(CANCER RESEARCH 46, 4904-4910, October 1986]

Antibody-directed Liposomes: Comparison of Various Ligands for Association,Endocytosis, and Drug Delivery1

Katherine K. Matthay,2 Timothy D. Heath,3 Christopher C. Badger,4 Irwin D. Bernstein, and

Demetrios PapahadjopoulosDepartments of Pediatrics ¡K.K. M.J and Pharmacology [T. D. H., D. P.] and the Cancer Research Institute fK. K. M., T. D. H., D. P.], University of California, SanFrancisco, California 94143, and the Fred Hutchinson Cancer Research Center [C. C. B., l. D. B.J, University of Washington, Seattle, Washington 98104

ABSTRACT

We have developed and compared the cytotoxicity of methotrexate-T-aspartate encapsulated in several liposome formulations which bindmouse monoclonal antibody in order to define conditions for screeningcell lines and antibodies for liposomal efficacy. Liposomes conjugated toStaphylococcus aureus Protein A were more potent than liposomes conjugated to either rabbit or affinity-purified goat anti-mouse immunoglob-ulin (Ig) when incubated with AKR/J SL2 cells sensitized with specificantibody. The antibody-directed Protein A liposomes were also 10-foldmore potent than liposomes conjugated directly to specific antibodyagainst the AKR/J SL2. We examined the effect of antibody specificity,concentration, and isotype on liposome-mediated drug delivery to AKR/J SL2 cells. The growth-inhibitory effect of the drug in the antibody-directed Protein A liposomes varied with the target antigen on the cell.The potency of the liposomes with a given antibody was proportional totheir relative binding and endocytosis by the cells, and to the reactivityof the particular antibody with the cell as demonstrated by indirectimmunofluorescence. The Protein A liposomes maintained maximal potency down to antibody concentrations as low as 1 MR/mlwith the anti-Thy 1.1-sensitized AKR/J SL2 cells, thus demonstrating the possible useof these liposomes for hybridoma screening. Use of isotype-switchedvariants of the anti-Thy 1.1 antibody with the AKR/J SL2 cells showedthe superior efficacy of the IgG 2a, IgG2b, and IgG3 isotypes to the IgGlwith the Protein A liposomes. The large differential potency of the freedrug and the drug encapsulated in antibody-directed Protein A liposomeswas maintained even at short incubation times, thus providing a systemwhich may be useful for eradication of tumor cells from bone marrow invitro.

INTRODUCTION

Drug delivery to tumor cells with antibody-directed liposomes is a promising method for increasing the specificity andefficacy of cancer therapy (1-5). We have previously demonstrated the specific cytotoxicity of antibody-conjugated liposomes containing methotrexate-7-aspartate for several cell lines(5,6). We and others have also shown that optimal drug deliveryto lymphoid cell lines is achieved with liposomes that are 0.05to 0.07 /¿min diameter (4, 5, 7).

As a logical extension of our previous work, we wished todevelop a method of testing small amounts of monoclonalantibodies rapidly for their use in directing specific liposomecytotoxicity. Moreover, we wanted a technique that would allowthe use of unpurified ascites or culture supernatants. Direct

Received 1/13/86; revised 6/10/86; accepted 6/13/86.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1Supported in part by USPHS Grant CA39448 (K. K. M.) awarded by theNational Cancer Institute, Department of Health and Human Services; Universityof California, San Francisco, Academic Senate grant (K. K. M.); and Universityof California, San Francisco, Faculty Development award (K. K. M.).

2 Recipient of American Cancer Society Junior Clinical Faculty Fellowship723 and a Leukemia Society Special Fellowship. To whom requests for reprintsshould be addressed, at Cancer Research Institute, 1282 M, University of California School of Medicine, San Francisco, CA 94143.

