[CANCER RESEARCH 5l. 4052-4058. August l, 1991)

Phase I Immunotoxin Trial in Patients with B-Cell Lymphoma1

Ellen S. Vitetta,2 Marvin Stone, Peter Amlot, Joseph Fay, Richard May, Mark Till, Joe Newman, Patty Clark,

Robert Collins, D. Cunningham, Victor Ghetie, Jonathan W. Uhr, and Philip E. ThorpeDepartment of Microbiology ¡E.S. V., R. M., M. T., V. G., J. W. U.J and Cancer Immunobiology Center [E. S. V.], University of Texas Southwestern Medical Center,Dallas, Texas 75235; Department of Oncology and Sommons Cancer Center [M. S., J. F., J. N., P. C., K. C.J, Baylor University Medical Center, Dallas, Texas 75246;Drug Targeting Laboratory [P. E. T.], ICRF, London, United Kingdom; Departments of Oncology and Immunology ¡P.A.], Royal Free Hospital, London, UnitedKingdom; and Royal Marsden Hospital fD. C.J, Sutton, Surrey, United Kingdom

ABSTRACT

Fifteen patients with refractory B-cell lymphoma were treated in a

Phase I dose escalation clinical trial with a highly potent immunotoxinconsisting of the Fab' fragment of a monoclonal anti-CD22 antibody

(RFB4) coupled to chemically deglycosylated ricin A chain. All patientshad low, intermediate, or high grade non-Hodgkin's lymphoma. The

immunotoxin was administered i.v. in two to six doses at 48-h intervals.

The peak serum concentration and the t,-_ were not dose dependent

among patients and averaged 1.3 Mg/ml and 86 min, respectively. Threepatients made antibody against A chain, and a fourth made antibodyagainst both A chain and mouse immunoglobulin. Antibody responseswere low (<85 /¿g/ml)in three patients and were not detected until 1 mo

after treatment. The maximum tolerated dose of the immunotoxin was75 nijVnr. Dose-related toxicities included vascular leak syndrome, fever,anorexia, and myalgia. Dose-limiting toxicities included pulmonary

edema and/or effusion, expressive aphasia, and rhabdomyolysis (resultingin reversible kidney failure). There was no evidence of liver dysfunction.Partial responses were achieved in 38% of évaluablepatients, and inthose patients who had >50% CD22* tumor cells, 50% of the patients

achieved a partial response. Clinical responses were not related to tumorgrade and were generally transient, lasting between 1 and 4 mo.

INTRODUCTION

Phase I/II clinical trials in which cancer patients have beentreated with ricin A chain-containing immunotoxins have thusfar yielded minimal clinical responses (1). Occasional transienttumor reductions have been reported in patients with melanoma (2, 3), breast cancer (4), colon cancer (5), and chroniclymphocytic leukemia (6). The lack of efficacy can be attributedto a number of factors beyond the obvious problem of treatingrefractory patients in Phase I dose-escalation studies; (a) nativericin A chain avidly binds to liver cells, via its carbohydrateresidues, resulting in hepatotoxicity and a reduced ability ofthe immunotoxin to reach tumor sites (7-10); (b) most immunotoxins were only moderately potent and required concentrations of 10~'°M or greater to kill 50% of tumor cells invitro (IC50)3(11); (c) melanoma, breast, and colon cancers are

not optimal for treatment because solid tumors may not bereadily accessible to circulating immunotoxin; (d) in the caseof accessible tumors such as chronic lymphocytic leukemia,the majority of the circulating tumor cells fail to internalizethe immunotoxin (12); and (e) antibody responses have oftenprecluded repeated courses of treatment (1).

Received 2/15/91; accepted 5/23/91.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1Supported by NIH Grants CA-28149 and CA-41081.2To whom requests for reprints should be addressed, at Department of

Microbiology, University of Texas Southwestern Medical Center, Dallas, TX75235.

'The abbreviations used are: IC50, 50% inhibitory concentration; IgGl-*,immunoglobulin Gì-*;dgA, deglycosylated ricin A chain; LD50, 50% lethal dose;MTD. maximum tolerated dose; HAMA, human anti-mouse immunoglobulinantibody; SCOT, serum glutamic-oxaloacetic transaminase; PBS, phosphate-buffered saline; HARÕ, human anti-ricin A chain antibody; CT, computerizedtomography; CR, complete response; PR, partial response; MR, mixed response;EKG, electrocardiogram; VLS, vascular leak syndrome; CK, creatine kinase.

Recently, we have developed "second generation" immu

notoxins that avoid some of these problems. We have focusedon malignant B-cell lymphoma as being particularly suited for

immunotoxin therapy because: (a) highly specific monoclonalantibodies against many B-cell surface markers are available;(b) the tumor cells are relatively accessible to blood-borneimmunotoxins; (c) elimination of normal B-cells would be

transient and should not constitute a clinical problem; (d) thepatients are unlikely to mount antiimmunotoxin responsessince they are immunosuppressed by their disease, by priortherapy, and by the killing of normal B-lymphocytes during

immunotoxin therapy (13); (e) the clonality and genetic variants of B-cells can be readily evaluated; and, (f) a murine

lymphoma model (BCL,) (14) was available for preclinicalstudies. The second generation immunotoxin used in thesestudies is a mouse IgGl-K anti-CD22 antibody, RFB4 (15),coupled to chemically dgA (16). One construct, Fab'-RFB4-

dgA (17), was used in the present Phase I study and the other,IgG-RFB4-dgA, is still undergoing clinical evaluation. TheRFB4 antibody is specific for human B-cells and shows noother reactivity as assessed by immunoperoxidase staining of37 normal human tissues (18). The antigen recognized byRFB4 is present on 15 to 100% of tumor cells in 60 to 70%of patients with B-cell lymphomas [(19) and present study].

