Prostaglandin D2 and histamine during the immediate and the late-phase components of allergic cutaneous responses

Prostaglandin D, and histamine during the immediate and the late-phase components af allergic cutaneous responses Marek M. Pienkowski, MD,* N. Franklin Adkinson, Jr., MD, Marshall Plaut, MD, Philip S. Norman, MD, and Lawrence M. Lichtenstein, MD ** Baltimore, Md., and Knoxville,

Tenn.

With u skin blister technique in which the mediators generated by the trauma of forming the blister are allowed to subside, we have collected human interstitial skin fluid during the course of allergic reactions to ragweed, and measured levels of histamine and prostaglandin DJ (PGD,). Of 18 ragweed-allergic individuals tested, I I developed both an immediate and u late-phase reaction (LPR) with fivefold-elevated levels of histamine (40 nglml) at 30 minutes and a peak level of PGD, (6.5 nglml) later at 2% hours after ragweed challenge. The other Seven allergic individuals had immediate reactions without an LPR lesion and demonstrated somewhat smaller elevations of histamine (25 nglml) but much lower levels of PGDZ (I.6 nglml; p < 0.05). The time course of appearance of these mediators was identical in both groups of patients. The fluids from unchallenged blisters of allergic and nonallergic patients and the fluids of nonallergic patients challenged with ragweed had similar levels of histamine, at the lower limit of detection, and undetectable PGD, levels. The peak levels of PGD, in allet@ individuals correlated with the size of the LPR lesion (p < 0.05). These data suggest thut the LPR involves the secondary elaboration of mediators different from mediators responsible for the immediate mamfestations of the allergic skin reaction. (J ALLERGYCLIN IMM~JNOL 1988;82:95-100.)

Challenges with several common allergens often evoke not only the familiar immediate response,but also an LPR, which typically peaks at about 8 hours and lasts 24 hours or more. In the skin, the LPR consists of subcutaneousedema with erythema and local warrnth,‘~’ whereasin the lung, there is a delayed increaseof airways resistance.3‘5In the nose, the LPR is also associatedwith an increase of airway resistance.6The LPR in the skin is IgE-dependent, since it can be passively transferredto normal skin with the serum of allergic patients, and passageof the serum over an anti-IgE solid-phase immunoabsorbent reFrom The JohnsHopkins University School of Medicine, Division of Clinical Immunology, Baltimore, Md., and the *University of Tennessee,Knoxville, Tenn. Supported by National Institutes of Health Grants AI04866, AIO8270, A121073, and AI07056. Received for publication July 13, 1987. Accepted for publication Jan. 20, 1988. Reprint requests:Philip S. Norman, MD, The JohnsHopkins University School of Medicine, Clinical Immunology Division, at the Good Samaritan Hospital, 5601 Loch Raven Blvd., Baltimore, MD 21239. **Recipient of a Pfizer Biomedical ResearchAward. Publication No. 722 of the O’Neill Research Laboratories, The Good SamaritanHospital, Baltimore, Md.

Abbreviations used LPR: LTC,: PGD, PGE, PGF: DAO: PNU:

Late-phase reaction Leukotriene C, Prostaglandins D, E, F Diamine oxidase Protein nitrogen unit

moves this ability. Furthermore, a similar reaction is evoked by anti-IgE antibody.‘. ’ Histopathologically, the LPR demonstratesedema associatedwith the degranulation of mast cells and infiltration of mononuclear cells, basophils, neutrophils, and eosinophils.‘, 2, ‘, ’ The same cellular pattern, however, can be found after an immediate wheal-and-flare reaction that did not lead to a macroscopic LPR.9. ‘OIntracutaneous injection of compound 48/ 80, a mast cell degranulator, produces a similar LPR, suggestinga primary role for the mast cell.2. ” Elevated histamine levels in skin blisters are found during the course of the immediate reaction” and remain elevatedfor at least4 hours.‘” LTs (mainly LTCJ 95

96

z s,

Pienkowski

0.5

J. ALLERGY CLIN. IMMUNOL. JULY 1988

et al.

