Abstract
Animals have been closely observed by humans for at least 17 000 years to gain critical knowledge for human and later animal survival. Routine scientific observations of animals as human surrogates began in the late 19th century driven by increases in new compounds resulting from synthetic chemistry and requiring characterization for potential therapeutic utility and safety. Statistics collected by the United States Department of Agriculture’s Animal and Plant Health Inspection Service and United Kingdom Home Office show that animal usage in biomedical research and teaching activities peaked after the mid-20th century and thereafter fell precipitously until the early 21st century, when annual increases (in the UK) were again observed, this time driven by expansion of genetically modified animal technologies. The statistics also show a dramatic transfer of research burden in the 20th and 21st centuries away from traditional larger and more publicly sensitive species (dogs, cats, non-human primates, etc) towards smaller, less publicly sensitive mice, rats, and fish. These data show that new technology can produce multi-faceted outcomes to reduce and/or to increase annual animal usage and to redistribute species burden in biomedical research. From these data, it is estimated that annual total vertebrate animal usage in biomedical research and teaching in the United States was 15 to 25 million per year during 2001–2018. Finally, whereas identification and incorporation of non-animal alternatives are products of, but not an integral component of, the animal research cycle, they replace further use of animals for specific research and product development purposes and create their own scientific research cycles, but are not necessarily a substitute for animals or humans for discovery, acquisition, and application of new (eg, previously unknown and/or unsuspected) knowledge critical to further advance human and veterinary medicine and global species survival.
INTRODUCTION
Human history is intertwined with the animal species sharing their environments, and throughout that history humans have gained critical knowledge for their survival, education, and health from observing both humans and animals with increasingly acumen and more powerful and sophisticated tools and techniques. The Lascaux Cave near the village of Montignac in southwestern France contains over 600 exquisite images of contemporary animals dating back 17 000 years (Figure 1). Given this observational acumen, the same humans—when butchering these animals—must have noticed gross anatomical features, including bones, body cavities, musculature, and internal organs similar to their own, thereby establishing original knowledge of comparative anatomy. Over 10 000 years later, the Ebbers Papyrus (1536 BCE) describes more than 800 preparations from animal, mineral, and vegetable sources used to treat Egyptian maladies.1 The first drug treatise from China (1st century CE) names 365 drugs with useful properties and parsed into 3 groups: considered non-toxic and supposedly useful, mildly toxic but with some efficacy to treat maladies, and toxic and effective to cure specific illnesses.2 Neither Egyptian nor Chinese document describes who and/or how these remedies were discovered/developed, but observation and perhaps intentional “rudimentary experimentation” must have been primary tools. Later Egyptian, Greek, Arab, and Chinese literatures detail the use of humans and animals to explore and develop knowledge of anatomy and rudimentary surgical procedures as well as potions and theriacs—some of which continue to be used in traditional medicine and are being investigated for new medical treatments today. In Western experience, Aristotle (4th century BCE) is acknowledged as the first to perform “experiments” or to have added manipulation to observation of living animals. An early example of such scientific manipulation was by Galen (129–216 CE); in his “De theriaca ad Pisonem,” Galen describes assigning roosters to control and theriac treatments, exposing both groups to venomous snakes, and observing that whereas control roosters died immediately after a snake bite, theriac-treated roosters survived.3 Ibn Zuhr (12th century CE) added manipulation when testing surgical procedures on animals before applying them to human patients.4,5 Andreas Vesalius (1514–1564), John Hunter (1728–1793), and others demonstrated animal observation and manipulation by conducting animal dissections, surgical procedures, and rudimentary experiments as academic and public demonstrations through the 19th century. In reviewing this literature, Kinter and DeGeorge concluded that humans were the species most frequently used in ancient scientific inquiries, there was sparse evidence of animal use in medicinal discovery, and no evidence of animal use to evaluate safety; they concluded that lack of use of animals as human surrogates in medical investigations was due in part to sufficient availability and tractability of human subjects and insufficient concepts of allometric scaling (eg, dose/unit body weight or body surface area).1
The relations between human observation and manipulation of animals sharing their environments. See text for details. Cave painting image courtesy of Prof Saxx—Public Domain, https://commons.wikimedia.org/w/index.php?curid=2846254.
