HYBRIDOMA Volume 9, Number s, 1990 Mary Ann Liebert, Inc., Publishers

Biochemical and Histological Characterization of Antigens Preferentially Expressed on the Surface and Cytoplasm of Breast Carcinoma Cells Identified by Monoclonal Antibodies Against the Human Milk Fat Globule JERRY A. PETERSON,' DAVID T. ZAVA,' AXEL K. DUWE,1 EDWARD W. BLANK,' HECTOR BATTIFORA,2 and ROBERTO L. CERIANI' 'John Muir Cancer and Aging Research Institute, 2055 N. Broadway, Walnut Creek, CA 2City of Hope National Medical Center, 1500 East Duarte Road, Duarte, CA

ABSTRACT The preparation of monoclonal antibodies (MAbs) against the human milk fat globule membrane with preferential binding to breast carcinoma cells is described. Using BALB/c mouse myeloma cells; inter-specific, intra-strain, and inter-strain hybridomas were isolated that identified three different components of the human milk fat globule of approximately 46,000, and 70,000 daltons and a mucin-like glycoprotein complex (NPGP) ranging from 400,000 to over a million daltons, respectively. Three MAbs (BrEl, BrE2, BrE3) identified the latter component which consists of at least three different size molecules for which the aforementioned MAb's have different binding specificities. MAbs, BrE2 and BrE3, bound to normal breast epithelial cells but to a lesser extent than to tumors and only at the apical surface facing the lumen, while they bound breast carcinomas strongly, and often in the cytoplasm as well as on the surface. Higher concentrations of BrE3 were required to stain normal breast compared to

breast tumors.

BrEl also stained breast carcinomas both

on

the surface and

cytoplasmically but did not stain normal breast tissue. The MAb, Mcl3, as well as the previously reported MAb McR2, both against the 70,000 dalton component, did not significantly stain either normal or cancerous breast tissue in histological sections but did bind significantly to cultured breast epithelial cells and to the milk fat globule membrane. The MAbs, Mc8 and Mc3, reported previously to be against the 46,000 dalton component, stained histologically only malignant breast tissue but only weakly; however, they bound strongly to intact breast carcinoma cells and breast cell membrane preparations with a radioimmunobinding assay. These MAbs should be useful in characterizing the surface of breast epithelial cells, studying surface alterations in malignancy, and possibly in breast cancer diagnosis and therapy. INTRODUCTION The human milk fat globule (HMFG) has been an excellent source of membrane material for the preparation of polyclonal (1) and monoclonal antibodies (MAbs) against surface antigens of normal and malignant human mammary epithelial cells. These antigens have found important use in breast cancer diagnosis 221

