IMMUNE CELL TREATMENT OF NERVE DAMAGE

Abstract

Provided is a composition for treating nerve injury, the composition including a natural killer cell, an immune cell, or a substance increasing activity thereof. According to the natural killer cell, the immune cell, or the substance increasing activity thereof according to an aspect, the natural killer cell may infiltrate into a nerve injury site to directly remove injured nerve cells, thereby being usefully applied to fundamental treatment of a nervous system disease caused by nerve injury or abnormal nerves.

Claims

1. A method for preventing or treating a nervous system disease caused by nerve injury or abnormal nerves, the method comprising: administering to a subject in need thereof an effective amount of an isolated natural killer cell, a cell population thereof, a substance increasing natural killer cell activity, or a combination thereof.

2. The method of claim 1, wherein the isolated natural killer cell or the cell population thereof has increased activity or is genetically modified to have increased activity, as compared with a natural killer progenitor cell.

3. The method of claim 2, wherein the genetically modified natural killer cell or cell population thereof is modified to express chimeric antigen receptor (CAR) or homing receptor.

4. The method of claim 3, wherein the chimeric antigen receptor comprises an extracellular domain, a transmembrane domain, and an intracellular stimulatory domain.

5. The method of claim 1, wherein the substance increasing natural killer cell activity is one or more selected from the group consisting of interferon, interleukin, an interleukin-antibody complex, an agonist of natural killer cell activating receptor, an antibody against immunoglobulin-like receptor, an antibody against glucocorticoid-induced TNF-related protein (GITR), an antibody against CD137, an antibody against CD27, an antibody against OX40, an antibody against CTLA-4, an antibody against PD-1, and an antibody against NKG2A.

6. The method of claim 5, wherein the interleukin is any one selected from the group consisting of IL-2, IL-5, IL-8, IL-12, IL-15, IL-18, IL-21, and a combination thereof.

7. The method of claim 5, wherein the agonist of natural killer cell activating receptor is bispecific killer engagers (BiKEs) or trispecific killer engagers (TriKEs).

8. The method of claim 1, wherein the nerve injury is neuropraxia, axonotmesis, or neurotmesis.

9. The method of claim 1, wherein the nervous system disease caused by nerve injury or abnormal nerves is a nervous system disease caused by nerve injury or abnormal nerves of the central nervous system or a nervous system disease caused by nerve injury or abnormal nerves of the peripheral nervous system.

10. The method of claim 9, wherein the nervous system disease caused by nerve injury or abnormal nerves of the central nervous system is any one selected from the group consisting of organic diseases and dysfunctions of the central nervous system, epilepsy, multiple sclerosis, amyotrophic lateral sclerosis, Alzheimer's disease, Lewy dementia, Huntington's disease, Parkinson's disease, schizophrenia, traumatic brain injury, stroke, Pick's disease, Creutzfeldt-Jakob disease, progressive supernuclear palsy, multiple system atrophy, corticobasal degeneration, spinocerebellar degeneration, cerebellar atrophy, post-traumatic stress disorder, amnesia, vascular dementia, and cerebral infarction.

11. The method of claim 9, wherein the nervous system disease caused by nerve injury or abnormal nerves of the peripheral nervous system is any one selected from the group consisting of peripheral neuropathy, diabetic neuropathy, peripheral neuropathic pain, chemotherapy-induced peripheral neuropathy, complex regional pain syndrome, optic neuropathy, mononeuropathy, mononeuropathy multiplex (mononeuritis multiplex), polyneuropathy, Guillain-Barre syndrome (acute inflammatory demyelinating polyneuropathy), chronic inflammatory demyelinating polyneuropathy, hereditary neuropathy, plexus disorder, glaucoma, macular degeneration, amyotrophic lateral sclerosis, progressive muscle atrophy, progressive bulbar palsy, polio, post-polio syndrome, stiff-man syndrome, Isaac's syndrome, myasthenia gravis, neonatal myasthenia, botulism, Eaton-Lambert syndrome, thoracic outlet syndrome, Charcot-Marie-Tooth disease, and spinal muscular atrophy.

12. The method of claim 11, wherein the peripheral neuropathic pain is any one selected from the group consisting of trigeminal neuralgia, diabetic neuropathic pain, phantom limb pain, viral infection, pain from trauma, pain following chemotherapy, atypical facial pain, and post herpetic neuralgia.

13. The method of claim 11, wherein the diabetic neuropathy is polyneuropathy or focal neuropathy.

14. The method of claim 13, wherein the polyneuropathy is any one selected from the group consisting of hyperglycemic neuropathy, distal symmetric polyneuropathy, autonomic neuropathy, acute sensory neuropathy, acute painful sensory neuropathy, and chronic sensorimotor neuropathy.

15. The method of claim 13, wherein the focal neuropathy is any one selected from the group consisting of cranial neuropathy, truncal neuropathy, limb neuropathy, thoracolumbar radiculoneuropathy, and lumbosacral radiculoplexus neuropathy.

16. The method of claim 1, wherein the isolated natural killer cell, the cell population thereof, the substance increasing natural killer cell activity, or the combination thereof removes injured or abnormal nerve cells.

Description

BRIEF DESCRIPTION OF DRAWINGS

[0068] FIGS. 1A to 1E show results of examining cytotoxicity of IL-12-stimulated natural killer cells against embryonic DRG neurons, wherein the natural killer cells were isolated, enriched, and activated from adult mouse spleen, in which FIG. 1A shows results of intracellular staining of granzyme B of non-stimulated natural killer cells (left) and IL-2 stimulated natural killer cells (right), FIG. 1B shows results of flow cytometry for examining changes in NKp46.sup.+DX5.sup.+ natural killer cells in splenic lymphocytes by negative MACS enrichment (top-right quadrant), FIG. 1C shows result of immunostaining (β-tubulin III) of embryonic DRG neurons after co-culture with IL-2 stimulated NK cells either by direct contact or separated by a transwell membrane, FIG. 1D shows results of ELISA detection of granzyme B in media of NK cell-DRG neuron co-cultures, and FIG. 1E shows results of examining effects of the NKG2D receptor on NK cell-mediated lysis of DRG neurons;

[0069] FIGS. 2A to 21 show results of examining susceptibility of cultured embryonic DRG neurons to NK-mediated cytotoxicity by RAE1, in which FIG. 2A shows immunolabeling of β-tubulin (magenta) and NKp46 (green) of embryonic DRG neurons or adult DRG neurons co-cultured with IL-12 stimulated natural killer cells, FIG. 2B shows results of LDH-release cytotoxicity assays of acutely cultured embryonic DRG neurons and adult DRG neurons at various Effector (NK):Target (DRG) (E:T) ratios, FIG. 2C shows images of in vitro time-lapse confocal Ca.sup.2+ imaging of rhodamine 3 AM-loaded embryonic DRG neurons or adult DRG neurons co-cultured with IL-2 stimulated NK cells isolated from adult male NKp46-YFP mice, FIG. 2D shows frequency over time of neurite Ca.sup.2+ events in embryonic DRG neurons or adult DRG neurons during IL-12 stimulated NK co-culture, FIG. 2E shows RT-PCR results of examining the expression of mRNA transcripts (Raet1, NK1.1, Advillin) in non-treated splenic NK cells, embryonic DRG neurons, or adult DRG neurons, FIG. 2F shows qRT-PCR results of examining Raet1 mRNA expression in embryonic DRG neurons or adult DRG neurons, FIG. 2G shows Western blot results of embryonic DRG neurons or adult DRG neurons with a pan-RAE1 antibody, FIG. 2H shows knockdown of Raet1 mRNA in embryonic DRG neurons using Raet1 selective siRNA, and FIG. 2I shows results of an LDH-release cytotoxicity assay of Raet1 mRNA knockdown embryonic DRG neurons;

[0070] FIGS. 3A to 31 show results of examining upregulation of Raet1 expression in injured adult DRG neurons and NK cell-mediated neurite fragmentation, in which FIG. 3A shows results of immunolabeling a neurite compartment with β-tubulin III after microfluidic culture of adult DRG neurons exposed to IL-2 stimulated NK cells, FIG. 3B shows results of quantification of DRG neurite density, FIG. 3C shows qRT-PCR results of examining time-dependent changes in Raet1 mRNA expression in adult DRG neuron cultures, FIG. 3D shows Western blot results of examining RAE1 protein expression in adult DRG cultures in vitro after 1 and 2 days, FIG. 3E shows results of β-tubulin III immunolabeling of an adult DRG culture alone (2 days in vitro) or an adult DRG culture co-cultured with IL-2 stimulated NK cells, FIG. 3F shows results of quantification of DRG neurite fragmentation, FIG. 3G shows results of knockdown of Raet1 mRNA in adult DRG neurons in vitro using Raet1 selective siRNA, FIG. 3H shows results of β-tubulin III immunolabeling of a Raet1 siRNA transfected adult DRG culture alone (2 days in vitro) or an adult DRG culture co-cultured with IL-2 stimulated NK cells, and FIG. 3I shows results of quantification of DRG neurite fragmentation;

[0071] FIGS. 4A and 4B show results of examining effects of blocking the NKG2D receptor of NK cells on neurite degeneration of adult DRG neurons, in which FIG. 4A shows results of β-tubulin III immunolabeling, after co-culturing anti-NKG2D antibody-pretreated adult DRG neurons with IL-12 stimulated natural killer cells, and FIG. 4B shows results of quantification of DRG neurite fragmentation;

