In Vitro Studies in : Diagnostic Value of a Dual Parameter Analysis

M ary von Blo mberg-van der Flier, M.Se., Cees K. H . van der Burg , M.D., Odette Pos, B.Se. , Ell a M . van de Pl assche-Boers, M .D ., D erk P. Bruynzeel, M.D., Ph.D., Gianni Garotta, Ph.D., and Rik J Scheper, Ph .D. Departlllellts of Patho logy (M vI3 -vd F, 0 1' , E M vd l'-B, RJ S) and Dermatology (CKH vdB, DPJ3), Free Uni ve rsity Hospital, Amsterdam, T he N etherbnds, and Centra l Research Units, F Hoffm ann- La Roche & Co., Ltd. (GG), Basel , Switze rl and

A compari son was m ad e be tween the diagn osti c value o f test reacti o ns . A ccurate d ete rmination of MIF becam e fea­ assaying nickel-induced ly mphocyte pro life ra ti o n (l y m­ sible b y using cell s fro m the human monocytoid cell line phocyte trans fo rmatio n test, L TT) and mig rati o n inhibi­ U 937 as ta rget cell s in a mic ro drople t agarose assay . Using tio n facto r (MIF) pro ducti o n in nickel contact sen sitivity. this MMIT, positive reacti o ns o ccurred in 13% of the healthy Altho u g h ly mphocyte pro life ratio n was sig nificantl y in­ contro ls and fa lse-negative reacti o n s w e re fo und in 26% o f creased in the g ro up of pa ti ents w ith skin test reactivity to pati ents w ith p ositive skin test reacti vity to nickel. nickel, positive L T T w ere also frequently fo und in skin A s LTT and MMIT data appeared to be only w eak ly test-negati ve s ubjects: in 63% o f subjects with and in 30 % correlated in the individua ls tested , a dual parameter an al­ 8 of subjects w itho ut a histo r y o f m e tal a llergy. T his would ysis was perfo rmed . An excell ent correlation [p = 1.8 (1 0 - ) ] limit the value o f the L TT as a n in vi tro correlate of skin w as fo und be tween s kin test and in vitro reactivity for test reacti v ity. H o w ever, in certain pati ents positive lym­ individua ls with m atching in vitro results (60% of all in­ phocyte tra nsformati o n may re veal nicke l sen siti zatio n at dividuals teste d). In those individuals with discordant i.n a time of undetectable skin reacti vity. v itro data, skin testing will rem ain indispen sable for di­ Data o b tained w ith the m acro phage mig ra tio n inhibitio n agnosing nickel a ll erg y. J filli es! D erwa!o/ 88:362- 368, 1987 test (MMIT) sh owed a good correlatio n with nickel patch

in ce nickel is the most common contact sensiti zer in theless, the individual diagnosti c va lue of this assay was limited, humans, ava il abi lity of di ag '.lOs ti c in vitro tes ts for ni ckel parti cularl y by a hi gh in cidcnce o f false-positi ve rcacti o ns. all ergy would be of g reat impo rta nce. Data o btain ed As antigen-induccd Iympho kine (LK) relcase represents a cru­ from the li terature regard in g the usc of lymphocyte cial stcp in the develo pmcnt o ftypc IV hypersensitivity reactions, transformation tes ts (LT T ) fo r this purpose, arc listed in one would ex pect assays measuring ni ckel-induced LK produc­ TSabl e I /1 - 151. Stati sti ca l differences between g ro ups o f ni ckcl­ ti on to be most suitabl e as diagnosti c tes ts in nickel all ergy. [t has sensitive pa ti ents and hea lthy contro ls were usuall y fo und. N ever- indeed been repo rted that lymphocytes from nickcl-allergic pa­ ti ents m ay produce leuk ocyte migrati o n inhibition factor (L1F) in vitro in the presencc of nickel-protcin conjugates [1 6, ·[ 7] or free ni ckel sa lts [1 8-20 1. Fro m T ablc [[ it is clear that no consistent results were o btained. This might be due to thc usc o fullseparated Man uscript received M ay 19, 1986; accepted for publi cation Septem ber bu ffy coats fo r simultaneous LlF production and mig ration (1 - 23, 1986. phase leucocyte migrati o n inhibition tes t, LM[T). In addition, This work was supported in part by Pracventicfonds, T he N etherl ands. the sho rt culture peri ods (up to 1 day) in these assays may not Reprint requests to: M ary von Blo mberg- va n der Fli er, Depa rtment o f have becn suffi cient fo r adequate ni ckcl prescntatio n by accessory Pathology, Free University Hospi ta l, De Boelela an 11 17, 1007 ME Am­ ce ll s, Iympho kinc release by effector cc ll s, and inhibition o f mi­ sterdam, T he Netherlands. g ratin g ta rget cell s. A bbreviatioll s: Con A: concanavalin A [n th e present study, therefo re, w e adaptcd a 2-phase protocol DTH : delayed-type hypersensiti vi ty in w hi ch peripheral bl ood lymphocytes (PBL) w ere precultured FCS: feta l ca lf serUIll to all ow fo r m acro phage mig rati o n inhibitio n facto r (MIF) pro­ L1 F: leukocyte migration inhi bition factor ducti on. Subsequentl y, MIF activity was measurcd on a ho m og­ LK: Iymphokine enous populatio n o f ta rgct cell s. In preliminary experimcnts 3 LMIT: leukocyte migrati on inhibition test(s) presently ava ilable hum.an m yelo/monocyti c ccll lin es (KG-l (22], LTT: lym phocyte transfo rmatio n tes t(s) HL60 r231, and U 937 r24,25]) w ere screened fo r M[F sensitivity. MI : m igration inhibitio n U937 ce ll s, found to provide o ptimal targets, w ere then used to MI F: macrophage migration inhibition fac tor eva lu ate the diagnosti c value o f a 2-phase macrophage mig rati on MMIT: macro phage migration inh ib ition test(s) PEL: peri pheral blood Iymphocyte(s) inhibition tes t (MM IT ) in ni ckel all ergy. The MMIT data w ere PBS: phosphate-buffe red saline compared w ith lymphocyte transfo rmatio n data as to their cor­ PM N : polymorphonuclea r relati on w ith skin test reactivity. [n addition, considering that SI: stimulatio n index both in vitro parameters (MIF producti on and lymphocyte pro-