dgA was used since studies in mice have demonstrated thatdgA-containing immunotoxins are not taken up by the liver,which significantly improves their antitumor activity (9, 20-22). The Fab'-RFB4-dgA immunotoxin has an IC5o of 10~"M on RFB4+ Daudi cells and an IC50 of >10~7 M on RFB4~

cells (17), making it equal in potency to ricin but > 10,000-foldmore specific. Preclinical studies in mice showed dose-relatedweight loss and myositis at doses above 10% of the LD50. Inrhesus monkeys doses of 5 to 15% of the calculated LD50induced appetite loss, myalgia, flu-like symptoms, and transient decreases in serum albumin. At 5% of the LD50, theimmunotoxin (but not Fab' RFB4 alone) selectively depletedRFB4* B-cells from the peripheral blood of the monkeys.4

In the present Phase I study, 15 patients with advancedrefractory B-cell lymphoma were treated with escalating dosesof Fab'-RFB4-dgA. The trial was carried out at the Baylor

University Medical Center, Dallas, the Royal Free Hospital,London, and the Royal Marsden Hospital, London, under aFood and Drug Administration-approved investigational newdrug application (INO) (BB2819). Doses as high as 50 mg/m2

were well tolerated and were only marginally immunogenic;the MTD was 75 mg/m2. Peak serum levels of immunotoxin

averaged 1.3 ¿¿g/ml,and the t]/2 averaged 86 min. Approximately 40% of the patients achieved partial remissions inwhich their overall tumor burden was reduced by 50% or more.Killing of tumor cells was rapid and was achieved within 1 wkafter completion of therapy.

4 E. S. Vitetta, unpublished observations.

4052

Research. on October 5, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

IMMUNOTOXINS IN B-LYMPHOMA

MATERIALS AND METHODS

Patients. Patients over 18 yr of age with histologically documented,refractory non-Hodgkin's B-cell lymphoma were eligible for entry into

the study. Other eligibility criteria included a Karnofsky performancestatus of £30%and a life expectancy of at least 2 mo. Patients had tohave failed to respond to at least one course of conventional therapyand had to have objectively measurable disease as determined by physical examination, laboratory techniques, or radiographie procedures. Inaddition, criteria for eligibility included absence of CNS disease ormeningea! involvement, severe infections, collagen vascular disease,vasculitis, inflammatory arthritis, cardiac disease or glomerulonephri-tis, and preserved renal and hepatic function (serum creatine, <2.0 mg/dl, and bilirubin, <1.5 mg/dl). Tumor cells, as assessed by flow cytom-etry, had to be >15% RFB4*. Patients had to have no history of allergy

to mouse immunoglobulin or ricin and to lack circulating levels ofHAM A (<1 Mg/ml). Chemotherapy was stopped at least 2 wk beforeentry. Patients on corticosteroids were maintained on the same dosesthroughout the treatment. The protocol was approved by the Institutional Review Boards of the participating institutions, and all patientssigned informed consent forms.

Immunotoxin. The Fab'-RFB4-dgA was prepared in a good laboratory practice laboratory as described previously (23). In brief, Fab'

fragments of RFB4 antibody (Abbott Laboratories, Needham Heights,MA) were generated by pepsin digestion and reduction. dgA wasprepared by Inland Laboratories, Austin, TX, according to previouslypublished procedures (24). The Fab' fragments were treated with Ell-man's reagent, chromatographed on Sephadex G-25M, mixed with anequal weight of freshly reduced dgA, and incubated at 25°Cfor 2 h.Free Fab' and free dgA were then removed from the immunotoxinpreparation by chromatography on Blue-Sepharose and Sephacryl S-200HR. The final immunotoxin preparation consisted of one dgAmolecule linked to one Fab' fragment. Other characteristics of theFab'-RFB4-dgA are summarized in Table 1. As determined by the

Limulus assay and by culture in a variety of media as well as analysisof viral DNA (Microbiological Associates), the patient formulation waslow in endotoxin and was free from bacteria and viruses. It inhibitedprotein synthesis by RFB4+ Daudi cells in vitro by 50% at 1.5 x 10~"

M. The LDso in mice was 1.5 mg/25-g mouse. Lots of 5 and 20 mg at1.0 mg/ml of 0.15 M NaCl were frozen in 10- and 30-ml vials andmaintained at -70°C. The shelf life of the material at -70°C is >2.5

yr (maximum time thus far tested). Small vials of each lot were frozenalong with the treatment vials so that the concentration of the immunotoxin could be determined before each patient was treated.

Treatment Plan. Before treatment, 30 to 40 ml of heparinized blood,a bone marrow aspirate, or an involved lymph node was obtained fromeach patient. Serum was also collected to evaluate HAMA. Tumor cellswere separated from blood or marrow by centrifugation over Ficoll-Hypaque or were teased from the lymph node biopsy. The tumor cellsand a reference cell line, Daudi, were stained with either RFB4 orMOPC-21 (control antibody). After incubation with a secondary fluo-rescein isothiocyanate-goat anti-mouse immunoglobulin, cells wereanalyzed on a flow cytometer to determine the percentage of positivecells. Patients having >15% RFB4+ cells at any tumor site (mean

fluorescence intensity, >10% Daudi) were eligible for entry into thestudy. A variety of other T- and B-cell markers (HLA-DR, CD2, CD3,CD4, CDS, CD8, CD 10, CD 19, CD20, CD21, CD22, CD38, K,A,M,«,7, and a) were also evaluated.

Before infusion, and to ensure access to blood, an indwelling centralvenous catheter with a triple lumen port was inserted. Immunotoxinwas filtered through a 0.22-Mm filter and diluted to a final volume of100 ml in saline solution. Before use, the test sample was filtered in asimilar manner, and its concentration was assessed before and afterfiltration by determining the A2so. Recovery was 87 to 98%, andadjustments were made for losses of protein. The immunotoxin wasadministered over a period of 4 h. Each patient received 2 to 6 infusionsof immunotoxin at 48-h intervals. The infusions (n = 2 to 6) given atany dose level were dependent upon the toxicity achieved after each.Hence, if Grade II toxicity was reached after any single infusion, furtherdoses were withheld unless toxicity improved (Table 2). The starting

Table 1 Patient formulation of Fab'-RFB4-dgA

SterilityEndotoxin (Limulus amebocyte lysate

assay)(23)Sodium dodecyl sulfate-polyacrylamide

gel electrophoresis (% as M, 80,000band)

Binding to thiopropyl-Sepharose 6B(free-sulfhydryl groups) (37)

A-chain activity relative to native dgA"Daudi-killing assay (IC5o)4

Antibody binding activity relative to native Fab'c

LD50 (mg/25-g mouse)''

Sterile2.0 Endotoxin units/ml

90%

None

80%1.5 x 10-"M

67%

1.5 mg" After release of dgA from Fab'-dgA by reduction with dithiothreitol. In

reticulocyte lysate assays, reduced immunotoxin and native dgA had IC5o valuesof 7.0 x 10~12M and 5.6 x IO"'2 M, respectively.