3 Present address: School of Pharmacy, University of Wisconsin, Madison, WI

53706.4 Recipient of an American Cancer Society Junior Faculty Clinical Fellowship.

conjugation of antibodies to liposomes requires the use of atleast several milligrams of purified material. Leserman et al.have successfully used liposomes conjugated to Staphylococcusaureus Protein A for delivery of methotrexate to cells (8). Inthis system, the cells are preincubated with small amounts ofmonoclonal antibody, washed, and then incubated with liposomes. The liposomes bind indirectly to the cells via the interaction of the Protein A with the Fc portion of the cell-associatedantibody. To allow use of the indirect targeting system inmedium containing heterologous immunoglobulin, anti-mouseimmunoglobulin could be substituted for Protein A to directliposomes to cells specifically sensitized with mouse antiserum.

In this paper, we have investigated the delivery of methotrex-ate-7-aspartate to cells by a variety of different indirect ligands.We find that Protein A is by far the most effective indirectligand, and we have extensively examined the variation in itspotency with antibody concentration, specificity, and isotype.We have also utilized the dissociation of Protein A from antibody at low pH to determine the extent of endocytosis of cell-associated liposomes.

MATERIALS AND METHODS

Phospholipids were obtained from Avanti (Birmingham, AL) andwere used without further purification. The MPB-PE5 was synthesized

and purified as described (9). Cholesterol was recrystallized 4 timesfrom methanol. All lipids were stored at —80°Cin chloroform solution

under argon in sealed ampuls until use. Methotrexate-7-aspartate wassynthesized and kindly provided by Dr. J. R. Piper (Southern ResearchInstitute, Birmingham, AL) (10). S. aureus Protein A was purchasedfrom Sigma (St. Louis, MO). Rabbit anti-mouse IgG and goat anti-mouse IgG were obtained from Miles (Naperville, IL); affinity-purifiedgoat anti-mouse IgG was from Kierkegaard-Perry (Gaithersburg, MD).Radiolabeled [3H]DPPC, ['"CJsucrose (673 mCi/mmol), and [3H]inulin(164 mCi/g) were obtained from New England Nuclear. [3H]DPPC waspurified by thin-layer chromatography within 1 mo prior to use.

Monoclonal Antibodies. Monoclonal antibodies were prepared andpurified as previously described (6). Hybridomas SS.l (IgG2a, anti-sheep erythrocyte) and 11-4.1 (IgG2a, anti-H-2Kk) (11) were supplied

by Dr. M. Cohn of the Salk Institute, San Diego, CA. Hybridomas19E12 (IgG2a, anti-Thy 1.1) (12), 16B7 (IgG2a, anti-gp70), and19VIIIE8 (IgG2b, anti-plSE) were provided by Dr. R. C. Nowinski andDr. M. Lostrom of Genetic Systems, Seattle, WA (13). MRCOX7(IgGl, anti-Thy 1.1) was provided by Dr. A. Williams, University ofOxford, Oxford, United Kingdom (14). RIO (IgGl, anti-human glyco-phorin) was provided by P. A. W. Edwards, Department of Pathology,Cambridge University, Cambridge, United Kingdom (15). The isotypeswitch variants of the anti-Thy 1.1 antibody were produced as described(16).

Cell Lines. The murine lines, L929, and the T-cell leukemias, Rl.l(17), a Thy 1.1-negative cell, and AKR/J SL2 (18), a Thy 1.1-positive

5The abbreviations used are: MPB-PE, 4-(p-maleimidophenyl)-butyryl-phosphatidylethanolamine; SPA-liposomes, Staphylococcus aureus Protein Aconjugated to small unilamellar vesicles; 1C.,,, concentration of drug required toachieve a 50% reduction in cell growth; DPPC, dipalmitoylphosphatidylcholine;gp70, M, 70,000 glycoprotein; pl5E, M, 15,000 protein; MES/HEPES buffer,50 mM 2-(/V-morpholino)ethanesulfonic acid/50 mM W-2-hydroxyethylpipera-zine-./v"-2-ethanesulfonic acid/2 min EDTA.

4904

on April 18, 2018. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ANTIBODY-DIRECTED LIPOSOMES

cell, were maintained as previously described (5, 6). The human T-cellleukemias, MOLT 3 and MOLT 4, were provided by Dr. M. Cowan atthe University of California, San Francisco, CA, and the CEM line wasfrom Dr. B. Sikic at Stanford University, Palo Alto, CA; all weremaintained in RPMI with 10% fetal calf serum in a 7% CO2 atmosphere. The human chronic myelogenous leukemia, K562, was maintained as previously described (1).

