- Email: [email protected]
Factors effective in reducing rat mortality due to acute liver failure as induced by d -galactosamine poisoning
JOURNAL
OF SURGICAL
Factors
Department
RESEARCH
36, 37 l-376
( 1984)
Effective in Reducing Rat Mortality Failure as Induced by D-Galactosamine PAULITA
LAPLANTE AND DAVID
of Surgery,
University
Presented
Due to Acute Poisoning’
O’NEILL, B.A., PETER L. BLANC, E. R. SUTHERLAND, M.D., PH.D. of Minnesota
at the Annual Syracuse,
Health
Sciences
Center,
M.D.,
Minneapolis,
Meeting of the Association of Academic New York, November 2-5, 1983
Liver
Minnesota
55455
Surgery,
Mortality from Bgalactosamine hydrochloride (DGalN)-induced acute liver failure (ALF) in rats can be reduced by ( 1) transplanting intact hepatocytes; (2) injecting cytosol from fractionated hepatocytes dispersed from a liver subjected to 70% hepatectomy 24 hr earlier (CYT-H); (3) injecting a cell-free supemate derived from cultured hepatocytes (SUP); or (4) injecting neuraminidase-treated plasma where the plasma is drawn 24 hr after donor rats are subjected to 70% hepatectomy (PHP-neu). These treatments are effective when administered 20-24 hr after DGalN poisoning, but experiments to determine the alterations in mortality rates as a function of time of treatment in relation to poisoning have not been performed. Two experiments are reported here. In the first the survival of lethally poisoned rats was compared after intravenously injecting either SUP prepared from cultured hepatocytes of syngeneic adult or fetal rat sources, or CYT-H from syngeneic adult rats 20 hr after poisoning. Untreated rats, rats treated with culture media alone, or rats treated with CYT from a nonregenerating source had an 88-100% mortality, with all deaths occurring within 72 hr following poisoning. Improved survival followed all other treatments: 55% of the rats receiving adult SUP, 70% of the rats receiving fetal SUP, and 80% of the rats receiving CYT-H survived. In the second experiment the survival of poisoned rats was compared after injecting them with PHP-neu, PHP not treated with neuraminidase (PHP without neu), and neuraminidase-treated plasma from sham-operated rats (SP-neu) or normal, nonoperated rats (NP-neu) 20 hr after poisoning. o-GalN rats not receiving any subsequent injection, DGalN rats receiving PHP without neu, NP-neu, or SP-neu, all experienced a 63-888 mortality within 72 hr. Improved survival occurred in rats receiving PHP-neu, and the mortality was only 43%. Treatment of PHP-neu with proteolytic enzymes partially abrogated the effect, and the mortality rate was 67% in this group. These studies indicate that the o-GalN-poisoned liver can recover sufficiently to sustain life when either (1) hepatectomy-stimulated humoral or intrahepatocellular substances or (2) liver cellculture-stimulated supemate is injected. Whether the factors active in the various preparations are the same or different is uncertain, and studies to isolate, purify, characterize, and compare the active compounds, in each preparation, are needed. Such studies may also be a prelude to development of a clinically useful treatment for ALF.
Acute liver failure (ALF) requires therapies to support or replace lost function until recovery occurs, but these therapies are not yet completely available or adequately defined in terms of specific action mechanisms. A variety of treatments have been found to be effective in one model of acute hepatic failure, the Dgalactosamine (D-GalN)-poisoned rat [9, 12, 141. The mechanism by which each of these treatments enhances survival may be similar, ’ Supported in part Medical Association.
by a grant
from
and the work presented here was undertaken in an attempt to develop a unifying hypothesis. In previous studies we found DGalN-poisoned rats injected with intact live hepatocytes had a higher survival rate than nontreated rats [ 141. The mechanism by which intact cells improved the survival rate was not clear, but, in experiments by Makowka et al. [ 121, it appeared that liver cells neither repopulated the recipient liver following intraportal injection nor did they establish a “temporary” organ upon intraperitoneal injection. In further experiments we found that cultured liver cells
the Minnesota 371
0022-4804/84
$1 SO
Copyright 8 1984 by Academic Press, Inc. All rights of reproduction in any form reserved.
372
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. 4, APRIL 1984
manufacture, or release, a factor(s) that could be injected to reduce the mortality of D-G~~N poisoning [9]. As a corollary, Makowka et al. [ 11, 121 found that injection of a cell-free cytosol preparation, obtained from fractionated, regenerating liver cells 16 hr after partial hepatectomy of otherwise normal donor rats, improved the survival of DGalN-poisoned rats. Experiments reported here compare the survival of D-GalN-poisoned control rats with poisoned rats receiving one or the other of the following injections 20 hr after D-GalN: (1) cell-free supernatant from adult or fetal rat liver cells in culture, (2) cell-free cytosol from regenerating liver cells, or (3) plasma drawn from rats 23 hr after partial hepatectomy and then incubated with neuraminidase. The latter treatment plan was devised because of the observation of Goldberg et al. [5] that such plasma stimulated DNA synthesis specifically in the liver (and not in other organs) when injected into normal rats. MATERIALS
AND
METHODS
Animals. Male F344 rats (Harlan SpragueDawley, Walkersville, Md.), 2 to 4 months old, weighing between 175 and 250 g were used in all experiments. Liver failure model. Rats were poisoned by a single, intraperitoneal injection of 0.75 g Dgalactosamine hydrochloride, Type I (Sigma Chemical Company, St. Louis, MO.) per kilogram body weight. In previous experiments, we found this dose to result in approximately a 90% mortality in rats receiving no further treatment [9]. Preparation of hepatocyte culture supemate.