* For free dgA chain, the values were 10~7 M. Neither Fab'-Ellman's reagentnor native Fab' was toxic to cells.

c Concentration giving 50% of the maximal fluorescence: Fab' = 1.34 x 10~8M; Fab'-dgA = 2 x 10~8M.

d As total protein. The proportion of dgA in Fab'-dgA is 40% by weight; thus,the LD50in terms of dgA is 0.6 mg/25-g mouse. The LD50of native (unconjugated)dgA is 0.5 mg/25-g mouse.

cumulative dose for the study was 12.5 mg/m2. This was calculated to

be 3.1% of the human LD50 by extrapolation from the mouse LD50 onthe basis of body surface area. Increasing doses of immunotoxin (12.5,25, 50, 75, and 100 mg/m2) were administered to groups of patients

until 3 patients in a group of 5 showed unacceptable Grade III or IVtoxicity in any category (Table 2 shows the toxicity grading particularlypertinent to immunotoxin therapy). When this dose was reached, 4patients received a dose that was '•'/>,as high and thus was defined as the

safe dose. In the first six patients, the course of 2 to 4 doses at eachpercentage of the LD50 was given in doubling doses. In 9 subsequentpatients, 2 to 6 equal doses were given.

Physical examination, tumor measurements, and laboratory testswere performed before treatment and at intervals up to 1 mo later.Laboratory tests included complete blood counts, serum creatinine,creatine kinase, electrolytes, urea, nitrogen, SCOT, bilirubin, albumin,and total serum protein. Serum samples taken up to 60 days aftertreatment were evaluated by radioimmunoassay for the presence ofimmunotoxin and antibodies against RFB4 and dgA.

If corticosteroid were increased or chemotherapy was initiated within30 days after treatment, patients were considered nonevaluable forclinical responses.

Pharmacokinetics and Clearance of Immunotoxins. To evaluate levelsof immunotoxins in serum, wells of a 96-well COSTAR microtiterplate were coated for 2 h at 25°Cwith 20 Mg/ml of affinity-purified

rabbit anti-mouse immunoglobulin in 100 fi\ of PBS. Plates wereblocked with 10% fetal calf serum in 100 n\ of PBS/azide for 16 h at4°C,washed in distilled H2O, and blotted dry. Known concentrations(0.001 to 10 Mg/ml) of Fab'-dgA and doubling dilutions of test serumin 100 M!of PBS/azide were added to plates for 16 h at 4°C.Samples

were removed, and plates were washed with distilled H2O and blotteddry. '"I-labeled rabbit anti-dgA (100,000 cpm) was added in 100 M!ofPBS/10% fetal calf serum/azide to samples for 2 h at 25°C.Samples

were removed, and wells were washed with distilled H2O and blotteddry. Wells were cut out and their radioactivity was counted. Concentrations of immunotoxins were determined from a standard curve constructed from known concentrations of Fab'-RFB4-dgA. The sensitivity

of the assay was 10 ng/ml.Measurement of HAMA and HARÕ. To detect HAMA, plates were

coated with 20 Mg/ml of Fab'-RFB4, and human antibody was detectedwith '-''I -rabbit anti-human immunoglobulin, using a protocol analo

gous to that described above. Affinity-purified anti-mouse immunoglobulin was used to construct the standard curve. The sensitivity ofthe assay was 1 ng/ml. To detect HARÕ, plates were coated with dgAand blocked with 10% normal goat serum/PBS/azide. HARÕ wasquantified with I25l-rabbit anti-human immunoglobulin. The standardcurve utilized affinity-purified anti-dgA diluted in 100 M' of normalhuman serum. The sensitivity of the assay was 10 ng/ml.

4053

Research. on October 5, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

IMMUNOTOXINS IN B-LYMPHOMA

Table 2 Grading of immunotoxin-related toxicityDose modification was as follows: Grade I toxicity, no change in the scheduled dosage; Grade II toxicity, 24-h delay of dosage, with dose given if toxicity improves;

Grade III toxicity, the next dose of immunotoxin was withheld and only given if toxic parameters have improved (halving of the dosage could be considered); GradeIV toxicity. no further dosage.

ParameterFever

Persistent"

