Effects of CXCR1/2 Blockade with Ladarixin on Streptozotocin-Induced Type 1 Diabetes Mellitus and Peripheral Neuropathy and Retinopathy in Rat
Article information
Abstract
Background
The CXC motif chemokine ligand 8 (CXCL8)-CXC motif chemokine receptor 1/2 (CXCR1/2) axis has been implicated in type 1 diabetes mellitus (T1DM). Its actions on non-immune cells may also contribute to T1DM-associated complications, including painful diabetic peripheral neuropathy (DPN) and diabetic retinopathy (DR).
Methods
We assessed the efficacy of early (4–8 weeks) or late (8–12 weeks) daily ladarixin (LDX) for the treatment of streptozotocin (STZ)-induced T1DM and the related complications of DPN or DR in male rats.
Results
Early LDX mitigated STZ-induced dysmetabolism (i.e., blood glucose, insulin), inflammation in dorsal root ganglion/sciatic nerve (interleukin-1β and tumor necrosis factor-α expression) and mechanical allodynia and thermal hyperalgesia, indicative of DPN. Moreover, vitreous citrullinated histone H3 (CitH3) and plasma GRO/CINC1 (CXCL8) increase were attenuated. Late LDX failed to reverse STZ-induced changes in metabolic parameters (i.e., blood glucose, insulin, C-peptide, pancreatic β-cell number and function). Strikingly, even in the absence of an effect on glycemic control, late LDX mitigated STZ-induced mechanical allodynia and thermal hyperalgesia and vitreous (CXCL8, CitH3) and retinal (CXCL8, CXCR1/2, myeloperoxidase, CitH3) inflammatory/pro-angiogenic (vascular endothelial growth factor, CD34) signs of DR.
Conclusion
These data confirm the efficacy of LDX in STZ-induced T1DM and provide evidence of a protective effect also against DPN and onset of DR which is independent of its effect on β-cell functionality preservation and glycemic control.
Highlights
• Early daily LDX 4–8 weeks after STZ protects against T1D in rats.
• Early daily LDX also prevents signs of STZ-induced DPN and DR.
• Late daily LDX 8–12 weeks after STZ does not reverse T1D, β-cell loss or dysfunction.
• Despite a lack of effect on T1D, late LDX prevents signs of STZ-induced DPN and DR.
INTRODUCTION
The chemokine CXC motif chemokine ligand 8 (CXCL8), or interleukin-8 (IL-8), and its receptors, CXC motif chemokine receptor 1 (CXCR1) and CXCR2, play a pivotal role in the initiation and amplification of inflammation [1]. Primarily produced by neutrophils and macrophages, CXCL8 is a potent neutrophil chemoattractant [2]. The CXCL8-CXCR1/2 axis has been implicated in several autoimmune- and inflammation-based diseases, including type 1 diabetes mellitus (T1DM) [3,4]. T1DM is an autoimmune disease in which autoreactive T-cells target and destroy pancreatic β-cells, leading to insulin deficiency, dysglycemia, and, ultimately, insulin dependence [5]. Importantly, secondary complications associated with T1DM can be severe and require therapeutic management.
Preclinical studies have reported upregulations in circulating CXCL1, the rodent homolog of CXCL8, at the onset of T1DM in spontaneously diabetic biobreeding rats and in streptozotocin (STZ)-induced T1DM [6,7]. A critical role for CXCR2-dependent pancreatic neutrophil recruitment in T1DM development has also been established in non-obese diabetic (NOD) mice [6,8]. Accordingly, the pharmacologic blockade of CXCR1/2 by ladarixin (LDX), a dual CXCR1/2 non-competitive allosteric inhibitor [9], has been shown to block or reverse T1DM in NOD mice [10]. LDX administration at diabetes onset resulted in the inhibition of insulitis and modification of leukocyte distribution in blood, spleen, bone marrow, and lymph nodes. Clinical trials of LDX for the treatment of T1DM are ongoing [11,12].
Correlative evidence suggests that the CXCL8-CXCR1/2 axis may also be involved in diabetic complications. A potential role for CXCL8 in diabetic peripheral neuropathy (DPN) has been suggested by reports that serum CXCL8 was elevated in patients with diabetic neuropathic pain relative to diabetic controls without neuropathic pain [13] and was directly correlated with cold perception threshold in childhood-onset T1DM patients [14]. Ocular CXCL8 has also been shown to be correlated with measures of disease severity and disease progression in diabetic retinopathy (DR) patients in several studies [15,16].
Here, we assessed the effects of repeated LDX on STZ-induced T1DM and signs of DPN and DR. The study seeks to determine whether LDX protective effects against DPN or DR in the context of STZ-induced T1DM, are primarily due to its ability to improve β-cell function and glycemic control, or if LDX directly influences the underlying pathophysiology of DPN or DR.
To this end, the study included two phases: an early-treatment phase, with repeated LDX administered 4 to 8 weeks post-STZ, and a late-treatment phase, with repeated LDX administered 8 to 12 weeks post-STZ, well after the establishment of STZ-induced β-cell dysfunction and hyperglycemia underlying T1DM.
METHODS
T1DM model and study design
Male Wistar rats (225 to 240 g) (Envigo, Udine, Italy), were used. Experimental procedures were approved by the Animal Ethics Committee of University of Campania “Luigi Vanvitelli” and the Ministry of Health (prot. 30/2021-PR). Following 1 week of adaptation, experimental T1DM was induced by a single intraperitoneal (i.p.) injection of STZ at 65 mg/kg in 0.1 M cold citrate buffer [17]. Control (CTRL) rats received 0.1 M citrate buffer (i.p.). Rats were allowed to drink 10% dextrose overnight and then placed back on the standard diet. Rats with blood glucose levels <300 mg/dL 1-week post-STZ were planned to be excluded, though no rats met this criterion and all rats were included in the analyses.
