Improving the treatment of non-Hodgkin lymphoma with antibody-targeted radionuclides

Authors

  • Robert M. Sharkey PhD,

    Corresponding author
    1. Garden State Cancer Center at the Center for Molecular Medicine and Immunology, Belleville, New Jersey
    • Center for Molecular Medicine and Immunology, Garden State Cancer Center 520, Belleville Avenue, Belleville, NJ 07109
    Search for more papers by this author
    • Fax: (973) 844-7020

  • Habibe Karacay PhD,

    1. Garden State Cancer Center at the Center for Molecular Medicine and Immunology, Belleville, New Jersey
    Search for more papers by this author
  • David M. Goldenberg ScD, MD

    1. Garden State Cancer Center at the Center for Molecular Medicine and Immunology, Belleville, New Jersey
    Search for more papers by this author

  • The articles in this supplement were presented at the “12th Conference on Cancer Therapy with Antibodies and Immunoconjugates,” in Parsippany, New Jersey, October 16-18, 2008.

Abstract

Radioimmunotherapy of non-Hodgkin lymphoma comprises a 90Y- or 131I-labeled murine anti-CD20 IgG, but both agents also include a substantial dose of unlabeled anti-CD20 IgG given immediately before the radioconjugate to reduce its uptake in the spleen (primary normal B-cell antigen sink); this extends its plasma half-life and improves tumor visualization. Thus, these treatments combine an effective anti-CD20 radioconjugate with an unconjugated anti-CD20 antibody that is also therapeutically active, but the large anti-CD20 IgG predose (∼900 mg) may diminish the tumor localization of the radioimmunoconjugate (eg, 10-35 mg). We have examined alternative approaches that enhance radionuclide targeting and improve antitumor responses. One uses a 90Y-labeled anti-CD22 IgG (epratuzumab) combined with an antibody therapy regimen of a humanized anti-CD20 IgG (veltuzumab). Pretargeted radionuclide therapy using a trivalent, humanized, recombinant bispecific anti-CD20 antibody with a 90Y-hapten-peptide is another highly effective method that is also less toxic than directly radiolabeled IgG. Finally, all approaches benefit from the addition of a consolidation-dosing regimen of the anti-CD20 IgG antibody. This article reviews these various options and discusses how some fundamental changes could potentially enhance the response and duration from radionuclide-targeted therapy Cancer 2010;116(4 suppl):1134–45. © 2010 American Cancer Society.

Radioimmunotherapy (RAIT) is an approved approach for the treatment of follicular and transformed non-Hodgkin lymphoma (NHL) with 2 radiolabeled anti-CD20 IgG products.1-390Y-ibritumomab tiuxetan (Zevalin; Cell Therapeutics, Seattle, Wash) and 131I-tositumomab (Bexxar; GlaxoSmithKline, Philadelphia, Pa) have shown improved objective responses over the corresponding anti-CD20 IgG used with each agent; namely, rituximab (Genentech, San Francisco, Calif; Biogen Idec, Cambridge, Mass) and tositumomab, respectively, which themselves have significant antitumor activity.4, 5 New clinical studies are lending increasing support for their use in frontline and consolidation settings,2, 6-22 but new laboratory investigations are providing insights into ways that could improve responses, and in some cases with less toxicity.

To improve these treatments, we need first to understand the current treatment regimens, and how each of these agents was developed. Monoclonal antibody-based therapeutics in hematopoietic malignancies started with an examination of several naked murine monoclonal antibodies in lymphoma and leukemias. Some objective responses were observed, but overall the antitumor effects were minimal, and in many respects limited by the amount of antibody and the number of treatments that could be given with the murine IgG.23-32 For example, in a small pilot study, Press et al reported progressively better objective antitumor responses in patients given a continuous infusion of the murine anti-CD20 IgG 1F5, but it required nearly 2 g of antibody to get a complete response. Today, this amount of antibody is commonplace for antibody therapeutics, but at that time, this was beyond the capability of most facilities to produce, and development of an antimouse antibody response further restricted the treatment. However, preclinical models often found that immunoconjugates (antibodies conjugated with radionuclide, drugs, or toxins) were much more active than the corresponding unlabeled antibody and required less antibody. After DeNardo et al reported encouraging responses with a radioiodinated anti-HLA-DR antibody in B-cell malignancies, several other studies using antibodies to CD37, CD20, and CD22 were pursued, first in a myeloablative setting with bone marrow support, and then with nonmyeloablative doses, to have more substantial antitumor activity than that reported with the naked antibodies.33-37 The initial studies performed with CD37 and CD20 antibodies found a more favorable biodistribution for the radioconjugates when additional unlabeled anti-CD20 IgG was mixed with the radioimmunoconjugate.34, 35 Imaging studies showed there was intense uptake of the radioimmunoconjugate in the spleen, accompanied by rapid blood clearance, when low protein doses were used.38 Because all of the above-mentioned antibodies bind to normal and malignant B-cells, the large repository B-cells in the spleen (normal and malignant B-cells, often resulting in splenomegaly) acted as a sink, taking up the small amount of radioconjugate from the blood before it had an opportunity to circulate and localize in all tumor sites. Administering additional antibody protein reduced splenic uptake and slowed the radioconjugate's clearance from blood. Antibodies against different antigens require different amounts of antibody for this to occur.34, 35, 39-44

