1. No category

From www.bloodjournal.org by guest on November 24, 2014. For personal use only.
1994 84: 650-656
Erythrocyte membrane proteins reactive with IgG (warm-reacting) antired blood cell autoantibodies: II. Antibodies coprecipitating band 3
and glycophorin A
JP Leddy, SL Wilkinson, GE Kissel, ST Passador, JL Falany and SI Rosenfeld
Updated information and services can be found at:
http://www.bloodjournal.org/content/84/2/650.full.html
Articles on similar topics can be found in the following Blood collections
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml
Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American
Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.
From www.bloodjournal.org by guest on November 24, 2014. For personal use only.
Erythrocyte Membrane Proteins Reactive With IgG (Warm-Reacting)
Anti-Red Blood Cell Autoantibodies: 11. Antibodies Coprecipitating
Band 3 and Glycophorin A
By John P. Leddy, Susan L. Wilkinson, Gregg E. Kissel, Sherry T. Passador, Josie
and Stephen 1. Rosenfeld
L. Falany,
In our initial immunochemical study of the red blood cell
(RBCI membrane proteins targeted in 20 cases of warmantibody autoimmune hemolyticanemia (AHA), RBC eluates
(IP) of both
of 6 patientsmediatedimmunoprecipitation
band 3 and glycophorin A (GPA). This dual IP pattern had
previously been observed with murine monoclonal antibodies (MoAbs) against the highfrequency blood group antigen,
Wrb (Wright), suggesting that the W 8 epitope may depend
on a band 3-GPA interaction. Earlier, anti-We hadbeen identified serologically as aprominent non-Rh specificity of AHA
autoantibodies. In the present study, 6 autoantibody eluates
immunoprecipitating band 3 and GPA from common Wr(b+)
RBCs were retested, in parallel with murine anti-WP MoAbs,
against very rare Wr(a+b-) En(a+) RBCs. One patient's autoantibodies were unreactive with the Wr(b-) RBCs by either
IP or indirect antiglobulin test(IAT) and were judgedt o have
"pure" anti-Wr" specificity. Two other patients' autoantibodies displayed both IP and serologic evidence for anti-Wrb
as a major component in combination with one or more
additional specificities. However, among 3 other patients
whose autoantibodies coprecipitated
band 3 and GPA, there
was no reduction in IP or IAT reactivity with Wr(b-) RBCs
in 2 and only slight reduction in the third. We conclude (1)
that human anti-Wrb autoantibodies, like their murine monoclonal counterparts, coprecipitate band 3 and GPA from human RBCs; but (2)
that not all antibodies with this IP behavior have anti-Wrb serologic specificity, as defined by this
donor's Wr(b-) RBCs. The possibility of an additional (nonWrb) RBC epitope dependent on a band 3-GPA interaction is
raised.
0 1994 b y The American Society of Hematology.
I
availability of very rare human RBCs deficient in the Wrh
antigen, ie, the Wr(a+b-) En(a+) phenotype."."
In a recent study from this laboratory using highly concentrated RBC eluates from 20 AHA patients in radioimmunoprecipitation assays," candidate RBC membrane autoantigens were identified in three major patterns: (1) a polypeptide
of apparent molecular weight (Mr) of 34 kD associated with
a polydisperse 37- to 55-kD glycoprotein, both of which
appeared to be members of the Rh protein family; (2) the
band 3 anion transporter alone; and (3), of particular interest
to the present study, band 3 in association with glycophorin
A (GPA; MN sialoglycoprotein). Coisolation of band 3 and
GPA from radiolabeled human RBCs had previously been
observed with several mouse monoclonal antibodies
(MoAbs) to Wrb or a Wrb-like antigen.23From this and other
evidence it was proposed that the Wrh epitope is dependent
on an interaction of band 3 and GPA in the RBC membrane.''
Itwas logical, therefore, to test for anti-Wrb specificity
among the AHA autoantibodies in our series producing dual
immunoprecipitation of band 3 and GPA. We now report
the results of both immunoprecipitation and serologic assays
withthe group 0 Wr(a+b-) En(a+) RBCs of the wellstudied donor, M.Fr., used in Issitt's classical study.'"
N KEEPING WITH heightened investigative interest in
the potential role of antigen drive in the genesis of many
types of human autoantibodies,"' systematic efforts have
been initiated in many laboratories to characterize relevant
protein autoantigens and their specific autoreactive epitopes
in a growing number of human autoimmune diseases.""
Although great strides have beenmade in the elucidation
of the biochemical nature ofmany important blood group
direct immunochemical studies of the red blood
cell (RBC) membrane proteins reactive with IgG warm-reacting autoantibodies from patients with autoimmune hemolytic anemia (AHA) are less advanced.'4"6 Earlier serologic
studies had provided impressive evidence that proteins of the
Rh complex and a variety of other RBC membrane proteins
carrying non-Rh bloodgroup antigens serve as target autoantigens inIn
the latter group, Issitt et alzohad identified the high frequency blood group antigen, Wl$ (Wright),
as an important autoantigen. Anti-Wrb autoantibodies were
detected as a single specificity or, more commonly, in the
company of RBC autoantibodies possessing other specificities." Demonstration of this specificity depended on the
From theDepartment of Medicine, University of Rochester School
of Medicine and Dentistry, Rochester, NY; and the Hoxworth Blood
Center, University of Cincinnati Medical Center, Cincinnati, OH.
