Properties of Somatic Cell Hybrids Between Mouse Cells and Simian Virus 40-Transformed Rat Cells

Download (0)

Full text

(1)

JOURNALOFVIROLOGY, JUIY 1972, p. 67-72

Copyright ( 1972 American Society for Microbiology Printed in U.S.A.Vol. 10,No. I

Properties of Somatic

Cell

Hybrids Between

Mouse

Cells and

Simian Virus 40-Transformed

Rat

Cells

J. VAN DER NOORDAA, ANNE VAN HAAGEN, J. M. M. WALBOOMERS

AND H. VAN SOMEREN'

LaboratoriumvoordeGezonidheidsleer, University of Amsterdam, Mauritskade57, Amsterdam, The Netherlands

Received for publication6 March1972

Hybrids between mouse cells and simian virus 40 (SV40)-transformed rat cells weremade, and their properties and chromosome constitutionwereinvestigatedover

manygenerations. Their hybrid naturewas confirmed byenzymestudies. Duringa

period of 1 year a loss of 10 to 20% of the total number of chromosomes was

observed. The SV40tumorantigenwaspresentand remainedpresentin the hybrids.

Theparental andhybrid cells were studiedfor agglutination with concanavalin A,

forgrowth in softagar,and forserumrequirement. These growth and surface

char-acteristics of the transformed cells appeared in the hybrids.

Several hybrid cell lines

originating

from the

fusion of normal and virus-transformed cells

have been described. Defendi et al.

(4)

have

studiedthepropertiesofhybridsbetween normal

mouse cells and mouse cells transformed

by

polyoma virus. The

hybrid

cells contained the

polyoma

virus-specific

tumorand

transplantation

antigens,

indicating

the

persistence

of the viral

genome inthe

hybrid

cell.

Hybrids between mouse cells and

polyoma-transformed hamstercellswereshown

by

Basilico

and Wang

(1)

to contain the viral genome and

appeared to havea number of

properties

ofthe

transformed cells. Hybrids between mouse cells

and simian virus 40

(SV40)-transformed

human

cells were described

by

Weiss

(13).

The

SV40-specific tumor

(T) antigen

was present in the

nuclei of thehybridcells. Thistypeof

interspecific

hybridundergoesa

rapid

lossofhuman

chromo-somes andtherefore was considered suited to a

study of the effect of chromosome loss on the

presence of the SV40 viral genome. A

positive

correlationwasfound between the loss ofhuman

chromosomesand the loss ofT

antigen

from the

hybrid cells, suggesting

integration

of the SV40

genome intothechromosomes of thetransformed cell.

In our study we report on the

properties

of

hybrid cells between SV40-transformed rat cells

andnormal mousecells.

According

toWeiss and Ephrussi (14), rat-mouse hybrids

preferentially

loserat chromosomes, and therefore this type of

IPresent address: Medical Biological Laboratory T.N.O.,

Rijswijk, andLaboratoryfor MolecularGenetics, University of Leiden.

hybrid was chosen to study the effect of chromo-some loss on theproperties ofthehybrid cell.

MATERIALS AND METHODS

Ratcells.Cultures of primary rat cells derived from

the brains of newborn BN/BI rats were inoculated

with SV40 (strain VA 45-54) at a multiplicity of

infection of about 10. A line of transformedcellswas

initiatedfrom a focus oftransformed cells. This line

will be referred to as rat-SV40cells. All cells of this

linecontained the SV40 T antigen. No infectious virus

could be recovered from the transformed cells after

either co-cultivation or fusion with monkeykidney

cells (BSC-1 andCVl).

Mousecells. Athymidinekinase-deficient derivative

from 3T3 cells was kindly provided by D. Bootsma

(Dept. of CellBiologyandGenetics,MedicalFaculty,

Rotterdam) andwillbereferred to as 3T3 TK-.

