Presented by Brian K. Link, MD, at the Radioimmunotherapy of Non-Hodgkin's Lymphoma symposium held at the American Society of Hematology 42nd Annual Meeting, December 1, 2000, in San Francisco, California.

Radiolabeled antibodies represent an incremental step forward in our goal of achieving functional immunotherapy in cancer. By way of perspective, if you look at immunotherapy highlights, the notion of using our immune system to treat cancer dates back now over a century, and in fact as far ago as 1904, when Ehrlich talked about "magic bullets." The dream became closer to reality in 1975 with Kohler and Milstein's work on the production of monoclonal antibodies, followed by clinical trials of the use of murine monoclonal antibodies in lymphoma.

The use of antibodies to treat cancer is based on the original hypothesis that the monoclonal antibodies can act as anti-cancer therapies by attaching to an external surface molecule and either delivering a toxin or inducing complement-mediated lysis or antibody-dependent cellular cytotoxicity. The underlying theme is that with the right tools, ultimately we can manipulate the immune system to fight cancer, much in the way it fights infections.

Much of the early work in this modality of therapy, just like the other modalities, radiotherapy and chemotherapy, has started and become most studied with the lymphomas. In part this is because lymphomas are potentially more sensitive to immunotherapy than other cancers, but ultimately because they are easy to work with in vitro, they have a number of well-defined antigens that are either cell- or tissue-specific and are biologically important, there are effective animal models in which to test them, and, importantly, they are sensitive to a variety of noxious agents.

The results of initial trials in which surface immunoglobulin was targeted using monoclonal murine antibodies were mixed. There were certainly several responders proving proof of concept, and some responses were clinically meaningful. However, the murine antibody was felt to be imperfect due to its immunogenicity, its imperfect activation of immune effector mechanisms, and what was felt to be suboptimal pharmacokinetics.

The decades following those initial trials focussed on ways to take incremental steps from those initial murine antibodies. One direction of improvement has been the humanization of unlabeled antibodies. Other directions studied include the use of immunotoxins, and, of course, the use of radiolabeled attachments to increase the potency of monoclonal antibodies.

Considering the variety of antibodies that are going to come into the clinical arena for testing and for clinical use, it is important to recognize that they are targeted against a variety of antigens, and that there are properties of the antigen that should be considered. Ideally we'd like them to be tumor-specific, or at least tissue-specific, and, for practical reasons, it would make sense if the antigen is preserved and somewhat ubiquitous throughout the species, so that individual antibodies don't have to be made.

Also, there are some disadvantages to targeted antigen that is shed out into the circulation. Ideally, the targeted antigen should be focussed on the cell surface. It is important to recognize that when the antibody interacts with the antigen, it has the potential to undergo immodulation and the antibody-antigen complex can either be internalized or shed out into the circulation, or it can remain fixed on the surface of the cell.

Finally, it is becoming important to realize the function of the antigen actually being targeted, because a number of antigens have important roles in cell cycle, cell growth, and cell death mechanisms.

In addition to considering the antigen that is targeted, the basic function of the construct of the antibody is important. Remember that these antibodies, as immunoglobulin molecules, have a hypervariable region which is responsible for the binding specificity. It's the other end of the antibody, the FC portion, that is generally accepted to mediate most of the biologic activity through either its complement binding or through binding to FC receptors.

Recall that the original antibodies were pure murine in nature, which caused the basis of a number of concerns. To alleviate these concerns, research has been conducted on the use of humanized antibodies, which are now technically available for production, but have not gone very far in clinical studies. When we talk about humanization of antibodies, we are often referring to a variable degree of chimerism.

To focus on the advances made with the unlabeled antibodies through the process of humanization, keep in mind that humanized antibodies are anticipated to have longer circulating half-life, be less immunogenic, and be more more potent in terms of antibody-dependent cellular cytotoxicity and complement-mediated lysis.

One hypothesis as to how these antibodies work, in simplistic terms, is that the antibody attaches to the antigen and then through its FC portion interacts with NK cells or the cellular effector mechanisms of the immune system. An alternative view is that when the antibody attaches to the antigen, it induces the complement cascade and ultimate cell death through that mechanism.

Mitch Reff demonstrated, at least with C2B8, the parental protein of rituximab, that when comparing the chimeric version of the antibody to the murine version of the antibody in assays that utilized ADCC of complement-mediated lysis of CD20 positive cells, the chimeric protein was much more efficient at inducing either ADCC or complement-mediated lysis of antigen-positive cell lines. And so it was with that general concept that these antibodies moved forward again into the clinic for another foray.

The first antibody to take advantage of that was the Campath-1H, which targets CD52, is expressed on B-cells and T-cells and monocytes, and is ubiquitous throughout the human population.

When the humanized version of Campath was given to patients with a variety of lymphoid malignancies, it was found to be very effective at depleting circulating tumor cells in the blood and bone marrow. But, for a variety of reasons, lymph nodes that were involved with tumors were relatively resistant to the effects of Campath. There was relatively profound monocyte depletion due to the distribution of CD52 with T-cells and B-cells and monocytes, which led initially to subsequent infections and required new strategies for infection control.

Rituximab was another antibody then to take advantage of the technology to make chimeric or humanized antibodies. Recall that it is chimeric with the variable region being murine and the backbone human IgG1. It targets CD20, which again, is expressed on all B-cells, but not T-cells, or NK cells, and is felt, to some extent, to be involved in regulating the cell cycle.

