Cell, Vol. 5, 11-17,
May 1975.
Copyright01975
by MIT
D-Valine as a Selective Agent for Normal Human and Rodent Epithelial Cells in Culture Scott F. Gilbert and Barbara Ft. Migeon Departments of Pediatrics and Biology Johns Hopkins University Baltimore, Maryland 21205
Summary A nutrient medium has been developed to enable the growth of normal epithellal cells while selectively lnhlbltlng flbroblast proliferation. In this medium, D-vallne Is substituted for L-valine; and only those cells containing D-amlno acid oxldase can convert the D-amino acid Into Its essential L-enantlomer. The ablllty to select for cells with this enzyme has enabled us to maintain eplthellal cell populatlons free from fibroblast overgrowth. The presence of D-amino acid oxidase has been hlstochemically conflrmed in the eplthellal cells selected from renal cell suspensions and explants. The ability to proliferate in the selective medlum Is transmitted to the clonal progeny of these cells. Moreover, epithelial cell proliferation In thls medlurn indicates the presence of D-amlno acid oxldase, which we have detected In tissues where it had not previously been reported-fetal human kldney, lung, and cord. Flbroblasts will not grow In the selective medlum, but will proliferate normally If the product of the D-amino acid oxldase reactlon Is supplied. Introduction The application of cell culture to the analysis of mammalian development has been severely limited because of the proliferative advantage of fibroblasts over the epithelial cells of an organ. Observations that outgrowths from mammalian organ explants may be initially epithelial but are soon obscured by stromal fibroblasts (Carrel and Burrows, 1910; Fleischer and Loeb, 1911; Champy, 1913) led
Medium:
D -va 1he
Cell:
w D-Valine
//“\H
Results Selection for Human Renal Eplthellal Cells Explants Fetal kidney explants in medium containing L-valine (either in the salt solution or in undialyzed fetal calf serum) produced outgrowths, initially
L-valine I =II 2-Ketoisovaleric
COOH I (CH3)2fi
Fischer (1946) to reflect that this was one of the major obstacles to the study of differentiated cells in culture. Studies of epithelial cells from explants can be performed only before fibroblastic outgrowth (Pyke and Gelfand, 1974). We have formulated a medium which makes essential an enzyme not found in fibroblasts, thus enabling the selective proliferation of epithelial cells from various organs. The selective medium is based on the requirement of the essential amino acid L-valine for cell growth (Eagle, 1955; Litwin, 1974) and on the occurrence of D-amino acid oxidase [D-amino acid: oxygen oxidoreductase (deaminating); EC 1.4.3.31 only in such specialized cells as the hepatocytes and renal tubular epithelium of many mammals, including man (Sallach and Fahien, 1969; Bliss, 1940). D-amino acid oxidase, which catalyzes the oxidative deamination of several D-amino acids, converts D-valine to 2-ketoisovaleric acid. This keto acid can then be converted to L-valine by the ubiquitous enzyme, branched-chain amino acid:2-oxoglutarate aminotransferase (Dancis et al., 1963; Ichihara and Koyoma, 1966; Shirai and Ichihara, 1971). Therefore as illustrated in Figure 1, cells with D-amino acid oxidase should be able to proliferate in nutrient medium where D-valine is substituted for L-valine (henceforth referred to as D-val; see Experimental Procedures), while fibroblasts, lacking a pathway for this transformation, should not survive. A control medium containing L-valine instead of D-valine, but otherwise identical to D-val, is referred to hereafter as L-val (see Experimental Procedures).
NH 2
FAD FADH -> D-amino acid oxida se
COOH I ’ =O I
acid
L-VZline COOH I c
H2N ‘b
(CH3) 2 g
HH
C(CH3)2
I Figure Essential D-valine
1. Basis
of Chemical
Selection
for Cells with
L-valine is removed from the nutrient to L-valine and therefore can proliferate
D-Amino
Acid
Oxidase
medium and D-valine is substituted. Epithelial while fibroblasts, lacking this enzyme, cannot.
cells
with
D-amino
acid
oxidase
convert
Cell 12
Selection 13
of Epithelial
Cells
by D. Valine
epithelial, which were subsequently obscured by the proliferation of fibroblasts. In D-val, however, fibroblast outgrowth was severely inhibited, and only epithelial populations proliferated from the explants (Figure 2). In valine-free medium, some epithelial cells proliferated for the first two weeks; but subsequently there were many less cells in the valine-free medium than in D-val. The temporary growth in valine-free medium is most likely due to a residual intracellular L-valine pool.
