Immunologic Studies of Autochthonous Cancer an Evaluation of Several Procedures*

Immunologic Studies of Autochthonous Cancer An Evaluation of Several Procedures*

ABRAHAM G. OSLER

(Department of Microbiology, The Johns tlol~ldns Univers4ly School of Medicine and School q[ Hygiene and Public Health, and the Department of Medicine, The Johns Ilopkins University School of Medicine, Baltimore, Md.)

INTRODUCTION sembling acquired immune tolerance or immuno- The unique and complex problems in studies of logic paralysis intervenes. spontaneous tumors of animals derive from the ~. Tumor-specific antibody is produced but is fact that the limiting frame of reference comprises not amenable to extensive in vitro investigation, a host response to an invasive, replicating cell pre- because it is of the cell-dependent type, which sumably of autologous origin. Unlike the situation mediates delayed hypersensitivity reactions, or of in many infectious diseases, immunologic phe- the reaginie, skin-sensitizing variety. nomena attributable to reactants of exogenous 3. Circulating tumor-specific antibody is syn- origin are not considered admissible in studies of thesized but is not uniformly detectable in the autochthonous cancer. As in the area of auto-antiserum because it is bound to the cells which incite body production, this limiting condition imposes its production. rigorous requirements for the immunologist. Evi- 4. Antibody is synthesized by the tumor cells dence for an immune response to spontaneous tu- and is directed against host tissue components. mors must be buttressed by the demonstration of 5. Circulating tumor-specific antibody is pro- antigenic activity which clearly differentiates the dueed which reacts extensively with a variety of neoplasm from normal host tissues. normal tissue components. The recurrent reports and disavowals of spe- The reInainder of this report will be based on cific antibody production to autologous cancer tis- the assumption that circulating antibody, en- sues have led to the consideration of several hy- dowed with varying degrees of tumor specificity, is potheses. They are listed, not in the sense of being present in sera of humans or animals bearing spon- all-inclusive, and without further comment as to taneous tumors. Several experimental applications their validity, since a critical evaluation has not to the problem will be discussed primarily in terms yet been undertaken and the probability exists of the interpretational restrictions imposed by the that a single mechanism may not be equally ten- methodology. able for all situations. A. COMPLEMENT FIXATION IN ANTI- 1. Tumor cell multiplication progresses in the GEN CHARACTERIZATION STUDIES absence of antibody production. a) The genetic changes which characterize neo- 1. ThE FIvE-UNIT METHOD IN STUDIES plastic cells are not reflected in antigenic differ- OF TUMOR-SPECIFIC ANTIGENS ences, morphologic and functional disparities not- The complement (C') fixation reaction has been withstanding. used extensively for the immunologic characteriza- b) Tumor cells are deficient in antigenic com- tion of a tissue lipide with potential tumor spec- ponents present in normal tissues. ificity (1~, 44). It may therefore be appropriate to c) Tumor cells proliferate only when a state re- consider the applicability of current C' fixation * The studies carried out in the author's laboratory have procedures to this problem. The investigator who been supported by the National Science Foundation and the contemplates this approach faces the choice be- National Institute of Allergy and Infectious Diseases, United tween one of two general types of C ' fixation tech- Public Health Service. Financial assistance was also available nics: the conventional method which utilizes five from the Office of The Surgeon General, Department of the Army, under the auspices of the Commission on Cutaneous 50 per cent units of C' (C'Hs0) such as described in Diseases of the Armed Forces Epidemiological Board. (38) and the so-called quantitative technic with 50 1187

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 1961 American Association for Cancer Research. 1188 Cancer Research Vol. ~1, October 1961 or more CtH~0. Certain more general aspects of the the reaction may be localized between the limits of two procedures have been evaluated previously 4.5 to 5.5 units, since the range of partial lysis re- (~0, 33). For the present purposes, it may be ap- quires 0.5 to 1.5 C'Hs0. propriate to discuss other features with specific c) Complete lysis.--Less than 4.5 CtH~0 were applicational reference to the estimation and char- fixed. acterization of lipide-soluble tissue extracts from It is thus readily apparent that the use of 6 normal or cancerous tissues. C'Hs0 leads to a quantitative estimate only within The technic used by Rapport and Graf in their a narrow range of C' action--namely, one C'Hs0. studies of tumor lipides is essentially that devel- The remainder of the C' serves to swamp various oped by Wadsworth, Maltaner, and Maltaner (54) anticomplementary effects which may be very and also studied extensively by Rice (~7, 48). The troublesome in tests with less than 5 or 6 C'Hs0. A results are expressed in terms of the quantity of consequence of this severe restriction in the range antigen or antibody required to yield 50 per cent of C' action is that heavy reliance must necessarily hemolysis in a C' fixation reaction with 6 C~Hs0, be placed upon the differences in test results with and a constant amount of antiserum or antigen in the twofold dilution sequences of antigen or anti- the presence of varying quantities of the other. body. It follows, then, that an antigenic variable which is expressed by less than a 50 or 100 per cent difference could be attributable to "experimental Z error," which is of the same order of magnitude in / n... these reactions. i- I / z . I / / Ld / A simple illustration may clarify this point. The identification by Pangborn of cardiolipin as the o I / I / major antigenic constituent in crude lipide ex- ~a- I / I // tracts containing the Wassermann antigen (39, 40) '~\ /// w led to the isolation of a similar phospholipide from wheat germ, called sitolipin (5~). Comparative TYPE I serum titer estimations with these two phospho-

