Small Molecule, Non-Peptide p75NTR Ligands Inhibit Aβ-Induced Neurodegeneration and Synaptic Impairment

The p75 neurotrophin receptor (p75NTR) is expressed by neurons particularly vulnerable in Alzheimer's disease (AD). We tested the hypothesis that non-peptide, small molecule p75NTR ligands found to promote survival signaling might prevent Aβ-induced degeneration and synaptic dysfunction. These ligands inhibited Aβ-induced neuritic dystrophy, death of cultured neurons and Aβ-induced death of pyramidal neurons in hippocampal slice cultures. Moreover, ligands inhibited Aβ-induced activation of molecules involved in AD pathology including calpain/cdk5, GSK3β and c-Jun, and tau phosphorylation, and prevented Aβ-induced inactivation of AKT and CREB. Finally, a p75NTR ligand blocked Aβ-induced hippocampal LTP impairment. These studies support an extensive intersection between p75NTR signaling and Aβ pathogenic mechanisms, and introduce a class of specific small molecule ligands with the unique ability to block multiple fundamental AD-related signaling pathways, reverse synaptic impairment and inhibit Aβ-induced neuronal dystrophy and death.


Introduction
Slowing the progression of Alzheimer's disease (AD) will likely require parallel strategies of managing amyloid-beta (Ab) levels and reducing neuronal vulnerability to Ab. With regard to neuroprotective strategies, the p75 neurotrophin receptor (p75 NTR ) is an attractive target [1,2]. Binding of neurotrophins, including nerve growth factor (NGF), to p75 NTR promotes proapoptotic or pro-survival signaling, depending on the recruitment of survival-versus death-promoting adaptors [3,4]. p75 NTR is expressed in adult brain primarily by basal forebrain cholinergic neurons, but also by hippocampal, entorhinal and neocortical neurons, each vulnerable in AD (reviewed in [5]). Moreover, p75 NTR expression is upregulated in cortical [6] and hippocampal [7] tissue in AD. Increased p75 NTR and decreased Trk neurotrophin receptor levels in AD, along with studies showing that increased p75 NTR /Trk ratios lead to neuronal degeneration, further encourage therapeutic targeting of p75 NTR [4].
Substantial overlaps exist between p75 NTR -mediated signaling and degenerative signaling in AD. In AD brain and in cultured neurons treated with Ab, there is excessive activation of calpain/cdk5 [8], GSK-3b [9], and JNK and its downstream transcriptional activator c-Jun [10]. p75 NTR stimulates calpain activation through its Chopper cell death domain, [11] and mediates NGF-induced inhibition of GSK-3b [12]; in addition, it can mediate induction of cell death through the activation of JNK [13]. Further, it has been reported that Ab can bind to p75 NTR , and the receptor mediates Ab-induced cell death, in part by induction of c-Jun activation [1,14,15]. Thus, there are numerous potential points of direct and indirect interactions between p75 NTR and AD pathogenic mechanisms. p75 NTR represents a significant target for AD therapeutic development from a number of perspectives. To the extent that direct interactions between p75 NTR and Ab contribute to AD, preventing those interactions could inhibit neurodegeneration. Further, modulating the receptor to reduce activation of c-Jun and calpain activity could counteract Ab activation of these signaling intermediates. Moreover, since Ab down-regulates trophic signaling, particularly the PI3K/AKT pathway which promotes survival and is important for synaptic function [16], the promotion of AKT activation by ligand binding to p75 NTR [3,17] may reduce the effects of Ab In addition, AKT down-regulates JNK [18] and GSK3b [19]. Thus, targeting p75 NTR could protect neurons from Ab by at least three possible mechanisms: i) blocking a deleterious interaction between Ab and p75 NTR ; ii) down-regulating deleterious signaling (calpain, GSK3b and c-Jun) which mediates Ab toxicity; iii) upregulating survival signaling (AKT) which is normally inhibited by Ab and which can antagonize Ab mechanisms. The latter two mechanisms could be operative even under circumstances in which Ab causes degeneration independent of p75 NTR .
Aberrant activation of the cdk5, GSK3b and JNK kinases leads to tau hyperphosphorylation, cytoskeletal disruption and neuritic dystrophy [20]. Also, excessive activation of these kinases along with Ab-induced inhibition of CREB activation causes synaptic dysfunction [21][22][23]. Thus, small molecules inhibiting activation of cdk5, GSK3b, JNK and/or c-Jun have become important therapeutic candidates [20,24,25]. However, it seems unlikely that modulating a single target will provide an effective therapy; in addition, given the ubiquity of these targets and their functions in a very broad range of cell types it may be anticipated that adverse effects will limit their utility. Therefore, the possibility of inhibiting Ab-induced excessive activation of cdk5, GSK3b and c-Jun through a receptor expressed by neurons particularly vulnerable in AD is an attractive strategy for therapeutic development.
Expanding on development of synthetic peptides modeled on loop 1 of NGF that prevent neuronal death through p75 NTRmediated mechanisms [26], we identified small molecule, nonpeptide ligands, with favorable pharmaceutical properties, that bind specifically to p75 NTR and activate survival-promoting signaling, including AKT, in hippocampal neurons ( [17], reviewed in [5]). In the present study, we tested the hypothesis that these ligands would interfere with deleterious Ab signalling and its functional consequences. We demonstrate that these compounds inhibit Ab-induced neuronal death, neuritic degeneration and activation of calpain/ cdk5, GSK3b, and c-Jun; and reverse Ab-mediated inhibition of AKT and CREB activation, and synaptic function. These findings suggest that the use of small molecule p75 NTR ligands may be a therapeutically feasible approach to AD capable of simultaneously targeting multiple underlying pathogenic mechanisms.