The method of hepatocyte preparation for culture by a collagenase perfusion digestion technique from adult rat livers has been previously described [9]. Fetal hepatocytes were prepared for culture by a completely different method. The livers of 30, 3-day-old male and female F344 donors were excised, minced, and repeatedly washed in a calcium-free solution containing 2% FCS and 0.6 mM EGTA. The minced tissue was divided into three aliquots of approximately 1 g each and suspended in
flasks containing 7 ml of 5 mM CaC12, 1.5 mg/ml collagenase Hanks’ solution. A thermoshake bath rotated the flasks at 39°C for approximately 25 mitt, but was stopped as soon as it was obvious to us that the fragments were disintegrating. Further manipulations of the fetal cells were the same as those described for adult cells [9]. After 36 hr of culture in Waymouth’s medium, 2 ml of culture supernatants from adult (Group I) or fetal (Group II) hepatocyte preparations were injected intravenously into poisoned rats. Preparation of hepatocyte cytosol. Twenty hours after subjecting hepatocyte donors to 70% hepatectomy by the method of Higgins and Anderson [ 81, we enzymatically dispersed intact hepatocytes from the remaining lobes by the collagenase perfusion technique as previously described (9). Approximately 10 ml of a cell suspension was obtained from each liver, and 70-95% of the cells were viable as determined by trypan blue exclusion. These cells were then fractionated using a sonicator (Heat Systems Ultra-Sonic, Plainville, N. Y.) so that no preparation contained intact cells upon microscopic examination. Using sucrose density gradient and centrifugation separations, we were able to obtain cell-free cytosol from normal (Group VI) or partially hepatectomized (Group V) donors for intravenous injection into poisoned rats (1 ml/rat). Preparation of plasma from partially hepatectomized rats (PHP). Modifying methods described by Goldberg et al. [5], we partially
hepatectomized rats under Nembutal narcosis, and then exsanguinated them 23 hr later. Using a ratio of 2 donors for each recipient, we pooled the donor plasma, acidified it to pH 5.5, and heated it to 95°C for 20 min to denature residual protein. After the plasma had cooled to room temperature, we pelleted denatured proteins by centrifuging the sample for 20 min at 3500g. We resuspended the pellet after decanting the supematant and subjected the pellet suspension to the same denaturation and centrifugation. Supematants from both runs were then pooled, lyophilized, and reconstituted in distilled water to half the original plasma volume. Plasma from sham-op
LA PLANTE,
BLANC, AND SUTHERLAND:
erated or normal rats was treated in a similar manner. Treatment of PHP. PHP extract treatment with neuraminidase was modified only slightly from that described by Goldberg et al. [5]. Briefly, we acidified and then incubated PHP with neuraminidase (Type V, Sigma) at 37°C for 1 hr prior to heat inactivating excess neuraminidase. The PHP-neu preparation was then centrifuged and the supematant was lyophilized. We reconstituted the resulting powder in bacteriostatic water, pH 7.2, so that 1 ml contained PHP-neu derived from 2 donors and this amount was injected intravenously into the poisoned rats of Group VIII. PHP not treated with neuraminidase is referred to as PHP without neu (Group IX). Plasma from sham (n = 4) or unoperated (n = 4) rats was treated with neuraminidase in the same manner (Group X). PHP not treated with neuraminidase (Group IX) was subjected to the incubation, heat inactivation, and centrifugation steps as in Group VIII. Proteolytic treatment of PHP-neu. After alkalinization to pH 7.6 with 0.1 NNaOH, we subjected PHP-neu to either of two enzymatic proteolytic treatments: simultaneous incubation with trypsin (Sigma, Type III) and chymotrypsin (Sigma, Type I-S), or serial incubations with trypsin and protease (Sigma, Type XIV). In both the trypsin/chymotrypsin (n = 4) and trypsin/protease (n = 5) groups, incubation was carried out at 37°C for a total of 2 hr; greater than 20 PM units (as defined by the manufacturer) of each enzyme per donor rat was used in either case. Following heat inactivation, centrifugation, and supematant lyophilization, we dissolved the lyophilizates in bacteriostatic water for intravenous injection into poisoned rats (Group XI). Gel filtration of PHP-neu. Using Sephadex G- 100 and G-200 (Pharmacia, Upsalla, Sweden) we gel filtered PHP-neu in an effort to establish molecular weight ranges of the factor(s) remaining in the plasma preparation and to compare PHP-neu gel filtration patterns obtained in our laboratory with those described by Goldberg et al. [ 51. We dissolved the lyophilized pooled extract from 6 rats in
D-GALACTOSAMINE
373
5 ml of 50 mM TRIS buffer (pH 7.6) before applying it to a column (2.5 cm X 90 cm) and eluting it with Tris (elution velocity, 50 ml/hr). We then collected 5-ml fractions, determined optical absorption at 280 nm, and concentrated pooled fractions from Peaks I, II, III over Diaflo membranes (Amicon, Lexington, Mass.) using PM-10 ( 10,000 Da exclusion limit membrane) for Peaks I, II, and UM-05 (500 Da exclusion limit membrane) for Peak III (Fig. 1). In a small pilot experiment, DGalN-poisoned rats received pooled, concentrated peak fractions intravenously. We calibrated column molecular weights by the method of Andrews [ 11 using the following reference proteins (Pharmacia, Uppsala, Sweden): bovine serum albumin (67,000 Da), ovalbumin from hen egg (43,000 Da), chymotrypsinogen A from bovine pancreas (25,000 Da), and ribonuclease A from bovine pancreas (13,700 Da). Statistical analysis. Fischer’s exact test was used to determine the significance of differences in the proportions of animals surviving between the various experimental groups. RESULTS
Rat survival as a function of SUP, CYT-H, or PHP-neu injection. In Table 1, data on the survival of lethally poisoned rats after injecting either SUP from syngeneic adult (Group II) or fetal (Group III) rats, or CYT-H (Group V) from syngeneic adult rats is given. Untreated rats (Group I), rats treated with culture
FIG. 1. Elution profile of neuraminidax-treated plasma from partially hepatectomized rats. G- 100 Sephadex column, VT = 438 ml (50 mm Tris buffer, pH 7.6).
