glomerular basement...

Glomerular Basement Membrane- Reactive

Antibody in Anti-Lymphocyte Globulin

CuRns B. WILSON, FRANKJ. DIXON, JOSEPHG. FORTNER,andG. JANms CEmUIa

From the Department of Experimental Pathology, Scripps Clinic and ResearchFoundation, 476 Prospect Street, La Jolla, California 92037

A B S T R A C T Equine immunoglobulin was detectedalong the glomerular basement membrane of three hu-man homograft recipients who had been treated withequine anti-lymphocyte globulin. Anti-lymphocyte globu-lins, given these patients, were obtained by immunizationof horses with lymphocytes from human spleens and/orlymph nodes and contained glomerular basement mem-brane-reactive antibodies. Quantitative paired-label iso-tope experiments (in rats) demonstrated that 30-170IPg/ml of kidney-fixing antibodies were present in thesepreparations. The anti-lymphocyte globulins formed aline of identity with a sheep anti-human glomerularbasement serum when reacted against collagenase-solu-bilized human glomerular basement membrane in doublediffusion in agar. The renal fixation of these antibodieswas blocked by absorption with human glomerular base-ment membrane, but not by buffy-coat leukocytes, indi-cating that they were directed specifically toward anti-gens in the basement membrane and were not cross-reacting anti-lymphocyte antibodies.

Anti-lymphocyte globulin preparations for human usewere studied for glomerular basement membrane-reac-tive antibodies by a direct immunofluorescent assay inrats. Anti-lymphocyte globulin from 13 of 20 horses, and7 of 10 serum pools from horses immunized with lympho-cytes derived from solid lymphoid organs (spleen, thy-mus, lymph node, tonsil), contained glomerular base-ment membrane-reactive antibodies. Sera from 18 horsesinjected with thoracic duct cells or cultured lymphoblasts

This is publication 465 from the Department of Experi-mental Pathology, Scripps Clinic and Research Foundation,La Jolla, Calif., and was presented in part at the AmericanSociety of Nephrology Meeting, 23-24 November 1970, inWashington, D. C.

Dr. Fortner is from the Department of Surgery, Memo-rial Sloan-Kettering Cancer Center, New York, and Dr.Cerilli, from the Department of Surgery, Ohio State Uni-versity Medical Center, Columbus, Ohio.

Received for publication 26 January 1971.

had no glomerular basement membrane-reactive anti-bodies.

An equine anti-human thymus serum containing glo-merular basement membrane-reactive antibodies, whichproduced fatal glomerulonephritis in monkeys, wasshown to cause both immediate and delayed glomerularinjury in monkeys after intravenous injection. Thereaction of this antibody with glomerular basement mem-brane in vivo was associated with little complement depo-sition in spite of the fact that the antibody could fixcomplement. This lack of glomerular complement fixa-tion resulted from almost complete in vivo decomplemen-tation of the monkeys receiving this anti-lymphocyteglobulin.

INTRODUCTIONAnti-lymphocyte globulin (ALG)1 has come into wide-spread clinical use as an immunosuppressant in humanorgan transplantation. The effectiveness of this agentin prolonging skin grafts in rats (1) initiated its clinicaluse. Subsequent observations have revealed the nephrito-genic potential of ALG. Lance reported nephritogeniceffects of rabbit anti-mouse thymus sera in mice (2).Iwasaki et al. reported production of glomerulonephritisof the serum sickness type after administration of equineanti-canine lymphoid sera in dogs (3). Serum sicknesshas also been reported in mice (4) and man (5) afterthe administration of ALG. In a series of papers, Gutt-mann, Carpenter, Lindquist, and Merrill reported glo-merulonephritis in rats was produced by rabbit anti-ratthymus serum, presumably mediated in part by anti-

'Abbreviations used in this paper: ALG, anti-lymphocyteglobulin; BUN, blood urea nitrogen; GBM, glomerularbasement membrane; GBMRAb, GBMreactive antibodies;Ig, immunoglobulin; PBS, phosphate buffered saline; PMN,polymorphonuclear leukocytes; SGOT, serum glutamic-ox-aloacetic transaminase; SGPT, serum glutamic-pyruvictransaminase.

The Journal of Clinical Investigation Volume 50 1971 1525

bodies capable of reacting with glomerular basementmembrane (GBM) (6-8). Antibodies also have beenfound in equine anti-canine lymph node serum whichfixed to dog glomeruli (9, 10).

To determine the nephritogenic potential of ALG inman, we studied renal tissues obtained from patientsundergoing treatment with ALG in conjunction withrenal or hepatic transplantation. Equine immunoglobulin(Ig) was demonstrated fixed in a linear fashion alongthe GBMin 3 of 35 patients in this group. Equine anti-human spleen and lymph-node lymphocyte sera adminis-tered to these patients were found to contain GBMreac-tive antibodies (GBMRAb). These antibodies werequantitated and their specificity demonstrated. A surveyof the incidence of GBMRAbin ALG preparations pro-duced for human use was done and correlated with thesource of lymphocyte antigen used for immunization.The nephritogenic properties of an equine anti-humanthymus antiserum in monkeys were defined.

METHODSImmunofluorescent studies. Renal tissues from 35 patients

(20 kidney and 15 liver homograft recipients) who had beentreated with equine ALG were examined by the direct im-munofluorescent technique (11). The reagents used werefluorescein isothiocyanate conjugated rabbit IgG shown tobe specific for human immunoglobulin (IgG, IgA, IgM),complement (C) component C3, fibrinogen, albumin, andequine Ig by means of immunoelectrophoresis and doublediffusion. The immunofluorescent microscopy techniques usedin this laboratory have been described previously (12).

