Idiotypic Cross-Reactivity of Immunoglobulins Expressed in Waldenstrom’s
Macroglobulinemia, Chronic Lymphocytic Leukemia, and Mantle Zone
Lymphocytes of Secondary B-Cell Follicles
By Ofra Axelrod, Gregg J. Silverman, Vip Dev, Robert Kyle, Dennis A. Carson, and Thomas J. Kipps
Monoclonal antibodies (MoAbs) specific for autoantibodyassociated cross-reactive idiotypes (CRls) of Waldenstrom’s
IgM react frequently with the surface lg (slg) expressed by
leukemia cells of patients with chronic lymphocytic leukemia
(CLL). Evaluation of the molecular basis for this crossreactivity indicates that such CRls are encoded by conserved
antibody variable region genes (V genes) that have undergone little or no somatic hypermutation. We find that such
anti-CRI MoAbs stain a subpopulation of cells within the
mantle zones surrounding the germinal centers of normal
human tonsil. In contrast, MoAbs specific for variable region
subgroup determinants react with cells in both the mantle
zones and germinal centers of secondary B-cell follicles. To
test whether mantle zone B cells not reactive with existing
anti-CRI MoAbs may express slg bearing as-yet-unrecognized CRls present on Igs produced by neoplastic cells of
some patients with Waldenstrom’s macroglobulinemia or
CLL, we immunized mice with purified Waldenstrom’s IgM
that have been characterized for their variable region sub-
groups using subgroup-specific antisera raised against synthetic peptides. The supernatants of hybridomas generated
from the splenocytes of immunized mice were screened for
their ability t o stain a subpopulation of mantle zone lymphocytes in human tonsil. With this approach, two new anti-CRI
MoAbs were identified, designated OAKl and VOH3. OAKl
binds t o a CRI present on a subset of K light chains of the V,1
subgroup. VOH3 recognizes a CRI determinant(s) present on
a subset of antibody heavy chains of the V3
, subgroup. Flow
cytometric analyses demonstrated that OAKl specifically
binds leukemia cells from 5 t o 20 patients (25%) with K light
chain expressing CLL. In addition, VOH3 reacted with the
leukemia cells from 1 of 17 (6%) patients tested. The success
of these methods demonstrates that the variable regions of
the Igs produced by mantle zone B cells share idiotypic
determinants with Igs expressed in B-cell CLL (B-CLL) and
Waldenstrom’s macroglobulinemia.
o 1991 b y The American Society of Hematology.
C
additional MoAbs specific for the protein products of
conserved hu-Ig V genes. Although “rescued” human
lymphoma antibody or myeloma paraproteins of unknown
specificity have been used to induce murine anti-CRI
Ig-producing cell^,"^'^ the V genes encoding such antibodies
may harbor numerous somatic mutation^.'^ These mutations may disrupt determinants encoded by V genes present
in the germline DNA. At least a subgroup of Waldenstrom’s tumors express hu-IgM proteins encoded by V
genes that apparently have not undergone substantial
somatic mutation, however.4 Thus, immunization of mice
with such hu-IgM proteins may induce antibodies that, like
17.109 or G6, can serve as useful serologic probes for
expression of conserved V genes.
We describe the production and characterization of
anti-CRI MoAbs against Waldenstrom’s hu-IgM paraproteins by a novel screening method. MoAbs directed against
major CRIs on hu-Ig stain a subset of lymphocytes within
the mantle zone surrounding the germinal centers of
human tonsil. Based on this observation, we screened our
hybridoma supernatants for their ability to stain tonsilar B
lymphocytes in fresh-frozen tissue sections. Hybridomas
producing antibodies that react with tissue sections in a
fashion simifar to that of other known anti-CRI MoAbs
were identified and isolated. We describe the identification
and characterization of two such hybridomas.
ROSS-REACTIVE idiotypes (CRIs) of human Ig
(hu-Ig) can be useful serologic markers for expression
of hu-Ig variable region genes (V genes). Two of the best
characterized human CRIs to date are defined by reactivity
with monoclonal antibodies (MoAbs), designated 17.109
and G6. These MoAbs were generated against CRIs present
on Waldenstrom’s hu-IgM paraproteins with anti-IgG or
rheumatoid factor (RF) binding activity.Ia2Subsequently,
they were shown to react frequently with the leukemia cells
from unrelated patients with B-cell chronic lymphocytic
leukemia (B-CLL) and related B-cell lymphomas,”” and
virgin B cells within the primary follicles of human fetal
spleen.6Investigations of the genetic basis for expression of
these CRIs show that each is encoded by a highly conserved
Ig variable region gene(s) (V gene) present in the germline
DNA.’.’’ As such, these CRIs apparently are serologic
markers for expression of these V genes without substantial
somatic mutation.
Waldenstrom’s IgM macroglobulins may be particularly
well suited to serve as immunogens for the generation of
From the Department of Medicine, University of Califomia, San
Diego, La Jolla, CA; and the Department of Medicine, Mayo Clinic,
Rochester, MN.
Submitted August 14, 1990; accepted November 21, 1990.
Supported in part by Grants No. CA49870, AR07144, AG04100,
and AR25443 from the National Institutes of Health, Bethesda, MD.
T.J.K. is a Scholar of the Leukemia Society, supported in part by the
Scoft Helping Hand Fund.
Address reprint requests to Thomas J. Kipps, MD, Department of
Medicine, University of Califomia, San Diego, CA 92093-0945.
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be hereby marked
“advertisement” in accordance with 18 U.S.C.section 1734 solely to
indicate this fact.