314
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. 4, APRIL 1984 TABLE 1
SURVIVAL OF DGalN-POISONED RATS TREATED wrr~ LIVER C~TOSOL EXTRACTS OR LIVER CULTURE SUPERNATANTS20 hr AFTER POISONING' Group I
Donor cell source and treatment
No. survived/ No. treated
No treatment
3125 (12%)
II
Adult hepatocyte culture supemate
6/11 (55%)
III
Fetal hepatocyte culture supemate
7/10 (70%)
IV
Culture media alone, no cells
l/14 (7%)
V
Regenerating liver cytosol
415 (80%)
Nonregenerating liver cytosol
O/5 (0%)
VI
’ Rats that died did so within 72 hr. P-c0.01 for Groupz II, III, V vs Groups I, IV, and VI.
media alone (Group IV), or rats treated with CYT from a nonregenerating liver source (Group VI) experienced an 88-100% mortality, with all deaths occurring within 72 hr after poisoning. Mortality was reduced following all other treatments (Groups II, III, V): only 45% of rats receiving adult SUP died, only 30% of rats injected with fetal SUP died, and only 20% of rats receiving CYT-H died. In Table 2, data on the survival of poisoned rats after injecting them with PHP-neu (Group VIII), PHP without neu (Group IX), NP-neu, or SP-neu (Group X), or PHP-neu digested with proteolytic enzymes (Group XI)
is given. The DGalN-poisoned rats not receiving any subsequent injection (Group VII) or those injected with plasma prepared from nonhepatectomized rats (Group X) had mortality rates of 85 and 88%, respectively. The mortality rate was significantly reduced to 43% in rats of Group VIII treated with PHP-neu. The mortality rate was also reduced in the rats treated with PHP without neu (Group IX), 63%, and in the rats given PHP-neu subjected to tryptic digestion (Group XI), 67% (2/4 rats given PHP-neu treated with trypsin/ chymotrypsin and 4/5 given PHP-neu treated with protease died). The overall results in the latter two groups were intermediate between those of Group VIII and Groups VII and X. Gel filtration results. We obtained a consistent elution profile (Fig. 1) for PHP-neu extract with both Sephadex G- 100 and G-200 gel filtration. Peak I eluted with the column void volume. Peak II represents a broad increase in eluent absorption over a 30,000- to 90,000-Da molecular weight range. Peak III falls in a molecular weight range less than 10,000 Da. Although we do not show the data here (manuscript in preparation), when we injected pooled fractions from Peak II into poisoned rats, the survival rate was enhanced to a degree comparable to that observed in poisoned rats injected with unfractionated PHP-neu extract (Group VIII). DISCUSSION
This study demonstrates that the injection of either ( 1) syngeneic cell culture supematant,
TABLE 2 SURVIVAL OF D-GalN-POISONED RATS TREATED WITH VARIOUS
PLASMA PREPARATIONS 20 hr AITER POISONING’
Group
Donor rat treatment
Plasma treatment
VII VIII IX X XI
None Hepatectomy Hepatectomy No operation or sham operation Hepatectomy
None Neuraminidase No neuraminidase Neuraminidase Neuraminidase + Proteolytic digestion
No. survived/ No. treated 3/20 8/14 318 l/8 3/9
(15%) (57%) (37%) (12%) (33%)
’ Rats that died did so within 72 hr. P < 0.05 for Group VIII vs VII and X. P z- 0.1 for Group IX and XI vs VII, VIII. and X.
