April 1, 2018 (Vol. 38, No. 7)
The Fittest Animal Models Will Advance Basic Research, Drug Testing, and Clinical Studies
Dramatic environmental change imposes new challenges. It also creates new opportunities, opening niches and potentially triggering a period of adaptive radiation, a rapid diversification of new forms. Although adaptive radiation is usually understood to be a process that occurs in nature, it may also occur in the laboratory—the environment in which animal models are developed.
The environment for animal models is being reshaped by technologies such as next-generation sequencing (NGS) and CRISPR/Cas genome editing. Whereas NGS exposes connections between genotype and phenotype, CRISPR/Cas9 makes it possible to exploit these connections and generate animal models that better recapitulate human diseases, even complex multifactorial diseases.
Many other factors are driving model development and creating, in effect, a new animal model ecology. It sounds almost too complicated. That’s why GEN has sought the guidance of experts. According to these experts, the laboratory, as a controlled environment, is just the place to generate animal models in a comprehensible way.
GEN: How will advances in next-generation sequencing and genome-editing techniques affect the market for animal models?
Dr. Morse: CRISPR/Cas9 technology has increased the speed of genome editing while driving down costs and allowing genomes to be altered in a wider range of species. With this technology, mutations known to exist in disease-associated genes of human patients can be inserted into homologous genes present in animal models, making it possible to study the effects of these mutations under controlled conditions.
We anticipate that new animal models will be engineered with greater precision, and that they will advance basic research and clinical studies. These new models should also support innovation in agricultural, vector control, disease modeling, and other applications.
Dr. Li: Next-generation sequencing (NGS) has succeeded traditional sequencing to become the standard profiling method, replacing competing technologies. NGS expedites and reduces the cost of genome-wide sequencing. It enables the rapid comprehensive genomic profiling of animal tumor models, which greatly enhances model classification and can lead to better model selection and the discovery of prognostic and predictive biomarkers. Together with NGS, recent advances in gene editing (particularly CRISPR) have greatly accelerated and simplified genome editing, creating novel disease models via germline, pluripotent stem cell, and somatic cell engineering.
Dr. Jelcick: We believe that both NGS and new techniques such as CRISPR/Cas9 have enhanced the animal model market’s potential. Recently, we have witnessed a surge in demand for new animal models, many of which had been specified based on clinical discoveries made through NGS. We have also seen that when CRISPR/Cas9 is used in model development, turnaround times and costs can be reduced.
For example, Cyagen Biosciences is deploying a sperm bank for which 10,000 newly established knockout alleles were introduced using CRISPR/Cas9 technology, a feat that required cooperation across continents. This resource could help researchers to validate their hypotheses in a timely manner.
Ms. Wildt: These advances have improved our ability to develop a broad assortment of custom animal models, including models that incorporate genetic predispositions for specific study objectives. Custom animal models, or precision animal models, have greater success in translational medicine. They can help bridge the gap between basic research discoveries and human disease applications.
For example, researchers can grow human tumors in immunodeficient mice models and trial different chemotherapies depending on the tumor genotype during preclinical research. The general market for some precision models could be quite large; however, some models are so specific to a particular set of phenotypes that there may be limited applications.
Although NGS is becoming more affordable, it does not answer every question facing developers of custom models. For example, NGS may not indicate how these models can strike a delicate balance between two capabilities: addressing specific research needs and adapting to different applications. Being overly specific (and insufficiently adaptable) is incompatible with economic viability. So, it’s very important that we recognize our customers’ needs and stay current with emerging research trends.
Dr. Forbes: The advancement of NGS may further facilitate its clinical application and its ability to establish genome-wide associations between genetic alterations and specific diseases. Ultimately, animal models will be indispensable to testing whether the newly discovered genetic alterations underpin the associated diseases.
The advancement of genome-editing technology will undoubtedly make the creation of such animal models faster and easier while also enabling the generation of difficult genomic modifications such as large targeted genomic inserts.
Dr. Soper: These resources will increase the numbers of strains created by researchers and increase the complexity of the models in terms of multiple genetic modifications. Why? Scientists are migrating from models with single-gene knockouts or mutations, and transgenes with unregulated but strong overexpression, to alleles that are “humanized” and/or expressed under native genetic control elements.
Also, the genetics underlying many human diseases are polygenic. Therefore, to properly recreate mouse models of human disease, multiple genetic alterations will be required. These alterations often go beyond changes to protein structure and include changes in gene expression across multiple alleles.
Dr. Seiler: Over the past 30 years, sequencing has paid more attention to the genome’s 30,000 or so coding sequences than to the noncoding regulatory sequences that surround key genes. Now that NGS is becoming more widely available, noncoding regulatory sequences are starting to get their due. At the same time, genome editing is becoming easier, faster, and more accurate thanks to advances in CRISPR/Cas9 technology. Together, these two trends—wide-ranging NGS and high-fidelity CRISPR/Cas9—are bound to have a major impact on the strategies for new model generation.