Single Cell Suspensions In growth medium, single cell suspensions from fetal kidney minces yielded clones at plating efficiencies from 0.30-0.95%. Although these were predominantly of the epithelial variety, fibroblast clones were also present. In regard to clonal morphology, each cell type “bred true.” Fibroblast clones always gave rise to progeny of fibroblastic morphology; and without exception, 160 subclones derived from 21 different epithelial clones maintained their epithelial morphology. In D-val, cells obtained from fetal human kidney minces yielded clones at an efficiency of approximately 0.1%. These were entirely of the epithelial
Figure
2. Human
Fetal Renal
(A and 8, opposite) (C, above) Epithelial of fibroblasts.
Epithelial
Cells
Selected
by D-Val
type. When transferred to growth medium by cloning cylinders, these clones maintained their epithelioid appearance. To determine if renal epithelial cells transmit the ability to grow in D-val, clones picked from D-val were maintained in growth medium until confluence. When these cells were replated into D-val, epithelial subclones were obtained. It is unlikely that proliferation in selective medium resulted from a large L-valine pool, as epithelial cells maintained in the selective medium for as long as four weeks produced clones when subsequently replated in D-val.
Growth lnhibifion
of Fibroblasfs
That fibroblast growth in D-val was inhibited at cell densities as great as 105/60 mm dish was demonstrated for adult human skin fibroblasts as well as for cells from the established rodent cell lines, CHO, 3T3, Cl lD, and RAG. The effect of selection was more striking in rodent cells because the inhibited cells, unlike human fibroblasts, quickly detached from the plate. Furthermore, human cells can divide once or twice before they cease to proliferate. When the intermediate product, 2-ketoisovaleric
(x 184)
Epithelial cell patterns which predominate during the first two to three weeks of culture. cells which predominate later. These patterns of cell growth are unique to cultured kidneys
and persist
in the absence
Cdl 14
acid (92 mg/l), was added to valine-free medium, both skin and renal fibroblasts were able to proliferate as well as they did in L-val. Enzyme Assay When assayed for FAD reduction using a D-amino acid as substrate, the kidney epithelial cytoplasm of D-val-selected cells displayed a strong purple color, whereas dermal fibroblasts could be seen in outline only. This is evidence that the pertinent enzyme is the D-amino acid oxidase rather than a heretofore undetected racemase able to convert D-valine into L-valine directly. Selection for Epithelial Cells from Other Sources The outgrowths from fetal lung and cord in growth medium, unlike those of the kidney, were predominantly fibroblastic, with only rare epithelial cells. In D-vat, however, fibroblasts were inhibited shortly after their appearance; and by the end of the second week in culture, large areas of epithelial cells could be observed (Figure 4). This is surprising, as neither organ has been thought to contain significant amounts of D-amino acid oxidase. Selection is not limited to human fetal tissues. D-val selected for epithelial cells, while L-val permitted the rapid proliferation of fibroblasts from adult mouse and rabbit kidney explants. In valinefree medium, proliferation from the explants was sparse. Discussion Several nonspecific chemical methods have been used to encourage the selective proliferation of epithelial cells in culture, including vitamin B complex (Heaton, 1926) hexenolactone (Medawar, Robinson, and Robinson, 1943), or steroids (Waymouth, Chen, and Wood, 1971) as components of the nutrient medium. Media have also been developed to favor the proliferation of cells with differentiated functions from tumor-derived HTC hepatoma (Kulka, Tomkins, and Crook, 1972) and murine neuroblastoma (Breakefield and Nirenberg, 1974). Normal hepatocytes have been selected from fetal rat liver by making essential the liver-specific enzymes of the urea cycle (Leffert and Paul, 1972). We have developed a nutrient medium specifically favoring the growth in vitro of normal epithelial cells from various human and rodent organs on the basis of the cell’s ability to utilize D-amino acids. It is clear that D-val is not acting as a selective poison for fibroblasts because these cells proliferate normally in medium containing D-valine as long as L-valine is included (Figure 3). This indicates that the absence of L-valine, rather than the presence of D-valine, inhibits fibroblast proliferation.
