Review article
Treatment resistant depression: A multi-scale, systems biology approach

https://doi.org/10.1016/j.neubiorev.2017.08.019Get rights and content

Highlights

  • Treatment resistant depression (TRD) is a major public health concern.

  • Yet little is known about how TRD differs from treatment-responsive depression.

  • A multi-scale, systems biology approach is needed to better understand & treat TRD.

  • This involves a coordinated study of depressed humans and improved animal models.

  • And integrated use of genetic, epigenetic, cellular, and circuit analyses.

  • The goal is a precision medicine approach to depression including TRD.

Abstract

An estimated 50% of depressed patients are inadequately treated by available interventions. Even with an eventual recovery, many patients require a trial and error approach, as there are no reliable guidelines to match patients to optimal treatments and many patients develop treatment resistance over time. This situation derives from the heterogeneity of depression and the lack of biomarkers for stratification by distinct depression subtypes. There is thus a dire need for novel therapies. To address these known challenges, we propose a multi-scale framework for fundamental research on depression, aimed at identifying the brain circuits that are dysfunctional in several animal models of depression as well the changes in gene expression that are associated with these models. When combined with human genetic and imaging studies, our preclinical studies are starting to identify candidate circuits and molecules that are altered both in models of disease and in patient populations. Targeting these circuits and mechanisms can lead to novel generations of antidepressants tailored to specific patient populations with distinctive types of molecular and circuit dysfunction.

Introduction

Depression, or Major Depressive Disorder (MDD), is characterized by the persistence of negative thoughts and emotions that disrupt mood, cognition, motivation and behavior. Depression is the leading cause of disability worldwide affecting over 300 million people (World Health Organization, 2017). This burden has been rising steadily, with an 18% increase in prevalence between 2005 and 2015. The illness occurs throughout the lifespan, from childhood through old age, is ∼two-fold more common in women than men, and has higher incidence during puberty, peripartum periods, and menopause. Depression in mothers has detrimental effects on the fetus and young children (O’Donnell and Meaney, 2016). Depression is chronic: half of those who experience one episode of depression have recurrent episodes, with increasing frequency and severity of episodes over time. Depression is a leading cause of suicide and is associated with several common medical conditions, such as obesity, diabetes, stroke, Parkinson’s disease and multiple sclerosis as well as a greater risk for Alzheimer’s disease and sudden cardiac death. The impact of depression on humanity cannot be overstated.

While depression is diagnosed as a single entity, MDD, by the Diagnostic Statistical Manual (DSM) (2013), there are 681 combinations of symptoms that could meet the DSM criteria, reflecting the heterogeneity of symptoms, etiologies and pathophysiologies. Several described subtypes of depression, notably melancholic, psychotic or atypical depression, are distinguished solely by self-report criteria with no objective biological indicators. Depression symptoms are also seen in several other DSM-defined psychiatric syndromes (Kozak and Cuthbert, 2016). Depression is moderately heritable with as much as ∼35% of the risk associated with genetic predisposition (Geschwind and Flint, 2015), but is also highly influenced by adverse life experiences (Otte et al., 2016).

Multiple modalities of treatment are effective for depression, including antidepressant medications, psychotherapies and various brain stimulation techniques. Nonetheless, fewer than half of MDD patients achieve full remission with a first treatment (Rush, 2007). Further, matching a patient to his/her optimal treatment generally requires multiple trials of different treatments, with the sobering observation that the more treatments tried without success, the less likely a successful outcome. In sum, there remains a huge unmet need for a “precision medicine” approach to depression, with an important next step requiring development of treatments designed selectively for biologically-defined subtypes of this broad, heterogeneous syndrome (Drysdale et al., 2017, Williams, 2016).

A significant percentage of all MDD patients exhibit resistance to all available standard treatments. The evolution of resistance can develop in patients previously responsive to treatment or as a progressive, deteriorating illness course over time (Thase and Schwartz, 2015). Resistance can manifest as the presence of residual depressive symptoms following treatment as well as loss of effectiveness with ongoing treatment. Treatment options with increasing resistance are limited and generally involve continued use of the same modalities, including combination, augmentation or switching medications, introduction of electroconvulsive therapy (ECT) or trials of other neurostimulation strategies. These approaches risk complications, including increased toxicity with higher medication dosages and combination regimens.