0022-202X/87/S03.S0 Copyrig ht © 1987 by T he Society for In vestigative Dermatology. In c.

362 VOL. 88. NO. 4 APRIL 1987 IN V ITRO TESTS IN NICKEL ALLERGY 363

Table I. Lymphocyte Transformati on in the Diagnosis of Nickel All ergy-Review of Literature

Results

Patients Control Group In Vitro Test % False % False Year First Author Reference Antigen Concentration" N i> N egative' N i> Positived Si gn'

1962 Aspegren [1] NiCIz 100 /L M 15 0 15 100 1970 Pappas [21 NiAc 112 '/LM 3 0 8 100 1970 M acLeod [3] NiSO. 50 /L M 12 17 14 0 + 1972 Form an 14] NiSO. 18 II 12 66 + 1972 Hutchinson 151 NiSO./ NiAc 50/LM 8 25 7 14 + 1973 Mill ikan [6 1 NiCL2 40 /L M 8 0 6 0 + 1975 G imenez 17] NiSO. 0.65/65 /L M 25 0 10 0 + 1976 Kim 181 NiSO./ NiAc 40 /L M 21 0 17 < 10 + 1978 Svejgaard 19] NiSO. 20 /L M 8 13 9 11 + 1979 Veien 110 1 NiSO. 36/72 /L M 8 13 4 25 + 1980 S i Iv en noincn-Kassi nen 111 J NiSO" 8/40 /L M 35 II 49 20 + 1981 AI-Tawil 11 2] NiSO, 24/LM 13 15 9 22 + 1982 MacLeod 11 3J NiSO., 50 /L M 6 33 6 33 1984 Nordlind 11 4] NiSO" 20 /L M 10 0 10 40 + 1985 AI -Tawil [1 5] NiSO" 80 /LM 41 15 34 32 + 1987 Blomberg IThis pa pcrl NiSO., 7- 80 /L M 27 7 23 39 +

"Nickel concentrations showin g :111 optimal difference between patient and co ntrol g roups. bNull1ber of individuals tested . 'Stimulati on ind ex (5 1) :s 2 in nickel-allergic patients. "51 > 2 in health y controls. 'Statisti cal differences between groups of nickel-a ll ergic pa tients and health y cont rols : +.

