CXCR1/2 Inhibition Blocks and Reverses Type 1 Diabetes in Mice

Type 1 diabetes (T1D) originates from an immunologic disorder that leads the immune system to attack pancreatic β-cells (1). At the onset of T1D, the inflammatory response is thought to be directed toward pancreatic islets (2), a process that seems to play a crucial role in maintaining the autoimmune response as well as representing a key feature of its pathology (3). Consequently, control of the inflammatory response is a strategic and somewhat underexplored action for influencing the disease (4–7).

Chemokines are a family of small chemoattractant cytokines that control through their receptors a wide variety of physiological and pathological processes, ranging from immune surveillance to inflammation, from viral infection to cancer (8–10). A logical model for initiating/maintaining leukocyte infiltration into mouse and human pancreatic islets is the secretion, by β-cells themselves, of chemokines favoring leukocyte recruitment (11–14). For instance, β-cell–derived CCL2 recruits monocytes and macrophages into pancreatic islets (15–18) and CXCL10 secreted by insulin-producing cells promotes T-cell infiltration (19,20).

We recently demonstrated that the inhibition of CXCR1/2 chemokine receptors is crucial for improving both human and murine islet survival after transplantation (21). In this study we hypothesized that CXCR1/2 inhibition may also be functional/effective in preventing inflammatory damage to pancreatic islets during diabetes development. In fact, pancreatic islets produce and secrete the CXCR1/2 chemokine ligands (named CXCL8, CXCL1, and CXCL2) in response to proinflammatory cytokines (12,15,17,22,23). Furthermore, the concentration of CXCR1/2 ligands is elevated in the blood of both rodents and humans with autoimmune diabetes (24–26); most important, recent reports support the notion that neutrophils (the major target of CXCR1/2 inhibitors) play a key role in the etiopathogenesis of T1D (27–29). We therefore extensively characterized the consequences of CXCR1/2 inhibition on inflammation- and autoimmunity-mediated diabetes in preclinical models. Reparixin (30) and ladarixin (31), CXCR1/2 noncompetitive allosteric inhibitors that have completed phase I studies and already entered phase II/III trials (32), were used to obtain the pharmacologic blockade of the CXCL1–CXCR1/2 axis. The demonstration that this strategy may be successful in preserving residual β-cells has the potential to make a significant change in the approach to managing human T1D.

Male C57BL/6 and female NOD/Ltj mice were purchased from Charles River Laboratories (Calco, Italy). All mice were bred and housed in specific pathogen-free conditions. The animals had free access to tap water and pelleted food throughout the course of the study. All experiments were in accordance with protocols approved by the Animal Care and Use Committee of San Raffaele Scientific Institute.

Reparixin (30) and DF1726A (structurally related to reparixin but not active on CXCR1/2 receptors at 10⁻⁵–10⁻⁸ mol/L) were provided by Dompè Farmaceutici S.p.A (L’Aquila, Italy) and administered by continuous subcutaneous infusion (Alzet osmotic pump; Alza Corporation, Palo Alto, CA) at a dose of 5.4 mg/h/kg for 7 days starting at day −1; ladarixin (31) was provided by Dompè Farmaceutici S.p.A and administered orally at a dose of 15 mg/kg daily for 14 days starting at day −1 or +5; 0.9% sodium chloride solution was used as vehicle.

Male C57BL/6 mice received treatment with multiple low doses (MLDs) of streptozotocin (STZ). STZ (Sigma-Aldrich, St. Louis, MO) was injected intraperitoneally at a dose of 40 mg/kg/day. This was carried out for 5 consecutive days. Nonfasting glucose concentrations in venous blood were measured every day starting from the first day of treatment. Blood glucose measurements were performed using an Ascensia Elite XL blood glucose meter (Bayer, Toronto, Ontario, Canada). Mice with nonfasting glycemia ≥250 mg/dL on two consecutive tests were considered diabetic; the first detection of hyperglycemia was considered the date of diabetes onset. Mice were followed for up to 60 days after the first STZ injection.

For studies of T1D prevention, female NOD/Ltj mice received ladarixin or vehicle by oral administration at 4, 8, or 12 weeks of age. Mice were monitored for blood glucose values twice a week. Diabetes was defined as two consecutive nonfasting blood glucose concentrations ≥250 mg/dL separated by 24 h. For studies of T1D diabetes reversion, female NOD mice with recent-onset diabetes received ladarixin or vehicle by oral administration. Animals were monitored two to three times per week for up to 10 weeks, with no exogenous insulin treatment. Remission of the disease in treated mice was defined as two consecutive glucose measurements <200 mg/dL. Relapse of the disease in cured mice was considered following two consecutive glucose measurements of ≥250 mg/dL.