I I I I lipides by conventional complement fixation and I 2 3 4 flocculation procedures supported the notion that RELATIVE ANTIGEN CONCENTRATION they were identical (46). 1 When quantitative CmaRT 1.--Two general types of C ~fixation reaction curves C' fixation technics with 50 C'Hs0 were applied to depicting the relative amounts of antigen and antibody re- this problem it was readily discernible that the quired to yield 50 per cent hemolysis (as adapted from [44]). phospholipide of plant origin was indeed similar but was not identical to those extracted from These, then, are two-dimensional block titrations, mammalian tissues. Sitolipin proved less efficient and the results are presented as curves drawn than the beef or human heart eardiolipins in pre- through the levels of immune reactants yielding 50 cipitating the Wassermann antibody or in C' fixa- per cent hemolysis as observed or interpolated tion studies with this antibody as derived from a values. The types of reaction curves that may be variety of sources including sera of human syphilit- obtained are given in Chart 1 as modified from ics (36, 37). (44). The data in Chart 1, which were obtained with The limitations of this method in studies of anti- the 6-unit method, describe the C' fixation results gen characterization may be summarized as fol- in the three different zones of antigen-antibody lows. interaction. The vertical line essentially parallel to When 6 units of C' are used in the initial fixation the ordinate roughly describes the region of anti- reaction, the subsequent addition of sensitized body" excess, becoming curved at equivalence zone erythroeytes to the antigen-antibody-C' mixture, ratios and asymptotic to the abscissa in the region for the lyric estimations, can lead to one of three of antigen excess. At antigen excess ratios, a di- results and implications, assuming that all the vergence may occur in that some immune systems controls are in good order. a) No lysis.--This finding means that at least exhibit more pronounced inhibition with excess 5.5 of the initial 6 C~Hs0 were consumed in reacting antigen (Type II) than do others (Type I). It is with the antigen, the antibody, or the specific ag- apparent from an inspection of these curves that gregate, since 0.5 C'Hs0 produces less than 10 per major segments of each type of curve contribute cent hemolysis. A. G. Osier, J. H. Strauss, and B. Lowenstein, unpublished b) Partial lysis.--The number of C'Hs0 fixed in observations.

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 1961 American Association for Cancer Research. OSLER--Immunologic Studies of A utochthonous Cancer 1189 little to the characterization of the antigen-anti- diolipin in aqueous solution can combine with body reaction under study. With a Type I curve, Wassermann antibody but does not fix C' in this marked variations in antigen or antibody concen- interaction unless cholesterol is added. The choles- trations produce no detectable change in the C' terol is considered to facilitate aggregation by pro- fixation pattern. It is only in the restricted cur- viding a particulate surface for the adsorption of vilinear portion of this curve that alteration in the cardiolipin, whose unit molecular weight approxi- amount of either antigen or antibody will seriously mates e000 or less. The tendency for cholesterol affect the outcome of the reaction. These consider- crystals to aggregate spontaneously in isotonic ations may account for the wide variations encoun- saline must also be controlled with judicious addi- tered in those experiments which "did not permit tion of lecithin in carefully calibrated quantities. the quantity (of antigen) to be assessed accu- This cardiolipin-cholesterol-lecithin complex may rately" despite "corrections for the fixability of C' then be used in studies of antigenic activity based in terms of the deviation of the standard antigen on aggregation with antibody. Cardiolipin alone from its assigned value" (45). In the case of Type will combine with its specific antibody and act as a II curves, the region of antigen excess depicted by hapten inhibitor. When new lipides with antigenic the right-hand linear portion can be applied to problems of antigen estimation. Unfortunately, NO. C'Hso IN TEST the Type II pattern does not describe all immune 4.5I// 6 ,5"" 24 / systems, and the factors which lead to either of 12- these two reaction patterns are not entirely clear. The inference to be drawn from these comments E IO- is that this C' fixation system provides a relatively d insensitive instrument for differentiating varia- 8- tions in antiserum or antigen activity when the in- crements are less than 50 per cent. As noted in references (~0) and (33), these difficulties emerge z from the design of this type of C' fixation procedure which utilizes a low level of C' and thereby circumscribes the limits within which only relatively narrow ranges of antigen and antibody interaction can be studied. The data shown in Chart i i i I i 4 8 12 16 20 e, which reproduce a portion of those in Figure 14 ANTIGEN ADOED, JJg.N of reference (44), corroborate this statement. It may be seen that, as the initial amount of C' is in- Cr~AR~ 9.--Influence of the level of available C' on the variations in antigen and antibody required to yield 50 per creased from 4.5 to s C'Hs0, each curve exhibits cent hemolysis (as adapted from [44]). greater discriminatory potential in terms of the range of antigen or antibody concentrations which activity like cytoplipin H are brought to investiga- alter the test result. tion, estimations of tumor specificity can be sub- Some of the advantages which accrue from the ject to meaningful interpretation only when each use of a relative excess of C' such as ~4, 50, or 100 tissue extract is carefully standardized with re- C'Hs0 in C' fixation studies designed to elucidate spect to the requirements for adjunct lipides. This certain aspects of antigen-antibody interaction in requirement renders studies of this type increas- highly dilute systems have been discussed pre- ingly tedious and complex, and even when fulfilled viously (33). The following section of the article may be applicable only to lecithin and cholesterol. will therefore consider additional applications par- The use of newer chromatographic methods for ticularly as they may be referable to immunologic lipide separation should be helpful in providing characterization of tissue antigens. more highly purified products so that the influence of lipide-lipide interaction on antigenic behavior ~. ThE QUANTITATIVEMETHOD may be further clarified. The experiments with a) Studies with lipide antigens.--One problem of cardiolipin now to be outlined may serve as one crucial importance with respect to lipide-soluble example of the applicability of C' fixation analyses antigens needs restatement. It has been amply to problems of organ and species specificity. Two demonstrated in (~6) that the specific antigenic ac- technical problems required resolution. The first tivity of these substances varies dramatically with dealt with a supply of purified cardiolipin prepara- the presence and concentration of auxiliary lipides, tions which were extracted from cardiac tissue of themselves lacking reactivity with antibody. Car- bovine and human origin, and from human liver

