Radioimmunotherapy of CD22-Expressing Daudi Tumors in Nude Mice with a ⁹⁰Y-Labeled Anti-CD22 Monoclonal Antibody

Radiolabeled monoclonal antibodies (MAb) represent promising new therapies for drug-refractory hematopoietic malignancies such as leukemias and lymphomas. This may be due to the radiosensitivity of lymphoid cells and the ability of antibodies to more readily access hematopoietic tumors compared to solid tumors. Radioimmunotherapy with ⁹⁰Y-labeled or ¹³¹I-labeled CD20 are reaching the mainstream as established therapies. The CD20-targeting ⁹⁰Y-based radiopharmaceutical, Zevalin, was the first Food and Drug Administration–approved radiolabeled radioimmunoconjugate. Although effective for lymphoma, CD20 is not broadly expressed on the less differentiated B cell leukemias, the most common childhood form of acute lymphocytic leukemia. Thus, MAbs recognizing other specificities must be considered. Anti-CD22 MAbs are candidates because studies show that these MAbs have therapeutic effects when clinically administered (1). Anti-CD22 immunotoxins have been used to successfully treat rare hairy cell leukemia (2), and have been used to treat B cell leukemia in mice (3) and in humans (4). Because CD22 is expressed in B cell leukemias, anti-CD22 is a logical choice for radioimmunotherapy of B-acute lymphocytic leukemia. Alternative agents for therapy of B-acute lymphocytic leukemia therapy are urgently needed because the American Cancer Society indicates that there are at least 2,000 new cases each year, and that chemotherapy-resistant disease is a frequent cause of treatment failure in leukemia patients (5). An anti-CD22–based radiopharmaceutical would be very useful.

CD22 is a 135-kDa B lymphocyte–specific glycoprotein and a member of the sialoadhesin family of molecules (6–8). It first appears at the late pro-B cell stage of B cell differentiation and is a key regulatory cytoplasmic protein that is coexpressed simultaneously with IgD on mature B cells (7). It is also expressed on 60% to 70% of B cell lymphomas and leukemias. The major function of CD22 is to regulate B cell responses, which is likely accomplished by recruiting key signaling molecules to the antigen receptor complex (9, 10). Experiments in knockout mice have established the importance of CD22 in modulating B cell responses in augmenting antibody responses, expanding peritoneal B-1 cell populations, and in increasing the levels of circulating autoantibodies (11–13).

For these studies, we chose ⁹⁰Y, which is a powerful β-emitting radionuclide widely accepted as a therapeutic targeting agent. ⁹⁰Y has a favorable maximum β energy of 2.3 MeV, a half-life of 2.7 days, and a short path length of 5 mm (14). Together, these features contribute to the well-known cross-fire effect of ⁹⁰Y. The fact that ⁹⁰Y, unlike ¹³¹I, has no gamma component means that the extensive precautions and isolation necessary for ¹³¹I administration are not needed for ⁹⁰Y administration, and in some instances, ⁹⁰Y-labeled antibodies are given on an outpatient basis.

Perhaps the single most important component of the radiolabeled antibody is the chelate, the isotope-binding molecule that is conjugated to the antibody. A stable chelate concentrates the therapy in the tumor site and prevents nonspecific irradiation of nontarget organs. The 1B4M-diethylenetriaminepentaacetic acid (DTPA)–based chelate was chosen for our studies because it is highly stable in vivo in the studies reported herein, and in our development of other radiopharmaceuticals (15, 16).

Studies previously indicated that CD22 served as a useful target for radioactive metals, but these studies were lacking in their evaluation of efficacy (17). In this report, we examined the anti-CD22 MAb, RFB4, as a vehicle for the delivery of the radionuclide ⁹⁰Y to B cell malignancies growing in nude mice. Radiolabeled anti-CD22 (Y22) displayed impressive and significant anticancer effects in the flank tumor model. Bone marrow studies indicated that the presence of CD22-expressing tumor might act to prevent radiolabeled antibody from entering the bone marrow compartment and destroying bone marrow. Comparisons of internalizing anti-CD22 to internalizing anti-CD19 and noninternalizing anti-CD45 revealed that all three ⁹⁰Y-labeled antibodies induced complete remissions and tumor-free, long-term survivors, arguing that both internalizing and noninternalizing antibodies can be equally effective for ⁹⁰Y radioimmunotherapy.

