Abstract
Indomethacin has been suggested for the treatment of Alzheimer's disease (AD), but its use is limited by gastrointestinal and renal toxicity. To overcome this limitation, D-Pharm Ltd. (Rehovot, Israel) developed DP-155 (mixture of 1-steroyl and 1-palmitoyl-2-{4-[1-(p-chlorobenzoyl)-5-methoxy-2-methyl-3-indolyl acetamido] hexanoyl}-sn-glycero-3-phosophatidyl choline), a lecithin derivative of indomethacin. Safety was tested by daily oral administration of DP-155 or indomethacin to rats in a dose range of 0.007 to 0.28 mmol/kg. The prevalence of gastrointestinal ulceration was significantly lower (10-fold) for DP-155 than for indomethacin, and the ulcerations were delayed. Signs of renal toxicity, namely reduced urine output and increased urine N-acetyl glycosaminidase to creatinine ratio, were 5-fold lower for DP-155. Indomethacin, but not an equimolar dose of DP-155, reduced urine bicyclo-prostaglandin E2. An equimolar oral dose of DP-155 or indomethacin, administered every 4 h for 3 days, was equally efficacious in reducing the levels of Aβ42 in the brains of Tg2576 mice. Indomethacin was the principal metabolite of DP-155 in the serum. After DP-155 oral administration, indomethacin's half-life in the serum and the brain was 22 and 93 h, respectively, compared with 10 and 24 h following indomethacin oral administration. The brain to serum ratio was 3.5 times higher for DP-155 than indomethacin. This finding explains the efficacy of DP-155 in reducing Aβ42 brain levels, despite the low systemic blood concentrations of indomethacin derived from DP-155. In conclusion, compared with indomethacin, DP-155 has significantly lower toxicity in the gut and kidney while maintaining similar efficacy to indomethacin in lowering Aβ42 in the brains of Tg2576 mice. This superior safety profile highlights DP-155's potential as an improved indomethacin-based therapy for AD.
Indomethacin belongs to the nonsteroidal anti-inflammatory drug (NSAID) family. The use of NSAIDs was found to have an inverse correlation with the prevalence and severity of Alzheimer's disease (AD) in epidemiological and clinical studies (in t' Veld et al., 2001). Alzheimer's disease is characterized by three pathological hallmarks: extracellular amyloid β (Aβ) plaques, intracellular neurofibrillary tangles (composed of hyperphosphorylated Tau protein), and neuronal and synaptic loss. The pathology development is accompanied by marked inflammatory processes (microglial and astrocytic reaction) (McGeer and McGeer, 1999). Indomethacin inhibits Aβ plaque formation via γ-secretase inhibition, which is a cyclooxygenase (COX)-independent process (Weggen et al., 2001). In addition, NSAIDs have COX-dependent anti-inflammatory and neuroprotective effects (Halliday et al., 2000; Weggen et al., 2001).
Long-term use of indomethacin for AD is limited by significant gastrointestinal (GI) and renal toxicities (Tabet and Feldman, 2002). Gastrointestinal complications are the most common side effect of NSAIDs and result in hospitalization and death in 1.5 and 0.2%, respectively, of patients consuming NSAIDs chronically for rheumatoid arthritis (Fries et al., 1998). The typical toxic lesions in the GI tract comprise mucosal damage and erosion, resulting in ulceration and hemorrhage, especially in the stomach. The toxic mechanism in these parts of the GI tract has been attributed mainly to the inhibitory effect of indomethacin on the activity of cyclooxygenase and the resulting reduction in prostaglandin E2 (PGE2) levels. PGE2 has a protective effect on the GI mucosa through inhibition of gastric acid secretion, stimulation of bicarbonate and mucus secretion, vasodilatation, and induction of ion exchange through the mucosa. In addition, some NSAIDs, including indomethacin, have direct local toxic effects on the GI mucosa (Whittle, 1998). Gastrointestinal toxicity is increased in old age and with higher doses of NSAIDs, but it is somewhat surprisingly not correlated with the duration of treatment (Tseng and Wolfe, 1998).
