AU645981B2 - Removal of human anti-murine and heterophilic antibody interference in double antibody EIA assays - Google Patents

Description

OP] DATE 11/11/91 PCT AOJP DATE 19/12/91 APPLN. in q9652 PCT NUMBER PCT/liS90/02239

INTERNATIONAL

(PCT)

(51) International Patent Classification 5 (11) International Publication Number: WO 91/16627 G01N 33/53, 33/543, 33/48 Al G01N 1/00, C12N 1/00 (43) International Publication Date: 31 October 1991 (31.10.91) (21) International Application Number: PCT/US90/02239 (81) Designated States: AT (European patent), AU, BE (European patent), CA, CH (European patent), DE (Euro- (22) International Filing Date: 25 April 1990 (25.04.90) pean patent)*, DK, DK (European patent), ES, ES (European patent), FR (European patent), GB (European patent), IT (European patent), JP, LU (European pa- (71) Applicant (for all designated States except US): CITY OF tent), NL (European patent), SE (European patent), US.

HOPE [US/US]; 1500 East Duarte Road, Duarte, CA 91010-0269 (US).

Published (72) Inventors; and With international search report.

Inventors/Applicants (for US only) BEATTY, John, David [CA/US]; BEATTY, Barbara, G. [CA/US]; 1532 Hyland Avenue, Arcadia, CA 91740 (US).

(74) Agent: IRONS, Edward, 919 18th Street, Suite 800, Washington, DC 20006 (US).

645981 (54)Title: REMOVAL OF HUMAN ANTI-MURINE AND HETEROPHILIC ANTIBODY INTERFERENCE IN DOUBLE ANTIBODY EIA ASSAYS (57) Abstract A procedure is described for the reduction or removal of human anti-murine and heterophilic antibody interference in double antibody assays by incubating a human serum or plasma sample with a mixture of monoclonal antibodies.

See back of page WO 91/16627 PCr/US90/02239 645 8 -1- REMOVAL OF HUMAN ANTI-MURINE AND HETEROPHILIC ANTIBODY INTERFERENCE IN DOUBLE ANTIBODY EIA ASSAYS Field of the Invention This invention relates to removal of interference observed in double antibody assays consequent from the presence of human anti-murine and heterophilic antibodies present in human plasma. This interference results in measurement of artifactual levels of analyte, carcinoembryonic antigen

(CEA).

Background of the Invention Clinical double antibody assays for CEA are used routinely to monitor recurrence of disease in colorectal cancer patients. Since treatment planning may be influenced by the results of this assay, accurate reliable determinations of this tumor associated antigen are extremely important.

Heterophilic antibody interference in solid phase enzyme immunoassays (EIAs) using monoclonal antibodies (MABs) has been observed for other blood analytes but has not been reported as such for the CEA enzyme immunoassays or radioimmunoassays. These heterophilic antibody interference problems as yet have no clearly defined immune response origin.

With the advent of murine monoclonal antibody production and use of these antibodies for targeting WO 91/16627 PCT/US90/02239 -2drugs and radioactivity to tumor tissue, a new problem has become evident. This is the production of human anti-mouse antibodies (HAMA) resulting from a natural immune response. As the administered doses f antibody increase (>20mg) or multiple injections are given, the incidence of HAMA increases. The interference of HAMA on the accurate measurement of plasma levels of CEA has been documented, see Dillman, et al., J. Biol. Response Mod., 5:394-410, (1986); Jaffers, et al., Transplant, 41:572-9 (1986); Courtenay-Luck, et al., Cancer Res., 46:6489-93 (1986). Since this assay is crucial to the management of patients with CEA-producing tumors, removal of this interference becomes very important. The present frequency of colorectal cancer patients falling into this classification at the City of Hope National Medical Center, Duarte, California, is 1-2 patients per week or 50-100 patients per year. As the use of murine monoclonal antibodies for diagnosis and therapy becomes more established, this patient population will increase, and this problem will become more universal.

Modifications of the immunoassay for CEA designed to obviate the effect of HAMA include preincubation of the plasma with polyclonal'murine IgG (see Morton, et al., Arch. Surg., 123:1242-6 (1988) and heat extraction of the plasma (see Hansen, et al., Clin. Chem., 35(1):146-51 (1989). Recently the commercially available CEA-Rocha EIA has been modified by the addition of a single MAB to reduce the interference of potential heterophilic antibodies. An evaluation of this modification and its effect on HAMA interference in 30 patients who had previously received indium-lll labeled anti-CEA murine monoclonal antibody for imaging studies was made at City of Hope National Medical Center. The overall incidence of false elevation of CEA values in the modified kit dropped to 47% from seen with the unmodified kit. The incidence of false positive CEA titers was reduced from 36% to 21%, but was not completely eliminated.

The existing practice with respect to the CEA-Roche® EIA itself involves the addition of a single monoclonal to the kit.

This still results in a 22% false positive CEA for patients who have received a murine monoclonal antibody for diagnosis or therapy, see Price, et al., Clin. Chem., (submitted (1990).

Heat extraction of serum appears to be effective for the removal of interference for CEA assays but may not be appropriate for other analytes which are heat sensitive. Addition of nonspecific immunoglobulin from different species with the patient specimens has been reported.

