Could we get new insights about the course of major mood disorders by counting single immune molecules and extracellular vesicles in your blood with innovative nano-sensors?
The molecular study of a longitudinal relationship between immune activation and the course of major mood disorders remains difficult, due in part to the molecular complexity that governs immune response (Haroon et al. Carpenter, Reference Carpenter, Crouse, Scott, Naismith, Wilson, Scott, Merikangas and Hickie2021; Miller et al. Reference Chandan, Schiffman and Balakrishna2009). At a proteomic level, this molecular complexity arises from the diversity of inflammatory molecules that are involved in orchestrating immune activation and induce low-level inflammation. Low-level inflammation can be transient or chronic and triggered by various factors (e.g., inactivity, smoking, weight gain, stress, sleep deprivation and toxicants). Interleukins (e.g., IL-1b, IL-6, IL-8 and IL-12), tumour necrosis factor-alpha (TNF-α), C-reactive protein (CRP) and interferons are involved in immune activation in response to many pathogens. Increasingly, however, they have been detected in various cohorts of people with major mood disorders, or those with other medical conditions and comorbid mood disorders (Coico Reference Coico2021; Furman et al. Reference Furman2019; McNamara and Lotrich Reference Dowlati2012; Mogensen Reference Drevets, Wittenberg, Bullmore and Manji2009; Miller and Raison Reference Farka2016) As these markers are typically only detected in very low concentrations, they are assumed to be a consequence of central nervous system dysregulation rather than being indicative of an alternative infective or immunopathological disorder. Consequently, their potential use as peripheral markers of brain change in mood disorders has been long proposed (Hickie et al. Reference Gooding and Gaus1990) recognising that their relationships with true impairment of immune function may be very limited or restricted to subgroups with the most severe mood disorders (Hickie and Wilson Reference Haroon, Miller and Sanacora1994). More recently (Drevets et al. Reference Hickie, Silove, Hickie, Wakefield and Lloyd2022), there has been a call for development of new antidepressant therapies that focus on modulation of specific immune targets.
Challenges of the heterogeneity of major mood disorders
Despite the last two decades of case–control studies, and even when focusing on more narrowly defined cohorts of people with major mood disorders (e.g., by age, age-of-onset, illness duration, proposed clinical subtype or patterns of physical or mental health comorbidity), there is a lack of consensus about the concentration or time-course of ‘low-level’ inflammatory markers in circulation; their physiological significance; their relationships with specific clinical features (e.g., anhedonia, motor or cognitive slowing, fatigue or other ‘sickness behaviour’-like features); or their utility as biomarker that can predict illness course or treatment response. Important other factors are ignored, such as the impact of the specific clock time or ‘circadian time’ of blood draw on results obtained – given that many immune measures present distinct circadian patterns (Scheiermann et al. Reference Hickie and Wilson2013) – or the body mass index or formal metabolic status of tested individuals. Accordingly, only very general conclusions are possible. For example, in a meta-analysis in people with a major depressive episode, TNF-α and IL-6 were found to be significantly elevated in the blood of people with a major depressive episode compared to unaffected controls (Dowlati Reference Hickie2010). Despite the statistical difference, the studies reviewed in the meta-analysis used different threshold concentrations of TNF-α and IL-6 to classify an individual as depressed or non-depressed. Consequently, the levels of inflammatory markers for some ‘depressed’ subjects in one study would classify as ‘non-depressed’ in another study. These differences can be attributed to many factors, notably the demographic and clinical heterogeneity in the cohorts examined. While it is common to run clinical studies with matched cohorts or account for this variable statistically, the technology used and (circadian) time of blood draw are typically less accounted for.