(1,2), breast cancer patient monitoring and follow-up (2,3), immunohistopathology (4), and treatment (5, 6). The first specific antisera against breast epithelial cells (1), using HMFG as the immunogen, were polyclonal antisera prepared in rabbits that after appropriate adsorptions with cells of other tissues were found to specifically identify breast epithelial cells in culture (7), in tissue sections (3), and permitted the discovery of these normal mammary epithelial antigens (HME-Ags) in sera of breast cancer patients (2). With the advent of hybridoma technology (8), we were the first to prepare MAbs that recognized a hitherto unidentified complex of large molecular weight mucin-like glycoproteins present on the surface of breast epithelial cells (9,10,11). MAbs to these mucin-like molecule(s) that we designated NPGP (9,11) were prepared by selecting for specificity for breast epithelial cell membrane compared to membranes from other cells such as breast fibroblasts and a lymphoblastoid cell line (10, 11). Since then, many investigators have prepared other MAbs, using both the human milk fat globule (10,11,12) and different breast tumor cell preparations (13-19) as immunogens. With an unusual regularity, MAbs were found to recognize large molecular weight glycoprotein(s) which have similar characteristics to what we refer to as the mucin-like glycoprotein complex NPGP. In fact, these MAbs were found to recognize a family of mucin-like glycoproteins that vary in size from about 250,000 daltons to over a million, depending on the antigen preparation (20). The frequency by which MAbs have been made against this molecule(s), using various normal and cancerous sources of immunogen, indicates that it is highly immunogeneic in the mouse, although its physiological function is not yet understood. In contrast to the BALB/c mouse, the rabbit polyclonal antisera prepared against HMFG and rendered specific for breast by adsorption with other tissues (anti-HME) does not appear to recognize this molecule as readily, but rather identifies primarily three components of the HMFG with smaller molecular weights: 150,000, 70,000, and 46,000, respectively (1). Still selecting for specificity to breast epithelial cells, a MAb, Mc3, was prepared that recognized the 46,000 dalton molecule, using the BALB/c mouse strain for immunization. By immunizing a rat, an interspecific hybridoma McR2 was selected that secreted a rat MAb against the 70,000 dalton molecule (11). However, obtaining the latter MAbs required screening thousands of hybridomas from many fusions while hybridomas for the mucin-like glycoprotein (NPGP) were found in almost every fusion. In an attempt to broaden the spectrum of MAbs to breast surface antigens, we have switched to a strain of mice that has a propensity for autoimmune disease (NZB strain) (21) using the BALB/c myeloma in the fusion. In this way, we have succeeded in obtaining MAbs that have a preferential binding to breast carcinoma cells, and which recognize new epitopes of the mucin-like molecule, as well as a The present communication describes new MAb against the 70 kDa glycoprotein. the antigens identified by these MAbs, the tissue and cell-type specificity of these MAbs, as well as of others previously reported (11) that recognize these breast associated components found in the human milk fat globule membrane. MATERIALS AND METHODS

Preparation of hybridomas. The preparation of the hybridomas was done with standard hybridization procedures using polyethylene glycol to aid in the cell fusion (11). Two different myeloma cell lines (P3-NS1/I-Ag.4-1 and P3-X63.Ag8.653) were used, both of which were deficient in hypoxanthine-guanine phosphoribosyl transferase, thus being unable to grow in HAT medium used for selection of the hybridomas (22). Myeloma cells were grown in growth medium made up of equal parts of RPMI 1640 and Dulbecco's modified Eagle's (DME) medium containing 10% fetal bovine The hybridomas were grown in the serum (FBS) and penicillin and streptomycin. same medium except the FBS concentration was increased to 20%. Cloning was done for the intrastrain hybridomas using spleen feeder cells for the BALB/c

222

compatible hybridomas, and for interstrain hybridomas, filtered medium

from two

day BALB/c spleen cultures was used. The cell fusion was done using 50% polyethylene glycol in RPMI 1640 medium and fused cells seeded in 6 to 10 96-well tissue culture plates and the next day HAT selective medium added.

The animals were immunized with whole delipidated human milk fat globule (HMFG). A primary injection of 50 \ig delipidated HMFG plus Freund's adjuvant was given intraperitoneally, the mice were boosted after three weeks with similar amounts also with adjuvant and then 3 or 4 days latter the spleens were taken for hybridization. The screening for selecting the hybridomas was done using a solid phase radioimmuno-plate binding assay or an ELISA assay with the HMFG, components thereof, and cell membrane preparations attached to 96-well microtiter plates as we have already described (11,23). The antibody isotypes were determined using a mouse immunoglobulin kit purchased from Zymed (South San Francisco, CA). Human cell lines used in this study were from breast carcinoma, MCF-7 (24), MDA-MB-157 (25), T47D (26); colon carcinoma, HT-29(27); cervical carcinoma, HeLa; lymphoma, Bris-8(28); and pancreatic carcinoma, PANC-K29). MAbs binding to intact cells in tissue culture. Cell lines were routinely cultured in DME or DME/RPMI-1640 medium (1:1) supplemented with 2-20% fetal bovine serum and penicillin/streptomycin. Cells

