Quantitative determination of humanized monoclonal antibody rhuMAb2H7 in cynomolgus monkey serum using a Generic Immunoglobulin Pharmacokinetic (GRIP) assay

Journal of Immunological Methods 335 (2008) 8 – 20 www.elsevier.com/locate/jim

Research paper

Quantitative determination of humanized monoclonal antibody rhuMAb2H7 in cynomolgus monkey serum using a Generic Immunoglobulin Pharmacokinetic (GRIP) assay Jihong Yang a,⁎, Carl Ng a , Henry Lowman b , Regina Chestnut a , Cheryl Schofield a , Bryan Sandlund a , James Ernst c,d , Gregory Bennett e , Valerie Quarmby a a

Department of Bioanalytical Research & Development, Genentech, Inc., South San Francisco, CA, United States b Department of Antibody Engineering, Genentech, Inc., South San Francisco, CA, United States c Department of Protein Engineering, Genentech, Inc., South San Francisco, CA, United States d Department of Protein Chemistry, Genentech, Inc., South San Francisco, CA, United States e Department of Assay & Automation Technology, Genentech, Inc., South San Francisco, CA, United States Received 14 September 2007; received in revised form 18 January 2008; accepted 23 January 2008 Available online 14 February 2008

Abstract Preclinical pharmacokinetic (PK) assays are important to help evaluate the safety and efficacy of a potential biotherapeutic before clinical studies. The assay typically requires a biotherapeutic-specific reagent to minimize matrix effects especially when the host species are non-human primates such as cynomolgus monkeys and the biotherapeutic is a humanized monoclonal antibody (MAb). Recombinant humanized mAb 2H7 (rhuMAb2H7) binds to the extracellular domain of CD20 that is expressed on B cells and results in B cell depletion. It is currently being evaluated for its therapeutic potential in rheumatoid arthritis (RA) in clinical studies. During the early development of rhuMAb2H7, a cynomolgus monkey PK assay was needed to help assess the pharmacokinetic parameters of rhuMAb2H7 in a pilot cynomolgus monkey study. However, development of a cynomolgus monkey PK assay was challenging due to lack of rhuMAb2H7-specific reagents. Here we describe an alternative method for detection of rhuMAb2H7 in cynomolgus monkey serum using polyclonal antibodies against human IgGs. This assay quantifies rhuMAb2H7 in 10% cynomolgus monkey serum with high sensitivity, accuracy, and precision. This assay successfully supported the rhuMAb2H7 development,

Abbreviations: ADCC, antibody dependent cell-mediated cytotoxicity; BGG, bovine γ-immunoglobulin; CHAPS, 3-[(3-cholamidopropyl) dimethylammonio]-1-propane-sulfonate; CDC, complement dependent cytotoxicity; CPG, controlled pore glass; ELISA, enzyme-linked immunosorbent assay; ECD, extracellular domain; GRIP assay, Generic Immunoglobulin Pharmacokinetic assay; HRP, horseradish peroxidase; LLOQ, lower limit of quantification; MAb, monoclonal antibody; PK, pharmacokinetic; PD, pharmacodynamic; PBS, phosphate-buffered saline; rhuMAb2H7, recombinant humanized mAb 2H7; RA, rheumatoid arthritis; TMB, 3,3′,5,5′-tetramethylbenzidine; ULOQ, upper limit of quantification; CV, coefficient of variation; POC, proof of concept. ⁎ Corresponding author. Bioanalytical Research & Development, Genentech, Inc., South San Francisco, CA 94080, United States. Tel.: +1 650 225 6638; fax: +1 650 225 1998. E-mail address: [email protected] (J. Yang). 0022-1759/$ - see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jim.2008.01.016

J. Yang et al. / Journal of Immunological Methods 335 (2008) 8–20

9

and has the potential to be used to quantify other humanized MAb biotherapeutics in serum from a variety of non-human species. © 2008 Elsevier B.V. All rights reserved. Keywords: Pharmacokinetic assay; Cynomolgus monkey serum; Humanized IgG; ELISA; Anti-CD20; rhuMAb2H7

1. Introduction rhuMAb2H7 is a humanized IgG1 kappa monoclonal antibody derived from the murine precursor 2H7 (Clark et al., 1985). The antibody is directed against the extracellular domain of the CD20 antigen, which is expressed on both normal and malignant B cells (Stashenko et al., 1980; Tedder and Schlossman, 1988; Clark and Ledbetter, 1989; Tedder and Engel, 1994; Riley and Sliwkowski, 2000). B cell depleting agents have been used successfully in treating B cell-mediated malignances such as non-Hodgkin's lymphoma (McLaughlin et al., 1998) and chronic lymphocytic leukemia (Jensen et al., 1998; Gopal and Press, 1999; von Schilling, 2003). In addition, B cells are also implicated in the pathophysiology of autoimmune diseases such as rheumatoid arthritis (Looney, 2002; Dorner and Burmester, 2003; Oligino and Dalrymple, 2003; Shaw et al., 2003), systemic lupus erythematosus (Anolik et al., 2003), and multiple sclerosis (Corcione et al., 2005). Treatment of non-human primates in vivo with rhuMAb2H7 results in binding of rhuMAb2H7 to the CD20 antigen on circulating B cells, a process that subsequently causes B cell depletion (Vugmeyster et al., 2005). Although the precise mechanism of B cell depletion by rhuMAb2H7 is not completely understood, it may include several different pathways such as antibody dependent cellmediated cytotoxicity (ADCC), complement dependent cytotoxicity (CDC), phagocytosis and apoptosis. Both in vitro and in vivo studies suggested that rhuMAb2H7, an anti-CD20 therapeutic, has a therapeutic potential for B cell-mediated autoimmune disorders and hematologic indications (Gopal and Press, 1999; Kosmas et al., 2002; Moore et al., 2004; Keystone, 2005; Panayi, 2005; Tuscano et al., 2005; Vugmeyster et al., 2005; Isenberg, 2006). Cynomolgus monkeys have been frequently used in preclinical studies due to their physiological and genetic similarities to humans (Da Silva et al., 2001; Hart et al., 2001; Kelley et al., 2001; Boon et al., 2002; Dodds et al., 2005; Ponce et al., 2005). During the early development of rhuMAb2H7 a proof of concept (POC) study was conducted in this species to assess the safety, pharma-

