Novel therapies for cancer-induced bone pain

Abstract

Cancer pain is a growing problem, especially with the substantial increase in cancer survival. Reports indicate that bone metastasis, whose primary symptom is bone pain, occurs in 65-75% of patients with advanced breast or prostate cancer. We optimized a preclinical in vivo model of cancer-induced bone pain (CIBP) involving the injection of Lewis Lung Carcinoma cells into the intramedullary space of the femur of C57BL/6 mice or transgenic mice on a C57BL/6 background. Mice gradually reduce the use of the affected limb, leading to altered weight bearing. Symptoms of secondary cutaneous heat sensitivity also manifest themselves. Following optimization, three potential analgesic treatments were assessed; 1) single ion channel targets (targeting the voltage-gated sodium channels Na_V1.7, Na_V1.8, or acid-sensing ion channels), 2) silencing µ-opioid receptor-expressing neurons by modified botulinum compounds, and 3) targeting two inflammatory mediators simultaneously (nerve growth factor (NGF) and tumor necrosis factor (TNF)). Unlike global Na_V1.8 knockout mice which do not show any reduction in CIBP-related behavior, embryonic conditional Na_V1.7 knockout mice in sensory neurons exhibit a mild reduction in CIBP-linked behavior. Modified botulinum compounds also failed to cause a detectable analgesic effect. In contrast, inhibition of NGF and/or TNF resulted in a significant reduction in CIBP-driven weight-bearing alterations and prevented the development of secondary cutaneous heat hyperalgesia. Our results support the inhibition of these inflammatory mediators, and more strongly their dual inhibition to treat CIBP, given the superiority of combination therapies in extending the time needed to reach limb use score zero in our CIBP model.

Keywords: Acid-sensing ion channels; Cancer-induced bone pain; Dorsal root ganglia; Modified botulinum compounds; Nerve growth factor; Tumor necrosis factor alpha; Voltage-gated sodium channels.

Fig. 1

The injection of LL/2 cells into the intramedullary space of the femur of mice caused a significant decline in the limb-use score (A) and the weight-borne on the ipsilateral limb (B). This decline was dependent on the number of LL/2 cells injected. (A) The mean limb-use score became significantly lower than the baseline on day 10 for the 2 × 10⁵ cancer cells group (p-value = 0.0095), day 15 for the 2 × 10⁴ cancer cells group (p-value 0.0007), and day 19 for the 2 × 10³ cancer cells group (p-value 0.0349). (B) The analysis of the weight-bearing results showed that the reduction in the weight-bearing became significant (compared to the baseline) on day 11 after surgery for the 2 × 10⁵ group (p-value = 0.0444), on day 12 after the surgery (p-value = 0.0456) in the mice that received 2 × 10⁴ cells and after 14 days in the mice that received 2 × 10³ cancer cells (p-value 0.0212). Analyses of the limb-use scoring, and weight-bearing results showed that the effect of the cancer cells is statistically significant, with the speed of reduction of limb-use and weight-borne on the ipsilateral limb being directly proportional to the number of cancer cells inoculated (asterisks of significance are shown on the graph (the mixed-effects model for the effect of cancer number over the whole study period)). At the baselines, n = 12 for the 2 × 10⁴ group and the 2 × 10³ group (equal mix of sexes), and n = 5 males for the 2 × 10⁵ cancer cells group. REML analysis (followed by Dunnett’s test for multiple comparisons) was used to estimate the time required to cause a significant drop in either the weight-bearing or the limb-use score compared to the baseline after inoculating a specific number of cancer cells in the femur.

Fig. 2

µCT imaging showing osteolysis induced by the growth of LL/2 cells in the femurs of C57BL/6 mice. The tumor growth causes bone degradation in the ipsilateral femur (A), leaving the contralateral femur (B) unaffected. Images were obtained at the end of the study when mice reached limb use score zero.

Fig. 3

An investigation of the potential significance of Na_V1.7 channels, Na_V1.8 channels, and acid-sensing ion channels as targets to treat CIBP. A) Conditional Na_V1.7 knockout mice in the DRGs showed a modest reduction in pain phenotype associated with CIBP. (A (i)) A comparison of the limb-use scores between the conditional Na_V1.7 knockout mice (Na_V1.7 cKO) and their littermates in the CIBP model. Results indicate that the decline in the use of the affected limb in the Na_V1.7 cKO mice was significantly slower than that of the control group (the REML, p-value = 0.0142). (A (ii)) Ipsilateral weight-bearing after the CIBP surgery in Na_V1.7 cKO mice and their littermates. The REML analysis revealed that the difference in the weight-bearing between the two groups was statistically significant (p-value = 0.0063), with the Na_V1.7 conditional knockout mice showing a significantly less reduction in the percentage of weight put on the ipsilateral limb over time compared to their littermates. No significant differences were detected between genotypes at any individual timepoints for the weight-bearing and the limb-use scoring test. N = 9 in the Na_V1.7 cKO group (five females and four males) and 14 for their littermates (seven males and seven females). B) Knocking out the expression of Na_V1.8 channels congenitally did not cause a statistically significant alteration in the pain phenotype associated with CIBP when assessed using limb-use score (B (i)) and weight-bearing (B (ii)). N = 8 for littermates (two males and six females) and 12 for Na_V1.8 knockout mice (Na_V1.8 KO, six males and six females). P-value for the genotype effect was 0.2715 for the limb-use scoring test and 0.4728 for the weight-bearing test. C) The intrathecal administration of mambalgin-1 (34 µM) did not alter the limb-use score (C (i)) or weight-bearing of mice with CIBP significantly (C (ii)). Behavioral tests were performed 30 min after mambalgin-1 or phosphate-buffered saline (PBS) administration. N = 10 for the mambalgin-1 group (five males and five females) and eight for the PBS group (four males and four females). For Figures A and B, data were analyzed using REML analysis, while for Figure C, data were analyzed using the two-way ANOVA test with multiple comparisons.

