CORRELATION OF IgG AND IgE ANTIBODY LEVELS ...

4 downloads 0 Views 4MB Size Report
Jan 1, 1981 - A group of 25 honey bee venom allergic patients were treated with commercial honey bee venom at a monthly maintenance dose of 100 ...
Reprinted

from ANNALS

of ALLERGY,

Volume

46, Number

1, January,

1981

CORRELATION OF IgG AND IgE ANTIBODY LEVELS TO HONEY BEE VENOM ALLERGENS WITH PROTECTION TO STING CHALLENGE DONALD R, HOFFMAN, Ph.D., SHERWIN A. GILLMAN, M.D., LEO H. CUMMINS, M.D., F.A.C.A., PETER P. KOZAK, JR., M.D., F.A.C.A., and ARTHUR OSWALD, M.D.

A group of 25 honey bee venom allergic patients were treated with commercial honey bee venom at a monthly maintenance dose of 100 micrograms for approximately one year. At the end of one year 24 patients were intentionally challenged and one was accidentally challenged. Three patients experienced significant systemic reactions to challenge and three experienced minor reactions. Sera obtained before commencing therapy, at maintenance and before challenge were tested by radioallergosorbent test (RAST). double antibody technique and protein A RAST for IgE and IgG antibody levels to all five known honey bee venom allergens. All of the treatment failures experienced at least a two-fold rise in IgG antibody against phospholipase. The ratios of IgG to IgE antibodies in the pre-challenge specimens were analyzed by a graphical method. Four patients had inadequate responses to at least three ofthefive allergens and three of these patients were those who experienced severe reactions to challenge. Sixteen patients had adequate responses to all five allergens, four patients to four allergens and one patient to three allergens: three of these patients experienced minor or local reactions to challenge and the remainder no reaction. No single allergen identified only the three severe reactors but three allergens identified all three reactors. The diagnostic efficiency of the criteria usedfor assessing protection was 0.96. The only non-correlating case was classifying a single nonreactor as at risk. No patients were misclassified as protected.

THE

CONCEPT OF A TYPE of immunity that could protect from allergy was proposed by Cooke, Barnard, and StulP in 1935. This protective immunity was called "blocking antibody" in distinction to the sensitizing antibody or "reagin." Since that time many investigators have attempted to correlate increases in "blocking antibody" with relief from symptoms. Soon after the discovery that reagin was IgP, it was established that much of Cooke's blocking antibody activity was IgG.3 Carefully designed and performed studies using the

From the Department of Pathology and Laboratory Medicine, East Carolina University School of Medicine, and S1. Joseph's Hospital, Orange, California. This paper was presented in part at the 36th annual meeting of the American Academy of Allergy, Atlanta, Georgia, February, 1980. This study was supported in part by grants from Pharmacia Inc. VOLUME

46. JANUARY.

1981

ragweed hay fever model were able to show some correlation of rise of blocking antibody titer during immunotherapy with improvement in symptom scores but the results were neither uniform nor definitive: However, it became clear that the ragweed hay fever system was not the best experimental choice because of the problems of variability of exposure, complex allergenic structure, the localization of the reaction in the nasal mucosa, the question of extraction of pollen grains and the possible existence of local rather than systemic immunity. In the last eight years several groups of investigators have turned to the hymenoptera venom allergy model to study many of the variables in pathogenesis and therapy of allergy. If one chooses to study only honey bee venom allergy, one has a very well characterized allergen 17

IgG & IgE ANTIBODY

LEVELS

& STING

system with five purified and defined allergens which account for more than 98% of the protein in the venom.5.6.,.8 The venom system has major advantages for the investigator. Exposure is known and challenge can be carried out in an experimental situation, although with some danger. The venom is injected by the insect subcutaneously or intradermally and is absorbed readily into the circulation so one need not be concerned with route of exposure and intervening membranes. The reactions are systemic and can be studied using blood by immunochemical methods as well as in vivo. This study uses the honey bee venom allergy system to study simultaneously the relationship of IgE and IgG antibodies to all five allergens of honey bee venom during immunotherapy and compares these immunochemical data to the results of sting challenges with honey bee venom. This is the first report in which the human response during immunotherapy to all of the significant allergens in a natural allergen mixture have been studied simultaneously. Material and Methods Antigens. Honey bee venom collected by electrical stimulation was purchased from Champlain Valley Apiaries, Middlebury, Vermont. Phospholiphase A and hyaluronidase were prepared by gel filtration on Sephadex G-75® followed by ion exchange chromatography on CM Sepharose CL-6B® as previously described.5.6 Melittin was prepared by repeated gel filtration and was free of phospholipase and hyaluronidase by enzyme assay (less than 0.1 %). Acid phosphatase and allergen C were prepared by solid phase immunoadsorption and gel filtration on Sephadex G-200®.' Each allergen was assayed for purity by ITable I. Immunotherapy Schedule. 11 100 I 2 18 15 40 4 0.01 1 20 3.0 6 8 100 80 12 13 0 3 .06 .03 5 10 .2 .4 Week 6.0 1.0 9 7 4.0 micrograms