Submitted September 30, 1993; accepted March 21, 1994.
Supported by US Public Health Service Research Grants No. 5ROl-AG-08178and 3-Pol-AI-29522,
and by the David Welk Immunology Research Fund.
Address reprint requests to John P. Leddy, MD, Box 695 (Clinical
Immunology), University of Rochester Medical Center, Rochester.
N Y 14642.
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 1994 by The American Sociery of Hematology.
0006-4971/94/8402-0023$3.00/0
650
MATERIALS AND METHODS
Erythrocytes. Citrated bloodwas obtained in Cincinnati from
the group 0 Wr(a+b-) En(a+) donor for express shipment to Rochester and, concurrently, from a normal group 0 Wrb(+) donor in
Rochester. Positivity of these Wr(b-) RBCs for theEn" antigen
reside,
indicates the presence of GPA, on which En" is known to
and distinguishes these cells from those of the Wr(a-b-) En(a-)
phenotype that are GPA-deficient.'2.2ZThe RBCs of both donors
were washed in 0.15 molL saline and the buffy coats were carefully
removed. A portion of the washed RBCs of both donors was stored
(not more than 1 week) in Alsever's solution at 4°C andusedas
fresh cells in indirect antiglobulin test (IAT) titrations and immunoprecipitation (IP) assays. The remaining RBCs of each donor were
frozen for later use in IP and IAT assays. In preparation for freezing,
0.5-mL aliquots of packed, washed RBCs were mixed with 0.5 mL
Blood, Vol 84, No 2 (July 15), 1994: pp 650-656
From www.bloodjournal.org by guest on November 24, 2014. For personal use only.
BAND 3-GPA COPREClPlTATlONBY AUTOANTIBODIES
40% sucrose and 0.5 mL 22% bovine serum albumin (BSA); these
RBC suspensions were then frozen as microdrops and stored in
liquid N2. Frozen RBCs were reconstituted in prewarmed (45°C)
0.15 mol/L saline and washed three times in room-temperature (RT)
saline. Recovery was excellent for both cell types. Critical results
with fresh or frozen RBCs were in good agreement.
A d - R R C antibodies. The human autoantibodies investigated
were highly concentrated RBC eluates prepared in Rochester as
described previously."."An
additional autoantibody source was a
known anti-Wf serum from anAHA patient studied originally in
Cincinnati (serum 2b); this serum served as a positive control in
serologic assays but was not sufficiently potent for IP analysis. Concentrated control eluates were also prepared from comparable volumes of normal, Coombs-negative RBCs, as described.'" Two murine MoAbs 4-21 and 10-22, that selectively immunoprecipitate band
3 and GPA from human RBCs, were generously supplied by Dr
Pablo Rubinstein (New YorkBlood Center, NewYork. NY) as
unprocessed culture supernatants. These MoAbs had been developed
and extensively characterized serologically by Dr Margaret Nichols
(then at the New York Blood Center, currently at Abbott Laboratories, Abbott Park, IL), and were further characterized by Telen
andChasis.'? The latter report provided evidence that, whereas
MoAb 10-22 behaved as a "pure" anti-Wrh antibody, as judged by
comparative analysis with human anti-Wf alloantibodies, MoAb 421 appeared to recognize a "Wf-like" epitope. The latter epitope
was band-3-dependent butnot totally lacking in GPA-deficient
RBCs ofthe Wr(a-b-) En(a-) phenotype." In our own IP and
IAT assays with 10-22 and 4-21, in which only frozen Wr(a+b-)
RBCs were used because these MoAbs were not available at the
time we received the fresh bloodof donor M.Fr., no differences
were evident between 10-22 and 4-21. Murine MoAb to human GPA
(IOF7MN)" was a culture supernatant of a hybridoma obtained from
American Type Culture Collection (Rockville, MD). Rabbit antiserum to human band 3 (15148) was generously donated by Dr Marguerite M.B. Kay (University of Arizona College of Medicine, Tucson, AZ).
