The Proton-Pump Inhibitor Lansoprazole Enhances Amyloid Beta Production

A key event in the pathogenesis of Alzheimer’s disease (AD) is the accumulation of amyloid-β (Aβ) species in the brain, derived from the sequential cleavage of the amyloid precursor protein (APP) by β- and γ-secretases. Based on a systems biology study to repurpose drugs for AD, we explore the effect of lansoprazole, and other proton-pump inhibitors (PPIs), on Aβ production in AD cellular and animal models. We found that lansoprazole enhances Aβ37, Aβ40 and Aβ42 production and lowers Aβ38 levels on amyloid cell models. Interestingly, acute lansoprazole treatment in wild type and AD transgenic mice promoted higher Aβ40 levels in brain, indicating that lansoprazole may also exacerbate Aβ production in vivo. Overall, our data presents for the first time that PPIs can affect amyloid metabolism, both in vitro and in vivo.


Introduction
Alzheimer's disease (AD) is the most prevalent neurodegenerative disorder worldwide. Many molecular lesions have been detected in AD, but one of the major pathological hallmarks is the extracellular deposition of amyloid-b (Ab) peptides in the brain [1], that results in oxidative and inflammatory damage, which in turn leads to energy failure and synaptic dysfunction [2]. Formation of Ab species is caused by the sequential cleavage of amyloid precursor protein (APP) by two proteases, b-secretase, also called b-site APP-cleaving enzyme 1 (BACE1), and csecretase. Ab40 is the major secreted form while Ab42 has been suggested to be the main pathological species in AD pathogenesis [3], although other truncated Ab peptides might also contribute substantially to toxicity and amyloidogenesis [4].
Several studies support the hypothesis that classic fibrillar amyloid plaques are deleterious to the brain, showing that the subpopulation of dense-core Ab plaques in particular, the so-called neuritic plaques, are intimately associated with local dendritic spine loss, changes in neurites, and gliosis in AD and mouse models [5,6]. However, the total number of amyloid plaques do not correlate well with the severity of illness [7] or with loss of neurons [8], arguing against a direct causal effect of plaques on cognition or neuronal cell death in AD.
More recently, an alternative hypothesis has been growing and gaining support, based on the idea that the toxic component is within the soluble fraction. There is data showing that soluble forms of Ab correlate more closely with dementia severity than fibrillar Ab [9,10,11,12] and that Ab oligomers alter dendritic spine density and affect hippocampal synaptic plasticity in vivo [13,14,15,16,17]. Furthermore, it has been demonstrated that brain oligomeric Ab, but not total amyloid plaque burden, correlates with neuronal loss and astrocyte inflammatory response in APP/Tau transgenic mouse model [18]. In this context, several studies show the presence of Ab oligomers in CSF of AD patients, while not in healthy individuals, and also the direct correlation of oligomers with cognitive impairment [19,20,21].
Despite all the efforts put on AD research over the past years, there are no effective treatments to prevent, halt or cure the disease. Indeed, there are only four FDA-approved drugs for AD treatment, although they mainly provide a symptomatic improvement and are unable to stop the disease progression [22]. This prompted us to use the therapeutic performance mapping system (TPMS) [23] to explore potential novel indications of marketed drugs to modify the biology of AD. In brief, the TPMS is a topdown systems biology approach with potential applications in drug repositioning [24]. Starting from the clinical effects produced by different therapeutic compounds (both, positive and negative), we first split them into causative physiological motifs and identified the responsible molecular effectors, which were then mapped onto the disease-related cell network. We afterwards established different relationships between drug targets and effector proteins in the network that are used for training a classifier, with capacity for predicting and scoring novel potential indications on AD of totally unrelated drugs. The complete results of this study will be published elsewhere [25]. Interestingly, the TPMS analysis suggested that lansoprazole could act as a strong potential modulator of the processes involved in amyloid-b pathology.
Lansoprazole is a proton pump inhibitor (PPI) drug widely used in the treatment of peptic ulcer disease and other conditions where inhibition of gastric secretion may be beneficial [26,27]. PPIs are generally well tolerated, and adverse effects are relatively infrequent [28,29]. Yet, chronic administration of PPIs is becoming increasingly common, and there is a growing concern about potential unexplored adverse effects from such long-term therapy [30].
In this study, we explore the effects of lansoprazole, and other PPIs, on b-amyloid production in a well-established cellular model of amyloid pathology, with special attention to the effect over the different Ab species. We assess the in vivo relevance of our findings in wild-type (wt) and AD triple transgenic (3xTg-AD) mice and we ultimately speculate about the potential mechanisms underlying the observed alterations.

