IMMUNOFLUORESCENT STAINING: FLUORESCENT ANTIBODYstaining. Klein and Burkholder (95) also report...

IMMUNOFLUORESCENT STAINING: THE FLUORESCENTANTIBODY METHOD'

ERNST H. BEUTNERUniversity of Buffalo, School of Medicine and Dentistry, Department of Bacteriology and Immunology,

Buffalo, New York

I. Principles of Tmmunofluorescent Staining..................................................A. Fluorescent Antibody Labeling........................................................B. Immunofluorescent Staining Procedures...............................................C. Controls..............................................................................D. Nonspecific Staining..................................................................

II. Antibody Synthesis and Reactions in Vivo.................................................A. Synthesis.............................................................................B. Hypersensitivity Reactions............................................................

III. Immunofluorescent Staining of Microorganisms ..........................................A. Viruses...............................................................................B. Bacteria, Protozoa, and Fungi..........................................................

IV. Immunofluorescent Staining of Tissue Antigens...........................................A. Heteroantibodies to Tissue Antigens...................................................B. Isoantibodies...........................................................................

C. Autoantibodies.........................................................................

V. Acknowledgments.........................................................................

TI. References ..............................................................................

4949515253545455575760626364656969

I. PRINCIPLES OF IMMUNOFLUORESCENTSTAINING

Methods employing fluorescent antibodies areeither the direct or indirect product of AlbertCoons' investigations. They are referred to hereas immunofluorescent staining or IF staining.In essence, IF staining consists of labeling anti-bodies with a fluorescent dye, allowing the labeledantibodies to react with their specific antigen,and observing the reaction product under thefluorescence microscope. The latter consists of astandard microscope with a dark-field condenserand an ultraviolet light source. Since Coons et al.(30) introduced the method in 1942, over 300reports have appeared. IF staining has added newdimensions to our understanding of such subjectsas antibody synthesis, viral infection, and cellularlocalization of tissue antibodies. Several reviewsof the subject are available (23-26, 28, 29, 113).The objectives of this review are: (a) to sum-marize the principles of staining with fluorescentantibodies, particularly as they have evolvedsince Coons' review of 1956 (24); (b) to sum-

1 This investigation was supported in part byresearch grant C-2357 from The National CancerInstitute, U. S. Public Health Service.

marize the major original contributions made bythis method since 1956 (older papers are reviewedonly as they relate to the more recent develop-ments); (c) to relate immunofluorescence to theclassic concepts of immunology.

A. Fluorescent Antibody Labeling

Compounds which emit light in one wavelength range when illuminated by light of ashorter wave length range are fluorescent. Mostfluorescent compounds that have been studiedare excited by light in the ultraviolet or nearultraviolet region and emit light in the visiblerange. Fluorescent dyes are called fluorochromes.The coupling of fluorochromes to antibodies

forms the basis of staining with labeled antibody.Briefly, three methods of coupling fluorescentdyes to proteins have been utilized to prepareconjugates for IF staining. (As used here the term"conjugates" refers to fluorochrome-labeled pro-teins.) According to the original procedure ofCoons, Creech, Jones, and Berliner (30) an iso-cyanate derivative of fluorescein is prepared byreaction of amino group of the fluor with phos-gene. The isocyanate reacts with free aminogroups of proteins to form a carbamido linkage:

49

on June 2, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

http://mmbr.asm.org/

ERNST H. BEUTNER

Fluor-N=C + N-protein >

H2fluor-N-C-N-protein

IIH O H

Later, Riggs et al. (135) introduced a procedurefor conjugating isothiocyanate derivatives offluorescent dyes (fluor-N=C= S) with pro-teins. Presumably the reaction with proteins issimilar to that of isocyanates.

Conjugates may also be prepared with sulfonylchloride derivatives of fluorochromes. Thismethod had been used by Weber (155) and othersto label proteins with the fluorochrome 1-di-methyl naphthalene. Though useful as a proteinlabel, this fluorochrome is of limited value in IFstaining because the intensity of its fluorescenceis low. According to data reported by Uehleke(151), the emission intensity of 1-dimethyl-5-naphthalene sulfonic acid when excited by lightat 365 mAu is about one-third that of fluorescein.Conjugation with sulfonyl chloride derivativesoccurs through free amino groups of the protein.

Fluor-SO2CI + NH2-protein

fluor-S02-NH-protein

According to Uehleke (152), sulfonyl chloridesalso react with secondary amines and alcoholgroups on proteins.The two fluorochromes used most widely thus

far are fluorescein and rhodamine. Fluoresceinhas a yellowish green fluorescence with a maxi-mum at about 550 mAu (48, 151), and lissaminerhodamine B 200 or rhodamine B has a reddishorange fluorescence with a maximum at about640 mu (151). These have been the fluorochromesof choice because the intensity or efficiency offluorescence of these compounds is greater thanthose of other fluorochromes that have beenstudied. The green fluorescence of fluoresceinoffers two important advantages over the redfluorescence of rhodamine: (a) although fluores-cein and rhodamine have about the same in-tensity of fluorescence, the eye is more sensitiveto apple green than to reddish orange; (b) redautofluorescence is more common in nature thangreen autofluorescence.

Conjugation of both fluorescein and rhodaminewith proteins has been carried out by the threemethods described above. According to the origi-nal procedure of Coons et al. (30), fluorescein

isocyanate is conjugated to antibody containingserum globulin. Hiramoto, Engel, and Pressman(77) describe a procedure for conjugating tetra-methylrhodamine to antibody globulins by anisocyanate linkage. Riggs et al. (135) introducedprocedures for conjugation of fluorescein andrhodamine derivatives to globulin with isothio-cyanate linkages. Isothiocyanate derivativeshave the advantage of being more stable andreadily available. J. D. Marshall, Eveland, andSmith (108) report a fourfold greater sensitivity offluorescent reactions with fluorescein isothiocya-nate than with fluorescein isocyanate, althoughthis report has not yet been confirmed. At presentthese are the most widely used labeling reagents.Chadwick, McEntegart, and Nairn (16) andUehleke (151) described methods for conjugatingrhodamine derivatives to proteins with sulfonylchloride linkages. The ratio of rhodamine to pro-tein in the procedure of Uehleke is higher; how-ever, Chadwick et al. report on staining withtheir conjugates, whereas Uehleke does not. Pre-liminary experiments in the writer's laboratoryindicate that the molar ratio of rhodamine toprotein is about 2:1 in conjugates preparedaccording to the procedure of Chadwick,McEntegart, and Nairn. Uehleke also describesa procedure for preparing a sulfonyl chloride de-rivative of fluorescein and for preparing a pro-tein conjugate with it. Although staining reac-tions were not carried out with this conjugate,the work is of interest because conjugation withsulfonyl chloride derivatives is more rapid thanconjugation with isocyanate or isothiocyanatederivatives (16). A promising flurochrome is 3hydroxypyren-5,8, 10-trisulfonic acid. Accordingto Uehleke (151) the wave length of its fluorescentemission maximum (530 m1u) is similar to that offluorescein, but its emission intensity is threetimes greater than that of fluorescein. This ob-servation has not as yet been confirmed. Trials ofIF staining with conjugates of hydroxypyren areindicated.Serum globulins for conjugation may be pre-

pared by fractionating serum with 50 per centsaturated ammonium sulfate (e.g., 30) or 23 percent sodium sulfate (e.g., 80) or by the cold alco-hol method of Cohn et al. (21). Chadwick,McEntegart, and Nairn (16) prepared conjugatesfor IF staining from whole serum.Some noteworthy modifications in the pro-

cedure of conjugating isocyanate and isothio-

50 [VOL. 25


mbr.asm

.org/D

ownloaded from


IMMUNOFLUORESCENT STAINING

cyanate derivatives of fluorochromes to serumglobulins have been introduced. According to theprocedure of Coons (26), serum globulins aremixed with acetone, dioxane, and carbonatebuffer (pH 9) before fluorescein isocyanate isadded. Goldman and Carver (67) conjugatedglobulins without the addition of acetone anddioxane. Fluorescein isocyanate solution wasdried on filter paper strips and stored in a desic-cator. Buffered serum globulins are conjugatedsimply by adding the paper impregnated withfluorescein isocyanate and shaking. Marshall,Eveland, and Smith (108) reported on the con-jugation with fluorescein isocyanate and isothio-cvanate with and without the addition of dioxaneand acetone. According to their data, conjugatesprepared without the addition of organic solventshave greater activity.

Although antibodies labeled with fluoresceinremain serologically active (30), some slightchanges in the properties of labeled proteins aremanifest. The disposition in vivo of labeled eggalbumin is not quite the same as that of the un-labeled egg albumin. Mayersbach and Pearse(112) called attention to a discrepancy betweenthe renal localization of unlabeled proteins asreported by Coons, Leduc, and Kaplan (33) andthe localization of fluorescein-labeled bovineserum albumin reported by Schiller, Schayer, andHess (139). Coons et al. reported demonstrationof egg albumin, bovine serum albumin, and hu-man y-globulin both on and in nuclei of renaltubular epithelium and other cells, whereasSchiller et al. reported localization of fluorescein-labeled bovine serum albumin only on the surfaceof the nuclear membrane. Mayersbach and Pearseconfirmed both observations with the use oflabeled and unlabeled egg albumin. The nuclearmembrane is impermeable to the fluorescein-labeled protein but is permeable to the unlabeledprotein. Another type of change resulting fromconjugation was demonstrated by Curtain (38).The electrophoretic mobility of fluorescein-labeled 72-globulin is greater than that of un-labeled globulin. Uehleke (152) also reported in-creased mobility of proteins conjugated withother fluorochromes.

B. Immunofluorescent Staining Procedures

Authoritative and detailed discussions offluorescent antibody staining techniques may befound in Coons' review of the subject (26) aswell as in various reports on observations made

DIeCr /MMUNOFRcOReSCeNr SM4//NNG

conjusate of specificantibodq globulin with fluorescent antigen -FLUOROCHROME antibodq complex

Figure 1. Outline of direct IF staining. Thisreaction sequence is usually carried out in vitro,sometimes in vivo.

by this method. The object here is to summarizethe concepts that have been evolved and appliedin fluorescent antibody studies. Antibody reac-tions have been localized in sections of tissue(31), smears of cells (e.g., 119), smears of anti-gens (e.g., 56), tissue blocks which are sectionedafter treatment with labeled antibodies (110),cell suspensions (e.g., 8), and living animals(e.g., 60). Several immunologic reaction sequenceshave been successfully utilized.

1. Coons' original procedure (30) consisted ofapplying fluorescein-labeled solutions of anti-body directly to preparations containing the anti-gen(s). This procedure is referred to as directimmunofluorescent staining (or direct IF stain-ing.) The procedure is illustrated in figure 1.

2. Indirect staining with fluorescent antibodyis carried out by treating the preparation con-taining the antigen first with unlabeled specificantiserum. The resulting antigen-antibody com-plex is rendered visible by treating it with afluorescent-labeled antibody to the specific (anti-body) globulin. For example, specific rabbit anti-bodies may be localized with labeled goat or horseantirabbit -y-globulin antibodies; specific humanantibodies may be localized with conjugates ofrabbit or goat antihuman -y-globulin antibodies,and so on. This procedure is referred to as indirectIF staining. It is illustrated in figure 2. Wellerand Coons (158) were the first to describe theprocedure. Watson, in Coons' laboratory was thefirst to demonstrate it. Indirect IF staining makesit possible to use a single conjugate to study avariety of antisera of a given species. The proce-dure is more readily controlled and more sensi-tive than direct IF staining. Indirect IF staininghas greatly expanded the utilization of the fluo-rescent antibody method. Conjugates of anti-globulins are now commercially available.

3. A complement-staining procedure was re-ported by Goldwasser and Shepard (71). This

511961]


mbr.asm

.org/D

ownloaded from


ERNST H. BEUTNER

/ND/P6'Ce /MMtN/vGOESceWr SM/N/NG

A,

i34

antigen specific antibodq8lobuln in whole serum antigen - antibodq

\\omplex0 0~complex

fluorescent antigen -~~'S' j!}!l~k'IjiL\! anibbodq antianlibodqantibodt to antibodq complexglobulin of step A. withFLU OROCHROMME LABEL

Figure 2. Outline of indirect IF staining. The second step with the labeled antiglobulin (e.g., labeledrabbit antibodies to human y-globulin for demonstration of human antibodies) is carried out in vitro.

technique is based on the fact that complement,if present, participates in immunologic reactions.By heating a serum to 56 C for 30 minutes itscomplement activity is destroyed. Guinea pigserum is rich in complement and is usually usedas a source of complement in serologic reactions.The procedure consists of applying a mixture of aspecific antiserum and whole fresh guinea pigserum to a section or smear containing the anti-gen. The complex of antigen, antibody, and com-plement which forms on the slide is visualized byapplying fluorescent-labeled antibodies to guineapig complement. Goldwasser and Shepard reportgreater sensitivity in detecting serologic reac-tions by complement staining than by indirect IFstaining. Klein and Burkholder (95) also reportspecific IF staining with fluorescein-labeled anti-complement antibodies.

Triple layer staining has been reported, thoughthe results are open to question. In our experi-ence, nonspecific staining is excessive. Still otherIF staining procedures have been applied instudies of antibody localization. These are sum-marized in the section on antibody synthesis andreactions in vivo.

