[CANCER RESEARCH 44. 5522-5531, December 19841

Effects of 1,2-Dimethylhydrazine on the Number of Epithelial Cells

Present in the Villi, Crypts, and Mitotic Pool along the Rat SmallIntestine1

G. G. Altmann2 and A. D. Snow

Department of Anatomy, The University of Western Ontario, London, Ontario, Canada N6A 5C1

ABSTRACT

Young adult male rats received 1,2-dimethylhydrazine (25 mg/kg) twice weekly for 2 months and once a week thereafter forup to 6 months. Histológica! samples of duodenum, jejunum,and upper, mid-, and terminal ileum were prepared from groupskilled at each month. Using cell counts, the average number ofepithelial cells was determined per representative section of villiand crypts and was used as an index of villus size and cryptsize, respectively. The average number of mitotic figures inrepresentative crypt sections was also determined.

All three parameters increased during treatment, and the increments showed a specific pattern in relation to time andintestinal region. Mitotic number showed a consistent change allalong the small intestine: close to 20% rise by 3 months; decrease to near control level by the fourth month; and a risethereafter. Probably, a systemic stimulation of mitotic activity by1,2-dimethylhydrazine took place. The crypt size index changedsimilarly, showing highly significant correlation with mitotic number. This correlation indicated that average mitotic time and cellcycle time remained unchanged and the number of divisionsincreased in progenitor cells. Calculations showed that only afraction of the progenitor cells was involved. These were probably "initiated" cells. There was a net increase of initiated cell

numbers with time, but a sharp drop at 4 months indicated thatthere is a mechanism inducing a regression of the initiated cellpopulation.

Villus size increased linearly in the duodenum and jejunum. Inthe ileum, there was also a net increase but with some initialfluctuation. In general, villus size seemed to increase so as tomaintain a fairly stable turnover time. This would mean that theincreased mitotic activity was balanced by increasing differentiation.

INTRODUCTION

Cell kinetics of the small intestinal epithelium under normaland experimental conditions has been studied by a large numberof investigators. However, the response of this rapidly renewingcell population to specific carcinogenic influences has receivedrelatively little attention thus far. The carcinogenic effect of DMH3

was first reported in 1967. DMH was found to be a rather specificcarcinogen for the epithelium of the large and small intestines(22). The morphological effect of DMH has been investigated

<Partially supported by the Medical Research Council of Canada and by the

National Cancer Institute of Canada.2To whom requests for reprints should be addressed, at the Department of

Anatomy, Health Sciences Centre, The University of Western Ontario, London,Ontario, Canada N6A5C1.

3The abbreviation used is: DMH, 1,2-dimethylhydrazine.

Received December 28. 1983; accepted August 28, 1984.

mainly with regard to the large intestine. (17,21, 27,28,36,37,48, 51), whereas only a few reports dealt with such effects onthe small intestine (46, 52). These latter reports showed that themitotic pool and the size of the crypts increased during DMHtreatment. The present study sets out to investigate the exactrelation of these changes to the time of the treatment, to intestinalregion, and to epithelial renewal. In this respect, we have investigated the rat duodenum recently and reported that the numberof epithelial cells in crypts and villi, the mitotic pool, and the sizeof the nucleoli in columnar cells increased in a progressivemanner in the duodenum during the initial 6 months of DMHcarcinogenesis. These changes appeared to be related to tumorformation since the increases were most pronounced in thetissue adjacent to the tumors (44). In the present investigation,this previous study was extended to include the various regionsof the small intestine: duodenum, jejunum, and upper, mid-, and

terminal ileum. The renewal of the normal epithelium in all theseregions has been studied previously (3, 8). In addition, specificlocal influences on the cell number in the villus epithelium havebeen demonstrated in these regions (2, 6, 9).

The parameters measured in the present work displayed patterns of increase which appeared to be specific to the time ofthe treatment as well as to the intestinal region. Combining thevarious data, some events of the carcinogenic process could bededuced.

MATERIALS AND METHODS

AnimalTechnique

Young male Wistar rats (Chartes River Canada, Inc., St.-Constant,

Quebec, Canada) approximately 4 weeks old and having an initial bodyweight of 100 to 120 g, were used in this study. The experimentalanimals were given s.c. injections with a solution of DMH (MathesonColeman and Bell Manufacturing Chemists, Norwood, OH) at a dosageof 25 mg DMH/kg of body weight. The DMH solution was prepared asfollows. 1,2-DMH (1000 mg) was dissolved in 200 ml of distilled water,

and 37 mg of EDTA were also added; the pH of this solution wasadjusted to 6.5 using 1 N sodium hydroxide (49). The control animalswere given s.c. injections of 0.9% NaCI solution containing EDTA (EDTA-

saline). Both experimental and control animals received injections twiceweekly during the initial 2 months and once a week thereafter for up to6 months. All animals were fed Purina laboratory chow and were allowedtap water ad libitum. They were housed in cages in groups of 4 to 6 andwere kept in a room maintained at a constant temperature (21 °)and at

a 12-hr day-night light cycle.

In preparation for the injections, the rats were anesthetized slightlywith methoxyflurane (Pitman-Moore), a mild inhalation anesthetic. The

animals were then weighed and given s.c. injections of the appropriateamount of 1,2-DMH (experimental group) or the EDTA-saline solution

(control group).Groups consisting of 4 to 6 DMH-treated rats were sacrificed monthly.

CANCER RESEARCH VOL. 44 DECEMBER 1984

5522

on March 4, 2021. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EFFECTS OF DMH ALONG RAT SMALL INTESTINE

The same number of control rats were also sacrificed at the same time.One week was allowed from the last OMH injection to the time of killingin order to avoid distortions arising from possible acute effects of DMHon cell proliferation (7). All animals were killed between 9 and 11 a.m. tominimize possible changes due to diurnal variation (1).

Histológica! Technique

All animals were killed by an overdose of chloroform. The smallintestine was immediately removed, pulled out to its full length, andmeasured from the pylorus to the ¡leocecaljunction. The distance of thevarious samples from the pylorus was expressed as the percentage ofthe total length. Thus, pieces of tissue 1 to 2 cm long were taken fromthe duodenum (0 to 1%), jejunum (25%), upper ileum (50%), midileum(75%), and terminal ileum (99 to 100%). The samples were cut openlongitudinally and flattened onto a piece of cardboard with the mucosa!surface facing upwards. The samples were then immersed into Camoy's

fixative for 5 to 6 hr. They were then placed in 70% alcohol, cleared inchloroform, and embedded in paraffin. The time between sacrifice andthe fixation of samples was kept no longer than 8 min/animai in order tominimize postmortem changes.

