(Received 30 September 1954)

Transcription

1 205 J. Physiol. (I955) I29, THE ROLE OF HISTAMINE IN CUTANEOUS ANTIGEN- ANTIBODY REACTIONS IN THE RAT BY W. E. BROCKLEHURST, J. H. HUMPHREY AND W. L. M. PERRY From the National Institute for Medical Research, Mill Hill, London (Received 30 September 1954) It is generally accepted that local cutaneous anaphylactic reactions in man are the result of histamine released in the skin as a result of antigen-antibody combination in the tissues. The position in the laboratory animal is more complex. There are few, if any, naturally occurring cutaneous anaphylactic reactions, and studies are therefore restricted to experimentally induced anaphylactic reactions. These are of two main types, both being defined as a result of an antigen-antibody union. The first type of response is characterized by increased capillary permeability, and can be mimicked by the intradermal injection of histamine. A typical example of this type of response is passive cutaneous anaphylaxis (P.C.A.) which has been produced in the guinea-pig (Ramsdell, 1928; Ovary & Bier, 1953) and in the rat (Ovary & Bier, 1952), and which, according to Ovary (1952) is blocked by mepyramine. The second type of response is characterized by infiltration of the skin with polymorphs and mononuclear cells with considerable accompanying cellular damage (Humphrey, 1955). Two different responses of this type are the tuberculin and Arthus reactions, which are not satisfactorily mimicked by the injection of histamine and are resistant to block by antihistamine substances. Of these distinct types of experimental anaphylaxis in the skin, only one was therefore thought to depend upon histamine release; and the recent introduction of histamine liberators (MacIntosh & Paton, 1949) seemed to offer another method of studying this presumptive function of histamine. The normal histamine content of the skin varies considerably in different animal species, being about 30,ug/g in the rat and only about 3,tg/g in the guinea-pig; yet the guinea-pig skin is much more reactive than the rat skin, in that experimental reactions are much easier to produce; this has been decisively demonstrated for P.C.A. by Ovary (1952). Feldberg & Talesnik (1953) were able to deplete rat skin of nearly 90% of its histamine content by treatment, either systemically or locally, with compound 48/80, one of the most 14 PHYSIO. CXXIX

2 206 W. E. BROCKLEHURST AND OTHERS potent histamine liberators (Paton, 1951); the histamine in the guinea-pig skin, however, is extremely resistant to liberation by these compounds (Perry, unpublished). Because of these apparent anomalies it was highly desirable that any comparison of the two types of cutaneous anaphylactic reaction should be carried out in parallel in the same animal species; and the species of choice was the rat, owing to the ease with which the skin histamine can be depleted. The Arthus reaction, unlike the tuberculin type of reaction, can be elicited by passive sensitization with serum (Beniacerraf & Kabat, 1950) and was thus more convenient for study. Ovary & Bier (1952) used rabbit antibodies to produce a 'reversed' P.C.A. reaction and a 'reversed' passive Arthus reaction in rat skin, and our methods were based on their work. The reversed passive type of reaction in which antibody is given intradermally in graded doses and antigen is administered intravenously, has advantages in this type of work over the direct type of reaction, in that the dose of antibody can be accurately graded and the lesions are readily measured. Thus a quantitative assessment of the response is possible without the complication of variation in antibody response which is inevitable when active immunization methods are used. This is particularly true in the rat, which is a poor producer of antibody. Furthermore, in the reversed passive type of reaction the animals are being exposed to the antigen for the first time, thus avoiding possible interfering mechanisms. In this paper we describe the development of quantitative methods, based on measurement of lesion diameters, of evaluating P.C.A. and Arthus reactions in the rat, and the effects on these reactions of antihistamine compounds and of depletion of the skin histamine. METHODS Animals Hooded black and white rats (N.I.M.R. strain), of both sexes and weighing between 100 and 120 g were used. At this weight range animals are about 7 weeks old and the skin histamine content has passed its maximum level (Hardwick, 1954), is of the order of 25,ug/g, and is still falling. Thus 3 weeks later the normal skin histamine (10 weeks) is about 18Mg/g (Hardwick, 1954), and our own results on normal control animals of each age group are in fair agreement with those of Hardwick, who used albino rats. The abdominal skin was clipped and depilated with barium sulphide paste 24 hr before intradermal injections were made. Depletion of skin histamine In early experiments a 0-1 % solution of compound 48/80 in physiological saline was given in a course of i.r. injections similar to that used by Feldberg & Talesnik (1953). A typical dosage scheme was 100, 140, 180, 250, 320, 400, 500 and 600,ug/100 g on successive days. On one occasion we continued the course for a further 7 days, increasing the dose as far as 800 Jhg/l00 g; in other experiments two injections were given daily. Neither of these modifications significantly increased the depletion of the skin histamine. In some experiments the course was shortened in order to obtain partial depletion of the skin histamine.

