Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Mar 4.
Published in final edited form as: Biol Psychiatry. 2020 Mar 29;88(8):649–656. doi: 10.1016/j.biopsych.2020.03.007

Translocator Protein Distribution Volume Predicts Reduction of Symptoms During Open-Label Trial of Celecoxib in Major Depressive Disorder

Sophia Attwells 1,2, Elaine Setiawan 1, Pablo M Rusjan 1, Cynthia Xu 1, Celeste Hutton 1, Dorsa Rafiei 1, Benjamin Varughese 1, Alan Kahn 3, Stephen J Kish 1,2,3, Neil Vasdev 1, Sylvain Houle 1,3, Jeffrey H Meyer 1,2,3
PMCID: PMC11878442  NIHMSID: NIHMS1796710  PMID: 32402468

Abstract

Background:

Gliosis is common among neuropsychiatric diseases but the relationship between gliosis and response to therapeutics targeting effects of gliosis is largely unknown. Translocator protein total distribution volume (TSPO VT), measured with positron emission tomography (PET), mainly reflects gliosis in neuropsychiatric disease. Here, the primary objective was to determine whether TSPO VT in the prefrontal cortex (PFC) and anterior cingulate cortex (ACC) predicts reduction of depressive symptoms following open-label celecoxib administration in treatment resistant major depressive disorder (TRD).

Methods:

41 TRD subjects underwent one [18F]FEPPA PET to measure PFC and ACC TSPO VT. Open-label oral celecoxib 200 mg bid was administered for 8 weeks. Change in symptoms was measured with the 17-item Hamilton Depression Rating Scale (HDRS).

Results:

Cumulative mean change in HDRS between 0 and 8 weeks of treatment was plotted against PFC and ACC TSPO VT, showing a significant non-linear relationship. At low TSPO VT values, there was no reduction in HDRS, but as TSPO VT values increased, there was a reduction in HDRS, which then plateaued. This was modelled with a 4-parameter sigmoidal model in which PFC and ACC TSPO VT respectively accounted for 84% and 92% of the variance.

Conclusions:

Celecoxib administration in the presence of gliosis labelled by TSPO VT is associated with greater reduction of symptoms. Given the predictiveness of TSPO VT on symptom reduction, this personalized medicine approach of matching a marker of gliosis to medication targeting effects of gliosis should be applied in early development of novel therapeutics, in particular for TRD.

Registry Name:

ClinicalTrials.gov

URL:

https://www.clinicaltrials.gov/ct2/show/NCT02362529?term=NCT02362529&cntry=CA&rank=1

Clinical Trial Registration Number:

NCT02362529

Keywords: Major Depressive Disorder, Translocator Protein, Positron Emission Tomography, Celecoxib, Treatment Response, gliosis

Introduction

Gliosis is frequently present among neuropsychiatric diseases and experts in the field often propose that there is an underutilized therapeutic opportunity to target effects of gliosis (1–3). However, targeting effects of gliosis in neuropsychiatric diseases in clinical trials is difficult because the magnitude of gliosis present varies within individual cases, depending on factors such as phase of illness and disease severity. Compounding this issue is that gliosis is difficult to quantitate routinely in vivo. Hence, to date, no study has addressed the question of whether the presence of gliosis in the brain is associated with a preferential response to a therapeutic that targets its effects of gliosis. The intent of this study is to investigate whether there is a relationship between translocator total distribution volume (TSPO VT), an index of translocator protein density, and response to celecoxib during major depressive episodes (MDE) of major depressive disorder.

Positron emission tomography (PET) imaging of TSPO is the most established marker of gliosis in vivo, but its interpretation has some complexity. Elevated TSPO binding is viewed as a measure of gliosis because TSPO is often strongly overexpressed in microglia and to a lesser extent astrocytes in neuropsychiatric diseases such as Alzheimer’s disease, Parkinson’s disease, Multiple Sclerosis, Huntington’s disease, HIV infection, and amyotrophic lateral sclerosis (4–9). However during health, TSPO expression is mainly attributed to its presence in endothelial cells (10). Over the past decade, a new generation of PET radiotracers were developed with superior quantification of TSPO binding. Among these, [18F]FEPPA has excellent properties including high, selective affinity for TSPO (11), increased binding during induced neuroinflammation (12), no detectable radioactive metabolites in brain in preclinical assessment (13) and a high ratio of specific binding relative to free and non-specific binding (13, 14).

MDD has enormous impact on society because 4% of the general population is in the midst of a major depressive episode (MDE) (15) and half of MDD is resistant to conventional treatment (16). PET studies of TSPO in MDE consistently report greater TSPO binding, in the prefrontal cortex (PFC) and anterior cingulate cortex (ACC). This is demonstrated across studies at four different sites totalling 142 patients and 93 controls (17–21). In these studies, ACC values are elevated 15% to 67% and PFC values are elevated 25% to 35%. One additional [11C]PBR28 study reported a negative result within a sample combining 5 MDE subjects and 4 recovered depressed subjects, however, the same group reports that the 5 subjects in a current MDE also had elevated TSPO VT values in gray-matter regions (22). The PFC and ACC are important regions in the pathophysiology of MDE because neuropathological abnormalities are frequently identified within these regions in MDE, subregions within these structures regulate affect, and these regions are functionally sensitive to depressed mood induction (23). Thus, MDE represents a neuropsychiatric disease for which elevated TSPO VT is often present in pathophysiologically important brain regions.

