Download Wall in Acute Kawasaki Disease Oligoclonal IgA Response in the

Oligoclonal IgA Response in the Vascular
Wall in Acute Kawasaki Disease
This information is current as
of June 15, 2017.
Subscription
Permissions
Email Alerts
J Immunol 2001; 166:1334-1343; ;
doi: 10.4049/jimmunol.166.2.1334
http://www.jimmunol.org/content/166/2/1334
This article cites 30 articles, 12 of which you can access for free at:
http://www.jimmunol.org/content/166/2/1334.full#ref-list-1
Information about subscribing to The Journal of Immunology is online at:
http://jimmunol.org/subscription
Submit copyright permission requests at:
http://www.aai.org/About/Publications/JI/copyright.html
Receive free email-alerts when new articles cite this article. Sign up at:
http://jimmunol.org/alerts
The Journal of Immunology is published twice each month by
The American Association of Immunologists, Inc.,
1451 Rockville Pike, Suite 650, Rockville, MD 20852
Copyright © 2001 by The American Association of
Immunologists All rights reserved.
Print ISSN: 0022-1767 Online ISSN: 1550-6606.
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
References
Anne H. Rowley, Stanford T. Shulman, Benjamin T. Spike,
Carrie A. Mask and Susan C. Baker
Oligoclonal IgA Response in the Vascular Wall in Acute
Kawasaki Disease1
Anne H. Rowley,2*†‡ Stanford T. Shulman,*‡ Benjamin T. Spike,*†, Carrie A. Mask,*† and
Susan C. Baker§
K
awasaki Disease (KD)3 is an acute vasculitis of young
childhood which may be fatal. KD is the leading cause
of acquired heart disease in children in developed nations (1), but little is known about the pathogenesis of the disorder.
The clinical features of KD include prolonged high fever, rash,
conjunctival injection, erythema of the oral mucosa, swelling and
redness of the hands and feet, and cervical adenopathy. These
acute features generally resolve within 1–3 wk after the onset of
illness; however, ⬃20% of untreated patients develop coronary
artery aneurysms (1). The majority of KD patients who are treated
within the first 10 days of illness with i.v. gammaglobulin and
aspirin show a rapid resolution of fever and inflammatory signs,
and this treatment reduces the prevalence of coronary artery abnormalities to ⬃5% (2). The mechanism of action of i.v. gammaglobulin in KD is unclear. The symptoms of KD resolve over
several weeks, even in untreated patients, and recurrences are unusual, distinguishing the illness from autoimmune vasculitic disorders. The sequelae of undiagnosed KD during childhood appears
to account for cases of acute myocardial infarction or sudden death
in young adults (3). Because the etiology of KD is unknown, no
Departments of *Pediatrics and †Microbiology and Immunology, Northwestern University Medical School, Chicago, IL 60611; ‡Children’s Memorial Hospital, Chicago,
IL 60614; and §Department of Microbiology and Immunology, Loyola University
Stritch School of Medicine, Maywood, IL 60153
Received for publication March 27, 2000. Accepted for publication October 16, 2000.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact. The costs of publication of
this article were defrayed in part by the payment of page charges. This article must
therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.
1
This work was supported by National Institutes of Health Grants K08HL03270 and
R01HL63771, the Charles E. Culpeper Foundation, the American Heart Association
of Metropolitan Chicago, and the Kawasaki Disease Research Fund of the Children’s
Memorial Hospital.
2
Address correspondence to Dr. Anne H. Rowley, Pediatrics W140, Ward 12-140,
Northwestern University Medical School, 303 East Chicago Avenue, Chicago, IL
60611. E-mail address: [email protected]
3
Abbreviation used in this paper: KD, Kawasaki Disease.
Copyright © 2001 by The American Association of Immunologists
diagnostic test is available, specific therapy cannot be developed,
and prevention is not feasible.
Clinical and epidemiologic features of KD strongly suggest an
infectious etiology (1). These include the young age group affected, the clinical features of the illness, the occurrence of epidemics with periodicity, and the geographic wave-like spread of
illness during epidemics (1). To date, traditional methods to identify a microbial agent have failed to clarify the etiology of KD.
Recently, there has been interest in a superantigen etiology of KD
based upon possible selective expansion of V␤2 and V␤8 T cell
receptor families in peripheral blood in acute KD (4, 5); however,
other investigators have been unable to confirm this finding (6 –
10). Toxic shock syndrome toxin-1 or streptococcal superantigens
were proposed to be related etiologically to KD based upon a single study (11) that has not been confirmed by others (7, 8, 12–14).
We have reported the novel finding that IgA plasma cells infiltrate the vascular wall during acute KD (15). In this study, we
examined the clonality of the IgA response in KD vascular tissue
by sequencing the CDR3 regions of ␣ genes isolated from vascular
tissues from three fatal acute KD patients. We report that the IgA
produced in acute KD vascular tissue is oligoclonal, consistent
with an Ag-driven immune response.
Materials and Methods
Analysis of ␣ genes from a KD vascular cDNA library
KD vascular cDNA library. Vascular tissue (750 mg) from the aorto-iliac
junction was used to isolate total RNA by the guanidium thiocyanate-acid
phenol method (16), and an aliquot of the total RNA was used to isolate
poly A⫹ RNA using Dynabeads oligo dT (Dynal, Oslo, Norway). The
tissue was obtained from patient A, a 10-year-old Caucasian male with
acute KD who died from a ruptured coronary artery aneurysm on day 13 of
illness (15). An oligo dT-primed, directional cDNA library in the phage
expression vector ␭ZAP was synthesized using a cDNA synthesis kit
(ZAP-cDNA synthesis kit; Stratagene, La Jolla, CA) as previously reported
(15). The primary, unamplified cDNA library was used to isolate ␣ clones
for sequencing of the VDJ junctions.