Closer to the modern era, advances in technology and economic growth fueled advances in scientific knowledge and welfare of both humans and animals.6 Commenting on the human condition prior to the 18th century, English philosopher Thomas Hobbes (1588–1679) observed that human life was “solitary, poor, nasty, brutish, and short,”7 and certainly the lives of animals were no better. The First Industrial Revolution (circa 1760–1840) unleashed unprecedented economic growth, technology, and improvements in living standards across Europe and North America. In addition to observations regarding basic anatomy and rudimentary surgical procedures, this period heralded advances in scientific knowledge, resulting in rudimentary explorations of the functions of body organs (medical physiology) using observation and increasingly sophisticated manipulations in both humans and animals: the monaural stethoscope and auscultation introduced in 1816 by René Laenne (1781–1826)8; William Harvey’s (1578–1657) demonstrations of functions of the heart and circulation using both humans and animals9; Stephen Hales’ (1677–1761) use of a column of water to measure arterial blood pressure in a horse10; Robert Hooke’s (1635–1703) observations of oxygenation of the blood within the dog lung5,11; and Antoine Lavoisier’s (1743–1794) observation using a guinea pig that respiration was a type of combustion.5,12 Also notable for this period were the first social efforts to recognize and promote the more humane treatment of animals, including the founding of the British Society for the Prevention of Cruelty to Animals in Britain in 1824.13
The Second Industrial Revolution (circa 1860–1914) brought new generations of technology and instrumentation with which to measure and record physiological functions, and new disciplines, including analytical, organic, and medicinal chemistry. August Waller (1856–1922) developed the first practical technology to record electrocardiograms using surface electrodes, conducting demonstrations using his pet bulldog dog “Jimmy”; arguably, Jimmy is the “poster dog” for use of canines in cardiovascular research. Willem Einthoven (1860–1927) showed that the electrocardiogram was useful in diagnoses heart conditions.14 Adolf von Baeyer (1835–1917) synthetized malonylurea in 1864, initiating a century of synthetic barbiturate analgesic therapies,15,16 and Paul Ehrlich (1854–1915) and Sahachiro Hata (1873–1938) synthesized arsenoxide and arsenobenzene derivatives of aminophenyl arsenic acid, making over 600 compounds before discovering dioxy-diamino-arsenobenzoldihydrochloride (“606”, arsphenamine), the first effective treatment for syphilis.17 Notably, Ehrlich observed that most of the arsenicals he experimented with were toxic to mice and rabbits and therefore unsuitable for human testing, marking an early example of product safety evaluation using animals. Kinter and DeGeorge concluded that the explosion of new synthetic chemistry in the late 19th and early 20th centuries necessitated large-scale pharmacological testing using animals as human surrogates for identification and characterization of potentially useful pharmacological properties and determination of tolerability, thereby founding the modern disciplines of experimental pharmacology and toxicology.1 Several professional organizations soon surfaced, including the American Physiological Society, founded in 1887; the American Society for Pharmacology and Experimental Therapeutics, founded in 1908 by John Jacob Abel of Johns Hopkins University; and the British Pharmacological Society in 1931. Notably, the first organizations dedicated to breeding and supplying animals explicitly for biomedical research were formed during this period, including Jackson Laboratories (1929), Harlan Laboratories (1931), Marshall BioResources (1939), and Charles River Laboratories (1947).
