Hemopoietic Colony Forming Cells in Regenerating Mouse Liver1 E S T H E R F . H A Y S , F R A N K C. FIRKIN,* Y O S H I K O KOGA AND D A N I E L M. H A Y S Laboratory of Nuclear Medicine and Radiation Biology, Department of Medi c i n e , University of California, Center for H e a l t h Sciences, Los Angeles, California, 90024, a n d t h e Department of Surgery, t h e Childrens Hospital of Los Angeles, University of S o u t h e r n California School of Medicine, Los Angeles, California, 90054
ABSTRACT
Cells from regenerating mouse liver removed L 2 5 days post 68% hepatic resection have been assayed for in vitro colony forming capacity in soft agar (CFU-C), proliferative capacity in liquid culture, and i n vivo spleen colony forming capacity (CFU-S). These studies demonstrated low concentrations of CFU-C and CFU-S in normal and sham-operated liver, with a n appreciable increase of both in regenerating liver, reaching maximum values i n tissue removed 5-7 days post hepatic resection. Colony formation in agar by regenerating liver cells occurred in the absence of exogenous colony stimulating factor. Separation of liver cells on the basis of adherent properties prior to culture indicated concentration of CFU-C in the nonadherent fraction, while cells producing colony stimulating factor were concentrated i n the adherent fraction. Foci of actively dividing cells of the macrophage and granulocyte series arose in liquid culture from preparations of sham operated and regenerating liver, although total cell formation was greater with regenerating liver. A small proportion of the colonies formed i n agar from regenerating liver consisted of cords of epithelioid cells, which resembled hepatocytes and differed from the macrophages or granulocytes found i n the majority of colonies, raising the possibility that regenerating hepatocytes form colonies in agar culture.
During the development of the mammalian embryo, hemopoietic stem cells migrate from blood islands of the yolk sac to the liver, which becomes the major site of embryonic blood cell formation. Bone marrow becomes the dominant site of hemopoiesis at about the time of birth, when hepatic hemopoiesis essentially ceases except under conditions of increased demand for blood formation (Moore and Metcalf, '70; Delmonte and Liebelt, '65). Although erythroid and granulocytic precursors disappear under normal conditions, a population of myeloid cells consisting of macrophages, or Kupffer cells, is maintained in adult liver and makes up about 30% of the total cell number. These cells have been shown to undergo DNA synthesis and mitosis during restoration of liver tissue following partial hepatectomy in the rat (Grisham, '62). Whether the macrophages constitute a truly self-sustaining population as suggested for hepatocytes, or are basically dependent on migration of precursors from bone marrow, J . CELL. PHYSIOL.,86: 21S22.0.
has not been determined (Kelly and Dobson, '71; Kinsky et al., '60). In this study the presence of hemopoietic stem cells in regenerating and normal mouse liver was evaluated by several methods. These techniques included a study of CFU-C in agar (Bradley and Metcalf, '66), CFU-S in spleens of lethally irradiated mice (Till and McCulloch, 'Sl), and cells in liquid culture (Sumner et al., '72). Another objective of the study was to determine whether hepatocytes under the stimulus of regeneration could continue to replicate sufficiently to form colonies in the culture systems employed. Received Oct. 14, '74. Accepted Feb. 6, '75. 1 Supported by Contract AT (04-1) GEN-12 between the Atomic Energy Commission and the University of California and by Public Health Service Grants CA 13666 from the National Cancer Institute, and AM 08879 from the National Institute of Arthritis, Metabolism and Digestive Diseases. 2 Frank C. Firkin was a recipient of the Bushell's traveling fellowship of the Royal Australian College of Physicians. Present address: Department of Medicine, St. Vincents Hospital, Fitzroy 3065, Victoria, Australia.
213
214
E. F. H A Y S , F . C . FIRKIN, Y. KOGA AND D. M. H A Y S
MATERIALS AND M E T H O D S
Mice. Two-month old female C57B16/J mice weighing 18-25 gm were employed in the study. They were supplied food and water ad libitum. Resection. Partial hepatectomy, performed under barbiturate anesthesia, resulted in the removal of about 68% of the liver tissue (Higgins and Anderson, '31). A sham hepatectomy, carried out under the same anesthesia, consisted of lifting liver lobes partially out of the abdomen (as in resection), then replacing the liver, and closing the abdomen. Preparation of cell suspensions. Cell suspensions from regenerating liver remnants, liver of sham operated animals, normal liver, and liver of 18-20 day embryos were prepared as follows: tissue was minced into 1-2 mm pieces, washed in Earle's balanced salt solution (EBSS). Pieces were then placed in 0.25% trypsin in EBSS supplemented with penicillin 100 unitslml and streptomycin (100 pglml). They were agitated with a magnetic stirrer in a 5 % COr incubator at 37" for ten minutes. The liver pieces were subjected to 10-minute trypsinization four times. The initial supernatant was discarded. Subsequent supernatants were pooled and centrifuged at 50 X g , after which the cell pellet was resuspended in culture medium. Pools of cells from 1 - 4 livers were used. Tissue from the regenerating liver remnants was always removed some distance from the line of resection in order to avoid damaged tissue infiltrated with inflammatory cells. Spleen cells were prepared by mincing the spleen in medium and bone marrow cells were prepared by expulsion of marrow from the femur with medium using a number 27 gauge needle. Peripheral blood leukocytes were obtained from heparinized axillary vein blood after sedimentation of erythrocytes with plasmagel. Spleen, bone marrow and peripheral blood leukocytes from normal animals and from animals after partial hepatectomy or sham operation were studied. All cell preparations were identically exposed to the trypsinization procedure. Preparations of adherent and non-adherent cell populations. Trypsinized cell suspensions were incubated in McCoy's 5a supplemented with 15% Fetal Calf Serum antibiotics (M5a) in 60 mm plastic petri dishes a t 37°C in a 5% COs atmosphere
