carcinoembryonic antigen production, secretion, and ... · radiology, university of california, san...

[CANCER RESEARCH 44,5475-5481, December 1984]

Carcinoembryonic Antigen Production, Secretion, and Kinetics in BALB/c Miceand a Nude Mouse-Human Tumor Model1

Kenneth W. Martin2 and Samuel E. Halpern3

Department of Radiology, University of California, University Hospital, San Diego, California 92103 [K. W. U., S. E. H.], and Nuclear Medicine Service, VeteransAdministration Medical Center, San Diego, California 92161 [S. E. H.]

ABSTRACT

Carcinoembryonic antigen (CEA) is currently being used as atarget antigen in the radioimmunodetection of cancer. CirculatingCEA may adversely affect the outcome of such studies byformation of intravascular immune complexes. The followingstudies were undertaken to expand our knowledge of the production, secretion, and pharmacokinetics of CEA, since thesefactors should have a direct bearing on the serum levels of CEAencountered in radioimmunodetection.

The production of CEA was assessed in nude mice givenimplants of the T-380 CEA secreting human colon tumor. SerumCEA rose linearly as the tumors enlarged; however, the concentration of CEA per g of extracted tumor remained constantthroughout the weight range studied.

The secretory rate of the T-380 tumor was determined bysurgically removing all blood flow to the liver and gastrointestinaltract of the nude mouse model. This procedure removes theknown sites of CEA degradation. Serum CEA levels rose progressively following surgery, the values being directly related tothe tumor size. The secretory rate was also proportional to tumorsize but was a constant 13.8 Â±3.6 (S.D.) ng/g tumor/hr whenexpressed on a per g tumor basis.

To determine if the serum levels of CEA observed in patientscould be due to unique differences in the clearance rates of eachpatient's CEA, serum from three patients with CEA levels of

2150,709, and 58 ng/ml was administered i.v. to groups of miceat the original and diluted concentrations. The kinetics of allsamples followed a single exponential clearance pattern with ahalf-time of about 2.5 hr. This was dramatically different fromthe kinetics of tumor-extracted CEA which exhibited a multiex-ponential pattern, the first component having a half-time of 3

min. These data suggest that CEA secreted by a tumor is insome way different from that adhering to the tumor. If thesecreted CEA truly has a monoexponential clearance with a fixedrate as the experiments suggest, the absolute values of serumCEA are either entirely a function of the tumor secretory rate, orelse the product having the short half-time is not measured in

serum samples obtained from patients.

INTRODUCTION

The radioimmunoassay of serum CEA4 has become an impor

tant method of assessing the progress of certain cancers. Many

1Research funded by the Veterans Administration and the Department of

Radiology, University of California, San Diego, CA.2 Present address: Mallinckrodt Institute of Radiology, 510 South Kingshighway,

St. Louis, MO 63110. This work was accomplished during his senior year of medicalschool at University of California, San Diego, CA.

3 Professor of Radiology, University of California San Diego, and Staff Physician,

Department of Nuclear Medicine, San Diego, CA.4The abbreviations used are: CEA, Carcinoembryonic antigen; HSA, human

serum albumin.Received June 5,1984; accepted August 31,1984.

tumors express this antigen (13,19); however, they do so to avariable degree (8, 15). This variability is observed even amongcancers arising from the same tissue (5). The extent to whichthis variability is due to differences in the specific CEA moietyproduced is unknown. Shuster ef al. (17), for example, havedemonstrated that the CEA obtained by extraction of a tumorexhibits multiexponential serum ?)/;>in experimental animals. Thesite of CEA removal was the liver, with 50% of the labeled CEAremoved in 5 min, and 80% by one-half hr after injection. The

implications of these data are of major significance when serumlevels of this antigen are used clinically. If CEA secreted by alltumors exhibited the intravascular kinetics described above,huge amounts of CEA would have to be produced and secretedto maintain serum levels in the range of 1000 ng/ml. Such levelsdo occur clinically. The further observation that high serum levelsof CEA can sometimes be produced by relatively small tumorsis even more intriguing, since that would require the small tumorto manufacture and secrete truly extraordinary amounts of CEA.

