Downloaded from
`
`http://jnci.oxfordjournals.org/
`
` by Abby Johnson on October 23, 2014
`
`(26) ISHIKAWA M, DENNIS JW, MAN S, ET AL: Isolation and characterization
`of spontaneous wheat germ agglutinin-resistant human melanoma mu-
`tants displaying remarkably different metastatic profiles in nude mice.
`Cancer Res 48:665-670, 1988
`(27) ISHIKAWA M, FERNANDEZ B, KERBEL RS: Highly pigmented human
`melanoma variant which metastasizes widely in nude mice, including
`to skin and brain. Cancer Res 48:4897-4903, 1988
`(28) YAGEL S, PARHAR RS, JEFFREY JJ, ET AL: Normal non-metastatic human
`trophoblast cells share in vitro invasive properties of malignant cells.
`J Cell Physiol 136:455-462, 1988
`(29) KHOKHA R, DENHARDT DT: On the use of anti-sense RNA: Down-
`regulation of mRNA encoding a metalloproteinase inhibitor. Anti-
`cancer Res 7:653-660, 1987
`(30) KHOKHA R, WATERHOUSE P, YAGEL S, ET AL: Modulation of the colla-
`genase inhibition (TIMP) mRNA levels by anti-sense RNA influences
`the invasiveness of mouse 3T3 cells. J Cell Biochem 12B:307, 1988
`(31) CARMICHAEL DF, SOMMER A, THOMPSON RC, ET AL: Primary structure
`
`and cDNA cloning of human fibroblast collagenase inhibitor. Proc
`Natl Acad Sci USA 83:2407-2411, 1986
`(32) KHOKHA R, WATERHOUSE P, YAGEL S, ET AL: Antisense RNA-induced
`reduction in murine TIMP levels confers oncogenicity on murine 3T3
`cells. Science 243:947-952, 1989
`(33) BARNER M, MAYER M, SCHIRRMACHER V, ET AL: Involvement of both
`heparanase and plasminogen activator in lymphoma cell mediated
`degradation of heparin sulfate in subendothelial extracellular matrix.
`J Cell Physiol 128:209-306, 1986
`(34) NICOLSON GL, WINKELHAKE JL: Organ specificity of blood-bome
`metastases as determined by cell adhesion. Nature 255:230-232, 1975
`(35) MURRAY JC, LIOTTA LA, RENNARD SI, ET AL: Adhesion characteristics
`of murine metastatic and non-metastatic tumor cells in vitro. Cancer
`Res 40:347-351, 1988
`(36) REICH R, THOMPSON EW, IWAMOTO Y, ET AL: Effects of inhibitors
`of plasminogen activator, serine proteinases, and collagenase IV on
`the invasion of basement membranes by metastatic cells. Cancer Res
`48:3307-3312, 1988
`
`Phase I Study of an Anti-Breast Cancer
`Immunotoxin by Continuous Infusion:
`Report of a Targeted Toxic Effect
`Not Predicted by Animal Studies
`
`Bruce J. Gould, Michael J. Borowitz, Eric S. Groves, Peter W.
`Carter, Douglas Anthony, Louis M. Weiner, Arthur E. Frankel*
`
`260F9 Monoclonal antibody-recombinant ricin A chain, an
`immunotoxin reactive with «50% of breast carcinomas, was
`given by continuous iv infusion at a dose of 50 ng/kg per
`day or 100 /ug/kg per day. Five patients with refractory
`breast cancer received treatment for from 6 to 8 days.
`Severe toxic effects, including marked fluid overload and
`debilitating sensorimotor neuropathies, occurred in most
`patients. Immunoperoxidase studies suggested that 260F9
`monoclonal antibody targeting of the Schwann cells may
`have induced demyelination and subsequent neuropathy.
`This is the first report of a targeted toxic effect due to an
`immunoconjugate. [J Natl Cancer Inst 81:775-781, 1989]
`
`Animal studies have demonstrated several problems that
`limit the efficacy of immunotoxins, particularly native RTA
`conjugates. These problems include rapid clearance by the
`liver (3), the generation of anti-immunotoxin antibodies (4),
`and poor tumor penetration (5). Rapid immunotoxin clear-
`ance is caused in part by the Kupffer cells, which bind the
`mannose moieties of native RTA. Although chemical degly-
`cosylation does not eliminate hepatic uptake, Blakey et al.
