Efficacy of Pentoxifylline for the Treatment of Bipolar I/II Patients with Treatment-Resistant Depression: A Proof-of-Concept, Randomized, Double-Blind, Placebo-Controlled Trial

Background: Immune dysregulation can play a role in depression pathophysiology, and immunological antagonists can improve depressive symptoms in treatment-resistant bipolar depression (TRD) patients according to studies. Objective: To evaluate the anti-depressant effects of the anti-inflammatory drug, pentoxifylline (PTX) in TRD bipolar I/II adult subjects. Methods: This 12-week, randomized, double-blind, placebo-controlled, parallel-group trial of 60 participants was conducted at Hawler Psychiatric Hospital and Private Clinic in Erbil, Iraq. Participants were confirmed as being qualified for bipolar I/II depression based on DSM-5 criteria. Data were analyzed using modified intent-to-treat analysis. Results: There were no significant differences between the two groups in Hamilton Rating Scale for Depression-17 (HAM-D-17) scores (χ 2 =1.9, P = .48) or a significant time × treatment interaction (χ 2 =7.1, P=.54). Nevertheless, a significant effect of time was observed with both groups’ reduction in HAM-D-17 scores from the start to the endpoint (χ 2 = 2.11, P= .002). Besides, a significant time × treatment × CRP interaction was found (χ 2 =3.1, P=0.016), where there was more reduction in HAM-D-17 score in PTX-treated subjects with CRP> 7.1 mg/L. The response rate difference between PTX and the placebo group did not reach a significance level (χ 2 =0.84, p=0.43). Furthermore, serum concentrations of TNF-α, CRP, and IL-6 significantly reduced at week 12 in the PTX group (P=


Introduction
Bipolar disorder (BD) is primarily recurrent chronic disorder characterized by depressive symptoms and fluctuations in mood state and energy, affecting more than 1% of the world's population, leading to cognitive and functional impairment and death by suicide (1,2).Accurate diagnosis of bipolar disorder is difficult in clinical practice and there are currently no valid biomarkers for the disorder as well.Based on the mood episodes of patients, BD is clinically categorized into two common subtypes: bipolar I disorder (BDI) and bipolar II disorder (BDII) (1).No existing treatment approved by the US Food and Drug Administration for bipolar disorder has been developed using an a priori disease model.
Over the recent decades, emerging evidence of the involvement of immune dysregulation in Major Depressive Disorder (MDD) and bipolar disorder has triggered an interest in exploring the potential anti-depressant characteristics of drugs with primary immune targets (3,4).
Interventional studies and meta-analytic reviews have reported that disparate anti-inflammatory agents have variable antidepressant effects in adults with unipolar and bipolar depressive disorders (5,6).Bipolar disorder has been linked to increased immunological markers such as cytokines, chemokines, adhesion molecules, and acute-phase proteins.In this regard, abnormal levels of tumor necrosis factor-alpha (TNF-α), interferon-gamma (INF-γ), interleukin-6 (IL-6), IL-2, IL-J o u r n a l P r e -p r o o f 18, soluble tumor necrosis factor receptor 2 (sTNF receptor 2), CC motif chemokine ligand 2 (CCL2), C-X-C motif chemokine ligand 4 (CXCL4), and CXCL7 showed to have a cross-sectional association with Bipolar disorder (7)(8)(9)(10)(11).These inflammatory mediators may directly or indirectly impact the central nervous system (CNS), by regulating monoamine transporters or precursors (12).Inflammation can negatively regulate the core co-factors essential for tryptophan hydroxylase and dopamine hydroxylase activity, decreasing the serotonin, norepinephrine, and dopamine synthesis rate (13,14).It was suggested that pro-inflammatory cytokines might lead to neurogenesis inhibition via the activation of the nuclear factor-kappa B (NF-κB) pathway (15).Furthermore, reducing glutamate transporter's expression upon inflammation negatively affects neurogenesis by the downregulation of brain-derived neurotrophic factor (BDNF) (16).
Researchers have examined whether reducing inflammatory cytokines might be used as a therapy for TRD, given the link between these cytokines and the disease.Numerous clinical studies have shown that lowering cytokines in depressed people improves mood and increases the effectiveness of anti-depressant interventions (17,18).In this regard, pentoxifylline (PTX), a methylated xanthine derivative with phosphodiesterase inhibitory characteristics, is particularly of great interest.It is an FDA-approved drug for treating intermittent claudication (19), but it also has antiinflammatory, immunomodulatory, and bronchodilatory effects.Cyclic adenosine monophosphate (cAMP) and cyclic guanine monophosphate (cGMP) are substrates of phosphodiesterases (PDEs) that are relatively upregulated in immune cells (20).PDEs inhibition positively regulates the cAMP level, which in turn activates protein kinase cascades suppressing the expression of TNFα, INF-γ, and other pro-inflammatory cytokines and immune cell migration (21,22).In a rat model of depression, PTX has been found to decrease the immobility time in the forced swim test and improve induced depression-like symptoms (21,23).Moreover, a previous proof-of-concept J o u r n a l P r e -p r o o f clinical trial evaluating adjunctive PTX to citalopram reported a significant reduction in the Hamilton Rating Scale for Depression-17 (HAM-D-17) score in patients with MDD, which was accompanied by a significant reduction in serum levels of TNF-α, IL-6, IL-10, 8-Hydroxy-2′deoxyguanosine  in the PTX group (24).Nevertheless, no previous study has evaluated the anti-depressive efficacy of PTX as a monotherapy approach to reverse depressive symptoms in bipolar patients with TRD.
In the present study, we sought to evaluate whether PTX would improve the depression indicators in adult subjects with TRD bipolar disorder I/II.The PTX anti-depressant effects have been assessed by evaluating changes in the HAM-D-17 score and various blood-based inflammatory markers.