After the GI tract, the kidney is the most frequent target of NSAIDs toxicity, reported to occur in 18% of NSAID-treated patients and to cause severe symptoms in 1% of this population (Stichtenoth and Frölich, 1998). The adverse effects are more severe and frequent in the elderly and in patients already suffering from renal failure. Renal toxicity is primarily due to the attenuation of the renal vessels' ability to dilate in reaction to hypovolemia, a response mediated by PGE2, which can lead to a reduced glomerular filtration rate (GFR). PGE2 is necessary for the optimal diuretic effect of furosemide, and reduced PGE2 can cause antidiuresis, sodium retention, and hyperkalemia. In the present study, the effect of indomethacin and DP-155 on renal PGE2 synthesis and excretion after treatment with furosemide and vasopressin was evaluated. Previous studies with this model showed that PGE2 can inhibit the vasoconstriction and antidiuretic effects of vasopressin and that indomethacin at 10 mg/kg decreases both urine PGE2 levels and the diuretic effect of furosemide in vasopressin-treated rats (Ayano et al., 1984b).
Different methods have been employed to avoid NSAIDs toxicity, including concurrent preventive medications, nitric oxide-releasing NSAIDs, enteric-coated preparations, and NSAIDs preassociated with zwitterionic phospholipids. All these methods have serious limitations (Tseng and Wolfe, 1998). Indomethacin inhibits both COX-1, an important constitutive “housekeeping” enzyme in the stomach, and COX-2, which is mostly inflammation-induced. Selective COX-2 inhibitors were thus considered very promising in terms of toxicity (MacDonald, 2000); however, this class of drugs was associated with serious cardiovascular complications, as well as some of the known GI and renal side effects (Brater et al., 2001; Psaty and Furberg, 2005).
DP-155 is a lipid-modified indomethacin and is comprised of an indomethacin molecule covalently attached via a linker to lecithin (Fig. 1). DP-155 was developed to overcome the limiting toxicity of indomethacin. DP-155 and indomethacin display similar analgesic and antipyretic effects in mice and rats, respectively (Duvdevani et al., 2003). The aim of this work was to compare some aspects of DP-155's safety profile and pharmacokinetics with those for indomethacin and to assess the ability of DP-155 to reduce brain Aβ levels. To that end, levels of SDS-soluble Aβ40 and Aβ42 were measured in the brains of Tg2576 hAPP transgenic mice following oral treatment with equimolar doses of DP-155 and indomethacin. The Tg2576 mouse model of AD expresses a mutation of the amyloid precursor protein (APP) known as the “Swedish mutation” (APPK670N,M671L). This mutation causes increases in secreted Aβ42 and Aβ40 in the brain, and the mice develop amyloid plaques and progressive cognitive deficits (Citron et al., 1994; Hsiao et al., 1996; Scheuner et al., 1996).
Materials and Methods
Safety Studies
All the studies were approved by the Institutional Animal Use and Care Committee, which works under the Israeli National Council for Experiments in Laboratory Animals, according to the Israeli law and the U.S. National Institutes of Health guidelines for the care and use of laboratory animals.
Safety studies were performed on male Sprague-Dawley rats (Harlan, Jerusalem, Israel), 7 to 9 weeks of age, weighing 200 to 250 g. The rats were acclimatized for 5 days before the study. The rats were divided into groups of four to five rats per dose tested. DP-155 was compared with commercially available indomethacin (Chemical Abstracts registry no. 53-89-4; Sigma-Aldrich, St. Louis, MO) and vehicle after single and repeated oral administrations. The vehicle for both drugs was a mixture of oil (4-10%) and water phases.
Gastrointestinal Safety
Single administration toxicity was assessed after 4 or 24 h. Repeated administration toxicity was assessed after three administrations at 24-h intervals. The rats were fasted overnight before the single administration 4-h observation study and were not fasted in the other longer studies.
The rats were observed daily for clinical signs of toxicity and were euthanized at the end of the observation period by either 200 mg/kg pentobarbitone i.p. or by CO2 asphyxiation. During euthanasia and under general anesthesia, blood was collected for serum biochemical analysis that included creatinine, urea, total bilirubin, alanine aminotransferase, aspartate aminotransferase, albumin, and total protein (analyzer, Cobas Mira, Roche Diagnostics, Basal, Switzerland; reagents, Sentinel Diagnostics, Milan, Italy). Serum sodium and potassium were also measured (Ektachem analyzer; Eastman Kodak, Rochester, NY). After euthanasia, all animals underwent necropsy that included inspection of the thoracic and abdominal organs. During necropsy, the stomach was opened along the major curvature, washed, and observed. On necropsy, 4 h after single administration, the stomach injury was scored as severe (wide erosion and hemorrhage or perforating ulcer), moderate (one to few subtle, focal lesions), or absent (no lesion or only hyperemia). The score was based on a scoring system used by Takagi and Okabe (1968). In the longer studies, the rats were fed, gastric ulcers were less visible, and this score could not be used. Selected organs were taken for histology including kidney, stomach, intestine, spleen, and liver. After the repeated administration study, the heart, lungs, and skin was also taken for histology.