Treatment Resistant Depression (TRD) represents a heterogeneous state with likely multiple causal mechanisms. TRD patients exhibit the same diversity of symptoms, course, history and co-occurring conditions as for treatment-responsive MDD. However, very little is known about what distinguishes patients who do or do not respond to treatment. The extent to which individuals with TRD versus treatment-responsive MDD differ in etiology or pathophysiology remains mostly obscure, although there are several reports that a history of early life stress increases treatment-resistance (Bernet and Stein, 1999; Nanni et al., 2012; Williams et al., 2016) and that individuals with TRD exhibit differences in brain circuit function (McGrath et al., 2016; Dunlop et al., 2017). Nevertheless, the underlying mechanisms are not known. Consequently, TRD remains an operational definition—and with several different definitions suggested, for example, referring to failure to respond to treatment within a depressive episode or failure to respond to a previously effective treatment in a subsequent episode (Fava, 2003, Conway et al., 2017). Thus, a major goal of current research is to establish more precise, biologically-based definitions of TRD as well as new antidepressant treatments targeting that underlying pathophysiology. Characterization of the biological heterogeneity of the TRD patient population is therefore both a necessity and a challenge. Strategies that consider biological subtypes rather than merely number and type of past treatments are needed. Yet, despite the compelling need for new treatments, especially for TRD, most pharmaceutical companies no longer prioritize depression given recent failures in drug discovery and the view that not enough is known about the underlying biology of depression to provide a rational path forward.

Animal models play an essential role in drug discovery in virtually all fields of medicine, but are particularly challenging in the case of depression (Nestler and Hyman, 2010). As with MDD itself, animal models must equally consider strategies that address symptom and etiological heterogeneity, while being mindful that only some human behaviors are amenable to study in non-human model systems: e.g., motivation, anhedonia, negative affect, hypothalamic dysregulation and homeostasis can be addressed, but not sadness, guilt, ruminations or suicidality. Several acute and chronic stress models have been used, but until recently it has been difficult to distinguish between adaptive vs. maladaptive responses to the stress. While stress is a risk factor for human depression, most individuals exposed to chronic stress do not develop depressive disorders. The issue of susceptibility is thus of paramount importance for animal studies of the biological basis for the relationship between stress and specific symptoms of depression. Additionally, many studies validate the models based on antidepressant response, thereby skewing away from identifying novel mechanisms of therapeutic actions. Indeed, there has not yet been a concerted effort to model the emergence of treatment resistance in rodents. This would require using animals with some genetic or developmental liability, exposing them to multiple bouts of stress and antidepressant treatment and characterizing a worsening course and the emergence to treatment resistance.

This review focuses on novel strategies in antidepressant drug discovery, particularly for patients with TRD. The authors came together four years ago to create the Depression Task Force sponsored by the Hope for Depression Research Foundation (http://www.hopefordepression.org/). Our goal is to use a reverse translation strategy to model the key features of depression, including its emergence, course and treatment response or resistance. We developed an interactive platform to test and integrate a set of complementary animal models of depression. These models are used to shed light on the pathophysiology of depression, including the relevant neural circuitry and the underlying genetic and molecular mechanisms. Anchored by an interactive “big data” analytic platform, we use a multi-scale, systems biology approach that leverages advances in genomics and neural circuitry, and integrates the discoveries in these animal models with findings from MDD patients and high-risk cohorts. Our goal is to identify biomarkers that will advance patient subtyping and treatment stratification, and facilitate the development of novel targeted interventions.

Section snippets

Brain plasticity and vulnerability in the context of brain-body interactions: historical overview

Studies of the neurobiology of depression, including TRD, focus largely on the association between stress and depression, with the hope that an understanding of the biological pathways that link stress to depression would inform on the pathophysiology of the disorder. The hypothalamic-pituitary-adrenal (HPA) axis, which controls secretion of both corticotrophin-releasing factor (CRF) and glucocorticoids, is central to the stress response. Mutations of several HPA genes have been used as genetic

Convergence and divergence across multiple rodent models of depression and treatment resistance

The heterogeneity of depression argues for the use of multiple animal models to capture both the diversity of the causes and symptoms as well as the common mechanisms that might underlie certain symptoms that are shared across all models. Each animal model likely recapitulates abnormalities seen in only a subset of patients or a subset of features of the broad syndrome, with many molecular-cellular mechanisms being unique to a given animal model. At the same time, other molecular-cellular

Neural circuitry of depression

Parallel but complementary studies using animal models, and both in vivo imaging and post-mortem studies in humans, suggest that depression does not arise through pathology in a single brain region or cell type, but instead is mediated by altered functioning across an integrated cortico-limbic circuit in the forebrain (Fig. 1) (Harris and Gordon, 2015, Heshmati and Russo, 2015). Key network nodes include regions of prefrontal cortex (PFC), connected with numerous subcortical structures

