Full text of DOI:10.1034/j.1399-3089.2001.00082.x

#
Munksgaard 2001
Xenotransplantation 2001: 8: 15±23
Printed in UK. All rights reserved
Copyright
XENOTRANSPLANTATION
ISSN 0908±665X
Removal of bowel aerobic gram-negative
bacteria is more effective than
immunosuppression with cyclophosphamide
and steroids to decrease natural a-Galactosyl
IgG antibodies
Manez R, Blanco FJ, Dõaz I, Centeno A, Lopez-Pelaez E, Hermida M, Rafael ManÄez,1 Francisco J. Blanco,1
Davies HFS, Katopodis A. Removal of bowel aerobic gram-negative
Inmaculada DõÂaz,1 Alberto Centeno,1
Eduardo Lopez-Pelaez,1
bacteria is more effective than immunosuppression with
cyclophosphamide and steroids to decrease nautral a-Galactosyl
Manuel Hermida,1 Hugh F. S. Davies2
and Andreas Katopodis3
#
antibodies. Xenotransplantation 2001; 8: 15±23.
Copenhagen
Munksgaard,
1Juan Canalejo Medical Center and University of La
CorunÄa Health Institute, 15006 La CorunÄa, Spain,
2Imutran Ltd (a Novartis Pharma A.G. company)
Abstract: Natural a-Galactosyl (Gal) antibodies play an important role
in the rejection of pig xenografts by humans and Old World monkeys. In
this study we investigate the ef®cacy of two different strategies to reduce
the serum level of natural anti-Gal antibodies. On the one hand, removal
of aerobic gram-negative bacteria from the intestinal ¯ora, because anti-
Gal antibodies appear to be produced as a result of the continuous
sensitization by these microorganisms. On the other hand, we studied the
effect on these antibodies of an immunosuppressive regimen of
cyclophosphamide and steroids. Ten baboons were treated for three
months with nor¯oxacin (Nor Group; n56) or cyclophosphamide and
steroids (CyP Group; n54). A further four baboons did not receive any
treatment (Control Group). Aerobic gram-negative bacteria became
negative in stools of the Nor Group after two weeks of treatment, and
remained undetectable until week 7. Thereafter, a gradual increase on the
fecal concentration of aerobic gram-negative bacteria was observed
despite the nor¯oxacin treatment. The mean anti-Gal IgG in the Nor
Group gradually declined from week 4 to 9 to a mean of 62.7¡18% of
the baseline level, and during this period were signi®cantly lower than in
the CyP (P,0.02) and the Control (P,0.05) groups. No differences were
observed between the three groups during the 16 weeks of follow-up in
serum levels of anti-Gal IgM, hemolytic anti-pig antibodies, total IgG,
IgM and IgA. In conclusion, removal of normal aerobic gram-negative
bacteria from the intestinal ¯ora is more effective than
3
Cambridge, UK, Novartis Pharma A.G., Basle,
Switzerland
Key words: baboon ± natural antibodies ±
rejection ± selective intestinal decontamination ±
xenotransplantation
Address reprint requests to Rafael ManÄez,
Transplantation Division, Juan Canalejo Medical
Center, Xubias de Arriba, 84, 15006 La CorunÄa,
Spain
immunosuppression with CyP and steroids in reducing the level of
natural anti-Gal antibodies, although there is no discernible effect on
IgM antibodies.
Received 26 April 2000;
Accepted 27 July 2000
Introduction
the recipient at the time of transplantation, which
bind xenograft endothelial cells and activate the
complement system [2,3]. For many years this
rejection was the most important barrier for the
success of xenotransplantation of pig organs in
Old World monkeys [4]. Recently, several meth-
ods have been developed to elude hyperacute
rejection. These include transient depletion of
Xenotransplantation of pig organs into humans,
apes and Old World monkeys has been precluded
so far by severe immunological reactions that lead
to a prompt xenograft failure. Hyperacute rejec-
tion is the earliest type of rejection, beginning
immediately upon reperfusion [1]. This reaction is
mediated by xenoreactive antibodies present in
15

ManÄez et al.
xenoreactive natural antibodies [4], inhibition of
complement by soluble factors [4], and particu-
larly the use of organs from pigs transgenic for
human complement regulatory proteins [5±7].
However, when hyperacute rejection is averted,
other forms of rejection emerge, destroying the
xenograft within days to a few weeks after
transplantation [7]. The most prevalent cause of
failure in these cases is an acute vascular rejection,
which appears to be mainly caused by xeno-
reactive antibodies [7], although NK cells and
macrophages may also be involved [8]. There-
fore, xenoreactive antibodies have a primary
role not only in hyperacute rejection but also in
acute vascular rejection of pig organs trans-
planted in Old World monkeys.
hyperacute rejection. To obtain an impact on
avoiding acute vascular rejection, a sustained
reduction of anti-Gal antibodies is required. If
natural antibodies do participate to some extent in
this reaction, a sustained decrease in their levels is
necessary. So far, most of the immunosuppression
protocols used in pig to non-human primate xeno-
transplantation include cyclophosphamide (CyP)
and steroids to inhibit antibody production [4±7].