Quantitation of GBMRAbin ALG preparations. GBM-RAb in seven ALG preparations were quantitated by thepaired label technique (13) and included those (H-ic, H-11,and H-50a)' given the three patients presented. 100 pg ofI or 'I-labeled ALG (14) with an equivalent amount of

normal equine-Ig labeled with the other isotope were givenintravenously to male Sprague-Dawley rats (150 g). Thenormal equine Ig was prepared by precipitation in a finalconcentration of 50% ammonium sulfate. Groups of four ratseach were injected and bled out 18 hr later. The liver,spleen, lungs, heart, and kidneys were taken after extensivein situ perfusion with 0.01 M sodium phosphate buffered0.15 m NaCl, pH 7.4 (PBS). The renal binding of ALGH-35 (Ps) and H-56b was also quantitated in Saimiri sci-urea monkeys (600 g). Pairs of monkeys were nephrecto-mized 18 hr after intravenous injection of 200 pug of ALGwith an equivalent amount of paired label normal equine Ig.Because of inability to perfuse the nephrectomy specimens,washed kidney homogenates were used. Radioactivity was

2The ALG preparations referred to throughout this paperwill be designated as in Table III. The source is indicatedby letter (H, horse; R, rabbit, and G, goat). Each animalis assigned a separate number. Different bleedings from thesame animal are designated a, b, c, d. (ALG H-ia wouldmean the first bleeding studied from horse 1.) If the ALGwas pooled sera from several animals, it is designated byP and the number of animals involved placed in paren-theses.

determined in a NaI crystal dual channel scintillation coun-ter. The specific amount of ALG bound to an organ wascalculated by the formula:Specific I*ALG organ bound = I*ALG organ bound

-I*Eq Ig (organ) X I*ALG (blood)L I*Eq Ig (blood) J

The results are presented for each group of rats or pairs ofmonkeys as the mean per cent of ALG administered whichwas specifically bound to an organ.

Specificity of binding of I*ALG. Four of the I*ALG'sthat exhibited specific renal binding were absorbed witheither human buffy coat leukocytes or human GBM todetermine whether the kidney-fixing antibody might be across-reacting antileukocyte antibody. Antibodies to serumproteins detectable by gel diffusion were removed by priorabsorption with pooled normal human serum. Four serialabsorptions of 450 pg of I*ALG were done with washedhuman buffy coat leukocytes. The leukocytes were preparedby sedimentation in one part 3% dextran T500 (PharmaciaFine Chemicals, Inc., Piscataway, N. J.) in PBS and twoparts heparinized peripheral blood. The cells were washedfour times in Hanks solution before absorption. The firstthree absorptions were done with 1 X 108 cells for 30 minat 37°C and the final absorption with a similar number ofcells overnight at 4°C with constant shaking. Likewise, 450pg of each I*ALG were absorbed with 25 mg human GBM(15) overnight at 4°C with constant shaking. The specificbinding of the I*ALG to leukocytes or GBMused for ab-sorption was calculated as described for organ binding.

The absorbed I*ALGs were used in in vivo paired labelexperiments identical with those used to quantitate theGBMreactivity of the original ALG, and the residual spe-cific organ binding was determined. The effect of the ab-sorptions on the binding of the I*ALG to buffy-coat leuko-cytes was also determined. 2 ug each of unabsorbed andabsorbed I*ALG were incubated with 1.2 X 1i0 washedbuffy-coat leukocytes for 1 hr at 37°C with frequent agita-tion. The specific binding of I*ALG to the leukocytes wascalculated as described in the paired-label studies after fourwashes of the cells in Hanks solution.

In vivo determination and incidence of GBMRAbin ALG.ALG preparations for human use originating from horses,rabbits, or goats were assayed for GBMRAbby direct im-munofluorescence. Sprague-Dawley rats (150-200 g) wereinjected intravenously with 1.0-2.0 ml of an ALG prepara-tion, housed in metabolic cages for urine collection, and bledout 18 hr after injection. Portions of the kidneys werefrozen in liquid nitrogen for immunofluorescence and theremainder placed in Bouin's fixative for routine histologicexamination. The presence of GBMRAbwas determined bydirect immunofluorescence using fluorescein isothiocyanateconjugated rabbit anti-equine Ig, sheep anti-rabbit IgG, orrabbit anti-goat Ig. Positive staining was shown to bespecific for Ig by absorption studies using homologous Ig.The ability of the GBMRAbto activate rat C was deter-mined by staining each positive tissue with fluorescein iso-thiocyanate conjugated rabbit anti-rat C3. The urine wasstudied for the presence of protein (12). Excretion ofgreater than 20 mg per 24 hr was considered abnormal.

Other measures of GBMreactivity in ALG preparations.The ALG preparations (H-ic, H-il, H-50a) were alsostudied for GBMRAbby indirect immunofluorescence usingcryostat sections of normal human kidney as a target. Thesections were incubated with dilutions of the ALG for 30min at room temperature. After washing, the sections were

1526 C. B. Wilson, F. J. Dixon, J. G. Fortner, and G. J. Cerilli

stained with fluorescein isothiocyanate conjugated rabbitanti-equine Ig.

The three ALG preparations which reacted in vivo withpatients' GBMwere studied for GBMRAbby double dif-fusion in 0.7% ion agar in PBS containing 3% glycine.The human GBMantigen (a gift from Dr. Hans Mar-quardt) was prepared from glomeruli obtained by the methodof Spiro (16) and solubilized by digestion with collagenase.Sheep anti-human GBMantibody (a gift from Dr. CharlesCochrane) was used for comparison. All precipitation linesto serum proteins were removed by prior absorption withpooled normal human serum.

Studies on the route of administration of the ALG prepa-rations. To study the effect of local absorption of GBM-RAb, the amounts of antibody fixed in the kidneys afterintravenous and intramuscular administration were com-pared. One or more intramuscular injections not exceed-ing 0.1 ml were compared to intravenous injections in 150 gSprague-Dawley rats. Fixation of Ig to the GBMwas de-tected by direct immunofluorescence.