In the present study, the objective was to evaluate the relationship between pre-treatment TSPO VT within the PFC and ACC and change in the 17-item Hamilton Depression Rating Scale (HDRS) (24) scores after an open-label trial of celecoxib, a cyclooxygenase-2 (COX-2) inhibitor. Our overall model is to assess whether targeting a downstream pathway of gliosis with celecoxib will have a greater clinical effect in MDE with greater level of gliosis marker TSPO VT in the PFC and ACC. Celecoxib is available for off-label use in clinical trials because it is approved for treating osteoarthritis in many countries, including United States, Canada, and most of Europe. COX-2 is often overexpressed during gliosis in chronic neuropsychiatric disorders, particularly at later phases of disease when gliosis is more prominent, although during disease states overexpression of COX-2, may also occur in injured neurons (1, 3). COX-2 is an important enzyme for producing a number of inflammatory cytokines, such as prostaglandin (PG) D2, PGF2α, PGI2, thromboxane A2 and PGE2; and is implicated in cytotoxic pathways (2, 3, 25). Moreover, it has also been proposed that greater expression of secondary inflammatory mediators, leads to further induction of COX-2 itself (2, 3, 25). Interestingly, overexpression of COX-2 in rodent cortex is associated with worsened depressive behaviors after chronic unpredictable stress (26). Celecoxib has been investigated to treat non-responders to antidepressants in MDE secondary to MDD in several small double-blind placebo-controlled studies, with most trials reporting significantly greater reduction in symptoms with celecoxib over placebo in at least one point over the course of the trial (27–30). Collectively these findings suggest that matching celecoxib administration in MDE to presence of gliosis may result in greater reduction of symptoms. Hence, it is hypothesized that there will be a relationship between TSPO VT in the PFC and ACC and reduction of MDE symptoms after celecoxib administration such that the reduction will be negligible when TSPO VT is low but of greater magnitude when TSPO VT is more elevated.

Methods

Participants

Forty-one participants with treatment resistant MDD (TRD) were recruited from the Greater Toronto Area and a tertiary psychiatric hospital, Centre for Addiction and Mental Health (CAMH), between February 2015 and December 2018 (Figure 1, Table 1). TRD was defined as a history of non-response to a clinically appropriate dose and duration of at least one antidepressant whose mechanism of action includes raising extracellular serotonin concentration and one or more antidepressants whose mechanism of action includes raising norepinephrine concentration; or non-response to at least one serotonin and norepinephrine reuptake inhibitor. To assess the diagnosis, all participants underwent the Structured Clinical Interview for DSM-IV with confirmation by a consultation with a psychiatrist (JHM). All participants reported receiving a stable clinical dose of an antidepressant medication for at least four weeks prior to PET scanning. Eligibility at the start of the celecoxib trial required, in addition to MDE diagnosis, being either a non-responder (HDRS ≥ 15 (n=34)) or a partial responder (HDRS ≥ 9 (n=7)). All were aged 18 to 58 years, non-cigarette smoking, and otherwise in good physical health.

Figure 1. Flow Diagram of Study Participants.

Figure 1.

Open in a new tab

Flow of participants through the study.

Abbreviations: PET, positron emission tomography.

Table 1.

Demographic Characteristics of Study Participants

Characteristics Treatment Resistant Depressed Participants (n=41)
Sex
 Female 28
 Male 13
Translocator Protein Genotype a
 High-Affinity Binders 24
 Mixed-Affinity Binders 17
Age, Year, Mean (SD) 32.9 (10.5)
Body Mass Index, Mean (SD) 24.8 (4.6)
17-Item HDRS Score, Mean (SD) b 19.7 (5.0)
Age at First MDE, Year, Mean (SD) 15.3 (5.8)
Number of MDEs, Mean (SD) 7.1 (8.2)
Total Antidepressant Treatment, Years, Mean (SD) 8.3 (7.0)
Years of Untreated MDD, Mean, (SD) 8.3 (6.5)
Open in a new tab
a

Single nucleotide polymorphism rs6971 of the TSPO gene known to influence [18F]FEPPA binding: HAB, high affinity binders; MAB, mixed affinity binders.

b

17-item HDRS; scores derived on the day of scanning.

Abbreviations: HDRS, Hamilton Depression Rating Scale; MDD, major depressive disorder; MDE, major depressive episode; SD, standard deviation; TSPO, translocator protein.

Exclusion criteria, which were mainly designed to rule out causes that might abruptly alter TSPO VT and create excessive variability in this measure, included history of, or current substance abuse disorder; current medical diseases associated with inflammation; electroconvulsive therapy or mechanical brain stimulation treatments within the previous six months; use of anti-inflammatory drugs lasting at least one week within the past month; and current hormone replacement therapy. In addition, pregnancy was also an exclusion criterion. The use of oral contraceptives was not an exclusionary criterion. All participants underwent urine drug screening and all women under age 65 years underwent a urine pregnancy test on both enrollment and PET scan days. To avoid potential bias of recent infection on TSPO VT, scheduling of PET scans was done in a manner such that all participants reported no recent infections at least four weeks prior to PET scanning. In addition, the binding affinity of [18F]FEPPA for TSPO is affected by a single-nucleotide polymorphism (rs6971; C→T) in exon 4 of the TSPO gene (NCBI Entrez Gene 706) (31, 32). Individuals with high-affinity binding (Ala147/Ala147) and mixed-affinity binding (Ala147/Thr147) account for more than 95% of the population in the greater Toronto area (17, 32). The rs6971 polymorphism was genotyped as described in Supplemental Information. Homozygotes for the low-affinity binding gene (Thr147/Thr147) were excluded from data analysis.

All participants provided written informed consent after all procedures were fully explained. The study, protocol, and informed consent forms were approved by the Research Ethics Board at CAMH.

Image Acquisition and Analysis

Each participant underwent one [18F]FEPPA (13) PET (HRRT; CPS/Siemens, Knoxville, TN, USA) and one two-dimensional axial proton density magnetic resonance imaging (MRI) scan (General Electric; Milwaukee, WI, USA) at the Research Imaging Centre at CAMH. [18F]FEPPA was administered intravenously as a bolus (mean 185.8 MBq [SD 13.5]). [18F]FEPPA was of high radiochemical purity (>99.7%) and high specific activity (mean 124.6 TBq/mmol [SD 117.2]). The PET scan duration was 125 minutes after the injection of [18F]FEPPA, and PET scans and image acquisition were acquired as described in Supplemental Information.