Design
The present study was a 12-week, single-center, randomized, double-blind, placebo-controlled, parallel-group, fixed-dose clinical trial of PTX versus placebo in treatment-resistant individuals with bipolar disease I/II.The diagnosis of bipolar I or bipolar II was confirmed using the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), evaluated by Structured Clinical Interview for DSM-5 (SCID-5) (25).This study was conducted in Hawler Psychiatric Hospital and Private Clinic, Erbil, Iraq.The research ethics committee of Hawler Medical University, Erbil, Iraq, approved the study protocol (approval code: HMU PE-EC 24112021/391).This study is registered on ClinicalTrials.gov(identifier: NCT05324735).Before enrolling in the study, all participants signed written informed consent forms.
J o u r n a l P r e -p r o o f

Participants
Eligibility criteria were as follows; female and male adult subjects (18-65 years), currently off an anti-depressant therapeutic regimen or on a consistent one for a minimum of 4 weeks before enrolling in the study, having moderate treatment resistance in the current episode as assessed by the Massachusetts General Hospital Staging (MGH-S) method for TRD (a score ≥2) (26), having a moderate episode of depression as assessed by the Quick Inventory of Depressive Symptomatology (QIDS)-SR-16 (a score ≥14) at the screening stage (26).Patients with hepatitis B or C, active bacterial or fungal infection, history of any malignancy, history of cardiovascular, hepatic, renal disorders, any autoimmune diseases, active inflammatory disorders, allergy to the used medication, Attention Deficit Hyperactivity Disorder (ADHD), personality disorders, drug dependency or abuse (determined by SCID), at risk of suicide or actively suicidal, and patients who had been under treatment with psychotropic agents, including anti-depressants (four weeks before the study), and electroconvulsive therapy, psychotherapy and physiotherapy in the last two months, and pregnant females.
Matched for age, gender and level of education Controls, were recruited from screened to rule out any history of neurodegenerative disorder, psychiatric disorder, mental retardation, infection or cancer.

Intervention
The participants were allocated randomly to receive either placebo tablets (control group) twice daily or the PTX (400 mg) twice daily for 12 weeks.

Outcome Measures
The primary outcome was the change in the depression severity which was assessed by HAM-D-17 score at the commencement of the intervention, weeks 2, 4, 6, 8, 10, and 12 after starting the J o u r n a l P r e -p r o o f intervention.Treatment response (a ≥50% decrease in HAM-D-17 score) and remission rate (a HAM-D-17 score of ≤7) were also assessed.In the present study, the HAMD-17 indicated good internal Cronbach's consistency reliability ( = 0.83).Additionally, side effects were monitored weekly using a side-effect checklist.As secondary outcome measures, serum concentrations of IL-6, CRP, and TNF-α were assessed at baseline and week 12 using the enzyme-linked immunosorbent assay (ELISA).Each participant had ten milliliters of peripheral blood drawn, and the sera were separated by centrifuging at 250 g for 15 minutes.Finally, the sera were kept at -30 °C until they were used.Samples were tested in parallel runs, including duplicate measurements for a set of standards.

Sample Size
By considering a standard deviation of 7 (taken from literature), 80% power, and a two-sided significance of 0.05, the study sample size of 60 subjects was calculated to evaluate differences in HAM-D-17 sores of 5 points between PTX and placebo.(24,27,28).

