Human Reproduction vol.15 no.6 pp.1235–1240, 2000
Extracorporeal perfusion of the human uterus as an
experimental model in gynaecology and reproductive
medicine
O.Richter1,5, E.Wardelmann2, F.Dombrowski2,
C.Schneider3, R.Kiel1, K.Wilhelm4, J.Schmolling1,
M.Kupka1, H.van der Ven1 and D.Krebs1
1Department of Obstetrics and Gynaecology, 2Department of
Pathology, 3Department of Cardio-Thoracic Surgery and
4Department of Radiology, University of Bonn, Faculty of
Medicine, Sigmund-Freud-Str. 25, 53105 Bonn, Germany
5To
whom correspondence should be addressed
Experimental perfusion of various organs has primarily
been used in transplantation medicine to study the physiology, pathophysiology and metabolism of tissues and cells.
The purpose of this study was to establish an experimental
model for the extracorporeal perfusion of the human
uterus with recirculation of a modified, oxygenated Krebs–
Henselait solution, in comparison with a non-recirculating
perfusion system. With consent of the patients we obtained
25 uteri after standard hysterectomy. We performed an
isovolumetric exchange of the perfusion medium at different intervals from 1 to 6 h and examined pH, pO2, pCO2,
lactate, lactate dehydrogenase and creatine kinase by taking
arterial and venous samples every hour for 24 h. We found
the perfusions to be adequate when maintaining flow rates
at 15–35 ml/min and at pressures ranging from 70 to
130 mmHg. Isovolumetric exchange of the perfusate
every 3–4 h was the maximum interval to keep pH, the
arterio-venous gradients of pO2 and pCO2, and the
other biochemical parameters in physiological ranges.
Examination by light and electron microscopy showed wellpreserved features of myometrial and endometrial tissue.
However, a 6 h exchanging interval led to increasing hypoxic
and cytolytic parameters during the whole perfusion period.
X-ray studies using digital subtraction angiography and
perfusion studies with methylene blue demonstrated the
homogeneous distribution of the perfusion fluid throughout
the entire organ.
Key words: extracorporeal perfusion/human uterus/physiological recirculation system
Introduction
Experimental perfusion models of various organs are a valuable
research tool in transplantation medicine and have primarily
been used to study the physiology, pathophysiology, and
metabolism of tissues and cells (Kamada et al., 1980; van der
Wjik et al., 1980; Toledo-Pereyra et al., 1982). Extracorporeal
perfusion of the human uterus, in particular, offers a new
experimental approach to studies of the myometrium, the
© European Society of Human Reproduction and Embryology
endometrium and the uterine vasculature. As demonstrated in
other organs such as heart, liver and kidney, it can be expected
that uterine perfusion with oxygenated media through the
uterine arteries could maintain tissue viability and responsiveness to hormones for prolonged periods of time (Iwasaki
et al., 1991; Marinelli et al., 1991; Bresticker et al., 1992;
Ohura et al., 1995; Balden et al., 1997). First investigations
have shown that physiological organ function of an extracorporeal perfused human uterus for longer periods following hysterectomy depends on a sufficient preservation of cell integrity
(Bulletti et al., 1986). Insufficient perfusion followed by
ischaemia causes cell damage in the perfused organ and
consecutive increase of hypoxic and cytolytic parameters.
The purpose of the present study was to establish a system for
extracorporeal perfusion of a human uterus with recirculation of
the perfusate and regeneration by interval exchange to guarantee tissue vitality of the perfused organ for up to 24 h.
Materials and methods
A total of 25 uteri obtained by abdominal or vaginal hysterectomy
for benign reasons was studied. Ethical approval by the EthikKommission of the University of Bonn was given under the number
146/97. Every patient signed an information sheet and gave signed
consent to the investigation. Patients with previous multiple abdominal
surgeries, pelvic endometriosis, pelvic inflammatory disease, large
fibroids, adenomyosis (⬎500 g) or malignant diseases were excluded.
The age of each patient and the stage of the cycle was established
by anamnestic means; in case of doubt hormone serum plasma
concentrations were investigated.
