Apical GLUT2 and Ca_v1.3: regulation of rat intestinal glucose and calcium absorption

Abstract

We have proposed a model of intestinal glucose absorption in which transport by SGLT1 induces rapid insertion and activation of GLUT2 in the apical membrane by a PKC βII-dependent mechanism. Since PKC βII requires Ca²⁺ and glucose is depolarizing, we have investigated whether glucose absorption is regulated by the entry of dietary Ca²⁺ through Ca_v1.3 in the apical membrane. When rat jejunum was perfused with 75 mm glucose, Ca²⁺-deplete conditions, or perfusion with the L-type antagonists nifedipine and verapamil strongly diminished the phloretin-sensitive apical GLUT2, but not the phloretin-insensitive SGLT1 component of glucose absorption. Western blotting showed that in each case there was a significant decrease in apical GLUT2 level, but no change in SGLT1 level. Inhibition of apical GLUT2 absorption coincided with inhibition of unidirectional ⁴⁵Ca²⁺ entry by nifedipine and verapamil. At 10 mm luminal Ca²⁺, ⁴⁵Ca²⁺ absorption in the presence of 75 mm glucose was 2- to 3-fold that in the presence of 75 mm mannitol. The glucose-induced component was SGLT1-dependent and nifedipine-sensitive. RT-PCR revealed the presence of Ca_vβ₃ in jejunal mucosa; Western blotting and immunocytochemistry localized Ca_vβ₃ to the apical membrane, together with Ca_v1.3. We conclude that in times of dietary sufficiency Ca_v1.3 may mediate a significant pathway of glucose-stimulated Ca²⁺ entry into the body and that luminal supply of Ca²⁺ is necessary for GLUT2-mediated glucose absorption. The integration of glucose and Ca²⁺ absorption represents a complex nutrient-sensing system, which allows both absorptive pathways to be regulated rapidly and precisely to match dietary intake.

We have proposed a new model for intestinal sugar absorption. When rat jejunum is challenged with glucose, the facilitative transporter, GLUT2, is rapidly inserted into the apical membrane (Kellett & Helliwell, 2000; Helliwell et al. 2000a; Affleck et al. 2003). In addition, the intrinsic activity of GLUT2 is rapidly up-regulated (Kellett & Helliwell, 2000; Helliwell et al. 2000b). Since GLUT2 is a high K_m, high capacity transporter compared with the Na⁺–glucose cotransporter, SGLT1, GLUT2 can provide a facilitated component several times greater than the active component at high glucose concentrations. Apical GLUT2 therefore provides a cooperative mechanism by which absorptive capacity is rapidly and precisely matched to dietary intake (Kellett, 2001; Kellett & Brot-Laroche, 2005). In this model, fructose absorption across the apical membrane is mediated by GLUT5 and GLUT2 (Cheeseman, 1993; Corpe et al. 1996; Helliwell et al. 2000a, 2000b). When intestine of wild-type, unanaesthetized mouse is challenged with fructose, glucose or sucrose by gastric intubation, large increases in GLUT2-mediated fructose absorption occur within minutes; the increase does not occur in GLUT2-null mice and the difference is attributable entirely to rapid, apical insertion of GLUT2 (Gouyon et al. 2003). Apical GLUT2 insertion is increased in response to enteroendocrine sensing (Au et al. 2002), energy sensing (Walker et al. 2004), experimental diabetes (Corpe et al. 1996; Marks et al. 2003), long-term dietary carbohydrate (Gouyon et al. 2003) and refeeding after phase 3 starvation (Habold et al. 2005). GLUT2 is present in the apical membrane of the midgut of larvae (Caccia et al. 2005) and increases in the apical membrane of rat after birth (Baba et al. 2005).

Two observations in particular point to a role for Ca²⁺ in apical GLUT2 regulation. First, regulation involves a PKC βII-dependent pathway, which is activated by glucose transport through SGLT1 and forms part of a sugar-sensing mechanism (Kellett & Helliwell, 2000; Helliwell et al. 2000a, Helliwell et al. 2003). PKC βII is a conventional PKC isoform dependent on Ca²⁺ for activity (Hug & Sarre, 1993). Second, Ca²⁺ is essential for the cytoskeletal rearrangement of the enterocyte accompanying glucose entry (Madara & Pappenheimer, 1987; Turner, 2000). It follows that there must be an apical mechanism for Ca²⁺ entry capable of operating under conditions of sustained depolarization.

However, these observations are not readily explained by the current view of transepithelial intestinal Ca²⁺ transport. Thus, active (transcellular) Ca²⁺ transport comprises three steps. In duodenum, absorption from the lumen across the apical membrane by epithelial Ca²⁺ channels TRPV5 (ECaC) and TRPV6 (CaT1) is strongly favoured by the electrochemical gradient (Ward & Boyd, 1980; Sharp & Debnam, 1994). Cytosolic diffusion of Ca²⁺ is enhanced by binding to calbindin-D_9K (Bronner et al. 1986; Feher et al. 1992). Finally, Ca²⁺ is transported actively across the basolateral membrane by the plasma membrane Ca²⁺-dependent ATPase (Bronner, 2003). TRPV5/6 are present predominantly in duodenum (Hoenderop et al. 2000; Zhuang et al. 2002), where there is little active absorption of glucose. Moreover, TRPV5/6 are activated by hyperpolarization, since they lack the S4 voltage sensor of L-type channels (Hoenderop et al. 1999, 2001; Peng et al. 1999). In contrast, apical GLUT2 insertion is concerned with Ca²⁺ absorption in jejunum under depolarizing conditions in the presence of high concentrations of nutrient at the apical membrane. Yet it is widely asserted that L-type channels are not present in intestine (Favus & Angeid-Backman, 1985; Fox & Green, 1986). Finally, we should note also that although active, saturable Ca²⁺ transport predominates in duodenum, the consensus is that at high Ca²⁺ concentrations the saturable component in the rest of the intestine is small compared with a non-saturable route of low permeability, which is attributed to paracellular flow (Pansu et al. 1983). When Ca²⁺ supply is plentiful, the active route apparently accounts only for ∼15% of overall absorption.

