TLR2 agonism reverses chemotherapy-induced neutropenia in Macaca fascicularis

Abstract

Neutropenia is a common consequence of radiation and chemotherapy in cancer patients. The resulting immunocompromised patients become highly susceptible to potentially life-threatening infections. Granulocyte colony-stimulating factor (G-CSF) is known to stimulate neutrophil production and is widely used as a treatment of chemotherapy-induced neutropenia. A small-molecule G-CSF secretagogue without a requirement for refrigerated supply chain would offer a more convenient and cost-effective treatment of chemotherapy-induced neutropenia. Bacterial lipopeptides activate innate immune responses through Toll-like receptor 2 (TLR2) and induce the release of cytokines, including G-CSF, from macrophages, monocytes, and endothelial. Pam₂CSK₄ is a synthetic lipopeptide that effectively mimics bacterial lipoproteins known to activate TLR2 receptor signaling through the TLR2/6 heterodimer. Substrate-based drug design led to the discovery of GSK3277329, which stimulated the release of G-CSF in activated THP-1 cells, peripheral blood mononuclear cells, and human umbilical vein endothelial cells. When administered subcutaneously to cynomolgus monkeys (Macaca fascicularis), GSK3277329 caused systemic elevation of G-CSF and interleukin-6 (IL-6), but not IL-1β or tumor necrosis factor α, indicating a selective cytokine-stimulation profile. Repeat daily injections of GSK3277329 in healthy monkeys also raised circulating neutrophils above the normal range over a 1-week treatment period. More importantly, repeated daily injections of GSK3277329 over a 2-week period restored neutrophil loss in monkeys given chemotherapy treatment (cyclophosphamide, Cytoxan). These data demonstrate preclinical in vivo proof of concept that TLR2 agonism can drive both G-CSF induction and subsequent neutrophil elevation in the cynomolgus monkey and could be a therapeutic strategy for the treatment of chemotherapy-induced neutropenia.

Graphical abstract

Figure 1.

Structure and in vitro activity of GSK3277329. (A) Chemical structure of GSK3277329. (B-C) TLR2 agonists stimulate G-CSF release in a concentration-dependent manner in PMA-differentiated human THP-1 cells (n = 2 replicates) (B) and HUVECs (GSK329 n = 6 replicates, PAM2CSK4 n = 2 replicates) (C). Graphs show mean ± standard error of the mean (SEM). Error bars are not shown if the error is smaller than the symbol size.

Figure 2.

Pharmacokinetics of GSK3277329 in M fascicularis. Compound levels were measured in plasma at the indicated times after subcutaneous injection of 0.2, 1, and 15 mg/kg. Mean ± SEM of 2 animals per group at 0.2 and 1 mg/kg and 3 animals for the 15-mg/kg dose group are shown. Error bars are not shown if the error is smaller than the symbol size. H, hours.

Figure 3.

Acute effects of TLR2 agonists on cytokine production and white blood cell counts in M fascicularis. (A-D) After a single subcutaneous injection of either Pam₂CSK₄ (red bars, 1-mg/kg dose) or GSK3277329 (blue bars, 15-mg/kg dose), circulating levels of G-CSF (A), MCP-1 (C), and IL-6 (D) and absolute neutrophil counts (B) were measured at either 6 or 24 hours after injection. Bars represent mean ± SEM (3 animals per group). Error bars are not shown for baseline IL-6 values, because the values shown are at the limit of quantification level. ANOVA analysis shows a significant effect of treatment on these graphed parameters (P < .05).

Figure 4.

Effect of repeat dosing of GSK3277329 (5 and 15 mg/kg subcutaneous injection once a day) on circulating neutrophil counts and the correlation with G-CSF levels in M fascicularis. (A) Neutrophil counts before, 6 hours after, and 24 hours after the first dose of GSK3277329 and the seventh dose. Significant difference compared with predose group (t = 0) determined by 2-way ANOVA followed by Sidak’s post hoc test (*P < .05, **P < .01). (B) Correlation plot between circulating G-CSF levels and neutrophil counts measured 6 hours after GSK3277329 administration, grouping all animals from both dosing groups and samples collected after the first and seventh dose (3 animals per dose group).

Figure 5.

The effect of a TLR2 agonist or G-CSF on IL-6 or G-CSF levels in cyclophosphamide-treated M fascicularis. (A) Circulating G-CSF levels in M fascicularis (n = 3 per group) after vehicle, cyclophosphamide (CYP), or CYP plus GSK3277329 (3 mg/kg) or CYP plus G-CSF (10 µg/kg) treatment. (B) Circulating IL-6 levels after vehicle, CYP, or CYP plus GSK3277329 (3 mg/kg) or CYP plus G-CSF (10 µg/kg) treatment. Samples were taken at the indicated times after the first or 14th dose. Mean ± SEM values are plotted. No error bars are shown if the error is smaller than the height of the symbol. Significant differences against the vehicle and CYP groups at each time point using 2-way ANOVA with Tukey’s post hoc test are noted by asterisks (**P < .0001, *P < .005).

Figure 6.

Effect of a TLR2 agonist or G-CSF on chemotherapy-induced neutropenia. Symbols indicate mean ± SEM values. Error bars are not shown if the error is smaller than the size of the symbol. (A-D) Neutrophil (A), lymphocyte (B), monocyte (C), and red blood cell (D) counts were measured in M fascicularis treated with vehicle, cyclophosphamide (CYP), or CYP plus GSK3277329 (3 mg/kg) or CYP plus G-CSF (10 µg/kg) (n = 3 per group). The normal range of neutrophil numbers between 1 and 10 × 10⁹ cells/L is noted by dashed lines (A). GSK3277329 or G-CSF were dosed daily for 14 days as indicated by the linked arrows. The first dose received was on day 5, and the last dose on day 19. Significant differences vs the CYP group were determined by 2-way ANOVA followed by Dunnett’s multiple comparisons post-hoc test (*P < .05, **P < .01, ***P < .005).

Figure 7.

Neutrophil counts in vehicle-treated animals or animals treated with 2 doses of cyclophosphamide. Cyclophosphamide (CYP)–treated animals received vehicle or 1, 3, or 7 daily doses of GSK3277329 (15 mg/kg) 2 days after the last cyclophosphamide dose. Neutrophil numbers were counted 7 days after the first dose. Bars represent mean ± SEM. *P < .05 vs vehicle by ANOVA and Dunnett’s multiple comparison post hoc test (n = 3 per group).