The surgical specimens were carefully prepared, avoiding traction
and laceration in order to get vascular stumps suitable for catheterization. Organ weights were always measured immediately before and
after perfusion. After cannulation of both uterine arteries with 14 G
bulb-headed cannulas, the uterus was flushed bilaterally with a
modified heparinized Krebs–Henseleit bicarbonate buffer at 37°C to
remove blood products and cell detritus caused by the operation
procedures. The composition of the perfusion medium was as follows:
pH 7.40, NaCl 6.89 g/l, KCl 0.37 g/l, MgSO4·7H2O 0.25 g/l,
CaCl2·2H2O 0.37 g/l, KH2PO4 0.14 g/l, NaHCO3 2.35 g/l, D(⫹)glucose 1.50 g/l, saccharose 0.70 g/l, glutatione 0.005 g/l, 1,4dithiothreitol 0.10 g/l, 100 IE insulin bolus at the beginning of
perfusion, refobacin 40 mg/l perfusate, heparin 150 IU/ml. Following
this ‘flush-perfusion’ the uterus was connected with the recirculating
perfusion system and placed in a cabinet (baby incubator, Dräger,
Köln, Germany) at 37°C with humidity values of ~97% on wire mesh
above a cylindrical teflon receptacle for perfusate collection under
sterile conditions. Oxygenation of the perfusion medium was accomplished by a membrane oxygenator (Jostra, Hirrlingen, Germany)
flushed with 95% O2 and 5% CO2 and monitored by gas flowmeters.
The perfusate was moved through silicone rubber tubing by separate
1235
O.Richter et al.
this recirculating perfusion model, isovolumetric exchange of the
perfusion medium after 1 (group 1), 2 (group 2), 4 (group 3) and 6
(group 4) h was compared to a non-recirculating system (group 0).
In each group five uterine perfusions were performed. After each
experiment we performed X-ray studies with opaque contrast medium
using digital subtraction angiography (DSA) and perfused each uterus
with methylene blue to show the distribution of the perfusion medium
throughout the organ. For statistical evaluation we used the Kruskal–
Wallis one-way analysis of variance for non-parametric data.
Figure 1. Schematic of the perfusion system. (1) Reservoir for
perfusion buffer. (2) Membrane oxygenator. (3) Gas mixture with
O2, CO2 and flowmeters. (4) Gas flowmeter. (5) Filter with bubble
trap. (6, 16, 17) Roller pumps with pressure transducers and
flowmeters. (7) Thermostatic heating element. (8) Temperature
transducers. (9) Reservoir for hormone solution. (10) Temperatureand humidity-controlled perfusion chamber. (11) Thermometer. (12)
Hygrometer. (13) Arterial catheters with sampling ports. (14)
Venous catheters with sampling ports. (15) Reservoir to collect
perfusate. (16) Recirculating pump. (17) Pump for perfusion with
different hormonal perfusion media. (18) Regulation element for
autoregulation between recirculation flow rate and effluate volume.
roller pumps (Storz, Tuttlingen, Germany) monitoring flow and
pressure rates throughout the perfusion period.
The recirculation was achieved by pumping the perfusion medium
from the teflon receptacle back to the perfusate reservoir with another
separate roller pump. In order to coordinate the effluate volume
and the recirculation flow rate, a regulating element triggered the
recirculation roller pump by measuring the filling level of the perfusate
in the collecting receptacle. For the description of the complete
perfusion system see Figure 1.
Arterial and venous samples of perfusion medium were taken every
60 min with syringes for measurements of pH, oxygen partial
pressure (pO2), carbon dioxide partial pressure (pCO2), lactate, lactate
dehydrogenase (LDH) and creatine kinase (CK).
Samples of pH, pO2 and pCO2 were analysed with an ABL 505
Blood Gas Analyzer (Radiometer, Copenhagen, Denmark); all other
biochemical parameters were measured using standard procedures
(Vitros-Analyzer, Johnson & Johnson, Rochester, NY, USA).
