Dual targeting of polyamine synthesis and uptake in diffuse intrinsic pontine gliomas

Abstract

Diffuse intrinsic pontine glioma (DIPG) is an incurable malignant childhood brain tumor, with no active systemic therapies and a 5-year survival of less than 1%. Polyamines are small organic polycations that are essential for DNA replication, translation and cell proliferation. Ornithine decarboxylase 1 (ODC1), the rate-limiting enzyme in polyamine synthesis, is irreversibly inhibited by difluoromethylornithine (DFMO). Herein we show that polyamine synthesis is upregulated in DIPG, leading to sensitivity to DFMO. DIPG cells compensate for ODC1 inhibition by upregulation of the polyamine transporter SLC3A2. Treatment with the polyamine transporter inhibitor AMXT 1501 reduces uptake of polyamines in DIPG cells, and co-administration of AMXT 1501 and DFMO leads to potent in vitro activity, and significant extension of survival in three aggressive DIPG orthotopic animal models. Collectively, these results demonstrate the potential of dual targeting of polyamine synthesis and uptake as a therapeutic strategy for incurable DIPG.

Fig. 1. Polyamine synthesis and catabolic genes in pediatric brain tumors.

Expression of polyamine (a) biosynthetic (p < 0.0001) and (b) catabolic enzyme genes (p < 0.0001) in DIPG (n = 49) compared to normal fetal brain (n = 11). Data were obtained from ZCC/ PRISM Clinical trial and McGill University. Protein (c) and gene (d) expression of the polyamine biosynthetic gene ODC1, and catabolic gene SAT1 (e) in patient-derived DIPG cell lines. Western blot is representative of two independent experiments. Putrescine concentrations (f) (p = 0.0087) and spermidine to spermine ratio (spd:spm) (g) (p = 0.003) in DIPG tumor injected brainstem region of mouse brain. h Exogenous polyamines were added to HSJD-DIPG007 cells at 10 μM. p-value was calculated by t-tests for control and addition of three polyamines at each time point. Day 2: p = 0.00171. Day 3: p = 0.0115. Day 5: p = 0.0194. (i) Patient-derived DIPG cell lines are sensitive to polyamine synthesis inhibition via DFMO, compared to normal healthy astrocytes (NHA) and normal lung fibroblast (MRC5) cells. Data are presented as mean values ± SEM of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. a, b, f–h Statistical analysis was performed by two-tailed t tests for normal and DIPG samples. d, e Statistical analysis was performed by one-way ANOVA between normal NHA cells and patient-derived DIPG cell lines. n = 1 from three independent experiments. d NHA vs SU-DIPGVI: p < 0.0001, NHA vs HSJD-DIPG007: p = 0.0478, NHA vs HSJD-DIPG008: p = 0.0239, NHA vs HSJD-DIPG011: p < 0.0001, NHA vs RA055: p = 0.0165. e NHA vs SU-DIPGVI: p = 0.0070, NHA vs HSJD-DIPG007: p = 0.0111, NHA vs HSJD-DIPG008: p = 0.0087, NHA vs RA055: p = 0.0066.

Fig. 2. Polyamine transporter expression in pediatric DIPG brain tumors.

Expression of polyamine transporter SLC3A2 in (a) DIPG tumors (n = 49) compared to normal fetal brain (n = 11) (p < 0.0001). Data were obtained from ZCC/PRISM Clinical trial and McGill University. b–c Examination of RNA expression levels in a cohort of high-risk childhood cancers showed that the polyamine transporter, SLC3A2, was significantly overexpressed in DMG (n = 28) compared with high-risk childhood cancers (n = 148) (p < 0.0001) and neuroblastoma (n = 17) (p < 0.0001). Treatment of HSJD-DIPG007 cells with 40 mM DFMO leads to increased SLC3A2 (d) protein and (e) gene expression, resulting in (f) increased uptake of radiolabeled spermidine (p = 0.0276). g Polyamine transport inhibition by AMXT 1501 decreased uptake of radiolabeled spermidine and (h) was cytotoxic against patient-derived DIPG cell lines. Data are presented as mean values ± SEM of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. a, b, c, f Statistical analysis was calculated using two-tailed t tests between sample cohorts. e, g Statistical analysis was calculated using one-way ANOVA between cohorts and for treated and untreated samples. e UT vs 24 h: p = 0.0005, UT vs 48 h: p < 0.0001, UT vs 72 h: <0.0001, 12 h vs 24 h: p = 0.0107, 12 h vs 48 h: p = 0.0001, 12 h vs 72 h: p < 0.0001, 24 h vs 48 h: p = 0.0142, 24 h vs 72 h: p = 0.0021. g Control vs 0.01: p < 0.0001, Control vs 0.05: p < 0.0001, Control vs 0.01: p < 0.0001, 0.01 vs 0.05: p = 0.0092. 0.01 vs 0.1: p = 0.0026.