Female NOD mice were killed at the end of the treatment. Collected pancreata were fixed in 10% buffered formalin and processed routinely for histology. Histology slides were stained with hematoxylin and eosin, anti-CD3 (rat antihuman CD3 IgG1 with mouse cross-reactivity, clone CD3–12, 1:1000; Serotec, U.K.), anti-F4/80 (rat anti-mouse F4/80 IgG2b, clone A3–1, 1:200; Serotec), and anti-myeloperoxidase (MPO; polyclonal rabbit antihuman; Dako, Denmark). Antigen was retrieved with Tris EDTA (pH 9) for anti-CD3 and MPO immunostaining and with pronase Bond Enzime diluent kit (Leica, Italy) for anti-F4/80 immunostaining. The degree of islet infiltration was analyzed by two independent observers who were blinded to the experimental conditions. Each observer assessed a minimum of 40 islets per animal. Insulitis was scored as previously described (33): 0, no insulitis; 1, peri-insulitis; 2, mild insulitis (<25% of infiltrated islets); 3, severe insulitis (25–75% of infiltrated islets). In terms of infiltration, for each parameter we also evaluated the morphology of the islet: 0, physiological islet; 1, amorphous islet. The total score is the addition of the infiltration and the morphological scores. The leukocyte infiltration score (CD3⁺ and F4/80⁺ cells) was quantified as follows: 0, no infiltration; 1, low peri-islet infiltration; 2, mild peri-islet infiltration; 3, high intraislet infiltration (Supplementary Fig. 1). Neutrophils infiltration was evaluated as the number of MPO⁺ cells per islet.

Single-cell suspensions were prepared from different tissues (i.e., pancreas, pancreatic lymph nodes, spleen, bone marrow, and blood) and were stained at 4°C in PBS containing 2% FBS and 0.5% EDTA. Cell surfaces were stained with fluorescein isothiocyanate–, phycoerythrin-, or allophycocyanin-labeled anti-CD4, anti-CD8 (clone 53–6.7), anti-CD3 (clone 145–2C11), anti-CD19 (clone 1D3), anti-Gr-1 (clone RB6–8C5), anti-CD11b (clone M1/70), anti-Ly6G (clone 1A8), anti-Ly6C (clone AL-21; BD Biosciences, San Diego, CA), and anti-CXCR2 antibodies (clone 242216; R&D Systems, Minneapolis, MN). Both Gr-1⁺Ly6C⁻CD11b⁺ and Ly6G⁺CD11b⁺ combinations were used to identify polymorphonuclear cells (PMNs). Linear regression analysis showed no statistically significant difference between the two strategies used for neutrophil identification (95% CI of the slope 0.9874–1.121) (data not shown). Consequently, on the basis of the expression of the mentioned markers, Gr-1⁺CD11b⁺Ly6C⁻ (mostly PMNs), CD3⁺CD4⁺ (mostly T-helper cells), CD3⁺CD8⁺ (mostly cytotoxic T lymphocytes), CD3⁺CD4⁻CD8⁻ (mostly NKT cells), CD19⁺ (mostly B lymphocytes), Gr-1⁻CD11b⁺Ly6C⁻ (mostly natural killer cells), and Gr-1⁻CD11b⁺Ly6C⁺ (mostly monocytes/macrophages) cells were identified. To exclude a possible contamination of T cells within the gate of MΦ, Gr-1 and Ly6C markers were evaluated on CD3⁺ and CD19⁺ cells. The frequencies of CD3⁺ and CD19⁺ cells that decrease within the macrophage gate were irrelevant (0.08% and 0.26%, respectively; data not shown). Samples were acquired on a FACSCantoII instrument (BD Biosciences). Rainbow calibration particles (Spherotech Inc., Lake Forest, IL) were used to calibrate and normalize acquisition settings in each experiment. Flow cytometry data were analyzed with FCS Express V4 (DeNovo Software, Glendale, CA).