TransientHepatic

BilirubinSCOTRenal

Serum creatinineCreatinineclearanceCardiac

RhythmI

(mild)99-101'F

>101-103'F<2.5

<2 xbaseline2.0-2.5

80- 100iil/minSinus

tachycardia >1 10 atII

(moderate)>101-103-F

>103-105°F2.5-3.0

2-5 xbaseline2.5-3.0

50-79ml/minUnifocal

PVC4 atrial ar-III

(severe)>103-105°F

>105-106°F3.0-4.0

5-10 xbaseline3.0-4.0

30-49ml/minMultifocal

PVCIV

(maximal)>105-106'F

>106°F>4.0

>10 xbaseline>4.0

<30ml/minVentricular

arhythmia; myo-

Pericardial effusion

VascularHypoalbuminemia

GastrointestinalNausea/vomiting

MuscleCreatine kinase

PulmonaryClinical

ABG

Chest X-ray

NeurotoxicityMental status

rest ythmiaAsymptomatic Symptomatic but no tap

required

<2.8

Nausea

Myalgia

Tachypnea respiration,<30

PAO2 <95 on room air

<10% of lung fields showinfiltrate or effusion

Transient lethargy, milddepression, confusion,severe headache

<2.4

Transient vomiting

Myalgia requiring treatment

Dyspnea on exertion

PAO2 <80 on room air

<20% of lung fields showinfiltrate or effusion

Somnolence or disorientation <50% of wakinghours

Tamponade and tap required

<2.0 with supplementation

Vomiting requiring hyperali-mentation

CPK and aldolase elevationsof >2-fold normal rangefor 48 h

Dyspnea at rest requiringsupplementary oxygen

PAO; on room air

<50% of lung fields show infiltrate or effusion

Aphasia, or disorientation>50% of waking hours

cardial infarction or ischemiaTamponade, hypotensives, and

tap required

<2.0 with edema (4+) requiring supplementation

Intractable vomiting

CPK and aldolase elevations of>5-fold

Assisted ventilation required

PAO2 <60 on supplementaryoxygen

>50% of lung fields show infiltrate or effusion

Coma, seizure, or psychosiswith abnormal EEC

" Persistent, >6 h; transient; up to 6 h.4 PVC, Premature ventricular contraction; CPK, Creatine phosphokinase; ABG, Arteriole blood gas; PAO2, Pulmonary artery oxygénation;EEC, electroenceph

alogram.

Evaluation of Tumor Responses. Patients were examined 48 h to 1wk and again at 30 days after the last infusion. When palpable diseasewas present, two perpendicular diameters of all masses were measureddaily during treatment, within 1 wk after completion of therapy, and at30 days. To measure nonpalpable disease, serial CT scans were takenat 1-cm intervals throughout the chest, abdomen, and pelvis at 48 h to1 wk and again at 30 days. Two perpendicular diameters of cross-sections of tumor were measured. Total disease was taken as the sumof the products of perpendicular diameters of all measurable lesions.Blood samples and bone marrow were evaluated histologically and byflow cytometry when disease was present in these sites. A CR wasdefined by the disappearance of all measurable tumor 1 mo aftertreatment. A PR was defined by a 50% or greater reduction of the sumof the products of perpendicular diameters of all évaluabletumornodules 1 mo after treatment, with no tumor sites showing enlargement.A MR was defined by a reduction of the product of perpendiculardiameters of all measurable lesions by 50% or greater 1 mo aftertreatment, with progression in one or more sites. No response wasdefined by overall reduction of measurable disease by less than 50% 1mo after treatment. Patients were removed from the study when theirdisease progressed or other treatments were initiated.

RESULTS

Patients. Fifteen patients were entered, and their demographics are detailed in Table 3. The majority of the patients hadbulky lymphadenopathy. Fourteen had Stage III or IV disease,4 had splenomegaly, 7 had blood involvement, and 11 hadmarrow involvement.

Pharmacokinetics and Serum and Immunotoxin Levels. Patients received 2 to 6 i.v. infusions of immunotoxin, and thecumulative dose was used to make comparisons of toxicity andclinical response. The mean peak serum concentrationsachieved for each fractionated dose of the cumulative dose andthe mean serum half lives are listed in Table 4. Fig. 1 showstwo typical blood clearance curves. In all patients, immunotoxinlevels in the serum peaked at the end of the infusion and fell tobaseline 8 to 12 h thereafter. The peak levels of immunotoxinreached in the serum varied among patients in a manner thatdid not correlate with the dose administered, although serumlevels were dose related among infusions within a single patient.The average serum level achieved (at all dose levels) was 1.3Mg/ml. The i,/2 value ranged from 48 to 121 min, averaging 86min. The rapid clearance rate and lack of dependence of peakserum levels on dose among patients have been observed inanother Phase I immunotoxin trial (25) and are probably explained by absorption of the immunotoxin onto tumor cells.Variations in the size of tumors among patients and accessibilityof tumor cells to the blood appear to have masked any changesrelated to the amount of immunotoxin infused and its rate ofclearance.

Antibody Responses. As summarized in Table 5, four of 14patients (29%) made antibody to the immunotoxin. One patientmade a significant IgG anti-dgA response which peaked at 28days. The magnitude (0.5 mg/ml) and isotype (IgG) of this

4054

Research. on October 5, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

IMMUNOTOXINS IN B-LYMPHOMA

Table 3 Patient demographics

ParameterNo.

treatedAge(yr)SexMaleFemaleNon-Hodgkin's

IvmphomagradeHighIntermediateLow-Patients

with % of tumor cells 2:50%RFB4*Patientswith % of tumor cells 20 to 49%RFB4+BoneGrade

Type (38) marrowBloodHighIBLf+IntermediateDL ——FLDL

+DSC+DL

+±'DSC

++DM+DSC

+LowFM +±FM

+±FSC++FM

±FMSL

+ +No.

(%)1548

(38-67)°1

1[731*4

[271im8

[53]6(40]12

[80]3[20]Splenomegaly-——_SPLx*—+++——++++———-

°Numbers in parentheses, range.* Numbers in brackets, percentage.' IBL. large cell, immunoblastic; DL. diffuse, large cell: FL, follicular, predom

inantly large cells; DSC, diffuse, small cleaved cell: DM, diffuse, mixed small andlarge cell; FM. follicular. mixed small cleaved cell and large cell; FSC. follicularsmall cleaved cell; SL, small lymphocytic.

d Prior splenectomy.' ±.small numbers of lymphoma cells in the blood, only detectable phenotyp-

ically.

Table 4 Serum levels and tv,of immunotoxins

Cumulative doseadministeredmg/m225

5075

100120No.

ofpatients92

11Mean

peak concentrationafter each

infusion(/¿g/ml)0.60.9

(0.3-2.2)*3.7(1.9-5.5)0.91.5t-/,"

(min)121

86(51-159)78 (69-87)48

102

Mean 1.3 86°A 0-phase of clearance was not discernible in any patient. Half-lives thus

refer to the average time for immunotoxin levels to decline by 50% during theinitial 8-h period after the infusion.