Adding additional antibody to an immunoconjugate raises concerns that the unlabeled antibody would compete for the binding sites in the tumor, reducing the amount of the radioimmunoconjugate localized in the tumor. However, in nude mice bearing Raji human B-cell lymphoma xenografts, Buchsbaum et al45 confirmed that tumor localization of the radiolabeled anti-B1 antibody (murine anti-CD20 IgG2a; tositumomab) could be enhanced by first predosing the animals with the unlabeled antibody before the radioimmunoconjugate. Although mouse B-cells do not have cross-reactivity with the anti-B1 antibody, nude mice have low levels of murine IgG2a.46, 47 As a consequence, when small amounts of exogenously added murine IgG2a are given to the mice, the antibody is rapidly bound by cells in the spleen having receptors for this IgG form, causing increased splenic uptake and rapid blood clearance, which subsequently reduced the amount of antibody available for tumor targeting. The predose preoccupies these receptors, allowing the radioimmunoconjugate to circulate longer in the blood, thus giving it more opportunity to bind to the tumor. Although this is not the same mechanism of antibody binding/clearance that occurs in humans, the result is similar. In this model, they showed that a predose of 0.1 mg reduced splenic uptake and increased the concentration of the radioimmunoconjugate in the tumor and blood, but when increased to 1.0 mg, tumor uptake decreased, a likely consequence of having too much unlabeled anti-CD20 that could compete with the radioimmunoconjugate for CD20 binding in the tumor.

Several groups independently examining 2 anti-CD20 IgGs confirmed the initial clinical findings supporting the use of a blocking dose of unlabeled antibody before the radioimmunoconjugate. Knox et al found a 1.0 mg/kg predose of unlabeled anti-CD20 reduced splenic uptake, improved tumor visualization, and slowed blood clearance for 90Y-ibritumomab tiuxetan, and a higher dose of 2.5 mg/kg improved the number of lesions detected by external scintigraphy.41 There was no apparent impact on the quantitative amount of the radioimmunoconjugate delivered to the tumor in patients given the 1.0 mg/kg predose compared with no predose, but tumor uptake was highly variable, which may have masked possible differences. Later studies examining 100 mg/m2 and 250 mg/m2 predose levels also observed the desired decrease in splenic uptake and slower blood clearance and concluded there was no apparent reduction in tumor uptake.48 Subsequently, investigators chose the predose of 250 mg/m2 because clinical trials with rituximab at the time were showing encouraging antitumor activity, and thus the additional amounts of rituximab given as a predose were considered a potential boost to the antitumor activity of the radioimmunoconjugate treatment. Kaminski et al had direct evidence that a 685-mg predose of the unlabeled anti-B1 given with a diagnostic injection of 131I-anti-B1 IgG caused tumor shrinkage even before the therapeutic radioimmunoconjugate was administered,35, 36 providing considerable incentive to add higher amounts of the unlabeled antibody with the radioimmunoconjugate. Interestingly, preclinical studies in mice bearing human B-cell lymphoma xenografts found better responses with the unlabeled B1 as compared with 131I-B1, lending additional support that the unlabeled antibody could benefit the overall response.49 Wahl et al later reported other studies that examined 95-mg and 450-mg predose amounts of unlabeled tositumomab, finding that the 95-mg predose substantially slowed the clearance of the radioimmunoconjugate. Blood clearance was not substantially different at the 450-mg predose, but the higher predose was preferred for patients with larger tumor burden or splenomegaly.42, 50

