Autism Spectrum Disorder (ASD)
Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder. Affected kids have impaired social communication, repetitive behavior, developmental delays, and other symptoms. The diagnosis is done at early childhood.
Our lab works with several monogenic forms of this disorder. In the lab, we reprogram fibroblasts or lymphoblasts from children with the disorder. We work in collaboration with Sheba Medical Center, Shaare Zadek Medical Center, Emek Medical Center, and sometimes directly with the parents. After reprogramming the patients' cells into induced pluripotent stem cells (iPSCs), we differentiate the iPSCs into hippocampal and cortical brain organoids and neuronal cultures. Currently we have in the lab iPSCs and brain organoids derived from patients with a variety of ASD related mutations, such as mutations in iQSEC2, SHANK3, SLC1A4, GRIN2B, RAI1, KAT6A, 7DUP, SMS, UBTF, and more.
​
Currently, there are no effective treatments that address the core deficits of this disorder. Electrophysiological characterization using whole-cell patch-clamp recordings, calcium imaging, and network activity analysis revealed a biphasic excitability phenotype in patient-derived neurons: hyperexcitability at early stages, followed by hypoexcitability and severely reduced synaptic currents in more mature neurons. To identify potential therapeutic strategies, we conducted a drug screen of over 20 repurposed, FDA-approved compounds on patient-derived neurons.
​
Several drugs significantly rescued key ASD-associated phenotypes, particularly the synaptic deficits. Notably, some compounds showed broad efficacy across neurons derived from several patients, suggesting the potential for mutation-independent therapeutic approaches. Encouragingly, a subset of these candidate drugs has already been administered to some patients, with clinicians and caregivers reporting marked symptomatic improvements. These findings highlight the promise of personalized, stem cell-based models for uncovering effective treatments for rare genetic neurodevelopmental disorders.
White blood cells from a child with ASD
Reprogrammed in the lab using Sendai Vectors
iPSCs from a child with ASD deletion syndrome
Differentiation into cortical neurons
Cortical neurons from a child with ASD
Parkinson’s disease (PD)
Parkinson’s disease was first described in 1817 by James Parkinson. He described it as a “shaking pulsy”; involuntary tremulous motion with lessened muscular power. Patients with PD experience movement difficulties with three cardinal signs: tremor, rigidity and bradykinesia. Other non-motor symptoms include loss of smell, dementia, sleep disorders, reduced bowel movement and more. Pathological examination of patients’ brains shows aggregates of the α-synuclein protein in the patients’ nerve cells. These aggregates are termed Lewy bodies and are considered a hallmark of Parkinson’s disease. These aggregates are hypothesized to cause the severe loss of dopaminergic neurons in the patients’ substantia nigra pars compacta, an area in the midbrain, very densely populated by dopaminergic neurons that indirectly regulate fine motor movement.
​
There are a few genes in which mutations were shown to have a causative effect of this disease. These include mutations in the alpha synuclein (SNCA), leucine rich repeat kinase 2 (LRRK2), parkin RBR E3 ubiquitin protein ligase (PRKN), PTEN-induced kinase 1 (PINK1), glucosylceramidase beta (GBA) and Parkinsonism associated deglycase (PARK7). There are many more genes associated with the disease, but without a definite causative role. However, these mutations account for about 15% of the Parkinson’s cases, while the rest of the cases are considered sporadic, and the disease cause is not known.
​
We have fibroblasts that were reprogrammed into iPSCs from human patients with different PD causing mutations, as well as sporadic PD. We differentiate them into dopaminergic neuronal cultures and into midbrain organoids enriched with dopaminergic neurons. We see a distinct electrophysiological phenotype that is common to dopaminergic neurons that were derived from PD patients, and we are investigating how these physiological alterations that we see when the neurons are young, trigger a cascade of events that eventually lead to cell death that is specific to dopaminergic neurons in PD.
​
Bipolar disorder (BD)
Already at the 1st century Aretaeus of Cappadocia, a Greek physician, wrote about a link between mania and depression “Melancholia is the beginning and a part of mania. The development of a mania is really a worsening of the disease (melancholia) rather than a change into another disease”. Many centuries later psychiatrists and physicians such as Teophilus Bonet (17th century), Jean-Pierre Falret (19th century) and Jules Baillarger (19th century) also described a disease where mania and depression cycle within the same patient. The term bipolar (two poles) first appeared in the American Psychiatric Association’s Diagonostic and Statistical Manual of Mental Disorders (DSM) in 1980.