cokinetics (PK) and pharmacodynamics (PD) of the molecule. Therefore a sensitive bioanalytical method that could accurately quantify rhuMAb2H7 concentrations in cynomolgus monkey serum was needed. In general, a PK assay that quantifies the concentration of a biotherapeutic in serum requires one or more biotherapeutic-specific reagents. Biological matrices tend to give high background due to nonspecific interactions between matrix components and assay reagents (Selby, 1999). Such specific reagents are highly desired for rhuMab2H7, a humanized IgG1 that shares a high sequence identity with cynomolgus monkey IgGs (Lewis et al., 1993; D'Ovidio et al., 1994; Pace et al., 1996). Alignment of the heavy chain sequences of both rhuMAb2H7 and the cynomolgus monkey IgG reveals an overall of 80% of amino acid identity and 85% of sequence similarity. In addition, the circulating levels of cynomolgus IgGs are usually in the range of 10–16 mg/ml, which is much higher than the concentration of the therapeutic to be quantified (Biagini et al., 1988; Tryphonas et al., 1991). In the rhuMAb2H7 cynomolgus POC study, a sensitivity of 20 ng/ml in serum was required. Quantitative enzyme linked immunosorbent assay (ELISAs) have been widely used to measure circulating levels of biotherapeutics in a biological matrix. These assays are typically based on analyte-specific reagents such as monoclonal and polyclonal antibodies directed against the therapeutics, target proteins or receptors (DeSilva et al., 2003). During the early development of rhuMAb2H7, the only available method that potentially could measure rhuMAb2H7 concentration was based on a WIL2-S cell line from American Type Culture Collection (Rockville, MD) that overexpresses CD20 on its surface (Hong et al., 2004). However, this assay was not suitable for developing a cynomolgus monkey PK assay with a high sensitivity and throughput. Specific polyclonal and monoclonal antibodies against rhuMAb2H7 were not available at the time of the POC pilot cynomolgus monkey study. CD20, the target of rhuMAb2H7, was only available in the form of peptides that resemble the C-terminal extracellular domain (ECD) of the molecule. The low binding affinities of these

10

J. Yang et al. / Journal of Immunological Methods 335 (2008) 8–20

peptides for rhuMAb2H7 precluded their utility as appropriate PK assay reagents. A stable full length CD20 assay reagent became available only after the cynomolgus monkey pilot study. To support the pilot cynomolgus monkey study we explored a bioanalytical PK assay development strategy that is independent of a rhuMAb2H7-specific reagent. The assay method of choice utilized sheep anti-human IgG that had been adsorbed against cynomolgus monkey serum for capture as well as detection of rhuMAb2H7. The assay is highly sensitive, accurate, specific, robust, and rugged, and was used successfully to support PK analysis for the rhuMAb2H7 pilot cynomolgus monkey study. In addition, this assay was also qualified for mouse and rat sera. Because the assay reagents are human IgG specific rather than only rhuMAb2H7 specific, the assay has more general utility and can potentially be used to quantify other human or humanized IgG molecules in non-human biological matrices including non-human primate and rodent serum. 2. Materials and methods 2.1. Materials The humanized monoclonal antibodies rhuMAb2H7, Herceptin®, Xolair®, Avastin® and Raptiva® were generated at Genentech, Inc. Goat anti-human IgG (H + L) HRP conjugate was purchased from American Qualex. Monkey serum adsorbed sheep anti-human IgG heavy and light (H + L) chain specific antibody (catalog no. CUS1684) and monkey serum adsorbed sheep anti-human IgG heavy and light (H + L) chain specific antibody conjugated to horse radish peroxidase (catalog no. CUS1684.H) were purchased from The Binding Site (San Diego, CA). Individual cynomolgus monkey sera were obtained from BioChemed (VA). Maxisorp Nunc-immuno 96-well microtiter plates were purchased from Nalge Nunc International (Rochester, NY). HRP substrate 3,3′,5,5′-tetramethylbenzidine (TMB) and hydrogen peroxide (H2O2) were purchased from KPL (Maryland). Bovine serum albumin (Bovuminar® Cohn Fraction V, pH 7) was obtained from Serologicals Corp (catalog no. 3322-90, Ontario, Canada) and Proclin 300 was from Supelco (Bellefonte, PA). A 20× solution of PBS that contains 1% Polysorbate 20 was purchased from MediaTech Cellgro, both bovine γ-immunoglobulin (BGG) and 3-[(3-cholamidopropyl)dimethylammonio]1-propane-sulfonate (CHAPS) were from Sigma. An EL 404 microplate autowasher from Bio-Tek Instruments, Inc. was used for all the washing steps in the ELISA. A SpectraMax 250 plate reader (Molecular Devices Cor-

poration, CA) was used to record ELISA signals using absorbance at 450 nm with background subtraction at 650 nm. 2.2. Generic Immunoglobulin Pharmacokinetic (GRIP) ELISA for measuring rhuMAb2H7 in cynomolgus monkey serum Each 96-well microtiter plate was coated with 100 µl of 1 µg/ml of sheep anti-human IgG (H + L) (monkey serum adsorbed) in sodium carbonate (pH 9.6) at 2–8 °C overnight. After washing with PBS/0.05% polysorbate20, the plate was incubated with 200 µl of a blocking buffer (PBS/0.5% BSA/0.05% P20/0.05% Proclin300/ 0.25% CHAPS/0.2% BGG/5 mM EDTA/0.35 M NaCl/ pH = 8.0) at room temperature for 2 h with gentle agitation. After washing, 100 µl of serially diluted assay standards, controls, serum blanks or samples with a 1:10 minimum dilution in sample/conjugate assay diluent (PBS/0.5% BSA/0.05% P20/0.05% Proclin300/0.25% CHAPS/5 mM EDTA/0.35 M NaCl/pH = 8.9) were added to the plate. After incubating at room temperature for another hour with gentle agitation, the plate was washed again before incubating with 100 µl of the antibody conjugate (monkey serum adsorbed sheep antihuman IgG (H + L) conjugated to HRP) at an optimized dilution in sample/conjugate assay diluent. The plate was incubated for 1 h with gentle agitation and was then washed. After that, 100 µl of freshly mixed TMB chromogen solution (from equal volumes of TMB and H2O2) was added. Color was allowed to develop for 15 min without agitation, and the reaction was stopped by the addition of 100 µl of 1 M H3PO4. The absorbance was read at a wavelength of 450 nm with 650 nm reference on a Spectra Max250 plate reader (Molecular Devices Corporation, CA), and the data were processed using the SoftmaxPro software provided by the manufacturer. Concentrations of rhuMAb2H7 in samples were interpolated from a four parameter fit of the standard curve on the same plate. 2.3. Adsorptions of the sheep anti-human IgG heavy and light (H + L) chain specific antibody with cynomolgus monkey IgGs to remove cross-reactivity Cynomolgus monkey IgGs were purified from cynomolgus monkey serum by a HiTrap Protein G column (Pharmacia) following the procedures recommended by the manufacturer. Briefly, each column was washed with water and then equilibrated with 20 mM sodium phosphate, pH 7.0. Approximately 1 ml aliquots of two individual cynomolgus monkey serum samples that gave

