Natural Killer Cells Degenerate Intact Sensory Afferents following Nerve Injury

doi:10.1016/j.cell.2018.12.022

. 2019 Feb 7;176(4):716-728.e18.

doi: 10.1016/j.cell.2018.12.022. Epub 2019 Jan 31.

Natural Killer Cells Degenerate Intact Sensory Afferents following Nerve Injury

Alexander J Davies¹, Hyoung Woo Kim², Rafael Gonzalez-Cano³, Jahyang Choi⁴, Seung Keun Back⁵, Seung Eon Roh⁶, Errin Johnson⁷, Melanie Gabriac⁸, Mi-Sun Kim⁴, Jaehee Lee⁹, Jeong Eun Lee⁵, Yun Sook Kim¹⁰, Yong Chul Bae¹⁰, Sang Jeong Kim⁶, Kyung-Mi Lee⁵, Heung Sik Na⁵, Priscilla Riva³, Alban Latremoliere¹¹, Simon Rinaldi¹², Sophie Ugolini⁸, Michael Costigan¹³, Seog Bae Oh¹⁴

Affiliations

¹ Dental Research Institute and Department of Neurobiology and Physiology, School of Dentistry, Seoul National University, Seoul 03080, Republic of Korea; Nuffield Department of Clinical Neurosciences, University of Oxford, John Radcliffe Hospital, Oxford OX3 9DU, UK.
² Department of Brain and Cognitive Sciences, College of Natural Sciences, Seoul National University, Seoul 08826, Republic of Korea.
³ Departments of Anesthesia and Neurobiology, Children's Hospital Boston and Harvard Medical School, Boston, MA 02115, USA.
⁴ Dental Research Institute and Department of Neurobiology and Physiology, School of Dentistry, Seoul National University, Seoul 03080, Republic of Korea.
⁵ Departments of Physiology, Biochemistry and Molecular Biology, College of Medicine, Korea University, Seoul 02841, Republic of Korea.
⁶ Department of Physiology, Seoul National University College of Medicine, Seoul 03087, Republic of Korea.
⁷ Sir William Dunn School of Pathology, University of Oxford, South Parks Road, Oxford OX1 3RE, UK.
⁸ Aix Marseille Univ, CNRS, INSERM, CIML, Centre d'Immunologie de Marseille-Luminy, Marseille, France.
⁹ Dental Research Institute and Department of Neurobiology and Physiology, School of Dentistry, Seoul National University, Seoul 03080, Republic of Korea; Departments of Physiology, Biochemistry and Molecular Biology, College of Medicine, Korea University, Seoul 02841, Republic of Korea.
¹⁰ Department of Anatomy and Neurobiology, School of Dentistry, Kyungpook National University, Daegu 700-412, Korea.
¹¹ Neurosurgery Department, Johns Hopkins School of Medicine, Baltimore, MD 21287, USA.
¹² Nuffield Department of Clinical Neurosciences, University of Oxford, John Radcliffe Hospital, Oxford OX3 9DU, UK.
¹³ Departments of Anesthesia and Neurobiology, Children's Hospital Boston and Harvard Medical School, Boston, MA 02115, USA. Electronic address: michael.costigan@childrens.harvard.edu.
¹⁴ Dental Research Institute and Department of Neurobiology and Physiology, School of Dentistry, Seoul National University, Seoul 03080, Republic of Korea; Department of Brain and Cognitive Sciences, College of Natural Sciences, Seoul National University, Seoul 08826, Republic of Korea. Electronic address: odolbae@snu.ac.kr.

PMID: 30712871
PMCID: PMC6418410
DOI: 10.1016/j.cell.2018.12.022

Natural Killer Cells Degenerate Intact Sensory Afferents following Nerve Injury

Alexander J Davies et al. Cell. 2019.

. 2019 Feb 7;176(4):716-728.e18.

doi: 10.1016/j.cell.2018.12.022. Epub 2019 Jan 31.

Authors

Affiliations

¹ Dental Research Institute and Department of Neurobiology and Physiology, School of Dentistry, Seoul National University, Seoul 03080, Republic of Korea; Nuffield Department of Clinical Neurosciences, University of Oxford, John Radcliffe Hospital, Oxford OX3 9DU, UK.
² Department of Brain and Cognitive Sciences, College of Natural Sciences, Seoul National University, Seoul 08826, Republic of Korea.
³ Departments of Anesthesia and Neurobiology, Children's Hospital Boston and Harvard Medical School, Boston, MA 02115, USA.
⁴ Dental Research Institute and Department of Neurobiology and Physiology, School of Dentistry, Seoul National University, Seoul 03080, Republic of Korea.
⁵ Departments of Physiology, Biochemistry and Molecular Biology, College of Medicine, Korea University, Seoul 02841, Republic of Korea.
⁶ Department of Physiology, Seoul National University College of Medicine, Seoul 03087, Republic of Korea.
⁷ Sir William Dunn School of Pathology, University of Oxford, South Parks Road, Oxford OX1 3RE, UK.
⁸ Aix Marseille Univ, CNRS, INSERM, CIML, Centre d'Immunologie de Marseille-Luminy, Marseille, France.
⁹ Dental Research Institute and Department of Neurobiology and Physiology, School of Dentistry, Seoul National University, Seoul 03080, Republic of Korea; Departments of Physiology, Biochemistry and Molecular Biology, College of Medicine, Korea University, Seoul 02841, Republic of Korea.
¹⁰ Department of Anatomy and Neurobiology, School of Dentistry, Kyungpook National University, Daegu 700-412, Korea.
¹¹ Neurosurgery Department, Johns Hopkins School of Medicine, Baltimore, MD 21287, USA.
¹² Nuffield Department of Clinical Neurosciences, University of Oxford, John Radcliffe Hospital, Oxford OX3 9DU, UK.
¹³ Departments of Anesthesia and Neurobiology, Children's Hospital Boston and Harvard Medical School, Boston, MA 02115, USA. Electronic address: michael.costigan@childrens.harvard.edu.
¹⁴ Dental Research Institute and Department of Neurobiology and Physiology, School of Dentistry, Seoul National University, Seoul 03080, Republic of Korea; Department of Brain and Cognitive Sciences, College of Natural Sciences, Seoul National University, Seoul 08826, Republic of Korea. Electronic address: odolbae@snu.ac.kr.

