This article focuses on how the immune system functions, the challenges that it presents to drug development and why there are such concerns around safety that need to be investigated in preclinical studies. In addition, it covers the role of checkpoint inhibitors as an attractive target for cancer immunotherapy.

The main role of the immune system is to protect the body from infection. It possesses a variety of cell types and systems to effectively identify and remove pathogens before they can cause damage to the immune system and protect the human body from infection.

There are two main arms to human immunity—the innate immune system (IIS) and the adaptive immune system (AIS). Innate responses are the foundation of immunity. All organisms, even bacteria, have innate defense mechanisms (IDM)—a combination of cell populations and soluble factors that work to damage and remove pathogens. They are an ancient system, with levels of conservation of lower order species, and take the form of compliment systems, toll like receptors, granulocyte populations, type I interferons and other cytokines. Innate defense mechanisms are the first wave of immune attack. IDMs have significant drawbacks—potentially causing collateral damage including temperature increase and the killing of human cells as well as pathogens, acute hypersensitivity and even death. They predominantly utilize pattern recognition to identify foreign material, in the form of pathogens, which they attack and try to clear.

However, there are limitations to the innate immune system. Parasitic agents grow more quickly than the IIS can control—with the pathogens overcoming an initial dampening of their immune response and growing back. Moreover, there is only limited amplification of the immune response with an inability to ramp-up beyond their initial impact on pathogens. The innate defense systems also have no memory; with the result that the response time on encountering the same pathogen is precisely identical each time. Consequently, higher organisms have evolved a more sophisticated adaptive immune system, forming the second wave of immune attack. There is some crossover between the two systems with some cell populations, such as NK cells, serving as a bridge between the initial innate response and the subsequent adaptive response.

The adaptive immune system has two arms: the cellular T cell immune response and the humoral antibody B cell response. It includes cells that  specifically target pathogens to minimize collateral damage. The AIS provides a mechanism for immune recognition—clonal selection—that can evolve as rapidly as the parasite and which can amplify up rapidly for efficient pathogen clearance. Most importantly, the immune system has a memory such that specific cell populations are retained after clearance of a pathogen, with the result that it is uniquely primed to remember and attack the same pathogen upon further encounter. Future exposure to the pathogen concerned is subclinical due to rapid, specific recognition and clearance.

Tolerance is vital to ensuring that there is an effective and controlled immune response and serves as an education mechanism for T cells and B cells in the immune system. It involves the concept of self vs non-self. For an adaptive immune system to be effective, it must only target foreign pathogens—minimizing collateral damage by not targeting a human’s own cells—and effectively clearing a pathogen before it can take hold and cause harm. There are two levels to the tolerance mechanism that mediate the education of the cells; central and peripheral.

Central tolerance ensures the selection during hematopoiesis of the best T cells and B cells—ones that meet specific criteria, mediated by particular antigens presenting cell populations in the thymus or bone marrow. The process ensures that cells are deleted that can’t identify and work against specific pathogens or which target self-antigens and consequently damage organs and tissues. Only appropriate cells are released into the periphery.

Though effective, central tolerance is still not perfect and multiple other peripheral control mechanisms exist in lymphoid organs to further refine responses to self and non-self antigens—further limiting collateral damage and ensuring that the immune system reverts to baseline once the pathogen is removed. These include exogenous control mechanisms—regulatory T cells inhibitory cytokines—and endogenous control mechanisms.

For effective immune function, it is critical that the right balance is struck between identifying pathogens and controlling subsequent responses—avoiding any over-reaction that will cause damage to self-tissues.

Cancer and the immune system
Tumors look inherently invisible to the immune system as a consequence of being derived from self. They also actively seek to avoid detection by the immune system by decreasing expression of markers that make them visible. They also utilize control mechanisms to dampen the immune response locally and this includes expression of inhibitory markers, production of suppressive cytotoxins, generation of cell populations with inhibitory properties and induction of anergy or exhaustion in effector T cell populations. The impact is the immune system lacking a prolonged or efficient tumor killing capability. Consequently a major focus of cancer immunotherapy is to increase tumor surveillance by preventing such evasion – tipping the balance towards activation to ramp up the way T cells see tumors and as a result increase their capability to kill them.

There are multiple anti-cancer immunotherapy platforms with differing target pathways. They range from adoptive cell transfer techniques that include TILs, TCR-transfected T-cells and CAR therapy, non-specific immunotherapies such as cytokines (IL-2, IFNa), CD40 agonist mAb and TLR agonists, vaccine strategies, for example, sipuleucel-T, MAGE-3 ASCL and OncoVEX, and immune checkpoint inhibitors, particularly CTLA-4 and PD-1.

