You have most surely seen an image of the iconic lab mouse, a pristinely white rodent with dark red eyes standing on its hindlegs while looking quizzically up at the camera. Less certain is what the mouse is used for, but at some level we imagine the mouse as a miniature human. Scientists use mice and other animal models to unravel the complexities of aging, perception, and memory. The insights gained from this work are then used to develop better treatments for breast cancer, back pain, and neurodegenerative disorders like Alzheimer's disease, Parkinson's disease, and the many other afflictions that plague our loved ones – and perhaps even ourselves. The focus of this article is to shine some light on the question – How valid are our models? – and the imperative to continuously improve upon them if we are to conduct the best science possible.

I am a scientist and work with mice, albeit not the white-furred, red-eyed version so often seen in movies. The mice I work with are the more traditional staple of a mouse laboratory known as C57BL/6 or colloquially as 'black 6'. All around the world, scientific researchers use black 6 to investigate a multitude of scientific questions (Fig. 1 - 3). One of the greatest achievements of modern science was to sequence all 3.2 billion nucleotides of the human genome and shortly thereafter geneticists moved on to decode our furry friend black 6's genetic code.

Figure 1. C57BL/6 or 'black 6‘, one of the commonly used mouse strains.

Molecular biology techniques allow scientists to add and remove most any of the ∼23,000 genes from the mouse genome to answer questions such as, 'Does this gene and subsequent protein influence neuronal death?' or 'If we target this enzyme can we reduce the chance of heart attack for at-risk patients?' Utilizing genetically modified mice, we can gain a glimpse into the complex workings of a living, breathing, moving organism. It is the ultimate goal of many scientists to translate the knowledge gained at their sterile lab bench to the ailing patient's bedside. This begs the question, ‘Does the knowledge gained from our rodent companions result in better treatments for grandma's breast cancer, Uncle George’s chronic back pain, and mom’s worsening memory?’ It seems like it should, right?

When developing a new drug, pharmaceutical companies run several animal tests to ensure both efficacy and safety before proceeding to human trials. Simply put, it is not common to test a drug in humans without it first entering the mouse. From this point of view, the mouse is clearly integral to drug development. However, it is important to realize that animal models do not always reliably predict the outcome in humans as tragically demonstrated in a recent Phase I human clinical trial conducted in France. Six people were hospitalized and one participant died from neuronal complications resulting from a new pain and anxiety medication even after rats, mice, dogs, and monkeys were previously treated with the medication without problems (though there were allegations that dogs suffered unreported side effects) [1].

Figure 2. The mouse and human brains have clear morphological differences (images not to scale).
Figure 3. A closer look at the difference between human and mouse brains (images not to scale).

Needless to say, working with biologically active compounds is dangerous business, but this work needs to be done if we as a society are to improve upon the current standard of treatment. On the other hand, mouse models of breast cancer have produced results of varying relevance for humans, [2] but despite missteps along the way, this work has resulted in improvements in human treatment. Mouse models of BRCA mutations (a breast cancer associated gene) have shed new light on mechanistic insights and potential treatments for patients. For example, there are currently inhibitors of DNA repair mechanisms (PARP inhibitors) in Phase III clinical trials for the treatment of breast cancer (see Talazoparib and Veliparib). Researchers believe that by inhibiting the DNA repair pathway, cancers cells die faster because these cells tend to divide faster than their healthy counterparts. Research into another breast cancer associated protein known as HER-2 has resulted in the production and approval of the monoclonal antibody Trastuzumab (an antibody is a molecule that binds to a protein). Scientists discovered that HER-2 is elevated in a subset of breast cancer patients’ tumors and then subsequently used laboratory mice to develop the first antibodies against the HER-2 protein [3]. This medication transformed a devastating diagnosis into one that is not without hope and is listed on the World Health Organization’s (WHO) model list of essential medicines.

These types of medicines demonstrate that mouse models of human disease are not without value. Pain research is another area that has gained new insight with the help of mice. Clearly different individuals have diverse pain tolerances. We all know that coworker who shrieks in agony from a paper cut and the contrasting stoic who hardly flinches after dropping an 80 kg machine on their foot. Relatedly in mice, there is interesting research led by Prof. Jeff Mogil of McGill University on the variability in pain tolerance across mouse strains [4]. Simply stated, on one end of the spectrum there are mouse strains with a high pain tolerance whose tolerance is further strengthened by analgesics, while on the other end exist mouse strains with a low pain tolerance regardless of the medications given. This work caught the field by surprise and brought a long-held belief into question, namely whether one type of mouse is representative of all the others (Fig. 4). Jeff Mogil’s comment about his own study speaks volumes, "Frankly, we just published a paper that scared the hell out of me.” In the field of pain research, researchers need to be cautious when generalizing the results from one mouse strain to another, let alone to humans. This same group of researchers later demonstrated that when male scientists worked with mice or rats a reduced pain response was observed. However, this effect was not observed when females handled the rodents. Most interestingly, the reduced pain response occurred again when females wore male T-shirts revealing that the sex of an experimenter can influence pain responses in rodents [5]. It appears scientific truth is not only obscured by different types of mice, but also by the scientists who interact with them.

