Enhancing fractalkine/CX3CR1 signalling pathway can reduce neuroinflammation by attenuating microglia activation in experimental diabetic retinopathy

Abstract The concept of diabetic retinopathy (DR) has been extended from microvascular disease to neurovascular disease in which microglia activation plays a remarkable role. Fractalkine (FKN)/CX3CR1 is reported to regulate microglia activation in central nervous system diseases. To characterize the effect of FKN on microglia activation in DR, we employed streptozotocin‐induced diabetic rats, glyoxal‐treated R28 cells and hypoxia‐treated BV2 cells to mimic diabetic conditions and explored retinal neuronal apoptosis, reactive oxygen species (ROS), as well as the expressions of FKN, Iba‐1, TSPO, NF‐κB, Nrf2 and inflammation‐related cytokines. The results showed that FKN expression declined with diabetes progression and in glyoxal‐treated R28 cells. Compared with normal control, retinal microglia activation and inflammatory factors surged in both diabetic rat retinas and hypoxia‐treated microglia, which was largely dampened by FKN. The NF‐κB and Nrf2 expressions and intracellular ROS were up‐regulated in hypoxia‐treated microglia compared with that in normoxia control, and FKN significantly inhibited NF‐κB activation, activated Nrf2 pathway and decreased intracellular ROS. In conclusion, the results demonstrated that FKN deactivated microglia via inhibiting NF‐κB pathway and activating Nrf2 pathway, thus to reduce the production of inflammation‐related cytokines and ROS, and protect the retina from diabetes insult.


| INTRODUC TI ON
Diabetic retinopathy (DR), one of the common complications of diabetes mellitus, is the leading cause of blindness among working people worldwide. 1 DR, traditionally considered as a microvascular disease, is also considered as neurodegeneration characterized by neuronal apoptosis and reactive gliosis. [2][3][4][5] Microglia, the resident monocytes of the retina, monitors the microenvironment of the retina and interact with other cells via numerous specific receptors. 6 As the main immune cells in retina, microglia participates the pathogenesis of DR by morphological transition from the resting-ramified morphology to activated-amoeboid phenotype and releasing the inflammation-related cytokines. 3,4,7 It is widely reported that microglia activation results in microvascular damage, neurodegeneration and retinal inflammation, 4,7 leading to the dysfunction of neurovascular unit in retina.
Fractalkine (FKN), also known as CX3CL1, exists as two forms of ligands, that is the membrane-bound and soluble forms. Both forms can bind the receptor CX3CR1. FKN is synthesized principally by neurons while CX3CR1 is merely expressed by microglia in retina. 8,9 A considerable amount of literatures shows that FKN/CX3CR1 signalling pathway, as an entry point of interaction between neurons and microglia, plays an important role in central nervous system diseases by modulating/inhibiting microglia activation. [10][11][12][13][14][15] In experimental DR, CX3CR1 deficiency potentiated retinal microglia changes, 16 enhanced the inflammatory response as well as neuronal damage. 17 In addition, CX3CR1 deficiency activated microglia, disrupted the vascular integrity and accelerated DR progression. 9,18 Furthermore, in mouse model of retinitis pigmentosa, FKN, delivered either by intravitreal injection or gene therapy, was protective for retinal neurons. 19,20 Based on the previous studies, we hypothesized that FKN administration might ameliorate microglia activation and proinflammatory cytokine release via CX3CR1 inhibitory signals in experimental DR, thus protecting the retina from diabetes insult by decreasing neuroinflammation.
To verify the hypothesis, we characterized the effect of FKN on microglia activation in experimental DR by employing streptozotocininduced diabetic rat, glyoxal-treated R28 cells and hypoxia-treated BV2 cells. The results showed that FKN treatment could ameliorate microglia activation via inhibiting NF-κB pathway and enhancing Nrf2 pathway, thus to reduce the production of inflammation-related cytokines and ROS, and protect the retina from diabetes insult.

| Reagents and antibodies
Recombinant rat fractalkine (E. coli-derived chemokine domain of rat fractalkine protein, Gln25-Gly100) was purchased from R&D were purchased from HyClone. Foetal bovine serum was purchased from Gibco (10091148, USA). Penicillin/streptomycin (15140155) was purchased from Invitrogen. Pierce BCA protein assay kit (23225) was purchased from Thermo Scientific. The information of the antibodies, used for Western blot and immunofluorescence, was provided in Table 1. TJHBLAC-2020-06). Male Sprague-Dawley rats with body weight 120-160 g were purchased from Slaccas, maintained under 12-hour light/dark cycle and were given ad libitum access to food and water.

| Experimental animals
To induce diabetes, the rats were injected intraperitoneally with STZ (60 mg/kg BW in citric acid buffer) after fasted for 24 h, while the normal control received an equal volume of citric acid buffer according to our previous study. 21 Blood glucose level was measured by glucometer for three consecutive days, and rats with blood glucose levels more than 16.6 mmol/L were included and considered as diabetic rats. Two hours after STZ injection, the right eyes of diabetic rats were injected intravitreally with recombinant FKN protein (0.2 mg/eye, 2 μl) as D+F group, while the left eyes were received the equivalent volume of phosphatebuffered saline (PBS) as D group. The age-matched normal control was injected intravitreally with equivalent volume of PBS and designated as N group. Rats were sacrificed at 4 to 12 weeks of diabetes.

| Retinal sample preparation
The rats were sacrificed with cervical dislocation after complete anaesthesia. Both eyes were enucleated and immediately fixed in 4% PBS-buffered paraformaldehyde overnight at 4℃.
For cryosection, the eyecups were dehydrated through a gradient concentration of sucrose from 10% to 30%. In addition, then, the eyecups were embedded in OCT compound (Tissue Tek; Sakura).
To prepare retinal flatmount, the neurosensory retina was isolated carefully and flattened on microscope slide with four radial cuts.