PMID: 30712871
PMCID: PMC6418410
DOI: 10.1016/j.cell.2018.12.022

Abstract

Sensory axons degenerate following separation from their cell body, but partial injury to peripheral nerves may leave the integrity of damaged axons preserved. We show that an endogenous ligand for the natural killer (NK) cell receptor NKG2D, Retinoic Acid Early 1 (RAE1), is re-expressed in adult dorsal root ganglion neurons following peripheral nerve injury, triggering selective degeneration of injured axons. Infiltration of cytotoxic NK cells into the sciatic nerve by extravasation occurs within 3 days following crush injury. Using a combination of genetic cell ablation and cytokine-antibody complex stimulation, we show that NK cell function correlates with loss of sensation due to degeneration of injured afferents and reduced incidence of post-injury hypersensitivity. This neuro-immune mechanism of selective NK cell-mediated degeneration of damaged but intact sensory axons complements Wallerian degeneration and suggests the therapeutic potential of modulating NK cell function to resolve painful neuropathy through the clearance of partially damaged nerves.

Keywords: Wallerian degeneration; autoimmunity; dorsal root ganglia; innate immunity; natural cytotoxicity; neurodegeneration; neuroimmune; neuropathic pain; peripheral neuropathy; sciatic nerve crush.

PubMed Disclaimer

Figures

**Figure 1**
Acutely Cultured Embryonic but Not Adult DRG Neurons Reveal Susceptibility to NK-Mediated Cytotoxicity by RAE1 (A) Immunolabeling of co-culture (4 h) between embryonic (top) or adult (bottom) DRG neurons (β-tubulin, magenta) and either freshly isolated (control) or IL-2-stimulated natural killer (NK) cells (NKp46, green). The inset shows a high-magnification image of NK cell in contact with embryonic DRG neurite. (B) LDH-release cytotoxicity assay of acutely cultured (1 day *in vitro*) embryonic (top) and adult (bottom) DRG at various Effector (NK):_target (DRG) (E:T) ratios. Matched two-way ANOVA: embryonic DRG, F(1,10) = 100.01, p < 0.0001); adult DRG, F(1,10) = 1.25, p = 0.2982). Three replicate co-cultures for each DRG group. (C) Still images of *in vitro* time-lapse confocal Ca²⁺ imaging of rhodamine 3 AM-loaded embryonic (top) and adult (bottom) DRG (magenta) co-cultured with IL-2-stimulated NK cells (green) isolated from adult male NKp46-YFP mice. (D) Frequency histogram (30 s time bins) of neurite Ca²⁺ events in embryonic (top) and adult (bottom) DRG during NK co-culture. Cumulative area under the curve (right). Student’s paired t test; t = 2.290, p = 0.045. n = 6 fields of view from two repeat co-cultures per group. (E) RT-PCR of mRNA transcripts in freshly isolated splenic NK cells and embryonic and adult DRG. (F) qRT-PCR shows higher *Raet1* mRNA expression in embryonic compared to adult DRG tissue. Student’s paired t test; t = 16.16, p < 0.0001. n = 5 mice, or replicates per group. (G) Western blot of embryonic and adult mouse DRG tissue (40 μg loading) with pan-RAE1 antibody and β-actin control. Images are representative of three independent experiments. (H) Selective siRNA knockdown reduces RAE1 protein (top) and *Raet1* mRNA (bottom) expression in embryonic DRG (2 d culture). Student’s unpaired t test; t = 9.060, p = 0.0008. n = 3 mice, or replicates per group. (I) LDH-release cytotoxicity assay of negative control or *Raet1*-selective siRNA knockdown embryonic DRG. Three replicate co-cultures for each siRNA group. Matched two-way ANOVA F(1,10) = 133.85, p < 0.0001). See also Figure S1.

**Figure 2**
*Raet1* Expression Is Upregulated in Dissociated Adult DRG *In Vitro* and Confers NK Cell-Mediated Neurite Fragmentation (A) Microfluidic culture of adult DRG (5 days *in vitro*) exposed to freshly isolated (control) or IL-2-stimulated NK cells (4 h) in the neurite compartment immunolabeled with β-tubulin III. (B) Quantification of DRG neurite density. n = 6–7 regions per microfluidic device, n = 3 devices per group, two independent experimental repeats. Student’s unpaired t test; t = 5.448, p < 0.0001. (C) *Raet1* mRNA expression in adult DRG cultures by qPCR. One-way ANOVA; F (3,15) = 25.94, ^∗∗∗p < 0.0001 with Bonferroni post-test: #p < 0.05, ^∗∗p < 0.001, ^∗∗∗p < 0.0001; n = 3–6 mice, or replicate cultures per time point. (D) RAE1 protein expression in adult DRG cultures 1 and 2 days *in vitro* (representative of three independent experiments). 25 μg protein loading. (E) Adult DRG culture (2 days *in vitro*) alone or co-cultured with IL-2-stimulated NK cells immunolabeled with β-tubulin III. (F) Quantification of DRG neurite fragmentation. Student’s unpaired t test; t = 12.74, p < 0.0001. n = 6–8 regions per coverslip, n = 3 coverslips per group, two independent experimental repeats. (G) siRNA knockdown of *Raet1* mRNA expression in adult DRG *in vitro*. Two-way ANOVA: effect of siRNA, F(1,16) = 54.02, ^∗∗∗p < 0.0001 with Bonferroni post-test: ^∗∗p < 0.001, ^∗∗∗p < 0.0001. n = 3–5 cultures per group. (H) siRNA-transfected adult DRG culture (2 days *in vitro*) alone or co-cultured with IL-2-stimulated NK cells (4 h) immunolabeled with β-tubulin III. (I) Quantification of DRG neurite fragmentation. One-way ANOVA; F (3,201) = 51.16, p < 0.0001 with Bonferroni post-test: ^∗∗∗p < 0.0001; n = 6–8 regions per coverslip, n = 6 coverslips per group, two independent experimental repeats. See also Figure S2.