[0072] FIGS. 5A to 5K show results of examining the effect of peripheral nerve injury on Raet1 expression and the effect of injured sensory neurons on neurite fragmentation by stimulated NK cells, in which FIG. 5A shows a schematic diagram of a spinal nerve transection site in relation to lumbar DRGs (L3, L4, and L5), FIG. 5B shows qRT-PCR results of examining time-dependent Raet1 mRNA expression in ipsilateral L5 DRG, after spinal nerve transection with respect to contralateral DRG, FIG. 5C shows qRT-PCR results of examining injury-related Raet1 mRNA expression in adult DRG neurons (1 day in vitro), FIG. 5D shows β-tubulin III immunolabeling of L5 DRG neurons isolated 7 days after co-culture (4h) of sham DRG neurons with IL-2 stimulated NK cells, FIG. 5E shows results of quantification of DRG neurite fragmentation, FIG. 5F shows β-tubulin III immunolabeling of L5 DRG neurons isolated 7 days after DRG neurons subjected to spinal nerve transection were co-cultured (4 hr) with IL-2 stimulated NK cells, FIG. 5G shows results of quantification of DRG neurite fragmentation, FIG. 5H shows results of anti-GFP immunolabeling of sciatic nerve tissue sections obtained from adult male NKp46-YFP mice after L5 spinal nerve transection injury, FIG. 5I shows results of quantification of the total number of YFP-positive events within a lymphocyte FSC/SSC gate from whole sciatic nerve homogenates, FIG. 5J shows results of ELISA quantification of granzyme B content in a whole sciatic nerve of a wild-type mouse after L5x injury, and FIG. 5K shows results of ELISA quantification of granzyme B content in a whole sciatic nerve of an NKp46-DTR mouse after L5x injury;

[0073] FIGS. 6A to 6C show results of characterization of NKp46-cre mice, in which FIG. 6A shows results of flow cytometry of peripheral blood lymphocytes from NKp46-YFP mice labelled with anti-NKp46 and anti-CD3 antibodies, FIG. 6B shows results of anti-GFP antibody (green) immunolabeling of spleen tissue sections from wild-type and NKp46-YFP mice, and FIG. 6C shows results of flow cytometry of peripheral blood lymphocytes from wild-type and NKp46-DTR mice 24 hr after intravenous treatment with DTx (100 ng);

[0074] FIGS. 7A to 7E show results of examining effects of sciatic nerve crush on RAE1 in peripheral nerve axons, in which FIG. 7A shows a schematic diagram of a sciatic nerve crush injury site in relation to lumbar DRGs (L3, L4, and L5), FIG. 7B shows qRT-PCR results of examining crush injury-related Raet1 mRNA expression in ipsilateral L3-5 DRG 3 days and 7 days post-surgery, FIG. 7C shows results of examining RAE1 protein expression in a sciatic nerve (3 days and 7 days) after peripheral nerve crush injury, FIG. 7D shows results of examining RAE1 protein expression in DRG neurons (3 days and 7 days) after peripheral nerve crush injury, and FIG. 7E shows results of examining RAE1 (green) co-localization with axonal marker β-tubulin III (magenta) in a sciatic nerve (3 days and 7 days) after tight ligation;

[0075] FIGS. 8A to 8G show results of examining effects of systemic NK cell depletion on sensory nerve injury after sciatic nerve crush, in which FIG. 8A shows results of examining sciatic nerve tissue sections from adult male NKp46-YFP mice 7 days after sciatic nerve crush injury, FIG. 8B shows flow cytometry results of sciatic nerve homogenates obtained from adult male NKp46-YFP mouse 3 days after sciatic nerve crush injury, FIG. 8C shows results of quantification of the total number of CD45.sup.+/YFP.sup.+ double-positive events within a lymphocyte FSC/SSC gate from whole sciatic nerve homogenates, FIG. 8D shows results of examining changes in daily pin prick response score in NKp46-DTR mice administered intravenously with DTx every 4 to 5 days, FIG. 8E shows the area under the curve (AUC) in relation to cumulative sensation following full sciatic nerve crush and during recovery, FIG. 8F shows a heat map showing mean sensitivity to pin prick along the lateral hind paw, and FIG. 8G shows results of quantification of NKp46.sup.+/DX5.sup.+ double-positive lymphocytes in peripheral blood 16 days after sciatic nerve crush;

[0076] FIGS. 9A and 9B show the results of quantification of granzyme B in a whole sciatic nerve after crush or sham injury, in which FIG. 9A shows the result of quantification of granzyme B in a whole sciatic nerve after crush or sham injury of wild-type mice, and FIG. 9B shows a result of quantification of granzyme B in a whole sciatic nerve after crush or sham injury of NKp46-DTR mice;

[0077] FIGS. 10A to 10H show the results of examining the effect of IL-2/IL-2 antibody complex treatment on NK cell-dependent acute sensory loss after partial sciatic nerve crush, in which FIG. 10A shows daily pin prick response scores for examining a transient reduction in sensitivity by treatment with an IL-2/IL-2 antibody complex, FIG. 10B shows AUC values related to loss of cumulative sensation in mice treated with an IL-2/IL-2 antibody complex, FIG. 100 shows a heat map showing mean sensitivity to pin prick along the lateral hind paw, FIG. 10D shows results of examining changes of NKp46.sup.+DX5.sup.+NK cells, CD3.sup.+CD8.sup.+ T cells, or CD3.sup.+CD4.sup.+ T cells in peripheral blood by treatment with an IL-2/IL-2 antibody complex, FIG. 10E shows daily pin prick response scores for examining a transient reduction in sensitivity by treatment with an IL-2/IL-2 antibody complex in mice that have received anti-NK1.1 to deplete NK cells, FIG. 10F shows AUC values related to loss of cumulative sensation in mice treated with an IL-211L-2 antibody complex, FIG. 10G shows a heat map showing mean sensitivity to pin prick along the lateral hind paw, and FIG. 10H shows results of examining changes of NKp46.sup.+DX5.sup.+NK cells, CD3.sup.+CD8.sup.+ T cells, or CD3.sup.+CD4.sup.+ T cells in peripheral blood by treatment with an IL-2/IL-2 antibody complex;

[0078] FIGS. 11A to 110 show results of examining the effect of NK cell depletion on the acute sensory loss by IL-2 complex treatment after partial sciatic nerve crush injury, in which FIG. 11A shows daily pin prick response scores for examining a transient reduction in sensitivity by treatment with an IL-2/IL-2 antibody complex in NKp46-DTR mice that have received a partial crush of the sciatic nerve with DTx treatment, FIG. 11B shows a heat map showing mean sensitivity to pin prick along the lateral hind paw, and FIG. 11C shows results of examining changes of NKp46.sup.+DX5.sup.+NK cells, CD3.sup.+CD8.sup.+ T cells, or CD3.sup.+CD4.sup.+ T cells in peripheral blood by treatment with an IL-211L-2 antibody complex;

[0079] FIGS. 12A to 12F show results of examining the effect of an IL-2/1L2 antibody complex on long-term mechanical thresholds by axon injury in the sciatic nerve, in which FIG. 12A shows results of β-tubulin III immunolabeling and quantification of full-length anisotropic sciatic nerve sections 6 days after partial crush in mice treated with an IL-2/IL-2 antibody complex (a-proximal region, b-crush site, and c-distal region), FIG. 12B shows results of Stathmin 2 immunolabeling and quantification of full-length anisotropic sciatic nerve sections 6 days after partial crush in mice treated with an IL-2/IL-2 antibody complex, FIG. 12C shows high-magnification images of β-tubulin III and STMN2 immunofluorescence in FIGS. 12A and 12B, respectively, FIG. 12D shows result of examining mechanical sensitivity thresholds of an ipsilateral hind paw 16 days after partial injury in mice treated with an IL-2/IL-2 antibody complex, FIG. 12E shows result of examining mechanical sensitivity thresholds of an ipsilateral hind paw 15 days after partial injury and IL-2/IL-2 antibody complex treatment of mice treated with anti-NK1.1 antibodies, and FIG. 12F shows results of examining a correlation between mechanical sensitivity outcomes in the injured limb and cumulative pin prick sensitivity;

[0080] FIG. 13 shows results of RAE1 (green) and STMN2 (magenta) co-immunolabeling in a chronically ligated sciatic nerve;

[0081] FIG. 14 shows results confirming that peripheral neuropathy caused by administration of an anticancer drug was induced since pain was induced on day 7 in an anticancer drug-administered group;

[0082] FIG. 15 shows results of examining the interaction between NK cells and peripheral sensory nerve cells which were cultured on day 7 after administration of the anticancer drug;

[0083] FIG. 16 shows results of examining ligand expression of NK cells in sensory nerve cell tissues on day 7 after administration of the anticancer drug;

[0084] FIGS. 17A to 17C show images of selective infiltration of NK cells into brain tissues in which excitotoxicity of the central nervous system was induced;

[0085] FIG. 18 shows results of analyzing behavioral responses occurring after induction of excitotoxicity and degeneration of the central nervous system in Nkp46-DTR transgenic mice; and

[0086] FIG. 19 shows results of measuring ligand expression of NK cells in hippocampal tissues on day 3 after induction of excitotoxicity and degeneration of the central nervous system.

MODE OF DISCLOSURE

[0087] Hereinafter, the present disclosure will be described in more detail with reference to exemplary embodiments. However, these exemplary embodiments are only for illustrating the present disclosure, and the scope of the present disclosure is not limited to these exemplary embodiments.

Reference Example 1. Experimental Materials

[0088] All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at Seoul National University (Approval number: SNU-121011-1) and were reported in accordance with the ARRIVE guidelines (Kilkenny et al., 2010). Adult male and pregnant female wild-type C57BL/6 mice were purchased from Dae Han Bio Link (Taconic, Korea). Rosa26eyfp (RRID:IMSR_JAX:006148) (Srinivas et al., 2001) and Rosa26dtr (RRID:IMSR_JAX:007900) (Buch et al., 2005) were purchased from Jackson Laboratories (USA). Ncr1icre mice (RRID:MGI:5308422) containing cre-recombinase inserted by homologous recombination at the 3′ end of Ncr1 (Nkp46) gene (Narni-Mancinelli et al., 2011) were a kind gift from Dr Eric Vivier. All mice were maintained as homozygous stocks in a specific pathogen free (SPF) facility. Double heterozygote Nrc1icre/wt;rosa26eyfp/wt (abbreviated to NKp46-YFP) and Nrc1icre/wt;rosa26dtr/wt (abbreviated to NKp46-DTR) mice were bred from single crosses in an SPF facility and transferred to a conventional room at least one week before experiments. Mice were maintained on a 12 h:12 h light/dark cycle (lights on at 8:00 a.m.), housed 4-6 mice per cage on wood chip bedding and provided with standard laboratory feed and water ad libitum. DRG neurons and natural killer cells were prepared from male C57BL/6 mice (6-8 weeks); embryonic DRG neurons were prepared from embryos on day 15 in utero (E15) removed from euthanized female C57BL/6 mice. Nerve injury experiments were performed on male mice aged 7-9 weeks of indicated genotype. Animals were killed in accordance with Schedule 1 of the UK Animals (Scientific Procedures) Act 1986 by inhalation of a lethal concentration of isoflurane, followed by a lethal concentration of carbon dioxide for tissue culture.

