The Expression of the CD44 Variant Exon 6 Is Associated with Lymph Node Metastasis … · Vol. 3. 1289-1297, August 1997 Clinical Cancer Research 1289 The Expression of the CD44 Variant

Vol. 3. 1289-1297, August 1997 Clinical Cancer Research 1289

The Expression of the CD44 Variant Exon 6 Is Associated with

Lymph Node Metastasis in Non-Small Cell Lung Cancer

Takanori Miyoshi, Kazuya Kondo,’ Naoki Hino,

Tadashi Uyama, and Yasumasa Monden

The Second Department of Surgery, School of Medicine, The

University of Tokushima, Kuramoto-cho, Tokushima 770, Japan

ABSTRACT

Recently, it was suggested that splice variants of the

surface glycoprotein CD44 (CD44v) were associated with

tumor metastasis in some cancers. We examined the expres-sion of variant forms of CD44 in 31 non-small cell lung

carcinomas (NSCLCs) and in 8 normal lung tissue samples

by reverse transcription-PCR (RT-PCR). CD44v3, CD44vS,

CD44v6, and CD44v7 were not expressed or were weakly

expressed in normal lung tissue (0 of 8). In contrast,

CD44v3, CD44v5, CD44v6, or CD44v7 was expressed in 28

of 31 (90.3%) NSCLCs. Additionally, we examined the ex-

pression of CD44v6, which has been shown to be related to

metastasis, in 5 normal lungs and 30 NSCLCs by RT-PCR

and immunohistochemical analysis to clarify which cellsexpress CD44v6 in NSCLC specimens. Thirty-six of 61

(59%) NSCLCs variably expressed CD44v6 by RT-PCR,

and cancer cells were selectively immunostained by anti-CD44v6 antibodies in 23 of 30 (76.7%) NSCLCs. The results

of immunohistochemical analysis almost correlated with

those of RT-PCR. NSCLCs with lymph node metastasis

expressed significantly more v6 exon than did those without

lymph node metastasis [23 of 29 (79.3%) versus 13 of 32

(40.6%); P < 0.01]. There was a significant association

between the intensity of v6 expression by RT-PCR and the

frequency of cases showing lymph node metastasis (Coch-

ran-Armitage’s test, P < 0.002). In conclusion, this studydemonstrated that in NSCLC, a number of variant forms of

CD44 are frequently expressed, although these variants are

infrequently expressed in normal lung tissue, and that the

expression of CD44v6 is particularly associated with lymph

node metastasis in NSCLC.

INTRODUCTION

The most life-threatening aspects of the oncogenic process

are invasion and metastasis. Lymphogenous or blood-borne

metastasis is an early event in patients with lung cancer. It is

difficult to treat lung cancer because of the ease of metastasis. A

search for tools that allow the metastatic potential of tumors to

be effectively assessed is very important for future lung cancer

therapies.

It is known that CD44 exists in a standard form, CD44s,2

and in multiple isoforms, which are generated by alternative

splicing of at least 10 variant exons (vl-vl0) encoding parts of

the extracellular domain. Recently, it was shown that a certain

variant exon was associated with tumor metastasis. Analyses of

several human carcinomas including colon cancer ( 1 ), breast

cancer (2), uterine cervical cancer (3), non-Hodgkin’s lym-

phoma (4), and gastric cancer (5, 6) revealed that the expression

of CD44v exons in tumors is related to tumor metastasis and

aggressiveness.

Herrlich and coworkers reported that the expression of one

particular exon (v6) was correlated with metastatic potential and

that transfection of several nonmetastatic tumor cell lines from

rats with v6-bearing forms of CD44 caused conversion to a

metastatic phenotype (7-9). In human colon cancer, the expres-

sion of CD44v6 in tumors was related to lymph node metastasis

(10).

Washimi et a!. ( I I ) found no correlation between the

expression patterns of the various forms of CD44 and the

clinicopathological data such as histological type, nodal in-

volvement, and disease stage in NSCLC (1 1 ). Givehchian el al.

( 12) also reported that there was no correlation between metas-

tasis and the expression of CD44v forms detected by IHC in

bronchial cancers.

We demonstrated that CD44s and CD44v molecules were

frequently expressed in NSCLC but were infrequently expressed

in small cell lung cancer.3 In this study, we examined the

expression pattern of CD44s and several variant forms of CD44

(CD44v3, CD44v5, CD44v6, CD44v7, and CD44vlO) in 31

NSCLCs and 13 corresponding normal lung tissue samples by

RT-PCR. We examined the expression of the v6 exon in 61

NSCLCs to clarify the relationship between the expression of

the v6 exon, which was shown to be related to metastasis by

Herrlich and coworkers, and lymph node metastasis. Moreover,

to clarify the relationship between the expression of CD44

molecules by RT-PCR and immunoreactivity by anti-CD44

antibodies, we immunostained 30 NSCLCs and S normal lungs

by anti-CD44s and v6 antibodies.

MATERIALS AND METHODS

Clinical Specimens. Sixty-one NSCLC samples and 13

corresponding normal lung tissue samples were obtained from

Received 10/7/96: revised 4/16/97; accepted 4/24/97.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby marked

advertisement in accordance with 18 U.S.C. Section 1734 solely toindicate this fact.

t To whom requests for reprints should be addressed. Phone: 8 1-886-

33-7143: Fax: 81-886-33-7144.

2 The abbreviations used are: CD44s, standard CD44: CD44v, CD44

variant: CD44v1-l0, CD44 with alternate splice variant exon 1-10.

respectively: NSCLC, non-small cell lung carcinoma: RT, reverse iran-

scription: IHC, immunohistochemical analysis: mAb, monoclonal anti-

body.3 K. Kondo, unpublished observations.

