Commentary on Some Recent Theses Relevant to Combating Aging: February 2019

Characterization and Modeling of Metabolic Stress Responses in Cellular Aging

David Alfego, PhD, Drexel University

Cellular aging describes the buildup of changes over time that affect normal mechanisms of cells, tissues and organisms throughout their lifespan, which can lead to any number of potential health risks, diseases or other disorders. One of the major causes of these changes is declining mitochondrial function, though the cause of this energy stress is still debated. The prevailing experimental model for aging studies examines cells in a senescent state as the hallmark of aging. Yet this permanent, post-mitotic phase is more commonly observed in vitro. Aged cells in vivo often retain their mitotic potential, indicative of a paused, quiescent state. This thesis proposes a new platform to study aging through perturbations of mitochondrial function via an experimental energy restriction in quiescence (ERiQ) model that may be more relevant to aging in tissues. This model causes adaptive changes in major stress response pathways for AKT, NF-κB, p53 and mTOR as a reaction to reduced ATP, NAD+ and NADP levels. The construction of a theoretical computational model, complementary to the experimental model, is based on feedback motifs that investigate the interplay between those key stress response pathways. The in silico model demonstrates adaptations to sudden energetic perturbations, promoting pro-survival phenotypes and recovery. This thesis hypothesizes that the very same survival mechanisms are chronically activated during aging, but also cause conflicting responses that actively suppress mitochondrial function to contribute to a lockstep progression of decline. The model makes predictions consistent with inhibitory and gain-of-function experiments in aging.

The relevance of ERiQ as a model to study aging is further emphasized by a transcription factor (TF) meta-analysis of gene expression datasets accrued from 18 tissues from individuals at different biological ages, which were compared to 7 different experimental platforms. Experimental datasets included replicative senescence and ERiQ, in which ATP was transiently reduced. TF motifs in promoter regions of trimmed sets of target genes were scanned using JASPAR and TRANSFAC motifs and TF signatures established a global mapping of agglomerating motifs with distinct clusters when ranked hierarchically. Remarkably, the majority of in vivo aged tissues correlated with the ERiQ profile instead of senescence, confirming its relevance as a new experimental model. Fitting motifs in a minimalistic protein-protein interaction (PPI) network model allowed us to probe for connectivity to distinct stress sensors, as well as identify novel targets of study in transcription factors that significantly switch enrichment between ERiQ and senescence. In the PPI, DNA damage sensors ATM and ATR linked to one subnetwork associated with senescence. By contrast, energy sensors PTEN and AMPK connected to the nodes in the ERiQ subnetwork. These data suggest that energy deprivation may be linked to transcriptional patterns characteristic of many aged tissues distinct from cumulative DNA damage associated with senescence. Finally, this thesis exemplifies the combined use of the predictive power of the computational model with experimental investigation in vitro. Preliminary experiments show how the model can be refined to reflect how certain conditions may alter metabolic output and offer intriguing insights into the future of cellular aging studies.

Comment: The SENS platform is designed to repair or obviate all known forms of age-related damage—defined as permanent alterations that occur largely at random over the life span, and which ultimately accumulate to levels that compromise cellular and organ-level functions, leading to the pathologies of old age. Crucially, the great majority of the dysfunction observed in aged tissue is not of this permanent class; rather, it is the result of ongoing responses to the presence of hard damage—responses which in general are adaptive and protective in an acute context, but which become pathological when an inability to remove the noxious stimulus leads to their chronic activation. It is hoped that at least in some cases, such maladaptive cellular states will return automatically to homeostasis once the underlying damage is effectively resolved. This thesis lends support to that idea, showing that much of the pattern of bioenergetic failure observed in aged tissue is of an essentially reversible nature (ERiQ, unlike senescence, reverts to a healthy phenotype when normal energy levels are restored)—consistent with the seminal finding in 2013 ¹¹ that a decline in NAD+ underlies a large part of age-related mitochondrial dysfunction, ¹² but that this phenotype can be effectively treated simply by boosting NAD+ levels. Therapies based on that premise ¹³ are being rapidly advanced toward the clinic and seem likely to be a boon to a large fraction of the populace, yet their benefit is likely to be rather limited in the long term since the hard damage responsible for declining NAD+ is not addressed (accordingly, preclinical data suggest that such therapies can greatly bolster health span, but not maximal life span; to some extent, the same benefits can be achieved through traditional means ¹⁴ ). What, then, is the real nature of that damage? Recent work has pointed to the enzyme CD38 as the dominant NAD+ recycler in mammalian tissues, with an age-related increase in expression and activity that parallels the observed decline in NAD+ levels (and specific CD38 inhibitors are already being assessed for their ability to block that decline ¹⁵ ). That increased expression, though, is driven by inflammatory signaling—the product (in the context of aging) of senescent cell accumulation, protein adducts such as AGEs, ¹⁶ and so forth. The development of therapies that comprehensively resolve those forms of damage will ultimately prevent their deleterious downstream effects, without the inevitable side effects of intervening in dynamic metabolic processes.

