of Human Allergies

of Human Allergiesby Byron Goldstein and Micah Dembo

A study of the interaction between allergen molecules and a cellular surfacebristling with antibodies has revealed details about the on-off switches for

release of chemicals that cause allergic symptoms.

w hen spring arrives in NewMexico, invisible, diffuseclouds of juniper pollen driftthrough the air leaving be-

hind a wake of red eyes, runny noses, andsneezing, suffering people. How is it thatsuch a tiny, seemingly harmless substancecan wreak such havoc? Why are somepeople victims, others not? How might thisand other allergies be brought under control?Regrettably, allergies are the result of com-plex biochemical reactions that start in thebody’s immune system, and answers to ourquestions require detailed knowledge of thesereactions.

The recent explosive growth in im-munology has provided some of the knowl-edge by uncovering many of the steps inallergic reactions, Lawrence Liechtensteinand Anne Kagey-Sobotka at the JohnsHopkins University School of Medicine and

20

we at Los Alamos have been conducting ajoint study of a step important in mostallergies—the release of histamine, one ofthe chemicals that cause the allergy sufferer’sdiscomfort. By coupling experimentalmeasurements of histamine release (JohnsHopkins) with theoretical modeling of thepertinent reactions (Los Alamos), it waspossible to map this step in detail. The studyalso provided important information aboutdesensitization, that is, the turning off of theallergic response. It is hoped that this studywill bring us closer to control of this com-mon, but often debilitating and sometimesdeadly, phenomenon.

The Immune Response

The human immune system is an amazingdefense system capable of distinguishingwhat is foreign from what is self. This system

responds to and destroys invading sub-stances that can cause infection or disease,but in some people it also responds to manyseemingly harmless foreign substances.Whether it be a virus, a bacterium, or onlythat troublesome pollen, any foreign materialthat triggers an immune response is called anantigen.

The immune system reacts to antigenswith one or both of two responses, a cellularresponse and an antibody response. Thecellular response involves production ofspecial white blood cells (T lymphocytes)that are capable of binding to and destroying

the antigen. A classic example of thisresponse is the rejection of organ trans-plants. Material from an organ grafted to agenetically different recipient of the samespecies acts as an antigen and elicits acellular immune response that destroys thegraft.

Fall 1982/LOS ALAMOS SCIENCE

In the second response the immune systemreacts to an antigen by producing and releas-ing into the blood special proteins calledantibodies. A popular immunology experi-ment exemplifies this response. If a mouse isinjected with red blood cells from a sheep,the mouse’s immune system produces anti-bodies that can bind to the sheep’s red bloodcells. The bound antibodies act as tags thatclearly advertise the foreign nature of theantigen. Once tagged, the antigen becomessubject to attack and destruction by othermolecules and cells of the body.

An important characteristic of the reac-tion between an antigen and an antibody orT lymphocyte is its specificity: an antibodyor a T lymphocyte binds to the antigen thattriggered its formation but with rare excep-tions does not bind to other antigens. Forexample, the antibody that binds to influenzaA virus does not bind to influenza B virus.

LOS ALAMOS SCIENCE/Fall 1982

As shown in Fig. 1, this specificity occursbecause the shape and charge distribution ofa particular molecular structure on the anti-gen—the binding site—are matched by acomplementary shape and charge distribu-tion of a binding site on the antibody or Tlymphocyte. The two structures mesh some-what like a lock and key.

Both immune responses possess the prop-erty of memory. If a rejected graft is fol-lowed by another from the same donor, itwill be rejected more rapidly than the first. Ifa mouse is injected a second time with redblood cells from a sheep, it will producegreater amounts of antibodies more rapidly,and these antibodies will bind more stronglyto the sheep’s red blood cells.

With these marvelous properties the im-mune system seems designed to protect usfrom disease and infection. However. as wehave said, not all immune responses are

directed against harmful substances and notall immune responses improve our well be-ing, as those of us who have allergies wellknow.

Allergic Reactions, IgE, and Histamine

There are two types of hypersensitive, orallergic. reactions, each involving one of thetwo immune responses. Delayed hyper-sensitive reactions (so called because theyevolve slowly, peaking in about 1 to 4 days)involve the cellular, or T lymphocyte,

response. Examples are allergies to poisonivy and industrial chemicals. Immediate hy-persensitive reactions, on the other handinvolve the antibody response. In the re-mainder of the article, we will discuss onlyimmediate hypersensitive reactions, and

there certainly are enough of them. Hayfever, hives. and asthma. as well as allergies

21

to grasses, dog and cat danders, certainfoods, bee venom, and penicillin, are allexamples of antibody-mediated hyper-sensitivity.

Human antibodies are grouped into fiveclasses (immunoglobulin A, D, E, G, and M)according to their biological functions, andthe immune system can produce all theseclasses in response to normally harmlessforeign substances. But no allergic symp-toms will result unless immunoglobulin E(IgE) antibodies are produced. An antigenthat causes the immune system to produceIgE antibodies is called an allergen.

IgE antibodies were discovered in 1967 byKimishige Ishizaka and Teruko Ishizakawhen they were studying the blood serum ofhay-fever patients at the Children’s AsthmaResearch Institute and Hospital in Denver.This class of antibodies was the last to bediscovered, in part because it is normallyproduced in very small amounts. The con-centration of all antibodies in the serum of anonallergic, healthy person is about 15 milli-grams per milliliter, but the concentration ofIgE antibodies is only about 0.0001 milli-grams per milliliter. Allergic individuals,however. tend to have higher concentrationsof IgE antibodies. (People suffering fromcertain parasitic infections also have elevatedlevels of IgE antibodies, but why this is so is

still an open question.)

All antibodies, regardless of class, aremade up of similar Y-shaped units. Each unitcontains two identical heavy polypeptidechains whose molecular weight depends onthe antibody class and varies from 55,000 to75,000 daltons. These heavy chains arejoined by one or more disulfide bonds alongsome portion of their lengths to form thebase of the Y. called the Fc region. Theremaining lengths of the heavy chains arefree and form flexible arms of the Y. Eacharm, or Fab region, also contains a lightpolypeptide chain joined to the heavy chainby a disulfide bond. The light chains in thetwo arms are identical and have a molecularweight of about 23.000 daltons. As shown in

22

Pollen Grain

BoundAntibody

Antibody

Fig. 1. Antibody-antigen binding. The “lock-and-key” model depicted here explainsthe high specificity of antibody-antigen binding. The bond between the two is notcovalent, but is due to a combination of ionic and induced-dipole dispersion forces. Asa result, a strong bond occurs when both the ionic charges and the shapes arecomplementary, the latter permitting the two surfaces to approach closely and thusmaximize the dispersion force. The Y-shaped antibody molecule has two identicalbinding sites, either or both of which can participate in the binding.

Fig. 2, IgE antibodies are monomers; that is,they consist of only one of these Y-shapedunits. (IgD and IgG are also monomers. IgMis a pentamer, and IgA exists as a monomer,dimer, or trimer.)