Results
p75 NTR small molecule ligands inhibit Ab-induced death of hippocampal, cortical and septal neurons Ligands LM11A-24 and LM11A-31 were selected from p75 NTR ligands developed by our laboratories based on chemical and pharmacological features favorable for drug development [17]. LM11A-31 is an isoleucine derivative (MW 243.3) and LM11A-24 is a caffeine derivative (MW 322.4). LM11A-36 is structurally similar to LM11A-24 but is inactive in neurotrophic assays [17] and served as a negative control (Fig. S1). 6-7 DIV hippocampal, cortical and septal neurons were treated with oligomeric Ab in the presence or absence of p75 NTR small molecule ligands or NGF. Preliminary experiments demonstrated that oligomeric Ab (derived from each of three oligomeric Ab protocols as described in Methods and Fig. S2) reached maximum toxicity (neuronal death) at 5-10 mM (data not shown), a concentration range similar to that reported by other laboratories. Exposure of neurons to Ab resulted in somal shrinkage and vacuolation along with neurite fragmentation and beading; whereas the majority of neurons co-treated with Ab and 100 nM LM11A-24 or LM11A-31 exhibited a normal appearance with intact cell bodies and neurites ( Fig. 1A-D). Quantitative analysis of hippocampal, cortical, and septal neuronal cultures demonstrated that Ab induced a 40-65% reduction in survival while simultaneous addition of LM11A-24 or LM11A-31 prevented 70-90% of the Abinduced decrease in survival. NGF and LM11A-36 had no protective effect except for a minor protective effect by NGF in hippocampal and septal cultures (Fig. 1E, G). In dose-response studies, LM11A-24 and -31 inhibited Ab-induced death with EC 50 values of ,20 nM, with the protective effect persisting to at least 500 nM (Fig. 1H, I). LM11A-31 was also found to be protective against Ab fibrillar preparations (Fig. 1J).
To confirm that p75 NTR ligands prevent Ab-induced neuronal death, hippocampal neuronal survival was further assessed using Terminal deoxynucleotidyl Transferase Biotin-dUTP Nick End Labeling (TUNEL)/DAPI staining. Neurons treated with Ab exhibited increased TUNEL staining whereas fewer neurons cotreated with Ab and LM11A-31 demonstrated TUNEL signal ( Fig. 2A-C). Quantitative analyses (Fig. 2D) demonstrated that in the absence of Ab, addition of LM11A-31 had no significant effect on neuronal survival while NGF was associated with a small but statistically significant increase in death. These findings were consistent with previous studies showing that under certain culture conditions, NGF promotes death of hippocampal neurons [27]. LM11A-31 inhibited Ab-induced death while NGF had no protective effect. The ability of LM11A-31 and LM11A-24 to block Ab-induced death was further confirmed using Hoechst staining (Fig. S3).