Fig. 3. Inhibition of synthesis and transport depletes intracellular polyamines.

(a) Putrescine, (b) Spermidine, and (c) Spermine levels in SU-DIPGVI cells treated with DFMO and AMXT 1501 alone and in combination. Data are presented as means ± SEM of three independent experiments. *p < 0.05, **p < 0.01. p-values were calculated using one-way ANOVA for untreated and treated samples. Synergy calculated using the Chou–Talalay method. a Vehicle vs DFMO: p = 0.0125, Vehicle vs AMXT 1501: p = 0.0072, Vehicle vs DFMO + AMXT 1501: p = 0.0041, DFMO vs DFMO + AMXT 1501: p = 0.0378, AMXT 1501 vs DFMO + AMXT 1501: p = 0.0028. b Vehicle vs DFMO: p = 0.0364, Vehicle vs AMXT 1501: p = 0.0152, Vehicle vs DFMO + AMXT 1501: p = 0.0047, DFMO vs DFMO + AMXT 1501: p = 0.0212, AMXT 1501 vs DFMO + AMXT 1501: p = 0.0027. c Vehicle vs DFMO: p = 0.0492, Vehicle vs AMXT 1501: p = 0.0351, Vehicle vs DFMO + AMXT 1501: p = 0.0077, DFMO vs DFMO + AMXT 1501: p = 0.0421, AMXT 1501 vs DFMO + AMXT 1501: p = 0.0026.

Fig. 4. The effect of combined inhibition of polyamine synthesis and uptake, using DFMO and AMXT 1501, on cell proliferation, colony formation and apoptosis.

HSJD-DIPG007 cells have synergistically decreased (a) cell survival and (b) form fewer colonies, upon combination treatment. c Synergy scores, calculated by Calcusyn, across a panel of neurosphere-forming DIPG cells treated with combination of both agents. d Annexin/PI staining of HSJD-DIPG007 cells treated with 40 mM DFMO, 5 µM AMXT 1501 or combination of both agents for 24 h. Vehicle vs DFMO: p = 0.0221, Vehicle vs AMXT 1501: p = 0.0003, Vehicle vs DFMO + AMXT 1501: p < 0.0001, DFMO vs DFMO + AMXT 1501: p < 0.0001, AMXT 1501 vs DFMO + AMXT 1501: p = 0.0005. e Apoptotic effects of DFMO and AMXT 1501 combination on protein expression of cleaved PARP and caspase 8. Representative blot from two independent experiments. Data are presented as mean values ± SEM of three independent experiments. a, b n = 3 wells examined over three independent experiments. d n = 3 independent experiments. **p < 0.01, ***p < 0.001. p-values were calculated using two-tailed t tests for treated and untreated cohorts.