Reference Example 2. Experimental Procedures

[0089] 2-1. Natural Killer Cell Depletion

[0090] NKp46-DTR mice were treated with diphtheria toxin (DTx; 100 ng) or sterile PBS solution (100 μl) intravenously by retro-orbital injection (Yardeni et al., 2011). Under isoflurane anaesthesia (3% induction, 1-2% maintenance in 100% O.sub.2), an ophthalmic solution (0.5% proparacaine, Alcon, Belgium) was applied to one eye as a local anaesthetic. Insulin (0.3 ml, BD Biosciences) was inserted into the retro-orbital sinus of the eye and slowly injected. Injections were alternated between eyes. Injections of DTx or a sterile PBS solution administered blind to the content of the syringe starting one day prior to surgery and continuing every four to five days for the duration of the study. All mice were blood sampled at the end of experiments for depletion efficiency check by flow cytometry. For antibody depletion of natural killer cells, wild-type C57BL/6 mice were injected with 100 μg of LEAF purified anti-mouse NK1.1 (clone PK136) (Biolegend, cat no. 108712, RRID:AB_313399) or LEAF purified IgG2a, κ isotype control (clone MOPC-173) (Biolegend, cat no. 400224, RRID:AB_326472) intravenously by retro-orbital injection one day before nerve injury.

[0091] 2-2. L5 Spinal Nerve Transection (L5x)

[0092] Male mice aged 7-9 weeks were placed under isoflurane inhalation, the dorsal lumbar region was shaved, treated with an iodine solution (Potadine) and a unilateral incision made parallel to the L6 vertebrate. Under a ×20 dissection microscope illuminated by a cold light source, the musculature was parted by blunt forceps dissection to reveal the L6 transverse process, which was then cut and removed. The L5 spinal nerve, which runs immediately below the L6 process, was carefully freed of connective tissue and cut with fine spring scissors; 1 mm of the nerve was removed to prevent nerve regeneration. The wound was irrigated with sterile saline and closed in two layers with 6-0 silk sutures (Ailee, Korea) and 9 mm skin clips (MikRon Precision, CA, USA). Mice were placed in a warm, darkened cage to recover from surgery.

[0093] 2-3. Sciatic Nerve Crush

[0094] Adult male mice (7-9 weeks old) received a single unilateral crush injury to the sciatic nerve (Bridge et al., 1994). Briefly, under isoflurane anaesthesia, the right thigh was shaved and iodine treated and an incision was made mid-thing length. The sciatic nerve was exposed as it emerges from the sciatic foramen by parting the muscle with blunt forceps dissection. The nerve was carefully freed of connective tissue and fully crushed for 15 sec using fine, mirrorfinishe forceps (No 5, Dumont, Fine Science Tools, Germany). The wound was closed in two layers with two sutures of the overlying muscle facia and a single skin clip to close. Complete crush of the sciatic nerve was deemed successful by a sensory score of zero using the pin prick assay on the day following surgery; any mice with a pin prick response less than 24 hr after surgery were excluded from analysis. For partial (moderate) crush, an ultra-fine haemostat (Cat no. 13020-12, Fine Science Tools, Germany) was fitted with a custom spacer created from two layers of aluminium foil (15 μm thick) to create a gap 30 μm thick when the sciatic nerve was fully closed. The sciatic nerve was carefully free of connective tissue and placed between spacers on the haemostat (2-3 mm from the tip) by gently lifting the nerve using a fire-polished glassed rod. The haemostat was then closed on the first locking position and held for 15 sec before careful release of the nerve. The wound was closed in two layers with two sutures of the overlying muscle facia and a single skin clip. Mice were recovered in a warm, darkened cage. All tools were autoclaved prior to surgery and strict aseptic was maintained throughout.

[0095] 2-4. IL-2/Anti-IL-2 Antibody Complex Treatment

[0096] Recombinant mouse IL-2 (Cat no. 212-12, lot no. 0608108; Peprotech, Rocky Hill, N.J., USA) was prepared as a stock at 0.1 mg/ml in PBS (without carrier protein) and stored at 4° C. for up to one week, according to the manufacturer's instructions. On the day of treatment, IL-2 (1.5 μg per mouse) was pre-mixed with anti-mouse IL-2 monoclonal antibody (50 μg per mouse) (54B6-1 clone) (BioXCell; RRID:AB_1107705) and incubated at room temperature for 15 min. The bound cytokine/antibody complex was further diluted in sterile PBS to a total volume in 500 μl and injected intraperitoneally. Injections were given once daily (evening time) for four days. For control experiments, mice were injected with an equal amount of rat IgG2a isotype peptide (clone 2A3) (BioXCell; RRID:AB_1107769). The efficacy of the IL-2 antibody complex treatment was confirmed by enlargement of the spleen compared to PBS injected mice one day after the final injection (Boyman et al., 2006).

[0097] 2-5. Behavioral Testing

[0098] All sensory testing was performed between the hours of AM 9 and PM 18 in an isolated room maintained at 22±2° C. and 50±10% humidity. For mechanical threshold (von Frey filament) testing, mice were brought from the animal colony and placed in transparent plastic boxes on a metal mesh floor with 5×5 mm holes (Ugo Basile, Italy). The mice were then habituated for at least 30 min prior to testing. To assess mechanical sensitivity, the withdrawal threshold of the affected hind paw was measured using a series of von Frey filaments (0.20, 0.40, 0.70, 1.6, 3.9, 5.9, 9.8 and 13.7 mN, Stoelting, Wood Dale, Ill., USA; equivalent in grams to 0.02, 0.04, 0.07, 0.16, 0.40, 0.60, 1.0 and 1.4). The 50% withdrawal threshold was determined using the up-down method as previously described (Chaplan et al., 1994). A brisk hind paw lift or flinch in response to von Frey filament stimulation was regarded as a withdrawal response. The 0.4 g filament was the first stimulus to be used, and when a withdrawal response was obtained, the next weaker filament was used. This process was repeated until no response was obtained, at which time the next stronger filament was administered. Interpolation of the 50% threshold was then carried out using the method of Dixon (Dixon, 1980). All behavioral testing was performed by an investigator who was blind to the treatment of the mice. Pin prick sensory recovery testing was performed as previously described (Ma et al., 2011) with slight modification. Mice were habituated on an elevated mesh in separate compartments. The lateral side of the affected hind paw was separated into five regions from the toe to the heel and stimulated with a stainless steel Austerlitz insect pin (Size 000, FST, Germany). A sensory response was confirmed by rapid lifting or flinching of the paw. The number of responses to two consecutive pin applications to the skin was recorded per region providing a score out of 10. Responses due to direct movement of the paw or hind limb (indicative of extraterritorial proprioception) were excluded. Testing was carried out daily until full sensory recovery (score of 10) was observed for all mice.

[0099] 2-6. DRG Neuron Culture

[0100] Adult or embryonic DRG were rapidly dissected on ice-cold Ca.sup.2+- and Mg.sup.2+-free Hank's Buffer Saline Solution (HBSS, Welgene) (including 20 mM HEPES) and digested 30-60 min in collagenase A (1 mg/ml) and dispase II (2.4 U/ml) (Roche, Switzerland) at 37° C. Additional digestion was carried out for 5-7 min in trypsin (0.25%) and stopped with a trypsin inhibitor (2.5 mg/ml) (Sigma, T9003) in PBS followed by washing in Dubellco's Modified Eagle Medium (Gibco, Life Technologies) containing 10% serum (Gibco, Life Technologies). DRG were dissociated by trituration with a fire-polished glassed pipette in DMEM containing DNase I (125 U/ml) and centrifuged at 200 g on a layer of Bovine Serum Albumin (15% BSA solution; Sigma) before re-suspension in neurobasal medium (Gibco, Life Technologies) with B27 supplement, L-glutamine (1 mM), penicillin (100 U/ml) and streptomycin (100 U/ml) supplemented with nerve growth factor (NGF 2.5S) at 50 ng/ml. Adult DRG (10.sup.3 cells) and embryonic DRG (8×10.sup.3 cells) were plated on 10 mm diameter glass coverslips, glass bottom dishes or 96 well flat-bottom culture plates (Nunclon, Thermo Scientific) previously coated with poly-D-lysine (10 μg/ml) and laminin (10 μg/ml) (Sigma). For injured DRG experiments, ipsilateral L5 DRG were rapidly dissected on ice-cold HBSS from adult mice 7 days after L5x or sham surgery, pooled (n=3 DRG per group) and dissociated to a single cell suspension as above. DRG were then seeded onto poly-D-lysine and laminin-coated glass bottom dishes (10.sup.3 cells per dish) and cultured overnight in neurobasal medium containing NGF (50 ng/ml). For microfluidic cocultures, DRG neurons isolated from adult mice (as above) were suspended in neurobasal medium and seeded (10.sup.4 cells) into the somal reservoir of a microfluidic device (Xona Microfluidics, CA, USA) previously coated with poly-D-lysine (10 μg/ml) and laminin (10 μg/ml) (Sigma). NGF (100 ng/ml) was added to the media in the neurite reservoir. Neurons were cultured for 5 days during which neurites grew along 3 μm×500 μm channels connected to the neurite reservoir.