Small molecule ligand protection is mediated through p75 NTR
Previous studies [17] demonstrated specificity of LM11A-24 and -31 for p75 NTR . The ability of p75 NTR ligands to induce PI3K/ AKT activation was entirely dependent on p75 NTR and these ligands did not activate Trk receptors [17]. These previous studies also demonstrated differential activities of neurotrophins and the small molecule ligands on p75 NTR -expressing cells. In cultures of oligodendrocytes, which express p75 NTR but not Trk receptors, NGF or proNGF can induce cell death; however, LM11A-24 and -31 not only failed to promote death but blocked neurotrophininduced death. In the current study, we reasoned that if a compound promotes its protective effect through a ligand-type interaction with p75 NTR , this effect should be blocked by the addition of a receptor-saturating concentration of a competing, non-protective ligand, such as NGF. The addition of NGF (100 ng/ml) together with LM11A-31 (100 nM) reversed entirely the protective effect of LM11A-31 ( Fig. 2D), consistent with a model in which these compounds act through a common receptor. An alternative explanation is that toxicity induced by NGF might mask a non-p75 NTR -dependent protective mechanism of LM11A-31. As a second approach for determining whether the LM11A-31 protective effect was mediated through p75 NTR , assays were conducted using p75 NTR +/+ and p75 NTR 2/2 hippocampal neurons derived from mice maintained on a C57Bl/6-strain background. Using neurons from that strain, maximum Abinduced death was reached at 2.5 mM Ab and this dose was therefore used for subsequent studies. In p75 NTR +/+ cultures, Ab triggered a 2-fold increase in death that was significantly inhibited by LM11A-31, a protective effect consistent with earlier studies (Fig. 2E). In p75 NTR 2/2 cultures, significant Ab toxicity was reduced to a 1.5-fold increase in death. The significant difference in the degree of cell death triggered in p75 NTR wildtype and mutant cultures indicated that the presence of wildtype p75 NTR contributes to Ab-induced toxicity. Notably, in p75 NTR2/2 cultures LM11A-31 had no protective effect, indicating a protective mechanism requiring p75 NTR .
p75 NTR small molecule ligands prevent Ab-induced neuronal death in organotypic hippocampal slice cultures Organotypic slice culture derived from postnatal brain and matured in vitro, is a widely used model for the study of  PLoS ONE | www.plosone.org neurodegenerative mechanisms and potential therapeutics. Addition of Ab to rat postnatal hippocampal slice cultures leads to death of pyramidal neurons as detected by propidium iodide (PI) uptake [28]. In our studies, 24 hour Ab exposure led to markedly increased in PI staining by pyramidal neurons which was blocked entirely by co-treatment with LM11A-31 (Fig. 3). had no effect on baseline death while NGF (100 ng/ml) was associated with a significant increase in death. Ab caused an approximate 2.0-fold increase in death that was significantly inhibited by LM11A-31 but not NGF. In the presence of NGF, LM11A-31 failed to prevent Ab-induced death (n = 98-131 fields derived from a total of 6 separate experiments). (E) In 6-7 DIV hippocampal neuronal cultures derived from C57Bl/6 p75 NTR +/+ and 2/2 mice, Ab triggered a 2.0-fold increase in death in p75 NTR +/+ cultures that was significantly inhibited by LM11A-31. In p75 NTR 2/2 cultures, Ab triggered a significant 1.5-fold increase in death, a degree of increase significantly less than that found in +/+ cultures. In p75 NTR   Organotypic slice cultures were prepared from PND-8 rat brain and allowed to mature in vitro for 11-19 days. Pyramidal neuron death was detected by propidium iodide (PI) staining, and all experimental conditions were compared to Ab treatment alone. Upper row, at baseline only trace levels of PI staining were detected. Middle row, PI staining after a 24 hour exposure to either culture medium (CM), Ab or Ab+LM11A-31 at 100 nM demonstrates readily apparent Ab-induced pyramidal neuron death that is inhibited in the presence of p75 NTR ligand.
Bottom row, PI staining shows maximum neuronal death after 24 hour treatment with NMDA. In the lower panel, quantitative analysis of PI staining demonstrates a significant reduction in Ab-induced neuronal death (n = 56-59 brain slices derived from 4 independent studies). doi:10.1371/journal.pone.0003604.g003 p75 NTR small molecule ligands prevent Ab-induced neuritic dystrophy in matured neurons Ab-induced tau/cytoskeletal derangement causes neuritic dystrophy, a process which occurs in early stages of AD [29] and is characterized by the appearance of varicosities and excessive tortuosity [30,31]. Hippocampal neurons kept in vitro for $3 weeks express mature isoforms of tau protein, and when exposed to Ab primarily demonstrate neuritic dystrophy rather than death [30]. Treatment of DIV 21-22 hippocampal neurons with 5 mM Ab induced dystrophic changes which were prevented almost entirely by LM11A-24 and -31 ( Fig. 4A-D). Assessment of dystrophy by visual criteria (Fig. 4D) and by quantitation of the neurite mean differential curvature, a measure of tortuosity ( Fig. 4E-H), showed that these ligands effectively blocked Abinduced dystrophy. This finding suggests that the compounds may interfere at an early/upstream stage with the complex cascade of Ab-induced signaling [20] which results in the disruption of neuritic integrity.

