Impact of Relapse in BDNF Receptors Expression in Patients With a First Episode of Schizophrenia

Abstract

Background and Hypothesis

Relapsing after a first episode of schizophrenia (FES) is a main predictor of clinical and functional prognosis. Brain-derived neurotrophic factor (BDNF) plays a critical role in neuronal development and plasticity, and its signaling may be altered by successive relapses.

Design

We assessed the impact of first relapse in the expression of the 2 isoforms of the BDNF tropomyosin-related kinase B (TrkB) receptor (active full-length TrkB-F and inactive truncated TrkB-T) in peripheral blood mononuclear cells from 53 FES patients in clinical remission followed up for 3 years.

Results

The group of participants that relapsed (n = 24) during the follow-up presented a significant decrease in the expression of the active TrkB-F receptor compared to baseline (M = 100 ± 28.13 vs. M = 83.42 ± 33.84, t = 2.5, P = .02), with no changes in the inactive TrkB-T receptor expression nor in BDNF plasma levels. This decrease also led to a significant decline in the F/T ratio (M = 1.13 ± 0.38 vs. 0.94 ± 0.36, t = 2.17, P = .041). No significant differences were found in the receptors’ expression nor in plasma levels in the group of cases that remained in remission (n = 29). These results were not associated with baseline differences between the groups in terms of the BDNF signaling pathway biomarkers, clinical or treatment variables.

Conclusions

These findings highlight the biological impact that a relapse produces over the systemic BDNF-TrkB signaling pathway, potentially undermining crucial neuronal functions. Identifying the actors involved can help design specific interventions for relapse prevention and improve the functional prognosis of people in the early stages of schizophrenia.

Introduction

Up to 3% of the general population suffers a psychotic episode in their lifetime.¹ After the first episode of psychosis (FEP), the clinical evolution is variable, mediated by a heterogenous treatment response.^2,3 In the 5 years following a FEP, up to 80% of cases present a second episode,⁴ and about 20% of patients are considered resistant to treatment from the first episode of schizophrenia (FES).⁵ Relapses increase the risk of developing a chronic psychotic disorder with poor clinical and functional outcomes, so effective relapse prevention in the early stages is an important intervention objective.^2,6 To this end, the identification of the biological underpinnings associated with relapses in the FEP population is necessary.²

Among the actors that have been linked to the biological basis of psychotic disorders are neurotrophins.^7,8 Neurotrophins, such as brain-derived neurotrophic factor (BDNF), play a crucial role in numerous neuronal processes within the central nervous system (CNS), including differentiation, maturation, and synaptic function.⁹ Research on schizophrenia and BDNF has advanced significantly in the last 20 years, particularly on its role as a neurobiological marker for understanding the pathogenesis, clinical trajectories, and treatment monitoring.^7,10–14 According to a recent network meta-analysis, BDNF plasma levels are decreased in patients with schizophrenia compared with controls and patients with bipolar disorder.⁸ Previous studies have also demonstrated an abnormal expression of BDNF and its receptor in the prefrontal cortex and corticolimbic system in postmortem brain samples from patients with schizophrenia.^15–17 Besides, decreased BDNF levels in cerebrospinal fluid of drug-naive first-episode psychotic has been directly correlated with plasma BDNF levels.¹⁸ Our group reported that BDNF plasma levels cannot serve as reliable state biomarkers of relapse after remission of an FES, pointing out that BDNF plasma levels may be considered a useful biomarker of long-term severity in schizophrenia and of the underlying illness traits, specially of negative symptomatology severity.¹⁹ In addition, numerous pre-clinical studies with different animal models have provided evidence that BDNF plays an important role in the pathophysiology of schizophrenia, pointing to the negative and cognitive symptom domains.^20–25 Finally, numerous studies have shown the effects of antipsychotic drugs on BDNF signaling, both in animal models and in clinical samples,^26–28 and a recent systematic review and meta-analysis showed a significant increase in serum but not in plasma BDNF concentrations after antipsychotic treatment.²⁹

BDNF has been shown to have the ability to activate tropomyosin-related kinase B receptors (TrkB), its own tyrosine kinase-type receptor.³⁰ Two forms of TrkB receptors have been described: an active full-length (TrkB-F) and a truncated (TrkB-T). TrkB-T lacks kinase activity and inhibits the TrkB-F function by competing for binding to BDNF.³¹ Because of this, TrkB-T is a negative effector of TrkB-F in the developing brain, and an excess of the TrkB-T1 isoform can lead to neuronal death.³² Our group reported systemic changes in the expression of BDNF receptors in peripheral blood mononuclear cells (PBMCs) in FEP patients during the first year after diagnosis.³³ Specifically, we observed an increase of the active TrkB-F receptor expression along with a decrease in TrkB-T1, while those patients in the lower quintile of the TrkB-F/TrkB-T receptor ratio associated a lack of antipsychotic treatment response. Besides, increases in 2 TrkB-T isoforms have been found in the prefrontal cortex of subjects with schizophrenia.³⁴ The expression of trkB was found increased in the piriform cortex and the striatum of Poly (I:C)-induced maternal immune activation rat offspring.³⁵