`(6) have shown in mice that serum levels of the partially
`deglycosylated RTA immunotoxins are sustained at higher
`levels than those formulated with native RTA. Recently, re-
`combinant RTA, which is not glycosylated but still has the
`same potency as the native molecule, has become available.
`Because recombinant ricin A chain (rRA) lacks carbohydrate
`
`Immunotoxins are a class of biological agents that have re-
`cently been used as cancer therapy in phase I trials. Immuno-
`toxins consist of two components: a carrier protein, usually a
`monoclonal antibody (MAb), and a cytotoxic agent such as
`ricin toxin A chain (RTA), diphtheria toxin, or Pseudomonas
`exotoxin. Ideally, the toxin is transported to the tumor cell
`surface by the antibody, which has specificity for a tumor
`antigen. Once bound to the cell surface, the complex is inter-
`nalized, the toxin escapes to the cytosol, and toxin-mediated
`inhibition of protein synthesis occurs. Several reviews detail-
`ing immunotoxin cytotoxicity have recently been published
`(A2).
`
`Received February 2, 1989; accepted March 1, 1989.
`B. J. Gould, P. W. Carter, A. E. Frankel, Department of Hematol-
`ogy/Oncology, Duke University Medical Center, Durham, NC.
`M. J. Borowitz, D. Anthony, Department of Pathology, Duke University
`Medical Center, Durham, NC.
`E. S. Groves, Cetus Corporation, Emeryville, CA.
`L. M. Weiner, Department of Medical Oncology, Fox Chase Medical
`Center, Philadelphia, PA.
`We thank Dr. Ira Pastan and Dr. Mark Willingham for giving us data
`that suggest that the 260F9 MAb binds to human nerve.
`*Correspondence to: Dr. Arthur E. Frankel, 050 CARL Bldg., Duke Uni-
`versity Medical Center, Durham, NC 27710.
`
`Vol. 81, No. 10, May 22, 1989
`
`ARTICLES 775
`
`IMMUNOGEN 2031, pg. 1
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`
`Downloaded from
`
`http://jnci.oxfordjournals.org/
`
` by Abby Johnson on October 23, 2014
`
`groups, it also has the theoretical advantage of being less im-
`munogenic than the native protein. When the processes that
`inactivate the immunotoxin are minimized, higher and more
`sustained immunotoxin levels should be achieved and better
`tumor penetration may become possible.
`The limited experience with immunotoxins in humans has
`been exclusively with native RTA containing immunotox-
`ins that have been administered by bolus infusions
`(7-9).
`Toxic effects seen in these studies in animal models have
`been largely predictable, and have consisted of fever, malaise,
`fluid retention, myalgias, hypoalbuminemia, electrocardio-
`gram changes, and hepatitis. All the toxic effects resolved
`when therapy was discontinued. Unfortunately, clinical re-
`sponses have been uncommon, with a complete response oc-
`curring in only two of 30 patients reported in the literature
`(7,9). In our study of patients with breast cancer, we report
`the first human use of an immunotoxin that contains rRA and
`the first study that uses continuous-infusion therapy. In addi-
`tion, we also report the occurrence of a unique, posttherapy
`toxic effect that was not predicted by primate trials.
`
`Patients and Methods
`Patients
`
`Five patients with biopsy-proven metastatic breast carci-
`nomas that were refractory to standard therapies were se-
`lected for treatment. A summary of patient characteristics,
`including sites of tumor involvement, hormonal receptors,
`and prior therapies is found in table 1. No patient received
`treatment for at least 1 month prior to the start of immuno-
`toxin therapy. All had a Karnofsky performance status of
`>80% and a life expectancy of >3 months. There was no ev-
`idence of cardiac, hepatic, renal, or neurologic dysfunction,
`nor were there any signs of an active infection. No patient
`had any major laboratory abnormalities.
`
`Immunotoxin Preparation
`
`The IgG, murine MAb 260F9, rRA, and the 260F9 MAb-
`rRA were produced by the Cetus Corporation, Emeryville,
`CA. A series of MAbs directed against breast cancer cell
`lines and breast carcinoma extracts were produced by stan-
`
`dard hybridoma techniques (10). The 260F9 MAb was cho-
`sen because of its affinity for «=50% of the breast cancer
`cell lines and for its limited cross-reactivity with normal tis-
`sues ( / / ). Genetic engineering techniques were used to man-
`ufacture sufficient quantities of pure, unglycosylated RTA.