Isolation of ␣ clones. ␣ clones were identified from plates of the primary
library by hybridization with an ␣ probe. Initial filter lifts were hybridized
with a VH-C␣ probe generated by RT-PCR of control spleen RNA as
0022-1767/01/$02.00
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
Kawasaki Disease (KD) is a potentially fatal acute vasculitis of childhood. Although KD is the leading cause of acquired heart
disease in children in developed nations, its pathogenesis remains unknown. We previously reported the novel observation that IgA
plasma cells infiltrate the vascular wall in acute KD. We have now examined the clonality of this IgA response in vascular tissue
from three fatal cases of KD to determine whether it is oligoclonal, suggesting an Ag-driven process, or polyclonal, suggesting
nonspecific B cell activation or a response to a superantigen. We first sequenced VDJ junctions of 44 ␣ genes isolated from a
primary, unamplified KD vascular cDNA library. Five sets of clonally related ␣ sequences were identified, comprising 34% (15
of 44) of the isolated ␣ sequences. Furthermore, point mutations consistent with somatic mutation were detected in the related
sequences. Next, using formalin-fixed coronary arteries from two additional fatal KD cases, we sequenced VDJ junctions of ␣
genes isolated by RT-PCR, and a restricted pattern of CDR3 usage was observed in both. We conclude that the vascular IgA
response in acute KD is oligoclonal. The identification of an oligoclonal IgA response in KD strongly suggests that the immune
response to this important childhood illness is Ag-driven. The Journal of Immunology, 2001, 166: 1334 –1343.
The Journal of Immunology
1335
Table I. Probes to identify ␣ variable region size and primers used for sequencing and to perform RT-PCR on paraffin-embedded, formalin-fixed KD
tissues
Primer
Oligonucleotide Sequence (5⬘33⬘)
Description
C␣
SK
C␣S
AR1
VH1
VH2
VH3
VH4
VH5
VH6
VH3FR3
VH4FR3
JH1,2,4,5
JH3
JH6
CTGGGATTCGTGTAGTGCTT
CGCTCTAGAACTAGTGGATC
AGGCGATGACCACGTTCCCAT
CTCCCACCTGCTGCTGGACG
CCATGGACTGGACCTGG
ATGGACATACTTTGTTCCAC
CCATGGAGTTTGGGCTGAGC
ATGAAACACCTGTGGTTCTT
ATGGGGTCAACCGCCATCCT
ATGTCTGTCTCCTTCCTCAT
MGATTCACCATCTCMAGRG
CGAGTCACCATRTCMGTAGAC
TGAGGAGACGGTGACCAGGGT
TACCTGAAGAGACGGTGACC
ACCTGAGGAGACGGTGACC
constant alpha primera
pBluescript SK primer
constant alpha sequencing primer
VH4-group 1 CDR3 probe
VH1 and VH7 leader primersb
VH2 leader primerb
VH3 leader primerb
VH4 leader primerb
VH5 leader primerb
VH6 leader primerb
VH3 framework 3 primer
VH4 framework 3 primer
JH1,2,4,5 primerb
JH3 primerb
JH6 primerb
a
b
Ref. 17.
Ref. 18.
Table II. Distribution of ␣ genes in KD cDNA library by VH family
VH
Family
VH1
VH2
VH3
VH4
VH5
VH6
VH7
a
No. of
Clones
Percentage of
Total
Percent Detected in Adult
Peripheral Blood B Cellsa
3
0
17
14
2
2
4
7
0
40
33
5
5
10b
13–19
2–4
54–56
20–22
4–6
1–2
0–2
Refs. 18 and 22.
Significantly different from adult control ( p ⫽ 0.05) as determined by Fisher’s
exact test.
b
(titer, 1 ⫻ 1010/ml) was subjected to PCR amplification using a VH4 leader
primer and a primer in the constant region of ␣ (Table I). Amplified product was cloned into the pGem-T vector (Promega, Madison, WI), and individual clones were subjected to colony hybridization with the oligonucleotide AR1 probe (Table I) using standard methods (19). Hybridizing
clones were sequenced as above to confirm that they contained the VH4group 1 CDR3 sequence. The variable region of the clones was assigned to
a germline sequence by comparing the sequence data with published human germline sequences. These clones are not included in the list of ␣
clones identified as they were not isolated from the primary library.
Analysis of ␣ genes from formalin-fixed, paraffin-embedded
coronary artery blocks
Tissue blocks. Coronary artery tissue blocks were obtained from patients
B (11-month-old Caucasian male) and C (4-mo-old Caucasian female), two
fatal KD cases in which death occurred within 2 mo of onset of the illness.
Patient B was not treated with i.v. gammaglobulin because the diagnosis of
KD was not made prior to death. Patient C received i.v. gammaglobulin
therapy. Ten 8-␮m sections from each block were processed for RNA
isolation. As a control source for the amplification of a polyclonal population of ␣ genes, ten 8-␮m sections of paraffin-embedded, formalin-fixed
spleen tissue from patient A were also obtained for RNA processing.
RNA isolation. Ten 8-␮m sections were processed for RNA isolation by
first deparaffinizing with xylene, and washing the tissue with 95% ethanol.
The tissue sections were resuspended in digestion buffer (1% SDS, 0.1 M
Tris-HC1, pH 7.3, 25 ␮M EDTA, and 1 mg/ml proteinase K) and incubated
at 50°C for 12 h with agitation. After phenol-chloroform extraction and
ethanol precipitation, the nucleic acid pellet was resuspended in solution D
(4 M guanidinium thiocyanate, 25 mM sodium citrate pH 7.0, 0.5% Nlauryl sarcosine, 0.1 M mercaptoethanol) and incubated at room temperature for 24 h. After phenol-chloroform extraction and ethanol precipitation,
the pellet was treated with 20 U of DNase I and incubated at 37°C for 30
min, followed by repeat extraction, precipitation, and resuspension of the
pellet in water for use in RT-PCR.