Despite all this evidence, there is a lack of studies on the impact that a second episode may have on the expression of these receptors in subjects who have remitted from an FES. The objective of the present study was to assess the potential impact of relapse in BDNF TrkB receptors expression in PBMCs in a cohort of FES patients during a 3-year follow-up period.

Methods

Study Setting and Participants

Participants came from the 2EPs Project, a naturalistic, multicenter study aimed to identify factors related to relapse in a cohort of patients with a first-episode non-affective psychosis in remission. After inclusion, participants were closely followed up for 3 years or until relapse, whichever happened first. The background, rationale, study design, and the clinical and treatment predictors of relapse identified were previously reported.^2,3

Briefly, the 2EPs Project inclusion criteria were: (1) age between 16 and 40 years; (2) meet DSM-IV-TR diagnostic criteria for schizophrenia or schizophreniform disorder³⁶; (3) fulfill the Remission in Schizophrenia Working Group (RSWG) criteria after having remitted from a first psychotic episode (which should have occurred within the previous 5 years before inclusion in the study)³⁷; (4) not having presented a previous relapse after remitting from the first episode; (5) Spanish spoken correctly; and (6) written informed consent.

Subjects with intellectual disability (defined by an estimated Intelligent Quotient < 70, together with malfunctioning and difficulties with adaptive process), a history of head trauma with loss of consciousness or with an organic condition with mental repercussions were excluded.

During the recruitment period (October 2012–December 2015) every subject who met the inclusion/exclusion criteria that was attended at one of the participating sites was invited to participate in the study.² Relapse was defined when a participant stopped fulfilling the RSWG remission criteria at any time during the 3-year follow-up. This meant scoring 4 or more during at least 1 week in any of the following 8 items of the Positive and Negative Syndrome Scale (PANSS)³⁸: delusions, unusual thought content, hallucinatory behavior, mannerisms/posturing, blunted affect, social withdrawal, lack of spontaneity. Hospitalization was also recorded in every follow-up visit, being considered a relapse when they were related to schizophrenia symptoms (and not to other causes). After inclusion, follow-up visits to detect relapses were scheduled every 3 months, where information was collected from the entire period between visits. Participants, family members, caregivers, or reference clinical teams could notify to the research team of an eventual relapse of a participant at any time. In case of relapse, the relapse evaluation and sample collection were performed as soon as possible, generally within hours or a few days after its detection by the reference clinical team.^2,3 In case of non-relapsing, participants were followed up for 3 years, with a clinical assessment and sample collection at the end of this period.

The study was approved by the investigation ethics committees of all centers. Informed consent was obtained from all participants or from parents or legal guardians in under-age subjects.

Clinical Assessment

Upon entering the study, personal and family medical histories were collected in a systematic interview, including semi-structured diagnostic interviews adapted to the patients’ age (SCID-I and II or K-SADS).^39–41 Besides, clinical and functionality scales were administered by expert clinicians at every study visit.

Psychotic symptoms were assessed using the validated Spanish version of the PANSS scale.^38,42 Depression symptoms were assessed with the Spanish-validated version of the Montgomery-Asberg Depression Rating Scale (MADRS).^43,44 The Global Assessment of Functioning Scale (GAF),⁴⁵ the Functioning Assessment Short Test (FAST),⁴⁶ and the Clinical Global Impression for Schizophrenia (CGI-SCH)⁴⁷ were also scored to assess the global clinical and functioning status.

Vital signs and anthropometric measures were recorded on every visit, together with a systematic register of drug misuse habits using a part of the European adaptation of a multidimensional assessment tool European Addiction Severity Index (EuropAsi).⁴⁸ Samples for urine drug test detection were also collected on every visit.

