Commentary on Some Recent Theses Relevant to Combating Aging: August 2018

There are many biological questions that require single-cell analysis of gene sequences, including analysis of clonally distributed dimeric immunoreceptors on lymphocytes (T cells and B cells) and/or the accumulation of driver/accessory mutations in polyclonal tumors. Lysis of bulk cell populations results in mixing of gene sequences, making it impossible to know which pairs of gene sequences originated from any particular cell and obfuscating analysis of rare sequences within large populations. Although current single-cell sorting technologies can be used to address some of these questions, such approaches are expensive, require specialized equipment, and lack the necessary high-throughput capacity for comprehensive analysis. Water-in-oil emulsion approaches for single-cell sorting have been developed but droplet-based single-cell lysis and analysis have proven inefficient and yield high rates of false pairings. Ideally, molecular approaches for linking gene sequences from individual cells could be coupled with next-generation high-throughput sequencing to overcome these obstacles, but conventional approaches for linking gene sequences, such as by transfection with bridging oligonucleotides, result in activation of cellular nucleases that destroy the template, precluding this strategy. Recent advances in the synthesis and fabrication of modular deoxyribonucleic acid (DNA) origami nanostructures have resulted in new possibilities for addressing many current and long-standing scientific and technical challenges in biology and medicine. One exciting application of DNA nanotechnology is the intracellular capture, barcode linkage, and subsequent sequence analysis of multiple messenger RNA (mRNA) targets from individual cells within heterogeneous cell populations. DNA nanostructures can be transfected into individual cells to capture and protect mRNA for specific expressed genes, and incorporation of origami-specific bowtie-barcodes into the origami nanostructure facilitates pairing and analysis of mRNA from individual cells by high-throughput next-generation sequencing. This approach is highly modular and can be adapted to virtually any two (and possibly more) gene target sequences, and, therefore, has a wide range of potential applications for analysis of diverse cell populations such as understanding the relationship between different immune cell populations, development of novel immunotherapeutic antibodies, or improving the diagnosis or treatment for a wide variety of cancers.

Comment: Aging is characterized by the accumulation of molecular damage throughout the body, produced largely by random side-effects of metabolism and kept in check effectively for most of the lifespan by a combination of endogenous repair mechanisms and pre-existing reserve capacity at the tissue and organ level. Particular combinations of damage can result in a permanently dysfunctional (senescent) state without leading to the death and replacement of the host cell, a phenomenon that has for some time been suspected to play a role in age-related pathology out of proportion to the actual number of cells afflicted; indeed, recent progress in the ablation of particular types of senescent cells has robustly confirmed that hypothesis, demonstrating remarkably significant improvements in health span with no detectable side-effects. However, we must not be blinded by that success—it would be very premature to assume that the cell types thus far targeted represent all of those that place a burden on elderly bodies! This exciting thesis introduces a general high-throughput mechanism for identifying gene coexpression in single cells within a population. Thus for the first time, it opens the door to the exhaustive investigation of single-cell level expression signatures, and, in turn, to a comparatively unbiased catalog of rare but potentially pathologically significant deviations from healthy youthful biology. We are delighted to see that the author has already begun the process of bringing this technology to market through Gemneo Bioscience, and we very much hope to see it applied to “normal” age-related decline in the very near future.

Development of Canine Chimeric Antigen Receptor T Cell Therapy for Treatment and Translation

Panjwani Mohammed PhD University of Pennsylvania

Chimeric antigen receptor (CAR) T cell therapy has had remarkable success targeting B cell leukemias in human patients, but unexpected toxicities and failures in other disease demonstrate the need for more predictive preclinical animal models than the murine ones currently used. Dogs develop spontaneous malignancies similar to humans in their tissues of origin, gene expression profiles, treatments, and disease courses, and have long been used as models for immunotherapy. I hypothesize that the development of CAR T cell therapy for dogs with spontaneous disease and that the treatment of these canine patients will recapitulate the observations found in human patients, and provide new insights into the safety and efficacy of this breakthrough therapy. To achieve this, I first established methods for growing primary canine T cells from healthy and disease-bearing donors to clinically relevant scale, developed RNA electroporation protocols to transiently express a CAR targeting the canine tumor-associated antigen CD20, demonstrated its function in vitro, and treated a relapsed canine B cell lymphoma patient with autologous CAR T cells as a proof of feasibility. I then developed methods to permanently express a second-generation cCD20-8-28-ζ CAR in canine T cells using lentiviral transduction, showed in vitro antigen-specific function and proliferation of CAR T cells, and treated four canine B cell lymphoma patients with CAR T cells. Based on my observations from those patients, I made iterative improvements to the T cell culture system and CAR construct, and treated a canine B cell lymphoma patient with cCD20-8-BB-ζ CAR T cells, whose tumor cells lost target antigen expression to avoid immune pressure. These results prove that it is not only possible to generate canine CAR T cell therapy, but that it recapitulates observations found until now only in human patients. In addition, novel findings regarding the recovery of T cells during ex vivo culture and the host immune response to the CAR demonstrate that this model can already inform human medicine.

