-
Clin Exp Immunol 1994; 95:49-55
Antineutrophil cytoplasm antibodies (ANCA) of IgA isotype
inadult Henoch-Schonlein purpura
N. RONDA. V. L. M. ESNAULT. L. LAYWARD*. V. SEPE. A. ALLEN*. J.
FEHHALLY* & C. M.LOCKWOOD Department of Medicine. School of
Clinical Medicine. University of Cambridge. Cambridge and
* Department of Nephrology. Leicester General Hospital.
Leieester. UK
(Accepted for publieation 21 September 1993)
SUMMARYANCA are associated with certain forms of systemic
vasculitis, and have been reported previously tobe of Ihe IgG and
IgM isotype. We examined ihc possible association between IgA ANCA
and theIgA-relaled diseases Henoch-Schonlein purpura (HSP) and IgA
nephropaihy (IgAN). IgA and IgGANCA were detected by
isotype-specifie solid-phase assays with a erude neutrophil
extract, and theirpresence was confirmed by antigen-specific
fluid-phase competitive inhibition tests and by
indirectinimunofiuorescence. The possible interference by IgA
rheumatoid faeior was excluded. IgA ANCAwere detected in sera from
11/I4 HSP patients (79%). from 1/30 IgAN paLicnts 0"A>), from
1/40patients with vasculitldes classically associated with IgG ANCA
(2-5%), and in none or60 sera fromhealthy blood donors. IgG ANCA
were present with IgA ANCA in three patients with HSP. Onlyone HSP
serum had anti-myeloperoxidase (MPO) activity by both IgA and IgG
isotypc-specificELISA, and none was positive for proteinase 3
(PR3). Western blot analysis performed withneutrophil extract
showed that the four strongest IgA ANCA-positive HSP sera reacted
with a 51-kDprotein; Western blot performed on cellular fractions
showed that this protein is primarilymembrane-assoeiated, and
different from fibronectin. Our study suggests that adult HSP is
eloselyassociated with cireulating IgA ANCA. which may be directed
against a different autoantigen thanthat recognized by IgG
ANCA.
Keywords antineutrophil eytoplasmic antibodies autoimmunity
IgAHenoeh Schontein purpura systemie vasculitis
INTRODUCTIONThe vasculitis syndromes comprise a spectrum of
diseases as yetstill defined by clinical and histological criteria
[1.2]. Theassociation of ANCA with Wegener's granulomatosis
[.1],microscopic polyarterilis [4] and renal limited vasculitides
[5]has indicated that autoimmune mechanisms may underlie
thedevelopment of these diseases, and increasing evidence points
totheir importance in pathogenesis 16]. Studies so far indicate
thatANCA are predominantly ofthe IgG isotype. However. ANCAof IgM
isotype (without IgG) have been found in patients with
apulmonary/renal, anti-glomerular basement membrane (GBM)negative
syndrome [7], as well as in association with IgG ANCAin patients
presenting with pulmonary haemorrhage [8],
Henoeh Schonlein purpura (HSP) is a systemic vasculitis.more
eommon in children, but well documented in adults [9],whieh is
associated with IgA abnormalities. These include raisedcirculating
IgA levels [10.11]. the presence in serum of IgA-containing immune
complexes [10.12.1.3] and IgA deposits in
Correspondence: Dr Nieolelta Ronda, Istiiulo di Clinica Mediea
eNefrologia. tJniversita degli Studi, v. Gramsci, 14.4:illOO Parma.
Italy.
skin [14.15]. as well as renal mesangium [14.16]. The same
renalimmunopathologie and serologic IgA abnormalities are
lypi-cally found in igA nephropathy (IgAN). and the two diseasescan
be distinguished only by the extrarenal vasculilic involve-ment in
HSP [15.17.18].
IgA ANCA have been reported lo occur in HSP [19]. but asyel this
finding has not been eonfirmed by others [20], and il hasbeen
pointed oul thai patienl IgA fibronectin complexes mayaccount, at
least in part, for some IgA ANCA activity [21]. Weexamined the
possible association of IgA ANCA wiih HSP andIgAN. excluding IgA
rheumatoid factor interference, andinvestigated autoantibody
specificity by means of antigen-speeifie solid-phase immunoassays
and Western blotting.
