Detection of pharmacovigilance-related adverse events using electronic health records and automated methods.

Electronic health records (EHRs) are an important source of data for detection of adverse drug reactions (ADRs). However, adverse events are frequently due not to medications but to the patients' underlying conditions. Mining to detect ADRs from EHR data must account for confounders. We developed an automated method using natural-language processing (NLP) and a knowledge source to differentiate cases in which the patient's disease is responsible for the event rather than a drug. Our method was applied to 199,920 hospitalization records, concentrating on two serious ADRs: rhabdomyolysis (n = 687) and agranulocytosis (n = 772). Our method automatically identified 75% of the cases, those with disease etiology. The sensitivity and specificity were 93.8% (confidence interval: 88.9-96.7%) and 91.8% (confidence interval: 84.0-96.2%), respectively. The method resulted in considerable saving of time: for every 1 h spent in development, there was a saving of at least 20 h in manual review. The review of the remaining 25% of the cases therefore became more feasible, allowing us to identify the medications that had caused the ADRs.

Biclustering of adverse drug events in the FDA's spontaneous reporting system.

In this article, we present a new pharmacovigilance data mining technique based on the biclustering paradigm, which is designed to identify drug groups that share a common set of adverse events (AEs) in the spontaneous reporting system (SRS) of the US Food and Drug Administration (FDA). A taxonomy of biclusters is developed, revealing that a significant number of bona fide adverse drug event (ADE) biclusters have been identified. Statistical tests indicate that it is extremely unlikely that the bicluster structures thus discovered, as well as their content, could have arisen by mere chance. Some of the biclusters classified as indeterminate provide support for previously unrecognized and potentially novel ADEs. In addition, we demonstrate the potential importance of the proposed methodology in several important aspects of pharmacovigilance such as providing insight into the etiology of ADEs, facilitating the identification of novel ADEs, suggesting methods and a rationale for aggregating terminologies, highlighting areas of focus, and providing an exploratory tool for data mining.

Prostaglandin E2 inhibits Na-K-2Cl cotransport in medullary thick ascending limb cells.

Prostaglandin E2 (PGE2) is known to inhibit transepithelial Cl transport in medullary thick ascending limb (mTAL), but the mechanism of inhibition or the transport pathway affected has not been identified. We undertook this study to examine the effect of PGE2 on Na-K-2Cl cotransport in mouse mTAL cells in culture. In nanomolar concentrations, PGE2 inhibited the Na- and Cl-dependent, bumetanide-sensitive K influx by 45%, and this inhibition was also observed in the presence of 3 mM ouabain. Although PGE2 also inhibited ouabain-sensitive K flux, that inhibition was abolished in the presence of apical nystatin, suggesting that the pump inhibition was secondary to diminished Na entry into the cells. The effect of PGE2 was concentration dependent. Inhibition was observed at a concentration of < 1 nM, and half-maximal effect was observed at 2.5 nM. The effect of PGE2 was not mediated by an action on cytosolic Ca because cytosolic Ca was unchanged after the addition of PGE2. PGE2 reduced the maximal velocity for the cotransporter but had no effect on the affinity of the cotransporter for external Na, K, or Cl. Specific [3H]bumetanide binding was reduced in the presence of PGE2, suggesting that PGE2 affected bumetanide-sensitive K influx by downregulating the number of functioning Na-K-2Cl cotransporters. These results suggest that Na-K-2Cl cotransport in the mTAL cells may be under tonic inhibitory control of PGE2.

Expression of 5-HT1C receptors in transfected MDCK cells is functionally and anatomically asymmetric.