3 HISTAMINE AND SKIN REACTIONS 207 Other attempts were made to reduce the skin histamine even further. (a) The skin area to be tested was infiltrated with 48/80 mixed with testicular hyaluronidase (0 05 mg/ml. 'Rondase', Evans Medical Supplies), to facilitate even spreading and distribution of the dose. Up to four such injections were given at 2-day intervals. (b) Octylamine (Light and Co.) converted to the hydrochloride was used as a histamine liberator (Mongar & Schild, 1953). A typical course of octylamine was 100, 300 and 1000,ug/100 g s.c. in a volume of 2 ml. on alternate days. (c) Combined courses of 48/80 and octylamine were given. The dose ranges used were the same as those given in the separate courses. (d) The histamine liberators were rubbed lightly over the skin area to be tested. 48/80 and octylamine were used separately and together, dissolved in 80% acetone to facilitate drying and penetration. (e) We thought it possible that in skin already depleted by a histamine liberator a, cutaneous anaphylactic reaction might itself produce further depletion. To test this we elicited a large P.C.A. response in depleted skin, using a pneumococcus antigen-antibody system. Two days later, we tested the skin at the same site, both by producing a P.C.A. reaction with an ovalbumin antigenantibody system and by estimating the skin histamine content. Arthus reaction Arthus reactions were produced by the reversed passive method described by Ovary & Bier (1952). Graded doses of the antibody were injected intradermally on the abdomen of the rat in volumes of 0-05 ml. and immediately thereafter excess of antigen was given intravenously in a volume of 0-5 ml. The doses of antibody normally used were yg. Antigen was given in a total dose of 2-10 mg. The animals were killed 4 hr later and the mean diameters of the areas of oedema and haemorrhage at the sites of the antibody injections were measured on the undersurface after reflecting the abdominal skin. The antigen usually used was ovalbumin, five times recrystallized by the method of Kekwick & Cannan (1936). The antibody was whole rabbit serum containing specifically precipitable anti-ovalbumin prepared by the Freund adjuvant technique (Freund & McDermott, 1942) or by other methods. In a few experiments, specifically noted in the text, we used as antigen pneumococcus type III polysaccharide with rabbit y-globulin of which 80% was specifically precipitable anti-pneumococcus antibody. Fig. 1 a shows the type of response that is obtained, and illustrates how a graded response to doses of antibody in the ratio 1: 2:4: 8 is obtained. Passive cutaneous anaphylaxis P.C.A. reactions were produced by the 'reversed' technique described by Ovary & Bier (1952). We used the same antigen-antibody systems as described for the Arthus reaction. Graded doses of the antibody were injected intradermally into the abdominal skin of the rat, in volumes of 0-05 ml. Three and a half hours later, excess antigen (3 mg) was given intravenously in a volume of 0-5 ml. of a 1-5% solution of Pontamine Blue 6 BX (George Gurr and Sons, London) in physiological saline. The doses of antibody were 10-80,ug. The animals were killed 25 min after the injection of antigen and the mean diameters of the areas of blueing measured on the undersurface after reflecting the abdominal skin. In two experiments almost pure anti-pneumococcus antibody gave P.C.A. responses which did not appear to differ in any way from those seen with cruder antibody; this indicates that the reaction depends only on the antibody content of the serum, even when crude preparations are used. The optimal time for reading the lesions after the injection of antibody was found to be min. Before this the responses were more variable, and after half an hour they became morediffuse and the edges were consequently difficult to determine. The interval between the injection of antigen and the injection of antibody was fixed at 3i hr. If the interval is as short as 10 minr there is a well-developed non-specific reaction which occurs also if serum containing no antibody is used. This non-specific reaction has almost disappeared if the interval is lengthened to 30 min. 14-2

4 208 W. E. BROCKLEHURST AND OTHERS The specific P.C.A. response, which is obtained only with the appropriate antigen, becomes wellgraded to the dose of antibody with intervals of 2 hr or more, but is too diffuse for easy measurement if the interval is 5 hr or longer. Typical graded responses are shown in Fig. l b. The procedures used in producing both Arthus and P.C.A. reactions are summarized in Table 1. Fig. 1. (a) Reversed passive Arthus reactions in normal rat skin, showing graded responses to doses of antibody (pg) given in pattern 25 l50) (b) Reversed passive cutaneous anaphylactic (P.C.A.) reactions in normal rat skin showing similar grading. Doses of antibody (jg) given in pattern Two reactions to intradermal injections of histamine (&g) are also shown in the right on pattern TABLE 1. Reaction Arthus P.c.A. Summary of methods of production of reversed P.C.A. and reversed passive Arthus reactions in rat skin Amount (sg) Interval Amount (mg) Interval of antibody before of antigen before (I.D.) antigen (I.V.) reading Lesion (diameter measured) Nil hr Haemorrhage and oedema j hr min Blueing due to increased capillary permeability Estimation of histamine content of skin Samples of the abdominal skin were taken in all experiments from sites adjacent to those used for the cutaneous anaphylactic reactions and their histamine content estimated. Histamine was extracted by a modification of the method used by Feldberg & Talesnik (1953). The skin sample