for two hours. The supernatants were removed and incubated in fresh petri dishes for an additional hour. The cells in the final supernatant were called the "non-adherent" cell population. Adherent cells were treated for 15 minutes with Alsever's solution to detach them and were then pooled. Preparation of conditioned medium. To test the capacity of cells to release colony stimulating activity (CSA), 2 X 10" cells from adherent or non-adherent fractions of liver and bone marrow were incubated for seven days, and 0.1 ml of conditioned medium assayed in duplicate soft agar plates with 5 X 104 adult C57B1 bone marrow cells (method described below). In uitro colony assay. A modification of the method described by Bradley and Metcalf, '66 was used. The various liver cell preparations were cultured in M5a in 0.3% agar. The cell concentration was adjusted to 5 X 10: or 2.5 X 105 per milliliter and pipetted into 35 mm plastic petri dishes in 1 ml volumes. Dishes contained 40 pl of endotoxin mouse serum (EMS), normal mouse serum (NMS), or no additive. Both sera (EMS and NMS) contain CSA, a factor which stimulates in vitro colony formation of bone marrow cells (Stanley et al., '71). EMS was obtained from mice four hours after a n intravenous inoculation of 5 kg of lipopolysaccharide, E. coli. The agar was allowed to solidify for 20 minutes at room temperature and the dishes were placed in a moist 5% COz atmosphere at 37°C for seven days. Colonies were counted with a dissecting microscope. Only colonies containing more than 50 cells were included. The colony number is reported as a mean of the value obtained by the counting of 2-9 replicate plates. In viwo liquid culture. Cell preparations were maintained in liquid culture by the technique described for murine bone marrow (Sumner et al., '72) using vessels obtained from Bio Research Glass, Inc., Vineland, New Jersey. Cells suspended in 1 ml M5a were placed in a chamber separated by dialysis membrane from a reservoir containing approximately 20 ml of medium. Culture conditions were identical to those employed for agar cultures. At the completion of the culture period, the cells had settled on the dialysis membrane and were examined in situ on the membrane, or in
215
COLONY FORMING CELLS I N REGENERATING LIVER
cytocentrifuge preparations after they had been dislodged by gentle suction. Phagocytic activity of cells was assessed by uptake of iron particles (Technicon lymphocyte separating reagent), which was added to the culture three hours before harvesting in amounts of approximately 0.05 ml per 10'; cells. In uiuo colony assay. The method described by Till and McCulloch, '61, was employed. Trypsinized liver or bone marrow cells were injected intravenously into mice irradiated with 800 rads f W o several hours previously. The animals were housed in sterile filter top cages, with autoclaved shavings and water which were changed daily. The animals were killed after ten days, the spleens removed and placed in Bouin's fluid, and colonies counted 24 hours later. Morphologic studies. Cytocentrifuge preparations of cell suspensions were stained with Wright's stain, and in certain instances with histochemical stains for CYnaphthyl acetate esterase and acid phosphatase. Aspirated agar colonies were stained with aceto-orecin. Liver, spleen and agar colonies were also fixed in Bouin's fluid and stained with hematoxylin and eosin after sectioning. Tissue samples for microscopic examination were taken from the periphery of the right lobe of the unresected liver remnant or from the same area of the whole liver. RESULTS
Liver tissue morphology. Vacuolization of hepatocytes was prominent a t two days post resection and diminished after this time. Mitotic figures were readily detected in the hepatocytes 2-10 days post resection, but were not observed in normal or sham operated liver. Numbers of macrophages were similar in all three situations. A few small (30 p diameter) foci of granulocytes and mononuclear cells were encountered in the liver parenchyma in 12 of 33 regenerating livers, 12 of 14 sham operated livers and 1 of 3 normal livers. Larger nodules and necrotic foci approximately 150 p in diameter were found in four of the regenerating liver specimens. On one occasion a megakaryocyte was noted in 10-day regenerating liver, and a small focus of erythroblasts was noted in a 6-, and 10-day regenerating liver. Fetal liver contained abundant erythroid and granulocytic precursors.
Identity of cells in trypsinized liver preparations. Intact hepatocytes made up only 5 to 15% of cells in preparations of liver from normal, sham operated and partially hepatectomized mice. The remaining cells consisted of 10-15% mature granulocytes, 25-50 % monocyte-macrophage series, and 30-50% mononuclear cells including lymphocytes. Monocytes and macrophages predominated in the adherent fraction, but amounted to only about 10% of the nonadherent fraction. Histochemical determinations of a-naphthyl acetate esterase and acid phosphatase revealed significant activity in hepatocytes, and thus these studies did not provide a n absolute distinguishing feature between hepatocytes and cells of the monocyte-macrophage series. Differentiation between these cell types were assisted by the failure of hepatocvtes to phagocytize iron particles, in contrast to monocytes and macrophages. In uitro colony forming cells in regenerating liver. Colonies developed in the agar culture system from cells prepared from regenerating liver three days or more after resection (table 1).Occasional colonies develTABLE 1
Cells f r o m regenerating liver tissue cultured i n soft agur Additive Interval post resection (days)
None
2 3 4 5 6 7 10 14 25
0,i
6 0 0 0
EMS
NMS
0,o 0,o 15,18,2 1,6,0 4,23,36 8,28,39 16,6,2 40, 10, 8 65,26 60,15 30,40 18,24 86 46 18 5 9 1
Interval post sham operation (days)
3 4 14 25
N o operation
1
1
4,4,2 0
5 6,15,1
Results are expressed in colonies per 10s liver cells plated. Tissue was removed at different intervals post hepatic resection and colony formation compared with that of cells from normal and sham operated liver. Each value is the mean number of colonieslplate from 2-9 identical cultures. N M S , normal mouse serum; EMS, endotoxin mouse serum.