2.5

5.5

8.5

24

TI ME (hours afler the challenge)

FIG. 1. Time course of the cutaneous reaction, expressed as diameter of erythema-edema (mean * SEM), in those 11 (group C) atopic individuals with a positive LPR. The reaction was induced with ragweed extract at a dose tenfold higher than that required for a 4+ immediate reaction for each individual.

also appearearly but rise from the secondto the fourth hours.14Injections of LTs into skin produce a reaction grossly resembling a late-phaseresponse.I5 Dorsch and RingI report increasesin thromboxane B2 and kallikrein activity in cutaneousblisters during the development of the LPR. Attempts, however, to reproduce the LPR in normal skin by injection of histamine or bradykinin have been unsuccessfu1,9’ I7 even though kallikrein, which generatesbradykinin, causesboth a wheal and flare and an LPR.” Calcitonin gene-relatedpeptide, a potent vasodilator, producesa late erythema and swelling without an immediate responsewhen it is injected into skin.18 The present investigation was designedto measure the kinetics of appearanceof the mediators released at reaction sites during the immediate and LPRs in the skin. We created skin blisters, induced a local reaction by the intradermal injection of antigen under the blister, and assayed for mediators entering the overlying blister as the reaction proceeded.By forming the blisters hours before initiation of the hypersensitivity reaction, we were able to dissociate mediator release caused by thermomechanical injury from that produced by antigen challenge. With this improved experimental technique, we followed the appearanceof two mediators, histamine (preformed in mast cells and basophils) and PGD, (presumably newly synthesized),during the courseof the cutaneous allergic reaction. We found that the patients who developed LPRs generatedmore PGD, and that the peak level of PGD, correlated with the size of LPR.

METHODS OF PROCEDURE Patient selection Eighteen individuals (eight were male and 10 were female), aged 18 to 45 years (median age 26 years), who had a history of hay fever and/or asthma during the ragweed season,were studied. Eleven had severesymptoms of allergic rhinitis during the ragweed season,and four of these individuals had concomitant asthma attacksrequiring both inhaled and oral medications. The other seven individuals had only mild symptomsof hay fever, not requiring medication. The control group consisted of eight individuals matchedby sex and agebut with no symptomsof hay fever and/or asthma.None had beentreatedwith immunotherapy, and nine were receiving oral or topical medications at the time of the study. The protocol was approved by The Johns Hopkins University Clinical Investigation Committee, and all individuals gave informed consent. Skin tests Three to four weeks before experiments, patients were intradermally tested with ragweed extract (short ragweed pollen; Center Laboratories, Port Washington, N.Y., lot 8221921), 7.4 antigen E units per milliliter of concentrate (20,000 PNU/ml), in serial dilutions ranging from 0.1 PNU/ml to 1000 PNU/ml, made with Alby (0.9% saline, 0.03% serum human albumin, and 0.4% phenol, HollisterStier, Spokane,Wash., lot 0472). For eachdilution, the size of erythema and wheal of the immediate reaction was recorded at 15 minutes, and the diameter of the less welldemarcatederythema and subcutaneousedemaof the LPR was measuredat 2.5, 5.5, 8.5, and 24 hours. Skin bullae The skin blisters were made on the volar surface of the forearm with a heating-suction device. As the result of application of 60 mm Hg/cmZ suction pressurethrough a hollow plate heated to 40’ C on precleaned skin, a circular blister (5 mm in diameter)was formed within approximately 1 hour. The blister resulted from separationof the dermis at the dermal-epidermaljunction and was filled with a clear transudatefluid.‘9.2”In order to removethe fluid, a 28-gauge needle was inserted into the skin 2 mm from the edge, advancedbetweenthe layers of skin of the base,and angled upward into the center of the cavity; the fluid was aspirated with a l-ml syringe. After the experimental procedure, the blister was cleaned with an alcohol swab and covered with sterile gauze. The lesions healed within a week without scarring, occasionally leaving an increased pigmentation that lasteda few weeks. None of the studieswere performed during the ragweed season. Measurement

of mediators

Blister fluid was collected periodically after the allergen challenge with an effort to evacuate the blister completely. The volume of fluid (approximately 0.1 ml) was measured, and the fluid was diluted with 1.O ml of precooled buffer