p75 NTR small molecule ligands prevent Ab-induced AKT inactivation and their protective effect is dependent, in part, on PI3K activity
In previous studies with hippocampal neurons, LM11A-24 and -31 were found to activate AKT in a p75 NTR -dependent manner and their neurotrophic effect was dependent upon PI3K [17]. These findings raised the possibility that these ligands might counteract the ability of Ab to decrease levels of AKT activation. Treatment of 21-22 DIV hippocampal neurons with Ab induced a significant decrease in AKT activity that was blocked by LM11A-24, -31 and NGF, although the effect of NGF was less than that of the small molecules (Fig. 5A). To determine whether the small molecule protective effect was dependent upon PI3K/AKT activity, we cotreated 6-7 DIV hippocampal neurons with Ab, LM11A-24 or -31, and LY294002, an inhibitor of PI3K. Under baseline conditions (Fig. 5B), LY294002 induced a 1.7-fold increase in cell death without reaching significance, a trend consistent with the known role of PI3K/AKT signaling in promoting neuronal survival [3]. In the presence of Ab without LY294002, LM11A-24 and -31 blocked the Ab-death promoting effect consistent with earlier results. However, in the presence of LY294002, the protective effect of LM11A-24 and -31 was significantly reduced. The ability of LY294002 to block the p75 NTR ligand protective effect, along with the finding that p75 NTR ligands reverse Ab-induced inhibition of AKT activation, suggest that maintenance of PI3K/AKT signaling by LM11A-24 and -31 contributes to their protective effect.
p75 NTR small molecule ligands inhibit Ab-induced activation of calpain/cdk5, GSK3b and c-Jun Activation of calpain results in the cleavage of the 250 kDa cytoskeletal protein a-fodrin to 145 and 150 kDa fragments, and the levels of these fragments serve as a measure of calpain activity [32]. Exposure of 21-22 DIV hippocampal neurons to Ab triggered a nearly 3-fold increase in the ratio of fragmented to full-length a-fodrin forms (Fig. 5C). Of note, a-fodrin is also a substrate for caspase 3 in degenerative states, with cleavage to 150 kDa and 120 kDa forms [33]. Quantitation of 120 kDa afodrin fragments revealed no significant differences across conditions (data not shown) suggesting that Ab-induced cleavage resulted from calpain rather than caspase activity. LM11A-24, -31 and NGF inhibited Ab-induced a-fodrin cleavage, indicating that they prevented Ab-induced calpain activation.
Activated calpain cleaves the cdk5 p35 regulatory subunit to the p25 constitutively active form promoting excessive cdk5 activation; the ratio of p25 to p35 reflects cdk5 activation [32]. Treatment of 21-22 DIV neurons with Ab resulted in markedly increased p35 cleavage, as evidenced by a significant increase in the ratio of p25 to p35 (Fig. 5D). Co-treatment with LM11A-24, -31 or NGF prevented the increase in p25/p35 ratio, demonstrating that p75 NTR ligands inhibit Ab-induced cdk5 activation.
Given that LM11A-24 and -31 promote PI3K/AKT signaling [17] and that this signaling inhibits GSK3b [19] and JNK activation [34], we determined whether these ligands, as well as NGF, prevent Ab-induced GSK3b and c-Jun activation. GSK3b activity is increased when the Ser 9 residue is dephosphorylated, providing a measure of the GSK3b activation state [19]. Addition of Ab to 21-22 DIV cultures induced significant dephosphorylation (i.e., activation) of GSK3b, while co-administration with LM11A-24 and -31, but not NGF, prevented this activation (Fig. 5E). An established method for monitoring JNK and c-Jun activation consists of quantitating neuronal nuclei positive for phospho-c-Jun immunostaining [35]. Treatment of 6-7 DIV neurons with Ab resulted in activation of c-Jun, which was largely inhibited by co-treatment with LM11A-24 or -31 but not by NGF (Fig. 5F).
In summary, LM11A-24 and -31, but not NGF, demonstrated significant inhibition of Ab-induced GSK3b and c-Jun activation. In contrast, LM11A-24, -31 and NGF each inhibited the ability of Ab to downregulate AKT signaling and to promote calpain/cdk5 signaling. These differences in signaling profiles between the small molecules and NGF might account, in part, for the greater protective effects of the small molecules and point to the importance of GSK3b and c-Jun activation in mediating Abinduced degeneration.
Since cdk5, GSK3b and JNK contribute to Ab-induced tau phosphorylation [20,24,25], we determined whether LM11A-24 and -31 might also prevent tau phosphorylation. In 21-22 DIV neurons, exposure to Ab resulted in a significant increase in tau phosphorylation at Ser 202 , a well characterized tau residue which is phosphorylated by each of these kinases and which is found to be phosphorylated in early AD [25,36,37]. Co-administration of LM11A-24 or -31 almost entirely prevented Ab-induced tau Ser 202 phosphorylation (Fig. 5G).

p75 NTR small molecule ligands prevent Ab-induced inactivation of CREB
In addition to aberrant activation of calpain/cdk5, GSK3b and JNK, another potential mechanism by which Ab inhibits synaptic function is through inhibition of CREB, a fundamental contributor to long-term potentiation (LTP) [21]. To further examine the effects of LM11A-24 and -31 on Ab-induced changes in signaling, we determined whether these ligands were capable of blocking Abinduced CREB deactivation. Treatment of 21-22 DIV hippocampal neurons with Ab for 3 hours resulted in a 43% reduction in phosphorylation of CREB, consistent with previous studies [21]. Notably, co-treatment with 100 nM LM11A-24 or -31 blocked the inhibitory effect of Ab on CREB phosphorylation (Fig. 5H).