RT-PCR. One-tenth volume of the isolated RNA was incubated with 120
pm random hexamers (Amersham Pharmacia, Piscataway, NJ) for 10 min
at 70°C, and first-strand buffer (Life Technologies), 500 ␮M dNTP, 10 mM
DTT, 27 U RNAguard (Pharmacia), and 200 U Superscript II (Life Technologies) were added. The reaction was incubated at room temperature for
10 min, followed by 1 h at 42°C. PCR was performed in a standard reaction
(19) using a primer in the constant region of ␣ and a primer in the third
framework region of VH3 or VH4 (Table I) for 40 cycles of denaturation at
94°C for 1 min, primer annealing at 58°C for 1 min, and extension at 72°C
for 2 min. A semi-nested PCR amplification followed, using the third
framework primer and three J region primers (Table I) for 40 additional
cycles using the same cycling parameters. Pfu Turbo (Stratagene) was used
to generate PCR products instead of Taq polymerase, to minimize copying
errors. After performing runoff reactions as described below, PCR products
were cloned into PCR Script (Stratagene) according to the manufacturer’s
instructions, and clones were sequenced as above.
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
previously described (15), in which products of six separate PCR reactions
using primers from the six leader sequences of the human heavy chain
variable regions with a primer from the constant region of ␣ (17, 18) were
mixed (Table I). Only five of 11 clones initially identified using this probe
contained variable region. Because use of this probe did not obviate the
isolation of clones containing the constant region of ␣ only, we incubated
additional filter lifts of primary library with an ␣ constant region probe
derived from a plasmid preparation of a cDNA library clone. PCR amplification of isolated plaques was then performed using a primer in the
pBluescript vector (SK; Table I) and a primer in the 5⬘ constant region of
␣ (C␣; Table I), and clones that yielded a product large enough to contain
variable region (⬎300 bp) were sequenced. Both probes were labeled with
[32P]dCTP using a random priming method (Random Primers DNA Labeling System; Life Technologies, Gaithersburg, MD). Positive clones
were plaque purified using standard methods (19). The majority (twothirds) of the phage clones were excised using ExAssist helper phage
(Stratagene) to yield pBluescript phagemid for sequencing; the remaining
one-third were PCR amplified and directly sequenced as described below.
DNA sequencing. Excised pBluescript phagemid obtained from twothirds of the clones was sequenced by standard dideoxynucleotide chain
termination methods using a kit (Sequenase version 2.0; U.S. Biochemical,
Cleveland, OH). The rest of the clones were sequenced using the PCR
product generated as above, using a kit (T7 sequenase PCR product sequencing kit; Amersham, Arlington Heights, IL). A primer in the 5⬘ constant region of ␣ was used to sequence the VDJ junctions (primer C␣S;
Table I).
Isolation of additional full-length VH4-group 1 clones. Additional fulllength clones containing the CDR3 sequence of the VH4-group 1 clones
were isolated from the amplified vascular cDNA library, to determine
whether other related clones demonstrating somatic mutation were present
in the library. The oligonucleotide AR1 corresponding to this CDR3 sequence (Table I) was synthesized and labeled with [␥-32P]dATP (Amersham) using T4 polynucleotide kinase (Life Technologies, Gaithersburg,
MD) by standard methods (19). One microliter of amplified cDNA library
1336
OLIGOCLONAL IgA RESPONSE IN ACUTE KD
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
FIGURE 1. Clonally related ␣ sequences from vascular cDNA library of KD patient A, showing evidence of somatic mutation. A, Nucleotide sequences of the five
clones that are members of the KD patient A VH4-group 1 set of sequences, and of the related germline sequence VH4.18 (21). Clone E2 is full-length. The other four clones
(5-1, 7-1a, A2, and C5v) are not full-length; the length of each is indicated in the figure. B, Amino acid sequences of the KD VH4-group 1 clones, showing amino acid
substitutions indicative of somatic mutation within the group. C, Nucleotide sequences of the two clones that are members of the KD patient A VH4-group 2 set of sequences
and of the related germline sequence DP-65 (21). Clone C2 is full-length. Clone 4–5 is not full-length; the length of the clone is indicated in the figure. D, Amino acid
sequences of the VH4-group 2 clones, showing amino acid substitutions indicative of somatic mutation within the group.
Runoff reactions. To obtain a rapid overall assessment of the distribution of CDR3 sizes in the population of ␣ genes amplified from the
tissues, we adapted a method previously reported for determining the
CDR3 sizes in a population of T cell receptor genes (20). This method uses
a fluorescent dye-labeled primer for a single cycle of primer annealing and
extension of the PCR products previously generated as above. Products of
The Journal of Immunology
1337
FIGURE 2. Nucleotide sequences of the
CDR3 regions of VH3-group 1 (clones D3,
11-2, and 11-5), VH3-group 2 (clones 2-2 and
13-1), and VH7-group 1 (clones 10-1, 7-1b,
and 10-4) from patient A, showing identical
sequences within each group.
Results
To determine if the IgA response in the KD vascular wall was
polyclonal or oligoclonal, we isolated ␣ gene sequences from vascular tissues from three fatal KD cases. First, we isolated ␣ sequences from a primary, unamplified cDNA library derived from
nonfixed KD vascular tissue (15). Because additional unfixed KD
vascular tissues were not available, we isolated ␣ sequences by
RT-PCR from formalin-fixed, paraffin-embedded vascular tissues
from two additional fatal KD cases.
Analysis of ␣ gene sequences from the primary, unamplified KD
vascular cDNA library
The primary library titer was 2.5 ⫻ 106 plaques, with ⬎90% recombinants and 0.1% ␤-actin clones. The library has at least 0.1%
immunoglobulin clones detected by immunoscreening with 125Ilabeled anti-human Ig (Amersham) as previously reported (15).
The dimensions of the tissue piece were approximately 5 ⫻ 3 ⫻ 30
mm (450 mm3). The mean number of IgA plasma cells in a 10
mm ⫻ 3 mm ⫻ 0.005 mm (0.15 mm3) piece of this tissue was 300,
as determined by calculating the average of IgA-positive cells in
eight separate FITC-stained sections of the tissue. Thus, a
0.005-mm tissue section contained ⬃2000 IgA plasma cells/mm3.