VOLUME NUMBER

82 1

Prostaglandin D2and histamine during immediate and late cutaneous responses

(25 mm01iL of piperazine-N-N’-bis-2-ethanesulfonic acid, 1I() mmol/L of NaCl, and 5 mmol/L of KCl, pH 7.3) and spun at about 500 g at room temperaturefor 5 minutes to remove debris. The clear supematant was transferred to separatetubes. The histamine concentration was measured with the automatedfluorometric methodof Siraganian” after precipitation of bullous fluid proteins with 8% perchloric acid. The sensitivity of this assayis 0.5 to 1.O rig/ml. To authenticatethe presenceof histamine, portions of several blister Huids were treated with DA0 (DAO, 1 U/ml, at 37” C. for 60 minutes), while anotherportion was incubated without DAO. The histamine was destroyed when it was compared,by the Ruorometricassay,to the untreatedfluids. PGD2was measuredby competitive radioimmunoassay,as previously described.‘2The anti-PGD, antiserum was prepared by immunizing rabbits with PGD,-bovine thyroglobulin conjugates, and cross-reactivity was < 1% for PGE,, PGE,, 6keto-PGF,,, PGF,,, PGF,,, and thromboxane B2. The assay was sensitive to about 40 pg/ml.” In someexperiments, sampleswere deproteinized by standardalcohol precipitation to assesspossible interference of transuded proteins in the PGD, assay,and no differences were found. Becauseof the approximately tenfold dilution of the blister fluid, the minimal detectable concentrations of histamine and PGDz in the undiluted fluid were about 8 rig/ml and 0.4 rig/ml, respectively.

Experimental

design

Our preliminary experiments demonstrated increased concentrations of both histamine (up to 100 rig/ml) and PGD, (up to 5 ngiml) in the fluid 30 minutes after blister formation. even without allergen challenge. These elevated values persistedfor up to 8% hours, presumably secondary to releasefrom cells damagedby thermomechanicalinjury, but were dissipatedby 14hours. Therefore, for our standard protocol, we produced two blisters, one on each forearm, 18 to 20 hours before the allergen challenge. For allergen challenge, 0.05 ml of ragweed extract, at a concentration 10 times higher than that which produced a 4-t reaction (wheal of diameter > 1 cm with pseudopods,and an erythema >4 cm), was injected intradermally into the baseof the blister on one arm. The other blister (control) was injected with 0.05 cc of the diluent. The size of erythemaedemawas measuredat 0.5, 2.5, 5.5, 8.5, and 24 hours after an allergen challenge,just before aspiration of the fluid from the blister. Basedon skin testsperformedbefore blister experiments were done, the subjects were divided into three groups. Group A (negative control subjects) consisted of eight individuals with both a negative immediate reaction (wheal <0.2 cm with ragweed extract at a concentration of 1000 PNUlml) and a negative LPR (erythema-edemalesion < 1 cm); group B (sevenindividuals) had a 4 + immediate reaction at a ragweedconcentration of s 100 PNUlml, but no LPR (at dosesof ragweedextract up to 1000 PNUlml); group C ( 11 subjects) had both a 4 + immediate reaction and a detectable late reaction. The results were grouped accordingly and analyzed with the Mann-Whitney U test.