Comment: Mice have been the preferred subjects of biomedical research for most of the modern era—despite their rather limited phenotypic similarity to humans—largely for simple reasons of economy, and although both naturally mutant and engineered murine models will surely remain scientifically valuable for decades to come, it is also becoming clear that mice are markedly less reliable in modeling diseases of a chronic and/or multigenic nature. Conversely, the use of species—such as chimpanzees—that exhibit the greatest similarity to humans, and thus to human disease, is an extremely expensive and lengthy process. In seeking a middle ground, dogs merit special consideration; not only do they naturally develop a wide range of human-relevant age-related diseases (and several fold more rapidly), they are also numerous (hundreds of millions of animals worldwide), well cared for (the United States spends >$40 billion per year on dog care), and, despite a more distant common ancestor, are on average more genetically similar to humans than are rodents. Their status as companion animals makes canines particularly viable as subjects in preclinical trials of novel therapeutics, with an expectation of direct benefit to the owner as well as the more abstract and thus sometimes less motivating potential to contribute to human medical advances. CAR-T therapy, first devised in 1989, involves modifying the T cell receptor of patient cells by introducing a fixed recognition domain (typically antibody derived) to enforce targeting against a specific antigen. It has since moved through successive generations implementing an increasing number of costimulatory modules to improve persistence in the body and clinical potency, with >200 clinical trials currently in progress; most recently, the introduction of “adaptor” small molecules as a bridge between the engineered TCR and an antigen has made it possible for the target to be changed even during therapy. This thesis recapitulates a significant part of that nearly 30-year history in dogs (possible thanks to the accelerated timeline of treatment, analysis, and refinement), and so lays essential groundwork for the use of canine subjects in the burgeoning field of CAR-T research and development. Meanwhile the discovery that the phenomena previously associated only with human treatment response are also present in our “best friends” will lend further credibility to their wider use in studies of chronic and age-related disease.

Endocytic Vesicle Rupture in the Pathogenesis and Propagation of Neurodegenerative Proteinopathies

Flavin William PhD Loyola University Chicago

Numerous pathological amyloid proteins spread from cell-to-cell during neurodegenerative disease, facilitating the propagation of cellular pathology and disease progression. Understanding the mechanism by which disease-associated amyloid protein assemblies enter target cells and induce cellular dysfunction is, therefore, key to understanding the progressive nature of such neurodegenerative proteinopathies. In this study, we utilized an imaging-based assay to monitor the ability of disease-associated amyloid assemblies to induce the rupture of intracellular vesicles following endocytosis, as well as to elucidate the cellular consequences of this damaging mechanism of invasion. We observed that the ability to induce vesicle rupture is a conserved feature of fibrillar amyloid assemblies of alpha-synuclein, tau, and polyglutamine-expanded huntingtin. In the case of alpha-synuclein amyloid assemblies, we determined that Serine 129 phosphorylation and strain conformation dictate the potency of endocytic vesicle rupture. We also demonstrated that vesicles ruptured by alpha-synuclein are lysosomes, and that these damaged vesicles are targeted to the autophagic degradation pathway. We observed that vesicles ruptured by alpha-synuclein can accumulate and fuse into large intracellular structures resembling Lewy bodies in vitro, and showed that the same markers of vesicle rupture surround Lewy bodies in brain sections from PD patients. These data underscore the importance of this conserved endocytic vesicle rupture event as a damaging mechanism of cellular invasion by amyloid assemblies of multiple neurodegenerative disease-associated proteins, and suggest that proteinaceous inclusions such as Lewy bodies form as a consequence of continued fusion of autophagic vesicles in cells unable to degrade ruptured vesicles and their amyloid contents.