0 I991 by The American Society of Hematology.
0006-4971l91l7707-0017$3.0OlO
1484
MATERIALS AND METHODS
Antibodies. An IgG, MoAb specific for a V,IIIb subgroup
determinant(s),” was obtained from Dr George Abraham (University of Rochester, Rochester, NY). B6, a murine IgG, MoAb
generated against the heavy chain of an IgM-RF paraprotein and
reactive with a VH3-subgroup-associated CRI,” and G6, a murine
IgG, MoAb; were provided by Drs Rizgar Mageed and Roy
Jefferis (University of Birmingham, Birmingham, England). MoAb
6B6.6, specific for a VK3a-associated CRI,” was provided by Drs
Ralph E. Schrohenloher and William J. Koopman (University of
Blood, Vol77, No 7 (April 1). 1991: pp 1484-1490
1485
HUMAN ANTIBODY CROSS-REACTIVE IDIOTYPES
Alabama, Birmingham, AL). MoAb 17.109 is as described previously.’ DA4-4, an IgG, anti-human p heavy chain,” was acquired
from the American Type Tissue Culture Collection. Anti-human K
or A light chain-producing hybridomas were as described previously.” Mouse MoAb was purified from ascites by ammonium
sulfate precipitation and either absorption with QAE (Pharmacia
Fine Chemicals, Upsala, Sweden) (for IgG, and IgG,) or protein A
Sepharose-column chromatography (BioRad, Richmond, CA).
hu-IgM paraproteins (Table 1) either were isolated from sera of
patients with Waldenstrom’s macroglobulinemia by 45% saturated
ammonium sulfate precipitation and Sephadex-G200 (Pharmacia)
column chromatography or were obtained from commercial sources
(eg, Caltag, San Francisco, CA; Jackson ImmunoResearch Laboratories, West Grove, PA, Binding Site, Birmingham, England; or
Tago, Burlingame, CA). Ig K light chain Bence Jones proteins were
provided by Dr A. Solomon (Department of Medicine, University
of Tennessee, Knoxville, TN). IgM proteins A224 and L16 (Table
2), which have known heavy chain variable regions of VH5and VH6,
respectively, were provided by Dr Ton Logtenberg (Academisch
Ziekenhaus, Utrecht, The Netherlands). The human paraproteins
used were analyzed by polyacrylamide gel electrophoresis (PAGE)
and immunoblotting with antisynthetic peptide antisera, each
specific for a primary structural determinant(s) of a given hu-Ig
variable region subgroup, as described previously?’ Daudi, a
Burkitt’s lymphoma cell line producing K light chains of the V,1
subgroup:’ was grown in serum-free tissue culture medium (HL-1,
Ventrex Laboratories, Portland, ME). hu-Ig produced by this cell
line was purified by precipitation of culture supernatants with 45%
ammonium sulfate. To isolate the K light chain of the IgM,
paraprotein MAR (Table 1) we reduced the purified IgM in 10
mmol/L dithiotrietol (Calbiochem, La Jolla, CA) at 37°C for 2
hours in 0.5 mol/L Tris HCl, 2 mmol/L EDTA (pH 7.6) before
adding iodoacetimide (Sigma Chemical, St Louis, MO) to a final
concentration of 25 mmol/L. After a 60-minute incubation at 4”C,
the antibody heavy and light chains were separated on an AcA34
column (LKB, Upsala, Sweden) equilibrated with 3 mol/L guanidine HCI and 0.25 mol/L ammonium bicarbonate at pH 8.2.
Table 1. Panel of hu-lgM Paraproteins
No.
MoAb
Designation
A1887
A4053
A0103
A8843
A0701
MAR
HEA
VIN
JB043
RJ293
IC461
KD477
ME591
WH951
EF985
Light
Chain
Heavy
Chain
VK1
vH3
vK3
VH4
VK2
VK1
vH4
17.109
vH3
vK4
vH4
VK1
vK3
VH1
VH3
vK3
vH3
v,
v*
v*
CRI
17.109
666.6
vH4
vn3
vH3
VK2
vn3
v*
vn3
VK1
VK 1
H
‘3
66
66
B6
vH3
The heavy and light chain subgroups determined for each paraprotein are shown. Proteins 1 through 5 were obtained from the Binding
Site (Birmingham, England), Calbiochem (La Jolla, CA), Caltag Laboratories (San Francisco, CA), Jackson ImmunoResearch Laboratories (West
Grove, PA), and Tag0 (Burlingame, CA), respectively. Paraproteins 6
through 8 and 9 through 15 were purified from serum samples of
patients with Waldenstrom‘s macroglobulinemia at the Scripps Clinic
(La Jolla, CA) and Mayo Clinic (Rochester, MN), respectively.
Table 2. VOH3 Binding t o Representative Paraproteins With Heavy
Chains of Different V, Subgroups
No.
Designation
13
1
14
11
4
7
8
10
15
6
ME591
A1887
WH951
IC461
A8843
HEA
VIN
RJ293
EF985
MAR
CESS
A0701
A224
L16
5
Light
Chain
Heavy
Chain
Cw* to
VOH3
C,,to
Anti-Cp
90
120
180
NRS
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
160
170
190
210
140
210
260
275
210
200
NTt
210
NT
NT
*Cs0 is the concentration (in ng/mL) of hu-Ig required to achieve 50%
of maximum binding to either VOH3 or an anti-Cp MoAb.
tNT indicates that the protein was nonreactive with VOH3 at an
excess saturating concentration (5 pglmL), and therefore was not
titrated on plates coated with anti-CRI or anti-constant region MoAb.
SNR indicates that the protein was nonreactive at the highest
concentration tested.