PATIENTS AND METHODSPatients
HSP (n= 14. mean age 45 years, age range 17-76 years, 12male and
two female). Palienls included in this group fulfilledthe criteria
of the American Rheumatism Association. Renalbiopsy was performed
in 12 patients and showed mesangial
49
-
50 .^ Ronda et al.
deposition of IgA in all cases. Sera were stored from all
patientsat presentation, or at times of clinical relapse. None of
thepatients was receiving treatment al the time serum was
obtained.
IgAN (/I = 30, mean age 36 years, age range 16-60. 24 maleand
six female). These patients had impairmcnl of renalfunction (or
haematuria/proteinuHa) and diffuse mesangial IgAdeposition, without
extrarenal involvement. Sera from all IgANpatients were taken at
onset of the disease, and in .six patientsduring remission, as well
as during episodes of maeroseopiehaematuria.
!gG ANCA-positive systemic lasciditis (/! 40).
Wegener'sgranulomatosis {/( = 9) or mieroscopic polyarteritis (/7 =
31).
Healthy donors (H = 60) . These were studied as negativecontrols
(age range 17-64 years, 38 male and 22 female).
ANCA solid-phase immunoa.ssaySereening on the crude neutrophil
extraet for IgG ANCA wasperformed as described [4]. Detection of
IgA ANCA was byELISA using the same neutrophil preparation,
blocking with2"A> bovine serum albumin (BSA); gelatin was used
as a blockingagent in a subsequent ELISA performed with the sera
whichshowed high reactivity, loexcludenon-specilic binding of IgA
toBSA. For this purpose, controls also included antigen-free
wellsblocked with BSA or gelatin and incubated with patient
andcontrol sera. Serum dilutions were calculated for each
samplefrom the total IgA concentrations in order to obtain a linal
IgAconcentration of 20 mg/dl in both patient and eontrol
sera.Binding was deteeted by adding alkaline
phosphatase-eonju-gated polyclona! goat anti-human IgA (Sigma,
A9669) diluted1:500. and alkaline phosphatase substrate bulTer.
After 40 minabsorbance was measured at 402 nm. Alt volumes were 100
ml/well and all incubations were for I h at 37 C, with triple
washesof PBS containing 0 1% Tween 20 (PBS-T) between stages.ELISA
posilivity was defined as absorbance > 2 s.d. over themean of
values obtained from :t panel of normal sera. IgAANCA-positive and
normal control sera were also tested in amodified ANCA solid-phase
immunoassay (SPIA), usingmouse anti-human IgA 1 and IgA2 MoAbs
(Unipath, M120 i Iand M270ID) and a rabbit anti-mouse
peroxidase-eonjugatedantibody (Dako, P260) as revealing system. In
this ease boundIgA was revealed with orthophenylenediamine nnd H;0;
incitric acid bufler, and reaction blocked after 20 min with 2
Msulphuric acid. Absorbance was read at 492 nm.
Antigen-speeifie assays were performed: igGanti-myeloper-oxidase
(MPO) as deseribed [22] and IgG anti-protcinase 3(PR3) using a
commercially available kit (Bio-Carb, Sweden)according to the
manufacturer's instructions. These assays wereadapted to deteci IgA
anti-MPO and IgA anti-PR3 by using thesame detection system as for
the total IgA ANCA SPIA,
Inhibition studiesThe specilicity of IgA ANCA binding in the
ANCA SPIA wastested by means of a fluid-phase inhibition test [4].
Aliquots ofsera were retested in the ANCA SPIA after incubation for
I h at37 C with respectively PBS, haemoglobin imd neutrophilextract
at proteineoncentrations ranging from 0 005 to 4 mg/ml.The test was
considered to be positive when the reduction inANCA binding was
greater than 20"/ii following incubation withneutrophil extract,
but not with the irrelevant protein, and whena dose-response
related to the neutrophil extract coneentrationwas observed.
Indirect immunojluorescenceIndirect immunofluoreseence (IIF) was
performed on ethanol-fixed human neutrophils, according to the
method previouslydescribed [4], modified for detection of IgA
antibodies by usingan anti-IgA lluorescein-eonjugatcd antibody
(Dako, F204)diluted 1:5 in PBS. Serum dilutions used were I; 2, 1:4
and 1:8for each patient and control sample. The fluorescence
wasassessed independently by three investigators.