liferation) relate to different phases of the cell-mediated immune Skin Tests Patch rests were performed w ith 5% NiS0 4 in response (representing effector and memory functions, respec­ petrolatum. Patches were removed after 48 h and the reactions tively), a dual parameter analysis was performed. were read 15 min later and graded according to Fregert [26]: negati ve (-), weak (nonvesicular) reacti ons (+), and strong MATERIALS AND METHODS (edematous and vesicular) reactions (+ +), or even bullous re­ Donors Twenty-seven ni ckel-all ergic indi viduals showin g pos­ actions (+ + +). itive ni ckel patch tests (13 + and 14 + + or + + +) were ran­ Lymphocyte Transformation Tests Mononuclea r cell s were domly selected from paticnts visiting our Allergy C linic, and from isolated fro l11 heparinized blood by Fi coll-isopaque centrifugati on junior hairdressers w ho had been skin tested in a parall el study prior to or 2-3 weeks after skin testing and cryopreserved. The on occupational all ergy. Thirty-one negative controls were avaiJ­ ce ll s were cultured in round-bottomed microtiter plates (105 cell s ab le from the same groups, including 8 skin test-negati ve indi­ in a final volume of 150 J.LI per well in sixfold) in RPMl 1640, viduals w ith a positive hi story of metal all ergy. Cord blood sam­ supplemented w ith HEPES, antibiotics, and either 20% in acti­ ples, kindly providcd by our Department of Obstetrics, were vated pooled human serum or 20% inactivated autologous serum. used as a source of nonsensitized control lymphocytes. NiSO",61-hO (Merck, purity > 99% ) was used as antigen in the

Table II. LlF/ MIF Production in the Diagnosis of Nickel All ergy-Review of Literature

Patients Control Groups In Vitro Test First % False % False Y ear Autho r Reference Assay Antigen C oncentration" N " Negati ve' N" Positived Sig n'

1975 Mirza [1 6] I-phase LMIT NiSO. 320 /L M II 73 6 17 Ni-BSA 294 /L M 15 0 13 31 + 1976 T hulin 11 7] I-phase LMIT NiC I2 844 /L M 12 58 II 55 Ni-HSA 844 /LM 10 30 9 22 + 1976 Jo rdan [1 8] I-phase LMIT NiSO" 468 /L M 10 0 10 0 + 1976 M acLeod rt 91 I-phase LMIT NiSO" 50 /L M 10 10 2-phasc LM IT N iSO" 50 /L M 3 3 1977 Swoboda 1201 I-phase LMIT NiSO" 400 /LM 21 14 21 10 + 1983 N o rdlind 121] I-phase LMIT NiSO " 36 /LM 10 100 10 0 1984 Blo mberg [T his paper] 2-phase MMIT NiSO" 7-40 /LM 27 26r 23 13r +

LIF = leuk ocyte mig ration inhibition f., cto r. M IF = ma cro pha ge mig ratio n in hibitio n f., cto r. LM IT = leukocy te migration inhibition test. MM IT = ma crophage migration inhi bition test. BSA = bo vine serum album in . H S A = 11ll rnan serum :1 lbumin. "' Nickel concentratio ns sho w in g :t il optimal difference between patient and control g roups. bNulllbcr o f individua ls tested. I" Migrati on inhibition < 20% in nickel-a llergic patients. "Mi gration in hibition 2: 20% in health y controls. 'Statisti ca l differen ces between groups of nickel-a ll ergic patients and healthy controls: +. f Mig ra ri oll inhibition > 10% lco nsid crcd as positive. 364 VON I3LO MI3EnG-VAN DEn FLIER lOT AL THE JOUnNAL OF INVESTIGATIVE DEnMATOLOGY

culturcs in fina l conccntrations of 0, 7, 14, 40, and 80 JL M. At day

6 rJH]-thymidine uptakc (1 JL C i/cuiturc with sp act of5 Ci/mm ol, 51 ') in 5 h) was mcasured usin g liquid scintillation counting. Nickel­ . " . " specifi c stimulation was expresscd as t · n

