| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Journal of Physiology (2002), 538.1, pp. 121-131
© Copyright 2002 The Physiological Society
DOI: 10.1113/jphysiol.2001.012969
 |
ABSTRACT |
|---|
 TopAbstract IntroductionMethodsResultsDiscussionReferences |
|---|
Fatty acids induce cholecystokinin (CCK) secretion both in humans and from murine enteroendocrine cell lines. In both cases, only fatty acids above a critical acyl chain length (C10) are capable of inducing a response. Using the enteroendocrine cell line STC-1, the aim of this study was to determine whether this acyl chain length dependency is related to the fact that longer chain fatty acids are relatively insoluble in aqueous solutions and, if so, whether it is insoluble aggregates of fatty acids rather than free fatty acids which evoke CCK secretion. Solutions of fatty acids (chain length C8-C14), which were judged by filtration and Zeta sizer measurement to contain no fatty acid aggregates, never evoked CCK secretion from STC-1 cells. Filtering fatty acid solutions (of chain length C10, C12 and C14) through polytetrafluoroethylene (PTFE) filters (0.45 µm pore size) revealed a narrow concentration range for each acid over which the amount of fatty acid removed from the solution increased sharply due to the formation of fatty acid aggregates. Filtration experiments, in which suspensions of C10, C12 and C14 fatty acids were passed through pore sizes of 0.2, 0.45 or 1.2 µm, suggested that STC-1 cells did not respond to fatty acid aggregates of greater than 1.2 µm, while at least 50 % of the CCK response was mediated by aggregates which were smaller than 0.45 µm. Fatty acids induce CCK secretion from STC-1 cells by elevating intracellular Ca2+ concentration ([Ca2+]i). We therefore measured the effects on [Ca2+]i of filtered C10, C12 and C14 fatty acids. In all cases, [Ca2+]i responses were closely correlated with CCK secretion. Interestingly, while filtrates of fatty acid solutions evoked CCK secretion and elevated [Ca2+]i, freshly prepared solutions of fatty acids at the same concentration as the filtrates did not. This suggested that fatty acid aggregates were not in equilibrium with the solvent after filtration. The observation that the ability of C10, C12 and C14 filtrates to elevate [Ca2+]i decayed with time was consistent with this hypothesis. Furthermore, sonication of the filtrates abolished their ability to elevate [Ca2+]i. These data further suggest that it is a physical property of the fatty acid solution (the presence of insoluble fatty aggregates) which is responsible for the observed cellular responses. We conclude that Ca2+ mobilisation and CCK secretion in STC-1 cells is driven by a signal transduction mechanism that senses insoluble fatty acid aggregates, rather than free fatty acids in solution.
(Resubmitted 10 July 2001; accepted after revision 26 September 2001)
Corresponding author R. Benson: G38 Stopford Building, University of Manchester, Manchester M13 9PT, UK. Email: rod.benson{at}man.ac.uk
* These authors contributed equally to this work.
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The signalling peptide cholecystokinin (CCK) plays a key role in regulating a range of intestinal responses to luminal nutrients, including stimulation of pancreatic secretion, gallbladder emptying and inhibition of gastric motility (Hopman et al. 1985; Liddle et al. 1986; Higham et al. 1997; Liddle, 1997). Collectively, these responses help to integrate and optimise the digestion and absorption of nutrients (Liddle, 1997). CCK also acts centrally as a satiety signal (Smith & Gibbs, 1994). The secretion of CCK is particularly responsive to luminal fatty acids rather than triglycerides (Guimbaud et al. 1997), an effect which is critically dependent on the acyl chain length; saturated fatty acids below an acyl chain length of 12 carbon atoms do not induce CCK secretion in humans (McLaughlin et al. 1999).
We have also shown that this chain length dependency of fatty acid-induced CCK release can be reproduced in two murine enteroendocrine cell lines (McLaughlin et al. 1998; Sidhu et al. 2000): one of colonic origin, GLUTag (Drucker et al. 1994), and one of duodenal origin, STC-1 (Rindi et al. 1990). These data suggest a signal transduction pathway that may involve a receptor specific for fatty acids. In these cells, fatty acid-induced CCK secretion is accompanied by a rise in [Ca2+]i, which is largely mediated by the influx of extracellular Ca2+, probably through L-type Ca2+ channels (McLaughlin et al. 1998; Sidhu et al. 2000).
In both enteroendocrine cell lines, the potency of fatty acids is also related to chain length; dodecanoic acid (C12) stimulates both CCK secretion and a rise in [Ca2+]i at a lower concentration than decanoic acid (C10). The limit of solubility for a fatty acid in an isotonic, divalent cation-containing solution is inversely related to its chain length. However, the concentrations of fatty acids required to induce CCK secretion are always greater than this solubility limit (Sidhu et al. 2000). Also, fatty acids which are presented to STC-1 cells on a carrier protein such as bovine serum albumin (BSA) do not elicit either CCK secretion or a calcium response (Sidhu et al. 2000). These observations raise the question of whether fatty acid-induced CCK secretion is signalled by insoluble fatty acid aggregates rather than by free fatty acids dissolved in solution.
