4r Clin Pathol 1994;47:4-8

ACP Broadsheet No 143 Jfanuary 1994

Detection of autoantibodies to neutrophilcytoplasmic antigens

R J Lock

IntroductionIndirect immunofluorescence remains thegold standard method for the detection ofantineutrophil cytoplasm antibody (ANCA).A recommended method for detection ofANCA by indirect immunofluorescence was

described in the First International Workshopon ANCA in 1988 and the approachdescribed here is based on that method.'Increasingly, solid phase assays (enzymelinked immunosorbent assay (ELISA),2 radio-immunoassay') are being applied for the defi-nition of reactions to specific antigens and forthe potential improvement in quantification.Several commercial assays are available forantibodies to myeloperoxidase, proteinase IIIand a granule extracts. However, standardisa-tion of these solid phase assays remains a

problem.2

Clinical correlationsUntil the discovery of this group of anti-bodies, there was no specific laboratorymethod for the investigation of systemicnecrotising vasculitis, other than histologicalexamination of biopsy material. In 1982Davies et al described ANCA in a small group

of patients with a necrotising glomerulo-nephritis.4 In 1985 this cytoplasmic stainingpattern (classic, c-ANCA) was recognised as

This Broadsheet has beenprepared by the author at theinvitation of the Association ofClinical Pathologists whoreserve the copyright. Furthercopies of this Broadsheet maybe obtainedfrom thePublishing Manager, 3rournalof Clinical Pathology, BMAHouse, Tavistock Square,London WClH 93R

Regional ImmunologyService, BloodServices South West,Southmead Road,Bristol, BSlO SNDR J LockAccepted for publication29 July 1993

Table 1 Major disease categories associated with the threemajor staining patterns

ANCA pattern Main clinical groups presenting

c-ANCA Wegener's granulomatosisMicroscopic polyarteritisChurg-Strauss syndrome

p-ANCA Microscopic polyarteritisClassic polyarteritis nodosaChurg-Strauss syndromeIdiopathic crescentic

glomerulonephritisSystemic lupus erythematosusRheumatoid arthritis

Atypical p-ANCA Ulcerative colitisCrohn's diseasePrimary sclerosing cholangitisAutoimmune hepatitisPrimary biliary cirrhosisSystemic lupus erythematosusRheumatoid arthritis

being associated with Wegener's granulo-matosis.5 In 1988 a second staining pattern(perinuclear, p-ANCA) was recognised inrenal patients with systemic vasculitis.6As well as proving useful in the diagnosis

of the necrotising vasculitides, ANCAs havebeen used to monitor disease activity.5'7The spectrum of ANCA related disorders

was further extended with the description of athird staining pattern associated with ulcera-tive colitis89 and later autoimmune hepatitis.'0This third pattern is referred to as atypical p-ANCA, x-ANCA or "snow drift" pattern. p-ANCA has also been reported in someconnective tissue disorders." The c-ANCApattern has retained a reputation for high sen-sitivity and specificity for necrotising vasculi-tis, but p-ANCA is much less specific. Severalgroups have described ANCA positivity ininfectious disorders.'2 13 Bacterial infectionmay be a particular problem.14The major associations of ANCA with

disease are summarised in table 1.

Indirect immunofluorescence assayThe assay described below is an immuno-fluorescence assay for the detection of IgGclass ANCA. In certain circumstances it maybe desirable to look for other classes of anti-body (IgA ANCA in Henoch Schonlein pur-pura,"4 IgM ANCA in Kawasaki disease.'5 Inthese cases the appropriate second antibodyshould be substituted. The reagents andequipment required are summarised in tables2 and 3.

PrincipleIn the first instance purified human neutro-phils fixed in ethanol are used as the substratein an indirect immunofluorescence assay.This will permit the detection of c-ANCA. Itwill also detect nuclear and perinuclear stains.This may be caused by (i) p-ANCA, (ii) anti-nuclear antibody (ANA), (iii) granulocytespecific ANA (GS-ANA) or a combination of(i) and (ii). Many sera that have p-ANCAhave a concurrent ANA. In our laboratoryabout 30% of p-ANCA sera fall into this

4

Detection of autoantibodies to neutrophil cytoplasmic antigens

Table 2 Reagents used in indirect immunofluorescenceassay

(1)(2)(3)

(4)