LA PLANTE,
BLANC, AND SUTHERLAND:
(2) cell-free cytosol of regenerating liver cells, or (3) neuraminidase-treated plasma from partially hepatectomized rats is effective treatment in reducing mortality induced by Dgalactosamine in rats. We deliberately avoid labeling any of the above treatments as hepatoregenerative, although we favor the hypothesis that the injected material stimulates or accelerates recovery of liver function. Conversely, injecting any of the isolates may merely provide or initiate endogenous production of some unknown factor(s) which prolongs life until the D-GalN-poisoned liver regenerates of its own accord. Support for the regenerative hypothesis comes from Goldberg et al. [5] in their assertion that PHP-neu is a liver-specific initiator of DNA synthesis and mitosis in normal rat liver. They also claim that the active factor is a neuraminidase induced p-galactose residue which binds to /3-galactose receptors on liver cells, thus causing proliferation. The results of our experiments, in a bioassay model, are consistent with their hypothesis, but we have not performed experiments that could prove or disprove the hypothesis. We agree that hepatectomy activates or stimulates assayable cellular and humoral growth factors. Even though Goldberg et al. [5] found neuraminidase treatment essential to show an effect of PHP in their model, it is not clear that neuraminidase treatment is necessary to improve survival in our liver poison model, since a slight effect on improving survival was seen even without neuraminidase treatment. Even though the mortality was significantly reduced only in the group receiving PHP-neu, we administered only one dose at one concentration in all groups, and a greater effect might be seen with an increase in dose or with multiple injections in both the PHP-neu group or the PHP without neu or PHP-neu proteolytic digestion groups. The validity of this conjecture remains to be tested experimentally. Factors that contribute to liver regeneration are well summarized by Golde et al. [6]. The discovery of these factors dates back to the classic cross circulation experiments performed by Fisher et al. [4], Lieberman [lo],
DGALACTOSAMINE
375
and Bucher et al. [2]. All demonstrated the existence of humoral factors in partially hepatectomized animals which stimulated DNA synthesis and mitosis in normal animals. These investigators isolated some of the cellular and humoral components elicited by partial hepatectomy, and demonstrated their effects on DNA synthesis and mitosis, but failed to implicate any one as the prime or sole initiator of liver regeneration in vivo. O’Keefe and co-workers [3, 131 exemplify this type of pursuit in their characterization and mechanistic studies of epidermal growth factor (EGF). EGF is one component of the complex signal which regulates liver regeneration. Similarly, our extracts may require the concerted action of other regulatory mechanisms in the recipient rats to produce increased DNA synthesis and mitosis. We need then, to purify and characterize these extracts with the aim of determining their similarity and of demonstrating a hepatoregenerative effect, or for that matter, a liver-specific effect of any kind, in vivo or in vitro. A reliable in vitro assay is needed; in vivo (bio)assays are complicated by the enormous number of factors which influence liver function and which are not easily excluded from impinging on the final outcome of an assay designed to study the effect of an isolated substance [7, 151. A necessary requirement for the production of factors which stimulate hepatic regeneration may be a liver deficiency state. In vitro hepatocyte culture may mimic the partial hepatectomized state in vivo, and if the active substances produced by the two techniques turned out to be the same it would indicate that the liver itself can be the source of regenerative stimuli. REFERENCES 1. Andrews, D. Estimation of the molecular weights of proteins by Sephadex gel-filtration. Biochem. .I 91: 222, 1964. 2. Bucher, N. L. R., Patel, U., and Cohen, S. Hormonal factors and liver growth. Advan. Enzyme Regulalion 16: 205,
1978.
3. Earp, H. S., and O’Keefe, E. J. Epidermal growth factor receptor number decreases during rat liver regeneration. J. Clin. Invest. 67: 1580, 1981.
376
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. 4, APRIL 1984
Fisher, B., Szuch, D., Levine, M., and Fisher, E. R. A portal blood factor as the humoral agent in liver regeneration. Science 171: 575, 1971. 5. Goldberg, M., Strecker, W., Feeny, D., and Ruhenstroth-Bauer, G. Evidence for and characterization of a liver cell proliferative factor from blood plasma of partially hepatectomized rats. Horm. Metab. Res. 12: 4.
94, 1980. 6.
12.
Golde, D. W., Herschman, H. R., Lusis, A. J., and Groopman, J. E. Growth factors. Ann. Intern. Med. 92: 650,
1980.
Greisler, H. P., Voorhees, A. B., and Price, J. B. The nonportal origin of the factors initiating hepatic regeneration. Surgery 86: 2 10, 1979. 8. Higgins, G. M., and Anderson, R. M. Experimental pathology of the liver of the white rat following partial surgical removal. Arch. Pathol. 12: 186, 1931. 9. LaPlante O’Neill, P. L., Baumgartner, D., Lewis, W. I., Zweber, B. A., and Sutherland, D. E. R. Cellfree supematant from hepatocyte cultures improves survival of rats with chemically induced acute liver failure. J. Surg. Res. 32: 347, 1982. 10. Lieberman, I. Studies on control of mammalian deoxyribonucleic acid synthesis. In R. Baserga (Ed.), 7.
11.
13.
14.
Biochemistry of Cell Division, Springfield, III.: Thomas, 1969. Makowka, L., Falk, R. E., Blandis, L. M., Langer, B., Falk, J. A., Rotstein, L. E., and Phillips, M. J. The use of liver cytosol fractions in the treatment of acute liver failure. Surgery 91(I): 170, 1982. Makowka, L., Rotstein, L. E., Falk, R. E., Falk, J. A., Zuk, R., Langer, B., Blendis, L. M., and Phillips, M. J. Allogeneic and xenogeneic hepatocyte transplantation. Transplant. Proc. 13: 855, 198 I. O’Keefe, E. J., Holler&r& M. D., and Cuatnecasas, P. Epidermal growth factor: Characteristics of specific binding in membranes from liver, placenta, and other target tissues. Arch. B&hem. Biophys. 164: 5 18, 1974. Sommers, B. G., Sutherland, D. E. R., Matas, A. J., Simmons, R. L., and Najarian, J. S. Hepatocellular transplantation for treatment of Dgalactosamine induced acute liver failure in rats. Transplant. Proc. 11: 578, 1979.
15. Statzl, T. E., Terblanche, J., Porter, K. A., Jones, A. F., Usui, S., and Mazzoni, G. Growth stimulating factor in regenerating canine liver. Luncet 1: 127, 1979.