washed twice with 20 ml buffer saline with 0.025% EDTA, and incubated for 10 minutes at 37°C. Cells were detached, resuspended in growth media, and 1 ml 5 x IO4 cells/well). After seeded into each well of 24-multiwell plates (2 allowing cells to attach and grow for 2-5 days, they were tested for direct MAb binding activity. Supematants from the ascites of hybridomas, shown to contain high binding titers at IO-4 dilution, generally were used as a source of antibody. In some cases, MAbs were purified on protein-A-Sepharose-4B or on a hydroxyapatite column using HPLC. Ascites were diluted 1:100 in fresh culture media containing 1% sodium azide and 0.1 ml added, in triplicate, directly to intact cells growing in 1 ml of growth media. Cells were incubated with MAbs for 2 hours at 22°C in a chamber gassed with 5% Œ>2-air mixture. Following incubation, each well was rinsed 4 times with 1 ml DME medium containing a 5% bovine serum and 0.1% sodium azide. Growth media (1 ml) containing 2 x 10^ CPM F(ab)2 125i_g0at anti-mouse Ig (Tago, Inc., Burlingame, CA) were added to each well and incubated 2 hours at 22°C again in a gassed chamber. Radiolabeled antibodies were prepared using the chloramine-T method (30). Wells were rinsed 4 times with 1 ml of 5% growth medium, cells disolved in 0.25ml 0.1N NaOH, and the samples neutralized with 0.1N HCI and phosphate buffered saline. Aliquots were counted for radioactivity and DNA content was determined fluorimetrically by the method of Labarca and Paigen (31), using Hoechst 35258 dye. were

-

Molecular weight determination. Molecular weights for the antigens detected by each MAb were determined with polyacrylamide gel electrophoresis (PAGE) by a solid-phase binding assay previously described by Ceriani (23). Briefly, HMFG was electrophoresed on polyacrylamide gels according to Laemmli (32). Gel lanes were then sliced into small sections and components eluted from the slices in electrode buffer for 48 hours. Fifty lambda aliquots were then dried onto microtiter plates (Dynatech, Alexandria, VA) which had been coated with methylated bovine serum albumin and treated with glutaraldehyde (23). After blocking for 1 hour with 0.5% buffered glycine, wells were incubated overnight at room temperature with an appropriate dilution of MAb and extensively washed. 125i_iaDeiecj g0at anti-mouse IgG was then added for 3 hours and the wells washed again and then cut out and counted for radioactivity. Molecular weights were determined by comparison to standards run and stained in the same gel.

Immunohi stochemi strv. The MAbs

tested by immunoperoxidase techniques for binding to of normal and cancerous human tissues using multi-tumor (sausage) blocks (33). This novel method involves the preparation of tissue blocks that are a collection of strips of the various fixed tissues that are wrapped in peritoneal membrane or intestine that are embedded in paraffin and then sliced. About a thousand sections can be prepared from a single sausage which allows the assessment of MAb binding to a large number of different tissues (routinely 70-80) with a single staining (33). For example, a breast sausage used in these studies contained 21 different specimens from normal breast, 22 fibroadenomas, and 33"breast carcinomas. Undiluted tissue culture supematants of each hybridoma were used for staining, unless otherwise indicated. Tissues were scored as positive if any staining was observed that was clearly not background when compared to a similar section stained with culture supernatant of the myeloma cell line P3-X63.Ag8.653. Positive controls were MAbs on sections that they were previously shown to stain. were

histological sections

TABLE 1 Mouse monoclonal antibodies selected for specificity for breast prepared against human milk fat globule membranes.

epithelial cells

HMFG Strain

Hybridoma Myeloma Parent

Component

Antibody

BALB/c

NPGPb

Immunized*

Identified

Subclass

BrEl

P3/NSl/l-Ag

BrE2

P3-X63-Ag8.653

NZB

NPGP

BrE3

P3-X63-Ag8.653

NZB

NPGP

IgG2a IgG] IgG]

