Poster Session K (+Lunch and Office Hours)

Acquired resistance to anti-cancer therapy is an enormous challenge. One of the main factors contributing to therapy resistance is tumor hypoxia. The oxygen and nutrient stress imposed by tumor hypoxia forces cancer cells to adapt in order to survive. These metabolically adapted cancer cells are often more invasive, more malignant, and more drug resistant. As a result, the cancer cells that emerge from hypoxic tumor regions are more likely to cause patient relapse. There is therefore a critical need to understand the mechanisms that promote the survival of cancer cells in stressfull tumor microenvironments. We previously showed that the enzyme acetyl-CoA synthetase 2 (ACSS2) supports cancer cell metabolism in hypoxic and nutrient-depleted environments. ACSS2 endows cancer cells with the ability to use acetate as an alternative nutrient source to drive acetyl-CoA biosynthesis during nutrient stress, and genetic silencing of ACSS2 inhibits human breast tumor growth in xenograft models. Given the important role of acetate metabolism in breast cancer we expanded upon our studies by using immunocompetent hosts and syngeneic mouse tumor models. Our results revealed a previously unknown role of ACSS2 in modulating host anti-tumor immunity. We found that ACSS2 deficient tumors are unable to grow when host immunity is intact. Depletion of host immunity (Tcells) using genetic or pharmacological models rescues the growth of ACSS2 deficient tumors. Pharmacological inhibition of ACSS2 in tumors in vivo displayed gene signatures associated with immune infiltration and activation within the tumor microenvironment. Our current research demonstrates a novel role for acetate metabolism in supporting tumor extrinsic modulation of host anti-tumor immunity. Since activation of acetate metabolism via ACSS2 is a near universal hallmark of metabolically stressed cancer cells, targeting acetate metabolism represents an unrealized opportunity with significant upside for improving current therapeutic modalities in breast cancer.

Click Here for Poster

The neural circuits underlying human speech exhibit a striking level of convergence with other vocal learning species, which all share the ability to produce novel vocalizations acquired through social experience. The presence of a direct cortico-motoneuronal projection–from the primary motor cortex to the brainstem motor neurons controlling the vocal organs–is thought to be a critical and highly derived feature found only in vocal learners. How direct projections for vocal learning form during development or, more broadly, how these novel cortico-motoneuronal circuits for speech evolved from pre-existing motor pathways in the neocortex, are unknown. Recent work has demonstrated that cortico-motoneuronal circuits for limb dexterity, which are highly abundant in humans but depauperate in nonhuman primates, are transiently formed during early postnatal development in rodents. Our group previously discovered that laboratory strains of mice, a species capable of ultrasonic (non-learning) vocalizations, also possess a sparse vocal cortico-motoneuronal (vCM) connection, which argues that the basic constituents of vocal-learning circuits may be more broadly conserved, and open to molecular genetic investigation. Given that projection neurons in the motor cortex are known to extend exuberant collaterals during development, which are subsequently pruned via regionalized and species-specific expression of axon guidance cues, we propose that vCM projections may emerge through developmental co-option of pathways regulating postnatal axonogenesis. We provide preliminary evidence of a population of vCM neurons in the neocortex of mice that make transient connections with motor neurons innervating the laryngeal muscles. Knocking down the axon guidance gene, PlxnA1, in layer V neurons of the neocortex alters juvenile ultrasonic vocalizations and increases the number of vCM neurons in subadults. We argue that interrogation of the largely underexplored diversity of transient neural circuits could contribute significantly to our understanding of pathways critical for human speech.

