Current Research and Scholarly Interests


The epigenetic landscape of a cell is largely determined by the organization of its chromatin and the pattern of DNA and histone modifications. These confer differential accessibility to areas of the genome and through direct and indirect regulation of all DNA related processes form the basis of the cellular phenotype. Global information about the epigenetic landscape, for example using ATAC- or histone ChIP- seq methods, have dramatically advanced our understanding of normal and neoplastic cellular processes. However, the current methods average the measurements over thousands of cells, provide snap shots instead of continuous output, and are poorly adapted for high content drug screening. We developed novel phenotypic screening platform, which interrogates the epigenetic landscape at single cell level using imaged-based machine learning. MIEL takes advantage of epigenetic marks such as histone methylation and acetylation, which are always present in eukaryotic nuclei and analyzes the patterns of epigenetic marks at the single-cell level using conventional image analysis methods. To focus the learning algorithm on the intrinsic pattern of epigenetic marks, we discarded the intensity and nuclear morphology features and used only texture-associated features (e.g. Haralick's texture features, threshold adjacency statistics, and radial features) for multivariate analysis. We have demonstrated MIEL’s utility for several applications (Fig. 1) including efficient detection and classification of epigenetically active molecules Farhy et al., 2018 (17) and identification of environmental chemical that are not cytotoxic, yet induce epigenetic changes, which we termed “epitoxic” chemicals. We have documented MIEL’s ability to accurately discriminate and classify multiple cell types and to distinguish young vs old cells from the same tissues potentially providing new tool for evaluating biological vs chronological aging. We have used MIEL to compute epigenetic signature of differentiated brain tumor cells, which is distinct from the tumor stem cell signature and enables identification of small molecule inducers of tumor stem cell differentiation Farhy et al., 2019 (18). Because MIEL is developed in a high throughput screening format (amenable to 384 and 1536 well plates), it represents an ideal screening platform to identify and characterize small molecule drug candidates. More recently, we developed live epigenetic activity probes (LEAPs) for dynamic analysis of epigenetic modifications in live cells. Combined with MIEL-based analysis of epigenetic signatures, such probes enable efficient tracking of cell fate in the dish in real time and thus well positioned to investigate the epigenetic changes and outcomes during consecutive cell divisions. In particular, MIEL-based analysis using LEAP reporters will enable us to address fundamental biological questions such as acquisition of asymmetry during mitotic division and the trajectories of cell fate change during the interphase.


Our long-term interest in this area is rooted in the discovery of common genes products and pathways operating in both hematopoietic and neural stem cells Terskikh et al., 2001 (1). Over the past decade, we have been increasingly focused on the mechanism of self-renewal and differentiation in neural stem/precursor cells, in particular using human pluripotent stem cell model. To facilitate such studies, we have established monolayer cultures of human Embryonic Stem Cells (ESCs) Bajpai et al., 2008 (2) and developed reproducible and robust differentiation of human ESCs into neuroepithelial cells Bajpai et al., 2009 (3), in particular, cranial neural crest stem cells Curchoe et al., 2010 (4) and characterized metabolic factors and pathways critical for the survival of such neuroepithelial cells Cimadamore et al., 2009 (5). Subsequently, we have established a physiological model of deficient hippocampal neurogenesis using postnatal ablation of primary cilia in granule neuron precursors Amador-Arjona et al., 2011 (6).These studies paved the road for discovering a key requirement of SOX2 function for the activation of proneural gene such as NeuroD during the onset of peripheral neurogenesis Cimadamore et al., 2011 (7). We have further demonstrated that an RNA-binding protein LIN28 is able to rescue SOX2 effect on neural precursor proliferation, but not its requirement for neurogenesis Cimadamore, 2013 et al., (8). Subsequent in vivo analysis of SOX2 role during adult hippocampal neurogenesis model revealed a novel epigenetic function of SOX2 in maintaining the bivalent chromatin of poised proneural genes Amador-Arjona et al., 2015 (9). Based on these studies, we propose that SOX2 plays critical role in orchestrating a robust activation of proneural genes at the onset of neuronal differentiation and such SOX2 function is required to develop fully functional neurons. Current efforts in the lab is focused on 1) understanding the role of SOX2 in a large population of astrocyte scattered throughout the brain parenchyma and their response to injury; 2) the function of SOX2 in modulating RNA processing.