Several investigators have labeled heterogeneous as well asmonoclonal antibodies targeted against CEA, and detected tumors expressing the CEA antigen (3, 4, 6,11,12, 20). Immunecomplex formation was observed in some cases (4), and questions have arisen concerning the effect of these complexes (6)on radioimmuno imaging. If a variety of CEA species is producedand secreted that has unique pharmacokinetics, the number ofimmune complexes formed following administration of radiola-

beled antibody could vary markedly from patient to patient. Thisin turn could alter image quality in an unpredictable manner. Forthese reasons, the following studies were performed to gathermore data on the pharmacokinetics of CEA.

MATERIALS AND METHODS

AnimalModels

The athymic mice used in these experiments were obtained from thenude mouse facility of the University of California, San Diego. They werefed water and food ad libitum prior to, and throughout the experiment.Tumor passage was by the mince-trocar technique. This method produced tumors of varying size; however, a 1-g tumor usually formed in 3weeks. The tumor, designated T-380 (7), was derived from a human

colon cancer. It remains mostly viable until it exceeds 1 g, followingwhich the quantity of necrotic tissue can vary. All inoculations were mades.c. onto the animal's back. At the time of transplantation, the tumor had

undergone between 20 and 30 passages; however, the tumor continuedto produce and secrete CEA.

Normal BALB/c mice (Simonsen Laboratory, Gilroy, CA) were fedwater and food ad libitum, and were used in all animal experiments thatdid not call for a tumor model.

Acquisition of Mouse Serum and QuantitÃ¤ten of Serum-derived CEA

The mice were anesthetized using diethyl ether and then pinned,ventral side up, to a styrofoam board. The axillary vessels were cut, and

CANCER RESEARCH VOL. 44 DECEMBER 1984

5475

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

http://cancerres.aacrjournals.org/

CEA PRODUCTION, SECRETION, AND KINETICS

the blood (generally over 1 ml) was collected in a syringe and centrifugeaat 12,000 rpm for 5 min. The serum was aspirated and assayed for CEAby a double-antibody approach (Hybritech, Inc., San Diego, CA), or by

the Abbott Kit (Abbott Laboratories, Chicago, IL). The tumors wereexcised from the animals, cleaned of all extraneous mouse tissue, and

immediately frozen.

QuantitÃ¤ten of Tumor-derived CEA

At the time of assay, the tumors were thawed and the CEA extractedin the following manner. About 0.2 g of the surface of the tumor wassliced off with a scalpel and transferred to a grinding tube together with0.9% NaCI solution (saline), 0.1% sodium azide, and 0.1% HSA. Thetissue was finely ground in a 4Â°environment. Following the grinding, thetubes were centrifugea at 4Â°for 15 min, and the supernatant was filtered

through a 0.22-^m Millipore filter. This was assayed for CEA using the

techniques described earlier. This method differs from those describedpreviously in that it does not use perchloric acid for extraction. Thischange in technique was used because it was thought to represent theleast traumatic conditions for CEA extraction. In all, about 10 ml ofsolution were used for each sample.

Determination of the Secretory Rate of the T-380 Tumors

Nude mice bearing 0.7- to 5.0-g tumors were administered atropinesulfate (0.1 mg/kg) s.c. and then ether-anesthetized and placed under a

lamp to ensure against hypothermia. The atropine was useful for supporting good blood pressure, although later it was determined not to benecessary for the survival of the animals. Two 50-//I capillary tubes werefilled with Wood from the tail, centrifuged, and broken at the RBC-serum

interface. The serum was saved for assay. The mice were then taped,ventral side up, to a styrofoam box, and an incision made from thexiphoid to above the urinary bladder. The entire bowel including thestomach and the spleen was tied tightly with 1-0 silk suture material and

excised. Next, the hepatic artery and portal vein were tied (Chart 1). Thisresults in total removal of the liver, spleen, and gastrointestinal tract fromthe circulation. Any CEA formed under these circumstances shouldremain in the serum. The mice recovered rapidly and were awake andactive at the time of sacrifice 30 to 65 min later. At that time, etheranesthesia was again induced, and blood was obtained by axillary cutdown. The tumors were excised, weighed, and examined for abscessformation. The latter is necessary, since we have noted that abscessedtumors produce very little CEA. The serum was obtained from the bloodas described previously, and the pre- and postsurgical serum CEA levels

were determined in the following manner.The length of the capillary tube was measured which allowed the

computation of the volume of serum, since volume per mm was known.A syringe was connected to the tube, and enough diluent aspiratedthrough it to bring the final volume to 100 Â¡A.The capillary tube was thenflushed with this solution several times to ensure that all of the samplewashed into the assay chamber. The change in serum levels of CEA inthe pre- and postsurgical samples represents the amount secreted

following surgery and is expressed on the basis of both time and tumorweight.