OF SURGICAL
Factors
Department
RESEARCH
36, 37 l-376
( 1984)
Effective in Reducing Rat Mortality Failure as Induced by D-Galactosamine PAULITA
LAPLANTE AND DAVID
of Surgery,
University
Presented
Due to Acute Poisoning’
O’NEILL, B.A., PETER L. BLANC, E. R. SUTHERLAND, M.D., PH.D. of Minnesota
at the Annual Syracuse,
Health
Sciences
Center,
M.D.,
Minneapolis,
Meeting of the Association of Academic New York, November 2-5, 1983
Liver
Minnesota
55455
Surgery,
Mortality from Bgalactosamine hydrochloride (DGalN)-induced acute liver failure (ALF) in rats can be reduced by ( 1) transplanting intact hepatocytes; (2) injecting cytosol from fractionated hepatocytes dispersed from a liver subjected to 70% hepatectomy 24 hr earlier (CYT-H); (3) injecting a cell-free supemate derived from cultured hepatocytes (SUP); or (4) injecting neuraminidase-treated plasma where the plasma is drawn 24 hr after donor rats are subjected to 70% hepatectomy (PHP-neu). These treatments are effective when administered 20-24 hr after DGalN poisoning, but experiments to determine the alterations in mortality rates as a function of time of treatment in relation to poisoning have not been performed. Two experiments are reported here. In the first the survival of lethally poisoned rats was compared after intravenously injecting either SUP prepared from cultured hepatocytes of syngeneic adult or fetal rat sources, or CYT-H from syngeneic adult rats 20 hr after poisoning. Untreated rats, rats treated with culture media alone, or rats treated with CYT from a nonregenerating source had an 88-100% mortality, with all deaths occurring within 72 hr following poisoning. Improved survival followed all other treatments: 55% of the rats receiving adult SUP, 70% of the rats receiving fetal SUP, and 80% of the rats receiving CYT-H survived. In the second experiment the survival of poisoned rats was compared after injecting them with PHP-neu, PHP not treated with neuraminidase (PHP without neu), and neuraminidase-treated plasma from sham-operated rats (SP-neu) or normal, nonoperated rats (NP-neu) 20 hr after poisoning. o-GalN rats not receiving any subsequent injection, DGalN rats receiving PHP without neu, NP-neu, or SP-neu, all experienced a 63-888 mortality within 72 hr. Improved survival occurred in rats receiving PHP-neu, and the mortality was only 43%. Treatment of PHP-neu with proteolytic enzymes partially abrogated the effect, and the mortality rate was 67% in this group. These studies indicate that the o-GalN-poisoned liver can recover sufficiently to sustain life when either (1) hepatectomy-stimulated humoral or intrahepatocellular substances or (2) liver cellculture-stimulated supemate is injected. Whether the factors active in the various preparations are the same or different is uncertain, and studies to isolate, purify, characterize, and compare the active compounds, in each preparation, are needed. Such studies may also be a prelude to development of a clinically useful treatment for ALF.
Acute liver failure (ALF) requires therapies to support or replace lost function until recovery occurs, but these therapies are not yet completely available or adequately defined in terms of specific action mechanisms. A variety of treatments have been found to be effective in one model of acute hepatic failure, the Dgalactosamine (D-GalN)-poisoned rat [9, 12, 141. The mechanism by which each of these treatments enhances survival may be similar, ’ Supported in part Medical Association.
by a grant
from
and the work presented here was undertaken in an attempt to develop a unifying hypothesis. In previous studies we found DGalN-poisoned rats injected with intact live hepatocytes had a higher survival rate than nontreated rats [ 141. The mechanism by which intact cells improved the survival rate was not clear, but, in experiments by Makowka et al. [ 121, it appeared that liver cells neither repopulated the recipient liver following intraportal injection nor did they establish a “temporary” organ upon intraperitoneal injection. In further experiments we found that cultured liver cells
the Minnesota 371
0022-4804/84
$1 SO
Copyright 8 1984 by Academic Press, Inc. All rights of reproduction in any form reserved.
372
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. 4, APRIL 1984
manufacture, or release, a factor(s) that could be injected to reduce the mortality of D-G~~N poisoning [9]. As a corollary, Makowka et al. [ 11, 121 found that injection of a cell-free cytosol preparation, obtained from fractionated, regenerating liver cells 16 hr after partial hepatectomy of otherwise normal donor rats, improved the survival of DGalN-poisoned rats. Experiments reported here compare the survival of D-GalN-poisoned control rats with poisoned rats receiving one or the other of the following injections 20 hr after D-GalN: (1) cell-free supernatant from adult or fetal rat liver cells in culture, (2) cell-free cytosol from regenerating liver cells, or (3) plasma drawn from rats 23 hr after partial hepatectomy and then incubated with neuraminidase. The latter treatment plan was devised because of the observation of Goldberg et al. [5] that such plasma stimulated DNA synthesis specifically in the liver (and not in other organs) when injected into normal rats. MATERIALS
AND
METHODS
Animals. Male F344 rats (Harlan SpragueDawley, Walkersville, Md.), 2 to 4 months old, weighing between 175 and 250 g were used in all experiments. Liver failure model. Rats were poisoned by a single, intraperitoneal injection of 0.75 g Dgalactosamine hydrochloride, Type I (Sigma Chemical Company, St. Louis, MO.) per kilogram body weight. In previous experiments, we found this dose to result in approximately a 90% mortality in rats receiving no further treatment [9]. Preparation of hepatocyte culture supemate.