A brain MRI was acquired using a Signa 3T MRI scanner (section thickness, 2 mm; repetition time, 6000 ms; echo time, 8 ms; flip angle, 90°; number of excitations, 1; acquisition matrix, 256 × 192; and field of view, 16.5 cm) for each participant. Regions of interest (ROIs) were generated using the in-house software, Regions of Mental Interest (ROMI), as described in Supplemental Information.

Celecoxib Dosing Regime and Compliance Calculation

All participants received celecoxib in an 8-week open-labelled trial. Dosing of celecoxib was 100 mg bid for week 1, then increased to 200 mg bid for weeks 2 to 8. After the final study visit, dosing was reduced to 100 mg bid for one week, and then stopped. Participants attended follow-up appointments with the treating physician (JHM) and research staff at 4–5 weeks and 8 weeks of treatment, and within 2 weeks of completing the celecoxib trial.

To assist and monitor compliance, celecoxib was given to participants in blister packs. The blister packs contained designated sealed compartments for the medication to be taken at times of the day (Monday through Sunday, breakfast and supper). To calculate compliance, participants were asked by study staff if they missed a dose. In addition, for a subsample of 32 subjects, blister packs were collected after use, and remaining pills were counted, and an average 97% compliance rate was found.

Mood Scales and Questionnaires

The 17-item HDRS (24) was the primary measure of depression severity change after the celecoxib trial. In addition, the Beck Depression Inventory (BDI) (33) and the Visual Analogs Scale (VAS) for Mood, Energy, and Anxiety were measured. These measures were obtained at celecoxib initiation, 4–5 weeks after starting celecoxib, and 8 weeks after starting celecoxib.

Statistical Analysis

Change in HDRS scores was calculated as the difference between weeks 0 and 8 of celecoxib treatment. TSPO VT in the PFC and ACC was plotted against cumulative mean change in HDRS scores after celecoxib treatment, showing a non-linear relationship of greater reduction in HDRS scores at moderately elevated TSPO VT values which plateaued as TSPO VT values were more elevated. To calculate the cumulative mean change in HDRS scores, the data was first sorted by ordering the participants by their regional TSPO VT values, from least to greatest. The mean cumulative change in HDRS score was calculated by taking the change in HDRS score of the participant of interest, and then averaging the score with the scores of participants with lower regional TSPO VT values. The cumulative mean change in HDRS measure reduces variance related to heterogeneity of response and enables a more sensitive assessment of the relationship of predictor variable to outcome of interest. More specifically, the cumulative mean change was applied to moderate potential bias of participants who are highly non-responsive, such as those with more than 8 previous antidepressant failures across multiple classes of antidepressant intervention, who are unlikely to respond to any oral antidepressant treatment. A non-linear curve fitting of a four parameter sigmoidal model (y=y0+(a/(1+exp(−(x-x0)/b)) was applied to the relationship between regional TSPO VT and change in HDRS scores. To evaluate secondary clinical outcome measures, such as BDI, suicidal ideation, and VAS scores for mood, anxiety and energy, a similar approach was applied.

Post-Hoc Analysis

While the primary hypothesis was that overall magnitude of TSPO VT would be predictive of treatment response, a secondary hypothesis was that there would be also be a maximum TSPO VT for which suprathreshold scores may not necessarily be associated with preferential response because the severity and breadth of the pathology then exceeds the potency of the intervention. In such a circumstance, an optimal window of TSPO VT may be identified that not only has a minimum value of TSPO VT but also a maximum value. As an exploratory analysis, participants were stratified into two groups based on visual identification of an optimal window of PFC TSPO VT and ACC TSPO VT; and a univariate analysis of variance (ANOVA) was performed to determine a difference between the mean change in HDRS between two groups following celecoxib treatment. Also, in a post-hoc analysis, participants with an initial HDRS score below 15 were excluded from analysis, and participants were re-stratified into two groups based on an optimal window of PFC TSPO VT. The same ANOVA analysis described above was performed.

Results

Participant Demographics

Forty-one out of 43 eligible participants completed the 8-week open-labeled celecoxib trial (Figure 1). The mean and standard deviation of the age of participants, age of first MDE onset, and number of previous non-responses to treatment were 32.9 ± 10.5, 15.3 ± 5.8, and 4.3 ± 2.7, respectively (Table 1). Assigned sex at birth of participants included 28 females and 13 males. Preferred genders of participants included 27 females, 12 males, 1 neutral, and 1 trans-female. Six participants responded (HDRS decrease ≥50%) and 3 participants remitted (final HDRS≤7) following the 8 week open-label celecoxib treatment.

Relationship of Translocator Protein Total Distribution Volume to Change in Hamilton Depression Rating Scale Scores Following Celecoxib Trial

There was a non-linear relationship between regional TSPO VT and cumulative mean change in 17-item HDRS scores which fit a four-parameter sigmoidal model (y=y0+(a/(1+exp(− (x-x0)/b)) reflecting a lesser reduction in HDRS scores at low TSPO VT, and a greater reduction in HDRS at higher TSPO VT values (Figure 2). The fitting was highly significant for the PFC (a=5.70, t(39)=7.84, P<0.0001; b=−0.59, t(39)=−3.73, P=0.0006; x0=9.93, t(39)=43.69, P<0.0001; y0=−3.55, t(39)=−23.37, P<0.0001; R2=0.84; Figure 2) and ACC (a=5.54, t(39)=17.76, P<0.0001; b=−0.079, t(39)=−3.07, P=0.004; x0=8.84, t(39)=458.40, P<0.0001; y0=−3.70, t(39)=−42.69, P<0.0001; R2=0.92; Figure 2). To assess whether there was a difference in fitting between the HAB and MAB the sum of squares of the residuals was compared between groups and there were no significant differences (PFC: F16,23=0.52, p>0.2; ACC: F16,23=0.29, p>0.2). A similar relationship was observed between clinical change and baseline TSPO VT in the subregions of the PFC, including the orbitofrontal, ventrolateral, dorsolateral and medial (Supplemental Figure 1).