LM11A-31 corrects deficits in synaptic transmission in Abtreated hippocampal slices
The effect of the ligands on CREB phosphorylation strongly suggested that they might be able to reverse the Ab-induced inhibition of LTP in the brain. Electrophysiological experiments were performed on hippocampal slices derived from adult mice that received a tetanus to produce LTP at the Shaffer collateral-CA1 connection. As previously shown [21], we found that the exposure of slices to Ab (200 nM) for 20 min before tetanization  . p75 NTR small molecule ligands prevent Ab-induced neuritic dystrophy. 21-22 DIV hippocampal neurons were exposed to the following conditions: (A) culture medium (CM); (B) 5 mM oligomeric-NaOH-derived Ab; or (C) 5 mM oligomeric-NaOH-derived Ab+100 nM LM11A-31. After 48 hours, cultures were fixed and immunostained for MAP2 to visualize dendrites. In the presence of CM alone, dystrophic changes including beading and tortuosity were rare. In contrast, Ab induced beading (arrowheads) and increased tortuosity (brackets), and each of these changes was markedly reduced with co-administration of LM11A-31. (D) Dystrophic neurites, defined as neurites exhibiting beading and/or multiple abrupt turns (i.e., tortuosity) were measured by a blinded observer. Data is expressed as average numbers of dystrophic neurites per field. All conditions were compared to Ab alone (n = 8 randomly chosen fields from 3 independent experiments). (E) Mean differential curvature (MDC) analysis in randomly selected fields demonstrated that oligomeric-NaOH-derived Ab induced a significant increase in MDC which was prevented by LM11A-31 (n = 7-9 fields per condition).  PLoS ONE | www.plosone.org led to significant impairment of LTP (Fig. 6). However, when slices were treated concomitantly with Ab and LM11A-31 (100 nM), LTP was normalized. LM11A-31 alone did not affect potentiation.

Discussion
There is no available therapy for AD that effectively targets underlying disease mechanisms. The present studies demonstrate that non-peptide small molecule ligands targeting p75 NTR , a receptor upregulated in neurons vulnerable in AD, prevent Abinduced neuritic dystrophy and cell death, while inhibiting Abinduced tau phosphorylation and the activation of several key signaling intermediates, each a candidate therapeutic target in its own right. The inhibition of Ab-induced and transgenic mouserelated synaptic impairment, the prevention of neuritic dystrophy and neuronal death, along with the modulation of multiple intracellular signaling mechanisms involved in AD is a novel activity profile for small molecule ligands acting at a known receptor target.
The protective effect of these compounds is likely mediated through p75 NTR . Their effects on neuronal signaling and survival are p75 NTR -dependent, they induce signaling-adaptor recruitment to p75 NTR and they fail to activate Trk receptors [17]. In addition, a standardized receptor binding screen conducted with LM11A-31 (Cerep panel, Table S1) failed to detect binding of p75 NTR ligands to other receptors. NGF lacked protective activity, and as expected for an 'inactive' ligand, inhibited the protective effect of p75 NTR small molecule ligands, consistent with their protective effect being mediated through p75 NTR . Moreover, in assays employing p75 NTR2/2 neurons in which Ab-induced death was reduced but still present to a significant degree, small molecule ligand protective activity was entirely absent. Taken together, our prior and current studies indicate that these ligands inhibit Abinduced degeneration by interacting with p75 NTR . The expression of p75 NTR by glial cells [38] raises the possibility that these ligands might also function through p75 NTR -mediated effects on nonneurons; however, the relative lack of non-neuronal cells in the culture protocols applied here and the effects evident in short term signaling studies make this possibility unlikely.
There are several mechanisms by which p75 NTR ligands might inhibit Ab toxicity. Ab(1-40) aggregates bind to p75 NTR and induce cell death and a modified NGF loop 1 synthetic peptide was found to block NGF binding to p75 NTR and prevent Ab(1-40)induced death of cultured cortical neurons [14]. It has yet to be determined whether Ab(1-42) oligomer species similarly bind to p75 NTR to induce neuronal degeneration. Our finding that Ab(1-42) oligomer-induced death is significantly, though incompletely, decreased in p75 NTR 2/2 cultures suggests that p75 NTR might be one of several targets mediating oligomeric Ab(1-42) effects. Thus, small molecule p75 NTR ligands might prevent degeneration, in part, by preventing Ab binding to p75 NTR . Alternatively, p75 NTR might not be an important target for Ab binding but might instead contribute to Ab toxic effects through indirect mechanisms. Its ability to affect deleterious or survival-promoting signaling (i.e. c-Jun and PI3K/AKT, respectively) might contribute an 'enabling' mechanism that augments Ab toxicity. In this scenario, the protective function of small molecule ligands would be derived from optimizing the degenerative-versus survival-promoting functions of p75 NTR rather than inhibiting Ab oligomer binding. The findings that NGF has no protective effect against Ab in some assays, and that the toxic effect of Ab is only partially lost in p75 NTR 2/2 neurons, suggest that the protective effect of p75 NTR small molecule ligands, especially in assays in which toxic effects of Ab are blocked entirely, is unlikely to be mediated solely by a simple mechanism of inhibiting Ab binding to p75 NTR .
Deleterious functional effects promoted by nanomolar, nonlethal, concentrations of Ab may occur through mechanisms in addition to those promoting gross neuronal dystrophy and death. Recent findings suggest that p75 NTR is highly enriched in postsynaptic densities and interacts with the PDZ3 domain of the PSD-95 scaffolding protein [39]. In addition, evidence continues to accumulate that Ab might impair synaptic function by perturbing PSD-95 related structures and function [40,41]. Thus, the finding that LM11A-31 is capable of blocking low concentration Abinduced impairment of synaptic function in adult hippocampal slices suggests additional pathways through which p75 NTR small molecule ligands may mitigate Ab effects. LM11A-31 is the first small molecule ligand shown to prevent Ab-induced synaptic LTP impairment when administered at nanomolar, rather than micromolar, concentrations.
The favorable drug development profile of LM11A-24 and -31, their low nanomolar potency and receptor selectivity, and their unprecedented ability to simultaneously inhibit Ab-induced activation of calpain/cdk5, GSK3b and c-Jun and to block Ab impairment of synaptic function, each of which are current pharmaceutical candidate targets for AD therapeutics, establish an important new class of candidate small molecule compounds for PLoS ONE | www.plosone.org AD therapeutics. Novel derivatives of these compounds with optimized pharmaceutical profiles are currently under study and will serve as important candidates for clinical development. Current studies in our laboratories are assessing the ability of p75 NTR ligands to decrease neurite degeneration and to correct behavioral deficits in transgenic mice that over-express Ab.