* Numbers in parentheses, range.

response are characteristic of a secondary response, suggestingthat this patient may have been exposed to ricin (castor beans?)or another cross-reactive antigen previously. Two other patientsmade small but measurable HARÕ responses that were detectable by 42 days after immunotoxin therapy. Both responseswere in the ng range (48 and 82 ¿¿g/ml).The fourth patientmade both HARÕ (26 Mg/ml) and HAMA (5 Mg/ml) at 28 days.

Toxicity. Table 6 lists the toxic effects observed in all 15patients. In 13 patients, side effects were transient and reversedwithin 1 to 2 wk after treatment; one patient had rhabdomy-olysis and took 6 wk to recover fully; one patient died duringtherapy of advanced lymphoma of the myocardium and pulmonary parenchyma complicated by pulmonary edema. Theother patients survived for 2 wk to >2 yr after completion oftherapy.

As in other immunotoxin trials, vascular leak syndromeappeared to be the major side effect. All patients showed the

01

E3L_

0

CO

o•oi*mILrrJQ

Time after starting infusion (hours)Fig. 1. Serum levels of immunotoxins in a patient receiving either 4 escalating

doses (A) or 4 equal doses (B) of immunotoxin. At the top, the patient received25 mg/m2 in doses of 1.7 mg/m2 (/), 3.4 mg/m2 (2), 6.8 mg/m2 (3), and 13.6mg/m2 (4). At the bottom, the patient received 50 mg/m2 in doses (/ to 4) of 12.5mg/m2.

Table 5 Antibody responses in treated patients

Response"HAMA

onlyHARÕ onlyHAMA + HARÕNo.

ofpatients0/14

3/14*1/14CPeak

antibodylevel(Mg/ml)HAMA

HARÕ48,

82, 5005 26

" Limits of detection of the assays were 1 ng/ml for HAMA and 10 ng/ml for

HARÕ.* Responses became detectable on Day 28 in the patient with the high liter

(500 ng/ml) and on Day 42 in the other 3 palients.' Detectable on Day 28.

Table6SymptomDecrease

in serum albumin(%)Weightgain(kg)Fever

(°F)Tachycardia

(beats/min)Decreasein EKG voltage(%)Myalgia*Peripheral

edemaAnorexiaandnauseaSymptoms

definingMTDPulmonaryedemaPulmonaryeffusionExpressiveaphasia'Rhabdomyolysis1'ToxicityNo.

ofpatients

(total =15)ISIS10667762121Medianchange28(14-45)°3.3(1.9-8.6)101

(99-102)125(100-135)51

(30-75)

" Numbers in parentheses, range.* With 3 patients having CK increases of 1.5-, 2-, and 415-fold above the upper

limit of the normal range.c Transient (3 to 8 h) in one patient. In the other patient, aphasia lasted 36 h

and was accompanied by mental obtundation.d Leading to acute but transient renal failure resolving within 10 days.

4055

Research. on October 5, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

IMMUNOTOXINS IN B-LYMPHOMA

Fig. 2. Abdominal CT scans of a patient with intermediate-grade non-Hodgkin's lymphoma before (top) and after (bottom) treatment with 75 mg/m1. A shows 3

lymph nodes to the left of the (calcified) aorta (arrow). B shows a node to the left of the aorta and a number of mesenteric nodes (arrows).

characteristic decrease in serum albumin and other proteinsand increase in body weight due to edema. Other toxicitiesincluded tachycardia and a reduction in EKG voltage in abouthalf the patients. In addition, about half the patients had fever,which ranged from mild to moderate, and myalgia, whichranged from mild to severe. At cumulative doses of 75 mg/m2

or higher, some patients developed toxic side effects (probablyrelated to vascular leak syndrome) that were severe enough towarrant discontinuation of therapy. These dose-limiting toxicities included pulmonary edema in two patients (with pulmonary effusion in one patient), expressive aphasia in one patient,and both aphasia and rhabdomyolysis in one patient, leadingto acute renal failure. The severity of the side effects was notalways dose related and did not correlate with either the performance status of the patient or the use of maintenance corti-costeroids. For example, one patient with a low dose (50 mg/m2) had severe myalgia, while two patients at higher doses (100to 120 mg/m2) had relatively mild myalgia. No patient showed

any evidence of direct hepatic or renal damage.Clinical Response. Fab'-RFB4-dgA induced rapid tumor

regression in nearly half of the 14 patients who were évaluable.Forty-three % showed greater than a 50% reduction in tumorwithin a week after completion of the therapy. Thirty-eight %of the patients achieved PRs according to standard WHOcriteria at 1 mo after the completion of therapy. No delayed

PRs were observed. Fig. 2 shows chest and pelvic CT scans ofone such patient. In the subset of patients whose tumor cellswere >50% RFB4+, 50% achieved PRs. There was also a trend

toward increasing responsiveness at higher dose levels: of 2patients receiving 12.5 to 25 mg/m2, neither achieved PR; of 7patients receiving 50 mg/m2, 2 achieved PRs; of 4 patientsreceiving 75 to 120 mg/m2, 3 achieved PRs. Responses were

transient, lasting an average of 1.8 mo (range, 1.25 to 4 mo) infive patients. There was no obvious correlation between theresponsiveness of any given patient to the immunotoxin andthe size or grade of the tumor, the presence of splenomegaly orof leukemic cells in the blood, the performance status of thepatient, or the use of maintenance corticosteroids. However, inpatients who had both peripheral adenopathy and retroperito-neal adenopathy, the latter generally showed less impressiveantitumor responses.

DISCUSSION

This study was designed to determine the toxicity, pharma-cokinetics, immunogenicity, and antitumor response of a highlypotent anti-CD22-dgA immunotoxin in B-cell lymphoma patients. This immunotoxin was chosen because it was extremelycytotoxic to neoplastic B-cells, surpassing all other B-cell-specific Fab'-dgA immunotoxins that we have evaluated in

4056

Research. on October 5, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

IMMUNOTOXINS IN B-LYMPHOMA

potency by at least 10-fold. In addition, the immunotoxin waseffective at killing pre-B-cell leukemia cells, despite the factthat these cells express low levels of CD22 on their surface(18).