C. ControlsIn staining with labeled antibodies, as in any

serologic method, controls constitute an integralpart of experimental procedures. Three types of

controls should be applied. (a) In all IF experi-ments it is necessary to identify the observed fluo-rescence as a serologic reaction. The precise natureof the necessary controls depends to some extenton the type of IF staining. A list of appropriatecontrols for direct IF staining is detailed by Coons(26). These include: inhibition of reaction byadsorption of conjugate with specific antigenand failure to inhibit by adsorption with heterolo-gous antigen; inhibition of staining by pre-treatment with unlabeled antiserum or by mixinglabeled with unlabeled antiserum; and negativereactions with labeled normal serum globulin orheterologous antiserum globulin. Each experi-ment carried out by indirect IF staining shouldinclude control preparations treated with con-jugate alone, as well as preparations treated withnormal serum or heterologous antiserum (inplace of the specific antiserum) and antiglobulinconjugate. Still other controls of the serologicnature of staining reactions are suggested in theliterature on IF studies. Immunofluorescentstaining by its very nature frequently providesintrinsic evidence of the serologic nature of stain-ing reactions, e.g., staining of selected bacteria ina mixed population or staining of specific histo-logic structures in tissue sections with tissueantibodies. (b) In any system that has not alreadybeen thoroughly investigated by immunofluores-cent methods, it is necessary to demonstrate the

52 (VOL. 25


mbr.asm

.org/D

ownloaded from



serologic specificity of the observed reaction. That is,the spectrum of cross reactions (if any) of theantibodies with related or different antigens mustbe ascertained. The serologic specificity of anti-bodies that have already been studied for crossreactions by other serologic methods should bereinvestigated by IF staining, since differentspectra of cross reactions are observed by differ-ent methods. For example, antibodies to group Af-hemolytic streptococci cross react with groupC by IF staining (118) but not by precipitation.As a rule, the number of control antigens thatshould be tested is one more than the numberalready tested. (c) Ultimate confirmation of IFstaining reactions lies in the demonstration of thegiven reaction by other serologic methods. Suchconfirmation is always desirable.

D. Nonspecific StainingMost fluorescent antibody studies are com-

plicated by nonspecific staining of the antigenwith the labeled globulins. Some recent studiesshed light on the nature of these nonspecificstaining reactions. Curtain (38) fractionatedfluorescein-labeled antisera by electrophoresisconvection. It is evident from his comparison ofantibody titers, nonspecific staining reactions,and protein concentration of the fractions, thatserologically inactive as well as active fractionsgive rise to nonspecific staining. Louis, Hughes,and their associates (see 106 for references) reportthat a variety of fluorescein- and rhodamine-labeled proteins, including normal serum glob-ulins and albumins as well as labeled egg al-bumin, give rise to nonspecific staining. Theprime interest of these investigators is thedifferentiation of normal and neoplastic cells bytheir staining with labeled proteins. Louis (106)summarizes the histologic distribution of thesenonimmunologic staining reactions with anextensive list of normal tissues which are stainedand neoplastic tissues which do not stain. Thecytoplasm of all normal epithelial cells and whiteblood cells in mammalian tissues stained, whereasconnective tissue, central nervous system tissue,and red blood cells did not.

Adsorption of conjugates with acetone-driedliver powders according to the method of Coons,Leduc, and Connolly (32) is one of the mostwidely used methods for removing nonspecificstaining. Frequently, only partial removal ofnonspecific staining is effected by this method(e.g., 38). Mayersbach (111) notes that freeze

drying, storage of frozen sections, formalinfixation, and pretreatment of tissue with alkalinebuffers reduce nonspecific staining. In my ex-perience, pretreatment with alkaline buffers orformalin not only inhibits nonspecific stainingbut also destroys certain antigens in epithelialcells.

Observations such as those of Louis (e.g.,106) suggest that normal serum proteins bindwith a variety of antigens. This is also manifestin controls of indirect IF stains. Control prepa-rations treated with normal serum and conjugateare consistently stained more strongly thanpreparations treated with only conjugates. Thus,Smith, Marshall, and Eveland (143) rendernonspecific staining red or brown by pretreatingpreparations with rhodamine conjugates ofnormal serum globulin. Similarly, in studies withindirect IF staining, nonspecific reactions areinhibited by pretreatment of preparations withnormal serum of another species (e.g., 81).Another concept which has been applied in thewriter's laboratory is that nonspecific binding ofserum proteins is different from binding ofspecific serologic reactions. Thus extraction ofpreparations with 50 per cent glycerol has beenused to remove nonspecific staining.

Nonspecific staining reactions are probably inpart a function of the amount of fluoresceinderivative used in the conjugation mixture. Coonsand his associates (30, 31) obtained optimalresults with 5 mg of fluorescein isocyanate(measured as fluorescein amine) per 100 mg ofcrude globulin protein. Ratios of molecularfluorescein to protein of about 2:1 are obtainedin such conjugates. Conjugates with lower ratiosof fluorescein to protein yield progressively lessIF staining. There is little information in theliterature relevant to the influence of largeramounts of fluor in the conjugation mixture onnonspecific and specific staining. Availableinformation suggests that nonspecific staining isincreased, whereas specific staining remains thesame or decreases. This increased nonspecificstaining appears to be caused by the increasedaffinity of the altered proteins for a variety ofsubstances.Two extreme examples serve to illustrate the

type of situations that may be encountered. Atone- extreme are leukocytes. Coons et al. (32),Louis (106), and others (40) who have encoun-tered them in fluorescent antibody studies havenoted their affinity for nonspecific staining. No

19611 53


mbr.asm

.org/D

ownloaded from


ERNST H. BEUTNER

amount of adsorption of conjugates will removethis nonimmunologic staining reaction. Danaher,Friou, and Finch (40) reported on indirect IFstaining of human white blood cells with isoanti-bodies. They were able to observe some specificstaining despite the strong nonspecific reaction.Indirect immunofluorescent staining with isoanti-bodies or autoantibodies is further complicated bythe fact that the conjugate reacts directly with-y-globulin and related antigens in the tissue aswell as specific antibodies. Thus Danaher et al. re-port observations which suggest that y-globulin ispresent in or on normal leukocytes. In sharpcontrast to the nonspecific staining problemsencountered with white cells is the total lack ofnonspecific staining of red blood cells. Cohen,Zuelzer, and Evans (19) report completely specificstaining of red cells with labeled blood groupisoantibodies without adsorption of conjugates.These authors report detailed observations onanother type of nonspecific staining frequentlyobserved: the trapping of conjugate under theantigen on the slide. Trapped conjugate, as theseauthors point out, can be readily distinguishedfrom specific staining.

II. ANTIBODY SYNTHESIS AND REACTIONSIN VIVO

A. Synthesis

The first direct evidence for the cellular site ofantibody formation came from studies of Coonset al. (32). Coons' review of 1956 (24) should beconsulted for an authoritative discussion. Basi-cally, four types of fluorescent antibody proce-dures have been brought to bear on this problem.Two of them yielded valuable results. Theprocedures and the essential results obtained wviththem may be summarized here.

1. Historically, the first approach used byCoons was direct staining for antigen e.g. (22).The globulin fractions of antisera to the tracerantigen were labeled. Animals treated by injec-tion with the tracer antigen were killed, and thelocalization of the antigen was determined by di-rect staining. By this method, a clear demonstra-tion of antigen localization was achieved. How-ever, it was manifest from the studies of Coonsand his associates that the cellular localization ofantigen and the site of antibody synthesis are notthe same.

2. The next logical step is to label antigenswith fluorescein. Information on antibody syn-

thesis has not been obtained by this method.However, Germuth et al. (60) and others havestudied antibody reactions in vivo by this method.It should be pointed out that not all antigenscan be labeled with fluorescein by the methodsused for labeling globulins.

3. The most significant results have beenachieved by a two-step reaction (2, 27, 32, 123,160, 164). Sections of tissue containing antibodcyare flooded first with the antigen, then with la-beled antibodies to this antigen (see figure 3). Itis by this method that the first direct demon-stration of antibody inside of plasma cells wasachieved (32) and the long standing controversyconcerning the site of antibody synthesis was fi-nally resolved.

4. More recently, direct IF staining for 'y-globulin with labeled antiglobulins has beenutilized to study antibody synthesis (127, 130).Results are in essential agreement with thoseobtained by the two-step method describedabove. However, direct staining for y-giobulin isa relatively nonspecific histochemical approach.The specificity of detected antibodies is unknown.Indeed, this method assumes that all y-globulinsare antibodies.The observations on antibody synthesis in the

cytoplasm of the lymphocyte-plasma cell series byCoons et al. (32) have been confirmed and ex-tended in several laboratories (2, 130, 160, 164).White (160) noted that Russell bodies withinplasma cells are composed largely of antibodies.Witmer (164) studied antibody formation in theeye by the double layer technique. Plasma cellswhich infiltrated the iris and ciliary body of theeye during inflammatory conditions were demon-strated to contain antibodies. Antibody titers asdetermined by tanned cell hemagglutinationroughly paralleled the appearance of antibody-containing plasma cells. Askonas and White (2)examined for formation of antibody to ovalbuminin guinea pigs. They correlated the number ofplasma cells stained by the fluorescent antibodymethod with the amount of anti-ovalbuminantibody produced in vitro in various tissues.With the exception of the bone marrow, thecorrelation was good.More recently, Coons (27) reported demon-

stration of occasional cells with intranuclearantibody. Such nuclei were observed to containbright granules with high antibody concentra-tions. These granules are interpreted to be nu-cleoli. He further reported that a large number of

54 [VOL. 25


mbr.asm

.org/D

ownloaded from



DOUBLE LAYeR IMMUNOFLUORESCeNT STAINING

FOR ANrIBODNOY LOCM4/ IZ/ON

~+ 0

specific intracellular anti- speczfic antigenbodies in tissue section

antigen - antibodqcomplex

specific antibodtwilth fluot-ochrorne

fluorescent antibod4 -

antigen - antibodqcomplex

Figure 3. Procedure for indirect fluorescent antibody staining for specific antibodies in tissue

nucleoli are found in association with active anti-body synthesis. The passive transfer of antibody-producing cells has been studied intensively byDixon and his associates (46, 137, 138). Theyinjected lymphatic cells (lymph node cells,thymus cells, and peritoneal exudates) from im-munized animals into recipient animals whichhad been rendered immunologically inert bywhole body irradiation. In a recent publication,Neil and Dixon (123) utilized the fluorescentantibody method together with other histologicmethods to trace the fate of cells producing anti-body after their transfer into recipient animals.Their data suggest that lymphocytes from donorsare transformed into plasma cells; that the de-velopment of plasma cells corresponds to theappearance of cells containing antibody; andthat secondary or anamnestic responses can beelicited in animals passively sensitized withcells of the lymphocyte series. In all these studies,the cellular localization of antibodies was carriedout by the two-step staining method. (Seefigure 3 and procedure 3 above.)

Since many antibodies are -y-globulins, intra-cellular antibody localization can also be achievedbv direct IF staining for y-globulin. (See proce-

dure 4 above.) Some noteworthy studies based on

the assumption that all y-globulin molecules are

antibodies have been reported. Studies of y-

globulin localization in human lymph nodes andspleen were carried out by Ortega and Mellors(130). Specific -y-globulin localization is reportedin plasma cells with and without Russell bodies,and in a cell type referred to as "intrinsic cells ofgerminal centers" of lymph nodes. According to

the authors, these are exactly the same cell typesthat were demonstrated by Coons et al. to pro-

duce specific antibodies. Localization of -y-glob-ulin in the spleen and liver of patients with hy-perglobulinemia was studied by Ohta et al. (127).According to their report, y-globulin was foundin plasma cells and related cell types as well as inlarge mononuclear splenic tumor cells.

B. Hypersensitivity ReactionsThe mechanism of hypersensitivity reactions

induced by reactions of circulating antibodiesin vivo has been elucidated by immunofluorescentstaining (17, 45, 60, 107, 114, 127, 142, 165). Theapproaches used in these studies include directIF staining for antigen and/or 'y-globulin on

sections of hypersensitivity lesions and theinjection of fluorescein-labeled antigens intosensitized animals, followed by observations on

the localization of the labeled antigen in lesions.Many investigators prefer to support IF stainingobservations with other serologic studies. Amongthe classic experimental hypersensitivity reac-

tions that have been studied with the aid ofstaining with labeled antibody are: local arthusreactions (local antibody-induced lesions), ana-

phylactic reactions (immediate systemic hyper-sensitivity reactions in a host previouslysensitized to the offending antigen), serum

sickness (systemic hypersensitivity reactions toforeign serum proteins which usually develop 5to 10 days after injection), glomerulonephritis,and nephrotoxic serum-induced kidney disease.Germuth et al. (60) demonstrated local Arthus

reactions in the avascular cornea of the eye.

7

19611 55


mbr.asm

.org/D

ownloaded from


ERNST H. BEUTNER

Reactions were produced by injecting fluorescein-labeled bovine serum albumin (BSA) into thecenter of the cornea of rabbits previously sensi-tized to BSA. Lesions took the form of an opaquearc around the site of inoculation. By examinationwith ultraviolet light, the are was demonstratedto correspond precisely to the localization of thelabeled antigen. Passive sensitization was demon-strated by injecting into normal rabbit corneaantisera to BSA on one side and fluorescein-labeled BSA on the other. An opaque linedeveloped between injection sites, again cor-responding to the antigen localization. Theseobservations leave little doubt that antigen-antibody complexes are directly associated withthe observed lesions of the Arthus reaction. Thepulmonary thrombi observed in active or passiveanaphylazis in rabbits were identified by McKin-non et al. (107) as antigen-antibody precipitate.The identity of these thrombi with immuneprecipitate was demonstrated by labeling thechallenging antigen BSA or bovine y-globulin(BGG) with fluorescein and observing the specificfluorescence of the immune precipitate. Theauthors suggest an association between thesethrombi and the pathogenesis of anaphylaxis inthe rabbit. Dixon et al. (45) investigated serumsickness reactions by injecting into rabbitsantigens (BSA and BGG) labeled with radioactiveiodine and fluorescein. Careful kinetic studies onthe disappearance of antigen and the developmentof antibodies clearly indicated that characteristicserum sickness lesions in the kidneys and arteriesappear during the time interval in which bothantigen and antibody are present. Specificlocalization of antigen and host 'y-globulin at thesite of the lesions was demonstrated by appro-priate direct IF staining for BSA and rabbity-globulin.