Research. on August 1, 2021. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

http://clincancerres.aacrjournals.org/

Table 1 Characteristics of the 61 NSCLC cases used to evaluate

CD44 expression

64.1 ± 6.8

(43-81)

47 cases

14 cases

25 cases

29 cases

7 cases

24 cases

4 cases

20 cases

8 cases

5 cases

13 cases

36 cases

6 cases6 cases

32 cases

7 cases

19 cases

3 cases

“ Sq, squamous cell carcinoma: Ad. adenocarcinoma: La. large cell

carcinoma.

1’ Union International Contre Cancer ( I 3).

1290 CD44v6 Relates to Lymph Node Metastasis

Age (yrs)

Average

RangeGender

Male

Female

Histology”

Sq

Ad

La

Clinical stage”

IIlIlAIIIBIV

T factor

T,T�T4

N factor

N,

N,

N,

N1

61 lung cancer patients, who had not received any preoperative

adjuvant therapy, during surgery at The Second Department of

Surgery, School of Medicine, The University of Tokushima

between March 1986 and November 1990. The characteristics

of these patients are listed in Table 1 . We used the Union

International Contre Cancer tumor-node-metastasis staging sys-

tern to stage the disease ( 13). Tumor histology was determined

according to the WHO classification of lung tumors (14). All

samples were obtained within 30 mm of surgical resection,

snap-frozen in OCT compound (Miles), and stored at -80#{176}C

until used. The corresponding normal lung tissues were obtained

from normal lung regions sufficiently far away from the can-

cerous lesions.

Evaluation of CD44s and CD44v by RT-PCR. After

confirmation of the cancerous lesions by H&E staining of a

section from frozen blocks, 20 sections ( 10 p.m) were cut from

each lesion and collected. Total RNA was extracted from the

frozen samples by a single-step procedure using an acid guani-

dinium thiocyanate-phenol chloroform mixture (15). Single-

stranded cDNA was synthesized with a CD44 exon-specific

oligonucleotide primer referring to the published CD44 nude-

otide sequence (16, 17). This was used for PCR amplification

and was homologous to positions 458-483 and 813-843 of the

human standard CD44 cDNA (Fig. 1; Ref. 16). These primers

anneal to an exon 3 region and an exon I S region of the CD44

gene (Fig. I). One of 20 pi of single-stranded template cDNA

was used. PCR was carried out for 30 cycles at 94#{176}C(denatur-

ation) for 30 s, 60#{176}C(annealing) for 30 s, and 74#{176}C(elongation)

for 1 mm. RT-PCR of 3-actin was used as the control for each

sample. The �3-actin-specific oligonucleotide primers used

were 5’-ATCACGATGCCAGTGGTACG-3’ (for RT) and 5’-

GTCAGAAGGATFCCTATGTG-3’ and S’-GCCTGGATAG-

CAACGTACAT-3’ (for PCR). PCR products from the corre-

sponding tumors and normal lungs were visualized by ethidium

bromide (0.5 jig/mi) staining under UV illumination after elec-

trophoresis in a 1.5% agarose gel (Seakem; FMC Bioproducts)

in 0.5% Tris-borate EDTA buffer, and then the products were

transferred to nylon membranes (Biodyne; Pall Bio Support,

East Hills, NY). After prehybridization, the membranes were

hybridized overnight at 42#{176}Cwith a 32P-labeled CD44-S probe

(Fig. 1 ; Refs. 16 and I 7). The CD44-S probe was a synthetic

oligonucleotide capable of annealing to exon 4, which is on the

S ‘ side of the alternative splicing site. The membranes hybrid-

ized with the CD44-S probe were washed in 2X SSC and 0.1%

SDS three times for S mm at room temperature and then washed

in 0. 1 X SSC and 0.1% SDS twice for 15 mm each time at 50#{176}C

and subjected to autoradiography. After washing out the

CD44-S probe, the same membranes were rehybridized with

probes CD44-v3, -v5, -v6, -v7, and -vlO to detect each variant

form by the same method (Fig. I ; Refs. 17 and 18).

Immunohistochemistry. Frozen sections (6 p.m) were

fixed in ice-cold acetone (10 mm), air-dried for 2 h, and then

washed in a Tris-buffered saline solution [0.05 M Tris and 0.15

M NaC1 (pH 7.6)] and preincubated with BSA for 5 mm. The

sections were removed by tapping the glass slides and then

incubated with primary antibodies SFF-2 (murine monoclonal

antibodies raised against human standard CD44) or VFF-7 (mu-

rime monoclonal antibodies raised against an epitope encoded by

exon v6 of human variant CD44; Bender, Vienna, Austria; Refs.

3 and 19) at 4#{176}Cfor 12-16 h. A 1 : 100 dilution was considered

optimal from the titration experiments, using a panel of control

tumors. The sections were incubated with the secondary bioti-

nylated antibody for 15 mm [antimouse F(ab’)2; DAKO Corp.]

to form the streptavidin-biotin peroxidase complex. Visualiza-

tion of the immunocomplex was performed by the immunoper-

oxidase-3,3’-diaminobenzidine method for 2 mm (20), and then

the reaction was stopped by adding H2O. The cells were coun-

terstained with hematoxylin, mounted with Malinol, and viewed

under the microscope. Background staining activity of the sec-

ondary antibody conjugate was excluded by staining control

slides without primary antibody. In this study, cases with more

than 50% positive tumor cells were defined as “diffusely

stained,” those with less than 50% positive tumor cells were

defined as “focally stained,” and those without positive tumor

cells were defined as “not stained.”