Mel 3

P3-X63-Ag8.653

NZB

70kDac

IgGi

Mc8

P3/NSl/l-Ag

BALB/c

46kDad

IgG]

aAll mice

4-1

4-1

immunized with delipidated HMFG (11). complex of mucin-like glycoproteins of HMFG with molecular weights from 250k to over a million daltons. c70kDa a component of HMFG of approximately 70k daltons. d46kDa a double component of HMFG of approximately 46k daltons. DNPGP

were

a

=

= =

RESULTS In the earlier efforts to prepare

hybridomas that secrete MAbs with

specificity for human breast epithelial cells (11), BALB/c mice were immunized with delipidated human milk fat globule, fused with the mouse myeloma P3/NSI/l-Ag4-l and screened initially for binding to HMFG and to membranes of two breast cell lines MCF-7 and MDA-MB-157, and then for no binding to membranes of colon carcinoma (HT-29), cervical carcinoma (HeLa) and lymphoma (Bristol-8) cell lines (11). Enriched membrane fractions were prepared as previously described (11). The majority of hybridomas prepared and selected in this way secreted MAbs against a large molecular weight mucin-like glycoprotein complex

(NPGP) (11). In one of the more recent fusions, using this same protocol, we have isolated a hybridoma (BrEl, see Table 1) that secreted MAbs that bound strongly to many breast carcinomas in immunohistopathology but did not bind to any of 22 normal breast tissue sections (Table 2). In a new series of fusions, using the autoimmune prone mouse NZB strain for immunization but using the same screening protocol, we have selected for characterization two new MAbs (BrE2 and BrE3) recognizing the breast mucin NPGP (Table 1). Also, with the BALB/c mouse strain for immunization and selecting with the 70,000 molecular weight component of the human milk fat globule recognized by the rabbit polyclonal anti-HME, we

isolated Mcl3, that recognized this component (Table 1). Recently we have demonstrated that 70kDa component is associated with the breast mucin NPGP by disulfide bonds indicating it to be possibly a linker protein (34). The molecular weights of the antigens identified by the MAbs were determined using PAGE and a solid-phase binding assay (23). The advantage of this technique is that it allows the analysis of several MAb binding profiles from a single polyacrylamide gel electrophoresis of an antigen preparation. When the HMFG was electrophoresed on a 7% polyacrylamide gel, the solid-phase binding assay showed that the MAbs, BrEl, BrE2, and BrE3, identified only high molecular weight material at the top of the gel (Fig. 1). This component stained strongly for carbohydrate but stained for protein only when the gel was heavily loaded (11). MAb Mel3 identified primarily the major 70kDa component of the human milk fat globule which was resolved into a doublet band and a minor binding in the region of 150kDa (Fig. 1). Mel3 must recognize a different epitope of the 70kDa component than the rat MAb McR2 reported previously, since the latter does not bind to a 150kDa component (11). MAb Mc8, that we have previously reported to be effective in MAb therapy of breast tumors in nude mice (5,6) and in detecting the 46kDa antigen and immune complexes in breast cancer patients (35), resolves a doublet band of the HMFG in the region of 46kDa (Fig. 1) similar to Mc3 reported previously (11). Since we (11) and others (12-19) have shown that NPGP is a complex of large mucin-like glycoproteins of at least 3 different components (20), we analyzed the three new MAbs against NPGP discussed above, and in addition, two that we have reported previously, Mc5 and Mel (10,11), for binding to HMFG that was 445

6691205 .9120

H

6691205

45

iU

M

116 97

66

45

Y

BrE

H

E

Q.