J. Yang et al. / Journal of Immunological Methods 335 (2008) 8–20

11

Fig. 1. Dose-dependent responses of rhuMAb2H7 in the GRIP assay. (a) Comparisons of rhuMAb2H7 standard curves in different concentrations of cynomolgus monkey serum; (b)–(e) standard curves of rhuMAb2H7 with either 0% (open square) or 10% cynomolgus monkey serum (cross) when the capturing reagent concentration was 1 (b), 2.5 (c), 5 (d), and 10 (e) µg/ml. All standard curves were fitted with a 4-parameter model using SoftMaxPro software. The r squared (r2) is 0.998 for the 0% serum standard curve at 1 µg/ml capturing concentration and 1.000 for all the other curves.

a high background during the initial screening were loaded onto the column with a syringe. The column was then washed with five column volumes of a 20 mM sodium phosphate buffer and then eluted with 0.1 M glycine, pH 2.7. The eluate was dialyzed against PBS

overnight at 4 °C and the concentration of the purified cynomolgus monkey IgG was measured by absorbance at 280 nm using an estimated extinction coefficient of 1.36. The purified cynomolgus IgG was then coupled to controlled pore glass (CPG) beads using a standard

12

J. Yang et al. / Journal of Immunological Methods 335 (2008) 8–20

Table 1a Spike recovery of rhuMAb2H7 at three concentrations in individual cynomolgus monkey serum using the GRIP assay n

5 5 4

rhuMAb2H7 concentration

Mean %Coefficient recovery of variance (CV) (%)

Target (ng/ml)

Mean (ng/ml)

Range (ng/ml)

15.6 300 1000

14.64 259.52 871.20

13.6–16.4 250.3–265.4 785.5–945.2

94 87 87

7 2 8

Samples were prepared by spiking rhuMAb2H7 into neat serum then diluted 1:10 with assay diluent to achieve the targeted concentrations.

procedure. In brief, the beads were first washed in distilled water and the supernatant was removed by vacuum suction and discarded. The wet bead volume was estimated and an equal volume of freshly prepared 1% sodium metaperiodate was added to the tube and the suspension was rotated gently at room temperature for 30 min to activate the beads. After the beads settled, the supernatant was decanted and the beads were washed with PBS five times to remove excess periodate. The purified cynomolgus monkey IgG was then added to the activated beads and the suspension was mixed thoroughly before the beads were allowed to settle. To the mixture was added 2 mg of solid sodium cyanoborohydride and the mixture was mixed at 4 °C for 40 h. The cynomolgus monkey IgG coupled resin was then washed in PBS several times, and then blocked with 1 M ethanolamine, pH 8.0 overnight. PBS was then used to wash and store the resin. The sheep anti-human IgG (H + L) obtained from The Binding Site (San Diego, CA) was further adsorbed against cynomolgus monkey IgG-CPG columns prepared as described above. This further purified sheep anti-human IgG (H + L) was used as a coat to determine if it would reduce assay background and the variation of the background signal that was originally obtained from individual cynomolgus monkey serum samples.

3. Results 3.1. Generic Immunoglobulin Pharmacokinetic (GRIP) assay for measuring rhuMAb2H7 in cynomolgus monkey serum A quantitative bridging ELISA was developed to measure rhuMAb2H7 in cynomolgus monkey serum without the utilization of rhuMAb2H7-specific reagents. Extensive assay characterizations and qualifications suggested the assay was accurate, sensitive, reproducible and easy to perform. 3.1.1. Standard curve range and assay sensitivity A series of standard curves for rhuMAb2H7 with concentrations ranging from 1.56 to 400 ng/ml were generated using either assay buffer (PBS/0.5% BSA/ 0.05% P20/0.05% Proclin300/0.25% CHAPS/5 mM EDTA/0.35 M NaCl/pH = 8.9) alone or with assay buffers that contained 1%, 5% and 10% cynomolgus monkey serum (Fig. 1a). All standard curves showed similar dose-dependent responses, suggesting minimum matrix effect at up to 10% serum concentration. We chose to use a buffer standard curve without addition of pooled cynomolgus monkey serum as it was convenient to maintain the assay and it also eliminated the potential need to find a future matching serum pool. The minimal matrix effect observed in the standard curves with 1 µg/ ml capturing reagent could also be extrapolated to other coat concentrations (Fig. 1b–e). Standard curves of rhuMAb2H7 in either an assay buffer alone or in a buffer that contained 10% cynomolgus monkey serum were compared in assays that used coat concentrations at 1, 2.5, 5, and 10 µg/ml. At each coat concentration, the serum curve overlapped with the buffer curve. Although a higher response was observed with an increasing coat concentration, we used 1 µg/ml in the assay to conserve the reagent. Because the standard curve was constructed

Table 1b Recovery of spiked rhuMAb2H7 at different concentrations in 10% cynomolgus monkey serum to determine the lower and upper limits of quantification (LLOQ and ULOQ, respectively) in the GRIP assay n

LLOQ

ULOQ

20 20 20 20 20 20 20 20

rhuMAb2H7 concentration Target (ng/ml)

Mean (ng/ml)

1.56 2.00 3.12 4.00 91 94 97 100

1.33 1.90 2.82 3.91 91.2 91.2 97.1 104.8

Recovery (%) 85.3 95.0 90.4 97.8 100.2 97.0 100.1 104.8

Variance components (%CV) Inter-assay precision (%)

Intra-assay precision (%)

Overall precision (%)

3.8 4.5 5.5 5.1 3.8 3.9 4.3 4.0

4.2 1.7 2.5 4.0 3.4 3.0 3.6 5.0

5.6 4.8 6.1 6.5 5.1 4.9 5.6 6.4

J. Yang et al. / Journal of Immunological Methods 335 (2008) 8–20

13

serum blanks and assay buffer blanks were monitored throughout the assay development, and they appeared to be close to each other. For example, the O.D. of 10% serum background from 20 cynomolgus monkeys ranged from 0.009 to 0.014 with a mean value of 0.011, close to the assay blank background which was 0.009. These results again confirmed the observed minimum matrix effect. In order to determine the quantifiable assay range including both lower and upper limits of quantification (LLOQ and ULOQ, respectively), samples of various concentrations of rhuMAb2H7 in 10% cynomolgus monkey serum were prepared and aliquoted. These aliquots

in an assay buffer that did not contain serum, we carefully evaluated the accuracy of the assay using rhuMAb2H7-spiked serum samples. In this experiment, rhuMAb2H7 was first spiked into individual neat serum from 4–5 cynomolgus monkeys, targeting at low (15.6 ng/ml), mid (300 ng/ml) or high (1000 ng/ml) concentrations. All samples were then diluted to 1:10 in assay diluent and analyzed. Recovery yields of rhuMAb2H7 in cynomolgus monkey serum samples are summarized in Table 1a. At the concentrations tested, the spike recovery yield of rhuMAb2H7 had a mean value of 94%, 87% and 87%, with a %CV ranging from 2% to 8%, respectively. The background from 10%