Engineering Immunity: Enhancing T Cell Vaccines and Combination Immunotherapies for the Treatment of Cancer

Kelly Moynihan, PhD, Massachusetts Institute of Technology

Checkpoint blockade with antibodies against CTLA-4 or PD-1 has demonstrated that an endogenous adaptive immune response can be stimulated to elicit durable tumor regressions in metastatic cancer, but these dramatic responses are confined to a minority of patients. This outcome is likely due in part to the complex network of immunosuppressive pathways present in advanced tumors, which necessitates the development of novel therapeutics and combination immunotherapies to generate a counter-directed network of pro-immunity signals. In Chapters 2 and 3 of this thesis, we describe methods for enhancing T cell priming against tumor antigens via covalent modification of molecular vaccines to enhance lymphatic drainage, serum stability, or cytosolic access to improve presentation on MHC class I. In Chapter 4, we demonstrate a combination immunotherapy that recruits a diverse set of innate and adaptive effector cells, enabling robust elimination of large tumor burdens that to my knowledge have not previously been curable by treatments relying on endogenous immunity. Maximal anti-tumor efficacy required four components: a tumor antigen targeting antibody, an extended half-life interleukin-2 (IL-2), antiPD-1, and a powerful T cell vaccine. This combination elicited durable cures in a majority of animals, formed immunological memory in multiple transplanted tumor models, and induced sustained tumor regression in an autochthonous BRrafvs00E/Pten melanoma model. Finally, in Chapter 5, we show preliminary data on combination immunotherapies used to treat antigenically heterogeneous tumors. Taken together, these data define design criteria for enhancing the immunogenicity of molecular vaccines and elucidate essential characteristics of combination immunotherapies capable of curing a majority of tumors in experimental settings typically viewed as intractable.

Comment: Since the theory of cancer immunosurveillance was proposed in the 1950s, our understanding of the relationship between tumors and the immune system has improved at an ever-accelerating rate—especially so in the last two decades. That most recent surge of progress has focused rather disproportionately on the adaptive arm of the immune system, prompted by early successes and a very reasonable desire to establish durable long-term control of tumors despite a changing antigenic profile. The element of this work that introduces methods to enhance the immunogenicity of peptide vaccines (to a quite remarkable degree, with serum lifetimes of up to a week compared with hours for unmodified peptides ¹⁷ ) is of course aligned with that focus. Unfortunately, in patients with established cancer, the tumor-specific repertoire of lymphocytes is almost, by definition, poorly expanded and actively immunosuppressed, while the tumor itself has adapted physiologically to resist effective surveillance. It is perhaps not so surprising that many adaptive–stimulatory therapeutics with a glowing preclinical record falter in such an unfavorable biological context! Synergistic activation of an innate immune response, however, naturally complements such therapies by immediately and potently altering the tumor microenvironment—even to the point of normalizing associated vasculature—to support adaptive cell function and recruitment, alongside the expression of a baseline cytotoxic profile (which can be dramatically enhanced by appropriate antibody engagement). The development of combination therapies faces the ongoing challenge of potential parallel increases in toxicity alongside enhancement of antitumor efficacy, but we sincerely hope that the increasingly compelling preclinical data, such as those presented in this excellent dissertation, will motivate urgent exploration and translation of this biologically rational approach. That hope seems well founded, with several combinations thought to have synergistic innate/adaptive stimulatory activity already in clinical investigation or positioned for testing in the near future and with rapid progress in the development of trial designs suited to identify safe and effective combinations. The very recent confirmation that temporally staggering elements of a combination therapy can substantially improve its overall efficacy, ¹⁸ while biologically unsurprising (it is wise to provide immature dendritic cells with time to process antigenic debris before stimulating them to mature, for example), possibly represents a further boon to translation—as combinatorial toxicity is most likely to manifest when agents are coadministered within a short time frame.