15

Mcl3 20-

BrE 3

10

30

40

Slice number

Antigen-binding profiles of mouse MAbs. Identical HMFG antigen generated by binding antigen eluted from slices of the same 7% PAGE to microtiter wells. MAbs were then incubated with each fraction and binding detected by addition of 125I labeled goat anti-mouse IgG. FIGURE 1.

profiles

were

225

20,000£

§

15,000-

53

§ 10,000-

o

o

5,000

20

10

30

40

FIGURE 2. Heterogeneity of the large molecular weight glycoprotein complex of NPGP. Identical HMFG antigen profiles were generated by binding aliquots of fractions eluted from slices of a 3% polyacrylamide-1% agarose gel in which HMFG MAbs were incubated with each fraction was electrophoresed to microtiter wells. attached to a solid phase, and binding of the MAbs detected by addition of 125I goat anti-mouse Ig. Abscissa slice number. =

electrophoresed on a 3% polyacryamide-1% agarose gel (Fig. 2). Four of these MAbs (Mel, Mc5, BrE2, BrE3) identify a major peak at fraction 10 of the gel and then a broad shoulder of smaller sized molecules (Fig. 2). Each seems to have a somewhat different profile, indicating that they are possibly identifying different epitopes. We consider the different profiles significant since they In contrast, BrEl were obtained on the same series of HMFG profile plates. binds to only a single component in the region of the larger size molecule (Fig. 2). The distinct profile of BrEl is especially significant because of the results of the immunohistopathology (see below). In order to characterize these MAbs with regard to tissue specificity, they were tested for binding to various normal and malignant tissues, using immunoperoxidase techniques. Initially, they were screened on a collection of normal breast, breast carcinomas, and fibroadenomas. Usually, when no binding was detected in this initial screen, further screening was not done. As a general rule, strong staining was obtained with all the MAbs against NPGP on at least some normal or malignant breast tissue (Tables 2 and 3). The MAbs (Mc3 and Mc8) that identified the 46kDa component, which have not previously been analyzed in histopathology, stained some breast tumors, but always weakly and in Neither stained normal breast nor any a diffuse cytoplasmic manner (Table 4). other non-breast tissues, normal or tumor. The MAbs that identified the 70kDa component, McR2 (11) and Mcl3 did not stain either normal breast tissue, or fibroadenomas. However, McR2 did stain 2 of 48 breast carcinomas but only with ethanol fixation (Table 4). Since the most significant histological staining was obtained with the MAbs that identified NPGP, they were screened more extensively. Mel, Mc5, and BrE2 stain virtually all breast tumors whether the tissues are fixed by formalin, ethanol or stained as frozen sections (Tables 2 and 3). The other major tumors of which a large percentage were stained by Mel and Mc5 were mesothelioma, and adenocarcinoma of lung, endometrium, cervix, ovary, pancreas, and stomach. As reported previously for Mel [formerly 3.14.A3 (10), and also referred to as HMFG-2 (21)], the tissue specificity of the MAb to NPGP suggests that the antigen is related to a secretory function, where it is present on the apical surface of the normal counterpart. As with all the MAbs against NPGP that we 226

TABLE 2

Staining distribution in histological sections of monoclonal antibodies against a breast mucin-like glycoprotein (NPGP), using Balb/c mice for immunization. Antigen

Mo Ab

Tissue Distribution

NPGP

Mel

Normal tissues: stains breast (apical), distal tubules of kidney, alveoli of lung, acini of pancreas.

breast, 48/48a, 155/163b, 40/40c, 3/5b; colon, 0/6a, l/5b; duodenum, 0/la, l/2b; ear, 0/1; endomedetrium, 3/3a, 4/5b; kidney, 2/8a, 0/3b; liver, 2/3a, 2/4b; larynx, 0/2; liver, 2/3b; lung, 6/7a, 6/6b; meningioma, 0/lb: ovary, 8/8a, 6/8b; pancreas, 2/2a, 2/2b; prostate, 0/2b, l/3b; salivary gland, 6/9b; stomach, 5/12b, thyroid, 0/2a, 2/3b; mesothelioma, 3/10b. Carcinomas of the:

cervix, 3/3a,

NPGP

Mc5

Normal tissue: breast (apical), upper layer of esophagus, distal tubules only of kidney, lung epithelial and ducts, acini and islets of pancreas, and thyroid.