Monoclonal antibodies and cell lines. The anti-CD22 hybridoma, RFB4, which secretes mouse IgG₁, has been previously described and used in clinical studies as a targeted immunotoxin conjugated to the ricin toxin A chain (18, 19). 3A1e, a control IgG₁, is a murine pan-T cell MAb recognizing human CD7. This hybridoma was provided by Dr. Barton Haynes, Duke University, Durham, NC (20) and was used as a negative control because it does not bind to Daudi cells. MAbs were purified from hybridoma supernatants by affinity chromatography using a protein A column. Fractions of the eluted MAb were pooled, concentrated, diafiltered into buffer and stored at −70°C until use. For comparisons to anti-CD22, a noninternalizing anti-CD45 (AHN-12), mouse IgG₁ was similarly radiolabeled as reported by our group (15). Also, an internalizing anti-CD19 (HD37), mouse IgG₁, was previously reported (16).

The CD22⁺, IgM⁺ human Burkitt's lymphoma cell line, Daudi (21), the CD22⁺ Raji B cell line, the CD22⁻ human T cell leukemia, HPBMLT (22), and the CD22⁻ mouse C57BL/6 myeloid leukemia, C1498 (23) were obtained from the American Type Culture Collection, Rockville, MD, and maintained in RPMI 1640 containing 10% fetal bovine serum, 100 units/mL penicillin, 100 μg/mL streptomycin, and 100 mmol/L l-glutamine. The cell lines were incubated at 37°C in a humidified atmosphere of 5% CO₂ in air. Viability was determined by trypan blue exclusion and viabilities of ≥90% were required for using cells in our experiments (24).

Chelation of the antibody. 1B4M is a modified DTPA chelate (25) obtained from Dr. Martin Brechbiel, NIH. We have used this IB4M-DTPA crosslinker (referred to as IB4M) and have described its use in previous studies (15, 16). Briefly, in this anti-CD22 study, 1 to 10 mg/mL antibody, in a 0.05 mol/L carbonate buffer (pH 8.6), and 0.15 mol/L NaCl was conjugated with a 10-fold molar excess of 1B4M overnight at room temperature. The conjugate was separated from unconjugated 1B4M and transferred to the chelation buffer, followed by six washes with 0.16 mol/L ammonium acetate buffer (pH 7.0). The conjugation buffer and the labeling buffer were passed through a Chelex 100 column to remove any extraneous metals and the final protein concentration was determined spectrophotometrically (absorbance = 280 nm). Previous studies showed that conjugated antibody was robust and stable even after several days in human serum (15).

Labeling efficiency of conjugated anti-CD22 monoclonal antibodies. Conjugated antibody was labeled with 5 to 10 μCi of carrier-free ¹¹¹In Cl₃ (MDS Nordion, Kanata, Ontario) in chelation buffer and EDTA was added to the tube, vortexed, and incubated for 5 minutes. The chelation mixture was then diluted with 1% bovine serum albumin in PBS. The ratio of bound protein versus free radiometal chelate was determined by TLC on silica gel–coated glass fiber paper (ITLC-SG Pall Life Science, East Hills, NY) as previously described (15). Labeling efficiency was expressed as (cpm origin) / (cpm origin + cpm front) × 100.

Binding and immunoreactivity assessment. The immunoreactivity of labeled RFB4 was evaluated using an established binding assay (26). Briefly, Daudi or C1498 cells were washed and plated with radiolabeled MAbs. After incubation, the total and the cell-bound radioactivity were determined using a gamma counter. Data was plotted as the percentage of binding versus increasing cell number. Immunoreactivity (immunoreactive fraction) was determined by nonlinear regression curve-fitting by plotting the inverse of the bound fraction compared with the inverse of the cell concentration, which is based on the assumption that the total antigen concentration (cell concentration) represents an accurate approximation of the concentration of free antigen. This calculates the B_max or immunoreactive fraction. GraphPad Prism software (San Diego, CA) was used for these calculations.

