Down-regulation of NOX4 expression in dorsal root ganglion of spinal cord could alleviate cancer induced bone pain in rats by reducing oxidative stress response

Recent studies have indicating their NADPH oxidases generate ROS such as H2O2, superoxide anion, nitric oxide etc. as byproducts of oxidative phosphorylation. In this study, we will analyze the expression levels of NOX4 and its mechanism of action in CIBP.

A CIBP model was established in rats and NOX4 was knocked down in the dorsal root ganglia via intrathecal injection of speci c lentivirus. The downstream effects of NOX4 knockdown were analyzed by RT-PCR, immuno uorescence and Western blotting.

Results
NOX4 was signi cantly upregulated in the rats with CIBP, and mainly localized in the microglia of the dorsal root ganglion of spinal cord. However, NOX4 knockdown alleviated CIBP by reducing oxidative stress and weakening spinal cord sensitization to pain.

Conclusion
Down-regulation of NOX4 in the dorsal root ganglion of spinal cord can alleviate CIBP by reducing oxidative stress.

Background
Cancer is one of the major causes of death worldwide. Although novel therapeutic strategies have signi cantly improved the survival of most patients, the pain caused by cancer seriously affects their life quality [1,2]. CIBP refers to the complex and intractable pain caused by bone metastasis of highly invasive malignancies such as breast cancer, prostate cancer, lung cancer etc. The pathogenesis of CIBP is complex and has not yet been fully elucidated. Studies show that sustained stimulation of peripheral nociceptors by secreted tumorigenic, osteolytic and in ammatory factors, as well as periosteal traction or compression of peripheral nerves, blood vessels and other tissues due to tumor invasion sensitizes the peripheral neurons to pain. In addition, central sensitization due to aberrant release of excitatory and inhibitory neurotransmitters and overactivation of corresponding receptors is also closely related to CIBP, although the speci c mechanism is still unclear [3][4][5].
Nicotinamide adenine dinucleotide phosphatase (NADPH) oxidase (NOX) is a group of protease complexes that produce reactive oxygen species (ROS) such as H 2 O 2 , superoxide anion, nitric oxide, etc. during oxidative phosphorylation [6,7]. NOX4 was the rst kidney-speci c NOX to be discovered, and has since been detected in nerve cells and vascular smooth muscle cells, and in organelles including mitochondria, nuclei and endoplasmic reticulum. Under normal circumstances, the baseline amounts of ROS are neutralized by superoxide dismutase (SOD), catalase (CAT), glutathione peroxidase (GP X ) and other antioxidants, which maintain a dynamic balance between ROS generation and elimination [6,7].
However, chronic and stressful stimuli can trigger a surge in ROS levels that overwhelm the antioxidant system, leading to oxidative stress and a pathological state. Recent studies show that ROS is involved in regulating sensory transmission and its high levels are responsible for nerve injury-related chronic pain. For instance, ligation of the L5 spinal nerves in a rat model of neuropathic pain signi cantly increased ROS production in the microglia of dorsal root ganglion (DRG) [8]. In addition, Zhou et al. [9] recently showed that the ROS scavenger N-tert-Butyl-α-phenylnitrone (PBN) and tetramethylpiperidine (4-hydroxy-2, 2,6,6-tetramethylpiperidine-1-oxyl, Tempol) reduced the mechanical pain in a rodent CIBP model by inhibiting DRG microglial activation. These ndings indicate that ROS is likely involved in the perception and regulation of pain through the spinal cord, and therefore a causative factor of CIBP. Accordingly, we hypothesized that down-regulating NOX4 in the DRG tissues would alleviate pain by inhibiting ROS production. To this end, we established a rat model of CIBP and speci cally knocked down NOX4 in the DRG neurons through intrathecal lentiviral injection. The changes in pain threshold, ROS production and other related indicators were measured in order to determine the role of NOX4 in CIBP and the underlying mechanisms.