Only epithelial cells proliferated in D-val, and this ability is transferred to the progeny of these cells. Furthermore, we have confirmed histochemically the presence of D-amino acid oxidase in the selected cells and have shown that fibroblasts supplied with the product of the D-amino acid oxidase reaction are able to proliferate. These findings suggest that the cells selected by D-val transmit to progeny not only their morphology, but also their ability to synthesize a “differentiated cell” enzyme in vitro. This system enabled us to obtain epithelial clones and proliferation of epithelial cells not only from renal explants, but also from nonrenal tissues (that is, lung and cord) which rarely produce outgrowths of epithelial cells. Above all, these epithelial cell populations can be maintained free from fibroblast overgrowth. D-val selection must be maintained to inhibit fibroblasts in mixed populations of fibroblasts and epithelial cells or in dishes containing explants. Although many of the fibroblasts in D-val detach from the dish, the few which remain, even after weeks of L-val deprivation, are capable of rapid proliferation when L-val is supplied. On the other hand, epithelial clones transferred as clones to dishes not containing fibroblasts may be maintained in nonselective medium. Cell growth has been shown to be a sensitive indicator for the presence of an enzyme or substrate (Guthrie and Susi, 1963; Spector and Bloom, 1973). It is not surprising that by manometric methods, Damino acid oxidase was not found in cultured mammalian kidney cells (Burlington, 1959), since the relative number of cultured cells containing this enzyme may be small. In vertebrates, D-amino acid oxidase was long thought to occur only in the kidney and liver (Krebs, 1935; Krebs, 1951) where racemation of D-amino acids has been observed (Gibson et al., 1954). However, this enzyme has more recently been detected in human cerebellum (Neims, Zieverink, and Smilack, 1966) and neutrophils (Cline and Lehrer, 1969). Our bioassay has enabled us to infer the presence of D-amino acid oxidase in the human fetal lung and cord-two places where it had not previously been found (Dunn and Perkoff, 1963)-and in the 10 week human embryonic kidney. This suggests that the enzyme may be more prevalent than is generally believed. Investigations are currently underway to determine if each of the epithelial cells favored by D-val is organ specific or if a common cell type (such as vascular endothelium) is being selected. Should the D-val-selected epithelial cells of renal origin have, in addition to a unique growth pattern, enzymes found specifically in renal parenchymal cells, they will provide a model for studies of differentia-
Selection 15
of Epithelial
Cells
by D-Valine
tion. Furthermore, the ability of renal epithelial cells to proliferate in a medium where established rodent cell lines cannot should permit the assignment of the gene for D-amino acid oxidase to a specific human chromosome. Experimental
organ minces into Lux plastic tissue culture dishes and flooding them with the appropriate medium. In the case of kidney samples, the capsule was removed prior to obtaining minces and cell suspensions. These ceil suspensions, where greater than 90% of the ceils were present as single cells, were prepared by the method of Leffert and Paul (1972) except that an equal volume of 0.25% trypsin (GIBCO) was added to the collagenase, and the cells were suspended in L-valine-free medium containing 15% dialyzed fetal calf serum. Established rodent cell lines assayed for proliferation in D-val were Chinese hamster ovary cell line CHO K-l (CCLGl) and mouse renal adenocarcinoma RAG (CCL 142) from the American Type Culture Collection, and mouse lines 3T3 and Cl 1D contributed by M. Weiss.
Procedures
Specimens Kidneys and samples from other organs were obtained aseptically from -the dead products of therapeutic abortions of lo-20 weeks’ gestational age, -two 4 week-old Swiss albino mice. -and an adult New Zealand white rabbit. These specimens were placed in ice-cold sterile Hanks’ salt soiution for less than an hour until cultures were initiated by cutting
I.-VALINE
Cl
Nutrient Media Growth medium was Eagle’s minimal salts) supplemented with nonessential
D,L-VALINE
essential medium (Earle’s amino acids and 15% fetal
D-VALINE
1D (IO31
RAG
Figure
3. Inhibition
of Growth
of Rodent
Clones were obtained after 10 days from obtained in D-val. even from kidney-derived
Established
Cell Lines
by D-Val
populations of 103 cells RAG cells.
plated
into L-val
or into a 1 :l mixture
of L-val
and D-vat
None
were
Cell 16
Figure
4. Outgrowths
Proliferation
from
of (A) fibroblasts
Lung
Explants
in L-val,
after
2 Weeks
and (6) epithelial
in L-Val cells
or D-Val
in D-val.
(x 230)
Phase-contrast
micrograph.
