NIH/KI/UCL Joint Neuroscience Symposium April 9-11, 2019

NIH/KI/UCL Joint Neuroscience Symposium

            April 9-11, 2019

A Computational Approach to Understanding Motivational Symptoms in Depression

J. Roiser

Motivational symptoms of depression are debilitating and associated with poor clinical outcome, but the
mechanisms underlying them are poorly understood. This talk will present data examining how a cognitive
processes related to effort-based decision making for reward are associated with depressive symptoms, using
a computational approach. Results from two studies, including 250 participants (healthy volunteers,
unmedicated depressed patients, first degree relatives and remitted depressed patients), will be presented.
Participants completed a rewarded physical effort task using a grip squeeze, and motivational symptoms were
assessed through questionnaires. Data were analysed using a hierarchical computational approach, with
model parameters estimated in a Bayesian framework using sampling. In the non-clinical study (N=90),
general depressive symptoms were associated with a reduction in reward sensitivity (P

Prefrontal cortical mechanisms of working memory in health and disease

J. Gordon

Advances in our genetic basis of psychiatric illness, coupled with novel technologies to study circuits, have
helped guide mechanistic studies of the neurobiology of behavior in health and disease. Dr. Gordon will
discuss research conducted in his lab and through collaborations, on the molecular and neural circuit basis of
working memory, and how this helps explain cognitive deficits in mouse models of the 22q11.2
microdeletion. His talk will focus on evidence that shows abnormal connectivity and miscommunication
between two brain structures (the prefrontal cortex and hippocampus). He will also discuss pharmacological
interventions during development that may reverse these neural connectivity and behavioral abnormalities.

Examining the effects of motivation and exploration on the neurorehabilitation of upper limb motor
control in patients of acute stroke.

W. Kistler

Understanding and characterizing impairment of the upper-limb (UL) post-stroke remains a fundamental
challenge for neurorehabilitation.1 UL movement is impaired in 50-80% of stroke survivors, and only 50% of
those patients experience substantial functional recovery.2 Behavioural rehabilitation is commonly leveraged
post-stroke as a means to recover from impairment but gains from current practices show no differential
benefits beyond the gains observed from spontaneous biological recovery, a phenomenon by which patients
who recover from impairment will do so to some degree within three months post-stroke. Drawing from
rodent and non-human primate models, neurologists have begun to examine the factors that mediate
spontaneous biological recovery and how they might be individually optimized for humans. Results in
non-human primate studies suggest that components of Enriched Environments over behavioral rehabilitation
may facilitate a vector of recovery similar to spontaneous biological recovery, but they lack a clear
differentiation of effects. Therefore, we have designed a proposal to examine the components of enriched
environments - motivation and motor exploration - and their independent effects on motor control in healthy
subjects and acute stroke patients throughout spontaneous biological recovery.

To measure a recovery of movement from impairment, researchers commonly assess motor control and a
retention of that control over time. In our research, we have reduced motor control to be meaningfully
represented by 3D kinematics, specifically the smoothness of movement trajectory as well as the
speed-accuracy trade-off during trial performance. Primary outcomes will be collected via an accelerometer
with sub-millimeter accuracy attached to the subject wrist. Common reach and navigation task trials will be
performed within a controlled, 3D virtual environment, allowing users mobility along all available degrees.
Secondary outcomes will be the Fugl-Meyer Upper Extremity Exam and the Action Research Arm Test. The
hypothesis presented here is that a focus on movement quality may be the training approach that best exploits
the sensitive period post-stroke to amplify the gains seen with spontaneous biological recovery. This
hypothesis is based on results that optimizing movement quality can enhance recovery for prehension in
rodents and nonhuman primates.3

   1. Tomita Y, Rodrigues MRM, Levin MF. Upper limb coordination in individuals with stroke: Poorly
      defined and poorly quantified. Neurorehabil Neural Repair. 2017; 31:885-897.
   2. Ward N. Restoring brain function after stroke - bridging the gap between animals and humans. Nature
      Reviews Neurology. 2017; 13:244-255.
   3. Krakauer, J. W., Carmichael, S. T., Corbett, D., & Wittenberg, G. F. Getting neurorehabilitation right
      what can be learned from animal models? Neurorehabilitation and Neural Repair. 2012; 26:923-931.

Breaking the neural code with brain perturbations

A. Afraz

A sensory stimulus such as the image of a face activates a cascade of neural responses in the visual system at
different temporal and spatial scales. Some features of this complex pattern of activity are read-out by
downstream neural structures, causally shape the perception of the stimulus and drive the behavior, I refer to
these unknown features as "the neural code". Other features of the neural response pattern are merely
epiphenomenal, driven by perturbing a complex system with a complex stimulus. Development of a rigorous
theory that explains visual perception based on its neural underpinning is impossible without exclusion of the
neural epiphenomena and establishment of quantitative links between causally relevant neural events and
perception. In this talk I will bring together evidence from electrical microstimulation, neuropharmacological
perturbations and optogenetics to investigate the neural code for the case of face recognition behavior in
non-human primate brain.