5 HISTAMINE AND SKIN REACTIONS (about 200 mg) was cleaned of subcutaneous tissue, weighed and immediately boiled for a few moments in 2 ml. of N/3-HC1. It was then ground with sand and boiled for a further 3 min. The ph was adjusted to 7*4 with N/3-NaOH, the suspension filtered and the filtrate made up to 10 ml. with Tyrode solution. The extracts were assayed on the atropinized guinea-pig ileum, and the results expressed in pg histamine base/g tissue. The usual control experiments were carried out to exclude from the assay substances other than histamine which cause contraction of the gut. In depleted skins the content of histamine was so low that part of the contraction was in fact due to such substances; and it was not possible wholly to eliminate these effects by dilution of the extract, since this reduced the histamine level below the limit of sensitivity of the ileum. Taking larger skin samples incurred the risk of using skin from areas known to have different histamine contents (Feldberg & Miles, 1953). Thus the histamine contents of the depleted skins may, in some cases, be slightly overestimated, but this does not materially affect the arguments put forward in this paper. Design of experiments In all experiments on cutaneous anaphylactic reactions, animals were randomly assigned to the different treatment groups; and the doses of antibody and of drugs injected intradermally were given in a random pattern over the skin area tested. In early experiments treatment groups were housed in separate cages, but this introduced a systematic error in several cases, since if, for instance, one cage were kept at a temperature slightly higher than another, e.g. in the sunlight, this was enough to alter markedly the anaphylactic responses. Both responses are dependent to a large extent on the skin blood flow and a small increase therein may well enhance them. At even higher temperatures, the skin blood flow may become so great as to diminish the responses, owing presumably to a washing out of the antibody. In all the experiments quoted, we therefore assigned animals to cages at random and in equal numbers from each of the treatment groups. Furthermore, it was found to be necessary to control the environmental conditions as closely as possible in order to minimize the variance of the responses between animals. This error variance was also greatly reduced with increasing technical skill, especially in the i.v. injection process, since handling of the animals more than is absolutely Aecessary always reduces the responses, presumably due to vasoconstriction of skin vessels induced by liberated adrenaline. The lesions were measured by an operator who was unaware of the treatment group to which any particular animal belonged. RESULTS The depletion of the histamine content of the skin of the rat Feldberg & Talesnik (1953) found that even prolonged courses of intraperitoneal injections of 48/80 failed to reduce the content of histamine of the abdominal skin of the rat below about 12% of its normal value. We have confirmed this finding, using doses of 48/80 by intraperitoneal injection somewhat smaller than those given by Feldberg & Talesnik. 209 In our hands, initial doses in excess of 110g/100 g rat were frequently lethal, whereas Feldberg & Talesnik were able to give rats of between 100 and 140.g as much as 200,ug of 48/80 without killing the animals. With dosage schemes rising from 100 to 800pg/100 g rat/day, we were unable to reduce the skin histamine below some 10 % of its initial value. We have attempted, using all the different techniques already described, to produce a further reduction of skin histamine. In Table 2 are shown the results

6 210 W. E. BROCKLEHURST AND OTHERS from two experiments in which many of these techniques were used. Octylamine was much less efficient than 48/80 in reducing the skin histamine, and even when given in combination with 48/80 did not increase its effect. In those animals which had undergone a P.C.A. reaction 2 days previous to the extraction of the skin histamine the results were inconclusive; if any further reduction did occur, it was incomplete and a considerable proportion of the extractable histamine remained unreleased. TABLE 2. Effects of various methods of depletion on the skin histamine content of rats No. of Mean skin histamine Expt. Treatment of skin animals Skin histamine levels (ttg/g) (Gg/g) I. (a) 48/80 +hyaluronidase s.c , 5, 4-6, 5 4, 054, (b) 48/80 +hyaluronidase s.c. + 48/80 painted 6 5-1, 4.7, 5-1, 5-8, 4-8, on (c) Octylamine +hyaluronidase s.c , 13-9, 27-1, 15-9, 18, (d) Octylamine +hyaluronidase s.c , 15-6, 28-5, 16, 24, octylamine painted on (e) 48/80 + octylamine + hyaluronidase 4 1-6, 2, 7-9, (f) Untreated controls 5 38, 20, 34, 30, II. (a) 48/80 +octylamine +hyaluronidase s.c , 6-7, 3-8, (b) 48/80 + octylamine + hyaluronidase s.c , 4-0, 3-6, P.C.A. 2 days previously (c) 48/80 + octylamine + hyaluronidase s.c , 2-3, 3.3, P.C.A. immediately before extraction (d) 48/80 + octylamine + hyaluronidase s.c , 2-4, 2-9, 3-6, 2-3, 3.5, P.C.A. 2 days previously +P.C.A. immediately before extraction (e) Untreated controls 6 28, 35-5, 32-5, 25-5, 34-5, When hyaluronidase was given, the average reduction in skin histamine was no greater than that produced in other experiments in which it was not used. In spite of the failure of hyaluronidase to increase the depletion of skin histamine, we used it routinely in most of our experiments since we considered that the depletion would probably be more even in the area of skin used for tests of anaphylactic response and for extraction of skin histamine. In order to test whether the 8-10% of skin histamine which we failed to remove was in fact resistant to liberation by 48/80, we injected 48/80 intradermally into such depleted skins and measured the response produced. The results of three such experiments are shown graphically in Fig. 2. In these experiments the normal animals had a skin histamine content of approximately 23,g/g, whereas the depleted animals had skin histamine contents ranging from 2 to 5,ug/g with a mean at about 3,ug/g, i.e. 13% of the initial histamine remained. In such animals the response to intradermal 48/80 has been reduced more than 90 %. It is impossible to give an accurate figure for the degree of reduction since the slopes of the dose-response lines are different. This difference in slope may in fact be attributable to a different mode of production of the reaction; in the depleted animals the dose of 48/80 required was so large

7 11 HISTAMINE AND SKIN REACTIONS E E x.o 47 x 0 5 ' Dose of 48/80 I.D. (j.g) Fig. 2. Effect of depletion of skin histamine on response in groups of six rats to intradermal (I.D.) injections of 48/80. Ordinate: mean lesion diameter (mm); abscissa: dose of 48/80 (,ug). 0-0, normal rats; x x, rats with skin depleted to 13% normal histamine E L- v E n Dose of histamine I.D. (jig) Fig. 3. Effect of depletion of skin histamine on response in groups of four rats to I.D. histamine. Ordinate: mean lesion diameter (mm); abscissa: dose of histamine (,jg); *-*, normal rats; x x, rats with skin depleted to 10% normal histamine. (Course of 48/80 to produce this depletion finished on day before test.)