For the early-treatment phase, rats were randomly assigned to one of three groups (CTRL, n=8; STZ/vehicle [VEH], n=12; STZ/LDX, n=12) and STZ-treated rats were administered LDX (15 mg/kg) or VEH via intragastric administration once daily 4 to 8 weeks post-STZ. Primary outcomes were blood glucose, body weight, insulin levels, and C-peptide levels. Blood glucose and body weights were assessed at day 0 and weekly. Additional diabetic/metabolic parameters, including insulin, triglycerides, total cholesterol, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), glycosylated hemoglobin (HbA1c), ketone bodies (acetoacetate [AcAc], β-hydroxybutyrate [BHB]), CXCL8, and citrullinated histone H3 (CitH3) were assessed in plasma at weeks 3, 5, and 7 post-STZ. Signs of DPN assessed included behavioral tests for mechanical allodynia and thermal hyperalgesia on day 0 and at weeks 2, 4, 5, 6, 7, and 8 and analyses of inflammation, or tumor necrosis factor-α (TNFα) and IL-1β expression, in the dorsal root ganglion (DRG) and sciatic nerve (SN) collected at week 8 post-STZ. Signs of DR assessed included vitreous levels of CXCL8 and CitH3 at weeks 3, 5, and 7 post-STZ.
For the late-treatment phase, rats were randomly assigned to one of three groups (CTRL, n=8; STZ/VEH, n=8; STZ/LDX, n=8) and STZ-treated rats were administered LDX (15 mg/kg) or VEH via intragastric administration once daily from 8 to 12 weeks post-STZ. Again, blood glucose and body weights were assessed at day 0 and weekly. For this phase, diabetic/metabolic parameters assessed at weeks 8, 9, and 12 included insulin, C-peptide, calculated insulin resistance, and insulin secretion (homeostasis model assessment for insulin resistance [HOMA-IR] and β-cell function [HOMA-β]), CXCL8 (weeks 9, 12 only) and CitH3 (weeks 9, 12 only). HOMA-IR and HOMA-β were calculated based on the following formula: HOMA-IR=serum insulin (mmol/L)×blood glucose (mmol/L)/22.5 and HOMA-β=[20×fasting insulin (μIU/mL)]/[fasting glucose (mmol/mL)−3.5]. Pancreatic tissue was harvested for analysis of numbers of insulin-positive β-cells and glucagon-positive α-cells at week 9 post-STZ. As signs of DPN, mechanical allodynia and thermal hyperalgesia behavior were assessed on day 0 and at weeks 2, 4, 6, 8, 9, and 12. Signs of DR assessed included vitreous CXCL8 and CitH3 and retinal expression of CXCL8, CXCR1/2, CitH3, myeloperoxidase (MPO), vascular endothelial growth factor (VEGF), and CD34 at weeks 9 and 12. Finally, plasmatic hyperproteinemia was assessed via plasma creatinine levels at weeks 9 and 12. The experimental design is given in Fig. 1. All experimental procedures were counterbalanced across treatment groups to minimize potential confounders and all experimenters were blind to the treatment groups until the completion of the data analysis.
Schematic flow chart of animals, grouping, and design. Experimental design showing model induction, treatments and behavioral and biochemical evaluations. i.g., intragastric; DRG, dorsal root ganglion; SN, sciatic nerve.
Body weight and blood glucose measurements
Body weight was assessed with a Ugo Basile scale (Ugo Basile S.R.L., Gemonio, Italy). Glucose concentrations were obtained from whole blood samples taken from rat tail vein and calculated using a standard clinical blood glucometer set for testing glycemia (GlucoMen LX2, A. Menarini Diagnostics, Winnersh, UK).
Pain behavioral assessments
Von Frey test
Mechanical allodynia was assessed using the von Frey test [18]. All animals were acclimated for about 30 to 45 minutes on an elevated mesh platform in an enclosure (Ugo Basile). Calibrated von Frey filaments (Stoelting, Wood Dale, IL, USA; ranging from 4 to 100 g bending force) were applied to the midplantar surface of the hind paw for 3–4 seconds [18]. The threshold was the lowest force (g) that evoked scratching or licking of the stimulated hind paw.
Tail-flick test
Thermal hyperalgesia was measured by the tail-flick apparatus (Harvard Apparatus, Holliston, MA, USA) [19]. A heat stimulus was applied 5 cm from the caudal tip of the tail. The reaction time between the onset of the heat stimulus and the movement of tail was determined by an automatic sensor and recorded as tail-flick latency (TFL, second). If the animal failed to flick its tail within 15 seconds (cut-off point), the tail was removed from the coil to prevent tissue damage.
Enzyme-linked immunosorbent assay
Plasma concentrations of AcAc, BHB, CitH3, CXCL8, insulin, C-peptide, triglycerides, total cholesterol, LDL-C, HDL‐C, and HbA1c were determined by enzymatic colorimetric methods using commercially available kits. The assay was performed according to the manufacturer’s instructions (Elabscience Biotechnology Inc., Huston, TX, USA; or R&D Systems, Minneapolis, MN, USA).