The method of hepatocyte preparation for culture by a collagenase perfusion digestion technique from adult rat livers has been previously described [9]. Fetal hepatocytes were prepared for culture by a completely different method. The livers of 30, 3-day-old male and female F344 donors were excised, minced, and repeatedly washed in a calcium-free solution containing 2% FCS and 0.6 mM EGTA. The minced tissue was divided into three aliquots of approximately 1 g each and suspended in
flasks containing 7 ml of 5 mM CaC12, 1.5 mg/ml collagenase Hanks’ solution. A thermoshake bath rotated the flasks at 39°C for approximately 25 mitt, but was stopped as soon as it was obvious to us that the fragments were disintegrating. Further manipulations of the fetal cells were the same as those described for adult cells [9]. After 36 hr of culture in Waymouth’s medium, 2 ml of culture supernatants from adult (Group I) or fetal (Group II) hepatocyte preparations were injected intravenously into poisoned rats. Preparation of hepatocyte cytosol. Twenty hours after subjecting hepatocyte donors to 70% hepatectomy by the method of Higgins and Anderson [ 81, we enzymatically dispersed intact hepatocytes from the remaining lobes by the collagenase perfusion technique as previously described (9). Approximately 10 ml of a cell suspension was obtained from each liver, and 70-95% of the cells were viable as determined by trypan blue exclusion. These cells were then fractionated using a sonicator (Heat Systems Ultra-Sonic, Plainville, N. Y.) so that no preparation contained intact cells upon microscopic examination. Using sucrose density gradient and centrifugation separations, we were able to obtain cell-free cytosol from normal (Group VI) or partially hepatectomized (Group V) donors for intravenous injection into poisoned rats (1 ml/rat). Preparation of plasma from partially hepatectomized rats (PHP). Modifying methods described by Goldberg et al. [5], we partially
hepatectomized rats under Nembutal narcosis, and then exsanguinated them 23 hr later. Using a ratio of 2 donors for each recipient, we pooled the donor plasma, acidified it to pH 5.5, and heated it to 95°C for 20 min to denature residual protein. After the plasma had cooled to room temperature, we pelleted denatured proteins by centrifuging the sample for 20 min at 3500g. We resuspended the pellet after decanting the supematant and subjected the pellet suspension to the same denaturation and centrifugation. Supematants from both runs were then pooled, lyophilized, and reconstituted in distilled water to half the original plasma volume. Plasma from sham-op
LA PLANTE,
BLANC, AND SUTHERLAND:
erated or normal rats was treated in a similar manner. Treatment of PHP. PHP extract treatment with neuraminidase was modified only slightly from that described by Goldberg et al. [5]. Briefly, we acidified and then incubated PHP with neuraminidase (Type V, Sigma) at 37°C for 1 hr prior to heat inactivating excess neuraminidase. The PHP-neu preparation was then centrifuged and the supematant was lyophilized. We reconstituted the resulting powder in bacteriostatic water, pH 7.2, so that 1 ml contained PHP-neu derived from 2 donors and this amount was injected intravenously into the poisoned rats of Group VIII. PHP not treated with neuraminidase is referred to as PHP without neu (Group IX). Plasma from sham (n = 4) or unoperated (n = 4) rats was treated with neuraminidase in the same manner (Group X). PHP not treated with neuraminidase (Group IX) was subjected to the incubation, heat inactivation, and centrifugation steps as in Group VIII. Proteolytic treatment of PHP-neu. After alkalinization to pH 7.6 with 0.1 NNaOH, we subjected PHP-neu to either of two enzymatic proteolytic treatments: simultaneous incubation with trypsin (Sigma, Type III) and chymotrypsin (Sigma, Type I-S), or serial incubations with trypsin and protease (Sigma, Type XIV). In both the trypsin/chymotrypsin (n = 4) and trypsin/protease (n = 5) groups, incubation was carried out at 37°C for a total of 2 hr; greater than 20 PM units (as defined by the manufacturer) of each enzyme per donor rat was used in either case. Following heat inactivation, centrifugation, and supematant lyophilization, we dissolved the lyophilizates in bacteriostatic water for intravenous injection into poisoned rats (Group XI). Gel filtration of PHP-neu. Using Sephadex G- 100 and G-200 (Pharmacia, Upsalla, Sweden) we gel filtered PHP-neu in an effort to establish molecular weight ranges of the factor(s) remaining in the plasma preparation and to compare PHP-neu gel filtration patterns obtained in our laboratory with those described by Goldberg et al. [ 51. We dissolved the lyophilized pooled extract from 6 rats in
D-GALACTOSAMINE
373
5 ml of 50 mM TRIS buffer (pH 7.6) before applying it to a column (2.5 cm X 90 cm) and eluting it with Tris (elution velocity, 50 ml/hr). We then collected 5-ml fractions, determined optical absorption at 280 nm, and concentrated pooled fractions from Peaks I, II, III over Diaflo membranes (Amicon, Lexington, Mass.) using PM-10 ( 10,000 Da exclusion limit membrane) for Peaks I, II, and UM-05 (500 Da exclusion limit membrane) for Peak III (Fig. 1). In a small pilot experiment, DGalN-poisoned rats received pooled, concentrated peak fractions intravenously. We calibrated column molecular weights by the method of Andrews [ 11 using the following reference proteins (Pharmacia, Uppsala, Sweden): bovine serum albumin (67,000 Da), ovalbumin from hen egg (43,000 Da), chymotrypsinogen A from bovine pancreas (25,000 Da), and ribonuclease A from bovine pancreas (13,700 Da). Statistical analysis. Fischer’s exact test was used to determine the significance of differences in the proportions of animals surviving between the various experimental groups. RESULTS
Rat survival as a function of SUP, CYT-H, or PHP-neu injection. In Table 1, data on the survival of lethally poisoned rats after injecting either SUP from syngeneic adult (Group II) or fetal (Group III) rats, or CYT-H (Group V) from syngeneic adult rats is given. Untreated rats (Group I), rats treated with culture
FIG. 1. Elution profile of neuraminidax-treated plasma from partially hepatectomized rats. G- 100 Sephadex column, VT = 438 ml (50 mm Tris buffer, pH 7.6).