Randomization and Blinding
A computer-generated code was utilized to split patients randomly into five unit blocks in a 1:1 ratio for enrolment in the placebo or PTX groups.For allocation concealment, opaque envelopes with sequential numbers were used.Distinct researchers conducted all allocation, randomization, and grading procedures.The assignment was concealed from patients, the psychiatrist, and the statistician.

Statistical Analysis
For all treatment efficacy tests, p<0.05 was considered significant.Categorical variables were reported as numbers (percentages), whereas continuous variables were reported as mean ± standard deviation (SD).Comparison of baseline HAM-D-17 score with each time point was analyzed using

J o u r n a l P r e -p r o o f mixed-effects model repeated measures (MMRM) analysis of covariance (ANCOVA). Least
Squares Means (LSM) was calculated instead of Mean when missing values do occur.LSM can be defined as a linear combination (sum) of the estimated effects (means, etc) from a linear model.
In the case where the data contains NO missing values, the results of the MEANS and LSM statements are identical.LSMEANS are used when a covariate(s) appears in the model such as in ANCOVA.Changes in the HAM-D-17 score were analyzed by a two-factor repeated measure analysis of variance (ANOVA).The same analysis was used to examine the effect of therapy and the time × treatment interactions.Time interval measurements were considered as a within-subject component, and the study groups were considered as a between-subject component.Paired t-test was employed for comparisons of biological markers at baseline and endpoint in each group.The qualitative variables and comparison of sociodemographic and clinical characteristics between two groups were evaluated using Fisher's exact test.Correlation test used for the relationship between biological markers and clinical response was performed using Pearson correlation coefficient test.
Analyses were performed using SPSS Statistic 26 (IBM Corp., NY, USA), and all graphs were generated using GraphPad Prism software version 8.3.0 (GraphPad Software, La Jolla CA, USA).

Demographics and Clinical Characteristics
A total number of 203 individuals were assessed for eligibility, 143 of which were excluded.Sixty participants were randomized to PTX group (n= 30 [50%]) and placebo group (n= 30 [50%]).Four subjects in the PTX group discontinued participation due to medical complications (at weeks 4, 7, and 10), and three subjects dropped out in the placebo group due to psychiatric complications (weeks 3, 7, and 8) (Fig 1).These patients' data were analyzed for HAM-D-17, but since there were no endpoint data available for them, cytokine marker analysis was not relevant.As a result, J o u r n a l P r e -p r o o f 53 individuals (88%) completed all 12 weeks of the study (PTX: 26 out of 30 [87%]; placebo: 27 out of 30 [90%], df= 1, P=.41); there was no significant difference in completion rates between the two groups.There were also no significant differences in sociodemographic and clinical factors across the research groups (Table 1).

Anti-depressant Efficacy
No statistically significant differences in HAM-D-17 scores were observed between the intervention groups (χ 2 =1.9, P = .48).Also, there was not a significant time × treatment interaction (χ 2 =7.1, P=.J o u r n a l P r e -p r o o f

J o u r n a l P r e -p r o o f
According to our analysis of comparison of treatment response between placebo and PTX in subgroups, the week 2 was the time point when the PTX differentiated from placebo.

Effect on serum markers
In either group, serum concentrations of CRP, TNF-α, and IL-6 reduced at week 12 (Fig 8A -C).
However, changes in the placebo group did not reach a significance level.Further analysis revealed that, at the end of the trial, the serum concentrations of selected immunological markers decreased significantly in both PTX-responders and PTX-non-responders except for CRP in non-responders (5.33 Vs. 3.88,P= .14).In contrast, neither placebo-responders nor non-responders saw substantial changes in their blood levels of the chosen indicators (Table 2).In both the PTX and placebo J o u r n a l P r e -p r o o f groups, the baseline concentrations of all three indicators were comparable across responders and non-responders.However, PTX-responders showed a significantly higher baseline CRP level than PTX non-responders (8.52 Vs. 5.33, P= .043)(Table 2).

