The major depressive disorder (MDD), or depression, is characterized by the loss of pleasure or interest in everyday activities, as well as other characteristics, including changes in sleep and appetite, low motivation, and suicidal ideation or action (Organization American Psychiatric, 2013; Otte et al., 2016). Current conventional antidepressants, mainly selective serotonin reuptake inhibitors (SSRIs), only have moderate rates of response and remission, with some unwanted side effects (Kogoj, 2014). Usually, long-term treatment period is required before the antidepressant effects of SSRIs can be elicited symptomatically (Agius and Bonnici, 2017). As depression is a progressive disease, the long lag period of the efficacy may lead to the miss of the optimal window for the intervention (Lane, 1998). Therefore, detection of early markers that can predict the treatment outcome is important for the improvement of the treatment. In the past decades, great efforts have been made to find the potential early biomarkers, including neurotropic factors, neurotransmitters, cytokines, as well as other molecules in the blood (Nutt, 2008; Dowlati et al., 2010; Tavakolizadeh et al., 2018; Xu et al., 2019). However, up to date, none of the individual predictors is robust enough to guide first-line treatment options (Kessler, 2018).

Brain-derived neurotrophic factor (BDNF) is a neurotrophic factor critical for the viability of neurons in neural circuit (Monteggia et al., 2007). A number of studies have shown that BDNF is profoundly implicated in depression (Peng et al., 2018; Koo et al., 2019; Szuhany and Otto, 2020). Studies have shown that BDNF levels are lower in depression patients than healthy people (Karege et al., 2002). Decreasing BDNF expression leads to deficient neural plasticity in the brain of patients with MDD, which makes it a potential biomarker to monitor these diseases (Peng et al., 2018). BDNF in serum or plasma is increasingly used as a putative biomarker for depression, and some studies revealed that lower levels of BDNF in the blood of depressed patients are improved after long-term treatment of certain antidepressants (Shimizu et al., 2003; Gazal et al., 2012; Vanicek et al., 2019). However, a considerable number of studies demonstrated no effect or even opposite findings (Kheirouri et al., 2016; Ryan et al., 2018). Therefore, whether the changes of blood BDNF levels at the early time of antidepressants treatment alone predict the long-term treatment efficacy remains elusive (Kessler, 2018; Xu et al., 2019).

Traditional medicine has been used as an alternative monotherapy or an adjunct to conventional antidepressant therapy, whose efficacy and safety has been supported on the basis of increasing number of clinical trials (Zhang et al., 2007; Qin et al., 2011; Wang et al., 2012; Yang et al., 2020). Yueju pill is a multiherbal pill formulated 800 years ago to treat syndromes associated with mood disorders (Ren and Chen, 2017; Wang et al., 2019). It contains multiple antidepressant components and some of the bioactive compounds have been identified (Wei et al., 2008; Zhang et al., 2015; Wu et al., 2016). More recently, a number of preclinical studies using various animal models demonstrated that a single high dose of Yueju was capable to elicit antidepressant activity in a fast-onset and persistent manner (Xue et al., 2013), hippocampus is one of many regions that are involved in depression-related behavior (Ren et al., 2016; Xia et al., 2016). A pilot perspective random controlled clinic trial showed that the use of high dose of Yueju contributed to the early onset of the alleviation of depressive symptoms, starting at 4 days after treatment together with fluoxetine (Wu et al., 2015). Additionally, chronic treatment of conventional low dose of Yueju pill also elicited antidepressant activity, with overall superior efficacy to fluoxetine, in an animal model of depression (Zou et al., 2017). Both single high dose and repeated low dose of Yueju pill significantly induce hippocampal BDNF synthesis through transcriptional and translational mechanisms (Xue et al., 2013).

Here, in a clinical observation of the effect of low dose of Yueju pill treatment, we disclosed that the ratio of serum BDNF level to the scale of depression using the scores of HAMD-24, defined as neuroplasticity index (NI), may be useful for prediction of the antidepressant treatment outcome. We thus carried out a prospective random placebo-controlled clinical trial, in which the monotherapy of Yueju pill was compared with the mainstream antidepressant escitalopram. We assessed the correlation of baseline NI with the changes of NI at different treatment times. The results indicated significant correlations of NI or their changes at different times post treatment of Yueju as well as escitalopram. Additionally, the change of NI at the early time may be used for prediction of the long-term antidepressant efficacy.

In this study, we investigated whether the ratio of serum BDNF level and depression scale of HAMD-24, or NI, was useful for prediction of the long-term antidepressant outcome, using Yueju and/or escitalopram in two trials. In the preliminary open-labeled trial, we found both symptom alleviation and serum BDNF enhancement as early as 1 week post a conventional low dose of Yueju. We also revealed significant correlation between NIs at baseline and at the early treatment time. In the confirmatory random controlled double-blinded clinical trial, we found both escitalopram and Yueju pill resulted in early improvement of depression symptom, and comparable antidepressant outcome after 4 weeks of treatment. Unlike Yueju, escitalopram treatment failed to change BDNF levels overtime. However, baseline NI, early treatment NI and final treatment NI were all significantly correlated for both Yueju and escitalopram. Furthermore, the early NI was also correlated to the final enhancement of NI for Yueju in both trials. Finally, ROC analysis indicated the final antidepressant efficacy of Yueju treatment was predictable on the basis of early NI enhancement. The results suggest that the NI may represent an individual’s trait that links BDNF with depression symptom, and early change of NI may predict long-term antidepressant efficacy.

