Cerebrospinal Fluid and Psychiatric Disorders (CSFPsych)

The relevance of cerebrospinal fluid to neurobehavioral traits
Being part of the central nervous system (CNS), cerebrospinal fluid (CSF) is
the body fluid in closest proximity to the brain. In clinical practice, CSF is
typically targeted in neurology to determine CNS inflammation (e.g. infections
and autoimmune diseases), exclude subarachnoid bleeding and aid in the
diagnosis of Creutzeldt-Jacob disease. Moreover, CSF constituents have proven
highly valuable diagnostic biomarkers for Alzheimer*s disease. CSF may also
provide a snapshot of other molecular processes with relevance to the
functioning of the CNS, including processes related to immunology,
neurotransmission and microRNA (miRNA).
With regard to the field of psychiatry, recent CSF studies have proven
informative for a range of biological and behavioral measures, including
seasonal variation in monoamine metabolites; genetic determinants of GABA,
monoamine metabolite concentrations and NMDAR coagonists; smoking behavior;
schizophrenia; and depressive symptoms. These studies demonstrate that CSF is a
valuable and promising body fluid that is in direct contact with the brain. It
may therefore provide insight into the neurobiological background of
neuropsychiatric (intermediate) traits. So far, studies have been carried out
in healthy subjects or in heterogeneous study populations of psychiatric
patients that have measured a range of constituents with diverging analytical
methods. On the other hand, we and others have demonstrated that by using
modern techniques many CSF constituents and their genetic underpinnings may be
reliably determined illustrating the feasibility and validity of CSF studies.

The neurobiology of psychotic disorders
From the previous paragraph, it is apparent that CSF may be informative for
neurobiological and immunological processes underlying psychiatric disorders.
One important group of incapacitating mental illnesses is psychotic disorders,
including schizophrenia, bipolar disorder and depression with psychotic
features. These disorders have far-reaching detrimental effects on mental
wellbeing and psychosocial functioning. Current treatments allow reduction of
psychotic symptoms but treatment options are limited due to i) incomplete
treatment response; ii) low treatment adherence (on average 26% at 18 months;
and iii) a high burden of adverse effects. Moreover, the underlying
neurobiology for these disorders remains largely unknown. Measuring CSF
constituents may elucidate CNS processes underlying psychotic disorders.
Therefore, this study aims to measure the following relevant constituents in
psychotic disorders and healthy controls, as mentioned below.

CSF constituents and the underlying neurobiology of psychotic disorders
A range of CSF constituents has been associated with psychosis.
First, it has recently become evident that autoimmune encephalitis caused by
autoantibodies may present with primarily psychotic symptoms, mimicking
psychiatric disorders such as schizophrenia and bipolar disorder.
Autoantibodies to a range of synaptic proteins, in particular to the
N-methyl-D-aspartate receptor (NMDAR), the α-
amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR), the GABAB
receptor, and components of the voltage-gated potassium channel complexes
(VGKCs) can trigger either isolated psychotic symptoms or neurological symptoms
(e.g. seizures and reduced consciousness), with or without delusions and/or
hallucinations. These subtypes of autoimmune encephalitis can cause psychosis,
which is often clinically indistinguishable from other psychotic disorders and
may respond well to immunotherapy. Since the outcome is dependent on the time
lag between appearance of first symptoms and treatment initiation, improved
diagnosis of this subgroup of patients with a treatable cause of psychosis is
essential. Importantly, it has been shown in patients with both neurologic and
psychiatric symptoms that autoantibodies may go undetectable in blood, while
they are present in CSF. However, it has not been investigated to which extent
auto-antibodies are present in the CSF of patients suffering from psychotic
disorders.
Second, several lines of evidence implicate immune pathways in the pathogenesis
of psychotic disorders. This hypothesis is supported by the observation that
genes with a role in the immune system (including the MHC region) are shared
between schizophrenia and bipolar disorder. On the other hand, CSF immune
system parameters have not been comprehensively investigated in psychotic
disorders.
Third, CSF miRNA levels may be informative for specific biological processes
underlying psychotic disorders. MiRNAs are small non-coding RNA molecules (~20
nucleotides) which regulate RNA silencing and gene expression. Evidence linking
miRNA to psychotic disorders stems from genome-wide association studies (GWAS),
including CSF-specific miRNAs 11. Other lines of research also support the role
of miRNAs in the pathophysiology of schizophrenia: postmortem studies, blood
plasma investigations and mouse models. Moreover, miRNAs influence dopamine
receptor 2 (DDR2) expression in humans and cognitive functioning in
schizophrenia. It was recently shown that the relatively small amounts of RNA
in human CSF yield sufficient RNA for both targeted expression profiling and
RNA sequencing. Taken together, although evidence pleads for a role of miRNA in
psychotic disorders, CSF miRNA levels have not been examined in psychotic
disorders.
Fourth, several major neurotransmitter systems that may be measured in CSF have
been linked to psychotic disorders, including the dopamine, GABA and glutamate
systems. Converging evidence points to a role for glutamate in schizophrenia,
and low CSF levels of glutamate have been found in schizophrenia patients
relative to healthy controls. Nevertheless, most later studies have been
negative, but those studies were underpowered and could not pick up small and
medium effect sizes. In bipolar disorder, no study has specifically
investigated glutamate or glutamine concentrations in CSF, but a small CSF
study in a heterogeneous group of patients with affective disorders and a
magnetic resonance spectroscopy (MRS) study detected aberrations in
glutamate/glutamine concentrations in bipolar disorder patients. Other lines of
evidence implicating glutamate in schizophrenia and bipolar disorder include
pharmacological studies demonstrating schizophrenia-mimicking symptoms for the
NMDA receptor antagonist ketamine; postmortem investigations showing NMDAR
NR-subunit RNA expression aberrations in brain tissue of both schizophrenia and
bipolar patients; and solid and consistent genetic findings. Evidence
implicating GABA dysfunction in psychotic disorders comes from preclinical,
electrophysiological and postmortem studies. GABAergic involvement in bipolar
disorder is illustrated by pharmacological studies demonstrating efficacy of
GABA-ergic anticonvulsants and by MRS. In the current study, we will therefore
measure glutamate, the metabolite glutamine and GABA in psychotic disorders and
compare these levels to those in controls.
Finally, several amino acids, e.g. D-serine and glycine, act as coagonists for
the NMDAR, and many of them have been implicated in schizophrenia bipolar
disorder and depression with psychotic features. Small and heterogeneous study
populations and outdated measurement methods with limited sensitivity have
resulted in inconsistent and underpowered studies into neurotransmitter
concentrations in psychotic disorders. In the present study we therefore
compare concentrations of D and L-isoforms of amino acids in psychotic patients
to controls.
Here, we hypothesize that at least 5% of individuals with psychosis and none of
the healthy controls have positive titers of one of the abovementioned
autoantibodies; that immune system parameters, miRNA levels, as well as GABA,
glutamate and amino acid concentrations differ between the two groups; and
finally that subjects with positive titers of autoantibodies carry genetic
mutations in pathways involved with NMDAR, AMPAR, VGKC, and GABA signaling.