Extracellular Inflammatory Signaling from Dysfunctional Telomeres

Zhuo Wang, PhD, University of the Sciences in Philadelphia

Telomere dysfunction describes the catastrophic damage at telomeres, which often leads to genomic instability at the cellular level. There is rising evidence showing that telomere dysfunction also influences the extracellular environment with the inflammatory response. However, little is known about the molecular mechanism of this dysfunctional telomere-associated inflammation. In this dissertation, we identified extracellular forms of Telomeric repeat-containing RNA (TERRA), and demonstrated it might play a role in mediating the crosstalk of telomere dysfunction and inflammation. We found this cell-free TERRA (cfTERRA) is present in mouse tumor and embryonic brain tissue, as well as in human tissue culture cell lines using RNA in situ hybridization. RNA-seq analyses revealed TERRA to be among the most highly represented transcripts in extracellular fractions derived from both normal and cancer patient blood plasma. By characterizing extracellular fractions of the human lymphoblastoid cell line culture media, cfTERRA is shown as a shorter form (∼200 nt) of cellular TERRA and co-purifies with CD63- and CD81-positive exosome vesicles that could be visualized by cryo-electron microscopy. Mass spectrometry and extracellular chromatin immunoprecipitation (ChIP) assays revealed that regular cfTERRA was physically interacting with histones and telomeric DNA. Incubation of cfTERRA-containing exosomes with peripheral blood mononuclear cells (PBMCs) stimulated transcription of several inflammatory cytokine genes, including TNFα, IL-6, and C-X-C chemokine 10 (CXCL10). Exosomes engineered with elevated TERRA or liposomes with synthetic TERRA further stimulated inflammatory cytokines, suggesting that exosome-associated TERRA augments innate immune signaling. The levels of cfTERRA and DNA damage marker γH2AX were increasingly incorporated into the exosomes during telomere dysfunction. These dysfunctional telomere-derived exosomes activated a more robust transcription of inflammatory cytokines in PBMCs. These findings imply a previously unknown extrinsic function of TERRA and a potentially molecular mechanism of communication between telomeres and innate immune signaling in tissue and tumor microenvironments.

Comment: One of the earliest theories of cellular aging, published by Olovnikov in 1971, posited that the incremental loss of telomeric repeats from chromosomal DNA during replication acted as an inherent limit on proliferation (in the absence of cancerous changes), driving cells to enter a senescent state when the remaining telomere length fell below a certain threshold. Although broadly true, it has now become clear that the relationship is more complex. Telomeres act as a sensor for genotoxic damage in general—being more susceptible to most forms thereof than primary chromosomal DNA—and sufficiently abnormal telomeres can induce senescence even when of normal length (as can epigenetic changes that appear to occur as a side effect of telomerase activity per se ¹⁹ ). TERRA is a family of long noncoding RNA (of 100 bp–9 kbp length in human cells) transcribed from the telomeric sequence, members of which play crucial roles in the routine maintenance and replication of telomeres, but which are also implicated in telomerase-independent telomere extension in various cancer lines, making attempts at therapeutic intervention within the cell liable to severe side effects. The discovery in this work that secreted forms of TERRA are responsible for at least part of the paracrine toxicity caused by cells with telomeric dysfunction, however, indicates a potential avenue for treatment (pending the ability to ablate and regenerate all such cells, which requires addressing a much larger number of specific cases) analogous to that previously proposed for the similarly proinflammatory, cell-free mitochondrial DNA. ²⁰ The distinctive and highly repetitive TERRA sequence is a prime target for silencing or degradation by a specific nuclease; the challenge will be to accomplish this in the extracellular milieu without introducing the same activity into the cell. Methods for introducing therapeutics into existing exosomes—exogenous loading—are under active development, but largely depend on isolation and return-based methods and are not yet suited for application in vivo (the exosomal membrane is nearly identical in composition to the plasma membrane, making it very difficult to independently target). Nonetheless, recent marked success in exogenously loading siRNA through conjugation to cholesterol ²¹ may point toward a route for delivery at least of nucleic acid agents (interfering RNA or even a ribozymal nuclease) if those agents can be engineered to degrade swiftly when subsequently taken up by cells. Endogenous loading—modifying cells to produce exosomes with specific added contents, an approach with generally much higher efficiency—may also be possible if a further technique can be devised to activate the therapeutic only during/after secretion.