Longitudinally oriented sections of the intestine were cut at 5 i<m,stained by the periodic acid-Schiff technique, and couterstained with

hematoxylin.

Cell Counts

In the histological sections of the 5 regions of the small intestine, 3parameters were determined. They were: (a) the number of epithelialcells per representative villus section ("villus size index"); (b) the numberof epithelial cells per representative crypt section ("crypt size index"); (c)the number of mitotic cells per representative crypt section ("mitoticnumber"). Each of the representative structures used for cell counts was

initially located under low-power magnification (x 100), and the cells were

then counted under high power (X1000).Villus Size Index. Villus size index was measured as the mean number

of epithelial cells in representative villus sections. The number of cellswas determined by counting the number of nuclei, since only mononu-

cleated cells are found in the small intestinal epithelium (34). All epithelialcells were included in these counts, I.e., columnar, goblet, and enteroen-

docrine cells. The selection of representative sections was critical. First,only those areas of the sections were used that were cut perpendicularto the intestinal surface and parallel to the long axis of the intestine. Inthese areas, mostly longitudinally sectioned villi were present. The villiselected for counting were the tallest to ensure that the entire length ofthe villus was included in the section. These villi also had to show asingle layer of epithelial nuclei. All nuclei within the epithelium werecounted. Sections of nuclei less than one-fourth of the average size of

the nucleus were not included in the counts. The counts were madeusing a hand tally counter after identifying each visible nucleus within thewhole thickness of the section by focusing up and down. The countingwas started at the crypt villus junction, which was recognized as thelowermost point of the intervillus spaces.

The minimum number of villi to be measured in each small intestinalregion (per animal) was assessed as that sample which yields a standarderror of less than 5% of the mean. The variation in villus size was foundto be quite small, so that the minimal necessary number was usually 4to 6. Nevertheless, 10 representative villi were measured per section,the results from all animals in the group were averaged, and the standarderror of the mean was calculated.

Crypt Size Index. Crypt size index was measured as the mean numberof epithelial cells (nuclei) in longitudinally cut crypt sections. All epithelialcell types were included in these counts, i.e., columnar (including mitoticones), goblet, enteroendocrine, and Paneth. A crypt section was regarded as representative when the plane of section passed along thelong axis of the crypt. Such crypts were found to extend from themuscularis mucosae to the base of the villi and consisted of a single

layer of epithelium on each side of the lumen.The minimum number of crypts required for counting in each small

intestinal section (per animal) was estimated at 4 to 6. A safe number,about 10 representative crypts, was therefore used for the counts persection. The results for each section were then averaged, and thestandard error of the mean was calculated.

Mitotic Number. Mitotic number denotes the mean number of mitoticfigures per representative crypt section. Each mitotic figure from pro-phase to telophase was counted, but early prophases and late telo-phases (45) were not included. Dividing cells were distinguishable fromnondividing cells mainly because of their chromosome aggregationswhich were usually closer to the crypt lumen than were the nonmitoticnuclei. The crypts were selected as above but, since larger numbers hadto be used, additional crypts sectioned near the lumen only in their lowertwo-thirds were also used. This lower part of the crypt is known to

contain all the mitoses. The number of mitoses was quite variable fromcrypt to crypt, and the minimum number of crypts to be measured wascalculated to be 10 to 15. In each section, a safe number, 25 to 30crypts, was measured. Average values for each animal and then for eachgroup were calculated along with the S.E.

RESULTS

Villus Size Index. The villus size index in both control andexperimental groups was always the largest in the duodenum. Itgradually decreased in the subsequent regions, reaching a minimum in the terminal ileum (Table 1). This phenomenon in normalanimals has been referred to previously as the villus size gradient(9). In the 0-month control animals, this gradient involved a

decrease from 272.0 ±1.8 (S.E.) in the duodenum to 133.0 ±0.6 in the terminal ileum. The values were similar in all the controlgroups during the 6-month period (Table 2). The villus size

became elevated during the treatment, and at the same time thegradient became steeper. The elevations were thus most pronounced in the proximal intestine. The villus size gradient observed at each month of treatment could be expressed as asignificant straight line by simple regression analysis (Table 2).The slope increased gradually in the treated groups from -1.408at 1 month to -1.928 at 6 months, whereas that of the controlgroups remained around -1.370 at all times of the treatment.The y-intercept (i.e., maximal villus size) also increased progres

sively in the treated groups, from 288.7 at 1 month to 334.5after 6 months, whereas in the control groups the average y-¡nterceptremained at 269.7. The minimal villus size did not showsignificant variation with time, being on the average 140.2 in thetreated groups and 132.6 in the control groups. The increase inthe slope (S) of the gradient with time (T) was also linear andcould be expressed as

S = 1.289 + 0.1037 (r = 0.916)

Another way to evaluate the effect of DMH on the villus sizeindex would be to analyze the increments of the villus size indexabove the control values (Table 1; Chart 1). In the duodenumand the jejunum (Chart 1), the increments (/) were found toincrease linearly with time:

/ = 4.667 + 10.857 for the duodenum (r = 0.911)

/ = 0.913 + 6.637 for the jejunum (r = 0.976)

In the ileal samples, the increase was not linear inasmuch asthere was a fluctuation in the increments; nevertheless, the net

CANCER RESEARCH VOL. 44 DECEMBER 1984

5523

on March 4, 2021. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EFFECTS OF DMH ALONG RAT SMALL INTESTINE

CO

aT-

o

S

l

II

i.2

co?<bCM•Hr»5CO

00öö•H-Hocn88CD0}fcr-I•Hco1h-

CMÖCM•H

-HCMCMgp:co•o10•HCM

CM•r^

Ö•H

-Ho>*SS**•o>•HÕin

oÖCM-H

+1(O00co

coCMCM«?

8"o.IO+1^I*CO

05r-

t—•H

+1ocoÕMÈMO

i-o>

r- coi-»° £8e*

o> tuinuiin CMco•H-H +1-H<°.