The effect of ketamine on cortico-striatal circuitry in depressed and healthy individuals

A. Mkrtchian, J. W. Evans, J. P. Roiser, C. A. Zarate Jr.

A sub-anesthetic dose of ketamine can improve depressive symptoms within hours in treatment-resistant
depressed (TRD) patients. These improvements have shown to occur specifically in symptom-clusters related
to lack of motivation or pleasure capacity, even over and above general depressive symptoms. There is also
some recent evidence that ketamine modestly increases similar symptoms for up to one day in healthy
controls, suggesting that ketamine might have different effects in health and disease. However, the precise
neural circuits driving these effects remain unclear. Here we examine if ketamine affects cortico-striatal
circuitry, which is well established in driving goal-directed behavior, in TRD patients and healthy controls.

These questions were explored in the context of a double-blind, placebo-controlled, crossover trial of
ketamine (0.5mg/kg), including 33 TRD patients (27 scans post-ketamine, 25 scans post-placebo) and 25
healthy controls (19 scans post-ketamine, 18 scans post-placebo). Resting-state functional magnetic
resonance imaging (fMRI) data were acquired two days following ketamine and placebo infusions, by which
time ketamine is fully metabolised. Functional connectivity across the entire brain was examined using four
seed regions in the caudate and putamen, reflecting functional subdivisions of the striatum that are connected
with different cortical regions involved in affective, cognitive and motor processes. Linear mixed-effects
models were conducted at the group level in AFNI. Correction for multiple comparisons was achieved using
an uncorrected cluster-forming of p

Observational fear learning: a behaviorally translational approach from mouse to man

S. Silverstein, T. Yoshida, V. Valton, J. Roiser, E. Viding, A. Holmes

  Learning from others' experiences is a highly evolutionarily conserved mechanism providing necessary
information about threat and safety. This behavior is observed across species, including humans. Various
behavioral tasks have been established to assay socially acquired fear, but the literature on observational fear
learning (OFL) remains scarce. First, we established a paradigm in mice using a modified version of standard
cued-Pavlovian fear conditioning, but with two mice - one is exposed to a footshock-tone pairing
(demonstrator), while the other watches (observer). We then assessed the associative memory of the
'observer.' Employing complementary behavioral, anatomical, and in vivo optogenetics, we defined a novel
mechanism subserving observational fear that has not previously been identified. We demonstrated a critical
contribution of the dorsomedial prefrontal cortex (dmPFC) to OFL and showed that ventral hippocampal
(vHPC) inputs to the dmPFC negatively gate OFL. This suggests that the inhibitory vHPC-dmPFC pathway
is critical for OFL.

  Second, we developed a human version of the OFL paradigm and applied detailed computational modelling
of trial-by-trial variation of behavior, supported by Bayesian model comparison, to examine learning rates
and variability during OFL. In this paradigm a video of a man in a similar setup to participants was presented
with two conditioned stimuli - one was associated with a high probability of shock, and another with a low
probability association. Participants were asked to predict whether each stimuli would result in a shock. The
Baysian model comparison revealed participants' choices were best characterized combining learning rates
and variability parameters as opposed to a more simplistic or separate model. We also observed a negative
correlation between learning rate and trait anxiety level indicating individuals with higher levels of anxiety
were slower to update their beliefs as to the likelihood a shock would be delivered. We are currently running
a fMRI study to assess the brain regions engaged during OFL compared to the circuit identified in mice. We
are coupling this information with a computational model to see if learning rate is modulating vHPC or
dmPFC activity. Together, our findings reveal a novel translational model of OFL in both mice and humans
with therapeutic implications for conditions associated with atypical fear learning, including generalized
anxiety and post-traumatic stress disorders.

Spatiotemporal control of ULK1 activation by NDP52 and TBK1 during selective autophagy

J. Vargas, Jose Norberto S. Vargas, Chunxin Wang, Eric Bunker, Ling Hao, Dragan Maric, Giampietro
Schiavo, Felix Randow, Richard J. Youle1