**Figure S2**
Antibody Blockade of NKG2D Receptor on NK Cells Attenuates Neurite Degeneration of Adult DRG Neurons *In Vitro*, Related to Figures 2 and 3 (A) Immunolabeling (β-tubulin III) of cultured adult DRG (10³ cells per coverslip) (2 days *in vitro*) co-cultured (4 h) with IL-2 stimulated NK cells (2.5x10⁵ cells per coverslip) pre-treated with either anti-NKG2D or IgG isotype control antibody. (B) Quantification of DRG neurite fragmentation after NKG2D receptor blockade. One way ANOVA, F(3,193) = 36.82, p < 0.0001 with Bonferroni post-test, ^∗∗∗p < 0.001. n = 6-8 regions per coverslip, n = 4-6 coverslips per group, two independent experimental repeats. (C) L5 DRG neurons were cultured 7 days after L5x injury (24 h in culture) and exposed to IL-2 stimulated NK cells in the presence of anti-NKG2D (CX5) blocking antibody or IgG1k isotype control (30 μg/ml) for 4 h. Images representative of β tubulin III labeling of fixed cultures in each condition. (D) Quantification of DRG neurite fragmentation in acute (< 24 h) cultures of L5x injured DRG neuron-NK co-cultures. One-way ANOVA; F (2,67) = 32.07, p < 0.0001 with Bonferroni post-test:: t = 7.999, t = 4.492; ^∗∗∗p < 0.0001; n = 6-8 regions per coverslip, n = 3 coverslips per group. (E) Quantification of total DRG cell body area in acute (< 24 h) cultures of sham and L5x injured DRG neuron co-cultured with or without control or IL-2 stimulated NK cells (4 h). One-way ANOVA: Sham, F(2,61) = 1.457, p = 0.2409; L5x, F(2,56) = 1.559, p = 0.2194. Bonferroni post-tests: ns, p > 0.05. n = 6-8 regions per coverslip, n = 3-6 coverslips per group, two independent experiments.

**Figure 3**
Peripheral Nerve Injury Regulates Raet1 Expression and Injured Sensory Neurons Show Increased Neurite Fragmentation by Stimulated NK Cells (A) Schematic diagram of spinal nerve transection injury site in relation to lumbar DRGs (L3, L4, and L5) (B) *Raet1* mRNA expression in ipsilateral L5 DRG after spinal nerve transection relative to contralateral DRG by qPCR. Two-way ANOVA: effect of injury, F(1,12) = 121.25, p = 0001; effect of time, F(1,12) = 5.08, p = 0.0438. Bonferroni post-test, ipsi versus contra: 4 days, t = 6.193, ^∗∗∗p < 0.001; 7 days, t = 9.380, ^∗∗∗p < 0.001. n = 4 samples per time point, n = 2 mice DRG pooled per sample. (C) qPCR shows injury-related increase in *Raet1* mRNA expression in adult DRG (1 day *in vitro*). Student’s unpaired t test, t = 4.994, p = 0.0075. n = 3 mice, 3 replicate cultures. (D) Immunolabeling (β-tubulin III) of cultured L5 DRG (1 day *in vitro*) isolated 7 d after sham surgery co-cultured (4 h) with either freshly isolated (control) or IL-2-stimulated NK cells. (E) Quantification of sham DRG neurite fragmentation. One-way ANOVA: F(2,61) = 7.171 p = 0.0016. Bonferroni post-test: DRG only versus control NK, t = 0.4349, nonsignificant (ns) p > 0.05; control NK versus IL-2-stimulated NK, t = 3.221 ^∗∗p < 0.01; DRG only versus IL-2-stimulated NK, t = 3.180, ^∗∗p < 0.01. (F) Immunolabeling (β-tubulin III) of cultured L5 DRG (1 day *in vitro*) isolated 7 days after L5 spinal nerve transection co-cultured (4 h) with either freshly isolated (control) or IL-2-stimulated NK cells. (G) Quantification of L5x DRG neurite fragmentation. One-way ANOVA: F(2,56) = 95.92, p < 0.0001. Bonferroni post-test: DRG only versus control NK, t = 0.5204, ns p > 0.05; control NK versus IL-2-stimulated NK, t = 11.45, ^∗∗∗p < 0.001; DRG only versus IL-2-stimulated NK, ^∗∗∗t = 11.86, p < 0.001. n = 6–8 regions per coverslip, n = 6 coverslips per group, two independent experimental repeats. (H) Sciatic nerve tissue sections from adult male NKp46-YFP mice 7 days after L5 spinal nerve transection injury immunolabeled with anti-GFP (NKp46, green). Arrows indicate NK cells in sciatic nerve. (I) Quantification of YFP⁺ events within lymphocyte gate from whole sciatic nerve homogenates. One-way ANOVA: F(3,16) = 10.61, p = 0.0004. Bonferroni post-test: L5x ipsi versus contra, t = 4.710, ^∗∗p < 0.01; sham ipsi versus contra, t = 0.053, ns p > 0.05; L5x ipsi versus sham ipsi, t = 4.526, ^∗∗p < 0.01 (n = 5 mice for each group). (J) ELISA quantification of granzyme B content in wild-type whole sciatic nerve after L5x injury. Student’s paired two-tailed t test: 7 days ipsi versus contra L5x, t = 8.088, p = 0.0002 (n = 7 mice). (K) ELISA quantification of granzyme B content in NKp46-DTR mice whole sciatic nerve after L5x injury. Student’s unpaired one-tailed t test: 7 days PBS versus DTx L5x ipsi, t = 11.03, p < 0.0001 (n = 5 mice per treatment). See also Figures S2 and S3.