IgA and IgG rheumatoidfacior ELISAA specific ELISA kit for tbe
detection oflgA rheumatoid factor(RF) was used on all IgA
ANCA-positive sera according to themanufacturer's instructions
(SKI.ISA; Walker LaboratoriesLtd., Ely, LIK). The standard curve of
IgA RF activity wasobtained using IgA RF-positive serum, provided
by the manu-facturer, from a patient with rheumatoid arthritis
(fulHIling thecriteria ofthe American Rheumatism and Arthritis
Council forRheumatoid Arthritis); the serum had been previously
cali-brated against the International Standard of IgA RF in order
todeteet from 3 0 to 600 IgA RF U/ml. The presence of IgG anti-IgA
RE was tested by ELISA on igA-coated plates.
Two HSP patients' sera and a serum positive for IgA RFwere
chromatographed on a Sepiiarose column coupled withProtein A at
dilutions of 1:10. 1:20 and I ;40, The loaded andthe effluent
fractions were tested for IgA ANCA by ELISA asdescribed.
Criteria for IgA ANCA positivilyPositivity was defined by the
association of all the followingcriteria: (i) positive IgA ELISA
and antigcn-spccilie dose-dependent fluid-phase inhibition; (ii)
positive IgA IIF; and (iii)absence of IgA RK activity.
Western blot for Ig.4 ANCANcutrophils were isolated on a Hypaque
gradient [4], resus-pended in PBS containing 0? niM MgCl:. i)*-)
niM CaCti andprotease inhibitor (phenylmethylsulfonyltluoride I mM)
andincubated at 37 C for 30 min with I mg/ml phorbol
myristateaeetate (PMA) to induee the release of secondary granules
[23|.After centrifugation the supernatant was removed and stored at
80 C until use in Western biol. nnd the cells washed threetimes
with PBS. The cells were then incubated in PBS with 1 "-Triton
XlOO. phenylmethylsulfonylfluoride 1 niM and EDTA2 mM at 4 C for 40
min. Following eentrifugation at ,S00 g for5 min at 4 C, the
supernatant was frozen at - 8 0 C until use inWestern blot,
SDS-PAGE was performed with: (i) the storedsupernatant after PMA
stimulation; and (ii) the neutrophilextract in reducing or
non-reducing buffer. Western blot wasperformed according to the
method described by Laemmli [24].IgA ANCA-positive and control sera
were diluted 1:20 in Tris-bulTered saline containing 1",. Tween 20
and 2"',) dried skimmedmilk (TBSTM) and incubated with the
nitrocellulose strips for 3h at room temperature. The strips were
then washed withTBSTM and incubated for 1 h with alkaline
phosphatase-conjugated anti-human IgA (Sigma, A9669), diluted
1:1000 inTBSTM. Binding was detected by adding alkaline
phosphatasesubstrate.
In order to study the antigen localization in the
neutrophils.cellular fractions were prepared from a total extract,
i.e. w ithoutPMA stimulation. Neutrophils were isolated on a
Hypaquegradient, washed twiee in PBS and ineubated in PBS
containing
-
IgA ANCA in HSP 51Table I. Orctirrence of IgA and IgG ANCA in
Hcnoch Schonleinpurptir;i(KSP). IgA nephropathy (IgAN). IgG
ANC'A-posilive vasculi-
tis (SV) patients and conlrols
HSP (%) IgAN (%) SV (%) Conlrols (%)
IgA ANCA
IgG ANCA
11/14(79)3/14(21)
1/30 (3)0/30 (0)
1/40(2-5)40/40 M (HI)
0/60 (0)0/60 (0)
T'/.i Triion XIOO. phenyimethylsulfonylfluoride 1 mM andEDTA 2
mM at 4 C for 40 min. Cellular extract was thencentriftigcd at
lOOOOj? for i 5 min at 4 C. The pellet was passedseveral times
through a 22 G needle and stored tintil use. Thesupernatant,
eontaining solubilized membranes and cytosoliematerial, was divided
into two aliquots. One aliquot was storeduntil use and the other
was centrifuged at I05000j?ror40minat4 C. Both the pellet and Ihe
supernatant obtained with this lasteentrifugation were saved for
use. ihe pellet being resuspendedin a volume equivalent to the
supernatant. SDS-PAGE wasperformed in reducing conditions with: (i)
pellet and super-natant after centrifugation at lOOflO^ (containing
respectivelynuclei + coarse fragments of membrane and solubilized
mem-branes + eytosolic material); (ii) pellet and supernatant
aftercentrifugation al 105 OOOjf (representing respectively
solubilizdmembrane-enriched material and soluble cytoplasm).
AfterWestern blot. IgA reactivity against each fraction was
testedusing an IgA ANCA-positive serum (whieh had previously
beenshown to react against the 51-kD protein) and a eontrol
serum.