Separated chains were dialyzed extensively against 0.25 mol/L
NH,CO, and then lyophilized.
Generation of monoclonal anti-idioiypes. F, (BALB/c x N J )
mice (Research Institute of Scripps Clinic, La Jolla, CA) were
immunized with 100 kg purified paraprotein emulsified in complete Freund’s adjuvant (Sigma) administered subcutaneously
(SC). After 3 weeks, these animals received two subsequent
booster injections of 100 pg purified paraprotein emulsified in
incomplete Freund’s adjuvant (Sigma) at intervals more than 1
week apart. Three days before fusion, mice were boosted intraperitoneally (IP) with 100 pg paraprotein dissolved in phosphatebuffered saline (PBS, pH 7.2). Fusion was performed as described;* with slight modifications. A non-Ig-producing variant of
P3-X-63-Ab8 resistant to
mol/L 8-azaguanine was fused with
mouse spleen cells using polyethylene glycol (PEG) 1500 (Boehringer Mannheim Biochemicals, Indianapolis, IN). Cells were
suspended in Dulbecco’s modified Eagle’s medium (DMEM)
supplemented with 15% fetal calf serum (FCS) (Hyclone, Logan,
UT) and 10% hybridoma growth factor (IGEN, Rockville, MD)
and seeded at 1 X 106/mL in 24-well plates. Twenty-four hours
later, selection medium (supplemented with hypoxanthineaminopterin-thymidine, Boehringer Mannheim) was added to
cultures. Ten to 14 days later, supernatants were collected and
checked for reactivity. Hybridoma cultures corresponding to positive supernatants were subcloned by limiting dilution. Positive
subclones were injected IP into pristane-primed syngeneic mice to
produce ascites.
Enzyme-linked immunoabsorbent assay (ELZSA). Polystyrene
microtiter plates were coated with hu-IgM at 5 pg/mL in boratebuffered saline (BBS), pH 8.2. Plates were washed free of unbound
material with borate buffer (BBS at pH 8.2) containing 1% bovine
serum albumin (BSA) to saturate residual plate protein binding
sites. Serum or supernatant samples diluted in 1% BSA in BBS
(BSABBS) were added to the plates and allowed to incubate
overnight. Plates then were washed with 0.05% Tween detergent in
BSABBS before incubation with alkaline-phospbatase-conjugated heterologous antibody specific for mouse Ig (Southern
Biotech, Birmingham, AL). After a subsequent 1-hour incubation,
1486
AXELROD ET AL
the plates were washed with BSA/BBS, developed with freshly
prepared p-nitrophenyl phosphate disodium (Sigma) in carbonate
buffer, and then monitored at OD, with an ELISA plate reader.
hu-IgM with variable region subgroups distinct from that of the
hu-IgM immunogen was added to hybridoma supernatants to a
final concentration of 100 pg/mL. After an overnight incubation at
4"C, these supematants were reexamined for reactivity with the
hu-IgM immunogen or an irrelevant hu-IgM.
Zmmunohistochemlstry. Residual lymphoid tissue from tonsillectomies were frozen in optimum cutting temperature medium
(Miles Laboratories, Naperville, IL). Four-micron sections were
prepared from the tissue blocks of frozen human tonsil for
immunohistochemical analyses with an avidin-biotin complex immunoperoxidase technique as described previously.z Hybridoma culture supernatants with and without exogenous hu-IgM (at 100
pg/mL) were incubated overnight at 4°C before use. Sixty microliters culture supernatant was added to each slide and allowed to
incubate at room temperature for 60 minutes. Slides were washed
with PBS (pH 7.2) before 60 pL biotinylated horse anti-mouse Ig
3.3 pg/mL was added (Vector Laboratories, Burlingame, CA).
After another 60-minute incubation at room temperature, the
slides were washed and then exposed to avidin D coupled to
horseradish peroxidase (Vector Laboratories) for 60 minutes at 10
pg/mL. After washing the slides with PBS, we developed the bound
peroxidase with 3-amino-9-ethylcarbazole (Sigma) at 0.4 mg/mL in
0.015% hydrogen peroxide (Sigma) in 0.1 molL sodium acetate
(pH 5.2) for approximately 15 minutes. Washed slides then were
stained with Mayer's hematoqlin (Sigma) for 2 minutes before
being mounted.
Zmmunoblot analyses. hu-Ig were analyzed for variable region
subgroup or for reactivity with identified anti-CRI MoAb using
immunoblotting as described previ~usly.~~~"
Test paraproteins were
run on polyacrylamide gel under denaturing and reducing conditions. Thereafter, separated heavy and light chains were transferred electrophoreticallyto a transfer membrane (Immobilon-P,
Millipore, Bedford, MA) and incubated with either anti-CRI
MoAb, antiheavy or antilight chain antibody (mouse anti-IgM or
goat anti-human K light chain, Calbiochem) or rabbit antisera
generated against peptides corresponding to variable region subgroups as described pre~iously.~"~"
Filters were developed with
'*'I-protein A (ICN Biomedicals,Costa Mesa, CA).
Flow cytometry. Leukemia cells from patients with CLL were as
described previously.' Indirect immunofluorescenceanalysesof the
sIg expressed by leukemia cells of patients with CLL with anti-CRI
MoAbs were performed using a FACScan flow cytometer (Becton
Dickinson, San Jose, CA). Leukemia cells were incubated in
staining medium (SM, consisting of RPMI 1640, 3% FCS, 100
mmol/L HEPES (pH 7.6) and 1 pg/mL propidium iodide) containing saturating amounts of anti-CRI MoAb or an irrelevant mouse
MoAb of the same isotype. After 20 minutes at 4"C, cells were
washed in propidium-iodidedeficient SM and then developed
with fluorescein-conjugated goat antibodies specific for mouse Ig
(Southern Biotechnology Associates, Birmingham,AL). Dead cells
with fluorescence at greater than 600 nm when excited at 488 nm
were excluded from the analyses.
