Blood Bank questions

N E W M E T H O D S A N D A P P R O A C H E S

Resolving the daratumumab interference with blood

compatibility testing

Claudia I. Chapuy,1 Rachel T. Nicholson,1 Maria D. Aguad,1 Bjoern Chapuy,2 Jacob P. Laubach,2

Paul G. Richardson,2 Parul Doshi,3 and Richard M. Kaufman1

BACKGROUND: Daratumumab (DARA), a promising

novel therapy for multiple myeloma, is an IgG1j

monoclonal antibody that recognizes CD38 on myeloma

cells. During routine compatibility testing, we observed

that the plasma of five of five DARA-treated patients

demonstrated a positive antibody screen and

panreactivity on red blood cell (RBC) panel testing. We

hypothesized that the observed panreactivity reflected

DARA binding to CD38 on reagent RBCs, and we

investigated methods to prevent this binding.

STUDY DESIGN AND METHODS: DARA binding to

CD381 or CD38– HL60 cells was assessed by flow

cytometry. To remove cell surface CD38, cells were

incubated with dithiothreitol (DTT) or trypsin. Soluble

CD38 or anti-DARA was used to neutralize DARA in

solution. Routine blood bank serologic methods were

used to test samples from DARA-treated patients and

normal plasma samples spiked with DARA and/or

alloantibodies.

RESULTS: Normal plasma samples spiked with DARA

(0.1-10 mg/mL) and incubated with reagent RBCs recapitulated the interference observed with samples

from DARA-treated patients. Flow cytometry experiments

confirmed DARA binding to CD381 HL60 cells, but not to

CD38– controls. DTT treatment of CD381 HL60 cells

reduced DARA binding by 92% by denaturing cell surface

CD38. Treating DARA-containing plasma with soluble

CD38 or anti-DARA idiotype also inhibited DARA binding.

CONCLUSION: DARA causes panreactivity in vitro by

binding to CD38 on reagent RBCs. Treating reagent

RBCs with DTT is a robust method to negate the DARA

interference, enabling the safe provision of blood to

DARA-treated patients. Because DTT denatures Kell

antigens, K– units are provided to these patients.

D aratumumab (DARA) is a promising novel

therapy for multiple myeloma (MM). DARA is

an IgG1j human monoclonal antibody (MoAb) that specifically targets human CD38,

which is highly expressed on myeloma cells. In preclinical

studies, DARA was highly cytotoxic to tumor cells via mul-

tiple mechanisms, including complement-dependent

cytotoxicity, antibody-dependent cellular cytotoxicity, and

apoptosis.1 In the first-in-human Phase I and II clinical

trial, DARA showed significant anti-MM activity as mono-

therapy in heavily treated patients with relapsed or refrac-

tory disease.2,3 Phase III trials of DARA are currently

getting under way in the United States and across several

countries internationally.

ABBREVIATIONS: DARA 5 daratumumab; GFP 5 green

fluorescent protein; MM 5 multiple myeloma.

From the 1Blood Bank, Department of Pathology, Brigham and

Women’s Hospital, and the 2 Department of Medical Oncology,

Dana-Farber Cancer Institute, Boston, Massachusetts; and the 3Janssen R&D, Spring House, Pennsylvania

Daratumumab and anti-DARA were provided by Janssen, Inc.

Address correspondence to: Richard M. Kaufman, MD, Blood

Bank, Amory 260, 75 Francis Street, Brigham and Women’s

Hospital, Boston, MA 02115; e-mail: [email protected]

The copyright line for this article was changed on 13 April

2015 after original online publication.

This is an open access article under the terms of the Crea-

tive Commons Attribution-NonCommercial License, which per-

mits use, distribution and reproduction in any medium,

provided the original work is properly cited and is not used for

commercial purposes.

Received for publication September 12, 2014; revision

received January 27, 2015; and accepted January 29, 2015.

doi:10.1111/trf.13069

VC 2015 The Authors Transfusion published by Wiley Peri-

odicals, Inc. on behalf of AABB

TRANSFUSION 2015;55;1545–1554

Volume 55, June 2015 TRANSFUSION 1545

On routine screening in the blood bank, we observed

that five of five patients who had received DARA in a

Phase I and II clinical trial had positive antibody screens.

The plasma of these patients was panreactive in routine

serologic tests, preventing the blood bank from providing

cross-match–compatible red blood cell (RBC) units.

Adsorptions using ZZAP-treated or untreated RBCs failed

to remove the interference.

The expression of CD38 on human RBCs has been

demonstrated previously. 4-7

For example, Albeniz and

coworkers7 performed Western blot analyses of CD38 on

human RBCs of cancer patients and healthy controls.

