Influenza and Cytokine Storm

The occurrence of a “cytokine storm” has been suggested as an explanation for the devastating nature of the 1918 flu and perhaps H5N1. So what is a cytokine storm? From a clinical perspective, a cytokine storm describes an immune system that has over-reacted and is damaging the body, causing failure of multiple organ systems. Ordinarily a cytokine storm is a rare event, which means there are few opportunities to study them, so we do not fully understand how they happen. The term “cytokine storm” is not precisely defined, referring particular kind of uncontrolled immune response. Cytokine storms can happen rapidly and patients who suffer them have high mortality. Because we lack knowlege, we don’t know the best way to treat the condition. Influenza is thought to be one of the rare conditions able to cause a cytokine storm.

The obvious, but important thing to remember: “cytokine storm” is a metaphor. It conjures up images of wild and chaotic weather events where nature’s fury is unleashed in an uncontrolled fashion. That may be one aspect of the phenomenon dubbed cytokine storm, but sometimes the problem with metaphors is that the unrelated analogy (here “storm”) carries with it images and inferences that may or may not be valid. As it turns out there is little to say about the “storm” part of this metaphor and much too much to say about the cytokine part. What follows is a general idea of what can be inferred about cytokine storms.

The immune system
Cytokines are important components of the immune system that act as messages between cells. In this context, the purpose of the immune system is to protect us from microbial infection (viruses, bacteria, parasites, fungi). It is impressively complex, with many moving parts interacting in complicated ways we are just beginning to understand. There are different ways to describe this system. One way is to talk about two major branches, (1) humoral immunity (via antibodies, which are large proteins that circulate in the blood and come from immune cells called B-cells); and (2) cell-mediated immunity, which relies on cell-to-cell contact between the target cell (e.g., a host cell infected by the influenza virus) and one of a set of immune system cells called T-cells. There is also another way to divide the pie: (1) innate immunity, which is the non-specific response to attack not directed against specific targets; and (2) adaptive immunity, which is a learned response that is specific to certain invaders and takes time to develop, but once developed is “remembered.” (Adaptive immunity is the mechanism by which vaccines work.)

Things that the immune system does to protect itself are frequently quite violent, taken on their own. Thus in cell-mediated immunity a common strategy is to kill the infected cell to save the rest of the body. In humoral immunity, the antibodies set up the invading organism for later attack by special killing (cytotoxic) cells or other systems; the antibodies stick to the invader like post-it notes telling the killer cells, “attack this!”. But like a rogue SWAT team, this arsenal of killing mechanisms would run rampant if left unregulated. The system needs ways to turn its aggressive defensive mechanisms off when they have done their job, or risk damage to the body.

So we need rules and regulations to sequence the various actions of the immune system and prevent them from getting out of hand. If you think in terms of a typical police action, you’ll get some of the flavor of this. First they wait for a 911 call, then they send a squad car to investigate, then apply force if certain conditions pertain—but only enough force to get the job done. Then they clean up and process the paper, etc. If these things are done in a completely different order and in addition are unconstrained, there will be trouble. The analogy with police response isn’t perfect, but even in this imperfect analogy if you were to catalog all the different possible actions, signals, responses, dispositions, regulations and much more it would make a fairly complicated system to describe. So it is with the immune system.

The immune system has many different kinds of cells that function in multiple ways in different circumstances. In order to keep things under control there have to be rules, and in order to follow the rules the cells of the immune system have to know what’s going on, i.e., they must communicate and coordinate between themselves and with the rest of the body, just as the police have to coordinate amongst themselves and with the public they serve and protect.

Cytokines are the signals that cells of the immune system use to talk to each other, coordinate their actions, and properly target and limit their violent defense mechanisms. In a proper police action, you stop firing your gun or swinging your baton when it isn’t necessary any longer; and you don’t do either if it isn’t required. Cytokines are like the commands which govern these actions.

At least that’s the way it is supposed to work. We all know it doesn’t always work that way, either in civil society or in the immune system.

Cytokines
Cytokines are soluble hormone-like proteins that signal cell to cell. In that regard they are like the more familiar hormones that signal organ to organ via the blood stream, but usually cytokine signalling is very local, sometimes even self-stimulatory. Some cytokines, however, can travel to distant sites through the blood stream and affect other organs, particularly the nervous system, where they can reset the body’s thermostat to cause a fever, or to the liver, where they cause the synthesis of substances used in fighting infection (such as ‘complement’). There are a large number of these substances now known, and more being described.

The names of cytokines can be confusing. Some cytokines have multiple names, in part because researchers in different labs were not at first aware they were studying the same entity as other researchers. Often the same cytokine does different things in different contexts—just as the word ‘fire’ might mean different things to a police officer and a fireman. Just like a conversation between people, what a signal means to cells depends on context. The same ‘sentence’ ‘spoken’ in cytokines might mean something different at different times of the process or in different circumstances. So it is not surprising that the cytokine story is complicated.

Cytokine storm
One of the many possible effects of cytokines is to summon other immune cells to the site of microbial attack and to activate those cells so that they, too, elaborate cytokines which in turn summon still more cells. This is a positive feedback loop and is ordinarily damped down by other cytokines signalling still other cells to elaborate still more cytokines that put the brakes on the process. Ordinarily this works well. But most regulatory mechanisms can get out of kilter and this can happen in a variety of different ways.

