four panels show a serial z-scan descending in 10 μm steps from the caecal crypt luminal openings through the tissue. The lumens were filled with FITC dextran from the mucosal surface. Note that there was rapid attenuation of the fluorescence signal with depth and little fluorescence outside the crypt lumens. Width of panels, 500 μm each. The lower six panels show a similar z-scan through descending colonic mucosa. Note that there is an increase in fluorescence intensity on descent from the luminal opening to about 40 μm and that much fluorescence is evident in the pericryptal zones in the upper 40–60 μm. Width of panels, 250 μm each.

Figure 9. Comparison of simultaneous uptake of Cy5 dextran and fluorescein acetate in rat caecal and descending colonic crypts.

A, uptake of Cy5 dextran into the crypt lumen (^) and pericryptal spaces (•) at varying depths along the caecal crypts. Uptake of fluorescein acetate is also shown in the lumen (□) and in the pericryptal space (▪). B, uptake of Cy5 dextran into the crypt lumen (^) and pericryptal spaces (•) at varying depths along the descending colonic crypts. Uptake of fluorescein acetate is shown in the lumen (□) and in the pericryptal space (▪). In descending colonic crypts Cy5 dextran is accumulated to a maximum at a depth of 20–40 μm. This is 5.7 ± 0.2-fold above the fluorescence intensity observed in the pericryptal space at the depth of the crypt opening, i.e. z= 0μm.

The caecal crypts showed only a slight accumulation of Cy5 dextran. The maximal luminal concentration rose only to 1.80 ± 0.17-fold above the concentration at the crypt opening. A much smaller accumulation of Cy5 dextran was observed relative to fluorescein acetate in the caecal crypt lumen (normalized Cy 5 dextran/fluorescein acetate ratio at 40 μm for descending colon is 11.4 ± 1.5, whereas in the caecum this ratio is only 2.27 ± 0.25; P < 0.001). These findings indicate that descending colon is able to concentrate dextran ca 22-fold more than the caecum.

The effect of low Na⁺ diet on dextran accumulation in rat descending colonic crypt lumens

FITC dextran was accumulated by descending colons from rats fed a low Na⁺ diet to 6.8 ± 0.3-fold above the concentration at the crypt luminal opening. The extent of dextran accumulation in the crypt lumens was not significantly higher than in controls (not shown).

DISCUSSION

Macroscopic differences in fluid and electrolyte absorption between rat caecum and descending colon

At a macroscopic level the results show that the descending colon is much better able to dehydrate against a raised luminal hydraulic resistance than the caecum (Figs 1–4). This localization of dehydration power to the distal colon shows that faecal consolidation occurs solely in the distal colon because of its capacity to exert a high suction force on its adjacent luminal content, and refutes the alternative possibility that faecal dehydration results from a progressive process along the entire length of the large intestine. Thus, as the proximal colon is unable to dehydrate against a significant hydraulic resistance, retrograde movement of faeces from distal to proximal colon will result in faecal rehydration. This could be of some practical importance in preventing stasis of digesta flow in the proximal colon.

These results demonstrate that there is a clearly differentiated functional difference between the proximal and distal colon.

Effects of low and high Na⁺ diets on colonic absorption

The effects of low and high Na⁺ diets on colonic absorptive function show that the capacity to absorb both fluid and Na⁺ against a high luminal hydraulic resistance in rat distal colons is increased (Figs 2–4) by raised aldosterone and also possibly angiotensin II (Sechi et al. 1993). In contrast no effect of Na⁺ restriction was observed on caecal absorption rates from 2.5 % gels.

These studies are consistent with the findings of Pacha & Pohlova (1995), who showed that a low Na⁺ diet both raises aldosterone and increases amiloride-sensitive Na⁺ absorption in rat colon. The confinement of the effects of a low Na⁺ diet to fluid absorption in the distal colon is also consistent with the known distribution of aldosterone-induced amiloride-sensitive Na⁺ conductance channels (Sandle & McGlone, 1987; Bridges et al. 1989).