Studies on the production of glomerulonephritis in ratsand monkeys with ALG preparations containing GBMRAb.Evidence for the immediate (heterologous) phase of nephro-toxic nephritis was sought in all rats used for the in vivoassay for GBMRAb. In addition, single intravenous injec-tions of 2.5-5 ml of selected ALG preparations were madeinto Sprague-Dawley rats (150 g). One kidney was removedfrom each animal 2.54 hr after injection; the second wasremoved at sacrifice 24-72 hr later. These kidneys werestudied by immunofluorescent and routine histologic means;the urinary excretion of protein was measured.

A sufficient quantity of equine anti-human thymus ALG(H-35(Pu)) was available for biologic studies in monkeys.Four 3-4 kg Macaca nemestrina burmensis monkeys weregiven 30 ml of this preparation intravenously and housedin metabolic cages for urine collection. Urinary protein wasdetermined. Base line protein excretion (4 days) was foundto average 18 mg/24 hr (range 5-32 mg). Renal biopsieswere taken 4 hr after injection. In one monkey, hepatic andsplenic biopsies were also obtained at 4 hr. In the two sur-viving monkeys, weekly renal biopsies were obtained. Snapfrozen renal tissue was studied by direct immunofluores-cence for equine Ig, monkey IgG, and C3. Bouin's fixed tissuewas used for routine histologic examination. Anti-bodyresponses to equine serum proteins were monitored by doublediffusion in ion agar. Serum transaminase (glutamic-oxalo-acetic and glutamic-pyruvic (SGOT, SGPT) determina-tions were performed by Dr. Thomas S. Edgington. Onecontrol monkey given 30 ml of normal equine Ig receivedcomparable studies.

Serum C determinations (CHao) (17) were done beforeinjection and at 5 min, 10 min, 2 hr, 4 hr, and 24 hr afterinjection in two of the four monkeys receiving ALG. Theability of the glomerular-bound equine ALG to fix C wasstudied by overlaying cryostat sections of the kidney withdilutions of fresh human serum and incubating for 30 min.Bound C was determined by immunofluorescence using an-tisera specific for human C3. The ability of ALG (H-35(Pn)) to inactivate C in normal serum in vitro was deter-mined. Residual CH50 levels in 1: 2, 1: 5, and 1: 10 dilutionsof normal human serum with either the native ALG giventhe monkeys or deaggregated (upper half of the tube afterultracentrifugation 100,000 g for 60 min) were measured.The CHin U remaining in the normal human serum after60 min incubation (37°C) with the ALG dilutions (nativeand deaggregated) were compared to those found after in-

cubation with the same dilutions of heat-inactivated (560CX 30 min), deaggregated, pooled normal equine serum.

RESULTS

Immunofluorescent studies. Equine Ig was found bydirect immunofluorescence in 1 of 20 renal homograftsand in the kidneys obtained at autopsy from 2 of 15hepatic homograft patients. The staining was a con-tinuous, smooth, linear pattern along the GBM. In ad-dition to the linear equine Ig, the renal homograft con-tained trace amounts of human IgG and C3 in a linearpattern along the GBM. Kidneys obtained from thetwo patients receiving hepatic homografts also had smallamounts of linear human IgG and C3 in the glomeruli.No deposits of human IgA, IgM, or fibrinogen werefound in these tissues.

Histologic sections from the renal homograft revealedevidence of rejection with mononuclear infiltratesaround venules and glomerular hypercellularity (Fig. 1).Histologic changes in the kidneys from the hepatichemograft patients consisted of only minimal glomerularhypercellularity.

Quantitation of GBMRAbin ALGpreparations. Theresults of the paired-label quantitation studies are pre-sented in Table I. Three of the ALGpreparations (H-ic,H-11, and H-50a) exhibiting GBMRAbwhen adminis-tered to the renal or hepatic homograft patients reportedin this paper showed specific binding of 0.43, 0.07, and0.07% to the kidneys of rats. The GBMRAbin ALGH-35 (Pn), which was used for the biological studies inmonkeys, was also quantitated in rats with a specificbinding of 0.13%. The only ALG (H-56b) induced byinnoculation of cultured lymphocytes having weak GBMreactivity with rat glomeruli on immunofluorescence wasalso studied quantitatively. It showed 0.01% and lessthan 0.01% binding to rat and monkey kidney, respec-tively. Two additional ALGpreparations (H-12a, H-53a)without immunofluorescent-rat GBMreactivity failed toshow specific renal binding. From the amount of pro-tein contained in the original ALG it could be calcu-lated that ALG H-ic, H-li, H-50a, and H-35 (P22)contained 170, 45, 30, and 65 tAg of kidney-fixing anti-body/ml (in the rat). ALG H-35 (Pn) contained 345l'g of antibody/ml capable of binding to Saimiri sciureamonkey kidneys.

Striking amounts of ALG H-ic and H-35 (Pn) werealso bound to liver and spleen as shown in Table I. Theliver and spleen from rats and/or monkeys injectedwith these ALG preparations had deposits of equineIg along the hepatic sinusoids and splenic reticulum whenstained with fluoresceinated anti-equine Ig (Fig. 2).

Specificity of binding of I*ALG. The amount ofALG that bound to isolated GBMduring the absorptionwas 4.9, 2.1, 1.3, and 1.0% of ALG H-ic, H-11, H-50a,and H-35 (Pn), respectively. The only ALG (H-56b)

Glomerular Basement Membrane-Reactive Antibody in ALG 1527

FIGURE 1 Histologic alterations in the renal homograft from J. N. removed 19 wk aftertransplantation. A, a venule in the juxtamedullary region is shown surrounded by a mono-nuclear cell infiltrate. B, a glomerulus from this kidney shows moderate hypercellularity. Hema-toxylin and eosin. X 250.