These drugs have demonstrated clinical ef®cacy in
several auto-immune diseases mediated by natural
antibodies [17±20]. However, in the xenotrans-
plantation of pig organs to non-human primates
the only effect demonstrated so far is a reduction
on the return rate of natural anti-Gal antibodies if
they are previously depleted with aGal columns
[21].
Xenoreactive antibodies are either natural, also
named pre-formed, because they are present at
the time of transplantation without a known
history of sensitization, or elicited, which are
those antibodies produced by xenograft stimula-
tion. Although both natural and elicited anti-
bodies may recognize the same antigens, they
appear to be produced by different immunologi-
cal mechanisms. Natural antibodies are generated
by a T-cell independent response, whereas the
production of elicited antibodies is T-cell depen-
dent [9]. The xenoreactive antibodies responsible
for hyperacute rejection of pig xenografts trans-
planted into humans, apes and Old World
monkeys are natural antibodies. These are
mainly IgM and their major target is the a-
Galactosyl (Gal) epitope [10±12], present in non-
primate mammals, prosimians and New World
monkeys, all of which lack anti-Gal antibodies
[13]. In acute vascular rejection it is unclear
whether the xenoreactive antibodies involved are
natural, elicited or both. However, humans and
baboons exposed in vivo or ex vivo to pig organs
or cells are clearly stimulated to produce anti-Gal
IgM and IgG [14±16]. The maintenance of the
main speci®city for xenoantibodies before and
after the exposure to pig antigens suggests that
natural xenoantibodies remain important in the
immunological reactions mediated by these anti-
bodies after transplantation.
While natural anti-Gal antibodies function in
the xenotransplantation setting to induce hyper-
acute rejection, its physiological role is in homeo-
stasis. These antibodies bind to carbohydrate
structures exposed by several aerobic gram-
negative bacterial strains [22,23], and share an
immunoglobulin (Ig) gene structure with anti-
bodies to common infectious agents [24], suggest-
ing that they may be involved in the defense
against infectious micro-organisms. Anti-blood
Group isohemagglutinins is another family of
natural antibodies that appear to be produced as a
result of constant immunization against intestinal
¯ora [9,25]. The similarities between natural anti-
Gal antibodies and anti-blood Group isohemag-
glutinins, along with the presence in normal intes-
tinal ¯ora of gram-negative bacterial strains that
interact with the former antibodies lead to the
hypothesis that these micro-organisms provide a
constant antigenic stimulation for the synthesis of
natural anti-Gal antibodies [23]. If this is true,
elimination from the bowel of micro-organisms
that exhibit Gal epitopes should remove the stim-
ulus for production of natural anti-Gal anti-
bodies, and thus to a sustained decrease in their
levels.
In the present study, we compared the effect of
two different strategies on the serum levels of
natural anti-Gal antibodies in baboons. On the
one hand, we removed aerobic gram-negative
bacteria from the intestinal ¯ora, using an oral
non-absorbable or poorly absorbable antibiotic
such as nor¯axacin [26,27], which has demon-
strated ef®cacy in producing a selective removal of
these micro-organisms in healthy subjects [28],
and in patients with granulocytopenia, cirrhosis
or liver transplantation [27,29,30]. On the other
hand, we tested the effect of an immunosuppres-
sion regimen of CyP and steroids.
Removal or neutralization of natural anti-Gal
antibodies has been achieved so far by several
methods. These include non-speci®c techniques
such as plasmapheresis, xenogenic organ perfu-
sion or infusion of monoclonal anti-idiotypic anti-
bodies, and speci®c techniques such as immu-
noadsorption with aGal columns or perfusion
with soluble aGal sugars [4]. However, all them
have the same problem. The effect on natural anti-
Gal antibodies lasts only a few hours after the
procedure ends, allowing only the prevention of
16

Natural aGal IgG antibodies
Materials and methods
log10 per g (wet weight) of feces (colony-forming
units). The lower limit of detection was 3.0 log10
colony-forming units (CFU) per g of feces.
Animals and study groups
A total of 14 baboons (Papio Anubis), purchased
from Consort Bioservices (Steyming, UK), were
used in this study. The animals, weighing 5 to
8 kg, were housed in a facility approved and
accredited by the Galicia Department of Agricul-
ture. The studies were approved and monitored by
the Research Committee of Juan Canalejo Medi-
cal Center.
