Rosette formation by peripheral lymphocytes. II. Inhibition of the phenomenon

In a previous paper it had been shown that up to 40% of human peripheral lymphocytes will make rosettes with sheep red cells. In this study the inhibitory effect of antilymphocyte serum (ALS) and of certain other preparations on this rosette phenomenon is described. ALS was prepared in horses by injection of human peripheral lymphocytes and rosette-inhibiting, cytotoxic and agglutinating titres were determined during immunization. The ALS had been tested in vivo by regional administration to vervet monkeys bearing skin allografts. It was found that the peak of rosette-inhibiting titre appeared to provide a good index of immunosuppressive activity. These peaks, and the accompanying periods of immunosuppressive activity, were of short duration, in contrast to the levels of cytotoxic and agglutinating titres. Antihuman lymphocyte sera had considerable rosette-inhibiting titres against the lymphocytes of the baboon (Papio papio) but little against those of the vervet monkey (Cercopithecus aethiops).

Inhibition by ALS of the phenomenon described by us is unaffected by complement. Antihuman globulin serum and phytohaemagglutinin had no inhibitory effect. Digestion of ALS with pepsin to obtain the F (ab')₂ fragment had only a slight effect on rosette-inhibiting titres, while it reduced the cytotoxic activity very considerably.

The phenomenon described by us is probably different from that reported by other workers, in which only a small proportion of lymphocytes form rosettes.

     

Organ specificity of antithymocyte globulin

Absorption of horse antimouse thymocyte globulin with either mouse kidney, liver, spleen, mesenteric lymph node, or bone marrow resulted in significant reduction of thymocytotoxicity titre. Absorption with thymus completely removed the thymocytotoxic effect. Cytotoxicity against spleen was removed by absorption with spleen, not by absorption with kidney, liver, or thymus.

Absorption by thymus resulted in complete removal of immunosuppressive potency, as judged by duration of survival of H-2 incompatible cardiac allografts, but absorption with the other lymphoid and non-lymphoid tissues did not significantly affect immunosuppressive potency.

These studies suggest that horse antimouse thymocyte globulin contained antibodies which were cross-reactive with non-thymic lymphocytes and non-lymphoid tissue but in addition contained antibodies specific for thymus. These thymus specific antibodies were of major importance in determining the immunosuppressive potency of horse antimouse thymocyte globulin.

     

Studies on antimacrophage globulin

The immunosuppressive and anti-inflammatory properties of rabbit antirat macrophage (AMG) and rabbit antirat thymocyte globulin (ALG) have been compared. In vitro experiments showed that both AMG and ALG contained antibodies directed against rat thymocytes and macrophages. When AMG was absorbed with large numbers of thymocytes, the final product was found to lack agglutinating and cytotoxic antibodies against such cells. However, AMG was cytotoxic to rat peritoneal macrophages at a dilution of 1/512 while the ALG cytotoxic titre was 1/64. It was also shown that AMG was able to suppress phagocytosis by peritoneal mononuclear cells while ALG was ineffective.

The immunosuppressive activity of AMG and ALG when given at the time of induction of the immune response was studied in vivo. Rats were pretreated with the antisera prior to induction of Freund's adjuvant arthritis and immunization with sheep erythrocytes. Administration of ALG prior to induction of adjuvant arthritis resulted in complete inhibition of the disease while AMG treated rats developed arthritis. ALG treated rats produced almost no haemagglutinating antibody to sheep erythrocytes used as antigen while the AMG treated group developed titres not significantly different from the control group. When the same animals were rechallenged with sheep erythrocytes 5 weeks later the response of the AMG treated group was similar to that of controls, while the ALG treated animals developed a primary antibody response.

When the antisera were given to animals coinciding with the development of generalized arthritis in order to study their anti-inflammatory activity, AMG treated rats showed a small decrease in the severity of arthritis. ALG was much more effective in reducing the inflammatory signs of the disease and its suppressive effects lasted for 2 weeks after the last injection. The anti-inflammatory properties of ALG were also evident in the marked inhibition of the granuloma in the adjuvant-injected paw which paralleled the effects of treatment on the polyarthritis. Similarly, ALG treatment produced a severe depression of the delayed hypersensitivity response to tuberculin while AMG was ineffective.

The above results indicate that ALG is a strong anti-inflammatory and immunosuppressive agent while AMG has no immunosuppressive properties and is only a mild anti-inflammatory agent.