[0101] 2-7. Natural Killer Cell Isolation and Stimulation

[0102] Natural killer cells were prepared from spleens of adult male C57BL/5 mouse (6-8 weeks old). Spleen cells were homogenized by sequentially passing through 70 μm and 40 μm cell strainers (Falcon, BD Biosciences). Red blood cells were lysed by incubation for 2 min in ACK lysis buffer (in mM: 150 NH.sub.4Cl, 10 KHCO.sub.3, 0.1 Na.sub.2EDTA, pH 7.3). Single-cell suspensions were then passed through nylon wool columns (Polysciences, Warrington, Pa.) for the depletion of adherent populations, consisting of B cells and macrophages. Eluted cells were resuspended in 0.01 M phosphate buffered saline (PBS)+2 mM EDTA and 2% FBS. Natural killer cells were enriched using a magnetic associated cell sorting (MACS) method in combination with a negative selection protocol (Mouse natural killer cell isolation Kit II, cat no. 130-096-892, Miltenyi Biotech GmbH, Germany) according to the manufacturer's instructions. Briefly, cell suspensions were sequentially incubated at 4° C. with a cocktail of biotin-conjugated monoclonal antibodies against non-natural killer cells followed by anti-biotin microbeads. The cell suspension was then passed through an LS column placed in a magnetic field (MidiMACS Separator, Miltenyi Biotech GmbH). Bead-conjugated non-natural killer cells remained in the columns while non-labelled natural killer cells passed through in the eluent. Enriched natural killer cells were either used directly (control) or stimulated with recombinant murine interleukin (IL)-2 (Cat no. 212-12, lot no. 0608108; Peprotech, Rocky Hill, N.J., USA) at 1000 U/ml for two days before use in experiments. Natural killer cells were cultured at 2×10.sup.6 cells per ml in RPMI 1640 medium (Gibco, Life Technologies) supplemented with fetal bovine serum (FBS) (10%) and penicillin/streptomycin (100 U/ml) in 96-well U-bottomed (Falcon, BD Biosciences) plates for 48 hr. The purity of natural killer cells (NKp46.sup.+DX5.sup.+) in the eluent was checked by flow cytometry to be consistently above 90%. The cells were then harvested and used as effector cells in co-culture and cytotoxicity experiments.

[0103] 2-8. DRG-NK Co-Cultures

[0104] Control or IL-2 stimulated natural killer cells were harvested, washed in RPMI and resuspended in neurobasal media. DRG were washed once in neurobasal media (NGF-free) and natural killer cells were added to the neurite compartment (5×10.sup.5 cells) for microfluidic cultures or seeded (2.5×10.sup.5 cells) directly over the DRG for glass coverslip cultures, and co-cultured for 4 hr at 37° C. and 5% CO.sub.2. Co-cultures were carefully washed once in warm HBSS and fixed with 2-4% PFA in 0.01M PBS (pH 7.4) for 30 min at room temperature, followed by washing in PBS (3×10 mins) at storage at 4° C. prior to immunolabeling for β-tubulin III and NKp46 (see Immunofluorescence). For trans-well experiments, glass coverslip cultures of DRG were transferred to a 24-well plate in 500 μl neurobasal media. Natural killer cells (2.5×10.sup.5 per well) were seeded onto 6.5 mm diameter polycarbonate trans-well membrane inserts with 0.4 μm pore size (Corning) and incubated for 4 hr at 37° C. and 5% CO.sub.2 before removal of the membrane and fixation as above. For antibody blocking of NKG2D function, natural killer cells were incubated with 30 μg/ml of LEAF purified anti-mouse CD314 (NKG2D) (CX5 clone) (Biolegend, cat no. 130204. RRID:AB_1227715) or LEAF purified rat IgG1 isotype control (Clone RTK2071) (Biolegend, cat no. 400414. RRID:AB_326520) for 15 min at room temperature (2.5×10.sup.6 natural killer cells per ml) before addition to target DRG neurons.

[0105] 2-9. LDH-Release Cytotoxicity

[0106] Effector natural killer cells were assayed for cytotoxicity against DRG neuron targets by measuring the release of lactate dehydrogenase (LDH) into the culture medium using an LDH Cytotoxicity Assay Kit (Thermo Scientific Pierce, IL, US). Control or IL-2 stimulated natural killer cells were harvested, washed in RPMI and added to DRG cultures in 96 well plates (Nunclon, Thermo Scientific) in neurobasal medium at various ratios and cultured for 4 hr before sampling of media supernatant which was assayed for LDH activity according to manufacturer's instructions. Absorbance values were acquired on a microplate spectrophotometer (BioTek Instruments, VT, US). Each ratio was determined in triplicate. Specific cytotoxicity was calculated as follows: % cytotoxicity=[(experimental release−spontaneous release)/(maximum release−spontaneous release)]×100. The optimum number of target cells (DRG neurons) for maximum LDH release detection was determined prior to experiments (8×10 embryonic DRG neurons; 10.sup.3 adult DRG neurons).

[0107] 2-10. Live Confocal Imaging

[0108] Adult or embryonic mouse DRG (10.sup.3 and 8×10.sup.3 neurons, respectively) cultured for one day on PDL and laminin-coated glass-bottom petri dishes were fluorescently labelled with Vybrant Dil (Molecular Probes, cat no. V-22886) or loaded with the Ca.sup.2+ indicator rhodamine 3-AM (Molecular Probes, cat no. R10145) according to the manufacturer's instructions. Natural killer cells previously isolated from NKp46-YFP mice and stimulated IL-2 (1000 U/ml) for 48 hr were suspended in neurobasal media and seeded onto the coverslip (2.5×10.sup.5 cells per dish). Dishes containing DRG-NK co-cultures were immediately transferred to a confocal microscope (LSM 700, Zeiss) and maintained in a humidified atmosphere at 37° C. and 5% CO.sub.2 (Live Cell Instruments, Seoul Korea). A time series of single z-section images (512×512) were acquired using a multitrack setting (488 nm and 555 nm fluorescence emission and differential interference contrast (DIC) bright-field) at 30-60 sec intervals and multiple positions under the control of Definite Focus. Images were acquired up to 3 hr and exported as a sequential time-lapse in AVI format.

[0109] 2-11. In Vivo Two Photon Sciatic Nerve Imaging

[0110] Male NKp46-YFP mice (8-9 weeks old) received unilateral sciatic nerve crush or sham surgery. On day 3, mice were anaesthetized with pentobarbital (80 mg/kg, i.p.) supplemented with 20 mg/kg immediately prior to recording. Some mice additionally received Dextran-Texas red (neutral 40,000 m.w.) (Molecular Probes) given via retro-orbital injection (100 μl, 10 mg/ml) to visualize the vasculature and confirm maintenance of blood flow during recording. The sciatic nerve was re-exposed and bathed in sterile saline. Mice were placed on a warm pad maintained at 35° C. for the duration of the imaging. The sciatic nerve was carefully lifted with two curved glass rods held via a micromanipulator at both distal sides of the exposed nerve to mechanically isolate the nerve from the body and minimize the breathing-induced motion artefacts. A W plan-Apochromat 20× water immersion lens was lowered to the nerve surface 1-2 mm distal from the crush site to identify the blood vessels and YFP-positive cells. After the Ti-sapphire laser (Chameleon, COHERENT) was tuned to the wavelength of 900 nm for excitation of both YFP and Texas Red, 3D time-lapse imaging was performed using a laser scanning confocal microscope (LSM 7MP, Zeiss) as follow: 1024×512 pixels per a single image, 0.79 μs of pixel dwell, 2 μm×25 sections per a z-stack (total 50 μm depth) at 30 sec interval. The two-photon laser power was compensated according to the depth across 5%-8% of total power. The image was rendered into 2D video using imaging software (Zeiss Efficient Navigation 2012).

[0111] 2-12. DRG Small Interference RNA Gene Knockdown

[0112] Acute isolated adult or E15 embryonic DRG were transfected with siRNA by electroporation (Neon, Invitrogen) according to the manufacturer's instructions. Briefly, single cell suspensions of DRG were suspended in an electroporation medium containing siRNA oligonucleotide and drawn into a tip containing a gold-plated electrode (5×10.sup.4 cells per 10 μl). Tips of cells were placed into a tube containing electrolytic buffer (1500 V, 20 ms), and electroporated (1500 V, 20 ms), and immediately ejected into penicillin/streptomycin-free neurobasal media containing 50 ng/ml NGF and cultured for 48 hr at 37° C., 5% CO.sub.2 until assessment of assessment of knockdown or functional experiments. Prior to experiments, transfection efficiency was optimized by electroporation of DRG with a cDNA plasmid encoding green fluorescent protein and examination of GFP fluorescence after two days in culture. Two siRNA oligonucleotides per target were tested for knockdown efficiency by real time PCR; siRNA oligos that reduced mRNA expression by more than 70% compared to a negative control siRNA oligonucleotide were used for functional experiments. Gapdh siRNA oligonucleotides (10 nM) (Silencer Select, Ambion, Life Technologies, cat no. 4390849) were used as a positive control.

[0113] 2-13. Peripheral Blood Lymphocyte Isolation

[0114] Systemic depletion of natural killer cells was confirmed at the end of experiments in a sample of peripheral blood obtained by retro-orbital bleed. Briefly, under isoflurane anaesthesia, a 15 ml glass Pasteur pipette was inserted into the retro-orbital sinus and gently twisted to disrupt the orbital venous plexus. 50 μl to 100 μl of blood was removed by capillary action and ejected into heparin-coated tubes (Idexx Laboratories, USA). Blood samples were diluted with an equal volume of serum-free RPMI, floated on top of a lymphocyte separation medium (Lympholyte Mammal, Cedarlane Labs, Canada) and centrifuged at 1000 g for 20 min with a slow acceleration gradient. Peripheral blood mononuclear cells were collected from the monolayer, washed in RPMI and suspended in FACS buffer (5% FBS, 0.002% NaN.sub.3 in 0.01 M PBS) for flow cytometry analysis.