For each compound, the dose was increased until pathology was observed. The dosing volume was limited to 10 ml/kg due to ethical considerations. In this volume, the highest soluble dose for DP-155 was 0.28 mmol/kg, and although this dose showed no toxicity 4 h after single administration, the dose was not increased further. In the single-administration 4-h observation study, the doses for indomethacin were 0.0028 to 0.028 mmol/kg (1-10 mg/kg, respectively) and for DP-155, 0.28 mmol/kg (270 mg/kg). In the single administration with 24-h observation, the doses for indomethacin were 0.007 to 0.14 mmol/kg and for DP-155, 0.028 to 0.28 mmol/kg. In the repeated administration study, the doses for indomethacin were 0.007 to 0.028 mmol/kg and for DP-155, 0.035 to 0.14 mmol/kg (the specific doses used in each study can be seen in Table 1).
Renal safety was studied in some of the dose groups in the GI toxicity studies: at the highest doses in the single administration 24-h observation study and in all the repeated administration studies. Urine output was measured during the last 24 h of the study, and creatinine clearance was determined based on the urine sample obtained within the last 4 h. In the repeated administration study, the urine N-acetyl glycosaminidase (NAG), a parameter for renal tubular damage, was also measured (analyzer, Cobas Mira; Roche Diagnostics; reagents, Kamiya Biomedical, Seattle, WA).
Renal safety was studied further according to a protocol described by Ayano et al. (1984b). Twenty rats were divided into four groups of five rats each: naive, vehicle, indomethacin, and DP-155 treated. On the night before the experiment, water was offered ad libitum, and food was withheld. On the study day, the rats were treated with 5 ml/kg vehicle, indomethacin, or DP-155 (0.028 mmol/kg) or not treated (naive). According to the model, 4 h post-treatment, all rats were treated with 2 U/kg vasopressin (Sigma-Aldrich) i.m. (for vasoconstriction and water reabsorption, mimicking hypovolemic stress), and water was withheld. Half an hour later, 5 mg/kg furosemide (Teva, Petah-Tikva, Israel) was administered i.m. (enhancing renal PGE2 synthesis and excretion), the rats were transferred to metabolic cages, and urine was collected for 1 h. After urine collection, the rats were euthanized with CO2. Urine samples were analyzed for the following parameters: volume, PGE2, creatinine, total protein, and NAG. Urine PGE2 was measured by quantifying the bicyclo-PGE2, a stable metabolite, using a commercial ELISA kit (Cayman 514531; Cayman Chemical, Ann Arbor, MI). Blood for creatinine levels was taken from the sino-orbital sinus. Necropsy was performed, and both kidneys were removed and put in formaldehyde for histology, prepared in a standard fashion, and stained with H&E and light green. Histology was performed by light microscopy. The kidneys were examined for casts and were scored as no casts, one to five casts, 5 to 50 casts, or >50 casts per kidney cross-section.
The differences between biochemical data of the study groups were analyzed by ANOVA followed by Student's t test, comparing each treatment group to vehicle. The differences in the prevalence of clinical and histological scores were analyzed by χ2 test. p < 0.05 was regarded as significant in all studies.
In Vivo Efficacy in an Alzheimer's Disease Transgenic Mouse Model
Female Tg2576 hAPP-transgenic mice at an age between 3 and 4 months received vehicle, indomethacin, or DP-155 in two doses, high and low. The high-dose study included 22 animals: seven vehicle-treated, eight DP-155-treated, and seven indomethacin-treated. Animals were administered indomethacin or DP-155 (0.14 mmol/kg/day) (50 or 135 mg/kg/day, respectively) for 3 days, in equally divided doses and volumes (0.111 ml/10 g/day) every 4 h. The low-dose study included 35 animals: 11 vehicle-treated, 12 DP-155-treated, and 12 indomethacin-treated. Animals were administered indomethacin or DP-155 (0.046 mmol/kg/day) (16.7 or 45 mg/kg/day, respectively) for 3 days, in equally divided doses and volumes (0.111 ml/10 g/day) every 12 h. Two hours after the final dose, animals were sacrificed. Brains were removed, weighed, and snap-frozen in liquid nitrogen, and the amount of brain Aβ was compared between treatment groups.