Intensity of CD44 Expression. Previously, we exam-

med the expression of the CD44s and CD44v exons in nine

NSCLC and five SCLC cell lines by RT-PCR. RNA of the

Ma-lO cell line (NSCLC), which strongly expressed CD44s

(380 bp) and CD44v exons (870 bp), was amplified by several

cycles of RT-PCR (IS, 20, 25, 30, 35, and 40 cycles). Because

the signals of the RT-PCR products grew exponentially between

25 and 35 cycles, we decided that 30 cycles was the most

suitable number of cycles for this study (Fig. 2). We quantified

the intensity of CD44s and CD44v6 expressions using NIH

image analysis (Version 1.56; Wane Rasband; NIH, Bethesda,

MD) to clarify the relationship between the frequency of lymph

node metastasis and the intensity of v6 expression. Because the

average intensity of the CD44s bands was 2529 ± 236 (mean ±



v2 v3 v4- - -

- - -

6B 7 8

Primer

CD44-F3

== = = = =

Probe Probe Probe Probe ProbeCD44-S CD44-v3 CD44-v6 CD44-v7 CD44-viO

CD44- Al 5

Primer CD44-RTI7Primer CD44-F3Primer CD44�R15Probe CD44-SProbe CD44-v6Probe CD44-v7Probe CD44-v3Probe CD44-v5Probe CD44-vi0

: 5’-TTGACTGCAATGCAAACTGC-3’: 5’-GACACATATTGCTTCAATGCTTCAGC-3’: 5’-TCAGATCCATGAGTGGTATGGGAC-3’: 5’-CCTGAAGAAGATTGTACATCAGTCACA-3’:5’-GGAAGAAACAGCTACCCAGAAGGAACAGTG-3: 5’-GCTCATACCAGCCATCCAATGCAA-3’: 5’-CGTCTTCAAATACCATCTCAGCAGGC-3’: 5.GAATGTGGGGTCTCTTCTTCCTCATG.3: 5-GATAAGGAACGATTGACATTAGAGTTGG-3’

Clinical Cancer Research 1291

Membranous Proximal Variable Domain Tail Exons

.�

variant exon vi

n-LI:j- �i-LJ�i-�

exon 1 2 3 4 5 6A

CD44cDNA

v5 v6 v7 v8 v9 viO

- � - - __L:9 iO ii i2 13 i4 i5

TM

1�E16 i7 18 19

PrimerCD44-RT1 7

Fig. I Schematic representation of CD44s and the CD44v isoforms as well as the probes used in this study. �, CD44s sequences: ope�i and hatched

bars, CD44v isoform sequences. Closed bars, primer CD44-RTI7 (position 965-984: Ref. 16): RT primer, primer CD44-F3 (position 458-483: Ref.

16): CD44-R15 (position 813-843: Ref. 16), PCR primer. Probe CD44-S (position 488-514: Ref. 16) detects both CD44s and CD44v isoforms.whereas probe CD44-v3 (position 4-29: Ref. 18) only hybridizes with mRNA species containing CD44v3. Probe CD44-v5 (position 321-346: Ref.

18) detects CD44v5, probe CD44-v6 (position 383-412: Ref. 18) detects CD44v6, probe CD44-v7 (position 495-518: Ref. 18) detects CD44v7. andprobe CD44-vlO (position 101 1-1038: Ref. 18) detects CD44vlO. TM, transmembrane region.

SD), the intensity of CD44s in each sample was used as an

internal control. Cases that did not express v6 were classified as

the none group. Cases with a numerical value of less than 2000

were classified as the weak group, cases with a numerical value

of more than or equal to 2000 and less than 5000 were classified

as the moderate group, and cases with a numerical value of more

than or equal to 5000 were classified as the strong group.

Statistical Analysis. The relationship between the mci-dence of CD44s and CD44v exon expression and several din-

ical factors was analyzed by x2 test, Fisher’s exact test, Stu-

dent’s t test (unpaired), or Welch test. Correlations between the

intensity of CD44v6 expression and the frequency of lymph

node metastasis were examined by using the Cochran-Armit-

age’s test. Correlations were considered to be statistically sig-

nificant at P < 0.05.

RESULTS

Expression of the CD44 Isoforms in Primary Lung

Cancers and the Corresponding Normal Lung Tissue. The

electrophoretic profiles of the PCR products of the CD44 gene

in eight lung cancer specimens and in the corresponding normal

lung tissue are shown in Fig. 3A. One PCR product had almost

the same intensity as the 382-bp fragment band, and a variety of

bands (greater than 382 bp) had variable intensities. Some bands

that were less than 382 bp were thought to be nonspecific bands

for the CD44 gene. Each sample was amplified with �3-actin

primers before CD44 amplification to check the quality and

abundance of the RNAs. PCR for �3-actin showed that the

amounts and quality of mRNA did not differ significantly

among the 16 samples.