Six baboons received nor¯oxacin 100 mg/day
orally for 3 months (Nor Group), whereas an-
other four baboons were treated during the same
period with cyclophosphamide (CyP) 500 to 1000
mg/m2 i.v. every 5 to 7 days, to maintain a white
blood cell (WBC) count of between 2000 and
3000 cells/mm3 (CyP Group). The CyP Group
animals also received during the ®rst two weeks,
three pulses of 15 mg/kg of methylprednisolone
i.v., followed by 5 mg/kg three times a week of
prednisolone tapered to a maintenance dose of
2 mg/kg three times a week over four weeks. A
further four baboons did not receive any treat-
ment (Control Group).
Measurement of anti-Gala1±3Gal antibodies by ELISA
Microtiter plates were coated with 50 mg/well
aGal(1±3)bGal(1±4)GlcNAc linked to human
serum albumin (HSA) (Dextra Labs, Reading,
UK) (5 mg/ml) in 0.1 m carbonate buffer pH 9.6
overnight at 4 uC. Coated plates were blocked
with 200 ml/well of 0.5% Tween/PBS for 1 h, at
room temperature. After washing, duplicate
dilutions of baboon serum prepared with 0.5%
Tween/PBS, ranging from 1:2.5 to 1:320 or 1:10
to 1:1280, were pipetted (50 ml/well) to the
antigen-coated wells and into uncoated, blank
wells serving as controls for non-speci®c antibody
binding. After an incubation of 1 h at room
temperature, plates were washed six times with
0.1% Tween/PBS and incubated with horseradish
peroxidase-labeled goat F(ab')2 anti-human IgG
(gamma chain speci®c) or IgM (mu chain
speci®c)-(Sigma, St Louis, MO, USA) at 1:200
in 0.5% Tween/PBS, for 1 h at room temperature.
After six washes with 0.1% Tween/PBS, the plate
was developed with a solution of orthophenylene
diamine (OPD, Sigma) prepared according to the
manufacturer's instructions and quenched with
1 m H2SO4 before reading absorbances at
492 nm. The levels of anti-Gal IgG or IgM in
different assays were normalized by comparison
with a pool of 50 human sera provided by the
local Blood Bank, which was included in all the
assays
Samples
Weekly stool samples were obtained from
baboons of the Control and Nor groups during
the 4 months. Serum samples for assaying levels
of anti-Gal IgG, IgM and hemolytic anti-pig
antibodies were taken at weekly intervals from
the three groups during the same period of time.
One additional sample was obtained from the
baboons of the Nor Group at week 20. Also,
serum samples were obtained before any treat-
ment and once a month for 4 months to study
total IgG, IgM and IgA levels. Finally, blood
samples were drawn every other day in the CyP
Group to follow the WBC count.
Measurement of hemolytic anti-pig antibody titer
Baboon clotted blood samples were spun for
5 min at 3000 r.p.m. at 4 uC. Sera were then
collected, transferred into Eppendorf tubes, and
heat inactivated at 56 uC for 30 min. Subse-
quently, each serum was serially diluted in a
96-well plate with complement ®xation diluent
(ICN, Costa Mesa, CA, USA), from 1:2.5 to
1:1280. After the addition of rabbit complement
(Serotec, Oxford, UK) and 1% porcine red blood
cells, the plates were placed in an orbital incubator
at 37 uC for 1 h. The plates were then centrifuged
for 10 min, and 100 ml of supernatant from each
well were transferred into reading plates. The
reading was performed by a Labsystems Multi-
skan Plus (Labsystems, Helsinki, Finland) at a
wavelength of 420 nm. The mean absorbance of
the two replicates was calculated for each dilution
and expressed in area under the curve (AUC)
units, considering the number of AUC units of the
pool of human control serum as equal to 1.
Collection and processing of specimen for fecal studies
Stool samples were collected in non-sterile plastic
containers and the specimen weight was recorded.
Fairly solid specimens were homogenized in a
stomacher; liquid samples were homogenized by
vortexing. A sample of 0.5 g of fresh feces was
diluted 1:100 in 0.9% NaCl with the use of a
mixer. Then 0.1-ml aliquots of this dilution and
further 100-fold dilutions (to 10±6) were spread
onto MacConkey agar (aerobic gram-negative)
and blood agar with nalidixic plates (aerobic
gram-positive) (Oxoid Ltd, Basingstoke, UK).
Plates were incubated aerobically at 36 uC for 24
to 48 h. Different colony types were identi®ed
through gram staining and counted as the viable
17

ManÄez et al.
Table 1. Baseline level of anti-Gal IgM, IgG and hemolytic anti-pig antibodies
mean¡S.D. shift from the individual baseline
levels of the baboons in each group was estimated
every week.