Primary neuronal cultures
Animal procedures were approved by each participating university's Committee on Laboratory Animal Care and were conducted in accordance with the NIH Guide for the care and use of laboratory animals. Cortical, hippocampal and septal cultures were prepared from embryonic day 16 (E16) CF1 mouse fetuses [42]. Tissue culture wells with or without coverslips were coated with 10 mg/mL poly-L-lysine in PBS. Cells were incubated in DMEM/ F12 containing 10% fetal bovine serum for the first 16-20 hours, and subsequently maintained in serum-free Neurobasal medium with 16 B27 supplement (Invitrogen). p75 NTR +/+ or 2/2 hippocampal neurons were derived from E15-17 mice maintained on a C57Bl/6 background and cultured under the same conditions, with the addition of 1 mM Glutamax supplement. For neuronal viability assays, neurons were seeded in 24-well plates at a density of 20,000-30,000 cells per well or in 12-well plates at 80,000-100,000 cells per well and allowed to mature 6-7 days. For neuritic dystrophy assays, 150,000-200,000 neurons per well were seeded into 6-well plates containing 25 mm coverslips and matured [21][22] days. For all cultures, medium was changed every 48-72 hours, prior to the addition of Ab and various compounds. MAP2 immunostaining demonstrated that each of the above protocols resulted in cultures containing 90-95% neurons with 100% of the MAP2-positive cells expressing p75 NTR . For Western blot signaling assays, the same protocol was followed except neurons were seeded at 450,000 cells per well in 6-well plates.

Ab preparations
Ab(1-42), referred to as ''Ab'', and Ab(S) (Ab(1-42) residues in a scrambled sequence) were obtained from rPeptide (Athens, GA). Ab oligomers were prepared using three different established methods. For the first two, Ab peptide was resuspended in 0.5 mM NaOH [43] or 0.2% NH 4 OH [44] at a concentration of 350 mM and stored at 270uC. For use in cell cultures, the stock solution was incubated at 37uC for 5-7 days. In the third method [45], 1.0 mg of Ab peptide was dissolved in 250 ml hexafluoroisopropanol (HFIP), aliquoted in sterile microcentrifuge tubes and HFIP was removed under vacuum in a Speed Vac. Resulting peptide films were stored desiccated at 220uC. Before use, the peptide was resuspended to 5 mM in dry dimethyl sulfoxide (Me 2 SO, Sigma), brought to 80 mM in PBS and incubated at 4uC for 16-24 hours. For the production of Ab fibrils, HFIP-prepared peptide was brought to 222 mM in 10 mM HCl.
Peptide preparations were characterized by Atomic Force Microscopy (AFM). A multimode scanning probe microscope controlled by a NanoScope IIIa controller was used in conjunction with an E-series piezoelectric scanner (Digital Instruments, Santa Barbara, CA). AFM probes were etched silicon micro cantilevers, model MPP-11100 (Veeco, Santa Barbara, CA). Samples were applied to freshly cleaved mica using a Langmuir-Schaffer horizontal transfer system, and were then rinsed with Milli-Q Ultrapure water (Millipore, Temecula, CA) [46]. Image data was acquired at a scan rate of 0.5-1 Hz. AFM analysis (Fig. S2) of Ab preparations demonstrated that NaOH-and NH 4 OH-based protocols resulted in primarily oligomeric species with occasional fibrils and that HFIP-based protocols resulted in primarily oligomeric species with rare or absent fibrils. The HCl-based protocol yielded primarily fibrillar species. Oligomeric Ab preparations were used in all studies except as indicated in Fig. 1J.