We have addressed this question by filtering fatty acid solutions through PTFE filters to remove fatty acid aggregates and testing the ability of the filtrates to elevate [Ca2+]i and evoke CCK secretion in STC-1 cells. Our data suggest that STC-1 cells responded to fatty acid aggregates rather than fatty acid monomers which are in solution.
|
METHODS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Cell culture consumables, Dulbecco's modified Eagle's medium (DMEM), horse serum, fetal bovine serum and penicillin- streptomycin solution were purchased from Gibco (Paisley, UK). Phosphate buffered saline (PBS) tablets were purchased from Oxoid (Basingstoke, UK). Saturated fatty acids, (C8 (octanoic), C9 (nonanoic), C10 (decanoic), C11 (undecanoic), C12 (dodecanoic), C13 (tridecanoic), C14 (tetradecanoic) and C18 (octadecanoic)) were purchased from Sigma-Aldrich (Poole, UK). All other materials were obtained from either Sigma or BDH (Merck, Poole, UK) with the exception of Fura-2 AM and Pluronic F-127 which were obtained from Molecular Probes (Leiden, Netherlands). 14C-labelled decanoic and dodecanoic acids and 3H-labelled tetradecanoic acid were purchased from Amersham (Bucks, UK).
Tissue culture
STC-1 cells (kindly supplied by D. Hanahan, University of California, USA) were grown in DMEM supplemented with 15 % horse serum and 2.5 % fetal bovine serum in an atmosphere of 5 % CO2-95 % air at 37 °C. Cells were routinely passaged upon reaching 80-90 % confluency by washing the cell layer repeatedly with PBS and incubating with a solution of 0.25 % trypsin for 1 min. Plating densities ranging from 1 
104 to 1 105 cells cm-2 were used for routine subculture in flasks.
For fluorescence studies, STC-1 cells were plated at densities between 8 104 and 1 105 cells cm-2 on coverslips coated with 0.025 % (w/v) poly-L-lysine in 24-well plates. Measurements of [Ca2+]i in STC-1 cells were carried out 24-48 h after seeding. Only cells of passages 4-16 were used in these experiments. Cell viability was estimated by trypan blue exclusion and was greater than 95 %.
Preparation of fatty acid solutions
Fatty acid solutions were prepared as described previously (Sidhu et al. 2000). In summary, fatty acid stock solutions were diluted into a buffered (pH 7.4) isotonic solution with the following composition (mM): 140 NaCl, 4.5 KCl, 10 Hepes acid, 10 Hepes salt, 1.2 CaCl2, 1.2 MgCl2 and 10 glucose, to produce working solutions with a fatty acid concentration of between 50 µM and 1.5 mM. For divalent cation-free solutions, both calcium and magnesium were replaced with 0.2 mM EDTA. Each fatty acid solution was sonicated just prior to use with a probe sonicator (Soniprep 150, Sanyo, Herts, UK) for 4 min at room temperature. For fatty acid solutions prepared in the presence of bile acid, 0.5 mM cholic acid and 5 mM 1,2-di-palmitoyl-sn-glycero 3-phosphocholine (DPPC) were added to the isotonic solution described above and sonicated for 10 min. Fatty acid at the required concentration was then dissolved in this solution, which was sonicated for a further 15 min on ice.
Calcium imaging studies and measurement of CCK
[Ca2+]i in STC-1 cells was determined using dual-excitation fluorescence microscopy with the calcium-sensitive ratiometric dye Fura-2 AM, as previously described (Sidhu et al. 2000). CCK concentration was determined using a radioimmunoassay, as previously described (McLaughlin et al. 1998).
Filtration experiments
A trace amount of 14C-labelled dodecanoic or tetradecanoic acid (0.5 µCi) or 3H-labelled decanoic acid (1 µCi; Amersham, Little Chalfont, UK) was added to its respective fatty acid solution and the solution was sonicated for 5 min. The solutions were then passed through 13 mm PTFE filters (Sartorius, Goettingen, Germany) of 0.2, 0.45 or 1.2 µm pore size using a 1 ml glass Hamilton syringe. The radioactivity of the solution was measured before and after filtration using a 
scintillation counter (Tricarb 460; Packard, Berks, UK) and the proportion of fatty acid remaining in the filtrate calculated. From these data it was possible to estimate the concentration range within which the fatty acid reached its limit of solubility by determining the concentrations at which the amount of fatty acid removed by filtration rapidly increased.
Measurement of fatty acid particle size
In order to confirm the solubility limits obtained by filtration, the average hydrodynamic size of particles in solution (Z average) was determined using a Malvern Zeta sizer 3000 (Malvern, UK). Due to the large polydispersity of fatty acid solutions, it was not possible to obtain accurate size measurements of fatty acid aggregates. However, this method was still suitable for detecting the rapid change in the size (and number) of particles as the solubility limit of the fatty acid was reached.
Data analysis
All data are expressed as means ± S.E.M. Unless otherwise stated statistical comparisons were made using one way ANOVA with a Dunnett post hoc test to compare filtered samples with unfiltered controls or a Newman-Keuls multiple comparison test to compare one filter pore size against another. P values of less than 0.05 were considered significant. In experiments which measured the change in [Ca2+]i, a change of 0.02 units in the 340/380 nm ratio was considered significant. Group comparisons for time course experiments were made using two-way ANOVA with a Bonferroni post hoc test.
|
RESULTS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Estimation of the solubility limit of fatty acids using filtration and particle sizing
Fatty acids induce CCK secretion and a rise in [Ca2+]i in STC-1 and GLUTag cells but these effects are only observed when the concentration of the fatty acid is greater than its solubility limit (Sidhu et al. 2000). As the concentration of fatty acid increases beyond its solubility limit, the solution will rapidly become populated with insoluble fatty acid aggregates. We used this fact to determine the approximate concentration at which each fatty acid reached its solubility limit. While, in our system, Zeta sizer measurement did not give accurate data on particle size, the appearance of insoluble particles as the solubility limit of the fatty acid was reached could easily be detected. A similar principle was applied when using PTFE filters: as the concentration of fatty acid reached its solubility limit, the proportion of fatty acid removed by filtration rapidly increased because the insoluble fatty acid aggregates were trapped by the filter. By examining either the average particle size in solution or the proportion of fatty acid removed by filtration we were able to obtain an estimate of the solubility of each fatty acid. In filtration experiments, we observed a narrow concentration range over which the proportion of fatty acid removed from the solution increased sharply, indicating a marked increase in the formation of fatty acid aggregates which are greater than the filter pore size (Fig. 1). These data were also corroborated by Zeta sizer measurement (data not shown).
