Phase I Therapy Study with ¹⁸⁶Re-labeled Humanized Monoclonal Antibody BIWA 4 (Bivatuzumab) in Patients with Head and Neck Squamous Cell Carcinoma¹

Squamous cell carcinoma, the predominant histological type among tumors of the head and neck, accounts for approximately 5% of all malignant tumors in Europe and the United States. In 2000, an estimated 551,100 new cases of HNSCC³ of the oral cavity, pharynx, or larynx were diagnosed worldwide (1). Early stages (stage I/II) of HNSCC generally have a good prognosis after surgery or radiotherapy. Unfortunately, this does not hold true for advanced-stage disease (stage III/IV) as present in about two-thirds of all HNSCC patients. Despite improvements in locoregional treatment modalities, the rate of locoregional recurrence still is near 40%, whereas approximately 25% of these patients develop distant metastases. With respect to the latter, autopsy studies have shown incidences up to 57% (2, 3, 4, 5). In particular, patients with multiple lymph node metastases are at risk for the development of locoregional recurrences and distant metastases (6, 7). Considering these figures, it is plausible that many of the advanced-stage HNSCC patients harbor residual tumor cells after surgery and radiotherapy. The role of adjuvant chemotherapy for this group of patients has not proven to be beneficial, and therefore the development of an effective adjuvant systemic treatment is a major challenge.

Selective targeting of radionuclides to HNSCC by use of MAbs as in RIT might contribute to a more effective treatment of advanced-stage disease. Because HNSCC is relatively radiosensitive compared with other solid tumors, RIT can become a realistic option. To date, most successes in the field of RIT have been achieved in patients with hematological malignancies, such as non-Hodgkin’s lymphomas, where high rates of partial and complete responses have been observed (8). In the treatment of B-cell non-Hodgkin’s lymphoma, most experience has been gained with iodine-131- and yttrium-90-labeled anti-CD20 MAbs such as tositumomab and ibritumomab, respectively (9, 10, 11, 12, 13).

These successes have motivated us to explore the additional value of RIT against HNSCC. In the selection of a suitable target antigen in HNSCC, CD44 splice variants containing the v6 domain appeared to be the most promising thus far. The CD44 protein family consists of isoforms, encoded by standard exons and up to nine alternatively spliced variant exons (v2–v10), which are expressed in a tissue-specific way and involved in physiological functions such as signal transduction, growth factor binding, cell adhesion, and cell migration. Expression of v6-containing CD44 variants has been related to aggressive behavior of various tumor types and was shown to be particularly high and homogeneous on the outer cell surface of HNSCC (14, 15).

In the past decade, the development of CD44v6-directed RIT in HNSCC has shown encouraging progress. For this purpose, two mMAbs became available, designated mMAb U36 and mMAb BIWA 1. Although BIWA 1 resembles U36, it binds to a different epitope with a 35-fold higher affinity (16). According to the numbering of Kugelman et al. (17), the epitope recognized by BIWA 1 consists of amino acids 360–370, and the epitope recognized by MAb U36 consists of amino acids 365–376.

Biodistribution studies with radiolabeled MAbs U36 and BIWA 1 in HNSCC patients undergoing surgery showed promising biodistribution patterns with selective tumor targeting and high tumor uptake for both MAbs (18, 19). Unfortunately, administration of BIWA 1 resulted in HAMA responses in 11 of 12 patients (19), a problem also observed after injection of mMAb U36, although to a lesser extent (18).

Subsequent Phase I RIT studies with chimeric (human/mouse) MAb U36 showed favorable targeting and (although not a primary objective) promising antitumor effects (20, 21). However, chimerization of MAb U36 had not resulted in a decrease of the immunogenicity because 5 of 12 patients showed development of HACAs (20). Formation of HAMA or HACA should be eliminated to prevent allergic reactions and rapid clearance of the radioimmunoconjugate, especially when multiple administrations of the radioimmunoconjugate are anticipated. The engineering of hMAbs, being almost completely human apart from a small part of the antigen-binding sites, became the next approach to deal with this problem. In several studies, hMAbs showed excellent tumor targeting in combination with low or no immunogenicity (22, 23).

To reduce immunogenicity, two humanized BIWA versions were developed, called BIWA 4 and BIWA 8. The intermediate-affinity MAb BIWA 4 was selected for further clinical evaluation rather than the high-affinity MAb BIWA 8 because of its superior tumor targeting in HNSCC-bearing nude mice (16).

