Targeting autotaxin impacts disease advance in the SOD1‐G93A mouse model of amyotrophic lateral sclerosis

1 INTRODUCTION

Despite the diverse aetiology of the disease, the sporadic and familial forms of amyotrophic lateral sclerosis (ALS) share excitability alterations in motor circuits as a central hallmark [1]. Early symptoms in both variants of the disease consist of fasciculation, cramps, hyper-reflexia and/or spasticity, which are commonly associated with motor neuron (MN) hyper-excitability [1-4]. In fact, hyper-excitability is a valuable diagnostic biomarker to differentiate a number of ALS-like patterns and to track disease advances in pre-symptomatic cases of familial ALS [4]. Furthermore, in the epidemiological framework, membrane hyper-excitability has been suggested as a consistent predictor for shorter survival rates in ALS patients [5]. Outstandingly, altered excitability of MNs is postulated to be causally related to their selective degeneration in ALS [2, 6]. In the arena of the current debate on the deleterious/protective impact of hyper-excitability in ALS, a line of argument supports that intrinsic hyper-excitability of MNs promotes the excitotoxic route [7, 8]. In this line, agents that reduce intrinsic membrane excitability (IME) were neuroprotective for MNs exposed to an excitotoxic challenge and also in ALS mice [8]. Hyper-excitability has been alternatively postulated as a beneficial transitory state in ALS, by contributing to the survival of S-type MNs, the MN subtype resistant to degeneration during disease progression [9-12]. In addition, an inconsistency with the hypothesis that hyper-excitability causes MN death in ALS is that FF-type MNs, the most vulnerable in ALS mouse models, exhibit hypo-excitability prior to degeneration [13]. Anyway, factors that impact MN excitability might represent potential targets to affect disease progression, partners in etiopathogenic mechanisms and/or biomarkers in ALS.

Against this background, the extracellular enzyme autotaxin (ATX), encoded by the ectonucleotide pyrophosphatase/phosphodiesterase 2 (enpp2) gene, emerges as a plausible candidate to control MN IME and, therefore, it can be postulated as a promising target to modulate disease progression in ALS. ATX is an ectoenzyme enriched in the brain with a pronounced lysophospholipase D activity, being the primary source of the lipid signalling molecule lysophosphatidic acid (LPA) [14, 15]. Thus, the majority of biological effects of ATX are associated with its ability to synthesize bioactive LPA from lysophosphatidylcholine (LPC) which, in turn, mainly stems from the cleavage of the membrane phospholipid phosphatidylcholine (PC) by phospholipase A2 (PLA2) [15]. In the CNS, ATX participates in several physiological processes such as brain and oligodendrocyte development and microglial activation. However, dysregulation of the ATX-LPA axis may contribute to many CNS disorders comprising inflammation, brain cancer, multiple sclerosis and Alzheimer's disease to quote some [15]. In the context of our working hypothesis, several findings underpin that ATX could contribute to disease pathogenesis: (i) levels of PC and LPC are augmented in the ventral horn of spinal cord tissue samples obtained from ALS patients [16]; (ii) PC is, indeed, the most abundant and discriminant lipid in cerebrospinal fluid of ALS patients [17]; (iii) obesity, a condition in which the ATX-LPA axis becomes dysregulated at both, systemic and CNS levels, is a predictor of better long-term survival of ALS patients [18-20]; (iv) the most prevalent LPA receptor in MNs, LPA1, is upregulated in the spinal cord and MNs from the transgenic SOD1-G93A mouse model of ALS preceding MN loss [21, 22] and (v) interfering with LPA1 affects MN IME and is neuroprotective in SOD1-G93A mice [22]. These findings, therefore, provide support to our main hypothesis and raise the possibility that targeting ATX could affect ALS progression.

In this report, we show that targeting ATX with an orally bioavailable specific inhibitor, PF-8380, leads to a reduction in MN IME in vitro by involving LPA1 signalling. In addition, we demonstrate that mRNAatx is upregulated in the spinal cord of SOD1-G93A mice before MN death and in cultured MNs from SOD1-G93A embryos. Finally, we provide evidence that chronic treatment with the ATX inhibitor slows down disease progression in ALS mice by a mechanism of action, at least in part, involving LPA1 knockdown in MNs.