Effects of paraffin on distal colonic fluid absorption

The experiments with paraffin were included to show the relationship between macroscopic fluid absorption in the distal colon and microscopic blockage of crypt lumens. The results show that paraffin selectively inhibits colonic fluid absorption when it is working against a large luminal hydraulic resistance. The reduction in fluid absorption from low resistance 2.5 % gels was 52 %, whereas the reduction of fluid absorption by paraffin from high resistance 10 % gels was 93 % (P < 0.01). These findings are reflected in the much reduced suction tension generated by the colon when treated with paraffin: 48 ± 48 vs. 725 ± 125 cmH₂O in control conditions (P < 0.001) (Figs 5 and 6). Na⁺ absorption is affected to a lesser extent than water absorption by raising the hydraulic resistance of the lumen, or by paraffin. These results suggest that the bulk of colonic Na⁺ absorption and isotonic fluid absorption from a fluid-containing lumen is via the colonic surface mucosa and, like the renal tubule, is unaffected by paraffin (Hellman et al. 1967), whereas hypertonic absorption via the crypts is completely inhibited by paraffin.

The absence of a significant effect of paraffin on Na⁺ absorption from high concentration gels suggests that the bulk of Na⁺ absorption is via the surface mucosa. Additionally, it indicates that the paraffin-dependent inhibition of water movement from 10 % gels cannot be ascribed to an unstirred layer at the surface mucosa, since this would affect water and Na⁺ to approximately the same extent.

Effects of paraffin on FITC dextran accumulation in descending colonic crypt lumens

The confocal microscopic evidence showing that paraffin droplets are in every crypt lumen and occlude the lumen to the extent that dextran accumulation is reduced by 52 ± 3 % indicates that there is a direct linkage between crypt luminal absorption and macroscopic absorption. The effects of paraffin on fluid absorption by descending colon only become obvious when the colon is required to dehydrate its luminal content against a large hydraulic resistance. Obstruction of the crypt lumen by paraffin droplets removes the capacity of the colon to absorb fluid from 10 % agarose gels and greatly reduces the capacity of mucosal crypts in vitro to concentrate FITC dextran (Fig. 9). Simulation shows that a 50 % reduction in concentration polarization of dye is obtained with a 70 % reduction in fluid inflow (Naftalin et al. 1995). Given that the rate of fluid entry into the crypt lumens in control tissue is approximately 1 × 10⁻³ cm s⁻¹, it follows that paraffin reduces the rate of water entry into the crypt lumen to approximately 3 × 10⁻⁴ cm s⁻¹.

Dextran accumulation

The accumulation of dextran by crypts from different regions or different animals can be assessed with some precision by using a double fluorescence labelling method to estimate dextran accumulation at differing depths in the lumens (Figs 8 and 9). This direct method rules out the possibility that the dextran accumulates within the crypt lumen as a result of binding, since if this were so then Cy5 dextran and fluorescein acetate should be accumulated equally. The method permits accurate assessment of the extent of dye accumulation at varying depths without the undue interference from tissue quenching.

The greater accumulation of the Cy5 dextran in the crypt lumens of descending colon than in the caecal crypts indicates that the ratio of fluid absorption (v) into crypt lumens to paracellular leakage across the crypt wall (P_Cy5), is much greater in descending colon than in caecum. This is corroborated by the lower crypt lumen/pericryptal space concentration ratio of both Cy5 and fluorescein acetate in the caecum than in descending colon.

Conclusions

The results presented in this paper show that there are large functional differences between caecal and descending colonic crypts. Only the descending colonic crypts can transport fluid against a significant hydraulic resistance and this capacity is enhanced by a low Na⁺ diet and reduced by a high Na⁺ diet. The structural and functional basis for these differences is presented in the following paper (Naftalin & Pedley, 1999).