The Role of Precision Medicine in Child Psychiatry: What Can We Expect and When?

In the United States, nearly 1 in 5 adults and 1 in 4 children suffer with a mental illness, and for most, these illnesses confer significant social, occupational, or academic impairment.^1,2 Whereas considerable progress has been made toward diminishing the impact of mental illness, in the last few decades, advances toward new, more effective, and more refined psychiatric treatments have slowed. First-line medications for depression, schizophrenia, bipolar disorder, and attention-deficit/hyperactivity disorder (ADHD) were all introduced decades ago,^3–6 and little progress has been made in augmenting their effectiveness.^7,8

DEFINING PRECISION MEDICINE

Precision medicine is advocated as a path forward, leading psychiatry away from stalled treatment development toward a more fruitful path paved by oncology with its successful development of highly refined and targeted therapeutics. Precision medicine is characterized by tailoring treatments to the specific disease process affecting a patient, parsing heterogeneous syndromes on the basis of etiology or underlying disease mechanisms. Complex syndromes, such as breast or ovarian cancer, are split into etiologically homogenous sub-types. Understanding the distinct etiology of a disease process can then lead to “precision medicines” that target the specific etiology, rather than broad-based, “one size fits all” treatments lacking mechanistic specificity. Personalized medicine denotes a concept similar to precision medicine, but suggests a more individualized approach, tailoring treatment to each individual patient, rather than disease mechanisms, which may be the same across groups of patients.

To clarify the approach of precision medicine, it is helpful to contrast an example of this approach successfully implemented in oncology with that of the empirical, algorithmic approach used in psychiatry. Imatinib, a ground-breaking antineoplastic agent, is one precision medicine success story. In 2001, the Food and Drug Administration (FDA) approved imatinib after clinical trials demonstrated its efficacy in treating chronic myeloid leukemia (CML). More than 90% of CML cases are caused by a chromosomal translocation, termed the Philadelphia chromosome, in which a portion of chromosome 9 is transferred to chromosome 22 and a fusion gene, BCR-ABL1, is formed.⁹ This genetic anomaly leads to constitutively active tyrosine kinase. Targeting this disease process, imatinib inhibits BCR-ABL1 tyrosine kinase, and in doing so, increases cell death. Imatinib has improved the 5-year survival rate of CML 3-fold from approximately 30% to 90%.¹⁰ It is important to underscore that imatinib is not a treatment for CML per se, but rather a treatment for a specific disease process or mechanism (ie, genetic mutations that lead to an over-expression of tyrosine kinase). This is the underlying cause of most CML cases, but it is also a cause of other malignancies, such as gastrointestinal stromal tumors, for which imatinib is also effective.¹¹ In other words, imatinib targets a cancer mechanism, not a specific tissue or organ with cancer.

Contrast this etiologic and mechanistic approach with psychiatric practice in which diagnostic nosologies are explicitly nonetiologic. For example, if a child presents with symptoms of anxiety or depression, it is the intensity, duration, and associated impairment that confer diagnosis—the causes of the symptoms are irrelevant to diagnosis. SSRIs may be prescribed for depression and/or anxiety disorders irrespective of whether the treating psychiatrist suspects that the syndrome arose because of a recent trauma, genetic pre-dispositions, head injury, or some combination or interaction of these etiologic factors. Validated approach linking etiology and treatment are scarce, and thus psychiatrists typically attempt to find what works best through a process of trial and error, zeroing in on the most efficacious treatment. Treatment selection is an empirical decision determined after one or more medication trials and is, to a large extent, unassisted by predictive analytics from the patient’s history, genetics, or environmental exposures. Conversely, a precision medicine approach within child psychiatry might entail parsing heterogeneous diagnoses such as autism spectrum disorders (ASD) into underlying etiologies or disease mechanisms (eg, de novo mutations). With sufficient treatment development, clinicians might then be able to select a treatment, not based strictly on the diagnosis of ASD or related symptoms, but rather on the underlying disease mechanism (eg, the pathophysiology related to a specific de novo mutation) (Figure 1). The ambition of bringing precision medicines to the child psychiatric clinic, although unarguably a laudable goal, also brings with it significant hurdles, as discussed in the proceeding sections.

FIGURE 1.

Comparison of Current Psychiatric Practice With Precision Medicine

Note: (A) In current psychiatric practice, treatments are selected on the basis of diagnoses or syndromes, which may encompass heterogeneous disease processes, or underlying pathophysiology. This heterogeneity is represented here by blue, red, and green stick figures. For example, children with autism spectrum disorders (ASD) are provided treatments based on their diagnosis and symptoms, not on the basis of the etiology of ASD. (B) Precision medicine aims to match treatments with disease processes. Although this goal has not yet been achieved, children with ASD could, in theory, be divided into subgroups based on de novo mutations that give rise to the disorder, represented here as stick figures sorted by color in the column labeled “Genetic variants.” Targeted treatments could then be offered to address the pathophysiology related to the de novo mutation, represented here in the column labeled “Precision medicines.”

