| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
-Galactosyl Antibody with Pig Tissues
From the Departments of Physiology,*
Surgery,
Microbiology,
and
Pathology,§
College of Physicians and Surgeons
of Columbia University, New York, New York; Laboratorio
Elettromicroscopia,¶
Ospedale S. Gallicano, Rome,
Italy; the Department of Biological and Medical
Research,||
King Faisal Specialist Hospital and Research
Centre, Riyadh, Saudi Arabia; and the Department of Microbiology and
Immunology,**
Allegheny University, Hahnemann
School of Medicine, Philadelphia, Pennsylvania
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-Galactosyl (Gal) and anti-porcine aortic
endothelial cell antibodies, similar to human natural
xenoantibodies and reactive with epitopes of thyroglobulin,
laminin, and heparan sulfate proteoglycans. When injected into
pigs with a protocol similar to that used in the rat to show the
nephritogenic potential of heterologous anti-laminin and anti-heparan
sulfate proteoglycan antibodies, baboon immunoglobulins bound
first to renal vascular endothelium, and later to interstitial
cells, especially fibroblasts and macrophages, and to
antigens in basement membranes and extracellular matrix, where
they colocalized with laminin- and heparan sulfate
proteoglycan-antibodies, and with bound Griffonia
simplicifolia B4. A similar binding was observed in other organs.
The pigs did not develop an acute complement-dependent
inflammation, but rather chronic lesions of the basement
membranes and the extracellular matrix. Incubation of renal fibroblasts
with baboon anti--Galactosyl antibodies resulted in increased
synthesis of transforming growth factor-ß and
collagen, suggesting a possible basis for the fibrotic
response. The results demonstrate that in this experimental model a
consequence of Gal antibody interaction with porcine
tissues, is immunoreactivity with Gal on matrix molecules
and interstitial cells, priming mechanisms leading to fibrosis
resembling that in chronic allograft rejection. The possibility that
similar lesions may develop in long-surviving pig xenografts is
discussed.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1–3Galß1–4 GlcNAc-R
(Gal) epitope5-7
present on glycoproteins and
glycolipids on pig endothelial cells.7
The rapid and destructive nature of unmodified xenograft rejection does
not allow determination of whether the antibodies that interact with
endothelium can bind and induce injury to other cells and extracellular
compartments. To obtain such additional information we have prepared in
baboons antibodies to Gal/bovine serum albumin (BSA) and to
porcine aortic endothelial cells (PAEC), termed baboon anti-Gal and
baboon anti-PAEC, respectively, similar to human xenoantibodies, and
reactive with epitopes of laminin and heparan sulfate proteoglycans
(HSPG). We injected these antibodies into pigs and studied the
consequences.
The results show that the baboon antibodies with Gal reactivity bind
first to endothelium, erythrocytes, and platelets, and later to Gal
epitopes on fibroblasts and macrophages and to the extracellular
matrix, especially to laminin and heparan sulfate proteoglycans, and
that the pigs develop basement membrane and fibrosclerotic lesions.
Incubation of pig fibroblasts with anti-Gal antibodies increased
their synthesis of transforming growth factor ß (TGF-ß) and
collagen, suggesting a basis for the fibrotic response. Therefore, in
this experimental model baboon antibody with Gal reactivity binds to
basement membranes and extracellular matrix, setting off mechanisms
that give rise to local fibrosclerotic lesions.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Baboons (Papio anubis), 25 to 31 kg body weight, were obtained from the Biomedical Research Foundation (Houston, TX) and from LEMSIP, New York University (Tuxedo, NY). Miniature pigs six week old were purchased from Harlan Sprague Dowley, Sinclair Research, Inc. (Columbia, MO) and used when their body weight was 3 to 13 kg. The protocol for animal experiments was formally approved by the Columbia University Review Board.
Antigen Preparation and Immunization
Four healthy baboons (two naive and two that had previously
received heterotopic pig heart transplants, subsequently removed) were
immunized with 1 mg Gal conjugated to BSA (14 atom spacer, termed
Gal/BSA; Dextra Laboratories, Reading, UK) in an equal volume of
incomplete Freund’s adjuvants (IFA) (Sigma) at multiple s.c.
and intradermal sites, boosted after 3 weeks by subcutaneous 500 µg
of Gal/BSA in IFA. A fifth (naïve) baboon was immunized with
3–4 million PAEC in IFA, and subsequently boosted as given above for
Gal/BSA. After the first bleeding, the baboons received booster
injections every 4 weeks and were bled every 6 weeks (<10% of blood
volume). The same protocol was used to immunize a sheep with BSA and
pigs with baboon 
-globulins. Anti-rabbit angiotensin-converting
enzyme (ACE) antibody was prepared in goats as previously
described.8
Pre-immune sera from naïve baboons,
normal baboon sera, and Gal antibody-negative sera from normal
rabbits and pigs were used as controls. The -globulin fractions were
obtained by precipitation in 50% ammonium sulfate; IgG was isolated
using ImmunoPure immobilized protein A column (Pierce, Rockford, IL).
The flow-through, devoid of IgG, was used as baboon anti-Gal IgM.
Total protein concentration was measured by a standard Bradford protein
assay (BioRad, Hercules, CA). The concentration of -globulins, IgG,
and IgM was determined by radial immunodiffusion,9
and that
of pig anti-baboon antibody in the sera of pigs injected with baboon
-globulin by Ouchterlony’s immunodiffusion. Gal antibody was
affinity purified from baboon anti-Gal serum on a 0.2 ml
Gal-conjugated silica beads column (Synsorb 115, Chembiomed,
Edmonton, AB, Canada) as previously described.10
Titers of
adsorbed and subsequently eluted antibody and of the flow-through were
determined by enzyme-linked immunosorbent assay (ELISA) on mouse
laminin and PAEC. Silica beads conjugated with Galß1–4-GlcNac-R
epitopes were used as specificity control. Other baboon anti-Gal
aliquots were depleted of Gal reactivity by extensive absorption
with -Gal/BSA and with Gal-rich rabbit
erythrocytes7,10
and were also tested by ELISA.