Relationships between plasma and central (CSF) levels
Biomarkers for psychotic disorders are currently unavailable, rendering
reliable diagnostics cumbersome given the clinical heterogeneity of these
conditions. A probable explanation behind such lack of reliable biomarkers for
the two maladies constitutes the scarcity of investigations into biomarkers
carried out to date. The focus of previous biomarker studies on peripheral body
fluids - mostly serum and plasma - further compounds this shortfall. Shortage
of data on the association between body fluid constituents and psychotic
disorders particularly applies to CSF. Most studies performed in this context
have targeted monoamine metabolites in patients suffering from a range of
psychiatric disorders. To our knowledge, no study to date has investigated a
comprehensive set of CSF constituents in psychotic disorders to detect possible
case-control differences in concentrations or ratios between constituents. For
some constituents (e.g. auto-antibodies), it is already known that measurements
in peripheral tissues may yield different results compared to CSF analyses.
Importantly, recent evidence shows that auto-antibodies in anti-NMDAR
encephalitis may be negative in serum while they are positive in 100% of the
anti-NMDAR cases. Moreover, changes in CSF titers correlate more strongly with
clinical outcome than serum titers. For psychotic disorders, it is therefore
also important to establish to which extent CSF concentrations are similar to
plasma levels of neurotransmitters and miRNA given the availability of plasma
and therefore the potential of developing blood biomarkers for psychosis.
Therefore, this study aims to compare CSF levels of the abovementioned
constituents to those in plasma levels.

Genetic background may determine CSF constituent levels
Although schizophrenia candidate genes have been replicated, the role of these
genes in the context of CSF intermediate phenotypes has not been investigated.
We previously demonstrated how genetic studies targeting CSF may further our
insight into both genetic determinants of CSF constituent concentration
variability and neurobehavioral traits. In analogy to these previous
hypothesis-driven and hypothesis-generating studies, here we aim to unravel
such genetic determinants in both patients and controls.