Glomerulonephritis is a prominent feature ofthe serum sickness syndrome. This is in part areflection of the fact that all manner of foreignantigens tend to become localized in the kidney.Kaplan's (89) fluorescent antibody studies onthe distribution of polysaccharides of group A3-hemolytic streptococci illustrate this fact.The observed sequellae are localization of host-y-globulin and a local Arthus reaction. Someinvestigators have focused their attention solelyon the kidney lesions following foreign antigeninjection. Direct association of localization of host-y-globulin in the kidney with BSA-inducedglomerulonephritis (in rabbits) was demonstrated

in fluorescent antibody studies of Mfellors et al.(114). In effect, this is part of the serum sicknesssyndrome. Bovine serum proteins are not uniquein their capacity to induce the development ofglomerulonephritis. A variety of other animalproteins and bacterial antigens have been used toinduce similar disease processes. Wood and White(165) studied glomerulonephritis produced inmice by repeated injections of heat-killed suspen-sions of Proteus mirabilis. The bacterial antigenwas demonstrated by direct IF staining to beassociated with kidney lesions. Agglutinationtiters with P. mirabilis appeared to be related tothe severity of the kidney lesions. Efforts toextrapolate observations to the pathogenesis ofhuman kidney diseases and other diseases havebeen reported from several laboratories, chieflyfrom the laboratories of Mellors (115), Dixon(44), and Gitlin (62). Localization of -y-globulinand other serum proteins in the kidney lesionsand other pathologic tissues has been investi-gated. These important studies are ably reviewedby Mellors (113). Such studies are for the mostpart histochemical in nature, since the immu-nologic basis of the observed localization of y-globulin and other serum proteins is not clearlyunderstood.

.More specific localization of foreign antigens inthe kidney is achieved with nephrotoxic antisera.These are usually antisera to the kidney of theanimal to which the serum was administered;e.g., rabbit antisera to rat kidney injected into therat produce nephrotoxic nephritis. The disease ischronic and sometimes fatal. The incidence ofkidney lesions following injection of nephrotoxicantisera is reported to be 100 per cent. However,the incidence of kidney lesions is reported to be ashigh as 80 per cent following the injection oflarge doses of other foreign antigens (114).Thus, as Ortega and Mellors point out (129),the consistent appearance of kidney lesionsfollowing injection of nephrotoxic sera is simplythe result of the more specific kidney localizationof the antibodies. This interpretation is consistentwith the views of Kay (91). His data suggestedthat kidney lesions and death in nephrotoxicnephritis resulted not from the action of thenephrotoxic sera on the kidney, but from thereactions of the host antibodies to the nephro-toxic antibodies in the kidney. It is noteworthythat not only kidney antisera but also antisera tolung and other organs are effective in producingkidney lesions (Baxter and Goodman (5)). This

56 [VOL. 25


mbr.asm

.org/D

ownloaded from


]IMMUNOFLUORESCENT STAINING

is a reflection of the fact that antibodies to kidneyare, in large part, not organ specific but speciesspecific. That is, they react with other organs andtissues of the antigen-supplying species. This wasalso demonstrated by localization in vivo and invitro of anti-kidney antibodies (Hiramoto et al.,79). Localization in vivo was demonstrated byintravenous injection of rabbit anti-rat kidneyantiserum into rats and staining rat organsections with anti-rabbit serum globulin conju-gate, whereas localization in vitro was demon-strated by indirect IF staining with the samecombination of antisera. Connective tissuestaining was demonstrated in kidneys, adrenals,ovaries, thyroid, spleen, lymph nodes, and liver.Under the conditions used, reactions in vitrowere stronger than those in vivo. According tothe authors, qualitative differences betweenreactions in vivo and in vitro are observed. Specificlesions have been reported only in the kidney,presumably because it is the most susceptibleorgan. The excellent reproducibility of nephro-toxic nephritis lesions has stimulated extensivestudies. Ortega and Mellors (129) reported onfluorescent anti-globulin staining, histologicobservations, and functional changes in nephro-toxic nephritis of rats induced by rabbit anti-ratkidney antibodies. Rabbit y-globulin remaineddemonstrable for 3 months. Increase in raty-globulin which was demonstrable in rat glome-ruli during the first postinjection week wasascribed to an accumulation of edema fluids,whereas further increases in glomerular concen-tration of rat y-globulin during the 2nd to 15thweek were attributed to the specific accumulationof autogenous rat antibodies to rabbit immuneglobulin. Similar observations are reported bySeegal (142). A delayed form of nephritis pro-duced by injection of small amounts of duckanti-rabbit kidney antisera into rabbits wasinvestigated. In contrast to the nephritis pro-duced by large doses of nephrotoxic sera, noevidence of immediate renal damage was mani-fest. Duck antibodies were demonstrable in thekidneys of the injected rabbits for 7 to 15 days.With the development of signs of nephritis,rabbit antibodies become demonstrable.

III. IMMUNOFLUORESCENT STAINING OFMICROORGANISMS

An extensive bibliography of this subject maybe found in Coons' recent review on the diagnosticapplications of the method (29). The objectives

here are to review the most significant contribu-tions that have been made since Coons' reviewof 1956 (24) and to discuss the literature interms of the various ways in which immuno-fluorescence has been brought to bear on micro-biological problems.

A. VirusesStaining with labeled antibody lends itself to

studies in situ of viral infections. The preciseintracellular localization of viral antigen isdemonstrable although the viral particles aresubmicroscopic. The method has been brought tobear at various levels of investigation: (a) in theinitial characterization of viral agents as inmeasles (20, 51) and the foamy agent of monkeykidney cell cultures (13); (b) in the intracellularlocalization of known viral agents (9, 10, 12,97, 124, 125); (c) as a potential diagnostic tool(18, 69, 70, 100); and (d) in more precise studiesof virus-host cell interaction (8, 75, 98, 102, 133,134, 154).

Enders' reports on the measles virus (50, 51)serve as a model of exploratory studies. A viralagent isolated in human kidney cell cultures wasidentified as the etiologic agent of measles bycytologic and serologic methods. Characteristiccytopathogenic changes in tissue culture wereassociated with virus by three serologic methods:(a) inhibition by human convalescent sera, (b)fixation of complement by the infected tissueculture and human convalescent sera, and (c)indirect IF staining of infected tissue with humanconvalescent sera. Enders reports cytoplasmiclocalization of the viral antigen. One of the basicprecepts of immunology illustrated by this studyis that all serologic reactions should be confirmedby the use of two or more test methods. Cohenet al. (20) arrived at similar conclusions inde-pendently. Viral agents were isolated on monkeykidney cell cultures from active cases of measles.They also demonstrated complement fixation andIF staining with infected tissue cultures andhuman convalescent sera. According to theirreport, the viral antigen was first demonstrablein the cytoplasm and later in the nucleus ofinfected cells. Baker et al. (3) later reported theobservation of intracytoplasmic and intranuclearinclusions by both IF staining and electronmicroscopy.

Accurate quantitation of the number ofinfectious viral particles has been achieved formany groups of viruses by plaque count methods.

19611 57


mbr.asm

.org/D

ownloaded from


ERNST H. BEUTNER

(A plaque is a discrete focus of infection in asusceptible cell population which may be pro-duced by a single viral particle.) The method ofdemonstrating plaques depends on the virus-host cell system under consideration. Rapp et al.(134) adapted fluorescent antibody staining forplaque counts of measles virus in tissue culture.Infected monolayers of human epidermoidcarcinoma cells grown on cover glasses werecovered with methyl cellulose to prevent thespread of virus. Infectious foci were demon-strable by the indirect IF method after 3 to 6days of infection under the conditions used.The number of foci was roughly proportional tothe dilution of the virus suspension, according tothese authors.Eaton et al. (47) isolated a virus which he

implicated as the etiologic agent of primaryatypical pneumonia. This virus can be carried inchick embryos but produces no demonstrablepathology. The only test system available inhis earlier studies was pathogenicity for thecotton rat. In a series of studies, Liu and Eaton(103), Liu (101), and Liu and Heyl (105) uti-lized immunofluorescent staining to investigatethis system. Viral antigen was demonstrated bythe indirect IF method with both convalescentsera from patients with primary atypical pneu-monia and with rabbit antisera. The antigen wasspecifically localized in the bronchi of the chickembryo near the bifurcation of the trachea. Thereaction of convalescent sera is of particularimmunologic significance. It takes place only inthe presence of complement-like activity of freshnormal serum. (Normal human and normalguinea pig sera were used by Liu.) This is, atpresent, the only reported IF staining reactionwhich is enhanced by complement, althoughundoubtedly others will come to light. Associatedwith primary atypical pneumonia is the develop-ment of cold agglutinins (antibodies whichagglutinate the patient's own red cells at refriger-ation temperatures) and agglutinins for strepto-coccus MG. Liu, Eaton, and Heyl (104) demon-strated that: (a) antibodies responsible forindirect staining of infected chick embryos wereserologically unrelated to antibodies involved incold agglutination or streptococcus MG aggluti-nation; (b) such IF reactions are a more sensitiveindicator of antibodies to primary atypicalpneumonia virus than pathogenicity for thecotton rat; (c) convalescent sera usually showsignificant increases in antibody titer over acute

sera; and (d) IF staining reaction titers paralleltiters of the neutralization test in cotton rats.Thus, labeled antibody staining is demonstratedto be specific for the virus. This is an example ofa system in which IF staining is the only specificserologic in vitro test presently available. Stillanother example of the use of antibody stainingin exploratory studies of viruses has arisen fromobservations on monkey kidney cell cultures.Laboratory workers were frequently hamperedby "foamy" degeneration of kidney cell cultures.Carski (13) demonstrated, by direct IF staining,specific reactions of selected monkey sera withkidney cell cultures manifesting foamy degenera-tion. By use of indirect IF staining, he was ableto determine which monkeys were free frominfection with the "foamy agent" virus.The elegant intracellular localization of viral

antigens by immunofluorescence has attracted anumber of workers to investigate known virusesby this method. Noyes and Watson (124) studiedvaccinia viral infection in a tissue culture strainof human epidermoid cancer. Viral antigen wasdetected by direct fluorescent antibody staining.Specific staining was reported to occur in the formin cytoplasmic granules. Late in the infectiousprocess, uniform staining extending into thenucleus was observed. Noyes (125) studied aneurotropic human virus, the Egypt 101 strain,with the aid of direct fluorescent antibodystaining. The viral antigen was demonstrated inboth infected mouse brain and in a tissue cultureof human epidermoid cancer to which Noyesadapted the virus. Specific fluorescent antibodystaining occurs only in the cytoplasm of infectedcells, according to this report. Lebrun (97)studied the development of herpes simplexvirus in tissue culture, chick embryo, andmouse brain by means of the fluorescent antibodymethod. In tissue cultures of human epidermoidcarcinoma, viral antigen appears as intranucleargranules 24 hours after infection; 48 hours afterinfection, nuclear staining decreases and granularcytoplasmic staining appears. Buckley (11)observed labeled antibody reactions on tissuecultures infected with poliovirus. Similar localiza-tion of reactions was observed with types I,II, and III on several types of human cellcultures. The need for careful control of thedosage of virus in studies of the sequence ofchanges during the course of infection is broughtout by Lebrun's report (98) on the poliovirus.When only a small proportion of the tissue

58 [VOL. 25


mbr.asm

.org/D

ownloaded from



culture cells are infected with viruses initially,virus liberated from these cells infects other cells.Subsequent IF staining reactions occur withcells in different stages of the infectious process.However, by using 100 per cent initial infection,Lebrun obtained complete destruction of cells in7 hours. Under these conditions, only cyto-plasmic staining occurs up to the late stages ofinfection (5th to 7th hour). The late appearanceof nuclear staining is ascribed to the developmentof nuclear lesions in degenerating cells.

Infection in vivo of wild and domestic rabbitswith Shope papilloma was studied with fluores-cein-labeled antibodies by Noyes and Mellors(126). Viral antigen could be detected only in asmall proportion of the domestic rabbits andthen only in the keratohyaline layers and in thekeratinized layers, not in the proliferating cellsor in the germinal cells. Specific fluorescentantibody staining was present only in the nucleiof infected cells. Viral antigen was more fre-quently detectable in wild cottontail rabbits.The authors postulate that a protein-deficientnucleic acid provirus which cannot be stained ispresent in the proliferating cells. Boyer, Denny,and Ginsberg (9) reported on careful studies of thedevelopment of adenovirus type 4 in HeLa cellcultures by indirect IF staining. Localization ofvirus was related to cytopathogenic changes asdemonstrated by histologic methods and byelectron microscopy. Conflicting results wereobtained by cell fractionation. The discrepancywas clarified by careful fluorescent antibodystudies.Moulton (121) studied naturally occurring

cases of canine distemper with the aid of directIF staining. Hyperimmune rabbit sera servedas a source of conjugate. Cerebella of infecteddogs were stained. Specific staining was reportedin nuclei of glial cells, whereas no specificstaining occurred in demyelinated nerve fibers.Liu and Coffin (102) elucidated the pathogenesisof canine distemper viral infection by studies ofthe progressive changes in virus localization.Viral antigen was demonstrated with conjugatesof globulins from convalescent dog sera. Ferretswere infected by intranasal inoculation. Thevirus invades the reticuloendothelial system viathe cervical lymph nodes and progresses to theviscera, brain, and other organs of the body.Viral antigen appeared as intracytoplasmicgranules. Coffin and Liu (18) propose a diagnosticprocedure using stains of conjunctival smears.

The viral nature of Negri bodies in rabies wasdemonstrated with fluorescein-labeled antibodies.by Goldwasser and Kissling (69, 70). The authorsreport specific staining of infected mouse brainswith a human antiserum by direct and indirectIF staining and complement-staining procedures.The structures stained by the fluorescent anti-body techniques were shown to correspond toNegri bodies by staining sections for visiblelight microscopy after examination of the fluores-cent stain. Diagnosis of rabies by direct IFstaining of salivary gland smears was attemptedby Goldwasser et al. (70). Among 56 positivespecimens on which detailed data are supplied,49 were positive by culture of salivary glandsmears. Of these 30 were positive by IF staining;18 yielded occasional positive smears, and onewas negative; a false positive reaction was re-ported in one of eight smears from one specimen.The authors report negative IF staining in 96negative specimens.

Influenza virus has been subjected to extensivestudies by the fluorescent antibody method (seeCoons, 24, for a detailed review of the subject).Liu proposed a rapid diagnostic procedure forinfluenzal infections utilizing the fluorescentantibody method (100). Smears of cells fromnasal washings are stained with fluorescein-labeled rabbit antiserum to influenza virus. Theprocedure is reported to be somewhat lesssensitive than the demonstration of a rise ofhemagglutination inhibition titer.

Immunofluorescence adds a new dimension tostudies of virus-host cell interaction. Watsoncarried out IF staining studies of mixed infection(154). Chick embryos were infected with influ-enza and mumps viruses under varying conditions.Her results suggest simultaneous infections ofcells with both viruses under certain conditions,though this is difficult to prove with antibodieslabeled with a single fluorochrome.