We have reported that the non-classical, neuroendocrine L-type channel, Ca_v1.3, is present in the apical membrane of rat intestine (Morgan et al. 2003). The level of Ca_v1.3 is negligible in duodenum, caecum and colon and maximal in distal jejunum and proximal ileum, that is, it is located in precisely the right region to play an important role in Ca²⁺ absorption during digestion and the associated generation of depolarizing nutrients. Ca_v1.3 is capable of operation under conditions of sustained, weak depolarization at low voltage thresholds and is therefore potentially ideal for intestine (Koschak et al. 2001; Lipscombe et al. 2004). Furthermore, in jejunum, we were able to monitor significant rates of Ca²⁺ absorption with L-type characteristics. Thus, unidirectional lumen-to-mucosa transport of 1.25 mm Ca²⁺ (the same as plasma free Ca²⁺) at 20 mm glucose was inhibited by nifedipine, Mg²⁺ and by repolarization of the membrane induced by blocking of glucose absorption with phloridzin; Ca²⁺ absorption was also activated by Bay K 8644. None of these four quite different conditions affect TRPV5/6. In the absence of evidence for any other known channel, we assume that Ca_v1.3 is functional and provides a significant route of transcellular absorption operating in times of dietary sufficiency.

We now report here that Ca²⁺ absorption across the apical membrane mediated by an L-type channel does indeed play an important role in the regulation of intestinal glucose absorption by controlling apical GLUT2 insertion. Furthermore, glucose stimulates Ca²⁺ absorption, most probably through Ca_v1.3.

Methods

Animals

All procedures used conformed to the UK Animals (Scientific Procedures) Act 1986. Male Wistar rats (240–270 g) were maintained on standard Bantin and Kingman (Hull, UK) rat and mouse diet ad libitum with free access to water.

Perfusion of jejunal loops

Rats were anaesthetized by an intraperitoneal injection of a mixture of 1.0 ml Hypnorm (Janssen Animal Health, High Wycombe, Bucks, UK) and 0.5 ml Hypnovel (Roche Diagnostics, Welwyn Garden City, Herts, UK) per kilogram body weight. Tail pinch, foot pinch and corneal reflexes were carefully monitored throughout the duration of the perfusion. Additional anaesthetic was administered by intramuscular injection of a mixture of 0.4 ml Hypnorm and 0.2 ml Hypnovel per kilogram body weight when required. Rats were humanely killed by exsanguination under anaesthetic at the conclusion of the experiment. A mid-to-distal loop of jejunum was cannulated at 10 and 35 cm from the Ligament of Treitz and perfused in vivo in single-pass mode with perfusate comprising nutrient at the stated concentration in modified Krebs–Henseleit buffer (KHB) containing 120 mm NaCl, 4.5 mm KCl, 1.0 mm MgSO₄, 1.8 mm Na₂HPO₄, 0.2 mm NaH₂PO₄, 1.25 mm CaCl₂, 25 mm NaHCO₃ gassed (19: 1, O₂: CO₂) to pH 7.4 before use (Kellett & Helliwell, 2000). ⁴⁵Ca²⁺ (0.35 kBq ml⁻¹) and ³H-inulin (0.70 kBq ml⁻¹) were also added for the determination of Ca²⁺ and water fluxes: the counting protocol automatically corrected for channel crossover. The flow rate of perfusate was 0.37 ml min⁻¹ and that of gas 0.19 ml min⁻¹. The perfusate flow rate of 0.37 ml min⁻¹ was determined by the need to measure both glucose and unidirectional ⁴⁵Ca²⁺ fluxes in a single experiment, when the rate of glucose absorption was some 70-fold greater than that of Ca²⁺. Thus Δc, the concentration difference across the loop, for ⁴⁵Ca²⁺ disappearance was measurable, yet glucose was not depleted by more than 25% in order to maintain apical membrane depolarization. For studies using a high calcium concentration (10.0 mm), we used a phosphate- and Mg²⁺-free buffer containing (mm): 140 NaCl, 3.4 KCl, 12 NaHCO₃ and 10.0 CaCl₂ (Auchere et al. 1997).

To test the effect of drugs on glucose and Ca²⁺ absorption, or to resolve the contributions of SGLT1 and GLUT2 to glucose absorption by selective inhibition of GLUT2 with phloretin, the loop was perfused with glucose for a control period of 0–40 min and then switched to perfusion with glucose and drug for an experimental period of 40–90 min. Perfusions were viable over a period of 90 min. All drugs were from Sigma (UK). The concentration of solvent (ethanol or DMSO) used for drugs or phloretin ranged from 0.1 to 0.3% v/v and solvent alone had no effect on any components of absorption. The unidirectional absorption of Ca²⁺ from lumen to mucosa was determined by disappearance of ⁴⁵Ca²⁺ from the lumen as described elsewhere (Auchere et al. 1998). Total Ca²⁺ was determined by ICP-OES (inductively coupled plasma optical emission spectroscopy) using a Spectro CIROS CCD instrument (Spectro Analytical Instruments, Kleve, Germany) and glucose was determined as previously described (Kellett & Helliwell, 2000). Perfusion rates for both glucose and Ca²⁺ were expressed in μmol min⁻¹ (g dry wt)⁻¹.