2 METHODS

Experiments were performed in neonatal (P5–P9, either sex) Wistar rats; young mice (1–2-month-old male) either SOD1-G93A (Jackson Laboratory, Bar Harbor, ME, USA) and non-transgenic (non-Tg) littermates; and embryos extracted from SOD1-G93A pregnant mice (12.5 days gestation). Animals were provided by the local Animal Supply Services (SEPA, University of Cadiz), and their care and handling followed the guidelines of the European Union Council (2010/63/EU, 86/609/UE). Experimental procedures were approved by the local Animal Care and Ethics Committee (University of Cadiz, Cadiz, Spain) and the Ministry of Agriculture, Fisheries and Rural Development (Junta de Andalucía, Spain). Animals were individually housed in cages with water and food pellets available ad libitum, at 21 ± 1°C, with a 12-h light/dark cycle. All efforts were applied to minimize the number of animals used and their suffering.

2.1 Whole-cell patch-clamp

Current-clamp recordings were performed in hypoglossal MNs (HMNs) from brainstem slices (P5–P9 rats) as previously described [22]. Briefly, pups were guillotined under anaesthesia by hypothermia (10–15 min at 4°C) and their brainstems rapidly removed and dissected in carbogen (95% O2, 5% CO2) bubbled artificial cerebrospinal fluid (aCSF) enriched with sucrose at 4°C (in mM: 26 NaHCO3, 10 glucose, 3 KCl, 1.25 NaH2PO4, 2 MgCl2 and 218 sucrose). Coronal slices (300–400-µm thick), obtained using a vibroslicer (NVSL; WPI), were transferred to normally oxygenated aCSF (in mM: 26 NaHCO3, 10 glucose, 3 KCl, 1.25 NaH2PO4, 6 MgCl2, 130 NaCl and 0.5 CaCl2) and stabilized at ~37°C for 1 h. Slices were then relocated to a recording chamber for electrophysiological assays.

Recordings of HMNs were carried out under perfusion (~3–4 ml min−1) with normally oxygenated aCSF at 31°C. The patch was performed at the HMN somata, using a Nikon (Tokyo, Japan) Eclipse CFI60 microscope equipped with an infrared differential interference contrast system. Patch pipettes (1.5–3 MΩ resistance) were filled with the following internal solution (in mM): 17.5 KCl, 122.5 K-gluconate, 9 NaCl, 1 MgCl2, 10 HEPES, 0.2 EGTA, 3 Mg-ATP and 0.3 GTP-Tris at pH 7.4. Recordings were obtained and low-pass Bessel filtered at 10 kHz with a MultiClamp 700B amplifier. Data were digitized at 20 kHz with a Digidata 1332A analogue-to-digital converter and acquired using pCLAMP 10.0 software (Molecular Devices, Foster City, CA). Only cells exhibiting small (<20 MΩ) and stable (change <20%) access resistance were analysed. Series resistance was usually compensated 65–75%. The pipette offset potential was neutralized just before HMNs were patched. Recordings were not adjusted for liquid junction potential.

To analyse HMNs IME, resting membrane potential (Vm), threshold current (Ith), threshold for action potential (AP) generation (AP threshold) and input resistance (RN) were measured [22]. Ith was determined as the lowest depolarizing current pulse (5 ms) required to evoke an AP in 50% of cases. The AP threshold was estimated as the membrane potential corresponding to 5% of the first peak amplitude in the second derivative (d2V/dt2) obtained from the waveform as previously described [22, 23]. RN was calculated from the current–voltage (I–V) plots obtained by injecting a series of depolarizing and hyperpolarizing current pulses (0.5 s, −0.2–0.2 nA). The resulting data points were then fitted with a regression line, and RN was the slope of the lines.

2.2 Quantitative real-time reverse transcriptase PCR (qRT-PCR)

These measures were performed following a well-established protocol in our lab [8, 21, 22]. Briefly, total RNA was extracted either from the lumbar spinal cord or primary cultures of embryonic spinal cord MNs (SMNs) (100,000 cells/well), using the TRIsure Isolation Reagent (BioLine). Ensuing treatment with the RNase-free DNase set (Invitrogen), following the manufacturer recommendations, was performed to minimize DNA impurities. The concentration and purity of RNA samples were determined by spectrophotometry at 260 and 280 nm. Retrotranscription was done using random hexamers, 500 ng of total RNA as a template and iScriptTM cDNA Synthesis Kit (Bio-Rad). For real time RT-PCR, each specific gene product was amplified with the MiniOpticonTM System (BIO-RAD) using iTaqTM Universal SYBR Green Supermix (Bio-Rad). The cDNA levels for the different samples were determined using the 2−ΔΔCt method, being gapdh the housekeeping gene. The PCR primers sequences were as follows: for atx, forward 5′-CAGGTATGTCTTGAGGGTCAGAA-3′, reverse 5′-CCACTACTACAGCATCATCACCA-3′; for gapdh, forward 5′-AGAACATCATCCCTGCATCCA-3′, reverse 5′-AGATCCACGACGGACACATTG-3′. All analyses were performed in triplicate, with each experiment repeated at least twice.