The total CEA in the circulation was determined by multiplying theserum CEA concentration by the total serum volume of the mouse. Thevolume factor had been derived previously by administering 12SI-HSAi.v.

to normal mice and assessing the value of distribution by comparing 100n\ of post-CEA injection serum with a standard of the injected dose. Thevolume of distribution was then expressed on a per-g-mouse basis. In 5

animals, this factor averaged 0.047 Â±0.003 ml/g (S.D.). Admittedly,there are errors inherent in such extrapolations; however, this should bemore accurate than estimations made on the basis of animal weight, theusual method of assessment.

Chart 1. Diagram of the surgical procedure used in the secretory rate preparation. A 1-0 silk ligature is placed around the abdominal contents in a manner thatoccludes the vascular supply to the spleen and entire abdominal gastrointestinaltract. A second ligature is placed around the hepatic artery and portal vein. Thegastrointestinal circulation must be occluded prior to tying the hepatic vessels toavoid pooling of blood in the gut and attendant shock. Exclusion of the liver fromthe circulation is followed by rising serum levels of CEA. When properly performed,this technique allows the determination of the quantity of CEA secreted by thetumor over a unit period of time.

CEA Clearance Studies

CEA from Patient Serum. The serum used in this experiment wasobtained from 3 patients with metastatic colon cancer. Radioimmunoas-

say indicated CEA levels of 2150, 709, and 58 ng/ml, respectively. Thefollowing clearance studies were performed in BALB/c mice. One-hundred fifty n\ of the 2150-ng/ml sample were administered at full

strength to 15 mice. Other aliquots of this serum were diluted to 103and 20 ng/ml with saline and 1% HSA such that 150 ^ I of solutioncontained 15.4 and 3 ng of CEA. Each of these aliquots was thenadministered to 15 mice. The second serum sample, containing 709 ngof CEA per ml, was administered in full strength, and also diluted to 20ng/ml. In each case, 150 pi were injected into 15 mice (106 and 3 ng,respectively). The third serum sample (58 ng/ml) was administeredundiluted in a 150-pl volume (8.7 ng) to 15 mice.

The mice from each lot of 15 were sacrificed in groups of 5 at 10,30,and 120 min postinjection, and the CEA in each mouse's serum was

assayed. Total serum volumes of the mice were calculated as describedpreviously. Corrections for the volume injected were also applied.

CEA Extracted from Tumors. The clearance kinetics of CEA obtainedfrom tumor extracts was determined in normal BALB/c mice. Twoextracts were studied. The first was a perchloric acid-derived CEA from

a liver metastasis of a patient with colon carcinoma. The second was asaline extract of a T-380 tumor removed from a mouse used in one of

our experiments.The perchloric acid extract of CEA (270 ng; 150 /d), which had been


5476




diluted 1:10 with saline, was injected i.v. into 20 mice. The mice weresacrificed in groups of 5 at 10, 20, 40, and 80 min. A Hamilton syringewas used for the injection to ensure accurate delivery of the dose. SerumCEA levels were determined for each of the mice at the various timeperiods, and the total remaining CEA in the serum determined byweighing the mouse and multiplying by the factor 0.047 Â± 0.003 mlserum/g mouse as described previously.

The extract from the T-380 tumor was prepared for injection by diluting

it a factor of 3 with each of the following solutions: human serum; mouseserum; and 0.9% saline with 1% HSA added. This was done to determineif the diluent affected the kinetics of the CEA. Three groups of 15 micewere given injections of 150 ml (44.1 to 55.4 ng) of CEA and sacrificedat 10,30, and 120 min. Serum was taken by axillary cut-down, and CEAassayed by the double-antibody technique. The corrections made for the

previous extract injections were also applied to these data.