To summarize a large body of work on rituximab in low-grade non-Hodgkin's lymphoma: it was observed initially, when given to patients with relapsed disease at 375mg/m2 weekly X four doses, that from a safety standpoint there was a unique infusion-related syndrome, while otherwise Grade III cytopenias were infrequent. There was B-cell depletion but no apparent resulting infections, and no immunogenicity. Importantly, a number of reports have described a 50% overall response rate, with median response duration of about one year.

It is important to recognize that these results represent proof of concept. They were not designed to show how to treat patients, but literally just to prove the concept that these antibodies have clinical activity. No matter how long these initial studies are followed, they will never give any insight in terms of how we have changed the natural history of those patients' disease.

When we go back and look at some of the subsets in these initial trials, it is important to recognize that the follicular histologies were very susceptible to rituximab, with response rates greater than 50% in a number of reports.

When patients who previously responded were re-treated, overall response rates were about 40%, but median time to progression was actually longer than was seen for the initial response, and that's something that is potentially unique to immunotherapy. More recently, rituximab has been studied as a first-line therapy for low-grade lymphoma patients. Results suggest that rituximab in this setting has an initial response rate of about 45%, but with increasing cycles can be brought to as high as 71%.

It is important in the field of oncology to consider how to combine rituximab with our other tools for the treatment of low-grade lymphomas. Czuczman reported in the Journal of Clinical Oncology on 40 patients with low-grade lymphoma who were largely previously untreated. When he combined CHOP and rituximab, he found an overall response rate of 100%, with the majority of those being complete responses. Twenty-eight of those patients remain in ongoing remission at 40 months.

The promising results of these Phase II studies, while not showing for sure that we have changed the natural history of low-grade lymphomas, have led to the generation of larger cooperative group studies, including those at SWOG and ECOG, both of which involve chemotherapy followed by rituximab. Additionally, Bob Marcus in the United Kingdom is looking at concurrent chemotherapy with CVP and rituximab versus CVP alone, and MD Anderson researchers are exploring a combination of rituximab and the FND regimen.

The field is moving forward now, targeting other molecules, such as CD52, CD22, and HLA-DR. Testing antibodies that target alternative molecules is not based on the hypothesis that ADCC or complement-mediated lysis may be improved, but rather that different molecules may initiate different intracellular signaling cascades that might influence cell growth, cycling, and death.

Epratuzumab is an antibody targeting CD22, which, again, is B-cell specific, is found on the surface and cytoplasmic cellular portions, and is not shed. It binds across a wide range of lymphoma subtypes and has minimal reactivity with normal tissues. Epratuzumab is the humanized version of the murine LL2 antibody. It is CDR-grafted into an IgG1 backbone. Radiolabeled and unconjugated constructs with epratuzumab are under evaluation. Reports on Phase I-II studies of this antibody in more than one hundred subjects show that the therapy is well tolerated, with no grade III-IV infusion reactions and no observed immunogenicity. It was tolerated well in patients previously treated with either rituximab or stem cell transplant.

Hu1D10 is an antibody against HLADR, which is a potentially attractive target because it is known to play a role in B-cell signaling and is not shed. The Hu1D10 antibody, again, is a humanized, CDR-grafted IgG1. It binds again across a relatively wide range of lymphoma subtypes, but is not as ubiquitously expressed across the species, with only about two-thirds of patients with B-cell lymphoma reacting with 1D10. In addition, it is not quite as tissue-specific as other antibodies, in that it also reacts with some benign B-cells and some interstitial cells in a variety of organs. A 20-patient Phase I study of Hu1D10 has been completed, with results suggesting that there may be a unique mechanism of action invoked by targeting HLADR.

Regarding use of the antibodies in aggressive non-Hodgkin's lymphoma, which has not been as extensively studied yet, Bertrand Coffier has reported on a series of 54 patients with intermediate and high-grade non-Hodgkin's lymphoma that were treated with rituximab as a single agent. This was a very heterogenous patient population here; some in first relapse, some in primary refractory, some PR, and some actually in the first line of therapy. Eight weekly infusions resulted in an overall response rate of 32%, again providing a proof of concept that there is some clinical activity.

In an upcoming article in the Journal of Clinical Oncology, Julie Vose describes a Phase II study in which rituximab plus CHOP was combined to treat newly diagnosed large-cell, or aggressive, lymphomas. Thirty-three patients with untreated aggressive non-Hodgkin's lymphoma all received six cycles of rituximab followed by CHOP, resulting in an overall response rate of 94% at 24 weeks. Serious adverse events in this study were no different than expected for CHOP alone.

ECOG, CALGB, and SWOG are cooperating in a study where elderly patients with newly diagnosed diffuse large-cell lymphoma are randomized to CHOP versus CHOP plus rituximab, and then responders are either randomized to observation or consolidative or maintenance therapy. Additionally, Bertrand Coffier and the GELA group will soon be presenting their first analysis of a study completed in Europe of CHOP versus rituximab plus CHOP.

To summarize, it appears fairly clear that we may have impacted the natural history of some of these patients with lymphoma through the use of antibodies, particularly with the chimeric CD20 antibody. Further maturation of randomized studies may demonstrate that such therapy can be clinically effective and useful.

Finally, continued development of this field may center on targeting other antigens, continuing to examine various antibody-plus-chemotherapy combinations, and exploring the potential of radioimmunotherapies.