High-Throughput Microfluidic Labyrinth for the Label-Free Isolation of Circulating Tumor Cells for Single-Cell Gene Expression Profiling

Eric Lin, PhD, University of Michigan

Circulating tumor cells (CTCs) present in the blood are the seeds of metastasis and are of high biological and clinical relevance. Single-cell technologies are playing an increasing role in profiling CTCs in the peripheral blood for detection and real time monitoring of cancer metastasis. CTCs also help in identifying distinct drivers of metastasis. However, current approaches are limited to subjective selection of CTCs based on biomarkers, which hinders the unbiased comprehensive study of CTCs on a single cell level. We present a unique label-free microfluidic “Labyrinth” device to isolate CTCs at a high throughput of 2.5 mL of blood per minute, offering the first biomarker independent single cell isolation and genomic characterization platform to study heterogeneous CTC subpopulations in cancer patients. The Labyrinth takes advantage of inertial forces on the microscale in curved geometries to differentially focus cells based on the size difference between CTCs (15–25 μm) and blood cells (2–12 μm). This novel strategy of multi-course path traversing across inner loops to outer loops yielding highest hydrodynamic path length enabling focusing of both CTCs and white blood cells (WBCs) differentially, leading to high recoveries (>90%) and efficient separation of WBCs from CTCs, resulting in high purity of CTCs in the final product (∼600 contaminating WBCs per mL remained), even in whole blood samples without any pre-processing. CTCs were successfully isolated from 56 breast cancer (9.1 CTCs/mL average, range 2–31/mL) and 20 pancreatic cancer (51.6 CTCs/mL average, range 11–115/mL) patients. We detected not only CTCs typically defined by epithelial markers, but also significant numbers of CTCs (>50%) lacking epithelial markers but expressing mesenchymal and cancer stem cell markers. Patient samples were then analyzed using single cell multiplex gene expression. Seventy single cells were successfully recovered and used to identify different subpopulations of CTCs based on their genetic signature, unlike other methods where a positive or negative selection based on protein expression is used. Interestingly, both inter- and intra-patient molecular heterogeneity at the single cell level in CTCs were observed with cells expressing genes uniquely related to epithelial, mesenchymal-epithelial transition, and epithelial-mesenchymal transition phenotypes. The Labyrinth platform allows a thorough molecular understanding of the heterogeneity among CTCs. This platform also shows CTCs potential as a biomarker to non-invasively evaluate tumor progression and response to treatment in cancer patients.

As a truly biomarker free isolation platform, Labyrinth is also adopted for the isolation of CTCs from various types of cancers, including but not limited to adenoid cystic carcinoma, hepatocellular carcinoma, and lung cancer. Besides enumeration, the isolated CTCs were used in a wide range of studies, such as ex vivo culture of live CTCs from pancreatic patients and generating xenograft tumor model in mice. Beyond CTC isolation, Labyrinth was applied in the study of skeletal muscle satellite cells in collaboration with Dr. Brian C. Syverud and Prof. Lisa M. Larkin.

All in all, Labyrinth device offers a microfluidic technology to address the need for efficient isolation of rare cells and enables downstream studies on the target cells.