 |
View larger version [in this window] [in a new window] |
|
|
Figure 1. The effect of concentration on the percentage of fatty acid removed by filtration The percentage of fatty acid removed when various concentrations of decanoic ( | ||
The relationship between fatty acid solubility, acyl chain length and divalent cations
Using a combination of filtration and particle size measurement, we examined the solubility of fatty acids having acyl chain lengths of C8 to C14 in the presence (Fig. 2A) or absence (Fig. 2B) of divalent cations. The presence of divalent cations in the form of Mg2+ (1.2 mM) and Ca2+ (1.2 mM) greatly reduced the solubility of fatty acids in aqueous solutions. For example, while dodecanoic acid was soluble at 1 mM in a Ca2+- and Mg2+-free isotonic solution, its solubility was lowered to just over 100 µM when divalent cations were present. There was also a close correlation between the presence of insoluble fatty acid aggregates and CCK secretion; CCK secretion was never observed when the fatty acid solution was devoid of fatty acid aggregates. Only four of the fatty acid preparations which contained insoluble particles did not induce CCK secretion (
, Fig. 2A) and these solutions were always at the lower limit of fatty acid aggregation.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 2. The effect of divalent cations, fatty acid chain length and concentration on the formation of fatty acid aggregates in solution Summary of the relationship between the concentration and chain length of fatty acids and their solubility in the presence (A) or absence (B) of divalent cations (1.2 mM Ca2+ and 1.2 mM Mg2+). | ||
The effect of cholic acid on CCK secretion in STC-1 cells
The presence of fatty acids in the gut induces the secretion of bile acids which solubilise them by the formation of micelles in preparation for absorption across the intestinal epithelia (Liddle, 1997). We therefore attempted to examine what effect the presence of cholic acid would have on fatty acid-induced CCK secretion in STC-1 cells. Preliminary experiments indicated that it was not possible to use physiologically relevant concentrations of cholic acid to obtain micellar fatty acid solutions because the thermodynamics controlling the formation of a dodecanoic-cholic acid micelle were unfavourable. We were able to address this problem by also including the phospholipid 1, 2-di-palmitoyl-sn-glycero 3-phospho-choline (DPPC) which causes the formation of a more complex fatty acid micelle that resembles a 'pill box' structure (Small et al. 1969). The formation of this entity occurs using concentrations of cholic acid and DPPC of 0.5 mM and 5 mM, respectively. Using this protocol, we examined the effect on CCK secretion of exposing STC-1 cells to this micelle for 1 h in the presence or absence of either dodecanoic (500 µM) or octadodecanoic (C18; 500 µM) acid. We found that the variability of CCK secretion was increased and differed markedly from preparation to preparation. In general, we noticed that the presence of the bile acid and phospholipid increased CCK secretion levels above those in untreated controls (0.8 ± 0.11 and 1.17 ± 0.24 pmol (mg protein)-1 for control and cholic acid, respectively) but, because of the large inter-experimental variation, these differences were not significant. Likewise, dodecanoic acid and octadodecanoic acid presented with the phospholipid- cholic acid micelle also appeared to increase CCK secretion (3.1 ± 0.9 and 1.9 ± 0.01 pmol (mg protein)-1 for C12 and C18, respectively) but the increase was only significant when C12-cholic acid data were compared with those of the untreated control (P < 0.05; one way ANOVA with Bonferroni's multiple comparison test); neither dodecanoic nor octadodecanoic acid caused CCK secretion significantly different from that in the bile acid control. Finally, 500 µM dodecanoic acid presented free in solution caused enhanced CCK secretion in this series of experiments (6.4 ± 1.6 pmol (mg protein)-1) and this value was significantly different from all other treatment conditions (P < 0.001). Taken together, these data appeared to indicate that the presence of bile acid reduced the effectiveness of the fatty acid to induce CCK secretion in STC-1 cells. However, given the chemical complexity of these solutions, concerns about the detergent effects of bile acids on the cell membrane and the observed inter-experimental variability, we decided not to pursue the bile acid presentation method further.