In vivo tumor studies. Female athymic nude mice were purchased from the National Cancer Institute, Frederick Cancer Research and Development Center, Animal Production Area (operated by Charles River Laboratories, Hartford, CT) and housed in an Association for Assessment and Accreditation of Laboratory Animal Care–accredited specific pathogen–free facility under the care of the Department of Research Animal Resources, University of Minnesota. Animal research protocols were approved by the University of Minnesota Institutional Animal Care and Use Committee. Animals were housed in microisolator cages to minimize the possibility of transmission of any contaminating virus. Daudi cells (5 × 10⁶), in 0.1 mL PBS, were injected s.c. into the right flank of the nude mice. For biodistribution studies, mice with palpable tumors were given 7 μCi ¹¹¹In-labeled MAb (i.p.). On day 5, organs were harvested (15, 16). Blood, tumor, spleen, liver, lung, kidney, muscle and bone were counted in a Packard Cobra 5002 Auto-Gamma well counter. Data was calculated as the percentage of injected dose per gram of tissue.

For therapy studies, tumors were grown in female athymic nude mice in the same way. When the tumors could be visualized, two perpendicular diameters and the height of the tumor were measured using calipers. Tumor volumes were estimated as a product of the three measurements, using the formula for the volume of an ellipse (r1 × r2 × r3) (4/3) (π). Animals were randomly assigned to treatment groups and received i.p. injections of the specified doses of ⁹⁰Y-labeled MAb or control MAb when the tumors were ∼0.4 to 0.6 cm³. Mice were observed for visible toxic signs, weighed, and tumor dimensions were recorded every 2 days.

Histology. Tissue specimens of liver, kidney, and intestine were obtained from mice, and histology studies were done as described (27). All samples were embedded in optimum cutting temperature compound (Miles, Elkhark, IN), snap-frozen in liquid nitrogen, and stored at −80°C until sectioned. Serial 4 μm sections were cut, thaw-mounted onto glass slides, and fixed for 5 minutes in acetone. The slides were then stained with H&E.

For bone marrow studies, femora were decalcified, embedded in paraffin, and cut at 5 μm (28). Histologic sections involved full-length coronal sections of each femur. Routine H&E staining was done and each of four replicate sections were analyzed. Digital images were acquired using a Spot 2 digital CCD mounted on an Olympus model BX51 microscope.

Statistical analysis. Groupwise comparisons of data were done using Student's t test.

Binding of radiolabeled anti-CD22 to Daudi cells as determined by flow cytometry. To determine whether 1B4M conjugation affected the ability of Y22 to bind to CD22⁺ Daudi cells, chelated and nonchelated RFB4 were analyzed by immunofluorescence and flow cytometry. To perform these studies, RFB4 and RFB4-1B4M were adjusted to the same saturating concentrations and then added to the Daudi cells. Prior testing of 1:25, 1:50, and 1:100 dilutions indicated that the correct saturating concentrations were chosen for fluorescence-activated cell sorting studies. Next, an identical quantity of a FITC-labeled goat anti-mouse secondary IgG₁ antibody was added to the cells and indirect immunofluorescent studies showed that the degree of binding (>95%) and the histograms were identical for RFB4, RFB4-1B4M, and RFB4-1B4M plus nonradioactive indium, indicating that chemically altering RFB4 by attaching 1B4M did not alter the binding site of the MAb to target cells (data not shown).

Binding analysis. Scatchard analysis previously showed that the affinity constant of anti-CD22 was 2.1 × 10⁸ mol/L (29). To further study the binding and determine the immunoreactivity, RFB4-1B4M was labeled with ¹¹¹In and then reacted with Daudi cells or control CD22⁻ C1498 cells. Indium-111 is typically used as a surrogate for ⁹⁰Y showing differences of only 10% to 15% in biodistribution (30). Figure 1 shows that when the percentage of binding was plotted against cell number that labeled RFB4 bound well to Daudi cells, but did not bind to CD22⁻ C1498. The percentages of binding for 1, 2, 4, 8, 16, or 32 million Daudi cells was 10%, 17%, 29%, 47%, 74%, and 89%, respectively. Nonlinear regression analysis was done and the fitted curve is shown in Fig. 1. The immunoreactive fraction calculated from these data and projected to infinite antigen excess by the method of Lindmo et al. (26) indicated that >100% of the agent was immunoreactive with target cells. In contrast, the negative control C1498, had no binding activity.