314
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. 4, APRIL 1984 TABLE 1
SURVIVAL OF DGalN-POISONED RATS TREATED wrr~ LIVER C~TOSOL EXTRACTS OR LIVER CULTURE SUPERNATANTS20 hr AFTER POISONING' Group I
Donor cell source and treatment
No. survived/ No. treated
No treatment
3125 (12%)
II
Adult hepatocyte culture supemate
6/11 (55%)
III
Fetal hepatocyte culture supemate
7/10 (70%)
IV
Culture media alone, no cells
l/14 (7%)
V
Regenerating liver cytosol
415 (80%)
Nonregenerating liver cytosol
O/5 (0%)
VI
’ Rats that died did so within 72 hr. P-c0.01 for Groupz II, III, V vs Groups I, IV, and VI.
media alone (Group IV), or rats treated with CYT from a nonregenerating liver source (Group VI) experienced an 88-100% mortality, with all deaths occurring within 72 hr after poisoning. Mortality was reduced following all other treatments (Groups II, III, V): only 45% of rats receiving adult SUP died, only 30% of rats injected with fetal SUP died, and only 20% of rats receiving CYT-H died. In Table 2, data on the survival of poisoned rats after injecting them with PHP-neu (Group VIII), PHP without neu (Group IX), NP-neu, or SP-neu (Group X), or PHP-neu digested with proteolytic enzymes (Group XI)
is given. The DGalN-poisoned rats not receiving any subsequent injection (Group VII) or those injected with plasma prepared from nonhepatectomized rats (Group X) had mortality rates of 85 and 88%, respectively. The mortality rate was significantly reduced to 43% in rats of Group VIII treated with PHP-neu. The mortality rate was also reduced in the rats treated with PHP without neu (Group IX), 63%, and in the rats given PHP-neu subjected to tryptic digestion (Group XI), 67% (2/4 rats given PHP-neu treated with trypsin/ chymotrypsin and 4/5 given PHP-neu treated with protease died). The overall results in the latter two groups were intermediate between those of Group VIII and Groups VII and X. Gel filtration results. We obtained a consistent elution profile (Fig. 1) for PHP-neu extract with both Sephadex G- 100 and G-200 gel filtration. Peak I eluted with the column void volume. Peak II represents a broad increase in eluent absorption over a 30,000- to 90,000-Da molecular weight range. Peak III falls in a molecular weight range less than 10,000 Da. Although we do not show the data here (manuscript in preparation), when we injected pooled fractions from Peak II into poisoned rats, the survival rate was enhanced to a degree comparable to that observed in poisoned rats injected with unfractionated PHP-neu extract (Group VIII). DISCUSSION
This study demonstrates that the injection of either ( 1) syngeneic cell culture supematant,
TABLE 2 SURVIVAL OF D-GalN-POISONED RATS TREATED WITH VARIOUS
PLASMA PREPARATIONS 20 hr AITER POISONING’
Group
Donor rat treatment
Plasma treatment
VII VIII IX X XI
None Hepatectomy Hepatectomy No operation or sham operation Hepatectomy
None Neuraminidase No neuraminidase Neuraminidase Neuraminidase + Proteolytic digestion
No. survived/ No. treated 3/20 8/14 318 l/8 3/9
(15%) (57%) (37%) (12%) (33%)
’ Rats that died did so within 72 hr. P < 0.05 for Group VIII vs VII and X. P z- 0.1 for Group IX and XI vs VII, VIII. and X.