Filtering dodecanoic acid preparations reduces CCK secretion
We evaluated the effect of insoluble dodecanoic acid aggregates on CCK secretion directly, by filtering the solutions and then adding the filtrates to STC-1 cells. Figure 3A shows the time course of CCK secretion by STC-1 cells exposed to 500 µM dodecanoic acid which had been filtered through PTFE filters of 1.2, 0.45 and 0.2 µm pore sizes. Filtering dodecanoic acid through a 1.2 or 0.45 µm filter did not significantly reduce the amount of CCK secreted over the time course of the experiment, while filtration through a 0.2 µm filter significantly reduced CCK secretion at both 30 (P < 0.05) and 60 min (P < 0.01) time points, compared to the unfiltered controls (Fig. 3A). We used the 60 min secretion values (Fig. 3B) to compare fatty acid-induced CCK secretion with fatty acid concentration after filtration.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 3. The effect of filtering dodecanoic acid on CCK secretion in STC-1 cells A, CCK secretion from STC-1 cells induced by dodecanoic acid filtrates. Dodecanoic acid solution (500 µM) was filtered through PTFE filters with pore sizes of 1.2 µm ( | ||
CCK secretion in STC-1 cells is caused by fatty acid aggregates which are larger than the monomeric size of the fatty acid and less than 0.45 µm
Using radiolabelled fatty acids, it was possible to examine simultaneously the effect of filtration on CCK secretion and fatty acid concentration. Figure 4 shows the relationship between filter pore size, fatty acid concentration and CCK secretion (normalised to the amount of CCK secreted in response to unfiltered fatty acids) for decanoic acid (Fig. 4A), dodecanoic acid (Fig. 4B) and tetradecanoic acid (Fig. 4C). The concentration of each fatty acid used was that which gave maximum CCK secretion in STC-1 cells. Filtering these solutions through a pore size of 1.2 µm greatly reduced the concentration of fatty acids in all solutions. The percentage of fatty acid removed by filtration increased with chain length, being 68.6 ± 0.3 % for decanoic, 88.4 ± 0.2 % for dodecanoic and 94.9 ± 0.2 % for tetradecanoic acid. The concentration of each fatty acid after filtration through any of the filters was always below that at which fatty acid aggregates were observed when the solution was freshly prepared (shown as a dashed line on Fig. 4). These data indicated that, in fatty acid suspensions which induced CCK secretion, at least two-thirds of the fatty acid was present as aggregates which could not pass through a 1.2 µm diameter filter pore.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 4. The relationship between fatty acid concentration and CCK secretion in STC-1 cells The effect of filtering decanoic (1000 µM, A), dodecanoic (500 µM, B) and tetradecanoic (500 µM, C) acid on CCK secretion ( | ||
Despite the removal of at least 68 % of the fatty acid by the 1.2 µm filter, STC-1 cells still responded by secreting CCK. For dodecanoic and tetradecanoic acid, the amount of CCK secreted compared with the unfiltered control was only significantly reduced (P < 0.05) by the 0.2 µm pore filter. In the case of tetradecanoic acid, CCK secretion appeared to be reduced by the 1.2 µm filtrate, but this was not significant. For decanoic acid the reduction in CCK secretion after filtration was insignificant for all pore sizes. More importantly, when a fresh solution of fatty acid was prepared at an equivalent concentration to that in the fatty acid filtrates, it never induced CCK secretion. While such fatty acid solutions had not been filtered, it was known that aggregates had not formed because the concentration of the fatty acid concerned was below that at which aggregation was observed to occur. Since the size of individual fatty acid molecules was much less than the diameter of the filter pore sizes used in these experiments, the value for these data points was set to just above zero on the horizontal axis of Fig. 4 (labelled 'FM'). Together these data suggested that STC-1 cells did not respond to fatty acid aggregates which are greater than 1.2 µm. Furthermore, for all fatty acids tested, over 50 % of the CCK response was mediated by particles which were less than 0.45 µm.
Filtering fatty acid solutions reduces the rise in [Ca2+]i in STC-1 cells
In our previous study (Sidhu et al. 2000) we showed that fatty acid-induced CCK secretion in STC-1 cells is mediated largely by an influx of extracellular Ca2+ through L-type Ca2+ channels, with stores also contributing a component to the observed rise in [Ca2+]i. Therefore, we examined the rise in [Ca2+]i induced by filtered fatty acid solutions. Figure 5A shows a typical [Ca2+]i response of STC-1 cells exposed to first unfiltered then filtered 500 µM dodecanoic acid solutions for 3 min. While the filtered solution still induced a rise in [Ca2+]i, the peak response was reduced. Exposing STC-1 cells to these solutions in reverse order gave similar results (Fig. 5B). Similar data were obtained from filtered decanoic and tetradecanoic acid solutions. The data for all three fatty acids are summarised in Fig. 6. The overall response patterns were similar to those observed for CCK secretion (Fig. 4). Thus, while filtration through a 1.2 µm filter removed over 68 % of the fatty acid under study (as described above), the rise in [Ca2+]i was still over 50 % of that observed with an unfiltered solution. Furthermore, no significant reduction was observed in the [Ca2+]i response when 1.2 µm and 0.45 µm filtrates were compared. However, filtrates obtained from a 0.2 µm filter showed a further reduction in the [Ca2+]i peak which was significant for dodecanoic and tetradecanoic acid (Fig. 6). Finally, freshly prepared solutions of each fatty acid at the same concentration as the filtrates did not induce a change in [Ca2+]i. Therefore, as with CCK secretion, the rise in [Ca2+]i in STC-1 cells appears to be caused by fatty acid aggregates which are greater than the monomeric size of the fatty acid but probably less than 0.45 µm.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 5. The effect of filtered and unfiltered dodecanoic acid solutions on [Ca2+]i in STC-1 cells Intracellular [Ca2+] was measured using dual excitation fluorescence microscopy (see Methods). Two solutions were presented: unfiltered dodecanoic acid (C12) and dodecanoic acid (500 µM) which had been passed through a 0.2 µm pore size PTFE filter (C12 (filtered)). The order of presentation was: unfiltered dodecanoic acid, 4 min recovery period, filtered dodecanoic acid (A) or filtered dodecanoic acid, 4 min recovery period, unfiltered dodecanoic acid (B). The traces represent the means ± S.E.M. of 19 cells and each trace is representative of 2 other experiments. | ||