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 1961 American Association for Cancer Research. 1190 Cancer Research Vol. ~1, October 1961 according to the method of Pangborn (40). The range of antigen concentrations that varied from activity of these preparations with lecithin and 0.07 to 11.0 t~g. In view of the relative excess of cholesterol was determined in preliminary trials C t, all regions of the antigen-antibody reaction and preceded the type of experiment such as de- were available for comparative studies. scribed in Chart 3 and in (87). It may be noted It became clear from these studies that beef from these data that the concentration of Wasser- heart, human heart, and human liver cardiolipins mann antibody is expressed in weight units. The were immunochemically indistinguishable, thus procedure for obtaining these values with lipide providing a rational basis for the so-called non- antigens was developed for this study and pro- specific Wassermann test. Oil the assumption that vided an independent means for confirming the re- the Wassermann antibody represents an immune sults of the C' fixation experiments. As indicated response to tissue rather than to Treponema pal- in the chart, two sets of reactions were set up, one lidum, it is clear that beef heart may therefore with varying levels of beef heart cardiolipin, the serve as an excellent and more readily available other with the human preparation, both reacting substitute for the human tissues in routine tests. with rabbit antibody to the beef heart phospho- b) Antigen estimation by means of C' fixation.- lipide. The antibody N levels obtained with this The data in Chart 3 also provide a sensitive and specific basis for estimating the cardiolipin content of different tissues, normal and neoplastic. As may FIXATICN OF C' BY RABBITANT~-BEEF HEART C,ARDIOLIPIN (055~g antibody N) WITH BEEF AND HUMAN HEART CARDIOUPINS be noted, microgram fractions of cardiolipin can FI XATION - 37~ 90 ram. WITH 50 C' HSO be assayed by means of a reference serum, with an accuracy of about 10 per cent. Further, some of the 20- BEEF HEART CARDIOLIPIN difficulties encountered in studies of lipide antigens i ...... 9 j/~ __ ...... 111 may be more amenable to resolution with an antigen assay system of this type through the utiliza- tion of inhibition studies with those lipides that fail to aggregate in the absence of cholesterol. As applied to cardiolipin, crude lipide extracts of vari- 5 ous tissues freed of cholesterol can be assayed for this phospholipide in terms of inhibition of C ~fixation with the cardiolipin-cholesterol-lecithin com- 0,25 1.0 20 50 4D II.0 plex. pg CARDIOLIPIN ADDED Levine and his colleagues have described extensive applications of the quantitative C' fixation CaART 3.--Each symbol (&O) indicates tile values obtained in one experiment with intestinal strips from a single method to the problem of antigen estimation and guinea pig. characterization. Among these may be mentioned the human low density lipoproteins (~8), a heat- serum were ~7.4 gg. for the homologous bovine labile a-~-glyeoprotein (4~), and the comparative preparation and ~8.3 for the human heart eardio- antigenic activities of proteins and deoxyribonu- lipin. The C' studies were carried out at 37 ~ C., cleic acid extracts obtained from ruptured T4 bac- with 1 nil. of a serum dilution calculated to yield teriophage (4, ~, ~4, 31). The enhanced C'-fixing 0.55 gg. of antibody N on the ~7.4 gg. basis. Ex- activity of thermally denatured DNA as compared cept for the appropriate controls, each tube in the with that in the unheated nucleic acid is also de- C' fixation experiment contained the same level of scribed. More recent studies with a mieromethod antibody and 50 C'Hs0, an amount chosen so as to have also been developed recently. An inter- leave a considerable excess. Of interest in this ex- esting application of the quantitative method periment are the identical values obtained with the in delineating those portions of a protein molecule two cardiolipins in the rising portion of the curves concerned with enzymatic and antigenic activity comprising the antibody excess and equivalence may be found in the review by Brown et al. (5). zone regions. A slight disparity, never exceeding Evidence is presented in this paper that disulfide more than 8 C'H~0, is apparent with increasing bonds play an important but quantitatively dif- quantities of antigen, attaining peak values of 19 ferent role in the enzymatic and antigenic activity and 17 for the beef and human eardiolipins, respec- of ribonuelease. When the protein was treated with tively. These differences barely exceed the experi- thioglycollate at pH 8.5 and iodoaeetic acid was mental error. With some sera, a reversal in rank added subsequently to prevent reoxidation of the was noted. This experiment, like others with many sulfhydryl residues, the derivative fixed half of the different immune systems, was carried out with a C' but possessed only 5 per cent of the initial enzy-

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 1961 American Association for Cancer Research. OsLER--Immunologic Studies of A utochthonous Cancer 1191 mic activity. The data as shown in Chart 4 (taken nomenon or immune fluorescence with cell nuclei, from that paper) allow for the further inference a nucleoprotein-~,-globulin complex is formed that the partial reactivity of the derivative cannot which is not disrupted by three cycles of washing be due to a mixture of inert and completely active and resuspension. These washings yield a final enzyme. Were this the case, the addition of this supernate containing less than 0.05 #g. of -/-globu- mixture in quantities greater than 0.5 gg. N would lin N. An aliquot of this nueleoprotein-~/-globulin fix as much C' as did the native ribonueleases. complex dissolved in alkaline buffer is added to a Since this was not observed, and on the basis of predetermined quantity (3.0 #g.) of rabbit anti- chemical analyses, it was concluded that two or human ~/-globulin N. In this fashion, the ~/-globu- more disulfide bonds are essential for complete lin bound to the nueleoprotein fixes C' in combina- antigenic activity. tion with its specific antibody in the rabbit serum. c) Quantitative estimates of the nucleoprotein-reac- The number of C'Hs0 fixed is then converted to give "y-globulin in lupus.--In a recent, as yet unpub- weight units of ~/-globulin on the basis of a calibra- lished study, the quantitative C t fixation proce- tion curve such as shown in Chart 5. To heighten dure has been used to estimate the amount of 3'- the specificity of this reaction for human ~,-globu- globulin in the sera of patients with systemic lupus erythematosus (S.L.E.) and other collagen dis- FIXATION OF GUINEA PIG C' AT 4~ eases, capable of combining with calf thymus nu- Antibody= 3.0 pg. Ro-Anti Human ~Globulin N. cleoprotein. The essential principle involves a C' Antigen= Human c~ Globulin fixation assay for human ~/-globulin as antigen 80- with specific rabbit antiserum as the antibody. In 7o~ Af j-j effect, these determinations can be regarded as providing reproducible estimates of lupus antibody 60- / / X o50- 40- 6O -o

30-

20- 4o %X 10-

i i ~ , , , i 1 b O. I 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Jag. ~/GLOBULIN NITROGEN

CH.at~T 5.v-Each symbol (A O l) indicates the values obtained in one experiment with intestinal strips from a single guinea pig.

0102 05 I0 2.0 lin, the rabbit antiserum was absorbed with human pg.ANTIGEN N a- and r human serum albumin, and C~T 4.--Fixation of C' with native RNase (A) and par- guinea pig serum proteins. The calibration curve, tially reduced RNase blocked by idioaeetie acid (taken from [51). which is duplicated with each experiment, was prepared with a lyophilized preparation of human under the anomalous circumstances in which the y-globulin. Preliminary studies with the eleetro- antigens have not yet been completely identified phoretieally separated "/1 and "r2 constituents of (51). this product showed little difference in the C' fix- This device is based on the well known fact that ing activity of these two proteins with the pooled sera from patients with S.L.E. react with a variety antisera used as the source of antibody. of preparations derived from cell nuclei, as shown This procedure has indicated that normal hu- by Miescher and Str~issle (~9), Holman, Kunkel man sara possess less than 4.0 #g. of nueleoprotein- and their colleagues (7, 18), and many others. A reactive ~,-globulin N/ml, whereas the levels in standardized suspension of insoluble nucleoprotein S.L.E. generally range from 9 to 80. In individual particles, prepared according to (18), is reacted patients, these levels vary with elinieal activity, with an appropriate dilution of the patient's being low in remissions which occur as a conse- serum. In the event that the serum contains reac- quence of steroid therapy. It has also been ob- tive y-globulin, such as will elicit the L.E. cell phe- served in patients with S.L.E. that serum C' levels

