Why it’s crucial to understand the limitations of in vivo preclinical models in immuno-oncology drug development

Drug developers are uncovering a wealth of new therapeutic opportunities in immuno-oncology (IO). However, bringing new treatments to market is proving to be particularly challenging in this field. A major reason for this is the extreme difficulty in predicting clinical success using data generated in early development. This means that many companies lose out by investing heavily in therapies that ultimately fail at a late stage.

But why is it so difficult to predict clinical success in IO? A key issue is that all available in vivo models have limitations, which can make it incredibly difficult to translate preclinical safety and efficacy data into reliable predictions of clinical effects. Here, we explain why in vivo modelling is so challenging in IO, present the most commonly used models, and explain how a thorough immunological understanding of these models could improve your chances of bringing new IO therapies to patients.

To learn how to navigate the risks in immuno-oncology drug development, download this free eBook: “Immuno-oncology: Five Tactics to enhance your Drug Development Strategy.”

Why is in vivo modelling so challenging in immuno-oncology drug development?

Since the murine immune system is so well-characterised, most IO drug developers use mouse models for in vivo studies. However, there are numerous differences between the murine and human immune systems, which makes it hard to gain translatable insights during preclinical studies. While you can mitigate this problem to some extent by using surrogate biologics to target the mouse homologs, the results from these studies won’t always reflect your drug’s effects in humans.

To complicate matters further, it’s also difficult to model the growth and structure of human tumours. For example, while tumours will arise spontaneously in human patients, certain mouse models require them to be induced artificially. Unfortunately, since the way in which a tumour develops can affect its interaction with the immune system, data gained from these models won’t be fully relevant to the clinical situation. What’s more, in many models the histological architecture may not be representative of that found in humans, which can again limit the physiological relevance of your results.

Taken together, these challenges mean that many developers struggle to use preclinical models to fully reflect the human tumour micro-environment (TME). Unfortunately, this is a major drawback in IO, as the effects of IO agents tend to be highly context dependent. This is because the intricate interaction between the growing tumour and the developing immune response is continually evolving, meaning that the importance of individual signalling pathways can rise and fall over time. As such, the effects of drugs targeting these pathways will depend heavily on the prevailing tissue conditions.

Since the efficacy of IO drugs is so dependent on the biological environment, it’s critical to be sure that your preclinical models can adequately mimic the clinical situation for your agent. To choose the most appropriate models, it’s essential to understand the advantages and limitations of commonly available options.

The most common mouse models in immuno-oncology drug development

Syngeneic mouse models (SMMs)

SMMs are generated by introducing murine cancer cell lines into immunocompetent mice. These are the most common mouse models in IO, as they provide a convenient way of testing the activity of your drug on an intact immune system. However, SMMs don’t fully reflect the human TME, as tumours arise artificially from a genetically homogenous population of cancer cells in these models, and therefore often grow very rapidly. As such, it can be difficult to translate the results of SMM studies into reliable clinical predictions.

Genetically engineered mouse models (GEMMs)

Thanks to advances in genetic engineering techniques, a range of GEMMs are now available. These incorporate one or more mutations found in human tumours, which can promote oncogene expression or inactivate tumour suppressor genes. A key advantage of GEMMs is that tumours arise spontaneously, providing a good model of the developing TME in relation to a competent immune system. However, since these mice are predisposed to cancer, it’s not unusual for multiple tumours to develop simultaneously. This can potentially overwhelm the immune response and make it difficult to interpret your data.

Patient-derived xenografts (PDXs)

Unlike SMMs and GEMMs, PDX models have the advantage of preserving the structural complexity of the human tumour. In these models, biopsy material from solid human tumours is implanted into immunodeficient mice. However, while the histological characteristics of the tumours reflect the human situation, the compromised immune status of the mice is not physiologically relevant. This means that these models can’t provide an accurate model of the complex interactions between immune and tumour cells in the human TME.

Humanised mouse models (HMMs)

HMMs have generated a surge of interest in IO, as they could provide a more accurate reflection of the interaction between a human tumour and the human immune system. Like PDX mice, these models employ human tumour biopsy material, but they also incorporate human immune cells. As such, they provide promising options to study mode of action and efficacy. However, despite the excitement generated by HMMs, they still have significant limitations. In particular, graft-versus-host disease (GvHD)—a condition in which the transplanted immune cells react against the host’s tissue—can be a major confounding factor. Since this usually develops within a few weeks, you’ll only have a limited time window in which you can use your HMM to test drug efficacy. What’s more, GvHD is hard to distinguish from immune-mediated toxicity, so these models aren’t always reliable for predicting toxicity in humans.

How understanding the limitations of in vivo models can boost your success in immuno-oncology drug development

To fully understand the advantages, limitations and applications of different mouse models, it’s essential to have in-depth immunological expertise. Accessing the right expertise will allow you to boost your chances of IO drug development success in two key ways:

  1. You can choose the most appropriate models to use as part of your preclinical studies. By understanding how the immunological environment in each model is likely to differ from the human TME, and knowing how this applies to your agent, you’ll be able to choose the best model to use for each study. This means that you’ll generate the most physiologically relevant data to help you predict clinical success, so you can better establish whether you should progress to expensive clinical trials.
  2. You’ll be able to understand the limitations of the data you produce with these models. Essentially, if you’re aware of the weaknesses of the in vivo models you use, you can determine how best to supplement your information with in vitro data from human cells. In this way, you’ll build up a comprehensive data package with the best likelihood of predicting clinical effects.

Ultimately, with a collection of well-chosen in vivo studies coupled with carefully selected supporting data, you’ll be well-equipped to avoid the costly risk of failure in clinical trials. Furthermore, you’ll also reduce the likelihood of prematurely discontinuing a promising candidate.

While understanding the available preclinical models is vital in IO drug development, you’ll also have to overcome many other significant challenges in this field. If you’d like to learn more about how to navigate the complexities of IO drug development, download our free eBook, “Immuno-oncology: Five Tactics to enhance your Drug Development Strategy.”

Download eBook