PCR of β-subunits

RT-PCR was performed for the four known auxiliary β-subunits using primers designed to rat subunit-specific sequences and cDNA from rat brain (Origene Technologies, Inc.) and a rat jejunal mucosal cDNA preparation provided by Dr D. Meredith, University of Oxford. Primers for rat β-subunits were designed using GenBank accession NM_017346 (Cacnb1), NM_053851 (Cacnb2), NM_012828 (Cacnb3) and XM_215742 (Cacnb4). The β₁-subunit primers were sense (5′-GAC TGG TGG ATC GGG AGG C-3′) and antisense (5′-CAA CCA GCT GCA AGG TCC GG-3′); β₂-subunit primers were sense (5′-ATC CAT CAC AAG AGT CAC TGC-3′) and antisense (5′-GGT GGG GCT CAG AGG TAA AG-3′); β₃-subunit primers were sense (5′-TCA GCC GAC TCC TAC ACC AG-3′) and antisense (5′-GAC GCG GGT GAT GGA GAT C-3′); and β₄-subunit primers were sense (5′-CGG AAG TAC AGA GTG AAA TTG AA-3′) and antisense (5′-ATA CGG TGA GAG AGC TGT GGA-3′). Cycling conditions were 95°C for 10 min followed by 35 cycles of 95°C for 1 min, 60°C (β₁) or 55°C (β₂, β₃, β₄) for 1 min and 72°C for 2 min with final extension at 72°C for 10 min using the Expand High Fidelity PCR System (Roche) and 1.25 mm Mg²⁺. Predicted PCR product sizes for β-subunits were 531 bp (β₁), 525 bp (β₂), 602 bp (β₃), and 409 bp (β₄). All PCR products were ligated directly into the pCR4.1-TOPO vector (Invitrogen) and then sequenced by the Oxford Biochemistry DNA sequencing facility to verify their identity.

Immunocytochemistry

Immunocytochemistry of Ca_v1.3 and β₃-subunits was performed as previously described. Briefly, unperfused jejunum or jejunum perfused as described above (75 mm glucose and 1.25 mm Ca²⁺) was fixed with paraformaldehyde lysine periodate fixative in vivo. Squares of tissue from the midpoint of the segment were fixed, cryoprotected and probed as previously described (Affleck et al. 2003). For Ca_v1.3, the primary antibody supplied by Calbiochem (1: 100 dilution) was an affinity purified rabbit antibody to residues 809–825 (DNKVTIDDYQEEAEDKD) recognizing all rat isoforms. For the β₃-subunit, the primary antibody supplied by Chemicon International (1: 40 dilution) was raised to residues 463–477 (DRNWQRNRPWPKDSY). The secondary antibody in each case was FITC-conjugated goat anti-rabbit IgG (1: 100 dilution, Sigma). In order to demonstrate specificity of labelling, some sections were treated with antibody that had been neutralized by incubation for 1 h with excess of antigenic peptide (antibody to peptide 1: 1 v/v, peptide 50 μg ml⁻¹). Fluorescence micrographs were taken using a Zeiss LSM 510 confocal microscope (Carl Zeiss GmbH). The intensities of apical membrane staining were quantified with LSM 510 Examiner software (v3.3).

Western blotting of membrane vesicles and nuclei

Brush-border and basolateral membrane (BBM and BLM, respectively) vesicles were prepared as previously described; the enrichments of sucrase and Na⁺/K⁺-ATPase were 20-fold and 14-fold, respectively (Corpe et al. 1996; Helliwell et al. 2000a). Every stage of the preparation was performed at 0–4°C to prevent changes in trafficking of proteins after the intestine had been excised. Enrichment of sucrase activity in these highly purified brush-border preparations ranged from 16- to 20-fold; there was no significant enrichment of Na⁺/K⁺-ATPase activity. For Western blots, protein (20 μg) was separated by SDS-PAGE using either 7.5% or 10% gels and transblotted onto polyvinylidene difluoride (PVDF): Ca_v1.3 and β₃-subunits were detected with the same antibodies as for immunocytochemistry; GLUT2, SGLT1 and PKC βII were detected by enhanced chemiluminescence (ECL) as previously described (Helliwell et al. 2000a). The same loading of protein was used for all samples: comparison of relative levels of GLUT2 was made on a protein basis in order to minimize potential complications that might be caused by the trafficking of other proteins in response to the same stimuli that affect GLUT2 trafficking. Specificity of staining was checked by neutralization of antibody with excess antigenic peptide as for immunocytochemistry. The major protease inhibitor in membrane preparation was phenyl methane sulphonyl fluoride (PMSF): the use of a commercial protease inhibitor cocktail (Sigma UK, P8340) did not produce any significant changes in banding patterns.

Statistical analysis

Values are presented as means ± s.e.m. and were tested for significance using paired or unpaired Student's t test as appropriate.