Here is an analogy, not meant to describe what is going on during a so-called “cytokine storm” (because in fact no one knows those details) but to illustrate one kind of easily visualizable dysregulation. When you drive a car down the highway you are continually making adjustments to the steering, accelerator and brakes to keep the car just to the correct side of the center line. You use your eyes as sensors and if you get too far from the center line they send a signal to your hands and feet to make the adjustment to get back to the center. Suppose that you close your eyes and only open them at certain intervals. If you aren’t going too fast and the intervals your eyes are closed aren’t too long, you’ll still be able to stay pretty much on the road. But if your eyes are closed too long you can steer past the center line and in the next interval do the same thing in the other direction, each time making wider and wider swings until you go off the road. It depends on how long your eyes are closed (or open) or perhaps how fast you are driving, or how responsive the car is to the wheel. In a complicated system, with many interactions, some kinds of problems (eyes or speed or steering sensitivity) can cause a well regulated system to become unstable, while other kinds of problems (something to do with the brakes or the rear view mirror, perhaps) are not as dangerous. In addition, some people’s eyesight, reaction times or driving skill may make them more or less vulnerable. Or another vehicle too close or giving your car a bump might send you out of countrol. Other kinds of dysregulation are also possible. Suppose the steering linkages got reversed and the harder we turned the wheel the more the car veered away from the center line. This would be a positive feedback, like reversing the wires on your thermostat. When the temperature increases, it tells the furnace to send more heat instead of less, resulting in a spiral out of control. These are but two examples. In a complex system many more are possible. In the case of the cytokine network, we are learning about some of these interactions but have a long way to go before we can understand what might make it unstable.

What does this background have to do with “cytokine storm”?

First it explains why it is hard to understand what “cytokine storm” means. It is a summary term for a very complicated situation. It also suggests we only have a vague idea of what is happening. This is indeed the case. Cytokine dysregulation is involved in other syndromes with symptoms much like those seen in complicated influenza (e.g., toxic shock syndrome or gram negative sepsis). In these cases the causes are more related to “always on” T-cell activation (stuck accelerator). Whether “always on” activation and thus continuous pro-inflammatory cytokine production, some other kind of cytokine dysregulation, or nothing to do with cytokines happens in influenza is still open to question. The strongest evidence comes from the clinical presentation of virulent influenza cases and the evidence in mice that infection with influenza virus carrying the HA gene from the 1918 virus seems to strongly activate some immune cells to over-produce a half dozen or more cytokines. An animation from The New England Journal article on H5N1 influenza by Mike Osterholm is meant to illustrate how a positive feedback could cause a cytokine storm, but it is only suggestive of one possibility, because we don’t know how a cytokine storm is produced in influenza (if indeed it is).

However, it is reasonable and plausible to say cytokine dysregulation might be involved in some virulent influenza infections. In desperation, clinicians have treated patients with potent anti-inflammatory drugs, usually steriods. There is no evidence that this helps. A “cytokine storm” of a more limited nature is sometimes seen in cancer chemotherapy patients, where it is treated in its earliest stages by iv. benadryl and steroids, with some success. However in these cases, there is no infectious agent involved; even if steroids worked for influenza-induced cytokine storm, they cause a general downshift of the immune system which might allow the virus to run rampant and kill the patient via ordinary viral pneumonia. In ordinary infection-related sepsis, steroids are shown to slightly increase mortality (Crit Care Med. 1995 Aug;23(8):1430–9.) This is but one of the complicating considerations that clinicians will have to navigate during an outbreak. An isolated study showed that in children with central nervous system (brain) symptoms—an early sign of cytokine storm—due to (human, not H5N1) influenza infection, mild and controlled reduction in body core temperature (hypothermia) seems to reduce damage to brain cells as well as reducing the progression to a full-blown cytokine storm and multi-organ failure. (Pediatrics International Volume 42 Issue 2 Page 197 - April 2000.)

In 2003, researchers at Imperial College London tested a drug that interferes with a “survival signal” that keeps activated T-cells working at the site of inflammation during influenza infection in mice. The signal, another cytokine designated OX40, essentially disables the brakes on the T-cell response. By blocking the OX40 receptor on T-cells, researchers were able protect mice from the serious symptoms of virulent flu (paper in J. of Experimental Medicine and reported in New Scientist). The drug, to be made by a company called Xenova Research, was supposed to be in phase I clinical trial in 2004, but we have no further information of its status (additional information solicited for this entry).

Cytokine Storms Venky Ramakrishna, International Society For Inerferon and Cytokine Research Jan 2006 pg 7–10.

Xenova retains all rights for the use of OX40 in up-regulation whilst Genentech Inc (since 2002) and Celltech Group (bought by UCB located in Belgium in 2005) have the rights for down-regulation. (maybe we should contact these companies and ask what is being done with these rights.)

Proinflammatory cytokine responses induced by influenza A (H5N1) viruses in primary human alveolar and bronchial epithelial cells Respiratory Research 2005, 6:135 Published Nov 2005

Cytokine Storm and the Influenza Pandemic Angela L. Petrosino, M.P.H., Medical College of Ohio

Inflammatory Response Current Concepts Edward R. Sherwood, MD, 56th Annual Referesher Course Lectures and Basic Science Reviews American Society of Anesthesiologists, 7 pgs

Storm on Wikipedia

Horst Ibelgrauft’s Cytokines & Cells Online Pathfinder Encyclopaedia (COPE) 18,000+ entries/pages | 48,800+ references | 215,500+ internal hyperlinks

Cytokine tutorial Dept of Veterinary Science and Microbiology, University of Arizona

Cell Interactions Cytokines Immunology Division, Department of Pathology, Cambridge University