Checkpoint Inhibitors
These drugs form one of the most exciting new classes of biologics and are a highly attractive target for cancer immunotherapy. They play a critical role in the maintenance of T cell tolerance. Checkpoint inhibitors (CPI) address a challenge for the immune system—the switching on and switching off of a T cell’s capability to kill tumor cells by signals from other cells of the immune system, namely, antigen presenting cells and the tumors themselves. Signaling via immune inhibitory checkpoint receptors on activated T cells effectively switches off the immune system’s T cells. However, T cells can be switched back on in lymph nodes or the tumor microenvironment by checkpoint inhibitors that target the checkpoint receptors and block them. There are many such drugs in development and others already on the market. Among the better-known checkpoint inhibitors are PD1, CTLA-4, PDL, KIR, IDO1, 4-1BB, OX40, LAG3, B7-H3, CD27, CD70, CD28 and CD30.

Immuno-inhibitory receptors are expressed after normal MHC class I/II adaptive T cell activation. Ligation of CTLA-4 by B7-1/B7-2 or PD1 by PD-L1 or PD- L2 attenuates normal T cell activation. Up-regulation of CTLA-4 and on chronically activated T cells makes them less responsive to antigenic stimulation—something the immune system originally intended these biomarkers to achieve. But equally, the blocking of these targets can restore T cell function and thereby enhance T-cell tumor killing capability.

The one drawback is that as drugs that stimulate the immune system, checkpoints have the potential to promote inflammation and autoimmunity. In particular, too much stimulation runs the risk of losing the immune system’s peripheral control mechanisms to limit immune responses and inducing auto-immune (AI) disease. AI can take the form of systemic disorders or localized diseases and leads to inflammation and the killing of pancreatic islet beta cells in T1D and destruction of the myelin sheath in MS. Multiple mechanisms are implicated in the induction of auto-immune disease leading to the escape of self-reactive T cells into the periphery, loss of or non-functional T cells and molecular mimicry. Mouse models that lack key checkpoint markers have demonstrated AI disease. It is imperative therefore that the potential to cause AI is investigated thoroughly in safety assessment studies in early stage development before a checkpoint inhibitor drug reaches the clinic.

The differing phenotypes of checkpoint-deficient models used in pre-clinical research have demonstrated that each checkpoint marker has a specific pattern of expression resulting in various effects when it is eliminated from the system. Furthermore, though the immune system may still be able to cope with the elimination of one control mechanism, there is potential for additive effects when multiple pathways are blocked. CTLA-4 marker deficiency leads to severe AI disease with the cause of death in mouse models thought to be caused by autoimmune-mediated cardiac failure. Checkpoint target studies using PD-1 knock out mice models demonstrate strain-specific autoimmunity disease later in life but enhanced proliferation of CD8+ T cells and increased cytotoxicity. PD-L1 knock out mice used in such research show increased susceptibility to AI disease but improved APC-mediated activation of T cells and increased CD4+ and CD8+ T cell activation.

Checkpoint inhibitors are showing tremendous promise in the clinic with CTLA-4 and PD-1 blockade significantly increasing patient survival in Phase III melanoma trials. But there is clear evidence of the drugs causing immune-mediated adverse events (AE) in the clinic—highlighting the need to characterize this in non-clinical safety packages. In a Phase III trial on ipilimumab, 77% of patients experienced AE of which 38% were severe including elevated liver enzymes, hepatitis and enterocolitis. Such advanced phenotypes clearly need to be controlled. Another key concern for the industry concerns is the onset of multiple CPI products. Given the AE impact experienced when removing control mechanisms in studies utilizing just single checkpoint inhibitors, scientists are faced with the question of what could be the adverse effects of knocking out multiple pathways. For effective safety assessment, it is critical to understand the effects of altering these signaling pathways, both individually and in combination.

Preclinical safety assessment challenges and strategies
There are a number of challenges which pharmacologists face in a pre-clinical safety assessment. They are tasked with taking the brakes off the immune system—preventing naturally inhibitory signals and activating the immune system in order to trigger the alarm bells on a drug being a high risk product. With most biopharmaceuticals, a delicate balance needs to be struck between therapeutically desired pharmacology and clinical dose limiting “exaggerated pharmacology.” The scientists concerned are trying to interfere with the balance that the immune system wants to maintain and push it instead towards an immuno-stimulated phenotype. The tricky question, which is especially pertinent to checkpoint inhibitors, is determining how far they can push this and at what point do adverse effects caused by the interference outweigh the benefits of obtaining non-clinical safety data designed to help biopharmaceutical companies make informed judgments about drug candidates and, so far as possible, de-risk their pipelines. Immunostimulatory drugs can be severe and life threatening. Though there is currently some scientific understanding of processes such as cytokine release, there is a distinct lack of knowledge of other acute responses including, for example, complement activation in type III hypersensitivity which can cause death and may not be treatable. Hence the need to be very careful during development.