Results

To test the hypothesis that an apical Ca²⁺ entry mechanism is important for the increased glucose absorption capacity at high (75 mm) glucose concentrations, rat mid to distal jejunum was perfused luminally in single-pass mode with KHB prepared with and without addition of Ca²⁺. The concentration of Ca²⁺ in modified KHB is normally 1.25 mm. Ca²⁺ chelators were not added to the Ca²⁺-free solution, as these might disrupt the integrity of the tight-junctions; this solution is therefore referred to as ‘Ca²⁺ deplete’. The intestine was pre-perfused for 20 min with 75 mm mannitol to allow any added effector time to act prior to perfusion with 75 mm glucose and effector; subsequent experiments, however, have shown this pre-perfusion to be unnecessary. To determine if any effects were indeed associated with the GLUT2 component, 1 mm phloretin, which inhibits GLUT2 but not SGLT1 in whole intestine, was added after an initial steady-state rate of glucose absorption was achieved (Fig. 1A). Experiments were also conducted without addition of phloretin to confirm that the steady state could be maintained for the duration of the experiment and was not affected by addition of solvent vehicle (data not shown). For control (1.25 mm Ca²⁺) perfusions, the initial steady-state rate was 39.9 ± 0.5 μmol min⁻¹ (g dry wt)⁻¹ and the SGLT1 component was 13.0 ± 0.3 μmol min⁻¹ (g dry wt)⁻¹ (Fig. 1A and B). These values are very similar to those previously reported; we have further established that use of cytochalasin B, or phloridzin or replacement of Na⁺ with choline (no inhibitors or solvents) give the same quantitative data (Kellett & Helliwell, 2000; Helliwell & Kellett, 2002). It is clear that the Ca²⁺ deplete condition affects the initial steady-state rate (24.2 ± 0.4 μmol min⁻¹ (g dry wt)⁻¹, P < 0.001, n=4), while the phloretin-insensitive rate is unaffected (13.9 ± 0.2 μmol min⁻¹ (g dry wt)⁻¹). Hence luminal Ca²⁺ concentration regulates the GLUT2- but not the SGLT1-mediated component.

Figure 1. The effect of luminal Ca²⁺ and L-type effectors on glucose absorption.

A, a representative time course showing the effect of Ca²⁺. Rat mid and distal jejunum was pre-perfused in vivo in single-pass mode with modified KHB + 100 mm mannitol containing either 1.25 mm Ca²⁺ (▪) or no Ca²⁺ (□, Ca²⁺-deplete perfusate); after 20 min the perfusate was switched to 75 mm glucose + 25 mm mannitol to achieve an initial steady state (30–50 min) before the addition of 1 mm phloretin at 50 min (arrow) to determine the phloretin-insensitive (SGLT1) steady-state rate (60–90 min). B, the effect of L-type effectors, which were present from the start of perfusions. Total (black bars) and phloretin-insensitive (hatched bars) steady-state rates were obtained for control, Ca²⁺ deplete, nifedipine (10 μm, Nif) and verapamil (100 μm, Ver) perfusions (n=4, each case). Rates are means ± s.e.m. expressed in μmol min⁻¹ (g dry wt)⁻¹. Student's t test was used to determine statistical significance. †††P < 0.001, unpaired test for the comparison of the initial steady-state rates of the Ca²⁺ deplete, nifedipine and verapamil to that of the control. ***P < 0.001, **P < 0.01, *P < 0.05, paired test comparing the initial steady-state rate with its phloretin-insensitive rate. The phloretin-insensitive rates are not statistically different from one another.

We have previously demonstrated that the non-classical L-type channel Ca_v1.3 is located in the apical membrane of jejunum, and have detected L-type activity pharmacologically in the presence of 20 mm glucose. We wished therefore to determine whether an L-type channel could provide the entry mechanism responsible for the Ca²⁺ effects at high glucose concentration. The rates of glucose and ⁴⁵Ca²⁺ absorption were determined in the presence and absence of two L-type channel inhibitors, nifedipine (10 μm) and verapamil (100 μm). The concentration of nifedipine was the same as that used for studies of Ca_v1.3 in INS-1 cells (Huang et al. 2004), while that of verapamil was the same as that widely used in the study of Ca²⁺ channels, such as TRPV5/6 (Peng et al. 2000). Figure 2 shows that both inhibitors caused significant reductions in both glucose absorption (51.8 ± 8.6% and 45.3 ± 1.0% for nifedipine and verapamil, respectively; n=4, P < 0.001 each) and ⁴⁵Ca²⁺ absorption (40.8 ± 4.5% and 38.1 ± 0.3%, P < 0.001). To confirm that the verapamil and nifedipine effects were associated solely with the GLUT2 component, paired perfusions with phloretin (1 mm) were also conducted. These show clearly that initial steady-state rates were reduced to a similar extent to that of the Ca²⁺ deplete experiments (Fig. 1B, 36.9 ± 2.4%, 50.4 ± 1.8% and 40.6 ± 3.0% for Ca²⁺ deplete, nifedipine and verapamil, respectively, P < 0.001), while the phloretin-insensitive rate was unaffected. Thus the GLUT2-mediated component of glucose absorption is regulated by Ca²⁺, while the SGLT1 component is not.

Figure 2. Inhibition of glucose absorption occurs concomitantly with inhibition of ⁴⁵Ca²⁺ absorption by L-type channel antagonists.

Jejunum was perfused in vivo with perfusate comprising modified KHB containing 75 mm glucose, 1.25 mm Ca²⁺ and ⁴⁵Ca²⁺ (0.35 kBq ml⁻¹) as a tracer to determine unidirectional lumen-to-mucosa ⁴⁵Ca²⁺ absorption. After a control period, the perfusion was switched (arrow) to an experimental period in which perfusate contained an L-type antagonist. Time courses are shown for: (A) the rate of glucose absorption when (C) ⁴⁵Ca²⁺ absorption was inhibited by nifedipine (10 μm) and for (B) the rate of glucose absorption when (D) ⁴⁵Ca²⁺ absorption was inhibited by verapamil (100 μm). Absorption rates are presented as μmol min⁻¹ (g dry wt)⁻¹ and means ± s.e.m., n=4.

