Male WKY and Wistar rats (250–300 gm; Charles River), 7 to 10-week-old male B6N(Cg)-Htr1a ^{tm1.1(KOMP)Vlcg} /J (5-HT1A^−/−) and C57BL/6NJ wild (WT) mice (Jackson Laboratory) were used. Animals were housed individually (21°C ± 2°C, relative humidity 50 ± 10%, 12 h light/dark cycle) with free access to water and food. Research protocols were approved by the Massachusetts General Hospital Institutional Animal Care and Use Committee. Experiments were conducted with the experimenter being unaware of group assignment.

Fluoxetine hydrochloride, the 5-HT1A receptor antagonist (N-[2-[4-(2-methoxyphenyl)-1-piperazinyl [ethyl]- N-(2-pyridyl) cyclohexane carboxamide (WAY100635)), and the IkB kinase inhibitor (2-amino-6-[2-(cyclopropylmethoxy)-6-hydroxylphenyl]-4-piperidin-4-yl-nicotinonitrile (ACHP)) were purchased from Tocris Bioscience. The IDO1 inhibitor L-1-methyl-tryptophan (1-MT) was from Aldrich-sigma. Fluoxetine and WAY100635 were dissolved in sterilized PBS, and PBS was used as vehicle control. ACHP was dissolved in 10% dimethyl sulfoxide (DMSO, Sigma-Aldrich) to prepare stock solutions (1 mM) and then diluted in PBS to 1μM, thus the final 0.01% DMSO in PBS was used as vehicle control. 1-MT was initially dissolved in 0.5 N HCl at 50 mg/ml, adjusted to 10 mg/ml with PBS, and PBS was used as control.

To induce depression-like behavior in rats, chronic social stress was introduced using a modified resident-intruder social interaction method as described previously (Kim et al., 2012). An experimental rat (275–300 gm), named as an intruder rat in this model, was transferred from its home cage into a cage of a resident rat (500–600 gm) for 1 h per day. The intruder rat and the resident rat were separated by a small round wire-mesh compartment (diameter 11 cm, height 14 cm) within the resident cage. After 1 h, the intruder rat was returned to its home cage. This procedure was carried out at the beginning of the dark cycle in a 24-h light cycle (light on and off for each 12 hour-period). An intruder rat was confronted with a different resident rat each day. This process lasted for 4 weeks. For controls, rats were placed in the same behavioral room but without the social interaction with a resident rat.

RN 46A cells (Public Health England), an immortalized serotonergic neuronal cell line derived from embryonic rat medullary raphe cells by infection with a retrovirus encoding the temperature-sensitive mutant of SV 40 large T-antigen (White et al., 1994), were cultured with 1:1 solution of DMEM/Ham’s F12 (GIBCO, Invitrogen), 100 U/ml penicillin, and 100 ug/ml streptomycin (Invitrogen) in a humidified atmosphere of 5% CO₂ and 95% air at 33°C.

Cells were fixed with 4% paraformaldehyde for 15 min at room temperature, rinsed with PBS. After fixation, cover cells with ice-cold 100% methanol, incubated cells for 10 min in freezer, rinsed with PBS and incubated in blocking solution (PBS containing 1% BSA and 0.3% Triton X-100) for 1 h. Cells were then incubated overnight at 4°C on a rocker with one of the following primary antibodies: 1:200 NF-κB p65 mouse monoclonal (Santa Cruz Biotechnology Inc.) (Cordoba-David et al., 2022; Zveik et al., 2022). Cells were then washed with PBS three times and incubated with 1:300 cyanine3-conjgated secondary antibody (Jackson Immuno Research Lab. Inc.) for 1 h at room temperature. Blue fluorescent DAPI (1:400; Invitrogen) was used to stain the nucleus of fixed cultured RN46A cells for 15 min. Moreover, for all antibodies used in the present study, we applied primary delete staining (only secondary antibody was used) to confirm the true positive. Images were examined with a fluorescence microscope (Olympus) and captured with a digital camera. The images were analyzed using Adobe PhotoShop (Version 7).