100 0.1 0.04 60 0.02 30 0.3 0.6 10 2.0

Dose in

CHALLENGE

-

HOFFMAN

ET AL

immunodiffusion using rabbit antisera, crossed immunoelectrophoresis, RAST using specific human sera, polyacrylamide gel electrophoresis, amino acid analysis, hemolytic activity and enzyme assays for phospholipase, hyaluronidase, acid phosphatase and protease. Subjects. Informed consent was obtained from each subject who entered the study. An additional informed consent was obtained for each intentional challenge. Eligible patients had experienced at least one recent serious systemic reaction to a honey bee sting without known sensitivity to other hymenoptera. Beekeepers were excluded. All patients admitted to the study were skin test positive to honey bee venom by intradermal test at I microgram/ml or less and almost all were negative to other venoms. All patients resided in Southern California where honey bee stings predominate.9 Both children and adults were admitted as long as they were candidates for venom immunotherapy. Two subjeCts refused sting challenge. Serum specimens were obtained at entrance upon reaching maintenance, at nine months and pre-challenge at one year. Post-sting specimens were also available for many subjects. All tests on each subject's serum specimen were performed at one time. During treatment patients were followed with routine hematology and serum chemistry tests. Immunotherapy. Patients were treated with Pharmacia Pharmalgen® honey bee venom. Patients numbered I through 23 were treated with a single lot, which was known to contain all five allergens. Patients 38, 53, 48 and 49 were treated with other lots of venom. The schedule is given in Table I. Several patients had slower dose progression because of severe reactions to therapy. All reached maximum dose within six months. All patients except number 48 were treated by the same physician. Sting Challenge Tests. A specific informed consent was obtained for each intentional sting challenge. Both an allergist and an anesthesiologist were present and remained with the subject until discharge. After a physical exam to establish that the subject was healthy and measurement of baseline pulmonary function, pulse and blood pressure, a honey bee brought in by a beekeeper was placed on the subject's arm until it completed a sting. The subject was observed and monitored. Any significant decrease in pulmonary function or drop in blood pressure was immediately treated with subcutaneous epinephrine. Resuscitation and emergency equipment were available in the room. The one case with an accidental sting was verified as a honey bee sting by the presence of the stinger and venom sac at the sting site. RAST. The radioallergosorbent test was performed as described by HoffmanJO using allergen coupled to paper discs by CNBr activation. Purified bee venom allergens were coupled to activated discs at a ratio of 0.5mg allergen per 0.2g discs. The same discs were used for IgE RAST and protein A RAST. A standardized lot of bee venom was used. ANNALS OF ALLERGY