Drug Treatments
For Ab production analysis, we seeded PS70 cells in 12-well plates at a density of 150,000 cells per well and we subsequently treated for 24 h with lansoprazole, omeprazole, pantoprazole or esomeprazole (Sigma-Aldrich) at different concentrations. We used the c-secretase inhibitor DAPT (N-[N-(3,5-Difluorophenacetyl)-Lalanyl]-S-phenylglycine t-butyl ester, Sigma-Aldrich) at 2 mM as a positive control. Lastly, we collected and stored cell culture supernatants at 280uC until further use.

Analysis of Ab Peptides by ELISA
We quantified human Ab42 and Ab40 levels by human bamyloid (1-42) and (1-40) ELISA kits (Wako Pure Chemical Industries), respectively. We loaded 100 mL of conditioned media on every assay. For 3xTg-AD mouse brain extracts we previously diluted them 1/40 in standard buffer and then we loaded following manufacturer's protocol. We measured rodent Ab40 and Ab42 levels with ELISA kits (IBL International) according to manufacturer's protocol.

Mass Spectrometry of Ab Species
We used W0-2antibody and Protein-G Sepharose beads to immunoprecipitate human Ab from conditioned media. We washed Sepharose beads twice in PBS and twice in 100 mM ammonium acetate. We then eluted Ab twice with 300 ml of 50% acetic acid and vacuum-dried. Finally, we resuspended the samples in 10 ml of 33% acetonitrile containing 0.1% tri-fluoric acetic acid and ultrasonicated. Afterwards we analyzed Ab species by MALDI-MS on sinapinic acid matrix with an UltraflexII TOF/ TOF (BrukerDaltonics).

Animals and Treatments
All animals were housed in an animal facility that is fully compliant with the European policy on the use of Laboratory Animals. Experimental protocols were approved by the Parc Científic of Barcelona Committee and meet the European and Spanish guidelines of animal experimentation.
We treated both female 3xTg-AD mice (Charles River) and non-transgenic mice (B6129SF1/J) at 7 month of age for 5 consecutive days with an intraperitoneal injection of lansoprazole. We diluted and administered lansoprazole in 10% DMSO and 18% of encapsin (2-hydroxipropil beta-cyclodextrine), at 20 mg/ kg or 100 mg/kg, respectively. At the end point, we sacrificed mice 5 h after the last treatment and we froze each hemisphere in liquid nitrogen and stored them at 280uC.

Brain Soluble Ab Extraction
We thawed non-transgenic mouse brains on ice in 36 (w/v) 0.2% of diethylamine (DEA) and 50 mM of NaCl buffer with a protease inhibitor cocktail (CompleteH EDTA-free, Roche), and then we homogenized them. We subsequently centrifuged homogenates at 100.000 g for 1 h at 4uC. Finally, we collected the supernatants, and neutralized them by addition of 1:10 volume of 0.5 M Tris-HCl pH 6.8. We stored samples at 280uC as DEAsoluble Ab fractions.
We thawed 3xTg-AD mouse brains on ice in 36(w/v) 2% SDS with a protease inhibitor cocktail (CompleteH EDTA-free, Roche) and then we homogenized them. Next, we centrifuged homogenates at 100.000 g for 1 h at 4uC. Finally, we collected the supernatants and we stored samples at 280uC as SDS-soluble Ab fractions.