Selective autophagy recycles damaged organelles and clears intracellular pathogens to prevent their aberrant
accumulation. Neuropathological aggregation of "prion-like" proteins are hallmarks of many
neurodegenerative disorders, such as Parkinson's Disease and Amyotrophic Lateral Sclerosis,
indicating a pivotal role of selective autophagy in disease etiology. ULK1 complex is the most upstream
autophagy machinery that functions to initiate autophagosome biogenesis. How ULK1 kinase is targeted and
activated during selective autophagic events remain to be elucidated. We used chemically-inducible
dimerization (CID) assays in tandem with CRISPR KO lines to systematically analyze the molecular basis of
selective autophagosome biogenesis. We demonstrate that ectopic targeting of autophagy receptor NDP52 to
mitochondria or peroxisomes is sufficient to initiate selective autophagy by focally localizing and activating
the ULK1 complex. The capability of NDP52 to induce mitophagy is dependent on its interaction with the
FIP200/ULK1 complex, which is facilitated by TBK1 kinase. Ectopically tethering ULK1 to cargo bypasses
the requirement for autophagy receptors and TBK1. The focal activation of ULK1 occurs independently of
AMPK and mTOR, suggesting that ULK1 cargo-dependent activation circumvents bioenergetic
sensors. These findings provide a parsimonious model for selective autophagy, which highlights the
coordination of ULK1 complex localization by autophagy receptors and TBK1 as principal drivers of
targeted autophagosome biogenesis.

AudioVisual Integration in Anterior Fundus Face Patch

A. Khandhadia, Aidan P. Murphy, Lizabeth Romanski, Jennifer K. Bizley, David A. Leopold

        The macaque superior temporal sulcus (STS) is a major area of convergence of socially relevant
stimuli, for example integrating visual and auditory signals. In the visual domain, the STS contains several
face patches, which are defined as brain regions more responsive to images of faces than non-face objects.
Within these patches, the influence of auditory information on neural responses is largely unknown, for
example during the issuance of an acoustic vocalization. In this experiment, we localized face patches using
fMRI and surgically targeted the anterior fundus face patch (AF), nested in the STS, for recording with
microwire bundles. We assessed the responses of single neurons of AF to fifteen audiovisual clips of
macaque vocalizations, as well as their audio and visual components in isolation. The results indicate a high
proportion of cells in AF are affected by auditory stimuli, which is often expressed principally as a
modulation of the visual response to the dynamic face sequence. Such modulation generally showed strong
stimulus specificity expressed both as amplification and suppression of the spike rate for different stimuli. A
relatively small subset of cells also responded to the auditory vocal stimuli alone. These results demonstrate
that neurons within the AF face patch incorporate auditory signals to shape spiking responses to observed
socially relevant facial behaviors.

RESCUE OF A42-INDUCED CELLULAR, NETWORK AND COGNITIVE DEFICITS BY
TARGETED ACTIVATION OF THE PROTEASOME

A. Fisahn, M Altun, D Papadia, C Marks, J Johansson, A Falk, A Acebes

Objectives: In the healthy brain excess peptides are cleared by the proteasomal system. Why it is unable to
clear aggregated amyloid-beta in AD is unclear. We tested whether targeted activation of the proteasome
rescues disease-typical impairments in several animal and human assays.

Methods: Two FDA-approved small molecule compounds that caused proteasome activation were tested for
their ability to rescue amyloid-beta-induced cellular, network and behavioural deficits in a Drosophila fly
model overexpressing amyloid-beta, in iPS cells derived from control and AD patients and in WT mouse
brain slices.

Results: Drosophila assay: behavioral climbing test showed functional improvement after treatment.
Number/size of amyloid-beta aggregates was reduced.

Human assay (iPSC): AD cell line exhibited lower resting membrane potential and reduced ability to fire
action potentials. Both parameters were rescued to control levels after treatment.

Mouse assay: Rhythmic electrical activity in the gamma-frequency range (30-80Hz; gamma oscillations)
underlies higher brain processes such as cognition and correlates with the progression of AD and impairment
of cognition. Therefor gamma oscillations can serve as a functional biomarker for diagnosis in the clinic. In
mice gamma oscillations and cellular/synaptic parameters are impaired by amyloid-beta and rescued by
subsequent treatment with the two proteasome activator compounds.

Conclusions: Activation of the proteasome leads to gain-of-function in relevant several assays. The
mechanism underlying the proteasome activation offers a novel therapeutic approach to AD and other
amyloid diseases. Behavioural experiments in an AD mouse model are ongoing. Patent pending.

A special relationship: my life as a transatlantic PhD program test pilot

R. Kaplan

I joined the NIMH as a post-baccalaureate IRTA fellow in order to gain more research experience for my
neuroscience PhD program applications. When I was subsequently applying for graduate programs, I was
pleasantly surprised that a partnership program was started between NIMH and University College London. I
immediately applied and ended up being the first PhD student in the program. In this talk, I'll detail my
experiences in the UCL-NIMH program and how some initiative from Jon Roiser and Barry Kaplan
helped me embark on the path that I'm on today.