216
E. F. HAYS, F. C. FIRKIN, Y. KOGA AND D. M. HAYS
oped from the liver of normal and sham operated mice, but the numbers were low in comparison with the levels in regenerating liver removed from 3-7 days after resection. Colonies from regenerating liver were larger than those of sham operated and normal livers. Colony formation was more frequent in the presence of EMS and NMS, although appreciable formation occurred in their absence when cells removed 3-7 days post resection were cultured. Some variation occurred between separate experiments, but results in individual experiments were consistent in up to nine identical cultures. There was no correlation between the microscopic appearance of liver sections and colony forming capacity of the cell suspension prepared from that tissue. Morphology of the colony cells. Colonies in agar which developed from all liver preparations were shown by aceto-orecin staining to be predominantly made up of cells of the macrophage or granulocyte series, or a mixture of both. A similar conclusion was reached following examination of thin sections of the colonies. However, in three instances, colonies contained cells which were epithelioid and resembled hepatocytes (5
days post-resection minus EMS and with EMS, and 4 days post-resection plus EMS). Thus a small number of colonies appeared to be made up of cells with morphologic features of hepatocytes. Cells of this type were not encountered in colonies derived from sham operated liver or bone marrow. Effects of fractionation on basis of adhesiveness. Separation of the trypsinized liver cell preparations on the basis of adhesiveness influenced colony forming capacity. CFU-C were concentrated in the nonadherent fraction (table 2). However, removal of adherent cells variably diminished the proportion of colonies formed from regenerating liver in the absence of NMS and EMS. This suggested that adherent cells may generate CSA in the agar cultures. Medium conditioned by adherent cells stimulated formation of 69 marrow colonies, demonstrating CSA production from these cells. Conditioned medium from non-adherent cells stimulated the formation of 39 colonies, also indicating the presence of CSA producing cells in this population, which contained some monocytes and macrophages. Marrow cells did not produce colonies in the absence of CSA, and condi-
TABLE 2
Colony f o r m a t i o n b y “ndherent” a n d “non-ctdherent” c e l l s f r o m regenerating liver, s h a m operuted liver arid bone Regenerating liver; interval post resection (days)
3 4
Adherent cells No additive
Non-adherent cells No
NMS
EMS
Colonies p e r lo5 Cells ctiltured 0 0 0 2 1 0 1 3 3
5
1
6
8
6 7 10 14 25
0
2
0 0
0 0
6 4 1
0
0
0
additive
NMS
0 2 4 9 2 20 4 0 14 5 1 1
1 10 13 58
26 48 10
EMS
6 44 71 65 21 104 112 51 102 47 40 5
Sham operated liver; interval post resection (days)
4 14 25
Bone marrow
0
17
NMS, normal mouse serum; EMS, endotoxin mouse serum
0
0
0 2
4 0 13
0
530
217
COLONY FORMING CELLS IN REGENERATING LIVER TABLE 3
Colony f o r m i n g potential of cells f r o m f e t a l liver, s p l e e n , m u r r o w , and peripheral blood l e u k o c y t e s c u l t u r e d i n soft agar Results, colonies per 105 cells cultured Tissue cultured
Fetal liver Spleen Marrow Marrow Marrow Peripheral blood Peripheral blood Peripheral blood Peripheral blood Peripheral blood
leukocytes leukocytes leukocytes leukocytes leukocytes
No additive
State of liver
Non-regenerating Non-regenerating Regenerating-5d Sham procedure-5d Non-regenerating RegeneratingAd Regenerating-5d Sham procedure4d Sham procedure-5d
13 1 0
NMS
EMS
29 2 148
41 3 350 422 490 3 1 10 2 2
0 0 0 0 0 0 0
0
N M S , normal mouse serum; EMS, endotoxin mouse serum
tioned medium from adherent bone marrow cells (macrophages) did not demonstrate CSA activity. Colony forming potential of fetal liver and non-hepatic tissues in normal and hepatectomized mice. Fetal liver was studied in order to compare the number of CFUC in a form of liver tissue with obvious microscopic hemopoiesis; with the number formed by regenerating liver tissue in which hemopoietic elements are rarely observed. As shown in table 3 , fetal liver formed colonies in the absence of CSA, colony numbers were increased with added CSA, and the overall colony number was similar to that found in 4-7-day regenerating liver. Table 3 also compares colony formation in marrow, peripheral blood, and spleen in normal animals; as well as marrow and peripheral blood leukocyte colony formation in animals after partial and sham hepatectomy. Colonies were not formed in the absence of CSA with these preparations. Hepatectomy resulted in some increase in CFU-C in the peripheral blood leukocytes at five days, but not at four days. No increase in marrow CFU-C was noted in hepatectomized animals. Serum from mice three days after hepatectomy did not demonstrate increased CSA. Cell behavior in liquid culture. Cell preparations from 4-day sham operated and 4-day regenerating liver yielded approximately equivalent numbers of viable cells after six days of culture in the absence of EMS (table 4). Increased numbers of cells developed in the presence of EMS only from
TABLE 4
Regenerating liver cells c u l t u r e d i n liquid m e d i a Nature of cultured cells
Number cultured X 105
Cells present after 6 days of culture X 1 0 5 No
EMS
EMS
1.0
0.93 0.84
1.15 1.75
2.5
1.68 2.60
4.90 4.30
4-day post sham operation
2.5
0.95 1.27
1.13 1.03
4-day post resection non-adherent fraction
1.0
0.24 0.11
1.45 0.96
4-day post resection
EMS, endotoxin mouse serum.
regenerating liver preparations. Cell yields were low when non-adherent cells from regenerating liver were cultured in the absence of EMS, but the marked decline in numbers was prevented by addition of EMS. The morphologic picture at the time of harvest was similar whether the cultured cells were from sham operated or regenerating liver, and consisted of approximately 35% cells of the granulocytic series and 55% cells of the monocyte-macrophage series. The remaining cells consisted of undifferentiated immature cells, cells in mitosis, and lymphocytes. Cells arising from regenerating liver preparations in the presence of EMS exhibited the same morphologic pattern as in its absence. Iron particles added to the cultures were preferentially
218
E. F . HAYS. F . C. FIRKIN. Y . KOGA AND D. M
ingested by cells with characteristic features of mature macrophages. The presence of immature granulocytes and macrophages represented the development of considerable numbers of cells which were not seen in the cell suspensions prior to culture. Examination of cells fixed in situ on the membrane revealed tightly packed focal aggregates of granulocytes and macrophages, with immature forms and cells in mitosis frequently occupying a central position. The areas between such foci were occupied by mature macrophages and granulocytes. Splenic colony forming potential of regenerating liver. Intravenous injection of trypsinized cells from regenerating liver resulted in splenic colony formation in irradiated mice (table 5). Fractionation of the liver cell suspensions on the basis of adhesiveness concentrated the CFU-S in the nonadherent population. The non-adherent fraction from the liver of sham operated mice contained some CFU-S, although considerably smaller numbers of colonies were produced. The concentration of CFU-S was greater in femoral bone marrow of non-resected animals than in any of the regenerating liver preparations. hlicroscopy of the splenic colonies revealed a n identical morphologic pattern when the source of CFU-S was bone marrow, regenerating liver, or sham operated TABLE 5