Cell lines and main reagents
Waiker256 rat breast cancer cells were purchased from Shanghai Cell Bank, Chinese Academy of Sciences. RPMI-1640 and fetal bovine serum (FBS) were purchased from Gibico. RIPA lysate and BCA protein quanti cation kit were purchased from Invitrogen, and the reverse transcription and real-time quantitative PCR kit from TaKaRa. The NOX4-targeting (RNAi-LV) and negative control (NC-LV) lentiviruses were purchased from Shanghai KeyGen Biotech. Rabbit anti-NOX4 and anti-GAPDH monoclonal antibodies were procured from Abcam, mouse anti-SOD and nNOS from Santa Cruz, goat anti-NMGAR2D and GABAA-γ2 from CST company, and the rabbit anti-NOX4, goat anti-NeuN, anti-Iba1 and anti-GAFP antibodies from Millipore. Alexa Fluor 488 and Alexa Fluor 555 labeled donkey antigoat/rabbit IgG were purchased from Guangzhou Jingcai Biological, and the HRP-labeled goat antirabbit/mouse IgG, immunocytochemical staining kit and HE staining kit from Wuhan Boster Biological Technology Co. Ltd. The Reactive Oxygen (ROS) Detection Kit was purchased from Nanjing Jiancheng Biological Engineering. All PCR primers were synthesized by Shanghai Biotech.
Intrathecal Catheterization And Lentiviral Injection SPF-grade healthy male SD rats weighing 180 ~ 220 g were purchased from Vital River Company and housed at 22-24 ℃, 55% humidity and a 12 h diurnal cycle, with ad libitum access to food and water. The rats were divided randomly into the sham operated (Sham), untreated CIBP (CIBP), empty vector (CIBP + NC-LV) and NOX4 knockdown (CIBP + RNAi-LV) groups (n = 8 in each group) and treated accordingly. All animal experiments were approved by the ethics committee of the Second Hospital of Jilin University, and conducted in strict accordance with the guidelines of the International Association for Study of pain on the use of animals for pain experimental research. The rats were anesthetized with iso urane (3% for induction and 2% for maintenance). Subarachnoid catheterization was performed in the lumbosacral region as reported previously [10,11]. Brie y, a PE-10 catheter was inserted into the subarachnoid space through the cerebello-medullary cistern to a depth of 7-7.5 cm along the direction of the spinal cord towards the intumescentia lumbalis of the tail. A 3 − 0 silk thread was then passed through the gap of the catheter to suture the muscle and x the catheter in position. After excluding rats with nerve dysfunction or catheter detachment 20 h after catheterization, the guide tube was injected into the remaining animals with 10 mL 2% lidocaine. Paralysis in the hind limbs con rmed correct positioning of the catheter, following which the rats were injected intraperitoneally with 5 mL normal saline. On the third day after intrathecal catheterization, 10 µL RNAi-LV or NC-LV was injected daily through the intrathecal catheter for 3 consecutive days. The sham-operated animals were injected with an equal volume of normal saline. All rats of the same groups were housed in a single cage after catheterization.
Viable cells were harvested, re-suspended in PBS at the density of 2 × 10 7 cells/mL, and inoculated into the abdominal cavity of rats at the volume of 500 µL. After one week of conventional feeding, 10 mL tumor ascites was extracted from each animal and centrifuged at 1000 rpm for 8 min. The pelleted cells were resuspended in PBS at the density of 2 × 10 7 /mL, and a small aliquot was boiled for 20 min for the Sham group rats. The CIBP model was established as previously described [12] . Brie y, the rats were anesthetized (see Sect. 1.2.1) on the third day after lentiviral injection and placed in supine position. The skin of the left hind limb was depilated and disinfected with iodine, and cut longitudinally at the joint of the left knee. The surface of tibial plateau was fully exposed by blunt dissection of the muscle, and the bone was drilled perpendicular to the surface with a No.5 needle till the bone marrow cavity. The needle was then pulled out, and 10 µL (~ 2 × 10 5 cells) of the live or dead Walker256 suspension was injected slowly into the cavity using a 25 µL microsyringe. The latter was drawn after 30 s, and the injection hole was closed with sterile bone wax. After sterilizing the wound, the muscle and skin layers were sutured.

Analgesic Behavioral Testing
The calibrated Von-Frey cilia were used to determine the paw withdraw threshold (PWT) of the hind foot on the affected side. As previously reported by Chaplan et al. [12] , the rats were acclimatized for 30 min prior to the experiment and the bottom of the third and fourth feet of the right hind foot were stimulated by the tip of Von-Frey cilia. The stimulation intensity was steadily increased from 0 to 40 g within 20 s. The time taken for the rat to withdraw or move the foot was recorded as the PWT. Paw withdrawal latency (PWL) of the middle and posterior 1/3rd of the right plantar was evaluated in response to thermal pain. The intensity of thermal irradiation was adjusted to 10 V and 50 W, and the diameter of the stimulated spot was 0.8 cm [13,14]. The reaction time of retracting, adding or lifting the foot, or biting was recorded as PWL. All measurements were taken between 8:00 a.m. to 12:00 noon by the same technicians, and each animal was stimulated thrice at 3 min intervals. The PWT and PWL scores were evaluated before and on days 0, 3, 7, 14 and 21 post-intrathecal catheterization.