In principle, inhibition of glucose and ⁴⁵Ca²⁺ absorption in in vivo perfusions could have arisen if Ca²⁺ channel inhibitors reduced water absorption, so increasing the volume of the perfusate in the experimental phase (containing L-type effectors) relative to that of the control phase and resulting in the reduction of luminal glucose and Ca²⁺ concentrations. A reduction in average luminal glucose concentration might act to repolarize the apical membrane and inhibit Ca²⁺ absorption. Water absorption is indeed inhibited (nifedipine 0.062 ± 0.006 ml min⁻¹ (g dry wt)⁻¹ and verapamil 0.05 ± 0.02 ml min⁻¹ (g dry wt)⁻¹ compared with 0.151 ± 0.02 ml min⁻¹ (g dry wt)⁻¹ in the control period; P < 0.001). In the control period of perfusions with an initial concentration of 75 mm glucose, the average concentration in the luminal effluent from single-pass experiments was 66.6 ± 0.5 mm and the average value throughout the loop was therefore 70.8 mm. However, inhibition of water absorption did not cause a reduction in the effluent glucose concentration. Indeed, the concentration was slightly increased in the presence of nifedipine and remained unaltered in the case of verapamil (70.8 ± 0.6 and 66.8 ± 1.4 mm, respectively, so that the average concentrations through the loop were 72.9 and 70.9 mm, respectively). Similarly, total [Ca²⁺] determined by ICP-OES remained constant within experimental error between the experimental and control periods. Changes in glucose and unidirectional Ca²⁺ absorption could not therefore be caused by changes in luminal concentration as a secondary consequence of effects on water transport. This conclusion is in agreement with that of Younoszai & Nathan (1985), who showed that the 5-fold increase in water absorption on switching from an isotonic and a hypotonic solution had little effect on Ca²⁺ absorption in rat jejunum.

Brush-border membrane vesicles were then prepared to determine whether the Ca²⁺ effects were associated with alterations in GLUT2 trafficking to the apical membrane (Fig. 3). Vesicles were prepared from rat jejunum initially perfused in vivo with 75 mm glucose in modified KHB (1.25 mm Ca²⁺). At 20 min, perfusates were switched to 75 mm glucose in KHB containing either 1.25 mm Ca²⁺ (control), 0 mm Ca²⁺ (Ca²⁺ deplete), or 1.25 mm Ca²⁺ containing nifedipine (10 μm). Vesicles were then immunoblotted to determine the relative levels of GLUT2, PKC βII and SGLT1; while a representative blot is shown in Fig. 3, we have quantified the data from three separate blots. There were significant reductions in apical membrane levels relative to control: GLUT2 49.7 ± 3.5% (Ca²⁺ deplete, P < 0.01), 37.3 ± 9.7% (nifedipine, P < 0.05) and PKC βII 44.1 ± 2.1% (Ca²⁺ deplete, P < 0.01), 44.9 ± 4.2% (nifedipine, P < 0.01). As expected the levels of SGLT1 remained unaltered. This confirms that inhibition of Ca²⁺ entry inhibits both the activation of PKC βII and the trafficking of GLUT2 to the apical membrane.

Figure 3. Dependence of GLUT2, PKC βII and SGLTI levels on luminal Ca²⁺ and nifedipine.

Brush-border membrane vesicles were prepared from rat jejunum initially perfused in vivo with 75 mm glucose in modified KHB (1.25 mm Ca²⁺). At 20 min, perfusates were switched to 75 mm glucose in KHB containing either 1.25 mm Ca²⁺ (control), 0 mm Ca²⁺ (Ca²⁺ deplete), or 1.25 mm Ca²⁺ containing nifedipine (10 μm). Vesicle protein (20 μg) was then separated by SDS-PAGE (10% gels), transblotted on to PVDF and Western blotted for GLUT2, SGLT1 and PKC βII. For full details, see Methods.

Since apical Ca²⁺ entry involves a channel with sensitivity to L-type effectors, it would be expected that membrane depolarization by Na⁺–glucose cotransport should increase Ca²⁺ absorption. This is very clearly demonstrated with 10 mm Ca²⁺ in the lumen, that is, when there is a substantial transepithelial gradient. In paired comparisons made within a single perfusion, 75 mm glucose was perfused for 0–40 min and was then substituted with 75 mm mannitol for 40–80 min; Fig. 4A shows that the rate of ⁴⁵Ca²⁺ absorption with glucose was 2-fold that with mannitol (P < 0.001, n=4). In unpaired comparisons, that is, when mannitol and glucose were used alone in separate perfusions (Fig. 4B), glucose increased the steady-state rate of ⁴⁵Ca²⁺ absorption 3-fold (P < 0.001, n=8).

Figure 4. Activation of ⁴⁵Ca²⁺ absorption by glucose.

A, the rates of ⁴⁵Ca²⁺ absorption in the presence of either mannitol or glucose were compared within a single perfusion (paired comparison): rat jejunum was perfused for a control period of 40 min with 75 mm glucose and 10 mm Ca²⁺ (n=4). After 40 min (arrow), perfusion was continued for an experimental period in which 75 mm mannitol was substituted for 75 mm glucose. B, separate perfusions (unpaired comparison): rat jejunum was perfused with 75 mm glucose in the presence of 10 mm Ca²⁺ for 40 min only (n=8); separate perfusions were undertaken in which mannitol replaced glucose (n=8). P < 0.001 for comparison of the effects of glucose and mannitol in A and B (†††).