(5)(6)

(7)

(8)

(9)

(10)

(1 1)

Polymorphoprep (Nycomed)Desiccant: silica gel (BDH)PBS: dissolve one tablet (Oxoid) in 100 ml distilled

water, pH7-2-7-4PBS-BSA: pipette 3 ml BSA (Boseral, Organon

Technica) into 100 ml PBS, pH to 7-2-7-4Ethanol fixative: absolute alcohol 100 (Hayman Ltd)Formalin fixative: mix 36 ml formaldehyde, 180 ml ace-

tone (both BDH) and 184 ml PBSLysing buffer: 8-32 g NH4C1 plus 0-84 g NaHCO3 (both

BDH) made to 1 litre with distilled waterFluorescein isothiocyanate (FITC) conjugated

antihuman IgG (y chain) (Dako Ltd) diluted in PBSActual dilution varies with batch, typically 1 in200-1 in 500

FITC conjugated antihuman IgA (a chain) (Dako),dilution as for (8) above

FITC conjugated antihuman IgM (a chain) (DakoLtd), dilution as for (8) above

Mountant: dissolve 2-5 g 1, 4-diazobicyclo[2,2,2]octane (DABCO; BDH) in 56°C water bathAdd 10 ml PBS and adjust to pH8-6

Table 3 Equipment used in indirect immunofluorescenceassay

(1)(2)(3)(4)

(5)(6)

Cytocentrifuge (Cytospin 2; Shandon)Fluorescence microscope (Labophot 2; Nikon)Bench centrifuge (Centra 8; IEC)37°C incubator (alternatively, a small drying oven set

to 37°C may be used)57°C water bathMagnetic stirrer (optional)

group. It is obviously critical to distinguishamong these various patterns. This isachieved by further investigation using forma-lin fixed neutrophils and by using an alterna-tive substrate for ANA, either rat tissue orHEp2 cells. p-ANCA, an artefactual patterncaused by charge dependent migration ofgranule components to the nucleus, will pre-sent with a cytoplasmic pattern on formalinfixed neutrophils. GS-ANA and ANA will benegative. ANA will be positive against rat orHEp2 cells, whereas GS-ANA and p-ANCAwill be negative. This approach is sum-marised in table 4.

Specimen collection and preparationSerum samples are preferred. These shouldbe stored at 4°C for up to 72 hours. If seraare to be stored for more than 72 hours

Table 4 Cross table of differential staining patterns seen on indirect immunofluorescence

Substrate

Ethanol Formalin Rat tissuefixed fixed or HEp2

Antibodies neutrophils neutrophils cells

c-ANCA Cytoplasmic Cytoplasmic Nonep-ANCA Nuclear/perinuclear Cytoplasmic NoneANA Nuclear/perinuclear None NuclearGS-ANA Nuclear/perinuclear None Nonep-ANCA + ANA Nuclear/perinuclear Cytoplasmic Nuclear

before analysis then storage at - 20°C or- 70°C is preferable. Alternatively, sodiumazide to a final concentration of 0 01 g/l maybe added to samples stored at 4°C when thesera will be stable for several weeks. Heatinactivated sera should not be used as bothfalse positive and false negative results canoccur. Lipaemic samples and those contain-ing particulate matter should be cleared bycentrifugation before testing.

Control seraControl sera for p-ANCA and c-ANCAshould be titred on ethanol fixed slides witheach batch of patient samples (see below). Anegative control serum should also beincluded. At the First InternationalWorkshop a c-ANCA reference serum (titre1/320) was made available.16 This is nowavailable from the Laboratory of Auto-immune Serology, Statens Seruminstitut,Copenhagen, Denmark. Commercial controlsfor both c-ANCA and p-ANCA are widelyavailable.

In our experience the between batch coeffi-cients of titre variation are 4-5% for c-ANCAand 6-6% for p-ANCA. Membership of anappropriate quality assurance scheme is rec-ommended (for example, the UnitedKingdom External Quality AssuranceScheme).

Titration on formalin fixed slides is unreli-able and is not recommended. For thepurpose of defining patterns of reactivity,controls are used at a dilution of 1 in 10.

Preparation of neutrophil suspensionNeutrophils are separated from fresh antico-agulated (heparin) blood by density gradientsedimentation.

