Commentary on Some Recent Theses Relevant to Combating Aging: December 2013

Asynchronous Bouts of Muscle Regeneration Lead to Features of Failed Regeneration in the Muscular Dystrophies

Sherry Dadgar, PhD, George Washington University

In many of the muscular dystrophies, age-related failure of regeneration is often seen with progressive replacement of the muscle with fibrofatty connective tissue. In Duchenne muscular dystrophy (DMD), lack of dystrophin is compatible with muscle function. However, failure in regeneration is associated with significant muscle wasting, weakness, and functional disability, suggesting that dystrophin deficiency is necessary, but not sufficient, for disability. The goal of this dissertation research was to develop a model that explains failed regeneration in DMD and other chronic inflammatory diseases. The mRNA profiling of 166 patient muscle biopsies from 12 disease groups showed a 56-member protein network centered on transforming growth factor-β (TGFβ) associated with severe pathology disease groups (fibrosis and failed regeneration). Studying the second independent muscle biopsy data set of four muscle diseases (49 biopsies) showed that indeed the TGFβ network is driven by tissue fibrosis and failed regeneration. Superimposing a TGFβ-associated network on a 27-time point murine normal muscle regeneration series showed that these 56 members of the TGFβ network are also expressed during normal regeneration process. However, the single network parsed into time-specific sub-networks, suggesting that failed regeneration may be due to inappropriate crosstalk between different temporal stages of regeneration happening within the same dystrophic microenvironment. To evaluate inappropriate crosstalk that may explain weakness and failed regeneration in muscular dystrophies, we developed an asynchronous remodeling model. According to this model, a 2-week regeneration process was predicted to be disrupted due to inappropriate crosstalk between neighboring injured myofibers that are at different stages of regeneration. To create an experimental mouse model with focal asynchronous bouts of muscle regeneration, we induced episodes of staged degeneration and regeneration in normal (wild-type) mouse muscle using localized notexin injection at two adjacent sites in different time points during the 2-week regeneration window. The first set of injections was separated by four days and the second set by 10 days. Laser capture microscopy was used to isolate each region of regeneration (first bout and second bout) and the myofibers in between the staged injection sites (crosstalk areas). The mRNA profiling and immunohistochemical studies showed that the crosstalk areas become inappropriately fixed in the developmental time point, by which the initial bouts were separated. Consequently, these conditions mediated a chronic inflammatory state and mitochondrial insufficiency in 4 days and a chronic pro-fibrotic state in 10 days in crosstalk areas. We showed that molecular networks associated with these localized areas of pro-inflammatory states were suppressed by treatment with glucocorticoids and VBP15. In conclusion, our data suggest that neighboring asynchronous bouts create inappropriate crosstalk between cells in different stages of the muscle regeneration, resulting in failed regeneration and the pro-fibrotic/ pro-inflammatory states. However, successful muscle remodeling is a synchronous process where all myofibers are at the same stage of regeneration and can complete the regeneration process in an orchestrated manner. Furthermore, our data support a model for failed regeneration and pathological fibrosis conditions in muscular dystrophies. Failure in the regeneration model can be generalized to chronic inflammatory states in other diseased tissues, where injured cells cannot successfully regenerate. This model also introduces a novel mechanism of action for glucocorticoids in many of these disorders in that they serve to resynchronize remodeling, much as diurnal cortisol fluctuations do in most animals.

Comment: The general loss of regenerative capacity observed in aging subjects contributes to frailty, loss of independence, and poorer outcomes after a wide range of therapeutic interventions. Although it is tempting to assume that cell-intrinsic deterioration of the patient's stem or progenitor cells is to blame, evidence is mounting that in many cases the accumulation of toxic damage to the microenvironment is the primary culprit. ⁹ However, the specific aspects of the aged milieu that mediate this impairment remain unclear. In this thesis, the regeneration of healthy skeletal muscle is shown to occur in distinct consecutive stages, with the signals that ensure proper regulation and progression of the program being vulnerable to interference from other nearby injuries or from more widespread chronic inflammation. Both factors are likely to contribute to the decline in regenerative potential in the elderly, and the feed-forward nature of this interference —where each injury slightly worsens the prognosis of every other—fits well with the accelerating decline observed in patients approaching the end of life, arguing strongly for therapeutic intervention as early as possible in the aging process.