The PCR products obtained from electrophoresis were

transferred onto nylon membranes. The membranes were hy-

bridized with the CD44-S probe that detected exon 4. After

washing out the CD44-S probe, the same membranes were

rehybridized with the probe for CD44-Vs to detect the variant

forms (Fig. 3B). All pairs of normal lung and lung cancer

specimens were found to express CD44s, which was the 382-bp

fragment, and the intensity of the CD44s expression of normal

lung tissue was almost the same as that in the lung cancer

specimens. But the expression of the variant forms of CD44

differed between normal tissue and the cancer specimens. With

the CD44-v3 probe, five of eight cancer cases expressed

CD44v3, and the intensity of CD44v3 expression varied. None

of the eight normal lung cases expressed it. With the CD44-v5

probe, none of the eight cancer or normal lung cases expressed

CD44v5. With the CD44-v6 probe, none of the eight normal

lung cases expressed CD44v6, but seven of eight cancer cases

expressed CD44v6. Although the intensity of expression varied

from case to case, CD44v7 expression paralleled CD44v6 ex-

pression in the same cases. With the CD44-v 10 probe, all of the

cancer cases expressed CD44vlO. Four of eight (50%) normal

lung cases expressed it, but CD44vIO expression in the normal

lungs was weaker than that in the corresponding cancers.

Expression of CD44 Isoforms in Primary Lung Can-

cers. We examined the expression of CD44s and the CD44v

exons in an additional 23 NSCLCs to determine the alternative

splicing pattern of the CD44 gene in lung cancer. All lung

cancer cases were found to express CD44s, and the intensity of

CD44s expression was almost the same in all of the lung cancer

cases. The frequency of variant exon expression in the 31



cycle

45 40 35 30 25 20 15 M bp

I

�Ii�

A

870

380

B 410

3� 10

� 20 10

,� 10

1353

1078

872

603

310281

25 30 35 40 45

PCR cycle


Fig. 2 A, RT-PCR amplification of CD44 tran-scripts derived from adenocarcinoma cell lines

(Ma-lO). Primers CD44-F3 and CD44-Rl5 (seethe legend to Fig. 1) were used to amplify CD44s

and CD44v transcripts by RT-PCR (15, 20, 25,

30, 35, and 40 cycles). PCR products were re-

solved in a I .5% agarose gel containing ethidium

bromide and visualized under UV light. The 380-

and 870-bp bands grew stronger by degrees as

the number of cycles increased. M, size marker

(4X174/HaeIII digested). B, the intensity of the

380- and 870-bp bands was evaluated by NIH

image analysis, and the values were plotted on a

logarithmic graph. We found that 30 cycles was

the most suitable number of cycles for this study.

. and solid line, 380-bp band: 0 and dotted line,

870-bp band.

NSCLC patients is shown in Fig. 4. With the CD44-v3 probe, 26

of 3 1 (83.9%) of the cases expressed CD44v3. The frequency of

CD44v3 expression in adenocarcinoma was higher than that in

squamous cell carcinoma (93.3 versus 66.7%). With the

CD44-vS probe, 12 of 31 (38.7%) of the cases expressed

CD44vS. With the CD44-v6 and -v7 probes, 26 of 31 (83.9%)

cases expressed CD44v6 and CD44v7. Although the intensity of

expression varied from case to case, CD44v6 expression paral-

leled that of CD44v7 in the same patients. With the CD44-vlO

probe, 30 of 31 (96.8%) cases expressed CD44v1O. The fre-

quency of CD44v5, CD44v6, CD44v7, and CD44vlO expres-

sion in adenocarcinoma was about the same as that in squamous

cell carcinoma.

Relationship between Clinical Findings and the Expres-

sion of Variant Exon 6 in NSCLC. In 3 1 NSCLCs that were

examined for expression of each variant exon by RT-PCR,

there was no relationship between v3, vS, v6, v7, or vl0

expression and clinicopathological features including age,

gender, Brinkman index, histology, T factor, and clinical

stage (data not shown). Only v6 and v7 expressions were

associated with lymph node metastasis. The expression of

other exons was not associated with lymph node metastasis

(Table 2). Because the expression of v6 paralleled that of v7,

the expression of CD44v6 was assayed by RT-PCR in an

additional 5 normal lung and 30 NSCLC specimens to con-

firm the association between v6 expression and lymph node

metastasis. In total, 2 of 13 (15%) normal lung samples

weakly expressed the CD44v6 exon. Conversely, 36 of 61

(59%) NSCLCs expressed CD44v6. Thirteen of 32 (40.6%)

NSCLCs without lymph node metastasis expressed v6, as

shown by RT-PCR, whereas 23 of 29 (79.3%) NSCLCs with

lymph node metastasis expressed v6. A statistically signifi-

cant association was seen only between the expression of the

v6 exon and lymph node metastasis (P < 0.01; Table 3).

There was no correlation between the expression of CD44v6

and the other clinicopathological parameters.

Correlation between Lymph Node Metastasis and the

Intensity of CD44v6 Expression. We quantified the intensity

of CD44s and CD44v6 expression by NIH image analysis to

clarify the association between the frequency of lymph node

metastasis and the intensity of v6 expression. Because the in-

tensity of CD44s expression in all lung cancers was almost the

same (Fig. 3A), that is, the average intensity of CD44s was

2529 ± 236 (mean ± SD), the intensity of CD44s was used as

an internal control. The 61 NSCLCs examined were divided into

four groups, according to the intensity of CD44v6 expression.

Six of 25 (25%) cases in the none group had lymph node

metastasis. Two of 5 (40%) cases in the weak group, 10 of 15

(67%) cases in the moderate group, and 1 1 of 16 (69%) cases in

the strong group had lymph node metastasis. There was a

significant association between the intensity of v6 expression

and the frequency of lymph node metastasis (Cochran-Armit-

age’s test, P < 0.002; Fig. 5).



A8 7 6 5 � 3 2

II II II II II II II

TNT N T N T N T NT N T NII

T NM

�.,.