Other Antibodies, Lectins, Components of the Extracellular Matrix, and Chemicals
All conjugated and some unconjugated affinity-purified antibodies were purchased from Sigma Chemical Co. (St. Louis, MO), except biotin-conjugated rat anti-mouse IgG (H+L). AP-conjugated goat anti-human IgG and AP-conjugated streptavidin which were from Zymed Laboratories, Inc. (South San Francisco, CA). Sheep anti-human C3c, and sheep anti-pig C3 were obtained from The Binding Site (San Diego, CA), and FITC-conjugated rabbit anti-human C3c from Dako (Carpinteria, CA). Anti-porcine IgA monoclonal antibody was obtained from Serotec (Oxford, UK). Rabbit anti-bovine type II collagen, rabbit anti-bovine type II collagen, and monoclonal antibodies to HSPG core protein (Perlecan), to rat fibronectin, and to human ß3 integrin, were purchased from Chemicon International, Inc. (Temecula, CA), and anti-mouse Englebert-Holm-Swerm (EHS) HSPG from Seikagaku (Tokyo). Biotin- and FITC-conjugated Bandeireea griffonia simplicifolia isolectin B-4, and FITC-avidin D were from Vector Laboratories (Burlingame, CA). The rabbit anti-rat basement membrane HSPG was a gift of Dr. Marilyn G. Farquhar.11 All polyclonal antisera to human proteins were crossreactive with baboon proteins in double immunodiffusion techniques. When appropriate, antibody and antisera were absorbed with pig serum before use.
Biocoat matrigel thin layer plates were from Becton Dickinson (Franklin Lakes, NJ). Human plasma fibronectin, bovine plasma fibronectin, mouse laminin from mouse EHS tumor, laminin from human placenta, HSPG from mouse EHS tumor, bovine chondroitin sulfate A, bovine hyaluronate, bovine type I, II, and III collagen, pig type I and II collagen, chicken type II collagen, and porcine thyroglobulin were all obtained from Sigma, and bovine type III collagen from Chemicon International (Temecula, CA). The 54-kd rabbit tubular basement membrane nephritogenic antigen12 was a gift of Drs. Butkowski and Charonis.
Agarose anti-human IgM (µ chain-specific), biotin (long arm) NHS
(Blanks) N-hydroxysuccininidyl-6-(biotinamide)hexanoate,
o-phenyledadiamine dihydrochloride, 3-3'-dimethoxybenzidine,
ovalbumin, pig albumin, BSA, pig thyroglobulin, bovine thyroglobulin,
protease inhibitors cocktail, and p-nitrophenylphosphate,
-galactosidase, and ß-galactosidase, and collagenase type VII were
obtained from Sigma. Other reagents were Accray Assay Human IgG
RID Kit, Accray Assay Human IgM RID kit, Micro-Ouchterlony Kit (ICN,
Costa Mesa, CA) and nitrocellulose membranes, gels, and Tris-glycine
(BioRad).
Blood Counts, Chemistry, and Urinalysis
Peripheral erythrocytes, lymphocytes, and platelets were enumerated, the glucose, BUN, creatinine, and CH50 measured, and the clotting profile was studied by the Bioveterinary Services Department of Roche Biomedical Laboratories (Raritan, NJ). Urinary protein excretion and sediments were examined using conventional methods.
Cultures of PAEC and Fibroblasts from the Renal Medulla
Porcine aortas and kidneys were obtained from a local abattoir. PAEC8 and fibroblasts13 were prepared and characterized by methods previously described. For fibroblasts, the inner stripe of the renal medulla was dissected in sterile conditions, minced, passed through a 106-µm mesh (Fisher Scientific, Pittsburgh, PA), and cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal calf serum (FCS).
Endothelial Cell Cytotoxicity
PAEC (third passage) were incubated at 37°C for 1 hour in 10
µmol/L Calcein AM (Molecular Probes, Eugene, OR), washed, and
incubated for 1 hour at 37°C with baboon anti-Gal or baboon
anti-PAEC -globulin or IgG fractions plus fresh or
heat-decomplemented rabbit serum. The supernatants were collected and
the absorbance was measured at 485 nm in a CytoFluor multi-well
fluorescent-plate reader (PerSeptive Biosystems, Framingham, MA).
Specific lysis was calculated subtracting the dye released by cells
incubated with the antibody and complement from that released from
cells incubated with phosphate buffered saline (PBS), divided by
the value of maximal dye release from cells incubated with 1%
saponin.
ELISA
Gal reactivity on immobilized Gal/BSA, mouse laminin,
porcine thyroglobulin, and single components of the extracellular
matrix was measured by ELISA according to Engvall and
Perlmann.14
Binding of antibody to PAEC was studied as
described by Platt et al.15
The reactivity of the antibody
with murine extracellular matrix was determined using Matrigel plates
(Becton Dickinson). The expression of Gal epitopes on porcine
cells or tissues was measured by inhibition ELISA with anti-Gal
monoclonal antibody (M86) prepared in 1–3,galactosyltransferase
knockout mice immunized with rabbit erythrocyte
membranes.16
Controls were normal rabbit and pig serum,
sheep anti-BSA serum, the baboon anti-Gal flow through of
Gal immunoadsorption column, and baboon anti-Gal absorbed with
Gal/BSA, or rabbit erythrocytes.
Western Blot Analysis
Sixty µg PAEC lysates, and 6 µg of single components of the
extracellular matrix were subjected to sodium dodecyl
sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) under reducing
conditions on 4–15% gradient gels.17
The proteins were
transferred to nitrocellulose membranes incubated with primary antibody
at concentrations indicated in the legends of Figures 1C and 2C
, followed by the appropriate
secondary biotin-conjugated antibody and AP-conjugated streptavidin.
Control reagents were as for ELISA.
|
|
The presence of antibodies against baboon IgG in sera of pigs
injected with baboon -globulin preparations was determined using
double immunodiffusion.