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 1961 American Association for Cancer Research. 1192 Cancer Research Vol. ~1, October 1961 vary inversely with the quantity of this 3,-globulin tion of tumor-specific antigens as distinguished demonstrable in the same specimen. This finding from normal host tissue constituents. In addition, provides a basis for further exploration of the there is the crucial question of selecting, from mechanism of tissue damage in S.L.E. These among the complex of antigens, the one or more studies confirm others in the demonstration that which are required for tumor cell viability. the fraction of nucleoprotein-reactive 7-globulin The parallelisms which may be drawn to analo- which reacts with purified preparations of calf gous studies of bacterial and viral antigenic com- thymus deoxynucleic acid represents only a minor position are many and obvious. A single illustra- portion of the total antibody response in some pa- tive example may be selected. The antibody re- tients and is entirely absent in others. sponse to an infection with a Type 1~ hemolytic C' fixation procedures of this type may be ap- streptococcus of Lancefield's Group A can be esti- plicable to studies of other auto-immune diseases mated by many methods. However, these meas- and possibly to investigations of spontaneous can- urements provide no insight as to the immune cer to the extent that the latter may be considered status of the host with respect to the specific M as an immunologic disease of autologous origin. protein characteristic of the Type 1~ organism. Thus, it would be of more than academic interest The amount of antibody produced to this antigen to determine whether neoplastic tissues or cells may be critical insofar as resistance to Type lfl contain larger quantities of bound ~-globulin than streptococcal infection is concerned, but probably their normal counterparts, when both are obtained comprises only a minor segment of the total im- from the same host. mune response to the many intra- and extracellu- lar antigens of this organism. B. CYTOTOXIC ACTION OF ANTIBODY It follows, then, that application of C' fixation, AND C' ON TUMOR CELLS hemagglutination, gel diffusion methods, etc., for The in vitro neutralization reaction represents studies of cancer immunology can be used for esti- another important method for studying the hu- mating the over-all immune response and in the moral antibody response in spontaneous cancer. enumeration of tile number of contributory anti- For these experiments, tumor cells are suspended gen-antibody systems. These procedures may also in a fluid reaction medium with varying dilutions be valuable as adjuncts in the selection of the anti- of antiserum and fresh guinea pig or other serum. gens which participate in the cytotoxic reactions. The effect of these additives on the viability of the However, identification of the latter will undoubt- tumor cells may be assessed by animal inoculation, edly constitute a major advance as a necessary" or by a variety of in vitro procedures such as de- prelude to rational efforts at immunotherapy. This scribed recently by Goldberg and Green (11, 13- information could conceivably lead to the produc- 15) and by Winn (57, 58) among others. Roizman tion of specific antibody less subject, perhaps, to and Roane 2 have recently devised an interesting the complex of cross-reactions with normal tissues. method for differentiation of viable from injured In view of the potential importance of this ap- or killed cells. Human epidermoid carcinoma cells proach, two aspects of the cytotoxic reactions in (H.Ep. #~) are infected with Herpes simplex virus vitro may be discussed briefly. and then treated with fresh guinea pig serum and Generally, the antisera used for these experi- rabbit antiserum to the H.Ep. ~ cells. Those tis- ments are obtained from tumor-bearing hosts (~, sue culture cells rendered nonviable through the 8, 9, 57, etc.) or following the injection of a heter- immune cytolytic process do not support replica- ologous animal (e.g., rabbit) with intact cells or tion of the virus and consequently fail to form their fractions (11, 13-15). The latter approach is plaques. The diminution of the plaque count as occasionally complicated by the fact that the anti- compared with the controls provides a quantita- body response of individual animals which follows tive estimate of the activity of the anti-H. Ep. #~ the injection of an antigenic mixture is highly vari- cell antibody. able with respect to any single component. This Two types of problems emerge in a considera- difficulty is often circumvented by absorption of tion of the many interesting studies in this direc- the immune serum with antigens other than those t.ion which bear on the role of C'. The first and under study and is based on the assumption that probably most important to the problem of cancer there is no immunologic overlap between the two. immunology relates to the detection and estima- The serum titers finally obtained after these ab- sorptions are often too low for definitive studies. B. Roizman and P. R. Roane, Jr. Quantitative Enumera- Difficulties of this nature have been encountered tion of Cells Killed by Antibody and Complement Based on Failure of Virus Multiplication in Non-Viable Cells, personal in attempts to produce antisera specific for indi- communication. vidual strains of Sahnonella. The magnitude of