S p l e n i c colony f o r m i n g potenhcil of regenertrti7ig l i v e r cells 1 ~~
~
Source of cells
Number iwected
'4-day post resection non-adherent 5-day post resection non-adherent Liver adherent 4-day sham operated non-adherent 5-day sham operated non-adherent
Normal
3 . 4 x 106
Spleen colonies at 10 days
106
22
106
106
25 8
106
1
106
6
105 -
0
33
0
Trypsinized cell preparations were injected i n t o the tail vein of mice irradiated with 800 rads two hours previously. 1
HAYS
liver. Colonies consisted of granulocytes, megakaryocytes, erythroid precursors, or varying combinations of these cell lines. 111 SC USSION
The present study has demonstrated that hemopoietic stein cells exist in low concentrations in adult mouse liver and increase during regeneration of the organ following partial hepatectomy. CFU-C and CFU-S have been previously described in newborn mouse liver, (Metcalf and Moore, .71), and in adult mouse liver (Silini et al., '67).Data from the present experiment reveal that the concentration of CFU-C and CFU-S is markedly increased three days after twothirds hepatic resection reaching maximum values 5-7 days after resection. Hepatocytes are known to commence active mitosis on the second, and reach maximal activity on the third day in vivo. Mitosis in macrophages develops more slowly and reaches peak activity on about the fifth day post resection (Yokoyama et al., '53; Grisham, '62), which approximates the time when CFU-C achieve maximum concentrations in the liver In this study. It is possible that CFU-S and CFU-C in regenerating liver serve as progenitors of this increasing macrophage population, as macrophages are among the progeny of CFU-C. It is suggested that the hemopoietic stem cells represent components of liver which are essentially dormant in vivo until stimulated to proliferate, either by removal from the micro-environment, or by changes associated with regeneration. The present study does not, however, exclude that the CFU-S aiid CFUC in question migrate from bone marrow. It has been shown that hepatic macrophages can be reconstituted from injected bone marrow cells (Boak et al., '68). However, the absence of a remarkable increase in numbers of CFU-C in marrow and peripheral blood in animals after partial hepatectomy suggests that we are not dealing with a n increase in CFU-C production and circulation, but an intra hepatic response to the insult of resection. Both CFU-C and CFU-S in liver cell preparations were concentrated in the nonadherent fraction. Adherent cells from liver produced CSA and those from bone marrow did not. Since macrophages, which are a major component of the adherent population, have been shown in other systems to
COLONY FORMING CELLS IN REGENERATING LIVER
produce CSA (Chervenick and LoBuglio, '72), this finding may indicate a functional difference of the macrophages derived from liver and bone marrow. Unfractionated preparations of 18-day fetal liver contained concentrations of CFUC similar to adult regenerating liver, and also supported spontaneous colony formation in agar, which was increased by addition of EMS. However, the CFU-C population in fetal liver occurs in the setting of obvious hemopoietic activity (Metcalf and Moore, '71), which is in sharp contrast to the low numbers of hemopoietic elements seen in microscopic sections during regeneration in the adult organ. Conditions of cell isolation from the liver tissue preferentially damages hepatocytes and resulted in low numbers of these cells in our preparations. A small number of colonies arising in agar from regenerating liver consisted of cords of apposed epithelioid cells which resembled hepatocytes. Cells of this type were not encountered in colonies originating from sham operated liver or bone marrow. This raises the possibility that an additional type of in vitro colony forming cell related to hepatic rather than hemopoietic tissue occurs in regenerating liver. No evidence of these cells was obtained in liquid culture or in colonies arising in spleen, and further investigation is required to establish their identity. ACKNOWLEDGMENTS
The authors wish to thank Dorothy Haskett, Melinda Newel1 and Wayne Shephard for their valuable technical assistance and Doctor Harry B. Neustein of the Department of Pathology, Childrens Hospital of Los Angeles, for his review of microscopic sections. LITERATURE CITED Boak, J. L., G. H. Christie, W. H. Ford a n d J . G . Howard 1968 Pathways in the development
219
of liver macrophages: alternative precursors contained in populations of lymphocytes a n d bone marrow-cells. Proc. Roy. SOC.B., 169: 307-327. Bradley, T. R., a n d D. Metcalf 1966 The growth of mouse bone marrow cells in nitro. Aust. J . Exp. Biol. Med. Sci., 44: 287-300. Chervenick, P. A , , and A. F. LoBuglio 1972 Hum a n blood monocytes, stimulators of granulocyte and mononuclear colony formation in nitro. Science, 1 78 : 64-1 78. Delmonte, L., a n d R. A. Liebelt 1965 Granulocytosis promoting extract of mouse tumor tissue. Partial purification. Science, 148: 521-523. Grisham, J. W. 1962 A morphologic study of deoxyribonucleic acid synthesis a n d cell proliferation i n regenerating rat liver, autoradiography with thymidine-HJ. Cancer Res., 22: 8 4 L 8 4 9 . Higgins, G. M., a n d R. M. Anderson 1931 Experimental pathology of the liver; restoration of the liver of the white rat following partial surgical removal. Arch. Pathol., 1 2 : 18CL-202. Kelly, L. S . , a n d E. L. Dobson 1971 Evidence concerning the origin of liver macrophages. Brit. J . Exp. Pathol., 52: 88-99. Kinsky, R. G., G. H. Christie, J . Elson and J. G. Howard 1969 Extra-hepatic derivation of Kupffer cells during oestrogenic stimulation of parabiosed mice. Brit. J. Exp. Pathol., 50: 438-447. Metcalf, D., and M. A . S. Moore 1971 In: Frontiers of Biology. Vol. 24. Haemopoietic Cells. A. Neuberger a n d E. L. Tatum, eds. North-Holland Amsterdam-London, pp. 71, 199. Moore, M. A. S . , a n d D. Metcalf 1970 Ontogeny of the haemopoietic system: yolk sac origin of in vivo a n d i n nitro colony forming cells in the developing mouse embryo. Brit. J . Haematol., 18: 279-296. Silini, G. L., V. Pozzi and S. Pons 1967 Studies o n the haemopoietic cells of mouse fetal liver. J . Embryol. Exp. Morphol., 17: 302-318. Stanley, E. R., T. R. Bradley and M. A. Sumner 1971 Properties of the mouse embryo condiditioned medium factor(s) stimulating colony formation by mouse bone marrow cells grown in vitro. J . Cell. Physiol., 78: 301-318. Sumner, M. A , , T. R. Bradley, G. S. Hodgson, M. J . Cline, P. A. Fry a n d L. Sutherland 1972 The growth of bone marrow cells i n liquid culture. Brit. J . Haematol., 2 3 : 221-234. Till, J. E., a n d E. A . McCulloch 1961 Direct measurement of the radiation sensitivity of normal mouse bone marrow cells. Rad. Res., 14: 212222. Yokoyama, H. O., M. E. Wilson, K. K. Tsuboi a n d R. E. Stowell 1953 Regeneration ofmouse liver after partial hepatectomy. Cancer Res., 1 3 : 8 0 85.