It is the Fc region of an antibody thatinteracts with cells and molecules of thebody and thus determines the antibody’sbiological functions. A class of antibodiesincludes all those antibodies with identical Fcregions. The name Fc arises from the fact

that these identical fragments can becrystallized from a sample of, say, IgEantibodies. (Crystallization would not occurif these fragments were heterogeneous. )

Identical antigen binding sites are locatednear the ends of the Fab regions. (The nameFab stands for antigen binding fragment.)Here the antibody’s specificity is determined:the binding sites have the correct three-

dimensional structure to attach to the siteson the antigen that triggered the antibody’s


the On & Off of Human Allergies

Fig. 2. Important features of an IgE antibody molecule. The molecule consists of twoidentical heavy polypeptide chains and two identical light polypeptide chains linked bydisulfide bonds into an overall Y shape. The configuration depicted here is onlysymbolic of complicated three-dimensional chains that intertwine and fold back onthemselves. Each arm is flexible and the angle between the arms of the Y can vary,possibly from O to 180 degrees. The Fc region is identical in all IgE antibodies. Thisregion interacts with the cells and molecules of the body and thus determines thebiological functions of the IgE class of antibodies. IgE antibodies specific to differentantigens (or allergens) differ in their Fab regions. At the end of each arm are bindingsites that match the binding site on the antigen (or allergen) that caused formation ofthe antibody. The binding sites are here depicted by concave half circles.

Fig. 3. Photomicrograph of a humanbasophil (7,000x). During an allergicreaction the granules of these cells re-lease histamine and other potentchemicals into the blood stream. (Thisphoto, kindly supplied by Ann M.Dvorak of the Beth Israel Hospital inBoston, is copyrighted by the UnitedStates-Canadian Division of the Inter-national Academy of Pathology and isreproduced with their permission.)

production. Thus, antibodies of the sameclass but specific to different antigens areidentical in their Fc regions and differ in theirFab regions.

IgE antibodies interact with two types ofcells—basophils and mast cells—that play acentral role in allergic reactions. Mast cellsare found in tissue and are most numerous inthe linings of the respiratory and gastro-intestinal tracts and under the skin. Ba-sophils (Fig. 3) are found in blood serumand make up approximately 1 per cent of thewhite blood cells. The surfaces of both thesetypes of cells are dotted with binding sites forthe Fc portion of IgE antibodies. These

epsilon subscript indicates specificity for IgEantibodies). The number of receptors per cellvaries widely. For example, on a humanbasophil there may be anywhere from 5000to 500,000. Basophils of allergic people tend

than do those of nonallergic people.

Stored in granules within basophil andmast cells are biologically potent molecules,including histamine and the recently dis-covered leukotrienes. When these moleculesare released from the cells they have severalimportant effects. They cause contraction ofsmooth muscles such as those surroundingblood vessels and air passages in the lungs. Ithas been shown that histamine contracts thesmooth muscles of the larger air passages inthe lungs, and the leukotrienes contract thesmaller peripheral airways. These moleculesalso affect the permeability of blood vesselwalls and other membranes and cause glan-dular hypersecretion. At the site of a punc-ture wound, these effects result in blood flowchanges, inflammation, fluid secretion, andthe passage through various membranes ofmolecules and cells that attack and destroyharmful substances. In an allergic reaction abody-wide release of these same chemicalsresults in symptoms such as fluid secretion inthe nose and throat and obstruction of airpassages in the lungs.

There are several steps leading up to

LOS ALAMOS SCIENCE/Fall 1982 23

histamine release; these are depictedschematically in Fig, 4. The first step is theinitial response of a person’s immune systemto an allergen: the production of IgE anti-bodies specific to the allergen. Some of theseIgE antibodies then bind through their Fc

basophils; the two arms of the Y and theallergen-specific binding sites projectoutward from the cell surface. The cells arenow “sensitized” to the allergen. Anotherexposure to the same allergen results inbinding between the allergen molecules andthe IgE antibodies that are bound to thesurfaces of the cells. It is this interaction that

Fig. 4. The allergic reaction. The body’simmune system may react to an allergenby producing IgE antibodies specific tothat allergen. Many of these antibodiesbind to the surfaces of basophils andmast cells through receptors that arespecific to the Fc region of IgE anti-bodies; the arms of the antibodies con-taining the binding sites specific to theallergen point outward. A cell stippledwith anti-bodies specific to an allergen issaid to be sensitized to that allergen.Another exposure to the same allergencan lead to the formation of crosslinks,or bridges, between the IgE antibodiesbound to the cell surfaces. Thesecrosslinks form as sites on an allergenbind to complementary sites on adjacentantibodies. (Note that the allergen musthave more than one binding site to formcrosslinks. For simplicity the allergenshown here is divalent; that is, it has twobinding sites.) Wherever a crosslink isformed a channel opens for transport ofcalcium ions (Ca++) into the cell in-terior. This influx triggers a mechanismin which the granules release their con-tents of histamine and other chemicalsinto the blood serum.

Calcium Ion Influx

Release of Histamineand OtherChemicals

●✎

24 Fall 1982/LOS ALAMOS SCIENCE


AllergenMolecular

ChainI

Rabbit AntibodySpecific to Human lgE

(c)

Fig. 5. A necessary condition for the release of histamine from

be brought and held in proximity. This condition is achieved asany of the following processes takes place: (a) allergens formcrosslinks between the arms of IgE antibodies bound to the

Antibody Specificto Receptors

(d)

receptors; (b) a molecular chain covalently binds two IgEantibodies bound to the receptors;(c) a rabbit antibody specificto human IgE antibodies binds to the Fc region of IgEantibodies bound to the receptors; and(d) a specially prepared

can trigger the release of histamine from thegranules of the cells. Here, then, is a criticalstep in an immediate hypersensitive allergicreaction.

In the remainder of the article we willdiscuss the histamine release mechanism inmore detail, in particular what turns it onand what turns it off. We will concentrate onbasophils, but almost everything we sayabout those cells holds true for mast cells aswell.

The Basophil’s “On” Signal

One condition necessary for histaminerelease from basophils is that the allergensmust bind to the IgE antibodies in such a


way that crosslinks, or bridges, are formedbetween the antibodies (Fig. 5a). Evidencethat this crosslinking is necessary lies in thefact that allergens with only one binding site,which are physically incapable of crosslink-ing two IgE antibodies, do not trigger his-tamine release. Chains, rings, or other con-figurations of many crosslinked IgE anti-bodies may form on the basophil surfaces,but such large aggregates of crosslinkedantibodies are not necessary for histaminerelease. Rather, the formation of crosslinkedantibody pairs is sufficient. David Segal, JoelTaurog, and Henry Metzger of the NationalInstitutes of Health demonstrated this suffi-ciency in 1978 by exposing basophils to

molecules consisting of two covalently linked

IgE antibodies (Fig. 5b). These permanentlylinked IgE antibody pairs caused histaminerelease in the absence of allergen.

Allergens crosslink IgE antibodies bybinding to their Fab regions, but othermolecules that crosslink IgE antibodies bybinding along the Fc regions in the base ofthe Y can also cause histamine release. Thisphenomenon was originally demonstratedwith molecules prepared by injecting rabbitswith human IgE antibodies; the rabbits

produced antibodies that bound specificallyto the Fc regions of human IgE antibodies.(Other animal antibodies specific to humanIgE antibodies are prepared in a similar

25

fashion.) When basophils sensitized withhuman IgE antibodies were exposed to theserabbit antibodies, histamine release occurredeven though the antibodies were crosslinkedthrough their Fc regions (Fig. 5c).