TABLE IBinding of I*ALG to Rat Organs in Vivo

Per cent of 100 yg ALGbound*

ALGt 2 kidneys Liver Spleen 2 lungs Heart

H-1c 0.43 1.85 0f3i 0.06 0.04Absorbed (cells)§ 0.36 1.45 0.24 0.06 0.03Absorbed (GBM)§ 0.13 0.45 0.15 0.06 0.02

H-l1 0.07 0.24 0.10 0.05 0.03Absorbed (cells) 0.05 0.30 0.06 0.03 0.01Absorbed (GBM) 0.00 0.21 0.04 0.02 0.01

H-50a 0.07 0.43 0.04 0.03 0.01Absorbed (cells) 0.06 0.35 0.04 0.02 0.01Absorbed (GBM) 0.04 0.39 0.03 0.01 0.01

H-35 (P12) 0.13 0.34 0.06 0.00 0.00Absorbed (cells) 0.09 0.34 0.06 0.00 0.00Absorbed (GBM) 0.05 0.10 0.01 0.00 0.00

H-12a 0.00 0.13 0.06 0.05 0.02H-53a 0.00 0.00 0.00 0.04 0.02H-56b 0.01 0.12 0.01 0.01 0.00

* The percentage represents the mean of four animals in each group.The ALG preparations are identified in accord with the descriptions in

Table III.I The I*ALG was absorbed with either buffy coat leukocytes (cells) or iso-lated human GBM(GBM) before in vivo study.

produced by immunization with cultured lymphoblasts,which had weak in vivo fixation to rat GBMby immuno-fluorescence, had no specific binding to isolated humanGBM. The percentage of I*ALG bound to 1 X 10' leu-kocytes during repeated absorptions is shown in TableII.

Repeated absorption with buffy-coat leukocytes re-sulted in an average reduction of 23% in the renal fixa-tion of I*ALG, while a single absorption with GBMcaused an average reduction of 69% (Table I). Thebinding of I*ALG to leukocytes was reduced 36% byprior absorption with GBMand 85% by absorption withperipheral leukocytes.

In vivo determination and incidence of GBMRAbinALG. The incidence of GBMRAbin 81 ALG prepa-rations for human use was determined by the fixation ofIg to rat GBMin vivo (Table III). The preparationsstudied consisted of 60 bleedings from 38 individualhorses, and 10 pooled ALG preparations from 56 horses(24 of which were, in addition, studied individually).The remaining 11 preparations studied were individualand pooled sera from 126 rabbits and 15 goats.

1528 C. B. Wilson, F. J. Dixon, l. G. Fortner, and G. J. Cerilli

kL-

FIGURE 2 Immunofluorescent pattern of equine Ig bound to the lining of the hepatic sinusoids(A) and to the splenic reticulum and vasculature (B) of a monkey 4 hr after intravenous

injection of ALG H-35 (Pu2). Fluorescein isothiocyanate conjugated rabbit anti-equine Ig.X 128.

The results were divided according to the source oflymphocytes used as antigens. Inoculation with lympho-cytes of splenic origin induced GBMRAbin 9 of 12individual horses and in one pool. Inoculation with thy-

mocytes produced GBMRAbin one of five individualhorses and in five of eight pools. Lymphocytes fromlymph nodes were used for inoculation in two horses,from tonsils in one horse, and mixed lymphoid organs

TABLE I IBinding of I*ALG to Peripheral Leukocytes during Absorption

Per cent of ALGbound to buffy coat cells*

1st absorption 2nd absorption 3rd absorption 4th absorption1 X 108 cells 1 X 108 cells 1 X 108 cells 1 X 108 cells

ALG$ 371C 30 min 370C 30 min 370C 30 min 4 ACovernight

H-ic 1.89 0.44 0.31 0.34H-11 0.69 0.30 0.23 0.19H-50a 1.01 0.70 0.38 0.35H-35 (P12) 0.47 0.22 0.13 0.35

* 450 ,ug of I*ALG with an equivalent amount of paired label normal equine Igused in the in vivo quantitative experiments were absorbed.

The ALG preparations are identified in accord with the descriptions in Table III.


TABLE I I IFixation of ALG to Rat GBMin vivo

GBMreactiveALG* Lymphocyte source Bleeding antibody

H-la, b, cH-2H-3H-4H-5H-6a, b, cH-7a, bH-8a, b, cH-9a, b, cH-10H-1IH-12a, b, c, dH-13 (P3)H-14a, b, cH-15H-16H-17H-18H-19 (P2)H-20 (P4)H-21 (P4)H-22 (P4)H-23 (P12)(H-23 -- H-34¶)H-35 (P12)(H-35 -+ H-46¶)H47 (P3)H-48 (P4)R-1R-2 (P40)R-3 (P40)R4 (P40)H-49a, b, cH-SOa, bH-51H-52 (P8)

G-1

H-53a, b, cH-54a, b, cH-55H-56a, bH-57, H-58H-59H-60 - H-70R-5R-6 -- R-9G-2 (P14)

SpleenSpleenSpleenSpleenSpleenSpleenSpleenSpleenSpleenSpleenSpleenSpleenSpleenThymusThymus§ThymusThymus§Thymus§ (membranes)Thymus§ (membranes)Thymus§Thymus§ThymusThymus§Thymus§Thymus§Thymus§Thymus§Thymus§ThymusThymus§Thymus§Thymus§Lymph nodeLymph nodeTonsilLymph nodeThymus,Thoracic duct§Spleen,Thymus,Peripheral bloodThoracic ductThoracic duct§Thoracic ductCultured lymphoblasts§Cultured lymphoblastsCultured lymphoblasts§Cultured lymphoblasts§Cultured lymphoblasts§Cultured lymphoblasts§Cultured lymphoblasts§

days69, 98, 1631182373277153, 72, 12074, 12376, 124, 150137, 143, 187754560, 120, 150, 180

90, 120, 15073030457065424290

44

21165180

28304345, 60, 150330, 365240

180

24045, 90, 210150, 300, 3659035, 63

6060-2126057120

* ALG. The source of the ALG is indicated by the letter (H, horse; R, rabbit; andG, goat); each animal is assigned a separate number. Different bleedings from the sameanimal are designated a, b, c, d (ALG H-la would mean the first bleeding studied fromhorse 1). ALGprepared from pooled sera are designated P and the number of animalsinvolved are placed in parentheses.t and 11 indicate if the ALGpreparation was effective (t) or not (11) in prolonging skingrafts in monkeys (18).§ Indicates the use of adjuvant for immunization.

