ABSTRACT
LPA1 is one of the six known receptors (LPA1-6) for lysophosphatidic acid (LPA). Constitutive Lpar1 null mutant mice have been instrumental in identifying roles for LPA-LPA1 signaling in neurobiological processes, brain development, and behavior as well as modelling human neurological diseases like neuropathic pain. Constitutive Lpar1 null mutant mice are protected from partial sciatic nerve ligation (PSNL)-induced neuropathic pain, however Lpar1 expressing cell types that are functionally responsible for mediating this protective effect are unknown. Here we report generation of a Lpar1flox/flox conditional null mutant mouse that allows cre-mediated conditional deletion combined with its use in a PSNL pain model. Lpar1flox/flox mice were crossed with cre transgenic lines driven by neural gene promoters for nestin (all neural cells), synapsin (neurons), or P0 (Schwann cells). CD11b-cre transgenic mice were also used to delete Lpar1 in microglia. PSNL-initiated pain responses were reduced following cre-mediated Lpar1 deletion with all 3 neural promoters but not the microglial promoter, supporting involvement of Schwann cells and central and/or peripheral neurons in mediating pain. Interestingly, rescue responses that were due to conditional deletion were non-identical, implicating distinct roles for Lpar1-expressing cell types. Our results with a new Lpar1 conditional mouse mutant expand an understanding of LPA1 signaling in the PSNL model of neuropathic pain.
Neuropathic pain is produced by nerve lesions or neurological conditions such as multiple sclerosis, diabetes, and cancer affecting an estimated 10% of the general population (1). Treatment options for individuals affected by neuropathic pain are limited and ineffective, often leading to a worsened condition and disability. Initiation and propagation of pain signaling occurs through afferent nerve fibers that relay peripheral signals through dorsal root ganglia (DRG) to signal centrally via the central nervous system (CNS) spinal cord dorsal horn and brain (reviewed in (2–4)). Neuropathic pain involves central sensitization, a process that results in allodynia (painful response to normally innocuous stimuli) and hyperalgesia (increased pain sensation to noxious stimuli) (5).
One identified modulator of neuropathic pain is the bioactive lipid lysophosphatidic acid (LPA). LPA normally signals through six known G protein-coupled receptors, Lpar1-6 (6), which are involved in myriad biological and pathological processes affecting most of the physiological systems in the body, including the nervous system (6–13). LPA1 is also expressed in the peripheral nervous system (PNS) and CNS. Schwann cells represent one of the LPA1 expressing cell types that may be involved in the induction of neuropathic pain. LPA signaling through this receptor influences Schwann cell morphology, migration, and survival (14,15). In-vivo, sciatic nerves of Lpar1 deficient mice show abnormalities including an increased number of apoptotic Schwann cells, reduced myelin thickness, and a proportionately lower number of small nerve fiber interacting Schwann cells (14,16). Neurons can also be affected through regulation of neuronal cell morphology, motility, growth cone collapse, calcium signaling, and proliferation (16–23). Mice deficient for this receptor display alterations in cortical development and neurogenesis and show behavioral abnormalities (22–24).
A role for LPA in pain sensation was first identified through intrathecal (i.t.) injection of LPA, where mice that received a single (i.t.) injection of LPA developed thermal hyperalgesia and mechanical allodynia (25). LPA-induced neuropathic pain was accompanied by other sequelae including demyelination in the dorsal root and increased expression of the pain associated markers, protein kinase C γ (PKCγ) in the spinal cord dorsal horn, and voltage-gated calcium channel Caα2δ1 in the DRG (25). Interestingly, i.t. injection of LPA also induced de novo production of LPA in the dorsal horn and dorsal root, implicating a feed-forward role in pain generation (26). De novo LPA production was also observed in the dorsal horn and dorsal root following PSNL (27–29). Wildtype mice subject to PSNL displayed pain behaviors similar to those of mice that received LPA i.t. and showed similar demyelination as well as upregulation of PKCγ and Caα2δ1 (25).