Western Blotting
We determined lysate protein concentration using the Bio-Rad D C protein assay (Bio-Rad Laboratories). We loaded 20-40 mg of each cell lysate and electrophoresed in 10% Tris-glycine gels for Western blot analysis. For Ab species detection, supernatants we run in 11% urea gels. For the immunoblotting we incubated overnight at 4uC with the following primary antibodies: rabbit polyclonal anti-C-terminal APP (Sigma-Aldrich), rabbit polyclonal anti-BACE (catalytic domain, Abcam), rabbit polyclonal anti-sAPPb (IBL), 6E10 (against Ab 1-16, Covance) or mouse monoclonal anti-actin (Sigma-Aldrich) antibodies. Afterwards, we incubated with either an HRP-conjugated secondary antibody plus enhanced chemiluminescence substrate (Millipore), or with an infrared fluorescent-labelled secondary antibody (IRDye, Rockland Immunochemicals, Gilbertsville, PA) for 1 h at room temperature.

Statistical Analysis
All data are shown as mean 6 SD. Statistical tests included oneway ANOVA for repeated measures and t-test when appropriate.

Lansoprazole and Other PPIs Increase Ab Levels in ADlike Cellular Models
To investigate the potential effect of lansoprazole on Ab production, we treated PS70 Chinese hamster ovary (CHO) cells stably expressing both wild-type human APP and presenilin 1 (PS1) with increasing concentrations of lansoprazole for 24 h. We first used the MTT reduction assay as a proxy to determine whether the tested concentrations caused cell toxicity, which was not the case below 50 mM (data not shown), in agreement with previous studies [32]. We subsequently measured the Ab40 and Ab42 levels in conditioned media of cultures by specific ELISA immunoassays. We observed that Ab40 levels in the conditioned media increased after treatments above 10 mM ( Figure 1A), up to 2-fold increase in the amount of Ab40 with respect to the vehicletreated cultured cells. Furthermore, we observed a dose-dependent increase in Ab42 between 5 mM and 50 mM ( Figure 1A), with more than 200% Ab42 increase at 25 mM and over 300% when treated with 50 mM. As expected, cells treated with the c-secretase inhibitor N-[N-(3,5-Difluorophenacetyl)-L-alanyl]-S-phenylglycine t-butyl ester (DAPT) showed negligible levels of Ab. These results indicate that lansoprazole treatment increase Ab production at commonly used concentrations [32,33,34,35,36], which in turn are comparable to other Ab inducers such as the calcium ionophore A23187, which increases the production of Ab approximately 3-fold [37], or caffeine (at millimolar concentrations), that increases Ab levels until 4-fold [38]. In addition, we have observed an Ab42 increase of over 250% compared to vehicle in cell cultures treated at 50 mM lansoprazole, while fenofibrate, a potent Ab42 raising Ab38 lowering compound, generates an increase of 125% at the same concentration [39]. Overall, these findings provide further evidence about the accuracy of the TPMS to predict pharmacological alterations in Ab metabolism.
To check whether these results could also be extrapolated to other PPIs, we examined the effect of compounds of the same drug class on Ab production. Omeprazole is one of the most widely prescribed drugs worldwide, which is administered as a racemic mixture of its two enantiomers, S-omeprazole (esomeprazole) and R-omeprazole. Similarly, pantoprazole is also broadly used because of its relatively long duration of action compared with other PPIs, and its lower propensity to become activated in slightly acidic body compartments [40]. Since the chemical structures and pharmacological targets of the four PPIs are very similar, our TPMS computational analyses also predicted an effect of these compounds on the Ab metabolism. Thus, we decided to investigate the effect of omeprazole, pantoprazole and esomeprazole within 5 mM and 50 mM range, based on our previous results. Although the increase in Ab40 was non-homogeneous for the 3 drugs, they did show a dose-dependent increase in Ab42 levels (Fig. 1B), achieving up to 175% increase. Hence, these results suggest that the modulation of Ab production is a shared feature among the PPI drug class.