8 212 W. E. BROCKLEHURST AND OTHERS that the 48/80 may itself have exerted a direct effect on the capillary permeability. Furthermore, the reduction in the response to 48/80 is proportionately greater than the reduction of the skin histamine content. We are consequently of the opinion that these experiments serve to confirm our conclusion that the remaining 10% of skin histamine is in fact resistant to liberation by histamine liberators. Effect of histamine in depleted skin In similar experiments we have tested the effect of depletion of skin histamine on the response to histamine injected intradermally. Depletion of the skin histamine by subcutaneous injections of 48/80 reduced the effect of intradermally injected histamine by approximately 50% (Fig. 3). It should be noted that this test was carried out on the day after the course of treatment with 48/80 was finished, and that this reduction in sensitivity to histamine may represent the refractoriness of skin which has been subjected to prolonged and intensive treatment with 48/80. No such refractoriness was observed 2 days later, or when depletion was produced by a course of intraperitoneal injections, although in fact the skin histamine was reduced just as much. 10 / E 5, E 4, 0~~~~~~~~ x.o /~~~~~~~~~~~~ / Dose of histamine I.D. (Ig) Fig. 4. Effect of mepyramine on response in groups of six rats to I.D. histamine. Ordinate: mean lesion diameter (mm); abscissa: dose of histamine (lg); 0 *, normal rats; x x, rats with skin depleted to 10% normal histamine. Solid lines before, dotted lines after mepyramine 50 mg/kg. Effects of mepyramine On response to intradermal histamine The effect of mepyramine on the response to intradermal injections of histamine in normal animals and in animals whose skin histamine has been depleted to about 10 % is shown in Fig. 4. In both groups there is a 64-fold reduction in the effect of histamine injected 45 min after treatment with

9 213 HISTAMINE AND SKIN REACTIONS mepyramine 50 mg/kg intraperitoneally. The dose-response lines are steeper after mepyramine. In this experiment (in which depletion was produced by subcutaneous 48/80) the histamine response was also depressed in the depleted animals, both with and without mepyramine, to the same relatively small extent (roughly two-fold). This is taken as a further indication of the refractoriness to histamine induced by recent subcutaneous injections of 48/ E, 5-o oe- O~~~~~~~~~~~ Dose of 48/80 I.D. g) Fig. 5. Effect of mepyramine on response in groups of four rats to I.D. 48/80. Ordinate: mean lesion diameter (mm); abscissa: dose of 48/80 (pg). 0-0, normal rats; *---., normal rats after mepyramine 50 mg/kg. 0-0, rats with skin depleted to 10% normal histamine; Ei---E], depleted rats after mepyramine 50 mg/kg. On response to intradermal 48/80 Fig. 5 shows the mean response of groups of four rats to intradermal injections of 48/80. In normal animals 0*006,ug of 48/80 produced a lesion of approximately 8 mm diameter. Mepyramine in a dose of 50 mg/kg caused a 64-fold reduction in the response, i.e. a dose of 0 4,g was necessary to produce a lesion of the same size. In animals depleted by 48/80, the response to intradermal 48/80 was reduced by about 16-fold, but even this reduced response was further depressed by mepyramine 64-fold (i.e. to the same extent as is the response in normal animals). Mepyramine depressed the responses to intradermal 48/80 and to intradermal histamine to about the same extent (Fig. 4). Miles & Miles (1952) found that mepyramine was 200 times less effective in blocking intradermal 48/80 than in blocking intradermal histamine, but they were using threshold doses of mepyramine in guinea-pigs, whereas our results were obtained using very large doses in rats.

10 214 W. E. BROCKLEHURST AND OTHERS On passive cutaneous anaphylaxis and Arthus reactions Arthus reaction. We found that, as described for guinea-pigs by Smith & Humphrey (1949), doses of mepyramine up to 50 mg/kg-the maximum tolerated dose-did not in any way influence the development of the Arthus reaction in the skin, the lesions being of the same size and intensity. Passive cutaneous anaphylaxis. Ovary (1952) found that P.C.A. in the rat skin was completely blocked by antihistamines. He used mepyramine in doses 15 _ Er E~~~~~ ( 10 1 E // ~0/ C~~~~~~~~~~~~~~~~ U) E 0~~~~~~~~~~~~~~ Fig. 6. (a) (b) Dose of histamine (jtg) Dose of antibody (ug) Effect of mepyramine on response in groups of eight rats to P.C.A. and to I.D. histamine. Ordinate: mean lesion diameter (mm); abscissae: (a) dose ofhistamine (pg), (b) dose ofantibody (jig); 0-0, P.C.A. response in normal rats; x--- x, P.C.A. response in normal rats after mepyramine 50 mg/kg; 0-0, histamine response in normal rats; A---A, histamine response in normal rats after mepyramine 50 mg/kg. of 50 mg/kg. One of the reasons for our selection of P.C.A. as a test response was this report of its sensitivity to antihistamines and our selection of mepyramine, like Ovary's, was based on its relatively high specificity and absence of side actions. In our hands, however, mepyramine in doses of 50 mg/kg had no effect on the P.C.A. response (Fig. 6); doses of 100 mg/kg were invariably fatal. Fig. 6 shows the striking difference in the slopes of the dose-response lines of histamine and of P.C.A., the latter being at least twice as steep as the former. More striking is the fact that the dose-response line for P.C.A. is almost identical in slope to that of the Arthus reaction (Fig. 7), both reactions having been read 4 hr after the injection of the antibody.