Summary of the Invention The present invention provides a method for reducing or eliminating interference consequent from the presence of human anti-murine or heterophilic antibodies in a double antibody assay of a human plasma or serum sample which includes incubating said sample with a plurality monoclonal antibody isotypes prior to conducting said assay.

The present invention further provides a method for reducing or eliminating interference consequence from the presence of human anti-murine and of heterophilic antibodies in a double antibody enzyme immunoasssay for CEA in a human plasma or serum sample which includes incubating said sample with a mixture of at least three antibody isotypes prior to conducting said assay.

Pursuant to this invention falsely elevated analyte levels in double antibody EIA assays are corrected by the addition of a mixture of IgGl, IgG2a and IgG2b to human plasma specimens.

The advantages of the addition of a mixture of monoclonal antibodies to remove interference found in double monoclonal antibody assays for blood analytes, including CEA are as follows: 3 1) it avoids heat extraction; 2) it completely removes interference of HAMA and heterophilic antibodies to the same degree as polyclonal IgG; and 3) monoclonal antibodies can be produced in large quantities cost effectively for commercial use whereas polyclonal antibodies (murine particularly) cannot.

3a WO 91/16627 PCT/US90/02239 -4- Description of the Figures Figure l(a) and Correlation of CEA values obtained using the modified and unmodified CEA-Rocha EIA: a) Non-imaged control patients who had never received murine monoclonal antibodies b) Imaged patients who had received 20-40mg 111 In-labeled anti-CEA murine monoclonal antibody (Observations below the limit of CEA detectability were omitted from this graph but used, as described in the text, for estimating the correlation.) Figure 2. Plasma CEA values (ng/ml) obtained using the unmodified EIA mIgG and the modified EIA M) were expressed as a ratio of the CEA values measured by the unmodified EIA alone 3) for each of the four patient groups. The patient groups were: group 1, non-imaged controls; group 2, imaged patients with true positive CEA titers; group 3, imaged patients with false positive CEA titers; and group 4, imaged patients with partially suppressible CEA titers.

Figure 3(a) and The percentage of CEA positive (>5ng/ml) specimens in the true positive false positive ES) and partially suppressible categories are shown for the unmodified CEA-Rochd EIA (n=58) and the modified CEA-RochiEIA (n=47) WO 91/16627 PCT/US90/02239 Detailed Description of the Invention Heterophilic antibody interference in double antibody immunoassays has been a recognized problem in the quantification of hormones, enzymes, and cell surface antigens for diagnostic purposes. This interference can produce either false positive or false negative results. Boscato and Stuart reported that 40% of normal plasma contain detectable non-analyte substances which are capable of broad specificity antibody binding and have the potential of interfering with double antibody immunoassays, see Boscato, et al., Clin. Chem., 32(8):1491-5 (1986).

The administration of murine monoclonal antibodies to patients for diagnosis and treatment of various cancers has augmented this problem. The development of human anti-murine antibodies has been demonstrated in patients receiving murine MABs for assessment or treatment of hematological tumors and solid tumors. Along with the potential for interference with the original MABs and possible anaphylactic complications, HAMA appears to be the cause of false positive results in the current clinically used double antibody immunoassays for CEA.

Pursuant to this invention HAMA and heterophilic antibody interference in double EIA assays is reduced or eliminated by incubating human plasma or serum specimens prior to assay with a mixture of IgG isotypes, .lj, j-Ga and I.G2b.

Assays are conducted in known manner, by utilization, in the manner set forth in the accompanying instructions, of a commercially WO 91/16627 PCT/US90/02239 -6available CEA-Roche EIA kit. A publicly available package insert from such a kit has been concurrently filed and is incorporated herein by reference. The CEA-Roche EIA kit includes a hybridoma or cell line on deposit with the ATCC, Accession No. HB8747, and described and claimed as T84.66 in U.S. Patent 4,873,313, issued on October 10, 1989.

In the practice of the invention the mixture of monoclonal IgG isotypes (about 15 to about 25 pl), Ig2a, IgEa and Ig2b, is incubated at about to 50"C, preferably 37*C, for an appropriate time, 15 to 90 minutes, preferably from about 30 to minutes, with from about 75 to about 150 pl of a diluted plasma or serum specimen. The monoclonal antibody mixture, preferably but not necessarily, contains equal proportions by weight of each MAB.

Patient Population Thirty colon cancer patients who had previously undergone diagnostic immunoscintigraphy with 111 In-labeled anti-CEA murine monoclonal antibody (Hybritech, Inc., San Diego, CA) presented with unexplained elevations of plasma CEA. The CEA levels were routinely measured using a double antibody CEA-Roch& EIA (Roche Diagnostic System, Nutley, NJ). Sixty-one plasma samples from these patients banked for HAMA determination were available for this retrospective study. Fifty-eight of these plasma samples were found to have positive CEA values Re-assay of the 58 CEA positive samples fcllowing the addition of 200pg polyclonal mIgG to the plasma prior to assay resulted in a separation of the plasma samples into three categories determined by the level of CEA in the presence of added mIgG.

WO 91/16627 PCT/US90/02239 -7- The three categories were: 1. True positive CEA: Plasma with elevated CEA values (>5ng/ml) which were unaffected by the addition of MigG.

2. False positive CEA: Plasma with elevated CEA values which were brought down to normal levels upon the addition of mIgG and thus were classified clinically to be CEA negative.