Several myometrium and endometrium biopsies for light and
electron microscopy were taken from the corpus uteri and the cervix
uteri with a biopsy needle at the beginning, during (at 6, 12 and 18 h
of perfusion) and at the end of perfusion.
For light microscopy the samples were fixed in 4% paraformaldehyde and then embedded in paraffin blocks. From these specimens
several sections of 4 µm thickness were stained with haematoxylin
and eosin (HE). For the electron microscopical investigations samples
were fixed in 2.2% glutaraldehyde and PBS buffer. Cubes of about
1 mm3 in size were cut from each specimen and embedded in Epon.
Thin sections were stained with uranyl acetate and lead citrate and
were examined with a Philips CM 10 electron microscope (Eindhoven,
The Netherlands).
To evaluate the best interval for regeneration of the perfusate in
1236
Results
The age of the hysterectomized patients was 28–56 years
(mean 42 ⫾ 8.6 years), all of them in a pre-menopausal status.
In 16 cases a hormone treatment was finished 2–6 weeks prior to
the operation; nine patients did not have any hormone therapy.
Perfusions were considered to be appropriate when constant
flow rates of 15–35 ml/min through each artery could be
maintained at pressure rates ranging from 70 to 130 mmHg.
The venous effluate was collected in a cylindrical teflon
reservoir from which the perfusate was recirculated by a
separate roller pump to the reservoir for the perfusion buffer
(Figure 1).
The biochemical parameters were initially elevated in all
groups due to preparation and cannulation time. Within the
first hours of perfusion, arterio-venous gradients of pH, pO2,
and pCO2 decreased significantly in groups 0, 1, 2 and 3 and
remained stable in the further course of perfusion as a sign of
physiological oxygen consumption of the perfused organ
(Figure 2a–c).
The concentrations of lactate, LDH and CK are shown in
Figure 3a–c. The initial hypoxia in all groups leading to
formation of lactic acid at the beginning of perfusion was
corrected in groups 0, 1, 2 and 3 as the perfusion proceeded
with oxygenated buffer (Figure 3a). Similarly, increased LDH
and CK levels as indicators for cytolytic tissue processes
occurring during the preparation and cannulation period subsided significantly in groups 0, 1, 2 and 3 with continuation
of the experiment (Figure 3b,c). In contrast, all parameters in
group 4 showed an adverse development due to the longer
perfusate exchanging period (Figures 2 and 3). These differences between groups 4 and 0, 1, 2 and 3 were significant
(α ⫽ 0.5; P ⬍ 0.0001).
Furthermore, examination by light and electron microscopy
showed well-preserved intracellular structures in myometrium
and endometrium without evidence of intracellular oedema in
comparison to the control tissue taken from corpus and cervix
immediately after hysterectomy in groups 0, 1, 2 and 3, but
not in group 4 (Figure 4a–g). There was no significant oedema
formation evaluated by comparison of the organ weights before
and after the perfusion (Table I).
In group 4 the light microscopic examination of endometrial
and myometrial samples taken after 24 h showed severe
regressive changes and partial necrosis with predominantly no
longer recognizable tissue structures (Figure 4c,f). Similiar
results were obtained by electron microscopical study of the
specimen. Up to a perfusate exchange interval of 4 h the
intracellular and intercellular tissue features did not significantly change in comparison to those seen in fresh samples
Extracorporeal perfusion of human uterus
Figure 2. (a) Arterio-venous gradients of pH in groups 1, 2, 3
and 4 (i.e. after 1, 2, 4 and 6 h respectively) during perfusion
compared to group 0 (a non-recirculating system). (b) Arteriovenous gradients of oxygen partial pressure (pO2) in each group
during perfusion. (c) Arterio-venous gradients of carbon dioxide
partial pressure (pCO2) in each group during perfusion.
Figure 3. (a) Concentrations of lactic acid in each group during
perfusion. (b) Concentrations of lactic dehydrogenase (LDH) in
each group during perfusion. (c) Concentrations of creatine kinase
in each group during perfusion.