Considering the naturalistic design of the study, participants received psychopharmacological treatment and psychosocial interventions based on clinical needs. The pharmacological, psychological treatment, and adverse drug reactions were recorded at every visit. Prescribed daily doses of antipsychotics were converted to an estimated equivalent amount of chlorpromazine to be able to compare them, following the method proposed by Leucht and collaborators.⁴⁹ Adherence to treatment was evaluated at each visit by the research team,² who used the Morisky-Green-Levine scale and consulted all available clinical information and external informants to identify those patients with poor therapeutic compliance.⁵⁰

Sample Collection and Biochemical Measurements

We collected 10 mL of anticoagulated venous blood samples after overnight fasting at baseline and the final visits (3-year follow-up visit or relapse visit). Blood tubes were immediately centrifuged (641 g × 10 min). The separated resultant plasma samples were stored at − 80 °C. The rest of the sample was 1:2 diluted in Roswell Park Memorial Institute Medium (RPMI) 1640 (LifeTech) culture medium. A gradient with Ficoll-Paque (GE Healthcare) was used to isolate PBMC by centrifugation at room temperature (800 g × 40 min). The PBMC layer was aspirated and resuspended in the culture medium, then centrifuged at room temperature (1116 g × 10 min). Once the supernatant layer was removed, the PBMC-enriched pellet was stored at − 80 °C at each site facilities. Once the whole recruitment was over, all samples were sent to the same laboratories for subsequent determinations.

Protein levels of TrkB-F and TrkB-T BDNF receptors in PBMCs were quantified by Western Blot analysis, following the methodology used by our group in previous studies.³³ In brief, 15 μg of cytosolic extracts were loaded onto electrophoresis gels. Protein samples were separated and transferred onto a nitrocellulose membrane (Transfer Pack, Biorad®). After blocking, membranes were incubated with specific antibodies: (1) TrkB-F, rabbit polyclonal antibody dilution of 1:750 in BSA 1% (sc12, SCB); (2) TrkB-T, rabbit polyclonal antibody dilution of 1:1000 in BSA 1% (ab18987, abcam); (3) β-actin mouse monoclonal in a dilution 1:10000 (A5441, Sigma). Proteins were then recognized by the respective horseradish peroxidase-linked secondary antibodies and visualized using an Odyssey® Fc System (Li-COR Biosciences®) and quantified by densitometry (NIH ImageJ® software).

All densitometry readings were initially obtained in arbitrary units of optical density and normalized to β-actin to account for loading differences across samples. To minimize potential variability due to experimental conditions, all samples from baseline and follow-up visits for each patient were analyzed concurrently and run on the same gel. All western blots were performed at least 3 times in separate assays to confirm reproducibility. To further enhance comparability across triplicates, the β-actin-normalized protein levels were transformed into percentages, using the baseline measurement as a reference. The statistical analysis was therefore performed on these percentage values to ensure that our results were accurate and comparable, minimizing the impact of experimental variability.

Given the counterbalancing effect of TrkB-T and TrkB-F, we decided to choose the ratio of TrkB-F to TrkB-T expression (hereafter F/T ratio) as our index variable for describing BDNF receptor expression.

Plasma levels of BDNF were determined using enzymatic assays (ELH-BDNF-1, RayBiotech®), according to the manufacturer’s instructions and as previously reported.¹⁹ All measurements were made in duplicate.

Statistical Analysis

The main objective of the analysis was to evaluate whether there were differences in the expression of TrkB-T and TrkB-F receptors, the F/T ratio values and BDNF plasma levels in patients who presented a psychotic relapse compared to those who maintained remission.

Baseline demographic, clinical, and treatment characteristics were compared between patients who relapsed and those who remained in remission during the follow-up were analyzed to better characterize subgroups and to identify potential sociodemographic or illness-related differences that might influence analyses on BDNF signaling pathway markers. Categorical variables were compared with Chi-square tests.

The differences on continuous variables with approximately normal distributions were assessed using a two-tailed Student’s t-test. A two-tailed nonparametric Mann-Whitney U test was used when continuous variables did not meet the assumption of normality in the Kolmogorov–Smirnov test (with Lilliefors correction). As this analysis involved multiple comparisons, we used the Bonferroni correction procedure (dividing the alpha level by the number of tests; 0.05/31 = P ≤ .001) to make our criterion of significance more conservative.

A paired-sample t-test was conducted to evaluate the effect of a relapse on BDNF receptors expression, the F/T ratio values, and BDNF plasma levels.

Finally, to determine whether any clinical variable was significantly associated with the expression of the BDNF signaling markers at relapse, we calculated Pearson correlation coefficients between BDNF receptors expression, the F/T ratio values, and BDNF plasma levels and clinical assessments (PANSS, MADRS, CGI-SCH, and GAF total scores), substance use (tobacco, alcohol, cannabis and stimulant use), total doses of antipsychotic treatment (in mg/day chlorpromazine equivalents) and body mass index in the relapse evaluation visit. Again, we applied a Bonferroni correction and established a P-value of ≤ .001 (0.05/40).