RESULTS

Effect of Tumor Size on Tumor and Serum Concentrationof CEA. Chart 2 shows the correlation of serum levels of CEAversus the tumor burden of the mouse. A continuous rise inserum CEA occurs between tumors of 0.2 and 1.7 g with acorrelation coefficient of 0.93. These data were corroborated by

<tLUO

SocÂ£

CORRELATIONCOEFFICIENT= 0.33

0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8

TUMOR MASS (grams)Chart 2. Correlation of serum CEA levels with each animal's tumor burden

indicates a direct relationship within the size limits of tumors studied.

z0pg

_1â€”Â§zoCJ"~z

xOpCD<

01LU\t

^O)ari

o:D

iâ€”5045403530252015105n.

CORRELATIONCOEFFICIENT=047.

O_0

0_o_o]Â°o

ooo--1

1 1 1 1 1 1 1 1 1 .,

0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

TUMOR MASS (grams)

Chart 3. Tumors responsible for the CEA levels shown in Chart 2 were extractedwith saline, and their CEA content plotted against tumor size. CEA concentrationranged from 16,000 to 45,000 ng/g of tumor and did not correlate with tumor size.

o, 50

40

CORRELATIONCOEFFICIENT= 0.92STANDARDERROR- 7 IS

Ãœ12345

TUMORMASS(grams)

Chart 4. Con-elation of CEA secretory rate and the mass of T-380 tumor bom

by the mouse. The secretory rate rises directly with the tumor weight.

50

40

i, 30

-5- 20

10

CORRELATIONCOEFFICIENT= 0.48

TUMORMASS (grams)Chart 5. Correlation of CEA secretory rate and mass of T-380 tumor burden

when the secretory rate is expressed on a per-g-of-tumor basis. Under thesecircumstances, the secretory rate of the tumor is sharply defined. The valueobtained for the 6 mice ranged from 10 to 18 ng/g/hr with a mean of 13.8 Â±3.6ng/g/hr.

measuring the serum CEA in the mice used for the secretoryrate experiments (data not shown). Chart 3 clearly indicates thatthe CEA content per g of the tumors in Chart 2 is independentof the tumor size, at least within the limits described.

Secretory Rate. Chart 4 shows the secretory rate of the T-

380 tumor in mice bearing steadily increasing tumor bulk. The0.92 correlation coefficient is almost identical to that observed inExperiment 1 where serum CEA was plotted against tumor size.Displaying the secretory rate of the T-380 tumor on a per-g basis

(Chart 5) indicates that this rate is fixed within the range of tumorsizes studied. The secretory rate in the 6 animals ranged from10 to 18 ng/hr/g of tumor with an average of 13.8 Â±3.6 ng/hr/9-

CEA Clearance Studies: Human Serum-derived CEA. Analysis of the kinetics of the serum-derived CEA (Table 1) indicates

only minor differences in clearance of CEA regardless of whichpatient it was obtained from, the original CEA concentration ofthe serum or the dilution of the antigen. The results are compat-


5477




Table 1Clearance of CEA obtained from serum of 3 patients

The serum was injected i.v. into BALB/c mice undiluted and diluted with 0.9% saline and human HSA. The mice were killed in groups of 5 at 10, 30, and 120 minafter injection. The CEA is removed by single exponential kinetics with half-timeof about 2.5 hr. These values were independentof the original serum values, its dilution,or the patient from whom the serum was obtained.

CEA in 1.0 ml ofpatientserum2150

ng21

50 ng + 0.9%saline+HSA21

50 ng + 0.9%saline+HSA709

ng709

ng + 0.9% saline+HSA58

ngBase

line foranalysis%

of total injecteddose%

of 10-minlevel%of 30-minlevel%of total injecteddose%




of 10-minlevel%of 30-mintevel%of total injecteddose%

of 10-minlevel%of 30-min levelCEA

injected(ng)32232232215.415.415.43331061061063338.78.78.7CEA

remaining after following samplingtimes (S.D.; n -1)10

min113.49.8100107.03.7100120.38.710096.210.8100128.77.010092.86.51

10030min106.67.494.010094.03.488.110099.713.982.810092.413.B96.1100106.313.682.610089.25.795.4100120min61.45.754.257.650.42.344.350.374.68.562.074.958.45.560.863.360.58.346.956.750.03.953.856.3K"

for various timeintervals0-120

min 10-120 min 30-120min0.24

0.330.370.34

0.440.460.15

0.260.190.27

0.270.300.25

0.260.380.35

0.34 0.38

where C0 is the initial CEA level and C is the final level for the time interval T expressed in hr.