Positron emission tomography of repetitive blast mild traumatic brain injury in mouse and human
brain

G. Terry

Blast-induced mild traumatic brain injury (mTBI, aka concussion) is the "signature injury" of the wars in Iraq
(OIF) and Afghanistan (OEF). Blasts account for 70-88% of mTBIs sustained in OEF/OIF, with affected
Servicemembers usually experiencing multiple blast mTBIs. Even years after returning from deployments,
many Veterans with blast-related mTBI report postconcussive cognitive, behavioral, and somatic symptoms,
leading to substantial disability and interference with job and family relationships. Individuals receiving
multiple mTBIs are generally young, by definition survive their initial injuries, and, as a result, few
postmortem brain samples are available to examine the neuropathology resulting from blast mTBI directly.
Effects of mTBI and post-concussive symptoms have been difficult to map to neuroanatomic correlates since
mTBI lacks any clinical or structural neuroimaging abnormalities by definition. Positron emission
tomography (PET) can assess functional and biochemical abnormalities, and may be a promising approach to
examine the chronic effects of blast mTBI on brain.

Our group's initial [18F]FDG-PET study demonstrated a consistent pattern of less [18F]FDG uptake in
infratentorial structures and medial temporal cortex in mTBI Veterans (n=12 males) compared to community
controls (n=7 males, 5 females). In our follow-up study of 33 Veterans with blast mTBI and 16 deployed
control Veterans with no lifetime history of mTBI, voxelwise analysis demonstrated hypometabolism in
bilateral parietal lobes, left somatosensory cortex, and right visual cortex, as well as in parahippocampus for
Veterans with 20 or more blast mTBI. There was a significant negative correlation between glucose
metabolism in the cerebellum and number of blast exposures (i.e., more blast exposures was associated with
less [18F]FDG uptake). A follow up analysis in a larger cohort of Veterans with mTBI (n=79; average blast
mTBI exposures 22) confirmed this correlation, and also revealed a positive correlation with [18F]FDG
uptake and time since last experienced blast mTBI (i.e., individuals imaged at longer intervals after exposure
had greater [18F]FDG uptake). As members of this cohort have had no measurable functional improvement
on longitudinal assessment, this greater glucose utilization over time might represent ongoing inflammatory
neuropathology.

The translational capability of PET also provides opportunity to compare neuroimaging outcomes to
neuropathology in our battlefield-relevant mouse model of blast mTBI. We have shown that blast mTBI in
mice is capable of provoking an early-occurring and persistent elevation of a large number of phosphorylated
and pathologically cleaved tau species in multiple brain regions, and persistently elevated IL-6, effectively
replicating our finding in human CSF. Furthermore, blast-exposed mice demonstrate persistent microglial
pathology strikingly similar to the neuropathology very recently identified in Veterans with blast-induced
mTBI. Evidence from our mouse model strongly suggest that following mTBI early-occurring
neuroinflammation mediated by microglia leads to a pathogenic cascade of long-term neuropathology.
Imaging microglia and/or neuroinflammation may provide insights into the regional and longitudinal
neuropathology in mouse brain, and could be translated to study in humans with blast mTBI.

PET imaging of phosphodiesterase 4 in brain and peripheral organs of McCune-Albright syndrome

L. Deuitch Weidner, Michael T. Collins, Alison Boyce, Yuichi Wakabayashi, Louise A. Stolz, Lori Guthrie,
Denise Rallis-Frutos, Sami S. Zoghbi, Victor Pike, Masahiro Fujita, Robert B. Innis

Objective: McCune-Albright syndrome (MAS) is a mosaic disorder arising from mutations of the GNAS
gene,which encodes the 3, 5-cyclic adenosine monophosphate (cAMP) pathway-associated G-protein, Gs
alpha.This mutation results in dysregulation of the cAMP signaling cascade, leading to upregulation of
phosphodiesterase type 4 (PDE4), an enzyme that catalyzes the hydrolysis of cAMP. Positron emission
tomography (PET) imaging of PDE4 using [11C](R)-rolipram has been successfully used to study the in vivo
activity of the cAMP cascade. Rolipram is a reversible inhibitor of PDE4, and binding of [11C](R)-rolipram
provides a measure of the activity of this enzyme in brain. Due to a feedback mechanism, in vivo binding of [
11C](R)-rolipram reflects the activity of the cAMP cascade; essentially, increased cAMP stimulates protein
kinase A (PKA), which phosphorylates PDE4 that, in turn, increases rolipram binding. Clinical
manifestations of MAS, such as fibrous dysplasia, in a given individual are determined by the timing of the
GNAS mutation during embryogenesis, the tissues involved, and the role of G protein, Gs alpha, in affected
tissues. Animal models of fibrous dysplasia have shown increases in PDE4 activity, however this correlation
has not been shown in humans with fibrous dysplasia. Therefore, it is unknown whether fibrous dysplasia,
and other symptoms of MAS, is related to increased PDE4 activity in humans. We hypothesize that subjects
with MAS would show greater rolipram binding than healthy controls in areas known to affected by the
disorder.