Data were managed and analyzed with the IBM SPSS Statistics v.29. A value of P < .05 was taken to be statistically significant in all analyses, except where noted.

Results

Fifty-three participants included in the 2EPs cohort had available data for all biomarker measurements and were included in the present study. Baseline sociodemographic, clinical, psychopathological, anthropometric, psychopharmacological, and substance use of the studied cohort are shown in Table 1, differentiating the patients who had presented a relapse during the 3-year follow-up (n = 24, 45.3%) of those who had not (n = 29, 54.7%). We did not find significant differences between both groups in the main sociodemographic and clinical variables at baseline except for the proportion of benzodiazepines use (20.8% in relapsed vs. 3.4% in non-relapsed, P = .047), but these differences did not reach the level of statistical significance when applying the Bonferroni’s correction procedure for multiple comparisons (P ≤ .001), see Table 1 for details.

Table 1.

Baseline Demoraphic and Clinical Characteristics Between Relapsed and Non-Relapsed Patients

	Relapsed (n = 24)	Non-relapsed (n = 29)	Statistic	P-value
Sex (female)—no. (%)	9 (37.5%)	15 (51.7%)	χ2 = 1.07	.30
Education level (< 10 years)—no. (%)	11 (45.8%)	13 (54.2%)	χ2 = 0.01	.94
Psychiatric comorbidities (yes)—no. (%)	4 (16.7%)	4 (13.8%)	χ2 = 0.09	.77
Personal psychiatric history (yes)—no. (%)	6 (25%)	6 (24%)	χ2 = 0.01	.93
Family psychosis history (yes)—no. (%)	4 (26.7%)	4 (28.6%)	χ2 = 0.01	.91
Obstetric complications (yes)—no. (%)	7 (29.2%)	7 (24.1%)	χ2 = 0.17	.68
Age—years [mean (sd)]	25 (5.4)	26.07 (5.9)	t = −0.67	.11
Age at psychosis onset—years [mean (sd)]	24.17 (5.4)	24.69 (6.1)	t = −0.33	.50
Duration of untreated psychosis—days [mean (sd)]	226.55 (381)	130.31 (214.2)	U = 294.5	.64
PANSS—total score [mean (sd)]	48.63 (16.2)	49.34 (14.8)	t = −0.17	.88
MADRS—total score [mean (sd)]	4.88 (5.4)	5.79 (5.8)	t = −0.59	.56
CGI-S—total score [mean (sd)]	3.08 (1.2)	3.21 (1.3)	t = −0.36	.72
GAF—total score [mean (sd)]	69.3 (15.7)	68.9 (15.3)	t = −0.1	.92
FAST—total score [mean (sd)]	21.04 (16.3)	19.83 (16.1)	t = −0.27	.79
Systolic blood pressure—mmHg [mean (sd)]	121.96 (16.6)	119.93 (13.8)	t = 0.48	.63
Diastolic blood pressure—mmHg [mean (sd)]	71.09 (9.7)	71.79 (10.3)	t = −0.25	.80
Body mass index—kg/m2 [mean (sd)]	24.5 (5.5)	26.76 (5.8)	t = −1.41	.16
Abdominal circumference—cm [mean (sd)]	87.88 (9.5)	92.71 (15.2)	t = −1.35	.18
Antipsychotic treatment (yes)—no. (%)	22 (91.7%)	24 (82.8%)	χ2 = 0.91	.34
Antipsychotic mean dosea—mg/day (sd)	358.45	279.37	U = 321.5	.63
Clozapine (yes)—no. (%)	(357.2)	(235.38)	χ2 = 3.13	.08
Anticholinergics (yes)—no. (%)	1 (4.2 %)	6 (20.7%)	χ2 = 0.06	.80
Antidepressants (yes)—no. (%)	2 (8.3%)	3 (10.3%)	χ2 = 3.18	.07
Mood stabilizers (yes)—no. (%)	2 (8.3%)	8 (27.6%)	–	–
Benzodiazepines (yes)—no. (%)	4 (16.7%)	0 (0%)	χ2 = 3.95	0.047*
Pre-study antipsychotic treatment duration (days) [mean (sd)]	5 (20.8%) 509,24 (395,35)	1 (3.4%) 311 (290,75)	U = 446	.08
Tobacco use (yes)—no. (%)	14 (58.3)	15 (48.3)	χ2 = 0.53	.47
Tobacco monthly use—cigarettes [mean (sd)]	327.5 (250.7)	341.43 (147.2)	t = −0.18	.86
Alcohol use (yes)—no. (%)	11 (45.8%)	13 (44.8%)	χ2 = 0.01	.94
Alcohol monthly use—SDU [mean (sd)]	23.09 (31.6)	25.92 (23.7)	t = −0.25	.80
Cannabis use (yes)—no. (%)	5 (20.8%)	3 (10.3%)	χ2 = 1.13	.29
Cannabis monthly use—cigarettes [mean (sd)]	38 (36.6)	70 (65.6)	t = −0.91	.40
Cocaine use (yes)—no. (%)	1 (4.2%)		–	–
Cocaine monthly use—grams [mean (sd)]	1	0	–	–
Other stimulants use (yes)—no. (%)	1 (4.2%)	0	–	–
Other stimulants monthly use—units [mean (sd)]	3	2 (6.9%) 3.5 (3.54)	–	–