Lansoprazole Might Modulate the c-secretase Complex
To gain a deeper insight in the relative formation of Ab species induced by lansoprazole, we conducted an Ab-immunoprecipitation coupled with mass spectrometry analysis of high dose lansoprazole-conditioned media. Different Ab species are generated because c-secretase has multiple APP cleavage sites. The main produced species is Ab40 and, to a lower extent, Ab38 and Ab42. Intriguingly, small molecule drugs called c-secretase modulators (GSM) are able to shift the c-secretase cleavage site, being classified as straight GSMs when they lower Ab42 and rise Ab38 [41], or as inverse GSMs (iGSM) when they do the opposite [39,42,43]. Interestingly, our MALDI-MS analysis revealed a considerably altered Ab peptide pattern in cells treated with lansoprazole at 50 mM. The relative levels of Ab42 increased whereas the relative levels of Ab38 decreased (Figure 2A). Intriguingly, there was also an increase in Ab37. As expected, the c-secretase inhibitor DAPT completely abrogated Ab production, showing no Ab peaks. We further confirmed the Ab42 increase and Ab38 decrease by Western blot ( Figure 2B). Therefore, attending to the results obtained with lansoprazole in the Ab species production shift, one possible explanation would be that lansoprazole might act as an inverse GSM.
To further investigate this hypothesis, we wanted to test if lansoprazole was able to counterbalancing the Ab42-lowering capability of R-flurbiprofen, a well characterized non-steroidal anti-inflammatory drug (NSAID) that acts as straight GSM [44,45]. NSAIDs are widespread used due to the prevalence of diseases in the aging population and to their crucial role as effective antipyretic analgesics in a wide spectrum of conditions and diseases ranging from a common cold to rheumatoid arthritis [46]. However, they are known to disrupt the mucosal resistance to gastric acid through several mechanisms including suppression of prostaglandin production and are thus associated with adverse events such as gastric or duodenal ulcers. For that reason, the coadministration with PPIs is strongly recommended in certain circumstances [47,48]. Notably, albeit we can find straight and inverse GSM modulators within NSAIDs, the former have been considered as very interesting therapeutic agents in AD, since they can lower Ab42 without perturbing the cyclooxygenase (COX) activity, the principal pharmacological target of NSAIDs [41,49].
As expected, cells treated with R-flurbiprofen showed decreased Ab42 levels while cells treated with lansoprazole increased Ab42 levels ( Figure 2C). Interestingly, when the two drugs were combined, lansoprazole blocked the Ab42 decrease induced by R-flurbiprofen. Therefore, these results suggest that concomitant administration of PPIs with NSAIDs may neutralize the Ab42 lowering effect provided by NSAIDs, at least in cell culture.
Taken together, these findings suggest that lansoprazole increase Ab42 production similarly to other described iGSMs. Nevertheless, alternative mechanisms related to APP dimerization processes could also play a role in the observed changes of Ab42 levels [50].