Summary
In summary, the rationale to investigate CSF in psychosis is supported by four
lines of reasoning:
First, it is currently unknown which percentage of psychotic patients may
suffer from an underlying autoimmune disorder, e.g. autoantibody encephalitis.
An LP is rarely performed in Dutch departments of psychiatry and autoantibodies
are not routinely investigated.
Second, CSF is the body fluid in closest proximity to the brain and constitutes
a valuable and promising target to gain insight into the neurobiological
background of psychotic disorders, including potential biomarkers, such as
GABA, glutamate, immune parameters and microRNAs.
Third, the genetic underpinnings of CSF variability are currently unknown and
may shed light on the genetic susceptibility for psychotic disorders and the
existing clinical heterogeneity.
Fourth, biomarker research in psychiatry has been hampered by clinical and
methodological heterogeneity, lack of power and absence of replication. As a
consequence, only one biomarker currently meets established biomarker criteria
for psychosis 58. CSF aliquots derived from psychotic patients would allow
researchers who have detected candidate biomarkers and autoantibodies to enable
replication of preliminary findings and cross-validate findings in other body
fluids (e.g. plasma) to CSF.

Potential yields of the study
The potential yields of the current project are:
1. Assessment of the percentage of subjects with positive autoantibodies
against the NMDAR, the AMPAR, the GABABR and components of the VGKCs so that in
the future such patients may be more readily recognized in clinical practice
and treated with immune modulatory agents according to established, efficacious
guidelines.
2. Detection of CSF constituents that are altered in psychotic disorders
compared to healthy controls and possible associations with clinical symptoms
(e.g. specificity to schizophrenia, bipolar disorder, or psychotic depression);
in the long run, this may yield biomarkers that differentiate patient from
control status and be informative for prognosis.
3. Homogeneously collected CSF aliquots for future research involving both
patients with psychotic disorders and healthy controls.

Primary objectives
- To investigate CSF autoantibodies (NMDAR, AMPAR, VGKC, and GABA receptors)
in patients with a psychotic disorder and healthy controls.
- To compare levels of immune parameters between patients with a psychotic
disorder and healthy controls.
- To compare neurotransmitter levels in CSF between patients with a psychotic
disorder and healthy controls.
- To compare microRNA (miRNA) levels in CSF between these patients and healthy
controls.

Secondary objectives
- To correlate CSF levels of the constituents mentioned under the first primary
objective with peripheral plasma levels.
- To unravel the contribution of genetic variation to the variation in the
primary outcomes.
- To relate temporal changes in CSF constituents to clinical outcome, i.e.
improvement in symptoms.
- To assess how specific the detected differences in these constituents are for
psychotic disorders.

This a case-control study. Inclusion is estimated to be completed in
approximately 7 years with an inclusion rate of 3-4 patients with a psychotic
disorder per month and 5-6 healthy controls per month. There are two visits for
patients, during which LPs and venipunctures are performed. CSF sampling at two
time points will allow for association testing with clinical improvement and
other prognostic factors. Healthy controls will also be asked to allow for an
additional LP, thus providing the necessary data to compare CSF concentrations
at repetitive time points between patients and controls.

This study does not include incapacitated subjects. Patients with psychotic or
affective disorders will be given a compensation fee of ¤20 for their
cooperation and time. In addition, possible travel costs will be reimbursed.
Risks for participants are minimal. The most common adverse reaction to a
lumbar puncture (LP) is post-puncture headache, transient nerve root
irritation, lower back pain and a local hematoma (none of which jeopardizes
patients* health). In the patient group we aim to minimize this risk by using
an a-traumatic needle and applying standardized operating procedures (SOPs),
including abiding by the UMCU LP protocol (see
*C1_Addendum_bij_Onderzoeksprotocol_Werkdocument Lumbaal punctie_herzien april
2015.pdf*), performance of LPs by experienced personnel and having the patient
rest for at least 10 minutes prior to standing after the LP. Rare adverse
reactions of LPs include hematoma (spinal, epidural, subdural or intracranial),
infections (meningitis and discitis) and transient brain nerve damage. The
prevalence of minor and transient neurological symptoms may be up to 15%, while
the most serious adverse reaction (paraparesis) is not associated with LPs in
patients who are not on anticoagulants during or shortly after the LP 1.
Patients will be educated in detail about these possible adverse reactions and
as mentioned above we minimize these risks by abiding to evidence-based
standards 2.
Healthy subjects will not be financially compensated. Our experience with these
procedures in previous studies has shown that none of the >550 subjects
included in these studies has suffered from adverse reactions to the lumbar
punctures performed for these studies 3-7. These subjects undergo spinal
anesthesia and a venipuncture for their minor surgical procedure, i.e.
independent of the current study design. They thus do not undergo invasive
procedures for the purpose of this study. Although suctioning of CSF in these
individuals entails a longer LP duration (on average 30 seconds) compared to
when they do not participate in the current project, no other types of burden
are expected for these individuals.
There are no direct benefits for participants. In the future, patients may
benefit from the current study in that treatment consisting of immunotherapy
may be investigated for those patients with both psychosis and positive
auto-antibody titers in their CSF. Moreover, possible aberrant neurotransmitter
or miRNA profiles in their CSF may spur further research into treatment targets
in psychosis.