Comment: A multitude of clinically significant neurodegenerative conditions display amyloid aggregates as a histopathological feature that correlates with disease progression, although the mechanisms by which such misfolded proteins exert their cytotoxic effects are still controversial. Many such aggregates are found in the extracellular space, and one seemingly rather straightforward therapeutic option was to introduce an agent capable of disrupting the plaque and thus perhaps permitting the clearance of its component peptides. Antibodies employed for this purpose showed promising results in vitro; unfortunately, clinical translation has been dogged by hazardous side-effects thought to be primarily due to transfer of amyloid to the cerebral vasculature, where it causes severe and often fatal microvascular disease (essentially exacerbating the cerebral amyloid angiopathy that already frequently copresents with plaques, and which is perhaps a sign of extant though inadequate clearance mechanisms), as well as to cross-reactivity with correctly folded amyloidogenic proteins expressed on cell membranes. Intriguingly, despite the extracellular location of mature plaques, recent studies have suggested that amyloids are not as toxic to external cellular membranes as was once assumed; amyloid-beta, in particular, accumulates on the plasma membrane, but may not actually disrupt it under reasonable conditions. ¹¹ Meanwhile, an earlier report from the author of this thesis indicated that downregulation of endocytosis can delay the spread of amyloid pathology, and vice versa ¹² —leading to the possibility, now herein confirmed across several relevant molecular species, that toxicity occurs within the endolysosomal system rather than (or at least in addition to) at the cellular surface. This is actually very exciting news; various methods to enhance the function of the lysosome are presently either in the clinic or in development, since a number of congenital and age-related pathologies are known or thought to involve failure of the cellular “waste disposal system.” It may, therefore, be possible to prevent vesicle rupture by delivering a potent antiamyloid protease to the lysosome, circumventing the side-effects of targeting extracellular amyloid. Indeed, the continued futile effort of the cell to degrade intracellular aggregates such as Lewy bodies through fusion with additional lysosomes might thus be made efficacious! Targeting earlier steps in the endosomal pathway may be necessary in the case of amyloids that induce rupture before fusion with a lysosome, but this is not an intractable problem—and demonstrating benefit in the simpler case of the lysosome will certainly incentivise its exploration.

Exploring Mechanisms of Metastasis Suppression in Metastatic Melanoma

Hampton Kelsey PhD University of Kansas

Melanoma is responsible for 76% of deaths from skin cancer, making it the deadliest form of commonly diagnosed skin cancer. The deadly nature of melanoma is due to its tendency towards rapid early metastasis. Metastasis, the process of cells exiting the primary tumor and forming secondary tumors in other parts of the body, accounts for the majority of morbidity and mortality associated with cancer. Therapeutically targeting and treating melanoma metastases is a challenging clinical goal, as metastatic cells are heterogeneous and can be morphologically and genetically distinct from the primary tumor. This dissertation examines two approaches for prevention or treatment of disseminated melanoma metastases: (1) reintroduction of metastasis suppressor protein fragments to prevent metastatic colonization and (2) treating disseminated metastases with a targeted small molecule treatment. By examining two discrete approaches of treating metastatic melanoma, this work sheds light on the clinical viability of using metastasis suppressors or metastasis-targeting drugs in patients with metastatic melanoma.