Quantitation of neuronal survival
p75 NTR ligands (LM11A-24 and LM11A-31), NGF or signaling inhibitors were added concomitantly with Ab. Ab preparations were added to 6-7 DIV cultures at a final concentration of 5 or 10 mM followed by 72 hours incubation. Following incubation with Ab neurons were stained with Syto 13 (Molecular Probes), Hoechst 33258 (Calbiochem); or TUNEL/DAPI, using the fluorescein-12-dUTP, DeadEnd TM Fluorometric TUNEL System (Promega Madison, WI), and VECTASHIELDH+DAPI (Vector Labs Burlingame, CA). Stained neurons were visualized under a fluorescence microscope (Leica DM IRE2) using 520 nm (TU-NEL, Syto13) or 460 nm (DAPI, Hoechst) filters. Survival of neurons was determined based on morphological criteria and Syto-13 (aids in cellular visualization) as assessed by phase contrast microscopy [47]. A dead or degenerating neuron was defined as one with a vacuolated cytoplasm, shrunken soma and/or beaded or retracted neurites. Data are expressed as percent of the total number of observed neurons that were scored as surviving. Neuron death was quantified with Hoechst 33258 by counting the number of cells containing condensed or fragmented nuclei as a percent of total observed neurons, and also with the TUNEL/ DAPI system, by dividing the number of nuclei exhibiting TUNEL staining by the total number of nuclei as identified by DAPI. Cell survival analyses were confirmed by blinded counts.
Quantitation of neuritic dystrophy [21][22] DIV hippocampal neuron cultures were treated with fresh Neurobasal B27 medium containing Ab at a final concentration of 5 mM in the presence or absence of LM11A-24 or -31 for 48 hours and then fixed in fresh 4% paraformaldehyde. Neurites were imaged by immunostaining with MAP-2 monoclonal primary followed by Cy3-conjugated anti-mouse secondary antibody (Jackson Immunoresearch). MAP2 positive dendrites were quantitated using the MEASURE COUNT OBJECT function of MetaMorph 5.0 (Universal Imaging Inc, West Chester, PA). Only dendrite branches longer than 10 mm were considered (processes shorter than 10 mm were considered as sprouts). Dendrites were considered dystrophic when they showed a persistent pattern of increased tortuosity (multiple abrupt turns). To more precisely quantitate the degree of neurite curvature, we modified an established method for assessment of neurite curvature [31]. In randomly selected fields, neurite courses were digitized and approximated by a series of n manually chosen connected line segments using NIH-Image (Fig. S4). Neurite analyses were confirmed in a blinded manner. Using a Sigmaplot macro, the angle of each segment was determined relative to a line connecting the endpoints of the neurite tracing (a), and, beginning at one end, successive angles were subtracted from the prior angle in the chain, and the results averaged to give the 'mean differential curvature' (MDC = S(a i+1 2a i )/n). This parameter reflects the degree of curvature over the course of the neurite with an increasing value indicating increased curvature.

Organotypic hippocampal slice cultures
Organotypic hippocampal slice cultures were prepared using previously described methods [48] with modification. Briefly, 350 mm thick hippocampal slices were prepared from postnatal 8day-old (PND-8) Wistar rats using a tissue chopper and separated in ice-cold Hank's balanced salt solution (HBSS) with 33.3 mM glucose, 4.2 mM NaHPO 4 , 10 mM MgSO 4 , 10 mM HEPES, 0.3% BSA and Penicillin-Streptomycin, pH 7.3. Slices were mounted on Millicell culture inserts and transferred to 6-well culture plates. Each well contained 1 mL of tissue culture medium consisting of 50% minimum essential medium, 25% HBSS, 25% heat inactivated horse serum and supplemented with 33 mM glucose, 12.5 mM HEPES and penicillin-streptomycin. Prior to the addition of Ab and small molecules, organotypic cultures were matured under tissue culture conditions for 11-19 days. Culture medium was changed three times per week.
Pyramidal neuron death in brain slices was quantitated by measuring propidium iodide (PI) uptake using a previously established protocol [48]. In brief, slices were treated with Ab6LM11A-31 for 24 hours and then with 50 mM N-Methyl-D-Aspartic acid (NMDA) to induce maximum pyramidal neuron death in each slice. Cellular damage was assessed by fluorescent image analysis of propidium iodide uptake (PI; 2 mg/mL), indicative of significant membrane injury. Before and after treatment, slices were observed and photographed with an inverted microscope (Nikon) coupled to a CCD camera, and mean fluorescent intensity in the CA1-3 regions was determined using OpenLab 4.04 software. Baseline (i.e. pre-treatment) fluorescence was subtracted from subsequent measurements and post-treatment cellular damage was normalized to the total neuronal density (determined by NMDA-induced PI intensity).

