Allergy to insect stings II. Phospholipase
A: The major allergen
Anne K. Sobotka, Martin Valentine,
Rudolph Harold
Baltimore.
in honeybee
M. Franklin, N. Franklin Baer,** and Lawrence
venom
Adkinson, Jr.,* M. Lichtenstein***
Md.
In order to determine the proteins of major allergenic importance in honeybee venom (Apis mellifera) it was chromatographed on G-50 Sephadex. The four major protein peaks eluted were identified as hyalzlronidase, phospholipase, melittin, and apamin. Testing these preparations on the leukooytes of 6 honeybee-sensitive patients, with the in vitro method of histamilte release, revealed that all individuals were most sensitive to phospholipase A. IgE antibodies against phosptilipase A (RAST) were found in the sera of honeybee-sensitive patients and IgG antibodies to this venom component were found in the sera from beekeepers and venomtreated patients. Melittin appeared to be allergenic in several patients, but the results were variable and were possibly due to contamination with phospholipase. All patients were insensitive to the hyaluronidose and apamin preparations. We conclude that phospholipase A is the major allergen of honeybee venom and, since this protein is readily available, it should be usefd for diagnostic and therapeutic studies as uell as for the standardization of materials used in the management of honeybee-sensitive patients.
The study of Hymenoptera sensitivity has been severely hampered by the use of impure and largely irrelevant extracts prepared from the whole ground bodies of these insects. Because of their widespread use, specific diagnosis was impossible’! 2 and even the “allergic” pathogenesis of the disease was legitimately called into question. Although some voices were raised against this practice, notably that of Loveless,3 this state of affairs has persisted for the last 30 years. We and others have recently used more modern tools to reinvestigate this problem and have been able to clearly demonstrate the utility of venoms in the diagnosis of Hymenoptera sensitivity. ** 5 It has, moreover, been possible to clearly implicate the IgE antibodies as the pathogenetic mechanism; preliminary therapeutic experience appears to confirm the importance of venoms.“9 i It is, of course, recognized that the venoms themselves contain multiple proteins From The Johns Hopkins University School of Medicine. Supported by Grants Nos. A107290, AI08270, and AI10304 from the National Institute of Allergy and Infectious Diseases, National Institutes of Health. Received for publication Sept. 12, 1974. Accepted for publication Jan. 10, 1975. Reprint requests to: Anne K. Sobotka, The Good Samaritan Hospital, 5601 Loch Raven Blvd., Baltimore, Md. 21239. *Allergic Diseases Academic Award. **From the Bureau of Biologics, Food and Drug Administration, Rockville, Md. ***Recipient of a Research Career Development Award from the National Institute of Allergy and Infectious Diseases, National Institutes of Health. Vol. 57, No. 1, pp. 29-40
30
Sobotka et al.
J. ALLERGY CLIN. IMMUNOL. JANUARY 1976
and therefore several potential allergens. To further assist diagnosis it is important to identify them; it is also imperative to have relatively clean allergens to use the RAST technique effectively and the same requirement exists for IgQ antibody measurements. We have begun this work with the best characterized venom, that of the honeybee (Rpis mellifera), in an effort to determine the major venom allergen and to obtain a preparation pure enough to carry out the in vitro diagnostic tests cited. Early studies of insect venoms by Jaques and Schaete9 revealed the presence of various vasoactive amines and kinins in wasp venom. Following the work of Benson and Semenov,B who found a beekeeper sensitive to both bee venom and whole insect extract, O’Connor, Rosenbrook, and Erickson1o and Shulman, Langlois, Arbesman, and their associateP’, b, c, d compared the antigenic content of whole insect bodies with that of venom sac preparations. Both groups found several antigens in the venoms not present in sacless bodies. Many investigators have studied the protein content of various pure venoms, especially that from the honeybee. The classic work of Habermann’2 led to the isolation and characterization of many of the components of the venoms of several Hymenoptera classes. In the case of the honeybee, hyaluronidase, melittin, phospholipase, and apamin make up great,er than ‘7070 of the venom protein. Several honeybee venom proteins have been well characterized and Shipolini and associatePa’ b have determined the primary amino acid sequence of the phospholipase A. in Melittin has also been sequencedI and then synthesized by investigators Gerrnany.15 This work has left unanswered, however, the question of which proteins in the venom were responsible for inducing the allergic response. The present study is directed to that question. With the use of simple chromatographic techniques and the leukoc.yte-histamine release model of “in vitro our results indicated that phospholipase A is the major allergen anaphylaxis,” in honeybee venom. METHODS Antigens A The antigens used for these studies were honeybee venom, commercial phospholipase (Sigma Chemical Company, Saint Louis, MO.) and 4 pools obtained from commercial honeybee venom by Sephadex G50 chromatography. The honeybee venom used was collected both by electrical stimulation and by allowing the insects to sting through a cellophane membrane.16 The dried venom was weighed in milligram quantities, diluted in Tris-buffered saline, dialyzed, and stored at -20” C as previously described.4 Phospholipase A from bee venom was treated similarly except that dialysis, used to remove substances from whole bee venom that interfere with the fluorometric assay for histamine, was not necessary. Separation of honeybee venom components was carried out as previously descrihed.17 A Scphadex G50 fine (Pharmacia, Piscatamay, N. J.) gel column was prepared and calibrated with blue dextran 2000 (Pharmacia) and with chymotrypsinogen, cytochrome C, and insulin as molreular weight markers. Then 500 mg of bee venom were placed on the column, the cluntes were collected at a constant volume of 4.8 ml, and the prot,ein content, of each tube was determined at an 0.D. of 280 nM. Alternatively (see below) the Folin-Ciocaltcu reaction was used to tstimate protein. Chromatography yielded 4 major protein peaks. Portions of the individual tubes comprising each peak mere pooled, equalized as to protein content., aliquoted, and stored at -20” C until assayed. The individual eluates were also aliquoted and frozen until used.
VOLUME 57 NUMBER 1
Allergy
to insect stings.
Previously described functional studies17 identified three pools as being largely of hyaluronidase (Pool I), phospholipase A (Pool II), and melittin (Pool III). Habermann’s data,12 Pool IV was assumed to contain apamin.
II
31
composed Based on
Antisera Antiserum to phospholipase A was prepared in two New Zealand white rabbits (Bunnyville Farms, R.D. 2, Littlestown, Pa.) by injecting 500 cg/ml in complete Freund’s adjuvant (Difco Laboratories, Detroit, Mich.) into both rear footpads. Two weeks later, the rabbits received an intramuscular injection of 500 pg of phospholipase A in saline and were bled 2 weeks later. The antiserum used in our in vitro studies was a pool made from this final bleeding; on immunodiffusion a single line of identity between purified phospholipase A and unfractionated honeybee venom was observed. Rabbit anti-whole bee venom was prepared according to a similar regimen; a total dose of 11 mg of venom was injected into each rabbit and the pool was made from several bleedings. Human antisera from 10 beekeepers was generously provided by Dr. A. W. Benton; sera from venom-treated patients were provided by Dr. Charles Reed and from our own practice. Drs. Halla Brown and Solomon Barr provided sera from patients on long-term therapy with whole-body insect extracts. Other sera from similar patients were obtained in our Allergy Clinic.