`By using 2-iminothiolane, we introduced sulfhydryl groups
`onto the MAb (12). The derivatized antibody was treated
`with DTNB [5,5'-dithio-bis-(2-nitrobenzoic acid)] and mixed
`with the rRa, which resulted in coupling of the molecules
`by a disulfide bond. The preparation was processed by size
`exclusion and affinity chromatography to remove the small
`amounts of free antibody and RTA. Clinical lots of 260F9
`MAb-rRA were prepared and tested for sterility, rabbit py-
`rogenicity, and murine virus contamination and found to be
`acceptable.
`
`Protocol
`
`All patients were treated in the Clinical Research Unit at
`Duke University Medical Center from June 1987 through
`August 1987. After we obtained informed consent, four pa-
`tients received a single, 8-day course of treatment. The only
`exception to this schedule was patient 5, who received 6 days
`of therapy. The immunoconjugate was diluted in 50 mL of
`5% human serum albumin and administered by continuous
`infusion through a central venous catheter. Patients 1, 2, and
`3 received 50 /ug of immunotoxin/kg per day for total daily
`doses of 2.8, 3.0, and 3.02 mg, respectively. Patient 4 and
`patient 5 received 100 /ug of immunotoxin/kg per day for
`total daily doses of 7.5 and 6.8 mg, respectively. Following
`completion of treatment, the patients were monitored in the
`hospital for an additional 48 hours. Two patients had biopsy
`specimens of chest wall tumors done 48 hours after com-
`pleting therapy to obtain tissue for immunotoxin penetration
`analyses. After treatment, the patients were followed regu-
`larly at Duke for 3 months; thereafter, they were followed
`by contact through their community physicians.
`All patients were monitored closely for side effects by daily
`physical examinations, frequent vital signs, daily weights, and
`daily fluid balances. Also, serial measurements were made
`of the cbc, electrolyte, creatinine, BUN, total protein, albu-
`min, triglyceride, cholesterol, calcium, and liver profile. Chest
`
`Table 1. Patient characteristics
`
`Estrogen receptor/
`progesterone receptor*
`
`Previous treatmentt
`
`Sites of disease
`
`— /-
`
`- /-
`
`- /-
`
`- /-
`
`CMF, tamoxifen.
`C.W. XRT
`CMF, doxorubicin.
`mitomycin, vinblastine.
`tamoxifen, C.W. XRT
`CAF, tamoxifen,
`C.W. XRT
`CAF, ABMT, tamoxifen.
`C.W. XRT
`CMFVP, megestrol,
`doxorubicin
`
`Chest wall, lungs
`
`Chest wall
`
`Right pleural disease.
`liver, cervical nodes
`Chest wall, lungs
`
`Chest wall, lungs,
`supraclavicular nodes
`
`Age
`(yr)
`
`69
`
`57
`
`77
`
`43
`
`55
`
`Patient No.
`
`1 2 3 4 5
`
`*— = absence of receptor.
`tC = cyclophosphamide; M = methotrexate; F = 5-fluorouracil; C.W. XRT = chest wall radiation therapy; A = doxorubicin; ABMT = autologous bone
`marrow transplantation; V = vincristine; P = prednisone.
`
`776
`
`Journal of the National Cancer Institute
`
`IMMUNOGEN 2031, pg. 2
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`
`Downloaded from
`
`http://jnci.oxfordjournals.org/
`
` by Abby Johnson on October 23, 2014
`
`using the avidin-biotin peroxidase complex (ABC) procedure
`as described (14). To determine 260F9 antigen expression,
`we tested serial dilutions of 260F9 MAb against known pos-
`itive breast carcinomas, and found the optimal concentration
`of antibody to be 25 jug/mL. We used irrelevant murine IgG!
`at 25 /xg/mL as a negative control.
`To determine if immunotoxin administered to patients had
`bound to the target, we used a modification of the ABC
`procedure. To detect bound mouse immunoglobulin, we in-
`cubated tissues from the patients with biotinylated horse
`anti-mouse immunoglobulin, followed by the avidin and
`biotinylated horseradish peroxidase complex. To detect
`bound ricin, we incubated tissue sections with rabbit anti-
`RTA followed by biotin-conjugated goat anti-rabbit im-
`munoglobulin and ABC. Prior to testing the patient speci-
`mens, we determined optimal dilutions of each reagent in
`a checkerboard fashion with known 260F9 antigen-positive
`breast tumor sections that had been preincubated in vitro
`with immunotoxin.