Baboon
Group
Anti-Gal IgM*
Anti-Gal IgG*
Hemolytic anti-pig{
1
Nor
Nor
0.62
0.69
0.90
0.72
0.79
0.83
0.72
0.71
1.06
0.45
0.79
0.78
0.55
0.65
0.69
0.29
0.93
0.62
0.77
0.34
0.71
0.93
0.86
0.39
0.56
0.72
0.36
0.59
1.35
0.95
1.20
0.93
1.09
0.90
1.10
1.24
1.28
0.8
2
Statistical analysis
3
Nor
4
Nor
ANOVA for repeated measures was used to compare
the level of anti-Gal IgM, IgG and hemolytic anti-
pig antibodies between the three study groups.
5
Nor
6
Nor
7
Control
Control
Control
Control
CyP
8
9
Measurement of total IgM, IgG and IgA levels
10
11
12
13
14
1.12
1.37
0.68
1.21
The concentrations of total IgM, IgG and IgA
were determined by nephelometry using a BN II
nephelometer (Behring, Marburg, Germany).
CyP
CyP
CyP
*Normalized by comparison with a pool of human sera (pool human sera 5 1).
{The values represent AUC units in comparison with the pool of human sera (pool
human sera 5 1 AUC unit).
Results
Effect of nor¯oxacin on fecal ¯ora
As illustrated in Fig. 1, the mean counts of aero-
bic gram-negative bacteria in the stools before
treatment were 5.37¡5.0 and 5.30¡5.18 (log10
CFU)/g in the Nor and Control Groups, respec-
tively. One week after nor¯oxacin treatment was
started, the mean count of aerobic gram-negative
bacteria declined to 3.98¡3.0 (log10 CFU)/g in
the Nor Group. Between weeks 2 and 7, no
aerobic gram-negative bacteria were found in the
feces. At week 8, a mean count of 3.18¡3.0 (log10
CFU)/g of aerobic gram-negative bacteria was
detected in the stool, despite the nor¯oxacin
treatment. Thereafter, a gradual increase of the
fecal concentration of aerobic gram-negative
bacteria occurred in the Nor Group. At week
12, within three weeks of the cessation of the
nor¯oxacin treatment, the counts of aerobic
gram-negative bacteria returned to pretreatment
Pretreatment serum level of anti-Gal antibodies and
hemolytic anti-pig antibody titer
In baboons, as in humans, the serum level of anti-
Gal IgG and IgM antibodies, and the serum hem-
olytic anti-pig antibody titer, show large varia-
tions between individuals [31]. This precludes the
use of average antibody titers to evaluate changes
in a group of several baboons. The use of a
shift from pretreatment level for each individual
baboon is necessary to give homogeneity to the
results.
Table 1 shows the pretreatment level of anti-
Gal IgG, IgM and hemolytic anti-pig antibody
titer of the 14 baboons. The weekly values
obtained from every baboon are expressed as a
percentage of the baseline level. To summarize
and compare the data from the three groups, the
Nor group
106
105
104
103
102
101
100
Control group
0
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16
Week
Fig. 1. Mean weekly counts of aerobic gram-negative bacteria in stools from the Nor and Control Groups.
18

Natural aGal IgG antibodies
1010
109
108
107
106
105
104
103
102
101
100
Nor group
Control group
0
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16
Week
Fig. 2. Mean weekly counts of aerobic gram-positive bacteria in stools from the Nor and Control Groups.
levels. During the same period, no change was
observed in the mean counts of aerobic gram-
negative bacteria in the Control Group.
Effect of cyclophosphamide on serum anti-Gal levels
Baboons treated with CyP received an average
daily dose of 7, 6.6, 6.9 and 5.9 mg/kg, respec-
tively. Figure 4(A) shows the mean levels of anti-
Gal IgG in the CyP and Control groups over the
16 weeks of the study period. No differences were
observed between these groups at any time.
However, the mean levels of anti-Gal IgG were
signi®cantly higher in the CyP Group than in the
Nor Group between weeks 4 and 9 (P,0.02).
The mean levels of anti-Gal IgM in the CyP and
Control Groups are illustrated in Fig. 4(B). There
were no differences between these two groups, or
with the Nor group, in any week of the follow-up.
At the beginning of the study, the mean baseline
counts of aerobic gram-positive bacteria in the
feces were 7.53¡7.07 and 7.35¡7.28 (log10
CFU)/g in the Nor and Control Groups, respec-
tively. In the Nor Group, the mean count
gradually increased to a maximum mean level of
9.56¡8.08 (log10 CFU)/g at week 3, remaining
above 9 log10 CFU/g until week 7, whilst no
change was observed in the Control Group during
the 16 weeks of the study period (Fig. 2).
Effect of nor¯oxacin on serum anti-Gal levels
Effect of nor¯oxacin and cyclophosphamide on the level of
hemolytic anti-pig antibodies
Figure 3(A) shows the levels of anti-Gal IgG
antibodies in the Nor and Control Groups.