Table 2

Kinetics of CEA derived by perchloric acid extraction of tumor

Clearanceof CEA obtained by the perchloricacidextraction of a liver metastasisobtained from a patient with colon carcinoma. The extract was injected i.v. intoBALB/c mice, and the animalswere killed in groups of 5 at 10,20,40, and 80 minafter injection. Ninety-five % of the CEA was cleared in the first 10 min. Amultiexponential curve is formed by these data compatible with a heterogeneouspopulation of molecules.

% of CEA remainingafter sampling time(S.D.;n-1)Base

line foranalysis%

of total injecteddose%

of 10-min levelCEA

injected(ng)270

27010

min4.9

0.810020

min5.6

1.010840min3.6

0.867.780min2.20.441.6Xa

for various

timeintervals0-10

40-80minmin18.1

0.75

â€¢K-GK)where Câ€žis the initial CEA level and C is the final level for the time interval Texpressed in hr.

Â¡blewith first-order kinetics and indicate that within the limits of

CEA concentrations observed, the clearance rate is not dosedependent. The half-time of this CEA in the serum is approxi

mately 2.5 hr.Tumor-derived CEA. Ninety-five % of the perchloric acid-

extracted material was removed from the vascular compartmentwithin 10 min of administration (Table 2). The remaining CEAshowed a disappearance rate similar to that of the serum-derivedCEA.

Table 3 shows the CEA kinetics when saline is used to extractthe tumors. It is apparent that neither the manner of extractionnor the solvent changes the basic disappearance pattern observed in Table 2. About 90% of the CEA is cleared by 10 min,95% by 30 min, and 98% by 2 hr. This is the equivalent of aclearance constant (K) of 13 for the first 10 min, 2 for the 10- to30-min interval, and 0.6 for the 30- to 120-min period. Expressedin terms of serum half-time the K translates to a half-time of 3

min for the first K, and 70 min for the third K.

DISCUSSION

The observation that serum CEA levels are directly related totumor size corroborates the work of Miwa ef al. (14), and iscontrary to that of Stragand et al. (18) and Lewis et al. (9). Withinthe 2-g range, the correlation of tumor size and serum CEA in

our experiments was high and continued to correlate well intumors up to 5 g. It is conceivable that in very large tumors(which tend to become necrotic) or in tumors that by their natureare necrotic, this linear correlation of serum CEA and tumormass may fail. Our group, as well as others, has observed aninverse relationship between tumor size and the percentage ofviable tissue in its makeup. Certainly, the Wood flow to largetumors is inferior to that of small tumors (1). This results in notonly frank necrosis but also cellular hypoxia in general (2, 21),which in turn leads to increased anaerobic metabolism and adecrease in intracellular pH (22). It is not clear what effect thesechanges have on CEA production and secretion. It is clear,however, that CEA continues to be produced by the tumor as it


5478




Table 3Kinetics of CEA derived by saline extraction of a T-380 tumor

The extract was administered diluted in mouse serum, 0.9% saline, or human serum to control for diluant interference in the CEA kinetics. The animals were giveni.v. injections of the preparations, and killed in groups of 5 at 10,30, and 120 min after injection. A rnultiexponential pattern is seen which is very similar to that producedby perchloric acid-extracted CEA and that observed by Shuster et al. (17).

Tumor extract + mouseserumTumor

extract + 0.9%salineTumor

extract + humanserumMean

of samples aboveBase

line foranalysis%

of totaldose%

of 10-minlevel%of 30-minlevel%

of totaldose%


of totaldose%


of totaldose%of 10-minlevel%of 30-min levelCEA

injected(ng)46.246.246.255.455.455.444.144.144.148.648.648.6%

of CEA remaining after samplingtime (S.D.; n -1)10

min10.32.21008.02.610015.32.010011.210030min1.41.514.11006.23.077.81009.93.464.51005.751.3100120min1.02.69.769.31.25.815.319.64.55.029.145.22.320.539.9K*

for various timeintervals0-10

min 10-30 min 30-120min13.7

5.90.2415.2

0.751.0811.3

1.310.5313.16

2.00 0.61

where C0 is the initial CEA level and C is the final level for the time interval T expressed in hr.

grows, and with within the 0.2- to 1.7-g range, this productionstays relatively constant if expressed on a per g basis; however,even in small tumors, we have observed some that contain twicethe concentration of CEA as another of the same size. Thissuggests that other factors effect CEA production as well asthose mentioned. Perhaps the inoculum used to initiate thetumors in some of the animals contained cells less able toexpress the antigen than did the inoculum transplanted into micewhose tumors produced large amounts of CEA.