Abbreviations: CGI-S: Clinical Global Impression for Schizophrenia; FAST: Functional Assessment Short Test; GAF: Global Assessment of Functioning Scale; MADRS: Montgomery-Asberg Depression Rating Scale; PANSS: Positive and Negative Symptom Scale; SDU: Standard drink units. ^aChlorpromazine equivalents. *P < .05.

There were no significant differences at baseline in BDNF receptors expression, the F/T ratio values and BDNF plasma levels, meaning that differences found between groups did not result from different starting points (see Table 2).

Table 2.

Baseline BDNF Receptors Expression, Receptors Activity Ratio, and Plasma Levels Comparisons Between Relapsed and Non-Relapsed Patients

Biomarker	Relapsed (n = 24)	Non-relapsed (n = 29)	Statistic	P-value
BDNF TrkB full-length receptors expression—% from control [mean (sd)]	101.24 (28.48)	97.02 (24.72)	t = 0.58	.57
BDNF TrkB truncated receptors expression—% from control [mean (sd)]	95.76 (29.43)	95.03 (26.95)	t = 0.09	.93
Full-length/Truncated receptors ratio	1.13 (0.38)	1.07 (0.33)	t = 0.60	.55
BDNF plasma levels—ng/mL [mean (sd)]	9301.48 (7456.96)	9280.27 (6980.49)	t = 0.01	.99

Abbreviations: BDNF: brain-derived neurotrophic factor. TrkB: Tropomyosin-related kinase B, *P < .05. All densitometry results are expressed as a percentage of the control.

A paired-sample t-test was conducted to evaluate the impact of relapse on BDNF receptors expression, the F/T ratio values and BDNF plasma levels. In the group of patients that relapsed, there was a statistically significant decrease in the expression of TrkB-F receptors from baseline (M = 101.24, SD = 28.48) to relapse (M = 83.42, SD = 33.84), t (24) = 2.5, P = .02. The mean decrease in TrkB-F receptor expression was 16.58 with a 95% confidence interval (CI) ranging from 2.89 to 30.29. The Cohen’s d statistic (0.51) indicated a medium effect size (see Figure 1).

Figure 1.

A: Neurotrophin Receptor Expression of Active Full-Length (F) TrkB Receptor; B: Neurotrophin Receptor Expression of Truncated (T) TrkB Receptor; C: F/T Ratio; D: Plasma Levels of BDNF in PBMCs From Remitted First Episode of Schizophrenia Psychosis Patients at Baseline and at Endpoint (Relapse or 3-Year Follow-Up). *P < .05 Endpoint vs. Baseline in Paired-Sample t-Tests.

As expected from the previous results, there was also a statistically significant decrease of the F/T receptors ratio from baseline (M = 1.13, SD = .38) to relapse (M = 0.94, SD = .36), mean decrease 0.19, 95% CI [0.009, 0.37], t (24) = 2.17, P = .041. The Cohen’s d statistic (0.44) also indicated a medium effect size.

Interestingly, no significant differences were found in the group of cases that remained in remission for 3 years of the study in the expression of any of the receptors. None of the 2 groups showed significant differences in plasma levels of BDNF.

In the evaluation visit of patients who relapsed during the follow-up, none of the clinical assessment scores (PANSS, MADRS, CGI-SCH, FAST, and GAF), substance use measures (tobacco, alcohol, cannabis, and stimulants use), total doses of antipsychotic treatment (in mg/day chlorpromazine equivalents) or body mass index was significantly associated with the expression of the BDNF receptors, the F/T ratio values, or the BDNF plasma levels.