Methods: [11C](R)-rolipram whole body PET scans were performed in two patients with MAS and six
healthy controls. Uptake of [11C](R)-rolipram was measured in organs such as the heart, kidneys, and liver,
as well as in the brain.

Results: Binding of [11C](R)-rolipram in the body correlated with known locations of fibrous dysplasia in
patients with MAS, with no uptake in the bones of healthy controls (Figure 1). SUV brain TACs showed a
two-fold difference in uptake in MAS patient #2 (SUVmax = 4) as compared to MAS patient #1 and healthy
controls (SUVmax = 2). While we could not correct for differences in brain exposure without arterial blood
samples, this study is ongoing and we will be obtaining dedicated brain scans. VT/fP values for healthy
control brain regions were similar to those previously reported.

Conclusion: Our results are consistent with animal studies showing that increased cAMP leads to activation
of PKA, which in turn increases the phosphorylation of PDE4 (which has increased activity and ten-fold
higher affinity for rolipram). This study revealed that rolipram binding does reflect increased cAMP pathway
activation. To determine if the increased binding is specific, we will perform whole-body blocking scans with
the PDE4 inhibitor roflumilast. Differences in rolipram binding in the brain will be determined by
radiometabolite-corrected arterial sampling.

Screening for novel inhibitors of glutathione peroxidase

D. Cheff, Q. Cheng, M.D. Hall, E.S.J. Arnér

Glioblastoma Multiforme (GBM) is an aggressive malignant tumor deriving from glial cells in the brain, with
poor prognosis and limited treatment options. Cancer cells have higher basal levels of reactive oxygen
species (ROS) production, and altered expression of antioxidant enzymes, when compared to normal cells. At
high levels, ROS can cause substantial cellular damage, often leading to cell death, and many cancer
therapeutics function through targeting these imbalanced redox systems. Glutathione peroxidases (GPx) are
selenoenzymes that protect cells from increased levels of ROS and oxidative stress by coupling the reduction
of hydrogen and organic peroxides with the oxidization of glutathione to glutathione disulfide [1]. GPx levels
have been found to be altered in GBM tissues, and additionally, GPx expression has been found to modulate
sensitivity to radio- and chemo-therapy, suggesting a possible protective role. [2-4] Because of this role of
glutathione peroxidases in controlling ROS levels, reducing oxidative stress, and regulating cell death
pathways, we are interested in identifying inhibitors of GPx1 and GPx4. Moreover, GPx4 has recently been
identified as a unique and important regulator of non-apoptotic cell death pathway, ferroptosis [5]. Using a
novel method of recombinant GPx expression [6] we are developing an assay amenable to high-throughput
screening of small molecules targeting GPx, and will characterize the identified inhibitors as potential
chemotherapeutics for treating Glioblastoma Multiforme.

[1] PMID: 13491573

[2] PMID: 28552540

[3] PMID: 22951908

[4] PMID: 11260655

[5] PMID: 24439385

[6] PMID: 28193838

Assessing embodiment in elite prosthesis users and expert tool users

H. Schone, R. Maimon, C. Baker, T. Makin

Advancements in robotics and information technology have led to the development of highly innovative
artificial prosthetic limbs. Despite these advancements, low rates of prosthetic hand usage and even complete
device rejection are commonplace amongst amputee populations (Cordella et al. 2016). While significant
efforts are being taken to develop technological solutions to combat these issues, little attention is given to
potential neurocognitive bottlenecks, such as embodiment (Makin et al. 2017). Recent evidence from our lab
has shown that the more amputees use a prosthesis, the less their brain represents the prosthesis as either tools
or hands. However, because the nature of a prosthesis is to act as a hand substitute, it should have a greater
propensity to be embodied. Therefore, we designed a study around individuals with the most successful
prosthesis usage, elite prosthesis users, to investigate if a prosthesis is represented as an expert tool or a hand
substitute. We included two control groups: able-bodied expert tool users (litter-pickers) and able-bodied
novices. Using fMRI, we evaluate the representation structure of hands, tools and prostheses, across the
body-representation and tool-use networks. To activate these representations, we presented first-person
perspective videos depicting actions performed by either a hand, a prosthesis, a litter-picker or an unknown
tool. Using a large battery of well-defined cognitive and motor tasks, we further evaluate the level of
embodiment and expertise of the prosthesis and tool users. By contrasting elite prosthesis users with tool
experts, our results help to elucidate the challenges and benefits of prosthesis usage on brain reorganisation
and embodiment.