3. Partially suppressible CEA: Plasma with elevated CEA values which were lowered by the addition of mIgG but still remained above the normal CEA values (>5mg/ml) and thus were classified clinically to be CEA positive.

Plasma samples of 13 patients who had never received murine antibodies and had CEA values ranging from normal (>5ng/ml) to 53.9ng/ml were used as controls.

CEA Assay The CEA level of each plasma sample was measured using both the standard (unmodified) CEA-Roche EIA kit (see the concurrently filed package insert) and a version of this kit modified by including a single MAB. For measurement of CEA using the unmodified EIA, all plasma samples were incubated with either PBS or 20pl (200pg) polyclonal mIgG for one hour at 37"C prior to performing the assay. The appropriate dilution corrections were included in the calculations. CEA measurements obtained using the modified EIA kit were performed on all plasma samples, according to the concurrently filed package instructions incorporated herein by reference. Each sample was tested in duplicate at two different dilutions. Seven plasma samples which continued to WO 91/16627 PCT/US90/02239 -8give falsely elevated CEA results with the Roche modified EIA were further tested following incubation with either polyclonal mIgG or irrelevant murine monoclonal antibodies of three different subclasses (IgG1, IgG2a, IgG2b) separately or as a mixture. In these assays the plasma samples were incubated with 100pg of mouse polyclonal IgG, 100g of each IgG subclass, or 100g of a 1:1:1 mixture of IgG1 IgG2a IgG2b prior to EIA. Samples were read using an automated EIA reader (Roche Diagnostic Systems, Nutley, NJ) and CEA values calculated according to the kit instructions. For purposes of this study, the lower limit of detection was considered to be

CEA.

HAMA Assay The level of human anti-murine antibody (HAMA) in each sample was measured using a non-competitive EIA. Plates were coated with the immunogen, murine monoclonal antibody ZCE025, and incubated for two hours at room temperature with serial dilutions of each plasma sample. This was followed by a one hour incubation at 37"C with affinity purified alkaline phosphatase-conjugated goat anti-human IgG. After the addition of the enzyme substrate, optical densities were read at OD 405 on an MR80 MicroELISA Reader (Dynatech, Alexandria, CA) interfaced with an Apple II Plus computer (Apple Computer Co., Cupertino, CA). The specific concentration of HAMA was then determined using a logit-log curve analysis computer progra previously developed, see Beatty, et al., J. Immunol. Meth., 100:161-72 (1987).

This program was modified to include an option for single point determination of values for those samples which did not provide a full curve for WO 91/16627 PCT/US90/02239 -9analysis. The positive control was plasma from a patient positive for HAMA which had been standardized against a human IgG standard of known concentration (amino acid analysis), see Beatty, et al., supra.

STATISTICS

Comparative statistical analyses of the data between pairs of groups were performed using Pearson's correlation coefficient and the paired Student's t test. Overall differences among the experimental groups were verified using repeated measures analysis of variance before the pairwise t tests were performed, and the pairwise tests were corrected for multiple comparisons. To satisfy assumptions necessary for the estimation of correlations, observations below the limit of CEA detectability (2.5ng/ml) were randomly assigned values in the range between zero and this limit, such that their expected value was the midpoint. All analyses were performed on the natural logarithms of the data to reduce skewness inherent in the CEA measurements.

RESULTS

Control Samples (non-imaged patients) Plasma samples obtained from a limited population of 13 patients that had a range of CEA values but who had never received a murine monoclonal antibody for diagnosis or therapy were run in the unmodified EIA mIgG and the modified EIA to compare CEA values and determine if addition of 200pg mIgG affected the measurement of CEA. The results are reported in Table 1.

o Table 1 Plasma CEA and HAMA Values in Non-Imaged Control Patients Plasma CEA a na/ml Plasma HAMAb ua/ml Unmodified Unmodifiedc Modified Pt. EIA EIA mIgG EIA 1 33.1 40.0 29.4 0.10 (0.01) 2 7.1 6.8 5.41 0.29 (0.02) 3 7.0 8.6 8.0 0.04 (0.00) 4 8.1 6.2 4.1 0.05 (0.00) 8.1 7.5 7.1 0.00 (0.00) 6 2.5 2.5 2.5 0.29 (0.01) 7 5.7 4.3 3.1 0.02 (0.00) 8 2.5 2.5 2.5 0.01 (0.00) 9 50.9 50.9 43.6 0.03 (0.02) 53.9 57.2 69.4 0.02 (0.00) 11 9.9 9.4 8.7 0.05 (0.00) 12 33.8 34.3 43.4 0.04 (0.00) 13 16.6 12.7 14.4 0.09 (0.00) a CEA measured using CEA-RocheR EIA.

b HAMA measured using indirect solid phase EIA (see text).

c Plasma preincubated with 200ug mIgG and assayed using unmodified Roche assay.

t~J "0 WO 91/16627 PCT/US90/02239 -11- A high degree of correlation between all three assays is demonstrated by the Table 1 data. The correlation between the CEA values obtained using the unmodified EIA mIgG was r=0.99, between the unmodified EIA mIgG and the modified EIA was r=0.99, and between the unmodified and the modified EIA was r=0.97 (Figure la). The mean CEA values S.D. obtained with the unmodified EIA, unmodified EIA mIgG and the modified EIA were 18.0 17.9, 18.3 19.2, and 18.1 20.8 respectively. These results indicate that adding 200pg mIgG to these patients' plasma prior to the CEA assay did not significantly alter the measured CEA levels.