Protein preparation and Western blotting
For assays of AKT, calpain, cdk5 and GSK3b activity, 21-22 DIV hippocampal neurons were incubated for 4 hours in fresh neurobasal/B27 culture medium containing 5 mM oligomeric Ab in the presence or absence of LM11A-24, -31 or NGF. Following incubation, cells were collected and washed in ice-cold PBS and then lysed in RIPA buffer (20 mM Tris, pH 8.0, 137 mM NaCl, 1% NP-40, 10% glycerol, 1 mM PMSF, 500 mM orthovanadate, 10 mg/ml aprotinin and 1 mg/ml leupeptin). Lysates were sonicated for 10 seconds, centrifuged at 14,0006g for 10 minutes at 4uC, and the supernatant was collected. For tau phosphorylation studies, heat-stable fractions were prepared as described previously [49]. In brief, cells were collected by scraping cells into ice-cold TBS and centrifuging at 140006g for 10 min at 4uC. The supernatant was discarded and the pellet was resuspended in 100 ml MES/NaCl buffer (100 mM MES, 1 M NaCl, 0.5 mM MgCl 2 , 1 mM EGTA, 2 mM DTT, 1 mM Na 3 VO 4 , 1 mM benzamidine hydrochloride, 5 mg/ml leupeptin, 2 mg/ml aprotinin, 1 mg/ml pepstatin, 0.2 mM PMSF) and immediately heated to 100uC for 10 min. These were then cooled on ice and centrifuged at 140006g at 4uC for 25 min. The tau and MAP2c-enriched supernatant was retained. Phosphorylated CREB (p-CREB) levels were determined as described previously [21]. Cultured 21-22 DIV hippocampal neurons were treated with 5 mM Ab in Neurobasal medium without B27 supplement or Lglutamine for 3 hours followed by the addition of 50 mM glutamine for 15 min in order to promote CREB activation (phosphorylation). Cells were harvested in 16 modified RIPA buffer as above with the addition of 80 mM glycerophosphate, and whole-cell extracts were then prepared. Protein concentrations were determined using the BCA Protein Assay Reagent (Pierce, Rockford, IL). Samples were electrophoresed through 4-20% Tris-HCl Linear Gradient Gels (Bio-Rad, Hercules, CA) and transferred to PVDF membranes. Western blots were processed using the ECL Chemiluminescence System (Amersham, Arlington Heights, IL) and bands were quantitated using densitometry.
Quantitation of c-Jun activation E16 hippocampal neurons harvested from CF-1 mice were plated on glass coverslips as described above or on Tissue Tek chamber slides and at 6-7 DIV were treated with 5-10 mM oligomeric Ab for 12 hours, then fixed in fresh 4% PFA and immunostained with phospho-c-Jun specific antibody and DAPI or Hoechst to label total nuclei [35]. Activation of c-Jun was quantitated by counting p-c-Jun positive nuclei as a percent of total neuronal nuclei.

Electrophysiological recordings
Hippocampal slice preparations were performed as described [21]. Hippocampi were harvested from 3 month old male mice (C57Bl/6; The Jackson Laboratory). Transverse hippocampal slices (400 mm in thickness) were cut and maintained in an interface chamber at 29uC, and perfused with saline solution (124.0 mM NaCL, 4.4 mM KCL, 1.0 mM Na 2 HPO 4 , 25.0 mM NaHCO 3 , 2.0 mM CaCL 2 , 2.0 mM MgSO 4 , and 10 mM glucose) continuously bubbled with 95% O 2 and 5% CO 2 . fEPSPs were recorded from the CA1 region of the hippocampus by placement of both the stimulating and the recording electrodes in the CA1 stratum radiatum. Basal synaptic transmission (BST) was assayed by plotting of the stimulus voltage (V) against slopes of fEPSP to generate inputoutput relations or by plotting of the peak amplitude of the fiber volley against the slope of the fEPSP to generate input-output relations. LTP was induced using theta-burst stimulation (4 pulses at 100 Hz, with the bursts repeated at 5 Hz, and each tetanus including 3 10-burst trains separated by 15 seconds). LM11A-31 (100 nM) (dissolved in H 2 O) or vehicle (H 2 O) was added to the bath solution for 20 minutes prior to the induction of LTP in studies in which Ab (200 nM) was added to slice preparations. Oligomeric Ab(1-42) was prepared using HFIP [45].

Statistical analysis
Statistical analyses applied ANOVA with Dunnett's correction except for the following: in Fig. 2 and Fig. 5B ANOVA with Tukey-Kramer correction was applied; and in Fig. 6 two-way ANOVA was applied. For all bar graphs, mean6SE is shown. For all figures: *p,0.05, **p,0.01, ***p,0.001. Figure S1 Small molecule structures. Structures of LM11A-31, LM11A-24 and LM11A-36 are shown. LM11A-36 is identical to LM11A-24 except that it contains two additional methyl groups and is inactive. These compound structures were published previously [17]. Found at: doi:10.1371/journal.pone.0003604.s001 (0.10 MB TIF) Figure S2 AFM imaging of Ab preparations. Images are representative ,161 mm scans with z-height 10 nm. Incubation Ab  peptide in NaOH (A) or NH4OH (B) demonstrates primarily oligomeric structures with occasional fibrillar structures. (C) Ab preparation derived from HFIP-processing of Ab followed by PBS incubation demonstrates oligomeric structure with no observed fibrils. (D) Ab fibril preparation derived from HFIPprocessing of Ab followed by HCl incubation demonstrates primarily fibril-like structures.  Figure S4 Illustration of neurite curvature quantitation. As described in Methods, neurites were traced manually in the form of a series of short vectors (small arrows along neurites) to determine the angles (ai, ai+1…) created by each vector and a line connecting the origin and termination of the measured segment. The differences between successive angles (e.g. ai+1-ai) were averaged to generate a mean differential curvature (MDC) score. The upper panel demonstrates a neurite in culture medium alone (CM) with a mean differential curvature of 10.7 and the lower panel demonstrates a neurite exposed to Ab with a mean differential curvature of 29.3. Found at: doi:10.1371/journal.pone.0003604.s004 (1.92 MB TIF)