In the present study, RIT with ¹⁸⁶Re-labeled hMAb BIWA 4 was evaluated in 20 head and neck cancer patients with regard to safety, MTD, pharmacokinetics, immunogenicity, and preliminary therapeutic potential. Three of these patients received a second dose, which was the same or slightly different from the first dose.

Patients with clinical evidence of recurrent locoregional and/or metastatic HNSCC, for whom no curative option was available, were candidates. A histologically confirmed HNSCC in the past was required for inclusion. Other eligibility criteria were an age between 18 and 80 years, a Karnofsky performance status above 60, and a life expectancy of at least 3 months. An interval of at least 4 weeks from the last chemotherapy or radiotherapy was required, as well as good recovery from prior treatment. Patients were excluded if their WBC count was less than 3000/μl, if their granulocyte count was less than 1500/μl, if their platelet count was less than 100,000/μl, if the bilirubin level was greater than 30 μm (upper limit of laboratory normals, 20 μm), or if the serum creatinine concentration was greater than 150 μm (upper limit of laboratory normals, 110 μm). Other exclusion criteria were pregnancy, life-threatening infection, severe allergic condition, organ failure, hematological disorders, recent myocardial infarction on electrocardiogram or unstable angina pectoris, congestive heart failure, and bronchial asthma. Also, premenopausal women were excluded when not surgically sterile or not practicing acceptable means of birth control. The study was reviewed and approved by the Institutional Ethics Committees of the VU University Medical Center (Amsterdam, the Netherlands) and UMC Nijmegen (Nijmegen, the Netherlands). All patients gave written informed consent after receiving a thorough explanation of the study.

mMAb BIWA 1 (Boehringer Ingelheim, Vienna, Austria), the parental MAb of BIWA 4, was generated by immunizing BALB/c mice with glutathione S-transferase fusion protein containing the human CD44 domains v3–v10 (14). The epitope recognized by BIWA 1 has been mapped to the v6 domain of CD44. Homogenous expression of v6-containing CD44 isoforms has been observed in squamous cell carcinoma of the head and neck, lung, skin, esophagus, and cervix, whereas heterogeneous expression was found in adenocarcinomas of the breast, lung, colon, pancreas, and stomach. In normal tissues, expression has been found in epithelial tissues such as skin, breast, and prostate myoepithelium and bronchial epithelium (24).

Generation, production, and characterization of BIWA 4 have been described previously (16). In short, mRNA was isolated from the BIWA 1 hybridoma cell line (Boehringer Ingelheim) using the QuickPrep mRNA Purification Kit (Pharmacia, Uppsala, Sweden). cDNA from heavy and light chain variable regions was generated by reverse transcription-PCR. Fragments were cloned into the TA cloning vector pCR II (Invitrogen, Groningen, the Netherlands) and sequenced. Two expression vectors derived from the plasmid pAD CMV1 (25) were constructed, carrying the constant regions of human γ-1 and of the human κ light chain, respectively. Subsequently humanized versions of the BIWA 1 heavy and light chain variable regions (generated by CDR grafting) were cloned in front of the immunoglobulin constant regions of the above-mentioned expression vectors, resulting in the hMAb BIWA 4.

BIWA 4 was stably expressed in Chinese hamster ovary cells, whereas high producers were selected as described previously (16). Cultivation was performed in proprietary serum-free medium TH-7 (Boehringer Ingelheim, Biberach an der Riss, Germany). Bulk substance from fermentation runs (titer after harvest, 90 mg/liter) was purified according to a generic downstream process for MAbs consisting of ultra/diafiltration, affinity chromatography on protein A-Sepharose, microwave treatment at 80°C, nanofiltration, ultra/diafiltration, anion exchange chromatography on Q-Sepharose, and ultra/diafiltration. The bulk substance was formulated in PBS (pH 7.4) containing 0.02% Tween 20, filter sterilized, and dispensed aseptically into vials. Each vial contained 5 ml of BIWA 4 at a concentration of 5 mg/ml.