Circadian alignment, synchrony and molecular fingerprinting
As the regulation of the immune system is controlled by circadian oscillation (Scheiermann et al. Reference Juncker, Bergeron, Laforte and Li2013), the time of sampling and number of time points of samples might be some of the most important variables that could lead to variations in concentration of peripheral inflammatory markers. To account for the temporal and person-to-person variations in inflammatory markers, profiling of a panel of inflammatory markers diurnally (over 24 h) and longitudinally (over weeks) to create an individualised molecular fingerprint and detect if the immune system is ‘out of sync’ with the light-dark cycle (i.e., external alignment) and other key internal circadian markers (i.e., timing and concentration of evening melatonin onset, core body temperature nadir, morning cortisol on wakening and cortisol after rising) is a highly attractive idea. Further, a ‘multiplex’ application of molecular fingerprinting to multiple inflammatory markers simultaneously could allow inferences to be made about whether a person is in a state of ‘internal desynchrony’, wherein the rhythms of these processes become uncoupled from each other.
Specifically, tracking such a molecular fingerprint is very well positioned to detect abnormalities that coincide with the transitions between the onset and offset of episodes of mood disorders—especially those characterised by highly recurrent patterns—irrespective of whether the detected markers are above or below those thresholds for ‘low-level’ inflammation that are primarily relevant to other primary infective or immune pathologies. As we are pursuing the clinical and neurobiological characterisation of a putative pathophysiological subtype of major unipolar and bipolar mood disorders that we have termed ‘circadian depression’ (Crouse Reference Li2021), the identification of biological markers that closely track the daily patterns or longer-term illness course of these disorders is a high priority (Hickie Reference Li2023). Characteristically, these ‘circadian mood disorders’ are associated phenotypically with factors such as prolonged fatigue, disrupted 24-hour sleep-wake cycles (e.g., phase-delayed, non-restorative sleep and reduced daytime activity) and weight gain. They have also been linked specifically to markers of metabolic dysfunction (e.g., increased insulin resistance) and non-specific markers of inflammation (e.g., raised CRP) or immune dysregulation (Carpenter Reference Li2021).
Nano-level monitoring of the immune/inflammatory state of the brain
Another challenge is the specificity of ‘general’ inflammatory markers such as peripheral cytokines and chemokines to indicate immune activation associated with major mood disorders. As the concentration of these markers change in response to a host of other medical and environmental factors, their unambiguous association with the course of mood disorders is problematic. The discovery and implementation of new and potentially brain-specific markers could provide the critical link to unravel the pathophysiology of at least some subtypes of major mood disorders. Small extracellular vesicles (sEVs) – biological nanoparticles of 30–200 nm in diameter – are a promising class of biomarkers that are known to carry a similar phenotype or ‘protein barcode’ to their cell of origin (van Niel et al. Reference Li2018; Wang Reference Li, Zhang, Trau and Wuethrich2021). sEVs are secreted from most cells and have the ability to cross the blood–brain barrier. Monitoring phenotypic changes in sEVs secreted from neurons, microglia and astrocytes might connect brain-related changes to peripheral immune activation (e.g., soluble inflammatory markers) and thus offer a unique window into the immune/inflammatory state of the brain.
The accurate measurement of trace-level inflammatory markers and novel markers such as sEVs in blood is technically challenging. From an analytical viewpoint, blood is a highly complex sample, and even when processed to plasma, it still constitutes a complex proteomic sample. While more abundant protein biomarkers can readily be measured by standard immunoassays and mass spectrometry, trace-level detection of biomarkers, particularly inflammatory markers, in plasma is difficult (Zhou Reference Lim2024; Li Reference Liu2024). The abundance of non-target molecules in plasma can suppress the detection signal of inflammatory markers and cause non-specific signals. Both factors increase the ‘background noise’ of an analytical method, which complicates accurate trace level analysis of inflammatory markers. Standard enzyme-linked immunosorbent assay (ELISA) typically has a limit of detection in the nanogram per mL range (10-9 g/mL), which can be multiple orders of magnitude above the concentration of some cytokines and chemokines found in low-level inflammation (Liu Reference Mattsson-Carlgren, Palmqvist, Blennow and Hansson2021). While more sensitive than ELISA, mass spectrometry requires specialised equipment and sample processing that can hinder its applicability in the clinical setting. New technologies need to be developed and tested on clinical samples, which can deliberately surmount all of these issues. As well as the issue of sensitivity, specificity of assays performed on trace biomarkers is also critical. This is especially important for highly multiplexed assays which utilise a single binding motif agents such as antibodies or aptamers – both of these capture agents are known for their tendency to cross-react with multiple epitopes and multiple proteins (Juncker et al. Reference McNamara and Lotrich2014; Tighe et al. Reference Miller, Maletic and Raison2013). If not properly addressed, cross-reactivity and non-specific binding can have detrimental effects on assay reproducibility and jeopardising the impact of large biomarker discovery programs (Williams Reference Miller and Raison2011; Mattsson-Carlgren et al. Reference Mogensen2020). Once again, for highly complex and heterogeneous samples, digital molecular counting approaches that can potentially surmount these issues (through multi-factor molecular interrogation approaches to improve the specificity of trace biomarker readout (Li Reference Möller and Lobb2023)) could provide an important advantage in this arena.