**Figure S3**
Characterization of NKp46-cre Mice, Related to Figures 3 and 4 (A) Flow cytometry of peripheral blood lymphocytes from NKp46-YFP mice labeled with anti-NKp46 and anti-CD3 antibodies. (B) Immunolabeling with an anti-GFP antibody (green) in spleen tissue sections from wild-type and NKp46-YFP mice. Note enhancement of signal only in labeled NKp46-YFP mouse spleen tissue. (C) Flow cytometry of peripheral blood lymphocytes from NKp46-DTR or wild-type (WT) mice 24 h after intravenous treatment with DTx (100 ng) or PBS vehicle (100 μl). Cells labeled with fluorescent-conjugated anti-NKp46 and anti-CD3 antibodies. (D) NKp46-DTR mice were chronically treated with DTx (n = 11) or PBS vehicle (n = 11) (i.v.) followed by L5 spinal nerve transection surgery (Day 0). Sensitivity to mechanical (von Frey filament) stimulation in the injured ipsilateral hind paw before (*naive*) and after injections (*baseline*) and at intervals after L5x injury. Ipsilateral: Two-way ANOVA (effect of time) F(6,140) = 62.24, #p < 0.0001; (effect of depletion) F(1,140) = 1.13, p = 0.2894 (ns, not significant). (E) Uninjured contralateral hind paw. Two-way ANOVA (effect of time) F(6,140) = 0.58, p = 0.7468; (effect of depletion) F(1,140) = 0.11, p = 0.7174.

**Figure 4**
Sciatic Nerve Crush Increases RAE1 in Peripheral Nerve Axons (A) Schematic diagram of sciatic nerve crush injury site in relation to lumbar DRGs (L3, L4, and L5). (B) *Raet1* mRNA expression in ipsilateral L3–5 DRG 3 and 7 days after surgery by qPCR. Two-way ANOVA: effect of injury, F(1,8) = 180.95, p < 0.0001; effect of time, F(1,8) = 8.04, p = 0.0220. Bonferroni post-test: 3 days, t = 7.507, ^∗∗∗p < 0.001; 7 days, t = 11.52, ^∗∗∗p < 0.001. n = 3 mice per surgery per time point. (C) *In situ* hybridization with *Raet1* probe (red) in L4 DRG immunolabeled for NeuN (blue) 6 days after partial sciatic nerve crush injury. Scale bar, 50 μm. (D) Distribution of Raet1 *in situ* spots in L4 DRG ipsi and contralateral to sciatic crush. Kolmogorov-Smirnov (KS) test: ^∗∗∗p < 0.0001, KS D-value 0.2297. n = 3 mice, n = 3 sections per ipsi and contralateral DRG per mouse. (E) (Top) Pan-RAE1 western blot of sciatic nerve tissue (40 μg protein loading) 3 days after sciatic nerve crush or sham surgery. Neuronal cadherin control is shown. (Bottom) Relative expression of RAE1 protein in ipsilateral versus contralateral sciatic nerves 3 and 7 days after sham or crush surgery. One-way ANOVA, 3 days: F(3,16) = 8.505, p = 0.0013. Bonferroni post-test: sham ipsi versus contra, t = 0.2025, ns p > 0.05; crush ipsi versus contra, t = 3.678, ^∗p < 0.05; ipsi sham versus crush, t = 4.438, ^∗∗p < 0.01. One-way ANOVA, 7 days: F(3,16) = 5.119, p = 0.0113. Bonferroni post-test: sham ipsi versus contra, t = 0.1438, ns p > 0.05; crush ipsi versus contra, t = 3.342, ^∗p < 0.05; ipsi sham versus crush, t = 3.043, ^∗p < 0.05 (n = 5 mice per surgery group per time point). (F) (Top) Pan-RAE1 western blot of DRG (L3–5) tissue (40 μg protein loading) 3 days after sciatic nerve crush or sham surgery. Neuronal cadherin control. (Bottom) Relative expression of RAE1 protein in ipsilateral versus contralateral DRG 3 and 7 days after sham or crush surgery. One-way ANOVA: 3 days, F(3,16) = 0.438, p = 0.7288; 7 days, F(3,20) = 0.053, p = 0.983 (n = 5–6 mice per surgery group per time point). (G) Composite images of RAE1 immunolabeling in full-length contralateral and ipsilateral sciatic nerves taken from three individual mice 3 days after tight sciatic nerve ligation. Arrow, ligation site; arrowhead, proximal to ligation. (H) Maximum projection images of RAE1 (green) co-localization with axonal marker β-tubulin III (magenta) in sciatic nerve 3 days after tight ligation. Scale bars, 50 μm. See also Figures S3 and S4.

**Figure S4**
Injury-Dependent Regulation of *Raet1* mRNA and RAE1 Protein Expression in Sensory Neurons *In Vivo*, Related to Figure 4 (A) The size of all NeuN+ neuronal profiles in each DRG are shown as a histograms in blue (50 μm² bins). The number of *in situ* ‘spots’ per NeuN+ neuronal profile are overlayed as a scatterplot. The majority of *de novo Raet1*-expressing NeuN+ profiles in ipsilateral DRG appear to be in the small-medium size range (200-500 μm²). A large number of *Raet1 in situ* spots are also seen NeuN+ profiles of 500 μm² and greater in ipsilateral DRG. The overall size distribution of NeuN+ profiles was not different between DRG (Kolmogorov-Smirnov test: p = 0.0513, KS D-value = 0.04502). (B) Maximum projection images of RAE1 (*green*) co-localized with neuronal injury marker STMN2 (*magenta*) in sciatic nerve 3 days after tight ligation. Note lack of STMN2 immunolabeling in uninjured contralateral nerves. Scale bars, 50 μm. (C) western blots of ligated and proximal regions of sciatic nerve tissue 7 days after tight ligation (35 μg protein loading) with pan-RAE1 antibody and neuronal cadherin as loading controls. Blots from two independent experiments.