(Figure 4g,h). Examination of the tissue samples of group 4,
however, showed irreversible damages e.g. degranulation of
rough endoplasmatic reticulum with disaggregation of free
polyribosomes and clarified mitochondrial matrix. Moreover,
frank cavitation with peripherally placed, disorientated, and
disintegrated cristae was observed.
1237
O.Richter et al.
Figure 4. Light and electron micrographs of myometrial and endometrial tissues. (a) Well-preserved corpus epithelium, group 3, intervalexchange after 4 h. Haematoxylin and eosin (HE); scale bar ⫽ 50 µm. (b) Control biopsy as a example for corpus tissue taken immediately
after removal of the uterus. HE; scale bar ⫽ 50 µm. (c) Autolytic corpus tissue, group 4, interval-exchange after 6 h. HE; scale bar ⫽ 50
µm. (d) Intact, mucinous cervix epithelium, group 4, interval-exchange after 4 h. HE; scale bar ⫽ 50 µm. (e) Control biopsy as an example
for cervix tissue taken immediately after removal of the uterus. HE; scale bar ⫽ 50 µm. (f) Dissociation of the cervix epithelium due to
autolysis, group 4, interval-exchange after 6 h. HE; scale bar ⫽ 50 µm. (g) Electron micrograph shows endometrial epithelial cells in the
proliferating phase of the cycle taken immediately after removal of the uterus. The fine chromatin structure and microvilli demonstrate wellpreserved epithelium; scale bar ⫽ 3.3 µm. (h) Specimen from the same uterus, group 3, interval exchange after 4 h. The electron
micrograph shows intact organelles of endometrial epithelial cells, i.e. mitochondria, rough endoplasmic reticulum, centrioles (arrow head),
intercellular junctions; scale bar ⫽ 1.8 µm.
The injected X-ray contrast fluid through both arteries showed
good distribution of the perfusate throughout the myometrium.
The uterus is well supplied with arcuate and radial arteries
crossing vertically and horizontally to form a network of small
1238
anastomoses which results in perfusion of the entire organ
(Figure 5a,b). Application of methylene blue demonstrated a
homogeneous distribution of the perfusion medium throughout
the corpus uteri including the endometrium (not shown).
Extracorporeal perfusion of human uterus
Discussion
Previous investigations emphasized the feasibility of perfusion
models for experimental examinations in the human placenta,
ovary and uterus (Bulletti et al., 1986, 1993; Page, 1991;
Brannstrom and Flaherty, 1995; Schmolling et al., 1997). The
experimental perfusion system of the human uterus described
by Bulletti et al. (1987a, 1988a) consisted of an open perfusion
circuit, i.e. without recirculation of the perfusate, with excessive
demand of perfusate volume, especially in the midterm and
longterm perfusion situation. In the present study we developed
an extracorporeal perfusion system of the human uterus with
recirculation of the perfusion medium allowing a reduction to
about a quarter of the necessary perfusate volume compared
with the previous model (Bulletti et al., 1987a, 1988a).
Table I. Numbers of the perfused uteri and their correponding organ
weights in grams before and after the perfusion experiment
No.
Before
After
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
78
67
98
77
105
212
87
138
178
94
162
81
124
72
107
68
88
150
60
138
93
65
245
116
89
82
69
96
85
112
217
96
151
190
103
177
89
136
74
110
77
90
162
64
146
104
71
258
119
94
In our modified experimental model, perfusions under
physiological flow and pressure conditions with modified
oxygenated Krebs–Ringer bicarbonate buffer could be maintained at 37°C for the entire perfusion period at a steady rate
of oxygen consumption, low output of lactate, LDH and CK
activity as markers for hypoxia and cytolytic processes and
preservation of tissue integrity. In contrast to current clinical
methods for extracorporeal preservation of organs by hypothermia, our purpose was to obtain an experimental model suitable
for metabolic studies. Cooling causes modifications of physiochemical functions and metabolic processes during prolonged
hypothermic storages: in particular, temperatures close to
freezing inactivate the sodium–potassium cell pump with
consecutive loss of cellular potassium and magnesium and a
corresponding sodium influx into the cell with cellular swelling.
