Preventive anti-NGF treatment suppresses alterations in functional connectivity imposed by cancer-induced bone pain in mice
David Buehlmann1,2, Giovanna Diletta Ielacqua1, Jael Xandry3, and Markus Rudin1,2,3

1Inst. for Biomedical Engineering, ETH & University of Zurich, Zürich, Switzerland, 2Neuroscience Center Zurich (ZNZ), Zürich, Switzerland, 3Inst. of Pharmacology & Toxicology, University of Zurich, Zürich, Switzerland


The efficacy of an anti-nerve growth factor (NGF) antibody in preventing rearrangements of whole-brain functional connectivity elicited by nociceptive input following bone metastases was evaluated in a mouse model of cancer-induced bone pain using longitudinal resting-state fMRI. ROI-based network and seed-based connectivity analysis approaches revealed major hubs of ascending and descending pain pathways to be affected by the developing pain. Functional rearrangements within these regions could be prevented by prospective application of anti-NGF antibody mAb911 indicating the efficacy of anti-NGF treatment in preventing, or at least delaying, adaptations of the brain circuitry associated with development of a chronic pain state.


Cancer-induced bone pain arises from a primary tumor in the bone or skeletal metastasis of common cancer types such as breast, lung or prostate cancer [4]. Therapies based on bisphosphonates, NSAIDs, radiotherapy as well as centrally active drugs such as opioids are effective in alleviating ongoing pain, but often exert severe side-effects [6]. Antibodies targeting nerve growth factor (NGF) have been shown to effectively relieve neuropathic aspects of cancer-induced bone pain both in mice and humans [2,3,5]. Yet, the impact of NGF inhibition on the central nervous system remains largely unknown. We therefore investigated the effects of anti-NGF treatment on functional connectivity in major ascending and descending pain pathways in a mouse model of cancer-induced bone pain using resting-sate fMRI (rs-fMRI).

Material & Methods

Mouse models of chronic pain from bone cancer were prepared by intramedullar injection of EO771 (tebu‐bio, Le-Perray-en-Yvelines, France) breast cancer cells into the tibia of female C57BL/6 (n=8). Functional changes were assessed longitudinally using a Biospec 94/30 small animal MR system (Bruker, Karlsruhe, Germany) with a cryogenic quadrature transmit/receive surface coil for excitation and signal reception. Rs-fMRI data was acquired using a gradient‐echo echo‐planar imaging (GE‐EPI) sequence: TR=1000ms, TE=12ms, FA=60°, number of averages=1, 900 repetitions. Eighteen consecutive slices of 0.5mm thickness with an in-plane resolution of 200mmx200mm have been recorded. Animals were anesthetized using medetomidine hydrochloride (0.05mg/kg as bolus, 0.1mg/kg/h maintenance, i.v.), intubated and ventilated with low dose isoflurane (0.5%) in 20%O2/80%air mixture at 80breaths/min. For immobilization, pancuronium bromide was administered s.c. in a bolus of 0.5mg/kg. Behavioral readouts of pain were assessed in terms of time spent guarding and number of flinches in a 2min period. Murine monoclonal anti-NGF antibody mAb911 (10mg/kg, i.p., Pfizer Inc.) was administered prospectively. Rs-fMRI data was analyzed using a ROI-based network analysis and seed-based connectivity. To test for differences in evolution of functional connectivity between groups a linear mixed model analysis was used, testing for significant group and session interactions. Four groups have been included: Tumor bearing animals receiving either mAb911 or vehicle (‘Tumor+antiNGF’, ‘Tumor+Vehicle’) as well as sham-operated mice receiving either of the treatments (‘Sham+anti-NGF’, ‘Sham+Vehicle).


ROI-based analysis of longitudinal rs-fMRI data revealed distinct rearrangement of functional connectivity (FC) patterns in Tumor+Vehicle as compared to Sham+Vehicle animals (Fig.2). ROI-based network analysis revealed interactions of amygdallar nuclei with thalamic as well as midbrain regions to be most significantly affected. In addition, FCs between amygdallar nuclei and cortical areas were found to be altered, in particular interactions involving the temporal associative/insular regions and the cingulate cortex. Furthermore, FC between the thalamus and cortical areas were specifically affected in somatosensory and temporal associative/insular regions. Less distinct changes were observed between striatal and midbrain nuclei as well as between midbrain and thalamic regions. Prospective treatment of tumor-bearing mice using mAb911 successfully prevented these network changes induced by persistent peripheral input. Seed-based connectivity analysis revealed significant effects on FC between the amygdallar seed contralateral to the tumor site and bilateral thalamic (p=0.004), dorsal hippocampal (p=0.006) and somatosensory (p=0.021) resting state networks (RSN) (Fig.3). RSNs constitute bilateral group-level independent components, extracted from naïve mice using independent component analysis (MELODIC [1]). Furthermore, FC between the anterior periaqueductal gray seed and barrelfield 2 cortex (p=0.032), posterior periaqueductal gray seed and limb cortex (p=0.032) as well as ipsilateral motor related superior collicular seed and the amygdala (p=0.035) RSNs were found to be significantly affected in the Tumor+Vehicle group. These changes in FC could be prevented by treatment with anti-NGF antibodies.