Histamine
release
studies
Patient sensitivity to honeybee venom, phospholipase A, and bee venom components was assessed with the leukocytes of known honeybee-allergic patients and normal control subjects as previously described.% 19 Briefly, leukocytes were collected by dextran sedimentation, washed, resuspended in TACM (Tris-buffered saline containing calcium, magnesium, and human albumin), and mixed with appropriate concentrations of each antigen, usually covering a range of 10-5-10 fig of protein per milliliter. After 60 min at 37” C the histamine release was determined fluorometrically. For passive sensitization experiments the recipient cells, previously shown to be nonreactive to honeybee venom, were sensitized following the techniques described by Levy and Osler.eo
Antibody
studies
Assessment of serum blocking antibodies using histamine release was carried out as previously described21 : appropriate dilutions of serum were incubated with the antigens for 30 min at 25” C prior to the addition of leukocytes and the amount of histamine release in these tubes compared with those in which antigen was incubated with control serum. Specific IgG anti-phospholipase A antibodies were measured with a double-antibody method based on the Farr technique with modifications including those described previously.22 la51-labeled phospholipase A was mixed with the serum to be tested, allowed to react for 4 hr at 25” C, and then the serum IgG precipitated with a specific goat anti-human y chain. After 18 hr at 4” C the precipitate was washed free of unbound antigen and the specific leaI-phospholipase A determined. Specific IgE antibodies were measured with the RAST technique described by Wide, Bennich, and Johanssonaa with few modifications.
RESULTS
Gel filtration chromatography of honeybee venom yielded 4 protein peaks, with molecular weights of approximately 30,000, 23,000, 4,000, and 1,500, respectively. A typical elution pattern is shown in Fig. 1. A fifth peak, seen at about tube 100, was not further studied since the low molecular weight substances, including histamine, which it contains are of little interest as allergens. Gel diffusion analysis of the first 80 tubes with a rabbit antiserum to whole bee
32
J. ALLERGY CLIN. IMMUNOL. JANUARY 1976
Sobotko et al.
Stgma 52 BeeVenom G-50 Fine Sephodex
30
40
50
60
70
80
90
iO0
I10
PO
TUBE NUMBER
FIG. 1. Elution pattern of honeybee venom on G50 Sephadex. Protein was measured the Folin Cittreleiu reaction. The maior protein component of each peak is indicated.
with
venom revealed the presence of an antigen between tubes 30 and 36, and a second in tubes 38 through 48, which overlapped into tubes 50 through 63. Elution tubes beyond the sixty-fifth showed only nonimmune precipita.tion24 as ascertained when normal rabbit serum was used as a control. With the use of information obtained from protein determinations and from the antigenic patterns, aliquots of the samples in each tube comprising the peaks were pooled. As noted above, they were largely made up of (I) hyaluronidase, (II) phospholipase A, (III) melittin, and (IV) apamin. These 4 pools were not further purified for this study; each probably contains trace amounts of the adjacent protein(s) . The leukocytes of 6 patients with clear-cut clinical and laboratory sensitivity to honeybee venom were challenged in vitro with unfractionated honeybee venom and each of the four chromatographic fractions. The results of two typical experiments are shown in Fig. 2. The leukocytes of both patients were most sensitive to Pool II containing phospholipase A; 50% histamine release occurred at 3 x 1O-5 and 5 x lo+ ag protein/ml, respectively. The same response to unfractioned venom required 3- to lo-fold more protein and challenge with Pools I, III, and IV showed that lOO- to l,OOO-fold more protein was required to elicit 50% histamine release. Some patients were insensitive to one or more venom components (e.g., Pool I, Fig. 2, upper panel). Similar patterns of release were found with the leukocytes of the 6 allergic patients tested, whereas the response of 6 normal control subjects to bee venom or to the protein pools was uniformly negative (Table I). Since Pool II appeared to be the most active we explored its homogeneity by studying aliquots of the individual elution tubes (38-56) for their ability to release histamine from sensitive leukocytes. Each tube was equalized as to protein content and tested over a concentration range from 1O-6-1Opg/ml. The results of experiments with 2 allergic patients are shown in Fig. 3. Two peaks of allergenic activity were found; one occurred in tube 42, the second in tubes 52 to 54.
VOLUME 57 NUMBER 1
Allergy
TABLE I. Summary
of histamine
release
data
Protein
Patient
S. M. P. p.* R. W. N. M. K. B.* J. V. ;. . zg.
Honeybee venom
2.0 1.0 ::X K > 10.000 3,&
Phospholipase Siama
(q/ml)
using bee venom for 50% histamine
to insect stings.
and bee venom
::t 2 0:04 > 10.000 10,ooo
Pool I
150 1,000 >l,ooo 1,000 ND > 100 > 10,000 10.000
Pool I I
Pool IV
Pool III
1.0
2.0
ND 3,000
3.0 20 5.0 > 10,000 10,ooo
NNE 1,200 50 >1o,ooo > 10,000
2: i: 2:o 1.0 0.05 >lO.ooo > 10,000
33
components
release
A
0.3
II
Yellow jacket venom
> 10,000 ND > 10,000 > 10,000 ND > 10,000 > 10,000 20
‘Passive sensitization. tVespid allergic. :Nonnllergic control.
These peaks bear little relation to the protein content but rather correspond to the ascending and descending limbs of the Pool II peak. Essentially the same protein concentration (S 3 x lo-* pg) was required to elicit a 50% response when either tube 42 or 52 was tested; when Pool II itself was studied on the leukocytes of the same patients, 5 x lo-” pg protein/ml was sufficient to elicit 50% histamine release. Commercially purified phospholipase A is readily obtained and was compared for potency with Pool II and the unfractionated venom. Two such assays are shown in Fig. 4. The concentrations of Pool II and of purified phospholipase A required for 507% histamine release were essentially the same; both patients were more sensitive to this venom component than to the unfractionated venom. Also shown are the results of exposing the cells of normal donors to phospholipase A; they were completely negative here as they were in 6 other experiments, in which concentrations of phospholipase A up to 10 ,ug/ml were used. Histamine release data obtained with the use of unfractionated venom, the 4 venom pools, and commercial phospholipase A are summarized in Table I. Results obtained with yellow jacket venom are included for comparison. All 6 honeybee-allergic patients reacted to nanogram amounts of honeybee venom, commercial phospholipase A, and Pool II. With the exception of one passive sensitization experiment (P. P.), all patients were ??lore sensitive to phospholipase A and Pool II than to unfractionated venom. Their sensitivity to the other 3 pools was variable; most patients required much higher protein concentrations for histamine release. Pool I containing hyaluronidase and Pool IV containing apamin caused histamine release from sensitive leukocytes but at protein concentrations = l,OOO-fold greater than that required to elicit a response when Pool II was used. Pool III containing melittin appeared to be of significance as an antigen in several patients studied. Patients N. M. and J. V. both reacted to nanogram amounts of Pool III. It is possible, however, that Patient J. V. was responding to a 1% contamination of phospholipase A present in Pool III. Moreover, the protein concentration of Pool III required for histamine release varied widely and in all but one patient (N. M.) a 50- to
34
Sobotka
J. ALLERGY CLIN. IMMUNOL. JANUARY 1976
et al.
