Senolytic treatment alleviates doxorubicin-induced chemobrain

Abstract

Doxorubicin (Dox), a widely used treatment for cancer, can result in chemotherapy-induced cognitive impairments (chemobrain). Chemobrain is associated with inflammation and oxidative stress similar to aging. As such, Dox treatment has also been used as a model of aging. However, it is unclear if Dox induces brain changes similar to that observed during aging since Dox does not readily enter the brain. Rather, the mechanism for chemobrain likely involves the induction of peripheral cellular senescence and the release of senescence-associated secretory phenotype (SASP) factors and these SASP factors can enter the brain to disrupt cognition. We examined the effect of Dox on peripheral and brain markers of aging and cognition. In addition, we employed the senolytic, ABT-263, which also has limited access to the brain. The results indicate that plasma SASP factors enter the brain, activating microglia, increasing oxidative stress, and altering gene transcription. In turn, the synaptic function required for memory was reduced in response to altered redox signaling. ABT-263 prevented or limited most of the Dox-induced effects. The results emphasize a link between cognitive decline and the release of SASP factors from peripheral senescent cells and indicate some differences as well as similarities between advanced age and Dox treatment.

Keywords: chemobrain; cognition; inflammation; oxidative stress; senolytic NMDA receptor; transcription.

FIGURE 1

Dox treatment increased and ABT‐263 treatment decreased expression of senescent and SASP genes in the periphery. (a) Time course of treatment, starting at 6 months of age. For the two groups receiving Dox treatment, 2 mg/kg dose of Dox was administered by injection once a week (arrows) for 4 weeks, while the group receiving the senolytic treatment received a 12 mg/kg dose of ABT‐263 by oral gavage for 5 consecutive days every other week (Weeks 2 and 4, green bars). The vehicle group also received oral gavage of vehicle for 5 consecutive days every other week. Starting 2 weeks after the last dose of Dox and 1 week after the last ABT‐263 treatment, animals were behaviorally characterized (n = 22 for Veh, n = 13 for Dox and n = 14 for Dox + ABT). (b) Change in body weight over the course of the study (n = 14 for Veh, n = 13 for Dox and n = 14 for Dox + ABT). (c) Grip strength normalized to weight (n = 14 for Veh, n = 13 for Dox and n = 14 for Dox + ABT). For a–c, data represented as mean ± SEM; a = significantly different (p < 0.05) from Veh and b = significantly different (p < 0.05) from Dox + ABT. Relative to Veh and Dox + ABT groups, treatment with Dox alone increased the expression level of Cdkn2a in (d) lung, (e) bone marrow, (f) liver, (g) kidney, and (h) spleen. Relative to Veh and Dox + ABT groups, treatment with Dox alone increased the expression level of (i) Cdkn1a, (j) IL‐6, (k) Mmp3, and (l) Tnfsf11 in spleen. Error bars denote SEM (n = 9 per group). a = significantly different (p < 0.05) from Veh, b = significantly different (p < 0.05) from Dox + ABT.

FIGURE 2

ABT‐263 treatment ameliorated Dox‐induced impairments in cognition. (a, d) latency, (b, e) swim speed and (c, f) escape distance for each training block for the cue discrimination a–c and spatial discrimination d–f tasks. The data are expressed as Mean ± SEM. Discrimination index (DI score) for the (g) acquisition and (h) retention probe trials on the watermaze. (i, j) Same data as g–h, except that data for the Dox group were limited to those that performed above chance (DI score >0) during acquisition testing. (k–m) Group mean heat maps for the animal's position during the acquisition and retention probe trials. (n, o) Box plots for latency to enter the dark compartment of inhibitory avoidance during (n) Day 1 training and (o) Day 2 retention testing (n = 22 for Veh, n = 13 for Dox and n = 14 for Dox + ABT); Data represented as mean ± SEM; a = significantly different (p < 0.05) from Veh, b = significantly different (p < 0.05) from Dox + ABT. Asterisks indicate a significant difference (p < 0.05) from chance performance (g–j).

FIGURE 3

ABT‐263 treatment preserved synaptic function in the hippocampus. (a) Examples for the total synaptic responses for stimulation intensity at 12 and 20 V for the three groups. (b) Input/output curve of the mean ± SEM slope of total excitatory postsynaptic field potentials recorded from hippocampal CA3‐CA1 synapses. (c) Examples for the NMDAR synaptic responses for stimulation intensity at 16 and 20 V for the three groups. (d) Input/output curve of the mean ± SEM slope of NMDAR synaptic field potentials recorded from hippocampal CA3‐CA1 synapses. (e) Examples for the NMDAR synaptic response before (1) and 1 h after (2) application of DTT. (f) Time course of the growth of the NMDAR‐mediated component of the EPSP during application of DTT (bar). Each point represents the mean relative to the baseline. (g) Mean ± SEM growth of the NMDAR‐mediated component of the EPSP, 1 h after application of DTT. Data are represented as mean ± SEM. n = 10/5 slices/animals for Dox; n = 8/4 slices/animals for Veh and n = 7/4 slices/animals for Dox + ABT. a = significantly different (p < 0.05) from Veh.

FIGURE 4

Effects of treatment on DG gene expression. Heatmap of expression patterns of DG genes under conditions of Veh, Dox, and Dox + ABT. (a) Heatmap of 37 genes that increased expression (p < 0.01) in Dox relative to Veh groups and were linked to lipopolysaccharide response, immune effector response or cellular senescence. (b) Heatmap of 41 synaptic component genes that decreased expression (p < 0.01) in Dox relative to Veh groups. Data are represented as z‐scores, normalized relative to the average expression across all animals (n = 10 for Veh; n = 8 for Dox and n = 8 for Dox + ABT).

FIGURE 5

Dox‐induced microglial activation was reduced by ABT‐263. (a) Representative cortex sections immunohistologically stained for Iba‐1. Mean ± SEM for microglial morphological parameters (b) soma size, (c) processes per cell, (d) branches per cell, and (e) the average length of the process. (f) Representative cortex sections immunohistologically stained for albumin. (g) Bars represent mean ± SEM of proportion of cells stained positive for albumin. n = 6 per group; a = significantly different (p < 0.05) from Veh, b = significantly different (p < 0.05) from Dox + ABT.