The individual horses (12 each) used in these pooled ALG preparations (H-23 (P12)and H-35 (P11)) were studied individually.

1530 C. B. Wilson, F. 1. Dixon, 1. G. Fortner, and G. 1. Cerilli

+t,+

+

++. +.1- -,

+,

+

+

+-t

+

-t

FIGuRE 3 Gel diffusion pattern of ALG H-ic, H-li, H-50a with in vivo GBMRAband H-53a which lacked in vivo GBMRAbreacted with collagenase-solubilizedhuman GBM(center well). Sheep anti-human GBMsera (Sh-a-GBM) was used asa reference. Note the double line of identity which developed with ALG H-ic, H-il,and Sh-a-GBM.

in pooled sera of eight horses, and all were positive.Three groups of 40 rabbits each given thymic lympho-cytes produced no detectable GBMRAb.

In contrast to the ALG prepared from lymphocytesderived from solid lymphoid organs, only ALG (H-56b)from 1 of 18 horses inoculated with lymphocytes col-lected from the thoracic duct or cultured lymphoblastshad a weak reaction with rat GBMby immunofluores-cence. Sera from 5 individual rabbits and a serum poolfrom 14 goats inoculated with cultured lymphoblasts alsohad no GBMRAb. Irregular granular and patchy stain-ing of the glomeruli seen in some rats could be easilydifferentiated from the smooth, continuous, linear stain-ing characteristic of antibodies reacting with the GBM.

The presence of GBMRAbin the series of ALG prep-arations studied was limited to horses immunized withlymphocytes derived from solid lymphoid organs. Inaddition, there seemed to be a correlation between thecrudeness of the inoculum and the development ofGBMRAb. For instance, in one laboratory, 12 of 12horses given thymic homogenate developed GBMRAbwhile 3 horses given purified thymocyte membranes didnot. There was no relationship between the use of ad-juvants and the development of GBMRAb. TheGBMRAbappeared as early as the 21st day and per-sisted in most animals studied serially. There was no

correlation between the presence of GBMRAband thein vivo immunosuppressive effect reported to us by thesuppliers of 20 ALG's tested (Balner [18]) (Table III).

Other measures of GBMRAbin ALG preparations.By indirect immunofluorescence the ALG preparations(H-ic, H-11, H-50a) could be shown to fix to humanGBM in vitro. When these ALG preparations werestudied in an undiluted state, fixation to the GBMwasobscured by intense staining of all cellular structures inthe sections. If the ALGpreparations were diluted, how-ever, the diffuse staining diminished and the presenceof linear staining along the GBMwas evident.

ALG's H-ic, H-il, and H-50a were studied for thepresence of anti-GBM antibody by double diffusion inagar. Double lines of identity developed between H-icand H-11, and an antibody developed in sheep to humanbasement membrane (Fig. 3) when reacted with col-lagenase-solubilized human GBM. These ALG prepa-rations contained 170 and 40 Ag, respectively, of kidney-fixing antibodies/ml (rat).

Studies on the route of administration of ALG prepa-rations. Comparison of intramuscular and intravenousinjections of the ALG preparation H-ic on in vivo glo-merular fixation in rats showed that 5 to 6 times asmuch ALG was needed intramuscularly as was requiredintravenously to produce detectable fixation.


FIGuRE 4 Smooth, continuous linear deposits of equine Ig are shown along theGBMof the glomerulus of a monkey 4 hr after intravenous ALG H-35 (P=). Flu-orescein isothiocyanate conjugated rabbit anti-equine Ig sera. x 250.

Studies on the production of glomerulonephritis inrats and monkeys with ALG preparations containingGBMRAb. No evidence of the heterologous phase ofnephrotoxic nephritis was seen in the rats used for thein vivo assay of GBMRAbas manifested by proteinexcretion or histologic alterations. The intravenous ad-ministration of 2.5-5 ml of ALG preparations H-ic,H-15, and H-52 (Ps) failed to elicit proteinuria orpolymorphonuclear accumulation in the glomeruli of ratsstudied at either 2.5-4 hr or 24-72 hr after injection.Heavy deposits of equine Ig found in these specimens at2.5, 4, and 24 hr disappeared by 72 hr. Rat C was notidentified in the glomeruli at any time.

ALG H-23 (Pu2) and H-35 (P12) produced a progres-sive and fatal glomerulonephritis when administered tomonkeys in doses of 120 mg/kg, three times per week(19). Frozen tissues from these monkeys were studiedin our laboratory and revealed linear fixation of equineIg, monkey IgG, and C3 along the GBM. To study themechanism of this glomerulonephritis, four monkeyswere given single intravenous injections of H-35 (Pu).Two monkeys died 6 and 12 hr after injection, appar-ently from the toxic effects of the ALG.