,o-6
I 10-5
1 10-4
I 10-3
I lo-2
I 10-l
I I
I
IO CONCENTRATION~pq/ml)
FIG. 2. Two experiments comparing histamine release to that released with the 4 protein pools from honeybee
elicited venom.
by unfractionated
venom
l,OOO-fold increase in protein over that required with Pool II was necessary to cause significant ( > 50%) release. Pool III was tested without effect on the leukocytes of control patients at concentrations up to 10 pg/ml to rule out the nonspecific histamine-releasing activity of melittin. This occurs only at concentrations 1 100 pg/ml. Further studies are necessary to determine whether the activity found in Pool III was due to melittin or to contamination of this pool with phospholipase A. Antibody
studies
If phospholipase A is the major allergen in honeybee venom, IgE antibodies against this antigen should be readily detectable. The sera of all 6 patients evaluated in this study were found to contain such antibodies. Three titration curves and appropriate controls are shown in Fig. 5. Control sera from patienm sensitive to other species of Hymenoptera and from nonallergic controls did not contain measurable IgE antibodies to phospholipase A. Similarly, if phospholipase A is the major allergen of honeybee venom, specific antibodies against this material should markedly decrease the activity of whole bee venom. Anti-phospholipase A markedly inhibited histamine release
VOLUME 57 NUMBER 1
Allergy
I-
38
40
PtwsphohpaseA
42
44
46
Peak-Pool II
4.3
50
52
to insect stings.
II
35
1
54
56
TUBE NUMBER FIG.
3. Histamine
release
with protein from the protein peak, visualized
from
individual at O.D.
the
leukocytes
of
tubes comprising 280 nM, is shown
2 patients
sensitive
Pool II (black triangles by the open circles.
to
honeybee and
venom
circles).
The
induced by unfractionated venom, Pool II, and phospholipase A. The results of 2 such experiments are shown in Fig. 6. A l/1000 dilution of the rabbit serum caused >SO% inhibition of antigenic histamine release from the cells of both patients. In both cases, a lower concentration of the monospecific antisera was required to cause 5070 inhibition of phospholipase A-induced histamine release than that necessary when the antigen was unfractionated venom. These results led us to determine whether IgG anti-phospholipase A antibodies were produced in human subjects as a result of bee stings. Consequently, we studied the sera of 10 multiply stung beekeepers and from two patients treated with whole bee venom. Antibodies of the IgG class able to bind 1251labeled phospholipase A were found in all of the sera studied. Sera from control patients who could not recall having ever been stung lacked measurable IgG antibodies against phospholipase A, as did the sera of 5 vespid-allergic patients. The results from a typical experiment are shown in Fig. 7. It is of interest that at a l/1000 dilution, sera from beekeepers and from venom-treated patients bind similar amounts of phospholipase A. Sera from honeybee-allergic patients, from patients treated with mixed insect extracts, and from some nonallergic control subjects who had recently been stung also bound measurable amounts of the labeled phospholipase A; however, all such sera contained lOO- to l,OOO-fold
36
Sobotka
J. ALLERGY CLIN. IMMUNOL. JANUARY 1976
et al.
I 2
.
dSigma
10-4
,WIl,lYl Siqmo Pbs A - ~ont,ol A----&----L\ I I I
” y --,,-“’ Phos-A I
I
Icy3
lo-z
I
l65
lo-4
PROTEIN
lo-3
lo-2
(pg/ml)
FIG. 4. Two experiments comparing histamine release from honeybee-sensitive leukocytes using unfractionated honeybee venom, commercial phospholipase A, and Pool II as the antigens. Note the lack of histamine release with the use of phospholipase A on normal control eel Is.
16 8
1 14 -
0 x
12-
F
IO -
= r
B-
5 G
6-
g
4-
-001
0.0 I
01 SERUM
FIG. 5. Measurement allergic, and normal
of IgE anti-phospholipase sera. Note the low level
IO
DILUTION
A antibodies in honeybee-allergic, of antibody in all control sera.
vespid-
less antibody than that from persons repeatedly injected with venom. These data will be presented in detail in a subsequent report. Sera from the beekeepers and the two venom-treated patients were also found to significantly inhibit both venom- and phospholipase-induced histamine release (data not shown). In most experiments, complete inhibition of histamine release was obtained with a l/1000 dilution of serum. DISCUSSION This report suggests that phospholipase A is the major allergen in honeybee venom. All allergic patients responded to this venom protein at concentrations
VOLUME 57 NUMBER 1
Allergy
,001
.ooot RABBIT
FIG. 6. Inhibition of histamine phospholipase A. The antibody phospholipase A, and Pool II.
100
ANTI-
.Oi
.00001
PHOS-A
.oool
to insect stings.
,001
II
37
.Oi
CONCENTRATION
release from bee-sensitive leukocytes by rabbit *abbit antiis effective in inhibiting honeybee venom venom, commercial comn iercial
-
Ig G Antlbody to PhosphollpaseA Total counts- 75,400 Blank 1,000
z! 2
80Venom Treoted
a .OOOl
,001 l/SERUM
FIG. 7. IgG anti-phospholipase A in sera from Note the lack of specific antibodies in control
.Ol
0.1
CONCENTRATION
multiply sera.
stung and venom-treated
patients.
equal to or less than that required to elicit histamine release with the unfractionated venom, while showing little sensitivity to the other venom proteins. The 6 patients were selected because they were found to be sensitive to honeybee venom and not to venoms from other Hymenoptera classes as judged by laboratory and clinical criteria. Control patients and patients sensitive to other Hymenoptera species were insensitive to bee venom and to its protein components. IgE antibodies specific for phospholipase A were also found in the sera of honeybee-allergic patients. Reisman and associates5 have data in support of this observation but also report a small number of patients whose serum contained IgE antibodies to unfractionated venom although no detectable
38
Sobotka
et al.
J. ALLERGY CLIN. IMMUNOL. JANUARY 1976
antibodies to phospholipase A could be found. This discrepancy is difficult to explain. It is to be noted, however, that the RAST test is difficult to interpret when the antigen used is not clean; indeed this is one of the reasons this study was initiated. The presence of specific IgG antibodies against phospholipase A in the sera of multiply stung persons also substantiates the importance of this venom component as a major antigen, as do our results with a monospecific phospholipase A antisera. Although, as expected, less antiserum was required to inhibit phospholipase A histamine release, inhibition of venom-induced release required only a small increase in antiserum concentration. This is a preliminary study. A more definitive investigation requires that each protein be isolated by multiple chromatographic steps until greater than 99% homogeneity can be demonstrated. Such studies are under way. Their importance is attested to by the clear demonstration of multiple allergenic forms of phospholipase A in Pool II. It has previously been reported that phospholipase A exists in several molecular species differing in molecular weight, probably on the basis of carbohydrate differences. I389I7 Two peaks of antigenic activity were found in Pool II that differed in size but not in antigenie or allergenic activity. Most major allergens, such as ragweed antigen E’” and rye grass Group I,26 exist in nature as a series of isoallergens of slightly different composition but with a complete antigenie and allergenic cross-reactivity. It appears that phospholipase A shares this characteristic. We feel that the establishment of the major allergenic component in bee venom is of importance as a diagnostic tool which can be rapidly exploited. Phospholipase A is readily available commercially in large quantities and could be used diagnostically in the RAST test or in skin testing. The ability to predict sensitivity to honeybees in the majority of patients by means of the RAST technique will serve to alleviate many of the diagnostic problems facing physicians and perhaps to prevent unnecessary treatment with irrelevant antigens. The same comments relate to the utility of phospholipase A in assessing the titer of IgG antibodies in monitoring the response to treatment. Recent reports from our laboratory6 and from Busse and associates7 support the concept of a role for IgG anti-phospholipase A antibodies in protection against anaphylaxis. Both venom-treated patients reached IgG anti-phospholipase A antibody titers equal to those found in the sera of beekeepers and were subsequently stung without reaction. Finally, the establishment of phospholipase A as the major allergen of honeybee venom is important for standardization. It has been shown that the potency of whole pollen extracts, such as that from grass or ragweed pollens, can be more meaningfully evaluated when the standardization is based on the content of a known major allergen than when it is judged with protein nitrogen units or per cent solids as criteria. 27 The phospholipase A content can be used to standardize and quantitate honeybee venom and, if necessary, to assess whether the whole body extracts have any venom proteins.