Lansoprazole Might Increase BACE1 and Meprin b Protease Activities
The putative iGSM activity of lansoprazole could explain the increase of Ab42 coupled to a decrease of Ab38, but it cannot account for the rise of Ab40. Hence, we wanted to investigate other possible effects of lansoprazole on Ab production. The APP can be processed by two different pathways: the non-amyloidogenic pathway, involving a-secretase and c-secretase activities, and the amyloidogenic pathway, requiring BACE1 and c-secretase [2]. If the first cleavage is performed by a-secretase, soluble APPa (sAPPa) and APP C83 C-terminal fragments are generated, and the consecutive cleavage by c-secretase produces the p3 peptides, which are non-amylodogenic. On other hand, when APP is first cleaved by BACE1, soluble APPb (sAPPb) and APP C99 fragments are otherwise generated, and c-secretase cleavage ultimately generates Ab peptides and the amyloid precursor protein intracellular domain (AICD) fragments.
To explore how lansoprazole was able to increase Ab37 and Ab40 levels, we first tested whether it increased APP or BACE1 protein levels, since higher amounts of Ab substrate or processing enzyme would certainly explain the increase in Ab production [1,51]. However, we did not observe any significant change in the protein expression levels of either protein ( Figure 2D). We then interrogated if lansoprazole could enhance BACE1 activity instead, and measured the generation of sAPPb, a BACE1 cleavage product. Interestingly, we observed that lansoprazole promoted sAPPb production, suggesting an increase of BACE1 activity. In contrast, sAPPa remained unaffected by lansoprazole treatment ( Figure 2E).
In addition, the mass spectra also revealed a slight increase of the Ab2-37, Ab2-40 and Ab2-42 species (Fig. 2A), which could be accounted for the meprin b metalloprotease, recently identified as an APP cleaving enzyme at the p2 position [52]. Variations in the media pH induced by lansoprazole could thus boost this protease activity, generating Ab2-x peptides.
Overall, these findings suggest that lansoprazole may not only modulate c-secretase, but also seems to increase BACE1 activity, boosting the Ab species production that would be eventually reflected in Ab40 and Ab37 increased levels. Nevertheless, since the PPIs are known to be irreversible inhibitors of H + /K + ATPase, lansoprazole-induced variations in the media pH may also affect the activity of other proteases, such as meprin b, and ultimately affecting Ab production, particularly Ab2-x species.

Acute Treatment with Lansoprazole Increases Ab Levels in Both wt and AD Mouse Models
Our results would not go beyond a cellular curiosity if lansoprazole could not cross the blood brain barrier. However, it has been reported that it can indeed cross it and exert its effects in brain tissue [53]. Thus, to determine whether lansoprazole is capable of altering Ab production in the brain, we conducted short-term intraperitoneal administration in wt and AD tripletransgenic (3xTg-AD) mouse models, like in previous studies [39] ( Figure 3). 3xTg-AD mice overexpress human tau and APP in a mutant PS1 knock-in background, and present both plaque and tangle pathologies in an age-related manner [54]. Even though, acute treatments do not enable an evaluation of histopathological or cognitive alteration, since changes in Ab plaque burden or cognitive impairment usually occur after a long and sustained treatment of at least 2-3 months. Yet, 8-month age transgenic mice typically contain few Ab plaques but they have significant amounts of intracellular and soluble Ab [54], being a suitable age to test changes in soluble Ab.
We used administration doses of 20 mg/kg/day and 100 mg/ kg/day. When compared to the equivalent doses in human [55], the administered concentrations are far below the LC50 (5000 mg/Kg) and comparable to the doses prescribed to treat certain pathologies, such as the Zollinger-Ellison syndrome, 180 mg/day [56]. We found that, indeed, acute treatment with 100 mg/kg/day intraperitoneally administered lansoprazole increased soluble Ab40 levels in healthy, non-transgenic mice ( Figure 3A). Levels of soluble Ab42 were also slightly increased, although they did not reach statistical significance. In the case of 3xTg-AD mice, short-term treatment with 20 mg/kg/day or 100 mg/kg/day of lansoprazole dramatically raised soluble Ab40 levels in a dose-dependent manner ( Figure 3B). Interestingly, Ab40 production was higher than in non-transgenic mice, attaining almost to the 250% rise at 100 mg/kg/day. Similarly to the nontransgenic mice, we also observed a moderate increase in soluble Ab42 levels, although they were not statistically significant either. In order to contextualize the relevance of the effects shown in mice at doses used in this study, it is worth to mention that a recent patent proposes a daily dose up to 400 mg of lansoprazole to inhibit tumor growth in humans [33]. Using the body surface area (BSA) normalization method [55], the conversion of the 20 mg/ Kg and 100 mg/Kg mouse doses into equivalent human doses results in around 100 mg/day and 486 mg/day, respectively. Interestingly, we found an increase of Ab levels in 3xTg-AD mice treated with 20 mg/Kg/day of lansoprazole, providing evidence of a clear effect at human prescribed equivalent doses.
The differences observed in the Ab40/Ab42 levels between cellular and mouse models could be partially explained by the Ab quantization, since we only measured extracellular Ab in cells, while we measured both extracellular and intracellular Ab species in brain homogenates. Despite these differences, our findings demonstrate that lansoprazole is able to augment Ab production both in vitro and in vivo models, with an exacerbated effect in AD models.