Hybrid cells. Theisolation of hybrid cells was

per-formed by the half-selective system of Davidson and

Ephrussi (3). The rat-SV40 (at passagelevel 13) and

3T3 TK-cells were mixed in aratio of 1:1,000 and

fused by Sendai virus (10) inactivated with

beta-propiolactone. After 24hr ofincubation at 37 C, the

mediumwasreplacedbyLittlefield'sselective medium

(7). This medium consists of standard medium

sup-plementedwith1 X 104 Mhypoxanthine,4 X 107M

aminopterin, and 1.6 X 105 M thynmidine (HAT medium). Aftertwoweeksof furtherincubation,two

types of colonies were observed: colonies with the

morphology of rat-SV40 cells and colonies with a

morphology different fromthe rat-SV40 cells. Three

of thelatter colonies wereisolated with the aid ofa

Pasteur capillary pipette, and one of these three

identical-looking colonies (H) was used for further study.

No macroscopically visible colonies of 3T3

TK-cellswerepresentatthattime,althoughsmall groups

67

on November 10, 2019 by guest

http://jvi.asm.org/

(2)

orsingle cells with themorphologyof3T3cells could

befound.

From colony H, four subclones were derived for

further experiments; these lines will bereferred to as

H2, H3, H8, andH1O.

Cultureconditions. Allcultures weremaintained in

Eagle basal medium supplemented with

10%O

calf

serum. Parental and hybrid lines were subcultured

twiceaweek.

Enzyme studies. Enzyme electrophoresis was

per-formed oncelluloseacetate gel (9) with celllysates of

parentalandhybridlines. Thefollowing enzymeswere

tested: lactate dehydrogenase Aand B (LDH Aand

B), 6-phosphogluconate dehydrogenase (6PGD), and

indophenol oxidase (IPO).

SV40Tantigen. The SV40 T antigen wasdetected

by the indirect immunofluorescence test (12). For all

tests, serumfrom the same pool was used. Theserum

was obtained from hamsters bearing SV40-induced

tumors. Fluorescein-labeledantihamster globulinwas

purchased from Nordic

Phiarmaceuticals

and

Diag-nostics.

Agglutination by concanavalin A. Cultures were

washed two times with Ca-and Mg-free

phosphate-buffered saline (PBS) and dispersed with 0.02%

ethylenediaminetetraacetic

acid at 37 C. Thedispersed

cells werethenwashed twice with PBS free fromCa2+

and Mg2+. To study the effect of trypsin on

ag-glutination, cells were treated for 2 min with 0.05%c

trypsin,centrifuged, washed two times with PBS, and

resuspended. The agglutination reaction was

per-formed in 25-mm petri dishes (Falcon) with 0.5 ml of

cellsuspension (106 to 3 X 101cells1/ml)and 0.5 ml of

asolutioncontaining5,25, 50, or 125pg of

concanava-lin A. Concanavalin A (Serva, Heidelberg) was

dis-solved in saline, pH 7. After gently shaking for 15 min

at 25 C,thedegree ofagglutinationwasscored by two

investigatorsas thepercentageofagglutinatedcells.

Serum requirement. To determine the serum

re-quirement, cultures were fed with medium supple-mented with 1, 5, and 101'" serum. An inoculum of

2 X 104 cells was seeded in 60-mm petri dishes

(Falcon), and the cellswerecountedafter 1, 2, and3

days.

Ability to grow in soft agar. Appropriate dilutions

of cells in 2ml of0.34%/ agar medium were planted

on abase layer of5ml of0.5%" agarmediuminsmall

flasks (Falcon); the number ofcolonies wascounted

after 2weeks.

:Nmmi-n

mm

A

_immim

-_

RESULTS

Enzyme and chromosome studies. The

zy-mogram patterns of the four hybrid lines for

LDH, 6PGD, and IPO) showed

intermediate

bands; Fig. 1 represents theresults for one of the

hybrids. These findings showed the hybrid nature

FIG. 1. Zymograms oflactate dehydrogenase (A),

6-phosphogltaconiate dehydrogenlase (B), and

indophe-nol oxidase (C)showing rat-SV40 (1), hybridH3 (2),

artificial mixture ofrat-SV40 and 3T3 TK- (3), and

3T3 TK- (4) patte(rtis.

... ..:....

...

.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...

*.:.:.:..

*~~~~~~~~~~~~~~~~~~~~~~

...:..:... .. .. .. ... ... ... ;.

68

on November 10, 2019 by guest

http://jvi.asm.org/

(3)

PROPERTIES OF SOMATIC CELL HYBRIDS

of the lines H2, H3, H8, and H10 at passage 14. Starting at passage 20, most of our studies were performed on cells cultured in standard medium and on cells cultured in HAT medium. At passage 45the hybrid cells were tested again; their hybrid naturehad been retained for the tested enzymes

in bothmedia used.