breast, 48/48a, 50/51b, 40/40c; Cervix, 2/3a, 3/4b; colon, 5/14b,; endometrium, 12/13b; kidney, 5/15°; liver, 2/3a; 2/4b; lung, 15/15b, l/lc; ovary, 15/17b; parathyroid, 0/la, 0/2b; prostate, 1/15°; ret. embryonal, l/la, l/lb; stomach, 7/9a, 8/12b; salivary, l/2b; salivary adenoma, 0/5b; thyroid, 6/6b; mesothelioma, 5/10b; Warthin's tumor, 2/2b; pancreatic ll/14b; undiff. ca. 10/29. Carcinomas of the:

NPGP

BrEl

Normal tissues: does not stain normal breast, 0/22b; stains alveolar lining of lung but not bronchus, brush border of renal tubules. Tumors:

See Table 5.

aethanol fixation.

bformalin

fixation.

cfrozen section

have tested, when they stain normal breast tissue they show only surface binding. However, in tumors, they can exhibit both cytoplasmic and surface staining, as has been reported for other MAbs that identify these mucin-like

molecules in breast carcinomas (13). The MAbs, Mel, BrE2, and BrE3 were similar in the spectrum of normal tissues that they stained. They stained all normal breast tissues tested, and of the non-breast tissues, they stained lung, kidney, and pancreas but were negative on adrenal, prostate, brain, bladder, colon,

lymph node, spleen, myocardia, testis, parathyroid, esophagus, muscle, stomach, thyroid, ovary, mesothelia, and liver. MAb Mc5 also stained normal breast tissue strongly, and of the non-breast tissue, it stained lung, kidney, esophagus, pancreas and thyroid but was negative on stomach, brain, bladder, muscle, skin, small intestine, and spleen. MAb BrEl was tested on a large panel of different tumors, because of the fact that it showed no binding to normal breast tissues, and of the other normal tissue, it bound only to lung epithelium and kidney tubules (Table 2). It did not stain adrenal, brain, bladder, colon, liver, lymph node, myocardium, pancreas, spleen, stomach, thyroid, testis, or ovary. It stained many breast adenocarcinomas but no adenomas, and the major non-breast tumors it stained were adenocarcinomas of ovary, endometrium, lung, and pancreas (Table 5). For most

TABLE 3 Tissue distribution in histological sections of monoclonal antibodies against the breast mucin-like glycoprotein (NPGP), using NZB mice for immunization.

Antigen NPGP

MoAb BrE2

Tissue Distributiona Normal tissue: breast, 6/11; renal tubules.

lung,

pancreas,

stomach, and

breast, 50/51; embryonal, 0/1; stomach, 8/13; cervix, 4/4; lung 54/58; pancreas, 4/17; endometrium, 11/12; ovary, 20/21; liver, 23/46; salivary gland, 7/12; colon, 23/32; thyroid, 0/4; parathyroid, 0/1; prostate, 1/1; larynx-pharynx, 5/5; pleura, 2/7; peritoneum, 3/4; skin, 1/2; kidney, 6/14; soft tissue, 14/37; nasopharyngeal, 2/4; rectum, 1/2; gallbladder, 2/4; melanoma, 0/3; neuroblastoma, 1/1. Tumors:

NPGP

Normal tissue: stains breast weakly, alveolar lining of lung, distal convoluted tubules of kidney, acini of pancreas,

BrE3

stomach

mucusa.

breast, 47/54; lung, 47/48; 6/9; endometrium, 9/9; mesothelioma, kidney, 6/13; liver, 3/28; lymphoma, pancreas, 12/15; rhabdosarcoma, 0/2; sarcomas, 8/43; thyroid, 3/9. Tumors:

23/25; bladder, 8/20; colon, 12/25; 0/20; merkel cell, 6/8; salivary gland, 3/9; ovary,

aformalin fixation.

TABLE 4

Staining distribution in histological sections

of monoclonal antibodies the 70kDa and 46kDa components of human milk fat globule membrane.