LPA’s effects in PSNL were shown to be receptor-dependent through the use of constitutive null receptor mutants. Lpar1 null mutant mice were protected from PSNL and i.t. LPA injection induced mechanical allodynia, and did not show accompanying increased levels of PKCγ and Caα2δ1 (25). Lpar5 null mutant mice were also protected from PSNL induced neuropathic pain, albeit through CNS mechanisms distinct from those of Lpar1 null mutants (30).
While Lpar1 null mutant mice are protected from PSNL-induced neuropathic pain, the cell types responsible for mediating this protection remain unclear. To address this issue, we generated an Lpar1 conditional null mutant mouse and targeted deletion of Lpar1 in all neural lineages, peripheral and CNS neurons, Schwann cells, and microglia/macrophages to identify the cell types responsible for mediating Lpar1’s protective effect in the PSNL neuropathic pain model.
Results
Generation of Lpar1 conditional null mutant mice
We selected a portion of the Lpar1 genomic locus for conditional gene targeting in embryonic stem (ES) cells resulting in the creation of a mutant mouse, designated Lpar1flox/flox, where Lpar1 exon 3 is selectively deleted in the presence of the cre recombinase (Fig. 1A). The targeting vector contained a loxP site that was introduced into a restriction enzyme site 5’ of exon 3, and a neomycin drug selection cassette flanked by loxP sites in a restriction enzyme site 3’ of exon 3. Following electroporation of the linearized Lpar1 targeting construct, drug selection, and screening of DNA isolated from selected ES cell clones by Southern blotting and hybridization, several clones with a homologously recombined Lpar1 allele were identified (Fig. 1B and 1C). PCR with primers flanking the 5’ loxP site was used to select ES cell clones for blastocyst injection. Mice positive for germline transmission of the recombined allele were then crossed with nestin-cre transgenic mice to produce Lpar1flox/flox-nestin-cre mice (31). Because nestin is expressed in the testis, male Lpar1flox/flox-nestin-cre mice were bred to C57BL/6J female mice to produce offspring with germline cre-mediated loxP site recombination. Selective deletion of the floxed neomycin cassette and retention of the 5’ loxP site in offspring were identified by PCR (Fig. 2A and 2B). Heterozygous Lpar1flox/+ mice with the correct recombination events were then crossed together to produce wildtype, Lpar1flox/+, and Lpar1flox/flox mice (Fig. 2C). A high level of embryonic lethality was observed for Lpar1 constitutive null mutant mice in a C57BL/6J background whereas Lpar1flox/flox mice in this background strain are healthy and are indistinguishable from wildtype littermates. Wildtype, Lpar1flox/+, and Lpar1flox/flox mice were identified by PCR (Fig. 2C) and are behaviorally the same.
Cre-mediated Lpar1 targeted deletion
To delete Lpar1 in all neural cell types, neurons, Schwann cells, and myeloid lineage cells, Lpar1floxflox mice were crossed to nestin, synapsin, P0, and CD11b-cre transgenic mice, respectively (31–34). To confirm that Lpar1 was deleted in the presence of cre, genomic DNA was isolated from DRG of Lpar1flox/flox and Lpar1flox/flox-nestin-cre mice and PCR was used to verify genomic recombination of the Lpar1 genomic locus to produce a null allele (Fig. 3A). DRG contain both neural and non-neural cells, with conditional deletion limited to neural cells, thus producing a recombined (neural) and unrecombined (non-neural) signal in conditional mutants. As expected, PCR products indicative of both an unrecombined and recombined Lpar1flox/flox allele can be amplified from genomic DNA isolated from Lpar1flox/flox-nestin-cre DRG, while only an unrecombined product can be produced from the DRG of control Lpar1flox/flox mice (Fig. 3A). In agreement with genomic deletion of Lpar1, RT-PCR showed Lpar1 mRNA transcripts are absent in Lpar1flox/flox-nestin-cre DRG (Fig. 3B). Following Schwann cell specific P0 cre crossing, PCR analyses of sciatic nerve showed deletion of Lpar1 (Fig. 3C), compared to wildtype. Neuronal deletion was confirmed in cerebral cortex (Ctx) of Lpar1flox/flox-synapsin-cre mice (Fig. 3D). Immunofluorescent labeling of peripheral myelinated axons for MBP (myelin in red) and satellite glia expressing LPA1 (green) in wildtype DRG (Fig. 3E) was not observed in Lpar1flox/flox-nestin-cre mice (Fig. 3F). These data demonstrate conditional deletion of Lpar1 in the presence of targeted cre recombinase expression.