Additional significant events of this period included the 1902 US Biologics Control Act, the 1937 US Pure Food and Drug Act, and the 1962 Kefauver Amendment, heralding a new era of regulation of medical products for humans and routine animal safety testing of new drugs as a prerequisite for US clinical trials and marketing approvals. Notably, Marshall BioResources initiated raising beagle dogs for research in 1962 in response to the changes imposed by the US FDA for non-clinical testing of candidate drugs.18 Other innovations occurring in this era included the electrical strain gauge by Edward E. Simmons and Arthur C. Ruge in 1938, enabling development of modern blood pressure and tissue contraction transducers.19
The period following World War II thru the end of the 20th century, sometimes referred to as the Third Industrial Revolution,6 witnessed the largest and most sustained advances in global economic growth and living standards, driven by some of the most remarkable advances and implementations of scientific knowledge and new technologies in human history.1 During this period, in vivo animal and later in vitro bioassays using animal tissues proliferated and became the research “backbone” advancing the disciplines of physiology, pharmacology, pathology, endocrinology, toxicology, and others for exploration of biological properties, characterization and categorization of pharmacological and toxicological properties of new chemistry, and selections of new drugs for clinical trials and marketing authorizations. By the early 1980s, there were hundreds of individual bioassays for pharmacological activities using dogs, cats, rabbits, guinea pigs, swine, and smaller rodent species.20 Technological advances in implantable catheters, sensors, and pumps, transducers, telemetry, and imaging technologies greatly expanded observational and manipulation assessments in animals. Additionally, this period witnessed development of conscious animal models acclimated to laboratory procedures to eliminate overlay of artifacts associated with stress and anesthesia.21,22 Further, miniaturization of technologies permitted collection of experimental endpoints in rats and mice that were previously limited to larger species.23,24 These advances were associated with the founding of the Society of Toxicology (1961), the British Society of Toxicology (1979), and the Safety Pharmacology Society (2000). However, despite these advances, uncertainty always remained as to whether and how results of animal studies would “predict” (or translate) to humans (eg, finding the “perfect animal model”).
Beginning in the 1970s and early 1980s, the primacy of observation/manipulation of animals in scientific research was challenged and progressively replaced by new concepts and technologies arising from new disciplines of cellular and molecular biology. The resulting paradigm shift—from phenomenological to mechanism-based characterizations of new chemistry and drug discovery based on specific interactions with human molecular targets (eg target-directed biology)—revolutionized both basic science and applied drug discovery disciplines. The impact was rapid replacement of the previous in vivo and in vitro animal model backbone of biomedical research with new cellular and molecular tools. This paradigm shift continues to evolve today with the use of stem cell models in cardiovascular risk assessment and in silico efforts to target specific mechanisms responsible for pharmacologic and toxic responses.
In a curious twist of history, a return to phenomenologically based animal observation and manipulation in the late 20th century was a result of perhaps the most transforming event of this period, Watson and Crick’s deciphering of the genetic code in the 1950s, leading to creation of genetically modified (GA) vertebrate organisms in the 1980s.25 Following the seminal discoveries of Watson and Crick, sequencing of the human genome in 2001 and later sequencing of the genomes of animal species commonly used in research by the beginning of the 21st century revealed 80% or greater commonality (eg, common orthologs) of genetic sequences.26 The genetic homologies confirmed prior centuries’ observational and phenomenological categorizations of relatedness of specific animal species and humans, for example, Linnaean taxonomy,27 with chimpanzees and other non-human primates (NHPs) being the most similar. Collectively, the genomic sequences and prior practical experience confirm that no animal species is a “perfect model” for humans and supports the concept that GA technologies can produce better models. To paraphrase Professor George Box: “All animal models are wrong, but some are useful.”28
Initially GA models (usually rodents) were applied to discovery and validation of new disease targets,29 across disease areas from HTT protein and mouse models of Huntington’s disease to oncology targets. Later, GA models were successfully applied to reduce non-clinical safety testing in normal healthy animals with the adoption of GA P53 and rasH2 transgenic mouse models.30–32 GA models have generally replaced the conduct of 2-year murine bioassays in non-clinical carcinogenicity testing with many fewer animals used. Likewise, functional defects observed in GA models also inform on potential safety risks from therapeutic interventions and thereby reduce animal safety testing.33 Such approaches have been adopted into recent international regulatory guidance.34 The impact has been the replacement of normal phenomenologically based animal, tissue, and cellular models to explore new biology and/or predict human responses with GA equivalents designed for the specific scientific purpose at hand.