97

RESULTS Injection of ragweed extract into the base of the blister at a concentration 10 times higher than that necessaryto produce a 4+ immediate:reaction in intact skin resulted in the prompt development of erythema in all allergic patients (groups B and C). The diameter of the erythema at 15 minutes in individuals who later developedthe LPRs (group C) was h.2 _tr 1.1 cm (N = 11) and was similar to that of the individuals who did not develop LPRL; (group B) (6.1 ? 1.2 cm) (N = 7). Nonallergic individuals (group A) (N = 8) did not develop erythema with ragweed at a concentration of 1000 PNl!!ml. concentrations that were as high or higher as those used in the patients of groups B and C. The wheai of the immediatereaction wasnot visible becausethe antigen was injected under the blister. Eleven of the 18 ragweed-allergic individuals developed late erythema-edemareactions after the injection of ragweed antigen (group C). The lesions, characteristic of the LPR, measured4.11z 1.0 cm (mean & SEM) 2% hours after challenge. The lesion size reachedan averagediameter of 6.6 t 0.8 cm at 5% hours (p < 0.01 versus2% hours) and 6.9 +_ 1.1 cm at 8% hours (p = not significant versus 5% hours), and then declined within 24 hours to an averageof 4.4 cm z? 0.9 cm (p < 0.01 versus55/z.and 8% hours) with no detectable 24-hour lesion in six of 11 individuals (Fig. 1). The remaining seven ragweed-allergic (group B) individuals, and all seven nonragweed-allergic individuals (group A) failed to develop any delayed skin lesion after challenge with ragweed extract. After the size of the skin lesion was determined, the skin bulla was aspirated, and the histamine and PGD, concentration of the fluid was measured.In the eight nonallergic individuals (group A), the concentration of histamine in the bullous fluid 30 minutes after the ragweedchallenge was below the detectable level of the assay.In thesenonragweed-sensitivesubjects, the histamine concentrations remamed unmeasurable during 24 hours in both the ragweed-injected and control bullous fluid (Fig. 2, group A 1. In ragweed-allergic subjectswho developedLPRs, the concentration of histamine in the fluid 30 minutes after the ragweed challenge was 40.5 t X.4 rig/ml (Fig. 2, group C). This level was approximately five times higher than the lower limit of detection. The histamine concentrationsgradually declined after ragweed challenge and, by 24 hours, were not significantly above controls. The ragweed-allergic patients without LPRs also exhibited elevated concentrations of histamine in the bullous fluid 30 minute’: after rag-

98

Pienkowski

J. ALLERGY CLIN. IMMUNOL. JULY 1988

et al.

CBA

I:., 0.5

TIME

( hours ofier the challenge 1

FIG. 2. Kinetics of histamine appearance in the bullaefluid during an allergic reaction induced with ragweed extract. A, nonatopic individuals (N = 8); B, atopic individuals without an LPR (N = 7); C, atopic individuals with an LPR (N = 11). Closed symbols indicate values from bullae fluids after ragweed challenge, and open symbols, values of control bullae fluids. Histamine levels are expressed as a mean 2 SEM. The dotted line indicates the lowest level of accurate measurement of histamine. *Significantly different from control (p < 0.05 by paired Student’s test).

t

weed challenge (25.6 + 7.8 rig/ml) (Fig. 2, group B). The histamine content of fluids of two of seven patients were not elevated with challenge compared to the levels of control fluids so that the mean increase in histamine was not significant by statistical test. In group C at 2% and 5 hours after challenge, the histamine levels were lower than at 30 minutes, but remained higher than control (p < 0.05). At 8% and 24 hours, bullous fluid from all antigen-challenged groups had levels of histamine not significantly higher than control. Levels of PGD2 in the blister fluid of nonallergic individuals, with or without ragweedchallenge (Fig. 3, group A), and in control blisters of allergic subjects were virtually undetectable(0.4 rig/ml) (Fig. 3). Four of I8 allergic patients had borderline detectableconcentrations of PGD, in the diluent-injected bullae, ranging from 1.7 to 0.4 ng/mI, 30 minutes and 2% hours after injections. Values for the 14 remaining individuals were CO.4 rig/ml. In the ragweedchallenged allergic patients, a marked elevation in PGD, concentration in the blister fluid occurred (Fig. 3, groups B and C). At 30 minutes after challenge, those who developed LPRs (group C) had elevated levels of PGD, (1.2 + 0.5 nglml), and these values were statistically different from saline-injected bullae (p. = 0.03) by paired Student’s t test. PGD, levels

0.5

2.5

5.5

TIME

(hoursoffer

8.5

24

the challenge)

FIG. 3. Kinetics of PGD, appearance in bullae fluid during allergic reaction induced with ragweed extract. A, nonatopic individuals; 9, atopic individuals without an LPR; C, atopic individuals with an LPR. Closed symbols indicate values from bullae after ragweed challenge, and open symbols, values of control bullae. *Significantly different different from LPR individuals from control; + significantly (p < 0.05 by paired Student’s t test). PGD, levels are expressed as mean ? SEM. The dotted line indicates the lowest level of accurate measurement of PGD2.