Fig. 4

Attempts to silence µ-opioid receptor-expressing neurons in the spinal cord using modified botulinum compounds failed to reduce pain-like behavior in a mouse model of CIBP. Intrathecal and subcutaneous morphine administration resulted in significant analgesia. A) Modified botulinum compounds to silence the µ-opioid receptor-expressing neurons failed to reduce pain-like behavior in our mouse model of CIBP. Mice were assessed for recovery from surgery on day 7 after the surgery by measuring the limb-use score (shown as BL on graph (i)) and no mice showed signs of limping on that day. After that, the limb-use was scored daily using the standard limb-use scoring system. If a mouse reached limb-use score ∼3 (slight limping), it was injected intrathecally with either 100 ng of dermorphine botulinum (Derm-BOT, shown in blue) or the molar-equivalent amount of the control (un-conjugated BOT lacking dermorphine, shown in pink). Following the intrathecal treatment, several behavioral tests were performed, including the limb-use score (i) and the static weight-bearing (ii) tests. (A (i)) The comparison of the average limb-use score between the two treatment groups over time highlighted no significant difference between the two groups when assessed using the REML analyses (p-value for the treatment effect was 0.7654). (A (ii)) The comparison of the average percentage of the weight-borne on the affected limb between the two treatment groups over time indicated no significant difference between them when assessed using the REML analyses (p-value for the treatment effect was 0.4330). n = 10 for each group at the baseline (five males and five females). B) Subcutaneous morphine (15 mg/kg) improved limb-use scores (i) and weight-bearing (ii) of mice with CIBP. Morphine was administered when mice reached limb-use score 3. Results were analyzed using repeated-measures two-way ANOVA test with multiple comparisons. N = 5 (three males and two females) for the morphine group and six for the PBS group (three males and three females). C) Intrathecal morphine improved limb-use scores (i) and weight-bearing (ii) of mice with CIBP. Morphine was administered intrathecally at a dose of 60 ng per mouse when mice reached limb-use score 3. Results were analyzed using repeated-measures two-way ANOVA test with multiple comparisons. N = 3 (two males and one female). P-values for the treatment effect in the two-way ANOVA tests are 0.0087for the limb-use score test and <0.0001for the weight-bearing test. The p-values for the multiple comparisons are shown in the figure. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 5

The dual inhibition of NGF and TNFα reduced heat hypersensitivity and pain-like behaviors associated with CIBP. A) Dual treatment using Etanercept and MEDI578 significantly increased the time needed to reach limb-use score zero of mice after CIBP surgery. The median time needed to reach limb-use score zero for the mice that received anti-NGF (MEDI578) + Etanercept (a protein that inhibits the binding of TNF to its receptors) was 17 days after treatment, while it was 12 days for the control group ((Gehan-Breslow-Wilcoxon test (p-value = 0.0026) and Log-rank Mantel-Cox test (p-value = 0.0357)). B) Antagonising the effects of NGF, TNF, or both slowed down the reduction in the weight put on the ipsilateral limb after CIBP. Compared to the control group, the etanercept-treated group had significantly higher mean weight-bearing results (the REML, 0.0113). Similarly, the mice treated with MEDI578 showed a significantly smaller drop in the weight-bearing (the REML, p-value < 0.0001). The combination of MEDI578 and Etanercept also had a higher fraction of weight put on the ipsilateral limb compared to the controls (p-value=<0.0001). Mice treated with the combination of Etanercept and MEDI578 had a statistically significant higher weight-bearing performance than the ones treated with Etanercept alone (p-value = 0.0098, the REML), but not compared to the mice treated with MEDI578 alone (p-value = 0.4159). C) Blocking the binding of NGF, TNF, or both to their receptors slowed down the reduction in the limb-use score associated with CIBP. Compared to the control group, the etanercept-treated group had a significantly higher mean limb-use score (the REML, p-value = 0.0024). Similarly, mice treated with MEDI578 showed better mean limb-use scores compared to control mice (the REML, p-value = 0.0029). The combination of MEDI578 and Etanercept also led to mice having better use of the affected limb compared to the controls (p-value < 0.0001). Mice treated with Etanercept and MEDI578 did not significantly improve the limb-use score statistically compared to those treated with MEDI578 alone (p-value = 0.1673, the mixed-effects model) or Etanercept alone (p-value = 0.0693). D) The concurrent application of MEDI578 and Etanercept prevented the development of secondary cutaneous heat hyperalgesia associated with CIBP. The application of MEDI578 significantly diminished heat hyperalgesia associated with CIBP compared to the control group (p-value < 0.0001). Similarly, Etanercept significantly lessened heat hyperalgesia associated with CIBP compared to the control group (p-value < 0.0001). The simultaneous application of MEDI578 and Etanercept prevented the development of secondary cutaneous heat hyperalgesia entirely and had a significantly improved effect compared to Etanercept alone (p-value = 0.0097) and MEDI578 alone (p-value = 0.0137). The asterisks of significance shown in the figure represent the post hoc analysis results comparing each treatment with the control group at each time point. n = 13 in each group at the baseline (except for the control group (n = 23)). Doses: Etanercept (10 mg/kg), MEDI578 (3 mg/kg), and control antibody (3 mg/kg). The drugs were injected intraperitoneally on day ten post-surgery. Control antibody (n = 10, five males and five females), PBS (n = 13, six males and seven females), MEDI578 (n = 13, six females and seven males), etanercept (n = 13, six males and seven females), and the combination (n = 13, six females and seven males).