Collagenase IV (Sigma-Aldrich, St. Louis, MO) solution (2 mg/mL; 2.5 mL) was injected into the bile duct. Pancreas samples were collected and digested at 37°C for 20 min. Obtained digested tissue was washed with cold RPMI 1640 medium and FBS (Lonza, Belgium) and resuspended in a mixture of Percoll (GE Healthcare/Biosciences AB, Sweden) and Histopaque 1077 (Sigma-Aldrich). After gradient centrifugation (2,000 rpm for 20 min), intrapancreatic leukocytes were collected from the interface. The collected cells were washed twice with RPMI 1640 medium and used for further analysis.

Serum samples were analyzed for 23 cytokines using the Bio-Plex Pro Mouse Cytokine Standard 23-plex (group 1; Bio-Rad Laboratories, Hercules, CA) according to the manufacturer’s protocol.

Variables are summarized as mean ± SD or median and interquartile range according to their distribution. Variables with a normal distribution were compared with one-way unpaired Student t test (two groups). Variables with a non-normal distribution were compared using the Mann-Whitney U test. Categorical variables were compared using the χ² test or Fisher exact test, as appropriate. The cumulative diabetes incidence or reversion was evaluated using Kaplan-Meier analysis, and the significance was estimated using the log-rank test. Statistical analysis of nonfasting glycemia during follow-up was performed using the general linear model for repeated measures. Hazard ratios for new-onset diabetes or reversed diabetes after treatment were estimated using Cox regression. All statistical analyses were performed using SPSS statistical software version 13.0 (SPSS Inc., Chicago, IL).

To determine the efficacy of CXCR1/2 inhibition as a strategy to prevent inflammation-mediated islet destruction, we tested the ability of two different CXCR1/2 inhibitors to prevent diabetes induction after MLD STZ injections. In the first set of experiments we tested reparixin, a noncompetitive allosteric CXCR1/2 inhibitor (30) that must be administered via continuous subcutaneous infusion because of its short half-life. A total of 34 male C57BL/6 mice were intraperitoneally injected with MLD STZ (40 mg/kg/day for 5 days), and reparixin (8 mg/h/kg; n = 17) or vehicle (n = 17) was administered starting from day −1 up to day +6 after the first STZ injection. Reparixin treatment significantly prolonged the timing of diabetes development. The median diabetes-free time was 12 ± 0.6 and 7 ± 0.6 days for reparixin- and vehicle-treated mice, respectively (P = 0.001; Fig. 1A). More important, glycemia remained constantly and significantly lower in the reparixin-treated group than in the vehicle-treated group after diabetes developed (P = 0.039; Fig. 1B). Notably, DF1726A, a compound structurally related to reparixin but not active on CXCR1/2 (34), did not prevent diabetes development when administered in the same experimental conditions (Supplementary Fig. 2A). Moreover, reparixin treatment at 15 mg/kg three times a day per os (a dose able to inhibit selectively CXCR1 but not CXCR2) did not prevent diabetes (Supplementary Fig. 2B).

In the second set of experiments we tested ladarixin, a potent blocker of both CXCR1 and CXCR2 and suitable for chronic oral administration because of its longer in vivo half-life (35). In a total of 52 male C57BL/6 mice that received MLD STZ treatment (Fig. 1C), ladarixin treatment significantly prolonged the duration of diabetes development, and its efficiency was strictly dependent on the time when treatment was started and the duration of treatment (Fig. 1C and D). The median diabetes-free time was 10 ± 2.5 days for vehicle-treated mice (n = 12). When started on day −1, the median diabetes-free time was 13 ± 2 days (P = 0.22 vs. control, n = 8) and 29 ± 16 days (P = 0.002 vs. control, n = 8), respectively, for 7 (data not shown) and 14 days of ladarixin treatment. When treatment with ladarixin for 14 days was started at day +5 or +10 after the first STZ injection, the median diabetes-free time was 16 ± 3 days (P = 0.031 vs. control, n = 12) and 10 ± 1 days (P = 0.61 vs. control, n = 12), respectively (data not shown). Similar to reparixin treatment, glycemic levels remained constantly and significantly lower in the ladarixin-treated group than in the vehicle-treated group after diabetes developed (Fig. 1D).

To determine whether pharmacologic blockade of CXCR1/2 prevents the development of spontaneous diabetes in NOD mice, female mice were randomized to receive ladarixin or vehicle for 2 consecutive weeks starting in the early or late preclinical stage of the disease.