Nerve growth factor-a potential therapy in Alzheimer´s disease

M. Eriksdotter

One of the most prevalent forms of dementia is Alzheimer's disease (AD), the symptoms of which are
deterioration of memory loss and cognitive functions brought about partly by a decrease in cholinergic
neurotransmission. At present there is no cure, but symptomatic treatment with cholinesterase inhibitors
(ChEI) is recommended. Loss of memory in AD is associated with central cholinergic dysfunction in the
basal forebrain, from where the cholinergic circuitry projects to cerebral cortex and hippocampus. The
neurotrophic protein, nerve growth factor (NGF), contributes to the survival and regeneration of cholinergic
neurons during aging and in age-related diseases such as AD. Studies on AD have found decreased levels of
nerve growth factor (NGF) in the basal forebrain and a dysregulation during AD development making NGF a
potential agent for a disease-modifying therapy. There have been several approaches for delivery of NGF,
many however hampered with side effects. To circumvent problems with negative side effects we have used
encapsulated cell biodelivery (ECB) of human NGF to the basal forebrain for 6-12 months in a pilot study
with ten patients suffering from mild to moderate AD. The results demonstrated safety and feasibility in all
patients and lower rates of brain atrophy, stable cognitive function, and improvement in cerebrospinal fluid
(CSF) cholinergic markers in a subset of patients. However, cell survival and NGF release differed between
devices and research is ongoing to study what factors affect cell survival and release.

Further development of therapeutic platforms such as the ECB and its efficacy in terms of cognitive
outcomes in AD and other neurodegenerative diseases is needed. A combined treatment with NGF and also
other neurotrophins important for cognition such as brain derived growth factor (BDNF) could be an
attractive therapeutic approach for AD neurodegeneration. As an additional option or in combination with the
ECB technique, synthetic modulators of NGF or BDNF signaling to stimulate neurotrophic signaling in AD
will also be discussed.

Stretching the Genetic Code for Improved Recombinant Selenoprotein Synthesis

Q. Cheng, DM. Cheff, MD. Hall, ESJ. Arnér

Keywords
Recombinant selenoprotein, Glutathione peroxidase, High-throughput screening, Glioblastoma Multiforme

Abstract
 Selenoproteins, containing one or several Sec residues, are found in all three domains of life (1). It is
important to characterize and understand selenoprotein function and role in life. We recently illustrated a
surprising flexibility of selenoprotein synthesis, which opens up a powerful method for selenoprotein
research using recombinant selenoprotein (2). In nature, Sec is co-translationally inserted at a predefined
UGA opal codon by specific translation machineries (3), here we found that Sec can be equally efficiently
incorporated at a predefined UAG amber codon, thereby competing with RF1 rather than RF2. By utilizing
this system within a RF1 depleted E. coli strain C321.deltaA (4) we could produce a number of
selenoproteins with unsurpassed purity and yield. We have produced several form of mammalian TrxR with
much higher specific activity compared to previous methods (5) and we believe this routine should be widely
employed for production of a wide variety of native selenoproteins carrying penultimate or ultimate Sec
residue. Interestingly, SECIS element was no longer required in such system, which allowed us to produce
selenoprotein with internal Sec, such as various isoforms of human glutathione peroxidase (GPx). We
subsequently confirmed a previously proposed catalytic tetrad in GPx1 (6) utilizing this methodology.
Detailed studies of most selenoproteins have hampered by major hurdles in their recombinant production, due
to the unique features of Sec incorporation machinery. We conclude that our new methodology should
constitute a powerful resource enabling many forthcoming studies of recombinant selenoproteins,
considering its versatility and comparable ease of use. As a collaboration between KI and NIH, we have
recently initiated a high-throughput project to screen small molecules targeting recombinant GPx4, which
could potentially sever as a chemotherapy for treating Glioblastoma Multiforme (GBM) utilizing novel GPx4
inhibitors.

Reference
1. GV. Kryukov et al. Science. 300, 1439-1443 (2003).
2. Q. Cheng, ESJ. Arnér. J Biol Chem. 292, 5476-5487 (2017).
3. M. Thanbichler, A. Böck. EMBO J. 21, 6925-6934 (2002).
4. MJ. Lajoie et al. Science. 342, 357-360 (2013).
5. ESJ. Arnér et al. J Mol Biol. 292, 1003-1016 (1999).
6. SC. Tosatto et al. Antioxidants and Redox Signalling. 10, 1515-1526 (2008).

Affiliation
Department of Medical Biochemistry and Biophysics, Karolinska Institutet. Stockholm, Sweden (QC, DMC and ESJA)
NCATS Chemical Genomics Center, National Institutes of Health. Maryland, United States (DMC and MDH)

How did we get there? Building a mood machine interface at the NIH

A. Stringaris, H. Keren

In this talk we describe a reward based closed-loop system designed to modify mood. This paradigm brings
engineering control theory methodologies to the study of mood and depression. While a typical experimental
strategy would be to measure mood fluctuations in response to changing stimuli values, this experiment's
target is to generate a predefined mood transition. By artificially shifting mood and tracking the individual
stimuli required to induce this mood change, we are able to quantify mood reactivity and its relations to
underlying behavior and neural signals.