The DRN was collected and homogenized in an SDS lysis buffer (10 ml/mg tissue, ingredients of SDS buffer) containing a cocktail of protease and phosphatase inhibitors (Sigma, St. Louis, MO, United States of America) at 4°C. Samples (50 μg per lane) were separated by SDS-PAGE gel (4%–15% gradient gel; Bio-Rad) electrophoresis and then transferred to polyvinylidene difluoride filter (PVDF) membranes (Millipore, MA). Membranes were blocked with 5% nonfat dry milk in Tris buffered saline with Tween 20 (TBST) for 1 h at room temperature and incubated over night at 4°C on a rocker with one of following primary antibody: 1:1000 5-HT1A receptor (abcam; rabbit polyclonal antibody) (Chao et al., 2020; Kondaurova et al., 2020), 1:200 IDO1 (Santa Cruz Biotechnology Inc.; rabbit polyclonal antibody) (Kim et al., 2012), and 1:1000 NF-κB p65 (Santa Cruz Biotechnology Inc.; mouse monoclonal antibody). After being washed with TBST x 4, membranes were incubated with 1: 8,000 HRP conjugated secondary antibody (Millipore) at room temperature for 1 h. After washing, protein bands were visualized in enhanced Chemiluminescent (ECL) solution (Pierce) for 5 min and exposed to hyperfilms for 15 min. Blots were incubated in a Stripping buffer (Pierce) for 15 min at room temperature and reprobed with 1:12,000 anti-β-actin antibody (Sigma) as a loading control. Western blots were made in triplicates. Band density was measured and normalized against a loading control band.

Total RNA was isolated from each group of samples using TRIzol (Invitrogen, Grand Island, NY). Then, 4 μg of total RNA was used to prepare cDNA according to the protocol for SuperScript III kit (Invitrogen, Grand Island, NY). TaqMan Gene Expression Assays (Applied Biosystems) containing primers and probes for 5-HT1A (Rn00561409_s1) and IDO1 (Rn00675146_ m1) were used to quantify candidate genes. The reaction was performed under the cycling conditions required in the protocol for TaqMan with a 7300 Real-Time PCR System (Applied Biosystems). The expression of candidate genes was normalized to GAPDH to obtain a ΔCt value and to calculate a 2^{−(mean ΔΔCt)}.

Contents of tryptophan, kynurenine, and serotonin in the DRN were determined by HPLC, as previously described (Kim et al., 2012). Samples were taken between 2 and 5 PM. Kynurenine was measured by a UV detector (Shimadzu, SPD-10Avp; 360 nm wavelength). Tryptophan and serotonin were measured by a fluorescence detector (Shimadzu, RF-10Axl; 286 nm excitation and 366 nm emission wavelength). External calibration was made daily to minimize day-to-day variations by using freshly prepared control samples of same concentration (100 nmol/L tryptophan, 100 μmol/L kynurenine, and 100 nmol/L 5-HT).

All data were expressed as mean ± SD. Sample size was estimated using power analysis based on the previous data on behavioral testing (Kim et al., 2012) and our pilot experiments. Behavioral tests at different time points in this study were performed within the same animals. For FST and TH in Figures 1A, B, unpaired 2-tailed Student t-test was used to compare differences between groups. For FST, TST and TH with different time points, repeated measures (RM) 2-way ANOVA and post hoc multiple comparison Bonferroni tests were used to compare the performances between groups in each time point. HPLC, western blot, real-time PCR and all other behavior data were analyzed using a 1-way or 2-way ANOVA and post hoc multiple comparison Bonferroni tests. The statistical analyses were performed using GraphPad (version 8.0) with the significance level set at 0.05.

To examine whether the fluoxetine effects on comorbid pain and depression would be diminished during maintenance treatment, we used three depression-pain comorbidity models (genetic predisposition, chronic social stress, CFA-induced hindpaw monoarthritis). (1) Wistar-Kyoto (WKY) rats, a genetic model of depression and hyperalgesia (Pare, 1994), (Millard et al., 2020; Ferdousi et al., 2021), had depression-like behavior shown as increased baseline immobility time in forced swimming test (FST) (Figure 1A, unpaired t-test, n = 8, t = 7.743, df = 14, p < 0.0001) and hyperalgesia-like behavior shown as decreased baseline paw withdrawal latency (PWL) in thermal hyperalgesia test (TH) (Figure 1B, unpaired t-test, n = 8, t = 6.420, df = 14, p < 0.0001) as compared with Wistar rats. WKY rats were genetically resistant to the fluoxetine (10 mg/kg, i. p.) effect on depression (Figure 1C, RM 2-way ANOVA, n = 8/group, F (1,14) = 0.399, p = 0.538) and hyperalgesia (Figure 1D, RM 2-way ANOVA, n = 8/group, F (1,14) = 0.0573, p = 0.814) as demonstrated following 14 consecutive days of treatment. (2) In a Wistar rat model of chronic social stress, comorbid depressive behavior (Figure 1E, RM 2-way ANOVA, n = 8/group, F (2,21) = 21.8, p < 0.0001) shown as increased immobility time in FST (Figure 1E, day 0, Control vs. CSS-Fluo) and hyperalgesia (Figure 1F, RM 2-way ANOVA, n = 8/group, F (2,21) = 47.6, p < 0.0001) shown as decreased PWL in TH (Figure 1F, day 0, Control vs. CSS-Fluo), were also demonstrated before fluoxetine treatment. Continuous daily treatment with fluoxetine (10 mg/kg i. p.) initially demonstrated antidepressant and antihyperalgesic effects when examined on day 14 shown as the decreased immobility time (Figure 1E, day 14, CSS-Veh vs. CSS-Fluo) and increased PWL (Figure 1F, day 14, CSS-Veh vs. CSS-Fluo). But the fluoxetine effect was substantially reduced when examined on day 28, as the immobility time was again increased (Figure 1E, day 28, Control vs. CSS-Fluo) and the PWL was again decreased (Figure 1F, day 28, Control vs. CSS-Fluo). (3) Wistar rats with CFA-induced inflammatory arthritic pain showed the same reduction of fluoxetine’s antidepressant and antihyperalgesic effects during maintenance treatment (Figures 1G, H, RM 2-way ANOVA, n = 8/group; 1G: F (2,21) = 13.7, p = 0.0002; 1H: F (2,21) = 81.5, p < 0.0001). Collectively, these results from three animal models of comorbid depression and pain demonstrate reduced fluoxetine’s antidepressant and antihyperalgesic effects during maintenance treatment.