Click Here for Poster

Background: Myocardial infarction (MI) causes pathological remodeling of the autonomic nervous system, which exacerbates heart failure and predisposes to ventricular arrhythmias. These changes, characterized by sympathetic activation and parasympathetic dysfunction (reduced vagal tone), act in concert to increase risk of death. The primary reasons and mechanisms for vagal withdrawal in this setting are currently unknown.
Methods: MI was induced percutaneously in pigs. Six-weeks post-MI (n=11) or age-matched normal animals (n=11) underwent functional nodose ganglia neuronal recordings using linear microarray electrodes in-vivo. After median sternotomy neuronal firing in response to changes in preload/afterload and application of epicardial mechanical or nociceptive stimuli (capsaicin, bradykinin, or vertradine) was assessed by spike sorting. In a different set of normal (n=22) and infarcted animals (n=31), nodose ganglia were rapidly excised for comparison of immunohistological changes.
Results: Functional recordings of cardiac chemosensitive neurons in normal animals showed expected increased firing rates in response to nociceptive stimuli. Paradoxically, greater numbers of nociceptive neurons and sensitivity of these neurons (defined as absolute changes in firing rates), infarcted animals demonstrated significant decreases in extracellular neuronal firing in response to nociceptive stimuli. Immunohistochemical analysis revealed increased calcitonin gene-related peptide (CGRP) expression, a surrogate for nociceptive neurons, similar to neural recordings. Both CGRP-positive and CGRP-negative neurons demonstrated increased expression of neuromodulators including glial activation and nitric oxide synthase post-MI. However, only CGRP-positive neurons showed inhibitory neurotransmission as indicated by increased expression of GABA and its enzymes.
Conclusions: Our results, for the first time, indicate significant changes in vagal cardiac afferent neurotransmission post-MII. Nociceptive neurotransmission is decreased, potentially through increased autoreceptor mediated inhibition by GABA. As nociceptive signaling through the vagal ganglia is known to increase vagal tone, decreased nociceptive neurotransmission post-MI may play an important role in occurrence of parasympathetic dysfunction.

Click Here for Poster

Vocalizations are a complex behavior that can be innate or learned. Vocal learners, animals that learn to imitate or modify vocalizations, use top-down input from the forebrain to provide fine motor control over vocal musculature. These forebrain circuits coordinate sub-cortical networks during, and after, learning to precisely execute these learned behaviors. Currently, we do not have genetically tractable mammalian models for speech and vocal learning. However, behavioral and anatomical findings from our lab and others' have shown that the ultrasonic song system of mice exhibits rudimentary features consistent with advanced features of vocal learners, including multisyllabic structure that varies with social context, and a pool of cortical neurons in posteromedial M1 that projects directly to brainstem vocal motor neurons. To test if these M1 neurons can functionally activate vocal muscles, and possibly provide fine motor coordination, we used intracortical microstimulation (ICMS) paired with electromyography (EMG) from vocal muscles in the larynx of anesthetized mice. We use latency from stimulation to EMG response to estimate the number of synapses between the cortex and the muscles, with shorter latencies meaning there are fewer neurons in the circuit (i.e. more direct connectivity). Using this approach, we found a representation of laryngeal musculature in posteromedial M1, consistent with our anatomical tracing, but also a second representation in anterolateral M1. The two regions have distinct latency and stimulation thresholds from one another. The anterolateral region, corresponding to orofacial motor cortex (OFC), has shorter latencies-to-response from stimulation and lower current injection thresholds. Conversely, the posteromedial region has longer latencies and higher thresholds. These two representations possibly underlie different functions of the laryngeal musculature in different behaviors, such as voluntary swallowing, breathing, or vocalizing. These results are the first characterization of the functional representations of laryngeal musculature in the rodent motor cortex.

Click Here for Poster

Cell surface glycans play critical roles in regulating extracellular signaling in embryonic differentiation and tissue development. Glycosaminoglycan (GAG) polysaccharides displayed on membrane-associated proteoglycans contain highly sulfated regions that serve as high affinity binding sites for a variety of growth factors and chemokines and are required for activation of cognate receptors at the cell surface. Synthetic materials that mimic the architecture and function of GAGs have recently been the subject of intensive research to establish chemical methods for controlling cellular differentiation in the laboratory. We have developed GAG-mimetic materials based on reactive polyacrylamide scaffolds generated by controlled polymerization techniques (e.g., RAFT) and functionalized with sulfated GAG disaccharide building blocks that show avidity for a broad range of growth factors. When introduced to the surfaces of stem cells, these materials were able to promote growth factor signaling and support cellular differentiation in numerous contexts, including neural, mesodermal or adipogenic differentiation. I will present the development of this concept toward reprogramming the adipogenic differentiation program to produce adipocytes with enhanced glucose clearance capacity. The change in the metabolic activity of the mature adipocytes was insulin-independent and resulted from a switch from fatty acid metabolism toward glycolysis determined early in differentiation. These findings are poised to open a new exciting avenue for the treatment of Type 2 diabetes.