,^"1 _°P"içgOTCMMggCO

CO COCMCMCM WW•H-H -H-Ht--r» ocn8*"

*" Sco cocoCO

I-.CM0* ">¿CM

CO•»CM ^°îi-

i- coin•H-H +1+1COCO CM •*

çôwgOTCOWOJin

in enineviCD öin+1+i -H+tcom inCMCÓW CMCO•«

; * COp«N

i «92a>v- eiv-•H-H -H-Htr- i- CDgéô'cMtô'gco'j:^

CM O*>»T^ CO^•H

-H -H-HpCO 0)CDO)

CT3 O)O>co

•«•T-cos

s sis^CO 0 V

W CÓ "? *»- CM CO«-+1

-H -H-H)00 ^ COCO-_"

C\J --.^ CM ^^CMCMco

to ooíCOCO CMCM•H-H -H+1inco CMin§

CM §C^T-

CM oina

a ssO>

CM 1^CM+)

-H -H-H^co ino)2?^ c/) escoö> o ~f v-5CMCOfcCOCOco

in CMCM•*'•* WCM+1

+1 -H-Hpi-; in^^

in CMiriCMCM CMCMCM

CO * mh»tT»^•HIs:ÙÔK^•HocSCD?CO•Hop'S^Õ—+11cq«>T-•HliCÔ41coeniv+1co&«?•HT-%a>

P!N!+iinÕCDCM•HCD£CD

change was an increase (Chart 1). Only after the fourth monthdid a trend toward a linear increase seem to prevail.

Crypt Size Index. In controls, the crypt size index averaged97.4 in all regions of the small intestine. In the DMH-treated

groups, the crypt size index rose above the control values in allregions except at 4 months when it was equivalent to the controlvalues (Table 3).

In the duodenum and the jejunum, the increments increasedprogressively during the first 3 months and then dropped to aminimum at 4 months. This was followed by a second progressive rise up to 6 months (Table 3; Chart 1).

The ileal samples showed a somewhat different pattern. Theminimum at 4 months and the subsequent trend of progressiverise were present as in the duodenum and jejunum. In the initial3 months, however, there was some fluctuation. Only after 4months did a trend of progressive increase prevail, especially inthe upper ileum. In the mid- and terminal ileum, the fluctuations

were still present but at a much reduced amplitude.Mitotic Number. The mitotic number, i.e., the number of

mitoses per crypt section (Table 4), in control groups showed nochange with time during the 6-month period. A slight regional

change along the small intestine was, however, present at alltimes: 6.0, 5.1, 5.0, 5.0, and 4.9, on the average, in the 5respective intestinal regions. Under DMH treatment, all regionsshowed an increase in mitotic number, except in the fourth monthwhen little or no increase was observed (Chart 1).

The pattern and magnitude of the increments were about thesame in all the 5 intestinal regions (Table 4). The incrementsincreased progressively up to 3 months and dropped to zero ornearly zero at 4 months. This was followed by a second progressive increase up to 6 months (Chart 1).

Comparison of the Increments of the 3 Parameters. Themost consistent pattern, similar in all 5 intestinal regions, wasdisplayed by the increments in mitotic number (Chart 1). Thispattern was characterized by a progressive increase up to amaximum at 3 months, a decrease to a minimum at the fourthmonth, and a progressive increase thereafter. All crypt size indexincrements followed this pattern. However, in the ileal samples,there was an initial fluctuation. This was also seen in the ilealvillus size increments. The villus size increments in the duodenumand jejunum followed a steadily progressing pattern and therefore appeared different from the pattern of all the other increments. A statistical analysis of these patterns is given in "Appendix A."

DISCUSSION

Indications from the Pattern of Increments. All parametersincreased in a special manner, and all the increments showed anet increase with the time of the treatment. The pattern ofincrease, however, was different for each of the 3 parametersused, and this pattern was also different in the various intestinalregions.

The increments of the mitotic number were similar in all intestinal regions for pattern as well as for magnitude (Chart 1). It islikely, therefore, that the mitosis-stimulating effect of DMH was

systemic. The previous belief was that the effective metabolitesof DMH reached the bowel through the bile from the liver. Thisview has, however, been challenged as systemic influences ofthe carcinogen have been reported (e.g., Refs. 16 and 53).

CANCER RESEARCH VOL. 44 DECEMBER 1984

5524

on March 4, 2021. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

on March 4, 2021. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EFFECTS OF DMH ALONG RAT SMALL INTESTINE

tp pr^ CM

•H +)

en o ;

00 0

-Hco

S

coco

o> * sr¿ m ^•H-H -H

« iv. en^ W ri T^ ö ö o•H-H +( -H -H -H -H»•^^ cn ^ o> rv cn

ei ö i-cn o) cncn

* * °?oí in h-!

co+1

CM N.W eó

T»+1CD

enri

CM

+1

.*- fc.cn fc.

oo cn co *™^ co *~^ inö<¿ ri W W CM ö•H-H +( -H -H -H -Hcp cn co OD p cn ^tS en ri ¿ CM ¿ rico oo cn en cn co cn

CM eori ri

eo »-TI TI

O CD TJ; Uf)W i-^ ri T-+1+1 +1 +1CMen

en cqri ^

Ö•H

CMci

T- cqri i^

CM T-

-H -H

CM <pW ci+< +1cq T-

cn cn en en

|v. CD

Si 800oi

•H -H«-;up

m CDri CM

1 ^^CM ^00+1O

. . . . . .W CM ri i- Ö Ö *-•H-H -H -H -H -H +1*r t co •* t oo rv.

cn en en en

en p N. co in oqin Ivi OD Tt W 00

_ „ ^ * î- •o cq en r~ CD »~ri •*' * ö W evi

+1 +1 +1 +1 +1 +1

co co co cn r^ CM coCM CM ¿ CO CM CO Tt

+1-H^ inoo »v

in cn T^ iv. iv.ö ö ö ö ö+1 +1 +1 +1 +1co cq in en CMW ó CM ó TtCM CM CM CM CM

COÖ

^ OO r-"

~ V 'CM

ooo

coIO

CO COO O

ooo

Wcxj£.10

00 O+1+1 -Hco co cn•*•»'•»

T PÖ T¿cnö

CM r-w i-

Ö Ö O•H -H +1cq cn p«f ^f vi

o o+)_+!evi Wpin 5-co

o•HenLO

r-_*" mif" inuri fc.in è-io

00 O+1+1 -H°PT p

LO LO

unÖ

O•Ho

o•Hoai

o+1en

co oao ö ö

T- CM U) CM CMo o ^ o o

in <p T^P -S*1?!u> in ivi ^ LO¿.uri:

oo o o o o•H-H +1 +1 -H -H

P T T T P Pin iñ io in in in

01

O

i-; ÃŽM |v. r-_ CM

Ö Ö Ö Ö Ö

TS.LO tD 5, IO 5. IO 5- LO

00 O O O O•H-H -H -H -H +1*". T 'S T T PLO LO LO 1/Ì LO LÒ

* coö T-:

00o

^ CMÖ Ö Ö Ö Ö Ö

io ^ V 'CM

ö ö> W'CM 01 <f>

i S.I-' OS S.