     

Rosette formation by peripheral lymphocytes

In preparations of human peripheral lymphocytes suspended in serum absorbed with sheep red cells, up to 30% of the lymphocytes may make rosettes with sheep erythrocytes.

Washed lymphocytes suspended in Hanks' solution make many rosettes if tested without delay. Such lymphocytes rapidly lose their capacity to make rosettes, but it can be restored by adding the serum of man or of the horse, rabbit or guinea-pig. The lymphocytes of three newborn babies, and of one adult who had no detectable antisheep agglutinin in the serum, made rosettes with sheep cells.

Rosette formation is uncorrelated with serum agglutinin levels. Many normal adults have far higher titres of agglutinins against the red cells of other animals than against sheep cells; yet their lymphocytes do not make rosettes with the cells of these other animals.

Sodium cyanide (0·01 M) abolished rosette formation, and horse antihuman lymphocyte globulin inhibits it.

It is concluded that sheep cell rosette formation by human peripheral lymphocytes is not due to humoral antibody or delayed hypersensitivity, because of the great proportion of lymphocytes that are capable of it. Its nature is obscure, but it is suggested that it may be due to a substance, not primarily an antibody, that is elaborated by a large proportion of circulating lymphocytes and cross-reacts with some red cell antigens as plant lectins do.

Caution is advised in using the system to test antihuman lymphocyte serum until more is known about it.

     

The antibody activities of 19S and 7S fractions from rabbit antisera to Bordetella pertussis

Bordetella pertussis antisera were fractionated on Sephadex G-200 to give 19S and 7S globulins, the latter refractionated on DEAE-cellulose, and the antibody properties of the two purified materials compared.

The minimal weights of 19S and 7S globulins required to neutralize haemagglutinin and to protect mice against an intranasal challenge causing lung infection were similar. In agglutination 7S globulin was 1.5–100 times more effective than 19S, the ratio depending on strain and serum, but there was no apparent correlation with specific agglutinogens.

7S globulin was at least 100 times more effective than 19S both in the complement-dependent bactericidal reaction in vitro and in the protection of mice against an intracerebral challenge. It is suggested that the two reactions measure the same antibody.

     

The effect of microsomal enzyme inhibition on the immunosuppressive and toxic effects of cyclophosphamide

Cyclophosphamide is activated in vivo to toxic alkylating materials by hepatic microsomal enzymes. In mice and rats, inhibition of these enzymes by chloramphenicol or SKF 525A reduces the toxicity of cyclophosphamide but has less effect in rats and no effect in mice on its immunosuppressive activity, so increasing its therapeutic effectiveness.

In contrast, vitamin A alcohol increases cyclophosphamide toxicity, again without affecting its immunosuppressive activity.

Hepatic activation of cyclophosphamide results in the sequential appearance of several metabolites of varying toxicity. Retardation of activation by microsomal enzyme inhibitors may alter the proportions of the various metabolites present and so lead to the differential effects observed here.

Alternatively, exposure of cells to lower concentrations of active products for longer times would favour survival of cells with substantial ability to repair alkylation damage to DNA but not that of cells with little repair ability. The shape of the dose-response curve for cyclophosphamide acting on immunocytes suggests that their repair abilities are limited. The cells essential for survival, which are differentially protected against cyclophosphamide by administering microsomal enzyme inhibitors, are probably not haemopoietic stem cells but may be intestinal epithelial cells.

The clinical relevance of these experiments is discussed.

     

Immunosuppression by mycobacterial arabinomannan

The means by which mycobacteria elude host defence mechanisms to cause human disease is unknown. This report provides the initial demonstration of a soluble product of Mycobacterium tuberculosis with immunosuppressive activity which may contribute to the pathogenicity of this organism.

Mycobacterial D-arabino-D-mannan (AM), purified from culture filtrates of M. tuberculosis by Concanavalin A-Sepharose affinity chromatography did not itself induce ³H-thymidine incorporation or migration inhibitory factor production by peripheral blood mononuclear cells from tuberculin skin-test-positive healthy humans although it produced weak delayed skin-test reactions in these subjects. However, this mycobacterial cell wall polysaccharide suppressed the response of human lymphocytes to mycobacterial and other antigens as assessed with these in vitro correlates of delayed hypersensitivity. Immunosuppression by arabinomannan was not contingent upon prior exposure of the individual to M. tuberculosis, since it occurred with lymphocytes from tuberculin skin-test non-reactive as well as reactive subjects. Nor was it due to cytotoxicity, an antimitotic effect of the polysaccharide preparation or generation of suppressor cells. Rather, AM seemed to suppress macrophage-dependent, antigen-induced activation of human lymphocytes. This immunosuppressive substance elaborated by mycobacteria has potential relevance to the pathogenesis of human tuberculosis and the nature and limitation of the resulting host response.