Biodistribution of ¹¹¹In-anti-CD22 in nude mice with flank Daudi tumors. In order to determine whether ¹¹¹In-labeled anti-CD22 was capable of reaching a tumor in vivo, biodistribution experiments were done. Figure 2 shows a significantly (P < 0.03) higher mean activity (23.5% injected dose/g) in the tumors of five mice treated with ¹¹¹In-RFB4 after 5 days. In contrast, only 2.8% activity of the injected dose reached tumor in a group of mice injected with an identical dose of the control ¹¹¹In-3A1e. Significantly less (P < 0.02) ¹¹¹In-RFB4 activity compared with ¹¹¹In-3A1e activity was found in kidney indicating that more of the specific agent was diverted to the CD22-expressing tumor target. The level of activity was still high in the blood after 5 days (9.1%). A comparison of ¹¹¹In-RFB4 and ¹¹¹In-3A1e in the other organs revealed no significant differences.

The in vivo efficacy of ⁹⁰Y-anti-CD22 in mice with Daudi flank tumors. MAb was labeled with ⁹⁰Y in an identical manner as described in the ¹¹¹In experiments above. The labeled MAb was passed through a spin column and ITLC indicated that 98% of radioactivity was protein bound. The specific activities were 4.65 mCi/mg for Y22 and 7.8 mCi/mg for ⁹⁰Y3A1e. Nude mice (n = 6/group) were injected in their flanks with Daudi cells and when tumors were about 0.4 to 0.6 cm³, mice were given a single i.p. injection of either 157 μCi of Y22, 295 μCi of ⁹⁰Y 3A1e or were not treated. In untreated mice, Fig. 3 shows that tumors grew rapidly, exceeding 2 cm³ in 12 days. In mice receiving 157 μCi Y22, all the tumors initially, completely regressed. However, two of the six mice relapsed between days 35 and 50. The other four mice remained tumor-free beyond day 200 when the experiment was terminated. Histologic examination of these mice showed no evidence of tumor. Two of six control mice injected with 295 μCi of ⁹⁰Y 3A1e died of early toxicity, indicating that the maximum tolerated dosage had been exceeded. The remaining four mice all died with tumors by day 45. Also, we observed that 75 μCi Y22 was not protective and that 310 μCi was too toxic (data not shown).

In the group of six mice that were treated with high-dose Y22 (310 μCi), four of six died between days 8 and 19 with no evidence of tumor, indicating that death was due to acute toxicity. Tissues from one of the mice were collected when the animal was preterminal and indicated extensive damage to liver due to fatty degeneration and necrosis. Because we were only able to observe a single animal, it is difficult to assess whether this toxicity was directly or indirectly related to the Y22 treatment and we were unable to assess bone marrow toxicity at this time. Two of six of the mice receiving 310 μCi Y22 were long-term survivors and survived >200 days. Together, these findings indicated that the antitumor effect of Y22 was specific and dose-dependent.

A second experiment was done based on the findings on the first experiment (Fig. 4). Groups of nine mice with tumors ∼0.4 to 0.6 cm³ were treated with either 196 μCi ⁹⁰Y-RFB4 (Y22) or 197 μCi of ⁹⁰Y 3A1e. No toxic deaths were observed in mice treated in this experiment. Figure 4A shows the mean tumor growth for the mice. In mice treated with ⁹⁰Y 3A1e, the mean tumor growth initially slowed, but the tumor quickly grew again and all mice were withdrawn from the experiment by day 30. In contrast, tumors initially regressed in all mice treated with Y22. The bump in the curve around day 70 represents the mean tumor growths of two mice that relapsed and were removed from the experiment. The remaining seven mice were tumor-free for the remainder of the experiment, which was terminated on day 195. Values between the ⁹⁰Y 3A1e and the Y22 mice varied significantly (P < 0.01) when compared by Student's t test on days 18 to 24. Figure 4B shows the individual data for the mice. Together, the data from these two experiments showed that Y22 was highly effective and selective in treating Daudi flank tumors.

Histology. To provide some insight into whether there were early toxic effects of the 196 μCi Y22 or 197 μCi ⁹⁰Y 3A1e dose regimen, two randomly selected mice were removed from each treatment group from the second tumor study (experiment 2) on day 12 and histology studies were done. Thus, the experiment proceeded with nine mice per group. Histologic examination of kidney tissue sections from treated mice revealed no major damage. The glomeruli were intact (Fig. 5A) and there was no overt damage to the proximal tubules and no signs of infiltration. The kidney responds late to radiation damage and one might not expect to see early damage. However, there was no evidence of renal toxicity in the long-term survivors in the Y22 group, even when the experiment was terminated on day 200. In the liver, there was evidence of only mild fatty changes, but no perivascular infiltrate, no hepatocyte damage, and bile ducts were intact and nonoccluded (Fig. 5B). Histology studies were done on small intestines and villi were normal, elongated, with a normal complement of goblet cells (Fig. 5C). Necropsies revealed no gross abnormalities and therefore correlated with our histology findings. These data indicated that the 196 μCi dose regimen did not damage critical nontarget organs.