 |
View larger version [in this window] [in a new window] |
|
|
Figure 6. The effect of filtration on fatty acid aggregation and Ca2+ mobilisation in STC-1 cells The effect on fatty acid-induced changes in [Ca2+]i of filtering decanoic (1000 µM, A), dodecanoic (500 µM, B) and tetradecanoic (500 µM, C) acid solutions through filters of different pore size is shown on the right ordinate ( | ||
The ability of fatty acid filtrates to induce a rise in [Ca2+]i decays with time
It is important to note that fatty acid filtrates could evoke responses in STC-1 cells whereas freshly prepared solutions containing fatty acids at the same concentration as in the filtrates could not. A simple model for this observation is that the filtrates contained a small number of aggregates which passed through the filter. If this was so, we hypothesised that the ability of a filtrate to induce a [Ca2+]i response would decline with time because any insoluble fatty acid aggregates which had passed through the filter would disaggregate and dissolve as they returned to equilibrium with the aqueous solvent. In contrast, the equilibrium for a supersturated (500-1000 µM) fatty acid solution will be a colloidal suspension of large fatty acid aggregates, which we have shown do not play a role in stimulating CCK secretion or Ca2+ mobilisation (Fig. 4 and Fig. 6). Sonication of either a freshly filtered or unfiltered fatty acid solution will lead to the mechanical dissipation of fatty acid aggregates. For unfiltered solutions, this dissipation is moving the system away from equilibrium, reducing the aggregate size to within the range needed to stimulate ST-C cells. This means that the ability of a supersaturated solution of fatty acid to induce Ca2+ mobilisation should decline with time after initial sonication. Furthermore, re-sonication of such a solution should once again restore the cellular response. However, for freshly filtered solutions, sonication will move the solution towards equilibrium and hence should accelerate the loss of the filtrate's ability to induce changes in [Ca2+]i. Figure 7 demonstrates that the predictions of this model hold when compared with experimental data. The efficacy of the unfiltered fatty acid solutions decayed over 10 min (B, Fig. 7) and this could be restored by sonication (C). Likewise, the efficacy of 0.2 µm decanoic, dodecanoic and tetradecanoic acid filtrates (D) to induce a change in [Ca2+]i declined with time so that after 10 min none of the filtrates were able to elicit a significant change in [Ca2+]i (E). Furthermore, sonication of the filtrate immediately abolished its ability to induce a [Ca2+]i response (F).
 |
View larger version [in this window] [in a new window] |
|
|
Figure 7. Analysis of filtered fatty acid solutions with respect to agonist stability The effect of time and sonication on the efficacy of filtered decanoic (A), dodecanoic (B) and tetradecanoic (C) acids to induce a rise in [Ca2+]i in STC-1 cells. Key to treatment protocols: A, freshly prepared unfiltered fatty acid solution; B, unfiltered fatty acid solution left for 10 min; C, unfiltered fatty acid solution left for 10 min and then re-sonicated; D, freshly prepared 0.2 µm (PTFE filter pore size) fatty acid filtrate; E, 0.2 µm fatty acid filtrate left for 10 min; F, 0.2 µm fatty acid filtrate, sonicated; G, freshly prepared fatty acid solution of the same concentration as the fatty acid filtrate. The concentration of unfiltered fatty acid was 500 µM for dodecanoic and tetradecanoic acid, and 1000 µM for decanoic acid. Each bar represents the means ± S.E.M. of 3 experiments. | ||
Removal of divalent cations from fatty acid solutions eliminates release of Ca2+ from intracellular stores
We showed previously (Sidhu et al. 2000) that while the fatty acid-induced rise in [Ca2+]i was mediated largely by an influx of extracellular Ca2+, release from intracellular stores also played a role. Given that the solubility of fatty acids is greatly altered by the presence of divalent cations (Fig. 2), we examined whether the removal of the other divalent cation (Mg2+) from our medium would eliminate Ca2+ release from stores when STC-1 cells are exposed to dodecanoic acid (Fig. 8). As expected, this was indeed the case: no change in [Ca2+]i was observed in the absence of Mg2+. This result suggested that both aspects of Ca2+ mobilisation were driven by a signal transduction mechanism that sensed insoluble fatty acid aggregates rather than fatty acid in solution and that this was upstream of all [Ca2+]i changes.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 8. The effect of divalent cations on dodecanoic acid-induced Ca2+ mobilisation The effect on [Ca2+]i in STC-1 cells of removing Ca2+ or Ca2+ and Mg2+ from the extracellular solution is shown in A and B, respectively. Cells were exposed to dodecanoic acid (C12, first horizontal bar) in the presence of Ca2+ (1.2 mM) and Mg2+ (1.2 mM) for 2 min. The cells were then washed twice and either a Ca2+-free (buffered with 0.2 mM EGTA) or a Ca2+- and Mg2+-free (buffered with 0.2 mM EDTA) solution was then applied (lower horizontal bars). While the cells were in this solution, dodecanoic acid was re-applied (C12, second upper horizontal bars). The traces represent the means ± S.E.M. of 17 and 20 cells for A and B, respectively. These traces are representative of at least 3 other experiments. | ||
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The filtration data described here led to the hypothesis that STC-1 cells secrete CCK following stimulation by fatty acid aggregates, rather than by free fatty acids in solution. From this it follows that only fatty acids of a given chain length (> 10 carbon atoms), which form aggregates in solution, can stimulate CCK secretion. The following points support this hypothesis.