It is now clear that the requirement forhistamine release of crosslinked IgE anti-

tors, which are mobile on basophil and mastcell surfaces, be brought and held in proxim-ity. This was verified in 1978 by theIshizakas and their collaborators at JohnsHopkins University, as well as a group at theNational Institutes of Health, when theymade an antibody that bound to Fcc recep-tors on mast cells and basophils of rats. Thisantibody triggered histamine release from the

close together (Fig. 5d).Histamine release from basophils also de-

mands another condition. If no calcium ionsare present in the medium surrounding thebasophils, no histamine will be released. Theproximity of Fcc receptors brought about bycrosslinked IgE antibodies somehow allowscalcium ions to cross the cell membrane, andthis influx of calcium ions is an essentialsignal for histamine release, If calcium ionscan be introduced into the cell in some otherway, crosslinked IgE antibodies are notneeded. For example, calcium ionophores,substances that cause calcium ion channelsto form in cell membranes, will induce his-tamine release in the absence of crosslinkingif calcium ions are available. Injectingcalcium ions directly into basophils alsoinduces histamine release. In a test tube thecalcium ion concentration can be manipu-lated, but in serum, the natural milieu of thebasophil, calcium ions are always present ata concentration (2 to 5 millimolar) that issufficient to insure histamine release.

In summary, an immediate hypersensitivereaction is turned on by the flow of calciumions into sensitized basophils made possibleby allergen-linked IgE antibodies on thebasophil surfaces. What turns the reactionoff?

26

1 0 0

Time before Addition of Calcium Ions (rein)

Fig. 6. Basophil desensitization. Basophils sensitized to the major ragweed allergenrelease a large percentage of their histamine when exposed to the allergen in thepresence of calcium ions. (The basophils studied in the experiment depicted herereleased almost 90 per cent of their histamine under these circumstances.) However, ifthe basophils are exposed to the allergen in the absence of calcium ions, the cellsgradually lose their ability to release histamine. The cells are said to be desensitizedwhen very little histamine is released in the presence of calcium ions.

A Nonspecific “Off" Signal

As early as 1964 Lawrence Liechtensteinand Abraham Osler at Johns Hopkins Uni-versity showed that the same agent thatcaused histamine release from basophils ofhay-fever patients, ragweed pollen, couldalso desensitize these cells, that is, block theirrelease of histamine, In 1971 Liechtensteininvestigated desensitization further with ex-periments on white blood cells, includingbasophils, of hay-fever patients. The cellswere exposed for various lengths of time to

the major ragweed allergen (a 38,000-daltonprotein isolated by T. P. King and hiscollaborators at Rockefeller University) in amedium containing no calcium ions. Cal-cium ions were then added, and the amountof histamine released by the cells during thefollowing 30 minutes was measured. Theseexperiments showed that the longer the cellswere exposed to the ragweed allergen in theabsence of calcium, the less histamine theywere capable of releasing in the presence of

calcium (Fig. 6). Liechtenstein obtainedsimilar results using as the crosslinking agent



Ragweed Goat RyegrassAllergen Antibody Allergen

Second Exposure Stimulus

Fig. 7. Nonspecific desensitization. The basophils studied in the experiment illustratedhere were from a donor who was allergic to both the major ragweed allergen and aryegrass allergen. On the left are the percentages of histamine released by thesedoubly sensitized basophils when exposed to either of the two allergens or to a goatantibody specific to human IgE antibodies. On the right are the percentages ofhistamine released by these cells when exposed to the ragweed or ryegrass allergens orthe goat antibody after being desensitized to the ragweed allergen. The negligiblerelease shows the cells to be nonspecifically desensitized to any allergen orcrosslinking agent.


a rabbit antibody specific to human IgEantibodies. It appears that the crosslinking ofIgE antibodies on the basophil surface in-itiates two signals: a signal for histaminerelease in the presence of calcium ions and asignal for blockage of histamine release inthe absence of calcium ions.

Liechtenstein performed similar experi-ments on white blood cells from a donor whowas allergic to both the ragweed allergen anda ryegrass allergen. Therefore, some of theFCC receptors on the donor’s basophils wereoccupied by IgE antibodies specific to theragweed allergen and some by IgE anti-bodies specific to the ryegrass allergen. Asbefore, after sufficient exposure to theragweed allergen in the absence of calciumions, addition of calcium ions caused nosignificant histamine release. The cells, nowdesensitized to the ragweed allergen, wereexposed in the presence of calcium ionseither to the ryegrass allergen or to a goatantibody specific to human IgE antibodies.Neither substance caused significant release

(Fig. 7). Desensitization to one allergen

turned the basophils off to other allergensand crosslinking agents as well. This “non-specific” desensitization occurred despite thefact that the desensitizing allergen interactedwith only a fraction of the IgE antibodies onthe cell surfaces, namely, those specific tothe ragweed allergen.

Although nonspecifically desensitizedbasophils cannot be made to release his-tamine by crosslinked IgE antibodies, theydo release histamine when exposed tocalcium ionophores or injected with calciumions. Nonspecific desensitization must there-fore involve the shutdown of a calcium iontransport mechanism. Apparently, crosslink-ing of IgE antibodies in the presence ofcalcium ions at first activates the transportmechanism, but with time that activationsomehow degrades.

Blocking histamine release by withholdingcalcium ions represents an artificial situation,Does nonspecific desensitization occur in thenatural milieu of the basophil, that is, in the

27

presence of calcium ions? Evidence for apositive answer comes from numerous in

vitro studies of histamine release frombasophils as a function of allergen concen-tration. Data from such studies are generallydisplayed as plots, known as dose responsecurves, of the percentage of histamine re-leased versus the logarithm of the allergenconcentration. Figure 8 shows dose responsecurves for basophils from three allergicdonors. The initial rise of the dose responsecurves reflects the increase in the number ofIgE antibodies that are crosslinked by theallergen. The more interesting feature of thecurves is their eventual fall, a phenomenonknown as allergen excess inhibition.

Karl Becker, Henry Metzger, and PhilipGrimley of the National Institutes of Health.

in collaboration with the Ishizakas, showedthat for basophils from allergic donors,which generally have large numbers of speci-

fic IgE antibodies on their surfaces, excessinhibition is accompanied by large numbersof crosslinks. They did so by studying thedistribution of IgE antibodies crosslinked bya fluorescent form of a sheep antibodyspecific to human IgE antibodies (Fig. 9). Atlow concentrations of the sheep antibody thedistribution of fluorescence was diffuse, but,at the concentrations at which excess inhibi-tion occurs, the distribution of fluorescencebecame patchy. These observations suggestthat at high concentrations of the crosslink-ing agent large aggregates of crosslinked IgEantibodies had formed.