`
`Results
`Patient Toxic Effects
`Four patients completed 8 days of the infusion, and pa-
`tient 5 requested that treatment be discontinued on day 6
`because of side effects. The 260F9 MAb-rRA shared many
`adverse effects observed with other immunotoxins (table 2).
`All the patients suffered fevers >38.4 °C, anorexia, malaise,
`arthralgias, and myalgias. Nausea and diarrhea were tran-
`sient problems for only one patient. Another patient devel-
`oped evidence of a drug reaction on day 6 when she de-
`veloped an erythematous-vesicular rash on her hands, finger
`arthritis, and a 7% eosinophilia. All the aforementioned side
`effects resolved when therapy was stopped.
`Fluid retention and peripheral edema were problems com-
`mon to all patients with maximal weight gains of 1 -5 kg oc-
`curring on day 8. Although several patients complained of
`dyspnea, none had chest x-ray evidence of congestive heart
`failure. Moderate-dose diuretics (Lasix, 60 mg) failed to re-
`verse the fluid retention, but the fluid gain did resolve in
`three patients when the immunotoxin was stopped. The other
`two patients required hospitalization 5 and 7 days after im-
`munotoxin treatment for pulmonary edema and symptomatic
`peripheral edema that was unassociated with acute cardiac
`diseases. The fluid overload of these two patients responded
`quickly and completely to moderate-dose diuretics, and they
`had no further volume problems.
`
`Table 2. Side effects of 260F9 MAb-rRA infusion
`
`Range
`
`>38 °C
`1.0-5.0 kg
`
`————
`
`Systolic blood pressure <
`
`Symptoms
`
`No. of
`patients
`
`5553111 1
`
`Fever
`Weight gain/edema
`Arthralgi as/myalgias
`Dyspnea on exertion
`Finger arthritis
`Rash
`Hypotension
`Nausea and diarrhea
`
`x rays, electrocardiograms, and chest and abdominal com-
`puter tomography scans were performed prior to study, as
`needed during study, and within 1 month after the comple-
`tion of therapy.
`Blood samples were obtained for pharmacokinetic studies,
`anti-rRA antibody levels, and anti-260F9 MAb levels. The
`samples were drawn contralateral to the infusion site prior to
`infusion, daily, and 0.5, 1, 2, 2.5, 4, 6, 12, and 18 hours after
`the start of the infusion and on the final day of therapy. Addi-
`tionally, blood was drawn at 32 and 48 hours and from 14 to
`32 days posttreatment. The blood samples were centrifuged
`within 2 hours of collection, the serum was removed, and the
`samples were stored at —70 °C.
`Pharmacokinetics
`Serum levels of 260F9 MAb-rRA were measured by an
`immunoradiometric assay. First, polyvinylchloride wells of
`a microtiter plate were incubated overnight with purified
`rabbit anti-RTA [10 pg/mL in phosphate-buffered saline
`(PBS)]. The wells were then washed three times with PBS
`and blocked for 1 hour with heat-inactivated fetal calf serum.
`Afterwards, patient sera, diluted 1:10 in PBS, were added
`to the wells. Various concentrations of the immunotoxin
`for use in formulating a standard curve were diluted in
`normal serum and added to the wells. All samples were
`tested in triplicate, and the plate was incubated for 2 hours.
`After another washing with PBS, a solution of 125I-sheep
`F(ab')2 anti-mouse immunoglobulin (DuPont, Wilmington,
`DE) in PBS with 2.5% heat-inactivated fetal calf serum
`was added to each well and incubated. The solution was
`formulated so that 3 X 106 cpm were added to each well.
`After 1 hour and a final washing, the wells were counted
`in a Packard Autogamma 5780 with the window settings
`for 125I. Nonspecifically bound radioactivity was determined
`from the control sera and subtracted before the immunotoxin
`concentrations were extrapolated from the standard curve.
`Anti-Murine and Anti-rRA Levels
`The polyvinylchloride wells of a microtiter plate were in-
`cubated overnight with either rRA or 260F9 MAb (both in
`10 jug of PBS/mL). The plate was washed with PBS and
`blocked for 1 hour with heat-inactivated fetal calf serum.