IgG & IgE ANTIBODY

LEVELS

& STING

Double Antibody Method for IgG Antibody to Hyaluronidase and Phospholipase. Purified bee venom hyaluronidase was labelled by reaction at pH 8.0 with N-succinimidyl 3-( 4-hydroxy, 5_['251] iodophenyl propionate) (Bolton-Hunter reagent). The excess reagent was reacted with glycine and the 1251_ hyaluronidase separated on a column of Sephadex G-25® in 0.1 % gelatifl in borate buffered saline. Phospholipase A was labelled by the method of Markwell and Fox11 using 1,3,4,6-tetrachloro-3, 6-diphenylglycoluril (IodoGen®) obtained from Pierce Chemical Co., Rockford, Illinois, as catalyst and the labelled enzyme was purified on a Sephadex G-25® column in borate buffered saline containing I mg/ml of human serum albumin. The assay was performed by placing 25 microliters of a I: I0 dilution of serum in a 1.5ml polypropylene centrifuge tube; 100 microliters of borate buffer pH 8.0 containing 1.25 microgram/ml of labelled hyaluronidase or 0.5 microgram/ml of labelled phospholipase. The tubes were allowed to stand at room temperature for three hours and then 100 microliters of a I :20 dilution of goat anti-human IgG were added (equivalence of second antibody was determined in a pilot experiment). The tubes were stoppered, mixed, allowed to stand one hour at room temperature and then overnight at 4°C. One ml of borate buffer, pH 8.0, was added to each tube and the tubes were spun for 15 minutes in an Eppendorf 5413 centrifuge. The supernatants were decanted; the precipitates washed with borate buffer and counted. Both positive and negative standards were run. Values are expressed as micrograms of antigen bound per ml of serum based upon a standard serum. Reproducibility of the assay is ± 5%. Protein A RAST. Protein A RAST was used to measure IgG antibodies against acid phosphatase, allergen C and melittin. It is difficult to prepare stable 1251-labelled acid phosphatase and allergen C to use in the double antibody method. The method used the same RAST discs used for measuring IgE antibody levels. Staphylococcal protein A, purchased from Pharmacia Fine Chemicals, Piscataway, New Jersey, was labelled as described by Langone et al.12 Sera were diluted 1:100 in borate buffer pH 8.0 containing I % Tween 20, 0.2% bovine serum albumin and I% pooled human cord serum. A RAST disc was placed in a 12 x 75mm tube and 100 microliters of serum dilution were added and incubated overnight at room temperature. The disc was then washed three times with 0.9% saline, containing I % Tween 20, and then 100 microliters of a solution of 160 ng/ml labelled protein A in borate buffer pH 8.0 containing I% Tween 20 and 0.2% BSA were added. After three hours incubation the disc was washed three times, changed over to a new tube and counted. Negative controls were cord serum only and results were expressed as a percentage of binding after subtraction of negative control binding. These assays were approximately 1/100 the sensitivity of IgE RAST and the reproducibility was ± 5% in the same VOLUMF

46. JANUARY.

19XI

CHALLENGE

-

HOFFMAN

ET AL

Table II. Description of Adverse Reactions from Honey Bee Sting Challenge after Maintenance Immunotherapy. Subject Reaction 6 Wheal followed by drop in peak flow progressing to overt wheezing: required epinephrine. 9 Local swelling and urticaria with skin tingling progressing to generalized urticaria and erythema: drop in peak flow, overt wheezing: required epinephrine. 11 Flushing, skin tingling, angioedema of lip, urticaria on neck: resolved spontaneously. 18 Throat tickle, cough. 48 Respiratory distress; loss of consciousness; required epinephrine (accidental sting). 49 Swelling on arm contiguous with sting site (large local).

run. The protein A RAST system was tested by comparison with the double antibody method. Results of linear regression analysis with 20 sera gave correlations of r = 0.94 (p< .00 1) for hyaluronidase and r = 0.91 (p c.') U;

~0

10

§

I

100

10

1.0

IgG Pre

Mluograms

Bound

Figure I. IgG antibody levels in micrograms antigen bound per ml of serum against phospholipase A for pre-treatment sera on the abscissa and sera at maintenance on the ordinate. The numbers refer to the subject. Triangles indicate severe reactions upon sting challenge and circles mild reactions. Plain numbers indicate subjects with no reaction to challenge. The lines indicate equal levels and doubling of levels.

CHALLENGE

ET AL

Table III. Patients with Inadequate IgGllgE of the Five Bee Venom Allergens.

Allergen

10

HOFFMAN

protection although the subjects who experienced severe reactions on challenge tended to have higher specific IgE levels. All of the subjects showed increases in specific IgG from pre-treatment to maintenance specimens. The levels of IgG against phospholipase A for the pre-treatment and maintenance specimens are shown in Figure I plotted by the method of Siber et al.14 Only two subjects did not show a two-fold increase in anti-phospholipase IgG at maintenance and neither of these experienced a severe reaction on challenge. The three subjects who did experience severe reactions on challenge showed 2,5-, 3.6- and 15-fold increases in specific IgG anti-phospholipase. Results were similar for the other four allergens. The data were then analyzed by looking at the ratios of IgG to IgE antibodies for each allergen in the pre-challenge specimens, These results are plotted for each allergen in Figures 2-6 on logarithmic scales. The subjects experiencing severe reactions are shown 'by triangles and those experiencing minor reactions by circles. The lines are drawn to best separate the severe reactors from the non-reactors. The minor reactors were not distinguishable from the non-reactors, The subjects whose specimens gave IgG / IgE ratios below the cut-off