Atrophied Thymus-Generated Th17 Cells Have Tissue-Specific Involvement in Autoimmunity

Jennifer Shaw, PhD, University of North Texas

The thymus is the organ of T cell development and is responsible for depleting self-reactive T clones via negative selection. However, natural aging involves the progressive loss of FoxN1, which leads to age-related thymic atrophy. Aging also causes an increased incidence of autoimmune disorders, in which Th17 cells are known to be involved. Because Th17 cells are confirmed to be generated in the thymus, it can be questioned whether atrophied thymus-generated auto-reactive Th17 cells cannot be completely depleted and therefore contribute to the greater risk of autoimmunity in the elderly. Using a mouse model (FoxN1 conditional knockout, FC) that mimics an aged thymus, we investigated this question. We found the proportion of natural Th17 (nTh17) cells, which are found in the thymus, tended to be higher after induction in an atrophied thymus compared to a normal thymus. However, peripheral Th17 cells generated from the atrophied thymus remained unchanged. Despite no change in peripheral Th17 proportions, when transferred into a young healthy host (RAG knockout mouse) lacking T and B cells, these peripheral Th17 cells caused inflammation and cellular infiltration in the colon and lung but not in the salivary or lacrimal gland. In addition, we induced experimental autoimmune encephalomyelitis (EAE), and we found no difference in disease onset or severity between mice with an atrophied thymus and those with a normal thymus. These data imply that self-reactive Th17 cells generated from an aged atrophied thymus do escape depletion and are released into the periphery. These cells have characteristics of tissue-specific self-reactivity, which may be related to the local microenvironment, such as existence of bacteria. In conclusion, thymic atrophy results in a defective process of negative selection and therefore allows auto-reactive Th17 cells to survive and leave the thymus. In the periphery, the auto-reactive Th17 cells can cause tissue-specific age-associated development of autoimmune phenotype. Our data can contribute to a mechanistic understanding of age-associated autoimmune disease and therefore is likely to contribute meaningfully to evidence-based therapeutically useful approaches to target self-reactive Th17 cells in age-associated autoimmune disease.

Comment: The blood-borne cells of the immune system are subject to strict regulation of their population numbers, with significant deviations occurring only transiently in response to an active pathogen. Although this homeostasis effectively prevents dangerous over- or under-proliferation of immune cells, difficulties can arise when the individual cells themselves become somehow defective, as occurs in aging. The contribution of hypo-reactive cells—particularly anergic T lymphocytes, apoptosis-resistant but functionally impotent clones induced by latent viral infection—to immunosenescence is well established, resulting in an impaired ability to respond to novel pathogens. This phenotype is, of course, absent in the chronologically youthful FC mice described. The increasing level of autoimmune disease associated with advancing age raises the possibility of a role for hyper-reactive immune cells, and this thesis establishes the identity of one such subset—natural Th17 cells (distinct from the more well-studied induced Th17 cells, which differentiate from CD4⁺ T cells in peripheral sites) subjected to inadequate negative selection against self-antigen due to the impaired function of the thymus where they develop. It will be important to determine whether distinct molecular markers exist to allow for the selective elimination of this population in elderly patients, similar to the planned therapy for anergic lymphocytes. ¹⁰ However, as in the latter case, the regeneration of a functional thymus ¹¹ (or, in younger patients, the prevention of involution ¹² ) is anticipated to be a requirement to fully regain youthful immunocompetence.