Figure 2. Cumulative Mean Change in Hamilton Depression Rating Scale Scores Based on Regional Translocator Protein Total Distribution Volume.

Figure 2.

Open in a new tab

Cumulative mean change in 17-item HDRS scores based on regional TSPO VT demonstrated greater change in HDRS scores at TSPO VT values greater than 10 to 11. Black line represents the 4-parameter sigmoidal function (y=y0+(a/(1+exp(−(x-x0)/b)) that provides the best fit for the data [PFC: Adjusted R2=0.84, y=−3.55+(5.70/(1+exp(−(x-9.93)/−0.59)); ACC: Adjusted R2=0.92, ACC: y=−3.70+(5.54/(1+exp(−(x-8.84)/−0.079))].

a To address the effect of the rs6971 genotype on TSPO VT, its differential effect on TSPO VT was found applying a linear regression. Then, the MAB TSPO VT values were adjusted by adding the differential effect of genotype to the TSPO VT values. (Adjusted PFC TSPO VT=unadjusted TSPO VT+b1*genotype, where b1=4.490; and adjusted ACC TSPO VT=unadjusted TSPO VT+b1*genotype, where b1=4.559).

Abbreviations: ACC, anterior cingulate cortex; HDRS, Hamilton Depression Rating Scale; MAB, mixed-affinity binder; PFC, prefrontal cortex; TSPO, translocator protein; vs., versus; VT, total distribution volume.

There was a similar non-linear relationship between regional TSPO VT and cumulative mean percent change in 17-item HDRS scores which fit a four-parameter sigmoidal model (y=y0+(a/(1+exp(−(x-x0)/b)) reflecting a lesser reduction in HDRS scores at low TSPO VT, and a greater reduction in HDRS at higher TSPO VT values (Figure 3). The fitting was highly significant for the PFC (a=−27.28, t(39)=−7.23, P<0.0001; b=0.61, t(39)=3.61, P=0.0009; x0=9.88, t(39)=39.53, P<0.0001; y0=14.11, t(39)=3.96, P=0.0003; R2=0.82; Figure 3) and ACC (a=25.59, t(39)=17.00, P<0.0001; b=−0.067, t(39)= −3.05, P=0.0042; x0= 8.84, t(39)=494.38, P<0.0001; y0=− 13.60, t(39)=−31.94, P<0.0001; R2=0.91; Figure 3). To assess whether there was a difference in fitting between the HAB and MAB the sum of squares of the residuals was compared between groups and there were no significant differences (PFC: F16,23=0.37, p>0.2; ACC: F16,23=0.33, p>0.2). A similar relationship was observed between clinical change and baseline TSPO VT in the subregions of the PFC, including the orbitofrontal, ventrolateral, dorsolateral and medial (Supplemental Figure 2).

Figure 3. Cumulative Mean Percent Change in Hamilton Depression Rating Scale Scores Based on Regional Translocator Protein Total Distribution Volume.

Figure 3.

Open in a new tab

Cumulative mean change in 17-item HDRS scores based on regional TSPO VT demonstrated greater per cent change in HDRS scores at TSPO VT values greater than 10 to 11. Black line represents the 4-parameter sigmoidal function (y=y0+(a/(1+exp(−(x-x0)/b)) that provides the best fit for the data [PFC: Adjusted R2=0.82, y=14.11+(−27.28/(1+exp(−(x-9.88)/0.61)); ACC: Adjusted R2=0.91, ACC: y= −13.60+(25.59/(1+exp(−(x-8.84)/−0.067))].

a To address the effect of the rs6971 genotype on TSPO VT, its differential effect on TSPO VT was found applying a linear regression. Then, the MAB TSPO VT values were adjusted by adding the differential effect of genotype to the TSPO VT values. (Adjusted PFC TSPO VT=unadjusted TSPO VT+b1*genotype, where b1=4.490; and adjusted ACC TSPO VT=unadjusted TSPO VT+b1*genotype, where b1=4.559).

Abbreviations: ACC, anterior cingulate cortex; HDRS, Hamilton Depression Rating Scale; MAB, mixed-affinity binder; PFC, prefrontal cortex; TSPO, translocator protein; vs., versus; VT, total distribution volume.

A post-hoc analysis comparing raw HDRS change and percent HDRS change in groups stratified by TSPO VT of 10 in the PFC, and 8.7 in the ACC was also applied. There was a trend level difference between groups using the PFC cutoffs (HDRS change, Mann Whitney U, p=0.09; per cent HDRS change, Mann Whitney U, p=0.08) and similar findings using the ACC cutoff (HDRS change, Mann Whitney U, p=0.08; per cent HDRS change, Mann Whitney U, p=0.04). However, if participants with greater than 8 previous non-responses, who are unlikely to respond to any oral medication are excluded (n=4), differences based on these thresholds are more significant, (HDRS change, Mann Whitney U, p=0.07; per cent HDRS change, Mann Whitney U, p=0.05) and for the ACC cutoff (HDRS change, Mann Whitney U, p=0.06; per cent HDRS change, Mann Whitney U, p=0.03).