Elevated plasma CEA (imaged patients) Plasma samples from 30 patients who had received 20-40mg 1 1 1 In-labeled anti-CEA MAB for tumor immunoscintigraphy and who developed a rising titer of CEA were used in this analysis. Of the 61 plasma samples obtained from these patients, 58 were found to have elevated (>5.0ng/ml) CEA values when measured using the unmodified EIA. Three samples had CEA values <5.0ng/ml and were not tested further. When these 58 samples were assayed using the modified EIA, the number of elevated CEA samples dropped to 47 Addition of 200kg mIgG to the plasma prior to running the unmodified EIA resulted in only 37 CEA positive (>5.0ng/ml) samples Thus, 21 samples, which were positive for CEA in the unmodified EIA, proved to be normal for CEA in the unmodiCied EIA run in the presence of MigG. Ten of these 21 samples continued to give positive CEA values in the modified CEA Roch_ EIA, although they displayed an overall average of 91.0% suppression when compared with the unmodified assay. The WO 91/16627 PCT/US90/02239 -12correlation coefficient between the unmodified and modified kits for this group of 58 patients (Figure Ib) was significantly lower (r=0.48; p<0.001) than for the control group.

The 58 plasma samples determined positive for CEA using the unmodified EIA were grouped into three categories based on the effect of MiGG incubation on the resultant plasma CEA value as described, infra.

The first category contained fifteen samples (26%) whose CEA values were not significantly (p>0.05) affected by the addition of mIgG, as shown by the data in Table 2.

Table 2 Plasma CEA and HAMA Valuesa for Imaged Patients: True Positive CEA Titers Plasma CEA nq/ml Plasma HAMA uq/ml Unmodified Unmodified Modified Pt. EIA EIA mIgG EIA 1157.0 45.0 19.2 28.3 45.0 81.7 44.5 540.0 92.9 16.9 38.1 20.2 529.0 23.1 (41.7) 1.6) 1.2) 1.0 1.5) 0.2) 2.5) 1.7) (38.3) 5.8) 0.1) 0.3) 0.9) (52.4) 0.4) 1141.0 44.0 17.3 12.1 41.5 49.0 93.3 46.6 586.0 91.3 17.0 37.8 21.9 500.0 22.8 (32.0) 0.2) 0.3) 0.3) 2.1) 0.8) 3.9) 3.0) (24.6) 1.7) 1.1) 0.9) 0.1) (11.0) 0.4) 1141.0 35.0 15.0 10.8 21.3 33.0 72.2 35.1 '500.0 93.8 14.5 33.4 17.0 515.0 18.2 (25.1) 9.4) 0.6) 0.2) 0.6) 4.5) 4.0) 0.7) (32.0) 2.7) 0.4) 0.1) 1.6) (14.9) 3.6) 0.03 0.05 0.03 0.07 0.07 0.04 0.02 0.09 0.08 0.04 0.08 0.08 0.08 0.07 0.02 (0.01) (0.01) (0.00) (0.01) (0.01) (0.01) (0.00) (0.00) (0.00) (0.00) (0.01) (0.00) (0.06) (0.00) (0.00) a CEA and HAMA were measured as described in text and Table WO 91/16627 PCT/US90/02239 -14- Addition of 200pg mIgG to these specimens resulted in a suppression of CEA values of less than 10% in all cases and gave an overall average increase of compared to the unmodified EIA. This gave a correlation of r=0.996 between the CEA values obtained using the unmodified kit with and without the addition of mIgG. These samples were termed "true positive" CEA samples. Similar good correlations were seen between the modified assay and the unmodified assay either with (r=0.994) or without (r=0.997) addition of mIgG. The paired t test analysis, however, showed that the CEA values obtained with the modified assay were significantly lower (p<0.001) than those obtained with the unmodified assay with or without addition of mIgG.

Twenty-one of the 58 elevated CEA samples (36%) were suppressed to normal levels ng/ml) upon the addition of mIgG when run in the unmodified assay, and were thus considered as false-positive CEA elevations. See Table 3.

Table 3 Plasma CEA and RAMA Valuesa for Imaged Patients: False Positive CEA Titers Plasma CEA nev/mi is n i Plasma RAMA un/mi IS.D.1 Unmodified Unmodified Modified EIA EIA mIgG Suppression EIA Suppression 1 228.0 6.4) 4.85 97.9 23.0 89.9 35.4 (2.23) 2 172.8 2.4) <2.5 100.0 2.5 100.0 8.73 (0.47) 3 134.3 (13.4) 4.0 97.0 10.6 92.1 12.3 (0.75) 4 129.1 3.4 97.4 2.9 97.8 11.8 (0.07) 134.4 <2.5 100.0 2.5 100.0 11.0 (0.97) 6 34.9 4.3 87.7 6.3 81.9 6.2 (0.01) 7 59.0 2.9) <2.5 100.0 3.4 94.2 0.49 (0.00) 8 457.4 <2.5 100.0 23.6 94.8 84.8 (0.59) 9 264.8 (6.1b 4.5 98.3 10.0 96.2 32.1 (1.44) 194.8 3.31 100.0 7.9 95.9 7.72 (0.25) 11 12.0 <2.5 100.0 2.5 100.0 27.8 (1.25) 12 307.1 (30.7) <2.5 100.0 14.4 95.3 1.50 (0.12) 13 106.5 1.1) <2.5 100.0 16.0 85.0 8.74 (0.73) 14 118.3 (14.7) <2.5 100.0 2.5 100.0 3.29 (0.00) 142.8 4.9) <2.5 100.0 2.5 100.0 0.28 (0.02) 16 93.0 3.7) <2.5 100.0 10.0 89.2 0.22 (0.02) 17 22.7 0.5) <2.5 100.0 2.5 100.0 0.39 (0.01) 18 38.6 <2.5 100.0 2.5) 100.0 5.95 (0.02) 19 61.7 <2.5 100.0 2.5) 100.0 0.17 (0.00) 70.4 <2.5 100.0 7.5 89.3 0.24 (0.01) 21 8.9 <2.5 100.0 2.5 100.0 0.06 (0.00) mean 98.9 mean 95.3 a CEA and RAMA were measured as described in text and Table 1.