2.3 Primary cultures of SMNs

Primary cultures of SMNs were prepared from the spinal cord of mouse embryos at 12.5 days of gestation (E12.5) following an established protocol [8, 21, 22]. Briefly, dissected ventral cords were first chemically (0.025% trypsin in glucose-HEPES buffer solution supplemented with 20 i.u. ml−1 penicillin and 20 mg ml−1 streptomycin) and mechanically dissociated to be subsequently collected under a 4% bovine serum albumin cushion. Centrifugation (10 min, 520 × g) on an Iodixanol density gradient (OptiPrep, Axis-Shield, Oslo, Norway) allowed to isolate the largest cells, which were once more centrifuged through a bovine serum albumin cushion. Isolated cells were pooled in a tube containing culture medium and plated at various densities depending on the type of experiment. Cultured SMNs were clearly identified by SMI32 immunofluorescence or by morphological criteria [8, 21] (Figure S1). SMNs were cultured in a Neurobasal medium (Gibco, Invitrogen, Paisley, UK) supplemented with B27 (Gibco; Invitrogen), horse serum (2% v/v), L-glutamine (0.5 mM) and 2-mercaptoethanol (25 µM; Sigma-Aldrich) and a cocktail of neurotrophic factors (NTFs): 1 ng ml−1 brain-derived neurotrophic factor, 10 ng ml−1 glial cell line-derived neurotrophic factor, 10 ng ml−1 ciliary neurotrophic factor and 10 ng ml−1 hepatocyte growth factor (PreProtech, London, UK). Isolated SMNs were plated in four well tissue culture dishes (Nunc, Thermo Fisher Scientific, Roskilde, Denmark).

2.4 Immunohistochemistry

As previously described [8, 21, 22], animals, under deep anaesthesia (7% chloral hydrate), were intraventricularly injected with heparin and immediately perfused transcardially with phosphate-buffered saline (PBS) followed by 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB), pH 7.4, at 4°C. Lumbar spinal cords (L3–L5) were rapidly dissected and post-fixed for 2 h in 4% PFA. Specimens were cryoprotected by immersion in 30% sucrose in 0.1 M PB (4°C, overnight). Serial coronal sections (30-µm thick), obtained by means of a cryostat, were kept at −20°C in a cryoprotectant solution (glycerol/PBS, 1:1 v/v) until processing.

For immunolabelling, after PBS washing, sections were immersed in a blocking solution [2.5% (w/v) bovine serum albumin, 0.25% (w/v) sodium azide and 0.3% (v/v) Triton X-100 in PBS] for 30 min, followed by incubation (4°C, overnight) with the polyclonal primary antibodies against LPA1 (1:200; Santa Cruz Biotechnology, Cat# sc-22207, RRID:AB_2135260) developed in goat and/or SMI32 (1:8,000; Covance Research Products Inc., Cat# SMI-32R-500, RRID:AB_509998) developed in mouse. Next, the tissue was rinsed with PBS and incubated for 1 h at room temperature with Cy3- and Cy5-conjugated anti-mouse and/or anti-goat IgGs or biotinylated anti-mouse IgG (1:250; Jackson ImmunoResearch Laboratories) as secondary antibodies. Sections for immunofluorescence were washed with PBS and mounted on slides with a solution containing propyl gallate (0.1 mM in PBS/glycerol, 1:9 v/v). Sections were analysed using an Olympus FV1000-MPE confocal microscope for fluorescence microscopy (Olympus, Japan). Images were acquired through a z-plane in which maximum antibody penetration was evidenced. The pinhole opening was 1 Airy unit. For comparison between different experimental conditions, the acquisition setting was kept identical. Animals and tissue were processed in parallel. Since these experiments were performed in parallel with others aimed to evaluate the effects of LPA1 inhibition, vehicle-treated animals were the same as in our previous report [22]. LPA1-immunoreactivity in the ventral horn was examined following analytical methods recently described [24]. Briefly, images were processed for background subtraction to obtain the maximum dynamic range of greyscale (from 0 to 250) and analysed using the software provided by Olympus. In all cases, the area delimiting the ventral horn of the lumbar spinal cord was manually traced, and for each region, the mean and integrated fluorescence intensity was measured. For a more detailed inspection of LPA1-immunoreactive patches, images for quantification were flattened and background-filtered to enhance cluster outlines and a user-defined intensity threshold was applied to select patches as previously described [22, 24, 25]. To delineate LPA1-immunoreactive structures, images were subjected to a smoothing filter (same for all) and subsequently were binarized based on local differences in intensity. Only patches reporting an area of >0.04 μm2 were taken as specific. To determine the proportion of patches that co-localize with SMI32-positive structures, a grid composed of 16 squares (10 µm × 10 µm) was distributed along with the ventral horn, and then, eight randomly selected squares were analysed in each section. Regions of interest were acquired in sections obtained from at least three animals per condition.