To examine strategies for metastasis suppression through metastasis suppressor proteins, we examined cleavage products of the metastasis suppressor KISS1, a metastasis suppressor protein. Expression of KISS1 inhibits metastatic colonization at secondary sites, rendering disseminated cells dormant. KISS1 must be secreted outside of the cell to suppress metastasis, where furin cleaves KISS1 into kisspeptins at three dibasic sites. This cleavage liberates an internal kisspeptin, Kisspeptin-54 (KP54, amino acid K⁶⁷ to F¹²¹), which is amidated and can bind a Gq/11-coupled protein receptor KISS1R. The mechanism of action for KISS1 metastasis suppression has long been assumed to be KP54 interacting with KISS1R. However, expression of KISS1R is not necessary for KISS1 metastasis suppression, and the extracellular processing of KISS1 hints at an alternative hypothesis: a different kisspeptin may be responsible for suppressing metastasis. To test this hypothesis, all possible kisspeptins (KISS1 Manufactured Peptides, or KMP) were generated based on known dibasic cleavage sites (M¹–Q¹⁴⁵; M¹–R⁵⁶; M¹–R⁶⁷; M¹–R¹²⁴; R⁵⁶–R⁶⁶; R⁶⁷–F¹²¹; R⁵⁶–F¹²¹; R⁵⁶–Q¹⁴⁵; R⁶⁷–Q¹⁴⁵; R¹²⁴–Q¹⁴⁵) and were used in an experimental metastasis assay to characterize their abilities to suppress metastasis. We found that while KP54 suppressed metastasis, additional KMP lacking the KISS1R binding site (LRF-NH₂) were able to completely suppress metastasis (p < 0.05). In particular, one kisspeptin (KMP2, M¹–R⁵⁶) suppressed metastatic traits in vitro as well as completely suppressing metastasis in vivo. To identify the signaling pathways used by KMP2 to suppress metastasis, a genome-wide CRISPR/Cas9 screen was performed in KMP2-expressing B16-F10 melanoma cells. As a whole, these data suggest that metastasis suppression by KISS1 is not necessarily contingent on KISS1R activation, and also supports investigation into additional receptors.

To investigate the efficacy of targeting metastases with small molecules, we also investigated the impact of ML246 (also known as metarrestin). Metarrestin was discovered by a high throughput assay for molecules which disassemble the perinucleolar compartment (PNC). PNCs are structures composed of RNA and RNA binding proteins near the nucleolus. These structures are enriched in metastatic cells and are druggable targets which target metastases and not normal epithelium. We examined the impact of metarrestin treatment on orthotopic tumor growth, microscopic metastasis formation, and macroscopic metastasis formation. We found that metarrestin treatment had no significant impact on metastatic outgrowth, but suppressed intradermal tumor growth. Based on these data, we can infer that PNC-positive metastases may be too small a population to effectively target in this model. This treatment paradigm may be more effective in conjunction with a more potent approach to metastasis suppression. Overall, the work in this dissertation identified a potent metastasis suppressing fragment of KISS1, KMP2, and described the efficacy of metarrestin treatment of disseminated metastases. The metastasis suppression induced by KMP2 expression was far more potent than the effects of metarrestin treatment on suppressing metastatic colonization and outgrowth, suggesting that treatment deliveries and targets are critical considerations in the development of antimetastatic therapeutics.

Comment: KISS1 was identified in 1996 in Dr. Danny Welch's laboratory in Hershey, Pennsylvania (hence the name—for Hershey's Kisses) during an effort to explain how the reintroduction of human chromosome 6 suppressed metastasis in melanoma cell lines by as much as 95% ¹³ —studies that were also among the first to concretely establish metastasis as a phenotype separable from tumorigenicity. Subsequent work soon determined that KISS1 itself resides on chromosome 1 (whereas its critical upstream regulator CRSP3 is indeed encoded on chromosome 6) and that its effects are observable in many other cancers, although it remains best studied in the context of melanoma. Unlike other known tumor suppressor genes, KISS1-expressing cells complete all the typical steps of the metastatic cascade, but once disseminated they become indefinitely dormant and fail to colonize their new environment—apparently at least, in part, due to a disruption of the Warburg effect and return to oxidative metabolism—whereas from a translational viewpoint, the fact that KISS1 exerts its antimetastatic effects from the extracellular space is obviously very advantageous. Yet KISS1 signaling also plays a range of important biological roles, especially in the context of reproductive tissues, and thus prolonged dosing with the protein carries a risk of side-effects that have impeded its clinical exploitation. This thesis (also from the Welch laboratory) revisits in more detail how KISS1's effect is mediated, highlighting a variant fragment of the peptide that is able to completely suppress metastatic growth in vivo—a remarkably noteworthy milestone! Meanwhile, the use of Genome-wide CRISPR Knock Out (GeCKO) (discussed further in the next commentary) to unpick the pathways by which KMP2 acts on cells will provide invaluable guidance to those working to develop a clinically viable version of the agent. Although such a therapeutic can only delay disease progression, the time bought by its employment will certainly allow physicians to characterize and eradicate the malignant population, both primary and disseminated, far more effectively than is currently the case.