tOO

ACID PHOSPHATASE

-

Acid phosphatase Allergen C Hyaluronidase Phospholipase A Melittin

Responses to Each

Patients with inadequate IgG/lgE responses

Severe systemic reactors 1 inadequate responses

6,6* ,9,23,38,48 9,11,22,23,48 9,38,48 1,6,9,38,48 6,9,38,48

3/6 2/5 2/3 3/5

3/4

10

ALLERGEN -C

10

1.0

1.0

01 01

IgE % Figure 2. Antibody levels against bee venom acid phosphate in pre-challenge sera. The IgE antibody measured by RAST is plotted on the abscissa in percentage of binding and the IgG antibody level by protein A RAST is plotted on the ordinate also in percentage of binding. Subjects with no reactions on challenge arc shown by plain numbers, those with severe reactions with triangles and those with mild reactions with circles. The asterisk indicates a specimen before a second challenge after doubling venom dose. The line indicates the "protective" cut-off level for IgGjlgE ratio. 20

0.1

0.1

1.0

10

IgE Figure 3. pre-challenge

100

%

Antibody levels against bee venom sera. Symbols as in Figure 2. ANNAI.S

allergen

C

OF ALI.ERGY

in

IgG & IgE ANTIBODY

LEVELS

& STING

CHALLENGE

-

HOFFMAN

ET AL

improved significantly by use of increasing or decreasing ratios (slope different from 1.0). Based upon these data it was proposed that an IgG/IgE ratio above the cut-off for three or more of the five allergens constituted a protective response. This criterion was analyzed by Bayes' theorem assuming the results of the sting challenges to be the true values. The sensitivity of the criterion was 1.00; no subject at risk

lines for each allergen are listed in Table III. Subject 6* is subject number 6 after doubling the venom dose to 200 micrograms and a rechallenge with no adverse reaction. The ratios for acid phosphatase were inadequate for six subjects, of which three were severe reactors, allergen C response was inadequate for five with two reactors, hyaluronidase response was inadequate for three with two reactors, phospholipase A response was inadequate for five with three reactors and melittin response was inadequate for four with three reactors. All three severe systemic reactors or "treatment failures" had inadequate responses to acid phosphatase, phospholipase A and melittin and two failures responded 0 inadequately to allergen C and hyaluronidase. (') ~t C, OJ However, 1.0 three protected subjects had inadequate response to acid phosphatase, three to allergen C, one to hyaluronidase, two to phospholipase and one to melittin. The inadequate responses to venom therapy for each allergen are summarized in Table IV. The three cases with severe systemic reactions on challenge had inadequate responses to three, five and five allergens respectively. One subject who had no reaction on challenge had inadequate response to four allergens. One subject had an inadequate response to two allergens and four subjects to one allergen each. The remaining 17 subjects had adequate responses to all five allergens. The cut-off values for IgG/IgE ratio were derived empirically from graphical analysis. The IgE RAST is approximately 100 times more sensitive than the IgG antibody assays. Since the parameter being evaluated was the ratio, only lines of slope equal to 1.00 were tested on the log-log plots. The results could not be

100

PHOSPHOLIPASE

-"§

CD

~

5323

10

-A

21 @ 221619 3 21~7 12

®

01 0.1

IgE % Figure 5. Antibody levels against bee venom phospholipase A in pre-challenge sera. IgG anti-phospholipase A is expressed in micrograms antigen bound per ml serum. Other symbols as in Figure 2.

100

100

HY ALURONIDASE

t0~E!'§ ~ (5 CD

0,

10

~l'4

MELITTIN

, 2117

OJ

lOt-

* (')

21 LO~

8

2

19

I

1.0

I 10

1.0

100

IgE %

I

Figure 4. Antibody levels against bee venom hyaluronidase in pre-challenge sera. IgG anti-hyaluronidase is expressed in micrograms antigen bound per ml serum. Other symbols as in Figure 2. VOLU~IE

41>. JANUARY.

19KI

3

/22381417 20

@10 1~

7

16

0' 0.1

1.0

10

100

IgE % Figure 6. Antibody levels against pre-challenge sera. Symbols as in Figure

bee 2.

venom

melittin

in

21

IgG & IgE ANTIBODY

LEVELS

& STING

CHALLENGE

-

HOFFMAN

ET AL

Table IV. Inadequate IgG/lgE Responses to Bee Venom Allergens in Patients With and Without Reactions on Challenge.