For MNs quantification in the L3–L5 segments, biotin was detected by means of the avidin–biotin–peroxidase system (Pierce, Rockford, IL, USA) using as chromogen 3,3-diaminobenzidine tetrahydrochloride. Sections were mounted on slides, dehydrated, covered with DePeX and visualized under light microscopy using an Olympus IX81 inverted microscope. The number of neurons with a well-defined nucleus was quantified in one of each three serial sections through the L3–L5 segments of the lumbar spinal cord. SMI32-identified MNs were counted in 30 coronal sections of the lumbar spinal cord from 4-month-old mice. All quantifications were carried out by researchers blinded to the treatment groups. In all cases, omission of the primary antibodies resulted in no detectable staining.

2.5 Assessment of phenotypic disease advance in SOD1-G93A mice

These procedures were performed as previously described [8, 22]. Briefly, transgenic males were received at the pre-symptomatic stage (P30). Then, habituation sessions for rotarod and grip strength began at P45. One or two valid measures were obtained before drug administration beginning at P60. Behavioural studies were carried out by researchers blinded for treatments. Animals receiving vehicles were the same that were used in our previous work [22] since both groups of experiments were performed in parallel.

The body weight for each mouse was weekly recorded. Individual weights were normalized taking the mean value of initial measures as 100%. At the beginning of the study, mean body weight was similar for the different cohorts used in each experimental series (8th week: vehicle: 25.8 ± 0.4 g, PF-8380: 25.1 ± 0.4 g, p = 0.218, Student t-test). Survival was analysed by means of Kaplan–Meier survival analysis with log-rank tests for statistical significance. The end-stage was defined as the day when mice were unable to right themselves 30 s after being placed on aside. At this time point, animals were euthanized using a carbon dioxide-enriched atmosphere.

Evaluation of motor function was assessed by rotarod, runtime and grip strength tests. Rotarod test was carried out to judge motor coordination, strength and balance. Animals were daily trained for 1 week before data acquisition. Mice were placed individually on a cylinder (3.4 cm in diameter), rotating at a constant speed of 15 rpm. The longest latency without falling for each animal was recorded once per week, after three attempts to keep onto the rotarod (Rotarod LE8500, Panlab SA, Barcelona, Spain). The cut-off time was taken at 180 s. In each experimental series, only mice that spent at least 180 s in rotarod before beginning drug administration were included in the study. Runtime was the time required by the mice to run along a track of 50 cm length and was weekly determined. Muscle strength of forelimbs was monitored on a grip strength meter (Grip Strength Meter, BIOSEB, Chaville, France) twice a week for each mouse. The animal apprehended by the tail, once confirmed that its forelimbs were gripping a T-bar, was pulled upwards at a 60° angle. The T-bar was coupled to a force transducer and grip strength was measured in grams. The highest outcome of three trials per mouse was recorded for each session. The mean value/week for each mouse was taken to construct the time course of grip strength for the different cohorts.