Selection 17
of Epithelial
Cells
by D-Valine
calf serum (GIBCO). Medium selective for the growth of epithelial cells (D-val) was prepared by reformulating the growth medium salt solution, substituting D-val (92 mg/l) for L-val(46 mg/l). This solution was supplemented with (a) nonessential amino acids and (b) 15% fetal calf serum which had been dialyzed 20-28 hr at 4°C against three changes of 50 vol each of 0.85% NaCI. A third medium, L-val, was prepared in the same way as D-val (including the dialyzed fetal calf serum), but with L-valine (46 mg/l) instead of D-valine. In addition, a medium composed of D-valine salt solution and undialyzed fetal calf serum was prepared as a control for some studies. Enzyme Assay Wohlrab’s method (1965) was used to stain for D-amino acid oxidase as modified for living cells in tissue culture by quick acetone fixation (Melnick, 1971). D-alanine was used as substrate since its reaction is catalyzed more efficiently than with D-valine (Krebs, 1951). Cloning Clones were 1962).
isolated
with glass
cloning
cylinders
(Ham
and Puck,
Acknowledgment This work was supported by a USPHS research grant. S.F.G. is a pre-doctoral fellow supported by an NIH training grant to the Department of Biology. Received
December
16, 1974;
revised
February
7, 1975
References Bliss,
S. (1941).
J. Biol. Chem.
137, 217-225.
Breakefield, X. 0.. and Nirenberg, Sci. USA 71, 2530-2533. Burlington, Carrel, 1379-I Champy,
H. (1959).
Am. J. Physiol.
A., and Burrows, 381. C. (1913).
Lehrer,
Anat.
Perkoff,
Eagle,
J. Biol. Chem.
H. (1955).
Fleisher, 8, 133-l
M. S., and 38.
Gibson, Q. H., Newley, J. Physiol. 125, 65~. Guthrie,
R., and Susi,
Ham, R. G., and 5, S. P. Colowick Press), p. 90. Heaton, Ichihara. Krebs,
Loeb,
A. (1963).
J. Path.
H. A. (1935).
55.
Sci.
USA
Nat. Acad. M. (1963).
Biochem.
Proc.
Sot.
Leffert,
Acta
Pediatrics
Batter.
Cambridge
Exp.
Biol. Med.
B. C. (1954).
32, 338-343.
In Methods eds. (New
of Enzymology, York: Academic
29, 293-306. J. Biochem.
59, 160-169.
J. 29, 1620-1644.
Krebs. H. A. (1951). In The Enzymes, 2, J. B. Sumner bath, eds. (New York: Academic Press), pp. 499-514. Kulka, R. G.. Tomkins. Biol. 54, 175-179.
Biochem.
Biophys.
D. H., and Whaler,
E. (1966).
Biochem.
Assoc.
Cells (Cambridge:
L. (1911).
H., Smyth.
A., and Koyoma,
Med.
274, 839-852.
of Tissue
Puck, T. T. (1962). and N. 0. Kaplan,
T. B. (1926)
Proc.
G. T. (1963).
Biology p. 95.
Nat. Acad.
23, 184-205.
J., and Levitz.
Dunn, J. T., and 73, 327-331. Fischer, A. (1946). University Press),
J. Amer.
Ft. I. (1969).
Dancis, J., Jensen, V., Hutzler, Biophys. Acta 77, 523-524.
Proc.
797, 68-70.
M. T. (1910).
Bibliog.
Cline, M. J., and 62, 756-763.
M. W. (1974).
G. M., and
H. L., and Paul, D. (1972).
Crook,
and K. Myr-
R. B. (1972).
J. Cell Biol. 52, 559-568.
J. Cell
Litwin,
J. (1974).
Medawar, 157, 195.
J. Cell Sci. 74, 671-680
P. B., Robinson,
Melnick,
P. J. (1971).
Neims. them.
A. H., Zieverink. 73, 163-168.
Pyke.
G. M.. and Robinson,
Progr.
Shirai, Spector,
Wohlrab.
E. W. (1974).
A. (1971).
E. B., and Bloom,
F. (1965).
2(l),
J. D. (1966). Nature
J. Biochem.
A. D. (1973).
Pediat.
H. W., and Wood, Histochemie
5, 311-325
Nature 11-15. J. Neuro-
257, 421-423.
Fahien, L. A. (1969). In Metabolic ed. (New York: Academic Press),
A., and Ichihara,
Waymouth, C., Chen, 371 (abstract).
Cytochem.
W. D., and Smilack.
K. W., and Gelfand.
Sallach, H. J., and 3, D. M. Greenberg,
Histochem.
R. (1943).
Pathways, p. 47.
70, 741-748. Res.
B. G. (1971).
7, 700-705. In Vitro
6,
May 1975.