Lpar1 expressing neural cell types contribute to PSNL-induced neuropathic pain phenotypes
To determine which Lpar1 expressing neural cell types mediate PSNL-induced neuropathic pain protection, paw withdrawal threshold responses following cre recombination for Lpar1flox/flox-nestin, Lpar1flox/flox-synapsin, Lpar1flox/flox-P0 and Lpar1flox/flox-CD11b-cre was assessed. No pain rescue was observed with Lpar1flox/flox-CD11b compared to wildtype or to Lpar1flox/flox transgenic mice (data not shown). All other genotypes rescued the pain phenotype compared to controls (Fig. 4A-D). Lpar1flox/flox-nestin-cre conditional mutant mice challenged with PSNL had similar paw withdrawal threshold responses compared to previously defined Lpar1 constitutive null mutant mice (25) (Fig. 4A). By contrast, Lpar1flox/flox-P0-cre mice initially responded like control mice at early time points (days 3 and 6), but then showed sustained protection at later time points (day 9 through day 21) (Fig. 4B and 4D). Lpar1flox/flox-synapsin-cre mice were initially refractory to PSNL-induced neuropathic pain (Fig. 4C) but lost protection over time (day 12 through day 21) (Fig. 4D). It is notable that the combined protection of P0 and synapsin-cre recombination approximated the protection produced by nestin-cre recombination (Fig. 4A), implicating an additive rescue effect produced by both Schwann cells and neuronal LPA1 activation in PSNL-initiated pain.
Discussion
Lpar1 conditional null mutant mice were generated and shown to undergo cre-mediated recombination, enabling identification of Lpar1-expressing neurons and Schwann cells as functionally important for the PSNL phenotype. In the absence of cre, Lpar1flox/flox mice developed a pain phenotype comparable to Lpar1 constitutive null mutant mice (25), demonstrating that this new floxed mutant gene functions normally in PSNL before cre crossing. Lpar1flox/flox-nestin-cre mice with a pan-neural lineage deletion of Lpar1 are protected from PSNL-induced neuropathic pain, supporting neural LPA1 signaling as important despite Lpar1’s ubiquitous tissue expression. By comparison, P0 and synapsin-cre recombination produced only partial rescue with complementary temporal phases of protection that appeared additive to account for the degree of rescue by nestin-cre recombination.
The actions of LPA1 in Schwann cells affecting PSNL phenotypes have not, to our knowledge, been previously reported, and the observed phenotype was unexpected with regard to the clear and differential time-dependence of the effect. Explanations for these temporal changes in pain protection may be due to differences in de novo synthesis of LPA and the varied activation states documented for LPA1 (8,11,35–38) that may occur in neurons and Schwann cells. Such LPA signaling effects could be altered by receptor removal to produce the time-course differences observed for PSNL-initiated pain rescue. Long-lasting protection from neuropathic pain at later time points may also reflect changes in nerve myelination that may interfere with the transmission of pain stimuli as previously suggested (14,25). Nerve fibers in Lpar1flox/flox-P0-cre mice may already be abnormally myelinated, and nerve injury induced demyelination may alter normal pain signal transmission. However, we note that the nerve fibers that respond to noxious stimuli are lightly myelinated Aβ fibers and unmyelinated C-fibers (2,3), requiring a more complex scenario that might involve central pain consolidation through myelinated fibers.