(1) 3-5-5 0 ml of anticoagulated bloodare carefully layered on to 3-5 ml Polymor-phoprep (Nycomed) and centrifuged at450-500 x g for 30 minutes.

(2) The lower leucocyte band, containingpolymorphonuclear cells, is harvested andsuspended in phosphate buffered salinebovine serum albumin (PBS-BSA) to a finalvolume of 10 ml.

(3) The suspension is centrifuged at 500x g for 10 minutes and the supernatant fluiddiscarded.

(4) Lysing solution is added, mixed gen-tly, and incubated at room temperature fortwo minutes. This will remove any contami-nating red cells.

(5) The suspension is centrifuged as in(3) (above) and the pellet washed twice in10 ml (PBS-BSA).

(6) The pellet is resuspended in PBS-BSA and the leucocyte count adjusted to0.1 x 109/I.

Preparation ofneutrophil substrateslidesUsing 300 ,1 of the above suspension, makecytocentrifuge preparations according to the

5

Lock

manufacturer's instructions. The cytocen-trifuge is run at 1500 rpm for three minutes.Slides should be rapidly air dried to preservecell morphology.A modified cytospin method using clotted

blood for the preparation of substrate slideshas been described.'7 18 Neutrophils areallowed to adhere passively to the slide andare then spread by centrifugation to improvevisualisation of the staining patterns. Thismethod permits the relatively rapid produc-tion of small batches of slides.

Ethanol fixationSlides are immersed in absolute ethanol, pre-viously cooled to 4°C, for 15 minutes andthen rapidly air dried.

Formalin fixationSlides are immersed in formalin fixative forexactly 50 seconds, drained, and transferredto absolute ethanol as above.

All slides are stored with desiccant at- 400C.

Immunofluorescence screenAll reagents and samples are allowed to reachroom temperature and mixed thoroughlybefore use.

(1) Remove ethanol fixed slides from thefreezer and rapidly transfer to 37°C incubatorfor 15 minutes to dry. It is important to avoidcondensation of moisture on to the slides asneutrophil morphology may be distorted.

(2) Prepare 1 in 10 dilutions of test andcontrol sera by dilution in PBS. Note, Wiik'prefers the use of sera diluted 1 in 20 toreduce background, but we have not foundthis to be a problem with the Dako conju-gates.

(3) Apply 100 ,ul of diluted sera to eachwell on the slides (sufficient to cover the neu-trophil preparation).

(4) Incubate in a moist box for 20 min-utes at room temperature. At no time duringthe following procedure should the preparations beallowed to dry out.

(5) Gently rinse the slides with PBS froma wash bottle, taking care not to disturb theneutrophils with the jet.

(6) Wash the slides by immersion in PBSfor 20 minutes. The PBS should be agitatedduring this time, preferably using a magneticstirrer.

(7) Remove the slides from the PBS oneat a time. Drain excess fluid from each slideand add 100 pl of diluted conjugate (Dako)to each well.

(8) Incubate in a moist box for 20 min-utes at room temperature.

(9) Wash as in (5) and (6) (above).Remove excess fluid and apply mountant.Seal under coverslips with nail varnish.

(10) Examine by indirect ultraviolet illu-mination with standard fluorescein isothio-cyanate filters, using 40 x air (or waterimmersion) objective and x 10 eyepieces.

Immunofluorescence titre forquantificationAll titrations should be performed on ethanolfixed slides. Titration on formalin fixed neu-trophils is unreliable and is not recom-mended.

Serial twofold dilutions in PBS are made ofthe initial 1 in 10 dilution, to give dilutions of1 in 20, 40, 80, 160 and 320. Where p-ANCA titres have a concurrent ANA thenthese dilutions should also be used to titre theANA. In this case the p-ANCA titre is validonly if it exceeds the ANA titre by at leasttwo dilution steps.

Verification ofp-ANCA on formalinfixed neutrophilsThe method is the same as for the immuno-fluorescence screen above, but substitutingformalin fixed slides for ethanol fixed slides.