First, when dodecanoic acid was filtered through a PTFE filter of 1.2 µm pore size the filtrate evoked CCK secretion from STC-1 cells but there was a significant reduction in CCK secretion in response to dodecanoic acid that had been filtered through a 0.2 µm filter. Second, dodecanoic acid-induced Ca2+ mobilisation was also reduced to a much greater extent when STC-1 cells were exposed to the 0.2 µm filtrate and this reduction was greater than that observed with the 0.45 or 1.2 µm filtrates. Third, similar reductions in CCK secretion and Ca2+ mobilisation were observed following filtration of both decanoic and tetradecanoic acid solutions, although decanoic acid-induced CCK secretion appeared to be less sensitive to filtration. Fourth, the radiolabelling experiments indicated that the concentrations of C10, C12 and C14 fatty acids in the filtrates were always below the concentration at which fatty acid aggregates appear. This finding is important because exposure of STC-1 cells to freshly prepared fatty acid solutions at concentrations commensurate with those obtained after filtration did not result in CCK secretion or a rise in [Ca2+]i. In agreement with this observation is the finding that the effects of filtered fatty acid solutions declined with time, presumably due to the progressive disaggregation of the few fatty acid aggregates which passed through the filter, as the fatty acid returned to equilibrium with its solvent. Unfortunately, our filtering experiments did not produce enough filtered solution to test whether the light scattering from such a solution decayed at a rate which was consistent with that observed for the biological response. However, we were able to measure the light scattering change which occurred when a sonicated dodecanoic acid solution was diluted by a factor of 10. Under these experimental conditions, fatty acid aggregates were detectable 10 min after dilution but not 1 hour after(data not shown).
Finally, the removal of Ca2+ from the extracellular solution eliminated the fatty acid-induced Ca2+ influx into cells, leaving the stores response intact. Removal of the second divalent cation, Mg2+, increased the solubility of all the fatty acids (Fig. 2B) and eliminated the stores response. A reasonable interpretation of these observations is that stores release and Ca2+ influx are part of the same signal transduction pathway which is activated by the presence of fatty acid aggregates. The removal of extracellular Ca2+ obviously prevents Ca2+ influx but does not disrupt other signalling events, one such event being stores release (Sidhu et al. 2000). However, removal of the second divalent cation results in the dissolution of fatty acid aggregates which are necessary for activation of the signal transduction pathway. In this instance all events associated with fatty acid signalling are abolished.
While our data indicated that STC-1 cells responded to fatty acid aggregates, the exact size of the aggregate was difficult to determine. It is dangerous to assume that the filters necessarily removed only fatty acid aggregates which were greater than the stated pore size of the filter. Fatty acids are notorious for binding to certain materials, a problem that was clearly evident in our preliminary studies with nitrocellulose filters (unpublished data). Likewise, Fig. 1 indicates that even PTFE filters removed 3-8 % of a fatty acid label that was completely dissolved in solution (as indicated by the data points on the ordinate, which represent the fatty acid label on its own). This finding also explained why the PTFE filters could lower the overall fatty acid concentration to below the limit of solubility for each fatty acid. In theory, if the fatty acid was in equilibrium with its solvent (a condition which probably was not met in our experiments), then filtering the solution should only lower the fatty acid concentration to above or near its solubility limit. The fact that the ability of the filtrates to produce a cellular response decayed with time further confirmed the fact that the filters removed fatty acid that was in solution, as this could not occur if the fatty acid solution was still saturated. However, the decay experiments also clearly demonstrated that it could not be this soluble fatty acid which induced the cellular response because, if it was, the filtered solutions would be stable over time. It is also possible that the size of the aggregate that is responsible for fatty acid-induced CCK secretion is much larger than is suggested by the filter pore sizes. If, for example, the aggregates are deformable or the pores, themselves, become stretched under pressure, aggregates which are greater than the pore size may pass through the filter and it could be this species which was responsible for the biological response observed in STC-1 cells. In summary, while the data suggest that fatty acids aggregates induced CCK secretion and [Ca2+]i mobilisation in STC-1 cells, the actual size of the aggregates involved remains uncertain.
Another important issue is how closely these experiments mimic what occurs at the apical surface of enteroendocrine cells in the gut lumen under physiological conditions. It is obviously very difficult to reproduce such conditions in a cultured cell line. One important difference between our experiment system and in vivo physiology is the absence of bile acids. We tried to address this issue by including bile acids in our preparations. If the particle hypothesis is correct, the model would predict that bile acids will decrease the response of STC-1 cells to fatty acids because they will solubilise the fatty acid aggregates. Our data are not inconsistent with this model. However, these experiments were unsatisfactory because of inter-experimental variability in the amount of CCK secreted and a concern that the bile acids may be damaging cell membranes. The only satisfactory way to assess the role of bile acids and the physiological relevance of our findings is to test the effects of fatty acid aggregates in vivo.
Although the hypothesis described above is perhaps surprising, the data presented in this paper are not in conflict with the current literature. It is clear that fatty acids in the gut lumen induce CCK release from intestinal I-cells (Liddle, 1995) and that this secretory response is dependent on the acyl chain length of the fatty acid (Shintani et al. 1995; Mclaughlin et al. 1998, 1999; Sidhu et al. 2000). The ability of enteroendocrine cells to detect a change in chain length of just one carbon atom has led to the suggestion that they may possess a transduction mechanism which is sensitive to the chemical structure of fatty acids. However, for such an explanation to be correct, only a few fatty acids in a restricted chain length range would be expected to induce CCK release from enteroendocrine cells. This is in contrast to what is actually observed, namely that all fatty acids above a chain length threshold induce CCK release to the same extent. This observation of a chain length threshold is more consistent with the concept of a change in physical property in fatty acids which occurs as the number of carbon atoms increases. One important physical property of fatty acids that varies with chain length is solubility in aqueous solutions, which decreases steeply as the acyl chain length increases (Vorum et al. 1992). In our studies, this was demonstrated by the fact that two fatty acids, differing in length by just one carbon, could be prepared at equimolar concentrations such that one was completely dissolved while the other formed a colloidal suspension (see Fig. 2).