Polymer and Nucleic Acid Self-Assemblies: Properties and Applications at the Biological Interface

Barnhill Sarah PhD University of California, San Diego

Ring opening metathesis polymerization (ROMP) was used to generate a variety of self-assembled nanostructures, including purely synthetic and biohybrid materials. The properties of polynorbornene amphiphilic block copolymer structures and their relationship to resulting morphology were explored, paving the way for nanoparticle design at the structural and processing level. In the context of biosynthetic polymer amphiphiles, hydrophobic ROMP polymers were attached to hydrophilic DNA strands to produce self-assembly micelles on the order of 20 nm in diameter. Herein, we explored the self-assembly properties, stability, and applications of these assemblies in pristine conditions and cellular environments.

Morphology plays an important yet poorly understood role in dictating how nanomaterials interact with cells and tissues. Uncovering this relationship relies on working out how to control particle morphology in the first place. We prepared purely synthetic amphiphiles using ROMP to prepare aqueous phase diagrams of block copolymer assemblies. By preparing polymers with varying properties, such as block lengths, block identity, and block ratios, the relationship between polymer structure and the resulting self-assembly nanostructure could be observed under certain conditions. Furthermore, by manipulating the assembly conditions of these polymers, we have shown that multiple stable morphologies can be generated from the same block copolymer starting material. This represents the first study of its kind for ROMP-derived amphiphilic assemblies, which exhibit variations in self-assembly dynamics compared with more traditional block copolymers.

Adding a level of complexity to our block copolymer system, we next explored more therapeutically relevant systems by conjugating DNA to a hydrophobic ROMP homopolymer and assembling them into DNA-displaying micelles. We determined the stability of the DNA on the micelle surface by treating the structures with various nucleases and human serum. The stability of the DNA on the micelle corona resisted degradation by nucleases in some circumstances, but not all, relative to the free DNA control, highlighting the importance of careful design of the amphiphile for a given application.

After determining the stability of DNA polymer assemblies (DPAs), antisense DPAs were designed against a known therapeutic target in cancer cells, MDR1. Importantly, these materials were designed with sequences not containing chemical modifications, such as locked nucleic acids or other backbone alterations. After treating MDR1-dependent doxorubicin-resistant cells with the antisense micelles, sensitivity to the chemotherapeutic could be restored to near-parental cell line IC₅₀ values.

Despite many desirable properties nanoparticles have in therapeutic applications, a major bottleneck in their development is the fact that very little is known about how they interact with cells and tissues. Next, a high-throughput whole-genome approach to elucidate the pathways responsible for nanomaterial uptake by cells was developed and tested using DPAs as a proof-of-concept. Using GeCKO, a population of cells representing knockouts across the entire genetic spectrum was tested against uptake of cyanine 5 labeled DPAs. Using this approach, we have identified the transmembrane protein SLC18B1, among a handful of other proteins, as candidate for mediating uptake; a previously unknown interaction by DNA-displaying nanomaterials with cell surfaces. By expanding this technique to other categories of nanoparticle medicines with different structures and surface modifications, the generation of new design rules for nanomaterial therapeutics may be prepared to help researchers avoid off-target accumulation and advance many more nanotherapies to the clinic.

Comment: The general class of nanoparticles includes any material with dimensions entirely on the <100 nm scale, at least two orders of magnitude smaller than a typical human cell. Research into the medical applications of nanoparticles has been ongoing for several decades, with major results in imaging (quantum dots, superparamagnetic iron oxide, etc.) and in drug delivery (polymer-based systems, artificial liposomes, etc.)—in many cases enabling experimental or clinical work that would otherwise be entirely impossible. Despite these successes, the lack of a robust theoretical basis for predicting nanoparticle properties and their interactions with living tissue has hindered both the iterative improvement of existing species and the design of new classes. This thesis contributes toward the development of that knowledge by investigating in detail the role of morphology in defining the properties of DPAs, but even more significantly so by demonstrating proof of principle for the use of a very widely applicable method to elucidate how novel nanoparticle species enter and interact with living cells. GeCKO, first published in late 2013, employs the CRISPR-Cas9 system in combination with a genome-wide array of oligonucleotide single guide RNAs (sgRNAs) to knock out distinct genetic loci in individual cells within a pooled population ¹⁴ ; furthermore, GeCKO vectors include the Cas9 nuclease along with its guide RNA, and thus do not require prior modification of the target cells. A subsequent positive or negative selection step is then followed by deep sequencing to reveal depletion of sgRNAs that target genes conferring a selective advantage. Although RNA interference has historically been applied toward the same goal, the inherent incompleteness of RNAi-based knockdown and its greater propensity for off-target effects renders GeCKO a substantially more reliable tool; thus we very much expect to see its broader use across the biomedical sciences in the coming years.