2.6 Oligonucleotides administration, drugs, and treatments

A small-interfering RNA directed against mRNA for lpa1 (siRNAlpa1) and a non-targeting siRNA (cRNA), taken as control, was administered as previously described [22]. The target sequence for the siRNAlpa1 was 5′-UCAUUGUGCUUGGUGCCUU-3′ the same as in our previous report [21]. This oligonucleotide robustly reduced mRNAlpa1 levels but did not affect the other isoreceptors (mRNAlpa2-6) [21]. Briefly, neonatal rat pups (P5) received a single injection of oligonucleotides [21, 22]. Anaesthetized animals (as above) were placed in a stereotaxic instrument (David Kopf Instruments, Tujunga, CA). siRNAlpa1 or cRNA (2 µg per rat; Accell, Dharmacon Inc., Lafayette, CO) dissolved in 2 µl of RNase-free PBS was injected at a rate of 0.5 µl min−1 by means of a microsyringe (5 µl, Hamilton Company, Tokyo, Japan). The needle end was visually guided by means of a surgical microscope to deliver oligonucleotides into the fourth ventricle. After the injection, the needle was left in place for 5 min, then slowly removed. Subsequently, the incision was sutured, cleaned with an aseptic solution (povidone-iodine) and the animals were allowed to survive 2–4 days for electrophysiological recordings. All animals received a post-operative injection of penicillin (20,000 i.u. kg−1; i.m.) in order to prevent infection. Pirazolone (0.1 mg kg−1; i.m.) was given on awakening for post-operative analgesia.

As previously described [22], for patch-clamp recordings, brainstem slices, at the level of the HN, were initially perfused for 10 min with normal aCSF to obtain baseline control data (Before condition). Next, slices were superfused for 10–15 min with aCSF supplemented with PF-8380 (1 µM) before voltage responses were acquired again. Finally, a last round of acquisition was taken after a 10 min washout with drug-free aCSF. Since electrophysiological effects of the drug on HMNs were all reversed to before-like values, these data have been omitted for clarity.

For treatment with PF-8380 (Sigma-Aldrich, St. Louis, USA) of adult mice, the drug was dissolved in sterilized drinking water supplemented with sucrose (1%). Each animal received daily 5 ml of the drug/vehicle solution to ensure that each mouse received the appropriate dosage (1 mg kg−1 d−1). Vehicle (1% sucrose) and PF-8380 containing solutions were freshly prepared each 3 days. Treatments began at 2 months of age.

2.7 Statistical analysis

Summary data are all presented as mean ± SEM. The number of analysed specimens per experimental paradigm, and statistical tests applied to each data set are indicated in figure legends or in Section 3. Statistical analysis was performed using SigmaPlot (Systat Software, Inc.). The minimum significance level was set at p < 0.05. Statistical tests were employed for all data sets with similar variance. For comparison between two groups, normally distributed data were analysed by unpaired or paired Student t-test, unless otherwise stated, while non-parametric data sets were assessed by Mann–Whitney U-test. One-way or two-way ANOVA followed by post hoc Holm-Sidak method was employed for comparison of three or more groups that passed the normality test. No data points were excluded from the statistical analysis unless otherwise noted.

3 RESULTS 3.1 PF-8380 reduces HMN IME via LPA1

Since baseline LPA1-mediated signalling maintains HMN IME in the brainstem slice preparation [22], it is highly probable that endogenous activity of the major LPA-synthesizing enzyme ATX also impacts HMN IME in this experimental condition. Thus, we first tested whether the ATX inhibitor PF-8380 [26] affects HMN IME in coronal slices obtained from rat pups (P5–P9). Whole-cell patch-clamp recordings of HMNs showed that the addition of the ATX inhibitor (1 µM for 10 min) to the bath solution produced a strong hyperpolarization (−11.2 ± 3.6 mV) of the membrane (Figure 1A,B; Table 1). This alteration was accompanied by an increase (+0.2 ± 0.1 nA) in the current threshold (Ith) to elicit an AP. However, the AP voltage threshold (AP threshold) was unaltered (+0.4 ± 2.1 mV) by the ATX inhibitor (Figure 1A,C,D; Table 1). These outcomes indicate that the PF-8380 induced reduction of Ith is mainly linked to the hyperpolarization shift observed for Vm rather than to a depolarization shift in the AP threshold. Finally, PF-8380 induced an abrupt reduction (−28.4 ± 6.0 MΩ) in HMN RN, a passive membrane property reflecting the electrical resistance of the whole surface membrane, a key determinant in the neuron's IME (Figure 1A,E; Table 1). Electrophysiological effects of PF-8380 on HMNs were all reversed after 10 min of washing and have been omitted for clarity. Therefore, PF-8380 led HMNs to a transitory hypo-excitable state, indicating that baseline activity of ATX, in our experimental conditions, contributes to maintain/modulate HMN IME.