Change in Beck Depression Inventory and Visual Analog Scale Scores Following Celecoxib Treatment

The relationship between PFC TSPO VT and cumulative mean change in BDI scores was plotted and no significant relationship was found using linear or sigmoidal models (linear regression: adjusted R2=0.0025, F1,40=1.1, P=0.30). The relationship between PFC TSPO VT and cumulative mean change in VAS mood scores was plotted and a significant linear relationship was found (linear regression: adjusted R2=0.60, F1,40=60.0, P<0.0001) (Supplemental Figure 3).

Discussion

The present study demonstrates that 8 weeks of oral celecoxib at a dose of 200 mg bid reduces depressive symptoms more prominently in TRD participants with a TSPO VT greater than 10 in the PFC and ACC. Similarly, post-hoc review of change in HDRS scores suggested that the greatest reduction in HDRS corresponded to a window of TSPO VT values between 10.0 and 15.5 in the PFC and ACC. This has major implications for matching therapeutics targeting aspects of gliosis to cases with a greater degree of brain gliosis in neuropsychiatric disease.

The relationship between baseline TSPO VT in the PFC and ACC with cumulative mean reduction in HDRS as shown with the fitting of the four-parameter sigmoidal model demonstrated a robust statistical relationship accounting for 84% and 92% of the variance. This level of statistical effect can be viewed as strong support for the concept of matching therapeutics targeting gliosis to gliosis in MDE. The overall magnitude of reduction in HDRS following celecoxib administration was approximately 4 points greater at more elevated TSPO VT as compared to low TSPO VT values. Such a differential effect may be considered meaningful in clinical trials, as, for example, a differential change of 2 HDRS points between the effect of monotherapy duloxetine 60 mg/day versus placebo was considered a positive finding, contributing towards approval for this commonly prescribed therapeutic (34, 35). However, the overall magnitudes of response in the present study were modest since most clinical trials in MDD prioritize definitions of response (50% reduction in HDRS score) or remission (final HDRS score of 7 or less). On the other hand, with the exception of the clinical trials of ketamine and s-ketamine or deep brain stimulation, such studies typically exclude TRD cases who have the level of TRD in the present study sample (the mean number of non-responses to previous antidepressant treatments was 4.3 in the present study). So, while the statistical robustness of the relationship between reduction in HDRS and baseline TSPO VT argues that matching celecoxib to presence of gliosis is meaningful conceptually, to obtain a more robust clinical change, additional markers and matching therapeutics may be needed since MDE is a heterogeneous illness.

Our post-hoc assessment suggested that clustering participants with either low or very high TSPO VT was associated with lesser change in HDRS, and we speculate that at low level of TSPO VT there is a poor match to the pathology present, but also at high level, under conditions of more extensive gliosis and potentially broader or more severe inflammatory change, there may also be a lesser match between the pathology present and COX-2 inhibition. In the context of the latter, it may be that celecoxib isn’t adequately potent relative to the level of inflammatory abnormality present. Celecoxib, primarily targeting COX-2, represents a relatively focused intervention in the context of broader inflammatory changes in MDD. During MDE of MDD a number of inflammatory abnormalities have been reported in the PFC and/or ACC which include: changes in markers indicative of greater microglial activation; reduction in markers of astroglial activation early in illness transitioning to a greater level of such markers late in illness (17–20, 36); greater gene and protein expression of cytokines such as IL-1β, IL-6, TNFα (37, 38); as well as altered expression of diverse genes with inflammatory and cytokine influencing functions in other brain regions such as the hippocampus (39).

There are advantages and disadvantages to the study design. One limitation of this study is that the evidence for brain penetration of celecoxib is not fully established with a PET imaging occupancy study as there are currently no PET COX-2 radioligands validated in humans. The most advanced such radiotracer is[11C]MC1 which did not demonstrate specific binding in healthy baboons (40). However, administration of intravenous [11C] radiolabelled celecoxib in humans was associated with a peak standard uptake value of 1.2 to 2 in the brain (41), which is similar to reports of some other radiolabelled drugs with high target occupancies like venlafaxine and citalopram (42). The ratio of cerebrospinal fluid (CSF) to unbound plasma concentration of celecoxib is reported as 2:1 which is favorable, but the proportion of unbound concentration of celecoxib is lower than some of other COX-2 inhibitors such as rofecoxib and valdecoxib (43) (neither of these latter medications are available for clinical use). Another limitation of the study is the interpretation of TSPO. As mentioned in the introduction, within the PFC and ACC, TSPO expression in health may be attributed to endothelial cell binding and the elevation of TSPO VT in disease is attributed to gliosis, with the main contribution from microglial activation and a modest contribution from activated astrocytes (4, 6–8). A third limitation is that this is an open-trial so placebo response could influence the overall result, although the likelihood of placebo responses is lower in TRD being zero to 10% in recent prominent investigations of esketamine and deep brain stimulation (44, 45). To minimize biasing of HDRS scores, study staff were blinded to TSPO VT values, as the PET data for each participant was processed following completion of the 8-week celecoxib trial. In addition, while we chose to assess TSPO VT as a predictor of change in symptoms, an alternative direction could have been to assess the change in TSPO VT after celecoxib administration as a predictor of change in symptoms, which could be examined in a future study. Also, to reduce potential bias of participants who are highly non-responsive such as those with more than 8 previous non-responses to antidepressant treatment, we applied a cumulative mean change to assess the relationship of TSPO VT to clinical change. An alternative approach would have been to exclude such participants.

To the best of our knowledge, this is the first investigation evaluating the relationship between degree of brain gliosis; and reduction in symptoms after administration of a medication targeting an effect of gliosis in neuropsychiatric disease. We found support for the concept of matching treatments targeting aspects of gliosis to cases with greater level of gliosis: In TRD, TSPO VT in the PFC and ACC were predictive of symptom reduction after administration of add-on celecoxib. This argues that future investigations of medications targeting aspects of gliosis in MDE should apply stratifying approaches to select cases with high TSPO VT in the PFC and ACC.

