Genome-wide identification and functional analyses of microRNA signatures associated with cancer pain

Abstract

Cancer pain remains a major challenge and there is an urgent demand for the development of specific mechanism-based therapies. Various diseases are associated with unique signatures of expression of microRNAs (miRNAs), which reveal deep insights into disease pathology. Using a comprehensive approach combining genome-wide miRNA screening, molecular and in silico analyses with behavioural approaches in a clinically relevant model of metastatic bone-cancer pain in mice, we now show that tumour-induced conditions are associated with a marked dysregulation of 57 miRNAs in sensory neurons corresponding to tumour-affected areas. By establishing protocols for interference with disease-induced miRNA dysregulation in peripheral sensory neurons in vivo, we functionally validate six dysregulated miRNAs as significant modulators of tumour-associated hypersensitivity. In silico analyses revealed that their predicted targets include key pain-related genes and we identified Clcn3, a gene encoding a chloride channel, as a key miRNA target in sensory neurons, which is functionally important in tumour-induced nociceptive hypersensitivity in vivo. Our results provide new insights into endogenous gene regulatory mechanisms in cancer pain and open up attractive and viable therapeutic options.

Keywords: Clcn3; bone metastatic pain; gene regulation; miRNA inhibitors; miRNA mimics.

Figure 1. Up- or down-regulation of microRNAs (miRNAs) in sensory neurons of the dorsal root ganglia (DRG) in a model of bone metastases pain

1. A. Increase in frequency of paw withdrawal to plantar application of a von Frey filament force of 0.07 g following induction of tumor growth in the calcaneous bone of the heel in mice as compared to sham surgery. * denotes p = 0.002 on PID-5, 6, 7 and <0.0001 from PID-8 through 15 as compared to basal and † denotes p < 0.001 on PID-5 and <0.0001 from PID-6 through 15 as compared to corresponding data point in the sham group, two-way ANOVA of repeated measures followed by Bonferroni's multiple comparisons post hoc test, n = at least 6 mice per group.

2. B. Mechanical response threshold calculated as von Frey filament strength required to achieve 50% withdrawal frequency. * denotes p < 0.001 from PID-4 through 15 as compared to basal and † denotes p = 0.004 on PID-5, 6 & 13, 0.006 on PID-7, 9 & 11, 0.005 on PID-8, 0.004 on PID-10, 0.0001 on PID-12 & 14, and 0.003 on PID-15 as compared to corresponding data point in the sham group, two-way ANOVA of repeated measures followed by Bonferroni's multiple comparisons post hoc test, n = at least 6 mice per group.

3. C,D. Heat maps of miRNAs found to be significantly up- or downregulated via microarray analysis in the ipsilateral lumbar DRG of tumor-bearing mice 4 days (C) or 8 days (D) post implantation as compared to sham surgery. Scale indicates expression intensities obtained from the microarray experiment.

4. E. Representation of examples of miRNAs showing up- or down-regulation following independent verification with quantitative RT-PCR analyses (left hand panel) and the original data from microarray analysis. *p = 0.001 for miR-544-3p, 0.003 for miR-1a-3p, 0.009 for miR-34c-5p, 0.04 for miR-370-3p, 0.03 for miR-291b-5p and 0.005 for miR-483-3p as compared to sham-treated group, ANOVA followed by post hoc Fischer's test, n = 3 mice per group.

Figure 2. Manipulation of miRNA expression in lumbar DRGs in mice in vivo via intrathecal application of miRNA mimics and inhibitors

1. Experimental scheme established in this study which enables effective knockdown/induction of miRNA expression in DRGs in vivo and analysis of tumour pain-associated behaviours.

2. Microscopic analyses of whole-mount DRGs or cryosections showing uptake of FAM-conjugated miRNA inhibitors. Scale bar is 50 µm in all panels.

3. Typical examples of qRT-PCR verification of efficacy of miRNA inhibitors in reversing tumour-induced upregulation of miRNAs in ipsilateral DRGs in vivo.

4. Typical examples of qRT-PCR verification of efficacy of miRNA mimics in reversing tumour-induced downregulation of miRNAs and inducing overexpression of miRNAs in ipsilateral DRGs in vivo. In panel (C), * denotes p = 0.02 for miR-1a-3p, 0.04 for miR-34c-3p as compared to corresponding mismatch inhibitor and in panel (D), * denotes p = 0.001 for miR-370-3p and <0.0001 for miR-291b-5p as compared to non-targeting mimic, ANOVA followed by post hoc Fischer's test, n = 3 per group.