| Cell viability assay
The

| Protein extraction and western blot
The samples, including the retinas, R28 cells and BV2 cells, were

| Statistical analysis
All experiments were repeated at least three times. Data were expressed as mean ± SE. The statistical analysis was carried out by using 2-tailed Student's t test; a P value of 0.05 or less was considered statistically significant.

| FKN expression declined with diabetes progression and the neuronal apoptosis increased in diabetic rat retinas
To detect the expression of FKN and apoptosis of retinal cells in experimental DR, we performed Western blot assay and TUNEL assay in diabetic rat retinas. Western blot densitometry analysis results showed the protein expression of FKN decreased significantly compared with that in normal control, that is decreased by 17.8% (4 weeks, n = 9, p=0.038), 41.4% (8 weeks, n = 9, p = 0.016) and 46.0% (12 weeks, n = 9, p = 0.022) respectively ( Figure 1C). We also detected the retinal cell apoptosis at week 4 after diabetes onset, and the result showed that apoptosis, especially in outer nuclear layer (ONL), increased compared with that in normal control ( Figure 1D).  (Figures 1-2).

| FKN attenuated retinal microglia activation in diabetic rats
Considering that the FKN receptor (CX3CR1) is uniquely expressed in microglia, which is activated in early DR, 3,7 the lack of inhibitory effect on CX3CR1 due to decreased FNK might activate microglia in diabetic rat retinas. We explored the effect of exogenous FKN on retinal microglia activation after intravitreal injection. According to Western blot densitometry analysis results, compared with that in normal control, the protein expressions of Iba-1 and TSPO (microglia markers) increased significantly by 55.3% (n = 8, p = 0.014, Figure 3A) and 84.7% (n = 8, p = 0.0033, Figure 3B). The surge of Iba-1 and TSPO in diabetic rat retinas was decreased by 31.0% (n = 8, p = 0.029, Figure 3A) and 42.7% (n = 8, p = 0.0045, Figure 3B) following FKN exposure. To further confirm the effect of FKN on microglia, we performed Iba-1 immunostaining on the retinal flatmount. As shown in Figure 3C, compared with the normal control, the microglia became activated in diabetic rat retinas with increasing number and more amoeboid morphologies, which was largely abolished by FKN. The above results indicated the inhibitory effect of FKN on microglia activation might be via CX3CR1.

| FKN treatment reduced retinal inflammation in diabetic rats
Microglia activation and its secretion of inflammation-related cytokines are important components of neuroinflammation in DR. 3,7 Meanwhile, the FKN/CX3CR1 signalling pathway, involved in the inflammatory networks, 23

| FKN suppressed NF-κB activation in hypoxiainduced microglia
To study the molecular mechanisms by which the FKN suppressed the inflammatory factors in microglia, we detected the activation of the transcription factor NF-κB, which is wildly reported to be related to microglia-associated neuroinflammation. 7,25 Compared with the normal control, the Western blot densitometry analysis result showed the ratio of phospho-NF-κB p65 to total NF-κB p65 was increased significantly in hypoxia group, increased by 36.4% (n = 8, p < 0.0001), which was decreased significantly by 22.6% (n = 8, p < 0.0001) after FKN treatment ( Figure 6A-B). To further characterize the activation of NF-κB, we performed immunostaining and monitored the translocation of NF-κB to the nucleus in BV2 cells. As shown in Figure 6C and Figure 6D, NF-κB was mainly distributed in cytoplasm in normal control group, while hypoxia induced a significant translocation of NF-κB from the cytoplasm to the nucleus, indicating the enhancing transcriptional activity for proinflammatory factors. Interestingly, we were able to observe that FKN treatment could largely inhibit the translocation of NF-κB.

| FKN enhanced the activation of Nrf2 and elimination of ROS in hypoxia-induced microglia
Nrf2 is a transcription factor which enters the cell nucleus to promote the expression of antioxidant-response genes leading to elimination of ROS 22,[26][27][28] in addition to its anti-inflammation effects. 29,30 Moreover, previous studies suggested that Nrf2 may be a relevant downstream target of FKN to limit microglial over-activation and decrease neuroinflammation. 26,31 Hence, we detected the changes of Nrf2 and ROS production in hypoxia-treated microglia with or without FKN.
As shown in Figure 7A and Figure (Figures 1-2).  22,26,31 Our study also demonstrated that FKN treatment enhanced the Nrf2 activation and accelerated ROS elimination in hypoxia-induced microglia (Figures 7 and 8). Nonetheless, sophisticated crosstalk between Nrf2 and NF-ĸB signalling pathways remains completely unknown 28,48 and deserved further study.

| DISCUSS ION
The limitation of the current study is lack of relevant inhibitors for proteins such as Nrf2, NF-ĸB and soluble CX3CR1, to elucidate the upstream and downstream molecular events from FKN/CX3CR1 to the expressions of the inflammation-related cytokines.

| CON CLUS IONS
The present study demonstrated that the decreased FKN expression was associated with the retinal neuronal cell apoptosis and microglia activation in experimental DR. Exogenous FKN via intravitreal injection attenuated microglia activation and inflammation-related cytokines in diabetic rat retina. Furthermore, FKN exerted its regulatory effect on microglia to reduce the neuroinflammation and ROS production through suppressing NF-κB pathway and enhancing Nrf2 pathway ( Figure 8D). Thus, FKN might be a potential candidate, which warrants further investigation for the treatment of DR and other neuroinflammatory diseases.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data available on request from the authors.