We analyzed antipsychotic treatment characteristics at the endpoint visit (see Table 3 for details). Since patients who relapsed during the follow-up left the study, it was expected that the duration of treatment after inclusion would be shorter in this subgroup compared to those who completed the entire follow-up. In the rest of the parameters, no significant differences were found between the groups.

Table 3.

Antipsychotic Treatment Characteristics at the Endpoint Visit

	Relapsed (n = 24)	Non-relapsed (n = 29)	Statistic	P-value
Antipsychotic treatment (yes)—no. (%)	12 (50)	18 (62.1)	χ2 = 0.39	.42
Antipsychotic treatment duration after inclusion (months) [mean (sd)]	13.73 (9.62)	27.83 (14.45)	U = 534.5	< .01*
Antipsychotic total mean dosea—mg/day (sd)	416.3 (355.16)	276.62 (282.27)	t = 1.56	.12
Clozapine (yes)—no. (%)	3 (13)	7 (31.8)	χ2 = 2.29	.16
Long-acting Injectable Antipsychotic (yes)—no. (%)	4 (17.4)	8 (36.4)	χ2 = 2.1	.19

^aChlorpromazine equivalents.

^* P < .05.

Discussion

In this study, we found evidence of systemic changes in the expression of BDNF TrkB receptors signaling pathway in a group of FES patients who presented a relapse but not in those maintaining remission during a 3-year follow-up. Specifically, presenting a second episode was associated with a decrease in the expression of the active isoform TrkB-F receptor in PBMCs. The second episode was not associated either with changes in the expression of the truncated TrkB-T receptor or in BDNF plasma levels. This decrease also led to a decline in the value of the F/T ratio. These findings were not due to baseline differences between the groups in terms of the BDNF signaling pathway biomarkers, clinical, or treatment variables.

Multiple evidence points to the negative impact on long-term outcomes of relapsing after presenting a FEP. Thus, successive episodes increase the risk of developing a chronic psychotic disorder with clinical and functional deterioration, brain tissue loss, and worse treatment response.^51,52 Relapses are also associated with premature mortality in this population,⁵³ in part due to an elevation of suicidal behaviors.^54,55 Relapses are also implicated in an increased risk of aggressive or violent behavior.⁵⁶ Multiepisodic forms are also associated with higher socioeconomic direct and indirect costs.⁵⁷ Altogether, these data are the bases that support recommending antipsychotic long-term maintenance treatment for relapse prevention,⁵⁸ reinforced by naturalistic studies that have followed national cohorts for years and decades.^59,60

To the best of our knowledge, this study shows for the first time the impact of a second episode on the BDNF signaling pathway, a key actor for the neuron normal development, function, and survival.⁹ While patients who remained in remission after the FES showed no change in BDNF receptor expression, those who relapsed showed a decrease in the expression in the functional isoform of the BDNF receptor. Even if, consistent with previous studies, BDNF plasma levels did not change significantly in patients experiencing a relapse,^19,61 decreased BDNF receptor expression could result in reduced BDNF activity, potentially compromising normal neuronal viability, connectivity, and short- and long-term plasticity, thereby affecting synaptic regulation.⁶² Seeing this BDNF signaling dysregulation in patients that presented a second episode, it is possible that subsequent relapses could lead to lower plasmatic BDNF levels in more advanced disorder stages. These mechanisms could underlie progression of the disorder and resistance to treatment, which are common signs of multiepisodic forms of schizophrenia.⁶³

These results are also in line with our previous findings in which we described that those patients in the lower quintile of the F/T ratio associate poorer antipsychotic treatment response,³³ which could be closely linked to higher relapse risk. The peripheral TrkB receptor down-regulation showed in relapsed FES subjects from our study also goes in the same direction as the reduced expression of the total TrkB receptor (without differentiating between TrkB-F and TrkB-T) described in the corticolimbic structures of postmortem brain samples from subjects with chronic, multiepisodic schizophrenia.¹⁶ Significant increases in mRNA expression of 2 TrkB-T isoforms have been described in the dorsolateral prefrontal cortex of people with schizophrenia compared to controls, also pointing to a brain overexpression of truncated isoforms in chronic, multiepisodic forms of schizophrenia.³⁴ A decreased TrkB signaling appears to underlie the dysfunction of inhibitory parvalbumin GABA neurons in the prefrontal cortex of subjects with schizophrenia.⁶⁴ Taken together with our results, all these findings point to an abnormal BDNF-TrkB signaling pathway associated with chronic forms of schizophrenia, probably starting from the first relapse. A decrease in the activity of the receptor’s functional isoform could impair some of the crucial functions of BDNF in neuronal differentiation, maturation, and synaptic function, implying a worse clinical evolution, typical of subjects with multiple episodes. The decrease in peripheral expression of TrkB-F with the first relapse could be related to this worse evolution linked to multiepisodic forms.