**Figure S5**
Systemic NK Cell Depletion Prevents Sensory Loss after Sciatic Nerve Crush, Related to Figure 5 (A) Sciatic nerve tissue sections from adult male NKp46-YFP mice 7 days after sciatic nerve crush injury. Inset shows higher magnification of ipsilateral nerve. Arrows indicate co-localization of NKp46-YFP (anti-GFP, *green*) and nuclear (DAPI, *blue*) labeling. (B) Flow cytometry scatterplots of sciatic nerve homogenates obtained from an NKp46-YFP mouse 3 days after sciatic nerve crush injury. (C) Quantification of total number of CD45+/YFP+ double-positive events within lymphocyte FSC/SSC gate from whole sciatic nerve homogenates. Two-way ANOVA, Crush: Effect of injury, F(1,16) = 168.63, p < 0.0001; Effect of time, F(1,16) = 8.10, p = 0.0117. Bonferroni post-test: 3 days ipsi versus contra crush, t = 7.426, ^∗∗∗p < 0.001; 7 days ipsi versus contra crush, t = 10.94, ^∗∗∗p < 0.001 (n = 5 mice per time point). (D) ELISA quantification of granzyme B content in whole sciatic nerve 7 days after crush injury in wild-type mice. ipsi *versus contra,* t = 7.112 (paired one-tailed t test, n = 8 mice); (E) NKp46+/DX5+ double-positive lymphocytes in peripheral blood 16 days after sciatic nerve crush. Student’s unpaired t test: t = 8.012, p < 0.0001. (F) NKp46 immunolabeling and nuclear (DAPI) staining in sciatic nerve sections (14 μm) 7 days after forceps crush injury in NKp46-DTR mice treated with DTx or PBS vehicle. (G) Quantification of NKp46+DAPI+ cells (white arrows) within a 0.4 mm² region of the nerve crush site were counted by an investigator blind to the treatment. Student’s unpaired t test, t = 2.898, p = 0.0442. n = 2-3 sections per nerve, n = 3 mice per treatment. (H) ELISA quantification of granzyme B content in whole sciatic nerve 7 days after crush injury in NKp46-DTR mice after treatment with PBS or DTx. t = 2.024 (unpaired one-tailed t test, n = 4 mice per treatment). (I) Daily pinprick response score in NKp46-DTR mice treated intravenously with DTx or PBS vehicle every 4 to 5 days. Full sciatic nerve crush on day 0. Two-way ANOVA: Effect of depletion, F(1,352) = 25.20, p < 0.0001). n = 12 mice per group. Bonferroni post-test, ^∗p < 0.05 (t = 3.048). (J) Heatmap showing mean sensitivity to pinprick along the lateral hind paw. (K) Peripheral blood sampled 16 days post-injury shows almost a complete loss of NKp46+DX5+ NK cells in DTx-treated NKp46-DTR mice compared to wild-type. Student’s unpaired t tests., t = 7.761, p < 0.0001. (L) CD3+CD8+ (t = 0.8590, p = 0.4030) and CD3+CD4+ (t = 1.971, p = 0.0662) T cell populations were not different between genotypes after DTx treatment (Student’s unpaired t tests). (M) 8 wild-type and 10 NKp46-DTR mice received partial crush of the sciatic nerve on day 0 as well as diphtheria toxin (DTx) i.v. starting one day before surgery and continuing every 4 to 5 days. All mice were treated with IL-2/anti-IL-2 antibody complex i.p. daily for four consecutive days from the evening of day 2 (arrows). Daily pinprick response shows a transient reduction in sensitivity in wild-type mice receiving the IL-2/anti-IL-2 antibody complex and DTx treatment but not NKp46-DTR mice. Two-way ANOVA. Effect of genotype: F(1,275) = 12.49, p = 0.0005). Bonferroni post-test, ^∗∗p < 0.01 (t = 3.789). (N) Heatmap showing mean sensitivity to pinprick along the lateral hind paw.

**Figure 5**
IL-2/Anti-IL-2 Antibody Complex Treatment Triggers NK Cell-Dependent Acute Sensory Loss after Partial Sciatic Nerve Crush (A) Peripheral blood sampled 16 days post-injury in IL-2 complex or IgG control mice. NKp46⁺DX5⁺ NK cells (U = 15.00, p = 0.0019); CD3⁺CD8⁺ T cells (t = 15.78, p < 0.0001); CD3⁺CD4⁺ T cells (t = 9.719, p < 0.0001). Mann-Whitney or Student’s unpaired t test was used. (B) Daily pinprick response. Male wild-type C57BL/6 mice received partial crush of the sciatic nerve on day 0 followed by daily injection of IL-2 complex or IgG control (i.p.) for 4 consecutive days (arrows). Two-way ANOVA was used. Effect of treatment: F(1,315) = 20.69, p < 0.0001. Bonferroni post-test ^∗∗p < 0.01 (t = 3.784), ^∗∗∗p < 0.001 (t = 4.741). (C) Heatmap showing mean sensitivity to pinprick along the lateral hind paw. Note the broad loss of sensitivity at day 6 in IL-2-complex-treated mice. (D) Area under the curve measurements in IL-2-complex-treated and IgG control mice. Days 1–4: p = 0.9513, t = 0.06181; days 5–10: ^∗∗∗p = 0.0083, t = 2.916; days 11–15: p = 0.4354, t = 0.7951. Student’s unpaired t test was used. (E) Effect of IL-2 complex treatment on NKp46-YFP⁺ cell infiltration to sciatic nerve 6 days after partial crush injury. Sciatic nerve sections (14 μm) were immunolabeled with anti-GFP and β-tubulin III antibodies. Scale bars, 100 μm. Arrows indicate individual YFP⁺ cells. (F) Quantification of NKp46-YFP⁺ cells per square millimeter in images of different regions of the nerve. Two-way ANOVA: effect of IL-2 complex, F(1,16) = 56.31, p < 0.0001; effect of region, F(3,16) = 23.73, p < 0.0001. Bonferroni post-tests: crush, t = 8.062; distal, t = 6.414 ^∗∗∗p < 0.001. ns, not significant. n = 3 sections per region, per mouse, per treatment. (G) Photograph of spleens isolated from mice 1 day after final injection of IgG or IL-2 complex. (H) Peripheral blood sampled 16 days post-injury in IL-2-complex-treated wild-type mice, which received either anti-NK1.1 antibody or isotype control. NKp46⁺DX5⁺ NK cells (t = 15.37, ^∗∗∗p < 0.0001), CD3⁺CD8⁺ T cells (U = 22.00, p = 0.1473), and CD3⁺CD4⁺ T cells (t = 3.035, p = 0.0079). Student’s unpaired t test was used. (I) Loss of NKp46-YFP⁺ cells from peripheral blood in anti-NK1.1-antibody-treated mice. Student’s unpaired t test, t = 15.51, ^∗∗∗p = 0.0001. (J) Wild-type mice received either anti-NK1.1 to deplete NK cells or isotype control antibody followed by partial crush of the sciatic nerve on day 0. All mice were treated with IL-2 antibody complex (arrows). Two-way ANOVA. Effect of antibody depletion: F(1,240) = 21.21, p < 0.0001. Bonferroni post-test, ^∗∗∗p < 0.001 (t = 4.542). (K) Heatmap showing mean sensitivity to pinprick along the lateral hind paw. Note the broad loss of sensitivity at day 6 in isotype control mice. (L) Area under the curve measurements in isotype control and anti-NK1.1-treated mice following IL-2 complex treatment. days 1–4: t = 0.1508, p = 0.8815; days 5–10: t = 2.390, p = 0.0295; days 11–15: t = 1.040, p = 0.3130). Student’s unpaired t test was used. See also Figure S5.