While they found increased CD38 expression on RBC

membranes of cancer patients, a signal confirming weak

expression on normal RBCs was also detected.7 We

hypothesized that the panreactivity observed in the blood

bank was caused by direct binding of DARA to endoge-

nous CD38 on reagent RBCs. Previous investigators

reported that CD38 is sensitive to denaturation by the

reducing agent dithiothreitol (DTT) and that enzymatic

digestion with trypsin can cleave CD38 from the cell sur-

face. 8,9

We explored methods to negate the DARA inter-

ference by removing RBC surface CD38 or by neutralizing

DARA in solution.

MATERIALS AND METHODS

Patient samples

The Brigham and Women’s Hospital blood bank received

whole blood (EDTA) samples collected as part of routine

clinical care from adult patients with refractory MM

receiving DARA. All of these patients were enrolled in a

Phase I and II clinical trial conducted at Dana-Farber

Cancer Institute (Protocol NCT00574288, Dana-Farber

Cancer Institute Protocol 10-429). Baseline (DARA-free)

samples were tested in all subjects (n 5 11). Subse-

quently, DARA-containing samples were tested in the

subset of enrolled subjects who required blood transfu-

sion (n 5 5). The samples were tested in the blood bank

using routine serologic methods, including solid-phase

(TANGO optimo automated blood bank system, Bio-

Rad, Hercules, CA) and tube testing (polyethylene glycol

[PEG], low-ionic-strength saline [LISS], or no enhance-

ment).10 Serologic tests performed on patient samples

were as follows: ABO/Rh type, antibody screen, RBC

panel, direct antiglobulin test (DAT), and antihuman

globulin (AHG) cross-match. All DATs were done with

monospecific testing for IgG and C3. Eluates were pre-

pared from patient samples with a positive DAT using

acid elution. Patient samples that were panreactive in

routine tests were further tested using DTT-treated rea-

gent RBCs (below). Institutional review board approval

was obtained to study methods to prevent DARA binding

on these patient samples.

Serologic testing of DARA-spiked plasma samples

or DARA-treated patient samples

Normal (DARA-free) plasma samples with or without RBC

alloantibodies were spiked with increasing concentrations

of DARA (0.1-10 mg/mL; provided by Janssen R&D, Spring House, PA). These samples were analyzed using routine

blood bank serologic testing with solid phase (TANGO

optimo, Bio-Rad) and tube testing using PEG, LISS, or no

enhancement.10 Agglutination was graded per routine (0

[no agglutination], M1 [macroscopically positive (weak)],

11, 21, 31, 41). Microscopic examinations were per-

formed only on DATs. For all serologic studies, 3% to 5%

cell suspensions of reagent RBCs (Bio-Rad, Biotestcell 1,

Ref. 816014100, Batch ID 959) in phosphate-buffered

saline (PBS), pH 7.3, were used. RBC panels used for anti-

body identification were from different manufacturers:

Bio-Rad, Immucor (Norcross, GA), Ortho Clinical Diag-

nostics (Raritan, NJ), and Medion Diagnostics (Miami,

FL). Cells were chosen depending on their specific antigen

expression profile. Panels included five to 12 cell lines.

RBC alloantibody identification required obtaining a posi-

tive agglutination reaction on three cells with the antigen

present (rule-in) and a negative agglutination reaction on

three cells with the antigen absent (rule-out). To demon-

strate removal of CD38 from reagent RBCs, RBCs were left

untreated, treated with 0.2 mol/L DTT (Sigma, St Louis,

MO; detailed protocol below), or treated with 1% trypsin

(incubated at 37�C for 30 min)11 and tested against DARA-

spiked plasma samples or plasma from DARA-treated

patients. For neutralization studies, DARA-spiked plasma

samples were incubated at room temperature for 15

minutes with recombinant human soluble CD38 (R&D

Systems [Minneapolis, MN], Cat. No. 2404-AC; final con-

centration, 0.05-5 mg/mL) or mouse anti-DARA idiotype (Janssen; final concentration, 5 mg/mL; reported by Oos- tendorp et al., submitted for publication). A mouse anti-

human antibody (mouse IgG1j, MOPC-21, Sigma-Aldrich, St Louis, MO) was used as an isotype control for the anti-

idiotype. Eluates were prepared from DARA-treated RBCs

using acid elution.10

DTT treatment of reagent RBCs

A detailed method for DTT treatment of RBCs is

described in the AABB Technical Manual.10 Briefly,

0.2 mol/L DTT was prepared by diluting 1 g of DTT in

32 mL of PBS, pH 8.0. K1, E1 control RBCs were used to

verify that DTT treatment had denatured the K antigen

while preserving the E antigen. Reagent and control RBCs

(100 mL of a 3%-5% suspension) were washed four times with PBS, pH 7.3, before adding 400 mL of 0.2 mol/L DTT to each tube. The RBCs were incubated at 37�C for 30

minutes with periodic mixing by inversion (three to four

times during incubation). The RBCs were washed four

times with PBS, pH 7.3, and used for subsequent testing.