this problem may be gauged from the report of assay system will consist of the following reagents, Habel et al. (16). In considering this line of ap- each of which must be carefully standardized for proach, it should be re-emphasized that the use of optimal activity. These are the tumor cells, guinea antigen dilution titers to estimate the potency of pig serum or other source of C', antibody, and an an antiserum or its degree of crossreactivity does adequate level of necessary co-factors, such as not adequately characterize the immune system divalent cations. To minimize the likelihood that under analysis unless two-dimensional titrations antibodies present in the serum used as a source of are also performed for comparative purposes. In C / might potentiate the cytotoxic activity of the studies with well characterized protein or poly- immune serum, the former should be absorbed saccharide antigens as well as purified poliovirus, with the cells under test prior to use. The require- it is usually observed that the antigen titer remains ment for divalent cations in the action C' 1 and relatively constant despite wide variations in anti- C' ~t has been amply demonstrated (~0, 33). It body levels, as may also be seen in (50, 59) (cf. also might, therefore, be anticipated that the enhance- reference [4]). It has been further shown that, in ment of antibody toxicity for lymphoma cells as the region of antigen excess, heterologous antigens reported in (57) could be demonstrated even more may yield higher C' fixation titers than those ob- effectively had the diluent contained these cations tained in the homologous reaction. These cannot at optimal levels. The presence of citrate or phos- be interpreted in any valid fashion unless quanti- phate in the buffer used as diluent in these studies tative considerations are noted and applied (35). may have led to still further difficulties when fresh In this light, antigen titers to one tissue prepara- guinea pig serum was added in varying amounts as tion following absorption with seemingly unre- a replacement for the buffer. Under these condi- lated crude tissue extracts of undetermined anti- tions, the cation levels were not uniform in the dif- genic content cannot be evaluated in the absence ferent tubes of the titration series, thus possibly of data pertaining to the respective serum anti- obscuring sharp end-point determinations. Verifi- body levels. cation of this effect can be established by hemolyt- An approach which might circumvent some of ic assays of the reaction mixtures, which should these technical obstacles would rely on the capac- comprise an essential control for this type of study. ity of tumor cell fractions to inhibit the cytotoxic The activity of the antibody-C' system in destroy- effect of C' and antibody. The combined use of in ing tumor cells should be checked for residual C' vitro measurements of cytopathic effects with ani- by means of hemolytic titrations. Assurance mal inoculation experiments might also expedite would thus be provided that the neutralization identification of specific cancer cell antigens of end point was not curtailed because of a C' deficit transmissible tumors, essential for cell survival in the assay reaction mixtures. and multiplication. Many investigators have relied solely on the dif- ferential action of fresh and heated sera for C' C. IDENTIFICATION OF C' identification. The inadequacy of this criterion Many in vitro studies of cytotoxicity have dealt may be emphasized in the statement that some with the role of serum C', and some of these have antibodies to tissue and other antigens, especially been discussed in (38). The facile assumption that those present in so-called normal sera, are also de- enhancement of the test effect by fresh guinea pig, stroyed in large measure by the thermal inactiva- rabbit, human, or other serum operates through tion conditions (56 ~ C., 30-rain.) used to destroy the C' system may be misleading in view of find- the heat-labile C' components, C I 1 and C' ~ (~7, ings such as those by Amano, Inoue, and their co- 36, 37). Specific decomplementation--i.e., removal workers (1, 19). These investigators concluded of C' components by absorption with a specific that the amount of lysozyme present in serum eon- precipitate formed by an unrelated immune sys- tributes in a striking fashion to spheroplast forma- tem-constitutes an additional experimental ma- tion of Gram-negative microorganisms. Metzger et neuver of value to this type of study provided al. have also reported that lysozyme enhances the titrations establish the simultaneous loss of hemo- immobilization of Treponema pallidum by anti- lyric and neutralizing activity. body and C1 (r The recent demonstrations that salicylaldoxime It may therefore be pertinent to mention some and phlorizin act as specific inhibitors of one of the of the criteria to be fulfilled in the identification of third C' components provide still another means C' as a participant in these reactions. On the prem- for verifying the assumption of C' participation ise that the C ~ requirements in these experiments (30, 49). Finally, in systems which utilize the will parallel those for immune hemolysis as dis- mouse as the test animal for cell viability, correla- cussed by Mayer in this symposium, the cytotoxie tion of in vitro with in vivo effects of C' should not

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 1961 American Association for Cancer Research. 1194 Cancer Research Vol. ~1, October 1961 prove a serious deterrent, since the total lyric ac- of normal tissues. The studies of Waksman et al. tivity for sensitized erythrocytes in the mouse is (55) and of Pearson (41) clearly demonstrate that probably equivalent to 0.1 ml. of undiluted guinea Freund's adjuvant is not an inert mixture. In rats, pig serum. In consequence, the addition of guinea a disseminated inflammatory reaction affecting pig serum may be expected to lead to marked in- many tissues may be produced following injection creases in neutralizing potency as demonstrated in of killed tubercle bacilli emulsified in oil. These (57) provided anticomplementary effects do not animals also showed marked skin reactions to tu- intervene and the evidence for a C' requirement is berculin which were attributed to dissemination of otherwise valid. The current availability of im- mycobacterial antigens. In view of these findings, proved methods for preparation and assay of indi- future studies of auto-antibodies in cancer might vidual C' components will greatly facilitate studies include immunization with incomplete adjuvant along these lines in providing additional criteria and hemagglutination studies with the patients' for unequivocal identification of C' in terms of well sera after absorption with normal and neoplastic characterized intermediates of immune hemolysis tissues, as well as with heat-killed tubercle bacilli, (see article by Ma~er, this volume and chapter on if these were present in the inoculation material. C' in [e0]). The incorporation of hyaluronidase, penicillin, and streptomycin introduces further interpretational D. OTHER STUDIES difficulties in that at least two of these prepara- Finney and co-workers have recently renewed tions contain several antigenic components of po- the claim that the serum of patients with terminal tential significance to the hemagglutination reac- malignancy may contain cytotoxic antibodies to tion. their own tumors (8, 9). The antibody response The observations drawn in these studies (8, 9) was estimated by hemagglutinin titrations with that an increase of hemagglutinating antibody tanned sheep erythroeytes. Although the technical may also follow deep x-ray therapy would seem to details of antigen preparation for coating the red vitiate the objections noted in the active immuni- cells are not given, it may be assumed that these zation experiments. A detailed confirmatory report were homogenized extracts of autologous tumor would provide a firmer basis for judgment. How- tissue. Similar preparations, derived from the pa- ever, the possibility must still be entertained that tient's own tumor tissue and incorporated in com- extensive irradiation and accompanying tissue plete Freund's adjuvants, were used for active im- damage may effectively release an intracellular munization. Many of the patients also received component which normally is but feebly antigenic, hyaluronidase in the injection material in order to if at all. Reference is made to cardiolipin which has minimize the formation of debilitating abscesses. been demonstrated in human tissues by means of There are several aspects of these studies which fluorescent antibody studies (~1) and which is require further consideration before these experi- markedly antigenic only in syphilis. The cellular ments can provide definitive evidence for auto- contituents reactive with sera in lupus provide antibody production in human cancer another such illustration. Finally, the possibility The data of the present experiments indicate must be eliminated that the hemagglutination re- enhanced hemagglutination activity following in- action involved a nontumor cell antigen that may jection of the patient with extracts of his own tu- have been adsorbed to the sheep erythrocytes mor tissue. The tumor tissue used for inoculation from the homogenized tissue extracts. This type of undoubtedly contained an undefined mixture of contamination has been noted in the case of bac- antigenic substances from normal and neoplastic terial agglutination due to the adsorption of lyso- tissues. The assumption that the serologic reaction zyme onto the surface of bacteria grown in broth was mediated only by the latter cannot be ac- enriched with serum. Antibody to lysozyme ag- cepted without further evidence. Controlled stud- glutinated tile microorganisms (3). ies with various normal tissues derived from the Among other reports ascribing antigenic spec- same or other patients are indicated to provide a ificity to tumors on the basis of anaphylactic ex- basis for estimating organ and individual spec- periments are the studies of Makari (~5) and of ificity and to preclude the interplay of viral or Zil'ber and his colleagues (53, 60). In the procedure bacterial antigens in these experiments. We have followed by Makari, female guinea pigs are given observed that rabbits given injections of complete injections of an extract of tumor tissue. Several Freund's adjuvant may respond with the produc- weeks later, the uterine horns of this and of a non- tion of Wassermann antibody. Further, sera of pa- immunized animal are placed in the same Schultz- tients with a variety of nonmalignant diseases can Dale bath and tested for reactivity with normal react with alcoholic (84) and aqueous extracts (10) serum. Following this treatment, the serum of the