IMPORTING PRECISION MEDICINE FROM ONCOLOGY TO PSYCHIATRY

Although advances in precision medicines in oncology are inspiring, their impact should not be overstated, nor should their applicability to psychiatry be assumed. Cancer diagnoses for which precision medicines are effective are quite limited. Among the most common malignancies (eg, breast, prostate, colorectal, and lung), some combination of surgery, radiation, and/or chemo-therapy is the primary treatment for more than 95% of patients.¹² These numbers will likely improve with time; however, until such broader applicability is achieved, child psychiatry should be careful not to assume that once precision medicines are developed for childhood mental illnesses, their widespread use will follow in lockstep; rather, once developed, the use of precision medicines in child psychiatry will likely remain circumscribed, similar to oncology, to cases in which rare genetic anomalous or other disease processes are closely linked to a clinical presentation.

There is a second, and more foundational, challenge to importing oncology as a model for treatment development in psychiatry. Cancers are heterogeneous diseases in terms of their histology, genetics, and environmental determinants; however, they are homogenous in regard to their cellular basis. Unrestrained mitosis due to genetic mutations defines all cancers, regardless of shape, size, tissue, etiology, prognosis, or prevalence, and this unifying feature makes the genome, and its expression of proteins, the categorical target for novel oncologic therapeutics. Even for environmental hazards, such as ionizing radiation, asbestos, or tobacco, carcinogenic effects are contingent upon entering the cell and mutating the genome. Some precision medicines, such as immunotherapies, target proteins rather than the genome itself, but the targeted proteins are excessively expressed because of a mutated gene. The etiology homogeneity of cancer therefore defines and limits the number of therapeutic targets against which precision medicines can be considered and tested.

There is no analogous, etiologic homogeneity within psychiatry. The etiological factors of psychiatric conditions, although largely unknown, are most likely drawn from multiple domains—from environmental (eg, trauma), to genetic (eg, rare variant mutations in autism), to brain/cognition (impaired executive functions) to developmental (eg, increases in psychopathology during adolescence). These etiologic domains may become linked to form causal chains; for example, genetics may predispose to trauma, and trauma may result in impaired executive functions; impaired executive functions may then curtail academic/vocational achievement, and academic/vocational failures may in turn affect mood. Causal chains may similarly exist in oncology; however, they ultimately spiral inward and coalesce around genetic mutations. If the genetic mutation(s) can be corrected, the disease process halts.

One might be tempted to identify the brain as the site of coalescence in the causal chains for psychiatric disorders, yet this may not translate directly to interventions. For instance, when brain and environment are causally linked, intervening at a level of the brain may not be the more efficacious route toward clinical improvement. For example, treating a traumatized child with psychopharmacology is unlikely to alleviate posttraumatic stress disorder if the trauma is ongoing. Similarly, a socially phobic child growing up in an environment steeped in humiliation and embarrassment is unlikely to benefit from targeting neural systems underlying fear, shame, or negative affect (even if such a medication were to exist). In both cases, interventions center on, or at least begin with, environmental manipulations. Unlike oncology, in which disease processes are circumscribed and centered on genetic mutations, there is no analogous restriction for mental illnesses, particularly in youth; disease processes can encompass any number of environmental, developmental, or somatic disturbances.

Even if we reject the above argument and assume that all disease mechanisms in psychiatry reside in the brain (arguably a more reductionistic stance), psychiatry’s route to precision medicine may still prove arduous. The etiologic homogeneity of oncology limits potential therapeutic targets to cellular and genomic foci. These foci are numerous, and their complexity undoubtedly gives way to the slow pace of oncological discovery. However, this complexity is still far more circumscribed than the potential therapeutic world of the brain, which including all cell types (eg, neurons, glia, microglia, astrocytes, etc.), extracellular compartments and signaling (eg, synapses, neuromodulators, vesicles, hormones, cytokines, etc.), and circuits (eg, phenotypes arising from large-scale network interactions). For child psychiatry, this playing field is made still more complex by the ontological unfolding of neurobiological processes in which the role of biological signals can vary depending on developmental stages. For example, γ-aminobutyric acid (GABA) has an inhibitory function in the adult brain and yet is excitatory to immature neurons; similarly, serotonin acts as a neurotransmitter in the adult brain but is a neurotropic factor in the developing brain.¹³ Precision medicines in child psychiatry must not only identify appropriate biological targets, but must also specify when in development these targets can or should be engaged.