Comment: The capacity for cancerous cells to spread from their tissue of origin to seed new tumors at remote locations in the body—metastasis—is already the main contributor to disease-associated mortality and will only become more significant as techniques for ablation of solid tumors continue to advance. Despite that significance, our understanding of metastasis, let alone our ability to intervene to prevent it, has thus far been severely limited both by the rarity of such cells and by their highly heterogeneous phenotypes. By efficiently isolating CTCs from whole blood using a fully label-free method, this dissertation introduces a genuinely groundbreaking tool for the study of the metastatic process. Yet, we believe “Labyrinth” has the potential to be more than a diagnostic tool—at least two therapeutic applications immediately spring to mind! In the first, isolated CTCs are loaded with a high-contrast imaging probe and returned to the circulation. This approach would enable the labeling of metastatic target sites throughout the body, circumventing our inability (except in certain well-defined models of minimal clinical relevance) to predict the “fertile soil” sought by those cells, and by employing a theranostic probe, or by coloading a suitable therapeutic agent, the returned cells could serve as molecular double agents. A second—perhaps more ambitious, but tractable given adequate motivation for its development—application would involve a very substantial scale-up of the device to the point of handling the 4–8 L/min flow of blood through the human circulatory system. At that scale, labyrinth becomes, in effect, an apheresis system specific for metastatic cells; such an apparatus could prove extremely valuable as a companion to chemotherapy, which often promotes metastatic spread as a side effect. ²² Granted, metastatic spread can occur through the lymph system as well as the blood, but as the lymphatic fluid drains into the primary circulation, this may still be an acceptable approximation. Similarly, a method would be required to return contaminating WBCs to the circulation, but that element of the system need not be label free. Thankfully, the cardiac output of a murine heart is only ∼20 mL/min, and thus the scale-up required to assess the viability of this concept in mice is substantially more tractable.

Oxygen Nanobubbles for Ultrasound-Guided Targeting of Cancer Hypoxia

Pushpak Bhandari, PhD, Purdue University

The effective targeting of hypoxia remains one of the most important and unsolved problems in cancer therapy. Hypoxia has been shown to correlate with significant unfavorable prognosis in several types of cancer. Hypoxia-adaptive pathways effect cellular and epigenetic changes in the tumor. Here, oxygen nanobubbles (ONBs) were designed to target the hypoxic tumor regions. Through in vitro and in vivo studies, we established the potential of ONBs in enhancing ultrasound imaging contrast, perturbing the microenvironment and epigenetic status of the hypoxic cell, and acting as in vitro and ex vivo hyperspectral imaging agents.

The first part of my thesis focuses on the development of ONBs to help regulate the epigenetic state and hypoxia-adaptive pathways of the tumor cell. Our technique enables us to significantly halt tumor progression and monitor the tumor using ultrasound and fluorescence imaging in vivo. Next, ONBs were utilized in increasing the efficacy of conventional methylating chemotherapeutic, temozolomide, for enhanced bladder cancer therapy. In the third chapter of this dissertation, ONBs were precisely delivered to orthotopic bladder cancer tumors in mice using Doppler ultrasound. It was found that ONB uptake in tumors can be significantly enhanced using ultrasound guiding. Further, reoxygenation because of ONBs led to enhanced efficacy of a conventional chemotherapeutic, mitomycin-C, in mice models. The fourth chapter of this dissertation focuses on the application of ONBs in dark-field microscopy to enable quantitative imaging and dynamic detection of the nanoparticles, in vitro and ex vivo. Finally, parameters are established to scale-up ONBs for evaluation in pet dogs with naturally-occurring urothelial carcinoma.

Comment: Almost all solid tumors exhibit a state of persistent hypoxia, with median O₂ levels 50%–95% lower than healthy tissue of the same type. This phenomenon is understood to result from rapid proliferation of cancerous cells coupled with an expansion of the associated vasculature, which, although extensive, is also highly disorganized and thus inefficient in performing gas exchange. While the natural response to acute hypoxia is protective, chronic hypoxia—particularly when cells cycle repeatedly between lower and higher oxygen concentrations, as inevitably occurs in growing tumors—leads to increased genomic instability and invasiveness while retaining the prosurvival signaling (and ensuing elevated resistance to radio- and photodynamic therapy) seen in the acute scenario. Low O₂ levels also significantly inhibit immune function, impairing the maturation of T cells and dendritic cells while promoting the function of regulatory T cells and tumor-associated macrophages—an accepted contributor to the difficulties observed in translating immunotherapies to the clinic—as well as reducing the effectiveness of the many chemotherapies with mechanisms of action dependent on oxygen. Efforts to reverse tumor hypoxia by increasing blood oxygenation (using hyperbaric O₂, artificial hemoglobin, liquid fluorocarbons, and other techniques) have thus far largely failed due to limited efficacy and adverse side effects. Nanobubbles—gas bubbles of <1 μm diameter contained within a minimal shell—can be remarkably stable in the body, with lifetimes of up to several hours, while being small enough to readily extravasate from the circulation. This dissertation substantially advances them toward clinical viability by demonstrating that oxygen-loaded nanobubbles can be steered with high efficiency using a focused ultrasound beam, allowing their marker-independent accumulation at a known tumor site (and subsequent controlled rupture). Unsurprisingly, that approach seems to present far fewer side effects than altering the effective O₂ concentration of the primary circulation! The near-ubiquity of the hypoxic phenotype implies that this pleasingly direct technique has great potential as an adjunct therapy for a majority of clinically relevant solid tumors; it may also find application in the treatment of chronic poorly healing wounds, another common clinical scenario in which persistent hypoxia contributes to poor prognosis.