36
J o u r n a l P r e -p r o o f

Discussion
Modern medication development methods have changed due to recent revelations that several pathophysiological pathways are involved in mood disorders and that targeting a particular pathogenic route is ineffective.Decreased neuroplasticity, immunological dysfunction, CBF, oxidative stress, and neuronal death are common dysregulated pathophysiological pathways to treat mood disorders (29)(30)(31).PTX is a pleiotropic drug having the ability to influence multiple pathways implicated in the dysregulated mechanisms of MDD, notably immunological dysfunction.As a result, the current study sought to evaluate PTX's anti-depressive efficacy, emphasizing its influence on inflammatory markers.
Although PTX was generally well-tolerated, it failed to show an overall advantage over placebo in improving depressive symptoms.Nevertheless, an association was found between the increased concentration of CRP and response to PTX.In fact, participants with a baseline CRP level of >7.1 mg/L demonstrated a higher reduction in HAM-D-17 score throughout the trial.An observed 4.9point decrease in HAM-D-17 means scores is in line with previously reported differences between placebo and anti-depressant agents (32).
Subjects with baseline CRP>7.1 demonstrated an improvement in multiple depressive symptoms, including suicidal ideation, agitation, work and activities, depressed mood, and psychic anxiety.Previous research has shown a correlation between a pattern of similar symptoms and neurocircuits that are altered by inflammatory cytokines.Hyper-inflammation cytokines have been demonstrated to cause psychomotor slowness, anhedonia, and reduced basal ganglia activity (32)(33)(34).Additionally, it was shown that the dorsal and subgenual region of the anterior cingulate cortex (ACC), which corresponds to the brain regions responsible for anxiety and depression, is activated when inflammatory cytokines are produced (35)(36)(37)(38).Concerning suicidal ideation, it has been J o u r n a l P r e -p r o o f suggested that microglial cells, with increased density in the ACC region, release cytokines that can modulate noradrenergic or serotonergic neurotransmission and potentially induce suicidality (39,40).
Intriguingly, CRP level was significantly higher in PTX responders than non-responders, implying a negative association of pretreatment hyper-inflammation with anti-depressant resistance, contradicting previous results (40)(41)(42).However, this finding could not be more established as the other pro-inflammatory markers showed high levels, but they were not statistically significantly different in responders and non-responders in the PTX group.On the other side, one may argue that a baseline inflammatory activity is required for PTX to start showing anti-depressant effects.
In this regard, our results showed that placebo outperforms PTX in subjects with CRP levels ≤7.1 mg/L.It is believed that inflammatory cytokines at the physiological level are required to maintain brain processes, including neuroplasticity and neurogenesis, which protect against depression (43)(44)(45)(46).Furthermore, it was shown that upregulation of CNS p11, a biomarker of anti-depressant response, which TNF-α and IFN-γ regulate, is required for serotonergic anti-depressants (47).
Besides, data indicated that nonsteroidal antagonists of immunological markers might impair behavioral symptoms of serotonergic via influencing p11 (47).The results of the current study did not provide solid evidence to support the claim that patients with a pretreatment hyperinflammatory profile are more responsive to PTX since the placebo responders also displayed higher levels of pro-inflammatory markers than the non-responders, even though this difference was not statistically significant.Therefore, these results should be interpreted cautiously.
Moreover our results showed that, serum concentrations of CRP, TNF-α, and IL-6 reduced at week 12 in placebo and PTX group.However, changes in the placebo group did not reach a significance level.Further analysis revealed that, at the end of the trial, the serum concentrations of selected J o u r n a l P r e -p r o o f immunological markers decreased significantly in both PTX-responders and PTX-non-responders except for CRP in non-responders.In contrast, neither placebo-responders nor non-responders saw substantial changes in their blood levels of the chosen indicators.In both the PTX and placebo groups, the baseline concentrations of all three indicators were comparable across responders and non-responders.However, PTX-responders showed a significantly higher baseline CRP level than PTX non-responders.The general inflammatory profile in the PTX-treated group decreased significantly over the course of the trial, and the anti-inflammatory effect of the PTX was significant compared to the placebo.This phenomenon can increase the serum level of BDNF and the bioavailability of serotonin (48,49).BDNF can exert its action by improving neurogenesis and protecting against neuronal damage, which provides a potential explanation for the observed antidepressant effect of PTX (50).PTX inhibits PDEs, which in turn increases the cAMP level, which may raise the BDNF level as a downstream element (51).In addition, a correlation between increased levels of pro-inflammatory cytokines and attenuation of BDNF has been reported, although this effect is yet to be clarified (52).In line with this, it has been shown that reduced proinflammatory cytokine concentrations increase the bioavailability of tryptophan, the main amino acid precursor for serotonin synthesis (53), and tryptophan hydroxylase co-factors (54).Therefore, the decreased pro-inflammatory profile may be the mechanism behind the increase in serotonin and BDNF as positive mood regulators.However, the current study did not find a correlation To the best of our knowledge, this is the first study evaluating the anti-depressant effect of PTX as monotherapy in treatment-resistant bipolar/depressed patients.Given the study's inclusion criteria, the participants' sample remained homogeneous compared to other studies on bipolar patients.However, certain limitations warrant comment.The focus of exploratory analyses on participants with CRP>7.1 mg/L heightened the efficacy of PTX over placebo.Given the low number of subjects in subgroups based on the CRP level, the study was underpowered to evaluate whether PTX is only effective in the patients with higher pretreatment CRP.Also, the underpowered analyses may increase the possibility of false positives in treatment × time × CRP assessment.Another limitation that may affect the power to evaluate the anti-depressant effect of PTX was a high rate of response rate in the placebo group, which is reported to be much lower in TRD patients.This high response rate might be due to the motivated and responsive sample of subjects (with current episodes over two years and no current active medicinal regimen).Lastly, a short follow-up duration could have prevented the ability to identify the anti-depressant effect of J o u r n a l P r e -p r o o f PTX as there was an apparent anti-inflammatory role for PTX.Moreover, in this study we focused on the differences between subgroups, this approach is a post-hoc approach which is a limitation of this study.Another limitation is that several statistical comparisons were performed, which may inflate the p-value.Another limitation of this study was related to the study measures.The outcome measures were self-report, highlighting the need for multimethod assessment to reduce reporter bias.Although HAMD-7 is a widely used scale for the assessment of depression, future research should use related scales and also incorporate interview-based measures to measure the broad spectrum symptoms of bipolar depression.