Figure 5 confirms that the channel operating at 75 mm glucose and 10 mm Ca²⁺ is under the control of SGLT1 and has L-type characteristics, that is, no new channel appears to have come into play. Thus 1 mm phloridzin inhibited ⁴⁵Ca²⁺ absorption by 72%, the phloridzin-insensitive component being similar to the rate of absorption in the presence of mannitol alone (Fig. 5C); note that glucose absorption was concomitantly inhibited by 80%, as phloridzin inhibits SGLT1 directly and also indirectly blocks that part of GLUT2 which rapidly traffics away from the membrane through inhibition of Ca²⁺ absorption (Fig. 5A; see Kellett & Helliwell, 2000). At 10 mm Ca²⁺, 10 μm nifedipine inhibited ⁴⁵Ca²⁺ absorption by 69% (Fig. 5D); glucose absorption was also concomitantly inhibited by 43% to give in Fig. 5B a rate of 20.13 ± 0.56 μmol min⁻¹ (g dry wt)⁻¹, similar to that at 1.25 mm Ca²⁺ seen under Ca²⁺-deplete conditions or with nifedipine or verapamil (Fig. 1); the nifedipine-insensitive component represents absorption mediated by SGLT1 plus that part of apical GLUT2 which does not traffic readily away from the membrane (see Fig. 3). This has been confirmed by showing that in the presence of nifedipine, phloretin diminishes the rate of glucose absorption from 21.45 ± 0.33 to 10.31 ± 0.26 μmol min⁻¹ (g dry wt)⁻¹ (P < 0.001, n=3); for comparison, phloretin diminishes the rate of glucose absorption at 75 mm glucose and 10 mm Ca²⁺ from 34.99 ± 0.68 to 9.98 ± 0.82 μmol min⁻¹ (g dry wt)⁻¹ (P < 0.001, n=3), the insensitive component being assigned to SGLT1. The magnitudes of the components of glucose absorption with either 20 mm or 75 mm glucose are not significantly different at 10 mm compared with 1.25 mm Ca²⁺.

Figure 5. Ca²⁺ absorption at high Ca²⁺ concentration also displays L-type characteristics.

A and C, jejunum was perfused for a control period of 40 min with 75 mm glucose and 10 mm Ca²⁺ (n=4). After 40 min (arrow), perfusion was continued for an experimental period with 1 mm phloridzin also present in the perfusate. A and C show the glucose and Ca²⁺ fluxes, respectively. B and D, jejunum was perfused for a control period of 40 min with 75 mm glucose and 10 mm Ca²⁺ (n=4). After 40 min (arrow), perfusion was continued for an experimental period with 10 μm nifedipine also present in the perfusate. B and D show the glucose and Ca²⁺ fluxes, respectively. Rates are presented as μmol min⁻¹ (g dry wt)⁻¹.

We have reported that the α₁ pore-forming subunit of the Ca_V1.3 channel is located in the apical membrane of rat jejunum (Morgan et al. 2003). It was therefore important to identify and localize the auxiliary β-subunit(s) in jejunum. The primary screen was by homology-based RT-PCR; the primers were designed to all four known rat β-subunits and spanned intronic sequences. As shown in Fig. 6, PCR detected all four β-subunit transcripts expressed in brain cDNA (Ludwig et al. 1997); sequencing confirmed their 100% identity with the reported rat sequences. A single β₃ PCR product of the correct size was detected in the rat jejunal mucosal cDNA preparation. Sequencing confirmed its 100% identity with the reported rat β₃ sequence. PCR screening revealed exclusive expression of the β₃-subunit transcript in jejunal mucosa. To determine the localization of β₃, Western blots of BBM and BLM vesicles were prepared after perfusion of jejunum with 75 mm glucose and probed with β₃ antibody. β₃ was detected only at the apical membrane as a single band of ∼64 kDa, consistent with the reported molecular mass of 67 kDa in neuronal cells (Pichler et al. 1997). Immunocytochemistry of jejunum confirmed that both Ca_v1.3 and Ca_vβ₃ are located at the apical membrane of rat jejunum (Fig. 7). Neutralization of antibody with excess peptide confirms that labelling is specific. There appears also to be significant intracellular labelling of Ca_v1.3 at the nucleus and in the cytosol, as well as labelling of both proteins within the lamina propria. Similar results were obtained whether jejunum was perfused or unperfused.

Figure 6. Rat jejunal mucosa expresses β₃ mRNA and β₃ protein is located in the apical membrane.

A, RT-PCR was performed for the four known auxiliary Ca_vβ subunits using primers designed to rat specific sequences and cDNA derived from rat brain and rat jejunal mucosa. B, jejunum was perfused for 30 min with 75 mm glucose; brush-border membrane (BBM) and basolateral membrane (BLM) vesicles were then prepared and Western blotted for β₃. The BBM band was eliminated by neutralization of antibody with excess peptide (data not shown).

Figure 7. Immunocytochemical localization of Ca_v1.3 and β₃-subunit in the proximal jejunum.

Sections of unperfused jejunum were labelled with a rabbit antibody that recognizes all forms of rat Ca_v1.3 (A and B) and with a rabbit antibody which specifically recognizes the β₃ subunit (C and D). The secondary antibody was FITC-conjugated goat anti-rabbit IgG. The peptide controls were sections treated with antibody to Ca_v1.3 (B) and β₃ (D) that had been pre-absorbed with excess antigenic peptide; non-specific staining in the lamina propria of these sections was seen with the secondary alone. All sections are at 63 × magnification and were taken at the same settings with a Zeiss LSM 510 confocal microscope.