PAl-1/Luciferase (PAIL) Assay for TGF-ß on Supernatants of Cultured Renal Fibroblasts
Fibroblasts (8th–10th passage) were placed into 6-well plates,
5 x 105
cells/well, and cultured in DMEM with 5%
FCS. After seeding, the fibroblasts were incubated for 24 hours in DMEM
with 0.5% FCS, followed by incubation for 60 hours in DMEM with 0.5%
FCS containing 1 mg/ml of the -globulin preparation. The
supernatants were collected and assayed for TGF-ß.18
Briefly, mink lung epithelial cells (MLECs-clone 32) transfected with a
fragment of the human plasminogen activator inhibitor-1 gene fused to
the firefly luciferase reporter gene from Dr. Daniel Rifkind
(Department of Cell Biology, New York University Medical Center) were
plated and incubated with test samples for 14 hours. Luciferase
activity was measured using the Luciferase Assay System E1500 (Promega,
Madison, WI) and Lumat LB 9501 (Wallac, Gaithersburg, MD). Standard
curve was with TGF-ß1 (R&D Systems, Minneapolis, MN).
Quantitation of Collagen Produced by Cultured Renal Fibroblasts
Collagen synthesis was quantified by proline
incorporation.19
Fibroblasts were prepared and made
quiescent as described in the previous paragraph, and incubated for 24
hours in DMEM/0.5% FCS, 50 µg/ml ascorbic acid, and 1 mg/ml of the
-globulin preparation. This medium was replaced for 12 hours with
DMEM/0.5% FCS, 50 µg/ml ascorbic acid, 50 µg/ml
ß-aminoproprionitrile, 1 mg/ml of the -globulin preparations, and
100 µCi/ml of L-[2,3,4,5-3H] Proline (Amersham, Little
Chalfont, UK). The supernatant proteins were precipitated with
trichloroacetic acid (TCA), dissolved in 0.2 mol/L NaOH,
neutralized, and incubated with or without 27 IU of collagenase VII
(Sigma), followed by the TCA precipitation. The radioactivity of
supernatant and pellet were counted using a liquid scintillation
counter 1219 Rackbeta (Wallac). Collagen synthesis was calculated on
the assumption that the proline content of collagen is 5.4 times that
of other proteins.
Tissue Preparations for Morphology and Immunohistochemistry
Autopsy or biopsy samples of lung, kidney, heart, liver, and
intestine were immediately fixed in 10% buffered formalin to be
processed for light microscopy. Others were fixed in 2% glutaraldehyde
and processed for electron microscopy.20
For
immunohistochemistry, fresh frozen sections or formalin-fixed paraffin
sections were used. Frozen sections were stained or double-stained with
FITC- or TRITC-conjugated antibody. For direct staining, purified IgG
were conjugated with FITC, TRITC, or biotin, then extensively absorbed
with normal pig serum. For indirect immunohistochemistry the sections
were first incubated with primary antibody, followed by appropriate
FITC-, TRITC-, or biotin-conjugated secondary antibody. The sections
were examined in a Nikon epifluorescence and phase contrast microscope,
or a Zeiss LSM 410 laser scanning confocal microscope. Controls were
done with baboon anti-Gal absorbed with Gal/BSA and rabbit
erythrocytes, sheep anti-BSA, substitution of PBS for the
primary antibody and digestion of tissue sections with - or
ß-galactosidase (1 U -galactosidase and 2 U ß-galactosidase in
100 mmol/L NaCl, 50 mmol/L sodium acetate, pH 5.0 at 37°C for 2
hours; control was buffer without enzyme). The ability of baboon
anti-Gal, baboon anti-PAEC to fix complement was evaluated in
vitro by indirect immunofluorescence.21
Experiments in Living Pigs (Table 1)
Five -globulin preparations were used: baboon anti-Gal,
baboon anti-PAEC, two controls (sheep anti-BSA and goat anti-ACE, which
cross-reacts with PAEC), and pig anti-baboon. The baboon anti-PAEC was
used to determine whether antibodies with multiple endothelial
specificities induced lesions more severe than or different from baboon
anti-Gal. Four groups of pigs received immunoglobulins as follows:
Group I, (pigs 1 and 2), injected over 6 hours to visualize early
binding site; pig 3 was the ACE control. Group II, (pigs 4, 5, and 6),
infused with antibody for 72 hours to determine the effect of the
longest possible Gal antigen-antibody interaction before additional
heterologous protein would have induced serum sickness; pigs 7 and 8
were the ACE and BSA controls, respectively. Group III, (pigs 9, 11,
and 12) that were mononephrectomized, infused with antibody for 72
hours, and studied 30 to 120 days later for long-term consequences. To
enhance cross-linking of tissue-bound antibodies in the autologous
phase,22
pigs were immunized s.c. with baboon -globulin
in IFA 2 days before the infusion, on a schedule like that used with
rats to test the pathogenetic potential of anti-laminin23
and anti-HSPG24
antibodies. In pig 9, injected with baboon
anti-Gal, vascular permeability was enhanced by concomitant
injection of an amount of goat anti-ACE comparable to that inducing in
rabbits a minimal increase in glomerular permeability without anatomic
lesions20; as control, pig 10 received only the same small
dose of goat anti-ACE; pigs 13 and 14 were ACE and BSA controls for
pigs 9, 11, and 12. Group IV, included pigs 15 and 16 used as positive
control for serum sickness, and to study complement fixation after
passive transfer of pig anti-baboon -globulins, as described
elsewhere23; baboon anti-Gal plus baboon anti-PAEC was
infused for 72 hours followed, 3 days later, by pig anti-baboon
-globulin.
|
Statistical analyses used ANOVA and unpaired t-test. P < 0.05 was considered a statistically significant difference between values. All measurements were in triplicate and all assays were repeated at least three times with similar results.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The concentration of baboon anti-Gal, baboon anti-PAEC, and
goat anti-ACE -globulins was 30 to 40 mg/ml. The highest dilution
producing unequivocal immunofluorescence in pig tissue sections was
0.63 µg/ml. Sheep anti-BSA (300 µg/ml) was the control. The
concentration of pig anti-baboon -globulin was 20 mg/ml and its
staining titer, on sections of pig kidney containing deposits of baboon
-globulin, was 8 µg/ml.