In order to investigate whether IgA-fibronectin
interactionscould be responsible for the 5i-kD band of reactivity
observedin the Western blot of" HSP IgA, Weslern blot experiments
witha total neutrophil extract and the membrane-rich fraction
werealso perfonned to test the reactivity of three
commerciallyavailable anti-libronectin antibodies (rabbit
polyclonal anti-human tibroneclin, mouse monoclonal anti-human
fibronectinand mouse monoclonal anti-cellular fibronectin: Sigma.
St.Louis. MO). Two IgA ANCA-positive HSP sera were used aspositive
controls. Revealing aniibodies were alkaline
phospha-tase-conjugated anti-rabbil IgG, anti-mouse IgG.
anti-mouseIgM and anti-human IgA.
RESULTSigA ANCA were deteeted in 11/14 (79%) of HSP patients, in
1/M) (3%) of IgAN. and in 1/40 (25%) systemic vasculitis
(SV)patienls. None of the sera from the healthy controls was
foundlo have IgA ANCA activity (Table 1). IgG ANCA was found
iniissocialion wiih IgA ANCA in three of the HSP patients, bulnol
in any of the patients with IgAN, nor in healthy controls.
ANCA SPIA with inhibition studyThe resull of the IgA ANCA SPIA
Is shown in Fig. I. High totalIgA levels are a common finding in
HSP and IgAN patients; wetherefore used for eaeh sample a different
serum dilution (from1:10 to 1;33) to obtain 20 mg/dl of total IgA
to exclude thepossibility of non-specific binding: this theoretical
possibilitywas also excluded by testing each serum on bolh
antigen-freewells bloeked with BSA and antigen-free wells blocked
withgelatin on the same microlltre plate. None of the positive
sera
800 r
600
400
200
Controls HSP IgAN/I = 30)
SV
Fig. 1. IgA ANCA ELISA resulls rehiive to palicnts with
HcnochSchonlein purpura (HSP). IgA nephropaihy (IgAN), lg(J
ANCA-positlvc sysictitic viisculitis (SV). and conlrols. On the
f-axis is theoptical density (OD). Bars show tneanis.d. for each
population. Thedotted line represents the upper normai value (mean
+ 2 s.d. of valuesfrom controls). Individual values wiihin the
normal range are notplotted.
Table 2. Maximal percent inhibition of IgA ANCAactivity in
ELISA. obtained by co-incubaiion ofIgA ANCA-posilive sera with
neutrophil extract
and with haemoglobin
HSPI23456789
1011
IgAN*SV*
Neutrophilextraet
8y52402430362661622133
4945
Haemoglobin
50020000203
80
* One serum oulor30 IgA nephropathy (IgAN)and one out of 40
systemic vasculitis (SV) sera wilhIgA ANCA positivity. Sec text.
HSP. Henoch-Schonlein purpura.
Sera were diluted 1:40 and competitor proteinconcentration was 2
mg/ml. The inhibition wasdosc-depcndent in all cases.
-
52 N. Ronda et al.
66
51
56
Z9
20
Fig. 2. IgA ANCA Western btol obtained with Triton XlOO
neutrophilextract after phorbol myrislate acetate (PMA)
stimuialion. run in SDSPAGE under reducing conditions. Lanes I }
were blotted with threeIgA ANCA Henoch-Schonlein purpura (HSP)
sera, and show IgAreactivity wilh a 51-kD protein. Lanes 4 and 5
were blotted with normalsera. The other 12 normal and live IgA
nephropathy control seraproduced no band (same appearance as in
lane 5). The faini band ol'Ume4 is discussed in the te\t.
t(D
97 -a
6B
I 2 3Fig. 3. IgA ANCA Western biot obtained with fractions of
theneulrophil extract. The neutrophil fraction containing
solubilizcdmembranes and cytosolic material, obtained as
supernatant aflcr acentrifugalion at 10()(H)A', was found lo
contain ihe 51-kD protein. LaneI is ihe cooitiassie blue-stained
nitrocellulose afler the protein transfer,Ljine 2 was blotted wilh
a normal serum. Lane .^ was blotted wiili ;m IgAANCA-positive
Henoch Scbonlein purpura (HSP) serum. IgA in thisserum reacted with
the 51-kD protein and also with a 57-kD protein.The band al 68 kli
in lanes 2 and 3 is non-spccilic and is due to direclbinding of the
alkaline phosphatase-conjugated anti-hunian IgA.
showed significant IgA binding to the aniigcn-free wells.