VOLUME 57
Allergy
to insect stings.
II
39
NUMBER 1
NOTE ADDED IN PROOF Our recent studies have revealed that hyaluronidase (Peak I) is rapidly denatured after isolation. If this is prevented, hyaluronidase is antigenic in most patients (King et al.: Allergens of honeybee venom. In press, Arch. Biochem. Biophys.) However, all honeybeeallergic patients are highly sensitive to Phos-A as judged by several immunologic parameters.
REFERENCES Schwartz, H. J.: Skin sensitivity in insect allergy, J. A. M. A. 194: 703, 1965. from honey bees, Benton, A. W., Morris, R. A., and Stua.rt, J. D.: Venom collection Science 142: 228, 1963. Loveless, M. H.: Immunization in wasp sting allergy through venom repositories and periodic insect sting, J. Immunol. 89: 204, 1962. Sobotka, A., Valentine, M. D., Benton, A., and Lichtenstein, L.: Allergy to insect stings. I. Diagnosis of IgE mediated Hymenoptera sensitivity by venom induced histamine release,J. ALLERGY CLIN. IMMUNOL.~~: 170,1974. 5 Reisman, R. E., Wypych, J. I., Yeagle, N., and Arbesman, S. E.: Stinging insect hypersensitivity. III. Detection and clinical significance of IgE antibodies to insect venom in in man, J. ALLERGY CLIN. IMMUNOL. 53: 110,1974. 6 Lichtenstein, L. M., Sobotka, A. K., and Valentine, M. D.: A case for venom treatment in anaphylactic sensitivity to Hymenoptera sting, N. Engl. J. Med. 290: 1223, 1974. 7 Busse, W., Reed, S. E., Lichtenstein, L. M., and Reisman, R. E.: Protection following honey bee venom immunotherapy in a case of bee sting anaphylaxis, J. ALLERGY CLIN.
IMM~NOL. 53: 104,1974 8 Jaques, R., and Schacter, M.: The presence of histamine,5-hydroxytryptamine and a potent, slow contracting substance in wasp venom, Br. J. Pharmacol. 9: 53, 1954. 9 Benson, R. L., and Semenov, H.: Allergy in its relation to bee sting, J. ALLERGY 1:
105,193o. 10 O’Connor, R., Rosenbrook, Jr., W. venoms and body proteins, Science lla Shulman, S., Langlois, C., and insects. I. Preparation of extracts 446, 1964. llb Shulman, S., Langlois, C., Miller, stinging insects. V. Fractionation
and Erickson, R.: Dioc electrophoresis of Hymenoptera 145: 1320, 1964. Arbesman, C. E.: The allergic response to stinging and their biochemical characterization, J. ALLERGY 35: J. B., and Arbesman, C. E.: of whole body and venom
ALLERGY 37: 350, 1966. 11~ Langlois, C., Shulman, S., Kozmycz, stinging
insects.
VI.
Fractionation
The allergic sac extracts
response to of bee, J.
S., and Arbesman, C. E. : The allergic of whole body and venom sac extracts
response to of wasp, J.
ALLERGY 37: 359, 1966. lld
12 13a 13b 14 15
16 17 18
Arbesman, C. E., Langlois, C., Bronson, P., and Shulman, S. : The allergic response to stinging insects. VII. Fractionation of whole body and venom sac extracts of yellow jacket, J. ALLERGY 38: 1, 1966. Habermann E.: Bee and wasp venoms, Science 177: 314, 1972. Shipolini, R. A., Callewaert, R. C., Doonan, S., Vernon, C. A., and Banks, B. E. C.: Phospholipase A from bee venom, Eur. J. Biochem. 20: 459, 1971. Shipolini, R. A.: The primary sequence of phospholipase A from bee venom, Febs. Letters 17: 39, 1971. Liibke, K., Matthes, S., and Kloss, G.: Isolation and structure of N-l formyl mellitin, Experimentia 27: 765, 1971. Schrader, E., Liibker, K., Lehmann, M., and Beetz, I.: Haemolytic activity and action on the surface tension of aqueous solutions of synthetic mellitins and their derivatives, Experientia 27: 764, 1974. from honey bees, Benton, A. W., Morris, R. A., and Stuart, J. D.: Venom collection Science 142: 228, 1963. Franklin, R., and Baer, H.: Comparison of honey bee venoms and their components from various sources, J. ALLERGY CLIN. IMMITNOL. 55: 285,1975, Lichtenstein, L. M., and Osler, A. G.: Studies on the mechanism of hypersensitivity phenomena. IX. Histamine release from human leukocytes by ragweed pollen antigen, J. Exp. Med. 120: 507, 1964.
40
Sobotka et al.
J. ALLERGY CLIN. IMMUNOL. JANUARY 1976
study 19 May, C. D., Lyman, M., Alberto, R., and Cheng, J.: Procedures for immunochemical of histamine release from leukocytes with small volume of blood, J. ALLERGY 46: 12, 1970. 20 Levy, D. A., and Osler, A. G.: Studies on the mechanisms of hypersensitivity phenomena. XIV. Passive sensitization in vitro of human leukocytes to ragweed pollen antigen, J. Immunol. 97: 203, 1966. 21 Lichtenstein, L. M., and Osler, A. Cr.: Studies on the mechanism of hypersensitivity study of the reaction between ragweed pollen antigen, phenomena. XII. An ia vitro allergic human serum and ragweed sensitive cells, J. Immuol. 96: 169, 1966. 22 Newcomb, R. W., and Ishizaka, K.: Human diptheria antitoxin in immunoglobulin classes IgG and IgA. J. Immunol. 99: 40, 1967. 23 Wide, L., Bennich, H., and Johansson, S. G. 0.: Diiagnosis of allergy by in vitro test for allergen antibodies, Lancet 2: 1105, 1967. 24 Franklin, R., and Baer, H.: Immune and nonimmune precipitates produced by honey bee venom and its components. In press, Proc. Sot. Exp. Biol. Med. 147: 585-588, 1974. 25 King, T. P., Norman, P. S., and Connell, J. T.: Isolation and characterization of allergens from ragweed pollen. II, Biochemistry 3: 458, 1964. 26 Johnson, I’., and Marsh, D. G.: Isoallergens from rye grass pollen, Nature 206: 935, 1965. 27 Baer, H., Godfrey, H., Malony, C. J., Norman, P. S., and Lichtenstein, L. M.: The potency and antigen E content of commercially prepared ragweed extracts, J. ALI,EK(:P 45: 347, 1970.