[0115] 2-14. Flow Cytometry

[0116] For surface labeling, cell suspensions were transferred to 5 ml round bottom tubes and Fc receptors were blocked with unconjugated rat anti-mouse CD16/CD32 (clone 2.4G2) monoclonal antibody (1:100; BD Biosciences, cat no. 553142. RRID:AB_394657) for 15 min at 4° C. and labelled with a combination of fluorescently-conjugated anti-mouse primary antibodies for 30 min at 4° C. After washing with FACS buffer, cell suspensions were run on a four-colour flow cytometer (FACSCalibur, BD Biosciences) and gated populations analyzed with Cell Quest software (BD Biosciences). Lymphocytes were initially gated according to the FSC-SSC scatter profile; cell populations were identified by fluorescence gating compared to unlabeled or IgG controls. Antibodies used were PE rat anti-mouse NKp46 (clone 29A1.4) monoclonal antibody (1:200; eBioscience, cat no. 12-3351. RRID:AB_1210743), APC rat anti-mouse CD49b (clone DX5) monoclonal antibody (1:500; eBioscience, cat no. 17-5971. RRID:AB_469484), FITC Armenian hamster anti-mouse CD3e (clone 145-2C11) monoclonal antibody (1:1000; eBioscience, cat no. 11-0031. RRID:AB_464881), APC rat anti-mouse CD45 (clone 30-F11) monoclonal antibody (1:200; eBioscience, cat no. 17-0451. RRID:AB_469393), PE rat anti-mouse CD4 (clone GK1.5) monoclonal antibody (1:1000; eBioscience, cat no. 12-0041. RRID:AB_465507), APC rat anti-mouse CD8a (clone 53-6.7) monoclonal antibody (1:1000; eBioscience, cat no. 17-0081. RRID:AB_469335). Titres of flow cytometry antibodies were determined prior to experiments by comparison to equivalent concentrations of fluorescence-conjugated IgG isotype controls. Percentages of peripheral blood cell populations were calculated from 20,000 gated lymphocyte events.

[0117] 2-15. Intracellular Staining

[0118] For intracellular labeling of granzyme B, cell suspensions were washed in RPMI, resuspended in FACS buffer, and Fc receptors were blocked for 15 min at 4° C. Cells then underwent fixation and permeablization in BD Cytofix/Cytoperm buffer (BD Biosciences, cat no. 554714) for 20 min at 4° C. followed by labeling with PE rat anti-mouse granzyme B (clone NGZB) monoclonal antibody (1:100; eBioscience, cat no. 12-8898. RRID:AB_10853811) or PE rat IgG2a isotype control (1:100; eBioscience, cat no. 12-4321. RRID:AB_470052) for 30 min at 4° C. and washing in BD Perm/Wash buffer.

[0119] 2-16. Preparation of Sciatic Nerve Suspensions and Flow Cytometry

[0120] For flow cytometry analysis of natural killer cells present in sciatic nerve, NKp46-YFP mice were deeply anaesthetized with pentobarbital (100 mg/kg, i.p.) followed by trans-cardiac perfusion with PBS (0.01 M, pH 7.4) at various time points after peripheral nerve injury to remove peripheral blood from the circulation. Bilateral sciatic nerves were rapidly removed to ice-cold Ca.sup.2+- and Mg.sup.2+-free HBSS (Welgene) (including 20 mM HEPES) and cut into 1-2 mm pieces. Tissues were transferred to 15 ml tubes and centrifuged at 500 g for 5 min. HBSS was replaced with collagenase A (1 mg/ml) and dispase II (2.4 U/ml) (Roche, Switzerland) and incubated for 90 min at 37° C. with frequent gentle agitation. Additional digestion was carried out for 5 min in trypsin (0.25%) and stopped with a trypsin inhibitor (2.5 mg/ml) (Sigma, T9003) in PBS followed by washing in RPMI containing 10% serum (Gibco, Life Technologies). Nerves were dissociated by trituration with a fire-polished glassed pipette in RPMI containing DNase I (125 U/ml), passed through a 30 μm separation filter to remove debris (Miltenyi) and resuspended in FACS buffer in 5 ml round bottom tubes. Prior to sciatic nerve flow cytometry, CD45.sup.+ lymphocyte and NKp46.sup.−YFP cell gating was set on peripheral blood lymphocytes from wild-type and NKp46.sup.−YFP mice. Sciatic nerve samples were run on the slow setting until the total event count was less than 50 per second, or gated events were less than one per 5 sec. Natural killer cells were identified by lymphocyte FSC-SSC scatter profile and YFP fluorescence; in some experiments, total lymphocytes were additionally labelled with an APC-conjugated anti-mouse CD45 antibody (1:400; eBioscience, cat no. 17-0451. RRID:AB_469393).

[0121] 2-17. Tissue Preparation

[0122] At specific time points after nerve injury or sham surgery, mice were deeply anaesthetized with pentobarbital (100 mg/kg, i.p.) followed by exsanguination by transcardiac perfusion with PBS (0.01 M, pH 7.4) containing heparin (500 U/L). Whole DRG (lumbar L3-L5) and full length sciatic nerve (spinal nerves to peripheral trifurcation) tissues were dissected and collected in sample tubes which were immediately frozen on liquid nitrogen and stored at −70° C. for later molecular analysis. For immunohistochemistry, PBS perfusion was followed by a paraformaldehyde fixative (4% PFA, 0.2% picric acid in 0.1M PBS, pH 7.4). Sciatic nerves were post-fixed up to 24 hr and cryoprotected in sucrose solution (30% sucrose in 0.01 M PBS, pH 7.4) at 4° C. Fixed sciatic nerve tissues were embedded in frozen section compound (Optimal Cutting Temperature, Leica), sectioned on a cryostat (14 μm) and thaw-mounted on glass microscope slides (Superfrost Plus, Fischer Scientific).

[0123] 2-18. Immunofluorescence

[0124] After washing in PBS (3×10 min), cells and tissues were blocked and permeabilized in 0.1%-0.3% triton-X 100 and 10% normal donkey serum (NDS, Jackson ImmunoResearch) in PBS for 1 hr at room temperature. Primary antibodies were applied in 1% NDS, 0.01%-0.03% triton-X 100 in PBS and incubated overnight at 4° C. in a humidified chamber. Primary antibodies used were as follows: Rabbit anti-β-tubulin III (1:400-500, Sigma, cat no. T2200. RRID:AB_262133), goat anti-NKp46 (1:200; R&D Systems, cat no. AF2225. RRID:AB_355192), rabbit anti-STMN2 (1:500; Novus Biologicals, cat no. NBP1-49461. RRID:AB_10011569), goat anti-mouse pan-RAE1 antibody (1:40; R&D systems, cat no. AF1136. RRID:AB_2238016. Fluorescence conjugated secondary antibodies were incubated in 1% NDS, 0.1%-0.3% triton-X 100 in PBS for 1 hr at RT the dark. Secondary antibodies used were as follows: Alexa Fluor 647 donkey anti-rabbit (1:200; Molecular Probes, cat no. A31573. RRID:AB_2536183) and Alexa Fluor 488 donkey anti-goat (1:200; Jackson ImmunoResearch, cat no. 705-545-003. RRID:AB_2340428). For RAE1 double labeling with β-tubulin III and STMN2, primary and secondary labeling was performed sequentially for each antibody with a repeat blocking stage. Fluorescence images (1024×1024) were acquired sequentially on a laser-scanning confocal microscope (LSM 700 Zeiss) with Zen 2012 software (v8.1 SP1, Zeiss). Coverslips were imaged in a single z-section. Tissue sections were mounted with DAPI-containing hardsetting aqueous mounting medium (Vectorsheild, Vector Laboratories) and imaged as a stack of 4-5×2 μm z-sections and exported as a maximum intensity projection TIFF.

[0125] 2-19. Western Blot

[0126] Tissues and cells were homogenized in RIPA buffer (Millipore, Cat #20-188) containing protease inhibitor cocktail (Sigma, P8340) and phosphatase inhibitor cocktail (Gendepot, Cat # P3200). Cultured cells were washed with warmed HBSS and collected in protein lysis buffer by scraping; frozen tissues were disrupted either in a Minilys bead homogenizer (Precellys, Bertin, France) or glass grinder. Homogenized samples were then sonicated (3×10 s, 25% amplitude) on ice and then spun at high speed (10,000 g) for 10 min at 4° C. after 40 min incubation on ice and pellet discarded. An equal volume of 5×SDS sample buffer was added to the sample lysates which were boiled at 9° C. heat block for 5 min. Protein content was determined by colorimetric assay (Lowry, BioRad). Equal amounts of protein (25 μg to 40 μg) and protein size markers were separated by SDS-polyacrylamide gel electrophoresis (5% stacking gel, 10% resolving gel) followed by transfer to a PVDF membrane. Membranes were blocked in a 5% skimmed milk solution containing Tris buffered saline and 0.1% tween-20 (TBS-T) at room temperature for 1 hr and subsequently incubated with goat anti-mouse Pan-RAE1 antibody (1:500; R&D systems, cat no. AF1136. RRID:AB_2238016) overnight at 4° C. in blocking solution. Blots were washed with TBS-T (3×10 min) and then incubated with anti-goat HRP-conjugated secondary antibody (1:10,000; Santa Cruz, cat no. 2020, RRID:AB_631728) for 1 hr at room temperature. After washing with TBS-T, blots were developed by application of western ECL substrate (BioRad, cat no. 1705061) according to the manufacturer's instructions and images of sequential exposure times were acquired digitally (ChemiDoc, BioRad). Blots were then stripped with stripping buffer at 50° C. for 30 min followed by TBS-T washing and subsequently incubated with mouse anti-beta-actin (1:10,000; Sigma, cat no. A5441. RRID:AB_476744) and rabbit anti-N-cadherin (1:10,000; Millipore, cat no. 04-1126. RRID:AB_1977064) in blocking solution. Goat anti-mouse polyclonal (1:10.000; Komabiotech, cat no. K-0211589. RRID:AB_2636911) or goat anti-rabbit (1:10,000; Santa Cruz, cat no. sc-2004. RRID:AB_631746) HRP-conjugated secondary antibodies were applied, respectively, and then exposed after ECL treatment. Embryonic mouse head tissue was used as positive control for RAE1.