This same period has witnessed the most pronounced global advances in animal welfare to date, beginning with the continuation of the concepts of Replacement, Refinement and Reduction (the 3Rs)35 and the adoption of the first standards for appropriate animal welfare in biomedical research36 and agriculture;37 the United States Laboratory Animal Welfare Act in 1966, subsequently amended and expanded in scope; and culminating with efforts to replace animal experimentation with non-animal alternatives and technologies.38 The period also witnessed global standardization for animal studies incorporating the principles of 3Rs and the attempt to eliminate unnecessary and/or duplicative studies (although sometimes at the expense of increasing animal usage in individual studies thru increasing the numbers and size of study groups), supporting global clinical trials and new product marketing authorizations.39
Documentation of Animal Usage in Biomedical Research
Prior to the 20th century, there was no formal accounting of the numbers of animals used in scientific research activities. In the US regulations promulgated to implement the US Animal Welfare Act of 1966, the US Department of Agriculture (USDA) required annual reporting of numbers of specific species (cats, dogs, rabbits, guinea pigs, hamsters, NHPs, and farm species) used by every academic, industrial, and government organizations engaged in scientific research, beginning in 1973 (Figure 2). The USDA data show that usage of these species peaked in the mid-1980s at over 2 million animals per year and, with 1 exception, the annual usage of all reported species has fallen by at least 50% since. Notably, these statistics are coincident with the paradigm shift away from phenomenological-based animal observation and manipulation to mechanism-based to cellular and molecular testing noted above as well as the international standardization of study requirements supporting clinical trials and new product registrations also noted above. The only exception in USDA statistics are NHPs, the least-used group of species in the 1970s but whose usage has approximately doubled since the 1980s. That growth has been driven by the perception based on phylogenetics and later comparative genomics that NHP species are more likely to mimic human responses, growth in development of new biopharmaceuticals and vaccines wherein NHPs are most appropriate species for non-clinical discovery and product development,40 and implementation of primate GA models (eg, marmosets) in discovery research.41
USDA/APHIS annual animal research data 1973–2019. Vertical axis: total numbers animals reported per year. Horizontal axis: years. (A) Total all reported species. (B) Totals individual reported species. The category of farm animals is a composite including pigs, goats, and sheep (1991–2007), after which pigs (including miniature pigs) and sheep were reported separately (2008–2018). 1973–2007 and 2019 data courtesy of USDA/APHIS staff; 2008–2018 data from USDA website.51
A major deficiency of the USDA statistics is that only certain mammalian species (notably excluding mice and rats) are included, and all other vertebrate species, for example, birds, fish, and amphibians, are excluded. The National Survey of Laboratory Animal Facilities and Resources: Fiscal Year 1978 contains similar usage statistics for USDA-reported species and adds 13 413 813 mice, 4 358 766 rats, and 450 352 birds for total of 19 956 388 animals used in laboratory facilities. Health Designs Inc. (Rochester, NY) conducted a Survey and Estimates of Laboratory Animal Use in the United States in 1983, estimating in addition to USDA species 8 500 000 mice, 3 700 000 rats, 100 000 birds, 500 000 amphibians, and 4 000 000 fish for total of 18 581 875 animals in laboratory facilities.42 Another recently published extrapolation concludes that between 2017 and 2018, over 100 000 000 mice and rats were used for research purposes in the United States.43 These estimates suggest that the large decreases in usage of USDA-reported species (Figure 2) were more than offset by increases in non-reported species in the late 20th and early 21st centuries.
The UK Home Office began reporting annual statistics on scientific procedures using living vertebrate animals (mammals, birds, reptiles, amphibians, and fish) in 1945, when total procedures were approximately 1 million per year (Figure 3). The numbers of procedures performed annually in the United Kingdom increased steadily, peaked in the late 1960s to early 1970s at approximately 6 million per year, and then declined precipitously to approximately 3 million per year by 2001. Since 2001, mice, rats, and fish have accounted for 85% or more of total procedures, with all other individual species ≤8%. Sensitive species (dogs, cats, horses, and NHPs) have collectively accounted for <1% of procedures in the United Kingdom since 2001.44
United Kingdom Home Office annual statistics of scientific procedures on living animals 1945–2019. Statistics are compiled from UK Home Office annual reports 2001–2019. Vertical axis: live animal procedures (millions/year). Horizontal axis: years. The disconnect between the 2 upper graphics (Experiments 1876 and 1986 Acts and Procedures 1986 Act, light and dark blue) reflect a recording difference occurring in 1987.44 The middle graphics reflect total procedures filtered for creation, breeding, and usage of GA animals (purple) and normal animals (green) 2001–2019. The lower graphic (red) reflects UK usage of specific species approximate to those reported by USDA/APHIS 2001–2019 (see Figure 1).