`Patient samples, diluted 1:10 in PBS, and various standard
`solutions of either anti-murine 260F9 MAb or anti-rRA an-
`tibodies (Cetus Corp.) were added in triplicate to each well.
`Following incubation for 2 hours, the wells were washed and
`then either l25I-260F9 MAb or 125I-rRA in PBS with 2.5%
`heat-inactivated fetal calf serum, adjusted to 5 X 106 cpm
`per well, was added to each well. After 1 hour, the wells
`were washed again and counted in a Packard Autogamma
`5780. Nonspecifically bound radioactivity was determined
`from pretreatment sera and subtracted before concentrations
`of the antibodies were extrapolated from the standard curves.
`260F9 MAb and rRA were iodinated as described by Fraker
`and Speck (73).
`Immunoperoxidase Studies
`Immunoperoxidase studies were performed on acetone-
`fixed frozen sections of tumor biopsy specimens and nerves,
`
`Vol. 81, No. 10, May 22, 1989
`
`ARTICLES 777
`
`IMMUNOGEN 2031, pg. 3
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`
`Downloaded from
`
`http://jnci.oxfordjournals.org/
`
` by Abby Johnson on October 23, 2014
`
`of tumor stabilization, tumor regression, a mixed response,
`or symptomatic improvement in any patient. Three patients
`died within 9 months following treatment, and a fourth pa-
`tient has progressive disease 1 year later.
`Tumor Penetration
`Immunoperoxidase analyses were done on biopsy speci-
`mens from the chest wall from two patients who were using
`antibodies against both murine immunoglobulin and RTA.
`No murine immunoglobulin or RTA, either free or as im-
`munotoxin, could be detected bound to the malignant cells,
`in the tumor matrices, or in the perivascular spaces. Both
`biopsy specimens from the chest wall bound 260F9 MAb in
`vitro.
`Nerve Studies
`A sural nerve biopsy was done in one patient during the pe-
`riod of profound neuropathic disturbance. The results of this
`biopsy showed marked axonal loss and segmental demyeli-
`nation. No inflammatory infiltrates or other abnormalities
`were found on microscopic examination. The biopsy failed
`to reveal any bound immunotoxin or 260F9 MAb; however,
`immunoperoxidase studies showed that the 260F9 antigen
`was present on the nerve in a distribution consistent with
`either Schwann cells or myelin (fig. 1). Seven of nine nor-
`mal human nerves and nerves from two untreated monkeys
`(one rhesus and one cynomologus) were also found to bind
`the 260F9 MAb. Irrelevant mouse MAb did not bind in any
`immunoperoxidase study. The finding of these positive im-
`munochemical results depended on the details of the fixation
`and drying techniques. Careful attention to these details was
`maintained for all the immunochemical procedures.
`Pharmacokinetics
`The 260F9 MAb-rRA serum levels were determined
`by immunoradiometric assay for the five infusions. The
`steady-state immunotoxin concentration for the group was
`reached at 24 hours. Even at similar doses, there was marked
`interpatient variability of steady-state levels. The steady-state
`levels ranged between .235 ± .011 ixg/mL for patient 1 and
`.905 ± .020 Mg/mL for patient 5. Once steady states were
`reached, the immunotoxin concentrations remained constant
`for 6 days in four patients. The steady-state level of patient
`2 remained constant for 3 days; afterwards, it dropped by
`one-third and remained constant until the end of the infu-
`sion. Assuming an immunotoxin distribution limited to a 3-L
`plasma volume, we calculated an immunotoxin half-life be-
`tween 4 and 6 hours for each patient.
`
`Anti-Immunotoxin Antibodies
`The levels of anti-rRa and anti-260F9 MAb antibodies
`were assayed by an immunoradiometric assay, which does
`not distinguish between IgG and IgM. The lower limit of
`sensitivity was 1 fig/mL for the anti-260F9 MAb assay and
`5 Mg/mL for the anti-rRA assay. Four patients developed
`anti-rRa antibodies, and three generated anti-260F9 MAb
`antibodies (table 4). Only patient 4, the patient with sepsis
`and the most heavily pretreated of the group, did not form
`antibodies to either component of the immunotoxin.