LA PLANTE,
BLANC, AND SUTHERLAND:
(2) cell-free cytosol of regenerating liver cells, or (3) neuraminidase-treated plasma from partially hepatectomized rats is effective treatment in reducing mortality induced by Dgalactosamine in rats. We deliberately avoid labeling any of the above treatments as hepatoregenerative, although we favor the hypothesis that the injected material stimulates or accelerates recovery of liver function. Conversely, injecting any of the isolates may merely provide or initiate endogenous production of some unknown factor(s) which prolongs life until the D-GalN-poisoned liver regenerates of its own accord. Support for the regenerative hypothesis comes from Goldberg et al. [5] in their assertion that PHP-neu is a liver-specific initiator of DNA synthesis and mitosis in normal rat liver. They also claim that the active factor is a neuraminidase induced p-galactose residue which binds to /3-galactose receptors on liver cells, thus causing proliferation. The results of our experiments, in a bioassay model, are consistent with their hypothesis, but we have not performed experiments that could prove or disprove the hypothesis. We agree that hepatectomy activates or stimulates assayable cellular and humoral growth factors. Even though Goldberg et al. [5] found neuraminidase treatment essential to show an effect of PHP in their model, it is not clear that neuraminidase treatment is necessary to improve survival in our liver poison model, since a slight effect on improving survival was seen even without neuraminidase treatment. Even though the mortality was significantly reduced only in the group receiving PHP-neu, we administered only one dose at one concentration in all groups, and a greater effect might be seen with an increase in dose or with multiple injections in both the PHP-neu group or the PHP without neu or PHP-neu proteolytic digestion groups. The validity of this conjecture remains to be tested experimentally. Factors that contribute to liver regeneration are well summarized by Golde et al. [6]. The discovery of these factors dates back to the classic cross circulation experiments performed by Fisher et al. [4], Lieberman [lo],
DGALACTOSAMINE
375
and Bucher et al. [2]. All demonstrated the existence of humoral factors in partially hepatectomized animals which stimulated DNA synthesis and mitosis in normal animals. These investigators isolated some of the cellular and humoral components elicited by partial hepatectomy, and demonstrated their effects on DNA synthesis and mitosis, but failed to implicate any one as the prime or sole initiator of liver regeneration in vivo. O’Keefe and co-workers [3, 131 exemplify this type of pursuit in their characterization and mechanistic studies of epidermal growth factor (EGF). EGF is one component of the complex signal which regulates liver regeneration. Similarly, our extracts may require the concerted action of other regulatory mechanisms in the recipient rats to produce increased DNA synthesis and mitosis. We need then, to purify and characterize these extracts with the aim of determining their similarity and of demonstrating a hepatoregenerative effect, or for that matter, a liver-specific effect of any kind, in vivo or in vitro. A reliable in vitro assay is needed; in vivo (bio)assays are complicated by the enormous number of factors which influence liver function and which are not easily excluded from impinging on the final outcome of an assay designed to study the effect of an isolated substance [7, 151. A necessary requirement for the production of factors which stimulate hepatic regeneration may be a liver deficiency state. In vitro hepatocyte culture may mimic the partial hepatectomized state in vivo, and if the active substances produced by the two techniques turned out to be the same it would indicate that the liver itself can be the source of regenerative stimuli. REFERENCES 1. Andrews, D. Estimation of the molecular weights of proteins by Sephadex gel-filtration. Biochem. .I 91: 222, 1964. 2. Bucher, N. L. R., Patel, U., and Cohen, S. Hormonal factors and liver growth. Advan. Enzyme Regulalion 16: 205,
1978.
3. Earp, H. S., and O’Keefe, E. J. Epidermal growth factor receptor number decreases during rat liver regeneration. J. Clin. Invest. 67: 1580, 1981.
376
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. 4, APRIL 1984
Fisher, B., Szuch, D., Levine, M., and Fisher, E. R. A portal blood factor as the humoral agent in liver regeneration. Science 171: 575, 1971. 5. Goldberg, M., Strecker, W., Feeny, D., and Ruhenstroth-Bauer, G. Evidence for and characterization of a liver cell proliferative factor from blood plasma of partially hepatectomized rats. Horm. Metab. Res. 12: 4.
94, 1980. 6.
12.
Golde, D. W., Herschman, H. R., Lusis, A. J., and Groopman, J. E. Growth factors. Ann. Intern. Med. 92: 650,
1980.
Greisler, H. P., Voorhees, A. B., and Price, J. B. The nonportal origin of the factors initiating hepatic regeneration. Surgery 86: 2 10, 1979. 8. Higgins, G. M., and Anderson, R. M. Experimental pathology of the liver of the white rat following partial surgical removal. Arch. Pathol. 12: 186, 1931. 9. LaPlante O’Neill, P. L., Baumgartner, D., Lewis, W. I., Zweber, B. A., and Sutherland, D. E. R. Cellfree supematant from hepatocyte cultures improves survival of rats with chemically induced acute liver failure. J. Surg. Res. 32: 347, 1982. 10. Lieberman, I. Studies on control of mammalian deoxyribonucleic acid synthesis. In R. Baserga (Ed.), 7.
11.
13.
14.
Biochemistry of Cell Division, Springfield, III.: Thomas, 1969. Makowka, L., Falk, R. E., Blandis, L. M., Langer, B., Falk, J. A., Rotstein, L. E., and Phillips, M. J. The use of liver cytosol fractions in the treatment of acute liver failure. Surgery 91(I): 170, 1982. Makowka, L., Rotstein, L. E., Falk, R. E., Falk, J. A., Zuk, R., Langer, B., Blendis, L. M., and Phillips, M. J. Allogeneic and xenogeneic hepatocyte transplantation. Transplant. Proc. 13: 855, 198 I. O’Keefe, E. J., Holler&r& M. D., and Cuatnecasas, P. Epidermal growth factor: Characteristics of specific binding in membranes from liver, placenta, and other target tissues. Arch. B&hem. Biophys. 164: 5 18, 1974. Sommers, B. G., Sutherland, D. E. R., Matas, A. J., Simmons, R. L., and Najarian, J. S. Hepatocellular transplantation for treatment of Dgalactosamine induced acute liver failure in rats. Transplant. Proc. 11: 578, 1979.
15. Statzl, T. E., Terblanche, J., Porter, K. A., Jones, A. F., Usui, S., and Mazzoni, G. Growth stimulating factor in regenerating canine liver. Luncet 1: 127, 1979.