Table 2 Linearity of recovery of rhuMAb2H7 targeted at two concentrations in individual cynomolgus monkey serum using the GRIP assay Target concentration (ng/ml)

Individual cynomolgus monkey

Dilution factor

Back calculated concentration (ng/ml)

% Recovery against target concentration

Range of % difference within dilutions

1000

Individual #1

1/10 1/20 1/40 1/80 1/10 1/20 1/40 1/80 1/10 1/20 1/40 1/80 1/10 1/20 1/40 1/80

874 845 866 897 786 803 869 946 880 836 921 961 945 891 975 1019

87 85 87 90 79 80 87 95 88 84 92 96 95 89 97 102

0–5

1/10 1/20 1/40 1/80 1/10 1/20 1/40 1/80 1/10 1/20 1/40 1/80 1/10 1/20 1/40 1/80 1/10 1/20 1/40 1/80

265 268 272 283 260 272 273 275 259 272 261 266 250 252 261 264 263 266 270 273

88 89 91 94 87 91 91 92 86 91 87 89 83 84 87 88 88 89 90 91

Individual #2

Individual #3

Individual #4

300

Individual #1

Individual #2

Individual #3

Individual #4

Individual #5

1–16

4–12

2–13

1–6

0–5

1–5

1–5

1–3

Samples were prepared by first spiking rhuMAb2H7 into neat serum to achieve the targeted concentrations followed by a 1:10 minimum dilution using assay diluent. A serially 1:2 fold diluted samples were then prepared and assayed.

14

J. Yang et al. / Journal of Immunological Methods 335 (2008) 8–20

were kept frozen at −70 °C until analyzed to mimic the storage conditions for study samples. Twenty aliquots of each sample were analyzed in duplicate over four days. Mean values for each sample were compared to targeted concentrations, and the results are summarized in Table 1b along with the variance components of the analyses. Recovery of rhuMAb2H7 at all conditions tested was within +/− 15% of the target concentration for each sample, and the overall precision was within 10%. These results suggested that the assay could readily quantify rhuMAb2H7 within the range of 1.56– 100 ng/ml. Because a minimum dilution of 1:10 of the serum spiked samples were used in the LLOQ experiment, the assay had a sensitivity of 15.6 ng/ml in neat serum.

IgGs in serum from various species. To explore broader applications of the GRIP assay, experiments utilizing serum from different species and various humanized IgGs were conducted and the results are summarized below. 3.2.1. Cross-reactivity of the GRIP assay reagent with various humanized IgGs and with serum of different species Standard curves in assay buffer of various humanized antibodies (IgG1) were generated using the GRIP assay and the results are shown in Fig. 2. All of the antibodies

3.1.2. Linearity of dilution Some samples from the PK studies might contain high concentrations of rhuMAb2H7, and need to be diluted beyond the minimum assay dilution of 1:10 before analysis. It is therefore important to evaluate if concentrations of rhuMAb2H7 can be accurately determined by the assay after dilution. In this experiment, rhuMAb2H7 was first spiked into individual cynomolgus monkey serum samples at targeted concentrations and then diluted to 1:10 with assay diluent. A series of 2fold dilutions were prepared from the initial dilutions for each sample. rhuMAb2H7 concentrations in the serial diluted serum samples were later determined in the assay and compared to each other (Table 2). Less than a 17% difference within serially diluted samples was observed (Table 2). These results suggested a minimum serum effect in the assay as the samples were diluted linearly within the tested range. 3.1.3. Assay variations To determine both the intra- and inter-assay precision, matrix control samples were prepared by diluting rhuMAb2H7 into neat cynomolgus monkey serum at concentrations of 30, 300, and 800 ng/ml. A total of twenty-two replicates of each set of controls were analyzed at a 1:10 dilution over three days. Variance components (%CV) were calculated for the control samples at each concentration, and were within 3–5% for both intraand inter-assays. 3.2. General applications of the GRIP assay Because the assay was independent of rhuMAb2H7specific reagents, it had the potential to become a Generic Immunoglobulin Pharmacokinetic (GRIP) assay and can be used to quantify other human/humanized

Fig. 2. Standard curves for rhuMAb2H7, Xolair® and Herceptin® (a) and Avastin® and Raptiva® (b) with MAbs being used as assay standards in the GRIP assay. All standard curves were fitted with a 4-parameter model using SoftMaxPro software. The r squared (r 2) is 0.999 for the Herceptin® standard curve and 1.000 for the others.

J. Yang et al. / Journal of Immunological Methods 335 (2008) 8–20

tested showed good dose-dependent responses as expected. It is worth mentioning that similar results were observed with other humanized IgG monoclonal therapeutics that contain frameworks other than IgG1, suggesting the assay can be used for quantifying different humanized IgGs (data not shown). To further evaluate the assay's suitability with serum from species other than cynomolgus monkey, a series of spike recovery experiments were carried out. Serum samples containing 100 ng/ml of rhuMAb2H7 were prepared from rodents and three other non-human primates, and the percent recovery of rhuMAb2H7 was determined. As summarized in Table 3, the recovery was within ± 20% of the target value for all species tested, suggesting the assay could accurately measure rhuMAb2H7 concentrations from serum samples from a number of species. 3.2.2. Comparison of a target-specific assay to GRIP assay In order to further assess the general application of the target-independent GRIP assay, we compared a target-specific assay and the GRIP assay side-by-side to quantify Herceptin® in both 1% and 10% cynomolgus monkey sera. The Herceptin®-specific direct ELISA used the extracellular domain of HER2 as the plate coat reagent, and goat anti-human IgG Fc HRP conjugate as the detecting antibody. The basic assay diluent (PBS/0.5% BSA, 0.05% Tween-20, and 0.05% Proclin-300) was used as blocking and washing buffer in the assay. Different concentrations of Herceptin® were spiked into cynomolgus monkey serum, and the samples were analyzed with both assays on the same day. Spike recovery of Herceptin® was calculated from the ratio of each experimentally determined Herceptin® concentration to its target value and duplicated results are summarized (Table 4). As shown in Table 4, both assays gave very comparable spike recoveries of Herceptin® in 1% and 10% cynomolgus monkey serum at all concentrations tested. Similar results were observed with another humanized antibody (data not shown).