In addition, NGS is allowing genome-wide association studies to go beyond single-gene disorders. For example, polymorphisms responsible for multifactorial diseases are being identified. When multifactorial diseases are better understood at the genetic level, they can be more readily replicated in animal models.
GEN: The right animal model balances multiple considerations. The most obvious is the need to recapitulate a particular disease or pathway. Please name a few others.
Dr. Morse: Genetic status, health status, and microbiome composition contribute to the etiology of disorders, and all can become confounding factors in animal studies if not properly controlled. But conversely, genetic and microbial variations among strains or study cohorts may provide important insights into disease mechanisms.
The effect of genetic background on disease onset and progression is well documented in the literature. Recent advances in sequencing and bioinformatics have dramatically increased the understanding of the role of microbiota in animal models. Since the reproducibility of research data is imperative, researchers need to be mindful of the genetics, health indicators, and microbiomes associated with their study cohorts.
Dr. Li: Traditionally, cancers were viewed as genetic diseases, and genomic profiles steered model selection. Cancers, however, are also immunological disorders. Consequently, cancer models need to emulate the immune environment, particularly in immuno-oncology research.
Cancer models vary greatly, so it is advisable to test a broad range of models across tumor types, mouse strains, and different degrees of immunity. Disease site and route of administration (such as orthotopic, subcutaneous, and spontaneous) should also be considered. For immuno-oncology studies, vendor selection can be critical. (For example, vendors may provide models that differ with respect to microbiota composition.) In addition, the availability of predictive biomarkers is also vitally important in guiding precision treatments.
Dr. Jelcick: Even with the appropriate gene or pathway, researchers should pay attention to different genetic backgrounds and technical approaches. Over time, spontaneous mutations have arisen in colonies, and the recent literature cites numerous phenotypic differences among strains. In models affected by multiple genes/pathways, genetic background may play a substantial role.
Additionally, while CRISPR/Cas9 approaches are fast and popular, embryonic stem cell–based approaches are still considered the gold standard for creating complex genetically engineered models. Many researchers spend extra time dealing with concerns from reviewers, including but not limited to off-target mutations and genome instability, when manuscripts specify CRISPR-based animal models.
Ms. Wildt: When selecting an animal model for a study, clients usually begin by focusing on gender, age, and model. These factors help determine whether a study will have appropriate statistical power. Health status is also a critical factor. It might determine whether an animal model will violate importation restrictions at the investigator’s institution. In some cases, it might affect the phenotype of the model being studied.
Additionally, some studies will require models that align with special diets or treatments, meet specific enrichment requirements, and/or possess certain surgical modifications. What researchers should pay closer attention to is the genetic background of their models—that is, the degree of heterogeneity (the level of inbreeding)—because this will affect the conclusions of the study.
Dr. Forbes: When an animal model is chosen, it is often necessary to answer questions that involve reproductive biology and the specific species appropriate to the disease or pathway: Does the model breed well? How long is its gestation period? When does it reach sexual maturity?
Other considerations include the relative advantages of using outbred or inbred animals; the complexity of a model’s genetic modifications (for example, whether knockouts are conditional or constitutive); maintenance costs (for example, whether models with similar capabilities differ with respect to manageability); intellectual property status; regulations/ethical issues.
Dr. Soper: One of the biggest challenges in using mouse models of human disease is clinical translation. Inbred strains with uniformity in genetic background are powerful and necessary tools to work out mechanisms of action and the first stages of in vivo preclinical efficacy. However, the human population is genetically diverse. Clinical translation may be improved by creating disease models in multiple inbred strain backgrounds and/or in strains designed to have high genetic diversity due to multiple segregating background alleles across the genome.
Dr. Seiler: The field of animal models and in vivo research is growing increasingly complex. Researchers and vendors need to deal with strain effects, comply with genetic integrity and maintenance programs (whether established in-house or by the vendor), and (most important) understand the impact of the microbiome on research outcomes.
GEN: Have there been or are there still disorders or disease classes that are refractory toward the development of animal models?
Dr. Morse: Studies involving research models are done with prediction in mind, but given the disparities between models and humans, models are not always capable of supporting the study of all clinical manifestations that may be experienced by humans.
This is particularly true with central nervous system diseases, such as Parkinson’s, where animal studies may not always translate well into clinical data. Conditions associated with multiple genetic or environmental factors, such as autism spectrum disorder, are also difficult to model in animals. Given the complexity of these conditions, when using animal models, researchers often focus on just one aspect of the condition in their studies.
Dr. Li: Certain disease models are difficult to establish, and many are unachievable at present. For example, patient-derived xenograft (PDX) models of acute myeloblastic leukemia (AML) or prostate cancer are extremely difficult to create, and models of multiple myeloma remain beyond reach, because of the intrinsic properties of these tumors, including their need to occupy a human niche.