11 HISTAMINE AND SKIN REACTIONS 215 Effects of other antihistamine compounds Because of the difference between our results with mepyramine and those of Ovary (1952), we repeated our experiments with a number of other antihistamines. These were chosen to be as diverse as possible both in their chemical _~~~~~~~~~~~~~~~~~~~~~~~~~~~~0 15 E A0, x E / ~~~~~~~~~~~~~~~0~~~~~~~~~~~~0 C ,_ 5 (a) (b) (c) (d) 3 1 Dose of histamine I.D. (jg) Dose of 48/80 I.D. ('g) Dose of antibody-arthus (pg) 80 Dose of antibody-p.c.a. (Ug) Fig. 7. Responses in normal rats in two experiments. Ordinate: mean lesion diameter (mm); abscissae: (a) dose of I.D. histamine (,ug), (b) dose of I.D. 48/80 (,ug), (c) dose of antibody (j,g) in Arthus reactions, (d) dose of antibody (lag) in P.C.A. reactions. Expt. 1: groups of twenty-four rats; 0-0, Arthus reaction; *-*, P.C.A. reaction; *l----l, histamine reaction; Expt. 2: groups of four rats; x x, histamine reactions; --- A---A, 48/80 reaction. structure and in their reported effects (Packman, Rossi & Harrisson, 1953). The results are shown in Table 3. Depression of the P.C.A. response occurred only with those drugs which, in the doses used, produced collapse of the rats. Furthermore, the degree to which the response to intradermal histamine was

12 216 W. E. BROCKLEHURST AND OTHERS blocked was in no way parallel to the reduction of the P.C.A. response. Thus, chlorcyclizine in a dose which produced a 256-fold reduction in a response to intradermal histamine failed completely to influence the P.C.A. response. Promethazine, 50 mg/kg, produced very severe collapse of the animals and we consider that the reduction in P.C.A. response which it produced is attributable to a general circulatory failure. Even in a lower dose of 20 mg/kg, it still caused collapse of the rats, yet reduced the P.C.A. response only four-fold, despite a depression of the histamine response by well over 256-fold. We conclude from these experiments that P.C.A. is not specifically inhibited by antihistamines. TABLE 3. Effects of antihistamine compounds on P.C.A. response Mean P.C.A. response Mean blue response (mm) to (mm) to A r histamine dose (Mg) A Dose high dose low dose, Group Antihistamine (mg/kg) antibody* antibody Non-specific effects I Nil Promethazine Profound circulatory collapse. Moribund 3 Antazoline Collapsed and convulsing 4 Chlorothen Nil obvious 5 Chloreyedizine Nil obvious 6t Nil t Promethazine Collapsed 8t Antazoline Slight * High dose of antibody =4 x low dose. t Groups 6, 7 and 8 in separate experiment, with different antigen (pneumococcus type III polysaccharide). Effects of depletion of skin histamine on Arthus reaction and passive cutaneous anaphylaxis Arthus reaction. The Arthus reaction was produced in rats after varying degrees of depletion of the skin histamine by 48/80. There was no clear quantitative or qualitative difference in the responses obtained in animals whose skin histamine content varied from 15 to 100% of normal. Atypical experiment is shown in Table 4. In addition to the similarity in the lesion diameters of the different groups shown in the table, the lesions were very similar in the degree of haemorrhage and oedema produced. Furthermore (Fig. 8a), the slope of the dose-response lines was the same in all three groups. It appears, therefore, that reduction of the skin histamine to 15 % of its normal value does not in any way influence the development of the Arthus response. These experiments were repeated in animals which had not only been depleted but which also received mepyramine 50 mg/kg intraperitoneally 45 min before the injection of the antibody. As in normal skin, mepyramine