Figures 2 and 3 show the results of the chro-mosomal studies of the parental mouse 3T3 TK-cells and rat-SV40 TK-cells. The modal number of chromosomes of the 3T3 TK- cells is 63; they are all telocentric or acrocentric. The rat-SV40 cells had a modal number of 42 chromosomes, of which about 24 were telocentric or acrocentric and about 18 were metacentric or submetacentric. The number and distribution of the four hybrid

cell lines are shown in Table 1. When tested at

about the 12th subculture after cloning, thecells

ofthe four hybrid lines contained 70 to 80 telo-centric or acrotelo-centric chromosomes and 11 to 16 metacentrics. Upon further subculturing, a small

cL

44 X.LLS LL

Zo 10-_

,

r rr;1f

m

40 50 60 70 >70

CHROMOSOMES/METAPHASE

FIG. 2. Histogram of clhromosomes/metaphase

againistiiumberofmetaphasesof3T3TK- cells.

e-1

,

2:e

co

Y-94 -tLLS

n 1;

F

ri

loss of telocentric or acrocentric chromosomes

andaminorchangeinthenumber ofmetacentric

orsubmetacentric chromosomes wasobserved. It

remains open to question whether rat or mouse chromosomes werelost.

Infectious virus and SV40 T antigen. The rat-SV40 and hybrid cells did not shed infectious

SV40, nor wasitpossible toinduce the production of SV40by means of fusionwith monkeykidney cells. The hybrid cells cultured in standard medium and HAT medium were first tested for the presence of the SV40 T antigen at about the 15th subculture. All cells from the four hybrid

lines contained the SV40 T antigen.The type of fluorescence staining was similar to that of the rat-SV40 cells. Everyfifth subculture of thehybrid

cells was investigated forTantigen; they remained

positive as faras tested, which includesthe 90th

subculture ofthelines H2 and H8 inbothmedia. The lines H3 and H10 have not been

systemati-cally investigatedfor T antigen beyond the 40th

subculture.

Agglutination. The results of agglutination by

concanavalin A with and without trypsin

treat-ment ofthe parental andhybiidcells areshown

inTable2.Without trypsin treatment,

agglutina-tion of 3T3 TK-cells occurred after addition of

25, 50,and 125 ,g of cancanavalin A. With 5

Ag,

noagglutination of 3T3 TK- cellswas observed;

at that concentration, 50% ofthe rat-SV40 and

25%ofthehybridcells hadagglutinatedafter 15

min.Withtrypsintreatment,anincreased

aggluti-nation of 3T3 TK- and hybrid cells was found. The degree ofagglutination of therat-SV40 cells

had not changed.

Growth in soft agar. Parental and

hybrid

cells

were tested for their ability to form colonies in

soft agar; the results are shownin Table 3. The

3T3 TK- cells formed no colonies at all; about

20%70

ofthetransformedcells gave risetocolonies

asdid 1 to 4% of the

hybrid

cells.

The reduction of theefficiencyofplating (EOP)

in soft agarseems specificsince the EOP in fluid medium of both rat-SV40 and hybrid cells was

approximately80c.

Serum requirement. The serum

requirement

of

thehybridlineswas

compared

tothatofthetwo

parental lines. Figure 4 contrasts the effect of

serum concentration on the growth rate of

rat-SV40 and 3T3 TK- cells. The transformed cells

grew at nearly the same rate in 10, 5, and

c,%o

serum concentration. No increase inthenumber

of 3T3 TK-cells was observed atthe1

%0

serum

concentration, and with a concentration of 5%

serum thegrowth rate was still lowerthan with

10%

serum. Figure 5 shows the response of the

hybridcellstoserumconcentration. Since all four

30 40 50 60 7r

CHROMOSOMES/METAPHASC

FIG. 3. Histogram of chromosomes/metaphase

againistnumberofmetaplhasesofrat-SV40cells.