In contrast, pronounced differences were apparent in coronal cross-sections of femora examined from the Y22-treated and ⁹⁰Y 3A1e-treated mice that were given nearly identical doses of labeled antibody. Figure 6A shows a day 12 bone section from one of the two Y22-treated mice. The histology was indistinguishable from that of a normal nonirradiated mouse. The most striking feature of the Y22 section was that the hematopoietic cells including megakaryocytes and erythrocytes were normal. Healthy endothelial cells were clearly visible. Findings were similar for the second mouse. On the other hand, one section from the ⁹⁰Y 3A1e-treated mouse shown in Fig. 6B revealed a complete absence of hematopoietic cells including megakaryocytes after 12 days. Only a background of adipocytes, histiocytes, and endothelial cells were visible indicating pronounced radiologic injury with resulting bone marrow aplasia. The second ⁹⁰Y 3A1e-treated mouse showed similar results, but there was some hematopoiesis in the proximal region of the bone.

Comparison of radioimmunotherapy with anti-CD22, anti-CD19, and anti-CD45. In Fig. 2, biodistribution studies were tried with ¹¹¹In-anti-CD22. In the same experiment, anti-CD19 (HD37) and anti-CD45 (AHN-12) were also labeled with ¹¹¹In and tested for tumor distribution in vivo. Anti-CD22 and anti-CD19 are established internalizing antibodies and anti-CD45 does not internalize (31). Interestingly, findings were similar with all three MAbs in that all showed significant tumor localization after 5 days. Only the negative 3A1e control antibody did not localize to tumor. The slight differences in tumor accumulation of radioisotope that were observed directly correlated with the known avidities of the antibodies (RFB4 > AHN-12 > HD37).

To determine the comparative ability of the three radiolabeled antibodies to produce long-term tumor-free survivors, mice were given lethal flank tumors. Experiments (n = 5) were comparable because all consistently induced large (2 cm³) flank tumors 10 to 20 days after an injection of 5 million Daudi cells (Table 1). Also, the negative controls produced identical results in all experiments in that either untreated mice or mice given negative control ⁹⁰Y 3A1e had no antitumor effect. In contrast, all three labeled antibodies, anti-CD22, anti-CD19, and anti-CD45, induced complete regressions of all tumors at mid and high doses. These treatments all produced tumor-free survivors in the majority of treated mice. No single antibody treatment produced antitumor results that were remarkably different from any of the others. Taken together, the data indicate that significant antitumor effects were obtained using any of the antibodies that reacted with Daudi, but not antibodies that were nonreactive. Despite whether the antibodies could or could not internalize, all three antibodies showed a similar ability to target tumor in vivo and induce antitumor responses. Because at least one toxic death occurred in the high-dose group for all the antibodies, it seemed that a maximum tolerated dose was reached.

The major findings to emerge from these studies is that Y22 given i.p. is highly effective in eliminating lethal tumors from Daudi-infected nude mice. A control-labeled antibody was not. Previous studies indicated that MAb labeled with radioactive yttrium showed potential (17, 32), but these were primarily biodistribution and not radioimmunotherapy studies. In this study, we analyzed efficacy and despite the fact that these were aggressive, established Daudi flank tumors, all tumors completely regressed after treatment with mid to high dosages of Y22 and most of the mice were long-term, tumor-free survivors. Although our results were improved by increasing the dose, we were approaching toxic dosages at or near the 300 μCi dose level and the therapeutic index was narrow in these mice. One must consider the limitations of the model in these evaluations because xenograft models are highly artificial and we are administering Y22 in a murine environment where human CD22 is not expressed on normal B cells. Despite this, the results were impressive against such a rapidly growing tumor.