Protein extraction and Western blot
Western blots were used to analyze IL-1β and TNFα expression in DRG and SN tissues. For protein extraction, tissues were homogenized, then suspended in lysis buffer (Roche, Mannheim, Germany); phosphatase inhibitor cocktail (Roche, Mannheim, Germany) and cleared by centrifugation (10 minutes at 10,000 ×g at 4°C). Lysates were centrifuged for 15 minutes at 13,000 ×g at 4°C, and the supernatants transferred into clear tubes and quantified by detergent compatible protein assay. Protein concentration was determined using Bradford’s method [20] and read by Bio-Rad iMarkTM (Hercules, CA, USA) microplate absorbance reader. Each sample was loaded (30 μg), electrophoresed in a 10% or 12% stands for sodium lauryl sulfate (SDS)-polyacrylamide gel and electroblotted onto a polyvinylidene difluoride membrane (EMD Millipore Corp., Billerica, MA, USA). The membrane was blocked in 3% bovine serum albumin, 1X Tris-buffered saline and 0.01% Tween-20. Primary antibodies against IL-1β (ab283818, Abcam, Cambridge, UK; 1:500) and TNFα (#PA5-19810, Invitrogen, Waltham, MA, USA; 1:500) were used according to the manufacturer’s instruction. Protein levels were normalized to anti-β-tubulin (Santa Cruz Biotechnology, Santa Cruz, CA, USA; 1:1,000 dilution). Reactive bands were detected by chemiluminescence (ECL or ECL-plus, Perkin-Elmer, Waltham, MA, USA). Images were analyzed on a Bio-Rad ChemiDoc XRS+System Reactive. Images were captured, stored, and analyzed using ‘Vision work software.’
Immunofluorescence
Pancreas and retinal tissues were analyzed using immunofluorescence. Anesthetized rats were transcardially perfused with phosphate-buffered saline followed by 4% paraformaldehyde and harvested pancreas and eyes were postfixed 4 hours, cryoprotected for 72 hours in 30% sucrose in 0.1 M phosphate buffer and frozen in optimal cutting temperature embedding compound. The tissue sections (10 µm) were cut using a cryostat (Leica Biosystems, Nussloch, Germany) and thaw mounted onto glass slides. The pancreas slides were incubated overnight with primary antibody against insulin (rabbit Alexa Fluor 647 anti-insulin antibody; 1:300; Abcam) and glucagon (mouse anti-glucagon antibody; 1:500; Abcam). Likewise, the eyes sections with antibodies against histone H3 (rabbit anti-histone H3 antibody; 1:200; Abcam), MPO (mouse anti-MPO antibody; 1:50; Abcam), CXCL1 (mouse CXCL8 monoclonal antibody; 1:400; Thermo Fisher Scientific, Waltham, MA, USA), CXCR1 (rabbit CXCR1 polyclonal antibody; 1:100; Elabscience, Houston, TX, USA), CXCR2 (rabbit CXCR2 polyclonal antibody; 1:100; Elabscience), VEGF (rabbit VEGF recombinant rabbit monoclonal antibody; 1:200; Thermo Fisher Scientific), and CD34 (rabbit CD34 polyclonal antibody; 1:100; Elabscience). Following incubation, sections were washed three times and incubated with the secondary Alexa Fluor conjugated antibodies (1:1,000; Invitrogen). Slides were washed, cover-slipped with Vectashield mounting medium (Vector Laboratories, Newark, CA, USA) and visualized under a Leica fluorescence microscope. Images were analyzed using Image J (National Institutes of Health, Bethesda, MD, USA) as previously described [21].
Statistical analysis
G*Power was used to perform statistical power analyses which supported the determination of sample size for these studies. Data were analyzed using GraphPad Prism version 9.4.0 (GraphPad Software Inc., San Diego, CA, USA). Experiments were analyzed using two-way analysis of variance (ANOVA) tests with treatment group and timepoint as factors with Tukey’s post hoc tests used to probe significant differences. One-way ANOVA tests followed by Dunnett post hoc comparisons were used to assess differences within groups (post-drug vs. pre-drug) for behavioral evaluations and to assess significant differences between groups for ex vivo biochemical evaluations. The level of significance was P<0.05. The Shapiro-Wilk test was used to verify the normality and homogeneity of the data.
RESULTS
Early LDX treatment mitigated both an STZ-induced T1DM phenotype and STZ-induced signs of DPN and DR
Early LDX treatment mitigated STZ-induced alterations in body weight, blood glucose, and other metabolic and inflammatory signs of T1DM
Basal blood glucose levels and body weights were 124.71±1.03 mg/dL and 229.79±1.84 g across groups (Fig. 2A and B). The STZ injection induced a significant increase in plasma glucose levels, emerging 2 weeks post-STZ, which persisted for the duration of the experiment. Early LDX partially mitigated hyperglycemia from 1-week post-LDX (5 weeks post-STZ) (425.2±10.4 mg/dL vs. STZ/VEH [501.69±11.11 mg/dL], P<0.0001) through the end of observations (414.2 mg/dL vs. STZ/VEH [562.76±10.2 mg/dL], P<0.0001). Four weeks post-STZ, CTRL rats exhibited higher body weights than STZ/VEH rats (347.5 g vs. 295.42 g, P<0.0001) and they exhibited a time-dependent weight gain throughout the experiment (8 weeks post-STZ, 471.62±7 g), whereas STZ/VEH remained stable (8 weeks post-STZ, 293.11±4.7 g) (Fig. 2B). In contrast, STZ/LDX rats exhibited a slight, but significant increase in body weight after 3- and 4-week of LDX treatment (8 weeks post-STZ, 325.38±11 g, P=0.04).