[0127] 2-20. Enzyme-Linked Immunosorbant Assay (ELISA)

[0128] Granzyme B content of sciatic nerves was determined by ELISA according to the manufacturer's instructions (DuoSet, R&D Systems, cat no. DY1865). Briefly, 96 well plates were coated with capture antibody overnight at room temperature, washed, and blocked 1 hr. Frozen tissues were homogenized on a shaker with 1.4 mm zirconium beads (Precellys, Bertin) in RIPA lysis buffer (Millipore) and lysates were centrifuged 13,000 rpm for 5 min. Supernatants were diluted 1:1 with reagent diluent, added in duplicate to coated wells along with a granzyme B standard series and incubated at room temperature for 2 hr. Plates were washed thoroughly before addition of Streptavidin-HRP, followed by substrate buffer. Stop solution was applied and absorbance (450 nm minus 540 nm correction) read on a microplate spectrophotometer (BioTek Instruments, VT, US). Granzyme B levels in samples were analyzed using the four parameter logistic curve fitting algorithm by reference to standard curve values and performed using online software (http://www.elisaanalysis.com/).

[0129] 2-21. Reverse Transcription Polymerase Chain Reaction (RT-PCR)

[0130] Total RNA was extracted from L5 DRG from L5x or sham mice (two mice pooled per sample) and L3-L5 DRG from sciatic crush or sham mice (one mouse per sample). Tissues were disrupted in RTL Plus lysis buffer (Qiagen) including 1% β-mercaptoethanol with a mini glass mortar and pestle on ice, and further homogenized using a Minilys bead homogenizer (Precellys, Bertin, France); DRG cultures were washed once in warm HBSS prior to lysis by pipetting; frozen cell pellets were vortexed in lysis buffer. RNA was purified from lysed samples, including genomic DNA elimination, by on-column extraction (RNeasy Plus, Qiagen) according to the manufacturer's instructions. RNA was eluted in RNase-free water and analyzed for purity (260/280 nm ratios of approximately 2.0 were considered acceptable) and nucleotide content on a spectrophotometer.

[0131] Equal amounts of RNA (150-250 ng from whole DRG, 500 ng from cell cultures) was reversed-transcribed using M-MLV (200 U/rxn), dNTPs and oligo(dT)12-18 primers (Invitrogen) in a 20 μl reaction volume according to the manufacturer's instructions. PCR reactions (25 μl) using Go Taq Flexi DNA polymerase (Promega) were then performed from cDNA on a thermal cycler (MJ Mini, BioRad). The PCR conditions were 95° C. (2 min), cycled 35 times from 95° C. (30 sec) to 60° C. (30 sec) to 72° C. (30 sec), 72° C. (5 min), 4° C. Reactions performed without cDNA served as negative control. PCR products were run on an agarose (1.5%) gel containing a DNA staining reagent (SafePinky, GenDepot, USA) and visualized on a UV transilluminator.

[0132] 2-22. Quantitative Real-Time PCR

[0133] Gene expression was analysis performed on cDNA from DRG using a Power SYBR Green PCR Master Mix (Applied Biosystems) and pairs of target-specific primers (500 nm) in MicroAmp optical tubes (20 μl reaction volume) on a 7500 Real-Time PCR system (Applied Biosystems). The PCR conditions were 50° C. (2 min), 95° C. (10 min), and cycled 40 times at 95° C. (15 sec) to 60° C. (1 min). Samples were run in triplicate. Data were analyzed using the built-in 7500 software (v2.0.4, Life Technologies) and expression was determined relative to a reference gene (Gapdh) in adult DRG for cultures or contralateral DRG for nerve injury experiments using the comparative Ct method (Schmittgen and Livak, 2008). Primers were designed using Primer-BLAST software (NIH) (Ye et al., 2012). Primers were selected based on specificity to the desired target gene upon BLAST search, overlap of the exon-exon boundary, lack of potential hairpin-forming or self-priming regions, single peak in the dissociation curve (single band PCR product) and equivalent amplification efficiency (linear shift in Ct value upon serial dilution of cDNA). Products of reverse transcription reactions omitting RNA or M-MLV (-RT) served as negative controls.

[0134] 2-23. Laser Scanning Confocal Imaging and Analysis

[0135] For analysis of neurite density in microfluidic co-cultures, single z-section low magnification (×10) confocal images at 0.5 zoom (LSM700, Zeiss) were acquired of β-tubulin III immunofluorescence (647 nm emission) along the full length of the neurite compartment. Gamma high channel images were exported as an unaltered TIFF using Zeiss imaging software (v8.1, ZEN 2012 SP1, Zeiss). Images were converted to black/white by adjusting to a set threshold, scale calibrated and horizontal pixel density was measured using the Plot Profile function in Image J (v1.46r, NIH). The distance to 50% neurite density was calculated from the normalized cumulative pixel density. 6-8 images per microfluidic device were acquired, n=3 devices per group were used. For analysis of neurite fragmentation in NK-DRG co-cultures, 5-10 fields of view were randomly selected from each coverslip in a non-systematic manner for acquisition in single or double z-section confocal images at low magnification (×20) and 0.5 zoom (LSM700, Zeiss). For total neurite fragmentation gamma, high channel images of β-tubulin III immunofluorescence were exported as an unaltered TIFF. In Image J, images were scale calibrated (1.6 pixels/μm) and brightness threshold set in black/white mode, creating a whole neuron silhouette for selection. Neurite fragments were selected using the particle analysis function (size 0.5-25 μm.sup.2, circularity 0-1) and saved as a drawing. The total area and particular area values (μm.sup.2) were obtained using the Measure function and used to calculate the percent neurite fragmentation for each field of view. For analysis of β-tubulin III and STMN2 labeling in sciatic nerve tissue, a composite of the full length nerve was created from individual z-section images acquired at ×10 magnification and 0.5 zoom along the length of the nerve (LSM700, Zeiss). Images containing crush site, as well as proximal and distal regions (±1 mm from the crush site) were exported to ImageJ, scale calibrated (0.8 pixels/μm) and brightness threshold set in black/white mode. The mean pixel density from a 9000 μm.sup.2 selection of the nerve mid-image was recorded from 3-4 images per nerve region, per mouse, per treatment. Co-localized regions of RAE1 double-labeling with β-tubulin III and STMN2 were exported using the co-localization function on the Zen 2012 software after setting for threshold image intensity in both fluorescence channels (1000 units per channel).

[0136] 2-24. Statistical Analysis

[0137] Comparisons between two groups of data were made with Student's t test (paired or unpaired) or Mann-Whitney test for non-normal distributions (confirmed with Kolmogorov-Smirnov test). Comparisons between three or more groups of data were made with One-way ANOVA. Two-way ANOVA tests were used to compare the effects of treatments on neuronal cytotoxicity and behavioral sensitivity assays. Data are presented as mean±standard error of the mean unless otherwise stated. Analysis were performed using GraphPad Prism version 5.00 for Windows, GraphPad Software, San Diego Calif. USA, www.graphpad.com. No statistical methods were used to predetermine sample sizes, which were based on previous literature and availability of animals. For behavior tests, experimenters were blinded either to the treatment or genotype of the mice during both surgery and sensory testing. Treatments were assigned to litter mates at random by an independent observer. p<0.05 was considered significant.

Example 1. Confirming that Activated Natural Killer Cells Induce Cytotoxicity in Embryonic Sensory Neurons by an RAEI-Mediated Mechanism

[0138] Natural killer cells are activated by cytokine IL-2, which induces cytotoxic attack by increasing intracellular content of granzyme B (FIG. 1A). In this Examples, the effects of IL-2-stimulated natural killer cells on DRG neurons acutely isolated from embryonic (E15) and adult mice (<24 hr in vitro) were examined using isolated, unstimulated splenic natural killer cells as a control (FIG. 1B). As reported, embryonic DRG neurons were highly susceptible to natural killer cell-mediated cytotoxicity (FIGS. 2A and 2B). Separating natural killer cells and embryonic DRG with a trans-well membrane prevented neurite fragmentation (FIG. 10) despite the presence of nanogram amounts of granzyme B in the culture media (FIG. 1D), indicating that the cytotoxic effect on DRG neurons requires direct contact with natural killer cells. In contrast, in acutely isolated adult DRG neurons co-cultured with IL-2-stimulated natural killer cells, no evidence of cell lysis by LDH release cytotoxicity assay was found even at much higher effector-to-target (E:T) ratios (FIG. 2B), although some neurites appeared truncated (FIG. 2A). Time-lapse confocal imaging of Dil-labelled DRG neurons co-cultured with IL-2 stimulated yellow fluorescent protein (YFP)-expressing natural killer cells revealed that stimulated natural killer cells were highly motile, enabling direct cell-cell contacts between natural killer cells and sensory neurons. In embryonic DRG, this led to neurite destruction and cell death (not shown 1); however, acutely cultured adult DRG neurite membranes remained largely intact with no attendant cell death despite similar direct cell-cell contacts (not shown 2). Next, in the present Example, DRG neurons were loaded with the Ca.sup.2+-sensitive indicator rhodamine 3-AM before applying IL-2 stimulated natural killer cells, and the time-lapse imaging was performed. Natural killer cell contacts were often synchronous with intracellular Ca.sup.2+ events in embryonic DRG neurites (FIGS. 2C, 2D, and not shown 3), leading to neurite fragmentation. While Ca.sup.2+ events were also observed in adult DRG after natural killer cell contact (FIGS. 2C and 2D), and occasionally, led to neurite fragmentation (not shown 4), they occurred at a much lower frequency than in embryonic neurons (FIG. 1D). The protein Retinoic Acid Early 1 (RAE1), encoded by the Raet1 gene family (α, β, γ, δ, ε), acts as a membrane-bound ligand for the mouse activating receptor NKG2D (Cerwenka et al., 2000), which has previously been implicated in natural killer cell-mediated lysis of embryonic DRG (Backstrom et al., 2003); however, a functional requirement for Raet1 in natural killer cytotoxicity against DRG neurons has yet to be shown. First, in this Example, it was confirmed that natural killer cytotoxicity against embryonic DRG neurons may be attenuated by an NKG2D receptor blocking antibody (FIG. 1D). Using RT-PCR with universal primers designed to detect all five Raet1 isoforms (α, β, γ, δ, ε), Raet1 transcripts were observed in acutely dissociated embryonic and adult DRG neurons (FIG. 2E). Quantitative RT-PCR (qRT-PCR) showed that Raet1 transcripts were 17 times more abundant in embryonic DRG relative to adult DRG (FIG. 2F). This difference was reflected at the protein level: a single large band at approximately 40 kDa to 50 kDa was detected by a pan-RAE1 antibody in western blot of embryonic but not adult DRG tissue (FIG. 2G). To assess the functional contribution of Raet1 in embryonic DRG neurons, all Raet1 isoforms were selectively knocked out using Raet1 siRNA (FIG. 2H). Compared to negative control siRNA, transfection of embryonic DRG neurons with Raet1-selective siRNA led to a 20% reduction in natural killer cell-mediated lysis (FIG. 2I). These results show that Raet1 expression in embryonic DRG neurons is involved in the cytotoxicity induced by direct engagement with stimulated natural killer cells.