For coupling of ¹⁸⁶Re to BIWA 4, the chelate MAG3 was used as described previously (26). ¹⁸⁶Re and MAG3 were obtained from Mallinckrodt, Inc. (Petten, the Netherlands). The mean specific activity of ¹⁸⁶Re at the time of labeling was 0.31 ± 0.07 mCi/nmol. After synthesis of ¹⁸⁶Re-MAG3, an esterification with tetrafluorophenol was performed. Subsequently, the ester was purified and conjugated to BIWA 4. Radiolabeled BIWA 4 was purified on a PD10 column (Pharmacia-Biotech, Woerden, the Netherlands) with 0.9% NaCl containing 5 mg/ml ascorbic acid (pH 5) as the eluent. Finally, the conjugates were filter sterilized. These procedures result in a sterile final product with endotoxin levels of <5 endotoxin units per ml, as demonstrated in five preceding validation runs. The molar ratio of ¹⁸⁶Re-MAG3 to BIWA 4 was always <4. The radiochemical purity of the conjugates as assessed by high-performance liquid chromatography and thin-layer chromatography was always >95% (mean, 98 ± 1.3%). After each preparation of ¹⁸⁶Re-BIWA 4, the immunoreactivity was determined by measuring binding to 0.1% paraformaldehyde-fixed cells of the HNSCC cell line UM-SCC-11B as described previously (16). The immunoreactive fraction of the ¹⁸⁶Re-BIWA 4 preparations, as determined by a modified Lineweaver-Burk plot, ranged from 81% to 100% (mean, 90 ± 6.9%).

The medical history and physical condition of all patients were examined, and routine laboratory analyses were performed, including complete blood cell counts, serum electrolytes, urine sediment, liver enzymes, and renal and thyroid functions. An electrocardiogram and a radiograph of the thorax were obtained. For all patients with locoregional tumor recurrence, a CT and/or MRI scan of the head and neck region was performed before treatment as has been described previously (18), and in cases with distant metastases, a CT scan of the region of interest was obtained. CT and/or MRI scans were also obtained at 6 weeks after administration for evaluation of RIT efficacy. To this end, perpendicular diameters of tumors were assessed by an experienced radiologist in both centers. In case of stable disease, follow-up assessments were, if possible, performed every 4 weeks.

For dosimetry, tumor volume was assessed by drawing regions of interest on consecutive CT or MRI slices with a VoxelQ system (Picker International, Highland Heights, OH), as described previously (20). This system enables three-dimensional reconstruction and automated volume calculation.

RIT was performed with a starting activity dose of 20 mCi/m², followed by escalation with 10-mCi/m² increments of ¹⁸⁶Re-labeled BIWA 4. All patients received 50 mg of BIWA 4, a dose that was found to result in optimal tumor uptake and tumor:nontumor ratios in a preceding MAb dose-finding study in which doses of 25, 50, and 100 mg BIWA 4 were evaluated (27). A complete pretreatment assessment was performed before the study, and blood samples were taken for pharmacokinetic analyses. Monitoring of vital functions (blood pressure, pulse rate, temperature, and breathing rate) was performed at the screening visit; before infusion; at 10, 60, 120, and 240 min after infusion; and 6 weeks after infusion. Patients who developed an objective response or stable disease after the first dose of ¹⁸⁶Re-BIWA 4 were eligible for a second administration. They underwent the same visit schedule as for the first administration. As a safety precaution, all repeated administrations were conducted at least 3 months after the first administration, and the same inclusion/exclusion criteria were used as for the first dose.

After treatment, patients were admitted for approximately 21 h in a special treatment room at the Department of Nuclear Medicine, and thereafter they stayed an additional 3 days in a single room at the Department of Otolaryngology/Head and Neck Surgery. Dose rates (in μSv/h) were measured 1 h after administration at a distance of 100 cm with a γ-radiation dose rate counter [at VU University Medical Center, Berthold LB 1230 (EG&G, Wildbad, Germany); at UMC Nijmegen, Automess 6150 AD 2 (Automess GmbH, Ladenburg, Germany)]. The patients were discharged 4 days after administration of ¹⁸⁶Re-BIWA 4, and the set of pretreatment laboratory analyses was repeated weekly for at least 6 weeks. The severity of toxicities was graded according to the National Cancer Institute CTC, version 2.0. The MTD was defined as the dose level at which grade 4 hematological or grade 3 nonhematological toxicity developed in not more than one of six patients.

Tumor response was assessed according to the WHO guidelines using physical examination or radiographic criteria. Stable disease was defined as a decrease of <50% or an increase of <25% of the sum of the perpendicular diameter products and no appearance of new lesions, which had to persist for at least 4 weeks. Progression was defined as a ≥25% increase in the size of one or more lesions or the appearance of a new lesion.