`
`Journal of the National Cancer Institute
`
`Patient 4, who had advanced pulmonary disease prior to
`therapy, died during the initial hospitalization. On day 3, she
`developed a temperature of 39.5 °C and hypotension. Two
`days later, coagulase negative staphylococcus was grown
`from her blood. Despite appropriate antibiotics, she devel-
`oped progressive pulmonary air space disease. On day 8
`coagulase-negative staphylococcus was again grown, and the
`central catheter was removed. Later that day she required
`intubation; a Swan-Ganz catheter showed low central venous
`pressures that were compatible with sepsis and adult respi-
`ratory distress syndrome. Although subsequent cultures were
`negative, the patient died of progressive pulmonary and car-
`diovascular failure on day 23.
`Several laboratory abnormalities developed in each patient
`during the course of the infusion (table 3). Two notable ab-
`normalities were the falls in the serum albumin and serum
`sodium that paralleled the fluid retention. The sodium levels
`fell as low as 126 meq/L and 120 meq/L in two patients, al-
`though neither patient manifested mental status changes. No
`proteinuria was found to explain the hypoalbuminema; urine
`sodium studies were not done. Minor elevations of the liver
`transaminases, serum glucose, and serum triglycerides also
`developed in several patients. All the abnormal laboratory
`values returned to normal within 72 hours of completion of
`the infusions.
`Debilitating plexopathies and neuropathies occurred in
`three patients, two at the 50-/ng/kg per day dose and one
`at the 100-/ug/kg per day dose. The neurologic symptoms
`began as a plexopathy on the side of previous chest wall
`irradiation about 2 weeks after the completion of the infu-
`sion. Over the ensuing 2 weeks, the patients developed typi-
`cal sensorimotor neuropathies of the other three extremities.
`The diagnoses were confirmed by neurologic examinations
`and nerve conduction studies in the three cases. Two of the
`three patients had normal brain and spinal cord examinations
`by computer tomography scans. The neuropathies were most
`severe at 2-3 months after treatment and resulted in a de-
`cline in Karnofsky status from 80% (normal activity with
`effort) to 40% (incapable of self care). During the following
`6 months, the patients had complete recovery of their mo-
`tor function, but they continued to suffer from paresthesias
`of their hands and feet.
`Tumor Response
`The clinical responses were followed by appropriate ra-
`diologic studies and by serial measurements of chest wall
`lesions in four patients and serial measurements of liver le-
`sions in the fifth patient. There were no findings suggestive
`
`Table 3. Major laboratory abnormalities associated
`with the 260F9 MAb-rRA immunotoxin
`
`Range
`
`120-200 mg/L
`246-1,244 mg/dL
`120-132 mEq/L
`2.6-3.2 g/dL
`80-149/65-73 U/L
`
`No. of
`patients
`
`55443
`
`Test
`
`Glucose
`Triglycerides
`Sodium
`Albumin
`SGOT/SGPT
`
`778
`
`IMMUNOGEN 2031, pg. 4
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`
`Figure 1. (A) Immunoper-
`oxidase stains show binding
`of the 260F9 MAb to hu-
`man nerve. There is distinct
`surface staining around the
`periphery of many nerves,
`which at high power (B) can
`be seen to have a distribu-
`tion consistent with either
`the Schwann cell membrane
`or the myelin sheath.
`
`Downloaded from
`
`http://jnci.oxfordjournals.org/
`
` by Abby Johnson on October 23, 2014
`
`*5 0Hm
`
`.,,*,-_ v
`
`Although the murine immunoglobulin (IgG,) used to make
`the 260F9 MAb-rRA does not ordinarily bind to human
`monocytes, Weiner et al. have submitted evidence showing
`that the 260F9 MAb-rRA and other RTA-containing im-
`munotoxins bind to the human monocyte Fc receptors.1 It
`is possible that monocyte activation or death may lead to
`the release of mediators that bring about the capillary leak-
`age. Although the pathophysiology of the fluid gain remains
`unclear, poor diuresis following moderate-dose diuretics dur-
`ing the immunotoxin infusion suggests that the fluid was ex-
`travascular.
`Two of our patients treated by continuous infusion and one
`treated by bolus infusions of 260F9 MAb-rRA developed
`symptomatic fluid overload after discharge, probably be-
`cause of the intravascular reaccumulation of the fluid. There
`was no indication by noninvasive studies of cardiac dam-
`age by the immunotoxin. Because the patients were elderly
`or were heavily pretreated with doxorubicin, they were less
`able to tolerate the rapid fluid shifts and developed symptoms
`that necessitated hospitalization. Fluid overload following
`immunotoxin therapy may not have been seen in other stud-
`
`Anti-rRA
`
`54(8)
`49(8)
`51 (32)
`0
`71(9)
`
`Table 4. Anti-immunotoxin antibody
`
`Patient No.