b t suppression 1 value of experimental EIA x 100 value of unmodified EIA WO 91/16627 PCT/US90/02239 -16- These same samples, when assayed using the modified EIA, also showed suppression of all 21 samples. The overall average percent suppression, when compared to the original unmodified EIA, was 95.3% for the modified Roche EIA versus the 98.9% suppression for the unmodified EIA with addition of mouse IgG. In this group of 21 patients, eleven samples demonstrated CEA values below 5ng/ml in the modified Roche EIA kit and the remaining ten samples continued to be falsely positive Thus, the CEA values obtained using either the unmodified assay alone or the modified assay correlated poorly with the unmodified assay in the presence of mIgG (r=0.20 and r=0.36, respectively).

In the remaining 21 elevated CEA plasma samples the addition of mIgG to the unmodified assay significantly lowered the CEA values (p<0.00001), but did not return them to normal levels. See Table 4.

Table 4 Plasma CEA and HlAMA Valuesa for Imaged Patients: :artially Suppressible CEA Titers Plasma CEA ng/ml Plasma HlAMA ug/ml Unmodified Unmodified Modif led Pt. EIA EIA MigG Suppresslonb EIA Suppressionb 1 170.5 (11.4) 65.7 5.7) 61.5 99.5 0.0) 41.6 3.78 0.35) 2 177.9 2.3) 66.4 4.0) 62.7 119.4 8.2) 32.9 8.02 1.03) 3 93.0 2.7) 48.4 0.7) 48.0 71.5 2.3) 23.1 7.98 0.83) 4 67.4 3.5) 46.9 0.7) 30.4 40.7 0.5) 39.6 6.12 0.51) 290.2 (22.6) 84.2 (22.3) 71.0 70.4 (20.6) 75.7 65.6 (21.78) 6 157.8 4.1) 6.3 0.8) 96.0 30.0 0.4) 81.0 39.2 9.92) 7 125.0 1.9) 89.0 2.5) 28.8 78.0 1.5) 37.6 4.10 0.10) 8 209.5 0.4) 165.9 20.8 163.0 22.2 4.53 (1.41) 9 34.0 2.9) 6.0 .82.4 5.0 85.3 10.8 (3.78) 28.1 1.0) 8.80 0.2) 68.7 13.4 52.3 0.93 (0.16) 11 22.4 0.2) .8.8 60.7 13.2 41.1 0.76 (0.05) 12 27.8 1.2) 10.2 63.3 17.7 C0.7) 36.3 0.61 (0.06) 13 201.6 (22.2) 6.9 0.1) 96.6 45.0 2.2) 77.7 7.47 (0.86) 14 493.7 (44.9) 11.4 0.4) 97.7 56.4 2.2) 88.6 14.5 (10.32) 53.0 1.3) 23.0 2.8) 56.6 19.7 0.7) 62.6 7.15 2.92) 16 159.6 2.7) 129.7 3.5) 18.8 131.9 1.9) 17.5 7.53 2.14) 17 280.3 (10.4) 25.2 2.6) 91.0 8.7 0.6) 96.9 7.24 0.43) 18 170.0 8.0) 8.3 1.9) 95.1 40.2 0.1) 76.4 0.22 0.06) 19 55 9 2.1) 8.2 0.2) 85.3 11.2 1.1) 80.0 9.91 4.50) 67.9 1.4) 13.6 0.5) 80.0 44.6 0.2) 34.3 0.38 0.06) 21 820.8 (31.2) 61.7 6.9) 92.5 47.8 7.2) 94.2 31.0 (10.32) 22 32.4 2.6) 21.3 0.1) 34.3 15.2 0.7) 53.1 0.05 0.00) mean 65.5 mean 56.8 a CEA and R-AMA were measured as described in text and Table 1.

b As described in Table 3.

WO 91/16627 PCT/US90/02239 -18- These samples were termed partially suppressible as they were CEA positive plasma but the actual levels of CEA were substantially lower than the apparent levels obtained using the unmodified assay.

With the addition of mIgG to the unmodified assay, samples in this group showed suppression of CEA values of >18% with an overall mean of 65.7%. The average percent suppression obtained with the modified kit for this group of samples was 56.8% which was significantly less (p=0.023) than the suppression achieved with the unmodified kit mIgG.