patient from whom the tmnor was derived or an In an effort to surmount some of the technical extract at the tumor is added to the bath fluid. In barriers associated with this procedure, the writer the author's hands a greater contraction of the and his colleagues have developed an isometric ap- uterine muscle from the immunized pig was ob- paratus for the study of antigen-antibody interac- served in more than 95 per cent of cancer patients. tion in guinea pig smooth muscle. Technical de- The potential utility of this approach, confirmed tails of the procedure are given in (3r 43). The by Burrows (6) and denied by Hackett and Gar- method utilizes 30-40 segments of guinea pig donyi (17), invites critical appraisal. Many ques- ileum, thus providing ample opportunity for repli- tions of a theoretical nature can be raised concern- cation of tests and for evading the problem of indi- ing the validity of these findings with respect to vidual variability. Measurements of reactivity to the untold numbers of antigens present in the histamine, serotonin, and to immune reagents with crude tissue extracts, the requirement for the

presence of the tumor-specific antigen in the pa- l LATENT PERIOD OFTHE RESPONSE OF SENSITIZED tient's serum, the failure to control many aspects / GUINEA PIG SMALL INTESTINE TO INCREASING of the procedure, the possible involvement of CONCENTRATIONS OF HOMOLOGOUSANTIGEN (Ea-Ro-onti-Ea blood group and plasma protein antigens, and the difficulties encountered by the author in dealing with autologous material.

r i RESPONSE OF SENSITIZED GUINEA PIG SMALL INTESTINE I00~ TO INCREASING CONCENTRATIONS OF HOMOLOGOUS i , ANTIGEN (Ea t Ra-anti Eo) [] 50- A 90-4 I []

804

7oJ E

6o~ Sensitization with Ra-a-Ea N,250#g/Kg Challenge- 18 hours later I0- 5o~ Summary of 4 experiments ,.,. , Each point refers to a mean of at least 6 replicates

3o~ 0.125- 0,25 d 5 1.0 ' i .0 8 i0 16.0 ' "32.0 ~~ with Ra-a-Ea N ,250jJglKg jJg EQ N /~/ Challenge - 18 hours later zoq 9 Summary of 4 experiments t / Each point refers to a mean of at least CHART 7.--Each symbol (A A 9I) indicates the values Io" // 6 replicates obtained in one experiment with intestinal strips from a single guinea pig. i i i, ~ i i 0325 0.25 0.5 I0 2.0 4.0 8.0 sensitized tissues yield a standard error of the ,ug Ea N mean approximating 10 per cent. Data typifying CHA[~;'r 6.--Each symbol (A Jk 9il) indicates the values the response of the passively sensitized guinea pig obtained in one experiment with intestinal strips from a single ileum to increasing levels of homologous antigen guinea pig. (egg albumin) are given in Chart 6, in which the A serious technical di~eulty which has plagued results of four experiments are smnmarized. The many students of hypersensitivity concerns the tracings which provided these data were also ana- lyzed for the duration of the latent, period which uneontrolled variability of the smooth muscle con- begins with the addition of antigen to the bath and traction when measured by means of the isotonic which terminates with contraction of the tissue. procedure used in the Sehultz-Dale type of reac- These data, given in Chart 7, show the inverse tion. This perhaps explains the necessity for re- linear relationship which characterizes the reaction lianee upon contractions of ~-5 ram. as an index of in the region of partial response. With further in- a positive result in some of the studies reported in crements of antigen, the latent period approaches (3). In view of the long-standing difl3eulties with a minimum of about 10-15 seconds for this im- this type of apparatus, the validity of these find- mune system. It was of interest to note that the ings might be subject to confirmation in a "double minimal latent period for histamine is about 1 see- blind" type of study. and, whereas that for anaphylatoxin, a physiologi-