Madin-Darby canine kidney (MDCK) cells were transfected with the cDNA for the rat 5-HT1C receptor (pMV7-SR1c) using electroporation. Cells that survived G418 selection medium were loaded with indo-1 and run through a fluorescence-activated cell sorter (FACS); 10% responded to serotonin (5-HT) with an increase in intracellular Ca2+ concentration ([Ca2+]i). Responding cells were separated with the FACS, grown to confluence, and resorted two more times until a clone of 100% respondents was obtained (SR-MDCK). In SR-MDCK cells grown on porous filters, [Ca2+]i increased only when 5-HT was applied to the basolateral membrane (change in [Ca2+]i = 190 +/- 43 nM); there was no response of [Ca2+]i to apical application of 5-HT. The asymmetric response to 5-HT was likely due to targeting of 5-HT1C receptors exclusively to the basolateral membrane of SR-MDCK cells; 125I-labeled lysergic acid diethylamide binding sites, a marker of high-affinity 5-HT receptors, were located only in the basolateral membrane. These experiments demonstrate that epithelial cells can be stably transfected to express G protein-linked, calcium-mobilizing receptors and that the receptors may be targeted asymmetrically to specific domains of the plasma membrane.

Protein kinase C does not participate in carbachol's secretory action in T84 cells.

We investigated the role of protein kinase C (PKC) in mediating carbachol's stimulation of transepithelial Cl- secretion in T84 cells. Direct PKC activation with phorbol 12-myristate 13-acetate (PMA) stimulated transepithelial Cl- transport (measured as the short-circuit current), demonstrating that PKC could interact with the secretory apparatus. Carbachol stimulated PKC activity, suggesting that the enzyme might participate in the hormone's action. Diacylglycerol metabolism inhibitors (DMIs), known to augment hormone-stimulated increases in diacylglycerol levels, potentiated the short-circuit current response to carbachol. The effect of DMIs was not due to amplification of carbachol-induced increases in PKC activity, however; PKC activity during carbachol stimulation was no higher in the presence of DMIs than in their absence. Augmentation of carbachol's action by DMIs appeared to be due to the direct activation of PKC which, like PMA, stimulated the Cl- conductance of the apical membrane (GCl). The effects of DMIs and carbachol on GCl were additive. Carbachol itself stimulated GCl but not by activating PKC; staurosporine did not blunt the effect of carbachol on GCl. Nor did staurosporine reduce the effect of carbachol on transepithelial Cl- secretion. These observations demonstrate that PKC does not participate in the secretory action of carbachol in T84 cells and suggest that direct PKC activation with DMIs and PMA stimulates an apical pool of PKC that is not accessible to carbachol applied to the basolateral membrane.

Changes in luminal flow rate modulate basal and bradykinin-stimulated cell [Ca2+] in aortic endothelium.

Hemodynamic forces influence many endothelial cell functions. The coupling between hemodynamic forces and cell function could be mediated by mechano-sensitive ion channels present in the plasma membrane of endothelial cells. Because one of these channels is permeable to Ca2+, we tested whether hemodynamic forces influence endothelial cell Ca2+ ([Ca2+]i). Bovine aortic endothelial cells were grown inside cylindrical glass tubes, loaded with fura-2, and perfused at different pressures and flow rates on the stage of a fluorescence microscope. Decreasing flow from 110 to 2 ml.min-1 raised [Ca2+]i from 57 +/- 11 to 186 +/- 29 nM (mean +/- SEM, p less than 0.01) by increasing the entry of extracellular Ca2+ into the cytoplasm. Increasing flow from 2 to 110 ml.min-1 transiently decreased [Ca2+]i from 62 +/- 3 to 33 +/- 5 nM (p less than 0.01) apparently due to reduced Ca2+ entry and concomitant extrusion by the plasma membrane Ca(2+)-ATPase. The rise in [Ca2+]i induced by bradykinin was magnified during a decrease in flow; in control cells, 10(-7) M bradykinin increased [Ca2+]i by 162 +/- 26 nM, whereas [Ca2+]i increased 350 +/- 67 nM (p less than 0.05) in cells previously exposed to 110 ml.min-1. These observations suggest that flow-induced changes in [Ca2+]i might be a signal-transduction mechanism for endothelial functions responsive to hemodynamic forces and may also modulate the magnitude of hormonally mediated increases in [Ca2+]i.

Participation of protein kinase C in desensitization to bradykinin and to carbachol in MDCK cells.