Controlling Micro- and Nano-Environment of Tumor and Stem Cells for Novel Research and Therapy of Brain Cancer

Christopher Smith, PhD, Johns Hopkins University

The use of modern technologies in cancer research has engendered a great deal of excitement. Many of these advanced approaches involve in-depth mathematical analyses of the inner working of cells via genomic and proteomic analyses. However, these techniques may not be ideal for the study of complex cell phenotypes and behaviors. This dissertation explores cancer and potential therapies through phenotypic analysis of cell behaviors, an alternative approach. We employ this experimental framework to study brain cancer (glioma), a particularly formidable example of this diverse ailment. Through the application of micro- and nanotechnology, we carefully control the surrounding environments of cells to understand their responses to various cues and to manipulate their behaviors. Subsequently, we obtain clinically relevant information that allows better understanding of glioma and enhancement of potential therapies. We first aim to address brain tumor dispersal through analysis of cell migration. Using nanometer-scale topographic models of the extracellular matrix, we study the migratory response of glioma cells to various stimuli in vitro. Second, we implement knowledge gained from these investigations to define characteristics of tumor progression in patients, and to develop treatments inhibiting cell migration. Next we use microfluidic and nanotopographic models to study the behaviors of stem cells in vitro. Here we attempt to improve their abilities to deliver therapeutic proteins to cancer, an innovative treatment approach. We analyze the multi-step process by which adipose-derived stem cells naturally home to tumor sites and identify numerous environmental perturbations to enhance this behavior. Finally, we attempt to demonstrate that these cell culture-based manipulations can enhance the localization of adipose stem cells to glioma in vivo using animal models. Throughout this work, we use environmental cues to analyze and induce particular behaviors in cells. We further demonstrate that this general technique can be used to determine clinically relevant tumor characteristics, to identify potential drug targets, and to enhance potential therapies. Therefore, this thesis illuminates a novel framework for experimentation into cancer and specifically advances two treatment approaches. We anticipate that the methodologies described in this study will prove useful to various branches of medicine and biological research.

Comment: One of the most common methods of identifying a particular cell population is through the detection of surface markers—molecules expressed on the plasma membrane—distinct to it. Although this approach is flexible and straightforward, it is a technique of convenience. In relatively few cases is the presence of a particular marker actually of direct therapeutic significance; rather, it signals the presence of a pathological phenotype. Additionally, it relies on prior knowledge of relevant surface marker(s), which may be unknown in poorly characterized diseases or those characterized by rapid unpredictable mutation. In this interesting thesis, the authors exploit modern imaging and micro-fabrication techniques to directly characterize cell behavior, yielding multiple insights into how environmental manipulations might serve to inhibit the migration of cancerous cells or better deliver therapeutic cells to a particular location. Such manipulations are likely to be effective against a wider range of cell populations and mutation states than those aimed at a particular biochemical target. Establishing reliable behavioral markers of healthy cell function, as well as of different pathologies, including those associated with “natural” aging, will enable the identification of anomalous phenotypes, even in the absence of a known molecular signature. Such cells can then be specifically isolated and subjected to thorough genetic and/or proteomic analysis, facilitating the detection of new disease-relevant pathways and interventions.

Genetic Engineering with Customizable Recombinases and Nucleases

Thomas Gaj, PhD, Scripps Research Institute

The development of new tools that allow for the targeted modification of genomes is providing researchers with previously unattainable forms of control over diverse biological processes. Site-specific recombinases represent one such class of tools (Chapter 1). The rigid specificities of many site-specific recombination systems have limited their adoption in fields that require site-specific recombinase specificity to be easily customized, however. To address this problem, we developed a robust strategy for evolving site-specific recombinases with novel substrate specificities (Chapter 2). This stringent selection system is based on enzyme-mediated reassembly of the gene encoding β-lactamase. This approach enables identification of site-specific recombinase variants that broadly react with sequences within the human genome after only three rounds of selection. By combining this strategy with site-saturation mutagenesis, we show that site-specific recombinases with comprehensively re-designed target specificities can be created (Chapter 3). These enzymes recognize unnatural sequences >10,000-fold more effectively than their parental counterparts and catalyze targeted integration into the human genome with >80% specificity. We expand on these studies by combining substrate specificity analysis and directed evolution to develop a diverse collection of recombinase catalytic domains capable of recognizing an estimated 3.77×10⁷ unique DNA sequences (Chapter 4). We show that zinc finger recombinases (ZFRs) assembled from these re-engineered catalytic domains recombine user-defined DNA targets with high specificity and that designed ZFRs integrate DNA into targeted endogenous loci in human cells.