Interpretation of resultsOn ethanol fixed neutrophils c-ANCA stainsin a granular pattern with a central accent-uation (fig 1). On formalin fixed slides thegranularity may be coarser and the centralaccentuation is less obvious (fig 2). Eosino-phils and lymphocytes, when present, do notstain.On ethanol fixed slides p-ANCA usually

shows as a sharply delineated perinuclearstain (fig 3). Note that with very strongly pos-itive sera, the antigens may locate to adjacentcells, mimicking the lymphocyte stain seen infig 5. On formalin fixed slides, p-ANCAreverts to cytoplasmic stain, similar to thatseen for c-ANCA (fig 4). In contrast, ANAstain (fig 5) is abolished by formalin fixation.

There are several other staining patternsthat can be detected, the most commonlyseen being the atypical p-ANCA associatedwith serum from inflammatory bowel diseasepatients (fig 6). Atypical c-ANCA patternsmay be seen also in sera from patients withinflammatory bowel disease and some con-nective tissue disease patients (fig 7). Thispattern is sometimes referred to as "flat" c-ANCA, emphasising that there is no centralaccentuation.

Atypical c-ANCA patterns may be seenalso in some infectious states, although theseare usually of low titre (fig 8).

Solid phase assaysMany solid phase assays for the detection ofANCA have been described and the use ofthese assays is likely to increase. Their theo-retical advantages over immunofluorescenceare: firstly, that they could be fully quanti-tative, which would offer a better tool formonitoring therapy; and secondly, thatdefined reactivity with specific antigens couldimprove the disease specificity of the assay.The major antigens associated with ANCA

have been well described. The major c-ANCA antigen is proteinase III.19 In thenecrotising vasculitides the major p-ANCA

6

Detection of autoantibodies to neutrophil cytoplasmic antigens

Figure 1 c-ANCA onneutrophils fixed in ethanolin serum sample from apatient with Wegener'sgranulomatosis. Notecentral accentuationbetween the lobes of theneutrophils.

Figure 2 c-ANCA onformalin fixed neutrophilsfrom same serum sampleshown in fig 1.Granularity is muchcoarser and centralaccentuation less obvious.

Figure 3 p-ANCA onneutrophils fixed in ethanolin serum sample frompatient with microscopicpolyarteritis.

Figure 4 p-ANCA onformalin fixed neutrophilsfrom same serum sample asshown in fig 3. Staining issimilar to that seen in fig2, often with afinergranularity.

Figure 5 Perinuclearanti-nuclear antibody onneutrophils fixed in ethanolin serum sample frompatient with connectivetissue disease. Serum didnot stain formalin fixedneutrophils but stronglystained in a perinuclearpattern for nuclei in rattissue. Note positive"dumbell" nucleus ofaneosinophil left of centre andthe two brightly positivelymphocytes.

Figure 6 Atypicalp-ANCA on neutrophilsfixed in ethanol in serumsample from a patient withulcerative colitis. Notebroad nuclear outline withpoor definition. Anunstained eosinophil can beseen top centre.

Figure 7 Atypical('flat") c-ANCA onneutrophils fixed in ethanolin serum sample from apatient with ulcerativecolitis. Stain is very finelygranular with no centralaccentuation.

Figure 8 Atypicalc-ANCA showing weakfine granular stain withlittle or no centralaccentuation in serumsample from a patient withparvovirus infection.A perinuclear "halo" isvisible on some cells. Theseparvovirus antibodiestypically are low titre andtransient.

antigen is myeloperoxidase.6 Other antigensassociated with the p-ANCA pattern includeelastase, lactoferrin, and cathepsin G.

There is little evidence to suggest thatELISAs that use crude neutrophil extract as a

substrate offer any improvement in sensitivityor specificity over immunofluorescence.Assays for antibodies to the individual anti-gens, however, offer high specificity, and formyeloperoxidase and lactoferrin, good repro-ducibility.'Many of the relevant p-ANCA anti-

gens (myeloperoxidase, lactoferrin, elastase,cathepsin G, fi glucuronidase, lysozyme) are

available from commercial sources and maybe used in standard ELISA systems, with theantigen coated directly on to the plate.Purified proteinase III has recently been

made available commercially. Kits for anti-myeloperoxidase and proteinase III are avail-able. Kits for anti-a granule antibodies areavailable, but may contain antigens otherthan proteinase III.

My thanks are due to Drs Tim Wallington and RebeccaMann for their helpful comments in the production of thismanuscript and to Arlene Cordwell for help with the photo-graphy. Thanks are also due to Dako Ltd, High Wycombe fortheir support of the colour reproductions.