Conclusions
Our results reveal that lansoprazole, in addition to its known inhibitory effect on gastric acid production, has an effect on Ab generation. Although the underlying mechanisms remain elusive, our observations show that lansoprazole increases Ab37, Ab40 and Ab42 and lowers Ab38 levels in an AD-like cell model. In addition, the increase of sAPPb and the lack of changes in APP and BACE1 protein levels seem to indicate that lansoprazole would not only modulate the c-secretase complex, but also increase BACE1 activity.
Taken together, we hypothesize that lansoprazole could inversely modulate the c-secretase activity by shifting the APP cleavage site, resulting in higher Ab42 and lower Ab38 levels ( Figure 4). Moreover, it might also increase the activity of other pH-dependent proteases, such as BACE1, raising total Ab production and particularly reflected in the raise of Ab37 and Ab40 levels, or meprin b, boosting Ab2-x species. Nevertheless, further experiments are needed to better understand the role of lansoprazole in Ab production and specifically to unveil its underlying mechanisms.
Notwithstanding, from a more clinical perspective, since PPIs are commonly used drugs, it would be interesting to perform epidemiologic studies to investigate whether the long-term use of PPIs could have any detrimental impact on AD, particularly in aged chronic recipients. Recent studies have actually reported potential inappropriate prescriptions (PIM) in aged people with dementia [57], where PPIs appeared among the most prevalent PIMs when used at maximum therapeutic dosage for more than 8 weeks [57]. . Lansoprazole raises Ab40 production in mice. A, Non-transgenic mice were treated 5 consecutive days with 100 kg/mg of lansoprazole (n = 10). Soluble Ab40 and Ab42 from brain extracts were measured by ELISA (n = 106 SD) p,0.05 (+). Lansoprazole increased Ab40 levels in non-transgenic mice. B, 3xTg-ADwere treated 5 consecutive days with 100 kg/mg of lansoprazole (n = 6). Soluble Ab40 and Ab42 from brain extracts were measured by ELISA (n = 66 SD) p,0.05 (+), p,0.01 (*). Lansoprazole increased soluble Ab40 levels in 3xTg-AD mice in a dosedependent manner. doi:10.1371/journal.pone.0058837.g003 We believe that our data demonstrate for the first time that lansoprazole and other PPIs can increase Ab not only in cell cultures but also in mice. These results can serve as a catalyst for further studies in order to evaluate whether the treatment with PPIs may have an impact on AD pathology. . Hypothetical mechanisms of lansoprazole on Ab production. Ab peptides are produced from the consecutive cleavage of APP by BACE1 (b-secretase) and c-secretase. The first cleavage generates soluble APPb (sAPPb) and the C99 C-terminal fragment, while the subsequent one releases Ab peptides and the amyloid precursor protein intracellular domain (AICD). In basal conditions (left), a variety of Ab species are formed. Conversely, when cells are treated with lansoprazole (right), BACE1 activity could be increased, generating more sAPPb and C99 fragments and subsequently increasing the overall Ab production. Lansoprazole also could act as an inverse GSM, shifting the c-secretase cleavage, augmenting Ab42 and reducing Ab38. Together, lansoprazole is able to increase Ab37, Ab40 and Ab42 species and decrease Ab38. doi:10.1371/journal.pone.0058837.g004