Towards a Scalable Biomimetic Antibacterial Coating

Dickson Mary PhD University of California, Irvine

Corneal afflictions are the second leading cause of blindness worldwide. When a corneal transplant is unavailable or contraindicated, an artificial cornea device is the only chance to save sight. Bacterial or fungal biofilm buildup on artificial cornea devices can lead to serious complications, including the need for systemic antibiotic treatment and even explantation. As a result, much emphasis has been placed on antiadhesion chemical coatings and antibiotic leeching coatings. These methods are not long-lasting, and microorganisms can eventually circumvent these measures. Thus, I have developed a surface topographical antimicrobial coating. Various surface structures including rough surfaces, superhydrophobic surfaces, and the natural surfaces of insects' wings and sharks' skin are promising antibiofilm candidates; however, none meet the criteria necessary for implementation on the surface of an artificial cornea device.

In this thesis I (1) developed scalable fabrication protocols for a library of biomimetic nanostructure polymer surfaces, (2) assessed the potential theses for poly(methyl methacrylate) (PMMA) nanopillars to kill or prevent formation of biofilm by Escherichia coli bacteria and species of Pseudomonas and Staphylococcus bacteria and improved upon a proposed mechanism for the rupture of Gram-negative bacterial cell walls, (3) developed a scalable commercially viable method for producing antibacterial nanopillars on a curved PMMA artificial cornea device, and (4) developed scalable fabrication protocols for implantation of antibacterial nanopatterned surfaces on the surfaces of thermoplastic polyurethane materials, commonly used in catheter tubings. This project constitutes a first step towards fabrication of the first entirely PMMA artificial cornea device.

The major finding of this work is that by precisely controlling the topography of a polymer surface at the nanoscale, we can kill adherent bacteria and prevent biofilm formation of certain pathogenic bacteria, without the use of any chemical antibiotic agents. Such nanotopographic coatings can be applied to implantable polymer medical devices with scalable commercializable processes, and may deter or delay biofilm formation, potentially improving patient outcomes. This thesis also opens the door for adaptation of antibacterial nanopillared surfaces for other applications, including other medical devices, marine applications, and environmental surfaces.

Comment: Aging results in the accumulation of molecular and cellular damage throughout the body and an ensuing loss of tissue function that leads eventually to degenerative disease. Although the ideal therapeutic approach to such deterioration will generally involve the direct repair of said damage, or at least the active exchange of the affected tissue for a bioidentical replacement, ¹⁵ the technologies required to do so are still very much in their infancy. In some cases, however—particularly where the damaged tissue has a more biophysical than biochemical function—a fundamentally more artificial substitute can serve as a valuable interim measure to sustain health, and indeed this basic strategy is widely used in contemporary clinical practice. However, such devices lack the protection against pathogenic infiltration that evolution has built in to healthy tissue; although the incorporation of active antimicrobial agents can improve matters in the short term, the absence of any means to renew those agents as they degrade means that the useful lifetime of such prosthetics rarely approaches the theoretical durability of their component materials. This thesis demonstrates proof-of-principle for an entirely different approach to antimicrobial defense, exploiting the relatively recent discovery that certain nanometer-scale topographic features of apparently arbitrary surfaces can effectively inhibit bacterial growth. (Intriguingly some such patterns are simultaneously highly permissive to the growth of larger and more flexible mammalian cells, a property that will undoubtedly be of value in the context of tissue engineering.) A variety of mechanisms are thought to play a role in this surface-based inhibition. The most direct is simple destabilization of the cell membrane, but another major factor is the inhibition of intercellular communication—which, in turn, blocks the signaling feedback loop required to trigger biofilm formation. The physical nature of nanotopographic effects suggests that the development of resistance to them may be much more difficult than has proven to be the case when antimicrobial agents act at a chemical level. The crucial question that remains to be answered, then, is whether such delicate structures are in practice sufficiently durable in the living body to retain their beneficial effects for longer than can be achieved by more traditional methods.