In this study it was our intention to simulate the in-vivo
situation as close as possible. Therefore, human specimens
were used and uterine perfusion was performed at body
temperature with the organ placed in an automatically thermoand humidity-controlled incubator.
With the use of autoregulated perfusion pumps, with adjustable minimum and maximum levels for perfusion flow and
perfusion pressure rates according to physiological conditions,
microvascular damages such as capillary destruction could
be avoided.
In a perfusion system with recirculation of the perfusion
medium, the composition of the perfusate is of particular
importance. From the biological point of view the use of
undiluted heparinized blood appears to be preferable to synthetic media. However, in-vitro perfusions using blood can
cause problems in mid-term and long-term perfusion experiments regarding its physiological instability, the probable
interaction with experimental pharmacological substances and
the large volume required for perfusion.
Currently various preservation solutions, for example Euro–
Collins solution, UW–Belzer solution, Bretschneider HTK
solution and Krebs–Henseleit bicarbonate buffer solution are
used experimentally and clinically for cold storage preservation
of organs in transplantation medicine or investigations in
oncology (Marinelli et al., 1991; Bresticker et al., 1992;
Figure 5. Sequences from perfusion study of whole uterus with X-ray material imaged by digital subtraction angiography. (a) Early phase,
after 5 s; (b) late phase, after 15 s. Scale bar ⫽ 1.5 cm.
1239
O.Richter et al.
Collins and Wicomb, 1992; Erhard et al., 1994; Blech et al.,
1997; Collins, 1997). Based on the experiences of Bulletti
et al. (1988b, 1993) we used a modified Krebs–Henseleit
bicarbonate buffer with saccharose, glutathione, 1,4-dithiothreitol, insulin and refobacin and demonstrated that exchanging the perfusate every 4 h seems to be sufficient to maintain
organ architecture and tissue vitality under optimal conditions
of oxygenation and nutrition, thus avoiding infection problems.
Under these circumstances the necessary volume of the
expensive perfusion medium could be reduced significantly.
Furthermore, observing uterine tissue fragments by light
and electron microscopy, which show good preservation of the
endometrial and myometrial intracellular structures as well as
the absence of inter- and intracellular oedema up to 24 h of
perfusion time, confirm the results of our investigations.
Performing X-ray studies with the DSA technique and perfusions with methylene blue demonstrated the homogeneous
distribution of the perfusion medium throughout the entire
organ.
By avoiding the degeneration of the organ by thrombus
formation, problems with organ and perfusate temperature
(hypothermia), increased capillary resistance or capillary
destruction, unphysiological perfusion flow and perfusion pressure, insufficient gas exchange and relatively activated anaerobic metabolism and bacterial infection, we could maintain
the viability and function of the organ.
Previous investigations by Bulletti et al. (1988a,b, 1993,
1997) have shown the feasibility of the extracorporeal perfusion
of the human uterus for various purposes. In conclusion, our
model for the extracorporeal perfusion of the human uterus
with recirculation of the perfusate by isovolumetric interval
exchange every 4 h represents a reproducible experimental
system which offers, especially in mid-term and long-term
perfusion experiments, a physiological and economical scientific approach for further observations on the endometrium and
myometrium in reproductive medicine.
Acknowledgements
This research is supported by the Deutsche Forschungsgemeinschaft
(DFG), Gz RI 958/1-1 and the BONFOR-Forschungskommission, Gz
103/17.
References
Balden, N., Toffano, M., Cadrobbi, R. et al. (1997) Kidney preservation in
pigs using celsior, a new organ preservation solution. Transplant. Proc.,
29, 3539–3540.
Blech, M., Hummel, G., Kallerhoff, M., Ringert, R.H. (1997) Electrolyte
equilibration of human kidneys during perfusion with HTK-solution
according to Bretschneider. Urol. Res., 25, 331–335.
Brannstrom, M. and Flaherty, S. (1995) Methodology and characterization of
an in vitro perfusion model for the mouse ovary. J. Reprod. Fertil., 105,
177–183.