After decades of decline, in the early 21st century, UK annual procedures began to increase, driven entirely by increased usage of mice, rats, and fish to generate GA animals (Figure 3). A comparable increase in usage in the United States is not reflected in the USDA annual statistics (Figure 2) because these do not include mice, fish, and rats. However, when the UK annual statistics are filtered for non-GA procedures, the same total downward trends are comparable in both the United States and United Kingdom. Additional insights provided by the UK statistics include: (1) use of genetically normal animals has continued to decrease in the 21st century; (2) use of GAs has increased and has accounted for approximately 50% of all reported procedures since 2012; and (3) the vast majority of reported GA procedures are for creation and breeding of GAs and the residual used in research programs. Enthusiasm for GA procedures is driven in part by the potential for inclusion of human response elements to improve efficacy, activity, and human translation of surrogate models for both therapeutic and safety endpoints, as noted above. The UK statistics demonstrate the substantial impact of a novel set of unanticipated transgenic and gene-editing technologies on usage of the traditional observation/manipulation animal research cycle, marking a fourth and current “Industrial Revolution” of genetic modification technologies in scientific research, product development, medicine, and animal welfare beginning in the first decade of the 21st century.
Filtering the 2001–2019 UK statistics for procedures with approximately the same species as reported in the 2001–2019 USDA statistics shows that UK procedures using these species has ranged from 70 000 to 280 000/year (Figure 3), whereas US usage has ranged from 740 000 to 1 240 000 per year (Figure 2). UK usage of these species was 7% to 16% of US usage during 2001–2013 and 27% to 36% during 2014–2018, reflecting an increase in UK and decrease in US usage during the latter interval. Between 2001 and 2019, procedures conducted in the United Kingdom using USDA-reported species ranged from approximately 2% to 7% of total reported procedures. Assuming that patterns and distributions of species usages in the United States and United Kingdom have been similar over this period, and USDA-reported species usage is approximately 5% of all vertebrate usage based on UK species procedure ratios during the same period, the estimated US annual usage (2001–2018) of all vertebrate species including fish has ranged from approximately 15 000 000 to 25 000 000/year (Figure 4), roughly similar to previous estimates of the 1978 ILAR and 1983 Health Designs reports (anon. 1991). The difference (fivefold) between Carbone’s 2017–2018 estimate and that reported in this paper is that the Carbone estimate is based on a survey of mouse and rat usage by 16 large US institutions,43 whereas the estimate in this paper is based on usage by all institutions of all vertebrate species by another country that is similar to the United States in scientific research activities.
Estimated US scientific usage of vertebrate species 2001–2019. Vertical axis: total estimated vertebrates/year. Horizontal axis: years. Analysis of UK Home Office statistics showing that specific species reported by USDA/APHIS 2001–2019 constitute approximately 5% of total UK scientific procedures on living animals during the period. See text for details.
Inferences to be drawn from the USDA and UK Home Office research animal statistics include:
Animals used in scientific research and product development increased progressively during the so-called third Industrial Revolution period, peaking between approximately 1970–1980 and likely reflecting the highest scientific annual use rates in US and UK history. Peak annual usage of all species in the United Kingdom was approximately 5 500 000 animals/year; peak annual usage for US AWA-reported species (eg excluding mice, rats, and all non-mammals) was approximately 2 000 000 animals/year, and undoubtedly many times more had all vertebrate species been included.
Animals used in scientific research and product development during the 20th century were predominantly normal or spontaneous mutants (eg, genetically hypertensive rats, immunodeficient mice) and used in primary physiological, pharmacological, and toxicological animal and animal tissue bioassays. Beginning in the late 1970s/early 1980s, primary animal and animal tissue bioassays were rapidly replaced by new molecular technologies that permitted target screening, target binding and activation, second messenger pathway activation, cellular pathways analyses, etc, or abandoned. The impact of the so-called “molecular revolution” is reflected in the dramatic decline (≥50%) in annual living animal procedures (UK) and usage of reported species (US) statistics into the 21st century.