image

PF-8380 affects the intrinsic membrane properties of HMNs. (A) Superimposed graded (AP failure) and AP voltage responses to depolarizing current pulses delivered at threshold intensities were recorded in a HMN from a P7 rat before and following 10 min exposure to PF-8380 (1 µM). The currents required to reach the threshold for AP generation (Ith) in each condition are shown at the bottom of the traces. Insets, voltage responses to step hyper-polarizing currents applied to determine RN in each condition. (B–E) Vm (B), Ith (C), AP threshold (D) and RN (E) data values of HMNs recorded before and after perfusion with PF-8380. The number of analysed HMNs in each group is in parentheses. Error bars, SEM. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant; by paired Student t-test

TABLE 1. Effect of the ATX inhibitor PF-8380 on intrinsic membrane properties of HMNs Treatment Vm (mV) Ith (nA) AP threshold (mV) RN (MΩ) Before −62.0 ± 2.6 0.824 ± 0.1 −22.4 ± 3.0 122.9 ± 9.1 PF-8380 −73.2 ± 3.9* 1.042 ± 0.1** −22.1 ± 3.5ns 94.6 ± 6.8*** *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant; by paired Student t-test.

As we previously reported [22], the ATX-synthesized lysophospholipid LPA affects HMN IME by acting on the G-protein-coupled receptor LPA1 which, subsequently, modulates the background K+ channel TASK1. Therefore, to reinforce the idea that baseline ATX activity controls HMN IME throughout LPA1-mediated signalling, we tested the effects of PF-8380, as above, on HMNs from rat pups that received at P5 microinjection of siRNAlpa1 or cRNA in the fourth ventricle. In the cRNA condition, PF-8380 induced expected Vm hyperpolarization, increase in Ith, and reduction in RN, without affecting AP threshold (Figure 2; Table 2). Strikingly, alterations induced by the ATX inhibitor were fully absent in HMNs from animals treated with siRNAlpa1 (Figure 2; Table 2). Given that this protocol efficiently knockdowns LPA1 without affecting the expression of the other isoreceptors [21, 22], these outcomes strongly support that endogenous ATX impacts HMN IME via, at least, LPA1.

image siRNAlpa1 avoids PF-8380 effects on analysed intrinsic membrane properties of HMNs. As in Figure 1, but recordings were performed in HMNs obtained from rat pups that received at P5 a single injection of cRNA or siRNAlpa1 into the fourth ventricle. Vm (B), Ith (C), AP threshold (D) and RN (E) data values of HMNs were recorded before and after perfusion with PF-8380 receiving the stated oligonucleotides. The number of analysed HMNs in each group is in parentheses. Error bars, SEM. *p p p t-test TABLE 2. Effect of the ATX inhibitor PF-8380 on intrinsic membrane properties of HMNs from rat pups receiving a single injection of the indicated oligonucleotides at P5 Treatment Vm (mV) Ith (nA) AP threshold (mV) RN (MΩ) cRNA Before −59.8 ± 2.0 0.661 ± 0.1 −31.3 ± 1.9 81.3 ± 6.4 PF-8380 −65.5 ± 3.7* 0.773 ± 0.1* −28.5 ± 2.5ns 67.2 ± 3.9* siRNAlpa1 Before −66.4 ± 2.3 0.705 ± 0.2 −40.0 ± 4.5 66.2 ± 8.8 PF-8380 −70.6 ± 3.4ns 0.790 ± 0.2ns −39.9 ± 5.8ns 64.7 ± 8.2ns *p < 0.05, ns, not significant; by paired Student t-test. 3.2 Upregulation of mRNAatx levels in the spinal cord of SOD1-G93A mice

Next, to gain support for a feasible contribution of ATX in disease progression in the mouse model of ALS, qRT-PCR analysis of mRNAatx expression was performed in the lumbar spinal cord of SOD1-G93A mice and their non-Tg littermates at different stages of disease progression. Noticeably, transcripts were already upregulated in transgenic samples extracted from 1-month-old mice (SOD1-G93A: 126.0 ± 5.8%, non-Tg: 102.1 ± 5.0%), the first time point tested (Figure 3A). mRNAatx levels rose in 2-month-old animals (SOD1-G93A: 159.2 ± 13.0%, non-Tg: 106.8 ± 7.8%) and declined to non-Tg-like levels in 3-month-old (SOD1-G93A: 110.2 ± 7.9%, non-Tg: 109.1 ± 8.2%) and 4-month-old (SOD1-G93A: 111.0 ± 4.9%, non-Tg: 110.0 ± 4.7%) ALS mice (Figure 3A). Thus, the increase of mRNAatx expression in the spinal cord of SOD1-G93A mice occurs at pre-symptomatic (1–2-month-old) stages, therefore preceding MN loss, but dropped to a control-like condition at early symptomatic (3-month-old) and symptomatic phases (4-month-old) (Figure 3A), when MN loss is already evident. At this point, it is interesting to remark that the increase in mRNAatx levels also occurred prior to the enhancement of mRNAlpa1 observed in these mice [22]. This supports a gain in the ATX-LPA1 signalling axis that precedes MN loss in ALS mice.