Histopathology
The rats were deeply anesthetized by intraperitoneal injection of 10% chloral hydrate on the 21st day after CIBP induction and dissected. The affected tibia was removed and decalci ed in 15% EDTA-2Na for 3 weeks, with the solution replaced weekly. After washing with PBS, the decalci ed tibia was embedded in para n and sectioned. The tissue sections were dewaxed, dehydrated through an alcohol gradient, and stained by HE before microscopic examination.

DRG Harvest And Immunohistochemistry (IHC)
The rats were anesthetized as above, and the dorsal region 2 cm on either side of the median line as well as both hind limbs was depilated. The animals were dissected along the median dorsal line from the tail to the upper part of the hind limbs, and the subcutaneous tissues and muscles were separated to expose the spine. The L4-6 section on the left side of the spinal cord harboring the DRG tissue was extracted, washed with chilled PBS and frozen in liquid nitrogen for molecular assays. For IHC, the chest wall of the anesthetized rats was opened to expose the heart, and 10 mL normal saline was injected into the aorta through the left ventricle following exsanguination. After the saline wash, the animals were perfused with 4% polyformaldehyde, and the DRG tissues were removed and xed in 4% paraformaldehyde for 2 h. The The DRG tissues were washed thrice with chilled PBS and homogenized in RIPA lysis buffer. The lysates were centrifuged at 12,000 rpm at 4 ℃ for 5 min and the supernatants were aspirated. The concentration of H 2 O 2 was measured using the speci c kit as per the manufacturer's instructions.

Real-time Fluorescence Quantitative Pcr (RT-PCR)
Total RNA was extracted from the DRG tissues using Trizol, and the purity and concentration were detected using a spectrophotometer. RNA was reverse transcribed into cDNA using a reverse transcription kit, and RT-PCR was performed according to the kit instructions and reaction conditions standardized by preliminary experiments. The RT-PCR primer sequences are as follows:

Western Blotting
The DRG tissues were washed thrice with chilled PBS, and homogenized with RIPA lysis buffer supplemented with protease inhibitor. Following quanti cation with the BCA kit, 35 µg of each protein sample was mixed with 5 × loading buffer at the ratio of 1:4, denatured by boiling for 10 min, and separated by polyacrylamide gel electrophoresis (SDS-PAGE). The protein bands were transferred to PVDF membrane by wet rotation method and blocked with 5% skim milk at room temperature for 2 h. The blots were then incubated overnight with primary antibodies against SOD (1: 500), nNOS (1: 500), GABAA-γ2 (1: 300), NMDAR2D (1: 500) and GAPDH (1: 1500) at 4 ℃ with constant shaking. After washing thrice with TBST solution for 5 min each time, the membranes were incubated with HRP-labeled secondary antibody (1: 5000) for 1 h at room temperature, and washed again as described. The positive bands were developed using an ECL reagent and exposed in a gel imager. The gray values of the bands were measured by Image J software, and the ratio of the target protein to the internal reference GAPDH was calculated. Three independent experiments were performed. Statistical analysis SPSS 19.0 and GraphPad Prism 5.0 were used for statistical analysis and all data were expressed as `X ± S. Two groups were compared using the independent sample t-test, and multiple groups by one-way ANOVA and Dunnett's or Bonferroni's ad-hoc test. The signi cance test level was α = 0.05, and P < 0.05 was considered statistically signi cant.

Successful construction of CIBP model
The induction of CIBP was veri ed in terms of pathological and behavioral parameters. As shown in Fig. 1a, the tibias of the sham-operated rats were intact with a clear boundary between the bone tissue and cavity, and abundant bone marrow. In contrast, inoculation of tumor cells resulted in visible deterioration of the tissue microstructure and replacement of the bone marrow cells with tumor cells within 7 days, which progressed to severe destruction of the bone matrix and trabeculae, obliteration of the boundary between tissue and marrow cavity, and excessive tumor cell proliferation in the bone marrow by the 14th and 21st day (Fig. 1a-d). The basal mechanical and thermal pain thresholds were similar across all groups. However, the pain threshold of the CIBP rats decreased signi cantly from day 7 post-inoculation, and sensitivity to pain gradually increased over the progression of disease. In contrast, the pain threshold of the sham-operated rats was relatively stable (Fig. 1e, f). Taken together, the CIBP model was successfully established.
NOX4 is upregulated in the DRG microglia upon CIBP induction NOX4 mRNA and protein expression increased signi cantly in the DRG tissues of the CIBP modeled rats compared to the sham-operated animals, and showed a time-dependent upregulation with the progression of the disease (Fig. 2a, b). Furthermore, co-staining of the DRG tissues with NeuN, GFAP or Iba 1 and NOX4 indicated that the latter was primarily localized to the microglia as opposed to the neurons and astrocytes (Fig. 2c-k).