`
`Anti-260F9 MAb
`
`100 (24)*
`31(18)
`10(9)
`
`00
`
`12345
`
`* First day detectable.
`
`Discussion
`The protocol was designed for serial dose escalations from
`50 /ug/kg per day to 450 Mg/kg per day. Despite expecta-
`tions based on animal studies that the immunotoxin would
`be well tolerated (data on file at the Cetus Corp.), the trial
`was stopped early because of severe toxic effects. Many of
`the acute side effects of the 260F9 MAb-rRA were simi-
`lar to those reported for other immunotoxins and by Weiner
`et al, who used bolus infusions of the 260F9 MAb-rRA in
`four patients.1 The pathophysiology of one of these side ef-
`fects, hypoalbuminemia, remains an unexplained phenom-
`enon of immunotoxin therapy. None of our patients had
`proteinuria, severe liver dysfunction, or a protein-losing en-
`teropathy, which would explain the low albumin. Also, the
`rapid return of the albumin after the immunotoxin was dis-
`continued is not compatible with protein loss caused by dys-
`function of a major organ. The association of the fluid re-
`tention and hypoalbuminemia is unclear. In our study and
`the study by Spitler et al. (7), fluid retention paralleled the
`decrease in albumin, unlike protein-losing diseases where hy-
`poalbuminemia precedes the fluid gain. Also, fluid gains of
`4 kg and 5 kg occurred in two patients who had only modest
`drops in their serum albumin to 3.2 mg/dL. While the low
`albumin may enhance accumulation of fluid, it is unlikely to
`be the primary cause.
`One mechanism that has been proposed to explain the fluid
`retention and hypoalbuminemia is a capillary leak syndrome.
`
`'Weiner LM, O'Dwyer J, Kitson J, et al. A phase I evaluation of the
`ricin A chain anti-breast carcinoma immunoconjugate 260F9 MAb-rRA;
`submitted for publication.
`
`Vol. 81, No. 10, May 22, 1989
`
`ARTICLES
`
`779
`
`IMMUNOGEN 2031, pg. 5
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`
`Downloaded from
`
`http://jnci.oxfordjournals.org/
`
` by Abby Johnson on October 23, 2014
`
`ies because they used younger patients. During the rehospi-
`talization, the patients had a rapid response to moderate-dose
`diuretics, which attests to the intravascular nature of the fluid
`at that time. Our experience indicates that patients with heart
`disease need their cardiovascular status very closely moni-
`tored following immunotoxin therapy. In several of our pa-
`tients, significant hyponatremia was also noted; therefore, we
`cannot exclude the possibility that syndrome of inappropriate
`antidiuretic hormone contributed to the fluid retention.
`This is the first study to report a major tissue-targeted
`toxic effect following immunotoxin administration. A debil-
`itating sensorimotor neuropathy afflicted three of our pa-
`tients. The time course is one factor which suggested that
`the neurotoxicity was caused by the immunotoxin. The onset
`of symptoms 2-3 weeks after exposure, maximal disability
`1 -2 months later, and gradual recovery is very similar to the
`acute demyelinating peripheral neuropathy of Guillain-Barre
`syndrome. On the other hand, the time course and recovery
`were not highly suggestive of either a chemotherapy-induced
`neuropathy or a paraneoplastic process. The onset of the gen-
`eralized symptoms was heralded by the development of a
`plexopathy on the side of the previously irradiated chest wall
`and axillary. Irradiation may have induced capillary damage
`that allowed for efficient passage of the immunotoxin, or al-
`ternatively, the previous irradiation injury of the plexus may
`have made it more susceptible to the cytotoxic effects of the
`immunotoxin. The neuropathy occurred at both the 50- and
`100-^g/kg per day doses, and the symptoms did not appear
`to be dose dependent. A nerve biopsy of one patient during
`the time of maximal debilitation helped clarify the nature of
`the neuropathy. Microscopic analyses revealed axonal loss
`and segmental demyelination, which are findings consistent
`with toxic injury to the Schwann cells. The 260F9 MAb was
`implicated when in vitro immunoperoxidase studies found
`intense staining of the nerve sheath, which suggested that
`the 260F9 epitope was present on either the Schwann cells
`or the myelin. The demonstration of 260F9 MAb binding
`on other human nerves provided indirect evidence that neu-
`ral targeting by the immunotoxin may have occurred in the
`other patients. The lack of demonstrable immunotoxin on
`the nerve does not exclude the possibility that the immuno-
`toxin may have produced the neurotoxicity. Letvin et al. (5)
`found in monkeys that target bound immunotoxin was re-
`moved within several days of treatment, while the biopsy
`specimen of our patient's nerve was done 1 month after the
`infusion. Interestingly, immunochemical staining of monkey
`nerves showed 260F9 MAb binding in vitro, although the
`animals did not develop a clinically evident neuropathy. The
`monkeys may have subtle differences in the 260F9 antigen
`that makes their Schwann cells less capable of internaliz-
`ing it; alternatively, their capillary structure may allow less
`efficient efflux of the immunotoxin.