ABSTRACT
Cells from regenerating mouse liver removed L 2 5 days post 68% hepatic resection have been assayed for in vitro colony forming capacity in soft agar (CFU-C), proliferative capacity in liquid culture, and i n vivo spleen colony forming capacity (CFU-S). These studies demonstrated low concentrations of CFU-C and CFU-S in normal and sham-operated liver, with a n appreciable increase of both in regenerating liver, reaching maximum values i n tissue removed 5-7 days post hepatic resection. Colony formation in agar by regenerating liver cells occurred in the absence of exogenous colony stimulating factor. Separation of liver cells on the basis of adherent properties prior to culture indicated concentration of CFU-C in the nonadherent fraction, while cells producing colony stimulating factor were concentrated i n the adherent fraction. Foci of actively dividing cells of the macrophage and granulocyte series arose in liquid culture from preparations of sham operated and regenerating liver, although total cell formation was greater with regenerating liver. A small proportion of the colonies formed i n agar from regenerating liver consisted of cords of epithelioid cells, which resembled hepatocytes and differed from the macrophages or granulocytes found i n the majority of colonies, raising the possibility that regenerating hepatocytes form colonies in agar culture.
During the development of the mammalian embryo, hemopoietic stem cells migrate from blood islands of the yolk sac to the liver, which becomes the major site of embryonic blood cell formation. Bone marrow becomes the dominant site of hemopoiesis at about the time of birth, when hepatic hemopoiesis essentially ceases except under conditions of increased demand for blood formation (Moore and Metcalf, '70; Delmonte and Liebelt, '65). Although erythroid and granulocytic precursors disappear under normal conditions, a population of myeloid cells consisting of macrophages, or Kupffer cells, is maintained in adult liver and makes up about 30% of the total cell number. These cells have been shown to undergo DNA synthesis and mitosis during restoration of liver tissue following partial hepatectomy in the rat (Grisham, '62). Whether the macrophages constitute a truly self-sustaining population as suggested for hepatocytes, or are basically dependent on migration of precursors from bone marrow, J . CELL. PHYSIOL.,86: 21S22.0.
has not been determined (Kelly and Dobson, '71; Kinsky et al., '60). In this study the presence of hemopoietic stem cells in regenerating and normal mouse liver was evaluated by several methods. These techniques included a study of CFU-C in agar (Bradley and Metcalf, '66), CFU-S in spleens of lethally irradiated mice (Till and McCulloch, 'Sl), and cells in liquid culture (Sumner et al., '72). Another objective of the study was to determine whether hepatocytes under the stimulus of regeneration could continue to replicate sufficiently to form colonies in the culture systems employed. Received Oct. 14, '74. Accepted Feb. 6, '75. 1 Supported by Contract AT (04-1) GEN-12 between the Atomic Energy Commission and the University of California and by Public Health Service Grants CA 13666 from the National Cancer Institute, and AM 08879 from the National Institute of Arthritis, Metabolism and Digestive Diseases. 2 Frank C. Firkin was a recipient of the Bushell's traveling fellowship of the Royal Australian College of Physicians. Present address: Department of Medicine, St. Vincents Hospital, Fitzroy 3065, Victoria, Australia.
213
214
E. F. H A Y S , F . C . FIRKIN, Y. KOGA AND D. M. H A Y S
MATERIALS AND M E T H O D S
Mice. Two-month old female C57B16/J mice weighing 18-25 gm were employed in the study. They were supplied food and water ad libitum. Resection. Partial hepatectomy, performed under barbiturate anesthesia, resulted in the removal of about 68% of the liver tissue (Higgins and Anderson, '31). A sham hepatectomy, carried out under the same anesthesia, consisted of lifting liver lobes partially out of the abdomen (as in resection), then replacing the liver, and closing the abdomen. Preparation of cell suspensions. Cell suspensions from regenerating liver remnants, liver of sham operated animals, normal liver, and liver of 18-20 day embryos were prepared as follows: tissue was minced into 1-2 mm pieces, washed in Earle's balanced salt solution (EBSS). Pieces were then placed in 0.25% trypsin in EBSS supplemented with penicillin 100 unitslml and streptomycin (100 pglml). They were agitated with a magnetic stirrer in a 5 % COr incubator at 37" for ten minutes. The liver pieces were subjected to 10-minute trypsinization four times. The initial supernatant was discarded. Subsequent supernatants were pooled and centrifuged at 50 X g , after which the cell pellet was resuspended in culture medium. Pools of cells from 1 - 4 livers were used. Tissue from the regenerating liver remnants was always removed some distance from the line of resection in order to avoid damaged tissue infiltrated with inflammatory cells. Spleen cells were prepared by mincing the spleen in medium and bone marrow cells were prepared by expulsion of marrow from the femur with medium using a number 27 gauge needle. Peripheral blood leukocytes were obtained from heparinized axillary vein blood after sedimentation of erythrocytes with plasmagel. Spleen, bone marrow and peripheral blood leukocytes from normal animals and from animals after partial hepatectomy or sham operation were studied. All cell preparations were identically exposed to the trypsinization procedure. Preparations of adherent and non-adherent cell populations. Trypsinized cell suspensions were incubated in McCoy's 5a supplemented with 15% Fetal Calf Serum antibiotics (M5a) in 60 mm plastic petri dishes a t 37°C in a 5% COs atmosphere
