The astonishing effect of monoclonal antibodies on B-cell tumors is there for everybody to see, but antibodies are really not designed to kill lymphocytes. They are splendid at killing bacteria, and good for killing cells without nuclei, like red cells and platelets, but nucleated cells have defences against them. Although there are natural autoimmune diseases like ITP and AIHA which are caused by antibody, it is very doubtful whether antibody has anything to do with autoimmune diseases where the damaged tissue has nucleated cells. Take pernicious anemia (PA) for example. A large proportion of the female population had natural antibodies against gastric parietal cells, but very few go down with PA. It’s the same with thyroid disease. Antibodies against thyroid microsomes are common, but myxedema is rare.
On the other hand we have now come to see that stem cell transplants work by T-cells attacking what they see as foreign. Graft versus leukemia is what saves lives, not rescuing the bone marrow following huge doses of chemo- or radiotherapy (as we used to think). The problem with allografts is that along with graft versus leukemia we have graft versus host disease which itself makes the procedure dangerous and gives us death rates of about 20% from the procedure.
It would be nice to use the efficiency of T-cell effectors coupled with the direction finding of monoclonal antibodies. In fact my daughter has just completed 4 years studying for a PhD on this very subject. The obvious way to attempt this is to couple together a monoclonal antibody that attacks the tumor cell with a monoclonal antibody that activates a T cell.
This technology has been in the public domain for nearly three years now and can be accessed through the Science website without charge (though it does require registration.
Briefly we are talking about single chain bispecific antibodies one end of which attacks the tumor cell (in this case an anti-CD19 and the other engages the T cell (an anti-CD3) They have been given the name BiTE (which stands for bispecific T cell engagers). The one that has drawn the interest of CLL patients is blinatumomab.
The Science paper reported on a phase I dose escalation clinical trial in 38 patients with non-Hodgkins lymphoma who had relapsed following standard therapies. They had had a median of three previous chemotherpeutic regimens and 87% had had rituximab. Most of the patients were follicular lymphomas or mantle cell lymphomas, but there were 3 with CLL/SLL. One important factor is that the dose used was minute - about 100,000 times smaller than the dose of rituximab and therefore it ought to be a lot cheaper to make (Ha! Ha!). Patients were offered 8 weeks of daily treatment as long as they had not progressed after four.
The reason for the very low dose was the fear of cytokine storm as happened a few years ago at Northwick Park. We had known that this was a risk when you activated T cells and I give an example our own use of a trispecific antibody back in the 1990s. The illustrations show a 'before and after' shot of a patient’s melanoma recurring in the skin of a previous excision. We treated him with a very low dose of an Anti-CD3/anti-CD2/anti-CD28 trispecific antibody with no attempt to direct the antibody at the melanoma, simply relying on the activated T cells to do their work. We had experience at the time of using systemic IL-2 to activate T cells but although this had been fairly successful with 50% response rates including 20% CRs, the toxicity was severe with cytokine storm like affects. For this reason we began using the trispecific antibody at doses of 200 nanograms – roughly 2 million times smaller than the dose of rituximab.
The Science paper gave two bits of useful information on this subject. First, they actually measured the release of cytokines by their treatment and second they have demonstrated that giving high doses of steroids does not detract from the killing ability of the bispecific antibody. In fact patients were given methylprednisolone and low molecular weight heparin during the first treatment days as prophylaxis against cytokine release problems. The side effects of the treatment were mostly well managed. One patient with a history of near fatal sepsis died after developing an infection on this treatment. Another patient with hypogammaglobulinemia had treatment discontinued because of the development of pneumonia. One patient with a history of renal insufficiency had the drug stopped because of metabolic acidosis accompanied by a seizure. Five patients had the drug stopped because of CNS-related events; two with confusion, two with cerebellar symptoms and one with the seizure already alluded to. Al these events were fully reversible after a couple of days. There has been a suspicion that CD19 might be on some cerebral tissues as a result of these side effects.
TNF-alpha was transiently increased in 6 patients. There was a transient elevation of IL-10 in 25/39 patients of whom 19 also showed transient elevations of IL-6 or gamma interferon.
Those patients with a peripheral lymphocytosis showed rapid and long-lasting clearance of lymphocytes from the blood. There was an increase in blood levels of effector memory CD8+ and CD4+ T cells but no change in naive, central memory and CD45RA+ effector memory T cells.
All seven patients who received more than 0.06 mg/sqm/d responded, with 2 CRs and 5 PRs. At smaller doses there were 2 CRs and 2 PRs (one of the PRs being in CLL/SLL). The three CLLs received very small doses of 5 and 15 micrograms/sq m/d and the one who responded received 150 micrograms/sq m/d. This is very preliminary but we should remember that CD19 is not low on the surface of CLL the way that CD20 is so there is no reason to believe that CLL would behave any differently than other forms of NHL. In fact, the one CLL patient who only received 5 micrograms/sq m/d had a minimal response (25% shrinkage of disease).