/Pof RRC membrane proteins. These ,methods closely followed
procedures previously described in detail.'" In all studies, Wr(b+)
and Wr(b-) RBCs were handled identically and concurrently; regardless of whichever concentration of antibody source was chosen
for testing (see below), an identical input of that antibody was tested
against each RBC phenotype. Briefly, S0 pL ( I .25 X 1 OK)of surfaceradioiodinated RBCs of each type was incubated on a rotator with
40 to 250 pL (depending on antibody potency) ofAHA eluate,
mouse MoAb, or rabbit antiserum, at37°C for 90 minutes, after
which the RBCs were washedand then lysed by resuspension in
hypotonic phosphate buffer containing I mmol/L phenylmethylsulfonyl fluoride (PMSF) and 10 mmol/L EDTA.'6 RBC ghosts were
pelleted, washed, and solubilized in peroxide-free 2% Triton X-100
(Boehringer-Mannheim, Indianapolis, IN) in cold isotonic buffer
containing PMSF and EDTA. After further centrifugation (35,OOOg
for 30 minutes at 4"C), the immune complex-containing supernatants
were exposed to prewashed, packedprotein G-Sepharose beads
(Zymed, South San Francisco, CA) or goat antihuman IgG-Sepharose beads (Zymed) overnight on a rotator at 4°C. (For precipitates
mediated by murine MoAb, only protein G beads were used.) The
beads were washedsix times,'" suspended inan equal volume of
sample buffer'6 containing 2% sodium dodecyl sulfate (SDS) and
20 mmol/L dithiothreitol (DIT), and placed at 100°C for 2 minutes.
After microcentrifugation, 20 ,uL of each supernatant, together with
radiolabeled molecular mass markers, was resolved in S% to 15%
or 8% to 16% gradient polyacrylamide slab gels in 0.1 % SDS'"
using a discontinuous buffer system."' Completed gels were dried
and analyzed by autoradiography as described.'"
Serologic procedrtres. Comparative reactivity of AHA eluates
651
A
B
C
A
D
B
C
D
20092.56946-
3021.5
B
A
Fig1.
Autoradiographs of SDS-PAGE separation of '%labeled
RBC membrane proteins from group 0 Wr(b+) or Wrlb-) human
RBCs. (A) Immunoprecipitation by mouse MoAb 4-21 (lanes A and
B); whole RBC lysates (no antibody added) (lanes C and D) (l-day
exposure). (B) Immunoprecipitation by AHA eluate GF (lanes A and
B) and by AHA eluate CM (lanes C and D).
or mouse MoAbs with electronically counted, matched 2.5% saline
suspensions of Wr(b+) and Wr(b-) RBCs was assayed by IAT.Z4
EachRBC suspension (0.1 mL volumes) was initially incubated
(37°C for I hour) with 0. I mL of serial twofold dilutions (in saline) of
each primary anti-RBC antibody. After three washes of the sensitized
RBCs, antiglobulin reactions were completed by the addition of
0.1 mL (undiluted) rabbit antihuman lgG Coombs antiserum (Ortho
Diagnostics, Raritan, NJ) or, for murine MoAbs, 0.1 mL 1 / 1 0 0 goat
antimouse IgG (Kirkegaard and Perry, Gaithersburg, MD). After
brief centrifugation, agglutination was read macroscopically. Neither
secondary antiserum agglutinated unsensitized humanRBCs. All
IAT reactions were performed with unmodified RBCs.
Partial absorption ofAHA eluate CMwith Wr(b-) RBCswas
performed by adding 200 pL eluate to an equal volume of packed,
washed RBCs followed by incubation (37°C for I hour) with frequent
mixing. After pelleting the absorbing RBCs, the supernatant was
absorbed with a second 200 p L packed Wr(b-) RBCs. Because of
the limited quantity of cells from M.Fr.. absorption of a second
eluate, ST, was performed three times with 1/10 vol packed Wr(b-)
RBCs by an otherwise similar protocol. Such absorbed eluates were
then tested by both IP and IAT titration as described above.
RESULTS
As shown in Fig 1A (lane A), mouse MoAb 4-21 (or
MoAb 10-22, data not shown) immunoprecipitated from
Wr(b+) RBCs two membrane proteins consistent with band
3 (- 100 kD) and GPA monomer (-4 I kD). However, from
concurrently tested Wr(b-) RBCs, MoAb 4-21 isolates
showed only a faint band 3 image (Fig I A , lane B) that may
be nonspecific (see Leddy et all"). Lanes C and D in Fig
I A , which are autoradiographic patterns of whole RBC lysates from the same experiment, demonstrate that the two
From www.bloodjournal.org by guest on November 24, 2014. For personal use only.
LEDDYETAL
652
Table 1. Serologic Reactions With Wr(b+)
and Wr(b-) Human RBCs
Indirect Antiglobulin Titer (reciprocal)*
Antibody Source
Group 0 Wr (b+)
Mouse MoAb (4-21)
Mouse MoAb (10-22)
AHA serum 2b
AHA eluate GF
AHA eluate SS
AHA eluate CM
AHA eluate CM absorbed
with Wr(b-) RBCs
AHA eluate ST
AHA eluate ST absorbed
with Wrlb-) RBCs
AHA eluate EH
AHA eluate CV
320
640
40
1,280
2,560
6,400
Group 0 Wr (b-)
0
640
6,400
640
2,560
80
2,560
0
0
0
80
640-1,280 (trace)
80
80
640
80
* Each antibody source, in serial twofold dilutions, was incubated
(37°C for 1 hour) with 2.5% suspensions of RBCs of each phenotype.