for two hours. The supernatants were removed and incubated in fresh petri dishes for an additional hour. The cells in the final supernatant were called the "non-adherent" cell population. Adherent cells were treated for 15 minutes with Alsever's solution to detach them and were then pooled. Preparation of conditioned medium. To test the capacity of cells to release colony stimulating activity (CSA), 2 X 10" cells from adherent or non-adherent fractions of liver and bone marrow were incubated for seven days, and 0.1 ml of conditioned medium assayed in duplicate soft agar plates with 5 X 104 adult C57B1 bone marrow cells (method described below). In uitro colony assay. A modification of the method described by Bradley and Metcalf, '66 was used. The various liver cell preparations were cultured in M5a in 0.3% agar. The cell concentration was adjusted to 5 X 10: or 2.5 X 105 per milliliter and pipetted into 35 mm plastic petri dishes in 1 ml volumes. Dishes contained 40 pl of endotoxin mouse serum (EMS), normal mouse serum (NMS), or no additive. Both sera (EMS and NMS) contain CSA, a factor which stimulates in vitro colony formation of bone marrow cells (Stanley et al., '71). EMS was obtained from mice four hours after a n intravenous inoculation of 5 kg of lipopolysaccharide, E. coli. The agar was allowed to solidify for 20 minutes at room temperature and the dishes were placed in a moist 5% COz atmosphere at 37°C for seven days. Colonies were counted with a dissecting microscope. Only colonies containing more than 50 cells were included. The colony number is reported as a mean of the value obtained by the counting of 2-9 replicate plates. In viwo liquid culture. Cell preparations were maintained in liquid culture by the technique described for murine bone marrow (Sumner et al., '72) using vessels obtained from Bio Research Glass, Inc., Vineland, New Jersey. Cells suspended in 1 ml M5a were placed in a chamber separated by dialysis membrane from a reservoir containing approximately 20 ml of medium. Culture conditions were identical to those employed for agar cultures. At the completion of the culture period, the cells had settled on the dialysis membrane and were examined in situ on the membrane, or in
215
COLONY FORMING CELLS I N REGENERATING LIVER
cytocentrifuge preparations after they had been dislodged by gentle suction. Phagocytic activity of cells was assessed by uptake of iron particles (Technicon lymphocyte separating reagent), which was added to the culture three hours before harvesting in amounts of approximately 0.05 ml per 10'; cells. In uiuo colony assay. The method described by Till and McCulloch, '61, was employed. Trypsinized liver or bone marrow cells were injected intravenously into mice irradiated with 800 rads f W o several hours previously. The animals were housed in sterile filter top cages, with autoclaved shavings and water which were changed daily. The animals were killed after ten days, the spleens removed and placed in Bouin's fluid, and colonies counted 24 hours later. Morphologic studies. Cytocentrifuge preparations of cell suspensions were stained with Wright's stain, and in certain instances with histochemical stains for CYnaphthyl acetate esterase and acid phosphatase. Aspirated agar colonies were stained with aceto-orecin. Liver, spleen and agar colonies were also fixed in Bouin's fluid and stained with hematoxylin and eosin after sectioning. Tissue samples for microscopic examination were taken from the periphery of the right lobe of the unresected liver remnant or from the same area of the whole liver. RESULTS
Liver tissue morphology. Vacuolization of hepatocytes was prominent a t two days post resection and diminished after this time. Mitotic figures were readily detected in the hepatocytes 2-10 days post resection, but were not observed in normal or sham operated liver. Numbers of macrophages were similar in all three situations. A few small (30 p diameter) foci of granulocytes and mononuclear cells were encountered in the liver parenchyma in 12 of 33 regenerating livers, 12 of 14 sham operated livers and 1 of 3 normal livers. Larger nodules and necrotic foci approximately 150 p in diameter were found in four of the regenerating liver specimens. On one occasion a megakaryocyte was noted in 10-day regenerating liver, and a small focus of erythroblasts was noted in a 6-, and 10-day regenerating liver. Fetal liver contained abundant erythroid and granulocytic precursors.
Identity of cells in trypsinized liver preparations. Intact hepatocytes made up only 5 to 15% of cells in preparations of liver from normal, sham operated and partially hepatectomized mice. The remaining cells consisted of 10-15% mature granulocytes, 25-50 % monocyte-macrophage series, and 30-50% mononuclear cells including lymphocytes. Monocytes and macrophages predominated in the adherent fraction, but amounted to only about 10% of the nonadherent fraction. Histochemical determinations of a-naphthyl acetate esterase and acid phosphatase revealed significant activity in hepatocytes, and thus these studies did not provide a n absolute distinguishing feature between hepatocytes and cells of the monocyte-macrophage series. Differentiation between these cell types were assisted by the failure of hepatocvtes to phagocytize iron particles, in contrast to monocytes and macrophages. In uitro colony forming cells in regenerating liver. Colonies developed in the agar culture system from cells prepared from regenerating liver three days or more after resection (table 1).Occasional colonies develTABLE 1
Cells f r o m regenerating liver tissue cultured i n soft agur Additive Interval post resection (days)
None
2 3 4 5 6 7 10 14 25
0,i
6 0 0 0
EMS
NMS
0,o 0,o 15,18,2 1,6,0 4,23,36 8,28,39 16,6,2 40, 10, 8 65,26 60,15 30,40 18,24 86 46 18 5 9 1
Interval post sham operation (days)
3 4 14 25
N o operation
1
1
4,4,2 0
5 6,15,1
Results are expressed in colonies per 10s liver cells plated. Tissue was removed at different intervals post hepatic resection and colony formation compared with that of cells from normal and sham operated liver. Each value is the mean number of colonieslplate from 2-9 identical cultures. N M S , normal mouse serum; EMS, endotoxin mouse serum.