. (mcan cpm w ith NiS04) Stimulation IJ1d cx (S I) = ---''-----'------'!....­ 14 . 14 (mean cpm w ithout NiSO,,) · Macrophage Migration Inhibition Test M ononuclear cells wcrc isolated as describcd abovc, and culturcd in conica l 15-ml 12 12 6 . tubes [2.5 (10 - ) ) in ] ml) in RPMI 1640, supplemcntcd with HE PES, antibio ti cs, and 10% fetal ca lf scrum (FCS). To the · 10 10 culturcs NiS04·6H 20 was added to g ive fin al concentrations of · 0, 7, 14, and 40 JLM. The MI F-containing supernatants for testin g 0 · · va ri ous target cel l lin es were prepared by addin g conca navalin A 0 · (Con A) (3 JL g/ ml) at 24 h of culture. Supern ata nts (10 min, 2000 8 r:- 8 g) were routincly harvested after 3 days ofculturc (3rC, 5% CO2 in air) and frozcn at -20°C. For assayin g MIF activity, super­ natants were d ilutcd 1:2 with fresh culture medium containing 6 · 6 10% FCS. · In order to identify the most suitable ta rgct ce ll s for mig ration · 'b · inhibition, the conventional capillary method was used [27). Three 4 0 r-· 4 human cel l lin cs werc tes tcd: KG-l , a myeloblast leukemia lin e of g ranulocyti c nature [22]; HL-60, a mo no myelocytic cell lin e , ·0o·. · ca pable of diffcrentiation to either g ranulocytes or macrophages :2 )-· · 2 [23]; and U 937, a relative ly well-differentiated monocyte-like cell line [251· · · Cell s were harvested in log-phase growth and washed twicc i"l ~ before setting up the mig ration assays. cord blood PBl PBl PBl For patient studies, the U937 cell line was selected and a fast IymPhocyt~ agarose-dro pl et as&ay was adapted according to T hurman et al Skin test') negative + ++/+++ [281. In brief, cell pellets were resuspended [2(107) ce ll s/ml, 37°C] NiS04 in 0.2% agarose. This was prepared by dissolving 20 mg sea­ plaque agarose of low gell in g temperature (Marine Coll oids) in n 6 3 1 13 14 1 ml phosphate-buffered sa lin e (PBS) at 120°C and di luting it 10 meant sd 1.1 ± 0.2 3.1 ± 3.3 6.4 ± 7. 8 8.9 ± 6.1 timcs w ith RPMI 1640, 10% FC5, and antibiotics. median 1.1 1.7 3.9 8 .1 From this suspension 1-JLI droplets were placed in fl at- bot­ tomcd micro plates usin g a Hamilton repeatin g dispenser with 3) I II I statistics 0.05-ml gas-tig ht syringe. After solidifyin g (10-20 min , 4°C) the P < O.Ol P < 0.0 5 droplets were overl aid w ith 100 JLI test fluid (a ll tests in fivefo ld , thc o uter well s were not used). Antigen-stimulated supernatants Figure 1. Ni ckel-specifi c lymphocyte transform ation in relation to skin and their contro ls were always tested in the sa me microtiter plate test res ul ts. '/) . Maximum SI from LTT with 7-80 J.tM NiSO.,. 2), Forty­ in order to reduce test variation. After in cubation of the covered eight hour readin g of patch tes ts with 5% NiSo •. 3), Wilcoxon test. pl ates during 21 h at 37°C w ith 5% CO2 in air, mig ration areas Indi viduals without a history of meta l all ergy arc represcnted by op ell were measured usin g a projection microscope. M IF production circles, patients with a hi story of mct3 1 all ergy by closed circles. was expressed as % migration inhibitio n (MI):

MI = 100 _ 100 X (mean area in cultures with NiSO,,) %

(mean area in cultures without NiS04) frequently observed in the skin test-positive patient groups, lead­ in g to statistical differences with the skin test- negative group 5 Statistics Differences between test res ults in different patient [Fisher's p = 3.9 (10 - ) ] . Moreover, patients with strong (+ + groups were considered to be sig ni ficant w hen p < 0.05 in the or + + +) skin test reactivity showed hi gher in vitro responses non para metrical Wilcoxon tes t. Differences in frequency were to nickel than did patients w ith w ea k (+) skin reactions, sug- evaluated usin g the Fi sher's exact test. For statistica l evaluation the LTT was considered positive w hen 51 > 2 and the MM IT w hen MI > 10%. Table III. Sensitivity of M yelo/M onocytic Cell Lines, RESULTS Human Polymorphonuclear (PMN) Cell s, and Guinea Pig Peritoneal M acrophages to Human Macrophage Migration Nickel-Specific LTT in Relation to Skin Test Results Lym­ Inhibition Factor phocyte transform ation tests were routinely performed in the presence of different concentrations (7-80 JLM ) ofNiSO" in pooled Dilution of Lymphokine homologous serum. Fo r each dono r the m aximum SI from this Preparation That Ma ximal Mi gration Produces 30% of dose-response curve is shown in Fi g 1. The concentrations of Target Ce ll Inhibition (%) Mi gration Inhibition NiS04 at w hich optimal lymphocyte proliferation occurred did no t differ between skin test-positive patients and healthy controls. U937 78 1 :50 A concentration of 40 JLM NiS04 resulted in an SI > 2 in 33 out KG-I 72 1 :26 of 38 LTT respo nders and was optimal in 20 o ut of these 38 HL-60 44 1:32 individuals. When cord blood lymphocytes were used to screen Hum3n PMN cc ll s 40 1 :20 for potential mitogenicity of NiSO'h no in creased thymidine up­ Guinea pi g peritoneal 56 1 :9 macropha gcs take could be detected. In contrast, hi gh stimulatio n indices were VOL. 88, N O. 4 APRI L 1987 IN VITRO TEST S IN NIC KEL ALLERGY 365

