Faculty Bibliography

The acceptance of students to a medical school places a considerable emphasis on performance in standardized tests and undergraduate grade point average (uGPA). Traditionally, applicants may be judged as a homogeneous population according to simple quantitative thresholds that implicitly assume a linear relationship between scores and academic success. This 'one-size-fits-all' approach ignores the notion that individuals may show distinct patterns of achievement and follow diverse paths to success. In this study, we examined a dataset composed of 53 variables extracted from the admissions application records of 1,088 students matriculating to NYU School of Medicine between the years 2006-2014. We defined training and test groups and applied K-means clustering to search for distinct groups of applicants. Building an optimized logistic regression model, we then tested the predictive value of this clustering for estimating the success of applicants in medical school, aggregating eight performance measures during the subsequent medical school training as a success factor. We found evidence for four distinct clusters of students-we termed 'signatures'-which differ most substantially according to the absolute level of the applicant's uGPA and its trajectory over the course of undergraduate education. The 'risers' signature showed a relatively higher uGPA and also steeper trajectory; the other signatures showed each remaining combination of these two main factors: 'improvers' relatively lower uGPA, steeper trajectory; 'solids' higher uGPA, flatter trajectory; 'statics' both lower uGPA and flatter trajectory. Examining the success index across signatures, we found that the risers and the statics have significantly higher and lower likelihood of quantifiable success in medical school, respectively. We also found that each signature has a unique set of features that correlate with its success in medical school. The big data approach presented here can more sensitively uncover success potential since it takes into account the inherent heterogeneity within the student population.

Melanoma brain metastasis is the largest cause of melanoma morbidity and mortality, and melanoma has the highest rate of brain metastasis of any cancer. The mechanisms that mediate melanoma brain metastasis remain poorly understood. We characterized patient-derived Short-Term Cultures (STCs) as a novel model system for the study of melanoma brain metastasis. Unbiased proteomics analysis of STCs revealed striking alterations in brain metastasis vs non-brain metastasis derived STCs in proteins related to neurodegeneration. Through in-vivo assays, we show that loss of Amyloid Precursor Protein (APP) in melanoma cells dramatically inhibits melanoma brain metastasis formation without affecting metastasis to other organs and that amyloid beta is the form of APP critically required for melanoma brain metastasis. Additionally, we demonstrate that APP is required for late growth and survival of melanoma cells in the brain parenchyma. Furthermore, we demonstrate that melanoma-derived amyloid beta polarizes astrocytes to an anti-inflammatory secretory phenotype that inhibits microglial phagocytosis of melanoma cells. Finally, we show that treatment of mice with a beta secretase inhibitor (LY2886721), which prevents amyloid beta production, decreases brain metastatic burden. Our results demonstrate a critical role for amyloid beta in melanoma brain metastasis, establish a novel connection between brain metastasis and neurodegenerative pathologies, and show that amyloid beta is a promising therapeutic target for brain metastasis treatment. Studies to further characterize how amyloid beta acts in the melanoma brain metastasis microenvironment are currently underway

Background: Albuminuria is a risk factor for kidney disease progression and CV disease. We examined the relative and absolute effects of CANA by baseline albuminuria among CREDENCE participants.
Method(s): CREDENCE was a double-blind, randomized study of 4401 participants with eGFR 30-<90mL/min/1.73m2 and uACR >300-5000mg/g who demonstrated that CANA significantly reduced renal and CV outcomes, including the primary composite of end-stage kidney disease, doubling serum creatinine, or renal or CV death. We analyzed the effect of CANA on renal, CV, and safety outcomes by baseline uACR.
Result(s): At baseline, 2348 (53.4%), 1547 (35.2%), and 506 (11.5%) participants had uACR <=1000, >1000-<3000, >=3000mg/g. Higher uACR was associated with higher event rates (Figure). CANA reduced renal and CV endpoints, with no statistical variation by uACR (all p heterogeneity >0.17). CANA led to a greater absolute reduction in renal events in those with higher uACR (number needed to treat to prevent 1 episode of the primary composite: 22 and 8 for uACR >1000-<3000 and >=3000mg/g). Rates of renalrelated adverse events were lower with CANA, and the relative reduction was greater with higher uACR (p heterogeneity=0.003). CANA had no significant effect on acute kidney injury, volume depletion, hyperkalemia, urinary tract infections or hypoglycemia, with no differences by uACR (all p heterogeneity >0.12).
Conclusion(s): CANA safely reduces renal and CV events in people with type 2 diabetes and substantial albuminuria, with the greatest absolute renal benefit in those with uACR of 3000-5000mg/g

The affective state is the combination of emotion and mood, with mood reflecting a running average of sequential emotional events together with an underlying internal affective state. There is now extensive evidence that odors can overtly or subliminally modulate mood and emotion. Relying primarily on neurobiological literature, here we review what is known about how odors can affect emotions/moods and how emotions/moods may affect odor perception. We take the approach that form can provide insight into function by reviewing major brain regions and neural circuits underlying emotion and mood, and then reviewing the olfactory pathway in the context of that emotion/mood network. We highlight the extensive neuroanatomical opportunities for odor-emotion/mood convergence, as well as functional data demonstrating reciprocal interactions between these processes. Finally, we explore how the odor- emotion/mood interplay is, or could be, used in medical and/or commercial applications.