216
E. F. HAYS, F. C. FIRKIN, Y. KOGA AND D. M. HAYS
oped from the liver of normal and sham operated mice, but the numbers were low in comparison with the levels in regenerating liver removed from 3-7 days after resection. Colonies from regenerating liver were larger than those of sham operated and normal livers. Colony formation was more frequent in the presence of EMS and NMS, although appreciable formation occurred in their absence when cells removed 3-7 days post resection were cultured. Some variation occurred between separate experiments, but results in individual experiments were consistent in up to nine identical cultures. There was no correlation between the microscopic appearance of liver sections and colony forming capacity of the cell suspension prepared from that tissue. Morphology of the colony cells. Colonies in agar which developed from all liver preparations were shown by aceto-orecin staining to be predominantly made up of cells of the macrophage or granulocyte series, or a mixture of both. A similar conclusion was reached following examination of thin sections of the colonies. However, in three instances, colonies contained cells which were epithelioid and resembled hepatocytes (5
days post-resection minus EMS and with EMS, and 4 days post-resection plus EMS). Thus a small number of colonies appeared to be made up of cells with morphologic features of hepatocytes. Cells of this type were not encountered in colonies derived from sham operated liver or bone marrow. Effects of fractionation on basis of adhesiveness. Separation of the trypsinized liver cell preparations on the basis of adhesiveness influenced colony forming capacity. CFU-C were concentrated in the nonadherent fraction (table 2). However, removal of adherent cells variably diminished the proportion of colonies formed from regenerating liver in the absence of NMS and EMS. This suggested that adherent cells may generate CSA in the agar cultures. Medium conditioned by adherent cells stimulated formation of 69 marrow colonies, demonstrating CSA production from these cells. Conditioned medium from non-adherent cells stimulated the formation of 39 colonies, also indicating the presence of CSA producing cells in this population, which contained some monocytes and macrophages. Marrow cells did not produce colonies in the absence of CSA, and condi-
TABLE 2
Colony f o r m a t i o n b y “ndherent” a n d “non-ctdherent” c e l l s f r o m regenerating liver, s h a m operuted liver arid bone Regenerating liver; interval post resection (days)
3 4
Adherent cells No additive
Non-adherent cells No
NMS
EMS
Colonies p e r lo5 Cells ctiltured 0 0 0 2 1 0 1 3 3
5
1
6
8
6 7 10 14 25
0
2
0 0
0 0
6 4 1
0
0
0
additive
NMS
0 2 4 9 2 20 4 0 14 5 1 1
1 10 13 58
26 48 10
EMS
6 44 71 65 21 104 112 51 102 47 40 5
Sham operated liver; interval post resection (days)
4 14 25
Bone marrow
0
17
NMS, normal mouse serum; EMS, endotoxin mouse serum
0
0
0 2
4 0 13
0
530
217
COLONY FORMING CELLS IN REGENERATING LIVER TABLE 3
Colony f o r m i n g potential of cells f r o m f e t a l liver, s p l e e n , m u r r o w , and peripheral blood l e u k o c y t e s c u l t u r e d i n soft agar Results, colonies per 105 cells cultured Tissue cultured
Fetal liver Spleen Marrow Marrow Marrow Peripheral blood Peripheral blood Peripheral blood Peripheral blood Peripheral blood
leukocytes leukocytes leukocytes leukocytes leukocytes
No additive
State of liver
Non-regenerating Non-regenerating Regenerating-5d Sham procedure-5d Non-regenerating RegeneratingAd Regenerating-5d Sham procedure4d Sham procedure-5d
13 1 0
NMS
EMS
29 2 148
41 3 350 422 490 3 1 10 2 2
0 0 0 0 0 0 0
0
N M S , normal mouse serum; EMS, endotoxin mouse serum
tioned medium from adherent bone marrow cells (macrophages) did not demonstrate CSA activity. Colony forming potential of fetal liver and non-hepatic tissues in normal and hepatectomized mice. Fetal liver was studied in order to compare the number of CFUC in a form of liver tissue with obvious microscopic hemopoiesis; with the number formed by regenerating liver tissue in which hemopoietic elements are rarely observed. As shown in table 3 , fetal liver formed colonies in the absence of CSA, colony numbers were increased with added CSA, and the overall colony number was similar to that found in 4-7-day regenerating liver. Table 3 also compares colony formation in marrow, peripheral blood, and spleen in normal animals; as well as marrow and peripheral blood leukocyte colony formation in animals after partial and sham hepatectomy. Colonies were not formed in the absence of CSA with these preparations. Hepatectomy resulted in some increase in CFU-C in the peripheral blood leukocytes at five days, but not at four days. No increase in marrow CFU-C was noted in hepatectomized animals. Serum from mice three days after hepatectomy did not demonstrate increased CSA. Cell behavior in liquid culture. Cell preparations from 4-day sham operated and 4-day regenerating liver yielded approximately equivalent numbers of viable cells after six days of culture in the absence of EMS (table 4). Increased numbers of cells developed in the presence of EMS only from
TABLE 4
Regenerating liver cells c u l t u r e d i n liquid m e d i a Nature of cultured cells
Number cultured X 105
Cells present after 6 days of culture X 1 0 5 No
EMS
EMS
1.0
0.93 0.84
1.15 1.75
2.5
1.68 2.60
4.90 4.30
4-day post sham operation
2.5
0.95 1.27
1.13 1.03
4-day post resection non-adherent fraction
1.0
0.24 0.11
1.45 0.96
4-day post resection
EMS, endotoxin mouse serum.
regenerating liver preparations. Cell yields were low when non-adherent cells from regenerating liver were cultured in the absence of EMS, but the marked decline in numbers was prevented by addition of EMS. The morphologic picture at the time of harvest was similar whether the cultured cells were from sham operated or regenerating liver, and consisted of approximately 35% cells of the granulocytic series and 55% cells of the monocyte-macrophage series. The remaining cells consisted of undifferentiated immature cells, cells in mitosis, and lymphocytes. Cells arising from regenerating liver preparations in the presence of EMS exhibited the same morphologic pattern as in its absence. Iron particles added to the cultures were preferentially
218
E. F . HAYS. F . C. FIRKIN. Y . KOGA AND D. M
ingested by cells with characteristic features of mature macrophages. The presence of immature granulocytes and macrophages represented the development of considerable numbers of cells which were not seen in the cell suspensions prior to culture. Examination of cells fixed in situ on the membrane revealed tightly packed focal aggregates of granulocytes and macrophages, with immature forms and cells in mitosis frequently occupying a central position. The areas between such foci were occupied by mature macrophages and granulocytes. Splenic colony forming potential of regenerating liver. Intravenous injection of trypsinized cells from regenerating liver resulted in splenic colony formation in irradiated mice (table 5). Fractionation of the liver cell suspensions on the basis of adhesiveness concentrated the CFU-S in the nonadherent population. The non-adherent fraction from the liver of sham operated mice contained some CFU-S, although considerably smaller numbers of colonies were produced. The concentration of CFU-S was greater in femoral bone marrow of non-resected animals than in any of the regenerating liver preparations. hlicroscopy of the splenic colonies revealed a n identical morphologic pattern when the source of CFU-S was bone marrow, regenerating liver, or sham operated TABLE 5