Antigen 70kDa

MoAb McR2

Component

Tissue Distribution Normal tissue: all negative

including

against

breast.

Carcinomas: only 2/48 breast tumorsa. All carcinomas of other tissues were negative, cervix, colon, duodenum, endometrium, kidney, liver, lung, ovary, pancreas, parathyroid,

prostate, stomach, thyroid.

70kDa

Mcl3

Negative

46kDa

Mc3

Normal tissue:

Component

46kDa

Component

for normal breast, breast

fibroadenomas.13

all

carcinomas, Component and

negative4,b

weak staining on breast tumors, 24/49a, 14/55°, 6/40c; negative on all other carcinomas tested: cervix, colon, duodenum, endometrium, kidney, liver, lung, ovary, pancreas, parathyroid, prostate, stomach, thyroid.

Carcinomas:

Mc8

Normal tissue: all negative3,b Tumors: similar staining as Mc3.

aethanol fixation.

Dformalin fixation.

cfrozen section

X* CO

CO I

-*

CD CO

CO

u_

co en CM

CM CM

r—

CM CM

r~

i i

CO «f I—

Ol

*v "V \ \ -^ \ ooooomoo O O O O O i—

CO

I

I u_

i—i

i—i

X > CD CO

X I

>

o

i

CM

•i-

O en

i—

CO CO

in

O O ul r*

r—

cm

c

10 «+J IO ¡A +J

r—

o

O" Il E i- > 0J -

__

E

c _:

-

>> u •r-

0J +J

-s

s

0)

o

I/»

E

(O

E

IO

10

E

Or— O Tuo

-i-

l/l

£

10 =>

O

E O

o m

+j i.

£_ll>l_l_00_.0_-CnO E

"O +->__>,f.T3er-aJ.r-l.+J+J_:-. O O • _3 OUOliOO_>>OiO>i/lf-f--- r- Q.r— ig f0) (0 _= •S-OI.O->Cr-Cr->,___tT+J->US-0JS._.>« "O Í-

E_J

> n-

_ZO--M

> •!-

U

1/1

Q.I— 11 aj O II x o> c - IO

>»_. c:

•

o en

aj _= s. c ia i/i >f— O (/) __

•r-_j>l__.x:i- c —IO_-l/lWh CU aj

r—

m

o> c C O

IO

E 0J _.

ia >>-o TO o

T3 C IO C O

< <

_.

"O l/l 01 >l "O S-UI/lf-ES-r—

o -o O oj aj a.r-

U-C_J__I-I_J________

lO l/l

r-

OÍ

C IO

01 CT > _.

C_ IO

II

c: cvi •r-

0J 0J IO S. +J +J OÍ oj m >i 3

i/i

t/i

10 l/l i.

f-

o o oír+j +-> _= c c aj 10 0J a. 10 -ro

•!-

>, 0J

•I-

T3

I

>> c

O o

l/l

>>

ÍÍ

__

O O w CO CO C_

UU

II

_3

X

-x (_>-

i—i

oo

IO O IO +- u-i o ai

__

x

>í

J3 >> E s. O IO aj aj aj s. E >

a>

I— 1_J|—

_)

i—

OÍ ¡31 O

—

CMinincO^*CMC001ri£> i— CM «J_

o

o

__

01 !->

IO

M-

IO

E CQ

__

o c 10 l/l E s. •

3 O

-=

O

i

l/> i-i

E -= > oo

_-

r—

cnOr— mmr-^-vO

r-

CM

_

_

CO

k.



E

3

0) _.

00 OÍ CM p- CO

O)

i—

0OCOCM

__onirir--J-cOr-o

o +-> o

-í

CM

t—

>

aj o

O) 10

B O

l/l

O. IO

IO

E E

T3

o o o r- r- co o to ao i— 0(vj_(.i_-r^-i—

co

i_> p~

f-

cm

i—

IO 4->

_.

o

o

£_

•o

Oi+J

10

10

0J

IO l/l

"O 0J OJ t-