During the ®rst three weeks, no differences were
found between the two groups. At week 4, anti-
Gal IgG levels dropped to give a mean of
74.3¡15.5% from the pretreatment levels in the
Nor Group. Thereafter, anti-Gal IgG gradually
declined to a minimum mean of 62.7¡18% of the
pretreatment level at week 9. Between weeks 5 and
9, the mean anti-Gal IgG level in the Nor Group
was signi®cantly lower than in the Control Group
(P,0.05). At week 10, two weeks after aerobic
gram-negative bacteria returned to the stool, anti-
Gal IgG levels started a progressive increase to
86.5¡15% of the pretreatment level at week 16
and reached the baseline pretreatment level at
week 20.
Figure 5(A) shows the mean levels of hemolytic
anti-pig antibodies in the Nor and Control
Groups. No differences were observed at any
time between these two groups. Also, the mean
levels of hemolytic anti-pig antibody titer from the
CyP Group were similar to the Control Group
during the 16 weeks of the study (Fig. 5B).
Effect of nor¯oxacin and cyclophosphamide on serum total
IgG, IgM and IgA
To investigate whether nor¯oxacin or CyP have
any effect on the total level of antibodies, total
IgG, IgM and IgA were measured before, during
and after the treatments. Table 2 shows the
mean¡S.D. of these immunoglobulins in the
three groups. No changes were observed on the
total level of any class of immunoglobulin before,
during or after treatments with nor¯oxacin or
CyP and steroids. Also, the level of IgG, IgM and
The mean levels of anti-Gal IgM antibody in
the Nor and Control Groups are shown in
Fig. 3(B). No differences were found between
these two groups in any of the 16 weeks of study
period.
19

ManÄez et al.
120
100
80
A
A
120
100
80
60
40
20
0
60
40
Control group
Nor group
20
Control group
Nor group
0
0
2
4
6
8
10
12
14
16
B
0
2
4
6
8
10
12
14
16
B
120
100
80
120
100
80
60
40
20
0
60
40
Control group
CyP group
20
Control group
CyP group
0
0
2
4
6
8
10
12
14
16
0
2
4
6
8
10
12
14
16
Week
Week
Fig. 3. Mean weekly anti-Gal IgG (A) and IgM (B) in the
Nor and Control Groups. The values represent the
percentage of the pretreatment level.
Fig. 4. Mean weekly anti-Gal IgG (A) and IgM (B) in the
CyP and Control Groups. The values represent the
percentage of the pretreatment level.
IgA were similar in the Nor, CyP and Control
Groups.
incomplete, because the lower level obtained was
63% of the baseline level. Several mechanisms may
explain the effect observed on natural anti-Gal
IgG. Nor¯oxacin could have a direct effect over
some transcriptional activator factors of B cells,
because it inhibits immunoglobulin production
[32]. However, suppression of production of anti-
Gal IgG by plasma cells or memory cells by
nor¯oxacin should be speci®c for this antibody
because in this study the total level of immuno-
globulins did not change with the treatment.
Another explanation may be the decrease of
serum natural anti-Gal IgG antibodies as a
consequence of the elimination of aerobic gram-
negative bacteria from the intestinal ¯ora. In this
case, a decrease in antigen stimulation may reduce
the production of anti-Gal IgG by plasma cells or
may lead to a failure of isotype Ig switching by
affecting factors such as gram-negative lypopoly-
saccharides (LPS) that play a critical role in this
process [33].
Discussion
The present study demonstrates that removal of
intestinal aerobic gram-negative bacteria from the
bowel with a non-absorbable antibiotic such as
nor¯oxacin is associated with a decrease in the
serum level of natural anti-Gal IgG antibody.
After two weeks of nor¯oxacin treatment aerobic
gram-negative bacteria became undetectable in
the fecal ¯ora. The decline of natural anti-Gal IgG
was signi®cant after week 4, with the lowest level
detected at week 9 of nor¯oxacin treatment, one
week after aerobic gram-negative bacteria
returned to the stool. The two-week delay
observed between the absence of aerobic gram-
negative bacteria and the lowest level of natural
anti-Gal IgG levels may be due to the relatively
long half-life of IgG antibody, which is around
23 days.