RESULTS
Characterizationof hu-ZgMparaproteins. By immunoblotting with heterologous antisera generated against synthetic
peptides corresponding to subgroup-specific residues, we
determined the variable region subgroups of the heavy and
light chains of each of the assembled Waldenstrom's
hu-IgM paraproteins (Table 1) and Bence Jones proteins.
hu-IgM heavy chains reacted with only one of the six
different antisera specific for each of the heavy chain
subgroups, allowing us to assign a single heavy chain
subgroup for each hu-Ig. Ten hu-IgM (67%) belonged to
the VH3 subgroup, four (27%) belonged to the V,4 subgroup, and only one (7%) belonged to the VH1subgroup. Of
the light chain variable regions of the 11 hu-IgM that had K
light chains, five (45%) belonged to the VK1subgroup, three
(27%) represented the VK3subgroup, two (18%) were V,2,
and only one belonged to the VK4 subgroup (Table 1).
Similarly, we could assign the light chains of K hu-Ig to one
of the four K light chain subgroups. Of the 18 K light chain
Bence Jones proteins available, 10 (55%) were assigned to
VK1, three (17%) to VK2, three (17%) to VK3, and two
(11%) to VK4. The variable region subgroups of h light
chains were not determined.
In addition, we tested the hu-IgM paraproteins for
reactivity with each of several existing anti-CRI MoAbs
(Table 1). Three of the 10 paraproteins characterized as
having heavy chain variable regions of the VH3 subgroup
reacted with B6, consistent with recognition of a VH3associated CRI by this MoAb.I6 Two of the three paraproteins with K light chains of the VK3subgroup reacted with
17.109. The other paraprotein with VK3 light chains was
recognized by 6B6.6. To produce MoAbs reactive with
previously undetected CRIs, we chose to generate MoAbs
against MAR (a VKIVHlhu-IgM,) (Table 1) and ME591 (a
V,VH3 hu-IgM,) (Table 1). Unlike other Waldenstrom's
hu-IgMs previously used to generate anti-CRI MoAbs,
these two paraproteins had neither RF activity nor antinuclear autoantibody activity (data not shown).
ELZSA screen for anti-idiotypic MoAbs. Ten to 14 days
after fusion, supernatants from growing hybridoma cultures
were assayed for anti-idiotypic activity in an ELISA. Supernatants were incubated on ELISA plates coated with the
hu-IgM immunogen. Because the hybridomas of each well
may be oligoclonal, hu-IgM of irrelevant heavy and light
chain variable region subgroups was added to the positive
supernatants to absorb possible activity against the constant
portion of the hu-IgM antibody molecule (Table 1). Supernatants with reactivity for the hu-IgM immunogen in the
presence of an irrelevant hu-IgM were scored as having
potential antiidiotypic activity.
Immunohistochemical detection of MoAbs specific for human CRI. Anti-CRI MoAbs can be detected by screening
hybridoma supernatants on tissue sections of human tonsil.
Antibodies specific for a major CRI stain a subpopulation
of tonsilar lymphocytes that have a distinctive histologic
distribution; eg, 17.109, specific for a K light chainassociated C R I encoded by a conserved VK3gene, labels a
subpopulation of lymphocytes that reside in the mantle
zone surrounding each of the germinal centers (Fig 1A).
This distinctive distribution of CRI-reactive cells has been
noted in sections of every tonsil specimen examined to date
(n > 24). In contrast, MoAbs directed against variable
region subgroup determinants stain a subpopulation of cells
in both the mantle zones and germinal centers (Fig 1B).
Antibodies specific for the constant region of hu-IgM react
with nearly all cells in the follicle (Fig 1C).
Supernatants with potential antiidiotypic reactivity by
HUMAN ANTIBODY CROSS-REACTIVE IDIOTYPES
1487
Table 3. OAK1 Binding to Paraproteins With RepresentativeK Light
Chains of Different V. Subgroups
No.
Designation
6
1
4
14
15
6
MAR
At887
A8843
WH951
EF985
MAR(LCt)
GREIBJS)
HAU (BJ)
AU (BJ)
SCW (BJ)
DAUDI
A0103
A4053
A0701
Light
Chain
Heavy
VK1
v, 1
VKl
VKl
VK1
VK1
V,1
V,1
VKl
V,1
b.1
Vnl
V"3
V.1
Chain
CWto
C,to
Anti-Cr
OAKl.
~
3
2
5
Flg 1. Immunohistochemicalanalyses of human tonsil. Comparable tissue sections were stained with either (A) 17.109, (E) anti-V,lllb
subgroup MoAb, (C) anti-human IgM MoAb, or (0) OAK1.
ELISA were used to stain sections of fresh-frozen human
tonsil. Supernatants were tested with and without an excess
concentration (100 p.g/mL) of the hu-IgM immunogen or
irrelevant hu-IgM protein. A few wells from each fusion
reacted with a subpopulation of cells within the mantle
zone in the presence of added irrelevant hu-IgM protein
(data not shown). Addition of the original hu-IgM immunogen to such supernatants completely blocked their reactivity for human tonsil (data not shown). From such tissue
reactivity, these supematants were assumed to have antiCRI activity.