     

Interaction of anti-thymocyte serum with haematopoietic stem cells: II. Stimulation of colony formation in vitro

When mouse haematopoietic cells are incubated in vitro in appropriate dilutions of horse anti-mouse thymocyte serum or globulin (HAMTS or HAMTG), the number of in vitro colonies formed by such cells is increased. There was a good correlation between the immunosuppressive potency and the colony-enhancing property of a given serum preparation. Correlation between the thymocytotoxicity titre and the colony enhancing property of the sera was poor. For colony enhancement, it was found necessary to have an added source of colony stimulating factor in the medium, whether in the form of mouse serum or human urine. Unlike the effect produced by exposure of bone marrow cells to certain antigens, HAMTG or HAMTS enhanced colonies in the absence of the alpha-globulin component of mouse serum. Thymus-derived immunocompetent cells are not involved in enhancement since this effect was also observed using foetal liver cells. It appears likely that the interaction of an antibody on the surface of some cell or cells involved in in vitro colony formation permits a larger fraction of the colony-forming cells to proliferate in vitro.     

Induction and suppression of the cytotoxic activity of human lymphocytes in vitro by heterologous anti-lymphocyte serum

Heterologous anti-lymphocyte sera were demonstrated to induce a cytotoxic potential in normal non-immunized human lymphocytes against allogeneic fibroblast target cells. The cytotoxicity-inducing capacity was restricted to certain dilutions of anti-lymphocytic serum above and below which no cytotoxic effect was obtained. This optimal concentration shifted towards higher dilutions in sera taken late during the immunization course. The antisera were shown to stimulate the DNA-synthesis in lymphocytes and to aggregate the lymphocytes to the target cells. The DNA-synthesis and the aggregation as well were maximal at the same dilution of anti-lymphocytic serum which induced cytotoxicity. No cytotoxic effect was demonstrable on sheep fibroblasts. It is, therefore, suggested that the anti-lymphocytic serum antibody induces lymphocyte-mediated cytotoxicity against allogeneic fibroblasts in a two step manner: it stimulates the lymphocytes into a cytotoxic state; it aggregates the human lymphocytes to the human fibroblasts by virtue of its bivalent structure.

Anti-lymphocytic serum was also found to suppress the cytotoxic activity of lymphocytes induced by various non-specific agents, such as phytohaemagglutinin, streptolysin O and anti-lymphocytic serum itself. The mechanism for this inhibition is extensively discussed and it is suggested that anti-lymphocytic serum suppresses the reaction by coating the lymphocytes, thereby preventing the intimate contact between effector and target cell. A similar mechanism may operate in vivo and could be a partial explanation of the in vivo immunosuppressive effect of anti-lymphocytic serum. Purified 7S γ-globulin possessed all activities of the whole antiserum.

     

A microdroplet assay for immune adherence

A microdroplet technique for assaying anti-lymphocyte antibody using rat peritoneal macrophage monolayers is presented. Its reliability is assessed with duplicate and repetitive performance using samples of normal and anti-lymphocytic equine globulin. The specificity of the test for antibody causing adherence of antigen to macrophages is indicated in a second series of tests by the use of globulin fractions with widely varing cytotoxic, lymph agglutinating and immunosuppressive activities.     

Effects of antimacrophage, antithymocyte, antilymphocyte and antispleen serum on the immune response in mice

The effect of antimacrophage serum (AMS) on antibody production was compared to that of antithymocyte (ATS) and antilymphocyte (ALS) serum. All three types of antisera inhibited antibody production to SRBC in mice. Antispleen serum was not immunosuppressive. Immunosuppression could best be demonstrated when the antisera were injected 3 days before a low dose of antigen (10⁷ or 5 × 10⁷ SRBC). None of the antisera affected secondary antibody production.