Nanotechnology enables the preparation of materials and systems at the nanoscale – a dimension at which materials can exhibit unique physico-chemical properties including quantum effects, increased surface area-to-volume ratios and enhanced mechanical, optical, electrical and chemical properties. These properties can be explored for method development for trace-level biomarker detection (Li Reference Scheiermann, Kunisaki and Frenette2020; Wang Reference Tighe, Negm, Todd and Fairclough2024; Yuan Reference van Niel, D’Angelo and Raposo2023). A recent and prominent example of nanomaterials, as corroborated by last year’s Nobel Prize in Chemistry, are quantum dots. One of the most striking features of quantum dots is their high quantum yield; an important property for achieving high detection sensitivity (Zhou et al. Reference Wang2020; Chandan et al. Reference Wang2018). Similarly, at the nanoscale, materials and structures have high surface area-to-volume ratios, which can speed up mass transfer and reduce assay time. Leveraging on these features of nanomaterials, research efforts have been geared towards the development of nanotechnologies with single protein detection sensitivity (Farka Reference Williams2020). An advantage of single molecule-level sensitivity is that it enables the ability to count molecules in a binary fashion, which in theory, enables absolute quantification without calibration (Gooding and Gaus Reference Yuan2016; Li et al. Reference Zhou, Chizhik, Chu and Jin2024). Traditional immunoassays generate a concentration-dependent signal based on the ensemble or average read-out of many proteins. A drawback of such ensemble measurements, particularly when it comes to trace-level detection, is that the quantification can be biased by signal fluctuations and background noise that can overpower the signal from a trace analyte. By contrast, digital single molecule sensitive technologies can largely avoid that bias and detect trace-level biomarkers much more precisely.
Over the past few years, our laboratory has focused on the development of a digital nanotechnology for detection of trace inflammatory markers and sEVs in clinical samples. The nanotechnology is based on a sandwich immunoassay using a silicon chip containing an array of 250,000 nanopillars and single particle active nanoparticle barcodes. Both the chip and barcodes are functionalised with antibodies against the inflammatory markers or sEVs to form an immune complex in the presence of the target species. At sufficiently low protein concentrations, the distribution of biomarkers on the nanopillar array is governed by probability driven distribution (i.e., Poisson distribution). For example, at a ratio of one biomarker molecule per 10 nanopillars, there is a probability of 99% to have either 1 or 0 molecules present per nanopillar. This binary distribution of molecules on nanopillars and use of single particle-active nanoparticle barcodes is critical for direct counting of single molecules that does not require a substrate-dependent amplification of the detection signal. Other assays for trace biomarker detection often require a substrate and enzyme to generate a sufficiently strong detection signal. Any non-specific binding that sets off the amplification process can lead to an overestimation of the analyte concentration. Conversely, if the amplification process is not triggered, an underestimation of the analyte concentration might be obtained. Direct (i.e., amplification-free) counting of molecules significantly reduces the risk of under- or overestimation of the analyte concentration.