I gesting a positive correlati on between nickel patch tes t and L T T Table IV. M acrophage Migration Inhibition Facto r data. Essentiall y similar results were o btain ed w hen autologous Production and Lymphocyte T ransfo rmation in Nickel-Allergic sera were used to supplem ent the culture media (d ata not shown) . Pati ents and H ealthy Controls However, a rem arkabl y high percentage (39%) of skin tes t­ Macroph age negati ve individuals showed significant lymphocyte stimulation Lymphocyte Mi gration in vitro. An importan t contributi on to this high percentage is Pati ent Transformati on Inhibition Test made by the group o f nickel skin tes t-negative individuals w ith Code Skin Tes t" Test (LTT)b (MMIT) ' a positive history of m etal all ergy. Abes 0 4.6 - 15 Sensitivity of Three Leukemic Cell Lines for Human MIF Liu 0 1.7 - 14 Conventional target cell s fo r clinica l LIF and MIF assays (human Worm 0 9.4 - 14 po lymorphonuclea r (PMN) cell s and guinea pig peritoneal m ac­ Wec r 0 1. 0 - 12 rophages, respectively) show considerable dono r-dependent vari­ Yd em 0 1. 4 - 10 ations in bo th their migrato ry capacity and responsiveness to Celi 0 2.6 6 LIF/ MIF. T herefo re, w e investi ga ted w hich of the human leu­ Timm 0 8.8 - 2 kemic ce ll lines-KG1, HL60, and U 937-could be applied in a Hoge 0 1.1 2 Vdlc 1. 2 0 m igration inhibition assay. Fo r this purpose MIF-sensitivity was 0 Go ud 0 2.8 1 evalu ated by testing the res ponsiveness to a con A-induced Iy m­ Hoek 0 1. 4 4 p hokine preparation. From Table III it is cl ea r that the mono­ Lank 0 4.5 4 cytoid cell line U 937 provides very sensiti ve target ce ll s. U sing TO Ll w 0 1.2 4 the U 937 cells in an aga rose microdro plet assay, test vari ati on Tbor 0 1. 8 4 was extrem ely lo w , allowing a migration inhibition exceeding Enge 0 1. 0 4 10% to be taken as significant. Buis 0 1. 4 5 Vge l 0 1.3 6 Nickel- Induced Production ofMIF in Relation to Skin Test Krij 0 1.3 7 Results The capacity o f lymphocytes fro m nickel-allergic pa­ Kemp 0 2.0 8 tients to produce MIF upon stimulati on w ith NiS0 4 was tes ted Mekk 0 1.1 10 using the m o nocytoid cell li ne U 937 as indica tor cell s. T he NiS0 4 La ut 0 1.7 13 concentrati o ns, rangin g from 0- 80 fJ-M , had no effect on either Poor 0 4. 1 17 migration or M IF sensitivity of the U 937 cell line (data not shown). Fo rt 0 1. 2 18

Schr 0" 1. 8 6 Pete 0" 2.5 - 1 Veer 0" 14.2 1 Vdij 0" 1. 0 2 ..- .. Stie 0,1 2.9 3 - O ud 0" 1. 5 3 c: 40 40 Vdme 0,1 2.9 4 ~ A li Sja u 0" 11. 5 6 :2 .5 30 30 c: • 32 3 0 Vers + :.- Kwa m + 4.7 7 ~ 01 • • Jagn + 2.9 4 'E 70 20 Debi + 2. 1 9 8 • Duri + 2. 1 11 *- 0 Fid a + 9.6 12 • 13 10 0 a 10 T hak + 3.9 • Dhaa + 4.3 13 • • Oost + 6. 1 14 ~~ 2.5 15 ( I o Mulk + 0 16 ( ) Pepe + 2.2 • Brou + 9. 1 16 De Zwan + 1.3 23 -1 0 0 -1 0 0 't9 Wiel ++ 2.9 Has +++ 7.8 2 Bakk ++ 11.2 6 Skin ted) negative 3) + ++/ +++ Stol +++ 17.2 16 NiS04 Dboe ++ 1. 4 15 Bons +++ 24.0 18 n 31 13 14 Pold ++ 5.4 23 Koet ++ 7.0 25 mean±sd 1.3 8.2 11.5 6.1 24.7 ± 15.3 ± ± Albe ++ 2.6 31 median 4 13 24 Eve! +++ 13.8 38 Vd am ++ 10. 1 38 I II I 8.3 39 statistics 4) P