The task works by tracking in real time a subject's momentary mood and then adapting respectively the next
reward stimulus, to push mood upwards or downwards. Based on prior research we devised an algorithm that
uses subjective mood rating as the output for control and Reward Prediction values as the driving stimulus.
We validate this paradigm on 68 adolescent subjects, who did the adaptive task in the fMRI scanner. We then
implement computational models to combine measures of mood reactivity, decision making, and neural
encoding of error signals. Moreover, we find that these parameters are informative about depression severity.
This development is relevant to control-theory applications in medicine and potentially to treatment through
closed-loop modification.

Targeting specific selenoproteins for drug therapies in human diseases

E. Arnér

Selenoproteins are a unique class of proteins containing a catalytic selenocysteine (Sec) residue, which is the
21st naturally occurring amino acid I protein synthesis. It is an analog of cysteine containing selenium instead
of sulfur, which makes Sec chemically more reactive than Cys. Humans have 25 genes encoding
selenoproteins, E. coli has three genes, C. elegans only one, with most plants and yeast completely lacking
selenoproteins. Sec is cotranslationally incorporated at a UGA codon by an extension of the genetic code,
which is enabled by a specific tRNA, a Sec-dedicated elongation factor interacting with specific secondary
structures of the selenoprotein-encoding mRNA., and accessory selenoprotein translation factors This
intricate translation machinery makes it difficult to express recombinant selenoproteins. We have nonetheless
developed methods to overcome the technical challenges of selenoprotein production, which has been the
basis for a long-standing collaboration with NIH/NCATS aimed at discovering novel drugs targeting
selenoproteins having therapeutic potentials. Previous such collaborative work includes discoveries of novel
drug leads for treatment of Schistosomiasis, targeting the TGR selenoprotein of Schistosoma mansoni [1-3],
as well as attempts to develop new anticancer compounds based upon irreversible inhibition of the human
selenoprotein TrxR1 [4-5]. Both of these selenoenzymes have their Sec residues close to their C-terminal
ends, which has facilitated their expression as selenoproteins in E. coli. Pitfalls, promises and further
developments of these projects will be discussed in this presentation. We furthermore recently developed
additional methods to express selenoproteins having internal Sec residues, as presented at this symposium by
Dr. Qing Cheng, and are currently working together with NIH/NCATS on the development of new
therapeutics for glioblastomas based upon inhibition of glutathione peroxidases, which is further presented at
this meeting by Ms. Dorian Cheff - a PhD student in the joint KI-NIH program studying with us at KI
together with Dr. Matthew Hall at NIH/NCATS.

[1] Silvestri I, et al. (2018) Fragment-Based Discovery of a Regulatory Site in Thioredoxin Glutathione
Reductase Acting as "Doorstop" for NADPH Entry. ACS Chem Biol. 13:2190-2202.

[2]: Rai G, et al (2009) Structure mechanism insights and the role of nitric oxide donation guide the
development of oxadiazole-2-oxides as therapeutic agents against schistosomiasis. J Med Chem. 52:6474-83.

[3]: Kuntz AN, et al (2007) Thioredoxin glutathione reductase from Schistosoma mansoni: an essential
parasite enzyme and a key drug target. PLoS Med. 4:e206.

[4]: Stafford WC, et al (2018) Irreversible inhibition of cytosolic thioredoxin reductase 1 as a mechanistic
basis for anticancer therapy. Sci Transl Med. 10: eaaf7444.

[5]: Prast-Nielsen S, et al (2011) Inhibition of thioredoxin reductase 1 by porphyrins and other small
molecules identified by a high-throughput screening assay. Free Radic Biol Med. 50:1114-23.

Transatlantic studies of neural activity during spoken interactions

K. Jasmin

The human voice is a remarkable tool that allows people to exchange ideas and form social bonds. However,
the bulk of neuroimaging studies on speech and language have used tasks involving only one person (the
participant in the scanner). In this talk, I will discuss two studies undertaken during my time in the
UCL-NIMH program which used live, real-time spoken interaction tasks involving two people. The first,
from Sophie Scott's lab at UCL, investigated joint, synchronized speech, i.e. that sort of speech used during
communal chants and prayers (Jasmin, et al., 2016, J Neurosci). The second study, from Alex Martin's lab at
LBC/NIMH, examined face-to-face conversation in participants in Autism Spectrum Disorder (ASD), and
matched controls (Jasmin et al., 2019, Brain). Neither study would have been possible without the generous
support and infrastructure supported by the UCL-NIMH program.