Example 2. Confirming that Nerve Injury Drives RAE1 Expression in Adult Sensory Neurons Allowing Cytotoxic Attack by Activated Natural Killer Cells

[0139] To investigate the potential effects of activated natural killer cells on adult sensory neurons, DRG neurons were cultured in a microfluidic chamber (5 days in vitro), which allows the selective exposure of their axons to stimulated natural killer cells in the neurite compartment. Following exposure to stimulated natural killer cells, 25% loss of neurite coverage relative to isolated control natural killer cells was observed (FIGS. 3A and 3B). Next, it was examined whether RAE1 plays a role in natural killer cell-mediated neurite fragmentation of these longer-term adult DRG cultures. Raet1 mRNA was time-dependently upregulated in adult DRG cultures (FIG. 3C), with corresponding de novo expression of RAE1 protein after 2 days in vitro (FIG. 3D). Consistent with a role for RAE1 in susceptibility to natural killer cell attack, adult DRG neurons cultured for 2 days displayed over a 10-fold increase relative to controls in neurite fragmentation in the presence of stimulated natural killer cells (FIGS. 3E and 3F). Transfection of dissociated adult DRG neurons with Raet1 siRNA prior to culture delayed Raet1 upregulation (FIG. 3G) and reduced the ability of stimulated natural killer cells to fragment DRG neurites relative to negative control siRNA (FIGS. 3H and 31). It was also confirmed that fragmentation of adult DRG neurites (2 days in vitro) can be reduced by blocking the NKG2D receptor on natural killer cells prior to co-culture (FIGS. 4A and 4B). Thus, de novo expression of RAE1 in long-term adult DRG cultures confers susceptibility to NKG2D-mediated neurite fragmentation by natural killer cells. It was also examined whether Raet1 upregulation in dissociated DRG in vitro occurs in injured cells in vivo. To address this, the fifth lumbar spinal nerve distal to the DRG (L5x) in adult mice was cut (FIG. 5A). An injury-dependent upregulation of Raet1 was revealed by qRT-PCR using RNA isolated from L5 DRG at 4 and 7 days after L5x injury (FIG. 5B). The difference in Raet1 transcript levels between ipsilateral and contralateral L5 DRG 7 days after L5x injury was maintained over the first 24 hr in culture (FIG. 5C). When IL-2 stimulated natural killer cells were introduced to these cultures, neurons from injured DRG (L5x) exhibited 3.5-fold more fragmentation of neurites (FIGS. 5F and 5G) than those from DRG from uninjured (sham) animals (FIGS. 5D and 5E). These results suggest that peripheral nerve injury drives Raet1 induction in adult sensory neurons and this expression correlates with natural killer cell-mediated neurite destruction.

Example 3. Confirmed that Stimulated Natural Killer Cells Degenerate Partially Injured Sensory Axons

[0140] To characterize the response and function of natural killer cells in vivo, a mouse expressing the inducible Cre recombinase (iCre) driven by the Ncr1 gene promoter (Ncr1.sup.icre) (Narni-Mancinelli et al., 2011) was used to cross with reporter mice which express either enhanced yellow fluorescent protein (Rosa26eYfP, YFP) (Srinivas et al., 2001) or the diphtheria toxin receptor (Rosa26dtr, DTR) (Buch et al., 2005) following cre-mediated recombination. Ncr1 encodes the NKp46 receptor, which is expressed on all natural killer cells, as well as tissue-resident ILCs, including group 1 ILCs (ILC1s) and a subset of group 3 ILCs (NCR.sup.1+ ILC3s). This genetic approach allowed to identify (via YFP expression) or systemically deplete (via DTR) NKp46.sup.+ cells in vivo (FIGS. 6A, 6B and 6C). YFP-positive natural killer cells were not observed in sciatic nerve in naive mice or sham surgery mice (FIGS. 5H and 5I). In contrast, spinal nerve transection (L5x) led to the marked recruitment of YFP-positive natural killer cells to the injured sciatic nerve (FIGS. 5H and 5I). Natural killer cell infiltration to the sciatic nerve was paralleled by a large increase in granzyme B levels at 7 days post-injury (FIG. 5J), which was significantly reduced by prior systemic depletion of natural killer cells by DTx treatment in NKp46-DTR mice (FIG. 5K).

[0141] To examine the potential role of natural killer cells in the degeneration of axons after injury, complete sciatic nerve crush injury was performed with fine forceps (FIG. 7A). Consistent with the effect of nerve transection, crush injury drove Raet1 expression in ipsilateral DRG (FIG. 7B). As RAE1 is known to be controlled by various forms of post-transcriptional modification (Raulet et al., 2013), Western blot of lysates of sciatic nerve and DRG tissue following crush and sham injury was performed using a pan-RAE1 antibody, which recognizes all isoforms of the protein. Crush injury induced a higher molecular weight band (approximately 45 kDa) in ipsilateral sciatic nerve but not in contralateral nerve or sciatic nerve from sham-injured mice. The 45 kDa size indicates an increase in the glycosylated mature form of RAE1 (Arapovic et al., 2009) in ipsilateral injured sciatic nerve (FIG. 7C) but not in the DRG (FIG. 7D). Furthermore, chronic ligation of the nerve to block axonal transport revealed RAE1 immunolabeling within axons proximal to the crush site (FIG. 7E). These data suggest that there is axonal-specific shipping and/or translation of RAE1 protein at the injury site. Crush injury in NKp46-YFP mice also caused recruitment of natural killer cells within damaged ipsilateral sciatic nerve (FIG. 8A), and flow cytometry showed that the number of CD45.sup.+YFP.sup.+ lymphocyte-gated events (indicative of natural killer cells) detected in ipsilateral sciatic nerve (FIG. 8B) increased by 30-fold 3 days after injury and a further 50% by day 7 (FIG. 8C). Crush injury also raised levels of granzyme B in the sciatic nerve (FIGS. 9A and 9B). Using two-photon imaging of the sciatic nerve in vivo 3 days after crush injury, NKp46-YFP cells rolling along the wall of intra-neural blood vessels (not shown 5) was observed. Within the sciatic nerve itself, natural killer cells were highly motile and displayed multipolar morphology, reminiscent of stimulated natural killer cells in vitro (not shown 6). These cells are ideally placed to interact functionally with damaged peripheral nerve axons that selectively express the natural killer cell target protein RAE1. No YFP.sup.+ cells were observed by two-photon imaging in the nerves of sham injured mice (Data not shown).

Example 4. Confirming that Endogenous NK Response Attenuates Post-Injury Sensitivity

[0142] Sciatic nerve crush injury induces acute loss of sensation and rapid Wallerian degeneration of primary afferent axons distal to the injury site, followed by sensory recovery within approximately two weeks (Ma et al., 2011). After crush with fine forceps, complete loss of response to pin-prick stimulation of the lateral hind paw was confirmed in all mice (FIGS. 8D and 8E). Control (PBS-treated) NKp46-DTR mice showed a general pattern of sensory recovery (Ma et al., 2011; Painter et al., 2014), remaining almost completely insensitive until a relatively rapid recovery beginning around 10 days to 11 days post injury (FIG. 8D, PBS). To investigate the role of natural killer cells in axonal integrity following nerve crush injury in vivo, peripheral sensitivity was assayed in NKp46-DTR mice chronically depleted of natural killer cells by diphtheria toxin (DTx) treatment (FIG. 8G). These mice displayed an early and persistent sensitivity to pin prick stimulation (FIG. 8D, DTx). In contrast to control mice, which showed a heel-to-toe pattern of sensory recovery (FIG. 8F, PBS), natural killer cell-depleted mice showed an anatomically broad pattern of responses to stimulation throughout the hind paw prior to full recovery (FIG. 8F, DTx), indicative of remaining sporadic innervation throughout the tissue.