Preparation of SDS-Soluble Aβ Samples. SDS-soluble Aβ protein samples were prepared as described previously (Kawarabayashi et al., 2001). In brief, brain tissues were sonicated for 35 s in 2% SDS solution (1 ml/150 mg wet weight) containing protease inhibitors (two tablets of Complete Mini-Protease Inhibitor Cocktail in 50 ml of solution, Roche Diagnostics). After centrifugation at 100,000g for 1 h at 4°C, the supernatant containing SDS-soluble Aβ protein was pipetted into new tubes.
Sandwich ELISA Assay for Aβ Proteins. Levels of SDS-soluble Aβ40 and Aβ42 were measured by the sandwich ELISA method (BioSource International, Camarillo, CA). The SDS extracts were diluted 1:20 or 1:40 (for Aβ42 or Aβ40, respectively) in the diluent solution containing protease inhibitors. Levels of Aβ were expressed as nanograms per gram of wet weight. Differences between groups were assessed by one-way ANOVA followed by post hoc Student-Newman-Keuls test.
Pharmacokinetic Studies
Pharmacokinetics after oral administration were studied after dosing four rats with 0.01 mmol/kg DP-155 or indomethacin. Blood samples were drawn through an in-dwelling cannula implanted in the right jugular vein a day before the experiment. The oral bioavailability and the subsequent brain concentration of DP-155 or indomethacin were determined for each treatment group (18 rats/group). The dose was 0.01 mmol/kg, and the drugs were administered by oral gavage. The time points for sample collection were: 2, 4, 8, 12, and 24 h. At each time point, the rat was anesthetized with ether, a systemic blood sample was withdrawn from the aorta, and then the animal was sacrificed. The whole brain was obtained and stripped of its external vasculature and meninges. Blood and brain samples were analyzed for DP-155 and indomethacin concentrations.
To determine brain levels of indomethacin, each brain sample was accurately weighed (0.75-1 g/brain sample), and the drug was extracted (Politron tissue homogenizer, 25,000 rpm) into 5 ml of ethyl acetate. A high-performance liquid chromatography system (Waters 2695; Waters, Milford, MA) with a UV detector (Waters 2996) was used to determine indomethacin and DP-155 concentrations in plasma or brain samples by a method described previously (Ioffe et al., 2002). Pharmacokinetics analyses were performed using the WinNonlin professional software (version 4.0.1; Pharsight, Palo Alto, CA) and the noncompartmental analysis model.
Results
Safety
Gastrointestinal Toxicity. Evaluation at 4 h after single administration of indomethacin revealed that 0.0028 mmol/kg indomethacin caused no ulcers, and the lowest dose that yielded ulcers was 0.0084 mmol/kg (moderate ulcers with a prevalence significantly higher than vehicle, p = 0.028, Table 1). Administration of DP-155 revealed that 0.28 mmol/kg (the highest dose that could be dissolved in a volume of 10 ml) did not cause any gastric ulcers (Fig. 2). This dose is 100-fold higher compared with the first nontoxic dose of indomethacin. The indomethacin-induced ulcers showed a dose response in both prevalence and severity. Although some of the ulcers were macroscopically wide and associated with severe hemorrhage, the histological evaluation of the stomach revealed relatively superficial lesions limited to the level of the mucosal capillary bed. Serum biochemistry results were unremarkable, with no differences between indomethacin or DP-155 and vehicle.
Evaluation at 24 h after single administration of DP-155 showed that the highest dose yielding no pathology was 0.14 mmol/kg. The next dose, 0.28 mmol/kg, yielded stomach ulcers, melena, enteritis (Table 1), and low serum albumin (Fig. 3). After treating with escalating doses of indomethacin, only 0.007 mmol/kg showed no pathology, and the first dose that yielded a higher prevalence of melena and enteritis and lower serum albumin than vehicle was 0.014 mmol/kg, a dose 20-fold lower than DP-155's equivalent toxic dose (Table 2). Histology revealed superficial mucosal injury and bleeding to the depth of the capillary bed, and only a few rats showed deeper injury. Melena is darkening of the feces caused by digested blood pigments, indicating bleeding into the upper GI tract (stomach or small intestine). The acute reduction in serum albumin most probably results from protein loss to the GI tract, not from reduced protein production or absorption, and the fact that GI blood loss was observed in the same treatment groups supports this assumption. A similar pattern was observed in total serum protein levels (data not shown). The rest of the serum biochemistry parameters were within normal range for all treatment groups.