Liposome Preparation. A mixture of phosphatidylcholine, cholesterol, and MPB-PE (10:5:1 molar ratio) was suspended in a solution of50 mM methotrexate-7-aspartate as previously described (5). The suspension was sonicated for l h in a bath sonicator (Lab Supplies,Hicksville, NY). Larger liposomes were removed by centrifugation in aBrinkmann Eppendorf 3200 centrifuge for 30 min, and unencapsulateddrug was removed by gel filtration on Sephadex G-75. The vesicleswere then conjugated for 18 h at 25°Cwith thiolated 5. aureus Protein

A (8) or immunoglobulin (9) in isotonic MES/HEPES buffer (pH 6.7).The protein and lipid concentrations during conjugation were, respectively, 0.25 g/liter and 2.6 mM. The conjugated vesicles were separatedfrom unbound protein by flotation on a metrizamide gradient using amodification of our previous method (19). Briefly, the vesicles in bufferwere mixed with metrizamide to a final concentration of 200 g/liter.Three ml of a metrizamide solution (100 g/liter) were carefully layeredover 1.5 ml of the vesicle-metrizamide (200 g/liter) solution in 5-mlcellulose nitrate tubes. The top was overlaid with 0.5 ml of MES/HEPES buffer, and the mixture was centrifuged at 35,000 rpm for 16h in a Beckmann L3-50 ultracentrifuge. Liposomes were sterilized bypassing them through a 0.2 /jm polycarbonate filter and analyzed forlipid, protein, and drug content as previously described (6). Fourseparate SPA-liposome batches had a mean protein:lipid ratio of 40 ±14 g/mol and a drug:lipid ratio of 0.016 ±0.003 mol/mol. Theprotein:lipid ratio for the goat anti-mouse conjugated liposomes was100 g/mol. Liposome size was estimated on a Coulter Model N4submicron particle analyzer (20). The mean liposome diameter was0.056 ±0.013 urn.

Cell Association. Cells were first incubated with the desired antibody(50 Mg/ml) at 0°Cfor 30 min. The unbound antibody was removed with

a single wash, and the cells were then incubated with the liposomes.Vesicles were labeled either by incorporation of the lipid label, [3H]-DPPC, or by encapsulation of an aqueous phase marker, [l4C]sucroseor [3H]inulin. All association experiments were done in duplicate, using2 x IO6cells per sample in a total volume of 0.2 ml of growth medium.

Unbound liposomes were removed at the end of the incubation periodby layering the incubation mixture over 2 ml of 10% (w/v) dextransolution in phosphate-buffered saline and pelleting the cells throughthe dextran (5). In certain experiments, cell surface-bound liposomeswere removed by incubation of the cells in isotonic sodium citratebuffer, pH 3, for 5 min, followed by a second pelleting through dextransolution. In this way, Protein A liposomes which were bound to surfaceantibody but not yet internalized could be removed and differentiatedfrom liposomes that had undergone endocytosis. In all experiments,calculations were based on the number of cells present at the beginningof the experiment. It was assumed that cell loss would be equivalent incontrols and experimental samples. The SD was less than 10%.

Cell Growth Inhibition. Cells were treated with the desired antibodyat 5 x IO6 cells/ml for 30 min at 0°C.They were washed once andplaced in 24-well Costar plates at 2 x IO5 cells/ml, 1 ml/well. They

were treated in triplicate with the desired concentration of drug orliposomes and then counted with a Model Fn Coulter Counter after 48h of incubation. The percentage of growth was calculated with thefollowing equation.

Final cell no. - initial cell no.Control cell no. - initial cell no. x 100

The lC5o is determined from a plot of logic drug concentration versuspercentage of growth.