To explore the possibility that protein kinase C (PKC) participates in desensitization to Ca(2+)-mobilizing hormones in MDCK cells, we measured intracellular free Ca2+ concentration ([Ca2+]i) using fura-2 and video microscopy. We first examined the response of MDCK cells grown on plastic dishes. Exposure of cells to bradykinin (BK) or to carbachol, followed by reexposure after washing off the hormone, revealed two features of hormone desensitization. First, the initial hormone-induced peak response of [Ca2+]i was transitory; [Ca2+]i returned to control levels despite continued presence of hormone. Second, cells remained refractory to hormone rechallenge for 5 min after washing off hormone; [Ca2+]i response on re-exposure was reduced 70% compared with initial hormone-stimulated peak. Subsequent experiments demonstrated involvement of PKC in both desensitization processes. Pretreatment with the phorbol ester, phorbol 12-myristate 13-acetate, significantly blunted initial response to BK and to carbachol by 70 and 86%, respectively. When hormone-stimulated C kinase activity was enhanced with the diglyceride lipase inhibitor, RG 80267, BK- and carbachol-induced increases in [Ca2+]i were blunted 50%. Pretreatment with sphingosine, an inhibitor of PKC, resulted in an amplification of initial hormone-stimulated increase in [Ca2+]i and restored the response to rechallenge. To examine the possible interaction between BK and carbachol,both of which use PKC to induce desensitization, we measured [Ca2+]i in cells grown as monolayers on permeable, collagen-coated supports. Both carbachol and BK induced desensitization to the other hormone (heterologous desensitization)provided that the two hormones were applied to the same side of the polarized monolayer (apical).(ABSTRACT TRUNCATED AT 250 WORDS)

Carbachol increases basolateral K+ conductance in T84 cells. Simultaneous measurements of cell [Ca] and gK explore calcium's role.

To explore the role of calcium in mediating the action of carbachol in chloride-secreting epithelia, we simultaneously measured intracellular free [Ca] ([Ca]i) and the potassium conductance (gK) of the basolateral membrane in T84 cells grown on collagen-coated filters. [Ca]i was measured with fura-2 and fluorescence microscopy and expressed as a relative value ([Ca]'i) normalized to control. To assess changes in basolateral gK, we measured the short circuit current (Isc) in the presence of luminal amphotericin and a transepithelial mucosa-to-serosa K+ gradient (Germann, W. J., M. E. Lowy, S. A. Ernst, and D. C. Dawson. 1986. J. Gen. Physiol. 88:237-251). Treatment of the monolayers with carbachol resulted in a parallel increase and then decrease in [Ca]'i and gK. The carbachol-induced changes in gK appeared to be dependent on the increase in [Ca]i because stimulation of gK was significantly diminished when the hormone-induced increase in [Ca]'i was blunted, either by loading the cells with BAPTA or by reducing the extracellular [Ca]. The carbachol-stimulated increase in gK appeared to be the direct result of the increase in steady-state [Ca]'i. The changes in gK and [Ca]'i after stimulation with carbachol were correlated and ionomycin also increased gK and [Ca]'i in a parallel manner. The carbachol-induced delta gK per delta[Ca]'i, however, was greater than that after ionomycin. Because ionomycin and carbachol appear to open the same channel, a conclusion based on inhibitor and selectivity experiments, carbachol may have a second action that amplifies the effect of calcium on gK.

Cell swelling increases intracellular free [Ca] in cultured toad bladder cells.

We examined the effect of cell swelling on intracellular free calcium concentration ([Ca]i) in cultured toad bladder cells (TB-M) grown as a polarized monolayer on collagen-coated filters. [Ca]i was measured by use of fura-2, fluorescence microscopy, and simple video imaging. In preliminary experiments we determined that reducing ionic strength by 15% had no effect on the Kd of fura-2, indicating that the dye could be used to examine the effects of cell swelling on [Ca]i. Reducing the osmolality of the serosal bathing medium by 15% caused [Ca]i to increase within 10 s from 97 +/- 9 to 354 +/- 88 nM (n = 5). By 2 min [Ca]i had declined to 163 +/- 22 nM. The increase in [Ca]i was not caused by a fall in Na concentration ([Na]) because isosmotic reduction of serosal [Na] did not result in an increase in [Ca]i. The swelling-induced increase in [Ca]i could be abolished by lowering serosal [Ca] to 200 nM and by addition of lanthanum. The calcium-channel blockers nitrendipine and verapamil also inhibited the swelling-induced increase in [Ca]i, although to different degrees. These experiments demonstrate that swelling of cultured toad bladder cells results in a significant increase in [Ca]i by enhancing the rate of calcium entry across the basolateral membrane, possibly through a calcium channel.