Toxicity was next evaluated daily following three repeated administrations at 24-h intervals. The lowest indomethacin dose in which any pathology was seen was 0.007 mmol/kg, with only microscopic signs of enteritis, no gross pathology findings, and normal serum albumin. The following higher dose of indomethacin, 0.014 mmol/kg, yielded microscopic and macroscopic pathology (Fig. 4) and low serum albumin (Fig. 5). Under DP-155 treatment, 0.035 mmol/kg showed no adverse effect, yet the following higher dose, 0.07 mmol/kg, yielded microscopic and macroscopic pathology with normal serum albumin. Only the highest dose, 0.14 mmol/kg, showed microscopic and macroscopic pathology and low serum albumin. This dose response indicates that indomethacin is 5- to 10-fold more toxic than DP-155 (Table 2) under these experimental conditions. The most prominent pathology observed was peritonitis, melena, and enteritis (Table 1; Fig. 4). Microscopically, with high doses of both drugs, perforating wounds leading to mesenteritis or peritonitis were evident. The similarity between the melena prevalence (Table 1) and the mean serum albumin (Fig. 5) supports the argument that the alterations were a result of blood loss to the GI tract. A similar pattern was observed with total serum protein levels (data not shown). The rest of the serum biochemistry parameters were within the normal range in all treatment groups.
Renal Toxicity. On necropsy 4 h after a single administration, all five rats in the 0.028 mmol/kg indomethacin treatment group had 5 to 50 tubular casts per kidney section. One rat in the 0.28 mmol/kg DP-155 group had one to five casts per kidney section. In the single administration study with necropsy after 24 h, tubular renal casts were observed in few treatment groups, including vehicle, and no treatment showed significantly higher prevalence compared with vehicle (data not shown). At the highest doses of indomethacin (0.14 mmol/kg) and DP-155 (0.28 mmol/kg), 24-h urine output was measured, and urine protein to creatinine ratio and creatinine clearance were determined from the urine sample obtained in the last 4 h. These results were compared with results obtained from four rats treated with vehicle. No statistically significant differences between the treatment groups were found in urine output; however, the results showed the following trend (mean ± S.D., milliliter per kilogram body weight): indomethacin (28.1 ± 7.42) < DP-155 (33.26 ± 11.66) < vehicle (50.35 ± 18.26). The creatinine clearance, a calculation reflecting the GFR, showed the same trend (mean ± S.D., milliliters per minute per kilogram): indomethacin (3.11 ± 1.09) < DP-155 (3.32 ± 0.88) < vehicle (3.90 ± 1.13). All measurements were markedly below the normal value of 7.2 ml/min/kg reported for Wistar rats (Edwards et al., 1971).
In the repeated administration study, urine output was measured in the last 24 h in all treatment groups. Indomethacin (0.028 mmol/kg) and 0.14 mmol/kg DP-155 showed markedly lower urine volume compared with vehicle (21.80 ± 8.15, 26.67 ± 10.09, and 49.62 ± 9.85 ml/kg/24 h ± S.D., respectively). Urine NAG/creatinine ratio was higher in the 0.028 mmol/kg indomethacin- and 0.14 mmol/kg DP-155-treated rats compared with vehicle (10.01 ± 6.46, 9.43 ± 3.28, and 4.47 ± 3.52 IU/mmol ± S.D., respectively); however, only DP-155 was significantly higher. Thus, the first dose of DP-155 with nephrotoxic signs was 5-fold higher compared with indomethacin (Table 2). Creatinine clearance was not different between the treatment and vehicle-treated groups. There was no significant difference in the prevalence of tubular casts in the kidneys between the different groups (data not shown).
Evaluating urine PGE2 after single administration of 0.028 mmol/kg indomethacin or DP-155 followed by vasopressin and furosemide revealed a clear significant reduction in PGE2 levels in the indomethacin-treated rats (28 ± 6.4 pg/ml) but no change in the DP-155-treated (51.2 ± 10.5 pg/ml) compared with vehicle-treated (55.6 ± 11.5 pg/ml) and naive (50.2 ± 13.9 pg/ml) rats. The indomethacin-treated rats had significantly (p = 0.03) lower urinary output (12.04 ± 4.02 ml/kg/h ± S.D.) compared with the DP-155 (19.18 ± 1.63), vehicle (17.17 ± 3.12), and naive (19.69 ± 3.07) groups. Creatinine clearance was higher in the indomethacin group (4.31 ± 1.51 ml/min/kg) compared with the DP-155 (3.36 ± 0.57 ml/min/kg), vehicle (2.49 ± 0.88 ml/min/kg), and naive (3.18 ± 1.73 ml/kg/h) groups. The NAG/creatinine ratio showed no differences between groups, and the same is true regarding serum biochemistry. No significant differences were found between the prevalence of casts in histological sections of the kidneys between the different treatment groups.