13 HISTAMINE AND SKIN REACTIONS 217 did not have any influence on the development of the Arthus reaction in histamine-depleted skin. Passive cutaneous anaphylaxis. We tested the effect of different degrees of reduction of the skin histamine on the P.C.A. response. The results of one such experiment are shown in Fig. 8 b. There were no significant differences between the responses at the different levels of skin histamine content (Table 5). The TABLE 4. Effect of depletion of skin histamine on reversed passive Arthus reaction Mean diameter +s.e. of lesion (mm) (haemorrhage Mean skin and/or oedema) produced by dose of antibody (pg) Group No. of histamine, no. animals (% normal) *2±0*53 9-2± ±0*49 14*1±0* *3±0*66 9*0±0*50 11*0±0*57 13*5± X2±0*72 9*3±0*34 11*5± ±0-34 TABLE 5. Effect of depletion of skin histamine on reversed P.C.A. reaction Mean diameter ±s.e. of lesion (mm) (blueing) Mean skin produced by dose of antibody (jug) Group No. of histamine, A no. animals (% normal) *4±0*84 8*0± *4±0*53 13*1+0* *5±2*0 6*8±0-75 9*8±0*8 12*94-0* ± ± (a) 15 (b) Arthus reactions P.C.A. reactions -0 ~~~.10 - ~~~~~~10 0) 9 5~~ - ~ 5 ~ ~ ~ ~ ~ ~ ~ - 01 I I 101 I I Dose of antibody (pg) Fig. 8. Effect of depletion of skin histamine on (a) Arthus reactions, and (b) r.c.a. reactions in groups of eight rats. Ordinates: mean lesion diameter (mm); abscissa: dose of antibody (isg). x - x, normal rats; 0--, rats with skin depleted to 25% normal histamine; , rats with skin depleted to 15% normal histamine.

14 218 W. E. BROCKLEHURST AND OTHERS lesion diameters were substantially the same as in normal skin and both the threshold doses of antibody and the slope of the dose-response line were unaffected by the reduction of skin histamine. The only difference consistently observed was that, in depleted skin, the edges of the lesion were more sharply defined and there was a lack of the diffuse, faint background blueing, often extending beyond the actual lesion, seen in the response in normal skin. The skin histamine content of each individual animal was determined and related to the P.C.A. response of that animal. Fig. 9 shows results from one > co x c c x0 x 0 x E 5 X E 0 X Skin histamine content (Ug/g) Fig. 9. Scatter diagram relating skin histamine content with P.C.A. response in individual rats. Ordinate: lesion diameter (mm) produced in P.CAs. by 40pg of antibody; abscissa: skin histamine content (tg/g); *,normal rats (8); x, rats with skin depleted to 13% normal histamine (13). such experiment. It is obvious that no correlation exrists between the two variables. Even if the normal animals are considered alone, there is no apparent correlation; and the same is true of the depleted animals considered alone. Depletion of the skin histamine does not affect the slope of the dose-response line of the P.C.A. response. This is shown in Fig. lo, which also shows th'at mepyramine is without effect on the response in depleted animals. The amount of intradermally injected histamine needed to produce a lesion of the same diameter as that produced by the P.C.A. can readily be calculated from these data. Thus Fig. lo shows that 0 07,ug histamine injected intradermally produces in normal animals a lesion of the same diameter as that

15 HISTAMINE AND SKIN REACTIONS 219 produced by the low dose of antibody used; whereas, in depleted animals treated with mepyramine, a dose of 20,ug histamine is necessary to produce a similar lesion, i.e. of diameter about 7-5 mm. The histamine content of the skin E 10 5 ol. (a) Dose of histamine in normal rats (jug) (b) Dose of histamine (jg) in depleted rats after mepyramine, 50 mg/kg (c) Dose of antibody (jig) Fig. 10. Effect of mepyramine on responses in groups of four rats to P.C.A. and I.D. histamine. Ordinate: mean lesion diameter (mm); abscissae: (a) dose of histamine (jig) in normal rats; (b) dose of histamine (j,g) in rats with skin depleted to 10% normal histamine and treated with mepyramine 50mg/kg; (c) dose of antibody (,ug). 0-0, normal rats; 0-0, rats with skin depleted to 10% normal histamine; A-A, normal rats after mepyramine; A-A, depleted rats after mepyramine. Solid lines, P.c.A. responses; dotted lines, histamine responses. is also known; and the total histamine in an area of skin of diameter 7-5 mm, weighing about 25 mg, has been shown in the same experiment to be normal skin (53-5,ug/g) = 1-33,ug depleted skin (61,ug/g)=0 = 15,g.

16 220 W. E. BROCKLEHURST AND OTHERS Thus, although in normal skin, in the absence of mepyramine, there is more than sufficient histamine to produce the P.C.A. lesion, in depleted skin treated with mepyramine, there is only 1/130 part (0-15/20) of the histamine which must be injected to produce a lesion of the same size as that produced by the P.C.A. response under these conditions. In other experiments similar calculations have given ratios as low as 1/1000. It should also be noted that, since the slope for the histamine dose-response line is very much shallower than that for the P.C.A. response line, the value for the ratio just calculated would be still smaller if the histamine equivalents of lesions of larger dimensions were calculated: thus in the experiment illustrated in Fig. 10, the amount of injected histamine needed to produce a 16 mm lesion in depleted skins treated with mepyramine can, by extrapolation, be estimated at about 1-25 mg. The histamine content of such an area of skin is approximately 0 75jg and the ratio is thus about 1/1700. DISCUSSION Our choice of cutaneous anaphylactic responses for this work was based on the published reports that P.C.A. could be mimicked by a histamine injection and was blocked by antihistamines, whereas the Arthus reaction could not be so mimicked and was not so blocked. Our results with a number of antihistamines, however, lead us to the conclusion that P.C.A. is not blocked by antihistamines unless the drugs are given in such massive doses as to produce complete circulatory collapse of the rats. It is possible that Ovary succeeded in blocking the P.C.A. response only by this means and not by virtue of specific antihistamine activity. Feldberg & Talesnik (1953) showed how the relatively high histamine content of rat skin could readily be depleted, by a histamine liberator, from about 30pg/g to about 3,ug/g. We have confirmed their findings in every respect. Moreover, we have tried by a wide variety of methods to deplete the skin of even more of its histamine and have in every case failed to do so. Because of these failures we have come to regard the residual 10% of skin histamine in the rat as resistant to histamine liberators. It is perhaps worth noting that this residual histamine is roughly equal in amount (3pg/g) to the normal skin histamine content of the guinea-pig; and that the skin histamine of the guineapig also appears to be resistant to histamine liberators (Perry, unpublished). Riley & West (1953) and Riley (1953) have suggested that histamine in tissues is largely located in the mast cells, from which it is liberated, with accompanying disruption of the cells or loss of their granules, by histamine liberators or by any form of mild trauma; but Fawcett (1954) found that 48/80, although causing loss of granules from mast cells, did not cause disruption of the cells. Limited histological examinations which we made similarly showed that, in rats whose skin had been depleted of histamine by 48/80, there