Figure 3. Functional validation of miRNAs upregulated in DRGs in tumor bearing mice with respect to tumor-induced mechanical hypersensitivity

1. Change in frequency of paw withdrawal to plantar application of a von Frey filament force of 0.07 g following induction of tumor growth in the calcaneous bone of the heel in mice receiving intrathecally delivered miR-1a-3p inhibitor (red symbols) or the corresponding mismatch inhibitors (green symbols) or vehicle (grey symbols). * denotes p < 0.0001 on PID-5 through 15 in the vehicle, mismatch inhibitor and miR-1a-3p inhibitor groups as compared to basal; † denotes p = 0.007 on PID-10 and <0.0001 on PID-12 & 14 as compared to corresponding data point in the mismatch inhibitor group; ‡ denotes p < 0.0001 on PID-10, 12 & 14 as compared to corresponding data point in the vehicle group.

2. Mechanical response thresholds calculated as von Frey filament strength required to achieve 50% withdrawal frequency following induction of tumor growth in the calcaneous bone of the heel in mice receiving intrathecally delivered miR-1a-3p inhibitor (red bars) or the corresponding mismatch inhibitors (green bars) or vehicle (grey bars). * denotes p < 0.0001 from PID-5 through 14 in the vehicle and mismatch inhibitor groups and on PID-10, 12, 14 in the miR-1a-3p-inhibitor groupas compared to basal; † denotes p = 0.004 on PID-5, 0.003 on PID-7, 0.05 on PID-10 and 0.01 on PID-12 as compared to corresponding data point in the mismatch inhibitor group; ‡ denotes p < 0.0001 on PID-5, 0.0002 on PID-7, 0.05 on PID-10, and 0.004 on PID-12 as compared to corresponding data point in the vehicle group.

3. Change in frequency of paw withdrawal to plantar application of a von Frey filament force of 0.07 g following induction of tumor growth in the calcaneous bone of the heel in mice receiving intrathecally delivered miR-34c-5p inhibitor (red symbols) or the corresponding mismatch inhibitor (green symbols) or vehicle (grey symbols). * denotes p < 0.0001 on PID-5 through 14 in the vehicle and mismatch inhibitor groups, 0.006 on PID-5 and < 0.0001 on PID-14 for miR-34c-5p- inhibitor group as compared to basal; † denotes p < 0.0001 on PID-7, 9 & 11 as compared to corresponding data point in the mismatch inhibitor group; ‡ denotes p < 0.0001 on PID-7, 9 & 11 as compared to corresponding data point in the vehicle group.

4. Mechanical response thresholds calculated as von Frey filament strength required to achieve 50% withdrawal frequency following induction of tumor growth in the calcaneous bone of the heel in mice receiving intrathecally delivered miR-34c-5p inhibitor (red bars) or the corresponding mismatch inhibitors (green bars) or vehicle (grey bars). * denotes p = 0.0003 on PID-7 & 9, < 0.0001 on PID-11 & 14 in the mismatch-inhibitor group; 0.0018 on PID-5 and < 0.0001 on PID-14 in the miR-34c-5p-inhibitor group and 0.0001 on PID-5 through 15 in the vehicle group; † denotes p < 0.0001 on PID-7, 0.0009 on PID-9, 0.05 on PID-11 as compared to corresponding data point in the mismatch inhibitor group; and ‡ denotes p < 0.0001 on PID-7 & 9, 0.0518 on PID-11 as compared to corresponding data point in the vehicle group.

5. Change in frequency of paw withdrawal to plantar application of a von Frey filament force of 0.07 g following induction of tumor growth in the calcaneous bone of the heel in mice receiving intrathecally delivered miR-544-3p inhibitor (red symbols) or the corresponding mismatch inhibitors (green symbols) or vehicle (grey symbols). * denotes p < 0.0001 from PID-3 through PID-14 in vehicle, mismatch-inhibitor and miR-544-3p inhibitor groups.

6. Mechanical response thresholds calculated as von Frey filament strength required to achieve 50% withdrawal frequency following induction of tumor growth in the calcaneous bone of the heel in mice receiving intrathecally delivered miR-544-3p inhibitor (red bars) or the corresponding mismatch inhibitors (green bars) or vehicle (grey bars). * denotes p < 0.0001 from PID-3 through PID-14 in mismatch-inhibitor and miR-544-3p inhibitor groups.

In all panels statistical significance was tested by two-way ANOVA of repeated measures followed by Bonferroni's multiple comparisons post hoc test, n = at least 6 mice per group. The experimental scheme employed is the same as described in Fig. 2A. The square box represents the time-course of miRNA-inhibitor or mismatch-inhibitor or vehicle application.