Preclinical and clinical data supported the value of a predose of the unlabeled anti-CD20 IgG to improve the targeting of the radioimmunoconjugates, but new preclinical data are raising questions regarding how much antibody should be given before the therapeutic radioimmunoconjugate doses. Currently, 2 doses of 250 mg/m2 of rituximab are given before 90Y-ibritumomab tiuxetan and 2 doses of 450 mg of tositumomab are given before 131I-tositumomab, amounting to nearly 900 mg of unlabeled antibody before the therapeutic radiolabeled antibody is administered. Indeed, the prerequisite established in the initial trials of a pretherapy imaging study performed several days in advance of the therapeutic treatment loads the patient with considerable unlabeled antibody well before the therapeutic radioimmunoconjugate treatment is given. This pretherapy imaging study is necessary to set the therapeutic dose of 131I-tositumomab, but the only requirement for the 111In-ibritumomab tiuxetan dose is to assess if there is an altered biodistribution, which appears to occur very rarely.51 In Europe, where the 111In-ibritumomab tiuxetan imaging study is not required before 90Y-ibritumomab tiuxetan, the predose of unlabeled antibody is nevertheless still given, attesting to the belief that the unlabeled antibody is an integral and important component of this treatment regimen.19

Gopal et al were the first to highlight the possibility of compromised targeting of an anti-CD20 radioconjugate (131I-tositumomab) when an unlabeled anti-CD20 predose was given.52 They reported a 55% decrease in tumor uptake when nude mice bearing human lymphoma xenografts were predosed with 2 doses of 0.4 mg of unlabeled tositumomab given 24 and 6 hours before 131I-tositumomab (0.2 mg 131I-tositumomab was coinjected with 0.4 mg of an irrelevant murine IgG2a to block Fc-receptor–mediated binding, as discussed above), and as a consequence, antitumor activity was also reduced significantly. We showed subsequently similar reductions (-47%) in tumor uptake in nude mice bearing established (0.5 cm3) Ramos human B-cell lymphoma xenografts that were predosed with 1.0 mg of unlabeled humanized anti-CD20 IgG1 (veltuzumab, hA20; Immunomedics, Inc., Morris Plains, NJ) given 1 day, or even just 1 hour, in advance of 111In-veltuzumab.53 Although there was a trend for tumors to progress more rapidly in animals given a 1-mg predose 1 day before 90Y-veltuzumab, the difference between no predose and predosing was not significant (median time to progress to 3.0 cm3, 7.7 vs 4.9 weeks, respectively; P = .189) (Fig. 1). Because 1 mg given to a 20-g mouse (ie, 50 mg/kg) is equivalent to 4 mg/kg in a human or about 280 mg (based on a FDA-recommended conversion factor54), we examined a 0.25-mg predose or 1-mg/kg human equivalent dose. This was the smallest amount of an unlabeled anti-CD20 IgG given as a predose that was tested clinically, which was sufficient with both 90Y-ibritumomab tiuxetan and 131I-tositumomab to “normalize” the blood clearance of the radioconjugate. With the lower predose, tumor uptake was decreased by only 25% and, as expected, the therapeutic response again was not significantly different from that of no predose (eg, median time to progress to 3.0 cm3, 8.9 vs 6.0 weeks, no predose vs with predose, respectively; P = .813; Fig. 2). In this study, a separate group of animals was given additional amounts of unlabeled veltuzumab starting 1 week after 90Y-veltuzumab. This treatment combination resulted in 100% of the animals having no evidence of tumor after 12 weeks. Additional observations of this combination therapy are discussed below.

Figure 1.

The effect of veltuzumab predosing on the therapeutic response of 90Y-veltuzumab. Groups of nude mice (n = 10) bearing subcutaneously. Ramos human B-cell lymphoma xenografts (average 0.65 cm3) were administered 0.115 mCi of 90Y-veltuzumab (0.05 mg) alone or they were given 1.0 mg of veltuzumab 1 day before 90Y-veltuzumab treatment. Survival based on time for tumors to progress to 3.0 cm3.

Figure 2.

90Y-veltuzumab therapy with consolidation veltuzumab therapy. Nude mice bearing subcutaneous Ramos tumors were given 90Y-veltuzumab alone (0.13 mCi/0.05 mg), or a reduced veltuzumab predose of 0.25 mg was given 1 day before receiving 90Y-veltuzumab, and another group received the predose, as well as a veltuzumab consolidation therapy starting 1 week after the radioconjugate using 1.0 mg veltuzumab followed by 3 more weekly injections of 0.5 mg veltuzumab. Survival based on time to progression to 3.0 cm3.