NOX4 Knockdown Signi cantly Alleviated CIBP
To further elucidate the biological relevance of NOX4 in CIBP, we knocked down the gene in the DRG tissues via intrathecal lentiviral injection. As shown in Fig. 3a-b, NOX4 mRNA and protein levels were signi cantly downregulated in the rats infected with RNAi-LV compared to the Sham and NC-LV groups. DRG-speci c knockdown of NOX4 signi cantly alleviated the mechanical and thermal pain sensitivity in the rats, thereby underscoring the causative role of NOX4 in CIBP (Fig. 3c, d).

NOX4 Knockdown Alleviated Oxidative Stress In DRG Tissue
NOX4 signi cantly decreased ROS levels in DRG tissue (Fig. 4a), and concomitantly increased that of the antioxidant enzyme SOD (Fig. 4b). ROS like the superoxide anions combine with NO to form nitrate, which inactivates SOD. In nerve tissues, NO is mainly produced by neuronal nitric oxide synthase (nNOS), which also plays an important role in neuropathic pain [15] . However, NO is di cult to detect in tissues and cells. Nevertheless, we found that down-regulation of NOX4 expression signi cantly inhibited nNOS in DRG tissues (Fig. 4b). Taken together, NOX4 promotes oxidative stress in the DRG microglia by enhancing ROS production and inhibiting SOD, which can be reversed by NOX4 knockdown.

NOX4 Knockdown Alters Neurotransmitter Receptor Activation
Since ROS accumulation can alter the activation status of the glutamate and gamma-aminobutyric acid (GABA) neurotransmitter receptors, we next analyzed the effect of NOX4 knockdown on these receptors. NR2D, the major subtype of the excitatory receptor NMDAR, was signi cantly elevated in the rats with CIBP, whereas the inhibitory receptor GABAA-γ2 was signi cantly decreased. Down-regulation of NOX4 reduced the activation levels of NR2D and signi cantly increased that of GABAA-γ2 (Fig. 5a). Similar results were seen with the mRNA and total protein expression levels of both receptors in the control and NOX4-knockdown DRG tissues (Fig. 5b-c).