Figure 4. Functional validation of miRNAs downregulated in DRGs in tumor bearing mice with respect to tumor-induced mechanical hypersensitivity

1. Change in frequency of paw withdrawal to plantar application of a von Frey filament force of 0.02 g following induction of tumor growth in the calcaneous bone of the heel in mice receiving intrathecally delivered miR-370-3p mimic (red symbols) or non-targeting mimic (green) or vehicle (grey symbols). * denotes p = 0.0192 on PID-13 in the vehicle group and <0.0001 on PID-8 through 15 in the miR-370-3p-mimic group as compared to basal; † denotes p = 0.0013 on PID-8 and <0.001 on PID-10, 13 & 15 as compared to corresponding data point in the non-targeting mimic group; ‡ denotes p = 0.0003 on PID-8, <0.0001 on PID-10 & 15 and 0.0044 on PID-13 as compared to corresponding data point in the vehicle group.

2. Mechanical response thresholds calculated as von Frey filament strength required to achieve 80% withdrawal frequency following induction of tumor growth in the calcaneous bone of the heel in mice receiving intrathecally deliveredmiR-370-3p mimic (red bars) or non-targeting mimic (green bars) or vehicle (grey bars). * denotes p = 0.0291 on PID-8, 0.05 on PID-10, 0.0192 on PID-13 and 0.05 on PID-15 in the vehicle group, 0.04 on PID-8, 0.002 on PID-10, 0.05 on PID-13 & 15 in the non-targeting mimic group; † denotes p = 0.0013 on PID-3, <0.0001 on PID-6, 8 & 10 as compared to corresponding data point in the non-targeting mimic group; ‡ denotes p = 0.0013 on PID-6 and <0.0001 on PID-8, 10 & 13 as compared to corresponding data point in the vehicle group.

3. Change in frequency of paw withdrawal to plantar application of a von Frey filament force of 0.07 g following induction of tumor growth in the calcaneous bone of the heel in mice receiving intrathecally delivered miR-483-3p mimic (red symbols) or non-targeting mimic (green) or vehicle (grey symbols). * denotes p = 0.0008 on PID-6, 0.0002 on PID-8 & <0.0001 on PID-10, 13, 15 in vehicle, non-targeting mimic and miR-483-3p mimic groups as compared to basal; † denotes p = 0.0206 on PID-10 & 0.0074 on PID-13 as compared to corresponding data point in the non-targeting mimic group; and ‡ denotes p = 0.0074 on PID-13 as compared to corresponding data point in the vehicle group.

4. Mechanical response thresholds calculated as von Frey filament strength required to achieve 50% withdrawal frequency following induction of tumor growth in the calcaneous bone of the heel in mice receiving intrathecally deliveredmiR-483-3p mimic (red bars) or non-targeting mimic (green bars) or vehicle (grey bars). * denotes p < 0.0001 from PID-3 through 15 in vehicle, non-targeting-mimic groups and miR-483-3p-mimic groups as compared to basal; † denotes p = 0.005 on PID-10 & 0.05 on PID-13 as compared to corresponding data point in the non-targeting mimic group; and ‡ denotes p = 0.031 on PID-10, 0.022 on PID-13 as compared to corresponding data point in the vehicle group.

5. Change in frequency of paw withdrawal to plantar application of a von Frey filament force of 0.07 g following induction of tumor growth in the calcaneous bone of the heel in mice receiving intrathecally delivered miR-291b-5p mimic (red symbols) or non-targeting mimic (green) or vehicle (grey symbols). * denotes p < 0.0001 from PID-3 through PID-14 in vehicle, non-targeting-mimic and miR-291b-5p-mimic groups.

6. Mechanical response thresholds calculated as von Frey filament strength required to achieve 50% withdrawal frequency following induction of tumor growth in the calcaneous bone of the heel in mice receiving intrathecally deliveredmiR-291b-5p mimic (red bars) or non-targeting mimic (green bars) or vehicle (grey bars). * denotes p < 0.0001 from PID-3 through PID-14 in vehicle, non-targeting-mimic and miR-291b-5p-mimic groups.

In all panels, statistical significance was tested by two-way ANOVA of repeated measures followed by Bonferroni's multiple comparisons post hoc test, n = at least 6 mice per group. The experimental scheme employed is the same as described in Fig. 2A. The square box represents the time-course of miRNA-mimic or non-targeting-mimic or vehicle application.

Figure 5. Analysis of miR-1a-3p target genes in the context of tumor-induced hypersensitivity in the bone metastatic pain model

1. QRT-PCR analysis demonstrating induction of predicted miR-1a-3p target genes in lumbar DRGs in vivo following suppression of miR-1a-3p expression via inhibitor delivery intrathecally in vivo. * denotes p <:0.0001 as compared to mismatch-inhibitor group.