We found that 5-HT1A receptors were predominantly expressed in the DRN but not in the magnus raphe nucleus (MRN) in WKY rats (Figure 2A, 2-way ANOVA, n = 6/group, F (1,20) = 5.801, p = 0.0258). Since IDO1 plays a key role in the comorbidity of pain and depression in our previous study (Kim et al., 2012), we examined whether fluoxetine treatment would change the expression of 5-HT1A auto-receptors and IDO1 in the DRN. Co-localization of 5-HT1A receptors with IDO1 was detected in the DRN of WKY rats (Figure 2B). Further, we found upregulation of 5-HT1A auto-receptors (Figure 2C, one-way ANOVA, n = 6/group, F (2,15) = 27.11, p < 0.0001) and IDO1 (Figure 2D, one-way ANOVA, n = 6/group, F (2,15) = 26.29, p < 0.0001) expression in the DRN of WKY rats after 14-days of fluoxetine exposure. Moreover, the ratio of kynurenine (KYN; product of IDO1) over tryptophan (TRP) was increased (Figure 2E, one-way ANOVA, n = 6/group, F (2,15) = 15.18, p = 0.0002), whereas the ratio of 5-HT over TRP was decreased (Figure 2F, one-way ANOVA, n = 6/group, F (2,15) = 21.87, p < 0.0001), in the DRN of the rats with upregulated IDO1 expression (Figure 2D), indicating a shift of tryptophan metabolism away from the 5-HT biosynthesis.

To examine a direct cellular effect of fluoxetine exposure on serotonergic neurons, we exposed cultured RN46A cells, an immortalized serotonergic neuronal cell line, to vehicle (PBS) or fluoxetine (10 μM) (Manjappa et al., 2021) for 0.5, 2, 8, or 24 h. Fluoxetine increased mRNA expression of 5-HT1A receptors (Figure 3A, one-way ANOVA, n = 6/group, F (4,25) = 6.116, p = 0.0014) and IDO1 (Figure 3B, one-way ANOVA, n = 6/group, F (4,25) = 7.364, p = 0.0005). Exposure to fluoxetine time-dependently induced the nuclear translocation (Figure 3C, one-way ANOVA, n = 6/group, Cytosol: F (4,25) = 15.52, p < 0.0001; Nucleus: F (4,25) = 7.713, p = 0.0003) of NF-kB, a downstream signaling element of 5-HT1A receptors contributory to transcription of IDO (Hemmati et al., 2019) because 1) NF-kB p65 immunoreactivity was detected in both the cytosol and nucleus of RN46A cells exposed to fluoxetine or LPS (positive control, 100 ng/ml) but only in the cytosol of the PBS control group (Figure 3D), and 2) inhibition of NF-kB with ACHP (an IkB kinase inhibitor, 1 μM) (Victoriano et al., 2006) prevented the upregulation of IDO1 in response to fluoxetine exposure (Figure 3E, one-way ANOVA, n = 6/group, F (4,25) = 18.42, p < 0.0001).

To assess the role of 5-HT1A auto-receptors in the reduced antidepressant and antihyperalgesic effects of fluoxetine, we microinjected a selective 5-HT1A receptor antagonist, WAY100635 (1 μg/0.5 μL once daily), into the DRN of WKY rats for 14 days (Carli and Samanin, 2000). Combined treatment of fluoxetine and WAY100635 decreased 5-HT1A auto-receptor and IDO1 expression at both protein levels (Figures 4A, B), 2-way ANOVA, n = 6/group, 4A: F (1,20) = 17.05, p = 0.0005; 4B: F (1,20) = 9.629, p = 0.0056) and mRNA levels (Figures 4C, D, 2-way ANOVA, n = 6/group, 4C: F (1,20) = 7.931, p = 0.0107; 4D: F (1,20) = 9.902, p = 0.0051) that were elevated by fluoxetine exposure. This same treatment regimen also restored the fluoxetine’s antidepressive (Figure 4E, RM 2-way ANOVA, n = 6/group, F (3,20) = 3.3, p = 0.04) and antihyperalgesic (Figure 4F, RM 2-way ANOVA, n = 6/group, F (3,20) = 4.6, p = 0.013) effects in WKY rats (Veh-Fluo vs. WAY-Fluo).