Heterologous or immediate phase of nephrotoxicnephritis manifested by immediate proteinuria occurredin all four monkeys, with the two that survived having393 and 821 mg in the first 24 hr. Microscopic hematuria

with large numbers of casts was also present. Immuno-fluorescent study of the 4 hr renal biopsies revealedheavy linear deposits of equine Ig along the GBMin allfour monkeys (Fig. 4) with minimal C3 deposits in thesame distribution in only one. Spleen and liver biopsiesat 4 hr and autopsy specimens at 6 and 12 hr revealedheavy accumulation of equine Ig along the splenicreticulum and hepatic sinusoids (Fig. 2). The levels ofSGOTand SGPT were elevated during the first 24 hrafter administration of this ALG in two of three monkeysstudied.

Hemolytic C activity in the two monkeys studied wasreduced 70% within 10 min of injection and reachedeven lower levels at 2, 4, and 24 hr. To explain this de-crease in circulating C activity with virtual absence ofdetectable C3 tissue localization, the ALG was mixed invarying dilutions with normal human serum. Nativeand deaggregated ALG H-35 (Pu2) diminished hemo-lytic C activity 75 and 70% respectively in a 1: 5 dilu-tion in normal serum, the same dilution as was achievedin the monkeys. The glomerular-bound GBMRAbfromthis ALG could be shown to fix C in tissue sections byindirect immunofluorescence. These findings suggestedthat the ALG might produce immediate damage to theGBMwithout significant C localization. The lack of Cfixation was not because equine antiserum was unableto fix C but apparently was caused by a reduction ofavailable C by another effect of the ALG.


Histologic observations within the first 12 hr afterALG administration revealed a variable accumulationof polymorphonuclear (PMN) leukocytes in the glo-meruli; a maximum of 10 per glomerulus was seen in theonly monkey having trace amounts of glomerular C3.PMNleukocytes were also present in hepatic sinusoids,splenic pulp, and alveolar septa. All these areas con-tained equine Ig by immunofluorescence.

Kidney biopsies, taken after circulating monkey anti-equine Ig antibody was detectable, revealed linear de-posits of monkey IgG and C3 in addition to equine Igalong the GBM. Histologic sections taken at this timerevealed a proliferative glomerular lesion with apparentthickening of the GBM. PMNleukocyte accumulationin the glomeruli was variable. Proteinuria, which wascontinuously present, declined initially but rose againcoincident with the accumulation of monkey IgG and C3and corresponding to the autologous (delayed) phase ofthe nephrotoxic nephritis.

At autopsy of the two monkeys, 17 and 58 days afterinjection, equine Ig was found on the GBM, hepaticsinusoids and splenic reticulum.

DISCUSSION

The identification of potentially nephritogenic GBMRAbin human ALG preparations similar to that responsiblefor anti-GBM antibody mediated glomerulonephritis(20) is not surprising and is consistent with observa-tions of GBMRAbin rabbit anti-rat thymus serum stud-ied by Guttmann, et al. (6). Binding of equine Ig tohuman homograft glomeruli has been reported (21, 22).Vascular elements in many tissues are capable of induc-ing nephrotoxic antibodies when injected into animals(23). Katz, Unanue, and Dixon reported the nephrito-genic properties of rabbit antibody produced by injec-tions of rat splenic homogenate (24). This anti-spleenantibody possessed properties similar to classic Masuginephritogenic antibodies and was capable of inducingboth the heterologous and autologous phases of nephro-toxic nephritis in rats. The anti-spleen antibody had 90%cross-reactivity with kidney but differed from anti-kidney antibody by disappearing from the kidney twiceas rapidly.

The three GBMRAbpositive ALG preparations ad-ministered to the patients reported here were induced byinjection of cells obtained from solid lymphoid organs.Of 20 horses injected with these lymphocytes, 13 madeantibodies capable of binding to the GBM. In four prep-arations studied in detail, fixation to GBMwas blockedby prior absorption with human GBMbut not by lympho-cytes, suggesting that the GBMRAbin these prepara-tions were directed toward basement membrane antigenscontaminating the inoculum and not a cross-react-ing anti-lymphocyte antibody. The lack of GBMRAb

in ALG induced by injection of thoracic duct cells orcultured lymphoblasts supported this assumption. TheGBMRAb-positive ALG preparations formed a precipi-tin line of identity with a sheep anti-human GBManti-body when reacted with collagenase-solubilized GBMdemonstrating further the specificity of the reaction. Fi-nally, two rabbit anti-human GBMsera did not possessdetectable lymphocytotoxicity. (D. Kallestad, KallestadLaboratories, Minneapolis, Minn.)

GBMRAbin equine anti-human thymic serum (H-35(Pu2)) were shown to produce progressive and fatal glo-merulonephritis in monkeys after repeated injections(19). These antibodies were also capable of causingimmediate glomerular injury accompanied 7-10 dayslater by increased proteinuria at the onset of the autolo-gous phase of nephrotoxic nephritis in monkeys. The im-mediate injury was produced in the virtual absence ofdetectable C localization in the glomeruli presumablybecause of the decomplementation caused by the ALGrather than inability of the equine Ig to fix C. Themeans of this decomplementation is the subject of an-other investigation and may involve an anti-complementantibody.

The pathologic significance of the GBMRAbgivenpatients is difficult to assess. The renal homograft ofJ.N. lost function after ALG administration (intramus-cularly) containing 19 mg of GBMRAb(as quantitatedin rats) to treat a third rejection episode. The glomer-ular pathology might have been associated with eitherrejection or GBMRAb.Any nephritogenic effect of thisGBMRAbwas presumably reduced by virtue of theintramuscular route of administration. Local absorptionof GBMRAbgiven intramuscularly may explain both theinfrequent observation of GBM-bound foreign Ig (25)and the severe local reactions (5, 26).