Hybridomas with supernatants having presumed antiCRI activity were cloned by limiting dilution. Two MoAbs
produced in this way, OAKl [specific for a CRI present on
MAR (Table l)] and VOH3 [specific for a CRI present on
ME591 (Table l)], retained the ability to stain a subpopulation of lymphocytes confined to the mantle zones of the
secondary follicles of human tonsil (eg, Fig 1D).
Spec.$city and cross-reactivityof OAKI and VOH3. The
reactivity of OAKl was assayed against each of the 15
characterized Waldenstrom's IgM proteins (Table l), 18 K
light chain Bence Jones proteins, and the separated heavy
and light chains of MAR, the hu-IgM paraprotein used to
induce OAK1. Each protein was screened at 5 &mL in an
ELISA using plates precoated with anti-CRI MoAb. In
addition, to compare the relative binding intensities of the
anti-CRI for the various Ig proteins, representative hu-Ig
were each titrated in parallel on plates precoated with
anti-CRI or anti-constant region MoAb. From these studies, we found that OAKl binds to the isolated K light chain
of MAR, the immunogen used to induce OAKl (Table 3).
Furthermore, MoAb bound to several hu-Ig with K light
chains of the VK1variable region subgroup, but did not bind
to the isolated heavy chain of MAR or to any hu-Ig with
either A light chains or K light chains belonging to the VK2,
v.3
v.3
None
None
None
None
None
ND
VH4
vn4
VI4
vK2
v,3
vK4
125
170
160
150
NRS
30
30
NR
NR
NR
NR
NR§
NR
NR
160
250
200
170
180
30
30
40
30
80
210
200
120
210
"C,is the concentration(in nglmL) of hu-lg required to achieve 50%
of maximum bindingto either OAKl or an antic, MoAb.
tLC in parentheses indicates that the sample was the K light chain
isolatedfrom the Waldenstr6m's paraprotein, MAR (no. 6, Table 1).
SBJ in parentheses designates that the protein tested was a Bence
Jones K light-chain myeloma paraprotein.
§NR indicates that the protein was nonreactive at the highest
concentration tested.
VK3, or VK4 subgroup (Tables 1 and 3). Titration of
OAK1-reactive hu-Ig showed each to have comparable
binding activities to plates coated with either OAKl or
anti+ constant region MoAb (Table 3), indicating that the
relative binding activities of OAKl for each of the OAKlreactive hu-Ig are similar. OAKl did not react with all hu-Ig
with K light chains of the V,1 subgroup, however; eg,
although the actual or deduced amino acid sequences of
AU, HAU, S C W , and DAUDI place these K proteins in the
VK1subgroup,2' these proteins are not bound by the OAKl
MoAb (Table 3). In all, OAKl recognizes 5 of 15 (33%) of
the tested hu-Ig with K light chain3 of the VK1subgroup.
From these data, we conclude that OAKl is specific for a
VK1-associatedCRI rather than a VK1-subgroupdeterminant(s).
Immunoblotting confirmed that OAKl is spec'ific for a
CRI present on the K light chains belonging to the VK1
variable region subgroup (Fig 2). OAKl reacted with the
separated K light chains of the hu-Ig immunogen. Furthermore, OAKl exclusively bound the separated K light chains
A
B
C
1 2 3 4 5 6
1 2 3 4 5 6
1 2 3 4 5 6
Fig 2 lmmunoblot analyses with OAK1 of electrophoretically
size-separated K light chains from the hu-lg preparations with V,
subgroups shown in parentheses: lane 1, MAR-K (V,l); lane 2, MAR
lV,l); lane 3. A1887 (V,l); lane 4, SCW B.J. (V,l); lane 5, A0103 (V,2);
lane 6, A0701 ( V p ) .Autoradiogram filters probed with (A) OAKl, (E)
peptide-induced anti-V,l antiserum, or (C) anti-lc light chain antlserum.
AXELROD ET AL
1488
A
1
2
3
4
5
C
B
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
patients, three had leukemia cells reactive with the 17.109CRI. Consistent with OAKl recognizing a V,l-associated
CRI, none of the three 17.109-reactive CLL expressed the
OAKl CRI. In addition, none of the leukemia cells from
patients with A light chain-expressing CLL reacted with
OAK1. Despite recognizing a CRI associated with heavy
chain variable regions of the relatively large VH3subgroup,
however, VOH3 reacted with the leukemia lymphocytes
from only 1 of 17 patients tested (data not shown). Again,
the leukemia cells reactive with VOH3 did not express the
B6-CRI.
---?!7&
Fig 3. lmmunoblot analyses with VOH3 of electrophoretically
size-separated heavy chains from the hu-lg preparations with V,
subgroups shown in parentheses: lane 1, ME591 (V,3); lane 2, A1887
(V,3); lane 3, MAR (V,l); lane 4, A4053 (V,,4); lane 5, IC461 (V,3); lane
6, KD477 (V,3); lane 7, normal serum IgM; lane 8, normal serum IgG.
Autoradiograms of filters probed with (A) VOH3, (B)anti-V,3 antiserum, or ( C )anti-p, heavy chain MoAb.
of some but not all hu-Ig with K light chains of the V,1
variable region subgroup (Fig 2).