In Vivo Efficacy in Alzheimer's Disease Transgenic Mice
The effect of DP-155 and indomethacin to reduce the level of the Aβ in the brain was investigated in female Tg2576 hAPP-transgenic mice with two doses. The high dose, 0.14 mmol/kg (50 mg/kg/day indomethacin or 135 mg/kg/day DP-155), was administered every 4 h, a total of 18 times. Two deaths occurred in the indomethacin-treated group, and three deaths occurred in the DP-155-treated group. In the vehicle-treated group all test animals remained healthy. The levels of SDS-soluble Aβ40 and Aβ42 in the different treatment groups are presented in Fig. 6, A and B, respectively. Neither treatment affected the level of Aβ40. A significant reduction in the level of SDS-soluble Aβ42 was found in the indomethacin- and DP-155-treated groups (-29.7 or -23.6%, respectively) compared with the vehicle-treated group.
The low dose, 0.046 mmol/kg indomethacin or DP-155 (16.66 or 45 mg/kg/day, respectively), was administered every 12 h over 3 days. Two deaths occurred in the indomethacin-treated group after the fifth dose. However, in the vehicle- or DP-155-treated groups, all test animals remained healthy. Neither treatment had an affect on the level of Aβ40 (Fig. 6C). A significant reduction in the level of SDS-soluble Aβ42 was found in the indomethacin- and DP-155-treated groups (-22.5 or -11.2%, respectively) compared with the vehicle-treated group (Fig. 6D).
Pharmacokinetics
The comparison between the pharmacokinetic profiles of indomethacin or DP-155 after oral administration revealed that after oral administration of DP-155 (10 mg/kg, n = 4), indomethacin was absorbed into the circulation, presumably after its cleavage from DP-155 in the rat gut, yet no absorption of DP-155 (as the whole complex) was detected (Fig. 7; Table 3). The serum concentration plot of indomethacin derived from DP-155 showed relatively constant concentrations that were maintained for 24 h, suggesting that indomethacin continues to be absorbed during this time, with no evidence for the high initial concentration observed following nonmodified indomethacin administration, which may be responsible for many adverse effects.
Indomethacin brain concentration-time profiles following oral administration of DP-155 or indomethacin are presented in Fig. 8. Brain and plasma pharmacokinetic parameters obtained following the administration of DP-155 or indomethacin are presented in Table 3.
The values of tmax were identical in the blood and brain for each experimental group (tmax = 4 h in brain and blood following DP-155 administration and 2 h in brain and blood following indomethacin administration). Hence, there was no lag time between the appearance of indomethacin in the blood and in the brain.
A surprising finding was revealed on comparing the indomethacin brain with plasma concentration ratio following administration of DP-155 or indomethacin. Although the indomethacin ratio of AUC0-24 h brain to AUC0-24 h plasma following oral administration of DP-155 was 0.075, this ratio following oral administration of indomethacin was 0.03, i.e., brain exposure was approximately 2.5-fold higher following DP-155 administration in comparison with indomethacin administration. The same phenomenon was observed also for Cmax values, with Cmax Brain to Cmax plasma ratio 3.5-fold higher following DP-155 administration (0.07 or 0.02 following administration of DP-155 or indomethacin, respectively). These ratios indicate that despite the higher concentrations in the plasma following the administration of indomethacin compared with DP-155, the former yields much lower relative concentrations in the brain.
Discussion
Indomethacin has been proposed to be a potential treatment of Alzheimer disease due to its ability to reduce Aβ42 production. A major drawback of chronic indomethacin therapy (as well as with other NSAIDs) is that severe gastrointestinal and renal adverse effects limit its use. To overcome this problem, a therapy is needed that will have the drug's beneficial effects in the brain while avoiding the consequent adverse effects. As shown in the current investigation, DP-155 seems to fulfill these requirements.