00 O+1+1 -Hco T^ cnLO CO LO

O O O O+1 +) -H -Hen CM T- CMLO CO CO CO îitll

v v

CANCER RESEARCH VOL. AA DECEMBER 1984

5526

on March 4, 2021. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EFFECTS OF DMH ALONG RAT SMALL INTESTINE

during the DMH treatment. This would mean that an expansionof the number of cells in mitosis would be associated with asimilar expansion of the number of cells in the other phases ofthe cell cycle. There is evidence from other reports that theaverage cell cycle time remains essentially unchanged duringDMH treatment in the small intestine (46), in the large intestine(39, 49), or even in the tumors arising (11). On the other hand,the crypt base cells may have differing cell cycle times (14,18),but this is probably not enough to affect the average value.

The "decision" to differentiate takes place in the upper crypt

region, and after this the cells cease to cycle (38). These cellsare contained approximately in the upper third of the crypt whichmay thus be called the differentiating zone. The rest of the cryptmay be considered as the cycling zone. This zone may containa few differentiating cells (e.g., Paneth, goblet, enteroendocrine),but the total number of these is negligible to the number ofcycling columnar cells for kinetic considerations (18). The average cell cycle time should be around 11.5 hr (20), and the averagemitotic time should be around 1 hr (18,29). It follows that mitoticnumber multiplied by 11.5 gives the approximate number of cellsin the cycling crypt zone, considering, in our case, a representative crypt section. Subtracting this number from the crypt sizeindex would give the approximate number of cells in the differentiating crypt zone.

When examining the increments in the crypt size indices (Table3; Chart 1), an important consideration is how much contributionwas made by the cycling and the differentiating crypt zones tothese increments. In other words, was the stimulus to proliferatemet by a proportional stimulus to differentiate? If the incrementsof mitotic number (Table 4) are multiplied by 11.5, the approximate number of cycling cells added to the crypt may be obtained.The difference between this number and the crypt size increment(Table 3) would indicate the increment in the differentiating cryptzone. After this calculation was completed, a distinct differencewas found between the duodenojejunal samples and the ilealsamples. In both, the cycling crypt zone changed according tothe pattern described for the increments of mitotic number (Chart1). In the duodenojejunal samples, the differentiating zone increased steadily during the 6-month period of treatment. On the

other hand, there was either no increase or a negative increaseat 2,3, and 4 months in the ileal samples. This is reflected in thefluctuation of the crypt size increments during the first 4 monthsin the ileal samples (Chart 1). This initial fluctuation was notpresent in the crypt size increments of the duodenojejunal samples (Chart 1). It thus appears that some local mechanisminterfered with the differentiation in the ileal samples during thefirst few months of the DMH treatment.

The differentiating crypt zone is probably maintained by afeedback from the mature villus cells (26, 50), possibly by achemical mediator, the "chalone" (13, 43). In our laboratory, we

have observed that the size of the mitosis-free upper crypt

portion was dependent on the degree of development of the villiin experiments in which the villi were damaged by methotrexateand were allowed to recover thereafter (5). Also, it has beenshown previously (2, 4, 9) that the villus size index is partiallyunder the control of factors contained in the intestinal chyme. Inthe upper small intestine, villus-enlarging factors predominate,whereas the ileal chyme contains mainly villus-reducing factors.An increase in the amount of villus-enlarging factors during DMH

treatment would facilitate villus growth and thereby the differ

entiation of crypt cells in the duodenum and the jejunum. On theother hand, increases in the villus-reducing factors may interfere

with the ileal villi and their effect on crypt cell differentiation. Inany case, the net change in the ileal villi was an increase so thatvillus growth was only temporarily interfered with. While thechanges in the villus size-controlling factors during DMH treat

ment are beyond the scope of the present work, it appearsprobable that the almost linear growth of villi in the duodenumand the jejunum during DMH treatment (Chart 1) enhanced acontinuous growth in the differentiating zone of the crypts.Inversely, the fluctuating growth pattern of the villi in the ileumprobably caused some temporary interference with the growthof the differentiating crypt zone. In any case, it seems certainthat, aside from the probable systemic influence of DMH onmitotic activity, some local effects also occur which may modifythe resulting tissue architecture and may have an influence onthe eventual appearance of neoplasm.

Probable Mechanism of Action of DMH. An increase in thenumber of mitotic cells recorded may result from a decreasedcell cycle time, an increased mitotic time, or an increased numberof cells entering the cell cycle.

Since there was a strong correlation between crypt size indexand mitotic number in practically all samples ("Appendix B"), it is

probable that the mitotic stimulus involved mainly an increase inthe number of cycling cells, whereas no major change occurredin either mitotic time or cell cycle time. In this, only the columnarepithelial cells are considered, because they constitute morethan 90% of the epithelium (18).

The cycling crypt zone contains mainly the stem and theprogenitor cells. These cells establish the cycling zone afterrepeated mitotic divisions, the total number of which for eachcell is believed to be 3 on the average (33). An increase in thenumber of cycling cells may come either from an increase in thenumber of stem cells by the conversion of resting (G0)stem cellsinto cycling ones (31) or from an increase in the number ofdivisions that the progenitor cells go through before differentiation. The former possibility is unlikely because resting stem cellsare probably not present in the small intestine (19).

