Early-life adversity increases morphine tolerance and persistent inflammatory hypersensitivity through upregulation of δ opioid receptors in mice

Abstract

Exposure to severely stressful events during childhood is associated with poor health outcomes in later life, including chronic pain and substance use disorder. However, the mediators and mechanisms are unclear. We investigated the impact of a well-characterized mouse model of early-life adversity, fragmented maternal care (FC) between postnatal day 2 and 9, on nociception, inflammatory hypersensitivity, and responses to morphine. Male and female mice exposed to FC exhibited prolonged basal thermal withdrawal latencies and decreased mechanical sensitivity. In addition, morphine had reduced potency in mice exposed to FC and their development of tolerance to morphine was accelerated. Quantitative PCR analysis in several brain regions and the spinal cords of juvenile and adult mice revealed an impact of FC on the expression of genes encoding opioid peptide precursors and their receptors. These changes included enhanced abundance of δ opioid receptor transcript in the spinal cord. Acute inflammatory hypersensitivity (induced by hind paw administration of complete Freund's adjuvant) was unaffected by exposure to FC. However, after an initial recovery of mechanical hypersensitivity, there was a reappearance in mice exposed to FC by day 15, which was not seen in control mice. Changes in nociception, morphine responses, and hypersensitivity associated with FC were apparent in males and females but were absent from mice lacking δ receptors or β-arrestin2. These findings suggest that exposure to early-life adversity in mice enhances δ receptor expression leading to decreased basal sensitivity to noxious stimuli coupled with accelerated morphine tolerance and enhanced vulnerability to persistent inflammatory hypersensitivity.

Figure 1.

Fragmented care reduces acute thermal and mechanical nociception in wild-type male and female C57BL/6 mice. (A) Dams housed in fragmented care (FC) cages made significantly more nest departures compared with dams housed in control cages on postnatal day (PD) 3 to 8 (cage condition: F_1,100 = 52.5, P < 0.0001; time F_5,100 = 1.7, P = 0.15; interaction: F_5,100 = 0.89, P = 0.49). (B) The time each dam spent interacting with their pups (expressed as a percentage of the total interaction time on each day) was similar between control and FC cages (cage condition: F_1,100 = 0.05, P = 0.82; time F_5,100 = 0.88, P = 0.5; interaction: F_5,100 = 1.9, P = 0.1). (C) Mice exposed to FC during PD 2 to 9 had significantly prolonged tail withdrawal latencies from warm water (48°C) at PD 60 (cage condition: F_1,27 = 8.1, P = 0.01; sex: F = 3.3, P = 0.08; interaction: F_1,27 = 0.01, P = 0.93). (D) A separate cohort of mice exposed to FC during PD 2 to 9 had significantly enhanced mechanical thresholds on PD 60 (cage condition: F_1,22 = 6.9, P = 0.02; sex: F = 1.6, P = 0.22; interaction: F = 0.04, P = 0.85). Statistical comparisons were performed using a 2-way ANOVA with Bonferroni correction. Data shown as mean ± SEM (bars) in all cases. Individual responses from each mouse are shown as separate datapoints. (A and B) n = 10 (control) and n = 12 (FC) separate cages. (C and D) n (control, then FC) = 14 and 13 (WT male), n = 15 and 15 (WT female). * P < 0.05 compared with control. ANOVA, analysis of variance; WT, wild-type.

Figure 2.

Fragmented care diminishes morphine antinociception and accelerates the development of tolerance in WT mice. Morphine antinociception in WT C57BL/6 control and fragmented care (FC) mice was assessed on postnatal day (PD) 60 using the warm-water (48°C) tail withdrawal assay. (A) Morphine cumulative dose–response relationship in control and FC mice on day 1. (B) Morphine (10 mg/kg) antinociception continued to be assessed using the tail withdrawal assay for the following 8 consecutive days. Separate Friedman tests comparing the maximum possible effect (%MPE) of morphine antinociception to day 1 within each group revealed a significant reduction in morphine antinociception after 8 days of injections in control mice (χ²(9) = 94.1, P < 0.0001). By contrast, morphine antinociception was significantly reduced in FC after just 5 days of injections (χ²(9) = 92.9, P < 0.0001). (C) Morphine cumulative dose–response relationships in control and FC mice on day 10. Estimates of morphine ED₅₀ in control and FC mice on day 1 and on day 10 are summarised in Table 1. Data shown as mean ± SEM in all cases. Control: n = 16 (comprised 8 males and 8 females), and FC: n = 15 (comprised 7 males and 8 females). *P < 0.05 compared with day 1. WT, wild-type.

Figure 3.

Fragmented care alters the expression of opioid pathway genes in the spinal cord of juvenile and adult C57BL/6 mice. Quantitative PCR was used to measure the expression of transcripts involved in opioid signalling (see Methods) in the spinal cords of (A) juvenile (postnatal day 11-18) and (B) adult (postnatal day 60) male and female WT C57BL/6 mice previously housed in control or fragmented care cages (see Methods). (A) Exposure to FC reduced the expression of Pomc (P = 0.015) mRNA and enhanced the expression of Pdyn (P = 0.017), Oprd1 (P = 0.042), and Arrb2 (P = 0.027) mRNAs in the spinal cord of juvenile mice relative to controls. (B) Adult mice previously exposed to FC similarly had enhanced expression of Pdyn (P = 0.01) and Oprd1 (P = 0.0005) mRNAs, and reduced expression of Pomc (P = 0.019) mRNA, in the spinal cord relative to control mice, whereas Arrb2 mRNA expression was significantly reduced (P < 0.0001). Data are normalised to GAPDH expression using the 2^−ΔΔCT method and expressed relative to control mice for each age separately. Each datapoint represents mRNA expression from an individual mouse with the group mean and SD shown. Juvenile mice: n = 9 (control and FC), and adult mice: n = 10 (control and FC). *P < 0.05 compared with control. FC, fragmented maternal care; WT, wild-type.