A: The major allergen
Anne K. Sobotka, Martin Valentine,
Rudolph Harold
Baltimore.
in honeybee
M. Franklin, N. Franklin Baer,** and Lawrence
venom
Adkinson, Jr.,* M. Lichtenstein***
Md.
In order to determine the proteins of major allergenic importance in honeybee venom (Apis mellifera) it was chromatographed on G-50 Sephadex. The four major protein peaks eluted were identified as hyalzlronidase, phospholipase, melittin, and apamin. Testing these preparations on the leukooytes of 6 honeybee-sensitive patients, with the in vitro method of histamilte release, revealed that all individuals were most sensitive to phospholipase A. IgE antibodies against phosptilipase A (RAST) were found in the sera of honeybee-sensitive patients and IgG antibodies to this venom component were found in the sera from beekeepers and venomtreated patients. Melittin appeared to be allergenic in several patients, but the results were variable and were possibly due to contamination with phospholipase. All patients were insensitive to the hyaluronidose and apamin preparations. We conclude that phospholipase A is the major allergen of honeybee venom and, since this protein is readily available, it should be usefd for diagnostic and therapeutic studies as uell as for the standardization of materials used in the management of honeybee-sensitive patients.
The study of Hymenoptera sensitivity has been severely hampered by the use of impure and largely irrelevant extracts prepared from the whole ground bodies of these insects. Because of their widespread use, specific diagnosis was impossible’! 2 and even the “allergic” pathogenesis of the disease was legitimately called into question. Although some voices were raised against this practice, notably that of Loveless,3 this state of affairs has persisted for the last 30 years. We and others have recently used more modern tools to reinvestigate this problem and have been able to clearly demonstrate the utility of venoms in the diagnosis of Hymenoptera sensitivity. ** 5 It has, moreover, been possible to clearly implicate the IgE antibodies as the pathogenetic mechanism; preliminary therapeutic experience appears to confirm the importance of venoms.“9 i It is, of course, recognized that the venoms themselves contain multiple proteins From The Johns Hopkins University School of Medicine. Supported by Grants Nos. A107290, AI08270, and AI10304 from the National Institute of Allergy and Infectious Diseases, National Institutes of Health. Received for publication Sept. 12, 1974. Accepted for publication Jan. 10, 1975. Reprint requests to: Anne K. Sobotka, The Good Samaritan Hospital, 5601 Loch Raven Blvd., Baltimore, Md. 21239. *Allergic Diseases Academic Award. **From the Bureau of Biologics, Food and Drug Administration, Rockville, Md. ***Recipient of a Research Career Development Award from the National Institute of Allergy and Infectious Diseases, National Institutes of Health. Vol. 57, No. 1, pp. 29-40
30
Sobotka et al.
J. ALLERGY CLIN. IMMUNOL. JANUARY 1976
and therefore several potential allergens. To further assist diagnosis it is important to identify them; it is also imperative to have relatively clean allergens to use the RAST technique effectively and the same requirement exists for IgQ antibody measurements. We have begun this work with the best characterized venom, that of the honeybee (Rpis mellifera), in an effort to determine the major venom allergen and to obtain a preparation pure enough to carry out the in vitro diagnostic tests cited. Early studies of insect venoms by Jaques and Schaete9 revealed the presence of various vasoactive amines and kinins in wasp venom. Following the work of Benson and Semenov,B who found a beekeeper sensitive to both bee venom and whole insect extract, O’Connor, Rosenbrook, and Erickson1o and Shulman, Langlois, Arbesman, and their associateP’, b, c, d compared the antigenic content of whole insect bodies with that of venom sac preparations. Both groups found several antigens in the venoms not present in sacless bodies. Many investigators have studied the protein content of various pure venoms, especially that from the honeybee. The classic work of Habermann’2 led to the isolation and characterization of many of the components of the venoms of several Hymenoptera classes. In the case of the honeybee, hyaluronidase, melittin, phospholipase, and apamin make up great,er than ‘7070 of the venom protein. Several honeybee venom proteins have been well characterized and Shipolini and associatePa’ b have determined the primary amino acid sequence of the phospholipase A. in Melittin has also been sequencedI and then synthesized by investigators Gerrnany.15 This work has left unanswered, however, the question of which proteins in the venom were responsible for inducing the allergic response. The present study is directed to that question. With the use of simple chromatographic techniques and the leukoc.yte-histamine release model of “in vitro our results indicated that phospholipase A is the major allergen anaphylaxis,” in honeybee venom. METHODS Antigens A The antigens used for these studies were honeybee venom, commercial phospholipase (Sigma Chemical Company, Saint Louis, MO.) and 4 pools obtained from commercial honeybee venom by Sephadex G50 chromatography. The honeybee venom used was collected both by electrical stimulation and by allowing the insects to sting through a cellophane membrane.16 The dried venom was weighed in milligram quantities, diluted in Tris-buffered saline, dialyzed, and stored at -20” C as previously described.4 Phospholipase A from bee venom was treated similarly except that dialysis, used to remove substances from whole bee venom that interfere with the fluorometric assay for histamine, was not necessary. Separation of honeybee venom components was carried out as previously descrihed.17 A Scphadex G50 fine (Pharmacia, Piscatamay, N. J.) gel column was prepared and calibrated with blue dextran 2000 (Pharmacia) and with chymotrypsinogen, cytochrome C, and insulin as molreular weight markers. Then 500 mg of bee venom were placed on the column, the cluntes were collected at a constant volume of 4.8 ml, and the prot,ein content, of each tube was determined at an 0.D. of 280 nM. Alternatively (see below) the Folin-Ciocaltcu reaction was used to tstimate protein. Chromatography yielded 4 major protein peaks. Portions of the individual tubes comprising each peak mere pooled, equalized as to protein content., aliquoted, and stored at -20” C until assayed. The individual eluates were also aliquoted and frozen until used.
VOLUME 57 NUMBER 1
Allergy
to insect stings.
Previously described functional studies17 identified three pools as being largely of hyaluronidase (Pool I), phospholipase A (Pool II), and melittin (Pool III). Habermann’s data,12 Pool IV was assumed to contain apamin.
II
31
composed Based on
Antisera Antiserum to phospholipase A was prepared in two New Zealand white rabbits (Bunnyville Farms, R.D. 2, Littlestown, Pa.) by injecting 500 cg/ml in complete Freund’s adjuvant (Difco Laboratories, Detroit, Mich.) into both rear footpads. Two weeks later, the rabbits received an intramuscular injection of 500 pg of phospholipase A in saline and were bled 2 weeks later. The antiserum used in our in vitro studies was a pool made from this final bleeding; on immunodiffusion a single line of identity between purified phospholipase A and unfractionated honeybee venom was observed. Rabbit anti-whole bee venom was prepared according to a similar regimen; a total dose of 11 mg of venom was injected into each rabbit and the pool was made from several bleedings. Human antisera from 10 beekeepers was generously provided by Dr. A. W. Benton; sera from venom-treated patients were provided by Dr. Charles Reed and from our own practice. Drs. Halla Brown and Solomon Barr provided sera from patients on long-term therapy with whole-body insect extracts. Other sera from similar patients were obtained in our Allergy Clinic.