Figure 2 shows the res ul ts of the MMIT, w hen testi ng nickel­ no obvious correlation between the degree of expression of these all ergic pati ents and hea lthy controls. Maximum MI indices, ob­ two ce ll fun ctions was observed when testing different nickel­

tained w ith 7-40 J.LM NiS04 , are given in relati on to the NiS04 specifi c clones [29]. In polyclonal cell suspensions like PBL, there­ patch test reactions. In patch tes t-positive patients more MIF was fore, this would res ult in a large overlap between proliferati ve produced than in patch test-negative individuals [Fisher'S p = and productive fun ctions. It ca n, however, not be excluded that 6 2.2 (10 - )]. M oreover, nickel-induced MIF production positively different pati ents preferentiall y express either o f these cell func­ correlated with the strength of the skin test reaction, as judged ti ons, depending for instance on their recent history of all ergenic by a 2-step statisti ca l analysis. In the skin test-negative individuals contacts. Obviously, technica l factors, li ke the duration of the w ith a positive hi sto ry of metal all ergy the MMIT data completely assay may create an additional di chotomy between results from reflected the lack of sk in reacti vity. tests for these fun ctions. In line with these considerations our With respect to the diagnosti c va lue of the presen t 2-phase present res ults indica te that LTT and MMIT data can show dif­ MMIT in nickel all ergy, our data show that in spite of the positive ferent patterns in the patient groups studied. correlation w ith skin test reactivity, false-negative reacti ons oc­ Although ni ckel-specific LTT correlated well w ith skin test curred in 7/27 ni ckel- hypersensitive donors (26%) and fa lse-pos­ reacti vity in groups of pati ents , the value of the LTT for individual iti ve reactions in 3/23 hea lth y controls (13%). diagnosis of ni ckel allergy was restricted by the hi gh percentage of positive tests found in hea lthy controls. As reported ea rlier by Diagnostic Value of Paired LTT and MMIT Data In Table AI-Tawil, Marc usson, and Mo ll er [1 2], NiS04 concentrations of IV , LTT and MMIT data are li sted fo r all individual donors, 7-80 J.LM were unable to stimulate proliferation of cord blood grouped according to skin test reactivity. Although no quanti­ lymphocytes in 6-day cultures. Using children's blood lympho­ tative correlati on was fo und between LTT and MMIT (,. = 0.16), cytes, known to be more sensitive to mitogens, N ordlind and the all -or-none correlati on was positi ve (F isher's p = 0.01). The Henze [30] observed a weak stimulati on (up to SI = 2. 1) under fact, however, that many individuals (22/58) showed discordant similar culture conditions, suggesting mitogenicity of NiS0 4 . In in vitro res ul ts prompted us to evaluate whether a combination our hands, NiS04 was not mitogenic in cultures of T-helper of LTT and MMIT data would improve th e di agnostic valu e of cl ones selected for nonrelated (malari a) antigens (data not shown). in vitro testing for nickel all ergy. Moreover the majority of hea lthy controls did not react to NiS04 Tabl e V shows the res ults of such a dual parameter analysis. at all. We must therefore conclude that in a substantial number [n this table the in vitro test results in the different patient groups of skin test-negative individuals nickel-sensitized lymphocytes are compared with those in hea lth y controls (both skin tes t and circul ate that proliferate upon contact w ith ni ckel but are not yet history negati ve, n = 23). By usin g 2 in vitro parameters, a better able to generate a clinically manifest all ergic reaction. A follow­ discrimination was o btained between ni ckel skin test- positive up study is now bein g ca rried out to in vestigate the predictive patients and health y controls (Table V) . va lue of the positive nickel LTT for future development of nickel A reliable individual diagnosis, however, could be made only contact . in patients w ith matchin g in vitro results: none of the all ergic As to the performance of the MMIT, it is cl ea r from our res ults patients showed 2 negative in vitro tes ts, whereas onl y 1 indi­ that the monocytoid ce ll lin e U 937 is most sensitive for human vidual in the hea lth y control group had 2 positive in vitro tests. MIF, as compared w ith the HL-60 and KG-1 ce ll lines . T his agrees with recent findin gs of Liu et al [31] who demonstrated the pres­ ence of substantial amounts of MIF- receptor molecul es on the D ISCUSSION U937 cell membrane. The central role of helper T lymphocytes both in the induction Using U937 target ce ll s in the microdroplet agarose MMIT, and elicitation phase of contact hypersensitivity is now becoming the assay is rapid and easy to perform, and requires only small w ell es tablished. All nickel-specifi c clones selected fr om periph­ amounts of blood (1 ml/antigen concentrati on). Reproducible eral blood of nickel-all ergic patients were of the OKT3 +, 4 + , 8 - , res ults are obtain ed, not only with nickel, but also with microbial Leu 7- phenotype 129]. T hese clones not only proliferated upon antigens [32]. Moreover, the 2-phase system all ows lymphocytes contact w ith ni ckel, th ey also produced va rious LK . Interes tingly, to be cryopreserved and MIF supernatants to be stored at - 200 e