Copyright01975
by MIT
D-Valine as a Selective Agent for Normal Human and Rodent Epithelial Cells in Culture Scott F. Gilbert and Barbara Ft. Migeon Departments of Pediatrics and Biology Johns Hopkins University Baltimore, Maryland 21205
Summary A nutrient medium has been developed to enable the growth of normal epithellal cells while selectively lnhlbltlng flbroblast proliferation. In this medium, D-vallne Is substituted for L-valine; and only those cells containing D-amlno acid oxldase can convert the D-amino acid Into Its essential L-enantlomer. The ablllty to select for cells with this enzyme has enabled us to maintain eplthellal cell populatlons free from fibroblast overgrowth. The presence of D-amino acid oxidase has been hlstochemically conflrmed in the eplthellal cells selected from renal cell suspensions and explants. The ability to proliferate in the selective medlum Is transmitted to the clonal progeny of these cells. Moreover, epithelial cell proliferation In thls medlurn indicates the presence of D-amlno acid oxldase, which we have detected In tissues where it had not previously been reported-fetal human kldney, lung, and cord. Flbroblasts will not grow In the selective medlum, but will proliferate normally If the product of the D-amino acid oxldase reactlon Is supplied. Introduction The application of cell culture to the analysis of mammalian development has been severely limited because of the proliferative advantage of fibroblasts over the epithelial cells of an organ. Observations that outgrowths from mammalian organ explants may be initially epithelial but are soon obscured by stromal fibroblasts (Carrel and Burrows, 1910; Fleischer and Loeb, 1911; Champy, 1913) led
Medium:
D -va 1he
Cell:
w D-Valine
//“\H
Results Selection for Human Renal Eplthellal Cells Explants Fetal kidney explants in medium containing L-valine (either in the salt solution or in undialyzed fetal calf serum) produced outgrowths, initially
L-valine I =II 2-Ketoisovaleric
COOH I (CH3)2fi
Fischer (1946) to reflect that this was one of the major obstacles to the study of differentiated cells in culture. Studies of epithelial cells from explants can be performed only before fibroblastic outgrowth (Pyke and Gelfand, 1974). We have formulated a medium which makes essential an enzyme not found in fibroblasts, thus enabling the selective proliferation of epithelial cells from various organs. The selective medium is based on the requirement of the essential amino acid L-valine for cell growth (Eagle, 1955; Litwin, 1974) and on the occurrence of D-amino acid oxidase [D-amino acid: oxygen oxidoreductase (deaminating); EC 1.4.3.31 only in such specialized cells as the hepatocytes and renal tubular epithelium of many mammals, including man (Sallach and Fahien, 1969; Bliss, 1940). D-amino acid oxidase, which catalyzes the oxidative deamination of several D-amino acids, converts D-valine to 2-ketoisovaleric acid. This keto acid can then be converted to L-valine by the ubiquitous enzyme, branched-chain amino acid:2-oxoglutarate aminotransferase (Dancis et al., 1963; Ichihara and Koyoma, 1966; Shirai and Ichihara, 1971). Therefore as illustrated in Figure 1, cells with D-amino acid oxidase should be able to proliferate in nutrient medium where D-valine is substituted for L-valine (henceforth referred to as D-val; see Experimental Procedures), while fibroblasts, lacking a pathway for this transformation, should not survive. A control medium containing L-valine instead of D-valine, but otherwise identical to D-val, is referred to hereafter as L-val (see Experimental Procedures).
NH 2
FAD FADH -> D-amino acid oxida se
COOH I ’ =O I
acid
L-VZline COOH I c
H2N ‘b
(CH3) 2 g
HH
C(CH3)2
I Figure Essential D-valine
1. Basis
of Chemical
Selection
for Cells with
L-valine is removed from the nutrient to L-valine and therefore can proliferate
D-Amino
Acid
Oxidase
medium and D-valine is substituted. Epithelial while fibroblasts, lacking this enzyme, cannot.
cells
with
D-amino
acid
oxidase
convert
Cell 12
Selection 13
of Epithelial
Cells
by D. Valine
epithelial, which were subsequently obscured by the proliferation of fibroblasts. In D-val, however, fibroblast outgrowth was severely inhibited, and only epithelial populations proliferated from the explants (Figure 2). In valine-free medium, some epithelial cells proliferated for the first two weeks; but subsequently there were many less cells in the valine-free medium than in D-val. The temporary growth in valine-free medium is most likely due to a residual intracellular L-valine pool.