increased to 6.5 +- 1.6 nglml at 2% hours (p = 0.004 against 30 minutes) and 5.2 + 1.4 nglml at 5% hours (p < 0.01 against 30 minutes), and then declined, becoming undetectableby 24 hours. Only one of the 11 patients in this group had no detectable PGD, in the fluid during the reaction, whereas the highest concentration of PGD, encounteredwas 17.0 rig/ml. The allergic individuals in group B had lower concentrations of PGD, than group C individuals at the peak of the reaction, at 2% hours (1.6 % 0.7 rig/ml) (p = 0.022 against 30 minutes), and at 5% hours (0.6 4 2) (p = 0.01 against 30 minutes). Theselevels were, however, higher than either control or thosein nonallergic patients. Three of eight patients in this group had undetectable PGD,. The highest concentration of PGD, measured in this group was 4.4 ng I ml. In group C, antigen-challenged individuals, there was a statistically significant (p < 0.05) correlation between the concentration of PGD, and the diameter of the late-phaseerythema at 2% hours (r = O-6), 5% hours (r = 0.7), and 8% hours (r = 0.7). Combined data for 5 Y2 hours in both groups B and C are presented in Fig. 4. Histamine levels at 30 minutes also correlated with the size of

VOLUME 82 NUMEEP 1

Prostaglandin

D, and histamine

during

immediate

and late cutaneoiJ:< responses

39

B

the late-phase erythema at 5% hours (r = 0.58; p -: 0.05). DISCUSSION By waiting 18 to 20 hours after the blister had been formed before initiating allergen challenge, we observed allergically generatedmediatorswithout interference by mediators generatedas a result of thermal and mechanical injury. Our results demonstratedan antigen-specificaccumulationof histamine and PGD?, but with a time course different from that observed during another in vivo human test system, allergen challenge with ragweed extract to the nose. When pollen extract is sprayedon the nasalmucosaandnasal secretionsare sampledserially, there is an immediate appearanceof both histamine and PGD, (along with a number of other mediators not testedfor in our skin experimentsfor lack of sufficient sample). Thesedisappearedin about 30 minutes. When late-phasesymptoms occurred some hours later, histamine and other mediators appearedagain, but no PGD,.‘4 In the skin, the apparent continuing production of histamine could also be due to a secondwave of histamine release. The depression between a first and second wave might be missed in the relatively long period between samples. To examine this issue, we areundertaking experimentswith a methodthat allows more frequent sampling. Late production of histamine could come from basophils attracted to the lesion (mast ceils and basophils are the only known sources of histamine). The difference in the time course of PGD, during the skin reaction may be due to the differences in the mast cell environment in respiratory mucosa and in skin. Mast cells isolatedfrom human lung releaseboth histamine and PGD, at 37” C in response to IgEmediated stimuli.*’ Adult skin mast cells isolated in our laboratory also releasePGD2,but this is probably degradedby other cells in the preparation.26This rapid degradation is not observed in lung mast cell preparations and may representthe presenceof additional cells unique to the skin that metabolize PGD,. Basophils, it should be noted, do not synthesizeor release PGD?.l’ The Langerhans cells of guinea pig skin do have a high capacity to generatePGD2,28therefore, it is possible that thesecells could be stimulated by the eventsof the immediate reaction to synthesize PGD, later. Whatever is the source, the time of PGD, appearancewas coincident with the development of the late reaction. Moreover, rhere is a significant pitive correlation betweenthe diameterof the erythema and edemaof the LPR and the concentration of PGD, in the bullae of those individuals who went on to developa late reaction. In addition, a topically applied

2

4

6

8

$0

!2

DIAMETER OF THE ERYTHEMA-EDEMA OF LATE-PHASE REACTION AT 5 5 hr !mcm)

FIG. 4. Correlation between the size of the cutaneous LPR at 5% hours after ragweed challenge, and PGD:, levels in the bullae fluid during allergic reactions in all 18 atopic individuals.