In the group of patients that presented a relapse, the decreased TrkB-F expression and F/T ratio values were not associated with any of the clinical assessment scores (PANSS, MADRS, CGI-SCH, and GAF), substance use measures (tobacco, alcohol, cannabis, and stimulant use), total doses of antipsychotic treatment (in mg/day chlorpromazine equivalents) or body mass index at relapse evaluation. Thus, more work is necessary to clarify a potential link between the BDNF-TrkB signaling and antipsychotic effects in schizophrenia patients,⁶² where precise spatial and temporal relevance of BDNF actions may play a crucial role to achieve the desired treatment outcomes.

As measuring BDNF levels in the human brain (central BDNF) is challenging,⁶⁵ it is worth pointing out that the results reported in this study come from measuring peripheral receptors expression and plasma levels as a proxy from the CNS. This method is supported by animal studies that demonstrate that BDNF can cross the blood-brain barrier in both directions, showing a correlation between peripheral and central BDNF levels,⁶⁶ and supporting the idea that peripheral BDNF levels could influence and reflect CNS BDNF activity.⁶⁷ Studies in animal models have also demonstrated that stress-induced reductions in brain BDNF are often mirrored by reductions in serum BDNF.⁶⁸ This strong correlation between serum/blood and brain BDNF levels in specific regions is also indirectly supported in different studies through neuroimaging techniques.⁶⁹ Since to date, it is not yet possible to get a direct measure of BDNF central levels via neuroimaging techniques, previous studies have correlated peripheral BDNF levels to certain brain structures volume in patients with schizophrenia.^70–72 There is also evidence that systemic inflammation can affect both peripheral and central BDNF levels. In this line, increased levels of inflammatory cytokines have been associated with reduced BDNF levels in both the blood and the brain.⁷³ Finally, more recent studies have demonstrated that physical activity increases BDNF levels both peripherally and centrally, showing that the peripheral increases in BDNF are paralleled by enhanced cognitive function and neurogenesis in the brain.⁷⁴ These findings suggest a possible functional, albeit indirect, relationship between the expression of TrkB receptors at the peripheral and central levels. Although there are indications that peripheral BDNF expression may reflect BDNF brain activity to some extent, the specific correlation between the expression of its TRKB receptor in both peripheral and central compartments requires further investigation. Several studies have measured separately the TrkB expression in PBMCs and in the CNS in different neuropsychiatric conditions, such as depression or Parkinson disease,^75,76 but a direct comparative study measuring TrkB expression levels in both compartments within the same experiment has not been yet reported.

It is also worth noting that some characteristics of the 2EPs project may be influencing the fact that the sample of participants included may be different in certain aspects from other with first psychotic episodes samples, for example with higher patients treated with clozapine (7 out of 53 [13.2%] at baseline; 10 out of 53 [18.8%] at endpoint). There are probably several explanations for these higher rates of clozapine use. First, the 2EPs project’s inclusion criteria required that participants met the remission criteria proposed by the RSWG after their FEP to be included in the study, so some participants had already needed to switch to clozapine for achieving remission from the first episode, prior to their inclusion in the project. Second, unlike other studies that allow the inclusion of patients with affective disorders with psychotic features, psychosis associated with the consumption of toxic substances or brief psychoses, only patients with a diagnosis of schizophrenia or schizophreniform disorder could be included in our study, which is also associated with a higher probability of ending up receiving this treatment. Finally, another possible explanation is the fact that the participating centers were tertiary centers, which usually tend to receive more severe or earlier-onset patients from other teams, and often find fewer barriers to starting clozapine treatments.

There are some limitations in this study that should be taken into consideration when analyzing these results. Firstly, the necessary biological samples were only obtained from 53 of the 223 participants initially included in the 2EPs Project, which may have limited the capability to detect differences between groups (relapsed vs. non-relapsed) at the end of the study. This low percentage of sample collection was mainly due to the technical difficulties of some participating centers in obtaining the PBMCs and to early dropouts in both groups.³ No significant clinical (including the % of relapse), demographic or treatment differences between the participants who were included in this analysis and those who could not enter. Secondly, all the BDNF signaling determinations were measured in peripheral blood samples, which may limit to reflect what is happening in the brain. Thirdly, all participant sites were tertiary university hospitals, so patient samples and therapeutic strategies may differ from those used in other areas. Finally, the 2EPs project was a naturalistic study, not a randomized controlled trial, so patients could be changing treatments during the follow-up period according to the clinician’s choice. In addition, total time exposure to antipsychotic was also different between groups, since patients who relapsed during the follow-up left the study. This significant differences in antipsychotic treatment duration between groups could be an alternative explanation for our findings, beyond the direct effect of that a relapse may cause on the systemic BDNF-TrkB signaling pathway. Besides, it is important to bear in mind the inherent difficulties in ensuring therapeutic compliance in this population, even though this study was focused on characterizing relapses, in which researchers recorded and measured adherence to treatment, also consulting the available clinical information and external informants. In short, it is worth considering these limitations regarding the effect that antipsychotic treatment could have on BDNF signaling.