Table 3 Recovery of 100 ng/ml rhuMAb2H7 spiked into 10% serum from different species as measured in the GRIP assay Species

Recovery (%)

Rat Mouse Baboon African green Rhesus monkey

97 99 81 85 93

15

Table 4 Duplicated spike recovery results of Herceptin® in cynomolgus monkey serum as determined by a Herceptin®-specific assay and the GRIP assay, where Herceptin® was used as the standard in both cases Serum Target concentration Herceptin® (%) concentration (ng/ml)

Spike recovery of Herceptin® with Herceptin®-specific assay (%)

Spike recovery of Herceptin® with GRIP assay (%)

10

91 89 90 93 102 106 99 106 106 104 114 104

88 85 84 90 95 94 105 107 105 100 108 103

75 25 5

1

75 25 5

3.3. Cynomolgus monkey serum adsorption to reduce serum background and inter-individual background variation The GRIP assay was developed using monkey serum adsorbed sheep anti-human IgG (H + L). To understand whether cynomolgus monkey serum adsorption is critical in reducing the serum background and the interindividual serum background variation, experiments were conducted using assay reagents with or without monkey serum adsorption (Table 5) in a model system to compare two assays. In the first experiment (Table 5a), antisera obtained from a goat immunized with rhuMAb2H7 were purified against a rhuMAb2H7 column and subsequently with or without a cynomolgus monkey IgG column. The cynomolgus monkey serum background was compared in two assays, where the first assay used reagents purified from the rhuMAb2H7 column only, and the second assay used reagents purified from both columns. The results showed that cynomolgus monkey serum adsorption significantly reduced serum background. Unfortunately the goat anti-rhuMAb2H7 only became available after the completion of the cynomolgus pilot study for rhuMAb2H7 and therefore was not used in the PK assay development. In a separate experiment, we kept the monkey serum adsorbed capture reagent the same in both assays but changed the detecting reagent (Table 5b). When the detecting reagent was not adsorbed against cynomolgus monkey serum, high serum background was observed. Due to the lack of non-adsorbed sheep anti-human IgG HRP conjugate, a related goat anti-human IgG HRP conjugate was chosen. Both experiments suggested that cynomolgus monkey

16

J. Yang et al. / Journal of Immunological Methods 335 (2008) 8–20

Table 5 Comparison of 10% cynomolgus monkey serum background using assay reagents with or without cynomolgus monkey serum adsorption (a) Assay #1

Assay #2

Purified by Purified by Capturing antibody: goat anti-rhuMAb2H7 rhuMAb2H7 column rhuMAb2H7 column only and cynomolgus monkey serum adsorption Secondary antibody: goat antirhuMAb2H7•biotin

Detection reagent: 10% cynomolgus serum background (O.D.450 nm–650 nm)

Purified by Purified by rhuMAb2H7 column rhuMAb2H7 column only and cynomolgus monkey serum adsorption

4.12

Streptavidin HRP conjugate 0.023

(b) Assay #3

Assay #4

Capturing reagent: cynomolgus monkey adsorbed sheep anti-human IgG Signaling reagent

Cynomolgus monkey serum adsorbed sheep anti-human IgG•HRP

Goat anti-human IgG•HRP

10% cynomolgus-A serum background (O.D.450 nm–650 nm)

0.036

0.237

10% cynomolgus-B serum background (O.D.450 nm–650 nm)

0.033

0.384

Difference between inter-individual serum background (O.D.450 nm–650 nm)

0.003

0.143

(a) Both capturing and secondary antibodies were adsorbed against cynomolgus monkey serum in Assay #2 but not in #1; (b) with the same capturing antibody, Assay #3 used monkey serum adsorbed secondary antibody while Assay #4 didn't.

serum adsorption can be used to reduce serum background in general. 3.4. Effects of BGG on serum background and rhuMAb2H7 signal 3.4.1. Inclusion of BGG/CHAPS/EDTA/NaCl in the blocking step lowers serum background variation from individual cynomolgus monkey In cases where it is not feasible to use a bridging ELISA with cynomolgus monkey serum adsorbed sheep

anti-human IgG as assay reagents, we evaluated a direct ELISA format with only the capturing reagent adsorbed against cynomolgus monkey serum. Commercially available goat anti-human IgG (H + L) HRP conjugate was used as the detection reagent in the direct ELISA. It was observed that the addition of BGG/CHAPS/EDTA/ NaCl in the blocking step significantly lowered the variation of the background from individual cynomolgus monkey sera. Although the plate coat reagent (sheep anti-human IgG) had already been adsorbed against cynomolgus monkey serum, a high and variable serum background was noticed in the initial testing when a basic assay buffer A (PBS buffer that contained 0.5% BSA, 0.05% Tween-20 and 0.05% Proclin-300) was used (Run 2 in Table 6). Additional purification of the coat reagent to further remove cynomolgus monkey IgG cross-reactivity did not decrease the mean serum background or the inter-individual background variation, suggesting serum proteins other than IgGs contributed to the matrix effect in the assay (Table 6). In addition, the inter-individual background variation remained high despite of multiple pre-adsorption processes (Table 6). Diluting the serum to 1% gave an even higher variation. Initial efforts with different block solutions failed to control the assay background variation. These solutions included: commercially available Superblock and Casein; solutions with higher concentrations of BSA or an addition of gelatin (Harlow and Lane, 1988; Pruslin et al., 1991); various detergents (data not shown). Further studies with higher concentrations of salt, metalchelating reagents, non-specific binding reagents, nondenaturing zwitterionic detergents, and different buffer systems were conducted and the results are summarized in Table 7. While buffer B didn't result in any improvement in lowering background variation, buffers C and D showed significantly decreased inter-individual background variations and an increase in background in general. The background between the assay blank and the mean serum samples became close to each other when buffer C or D was used, which was important for an assay that used a buffer-based standard curve to quantify serum samples. While the precise mechanism of an increased background and a decreased variation is not clear, one potential explanation is that weak interactions between the plate coat material and the additional additives in buffers C and D enhanced the assay background to a maximum value, resulting in a masked or reduced inter-individual difference. A slightly high pH in buffer D as compared to buffer C gave an even greater reduction in the inter-individual background variation from cynomolgus monkey serum.