._� �

� �=‘ �

IB

8 7

TNT N

6 5 4. 3 2 1I’ II II II II II

TN T N T NT N T N T N

�6 iI.4sprobe

yb

y7

y6

v5870-640

‘‘p

� S�

- 870

I 640

. �‘�Lt�L

S


antibodies. In the normal lung, all five specimens expressed

Fig. 3 Detection of CD44 mRNA in pri-

mary lung cancer and the corresponding

normal lung tissues. A, mRNA from eightpairs of corresponding tumor and normal

lung tissue. PCR products were visualized

by ethidium bromide staining under UV

illumination after electrophoresis. A con-

trol PCR reaction with 3-actin primers was

carried out, and the size markers (right) are

also shown. B, expression of mRNA spe-

cific for CD44s and variant forms of CD44

in eight pairs of corresponding tumor andnormal lung tissue. The membranes usedfor hybridization with 32P-labeled CD44-S

probe were washed to strip off the probe

and rehybridized with 32P-labeled

CD44-Vs probe. All pairs of samples cx-pressed CD44-S in normal lung tissues at

levels similar to those in the corresponding

primary tumors. However, expression of

CD44-Vs varied. M, size marker (�Xl74/

HaeIII digested); N, normal lung: T, pri-

mary tumor; underline, small cell carci-

noma.

Immunohistochemical Detection of CD44 in the Normal

Lung and in Cancerous Lung Tissue. To clarify the associ-

ation between CD44s and v6 expression of RT-PCR and that of

immunostain, cryostat sections from 5 of 13 normal lungs cx-

amined by RT-PCR and 30 of 61 NSCLCs examined by RT-

PCR were immunostained with anti-CD44s- and v6-specific

CD44s by RT-PCR; type I and II pneumocytes, interstitial

tissues, and macrophages were stained by anti-CD44s antibodies

(Fig. 6A). Three normal lungs that did not express CD44v6 by

RT-PCR were not stained by anti-CD44v6 antibodies, but two

normal lungs that expressed small amounts of v6 by RT-PCR

showed focal staining of type II pneumocytes by anti-CD44v6

antibodies (Fig. 6B). In the NSCLC cases, the tumor cells as



�#{149}:. Squamous cell ca.::�: (n=12)

. Adenoca.

(n=15)

&� Large cell ca.--� (n=4)

� Non-small cell ca.r”4 (n=31)

Fig. 4 The frequency of CD44v

exon expression in 3 1 NSCLC cases.

V3 VS V6 V7 V10

Probes


100

90

80

�70

I 60

� 20

10

0

well as the interstitial tissue were stained by anti-CD44s anti-

bodies, and the intensity of CD44s immunostain was almost the

same in all NSCLC cases (Fig. 6, C and E). However, seven

cases that did not express CD44v6 by RT-PCR did not show any

immunostain with anti-CD44v6 antibodies. Twenty-three cases

that expressed CD44v6 by RT-PCR were focally or diffusely

immunostained. The membrane of the tumor cells was selec-

tively stained by anti-CD44v6 antibodies. The fibroblast and

infiltrating leukocytes in the stroma were not stained. The three

cases in the weak group on RT-PCR were focally immuno-

stained. In the moderate and strong groups, 14 cases were

focally immunostained, and 6 cases were diffusely stained (Fig.

6, D and F; Table 4). There was a correlation between the

immunoreactivity detected by anti-CD44s and v6 antibodies and

the expression of CD44s and v6 on RT-PCR.

DISCUSSION

CD44 is an integral membrane glycoprotein that possesses

a number of isoforms generated by alternative splicing of 10

variant exons. Several studies investigated the expression of

CD44s and variant exons in lung cancer (21, 22). We also

analyzed the expression of CD44s and variant exons in lung

cancer cell lines and specimens (data not shown). Our results

were similar to those of previous studies (2 1 , 22). These results

suggested that CD44s and variant exons were frequently cx-

pressed in NSCLC but were not expressed in small cell lung

cancers and that the metastasis of NSCLC may be related to

CD44 molecules. In this study, we focused on the expression of

CD44s and CD44v in NSCLC and examined the relationship

between CD44v6 expression and lymph node metastasis in

NSCLC.

Interestingly, CD44v exons are frequently expressed in

lung cancer but not in the normal lungs, whereas CD44s is

expressed to the same degree not only in normal lungs but also

in cancerous lungs. CD44v exons, except vl0, were not cx-

pressed or were weakly expressed in normal lung tissue, as

shown by RT-PCR, although 2 of 13 normal lungs very weakly

expressed CD44v6. In contrast, CD44v3, CD44vS, CD44v6, or

CD44v7 was expressed in 28 of 3 1 (90.3%) NSCLCs. This

study demonstrated that in two normal lungs that weakly cx-

pressed CD44v6, as shown by RT-PCR, some type II pneumo-

cytes were immunostained by anti-CD44v6 antibodies. Using a

different antibody than we did, Kasper et a!. (23) reported that

the CD44v6 isoform was detectable in type II pneumocytes and

in basal cells of the bronchial epithelium. Although some tissues

in normal lung express CD44v, as shown by IHC, the RT-PCR

conditions in this study detected little or no CD44v expression

in normal lungs. However, we detected variable CD44v expres-

sion in many of the NSCLCs. Thus, our results suggest that the

expression of CD44v on RT-PCR may be a useful marker for

the early diagnosis of primary NSCLC or of local recurrence.

NSCLC had a wide variety of CD44v expression patterns.