cal mediator of histamine release, is abou~ o sec- Immune Gamma Globulin and Complement on Krebs onds. Further exploration of the factors which Ascites Tumor Cells. I. Ultrastructural Studies. J. Exp. YIed., 109: 505-10, 1959. influence the duration of this initial phase of 1~. GRAF, L., and RAPPORT, M. M. Immunochemical Studies smooth muscle contraction, and of the events at of Organ and Tumor Lipides. VII. The Reactivity of Anti- the cellular level, are required. The data show that Human Tumor Sera with Cytolipin H, Cardiolipin, and careful measurements of this interlude provide an Forssman Haptens. Cancer Research, 20:546-50, 1960 (and earlier papers). additional and sensitive estimate of the response 13. GREEN, II.; BARROW, P.; and GOLDBERG, B. Effect of An- of sensitized tissue to specific antigen. tibody and Complement on Permeability Control in It may be anticipated that the degree of repro- Ascites Tumor Cells and Erythrocytes. J. Exp. Med., 110: ducibility and replication attainable with this as- 699-713, 1959. say system can be helpful in the identification of 14. GREEN, H.; FLEISCHER, R. A.; BARROW, P.; and GOLD- BERG, B. The Cytotoxic Action of Immune Gamma Globu- tumor-specific antigens. Since the method permits lin and Complement on Krebs Ascites Tumor Cells. the use of 30 or more intestinal segments from a I[. Chemical Studies. J. Exp. Med., 109:511-~1, 1959. single guinea pig and requires minute quantities of 15. GREEN, H., and GOLDBERG, B. The Action of Antibody reagents, it becomes possible to carry out all the and Complement on Mammalian Cells. Ann. New York necessary control experiments and to replicate the Acad. Sci., 87:35~-61, 1960. 16. HABEL, K.; HORNIBROOK, J. W.; and GREGG, N. C. Cyto- observations with the tissues of a single animal. toxic Effects of Antisera against Human Epithelial Cells This degree of flexibility in the experimental design Grown in Tissue Culture. Ann. New York Acad. Sc., 69: should prove valuable in attempts to confirm the 801-3, 1957. studies reported by Zil'ber on systemic anaphy- 17. IIACKETT, E., and GARDONYI, E. Serum Detection of Car- laxis in guinea pigs (53, 60) In these experiments, cinoma Experience with the Schultz-Dale Technique. Brit. M. J., 1:1785-87, 1960. it would be of interest to compare the results ob- 18. HOLMAN, H. R., and DEICHER, H. R. The Reaction of the tained with neoplastic and normal tissues of L. E. Cell Factor with Deoxyribonucleoprotein of the Cell autologous origin with those reported in (53) where Nuclei. J. Clin. Investigation, 38:~059-72, 1959. homologous extracts of tumors and normal tissue 19. INOUE, K. ; TANIGAWA, W. ; TAKUBO, M. ; SATANI, M. ; and AMANO, W. The Role of Lysozyme in Immune Bacteriol- were employed. ysis. Biken's J., 2:1-~0, ]959. O-0. KARAT, E. A., and MAYER, M. M. Experimental hnmuno- REFERENCES chemistry. O-d ed. Springfield, Illinois: Charles C Thomas, 1. AMANO, W.; FUJIKAVCA, K.; MORIOKA, T.; MIYAMA, A.; 1961. and ICHIKAWA, S. Quantitative Studies on the Immune ~1. KAPLAN, M. H.; MEYESEmAN, M.; and KUSIfNER, I. Im- Bacteriolysis. 2. A New Method of Quantitative Estima- munologic Studies of Heart Tissue. IV. Serologic Reactions tion. BikeR's J., 1:13-O-5, 1958. with Human Heart Tissue as Revealed by hnmunofluores- O-. AMOS, D. B. Possible Relationships between the Cytotoxic cent Methods: Isoimmune, Wassermann, and Autoim- Effects of Isoantibody and Host Cell Function. Ann. New mune Reactions. J. Exp. Med., 113:17-36, 1961. York Acad. Sci., 87: ~73-90, 1960. O-O-. LEVINE, L. ; BARLOW, J. L.; and VAN VUNAKIS, H. An In- 3. BAUMANN-GRAcE,J. B., and TOMCSIK, J. Fixation of Cer- ternal Protein in T ~- and T4 Bacteriophages. Virology, 6: tain Heterogeneous Antigenic Substances on Bacterial 700--17, 1958. Cells and Endospores. Experientia, 15:305, 1959. 0~3. LEVINE, L.; KAVFFMAN, D. L.; and BROVCN, R. K. The An- 4. BROWN, R. K.; DELANEY, R.; LEWNE, L.; and VAN VVNA- tigenic Similarity of Human Low Density Lipoproteins. KIS, H. Studies on the Antigenic Structure of Ribonuclease. J. Exp. Med., 102:105-18, 1955. J. Biol. Chem., 234: O-043-49, 1959. O-4. LEVINE, L.; MURAKAMI, W. W.; VAN VUNAKIS, H.; and 5. BROWN, R. K.; DUmEVX, J.; DELANEu R.; LEIKIIIM, E.; L. GROSS~tAN. Specific Antibodies to Thermally Denatured and CLARK, B. J. Immunochemical Reactions Involving Deoxyribonucleic Acid of Phage T4. Proc. Nat. Acad. Sci., Ribonuclease. Ann. New York Acad. Sci., 81 : 5O-4-39, 1959. 46:1038-43, 1960. 6. BVRROWS, D. Schultz-Dale Test for Detection of Specific O-5. MAKARI, J. G. Detection of Antigens in Sera of Patients Antigen in Sera of Patients with Carcinoma. Brit. M. ,l., with Neoplastic Disease by Schultz-Dale Test. Brit. Med. :t : ~68-70, 1958. J., 2:858-61, 1958. 7. DmCHER, H. R. G.; HOLMAN, H. R. ; and KVNKEL, H. G. O-6. ]~/[ALTANER,E., and MALTANER, F. The Standardization of The Precipitin Reaction between DNA and a Serum Factor the Cardiolipin-Lecithin-Cholesterol Antigen in the Com- in Systemic Lupus Erythematosus. J. Exp. Med., 109:97- plement Fixation Test for Syphilis. J. Immunol., 51:195- 114, 1959. O-14, 1945. 8. FINNEY, J. W.; BYERS, E. H.; and WILSON, R. H. Studies O-7. ~IAURER, P. H., and THORPE, R. M. The Effect of Heat on in Tumor Autoimmunity. Cancer Research, 20:351-57, the Immunochemical and Biologic Properties of Rabbit 1960. Antisera. J. Immunol., 84:318-O-4, 1960. 9. FINNEY, J. W.; WILSON, R. H.; and BYERS, E. H. Studies 28. METZGER, M.; HARDY, P. H., JR.; and NELL, E. E. Influ- in Tumor Autoimmunity. J. Lab. & Clin. Med., 54:814, ence of Lysozyme upon the Treponeme Immobilization 1959. Reaction. Am. J. Hyg., 73:~36-44, 1961. 10. GAJDUSEK,D. C. An "Autoimmune" Reaction against Hu- 29. MIESCIIER, P., and STnXSSLE, R. The Pathogenesis of man Tissue Antigens in Certain Acute and Chronic Dis- Visceral Lupus Erythematosus as Reflected in the Nero- eases. I. Serological Investigations. Arch. Int. Med., 101: Reactions. Immunopathology, 1st International Sympo- 9-29, 1958. sium, pp. 454-6O-. Basle, Switzerland: Benno Schwabc & 11. GOLDBERG, B., and GREEN, II. The Cytotoxic Action of Co., 1959.