Sometimes samples for model building are simply unavailable. This difficulty may arise, for example, if a PDX model of terminal disease and resistance to a specific drug is required. Humanization of mouse models with human immunity has yet to demonstrate full human immune functions, hindering much immuno-oncology, infectious diseases, and immunology research.
Dr. Jelcick: In principle, all disorders can be mimicked by animal models to some extent. Indeed, animal models in therapeutic areas have been increasingly criticized for their limited ability to predict efficacy, safety, or toxicity in humans.
Additionally, we have found researchers to have limited access to certain animal models based on surgery, xenografts, or germ-free environments. However, rather than dismiss animal models, researchers should consider whether existing animal models could be improved, or whether new and more relevant animal models could be developed.
Ms. Wildt: The research community has long doubted the predictive validity of animal models of certain psychiatric, neurodegenerative, and neurological disorders. Some scientists believe that animal models are poorly suited to the study of central nervous system disorders that affect the higher cognitive abilities, which are unique to the human species. Animal models of autism or Huntington’s disease, for example, may fail to inspire confidence.
In addition, mouse models have historically had difficulty replicating some types of cancers, such as liver or ovarian cancer. Such models may do a poor job of reproducing the tumor microenvironment or representing how processes such as tumor growth and metastasis occur in humans.
Dr. Forbes: Rare diseases are undersupplied in terms of animal models. Whether that is indicative of funding constraints, too few researchers specializing in rare diseases, or poor target identification remains to be seen. Amyotrophic lateral sclerosis (ALS) is a good example. The current animal models for ALS have not performed well, and new genetic models of the disease are needed.
Now that NGS is more accessible, it can be of more use in establishing the genetic causes of rare diseases. The work of aligning genotypes and phenotypes in rare diseases can be taken even farther if animal models are used to support validation efforts. The ultimate goal for Horizon Discovery and its partners is to improve clinical outcomes.
Dr. Soper: The modeling of all aspects of human cardiovascular disease along with the associated metabolic disorders of obesity and type 2 diabetes in a single mouse strain remains a challenge. Many strains have been created that successfully model specific components of this complex disease, but a single strain with obesity, hypercholesterolemia, and/or hyperglycemia, as well as true cardiovascular plaque formation with true arteriosclerosis remains elusive.
Dr. Seiler: In preclinical research, animal models often face physical and biological restrictions. For example, animal models pose unique challenges for testing immuno-oncology therapies. Traditional models with a functional murine immune system may be incapable of evaluating established clinical therapies, many of which are fully humanized monoclonal antibodies. Such antibodies recognize only human cell targets.
Interspecies variation poses significant constraints on testing combinations of therapies. For example, a clinical moiety cannot effectively return to a rodent system and still recognize the mouse immune cells. Humanized immune system models have brought researchers closer to overcoming these constraints, but challenges remain in fully reconstituting the complete immunologic repertoire and function of the human immune system in rodents.
Animal models also have difficulty recapitulating multifactorial diseases. Even the multifactorial diseases that have been illuminated by genome-wide association studies have yet to be adequately recapitulated.
Porcine Models Ready to Take Flight
The unique opportunities presented by porcine models inspire translational scientists, including scientists who study musculoskeletal regenerative medicine and develop biomechanical applications. For example, a team of scientists affiliated with North Carolina State University and the University of North Carolina at Chapel Hill recently described the advantages of porcine models in Tissue Engineering Part C: Methods, a peer-reviewed journal from Mary Ann Liebert, Inc.
In the November 2017 issue, senior author Matthew B. Fisher, Ph.D., and colleagues published a review entitled, “Rise of the Pigs: Utilization of the Porcine Model to Study Musculoskeletal Biomechanics and Tissue Engineering during Skeletal Growth.” The review discussed interspecies comparative studies, highlighted biological considerations such as skeletal growth, and presented challenges and emerging opportunities for porcine models.
“Both a variety of strains and a range of genetic haplotypes within strains are available through commercial vendors,” the article detailed. “In addition, several closed herds of specialized strains and crossbreeds are maintained through academic research laboratories. These strains have been used to study graft remodeling following [anterior cruciate ligament] reconstruction, macular degeneration, and intestinal epithelial stem cell–driven regeneration in the porcine intestinal epithelium.
“Of particular note is the ability to create genetically modified pigs. Such animals can be used as genetic tools similar to those in rodent models, but in a large animal model, allowing more relevant physiology and surgical procedures similar to humans.”
Custom pigs, the authors continued, exist that can express green fluorescent protein. The authors also noted that immunocompromised and “humanized” animals could be created to tackle issues surrounding allotransplantation or xenotransplantation. Recent work, the authors maintained, has suggested that genetically modified “designer” pigs could be raised to produce organs with reduced potential for immune rejection for human transplantation.
“Xenografts are already utilized for rotator cuff repair,” the authors pointed out, and some are advocating their use as an alternative to human autografts and allografts, although acceptable results using these grafts are still unclear.