The incomplete effect observed on natural anti-
Gal IgG, and the lack of effect on natural anti-Gal
IgM suggests the production of these antibodies,
to some extent in the former and completely in the
latter, through mechanisms independent of intest-
Elimination of aerobic gram-negative bacteria
from intestine was more effective than immuno-
suppression in reducing the serum level of natural
anti-Gal IgG antibodies. However, the effect was
20

Natural aGal IgG antibodies
120
100
80
60
40
20
0
A
spectrum of antibiotic therapy might have a major
impact on anti-Gal antibodies. However, the lack
of evidence that other micro-organisms from nor-
mal intestinal ¯ora carry the Gal antigen besides
aerobic gram-negative bacteria, along with the
risk of rising resistant micro-organisms with a
broad-spectrum treatment, lead us to choose a
single antibiotic treatment. Nor¯oxacin has
proven clinical ef®cacy on the prevention of
gram-negative infections in cirrhotic and trans-
plant patients [29,30]. Despite this, it was associ-
ated with an augmentation of aerobic gram-
positive, and the total elimination of aerobic
gram-negative micro-organisms for only 6 weeks.
Thereafter, the isolation of aerobic gram-negative
bacilli from stools, despite the maintenance of the
antibiotic treatment, suggests the appearance of
nor¯oxacin resistant micro-organisms.
The immunosuppressive regimen used in this
study, which included CyP and steroids, did not
have any effect on the level of natural anti-Gal
IgG or IgM. Previous observations in humans
showed that primary and secondary humoral
immunode®ciencies, as well as active immuno-
suppression with steroids, are associated with a
decline in natural anti-Gal IgG titers [37]. The
lack of ef®cacy of CyP and steroids on natural
anti-Gal antibody titer, despite using an average
dose of CyP and steroids signi®cantly higher than
that recommended for immunosuppression in
humans, suggests that baboon B cells may be
more resistant than human B cells to these drugs
or that the pharmacokinetics and bioavailability
of the active CyP metabolite is different in
baboons. This was corroborated in our study by
the lack of effect observed with CyP and steroids
on total IgG, IgM and IgA levels during the three
months of treatment. When the same treatment is
used for auto-immune diseases, clinical improve-
ment is associated with a decline in the total level
of IgM, IgG and IgA [19]. The results of this study
also differ from the ef®cacy observed with high
doses of CyP in prolonging xenograft survival in
pig to non-human primate xenotransplantation
[38]. This may indicate that the xeno-antibodies
responsible for acute vascular rejection are mainly
elicited instead of natural antibodies, and for this
Control group
Nor group
0
2
4
6
8
10
12
14
16
B
120
100
80
60
40
20
0
Control group
CyP group
0
2
4
6
8
10
12
14
16
Week
Fig. 5. Mean weekly hemolytic antibody titer in the Nor
(A) and CyP (B) Groups. The values represent the
percentage of the pretreatment level.
inal bacteria stimulation or by micro-organisms
other than aerobic gram-negative bacteria. Gal
epitopes have been described in the carbohydrate
portion of LPS from gram-negative bacteria [23],
in parasites and in retroviral envelopes [34,35]. Of
these, only gram-negative bacteria are present in
physiological conditions in humans. In baboons,
we cannot rule out the fact that parasites also play
a role in the production of anti-Gal IgG and IgM
antibodies, because detection of these micro-
organisms in stools is associated with higher
level of antibodies [36]. However, the 14 animals
included in this study were totally free of parasites
and therefore similar to the human situation.
The oral non-absorbable antibiotic used in the
study was only active against aerobic gram-nega-
tive bacteria. It cannot be ruled out that a broader
Table 2. Total levels of immunoglobulins before, during and after nor¯oxacin and cyclophosphamide treatments
IgG (mg/dL)*
During
IgM (mg/dL)*
During
IgA (mg/dL)*
During
Group
Before
After
Before
After
Before
After
Nor
1020¡241
969¡214
896¡231
902¡176
834¡153
795¡193
855¡108
786¡186
838¡97
62¡32
74¡16
68¡28
75¡11
69¡26
77¡18
78¡15
82¡20
71¡24
25.3¡4.8
21.8¡6.6
26.8¡2.9
25.0¡1.7
30.5¡10.8
27.1¡5.3
25.9¡4.2
22.2¡7.1
27.8¡4.6
CyP
Control
*Mean¡S.D.
21

ManÄez et al.
6. SCHMOECKEL M, BHATTI FNK, ZAIDI A et al. Heart trans-
plantation in a transgenic pig-to-primate model. Trans-
plantation 1998; 65: 1570.
7. ZAIDI A, SCHMOECKEL M, BHATTI FNK et al. Life-support-
ingpig-to-primaterenalxenotransplantationusinggenet-
ically modi®ed donors. Transplantation 1998; 65: 1584.
8. BACH FH, ROBSON SC, WINKLER H et al. Barriers to Xeno-
transplantation. Nature Med 1995; 1: 869.
9. PARKER W, BRUNO D, PLATT JL. Xenoreactive natural
antibodies in the world of natural antibodies: typical or
unique? Transplant Immunol 1995; 3: 181.