After washing of the test RBCs, rabbit antihuman IgG (or goat antimouse IgG for the mouse MoAbs) was added and, after brief centrifugation, agglutination was readmacroscopically (see Materials and
Methods for details).
cell types had been comparably radiolabeled and exhibited
the same major bands: aggregated band 3 at greater than 200
kD (see Steck”), band 3 monomer at 100 kD, GPA dimer
just below band 3 monomer, and GPA monomer at -41 kD.
Eachof the human AHA autoantibody eluates included
for study was selected for sharing this property of coprecipitation of band 3 and GPA from common Wr(b+) RBCs. As
demonstrated in a second gel in the same experiment (Fig
IB), AHA eluate GF isolated no discernable bands from
Wr(b-) RBCs (lane A)
but
immunoprecipitated from
Wr(b+) RBCs an intense -100-kD band consistent with
band 3 and a strong -41-kD band consistent withGPA
monomer (lane B). Higher molecular weight (HMW) aggregates (>200 kD)were also evident. Wehad previously
shown by specific immunoblotting that these 100-kD and
41-kD bands were, indeed, band 3 and GPA, respectively.“
Extension of autoradiographic exposure of Fig I B to 3 weeks
still showed no IP in lane A,and this lack of reactivity
of eluate GF with Wr(b-) RBCs was confirmed in other
experiments. Thus, the autoantibodies of patient GF yielded
IP results with Wr(b-) and Wr(b+) RBCsthat are very
similar to those obtained with two mouse anti-Wf MoAbs
(except for the additional presence of HMW aggregates).
Table 1 presents IAT titers of the autoantibody eluates
plus the two anti-Wrh MoAbs and serum 2b, an AHA serum
with previously defined anti-Wf specificity, each assayed
concurrently against Wr(b+) and Wr(b-) RBCs. AHA patient GF, whose eluate mediated no discernible protein isolationfrom Wr(b-) RBCs (Fig IB). resembled the mouse
monoclonals and reference serum 2b in having no detectable
serologic reactivity with Wr(b-) RBCs. The combined IP
and IAT data on GF suggest that this patient’s autoantibodies
possessed “pure” anti-Wrh specificity.
In IP against Wr(b-) RBCs, AHA eluates CM (Fig IB,
lanes C andD)and S S (Fig 2, lanes A and B) exhibited
-
much weaker precipitation of band 3 and either no definite
IP of GPA (eluate S S ) or a very weak IP of GPA (eluate
CM), in contrast to strong IP of both proteins from Wr(b+)
cells by both eluates. Thus, these two eluates appeared to
contain major antibody populations reactive with one or
more epitopes that are lacking or deficient in Wr(b-) RBCs.
AHA eluate S S also exhibited a fourfold reduction in IAT
titer against Wr(b-) compared with Wr(b+) RBCs (Table
l). Together with the IP result (Fig 2), this finding is consistent with a major anti-Wrh specificity in eluate S S , in combination with other autoantibodies of undefined specificity.
Given the quantitative differences in IP between Wr(b+)
and Wr(b-) RBCs obtained with eluate CM (Fig IB), the
equal IAT titers with the two cell types (Table I ) was surprising, but was reproducible. However, absorption of CM eluate
with Wr(b-) RBCs (see Materials and Methods) to reduce
the quantity of autoantibodies reactive with antigens other
than Wrh did show the presence of an autoantibody population withpreferential serologic reactivity with Wr(b+) RBCs
(Table 1) and IP behavior consistent with anti-Wrh(Fig 3A).
In the latter assay, absorbed CM eluate (lane D) retained the
capacity for strong coprecipitation of band 3 and GPA from
Wr(b+) RBCs. In contrast, only a relatively weak isolation
of band 3 was obtained with Wr(b-) cells (Fig 3A. lane C),
and no GPA band was identified, even with an additional 7day autoradiographic development (data not shown). The
remaining volume of eluate S S was insufficient to permit
similar absorption studies.
In Fig 2, mouse MoAb to GPA, rather than to Wrh, was
run asa known marker antibody (lanes E and F), again
illustrating that both cell types had been effectively radiola-
BAND 3
GPA2
A
B
C
D
E
F
Ab SS
SS
ST
ST GPA
GPA
G
-
GPA m
cOnl
Fig 2. SDS-polyacrylamide gel electrophoresis (SDS-PAGE) autoradiograph of RBC membrane proteins immunoprecipitated from
Wr(b+) and Wr(b-) RBCs by AHA eluate SS (lanes A and B): AHA
eluate ST (lanes C and D);mouse anti-GPA MoAb lOF7MN (lanes E
and F); and control eluate from normal RBCs (lane GI.
From www.bloodjournal.org by guest on November 24, 2014. For personal use only.
BAND 3-GPACOPRECIPITATION BY AUTOANTIBODIES
653
beled. The principal higher Mr protein isolated by anti-GPA
MoAb is GPA dimer (GPA,); only a trace of band 3 is visible
just above GPA dimer (lane E).