Although these human CD20-models are unable to account for how much residual unlabeled antibody might be available to compete for tumor binding with the radioconjugate had there been an antigen sink, it was interesting to find, at least in our experience, that antitumor responses were not more obviously affected by predosing. Several reports have highlighted the ability of the naked antibody to induce apoptosis or influence the cell cycle that can enhance the sensitivity of the tumor to radiation,55-59 so these results might reflect the additive effects achieved by combining the anti-CD20 therapy with a radioimmunoconjugate Still, one would expect that there is predose level that, if exceeded, the combined effect would not be able to make up any reductions in radionuclide delivery and, therefore, the net response would be diminished had the radioimmunoconjugates uptake been optimized. Certainly the experience of Gopal et al52 confirmed that this can occur with a 0.4 mg predose of tositumomab before its 131I-radioimmunoconjugate, and while there was trend to a significant reduction in antitumor uptake with a 1-mg predose of veltuzumab before 90Y-veltuzumab, it did not reach a statistically significant level in the same Ramos xenograft model. Although we believe both studies support the view that excessive predosing can be harmful to a radioimmunoconjugate therapy, because anti-CD20 antibodies have at least 2 different mechanisms of action, it is possible that the effects might differ when using a type I or type II antibody.60, 61 However, it is difficult to predict what the potential boost in antitumor response provided by the unlabeled antibody might be when added to the radioimmunoconjugate, because this will undoubtedly be variable in different cell lines and in different targeting situations. Nevertheless, given the sensitivity of hematopoietic malignancies to radiation, it is reasonable to expect that responses would improve if more radiation could be selectively targeted to the tumor. This leads to a consideration of how to optimize the delivery of radiation to tumor, and how might we achieve the maximum benefit from the very effective anti-CD20 antibody immunotherapy when used in combination with a radioimmunoconjugate. Several different approaches are discussed below.

Anti-CD22 Radioconjugate Combined With Anti-CD20 Immunotherapy

Although naked anti-CD20 antibodies remain the most potent for NHL, there are several other antigens that can be used for targeting radionuclides. Some of the more commonly expressed antigens include CD19 and CD22 expressed almost exclusively on B-cells, and importantly not on progenitor cells. Many of the other antigens are also expressed on other hematopoietic cells from different lineages, such as CD37, HLA-DR, CD74, CD138, and CD45. For a variety of reasons, it is generally better to select an immunoconjugate with a more limited cell-binding profile.

Our group has studied an anti-CD22 antibody, known as LL2 or epratuzumab (Immunomedics, Inc., Morris Plains, NJ). First used as a radioimmunoconjugate, 99mTc- and 131I-labeled murine LL2 had excellent targeting properties even at low protein doses, and 1 of the initial patients given a small dose of protein and radioactivity experienced a measurable antitumor response.37, 62, 63 Subsequent studies revealed the antibody was internalized, and, therefore, tumor uptake was improved by using a residualizing radiometal, such as 90Y, over tyrosine-linked 131I.64 Predosing or even coadministration of unlabeled antibody did help reduce splenic uptake, but only 50 mg was sufficient, most likely because of the lower antigen density of CD22 on most malignant and nonmalignant B-cells.40, 43 Objective responses have been observed in phase 1 trials with 90Y-epratuzumab given as a single or fractionated dosing regimen, and antitumor responses were observed in lesions that were not observed by external scintigraphy of the radioimmunoconjugate.65-67 Recently, a multicenter phase 1 and 2 study using 2 fractionated doses given 1 week apart showed a much higher cumulative dose could be tolerated (2 × 20 mCi/m2), with an encouraging response rate in relapsed follicular NHL treated at this dose level (100% overall response, 10 of 11 having CR/CRu).68

Unconjugated epratuzumab was also found to have antitumor activity in patients, alone or when combined with rituximab,69-71 but it has not been active in animals.72 Because anti-CD20 antibodies are effective in animal models, and because there are encouraging clinical responses with 90Y-epratuzumab, we asked if 90Y-epratuzumab could be used in conjunction with anti-CD20 immunotherapy for improved activity over either agent alone. Such an approach would allow an effective radioimmunotherapeutic to be combined with a highly effective antibody therapy without concern that the radioconjugate's targeting would be compromised by the unlabeled antibody. We reported activity of a humanized anti-CD20 IgG, veltuzumab, in animal models that was similar to rituximab,72, 73 and early results from ongoing clinical studies have confirmed its activity in patients.74 Thus, the studies examined the combination of 90Y-epratuzumab with veltuzumab.