About 80% of NSCLCs expressed variant exons v3, v6, and/or

v7 variably. The frequency of CD44v5 expression in NSCLC

was less than that of the other variant exons. Conversely,

CD44vlO, which was one exon of the epithelial form, was

expressed in most NSCLCs (30 of 31 , 96.8%). It is known that

CD44 molecules have various functions that are attributed to the

various isoforms. Although no ligand has yet been identified for

the region encompassing variant exons vl-vlO, a hydrophilicity

plot of the CD44 molecule revealed that the hydrophilic area

was predominantly concentrated in variant exons v 1-v9. Be-

cause the hydrophilic regions of a protein usually face outward,

these regions are potential candidates for regions interacting

with ligands (24). This review leads us to the assumption that

NSCLC cells may have some advantages for growth, invasion,

and metastasis by gaining the multifunctions for cell-cell and/or

cell-matrix interactions induced by a wide variety of alternative

splicing patterns.

The frequency of v6 expression in NSCLCs with lymph

node metastasis was significantly more than that in NSCLCs

without lymph node metastasis [23 of 29 (79.3%) versus 13 of

32 (40.6%); P < 0.01]. Furthermore, there was a correlation

between the intensity of CD44v6 expression in NSCLCs and



Table 2 Correlation between N factor and the expression of CD44s, CD44v3, CD44vS. CD44v6, CD44v7, and CD44vIO by RT-PCR in 31

NSCLCs

CD44s” CD44v3” CD44v5” CD44v6 CD44v7 CD44v1O”

N factor +h + - + - + - + - + -

N0

N,

N2

N3

a NS, not significant. Fisher’s exact test.

b � expression by RT-PCR.C _� no expression by RT-PCR.

d Fisher’s exact test.

13 0 9 4 5 8 8 5� 8 5� 12 15

103

000

4103

100

241

362

5103

0 7 � p<o.oi” �0-4J 10

oJ 3

o-� � P<0.0l” s04-i 10oJ 3

o0

0

Table 3 Correlation between several factors and the expression ofthe CD44v6 in NSCLCs

a�1)

a

E

3

100

80

60

40

20

0

none weak moderate strong

a NS, not significant: t-test.

“ x2 test.C Welch test.


CD44v6

Positive Negative

Age (yrs: average) 65.5 ± 8.0 62.1 ± 9.0

NS”Gender

MaleFemale

28 (59.6%)

8(57.1%)

19 NS”6

BI (average) 835 ± 696 832 ± 416

NS’Stage

I 11(45.8%) 13

II 4(100%) 0

lIlA 12(60%) 8 NS”IIIB 5 (62.5%) 3IV 4(80%) 1

T factorT, 7(54%) 6T2 24(67%) 12

NS”T3 2 (33%) 4T4 3 (50%) 3

N factor

N0

N,

N,N3

13(40.6%)

7(100%)13 (68.4%)

3(100%)

190 ‘� bP<0.0l”

6 3 J0

lymph node metastasis. Reber et a!. (25) reported that the

application of the mAb 1 . 1ASML, which recognizes an epitope

of v6 in rats, to a highly metastatic rat cell line considerably

blocked metastasis in the lungs (25). cDNA transfection of

variant exons v4, v5, v 6, and v7 or v6 and v7 into nonmetastatic

cell lines conferred metastatic potential to these transfectants

(7), and the formation of metastasis in these transfectants could

be dramatically blocked with monoclonal antibodies

(1.IASML) when given before lymph node colonization (26).

This study focused on the relationship between v6 expression

and lymph node metastasis in NSCLC. Thus, our results suggest

that the expression of CD44v6 by RT-PCR may be a useful

marker for the diagnosis of NSCLC with lymph node metastasis.

Intensity of CD44v6 expression

Fig. 5 Correlation between the frequency of cases with lymph node

metastasis and the intensity of CD44v6. The number of cases with

lymph node metastasis increased in proportion to stronger CD44v6

expression. Cochran-Armitage’s test, P < 0.002. NIH image analysis:none, no expression: weak, <2000: moderate. �2000 and <5000:strong, �5000.

For instance, if a case is preoperatively diagnosed as NSCLC

and shows strong expression of CD44v6 on a biopsy specimen,

that case should undergo adequate dissection of the mediastinal

and hilar lymph nodes.

In lung cancer, Washimi et a!. ( I I ) reported no correlation

between the expression patterns of CD44v forms and nodal

involvement in NSCLC. The results of Washimi et a!. (1 1) do

not concur with our results. They used a probe corresponding to

CD44v4-v7 and demonstrated that not only NSCLC but also

normal lung tissue strongly expressed either CD44v4, CD44vS,

CD44v6, or CD44v7. In their study, the contamination with

normal cells may have caused false positive results. In contrast,

we demonstrated the specific expression of CD44v6 in NSCLC.

In our study, CD44v6 expression in normal lung tissue was

either nonexistent or very weak, as shown by RT-PCR. Only

tumor cells in NSCLC expressed CD44v6, but the fibroblasts

and infiltrating leukocytes in the stroma of NSCLC did not



_,f’_ � -

Bi4

F,

�. 0

� .�

� �� ,,,,�,

c�..., #{149} :- . �

-- . :- . � -#{149}‘ .,- � .;i’:’�

,,.‘ ,� � .. �-

-.

.--.f�,

‘ :�:�‘

,‘ - -,‘�

:‘j�: � � ,

�. ,t.’ . ..-‘.

,.�,3 #{149} 1�� #{149}� . _ ;

: � � �#{149}‘�?�/,

,�‘

.- - - ,-� ‘

� ..�-.-=.

t

- I’

. � � �-

�. --“.-. -

-3’� . - �

� -�

.-.‘ 1�

� � -- � - -. �

X�4�:’

� - .