image

Changes in the expression levels of mRNAatx in the spinal cord of ALS mice and in SMNs from SOD1-G93A embryos. (A) Time course of changes in mRNAatx expression in the spinal cord of SOD1-G93A mice and their non-Tg littermates. Values were normalized taking the mean value for 1-month-old non-Tg mice as 1. (B) mRNAatx expression at the indicated days-in-vitro, determined by qRT-PCR, in SMNs isolated from non-Tg and SOD1-G93A embryos. gapdh was the housekeeping gene. mRNAatx values were normalized relative to the mean value obtained at 4 DIV for non-Tg SMNs. The number of mice (A) or SMNs extractions (B) in each group is in parentheses. Error bars, SEM. *p < 0.05, **p < 0.01, ***p < 0.001; by two-way ANOVA with post hoc Holm-Sidak method. Statistic outputs between genotypes are displayed in plots

To explore the possibility that MNs were feasible sources for mRNAatx enhancement, and given that ATX is an ectoenzyme, we measured and compared transcript levels in primary cultures of SMNs isolated from SOD1-G93A and non-Tg mouse embryos at two different differentiation days in vitro (4 DIV and 6 DIV). These time points were selected based on our preceding report, which shows that in SMNs isolated from SOD1-G93A mice the expression levels of mRNAlpa1 are normal at 4 DIV and abruptly increase at 6 DIV relative to non-Tg-cultured cells [22]. Interestingly, mRNAatx levels were significantly higher in SMNs isolated from SOD1-G93A (156.1 ± 7.7%) than in those extracted from non-Tg (100.0 ± 16.6%) embryos at 4 DIV (Figure 3B). The difference was even more pronounced (non-Tg: 91.6 ± 3.9%, SOD1-G93A: 224.2 ± 20.5%) at 6 DIV (Figure 3B). Altogether, the outcomes indicate that alterations in mRNAatx levels precede changes in mRNAlpa1 expression both, in the lumbar spinal cord of SOD1-G93A mice and in SMNs extracted from transgenic embryos during the differentiation progress [22]. Furthermore, it can be suggested that MNs could be a feasible origin for mRNAatx upregulation observed in ALS mice.

3.3 PF-8380 delays MN loss in the ALS SOD1-G93A model

The increase of LPC, the main ATX substrate, in the ventral horn of ALS patients [16], together with mRNAatx over-expression in the spinal cord of SOD1-G93A mice, highlights ATX as a pivotal partner in the pathogenesis of disease and as a target for further preclinical assessment. To explore this issue, we examined the potential neuroprotective effect of chronic in vivo treatment with the orally bioavailable ATX inhibitor PF-8380 in the SOD1-G93A mouse model. In this murine line, disease progression predominantly courses from hind to forelimbs. Therefore, we focused our in vivo treatment studies on MN survival in the motor pools located at the lumbar spinal cord. Feasible neuroprotection against MN degeneration was analysed by quantifying the number of SMI32-positive MNs in the L3–L5 lumbar segment from 4-month-old mice [8, 22]. Chronic administration of PF-8380 (1 mg kg−1 d−1) in the drinking water beginning at 2 months of age exerted beneficial effects on MNs in the ALS model. Under PF-8380 treatment, the number of MNs present in the lumbar spinal cord of SOD1-G93A mice was higher (+29.2 ± 11.5%) than in age-matched transgenic animals treated with vehicle (Figure 4). This finding indicates that treatment with the ATX inhibitor delays loss of lumbar MNs in the ALS model.

image Treatment with PF-8380 delays MN loss in ALS mice. (A) Representative microphotographs of lumbar spinal cord sections immunostained for SMI32 from 4-month-old SOD1-G93A mice that received vehicle or PF-8380 (1 mg kg−1 d−1) in the drinking water beginning at P60. Boxed areas indicate SMI32-positive MN pools. Scale bar: 65 µm. (B) Quantification of the number of SMI32-identified MNs in the L3–L5 spinal cord segments for each condition. Since these experiments were performed in parallel with others intended to evaluate the effects of the administration of the LPA1 inhibitor AM095 [22], the vehicle group is the same for both assays. The number of analysed mice in each group is in parentheses. Error bars, SEM. *p p p t-test 3.4 Benefits of PF-8380 on phenotypic disease progression in SOD1-G93A mice