VOL. 10, 1972 69

on November 10, 2019 by guest

http://jvi.asm.org/

(4)

TABLE 1. Number anddistribution ofchromosomes ofthe hybridcellsa Cell line Hybrid H2 Hybrid H3 Hybrid H8 Hybrid HIO Sub-culture level 12 24 62 87 11 23 39 72 11 26 70 89 12 19 26 42 Standard medium

Totalno.of chromosomes 96.4 (89-98) 88.9 (84-97) 80.4 (52-110) 83.9 (64-88) 85.3 (65-96) 75.3 (59-102) 69.6 (54-83) 74.3 (60-85) 95.8 (82-105) 87.0 (78-94) 73.9 (60-83) 78.0 (68-86) 92.2 (88-103) 90.4 (83-96) 89.4 (83-98) 75.3 (51-89) Telo-centric/ acro-centric 80.6 73.3 65.7 69.4 69.1 58.7 53.8 62.5 80.3 71.2 59.8 60.9 77.4 75.6 75.3 61.1 Meta-centric/ submeta-centric 15.7 13.6 12.7 13.8 15.6 15.5 14.5 10.3 15.3 14.2 13.8 16.3 14.4 13.5 12.9 13.3 HATmediumb

Totalno.of chromosomes NTc 90.5 (77-100) 83.0 (71-94) 82.3 (53-134) NT 78.0 (55-92) 68.8 (52-85) NT NT 88.2 (79-94) 77.0 (59-89) 74.7 (62-138) NT 88.5 (82-94) 86.9 (73-94) 78.1 (68-92) Telocentric! acrocentric NT 74.9 66.8 66.3 NT 60.1 51.8 NT NT 71.9 58.5 55.0 NT 74.6 74.2 64.3 Meta-centric,' submeta-centric NT 13.8 14.6 15.2 NT 16.7 16.0 NT NT 14.9 17.2 19.2 NT 12.9 11.6 12.9

aThenumber of chromosomesrepresents themeanvalue of 20metaphases.Thefiguresinparentheses

denote the rangeofchromosomal numbers. The difference between the sumoftelocentric/acrocentric

andmetacentric/submetacentric chromosomes and the total number is caused by chromosomes that

cannot be identified.

bSeeMaterials and Methods.

cNT = nottested.