CHAPUY ET AL.

1546 TRANSFUSION Volume 55, June 2015

Cell culture

Human HL60 cells were propagated in RPMI supple-

mented with 10 mmol/L HEPES buffer, 2 mmol/L L-gluta-

mine, 50 U/mL penicillin, 50 U/mL streptomycin, and

10% heat-inactivated fetal bovine serum (FBS; all from

Life Technologies, Grand Island, NY).

Generation of CD381 HL60 cells and CD38–, green

fluorescent protein–positive control HL60 cells

HL60 cells were transduced with human CD38 or green

fluorescent protein (GFP), the latter serving as a negative

control for DARA binding. The cDNA of human CD38 was

obtained in pDONR221 (HsCD00045212) from the DF/

HCC DNA Resource Core (http://plasmid.med.harvard.

edu/PLASMID/Home.jsp). Sanger sequencing was used to

confirm that the plasmid contained the full-length open

reading frame of human CD38. The human CD38

sequence was cloned into pMSCV-puro using a standard

Gateway LR reaction according to the manufacturer’s

directions (Life Technologies). Generation of the control

vector pMSCV-puro-GFP, packaging of retroviral particles,

and infection were performed as previously described. 12

Two days after transduction, HL60 cells were selected for

Fig. 1. Generation of a one-tube HL60 cell model system to study DARA binding. (A) Flow cytometric assessment of CD38 expres-

sion on patient RBCs was performed after gating on GlyA1 RBCs using an anti-CD38 FITC–conjugated antibody (black) or an

isotype control (gray). (B) Flow cytometric assessment of CD38 surface expression in stably transduced HL60-CD38 cells (black)

compared to nontransduced HL60 cells (gray). Human CD38 was detected using mouse anti-CD38 directly conjugated to allo-

phycocyanin. (C) Flow cytometric assessment of GFP expression in stably transduced HL60-GFP cells (black) compared to non-

transduced HL60 cells (gray). (D) For DARA-binding studies, HL60-CD38 cells and control HL60-GFP cells (2.5 3 105 cells each)

were incubated with increasing concentrations of DARA. DARA binding was assessed by flow cytometry using an anti-human

antibody labeled with PE. Separate gating on GFP1 cells (HL60-GFP) and GFP– cells (HL60-CD38) allowed the assessment of

DARA binding to each cell type in one tube. (E) Quantification of dose-dependent binding of DARA to HL60-CD38 cells, meas-

ured by flow cytometry (black; y-axis reflects the geometric mean) and compared to isotype control (gray). Plasma from a

DARA-treated patient (Patient 3, Table 1) also showed significant binding to HL60-CD38 cells. (F) Flow cytometric assessment of

DARA binding to HL60-CD38 (black) compared to HL60-GFP cells (gray). All data show a representative example of at least three

independent experiments. Error bars indicate SD.

DARATUMUMAB BLOOD BANK INTERFERENCE

Volume 55, June 2015 TRANSFUSION 1547

72 hours with 1 mg/mL puromycin (Sigma-Aldrich). Trans- duction efficiency of HL60-GFP and HL60-CD38 cells was

assessed by flow cytometry (Fig. 1).

Detection of CD38 on RBCs and on transduced

HL60 cells by flow cytometry

Reagent RBCs (Ortho Clinical Diagnostics) were incubated

for 30 minutes with anti-CD235A-phycoerythrin (PE; anti-

GlyA, eBioscience [San Diego, CA], 12-9987-82, 1:1000)

and either a monoclonal mouse anti-CD38-fluorescein

isothiocyanate (FITC; BD Biosciences [Sparks, MD], Clone

HIT2) or a monoclonal mouse anti-human FITC-labeled

IgG1j isotype control (BD, Clone 15H6). Enforced expres- sion of human CD38 on HL60 cells after transduction was

detected using a monoclonal mouse anti-CD38 directly

conjugated to allophycocyanin (BD, 560980). This was

done as follows: 5 3 10 5

cells were washed twice with

PBS, resuspended in 10 mL of antibody, and 90 mL of PBS and incubated for 30 minutes at room temperature in the

dark. Per sample, 20,000 events were recorded. All flow

cytometry data were acquired on a cell analyzer (BD LSR

Fortessa, BD Biosciences). Data analysis and graphics

generation were performed with computer software

(FlowJo, V10.0.6 for MacOS, Tree Star, Ashland, OR).

Detection of DARA binding to transduced HL60

cells by flow cytometry

For DARA-binding studies, a mixture of HL60-CD38 and

HL60-GFP cells (2.5 3 105 each) was added to the same

tube and incubated with increasing concentrations of

DARA (0.1-1 mg/mL) for 1 hour at room temperature. An IgG isotype antibody (SouthernBiotech [Birmingham, AL];

human IgG1 kappa-UNLB, Cat. No. 0151K-01) in equal

concentrations to DARA was used as a negative control.