Ladarixin significantly delayed and prevented diabetes onset when administered at 12 weeks of age (P = 0.007; Fig. 2A), but not at 4 weeks of age, even if a trend was evident (P = 0.175). Specifically, 22% of the mice receiving CXCR1/2 inhibitor from 12 weeks of age developed diabetes during the follow-up (i.e., 62 weeks of age; mean time to diabetes 25.8 ± 9.4 weeks), whereas 78% of the vehicle-treated mice developed diabetes during the same period (mean time to diabetes 17.8 ± 4.1 weeks; Fig. 2A). As for 12-week-old NOD mice, treatment of 8-week-old mice resulted in significantly delayed diabetes onset (P = 0.046; data not shown).

To determine the progression of islet mass loss and insulitis in NOD mice and the effect of CXCR1/2 blockade on such events, we analyzed islet numbers and insulitis in ladarixin- and vehicle-treated mice at the end of the treatment (Fig. 2B). There was no significant difference in islet numbers between ladarixin- and vehicle-treated groups or between 4- and 12-week-old mice. However, insulitis score (Fig. 2C) and infiltrating CD3⁺ T cells, F4/80⁺ macrophages, and MPO⁺ neutrophils were significantly skewed toward lower values (Fig. 3).

Pancreas-infiltrating leukocytes were isolated from 12-week-old mice and characterized by flow cytometry (CD45, Gr-1, Ly6C, CD11b, CXCR2) at the end of the treatment. The number of leukocytes was significantly smaller in ladarixin-treated than in vehicle-treated mice (250,000 ± 57,735 vs. 440,000 ± 89,814 cells/pancreas; P = 0.012). Among all CD45⁺ cells, CXCR2⁺ cells were less abundant in ladarixin-treated than in vehicle-treated mice (17.8 ± 5.1% vs. 10.7 ± 2.3%; P = 0.045) (Supplementary Fig. 3). CD45⁺ and CD45⁺CXCR2⁺ cells were analyzed separately for their expression of Gr-1, Ly6C, and CD11b (Fig. 4A and B). Among CD45⁺, lymphoid cells (Gr-1⁻Ly6C⁻CD11b⁻) were the most abundant leukocytes (65 ± 17%), followed by macrophages (Gr-1⁻Ly6C⁺CD11b⁺, 23 ± 7%) and neutrophils (PMNs; Gr-1⁺Ly6C⁻CD11b⁺, 1.5 ± 1.1%). Ladarixin treatment significantly reduced the percentage of PMNs and the absolute number of PMNs and lymphoid cells.

Among CD45⁺CXCR2⁺cells, macrophages (94 ± 3%) were the most abundant, followed by lymphoid cells (4 ± 2%) and PMNs (1.3 ± 1.3%). Ladarixin treatment significantly reduced the absolute number of macrophages.

CD45⁺ leukocytes were analyzed by flow cytometry in different tissues (i.e., blood, spleen, pancreatic lymph nodes, and bone marrow) after 14 days of vehicle or ladarixin treatment. The analysis showed myeloid cells and, in particular, PMNs, readily identified by expression of Gr-1 and/or Ly6C, as the predominant CXCR2⁺ cells among blood, spleen, and bone marrow leukocytes (Supplementary Table 1). CXCR2 is also expressed, depending on the tissue context, by part of CD4⁻CD8⁻ double negative CD3⁺ lymphocytes and by few CD19⁺ B lymphocytes and CD3⁺CD4⁺ or CD3⁺CD8⁺ T lymphocytes.

In 4-week-old mice (Supplementary Table 2) the CXCR1/2 inhibitor induced mild modification of white blood cells, whereas spleen and pancreatic lymph nodes were unaffected. A significant increase in total circulating leukocytes was evident and, among CD45⁺ cells, the proportions of CXCR2⁺ cells and PMNs increased.

In 12-week-old mice (Supplementary Table 2), major and more extensive changes were observed. Total blood circulating leukocytes were greatly increased; cellularity was augmented in pancreatic lymph nodes and reduced in bone marrow. Ladarixin treatment decreased significantly the proportion of CD45⁺CXCR2⁺ cells in blood and lymph nodes. At least one of the myeloid cell fractions (Gr-1⁺Ly6C⁻CD11b⁺, Gr-1⁻CD11b⁺Ly6C⁺) was consistently reduced by CXCR1/2 inhibition in blood, spleen, and bone marrow. A surprising increase in CD45⁺/CD19⁺ B lymphocytes was evident in blood and spleen, whereas ladarixin treatment differed in its effects on T lymphocytes depending on site: fractional expression of CD45⁺/CD3⁺ T lymphocytes increased in blood and bone marrow and decreased in spleen. CD4⁺CD25⁺Foxp3⁺ regulatory T cells in blood and spleen were not affected by CXCR1/2 treatment (data not shown).