Supplementary Material

1

KEY RESOURCES TABLE

Resource Type Specific Reagent or Resource Source or Reference
Add additional rows as needed for each resource type Include species and sex when applicable. Include name of manufacturer, company, repository, individual, or research lab.
Include PMID or DOI for references; use “this paper” if new.
Antibody
Bacterial or Viral Strain
Biological Sample
Cell Line
Chemical Compound or Drug
Commercial Assay Or Kit
Deposited Data; Public Database
Genetic Reagent
Organism/Strain
Peptide, Recombinant Protein
Recombinant DNA
Sequence-Based Reagent
Software; Algorithm Two tissue compartment model PMID: 21522163
Transfected Construct
Other
Radiotracer [18F] FEPPA PMID: 1835586
Identifiers Additional Information
Include catalog numbers, stock numbers, database IDs or accession numbers, and/or RRIDs. RRIDs are highly encouraged; search for RRIDs at https://scicrunch.org/resources. Include any additional information or notes if necessary.
modeling reference cited in PMID
synthesis reference cited in PMID
Open in a new tab

Translocator Protein to Predict Improvement After Celecoxib.

Translocator protein (TSPO) often labels an inflammatory response of gliosis in neuropsychiatric illness. Celecoxib, an inhibitor of inflammatory mediator COX2, shows promising effects in small clinical trials of major depressive disorder. Attwells et al. demonstrate an example of personalized medicine: The positron emission tomography imaging index of TSPO level in the prefrontal and anterior cingulate cortex regions is predictive of reduction in symptoms in treatment resistant depression after 6 weeks of celecoxib administration.

Acknowledgements

Alvina Ng, B.S.c; Laura Nguyen, B.Sc.; and Colin Cole, B.Sc. worked as study PET technicians. Nathan Kolla, M.D. Ph.D.; and Antonio Strafella, M.D. provided medical coverage for the PET scans. Jun Parkes, B.Sc.; Armando Garcia, B.Sc.; and Michael Harkness, M.Sc. of Research Imaging Centre, served as PET chemistry staff. Anusha Ravichandran, B.Sc.; Garfield E Detzler, B.Sc.; and Hillary Bruce, B.Sc. worked as study MRI technicians. Apitharani Santhirakumar and Ashmi Sharma assisted with the data analysis. All contributors are paid employees of the Centre for Addiction and Mental Health.

Financial Disclosures

This study received funding support from the Canadian Institutes of Health Research [Canada Research Chair and Operating Grant (MOP-136955), JHM; Doctoral Award (GSD-157948), SA; Fellowship Award, ES; Canada Research Chair, NV], the Brain and Behavior Foundation, and the neuroscience catalyst fund (from the Government of Ontario and Janssen). None of the funding sources participated in the execution of the project, generation of results, interpretation of data nor drafting of the manuscript. Funding for infrastructure was from the Azrieli Foundation, the Canadian Foundation for Innovation and the Ontario Ministry for Innovation.

Competing Interests

SH and JHM have received operating grant funds from Janssen in the past 5 years. JHM has been a consultant to Lundbeck and Takeda, in the past 5 years. JHM is an inventor on five patents of blood and/or clinical markers to predict brain inflammation or to diagnose affective disorders, and a dietary supplement to reduce depressed mood post-partum. JHM is arranging collaborations with nutraceutical companies for the dietary supplement to prevent post-partum depression. SK has operating grant funds from US National Institutes of Health National Institute on Drug Abuse (DA04066) to measure microglial status in the brains of methamphetamine users, and from Jazz Pharmaceuticals for an unrelated study. All other authors report no biomedical financial interests or potential conflicts of interest.