Bresticker, M.A., LoCicereo, J., Oba, J. and Greene, R. (1992) Successful
extended lung preservation with UW solution. Transplantation, 54, 780–784.
Bulletti, C., Jasonni, V.M., Lubicz, S. et al. (1986) Extracorporeal perfusion
of the human uterus. Am. J. Obstet. Gynecol., 154, 683–688.
Bulletti, C., Jasonni, V.M., Martinelli, G. et al. (1987) A 48-hour preservation
of an isolated human uterus: endometrial responses to sex steroids. Fertil.
Steril., 47, 122–129.
Bulletti, C., Jasonni, V.M., Tabanelli, S. et al. (1988a) Early human pregnancy
in vitro utilizing an artificial perfused uterus. Fertil. Steril., 49, 991–996.
1240
Bulletti, C., Jasonni, V.M., Ciotti, P.M. et al. (1988b) Extraction of estrogens
by human perfused uterus: Effects of membrane permeability and binding
by serum proteins on differential influx into endometrium and myometrium.
Am. J. Obstet. Gynecol., 159, 509–515.
Bulletti, C., Prefetto, R.A., Bazzocchi, G. et al. (1993) Electromechanical
activities of human uteri during extra-corporeal perfusion with ovarian
steroids. Hum. Reprod., 8, 1558–1563.
Bulletti, C., de Ziegler, D., Flamigni, C. et al. (1997) Targeted drug delivery
in gynaecology: the first uterine pass effect. Hum. Reprod., 12, 1073–1079.
Collins, G.M. (1997) What solutions are the best? Overview of flush solutions.
Transplant. Proc., 29, 3543–3544.
Collins, G.M. and Wicomb, W.N. (1992) New organ preservation solutions.
Kidney Int., 38 (Suppl.), 197–202.
Erhard, J., Lange, R., Scherer, R. et al. (1994) Comparison of histidinetryptophan-ketoglutarate (HTK) solution versus University of Wisconsin
(UW) solution for organ preservation in human liver transplantation. A
prospective, randomized study. Transplant. Int., 7, 177–181.
Iwasaki, S., Araki, H., Miyauchi, Y. and Nishi, K. (1991) 24-hour preservation
of isolated rat hearts perfused with Krebs–Henseleit solution and
pyridoxalated hemoglobin polyoxyethylene conjugate PHP solution at low
temperature. Artif. Organs, 15, 78–85.
Kamada, N., Calne, R.Y., Wight, D.G.D. and Lines, J.G. (1980) Orthotopic
rat liver transplantation after long-term preservation by continuous perfusion
with fluocarbon emulsion. Transplantation, 30, 43–48.
Marinelli, A., Dijkstra, F.R., van Dierendonck, J.H. et al. (1991) Effectiveness
of isolated liver perfusion with mitomycin C in the treatment of liver
tumors of rat colorectal cancer. Br. J. Cancer, 64, 74–78.
Ohura, H., Kondo, T., Handa, M. et al. (1995) Functional and histopathologic
studies of primate pulmonary allografts preserved for 24 h with a form of
extracellular solution. J. Heart Lung Transplant., 14, 493–504.
Page, K.R. (1991) Perfusion of isolated human placenta. Proc. Nutr. Soc., 50,
345–347.
Schmolling, J., Jung, S., Schlebusch, H. et al. (1997) Modification of
transplacental digoxin transfer in the isolated placental lobule. Z.
Geburtshilfe Neonatol., 201 (Suppl. 1), 9–12.
Toledo-Pereyra, L.H. (ed.) (1982) Basic Concepts of Organ Procurement,
Perfusion and Preservation for Transplantation. Academic Press, New York.
Van der Wjik, J., Sloof, M.J.H., Rijkmans, B.G. and Koostra, G. (1980)
Successful 96 and 144 h experimental kidney preservation and newly
developed normotermic ex vivo perfusion. Cryobiology, 17, 473–482.
Received on September 14, 1999; accepted on March 1, 2000