The correlation coefficient between the CEA values obtained using the modified EIA and the unmodified EIA mIgG was 0.73. This correlation was better than that found in the false-positive group Using the paired t-test, 12 of the 22 samples (54.5) in this group of patients had significantly higher CEA values (p<0.05) with the modified assay than with the unmodified assay mIgG. Thus approximately half of the partially suppressible samples continued to give falsely elevated CEA titers with the modified EIA.

Polyclonal mIgG was added to 16 plasma samples which showed significantly higher CEA values (p<0.001) when assayed using the modified EIA compared to those obtained with the unmodified EIA mIgG. See Table Table Effect of Polyclonal MigG on Plasma CEA Values Measured by the modified EIA: False Positive and Partially Suppressible Plasma Specimens Unmodified Modified Modified Pt. Category EIA mIgGa EIA EIA mIgGa ng/ml ng/mi ng/ml f.p.

f.p.

f.p.

f.p.

f.p.

f.p.

f.p.

f.p.

f.p.

p.s.

p.s.

p.s.

p.s.

p.s.

p.s.

p.s.

4.9 4.3 2.5 4.5 3.3 2.5 2.5 2.5 2.5 65.7 66.4 48.4 6.3 6.9 11.4 8.3 (0.8) (0.1) (0.0) (1.2) (0.8) (0.0) (0.0) (0.0) (0.0) (5.7) (4.0) (0.7) (0.8) (0.1) (0.4) (1.9) 23.0 6.3 23.6 10.0 7.9 14.4 16.0 10.0 7.5 99.5 119.4 71.5 30.0 45.0 56.4 40.2 (0.4) (0.0) (1.7 (0.6) (0.1) (0.1) (0.1) (0.3) (0.7) (0.0) (8.2) (2.3) (0.4) (2.2) (2.2) (0.1) 2.8 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 62.2 31.3 35.9 9.3 5.0 10.3 2.5 0.0) 0.0) 0.0) 0.0) 0.0) 0.0) 0.0) 0.0) (22.3) 1.1) 4.9) 0.7) 0.3) 0.8) 0.0) f.p. False positive CEA specimens p.s. Partially suppressible CEA specimens a 200ug mIgG were added to plasma samples prior to assay using either unmodified or modified Roche EIA.

WO 91/16627 PCT/US90/02239 These samples which continued to demonstrate HAMA interference with the modified EIA contained both false positive CEA and partially suppressible CEA samples. When these samples were reassayed with the modified EIA following incubation with mIgG the CEA titers were lowered to values which were comparable to or lower than those seen with the unmodified mIgG assay. All the false positive CEA plasma which had remained positive with the modified EIA were shown to be negative with the further addition of polyclonal mIgG. Similarly, the partially suppressible CEA specimens gave significantly lower (p<0.025) CEA titers with the addition of mIgG than without. These corrected values (modified EIA mIgG) were not significantly different from the values obtained with the unmodified EIA mIgG (p=0.12).

The question of whether the addition of a single irrelevant monoclonal murine IgG could be as effective as that of the polyclonal mIgG in reducing interference still evident in the modified assay was examined in 7 samples. Again, see Table 5. These specimens were comprised of 5 false positive and 2 partially suppressible plasma samples. Equal amounts (100pg) of either polyclonal mIgG or irrelevant monoclonal antibodies of different subclasses (IgG1, IgG2a, IgG2b) alone or mixed together were then incubated with the plasma prior to assay. The results shown in Table 6 indicate that the most consistent removal of the artifactual CEA values remaining in the modified EIA was achieved with either the addition of polyclonal mIgG or the combination of the IgG1, IgG2a and IgG2b monoclonal antibodies. The addition of the individual monoclonal antibodies also decreased the CEA levels measured in each case; however, the effectiveness was more variable.

Table 6 Effect of Plyclonal xnlgG vs. Monoclonal mIgG on the Modified EIA Kit Determination of CEA in False Positive and Partially Suppressible Plasma Specimens Plasma CEA ng/ml Polyclonal Monoclonal Monoclonal Monoclonal mIgG mIgGi mIgG2a mIgG2b Monoclonal Sample PBS (lO0ug) (1O0ug) (0oug) (0oug) Mixtured, ,a 12.1 6 )c <2.5 5.0 <2.5 <2.5 <2.5 (.04) 2a 9. 4 <2.5 5.3 <2.5 3.0 <2.5 3 a 12.8 <2.5 2.5 <2.5 <2.5 <2.5 4a 6.6 <2.5 <2.5 <2.5 <2.5 <2.5 sa 15.9 <2.5 5.0 3 8.6 8.8 3.7 6 b 23.3 8.4 (1.1 8.5 (1.1 7.1 9.5 8.4 7b 28.2 10.0 (2.0 11.8 (1.5 11.2 10.0 10.0 (A4 a False positive CEA specimens b Partially suppressible CEA specimens c S.D.

d mIgGl mIgG2a mIgG2b lO0ug mixture) WO 91/16627 PCT/US90/02239 -22- HAMA Values The normal range of HAMA for the indirect EIA used in this study was determined from a sampling of 44 patients who had never received murine antibodies and who had normal plasma CEA levels. The normal range, obtained from the mean HAMA value plus two standard deviations, was 0.36pg/ml. As seen in Table 1, all of the non-imaged control patients had HAMA values in the normal range. All samples in the true positive CEA group (Table 2) also showed normal HAMA values.