Several features of the immunotoxin construct used in thepresent study should be noted. It consists of a univalent Fab'

fragment of RFB4 antibody coupled by a cystine bond to dgA.The use of the Fab' fragment as compared to intact IgG offersa number of potential advantages: (a) the Fab' immunotoxin is

smaller (M, 80,000) and is probably more effective in thepenetration of tumor masses; (b) any Fab' released from theimmunotoxin in vivo will be rapidly cleared (21, 26-29) andtherefore will not compete with the circulating immunotoxinfor binding to tumor cells; (c) Fab' cannot bind to Fc receptorsthat are present on a variety of cells; (d) the Fab' immunotoxin

lacks the relatively immunogenic Fc portion and might havelower immunogenicity than an intact antibody construct; and,(e) possibly because of its more rapid clearance, the Fab'-dgA

has an LD50 in mice which is 5 to 8 times higher (i.e., less toxic)than its more stable and long-lived IgG-dgA counterpart. Itsdisadvantages, as compared to IgG-dgA, include a shorter invivo half-life, 10-fold lower potency (17, 21, 30, 31), lower

stability (21), and a higher cost of production.The VLS appears to be the major side effect of this as well

as other types of ricin A chain-containing immunotoxins (1,11). The physical signs of the VLS were weight gain, peripheraledema, decrease in blood pressure, hypoalbuminemia, and smallpleural effusions. Several other toxicities could also be explained by VLS: tachycardia was probably related to the ensuingintravascular hypovolemia; decrease in EKG voltage was aconsequence of edema in the chest wall and/or the pericardium/myocardium; anorexia and nausea could relate to gastric mu-

cosal edema; and cerebral edema could be responsible for the 2transient aphasias seen. In no case was myocardial damageestablished by elevation in myocardial isoenzymes of CK. In itsmost severe form, encountered at doses of 100 mg/m2 and

greater, VLS resulted in pulmonary edema and gross pleuraleffusions. Since VLS is a major side effect of interleukin 2 and7-interferon therapy (32-34), it is tempting to invoke the samecytokine effects to explain the immunotoxin-mediated VLS.Either activated cells or killed lymphoma cells are capable ofreleasing cytokines. VLS can be controlled by glucocorticoids(33), which could be used prophylactically during immunotoxintherapy. It remains to be determined whether glucocorticoidswould augment or negate the antitumor effect of immunotoxins,since vascular leakage might enhance the extravasation of immunotoxin into tumor sites.

Myalgia occurred in nearly half the patients and, althoughits mechanism is not understood, it was expected from previousimmunotoxin studies (1, 11) and from the myositis found inmice dying from dgA- and ricin A chain-containing immunotoxins.4 In 3 of 7 cases myalgia was accompanied by evidence

of muscle damage in the form of elevated levels of CK in theblood. In one instance, the muscle damage resulted in severerhabdomyolysis, leading to renal failure. In this patient, amuscle biopsy (performed 7 days after the initial event andwhile CK levels were still grossly elevated) was not informative,and there was only a moderate increase in macrophages betweenmuscle fibers with no other leukocytic infiltrate and no changein staining patterns of muscle enzymes. This patient made afull recovery from both the ensuing acute renal failure (15 days)and muscle weakness (6 wk). As mentioned above, there wasno myocardial damage even in the patient who developed

rhabdomyolysis (myocardial CK isoenzyme remained normal)and, in the same patient, smooth muscle function was notaffected (vascular or gastrointestinal). It therefore appears thatthe toxicity of dgA-containing immunotoxins is restricted tostriated muscle.

As predicted from the immunosuppressed state of the patientpopulation (13, 14) and from the fact that the ¡mmunotoxinprobably kills normal B-cells, antibody responses to the immunotoxin were minor and less frequent than reported in otherimmunotoxin trials (26, 35). Only 4 patients raised antiimmu-notoxin antibodies and, in three of these, the antibody levelswere low. Three of the four antibody responses that did developoccurred late (28 to 42 days after completion of the treatment).This raises the question of how such late responses are mountedto an immunogen that is rapidly cleared. Perhaps there are asmall number of slowly dividing B-cells in these patients thattake a long time to expand and then differentiate into plasmacells. Alternatively, the antigen is anatomically sequestered andstimulates a late response as it is released into the circulation.In either case, there appears to be a window of about 2 to 3 wkin the patients mounting responses in which retreatment couldbe considered.

The mean serum half-life of the immunotoxin was short (1.5h) and similar to that seen in mice (23). Hence, therapeuticconcentrations of immunotoxin could not be sustained betweendoses. A direct correlation between the in vivo half-life of animmunotoxin and the resultant antitumor activity has beendemonstrated in murine lymphoma models (22). Thus, in PhaseII patients, a continuous infusion protocol may have greaterefficacy than fractionated doses.

The clinical responses of the patients in this study were higherthan would have been predicted in a Phase I dose escalationtrial (36). PRs averaged 38% for all évaluablepatients and weremarginally higher (50%) in patients whose tumor cells were>50% RFB4+. In general, patients with peripheral but not

retroperitoneal adenopathy showed better responses. It is unclear whether this difference was related to the degree of tumorpenetration by the immunotoxin or to the fact that tumor cellsin different sites were heterogeneous with respect to RFB4positivity, susceptibility to immunotoxin, or accessibility oftumor cells to the blood. These issues will be addressed in PhaseII trials by phenotyping residual tumor cells after treatment. Inpatients receiving doses of 75 mg/m2 or greater, 3 of 4 (75%)

attained PR. Taken together, these results suggest that dosesof 50 mg/m2 should be effective in treating minimal disease,

particularly in situations where 50 to 100% of the cells in thetumor are RFB4+.

Finally, it should be emphasized that the level of expressionof CD22 on tumor cells in a given patient was heterogeneous,with some cells staining strongly and others not at all. Thisobservation suggests that a cocktail of immunotoxins directedagainst different B-cell antigens combined with conventionaltherapy will be necessary to achieve the maximal therapeuticbenefit in B-lymphoma patients.