17 HISTAMINE AND SKIN REACTIONS 221 was a loss of granules from the numerous large deeply staining mast cells normally present. Guinea-pig skin is not nearly so rich in mast cells as rat skin, and those that are found are smaller and less deeply staining. In view of the ease with which histamine is liberated from mast cells (Riley & West, 1953) it is possible that the depletion in rat skin is largely due to its loss from these cells; and it may well be that the residual 10% of histamine is resistant to liberation because it is present in other types of cell. In fact, Graham, Lowry, Wahl & Priebat (1953) found suggestive evidence that some 10% of skin histamine in the dog was in cells other than mast cells. The fact that in depleted skins 48/80 injected intradermally will still produce a response if the dose used is sufficiently large, suggests that such doses produce a lesion by some mechanism other than the release of histamine. Yet the fact that mepyramine blocks the response to 48/80 to the same degree as it does in normal skin is not in agreement with this hypothesis, unless the assumption is made that the antihistamines also antagonize the direct effects of 48/80 to the same degree as they antagonize released histamine. The dose-response lines (relating log. dose to lesion diameter) of intradermally injected histamine and 48/80 have almost the same slope. Yet when a P.C.A. response is elicited in the skin, the corresponding dose-response line (relating log. dose of antibody to the lesion diameter) has a slope which is twice as steep. Furthermore, although the Arthus phenomenon has never been attributed to the release of histamine, the slope of the dose-response line for the Arthus reaction (Figs. 7, 8) is almost identical with that for P.C.A. This suggests that the two phenomena may be linked by a common aetiological factor. Thus, we believe that the governing factor in the production of the P.C.A. response is a sufficient concentration of antigen-antibody complex, and that consequently the area of blueing is determined by the area over which the antibody has diffused to give a supra-threshold concentration; within this area the response appears to be uniform. It is probable, however, that some histamine is released as a secondary result of the antigen-antibody reaction; and this histamine may account for the diffuse blurring of the edges of the P.C.A. reaction in normal animals. Thus, in the absence of releasable skin histamine (in depleted animals) we observed the edges of the P.C.A. response to be more sharply defined. Antihistamines at high dose levels block the effects of intradermal injections of histamine and of 48/80 to a similar extent. Thus not only do the antihistamines block injected histamine, they also block the action on capillary permeability of released histamine. Yet they fail to affect the P.C.A. response. Feldberg & Talesnik (1953) found that in rats depleted of histamine by 48/80, the characteristic oedema following the injection of whole egg-white did not occur. On the assumption that the effect in unsensitized animals of whole eggwhite (purified ovalbumin being inactive) depends upon a direct liberation of 15 PHYSIO. CXXIX

18 222 W. E. BROCKLEHURST AND OTHERS histamine, and not on any form of antigen-antibody reaction, the results of Feldberg & Talesnik agree with our findings with 48/80. In depleted skin (10% normal histamine) the response to intradermally injected histamine is little affected; but the response to intradermally injected 48/80 is reduced 256-fold or more. Yet the P.C.A. response is unchanged. This is even more striking than at first appears, for three reasons. First, if an attempt were made to produce blueing in depleted skins by intradermal injection of histamine so as to correspond to a P.C.A. response, the dose of histamine needed to produce even the faintest blueing would be greater than the total free histamine extractable from the corresponding area of skin-and in mepyramine-treated animals it would be 1000 times greater. Furthermore, it is probably greater than the amount of histamine which could be introduced into the area by the formed elements of blood. Secondly, in depleted animals neither the threshold dose of antibody nor the grading of the response to increasing doses of antibody is changed. Finally, the histamine content of areas of skin in which a P.C.A. reaction has taken place is little different from the histamine content of similar areas of skin in which no reaction has occurred. (This is only true when the histamine contents are expressed in,ug/unit area; if expressed in,ug/g the content falls, due presumably to the increased weight of the skin because of oedema formation.) For all these reasons we conclude that P.C.A., for all its superficial resemblance to the effects of injected histamine, is brought about by a mechanism which is independent of the release of intrinsic skin histamine; and in spite of our failure to obtain complete depletion of the skin histamine, we do not consider that the residual 10% can be involved. The implications of any other conclusion are that the residual histamine, which we believe to be located in cells other than the mast cells, is not liberated by 48/80 or octylamine but is liberated by antigen-antibody reactions; and that after release it is not blocked by antihistamine compounds, in the presence of which it is at least 1000 times as effective as histamine released from mast cells or given by intradermal injection. These conclusions apply with equal force to the Arthus reaction, and thus one of the criteria for differentiating between the two types of reaction is removed. Furthermore, we noted that, when a P.C.A. reaction was produced with a large dose (80.Lg) of purified antibody, the area of blueing often showed a central patch of haemorrhage which was never seen as a result of large intradermal doses of histamine or 48/80. Thus, the distinction between P.C.A. reactions and the early stages of the Arthus reaction may well be quantitative rather than qualitative.