IgAANCA-positive HSP sera retained positivity with scrum dilu-tions
ranging from 1:40 to 1:80. In all but one IgAN patientsIgA ANCA
were not detectable, regardless of the disease phase.
All IgA ANCA-positive sera were also positive in the IgAlANCA
SPIA; of these three HSP. one IgAN and one SV seraalso showed
raised levels of IgA2 ANCA.
The inhibition of IgA ANCA activity obtained for eachpositive
paiient by preinctibation of scrum with neutrophilextract and with
haemoglobin is shown in Table 2.
Indirect immunqfluorescencePositive IgA IIF was obtained with
all the positive seradescribed above, and produced a diffuse
cytoplasmic staining,which did not have the granular appearance
typically observedwith IgG ANCA IIF, Serum from one HSP patient
producedboth cytoplasmic and perinuclear (PN) staining: this serum
wasalso IgG ANCA-positive and produced a PN pattern in IgGIIK.
Cytoplastnic staining was obtained by testing the two otherIgG
ANCA-positive HSP sera in IgG-specific IIF.
IgA RF FUSANone of the HSP and IgAN sera showed IgA RF or IgG
anti-IgA RF activity. IgA RF was found in one SV patient,
positivefor both IgA and IgG ANCA: this patient was not considered
asbeing positive for IgA ANCA.
To further exclude IgA RF interference in IgA ANCAmeasurements,
sera from two IgA ANCA-positive HSP patientsand from the IgA
RF-positive patient used as control werediluted l:i(). 1:20 and
1:40 and chromatographed on aSepharose column coupled to Protein A.
which binds IgG, docsnot bind IgAl.and binds very weakly lgA2. IgA
ANCA activityin the loaded and in the IgG-depletcd material was
then assessedby ELISA, In the case of the positive IgA RF serum, a
reductionin IgA ANCA activity of 29"-^ !, 55'V.i and 85"''..,
compared withthe loaded material, was obtained by testing the
eliluenlrespectively at 1:10. 1:20 and I :40. No difference was
observedin IgA ANCA aetivity between the IgG-depleted and the
loadedHSP sera at any of the dilutions.
Antigen-specific SPIATwo patients were found to have IgA and IgG
anti-MPO
-
IgA ANCA in HSP 53
kD
68
29
('ig. 4. IgA ANCA Western blnl obiained wiili fractions of
theiicuirophil extract. The ncutrDphil fraclion used in the
cxpcrimetilsliown in Fig, } was further fraclionalcd by
cenirifiigalion al 105000^and ihc same Henoch Schtinlcin purpura
(HSP) and conirol sera as inlig, .1 were used for the immunoblot.
Results obtained wilh ihesupernatanl. thai is soluble cytoplasmie
material, are shown in lanes 1. 2and 3; results relative to the
pellet. Ihat is the membrane-rich fratlion.are shown in lanes 4. 5
and 6. Lanes I and 4 show ihe coomassie blue-slained nitrocellulose
after prolein transfer; the membrane-rich fraction(lane 4) is
relatively depleted in prolein contenl. Lanes 2 and 5
wereiiicubilled wilh ihe healthy control serum and lanes 3 and 6
wilh the HSPpatient serum. The almost complete lack of reactivity
of patient's IgAagainst ihe soluble cytoplasmie fraetion ofthe
neutrophil extrnci and theconserved reactivity against the
membrane-rich fraclion, relativelydepleleJ in prolein content,
indicate ihat the followed procedureprovides a partially purified
antigen preparation and that the 51-kDprotein recognized by ihe
four strongest IgA ANCA-positive HSP serais probably
membrane-associaled.
97
68
51
43
2 9 -
1 2 3 4 5Fig. 5. Western blot on total neiitrophil extract to
test reaclivity of anti-fibroneclin antibodies. Lanes 1 and 2 were
bloited with IgA ANCA-positive Henoch-Schonlein purpura (HSP) sera
as positive controls.Lanes 3. 4 and 5 were bloited respectively
with rabbit polyclonal anti-human fibroneclin, mouse monoclonal
anti*human fibroncctm andmouse monoclonal anti-cellular
fibroncclin. None ofthe anti-fibronec-lin antibodies reeognized the
51-kD prolein whieh is recognized by HSPIgA, The three antibodies
reacted with a 25-kD protein in the neulrophilextract and two of
them with a 4()-kD prolein, IgA from one of thepatients produced a
faint band of reaclivily at 25 kD. Anii-tibronectinantibodies
reacted with the same molecular species in experiments (notshown)
performed using the membr:ine-rieh fraction of neulrophilextract,
but the reactivity was much weaker. In this case HSP IgA whichhad
weakly reacted with the 25-kD prolein did not produce anydetectable
band al ihat molecular weight, whilst both HSP IgA stronglyreacted
wilh the 51-kD prolein.
aniibodies (one HSP and one SV), NeJiher IgG nor IgA anii-PR3
aclivity was delected in the sera with IgA ANCA positivity.