Histamine
release
studies
Patient sensitivity to honeybee venom, phospholipase A, and bee venom components was assessed with the leukocytes of known honeybee-allergic patients and normal control subjects as previously described.% 19 Briefly, leukocytes were collected by dextran sedimentation, washed, resuspended in TACM (Tris-buffered saline containing calcium, magnesium, and human albumin), and mixed with appropriate concentrations of each antigen, usually covering a range of 10-5-10 fig of protein per milliliter. After 60 min at 37” C the histamine release was determined fluorometrically. For passive sensitization experiments the recipient cells, previously shown to be nonreactive to honeybee venom, were sensitized following the techniques described by Levy and Osler.eo
Antibody
studies
Assessment of serum blocking antibodies using histamine release was carried out as previously described21 : appropriate dilutions of serum were incubated with the antigens for 30 min at 25” C prior to the addition of leukocytes and the amount of histamine release in these tubes compared with those in which antigen was incubated with control serum. Specific IgG anti-phospholipase A antibodies were measured with a double-antibody method based on the Farr technique with modifications including those described previously.22 la51-labeled phospholipase A was mixed with the serum to be tested, allowed to react for 4 hr at 25” C, and then the serum IgG precipitated with a specific goat anti-human y chain. After 18 hr at 4” C the precipitate was washed free of unbound antigen and the specific leaI-phospholipase A determined. Specific IgE antibodies were measured with the RAST technique described by Wide, Bennich, and Johanssonaa with few modifications.
RESULTS
Gel filtration chromatography of honeybee venom yielded 4 protein peaks, with molecular weights of approximately 30,000, 23,000, 4,000, and 1,500, respectively. A typical elution pattern is shown in Fig. 1. A fifth peak, seen at about tube 100, was not further studied since the low molecular weight substances, including histamine, which it contains are of little interest as allergens. Gel diffusion analysis of the first 80 tubes with a rabbit antiserum to whole bee
32
J. ALLERGY CLIN. IMMUNOL. JANUARY 1976
Sobotko et al.
Stgma 52 BeeVenom G-50 Fine Sephodex
30
40
50
60
70
80
90
iO0
I10
PO
TUBE NUMBER
FIG. 1. Elution pattern of honeybee venom on G50 Sephadex. Protein was measured the Folin Cittreleiu reaction. The maior protein component of each peak is indicated.
with
venom revealed the presence of an antigen between tubes 30 and 36, and a second in tubes 38 through 48, which overlapped into tubes 50 through 63. Elution tubes beyond the sixty-fifth showed only nonimmune precipita.tion24 as ascertained when normal rabbit serum was used as a control. With the use of information obtained from protein determinations and from the antigenic patterns, aliquots of the samples in each tube comprising the peaks were pooled. As noted above, they were largely made up of (I) hyaluronidase, (II) phospholipase A, (III) melittin, and (IV) apamin. These 4 pools were not further purified for this study; each probably contains trace amounts of the adjacent protein(s) . The leukocytes of 6 patients with clear-cut clinical and laboratory sensitivity to honeybee venom were challenged in vitro with unfractionated honeybee venom and each of the four chromatographic fractions. The results of two typical experiments are shown in Fig. 2. The leukocytes of both patients were most sensitive to Pool II containing phospholipase A; 50% histamine release occurred at 3 x 1O-5 and 5 x lo+ ag protein/ml, respectively. The same response to unfractioned venom required 3- to lo-fold more protein and challenge with Pools I, III, and IV showed that lOO- to l,OOO-fold more protein was required to elicit 50% histamine release. Some patients were insensitive to one or more venom components (e.g., Pool I, Fig. 2, upper panel). Similar patterns of release were found with the leukocytes of the 6 allergic patients tested, whereas the response of 6 normal control subjects to bee venom or to the protein pools was uniformly negative (Table I). Since Pool II appeared to be the most active we explored its homogeneity by studying aliquots of the individual elution tubes (38-56) for their ability to release histamine from sensitive leukocytes. Each tube was equalized as to protein content and tested over a concentration range from 1O-6-1Opg/ml. The results of experiments with 2 allergic patients are shown in Fig. 3. Two peaks of allergenic activity were found; one occurred in tube 42, the second in tubes 52 to 54.
VOLUME 57 NUMBER 1
Allergy
TABLE I. Summary
of histamine
release
data
Protein
Patient
S. M. P. p.* R. W. N. M. K. B.* J. V. ;. . zg.
Honeybee venom
2.0 1.0 ::X K > 10.000 3,&
Phospholipase Siama
(q/ml)
using bee venom for 50% histamine
to insect stings.
and bee venom
::t 2 0:04 > 10.000 10,ooo
Pool I
150 1,000 >l,ooo 1,000 ND > 100 > 10,000 10.000
Pool I I
Pool IV
Pool III
1.0
2.0
ND 3,000
3.0 20 5.0 > 10,000 10,ooo
NNE 1,200 50 >1o,ooo > 10,000
2: i: 2:o 1.0 0.05 >lO.ooo > 10,000
33
components
release
A
0.3
II
Yellow jacket venom
> 10,000 ND > 10,000 > 10,000 ND > 10,000 > 10,000 20
‘Passive sensitization. tVespid allergic. :Nonnllergic control.
These peaks bear little relation to the protein content but rather correspond to the ascending and descending limbs of the Pool II peak. Essentially the same protein concentration (S 3 x lo-* pg) was required to elicit a 50% response when either tube 42 or 52 was tested; when Pool II itself was studied on the leukocytes of the same patients, 5 x lo-” pg protein/ml was sufficient to elicit 50% histamine release. Commercially purified phospholipase A is readily obtained and was compared for potency with Pool II and the unfractionated venom. Two such assays are shown in Fig. 4. The concentrations of Pool II and of purified phospholipase A required for 507% histamine release were essentially the same; both patients were more sensitive to this venom component than to the unfractionated venom. Also shown are the results of exposing the cells of normal donors to phospholipase A; they were completely negative here as they were in 6 other experiments, in which concentrations of phospholipase A up to 10 ,ug/ml were used. Histamine release data obtained with the use of unfractionated venom, the 4 venom pools, and commercial phospholipase A are summarized in Table I. Results obtained with yellow jacket venom are included for comparison. All 6 honeybee-allergic patients reacted to nanogram amounts of honeybee venom, commercial phospholipase A, and Pool II. With the exception of one passive sensitization experiment (P. P.), all patients were ??lore sensitive to phospholipase A and Pool II than to unfractionated venom. Their sensitivity to the other 3 pools was variable; most patients required much higher protein concentrations for histamine release. Pool I containing hyaluronidase and Pool IV containing apamin caused histamine release from sensitive leukocytes but at protein concentrations = l,OOO-fold greater than that required to elicit a response when Pool II was used. Pool III containing melittin appeared to be of significance as an antigen in several patients studied. Patients N. M. and J. V. both reacted to nanogram amounts of Pool III. It is possible, however, that Patient J. V. was responding to a 1% contamination of phospholipase A present in Pool III. Moreover, the protein concentration of Pool III required for histamine release varied widely and in all but one patient (N. M.) a 50- to
34
Sobotka
J. ALLERGY CLIN. IMMUNOL. JANUARY 1976
et al.