30. MILLS, S. E., and LEVINE, L. The Inhibition of Immune 46. REIN, C. R.; KELCEC, L. C.; and ROSENFIELD, T. M. The Haemolysis by Salicylaldoxime. Immunology, 2:868-88, Role of Sitolipin in Serologic Tests for Syphilis. Venereal 1959. Diseases Symposium, Washington, D.C., 1951. ,~1. MURAKAMI, W. T.; VAN VUNAKIS, H.; and LEVINE, L. 47. RICE, C. E. Some Factors Influencing the Selection of a Synthesis of T~ Internal Protein in Infected Escherichia Complement Fixation Method. I. A Comparison of Two coli, Strain B. Virology, 9:624-35, 1959. Quantitative Technics and an Alternative Method of Ex- 32. NEU, H. C.; RANDALL, H. G.; and OSLER, A. G. Studies on pressing Serum Dilution Titer. J. Immunol., 59:95-107, the Mechanism of Hypersensitivity Phenomena. V. Anti- 1948. gen-Antibody Interaction in the Guinea Pig Small Intes- 48. ------. Complement Fixation Tests of Multiple Anti- tine. Immunology (in press). gen-Antibody Systems. I. Difficulties in the Interpretation ~,~. OSLER, A. G. Quantitative Studies of Complement Fixa- of Diagnostic Tests. Canad. J. Comp. Med., 23:112-15, tion. Back Rev., 22:246-66, 1958. 163-71, 360-68, 1959. 34. OSLER, A. G.; ItAnDY, P. H.; and S~ARP, J. T. The Fixa- 49. RODRIGUEZ, E., and OsL~m, A. G. Studies on the Inhibi- tion of Complement by Human Sera and Alcoholic Ex- tion of Immune Hemolysis by Phlorizin. J. Immuno]., 85: tracts of Human Cardiac Tissue. Am. J. Syphilis, Gonor- 347-55, 1960. rhea, Vetmreal Diseases, 38: 55~-66, 1954. 35. OSLE~, A. G., and HEIDELBEnGEn, M. Quantitative 50. RosE, N. R.; METZGAH, R. S.; and WITEBSKY, E. Studies Studies of Complement Fixation. III. Homologous and on Organ Specificity. XI. Isoantigens of Pancreas Extracts. Cross-Reactions in Pneumococcal Type III and Type VIII J. Immunol., 85: 575-87, 1960. Systems. J. Immunol., 60:317-25, 1948. 51. TOWNES, A. S.; STEWART, C. R., JR.; and OSLER, A. G. 36. OSLER, A. G., and KNIPP, E. A. Quantitative Studies of Quantitative Estimates of tim Nucleoprotein-Reactive Lipid-Soluble Tissue Antigens as Exemplified by the Was- Gamma Globulins in Systemic Lupus Erythematosus. 7th sermann Antigen-Antibody System. I. Estimation of the Interim Sci. Session Am. Rheumatism Assoc., Dallas, Wassermann Antibody in Weight Units. J. Immunol., 78: Texas, 1960. 19-29, 1957. 5~2. U~OMA, E., and TuomoaA, M. The Titer Values of the 37. ------. Quantitative Studies of Lipid-Soluble Tissue Anti- VDRL Test with Sitolipin as the Antigen. Ann Med. Exp. gens as Exemplified by the Wassermann Antigen-Antibody Fenn., 9.9:311-14, 1951. System. II. Comparative Studies with Mammalian and 53. VYGODCIIIKOV,G. V. (ed.), Pathogenesis and hnmunology Plant Tissue Phospholipids. Ibid., pp. 30-44. of Tumors. New York: Pergamon Press, 1959. 38. OSLER, A. G. ; STRAUSS, J. H. ; and MAYER, M. M. Diagnos- 54. WADSWORTH, A.; MALTANER, F.; and MALTANER, E. tic Complement Fixation. I. A Method. Am. J. Syphilis, Quantitative Studies of the Complement Fixation Reac- Gonorrhea and Venereal Diseases, 36:140-58, 195r tion with Syphilitic Serum and Tissue Extract: Technic of 39. PANGBORN, M. C. The Composition of Cardiolipin. J. Biol. the Practical Quantitative Test. J. Immunol., 35: 217-54, Chem., 168:351-61, 1947. 1938 (and earlier papers). 4o0. PANGBORN, M. C.; MALTANER, F.; TOMPKINS, V. N.; 55. WAKSMAN, B. H.; PEARSON, C. M.; and SHAm', J. T. BEECHER, T. ; THOMPSON, W. R. ; and FLYNN, M. R. Cardi- Studies of Arthritis and Other Lesions Induced in Rats by olipin Antigens. World Health Organization, Monograph Injection of Mycobacterial Adjuvant. II. Evidence That Series No. 6, 1951. the I)isease Is a Disseminated Immunologic Response to 41. PEAESON, C. M. Development of Arthritis, Periarthritis Exogenous Antigen. J. Immunol., 85:403-17, 1960. and Periostitis in Rats Given Adjuvants. Proc. Soc. Exp. 56. WALLACE,A. L. ; OSLER, A. G. ; and MAYER, M. M. Quanti- Biol. & Med., 91:95-101, 1956. tative Studies of Complement-Fixation. V. Estimation of 4~. PETERKOFSKY, A.; LEVINE, L.; and BROWN, R. K. Quan- Complement-Fixing Potency of Immune Sera and Its Rela- titative Estimation of Antigens by Complement Fixation. tion to Antibody-Nitrogen Content. J. Immunol., 65: 661- Studies on the Heat-Labile Alpha-~ Glycoprotein. J. Im- 73, 1950. munol., 76:237-42, 1956. 57. WINN, H. S. Immune Mechanisms in Homotransplanta- -~3. RANDALL, H. G.; TALBOT, S. L.; NEU, H. C.;and OSLER, tion. I. The Role of Serum Antibody and Complement in A. G. Studies on the Mechanism of Hypersensitivity Phe- the Neutralization of Lymphoma Cells. J. Immunol., 84: nomena. IV. An Isometric Smooth Muscle Assay System. Immunology (in press). 530-38, 1960. 44. RAPPORT, M. M., and GrufF, L. Immunochemical Analysis 58. ------. The Immune Response and the Homograft Reac- Based on Complement Fixation. Ann. New York Acad. tion. Nat. Cancer Institute Monograph, 2:113-40, 1960. Sci., 69:608-32, 1957. 59. WITEBSKY, E. ; ROSE, N. R.; and NADEL, H. Studies on -~5. RAPPORT, M. M.; GnAF, L.; and ALONZO, N. Immuno- Organ Specificity. X. The Serologic Specificity of Pancreas chemical Studies of Organ and Tumor Lipids. II. Organ Extracts. J. Immunol., 85:568-74, 1960. and Species Specificity of the Lipid Antigens of the Rat 60. ZIL'BEn, L. A. Specific Tumor Antigens. Adv. Cancer Re- Lymphosarcoma. Cancer, 8:546-51, 1955. search, 5: 291-,o5, 1958.

Abraham G. Osler

Cancer Res 1961;21:1187-1197.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/21/9/1187.citation

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/21/9/1187.citation. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.