Role of calcium in mediating action of carbachol in T84 cells.

To examine the role of calcium in mediating carbachol's action in secretory epithelia, we simultaneously measured intracellular free [Ca] [( Ca]i) and transepithelial chloride transport in T84 cells grown on collagen-coated filters. [Ca]i was measured with fura-2 and fluorescence microscopy and expressed as a relative value [( Ca]'i) normalized to control. Chloride transport was measured as the short-circuit current (Isc) with a voltage clamp. Monolayers were pretreated with cyclic AMP to augment the response of Isc to carbachol, a procedure that did not qualitatively change the response of the monolayer to carbachol. The carbachol-induced changes in Isc appeared to be dependent on the increase in [Ca]i. First, carbachol caused both Isc and [Ca]'i to increase in parallel. Isc increased from 32 +/- 5 to 70 +/- 9 microA and then declined to 57 +/- 16 microA while [Ca]'i increased from 72 +/- 14 to 156 +/- 22 nM and then declined to 133 +/- 45 nM. Second, the carbachol-induced increases in Isc and [Ca]'i were correlated. The greater the hormone-stimulated rise in [Ca]'i, the higher the increase in Isc. Third, carbachol's stimulation of Isc was blunted by preventing the calcium spike with the cellular calcium buffer 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetra-acetic acid (BAPTA). Although the carbachol-induced increase in [Ca]'i appeared necessary for the increase in Isc, it was not clear if carbachol's action was solely the result of an increase in [Ca]'i. Increasing [Ca]'i with ionomycin, although causing Isc and [Ca]'i to increase in parallel, failed to increase Isc to the levels observed with carbachol. These experiments suggest that although the carbachol-induced increase in Isc is dependent on the increase in [Ca]i, the hormone may activate a second process that increases the sensitivity of the calcium-activated transport process to changes in [Ca]i.

Effect of vasopressin on intracellular [Ca] and Na transport in cultured toad bladder cells.

Intracellular free [Ca] [( Ca]i) and transepithelial sodium transport were measured simultaneously in cultured toad bladder cells grown on collagen-coated filters. [Ca]i was measured with fura-2 and fluorescence microscopy while sodium transport was measured as the short-circuit current (Isc) with a voltage clamp. Following stimulation with vasopressin [Ca]i and Isc rose in parallel to maximal values within 10 min. [Ca]i increased from 65 +/- 5 to 123 +/- 12 nM and Isc, from 11 +/- 3 to 25 +/- 6 microA (n = 4). The vasopressin-induced rise in [Ca]i correlated significantly with the increase in Isc, suggesting that the rise in [Ca]i might be necessary for the increase in Isc. If so, then adenosine 3',5'-cyclic monophosphate (cAMP), which mimics the natriferic action of vasopressin, should also increase [Ca]i. Although cAMP increased [Ca]i to a peak value of 32 +/- 13% (P less than 0.05) above control at 10 min, the rise in Isc did not parallel the increase in [Ca]i. Isc peaked instead at 20 min, rising to 114 +/- 25% (P less than 0.05) over control, during which time [Ca]i returned to base line. This result suggested that a steady state increase in [Ca]i was not necessary for the natriferic action of cAMP. This notion was confirmed in experiments in which the vasopressin-induced increase in [Ca]i was prevented by bathing the tissue in a low-[Ca] buffer. Under these conditions, Isc increased 37 +/- 9% above control (P less than 0.05, n = 4) even though [Ca]i remained largely unchanged. Our results suggest that although vasopressin increases [Ca]i in toad bladder cells, the rise in [Ca]i does not seem to play a role in the natriferic response. These experiments also demonstrate the utility of making simultaneous measurements of ion transport and [Ca]i, which allow direct examination of calcium's role in mediating ion transport.