Depression genetics and genomics

Genetics and genomics influence the risk for depression. What is unclear is: 1) What is the nature of genetic vs. environmental influences, and their interactions, on the propensity for depression? And 2) What are the genes and gene networks involved in defining that heritable propensity for depression, and what biological systems do they affect? Strands of evidence are beginning to shed light on the answers. A third question is the extent to which genetic factors that contribute to MDD overall

Transcriptomic and epigenomic data across multiple animal models

Initial work in animal models of depression utilized DNA microarray technology, with RNA-seq increasingly used in recent years. The latter has the marked advantage of quantifying expression changes in individual splice variants of a gene as well as in several types of non-coding RNAs (e.g., microRNAs, long non-coding RNAs), which we now know play important roles in cell regulation. Moreover, transcripts revealed by RNA-seq can be better aligned with GWAS results, which often identify variants

Convergence across animal models and human depression

Given the inherent limitations of animal models of depression, it is imperative to correlate the exploration of transcriptional and epigenetic mechanisms in animals with abnormalities seen in the depressed human brain. The latter approach is fraught with its own limitations, such as the difficulty in obtaining high quality tissue with short post-mortem intervals, and the complications of comorbid conditions and variable, complex personal histories of stress and treatment exposures. Thus,

Gene x Environment Interactions

Environmental influences such as developmental history, early life stress and repeated exposure to trauma, as well as perhaps stochastic events during development, contribute almost two thirds of the variance to depression (Otte et al., 2016). Moreover, the illness has a dynamic course, whereby exposure of a vulnerable individual to negative or traumatic events can trigger of initial episodes of depression. However, the association between a triggering event and a depressive episode becomes

9 How can we use this multi-scale information to generate novel antidepressants?

Integration of multi-scale analyses are already providing a compelling algorithm to help guide antidepressant drug discovery efforts. We are proposing a reverse translation approach that starts by recognizing the complexity of the human syndrome with a focus on TRD, employs a broad range of animal models, overlays gene regulatory data from these models with studies of the post-mortem depressed human brain and GWAS and integrates studies using human and animal neuroimaging together with genetic

10 Conclusion and future directions

Recent advances in methodologies to study genetic and epigenetic mechanisms, as well as the functioning of precise brain microcircuits, prompt new optimism for our ability to parse the broad, heterogeneous syndrome of human depression into biologically-defined subtypes and to generate more effective and rapidly-acting treatments based on a knowledge of disease etiology and pathophysiology and circuit dynamics. We outline here a multi-dimensional systems biology approach to depression and TRD

Acknowledgements:

Preparation of this review was supported by the Hope for Depression Research Foundation (HDRF). The authors thank Dr. Steven Roose (Columbia University) for helpful discussions.

References (214)

  • C. Hamani et al.

    Deep brain stimulation in rats: different targets induce similar antidepressant-like effects but influence different circuits

    Neurobiol. Dis.

    (2014)
  • A. Heck et al.

    Converging genetic and functional brain imaging evidence links neuronal excitability to working memory, psychiatric disease, and brain activity

    Neuron

    (2014)
  • S.R. Heilbronner et al.

    Circuit-based corticostriatal homologies between rat and primate

    Biol. Psychiatry

    (2016)
  • R. Hultman et al.

    Dysregulation of prefrontal cortex-mediated slow-evolving limbic dynamics drives stress-induced emotional pathology

    Neuron

    (2016)
  • N. Insel et al.

    Chronic deep brain stimulation of the rat ventral medial prefrontal cortex disrupts hippocampal-prefrontal coherence

    Exp. Neurol.

    (2015)
  • M.A. Kheirbek et al.

    Differential control of learning and anxiety along the dorsoventral axis of the dentate gyrus

    Neuron

    (2013)
  • M.A. Kohli et al.

    The neuronal transporter gene SLC6A15 confers risk to major depression

    Neuron

    (2011)
  • J. Abbasi

    23 and Me, big data, and the genetics of depression

    JAMA

    (2017)
  • H. Akil et al.

    The future of psychiatric research: genomes and neural circuits

    Science

    (2010)
  • H. Akil

    Stressed and depressed

    Nat. Med.

    (2005)
  • P.R. Albert

    Transcriptional regulation of the 5-HT1A receptor: implications for mental illness

    Phil. Trans. R. Soc. London Ser. B Biol. Sci.

    (2012)
  • C. Anacker et al.

    Adult hippocampal neurogenesis and cognitive flexibility – linking memory and mood

    Nat. Rev. Neurosci.

    (2017)
  • C. Anacker et al.