References

  • 1.Minghetti L (2004): Cyclooxygenase-2 (COX-2) in inflammatory and degenerative brain diseases. Journal of neuropathology and experimental neurology. 63:901–910. [DOI] [PubMed] [Google Scholar]
  • 2.Block ML, Zecca L, Hong JS (2007): Microglia-mediated neurotoxicity: uncovering the molecular mechanisms. Nature reviews Neuroscience. 8:57–69. [DOI] [PubMed] [Google Scholar]
  • 3.Rojas A, Chen D, Ganesh T, Varvel NH, Dingledine R (2019): The COX-2/prostanoid signaling cascades in seizure disorders. Expert opinion on therapeutic targets. 23:1–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Cosenza-Nashat M, Zhao ML, Suh HS, Morgan J, Natividad R, Morgello S, et al. (2009): Expression of the translocator protein of 18 kDa by microglia, macrophages and astrocytes based on immunohistochemical localization in abnormal human brain. Neuropathology and applied neurobiology. 35:306–328. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Dupont AC, Largeau B, Santiago Ribeiro MJ, Guilloteau D, Tronel C, Arlicot N (2017): Translocator Protein-18 kDa (TSPO) Positron Emission Tomography (PET) Imaging and Its Clinical Impact in Neurodegenerative Diseases. International journal of molecular sciences. 18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Banati RB, Myers R, Kreutzberg GW (1997): PK (‘peripheral benzodiazepine’)--binding sites in the CNS indicate early and discrete brain lesions: microautoradiographic detection of [3H]PK11195 binding to activated microglia. Journal of neurocytology. 26:77–82. [DOI] [PubMed] [Google Scholar]
  • 7.Martin A, Boisgard R, Theze B, Van Camp N, Kuhnast B, Damont A, et al. (2010): Evaluation of the PBR/TSPO radioligand [(18)F]DPA-714 in a rat model of focal cerebral ischemia. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism. 30:230–241. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Cosenza-Nashat M, Zhao ML, Marshall HD, Si Q, Morgello S, Lee SC (2007): Human immunodeficiency virus infection inhibits granulocyte-macrophage colony-stimulating factor-induced microglial proliferation. Journal of neurovirology. 13:536–548. [DOI] [PubMed] [Google Scholar]
  • 9.Venneti S, Wang G, Nguyen J, Wiley CA (2008): The positron emission tomography ligand DAA1106 binds with high affinity to activated microglia in human neurological disorders. Journal of neuropathology and experimental neurology. 67:1001–1010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Betlazar C, Harrison-Brown M, Middleton RJ, Banati R, Liu GJ (2018): Cellular Sources and Regional Variations in the Expression of the Neuroinflammatory Marker Translocator Protein (TSPO) in the Normal Brain. International journal of molecular sciences. 19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Bennacef I, Salinas C, Horvath G, Gunn R, Bonasera T, Wilson A, Gee A. and Laruelle M. (2008): Comparison of [11C]PBR28 and [18F]FEPPA as CNS peripheral benzodiazepine receptor PET ligands in the pig. J Nucl Med. 49:81P. [Google Scholar]
  • 12.Kudo G, Toyama H, Hatano K, Suzuki H, Wilson AA, Ichise M, et al. (2008): In-vivo imaging of microglial activation using a novel peripheral benzodiazepine receptor ligand, 18F-FEPPA and animal PET following 6-OHDA injury of the rat striatum; A comparison with 11C-PK11195. NeuroImage. 41:T94–T94. [Google Scholar]
  • 13.Wilson AA, Garcia A, Parkes J, McCormick P, Stephenson KA, Houle S, et al. (2008): Radiosynthesis and initial evaluation of [18F]-FEPPA for PET imaging of peripheral benzodiazepine receptors. Nuclear medicine and biology. 35:305–314. [DOI] [PubMed] [Google Scholar]
  • 14.Rusjan PM, Wilson AA, Bloomfield PM, Vitcu I, Meyer JH, Houle S, et al. (2011): Quantitation of translocator protein binding in human brain with the novel radioligand [18F]-FEPPA and positron emission tomography. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism. 31:1807–1816. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.WHO (2001): The World Health Report 2001: Mental Health: New Understanding, New Hope. Geneva, Switzerland: World Health Organization, pp 169. [Google Scholar]
  • 16.Trivedi MH, Rush AJ, Wisniewski SR, Nierenberg AA, Warden D, Ritz L, et al. (2006): Evaluation of outcomes with citalopram for depression using measurement-based care in STAR*D: implications for clinical practice. The American journal of psychiatry. 163:28–40. [DOI] [PubMed] [Google Scholar]
  • 17.Setiawan E, Wilson AA, Mizrahi R, Rusjan PM, Miler L, Rajkowska G, et al. (2015): Role of translocator protein density, a marker of neuroinflammation, in the brain during major depressive episodes. JAMA psychiatry. 72:268–275. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Setiawan E, Attwells S, Wilson AA, Mizrahi R, Rusjan PM, Miler L, et al. (2018): Association of translocator protein total distribution volume with duration of untreated major depressive disorder: a cross-sectional study. The lancet Psychiatry. 5:339–347. [DOI] [PubMed] [Google Scholar]
  • 19.Richards EM, Zanotti-Fregonara P, Fujita M, Newman L, Farmer C, Ballard ED, et al. (2018): PET radioligand binding to translocator protein (TSPO) is increased in unmedicated depressed subjects. EJNMMI research. 8:57. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Li H, Sagar AP, Keri S (2018): Translocator protein (18kDa TSPO) binding, a marker of microglia, is reduced in major depression during cognitive-behavioral therapy. Progress in neuro-psychopharmacology & biological psychiatry. 83:1–7. [DOI] [PubMed] [Google Scholar]
  • 21.Holmes SE, Hinz R, Conen S, Gregory CJ, Matthews JC, Anton-Rodriguez JM, et al. (2018): Elevated Translocator Protein in Anterior Cingulate in Major Depression and a Role for Inflammation in Suicidal Thinking: A Positron Emission Tomography Study. Biological psychiatry. 83:61–69. [DOI] [PubMed] [Google Scholar]