CONCLUSIONS

The arguments presented above are not meant to suggest that the challenges to importing precision medicine into child psychiatry are insurmountable. It is quite likely that with time, persistence, and some tincture of luck, this approach will yield new treatments. However, before precision medicines are integrated into child psychiatrists’ therapeutic landscape, a good deal of patient waiting may be required. In the next section, I discuss complementary paths toward treatment development that may help advance treatment in the nearer term—paths that can be pursued in parallel with precision medicines with the goal of providing therapeutic developments in both the short and long term with a balanced portfolio of investigative approaches:

1. Repurposing: Observing and leveraging shared mechanisms of action across medication classes remains an important, and potentially under-utilized, route to new treatments.¹⁴ For example, atomoxetine was initially developed as a treatment for depression. After proving ineffective for this disorder, it was noted that atomoxetine’s mechanistic similarities to desipramine, an effective treatment for ADHD.^15,16 This led to clinical trials and subsequent FDA approval for atomoxetine in the treatment of ADHD.¹⁶ Unlike precision medicine, which focuses on understanding disease mechanisms, repurposing atomoxetine hinged on understanding drug mechanisms. A recent review indicates that across medical disciplines, repurposing has had its greatest success in the CNS domain.¹⁷ This approach can continue to be advanced by increasing the availability and transparency of existing compounds with CNS properties, many of which are unfortunately difficult to access because of intellectual property laws.

2. Collaborative research participation: Over the past 35 years, survival rates for children with acute lymphoblastic leukemia (ALL) have increased from less than 10% to more than 85%. This success is in large measure attributable to multisite clinical trials fostered by international research partnerships, such as the Children’s Oncology Group (COG), that promote multisite collaborative clinical studies.¹⁸ As a result of these efforts, more than 85% of children with cancer diagnoses participate in clinical research.¹⁹ Rates of research participation have not been quantified in psychiatry; however, increasing research participation is clearly an area in which improvements can be made. Stigma attached to psychiatric illness is a barrier, but not an insurmountable one. For example, the Simons Foundation has launched the Simons Foundation Powering Autism Research for Knowledge (SPARK) to bring together more than 20 research clinics and more than 50,000 individuals with autism, with the goal of making extensive genotypic and phenotypic data available to the scientific community. Similar initiatives could be conducted with other disorders, particularly those such as ADHD in which significant progress has been made in reducing stigma.²⁰

3. Psychotherapy: Although not developed with precision medicine in mind, many psychotherapeutic approaches arguably encompass several of its key characteristics. A mechanistic understanding of a patient’s distress is formulated and treatment is tailored to redress that disease process—for example, a socially anxious child is noted to overestimate social rejection; this begets avoidance of social interactions, which then exacerbates feelings of isolation. A cognitive-behavioral therapist might address these maladaptive thoughts and behaviors by restructuring the cognitive distortions and using exposures to confront avoidance. Importantly, these psychotherapeutic techniques fit within one psychotherapy approach (ie, cognitive-behavioral therapy), yet the specific interventions selected for a given socially anxious child may differ from those selected for another child. The therapy attempts to match “precision interventions” to the salient cognitive and behavioral factors contributing to the clinical presentation.

Current psychotherapeutic practice differs from precision medicine is that its inherent mechanistic understanding is infrequently tested or conclusive. In other words, the clinician (or psychotherapy researcher) is seldom able to establish whether a patient has improved because a hypothesized cognitive distortion has been redressed, or whether other aspects of the psychotherapeutic engagement are in fact accounting for the reduction in symptoms. Novel empirical approaches may be helpful in this regard by more directly focusing on, and testing, the mechanistic frame-work (eg, NIMH R61/R33 studies: https://grants.nih.gov/grants/guide/rfa-files/RFA-MH-17-604.html). Likewise, digital technologies, such as ecological momentary assessments with smartphones and wearable devices, may provide phenotypic characterization with much finer timescales than were previously practical. This capacity to examine change following treatment with highly granular and dynamic data may help advance, and empirically validate, mechanistic understandings of psychotherapy.

CONCLUSION

In conclusion, precision medicine presents a route to treatment discovery that hinges on understanding how psychiatric disorders emerge and progress, and by developing novel interventions that target these underlying pathophysiologic pathways. This approach will almost certainly bear fruit one day, yet given the significant hurdles that stand in the way, including the heterogeneity and complexity of the determinants that contribute to mental illness, a good deal of patience may still be necessary. As we wait for precision medicines to unfold, we are wise to maintain a balance of methodologies toward treatment development, pursuing investigative leads as they arise regardless of the precision they may confer, and sustaining a healthy, albeit reserved, optimism toward treatments of the future.