cyclooxygenase-inhibiting agent, indomethacin. impairs the late reaction.29As a number of other mediators are found in allergic reactions: it would be prematureto suggesta cause-and-effectrelationship. Indeed, injection of PGD, into human skin induces immediate erythema without the late-phase component.” In the rat, intracutaneousinjection of histamine induced an increasein vascular permeability that was potentiated, in excess of simple summation, by simultaneously administered PGD,; however, an LPR did not develop.3o In this modeI, Lemanske and Kaliner” isolated a specific factor from mastcell granules responsible for cellular changesassociatedwith the LPR. Without naming all the possibilities, it is evident that there are other possible mediators that might be associatedwith the evolution of the LPR. It is generally believedthat mastcell chemotacticfactors may act to bring in or activate other inflammatory cells that participate in the generation of the late Iesion.32.35The positive correlation between histamine at 30 minutes and the eventual maximum size (at S’/2 hours) of the late-phase erythema is in agreement with this idea as in vitro studies indicate that mast cells secretetheir several mediators almost simuttaneously.2s Histologic observations in skin biopsy specimens and mediator studies in both skin and other organs indicate clearly that the subsequentlate-phaseis too complicated to be accountedfor by the elaboration of a single mediator, such as PGD,. Nevertheless, the correlation between the late lesion size and simulta-

100

Pienkowski et al.

neous levels of PGD, is impressive and suggeststhat PGD, may be a marker for whatever process causes the visible lesion. Sampling fluid from blisters overlying reactions representsa useful way to assessmediator release.We are initiating studiesthat will allow more frequent sampling of fluids from blisters with measurementof additional mediatorsto help elucidate the problem. REFERENCES

1. Dolovich J, Hargreave FE, ChalmersR, Shier KJ, Gauldie J. Late cutaneous allergic responsesin isolated IgE-dependent reactions. J ALLERGYCLM IMMUNOL1973;52:38. 2. Solley GO, Gleich GJ, Jordon RE, S&meter AL. The late phase of the immediate wheal and flare skin reaction. J Clin Invest 1976;58:408. 3. Killian D, Cockcroft DW, Haigreave FE, Dolovich J. Factors in allergen-induced asthma: relevance of the intensity of the airways allergic reaction and nonspecific bronchial reactivity. Clin Allergy 1976;6:219. 4. Davies RJ, Green M, Schofield NM. Recurrent nocturnal asthma after exposure to grain dust. Am Rev Respir Dis 1976;114:1011. 5. Bush RK. Cohen M. Immediate and late onset asthma from occupational exposure to soybean dust. Clin Allergy 1977; 7:369. 6 Tayor G, Shivalker PR. “Arthus-type” reactivity in the nasal airways and skin in pollen-sensitive subjects. Clin Allergy 1971;1:407. 7. Zetterstriim 0. Dual skin test reactions and serum antibodies to subtilisin and Aspergih fumigatus extracts. Clin Allergy 1978;8:77. 8. Atkins P, Green GR, Zweiman B. Histologic studiesof human skin test responsesto ragweed, compound 48/80, and histamine. J ALLERGYCLM IMMUNOL1973;51:263. 9. deShazoRD, Levinson AI, Dvorak HF, Davis RW. The latephaseskin reaction: evidencefor activation of the coagulation system in an IgE-dependent reaction in man. J Immunol 1979;122(2):692. 10. Richerson HB, Rajtora DW, Penick GD, Dick FR, Yoo TG, Kammermeyer JK, Anura JS. Cutaneous and nasal allergic responsesin ragweed hay fever: lack of clinical and histopathologic correlations with late-phasereactions. J ALLERGY CLIN IMMLINOL1979;64:67. 11. Dor PJ, Vervloet D, SapeneM, Andrac L, BonerandlJJ, Charpin J. Induction of late cutaneousreaction by kallikrien injection: comparisonwith allergic-like late responseto compound 48180. J ALLERGYCLIN IMMUNOL1983;71(4):363. 12. Ting S, Zweiman B, Lavker R, Dunsky EM. Histamine suppressionof in vivo eosinophil accumulation and histamine releasein human allergic reactions. J ALLERGYCLINIUMUNOL 1981;68:65. 13. Talbot S, Atkins P, Zweiman B. Prolonged histamine release in cutaneous allergic reactions. J ALLERGYCLIN IMMUNOL 1984;73:147. 14. Talbot SF, Atkins PC, Goetzl EJ, Zweiman B. Accumulation