Beginning in the early 21st century, the downward trend in the UK Home Office statistics unexpectedly reversed and over the following years rebounded to levels nearly as high as peak level procedures per year in the prior century. The rebound is almost entirely accounted for by mice, fish, and rats used for creating, breeding, and usage of GA animals. No rebound is apparent in the USDA statistics because the USDA-reported species have not generally been used for GA purposes. However, it is reasonable to surmise a similar pattern regarding GA species has occurred in the United States and that current levels of animal usage (all species) in the United States are also approaching/exceeding peak levels during the 20th century. This unanticipated increase in annual animal use in scientific research was the result of the uptake of unanticipated new technologies (eg, transgenics, gene editing/CRISPR-cas9). Notably, Russell and Birch predicted that advances in technology would create opportunities to enhance animal welfare thru the 3Rs,35 and the USDA and UK Home Office statistics for the 20th century support their prediction. However, Russell and Birch apparently did not envision a potential impact of new and unanticipated technologies to rapidly generate new demands for animal usage in scientific research, for example, GAs. The UK Home Office statistics for the 21st century demonstrate the impact of new and unanticipated technologies driving new demand for animal usage in scientific research.
During the years of peak animal procedures (United Kingdom) and usage (United States) in the 20th century, the animal burden was shared by both rodent and non-rodent mammalian species, including mice, rats, dogs, cats, guinea pigs, rabbits, hamsters, and farm animals. In the current “Fourth Era,” the burden has shifted almost entirely towards species that reproduce rapidly, produce high numbers of offspring, and are subject to easy genetic manipulation, for example, mice, rats, and fish. In 2019 the UK Home Office reported 86% of experimental procedures used mice, fish, or rats; no other species accounted for >8%; dogs, cats, horses, and NHPs collectively <1%; and dogs and NHP usage was primarily for regulatory product testing. A similar conclusion was recently found for US dog and NHP usage using 2017 USDA statistics.40 The USDA and UK Home Office statistics for the 21st century demonstrate the potential impact of new and unanticipated technologies to change the species distribution burden in scientific research. In addition, the shift in the discovery research paradigm away from animal models has enabled proliferation of both academic and venture capital biotechnology organizations to leverage new science for new potential therapies and to engage contract research organization for non-clinical development studies, contributing to increased animal usage and further underscoring the need to collect annual usage statistics for all species.
Non-animal Alternatives
Throughout history, humans—using their powers of observation and manipulation—have acquired critical knowledge leading to novel applications and thereby allowing survival and success of individuals and the species. Scientific research is just the most recent iteration of this observation/acquisition/application cycle (Figure 5), and development of the scientific method-applied rigor, process, and new technologies has only facilitated generation of new knowledge leading to new applications. The iterative process also generates opportunities for the 3Rs (with another less-sentient species or tissues), opportunities for necessary animal research to achieve its goals.35 However, replacement of animals and animal tissues with non-animal alternatives is highly sought and considered by some the ultimate goal. Non-animal replacements are applications derived from knowledge acquired through observation/manipulation using animals, and their utility is to replace for defined and specific purposes the necessity of further observations/manipulations using animals. Non-animal replacements offer greater efficiency, sensitivity, accuracy, cycle-time and productivity, and reduced cost for their specific purposes and should be identified, extracted, developed, and implemented whenever possible. However, non-animal replacements are products of, and not integral components of, the animal observation cycle, and although they will generate new knowledge within their alternative cycles, in the near term they are unlikely to replace the primacy of new observations using animals (including humans) to generate new previously unknown and unsuspected knowledge leading to new novel applications for clinical and veterinary healthcare and species survival. Consider the below examples of the animal observation/acquisition/application cycle leading to implementation of non-animal alternatives.
Depicts the modern relations between human observation and manipulation of animals for acquisition of new knowledge supporting new applications for benefit of clinical and veterinary medicine and species survival. See text for details.