Amounts of GBMRAbin the ALG given intrave-nously to the two liver homograft recipients (27) wereestimated from the rat isotopic studies. Patient G. B. re-ceived a total of 2.8 mg of GBMRAb20 days beforedeath; patient C. V. received 2.0 mg, 165 days and 0.5mg, 4 days before death. Only minimal glomerular al-terations were observed in the kidneys of these patientsat autopsy. Any nephritogenic effects of the GBMRAbmight have been modified by the concomitant use ofchemical immunosuppressive regimens in both patients.The rather broad basement membrane reactivity of someALG preparations with quantitatively greater hepaticthan renal fixation should be considered in evaluatinghepatic injury as was seen in the above two cases.

APPENDIXCase G. B. This 42 yr old Caucasian male underwent total

hepatectomy and hepatic homotransplantation for unresect-able bile duct carcinoma. His initial immunosuppression pro-gram consisted of azathioprine, prednisolone, and ALG, 15-


20 ml daily intravenously. Hepatic dysfunction persistedwith elevations in SGOT, alkaline phosphatase, and totaland conjugated bilirubin in spite of increased amounts of allthe above immunosuppressive agents in addition to localradiation therapy (100 R daily for 5 days). The patientexpired 37 days after transplantation from bilateral Gram-negative pneumonia. No proteinuria or elevation in serumBUNor creatinine were observed. During the posttransplan-tation period, the patient received the following amounts ofALG: H-12a-72.5 ml, H-17-285.0 ml, H-SOa-92.5 ml, andH-54a-55.0 ml.

Case C. V. This 27 yr old Caucasian female underwenthepatectomy and hepatic homotransplantation for a primaryhepatoma. The patient was preterminal at time of trans-plantation after chemotherapeutic and radiation therapy forthe tumor. Transplantation immunosuppressive therapy con-sisted of azathioprine, prednisolone, and intravenous ALGdaily for 6 wk supplanted by varying doses and schedules.Persistent elevations of total and conjugated bilirubin,SGOT, and alkaline phosphatase were noted after transplan-tation. Her condition stabilized sufficiently 3 months aftertransplantation to allow discharge on azathioprine, predniso-lone, and ALG, until her death occurred in hepatic comaand terminal bronchopneumonia 170 days after transplanta-tion. No proteinuria or elevation in serum BUN or creati-nine were observed in the postoperative period. During theposttransplantation period, the patient received the followingamounts of ALG: H-11-45 ml, H-12a-22.5 ml, H-16-6.25 g,H-17-100 mg, H-50a-15 ml, H-53a-183 ml, H-53b-125 ml,and H-58-11.8 g. Two additional ALG preparations notavailable for study were also given in 15 and 52.5 mlamounts.

Case J. N. This 23 yr old Caucasian female had a 10 yrhistory of glomerulonephritis culminating in renal failurerequiring hemodialysis for 2 months. She underwent bilateralnephrectomy with concurrent insertion of a renal homo-graft obtained from her brother (Teraski C minus match).The kidney functioned promptly with a creatinine clearanceof 81 ml/min on the 2nd postoperative day. Her immuno-suppressive regimen consisted of azathioprine, prednisone,and intramuscular ALG. The dosage of ALG was 2 mldaily for 5 days pre- and post-transplantation and then onalternate days. Rejection episodes occurred on the 3rd and28th postoperative days and responded to increased steroids,local radiation and, actinomycin C. A 3rd rejection episodewith increased proteinuria (500 mg/24 hr) occurred on the63rd posttransplantation day. It was elected not to increasethe prednisone but to treat the patient with high doses ofintramuscular ALG H-lc-4 ml, plus 2-4 ml of ALG R-1daily for 4 wk. Renal function improved over the next 10days and then deteriorated. Proteinuria remained unchanged.The kidney was removed 19 wk after transplantation atwhich time the creatinine clearance had fallen to 8 ml/min.The patient received ALG: H-lc-190 ml and approximately100 ml of ALGR-1.

ACKNOWLEDGMENTSWe wish to acknowledge the kind cooperation of the fol-lowing investigators for providing the ALG preparationsused in this study: Dr. G. Gray, Hypersensitivity DiseasesResearch, The Upjohn Company, Kalmazoo, Mich.; DoctorsK. Heide and F. Seiler, Behringwerke AG, Marburg, Ger-many; Mr. Donald Kallestad, Kallestad Laboratories, Inc.,Minneapolis, Minn.; Dr. D. E. Kayhoe, National Instituteof Allergy and Infectious Diseases, National Institutes of

Health, Bethesda, Md.; Dr. G. S. LaFontaine, Bionetics Re-search Laboratories, Bethesda, Md.; Dr. J. S. Najarian, De-partment of Surgery, University of Minnesota MedicalSchool, Minneapolis, Minn.; Dr. T. E. Starzl, Departmentof Surgery, University of Colorado Medical Center, Denver,Colo.; Dr. H. E. Taylor, Medical Research Council, Ot-tawa, Canada; and Dr. E. R. Unanue, Department of Path-ology, Harvard Medical School, Boston, Mass.

This research was supported by United States PublicHealth Service Grant AI-07007, United States Public HealthService Contract PH 43-68-621, and United States AtomicEnergy Commission Contract AT (04-3)-410.

REFERENCES1. Woodruff, M. F. A., and N. A. Anderson. 1962. Effect

of lymphocyte depletion by thoracic duct fistula andadministration of anti-lymphocyte serum on the survivalof skin homografts in rats. Nature (London). 200: 702.

2. Lance, E. M. 1968. The effects of chronic ALS ad-ministration in mice. In Advance in TransplantationProceedings of the First International Congress of theTransplantation Society. J. Dausset, J. Hamburger, andG. Mathe, editors. Munksgaard, Copenhagen S. 107.

3. Iwasaki, Y., K. A. Porter, J. R. Amend, T. L. Marchi-oro, V. Zuhlke, and T. E. Starzl. 1967. The preparationand testing of horse anti-dog and anti-human anti-lymphoid plasma or serum and its protein fractions.Surg. Gynecol. Obstet. Int. Abstr. Surg. 124: 1.