Careful studies of phagocytosis of PR8influenza virus in vitro were reported by Boand,Kempf, and Hanson (8). Virus in leukocytes wasdemonstrated by direct staining with reagentsprepared from hyperimmune rabbit sera. Obser-vations of staining reactions were confirmed byhemagglutination and chick cell agglutinationreactions. Normal leukocytes in the presence ofnormal serum phagocytized no virus. Only surfaceadsorption was observed. Leukocytes from im-mune donors suspended in normal serum, aswell as normal leukocytes suspended in im-

1961]


mbr.asm

.org/D

ownloaded from


ERNST H. BEUTNER

mune serum, phagocytized some virus. Optimalphagocytosis occurred with leukocytes fromimmune donors in the presence of immuneserum. Phagocytosis in vitro, when observed,occurred within 30 minutes. These studies enabledHanson, Kempf, and Boand (75) to undertakepreliminary studies in vivo of virus phagocytosis.Mice treated by intraperitoneal injection withPR8 influenza virus were killed at time intervalsup to 28 hours to demonstrate phagocytosis inperitoneal exudates. Some important differencesfrom phagocytosis in vitro were reported. Leuko-cytes in normal mice and mice immunized withtwo injections exhibited phagocytosis after 16hours but not after 8 hours, whereas leukocytesfrom mice immunized with three injectionsexhibited phagocytosis after 8 hours. This is incontrast to phagocytosis in vitro in 30 minutes.Further IF studies of phagocytosis in vivo andin vitro and its relation to immunity seemindicated.

Prince and Ginsberg (133) detected by IFstaining a form of Newcastle disease virus inEhrlich ascites tumor cells which was noninfec-tious, did not hemagglutinate, and did not fixcomplement. Infected cells were inoculated intomice. Noninfectious antigen was detected inharvested cells by indirect IF staining aboutthree hours after infection. Staining appeared ascytoplasmic granules. One of the interpretationsproposed by the writers is that incomplete viralantigen may be synthesized in the tumor cells.

B. Bacteria, Protozoa, and FungiMost reports on the application of IF staining

to bacteriology deal with potential diagnosticuses of the method (41, 74, 86, 109, 118, 146-148).Also, the fate of injected bacterial antigens hasbeen traced in experimental animals (35, 48, 89,140, 141). A few observations on bacterialcytology have come to light in the course ofstudies of fluorescent antibody (42, 43, 148).Immunofluorescent staining has been usedto advantage in relating the proportion of givenvarieties of bacteria on direct smears to theorganisms isolated culturally (80).A review of the literature on applications of

the fluorescent antibody method to diagnosticbacteriology must be preceded by a discussion ofsome basic immunologic concepts. Immuno-fluorescent staining is a serologic test. Specificityand sensitivity are the principal factors to beconsidered in evaluating any serologic method.

Specificity is of prime importance. In diagnosticbacteriology, serologic tests with known antiserausually follow cultural isolation and identifica-tion. This limits the scope of cross reactions thatmust be considered. If crude mixtures of bacteriaare used, a much wider range of cross reactionsmay come into play. The range of cross reactionswith a given type of antiserum may vary with themethod of testing. Thus the specificity of eachnew serologic test method must be evaluated;i.e., extensive studies of cross reactions withvarious types of bacteria must be carried out.Similar specificity controls must be carried outfor the Widal-Wassermann type of diagnostictest (i.e., tests with known bacterial antigensand unknown patients' sera). The sera frompatients with other clinically diagnosed diseasesmust be tested with the antigen by establishedtest methods as well as by the proposed newmethods.The term "sensitivity" in the classical sense

refers to the capacity of a serologic technique todetect the presence of antibodies. A techniquewhich detects antibody reactions at higher serumdilutions than other methods or one whichdetects reactions not detectable by other methodsis said to be the most sensitive. In my experience,the sensitivity of indirect IF staining (in termsof the highest positive serum dilutions) is usuallyapproximately the same as complement fixation,greater than precipitation, and less than aggluti-nation reactions. Some notable exceptions maybe found, e.g., nuclear autoantibody reactions.Estimates of the amount of antibody per milli-liter of serum demonstrable by IF stainingare not available at present. The capacity ofa serologic test to detect small amounts ofantigen is also of value in some situations. Inthis respect, IF staining is outstanding. Coonspoints out (24) that a single pneumococcuswhich contains about 5 X 10-11 mg of nitrogencan be readily detected. Small numbers of bacteriacan be detected with a high degree of precision(120, 149). This becomes particularly valuable indealing with organisms which cannot be culturedon artificial media (41).

Diagnostic use of fluorescent antibodies in thedetection of enteric pathogens has been carefullyinvestigated by Thomason and her associates atthe United States Communicable Disease Center.The sensitivity of direct IF staining in thedemonstration of antigenic components ofSalmonella spp. is comparable to slide agglutina-

60 [VOL. 25


mbr.asm

.org/D

ownloaded from



tion but less than the sensitivity of test tubeagglutination according to Thomason, Cherry,and Moody (148). Similarly, Kabanova andGlubokina (86) in a study of unidentified Shigellaspecies, reported greater sensitivity of aggluti-nation reactions than of fluorescent staining.The problem is rendered even more difficult bythe numerous cross reactions among the gram-negative enteric bacilli. Thomason, Cherry, andEdwards (146) report on a number of such crossreactions as observed by direct IF stainingreactions. Thus stains of fecal smears withlabeled antibodies to enteric pathogens yieldunsatisfactory results (68 per cent of normalstool smears and 68 per cent of smears fromtyphoid carriers were positive with labeledtyphoid antibodies). Several noteworthy crossreactions among the enterobacteriaceae aredetailed by the authors. Thomason, Cherry, andEwing (147) investigated the specificity of directIF staining reactions with antisera to entero-pathogenic Escherichia coli. Some cross reactions(three among 60 serotypes of E. coli and sixamong 102 serotypes of other enteric bacilli)were observed with nine labeled antisera. Thus,some false positive staining reactions would beexpected with normal stools, although the numberof organisms stained in specimens from cases ofinfantile diarrhea is much greater, according tothe authors.

13-Hemolytic streptococci are classified intogroups on the basis of precipitin reactions of thepolysaccharides in the cell wall. Cross reactionsbetween groups are not usually observed byprecipitin tests. With fluorescent antibodystaining, however, antibodies to group A poly-saccharide react with group C f3-hemolyticstreptococci. Moody et al. (118) achieved com-plete specificity by adsorbing group A antiserawith group C streptococci. Halperen et al. (74)stained throat smears for group A 3-hemolyticstreptococci. According to this report, 30 of 33smears from culturally proven cases were positive,whereas none of 50 smears from negative casesreacted.

Erysipelothrix insidiosa, the causative agentof hog erysipelas, was subjected to some pre-liminary antibody staining studies. Dacres andGroth (39) observed no cross reactions by directIF staining, whereas J. D. Marshall et al. (109)reported a cross reaction with Bacterium ani-tratum.

Indirect IF staining for Treponema pallidum

with patients' sera holds promise as a diagnostictool (41). Antigens for conventional serologictests for syphilis are mixtures of lecithin, cho-lesterol, and cardiolipids. These antigens are theproduct of several decades of careful research.At present there is a trend toward the develop-ment of tests utilizing T. pallidum as antigen.IF staining offers important advantages here;the method is usually at least as sensitive ascomplement fixation and requires less antigen.A few thousand bacteria per slide can easily bedetected (14). Deacon, Falcon, and Harris(41) compared indirect IF staining for T. pallidumwith T. pallidum complement fixation test,T. pallidum immobilization test, and a conven-tional type test (Venereal Disease ResearchLaboratory) with rabbit and human sera. Im-munofluorescence and complement fixation testswere essentially of equal sensitivity in this study.Other test methods were less sensitive. Thespecificity of the method remains to be deter-mined. The potential value of antibody stainingfor T. pallidum in the serodiagnosis of syphilis isborn out by the reports of Olansky and McCor-mick (128) and Borel and Durel (8a). It seemslikely that this will become the method of choicein the serodiagnosis of syphilis.Most reports on the demonstration of proto-

zoan parasites by immunofluorescence havecome from Goldman's laboratories at the UnitedStates Communicable Disease Center. His earlierstudies on Endamoeba coli and Endamoebahistolytica (64) were reviewed by Coons (24).More recently, Goldman studied Toxoplasmagondii by a method employing inhibition ofstaining (65, 66). A fluorescein conjugate wasprepared from a human serum with a highmethylene blue reduction titer for T. gondii.Sera containing antibody inhibited the specific IFstaining of toxoplasmas. Carver and Goldman(15) described improved methods of preparingtissues infected with T. gondii for IF staining.Although the method holds promise as a diag-nostic tool, the authors observed, "It would beunrealistic to believe that specific staining withfluorescent antibody can solve all problems ofetiologic diagnosis in pathology when Toxoplasmaorganisms are suspected."

Preliminary studies on the specificity of directIF staining reactions for Candida albicans arereported by Gordon (73). All strains of C.albicans could be specifically stained. In addition,all strains of Candida tropicalis and some strains

1961] 61


mbr.asm

.org/D

ownloaded from


ERNST H. BEUTNER

of other Candida species reacted with conjugatesprepared from rabbit antisera to C. albicans.Also 5 of 46 strains of unrelated fungi stained withthese conjugates. All the latter cross reactionsand some of the cross reactions with otherspecies of Candida were eliminated by adsorptionwith Candida parakrusei. The need for carefulspecificity controls on each serum is emphasizedby the difference in the spectrum of cross reac-tions between the two reagents prepared fromrabbit antisera to C. albicans.The potential value of immunofluorescence in

microbiology goes far beyond diagnostic appli-cations. Some incidental observations suggestthe value of the method in cytologic studies.The localization of flagellar and somatic antigensof Salmonella spp. is clearly illustrated in studiesof Thomason, Cherry, and Moody (148). DeRepentigny (42) demonstrated capsular reactionsin pneumococci by direct IF staining. Similarobservations had been made earlier by Coonset al. (30). De Repentigny and Frappier (43) alsoattempted to determine the location of theprotective antigen of Haemophilus pertussis onthe bacilli with labeled antibodies. The potentialvalue of such studies is limited by the fact thatbacteria are near the limit of resolution of thelight microscope.The studies of Hobson and Mann (80) illustrate

another unique potential of immunofluorescentstaining to relate organisms seen on direct smearswith organisms isolated in pure culture from thesame source. It is usually impossible to do thisby other methods since morphology alone isinconclusive. Hobson and Mann describe methodsfor relating organisms isolated from the rumencontent of cattle with organisms isolated in pureculture. As the authors point out, the methodcannot be guaranteed to be entirely specific.However, by suitable adsorption and dilutionprocedures, cross reactions can be minimized.The distribution of injected bacterial antigens

has been subjected to close scrutiny by severalinvestigators (35, 49, 89, 140, 141). Because oftheir medical significance, attention has beenfocused primarily on the antigens of 13-hemolyticstreptococci. In the studies of Schmidt (140, 141)protein and polysaccharide antigens were injectedinto mice. The distribution and persistence ofthese antigens in tissues were studied by directIF staining. Carbohydrate antigens remaineddetectable for only 30 minutes after injection,whereas the M proteins were detectable for up

to 60 hours. Kaplan (89) reported the detectionby indirect IF staining of M proteins of group Astreptococci in the tissues of mice up to 8 daysafter injection. In both studies the antigenremained in the kidney for the longest period oftime. The M protein of a type 12 strain ofstreptococci from an outbreak of pyelonephritispersisted in the kidney for the same length oftime as other types of M proteins. These studieswere carried out with mice treated by intra-peritoneal injection with purified antigens. Asomewhat different distribution of XVI proteinantigen was observed in mice infected withgroup A f3-hemolytic streptococci.

Localization of streptococcal hyaluronidase intissue has aroused interest because antibodies tothis enzyme are found in the serum of patientswith the rheumtic fever and glomerular nephritis.The problem was investigated by injecting con-centrated preparations of this enzyme intraven-ously into mice; frozen sections of mouse organswere stained with a fluorescein-labeled rabbitantiserum to the streptococcal hyaluronidase(Emmart et al., 49). Staining reactions arereported to occur in all organs studied. Factorsinfluencing the elimination of Salmonella typhosaendotoxin from the reticuloendothelial systemwere investigated by Cremer and Watson (35).Fatalities following parenteral administration ofendotoxin to experimental animals were increasedby x-ray irradiation, cortisone, thorotrast, andby two consecutive injections of the endotoxin.Cremer and Watson studied the tissue localiza-tion of the endotoxin by direct IF staining.Small rabbits were given intravenously 2 mg ofthe endotoxin. The authors report the finding ofendotoxin in the reticuloendothelial system inlungs, liver, and spleen. After single injections,the endotoxin was largely eliminated in 12 to 2days. Cortisone or x-irradiation delayed theelimination of endotoxin, whereas thorotrast anddouble injections of the toxin markedly inhibitedits elimination. The studies of Cremer andWatson deserve confirmation and extension.

IV. IMMUNOFLUORESCENT STAINING OF TISSUEANTIGENS

For the purpose of this review, fluorescentantibody studies of tissue antigens are classifiedinto three groups on the basis of the typesof antibody responses involved. Heteroantibodiesas defined here are antibodies of one speciesdeveloped against antigens of another species.

62 [VOL. 25


mbr.asm

.org/D

ownloaded from



Isoantibodies react with tissue antigens of somemembers of the species of the antibody producer,but not with antigens of the antibody-producingindividual. Autoantibodies react with tissueantigens of the antibody-producing individual.

A. Heteroantibodies to Tissue AntigensThe two major types of antigens considered in

this category are (a) biochemically isolated tissueantigens and (b) the classic heterophile, speciesspecific, and organ specific antigens.

Biochemically isolated products of tissue havebeen studied, of which hormones and musclemyosin are two prime examples (36, 52, 53, 84,94, 96, 110, 150). In such studies, the fluorescentantibody method is used essentially as a histo-chemical tool. Although the method is potentiallyof unique specificity, each antiserum must betested for cross reactions. Minor impurities may,and often do, stimulate antibody formation.