Isoproterenol and cyclic AMP increase intracellular free [Ca] in MDCK cells.

We examined the relationship between cyclic AMP and the intracellular free calcium concentration ([Ca]i) in MDCK cells, a hormonally responsive and chloride-secreting cell line. We measured [Ca]i using the calcium-sensitive dye fura-2, fluorescence microscopy, and a silicon intensifier target camera to amplify the signal. Isoproterenol, known to stimulate chloride transport via cyclic AMP, increased [Ca]i from 112 +/- 17 to 373 +/- 53 nM. The rise appeared due to an increase in cyclic AMP: isobutylmethylxanthine enhanced the effect of a submaximal dose of isoproterenol on [Ca]i, the cyclic AMP analogue 8-bromo-cyclic AMP caused [Ca]i to increase from 100 +/- 8 to 278 +/- 40 nM, and direct activation of adenylate cyclase with forskolin increased [Ca]i from 192 +/- 29 to 279 +/- 35 nM. The rise in [Ca]i after cyclic AMP may be due to an influx of calcium from the outside: the cyclic AMP-induced increase in [Ca]i was prevented either by lowering extracellular [Ca] or by addition of 1 mM lanthanum. The mechanism by which calcium enters the cell may not be a calcium channel because neither verapamil nor nitrendipine prevented cyclic AMP from increasing [Ca]i. Cyclic AMP also does not appear to act directly on sodium-calcium exchange. [Ca]i increased as well in low [Na] as in high [Na] following addition of the nucleotide. Thus, isoproterenol, acting through an increase in [cyclic AMP] causes an increase in [Ca]i in MDCK cells. The source and route of entry of calcium into the cytoplasm remain uncertain.

Role of intracellular calcium in cellular volume regulation.

We investigated the role of intracellular calcium in epithelial cell volume regulation using cells isolated from the toad urinary bladder. A suspension of cells was prepared by treatment of the bladder with collagenase followed by ethyleneglycol-bis(beta-aminoethylether)-N,N'-tetraacetic acid. The cells retained their ion-transporting capabilities: ouabain (1 mM) and amiloride (10 microM) inhibited cellular uptake of 86Rb and 22Na, respectively. Using a Coulter counter to measure cellular volume, we found that we could swell cells either by reducing the extracellular osmolality or by adding the permeant solute urea (45 mM) isosmotically. Under both conditions, cells first swelled and then returned to their base-line volume, in spite of the continued presence of the stimulus to swell. Volume regulation was inhibited when cells were swelled at low extracellular [Ca] (100 nM) and was retarded in cells preloaded with the calcium buffer quin 2. Swelling increased the intracellular free calcium concentration ([Ca]i), as measured by quin 2 fluorescence: [Ca]i increased 35 +/- 9 nM (n = 6) after hypotonic swelling and 42 +/- 3 nM (n = 3) after urea swelling. Reducing extracellular [Ca] to less than 100 nM prevented the swelling-induced increase in [Ca]i, suggesting that the source of the increase in [Ca]i was extracellular. This result was confirmed in measurements of cellular uptake of 45Ca: the rate of uptake was significantly higher in swollen cells compared with control (1.1 +/- 0.2 vs. 0.4 +/- 0.1 fmol . cell-1 X 5 min-1). Our experiments provide the first demonstration that cellular swelling increases [Ca]i. This increase is likely to play a critical role in cellular volume regulation.

Does calcium couple the apical and basolateral membrane permeabilities in epithelia?