Early ladarixin (LDX) mitigated streptozotocin (STZ)-induced type 1 diabetes mellitus and associated inflammatory/metabolic parameters in rats. Effects of LDX treatment (daily from week 4 to 8) in control (CTRL), STZ/vehicle (VEH) and STZ/LDX groups on (A) glucose level (mg/dL), (B) body weight (g), (C) insulin (pg/mL), (D) glycosylated hemoglobin (HbA1c; ng/mL), (E) triglycerides (mg/dL), (F) cholesterol (mg/dL), (G) high-density lipoprotein cholesterol (HDL-C; mg/dL), (H) low-density lipoprotein cholesterol (LDL-C; mg/dL), (I) acetoacetate (AcAc; µM), and (J) β-hydroxybutyrate (BHB; nM) measured at different time points post-STZ injection. Data are represented as mean±standard error of the mean (n=8 to 12). The threshold of 0.05 was used for determining statistical significance. aP<0.05, bP<0.01, cP<0.001, and dP<0.0001 and eP<0.05, fP<0.01, gP<0.001, and hP<0.0001 indicate significant differences vs. CTRL and STZ/VEH, respectively.
Consistent with blood glucose, significant reductions in plasma insulin levels were observed 3 weeks post-STZ and persisted through 5 and 7 weeks post-STZ in STZ/VEH rats relative to CTRL rats (Fig. 2C). LDX treatment normalized insulin levels after 1 and 3 weeks of treatment (5 and 7 weeks post-STZ) (7 weeks post-STZ, STZ/VEH [59.72±14.94 pg/mL] vs. STZ/LDX [135±13 pg/mL], P=0.025) (Fig. 2C). Increases in both HbA1c and triglycerides emerged at 5 weeks post-STZ and persisted through 7 weeks post-STZ in STZ/VEH rats, and this was mitigated by LDX treatment (Fig. 2D and E) (7 weeks post-STZ, HbA1c: STZ/VEH [34±6 ng/mL] vs. STZ/LDX [18±3 ng/mL], P=0.008; triglycerides: STZ/VEH [131±10 mg/mL] vs. STZ/LDX [71±9 mg/mL], P=0.0006). As expected, STZ increased plasma total cholesterol and influenced HDL-C/LDL-C levels as early as 5 weeks post-STZ (Fig. 2F-H), and this was also mitigated by early LDX. Specifically, 7 weeks post-STZ, LDX mitigated increases in total cholesterol and LDL-C and a decrease in HDL-C. LDX restored cholesterol metrics to levels comparable to CTRL rats (total cholesterol: STZ/VEH [110±8 mg/mL] vs. STZ/LDX [87±8 mg/mL], P=0.049; LDL-C: STZ/VEH [60±2 mg/mL] vs. STZ/LDX [37±4 mg/mL], P=0.0001; HDL-C: STZ/VEH [13±2 mg/mL] vs. STZ/LDX [33±4 mg/mL], P=0.0001).
Finally, plasma ketone body concentration was significantly higher in STZ-treated rats (Fig. 2I and J). Increases emerged at 3 weeks post-STZ for AcAc and 5 weeks post-STZ for BHB. LDX treatment mitigated these effects (AcAc, 3 weeks post-STZ, STZ/VEH [557.33 ±22.9 µM] vs. STZ/LDX [287.7 ±22 µM], P=0.0018; BHB, 5 weeks post-STZ, STZ/VEH [709.23±80.1 nM] vs. STZ/LDX [246.7±15.7 nM], P=0.0006).
Early LDX protected against STZ-induced mechanical allodynia and thermal hyperalgesia and markers of inflammation in the DRG, SN, vitreous, and plasma
No mechanical allodynia was observed over time in CTRL rats (95±5 g, 5 weeks) (Fig. 3A). In contrast, mechanical allodynia was observed as early as 2 weeks post-STZ in STZ-treated rats, with a peak 5 weeks post-STZ (18.11±3.86 g, P<0.0001 vs. CTRL) (Fig. 3A). This pattern is notably similar to that of blood glucose levels over time following STZ (Fig. 2A). LDX treatment reversed STZ-induced mechanical allodynia at all timepoints (i.e., 1–4 weeks of LDX, 5–8 weeks post-STZ) (5 weeks post-STZ, 58.58±5.54 g, P=0.0003 vs. STZ/VEH). The peak anti-allodynic effect of LDX was observed at 2 weeks post-LDX, when the mean paw threshold in STZ/LDX rats (68.55±7.03 g) was comparable to that observed in CTRL rats (95±5 g).
Early ladarixin (LDX) reduced streptozotocin (STZ)-induced signs of painful diabetic peripheral neuropathy and CXC motif chemokine ligand 8 (CXCL8) and citrullinated histone H3 (CitH3) levels in STZ-treated rats in plasma and vitreous. Effects of LDX (15 mg/kg, intragastric [i.g.]) treatment (from week 4 to 8) in control (CTRL), STZ/vehicle (VEH), and STZ/LDX groups on (A) Paw withdrawal response (g) and on (B) thermal withdrawal latency (tail-flick latency; second). (C) Representative Western blot images of tumor necrosis factor-α (TNFα) 17 KDa and β-actin in dorsal root ganglion (DRG) and sciatic nerve (SN). (D, E) Bar graphs showing relative quantification of TNFα protein levels in DRG and SN in CTRL, STZ/VEH, and STZ/LDX rats. (F) Representative Western blot images of interleukin-1β (IL-1β) 18 kDa and β-actin in DRG and SN. (G, H) Bar graphs showing relative quantification of IL-1β protein levels in DRG and SN in CTRL, STZ/VEH, and STZ/LDX rats measured at 8 weeks post-STZ injection. Data are represented as mean±standard error of the mean (n=8 to 12). Effects of LDX (15 mg/kg, i.g.) treatment (from week 4 to 8) in CTRL, STZ/VEH, and STZ/LDX groups on (I, J) CXCL8 at 3, 5, and 7 weeks (ng/mL), (K, L) CitH3 at 3, 5, and 7 weeks (pg/mL) measured at different time points post-STZ injection. Data are represented as mean±standard error of the mean (n=4). The threshold of 0.05 was used for determining statistical significance. GRO/CINC1, growth-regulated gene product/cytokine-induced neutrophil chemoattractant 1. aP<0.05, bP<0.01, cP<0.001, and dP<0.0001 and eP<0.05, fP<0.01, gP<0.001, and hP<0.0001 indicate significant differences vs. CTRL and STZ/VEH, respectively.