HAMA values for the plasma in the false positive CEA group (Table 3) showed a broad range from normal levels (<0.36yg/ml to 84.5pg/ml. Five samples were <0.36pg/ml. The degree of false elevation of CEA in the unmodified EIA did not show a significant relationship to the measured level of plasma HAMA using a linear regression analysis, p=0.065. The titer of HAMA in the samples that continued to give positive CEA (>5.0ng/ml) values with the modified EIA were on the whole higher (18.9 24.9) than those that gave negative (<5.0ng/ml) CEA values (6.99 8.17). However, this difference was not statistically significant due to the wide spread of values and concomitant high standard deviations.

Thus no statistical correlation between HAMA level and ability of the modified EIA to correct the false positive CEA values was obtained.

In the partially suppressible CEA group (Table 4) the HAMA values also ranged from normal levels (two samples) to high values Gpg/il). In this patient population, there was a significant relationship between the level of HAMA and the degree of false elevation of CEA in the unmodified EIA WO 91/16627 PCT/US90/02239 -23relative to both the unmodified EIA mIgG (p=0.013) and to the modified EIA (p=0.007). The degree of HAMA interference in the unmodified kit was also shown to have a significant relationship with the level of HAMA when the samples from Tables 3 and 4 were combined and similarly analyzed by a linear regression analysis (p=0.0019 with unmodified EIA mIgG and p=0.0013 with the modified EIA, n=43).

Interestingly, no statistically significant relationship was found between the level of HAMA and the difference in suppression of CEA levels obtained with the modified assay and the unmodified assay mIgG In those partially suppressible samples in which the CEA values were significantly higher in the modified EIA than the unmodified EIA mIgG, the HAMA values were lower but not significantly (p>0.15) than those whose CEA values correlated well between these assays. Thus the failure of the modified EIA to correctly lower the CEA values in these specimens was not related to the level of HAMA.

The CEA values of the samples obtained with modified EIA and the unmodified EIA mIgG were expressed as a ratio of the corresponding values obtained in the standard unmodified EIA alone (Figure This was done for each of the four patient groups. Plasma CEA values were similar in all three assays for patients who had not been given murine antibodies (control group or who had little or no measurable HAMA (true positive group However, for patients with measurable HAMA (groups 3 and 4), there was a substantial difference between the CEA values obtained using the unmodified assay and those obtained with the other two assays. The most marked WO 91/16627 PCT/US90/02239 -24differences were seen in group 3, the false positive group in which an overall average percent suppression of 95.3% and 98.9% was obtained with the modified Roche EIA and the unmodified Roche EIA mIgG respectively (Table 3).

It has been shown that exogenous addition of 200g polyclonal mouse IgG (mIgG) to patients' plasma samples (after dilution for the assay) will remove any HAMA effects and result in a true measurement of CEA. This amount of mIgG was chosen to be well in excess of any HAMA levels present, as the plasma samples are diluted according to their apparent CEA titer so they fall between 5 and 20ng/ml CEA (linear portion of the standard curve) prior to addition of the mIgG. Even if the apparent CEA is entirely due to HAMA, this amount of mIgG is far in excess of the HAMA present in the diluted plasma samples. The level of mIgG added to the samples is therefore not deemed to be a limiting factor in the ability to reduce apparent CEA values.

It has also been shown that addition of mIgG does not affect the assay itself (Tables 1 and 2) in the absence of HAMA. The CEA values remained virtually unchanged (r=0.99, p>0.05) for both the control group (n=13) and the true positive group with addition of mIgG to the unmodified assay. However, in the presence of HAMA, the addition of mIgG to the unmodified assay to block the HAMA interference reduced the cEA values significantly resulting in 21/58 false positive specimens and 22/58 partially suppressible specimens (Figure 3a).

Previously published results report the incidence of falsely elevated CEA values in colon WO 91/16627 PCT/US90/02239 cancer patients who had received indium-ll labeled anti-CEA monoclonal antibodies for diagnostic imaging, see Morton, et al., Arch. Surq..

123:1242-6 (1988). Six of eight patients given who demonstrated a false elevation of CEA (with no other indication of disease) had measurable HAMA. The CEA assay employed was the Roche double antibody EIA. In this study 36/43 of the patients who demonstrated a false elevation of CEA titer in the unmodified EIA had measurable HAMA. As noted previously, Morton, et al., supra, this shows a strong correlation between the presence of measurable HAMA and false elevation of CEA levels. As well, a significant relationship between the level of HAMA and the degree of false elevation of CEA was demonstrated for the combined group of false positive and partially suppressible CEA samples (p<0.005).

This relationship was much stronger in the partially suppressible samples and was actually just shy of being significant in the false positive samples (p=0.065). A larger sample size will be necessary to look at this relationship more carefully. However, these observations agree with those of Hansen, H.J., et al., Clin. Chem., 35(1):146-51 (1989), who also showed a direct correlation between HAMA titer and degree of interference of CEA determination for three commercially available CEA immunoassays. This group found the unmodified Roche EIA to be the most sensitive to HAMA interference. The data set forth in this application does not establish any relationship between the level of HAiA and the ability of the modified EIA to correct the falsely elevated CEA values in either the false positive CEA group or the partially suppressible group.