Effects of Lpar1 deletion from neurons in Lpar1flox/flox-synapsin-cre mice showed early protection in PSNL, contrasting with later protection of Schwann cell receptor deletion, while supporting the involvement of neurons in LPA1-mediated PSNL-induced pain. Synapsin-cre deletion is effective in deleting genes from CNS neurons (39) but less effective in peripheral (DRG) neurons (33,40), implicating central neuronal mechanisms. A possible explanation for rescue at early timepoints could involve a lack of de novo LPA synthesis from Lpar1 deficient neurons. LPA can be released by neurons following nerve transection and neurons can synthesize LPA de novo through an LPA receptor dependent feed-forward mechanism (as evidenced by LPA3) (15,26,41). De novo LPA synthesis from other cell types following PSNL may result in LPA accumulation to drive neuropathic pain at later time points, particularly through activation of Schwann cell receptors in the neuron-specific mutants. Alternatively, PSNL may cause damage and vascular leakage that exposes peripheral nerves to LPA, activating cognate receptors to produce aberrant pain signaling (9,42–44).
Other LPA receptor subtypes can contribute in distinct ways to neuropathic pain based on analyses of different LPA receptor-null mutants (13,16,25,30,37,45,46). Similar to Lpar1 null mutant mice, Lpar5 null mutant mice are protected from PNSL-induced neuropathic pain and also show decreased sensitivity to acute pain stimuli and faster recovery responses when challenged in an inflammatory pain model (30,47). Additionally, deletion of Lpar3 in mice prevents i.t. LPA-induced de novo production of LPA in the dorsal horn and dorsal root and also prevents LPA-induced allodynia and hyperalgesia (26), suggesting an Lpar3 mediated feed-forward mechanism for LPA in neuropathic pain initiation. Prevention of LPA de novo synthesis and neuropathic pain in the i.t. LPA and PSNL neuropathic pain models using minocycline combined with Lpar3 expression in microglia indicate that this feed-forward mechanism is likely mediated by microglia (48,49). In the present study, we did not observe a rescue effect of Lpar1 loss from microglia, suggesting that maintained Lpar3 could sustain PSNL-initiated pain. The actual cell types involved in the functions of these and other LPA receptor subtypes remain to be determined but is experimentally tractable through generation of conditional mutants.
The generated Lpar1 conditional mutant mice will be useful in identifying cell types involved directly with LPA1 signaling in neuropathic pain as well as many other conditions and disease models (6–12,37,45,50,51). The tractability of LPA1 as a member of the lysophospholipid receptor family supports its potential as a druggable GPCR target (8,10,35,36) for the development of improved therapies targeting specific Lpar1 expressing cell types.
Experimental procedures
Mice
All procedures performed on animals were IACUC approved and performed in accordance with the regulations of The Scripps Research Institute (TSRI) Department of Animal Resources and the Sanford Burnham Prebys Medical Discovery Institute animal care and use committees. Mice used in this study were nestin-cre (Jackson Laboratory Stock Number 003771), P0-cre (Jackson Laboratory Stock Number 017927), synapsin-cre (Jackson Laboratory Stock Number 003966), and CD11b-cre (obtained from Don Cleveland) transgenic lines.
Synthesis of the Lpar1 conditional gene targeting vector
Creation of the Lpar1 conditional gene targeting vector was accomplished by PCR amplification of mouse Lpar1 genomic fragments using a bacterial artificial chromosome (BAC RP23-149020 Children’s Hospital Oakland Research Institute (CHORI)) containing the Lpar1 genomic locus as a template. PCR amplification was performed using Pfx50 DNA polymerase (Invitrogen) and amplified genomic fragments were assembled into pBluescript II. During the process of assembly, a loxP site was inserted into a HindIII site 5’ of Lpar1 exon 3 and a neomycin cassette under the control of the phosphoglycerate kinase promoter (PGK-neo) flanked by loxP sites was inserted directionally (all loxP sites in the same orientation) into an XbaI site 3’ of Lpar1 exon 3 (Fig. 1A). The construct was engineered so that 3.4 and 6.7 kb of Lpar1 genomic DNA flanked the PGK-neo insertion site. To aid in cloning, BamHI and AatII restriction enzyme sites were added to the distal 5’ and 3’ ends of the Lpar1 genomic segment chosen for targeting vector design. An EcoRI restriction enzyme site was included in the loxP flanked PGK-neo cassette to identify ES cell clones containing an allele that recombined homologously with the targeting vector.