These studies were also novel in that simultaneous bone studies were done because of the concerns of using ⁹⁰Y which has historically been considered a bone-seeking isotope (33, 34). Importantly, a clear difference in bone histology of Daudi tumor–bearing mice was observed in mice that received Y22 in comparison with the mice that received control ⁹⁰Y 3A1e. Bone marrow damage in the form of a complete absence of hematopoietic cells was severe in the control ⁹⁰Y 3A1e-treated mice and nearly nonexistent in the Y22-treated mice. Because the outcome of the tumor experiments revealed that Y22 selectively targets CD22-positive tumor cells and protects, and ⁹⁰Y 3A1e does not, it is likely that the CD22⁺ tumor load reduced bone marrow–related radiation damage. This may argue in favor of using a high avidity antibody such as RFB4 with ⁹⁰Y for selectively targeting tumor cells in humans because less radiolabeled antibody may nonspecifically collect in the bone marrow. This also implies that CD22⁺ leukemia cells, as well as CD22⁺ normal B cells, will bind Y22 and reduce bone marrow damage. However, a fraction of cells in the bone marrow are CD22⁺ cells (35) and leukemia cells are commonly found in the bone marrow in afflicted patients. Thus, the effects of these cells will have to be considered. Also notable from these studies is the fact that ⁹⁰Y 3A1e treatment caused more injury in the distal region of the bone as compared with the proximal region where the blood flow is greater.

Another unique aspect of these studies was our comparison of anti-CD22 radioimmunotherapy to radioimmunotherapy with anti-CD19 and anti-CD45 MAbs. When labeled in an identical manner and compared in the same experiment, all three antibodies similarly localized in tumor. Significantly more radiolabel was selectively delivered to tumor in comparison to the negative control 3A1e antibody. There was a correlation between antibody avidity and the amount of radioisotope delivered to tumor, although all antibodies delivered well. All antibodies were remarkable in producing complete remission at mid and high doses and most of these mice became tumor-free, long-term survivors. The clinical use of ⁹⁰Y-anti-CD45 will be much different than Y22. The anti-CD45 agent targets all hematopoietic cells including the majority of leukemias/lymphomas. Although more broadly reactive, CD45 is also expressed on stem cell progenitor cells and thus this approach must include stem cell infusion/bone marrow transplantation. The Y22 approach presented in this study offers an alternative therapy that does not have to be used in combination with a bone marrow transplant. This would offer treatment options to those patients who are not eligible for aggressive bone marrow transplant protocols.

Some believe that internalizing determinants are not as desirable for radiotherapy, particularly because earlier studies with iodinated antibodies showed that they are metabolized quickly with a subsequent release of low molecular mass targets from the cell (36, 37). Studies by Press et al. comparing radioiodinated noninternalizing anti-CD45 and internalizing anti-CD19 indicate that noninternalizing antibodies such as anti-CD45 are excellent choices for ¹³¹I radioimmunotherapy because they are resistant to intracellular degradation and deiodination whereas internalizing antibodies are not (31). Our data indicate that regardless of whether we used internalizing anti-CD22 and anti-CD19 or noninternalizing anti-CD45, ⁹⁰Y radioimmunotherapy was highly effective in inducing an anticancer effect. Our data support the contention that radioactive metals are retained intracellularly (38–41) and that radiometals such as ⁹⁰Y that are known to be residualized, particularly when combined with an internalizing target, result in a beneficial clinical effect. Our visual analysis of bone, in conjunction with the impressive efficacy observed in these studies, further support the fact that ⁹⁰Y delivered with internalizing antibodies localized in its intended target (tumor) as opposed to nontarget tissue (bone, liver, kidney).

Our group is most interested in alternative therapies for drug-refractory leukemia. CD20 is ontogenically expressed later than CD22 (42). Whereas CD20 is expressed on only a subpopulation of precursor B cells, CD22 is expressed on all precursor B cells. We already know that RFB4 reactivity is highly B lineage–restricted because in a panel of >40 human tissues, it recognized only B cells (29, 43, 44), and is highly expressed on leukemias. Thus, anti-CD22 is a better candidate for treating leukemia than anti-CD20. Several successful studies now indicate that CD20 targeting is highly effective for lymphoma. Unfortunately, diseases such as pre–B cell leukemia are less likely to express CD20, hence, alternative radiolabeled antibodies are urgently needed.

The primary purpose of these studies was to assess the efficacy of ⁹⁰Y-labeled anti-CD22 in an established animal therapy model because only a small number of efficacy studies regarding ⁹⁰Y-labeled anti-CD22 administration have been published. A secondary goal was to compare ⁹⁰Y-labeled anti-CD22 to ⁹⁰Y-labeled anti-C19 because CD19, another highly internalized B cell marker, is also widely expressed on B leukemia cells. Previous studies from our laboratory showed that ⁹⁰Y-labeled anti-CD19 antibodies were highly effective for targeting Daudi cells in this same animal model (16). No clinical studies using anti-CD19 ⁹⁰Y radioimmunotherapy have been published, so whether it would be better to target CD22 or CD19 is unknown.