Conclusion
This proof of concept study demonstrated that regardless of effective reduction of proinflammatory markers, PTX does not have overall efficacy in TRD bipolar patients.Nevertheless, novel interactions between treatment response and inflammation status were revealed.PTX-treated patients with higher baseline CRP levels showed more improvement in depression symptoms, while the effect in low baseline CRP level patients was worse than placebo.While other inflammatory indicators in PTX and placebo-responders were typically high, baseline CRP in PTX-responders was considerably greater than in non-responders.Overall, PTX showed antiinflammatory benefits and may be more beneficial in a subset of patients with hyper-inflammation status, although further studies are required to address the anti-depressant efficacy of PTX in these patients.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
J o u r n a l P r e -p r o o f
54).Nevertheless, a significant effect of time with a reduction in HAM-D-17 scores was revealed in both groups (χ 2 = 2.11, P= .002)(Fig2).Regarding clinical and sociodemographic J o u r n a l P r e -p r o o f covariates, no significant time × treatment interaction was found.In contrast, by adding CRP to the model, a significant time × treatment × CRP interaction was observed (χ 2 =3.16,P=0.016).

Fig 2 .
Fig 2. Change in HAM-D-17 scores from baseline to week 12 in patients randomly assigned

Fig 3 .
Fig 3. Correlation between changes in HAM-D-17 scores and baseline CRP level.A.

Fig 4 .
Fig 4. Change in HAM-D-17 scores from baseline to week 12 (PTX-Placebo) in subjects

Fig 5 .
Fig 5. Change in HAM-D-17 scores from baseline to week 12 in PTX and placebo groups A.

Fig 6 .
Fig 6.Median change in individual HAM-D-17 items from baseline to week 12 in PTX Vs.

Fig 7 .
Fig 7. Changes in the response rate within the subgroup of patients with baseline CRP≤7.1

Fig 8 .
Fig 8. Change in the serum concentration of the selected inflammatory markers.A. Change between a patient's decreased pro-inflammatory profile and a change in HAM-D-17 score.Recently, our group has evaluated the PTX as an Add-on Therapy for MDD Patients demonstrating a significant reduce in HAM-D-17 score in the patients who received PTX(55).Moreover, we observed a reduction in serum concentrations of pro-inflammatory factors and an increase in serotonin and BDNF in the patients who received PTX.In contrast with the results of current study J o u r n a l P r e -p r o o f on TRD bipolar I/II patients, a significant positive correlation between IL-1-β, TNF-α, IL-10, CRP, IL-6, and HAM-D-17 scores was observed either before or after treatment.Moreover, BDNF and serotonin levels were negatively correlated with HAM-D-17 scores before and after treatment.In line with this, Farajollahi-Moghadam et al have evaluated the efficacy of pentoxifylline combined with sertraline in the treatment of MDD.They have reported that the sertraline plus pentoxifylline improves HAM-D scores and response to treatment rate supporting our previous study on MDD(56).The results found in the current study on TRD bipolar I/II patients are in contrast with the above mentioned results showing no significant improvement in depressive symptoms and no correlation between patient's decreased pro-inflammatory profile and a change in HAM-D-17 score.

Table 1 . Sociodemographic and clinical characteristics of the study sample.
Marital status -no.(%)