Single Cell Suspensions In growth medium, single cell suspensions from fetal kidney minces yielded clones at plating efficiencies from 0.30-0.95%. Although these were predominantly of the epithelial variety, fibroblast clones were also present. In regard to clonal morphology, each cell type “bred true.” Fibroblast clones always gave rise to progeny of fibroblastic morphology; and without exception, 160 subclones derived from 21 different epithelial clones maintained their epithelial morphology. In D-val, cells obtained from fetal human kidney minces yielded clones at an efficiency of approximately 0.1%. These were entirely of the epithelial
Figure
2. Human
Fetal Renal
(A and 8, opposite) (C, above) Epithelial of fibroblasts.
Epithelial
Cells
Selected
by D-Val
type. When transferred to growth medium by cloning cylinders, these clones maintained their epithelioid appearance. To determine if renal epithelial cells transmit the ability to grow in D-val, clones picked from D-val were maintained in growth medium until confluence. When these cells were replated into D-val, epithelial subclones were obtained. It is unlikely that proliferation in selective medium resulted from a large L-valine pool, as epithelial cells maintained in the selective medium for as long as four weeks produced clones when subsequently replated in D-val.
Growth lnhibifion
of Fibroblasfs
That fibroblast growth in D-val was inhibited at cell densities as great as 105/60 mm dish was demonstrated for adult human skin fibroblasts as well as for cells from the established rodent cell lines, CHO, 3T3, Cl lD, and RAG. The effect of selection was more striking in rodent cells because the inhibited cells, unlike human fibroblasts, quickly detached from the plate. Furthermore, human cells can divide once or twice before they cease to proliferate. When the intermediate product, 2-ketoisovaleric
(x 184)
Epithelial cell patterns which predominate during the first two to three weeks of culture. cells which predominate later. These patterns of cell growth are unique to cultured kidneys
and persist
in the absence
Cdl 14
acid (92 mg/l), was added to valine-free medium, both skin and renal fibroblasts were able to proliferate as well as they did in L-val. Enzyme Assay When assayed for FAD reduction using a D-amino acid as substrate, the kidney epithelial cytoplasm of D-val-selected cells displayed a strong purple color, whereas dermal fibroblasts could be seen in outline only. This is evidence that the pertinent enzyme is the D-amino acid oxidase rather than a heretofore undetected racemase able to convert D-valine into L-valine directly. Selection for Epithelial Cells from Other Sources The outgrowths from fetal lung and cord in growth medium, unlike those of the kidney, were predominantly fibroblastic, with only rare epithelial cells. In D-vat, however, fibroblasts were inhibited shortly after their appearance; and by the end of the second week in culture, large areas of epithelial cells could be observed (Figure 4). This is surprising, as neither organ has been thought to contain significant amounts of D-amino acid oxidase. Selection is not limited to human fetal tissues. D-val selected for epithelial cells, while L-val permitted the rapid proliferation of fibroblasts from adult mouse and rabbit kidney explants. In valinefree medium, proliferation from the explants was sparse. Discussion Several nonspecific chemical methods have been used to encourage the selective proliferation of epithelial cells in culture, including vitamin B complex (Heaton, 1926) hexenolactone (Medawar, Robinson, and Robinson, 1943), or steroids (Waymouth, Chen, and Wood, 1971) as components of the nutrient medium. Media have also been developed to favor the proliferation of cells with differentiated functions from tumor-derived HTC hepatoma (Kulka, Tomkins, and Crook, 1972) and murine neuroblastoma (Breakefield and Nirenberg, 1974). Normal hepatocytes have been selected from fetal rat liver by making essential the liver-specific enzymes of the urea cycle (Leffert and Paul, 1972). We have developed a nutrient medium specifically favoring the growth in vitro of normal epithelial cells from various human and rodent organs on the basis of the cell’s ability to utilize D-amino acids. It is clear that D-val is not acting as a selective poison for fibroblasts because these cells proliferate normally in medium containing D-valine as long as L-valine is included (Figure 3). This indicates that the absence of L-valine, rather than the presence of D-valine, inhibits fibroblast proliferation.