Because anti-CD20 therapy can also induce apoptosis or induce cell cycle arrest, which could enhance antitumor activity of a radioimmunoconjugate,55-59 there is a rationale for starting the anti-CD20 therapy in advance of RAIT. Clinically, starting treatment with the anti-CD20 antibody would reduce the B-cell sink that could potentially affect tumor uptake of the anti-CD22 radioconjugate, because it too would be targeted to normal B-cells. Although this is not a factor in mice, we decided the treatment regimen should start with the anti-CD20 IgG in advance of the anti-CD22 radioconjugate. We chose nude mice, because they are less sensitive to radiation than SCID mice, and used the Ramos human B-cell lymphoma cell line inoculated subcutaneously, because tumor growth was dependable and predictable. Once this tumor begins to grow, it progresses rapidly, and occasionally, tumors may spontaneously regress, but most often only after they have already grown to ≥2.0 cm3. Although we would have preferred to initiate veltuzumab several days in advance of the 90Y-epratuzumab treatment, because the tumors grew rapidly, veltuzumab was given only 1 day in advance of the 90Y-epratuzumab, when the tumors were first visibly apparent, with 1.0 mg and then 3 subsequent doses, each of 0.5 mg, given at 1-week intervals.75

Biodistribution studies confirmed that 1 mg of veltuzumab given 1 day in advance of an 111In-epratuzumab injection had no impact on tumor or other normal tissue uptake measured 3 days after the RAIT.75 Tumors given veltuzumab alone progressed as quickly as untreated animals (Fig. 3), but after an initial growth spurt over the first week, animals given a maximum dose of 90Y-epratuzumab alone experienced an initial partial (ie, ≥50% reduction in size) or even a complete (no measurable tumor seen) response. However, this effect was short-lived, and within a few days to a few weeks, most tumors progressed rapidly. In contrast, 80% of the animals given the veltuzumab treatment with 90Y-epratuzumab had a complete response and were tumor-free for more than 5 months, when the study was terminated. Another study showed a similar result when veltuzumab was delayed for 6 days after 90Y-epratuzumab treatment (Fig. 4). These results suggest that the radioimmunoconjugate was instrumental in first ablating growth of the tumor, at which point the effect of unlabeled antibody could be seen. Interestingly, the radioimmunoconjugate dose used in these studies causes severe hematologic toxicity that would require a minimum of 6 weeks before full recovery occurred, which could have compromised the number of effector cells available for antibody-dependent cellular toxicity (ADCC). Although the underlying mechanism of action behind the enhanced activity with the veltuzumab consolidation dose is not known, there are several possibilities including: (a) cell signaling (eg, apoptosis), (b) complement pathway, (c) effector cells capable of eliciting ADCC, perhaps in the tumor, were in sufficient number, (d) extending veltuzumab therapy 3-4 weeks after treatment may have allowed enough antibody to be present when effector cells were repopulated to exert ADCC activity.

Figure 3.

Combination radioimmunotherapy and antibody therapy using 90Y-epratuzumab and unlabeled veltuzumab. Treatment was initiated once the subcutaneous Ramos xenographs were visible in BALB/c nude mice. Treatment comprised 90Y-epratuzumab (90Y-E) alone, veltuzumab (V) alone (1 mg followed 1 week later with 3 more weekly injections of 0.5 mg), or 90Y-E combined with V therapy that was started 1 day before the radioconjugate. An irrelevant 90Y-anti-CEA IgG was given to assess nonspecific radiation effects.

Figure 4.

Veltuzumab antibody therapy enhances the response of 90Y-epratuzumab even when started 1 week after the radioconjugate treatment. Growth curves of the individual animals with subcutaneous Ramos xenografts treated with 90Y-epratuzumab alone or with veltuzumab therapy initiated either 1 day before (day -1) or 6 days after the radioconjugate dose.