Simultaneously acquired planar anterior and posterior whole-body scans were obtained within 1 h after administration of the radioimmunoconjugate; 21, 48, 72, and 144 h after administration of the radioimmunoconjugate; and, if feasible, for up to 2 weeks after administration of the radioimmunoconjugate. For all scintigraphic studies, a large field-of-view gamma camera [at VU University Medical Center, Dual Head Genesys Imaging System (ADAC Laboratories, Milpitas, CA); at UMC Nijmegen, Siemens Multispect 2 (Siemens, Erlangen, Germany)] equipped with low-energy high-resolution parallel-hole collimators and connected to a computer system [at VU University Medical Center, Pegasys (ADAC Laboratories); at UMC Nijmegen, ICON version 7.1 (Siemens)] was used. Camera quality control was performed at each imaging session. With the aliquot retained from the radioimmunoconjugate preparation for injection, a weighed dilution of the injected patient dose was prepared as standard, put in an Adams phantom, and measured simultaneously during whole-body imaging. Lateral, anterior, and posterior planar images of the head and neck region were acquired at 21, 48, 72, and 144 h. SPECT images were acquired at 72 h. Acquisition parameters for planar and SPECT images were as described previously (18), with a Hanning filter used for postfiltering (cutoff at 1 cycle/cm). In each of the study centers, all images were interpreted by one experienced examiner, who was unaware of the site of recurrence and/or the presence of distant metastases.

Dosimetry to assess dose delivery to whole body, bone marrow, organs, and tumor was performed as described previously (20). Patients 5 and 17 were excluded from dosimetric analysis due to insufficient data. In short, whole-body scintigraphy was used to draw regions around organs and sites of interest. Residence times in these organs and sites were calculated and entered in the MIRDOSE3 computer program, version 3.1 (Oak Ridge Associated Universities, Oak Ridge, TN). If CT/MRI scanning before treatment with ¹⁸⁶Re-BIWA 4 showed a measurable tumor lesion, and if this lesion was visible on scintigraphy after RIT, dosimetry of the tumor was performed. The red marrow dose was calculated using a blood-derived method as described previously (20, 28).

Serial blood samples were taken from a peripheral vein of the arm opposite to the infusion site at the following time points: preinfusion; at the end of the infusion; and at 5 and 30 min and 1, 2, 4, 16, 21, 48, 72, and 144 h after administration of ¹⁸⁶Re-BIWA 4. Preliminary results obtained from the first four patients showed that blood sampling at additional time points in the second week after administration would improve the characterization of the ¹⁸⁶Re-BIWA 4 pharmacokinetics. Therefore, for patients 5–20, blood samples were also taken at 240 and 336 h after administration. Plasma samples were prepared for determination of the concentration of immunoreactive BIWA 4 by validated ELISA methods, essentially as described previously (19). In this sandwich-type ELISA, microtiter plates were absorptively coated with the fusion protein GST-CD44v6. In a first step, BIWA 4 in the plasma sample bound to the immobilized antigen CD44v6. After incubation, unbound BIWA 4 and plasma components were removed by washing. In a second step, rabbit antihuman IgG labeled with horseradish peroxidase (DAKO, Copenhagen, Denmark) was used for detection of bound BIWA 4, and tetramethyl-benzidine (Kirkegaard & Perry, Gaithersburg, MD) was used for detection of enzyme activity.

In addition, whole-blood and serum samples were prepared and counted in a gamma counter [at VU University Medical Center, 1470 Wizard (Wallac, Turku, Finland); at UMC Nijmegen, 1480 Wizard (Wallac)], and radioactivity levels were expressed as %ID/kg. Background activity and decay were corrected for, and the %ID/kg was determined by comparison with an aliquot retained from the conjugate preparation for injection.