TABLE 2. Agglutinationbyconcanavalin A ofparenital andhybridcells

Subculture

~~~~~~~~CoDcanavalin

Aadded(pug)

Cellline cl Trypsintreatment

5 25 50 125

3T3 TK- Without Oa 1+ 2+ 2-3±

Rat-S V40 42 Without 2+ 3-4+ 4+ 4+

HybridH3 38 Without 1+ 2± 3+ 4+

3T3 TK- With 0 3+ 3-4+ 4+

Rat-S V40 42 With 2+ 3-4+ 4+ 4+

Hybrid H3 38 With 2+ 3-4+ 4+ 4+

aU = N~o

agglutination;

agglutinated.

1+=

LY70/

01 the ceiis agglutinateul; 2-j- =

JUu/0;

35f = i;'4-1 =

IVUIJy-hybrid lines had about the same

dependence

on

serum,the results withonlyoneofthemaregiven.

Although the hybrid cells had a reduced growth

capacityin 1%serum,therestillwas anincrease in

the number of cells.

DISCUSSION

The results of the chromosomal studies ofthe

hybrid

cells show that over a

period

of about a

year

only

10 to 20% of the chromosomes has

been lost. The continuous presence of the three

enzymemarkers alsoindicates the

stability

of the

hybrids.

During this

period

of about a year, the cellsweresubculturedtwiceaweek.Thesmallloss of chromosomes isinagreement with thefindings

of Weiss and

Ephrussi (14).

In the

study

of

rat-mouse

hybrids,

5to

10%l

ofthechromosomeswas

lostover aperiodof 8months;thepercentloss of

rat marker (metacentric) chromosomes was

greater than thatofthe total number of

chromo-J. VIROL.

on November 10, 2019 by guest

http://jvi.asm.org/

(5)

PROPERTIES OF SOMATIC CELL HYBRIDS

TABLE 3. Efficiencyof colony formation in agar of

parental and hybrid cells

Cell lineCellline Subculture

~level

Coloncells-forming(%a

3T3 TK- 0

Rat-SV40 78 20.0

HybridH2 33 1.0

78 1.2

Hybrid H8 35 3.2

80 4.0

a Percentageis calculated from cell inocula of

2,000.

24 48 72

TIME IN HOUR5

FIG.4. The growth of3T3TK-and rat-S V40 cellsas

afinctionof theserumconcentration.

24 48

TIME 1N HOUR5

FIG. 5. Growth of hybridcells as afunction ofthe

serumconcentration.

somes, suggesting a preferential loss of rat

chromosomes. In our studies, no preferential

loss of metacentrics was observed, and most of

the chromosomal loss occurred among the

telo-centric and acrotelo-centric chromosomes. Further

subculturingof thehybridcellsisnecessaryto ob

tain further loss of chromosomes; this will per-haps be enhanced by changing the experimental conditions.

Because identification of chromosomes re-mains difficult, the use of more enzyme markers might be valuable for studying the loss of parental genomesfrom the hybridcells.

Allproperties of the SV40-transformed rat cells appear in the hybrid cells. The SV40 T antigen has been studied over manycell generations, and no loss of T antigen from the hybrid cells was observed. This is not surprising in view of the findings of Weiss (13) who studied the loss of Tantigen fromhybrids between mouse and SV40-transformed human cells. Cells became negative

for T antigen only after most of the human

chromosomeswerelost.Since inourstudy onlya

very limited loss of chromosomes has occurred,

the continuouspresenceofTantigen in the hybrid

cellsisnot unexpected.

During further subculturing, the hybrid cells

willbestudied for the presenceofTantigenand

thenumberofratchromosomes. The presence of

Tantigen indicates that the hybrid cellscontain

at least an early function of the viral genome.

Since no infectious virus couldbe recovered it is

not known whether the whole virus genome is

present.

Thefact that theSV40-transformed parentcells

did not yield infectious virus was considered

advantageous.The presence of infectiousvirusin

oursystemmight, after

superinfection

ofnegative

cells, lead to the induction of T

antigen

or to

transformation of themousecomponentof

hybrid

cells.

The altered structural organization of the

sur-facemembraneasdemonstrated by agglutination

with concanavalin A was found in

SV40-trans-formed cells (6). The degree of

agglutination

of

ourhybrid cellssuggests that thesurface

proper-ties of the transformedcells are atleastpartially

present in the hybrid cells. Trypsin treatmentof

thehybridcells resulted inanenhanced

agglutina-tion by concanavalin A. Ithas been shown that

trypsin treatmentofnormal cells leads to an

in-creased agglutinationbyglycoproteins (2, 6, 11).

Fromthis it seemsreasonabletosupposethat the

hybrid cells also have surface

properties

of

normal cells. It is not clear whether the surface

propertiesof thehybridcellsresult from

suppres-sion of the transformedphenotype

by

the normal

mousegenome.

In addition, wehavestudied the

growth

char-acteristics (5)ofthe

hybrid

cells in

comparison

to

those of the parental cells. All

growth

character-istics of the transformed cells were

expressed

in