A fourth AHA eluate, ST, also immunoprecipitated band
3 and GPA, butwithno diminution when Wr(b-) RBCs
were tested (Fig 2, lanes C and D. and Fig 3B. lanes C and
D). Based on multiple studies with this eluate, an artifactual
partial loss of band 3 protein (but notof GPA) occurred
during the IP procedure with normal Wr(b+) cells (lane C,
Fig 2 ) . in part probably because of the formation of HMW
band 3 aggregates. In any event, strong IP of both band 3
and GPA from Wr(b-) RBCs is clearly demonstrated in lane
D of Fig 2. More typical IP patterns with ST eluate are also
shown in Fig 3B (lanes C and D), which confirms equally
strong reactions with Wr(b+) and Wr(b-) RBCs. IAT titers
against the two cell types were also equal (Table l). Absorption of eluate ST with Wr(b-) RBCs yielded a symmetrical
reduction in IP reactivity withboth Wr(b+) and Wr(b-)
RBCs (Fig 3B, lanes A and B), and postabsorption IAT titers
were greatly reduced but remained equal against both cell
types (Table I ) . Patient ST had been observed at this center
for many years and eluates were available from 6 bleeding
dates spanning a 9-year period. As shown in Fig 4,in which
only Wr(b+) RBCs were studied, eluates from only 2 of
these 6 dates (in chronologic sequence, left to right) isolated
strong GPA bands (lanes A and F), despite very heavy immunoprecipitation ofband 3 throughout. (This autoradiograph was deliberately overexposed to show weaker bands.)
Differences in total IgG concentrations in the 6 ST eluates,
given in the legend to Fig 4, didnot correlate with the
presence or absence of a GPA band. These findings suggest
that, in this patient, antibodies to band 3 and those producing
the GPA band varied independently (see Discussion). This
A
B
CA DB
C
D
-.
BAND,
3
GPA-
B
A
unab
unab
RBC Wr
(+)(-1(-1
abs abs
(+)
STSTSTST
(+l
(-1(-1
(+l
Fig 3. Effect of eluate absorption by Wrlb-) RBCs. (A) SDS-PAGE
autoradiograph (60 hours of exposure)after immunoprecipitationby
CM eluate unabsorbed (lanesA and B) and absorbedby Wr(b-) RBCs
(lanes C and D). (B) Autoradiograph after immunoprecipitation by
absorbed (lanes A and B) and unabsorbed ST eluate (lanes C and D).
Details of absorptions are given in Materials and Methods.
A
B
C
D
E
F
""A-
-
BAND 3 t
I
64
50 t
GPA
Fig 4. Overexposed autoradiograph of the membrane proteins
concurrently isolated in the same experiment from a single radiolabeled group 0 Wrlb+) RBC sample bythe RBC eluates of patient ST,
prepared from 6 different bleeding dates, shown in chronologic order
from the left to right (lanesA through F). The total IgG concentrations
(in microgramsper milliliter) by nephelometryin these 6 eluates were
ST1.213; ST2.346; ST3, 153; ST4.270; ST5, 156; and ST6,181.
would be consistent with the evidence in Table 1, Fig 2, and
Fig 3B, all of which were performed with the eluate designated ST1 in Fig 4,that this patient's autoantibodies do not
have anti-Wrh specificity.
The autoantibody eluate of patient CV was much weaker
than that of the others. By extending the autoradiographic
exposure time to 5 days, this eluate was also found to mediate
equal immunoprecipitation of band 3 and GPA fromWr(b-)
and Wr(b+) RBCs (Fig 5A, lanes C and D). The relatively
dense band 3 image (but without GPA) produced by the antiWrh reference MoAb with Wr(b-) RBCs in lane A is atypical for this MoAb and may be attributable, at least in part,
to nonspecific precipitation magnified by the 5-day exposure
(see Fig IA, l-day exposure; and Fig 5B, %day exposure).
A sixth AHA patient, BH, exhibited major IP reactivity
with band 3 (orits breakdown fragments at -64 kD and -50
kD) plus much weaker reactivity with GPA that consistently
appeared to be sharper with Wr(b+) than with Wr(b-) RBCs
(Fig 5B, lanes C and D). However, this difference in IP
results between Wr(b+) and Wr(b-) RBCs was clearly less
striking than with eluates GF, S S , and CM presented above.
IAT titers against the two cell types were equal (Table I ) .
Insufficient Wr(b-) RBCs were available atthispointto
undertake absorption studies.
IP experiments conducted with rabbit anti-band 3 serum
against Wr(b+) RBCs are not shown, and only one of the
multiple studies with mouse anti-GPA MoAb is presented
in Fig 2. These studies demonstrated that '251-labeled GPA
was not detectably coprecipitated with band 3 when specific
anti-band 3 antibody was used. Conversely, when anti-GPA
From www.bloodjournal.org by guest on November 24, 2014. For personal use only.