Evaluation of the immunogenicity of the BIWA 4 radioimmunoconjugate was conducted by determination of human anti-hMAbs in serum samples taken before administration of ¹⁸⁶Re-BIWA 4 and 1 and 6 weeks after administration, using a validated bridging ELISA. In short, BIWA 4 antibodies labeled with the chelator MAG3 were immobilized onto microtiter plates. Antibodies to BIWA 4-MAG3 present in the serum sample bound simultaneously to the solid phase and to biotinylated BIWA 4-MAG3 that was added at the same time in a fixed amount. Streptavidin coupled to horseradish peroxidase (Boehringer Mannheim, Mannheim, Germany) was used for signal generation by the peroxidase/tetramethyl-benzidine system. Due to the lack of a positive human anti-BIWA 4 serum, which would be desirable as a standard, the validation of the present assay was done with a rabbit antiserum from an animal that was previously immunized with nonradioactive ¹⁸⁵Re-MAG3-BIWA 4. The assay only enabled the relative measurement of the HAHA titer in equivalents of the anti-BIWA 4 titer of the reference rabbit serum. For an easier handling of data, the dilution factors of the rabbit reference serum were converted into Boehringer Ingelheim titer units. The titer unit was defined as dilution of the rabbit reference serum (1:x) multiplied by 18,000. Assay performance during the study was assessed by back-calculation of calibration standards, tabulation of the standard curve fit function parameters, measurement of quality control samples, and construction of a precision profile. Triplicate determinations in three randomly assigned wells of a microtiter plate were used for unknown samples, calibration standards, and quality controls. The lower and upper limits of determination of the assay were 0.180 and 60 Boehringer Ingelheim titer units (1:100,000 to 1:300 dilution of the rabbit reference serum). A serum sample was considered HAHA positive only if all of the following criteria were met: (a) the measured titer was above the lower limit of determination/quantification; (b) the titer in the postinjection sample was at least two times higher than that in the preinjection sample; (c) the sample could be serially diluted with a strong correlation between titer and dilution factor; (d) the titer could be blocked by addition of BIWA 4; and (e) the titer could not be blocked by a nonspecific IgG MAb.

Twenty patients (14 males and 6 females; age range, 42–73 years; mean age, 58 years) entered this study. Their characteristics are shown in Table 1. Thirteen patients had received surgical treatment of their disease at an earlier stage. All patients had been treated previously with radiotherapy, and some of them had also been treated with chemotherapy. Two patients were treated at the lowest dose level of 20 mCi/m², four patients were treated at 30 mCi/m² (one patient more than planned because of death of one of the patients during follow-up), three patients were treated at 40 mCi/m², six patients were treated at 50 mCi/m², and five patients were treated at 60 mCi/m². Three patients received a second administration of 50 mCi/m^{2 186}Re-BIWA 4 (Table 2).

Administration of ¹⁸⁶Re-BIWA 4 was tolerated well, and no side effects were observed directly after and during the first few days after treatment. No changes occurred in the physical parameters during the first few days after administration, and no changes in blood parameters indicated toxicity of organs such as the liver or kidneys. Nonhematological toxicity defined as drug-related CTC grade 3 or 4 occurred in four patients (Table 2). One patient (patient 5) treated at the 30-mCi/m² dose level was admitted to the hospital because of unexplained seizures 5 days after treatment. During hospitalization, Quincke’s edema or angioedema caused respiratory distress 8 days after the start of RIT, which was treated with infusion of prednisolone and inhalation of bronchodilatators. After treatment for low serum sodium levels and treatment with antiepileptic drugs, the patient left the hospital again 11 days after the start of RIT. She was found dead at home 22 days after start of treatment. No signs indicating a relationship between antibody administration and death were noticed, but the possibility of a relationship could not be excluded. Permission for autopsy was not given by the relatives. Analysis of pre- and posttreatment serum samples showed absence of HAHA. Because this patient was not evaluable for hematological toxicity and antitumor response, an extra patient was treated at the 30-mCi/m² dose level.

Febrile neutropenia occurred in patients 13 and 16 (both CTC grade 3) and in patient 17 (CTC grade 4), all of whom were treated at the 60-mCi/m² dose level. All three patients received empirical antibiotic and antimycotic treatment until WBC counts had recovered and the patients were afebrile. CTC grade 3 fatigue occurred in patient 17 at approximately 4 weeks after administration of 60 mCi/m^{2 186}Re-BIWA 4. The fatigue was most likely related to the aggravation of the pre-existing anemia. Main nonhematological toxicity consisted of mild (maximum CTC grade 2) mucositis (Fig. 1) in two of the six patients treated at the 50-mCi/m² dose level and in three of the five patients treated at the 60-mCi/m² dose level. This form of toxicity is most likely drug related because the CD44v6 target antigen is known to be expressed in normal mucosa. The mucositis usually started 1–2 weeks after administration of the conjugate and disappeared 2 weeks after onset. Most other toxicity was not related to the conjugate, except for reversible loss of taste in one patient, observed after both injections of ¹⁸⁶Re-BIWA 4.