2. QRT-PCR analysis representing changes in expression of miR-1a-3p target genes in ipsilateral lumbar DRGs in vivo at day 8 in the bone metastases model. * denotes p = 0.041 as compared to sham group.

3. Time course of tumor-induced downregulation of Clcn3 expression in ipsilateral lumbar DRGs via QRT-PCR analysis. * denotes p = 0.002 on PID-4, 0.003 on PID-8, 0.05 on PID-10 and 0.04 on PID-15 as compared to sham group.

4. Time course of tumor-induced changes in miR-1a-3p expression in ipsilateral lumbar DRGs via QRT-PCR analysis. * denotes p = 0.04 on PID-4, 0.002 on PID-8, and 0.04 on PID-10 as compared to sham group.

5. In panels A, B, C &D statistical P value was calculated by ANOVA followed by post hoc Fischer's test, n = 3 mice per group.

6. Immunohistochemical analysis of expression of Clcn3 protein in mouse DRG and colabeling with isolectin-B4-binding (IB4) non-peptidergic nociceptors and peptidergic (CGRP-positive) nociceptors. Scale bars represent 50 µm.

Figure 6. In vitro and in vivo validation of Clcn3 as a miR-1a-3p target gene and its functional contribution to tumor-induced mechanical hypersensitivity in the bone metastatic pain model employed

1. Binding sites for miR-1a-3p (mmu-miR-1) on the 3' untranslated region (UTR) of the mouse Clcn3 gene.

2. Luciferase-reporter based assay in HEK293 cells demonstrating changes in translation of the Clcn3 gene following suppression of miR-1a-3p expression via graded delivery of the specific inhibitor. *p = 0.04, 0.02 and 0.05 for 6, 4 & 2 µg/L miR-1a-3p inhibitor groups respectively as compared to control, Student's t-test, n = 3 independent experiments.

3. Luciferase-reporter based assay in HEK293 cells demonstrating changes in translation of the Clcn3 gene following suppression or induction of miR-1a-3p expression via graded delivery of the specific mimic. *p = 0.03, & 0.02 for 4 & 2 µg/L miR-1a-3p inhibitor groups respectively as compared to control, Student's t-test, n = 3 independent experiments.

4. QRT-PCR analysis demonstrating significant knockdown of Clcn3 expression in lumbar DRG following intrathecal delivery of a siRNA directed against Clcn3 or the corresponding control non-targeting siRNA, *p = 0.001 as compared to naive group, ANOVA followed by post hoc Fischer's test, n = 3 mice per group.

5. Change in frequency of paw withdrawal to plantar application of von Frey filament forces of 0.02 g following induction of tumor growth in the calcaneous bone of the heel in mice receiving intrathecally delivered siRNA directed against Clcn3 (red symbols) or the corresponding non-targeting siRNA (green symbols) or vehicle (grey symbols). * denotes p = 0.0006 on PID-11, and <0.0001 on PID-13 & 15 in the vehicle group, <0.0001 on PID-15 in the non-targeting siRNA group, 0.0037 on PID-5, and <0.0001 on PID-7 through 15 for the Clcn3-siRNA group as compared to basal, † and ‡ denotes p = 0.0202 on PID-5, 0.0011 on PID-7 and <0.0001 on PID-9, 11, 13 & 15 as compared to corresponding data point in the non-targeting siRNA or vehicle groups respectively. The square box represents the time-course of Clcn3-siRNA or non-targeting-siRNA or vehicle application.

6. Mechanical response thresholds calculated as von Frey filament strength required to achieve 80% withdrawal frequency following induction of tumor growth in the calcaneous bone of the heel in mice receiving intrathecally deliveredsiRNA directed against Clcn3 (red bars) or the corresponding non-targeting siRNA (green bars) or vehicle (grey bars). * denotes p = 0.0003 on PID-5 & 9, 0.0233 on PID-7, and <0.0001 on PID-11, 13 & 15 in the vehicle group, 0.0003 on PID-5, 0.002 on PID-7 and <0.0001 on PID-9, 11, 13 & 15 for the non-targeting-siRNA group and <0.0001 from PID-5 through PID-15 for the Clcn3-siRNA group as compared to basal; † denotes p = 0.0122 on PID-5, 0.0024 on PID-7, 0.0127 on PID-9, 0.05 on PID-11 as compared to corresponding data point in the non-targeting-siRNA group and ‡ denotes p = 0.0122 on PID-5, 0.0002 on PID-7, 0.0029 on PID-9, and 0.0401 on PID-11 as compared to corresponding data point in the vehicle group.