ACKNOWLEDGMENTS

We thank Dr. A. Bailey, A. Brown, Y. Chinn, Dr. M. Ghetie, D.Grennan, P. Holder, P. Knowles, L. Li, M. M. Liu, M. McCarthy, R.Nisi, E. Rawlings, M. Reif, D. Scott, L. Traban, and C. Viner forskilled and dedicated technical and clinical assistance. We thank Dr.M. Chandler, Inland Laboratories, and Dr. A. Jarvis, Abbott Laboratories, for working with us on the preparation and testing of the RFB4and dgA. We are indebted to Dr. T. Zaremba and Dr. K. Scribner,

4057

Research. on October 5, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

IMMUNOTOXINS IN B-LYMPHOMA

Food and Drug Administration, for their valuable advice. We thankDr. W. Neaves and Dr. K. Wildenthal, University of Texas, for constantsupport and help in the construction of the scale up laboratory. Finally, igwe acknowledge C. Baselski, R. Reiber, A. Beckett, S. Richardson, andD. Thomas for expert secretarial assistance.

19.REFERENCES

1. Byers, V. S.. and Baldwin, R. W. Therapeutic strategies with monoclonal 20.antibodies and immunoconjugates. Immunology, 65: 329-335, 1988.

2. Oratz, R., Speyer, J. L., Wernz, J. C., Höchster, H.. Myers, M.. Mischak,R., and Spitler, L. E. Anti-melanoma monoclonal antibody ricin A chain 21immunoconjugate(XMMME-OOI-RTA)pluscyclophosphamidein the treatment of metastatic malignant melanoma: results of a Phase II trial. J. Biol.Response Modif, 9: 345-354, 1990.

3. Spitler, L. E., de! Rio, M., Khentigan, A., Wedel, N. I., Brophy, N. A., 22Miller, L. L., Harkonen, W. S., Rasendorf, L. L.. Lee, H. M., Mischak, R.P., Kawahata, R. T.. Stoudemire. J. B., Fradkin, L. B.. Bautista, E. E., andScannon, P. J. Therapy of patients with malignant melanoma using amonoclonal anti-melanoma antibody-ricin A chain immunotoxin. Cancer 23Res., 47: 1717-1723, 1987.

4. Weiner, L. M., O'Dwyer. J.. Kitson, J., Comis. R. L., Frankel, A. E., Bauer,R. J., Konrad, M. S., and Groves. E. S. Phase I evaluation of an anti-breastcarcinoma monoclonal antibody '260F9-recombinant ricin A chain immu- 24noconjugate. Cancer Res., 49:4062-4067. 1989.

5. Byers, V. S., Rodvien, R., Grant, K., Durrani, L. G., Hudson. K. H., Baldwin.R. W., and Scannon, P. J. Phase I study of monoclonal antibody ricin A 25.chain immunotoxin XOMAZYME-791 in patients with metastatic coloncancer. Cancer Res., 49: 6153-6160. 1989.

6. Hertler, A. A., Schlossman, D. M., and Borowitz, M. J. A Phase I study ofTlOl-ricin A chain immunotoxin in refractory chronic lymphocytic leukemia. 15J. Biol. Response Modif, 7: 97-113, 1988.

7. Blakey, D. C., Skilleter, D. N., Price, R. J.. and Thorpe, P. E. Uptake of 27.native and deglycosylated ricin A-chain immunotoxins by mouse liver paren-chymal and non-parenchymal cells in vitro and in vivo. Biochim. Biophys.Acta, 968: 172-178, 1988.

8. Skilleter. D. N., Paine, A. J., and Stirpe. F. A comparison of the accumulationof ricin by hepatic parenchyma! and non-parenchymal cells and its inhibition 28.of protein synthesis. Biochim. Biophys. Acta. 677:495-500. 1981.

9. Blakey, D. C, Watson, G. J., Knowles, P. P.. and Thorpe, P. E. Effect ofchemical deglycosylation of ricin A chain on the in vivo fate and cytotoxic 29.activity of an immunotoxin composed of ricin A chain and anti-Thy 1.1antibody. Cancer Res., 47: 947-952, 1987.

10. Bourrie, B. J., Casellas, P., Blythman, H. E., and Jansen, F. K. Study of the 30.plasma clearance of antibody-ricin-A-chain ¡mmunotoxins. Evidence for specific recognition sites on the A chain that mediate rapid clearance of theimmunotoxin. Eur. J. Biochem., 155: 1-10, 1986. 3]

11. Hertler. A. A., and Frankel, A. E. Immunotoxins: a clinical review of theiruse in the treatment of malignancies. J. Clin. Oncol., 7: 1932-1942, 1989. 32.

12. Laurent. G., Kuhlein, E., Casellas, P.. Canat, X., Carayon, P.. Poncelet. P..Correli. S.. Rigai, F.. and Jansen, F. K. Determination of sensitivity of fresh 33leukemia cells to immunotoxins. Cancer Res.. 46: 2289-2294, 1986.

13. Scharff, M. D., and Uhr, J. R. Immunologie deficiency disorders associatedwith lymphoproliferative diseases. Semin. Hematol., 2:47-67, 1965. 34.

14. Krolick, K. W., Uhr. J. W., Slavin. S., and Vitella. E. S. In vivo therapy of amurine B cell tumor (BCL,) using antibody-ricin A chain immunotoxins. J. 35.Exp. Med., 155: 1797-1809, 1982.

15. Campana, D., Janossy, G.. Bofill. M., Trejdosiewicz, L. K., Ma. D., Hoff-IM.nul. A. V., Mason. D. Y.. LeBacq, A-M., and Forster. H. K. Human B 36.cell development. I. Phenotypic differences of B lymphocytes in the bonemarrow and peripheral lymphoid tissue. J. Immunol., 134:1524-1529,1985. 37.

16. Thorpe, P. E., Detre, S. I.. Foxwell. B. M. J.. Brown, A. N. F., Skilleter, D.N., Wilson, G., Forrester, J. A., and Stirpe. F. Modification of the carbohy- 38.drate in ricin with metaperiodate-cyanoborohydride mixtures. Effects oftoxicity and in vivo distribution. Eur. J. Biochem., 147: 197-206, 1985.