Interestingly, studies argue in favor of targeting CD22 and CD19 simultaneously (19). Preclinical studies with a combination of both anti-CD19 and anti-CD22 antibodies labeled with toxins or with radionuclide indicate that the combined use of these agents may have pronounced advantages over the use of single-agent therapy (3, 45). Recent studies with recombinant antibody fragments in which anti-CD22 and anti-CD19 sFv are engineered on the same molecule indicate that the future may yield important molecules which can be used for bispecific targeting, and can be bioengineered to address important shortcomings such as toxicity, immunogenicity, and rapid clearance (46).

Others are using MAbs of murine origin for radioimmunotherapy. For example, ¹³¹I-labeled Tositumomab is a murine anti-CD20 MAb and used for therapy of follicular lymphoma (reviewed in ref. 47). Short-term toxic effects are mild, including immediate infusion reactions, moderate myelosuppression, and an influenza-like reaction, which are managed on an outpatient basis. Although radiation from the conjugate may pose risks to those in physical contact with the patient immediately posttreatment, straightforward protocols for the safe outpatient administration of drug can be followed. Long-term toxic effects such as hypothyroidism occurring in ∼10% of patients are easily managed. No cases of myelodysplasia have been noted, but only a small number of patients have been treated thus far. No serious infections have been observed. Normal B lymphocytes were only temporarily depleted, and there was no evidence of an effect on overall antibody levels. The approach is promising because an effective systemic treatment for disseminated follicular lymphoma can be completed entirely within a few weeks on a convenient outpatient basis with modest toxicity. The clinical results obtained with this anti–B cell murine monoclonal antibody encourage the clinical testing of Y22 which would have the added advantage of use against less differentiated B cell malignancies such as B-acute lymphocytic leukemia.

In the future, if these molecules are to become mainstream therapies, important issues will need to be addressed. Anti-CD22 is a mouse antibody and this will likely mean a rapid clearance when administered to humans. If they do show clinical promise, then chimerization or humanization to improve clearance may be a desirable option. This will also reduce their immunogenicity. However, it is important to keep in mind that immunogenicity may not be a problem in patients because a highly effective radiolabeled antibody against B cells could suppress the production of neutralizing antibodies by killing highly radiosensitive B cells. Also, by the very nature of their treatment, leukemia patients will likely be highly immunosuppressed from prior chemotherapy. Regardless, RFB4, has been chimerized and is available for testing.

The 1B4M chelate was chosen for these studies because, in the case of metals such as ⁹⁰Y, it is critical that the chelate retain the radiometal and not permit its release from the complex in vivo (reviewed in refs. 48, 49). Earlier generation of C-functionalized DTPA compounds (50) were improved by adding a methyl group to the structure of the base compound producing the M-DTPA producing a more stable configuration (25). This agent has already done well in preclinical and clinical radioimmunotherapy trials using ⁹⁰Y (51–54) and is similar to the chelate used in Zevalin. Brechbiel et al. have developed a simple, dependable, and reproducible process for the preparation of 1B4M-DTPA in large clinical scale batches (25). Currently, an Investigational New Drug–approved study of an ⁹⁰Y-labeled anti-CD45 antibody is under way at our institution using this chelate and performed optimally, consistently yielding high labeling efficiencies and immunoreactive fractions in clinical labelings. In our studies, as in others (30), ¹¹¹In was used as a surrogate marker for ⁹⁰Y because studies by Carrasquillo et al. showed that the differences in biodistribution were only 10% to 15%.

1,4,7,10-tetra-azacylododecane −N,N′,N″,N′″-Tetraacetic (DOTA)-based linkers have been successfully used in other studies (55). Interestingly, these studies also showed that anti-CD22 MAbs can have independent lymphomacidal properties and complement the activity of radiolabeled antibodies.

In summary, the high affinity, internalizing Y22 is highly effective against human tumors in a mouse xenograft model. These studies indicate that when a residualizing radiometal is combined with a highly internalizing antibody on the same radioconjugate, the result is a high level of distribution of the drug to tumor and an impressive antitumor response. This warrants further consideration of Y22 for clinical trials.

We thank Michael Elson for excellent technical assistance in these studies.