S p l e n i c colony f o r m i n g potenhcil of regenertrti7ig l i v e r cells 1 ~~
~
Source of cells
Number iwected
'4-day post resection non-adherent 5-day post resection non-adherent Liver adherent 4-day sham operated non-adherent 5-day sham operated non-adherent
Normal
3 . 4 x 106
Spleen colonies at 10 days
106
22
106
106
25 8
106
1
106
6
105 -
0
33
0
Trypsinized cell preparations were injected i n t o the tail vein of mice irradiated with 800 rads two hours previously. 1
HAYS
liver. Colonies consisted of granulocytes, megakaryocytes, erythroid precursors, or varying combinations of these cell lines. 111 SC USSION
The present study has demonstrated that hemopoietic stein cells exist in low concentrations in adult mouse liver and increase during regeneration of the organ following partial hepatectomy. CFU-C and CFU-S have been previously described in newborn mouse liver, (Metcalf and Moore, .71), and in adult mouse liver (Silini et al., '67).Data from the present experiment reveal that the concentration of CFU-C and CFU-S is markedly increased three days after twothirds hepatic resection reaching maximum values 5-7 days after resection. Hepatocytes are known to commence active mitosis on the second, and reach maximal activity on the third day in vivo. Mitosis in macrophages develops more slowly and reaches peak activity on about the fifth day post resection (Yokoyama et al., '53; Grisham, '62), which approximates the time when CFU-C achieve maximum concentrations in the liver In this study. It is possible that CFU-S and CFU-C in regenerating liver serve as progenitors of this increasing macrophage population, as macrophages are among the progeny of CFU-C. It is suggested that the hemopoietic stem cells represent components of liver which are essentially dormant in vivo until stimulated to proliferate, either by removal from the micro-environment, or by changes associated with regeneration. The present study does not, however, exclude that the CFU-S aiid CFUC in question migrate from bone marrow. It has been shown that hepatic macrophages can be reconstituted from injected bone marrow cells (Boak et al., '68). However, the absence of a remarkable increase in numbers of CFU-C in marrow and peripheral blood in animals after partial hepatectomy suggests that we are not dealing with a n increase in CFU-C production and circulation, but an intra hepatic response to the insult of resection. Both CFU-C and CFU-S in liver cell preparations were concentrated in the nonadherent fraction. Adherent cells from liver produced CSA and those from bone marrow did not. Since macrophages, which are a major component of the adherent population, have been shown in other systems to
COLONY FORMING CELLS IN REGENERATING LIVER
produce CSA (Chervenick and LoBuglio, '72), this finding may indicate a functional difference of the macrophages derived from liver and bone marrow. Unfractionated preparations of 18-day fetal liver contained concentrations of CFUC similar to adult regenerating liver, and also supported spontaneous colony formation in agar, which was increased by addition of EMS. However, the CFU-C population in fetal liver occurs in the setting of obvious hemopoietic activity (Metcalf and Moore, '71), which is in sharp contrast to the low numbers of hemopoietic elements seen in microscopic sections during regeneration in the adult organ. Conditions of cell isolation from the liver tissue preferentially damages hepatocytes and resulted in low numbers of these cells in our preparations. A small number of colonies arising in agar from regenerating liver consisted of cords of apposed epithelioid cells which resembled hepatocytes. Cells of this type were not encountered in colonies originating from sham operated liver or bone marrow. This raises the possibility that an additional type of in vitro colony forming cell related to hepatic rather than hemopoietic tissue occurs in regenerating liver. No evidence of these cells was obtained in liquid culture or in colonies arising in spleen, and further investigation is required to establish their identity. ACKNOWLEDGMENTS
The authors wish to thank Dorothy Haskett, Melinda Newel1 and Wayne Shephard for their valuable technical assistance and Doctor Harry B. Neustein of the Department of Pathology, Childrens Hospital of Los Angeles, for his review of microscopic sections. LITERATURE CITED Boak, J. L., G. H. Christie, W. H. Ford a n d J . G . Howard 1968 Pathways in the development
219
of liver macrophages: alternative precursors contained in populations of lymphocytes a n d bone marrow-cells. Proc. Roy. SOC.B., 169: 307-327. Bradley, T. R., a n d D. Metcalf 1966 The growth of mouse bone marrow cells in nitro. Aust. J . Exp. Biol. Med. Sci., 44: 287-300. Chervenick, P. A , , and A. F. LoBuglio 1972 Hum a n blood monocytes, stimulators of granulocyte and mononuclear colony formation in nitro. Science, 1 78 : 64-1 78. Delmonte, L., a n d R. A. Liebelt 1965 Granulocytosis promoting extract of mouse tumor tissue. Partial purification. Science, 148: 521-523. Grisham, J. W. 1962 A morphologic study of deoxyribonucleic acid synthesis a n d cell proliferation i n regenerating rat liver, autoradiography with thymidine-HJ. Cancer Res., 22: 8 4 L 8 4 9 . Higgins, G. M., a n d R. M. Anderson 1931 Experimental pathology of the liver; restoration of the liver of the white rat following partial surgical removal. Arch. Pathol., 1 2 : 18CL-202. Kelly, L. S . , a n d E. L. Dobson 1971 Evidence concerning the origin of liver macrophages. Brit. J . Exp. Pathol., 52: 88-99. Kinsky, R. G., G. H. Christie, J . Elson and J. G. Howard 1969 Extra-hepatic derivation of Kupffer cells during oestrogenic stimulation of parabiosed mice. Brit. J. Exp. Pathol., 50: 438-447. Metcalf, D., and M. A . S. Moore 1971 In: Frontiers of Biology. Vol. 24. Haemopoietic Cells. A. Neuberger a n d E. L. Tatum, eds. North-Holland Amsterdam-London, pp. 71, 199. Moore, M. A. S . , a n d D. Metcalf 1970 Ontogeny of the haemopoietic system: yolk sac origin of in vivo a n d i n nitro colony forming cells in the developing mouse embryo. Brit. J . Haematol., 18: 279-296. Silini, G. L., V. Pozzi and S. Pons 1967 Studies o n the haemopoietic cells of mouse fetal liver. J . Embryol. Exp. Morphol., 17: 302-318. Stanley, E. R., T. R. Bradley and M. A. Sumner 1971 Properties of the mouse embryo condiditioned medium factor(s) stimulating colony formation by mouse bone marrow cells grown in vitro. J . Cell. Physiol., 78: 301-318. Sumner, M. A , , T. R. Bradley, G. S. Hodgson, M. J . Cline, P. A. Fry a n d L. Sutherland 1972 The growth of bone marrow cells i n liquid culture. Brit. J . Haematol., 2 3 : 221-234. Till, J. E., a n d E. A . McCulloch 1961 Direct measurement of the radiation sensitivity of normal mouse bone marrow cells. Rad. Res., 14: 212222. Yokoyama, H. O., M. E. Wilson, K. K. Tsuboi a n d R. E. Stowell 1953 Regeneration ofmouse liver after partial hepatectomy. Cancer Res., 1 3 : 8 0 85.