**Figure S6**
NKG2D Antibody Block Attenuates Effect of IL-2 Complex Treatment on Sensory Recovery after Partial Sciatic Nerve Crush, Related to Figure 6 (A) Adult male C57BL/6 mice received a total of 200 μg anti-mouse NKG2D (CX5 clone) (n = 3) or IgG1k isotype control (n = 3) via retroorbital injection (i.v.) over 4 days (Ogasawara et al., 2004): 50 μg (Day 0), 50 μg (Day 2), 100 μg (Day 4). One day later (Day 5) PBMCs were isolated and labeled with PE-conjugated anti-mouse NKG2D or isotype controls, as well as cell surface NK cell markers NKp46 and DX5. Flow cytometry data were gated on NKp46+DX5+ lymphocytes from 70,000 events. (B) Histogram of NKG2D-PE fluorescence labeling on NKP46+DX5+ gated lymphocytes relative to isotype-PE controls (gray filled). Blue, IgG1k isotype control injection; Red, NKG2D blocking antibody injection. (C) Quantification of mean NKG2D immunofluorescence (IgG isotype subtracted) in NKp46+DX5+ cell population. Student’s t test, t = 9.763, p = 0.0006; n = 3 mice per treatment. (D) Daily pinprick responses in wild-type mice which received either blocking anti-NKG2D antibody (n = 12) or IgGk1 isotype control (n = 7) on days 0, 2 and 4 following partial (moderate) crush of the sciatic nerve. All mice were treated with IL-2/anti-IL-2 antibody complex i.p. daily for four consecutive days from the evening of day 2 (arrows). Two-way ANOVA. Effect of antibody: F(1,187) = 8.03, ^∗∗p = 0.0051). (E) Area under the curve measurements. Days 1-4: t = 0.1739, p = 0.8640; Days 5-10: t = 2.13, p = 0.0451; Days 11-15: t = 1.059, p = 0.3044. Student’s unpaired t test. (F) Mechanical sensitivity thresholds of ipsilateral (injured: U = 12.00 p = 0.0122, Mann-Whitney U test) and contralateral (uninjured: U = 39.00, p = 0.8303, Mann-Whitney U test) hind paw 16 days after partial crush injury.

**Figure 6**
IL-2/Anti-IL2 Antibody Complex after Partial Crush Results in Loss of Myelinated Fibers in Sciatic Nerve (A) β-tubulin III immunolabeling of full-length ipsilateral sciatic nerve sections (14 μm) 6 days after partial crush in mice treated with IgG (top) and IL-2 complex (bottom). Insets a, b, c represent proximal, crush site, and distal regions, respectively. (B) High-magnification images of β-tubulin III immunofluorescence from insets in (A). Scale bars, 50 μm. (Right) Quantification of β-tubulin III fluorescence density of 9,000 μm² area of proximal, crush site, and proximal regions. Two-way ANOVA: effect of region: F(2,64) = 57.39 (p < 0.0001). Effect of IL-2 complex treatment: F(1,64) = 7.36 (p = 0.0106). Bonferroni post-test ^∗p < 0.05 (t = 2.558), ^∗∗∗p < 0.001 (t = 3.981). (n = 5 mice per treatment, 3–4 sections per mouse per region, two experimental repeats). (C) Stathmin 2 immunolabeling of full-length ipsilateral sciatic nerve sections 6 days after partial crush in mice treated with IgG (top) and IL-2 complex (bottom). (D) High-magnification images of STMN2 immunofluorescence from insets in (B). Scale bars, 50 μm. (Right) Quantification of STMN2 fluorescence density. Two-way ANOVA: effect of region: F(2,58) = 39.95 (p < 0.0001). Effect of IL-2 complex treatment: F(1,58) = 10.24 (^∗∗p = 0.0033). Bonferroni post-test ^∗∗p < 0.01 (t = 3.346), ^∗∗∗p < 0.001 (t = 3.846). (n = 5 mice per treatment, 3–4 sections per mouse per region. (E) Transmission electron micrographs of cross-sections of ipsilateral sciatic nerve in IgG, IL-2 complex, and anti-NK1.1 antibody treated mice 6 days after partial crush injury. Scale bars, 10 μm. (F) Higher-magnification TEM images showing examples of “normally” myelinated, “abnormally” myelinated, and degenerated axons (denoted by asterisks). Scale bars, 2 μm (black) and 5 μm (white). (G) Quantification of normal myelinated, abnormal myelinated, and degenerated fiber classifications in the different treatment groups. Kruskal-Wallis one-way ANOVA with Dunn’s multiple comparison test. Kruskal-Wallis statistic: normal (26.50), abnormal (25.03), and degenerated (51.81),^∗p < 0.05, ^∗p < 0.01, ^∗∗∗p < 0.001. n = 10 fields of view per nerve section, n = 3 mice per treatment. See also Figure S6.