We also compared the levels of systemic cytokines between vehicle- and ladarixin-treated mice at different ages. With the exception of a mild significant decrease of interleukin (IL)-9 and IL-6 in 4- and 12-week-old mice treated with the CXCR1/2 inhibitor, respectively, the cytokine profiles did not change (Supplementary Table 3).

NOD female mice with recent-onset diabetes received either ladarixin (15 mg/kg/day; n = 28) or vehicle (n = 22) for 14 days. Strikingly, diabetes was reversed in 22 of 28 ladarixin-treated NOD mice (78%) (Fig. 5A). Remission was rapid (detected in 20 of 22 NOD mice [91%] within 72 h of treatment), and 5 of 22 remission NOD mice (23%) remained diabetes free >10 weeks after onset (Fig. 5B). The majority of NOD mice receiving vehicle, however, failed to undergo remission or, if induced (2 of 22 [9%]; P < 0.001), remission was short (7 and 17 days). Diabetes reversion was not associated with modification of CD4⁺CD25⁺Foxp3⁺ regulatory T cells in blood and spleen (data not shown) or systemic cytokine profile changes, with the exception of a mild IL-12p70 increase (data not shown).

Notably, the average blood glucose concentration in ladarixin-treated mice remained significantly lower than that of vehicle-treated mice during the entire follow-up (Fig. 5C). Lower glucose concentrations at diabetes onset correlated with a higher efficacy of ladarixin treatment. In fact the CXCR1/2 inhibitor maintained the ability to reverse diabetes (43%) in highly hyperglycemic NOD mice (glucose concentrations ≥350 mg/dL), but diabetes reversion did not demonstrate a long-term benefit (Supplementary Fig. 4). Similarly, late diabetes onset (>21 weeks of age) did not affect ladarixin-induced reversion of diabetes but correlated with an earlier recurrence of the disease (Supplementary Fig. 4).

Histological analysis was performed to determine the progression of islet mass loss and insulitis and the effect of CXCR1/2 blockade on such events (Fig. 6). According to blood glucose concentration, a significantly higher number of islets were counted in the pancreas of ladarixin-treated mice compared with vehicle-treated mice and, among ladarixin-treated mice, in the pancreas of mice with diabetes remission. Both insulitis (Fig. 6) and leukocyte infiltration scores (Fig. 7) were significantly decreased toward lower values in ladarixin-treated mice.

We recently reported that the inhibition of CXCR1/2 chemokine receptors is relevant for improving human and murine islet survival after transplantation (21). Here we demonstrate that the same pathway seems to be relevant for islet survival after inflammation- and autoimmunity-mediated damage in two preclinical models. To our knowledge this is the first study demonstrating success in preventing and reversing diabetes by inhibiting CXCR1/2 chemokine receptors, identifying this pathway as a “master regulator” of diabetes pathogenesis. Concordantly, Diana and Lehuen (36) recently reported that treatment with the CXCR2 antagonist SB225002 at early ages dampens the later development of autoimmune diabetes in NOD mice. Even if SB225002 seemed to be ineffective when used at a late age (36), the results strongly support the therapeutic potential of CXCR1/2 chemokine receptors. More than one-half of the ∼50 human/mouse chemokines have been previously associated with, or implicated in, the pathogenesis of T1D (12), but the identification of a single specific chemokine/receptor pathway that may constitute a suitable target for the development of therapeutic interventions is still lacking (37). Chemokines such as CCL2, CCL3, CCL5, CCL22, and CXCL10 and their receptors were suggested to directly affect T1D development. However, disease prevention in the NOD model can be achieved only by interfering with more than one of them (37,38). Individual treatments reduced inflammatory infiltrates without effectively preventing diabetes development in some experimental settings (12) or even worsening the disease in others. For example, NOD mice deficient in CXCR3 (CXCL19/CXCL10 receptor) or CCR5 (CCL3/CCL4/CCL5 receptors) developed spontaneous diabetes earlier than wild-type NOD mice (39,40).