Anti-CD20 Radioconjugate Therapy With Unlabeled Anti-CD20 Consolidation Therapy

With evidence that immunotherapy with veltuzumab anti-CD20 IgG can contribute to the antitumor response of radiolabeled anti-CD22 IgG, it was important to assess whether the addition of an anti-CD20 immunotherapy combined with an anti-CD20 radioconjugate enhances antitumor activity. Borrowing from the same treatment regimen used in combination with 90Y-epratuzumab, veltuzumab was initially configured for administration at 1.0 mg 1 day in advance of 90Y-veltuzumab, continuing with 3 sequential weekly doses of 0.5 mg. As shown previously in Figure 1, animals receiving a 1.0-mg predose before 90Y-veltuzumab, with no additional treatments, had no improvement in response over animals given only 90Y-veltuzumab. In this study, when the veltuzumab treatment regimen started with 1 mg given as a predose before the 90Y-veltuzumab therapy, there was no improvement in response over that seen with 90Y-veltuzumab alone, despite the continuation of additional veltuzumab injections. Thus, continued consolidation therapy with veltuzumab had no impact on response. However, if this same dosing regimen was shifted so that veltuzumab therapy started 1 week after the 90Y-veltuzumab dose (ie, no veltuzumab given before the 90Y-veltuzumab), a significantly improved median time to progression and several cures were observed (Fig. 5). These results suggest that the 1.0-mg predose may have reduced the radioimmunoconjugate uptake just enough to impede the ability of the consolidation therapy to “drive the therapeutic response to completion.” As shown earlier in Figure 2, if the veltuzumab predose was reduced to 0.25 mg, and the animals received a full consolidation treatment of veltuzumab (4 weekly doses; 1.0 mg + 3 × 0.5 mg) starting 1 week after the 90Y-veltuzumab treatment, we again observed a significant enhancement in antitumor activity. Thus, consolidation antibody therapy could be a very effective way to improve the objective response achieved with a radioimmunoconjugate therapy, as well as to extend the response duration. To achieve the best response, the radioimmunoconjugate must be given an opportunity to reach its highest possible concentration in the tumor, another reason to moderate the amount of an unlabeled antibody that can compete with the radioimmunoconjugate's binding to the tumor.

Figure 5.

Consolidation veltuzumab therapy improves response of 90Y-veltuzumab, but only when given in the absence of a predose, or as in Figure 2, when the predose is minimized. Ramos-bearing nude mice were given 0.13 mCi of 90Y-veltuzumab alone or animals received 1.0 mg of veltuzumab before the radioconjugate, which was then followed with 3 additional weekly injections of 0.5 mg of veltuzumab (pre & post hA20). Other animals received the same veltuzumab regimen, but it was started 1 week after the radioconjugate was given.

In retrospect, the initial clinical data showed predoses of about 1.0 to 1.5 mg/kg would have been sufficient to reduce splenic uptake and increase the radioimmunoconjugate's residence time in the blood, at least in patients without excessive tumor burden or splenomegaly. Instead, with both of the currently approved RAIT products, patients now receive 2 doses of unlabeled anti-CD20 IgG amounting to about 6.5 mg/kg (a total of 13 mg/kg) before the therapeutic radioimmunoconjugate is given. Perhaps in patients with large tumor burdens or with enlarged spleens, a substantial portion of this predose amount might be consumed by the antigen sink, leaving relatively small amounts of residual antibody in the blood to compete with the radioimmunoconjugate for tumor binding. However, given the current interest in moving RAIT to a frontline setting, and particularly in the situation now where 90Y-ibrituzumab is being used in a consolidation therapy after patients have already achieved a complete response,18 one needs to rethink how to best deliver the radioimmunoconjugate portion of this regimen. Minimizing the predose would be an important first step, but consideration should be given to examining other radioimmunoconjugates that do not compete with anti-CD20 antibody therapeutic responses, allowing each treatment to build a better response. Although animal data suggest that a consolidation treatment with an anti-CD20 IgG would be useful, adding consolidation therapy to the current treatment regimens would not likely bring much additional benefit because, as our data suggest, when a higher predose is given, additional consolidation treatments did not improve responses. Although we did not test more aggressive consolidation treatment regimens (eg, higher doses) in the animals given the 1.0 mg predose, under the current clinical procedures, patients already have a considerable load of unlabeled antibody, so that consolidation therapy would likely be best given several weeks later.

Pretargeted Radioimmunotherapy

Thanks in large part to the predose of the unlabeled anti-CD20 antibody, the radioimmunoconjugate has an extended half-life in the blood, which increases radiation exposure of the red marrow that contributes the dose-limiting hematological toxicity, and this, in turn, reduces the amount of the radioimmunoconjugate that can be given, thus reducing the amount of radiation delivered to tumor. Fortunately, NHL is very sensitive to radiation and impressive responses at the maximum dose allowed for each of the approved radioimmunoconjugates have been achieved. However, there is another approach that could be used to target radiation selectively, and which can deliver more radiation to the tumor with much less severe hematological toxicity.