LEDDY ET AL
654
A
BAND 3
B
CA DB
C
D
-
A
RBC Wrb
B
(-1 (+l (+l (-1
(+l
(-1 (+l (-1
Fig 5. SDS-PAGE autoradiograph of RBC membrane proteins immunoprecipitated by (A)mouse MoAb 4-21 (lanes A and B) and AHA
eluate CV (lanes C and D)(S-day exposure); and (B) mouse MoAb 421 (lanes A and B) and AHA eluate BH (lanes C and D) (3-day exposure). The Mr 64- and 50-kD bands (indicated by arrows in the left
margin) were present only in lanes C and D of (B).
MoAb was tested, typically no band 3 was detected. In overdeveloped autoradiographs, as in Fig 2 (lane E), a trace band
3 image sometimes appeared slightly above the strong GPA
dimer band; however, band 3 images of this very low intensity have also been observed as apparently nonspecific backgroundwhen anti-Rh(D) alloantibodies were tested in IP
against Rh(D)-negative RBCs."This stands in contrast to the
potent and consistent coimmunoprecipitation of both band 3
andGPA either by mouse anti-Wrh or by several of the
human autoantibody eluates in this study.
DISCUSSION
This report is the first to demonstrate thathuman antiRBC autoantibodies that exhibit serologic evidence for antiWrh specificity mimic their murine MoAb counterparts'3 in
coprecipitating band 3 and GPA from radiolabeled human
RBCs. This was most clearly observed with the autoantibodies of patients GF and S S , for which both IP and serologic
studies of unabsorbed eluates supported anti-Wrh specificity
as a major component. A significant anti-Wrh specificity was
also demonstrated in a third AHA patient, CM, after absorption of the eluate with Wr(b-) RBCs.
Although there is biophysical evidence for the existence
of band 3-GPA complexes in the human RBC membrane**
and evidence for facilitated membrane expression of band 3
in the presence of GPA,29 we do not believe that either
proteinis a passive participant in their coprecipitation by
these anti-Wrh antibodies. This interpretation is based onour
parallel observation that monospecific anti-band 3 or antiGPA antibodies each precipitate their respective target proteins without significant coprecipitation of the other RBC
protein, confirming a similar finding by Telen and Chasis.'3
We have also observed that several other AHA eluates that
strongly immunoprecipitate band 3 do not detectably coisolate GPA.'" These IP findings suggest, therefore, ( I ) that,
whatever the normal association is between band3 and GPA,
their affinity for one another is not sufficientforthetwo
proteins to remainbonded during theimmunoadsorption
phase of immunoprecipitation, except in the presence of antibody of a particular specificity; and, conversely, (2) that the
consistent coprecipitation of these two proteins by murine
or human anti-Wrhor anti-"Wrh-like" antibodies is specific.
This concept is consistent with the proposal that anti-Wrh
antibodies have affinity for one or more epitopes resulting
from a band 3-GPA interaction in the RBCmembrane"
rather than for an epitope(s) dependent on GPA alone, even
though a GPA peptide sequence positioned just exterior
to the membrane appears tobe critical to the antigenic
site. 1?.3n-x Anti-Wrh antibody may stabilize a band3-GPA
interactive epitope thatis otherwise transient or unstable.
Remarkable as it would be for mouse and human antibodies
to recognize the very same epitope that is genetically deficient or defective in the Wr(a+b-) RBCs of donor M.Fr.,
the plausible number of epitopes that could fulfill these conditions must be small.
Our overall findings, however, on the 6 AHA patients
whose autoantibodies coprecipitate band 3 and GPA clearly
indicate additional complexity. In 2 of the 3 patients who
appeared to have a major anti-Wrhcomponent ( S S and CM),
the presence of one or more additional autoantibody specificities was evident by both IP and IAT assays. This was
also true in the large serologic study of Issitt et al."' The
findings with eluate CM suggested that IP and serologic
methods differed in their ability to show the anti-Wrhspecificity. That is, IATtitration of unabsorbed CMeluate showed
no difference in reactivity with Wr(b+) and Wr(b-) cells
(Table 1) despite quite impressive differences by IP (Fig
IB). Similarly, with eluate S S , the IAT titer with Wr(b-)
RBCs (Table 1) was higher than expected, given the striking
difference in IP patterns with the two cells (Fig 2). Probably
the simplest explanation would be the failure to radiolabel
certain RBC autoantigens reactive with the additional (nonWrh)antibody specificities in CM and S S eluates. An alternative explanation also deserves consideration. Our experience
with IP has turned up a number of instances in which the
IAT titer of an anti-RBC antibody of known specificity, eg,
allo-anti-Rh(C) or anti-Rh(E), and its effectiveness in IP with
the same donor's RBCs have not correlated well, leading us
to suspect that IP is dependent onrelativelyhigh-affinity
antibodies (capable of remaining bound to antigen through
the immunoadsorption procedure), whereas IAT may be responsive to both low- and high-affinity antibodies. By this
hypothesis, patients CM and S S may have sufficient antibodies of lowaffinityto
mediate higher IAT titers against
Wr(b-) RBCs than predicted by IP results.