In our study we established a relationship involving the CXCR1/2 chemokine receptors and the development of insulitis and diabetes in mice. We first demonstrated that transient CXCR1/2 inhibition prevents inflammation-mediated islet damage in the MLD STZ model. MLD STZ–induced β-cell destruction is associated with a cell-dependent inflammatory reaction and with a central pathogenic role of proinflammatory cytokines, such as IL-1β and interferon-γ (41), but does not require functional T and B cells (42). In this model two different CXCR1/2 inhibitors were effective in protecting islets, whereas a nonactive, structurally related compound was not, confirming that the effect was not caused by multiple off-target actions. Second, we demonstrated that transient CXCR1/2 blockade inhibits autoimmune insulitis in NOD mice, clearly reducing neutrophil, macrophage, and T-cell infiltration. This effect occurred independent of age and, starting from 8 weeks, was associated with the prevention of autoimmunity-mediated islet damage. Of note, even if batches of control and treatment groups run in parallel at all ages, the rate of T1D onset in 4-week-old NOD mice in the control group was relatively lower than in the 12-week-old mice, and this pattern may have contributed to reducing the impact of the intervention started at the earliest age. Third, we demonstrated that transient CXCR1/2 inhibition is able to reverse diabetes, preserving islet mass after onset (21). This is in agreement with indirect evidence previously reported. Human pancreatic islets produce and secrete the proinflammatory CXCL8 and CXCL1 ligands (12,15,17,22,23), whereas lipopolysaccharide-induced production of CXCL8 by neutrophils is increased in prediabetic and T1D patients. In parallel, circulating concentrations of CXCL8 are elevated in children with T1D compared with nondiabetic controls (24–26). Specifically, concentrations of CXCL8 correlate with glycemic control; higher concentrations are associated with poor or inadequate glucose control. Finally, and most important, recent reports demonstrate that neutrophils, the major target of CXCR1/2 inhibitors, play a key role in the etiopathogenesis of T1D in mice and humans (27–29,36).

The CXCR1/2 noncompetitive allosteric inhibitors used in our work have different pharmacologic characteristics (31). Ladarixin is a potent blocker of both CXCR1 and CXCR2, with a half-maximal inhibitory concentration in the range of 1 to 2 nmol/L (31). In contrast, reparixin exhibits ∼400-fold greater efficacy in blocking CXCR1 compared with CXCR2 (30). We imputed the effect mainly to the inhibition of CXCR2 and not of CXCR1 in mice. Reparixin treatment at a dose able to inhibit mainly CXCR1, but not CXCR2, was unable to prevent diabetes in the MLD STZ model. Moreover, despite recent studies reporting the existence of a mouse homolog of human CXCR1 (43–45), none of the known CXCR1/2-related chemokines have been shown to activate this putative receptor (43–45). Consequently, mice are considered to have only functional CXCR2 and, although we cannot completely exclude a possible action mediated by the inhibition of some not yet discovered CXCR1 functions, CXCR2 should be considered as the unique target of CXCR1/2 inhibitors in mice.

To further investigate the protective mechanism(s) mediated by the CXCR1/2 inhibitor in NOD mice, we analyzed CXCR2 expression, systemic leukocytes by flow cytometry, and the systemic cytokine profile by a multiplex approach. CXCR2 was confirmed to be highly expressed on myeloid cells. Nevertheless, CXCR2 also was detected on lymphoid cells (natural killer T cells, B cells, and CD8⁺ and CD4⁺ T cells). In concordance with this, intrapancreatic myeloid CD45⁺CXCR2⁺ cells were significantly reduced by CXCR1/2 blockade. Other significant modifications observed in the later phase of the disease involved not only the myeloid subpopulation in bone marrow and spleen but also the circulating lymphoid subpopulation. Of note is the observation of increased B cells in ladarixin-treated NOD mice. The role of B cells in the modulation of islet immune cell responses has been recently recognized in models of autoimmune diabetes (46,47) and immunomodulatory effects of Gr-1⁺CD11b⁺ cells were reported in NOD mice receiving anti-CD20 treatment (48). In addition, IL-9, IL-6, and IL-12p70 were modified by CXCR1/2 inhibitor treatment in relation to the different disease phases. Currently, determining which of these modifications have the most relevant causal relationship with the prevention of diabetes is not possible. Moreover, we cannot exclude that diabetes prevention could be caused by the combined activities of several leukocyte subpopulations. The picture is further complicated by the fact that the CXCR1/2 ligands are not limited to CXCL1/2/8 produced by islets but also include CXCL3/5/7 in mice (which lack CXCL8) and CXCL3/5/6/7, as well as some other ligands, in humans (49). Additional studies, including selective CXCR2 knockout in leukocyte populations, are required to ascertain the impact of individual changes. Since depletion of neutrophils, the major CXCR2-positive circulating leukocyte population, has recently been suggested to reduce spontaneous T1D incidence in NOD mice (36), we can speculate that they could constitute the more relevant target of CXCR1/2 inhibitors. On this basis preliminary experiments with anti-Ly6G, a neutrophil-depleting agent, in association or comparison with the CXCR1/2 inhibitor, were performed both in the MLD STZ model and 12-week-old NOD mice (data not shown). Anti-Ly6G alone was inefficient and combination with the CXCR1/2 inhibitor was as efficient as the CXCR1/2 inhibitor alone in preventing diabetes. These results suggest that the action of the CXCR1/2 inhibitor on neutrophils could be necessary but not sufficient to prevent diabetes in these models.