Two recent papers also suggest that it is the physical properties of fatty acids which determine their efficacy at inducing CCK secretion. In one study on humans it was found that the saturation of the fatty acid was an important determinant of how potent it was at inducing CCK secretion and the onset of satiety (French et al. 2000). These authors concluded that a change in fatty acid saturation probably influenced a physical property (such as melting point) and it is this property that is sensed by the CCK signal transduction mechanism. In the second study, CCK secretion in rats was suggested to depend on chylomicron formation (Raybould et al. 1998), a conclusion which was based on the use of a surfactant (Pluronic L-81) perfused into the intestine with the test lipids. It would be interesting to know the effect of this agent on the formation of insoluble fatty acid aggregates.
Our conclusion that enteroendocrine cells respond to fatty acid aggregates rather than free fatty acids suggests that the signal transduction pathway begins at the cell surface. While we have previously observed that fatty acids rapidly permeate the cell membrane (Sidhu et al. 2000) it is hard to envisage that fatty acids, once they have crossed the cell membrane (by either simple diffusion or facilitated transport), would re-form into insoluble aggregates because, in theory, they should be bound by various proteins including fatty acid binding protein (FABP) (Kaikaus et al. 1990; Coe & Bernlohr, 1998). Also, in our studies on fatty acid permeation, we used concentrations which were below the solubility limit of the fatty acid. At these concentrations there was no physiological response (i.e. CCK secretion or elevation of [Ca2+]i) despite the fact that the cell was heavily loaded with the fatty acid in question. One important caveat is that endocytosis of fatty acid aggregates may cause a rise in [Ca2+]i and CCK secretion.
However STC-1 cells sense fatty acid aggregates, the data in this paper suggest there is not a specific fatty acid receptor. Rather, it appears that the cell detects a variety of particulate substances of a given size. Such a function would be consistent with the requirements of fat digestion (achieved via CCK secretion) which are presumably solubilisation of fatty acid aggregates. It is also interesting to note that the presence of high concentrations of lipid in the duodenum produces feelings of nausea (Baldwin et al. 1998; Greenough et al. 1998). This effect can be partially blocked by the CCK-A antagonist loxiglumide and mimicked by exogenous CCK (Feinle et al. 1996; Baldwin et al. 1998). These findings suggest that in some circumstances CCK acts as a negative stimulus to limit ingestion of certain substances. If this is so, then the prolonged presence in the gut of an insoluble particulate substance might stimulate I-cells to produce elevated levels of CCK which will induce nausea and so evoke behavioural changes which prevent ingestion of the substance a second time. In conclusion, our studies suggest that, at least in STC-1 cells, fatty acid-induced CCK secretion is not evoked by free fatty acids but rather by insoluble fatty acid aggregates.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
| BALDWIN, B. A., PARROTT, R. F. & EBENEZER, I. S. (1998). Food for thought: a critique on the hypothesis that endogenous cholecystokinin acts as a physiological satiety factor. Progress in Neurobiology 55, 477-507 | [Medline] |
| COE, N. R. & BERNLOHR, D. A. (1998). Physiological properties and functions of intracellular fatty acid-binding proteins. Biochimica et Biophysica Acta 1391, 287-306 | [Medline] |
| DRUCKER, D. J., JIN, T., ASA, S. L., YOUNG, T. A. & BRUBAKER, P. L. (1994). Activation of proglucagon gene transcription by protein kinase-A in a novel mouse enteroendocrine cell line. Molecular Endocrinology 8, 1646-1655 | [Medline] |
| FEINLE, C., D'AMATO, M. & READ, N. W. (1996). Cholecystokinin-A receptors modulate gastric sensory and motor responses to gastric distension and duodenal lipid. Gastroenterology 110, 1379-1385 | [Abstract] |
| FRENCH, S. J., CONLON, C. A., MUTUMA, S. T., ARNOLD, M., READ, N. W., MEIJER, G. & FRANCIS, J. (2000). The effects of instestinal infusion of long-chain fatty acids on food intake in humans. Gastroenterology 119, 943-948 | [Abstract/Full Text] |
| GREENOUGH, A., COLE, G., LEWIS, J., LOCKTON, A. & BLUNDELL, J. (1998). Untangling the effects of hunger, anxiety, and nausea on energy intake during intravenous cholecystokinin octapeptide (CCK-8) infusion. Physiology and Behaviour 65, 303-310 | |
| GUIMBAUD, R., MOREAU, J. A., BOUISSON, M., DURAND, S., ESCOURROU, J., VAYSSE, N. & FREXINOS, J. (1997). Intraduodenal free fatty acids rather than triglycerides are responsible for the release of CCK in humans. Pancreas 14, 76-82 | [Medline] |
| HIGHAM, A., VAILLANT, C., YEGEN, B., THOMPSON, D. G. & DOCKRAY, G. J. (1997). Relation between cholecystokinin and antral innervation in the control of gastric emptying in the rat. Gut 41, 24-32 | [Abstract] |
| HOPMAN, W. P. M., KERSTENS, P. J. S. M., JANSEN, J. B. M. J., ROSENBUSCH, G. & LAMERS, C. B. H. W. (1985). Effect of graded physiologic doses of cholecystokinin on gallbladder contraction measured by ultrasonography. Gastroenterology 89, 1242-1247 | [Abstract] |