Detection of DARA or control antibody binding was

assayed using a PE-labeled goat anti-human IgG (South-

ernBiotech, Cat. No. 2040-09, final concentration, 0.1 mg/ 106 cells) incubated with HL60-CD38 and HL60-GFP cells

for 30 minutes at room temperature in a final volume of

100 mL. Separate gating on HL60-CD38 and HL60-GFP

cells was performed to assess specific DARA binding to

each cell population.

Negation of DARA binding to transduced HL60

cells

To remove cell surface CD38, mixtures of HL60-CD38 and

HL60-GFP cells (2.5 3 105 each) were incubated for 30

minutes at 37�C with DTT (0.1-10 mmol/L) or trypsin

(1%-2%) before incubation with DARA. The DTT concen-

tration was optimized for HL60 cells. Cell death was

observed when HL60 cells were treated with the standard

DTT concentration used to treat RBCs (0.2 mol/L). After

DTT incubation, cells were washed twice with PBS; after

trypsin incubation cells were washed once with growth

medium to inactivate trypsin and then twice with PBS.

The subsequent incubation with DARA and detection of

DARA with PE-labeled goat anti-human IgG followed the

same steps as described. For neutralization studies,

DARA-containing plasma (final concentration, 0.5 mg/mL) was incubated for 15 minutes with either anti-DARA idio-

type (final concentration, 5 mg/ml) or a control mouse anti-human antibody (mouse IgG1j, MOPC-21, Sigma- Aldrich) at room temperature. Likewise, increasing con-

centrations (0.05-5 mg/mL) of recombinant soluble human CD38 (R&D Systems) or identical concentrations of bovine

serum albumin (BSA, Sigma-Aldrich) were incubated with

DARA before mixing with HL60 cells for 15 minutes at

room temperature.

Statistical analysis

Comparison of groups was performed with a two-sided

unpaired t-test using computer software (GraphPad

Prism, Version 6.0c for Mac, GraphPad Software, La Jolla,

CA; www.graphpad.com).

RESULTS

DARA causes plasma panreactivity in vitro

On routine screening in the blood bank, five of five

patients with MM receiving DARA were observed to have

a positive antibody screen and panreactive plasma in RBC

TABLE 1. Summary of blood bank testing of DARA-treated patients

Patient (number of samples tested) Sex

Age (years)

DARA dose/ week (mg/kg)

Period from last DARA dose to BB screen (days)

Result of antibody screen and RBC panel DAT

Result of RBC panel using DTT-

treated cells

1 (3) Male 48 8 0-7 Pan reactive Positive Negative 2 (3) Male 68 8 7 Pan reactive Positive Negative 3 (10) Female 44 8 6-13 Pan reactive Negative Negative 4 (1) Female 66 16 0 Pan reactive Positive Negative 5 (1) Male 59 16 0 Pan reactive Not done Negative

BB 5 blood bank.

CHAPUY ET AL.

1548 TRANSFUSION Volume 55, June 2015

Fig. 2. Evaluation of methods to prevent DARA binding to CD38 on transduced HL60 cells. (A) DARA (0.5 mg/mL) binds selectively

to HL60-CD38 cells (black) and not control HL60-GFP cells (gray). Treating the cells with increasing concentrations of DTT (0-10

mg/mL) resulted in a dose-dependent reduction in DARA binding. (B) Quantification of reduced DARA binding after pretreating

HL60-CD38 cells with increasing concentrations of DTT or (C) trypsin compared to isotype control binding (gray). (D, E) Incuba-

tion of DARA-spiked plasma (0.5 mg/mL) with soluble CD38 (D) or anti-DARA idiotype (E) caused reduced DARA binding to

HL60-CD38 cells, while negative controls using BSA (D) or a mouse anti-human IgG (mah control; E) showed no effect on DARA

binding. (F) Quantitative assessment of DARA adsorption by HL60 CD381 cells. DARA-spiked plasma (0.5 mg/mL) was incubated

with HL60 CD381 cells or HL60 CD38– cells. After incubation, the plasma supernatants were assayed for residual DARA using

flow cytometry. Significant adsorption of DARA was detected after incubation with the higher dose (5 3 106 cells), but not the

lower dose (1 3 106 cells) of CD381 adsorbing cells. All data show a representative example of at least three independent experi-

ments. Error bars indicate SD.