Figure 4.

Fragmented care has no effect on acute nociception or morphine antinociception in δ−/− and β-arrestin2−/− mice. Acute thermal nociception and morphine antinociception were assessed using the warm water (48°C) tail withdrawal assay in adult (postnatal day 60) control or fragmented care (FC) δ−/− and β-arrestin2−/− mice (see Supplementary Fig. 2, available at http://links.lww.com/PAIN/B818). Mechanical nociception was assessed using automated von Frey. (A and E) FC had no effect on acute thermal nociception in (A) δ−/− (P = 0.98) or (E) β-arrestin2−/− (P = 0.76) mice. (B and F) FC similarly had no effect on acute mechanical nociception in either (B) δ−/− (P = 0.1) or (F) β-arrestin2−/− (P = 0.92) mice. (C and G) Morphine cumulative dose–response relationships in control and FC δ−/− (C) and β-arrestin2−/− (G) mice on day 1. (D and H) Morphine (10 mg/kg) antinociception continued to be assessed using the tail withdrawal assay for the following 8 consecutive days in control and FC δ−/− (D) and β-arrestin2−/− (H) mice. Separate Friedman tests with Bonferroni corrections comparing the maximum possible effect (%MPE) of morphine antinociception to day 1 within each group detected no significant reduction in morphine antinociception in either control or FC (D) δ−/− or (H) β-arrestin2−/− mice. Estimates of morphine ED₅₀ in control and FC δ−/− and β-arrestin2−/− mice on day 1 and on day 10 are summarised in Table 1. Data shown as mean ± SEM in all cases. δ−/− mice: n = 8 (control) and 11 (FC), and β-arrestin2−/− mice: n = 14 (control) and 11 (FC).

Figure 5.

Fragmented care paradigm is not obviously impaired in the absence of δ receptors or β-arrestin2. (A) Total number of nest departures (termed sorties) made by C57BL/6 WT, δ−/−, and β-arrestin2−/− control or fragmented care dams (Fig. 1A and Supplementary Figs. 2A and E, available at http://links.lww.com/PAIN/B818) during postnatal days (PD) 3 to 8. A 2-way ANOVA comparing the total number of sorties, summed across all days with genotype and cage conditions as factors, detected a significant effect of cage condition (F_1,35 = 89.3, P < 0.0001), although there was no significant effect of genotype (F_2,35 = 1.4, P = 0.26) or an interaction (F_2,35 = 0.9, P = 0.4). (B) Total interaction time between control or FC WT, δ−/−, and β-arrestin2−/− dams and their pups (Fig. 1B and Supplementary Figs. 2B and F, available at http://links.lww.com/PAIN/B818) during PD 3 to 8. A 2-way ANOVA comparing interaction times with genotype and cage conditions as factors detected no significant effect of genotype (F_2,35 = 0.54, P = 0.59), cage condition (F_1,35 = 0.02, P = 0.89), or an interaction (F_2,35 = 0.54, P = 0.59). Together, these data demonstrate that fragmented care caused by limited bedding in WT mice is not impaired in δ−/− or β-arrestin2−/− mice. WT mice: n = 10 and 12 (separate control and FC cages, respectively), δ−/− mice: n = 6 and 5 (separate control and FC cages, respectively), and β-arrestin2−/− mice: n = 4 and 4 (separate control and FC cages, respectively). ANOVA, analysis of variance; C, control; FC, fragmented maternal care; WT, wild-type.

Figure 6.

Fragmented care causes persistent mechanical hypersensitivity that is dependent on δ receptors and β-arrestin2. Mechanical sensitivities of control and fragmented care (FC) WT (A), δ−/− (B), and β-arrestin2−/− (C) mice, expressed relative to baseline, were assessed for 30 days after a unilateral injection with 10-µL undiluted (1 mg/mL) complete Freund's adjuvant (CFA) in 1 hind paw. Statistical comparisons were performed using a 2-way ANOVA with repeated measures for each genotype separately using cage condition and time as factors. (A) Recovery from CFA-evoked hypersensitivity was complete and sustained by day 7 in WT control mice. By contrast, whereas recovery occurred by day 7 in WT FC mice, mechanical hypersensitivity recurred in these mice on day 11 and beyond (time: F_9,216 = 9.8, P < 0.0001; cage condition: F_1,24 = 4.3, P = 0.05; interaction: F_9,216 = 2.0, P = 0.04). FC did not affect recovery in (B) δ−/− or (C) β-arrestin2−/− mice. Recovery occurred by day 15 in δ−/− mice (time: F_9,162 = 70.5, P < 0.0001; cage condition: F_1,18 = 0.3, P = 0.57; interaction: F_9,162 = 0.8, P = 0.65) and by day 9 in β-arrestin2−/− mice (time: F_9,153 = 25.3, P < 0.0001; cage condition: F_1,17 = 0.3, P = 0.61; interaction: F_9,153 = 0.8, P = 0.59) and was sustained on day 30. Data are expressed as mean ± SEM. Absolute mechanical sensitivities are summarised in Supplementary Table 3 (available at http://links.lww.com/PAIN/B818). *P < 0.05 compared with baseline, and #P < 0.05 compared with control. n (control, then FC) = 13 and 13 (WT), n = 9 and 11 (δ−/−), and n = 9 and 10 (β-arrestin2−/−). WT, wild-type.