Despite these limitations, our study also has strengths to highlight. Firstly, the study population was very homogeneous at the beginning (FES in remission), unlike other studies with first psychotic episodes, that include patients with affective disorders or with differing clinical status at baseline. Secondly, all participants were in an early stage of the disorder (first 5 years), which helps to avoid confounding variables such as prolonged antipsychotic exposition, comorbidities, or chronicity.⁷⁷ Thirdly, considering the naturalistic design, our sample is representative of the real-world population with an FES in our context. Finally, we used a very comprehensive diagnostic and clinical assessment protocol, with strict inclusion-exclusion criteria, a long period of follow-up and deep clinical and biological characterization.²

In conclusion, these findings highlight the biological effects of a psychotic relapse on the systemic BDNF-TrkB signaling pathway, potentially negatively influencing crucial BDNF functions at the neuronal level. Future replication studies in larger cohorts incorporating the different BDNF signaling pathway biomarkers are needed to better understand the biological correlates that produce relapses in psychotic patients on this pathway. It will also be necessary to better understand the role of certain potentially confounding factors on this pathway, such as the use or discontinuation of antipsychotic treatment or the differences between different antipsychotics such as clozapine. Identifying the factors involved can help design specific interventions for relapse prevention and improve the functional prognosis of people with schizophrenia.

Author Contributions

MBe obtained funding for the study. MBi conceptualized and proposed the present investigation. GM performed the project administration. KSM and AMP made the laboratory analysis. MBi and KSM acquired, analyzed, and interpreted the data, drafted the original article created tables and figures, and critically revised the manuscript for intellectual content. All authors have participated in the recruitment. All authors have read, reviewed, edited, and approved the final manuscript.

Funding

This study is part of the coordinated-multicentre project “Clinical and neurobiological determinants of second schizophrenia episodes. Longitudinal study on first-episode psychosis” (2EPs Project), funded by the Ministerio de Economía y Competitividad, Instituto de Salud Carlos III (PI11/00325).

Conflicts of Interest

Dr Amoretti has been a consultant to and/or has received honoraria/grants from Otsuka-Lundbeck, with no financial or other relationship relevant to the subject of this article. Dr Bernardo has been a consultant for, received grant/research support and honoraria from, and been on the speakers/advisory board of Abartis Pharma, ABBiotics, Adamed, Angelini, Casen Recordati, Esteve Pharmaceuticals, Janssen-Cilag, Menarini, Rovi, and Takeda. Dr Bioque has been a consultant for, received grant/research support and honoraria from, and been on the speakers/advisory board of has received honoraria from talks and/or consultancy of Adamed, Angelini, Casen Recordati, Exeltis, Ferrer, Janssen, Lundbeck, Neuraxpharm, Otsuka, Pfizer, Rovi, and Sanofi. Dr Díaz-Caneja has received honoraria from Angelini and Viatris, and travel support from Janssen and Angelini. Dr Ibañez has received research support from or served as speaker or advisor for Janssen-Cilag, Lundbeck, Otsuka Pharmaceutical SA, Alter and Rovi, with no financial or other relationship relevant to the subject of this article. Dr Mané has received funds from Otsuka, Lundbeck, Janssen, Neuraxpharm, and Angelini as a speaker and for congress registration and travel expenses. Dr Mezquida has been a consultant to and/or has received honoraria/grants from Boehringer Ingelheim, with no financial or other relationship relevant to the subject of this article. Dr Roldán has been advisor or speaker for the companies Otsuka, Angelini, Rovi, and Casen Recordati (unrelated to the present work). Dr Sáiz has been a consultant to and/or has received honoraria or grants from Adamed, Alter Medica, Angelini Pharma, CIBERSAM, Ethypharm Digital Therapy, European Commission, Government of the Principality of Asturias, Instituto de Salud Carlos III, Johnson & Johnson, Lundbeck, Otsuka, Pfizer, Plan Nacional Sobre Drogas, and Servier. The rest of authors declare no conflict of interest.