17. Shen, G-L., Li, J-L., Ghetie M-A., May, R. D., Till, M., Brown, A. N.. Reif.

M.. Knowles. P., Uhr, J. W.. Janossy. G., Amlot, P., Vitella. E. S., andThorpe, P. E. The evaluation of four anti-CD22 antibodies as ricin A chain-containing immunotoxins. Int. J. Cancer. 42: 792-797, 1988.Li, J-L., Shen. G-L.. Ghetie. M-A.. May, R. D., Till, M.. Ghetie, V., Uhr, J.W., Janossy, G., Thorpe, P. E., Amlot. P., and Vitetta. E. S. The epitopespecificity and tissue reactivity of four murine monoclonal anti-CD22 antibodies. Cell Immunol.. 118: 85-99. 1989.Anderson. K. C., Bates, M. P.. Slaughenhoupt, B. L., Pinkus. G. S.. Schlossman, S. F., and Nadler, L. M. Expression of human B cell-associated antigenson leukemias and lymphomas: a model of human B cell differentiation.Blood. 63: 1424-1433. 1984.Fulton. R. J., Uhr, J. W.. and Vitetta. E. S. In vim therapy of the BCL,tumor: effect of immunotoxin valency and deglycosylation of the ricin Achain. Cancer Res., 48: 2626-2631. 1988.Fulton. R. J.. Tucker. T. F.. Vitetta. E. S.. and Uhr, J. W. Pharmacokincticsof tumor-reactive immunotoxins in tumor-bearing mice: effect of antibodyvalency and deglycosylation of the ricin A chain on clearance and tumorlocalization. Cancer Res., 48: 2618-2625, 1988.Thorpe. P. E., Wallace, P. M.. Knowles. P. P., Reif, M. G., Brown, A. N.F., Watson, G. J.. Blakey, D. C., and Newell, D. R. Improved anti-tumoreffects of immunotoxins prepared with deglycosylated ricin A chain andhindered disulfide linkages. Cancer Res.. 48: 6396-6403, 1988.Ghetie, V., Ghetie, M-A., Uhr, J. W., and Vitetta, E. S. Large scale preparation of immunotoxins constructed with the Fab' fragment of IgGl murine

monoclonal antibodies and chemically deglycosylated ricin A chain. J. Immunol. Methods. 112: 267-277. 1988".

Fulton. R. J.. Blakey. D. C., Knowles, P. P.. Uhr. J. W., Thorpe, P. E., andVitetta. E. S. Purification of ricin Al, A2, and B chains and characterizationof their toxicity. J. Biol. Chem.. 261: 5314-5319. 1986.Gould. B. J., Borowitz. M. J.. Groves, E. S., Carter, P. W.. Anthony, D.,Weiner, L. M., and Frankel. A. E. Phase I study of an anti-breast cancerimmunotoxin by continuous infusion: report of a targeted toxic effect notpredicted by animals studies. J. Nati. Cancer Inst.. */: 775-781, 1989.Spiegelberg. H. L., and Weigle. W. O. The catabolism of homologous andheterologous 7S 7-globulin fragments. J. Exp. Med., 121: 323-338, 1965.Thorpe, P. E.. Wallace. P. M., Knowles, P. P.. Reif. M. G., Brown, A. N.F., Watson, G. J., Knyb, R. E.. Wawrzynczak. E. J., and Blakey, D. C. Newcoupling agents for the synthesis of immunotoxins containing a hindereddisulfide bond with improved stability in vivo. Cancer Res., 47: 5924-5931,1987.Wochner, R. D., Strober, W.. and Waldmann, T. A. The role of the kidneyin the catabolism of Bence-Jones proteins and immunoglobulin fragments.J. Exp. Med., 126: 207-221, 1967.Covell, D. G.. Barbet, J.. Holton, O. D., Black, C. D., Parker, R. J., andWeinstein, J. N. Pharmacokinetics of monoclonal immunoglobulin G F(ab')¡and Fab' in mice. Cancer Res.. 46: 3969-3978, 1986.

Masuho. Y.. Kishida, K., Saito. M., Umeto, N., and Hará.T. Importance ofthe antigen-binding valency and the nature of cross-linking bond in ricin Achain conjugates with antibody. J. Biochem.. 91: 1583-1591. 1982.Raso, V.. and Griffin. T. Specific cytotoxicity of a human immunoglobulindirected Fab'-ricin A conjugate. J. Immunol.. 725: 2610-2616. 1980.Parkinson, D. R. Interleukin-2 in cancer therapy. Semin. Oncol.. 10: 10-26,1988.Rosenberg, S. A., and Lotze, M. T. Cancer immunotherapy using interleukin-2 and interleukin-2-activated lymphocytes. Annu. Rev. Immunol., 4: 681-709, 1986.Fahey, J. L., Sarna. G.. Gale, R. P.. and Seeger. R. Immune interventions indisease. Ann. Intern. Med.. 106: 257-274, 1987.Hertler. A. A.. Spitler. L. E.. and Frankel, A. E. Humoral immune responseto a ricin A chain immunotoxin in patients with metastatic melanoma. CancerDrug Delivery. 4: 245-253, 1987.Popcock, S. J. Clinical Trials: A Practical Approach. New York: John Wiley&Sons. 1984.Hilson, D. A. Resolution of thiol-containing proteins by sequential-elutioncovalent chromatography. J. Immunol. Methods. 4: 101-111, 1981.The Non-Hodgkin's Lymphoma Pathologic Classification Project: NationalCancer Institute sponsored study of classifications of non-Hodgkin's lym

phomas. Summary and description of a working formulation for clinicalusage. Cancer (Phila.), 49: 2112-2135. 1982.

4058

Research. on October 5, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

1991;51:4052-4058. Cancer Res   Ellen S. Vitetta, Marvin Stone, Peter Amlot, et al.   Phase I Immunotoxin Trial in Patients with B-Cell Lymphoma

  Updated version

  http://cancerres.aacrjournals.org/content/51/15/4052

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/51/15/4052To request permission to re-use all or part of this article, use this link

Research. on October 5, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from