**Figure 7**
NK Cell Stimulation Post-injury Rescues Long-Term Mechanical Hypersensitivity after Partial Nerve Crush (A) Effect of partial or full crush of the sciatic nerve on daily pinprick responses. Wild-type mice received either partial crush or full crush on day 0. Two-way ANOVA. Effect of crush: F(1,165) = 186.25, p < 0.0001. Bonferroni post-test ^∗∗p < 0.01 (t = 3.68, 4.07), ^∗∗∗p < 0.001 (t = 4.19–6.14). (B) Area under the curve measurements showing cumulative difference in pinprick sensitivity between partial and fully crushed sciatic nerve. Student’s unpaired t test: days 1–4, t = 4.176, p = 0.0.0015; days 5–10, t = 4.674, p = 0.0007; days 11–15, t = 2.502, p = 0.0294. (C) Heatmap showing mean sensitivity to pinprick along the lateral hind paw. (D) Mechanical sensitivity thresholds of the ipsilateral hind paw after partial or full crush injury at day 16 (t = 4.550 p = 0.0008, unpaired t test) and 30 days (U = 3.50, p = 0.0149, unpaired Mann-Whitney test) after injury. (E) Mechanical sensitivity thresholds of ipsilateral hind paw 16 days after partial crush injury in mice treated with IL-2 complex or IgG control (U = 32.00, p = 0.0370, Mann-Whitney test). (F) Mechanical sensitivity thresholds of ipsilateral hind paw 15 days after partial crush injury and IL-2/anti-IL-2 complex treatment in mice that received anti-NK1.1 antibody or isotype control (U = 7.000, p = 0.0049, Mann-Whitney test). (G) Mechanical sensitivity outcomes in the injured limb correlated with the cumulative pinprick sensitivity (area under curve, days 5–10) during the peak effect of treatment (Spearman’s correlation: r = −0.5958, p < 0.0001). See also Figure S7.

**Figure S7**
Specific Overexpression of *Raet1* in Nociceptive (TrpV1-Expressing) Neurons Reduces Sensitivity to Noxious Heat but Not Touch, Related to Figure 7 (A) qPCR shows increase in *Raet1* mRNA expression in DRG from TrpV1-Rae1 mice. One-way ANOVA; F (2,18) = 9.346; p = 0.0016 with Bonferroni post-test, t = 3.885, ^∗∗p < 0.01. n = 5 (C57BL/6), 8 (littermates and TrpV1-Rae1 mice). (B) Latency to withdrawal from a 49°C hotplate. Cut-off time was set at 60 s. TrpV1-Rae1 mice showed reduction on heat pain perception compare to littermate controls. One-way ANOVA; F (2,21) = 22.31; p < 0.0001 with Bonferroni post-test, t = 4.831, ^∗∗∗p < 0.001). n = 8 mice per genotype. (C) TrpV1-Rae1 mice did not differ from littermate or wild-type controls in their mechanical sensitivity, as measured by the 50% threshold with von Frey filaments of different forces applied to the plantar surface of the hind paw. One-way ANOVA; n = 8 mice per genotype (ns, not significant).

See this image and copyright information in PMC

Comment in

Pain killers.
Bordon Y. Bordon Y. Nat Rev Immunol. 2019 Mar;19(3):136-137. doi: 10.1038/s41577-019-0137-4. Nat Rev Immunol. 2019. PMID: 30733595 No abstract available.

Cited by

Inducible knockout of Clec16a in mice results in sensory neurodegeneration.
Hain HS, Pandey R, Bakay M, Strenkowski BP, Harrington D, Romer M, Motley WW, Li J, Lancaster E, Roth L, Grinspan JB, Scherer SS, Hakonarson H. Hain HS, et al. Sci Rep. 2021 Apr 29;11(1):9319. doi: 10.1038/s41598-021-88895-0. Sci Rep. 2021. PMID: 33927318 Free PMC article.
The Genetics of Neuropathic Pain from Model Organisms to Clinical Application.
Calvo M, Davies AJ, Hébert HL, Weir GA, Chesler EJ, Finnerup NB, Levitt RC, Smith BH, Neely GG, Costigan M, Bennett DL. Calvo M, et al. Neuron. 2019 Nov 20;104(4):637-653. doi: 10.1016/j.neuron.2019.09.018. Neuron. 2019. PMID: 31751545 Free PMC article. Review.
Mechanisms, Characteristics, and Treatment of Neuropathic Pain and Peripheral Neuropathy Associated with Dinutuximab in Neuroblastoma Patients.
Mastrangelo S, Rivetti S, Triarico S, Romano A, Attinà G, Maurizi P, Ruggiero A. Mastrangelo S, et al. Int J Mol Sci. 2021 Nov 23;22(23):12648. doi: 10.3390/ijms222312648. Int J Mol Sci. 2021. PMID: 34884452 Free PMC article. Review.
Natural killer cells modulate motor neuron-immune cell cross talk in models of Amyotrophic Lateral Sclerosis.
Garofalo S, Cocozza G, Porzia A, Inghilleri M, Raspa M, Scavizzi F, Aronica E, Bernardini G, Peng L, Ransohoff RM, Santoni A, Limatola C. Garofalo S, et al. Nat Commun. 2020 Apr 14;11(1):1773. doi: 10.1038/s41467-020-15644-8. Nat Commun. 2020. PMID: 32286313 Free PMC article.
NK cells associate with ALS in a sex- and age-dependent manner.
Murdock BJ, Famie JP, Piecuch CE, Pawlowski KD, Mendelson FE, Pieroni CH, Iniguez SD, Zhao L, Goutman SA, Feldman EL. Murdock BJ, et al. JCI Insight. 2021 Jun 8;6(11):e147129. doi: 10.1172/jci.insight.147129. JCI Insight. 2021. PMID: 33974561 Free PMC article.

