Nature: unraveling the lethal cytokine storm behind CAR-T toxicity and a potential preventive role for existing hypertension drugs

Humanity's war against cancer has entered the immune level. The emergence of many immunotherapy approaches, including CAR-T therapy, has filled us with confidence about "curing cancer". But they also have an urgent problem to solve: severe side effects. On one hand, immunotherapy activates the body's immune system to fight cancer cells; on the other hand, the massive secretion of cytokines can very easily trigger a cytokine storm (CRS). Fever, low blood pressure and heart problems are considered mild; in severe cases, CRS can cause multi-organ failure and even threaten life. At present, we can only rely on the treatment experience of clinical teams to control CRS, and scientists are still continuously searching for simpler and more feasible methods.

That method is coming soon! This week's Nature published a result from a Johns Hopkins University team [1]. The researchers discovered that cytokine release is related to catecholamines such as adrenaline, and limiting their synthesis can effectively control the occurrence of CRS! In multiple mouse models of cancer immunotherapy, this approach reduced cytokine production and increased survival, without affecting the treatment effect.

Even more exciting is that the substances used by the researchers, such as metyrosine and prazosin, have already been approved by the FDA for the treatment of hypertension, so clinical translation is very close. If the same results can be achieved in clinical research, CAR-T therapy will surely become even more powerful, which would be a tremendous achievement!

Catecholamines are like an amplifier before the cytokine storm arrives. Before this study, the team had been researching a new oncolytic bacterial therapy. They selected a bacterium called Clostridium novyi, an anaerobe; the researchers wanted to use its anaerobic nature to track the hypoxic environment of tumors and destroy tumors through spores. The idea was good, but in practice this therapy, named C.novyi-NT [2], encountered a huge problem. The amount of bacteria used had to vary with the tumor burden: whenever there was a large tumor, a large number of spores had to be used. On one side were tumors collapsing under the action of the spores; on the other side was sepsis caused by the large number of bacteria. The cytokine levels in the experimental animals soared, and the animals often died within a few days; antibiotics were of no help.

This was fatal for a new therapy! The researchers began to look for ways to suppress cytokines. Antibodies against various cytokines were tried one after another, and the anti-inflammatory dexamethasone was also used, but none worked. The researchers decided to start by modifying the bacteria, thinking that perhaps transferring genes for some anti-inflammatory proteins could reduce the biological toxicity at its root.

This idea proved feasible. After trying a series of anti-inflammatory proteins, the researchers found that a protein called atrial natriuretic peptide (ANP) could lower cytokine levels without affecting the oncolytic effect and improve the survival of mice.

Whether endogenous ANP or injected ANP, both could protect mice from CRS. ANP is a small protein secreted by cardiomyocytes that maintains electrolyte balance and is a commonly used blood pressure indicator. Previous studies have also proven that ANP has anti-inflammatory properties and can reduce lipopolysaccharide (LPS)-induced cytokine release [3]. The protective effect of ANP was obvious. In mice injected with ANP-modified bacteria, plasma ANP levels were 2-4 times those of the control group, 80% of the mice survived after treatment, and tumors completely dissolved in 84% of them, whereas the control group was entirely wiped out. Directly injecting ANP into mice receiving C.novyi-NT treatment also worked well: 75% of mice survived, with complete tumor dissolution in 77%.

At the same time, these mice with more ANP had milder tissue damage and inflammation, fewer immune cells infiltrating unrelated tissues, and lower cytokine levels. The experimental results were consistent in two cancer models: colon cancer and malignant glioma. Interestingly, their levels of catecholamines such as adrenaline, noradrenaline and dopamine were also lower.

Previously, scientists believed that the anti-inflammatory effect of ANP was related to the NF-κB pathway, but after testing inhibitors of related proteins, the team found that this was not the case at all, so they naturally turned their attention to catecholamines. After all, previous studies had confirmed that macrophages, the main producers of cytokines, can both secrete catecholamines and respond to catecholamines after bacterial stimulation, which in turn stimulates cytokine production [4]. The researchers first tested several catecholamines separately and found that adrenaline was the culprit. In LPS-induced inflammatory mice, injecting adrenaline led to worsening of the disease and increased mortality, and levels of cytokines such as IL-6 and TNF-α all rose. ANP, on the other hand, works by limiting catecholamine synthesis.

Then things became much simpler. The key rate-limiting enzyme of catecholamine synthesis, tyrosine hydroxylase (TH), has long had an antagonist: alpha-methyltyrosine (MTR). Drug antagonism was then applied. After experiments in several infection animal models, the results were also very encouraging: MTR did provide good protection against bacterial infection-induced CRS.

It is not surprising that cytokines dropped, and adrenaline and noradrenaline also dropped; MTR and ANP could both control LPS-induced CRS. But the story does not end here. Although similar in nature, whether CRS caused by biological products and CRS caused by bacterial infection can benefit in the same way remains unclear. Here the researchers selected two immunotherapies: one was the CD3 antibody, the first monoclonal antibody OKT3, and the other was the much-anticipated CAR-T. Using the CD3 antibody clearly increased catecholamine synthesis, while MTR could offset most of the increase in synthesis and the rise in cytokine levels, and mouse mortality dropped significantly.

The effect on CAR-T therapy was also good. Both in vitro and in vivo, MTR could reduce the catecholamine elevation and cytokine levels brought by CD19-CAR-T, and ANP could achieve similar effects. In addition, the researchers also tested the impact of MTR on treatment efficacy. The experimental results showed that, for different tumor burdens, MTR did not affect the therapeutic effect of CAR-T.

The researchers started treatment at different times after transplantation to simulate different tumor burdens, showing that MTR did not make the therapeutic effect of CAR-T worse. The accompanying commentary published in the same issue of Nature concluded [5] that catecholamines are like a self-amplifying signal releaser at the onset of the cytokine storm, but the details of this loop still need further exploration. For example, how does the activation of immune cells lead to the increase of catecholamine levels? How do catecholamines promote cytokine production? How does ANP suppress catecholamine production? In addition, which of the various adrenergic receptors matters most is also a question that needs to be clarified. At a time when immunotherapy is becoming more and more accessible, this study is undoubtedly good news. The alpha-adrenergic receptor blocker prazosin, proven effective in the study, is a classic antihypertensive drug, and MTR can be used to treat hypertension in patients with pheochromocytoma.

We hope that these clinically validated drugs will also shine in the field of immunotherapy. (December 2018) [Source: This article is an original article by GeekHeal (WeChat official account: geekheal_com), originally published in March 2017.]