Diabetes and Insulin
In 1889 Oskar Minkowski and Joseph von Mering performed manipulations and observed that pancreatectomized dogs developed diabetes mellitus whereas pancreatic duct ligation alone did not, thereby acquiring new knowledge of a hitherto unknown pancreatic function and establishing the dog as the primary animal model in early diabetes research.45,46 Continuing the cycle, in 1920 Fredrick Banting and his student Charles Best prepared dog pancreas extracts and observed that upon injection into pancreatectomized dogs, the symptoms of diabetes mellitus (Type I) were relieved, thereby acquiring new knowledge that their pancreatic extract contained a hitherto unknown active substance or activity (they named “insletin”) critical for physiological glucose regulation. In 1921, Banting, Best, and their Department Chair John McLeod observed that the new knowledge acquired with canine insletin (renamed insulin) was reproduced in a Type I diabetic patient, thereby establishing cross-species translation of the therapeutic potential of insulin in Type I diabetes management. In retrospect, the dog model was fortuitous for human translation because diabetes Type I blood glucose regulation and pathophysiology are highly conserved between the 2 species and there is only 1 amino acid difference between canine and human insulins.46,47
Recognizing that dog-sourced insulin was not logistically viable as a therapeutic for human diabetes, Best and David Scott developed procedures by which insulin could be isolated from commercial porcine and bovine sources,48 thereby invoking 3Rs Reduction and Refinement principles in animal research by replacing use of a sensitive species with tissues obtained as a bi-product of agricultural processes. Eli Lilly & Co. then applied the new knowledge and commercialized bovine and porcine insulin extracts from 1923 to the end of the 20th century for human and veterinary diabetes therapy. In the 1980–1990s, recombinant DNA technology was applied for commercial manufacture of human insulin, creating a non-animal alternative (Replacement) for animal-sourced insulin in diabetes therapy.
McLeod and Banting (but not Best or Scott) were awarded the Nobel Prize in 1923 for their discovery of insulin—perhaps the most consequential Prize of the 20th century for its impact on diabetes management in human and veterinary medicine and for launching a century of advances in chemistry; the commercialization of insulin and other peptide hormones; and biology, pathophysiology, and therapeutics of diabetes and other endocrine disorders. The Prize has also been consequential many of the non-animal replacements for traditional animal sourcing of peptides and exploration (eg insulin, vasopressin, oxytocin, thyroid hormone, etc), identification and investigation of cellular and molecular diabetes mechanisms, and for identification of spontaneous, diet/nutrition, chemical and surgical induction, and genetically-altered (eg transgenic, knock-in, knock-out, etc.) animal models, diagnostic assays (eg glucose clamp, ELIZA), and therapeutic advances (eg feed-back controlled insulin pumps, synthetic hormone antagonists).47,49,50
In summary, while the original discoveries of Banting, Best, Scott, and McLeod—following on observations of predecessors including Minkowski, von Mering, and others—laid the foundations for non–animal-sourced synthetic human insulin for diabetes management, they also laid the foundations for countless other observation cycles initially using animals for discovery, characterization, and therapeutic utilization of other peptide hormones and leading to new non-animal replacements for the detection, quantification, and production of said hormones and analogs for therapeutic applications. The research and discovery cycle is exemplary of so many current areas of biomedical research.
Endotoxin Detection
Parenterally administered drug formulations are routinely tested for presence of endotoxin, which can be inadvertently introduced through bacterial contamination during manufacture and packaging processes. Endotoxins (lipopolysaccharide complex [LPS], LPS A) are part of the outer membrane of the bacterial cell wall of Gram-negative pathogens such as Escherichia coli, Salmonella, Shigella, Pseudomonas, Neisseria, Haemophilus influenzae, Bordetella pertussis, and Vibrio cholerae. In humans, LPS binds to a lipid binding protein and triggers the signaling cascade for macrophage/endothelial cells to secrete pro-inflammatory cytokines, including IL-1 (“endogenous pyrogen”). IL-1 stimulates the hypothalamus to increase body temperature. The traditional method used to detect endotoxin is the rabbit pyrogen test (RPT), in which groups of rabbits are restrained and rectal probes inserted before intravenous test article administration; the primary experimental endpoint is rectal temperature change. Any agent that stimulates production of IL-1 or any other mechanism that stimulates hypothalamus to increase body temperature will be detected in the RPT. As late as the late 20th century, CROs specialized in performing RPTs and the RPT was probably second only to the Draize eye irritation test in generating pressure for a non-animal alternative method.