The results we obtained in preclinical models offer important insights in the context of the literature and suggest possible avenues and expectations in future clinical trials. Preserving residual β-cell function has been associated with a reduced rate of microvascular complications and hypoglycemia, an improved quality of life, and an overall reduction in morbidity and disease-associated management costs (50). Therefore, approaches aimed at controlling the autoimmune response and restoring self-tolerance to pancreatic β-cells have attracted much clinical and scientific interest (6). Among the many pharmacologic interventions directed at this purpose, rituximab (51), abatacept, CD3-specific monoclonal antibodies (tepluzimab, otelixizumab) (52), GAD65 (Diamyd) (53), and HSP p277 (DiaPep) have progressed to phase IIB/III clinical trials (7). Other agents, including cytokine modulators (e.g., anti-tumor necrosis factor– or anti-IL-1–based agents) (54) are also under consideration for large-scale evaluation. Unfortunately, despite some positive outcomes, none of the pharmacologic approaches tested thus far have resulted in a consensus agreement that acknowledges an acceptable means for preserving C-peptide after diabetes onset. In addition, when these trials are considered together, it seems that therapies based on a single drug are unlikely to have substantial effects on the prevention or amelioration of hyperglycemia. Our data support the idea that the inclusion of CXCR1/2 inhibitors in future trials of multiple-drug regimens for T1D may provide therapeutic benefit by preserving β-cell mass. Interestingly, in the current clinical experience, ladarixin was already tested in humans in three phase I pharmacokinetic/safety studies, including repeated oral administrations, and it was safe and well tolerated. On this basis a trial of ladarixin in patients recently diagnosed with T1D will be started in the next months.

In conclusion, we identified CXCR1/2 chemokine receptors and their ligands as “master regulators” and demonstrated that CXCR2 blockade could be a useful strategy in preventing and/or reversing diabetes in mice. The demonstration that the inhibition of this pathway may be successful in preserving residual β-cells in patients with new-onset T1D holds the potential to significantly change our approach to the management of this disease. This treatment, focused on inhibiting inflammatory response by targeting leukocyte migration, could represent an ideal complement to any intervention acting on the interdiction of the pathogenic autoimmune process.

Acknowledgments. This work was performed in partial fulfillment of the requirements for a PhD degree for A.V. and E.C.

Funding. This research was supported by the European Union (HEALTH-F5-2009-241883-BetaCellTherapy) and was partially funded by Dompè Farmaceutici S.p.A in the fiscal year preceding the date of submission.

Duality of Interest. L.D., O.K., P.A.R., and M.A. are Dompè Farmaceutici S.p.A employees. L.P. received grants for research by Dompè Farmaceutici S.p.A in the fiscal year preceding the date of submission and is involved in two clinical trials supported by Dompè Farmaceutici S.p.A (NCT01220856 and NCT01817959). No other conflicts of interest relevant to this article were reported.

Author Contributions. A.C. and E.C. performed the in vivo mouse studies. A.C. and A.V. performed flow cytometry immunophenotyping. A.C., A.V., A.M., and S.P. performed the histopathological analysis of the pancreas. D.L. performed multiplex analysis of cytokines. L.D., M.B., and M.A. reviewed and edited the manuscript and contributed to the discussion. O.K. and P.A.R. developed and provided CXCR1/2 inhibitors. L.P. developed the concept, designed the experiments, wrote the manuscript, promoted the study, and researched data. L.P. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.