Immunoreactivity on Gal/BSA was studied by ELISA; IgG titers were
increased 32- to 200-fold, and IgM titers were increased 2- to 32-fold
as compared to naive baboon or preimmune sera. Similar results were
obtained with Gal-bearing proteins, such as pig thyroglobulin and
mouse laminin. Sheep anti-BSA was not reactive (Figure 1A)
. Complementary assays were developed
to evaluate immunoreactivity and lysis of PAEC. The two -globulin
fractions had similar titers, which were 20 to 27 times higher than in
normal baboon sera (Figure 1B)
. Baboon anti-PAEC and baboon anti-Gal
induced 54- and 20-fold more cell lysis than normal baboon sera,
respectively, but none with heat-inactivated rabbit serum (not shown).
Baboon anti-Gal and baboon anti-PAEC also fixed complement on kidney
sections; sections incubated with the tested -globulin preparations
and rabbit or porcine serum, then stained for rabbit or pig C3,
revealed immune deposits in the brush border of proximal tubules, in
glomerular and tubular basement membranes, and in the extracellular
matrix, but not when heat-inactivated rabbit or porcine serum was used
(not shown). Immunoblotting of PAEC with baboon anti-Gal IgG, or
affinity-purified baboon anti-Gal IgG, displayed several
glycoproteins with apparent molecular weights of 70, 125–135,
and 225 (Figure 1C
, Lanes 1 and 6, respectively). Baboon anti-PAEC
displayed more intense staining of immobilized PAEC extract, and a
prominent 44-kd band, not observed with baboon anti-Gal (Figure 1C
,
Lane 2). As expected, pre-immune baboon serum showed a similar pattern
of immunoreactive bands, which was prevented by absorption with rabbit
erythrocytes (Figure 1C
, Lanes 4 and 5, respectively). Normal rabbit
serum revealed no bands (Figure 1C
, Lane 3).
Reactivity with the extracellular matrix was studied by ELISA on murine
basement membrane-like matrix of EHS tumor (Matrigel); baboon
anti-Gal, baboon anti-PAEC, and baboon pre-immune -globulins
showed reactivity, though baboon anti-Gal had the highest titer
(Figure 2A)
. The presence of Gal
epitopes in components of the extracellular matrix was also studied by
ELISA; mouse laminin, and mouse HSPG were most reactive and,
minimally, bovine fibronectin (Figure 2B)
. In immunoblotting Gal
antibody was bound to immobilized mouse laminin (Figure 2C
, Lane 1),
bovine fibronectin (Figure 2C
, Lane 4), and mouse HSPG (Figure 2C
, Lane
7), but not to human laminin and fibronectin (Figure 2C
, Lanes 3 and 6,
respectively). Absorption of baboon anti-Gal with Gal/BSA or
rabbit erythrocytes almost completely abolished this binding (Figure 2C
, Lanes 2, 5, and 8).
Binding of Baboon Anti-Gal and Baboon Anti-PAEC to Normal Pig
Tissues
On kidney sections as substratum, the staining titer of baboon
anti-Gal and baboon anti-PAEC IgG was 200-fold higher than that of
preimmune baboon IgG, which bound mainly to tubular brush border. The
staining pattern with baboon anti-PAEC was comparable to that with
baboon anti-Gal, although weaker in epithelial cells. The IgM
fractions of naïve pigs or pre-immune sera did not stain the
tissues, whereas those of baboon anti-Gal and baboon anti-PAEC
stained similarly to IgG, but weaker.
In kidney, lung, heart, liver, and intestine, dual fluorescence
confocal microscopy, colocalized Gal with laminin and HSPG in
the basement membranes. Fibronectin partially colocalized with Gal
in the basement membranes, but was mainly expressed in the septa and in
the media of the arteries, where Gal was not (or only weakly)
detected. Moreover, baboon anti-Gal bound to several epithelial
cells, fibroblasts, macrophages, and chondrocytes, whereas baboon
anti-PAEC bound to the same cells and structures, but provided a
stronger staining of endothelial cells (not shown).
The specificity of baboon anti-Gal and baboon anti-PAEC for Gal
was shown by their colocalization with Griffonia
simplicifolia B4, and by examination of sections predigested with
- or ß-galactosidase; digestion with -galactosidase, but not
with ß-galactosidase, abolished staining. In contrast,
-galactosidase digestion only partially reduced the staining
capacity of baboon anti-PAEC. Moreover, absorption with Gal/BSA and
Gal-rich rabbit erythrocytes abolished the ability of baboon
anti-Gal to stain and reduced that of baboon anti-PAEC. Sheep
anti-BSA did not stain (not shown).
Clinical Observations after Antibody Infusion
Shortly after the beginning of the injection of baboon
anti-Gal, baboon anti-PAEC, and goat-anti-ACE, the pigs developed
episodes of tachypnea, agitation, and occasionally, pulmonary edema.
With the exception of pigs 6 and 16 (which died at the end of day 4 and
after beginning of passive transfer of pig anti-baboon
-globulin, respectively) these respiratory crises were overcome by
decreasing or stopping the delivery of antibody. Infusion of sheep
anti-BSA (pigs 8 and 14) did not induce respiratory distress. In pigs
injected with baboon anti-Gal and baboon anti-PAEC the signs of
pulmonary distress were associated with decreased number of circulating
erythrocytes (23% pig 1, 84% pig 4), leukocytes (72% pig 1 and 60%
in pig 6), and platelets (83% pig 4). In contrast, blood cell counts
remained normal in pig 8, injected with sheep anti-BSA.