In the present study, a similar effect of DP-155 and indomethacin on reduction of soluble Aβ42 levels in the brains of 3-month-old female Tg2576 is demonstrated. Amyloid β42 accumulation in the brain has a central role in AD neuroinflammation and neurodegenerative processes. Comparing different NSAIDs, indomethacin was found to have the highest anti-Aβ aggregation potency (Gasparini et al., 2004). It was shown that blocking the activation of NF-κB by indomethacin was associated with significantly reduced insoluble Aβ42 and amyloid plaque burden in Tg2576 mice (Sung et al., 2004). The well described inhibitory effect of indomethacin on γ-secretase cleavage of APP to Aβ42 (Eriksen et al., 2003; Weggen et al., 2003) is most likely to be responsible for the reduction in soluble Aβ42 after a short-term treatment in young mice (Gasparini et al., 2004). Reducing soluble Aβ42 accumulation at a very early phase can prevent the formation of amyloid oligomers, the main neurotoxic species, and plaque deposition (Gasparini et al., 2004). Weggen et al. (2001) demonstrated that selective NSAIDs induce alteration of the γ-secretase cleavage pattern, which leads to reduced production of Aβ42, without affecting Aβ40 levels. Selective reduction of Aβ42 was also seen in our study but not in some other studies where non-indomethacin, γ-secretase inhibitors were employed (Lanz et al., 2005). We used the same model and dosing paradigm (high doses for a very short period) reported by Weggen et al. (2001) since it resulted in consistent reduction in Aβ42. In our high-dose study, 0.14 mmol/kg/day DP-155 or indomethacin was toxic. In the low-dose study, only indomethacin (0.046 mmol/kg/day) was toxic. This is in agreement with the rat toxicity studies and illustrates the superior therapeutic index of DP-155. We assume that the observed toxicity did not affect the results and since the favorable effect of indomethacin was Aβ42 specific, it is in agreement with previous studies at lower doses of indomethacin (Eriksen et al., 2003; Sung et al., 2004). Moreover, DP-155 showed an effect also at a nontoxic dose. This model, however, does not always give favorable results when non-indomethacin NSAIDs are used (Lanz et al., 2005; Peretto et al., 2005). The therapeutic effect should be further investigated in long-term studies to evaluate the relationship between the structure of various NSAIDs and anti-Aβ42 function, together with adjustments for the individual PK and toxicity profile of each compound.
The pharmacokinetic analysis of the brain indomethacin concentrations in correlation with the blood indomethacin levels reveals the surprising phenomenon of higher brain to plasma ratio of indomethacin following DP-155 oral administration in comparison with oral administration of indomethacin. This explains the equivalent central nervous system activity of DP-155 in the AD transgenic mice compared with indomethacin while displaying significantly lower systemic indomethacin concentrations than obtained with indomethacin administration. These data suggest that there is a threshold of plasma indomethacin, above which the entrance of indomethacin into the brain is blocked. This novel finding may result from a self-limiting mechanism of indomethacin brain uptake due to the cerebral vasoconstriction and consequent reduction in blood flow that is known to be induced by indomethacin (McCulloch et al., 1982; Markus et al., 1994). The relatively slow elimination of indomethacin from the brain (either following its administration as the drug or the prodrug) is favorable compared with that described for other NSAIDs (Gasparini et al., 2004), indicating the suitability of this drug for selective central nervous system activity.
Toxicity remains the limiting factor in the use of NSAIDs and the development of COX-1-sparing drugs has not entirely overcome this obstacle since they share a similar risk for adverse renal effects as COX-1 inhibitors (Brater et al., 2001). The safety studies in the present work reveal no adverse reaction 4 h postadministration of the highest administered dose of DP-155 (0.28 mmol/kg), whereas for indomethacin at a 33-fold lower dose, gastric ulcers were observed. This difference is probably directly related to the dissimilar pharmacokinetic profiles with a significantly higher indomethacin Cmax after indomethacin ingestion in comparison with the prodrug-cleaved product, as well as a delayed time to achieve maximal concentration with the latter. Indomethacin reaches its maximal concentration at 2 h; therefore, at 4 h, both systemic and local effects can be observed. Indomethacin derived from DP-155, on the other hand, only reaches its maximal concentration at approximately 4 h, leaving less time for pathology secondary to its systemic exposure to develop. This study, however, shows clearly that DP-155 lacks local toxicity, whereas local toxicity for indomethacin, which is well described and considered very high compared with other NSAIDs (Ivanov and Tzaneva, 2002), was clearly seen. DP-155 produces less toxic effects than indomethacin, which only become evident 24 h after a single dosing and following repeated administrations, when sufficient time for systemic exposure is allowed.