There was no correlation between the immunosuppressive potency of the antisera and their in vitro cytotoxic or in vivo lymphopenic acitivity. AMS inhibited phagocytosis, whereas ATS and ALS enhanced phagocytosis. So far we have been unable to absorb out immunosuppressive activity of the antisera but have been able to absorb out cytotoxic activity. The significance of these findings is discussed.

     

The significance of the bursa of Fabricius in relation to the synthesis of 7S and 19S immune globulins and specific antibody activity in the fowl

Fowls in which the development of the bursa of Fabricius was suppressed, by the injection of testosterone in ovo, failed to develop normal serum levels of both 7S and 19S immune globulins. In some instances the 7S immune globulin, in the serum of bursaless fowls, suggested the survival of clones of cells synthesizing 7S immune globulin with a sharply defined electrophoretic mobility. Thus the bursa may make some fundamental contribution to immunological competence and when this effect is markedly reduced by testosterone treatment the few cells which are capable of globulin synthesis subsequently proliferate and produce globulin lacking the usual electrophoretic heterogeneity of mobility.

The indirect haemagglutination test showed that normal fowls developed 7S and 19S antibodies 7 days after the injection of BSA. At this time most of the activity was in the 19S fraction. However, 35 days after the primary injection and 7 days after a second stimulus, highest activity was in the 7S fraction. Precipitin activity was shown at 7 and 35 days in the 7S but not the 19S fractions. The 19S but not the 7S immune globulins were sensitive to mercaptoethanol.

Bursaless fowls, when given similar BSA injections, failed to show any antibody activity in their sera and the antigen was circulating 7 days later, when normal fowls had developed antibody. When time was given for antigen clearance, bursaless fowls still failed to respond to a secondary BSA stimulus. Surgical bursectomy at hatching reduced, but did not eliminate, specific antibody production.

Although there was a significant decrease in thymus weight in testosterone treated fowls, this was not observed in surgically bursectomized fowls. Nevertheless, in both these groups of fowls there was a significant decrease in secondary lymphoid foci in the spleen and in spleen weights. Thus, secondary lymphoid foci appeared to be more dependent on the bursa than on the thymus.

The significance of the synthesis of 7S immune globulin, which occasionally has a sharply defined electrophoretic mobility, by bursaless fowls together with their inability to show any specific antibody activity, is discussed in relation to the function of the bursa.

     

The Significance of Multiple Antibody Components in Serum of Immunized Rabbits

The electrophoretic patterns of six sera from rabbits immunized by two or more courses of intravenous injections of killed pneumococci type III showed multiple peaks in the γ-globulin region. Such sera contained large amounts of antibody (up to 85 per cent of the total γ globulin) against the capsular polysaccharide. One serum contained a cryoglobulin, which contained almost as great a proportion of specific antibody as did the remaining γ globulin.

The electrophoretic patterns and antibody contents were similar in the water-soluble and water-insoluble fractions of γ globulin.

The sedimentation constant and diffusion coefficient of a water-soluble fraction of γ globulin, containing 85 per cent specific antibody, were measured. The values, at 0.4 per cent protein concentration, were S_20.w = 6.97×10^-13 and D_20.w = 4.16×10^-7 cm.² sec.^-1, corresponding to molecular weight 159,000.

The antibody-containing globulin from one serum was separated by zone electrophoresis into three fractions with different electrophoretic mobilities. These contained 53–71 per cent of antibody precipitable by type III pneumococcus capsular polysaccharide. Only doubtfully significant differences were found in respect of amino-acid composition, hexose and hexosamine contents, or antigenic characteristics.

A method was devised for detecting small amounts of antibody against capsular polysaccharide by means of red cells sensitized with culture filtrates of capsulated pneumococci.

The antibody was also fractionated by chromatography on anion-exchange cellulose, and numerous fractions with antibody activity were obtained. It was shown by labelling the γ globulin with ¹³¹I that similar fractionation occurred both in the presence and absence of other serum components. All the chromatographic fractions of γ globulin were found to contain approximately similar proportions of antibody. By electrophoresis in starch gel the fractions were found to differ from one another and to be heterogeneous.

The implications are discussed of the finding that antibody against type III pneumococcus capsular polysaccharide can occur over the entire range of γ-globulin molecules.