J. Yang et al. / Journal of Immunological Methods 335 (2008) 8–20

17

Table 6 Effects on cynomolgus monkey serum background through additional pre-adsorption of capturing reagents Run Capture reagent pre-treatment

Serum concentration (%)

Number of individuals

Mean background CV% of (O.D. 450–650 nm) background

1 2 3

1 10 10

10 8 8

0.181 0.308 0.314

85 41 42

10

8

0.299

46

10

8

0.261

47

4 5

Monkey serum adsorption Monkey serum adsorption Monkey serum adsorption plus pooled cynomolgus monkey IgG fractions Monkey serum adsorption plus one problematic individual and pooled fractions Monkey serum adsorption plus two problematic individual and pooled fractions

Cynomolgus monkey IgG columns were generated with commercially available pooled monkey IgGs or IgGs purified from individual cynomolgus monkey sera that showed high assay background in the initial screening (Run 2). The sheep anti-human IgG (H + L) coat reagent that has been monkey serum adsorbed was further adsorbed with either the pooled monkey IgG column or in combination with individual cynomolgus monkey IgG columns. The experiments were carried out using an direct ELISA where the goat anti-human IgG HRP conjugate was used as detection reagent and the basic assay diluent (PBS/0.5% BSA, 0.05% Tween-20, and 0.05% Proclin-300) was used in blocking, washing and dilution steps.

3.4.2. Replacing BGG with other mammalian IgGs or fish gelatin Further experiments were conducted to determine if BGG could be replaced in the assay methods by fish Table 7 Effects of assay buffer components on cynomolgus monkey serum background Buffer Buffer additives

(O.D.

450 nm–650 nm)

240 ng/ml Assay 10% serum rhuMAb2H7 Blank background (mean of eight individuals)

CV %a

Ab

none

2.041

0.012 0.083

47

Bb

5 mM EDTA 0.35 M NaC1

1.911

0.007 0.058

62

Cb

5 mM EDTA 1.878 0.35 M NaC1 0.25% CHAPS 0.2% BGG, pH 8.0

0.113 0.142

22

5 mM EDTA 1.724 0.35 M NaC1 0.25% CHAPS 0.2% BGG, pH 8.9

0.142 0.165

none

0.009 0.202

Db

Ec

2.098

gelatin or other mammalian immunoglobulins, including mouse IgG, rabbit IgG and donkey IgG. The assay was carried out similarly to the procedure described early. Sheep anti-human IgG (1 µg/ml, monkey serum preadsorbed) was used as the capturing reagent. The HRP conjugate of the capturing reagent was used as the detecting antibody. Fish gelatin or another mammalian immunoglobulin was used in the blocking buffer, sample buffers and detection agent buffers, in place of BGG. The results were summarized in Table 7. The rhuMAb2H7 signal produced when these agents were used was substantially less than that obtained with the use of BGG in the assay buffer. The data also revealed that the background variation as measured by the %CV increased significantly when BGG was absent from the assay buffer (Table 8). Table 8 Replacing BGG with other mammalian IgGs or fish gelatin in blocking and assay diluents Blocking and assay diluent

D

D1+ 0.2% rabbit IgG

D1+ D1+ 0.2% 0.2% fish gelatin donkey IgG

240 ng/ml rhuMAb2H7 (O.D. 450 nm–650 nm) Mean O.D. of 10% cynomolgus monkey serum background (20 individuals) CV% of serum background S/N

1.15 0.09

0.06

0.38

0.34

.015 0.033

0.035

0.026

0.045

7

13

30

23

35

77

3

2

15

8

18

48

The experiments were carried out with rhuMAb2H7, assay blank and eight individual cynomolgus monkey sera using an indirect ELISAwhere the goat anti-human IgG HRP conjugate was used as detection reagent. a CV% of the cynomolgus monkey serum background. b Basic buffer components were PBS/0.5% BSA, 0.05% Tween-20, and 0.05% Proclin-300. c Basic buffer components were 55 mM HEPES/0.5% BSA, 25 mM HEPES sodium salt, 2% Triton X-100, and 0.05% Proclin-300.

D1+ 0.2% mouse IgG

Buffer D is composed of PBS/0.5% BSA/0.05% P20/0.05% Proclin300/ 0.25% CHAPS/0.2% BGG/5 mM EDTA/0.35 M NaCl; Buffer D1 is the same as Buffer D but without BGG.

18

J. Yang et al. / Journal of Immunological Methods 335 (2008) 8–20

4. Discussion The accuracy of an assay can often be compromised by interference, which has several slightly different definitions (Anonymous, 1976; Selby, 1999). We here adopt the definition by Selby as it better reflects our experience in the development of the GRIP assay. It states that interference is “the effect of a substance present in an analytical system which causes a deviation of the measured value from the true value”. The direction and magnitude of the effect of an interfering substance depends on the source of the interference, the concentration of each interfering substance and the overall assay condition. Techniques to eliminate or minimize interference in an assay vary greatly based on the nature of the interference, and some commonly used methods have been discussed (Levinson and Miller, 2002). One obvious approach that can be effectively used in general is to use analyte-specific reagents in the assay. To quantify rhuMAb2H7 in cynomolgus monkey serum, we first sought to use rhuMAb2H7-specific molecules, including CD20, the antigen that rhuMAb2H7 recognizes. However, available CD20 molecules were not suitable for developing a highly sensitive PK assay in cynomolgus monkey serum as discussed earlier. Other rhuMAb2H7specific reagents such as monoclonal and polyclonal antibodies against rhuMAb2H7 were not available at the time of this work. Because there are several murine residues preserved in rhuMAb2H7 during the humanization process, we tried to use antibodies against murine IgGs as capturing reagents but found that the rhuMAb2H7 signal was too low for these reagents to be useful in developing a sensitive PK assay (data not shown). Sheep anti-human IgGs, similar to other polyclonal anti-human IgG antibodies, tend to cross-react with cynomolgus monkey IgGs. Specificity of the polyclonal antibody to rhuMAb2H7 was improved using affinity purification to remove cross-reactivity. When both capturing and detecting antibodies were adsorbed against a cynomolgus monkey serum protein column, cynomolgus monkey serum background was similar to that of buffer blank (Fig. 1 and Table 5). When only the capturing antibody was adsorbed against a cynomolgus monkey serum column (i.e., the detecting antibody was goat antihuman IgG•HRP), it was noted that serum background variation could be suppressed through additives such as BGG in the assay buffers (Table 7). This might result from the weak interaction between BGG and the capturing antibody (sheep anti-human IgG) that masked out the inter-individual serum background variation. Even though all are mammalian IgGs (bovine IgG1 shares 62% and 66% sequence identities with mouse IgG1