Do as I say, Not(ch) as I do: dissecting mechanisms governing cell fate choices

S. de Haan, M.W. Kelley, E.R. Andersson

The Notch signalling pathway is a highly conserved signalling pathway comprised of membrane bound
ligands and receptors which interact to influence different aspects of development in multicellular organisms,
including cellular patterning, proliferation, and differentiation. Two examples of Notch-mediated processes
are lateral induction and lateral inhibition. Inductive interactions typically lead to both cells adopting the
same fate while inhibitory interactions have the opposite outcome. While Notch-mediated lateral induction
and inhibition have been widely studied, the molecular factors that determine whether an inhibitory or
inductive response will occur remain to be elucidated. A classic example in which functionally different
outcomes of the single Notch signalling pathway become evident is the highly stereotypical patterning of the
developing inner ear. Several rounds of lateral induction and inhibition, each regulated by Notch signalling,
are required for development of a functioning inner ear. In this project we aim to use the developing inner ear
as a model to dissect the roles of Notch-components and regulators during different inductive or inhibitory
interactions. To do this, we will analyse RNA-sequencing data of the developing inner ear form the Kelley
lab (NIDCD/NIH) to identify Notch signalling pathway components potentially involved in lateral inhibition
and induction during development. The effects of specific candidates will then be tested by conducting loss
and gain of function experiments in vivo using ultrasound guided nano-injections targeting the inner ear at the
ERA lab (KI). Together, these experiments will increase the basic understanding of Notch signalling
mechanisms governing cell fate decisions.

Sex Differences in Functional Connectivity of the Cerebellum in Autism Spectrum Disorder

R. Smith, Jason A. Avery, Gregory L. Wallace, Lauren Kenworthy, Stephen J. Gotts, & Alex Martin

Autism Spectrum Disorder (ASD) is more prevalent in males than females, but the
underlying neurobiology of this sex bias remains unclear. Given its involvement in
ASD, its role in sensorimotor, cognitive, and socio-affective processes, and its
developmental sensitivity to sex hormones, the cerebellum is a candidate for
understanding this sex difference. The current study used resting state functional
magnetic resonance imaging (fMRI) to investigate sex-dependent differences in
cortico-cerebellar organization in ASD. We collected resting-state fMRI scans from 47
females (23 ASD, 24 controls) and 120 males (56 ASD, 65 controls). Using a measure
of global functional connectivity, we ran a linear mixed effects analysis to determine
whether there was a sex-by-diagnosis interaction in resting-state functional
connectivity. Subsequent seed-based analyses from the resulting clusters were run to
clarify the global connectivity effects. Two clusters in the bilateral cerebellum
exhibited a diagnosis-by-sex interaction in global connectivity. These cerebellar
clusters further showed a pattern of interaction with regions in the cortex, including
bilateral fusiform, middle occipital, middle frontal, and precentral gyri, cingulate
cortex, and precuneus. Post-hoc tests revealed a pattern of cortico-cerebellar
hyperconnectivity in ASD females and a pattern of hypoconnectivity in ASD males.
Furthermore, cortico-cerebellar functional connectivity in females more closely
resembled that of control males than that of control females. These results shed light on
the sex-specific pathophysiology of ASD, and are indicative of potentially divergent
neurodevelopmental trajectories for each sex. This sex-dependent, aberrant cerebellar
connectivity in ASD might also underlie some of the motor and/or socio-affective
difficulties experienced by members of this population, but the symptomatic
correlate(s) of these brain findings remain unknown.

Making sense of real world scenes

C. Baker

The goal of the human visual system is to extract behaviorally-relevant information from complex visual
environments (scenes) to perform a multitude of tasks. In this talk I will present recent work from my lab
investigating both the perception and memory of such real world scenes. First, I will show how we have used
variety of different neuroimaging and computational approaches to reveal what visual and functional
properties are represented in different brain regions that have been implicated in scene perception. Second, I
will discuss scene memory and how we can uncover what aspects of real world scenes are held in memory by
analyzing thousands of drawings from a free recall task. Collectively, these studies provide important insights
into how we understand real world scenes.

Development of psychopathology: How can neurocognitive research improve our understanding of
environmental risk?

E. Viding

Much of the neurocognitive research into various psychiatric disorders has been cross-sectional and has
focused on specific processes that are atypical in a given disorder. To progress our understanding of how
psychopathology develops, we need to combine different analytical approaches within a longitudinal,
developmental, genetically informative framework. This can help us understand phenomena of
gene-environment and environment-environment correlation.

In this talk I will provide a brief overview of neurocognitive and genetically informative research into
developmental risk for conduct disorder. I will use this overview as a framework for considering
how atypical neurocognitive functioning may serve to generate and maintain maladaptive social interactions.
I will argue that neurocognitive studies can inform our understanding of individuals as active agents in the
generation of particular social ecologies and that unlocking the mechanisms of gene-environment and
environment-environment correlation will be of key importance. Advances in this area of research have scope
to inform theoretical understanding, as well as interventions designed to help children at risk of developing a
disorder and their families.