Site-specific nucleases are another class of enzymes that allow for the introduction of a broad range of custom alterations (Chapter 5). These customizable enzymes are based on programmable, sequence-specific DNA-binding modules linked to a non-specific DNA cleavage domain. In particular, zinc finger nucleases (ZFNs) have enabled highly efficient gene disruption in numerous cell types and model organisms and have facilitated the progress of targeted gene therapy in humans. However, contemporary DNA and mRNA-based ZFN delivery methods are associated with undesirable side effects that may limit the continued advancement of this technology. To address this problem, we investigated the direct delivery of purified ZFN proteins into cells (Chapter 6). We demonstrate the intrinsic cell-penetrating capabilities of ZFN proteins and show that direct delivery of ZFN proteins leads to highly efficient endogenous gene disruption in a variety of mammalian cell types. We also show that ZFN protein delivery leads to comparatively fewer off-target effects than ZFN gene delivery systems that rely on expression from nucleic acids.

Together, these studies demonstrate the feasibility of generating customizable site-specific recombinases for targeted genetic engineering and provide researchers with a new means for inducing targeted alterations in a safe and efficient manner.

Comment: A great majority of diseases seem to be prospectively curable or preventable through the rational editing of host DNA, either to repair a mutated gene or to introduce novel genetic material enabling a new or improved biochemical capability. However, the high risk that errors arising during insufficiently accurate editing will induce cancer has largely prevented such therapies from being brought to market, or even to clinical trials. The recent advent of engineered versions of the bacterially derived CRISPR/Cas9 system, ¹³ the selectivity of which is dependent on a guide RNA supplied with the nuclease, has been hailed as a breakthrough in this area but remains largely untested in humans and suffers from some limitations that may restrict its long-term therapeutic utility. Particularly, it is restricted to recognizing sequences of 12–20 nucleotides in length, fails to function with some potential guide sequences for as-yet-unknown reasons that may relate to RNA secondary structure, and is dependent on the target DNA already containing an (admittedly short) pre-existing motif directly adjacent to the editing site. Zinc finger nucleases (ZFN) have none of these limitations and thus are likely to remain therapeutically relevant for many years to come. This thesis makes progress in addressing two of the major challenges in their further development—the relatively slow and laborious process of optimizing a given enzyme's specificity against its intended target sequence and the compounding of targeting errors that occurs when the ZFN is delivered via a nucleic acid vector rather than as pure protein.

Intranasal Delivery of pGDNF Nanoparticles for Parkinson Disease

Brendan Harmon, PhD, Northeastern University

Parkinson disease (PD) is a progressive neurodegenerative disorder that primarily affects the dopaminergic A9 nigrostriatal tract. For dopamine neurons specifically, glial cell-derived neurotrophic factor (GDNF) has been shown to promote their survival and proliferation both in culture and in vivo. GDNF has also proven to be neuroprotective and restorative in various animal models of PD and some human clinical trials. However, its delivery to the brain has required invasive surgical routes that are not clinically practical for many patients. The main objective of this project was to test intranasal delivery to the brain of a nanoparticle vector incorporating an expression plasmid for GDNF (pGDNF). The intranasal route circumvents the blood–brain barrier, allowing larger-sized vectors into the central nervous system while avoiding peripheral distribution. This approach would provide a renewable source of GDNF within the target areas of the brain, the striatum and the substantia nigra (SN), without the need for surgical injections or frequent re-dosing. A PEGylated polylysine compacted plasmid nanoparticle vector (PEG-CK30), developed by Copernicus Therapeutics, Inc., has been shown to transfect neurons and glial cells in vivo while lacking the safety issues present with other vectors.

The first goal of this work was to determine if these PEG-CK30 compacted plasmid nanoparticles can successfully transfect cells and express the reporter protein, enhanced green fluorescent protein (eGFP) in the rat brain after intranasal administration. Initial in vivo experiments used the expression plasmid pCG, expressing eGFP under the fast-acting cytomegalovirus (CMV) promoter. Intranasal administration of pCG nanoparticles resulted in evidence of transfection of brain cells, as shown both qualitatively, by GFP immunohistochemistry, and quantitatively, by GFP–enzyme-linked immunosorbent assay (ELISA). Expression was detected throughout the rat brain 2 days post-administration.