and rabbit IgG, respectively), we found that BGG could not be replaced by other non-specific immunoglobulins, including rabbit, donkey and mouse IgGs. It is therefore important to include BGG in assays where not all reagents are absorbed against cynomolgus monkey serum. Other commonly used techniques to minimize interference include: using assay buffers that are supplemented with salt (Hashida et al., 1983), EDTA (Kapyaho et al., 1989; Selby, 1999) and CHAPS (Dubois et al., 1987; Krummen et al., 1993), pre-treatment of samples to remove potential interfering substances, and addition of various concentrations of serum in assay diluents and standard curves. We try to avoid sample pre-treatment in a PK assay for two reasons. First, because of the small volume of each sample, potentially errors could be introduced during additional pre-treatment steps. In addition, introducing sample pre-treatment could be laborious and time consuming. Although including serum in assay diluents and standard curves may be more representative of the study samples, it requires the maintenance of serum pools for long-term usage. We therefore focused on optimizing the assay reagents and buffer conditions to minimize the interference. We found that a combination of reagent preadsorption to remove cross-reactivity and buffer optimization could effectively minimize the overall serum background and inter-individual background variation. In addition, the mean background from cynomolgus monkey serum became very close to that of the assay blank. Therefore, using a standard curve in assay diluent alone without the addition of cynomolgus monkey serum enabled us to achieve high accuracy with rhuMAb2H7spiked samples. The conditions we found greatly simplified the operational procedure and allowed for standards to be easily prepared without serum supplement. This assay can potentially be used for similar assay development needs. The GRIP assay developed in our lab is sensitive, accurate, and reproducible. The assay also has a high throughput and can be maintained easily. We successfully used the assay to support rhuMAb2H7 cynomolgus monkey studies. The assay has also been applied to quantify other humanized IgGs in serum from species besides cynomolgus monkey. Because the GRIP assay reagents are human IgGspecific, it has the potential to recognize various forms of humanized or human monoclonal therapeutics even when the therapeutic may be in association with its target. Therefore, the GRIP assay offers a potential advantage to measure the total therapeutic level in serum and tests are underway to confirm this. Additionally, the conjugate used in the GRIP assay could also be combined with a target-specific reagent for developing alternative PK assays.

J. Yang et al. / Journal of Immunological Methods 335 (2008) 8–20

In summary, we described in this paper a pharmacokinetic assay that can quantify humanized IgGs in cynomolgus monkey serum without analyte-specific reagents. This assay could have broader applications to support human/humanized antibody biotherapeutics development, and offers a potentially useful tool to evaluate the total biotherapeutic level in vivo. Acknowledgements We are grateful to Cliff Quan and Kurt Deshayes for generating CD20 peptides for initial assay development testing. Thanks to Randy Dere and John Lowe for their technical support and to Eric Wakshull and Jane Ruppel for helpful discussion. References Anolik, J., Sanz, I., Looney, R.J., 2003. B cell depletion therapy in systemic lupus erythematosus. Curr. Rheumatol. Rep. 5, 350. Anonymous, 1976. IFCC provisional recommendation on quality control in clinical chemistry. J. Clin. Chem. Biochem. 14, 270. Biagini, R.E., Moorman, W.J., Lal, J.B., Gallagher, J.S., Bernstein, I.L., 1988. Normal serum IgE and IgG antibody levels in adult male cynomolgus monkeys. Lab. Anim. Sci. 38, 194. Boon, L., Laman, J.D., Ortiz-Buijsse, A., den Hartog, M.T., Hoffenberg, S., Liu, P., Shiau, F., de Boer, M., 2002. Preclinical assessment of anti-CD40 Mab 5D12 in cynomolgus monkeys. Toxicology 174, 53. Clark, E.A., Ledbetter, J.A., 1989. Structure, function, and genetics of human B cell-associated surface molecules. Adv. Cancer Res. 52, 81. Clark, E.A., Shu, G., Ledbetter, J.A., 1985. Role of the Bp35 cell surface polypeptide in human activation. Proc. Natl. Acad. Sci. U. S. A. 82, 1766. Corcione, A., Aloisi, F., Serafini, B., Capello, E., Mancardi, G.L., Pistoia, V., Uccelli, A., 2005. B-cell differentiation in the CNS of patients with multiple sclerosis. Autoimmun. Rev. 4, 549. D'Ovidio, M.C., Pace, M., Via, F., Strom, R., Turillazzi, P.G., 1994. Characterization of kappa and lambda light immunoglobulin chains in adult macaca fascicularis. Folia Primatol. 63, 221. Da Silva, M., Petruzzo, P., Virieux, S., Tiollier, J., Badet, L., Martin, X., 2001. A primate model of renal ischemia–reperfusion injury for preclinical evaluation of the antileukocyte function associated antigen 1 monoclonal antibody odulimonab. J. Urol. 166, 1915. DeSilva, B., Smith, W., Weiner, R., Kelley, M., Smolec, J., Lee, B., Khan, M., Tacey, R., Hill, H., Celniker, A., 2003. Recommendations for the bioanalytical method validation of ligand-binding assays to support pharmacokinetic assessments of macromolecules. Pharm. Res. 20, 1885. Dodds, M.G., Visich, J.E., Vicini, P., 2005. Population pharmacokinetics of recombinant factor XIII in cynomolgus monkeys. AAPS J. 7, E693. Dorner, T., Burmester, G.R., 2003. The role of B cells in rheumatoid arthritis: mechanisms and therapeutic targets. Curr. Opin. Rheumatol. 15, 246. Dubois, D.Y., Cantraine, F., Malmendier, C.L., 1987. Comparison of different sandwich enzyme immunoassays for the quantitation of human apolipoproteins A-I and A-II. J. Immunol. Methods 96, 115.