Brain-derived neurotrophic factor is a neurotrophin that has been linked to the viability of neurons in brain circuits for depression regulation (Molendijk et al., 2011), and it has been argued that BDNF is a key player for neural plasticity underlying depression and antidepressant activity (Warner-Schmidt and Duman, 2006). Several lines of studies demonstrated the reduced levels of BDNF in MDD patients (Dreimüller et al., 2012). Consistent with this, the present study showed lower serum BDNF levels in MDD patients. The serum BDNF levels of depression patients also were negatively correlated with the degree of depression, when the data from the two trials were collapsed. Furthermore, we found MDD can be reliably predicted with high sensitivity and specificity on the basis of serum BDNF levels. Despite the utility of BDNF for prediction of MDD, our data indicated that it was not useful for prediction of antidepressant efficacy (Dreimüller et al., 2012; Wilkinson et al., 2018). Unlike the increase of BDNF level post treatment of Yueju in both trials, BDNF levels were not changed in the escitalopram treatment group even escitalopram induced significant symptom improvement, consistent with a finding that escitalopram did not affect BDNF levels at either early and late time points (Matrisciano et al., 2009).

The recovery of neuroplasticity is considered to be an important neural mechanism for antidepressant activity (Castrén and Antila, 2017; Koopman and El Aidy, 2017; Franklin et al., 2018; Fox and Lobo, 2019). NI is used to characterize the relative contribution of neuroplasticity indicated by the BDNF level to depression symptoms indicated by the score of HAMD-24. The differences of NI in individuals may reflect different levels of sensitivity of depression symptoms to BDNF levels. NIs at the baseline, the early time of treatment and the final time of treatment were all correlated. Importantly, this correlation was found for both Yueju and escitalopram treatment, even though escitalopram group did not show increased BDNF levels or different doses of Yueju was used overtime. These suggest that NI may represent a trait constantly determining the role of BDNF in an individual’s depressive symptoms both at the disease baseline condition and following a certain antidepressant treatment. These intrinsic characteristics also led to the association between early or final treatment time NI with early or final NI enhancement, respectively. It is also noticeable that the measurement of HAMD-17 was not as sensitive as HAMD-24, which may indicate the importance of completeness of the depression symptom for NI. As HAMD-24 scaling and BDNF Elisa measurement are well developed techniques, NI can be reliably calculated and used in the clinical setting.

Partly due to the correlation of baseline NI with the NI at early or final treatment time, it was not surprising that early/final NI was significantly correlated with early/final NI enhancement for either Yueju or escitalopram treatment. This may contribute to the predictability with early NI enhancement to the final treatment efficacy of Yueju. The present study was unable to predict the treatment efficacy of escitalopram. This is technically ascribed to the insufficient number of subjects that met the criteria of treatment efficacy after 4 weeks of treatment of escitalopram. Other factors may also account for the difficulty: in contrast to the general increase of NI at both early and final treatment time of Yueju, NI was mostly reduced at the early treatment time, but augmented by the final treatment time of escitalopram. Consistently, we found the early treatment NI was correlated with the final NI enhancement for Yueju, but not for escitalopram. The definition of NI or related index may have some applications in testing other antidepressants. For example, ketamine is known for the neuroplasticity-dependent mechanism. For escitalopram or other 5-HT based antidepressants, we speculate to use the index like the ratio of the 5-HT improvement value to HAMD-24 score at somehow early stage, to predict the outcome from the contribution of early 5-HT response to behavioral improvement. Certainly, these warrant further investigations. These possibilities should be addressed in future studies by use of a variety of antidepressants for a longer treatment time with a larger number of MDD subjects.

In the present study, we also revealed that the clinical symptom improvement as early as 4 days after treatment of escitalopram. To the best of our knowledge, this finding was reported for the first time. It warrants further clinical and preclinical investigations as escitalopram has become one of the most frequently prescribed antidepressants for its relatively fast action among SSRIs (Pastoor and Gobburu, 2014). Consistent with the finding on combination of the high dose of Yueju pill with fluoxetine achieved early symptom improvement, the present study provided evidence directly showing this dose of Yueju pill by itself was capable to elicit early-onset antidepressant effects in patients (Wu et al., 2015). It appeared there was a larger attenuation of the depressive symptoms by Yueju, compared to escitalopram, although it did not reach the statistic significance. These are in agreement with the findings using different animal models that displayed an immediate and lasting antidepressant effect following a single high dose of Yueju pill (Tang et al., 2015). Additionally, the present study also showed the long-term antidepressant efficacy of conventional low dose of Yueju in both trials. As low dose of Yueju has been used safely over 800 years clinically, it may be useful for a long-term treatment that many times is required for prevention of recurrence of the disorder.

In conclusion, we demonstrate that NI characterized the role of neuroplasticity in antidepressant treatment. The change of the NI after a short period time of treatment may indicate how the neural plasticity contributes to symptom improvement, which is helpful for the prediction of long-term antidepressant efficacy. Whether this measurement can be generalized to other antidepressant treatment remains to be determined. Nonetheless, the present study indicates the putative composite biomarker at the early time to predict the long-term antidepressant response, which offers novel insights to develop effective biomarker for antidepressant treatment efficacy to improve the treatment outcomes.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the Institutional Review Board of The Fourth People’s Hospital of Taizhou, Jiangsu Province. The patients/participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

YZ, BC, TW, YL, ZC, ZZ, JM, XZ, YY, HW, and GC performed the experiments and analyzed the data. YZ and GC conceived the experiments and contributed to the interpretation and to writing the manuscript. All authors revised the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.