VOH3, on the other hand, binds an Ig heavy chainassociated CRI. Similar analyses with VOH3 showed that
this anti-CRI has binding activities for hu-IgM, paraproteins, A1887 or WH951, that are comparable to that for the
hu-IgM, protein, ME591, used to generate this MoAb
(Table 2). All three of these Waldenstrom’s paraproteins
have heavy chain variable regions belonging to the VH3
subgroup. Immunoblotting demonstrated that VOH3 binds
to the separated heavy chains of VOH3-reactive paraproteins (Fig 3). However, VOH3 did not bind to any of the
other 12 Waldenstrom’s IgM paraproteins (Table 1) or to
representative IgM proteins of all six V, subgroups other
than VH3(Table 3). Moreover, VOH3 did not react with all
hu-Ig of the VH3subgroup; eg, VOH3 did not react with any
of the hu-Ig found positive for B6, a CRI associated with Ig
heavy chain variable regions of the VH3subgroup‘6(Tables
1 and 2 and Fig 3). In addition, although the separated
heavy chains of either serum IgM or IgG react with the
antipeptide antisera specific for Ig V,3 subgroup (Fig 3,
group B, lanes 7 and 8, respectively), they do not have
detectable reactivity for VOH3 by immunoblot analysis (Fig
3, group A, lanes 7 and 8, respectively). Collectively, these
studies indicate that VOH3 is specific for a CRI that is
distinct from B6 and present on a subset of heavy chain
variable regions of the V,3 subgroup.
Reactivity of OAKl and VOH3 with lymphocytes of CLL
patients. The newly generated anti-CRI MoAb reacted
with the leukemia cells of unrelated patients with CLL. By
flow-cytometric analysis, OAKl was shown to stain the
leukemia cells from 5 of 20 patients (25%) with K light
chain-expressing CLL (eg, Fig 4). In this group of 25
A
B
DISCUSSION
MoAbs generated against Waldenstrom’s macroglobulins specific for prominent CRIs expressed in CLL stain a
subpopulation of lymphocytes within human tonsil. The
tonsillar B cells expressing these CRIs are confined to the
mantle zones surrounding the germinal centers of secondary B-cell follicles. Taking advantage of this anatomic
distribution of CRI-reactive B cells, we developed methods
to screen for additional anti-CRI/MoAb-producing hybridomas. This allowed us to identify two new anti-CRI/
MoAbs, designated OAKl and VOH3, that bind exclusively
to a subset of mantle zone lymphocytes. Immunochemical
studies demonstrated that these MoAbs recognize a CRI
determinant(s) present on a subset of K light chains of the
VK1subgroup or a subset of Ig heavy chains of the V,3
subgroup, respectively. Including the hu-IgM used as immunogen, OAKl or VOH3 reacted with 27% (4 of 15) or 20%
(3 of 15) of the hu-IgM paraproteins assembled for this
study, respectively (Table l), suggestingthat these CRIs are
frequently present on the IgM paraproteins of patients with
Waldenstrom’s macroglobulinemia. Furthermore, a screening of unrelated patients with CLL suggests that the OAKl
and VOH3 CRIs also may be expressed by the leukemia
cells in this disease. The overall success of these methods
indicates that variable regions of the Igs produced by
mantle zone B cells share idiotypic determinants with Igs
expressed in Waldenstrom’smacroglobulinemiaand B-CLL.
Certainly the expression frequencies of these CRIs may
differ among patients with Waldenstrom’s macroglobulinemia or CLL. In a previous study of IgM paraproteins mostly
from patients with Waldenstrom’s macroglobulinemia, 10
of 23 (43%) were typed as having heavy chain variable
C
Fig 4. Representative flow cytometric analyses of leukemic cells examined for reactivity with anti-CRI MoAbs. The logarithmic fluorescence
intensity at 540 nm of cells excited at 480 nm is shown on the abscissa; the relative cell number is shown on the ordinate. Leukemic cells were
stained with either (A) an lgGl control MoAb of nonspecific activity, (E) goat anti-human K light chain MoAb, or (C)OAKl. The proportion of cells
labeling with anti-K MoAb (94%) was comparable to that reacting with the OAKl MoAb (92%).
1489
HUMAN ANTIBODY CROSS-REACTIVE IDIOTYPES
regions that were most likely encoded by V, genes of the
V,3 subgroup, the largest of the V, gene subgroups in the
human heavy chain locus.26 Of the 15 Waldenstrom’s
proteins typed in these studies, two thirds had heavy chain
variable regions of the V,3 subgroup (Table 1). Although
V, genes belonging to the V,3 gene family may encode
polyreactive antibodies produced by CD5 B cells:7328antibodies of the V,3 subgroup may be proportionately underrepresented among the Igs expressed in CLL.439s29-3’
Consistent
with this notion, VOH3, although reactive with 3 of 15
(20%) of the tested Waldenstrom’s proteins, reacted only
with the leukemia cells from 1of 17 (6%) randomly selected
patients with CLL. This may reflect a difference in V, gene
subgroup use between these two B-cell malignancies.
In addition, the results of our limited survey suggest that
the expression frequencies of these CRIs may differ between patients with Waldenstrom’s macroglobulinemia or
multiple myeloma. Although the proportion of K light chain
Waldenstrom paraproteins typed as V,1 (45%) was comparable with that of V,1 K Bence Jones paraproteins (55%),
OAK1 reacted with 80% (4 of 5) of the Waldenstrom’s
hu-Ig with K light chains of the VK1subgroup, but only 10%
(1 of 10) of the V,1 myeloma light chains. Conceivably, the
V, genes encoding many of these VK1Bence Jones proteins
may have incurred somatic mutations, ultimately disrupting
the CRI determinant(s) recognized by OAK1. Alternatively, although not excluding the former notion, the V,1
genes expressed in multiple myeloma may differ from those
used in Waldenstrom’s macroglobulinemia.