The recent overwhelming experience shows that ZIKV and the associated neurological complications represent a long-term world-wide challenge to public health. Although the number of ZIKV cases in the Western Hemisphere have dropped since 2016, the need for basic research and anti-ZIKV drug development remains strong. Vertical transmission of ZIKV is linked to a disruption in the development of the brain and the eye, leading to severe neurologic and other dysfunctions called the Congenital Zika Syndrome. Neurologic dysfunction includes abnormal movements and hyperexcitability, sensory deficits, and severe cognitive deficits. Microcephaly can be present at birth, can develop postnatally or be absent. Understanding and preventing vertical transmission of ZIKV are therefore urgent medical needs in the face of continuing global epidemics Gorshkov et al., 2018 (10).Because wild-type mice (e.g. C57Bl/6) are resistant to ZIKV, interferon (INF)-deficient mice are widely used in ZIKV research. However, such mice succumb to ZIKV in 6-7 days post-infection limiting their utility to study vertical transmission. To overcome this problem, we established a chronic model of ZIKV infection in SJL mice Shiryaev et., 2017, Mesci et al., 2018 (11, 12). Similar to humans, SJL mice sustain ZIKV for weeks or even months. We have demonstrated vertical transmission from both infected males and females in SJL mice which, to the best of our knowledge, is the first model of sexual transmission of ZIKV. Pregnant SJL mice infected with ZIKV at E12.5 deliver newborns with cortical malformations in the surviving animals including reduced layer thickness, a sign associated with microcephaly in humans. Perinatal microcephaly caused by ZIKV is just the tip of the iceberg. Some babies born with a normal head circumference develop microcephaly as they grow older. Maternal ZIKV infection in macaque results in multiple subtle alterations in fetal brain suggesting that all children exposed to ZIKV in utero could be affected. We have observed that ZIKV infection in adult mice nearly ablate proliferating neural progenitors in the germinal zones Li et al., 2016 (13). In addition to developing preventive ZIKV vaccines, there is a need for small molecules that treat ZIKV. Our team documented that Chloroquine (CQ) Shiryaev et al., 2017a (11), Sofosbuvir (SOF) Mesci et al., 2018 (12), Emetine, Yang et al., 2018 (14), and inhibitors of ZIKV protease Shiryaev et al., 2017b (15) are potent suppressors of ZIKV infection in vivo as monotherapies. However, combination treatment is significantly more efficacious for many viral infections (for example HIV and hepatitis C). Our current approach takes advantage of the sustained ZIKV infection in SJL mice, thus enabling us to determine the structural and behavioral outcomes of the physiological congenital ZIKV infection for the fetal and adult brain. We propose to develop drug cocktails to efficiently prevent vertical transmission of ZIKV


Hair loss is a relevant social and health issue associated with diminished self-esteem and emotional distress (especially in youth), often leading to reduced quality of life, secondary morbidity, and depressive episodes. Hair loss in women is particularly mentally devastating and often psychologically debilitating. In female patients, alopecia is often the most traumatic immediate consequence of chemotherapy. Androgenetic alopecia (inherited pattern) is the most common form of hair loss, affecting nearly 35 million men and 21 million women in the United States alone ( Androgenic alopecia, alopecia areata, and scarring alopecia are the most common progressive and incurable types of clinical hair loss with no treatments available to stop or reverse the loss. Finasteride and minoxidil the only FDA-approved drugs that claim to slow down hair loss, are problematic due to cost, side effects, and inadequate performance. Conventional hair transplantation in man moves limited numbers (10%) of pre-existing hair follicles from the back of the scalp to the front/top and the effect is temporary due to the progressive nature of genetic hair loss. Despite these glaring inadequacies, ~133,000 and ~80,000 hair restoration procedures are performed annually in the US and Europe, respectively and represent a significant total current market (~$3.5 billion in the US and ~$2 billion in Europe). In most cases, hair loss is caused by progressive loss of DPC number and function. The approach pioneered by the Terskikh laboratory Gnedeva et al., 2015 (16), enables differentiation of human iPSCs into folliculogenic DPCs (iPSC-DPCs). More recently, we have developed a novel biodegradable microscaffold, termed “lolliup,” designed to directionally control cell transplantation and hair growth. We are currently focused on generating fully human hair follicles in immunodeficient (Nude) mice. The transplanted construct, designed as a unit of hair transplantation in humans, will comprise patient-specific iPSC-DPCs and iPSC-EpSCs combined inside lolliup.