4. Cohen, B. J., M. J. de Vries, M. J. van Noord, andF. H. Lubbe. 1970. Strain-specific renal toxicity ofheterologous anti-lymphocyte 'y-globulin in mice. Trans-plantation. 10: 1.

5. Monaco, A. P., M. L. Wood, and P. S. Russell. 1967.Some effects of purified heterologous anti-human lym-phocyte serum in man. Transplantation. 5: 1106.

6. Guttmann, R. D., C. B. Carpenter, R. R. Lindquist, andJ. P. Merrill. 1967. Renal transplantation in the inbredrat. III. A study of heterologous antithymocyte sera.J. Exp. Med. 126: 1099.

7. Guttmann, R. D., C. B. Carpenter, R. R. Lindquist, andJ. P. Merrill. 1967. Treatment with heterologous anti-thymus sera: nephritis associated with modification ofrenal allograft rejection and the immune status of thehost to the foreign protein. Transplantation. 5: 1115.

8. Lindquist, R. R., R. D. Guttmann, C. B. Carpenter, andJ. P. Merrill. 1969. Nephritis induced by anti-lympho-cyte serum. Transplantation. 8: 545.

9. Diethelm, A. G., G. J. Busch, J. M. Dubernard, W.McN. Orr, R. J. Glassock, A. G. Birtch, and J. E.Murray. 1969. Prolongation of canine renal allografts byhorse anti-dog lymphocyte serum and globulin. Ann.Surg. 169: 569.

10. Orr, W. McN., A. G. Birtch, A. G. Diethelm, J. M.Dubernard, J. Duquella, and R. J. Glassock. 1970. Astudy of the potential nephrotoxicity of heterologousanti-lymphocyte serum. Clin. Exp. Immunol. 6: 305.

11 Coons, A. H., and M. H. Kaplan. 1950. Localization ofantigen in tissue cells. II. Improvements in a methodfor the detection of antigen by means of fluorescentantibody. J. Exp. Med. 91: 1.

12. Wilson, C. B., and F. J. Dixon. 1970. Antigen quantita-tion in experimental immune complex glomerulonephritis.I. Acute serum sickness. J. Immunol. 105: 279.

13. Pressman, D., E. D. Day, and M. Blau. 1957. The useof paired labeling in the determination of tumor-local-izing antibodies. Cancer. Res. 17: 845.


14. McConahey, P. J., and F. J. Dixon. 1966. A method oftrace iodination of proteins for immunologic studies.Int. Arch. Allergy Appl. Immunol. 29: 185.

15. Krakower, C. A., and S. A. Greenspon. 1951. Localiza-tion of the nephrotoxic antigen within the isolated renalglomerulus. Arch. Pathol. 51: 629.

16. Spiro, R. G. 1967. Studies on the renal glomerular base-ment membrane. Preparation and chemical composition.J. Biol. Chem. 242: 1915.

17. Mayer, M. M. 1967. Complement and complement fixa-tion. In Experimental Immunochemistry. E. A. Kabatand M. M. Mayer, editors. Charles C. Thomas, Spring-field, Ill. 2nd edition. 133.

18. Balner, H., U. P. Eysvoogel, and F. J. Cleton. 1968.Testing of anti-human lymphocyte sera in chimpanzeesand lower primates. Lancet. 1: 19.

19. Taylor, H. E. 1970. The occurrence of glomerular base-ment membrane reactive antibodies in anti-human anti-thymocyte globulin. Transplant Proc. 2: 413.

20. Lerner, R. A., R. J. Glassock, and F. J. Dixon. 1967.The role of antiglomerular basement membrane anti-body in the pathogenesis of human glomerulonephritis.J. Exp. Med. 126: 989.

21. Carpenter, C. B., A. Rashid, A. I. Katz, G. J. Busch,A. L. Sheffer, M. D. Valentine, A. G. Birtch, R. E.Kay, K. F. Austen, and J. P. Merrill. 1968. Adversereactions to the use of horse anti-human lymphocyte

globulin (ALG) in man. American Society NephrologyMeeting, Washington, D. C., 25-26 November 1968.

22. Thiel, G., J. Moppert, J. Mahlich, F. Buhler, T. Vischer,F. Enderlin, H. Weber and H. U. Zollinger. 1971.Glomerular damage after intravenous administration ofantilymphocyte globulin (ALG) in man and rhesusmonkey. Transplant Proc. 3: 741.

23. Unanue, E. R., and F. J. Dixon. 1967. Experimentalglomerulonephritis: Immunological events and patho-genetic mechanism. Advan. Immunol. 6: 1.

24. Katz, D. H., E. R. Unanue, and F. J. Dixon. 1967.Nephritogenic properties of cross-reacting kidney-fixingantibodies to heart, spleen and muscle. J. Immunol. 98:260.

25. Traeger, J., D. Fries, J. P. Revillard, J. Brochier, andN. Brunat-Blanc. 1969. Antilymphocyte globulins in kid-ney transplantation: effects on glomerular disease. Trans-plant Proc. 1: 1006.

26. Kashiwagi, N. B. A. B., L. Brettschneider, C. G. Groth,and T. E. Starzl. 1968. Clinical reaction and serologicchanges after the administration of heterologous anti-lymphocyte globulin to human recipients of renal homo-grafts. Ann. Intern. Med. 68: 275.

27. Fortner, J. G., M. H. Shiu, H. Balner, C. B. Wilson, G.Sichuk, N. Kawano, J. T. Holmes, and E. J. Beattie,Jr. 1971. Observations on prolonged immune suppressionfor human liver homografts. Transplant Proc. 3: 383.


glomerular basement...

Documents