Another challenge is that healthy SPF used in safety assessment testing of checkpoint inhibitors, which have the benefits of being healthy and young, may have relatively clean immune systems that lack or have limited antigen exposure. The target expression of checkpoint molecules in these animals can be very different to humans—animal immune systems are very different to the ones in humans and healthy subjects respond differently to those that are diseased. It is imperative that these factors are taken on board when running such non-clinical toxicology studies.

ICH S6 (R1) stipulates that CPI products must be assessed in pharmacologically relevant animal species. The regulations go beyond just confirming that the product can bind the target and induce relevant pharmacology. Scientists must ensure that the immune status in non-clinical species, compared to the human population, is clearly understood. This includes density of target expression by the immune system, the extent of antigen experience in the animal models—including whether they have been vaccinated or have any activated T cells—and, in clinical studies, whether patients have had chemotherapy and consequently whether they will have a fully functioning immune system at the time that they are treated.

Non-clinical safety combination and single specied studies—carried out over six months into the Yervoy (Ipilimumab) anti-CTLA-4 antibody—produced some notable conclusions.
Adverse findings were seen in combination toxicity and exploratory (TDAR) pharmacology studies but not in the anti-CTLA-4 alone studies. There was a low incidence of immune-meditated toxicities such as colitis, dermatitis and infusion reactions—consistent with the proposed mode of action of CTLA-4 checkpoint inhibitors in maintaining self-tolerance. Overall, the findings in cynomolgus monkeys correlated with findings in humans, although with less frequency and therefore potentially under predicting the findings in the clinic—linking back to the status of the immune system in animal models compared to humans in clinical studies. Blocking CTLA-4 did result in an over stimulation of the T cell compartment with few meaningful changes to the immunophenotype or autoimmune organ pathology (apart from the colitis and dermatitis incidences). Increased TDAR responses were observed as a consequence of the enhanced functionality of CTLA-4, demonstrating the expected pharmacodynamics.

As a top tips overview summary, it is important to have an eye to all the following factors when developing non-clinical safety assessment packages for CPIs:

• Know that there are likely to be autoimmune (AI) consequences in the clinic and aim to generate data that characterizes the full extent of any autoimmunity in clinical studies;
• Consider activation of the immune response;
• Devote time and resources to learning more about the immune system given that it is far from fully understood at present and scientists are not yet able to explain in full when and why adverse reactions will occur or why;
• Assess key tissues and organs for autoimmune effects; and
• Be vigilant to the fact that animal models used have had less exposure to antigens than humans.

In conclusion, the immune system is a complex interaction of multiple cell populations and control mechanisms that provide sufficient sensitivity to identify and eliminate pathogens, while limiting more extensive damage. Scientists have the potential to adjust the extent of an immune response by removing elements of the peripheral tolerance system—effectively taking the brakes off the immune response and amplifying the magnitude and duration of the response and increasing tumor surveillance.

However, on the downside, this also has the potential to increase anti-self responses that can lead to autoimmune diseases. It is important to understand the magnitude and downstream effects of any increases in immune activation when designing non-clinical safety assessment packages, giving consideration to comparative immune function and measuring functional responses prior to clinical studies. Scientists should also be mindful to the relative immune status of animal models used in toxicology studies compared to clinical populations. The biology will be different if the target density of the non-clinical species is different.

It is also important to consider the immune effects that may be mediated by combinations of checkpoint inhibitor products. Layers of redundancy are built into the immune system so that it can survive if one cellular population or pathway is removed. However, it is generally accepted that blocking two CPIs can generate serious immunological problems as research into checkpoint inhibitors like ipilumimab clearly demonstrates. Comparing research findings with existing products and combinations, both clinically and non-clinically, can enhance understanding of immune system response to a drug. It is crucial as well to understand the pathways that are being blocked and their relative contributions to peripheral tolerance. For example, CTLA-4 is constitutively expressed on regulatory T cells and blocking these CPIs will prevent the T cell population from functioning. In the case of other checkpoint markers that are linked with T cell activation, blockage may lead to enhanced or prolonged cell activation, though, they can eventually be brought back under control through mechanisms like regulatory T cells.

Above all, at the core of non-clinical safety assessment is biology – understanding as many pathways as possible, and applying measurement of those pathways, will be hugely informative to the biopharmaceuticals industry moving forwards, especially in fields like checkpoint inhibitors. Scientists must pay particularly strong attention to the risks of potent adaptive immune activation caused by CPIs targeting drugs and should fully embrace the renewed need to better understand the biology through their early stage research and tests. 

---

Greg Bannish is vice president of biopharmaceutical development at Envigo. In this role he is responsible for the development of biopharmaceutical products within the U.S., providing advice within the senior management team, and working with internal and external customers.