     

The Role of Penicillenic Acid as an Antigenic Determinant in Immediate-Type Sensitivity to Penicillin

Rabbits injected with a pre-incubated mixture of benzyl penicillin and normal rabbit serum in Freund's complete adjuvant produced an antibody which reacted by tube precipitation, agar-gel diffusion precipitation, and passive cutaneous anaphylaxis with (1) a conjugate formed between benzyl penicillenic acid and human γ globulin, and (2) one of four preparations of a mixture of benzyl penicillin and human γ globulin. Evidence was obtained that the penicillin—protein preparation was effective by virtue of penicillenic acid—protein groups that it contained.

The theory that systemic penicillin sensitivity is due to allergy to conjugates formed in vivo between the benzyl penicillin derivative, benzyl penicillenic acid and plasma protein is reviewed. It is suggested that the experimental results obtained support this theory.

     

Analysis of the biological activity of antilymphocyte serum: III. Concentration of immunosuppressive activity on one IgG subfraction and its inhibition by another

Two IgG subfractions of horse antilymphocyte serum (ALS) were obtained by DEAE Sephadex chromatography. Although the fractions did not differ antigenically, they differed on amino acid and carbohydrate analysis, and in electrophoretic mobility. As demonstrated by binding studies, only the most positively charged population of IgG molecules (fraction 1) obtained from anti-lymphocyte serum had specificity for the small lymphocyte; 50 per cent of the molecules in this population bound specifically to lymphocytes in vitro. As determined by an in vitro correlate of immunosuppressive potency (rosette inhibition), fraction 1 (F₁) IgG from ALS contained approximately 4 times the specific activity of fraction 2 (F₂). F₁ was significantly more effective in prolonging skin graft survival than F₂, whereas F₂ contained the major component of the non-specific anti-inflammatory activity of serum. The anti-inflammatory effect was mediated by anticomplement activity.

F₂ was found to be an effective inhibitor of the immunosuppressive activity of F₁ both in vivo and in vitro. Quantitative studies indicated that 1 part of F₂ could maximally inhibit 4 parts of F₁. The percentage of F₂ present in serum IgG was inversely related to the skin graft survival elicited by the serum, which indicated that F₂ was active as an inhibitor when tested as purified fraction as well as in unfractionated serum. Following immunization when F₁ gained immunosuppressive potency, it lost non-specific anti-inflammatory activity. These observations indicated that not only was there a quantitative, as well as a qualitative concentration of immunosuppressive antibodies in F₁, but also that this activity was controlled by the concentration of F₂. This report, therefore, describes an IgG control mechanism which can limit the expression of antibody induced biological activity.

It is suggested that in ALS the immunosuppressive antibody molecules possess a greater net positive charge than the remaining population, and that this is due to the degree of the negative charge on the immunizing antigen. Using DEAE Sephadex chromatography, these populations could be separated into two differently charged populations of molecules, only one of which had significant immunosuppressive capability. This increase in activity resulted from the increase of specific molecules, the loss of non-specific molecules, and was manifest upon the removal of an IgG inhibitor.

     

Eimeria tenella (Gallus domesticus): Attempts to Induce a Passive Immunity to • in Young Fowls

Whole serum or γ globulin derived from fowls, either susceptible or immune to coccidiosis (Eimeria tenella), was injected into fully susceptible fowls.

The serum proteins were given by intravenous or intraperitoneal injection and subsequent attempts were made to infect these fowls by giving either a moderate dose of sporulated oocysts per os or a suspension of viable merozoites per rectum.

In the agar-gel double diffusion test, the serum from the donor birds, resistant to E. tenella, formed only weak lines of precipitate when reacting against an antigen prepared from the second schizont stage of the life cycle. However, the immune serum was considered satisfactory because the donor birds on challenge were immune.

Although relatively large amounts of γ globulin were injected into the susceptible fowls (up to 0.88 g. per kg. body weight) and subsequently only mild infections were given, no passively acquired resistance was shown either from the results of the oocyst counts on the faeces or by a macrosopic or microscopic examination of the caeca.

These results are discussed in relation to earlier studies; they show that passively acquired serum antibody at these dose levels did not provide protection.