The Eye as a Window to the Alzheimer's Disease Brain

Fred Souza, PhD, The University of Alabama at Birmingham

Alzheimer's disease (AD) is a debilitating, and the most prevalent, type of dementia that is manifested by cognitive deficits, anomalous protein metabolism, cell loss, and pathological alterations in several neurotransmitter systems, particularly the cholinergic and glutamatergic systems. Moreover, AD is associated with visual deficits that have been reported to occur even in the early stages of the disease and may precede conspicuous cognitive impairment. To date, the underlying causes of the visual deficits and whether they stem from retinal or cortical abnormalities remain poorly understood. The following studies aimed at establishing whether the pathological changes observed in the cerebrum are also present in the retina and assessing AD's influence in retina's physiological responses. We used quantitative polymerase chain reaction and immunohistochemistry to assess changes in acetylcholine receptor (AChR) gene expression, gliosis, retinal cell number in the Tg-SwDI mouse model as compared to age-matched wild-type (WT). Young adults and middle-aged adults Tg-SwDI mice exhibited initial upregulation of AChR gene expression, but downregulation in old adults. Furthermore, young adult transgenic mice displayed significant cell loss in the inner retina and photoreceptor layer. Middle-aged adult and old adult mutants exhibited increased astrocytic gliosis and cholinergic cell loss. Electroretinography (ERG) was employed to measure the amplitude and implicit time of retinal responses from TgF344-AD rat model and age-matched WT at 9 and 16 months of age. Nine-month mutants exhibited higher responses from several retinal cells, but lower responses from off bipolar cells and Müller cells. Sixteen-month TgF344-AD rats displayed lower scotopic critical flicker fusion threshold and photoreceptor responses, and slower implicit time for on bipolar cell responses, at several light intensities. These data collectively indicate that AD-related changes observed in the cerebrum are also present in the retina and may be, at least in part, responsible for the visual deficits associated with the disease. Furthermore, we demonstrated that AD pathology affects retinal cells' physiological responses and that ERG can be employed as a suitable means to detect AD-related visual changes to ultimately serve as an efficacious diagnostic tool to identify the disease in its earlier stages, thus improving treatment efficacy.

Comment: The retina is the only part of the central nervous system that is directly visible and hence presents a unique opportunity to assess CNS-specific health without the need for invasive measures. Although remote from the areas of the brain where neurodegenerative disease is thought to begin, its extracellular milieu is continuous with that of the cerebrum and thus its properties, in part, reflect changes to that environment (notably including changes involving a well-known secretory phenotype ²³ ). This thesis adds to a body of evidence that various retinal imaging techniques can indeed be used to detect prodromal AD (up to two decades before clinical diagnosis, a point at which disease-modifying intervention is expected to be much more feasible) with excellent sensitivity, although in many cases, those same tests are prone to return a false positive in the presence of other neurodegenerative and/or retinal conditions. Critically, such methods are typically cheap and easy to conduct and yield datasets of manageable size, making them excellent candidates for use in a population-level screening context, where their promiscuity is not necessarily disadvantageous (since the key goal is to highlight the existence of pathology at all, not to arrive at a precise diagnosis). It would be reasonable to expect that longitudinal records generated by such screening will allow for refinement of the tests themselves to better control for interindividual variation, a major source of remaining errors, and perhaps lead to identification of still earlier biomarkers currently masked by that variation. We would suggest that initiation of just such a screening program is now long past due!