Excess inhibition can be understood interms of the effect of crosslinking on thecalcium ion transport mechanism. Lownumbers of crosslinks (low allergen concen-trations) activate the transport mechanismand histamine release occurs. Degradation ofthe transport mechanism takes place slowly,and histamine release can be blocked only bywithholding calcium ions until this gradualshutdown has been completed. As the num-ber of crosslinks increases (higher allergenconcentrations), the transport mechanismdegrades more rapidly. Eventually, degrada-

tion dominates and histamine release is

28

100

50

01 0- 5 1 0- 3 10-1

Fig. 8. Typical dose response curves. These curves show the percentages of histaminereleased by basophils from three ragweed-pollen-allergic donors when exposed tovarious concentrations of ragweed pollen extract. In all cases the percentage ofhistamine released first rises with increasing concentration but then peaks anddeclines. This decline at high allergen concentrations is called allergen excessinhibition. The ragweed pollen extract used in the experiment contained five ragweedallergens, From Lawrence M. Liechtenstein and Abraham G. Osler, Journal ofExperimental Medicine 120,507 (1%4).

100

50

00.01 0.1 1.0 10.0

Fig. 9. A t low concentrations of a sheep antibody specific to human IgE antibodies, thedose response curve (black) is rising and, as shown by fluorescence microscopy (red),the distribution of IgE antibodies on the basophil surface is mainly diffuse. Atconcentrations above that for maximum release of histamine, the distribution of IgEantibodies becomes mainly patched. The patches of fluorescence indicate theformation of large aggregates of crosslinked antibodies. Based on data from Karl E.Becker, T. Ishizaka, H. Metzger, K. Ishizaka, and Philip M. Grimley, Journal ofExperimental Medicine 138,394 (1973).



Basophil from Incubated in Blood Becomes PassivelyNonallergic Donor Serum from Sensitized

Allergic Donor

allergen. This passive sensitization occurs because IgE anti-receptors can be sensitized to an allergen by incubating thecells in blood serum from a donor who is allergic to the receptors.

blocked even in the presence of calciumions.

Only on basophils with large numbers ofthe allergen-specific IgE antibodies bound totheir surfaces can there be sufficientcrosslinks for nonspecific desensitization tooccur in the presence of calcium ions. There-fore, generally only basophils from allergicdonors exhibit allergen excess inhibition be-cause a substantial fraction of their large

allergen-specific IgE antibodies.What happens, however. when there are

only small numbers of the allergen-specificIgE antibodies on the basophil surfaces? Theanswer to this question was provided by ourjoint research with Liechtenstein and KageySobotka. What happens proved to be notonly something different but also quite inter-esting. In particular, we observed anothermechanism for turning off the allergic reac-tion.

The Experiments

The goal of our initial collaboration withLiechtenstein and Kagey-Sobotka in 1977was to test our theoretical predictions, inparticular, predictions we had made about

the dose response curves as the number of


allergen-specific IgE antibodies on the

basophil surfaces is increased. To eliminateextraneous variables that would blur a cor-relation between theory and experiment, weneeded to use the same allergen, the sameallergen-specific IgE antibody, and basophilsfrom the same donor. Moreover. we needed

be filled with different amounts of the IgEantibody of our choice. This requirementforced us to use basophils from nonallergicdonors, because basophils from allergic

receptors tilled. However, the total numberof receptors is much smaller on basophilsfrom nonallergic donors. As a result, ourstudy was of basophils with relatively smallnumbers of the allergen-specific IgE anti-bodies on their surfaces, usually less than10,000 per cell. Serendipitously, this circum-stance led to our discovery of a secondmechanism for desensitization.

Other experimental restraints resultedfrom the fact that no one had, or has yet,learned how to keep human basophils aliveoutside the body for longer than a day or so.The research thus depended on the avail-ability of donors, both of basophils andblood serum. Another difficulty was the day-to-day variation in the basophil donor’sexposure and immune system response that

on the basophil surfaces.To obtain basophils with different and

well-characterized numbers of the allergen-specific IgE antibody on their surfaces, we

serum from a donor who was extremelyallergic to penicillin. (Ninety per cent of thepenicillin-allergic donor’s IgE antibodies

were specific to the benzylpenicilloyl, orBPO, group.) The length of the incubation orthe dilution of the penicillin-allergic donor’sserum determined how many BPO-specific

basophil surfaces.This technique for sensitizing cells is

called passive sensitization (Fig. 10) and hasbeen known since 1921 when Carl Prausnitzinjected into the skin of a nonallergic subject

a small quantity of serum from HeinzKustner, who was extremely allergic to fish.Twenty-four hours later fish extract wasinjected into the same area of the nonallergicsubject’s skin. Immediately a wheal ap-

peared. We now know that Kustner’s serumcontained IgE antibodies specific to an al-lergen found in fish. When transferred to theskin of the nonallergic subject, these anti-bodies passively sensitized his mast cells.Then, when exposed to the allergen in thefish extract, the sensitized mast cells released

29

Fig. 11. Synthetic benzylpenicilloyl (BPO) allergens. The arms of the IgE antibodies. The monovalent molecules, withbivalent molecules, with a BPO group (circle) at each end of only one BPO group, cannot form such crosslinks.the chain, are able to form bridges between binding sites on the

histamine and other chemicals that producedthe wheal. The PK test for allergy is namedfor these early investigators.

Our choice of using IgE antibodies from apenicillin-allergic person was based on theavailability of simple, well-defined syntheticpenicillin allergens. Bernard Levine at theNew York University School of Medicinehad first prepared a series of such allergensin 1967. Some of these are illustrated in Fig.11. Each is a linear chain of different lengthwith either a BPO group at each end(bivalent) or a BPO group at one end only(monovalent).

We resurrected the bivalent and mono-valent synthetic penicillin allergens becausethey were ideal for our purpose: to start tobuild a mathematical model of histaminerelease from basophils. In (BPO)2 we had the

30

simplest possible crosslinking agent, a sym-metric linear molecule with two identicalbinding sites. In (BPO) l we had a tool fortesting our ideas about what happens whenthe number of crosslinks is reduced.

The final step of the experiments entaileddetermining the dose response curves for thepassively sensitized basophils, that is,measuring the percentages of histamine re-leased by the basophils when exposed to

various concentrations of (BPO) 2 a n d(BPO) l allergens. We hoped that by compar-ing the dose response curves with our theo-retical calculations about crosslinks we couldlearn something new about the role of

crosslinks in histamine release.We shall first present the results of our

calculations and then a comparison of theseresults with those of the experiments.

Results of Crosslinking Calculations

We learned experimentally that the bind-ing between the IgE antibodies used topassively sensitize the basophils and the(BPO) 2 or (BPO)1 allergens came to equi-librium within seconds after the basophilswere exposed to the allergens and well beforeany measurable histamine release or de-sensitization had occurred. Thus, histaminerelease and desensitization were governed bythe equilibrium concentration of crosslinks.For these experiments we could neglect thedetails of the binding during the first fewseconds of exposure and instead treat thebasophils as if the crosslinks formed instan-taneously. Thus, we calculated, for a givennumber of BPO-specific IgE antibodies perbasophil, the equilibrium concentration of



Fig. 12. According to our calculations, the fraction of IgE antibodies crosslinked by(BPO)2 varies as shown with the concentration of (BPO)2 to which the basophils areexposed. The curve has a single maximum and is symmetric about that maximum. Theinitial rise in the curve is expected; the decline is due to a saturation effect depicted inFig. 13.