For patients treated at the MTD level, mean radiation dose rate measured 1 h after injection at a distance of 100 cm was 4.9 μSv/h. These dose rates would result in cumulative doses far less than the annual limit of 2 mSv, which, according to Dutch guidelines for radionuclide therapy, is considered to be acceptable for medical personnel and family members.

The bone marrow appeared to be the dose-limiting organ. A significant decrease of platelet and WBC counts was observed for patients treated at the highest dose levels, with a nadir of 4–5 weeks after injection of the conjugate (Fig. 2). Hematological dose-limiting toxicity defined as drug-related CTC grade 4 was reported in three patients (Table 2). Reversible CTC grade 4 leukocytopenia and granulocytopenia were observed in two patients (patients 13 and 17) treated at 60 mCi/m², at 4 and 5 weeks after injection, respectively. In addition, the anemia present before entering the study worsened slowly in patient 17 and finally resulted in a CTC grade 4 anemia, which was treated with blood transfusions. Reversible CTC grade 4 thrombocytopenia and grade 4 leukocytopenia were observed in one patient (patient 18) treated at 50 mCi/m², who received concomitant treatment with methotrexate for rheumatoid arthritis. Clinical symptoms of thrombocytopenia in this patient consisted of mild petechiae. Patient received a total of two pooled donor units of platelets on day care basis. All three patients with grade 4 hematological toxicity received antibiotic and antimycotic treatment until WBC counts had recovered and the patient was afebrile. MTD was established at 50 mCi/m². Nadirs of each individual patient are listed in Table 2. Drug-related thrombocytopenia and leukocytopenia occurred about 21 days after start of RIT, and the patients started recovering from these conditions before 6 weeks postinjection.

Whole-body scans obtained directly after administration of ¹⁸⁶Re-BIWA 4 showed mainly blood-pool activity, which decreased over time. Representative whole-body scans are shown in Fig. 3. At later intervals, a homogenous distribution was observed in the lungs, liver, spleen, and kidneys, whereas relative uptake at tumor sites gradually increased. In general, no selective accumulation at nontumor sites was visible except for uptake in squamous epithelia (oral mucosa, nasal mucosa, and skin), testes, feces, and urine (bladder). Planar and SPECT images obtained showed improved delineation of small tumors in the head and neck region (see Fig. 4). However, visualization of distant metastases, concerning mainly lung metastases, was in most cases hampered by the presence of blood-pool activity and the small size of the lesions.

Concentrations of ¹⁸⁶Re-BIWA 4 were assessed by measuring immunoreactive BIWA 4 in plasma with an ELISA assay as well as by measuring radioactivity in serum. Because each patient received BIWA 4 as a 50-mg infusion, the immunoreactive BIWA 4 concentrations are presented as a group to demonstrate the consistency of data across patients (Fig. 5,A). In a similar way, the normalized (%ID/kg) radioactivity data are presented (Fig. 5,B). Overlaying the immunoreactivity and radioactivity data demonstrates that plasma BIWA 4 concentrations and radioactivity serum concentrations track each other well, indicating that the ¹⁸⁶Re-BIWA 4 conjugate remained stable after i.v. injection. This graphical consistency between plasma BIWA 4 concentrations and serum radioactivity concentrations was observed for each patient and within the three patients receiving ¹⁸⁶Re-BIWA 4 twice (Fig. 6). However, modeling of the data revealed that the geometric mean serum half-life for the radioactive portion of ¹⁸⁶Re-BIWA 4 was significantly longer than the plasma half-life of BIWA 4 portion: 126.5 ± 36.3 versus 95.1 ± 15.9 h (P < 0.001). Differences noted in the modeling occurred at late time points, when most of the radioactivity was already eliminated or physically decayed. These data might indicate that a small proportion of injected radiolabeled BIWA 4 loses binding capacity at late time points after injection, by which detection in ELISA becomes impossible.