19

Gopal, A.K., Press, O.W., 1999. Clinical applications of anti-CD20 antibodies. J. Lab. Clin. Med. 134, 445. Harlow, E., Lane, D., 1988. Antibodies, a Laboratory Manual. Cold Spring Harbor Press, NY. Hart, T.K., Cook, R.M., Zia-Amirhosseini, P., Minthorn, E., Sellers, T.S., Maleeff, B.E., Eustis, S., Schwartz, L.W., Tsui, P., Appelbaum, E.R., Martin, E.C., Bugelski, P.J., Herzyk, D.J., 2001. Preclinical efficacy and safety of mepolizumab (SB-240563), a humanized monoclonal antibody to IL-5, in cynomolgus monkeys. J. Allergy Clin. Immunol. 108, 250. Hashida, S., Nakagawa, K., Imagawa, M., Inoue, S., Yoshitake, S., Ishikawa, E., Endo, Y., Ohtaki, S., Ichioka, Y., Nakajima, K., 1983. Use of inorganic salts to minimize serum interference in a sandwich enzyme immunoassay for human growth hormone using Fab'horseradish peroxidase conjugate. Clin. Chim. Acta 135, 263. Hong, K., Presta, L.G., Lu, Y., Penn, A., Adams, C., Chuntharapai, A., Yang, J., Wong, W.L., Meng, Y.G., 2004. Simple quantitative live cell and anti-idiotypic antibody based ELISA for humanized antibody directed to cell surface protein CD20. J. Immunol. Methods 294, 189. Isenberg, D.A., 2006. B cell targeted therapies in autoimmune diseases. J. Rheumatol., Suppl. 77, 24. Jensen, M., Winkler, U., Manzke, O., Diehl, V., Engert, A., 1998. Rapid tumor lysis in a patient with B-cell chronic lymphocytic leukemia and lymphocytosis treated with an anti-CD20 monoclonal antibody (IDEC-C2B8, rituximab). Ann. Hematol. 77, 89. Kapyaho, K., Tanner, P., Weber, T., 1989. Effect of complement binding on a solid-phase immunometric TSH assay. Scand. J. Clin. Lab. Invest. 49, 211. Kelley, S.K., Harris, L.A., Xie, D., Deforge, L., Totpal, K., Bussiere, J., Fox, J.A., 2001. Preclinical studies to predict the disposition of Apo2L/tumor necrosis factor-related apoptosis-inducing ligand in humans: characterization of in vivo efficacy, pharmacokinetics, and safety. J. Pharmacol. Exp. Ther. 299, 31. Keystone, E., 2005. B cell targeted therapies. Arthritis Res. Ther. 7 (Suppl 3), S13. Kosmas, C., Stamatopoulos, K., Stavroyianni, N., Tsavaris, N., Papadaki, T., 2002. Anti-CD20-based therapy of B cell lymphoma: state of the art. Leukemia 16, 2004. Krummen, L.A., Woodruff, T.K., DeGuzman, G., Cox, E.T., Baly, D.L., Mann, E., Garg, S., Wong, W.L., Cossum, P., Mather, J.P., 1993. Identification and characterization of binding proteins for inhibin and activin in human serum and follicular fluids. Endocrinology 132, 431. Levinson, S.S., Miller, J.J., 2002. Towards a better understanding of heterophile (and the like) antibody interference with modern immunoassays. Clin. Chim. Acta 325, 1. Lewis, A.P., Barber, K.A., Cooper, H.J., Sims, M.J., Worden, J., Crowe, J.S., 1993. Cloning and sequence analysis of kappa and gamma cynomolgus monkey immunoglobulin cDNAs. Dev. Comp. Immunol. 17, 549. Looney, R.J., 2002. Treating human autoimmune disease by depleting B cells. Ann. Rheum. Dis. 61, 863. McLaughlin, P., White, C.A., Grillo-Lopez, A.J., Maloney, D.G., 1998. Clinical status and optimal use of rituximab for B-cell lymphomas. Oncology (Williston Park) 12, 1763 discussion 1769–70, 1775–7. Moore, J., Ma, D., Will, R., Cannell, P., Handel, M., Milliken, S., 2004. A phase II study of Rituximab in rheumatoid arthritis patients with recurrent disease following haematopoietic stem cell transplantation. Bone Marrow Transplant. 34, 241. Oligino, T.J., Dalrymple, S.A., 2003. Targeting B cells for the treatment of rheumatoid arthritis. Arthritis Res. Ther. 5 (Suppl 4), S7.

20

J. Yang et al. / Journal of Immunological Methods 335 (2008) 8–20

Pace, M., D'Ovidio, M.C., Iale, E., Strom, R., Turillazzi, P., 1996. Serum kappa and lambda light immunoglobulin chains in cynomolgus macaques (Macaca fascicularis) during the first twenty months of age. Immunol. Lett. 50, 139. Panayi, G.S., 2005. B cell-directed therapy in rheumatoid arthritis— clinical experience. J. Rheumatol., Suppl. 73, 19 discussion 29–30. Ponce, R., Armstrong, K., Andrews, K., Hensler, J., Waggie, K., Heffernan, J., Reynolds, T., Rogge, M., 2005. Safety of recombinant human factor XIII in a cynomolgus monkey model of extracorporeal blood circulation. Toxicol. Pathol. 33, 702. Pruslin, F.H., To, S.E., Winston, R., Rodman, T.C., 1991. Caveats and suggestions for the ELISA. J. Immunol. Methods 137, 27. Riley, J.K., Sliwkowski, M.X., 2000. CD20: a gene in search of a function. Semin. Oncol. 27, 17. Selby, C., 1999. Interference in immunoassay. Ann. Clin. Biochem. 36 (Pt 6), 704. Shaw, T., Quan, J., Totoritis, M.C., 2003. B cell therapy for rheumatoid arthritis: the rituximab (anti-CD20) experience. Ann. Rheum. Dis. 62 (Suppl 2), ii55. Stashenko, P., Nadler, L.M., Hardy, R., Schlossman, S.F., 1980. Characterization of a human B lymphocyte-specific antigen. J. Immunol. 125, 1678.

Tedder, T.F., Engel, P., 1994. CD20: a regulator of cell-cycle progression of B lymphocytes. Immunol. Today 15, 450. Tedder, T.F., Schlossman, S.F., 1988. Phosphorylation of the B1 (CD20) molecule by normal and malignant human B lymphocytes. J. Biol. Chem. 263, 10009. Tryphonas, H., Karpinski, K., O'Grady, L., Hayward, S., 1991. Quantitation of serum immunoglobulins G, M, and A in the rhesus monkey (M. mulatta) using human monospecific antisera in the enzyme-linked immunosorbent assay: developmental aspects. J. Med. Primatol. 20, 58. Tuscano, J.M., Martin, S., Song, K., Wun, T., 2005. B cell reductive therapy in the treatment of autoimmune diseases: a focus on monoclonal antibody treatment of rheumatoid arthritis. Hematology 10, 521. von Schilling, C., 2003. Immunotherapy with anti-CD20 compounds. Semin. Cancer Biol. 13, 211. Vugmeyster, Y., Beyer, J., Howell, K., Combs, D., Fielder, P., Yang, J., Qureshi, F., Sandlund, B., Kawaguchi, L., Dummer, W., Lowman, H., McKeever, K., 2005. Depletion of B cells by a humanized anti-CD20 antibody PRO70769 in Macaca fascicularis. J. Immunother. 28, 212.