Following the proof-of-principle study with pCG, a new plasmid was created by Copernicus Therapeutics, Inc. to better mimic their long-lasting pGDNF plasmid while providing both GDNF as well as the reporter function of eGFP. This eGFP-GDNF plasmid was used to monitor expression and cell-types transfected. This expression plasmid, called pUGG, was first characterized in vitro to verify protein expression. Transfection experiments in SHEP-1 neuroblastoma cells, ventral midbrain cultures, and N27 dopaminergic cells all demonstrated that pUGG expressed bioactive eGFP and GDNF. However, cleavage of the two proteins did not occur and the expressed protein emerged as a fusion construct that was not detectable by GDNF-ELISA, although it was detected by GFP-ELISA.

The next goal was to determine if pUGG was able to transfect cells in vivo in rat brain. Direct striatal injection of pUGG nanoparticles showed significant eGFP expression at the site of injection both 7 and 14 days post-administration with no difference in eGFP expression between the two time points. GFP immunohistochemistry at the striatal injection site revealed expression of eGFP-positive cells as well as evidence of GDNF's bioactivity as indicated by neurite outgrowth. Moving forward, we administered pUGG nanoparticles intra-nasally to rats and found significant expression seven days later throughout the brain, with highest levels in the forebrain areas (olfactory bulb and frontal cortex). Significant expression was also seen along the rostral–caudal axis of the brain compared with naked pUGG plasmid.

The final goal of this work was to examine whether intranasal pGDNF pre-treatment could generate sufficient GDNF to protect SN dopamine neurons after a unilateral 6-hydroxydopmaine (6-OHDA) lesion, a common animal model for PD. Copernicus' pGDNF plasmid was used for the neuroprotection experiments to avoid possible confounds due to the GFP fusion produced by pUGG. Tyrosine hydroxylase immunostaining density was used as a marker for dopamine neurons in the SN and their nerve terminals in the striatum. Dopamine cell counts were also performed in the SN.

Intranasal delivery of pGDNF significantly protected dopamine neurons in the rat 6-OHDA model of PD. This was revealed in three ways. First, pGDNF treatments reduced amphetamine-induced circling behavior, suggesting a prevention of dopamine loss on the 6-OHDA–lesioned side. Second, pGDNF increased TH staining density and dopamine cell counts in the SN on the 6-OHDA–lesioned side. This result was direct evidence of neuroprotection of dopamine cell bodies. Third, pGDNF increased TH staining density in the striatum on the 6-OHDA–lesioned side. This result was direct evidence of protection of dopaminergic nerve terminals. Intranasal pGDNF nanoparticles provided greater neuroprotection than naked pGDNF for all measures. This result was consistent with our previous findings that pGDNF nanoparticles produce more GDNF in brain than the naked plasmid.

Collectively, these results demonstrate that intranasal delivery of Copernicus' pGDNF nanoparticles has great clinical potential as a new, non-invasive and non-viral gene therapy approach for early-stage PD. By promoting recovery of damaged neurons and preventing further cell loss, symptoms may be reversed and disease progression may be stopped.

Comment: Glial cell-derived neurotrophic factor (GDNF) has a multitude of roles during embryonic development, but in the adult brain it is thought to have substantial effects only on dopaminergic and motor neurons, being in both cases potently neuroprotective (in contrast to some of the less benevolent products of aging glial cells ¹⁴ ). Continuous intra-ventricular administration of the protein has been shown to delay and to some extent reverse symptoms of Parkinson disease (PD) in primates, ¹⁵ but this modality is clinically impractical for patients. Unfortunately, bolus delivery in the human brain has not proven beneficial, and the effects of GDNF on the kidneys, gonads, and other locations plus its poor ability to cross the blood–brain barrier (BBB) make systemic infusions hazardous. Intranasal delivery is becoming well established as a practical option for delivery of genetic material (and cells ¹⁶ ) to the central nervous system without the need to cross the BBB, also benefiting from a reduced risk of immune reaction. This encouraging study shows that an intra-nasally administered GDNF-bearing vector can become stably expressed in the brain for extended periods, and functions to protect dopaminergic neural cell bodies and synapses from 6-hydroxydopmaine (6-OHDA)-induced damage. Like many neurodegenerative diseases, PD is characterized by aggregates, in this case, alpha-synuclein inclusions called Lewy bodies. Although their pathological role is not well understood, the presence of Lewy bodies associates strongly with neuron loss and with specific cognitive dysfunctions. As in other such cases, the complete removal of such intracellular junk is anticipated to prevent further disease progression and introduce the possibility of regeneration, but may also temporarily place additional stress on the afflicted cells—a risky prospect for post-mitotic neurons. Consequently, the therapy described is likely to be beneficial in at least three cases: In early-stage and pre-clinical patients, to delay disease progression; at the time of administration of an aggregate-clearing approach, to reduce collateral damage; and subsequently, to encourage the surviving cells to form additional synaptic connections, helping to compensate for those lost to disease.