DARATUMUMAB BLOOD BANK INTERFERENCE

Volume 55, June 2015 TRANSFUSION 1549

panel testing (Table 1). All samples were initially screened

using a solid-phase method (TANGO optimo); confirma-

tory testing was performed by tube testing with PEG

enhancement. Patients requiring transfusion on multiple

occasions had several longitudinal samples sent to the

blood bank. In total, the blood bank observed the pan-

reactivity in 18 of 18 samples from these patients. The

majority of the patients (3/5) had a positive DAT (IgG

only) and positive autocontrol. None of the five DARA-

treated patients showed signs of hemolysis. Typically,

Fig. 3. Eluate experiments: DARA binds specifically to RBC CD38. (A) Untreated or DTT-treated RBCs were incubated with DARA

and then washed. Eluates were then prepared by the acid elution technique.9 The eluates were added to a mixture of HL60-

CD38 cells and control HL60-GFP cells, and DARA binding was assessed by flow cytometry. (B, left panel) An eluate of untreated

RBCs contained recovered DARA that bound to HL60-CD38 cells (black) but not to HL60-GFP cells (gray). (Right panel) An elu-

ate of DTT-treated RBCs did not contain CD38-binding activity (i.e., did not contain recovered DARA). (C, left panel) An RBC elu-

ate prepared from a DAT-positive DARA-treated patient (Patient 4, Table 1) bound to HL60-CD38 cells (black) but not to HL60-

GFP control cells (gray). (Right panel) Incubating the patient eluate with anti-DARA idiotype reduced binding to HL60-CD38

cells. (D) Quantification of flow cytometry studies of the patient eluate. Incubating the patient eluate with anti-DARA idiotype

reduced binding to HL60-CD38 cells, while incubating with a mouse anti-human IgG control antibody did not have a substantial

effect. All data show a representative example of at least three independent experiments. Error bars indicate SD.

CHAPUY ET AL.

1550 TRANSFUSION Volume 55, June 2015

reaction strengths of the DARA-treated patient samples

were graded as weakly positive (M1 [macroscopically

positive] to 11) in both solid-phase and tube. When heter-

ologous adsorption studies were performed, the panreac-

tivity persisted after three passes using untreated RBCs.

ABO/Rh typing of patient RBCs was unaffected.

To verify that DARA was causing the observed agglu-

tination reactions, we first confirmed by flow cytometry

that CD38 is weakly expressed on human RBCs, as

reported previously 4-7

(Fig. 1A). Next, we spiked normal

plasma samples with increasing concentrations of DARA

(0.1-10 mg/mL). When these samples were incubated with reagent RBCs, panreactivity was observed at anti-

human globulin phase using no enhancement, PEG, and

LISS, at all concentrations tested. Reaction strengths

were graded as M1 (0.1-0.5 mg/mL) or 11 (1.0-10 mg/ mL). Six cycles of heterologous adsorptions10 using

untreated RBCs failed to eliminate the interference.

DATs performed on reagent RBCs incubated with

DARA-spiked plasma were positive (IgG only) at all

concentrations tested (0.1-10 mg/mL), with strengths varying between microscopically positive for concentra-

tions of 0.1-0.25 mg/mL and 11 for all higher concentrations.

Negating the DARA interference in an HL60 cell

model system

To study methods of eliminating the DARA interference in

the blood bank, we established a model system in HL60

cells (Figs. 1B-1D). The use of transfected HL60 cells to

study CD38 function was reported previously.8 Here, HL60

cells were stably transduced with either CD38 (Fig. 1B) or

GFP, which served as a CD38– control (Fig. 1C). Flow

cytometry confirmed specific, dose-dependent binding of

DARA in spiked plasma to CD381 HL60 cells but not to

CD38– controls (Figs. 1E and 1F).

Using this system, we evaluated methods to remove

CD38 antigen from the cell surface. Incubating CD381

HL60 cells with 10 mmol/L DTT reduced DARA binding

by 92% (p < 0.001, Figs. 2A and 2B). Treating CD381 HL60

cells with 2% trypsin reduced DARA binding by 40%

(p < 0.001, Fig. 2C).

In addition to CD38 antigen removal, we attempted

to inhibit DARA binding to CD381 HL60 cells by neu-

tralizing DARA in plasma. Soluble CD38 added to

DARA-spiked plasma reduced DARA binding to CD381

HL60 cells in a dose-dependent manner (Fig. 2D). Simi-

larly, addition of a neutralizing mouse anti-DARA idio-

type antibody decreased DARA binding by 95%. A

nonspecific control antibody had no effect on DARA

binding (Fig. 2E).

We also investigated using CD381 HL60 cells as

DARA-adsorbing cells (Fig. 2F). DARA-spiked plasma was

incubated with either HL60 CD381 cells or HL60 CD38–

control cells. The adsorbed plasma was then assayed for

the presence of residual DARA by flow cytometry. While

plasma adsorbed with 1 3 10 6

HL60 CD381 cells con-

tained high levels of residual DARA, plasma adsorbed with

5 3 106 HL60 CD381 cells contained significantly reduced

levels of residual DARA (p < 0.001). Adsorbing plasma

with HL60 CD38– control cells did not significantly reduce

the level of DARA detected.