19 HISTAMINE AND SKIN REACTIONS 223 SUMMARY 1. Methods are described for producing Arthus and P.C.a. reactions quantitatively by a reversed passive technique in rat skin. 2. Various techniques were used in an attempt to deplete the rat skin of all its free histamine. None effected reduction of more than 90 %. Reduction of the skin histamine to 10% of normal was accompanied by disappearance of the granules from the skin mast cells. 3. The residual 10% of skin histamine is not liberated by intradermal injections of compound 48/80, nor is it greatly reduced by antigen-antibody reactions. 4. Arthus and P.C.A. reactions in rat skin are not blocked by antihistamine compounds unless these are given in doses sufficient to produce signs of collapse in the rats. 5. Arthus and P.C.A. reactions in rat skin depleted to 10% of its normal histamine content are of the same size and intensity as those in normal rat skin. 6. Antihistamines do not modify these reactions in depleted rat skin, although they still block intradermally injected histamine. The P.C.A. reaction under such conditions is equivalent in size and intensity to the effect of the intradermal injection of more than 1000 times the quantity of histamine present in the corresponding area of skin. 7. These findings are incompatible with the view that histamine release is the immediate consequence of antigen-antibody combination in the skin and the cause of the observed phenomena of cutaneous anaphylaxis. We are indebted to Messrs Burroughs Welicome and Co. for a supply of compound 48/80; and to Messrs May and Baker Ltd, for samples of mepyramine maleate and promethazine hydrochloride. REFERENCES BENACERRAF, B. & KABAT, E. A. (1950). A quantitative study of the Arthus phenomenon induced passively in the guineapig. J. Immunol. 64, FAWCETT, D. W. (1954). Cytological and pharmacological observations on the release of histamine by mast cells. J. exp. Med. 100, FELDBERG, W. & MILES, A. A. (1953). Regional variations of increased permeability of skin capillaries induced by a histamine liberator and their relation to the histamine content of the skin. J. Physiol. 120, FELDBERG, W. & TALESNIK, J. (1953). Reduction of tissue histamine by compound 48/80. J. Physiol. 120, FREUND, J. & MCDERMOTT, K. (1942). Sensitization to horse serum by means of adjuvants. Proc. Soc. exp. Biol., N. Y., 49, GRAHAM, H. T., LowRY, 0. H., WAHL, N. & PRIEBAT, N. K. (1953). Localization of tissue histamine in mast cells. Abstr. XIX int. phy8iol. Congr. p HARDWICK, D. G. (1954). Age changes in the histamine content of rat skin. J. Physiol. 124, HuMPHREY, J. H. (1955). The mechanism of Arthus reactions. I. The role of polymorphonuclear leucocytes and other factors in reversed passive Arthus reactions in rabbits. Brit. J. exp. Path. (in the Press). 15-2

20 224 W. E. BROCKLEHURST AND OTHERS KEKWICK, R. A. & CANNAN, R. K. (1936). The hydrogen ion dissociation curve of the crystalline albumin of the hen's egg. Biochem. J. 30, MACINTOSH, F. C. & PATON, W. D. M. (1949). The liberation of histamine by certain organic bases. J. Phy8iol. 109, MIEs, A. A. & MILES, E. M. (1952). Vascular reactions to histamine, histamine-liberator and leukotaxine in the skin of guinea-pigs. J. Physiol. 118, MONGAR, J. L. & SCHLD, H. 0. (1953). Quantitative measurement of the histamine releasing activity of a series of mono-alkylamines using minced guineapig lung. Brit. J. Pharmacol. 8, OvARY, Z. (1952). Cutaneous anaphylaxis in the albino rat. Int. Arch. Allergy, N. Y., 3, OvARY, Z. & BiER, 0. G. (1952). Quantitative study of Arthus reaction and of cutaneous anaphylaxis induced passively in the rat. Proc. Soc. exp. Biol., N. Y., 81, OvARY, Z. & BIBR, 0. G. (1953). Quantitative studies on passive cutaneous anaphylaxis in the guineapig and its relationship to the Arthus phenomenon. J. Immunol. 71, PACKMAN, E. W., Rossi, G. V. & HARRISSON, J. W. L. (1953). The effect of histamine and antihistamines on body temperature. J. Pharm. 5, PATON, W. D. M. (1951). Compound 48/80: a potent histamine liberator. Brit. J. Pharmacol. 6, RAMSDELL, S. G. (1928). The use of trypan blue to demonstrate the immediate skin reaction in rabbits and guineapigs. J. Immunol. 15, RILEY, J. F. (1953). The effects of histamine liberator on the mast cells in the rat. J. Path. Bact. 65, RILEY, J. F. & WEST, G. B. (1953). The presence of histamine in tissue mast cells. J. Physiol. 120, SMITH, W. & HUMPHREY, J. H. (1949). The effect of sodium salicylate upon hypersensitivity reactions. Brit. J. exp. Path. 30,