Only epithelial cells proliferated in D-val, and this ability is transferred to the progeny of these cells. Furthermore, we have confirmed histochemically the presence of D-amino acid oxidase in the selected cells and have shown that fibroblasts supplied with the product of the D-amino acid oxidase reaction are able to proliferate. These findings suggest that the cells selected by D-val transmit to progeny not only their morphology, but also their ability to synthesize a “differentiated cell” enzyme in vitro. This system enabled us to obtain epithelial clones and proliferation of epithelial cells not only from renal explants, but also from nonrenal tissues (that is, lung and cord) which rarely produce outgrowths of epithelial cells. Above all, these epithelial cell populations can be maintained free from fibroblast overgrowth. D-val selection must be maintained to inhibit fibroblasts in mixed populations of fibroblasts and epithelial cells or in dishes containing explants. Although many of the fibroblasts in D-val detach from the dish, the few which remain, even after weeks of L-val deprivation, are capable of rapid proliferation when L-val is supplied. On the other hand, epithelial clones transferred as clones to dishes not containing fibroblasts may be maintained in nonselective medium. Cell growth has been shown to be a sensitive indicator for the presence of an enzyme or substrate (Guthrie and Susi, 1963; Spector and Bloom, 1973). It is not surprising that by manometric methods, Damino acid oxidase was not found in cultured mammalian kidney cells (Burlington, 1959), since the relative number of cultured cells containing this enzyme may be small. In vertebrates, D-amino acid oxidase was long thought to occur only in the kidney and liver (Krebs, 1935; Krebs, 1951) where racemation of D-amino acids has been observed (Gibson et al., 1954). However, this enzyme has more recently been detected in human cerebellum (Neims, Zieverink, and Smilack, 1966) and neutrophils (Cline and Lehrer, 1969). Our bioassay has enabled us to infer the presence of D-amino acid oxidase in the human fetal lung and cord-two places where it had not previously been found (Dunn and Perkoff, 1963)-and in the 10 week human embryonic kidney. This suggests that the enzyme may be more prevalent than is generally believed. Investigations are currently underway to determine if each of the epithelial cells favored by D-val is organ specific or if a common cell type (such as vascular endothelium) is being selected. Should the D-val-selected epithelial cells of renal origin have, in addition to a unique growth pattern, enzymes found specifically in renal parenchymal cells, they will provide a model for studies of differentia-
Selection 15
of Epithelial
Cells
by D-Valine
tion. Furthermore, the ability of renal epithelial cells to proliferate in a medium where established rodent cell lines cannot should permit the assignment of the gene for D-amino acid oxidase to a specific human chromosome. Experimental
organ minces into Lux plastic tissue culture dishes and flooding them with the appropriate medium. In the case of kidney samples, the capsule was removed prior to obtaining minces and cell suspensions. These ceil suspensions, where greater than 90% of the ceils were present as single cells, were prepared by the method of Leffert and Paul (1972) except that an equal volume of 0.25% trypsin (GIBCO) was added to the collagenase, and the cells were suspended in L-valine-free medium containing 15% dialyzed fetal calf serum. Established rodent cell lines assayed for proliferation in D-val were Chinese hamster ovary cell line CHO K-l (CCLGl) and mouse renal adenocarcinoma RAG (CCL 142) from the American Type Culture Collection, and mouse lines 3T3 and Cl 1D contributed by M. Weiss.
Procedures
Specimens Kidneys and samples from other organs were obtained aseptically from -the dead products of therapeutic abortions of lo-20 weeks’ gestational age, -two 4 week-old Swiss albino mice. -and an adult New Zealand white rabbit. These specimens were placed in ice-cold sterile Hanks’ salt soiution for less than an hour until cultures were initiated by cutting
I.-VALINE
Cl
Nutrient Media Growth medium was Eagle’s minimal salts) supplemented with nonessential
D,L-VALINE
essential medium (Earle’s amino acids and 15% fetal
D-VALINE
1D (IO31
RAG
Figure
3. Inhibition
of Growth
of Rodent
Clones were obtained after 10 days from obtained in D-val. even from kidney-derived
Established
Cell Lines
by D-Val
populations of 103 cells RAG cells.
plated
into L-val
or into a 1 :l mixture
of L-val
and D-vat
None
were
Cell 16
Figure
4. Outgrowths
Proliferation
from
of (A) fibroblasts
Lung
Explants
in L-val,
after
2 Weeks
and (6) epithelial
in L-Val cells
or D-Val
in D-val.
(x 230)
Phase-contrast
micrograph.