To our knowledge, no information has been published on thesecretory rate of a tumor other than in cell culture or suspension(16).The method we havedescribed, while admittedly associatedwith alterations of normal physiology, provides information thatcorrelates well with the serum CEA levels we have observed inthe unmanipulated state of the animals. It was possible todemonstrate that while small tumors produced low serum CEAlevels, and large tumors much higher levels, the secretory rate,which is based upon time and tumor mass, remained constant.This is in keeping with the data from clinical studies that indicatea slow rise in serum CEA levels as the total body tumor burdenrises (10). It also appears to be in keeping with the phenotypicexpression of the tumor as this technique was applied to 3 othernude mouse-tumor models that produced progressively greaterserum levels of CEA than the one described (6). In each case,the secretory rate fell into the range expected of a tumor responsible for such values.

When the amount of CEA in the T-380 tumor is compared tothe amount secreted in a 24-hr period, it becomes obvious thatvery little of the antigen present is secreted. A 1-g T-380 tumorcontains about 20,000 ng of CEA by our assay methods; yet, ata secretory rate of 13.8 ng/hr, it will secrete only 331 ng, or1.7% of the total immunoreactive substance available in oneday. Why such small amounts are liberated can only be speculated; however, what is implied is that more than one CEA

species exists, and that only one is secreted in significantamounts. That product would appear to be recognized by theliver in a different manner than CEA obtained by the extractionof CEA-producing tumors. The clearance kinetics of both theperchloric acid- and saline-extracted CEA show a pattern verysimilar to that observed by Shuster ef al. (17), even though theexperimental animal used by those investigators was differentthan the one used in our experiments.

A mathematical model which described the clearance of asubstance from the blood was used to better describe ourresults. The basic equation (Equation A) assumes, in accordancewith the work of Shuster ef a/. (17), that the liver is responsiblefor the removal of CEA:

r r^C = Coe- â€”FkT(A)

where V is serum volume in ml, F is serum flow to the tumor inml/hr, 7 is time in hr, k is fraction of CEA cleared in one passthrough the liver, C0 is initial CEA concentration, and C isconcentration of CEA at time 7. If it is assumed that F and V areconstant (which should be the case), and that the same fractionof available CEA is cleared by each pass through the liver, asingle constant K can be substituted as shown in Equation B toyield Equation C.

K = 7 <B)

C = COB~KT (C)

This constant K can be determined experimentally by rearrangingthis equation, substituting in the experimental values, and solvingfor K as in Equation D.

K = -I in Â£ (D)


5479




This clearance constant K was determined for all of the solutionsinjected. All of the patients' serum-derived CEA had K values

that were approximately equal, and there was no evidence ofheterogeneity of the material.

Tumor-derived CEA, whether perchloric acid or saline-ex

tracted, showed an early rapid clearance of a second fraction. Ifthe data from the 3 samples of extract used in Table 3 arepooled and all the values are averaged, K constants can bedetermined for each time interval. In the first 10 min, a K of 13.16was observed, and 89% of the CEA had been cleared. The Kfrom 10 to 30 min is 2.0, and 94% of the CEA has been cleared.Finally, the 30- to 120-min interval shows a K of 0.6, and 98%

of the CEA has been removed. This is clear evidence of aheterogeneous population of CEA molecules, since a K of 13.16,if applied to the entire first 30 min, would dictate that only 0.1%of the dose injected remains in the blood instead of the nearly6% obtained experimentally.