Table V. T he Diagnosti c Va lu e of Combining Res ults from Two In Vitro Tests in Nickel All ergy (L TT and MMIT) As Compared to Resul ts from Either Test Al one

Sin g le Para meter Anal ysis Dual Parameter Analysis

LTT Pas LTT andlor LTT N eg Patient Groups LTT Pas" MMIT POSh and MMIT Pas MMIT Pas and MMIT Neg

Skin test nega tive History ncg' n = 23 7/23 3/23 \ /23 9/23 14/23 History pas n = 8 518 0/8 0/8 5/8 3/8 I1. S. d I1 .S. I1.S. n.s. n.s. Total n = 31 12/3 1 3/13 1/31 14/31 17/31 Skin tcst positive" + n = 13 12/13 9/13 8/13 13113 0/13 4 3 4 4 [4(10- ) ] " [10 - ] 13(10 - )1 [2(1 0- ) ] [2(10 - ") ] ++ 1+++ n 14 13/14 11 / 14 10/14 14 /14 0113 4 5 12(10 - ")]" 11 0 - ] [3(10- )1 [1 0- "] [1 0- "] Total 25/27 20/27 18/27 27/27 0/27 7 7 15(10- ")1 [2( 10- S) ] [4(10 - (') ] 19(10- )1 19(10 - ) ]

' Lymphocyte transformati on test (LTT ) was considered positive when stimulation index > 2.0. bMacrophagc mig rat ion jnhibition tcst (MMIT) was consid ered positive w hen mi gration inh ibition > 100/0. 'Histo ry of metal all erg y. ' All p va lues given between brackets arc obtained by comparing the test res ults from the different patient groups with those from skin test-negati ve, history-nega ti ve hC:l lth y contro ls. lI sing th e Fi sher exact test. ' Fo rty-eight hour reading of NiSO, patch tes t reaction. VOL. 88, NO.4 APR IL 1987 IN VIT RO TEST S IN NIC KEL ALLERGY 367