Pretargeting procedures have been shown in many tumor models to outperform a directly radiolabeled antibody.76 Several pretargeting methods have been investigated, but each system is based on the fundamental principle that the slow antibody-targeting step should be separated from radionuclide targeting, which should be able to localize quickly in the tumor and then just as quickly be removed from the body.77 Our group has focused on a pretargeting system based on recombinant, humanized bispecific antibodies (bsMoAb) that subsequently bind a radiolabeled hapten-peptide, primarily to avoid the use of foreign immunogenic proteins, such as streptavidin, which effectively pretargets radiolabeled biotin.76 All of these pretargeting systems are so efficient at trapping the radionuclide that in some model systems, the radiolabeled effector is taken up by the tumor at a maximum level within an hour, and the percentage uptake of pretargeted hapten-peptide or other small molecules delivered to the tumor can equal that of a directly targeted IgG, yet the rapid clearance of the radiolabeled effector molecule from the blood and body greatly reduces radiation exposure.78-82

We showed a bsMoAb-pretargeting system based on veltuzumab and paired with a 90Y-labeled hapten-peptide had significantly enhanced antitumor responses as compared with 90Y-veltuzumab (Fig. 6).82 With pretargeting, not only did we increase the time to progression, but we were able to completely ablate a large number of the tumors, even at doses well below the maximum tolerated dose for this procedure; yet with the 90Y-veltuzumab direct conjugate, only a short temporary response could be achieved.53, 82 Because the recombinant humanized bsMoAb lacks the Fc-portion of the immunoglobulin, we next asked whether this pretargeting system could benefit from the addition of the veltuzumab IgG immunotherapy. Because it might also be necessary clinically to predose patients with unlabeled anti-CD20 IgG to allow better pretargeting with the bsMAb, we also determined how a predose given 1 day in advance of the bsMAb impacted this pretargeting procedure.

Figure 6.

Veltuzumab bsMoAb pretargeted 90Y-hapten-peptide improves therapeutic response compared with directly radiolabeled 90Y-veltuzumab. Nude mice bearing subcutaneous Ramos tumors were treated with the maximum tolerated dose of 90Y-veltuzumab (0.13 mCi/0.05 mg) or a pretargeting procedure using a humanized, recombinant bispecific antibody containing 2 veltuzumab Fab fragments and 1 anti-hapten Fab fragment capable of binding the 90Y-hapten-peptide. 0.5 or 0.7 mCi of the pretargeted 90Y-hapten-peptide-only results in ∼50%-70% loss in white blood cells (WBC) that recovers in several weeks, whereas 0.13 mCi of 90Y-veltuzumab causes >90% loss in WBC that requires 7 weeks to recover. Growth curves and the subsequent survival curve show the improved responses in the pretargeted animals with the ability to completely ablate 9 of 10 tumors at the 0.7 mCi 90Y-hapten-peptide dose.

Unlike the 90Y-veltuzumab direct conjugate whose antitumor activity was not affected significantly by a 1.0-mg veltuzumab predose, antitumor responses with the pretargeting procedure were seriously impaired.53 However, biodistribution studies indicated a higher loss (∼75%) in the tumor uptake of the radiolabeled hapten-peptide under these conditions. Additional consolidation therapy (3 × 0.5 mg) after the 1-mg predose had no effect, because without adequate uptake of the pretargeted 90Y-hapten-peptide, tumors progressed rapidly (Fig. 7A). When the predose was reduced to 0.25 mg, tumor uptake was 50% lower than when no predose was given, but antitumor responses showed some evidence of improvement over the pretargeting procedure in the absence of a predose (Fig. 7B and C). Consolidation anti-CD20 veltuzumab therapy further enhanced the therapeutic effects when a 0.25-mg predose was given (Fig. 7C).

Figure 7.

The effect of veltuzumab predosing and consolidation therapy on a veltuzumab-bsMoAb pretargeting procedure. (A) Administering 1.0 mg of veltuzumab 1 day before the injection of the veltuzumab bsMoAb reduces efficacy of the pretargeting procedure using 0.25 mCi of the 90Y-hapten-peptide, despite the use of additional veltuzumab therapy after the pretargeted 90Y-hapten-peptide was given. (B) Minimizing the veltuzumab predose to 0.25 mg restores the antitumor effects of the pretargeting procedure, even improving it in this study. (C) Adding a full veltuzumab consolidation therapy after the pretargeted therapy (ie, 1.0 mg veltuzumab starting 1 week after the 90Y-hapten-peptide, followed by 3 weekly doses of 0.5 mg veltuzumab) improves response, even when animals were predosed with 0.25 mg veltuzumab.