Considering now a representative crypt section, s number ofstem cells would be present, and these would go through nnumber of divisions, on the average, to reach a number of cellswhich would then constitute the cycling zone of the crypts. Thefirst division would double s, the second division would quadrupleit, and so on. It follows that

s2" (A)

where a is the number of cells present in the cycling crypt zone.But a can also be calculated by multiplying the mitotic number(M) with the cell cycle time (7).

S2" (B)

for controls. For the treated animals, there will be an increase (/)in n, and M changes to M':

S2"**= TM'

Dividing now Equation C with Equation B,

2^ = AT2" ~ M

(C)

(D)

CANCER RESEARCH VOL. 44 DECEMBER 1984

5527

on March 4, 2021. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EFFECTS OF DMH ALONG RAT SMALL INTESTINE

Taking the logarithm of both sides,

Rearranging and substituting 0.30 for log 2:

0.30/ = tog Aj

aría

/ = 3.333 tog ^-

(E)

(F)

This equation provides a way to estimate the increase (/) in thenumber of cycles. This increase, as shown in Table 5, variedbetween 0.02 and 0.47 and was apparently not dependent toany great degree on the intestinal region but rather on the timeof the treatment, being smallest at 1 and 4 months and highestat 3 and 6 months. This would imply that during the treatmentnot only stimulation but also a repression of the previouslyincreased mitotic activity took place.

If / were equal to 1.0, it would mean that each progenitor cellgoes through one extra division before differentiation begins.Since / was consistently below 1.0, only a fraction of the progenitor cells went through one additional cycle. This fraction isprobably identical to the so-called "initiated" cells, which are

known to appear in the early stages of carcinogenesis (24). Thepresent data provide for the possibility of estimating the approximate number (IV) of these cells from the difference between thecycling crypt portion in control and treated rats:

W = T(M' - M)

The values for W - M are given in Table 4, the value of T isassumed to be constant. The variation of M' - M would thus

reflect the variation of the number of initiated cells. This valuecorrelates well with the increments of crypt size values becausethe variation with the time of treatment was the same in the 2sets of values. In general, since mitotic number and crypt sizeindex were well correlated, it may be inferred that most of thechanges observed in crypt size came from changes in ; as wellas from the corresponding changes in w. The most likely possibility then is that the number of divisions increased by only one(on the average) in a fraction of progenitor cells and that the size

TablesIncreases in the number of cycles

Usinga relativelysimplemathematicalmanipulation,the amountof increaseinthe numberof cyclesafter DMH was calculated.AHthe valuesare below1, whichprobably means that only a fraction of the crypt cell population entered an additionalcycle before differentiation into nonmitotic absorptive cells. The highest increasestook place in the third month, whereas the lowest ones occurred in the fourthmonth of the DMH treatment.

Time ofDMH treatment(mo)1

23

456Cycle

increaseDuodenum0.0236

0.09490.288300.17830.1756Jejunum0.0556

0.18290.350900.11130.1639Upperiteum0.1093

0.13530.478900.08430.1639Mid-

ileum0.0279

0.02630.23960.02860.11360.1379Terminal

iteum0.0859

0.11360.27390.05890.08590.0843Av.0.0605

0.11060.32630.01750.11470.1451

of this fraction fluctuated with time. In other words, the numberof "initiated" cells would be involved in marked time-related

fluctuation, implying that certain factors could increase and others could decrease their number. The cumulative effect on mitoticactivity seen between 1 and 3 months and between 4 and 6months of DMH treatment was probably due then to a gradualincrease in the number of initiated cells. At 4 months of treatment,no significant mitotic stimulus was observed, which may havebeen due to a mechanism which counteracted DMH, either byremoving the stimulus or by eliminating the initiated cells. Theapproximate percentages of initiated cells in the crypts were 3,7,15,1,5, and 7 in the subsequent months of treatment. Thetime-dependent variation was thus large, whereas the regional

variation along the intestine was relatively small.Renewal of the Crypt Epithelium. If a cells are in the cycle

per representative crypt section, a number of new cells will beproduced during the elapse of each cell cycle time, i.e., during11.5 hr. To keep the steady state, the same number of cells willhave to enter the villus epithelium. This will thus determine therate of flow of cells through the epithelium. Concerning thedifferentiating zone of the crypts, if a cells are moved within 11.5hr, C - a cells will be moved within [11.5 (C - a)]/a hr where C

is the crypt size index. The formula can be rewritten

On the other hand, c/a turned out to be a fairly constant number(k) with values of 1.56 in duodenum, 1.47 in jejunum, and 1.42in ileum in all control and treated samples. The overall formulathen for the transit time of the differentiating zone is

ta= 11.5Õ./C-1) (G)

Using this formula, 6.44,5.4,4.60,4.83, and 4.83 hr are obtainedfor the 5 respective intestinal regions. This holds for control aswell as treated animals because the value of k does not changewith treatment. The total crypt transit time is obtained by adding11.5 to td. The transit times thus ranged between 16 and 18 hr.Another technique yielded 24 hr on the average for the cryptturnover time in normal adult rats (8). Transit times determinedby the radioautographic technique were, however, close to thepresent findings (14, 18). In general, transit times are shorterthan turnover times because they indicate the travel time of themost typical migrating cells. Turnover times, on the other hand,include slow cells as well as nonmigrating cells such as, e.g., thePaneth cells.

The fact that DMH did not increase the crypt transit time maybe explained by the fact that the differentiating zone tended tobe a constant fraction of the crypt. As the cycling portionincreased, there was also an increase in the differentiating portion. On the other hand, at 2, 3, and 5 months in the ¡lealsamples, there must have been some slight deviation from theexpected crypt transit times as the differentiating zone wasreduced. This reduction, however, may be negligible when compared to the whole cell number of the crypt.