1 Wiik A. Delineation of a standard procedure for indirectimmunofluorescence detection of ANCA. Acta PatholMicrobiol Immunol Scand 1989;97(Suppl 6):1L2-13.

2 Hagen EC, Andrassy K, Chemok E, et al. The value ofindirect immunofluorescence and solid phase tech-niques for ANCA detection. A report on the first phaseof an international cooperative study on the standardiza-tion of ANCA assays. J Immunol Methods 1993;159:1-16.

3 Savage COS, Winearls CG, Jones S, Marshall PD,

7

Lock

Lockwood CM. Prospective study of radioimmunoassayfor antibodies against neutrophil cytoplasm in diagnosisof systemic vasculitis. Lancet 1987;i: 1389-93.

4 Davies DJ, Maran JE, Niall JF, Ryan GB. Segmentalnecrotizing glomerulonephritis with antineutrophil anti-body: possible arbovirus aetiology? Br Med J 1982;285:606.

5 van der Woude FJ, Rasmussen N, Lobatto S, et al. Auto-antibodies against neutrophils and monocytes: tool fordiagnosis and marker for disease activity in Wegener'sgranulomatosis. Lancet 1985;i:425-9.

6 Falk RJ, Jennette CJ. Anti-neutrophil cytoplasmic auto-antibodies with specificity for myeloperoxidase inpatients with systemic vasculitis and idiopathic necro-tizing and crescentic glomerulonephritis. N Engl J Med1988;318:1651-7.

7 Egner W, Chapel HM. Titration of antibodies againstneutrophil cytoplasmic antigens is useful in monitoringdisease activity in systemic vasculitides. Clin ExpImmunol 1990;82:244-9.

8 Saxon A, Shanahan F, Landers C, Ganz T, Targan S. Adistinct subset of antineutrophil cytoplasmic antibodiesis associated with inflammatory bowel disease. J AllergyClin Immunol 1990;86:202-10.

9 Seibold F, Weber P, Klein R, Berg P, Wiedemann K.Clinical significance of antibodies against neutrophils inpatients with inflammatory bowel disease and primarysclerosing cholangitis. Gut 1992;33:657-62.

10 Snook JA, Chapman RW, Fleming K, Jewell DP. Anti-neutrophil nuclear antibody in ulcerative colitis,Crohn's disease and primary sclerosing cholangitis. Clin

Exp Immunol 1989;76:30-3.11 Lesavre P. Antineutrophil cytoplasmic autoantibodies

antigen specificity. Am Y Kid Dis 1991;18:159-63.12 De Clerk LS, Van Offel JF, Smolders WA, et al. Pitfalls

with anti-neutrophil cytoplasmic antibodies (ANCA).Clin Rheumatol 1989;8:512-6.

13 Koderisch J, Andrassy K, Rasmussen N, Hartman M,Tiglen W. "False-positive" anti-neutrophil cytoplasmicantibodies in HIV infection. Lancet 1990;i: 1227-8.

14 van den Wall Bake AWL, Lobatto S, Jonges L, Daha MR,van Es LA. IgA antibodies directed against cytoplasmicantigens of polymorphonuclear leucocytes in patientswith Henoch-Schonlein purpura. Adv Exp Med Biol1987;216b: 1593-8.

15 Savage COS, Tizard EJ, Jayne D, Lockwood CM, DillonMJ. Antineutrophil cytoplasm antibodies in Kawasakidisease. Arch Dis Child 1989;6:360-3.

16 The international serum standard of anti-neutrophilcytoplasm antibodies (ANCA) according to the 1 stinternational workshop on ANCA, 1988. Acta PatholMicrobiol Immunol Scand 1989;97 (Suppl 6):30.

17 Roberts DE, Peebles C, Daggett R. Simplified method ofpreparing neutrophil slides to examine antibodies tocytoplasmic antigens. J Clin Pathol 1990;43:83-4.

18 Roberts DE, Rubin RL. Anti-neutrophil cytoplasmicautoantibodies. In: Rose NR, ed. Manual of clinicallaboratory immunology. 4th edn. Washington, DC:American Society for Microbiology, 1992:781-4.

19 Ludemann J, Utect B, Gross WL. Anti-neutrophil cyto-plasm antibodies in Wegener's granulomatosis recognizean elastinolytic enzyme. J Exp Med 1990;171:357-62.

8