`There was no evidence of tumor regression despite the
`260F9 antigen's being present on at least two breast tumors.
`By immunoperoxidase methods, no component of the im-
`munotoxin was found (a) bound to the malignant cells, (b)
`in the connective tissue, or (c) in the perivascular space at
`completion of therapy. Although we were able to achieve
`sustained immunotoxin levels, it is likely that the levels were
`
`too low to result in meaningful diffusion into the tumor bed.
`Other factors that may have hindered tumor binding include
`insufficient tumor vascularity, abnormal vasculature that did
`not allow for immunotoxin passage, and poor ability to dif-
`fuse through the tumor matrix. On the other hand, the neuro-
`toxicity provides good circumstantial evidence that the im-
`munotoxin has potent biological effects once it binds to its
`cellular target.
`The pharmacokinetic studies showed that within 24 hours
`a steady-state level was reached. Although there was signif-
`icant interpatient variability of steady-state levels for each
`dose, four patients maintained consistent levels throughout
`6 days of treatment. The calculated half-lives of the 260F9
`MAb-rRA were between 4 and 6 hours, which were substan-
`tially longer than the half-life of 43 minutes found for the
`T101-RTA in Herder's study of patients with chronic lym-
`phocytic leukemia (Hertler AA: unpublished observations).
`The uptake of the T101-RTA by circulating lymphocytes
`probably contributed to the short half-life in that study. No
`published data, however, are available on the half-life of na-
`tive RTA containing immunotoxins in humans.
`Antibodies directed against both immunotoxin compo-
`nents were found in three patients, while a fourth devel-
`oped only anti-rRA antibodies. The development of antibod-
`ies may have been one factor in the fall in the steady-state
`immunotoxin levels of patient 2, especially since she had
`no organ dysfunction to explain the change in clearance.
`Although her anti-immunotoxin antibody titers were unde-
`tectable during the infusion, the titers may have been suf-
`ficient to alter her pharmacokinetics. An adverse impact of
`anti-immunotoxin antibodies on pharmacokinetics has been
`previously demonstrated by Hertler et al. (8) in a patient
`treated with bolus infusions of T101-RTA. With each suc-
`cessive treatment, there was a fall in the peak immunotoxin
`concentrations that correlated with a rise in anti-TlOl-RTA
`antibody levels. In addition to accelerating clearance, anti-
`bodies may inhibit immunotoxin function in several ways:
`(a) by binding to the Fab region and blocking the antigen
`recognition site; (b) by binding to non-Fab sites and hin-
`dering internalization; and (c) by binding to the toxin and
`inhibiting its catalytic activity. Because most of our patients
`developed detectable antibodies after treatment, the signif-
`icance of these antibodies remains undetermined. Only pa-
`tient 4 did not develop any anti-immunotoxin antibodies. She
`may have been immunosuppressed because of the prior bone
`marrow transplantation.
`Our anti-immunotoxin antibody results are similar to those
`found in patients treated with various doses of the antimel-
`anoma immunotoxin, XMME-001-RTA (75). Twenty-two
`patients were treated with single, 5-day courses of the
`XMME-001-RTA, which were given as bolus infusions. Both
`studies found a 60% incidence of anti-murine antibodies;
`however, our study found an 80% incidence of anti-rRA an-
`tibodies compared to the melanoma study, in which there
`was a 40% incidence of anti-RTA antibodies. Evi