J. M. Marshall (110) was the first to reporton this type of study. Using a modified direct IFstaining procedure, he investigated the pituitarylocalization of antibodies to adrenocorticotropichormone (ACTH). Reactions were reported to belocalized in the 13 cells of the anterior pituitary.As Coons indicates in his review (24) of Marshall'spaper, further specificity controls of this reactionare indicated. More recently, Cruickshank andCurrie (36) studied localization of pituitaryhormones. Reactions of most antisera to humanpituitary hormones occurred in the j3 cells of theanterior pituitary. However, staining was notabolished by adsorption with specific antisera.It is noteworthy that they demonstrated exten-sive cross reactions between antisera to humanpituitary ACTH, growth hormone, and thyro-tropic hormone preparations by precipitin tests.The identity of reactions was further demon-strated by adsorption experiments. Similarly,antisera to postmenopausal gonadotropin andthyrotropic hormone preparations extracted fromurine cross reacted completely with each otherbut failed to cross react with hormone prepara-tions made from pituitaries; i.e., antisera tothyrotropic hormone preparations from pituitaryand urine failed to cross react. These studiesillustrate the need for careful specificity controls.Lacy and Davies (96) studied reactions of anti-sera to commercial preparations of insulin (frombeef and pork) by direct IF staining. Islet cells ofbeef, cat, and mouse pancreas sections stained,whereas guinea pig, rabbit, and human islet cells

failed to stain, according to this report. Adsorp-tion of labeled antibodies with pork insulinabolished staining, whereas adsorption withglucagon (a pancreatic hormone with insulin-antagonizing activity) failed to inhibit staining.

Rabbit antisera to myosin prepared from chickembryo skeletal muscle have been studied forseveral years in the laboratories of J. M. Marshall(52, 53, 84, 94, 150). Immunofluorescent reactionsare localized in the A bands of skeletal muscleaccording to Finck, Holtzer, and Marshall'sreport (53). Holtzer, Marshall, and Finck (84)reported on studies of embryonic myogenesis.Myosin was demonstrable by IF staining beforethe development of histologically demonstrablebands. According to Finck and Holtzer's recentstudies (52) of the specificity of antimyosinantibodies, no cross reactions are demonstrablewith chick smooth muscle or other cell typesstudied. Klatzo, Horvath, and Emmart (94)reported on reactions of labeled antibodies tohuman myosin. Aggregation of specific stainingwas reported to occur in atrophic muscle fibersfrom cases of myotonic dystrophy.Another example of the histochemical localiza-

tion of a partially purified tissue antigen by IFstaining is the study of Nairn, Fraser, and Chad-wick (122). They attempted to localize hog reninin hog organs by direct staining. However,antibodies formed to impurities in the hog reninpreparation. Although cross-reacting antibodiescould be removed by adsorption, localizationbecomes quite difficult in such situations. Exten-sive studies on the tissue localization of y-globulinand other serum proteins have been reportedfrom the laboratories of Dixon (45), Mellors(116), Gitlin (61), and others. Some of thesestudies are considered with reports on antibodysynthesis and reactions in vivo. An authoritativediscussion of these papers may be found in therecent review by Mellors (113).

Heterophile, species specific, and organ specificantigens are identified on the basis of classicserologic tests. Heterophile antigens are foundin tissues of many species. Species specificantigens are found in all organs of a given species,and organ specific antigens are found in a givenorgan of many species. Fluorescent antibodystudies extend these concepts to the cellular ortissue component level. The classic example of aheterophile antigen is the Forssman antigen.Tanaka and Leduc (145) reported that Forssmanantigen is found in the form of droplets in the

1961] 63


mbr.asm

.org/D

ownloaded from


ERNST H. BEUTNER

endothelium and adventitial connective tissues ofblood vessels in all organs of guinea pigs, cats,dogs, mice, and chickens. The distribution of thisantigen in other types of cells was reported to bevariable in different species.

Species specific antibodies react with all organsof the given species. Most reports on IF stainingwith species specific antibodies have dealt withkidney antisera. Cruickshank and Hill (37) werethe first to demonstrate the localization of speciesspecific antibodies in the basement membranesof various organs. They used rabbit antisera torat kidney and rat organs as a test system.Roberts (136) reported specific reactions of rabbitantisera to rat glomeruli with the lens capsule ofthe eye as well as with the lining of blood vessels.Hiramoto, Goldstein, and Pressman (78) studiedrabbit antisera to two human tissue culture celllines (HeLa and fetal liver). Stromal elementsand connective tissues of various organs of thebody reacted with these sera by indirect IFstaining. Conversely, Goldstein, Hiramoto, andPressman (68) stained two permanent cell lines,HeLa and 407 liver, with rabbit antisera tohuman thyroid and to human melanoma. Thesespecies specific reactions indicate that the celllines are of connective tissue origin. Hiramoto,et al. (79) confirmed and extended the observa-tions of Cruickshank and Hill (37). Speciesspecific reactions were demonstrated in basementmembranes, capsules, blood vessel walls, andother connective tissue elements. Localizationin vivo and in vitro was similar, but the latterreactions were considerably stronger.

Hill and Cruickshank (76) also demonstratedkidney specific antibodies. A conjugate of anantiserum to whole kidney stained not onlybasement membranes of tubules, glomeruli, andblood vessels (species specific) but also cytoplasmof tubular epithelium (organ specific), whereasconjugates of antisera to rat glomeruli and lungstained only basement membranes and bloodvessels. The kidney specific nature of the epi-thelial reaction was further demonstrated byadsorption experiments. Weiler (157) alsodemonstrated localization of kidney specificantibodies in the cytoplasm of tubular epithelium.Specificity of the reactions was adequatelyconfirmed. In the earlier papers of Louis and hisassociates (93), the authors maintained that thereactions reported by Weiler (156) were non-immunologic in nature. This view was based onthe histologic congruence of their nonimmuno-

logic reactions with the fluorescent antibodyreactions reported by Weiler. However, suchhistologic coincidence of reactions does notnecessarily prove their identity. It is noteworthythat Fisher and Fisher (54) failed to demonstratekidney-localizing antibodies in the sera of dogsthat had rejected homotransplanted kidneys.Humphrey (85) demonstrated the origin of

platelets by IF staining. Labeled rabbit anti-bodies to guinea pig platelets were used to stainguinea pig bone marrow and spleen smears.Only megakaryocytes stained specifically.

B. IsoantibodiesSeveral investigators have studied blood group

isoantibodies by IF staining (1, 19, 63, 144, 159).However, only limited success has been achieved.Heteroantibodies to blood group substances werethe first to yield positive fluorescent antibodyreactions. The distribution of group A substance,H substance, and Lewisa substance in humangastric mucosa was studied by Glyn, Holborow,and Johnson (63) by direct IF staining withconjugates derived from rabbit sera. Specificlocalization and differences in resistance to 90per cent ethanol fixation in different sites aredetailed by the authors. Alexander (1) reportedon agglutination titers and IF staining reactionsof several red cell-agglutinating systems. Negativeresults were reported with labeled isoantibodies,including human anti-A, although the conjugateshad high agglutinating titers. Only certain rabbitantisheep and rabbit antihuman A substanceantisera yielded positive staining reactions.Whitaker et al. (159) were the first to report the

staining of red cells with anti-A and anti-Bisoantibodies. They reported positive IF reactionsonly with cell suspensions pretreated withethylenediaminetetraacetic acid (EDTA). Cohen,Zuelzer, and Evans (19) reported on furtherstudies of red cells with isoantibodies. Strongpositive staining with A and B isoantibodies wasachieved by the direct IF method. Reactions ofA and A, antibodies were distinguishable. Stain-ing reactions were less sensitive than agglutina-tion in detection of blood group isoantibodies.Direct staining, according to this report, was notsufficiently sensitive to detect Rh(D) isoanti-bodies. These were, however, demonstrable byindirect IF staining. The observed reactions wereweak but specific. By these methods, the authorsdemonstrated fetal red blood cells in the maternalcirculation during the last trimester of pregnancy.

64 [VOL. 25


mbr.asm

.org/D

ownloaded from



This important report will undoubtedly bechecked in many laboratories.Szulman (144) studied the distribution of the

A and B blood group antigens throughout thebody. The specific antigens were found in threelocations according to this report: in the cellsand secretions of mucous secreting glands, in thestratified epithelium, and in the vascular endo-thelium.

According to a preliminary report of Danaher,Friou, and Finch (40), specific fluorescent anti-body reactions are demonstrable with leukocyteisoantibodies by indirect IF staining. This is adifficult system to study because of the non-specific staining problems involved.

C. AutoantibodiesBy definition, autoantibodies react with some

tissue component(s) of the individual producingthe antibody. They might be associated withautoimmune diseases. The criteria for establishingthe autoimmune etiology of a disease are setforth in Witebsky's postulates (162): "(1) thedirect demonstration of free, circulating anti-bodies that are active at body temperature or ofcell-bound antibodies by indirect means; (2) therecognition of the specific antigen against whichthis antibody is directed; (3) the production ofantibodies against the same antigen in experi-mental animals; (4) the appearance of pathologi-cal changes in the corresponding tissues of anactively sensitized experimental animal that arebasically similar to those in the human disease."Immunofluorescence is a useful tool for demon-strating the autoantibody nature of serologicreactions, determining the specificity of thereaction, and characterizing the antigen.

Chronic thyroiditis is the best documentedexample of an autoimmune disease (162). Severalfluorescent antibody studies of this disease havebeen reported (6, 7, 77, 81, 161). Figures 4 and 5illustrate the appearance of indirect IF stainingreactions of a normal serum and a chronicthyroiditis serum on sections of a monkeythyroid.White (161) reported localization of thyroiditis

autoantibodies in the colloid of the thyroid andthe glandular epithelium. He also reported ondemonstrations of antibodies against thyroidextracts in the regional lymph nodes. Hiramoto,Engle, and Pressman (77) studied the localizationof autoantibodies in the serum of a patient withchronic thyroiditis with the use of both fluores-

cein- and rhodamine-labeled antiglobulin anti-sera. They reported staining of the thyroidcolloid and the edge of the glandular epithelium.Beutner et al. (7) studied the reactions of rabbitthyroid autoantibodies and rabbit spinal cordautoantibodies by both the direct and indirectfluorescent antibody method. The specificity ofthe reactions was demonstrated by cross reactionstudies with rabbit thyroid and rabbit spinalcord sections. They observed staining of thyroidcolloid and a thin rim of staining at the peripheryof the glandular epithelium with rabbit thyroidautoantibodies. No cross reaction occurred withspinal cord autoantibodies. Further studieswere carried out by the same group (6) on human,rabbit, and dog thyroid autoantisera by theindirect staining method. They report limitedcross reactions between the antisera and thyroidsof other species. The reactions are predominantlyspecies specific as well as organ specific. In allcases, staining reactions occurred in the thyroidcolloid. The specificity of the reaction is amplydocumented by other serologic studies (162).A fluorescent antibody spot test for thyro-

globulin autoantibodies of thyroiditis patientswas reported by Crawford, Wood, and Lessof(34). Fixed smears of human thyroglobulin arestained by the indirect IF method. Reactions areread with a Wood light, i.e., a filtered ultravioletlight. The nucleohistone spot test of Friou (57)for antibodies of lupus erythematosus patients(see below) served as a specificity control. Inthe hands of these investigators (99) the spottest proved to be more sensitive than the passivecutaneous anaphylaxis test of Ovary et al. (131)although this has not been confirmed yet.Holborrow et al. (81) reported another type of

autoantibody reaction with the thyroid, a reactionwith the glandular epithelium. Figures 6 and 7illustrate the appearance of a control and thecellular reaction. It is reported to be associatedwith antibodies which fix complement with amicrosome fraction of the thyroid. Antibodies tothe cells are usually associated with the colloidantibodies of thyroiditis. This cellular stainingreaction was observed independently in thewriter's laboratory. The nature of the diseaseprocesses with which they are associated, if any,is not yet established. The experimental produc-tion of such autoantibodies has not as yet beenreported. Studies of this reaction are complicatedby the affinity of epithelial cells for nonspecificstaining.

1961] 65


mbr.asm

.org/D

ownloaded from


ERNST H. BEUTNER

Figures 4 and 5. Indirect immunofluorescent staining of monkey thyroid sections with human seradiluted 1:10. Figure 4. Sections treated with a control serum (tuberculosis patient). This is a negativereaction. Slight nonspecific staining of cellular elements was observed in the original preparation. Figure5. Section treated with serum of a thyroiditis patient. Immunofluorescent staining of the colloid is ob-served. This is an autoantibody reaction associated with thyroiditis (6, 161, E. Witebsky and E. H.Beutner, unpublished data).

Hargraves, Richmond, and Morton (75a) werethe first to observe antinuclear antibodies in thesera of patients with lupus erythematosus (L.E.)(see Miescher, 117, for review and references).Friou, Finch, and Detre (58, 59) demonstratednuclear staining with L.E. sera by the indirectIF procedure. This reaction occurred with allvertebrate cells studied by these authors. Onlvsera from L.E. cases and a few rheumatoidarthritis sera reacted, whereas all other seratested were negative. These observations werequickly confirmed in other laboratories. Hol-borow, Weir, and Johnson (83) observed IFstaining of nuclei with L.E. sera in severalhuman tissues, including the thyroid. Figures 8and 9 illustrate the appearance of the latterreaction. Bardawil et al. (4) observed nuclearreactions with L.E. sera by both direct and

indirect IF staining. These investigators reportedinhibition of the reaction with the enzyme deoxy-ribonuclease. At approximately the same timeMellors, Ortega, and Holman (116) as well asVazquez and Dixon (153) reported localization ofhuman y-globulin in the nuclei of leukocytes ofL.E. patients. This was demonstrated by directIF staining for human y-globulin. -y-Globulinlocalization was substantiated, but the functionor specificity of the localized y-globulin couldnot be determined from these observations; i.e.,such staining reactions offer only presumptiveevidence for the localization of some antibody ofunknown specificity.