     

Immunosuppressive properties of a ribonuclease-containing fraction from bacterial cultures

The supernatant fluid from cultures of an organism resembling Bacillus cereus is immunosuppressive to mice given intravenous sheep red blood cells, and inhibits the proliferative response in vitro of human or mouse lymphocytes to phytohaemagglutinin. Immunosuppressive activity in such culture fluids is proportional to the amount of ribonuclease activity present. Adsorption and elution from DEAE cellulose results in two peaks of ribonuclease activity, only one of which is immunosuppressive in vivo and in vitro. Gel filtration on Biogel P-10 produces a peak containing all of the ribonuclease activity and immunosuppressive potency, and has an estimated mol. wt of 9000–10,000. In its partially purified form, enzymatic activity is of the order of 100 units ribonuclease per microgram of protein; less than 100 units are required to completely suppress the immune response to sheep red blood cells in a mouse. The immunosuppressive activity of the bacterial product may be analogous to that of the serum alpha-globulin (Fraction C) which also possesses ribonuclease activity.     

Passive anaphylaxis in mice with γG antibodies: V. Competitive effects of different immunoglobulins and inhibition of reactions with antiglobulin sera

Mouse antisera to DNP conjugates were fractionated on agar block electrophoresis, and each fraction was tested for ability to give anaphylactic reactions in vivo in mice and guinea-pigs, and in vitro on mouse mast cells. Concentrations of IgG₁ and IgG₂ globulin were measured for each fraction, and the results indicate that mouse IgG₁ globulin mediates both the in vivo PCA reaction in the mouse and the in vitro mast cell sensitization.

PCA reactions in guinea-pigs with mouse antisera are mediated only by electrophoretic fractions containing IgG₂ globulin. PCA reactions in guinea-pigs could be blocked by the addition of IgG_2a myeloma containing sera to the sensitizing anti-DNP preparation, but not by addition of IgG_2b or IgG₁ myeloma sera.

From six different IgG₁ myeloma sera tested on their capacity to block PCA reactions in mice, five showed strong inhibitory activity whereas the sixth was inactive. However, also IgG_2a and IgG_2b myelomas were found capable of blocking PCA reactions in mice. The passive in vitro sensitization of mouse mast cells by purified mouse anti-DNP antibodies could also be blocked by the addition of specific anti-IgG₂ sera. Possible interpretations of these findings are discussed.

     

Immunosuppresive effect of a human hepatic glycoferroprotein, alpha2H globulin. A study on the transformation of normal human lymphocytes.

A macroglycoferroprotein of hepatic orgin, alpah2H globulin, the serum level of which increases a few weeks or months before local recurrence of metastases, has been essayed for its immunosuppressive activity. The study was carried out using the lymphoblastic transformation test and was judged by tritiated thymidine incorporation and microscopic examination. PHA-induced blast transformation of 97 per cent of normal donor lymphocytes is inhibited by 100 mug/ml of alpha2H globulin. This inhibitory effect is proportional to the quantity of added alpha2H globulin. In is obvious with a concentration of 2-5 mug/ml, a frequently observed level in the serum of patient with tumours. Preincubation of lymphocytes with alpha2H globulin renders more effective the inhibitory action on PHA-induced transformation. A mechanism of competition between PHA and alpha2H globulin is suggested by preincubation and the inhibitory effect related to the doses. However, microscopic observation shows that alpha2H globulin acts on the earliest events occurring to the stimulated lymphocytes, by inhibiting cytoplasmic RNA and protein synthesis. The alpha2H globulin effect may not only have an immunosuppresive activity but it may have a more general effect, for example blocking or modifying cellular respiration.     

Study of the Antigenic Constituents of Sera from Mouse/Rat Chimaeras

The sera of CBA mice lethally irradiated and restored with rat bone marrow or foetal liver have been examined by immunoelectrophoresis at various times after irradiation. Such sera contain several specific rat proteins including γ globulin, at least two α₂ globulins, one or two β₁ globulins and one β₂ globulin. A large number of specific mouse proteins are also present, but no mouse γ globulins have been detected.

The rat constituents appear progressively and persist for long periods; this implies that they are synthesized by rat cells.

There is a close correlation between the presence of rat γ globulin in the serum and the presence of dividing rat cells in the bone marrow and spleen.

The fact that γ globulin is solely of rat specificity does not necessarily rule out the possibility of immune activity by the host against the graft.

Several of the normal serum proteins, in addition to the γ globulins, originate from the injected cells, and are therefore presumably not synthesized by the liver.

In chimaeras grafted with rat skin, the number of specific rat serum proteins is often larger than the number in chimaeras which have not been grafted.