,o-6
I 10-5
1 10-4
I 10-3
I lo-2
I 10-l
I I
I
IO CONCENTRATION~pq/ml)
FIG. 2. Two experiments comparing histamine release to that released with the 4 protein pools from honeybee
elicited venom.
by unfractionated
venom
l,OOO-fold increase in protein over that required with Pool II was necessary to cause significant ( > 50%) release. Pool III was tested without effect on the leukocytes of control patients at concentrations up to 10 pg/ml to rule out the nonspecific histamine-releasing activity of melittin. This occurs only at concentrations 1 100 pg/ml. Further studies are necessary to determine whether the activity found in Pool III was due to melittin or to contamination of this pool with phospholipase A. Antibody
studies
If phospholipase A is the major allergen in honeybee venom, IgE antibodies against this antigen should be readily detectable. The sera of all 6 patients evaluated in this study were found to contain such antibodies. Three titration curves and appropriate controls are shown in Fig. 5. Control sera from patienm sensitive to other species of Hymenoptera and from nonallergic controls did not contain measurable IgE antibodies to phospholipase A. Similarly, if phospholipase A is the major allergen of honeybee venom, specific antibodies against this material should markedly decrease the activity of whole bee venom. Anti-phospholipase A markedly inhibited histamine release
VOLUME 57 NUMBER 1
Allergy
I-
38
40
PtwsphohpaseA
42
44
46
Peak-Pool II
4.3
50
52
to insect stings.
II
35
1
54
56
TUBE NUMBER FIG.
3. Histamine
release
with protein from the protein peak, visualized
from
individual at O.D.
the
leukocytes
of
tubes comprising 280 nM, is shown
2 patients
sensitive
Pool II (black triangles by the open circles.
to
honeybee and
venom
circles).
The
induced by unfractionated venom, Pool II, and phospholipase A. The results of 2 such experiments are shown in Fig. 6. A l/1000 dilution of the rabbit serum caused >SO% inhibition of antigenic histamine release from the cells of both patients. In both cases, a lower concentration of the monospecific antisera was required to cause 5070 inhibition of phospholipase A-induced histamine release than that necessary when the antigen was unfractionated venom. These results led us to determine whether IgG anti-phospholipase A antibodies were produced in human subjects as a result of bee stings. Consequently, we studied the sera of 10 multiply stung beekeepers and from two patients treated with whole bee venom. Antibodies of the IgG class able to bind 1251labeled phospholipase A were found in all of the sera studied. Sera from control patients who could not recall having ever been stung lacked measurable IgG antibodies against phospholipase A, as did the sera of 5 vespid-allergic patients. The results from a typical experiment are shown in Fig. 7. It is of interest that at a l/1000 dilution, sera from beekeepers and from venom-treated patients bind similar amounts of phospholipase A. Sera from honeybee-allergic patients, from patients treated with mixed insect extracts, and from some nonallergic control subjects who had recently been stung also bound measurable amounts of the labeled phospholipase A; however, all such sera contained lOO- to l,OOO-fold
36
Sobotka
J. ALLERGY CLIN. IMMUNOL. JANUARY 1976
et al.
I 2
.
dSigma
10-4
,WIl,lYl Siqmo Pbs A - ~ont,ol A----&----L\ I I I
” y --,,-“’ Phos-A I
I
Icy3
lo-z
I
l65
lo-4
PROTEIN
lo-3
lo-2
(pg/ml)
FIG. 4. Two experiments comparing histamine release from honeybee-sensitive leukocytes using unfractionated honeybee venom, commercial phospholipase A, and Pool II as the antigens. Note the lack of histamine release with the use of phospholipase A on normal control eel Is.
16 8
1 14 -
0 x
12-
F
IO -
= r
B-
5 G
6-
g
4-
-001
0.0 I
01 SERUM
FIG. 5. Measurement allergic, and normal
of IgE anti-phospholipase sera. Note the low level
IO
DILUTION
A antibodies in honeybee-allergic, of antibody in all control sera.
vespid-
less antibody than that from persons repeatedly injected with venom. These data will be presented in detail in a subsequent report. Sera from the beekeepers and the two venom-treated patients were also found to significantly inhibit both venom- and phospholipase-induced histamine release (data not shown). In most experiments, complete inhibition of histamine release was obtained with a l/1000 dilution of serum. DISCUSSION This report suggests that phospholipase A is the major allergen in honeybee venom. All allergic patients responded to this venom protein at concentrations
VOLUME 57 NUMBER 1
Allergy
,001
.ooot RABBIT
FIG. 6. Inhibition of histamine phospholipase A. The antibody phospholipase A, and Pool II.
100
ANTI-
.Oi
.00001
PHOS-A
.oool
to insect stings.
,001
II
37
.Oi
CONCENTRATION
release from bee-sensitive leukocytes by rabbit *abbit antiis effective in inhibiting honeybee venom venom, commercial comn iercial
-
Ig G Antlbody to PhosphollpaseA Total counts- 75,400 Blank 1,000
z! 2
80Venom Treoted
a .OOOl
,001 l/SERUM
FIG. 7. IgG anti-phospholipase A in sera from Note the lack of specific antibodies in control
.Ol
0.1
CONCENTRATION
multiply sera.
stung and venom-treated
patients.
equal to or less than that required to elicit histamine release with the unfractionated venom, while showing little sensitivity to the other venom proteins. The 6 patients were selected because they were found to be sensitive to honeybee venom and not to venoms from other Hymenoptera classes as judged by laboratory and clinical criteria. Control patients and patients sensitive to other Hymenoptera species were insensitive to bee venom and to its protein components. IgE antibodies specific for phospholipase A were also found in the sera of honeybee-allergic patients. Reisman and associates5 have data in support of this observation but also report a small number of patients whose serum contained IgE antibodies to unfractionated venom although no detectable
38
Sobotka
et al.
J. ALLERGY CLIN. IMMUNOL. JANUARY 1976
antibodies to phospholipase A could be found. This discrepancy is difficult to explain. It is to be noted, however, that the RAST test is difficult to interpret when the antigen used is not clean; indeed this is one of the reasons this study was initiated. The presence of specific IgG antibodies against phospholipase A in the sera of multiply stung persons also substantiates the importance of this venom component as a major antigen, as do our results with a monospecific phospholipase A antisera. Although, as expected, less antiserum was required to inhibit phospholipase A histamine release, inhibition of venom-induced release required only a small increase in antiserum concentration. This is a preliminary study. A more definitive investigation requires that each protein be isolated by multiple chromatographic steps until greater than 99% homogeneity can be demonstrated. Such studies are under way. Their importance is attested to by the clear demonstration of multiple allergenic forms of phospholipase A in Pool II. It has previously been reported that phospholipase A exists in several molecular species differing in molecular weight, probably on the basis of carbohydrate differences. I389I7 Two peaks of antigenic activity were found in Pool II that differed in size but not in antigenie or allergenic activity. Most major allergens, such as ragweed antigen E’” and rye grass Group I,26 exist in nature as a series of isoallergens of slightly different composition but with a complete antigenie and allergenic cross-reactivity. It appears that phospholipase A shares this characteristic. We feel that the establishment of the major allergenic component in bee venom is of importance as a diagnostic tool which can be rapidly exploited. Phospholipase A is readily available commercially in large quantities and could be used diagnostically in the RAST test or in skin testing. The ability to predict sensitivity to honeybees in the majority of patients by means of the RAST technique will serve to alleviate many of the diagnostic problems facing physicians and perhaps to prevent unnecessary treatment with irrelevant antigens. The same comments relate to the utility of phospholipase A in assessing the titer of IgG antibodies in monitoring the response to treatment. Recent reports from our laboratory6 and from Busse and associates7 support the concept of a role for IgG anti-phospholipase A antibodies in protection against anaphylaxis. Both venom-treated patients reached IgG anti-phospholipase A antibody titers equal to those found in the sera of beekeepers and were subsequently stung without reaction. Finally, the establishment of phospholipase A as the major allergen of honeybee venom is important for standardization. It has been shown that the potency of whole pollen extracts, such as that from grass or ragweed pollens, can be more meaningfully evaluated when the standardization is based on the content of a known major allergen than when it is judged with protein nitrogen units or per cent solids as criteria. 27 The phospholipase A content can be used to standardize and quantitate honeybee venom and, if necessary, to assess whether the whole body extracts have any venom proteins.