However, because the maximal diameter of a large plasma cell
is ⬃0.025 mm, the same plasma cell might appear in up to five
successive 0.005-mm sections. Therefore, the number of IgA
plasma cells in the tissue more closely approximated 400 per mm3
(2000 ⫼ 5). Thus, in the original 450-mm3 piece of tissue, there
would be ⬃180,000 IgA plasma cells (400 cells/mm3 ⫻ 450 mm3).
We isolated 88 ␣ clones from the primary, unamplified library
by hybridization with an ␣ probe as described in Materials and
Methods. Forty-four of these 88 clones included the CDR3 region
and were sequenced. Ten clones were sequenced twice, using excised phagemid and PCR, to ensure that Taq polymerase error did
not introduce random mutations into the products; identical CDR3
sequence was obtained by both methods in all cases.
Distribution of VH␣ genes in the KD vascular cDNA library by
family. To determine whether expansion or deletion of a heavy
chain variable region (VH) family occurred in the IgA genes in the
KD tissue, isolated clones were grouped by VH family by examination of the CDR1 and 2 regions and the framework regions and
comparing them to published VH family sequences (21). Of the 44
␣ clones, 42 were members of the VH1, VH3, VH4, VH5, VH6, and
VH7 families as shown in Table II. The distribution of VH family
members in the ␣ genes in the KD cDNA library was similar to
that reported in the literature for adult peripheral blood B cells (18,
22). More VH7 clones were present in the KD vascular cDNA
library than would be expected by comparison to published data
for adult controls (10 vs 2%, p ⫽ 0.05 by Fisher’s exact test);
however, VH7 has been reported to be frequently used in the infant
cord blood repertoire (23), and data for its use in childhood peripheral blood B cells are lacking. The remaining two ␣ clones did
not contain sufficient framework 3 region to allow assignment to a
family. Overall, the distribution of VH families in KD vascular
tissue appears to be similar to that found in normal peripheral
blood B cells.
Identification of clonally related sequences. To determine
whether ␣ genes in the vascular cDNA library were polyclonal or
oligoclonal, we sequenced the CDR3 regions. Sequence identity
between B cells in this region indicates clonal relatedness (24, 25);
expansion of B cells with the same VDJ rearrangements and with
evidence of somatic mutation are characteristic features of an immune response to Ag (24, 25).
Sequence analysis of the 44 ␣ clones revealed two groups of
clonally related VH4 sequences (Fig. 1). The VH4-group 1 sequences contain five members (clones C5v, E2, 5-1, 7-1a, and A2)
and the VH4-group 2 sequences contain two members (clones C2
and 4–5). As can be seen in Fig. 1A, the VH4-group 1 sequences
show evidence of somatic mutation compared to germline sequence VH4.18, and clone A2 shows evidence of somatic mutation
compared to clones E2, 5-1, 7-1a, and C5v. Amino acid sequences
for this group are seen in Fig. 1B. Similar evidence of somatic
mutation is seen in the VH4-group 2 sequences (Fig. 1C). Clones
C2 and 4–5 show somatic mutation compared to germline DP-65
and to each other. Fig. 1D shows amino acid sequences for this
second group. The identification of multiple clones with identical
VDJ rearrangements and evidence of somatic mutation strongly
suggests an oligoclonal, Ag-driven immune response in acute KD.
Three additional sets of identical sequences were identified in
the KD vascular library (Fig. 2). The VH3-group 1 has three members (clones D3, 11-2, and 11-5), the VH3-group 2 has two members (clones 2-2 and 13-1), and the VH7-group 1 contains three
members (clones 10-1, 7-1b, and 10-4). The members of each
group were identical to each other in the regions sequenced. Therefore, these clones likely represent sequences isolated from plasma
cells that are highly represented in the vascular tissue and were
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
the runoff reaction were analyzed on a DNA sequencer (PE Biosystems
377XL, ABI Prism, Genescan 3.0, run mode 36C-2400, well-to-read distance 36 cm). The runoff products migrate by size, which depends on the
length of the CDR3 region(s) amplified. Size standards were made by
performing runoff reactions on PCR products of ␣ genes of known CDR3
size using a primer labeled with a different fluorescent dye. Size standards
were run in the same well with the sample to be assayed. The primers used
were designed at the start of the third framework region (Table I). Primers
for samples were labeled with 6-FAM dye, and primers for size standards
were labeled with HEX dye.
1338
OLIGOCLONAL IgA RESPONSE IN ACUTE KD
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
FIGURE 3. Nucleotide sequences of the CDR3 regions of additional ␣ sequences identified in the KD patient A vascular cDNA library. These clones
include members of the VH1, VH3, VH4, VH5, VH6, and VH7 families. Clones 10-11 and 6-4 are shorter clones with insufficient framework 3 region to allow
assignment to a VH family.
derived from the same original B cell that underwent clonal
expansion.
VH4-group 1 sequences. The VH4-group 1 sequences were investigated further, as this set contained the largest number of members of the five clonally related sets. These clones use a VH4 gene
most closely related to the germline VH4.18 gene, as determined
by comparing the obtained sequence to published germline sequences (21) and by comparing the sequence to immunoglobulin
sequences entered in the Genbank/EMBL database. The CDR3
region contains sequence most similar to the DN1 gene and uses a
JH1 gene. JH1 has been reported to be used by only 1% of all
human peripheral blood B cells (22, 26). Clones C5v, E2, 5-1, and
7-1a have identical CDR3 sequence, and clone A2 has one nucleotide difference (G3C) in CDR3 that confers an amino acid substitution (G3A) (Fig. 1, A and B). Clone A2 extends only to the
start of FR3 and has two additional nucleotide differences in FR3
compared to the other clones. Sequencing of the VH4-group 1 was
performed on excised phagemid rather than on PCR-amplified
The Journal of Immunology
1339
FIGURE 4. Dye-labeled runoff reactions of paraffin-embedded, formalin-fixed
spleen ␣ gene RT-PCR products from patient
A run on DNA sequencer. Positions of
CDR3 size standards (29, 33, and 42 nucleotides) are indicated with arrows. Top, VH3␣
products showing a polyclonal CDR3 size
usage. Bottom, VH4␣ products showing a
polyclonal CDR3 size usage.
and have been clonally expanded, consistent with an Ag-driven
immune response.