Background: The COVID-19 pandemic was associated with a greater incidence of AKI than expected. At the NY Harbor VA we faced an overwhelming number of AKI patients who were critically ill with multi-organ failure. We needed to invoke new mechanisms of providing kidney replacement therapy (KRT).
Method(s): We obtained 3 Tablo systems in late March, 2019. The machines have selfcontained reverse osmosis capabilities and so do not require other equipment to operate. They can make dialysate from concentrate and tap water and so do not require special plumbing adaptation. Their self-contained step-by-step procedures are relatively simple to follow and allow rapid training of previously unskilled personnel. Tablo generates 300 ml dialysate per minute, and blood flow was increased to up to 400 ml/min as tolerated.
Result(s): Training was completed by 2 nephrologists and 2 RNs without previous dialysis experience. We used the Tablo Hemodialysis System to provide KRT to critically ill patients. In the first week we demonstrated that water cultures and endotoxin testing were negative, and that AAMI water tests were acceptable. We used the machines to provide KRT for ICU patients with double-lumen dialysis catheters. In addition we used the machines on hospital wards where KRT had not been provided before because of a lack of the plumbing needs of conventional HD machines. We provided multiple treatments 3-6 times per week for 15 AKI patients, mean age 65 years. The mean of the best urea reduction ratio achieved in the first 1-4 treatments, if available, was 41% (often limited by hypotension and fulfillment of ultrafiltration, UF, needs). Most treatments were successful and were slowed for hypotension or tachycardia. Some were aborted because of water pressure alarms if sediment filters needed replacement, or lines clotted due to hypercoagulability associated with COVID-19. Personnel availability dictated that most treatments were 3-4 hours (and up to 8h), and generally achieved UF goals. Later HD nurses cannulated arteriovenous fistulas in ESKD patients and left treatment to non-HD nurses to complete.
Conclusion(s): By incorporating a user-friendly platform and an accelerated training program including nephrologists and RNs without previous dialysis experience, we were able to nearly double our capacity to deliver KRT during the surge

Normative models of neural computation offer simplified yet lucid mathematical descriptions of murky biological phenomena. Previously, online Principal Component Analysis (PCA) was used to model a network of single-compartment neurons accounting for weighted summation of upstream neural activity in the soma and Hebbian/anti-Hebbian synaptic learning rules. However, synaptic plasticity in biological neurons often depends on the integration of synaptic currents over a dendritic compartment rather than total current in the soma. Motivated by this observation, we model a pyramidal neuronal network using online Canonical Correlation Analysis (CCA). Given two related datasets represented by distal and proximal dendritic inputs, CCA projects them onto the subspace which maximizes the correlation between their projections. First, adopting a normative approach and starting from a single-channel CCA objective function, we derive an online gradient-based optimization algorithm whose steps can be interpreted as the operation of a pyramidal neuron. To model networks of pyramidal neurons, we introduce a novel multi-channel CCA objective function, and derive from it an online gradient-based optimization algorithm whose steps can be interpreted as the operation of a pyramidal neuron network including its architecture, dynamics, and synaptic learning rules. Next, we model a neuron with more than two dendritic compartments by deriving its operation from a known objective function for multi-view CCA. Finally, we confirm the functionality of our networks via numerical simulations. Overall, our work presents a simplified but informative abstraction of learning in a pyramidal neuron network, and demonstrates how such networks can integrate multiple sources of inputs.

Seizure incidence is increased in Alzheimer's disease (AD) patients and mouse models, and treatment with the antiseizure drug levetiracetam improves cognition. We reported that one mechanism by which seizures can exert persistent effects on cognition is through accumulation of ΔFosB, a transcription factor with a long half-life. Even the infrequent seizures that spontaneously occur in transgenic mice expressing human amyloid precursor protein (APP) lead to persistent increases in ΔFosB in the hippocampus, similar to what we observed in patients with AD or temporal lobe epilepsy. ΔFosB epigenetically regulates expression of target genes, however, whether ΔFosB targets the same genes when induced by seizures in different neurological conditions is not clear. We performed ChIP-sequencing to assess the repertoire of ΔFosB target genes in APP mice and in pilocarpine-treated wildtype mice (Pilo mice), a pharmacological model of epilepsy. These mouse models allowed us to compare AD, in which seizures occur in the context of high levels of amyloid beta, and epilepsy, in which recurrent seizures occur without AD-specific pathophysiology. Network profiling of genes bound by ΔFosB in APP mice, Pilo mice, and respective control mice revealed that functional domains modulated by ΔFosB in the hippocampus are expanded and diversified in APP and Pilo mice (vs. respective controls). Domains of interest in both disease contexts involved neuronal excitability and neurotransmission, neurogenesis, chromatin remodeling, and cellular stress and neuroinflammation. To assess the gene targets bound by ΔFosB regardless of seizure etiology, we focused on 442 genes with significant ΔFosB binding in both APP and Pilo mice (vs. respective controls). Functional analyses identified pathways that regulate membrane potential, glutamatergic signaling, calcium homeostasis, complement activation, neuron-glia population maintenance, and chromatin dynamics. RNA-sequencing and qPCR measurements in independent mice detected altered expression of several ΔFosB targets shared in APP and Pilo mice. Our findings indicate that seizure-induced ΔFosB can bind genes in patterns that depend on seizure etiology, but can bind other genes regardless of seizure etiology. Understanding the factors that underlie these differences, such as chromatin accessibility and/or abundance of co-factors, could reveal novel insights into the control of gene expression in disorders with recurrent seizures.