Differences in the expression frequencies of CRIs in
B-cell malignancies may reflect apparent differences in the
CRI expression frequencies of various subpopulations of
normal B cells. Each of these anti-CRI MoAbs stains a
subpopulation of mantle zone lymphocytes but rarely any
germinal center B cells. Conceivably, germinal center B
cells may express V genes that differ from those expressed
by mantle zone B cells. Alternatively, although not excluding the former hypothesis, germinal center B cell may
express V genes that have undergone somatic hypermutation. In this regard, the scarcity of CRI’ cells in the
germinal centers may mimic the low frequencies at which
such CRIs are detected on Ig expressed by non-Hodgkin’s
lymphomas of presumed follicular center cell rigi in.^ The
Ig V genes expressed by such lymphomas demonstrate
intraclonal diversity in their expressed Ig V genes indicative
of ongoing somatic m~tation.~’
Somatic mutation of the V
genes expressed by these lymphomas, and the germinal
center cells from which they are derived, may disrupt and
destroy variable region determinants that are recognized by
certain anti-CRI MoAbs. Consistent with this notion,
MoAbs specific for subgroup determinants that are more
resilient to the structural permutations induced by somatic
hypermutation stain cells in both the germinal centers and
the surrounding mantle zones.
The use of typed Waldenstrom’s IgM paraproteins together with the screening methods we describe may facilitate detection of additional MoAbs specific for important
but as yet undetected CRIs. Previous methods required
testing anti-idiotypic antibodies on a panel of purified hu-Ig
or assaying for inhibition of anti-idiotypelidiotype binding
with pooled normal hu-IgM or
In the former
method, the panel of hu-Ig had to be large enough to detect
less common CRIs. Using the latter method, less common
CRIs or CRIs not present on secreted hu-Ig may not be
detected. By analyzing the staining pattern of each hybridoma culture supernatant on human tonsil, however, we
can distinguish between MoAbs having antiprivate idiotype, anti-CRI, antisubgroup, or anticonstant region binding activity. As we have shown, newly generated anti-CRI
MoAbs may react with leukemia cells from patients with
CLL. Batteries of MoAbs recognizing disparate CRIs may
allow early detection and possible immunotherapy of B-cell
neoplasia in a large proportion of patient^.",'^,^^"^ Finally, as
the molecular basis for idiotypic cross-reactivity is resolved,
such reagents may be useful for probing V gene expression
in B-cell lymphoproliferative, immunodeficiency,or autoimmune diseases.
ACKNOWLEDGMENT
We thank Drs Rizgar Mageed and Roy Jefferis (University of
Birmingham, Birmingham, England), George Abraham (University of Rochester, Rochester, NY),Ralph E. Schrohenloher and
William J. Koopman (University of Alabama, Birmingham, AL),
and Ton hgtenbeig (Academisch Ziekenhaus, Utrecht, Netherlands) for providing us with some of the MoAbs used in these
studies. We also thank Dr Alan Solomon (University of Tennessee
Medical Center, Knoxville, TN) for providing us with the Bence
Jones proteins used in this study.
REFERENCES
1. Carson DA, Fong S: A common idiotope on human rheuma5. Kipps TJ, Robbins BA, Tefferi A, Meisenholder G, Banks
toid factors identified by a hybridoma antibody. Mol Immunol
PM, Carson D A CD5-positive B-cell malignancies frequently
20:1081,1983
express cross-reactive idiotypes associated with IgM autoantibod2. Mageed RA, Dearlove M, Goodall DM, Jefferis R: Immunoies. Am J Pathol 136:809,1990
genic and antigenic epitopes of immunoglobulins XVII-Mono6. Kipps TJ, Robbins BA, Carson DA. Uniform high frequency
clonal antibodies reactive with common and restricted idiotopes to
expression of autoantibody-associated cross reactive idiotypes in
the heavy chain of human rheumatoid factors. Rheumatol Int
the primary B cell follicles of human fetal spleen. J Exp Med
6:179, 1986
171:189, 1990
3. Kipps TJ, Robbins BA, Kuster P, Carson D A Autoantibody7. Kipps TJ, Tomhave E, Chen PP, Carson D A Autoantibodyassociated cross-reactive idiotypes expressed at high frequency in
associated kappa light chain variable region gene expressed in
chronic lymphocytic leukemia relative to B-cell lymphomas of
chronic lymphocytic leukemia with little or no somatic mutation.