A similar pattern was observed for STZ-induced thermal hyperalgesia. Though CTRL rats did not exhibit thermal hyperalgesia (7.86±0.33 seconds, 5 weeks), starting 2 weeks post-STZ, the thermal pain threshold (PT) was significantly reduced in STZ/VEH rats relative to CTRL rats (5 weeks post-STZ, 4.7±0.43 seconds, P=0.0020) and this persisted through the end of treatment. STZ-induced thermal hyperalgesia was mitigated by just one week of LDX treatment (5 weeks post-STZ, 9.03±0.39 seconds vs. STZ/VEH, P<0.0001) and this protective effect persisted through weeks 2 and 3 of LDX treatment.
Inflammation in the DRG and SN may be indicative of DPN. As such, IL-1β and TNFα protein expression in the L4–L6 (lumbar) DRG and in the SN was assessed via Western blot (Fig. 3). In both the DRG and SN, IL-1β and TNFα were significantly higher in STZ/VEH rats relative to CTRL rats and this effect was mitigated by LDX treatment (IL-1β, P=0.028 and P=0.014; TNFα, P=0.046 and P=0.0008, in DRG and SN respectively) (Fig. 3C-H).
As previously mentioned, elevated plasma CXCL8 or CXCL1 have been observed in T1DM patients and preclinical T1DM models, respectively [4,6,22]. Elevated circulating CXCL8, or other markers of inflammation, could be indicative of immune dysfunction contributing to T1DM progression or of effects on non-immune cells contributing to related complications. Vitreous inflammation in particular could be associated with DR. To this end, we assessed CXCL8 and CitH3, a marker neutrophil extracellular trap (NET) formation, in the plasma and vitreous. We observed increased levels of CXCL8 and CitH3 in both plasma and vitreous in STZ-treated rats at 5 and 7 weeks post-STZ relative to CTRL rats (Fig. 3I-L) and these effects were significantly mitigated after 3 weeks of LDX in plasma (7 weeks post-STZ, STZ/VEH vs. STZ/LDX: CXCL8, P=0.049; CitH3, P=0.042) (Fig. 3I and K) and after 1 or 3 weeks of LDX in vitreous (1 week post-LDX, 5 weeks post-STZ, STZ/VEH vs. STZ/LDX: CXCL8, P=0.033; CitH3, P=0.012) (Fig. 3J and L).
Late LDX treatment maintained a protective effect on STZ-induced signs of DPN and DR even after irreversible β-cell death and dysfunction
Late LDX treatment had no effect on STZ-induced alterations in blood glucose, metabolic or inflammatory signs of T1DM or β-cell number and functions
The effect of late LDX treatment, with daily LDX 8–12 weeks post-STZ, was assessed using additional diabetic/metabolic parameters, including assessments of pancreatic β-cell numbers and functions. Again, there was a significant increase in plasma glucose as early as 2 weeks post-STZ which persisted for the duration of experiments (Fig. 4A). In contrast to early LDX, late LDX had no effect on hyperglycemia in STZ-treated rats after 1 or 3 weeks of LDX treatment (9 or 12 weeks post-STZ) (blood glucose, 9 weeks post-STZ, STZ/VEH [574.6±22.06 mg/dL] vs. STZ/LDX [550.4±23.04 mg/dL], P=0.73) (Fig. 4A). Notably, LDX did protect against a reduction in body weight in STZ-treated rats (Fig. 4B). Significant reductions in insulin and C-peptide were observed in STZ-treated rats at 8, 9, and 12 weeks post-STZ relative to CTRL rats. However, late LDX was insufficient to correct either parameter (9 weeks post-STZ, insulin: STZ/VEH [107.88±14.23 pg/mL] vs. STZ/LDX [165.56±12.39 pg/mL], P=0.08; C-peptide: STZ/VEH [300.44±38.78 pg/mL] vs. STZ/LDX [434.18±27.95 pg/mL], P=0.095) (Fig. 4C and D). Calculations of HOMA-IR and HOMA-β confirmed the development of STZ-induced insulin resistance and insulin secretion dysfunction, respectively, in STZ/VEH rats versus CTRL rats (Fig. 4E and F) and LDX treatment did not significantly improve these parameters.
Late ladarixin (LDX) treatment was insufficient to mitigate streptozotocin (STZ)-induced type 1 diabetes mellitus and associated inflammatory/metabolic parameters but prevented STZ-induced signs of painful diabetic peripheral neuropathy. Effects of LDX treatment (from week 8 to 12) in control (CTRL), STZ/vehicle (VEH), and STZ/LDX groups on (A) glucose level (mg/dL), (B) body weight (g), (C) plasmatic insulin (pg/mL), (D) plasmatic C-peptide (pg/mL), (E) homeostasis model assessment for insulin resistance (HOMA-IR) and (F) homeostasis model assessment for β-cell function (HOMA-β). (G) Immunofluorescence staining at 9 weeks post-STZ of insulin (red) or glucagon (green) in pancreatic tissues (×20). Mean of insulin (H) or glucagon (I) positive cells per islet in CTRL, STZ/VEH, and STZ/LDX groups. (J) Paw withdrawal threshold (g) and (K) thermal withdrawal latency (tailflick latency; second). Data are represented as mean±standard error of the mean (n=8 for behavior and n=4 for immunofluorescence). The threshold of 0.05 was used for determining statistical significance. DAPI, 4´,6-diamidino-2-phenylindole. aP<0.05, bP<0.01, cP<0.001, and dP<0.0001 and eP<0.05 and fP<0.01, indicate significant differences vs. CTRL and STZ/VEH, respectively.