See all "Cited by" articles

References

1. Arapovic J., Lenac T., Antulov R., Polic B., Ruzsics Z., Carayannopoulos L.N., Koszinowski U.H., Krmpotic A., Jonjic S. Differential susceptibility of RAE-1 isoforms to mouse cytomegalovirus. J. Virol. 2009;83:8198–8207. - PMC - PubMed
1. Backström E., Chambers B.J., Kristensson K., Ljunggren H.G. Direct NK cell-mediated lysis of syngenic dorsal root ganglia neurons in vitro. J. Immunol. 2000;165:4895–4900. - PubMed
1. Backström E., Chambers B.J., Ho E.L., Naidenko O.V., Mariotti R., Fremont D.H., Yokoyama W.M., Kristensson K., Ljunggren H.G. Natural killer cell-mediated lysis of dorsal root ganglia neurons via RAE1/NKG2D interactions. Eur. J. Immunol. 2003;33:92–100. - PubMed
1. Backström E., Ljunggren H.G., Kristensson K. NK cell-mediated destruction of influenza A virus-infected peripheral but not central neurones. Scand. J. Immunol. 2007;65:353–361. - PubMed
1. Bennett G.J., Doyle T., Salvemini D. Mitotoxicity in distal symmetrical sensory peripheral neuropathies. Nat. Rev. Neurol. 2014;10:326–336. - PMC - PubMed

Publication types

Actions
Actions

MeSH terms

Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions

Substances

Actions
Actions
Actions
Actions
Actions

Supplementary concepts

Actions

Grants and funding

LinkOut - more resources

Full Text Sources
Other Literature Sources
- H1 Connect
- The Lens - Patent Citations Database
Medical
- MedlinePlus Health Information
- The YODA Project
Molecular Biology Databases
- Mouse Genome Informatics (MGI)
Miscellaneous
- SciCrunch

[1] Arapovic J., Lenac T., Antulov R., Polic B., Ruzsics Z., Carayannopoulos L.N., Koszinowski U.H., Krmpotic A., Jonjic S. Differential susceptibility of RAE-1 isoforms to mouse cytomegalovirus. J. Virol. 2009;83:8198–8207. - PMC - PubMed

[2] Arapovic J., Lenac T., Antulov R., Polic B., Ruzsics Z., Carayannopoulos L.N., Koszinowski U.H., Krmpotic A., Jonjic S. Differential susceptibility of RAE-1 isoforms to mouse cytomegalovirus. J. Virol. 2009;83:8198–8207. - PMC - PubMed

[3] Backström E., Chambers B.J., Kristensson K., Ljunggren H.G. Direct NK cell-mediated lysis of syngenic dorsal root ganglia neurons in vitro. J. Immunol. 2000;165:4895–4900. - PubMed

[4] Backström E., Chambers B.J., Kristensson K., Ljunggren H.G. Direct NK cell-mediated lysis of syngenic dorsal root ganglia neurons in vitro. J. Immunol. 2000;165:4895–4900. - PubMed

[5] Backström E., Chambers B.J., Ho E.L., Naidenko O.V., Mariotti R., Fremont D.H., Yokoyama W.M., Kristensson K., Ljunggren H.G. Natural killer cell-mediated lysis of dorsal root ganglia neurons via RAE1/NKG2D interactions. Eur. J. Immunol. 2003;33:92–100. - PubMed

[6] Backström E., Chambers B.J., Ho E.L., Naidenko O.V., Mariotti R., Fremont D.H., Yokoyama W.M., Kristensson K., Ljunggren H.G. Natural killer cell-mediated lysis of dorsal root ganglia neurons via RAE1/NKG2D interactions. Eur. J. Immunol. 2003;33:92–100. - PubMed

[7] Backström E., Ljunggren H.G., Kristensson K. NK cell-mediated destruction of influenza A virus-infected peripheral but not central neurones. Scand. J. Immunol. 2007;65:353–361. - PubMed

[8] Backström E., Ljunggren H.G., Kristensson K. NK cell-mediated destruction of influenza A virus-infected peripheral but not central neurones. Scand. J. Immunol. 2007;65:353–361. - PubMed

[9] Bennett G.J., Doyle T., Salvemini D. Mitotoxicity in distal symmetrical sensory peripheral neuropathies. Nat. Rev. Neurol. 2014;10:326–336. - PMC - PubMed

[10] Bennett G.J., Doyle T., Salvemini D. Mitotoxicity in distal symmetrical sensory peripheral neuropathies. Nat. Rev. Neurol. 2014;10:326–336. - PMC - PubMed

Save citation to file

Email citation

Add to Collections

Add to My Bibliography

Your saved search

Create a file for external citation management software

Your RSS Feed

Natural Killer Cells Degenerate Intact Sensory Afferents following Nerve Injury

Affiliations

Natural Killer Cells Degenerate Intact Sensory Afferents following Nerve Injury

Authors

Affiliations

Abstract

Figures

Comment in

Similar articles

Cited by

References

Publication types

MeSH terms

Substances

Supplementary concepts

Grants and funding

LinkOut - more resources

Full Text Sources

Other Literature Sources

Medical

Molecular Biology Databases

Miscellaneous

Abstract

Figures

Comment in

Similar articles

Cited by

References

Publication types

MeSH terms

Substances

Supplementary concepts

Related information

Grants and funding

LinkOut - more resources

Full Text Sources

Other Literature Sources

Medical

Molecular Biology Databases

Miscellaneous