follicular center cell origin. Blood 72:422, 1988
Implications for etiology and immunotherapy. J Exp Med 167:840,
4. Kipps TJ: Immunoglobulin idiotypes in human B cell neopla1988
sia. Chem Immunol48:167,1990
8. Chen PP, Albrandt K, Kipps TJ, Radoux V, Liu F T , Carson
1490
DA: Isolation and characterization of human VkIII germ-line
genes. Implications for the molecular basis of human VkIII light
chain diversity. J Immunol 139:1727, 1987
9. Kipps TJ, Tomhave E, Pratt LF, Duffy S, Chen PP, Carson
DA: Developmentally restricted V H gene expressed at high
frequency in chronic lymphocytic leukemia. Proc Natl Acad Sci
USA 865913,1989
10. Chen PP, Liu MF, Glass CA, Sinha S, Kipps TJ, Carson D A
Characterization of two immunoglobulin VH genes that are homologous to human rheumatoid factors. Arthritis Rheum 32:72,1989
11. Stevenson FK, Wrightham M, Glennie MJ, Jones DB,
Cattan AR, Feizi T, Hamblin TJ, Stevenson GT: Antibodies to
shared idiotypes as agents for analysis and therapy for human B
cell tumors. Blood 68:430,1986
12. Kiyotaki M, Cooper MD, Bertoli LF, Kearney JF, Kubagawa
H: Monoclonal anti-Id antibodies react with varying proportions of
human B lineage cells. J Immunol138:4150,1987
13. Miller RA, Hart S, Samoszuk M, Coulter C, Brown S,
Czerwinski D, Kelkenberg J, Royston I, Levy R: Shared idiotypes
expressed by human B-cell lymphomas. N Engl J Med 321:851,
1989
14. Levy R, Levy S, Cleary ML, Carroll W, Kon S, Bird J, Sklar J:
Somatic mutation in human B-cell tumors. Immunol Rev 96:43,
1987
15. Greenstein JL, Solomon A, Abraham G N Monoclonal
antibodies reactive with idiotypic and variable-region specific
determinants on human immunoglobulins. Immunology 51:17,
1984
16. Crowley JJ, Mageed RA, Silverman GJ, Chen PP, Kozin F,
Erger RA, Jefferis R, Carson DA: The incidence of a new human
cross-reactive idotype linked to subgroup VHIII heavy chains. Mol
Immunol27:87,1990
17. Crowley JJ, Goldfien RD, Schrohenloher RE, Spiegelberg
HL, Silverman GJ, Mageed RA, Jefferis R, Koopman WJ, Carson
DA, Fong S: Incidence of three cross-reactive idiotypes on human
rheumatoid factor paraproteins. J Immunol140:3411,1988
18. Maruyama S, Kubagawa H, Cooper MD: Activation of
human B cells and inhibition of their terminal differentiation by
monoclonal anti-mu antibodies. J Immunol 135:192,1985
19. Kipps TJ, Fong S, Tomhave E, Chen PP, Goldfien RD,
Carson D A Immunoglobulin V gene utilization in CLL, in Gale
RP, Rai KR (eds): Chronic Lymphocytic Leukemia: Recent
Progress and Future Direction. New York, Liss, 1987, p 115
20. Silverman GJ, Carson DA, Solomon A, Fong S: Human
kappa light chain subgroup analysis with synthetic peptide-induced
antisera. J Immunol Methods 95:249,1986
21. Klobeck HG, Meindl A, Combriato G, Solomon A, Zachau
HG: Human immunoglobulin kappa light chain genes of subgroups
I1 and 111. Nucleic Acids Res 13:6499, 1985
22. Kipps TJ, Herzenberg LA: Schemata for the production of
AXELROD ET AL
monoclonal antibody producing hybridomas, in Weir DM, Herzenberg LA, Blackwell CC (eds): The Handbook of Experimental
Immunology, Edinburgh, Blackwell Scientific, 1986, p 108.1
23. Robbins B A Diagnostic immunohistochemistry of lymphoma and related disorders: Practical aspects of frozen section
technique and interpretation. J Clin Lab Anal 1:104,1987
24. Silverman GJ, Schrohenloher RE, Accavitti MA, Koopman
WJ, Carson D A Structural characterization of the second major
cross-reactive idiotype group of human rheumatoid factors. Arthritis Rheum 33:1347, 1990
25. Kabat E, Wu 'IT, Reid-Miller M, Perry HM, Gottesmann
KS: Sequences of Proteins of Immunological Interest (ed 4).
Bethesda, MD, U.S. Government Printing Office, 1987
26. Berman JE, Mellis SJ, Pollock R, Smith CL, Suh H, Heinke
B, Kowal C, Surti U, Chess L, Cantor CR, Alt FW: Content and
organization of the human Ig VH locus: definition of three new VH
families and linkage to the Ig CH locus. EMBO J 7:727,1988
27. Sanz I, Casali P, Thomas JW, Notkins AL, Capra JD:
Nucleotide sequences of eight human natural autoantibody VH
regions reveals apparent restricted use of VH families. J Immunol
142:4054,1989
28. Casali P, Notkins AL: Probing the human B-cell repertoire
with EBV: Polyreactive antibodies and CDS+B lymphocytes. Annu
Rev Immunol7:513,1989
29. Humphries CG, Shen A, Kuziel WA, Capra JD, Blattner
FR, Tucker PW: A new human immunoglobulin VH family
preferentially rearranged in immature B-cell tumours. Nature
331:446,1988
30. Mayer R, Logtenberg T, Strauchen J, Dimitriu-Bona A,
Mayer L, Mechanic S, Chiorazzi N, Borche L, Dighiero G,
Mannheimer-Lory A, Diamond B, Alt F, Bona C: CDS and
immunoglobulin V gene expression in B-cell lymphomas and
chronic lymphocytic leukemia. Blood 75:1518,1990
31. Borche L, Lim A, Binet JL, Dighiero G: Evidence that
chronic lymphocytic leukemia B lymphocytes are frequently committed to production of natural autoantibodies. Blood 76562,1990
32. Levy R, Levy S, Brown SL, Kon S, Carroll W: Anti-idiotype
antibodies reveal the existence of somatic mutation in human B cell
lymphoma. Monogr Allergy 22:194,1987
33. Miller RA, Maloney DG, Warnke R, Levy R: Treatment of
B-cell lymphoma with monoclonal anti-idiotype antibody. N Engl J
Med 306517,1982
34. Dillman RO, Shawler DL, Sobol RE, Collins HA, Beauregard JC, Wormsley SB, Royston I: Murine monoclonal antibody
therapy in two patients with chronic lymphocytic leukemia. Blood
59:1036,1982
35. Meeker TC, Lowder J, Maloney DG, Miller RA, Thielemans K, Warnke R, Levy R: A clinical trial of anti-idiotype therapy
for B cell malignancy. Blood 65:1349,1985