Renewal of the Villus Epithelium. The-rate of flow from thecrypts to the villi would be a for every 11.5 hr from eachrepresentative crypt section. In a section, however, on the average, 1.8 to 2.0 crypts feed a villus (3). This crypt frequencyper villus (0 was found not to change during DMH treatment

CANCER RESEARCH VOL. 44 DECEMBER 1984

5528

on March 4, 2021. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EFFECTS OF DMH ALONG RAT SMALL INTESTINE

(44). If fa cells move to each representative villus section per11.5 hr, V cells (V is the villus size index) will renew within11.5(V/fe) hr, which is then the transit time of the villus epithelium(tv). Sincea = 11.5M,

(H)

Assuming that f is constant, V/M may be used as provisionalturnover time for comparative purposes (3). For actual turnovertime, a correction factor may be calculated by dividing turnovertimes, obtained by another technique, with the provisional turnover times (3). The turnover times shown in Table 6 wereobtained in this manner. It can be seen that the villus turnovertimes did not decrease as the mitotic activity increased. Thevillus size index increased so that it compensated or evenovercompensated for the increase of mitotic activity. As a result,villus turnover times did not change or became slightly elongated.If, then, tv tends to remain reasonably constant, the value of V isessentially determined by the variation in M as can be seen fromEquation H. The DMH-induced changes in villus size index and

in mitotic number should therefore be interrelated.General Conclusions. There was a tendency of increase in

all the 3 parameters measured, in every intestinal region underthe influence of DMH. After analyzing the pattern of increases,several additional aspects of the carcinogenic process wererevealed.

Since the pattern and magnitude of the increments of mitoticnumber were similar in each intestinal region, a systemic effectof DMH was thought to be involved. It appeared that thisstimulation of mitotic activity occurred only on a select numberof cells which may be referred to as "initiated." These cells

probably increased their number of cycles by one, on the average, before differentiating. The results indicated furthermore thatthere was a considerable variation in the number of initiated cellsin a definite pattern which in turn implied that the cumulativeeffect of the carcinogen was basically an increase in the numberof these cells. It also indicated that the animal was able at timesto counteract this increase and to reduce the number of thesecells. Occasional as yet unexplained reductions in the number ofinitiated cells during carcinogenesis have been reported, forexample, in skin and liver (23, 30,32, 41,42).

The increase in crypt size and villus size differed with intestinalregion and were therefore influenced by local factors. The picturewhich is emerging is that a primary stimulus of mitotic activitybrings about a corresponding change in crypt size index mainly

by the increase of the zone of cycling cells within the crypts. Thesteady change in the differentiating zone would represent thecompensatory reaction of the villi which after their increase insize probably increase their "negative feedback" on the crypts.

This negative feedback may be a chalone which would thusexert a balancing influence so that the increased mitotic activityis met by increased differentiative activity. As a result, the steadystate, including villus and crypt turnover time, would not essentially be altered.

The concept of villus-enlarging and -reducing factors of thechyme which was postulated over 10 years ago proved to beuseful at this point. If the activity of villus-enlarging factorsincreases during carcinogenesis, this would enhance the adaptive response of the villi. On the other hand, if the activity ofvillus-reducing factors also increases, this would interfere with

the adaptive response in the ileum. The total effect of theincrease in the 2-factor systems would be the steepening of the

villus size gradient during treatment which was indeed the observation. The ¡lealvillus size increments all showed some initialfluctuation, presumably due to villus-reducing factors. The netchange of the villi, however, in the ileum was an increase whichseems to indicate that the adaptive increase overrid the activityof the villus-reducing factors.

The need for detailed "sequential analysis" of chemical carcin

ogenesis has been emphasized (25). A portion of our numericalresults was reported earlier concerning the duodenum only (44).This previous as well as the present work may be regarded asa form of sequential analysis using the method of morphometry.The results commensurate with the current concept that carcinogenesis produce an initiated cell population. The time periodcovered by the present study probably corresponds to Phase Iof carcinogenesis described for the colon (35). In the smallintestine, Phase I seems to be mainly concerned with the establishment, maintenance, and an eventual increase of the initiatedcell population. These events seem to take place all along thesmall intestine and not localized to specific regions. The initiatedcells as expected are able to differentiate and to respond to thepresumed negative feedback or chalone originating from the villi.As a result, renewal seems to proceed at a normal rate. Thefurther events leading to carcinoma formation are not known.Whether they involve some disturbance of feedback and renewal,perhaps only in localized areas, or the appearance of cells whichdo not respond well to the feedback will have to be decided byfurther research.

APPENDIX A

To analyze the time-related patterns of the increments of the 3 parameters, theprocedure for comparing polynomial regression models was used (54). The essence

Tabte6Villus turnover times (hr)

The turnover times for the villus epithelium are given for control and DMH-treated rats. The so-called 'provisional turnover times* were first obtained by dividing villus

size index (V) with mitotic number (M) according to a method described in an earlier work (3). These V/M values were then multiplied by a constant, i.e., 0.85,0.65,0.68,0.61,0.56 for duodenum, jejunum, and upper, mid-, and terminal ileum, respectively. These constants were obtained by dividing known turnover time values (8) with theV/M values. In this manner, values dose to actual turnover times were obtained. These are presented here in order to show that turnover time remained unchanged orbecame slightly elongated during DMH treatment.

DuodenumJejunumUpperileumMidileum

Terminal ileumOM40.0

29.426.921.915.31Control38.5

29.526.824.115.4MTreated41.2

29.426.822.716.12MControl39.6

28.726.721.515.2Treated40.5

26.924.317.915.83MControl39.8

29.026.921.215.4Treated35.6

25.120.419.613.84MControl37.4

29.127.321.315.5Treated42.7

31.928.121.616.85MControl38.5

30.327.222.717.4Treated42.5

32.228.221.016.36MControl37.9

30.027.221.115.1Treated42.5

31.728.520.816.1

CANCER RESEARCH VOL. 44 DECEMBER 1984

5529

on March 4, 2021. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EFFECTS OF DMH ALONG RAT SMALL INTESTINE

of the method is to find the best equation expressing the pattern and to see howsignificantlythe various patterns differ from the generalizedone. Whenconsideringthe changes between 0 and 6 months, only 2 significant patterns could beestablished, Pattern 1 (rise, decrease, and rise) and Pattern 3 (linear rise). Thevilius size changes in duodenum and jejunum were dearly following Pattern 3. Inthe combined samplesof the iieum. this analysis also favored Pattern 3, this beingthe more significant.

For crypt size index and mitotic index, Pattern 1 was most significant in theduodenum and jejunum as well as in the ileum. This method was not sensitiveenough to account for the fluctuations in crypt size index in ileum during the first 3months. The method was therefore used again but for only the first 3 months.

This latter analysis accounted for a linearly increasing pattern and a fluctuatingpattern, i.e.. a rise in the first month, a decrease in the second month, and a risein the third month. The former pattern was significant for all mitotic number dataas expected. A steeper rise fitted the vilius size index data in duodenum andjejunum. This latter pattern tended to prevail for the vilius size indices of the ileurn.but here the other, i.e., the fluctuating pattern, was also at the borderline ofsignificance.