A candidate alternative method was limulus amoebocyte lysate (LAL), an in vitro test arising from observations that circulating blood cells in Atlantic horseshoe crabs (Limulus polyphemus) produce enzymes that bind and inactivate LPS from invading bacteria. The LAL test was accepted by the FDA as an alternative test method for endotoxin/LPS in 1983. It could be argued that the LAL test is not a true non-animal alternative because LAL is harvested from crabs, which originally did not survive the procedure; however, methods were quickly developed permitting the return of the crabs to the ocean unharmed following their blood donations. Today, LAL has almost completely replaced RPT and detects the presence of LPS by a mechanism entirely different from the RPT.
In the late 1990s, 2 authors (L.K. and D.J.) provided non-clinical support for the clinical development of a novel high-volume parenteral X-ray contrast agent. The formulation included a nanoparticulate encapsulated in egg-sourced lecithin-based liposomes. Clinical trials showed that volunteers receiving >100 mg of formulation experienced a febrile reaction, characteristic of LPS exposure. The formulation had been tested using LAL and was negative for LPS. Based on a finding from a non-rodent toxicology study that also suggested a febrile response, we evaluated the formulation using the RPT at a CRO; it was pyrogenic. We tested additional clinical formulations, all pyrogenic. We made a test formulation using the same nanoparticulate and synthetic lecithin liposomes, which tested negative for LPS in the LAL and was not pyrogenic in RPT. We concluded the following: (1) the febrile response observed in volunteers was due to LPS contamination of the test article, (2) the egg-sourced lecithin was the source of the LPS contamination, and (3) the liposome formulation effectively “masked” the LPS so that it was not detected in the LAL test. However, in the RPT, the liposomes were metabolized releasing the LPS, stimulating secretion of pro-inflammatory cytokines, including IL-1, and the pyrogenic response. Mechanisms matter, and tests based on different mechanisms may or may not provide congruent results, not because either test is “wrong” but because the tests measure different endpoints and/or use different conditions.
In summary, although non-animal replacements focusing on a specific mechanism do create their own “micro-cycles” in terms of new knowledge creation, these are separated from larger and more general endpoints and new knowledge opportunities associated with more mechanistically generalized animal research cycles.
CONCLUSIONS
Human ancestors recording their observations of animals as images on cave walls must have experienced a similar sense of accomplishment as modern scientists reporting their findings in peer-review journals or professional meetings; and we can surmise that the relationships between human observation and acquisition of knowledge leading to useful applications were as laborious, time-consuming, non-linear, and influenced by serendipity in ancient times as is for scientific research today. Yet despite that humans have benefited from observing and manipulating animals throughout their history, systematic use of animals as “human surrogates” in fundamental and applied scientific research is a modern phenomenon dating from the 19th century and reaching its peak in the latter half of the 20th century as documented by USDA and UK Home Office annual statistics. The dramatic decrease in annual animal usage for biomedical purposes in the latter 20th century was the result of multiple factors, most importantly discovery and implementation of new cellular- and molecular-based technologies that were more specific, analytic, efficient, productive, and cost-effective than then-current phenomenological-based animal and animal tissue bioassays. However, the UK Home Office annual statistics also show that implementation of new technology can be “multifaceted” and unpredictable. Implementation of new transgenics and gene-editing technologies in the early 21st century have been associated with sizable increases in annual animal usage to levels approaching or exceeding previous peak levels. The increases were largely due to necessary animal usage for creation and production of genetically modified animal constructs. The research burden imposed by GA animals has fallen almost exclusively on mice and fish, and to a lesser extent rats, demonstrating that redistribution of research burden is dependent on the new technology and is inherently unpredictable. Finally, whereas identifying and applying non-animal alternatives is a critical product of the observation/acquisition/application animal research cycle that replaces further use of animals for specific purposes, non-animal alternatives are generally not a substitute for animals or humans for future discovery, acquisition, and application of new knowledge critical to further advance human and veterinary medicine and global species survival.
Acknowledgments
The authors acknowledge the courtesy of Dr. Betty Goldentyer (Deputy Administrator, USDA/ APHIS, Animal Care) and members of the USDA/APHIS staff for providing 1973–2007 and 2019 summary statistics.
Potential conflicts of interest. All authors: No reported conflicts.
References
Anon.
Van der