Production of Lpar1flox/flox and Lpar1flox/flox-cell type specific null mutant mice
To create the Lpar1flox/flox mice, 1 x 107 R1 ES cells were mixed with 50 μg of linearized Lpar1 targeting vector in a 0.4 cm electroporation cuvette and the cells were pulsed with a Bio-Rad Gene Pulser II (200 mVolts x 800 μF capacitance). The electroporated ES cells were plated on mitotically inactive mouse feeder cells and allowed to recover for 24 hrs at 37°C; 24 hrs after electroporation and plating, 150 μg/ml Geneticin (Invitrogen) was added to the ES cell medium and the cells were placed under selection for 7 days. ES cell clones were then isolated and grown individually for subsequent DNA isolation and screening for homologous recombination events by Southern blotting and hybridization with an Lpar1 DNA probe containing sequence external to that of the 5’ end of the Lpar1 targeting vector. Clones with homologous recombination events were then screened for retention of the loxP site 5’ to Lpar1 exon 3 with the following primers: 5’ loxP Forward 5’-gttgggacatggatgctattc-3’ and 5’ loxP Reverse 5’-aatctgttctcatcccacacg-3’. Correctly targeted ES cell clones were then injected into C57BL/6J blastocysts at the TSRI Murine Genetics Core.
To delete the loxP flanked PGK-neo cassette in-vivo, gene targeted mice were crossed to nestin-cre transgenic mice and resultant males were then bred to C57BL/6J female mice. Male mice were chosen because cre is expressed in the germline of nestin-cre male mice. Offspring were then screened by PCR for the presence of the 5’ loxP site with the primers listed above, in the presence or absence of the PGK-neo cassette with primers A1 Exon 3 Forward 5’-agactgtggtcattgtgcttg-3’ and Neo Reverse 5’-tggatgtggaatgtgtgcgag-3’, and for retention of the loxP site 3’ to Lpar1 exon 3 with primers 3’ loxP Forward 5’-tgcagaattatgagtggacagg-3’ and 3’ loxP Reverse 5’-ggtttagtggtgtgggatcg-3’. Mice that retained the loxP sites 5’ and 3’ to Lpar1 exon 3 but deleted the PGK-neo cassette were selected for propagation and crossing with nestin-cre, P0-cre, synapsin-cre, and CD11b-cre transgenic mice (31–34).
PCR genotyping of the Lpar1 conditional mutant mice was done with the following primers: 5’ loxP Forward 5’-gttgggacatggatgctattc-3’, 3’ loxP Reverse 5’-ggtttagtggtgtgggatcg-3’, and A1 Exon 3 Forward 5’-agactgtggtcattgtgcttg-3’. PCR amplification of genomic DNA with these primers identified wildtype, Lpar1flox, and Lpar1 deleted products of 316, 354, and 242 bp, respectively. Synapsin-cre, CD11b, P0-cre, and nestin-cre transgenes were identified by PCR amplification of genomic DNA with a common reverse PCR primer, (Cre Reverse 5’-CAG CAT TGC TGT CAC TTG GTC-3’), and forward primers specific for synapsin (SynCreForward 5’-CCCAAGAAGAAGAGGAAGGTG-3’), CD11b (CD11b Forward 5’-ACACCTCAGCCTGTCCAGTAG-3’), P0 (MPZ Forward (P0 Cre) 5’-ATT GGT CAC TGG CTC AAG AC-3’), and nestin (Nestin Prom: 5’-ACT CCC TTC TCT AGT GCT CCA-3’) yielding products of 350 bp, 1 kb, 525 bp, and 550 bp respectively.