10. GOOD AH, COOPER DK, MALCOLM AJ et al. Identi®cation
of carbohydrate structures that bind human antiporcine
antibodies: implications for discordant xenografting in
humans. Transplant Proc 1992; 24: 559.
reason CyP is more effective. However, it should
be noted that the average dose of CyP during the
®rst 9 days of transplantation required to prolong
pig xenograft survival is more than three times the
average dose given in this study [38]. This dosage
is associated with signi®cant toxicity and when it
was decreased to an average of twice the dose used
in our protocol, all the xenografts underwent
acute vascular rejection. Thus, it cannot be ruled
out that keeping lower WBC counts through high
CyP doses would have reduced the total level of Ig
and particularly natural anti-Gal antibodies.
Nevertheless, this type of treatment cannot be
main-tained for very long and the objective of
the present study was to obtain a sustained
decrease of natural anti-Gal antibodies with a
protocol that could be used clinically for long
periods of time. In these circumstances the level of
WBC must be maintained between 2000 and 3000
cells/mm3 to avoid side-effects, particularly infec-
tious complications.
In summary, removal of aerobic gram-negative
bacteria from the bowel with nor¯oxacin is associ-
ated with a decrease in natural anti-Gal IgG. The
intrinsic mechanisms responsible of this response
remain to be elucidated. However, elimination of
these micro-organisms with nor¯oxacin is the
only method, besides the immunoadsorption tech-
niques, that has been associated with a sustained
reduction of at least some of natural anti-Gal
antibodies.
11. PARKER WR, BRUNO D, HOLZKNECHT ZE, PLATT JL. Xeno-
reactive natural antibodies: isolation and initial char-
acterization. J Immunol 1993; 153: 379.
12. SANDRIN MS, VAUGHAN HA, DABKOWSKI PL, MCKENZIE
IFC. Anti-pig IgM antibodies in human serum react
predominantly with Gala (1±3) Gal epitopes. Proc Natl
Acad Sci USA 1993; 90: 11391.
13. GALILI U. Interaction of the natural anti-Gal antibody
with a-galactosyl epitopes: a major obstacle for xeno-
transplantation in humans. Immunol Today 1993; 14:
480.
14. SATAKE M, KAWAGISHI N, RYDBERG L et al. Limited speci®-
city of xenoantibodies in diabetic patients transplanted
with fetal porcine islet cell clusters. Main antibody react-
ivity against a-linked galactose-containing epitopes.
Xenotransplantation 1994; 1: 89.
15. BAQUERIZO A, MHOYAN A, KEARNS-JONKER Met al. Charac-
terization of human xenoreactive antibodies in liver
failure patients exposed to pig hepatocytes after bioarti-
®cial liver treatment. Transplantation 1999; 67: 5.
16. MANEZ R, CRESPO F, GONZALEZ E et al. Neutralization of
anti-agalactosyl antibodies without immunosuppression
prevents hyperacute rejection but not acute vascular
rejection of pig organs transplanted into baboons.
Transplant Proc 2000; 32.
Acknowledgments
17. FAUCI AS, KATZ P, HAYNES BF, WOLFF SM. Cyclophos-
phamidetherapyof severesystemic necrotizingvasculitis.
N Engl J Med 1979; 301: 234±238.
18. HOFFMAN GS,LEAVITT RY,FLEISHER TA,MINOR JR,FAUCI
AS. Treatment of Wegener's granulomatosis with inter-
mittent high-dose intravenous cyclophosphamide. Am
J Med 1990; 89: 403.
We thank Dr Francisco Javier MunÄiz for his assis-
tance in the statistical analysis.
This work was funded in part by Novartis
Pharma AG, Basel, Switzerland.
19. CUPPS TR, EDGAR LC, FAUCI AS. Suppression of human B
lymphocyte function by cyclophosphamide. J Immunol
1982; 128: 2453.
20. FAUCI AS, HAYNES BF,KATZ P.Thespectrumofvasculitis:
clinical,pathologic, immunologic and therapeutic con-
siderations. Ann Intern Med 1978; 89: 660.
References
1. PLATT JL, VERCELLOTTI GM, DALMASSO AP et al. Trans-
plantation of discordant xenografts: a review of progress.
Immunol Today 1990; 11: 450.
2. ROSE AG, COOPER DKC, HUMAN PA, REICHENSPURNER H,
21. LAMBRIGHTS D, VAN
CALSTER P, XU Y et al. Pharmacologic
R
EICHART B. Hispotahology of hyperacute rejection of the
immnosuppressive therapy and extracorporeal immu-
noadsorption in the suppression of anti-aGal antibody in
the baboon. Xenotransplantation 1998; 5: 274.
heart: experimental and clinical observations in allografts
and xenografts. J Heart Lung Transplant 1991; 10: 223.