Click Here for Poster

An expanding atlas of single-cell organization in tumor tissue offers insights into how malignant and stromal cells may communicate with each other to influence disease progression. Nonetheless, it remains difficult to understand functional cause-and-effect relationships between cells from static maps of tissue composition. To overcome this limitation, we have developed strategies for high-resolution in vivo confocal microscopy to directly monitor in situ signaling dynamics and perturbation in live tumor models. We present recent progress in applying this approach to study the spatially-dependent activities of multiple mitogen activated protein kinase (MAPK) pathways, using mouse models of cancer driven by constitutively activating BRAF or KRAS mutations. BRAF mutation is especially common in malignant melanoma, and targeted kinase inhibitors have been developed to inhibit BRAF and downstream ERK signaling activities. However, drug resistance frequently emerges in patients and new treatment strategies are urgently needed.
This work builds on a series of collaborative studies, in which our team found that efficient MAPK-ERK kinase inhibition and resultant cancer cell killing could elicit an immunogenic wound-healing response in tumors, leading to recruitment of innate immune cells (namely macrophages) into the tumor microenvironment. Time-lapse microscopy in mice revealed that macrophages could limit the ability of MAPK-ERK targeted kinase inhibitor to block signaling in adjacent tumor cells, in a highly localized manner. Further analysis implicated bi-directional signaling between tumor cells and macrophages through a family of immunosuppressive receptors involved in clearing dead cell debris from the body. These results helped demonstrate how a feedback of reciprocal tumor-immune signaling can locally amplify kinase inhibitor resistance in the tumor microenvironment. Furthermore, they set the stage for future work on this project to dissect spatial regulatory relationships and potential therapeutic avenues by locally manipulating immune signaling dynamics.

Click Here for Poster

While most intracellular pathogens export a limited repertoire of effector proteins to co-opt existing host-cell metabolic machineries, the Malaria-causing parasite Plasmodium falciparum exports more than 10% of its proteome into host human red blood cells, which are highly specialized for carrying hemoglobin and lack the resources to support the active growth and replication of the parasites. The hundreds of proteins in the P. falciparum exportome extensively remodel host erythrocytes, creating the infrastructure needed to import nutrients, export waste, and evade the host immune system. The complexity and breadth of its host-cell remodeling machinery make P. falciparum a rich and exciting system for the study of host-pathogen interfaces. Unfortunately, many of the molecular mechanisms underlying this parasite's ability to hijack human red blood cells remain enigmatic, as much of the P. falciparum proteome has proven recalcitrant to structural and biochemical characterization using traditional recombinant approaches. This paucity of high resolution structural and functional information is compounded by the fact that 50% of the P. falciparum proteome is novel.
To overcome these barriers to structural study of malaria parasites and address the gaps in our understanding of the molecular mechanisms underpinning host-pathogen interactions in parasite-infected red blood cells, our lab develops and implements methodologies for endogenous structure determination from P. falciparum. We combine CRISPR-Cas9 parasite gene editing, single particle cryoelectron microscopy (cryoEM), and in situ cryoelectron tomography (cryoET) to determine near-atomic resolution structures of previously intractable protein complexes enriched directly from endogenous P. falciparum parasites and directly visualize the host-pathogen interface in intact parasite-infected red blood cells at sub-nanometer resolutions.

Click Here for Poster

The algorithms used to process sensory information dictate the structure of underlying neuronal circuits. This circuit connectivity in turn often provides insight into the relevant stimulus features, such as topographic location, orientation or sound frequency. In the olfactory system, anatomical studies have found that principal neurons of the olfactory bulb (OB) innervate the piriform cortex (PC) in a broad and seemingly unstructured fashion. The apparent absence of orderly connectivity from OB to PC together with reports of distributed connections within PC inspired computational models of circuit function that rely on random connectivity. Here we have exploited the high throughput of Multiplexed Analysis of Projections by sequencing (MAPseq) and Barcoded Anatomy Resolved by sequencing (BARseq) to obtain the projections of 5,309 OB and 30,433 PC output neurons in the mouse at single-cell resolution. Analysis of this data set reveals previously unrecognized spatial structure in the connectivity of both PC inputs and outputs. We identify specific projection gradients in the OB output neurons co-innervating PC and subsets of extra-piriform OB target regions. Distinct populations of narrowly and broadly-projecting OB cells tile differentially the anterior-posterior (A-P) axis of PC. Furthermore, characteristic groups of PC output neurons, also organized along its A-P axis engage local intra-piriform connectivity and innervate specific sets of brain regions. Strikingly, input-output parallel circuit motifs spanning the A-P axis of PC emerge: a given PC output neuron appears to be wired up such that the strength of its projection to specific extra-piriform OB targets matches the strength of its dominant OB input co-projection to the same region. Our findings suggest an organizing principle of matched, direct and indirect pathways in the olfactory system and challenge the canonical model of piriform cortex as a random network.

Click Here for Poster