Figure 4. Pain sensitivity varies across mouse strains at an order of 1.2 to 54-fold

The days of debating nature versus nurture are behind us. It is both the interaction of genes and the external environment that produce a given outcome. But what about our internal environment? A recent paper highlighted another aspect of animal research, namely the sterility of it all. Laboratory mice are housed in pristinely clean environments, especially when compared to their wild counterparts. Recently, there has been growing interest in the microbiome and the role that gut bacteria play in health and disease. In a study investigating exactly this, researchers colonized laboratory mice with gut bacteria from that of wild mice and subsequently demonstrated that laboratory mice harboring the ‘wild bacteria’ were more resistant to viral infection and tumorigenesis models [6]. To demonstrate the impact of this experiment, let’s talk numbers. Of the normal laboratory mice, 17% were alive 18 days after intranasal infection with the influenza virus, while 92% of the bacterial carrying lab mice were still squeaking around! Clearly, our animal models are kept under artificial conditions that appear to have a strong influence on disease susceptibility, at least in our rodent friends.

Neurology is another field where treatments have not much improved despite the multimillions of dollars, euros, pounds, and other currencies invested in this research area. Alzheimer’s disease has frustrated many researchers as it still lacks any treatment despite many transgenic mouse models developed to study this disease (although one could argue that a true mouse model of Alzheimer’s does not exist). The most accepted explanation of Alzheimer’s disease is the buildup of a protein in brain cells known as amyloid beta (Fig. 5) (See Jay Rasumssen's Alzheimer's research article). Unfortunately, the majority of attempted Alzheimer’s treatments have failed and none have made it to market. Currently, there is not an approved treatment that works on amyloid beta, but there are Phase III clinical trials (the final phase) in process to test an antibody treatment that reduces the amount of amyloid beta. This study should be completed by 2022.

Figure 5. Amyloid plaques are commonly found in Alzheimer’s patient neural tissue

Mice do allow for better science and have helped advance our knowledge of biological processes. In the end though, mice are not humans and their utility as guideposts along the road to developing new medications for humans can be misleading. Our research is only as good as our tools. Moreover, it is not only mouse models that require rigorous scrutiny. The validity of cell culture models and antibodies have been questioned as well. Recently, researchers demonstrated the poignancy of antibody validation by bringing 20 years of research into question [7]. Their paper validated 13 antibodies for a breast cancer associated protein (oestrogen receptor β) and concluded only one antibody to be truly specific. Furthermore, there was no expression detected in cancerous human breast tissue (nor in healthy tissue) meaning that the association of this protein with breast cancer is tenuous as best. Researchers have also found that cell culture lines are contaminated with other cell lines, meaning sometimes researchers believe they are working with lung cells when they are in fact working with liver cells, or human cells that are actually rat cells [8]. Proper characterization of our tools is clearly an issue not to be taken lightly.

The situation is not black and white, and I do not imagine an easy fix for such difficult and multidimensional problems. What I hope to have accomplished in this short article is to highlight the need to question the validity of our models and the imperative to continuously improve upon them if we are to conduct the best science possible. This means rigorous characterization of our tools and a readiness to return to the drawing board when the data from our models do not provide predictive value in our own species. Only then can we conduct the best science and further unravel the true underpinnings of grandma's breast cancer, Uncle George’s chronic back pain, and mom’s worsening memory.


Michael Paolillo is currently a doctoral student at the Interfakultäres Institut für Biochemie (IFIB) in the lab of Prof. Dr. Robert Feil.

Citations:

  1. theguardian.com/science/2016/mar/07/french-drug-trial-man-dead-expert-report-unprecidented-reaction. Accessed 06 Nov 2017
  2. Dine J, Deng CX. 2. Cancer Metastasis Rev. 2013;32:25-37
  3. Harries M, Smith I. The development and clinical use of trastuzumab (Herceptin). Endocr Relat Cancer. 2002;9:75-85
  4. William R et. al. Transgenic Studies of Pain and Analgesia: Mutation or Background Genotype? Journal of Pharmacology and Experimental Therapeutics. 2001;297:467-473
  5. Sorge RE, et. al. Olfactory exposure to males, including men, causes stress and related analgesia in rodents. Nat Methods. 2014;11:629-32.
  6. Rosshart SP, et al. Wild Mouse Gut Microbiota Promotes Host Fitness and Improves Disease Resistance. Cell. 2017;pii: S0092-8674(17)31065-6
  7. Andersson S. Insufficient antibody validation challenges oestrogen receptor beta research. Nat Commun. 2017;8:15840
  8. Horbach SPJM, Halffman W. The ghosts of HeLa: How cell line misidentification contaminates the scientific literature. PLoS One. 2017 Oct 12;12(10):e0186281

Image sources:

Title image - Jax.org
Figure 1 - Charles River Laboratories International
Figure 2 - Elizabeth Atkison, Washington University in St. L
Figure 3 - Cryan JF, Holmeands A. The ascent of mouse: advances in modelling human depression anxiety. Nat Rev Drug Discov. 2005. Sep;4:775-90
Figure 4 - Mogil JS et. al. Heritability of nociception I: responses of 11 inbred mouse strains on 12 measures of nociception. Pain. 1999;80:67-82
Figure 5 - BrightFocus Foundation
Photo of four mice - Stanton Short / Jennifer L. Torrance Jax.org
Photo of cell culture, antibodies, and mouse - labdepotinc.com; shutterstock.com; Jax.org

Categories: Science perspectives

Leave a Reply

Your email address will not be published. Required fields are marked *