VOLUME 57
Allergy
to insect stings.
II
39
NUMBER 1
NOTE ADDED IN PROOF Our recent studies have revealed that hyaluronidase (Peak I) is rapidly denatured after isolation. If this is prevented, hyaluronidase is antigenic in most patients (King et al.: Allergens of honeybee venom. In press, Arch. Biochem. Biophys.) However, all honeybeeallergic patients are highly sensitive to Phos-A as judged by several immunologic parameters.
REFERENCES Schwartz, H. J.: Skin sensitivity in insect allergy, J. A. M. A. 194: 703, 1965. from honey bees, Benton, A. W., Morris, R. A., and Stua.rt, J. D.: Venom collection Science 142: 228, 1963. Loveless, M. H.: Immunization in wasp sting allergy through venom repositories and periodic insect sting, J. Immunol. 89: 204, 1962. Sobotka, A., Valentine, M. D., Benton, A., and Lichtenstein, L.: Allergy to insect stings. I. Diagnosis of IgE mediated Hymenoptera sensitivity by venom induced histamine release,J. ALLERGY CLIN. IMMUNOL.~~: 170,1974. 5 Reisman, R. E., Wypych, J. I., Yeagle, N., and Arbesman, S. E.: Stinging insect hypersensitivity. III. Detection and clinical significance of IgE antibodies to insect venom in in man, J. ALLERGY CLIN. IMMUNOL. 53: 110,1974. 6 Lichtenstein, L. M., Sobotka, A. K., and Valentine, M. D.: A case for venom treatment in anaphylactic sensitivity to Hymenoptera sting, N. Engl. J. Med. 290: 1223, 1974. 7 Busse, W., Reed, S. E., Lichtenstein, L. M., and Reisman, R. E.: Protection following honey bee venom immunotherapy in a case of bee sting anaphylaxis, J. ALLERGY CLIN.
IMM~NOL. 53: 104,1974 8 Jaques, R., and Schacter, M.: The presence of histamine,5-hydroxytryptamine and a potent, slow contracting substance in wasp venom, Br. J. Pharmacol. 9: 53, 1954. 9 Benson, R. L., and Semenov, H.: Allergy in its relation to bee sting, J. ALLERGY 1:
105,193o. 10 O’Connor, R., Rosenbrook, Jr., W. venoms and body proteins, Science lla Shulman, S., Langlois, C., and insects. I. Preparation of extracts 446, 1964. llb Shulman, S., Langlois, C., Miller, stinging insects. V. Fractionation
and Erickson, R.: Dioc electrophoresis of Hymenoptera 145: 1320, 1964. Arbesman, C. E.: The allergic response to stinging and their biochemical characterization, J. ALLERGY 35: J. B., and Arbesman, C. E.: of whole body and venom
ALLERGY 37: 350, 1966. 11~ Langlois, C., Shulman, S., Kozmycz, stinging
insects.
VI.
Fractionation
The allergic sac extracts
response to of bee, J.
S., and Arbesman, C. E. : The allergic of whole body and venom sac extracts
response to of wasp, J.
ALLERGY 37: 359, 1966. lld
12 13a 13b 14 15
16 17 18
Arbesman, C. E., Langlois, C., Bronson, P., and Shulman, S. : The allergic response to stinging insects. VII. Fractionation of whole body and venom sac extracts of yellow jacket, J. ALLERGY 38: 1, 1966. Habermann E.: Bee and wasp venoms, Science 177: 314, 1972. Shipolini, R. A., Callewaert, R. C., Doonan, S., Vernon, C. A., and Banks, B. E. C.: Phospholipase A from bee venom, Eur. J. Biochem. 20: 459, 1971. Shipolini, R. A.: The primary sequence of phospholipase A from bee venom, Febs. Letters 17: 39, 1971. Liibke, K., Matthes, S., and Kloss, G.: Isolation and structure of N-l formyl mellitin, Experimentia 27: 765, 1971. Schrader, E., Liibker, K., Lehmann, M., and Beetz, I.: Haemolytic activity and action on the surface tension of aqueous solutions of synthetic mellitins and their derivatives, Experientia 27: 764, 1974. from honey bees, Benton, A. W., Morris, R. A., and Stuart, J. D.: Venom collection Science 142: 228, 1963. Franklin, R., and Baer, H.: Comparison of honey bee venoms and their components from various sources, J. ALLERGY CLIN. IMMITNOL. 55: 285,1975, Lichtenstein, L. M., and Osler, A. G.: Studies on the mechanism of hypersensitivity phenomena. IX. Histamine release from human leukocytes by ragweed pollen antigen, J. Exp. Med. 120: 507, 1964.
40
Sobotka et al.
J. ALLERGY CLIN. IMMUNOL. JANUARY 1976
study 19 May, C. D., Lyman, M., Alberto, R., and Cheng, J.: Procedures for immunochemical of histamine release from leukocytes with small volume of blood, J. ALLERGY 46: 12, 1970. 20 Levy, D. A., and Osler, A. G.: Studies on the mechanisms of hypersensitivity phenomena. XIV. Passive sensitization in vitro of human leukocytes to ragweed pollen antigen, J. Immunol. 97: 203, 1966. 21 Lichtenstein, L. M., and Osler, A. Cr.: Studies on the mechanism of hypersensitivity study of the reaction between ragweed pollen antigen, phenomena. XII. An ia vitro allergic human serum and ragweed sensitive cells, J. Immuol. 96: 169, 1966. 22 Newcomb, R. W., and Ishizaka, K.: Human diptheria antitoxin in immunoglobulin classes IgG and IgA. J. Immunol. 99: 40, 1967. 23 Wide, L., Bennich, H., and Johansson, S. G. 0.: Diiagnosis of allergy by in vitro test for allergen antibodies, Lancet 2: 1105, 1967. 24 Franklin, R., and Baer, H.: Immune and nonimmune precipitates produced by honey bee venom and its components. In press, Proc. Sot. Exp. Biol. Med. 147: 585-588, 1974. 25 King, T. P., Norman, P. S., and Connell, J. T.: Isolation and characterization of allergens from ragweed pollen. II, Biochemistry 3: 458, 1964. 26 Johnson, I’., and Marsh, D. G.: Isoallergens from rye grass pollen, Nature 206: 935, 1965. 27 Baer, H., Godfrey, H., Malony, C. J., Norman, P. S., and Lichtenstein, L. M.: The potency and antigen E content of commercially prepared ragweed extracts, J. ALI,EK(:P 45: 347, 1970.