Fig. 13. At high (BPO)2 concentrations most of the binding sites on the antibodies arefilled with (BPO)2 molecules. Because there are few empty sites available, thesebivalent (BPO)2 molecules are unable to form bridges between the antibodies.

crosslinks on the basophil surfaces for vari-ous concentrations of (BPO)2 and (BPO) l

allergens.Our method of attacking this equilibrium

problem is described in the sidebar“Crosslinking-A Theoretical Approach.”The results of the calculations revealed inter-esting predictions about crosslinking. Forexample, Fig. 12 shows a plot, called acrosslinking curve, of the fraction ofcrosslinked antibodies versus the logarithmof the (BPO)2 concentration. The curve risesfrom zero to a maximum and then decreasesto zero. The curve falls because, as the(BPO) 2 concentration increases, more andmore of the binding sites on the antibodiesare occupied by (BPO)2 allergens with oneunbound BPO group (Fig. 13). With fewerpotential sites available for that unboundBPO group the number of crosslinks de-creases.

Another property of the crosslinkingcurves is symmetry about their maxima.That is, if the maximum number ofcrosslinks occurs at a (BPO)2 concentrationof 10–7 molar (as it does in Fig. 12), therewill be just as many IgE antibodiescrosslinked at 10–6 molar as at 10–8 molar.

The (BPO)2 concentration at which themaximum of the crosslinking curve occurs isdetermined by only two parameters: K, theequilibrium constant for the binding betweena BPO group and a binding site on anantibody, and the concentration of (BPO)l

allergen. If [(BPO)2] max is the (BPO)2 concen-tration at which crosslinks are a maximumand [(BPO)1] is the (BPO)l concentration to

which the basophils are exposed, then

One of the most interesting predictions ofour calculations is that [(BPO)2] max does notdepend on the total number of BPO-specificantibodies on the basophil surfaces. In otherwords, if the number of BPO-specific anti-


Sidebar

crosslinking— a theoretical

T he problem at hand was to calculatethe equilibrium concentration ofcrosslinks formed when BPO-specific

IgE antibodies on basophil surfaces areexposed to the monovalent and bivalentsynthetic penicillin allergens (BPO)l and(BPO)2. We began by constructing a modelconsisting of all the binding reactions thatcan occur in this situation. Crucial to thecalculations is knowledge of the equilibriumconstants for these binding reactions. Al-though the model includes an infinite numberof reactions, some reasonable assumptionsreduce to a manageable number the equi-librium constants that must be known.

First, we let K be the equilibrium constantfor the binding between a monovalent al-lergen and a “monovalent antibody.” (Amonovalent antibody does not, of course,exist but is useful as a theoretical constructbecause the equilibrium constant for its reac-tion with a monovalent allergen is indicativeof the basic strength of the forces betweenthe binding sites.) Consider now the simplestof the reactions depicted in the accompany-ing figure, those initial reactions in which a(BPO)l or (BPO)2 allergen binds to one ofthe two sites on a free antibody (that is, anantibody bound through its Fc region to thecell surface but with each of the binding sitesin its Fab regions free). The equilibriumconstant for the reaction when (BPO)l isinvolved is 2K since the antibody offers twopossible binding sites. Similarly, the equi-librium constant for the reaction involving(BPO)2 is simply 4K because in this case theallergen also offers two possible bindingsites. We assume that the equilibrium con-

32

stants for these two reactions are unaffectedif the free antibody is replaced by a chain ofcrosslinked antibodies with a free bindingsite on the antibody at each end of the chain.

Considering next the binding of a (BPO)l

or (BPO)Z allergen to the single available siteon the products of the initial reactions, weassume that the equilibrium constants forthese reactions are also related to K byappropriate statistical factors.

Next, we let Kx be the equilibrium con-stant for the basic crosslinking reaction, thebinding of the complex containing one anti-body and one (BPO)2 allergen, each with afree binding site, to a free antibody. Againwe assume that the equilibrium constant isunaffected if the free antibody is replaced bya chain of crosslinked antibodies with a freebinding site on the antibody at each end.

Finally, a ring containing i antibodies isformed when a free BPO group of a (BPO)2

allergen bound to one end of a chain of icrosslinked antibodies binds to the free siteon the antibody at the other end of the chain.

constant Ji for such a reaction is inverselyproportional to i2. Therefore, J1 = 4J2/i2 for

for formation of the “ring” consisting of asingle (BPO)2 allergen spanning the sites on asingle antibody.

Armed with the four equilibrium constantsK, Kx, Jl, and J 2, we can calculate theequilibrium concentrations of all possiblereaction products, (Reasonable estimates forthe magnitudes of these constants can beobtained from various experimental data,)We will not present details of the calculations

but rather the general concepts on whichthey are based.

The accompanying figure shows thatseven complexes contain one antibody. Theequilibrium concentrations of each of thesecomplexes can be expressed as a function ofK, J1, and the (BPO)l and (BPO)Z concentra-tions multiplied by the equilibrium concen-tration of free antibody. Therefore, W1, thetotal equilibrium concentration of complexescontaining one antibody, is obtained simplyby adding together the equilibrium concen-tration of each of the complexes, We findthat

where

In these equations [(BPO)1] and [(BPO)2] arethe concentrations of (BPO)l and (BPO)2,respectively, and X, the only unknown, is theequilibrium concentration of free antibody.

The accompanying figure also shows thatseven complexes contain two antibodies. Wecan express the concentration of each ofthese complexes as a function of K, J2,[(BPO) 1], and [(BPO)2] multiplied by theconcentration X2 of the crosslinked chaincontaining two antibodies with a free binding



Sidebar

Shown here are all the binding reactions that can occur when BPO-specific IgEantibodies on basophils are exposed to the monovalent and bivalent synthetic penicillinallergens (BPO)l and (BPO)2. The equilibrium constants for each reaction are alsogiven.

site on the antibody at each end. But X2 is in We can continue this process iteratively andturn a function of K, [(BPO) 2], and X, develop a general expression for Wn, thenamely equilibrium concentration of complexes con-

taining n antibodies, as a function of[(BPO)1], [(BPO)2], K, Kx, Jn, and X:


The conservation law for total antibodyconcentration XT leads to the equation

When we express all the Wn in this infiniteseries in terms of X, the infinite series can besummed, and we obtain an algebraic equa-tion that can be solved for X. By substitutingthe solution for X into the expression for aparticular Wn or for the equilibrium concen-tration of a particular complex, we cancompute values for these expressions.

Because of the central role of crosslinks inthe activation and desensitization ofbasophils, we are particularly interested inX

poly, the fraction of antibodies incorporatedat equilibrium into complexes containingmore than one antibody, that is, in thefraction of crosslinked antibodies. An ex-pression for Xpoly is easily derived since

Since this calculation was first presented,much theoretical work has been done onboth the equilibrium and the kinetic theory ofbinding of multivalent antigens to antibodieson cell surfaces. As a result primarily of thework of Alan Perelson (Los Alamos),Charles DeLisi (National Institutes ofHealth), and Catherine Macken (LincolnCollege, New Zealand), much progress hasbeen made toward understanding the bond-

33

bodies is increased (by, for example, longerincubation), the (BPO)2 concentration atwhich the maximum occurs does not change(Fig. 14a). On the other hand, increasing theconcentration of (BPO)1, which reduces thenumber of potential crosslinks, causes themaximum to shift to higher (BPO)2 concen-trations (Fig. 14 b).