RESULTS

Relative Potency of Various Targeted Liposome Preparations.Fig. \A shows the relative potency of the 5. aureus Protein A

100

50

* 02 100

50

x i x

0.01 0.1 1.0

DruftuMFig. 1. Growth-inhibitory effect of various types of antibody-directed lipo

somes on AKR/J SL2. Cell growth was calculated from triplicate measurementsof cell number compared to control cell number after a 48-h incubation in thepresence of free methotrexate--y-aspartate or drug-containing liposomes. Drugconcentration is shown on a logarithmic scale. 1,comparative growth inhibitionby SPA-liposomes with specific and nonspecific antibody, x, free methotrexate-7-aspartate; •,SPA-liposomes plus cells pretreated with specific anti-Thy 1.1antibody (19E12); O, SPA-liposomes plus cells pretreated with a nonspecificantibody (SS.l). B, comparative growth-inhibitory effect of three different typesof antibody-directed liposomes, x, free methotrexate--x-aspartate: O, liposomesdirectly conjugated to anti-Thy 1.1 antibody (19E12); Q, liposomes directlyconjugated to affinity-purified goat anti-mouse immunoglobulin plus cells pretreated with specific anti-Thy 1.1 antibody: •,SPA-liposomes plus cells pretreatedwith specific anti-Thy I. I antibody.

liposomes (SPA-liposomes). Free methotrexate-7-aspartate hasan ICso of 0.78 MMwith the AKR/J SL2 cells. Methotrexate-7-aspartate encapsulated in SPA-liposomes has an IC50 of 0.008MMwith AKR/J SL2 cells pretreated with anti-Thy 1.1 antibody, but only 0.27 MMwhen the AKR/J SL2 cells are pretreated with a nonspecific antibody (SS.l, anti-sheep erythro-cyte). The SPA-liposomes with the specific antibody are 150-fold more potent than the free drug and 50-fold more potentthan they are with the nonspecific antibody. The mean IC50 forthe SPA-liposomes with the anti-Thy 1.1-coated AKR/J SL2cells was 0.012 ±0.004 (SD) MMfor 15 separate experimentsdone in triplicate using 4 separate liposome preparations. Themean IC50 for the free methotrexate-7-aspartate for the sameexperiments was 0.856 ±0.327 MM,approximately 70-fold lesseffective than the encapsulated form of the drug.

Fig. IB compares the relative potency of methotrexate-7-aspartate encapsulated in SPA-liposomes to liposomes directlyconjugated to anti-Thy 1.1 antibody and to liposomes conjugated to affinity-purified goat anti-mouse immunoglobulin. Thetwo indirect liposome preparations were tested on AKR/J SL2cells pre-treated with anti-Thy 1.1 antibody. The SPA-liposomes are clearly the most potent of the three preparations,showing an IC50 9-fold lower than either the direct anti-Thy1.1-liposome conjugate or the anti-mouse immunoglobulin-liposome conjugate. All three preparations are more potentthan either free drug or than nonspecific liposomes. However,a third indirect system involving liposomes conjugated to rabbitanti-mouse immunoglobulin was no more potent than free drug(not shown), presumably because this preparation of antiserumcontained less effective ligand than the affinity-purified anti-serum.

SPA-liposomes have been used by Machy and Leserman (4)

4905

on April 18, 2018. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ANTIBODY-DIRECTED LIPOSOMES

to deliver methotrexate to cells. However, they have not previously been used to deliver methotrexate--y-aspartate, nor has

their potency been compared to liposomes prepared with othertargeting methods. Therefore, we chose to investigate SPA-liposomes in more detail. The percentage of the SPA-liposomesthat associate with AKR/J SL2 cells is 5 to IS times greaterwhen the cells are sensitized with anti-Thy 1.1 than it is whenthey are incubated with a nonreactive antibody (Fig. 3). Thisdifference in cell association correlates qualitatively with therelative potency of the drug-loaded SPA-liposomes.

The effect of time of exposure of the AKR/J SL2 cells to theSPA-liposomes was examined by growth inhibition studies inthe following manner. AKR/J SL2 cells were coated as usualwith anti-Thy 1.1 antibody and treated with the desired concentration of liposomes. The liposomes were removed by washingat timed intervals over the 48-h incubation. Fig. 2 shows theincrease in efficacy that occurs with increasing time of exposureof the cells to the liposomes. However, it also demonstratesthat, even with a 2-h exposure, a 120-fold enhancement ofpotency compared to free methotrexate-7-aspartate occurs. Theenhancement is 250-fold after 12 h of exposure, 90-fold at 24h, and 70-fold at 48 h of exposure.