Patient 6 9 48 11 18 49 1 6' 22 23 38

Reaction on challenge Systemic Systemic Anaphylactic Minor Minor Large local None None None None None

Allergens with inadequate IgG/lgEa A.P., PLA, Mel A.P., C, Hyal, PLA, Mel A.P., C, Hyal, PLA, Mel C None None PLA A.P. C A.P., C A.P .. Hyal, PLA, Mel

aAbbreviations: AP-acid phosphatase, PLA-phospholipase A. Mel-Melittin.

was classified as protected. The specificity was 0.95; one individual who was protected was classified as at risk. The predictive value of a negative (individual was protected) was 1.00 and the predictive value of a positive (individual at risk) was 0.75. The only error was in classifying one individual who was protected on challenge as at risk. The efficiency of the criterion was 96%. The evaluation using the ratios for any single allergen were only as good for melittin; however, many of the subjects had very low levels of IgE antibody against melittin. Discussion The results of this study suggest that IgG antibody against honey bee venom allergens, the so-called "blocking antibody," is important to the development of protection to subsequent stings. It was not possible to define a level of IgG antibody against each allergen necessary for protection but the ratios of IgG to IgE antibodies against the various allergens gave an excellent correlation with the existence and non-existence of protection. The three subjects who experienced severe systemic reactions on post-treatment sting challenge all had inadequate IgG j IgE responses to at least three of the five honey bee venom allergens while only one protected (non-reactive) subject was in this group. The other 2 I protected subjects all had adequate IgG j IgE ratios for three to five allergens. The failure of one individual to experience a systemic reaction although his IgG j IgE ratios were low is not surprising since only about 60% of untreated patients experience a reaction on sting challenge.15 The criterion for protection suggested by this study is an adequate response to at least three of the five honey bee venom allergens. An adequate response is defined by graphical analysis of the data and consists of an IgGjlgE ratio of the order of 200 to one. Since the methods used were RAST-based, it was not possible to calculate the exact ratio. This criterion gave a sensitivity of 100% and a specificity of 95%. No individual was misclassified as protected and only one was misclassified as at risk. 22

C-allergen

C.

Number of allergens with inadequate responses

3 5 5 1

o o 1 1 1 2

4 Hyal-hyaluronidase,

Another criterion has been proposed by Lichtenstein et al.16 They proposed that doubling of the IgG antibody. level against phospholipase A from pre-treatment to maintenance correlates with protection. When this test is applied to our subjects, as shown in Figure I, all three severe reactors are classified as protected. Only two subjects fail this criterion, one non-reactor and one who experienced a large local on challenge. Yunginger has reported two patients who tolerated sting challenge despite showing no increase in anti-phospholipase IgG.1J

Two other studies support the concept that IgG blocking antibodies are important in protection against honey bee stings. Many beekeepers who have high levels of IgE antibodies against honey bee venom allergens, positive skin test and RAST are stung regularly without adverse reaction. High levels of IgG antibody against phospholipase A are found in sera from these beekeepers. Sensitive beekeepers had low or absent levels of IgG anti-phospholipase in their sera.IS A more direct study was reported by Lessof et al19 who sting-challenged a group of 12 honey bee venom allergic individuals of whom five reacted systemically. These five subjects were then given infusions of gamma globulin prepared from hyperimmune beekeepers and challenged again. All five patients tolerated greater venom doses than had induced the initial reaction. The Lessof study is the only direct evidence available for demonstrating the role of IgG antibody in protection from venom anaphylaxis. The data presented here are the first study of the response to all of the significant allergenic components of a natural allergen system during immunotherapy. The responses of a single individual to the various component allergens are not always parallel, so one cannot study the response to a single allergenic component and assume that it is typical of the response to the other allergenic components present in the natural allergen mixture. Two of the three treatment failures discussed in this paper have since been treated with double the usual maintenance dose of venom. One of them consented to A:-;:-;AI.S