Discussion

Regulation of glucose absorption by Ca²⁺

When Ca²⁺ was omitted from the luminal perfusate at high glucose concentrations, glucose absorption was inhibited (Fig. 1A). Inhibition occurred even though Ca²⁺ supply from the blood was still maintained in perfusions in vivo. Selective inhibition with phloretin showed that only the apical GLUT2 component of glucose absorption was inhibited, by 60%; the SGLT1 component was unaffected. Apical GLUT2 levels were decreased to a similar extent, but SGLT1 levels remained unaltered (Fig. 3). A possible reason for the lack of an effect on SGLT1 is that the chow diet used to feed the rats up-regulates SGLT1 (Kellett & Barker, 1989). That GLUT2 trafficking and absorption depend on luminal Ca²⁺ is consistent with the observation that nifedipine and verapamil inhibited the GLUT2 component of absorption and apical GLUT2 insertion in a manner very similar to that caused by omission of luminal Ca²⁺; there was again no effect on SGLT1 (Figs 2 and 3). Moreover, nifedipine and verapamil inhibited over 40% of ⁴⁵Ca²⁺ absorption from the lumen at 75 mm glucose and the inhibition of ⁴⁵Ca²⁺ absorption was concomitant with that of glucose absorption (Fig. 2).

The link between Ca²⁺ entry and glucose absorption must be cytosolic Ca²⁺, but there appears to be no way of measuring this directly during a perfusion in vivo. Fortunately, there are two established biochemical markers, which show that increased Ca²⁺ absorption goes with increased cytosolic Ca²⁺ and increased glucose absorption at high concentrations. Thus, apical GLUT2 insertion correlates with activation of PKC βII, which requires an increase in cytosolic Ca²⁺. Moreover, glucose induces cytoskeletal rearrangement of the enterocyte, which is dependent on an increase in cytosolic Ca²⁺ to stimulate myosin phosphorylation (Turner et al. 1997). Figure 3 shows that PKC βII is inactivated in Ca²⁺-deplete conditions or by luminal application of nifedipine. The accompanying paper shows that these two conditions, as well as replacement of glucose by mannitol or Ca²⁺-deplete conditions, block myosin phosphorylation in the terminal web and hence cytoskeletal rearrangement and apical GLUT2 insertion (Mace et al. 2007).

We have already reported that at 20 mm luminal glucose in vivo a significant pathway of jejunal Ca²⁺ absorption is mediated by a channel with L-type characteristics (Morgan et al. 2003). Although L-type effectors, namely nifedpine, Mg²⁺ and Bay K 8644, caused marked changes in Ca²⁺ absorption, there was little change in glucose absorption. In addition, repolarization of the membrane with phloridzin rapidly inhibited ⁴⁵Ca²⁺ absorption; this observation is consistent with the fact that the apparent K_m of glucose transport by SGLT1 is 23–26 mmin vivo (Debnam & Levin, 1975; Kellett & Helliwell, 2000). In those experiments therefore it appeared that glucose regulated Ca²⁺ absorption, but not that Ca²⁺ regulated glucose absorption. The reason is that at low glucose (20 mm and below, or mannitol), there is a basal level of GLUT2; insertion of additional GLUT2 only occurs at 30 mm and above. This statement applies only to fed rats maintained on Bantin & Kingman (UK) chow diet; the situation is different in other conditions, for example, overnight starved mice maintained on different chow diets (Gouyon et al. 2003). In the absence of luminal Ca²⁺ (Ca²⁺ deplete), only the basal level of GLUT2 is observed at 75 mm glucose, when addition of 1.25 mm Ca²⁺ to the lumen increases insertion (Figs 1 and 3).

We therefore conclude that the presence of luminal Ca²⁺ is a prerequisite for apical GLUT2 insertion in vivo. However, we note that apical GLUT2 insertion is not detectable in our experiments until 30 mm glucose, whereas repolarization of the membrane with phloridzin inhibited ⁴⁵Ca²⁺ absorption at 20 mm glucose (Morgan et al. 2003). Also, increasing luminal Ca²⁺ from 1.25 mm to 10 mm increases the total rate of Ca²⁺ absorption almost 8-fold, but does not affect the rate of glucose absorption or the two components, either at 75 mm (see above) or at 20 mm glucose (Morgan et al. 2003). Luminal Ca²⁺ is therefore a necessary, but not a sufficient, requirement for insertion; a second, unknown signal is required at high glucose concentrations.

Regulation of Ca²⁺ absorption by glucose

As noted, previous work indicated that glucose regulated Ca²⁺ absorption. In addition, Western blotting and immunocytochemistry revealed Ca_v1.3 in the apical membrane of jejunum (see also Fig. 7). Since no other known channel in intestine can explain these observations, the L-type activity was attributed to Ca_v1.3 (Morgan et al. 2003). The present perfusion data at 75 mm glucose confirm the existence of an apical route of Ca²⁺ entry with L-type characteristics. Moreover, following identification of transcripts for Ca_vβ₃ L-type channel subunits in rat jejunal mucosal cDNA, Ca_vβ₃ was localized to the apical membrane by Western blotting and immunocytochemistry (Figs 6 and 7). This first demonstration of the presence of Ca_v1.3 and its auxiliary subunit Ca_vβ₃ provides strong evidence for the existence of an L-type channel in the apical membrane of the mucosal epithelium. Auxiliary β-subunits direct α1-subunits to the plasma membrane by concealing an endoplasmic reticulum retention signal (Bichet et al. 2000) and appear important for functional expression of the α1-subunit at the cell surface (Gao et al. 1999). Co-expression studies in oocytes show that four different β isoforms, including Ca_vβ₃, enhanced Ca²⁺ currents compared with Ca_v1.3 alone, but had no effect on channel activation threshold (Xu & Lipscombe, 2001).