Discussion
CIBP is generally the earliest palpable symptom in various malignancies, and most patients seeking treatment for the pain are already in the advanced stages of cancer. However, the pathogenesis of CIBP is complex and ambiguous, leading to unsatisfactory clinical outcomes. Therefore, it is essential to identify the molecular mechanisms underlying CIBP in order to improve treatment strategies [16].
Peripheral sensitization due to continuous stimulation by various in ammatory cytokines, the acidic tumor-bone microenvironment and altered ROS levels, and central sensitization caused by synaptic plasticity are closely related to the occurrence of CIBP [17−19] . Direct injection of cancer cells into the bone marrow cavity can simulate the clinical characteristics of CIBP and therefore help explore its molecular mechanisms. Intrathecal injection of rat breast cancer cells into the tibia marrow cavity of adult female rats resulted in mechanical and thermal hyperalgesia within a week, and increased pain sensitivity over a period of 3 weeks. This corresponded to progressive deterioration of the bone trabeculae and the gradual colonization of the marrow cavity with highly differentiated tumor cells. These results con rmed that CIBP was successfully established in the animal model.
The NOX family of proteases are the major catalysts of ROS generation and oxidative stress. NOX4 is distinct from other NOX isomers on account of its tissue-speci c expression and unique molecular structure. In addition, while activation of NOX1-3 depends on the phosphorylation of cellular subunits and that of NOX5 on Ca 2+ signaling, NOX4 harbors an intrinsic activated dehydrogenase domain that can generate ROS through the electron transfer from NADPH to FDA. However, due to the highly conserved histidine residue in its third extracellular ring, NOX4 could induce superoxide to spontaneously disproportionate into H 2 O 2 in the enzyme, and therefore the ROS produced by NOX4 is mainly H2O2 [20,21]. In this study, we detected a signi cant upregulation of NOX4 mRNA and protein expression during disease progression. In addition, NOX4 was mainly localized in the microglia of DRG tissues. Microglia are the resident macrophages of the central nervous system (CNS), and account for more than 5% of the nerve cells in the brain. When the CNS is subjected to pessimal stimulation, the microglia are rapidly activated and release effectors like ROS, NO and in ammatory factors (e.g. TNF -α, IL-1 β and COX-2), which can signi cantly damage peripheral nerve cells, especially the neurons. Recent studies show that the interaction between neurons and microglia is crucial to central sensitization and the feeling of pain [20,21] . To elucidate the mechanistic role of microglial NOX4 in CIBP, we speci cally knocked it down in the DRG tissues, and observed that the animals were signi cantly sensitized to both mechanical and thermal pain. These preliminary results indicate that NOX4 is a causative factor in CIBP and therefore can be targeted to alleviate bone pain in cancer patients.
The CNS is metabolically demanding due to the high proportion of unsaturated fatty acids, which increases its susceptibility to oxidative stress. The neurons in particular are highly sensitive to ROSinduced injury, which is the mechanistic basis of neurodegenerative disorders like Parkinson's disease, Alzheimer's disease, stroke and other neurological lesions [22] . Hassler et al. [8] found that ROS promoted chronic neuralgia after spinal cord injury, and its neutralization signi cantly relieved the intensity of pain.
Little et al. [23] have postulated that ROS and NO are key players in the development of pain via central sensitization. H 2 O 2 plays an important role in pain perception and regulation as a diffusive signaling molecule. Indeed, H 2 O 2 levels increased signi cantly in the DRG of a rat neuropathic pain model established via spinal cord transection, and induced neuronal apoptosis through the PI3K/AKt signaling pathway. SOD is closely related to the perception and regulation of pain. Clinical studies show that SOD levels are signi cantly lower in patients with chronic migraine compared to healthy controls. Consistent with this, the CIBP rats also expressed lower levels of SOD compared to the sham-operated animals. This could be due to the accumulation of large amounts of ROS that combines with nitric oxide to form peroxynitrite, which in turn reduces the content of SOD through nitri cation. NOX4 knockdown in the DRG signi cantly upregulated SOD, likely by inhibiting the production of H 2 O 2 and nNOS.
Glutamate and γ-aminobutyric acid (GABA) are respectively the major excitatory and inhibitory neurotransmitters in the CNS, and play a key role in in ammation-related pain as well as CIBP [24]. The considerable increase in ROS levels in the bone tumor microenvironment leads to the production and secretion of excessive amounts of glutamic acid by the tumor cells. The accumulation of glutamate increases the excitability of afferent nerves by activating N-methyl-D-aspartic acid (NMDA) receptors on peripheral nerve endings, which in turn alters synaptic plasticity in the DRG and enhances pain responsiveness and central sensitization [24] . In fact, direct injection of ROS into the spinal cord of rats with neuropathic pain not only aggravated the pain response, but also signi cantly reduced the expression of GABA. Introducing the ROS scavenger PBN signi cantly alleviate the above [25] . Consistent with this, knocking down NOX4 in the CIBP model signi cantly downregulated the excitatory receptor NMDAR in the DRG tissue and upregulated the inhibitory receptor GABAA-γ2.
NOX4 is signi cantly upregulated in the DRG microglia of rats with CIBP, and its inactivation alleviated pain by reducing oxidative stress and weakening spinal cord sensitization. Thus, NOX4 is a potential therapeutic target in CIBP and its role needs to be investigated further, along with the additional mechanisms of CIBP.
Abbreviations CIBP, Cancer-induced bone pain; ROS: reactive oxygen species; NOX: NADPH oxidases; CNS: central nervous system; GABA: Glutamate and γ-aminobutyric acid; NMDA: N-methyl-D-aspartic acid Declarations Ethics approval and consent to participate All animal experiments were approved by the ethics committee of the Second Hospital of Jilin University, and conducted in strict accordance with the guidelines of the International Association for Study of pain on the use of animals for pain experimental research.

Availability of data and materials
All data are available in the included gures.

Competing interests
The authors declare that they have no conflict of interests.  The expression and cellular localization of NOX4 in DRG tissues. a-b: NOX4 mRNA (a) and protein (b) levels in the DRG of Sham and CIBP group rats; c-k: Representative immuno uorescence images showing co-staining of NOX4 (green) with NeuN+ (red) neurons, GFAP+ (red) astrocytes or Iba1+ (red) microglia in the DRG tissue of the affected side on the 21st day after tumor inoculation. Compared with Sham group, *P < 0.05, * *P < 0.01.