I�

.-4

-F--

..� ._�,_4.

:�

�4:�� -�-

�

: �.

‘..4;y�

. �

. �

-�

�.

.� �p �

:.* ��“ ‘.

- ‘.,.“--#{176}.

-

-

) .- . . -

- -


-‘. � -

-‘i-. ;dT�

� .‘,‘-‘

�.# �‘ �� .

0 #{149} :::� �‘

. .4.‘#{149},.4,.#{149}. � -

E f�� ‘.

�?/- ‘- -

� 4 �i.

: �

4#{149} - ,,, �,

. . 0 �

4 ,_,.

Fig. 6 Immunohistochemical staining of normal lung and lung cancers was analyzed for the presence of CD44s and all variant exon-encoded epitopesusing mAb SFF-2 (A, C, and E) and for the presence of CD44v6 using mAb VFF-7 (B, D, and F). In the normal lung, all five tissues including typeI and II pneumocytes, interstitial tissues, and macrophages were stained by mAb SFF-2 (A). In NSCLCs, both tumor cells and interstitial tissues werestained by mAb SFF-2 (C and E). In two normal lungs with weak CD44v6 expression on RT-PCR, type II pneumocytes were focally stained by mAbVFF-7 (B). The membranes in the tumor cells were selectively stained by mAb VFF-7 in squamous cell carcinomas (D) and in adenocarcinomas (F)with CD44v6 expression on RT-PCR. Labeled streptavidin biotin method. A and B, normal lung, X400; C and D, squamous cell carcinoma, X200:

E and F, adenocarcinoma, X 200: counterstain, hematoxylin.

express CD44v6 on IHC, as shown in Fig. 6, D and F. If there Several studies have demonstrated that the expression of

was contamination with normal cells, our assay did not detect the v6 exon is correlated with the lymphatic metastatic potential.

the false positive. Therefore, we think that the results of In colorectal tumors, an epitope encoded by CD44v6 correlates

Washimi et a!. (I I) did not pinpoint CD44v6 expression in with tumor progression, being the highest in metastasizing tu-

NSCLC cancer cells and that their findings were influenced by mors (Dukes’ stages C and D) and has a poor prognosis (10). In

exons other than v6 and by normal components in lung cancers. breast cancers with lymph node metastasis, expression of 5ev-




9. Rudy, W., Hofmann, M., Schwartz, A. R., Zoller, M., Heider, K. H.,Ponta, H., and Herrlich, P. The two major CD44 proteins expressed on

Table 4 The expression of the CD44v6 by RT-PCR and IHCin NSCLCs

IHC”

Intensity of RT-PCR”

None Weak Moderate Strong

- 7 0 0 0+ 0 3 5 9

++ 0 0 3 3a Intensity quantified by NIH image as:none, no expression: weak,

<2000; moderate, �2000, <5000;strong, �5000.b no stain; +, focally stained (positive cells are less than 50% of

cancer cells); + +, diffusely stained (positive cells are more than 50% of

cancer cells).

eral splice variants was strong, but in tumors without lymph

node metastasis, signal intensity was weaker (2).

In conclusion, this study demonstrated that in NSCLC, a

variety of CD44v exons are frequently expressed, but the normal

lung infrequently expresses them, and that the expression of

CD44v6 is associated with lymph node metastasis. The func-

tions of the CD44v exons, including v6, need to be examined

further to elucidate the adhesive features of NSCLC. The dis-

covery of the CD44v6 ligand may also provide a strategy for

overcoming the metastasis of NSCLC cells in the lymph nodes.

ACKNOWLEDGMENTS

We thank Prof. T. Sano (First Department of Pathology, School of

Medicine, The University of Tokushima) for his pathological analysis

and helpful discussion.

REFERENCES

1. Tanabe, K. K., Ellis, L. M., and Saya, H. Expression of CD44RIadhesion molecule in colon carcinomas and metastases. Lancet, 341:

725-726, 1993.

2. Matsumura, Y., and Tam, D. Significance of CD44 gene productsfor cancer diagnosis and disease evaluation. Lancet, 340: 1053-1058,

1992.

3. Dall, P., Heider, K. H., Hekele, A., von Minckwitz, G., Kaufmann,

M., Ponta, H., and Herrlich, P. Surface protein expression and messen-ger RNA-splicing analysis of CD44 in uterine cervical cancer andnormal cervical epithelium. Cancer Res., 54: 3337-3341, 1994.

4. Pals, S. T., Koopman, G., Heider, K. H., Griffloen, A., Adolf, G. R.,van den Berg, F. M., Ponta, H., Herrlich, P., and Horst, E. CD44 splice

variants: expression during lymphocyte activation and tumor progres-sion. Behring Inst. Mitt., 92: 273-277, 1993.

5. Heider, K. H., Dammrich, J., Skroch, A. P., Muller, H. H., Vollmers,H. P., Herrlich, P., and Ponta, H. Differential expression of CD44 splice

variants in intestinal- and diffuse-type human gastric carcinomas and

normal gastric mucosa. Cancer Res., 53: 4197-4203, 1993.

6. Yokozaki, H., Ito, R., Nakayama, H., Kuniyasu, H., Taniyama, K.,and Tahara, E. Expression of CD44 abnormal transcripts in humangastric carcinomas. Cancer Lett., 83: 229-234, 1994.

7. Gunthert, U., Hofmann, M., Rudy, W., Reber, S., Zoller, M.,Haussmann, I., Matzku, S., Wenzel, A., Ponta, H., and Herrlich, P. A

new variant of glycoprotein CD44 confers metastatic potential to ratcarcinoma cells. Cell, 65: 13-24, 1991.