If the data derived in the mouse model hold true for the human,it implies that the measurement of circulating CEA, while anindicator of CEA secretion, gives no indication of its production.This could have far-reaching consequences for radioimmunode-

tection. An abundance of CEA might be present on a tumor, butvery little might be secreted. Such a tumor presents a legitimatetarget for radioimmunodetection using either heterogeneous ormonoclonal antibodies. Its detection would not be inhibited bythe presence of circulating antigen. Conversely, a tumor couldtheoretically produce only moderate or small amounts of CEA,yet have a high secretory rate. This would result in high serumlevels of circulating antigen with little CEA in the tumor to serveas a target for the labeled antibody. Animal work in our laboratoryindicates that immune complexes would occur in these circumstances, and that the complexes would be cleared by the liverwith an attendant increased radiation dose to that organ, and adecrease in tumor concentration of radiopharmaceutical (6). Gol-

denberg ef al. (4) and Mach et al. (11) have observed theformation of CEAiantibody complexes in patients subjected toradioimmunodetection scanning and report that tumor detectionwas not inhibited. It should be noted, however, that both investigators used radioiodinated materials. Our work indicates thatradioiodinated murine complexes will be cleared by the liver inexperimental animals, and quickly dehalogenated (6). The freeiodine leaves the liver and is eliminated in the urine. As aconsequence, the liver does not become intensely radioactive,so the evidence of immune complex acquisition by the liver isnot observed. When nun is used as the radiolabel for theantibody, there is a dramatic rise in liver activity which is proportional to the serum CEA levels. The tumor continues to concentrate some of the radioiodinated antibody; however, the absoluteamount is less than that accumulated in the absence of immuneCEA complexes. It is possible that, in those cases where extraordinarily high serum CEA levels occur (with nearly all of theantibody being complexed), the complexes not only form butbreak apart, at least to the extent that some accumulation in thetumor does occur. The end result, however, of high circulatingCEA (if the animal model studies are valid evidence of whatoccurs in the human) will be a decrease in the absolute uptakeof labeled antibodies by the tumor and, in all probability adiminished chance of detecting a lesion. This reasoning impliesthat the most appropriate patient to select for radioimmunode

tection may be those with near-normal levels of CEA in their

serum.

ACKNOWLEDGMENTS

The authors wish to thank the HybritechCorp. for supplyingpart of the materialfor the CEA assays.

REFERENCES

1. Catalana, S., Cohen, C., and Sapirstein, L. A. Relationshipbetween size andperfusion rate of transplanted tumors. J. Nati. Cancer Inst., 29: 389-391,1962.

2. Constable,T. B., Rogers, M. A., and Evans, N. T. S. Comparisonbetween theoxygen removalrate and the histologicalstructure of normaland tumor tissues.PfluegersArch. Eur. J. Phystol.,373:145-151,1978.

3. Goldenberg,D. M., Deland, F. H., Kim, E. E., Bennett, S., Primus, F. J., VanNagel,J. R., Jr., Estes, N., Desiinone,P., and Raybum, P. Useof radiolabeledantibodies to carcinoembryonic antigen for the detection and localization ofdiversecancers by external photoscanning.N. Engl.J. Med., 298:1384-1388,1978.

4. Goldenberg,D. M., Kim, E. E., Deland, F. H., Bennett, S., and Primus, F. J.Radioimmunodetectionof cancer with radioactive antibodies to carcinoem-bryonic antigen. Cancer Res., 40:2984-2992,1980.

5. Gosland, R.. O'Brian, M. J., Steel, G., Mayer, R . Wilson, R., Corson, J. N.,

and Zamcheck, N. Correlation of plasma CEA and CEA tissue staining inpoorly differentiated colorectal cancer. Am. J. Med., 71: 246-253,1981.

6. Hagan, P. L., Halpem, S. E , Chen, A., Krishnen, L., Frincke, J. M., Bartholomew, R. M., David, G. S., and Carlo, D. J. Effect of tumor size, caranoembryonic antigen (CEA) production and secretion on the distribution of 111lnand 1251monoclonal antitumor antibodies (MoAb) in animal models. J. NucÃ.Med., 24: 77,1983.

7. Halpem, S. E., Hagan, P. L., Garver, P. R., Koztol, J. A., Chen, A. W. N.,Frincke, J. M., Bartholomew, R. M., David, G. S., and Adams, T. H. Stability,characterization,and kinetics of '"In-labeled monoclonalantitumor antibodiesin normal animals and nude mouseâ€”humantumor models. Cancer Res., 43:5347-5355,1983.