Other VDJ Rearrangements among isolated ␣ genes. CDR3
sequence data for additional ␣ genes isolated from the vascular
cDNA library are given in Fig. 3. They consist of three VH1 genes,
12 VH3 genes, seven VH4 genes, two VH5 genes, two VH6 genes,
and one VH7 gene.
Summary of clonally related sequences in KD vascular library.
Overall, five sets of clonally related sequences were identified. The
VH4 group 1 family, consisting of clones E2, A2, C5v, 5-1, and
7-1a; and the VH4 group 2 family, consisting of clones C2 and 4–5,
show evidence of somatic mutation within the group. We also
identified clonally related VH3 group 1, consisting of clones D3,
11-2, and 11-5; VH3 group 2, consisting of clones 2-2 and 13-1;
and VH7 group 1, consisting of clones 10-1, 7-1b, and 10-4. These
five groups of CDR3 sequences accounted for 15 of 44 (34%) of
the ␣ clones (five VH4-group 1, two VH4-group 2, three VH3group 1, two VH3-group 2, and three VH7-group 1) in the primary
cDNA library. None of the clonally related CDR3 regions had
been entered previously into the GenBank/EMBL database. All 44
sequences are now entered into the GenBank/EMBL database under accession numbers AF064878–AF064921.
Additional unfixed tissues were not available from other fatal
KD cases to allow for construction of other KD vascular libraries.
Therefore, methodology was developed to isolate and analyze ␣
gene sequences from formalin-fixed, paraffin-embedded coronary
artery tissues from other fatal KD cases.
Analysis of formalin-fixed, paraffin-embedded KD coronary
artery tissue blocks
We chose to amplify VH3␣ and VH4␣ gene sequences because
these families are most represented in the peripheral blood B cell
repertoire and consistently gave polyclonal CDR3 size distribution
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
product, ruling out PCR as a cause of these mutations. Therefore,
the sequence differences in these clones likely represent point mutations introduced during somatic mutation in B cells undergoing
clonal expansion.
In a separate experiment, we performed PCR amplification of
the amplified cDNA library using a VH4 leader primer and a
primer in the constant region of ␣, cloned the PCR products, and
hybridized clones with the oligonucleotide probe AR1 derived
from the VH4-group 1 CDR3 sequence. We obtained three fulllength clones with the expected CDR3 sequence, designated clones
17PCR, 28PCR, and 39PCR (GenBank accession numbers
AF247737–AF247739), all of which had VH4 gene sequence more
closely related to the germline VH4.18 than the clones isolated
from the primary library. For example, clone 28PCR demonstrated
95% homology to germline VH4.18, whereas the VH4-group 1
sequence E2 demonstrates 85% homology to germline VH4.18.
The CDR3 sequence of the clones identified by PCR was similar
to that of the VH4-group 1 sequences, except for the last two amino
acids (HY in VH4-group 1, QH in the PCR clones). This provides
further support for the assignment of the VH4-group 1 sequences to
the germline VH4.18, as the PCR clones had the same VDJ rearrangement but higher homology to germline VH4.18 than the VH4group 1 set of clones. This indicates the presence of additional
clones of the same B cell lineage in the cDNA library and provides
further support for clonal expansion with selection by Ag.
The VH4-group 1 sequences have undergone significant somatic
mutation; the R (replacement)/S (silent) ratio in these sequences
when compared to germline is 12:1 in the CDR regions compared
to 13:7 (1.9) in the FR regions. This compares to the calculated
R/S ratios for the VH4.18 germline gene of 4.5 for the CDR regions and 2.6 for the FR regions (27); observed number of CDR R
mutations is significantly higher than expected, p ⬍ 0.001 using
formula in Ref. 27. These data collectively indicate that the VH4group 1 sequences have undergone significant somatic mutation
1340
OLIGOCLONAL IgA RESPONSE IN ACUTE KD
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
FIGURE 5. Nucleotide and amino
acid sequences of CDR3 regions of VH3
and VH4␣ genes in paraffin-embedded,
formalin-fixed spleen from patient A.
patterns when RT-PCR and runoff reactions were performed on ␣
genes from peripheral blood B cell RNA from controls (data not
shown). Moreover, RT-PCR amplification of RNA isolated from
paraffin-embedded, formalin-fixed spleen tissue from KD patient
A yielded polyclonal VH3␣ and VH4␣ CDR3 size profiles (Fig. 4)
as expected. Sequencing of the spleen VH3␣ and VH4␣ PCR products revealed a diverse repertoire of CDR3 genes (Fig. 5). These
experiments demonstrate that our procedure will detect a polyclonal distribution of ␣ sequences if such sequences are present in
the sample.
The Journal of Immunology
1341
FIGURE 6. Dye-labeled runoff reactions of VH4␣ gene RT-PCR products of
paraffin-embedded, formalin-fixed coronary artery tissue from patient B run on
DNA sequencer, showing an oligoclonal
CDR3 size usage. Positions of CDR3
size standards (29, 33, and 42 nucleotides) are indicated with arrows.
Discussion
The presence of IgA plasma cells as a prominent component of the
vascular immune response in young infants with KD is interesting.