8. Herrlich, P., Zoller, M., Pals, S. T., and Ponta, H. CD44 splicevariants: metastases meet lymphocytes. Immunol. Today, 14: 395-399,1993.

a metastatic rat tumor cell line are derived from different splice variants:

each one individually suffices to confer metastatic behavior. Cancer

Res., 53: 1262-1268, 1993.

10. Wielenga, V. J., Heider, K. H., Offerhaus, G. J., Adolf, G. R., vanden Berg, F. M., Ponta, H., Herrlich, P., and Pals, S. T. Expression of

CD44 variant proteins in human colorectal cancer is related to tumorprogression. Cancer Res., 53: 4754-4756, 1993.

I 1. Washimi, 0., Ueda, R., Ariyoshi, Y., Suyama. M., Seki, T.,Takahashi, T., and Takahashi, T. Expression of CD44 variant isoforms

in normal and neoplastic cells of the lung. Jpn. J. Cancer Res., 85:

1112-1116, 1994.

12. Givehchian, M., Woerner, S. M., Lacroix. J., Zoller, M., Drings, P.,

Becker, H., Kayser, K., Ridder, R., and Doeberitz, M. K. Expression ofCD44 splice variants in normal respiratory epithelium and bronchial

carcinomas: no evidence for altered CD44 splicing in metastasis. On-

cogene, 12: 1137-1144, 1996.

I 3. Union Internationale Contre Cancer. Manual of Clinical Oncology.4th ed., pp. 181-189. Berlin: Springer-Verlag. 1987.

14. WHO. Histological Typing of Lung Tumours, 2nd ed., pp. 25-26.

Geneva: WHO, 1981.

15. Chomczynski, P., and Sacchi, N. Single-step method of RNA iso-lation by acid guanidinium thiocyanate-phenol chloroform extraction.

Anal Biochem., 162: 156-159, 1987.

16. Stamenkovic, I., Amiot, M., Pesando, J. M., and Seed, B. A lym-phocyte molecule implicated in lymph node homing is a member of thecartilage link protein family. Cell, 56: 1057-1062, 1989.

17. Screaton, G. R., Bell, M. V., Jackson, D. G., Cornelis, F. B.. Gerth,U., and Bell, J. I. Genomic structure of DNA encoding the lymphocyte

homing receptor CD44 reveals at least 12 alternatively spliced exons.

Proc. NatI. Acad. Sci. USA, 89: 12160-12164, 1992.

18. Hofmann, M., Rudy, W., Zoller. M., ToIg, C., Ponta, H., Herrlich,P., and Gunthert, U. CD44 splice variants confer metastatic behavior inrats: homologous sequences are expressed in human tumor cell lines.

Cancer Res., 51: 5292-5297, 1991.

19. Koopman, G., Heider, K. H., Horst, E., Adolf, G. R., Berg, F.,Ponta, H., Herrlich, P., and Pals, S. T. Activated human lymphocytesand aggressive non-Hodgkin’s lymphomas express a homologue of therat metastasis-associated variant of CD44. J. Exp. Med., 177: 897-904,1993.

20. Grube, D. Immunoperoxidase methods: increased efficiency using

fluorescence microscopy for 3,3 ‘-diaminobenzidine (DAB)-stained

semithin sections. Histochemistry, 70: 19-22, 1980.

21. Penno, M. B., August, J. T., Baylin, S. B., Mabry, M., Linnoila,R. I., Lee, V. S., Croteau, D., Yang, X. L., and Rosada. C. Expressionof CD44 in human lung tumors. Cancer Res., 54: 1381-1387,

1994.

22. Jackson, D. G., Schenker, T., Waibel, R., Bell, J. I., and Stahel,R. A. Expression of alternatively spliced forms of the CD44 extracel-lular matrix receptor on human lung carcinomas. Int. J. Cancer, 8

(Suppl.): 110-115, 1994.

23. Kasper, M., Gunthert, U., DalI, P., Kayser, K., Schuh, D., Haroske.0., and Muller, M. Distinct expression patterns of CD44 isoforms

during human lung development and in pulmonary fibrosis. Am. J.Respir. Cell Mol. Biol., 13: 648-656, 1995.

24. Chou, P. Y., and Fasman, G. D. Prediction of protein conformation.Biochemistry, 13: 222-244, 1974.

25. Reber, S., Matzku, S., Gunthert, U., Ponta, H., Herrlich, P., andZoller, M. Retardation of metastatic tumor growth after immunizationwith metastasis-specific monoclonal antibodies. mt. J. Cancer. 46: 919-927, 1990.

26. Seiter, S., Arch, R., Reber, S., Komitowski, D., Hofmann, M.,Ponta, H., Herrlich, P., Matzku, S.. and Zoller, M. Prevention of tumormetastasis formation by anti-variant CD44. J. Exp. Med., 177: 443-455,

1993.



1997;3:1289-1297. Clin Cancer Res T Miyoshi, K Kondo, N Hino, et al. lymph node metastasis in non-small cell lung cancer.The expression of the CD44 variant exon 6 is associated with

Updated version

http://clincancerres.aacrjournals.org/content/3/8/1289

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://clincancerres.aacrjournals.org/content/3/8/1289To request permission to re-use all or part of this article, use this link



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

mailto:[email protected]



Documents

The Expression of the CD44 Variant Exon 6 Is Associated with Lymph Node Metastasis … · Vol. 3. 1289-1297, August 1997 Clinical Cancer Research 1289 The Expression of the CD44 Variant