3. DALMASSO AP, VERCELLOTTI GM, FISCHEL RJ, BOLMAN
RM, BACH FH, PLATT J. Mechanism of complement
activation in the hyperacute rejection of porcine organs
transplanted into primate recipients. Am J Pathol 1992;
140: 1157.
22. SPRINGER G. Blood groups and Forssman antigenic
determinanats shared between microbes and mammalian
cells. Prog Allergy 1971; 15: 9.
23. GALILI U, MANDRELL RE, HAMADEH RM, SHOHET SB,
G
RIFFISS JM. Interaction between human natural anti-a-
4. LAMBRIGTS D,SACHS DH,COOPER DKC.Discordantorgan
xenotransplantation in primates: World experience and
current status. Transplantation 1998; 66: 547.
galactosyl immunoglobulin G and bacteria of the human
¯ora. Infect Immun 1988; 56: 1730.
5. LIN SS, WEIDNER BC, BYRNE GW et al. The role of anti-
bodies in acute vascular rejection of pig-to-baboon
cardiac transplants. J Clin Invest 1998; 101: 1745.
24. KEARNS-JONKER M, FRAIMAN M, CHU W et al. Xenoanti-
bodies to pig endothelium are expressed in germline
con®guration and share a conserved immunoglobulin VH
22

Natural aGal IgG antibodies
gene structure with antibodies to common infectious
agents. Transplantation 1998; 65: 1515.
to serum cytotoxicity against pig cells. Xenotrans-
plantation 1994; 1: 58.
25. SPRINGER GF,HORTON RE.Bloodgroupisoantibodystim-
ulation in man by feeding blood group-active bacteria.
J Clin Invest 1969; 48: 1280.
32. FORSGREN A, BREDBER A, RIESBECK K. New quinolones: in
vitro effects as potential source of clinical toxcicity. Rev
Infect Dis 1989; 11: S1382.
26. VANDER
WAAIJ D. Colonization resistance of the digestive
33. ZELAZOWSKI P, CARRASCO D, ROSAS FR, MOORMAN MA,
tract: clinical consequences and implications. J Anti-
microb Chemother 1982; 10: 263.
BRAVO R, SNAPPER CM. B cells genetically de®cient in the
c-Rel transactivation domain have selective defects in
germline CH transcription and Ig class switching.
J Immunol 1997; 159: 3133.
27. D
W
E
V
RIES-HOSPERS HG, SLEIJFER DT, MULDER NH, VANDER
AAIJ D, NIEWEG HO, vAN
SAENE HK. Bacteriological
aspects of selective intestinal decontamination of the di-
gestive tract as method of infection prevention in granu-
locytopenic patients. Antimicrob Agents Chemother
1981; 19: 813.
34. TOWBIN H, ROSENFELDER G, WIESLANDER J et al. Circu-
lating antibodies to mouse laminin in Chagas disease,
American cutaneous leishmaniasis and normal indivi-
duals recognize terminal galactosyl (a1R3) galactosyl
epitopes. J Exp Med 1987; 166: 419.
Á
28. PECQUET S, ANDREMONT A, TANCREDE C. Selective anti-
microbial modulation of the intestinal tract by nor¯ox-
acin in human volunteers and in gnotobiotic mice
associated with human faecal ¯ora. Antimicrob Agents
Chemother 1986; 29: 1047.
35. TAKEUCHI Y, PORTER CD, STRAHAN KM et al. Sensitization
of cells and retroviruses to human serum by (a1±3)
galactosyltransferase. Nature 1996; 379: 85.
36. LOPEZ E, CENTENO A, CRESPO F et al. Augmented anti-
agalactosyl antibodies in adult baboons in the setting of
intestinal parasitic infections are not associated with
higher titers of anti-pig hemolytic antibody. Transplant
Proc 2000; 32.
29. GINES P, RIMOLA A, PLANAS R et al. Nor¯oxacin prevents
spontaneous bacterial peritonitis recurrence in cirrhosis:
Resultsofadouble-blindplacebo-controlledtrial. Hepat-
ology 1990; 12: 716.
30. LUMBRERAS C, LIZASOAIN M, MORENO E et al. Major
bacterial infections following liver transplantation: a
prospective study. Hepatogatroenterology 1992; 39:
362.
31. KUJUNDZIC M, KOREN E, NEETHLING FA et al. Variability
of anti-Gal antibodies in human serum and their relation
37. GALILI U, RACHMILEWITZ EA, PELEG A, FLECHNER I. A
unique natural human IgG antibody with anti-a-galac-
tosyl speci®city. J Exp Med 1984; 160: 1519.
38. WATERWORTH PD, COZZI E, TOLAN MJet al. Pig-to Primate
cardiac xenotransplantation and cyclophosphamide
therapy. Transplant Proc 1997; 29: 899.
23