Histamine Release vs Crosslinking

How do the experimental dose responsecurves and the theoretical crosslinkingcurves compare? Figure 15 shows that thedose response curves exhibit the sameproperties as the crosslinking curves: the(BPO)2 concentration at which the maximumoccurs does not change with the number ofBPO-specific antibodies on the basophil sur-faces (Fig, 15a) but does change with the(B P O) l concentration (Fig. 15 b). There is

other agreement as well. For example, theproperties of the basophil do not appear inEq. 1; only the binding between the BPOgroup and the antibody is important. There-fore, we predict that basophils from differentdonors should exhibit maxima in theircrosslinking curves at the same (BPO)2 con-centration. This prediction is in agreementwith the observation that each of three doseresponse curves for basophils subjected tothe same experimental treatment but fromdifferent nonallergic donors had its max-

Fig. 14. Behavior of theoret icalcrosslinking curves. (a) As the amount ofBPO-specific IgE antibodies on thebasophil surfaces is increased, thecrosslinking curves exhibit greater max-ima, but the allergen concentration atwhich the maxima occur remains con-stant. (b) As the concentration of mono-valent (BPO)l to which the BPO-specificIgE antibodies are exposed increases,the number of crosslinks decreases andthe position of the maximum shifts tohigher (BPO)2 concentrations.



imum at the same (BPO)2 concentration.From these and other similar experiments

on basophils from nonallergic donors, we

100 concluded that the percentage of histaminereleased rises and falls directly with the riseand fall of the fraction of crosslinked anti-bodies. This behavior is quite unlike that ofbasophils from allergic donors. Recall that

Large Amount of IgE for these cells the fall in the dose response

on Cell curve occurs because the number ofcrosslinks has increased to the point wherenonspecific desensitization dominates overhistamine release. In other words, too manycrosslinks exist rather than too few. Thedifference in behavior is surely related to thefact that basophils from nonallergic donorshave relatively small numbers of specific IgEantibodies bound to their surfaces, whereas

1 0-11 1 0- 9 1 0- 7

1 0- 5cells from allergic donors have relativelylarge numbers.

(BPO)2 Concentration (M) Basophils from allergic and nonallergicdonors should manifest other differences as

(a) well. In particular, can basophils fromnonallergic donors be desensitized bywithholding calcium ions?

100 NO (BPO)1 Binding Another “Off’ Signal:

to lgE Specific Desensitization

We found that passively sensitized

Fig. 15. Behavior of experimental doseresponse curves. Like the crosslinkingcurve shown in Fig. 12, the doseresponse curves for basophils fromnonallergic donors exhibit maxima andare symmetric about those maxima.Further, the dependence of the dose

Considerable response curves on the amount of BPO -specific antibodies on the basophil sur-faces (a) and on the (BPO) concentra-

1 0- 1 1 1 0- 9 1 0- 7

1 0- 5

tion (b) is similar to that of the crosslink-

(BPO) 2 Concentration {M)ing curves (Fig. 14). The high concentra-tion wings of the dose response curves

(b) are missing because of the difficulty inkeeping high concentrations of (BPO)2

in solution.


basophils from nonallergic donors do, infact, desensitize. But they do so in a way thathas not before been observed—they de-

sensitize specifically. When exposed to(BPO) 2 in the absence of calcium ions, theybehave in one respect just like basophilsfrom allergic donors: they progressively losetheir ability to respond to (BPO)2 (by releas-ing histamine) when calcium ions are added.However, when exposed in the presence ofcalcium ions to a rabbit antibody specific tohuman IgE antibodies, they release his-tamine normally. Desensitizing with one al-lergen or crosslinking agent affected only thehistamine release that is triggered by thatallergen or crosslinking agent.

Our co l leagues a t Johns Hopkinsperformed an experiment showing thatbasophils from some allergic donors alsodesensitize specifically. After screening anumber of individuals who were allergic tosubstances other than penicillin, they foundone whose basophils had some fraction of

They passively sensitized these basophilswith BPO-specific IgE antibodies and thenexposed them in the absence of calcium ionsto (BPO)2 at the concentration for maximumcrosslinking of the antibodies. The cells werethen exposed in the presence of calcium ionst. either (BPO)2, ryegrass allergen (the indi-

vidual’s natural allergen), or a goat antibodyspecific to IgE antibodies. The results (Fig,16) show that the cells desensitized only to(BPO)2. In this case, even though the numberof crosslinks between BPO-specific IgE anti-bodies was at a maximum, the total numberwas too small for nonspecific desensitizationto play a significant role, and another type ofdesensitization came into play.

In a subsequent study we found thatallergic individuals whose basophils de-sensitize specifically have much lower con-centrations of IgE antibodies in their serumthan do allergic individuals whose basophilsdesensitize nonspecifically. This fact isfurther support for the suggestion that speci-fic desensitization is a phenomenon as-

36

100

Ryegrass

Fig. 16. Specific desensitization. The basophils studied in the experiment illustrated

receptors on the donor’s basophils were filled with IgE antibodies specific to the

basophils could be passively sensitized to the BPO group, On the left are thepercentages of histamine released by the passively sensitized basophils when exposedto either the synthetic (BPO)2 allergen, a goat antibody specific to human IgEantibodies, or the ryegrass allergen. On the right are the percentages of histaminereleased by these cells when exposed to the allergens or the goat antibody after beingdesensitized to (BPO)2. Only the histamine release triggered by the desensitizing agentis significantly affected.

sociated with small numbers of specific IgEantibodies on the basophil surfaces.

Because (BPO)l, which can bind but notcrosslink IgE antibodies, does not desensitizebasophils, we know that specific desensitiza-tion is also triggered by crosslinks. Wetherefore predict that basophils fromnonallergic donors will undergo the greatest

specific desensitization at the allergen con-centration producing the maximum numberof crosslinks, that is, the concentration atwhich histamine release is a maximum. Totest this prediction we performed the follow-ing experiment. Again, basophils fromnonallergic donors were passively sensitizedto (BPO)2. The dose response curve for these



Fig. 17. The basophils studied in the experiment illustrated here were from anonallergic donor and had been passively sensitized to the BPO group. The doseresponse curve for these cells exhibited a maximum at a (BPO)2 concentration of 10molar. The cells were then specifically desensitized to (BPO)2 by exposing them in theabsence of calcium ions to various (BPO)2 concentrations for either 15 or 30 minutes.Finally, the percentages of histamine released by the cells when exposed in thepresence of calcium ions to a (BPO)2 concentration of 10-7 molar was measured. Thedifference between the maximum histamine released (dashed line) and the measuredhistamine release (lower curves) is a measure of the amount of specific desensitizationundergone by the cells at that (BPO) concentration. Note that the desensitization isgreatest for a desensitizing (BPO)2 concentration of 10 -7 molar, the (BPO)2

concentration for maximum histamine release.


cells exhibited a maximum at a ( B P O )2

concentration of 10-7 molar. Groups of thesensitized cells were then desensitized byexposing them for a fixed time to (BPO)2 inthe absence of calcium ions. The (BP0)2

concentration was varied from group togroup. Finally, the desensitized cells wereexposed in the presence of calcium ions to(BPO) 2 at the concentration for maximumhistamine release, or 10 -7 molar. If nodesensitization had taken place, all thegroups of cells would have released themaximum amount of histamine. But de-sensitization did, in fact, occur and wasevidenced by the cells’ release of less than themaximum amount of histamine. Further, thegroup of cells desensitized at a (BPO)2

concentration of 10–’ molar released theleast amount of histamine and thus under-went the greatest amount of desensitization(Fig. 17).