Potency of SPA-Liposomes with Different Cell Lines. TheSPA-liposomes were also tested for their cytotoxicity to othercell lines using the appropriate antibody (Table 1). R 1.1 T-celllymphoma cells showed marked growth inhibition when incu-

100

10

0.1

0.01

12 24 36 48

Exposure time, hrFig. 2. Growth-inhibitory effect of varying exposure time of AKR/J SL2 cells

to SPA-liposomes. AKR/J SL2 cells pretreated with specific anti-Thy 1.1 antibody were exposed to methotrexate-7-aspartate (x) or SPA-liposomes (•)for thedesignated period of time, washed, and then incubated for the remainder of the48-h period prior to counting the cells. 1C»was determined as described in"Materials and Methods."

Table I Efficacy of antibody-directed SPA-liposomes with various cell lines

Antibody concentration equals SOMg/ml.

CellAKR/JSL2

Rl.lL929CEMMOLT 3MOLT 4K562OriginHi

rnmhhhhSpecific

monoclonal19E12

11-4.111-4.1Leu 1OKT 11OKT3RIOFree

drug0.78

0.581.000.290.760.700.601C»

(MM)Nonspecific

antibody+ SPA-

liposome0.27

0.440.300.280.220.220.16Specific

antibody+ SPA-

liposome0.005

0.0030.0230.210.170.200.16

bated with SPA-liposomes directed by anti-H-2Kk (11-4.1monoclonal) antibody. The SPA-liposomes were again morethan 100-fold more potent than unencapsulated methotrexate-•y-aspartateand 90-fold more potent than SPA-liposomes directed by nonspecific antibody. The SPA-liposomes were 50-fold more potent than free methotrexate-7-aspartate againstL929 fibroblasts coated with anti-H-2Kk antibody and 3-fold

more potent than the direct conjugate against this cell line. TheSPA-liposomes directed by specific antibody were not morepotent than nonspecific SPA-liposomes for growth inhibitionof the human T-cell leukemias, CEM, MOLT 3, and MOLT 4,and a human leukemia line derived from a patient with chronicmyelogeneous leukemia in blast transformation, K562. It ispossible that differential susceptibility of human cell lines tothe drug may influence their response to the antibody-directedliposome-encapsulated drug. However, the differences in freedrug ICso among cell lines were small and not sufficient toexplain the discrepant targeting results of the AKR/J SL2 cellscompared to the human T-cell leukemias. It is more likely thatthe failure of specific enhancement in the human cell lines isdue to inadequate cell association of the SPA-liposomes withthe cells pretreated with the antibodies tested. Table 2 showsthat the antibody-directed SPA-liposomes labeled with [3H]-inulin bound specifically to the cells in 5-fold greater amountthan cells treated with nonspecific antibody. However, theabsolute quantity of liposomes bound per cell was 8- to 10-foldlower than with the AKR/J SL2 cells and anti-Thy 1.1 antibody,where 25% of the liposomes became cell associated at equivalentlipid concentrations.

Effect of Antibody on S l'A-1 ,¡posómePotency. We wished to

gain greater insight into the factors affecting the role of antibody in promoting drug delivery. Therefore, we examined thegrowth-inhibitory potency of SPA-liposomes with four differentantibodies against three different antigenic sites on the AKR/JSL2 leukemia cells: 19E12 (anti-Thy 1.1, IgG2a); MRC OX7(anti-Thy 1.1, IgGl); 16B7 (anti-gp70, IgG2a); and 19VIHE8(anti-pi5E, IgG2b). Table 3 summarizes the comparative efficacy of the various antibodies with the SPA-liposomes. Cellswere exposed to the respective antibodies at concentrations of100 fig/ml to ensure saturation of cell surface antigenic sites.The most potent antibodies for sensitizing the cells to liposome-mediated drug delivery were 19E12 and 16B7. Both MRC OX7and 19VIIIE8 were 7- to 10-fold less potent than 19E12 and16B7. The effect of a combination of the three antibodies,19E12, 16B7, and 19VIHE8, results in slight enhancement ofpotency over any single antibody. Indirect immunofluorescencemicroscopy of AKR/J SL2 cells with antibodies against thethree separate sites showed the greatest fluorescence with19E12 and MRC OX7, intermediate fluorescence with 16B7,