As expected, immunofluorescence analyses revealed a decrease in the number of insulin-positive cells in pancreatic tissue from STZ-treated rats relative to CTRL rats (12 weeks post-STZ: STZ/VEH [40.17±11.42] vs. CTRL [278.5±63.44], P=0.001), suggesting a reduction in the number of β-cells (Fig. 4G and H). Late LDX treatment had no effect on the number of pancreatic insulin-positive cells in STZ/LDX rats (12 weeks post-STZ: STZ/LDX [82.0±8.56] vs. STZ/VEH [40.17±11.42], P=0.71) (Fig. 4G and H). There was a mild effect of STZ on the number of glucagon-positive cells, indicative of number of α-cells, with a trend towards reduction in STZ/VEH rats relative to CTRL rats and no effect of late LDX (12 weeks post-STZ: STZ/VEH [64.83±13.93] vs. STZ/LDX [75.50±14.58], P=0.888) (Fig. 4G and I).
Late LDX treatment protected against STZ-induced mechanical allodynia and thermal hyperalgesia and markers of inflammation in the vitreous, retina, and plasma even in the absence of an effect on the STZ-induced T1DM phenotype
Strikingly, despite the lack of effect of late LDX treatment on STZ-induced dysglycemia and β-cell function, late LDX improved STZ-induced signs of DPN and DR. Significant reductions of mechanical and thermal thresholds relative to CTRL rats emerged at 2 weeks post-STZ (Fig. 4J and K) and persisted for the duration of the experiment. Late LDX treatment showed a trend towards improvement in the mechanical and thermal PT with 1 week of LDX treatment and a significant improvement in these metrics after 4 weeks of LDX treatment (PT: STZ/VEH [18±10.56 g] vs. STZ/LDX [66.8±8.38 g], P=0.018; TFL: STZ/VEH [4.44±0.65 seconds] vs. STZ/LDX [7.34±0.88 seconds], P=0.021) (Fig. 4J and K).
We found higher levels of CXCL8 and CitH3 at 9 and 12 weeks post-STZ in STZ/VEH rats relative to CTRL rats in plasma (Fig. 5). Intriguingly, LDX treatment restored basal levels of both parameters at both 1 and 3 weeks of LDX treatment (9 and 12 weeks post-STZ) (9 weeks post-STZ, CXCL8: STZ/VEH vs. STZ/LDX, P=0.0009; CitH3: STZ/VEH vs. STZ/LDX, P=0.047).
Late ladarixin (LDX) mitigated CXC motif chemokine ligand 8 (CXCL8) and citrullinated histone H3 (CitH3) in streptozotocin (STZ)-treated rats in plasma and vitreous. Effects of LDX (15 mg/kg, intragastric) treatment (from week 8 to 12) in control (CTRL), STZ/vehicle (VEH), and STZ/LDX groups on (A) plasma CXCL8 (pg/mL), (B) vitreous CXCL8 (pg/mL), (C) plasma CitH3 (ng/mL), and (D) vitreous CitH3 (ng/mL) measured at 9- and 12-week post-STZ injection. Data are represented as mean±standard error of the mean (n=4). GRO/CINC1, growth-regulated gene product/cytokine-induced neutrophil chemoattractant 1. aP<0.05 and bP<0.01, and cP<0.05 and dP<0.01 indicate significant differences vs. CTRL and STZ/VEH, respectively.
Inflammatory parameters in the vitreous and in retinal tissue were assessed as signs of DR. STZ-induced increases in vitreous CXCL8 and Cit3H were observed at 9 weeks post-STZ in STZ/VEH (Fig. 5). CTRL rats and this was prevented by late LDX (CXCL8: STZ/VEH vs. STZ/LDX, P=0.008; CitH3: STZ/VEH vs. STZ/LDX, P=0.0027) (Fig. 5). The effect persisted through 12 weeks post-STZ only for CitH3, not CXCL8, and the protective effect of LDX was maintained for CitH3.
Similarly, we observed at 9 weeks post-STZ increased immunoreactivity of CXCL8 (Fig. 6C and J) and CXCR1/2 (Fig. 6D, E, K, and L) in the retina of STZ-treated rats relative to that of CTRL rats at the end of the experiment. Moreover, we found increased expression of CitH3 (Fig. 6A and H) and MPO (Fig. 6B and I), both biomarkers of NET formation, in STZ/VEH versus CTRL rats. LDX treatment mitigated STZ-induced increases in retinal CXCL8, CXCR1, CitH3, and MPO (STZ/LDX vs. STZ/VEH, all P<0.0001), but not CXCR2 (STZ/LDX vs. STZ/VEH, P=0.9955).