A second area of complexity arises from our observation
that some human autoantibodies that share with more typical
anti-Wrh antibodies the ability to coprecipitate band 3 and
GPA, eg, eluates ST and CV, apparently do not have antiWrh specificity, at least as defined by the Wr(b-) RBCs of
From www.bloodjournal.org by guest on November 24, 2014. For personal use only.
BAND 3-GPA COPREClPlTATlON BY AUTOANTIBODIES
donor M.Fr. The autoantibodies of ST and CV gave no hint
of weaker reactivity with Wr(b-) cells by either assay
method, and absorption of ST eluate with Wr(b-) RBCs
produced a symmetrical loss of IP and IAT reactivity with
both cell types. Eluate BH consistently produced a slightly
weaker IP pattern with Wr(b-) RBCs, but retained clearly
evident IP and full IAT reactivity against these cells. AntiWrb specificity in eluate BH, if it is present at all, may be
confined to a minor subpopulation that is strongly overshadowed by other specificities. Unfortunately, there were insufficient Wr(b-) RBCs for absorptions of eluates CV and BH
to permit a more complete assessment of minor anti-Wrb
antibody populations. However, in the case of ST, the absorption studies clearly argue against anti-Wrb specificity.
Moreover, we believe that the serial IP results in Fig 4 favor
independent antibody populations reactive with band 3 and
with either GPAitself (anti-Ena?) orwith a band 3/GPA
interactive epitope that is preserved in the RBCs of M.Fr.
(see below), and that the relative abundance of the GPAreactive antibody subpopulation in the ST eluates apparently
shifts over time. In any event, if the RBCs of donor M.Fr.
are accepted as the current “gold standard” to define antiWrb specificity, our observations provide a caveat that antiWrb specificity cannot necessarily be inferred from the IP
pattern with common Wr(b+) erythrocytes.
The possibility that there is a distinct, non-Wrb epitope
resulting from a band 3-GPA interaction deserves consideration. By this hypothesis, the Wr(a+b-) RBCs of M.Fr.
would be deficient in only one of the putative band 3/GPAdependent epitopes. However, the autoimmune response
would make this distinction inconsistently, resulting in some
instances (eg, patient GF) in typical anti-W? specificity, in
other instances in specificity for a (putative) non-Wf band
3/GPA interactive epitope (possibly patients BH and CV),
and in still other instances in a mixture of antibodies reactive,
with differing affinities, against W f and non-Wrb epitopes
that arise from band 3/GPA interaction (patients CM and
SS). In this connection it would be interesting to determine
the IP behavior of the human alloantibody, anti-En”FR,
which is thought to be closely related to anti-Wrb but is
known to react normally with the RBCs of M.Fr.’*,3’Could
En‘FRbe a band 3/GPA-interactive epitope distinct from
Wrb? However, wemust acknowledge that the additional
(non-W?) specificities in the autoantibody eluates under
study could be reactive with totally unrelated antigens.”
ACKNOWLEDGMENT
We thank Lynn Kosarko for skillful preparation of the manuscript.
Generous gifts of antisera from Dr Pablo Rubinstein (New York
Blood Center) and from Dr Marguerite M.B.Kay (University of
Arizona) are acknowledged with gratitude.
REFERENCES
1. Capra JD, Natvig JB: Is there V region restriction in autoimmune disease? Immunologist l:l6, 1993
2. Craft JE, Hardin JA: Linked sets of antinuclear antibodies:
What do they mean? J Rheumatol 14:106, 1987
3. Bini P, Chu J-L, Okolo C, Elkon K Analysis of autoantibodies
to recombinant La(SS-B) peptides in systemic lupus erythematosus
and primary Sjogren’s syndrome. J Clin Invest 85325, 1990
655
4. St Clair EW, Burch JA Jr, Ward MM, Keene JD, Pisetsky DS:
Temporal correlation of antibody responses to different epitopes of
the human La autoantigen. J Clin Invest 85:515, 1990
5. Miller F W , Twitty SA, Biswas T, Plotz PH: Origin and regulation of a disease-specific autoantibody response. Antigenic epitopes,
spectrotype stability, and isotype restriction of anti-Jo-l autoantibodies. J Clin Invest 85:468, 1990
6. Kekomaki R, Dawson B, McFarland J, Kunicki TJ: Localization of human platelet autoantigens to the cysteine-rich region of
glycoprotein IIIa. J Clin Invest 88347, 1991
7. Kaufman DL, Erlander MD, Clare-Salzler M, Atkinson MA,
Maclaren NK, Tobin AJ: Autoimmunity to two forms of glutamate
decarboxylase in insulin-dependent diabetes mellitus. J Clin Invest
89:283, 1992
8. AmagaiM,
Klaus-Kortun V, Stanley JR: Autoantibodies
against a novel epithelial cadherin in pemphigus vulgaris, a disease
of cell adhesion. Cell 67:869, 1991
9. Finke R, Set0 P, Ruf J, Carayon P, Rapoport B: Determination
at the molecular level of a B-cell epitope on thyroid peroxidase
likely tobe associated with autoimmune thyroid disease. J Clin
Endocrinol Metab 73:919, 1991
IO. KoshakaH, Yamamoto K, Fujii H, Miura H, Miyasaka N.