CH50 decreased 35% in pig 6, injected with baboon
anti-PAEC, 65% in pig 4, injected with baboon anti-Gal, and 100%
in pig 7, injected with goat anti-ACE. After day 4, slight proteinuria
(+ to ++) developed in all pigs injected with baboon anti-Gal,
baboon anti-PAEC, and goat anti-ACE, but not with sheep anti-BSA. Serum
creatinine, BUN, and liver enzymes remained normal in all animals.
Salient features of results for each group are described below.
Group I
Pigs Injected for 6 Hours with Baboon Anti-Gal (Pig 1) or
Baboon Anti-PAEC (Pig 2)
Lungs showed congestion and patchy edema. Some alveolar
capillaries were occluded by aggregated erythrocytes and platelets.
Baboon IgG, but not IgM, was bound to endothelium and to platelets
(Figure 3A)
. Pig C3 was present only in
aggregated platelets. Kidneys showed only aggregated platelets and
erythrocytes in capillaries and venules. Baboon IgG, but not IgM,
stained the endothelium of peritubular capillaries, the glomerular
capillary walls, and, less intensely, tubular basement membranes
(Figure 3, B and C)
. Pig C3 was deposited only in glomeruli of the pig
injected with baboon anti-PAEC. Pig IgG was absent. In other organs
aggregated erythrocytes and platelets were in the capillaries, and
baboon IgG was deposited on endothelium.
|
In lungs there was severe edema, hemorrhage, and accumulation of inflammatory cells. Fine granular deposits of goat IgG and pig C3 were in the alveolar endothelium, and fibrin thrombi occluded the capillaries. The kidneys had granular deposits of ACE and goat IgG in glomerular and some tubular basement membranes.
Group II
Pigs Injected for 72 Hours with Baboon Anti-Gal (Pig 4)
or Baboon Anti-PAEC (Pigs 5 and 6)
Circulating antibodies to pig IgG were not detectable. In
lungs there was alveolar, subpleural, and perivascular edema and focal
hemorrhage, greater after baboon anti-PAEC (Figure 3, J and K)
.
Mononuclear cells and occasional erythrocytes were present in the
bronchial walls and lumens. Both baboon anti-Gal and baboon
anti-PAEC bound at the same sites, but baboon anti-PAEC stained
endothelia more strongly. By confocal microscopy baboon IgG
co-localized with laminin in pulmonary basement membranes and
peribronchial extracellular matrix; it was also bound to alveolar
epithelium, capillaries and venules in the visceral pleura, alveolar
septa, peribronchial vessels and plasma membrane of bronchial
epithelial cells, chondrocytes, and matrix of the cartilage. Baboon IgG
and Gal colocalized in alveolar macrophages and in peribronchial and
endobronchial mononuclear cells, presumably monocyte/macrophages
(Figure 3, F–H)
. The kidneys were morphologically normal. There were
deposits of baboon IgG in glomerular basement membrane and at the
surface of fibroblast-like cells in the interstitium of the medulla,
and lesser ones in tubular basement membranes and basolateral
compartments of proximal tubules, but, surprisingly, not in the brush
border of proximal tubules, which contain large amount of Gal.
Baboon IgG colocalized with Gal in the walls of sinusoids and
hepatic septa; in the intestine in the lamina propria of villi, the
basement membranes and the lamina propria of choroid plexus, ciliary
body, aortic endothelium, and media were also stained.
Pig Injected for 72 Hours with Goat Anti-ACE (Pig 7)
Changes in lungs were similar to but milder than in the pig injected with goat anti-ACE for 6 hours (Group I, pig 3). In kidney, glomeruli had minimal morphological changes but, by immunofluorescence, diffuse granular deposits of ACE and goat IgG were seen in the capillary walls, corresponding to discrete dense deposits in the filtration slits, by electron microscopy.
Pig Injected for 72 Hours with Sheep Anti-BSA (Pig 8)
All organs appeared normal.
Group III
Pigs Injected for 72 Hours with Baboon Anti-Gal (Pigs 9 and 11)
and Baboon Anti-PAEC (Pig 12), Sacrificed 4 to 17 Weeks Later
All pigs had low levels of circulating antibodies against baboon
IgG. The lesions were more severe in pig 9, to which a small dose of
goat anti-ACE was given in order to make extravascular Gal more
accessible to antibody. In the lungs of pig 9 there was severe
fibrosclerosis around bronchioles, bronchi, medium-sized vessels, and
in interlobular septa. The alveolar basement membrane was focally
thickened, the alveolar septa enlarged, with collagen fibrils (Figure 4, A–D)
. Alveolar basement membranes
were similarly affected in pigs 11 and 12. In sclerotic areas the
perineurium of nerves was thickened and collagenous (Figure 4H)
. Baboon
IgG was bound to the alveolar capillary walls, alveolar septa,
peribronchial matrix, bronchial capillaries, perineurium, and
endoneurium (Figure 4, F and G)
. Some deposits of pig IgA were present
in the bronchi, but there was no baboon IgM or pig C3. Deposits of pig
IgG were absent or minimal and focal. All fibrotic tissues were stained
by anti-type I collagen (Figure 4E)
, and less intensely, by anti-type
III collagen antibody. In kidneys of all three pigs, glomeruli had
increased mesangial matrix with distorted or collapsed capillary walls,
glomerulo-capsular adhesions, and focal sclerosis. The glomerular
basement membranes were focally thickened, with formation of small
spikes and deposits of foreign material between the endothelium and the
basement membrane, in the mesangium, and, more rarely, between
epithelium and basement membrane. There was periglomerular and
medullary sclerosis, especially in the inner stripe. Some arteries were
thickened and sclerotic (Figure 5, A, C, D
–F, H and I). Baboon IgG was
bound to glomerular capillary walls (Figure 5B)
, mesangium, and Bowman’s capsule.
Baboon IgG was also bound to fibroblasts in the interstitium of the
medulla (Figure 5G)
, whereas baboon IgM and pig C3 were absent.