To further determine the role of 5-HT1A auto-receptors in this process, we used 5-HT1A receptor-knockout (KO) mice (Figure 5B, 2-way ANOVA, n = 6/group, 5A: F (1,20) = 5.298, p = 0.0322; 5B: F (1,20) = 43.13, p < 0.0001) with CFA-induced comorbid pain and depression. 5-HT1A receptor-KO showed the reduced expression of IDO1 in the DRN at the baseline (Figures 5C, D), 2-way ANOVA, n = 6/group, 5C: F (1,20) = 10.01, p = 0.0049; 5D: F (1,20) = 7.325, p = 0.0136) (WT-Veh vs. KO-Veh, WT-Fluo vs. KO-Fluo). Fluoxetine treatment increased the expression of 5-HT1A receptors and IDO1 in the DRN of wild-type mice (WT-Veh vs. WT-Fluo), but not in 5-HT1A KO mice (Figures 5A–D) (KO-Veh vs. KO-Fluo). Consistently, 5-HT1A KO mice demonstrated the sustained antidepressant and antihyperalgesic effects following 28 consecutive days of fluoxetine treatment (Figures 5E–G) and also an earlier onset (on day 7) of the antidepressant (Figures 5E, F), RM 2-way ANOVA, n = 6/group, 5E: F (3,20) = 4.57, p = 0.01; 5F: F (3,20) = 11.08, p = 0.0002) and antihyperalgestic (Figure 5G, RM 2-way ANOVA, n = 6/group, F (3,20) = 10.1, p = 0.0003) effects as compared with WT mice under the same fluoxetine treatment regimen. There were no differences in depressive and hyperalgesic behaviors between WT-Veh group mice and KO-Veh group mice (Figures 5E–G).

To examine whether 5-HT1A-mediated IDO1 upregulation would be important for the reduced fluoxetine therapeutic effects, we administered L-1-methyl-tryptophan (1-MT, IDO1 inhibitor, 10 mg, i. p.) once daily for 28 consecutive days according to a previous report (O'Connor et al., 2009) in models of comorbid pain and depression. In chronic social stress rats with comorbid pain and depression, 1MT alone or in combination with fluoxetine decreased immobility time in FST when examined on day 28 as compared to vehicle treated group (Figure 6A, 2-way ANOVA, n = 6/group, F (1,20) = 23.41, p < 0.0001). The effect of combination treatment was more effective than the fluoxetine or 1MT alone treatment, shown as combined treatment with fluoxetine and 1-MT increased PWL in TH as early as day 7 (Figure 6B, RM 2-way ANOVA, n = 6/group, F (3,20) = 5.76, p = 0.005) and maintained the fluoxetine’s antidepressant and antihyperalgesic effects up to day 28 (Figures 6A, B). Moreover, 1-MT did not change the upregulation of 5-HT1A auto-receptors (Figure 6C, 2-way ANOVA, n = 6/group, F (1,20) = 0.02633, p = 0.8727) but reduced the expression of IDO1 (Figure 6D, 2-way ANOVA, n = 6/group, F (1,20) = 41.36, p < 0.0001), supporting the notion that IDO1 upregulation was downstream to the 5-HT1A upregulation induced by fluoxetine exposure. Importantly, inhibition of IDO1 with 1-MT also restored normal kynurenine/tryptophan and 5-HT/tryptophan ratios that were altered by fluoxetine exposure (Figures 6E, F), 2-way ANOVA, n = 6/group, p < 0.0001, 6E: F (1,20) = 35.87, 6F: F (1,20) = 42.75). Similar outcomes were observed in rats with CFA-induced comorbid pain and depression in which 1-MT combined with fluoxetine prevented the reduced antidepressant and antihyperalgesic effects and restored normal kynurenine/tryptophan and 5-HT/tryptophan ratios induced by fluoxetine exposure alone (Figures 6G–L, 2-way ANOVA, n = 6/group, 6G: F (1,20) = 16.14, p = 0.0007, 6I: F (1,20) = 1.246, p = 0.2775, 6J: F (1,20) = 14.69, p = 0.0010, 6K: F (1,20) = 12.67, p = 0.0020, 6L: F (1,20) = 6.013, p = 0.0235; 6H: RM 2-way ANOVA, n = 6/group, F (3,20) = 6.27, p = 0.0036).