Systemic IgA responses appear to be immature in young infants
(28). However, mucosal IgA responses in this age group are mature (29), and infants presented with mucosal Ag such as poliovirus mount excellent Ag-specific IgA Ab responses (30). If IgA
plasma cell infiltration in KD is the result of stimulation of the
systemic IgA immune response, the IgA Abs produced may be
oligoclonal as a result of Ag stimulation, or polyclonal as a result
of generalized B cell activation or a response to a superantigen. In
young infants, polyclonal stimulation of peripheral blood induces
B cells to secrete predominately IgM, with fewer cells secreting
IgG and only very few secreting IgA (28, 31). A predominant IgA
immune response is also unlikely to be the result of stimulation of
B cells by a superantigen; B cell superantigens appear to bind
preferentially to the framework regions of germline immunoglobulins, and are thus more likely to bind to the IgM repertoire because
it generally contains less somatic hypermutation than does the IgG
and IgA repertoire (32). Therefore, it seems most plausible that the
IgA response in KD is Ag-driven, likely as a result of stimulation
of the mucosal IgA system by a specific etiologic agent. If this
hypothesis is correct, the IgA response in KD is likely to be oligoclonal, directed primarily at various epitopes of the etiologic
agent.
We have demonstrated in this study that clonally related ␣ genes
are present in a primary, unamplified vascular cDNA library from
patient A, a child who died of fatal acute KD. Use of the primary,
unamplified cDNA library avoided potential amplification bias that
might occur with amplification of the primary cDNA library or
copying errors that might occur with use of PCR methods to generate the population of ␣ genes to be studied. Strikingly, five
clonally related sets of ␣ sequences account for 34% of the ␣
sequences characterized from the primary, unamplified vascular
FIGURE 7. Nucleotide
and amino acid sequences of
CDR3 regions of predominant
VH3 and VH4␣ genes in paraffin-embedded, formalin-fixed
coronary arteries from patients
B and C.
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
IgA plasma cells were counted in 5-␮m tissue sections from
these blocks. Using the same calculation as described for the unfixed vascular tissue and assuming that the same plasma cell might
appear in five successive sections, we calculated that there would
be at least three times as many plasma cells in ten 8-␮m tissue
sections as in one 5-␮m tissue section. Sections from blocks from
both patients had ⬃400 IgA plasma cells in a 5-␮m tissue section
or, at a minimum, 1200 IgA plasma cells in the ten 8-␮m tissue
sections.
RT-PCR amplification of the coronary artery block from patient
B did not yield any visible VH3␣ product after reverse transcription and nested PCR, but a reproducible VH4␣ product was obtained. The runoff reaction products of this VH4␣ product are seen
in Fig. 6; one single prominent band corresponding to a CDR3 size
of ⬃33 nucleotides is seen. Nine clones obtained from this PCR
product were sequenced in the CDR3 region and were identical
(Fig. 7).
RT-PCR amplification of the coronary artery block from patient
C yielded visible VH3␣ and VH4␣ products; runoff reaction products are seen in Fig. 8. Eleven VH3 clones and 12 VH4 clones were
sequenced; each set of sequences within the amplified VH␣ family
was identical in the CDR3 region (Fig. 7). The CDR3 sequences
from patients B and C are entered into the GenBank database under accession numbers AF237658 –AF237660. Thus, in patients B
and C, the usage of predominant VH3 and VH4␣ sequences provides further evidence of an oligoclonal response in the KD
vascular wall.
One CDR3 sequence cloned from the VH4␣ PCR products from
paraffin-embedded, formalin-fixed spleen from patient A was identical to CDR3 sequence E1, which was isolated from the primary,
unamplified vascular cDNA library from KD patient A (Figs. 3 and
5). Isolation of clones from the vascular library and from the formalin-fixed, paraffin-embedded tissues were performed by different methods as described above, at different times, and in different
laboratories.
1342
OLIGOCLONAL IgA RESPONSE IN ACUTE KD
FIGURE 8. Dye-labeled runoff reactions
of ␣ gene RT-PCR products of paraffin-embedded, formalin-fixed coronary artery tissue
from patient C run on DNA sequencer. Positions of CDR3 size standards (29, 33, and
42 nucleotides) are indicated with arrows.
Top, VH3␣ products showing oligoclonal
CDR3 size usage. Bottom, VH4␣ products
showing oligoclonal CDR3 size usage.
(32). Rather, we found evidence of clonally related sequences in
the VH3, VH4, and VH7 families, consistent with an oligoclonal Ab
response directed at various antigenic epitopes of an etiologic
agent.
Although the presence of a set of two clonally related sequences
from our 44 clones isolated from the unamplified vascular cDNA
library might simply be due to isolation of the sequence from the
same plasma cell by chance, we believe this to be unlikely. First,
the cDNA library was prepared from total RNA originally isolated
from tissue that contained ⬃180,000 IgA plasma cells, as described in Materials and Methods. Second, sequence data clearly
indicate that the VH4-group 2 set consists of two different members, as shown in Fig. 1 C and D. These clones were sequenced as
excised pBluescript phagemids; we performed no PCR amplification with its potential to introduce nucleotide replacement by Taq
polymerase error. Clone C2 is full length and clone 4-5 includes
most of FR3. The clones differ in sequence by four nucleotides in
FR3 and one in CDR3, indicating that they were produced by two
different plasma cells (Fig. 1C). We cannot rule out that VH3group 2 clones 2-2 and 13-1 are from the same plasma cell; clone
13-1 extends only to the end of CDR2 and has the same nucleotide
sequence as clone 2-2 throughout FR3.
We cannot exclude that a superantigen-driven T cell response
occurs concurrently with an Ag-driven B cell response in acute
KD, but we believe this to be unlikely. Data indicate that clonal
expansion of CD8⫹ T cells occurs in acute KD (6, 9). As previously suggested (6), an apparent polyclonal increase of V␤2 and
V␤8 T cells in acute KD, suggesting activation by a superantigen
(4), may have actually resulted from failure to separate CD4⫹ and
CD8⫹ T cells, with the presence of CD4⫹ T cells obscuring clonal
expansion of CD8⫹ cells. The data presented in this paper support
an Ag-driven immune response in KD. We hypothesize that the
KD infectious agent enters the host through the respiratory route
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
cDNA library. The most frequently isolated sequence, the VH4group 1, shows evidence of extensive somatic mutation, with a R/S
ratio of 12 in the CDRs and 1.9 in the FR regions. It uses a JH1
gene, which has been reported to be used by only 1% of human
peripheral blood B cells (22, 26). The presence of these ␣ clones,
which have undergone somatic mutation and clonal expansion,
supports an Ag-driven B cell immune response in the vascular wall
in acute KD.