Efforts to assess the broader applicability of pretargeted radioimmunotherapy/veltuzumab consolidation therapy have been hampered by the difficulty to get a wide range of lymphoma cell lines to grow consistently in unconditioned nude mice. Although severe combined immunodeficient (SCID) mice are more receptive to tumor xenograft growth, they are far more sensitive to radiation than nude mice, requiring extensive reduction in the administered activity to avoid fatal events. SCID mice are often able to tolerate a pretargeting treatment of just 0.125 mCi. Animals bearing RL xenografts (0.6 cm3) experience a significant, yet modest improvement in survival over the untreated animals, but when combined with the veltuzumab consolidation therapy, the response is substantially and significantly enhanced (Fig. 8).

Figure 8.

Veltuzumab consolidation therapy enhances antitumor response in SCID mice bearing RL human lymphoma xenografts. SCID mice received 0.125 mCi of veltuzumab bsMoAb pretargeted 90Y-hapten-peptide alone, or they were given a 0.25 mg predose of veltuzumab 1 day before the bsMoAb injection, as well as receiving a full veltuzumab consolidation therapy (1.0 mg + 3 × 0.5 mg) starting 1 week after the 90Y-hapten-peptide (treatment-related toxicity in pretargeting alone group).

Conclusions

Radioimmunotherapy, an extremely effective treatment for follicular non-Hodgkin lymphoma, is currently underused. Data from new clinical trials may eventually overcome the reluctance for its use, and preclinical data are now revealing ways that this procedure can be improved, which could provide additional incentives for wider acceptance. Cancer is rarely treated with a single agent or administration, but usually involves combined modalities or agents given repeatedly. Because tumors rarely have unique signatures that allow exclusive uptake when using targeted therapies, in moving forward, when using a more toxic antibody conjugate, we need to learn how to best combine different strategies that foster the highest localization ratio of the immunoconjugate, and to ensure that they do so in concert, not in competition when both target the same antigen. By using antibodies that react with different antigens, we can induce responses by different mechanisms of action, thus providing additive or synergistic effects. Furthermore, pretargeting methods are now being recognized for their ability to improve radionuclide targeting with reduced hematological toxicity, making these procedures very promising in an environment where the current chemotherapeutic agents are associated with major hematopoietic toxicity. Because pretargeting can deliver the same radiation doses to tumor as direct RAIT while not having a commensurate level of myelotoxicity, this lends itself to combination therapies with cytotoxic drugs. Other issues of importance, and discussed elsewhere,83 include fractionating RAIT and the use of multiple treatment cycles. If radioimmunotherapy is to be used in a consolidation setting, then we will need to ensure that the radionuclide is well matched for the disease setting, involving either bulky tumors or minimal disease. Thus, targeting (or pretargeting) of radionuclides for the treatment of hematological malignancies remains a very promising modality, with several new opportunities to enhance efficacy.

Acknowledgements

We thank M. Jules Mattes, PhD (Center for Molecular Medicine and Immunology), Edmund A. Rossi, PhD (IBC Pharmaceuticals, Inc.), William J. McBride, PhD (Immunomedics, Inc.), and Chien-Hsing Chang, PhD (Immunomedics, Inc.) for collaborations and provision of certain critical reagents.

CONFLICT OF INTEREST DISCLOSURES

The articles in this supplement represent proceedings of the “12th Conference on Cancer Therapy with Antibodies and Immunoconjugates,” in Parsippany, New Jersey, October 16-18, 2008. Unrestricted grant support for the conference was provided by Actinium Pharmaceuticals, Inc., Bayer Schering Pharma, Center for Molecular Medicine and Immunology, ImClone Systems Corporation, MDS Nordion, National Cancer Institute, NIH, New Jersey Commission on Cancer Research, and PerkinElmer Life & Analytical Sciences. The supplement was supported by an unrestricted educational grant from ImClone Systems Corporation, a wholly owned subsidiary of Eli Lilly and Company, and by page charges to the authors. This work was supported in part by USPHS grant P01 CA103985 from the National Cancer Institute, NIH (DM Goldenberg). DM Goldenberg is an officer and shareholder of Immunomedics, Inc., and IBC Pharmaceuticals, Inc. RM Sharkey and H Karacay declare no financial conflicts.

Ancillary