DP-155 does not alter COX-1 activity (data not shown), which may contribute to its local safety. Regulation (e.g., reduction) of gastric epithelial permeability in response to acid exposure is an important local protection mechanism, mediated by PGE2, which is inhibited by COX-1 inhibitors, including indomethacin, but not COX-2 inhibitors (Takezono et al., 2004). It is also known that unsaturated phospholipids reduced gastrotoxicity (Leyck et al., 1985), and the DP-155 structure is ideal in this respect. Indomethacin is also known for inducing a high prevalence of lesions in the small and large bowels (Whittle, 1998; Rainsford et al., 2003), including ulceration and inflammation that can cause hypoproteinemia, anemia, and perforation of the gut (Bjarnason et al., 1987; Hawkey, 1998). In the current study, there were indeed signs of intestinal toxicity, namely enteritis accompanied by blood and protein loss to the intestine. DP-155 was 10-fold safer than indomethacin in this respect. Indomethacin as a DP-155 metabolite had a relatively constant serum concentration over 24 h that would allow less frequent dosing; this might improve its intestinal safety further.
Repeated administration of 0.028 mmol/kg indomethacin and 0.14 mmol/kg DP-155 yielded reduced urine output and an increased NAG to creatinine ratio. Equivalent nephrotoxic signs were observed with DP-155 at a dose 5 times higher, indicating 5-fold better renal safety. The tubular damage expressed by increased NAG to creatinine ratio is most likely due to ischemia and this may indicate reduction in renal blood flow (Arnold et al., 1974). Renal safety was further studied by a more specific model. In this model, 4 h after treatment with indomethacin, DP-155, or vehicle, rats were treated by vasopressin followed by furosemide. Vasopressin induces renal vasoconstriction and mimics hypovolemia or dehydration, and furosemide induces diuresis through PGE2 synthesis with a subsequent increase PGE2 urinary excretion. Indomethacin abolishes the synthesis of PGE2 and its consequent natriuretic effect at the loop of Henle and thus removes PGE2's inhibitory effect on vasopressin's urine concentration (Ayano et al., 1984b). The 4-h timing in the present study was chosen since it is the time when the DP-155-cleaved indomethacin reaches its peak serum concentration, whereas the administered indomethacin is close to its peak. In previous studies using this model, indomethacin was found to decrease PGE2 synthesis, diuresis, and PGE2 excretion (Ayano et al., 1984'02; Stichtenoth and Frölich, 1998). In the current study, indomethacin at a dose of 0.028 mmol/kg produced a similar effect, but an equimolar dose of DP-155 had no such effect. Volume retention and antidiuretic effects are serious complications in the treatment of renal and heart failure, and DP-155 has the potential to decrease this risk in the treatment of AD patients. In this model, the antidiuretic effect of indomethacin was described as limited to the renal tubules and independent of GFR (Ayano et al., 1984a). In the current study, the creatinine clearance with DP-155 was similar to vehicle, whereas in the indomethacin-treated group, it was increased, which is difficult to explain. It may, however, require higher doses or prolonged treatment with indomethacin for a sufficient decrease in PGE2 that will elicit hemodynamic effects leading to reduced GFR. The question of the effect of indomethacin on the GFR is, therefore, awaiting further investigation with more sensitive methods and prolonged treatment time.
DP-155 has a superior safety profile compared with indomethacin, with similar efficacy in reducing Aβ42 levels in brains of Tg2576 mice. This is probably due to its unique pharmacokinetics that lead to increased brain exposure while limiting systemic exposure to indomethacin. DP-155 is, therefore, a potential candidate for chronic use, particularly in the treatment of AD. Its safety profile should be further investigated for prolonged periods, including additional renal safety studies. Further prolonged efficacy studies should include behavioral and cognitive studies such as memory and learning.
Footnotes
-
This work was supported by the Israeli Consortium of Pharmalogica and by D-Pharm Ltd.
-
doi:10.1124/jpet.106.103184.
-
ABBREVIATIONS: NSAID, nonsteroidal anti-inflammatory drug; AD, Alzheimer's disease; Aβ, amyloid β; COX, cyclooxygenase; GI, gastrointestinal; PGE, prostaglandin E; GFR, glomerular filtration rate; DP-155, mixture of 1-steroyl and 1-palmitoyl-2-{6-[1-(p-chlorobenzoyl)-5-methoxy-2-methyl-3-indolyl acetamido] hexanoyl}-Sn-glycero-3-phosophatidyl choline; APP, amyloid precursor protein; NAG, N-acetyl glycosaminidase; ELISA, enzyme-linked immunoassay; ANOVA, analysis of variance; AUC, area under the curve.
-
↵1 This work is a part of Ph.D. dissertation.
- Received February 20, 2006.
- Accepted June 7, 2006.
- The American Society for Pharmacology and Experimental Therapeutics