Deposition of pig IgG was absent or minimal and focal. Increased
collagen I and less collagen III expanded the interstitium of the
medulla and the adventitia of vessels (Figure 5J)
. In the small
intestine of pig 9 severe fibrosis had developed, mainly in the lamina
propria, and less in the submucosa, with great distortion of the
mucosal structure. The fibrotic lamina propria was strongly stained by
type I and type III collagen antibody (Figure 6)
. In all pigs, baboon but not pig IgG
was bound to intestinal basement membranes and extracellular matrix,
especially in the small intestine.
|
|
|
The organs of pig 10 were normal. Pig 13 had developed some small subepithelial granular deposits of goat IgG and pig IgG; all other organs were normal.
Pig Injected with Sheep Anti-BSA (Pig 14) for 72 Hours and Sacrificed 13 Weeks Later
All organs were normal.
Group IV
Pigs Injected with Baboon Anti-Gal/Baboon Anti-PAEC and Pig
Anti-Baboon -Globulin (Pigs 15 and 16, Passive Transfer)
The lesions were more severe in pig 16, which died, probably of serum sickness. The lungs were diffusely edematous and hemorrhagic. The alveolar capillaries were occluded by erythrocytes, polymorphonuclear leukocytes, and mononuclear cells. The kidneys exhibited proliferative and exudative glomerulonephritis. Diffuse, coarse/granular deposits of baboon IgG, pig IgG, and C3 were present in alveolar and glomerular capillary walls and in the walls of small vessels (not shown).
Effect of Baboon Anti-Gal on Renal Fibroblast Production of
TGF-ß and Collagen
Because baboon anti-Gal bound to fibroblasts and to the
extracellular matrix, and widespread fibrotic lesions were observed in
pigs infused with baboon anti-Gal and baboon anti-PAEC, we tested
the hypothesis that the interaction of Gal antibodies with
fibroblasts elicits expression of profibrotic cytokine/growth factors,
such as TGF-ß, and/or causes increased production of collagen. Cells
cultured from the inner stripe of the renal medulla had morphological
aspects and a growth cycle characteristic of fibroblasts, were stained
by baboon anti-Gal, by anti-ß3 integrin and anti-vimentin
antibodies, but not cytokeratin antibody, which is reactive with
renal tubular epithelial cells.25
In culture supernatants
of fibroblasts incubated with baboon anti-Gal and anti-PAEC, total
TGF-ß was twofold (60 hours) higher than in supernatants of
fibroblasts incubated with normal rabbit -globulins. This effect was
reduced by absorption of baboon antibodies with rabbit erythrocytes.
Anti-ß3 integrin antibody did not increase the production of
TGF-ß (Figure 7A)
. Baboon anti-Gal,
but not anti-ß3 integrin antibody, also induced increased production
of collagen. This effect of baboon anti-Gal was blocked by
preabsorption with rabbit erythrocytes (Figure 7B)
.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Gal antibody
with pig tissues in a nondestructive model that allows assessment of
long-term consequences of the antigen-antibody reaction. We observed
fibrosclerotic lesions associated with, and probably consequent upon,
deposition of the antibodies in basement membranes and extracellular
matrix. We demonstrate Gal epitopes in the extracellular matrix,
especially in laminin and HSPG. We will consider how the results of
this model may relate to prolonged survival of pig organs in primates. Characteristics of Baboon Antisera and Their Suitability as a Model for Elicited Human Xenoantibodies
Immunization of baboons with Gal/BSA or PAEC resulted in
increased levels of Gal IgG and IgM. This is consistent with the
immune response observed in cynomolgus monkeys transplanted with
porcine or bovine cartilage,26
patients transplanted with
fetal porcine islet cells,27
or those treated with porcine
liver perfusion.28
The observation that baboon anti-PAEC
had high Gal titer confirms that Gal is a major xenoantigen of
PAEC.4–7,29–31 ELISA reactivity was also
increased for Matrigel, which contains laminin, with 50 to 60 Gal
epitopes per molecule16; type IV collagen; and HSPG, with a
ratio of 1:0.6:0.03, respectively.32
By
immunohistochemical titration, using a section of pig kidney as
substratum, the reactivity of baboon anti-Gal and baboon anti-PAEC
IgG was about 200-fold greater than that of human natural
xenoantibodies.
Western blot analysis of PAEC extracts using baboon anti-Gal and
baboon anti-PAEC identified bands with masses similar to those
identified with human natural xenoantibodies. These are 125 to 135-kd,
presumably integrins; 230-kd likely either von Willebrand factor or
fibronectin; and bands of 34- to 76-kd glycoproteins.33,34
Moreover, baboon anti-PAEC immunoprecipitates another abundantly
glycosylated 44-kd protein.35
Baboon anti-Gal, baboon
anti-PAEC, and human natural xenoantibodies may have similar
specificities, because the specificity of natural xenoantibodies and
that of sera of patients immunized by cross-perfusion with pig liver is
the same.28
The reactive epitope of the glycoproteins
identified by human natural xenoantibodies is
Gal,4–7,30,34 which is broadly represented in
phylogeny.7,10
Fixation of xenoantibodies to PAEC, however,
is not solely dependent on Gal, since removal of Gal decreases
its fixation only by 75 to 80%.34
That the baboons had
developed a sustained antibody response was shown by the increased
binding of baboon anti-Gal and baboon anti-PAEC to PAEC, and by
greater complement-mediated cytotoxicity, than normal baboon sera. Both
antisera could fix complement in vitro as in sections of
normal pig tissue.
