Gilenya and PML: cause, effect and risk

Today the FDA followed up on a case of progressive multifocal leukoencephalopathy (PML) in a multiple sclerosis (MS) patient, first reported in July, by announcing an investigation. FiercePharma spelled it out as follows: 

Novartis’ MS drug Gilenya tied to rare brain infection FDA sends out alert
That’s a pretty scary headline, and the news drifted across twitter streams like a dry fly on a light tippet. Novartis quickly declared that there was no indication that Gilenya caused the rare brain infection, the patient was on co-medications, had been given intravenous steroid treatment, etc. etc. 

Lets back up. 

PML is caused by an infection of the JC virus within the central nervous system (CNS). PML is common and almost always benign, unless it occurs in a patient who in severely immunosuppressed. Of note, PML first came to public attention in the early days of the AIDs epidemic, appearing in HIV patients whose disease had progressed to the point that they had no more T cells in circulation. T cells are a critical arm of the immune system, that, along with B cells and other cells, control pathogen infections, including viral infections.   

MS, as readers know, is a chronic inflammatory autoimmune disease in which T cells, B cells and other components of the immune system engage in a deranged attack on the CNS, particularly targeting the myelin and other proteins that encase neurons in a sheath (thus, the myelin sheath). Neurons stripped of myelin do not signal properly and die off pretty quickly. The disease occurs in flares, geographically localized. The product of a flare is a “lesion” within the CNS, where the effect of demyelination and neuronal die back can be visualized by an MRI image. 

So, immune system cells, particularly T cells, are necessary for protection from JC-virus induced PML on the one hand, and cause MS flares and lesions on the other hand. It should be noted that more flares = more lesions = MS relapse or progression = more disability. 

The interplay of immune defense or immune surveillance and the pathology of MS first came to light with the development of two powerful immune modulating agents: Tysabritmand Rituxantm. Tysabri is a monoclonal antibody to the alpha4beta1 integrin (thus, a4b1) developed by Elan and licensed to Biogen. Tysabri is the single most potent agent developed to date for the treatment of MS, and its mechanism of action suggests why this is so. T cells use the a4b1 integrin to migrate out of circulation and across cell layers like endothelium. To gain access to the CNS, T cells must cross an endothelial cell membrane, a layer of basement membrane and then an epithelial cell membrane. This collection of barriers between the circulation and the CNS is called the blood-brain barrier. Importantly, a4b1 is necessary for efficient trafficking of T cells out of circulation and into the CNS (also into the lung, the gut mucosa and other tissues). When a4b1 is blocked by treatment with Tysabri, which binds and neutralizes a4b1, T cell movement out of circulation is effectively stopped. Indeed one of the first effects of Tysabri treatment is a rapid increase on the white blood cell count, as T cells excluded from tissues begin to accumulate in the bloodstream (and lymph). 

Tysabri treatment reduces signs and symptoms of MS, reduces relapses, number of flares, time to flare, number of lesions, and disease progression. Tysabri treatment also puts patients at a low, but still troubling, risk of developing PML, since the population of T cells excluded from the CNS also include those T cells that normally would control JC virus. In this case the mechanism of action (MOA) of Tysabri contributes to the risk of PML, and this can be seen in patients treated only with Tysabri (as monotherapy) as well as in patients taking multiple medications. The risk is handled in several ways: patients are monitored for the presence of antibodies to PML, that indicate exposure and likely the presence of quiescent virus in the patient, and by temporarily suspending Tysabri use for several months, usually after 2 years and intermittently thereafter. Its not a well defined art at this point, but because of the efficacy achieved with use of Tysabri, most patients and their physicians tolerate the risk, now estimated at about 1% across all patient groups. 