Our study revealed FC alterations involving the major hubs of ascending and descending pain pathways in a mouse model of persistent pain from bone metastasis. We found profound alterations affecting connections form the contralateral amygdallar nuclei to cortical somatosensory (B1c, B2c, Lc) and limbic structures such as cingulate, temporal associative and insular cortex and midbrain. Seed-based analysis supported these findings and furthermore indicated altered connectivities between the anterior and posterior periaqueductal gray and cortical structures. Cancer pain associated FC alterations could be successfully prevented through the preventive administration of anti-NGF antibody mAb911 with only few residual interactions. Seed-based analysis indicated no changes in connectivity strength as a function of time in Tumor+anti-NGF mice as compared to Sham+Vehicle animals throughout the observation period, indicating efficacy of anti-NGF treatment in preventing FC rearrangements in response to cancer-induced bone pain.


Using longitudinal rs-fMRI readouts in a mouse model of cancer-induced bone pain we were able to show efficacy of anti-NGF treatment in preventing pain-induced alterations in FC involving major nuclei of ascending and descending pain pathways. Monitoring the adaptation of the brain circuitry in such models may yield information on mechanisms contributing to the chronification of pain during persistent bone cancer-induced nociceptive input.


No acknowledgement found.


[1] Beckmann CF, Smith SM. Probabilistic Independent Component Analysis for Functional Magnetic Resonance Imaging. IEEE Trans. Med. Imaging 2004;23:137–152. doi:10.1109/TMI.2003.822821.

[2] Cattaneo A. Tanezumab, a recombinant humanized mAb against nerve growth factor for the treatment of acute and chronic pain. Curr. Opin. Mol. Ther. 2010;12:94–106. Available: http://www.ncbi.nlm.nih.gov/pubmed/20140821. Accessed 6 Oct 2017.

[3] Jimenez-Andrade JM, Ghilardi JR, Castañeda-Corral G, Kuskowski MA, Mantyh PW. Preventive or late administration of anti-NGF therapy attenuates tumor-induced nerve sprouting, neuroma formation, and cancer pain. Pain 2011;152:2564–2574. doi:10.1016/j.pain.2011.07.020.

[4] Mantyh P. Bone cancer pain: Causes, consequences, and therapeutic opportunities. Pain.2013, Vol. 154.

[5] Sevcik MA, Ghilardi JR, Peters CM, Lindsay TH, Halvorson KG, Jonas BM, Kubota K, Kuskowski MA, Boustany L, Shelton DL, Mantyh PW. Anti-NGF therapy profoundly reduces bone cancer pain and the accompanying increase in markers of peripheral and central sensitization. Pain 2005;115:128–141. doi:10.1016/j.pain.2005.02.022.

[6] World Health Organization. Cancer pain relief : with a guide to opioid availability. World Health Organization, 1996 p.


Behavioral readouts of spontaneous pain behavior. Guarding and flinching was assessed during a period of 2min. Statistical significance was tested per timepoint compared to Sham+Vehicle using t-tests accounting for multiple comparisons (Holm-Sidak method). Brackets indicate significant treatment-effects, comparing Tumor+anti-NGF to Tumor+Vehicle (* p<0.05, ** p<0.01 and *** p<0.001)

Longitudinal Roi-based network analysis indicating significant differences between groups. A) Upper triangle indicates significant differences between Tumor+Vehicle and Sham+Vehicle animals. The lower triangle represents the comparison between Tumor+anti-NGF and Sham+Vehicle animals indicating only little differences between the two groups. B) Upper triangle comparing Tumor+Vehicle and Tumor+anti-NGF animals, indicating significant treatment effects. Lower triangle indicates effect of treatment comparing Sham+Vehicle and Sham+anti-NGF animals.

Seed based correlations were analyzed in regions of interest, plotting Z-scores as group averages. Projections of the left amygdallar seed (L Amg) to the thalamus (Th), dorsal hippocampus (dHp) and the barrelfield 1 cortex (B1c) indicate FC changes in Tumor+Vehicle animals, while other groups were not altered. Furthermore, projections of the right superior colliculus seed (R SCm) to the amygdalar RNS, anterior periaqueductal gray (aPAG) seed to the barrelfield 2 cortex (B2c) RNS and posterior peraqueductal gray (pPAG) seed to the limb cortex RSN were found to be altered in Tumor+Vehicle animals only (* p<0.05, ** p<0.01).

Proc. Intl. Soc. Mag. Reson. Med. 26 (2018)