Binding Sites of Baboon Anti-Gal and Baboon Anti-PAEC to
Sections of Normal Pig Tissues
Immunoreactive epitopes were identified in plasma membranes of
endothelial and some epithelial cells, blood cells, macrophages,
chondrocytes, and fibroblasts. In solid tissues they stained vascular
endothelia, as previously shown with Bandeiraea (Griffonia)
simplicifolia 1 isolectin B4 and Gal antibody
isolated from normal human sera.36-38
However, baboon
anti-Gal and baboon anti-PAEC also stained the basement membranes
and the extracellular matrix, where they colocalized with antibodies to
laminin, HSPG, perlecan, and partially to fibronectin. The principal
immunological target was Gal, because the staining was identical to
that obtained with Bandeiraea (Griffonia) simplicifolia B4,
which is specific for Gal epitopes39; was eradicated by
digestion of tissues with -galactosidase, but not by
ß-galactosidase; was abolished by absorption of baboon anti-Gal
with Gal/BSA, but not with BSA; and was also abolished by absorption
with Gal-rich rabbit erythrocytes. In contrast, the
-galactosidase digestion markedly inhibited, but did not abolish,
the staining with baboon anti-PAEC, consistent with recognition of
epitopes other than Gal. Differences between our immunohistochemical
findings and previous reports may be ascribed mostly to the high titer
and avidity of the baboon antisera, which were used at optimal staining
concentrations 350 to 400 times lower than those of Gal antibodies
isolated from normal human sera.37
It could be possible
that, in vivo, not all the epitopes visualized in tissue
sections are readily accessible to antibody.
Usefulness and Limitations of the Protocol Used for the Experiments Performed in Vivo
One of aims of this study was to find out whether baboon
anti-Gal, which react with Gal epitopes on matrix
glycoproteins of several animal species, induced in pig matrix
lesions similar to those described in animals injected with
heterologous anti-matrix antibodies.23,24,40-45
Baboon
anti-PAEC was used to determine whether an antiserum with multiple
specificities, including Gal, induced lesions similar to or
different from baboon anti-Gal. We infused baboon -globulins i.v.
for a relatively short period of time in order to prevent acute serum
sickness, and serum sickness never occurred in pigs of Groups I, II,
and III, as shown by the sharp linear immunofluorescence binding of
baboon antibody to pig basement membranes,46
without pig
IgG, and by absence of acute inflammatory lesions.47
And we
could also compare the effects of baboon anti-Gal and baboon
anti-PAEC with those natural xenoantibodies.30,48
The major limitation was that, aside from antibody, pig rather than baboon immune reactants were involved. Moreover, the effect of baboon antibodies was diluted in the entire blood and body of the pig instead of being focused on a single organ. Lastly, the binding of antibodies to blood cells and their agglutination at the beginning of the injections could cause acute pulmonary symptoms and release of mediators and cytokines in the circulation.
Clinical Observations
In pigs injected with baboon anti-Gal and baboon anti-PAEC the
early signs of pulmonary distress were mainly due to erythrocyte and
platelet agglutination and pulmonary small vessel embolism, as shown in
the pulmonary morphology of pig 6, which died suddenly at the beginning
of the infusion, or pigs that were sacrificed after 6 hours (Group I).
This is consistent with the observed decrease in blood cell counts.
Similar pulmonary symptoms developed in control pig 3, injected with
goat anti-ACE, in which the binding of antibody to pulmonary
endothelium with local fixation of complement induced local
accumulation and degranulation of inflammatory cells, especially
platelets and monocyte/macrophages, as in rabbits injected with goat
anti-ACE antibody.49,50
Control pigs 7 and 13, also
injected with goat anti-ACE, developed pulmonary signs analogous to pig
3. In contrast, pigs injected with sheep anti-BSA remained normal in
all these respects.
In Vivo Localization of Baboon Anti-Gal and Baboon
Anti-PAEC
In pigs injected for 6 hours (Group I) baboon IgG, most strongly baboon anti-PAEC IgG, was bound mainly to the endothelium, as seen in pig organs transplanted into unmodified baboons, confirming that the endothelium is the principal and immediate target of circulating xenoantibodies.1,48 In the kidney, baboon IgG was bound to the endothelium, but also to glomerular and tubular basement membrane.
After 72 hours (Group II), baboon IgG, but not pig C3, was
localized in alveolar basement membranes, in the alveolar septa, and on
the plasma membrane of alveolar epithelial cells. This delayed fixation
of the IgG is probably due to restrictive blood to airspace barrier
posed by the alveolar capillary walls; when the permeability is not
increased, about 24 hours are required for the transfer of IgG from the
circulation to the alveolar space.51
Baboon IgG was also
bound to renal basement membranes, to the basolateral compartment of
the cells of proximal tubules, and to Gal-positive cells in the
medulla. These findings are like those in other model systems. Fixation
of IgG and IgM to basement membranes and extracellular matrices of
heart, lung, liver, pancreas, intestine, and aortic adventitia was
described in pig organs which had been extracorporeally perfused with
human blood,52
fixation of rat IgG and IgM to the
extracellular matrix of hamster or guinea pigs aortas was shown to
occur after transplantation into rats,53
and rabbit IgG was
bound to fibroblasts and extracellular matrices of pig patellar tendon
and cartilage transplanted into rabbits.54
It is notable
that in rats with allograft rejection glomerulosclerosis, serum
antibodies reactive with cryptic basement membrane antigens and
proteoglycans (biglycan and decorin) are bound to the basement
membranes of the kidney.55,56
Four weeks after infusion of baboon anti-Gal and baboon anti-PAEC
(Group III), baboon IgG was bound to pig lung and renal basement
membranes in a pattern like that of animals given antibodies reactive
with basement membranes or some of their purified components,
laminin,23,40-43
type IV collagen,40,41
HSPG,24,44
or fibronectin.45
It is well
established that damage sufficient to alter morphology and function of
basement membranes and extracellular matrices can be induced by
antibody deposition alone, independent of complement activation and
inflammation.57
Consequences of a Prolonged Antigen-Antibody Interaction in Vivo and Mechanisms of Tissue Damage
That baboon anti-Gal and baboon anti-PAEC induce a similar
effect is in agreement with previous studies showing that the lesions
in pig cells or tissues caused by human natural xenoantibodies, or sera
of baboons or man sensitized by porcine grafts, are mainly due to
Gal antibodies.5-7,29-31
With the exception of antibody to 3 chain type IV collagen, the
Goodpasture’s antigen,58
all other antibodies to defined
components of the extracellular matrix provoke mild tissue lesions,
detectable only by electron microscopy.24,42,44
This is
probably due to the abundance and very widespread distribution of
matrix antigens, the amount of tissue-fixing antibody being too low to
reach the threshold necessary to fix complement.59
For
example, sheep anti-laminin antibody can fix C in vitro but,
when injected i.v. into rats, does not induce morphological and
functional changes, even when the autologous phase is actively or
passively boosted. Severe complement-dependent lesions, however,
develop with a more focused immunological attack, as when kidneys with
planted sheep anti-laminin antibodies are transplanted into
naïve recipients passively immunized with rat anti-sheep
IgG.23
Likewise, the effects of baboon anti-Gal and
baboon anti-PAEC antibodies are dissipated throughout the pig body, and
early fixation to blood cells may remove some high-affinity antibodies,
so that antibodies bound to any tissue were insufficient in quality,
amount, and concentration to fix C1q. That insufficient complement
fixation was due to a limitation of the model, and not to intrinsic
deficiency of baboon antibodies, was shown by studies in
vitro, and by the observation that when a control passive acute
serum sickness was induced (Group IV), complement was locally activated
and acute pulmonary and glomerular lesions ensued.