PML in Rituxan treated patients presents very differently. Rituxan is an antibody that binds to CD20 expressed on the surface of B cells, and this binding causes the B cells to be depleted. The antibody was developed by Idec and acquired by Biogen when those two companies merged to become Biogen Idec. Before Rituxan was essentially removed from the clinical MS landscape – a truly scandalous story for another time – it had undergone clinical trials in MS, showing excellent efficacy and limited toxicity, which did not include PML. Rituxan causes PML very rarely, and almost exclusively in lymphoma patients, who have been treated previously or concurrently with chemotherapy. In other words, PML was seen in heavily immunosuppressed cancer patients. Rituxan is also used in rheumatoid arthritis (RA) and a host of other inflammatory and autoimmune indications, and the risk of PML in these patients taking Rituxan is at least 10-fold lower than in lymphoma patients taking Rituxan. This is because RA patients are much less immunosuppressed than cancer patients. What we see then with the risk of PML from Rituxan is that the risk is really dependent on the prior immune status of the patient. In this sense you would not make the claim that Rituxan causes PML, but rather that it contributes to the risk of developing PML. IN the same way, Tysabri treatment doesn’t cause PML, but also contributes the risk of developing the disease. It is the case that the risk is higher with Tysabri than with Rituxan. 

Why Rituxan treatment allows JC viral escape and the occurrence of PML is a result of the depletion of the B cells that Rituxan targets. B cells make antibodies to pathogens, including anti-JCV antibodies, and this contributes to the control of virus by clearing both free viral particles and infected cells. Further, B cells supply “help” to T cells by secretion of cytokines and other factors that induce T cell survival, proliferation and function. So, T and B cells together are required for efficient JC virus control, although the available data suggest T cells are more directly critical. 

OK, back to Gilenya. I guess it was 4 or so years when the phase 3 data was coming out that I had a conversation with Jeff Browning, then at Biogen Idec and now at BU, about the results. He said something along the lines of, “sure, impressive results, but wait until they start to get PMLs.” 

Prescient, and certainly a well informed view. Jeff, myself, Paula Hochman (all from Biogen) and Reina Mebius (a key collaborator) along with a host of colleagues both inside and outside the company, had reinvigorated the study of lymph node biology in the mid-90s by unraveling the mechanisms by which these organs develop (search PubMed if interested), and as the field literally burst with new findings, attention turned to the control of lymphocyte (T cell and B cell) movement through the lymphatic system, within which lymph nodes are the critical organs. Lymph nodes (LN) provide the essential microenvironment for surveillance of tissues, including the CNS. Many labs unraveled the biology of cell traffic through LN, but it was Jason Cyster at UCSF who discovered that an agent in clinical trials called FTY720 functioned by shutting down the molecular pathway that allows T cells to exit LNs. The molecular pathway was driven by a bioactive lipid called sphingosine 1 phosphate (S1P). S1P bound and signaled through a class of receptors (the S1P receptors) and S1P receptor 1 was shown to be critical for T cell egress from LN. 

FTY720 became Gilenya. Thus, the mechanism of action of Gilenya is eerily similar to that of Tysabri – it controls the trafficking of T cells, in this case by trapping them inside of lymph nodes (B cells as well, although not so completely). The fact that Gilenya treated patients do not develop lymphadenopathy(LN enlargement) suggests that these trapped T cells die off. Gilenya treated patients experience a rapid drop in circulating white blood cell number – the opposite of what is seen in Tysabri treated patients – but the net effect is the same: T cells do not access the CNS, and therefore the signs and symptoms of MS are reduced, the relapse rate drops, the number of new lesions drops, etc, more or less similar to Tysabri treatment, albeit somewhat less effective. Gilenya has a host of non-immune system toxicities, which we won’t discuss here, but until July, Novartis, who developed the drug, had reported no patients with PML. 

Now, back to the headline. What Jeff Browning had recognized was that by its’ mechanism of action of Gilenya should increase the risk of PML, just as Tysabri increased that risk. Both stopped T cells from entering the CNS, by different mechanisms, and therefore both should prevent immune defense against JC virus in the CNS. Furthermore, the degree of risk should be proportional to the extent to which the exclusion of T cells was achieved. Since the efficacy of the drug should also be proportional to the extent to which t cells were excluded from the CNS . . . it follows that efficacy and risk of PML should be correlated. 

When Novartis states that Gilenya does not cause PML, they are being accurate, as accurate as when we state that Tysabri and Rituxan do not cause PML. More importantly and usefully, Gilenya is now hypothesized to be associated with the risk of developingPML, although it’s certainly a low risk. More specifically, the null hypothesis should be that Gilenya is associated with PML risk, based on its’ mechanism of action, and regulatory agencies and physicians should take note accordingly. 

We’ll watch out for other cases and comment as the clinical landscape develops. 

As always, stay tuned.