Late ladarixin (LDX) normalized signs of streptozotocin (STZ)-induced diabetic retinopathy and diabetic nephropathy: retinal expression of CXC motif chemokine ligand 8 (CXCL8), CXC motif chemokine receptor 1 (CXCR1), citrullinated histone H3 (CitH3) myeloperoxidase (MPO), CD34, and vascular endothelial growth factor (VEGF) and of plasma creatinine levels in STZ-treated rats. Effects of LDX (15 mg/kg, intragastric [i.g.]) treatment (from week 8 to 12) in control (CTRL), STZ/vehicle (VEH), and STZ/LDX groups on the representative images of (A, H) CitH3, (B, I) MPO, (C, J) CXCL8, (D, K) CXCR1, and (E, L) CXCR2 expression in the retina layers and relative quantification measured as fluorescence intensity per area at different time points post-STZ injection. Effects of LDX (15 mg/kg, i.g.) treatment (from week 8 to 12) in CTRL, STZ/VEH, and STZ/LDX groups on the representative images of (F, M) CD34 and (G, N) VEGF expression in the retina layers and relative quantification measured as fluorescence intensity per area at different time points post-STZ injection. Data are represented as mean±standard error of the mean (n=3). (O) Effects of LDX (15 mg/kg, i.g.) treatment (from week 8 to 12) in CTRL, STZ/VEH, and STZ/LDX groups on creatinine (pg/mL) measured at 9- and 12-week post-STZ injection. Data are represented as mean±standard error of the mean (n=4). The threshold of 0.05 was used for determining statistical significance. The magnification of immunofluorescence picture is of ×20. GRO/CINC1, growth-regulated gene product/cytokine-induced neutrophil chemoattractant 1. aP<0.0001, bP<0.001, and cP<0.0001 indicate significant differences vs. CTRL and STZ/VEH.
In DR, VEGF plays a major role in the processes of cell migration, proliferation, proteolysis, survival, and vascular permeability [23]. Therefore, the expression of VEGF along with CD34, a marker of cell proliferation, in the retina was also investigated. VEGF expression was located in retinal ganglion cell layer and the outer plexiform layer (Fig. 6F and M). CD34 expression was located in the vascular endothelial cells (Fig. 6G and N). Nine weeks post-STZ, VEGF and CD34 were significantly increased in the STZ-treated rat retinas relative to those of CTRL rats. The enhancement of these markers was prevented by 1 week of LDX treatment (STZ/VEH vs. STZ/LDX: VEGF, P<0.0001; CD34, P<0.0001).
Lastly, given the striking protective effect of late LDX treatment on signs of STZ-induced DPN and DR and the prevalence of diabetic nephropathy (DN) also associated with T1DM, we evaluated plasmatic creatinine levels to assess the kidney function (Fig. 6O). As expected, STZ-treated rats showed high levels of creatinine in plasma at 9 and 12 weeks post-STZ relative to CTRL rats. Late LDX treatment significantly reduced the hypercreatinemia at both time points (9 weeks post-STZ: STZ/VEH [68.46±2.12 pg/mL] vs. STZ/LDX [44.94±3.2 pg/mL], P<0.0001).
DISCUSSION
The CXCL8-CXCR1/2 axis has been proposed as a master mediator in the leukocyte recruitment involved in insulitis and pancreatic damage preceding T1DM [3,11]. Our report supports this hypothesis in that LDX 4–8 weeks post-STZ protected against the development of STZ-induced T1DM in rats. Early LDX treatment was also sufficient to improve STZ-induced signs of DPN and DR. Importantly, though late LDX 8–12 weeks post-STZ was insufficient to reverse dysglycemia and β-cell damage, it was still sufficient to directly improve signs of STZ-induced DPN and DR. A T1DM therapeutic approach which can act directly on pathological mechanisms underlying the most common T1DM complications in addition to its actions to prevent or reverse β cell loss would significantly improve patient outcomes. The design of this study has revealed a distinct beneficial effect of LDX on STZ-induced T1DM complications in rats, independent of its protective effect on T1DM.
In line with the results reported in the NOD mouse [11], early LDX treatment resulted in a profound improvement in the classic symptoms and biochemical parameters of T1DM. Daily administration of LDX 4–8 weeks post-STZ mitigated persistent hyperglycemia, prevented the reduction in insulin levels, and restored dysregulation of HbA1c, total cholesterol, LDL-C, HDL-C, and triglycerides. T1DM patients also frequently experience ketosis (hyperketonemia), which is known to accelerate microangiopathy and to underlie vascular disease and precipitate neuropathy in T1DM patients [24]. This study is also the first to show that circulating ketone levels significantly increased in the early phases STZ-induced T1DM in rat and, subsequently, that LDX treatment restored normal ketone levels.
Eight weeks after STZ administration, a dramatic reduction in β-cell mass and function was confirmed in the late-treatment experiment. LDX initiation in late disease progression strongly reduces the possibility that LDX-mediated protection of the residual β-cells may reverse the established dysglycemia. Indeed, late LDX was not effective in reverting STZ-induced impairments in β-cell mass, glucose tolerance, or insulin secretion and did not restore glucose, insulin, or C-peptide levels.
Plasma CXCL8 levels were elevated throughout course of disease development and this was paralleled by a sustained increase of CitH3 and MPO [25], indicative of NET formation. Both early and late LDX treatment mitigated these effects, probably due to the decoy function of CXC receptors that mediate the dampening of cytokine signaling by internalizing and translocating functional CXCL8 [26]. This novel finding is particularly relevant because previous studies linked augmented NETosis not only with β-cell autoimmunity but also with long-term diabetic complications [27]. It may be that alterations in circulating CXCL8 and/or inflammation reflect enhanced actions both on immune cells and on non-immune cells involved in T1DM complications.