Twenty-one patient studies [5 female, 13 male (of which 3 patients received two administrations)] could be used for dosimetry. The mean whole-body absorbed self-dose was 1.18 ± 0.10 cGy/mCi for males and 0.95 ± 0.25 cGy/mCi for females. The mean red marrow dose was 1.82 ± 0.11 cGy/mCi for males and 2.35 ± 0.10 cGy/mCi for females. The normal organ with the highest absorbed dose appeared to be the kidneys (mean dose in males, 5.95 ± 2.78 cGy/mCi; mean dose in females, 9.62 ± 3.47 cGy/mCi, which is not expected to lead to renal toxicity). The doses delivered to the tumors (n = 15), recalculated to the MTD level of 50 mCi/m², ranged from 3.8 to 76.4 Gy for tumors ranging in size from 285.9 to 5.1 cm³. The median absorbed dose was 12.4 Gy. The largest tumor (285.9 cm³) showed the lowest absorbed dose (3.8 Gy), whereas the two tumor localizations with the highest absorbed dose (65.1 and 76.4 Gy) appeared to be small lesions (11.5 and 5.1 cm³, respectively). However, no clear correlation was found between tumor absorbed dose and tumor size.

HAHA analyses were evaluable for all patients. Two of the 20 patients (patients 10 and 17) showed a HAHA response (Table 3). For both patients, the titer of the week 6 sample was elevated and regarded as HAHA positive.

All patients treated with doses of up to 40 mCi/m^{2 186}Re-BIWA 4 showed progressive disease upon RIT. One patient (patient 5) was not evaluable for tumor size measurements, due to intercurrent death (see “Safety and Nonhematological Toxicity”).

Three of the six patients (patients 10, 11, and 12) treated with 50 mCi/m^{2 186}Re-BIWA 4 developed stable disease after the first administration (Table 4). As a result, patients 11 and 12 received a second administration, and progressive disease was observed after a total of 18 weeks (patient 11) and 21 weeks (patient 12) after first administration of the trial drug.

One patient (patient 14) treated with 60 mCi/m^{2 186}Re-BIWA 4 experienced stable disease after the first treatment. Progressive disease was observed after a second treatment with 50 mCi/m^{2 186}Re-BIWA 4, approximately 24 weeks after the first administration of the trial drug.

The overall time to progression for patients who did not respond to therapy ranged from 0 to 55 days, with a mean of about 5–6 weeks, irrespective of the treatment group. The survival time of the patients who did not obtain stable disease upon ¹⁸⁶Re-BIWA 4 treatment ranged from 21 to 219 days. For the four patients with stable disease, the survival time ranged from 217 to 403 days.

RIT with ¹⁸⁶Re-BIWA 4 seems to be safe, and no severe side effects were observed in this Phase I study besides dose-limiting myelosuppression at a ¹⁸⁶Re dose level of 60 mCi/m². Furthermore, mild mucositis was observed at the highest ¹⁸⁶Re dose levels of 50 and 60 mCi/m². Mucositis is most probably caused by binding of ¹⁸⁶Re-BIWA 4 to the CD44v6 target antigen as present in oral mucosa. The MTD of ¹⁸⁶Re-BIWA 4 was established at 50 mCi/m², which seems to be higher than the 27 mCi/m² found in a previous Phase I trial with the other anti-CD44v6 conjugate ¹⁸⁶Re-cMAb U36 (20). In the trial with ¹⁸⁶Re-cMAb U36, no mucositis became apparent. Because BIWA 4 and cMAb U36 have a comparable biological half-life in blood, differences in MTD cannot be explained by differences in pharmacokinetics. We believe that the large increments of the radiation dose-escalation steps used in the ¹⁸⁶Re-cMAb U36 Phase I trial can partly explain the difference in MTD and that the use of smaller dose-escalation steps would most likely have resulted in a higher MTD for ¹⁸⁶Re-cMAb U36. Also, the heterogeneity of the HNSCC patient population might have contributed to this difference.

Of great importance is the low immunogenicity of this humanized MAb directed against the CD44v6 antigen. Ten percent of the patients treated in the present study developed HAHAs. In the meantime, 28 more patients have been treated with 25–100 mg of radiolabeled BIWA 4 in three parallel radioimmunoscintigraphy/biodistribution studies with HNSCC, non-small cell lung cancer, and breast cancer patients. HAHA responses were observed in none of these patients, resulting in an overall HAHA response rate of 4% for all four studies. This is considerably lower than the 40% HACAs as found with the cMAb U36 (20) and the 90% HAMAs as found with the parental mMAb BIWA 1 (19). None of the three patients who received a second administration developed HAHAs, which illustrates the possibility for repeated treatment with BIWA 4.