8. Lavin, P. T., Day, J., Holyoke. A. 0., Mittleman, A., and Ming Chu, T. Astatistical evaluation of baseline and followup carcmoembryonic antigen inpatients with resectable colorectal carcinoma. Cancer (Phila.),47: 823-826,1981.

9. Lewis, J. C. M., and Keep, P. A. Relationship of serum CEA levels to tumorsize and CEA content in nude mice bearing colonic-tumor xenografts. Br. J.Cancer,44: 381-387,1981.

10. Mach, J-P., Bertholet, M. M., Jaeger, P., Ruegsegger, C-H., Pettavel, J., andLoosli, R. M. Detection of recurrence of large bowel carcinoma by radioim-munoassay of circulating carcinoembryonic antigen (C.E.A.) Lancet, 2: 535-540,1974.

11. Mach. J. P., Buchegger, F., Forni, M., Ritschard, J., Berche, C., Umbroso, J.D., Schreyer,M., Girardet,C., Accolla,R. S., and Carrel,S. Useof radiolabeledmonoclonal anti CEA antibodies for the detection of human carcinomas byexternal photoscanning and tomoscintigraphy. Immunol. Today, 2: 239-249,1981.

12. Mach, J-P., Carrel, S., Forni, M., Ritschard, J., Donath, A., and Alberto, P.Tumor localizationof radiolabeledantibodiesagainst carcinoembryonicantigenin patients with carcinoma. N. Engl. J. Med., 303: 5-10,1980.

13. Madeddu, G., Langer, M., Dettori. G., and Costanza, C. Role of serumcarcinoembryonicantigen in ore-operativediagnosis of cancer in patients withthyroid nodules. Cancer (Phila.)45: 2607-2610,1980.

14. Miwa, N., Hirohashi,S., Shimosato, Y., Mukojima, T., Ohkura, H., Sakura, H.,Kawachi,T., and Sugimura,T. Co-production of carcinoembryonicantigenanda/pAia-feto-proteinin transplantable human colon cancer in nude mice. Proceedings of the Second InternationalWorkshop on Nude Mice, pp. 435-439.Tokyo: Universityof Tokyo Press, 1977.

15. Moertel, C. G., Schutt, A. J., and Go, V. L. W. Carcinoembryonicantigentestsfor recurrent colorectal carcinoma: inadequacy for early detection. J.A.M.A.,239:1065-1066,1978.

16. Rosenthal, K. L, Tompkins, W. A. F., and RaHs,W. E. Factors affecting theexpressionof carcinoembryonicantigenat the surfaceof cultured humancoloncarcinomacells. Cancer Res., 40: 4744-4750,1980.

17. Shuster,J., Silverman,N.,and Gold,P. Metabolismof humancarcinoembryonicantigen in xenogeneicanimals. Cancer Res., 33: 65-68,1973.

18. Stragand, J. J., Yang, L-Y., and Drewinko, B. Serum CEA levels in a humancolonie adenocarcinoma(LOVO) xenogran system. Cancer Lett. 70: 45-50,


5480




1980. primate anti carcinoembryonic antigen (CEA) antibodies in patients with cancer.19. Sugarbaker, P. H., Skarin, A. T., and Zamcheck, N. Patterns of serial CEA Invest. Radiol.. 17:350-355,1982.

assays and their clinical use in the management of colorectal cancer. J. Surg. 21. Vaupel, P. Hypoxia in neoplastia tissue. Microvasc. Res., 73:399-408,1977.Oncol., 8:523-537,1976. 22. Vaupel, P.. Thews. G., and Wendung. P. Kritische Sauerstoffâ€”und glucoseâ€”

20. Sullivan, D. C., Suva, J. S., Cox, C. E., Haagensen, D. E., Jr., Harris, C. C., Versorgung maligner tumoren. Dtsch. Med. Wochenschr., 707. 1810-1816,Briner, W. H., and Wells, S. A., Jr. Localization of 1-131 labeled goat and 1976.


5481



1984;44:5475-5481. Cancer Res Kenneth W. Martin and Samuel E. Halpern in BALB/c Mice and a Nude Mouse-Human Tumor ModelCarcinoembryonic Antigen Production, Secretion, and Kinetics

Updated version

http://cancerres.aacrjournals.org/content/44/12_Part_1/5475

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

[email protected] 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_1/5475To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



carcinoembryonic antigen production, secretion, and ... · radiology, university of california, san...

Documents