Fig. 5. Evans Blue (EB) extravasation assay in three orthotopic models of DIPG.

a Representative images of Balbc/Nude mice injected with Evans Blue dye. b Representative images of brains and spleens harvested from animals intracranially injected with matrigel and RA055 cells. c Brainstem region shows low EB extravasation with no difference among the uninjected, matrigel injected and DIPG injected animals (RA055, SU-DIPGVI, HSJD-DIPG007) and no differences were observed with the tumor free cortical region; Splenic tissue shows higher EB extravasation compared to both brain regions. Data are presented as mean values ± SEM of samples collected from three mice in each cohort. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Statistical analysis was performed using two-way ANOVA. Spleen-uninjected vs matrigel: p = 0.0012, uninjected vs SU-DIPGVI: p = 0.0040, uninjected vs HSJD-DIPG007: p < 0.0001, uninjected vs RA055: p = 0.04114. Brainstem vs Spleen-uninjected: p < 0.0001, matrigel: p = 0.0001, SU-DIPGVI: p < 0.0001, HSJD-DIPG007: p = 0.0005, RA055: p < 0.0001. Cortical vs Spleen-uninjected: p < 0.0001, matrigel: p < 0.0001, SU-DIPGVI: p < 0.0001, HSJD-DIPG007: p = 0.0003, RA055: p = 0.0002.

Fig. 6. Therapeutic efficacy of DFMO, AMXT 1501 and combination treatment in orthotopic models of DIPG.

Mice were intracranially injected with DIPG cells and 4–8 weeks post injection treatments commenced. Mice were humanely euthanised when they displayed severe neurological decline and/or weight loss or reached maximum holding time (MHT). a Survival curve of SU-DIPGVI-Luciferase with DFMO/AMXT 1501 treatment. Median survival of cohorts (days): Vehicle = 95, DFMO = 101, AMXT 1501 = 96, DFMO + AMXT 1501 = undefined. Exact p-values listed in Supplementary Table 7. b–c Xenogen imaging showed the cytostatic effect of DFMO/AMXT 1501 on SU-DIPGVI DIPG cells. Vehicle vs DFMO/AMXT1501—Day 63: p = 0.02301, Day 76: p = 0.0024, Day 77: p < 0.0001, Day 81: p < 0.0001, Day 98: p < 0.0001. DFMO vs DFMO/AMXT1501—Day 98: 0.0050. d Survival curve of HSJD-DIPG007 with DFMO/AMXT 1501 treatment. Median survival of cohorts (days): Vehicle = 55, DFMO = 65, AMXT 1501 = 59, DFMO + AMXT 1501 = undefined. Exact p-values listed in Supplementary Table 8. e Ki67 staining of DIPG tumors post 4 weeks of treatment in HSJD-DIPG007 model. Three images were taken from brain samples collected from three mice in each cohort. Scale bars: black 50um. Vehicle vs DFMO: p = 0.0005, Vehicle vs DFMO/AMXT1501: p < 0.0001, DFMO vs DFMO/AMXT1501: p = 0.0099, AMXT1501 vs DFMO/AMXT1501: p < 0.0001. f Survival curve of RA055-xenografted mice with DFMO/AMXT 1501 treatment. Median survival of cohorts (days): Vehicle = 44, DFMO = 53, AMXT 1501 = 48, Irradiation = 61, DFMO + Irradiation = 66 days, AMXT 1501/Irradiation=69 days, DFMO + AMXT 1501 = 136, DFMO + AMXT 1501+Irradiation = 158. Exact p-values listed in Supplementary Table 9. g Putrescine levels in the tumor-ridden brainstem region in SU-DIPGVI-LUC model, post 1 week treatment. Brain samples were collected from three mice in each cohort. Vehicle vs DFMO: p = 0.0005, Vehicle vs DFMO/AMXT1501: 0.0004, DFMO vs DFMO/AMXT1501: p = 0.0315, AMXT1501 vs DFMO/AMXT1501: p = 0.0021. h Spermidine to spermine ratio (spd:spm). Brain samples were collected from three mice in each cohort. Vehicle vs DFMO/AMXT1501: p = 0.0110, AMXT1501 vs DFMO/AMXT1501: p = 0.0146. a, d, f Shaded areas indicate treatment period of combination treatment, which was continuous in RA055 model. Statistical analysis has been performed using the Log-Rank (Mantel–Cox) with multiple test correction applied. c, e, g, h Data are presented as mean values ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. p-values were calculated using two-way ANOVA (c) and one-way ANOVA (e, g, h) for treated and untreated cohorts.