Recently, in collaboration with HenryMetzger of the National Institutes of Health,we used covalently linked pairs of IgE anti-

bodies to show that large aggregates ofcrosslinked IgE antibodies are not requiredto induce specific desensitization. Just as forhistamine release, the formation of linkedpairs of IgE antibodies is the “unit” signalfor specific desensitization.

The two modes of desensitization arecompared in Table I. Are they related or arethey independent mechanisms?

37

Fig. 18. Desensitization model. In this model an open calcium the channel results in loss of the gating factor (Gi), which canchannel (G*) is formed when an allergen crosslinks two IgE account for nonspecific desensitization, and alterations in theantibodies and a gating factor (G) combines with apart@) of

desensitization.

The Transition from SpecificNonspecific Desensitization

to

We have evidence that basophils withrelatively small numbers of specific IgEantibodies on their surfaces desensitizespecifically and that basophils with relativelylarge numbers of specific IgE antibodies ontheir surfaces desensitize nonspecifically.Perhaps if the number of specific IgE anti-bodies on the basophil surfaces could bevaried over a large enough range, the type ofdesensitization exhibited by the basophilswould undergo a smooth transition fromspecific through partially nonspecific to com-pletely nonspecific.

In 1981 Donald MacGlashan and Law-rence Liechtenstein performed an experimentto test this suggestion. A major difficultythey faced was finding a donor whose

eventually one was located whose basophilshad close to 20,000 per cell. To take fulladvantage of the free sites available, theyvery carefully purified a BPO-specific anti-body preparation before passively sensitizingthe basophils. They estimated that in theirexperiments the number of BPO-specific IgEantibodies per basophil was varied from

38

approximately 800 to 14,000. Over thisrange they observed a smooth transition inthe type of desensitization from specific tononspecific, although the nonspecific de-sensitization was not complete. Our expecta-tion is that if more BPO-specific IgE anti-bodies could be placed on the basophils,complete nonspecific desensitization wouldbe achieved.

A Speculative Model for Desensitization

Crosslinking of IgE antibodies on thesurface of a basophil causes a flow ofcalcium ions into the cell that triggers therelease of histamine-containing granulesfrom the cell. But crosslinking also leads todesensitization of a type determined by thenumber of IgE antibodies on the basophilsurface.

How does this all come about? We’re notcertain, but we have some ideas that we haveformalized in a mathematical model. Themodel can explain what has been observedso far and can also make predictions thatcan be tested experimentally. The basic fea-tures of the model are sketched in Fig. 18.

brought into proximity by crosslinked IgEantibodies, they combine with a “calcium

gating factor” in the cell membrane. Thisreaction forms a channel through whichcalcium ions flow into the basophil. We haveshown experimentally that channels formedby crosslinked IgE antibodies are short-lived.We have incorporated this observation intoour model by assuming that a channelrapidly decays to an inactive form. We alsoassume that only a limited amount of gatingfactor is available.

This model explains nonspecific de-sensitization as follows. As basophils withlarge numbers of IgE antibodies specific to aparticular allergen are exposed to the al-lergen in the absence of calcium ions, somany calcium channels are formed that thesupply of gating factor is exhausted. Whenlater exposed to any allergen in the presenceof calcium ions, no further calcium channelscan be formed and those formed duringdesensitization have decayed. Hence, no his-

tamine is released.Calcium ions act as stimulator signals

not only for basophils and mast cells but alsofor a variety of other cells, some of whichundergo processes similar to desensitization.For example, cells in the blowfly’s salivarygland can be stimulated to secrete by aninflux of calcium ions triggered by themolecule 5-hydroxytryptamine. Continued



LargeAmount

on Cell

/

Fig. 19. These two graphs illustrate the relationship predicted by the model of Fig. 18between crosslinks and histamine release. The upper graph shows that very largenumbers of crosslinks can form on basophils with large numbers of BPO-specific IgEantibodies on their surfaces. These large numbers of crosslinks are accompanied, at(BPO) 2 concentrations around the optimum for release, by reductions in thepercentage of histamine released (lower graph). This reduction represents thetransition from specific to nonspecific desensitization.

treatment with 5-hydroxytryptamine leads to only a minor constituent of the cell mem-a shutdown of the mechanism for calcium branes and is eventually used up. Whenion transport and, hence, to a shutdown of phosphatidylinositol is restored to the cells,secretion. The shutdown occurs because the desensitization is reversed, and the cells oncecalcium gating factor, which has been iden- again secrete normally.tified in this case as phosphatidylinositol, is What the calcium gating factor is in


human basophils, if there is one, is notknown. We have attempted to identify agating factor by incorporating phos-phatidylinositol and other likely candidatesinto nonspecifically desensitized basophils,but none of these substances caused theirrelease of histamine to revert to normal.

Another feature of the model is invoked toexplain specific desensitization. We proposethat decay of a calcium channel is accom-

between which the calcium channel wasformed. This inactivation affects the courseof events as follows. As basophils with smallnumbers of IgE antibodies specific to aparticular allergen are exposed to that al-lergen in the absence of calcium ions, only asmall number of calcium channels areformed and the supply of gating factor is notexhausted. But as the calcium channels

these crosslinked antibodies. This inactiva-tion of the receptors in effect inactivates allthe IgE antibodies specific to the desensitiz-ing allergen, and later exposure to that

allergen in the presence of calcium ions doesnot cause histamine release. On the otherhand, later exposure in the presence ofcalcium ions to a different allergen can causehistamine release because that allergencrosslinks IgE antibodies specific to itself.These crosslinked antibodies can then com-bine with the remaining gating factor to formcalcium channels.

The transition between the two types ofdesensitization occurs when the number ofcrosslinks becomes significant compared tothe amount of gating factor. These ideas alsoare summarized in Table I. Figure 19 showsthe relationship between crosslinks and his-tamine release predicted by the model. Thelower three curves in both graphs essentiallyduplicate the crosslinking and dose responsecurves for basophils with limited numbers ofspecific IgE antibodies on their surfaces(Figs. 14a and 15a). The upper two curves inboth graphs show the effect of large numbersof specific IgE antibodies. In particular, the

39

onset of nonspecific desensitization is in-dicated at the middle (BPO)2 concentrationsby the continued rise (off scale) in thenumber of crosslinks accompanied by adecrease in the percentage of histamine re-leased.

Our model couples specific and non-specific desensitization. Because of thiscoupling the model makes strong predictionsabout the relationship between the timecourse of specific and nonspecific de-sensitization for experiments carried out withcells from the same donor. It also makespredictions about what happens whenbasophils desensitized with one allergen arelater exposed to the same or different al-lergens. Experiments to test these predictionsare in progress at Johns Hopkins.