We subsequently evaluated whether the neuroprotective effect of PF-8380 on SOD1-G93A MNs translates into a change in the phenotypic progression of the disease. For that purpose, forelimb grip strength, latency to fall from a rolling cylinder (Rotarod), time to cover a distance of 50 cm (Runtime), time for symptom onset, the evolution of body weight and survival time were analysed in the vehicle and PF-8380 treated groups of SOD1-G93A mice (Figure 5). As a starting point to construct the cumulative probability curve of symptom onset, we used the first day in which a mouse showed motor function deficits on the rotarod test (≥10% reduction in the time to fall) followed by progressive performance deterioration in the next sessions [22]. Noticeably, the PF-8380-treated group displayed better outputs in these tests in comparison with the vehicle-treated SOD1-G93A cohort (Figure 5). For example, taken 18-week-old point for comparison, PF-8380 significantly improved grip strength (+16.3 ± 6.7%), time in rotarod (+105.1 ± 23.8 s) and runtime (−10.2 ± 3.7 s) in comparison to outputs displayed by vehicle-receiving ALS mice (Figure 5A–C). In addition, PF-8380 delayed substantially symptom onset (+20.2 ± 2.9 days, difference between mean values), improved weight loss advancement and prolonged lifespan (+5.6 ± 3.0 days, difference between mean values) in SOD1-G93A mice (Figure 5D,E). In studies parallel to those presented herein, analysis of transgene expression in a random sampling of several SOD1-G93A cohorts subjected to diverse treatments did not show, in any case, differences in the levels of SOD1-G93A transcripts [8, 22]. It is therefore improbable that positive effects observed here can be as a result of different expression levels of the human SOD1-G93A in both cohorts. These findings support a critical role of ATX in the pathophysiology of this lethal neurodegenerative disease providing also preclinical validation of ATX knockdown and/or function inhibition as a potential therapeutic strategy in ALS.

image Neuroprotective effects of PF-8380 in the SOD1-G93A mouse model of ALS. (A) Time course of forelimb grip strength (in per ent) decline in vehicle- and PF-8380-treated SOD1-G93A mice. Strength measured at 9 weeks of age/mouse was taken as 100%. (B) Age-related changes in rotarod performance for both treatment groups. Rotarod test was performed to evaluate the time course of deficits in motor coordination, strength and balance as well as to construct the cumulative probability curve of symptoms onset in (D). (C) PF-8380 delayed runtime (time animals took to run a track of 50 cm) decline compared with vehicle. (D–F) Cumulative probability curves of symptom onset (D), the time course of mean body weight (E) and survival (F) for vehicle- and PF-8380-treated SOD1-G93A mice. The number of mice in each group is in parentheses. The vehicle groups are the same as those in Gento-Caro and colleagues [22]. Error bars, SEM. *p p p post hoc Holm-Sidak method (A–C, E). Log-rank test, Kaplan–Meier analysis was applied to (D and F). Statistic outputs comparing both treatments are displayed in plots

Whether higher doses of PF-8380 may lead to further improvement in disease progression in this transgenic mouse model of ALS merits further investigation. In conclusion, these outcomes point to ATX inhibitors as potential pharmacological candidates to be tested in translational clinical assays in ALS.

3.5 PF-8380 reduces LPA1 expression in SOD1-G93A MNs

It is interesting to remark at this point that LPA1 is upregulated in the ventral horn and in MNs at the lumbar spinal cord of this ALS model [22]. When comparing the time course of mRNAatx and mRNAlpa1 expression in ALS mice, it becomes apparent that mRNAatx over-expression precedes mRNAlpa1 upregulation in the lumbar spinal cord of SOD1-G93A mice (Figure 6A). Strikingly, rising levels of LPA following its intrathecal administration leads to LPA1-dependent LPA1 upregulation in dorsal root ganglions [27]. In addition, two different strategies, both leading to LPA1 knockdown in MNs, were neuroprotective in the ALS mouse model [22]. These findings led us to hypothesize that the mechanism of action by which chronic treatment with PF-8380 slows down disease progression in SOD1-G93A mice might, at least in part, involve LPA1 knockdown in the ventral horn and in MNs.

留言 (0)

沒有登入
gif