​
Bipolar disorder (BD) is a progressive psychiatric disorder with 1-2% prevalence worldwide. Patients experience recurrent episodes of depression and mania and are at a high risk of suicide. The first line of treatment usually involves lithium derivatives. While lithium helps many BD patients, unfortunately not all respond with a good outcome to lithium treatment. The genetics of BD is complex with many associations, but no genes that were identified as causative. This genetic heterogeneity has challenged the development of animal and cellular models, and the available animal models do not fully recapitulate the disease phenotype. The use of neurons derived from iPSCs of human patients has allowed us to identify a physiological phenotype of hyperexcitability in hippocampal neurons that were derived from BD patients. BD hippocampal neurons have a remarkable ability to sustain activity over long periods of time.
Interestingly we had shown that BD neurons share a common phenotype of hyperexcitability but divide into sub-populations with distinct features. Neurons derived from patients who respond to lithium treatment are fast spiking, while those derived from patients who do not respond to lithium are physiologically unstable and strongly shift their excitability with small changes in their environment. Overall these two sub classes are so different from one another, that it is possible to predict directly from physiological measurements of these neurons which of the patients will respond to lithium.
Stern et al. Molecular Psychiatry 2018
After measuring the neurophysiological properties of dentate gyrus and CA3 pyramidal hippocampal neurons, we were able to reveal alteration in ionic currents that cause the hyperexcitability. Further, we observed that hippocampal neurons derived from BD patients who do not respond to lithium, possess a physiological instability causing them to shift drastically between hypoexcitable and hyperexcitable states, leaving the neuronal network in a multi-excitatory state and possibly contributing to affective lability in BD patients. Strikingly, over prolonged recordings, the BD network sometimes self-organizes in a global hypoexcitability state, and sometimes in a global hyperexcitability state.
While the healthy brain consists of neurons with similar excitability, the BD brain consists of a large proportion of neurons that are very hyperexcitable or hypoexcitable. Conditions occur occasionally that causes a global shift towards super activity and sometimes towards complete inactivity, suggesting behavioral states such as mania and depression
Stern et al. Biological Psychiatry 2019
Stern et al. Biological Psychiatry 2020
​
Biomarkers for drug response
Precision medicine has revolutionized cancer treatment by the use of new techniques such as DNA sequencing. This revolution has increased survival rates dramatically. Other fields of medicine, such as psychiatric disorders are lagging far behind due to lack of investment along with the complexity of these disorders. Owing to a limited understanding of the pathophysiology of these disorders, treatment options follow a trial-and-error strategy, sometime having months and years pass until finding a suitable drug. During this time, the patients are practically untreated, and often deteriorate to a point of no return, losing their work place, their friends and their ability to function in society. This is truly unfortunate, since especially in psychiatric disorders, sometime small interventions can save a life both figuratively and literally.
​
Multiple studies suggest sub populations within different psychiatric disorders. Within each sub-population, the symptoms are more homogeneous, sometimes they share familial genomic associations, and their response to certain drugs may be similar. Our previous work with bipolar disorder also suggests sub populations in which the physiology is so different that drug
response can be predicted. We have a large set of immortalized lymphoblasts from patients with characterized drug response and we use different biological features from these lymphoblasts to predict drug response in psychiatric patients. We are developing computational models to build classification schemes to find the most suitable drugs.
Stern et al. Open Biology 2018
​
​
We have developed a neural network-based algorithm that can precisely predict how a bipolar disorder patient would respond to lithium treatment. This method is cheap, fast, and reliable. An invention disclosure was submitted. Check out the exciting results in Mizrahi at al.
_edited.jpg)
_edited.jpg)
_edited.jpg){kind=small}
_edited.jpg){kind=small}
LOG2FoldChange
Analysis of RNA sequencing results confirms that there are many differentially expressed genes between LCLs of bipolar disorder patients who respond to lithium and those who do not respond to lithium. We have used these genes to construct an algorithm that can predict if a bipolar disorder patient would respond to lithium treatment.
Developing Methods for Improving Brain Organoids
The development of brain organoids of the past few years allows us to study brain disorders in these excellent models for the human brain. Brain organoids self-assemble and create brain-like structures that mimic the human brain. They are made standardly by embedding three-dimensional embryoid bodies in droplets of Matrigel that provide a scaffold for more complex tissue growth. However, as they grow in size, the diffusion of nutrients to their centers reduces and they suffer from necrotic cores. We are developing organoids in synthetic gels with lower strain modulus that allow for better diffusion and therefore there is reduced cell death in the cores of the organoids, and in addition the neurons in the organoids are much more active.
Improved brain organoids. a. A brightfield image of an organoid that was developed in a synthetic gel. b. A fluorescence image of neurons inside an organoid. c. Electrical activity of neurons inside a synthetic gel. The neurons fire repetitively and synchronically creating periodic network bursts. d. In Matrigel imbedded organoids the activity is more sparse and less synchronized.