In addition, we have shown an oligoclonal pattern of VH3 and
VH4␣ gene usage in paraffin-embedded, formalin-fixed coronary
arteries from two additional fatal KD patients in whom one CDR3
sequence predominated in the VH3␣ or VH4␣ genes in the coronary artery tissues. Thus, evidence of an oligoclonal IgA response
was detected in vascular tissue of all three KD patients. Using the
same method, a polyclonal profile of VH3␣ and VH4␣ gene usage
in paraffin-embedded, formalin-fixed KD spleen was demonstrated. However, it is possible that the clonality of ␣ genes in KD
spleen is somewhat more restricted than are those in a spleen from
a healthy patient; this may be the case in spleen from any patient
with an overwhelming infection. To support this point, one identical CDR3 sequence was obtained from 44 ␣ sequences isolated
from a primary, unamplified vascular cDNA library from KD patient A and from 20 VH4␣ sequences cloned from a paraffin-embedded, formalin-fixed spleen sample from the same patient. The
fact that an identical CDR3 sequence was identified in a small
number of ␣ sequences examined from vascular tissue and spleen
from the same patient implies that the KD spleen is producing
some of the same oligoclonal IgA Ab that are being made in the
vascular wall.
The representation of different VH family members in our vascular cDNA library is similar to that reported for human peripheral
blood B cells (18, 22). We did not observe a significant expansion
of a VH family without clonality nor depletion of a VH family,
findings that might have suggested stimulation by a superantigen
The Journal of Immunology
Acknowledgements
We thank Dr. Katherine L. Knight and Dr. Stephen D. Miller for helpful
discussions regarding this research, and Danuta Wronska for technical
assistance.
References
1. Rowley, A. H., and S. T. Shulman. 1999. Kawasaki syndrome. Pediatr. Clin.
N. Am. 46:313.
2. Newburger, J. W., M. Takahashi, A. S. Beiser, J. C. Burns, J. Bastian, K. J.
Chung, S. D. Colan, C. E. Duffy, D. R. Fulton, M. P. Glode, et al. 1991. A Single
intravenous infusion of gamma globulin as compared with four infusions in the
treatment of acute Kawasaki syndrome. N. Engl. J. Med. 324:1633.
3. Burns, J. C., H. Shike, J. B. Gordon, A. Malhotra, M. Schoenwetter, and
T. Kawasaki. 1996. Sequelae of Kawasaki disease in adolescents and young
adults. J. Am. Coll. Cardiol. 28:253.
4. Abe, J., B. L. Kotzkin, K. Jujo, M. E. Melish, M. P. Glode, T. Kohsaka, and
D. Y. M. Leung. 1992. Selective expansion of T cells expressing T-cell receptor
variable regions V␤2 and V␤8 in Kawasaki disease. Proc. Natl. Acad. Sci. USA
89:4066.
5. Curtis, N., R. Zheng, J. R. Lamb, and M. Levin. 1995. Evidence for superantigen
mediated process in Kawasaki disease. Arch. Dis. Child. 72:308.
6. Pietra, B. A., J. De Inocencio, E. H. Giannini, and R. Hirsch. 1994. TCR V␤
family repertoire and T-cell activation markers in Kawasaki disease. J. Immunol.
153:1881.
7. Melish, M. E., J. Parsonett, and N. Marchette. 1994. Kawasaki syndrome (KD) is
not caused by toxic shock syndrome toxin-1 (TSST-1)⫹Staphylococci. Pediatr.
Res. 35:187A.
8. Nishiyori, A., M. Sakaguchi, H. Kato, K. Sagawa, K. Itoh, K. Miwa, and
H. Igarashi. 1995. Toxic shock syndrome toxin and V␤2 expression on T cells in
Kawasaki disease. In Kawasaki Disease. H. Kato ed., Elsevier Science, Amsterdam, p. 127.
9. Choi, I. H., Y. J. Chwae, W. S. Shim, D. S. Kim, D. H. Kwon, J. D. Kim, and
S. J. Kim. 1997. Clonal expansion of CD8⫹ T cells in Kawasaki disease. J. Immunol. 159:481.
10. Mancia, L., J. Wahlstrom, B. Schiller, L. Chini, G. Elinder, P. D’Argenio, D.
Gigliotti, H. Wigzell, P. Rossi, and J. Grunewald. 1998. Characterization of the
T-cell receptor V-␤ repertoire in Kawasaki disease. Scand. J. Immunol. 48:443.
11. Leung, D. Y. M., H. C. Meissner, D. R. Fulton, D. L. Murray, B. L. Kotzkin, and
P. M. Schlievert. 1993. Toxic shock syndrome toxin-secreting Staphylococcus
aureus in Kawasaki syndrome. Lancet 342:1385.
12. Terai, M., K. Miwa, T. Williams, W. Kabat, M. Fukuyama, Y. Okajima,
H. Igarashi, and S. T. Shulman. 1995. The absence of evidence of staphylococcal
toxin involvement in the pathogenesis of Kawasaki disease. J. Infect. Dis. 172:
558.
13. Morita, A., Y. Imada, H. Igarashi, and T. Yutsudo. 1997. Serologic evidence that
streptococcal superantigens are not involved in the pathogenesis of Kawasaki
disease. Microbiol. Immunol. 41:895.
14. Marchette, N. J., X. Cao, S. Kihara, J. Martin, and M. E. Melish. 1995. Staphylococcal toxic shock syndrome toxin-1, one possible cause of Kawasaki syndrome? In Kawasaki Disease. H. Kato, ed. Elsevier Science, Amsterdam, p. 149.
15. Rowley, A. H., C. A. Eckerley, H. M. Jäck, S. T. Shulman, and S. C. Baker. 1997.
IgA plasma cells in vascular tissue of patients with Kawasaki syndrome. J. Immunol. 159:5946.