Hestetn hint for IgA ANCAThe three strongest IgA ANCA-positive
HSP sera reacted wilha 51-kD protein in the Triton XIOO neutrophil
extract afterPMA stimulation, under reducing conditions (Fig. 2),
The samereaciion was observed at a lesser degree of intensity by
testingthe sera on the supernatant obtained after PMA
slimulalion.Another IgA ANCA-posilive HSP serum reacted against the
51 -kD and against another protein of 57 kD (Fig. 3), A very
faintband at 51 52 kD was produced by one normal serum. None ofthe
other 13 normal and five IgAN conirol sera showed anyreactivity.
Western blot performed wiih ihe various cellularfractions showed
that the membrane-rich fraction, that is thepellet obtained afler
centrifugation at 105000 g. is stronglyenriched in the 51-kD
prolein recognized by IgA ANCA-positive HSP sera (Fig, 4), The
three dJITerenl anti-libronectiniinlibodies tested by Western blot
on a total neutrophil extractand on the membrane-rich fraction did
not show any reactivityut 51 kD (Fig, 5), All three produced a band
of reactivity at 25kDand twoof them a second band al 40 kD, IgA
from one of theHSP sera tested produced a very faini band at 25 kD
in Westernblot performed with total neutrophil extract (Fig,
5).
DISCUSSIONThe high prevalence of IgA ANCA in HSP (79%) but not
inIgAN {yv
-
54 A'. Ronda et a].frequently recognized by IgG or IgM ANCA:
only one HSPserum showed IgA and IgG reactivity against MPO; this
serumalone produced both cytoplasmic and perinuclear staining
inIgA-specific HF. None of the sera reacted against PR3-
Thereactivity of the four strongest IgA ANCA-positive HSP
seraagainst a 51-kD protein in the Western blot suggests that a
newautoantigen might be the target of IgA ANCA in this disease.We
have also included in Kig. 2 the Western blot obtained withthe only
control serum, of 14 normal and live IgAN tested.which produced a
faint band at 51-52 kD. If the specificity ofthis binding is
ultimately shown to be the same as that obtainedwith HSP sera, the
finding could be explained by the occurrenceof natural ANCA
activity in the sera of some normal indi-viduals. Such activity has
already been described [27] and is notsurprising, sinee
autoantibodies of several specificities consti-tute a substantial
part of normal circulating immunoglobulin.The Western blot linding
that Ihe 51-kD protein reeognized inthe neutrophil extract by four
HSP sera is present in themembrane-rich fraction, which is
relatively depleted in proteincontent, and is almost completely
absent from the cytoplasmicsoluble compartment, indicates that the
IgA ANCA target isprobably a membrane-associated protein.
Characterization ofthis antigen by amino acid sequence analysis is
currently theobjective of a separate study.
One of the most controversial issues about IgA ANCA
inIgA-related renal diseases is the possibility of obtaining
falsepositive results in ELISA or immunoRuorescence tests, due
tothe inereased levels of circulating IgA. or to
physicochemicalchanges of circulating IgA in these patients, which
couldincrease the ability of IgA to bind to several molecules
through"non specific" interactions [28]. In particular it has been
pointedout that, because fibroneetin is found in small amounts
inneutrophil extracts [21,29] and because IgA-fibronectin
com-plexes are increased in IgAN [30], IgA fibronectin
complexescould account, at least partly, for the IgA ANCA
reactivitydetected in IgAN and HSP [21]. We have performed
Westernblot experiments to test whether IgA reactivity towards
fibro-nectin could be responsible for the IgA ANCA activity that
wedetected in HSP sera. Threedifferent anti-fibronectin
antibodiesdid not react with the 51-kD protein in the neutrophil
extractthat appears to be the major target for HSP IgA. All
anti-fibronectin antibodies reacted with a 25-kD protein and two
ofthem with a 40-kD protein, supporting previous studies
indicat-ing that fibronectin fragments could be present in the
neutrophilextract. We also found that one of the HSP sera tested
produceda weak band of reactivity at 25 kD. Taken together,
thesefindings support the view that some antibody speeies within
theIgA circulating pool in HSP patients may be able to interactwith
neutrophil-assoeiated fibronectin [21]. but also confirmthat within
such pool other antibodies specifically associatedwith HSP are able
to react with a neutrophil antigen, anddeserve to be further
studied. A matter of debate eould now bewhether all these
antibodies should be ealled IgA ANCA. Untilwe have more information
about fibroneetin-reaetive anti-bodies, at present we call IgA ANCA
the antibody activitywhich we detected specifically in HSP patients
and whichappears to be directed against a 51-kD protein.