OF Al.LERGY

IgG & IgE ANTIBODY

LEVELS

& STING

rechallenge and had no reaction. Her serum specimen obtained prior to the second challenge showed an adequate response to four of five allergens and is shown as 6* in Figures 2-6. The third patient became pregnant and was discontinued from the study until after parturition. Acknowledgements The authors thank Dr. Michael Woehler and Mrs. !Ida DeFilippi of Pharmacia for their support and help with this project. Dr. Bernard Geller generously provided sera from a case of venom treatment failure. Mr. and Mrs. Don Gallup provided the honey bees and assisted with the sting challenges. James Sutton and Jeffrey Miller rendered superb technical assistance in the laboratory studies and Mrs. Jo Gillin and Ruth Carson provided excellent secretarial assistance. References I. Cooke RA. Barnard JH. Hebald S and Stull A: Serologic evidence of immunity coexisting sensitization in a type of human allergy (hay fever). J Exper Med 62: 733. 1935. 2. Ishizaka K. Ishizaka T and Hornbrook M M: Physicochemical properties of human reaginic antibody: presence of a unique antibody as carrier of reaginic activity. J Immunol 96: 75 1966. 3. Lichtenstein LM. Holtzman NA and Burnett LS: A quantitative in vitro study of the chromatographic distribution and immunoglobulin characteristics of human blocking antibody. J Immunol 101: 317. 1968. 4. Lichtenstein LM. Ishizaka K. Norman PS. Sobotka AK and Hill BM: IgE antibody measurements in ragweed hay fever: Relationship to clinical severity and the results of immunotherapy. J Clin Invest 52: 472. 1973. 5. Hoffman DR and Shipman WH: Allergens in bee venom. I. Separation and identification of the major allergens. J Allerg & Clin Immunol 58: 551. 1976. 6. King TP. Sobotka AK. Kochoumian L and Lichtenstein LM: Allergens of honeybee venom. Arch Biochem & Biophys 172: 661, 1976.

Printed

VOLUME

-16. JANUARY.

19KI

.----------------------------------~

CHALLENGE

-

HOFFMAN

ET AL

7. Hoffman DR. Shipman WH and Babin D: Allergens in bee venom. II. Two new high molecular weight allergenic specificities. J Allerg & Clin Immunol 59: 147. 1977. 8. Hoffman DR: Allergens in bee venom. III. Identification of allergen B of bee venom as an acid phosphatase. J Allergy & Clin Immunol 59: 364. 1977. 9. Hoffman DR. Sutton JL and Miller JS: Hymenoptera venom allergy: a geographic study. Ann Allerg 45: 276, 1980. 10. Hoffman DR: The use and interpretation of RAST to stinging insect venoms. Ann Allerg 42: 224. 1979. 11. Markwell MAK and Fox CF: Surface-specific iodination of membrane proteins of viruses and eucaryotic cells using 1.3.4.6-tetrachloro-3. 6 .-diphenyl glycoluril. Biochem 17: 4807. 1978.

n.

12. Langone Boyle M DP and Borsos T: "'J Protein A: Applications to the quantitative determination of nuid and cell-bound IgG. J Immunol Meth 18: 281. 1977. 13. Hamilton RG. Sobotka AK and Adkinson NF: Solid phase radioimmunoassay for quantitation of antigen-specific IgG in human sera with "'I-protein A from Staphylococcus aureus. Immunol 122: 1073. 1979. 14. Siber GR. Ransil BJ and Schiffman G: Graphical method for evaluation antibody response to vaccines. Infect Immunity 28: 641, 1980. 15. Hunt KJ. Valentine MD. Sobotka AK. Benton A W. Amodio FJ and Eichtenstein LM: A controlled trial of immunotherapy in insect hypersensitivity. N lOng .J Med 299: 157, 1978. 16. Lichtenstein LM. Valentine MD and Sobotka AK: Insect allergy: the state of the art. J Allerg & Clin Immunol 64: 5. 1979. 17. Yunginger .JW: The sting - revisited . .J Allerg & Clin Immunol 64: I, 1979. 18. Yunginger JW. Jones RT. Leiferman KM. Paull BR, Welsh PW and Gleich GJ: Immunological and biochemical studies in beekeepers and their family members . .J Allerg & Clin Immunol 61: 93. 1978. 19. Lessof MH. Sobotka AK and Lichtenstein LM: Effects of passive antibody in bee venom anaphylaxis. Johns Hopkins Med .J 142: I 1978. Requests for reprints should be addressed to: Donald R. Hoffman. Ph.D. Associate Professor of Pathology and Laboratory East Carolina University School of Medicine Greenville. North Carolina 27834

Medicine

in U.S.A.

2.1