    Neuroanatomical differences associated with stress susceptibility and resilience

    Biol. Psychiatry

    (2015)
  • M.S. Ansorge et al.

    Early-life blockade of the 5-HT transporter alters emotional behavior in adult mice

    Science

    (2004)
  • R.C. Bagot et al.

    Epigenetic signaling in psychiatric disorders: stress and depression

    Dialogues Clin. Neurosci.

    (2014)
  • R.C. Bagot et al.

    Ventral hippocampal afferents to the nucleus accumbens regulate susceptibility to depression

    Nat. Commun.

    (2015)
  • R.C. Bagot et al.

    Ketamine and imipramine reverse transcriptional signatures of susceptibility and induce resilience-specific gene expression profiles

    Biol. Psychiatry

    (2016)
  • R. Bernard et al.

    Altered expression of glutamate signaling, growth factor, and glia genes in the locus coeruleus of patients with major depression

    Mol. Psychiatry

    (2011)
  • C.Z. Bernet et al.

    Relationship of childhood maltreatment to the onset and course of major depression in adulthood

    Depress. Anxiety

    (1999)
  • O. Berton et al.

    New approaches to antidepressant drug discovery: beyond monoamines

    Nat. Rev. Neurosci.

    (2006)
  • J.M. Bessa et al.

    The mood-improving actions of antidepressants do not depend on neurogenesis but are associated with neuronal remodeling

    Mol. Psychiatry

    (2009)
  • B. Bigio et al.

    Epigenetics and energetics in ventral hippocampus mediate rapid antidepressant action: implications for treatment resistance

    Proc. Natl. Acad. Sci. U. S. A.

    (2016)
  • B. Bunney et al.

    Circadian dysregulation of clock genes: clues to rapid treatments in major depressive disorder

    Mol. Psychiatry

    (2015)
  • C. Buss et al.

    Maternal care modulates the relationship between prenatal risk and hippocampal volume

    J. Neurosci.

    (2007)
  • C. Buss et al.

    Maternal cortisol over the course of pregnancy and subsequent child amygdala and hippocampus volumes and affective problems

    Proc. Natl. Acad. Sci. U. S. A.

    (2012)
  • A. Caspi et al.

    Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene

    Science

    (2003)
  • S. Chattarji et al.

    Neighborhood matters: divergent patterns of stress-induced plasticity across the brain

    Nat. Neurosci.

    (2015)
  • D. Chaudhury et al.

    Rapid regulation of depression-related behaviours by control of midbrain dopamine neurons

    Nature

    (2013)
  • P.V. Choudary et al.

    Altered cortical glutamatergic and GABAergic signal transmission with glial involvement in depression

    Proc. Natl. Acad. Sci. U. S. A.

    (2005)
  • D.J. Christoffel et al.

    I(B kinase regulates social defeat stress-induced synaptic and behavioral plasticity

    J. Neurosci.

    (2011)
  • C.R. Conway et al.

    Toward an evidence-based, operational definition of treatment-resistant depression. When enough is enough

    JAMA Psychiatry

    (2017)
  • H.E. Covington et al.

    Antidepressant effect of optogenetic stimulation of the medial prefrontal cortex

    J. Neurosci.

    (2010)
  • R.C. Craddock et al.

    Disease state prediction from resting state functional connectivity

    Magn. Reson. Med.

    (2009)
  • Cross-Disorder Group of the Psychiatric Genomics Consortium

    Identification of risk loci with shared effects on five major psychiatric disorders: a genome-wide analysis

    Lancet

    (2013)
  • Cross-Disorder Group of the Psychiatric Genomics Consortium

    Genetic relationship between five psychiatric disorders estimated from genome-wide SNPs

    Nat. Genet.

    (2013)
  • R.C. Culverhouse et al.

    Collaborative meta-analysis finds no evidence of a strong interaction between stress and 5-HTTLPR genotype contributing to the development of depression

    Mol Psychiatry Mol. Psychiatry

    (2017)
  • N.A. Datson et al.

    Prior history of chronic stress changes the transcriptional response to glucocorticoid challenge in the dentate gyrus region of the male rat hippocampus

    Endocrinology

    (2013)
  • E.R. De Kloet et al.

    Stress and the brain: from adaptation to disease

    Nat. Rev. Neurosci.

    (2005)
  • S. De Rubeis et al.

    Synaptic, transcriptional and chromatin genes disrupted in autism

    Nature

    (2014)
  • K. Dean et al.

    Full spectrum of psychiatric outcomes among offspring with parental history of mental disorder

    Arch. Gen. Psychiatry

    (2010)
  • Cited by (0)

    View full text