IgA ANCA are probably autoantibodies with a relativelylow
affinity for the antigen or/and present in small amounts inihe
circulation, as suggested by the fact that they can be detectedin
ELISA and in immunofluorescence within a relativciv narrow
range of serum dilutions. Moreover, at least a subset of IgAANCA
reaets with a neutrophil protein which is probably notabundant in
the cells. For these reasons, and for ihc multiplecriteria that we
set to define positivity in order to minimize therisk of false
positive results, we believe that IgA ANCA are notyet useful for
diagnosis of HSP. at least until the antigen isidentified. However,
in the case of particularly aggressivedisease. IgA ANCA monitoring
might soon be a useful tool forelinicians, as suggested by a case
report recently published [31],in which IgA ANCA titres correlated
strictly with diseaseactivity.
We have shown that IgA ANCA occur in a high percentageof adult
patients with HSP, but not with tgAN or SV. IgAANCA specificity in
most cases appeared to be different fromthe target reported to be
most commonly recognized by IgGANCA, This information may
contribute to an understandingof the pathogenesis of Henoch
Schonlcni purpura and may beuseful for the classification of
systemic vasculitides.
ACKNOWLEDGMENTSN,R. is a recipient of a gram from tlie
Universily of Roma Tor VergaUi.Italy.
REFERENCES! Fauci AS, Hayncs BF. K.al7 P, The spectrum of
vasculitis, Ann lnt
Med 1978; 89:660 78.2 Lie JT. Members and Consuluints of ihe
American College ol'
Rheumatology. Ittustruted iiislopatholoijicclassilication
criteria forselected vasculilis syndromes. Arlhriiis Rheum IWI):
.1.^1074 87,
}> Van der Woude FJ, Rasmusscn N. Lobatlo S ct al.
Autountibodiesagainst neutrophils and monocytes: lool tor diagnosis
;md murker ofdisease activity in Wegencr's granulomalosis. Lancet
1985; i:425 9.
4 Savage COS, Winearls CG. Jones S et al. Prospective study
ofradioimmunoassay for aniibodies against neulrophil eytoplasm
indiagnosis of systemic vasculitis. Lancet 1487: i:t3S9 93.
5 Fiilk RJ. Jenncltc JC. Anti-ncutrophil cytoplasmic
autoantibodieswith specificity for myclopcroxidasc in palients wilh
systemicvasculilis and idiopathic necroli?:ing and crescentie
gtomeruloncph-ritis. N Engl J Mod t9HK; 318:1651 7,
6 Ewcrt BH, Jcnnctie JC, Falk RJ. .^nti-mycloperoxidase
antibodiesstimulate neutrophils todamagc human cndotheliai cells.
Kidney lnt1992:4l;.175 8.1,
7 Jayne DRW. Jones SJ. Severn A et al. Severe pulmonary
hemor-rhage and systemic vasculitis in association wilh circulating
unti-neulrophil cytoplasm antibodies of IgM clas.s only. Clin
Nephrol1989:32:101 6,
8 Esnault VLM, Soieimani B. Kcogaii MT ct ol. Association of
IgMwilh IgG ANC,\ in patients presenting wilh pulmonary
hemor-rhage. Kidney Inl 1992; 41:1.^04 10.
9 Cream JJ. Giimpel JM. Peachy RDG. Schonlein Henoch purpurnin
aduits: a siudy of 77 adulls wilh aniiphyiacloid or SchonleinHcnoch
purpura- Q J Med 1970: 39:461 84.
10 Levinsky RJ, Barratt TM. IgA immune complexes in
Henoch-Schonlein purpura. Lancet 1979; ji:l 100 3.