Example 5. Confirming that Stimulated Natural Killer Cells Degenerate Partially Injured Sensory Axons in Traumatic Peripheral Neuropathy Mouse Model

[0143] To determine whether natural killer cells is able to selectively degenerate partially injured axons, a consistent partial sciatic crush was delivered by using a fine hemostat with 30 μm spacer. Control animals generally showed some residual pin prick sensitivity one day after crush, followed by a rapid functional recovery to over 60% of maximum by day 6 (FIG. 10A, IgG), suggesting that a proportion of fibers were incompletely axotomized. This partial sensitivity was maintained until 11 days post-crush, at which point sensory recovery occurred in a heel-to-toe pattern, reaching full recovery by day 15 (FIG. 10B, IgG), consistent with regenerating fibers. To replicate the stimulation of natural killer cells by I L-2 in vivo, mice were systemically treated with an IL-2/anti-IL-2 monoclonal (S4B6) antibody complex. IL-2/anti-IL-2 antibody complex treatment is known to enhance both NK and CD8.sup.+ T cell populations (Boyman et al., 2006). Indeed, both NKp46.sup.+DX5.sup.+ and CD3.sup.+CD8.sup.+ (FIG. 10D) cell populations were enriched in the blood of IL-2 complex-treated mice at the end of the experiment (Day 16) compared to controls, whereas the proportion of CD3.sup.+CD4.sup.+ T cells was decreased (FIG. 10D). Compared to IgG-treated controls, animals treated with the IL-2 complex after injury showed a significant acute deterioration in pin prick response (FIGS. 10A and 10C, IL-2 complex); however both groups of animals regained full pin prick sensitivity at around two weeks (FIG. 10B). Further, the specific contribution of natural killer cells in this process was determined. The effect of IL-2 complex was examined in the mice, after pre-treatment with an anti-NK1.1 antibody to systemically deplete natural killer cells. In isotype control-treated mice, IL-2 complex treatment again resulted in acute loss of the pin prick response (FIGS. 10E, 10F, and 10G, Isotype), but natural killer cell-depleted mice maintained a constant level of ‘retained’ sensitivity throughout the paw until regeneration continued from 11 days after injury (FIGS. 10E, 10F, and 10G). The dependence of natural killer cells on the effect of IL-2 complex was further confirmed in DTx-treated wild-type and NKp46-DTR mice (FIGS. 11A and 11B). Flow analysis of peripheral blood after recovery (Day 16) confirmed depletion of NKp46.sup.+DX5.sup.+ natural killer cells by anti-NK1.1 in wild-type mice (FIG. 10H) and by DTx in NKp46-DTR (FIG. 11C). No significant difference was observed in CD3.sup.+CD8.sup.+ T cell populations, although the proportion of CD3.sup.+CD4.sup.+ T cells was 20% higher in the antibody depleted group compared to isotype controls (FIG. 10H). To determine the mechanism behind the natural killer cell-dependent sensory loss after IL-2 complex treatment, axon fibers within the sciatic nerve were examined 6 days after partial crush. The increase in density of β-tubulin III-labelled axon fibers from the crush site extending distal throughout the sciatic nerve in control (IgG-treated) mice was reduced by IL-2 complex-treatment (FIGS. 12A and 12C, insets B and C). Axon density in the proximal nerve region, however, was not different between treatment groups (FIGS. 12A and 12C, inset A). The distribution of the microtubule-associated protein stathmin 2 (STMN2), which is specifically expressed by injured DRG neurons, was also examined (Cho et al., 2013). Similar to the distribution of β-tubulin III, STMN2 labeling was observed throughout the crushed nerve, and was reduced by IL-2 complex treatment at the crush site and distal regions relative to IgG controls (FIGS. 12B and 12C, insets B and C). STMN2 also co-localized with RAE1 immunolabeling at the injury site (FIG. 13). These results suggest that the temporary sensory loss observed with IL-2 complex treatment after partial crush injury is due to removal of injured axon fibers within the sciatic nerve.

Example 6. Confirming that Clearance of Partially Damaged Axons Alleviates Pain Hypersensitivity after Injury in Traumatic Peripheral Neuropathy Mouse Model

[0144] In this Example, the mechanical threshold after recovery from partial crush injury (Days 15-16) was assessed by measuring responses to von Frey hair stimulation of the lateral hind paw. IL-2 complex treatment resulted in significantly higher mechanical thresholds in the hind paw of the previously crushed (ipsilateral) limb, compared to control (IgG) animals, which displayed relative hypersensitivity (FIG. 12D). Furthermore, normalization of mechanical thresholds by IL-2 complex treatment after partial crush was prevented by prior depletion of natural killer cells (FIG. 12E). Combining data from all animals revealed that the cumulative pin prick sensitivity (area under the curve) during the peak effect of treatment (days 5 to 10) correlates with the resulting mechanical sensitivity of the injured limb after recovery from a partial crush (FIG. 12F). Taken together, these data suggest that effective clearance of partially damaged axons by natural killer cells following nerve injury is likely to be important to prevent the appearance of persistent mechanical sensitivity from retained partially damaged axons within the sciatic nerve.

Example 7. Confirming that Natural Killer Cells have Function to Remove Nerves in Mouse Model of Peripheral Neuropathy Caused by Administration of Anticancer Drug

[0145] To examine whether natural killer cells are also able to selectively remove injured or abnormal nerves in peripheral neuropathy caused by administration of an anticancer drug, interactions between natural killer cells and anticancer drug-exposed sensory nerve cells were observed and ligand expression of natural killer cells was measured in mouse models of peripheral neuropathy caused by administration of the anticancer drug.

[0146] In detail, an anticancer drug (oxaliplatin, 10 mg/kg) was injected into the intraperitoneal cavity of mice to induce peripheral neuropathy due to administration of the anticancer drug, and the injured peripheral sensory nerve cells were isolated on day 7, primarily cultured, and then co-cultured with natural killer cells to examine the interaction between the two cells. As a control, peripheral sensory nerve cells of a normal mouse were used. FIG. 14 shows results confirming that peripheral neuropathy caused by administration of the anticancer drug was induced since pain was induced on day 7 in the anticancer drug-administered group. FIG. 15 shows results of examining the interaction between natural killer cells and peripheral sensory nerve cells which were cultured on day 7 after administration of the anticancer drug. FIG. 16 shows results of examining ligand expression of natural killer cells in sensory nerve cell tissues on day 7 after administration of the anticancer drug.

[0147] As a result, as shown in FIG. 15, it was confirmed that natural killer cells selectively removed the anticancer drug-exposed sensory nerve cells.

[0148] Further, as shown in FIG. 16, no changes in natural killer cell-activating ligands Raet1 and Mult1 were observed, but an inhibitory ligand Qa1b was significantly reduced, indicating that attack of natural killer cells was induced due to reduction of the inhibitory ligand in the target cells.

[0149] These results suggest that natural killer cells can also be used to remove injured nerves in peripheral neuropathy caused by administration of the anticancer drug, which may be mediated by Qa1b.

Example 8. Confirming that Infiltration of Natural Killer Cells into Injured Nerve Site Increases in Mouse Models with Central Nervous System Excitotoxicity

[0150] To examine whether natural killer cells are able to selectively remove injured or abnormal nerves in the central nervous system, infiltration of natural killer cells into the central nervous system was examined in central nervous system excitotoxicity-induced mice.

[0151] In detail, Kainic acid (35 mg/kg) was injected into the abdominal cavity of mice to induce excitotoxicity of the central nervous system, and saline-injected mice were used as a control. Central nervous system excitotoxicity is a phenomenon observed in representative brain diseases such as Alzheimer's disease, epilepsy, Parkinson's disease, etc. (Dong et al., 2009), and a central nervous system excitotoxicity model is also known as a representative seizure model. Kainic acid is a potent neuroexcitatory amino acid agonist that acts by activating the glutamate receptor, which is a major excitatory neurotransmitter in the central nervous system, and treatment of mice with high concentrations of kainic acid leads to excitotoxicity and neurodegeneration in the central nervous system (Levesque et al., 2013). FIGS. 17A to 17C show images of selective infiltration of NK cells into brain tissues in which excitotoxicity of the central nervous system was induced.

[0152] As a result, as shown in FIGS. 17A to 17C, it was confirmed that natural killer cells (NKp46 expression) infiltrated into the brain tissues only in the kainic acid-treatment group.

Example 9. Confirming that Natural Killer Cells have Function to Remove Nerves in Central Nervous System Excitotoxicitv Mouse Model

[0153] To examine whether natural killer cells are also able to selectively remove injured or abnormal nerves in the central nerve system, analysis of behavioral responses and measurement of ligand expression of natural killer cells were performed in central nervous system excitotoxicity-induced mice.

[0154] In detail, Nkp46-DTR transgenic mice that conditionally lack natural killer cells (normal when treated with PBS, and lack of natural killer cells when treated with Diphtheria toxin (DTx)) were treated with a high concentration of kainic acid (35 mg/kg) which acts on the central nervous system to cause excitotoxicity and to induce degeneration, and behavioral responses were analyzed based on Racine's stage criteria. Criteria used when seizure behaviors were observed are shown in [Table 1] below, and a higher stage indicates more serious.

TABLE-US-00001 TABLE 1 Stage Behavior 1 Mouth and facial movement 2 Head nodding 3 Forelimb clonus 4 Rearing with forelimb clonus 5 Rearing and falling with forelimb clonus (generalized motor convulsions)

[0155] FIG. 18 shows results of analyzing behavioral responses occurring after induction of excitotoxicity and degeneration of the central nervous system in Nkp46-DTR transgenic mice. FIG. 19 shows results of measuring ligand expression of natural killer cells in hippocampal tissues on day 3 after induction of excitotoxicity and degeneration of the central nervous system.

[0156] As a result, as shown in FIG. 18, it was confirmed that the behavioral responses due to neurodegeneration appeared less when natural killer cells were deficient.

[0157] Further, as shown in FIG. 19, increase of Raet1 and Mult1, which are activating ligands of natural killer cells, was observed, and expression of an inhibitory ligand, Qa1b, was also slightly increased. However, since target cell attack of natural killer cells is regulated by the relative expression levels of the activating ligands and the inhibitory ligand of the cells, these result means that the injured nerve-removing function of natural killer cells is also activated in the central nerve system.

[0158] The above results suggest that natural killer cells are able to degenerate the injured central nerve causing excitotoxicity, which may be mediated by the activating ligands (RAE1 and MULTI) highly expressed in the injured nerve. It means that natural killer cells may be usefully applied in the treatment of seizures or symptoms caused by nerve degeneration, which occurs due to nervous system diseases caused by nerve injury or abnormal nerves in the central nervous system.