Lesions of basement membranes and the extracellular matrix were
similar, though more severe, than those of animals given
anti-laminin42
or anti-HSPG44
antibodies which
do not, or only poorly, fix complement in vivo. These
antibodies also bind to fibroblast surface and probably affect their
function. A similar effect seems to have been exerted by baboon
anti-Gal and baboon anti-PAEC when they are bound in vivo
to Gal epitopes on laminin and HSPG, interfering with the critical
function of these two glycoproteins in the supramolecular assembly of
an integrated basement membrane and matrix network.60
The
autologous phase, with in situ formation of immune
complexes, might have contributed to the development of the lesions. In
pigs, however, deposits of pig IgG in tissues were either absent or
minimal and focal. Therefore, even if a cross-linking effect of pig IgG
cannot be excluded, there is no evidence of it.
The possibility that cytokines/mediators released from platelets,
lymphocytes, and monocytes during their early and transitory
agglutination in pulmonary capillaries might have contributed to the
pathogenesis of the lesions deserves consideration. The control
experiments most appropriate to rule out the possible effect of these
mediators would have required injection of heterologous antibodies to
pig blood cell surface antigens, but devoid of Gal reactivity, but
such antibodies are not available. However, pigs injected with anti-ACE
antibodies also developed acute pulmonary symptoms, due to interaction
with pulmonary endothelium, activation of complement, and local
accumulation and degranulation of blood cells, especially platelets and
monocyte/macrophages. Absence of basement membrane and fibrosclerotic
lesions in these control pigs support the interpretation that fixation
of Gal antibodies to the extracellular matrix is the primary
pathogenetic mechanism. This interpretation is strengthened by the
similarity with lesions induced in the rat by heterologous
anti-laminin42
and anti-HSPG44
antibodies, and
by apparent absence of matrix lesions induced by cytokines/mediators
released in the circulation when antibody or immune complexes are not
bound to tissues.
Binding of Gal antibodies to Gal on fibroblast and macrophage
surfaces, might stimulate a fibrogenic response, consistent with our
observation that binding of baboon anti-Gal to renal fibroblasts
increased their production of active TGF-ß. Enhanced production
of collagen could result from stimulation by TGF-ß,61-63
and/or other effects of antibody engagement of cell surface Gal.
Similar mechanisms were invoked to explain the fibrosclerotic response
of rats injected with anti-thymocyte63
or anti-glomerular
basement membrane64
antibodies.
How the Results May Relate to Transplanted Pig Organs Surviving in Primates
The probability that in acute vascular and chronic rejection
Gal antibodies may bind to Gal epitopes on laminin, HSPG, and
fibronectin should be contemplated, as should their perturbing effect
on Gal-positive cells involved in the synthesis and remodeling of
the extracellular matrix. This hypothesis gains additional credibility
from the consideration that pigs were injected with Gal antibodies
only for 72 hours, whereas functioning xenografts may be exposed to a
longer and more focused attack. Thus, overcoming the hurdles of
hyperacute and acute vascular rejection would be insufficient to
prevent the subsequent development of fibrosclerotic lesions similar to
those restricting the survival of allografts,65,66
unless
the expression of Gal epitopes in the xenograft67
or the
ability of the host to generate Gal antibodies68
are
substantially inhibited or eradicated.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by National Institutes of Health grants DK-36807-25/27 (to G. A.) and HL-42507-PERC (to D. S.).
The first two named authors contributed equally to this work.
Accepted for publication July 27, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1-3Gal epitopes. Proc Natl Acad Sci USA 1993, 90:11391-11395 Galactosyl epitope (Gal1-3Galß1-4GlcNAc-R) and the natural anti-Gal antibody. Blamcher A Klein J Socha WW eds. Molecular Biology and Evolution of Blood Groups and MHC antigens in Primates. 1997, :pp 236-253 Springer, Berlin
-galactosyl epitopes on nucleated cells. J Biol Chem 1988, 263:17755-17762-gal epitope expression on cells by monoclonal anti-Gal antibody. Transplantation 1998, 65:1129-1132[Medline]
(1,3)Gal is the major xenoepitope expressed on pig endothelial cells recognized by naturally occurring cytotoxic human antibodies. Transplantation 1994, 58:879-882[Medline]
-Galactosyl determinant in hyperacute rejection. J Immunol 1995, 154:5500-5510[Abstract]
) and baboon
anti-PAEC IgG (•) are
increased 32- to 200-fold as compared to normal baboon IgG
(
). The titers of
baboon anti-
) and baboon
anti-PAEC IgM (
) are
increased 2- to 32-fold.
(
), normal baboon IgM;
(X) sheep anti-BSA IgG.
B: Binding of baboon anti-
) is like that of
normal baboon serum. C: SDS-PAGE analysis of PAEC. The Western
blots were probed with 10 µg/ml of the IgG fractions of baboon
anti-