DARA is present in eluates of untreated, but not

DTT-treated, RBCs

Untreated or DTT-treated RBCs were incubated with

DARA and then washed (Fig. 3A, Steps 1 and 2). Eluates

were then generated from untreated or DTT-treated RBCs

and used for flow studies on HL60 CD381 cells and

CD38– control cells (Fig. 3A, Steps 3-5). Eluates prepared

from the untreated RBCs contained specific IgG binding

to CD381 HL60 cells, while eluates prepared from DTT-

treated RBCs contained no detectable IgG binding to

CD381 HL60 cells (Figs. 3A and 3B). These eluates were

further serologically tested against a minipanel of five

RBC lines. No reactivity was observed in the eluate of

DTT-treated RBCs, but the eluate of untreated RBCs was

TABLE 2. Representative serology results: identification of anti-E in the presence of DARA*

Screening cell Plasma Alloantibody Antibody screen result Panel cells Panel result

Cell 1 No DARA – 0 Untreated No reactivity Cell 2 0 Cell 1 No DARA Anti-E 0 Untreated Anti-E Cell 2 11 Cell 1 1 DARA – 11 Untreated Panreactivity Cell 2 11 Cell 1 1 DARA Anti-E 11 Untreated Panreactivity Cell 2 11 Cell 1 1 DTT 1 DARA – 0 DTT-treated No reactivity Cell 2 1 DTT 0 Cell 1 1 DTT 1 DARA Anti-E 0 DTT-treated Anti-E Cell 2 1 DTT 11

* Screening Cell 1 5 phenotype R1R1, DCe; Screening Cell 2 5 phenotype R2R2, DcE.

DARATUMUMAB BLOOD BANK INTERFERENCE

Volume 55, June 2015 TRANSFUSION 1551

panreactive. These results are consistent with DTT dena-

turing RBC CD38 epitopes, preventing DARA binding.

An RBC eluate was prepared from a DARA-treated

patient sample (Patient 4, Table 1). The eluate was pan-

reactive on RBC panels. Flow cytometry confirmed that

the eluate contained binding activity to CD381 HL60 cells

but not to CD38– controls (Fig. 3C). Binding to CD381

HL60 cells was specifically inhibited by the addition of

anti-DARA idiotype (Figs. 3C and 3D), confirming the

presence of DARA in the patient sample.

Negating the DARA interference with blood bank

tests

We performed a series of experiments on DARA-spiked

plasma samples (1.0 mg/mL) using routine blood bank serologic assays. Treating reagent RBCs with DTT or

trypsin eliminated the panreactivity with these samples.

This allowed identification of underlying clinically sig-

nificant alloantibodies (anti-E, anti-Fya, anti-Jka, or anti-

s) in the presence of DARA. Representative blood bank

serology results for identifying anti-E are shown in

Table 2. Similar results were obtained by neutralizing

DARA-spiked samples with either anti-DARA idiotype

or soluble hCD38 (data not shown).

Adding anti-DARA idiotype to the plasma of DARA-

treated patients specifically eliminated positive antibody

screen reactions (Table 3), confirming that the positive

antibody screens seen in the clinical samples were directly

caused by DARA. Finally, using DTT-treated reagent RBCs,

the DARA interference was completely eliminated from

the plasma of all five DARA-treated patients and all 18

patient samples, allowing the safe release of blood prod-

ucts for these patients.

DISCUSSION

The use of anti-CD38 is a promising treatment for patients

with MM. 2,3

A problem with DARA is that it interferes with

blood compatibility testing, complicating the safe release

of blood products. Here, we show that direct binding of

DARA to endogenous CD38 on RBCs causes the panreac-

tivity observed in antibody screens and other pretransfu-

sion tests. Several lines of evidence support this

conclusion. First, a fluorescently labeled anti-CD38 was

shown by flow cytometry to bind directly to reagent RBCs.

Second, normal plasma samples spiked with DARA and

incubated with reagent RBCs recapitulated the interfer-

ence observed in the blood bank with samples from

DARA-treated patients. Third, an eluate prepared from the

RBCs of a DARA-treated patient bound only to CD381

cells and not to CD38– control cells. Fourth, an eluate pre-

pared from untreated RBCs that had been incubated with

DARA likewise bound only to CD381 cells and not to

CD38– control cells (i.e., contained recovered DARA).

Finally, when a specific anti-DARA neutralizing antibody

was added to DARA-treated patient samples or DARA-

spiked samples, RBC agglutination reactions were pre-

vented. Surface expression of RBC CD38 appears to be rel-

atively low, potentially explaining the weak agglutination

reactions seen in vitro. Despite binding to RBCs, DARA did

not cause significant hemolysis in the five treated patients.