Perfusion of 75 mm glucose at 10 mm Ca²⁺ increased the total rate of ⁴⁵Ca²⁺ absorption up to 3-fold compared with 75 mm mannitol (Fig. 4B). Thus, when there was a large transepithelial gradient of Ca²⁺ under these conditions, there was a large glucose-induced ⁴⁵Ca²⁺ flux, which was SGLT1-dependent since it was abolished by phloridzin. In principal, such a flux could have arisen by paracellular flow. However, the glucose-induced ⁴⁵Ca²⁺ flux was also abolished by nifedipine, indicating that it was mediated by an L-type channel. Moreover, at 20 mm glucose, nifedipine inhibits absorption of 10 mm Ca²⁺ in the lumen with a time course identical to that at 1.25 mm Ca²⁺ (Morgan et al. 2003). In the latter case, any absorption by paracellular flow should be minimal because plasma free Ca²⁺ is 1.25 mm. Of particular interest, our data suggest a signalling role for Ca_v1.3 in intestine similar to that in pancreatic β cells. Thus the cultured cell line INS-1 contains both Ca_v1.2 and Ca_v1.3, yet glucose-stimulated insulin secretion is preferentially coupled to Ca_v1.3 (Liu et al. 2003); moreover, glucose also induces a nifedipine-sensitive Ca²⁺ influx through Ca_v1.3 (Huang et al. 2004). The nifedipine- and phloridzin-insensitive component of Ca²⁺ absorption is most likely channel-mediated, probably by TRPV5/6 (Mace et al. 2007).

Nutrient sensing: the integration of glucose and Ca²⁺ absorption

Glucose regulates Ca²⁺ absorption and Ca²⁺ regulates glucose absorption. The signalling and absorptive roles of Ca²⁺ and glucose are not distinct, rather they are integrated. The primary signal for regulation of both glucose and Ca²⁺ absorption is glucose. High concentrations depolarize the apical membrane by transport through the Na⁺–glucose cotransporter, which exerts an important regulatory role, so that Ca²⁺ absorption is increased up to 3-fold by induction of a nifedipine-sensitive component through Ca_v1.3. Binding of Ca²⁺ to calbindin-D_9K prevents the increase in Ca²⁺ absorption from flooding the absorptive cell and stimulates transcellular transport, which would in principle act to attenuate potential increases in cytosolic Ca²⁺. However, studies in kidney have revealed that the kinetics of Ca²⁺ binding to calbindin-D_28K are slow relative to the initiation of the rise in cytosolic Ca²⁺ (Koster et al. 1995). Ca²⁺ is therefore able to exert a signalling role with respect to glucose absorption in intestine, despite the presence of calbindin-D_9K, which is necessary for normal absorptive function.

Although we have focused on glucose and Ca²⁺, the transport of other nutrients, such as Na⁺-dependent amino acids or H⁺-dependent peptides, may also depolarize the apical membrane. There exists, therefore, a network of nutrient interactions beyond glucose, which has the potential to modulate Ca²⁺ absorption. This does not mean that these other nutrients are necessarily capable of promoting apical GLUT2 insertion. Their ability to do so will depend ultimately on whether they are also able to provide the second signal, which glucose does so effectively at high but not low concentrations, and also on their rate of delivery to the jejunum. Potential nutrient interactions are also modulated by long-term diet. Maintenance of mice on a low carbohydrate/high protein diet prevents apical GLUT2 formation and blocks its induction by a bolus of sugar (Gouyon et al. 2003). This observation may explain why patients with glucose galactose absorption syndrome, who are required to avoid dietary sugar, cannot absorb glucose even in the presence of amino acids.

Our observations emphasize the importance of studying the absorption of one nutrient in the presence of another. Nevertheless, the number of papers in the literature in which the effect of glucose on Ca²⁺ absorption has been studied are small (Younoszai & Nathan, 1985; Carroll et al. 1988). The majority of work on Ca²⁺ absorption has been done on starved rats in the absence of nutrient, conditions which cause membrane hyperpolarization and diminish the contribution of L-type channels, so that they may not be functionally detectable with L-type antagonists (Favus & Angeid-Backman, 1985; Fox & Green, 1986). Other possible reasons why the role of Ca_v1.3 has been overlooked include the fact that the majority of work has been in duodenum, where active sugar transport and Ca_v1.3 are low, but TRPV5/6 are high. In addition, Ca_v1.3 is 10-fold less sensitive to L-type channel blockers than are classical channels (Xu & Lipscombe, 2001).

The physiological significance of the regulatory and absorptive mechanisms described seems clear. Ca²⁺ is normally plentiful when food is plentiful, so that the soluble Ca²⁺ concentration in the lumen is about 5–10 mm after a meal (Bronner, 2003). The initial digestion products along with free and complexed dietary Ca²⁺ pass into the duodenum, where active transport of glucose is low, as are the level and activity of Ca_v1.3. Even though the activities of TRPV5/6 in duodenum are at their highest in the gastrointestinal tract, Ca²⁺ absorption is very restricted because the transit time through the duodenum in rat is just 2.5 min (Bronner, 2003). In contrast, the transit time from the proximal jejunum to the mid-ileum, where apical Ca_v1.3 is located (Morgan et al. 2003), is 126 min. In this region, the final digestion products, namely glucose, amino acids and oligopeptides, are generated in high local concentrations at the apical membrane by membrane-bound hydrolases. The transport of these products depolarizes the apical membrane and induces the nifedipine-sensitive component of Ca²⁺ absorption through Ca_v1.3, helping to clear most of the free or loosely complexed Ca²⁺. Glucose absorption is up-regulated by Ca²⁺-dependent apical GLUT2 insertion, which seems to require another signal in addition to Ca²⁺. As glucose is absorbed, the process is reversed; Ca²⁺ absorption is down-regulated as the apical membrane is repolarized and glucose absorption is down-regulated by loss of apical GLUT2. The integration of glucose and Ca²⁺ absorption represents a complex nutrient sensing system, which allows both absorptive pathways to be regulated rapidly and precisely to match dietary intake.