| KAIKAUS, R. M., BASS, N. M. & OCKNER, R. K. (1990). Functions of fatty acid binding proteins. Experientia 46, 617-630 | [Medline] |
| LIDDLE, R. A. (1995). Regulation of cholecystokinin secretion by intraluminal releasing factors. American Journal of Physiology 269, G319-327 | [Medline] |
| LIDDLE, R. A. (1997). Cholecystokinin cells. Annual Review of Physiology 59, 221-242 | [Abstract/Full Text] |
| LIDDLE, R. A., MORITA, E. T., CONRAD, C. K. & WILLIAMS, J. A. (1986). Regulation of gastric emptying in humans by cholecystokinin. Journal of Clinical Investigation 72, 992-996 | |
| MCLAUGHLIN, J. T., LOMAX, R. B., HALL, L., DOCKRAY, G. J., THOMPSON, D. G. & WARHURST, G. (1998). Fatty acids stimulate cholecystokinin secretion via an acyl chain length-specific, Ca2+-dependent mechanism in the enteroendocrine cell line STC-1. Journal of Physiology 513, 11-18 | [Abstract/Full Text] |
| MCLAUGHLIN, J., LUCA, M. G., JONES, M. N., DOCKRAY, G. J. & THOMPSON, D. G. (1999). Fatty acid chain length determines CCK secretion and different effects on proximal and distal gastric motility. Gastroenterology 116, 46-53 | [Abstract/Full Text] |
| RAYBOULD, H. E., MEYER, J. H., TABRIZI, Y., LIDDLE, R. A. & TSO, P. (1998). Inhibition of gastric emptying in response to intestinal lipid is dependent on chylomicron formation. American Journal of Physiology 274, R1834-1838 | [Medline] |
| RINDI, G., GRANT, S. G. N., YIANGOU, Y., GHATEI, M. A., BLOOM, S. R., BATCH, V. L., SOLCIA, E. & POLAK, J. M. (1990). Development of neuroendocrine tumors in the gastrointestinal tract of transgenic mice. American Journal of Pathology 136, 1349-1363 | [Abstract] |
| SHINTANI, T., TAKAHASH, N., FUSHINKI, T. & SUGIMOTO, E. (1995). The recognition system of dietary fatty acids by the rat small intestinal cells. Biosciences Biotechnology Biochemistry 59, 1428-1432 | |
| SIDHU, S. S., CASE, R. M., WARHURST, G., THOMPSON, D. G. & BENSON, R. S. P. (2000). Fatty acid-induced cholecystokinin secretion and changes in intracellular Ca2+ in two enteroendocrine cell lines, STC-1 and GLUTag. Journal of Physiology 528, 165-176 | [Abstract/Full Text] |
| SMALL, D. M., PENKETT, S. A. & CHAPMAN, D. (1969). Studies on simple and mixed bile salt micelles by nuclear magnetic resonance spectroscopy. Biochimica et Biophysica Acta 176, 178-189 | [Medline] |
| SMITH, G. P. & GIBBS, J. (1994). Satiating effect of cholecystokinin. Annals of the New York Academy of Sciences 713, 236-241 | [Medline] |
| VORUM, H., BRODERSEN, R., KRAGH-HANSEN, U. & PEDERSEN, A. O. (1992). Solubility of long-chain fatty acids in phosphate buffer at pH 7. 4. Biochimica et Biophysica Acta 1126, 135-142 | [Medline] |
Acknowledgements
We would like to thank Professor Graham Dockray, University of Liverpool, for the use of the CCK antibody and Drs Annette Fillery-Travis and Martin Wickham of the Institute of Food Research, Norwich for assisting with light scatter measurements. This work was supported by a BBSRC grant (34/FO8146). Satjinder Sidhu was an MRC student.
This article has been cited by other articles:
|
|
 |
|
  T. Hira, A. C. Elliott, D. G. Thompson, R. M. Case, and J. T. McLaughlin Multiple Fatty Acid Sensing Mechanisms Operate in Enteroendocrine Cells: NOVEL EVIDENCE FOR DIRECT MOBILIZATION OF STORED CALCIUM BY CYTOSOLIC FATTY ACID J. Biol. Chem., June 18, 2004; 279(25): 26082 - 26089. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
), dodecanoic (
) and tetradecanoic (
) acid were filtered through PTFE filters (0.45 µm pore size). Each fatty acid preparation contained trace amounts of radiolabelled fatty acid (0.5 µCi 14C-labelled dodecanoic or tetradecanoic or 1 µCi 3H-labelled decanoic acid). The percentage of fatty acid removed by each filter was calculated by comparing the radioactivity of the solution before and after filtration. The data points located on the ordinate were obtained by filtering the radioactive fatty acid label on its own ( ~2 µM). Data points represent the mean ± S.E.M. of 4 experiments; error bars fall within the data symbols.
, solutions which were judged by filtration or Zeta sizer measurement to contain no fatty acid aggregates; 
, solutions which contained aggregates and elicited CCK secretion; 
, left ordinate). CCK secretion has been normalised against the amount of CCK secreted from an unfiltered fatty acid solution and is labelled '% control response'. The filter pore size (µm) is plotted on the abscissa with 'FM', (before the first axis break) representing a solution of fatty acid which was freshly made up to the concentration obtained after filtration. Under these conditions there were no measurable fatty acid aggregates (the dashed lines represent the concentration at which the Zeta sizer detected fatty acid aggregates) and so the size of the fatty acid in these solutions is much less than the filter pore sizes on the abscissa. UF, unfiltered fatty acid solution (to the right of the axis break); its % control response is 100 % by definition. Fatty acid concentrations were determined as described in Fig. 1. For CCK secretion each data point represents the mean ± S.E.M. of 4, 9 and 4 experiments for decanoic, dodecanoic and tetradecanoic acid, respectively. For filtration work each data point represents the mean ± S.E.M. of 4 filtrates.