Druggable Targets in Cytokine Release Syndromes

In this issue of Clinical Cancer Research, Leclercq-Cohen and colleagues experimentally explore targetable mechanisms to prevent or mitigate the effects of cytokine release syndrome (CRS) as induced by a CD20-CD3 T-cell engager (1). Systemic inflammatory responses mediated by cytokines often compromise the safety of immunotherapy based on chimeric antigen receptor (CAR) T cells and T-cell engagers.

Inflammation when local constitutes an important mechanism to deal with infections and tissue injury. However, if inflammation becomes systemic, it can be life threatening as observed in septic shock (2). Indeed, overt T-cell or macrophage activation can lead to multiorgan failure and death. In the case of T cells, acute polyclonal activation, for instance with bacterial superantigens, can also be life threatening as observed in cases of toxic shock syndrome (3).

Harnessing T-cell activation and redirecting it to tumor antigens has prompted an ongoing therapeutic revolution for hematologic B-cell malignancies. Two main approaches are followed: CD3 T-cell engagers and CAR T cells redirected to surface antigens of malignant B cells (4). In both instances, an overt activation of T cells occurs as a result of eliciting or mimicking TCR-CD3 signaling. This is an on-target mechanism of action that intends to redirect to tumor cells the cytotoxic mechanisms with which certain subsets of T cells, mainly CD8⁺, are endowed. Some of the ample arrays of cytokines that T cells can produce upon activation may help in tumor cell destruction, especially IFNγ (5). However, this comes at a cost because various of these soluble mediators, perhaps acting in concert, can result in systemic damage involving high fever, low blood pressure due to overwhelming vasodilation, respiratory failure, and generalized endothelial damage (6). Systemic inflammation may also underlie acute neurologic syndromes termed ICANs (immune effector cell-associated neurotoxicity syndrome; ref. 7) and, if sustained, may result in lasting macrophage activation conditions such hemophagocytic syndromes.

In this field, there was a landmark discovery when the group of S. Grupp and C. June first used CAR T cells for pediatric acute B-cell leukemia. These authors reported that an anti-IL6R mAb (tocilizumab) and anti-TNF (etanercept) saved the acute critical condition of a child who eventually went on to develop a durable complete response (8). Since then, IL6R blockade is used to treat severe CRS in addition to steroids and even to prevent this condition. However, IL6/IL6R is not the only cytokine whose circulating levels are raised in CRS and hence multiple pathways could be interfered with to mitigate this condition, which is most often transient and dependent on the burden of disease. In this setting, disease burden is proportional to the abundance of antigen and thus is associated with the number of T cells simultaneously engaged in the organism. In fact, tumor debulking, if feasible, is advisable to mitigate this problem.

IL6, IL1β, TNFα, IL18, IL2, IFNγ, and various chemokines are part of a long list of likely pathogenic substances. Establishing a pathogenic hierarchy of these mediators makes sense to define the most suitable targets to act upon provided that we already have either clinical approvals or experimental agents for some of these inflammatory mediators. These facts offer obvious opportunities to repurpose cytokine-neutralizing drugs.

Leclercq-Cohen and colleagues (1) used human whole blood exposure to a CD20 T-cell engager (CD20-TCB) to explore the cytokines that are induced and which leukocytes produce them. They discovered that T cells release TNFα, IFNγ, IL2, IL8, and CCL4 which therefore are the primary mediators (Fig. 1), that in turn act on monocytes, neutrophils, dendritic cells, and natural killer cells. As a result, there is a secondary production of TNFα, IL8 and the induction of IL6, IL1β and other chemokines. The power of the scRNA-seq tool is used to provide unequivocally conclusive results attributing cytokine production to specific cell subsets in very detailed time course experiments with the only caveat that transcripts are measured rather than the proteins, which in some instances would need posttranslational processing. Moreover, costudying the cellular distribution of the transcripts for the corresponding cytokine receptors, important information is inferred regarding autocrine effects and cytokine loops (Fig. 1).

Of much interest and novelty are also the observations on the role played by neutrophils in cytokine sensing and production, which strongly suggest a role for polymorphonuclear leukocytes in the pathogenesis of CRS that needs to be further studied. Using human umbilical vein endothelial cell cultures with conditioned media from CD20-TCB–stimulated leukocytes, the authors established that endothelial cells contribute to IL6 and IL1β release, while also upregulating ICAM-1 and VCAM adhesion molecules that would conceivably in turn mediate adhesion and extravasation of various activated leukocyte types.

Next, the authors used immunodeficient NOD SCID mice pre-engrafted with human CD34⁺ stem cells and then subcutaneously challenged with a human B-cell lymphoma cell line. In these mice, human T cells are functional and can be induced to reject tumors if the CD20-TCB bispecific construct is infused. In this mouse model, several agents neutralizing cytokines were tested including dexamethasone, IL6R blockade, IL1R blockade, and NLRP3 inflammasome inhibition. It is important to acknowledge that the readout was cytokine levels in serum rather than the severity of the sickness in mice that cannot be modeled. In this setting, none of these agents ablated antitumor activity suggesting that these targets can be tested against human CRS. However, it is noteworthy that neutralization of TNFα with adalimumab in vivo seems to slightly worsen tumor outcome at least in this xenografted lymphoma model (1), arguing against the use of this type of agent at least as preventive premedications.

It would also be nice to know whether the neutralization of IFNγ or the inhibition of JAK kinases are suitable targets that may not compromise the antitumor effects if the corresponding agents are used transiently. Because approved neutralizing agents are also available for IFNγ and JAK kinases, the knowledge that they would not spoil the CD20-TCB antitumor activity would be of much interest and their effects could be tested in the same experimental setting of humanized mice.

All in all, CRSs might be therapeutically addressed acting on several inflammatory mediators with prominent pathogenic roles in the origin and self-amplification mechanisms of this acute condition. Perhaps, we will be able to define personalized biomarkers such as the serum cytokine profile of each patient, that will tell us which neutralizing agent or which combination of agents is to be used in each particular case. Preclinical knowledge that the antitumor outcome would not be compromised by some of these neutralizing agents is of great value to guide clinical development. Hence, multiple umbrellas seem to be available for cytokine storms.

I. Matos reports grants from ESMO and BeiGene; and personal fees from Gilead, Servier, and MSD outside the submitted work. I. Melero reports grants and personal fees from BMS, Roche, AstraZeneca, Genmab, F-Star, and Boehringer Ingelheim; and personal fees from PharmaMar, Pierre Fabre, Merus, BioLineRx, BioNTech, Amunix, Boston Therapeutics, Pieris, Servier, and Bright Peaks outside the submitted work. No disclosures were reported by the other authors.

The team is supported by the Spanish Ministry of Economy and Competitiveness and Spanish Ministry of Research (MINECO SAF2014-52361-R and SAF 2017-83267-C2-1R MCIN/AEI/10.13039/501100011033/ and by FEDER Una manera de hacer Europa and PID2020-112892RB-100, PID2020-113174-RA-100 MCIN/AEI/10.13039/501100011033, Cancer Research Institute under the CRI-CLIP, Instituto de Salud Carlos III (AC16/00015) and European Funds for Regional Development (EFRD) under the TRANSCAN-2 Programme, and Gobierno de Navarra Proyecto LINTERNA Ref: 0011-1411, and Mark Foundation. Our group is also supported by a grant from the Scientific Foundation of the Spanish Association Against Cancer (AECC) jointly with Instituto de Salud Carlos III (AC22/00026) and European Union – NextGeneration EU funds under the Plan de Recuperación, Transformación y Resiliencia (PRTR).