the hybridcells,

although

toalesser

degree.

This

mightenableustodeterminewhethertheeventual

VOL. 10, 1972 71

-1 I

xz

I I

---l<

I --I

i

---4

I. II

C:

r

.9

t

9

"I

-1 --l

u 0 0.1 (t x

:1r-10% 5 %

I %

on November 10, 2019 by guest

http://jvi.asm.org/

(6)

changes of the transformed phenotype of the hybrid are related to loss of chromosomes and

enzymemarkers.Ithas been shownby Marinand

Littlefield (8) that one of the growth

character-istics of polyoma-transformed hamster cells, growth in soft agar, was reduced after loss of chromosomes. As far as studied, no changes in

growthcharacteristics have occurredinourhybrid

cells, which isnotunexpectedinview of the small

loss of chromosomes. Future studies will be

directed towardrelatingloss ofchromosomesand enzymemarkersto propertiesof thehybridcells.

ACKNOWLEDGMENTS

WeareindebtedtoIneHassink for excellent technical

assist-anceandtoMeeraKhan(Departmentof HumanGenetics,

Uni-versity ofLeiden) for his advice.

LITERATURE CITED

1. Basilico,C., and R.Wang.1971.Susceptibilityof superinfec-tion of hybridsbetweenpolyoma"transformed" BHK and

"normal" 3T3 cells. NatureN. Biol. 230:105-108.

2. Burger, M.M. 1969. Adifference in thearchitectureofthe

surface membraneofnormalandvirallytransformedcells.

Proc. Nat. Acad. Sci.U.S.A.62:944-1002.

3. Davidson, R. J., andB.Ephrussi. 1965.Aselectivesystemfor theisolationofhybridsbetween L cells and normal cells.

Nature(London) 205:1170-1171.

4. Defendi, V., B. Ephrussi, H. Koprowski,andM.C.Yoshida. 1967. Properties of hybrids between polyoma-transformed

and normalmousecells.Proc.Nat. Acad. Sci. U.S.A. 57: 299-306.

5. Eagle, H., G. E. Foley, H. Koprowski, H. Lazarus,E. M. Levine, and R. A. Adams. 1970. Growth characteristics of virus-transformed cells. J. Exp. Med.131:863-879. 6. Inbar,M.,andL. Sachs. 1969. Interaction of the

carbohydrate-binding protein concanavalin A with normal and

trans-formedcells. Proc. Nat. Acad. Sci.U.S.A. 63:1418-1426. 7. Littlefield,J. W. 1964.Selection of hybrids from matings of

fibroblasts in vitro and their presumed recombinants. Science145:709-711.

8. Marin, G., and J. W. Littlefield. 1968.Selectionof

morpho-logically normal cell lines from pclyoma-transformed BHK21/13 hamster fibroblasts. J. Virol. 2:69-78. 9. Meera Khan, P. 1971. Enzymeelectrophoresis in cellulose

acetate gel:zymogram patternsin man-mouseand

man-chinese hamster somatic cell hybrids. Arch. Biochem. Biophys. 145:470-483.

10. Neff, J.M., and J. F. Enders. 1968. Poliovirus replication and cytopathogenicity in monolayer hamster cell cultures fused with betapropiolactone-inactivated 'Sendai virus. Proc. Soc. Exp. Biol. Med. 127:260-268.

11. Ozanne, B., and J. Sambrook. 1971. Binding of radioactively labelled concanavalin A and wheat germ agglutinin to normal and virus-transformedcells. Nature N. Biol. 232: 156-161.

12. Pope, J. H., and W. P. Rowe. 1964. Detection of specific anti-geninSV40-transformed cells by immunofluorescence. J. Exp. Med. 120:121-123.

13. Weiss, M. C. 1970.Further studiesonloss ofT-antigen from somatic hybridsbetweenmousecellsandSV40-transformed human cells. Proc. Nat.Acad. Sci. U.S.A. 66:79-86. 14. Weiss, M. C., and B. Ephrussi. 1966. Studies of interspecific

(ratxmouse) somatic hybrids.I.Isolation,growth and

evo-lution ofthekaryotype. Genetics 54:1095-1109.

on November 10, 2019 by guest

http://jvi.asm.org/

Figure

FIG.1.3T3artificialnol6-phosphogltaconiate Zymograms of lactate dehydrogenase (A), dehydrogenlase(B), and indophe- oxidase (C) showing rat-SV40 (1), hybrid H3 (2), mixture of rat-SV40 and 3T3 TK- (3), and TK- (4) patte(rtis.
FIG.1.3T3artificialnol6-phosphogltaconiate Zymograms of lactate dehydrogenase (A), dehydrogenlase(B), and indophe- oxidase (C) showing rat-SV40 (1), hybrid H3 (2), mixture of rat-SV40 and 3T3 TK- (3), and TK- (4) patte(rtis. p.2
FIG.2.againistHistogramofclhromosomes/metaphase iiumber ofmetaphases of3T3 TK- cells.F
FIG.2.againistHistogramofclhromosomes/metaphase iiumber ofmetaphases of3T3 TK- cells.F p.3
TABLE 2. Agglutination by concanavalin A of parenital and hybrid cells

TABLE 2.

Agglutination by concanavalin A of parenital and hybrid cells p.4
TABLE 1. Number and distribution of chromosomes of the hybrid cellsa

TABLE 1.

Number and distribution of chromosomes of the hybrid cellsa p.4
FIG. 4.a finction The growth of3T3 TK- and rat-S V40 cells as of the serum concentration.
FIG. 4.a finction The growth of3T3 TK- and rat-S V40 cells as of the serum concentration. p.5
TABLE 3. Efficiency of colony formation in agar ofparental and hybrid cells

TABLE 3.

Efficiency of colony formation in agar ofparental and hybrid cells p.5

References