J. ALLERGY CLIN. IMMUNOL. JULY 1988

of leukotriene C, and histamine in human allergic skin reactions. J Clin Invest 1985;76:650. 15. Soter NA, Lewis PA, Corey EJ, Austen KA. Local effects of synthetic leukotrienes(LTC,, LTD.,, LTE,, and LTB,) in human skin. J Invest Dermatol 1983;80:115. 16. Dorsch W, Ring J. Induction of late cutaneousreactions by skin-blister fluid from allergen-testedand normal skin. J ALLERGYCLIN IMMUNOL 198 1;67: 117. 17. Juhlin L, Michaelsson G. Cutaneousreactions to kalliiein, bradykinin, and histamine in healthy subjects and in patients with urticaria. Acta Derm Venereol 1969;49:26. 18. Piotrowski W, Fonnan JC. Some effects of calcitonin generelated peptide in human skin and on histamine release. Br J Dermatol 1986;114(1):37. 19. Kiistala U, Mustakallio KK. In vivo separation of epidermis by production of sunction blister. Lancet 1964;1:1444. 20. Willebrand EV, HorsmanheimoM, Kiistala U, Havry P. Analysis of intracutaneousinflammatory lesions with skin blisters. Clin Immunol Immunopathol 1978;11:445. 21. Siraganian RP. An automatedcontinuous flow system for the extraction and fluorometric analysis of histamine. Anal Biothem 1974;57:383. 22. Adkinson NF Jr. Prostaglandinproductionby humanperipheral blood cells in vitro. J Lab Clin Med 1977;90:1043. 23 SchulmanES, Newball HH, DemersLM, Fitzpatrick FA, Adkinson NF Jr. Anaphylactic releaseof thromboxane A*, prostaglandinsD2, and prostacyclin from humanlung parenchyma. Am Rev Respir Dis 1981;124:402. 24 Naclerio RM, Proud D, Togias AG, Adkinson NF Jr, Meyers DA, Kagey-Sobotka A, Plaut M, Norman PS, Lichtenstein LM. Inflammatory mediators in late antigen-inducedrhmitis. N Engl J Med 1985;313:65. 25. PetersSP, Schulman ES, Schleimer RP, MacGlashan DW Jr, Newball HH, Lichtenstein LM. Dispersed human lung mast cells: pharmacologicaspectsand comparisonwith human lung tissue fragments. Am Rev Respir Dis 1982;126:1034. 26. LawrenceID, WarnerJA, BascomR, Kagey-Sobotka A, Lichtenstein LM. Heterogeneity of human mast cells. J ALLERGY Crm IMMLJNOL 1987;79:179. 27. SchulmanES, MacGlashan DW Jr, Schleimer RP, Peters SP, Kagey-Sobotka A, Newball HH, Lichtenstein LM. Purified human basophils and mast cells: current conceptsof mediator release. Eur J Respir Dis 1983;64(128):53. 28. Ruzicka T, Aubijck J. Arachidonic acid metabolism in guinea pig Langerhanscells: studies on cyclooxygenase and lipoxygenasepathways. J Immunol 1987;138:539. 29. Dorsch W, Baur X. Effect of topical indomethacinon allergeninduced dual skin reactions. Allergy 1980;35:503. 30. Flower RJ, Harvey EA, Kingston WP. Inflammatory effects of prostaglandin D, in rat and human skin. Br J Pharmacol 1976;56:229. 31. LemanskeRF, Kaliner M. Mast cell-dependentlate-phasereactions. Clin Immunol Rev 1981;1(4):547. 32. Gleich GJ. The late phaseof the immunoglobulin E-mediated reaction: a link between anaphylaxis and common allergic disease?J ALLERGYCLIN IMMUNOL1982;70(3):160. 33. Atkins P, Zweiman B. The IgE-mediated late-phase skin response-unraveling the enigma [Editorial]. J ALLERGYCLIN IMMIJNOL1987:79:12.