Friou (56) also reported that the extraction ofcell smears with water (presumably distilledwater) or 1 M NaCl removed staining reactionsof the lupus sera tested, whereas extraction with

66 [VOL. 25


mbr.asm

.org/D

ownloaded from



Figures 6 and 7. Sections of monkey thyroid stained as for figures 4 and 5 with human sera by the in-direct immunofluorescent method. Figure 6. Negative reaction of a human serum. Figure 7. Immuno-fluorescent staining of the cytoplasm of the glandular epithelium obtained with serum from a patientwith thyroid disease. This is an autoantibody reaction, distinct from the colloid reaction (see figures 4and 5) but frequently associated with it (81, E. Witebsky and E. H. Beutner, unpublished data).

0.15 M NaCl did not. This he associated with thesolubility characteristics of deoxyribonucleicacid (DNA) histones. He took further advantageof these solubility characteristics to preparesmears of DNA histones on slides for reactionswith fluorescent antibody. Nucleohistone spotsstained by the indirect method served as ascreening test. Of 34 sera that were positive byIF staining, only 20 were positive by the L.E.cell test (phagocytosis of leukocytes in vitro). Itis noteworthy that discrepancies were indepen-dent of titers as determined by IF staining.Recently Goodman et al. (72) reported on thefailure of nuclear antibodies in the S 18 globulinfraction (see below) to give positive L.E. reac-tions. This may account for the lack of correla-tion between L.E. test results and IF stainingreactions observed by Friou (55, 56). Crawford,Wood, and Lessof (34) report on the use of this

spot test. Also, Miescher and I have used it.Friou identified the antigen which participates inthe nuclear staining reaction with antibody asnucleohistone. This contention is supported bythe observation of Holborow and Weir (82).According to their report, nuclear staining isdemonstrable in guinea pig tissues, includingimmature germ cells, but not in the heads ofmature sperms. The authors point out thatnucleohistones are transformed into protamine-like substances in the maturation of sperms.Some doubt is cast on the identity of the antigenby the observation that the enzyme deoxyribo-nuclease inhibits the reaction (4). It may be thatIF staining antibodies are directed against theDNA-protein complex rather than the nucleohis-tones alone. Goodman et al. (72) fractionated serafrom lupus patients on diethylaminoethylcellulose columns. Ten fractions were tested for

19611 67


mbr.asm

.org/D

ownloaded from


ERNST H. BEUTNER

Figures 8 and 9. Indirect IF staining of normal human thyroid sections with human sera as describedfor figures 4 and 5. Figure 8. A negative reaction with a tuberculous serum. Note the nonfluorescentnuclei. Figure 9. Section treated with the serum of patient with lupus erythematosus. Serologicallyspecific staining of the nuclei is observed. This is an autoantibody reaction observed in all cases of lupuserythematosus and occasionally in related diseases (58, 59, 83).

total protein, y-globulin, IF staining, tanned cellhemagglutination, and L.E. cell formation. Twoactive components were identified by both thetanned cell hemagglutination and IF staining.Goodman et al. regard these as the heavy (S 18)and light (S 6.6) globulins. Only the light globulinfraction was active in L.E. cell formation.

In brief, antinuclear autoantibodies have beendemonstrated in the sera of patients with dis-seminated lupus erythematosus. The antigen issome part of the DNA-protein complex. However,autoantibodies which react with nuclei by stain-ing, and lesions characteristic of the disease havenot been produced in experimental animals.Thus the third and fourth postulates have notbeen fulfilled.Kaplan (87) noted that rabbits immunized with

beef heart infusion broth cultures of streptococcideveloped antibodies which react with hearttissue by the fluorescent antibody method. Hiscareful studies of this phenomenon indicatedthat similar antibody responses were obtained inanimals immunized with sterile broth, or with

beef heart, rat heart tissue, or human hearthomogenates. Such antisera react with rabbitheart, including the heart of the rabbit producingthe antibody. Staining reactions occurred betweenmyofibrils and sarcoplasm, according to thisreport. They were paralleled by complement-fixing activity of the sera. The alcohol-solubleantigen involved in these immunohistologicalstaining reactions is distinct from the cardiolipinantigen used in serologic tests for syphilis. Kaplan(88, 90) also demonstrated two types of auto-antibodies to heart muscle in human sera afterheart surgery and in sera of patients with rheu-matic heart disease. Certain pathologic changesin heart muscle which appear after heart surgerymay be associated with these autoantibodies.The first three of Witebsky's postulates are, atleast in part, fulfilled by these outstandingstudies.

Experimental allergic encephalomyelitis isinduced by immunization with brain suspensionsin Freund adjuvant. The disease is characterizedby the development of paralysis, brain lesions,

68 [VOL. 25


mbr.asm

.org/D

ownloaded from



and sometimes complement-fixing autoantibodies.The antigen with which the circulating antibodiesreact is in the alcohol-soluble fraction (163).Immunofluorescent staining with rabbit brainautoantibodies is localized in the myelin sheathof the white matter (7). Subsequent unpublishedobservations revealed staining in the gray matter.This is consistent with the view that submicro-scopic deposits of myelin surround "unmyeli-nated" nerve fibers. In short, circulating anti-bodies react with an alcohol-soluble antigen inthe myelin sheath. However, the nature of theencephalitogenic antigen(s) is a controversialsubject (92). Furthermore, circulating antibodiesplay a minor role in the pathogenesis of allergicencephalomyelitis. Production of characteristiclesions by passive transfer of hypersensitive cellshas been reported (132). If confirmed, the firstof the Witebsky postulates would be fulfilledinsofar as cellular hypersensitivity reactions areconcerned. Thus far, cellular or delayed typehypersensitivity reactions have not been demon-strable by the fluorescent antibody method. Likeother serologic methods, immunofluorescentstaining reveals only circulating type antibodies.

V. ACKNOWLEDGMENTSThe writer is indebted to Dr. Ernest Witebsky

not only for the encouragement which made thisreview possible but also for his guidance in studiesof immunofluorescence and instruction in thebasic principles of immunology. The writer is alsoindebted to Dr. Albert Coons for his personalinstruction in certain aspects of his technique andfor his assistance in some of the initial studies.The writer gratefully acknowledges Dr. Witeb-sky's and Dr. Coons' thoughtful critiques ofthis review.

VI. REFERENCES1. ALEXANDER, W. R. M. 1958 The applica-

tion of the fluorescent antibody techniqueto hemagglutinating systems. Immu-nology, 1, 217-223.

2. ASKONAS, B. A. AND WHITE, R. G. 1956Sites of antibody production in the guinea-pig. The relation between in vivo synthesisof anti-ovalbumin and globulin and dis-tribution of antibody containing plasmacells. Brit. J. Exptl. Pathol., 37, 61-74.

3. BAKER, R. F., RAPP, F., GROGAN, E., ANDGORDON, I. 1957 Visualization of measlesvirus in human cells. Bacteriol. Proc.,76.

4. BARDAWIL, W. A., Toy, B. L., GALINS, N.,AND BAYLES, T. B. 1958 Disseminatedlupus erythematosus, scleroderma anddermatomyositis as manifestations ofsensitization to DNA protein. Am. J.Pathol., 34, 607-629.

5. BAXTER, J. H. AND GOODMAN, H. C. 1956Nephrotoxic serum nephritis in rats. I.Distribution and specificity of the antigenresponsible for the production of nephro-toxic antibodies. J. Exptl. Med., 104,467-485.

6. BEUTNER, E. H., WITEBSKY, E., AND GER-BASI, J. R. 1959 Organ specific reactionsof thyroid autoantisera from rabbits, dogsand humans as determined by the fluor-escent antibody technic. FederationProc., 18, 559.

7. BEUTNER, E. H., WITEBSKY, E., ROSE, N. R.,AND GERBASI, J. R. 1958 Localizationof thyroid and spinal cord autoantibodiesby fluorescent antibody technic. Proc.Soc. Exptl. Biol. Med., 97, 712-716.

8. BOAND, A. V., JR., KEMPF, J. E., ANDHANSON, R. J. 1957 Phagocytosis ofinfluenza virus. I. In vitro observations.J. Immunol., 79, 416-421.

8a. BOREL, L. J. AND DUREL, P. 1959 L'im-muno-fluorescence appliquee au diagnosticde la syphilis. Pathol. Biol., 7, 2317-2324.

9. BOYER, G. S., DENNY, F. W., JR., AND GINS-BERG, H. S. 1959 Intracellular localiza-tion of type 4 adenovirus. II. Cyto-logical and fluorescein-labelled antibodystudies. J. Exptl. Med., 109, 85-96.

10. BUCKLEY, S. M. 1956 Visualization ofpoliomyelitis virus by fluorescent anti-body. Arch. ges. Virusforsch., 6, 388-400.

11. BUCKLEY, S. M. 1957 Cytopathology ofpoliomyelitis virus in tissue culture.Fluorescent antibody and tinctorial stud-ies. Am. J. Pathol., 33, 691-707.

12. BUCKLEY, S. M., WHITNEY, E., AND RAPP,F. 1955 Identification by fluorescentantibody of developmental forms ofpsittacosis virus in tissue culture. Proc.Soc. Exptl. Biol. Med., 90, 226-230.

13. CARSKI, T. R. 1959 Fluorescent antibodystudy of the simian foamy agent. Fed-eration Proc., 18, 561.

14. CARTER, C. H. AND LEISE, J. M. 1958Specific staining of various bacteria witha single fluorescent antiglobulin. J. Bac-teriol., 76, 152-154.

15. CARVER, R. K. AND GOLDMAN, M. 1959Staining Toxoplasma gondii with fluores-cein-labeled antibody. III. The reaction

6919611


mbr.asm

.org/D

ownloaded from


ERNST H. BEUTNER

in frozen and paraffin sections. Am. J.Clin. Pathol., 32, 159-164.

16. CHADWICK, C. W., MCENTEGART, M. G., ANDNAIRN, R. C. 1958 Fluorescent proteintracers. A trial of new fluorochromes andthe development of an alternative tofluorescein. Immunology, 1, 315-327.

17. COCHRANE, C. G., VAZQUEZ, J. J., ANDDIXON, F. J. 1957 The specific localiza-tion of antigen in lesions of experimentalserum sickness. Am. J. Pathol., 33,593-594.

18. COFFIN, D. L. AND LIu, C. 1957 Studieson canine distemper infection by meansof fluorescein-labeled antibody. II. Thepathology and diagnosis of the naturallyoccurring disease in dogs and the antigenicnature of the inclusion body. Virology.3, 132-145.

19. COHEN, F., ZUELZER, W. W., AND EVANS,M. M. 1960 Identification of bloodgroup antigens and minor cell populationsby the fluorescent antibody method.Blood, 15, 884-900.

20. COHEN, S. M., GORDON, I., RAPP, F., MACAU-LAY, J. C., AND BUCKLEY, S. M. 1955Fluorescent antibody and complement-fixation tests of agents isolated in tissueculture from measles patients. Proc. Soc.Exptl. Biol. Med., 90, 118-122.

21. COHN, E. J., GURD, F. R. N., SURGENOR,D. M., BARNES, B. A., BROWN, R. K.,DEROUAUX, G., GILLESPIE, J. M., KAHNT,F. W., LEVER, W. F., LIu, C. H., MITTEL-MAN, D., MOUTON, R. F., SCHMID, K.,AND UROMA, E. 1950 A system for theseparation of the components of humanblood; quantitative procedures for theseparation of the protein components ofhuman plasma. J. Am. Chem. Soc., 72,465-474.

22. COONS, A. H., LEDUC, E. H., AND KAPLAN, M. H.1951 Localization of antigen in tissue cells.VI. The fate of injected foreign proteins inthe mouse. J. Exptl. Med. 93: 173-188.

23. COONS, A. H. 1954 Labeled antigens andantibodies. Ann. Rev. Microbiol., 8,333-352.

24. COONS, A. H. 1956 Histochemistry withlabeled antibody. In International reviewof Cytology, Vol. 5. pp. 1-23. Edited byG. H. Bourne. Academic Press Inc.,New York.

25. COONS, A. H. 1957 The application offluorescent antibodies to the study ofnaturally occurring antigens. Ann. N. Y.Acad. Sci., 69, 658-662.

26. COONS, A. H. 1958 Fluorescent antibody

methods. In General cytochemical methods,Vol. I, pp. 399-422. Edited by J. F. Danielli.Academic Press, Inc., New York.

27. COONS, A. H. 1959 The cytology of anti-body formation. J. Cellular Comp. Phys-iol. (Suppl.), 52, 55-67.

28. COONS, A. H. 1959 Antibodies and anti-gens labelled with fluorescein. Schweiz.Z. Pathol. u. Bakteriol., 22, 693-699.

29. COONS, A. H. 1959 The diagnostic ap-plication of fluorescent antibodies.Schweiz. Z. Pathol. u. Bakteriol., 22,700-723.

30. COONS, A. H., CREECH, H. J., JONES, R. N.,AND BERLINER, E. 1942 The demon-stration of pneumococcal antigens intissues by the use of fluorescent antibody.J. Immunol., 45, 157-170.

31. COONS, A. H. AND KAPLAN, M. H. 1950Localization of antigen in tissue cells.II. Improvements in a method of thedetection of antigen by means of fluores-cent antibody. J. Exptl. Med., 91, 1-13.

32. COONS, A. H., LEDUC, E. H., AND CONNOLLY,J. M. 1955 Studies on antibody pro-duction. I. A method for the histo-chemical demonstration of specific anti-body and its application to a study of thehyperimmune rabbit. J. Exptl. Med., 102,49-60.

33. COONS, A. H., LEDUC, E. H., AND KAPLAN,M. H. 1951 Localization of antigen intissue cells. VI. The fate of injectedforeign protein in the mouse. J. Exptl.Med., 93, 173-188.

34. CRAWFORD, H. J., WOOD, R. M., AND LESSOF,M. H. 1959 Detection of antibodies byfluorescent-spot technique. Lancet, 2,1173-1174.

35. CREMER, N. AND WATSON, D. W. 1957 In-fluence of stress on distribution of endo-toxin in RES determined by fluoresceinantibody technic. Proc. Soc. Exptl. Biol.Med., 95, 510-513.

36. CRUICKSHANK, B. AND CURRIE, A. R. 1958Localization of tissue antigens with thefluorescent antibody technique: applica-tion to human anterior pituitary hormones.Immunology, 1, 13-26.

37. CRUICKSHANK, B. AND HILL, A. G. S. 1953The histochemical identification of aconnective-tissue antigen in the rat. J.Pathol. Bacteriol., 66, 283-289.