Multiple rounds of adsorption with untreated or

ZZAP-treated RBCs failed to remove the panreactivity

from the plasma of DARA-treated patients. ZZAP con-

tains DTT, so ZZAP-treated RBCs are predicted to have

denatured CD38 surface antigen that would fail to bind

DARA. In contrast, we speculate that adsorptions using

untreated RBCs failed to remove the panreactivity from

DARA-treated patient samples due to low expression of

intact CD38 antigen on the adsorbing RBCs. The low

expression of CD38 on RBCs is reflected in the flow

cytometry results shown in Fig. 1A. Experiments using

transduced CD381 HL60 cells as adsorbing cells (Fig.

2F) demonstrated detectable removal of DARA from

spiked plasma only when a high number of adsorbing

cells was used. RBCs appear to express considerably

less surface CD38 than the transduced CD381 HL60

cells; thus we would expect RBCs to function relatively

poorly as DARA-adsorbing cells.

We evaluated a number of potential methods to neg-

ate the DARA interference in the blood bank. The extrac-

ellular domain of human CD38 contains six disulfide

bonds that are critical to the protein structure. 13

Previous

TABLE 3. Neutralization of DARA in patient plasma samples

Plasma sample DARA dose Days from last DARA infusion

Initial antibody screen result

(Cells 1 and 2) Neutralization

Antibody screen result (Cells 1 and 2)

after neutralization

Normal plasma spiked with DARA

1 mg/mL NA Positive Anti-DARA idiotype (10 mg/mL) Negative Isotype control (10 mg/mL) Positive

DARA-treated Patient 3 8 mg/kg/week 7 Positive Anti-DARA idiotype (100 mg/mL) Negative Isotype control (100 mg/mL) Positive

DARA-treated Patient 5 16 mg/kg/week 0 Positive Anti-DARA idiotype (100 mg/mL) Negative Isotype control (100 mg/mL) Positive

CHAPUY ET AL.

1552 TRANSFUSION Volume 55, June 2015

investigators reported that the enzymatic activity of CD38

was highly sensitive to reducing agents such as DTT 14

and 2-mercaptoethanol. 9

Berthelier and colleagues 8

reported that treating CD381 HL60 cells with DTT

decreased the binding of specific MoAbs to CD38 by

denaturing the protein. Additionally, trypsin was reported

to cleave the ectodomain of CD38 from the cell mem-

brane. 8

These findings provided the basis for our hypoth-

esis that DTT or trypsin could prevent DARA binding by

disrupting the extracellular domain of CD38 on RBCs.

Using the HL60 model system, we confirmed that DTT is

highly effective in denaturing CD38 and preventing DARA

binding. CD38 binding activity was completely absent

from an eluate prepared from DTT-treated RBCs incu-

bated with DARA. We further showed using samples from

DARA-treated patients that DTT pretreatment of RBCs

eliminated the DARA interference with blood bank tests.

Trypsin pretreatment of HL60 cells and of RBCs was also

successful in reducing DARA binding, but was less effi-

cient than DTT treatment.

Another approach to prevent DARA binding was

neutralization of free DARA in plasma by adding solu-

ble CD38 or an anti-DARA idiotype. Both methods

were highly effective in preventing DARA binding, and

DARA neutralization in solution is simpler to perform

than DTT treatment of RBCs. Disadvantages of these

neutralization methods, however, are higher costs and

a lack of widespread availability of the reagents. Large

quantities of soluble CD38 would be needed to treat

clinical samples from DARA-treated patients. [Correc-

tion added on 21-April-2015, after first online publica-

tion: "DTT-treated" changed to "DARA-treated" in the

preceding sentence] In contrast, DTT is very inexpen-

sive and is already used by blood banks.10

A potential drawback of DTT treatment is the dis-

ruption of a limited number of blood group antigens

(Table 4). The sensitivity of virtually all clinically signifi-

cant RBC antigens to DTT or trypsin has previously been

defined. 10

In routine clinical practice, anti-K is the only

commonly encountered, clinically significant antibody to

a DTT-sensitive RBC antigen. In practice, this issue is

readily addressed by providing K– units to DARA-treated

patients. More than 90% of all donated RBC units are K–

. 15

It is possible that using the DTT method to evaluate

a DARA-treated patient could result in a potentially sig-

nificant RBC alloantibody (e.g., anti-k, anti-Yta) being

missed, but this would be a very rare event. Here, we

have shown that the DTT method allows for the detec-

tion of alloantibodies in the presence of DARA from

those blood group systems—aside from Kell—that

account for the majority of clinically significant hemo-

lytic reactions: Rh, Duffy, Kidd, and MNS.

In addition to DARA, many antibody-based cancer

therapies are in various stages of development that might

similarly interfere with routine blood bank tests. In the

future, the DTT-based method described here may be use-

ful to eliminate novel in vitro interferences in the blood

bank, in those cases where the involved RBC antigen con-

tains extracellular disulfide bonds.