Tight epithelial cells actively transport sodium against steep electrochemical gradients. To maintain a constant internal ionic content and volume, they must continuously adjust the passive cation permeabilities of their membranes as the rate of transport varies. There is evidence suggesting that changes in cell calcium may accomplish this task. An increase in cell calcium reduces the luminal sodium permeability and increases basolateral potassium permeability. There is basolateral sodium-calcium exchange through which changes in the rate of sodium transport, reflected in the cell sodium activity, are translated into changes in cell calcium. To demonstrate that cell calcium couples the permeability of the cell membrane requires obtaining measurements of cell calcium activity under physiologically relevant conditions, and, to date, there are no measurements during spontaneous changes in the rate of transport. However, there are measurements following ouabain inhibition of the pump indicating that the increase is sufficient to account for the reduction in luminal sodium permeability observed in intact tissues.

Calcium reduces the sodium permeability of luminal membrane vesicles from toad bladder. Studies using a fast-reaction apparatus.

Regulation of the sodium permeability of the luminal membrane is the major mechanism by which the net rate of sodium transport across tight epithelia is varied. Previous evidence has suggested that the permeability of the luminal membrane might be regulated by changes in intracellular sodium or calcium activities. To test this directly, we isolated a fraction of the plasma membrane from the toad urinary bladder, which contains a fast, amiloride-sensitive sodium flux with characteristics similar to those of the native luminal membrane. Using a flow-quench apparatus to measure the initial rate of sodium efflux from these vesicles in the millisecond time range, we have demonstrated that the isotope exchange permeability of these vesicles is very sensitive to calcium. Calcium reduces the sodium permeability, and the half-maximal inhibitory concentration is 0.5 microM, well within the range of calcium activity found in cells. Also, the permeability of the luminal membrane vesicles is little affected by the ambient sodium concentration. These results, when taken together with studies on whole tissue, suggest that cell calcium may be an important regulator of transepithelial sodium transport by its effect on luminal sodium permeability. The effect of cell sodium on permeability may be mediated by calcium rather than by sodium itself.

Regulation of the sodium permeability of the luminal border of toad bladder by intracellular sodium and calcium: role of sodium-calcium exchange in the basolateral membrane.

Sodium movement across the luminal membrane of the toad bladder is the rate-limiting step for active transepithelial transport. Recent studies suggest that changes in intracellular sodium regulate the Na permeability of the luminal border, either directly or indirectly via increases in cell calcium induced by the high intracellular sodium. To test these proposals, we measured Na movement across the luminal membrane (th Na influx) and found that it is reduced when intracellular Na is increased by ouabain or by removal of external potassium. Removal of serosal sodium also reduced the influx, suggesting that the Na gradient across the serosal border rather than the cell Na concentration is the critical factor. Because in tissues such as muscle and nerve a steep transmembrane sodium gradient is necessary to maintain low cytosolic calcium, it is possible that a reduction in the sodium gradient in the toad bladder reduces luminal permeability by increasing the cell calcium activity. We found that the inhibition of the influx by ouabain or low serosal Na was prevented, in part, by removal of serosal calcium. To test for the existence of a sodium-calcium exchanger, we studied calcium transport in isolated basolateral membrane vesicles and found that calcium uptake was proportional to the outward directed sodium gradient. Uptake was not the result of a sodium diffusion potential. Calcium efflux from preloaded vesicles was accelerated by an inward directed sodium gradient. Preliminary kinetic analysis showed that the sodium gradient changes the Vmax but not the Km of calcium transport. These results suggest that the effect of intracellular sodium on the luminal sodium permeability is due to changes in intracellular calcium.

Phenformin-associated pancreatitis.

Although phenformin has been previously reported to be associated with acute pancreatitis, little emphasis of this association has been made in the literature. We report the case of a 70-year-old diabetic man who developed acute hemorrhagic pancreatitis and severe lactic acidosis while taking phenformin. The patient was not taking any other medications, nor did he have any of the known metabolic conditions associated with pancreatitis. We review the four previously published cases of patients who developed acute pancreatitis while taking phenformin. Three of those patients also developed lactic acidosis, a well-known complication of phenformin therapy. Although phenformin has been reported to increase the serum amylase activity and to alter the content of the pancreatic secretions in response to various stimuli, the manner in which the drug might cause acute pancreatitis remains completely unknown.