Table 2 Association of SPA-liposomes with human T-cell leukemiaLeukemia cells (2 x IO7/ml) were sensitized with monoclonal antibody,

washed, and then incubated 2 h with [}H]inulin-labeled SPA-liposomes (25 «IM).After removal of unbound liposomes, cell-associated radioactivity was measured.

" m, murine; h, human.

Cell-associatedliposomesCellCEMMOLT

3MOLT

4Antibody

SpecificityLeu-1

+Leu-9 +SS.1OKT11

+SS.1OKT3

+SS.lLipid

(nmol)0.1136

0.14490.00870.1781

0.04450.0724

0.0295%of

total2.2

2.90.23.6

0.91.4

0.6

4906

on April 18, 2018. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

sANTIBODY-DIRECTED LIPOSOMES

and the least with 19VIIIE8. The potency of 19VIIIE8 may below because the level of binding to cells, shown by immunoflu-

orescence, is low. However, the poor potency of OX7 cannotbe explained in this way and led us to examine the effects ofantibody isotype on SPA-liposome-mediated drug delivery.

The effect of antibody isotype on liposome efficacy and cellassociation was examined using a panel of anti-Thy 1.1 antibodies of different isotype but apparently identical bindingregions. Anti-Thy 1.1 antibody of isotypes IgGl, IgG2a, andIgG2b was derived from the parent hybridoma 31C5.31 producing IgG3 (16). These antibodies were used to coat the AKR/J SL2 cells prior to SPA-liposome treatment (Table 3). At bothantibody concentrations, the IgG2a, IgG2b, and the IgG3 wereapproximately 15-fold more effective than the IgGl for sensi-tization of the cells. Cell association with the SPA-liposomesagreed with growth inhibition results, with excellent associationfor the IgG2a, IgG2b, and IgG3 and lower association for theIgGl (Fig. 3). These results are consistent with the aboveobservations of cell association and cytotoxicity using SPA-liposomes with 19E12 and MRC OX7. Moreover, these resultsare consistent with the known affinity of Protein A for thevarious murine IgG subclasses (21). Protein A is known to bindweakly to IgGl, which would therefore be a less effectiveantibody than the other subclasses for cell sensitization.

Our major interest in SPA-liposomes was their possible use

Table 3 Effect of monoclonal antibody isotype and antigenic specificity on SPA-liposomes potency against AKR/J SL2

Monoclonal"MRC

OX719E12I6B719VIIIE8Combination*31C5.31rIA12

1A1431ASS.1Free

drugIso

typeIgGl

IgG2aIgG2aIgG2bIgG3

IgG2bIgG2aIgGlIgG2aAntigenThy

I.IThy1.1

gp70pl5EThy

1.1Thy 1.1Thy 1.1Thy1.1Sheep

RBC1C»(UM)0.042

0.0050.0050.0520.0030.011

0.0130.0120.160.270.74

" Antibody concentration, 100jjg/ml.* 19E12, 16B7, and 19VIIIE8 together at a concentration of 100 Mg/ml each.'31C5.31, 1A12, 1A14, and 31A are anti-Thy 1.1 antibodies of varying

isotypes but identical binding sites.