DPN and DR, in particular, are two of the most common and devastating complications of T1DM with unmet therapeutic need [28]. It has been suggested that the severity of neuropathy may be influenced by the degree of metabolic dysfunction and hyperglycemia, with several clinical studies suggesting stabilization of blood glucose as the first step in the management of painful DPN [29]. However, there is also evidence that proinflammatory cytokines/chemokines involved in the recruitment and activation of immune cells are key mediators of neuroimmune interactions in pain control [30]. Specifically, the development of chronic pain in T1DM has been attributed to neuronal dysfunction sustained by an aberrant hyperexcitability of nociceptive spinal dorsal horn neurons and a loss of inhibitory interneurons [31]. Activation of CXCR1 and CXCR2 on the surface of peripheral nociceptors by CXCL1 may contribute to this mechanism directly by increasing excitability and sensitivity to noxious stimuli [32]. Spinal CXCL1 produced by microglia following nerve injury or peripheral inflammation has been shown to enhance inflammatory pain through CXCR2 activation [33]. Moreover, there is also evidence for local indirect effects of CXCL1 on hyperalgesia mediated by neutrophil/leukocyte migration following SN injury [34].
Here, we are the first to report that LDX conveys protection against a preclinical model of DPN and that the anti-nociceptive effect of LDX is independent of its ability to preserve β-cell function. We showed that, consistent with another study [35], STZ-induced T1DM in rats is characterized by a concomitant DRG and SN overexpression of IL-1β and TNFα, which is prevented by LDX. Both early and late LDX also mitigated STZ-induced mechanical allodynia and thermal hyperalgesia. These results support the hypothesis that the protective effect of LDX on DPN occurred, at least in part, independent of its effects on β-cell function. A direct analgesic effect of LDX is consistent with other studies showing that CXCL1 neutralizing antibody or CXCR2 inhibition can counteract DRG neuron hyperexcitability and reduce pain [36].
DR is another common complication exacerbating the T1DM patient burden. When T1DM advances to ocular complications, chronic hyperglycemia can lead to leukocytosis and vascular complications. Retinal inflammatory reactions and pro-angiogenic factors have been linked to neuronal and glial damage involved in vascular changes potentially contributing to DR CXCL8 and VEGF contribute to blood-retinal barrier breakdown, vascular damage, and neuroinflammation. Thus, while there is evidence that CXCR1/2 blockade may influence DR, whether this is through direct effects or second to its effect on glycemic control has not been elucidated [15].
Here, we observed LDX-dependent normalization of STZ-induced signs of DR in the vitreous and in retinal tissue following both early and late LDX, namely, normalization of vitreous CXCL8 and CitH3 and of retinal CXCL8, CitH3, and MPO. LDX also prevented STZ-induced increases in VEGF and CD34 expression, typical markers of DR. Interestingly, the protective effect of LDX on CXCL8 was accompanied by a normalization of CXCR1 expression, but not CXCR2 expression in the retina. While the LDX allosteric mechanism on CXCR1 and CXCR2 is comparable [9], the two receptors show distinct expression, function, affinity for CXC chemokines, and internalization [37]. It may be that other proinflammatory cytokines involved in the retinal immune response might have a tonic effect contributing the differential effect of LDX on CXCR1 and CXCR2. Though the cellular cytotype of CXCR1 and CXCR2 was not examined, it is well known that CXCR1 is expressed primarily on retinal microglia, astrocytes, and Müller glia, which are activated to secrete inflammatory mediators, aggravating cell apoptosis and subsequent vascular leakage in DR. Novel contributions of infiltrating peripheral immune cells recruited by the increased CXCL1 levels might complicate the picture in DR [38].
Finally, we assessed a metric indicative of DN. Renal leukocyte accumulation and rapid increases in CXCL8 urinary levels have been reported in patients with DN [39]. Previous studies have also suggested the therapeutic potential of LDX in preventing DN by mitigating podocyte damage associated with CXCL8-CXCR1/2 signaling [40]. Here, plasmatic creatinine levels were elevated in STZ-treated rats at 9 and 12 weeks post-STZ and this, too, was mitigated by late LDX.
In conclusion, LDX is currently under clinical investigation (NCT04628481) in patients with new-onset T1DM and low residual β-cells function. However, little is known about CXCR1/2 pathway in the control of diabetes comorbidities. This study strongly reinforces the concept that the CXCL8-CXCR1/2 axis is a key mediator in T1DM and although further analysis is needed, we provides for the first time initial evidence for a direct role of CXCR1/2 also in diabetic complications, namely DPN and DR. These results highlight an important potential clinical advantage of LDX for the treatment of T1DM in that, in addition to its protective effect on T1DM pathology, it may also provide protection against advanced diabetes complications which are particularly dangerous for long-term T1DM patients.
Notes
CONFLICTS OF INTEREST
Serena Boccella, Cristina Giorgio, Andrea Aramini, Marcello Allegretti, Meghan Jones, and Laura Brandolini are Dompé Farmaceutici S.p.A employees and Andrea Maria Morace, Francesca Guida, Michela Perrone, Iolanda Manzo, Carmela Belardo, Sabatino Maione, and Livio Luongo are University of Vanvitelli in Naples employees. The authors declare no competing interests in relation to the contents of this work. The authors declare that this study received funding from Dompé Farmaceutici SpA.
AUTHOR CONTRIBUTIONS
Conception or design: S.B., C.G., F.G., S.M., A.A., M.A., L.L., L.B.
Acquisition, analysis, or interpretation of data: S.B., A.M.M., C.G., M.P., I.M., C.B.
Drafting the work or revising: S.B., C.G., F.G., M.J., L.L., L.B.
Final approval of the manuscript: S.M., A.A., M.A., L.L., L.B.
FUNDING
Pharmacological studies on inflammatory models of Netosis—Project co-financed entitled: Ladarixin as new Juvenile Diabetes Inhibitory Agent—LIDIA under the Sustainable Growth Fund (Proposal No. 1410 MISE call 02/08/2019 and subsequent DD 02/10/2019).
ACKNOWLEDGMENTS
None