16. Chomczynski, P., and N. Sacchi. 1987. Single-step method of RNA isolation by
acid guanidium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162:
156.
17. Islam, K. B., B. Baskin, B. Christensson, L. Hammarstrom, and C. I. E. Smith.
1994. In vivo expression of human immunoglobulin germ-line mRNA in normal
and in immunodeficient individuals. Clin. Exp. Immunol. 95:3.
18. Brezinschek, H. P., R. I. Brezinschek, and P. E. Lipsky. 1995. Analysis of the
heavy chain repertoire of human peripheral B cells using single-cell polymerase
chain reaction. J. Immunol. 155:190.
19. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular Cloning. Cold
Spring Harbor Laboratory Press, Plainview, NY.
20. Pannetier, C., M. Cochet, S. Darche, A. Casrouge, M. Zöller, and P. Kourilsky.
1993. The sizes of the CDR3 hypervariable regions of the murine T-cell receptor
␤ chains vary as a function of the recombined germ-line segments. Proc. Natl.
Acad. Sci. USA 90:4319.
21. Tomlinson, I. M., G. Walter, J. D. Marks, M. B. Llewelyn, and G. Winter. 1992.
The repertoire of human germline VH sequences reveals about fifty groups of VH
segments with different hypervariable loops. J. Mol. Biol. 227:776.
22. Brezinschek, H. P., S. J. Foster, R. I. Brezinschek, T. Dörner, R. Domiati-Saad,
and P. Lipsky. 1997. Analysis of the human VH gene repertoire. J. Clin. Invest.
99:2488.
23. Mortari F., J. A. Newton, J. Y. Wang, and H. W. Schroeder. 1992. The human
cord blood antibody repertoire: frequent usage of the VH7 family. Eur. J. Immunol. 22:241.
24. French, D. L., R. Laskov, and M. D. Scharff. 1989. The role of somatic hypermutation in the generation of antibody diversity. Science 244:1152.
25. Jacob, J., G. Kelsoe, K. Rajewsky, and U. Weiss. 1991. Intraclonal generation of
antibody mutants in germinal centres. Nature 354:389.
26. Yamada, M., R. Wasserman, B. A. Reichard, S. Shane, A. J. Caton, and
G. Rovera. 1991. Preferential utilization of specific immunoglobulin heavy chain
diversity and joining segments in adult human peripheral blood B lymphocytes.
J. Exp. Med. 173:395.
27. Chang, B., and P. Casali. 1994. The CDR1 sequences of a major proportion of
human germline Ig VH genes are inherently susceptible to amino acid replacement. Immunol. Today 15:367.
28. Andersson, U. 1985. Development of B lymphocyte function in childhood. Acta
Paediatr. Scand. 74:568.
29. Mestecky, J. and J. R. McGhee. 1993. The secretory IgA system. In Local Immunity in Reproduction Tract Tissues. P. D. Griffin and P. M. Johnson, eds.
Oxford University Press, New York, p. 53.
30. Faden, H., J. F. Modlin, M. L. Thomas, A. M. McBean, M. B. Ferdon, and
P. L. Ogra. 1990. Comparative evaluation of immunization with live attenuated
and enhanced-potency inactivated trivalent poliovirus vaccines in childhood: systemic and local immune responses. J. Infect. Dis. 162:1291.
31. Andersson, U., A. G. Bird, S. Britton, and R. Palacios. 1981. Humoral and cellular immunity in humans studied at the cell level from birth to two years of age.
Immunol. Rev. 57:5.
32. Silverman, G. J. 1995. Unconventional B-cell antigens and human immune repertoires. Ann. NY Acad. Sci. 764:342.
33. Owens, G. P., R. A. Williamson, M. P. Burgoon, O. Ghausi, D. R. Burton, and
D. H. Gilden. 2000. Cloning the antibody response in humans with chronic inflammatory disease: immunopanning of subacute sclerosing Panencephalitis
(SSPE) brain sections with antibody phage libraries prepared from SSPE brain
enriches for antibody recognizing measles virus antigens in situ. J. Virol. 74:
1533.
34. Vandvik, B., E. Norrby, J. Steen-Johnson and K. Sensvold. 1978. Mumps meningitis: prolonged pleocytosis and occurrence of mumps virus-specific oligoclonal IgG in the cerebrospinal fluid. Eur. Neurol. 17:13.
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
and is recognized and processed in the bronchus-associated lymphoid tissues. Epidemiologic data in KD are consistent with a respiratory portal of entry of the etiologic agent (1). B cells then
switch to IgA and enter the general circulation. The pathogen may
enter the bloodstream, resulting in the multisystem involvement
characteristic of KD (1), and also enter vascular tissues. IgA B
cells may enter the vascular walls with other inflammatory cells
and there differentiate into IgA plasma cells, probably under the
influence of locally produced cytokines. Our findings strongly support the hypothesis of an Ag-driven immune response in KD, because we found both clonally related ␣ sequences and evidence of
somatic mutation within the related clones in the vascular cDNA
library. These findings are characteristic of an Ag-driven immune
response in a germinal center (25). By probing the KD vascular
cDNA library with an oligonucleotide probe corresponding to the
CDR3 sequence of the VH4-group 1 set of ␣ genes, we identified
additional sequences of the same B cell lineage in the library,
findings again typical of an Ag-driven immune response in a germinal center. Because germinal centers are not present in KD vascular tissue (15), this suggests that IgA cells are undergoing somatic mutation in the bronchus-associated lymphoid tissue or other
lymphoid tissue in response to Ag prior to migrating to the vascular tissue.
Our finding of an oligoclonal IgA response in the vascular wall
in acute KD supports an Ag-driven immune response. The production of oligoclonal Abs in human tissues that are directed at the
disease-causing pathogen is a feature of many infectious processes
(33, 34). Our data support the hypothesis of a conventional Ag in
the pathogenesis of KD.
1343