wi thout loss of activity for several months, enabling repea ted 9. Svejgaard E, Morling N , Svejgaard A, Veien NK: Lymphocyte testing and longitudinal studies. transformati on induced by nickel sulphate: an in vitro study of A hi ghl y significant correlation was found between ni ckel-in­ subjects with and w ithout a positive ni ckel patch tes t. Acta Derm duced MIF release and skin tes t rea ctivity to nickel. Regarding Venereo l (S tockh ) 58:245-250, 1978 the extremely low intraexperimental va ri ation in migration o f the 10. Veien NK, Svejgaard E, Menne T: In vitro lymphocyte transfor­ U937 cell s, an inhibition exceedin g 10% should be considered as mation to nickel: a stLId y of nickel sensitive patients before and significant. Usin g this criterion, the MMIT di scriminated better after epicutaneous and oral challenge with ni ckel. Acta Derm Ve­ between skin test-positive and skin test-negative patients than nereol (Stockh) 59:447-451, 1979 did the LTT. 11 . Sil vennoinen-Kassin en S: Lymphocyte transformati on in ni ckel al­ A dual parameter anal ys is, by combining the results of both lergy: amplifi cation ofT lymphocyte res ponses to ni ckel sulphate LTT and MMIT, all ows a reliable diagnosis in about 60% of the by macropha ges in vitro. Sca nd] Immunol 12:6 1-65, 1980 individuals tes ted, ni ckel all ergy bein g either excluded (both in 12. AI-Tawil NG, Marcusson ]A, Moller E: Lymphocyte transforma­ vitro assays negative) or demonstrated (both assays positive) (Ta­ ti on tes t in patients w ith ni ckel sensiti vi ty: an aid to diagnosis. Acta Derm Venereol (Stockh) 61:511 -515, 198 '1 ble V) . However, in 40% of the individuals showin g onl y 1 pOsitive tes t, the in vitro studies did not permit definite conclu­ 13. MacLeod TM, Hutchinson F, RafRe EJ: In vitro bla stogeni c Iym­ sions. phokine activity in ni ckel all ergy. Acta DCI'm Venereol (Stock h) 62:249-250, 1982 Discrepancies between in vivo skin reactivity and in vitro blood lymphocyte res ponses are also encountered with various micro­ 14. N ordlind K: Lymphocyte transformation test in di agnosis of ni ckel all ergy. A comparison between the separation of peripheral blood bial antigens [32] and are probably related to compartmentali­ lymphocytes on Ficoll-paque, Percoll or by gravity sedimentation. za tion of immune res ponsiveness: local retention of nickel-spec ifi c lnt Arch Allergy Appl Im munol 73: 151-154, 1984 ly mphocytes at distinct anatomic sites might cause fa lse nega tive in vitro reactions of PBL. False negative skin reactions on the 15. AI-Tawil N G, Berggren G, Emtcs tam L, Fransson], ] ernselius R, Marcusson ]A: Correlation between quantitati ve in vivo and in Other hand may be due to local or systemic desensitiza tion [33]. vitro res ponses in ni ckel-a llergic patients. Acta Derm Venereol Differences in sensitivity of the in vitro and in vivo assays ma y (Stockh) 65: 385-389, 1985 fu rther add to such discrepancies . In this context it should be 16. Mirza AM, Perera MG, Maccia C A, Dziubynskyj OG, Bern stein stressed that in the present study a routine patch tes t procedure IL : Leuk ocyte migrati on inhibition in ni ckel dermatitis. Int Arch was used with only a sin gle concentration of NiS04 , as recom­ All ergy Appl Immunol 49:782-788, 1975 mended by the IC DRG (International Re­ 17. Thulin H: The leukocyte migrati on test in nickel contact dermatitis. search Group). The vehicle used in this procedure (petrolatum) Ac ta DCI'm Venereol (Stock h) 56:377-380, 1976 was recently suggested not to be optimal for NiSO" skin testing 18. Jordan WI', Dvorak J: Leukocyte migrati on inhibition ass ay (L1 F) [34,35] . in ni ckel contact dermatitis. Arch DcrmatoI 112:1 741-1744, 1976 From our study the following conclusions can be drawn: 19. MacLeod T M, Hutchinson F, RafR e EJ: The leukocyte migration 1. Assessment o f peripheral blood lymphocytes for ni ckel-in­ inhi bition tes t in all ergic ni ckel contact dermatitis. Br] DerI1latol duced MIF release has become feasible by the use of a 2- 94 :63- 64, 1976 phase system with U937 cells as MIF targets. 20. Von Swoboda B, FritzJ , Ludvan M: Nickelallergie und Leukozyten­ 2. The MMIT is less sensitive, but more specifi c, than the LTT Migrations-Hemmung. Dermatol Mona tsschr 163:208-2'12, 1977 for predicting skin reactivity to ni ckel. 21. N ordlind K, Sandberg G: Leukocytes from patients allergic to chro­ 3. Dual parameter analysis, by combining L TT and MMIT mium and ni ckel examined by the scaled ca pi ll ary migrati on tech­ results, all ows a reli able diagnosis in the majority of indi­ ni que. Int Arch Allergy Appl Immunol 70:30- 33, 1983 viduals tes ted for nickel all ergy. 22. KoefRer HI', Gold e DW: Acute myel ogenous leukemia: a human ce ll line res ponsive to colony stimulating activity. Science 200: 11 53-1154,1978 We wish to thallk Sha kllmala SallJpa t- Sardjoeperslld alld Barbara Sylwa schy for 23. Coll ins S], Galpo RC, Ga ll agher RE: ContinLi oLis growth and dif­ excel/wt techllical assisttlll Ce , Pro!). Oort for his critical rel/ielll ofth e ll l

glycolipids which both bind MIF and convert nonres ponsive cells 33. Boerrigter G H, Scheper RJ: Loca l and systemic desensitiza ti on in­ to res ponsive. Cell Immunol 90:605-613, 1985 duced by repeated epiclltaneous application. J In vest Der­ 32. Van de Plassche-Boers EM, Drexhage HA, Kokje-Kleingeld M, matol 88:3-7, 1987 Leezenberg HA: Parameters ofT-ceil mediated immunity to com­ 34. Fi scher T I, Maibach HI: T he thin layer rapid use epicutaneous test mensal micro-organisms in patients with chronic purulent rhi­ (TRUE-tes t), a new patch tes t· method w ith high accuracy. Br J nosinusitis: a co mparison between delayed type hypersensitivity Dermatol 11 2:63-68, 1985 (DTH) skin test, lymphocyte transformation tes t (LTT) and mac­ rophage migrati on inhibition fac tor (MIF) assay. C lin Exp Im­ 35 . Van Ketel WG : Petrolatum-a reliable vehicle for metal all ergens? munol 66: 1986 . (letter to the editor). Contact Dermatitis 7:60, 1981