Of course it could be that specific andnonspecific desensitization are not coupledat all. Indeed, our first guess was thatnonspecific desensitization came about be-cause some gating factor was used upwhereas specific desensitization was the re-sult of a totally independent process calledreceptor-mediated endocytosis. This processwould literally transfer the allergen-linkedIgE antibodies from the cell surface to thecell interior. This guess was based on anumber of examples of endocytosis in whichcells internalize a variety of their own surfacereceptors. In some cases the internalizationis triggered by the binding of a molecule tothe receptor and in other cases by thecrosslinking of molecules bound to the recep-tors. However, Donald MacGlashan hasrecently shown that endocytosis does notoccur in specifically desensitized basophils.He passively sensitized basophils with BPO-specific IgE antibodies and specifically de-sensitized these cells by exposing them to(BPO)2 in the absence of calcium. He then

washed off the (BPO)2 and exposed the cellsto radioactively labeled molecules containingmany BPO groups. He observed that thebasophils specifically bound the radioactivelabel. Therefore, the BPO-specific IgE anti-bodies still remained on the basophil surfacesafter specific desensitization.

Conclusion

Allergic reactions of the immediate typearise because of a complicated chain ofevents: exposure to an allergen, recognitionof the allergen by the immune system, pro-duction of IgE antibodies, sensitization ofbasophils or mast cells, re-exposure to theallergen, triggering of basophils or mast cellsto release histamine and other chemicals,and response of the body to those chemicals.Treatments of allergies are designed to breakthis chain. Of course, the best thing to do isto avoid the allergen. This is straightforwardif you are allergic to cod fish, but impossibleif you are allergic to juniper pollen and insiston living in New Mexico.

If you can’t stay away from the allergen,you can try to break the chain by manipulat-ing the immune response. Almost all allergy“shots” are directed toward this end; theyattempt to affect, not the basophils or themast cells, but the cells of the immunesystem that produce the antibodies. Ex-posures to low concentrations of an allergen(the shots) over long periods can sometimesdesensitize these cells. The cells are then saidto be tolerized because they no longer pro-duce antibodies that bind to the allergen butrather tolerate its presence. If tolerance canbe maintained so that the IgE response isconstantly blocked, the chain is broken andthe patient is free of symptoms, Un-fortunately, a substantial fraction of those

treated with low doses of allergen do notbecome tolerant to the allergen.

A second type of treatment involves in-creasing the immune response rather thandecreasing it. The idea is to produce in anallergic individual such high concentrationsof antibodies of classes other than IgE thatthese antibodies, flowing in the blood, bind tothe allergen and prevent it from triggeringbasophils and mast cells. This raising of the“blocking” antibody concentration has beenhighly successful in the treatment of beesting allergy. Allergic individuals receive reg-ular injections of pure bee venom at concen-trations that raise and maintain a high IgGresponse but do not trigger histamine releasefrom basophils. If such an individual is stungby a bee, the IgG antibodies in solution bindto the bee venom and prevent it fromcrosslinking the venom-specific IgE anti-bodies on basophils. This treatment, how-ever, is much less successful for allergensthat are inhaled than for allergens that areinjected.

An alternative to immunotherapy is drugtherapy. Drugs such as the antihistamineshave been used for some time in attempts toprevent the body from responding to thechemicals released during allergic reactions.Some of these drugs also inhibit, at least tosome extent, the release of chemicals fromthe basophil and mast cell granules.

These are the major approaches to thetreatment of immediate hypersensitivity. Atthe moment there is no therapy in use that isdesigned to bring about basophil and mastcell desensitization. Some cases have beenreported in which immunotherapy produceddesensitization of basophils, but this resultwas fortuitous rather than by design. Per-haps our expanding grasp of the desensitiza-tion process will alter this situation. ■


. . . .


AUTHORS

Byron Goldstein was born and grew up in New York City. He received hisBachelor of Science in physics from the City College of New York in 1961and his Ph.D. in theoretical physics from New York University in 1967. Hebegan to work seriously on problems related to biology while he was aNational Institutes of Health postdoctoral fellow at the University ofCalifornia, San Diego. He joined the Theoretical Biology and BiophysicsGroup at the Laboratory in 1975. Before coming to Los Alamos he was aProfessor of Physics and member of the Biophysics Group at FairleighDickinson University.

Micah Dembo earned his Bachelor of Science in mathematics fromAllegheny College in 1972 and his Ph.D. in biomathematics from CornellUniversity Medical College in 1977. After finishing graduate work, he cameto Los Alamos as a postdoctoral fellow in the Theoretical Biology andBiophysics Group and remained as a staff member after the postdoctoralappointment ended. During his years in the group he has worked on anumber of theoretical problems of importance in biology. In addition todeveloping mathematical models of cell activation and desensitization, he hasworked on the modeling of cooperative interactions in proteins, on diffusionreaction problems, particularly with regard to membrane transportphenomena, and on fluid mechanical models of cell motility.

Further Reading

Kimishige Ishizaka and Teruko Ishizaka, “Mechanisms of Reaginic Hyper-sensitivity and IgE Antibody Response,” Immunological Reviews 41,109-148 (1978).

Anne K. Sobotka, Floyd J. Malveaux, Gianni Marone, Larry I.. Thomas,and Lawrence M. Liechtenstein, “IgE-Mediated Basophil Phenomena: Quan-titation, Control, Inflammatory Interactions,” Immunological Reviews 41,171-185 (1978).

Micah Dembo, Byron Goldstein, Anne K. Sobotka, and Lawrence M.Liechtenstein, “Histamine Release Due to Bivalent Penicilloyl Haptens:Control by the Number of Cross-Linked IgE Antibodies on the BasophilPlasma Membrane,” Journal of Immunology 121, 354-358 (1978).

Lawrence M. Liechtenstein, Martin D. Valentine, and Anne K. Sobotka,“Insect Allergy: The State of the Art,” Journal of Allergy and ClinicalImmunology 64, 5-12 (1979).

Anne K. Sobotka, Micah Dembo, Byron Goldstein, and Lawrence M.Liechtenstein, “Antigen-Specific Desensitization of Human Basophils,”Journal of Immunology 122, 511-517 (1979).

Micah Dembo and Byron Goldstein, “A Model of Cell Activation andDesensitization by Surface Immunoglobin: The Case of Histamine Releasefrom Human Basophils,” Cell 22, 59-67 (1980).

Ann M. Dvorak, H. H. Newball, H. F. Dvorak, and L. M. Liechtenstein,“Antigen-Induced IgE-Mediated Degranulation of Human Basophils,” Lab-oratory Investigation 43, 126-139 (1980).

Teruko Ishizaka, “Analysis of Triggering Events in Mast Cells for Im-munoglobulin E-Mediated Histamine Release,” Journal of Allergy andClinical Immunology 67, 90-96 (1981).

Donald W. MacGlashan, Jr., and Lawrence M. Liechtenstein, “The Tran-sition from Specific to Nonspecific Desensitization in Human Basophils,”Journal of Immunology 127, 2410-2414 (1981).

Paul D. Buisseret, “Allergy,” Scientific American Vol. 247, No. 2, 86-95(1982).

B. Goldstein and M. Dembo, “The IgE-Mediated Activation and De-sensitization of Human Basophils,” in Cell Surface Phenomena, A. S.Perelson, C. DeLisi, and F. Wiegel, Eds. (Marcel Dekker, New York, inpress).


Documents

of Human Allergies