Initially, we investigated the digital nanotechnology for trace detection of cytokines in melanoma patients treated with immune checkpoint inhibitors (ICI) (Li Reference Zhou2021). Melanoma patients frequently experience immune toxicity as a side effect of the inhibitory therapy. Inflammatory cytokines including FGF-2, G-CSF, GM-CSF and CX3CL1 have previously been implicated in ICI-treated melanoma patients with acute immune toxicities (Lim Reference Zhou2019). We hypothesised that trace level detection of these inflammatory markers could indicate the onset of immune toxicities in response to ICI. In the tested cohort, half of the patients showed high grade immune toxicities, while the other half had low grade or no immune toxicity. We found that patients with high-grade immune toxicity showed a steady increase in trace-level cytokines over the course of ICI, while the cytokine levels in patients with low/no immune toxicity remained relatively stable. Interestingly, the detected cytokine levels were in the femtomolar range (10–15 M), which was approximately a 1000 times lower than the limit of detection of most conventional ELISAs. Encouraged by the potential of trace biomarker analysis, we were interested in profiling of sEVs as a tool for non-invasive lung cancer screening. Cancer-associated sEVs can be present at concentrations many orders of magnitude lower than the total sEVs concentration in human plasma, rendering their study similar to the search for a needle in a haystack. For example, at early stages of cancer, the amount of tumour-specific sEVs secreted into circulation is relatively small, requiring both high target specificity and sensitivity (Möller and Lobb Reference Williams2020). We re-programmed the nanopillar chip to target multiple sEVs markers associated with lung cancer and profiled sEVs in patients with benign and malignant lung lesions (Li Reference Möller and Lobb2023). While the patients in the benign and malignant cohorts showed similar features in the computed tomography images and thus were difficult to discern, the sEVs profiles showed clear differences between the two cohorts.
Applications to modelling the illness course of major mood disorders
From our perspective, actively profiling soluble inflammatory markers and innovative tissue-specific markers like sEVs shows promise for creating a detailed temporal inflammatory signature. This approach could significantly enhance our understanding of how immune activation relates to mood disorders. Achieving precise analysis of trace biomarkers demands the use of highly sensitive and highly specific technologies. Digital nanotechnologies, which can count individual inflammatory molecules and sEVs, open up novel and exciting opportunities to explore the intricate connection between the brain and immune system, with potential applications in diagnosing and monitoring the illness course and response to treatment of at least some common types of major mood disorders. This is the major vision of our NHMRC Synergy Grant program (APP2019260). The Synergy Grant investigates the dysfunction of 24-hour body clocks as a potential mechanism underlying mood disorders and that alleviation of this circadian dysfunction requires highly personalised therapies. To gain new insights and delivery personalised care, the Synergy Grant develops nanotechnologies for monitoring blood-based biomarkers to measure circadian dysfunction in patients and evaluate the effect of treatment on circadian system in patients with mood disorders.
Data availability statement
The authors confirm that the data supporting the findings of this study are available within the article.
Author contributions
A.W. and M.T. have contributed equally to the conceptualisation, writing, editing and reviewing of this manuscript. C.V., Y.Z., F.F., J.J.C. and I.B.H. contributed to editing and writing after the original draft. Matt Trau can also be contacted for correspondence, m.trau@uq.edu.au.
Financial support
A.W., C.V., Y.Z., F.F. and M.T. acknowledge the support by National Health and Medical Research Council (APP2019260, APP1173669), Cancer Australia (2010799) and Australian Research Council (DP210103151, FL220100059). J.J.C. and I.B.H. acknowledge support by National Health and Medical Research Council Investigator Grants (APP2016346, APP2008196). All authors acknowledge support by a National Health and Medical Research Council Synergy Grant (APP2019260).
Competing interests
A.W., C.V., Y.Z., F.F., J.J.C. and M.T. declare no conflict of interest.
I.B.H. is the Co-Director, Health and Policy at the Brain and Mind Centre (BMC) University of Sydney. The BMC operates an early intervention youth service at Camperdown under contract to headspace. He is the Chief Scientific Advisor to, and a 3.2% equity shareholder in, InnoWell Pty Ltd., which aims to transform mental health services through the use of innovative technologies.
Ethical standards
Ethical approval and consent are not relevant to this article type.