Humanization of other MAbs has been performed with variable success. For instance, hMAb M195 elicited no immune response in a study group of 14 patients (23), compared with 37% HAMAs found with mMAb M195 (29). Furthermore, hMAb BrE-3 showed 14% HAHAs (22), compared with mMAb BrE-3, which showed immunogenicity in 46.7% (30) and 83% (31), respectively. On the other hand, hMAb A33 elicited HAHAs in 26 of 41 patients (32). Although significantly lower as compared with its parental mMAb A33, for which HAMAs were found in all treated patients (33), this example illustrates that humanization does not always solve the problem of immunogenicity.

A secondary objective in this Phase I study was to determine the therapeutic effects of RIT with ¹⁸⁶Re-BIWA 4. Stabilization of tumor growth was observed in three of six patients treated at the MTD level, with durations ranging from 6 to 21 weeks. This is consistent with efficacy results obtained in previous Phase I RIT trials with ¹⁸⁶Re-cMAb U36. In a first Phase I trial with ¹⁸⁶Re-cMAb U36, stable disease was observed in one of six patients treated at MTD. In a second Phase I trial with ¹⁸⁶Re-cMAb U36, reinfusion of granulocyte colony-stimulating factor-stimulated unprocessed whole blood was used to reduce myelotoxicity and to increase the MTD (21). In this procedure, granulocyte colony-stimulating factor is administered s.c. at home during 5 days before the start of RIT. On day 0, just before administration of ¹⁸⁶Re-labeled MAb, 1 liter of whole blood is harvested and kept unprocessed at 4°C until reinfusion after 72 h. Indeed, by using this facile procedure, the administered dose could be increased from 27 to 54 mCi/m², without exceeding grade 3 myelotoxicity and grade 2 mucositis. In this particular study, stable disease was observed in five of nine patients, for a period ranging from 3 to 7 months. All three patients treated at the highest dose level showed stable disease. Whether such a procedure for further dose escalation is also applicable for ¹⁸⁶Re-BIWA 4 remains to be established. It might well be that a second dose-limiting toxicity, most probably mucositis, will become manifest. Furthermore, it has to be considered that all HNSCC patients included in the Phase I RIT trials described above had a history of external beam irradiation, which had caused severe mucositis in some of the patients.

The observation of antitumor effects in patients with bulky disease offers opportunities for further development of RIT with single or multiple doses of ¹⁸⁶Re-labeled BIWA 4 in an adjuvant setting. The tumor absorbed doses, recalculated to MTD level, ranged from 3.8 to 76.4 Gy, whereas the median absorbed dose was 12.4 Gy. The tumor with the highest absorbed dose of 76.4 Gy appeared to be a small lesion of 5.1 cm³, which remained stable in size upon RIT (patient 10). It is unlikely that antitumor effects are caused by immune modulating effects because BIWA 4 lacks antibody-dependent cell-mediated cytotoxicity or complement-dependent cytotoxicity-mediating activity in vitro. Recently, we reported on the relationship between HNSCC size and MAb accumulation (34). Data for this report were obtained from several radioimmunoscintigraphy and biodistribution studies with the anti-HNSCC MAbs E48 and U36 in HNSCC patients. MAb uptake in small-volume tumors (1 cm³) was approximately 4 times higher than uptake in tumors with a large volume (>50 cm³). Therefore, in our view, these data justify evaluation of RIT with ¹⁸⁶Re-BIWA 4 as a systemic adjuvant treatment.

In conclusion, this study shows that RIT with ¹⁸⁶Re-BIWA 4 is safe. By using BIWA 4, a humanized anti-CD44v6 MAb, the rate of HAHA responses was reduced to a minimum, and repeated administrations appeared possible. The MTD was established at 50 mCi/m², at which dose level stabilization of disease was observed in three of six inoperable HNSCC patients with bulky disease. The application of RIT with ¹⁸⁶Re-BIWA 4 therefore holds promise, especially in an adjuvant setting.

We thank F. G. van Schaijk, M. Siegmund, and M. J. W. D. Vosjan for radiolabeling support and pharmacokinetic determinations; Dr. U. Kunz for pharmacokinetic and HAHA analyses; Prof. J. A. Castelijns and Dr. F. Joosten for CT and MRI examinations; M. van der Vlies for supervision on radiation safety issues; and Drs. E. F. I. Comans, R. Pijpers, P. G. H. M. Raymakers, M. A. W. Merkx, C. M. L. van Herpen, and Prof. P. C. Huijgens for clinical support.