  • 22.Holmes SE, Scheinost D, Finnema SJ, Naganawa M, Davis MT, DellaGioia N, et al. (2019): Lower synaptic density is associated with depression severity and network alterations. Nature communications. 10:1529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Ressler KJ, Mayberg HS (2007): Targeting abnormal neural circuits in mood and anxiety disorders: from the laboratory to the clinic. Nature neuroscience. 10:1116–1124. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Hamilton M (1960): A rating scale for depression. Journal of neurology, neurosurgery, and psychiatry. 23:56–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Iadecola C, Niwa K, Nogawa S, Zhao X, Nagayama M, Araki E, et al. (2001): Reduced susceptibility to ischemic brain injury and N-methyl-D-aspartate-mediated neurotoxicity in cyclooxygenase-2-deficient mice. Proceedings of the National Academy of Sciences of the United States of America. 98:1294–1299. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Chen Q, Luo Y, Kuang S, Yang Y, Tian X, Ma J, et al. (2017): Cyclooxygenase-2 Signalling Pathway in the Cortex is Involved in the Pathophysiological Mechanisms in the Rat Model of Depression. Scientific reports. 7:488. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Majd M, Hashemian F, Hosseini SM, Vahdat Shariatpanahi M, Sharifi A (2015): A Randomized, Double-blind, Placebo-controlled Trial of Celecoxib Augmentation of Sertraline in Treatment of Drug-naive Depressed Women: A Pilot Study. Iranian journal of pharmaceutical research : IJPR. 14:891–899. [PMC free article] [PubMed] [Google Scholar]
  • 28.Abbasi SH, Hosseini F, Modabbernia A, Ashrafi M, Akhondzadeh S (2012): Effect of celecoxib add-on treatment on symptoms and serum IL-6 concentrations in patients with major depressive disorder: randomized double-blind placebo-controlled study. Journal of affective disorders. 141:308–314. [DOI] [PubMed] [Google Scholar]
  • 29.Akhondzadeh S, Jafari S, Raisi F, Nasehi AA, Ghoreishi A, Salehi B, et al. (2009): Clinical trial of adjunctive celecoxib treatment in patients with major depression: a double blind and placebo controlled trial. Depression and anxiety. 26:607–611. [DOI] [PubMed] [Google Scholar]
  • 30.Muller N, Schwarz MJ, Dehning S, Douhe A, Cerovecki A, Goldstein-Muller B, et al. (2006): The cyclooxygenase-2 inhibitor celecoxib has therapeutic effects in major depression: results of a double-blind, randomized, placebo controlled, add-on pilot study to reboxetine. Molecular psychiatry. 11:680–684. [DOI] [PubMed] [Google Scholar]
  • 31.Owen DR, Yeo AJ, Gunn RN, Song K, Wadsworth G, Lewis A, et al. (2012): An 18-kDa translocator protein (TSPO) polymorphism explains differences in binding affinity of the PET radioligand PBR28. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism. 32:1–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Mizrahi R, Rusjan PM, Kennedy J, Pollock B, Mulsant B, Suridjan I, et al. (2012): Translocator protein (18 kDa) polymorphism (rs6971) explains in-vivo brain binding affinity of the PET radioligand [(18)F]-FEPPA. Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism. 32:968–972. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J (1961): An inventory for measuring depression. Archives of general psychiatry. 4:561–571. [DOI] [PubMed] [Google Scholar]
  • 34.Goldstein DJ, Mallinckrodt C, Lu Y, Demitrack MA (2002): Duloxetine in the treatment of major depressive disorder: a double-blind clinical trial. The Journal of clinical psychiatry. 63:225–231. [DOI] [PubMed] [Google Scholar]
  • 35.Detke MJ, Lu Y, Goldstein DJ, Hayes JR, Demitrack MA (2002): Duloxetine, 60 mg once daily, for major depressive disorder: a randomized double-blind placebo-controlled trial. The Journal of clinical psychiatry. 63:308–315. [DOI] [PubMed] [Google Scholar]
  • 36.Moriguchi S, Wilson AA, Miler L, Rusjan PM, Vasdev N, Kish SJ, et al. (2019): Monoamine Oxidase B Total Distribution Volume in the Prefrontal Cortex of Major Depressive Disorder: An [11C]SL25.1188 Positron Emission Tomography Study. JAMA psychiatry. 76:634–641. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Wang Q, Roy B, Turecki G, Shelton RC, Dwivedi Y (2018): Role of Complex Epigenetic Switching in Tumor Necrosis Factor-alpha Upregulation in the Prefrontal Cortex of Suicide Subjects. The American journal of psychiatry. 175:262–274. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Pandey GN, Rizavi HS, Zhang H, Bhaumik R, Ren X (2018): Abnormal protein and mRNA expression of inflammatory cytokines in the prefrontal cortex of depressed individuals who died by suicide. Journal of psychiatry & neuroscience : JPN. 43:376–385. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Mahajan GJ, Vallender EJ, Garrett MR, Challagundla L, Overholser JC, Jurjus G, et al. (2018): Altered neuro-inflammatory gene expression in hippocampus in major depressive disorder. Progress in neuro-psychopharmacology & biological psychiatry. 82:177–186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Kim MJ, Shrestha SS, Cortes M, Singh P, Morse C, Liow JS, et al. (2018): Evaluation of Two Potent and Selective PET Radioligands to Image COX-1 and COX-2 in Rhesus Monkeys. J Nucl Med. 59:1907–1912. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Kumar JSD, Bai B, Zanderigo F, DeLorenzo C, Prabhakaran J, Parsey RV, et al. (2018): In Vivo Brain Imaging, Biodistribution, and Radiation Dosimetry Estimation of [(11)C]Celecoxib, a COX-2 PET Ligand, in Nonhuman Primates. Molecules. 23. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Schou M, Varnas K, Lundquist S, Nakao R, Amini N, Takano A, et al. (2015): Large Variation in Brain Exposure of Reference CNS Drugs: a PET Study in Nonhuman Primates. The international journal of neuropsychopharmacology. 18:pyv036. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Dembo G, Park SB, Kharasch ED (2005): Central nervous system concentrations of cyclooxygenase-2 inhibitors in humans. Anesthesiology. 102:409–415. [DOI] [PubMed] [Google Scholar]
  • 44.Singh JB, Fedgchin M, Daly E, Xi L, Melman C, De Bruecker G, Tadic A, Sienaert P, Wiegand F, Manji H, Drevets WC, Van Nueten L. (2016): Intravenous Esketamine in Adult Treatment-Resistant Depression: A Double-Blind, Double-Randomization, Placebo-Controlled Study. Biological Psychiatry. 80:424–431 [DOI] [PubMed] [Google Scholar]
  • 45.Holtzheimer PE, Husain MM, Lisanby SH, Taylor SF, Whitworth LA, McClintock S, Slavin KV, Berman J, McKhann GM, Patil PG, Rittberg BR, Abosch A, Pandurangi AK, Holloway KL, Lam RW, Honey CR, Neimat JS, Henderson JM, DeBattista C, Rothschild AJ, Pilitsis JG, Espinoza RT, Petrides G, Mogilner AY, Matthews K, Peichel D, Gross RE, Hamani C, Lozano AM, Mayberg HS (2017): Subcallosal cingulate deep brain stimulation for treatment-resistant depression: a multisite, randomised, sham-controlled trial. Lancet Psychiatry. 11:839–849 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1796710-supplement-1.pdf (402.7KB, pdf)

RESOURCES