The crypt size indices for the first 3 months showed the linearly increasingpattern for the duodenumand jejunum but were highlysignificant for the fluctuatingpattern seen in the combined ilealsamples.

APPENDIX B

The correlation coefficients for crypt size index and mitotic number were 0.89in duodenum, 0.91 in jejunum, 0.82 in upper ileum, 0.63 in midileum,and 0.74 interminal ileum. The number of observations in each of these 5 groups rangedbetween 51 and 57. Thep value in each group was 10 5or less, indicatinga highlysignificant positive correlation between the 2 variables. The minimal level of significance is usually taken at p = 0.05.

REFERENCES

1. AI-Dewachi,H. B , Wright. N. A., Appleton, D. R., and Watson, A. J. Studieson the mechanismof diurnal variationof proliferative indices in the small bowelmucosa of the rat. Cell Tissue Kinet., 9:459-467,1976.

2. Altmann, G. G. Influence of bile and pancreatic secretions on the size of thevaiiof the small intestine in the rat. Am. J. Anat, 732:167-178,1971.

3. Altmann, G. G. Influence of starvation and refeeding on mucosa! size andepithelial renewal in the rat small intestine. Am. J. Anat., 733:391-400,1972.

4. Altmann, G. G. Demonstration of a morphological control mechanism in thesmall intestine. Role of pancreatic secretions and bile. In: R. Dowling and E.Richken, (eds.). Intestinal Adaptation, pp. 75-86,1974.

5. Altmann, G. G. Changes in the mucosa of the small intestine following meth-otrexate administration or abdominal irradiation. Am. J. Anat., 140:263-2180,1974.

6. Altmann, G. G. Factors involved in the differentiation of the epithelial cells inthe adult rat small intestine, in: A. B. Cairnie. P. K. Lala, and O. G. Osmond(eds.), Stem Cells of Renewing Cell Populations, pp. 51-65. New York:Academic Press, Inc., 1976.

7. Altmann, G. G. Evidence for a chalone-type effect on the mitotic activity ofintestinal crypts after recovery from methotrexate. in: Proceedings of theCanadian Federation of Biological Sciences, p. 142,1978.

8. Altmann, G. G., and Enesco, M. Cell number as a measure of distribution andrenewal of epithelial cells in the small intestine of growing and adult rats. Am.J. Anat., 721: 319-335,1967.

9. Altmann, G. G., and Leblond. C. P. Factors influencing vilius size in the smallintestine of adult rats as revealed by transposition of intestinal segments. Am.J.Anat.. 727:15-36,1970.

10. Argyris, T. S. The regulation of epidermal hyperplastic growth. Crit. Rev.Toxfcol., 151-197.1981.

11. Baserga, R. Relationshipof the cell cycle to tumor growth and control of celldivision: a review. Cancer Res., 25: 581-595,1965.

12. Baserga, R., and Kisieleski. W. E. Comparative studies of the kinetics ofcellular proliferation of normal and tumorous tissues with the use of tritiatedthymidine. I. Dilution of the label and migration of labeledcells. J. Nati. CancerInst., 28: 331-332,1962.

13. Brugal. G. Presenceof intestinal chalones. In: A. B. Cairnie,P. K. Lala, and D.P. Osmond (eds.).Stem Cells of Renewing Ce«Populations, pp. 41-50. NewYork: Academic Press, Inc., 1976.

14. Cairnie,A. B., Lamerton, L. F., and Steel, G. G. Cell proliferation studies in theintestinal eiptheliumof the rat. 1. Determinationof the kinetic parameters.Exp.Cell Res., 39: 528-538,1965.

15. Cameron, R., and Farber, E. Some conclusions derived from a liver model forcarcinogenesis.Nati. Cancer Inst. Monogr., 58:49-53,1981.

16. Campbell, R. L. Singh. D. V., and Nigro. N. D. Importanceof the fecal streamon the induction of colonie tumors by azoxymetnane in rats. Cancer Res., 35:1369-1371,1975.

17. Chang, W. W. Histogenesis of symmetrical 1,2-dimethylhydrazine-inducedneoplasms of the colon in the mouse. J. Nati. Cancer Inst., 60: 1405-1418,

1978.18. Cheng, H., and Leblond. C. P. Origin, differentiation and renewal of the four

main epithelial cell types in the mouse small intestine. I. Columnarcell. Am. J.Anat., 747: 461-480,1974.

19. Cheng, H., and Leblond, C. P. Origin, differentiation and renewal of the fourmain epithelial cell types in the mouse small intestine. V. Unitariantheory ofthe origin of the four epithelial cell types. Am. J. Anat., 717: 537-562,1974.

20. Cowdry, E. V., and Paletta, F. X. Changes in cellular, nuclear and nucleolarsizes during methylcholanthrene epidermal carcinogenesis. J. Nati. CancerInst, 7:745-759,1941.

21. Deschner, E. E. Experimentally induced cancer of the colon. Cancer (Phila.),34:824-828,1974.

22. Druckrey, H. Production of colonie carcinomas by 1,2-dimethylhydrazineandazoxyalkanes. in: W. J. Burdette and C. Thomas (eds.), Carcinoma of theColon and Antecedent Epithelium, p. 261-279. Springfield, IL: Charles CThomas, Publisher, 1970.

23. Farber, E. The sequential analysisof liver cancer induction. Biochim. Biophys.Acta, 605:149-16,1980.

24. Farber, E. Chemical carcinogenesis. A biologic perspective. Am. J. Pathoi.,706: 271-294,1982.

25. Farber, E., and Cameron, R. The sequential analysis of cancer development.Adv. Cancer Res., 37:125-226,1980.

26. Galjaard, H., Van Der Meer-Fieggen. and Giesen, J. Feedback control byfunctional viliuscells on cell proliferationand maturation in intestinalepithelium.Exp. Cell Res., 73:197-207,1972.

27. Haase, P., Cowen, D. M., Knowles, J. C., and Cooper, E. H. Evaluation ofdimethylhydrazine induced tumors in mice as a model system for colorectalcancer. Br. J. Cancer, 28: 530-537,1973.