38. CURTAIN, C. C. 1958 Electrophoresis offluorescent antibody. Nature, 182, 1305-1306.

39. DACRES, W. G. AND GROTH, A. H., JR. 1959Identification of Erysipelothrix insidiosa

70 [VOL. 25


mbr.asm

.org/D

ownloaded from



with fluorescent antibody. J. Bacteriol.,78, 298-299.

40. DANAHER, T. H., FRIOU, G. J., AND FINCH,S. C. 1957 Fluorescein labeled anti-globulin test applied to leukocyte im-munology. Clin. Research, 5, 9.

41. DEACON, W. E., FALCONE, V. H., AND HARRIS,A. 1957 A fluorescent test for treponemalantibodies. Proc. Soc. Exptl. Biol. Med.,96, 477-480.

42. DEREPENTIGNY, J. 1958 Some limitationsin the use of the fluorescent antibodytechnique. Rev. can. biol., 17, 492-502.

43. DE REPENTIGNY, J. AND FRAPPIER, A. 1956Studies on Hemophilus pertussis liquidcultures. III. Localization of surfaceantigens by means of fluorescent antibody.Can. J. Microbiol., 2, 677-683.

44. DIXON, F. J. AND VAZQUEZ, J. J. 1959 Im-munohistochemical analysis of hyper-sensitivity and related lesions. In Mecha-nisms of hypersensitivity, pp. 191-196.Edited by J. Shaffer. Little, Brown andCo., Boston.

45. DIXON, F. J., VAZQUEZ, J. J., WEIGLE, W. O.,AND COCHRANE, C. G. 1958 Immunologyand pathogenesis of experimental serumsickness. Arch. Pathol., 65, 18-28.

46. DIXON, F. J., WEIGLE, W. O., AND ROBERTS,J. C., JR. 1957 Comparison of antibodyresponses associated with the transfer ofrabbit lymph node, peritoneal exudate, andthymus cells. J. Immunol., 78, 56-62.

47. EATON, M. C., MEIKELJOHN, G., AND VANHERICK, W. 1944 Studies on the etiologyof primary atypical pneumonia, a filterableagent transmissible to cotton rats, hamstersand chick embryos. J. Exptl. Med., 79,649-668.

48. EMMART, E. W. 1958 Observations on theabsorption spectra of fluorescein, fluores-cein derivatives and conjugates. Arch.Biochem. Biophys., 73, 1-8.

49. EMMART, E. W., COLE, R. M., MAY, E. L.,AND LONGLEY, J. B. 1958 Studies onstreptococcal hyaluronidase and anti-hyaluronidase. II. The localization ofsites of absorption of streptoccal hyaluron-idase (Group C) with fluorescent antibody.J. Histochem. and Cytochem., 6, 161-173.

50. ENDERS, J. F. 1954 Recent observationson the behavior in tissue culture of certainviruses pathogenic for man. Natl. Insts.Health Ann. Lectures, 121-135.

51. ENDERS, J. F. 1956 Observations on cer-tain viruses causing exanthematousdiseases in man. Am. J. Med. Sci., 231,622-637.

52. FINCK, H. AND HOLTZER, H. 1960 Animmunochemical comparison of variousmotile systems. Anat. Record, 136, 193-194.

53. FINCK, H., HOLTZER, H., AND MARSHALL,J. M., JR. 1956 An immunochemicalstudy of the distribution of myosin inglycerol-extracted muscle. J. Biophys.Biochem. Cytol. Suppl., 2, 175-178.

54. FISHER, E. R. AND FISHER, B. 1959 In-vestigation of rejection of canine renalhomotransplants by fluorescent antibodytechnic. Proc. Soc. Exptl. Biol. Med.,101, 259-261.

55. FRIOU, G. J. 1958 Clinical application ofa test for lupus globulin-nucleohistoneinteraction using fluorescent antibody.Yale J. of Biol. Med., 31, 40-47.

56. FRIOU, G. J. 1958 Identification of thenuclear component of the interaction oflupus erythematosus globulin and nuclei.J. Immunol., 80, 476-481.

57. FRIOU, G. J. 1958 The significance of thelupus globulin-nucleoprotein reaction.Ann. Internal Med., 49, 866-875.

58. FRIOU, G. J., FINCH, S. C., AND DETRE,K. D. 1957 Nuclear localization of afactor from disseminated lupus serum.Federation Proc., 16, 413.

59. FRIOU, G. J., FINCH, S. C., AND DETRE, K. D.1958 Interaction of nuclei and globulinfrom lupus erythematosus serum demon-strated with fluorescent antibody. J.Immunol., 80, 324-329.

60. GERMUTH, F. G., JR., MAUMENEE, A. E.,PRATT-JOHNSON, J. A., SENTERFIT, L. B.,VAN ARNAM, C. E., AND POLLACK, A. D.1958 Observations on the site and mecha-nisms of antigen-antibody interaction inanaphylatic hypersensitivity. In Mecha-nisms of hypersensitivity,pp. 155-162. Editedby J. Shaffer. Little, Brown and Co.,Boston.

61. GITLIN, D. AND CRAIG, J. M. 1957 Varia-tions in the staining characteristics ofhuman fibrin. Am. J. Pathol., 33, 267-283.

62. GITLIN, D., CRAIG, J. M. AND JANEWAY,C. A. 1957 Studies on the nature offibrinoid in the collagen diseases. Am.J. Pathol., 33, 55-77.

63. GLYNN, L. E., HOLBOROW, E. J., AND JOHN-SON, G. D. 1957 The distribution ofblood-group substances in human gastricand duodenal mucosa. Lancet, 2, 1083-1088.

64. GOLDMAN, M. 1954 Use of fluorescein-tagged antibody to identify cultures of

1961] 71


mbr.asm

.org/D

ownloaded from


ERNST H. BEUTNER

Endamoeba histolytica and Endamoeba coli.Am. J. Hyg., 59, 318-325.

65. GIOLDMAN, M. 1957 Staining Toxoplasmagondii with fluorescein-labelled antibody.I. The reaction in smears of peritonealexudate. J. Exptl. Med., 105, 549-556.

66. GOLDMAN, M. 1957 Staining Toxoplasmagondii with fluorescein-labelled antibody.II. A new serologic test for antibodies totoxoplasma based upon inhibition of spe-cific staining. J. Exptl. Med., 105, 557-573.

67. GOLDMAN, M. AND CARVER, R. K. 1957Preserving fluorescein isocyanate forsimplified preparation of fluorescent anti-body. Science, 126, 839-840.

68. GOLDSTEIN, M. N., HIRAMOTO, R., ANDPRESSMAN, D. 1959 Comparative fluor-escein labeling and cytotoxicity studieswith human cell strains HeLa, Raos, and407-liver and with fresh surgical specimensof cervical carcinoma, osteogenic sarcoma,and normal adult liver. J. Natl. CancerInst., 22, 697-705.

69. GOLDWASSER, R. A. AND KISSLING, R. E.1958 Fluorescent antibody staining ofstreet and fixed rabies virus antigens.Proc. Soc. Exptl. Biol. Med., 98, 219-223.

70. GOLDWASSER, R. A., KISSLING, R. E.,CARSKI, T. R., AND HOSTY, T. S. 1959Fluorescent antibody staining of rabiesvirus antigens in the salivary glands ofrabid animals. Bull. World Health Or-ganization, 20, 579-588.

71. GOLDWASSER, R. A. AND SHEPARD, C. C.1958 Staining of complement and modifi-cations of fluorescent antibody procedures.J. Immunol., 80, 122-131.

72. GOODMAN, H. C., FAHEY, J. L., MALMGREN,R. A., AND BRECHER, G. 1959 Separa-tion of factors in lupus erythematosusserum reacting with components of cellnuclei. Lancet, 2, 382-383.

73. GORDON, M. A. 1958 Differentiation ofyeasts by means of fluorescent antibody.Proc. Soc. Exptl. Biol. Med., 97, 694-698.

74. HALPEREN, S., DONALDSON, P., AND SULKIN,S. E. 1958 Identification of streptococciin bacterial mixtures and clinical specimenswith fluorescent antibody. J. Bacteriol.,76, 223-224.

75. HANSON, R. J., KEMPF, J. E., AND BOAND,A. V., JR. 1957 Phagocytosis of in-fluenza virus. II. Its occurrence innormal and immune mice. J. Immunol.,79, 422-427.

75a. HARGRAVES, M. M., RICHMOND, H., ANDMORTON, R. 1958 Presentation of two

bone marrow elements: the "tart" celland the "L.E." cell. Proc. Staff MeetingsMayo Clinic, 23, 25-28.

76. HILL, A. G. S., AND CRUICKSHANK, B. 1953A study of antigenic components of kid-ney tissue. Brit. J. Exp. Pathol., 34,27-34.

77. HIRAMOTO, R., ENGEL, K., AND PRESSMAN,D. 1958 Tetramethylrhodamine as im-munohistochemical fluorescent label inthe study of chronic thyroiditis. Proc.Soc. Exp. Biol. Med., 97, 611-614.

78. HIRAMOTO, R., GOLDSTEIN, M., AND PRESS-MAN, D. 1958 Reactions of antisera pre-pared against HeLa cells and normal fe-tal liver cells with adult human tissues.Cancer Res., 18, 668-669.

79. HIRAMOTO, R., JURANDOWSKI, J., BERNECKY,J., AND PRESSMAN, D. 1959 Precise zoneof localization of anti-kidney antibody invarious organs. Proc. Soc. Exptl. Biol.Med., 101, 583-586.

80. HOBSON, P. N. AND MANN, S. O. 1957 Somestudies on the identification of rumenbacteria with fluorescent antibodies. J.Gen. Microbiol., 16, 463-471.

81. HOLBOROW, E. J., BROWN, P. C., ROITT,I. M., AND DONIACH, D. 1959 Cyto-plasmic localization of "complement-fixing" auto-antigen in human thyroidepithelium. Brit. J. Exptl. Pathol., 40,583-588.

82. HOLBOROW, E. J. AND WEIR, D. M. 1959Histone: an essential component for thelupus erythematosus antinuclear reaction.Lancet, 1, 809-810.

83. HOLBOROW, E. J., WEIR, D. M., AND JOHN-SON, G. D. 1957 A serum factor in lupuserythematosus with affinity for tissuenuclei. Brit. Med. J., 2, 732-734.

84. HOLTZER, H., MARSHALL, J. M., AND FINCK,H. 1957 An analysis of myogenesis bythe use of fluorescent antimyosin. J.Biophys. Biochem. Cytol., 3, 705-724.

85. HUMPHREY, J. H. 1955 Origin of bloodplatelets. Nature, 176, 38.

86. KABANOVA, E. A. AND GLUBOKINA, A. I.1958 The use of antibodies labelled withfluorescein for the detection of dysenterybacilli-I. Zhur. Mikrobiol. Epidemiol.Immunobiol, 1, 5-8. Reprinted in J. Micro-biol., Epidemiol. Immunobiol., 29, 3-6.

87. KAPLAN, M. H. 1958 Immunologic studiesof heart tissue-I. Production in rabbitsof antibodies reactive with an autologousmyocardial antigen following immuniza-tion with heterologous heart tissue. J.Immunol., 80, 254-267.

72 [VOL. 25


mbr.asm

.org/D

ownloaded from



88. KAPLAN, M. H. 1958 Immunologic studiesof heart tissue. II. Differentiation of amyocardial sarcoplasmic antigen andcardiolipin. J. Immunol., 80, 268-277.

89. KAPLAN, M. H. 1958 Localization of strep-tococcal antigens in tissues. I. Histologicdistribution and persistence of M proteinstypes 1, 5, 12 and 19 in the tissues of themouse. J. Exptl. Med., 107, 341-352.

90. KAPLAN, M. H. 1959 Autoantibodies toheart tissue in the sera of certain patientswith rheumatic fever. Federation Proc.,18, 576.

91. KAY, C. F. 1942 The mechanism of a formof glomerulonephritis. Nephrotoxic neph-ritis in rabbits. Am. J. Med. Sci., 204,483-490.

92. KIES, M. W. AND ALVORD, E. C. (editors).1959 In Allergic encephalomyelitis, Chap.8-13. Charles C Thomas, Publisher, Spring-field, Illinois.

93. KING, E. S. J., HUGHES, P. E., AND Louis,C. J. 1958 The species non-specificityof globulins in the globulins-fluoresceinstaining of tissues. Brit. J. Cancer, 12,5-13.

94. KLATZO, I., HORVATH, B., AND EMMART,E. W. 1958 Demonstration of myosinin human striated muscle by fluorescentantibody. Proc. Soc. Exptl. Biol. Med.,97, 135-140.

95. KLEIN, P. AND BURKHOLDER, P. 1959 DieDarstellung von fixiertem Komplement mitmarkiertem Antikomplement. Schweiz.Z. Pathol. u. Bakteriol., 22, 729-731.

96. LACY, P. E. AND DAVIES, J. 1957 Prelim-inary studies on the demonstration ofinsulin in the islets by the fluorescentantibody technic. Diabetes, 6, 354-357.

97. LEBRUN, J. 1956 Cellular localization ofherpes simplex virus by means of fluores-cent antibody. Virology, 2, 496-510.

98. LEBRUN, J. 1957 L'antigene poliomyeliti-que au cours du developpement intracellu-laire du virus. Ann. inst. Pasteur, 93,225-229.

99. LESSOF, M. H., CRAWFORD, H. J., AND WOOD,R. M. 1959 Antibody to thyroglobulinin patients with thyroid disease. Lancet,2, 1172-1173.

100. Liu, C. 1956 Rapid diagnosis of humaninfluenza infection from nasal smears bymeans of fluorescein-labeled antibody.Proc. Soc. Exptl. Biol. Med., 92, 883-887.

101. Liu, C. 1957 Studies on primary atypicalpneumonia. I. Localization, isolationand cultivation of a virus in chick embryos.J. Exptl. Med., 106, 455-466.

102. Liu, C. AND COFFIN, D. L. 1957 Studieson canine distemper infection by meansof fluorescein-labeled antibody. I. Thepathogenesis, pathology, and diagnosisof the disease in experimentally infectedferrets. Virology, 3, 115-131.

103. Liu, C. AND EATON, M. D. 1955 Study andisolation of primary atypical pneumoniavirus in chick embryos by means of fluores-cein-labelled antibody. Bacteriol. Proc.,61.