Selection 17
of Epithelial
Cells
by D-Valine
calf serum (GIBCO). Medium selective for the growth of epithelial cells (D-val) was prepared by reformulating the growth medium salt solution, substituting D-val (92 mg/l) for L-val(46 mg/l). This solution was supplemented with (a) nonessential amino acids and (b) 15% fetal calf serum which had been dialyzed 20-28 hr at 4°C against three changes of 50 vol each of 0.85% NaCI. A third medium, L-val, was prepared in the same way as D-val (including the dialyzed fetal calf serum), but with L-valine (46 mg/l) instead of D-valine. In addition, a medium composed of D-valine salt solution and undialyzed fetal calf serum was prepared as a control for some studies. Enzyme Assay Wohlrab’s method (1965) was used to stain for D-amino acid oxidase as modified for living cells in tissue culture by quick acetone fixation (Melnick, 1971). D-alanine was used as substrate since its reaction is catalyzed more efficiently than with D-valine (Krebs, 1951). Cloning Clones were 1962).
isolated
with glass
cloning
cylinders
(Ham
and Puck,
Acknowledgment This work was supported by a USPHS research grant. S.F.G. is a pre-doctoral fellow supported by an NIH training grant to the Department of Biology. Received
December
16, 1974;
revised
February
7, 1975
References Bliss,
S. (1941).
J. Biol. Chem.
137, 217-225.
Breakefield, X. 0.. and Nirenberg, Sci. USA 71, 2530-2533. Burlington, Carrel, 1379-I Champy,
H. (1959).
Am. J. Physiol.
A., and Burrows, 381. C. (1913).
Lehrer,
Anat.
Perkoff,
Eagle,
J. Biol. Chem.
H. (1955).
Fleisher, 8, 133-l
M. S., and 38.
Gibson, Q. H., Newley, J. Physiol. 125, 65~. Guthrie,
R., and Susi,
Ham, R. G., and 5, S. P. Colowick Press), p. 90. Heaton, Ichihara. Krebs,
Loeb,
A. (1963).
J. Path.
H. A. (1935).
55.
Sci.
USA
Nat. Acad. M. (1963).
Biochem.
Proc.
Sot.
Leffert,
Acta
Pediatrics
Batter.
Cambridge
Exp.
Biol. Med.
B. C. (1954).
32, 338-343.
In Methods eds. (New
of Enzymology, York: Academic
29, 293-306. J. Biochem.
59, 160-169.
J. 29, 1620-1644.
Krebs. H. A. (1951). In The Enzymes, 2, J. B. Sumner bath, eds. (New York: Academic Press), pp. 499-514. Kulka, R. G.. Tomkins. Biol. 54, 175-179.
Biochem.
Biophys.
D. H., and Whaler,
E. (1966).
Biochem.
Assoc.
Cells (Cambridge:
L. (1911).
H., Smyth.
A., and Koyoma,
Med.
274, 839-852.
of Tissue
Puck, T. T. (1962). and N. 0. Kaplan,
T. B. (1926)
Proc.
G. T. (1963).
Biology p. 95.
Nat. Acad.
23, 184-205.
J., and Levitz.
Dunn, J. T., and 73, 327-331. Fischer, A. (1946). University Press),
J. Amer.
Ft. I. (1969).
Dancis, J., Jensen, V., Hutzler, Biophys. Acta 77, 523-524.
Proc.
797, 68-70.
M. T. (1910).
Bibliog.
Cline, M. J., and 62, 756-763.
M. W. (1974).
G. M., and
H. L., and Paul, D. (1972).
Crook,
and K. Myr-
R. B. (1972).
J. Cell Biol. 52, 559-568.
J. Cell
Litwin,
J. (1974).
Medawar, 157, 195.
J. Cell Sci. 74, 671-680
P. B., Robinson,
Melnick,
P. J. (1971).
Neims. them.
A. H., Zieverink. 73, 163-168.
Pyke.
G. M.. and Robinson,
Progr.
Shirai, Spector,
Wohlrab.
E. W. (1974).
A. (1971).
E. B., and Bloom,
F. (1965).
2(l),
J. D. (1966). Nature
J. Biochem.
A. D. (1973).
Pediat.
H. W., and Wood, Histochemie
5, 311-325
Nature 11-15. J. Neuro-
257, 421-423.
Fahien, L. A. (1969). In Metabolic ed. (New York: Academic Press),
A., and Ichihara,
Waymouth, C., Chen, 371 (abstract).
Cytochem.
W. D., and Smilack.
K. W., and Gelfand.
Sallach, H. J., and 3, D. M. Greenberg,
Histochem.
R. (1943).
Pathways, p. 47.
70, 741-748. Res.
B. G. (1971).
7, 700-705. In Vitro
6,