28. Hagihara,P. F., Yoneda, K., Sachatello.C. R., Hedgecock, H., Flesher,J. W.,Ram, M. D., Griffen, W. O., and Goldenberg, D. M. Colonie tumorigenesis inrats with 1,2-dimethylhydrazine.Dis. Colon Rectum, 23:137-140,1980.

29. Hanson,W. R., and Osbome, J. W. Epithelialcell kinetics in the small intestineof the rat 60 days after resection of 70 percent of the ileum and jejunum.Gastroenterology,60:1087-1098,1971.

30. Henderson, J. S., and Rous, P. The plating of tumor components on thesubcutaneous expanses of young mice. Findings with benign and malignantepidermalgrowths and with mammary carcinomas. J. Exp. Med., 775:1211-1230,1962.

31. Lajtha, L. G. Stem cell kinetics. In: Gordon (ed.), Regulationof Hematopoiesis,pp. 111-131. New York: Appteton-Century-Crofts, 1970.

32. Lappe, M. A. Evidencefor trie antigenicity of papiltomasinduced by 3 methylcholanthrene.J. Nati. Cancer Inst., 40:823-846,1968.

33. Leblond, C. P. The life history of cells in renewing systems. Am. J. Anat., 760:113-158,1981.

34. Leblond, C. P., and Walker, B. E. Renewalof cell populations. Phys. Rev., 36.255-276,1956.

35. Lipkm, M. PhaseI and Phase2 proliferative lesions of colonieepithelialcells indiseases leading to coloniecancer. Cancer (Phila.),34: 878-888,1974.

36. Maskens, A. P. Histogenesis and growth pattern of 1,2-dimethylhydrazine-induced rat colon adenocarcinoma.Cancer Res., 36: 1585-1592,1976.

37. Pozhansski, K. M., and Klimashevski, V. F. Comparative morphological andhistoautoradiographic study of multiple experimental intestinal tumors. Exp.Pathoi., 9: 8-98,1974.

38. Quastler, H., and Sherman, F. G. Cell population kinetics in the intestinalepitheliumof the mouse. Exp. Cell Res., 77:420-438,1959.

39. Richards, T. C. Early changes in the dynamics of crypt cell populations inmouse colon following administration of 1,2-dimethylhydrazine.Cancer Res.,37: 1680-1685,1977.

40. Richards,T. C. Theeffects of the carcinogen1,2-dimethylhydrazineon turnoverof 3H-thymk»nelabeledcells from mucosal glands of mouse colon. Anat. Ree.,200:299-308,1981.

41. Roof, B. S. Generalcharacter of epidermalpapiUomasinduced by carcinogenson mouse skin as disclosed by transplantation. Proc. Sec. Exp. Biol. Med.,702: 41-42,1959.

42. Rous, P., and Allen, R. A. Fatal keratomas due to deep homecrafts of thebenign papillomas of tarred mouse skin; normal proclivities and neoplastiadisabilitiesas determinants of tumor course. J. Exp. Med., 707: 63-86,1958.

43. Sassier, P., and Bergeron, M. Cellularchanges in the small intestine epitheliumin the course of cell proliferation and maturation. Subcell. Biochem., 5: 129-185,1978.

44. Snow, A. D., and Altmann, G. G. A morphometric study of the rat duodenalepithelium during the initial six months of the 1,2-dimethylhydrazinecarcinogenesis. Cancer Res.. 43: 4838-4849,1983.

45. Stevens Hooper, C. E. Use of colchicine for the measurement of mitotic ratein the intestinal epithelium.Am. J. Anat., 708: 231-244,1961.

46. Sunter, J. P., Appteton, D. R., Wright, N. A., and Watson, A. J. Kinetics ofchanges in the crypts of the jejunum mucosa of dimethylhydrazine-treatedrats. Br. J. Cancer, 37: 662-672,1978.

47. Sunter, J. P., Appteton, D. R., Wright, N. A., and Watson, A. J. Pathologicalfeatures of the colonie tumours induced in rats by the administration of 1,2-dimethylhydrazine.Virchows Arch. B Zeilpathol., 29: 221-223,1978.

48. Themherr, N., Deschner, E. E., Stonehill, E. H., and Lipkin, M. Induction of

CANCER RESEARCH VOL. 44 DECEMBER 1984

5530

on March 4, 2021. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

EFFECTS OF DMH ALONG RAT SMALL INTESTINE

adenocarcinomas of the colon in mice by weekly injections of 1,2-dimethylhy- 52. Wiebecke, B., Krey, V., Lohrs, V., and Eder, M. Morphological and autoradi-drazine. Cancer Res., 33:940-945,1973. ographical investigations on experimental caroinogenesis and polyp deveiop-

49. Tutton, P. J. M., and Barkla, D. H. Cell proliferation in the descending colon of ment in the intestinal tract of rats and mice. Virchows Arch. Abt. A Pathol.dimethylhydrazine treated rats and in dimethylhydrazine induced adenocarci- Anat., 360:179-193,1973.nomata. VirchoWs Arch. Abt. B Zellpathol., 21:147-160,1976. 53. Wittig, G., Wildner, G. P., and Ziebarth, D. Der Einfluss der Ingesta auf

50. Van der Meer-Fieggen, W. Regulation of cell proliferation and differentiation in Kanzerisierung des Rattendarms durch Dimetrtylhydrazin. Arch. Geschwulst-intestinal epithelium. Thesis, Rotterdam, 1973. forsch., 37:105-115,1971.

51. Ward, J. M. Morphogenesis of chemically induced neoplasms of the colon and 54. Zar, J. Biostatistical Analysis, pp. 268-273. Englewood Cliff, NJ: Prentice Hal,small intestine in rats. Lab. Invest., 30:505-513,1974. Inc., 1974.

CANCER RESEARCH VOL. 44 DECEMBER 1984

5531

on March 4, 2021. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1984;44:5522-5531. Cancer Res   G. G. Altmann and A. D. Snow  Small IntestineCells Present in the Villi, Crypts, and Mitotic Pool along the Rat Effects of 1,2-Dimethylhydrazine on the Number of Epithelial

  Updated version

  http://cancerres.aacrjournals.org/content/44/12_Part_1/5522

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/44/12_Part_1/5522To request permission to re-use all or part of this article, use this link

on March 4, 2021. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from