Southern blotting and DNA hybridization
ES cell clones were screened for homologous recombination by digesting 10 μg of ES cell DNA with EcoRI, running the DNA on a 0.8% 1 x TAE agarose gel, and transferring the digested DNA to Nytran SuPerCharge membrane (GE Healthcare Life Sciences) in 20 x SSPE. Transferred DNA was UV crosslinked to the membrane and hybridized with a 32P-labeled (Prime-It II Random Primer Labeling Kit, Agilent) Lpar1 probe with sequence external to the 5’ end of the targeting vector. The 800 bp probe was produced by PCR from a BAC containing Lpar1 with the following primers: A1 Ext Forward 5’-actgaggtcacttactcagag-3’ and A1 Ext Reverse 5’-gtctatggctgtggaattcaag-3’. Probe hybridization was carried out overnight at 42°C in a .05 M pH 7.4 phosphate buffer containing 50% formamide, 5 x SSPE, 1 x Denhart’s, 1% SDS, containing .1% denatured 10 mg/ml salmon sperm DNA following a 1 hr pre-hybridization. Blots were washed and visualized using a phosphorimager. The presence of a 4.2 kb recombined band and a 7.9 kb wildtype band was indicative of ES cells with homologous recombination events.
Partial sciatic nerve ligation and behavioral testing
The partial sciatic nerve ligation (PSNL) procedure was performed as described (30). Adult Lpar1flox/flox and Lpar1flox/flox-cre transgenic mice in a C57BL/6J background were anesthetized via nosecone delivery of isoflurane and the right limb sciatic nerve exposed and tightly ligated with 10-0 fine sutures. The wound and skin were closed and stitched, and the animals allowed to recover. For behavioral testing, animals were acclimated in cages with wire mesh bottoms for one hour prior to testing in an environmentally controlled testing room. Paw withdrawal threshold (gram (g)) against increasing mechanical stimuli (0-50 g in 20 s) were measured before and following PSNL surgery with tests conducted four separate times with at least a 1 min interval between tests. The average response was normalized to pre-surgery controls +/- SEM.
Immunohistochemistry
DRG were isolated from the lumbar region of Lpar1flox/flox control and Lpar1flox/flox-conditional null-mutant mice. Tissues were embedded in OCT compound and 5 μM sections were cut and immunolabeled with antibodies to mouse LPA1 (PA1 10401, Thermo Fisher Scientific), MAG (clone 513 MAB-1567, Chemicon), MBP (ab134018 Abcam), and TUJ1 (MMS-435P, Covance). Secondary antibodies were used against the listed primary antibodies and 60x images were acquired on a Zeiss Axio Imager.D2 microscope.
Reverse transcription PCR
DRG were isolated from the lumbar region of Lpar1flox/flox and Lpar1flox/flox-nestin-cre conditional null-mutant mice. DRG were placed in 1 ml of TRIzol Reagent (Thermo Fisher Scientific) and total RNA was isolated according to the manufacturer’s directions. cDNA was synthesized from total RNA using a Bio-Rad iScript cDNA synthesis kit and β-actin and Lpar1 specific oligonucleotide primer pairs were used to amplify target gene transcripts. Primers used to amplify a 350 bp product from β actin cDNA were M β Actin Forward 5’-tggaatcctgtggcatccatg-3’ and M β Actin Reverse 5’-aaacgcagctcagtaacagtc-3’; primers used to amplify a 194 bp product from Lpar1 cDNA were M LPA1 Forward RT 5’-gacaccatgatgagccttctg-3’ and M LPA1 Reverse RT 5’-tcgcggtaggagtagatgatg-3’. An equivalent amount of cDNA from each sample, calibrated to produce equal amounts of β-actin PCR product, was used to amplify the Lpar1 cDNA.
Conflict of interest
The authors declare no conflicts of interest with the contents of this article.
FOOTNOTES
Funding was provided by the NIMH of the National Institutes of Health under award number R01MH051699 to J.C. and non-Federal funds from a predoctoral fellowship from Amira Pharmaceuticals to M.L. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Acknowledgments
We thank Dr. Andras Nagy for the R1 ES cells used for gene targeting, Grace Kennedy for histology expertise, Dr. Gwendolyn Kaeser for statistical analysis, and Dr. Gwendolyn Kaeser and Danielle Jones for editorial assistance.
The abbreviations used are:
- DRG
- dorsal root ganglia
- CNS
- central nervous system
- LPA
- lysophosphatidic acid
- PNS
- peripheral nervous system
- PNSL
- partial sciatic nerve ligation.