In conclusion, we showed that DARA potently inter-

feres with routine blood bank serologic tests by directly

binding to CD38 on RBCs. DTT treatment of reagent RBCs

is a robust method to negate the DARA interference,

allowing the safe provision of RBC units to DARA-treated

patients. Before patients are started on DARA, RBC anti-

gen phenotyping or genotyping is recommended.

ACKNOWLEDGMENTS

The authors thank William Savage and Leslie Silberstein for their

reviews of the manuscript, and the blood bank technologists at

Brigham and Women’s Hospital for their help with developing

and validating the DTT method. We also thank Barbara Bierer for

making this study possible.

CONFLICT OF INTEREST

PD is a full-time employee of Janssen. All other authors have dis-

closed no conflicts of interest.

REFERENCES

1. de Weers M, Tai YT, van der Veer MS, et al. Daratumumab, a

novel therapeutic human CD38 monoclonal antibody,

TABLE 4. DTT-sensitive blood group systems*

Blood group system name ISBT symbol Transfusion reaction potential

Dombrock DO Immediate/delayed, mild to severe Indian IN Very rare, decreased cell survival with IN1 John Milton Hagen JMH Delayed (rare) Kell KEL Immediate/delayed, mild to severe Knops KN No Landsteiner-Wiener LW Delayed, none to mild Lutheran LU No to moderate Raph RAPH No to moderate Cartwright YT Delayed (rare); mild

* Adapted from the Blood Group Antigen Facts Book. 15

DARATUMUMAB BLOOD BANK INTERFERENCE

Volume 55, June 2015 TRANSFUSION 1553

induces killing of multiple myeloma and other hematologi-

cal tumors. J Immunol 2011;186:1840-8.

2. Plesner T, Lokhorst H, Gimsing P, et al. Daratumumab, a

CD38 monoclonal antibody in patients with multiple

myeloma-data from a dose-escalation phase I/II study. ASH

Annual Meeting Abstracts 2012;120:73.

3. Lokhorst HM, Plesner T, Gimsing P, et al. Phase I/II dose-

escalation study of daratumumab in patients with relapsed

or refractory multiple myeloma. ASCO Meeting Abstracts

2013;31(Suppl 15):8512.

4. Deaglio S, Canella D, Baj G, et al. Evidence of an immu-

nologic mechanism behind the therapeutical effects of

arsenic trioxide (As(2)O(3)) on myeloma cells. Leuk Res

2001;25:227-35.

5. Mehta K, Shahid U, Malavasi F. Human CD38, a cell-

surface protein with multiple functions. FASEB J 1996;10:

1408-17.

6. Zocchi E, Franco L, Guida L, et al. A single protein immuno-

logically identified as CD38 displays NAD1 glycohydrolase,

ADP-ribosyl cyclase and cyclic ADP-ribose hydrolase activ-

ities at the outer surface of human erythrocytes. Biochem

Biophys Res Commun 1993;196:1459-65.

7. Albeniz I, Demir O, T€urker-Sener L, et al. Erythrocyte

CD38 as a prognostic marker in cancer. Hematology

2007;12:409-14.

8. Berthelier V, Laboureau J, Boulla G, et al. Probing ligand-

induced conformational changes of human CD38. Eur J Bio-

chem 2000;267:3056-64.

9. Guida L, Franco L, Zocchi E, et al. Structural role of disulfide

bridges in the cyclic ADP-ribose related bifunctional ectoen-

zyme CD38. FEBS Lett 1995;368:481-4.

10. Roback JD, Grossman BJ, Harris T, et al. Technical manual.

17th ed. Bethesda (MD): AABB; 2011.

11. Judd WJ, Johnson ST, Storry JR. Judd’s methods in

immunohematology. 3rd ed. Bethesda (MD): AABB;

2008.

12. Chapuy B, McKeown MR, Lin CY, et al. Discovery and char-

acterization of super-enhancer-associated dependencies

in diffuse large B cell lymphoma. Cancer Cell 2013;24:

777-90.

13. Zhao YJ, Lam CM, Lee HC. The membrane-bound enzyme

CD38 exists in two opposing orientations. Sci Signal 2012;5:

ra67.

14. Kontani K, Nishina H, Ohoka Y, et al. NAD glycohydrolase

specifically induced by retinoic acid in human leukemic

HL-60 cells. Identification of the NAD glycohydrolase as leu-

kocyte cell surface antigen CD38. J Biol Chem 1993;268:

16895-8.

15. Reid ME, Lomas-Francis C, Olsson ML. The blood group

antigen facts book. 3rd ed. Waltham (MA): Academic Press;

2012.

CHAPUY ET AL.

1554 TRANSFUSION Volume 55, June 2015