Nishioka K, Miyamoto T: Fine epitope mapping of the human SSB L a protein. Identification of a distinct autoepitope homologous to
a viral Gag polyprotein. J Clin Invest 85:1566, 1990
11. Manfredi AA, Protti MP, Bellone M, Moiola L, Conti-Tronconi BM: Molecular anatomy of an autoantigen: T and B epitopes
on the nicotinic acetylcholine receptor in myasthenia gravis. J Lab
Clin Med 120:13, 1992
12. Anstee DJ: Blood group-reactive surface molecules ofthe
human red blood cell. Vox Sang 58:1, 1990
13. Agre P, Cartron J-P: Molecular biology of the Rh antigens.
Blood 78:551, 1991
14. Victoria EJ, Pierce SW, Branks MJ, Masouredis SP: IgG red
blood cell autoantibodies in autoimmune hemolytic anemia bind to
epitopes onred blood cell membrane band 3 glycoprotein. J Lab
Clin Med 115:74, 1990
15. Barker R N , Casswell KM, Reid ME, Sokol W,ElsonCJ:
Identification of autoantigens in autoimmune haemolytic anaemia
by a non-radioisotope immunoprecipitation method. Br J Haematol
82: 126, 1992
16. Leddy JP, Falany JL, Kissel GE, Passador ST, Rosenfeld
SI: Erythrocyte membrane proteins reactive with human (warmreacting) anti-red cell autoantibodies. J Clin Invest 91:1672, 1993
17. Petz JD, Garratty G: Acquired Immune Hemolytic Anemias.
New York, NY, Churchill Livingstone, 1980, p 232
18. Issitt PD: Applied Blood Group Serology (ed 3). Miami, F L ,
Montgomery Scientific, 1985, p 523
19. Dacie JV: The Haemolytic Anaemias, v01 3: The Autoimmune
Haemolytic Anaemias (ed 3). New York, NY, Churchill Livingstone,
1992
20. Issitt PD, Pavone BG, Goldfinger D, Zwicker H, Issitt CH,
Tessel JA, Kroovand SW, Bell CA: Anti-W?, andother autoantibodies responsible for positive direct antiglobulin tests in 150 individuals. Br J Haematol 34:5, 1976
21. Adams J, Broviac M, Brooks W, Johnson NR, Issitt PD: An
antibody, in the serum of Wr(a+) individual, reacting with an antigen
of very high frequency. Transfusion 11:290, 1971
22. Issitt PD, Pavone BG, Wagstaff W, Goldfinger D: The phenotypes En(a-), Wr(a-b-), and En(a+), Wr(a+b-), and further studies
on the Wr and En blood group systems. Transfusion 16:3%, 1976
23. Telen MJ, Chasis JA: Relationship of the human erythrocyte
Wrb antigen to an interaction between glycophorin A and band 3.
Blood 76:842, 1990
24. Leddy JP, Peterson P, Yeaw MA, Bakemeier RF: Patterns
From www.bloodjournal.org by guest on November 24, 2014. For personal use only.
656
of serologic specificity of human yG erythrocyte autoantibodies. J
Immunol 105:677, 1970
25. Bigbee WL, Vanderlaan M, Fong SSN, Jensen RH: Monoclonal antibodies specific for the M- and N- forms of human glycophorin A. Mol Immunol 20:1353, 1983
26. Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680, 1970
27. Steck TL: The band 3 protein of the human red cell membrane: A review. J Supramol Struct 8:311, 1978
28. Nigg EA, Bron C, Girardet M, Cherry RJ: Band 3-glycophorin
A association in erythrocyte membranes demonstrated by combining
protein diffusion measurements with antibody-induced cross-linking.
Biochemistry 19: 1887, 1980
29. Groves JD, Tanner MJA: Glycophorin A facilitates the ex-
LEDDY ET AL
pression of human band 3-mediated anion transport in Xenopus 00cytes. J Biol Chem 267:22163, 1992
30. Ridgwell K, Tanner MJA, Anstee DJ: The Wrh antigen, a
receptor for Plasmodium falciparum malaria, is located on a helical
region of the major membrane sialoglycoprotein of human red blood
cells. Biochem J 209:273, 1983
3 1 . Dahr W, Wilkinson S, Issitt PD, Beyreuther K, Hummel M,
Morel P: High frequency antigens of human erythrocyte membrane
sialoglycoproteins. 111. Studies on the En"FR, Wrh and WS antigens.
Biol Chem Hoppe Seyler 367:1033, 1986
32. Reardon A: Heterogeneity inthe specificity of Wrh monoclonal antibodies, in Rouger PG, Salmon C (eds): Monoclonal Antibodies Against Red Blood Cell and Related Antigens. Paris, France,
Arnette, 1987, p 261