11 Trygstad CW. Stichn ER. Elevaled serum IgA globulin in
anaphy-lactoid purpura. Paediatrics 1971: 47:lt)23 8.
12 Coppo R. Basolo B. Martina G ct al. Circulating immune
complexescontaining IgA. IgG and IgM in palients wilh primary
!gAnephropathy and with Henoch Schonlein nephritis. Correialionwith
clinical and histologic signs of activity. J Clin Nephrol
1982;18:230 9,
13 KaulTmann Rll. Herrmann WA. Meyer CJLM et al. CirculatingIgA
immunocomplexcs in Hcnoch Schonlein purpura. Am J Med1980:69:859
65.
-
IgA ANCA in HSP 55
14 Giangiacomo J. Tsai CC. Dermal and glomerular deposition of
IgAin anaphylattoid purpura. Am J Dis Child 1977: 131:981 3.
15 Baart dc la Faille-Kuyper EH. Kater L. Kuijten RH et
aiOcuurrence of vascular IgA deposits in clinically normal skin
ofpalicnts with renal disease. Kidney Int 1976: 9:424 9.
16 Sinniah R. Fend PH. Chen BTM et ai. A clinical and
morphologicalstudy or renal biopsies. Clin Nephroi 1978;
9:219-28,
17 Nakonioto Y. Asamo Y, Dohi K et al. Primary IgA
glomerutoneph-ritis and Schoniein-Henoch purpura nephritis:
clinicopathologicaland immunohistological characteristics, QJ Med
1978:47:495-516,
I Whitworth JA. Lcibowits S. Kennedy MC et al. IgA and
glomerulardisease. Clin Nephrol 1976; 5:33-36,
19 Van der Wall-Bake AWL, Lobatlo S. Joges L et al. IgA
antibodiesagainst cytoplasmic antigens of polymorphonuclear
leukocytes inpatients with Henoch-Schonlcin purpura. Adv Exp Med
Biol 1987:216:1593-8.
20 O'Donoghuc DJ. Nusbaum P. Noel LH PI at.
Antineutrophilcytoplasmic antibodies in IgA nephropathy and
Henoch-Schonleinpurpura- Nephrol Dial Transplani 1992: 7:534 8,
21 Tadros M, Radice A. Pozzi C et al. igA-ANCA in patients
withHenoch-Schonlein purpura (HSP). JASN 1992; 3:666,
22 Esnaut VLM. Jayne DRW. Woetman AP et at. IgG
subclassdistribution and relative functional affinity of
antimyeloperoxidaseantibodies in systemic vaculitis at presentation
and during foUow-up. Immunology 1991: 74:714-8.
1^ Wright DG. Bralove DA. Gallinn JI. The differential
mobilization of
human neutrophil granules: effects of phorbol myristate acetate
andionophorcA23l87. Am J Pathol 1977; 87:273 83,
24 Laemnii UK. Cleavage of structural proteins during the
assembly ofIhe head of bacleriophage T4, Nature 1970; 227:6S0
5.
25 Mills JA, Michel BA. Bloch DA et al. The American College
ofRheumatology 1990 criteria for ihe classification of
Hcnoch-Schonlein purpura. Arthritis Rheum 1990; 33:1114 21,
26 Woodworth TG. Abvclo JG. Austin HA et ai. Severe
glomerulo-nephritis with late emergence of classic Wegener's
granulomatosis.Medicine 1987:66:181-91.
27 Harrison DJ, Kharabanda R. Autoantibodies to neutrophit
cyto-plasm antigens in systemic vasculitis have the same target
specificity.J Pathnl 1989; 158:233-238.
28 Coppo R. Amore A, Gianogho B et iil. IgA-ANCA in children
withHenoch-Schonlein (HSP) and primary IgA nephropathy (pIgAN).JASN
1992; 3:652.
29 Weissmann G. Pearlstein E. Perez HD ct at. Neutrophils
synthesizeand deposit libroncctin on surfaces to which they attach.
TransAssoc Am Phys 1986; 93:72-84,
30 Cederholm B. Wieslander J. Bygren P et al. Circulating
complexescontaining IgA and fihroncctin in patients with primary
IgAnephropathy, Proc Natl Acad Sci USA 1988: 85:4X65 8.
31 Shaw G. Ronda N. Bevan JS et al. Antineiilrophil
cytuplasmicantibodies (ANCA) of IgA class correlate with disease
activity inadult Henoch-Schonlein purpura. Nephrol Dialysis Transpl
1992;7:1238-41,