WO 91/16627 PCT/US90/02239 -26- The data gathered from this study clearly indicate that while the newly revised CEA-EIA assay kit is much improved in its incidence of false-positive results, it is unable to satisfactorily suppress HAMA interference to the degree that the unmodified assay with the addition of mIgG proves is possible. Overall, the modified assay gave 10/47 false positive CEA determinations and 12/47 partially suppressible CEA titers (Figure 3b). In the group of 21 samples which were found to be false positive for CEA using the unmodified assay, 10 continued to show false positive CEA values with the modified assay. It is this sub-group of imaged patients which continued to give positive CEA titers (>5ng/ml) with the modified Roche EIA but were negative (<5.0ng/ml) on addition of mIgG that is the critical patient group.

Clinically, plasma levels of CEA are used in the follow-up management of patients who have undergone a potentially curative operative procedure. Increases in the CEA titer are an early indicator of residual or recurrent colorectal cancer and as such may result in further diagnostic tests or surgical explorations second look laparotomy) for localization of the site of the disease, see Shively, et al., Crit. Rev. Onc./Hem., 2:355-399 (1985). Thus, a false elevation of the plasma CEA may lead to extensive unnecessary testing or even unnecessary surgical procedures. Therefore, elimination of false positive CEA values is clinically most desirable.

In the group of 22 partially suppressible specimens, 12 continued to show significantly higher CEA values (p<0.05) using the modified assay when compared to the values shown with the unmodified WO 91/16627 PCT/US90/02239 -27assay mIgG, although interestingly, this difference was not directly related to the level of HAMA. Thus for both the false positive and partially suppressible groups, the modified EIA corrected the HAMA interference in approximately half of the samples tested. Sixteen of these patient plasma samples in which there was a discrepancy in value between the unmodified assay mIgG and the modified assay were reassayed with the addition of polyclonal mIgG to the plasma prior to running the modified kit. In each case, addition of the mIgG to the modified assay returned the CEA to levels comparable to those obtained in the unmodified assay plus mIgG.

As a result of this observation, it is desirable that all patients who have received radiolabeled anti-CEA monoclonal antibodies for tumor scintigraphy prior to surgery and who demonstrate apparent elevated CEA values post surgery have the modified EIA rerun with the addition of mIgG. In this way falsely elevated plasma CEA values may be avoided.

Since the imaging agent administered to the patients was a murine IgG1 monoclonal antibody, the effectiveness of adding murine monoclonal antibodies to correct the remaining falsely elevated CEA samples in the modified EIA was examined. It was found that addition of each monoclonal antibody alone had some effect on decreasing the CEA levels. However, the exemplified combination of the three monoclonals 100/g total) gave CEA values which most consistently reflected those produced with the addition of polyclonal mIgG. Interestingly, the addition of 100 g of IgGl MAB was not as effective as the mixture in eliminating HAMA interference, even though the immunogen given to the patients was an WO 91/16627 PCT/US90/02239 -28- IgG1 monoclonal antibody. Why removal of HAMA interference is more effective using a polyclonal mIgG or a mixture of monoclonal IgG subclasses remainJ unclear.

The clinical significance of these data is that for this unique population of patients previously exposed to murine antibodies there remains a 47% error in CEA determinations with a 21% false positive incidence when measured with the current CEA-Roche EIA. Since CEA levels in plasma or serum are used as a diagnostic tool for assessment of recurrent disease in colorectal cancer patients, it is crucial that the assay for this tumor associated antigen be as accurate as possible. The false positive results produced by the presence of HAMA in patients who have received murine monoclonal antibodies and whose CEA titers are measured by double antibody assay can be eliminated by blocking with addition of MIgG. The results presented here indicate that effective blocking of HAMA is not achieved by a single monoclonal antibody IgG but requires a mixture of IgG subclasses, or a polyclonal IgG. The advantage of blocking HAMA interference with mIgG over the heat extraction method is that it can be used in other double antibody assays when the analyte may not be as heat resistant as CEA.

Addition of high concentrations of polyclonal mIgG to kit components is not commercially feasible.

Hence, the use of a mixture of irrelevant monoclonal IgGs, pursuant to this invention, provides a more appropriate procedure.

Claims (8)

1. A method for reducing or eliminating interference consequent from the presence of human anti-murine or heterophilic antibodies in a double antibody assay of a human plasma or serum sample which includes incubating said sample with a plurality monoclonal antibody isotypes prior to conducting said assay.

2. A method as defined by claim 1 in which said assay is an enzyme immunoassay.

3. A method as defined by Claim 1 or Claim 2 in which said assay is for carcinoembryonic antigen (CEA).

4. A method for reducing or eliminating interference consequent from the presence of human anti-murine and of heterophilic antibodies in a double antibody enzyme immunoassav for CEA in a human plasma or serum sample which includes incubating said sample with a mixture of at least three antibody isotypes prior to conducting said assay.

A method, as defined by Claim 4, in which said isotypes are IgGi, IgG2a and IgG2b.

6. A method, as defined by Claim 5, in which from about 100 to 1501p of said plasma or serum sample are incubated at a temperature of from about to about 50'C for about 15 to about 90 minutes with from about 75 to about 150pl of said mixture of antibody isotypes.

7. A method, as defined by Claim 6, in which each antibody isotype is present in said mixture in about equal proportions by weight.

8. A method according to claim 1 substantially as hereinbefore described with reference to any one of the examples. DATED: 22 November, 1993 PHILLIPS ORMONDE FITZPATRICK Attorneys for: E CITY OF HOPE u r o I r