Potential Role of Tazarotene-Induced Gene 3 and Transglutaminase 1 in Tauopathies

Yasin Kizilyer, PhD, University of Maryland

The family of neurodegenerative diseases, called tauopathies, includes Alzheimer disease, progressive supranuclear palsy, and Pick disease. These diseases are associated with the formation of tau inclusions and neuronal cell death. Tau inclusions are characterized by the presence of neurofibrillary tangles (NFTs), where the protein Tau, a member of microtubule-associated proteins (MAPs), is found to be hyper-phosphorylated and aggregated. The formation of these aggregates leads to microtubule instability and cellular toxicity. There are several lines of evidence that demonstrate this aggregation is due to transglutaminase mediated cross-linking of Tau protein. However, little is known about the mechanism of the hyper-phosphorylation, aggregation, and Tau-associated toxicity. Analysis of brains with Tau pathologies showed the co-localized expression of transglutaminase 1 (TG1) and its activator, tazarotene-induced gene 3 (TIG3) in NFTs, suggesting functional relevancy. Therefore, we hypothesized that TIG3 regulates TG1-catalyzed cross-linking of Tau, and therefore the formation of Tau inclusions in neurons. To test this hypothesis, neurons differentiated from human embryonic stem cells used to develop an in vitro tauopathy model. Endogenous expression analyses of TIG3, TG1, and Tau were performed and induction of Tau occlusions via overexpression of TIG3 studied. Results have shown endogenous TIG3 mRNA and protein were expressed in neurons derived from embryonic stem cells. Endogenous TIG3, TG1, and TAU have shown co-localization. Cells overexpressing TIG3 have shown a higher level of co-localization and increased Tau aggregation, which confirmed our hypothesis. The results of this study provide new mechanistic insights into the formation of neurofibrillary tangles and the resulting pathology.

Comment: Transglutaminases (TGs) are a family of enzymes that catalyze the formation of cross-links between a glutamine residue and an amine-terminated side chain, such as a lysine residue. The bonds formed are extremely stable and resistant to proteolytic degradation and are widely used by the body to produce robust polymers, notably in the clotting and healing process; consequently TG up-regulation is a common response to injurious stimuli. The influence of TGs on the aging of the cardiovascular system is well established. ¹⁷ Although numerous proteins implicated in neurodegenerative disease have been shown to be TG substrates, the role of these enzymes in the disease course remains unclear. There is evidence that TG activity can either promote or inhibit aggregation of susceptible peptides, ¹⁸ depending on the precise cross-links formed, which in turn depends on the molecular environment of the reaction. Given the increasing evidence that at least some aggregate-prone proteins exhibit their highest toxicity in small-molecular-weight oligomers, and that precipitation of these species can reduce toxicity, ¹⁹ this data seems to fit a pleiotropic situation where low levels of TG activity are protective (by sequestering oligomers into less harmful and amyloidogenic forms), whereas excessively high/chronic expression or poor conditions could instead promote runaway aggregation. TG inhibition may thus prove clinically useful either by delaying the progression of existing disease, or by enhancing the effectiveness of aggregate-clearing therapies such as lysosomal modulation ²⁰ or up-regulation of autophagy. ²¹