104. Liu, C., EATON, M. D., AND HEYL, J. T.1959 Studies on primary atypical pneu-monia. II. Observations concerning thedevelopment and immunological charac-teristics of antibody in patients. J.Exptl. Med., 109, 545-556.

105. Liu, C. AND HEYL, J. T. 1957 Serologicalstudy of primary atypical pneumonia.Federation Proc., 16, 423.

106. Louis, C. J. 1958 The significance of thecell type in the fluorescein-globulin stain-ing of tissues. Brit. J. Cancer, 12, 537-546.

107. MCKINNON, G. E., ANDREWS, E. C., JR.,HEPTINSTALL, R. H., AND GERMUTH, F. G.,JR. 1957 An immunohistologic study onthe occurrence of intravascular antigen-antibody precipitation and its role inanaphylaxis in the rabbit. Bull. JohnsHopkins Hosp., 101, 258-280.

108. MARSHALL, J. D., EVELAND, W. C., ANDSMITH, C. W. 1958 Superiority of fluores--cein isothiocyanate (Riggs) for fluorescent-antibody technic with a modification ofits application. Proc. Soc. Exptl. Biol.Med., 98, 898-900.

109. MARSHALL, J. D., EVELAND, W. C., SMITH,C. W., AND HARR, J. R. 1959 Fluorescentantibody studies of Erysipelothrix insi-diosa. Bacteriol. Proc., 90.

110. MARSHALL, J. M., JR. 1951 Localization ofadrenocorticotropic hormone by histo-chemical and immunochemical methods.J. Exptl. Med., 94, 21-29.

111. MAYERSBACH, H. 1959 Unspecific inter-actions between serum and tissue sectionsin the fluorescent-antibody technic fortracing antigens in tissues. J. Histochem.and Cytochem., 7, 427.

112. MAYERSBACH, H. AND PEARSE, A. G. E.1956 The metabolism of fluorescein-labelled and unlabelled egg-white in therenal tubules of the mouse. Brit. J.Exptl. Pathol., 37, 81-89.

113. MELLORS, R. C. (editor). 1959 Fluo-rescent-antibody method. In Analyticalcytology, pp. 1-67. McGraw-Hill Book Co.New York.

731961]


mbr.asm

.org/D

ownloaded from


ERNST H. BEUTNER

114. MELLORS, R. C., ARIAS-STELLA, J., SIEGEL,M., AND PRESSMAN, D. 1955 Analyticalpathology. II. Histopathologic demonstra-tion of glomerular-localizing antibodies inexperimental glomerulonephritis. Am. J.Pathol., 31, 687-715.

115. MELLORS, R. C. AND ORTEGA, L. G. 1956Analytical pathology. III. New observa-tions on the pathogenesis of glomerulo-nephritis, lipid nephrosis, periarteritisnodosa, and secondary amyloidosis in man.Am. J. Pathol., 32, 455-499.

116. MELLORS, R. C., ORTEGA, L. G., AND HOL-MAN, H. R. 1957 Role of gamma globu-lins in the pathogenesis of renal lesions insystemic lupus erythematosus and chronicmembranous glomerulonephritis, with anobservation on the lupus erythematosuscell reaction. J. Exptl. Med., 106, 191-202

117. MIESCHER, P. (editor). 1957 Die Immuno-pathologie in Klinik und Forschung, pp. 79-84. Georg Thiema Verlag, Stuttgart,Germany.

118. MOODY, M. D., ELLIS, E. C., AND UPDYKE,E. L. 1958 Staining bacterial smearswith fluorescent antibody. IV. Groupingstreptococci with fluorescent antibody.J. Bacteriol., 75, 553-560.

119. MOODY, M. D., GOLDMAN, M., AND THOM-ASON, B. M. 1956 Staining bacterialsmears with fluorescent antibody. I. Gen-eral methods for Malleomyces pseudomallei.J. Bacteriol., 72, 357-361.

120. MOODY, M. D., THOMASON, B. M., WINTER,C. C., AND HALL, A. D. 1959 Sensitivityof bacterial agglutination and fluorescentantibody reactions. Bacteriol. Proc., 89.

121. MOULTON, J. E. 1956 Fluorescent antibodystudies of demyelination in canine dis-temper. Proc. Soc. Exptl. Biol. Med., 91,460-464.

122. NAIRN, R. C., FRASER, K. B., AND CHADWICK,C. S. 1959 The histological localizationof renin with fluorescent antibody. Brit.J. Exptl. Pathol., 40, 155-163.

123. NEIL, A. L. AND DIXON, F. J. 1959 Im-munohistochemical detection of antibodyin cell-transfer studies. A.M.A. Arch.Pathol., 67, 643-649.

124. NOYES, W. F. AND WATSON, B. K. 1955Studies on the increase of vaccine virus incultured human cells by means of the fluo-rescent antibody technique. J. Exptl.Med., 102, 237-242.

125. NOYES, W. F. 1955 Visualization of Egypt101 virus in the mouse's brain and in cul-tured human carcinoma cells by means of

fluorescent antibody. J. Exptl. Med., 102,243-248.

126. NOYES, W. F. AND MELLORS, R. C. 1957Fluorescent antibody detection of theantigens of the Shope papilloma virus inpapillomas of the wild and domestic rab-bit. J. Exptl. Med., 106, 555-562.

127. OHTA, G., COHEN, S., ZAK, F. G., AND SINGER,E. J. 1959 Immunohistochemical studiesof spleen and liver. Federation Proc., 18,497.

128. OLANSKY, S. AND MCCORMICK, G. E. 1960Experiences with the fluorescent trepo-nemal antibody test for syphilis. A.M.A.Arch. Dermatol., 81, 59-66.

129. ORTEGA, L. G. AND MELLORS, R. C. 1956Analytical pathology. IV. The role oflocalizing antibodies in the pathogenesis ofnephrotoxic nephritis in the rat. A fluo-rescent antibody study. J. Exptl. Med.,104, 151-170

130. ORTEGA, L. G. AND MELLORS, R. C. 1957Cellular sites of formation of gammaglobulin. J. Exptl. Med., 106, 627-640.

131. OVARY, Z., RANDALL, H., WITEBSKY, E.,ROSE, N., SHULMAN, S., AND METZGAR,R. 1958 Thyroid-specific autoantibodiesstudied by passive cutaneous anaphylaxisof guinea pig. Proc. Soc. Exptl. Biol.Med., 99, 397-401.

132. PATERSON, P. Y. 1959 Transfer of allergicencephalomyelitis in rats by means ofimmune lymph cells. Federation Proc., 18,591.

133. PRINCE, A. M. AND GINSBERG, H. S. 1957Immunohistochemical studies on the inter-action between Ehrlich ascites tumor cellsand Newcastle disease virus. J. Exptl.Med., 105, 177-188.

134. RAPP, F., SELIGMAN, S. J., JAROSS, L. B.,AND GORDON, I. 1959 Quantitative de-termination of infectious units of measlesvirus by counts of immunofluorescent foci.Proc. Soc. Exptl. Biol. Med., 101, 289-294.

135. RIGGS, J. L., SEIWALD, R. J., BURCKHALTER,J. H., DOWNS, C. M., AND METCALF, T. G.1958 Isothiocyanate compounds as fluo-rescent labeling agents for immune serum.Am. J. Pathol, 34, 1081-1097.

136. ROBERTS, D. S. C. 1957 Studies on theantigenic structure of the eye using thefluorescent antibody technique. Brit. J.Ophthalmol., 41, 338-347.

137. ROBERTS, J. C., JR. AND DIXON, F. J. 1955The transfer of lymph node cells in thestudy of the immune response to foreignproteins. J. Exptl. Med., 102, 379-392.

138. ROBERTS, J. C., JR., DIXON, F. J., AND

74 [VOL. 25


mbr.asm

.org/D

ownloaded from


1IMMUNOFLUORESCENT STAINING

WEIGLE, W. 0. 1957 Antibody-producinglymph node cells and peritoneal exudatecells. A.M.A. Arch. Pathol., 64, 324-332.

139. SCHILLER, A. A., SCHAYER, R. W., ANDHEss, E. L. 1953 Fluorescein-conjugatedbovine albumin. Physical and biologicalproperties. J. Gen. Physiol., 36, 489-505.

140. SCHMIDT, W. C. 1952 Group A strepto-coccus polysaccharide: studies on itspreparation, chemical composition, andcellular localization after intravenous in-jection into mice. J. Exptl. Med., 95, 105-119.

141. SCHMIDT, W. C. 1954 The cellular localiza-tion of injected streptococcal antigens.In Streptococcal infections, p. 87. Edited byC. M. McCarty. Columbia UniversityPress, New York.

142. SEEGAL, B. C. 1958 In vivo localization ofspecific antiorgan sera: relation to occur-rence of renal lesions. In Mechanisms ofhypersensitivity, pp. 143-154. Edited by J.Shaffer. Little, Brown and Co., Boston.

143. SMITH, C. W., MARSHALL, J. D., AND EvE-LAND, W. C. 1959 Investigation of Lis-teria monocytogenes by the fluorescentantibody technique. Bacteriol. Proc., 91.

144. SZULMAN, A. E. 1959 Histochemical dis-tribution of blood group antigens A and Bin human tissues. Federation Proc., 18,509.

145. TANAKA, N. AND LEDUC, E. H. 1956 Astudy of the cellular distribution of Forss-man antigen in various species. J. Im-munol., 77, 198-212.

146. THOMASON, B. M., CHERRY, W. F., ANDEDWARDS, P. R. 1959 Staining bacterialsmears with fluorescent antibody. VI.Identification of Salmonellae in fecalspecimens. J. Bacteriol., 77, 478-486.

147. THOMASON, B. M., CHERRY, W. B., ANDEWING, W. H. 1959 Rapid identificationof serotypes of Escherichia coli with poly-valent antibody solutions labeled withfluorescein. Bacteriol. Proc., 90.

148. THOMASON, B. M., CHERRY, W. B., ANDMOODY, M. D. 1957 Staining bacterialsmears with fluorescent antibody. III.Antigenic analysis of Salmonella typhosa bymeans of fluorescent antibody and aggluti-nation reactions. J. Bacteriol., 74, 525-532.

149. THOMASON, B. M., MOODY, M. D., ANDGOLDMAN, M. 1956 Staining bacterialsmears with fluorescent antibody. II.Rapid detection of varying numbers ofMalleomyces pseudomallei in contaminated

materials and infected animals. J. Bac-teriol., 72, 362-367.

150. TUNIK, B. D. AND HOLTZER, H. 1960 Local-ization of muscle antigens in contractedsarcomeres by means of fluorescein labelledantibodies. Anat. Record, 136, 294-295.

151. UEHLEKE, H. 1958 Neue Moglichkeiten zurHerstellung fluoreszenzmarkierter Pro-teine. Z. Naturforsch., 13B, 722-724.

152. UEHLEKE, H. 1959 Untersuchungen mitfluoreszenz-markierten Antikorpern. IV.Die Markierung von Antikorpern mitSulfochloriden fluoreszierender Farbstoffe.Schweiz. Z. Pathol. u. Bakteriol., 22, 724-729.

153. VAzQUEz, J. J. AND DIXON, F. J. 1957Immunohistochemical study of lesions inrheumatic fever, systemic lupus erythema-tosus and rheumatoid arthritis. Lab.Invest., 6, 205-217.

154. WATSON, B. K. 1956 Mixed infections withinfluenza and mumps viruses in the chickembryo. Bull. N. Y. Acad. Med., 32, 169.

155. WEBER, G. 1952 Polarization of the fluo-rescence of macromolecules. II. Fluorescentconjugates of ovalbumin and bovine serumalbumin. Biochem. J., 51, 155-167.

156. WEILER, E. 1956 Die Xnderung derserologischen Spezifitat von Leberzellender Ratte wahrend der Cancerogenesedurch p-Dimethylaminoazobenzol. Z.Naturforsch., llB, 31-38.

157. WEILER, E. 1956 Antigenic differences be-tween normal hamster kidney and stil-boestrol-induced kidney carcinoma. Histo-logical demonstration by means offluorescing antibodies. Brit. J. Cancer,10, 560-563.

158. WELLER, T. H. AND COONS, A. H. 1954Fluorescent antibody studies with agentsof varicella and herpes zoster propagatedin vitro. Proc. Soc. Exptl. Biol. Med., 86,789-794.

159. WHITAKER, J., ZUELZER, W. W., ROBINSON,A. R., AND EVANS, M. 1959 The use ofthe fluorescent antibody technique for thedemonstration of erythrocyte antigens.J. Lab. Clin. Med., 54, 282-283.

160. WHITE, R. G. 1954 Observations on theformation and nature of Russell bodies.Brit. J. Exptl. Pathol., 35, 365-370.

161. WHITE, R. G. 1957 An immunologicalinvestigation of Hashimoto's disease.Proc. Roy. Soc. Med., 50, 953-954.

162. WITEBSKY, E., ROSE, N. R., TERPLAN, K.,PAINE, J. R., AND EGAN, R. W. 1957Chronic thyroiditis and autoimmuniza-tion. J. Am. Med. Assoc., 164, 1439-1447.

751961]


mbr.asm

.org/D

ownloaded from


ERNST H. BEUTNER

163. WITEBSKY, E. AND STEINFELD, J. 1928Untersuchungen uber Spezifische Antigen-funktionen von Organen. Z. Immun-forsch., 58, 271-296. See also chap. 14 inreference 92.

164. WITMER, R. H. 1955 Antibody formationin rabbit eye studied with fluorescein-

labeled antibody. A.M.A. Arch. Ophthal-mol., 53, 811-816.

165. WOOD, C. AND WHITE, R. G. 1956 Experi-mental glomerulo-nephritis produced inmice by subcutaneous injections of heat-killed Proteus mirabilis. Brit. J. Exptl.Pathol., 37, 49-60.

76 [VOL. 25


mbr.asm

.org/D

ownloaded from


IMMUNOFLUORESCENT STAINING: FLUORESCENT ANTIBODYstaining. Klein and Burkholder (95) also report...

Documents

Transcript of IMMUNOFLUORESCENT STAINING: FLUORESCENT ANTIBODYstaining. Klein and Burkholder (95) also report...