20

O(0a

10CO'5o«009

Õ<S o

125 250

Liposomes added,Fig. 3. Effect of antibody isotype on association of SPA-liposomes with AKR/

J SL2. Cells were preincubated with anti-Thy 1.1 antibody of varying isotypesbut apparently identical binding sites, washed, and then incubated with the I'll]inulin-labeled SPA-liposomes for 30 min. Cell-associated liposomes were measured as before, x, IgG2b (1A12); •,IgG2a (1A14); O, IgG3 (31C5.31); A, IgGl(31 A); a nonspecific antibody, IgG2a (SS.l).

for screening many different monoclonal antibodies for maximal efficacy with liposomes. An important concern for such anapplication is the amount of antibody needed to sensitize thecells. The growth inhibition of AKR/J SL2 is shown in Fig. 4for several concentrations of various antibodies reactive withAKR/J SL2 cells. The IC50 of the SPA-liposomes is fairlyconstant as the anti-Thy 1.1 antibody concentration is variedbetween 10 and 100 ng/m\ (Fig. 4, A and C). However, as onewould expect, the SPA-liposomes are more effective with theIgG2a antibody, 19E12, than with the IgGl, MRC OX7. Asignificant drop in the potency of the SPA-liposomes occurswith 19E12 concentrations below 1.0 ng/mi. The other IgG2aantibody, 16B7 (Fig. 4Ä), requires higher concentrations formaximal effectiveness, in the range of 30 ng/m\. The IgG2banti-pi5E antibody, 19VIIIE8 (Fig. 40), must be present at100 ííg/mlto produce any significant effect with the SPA-liposomes. Fig. 5 shows in more detail the effect of antibodyconcentration, examined by holding SPA-liposome concentration constant and measuring growth inhibition at various levelsof antibody. Specific growth inhibition of the AKR/J SL2 cells

100

so

0100

50

19E12

MRCOX7

16B7

19VIIIE8

0.01 0.1 1.0 0.01 0.1 1.0

Drug.uW

Fig. 4. Effect of antibody concentration on growth inhibition of AKR/J SL2by SPA-liposomes. Growth inhibition was measured as before, but preincubatingthe cells at varying concentrations of antibody. The concentration (jig/ml) of theantibody is shown next to its respective curve. A, 19E12 (anti-Thy 1.1, IgG2a); II.16B7 (anti-gp70, IgG2a); C, MRC OX7 (anti-Thy 1.1, IgGl); D, 19VIIIE8 (anti-pi 5E,IgG2b).

100

50

0.1 10 100

Antibody, ug/mlFig. 5. Antibody titration of growth-inhibitory effect of SPA-liposomes on

AKR/J SL2 with 19E12 and 16B7. Cells were preincubated with varying concentrations of specific antibody, then incubated for 48 h at a constant SPA-liposomeconcentration (0.03 *JM),and counted as before. 0,19E12 (anti-Thy 1.1); •,16B7(ami-gp70).

4907

on April 18, 2018. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ANTIBODY-DIRECTED LIPOSOMES

does not occur with the two IgG2a antibodies below an antibodyconcentration of 0.1 Mg/ml. Growth inhibition using 19E12increases with antibody concentration to 1.0 Mg/ml and thenremains fairly constant from 1 to 100 ng of antibody per ml.Growth inhibition with 16B7 increases gradually between 0.1and 30 ng of antibody per ml and then remains constant.

Endocytosis of SPA-Liposomes. The binding of Protein A tomurine IgG occurs at pH 7.0 and above and is reversed byexposure of the complex to lower pH (21). The phenomenonprovides an opportunity to study the extent of endocytosis ofcell-associated SPA-liposomes. Exposure of the cells to pH 3.0buffer should rapidly elute surface-bound liposomes, and theresidual cell-associated material should reflect the extent ofendocytosis. Several different liposome labels were used forindicators of cell association: [3H]DPPC, a phospholipidmarker; [l4C]sucrose, an aqueous phase marker; and [3H]inulin,

another aqueous phase marker. Liposomes labeled with any ofthese molecules showed a similar degree of cell association, butdiffering degrees of elution at pH 3 (Table 4). The liposomeslabeled with [3H]DPPC were only partially eluted from the cellsurface by low pH, even from cells that had been kept at 0°Cto inhibit endocytosis. [l4C]Sucrose was unsatisfactory as a

marker because it leaked rapidly from the liposomes and therefore could not be used to demonstrate endocytosis by the cellsover time. Repeated Chromatographie separations of the[l4C]sucrose SPA-liposomes showed that 30% of the radioactiv

ity was not liposome associated in material that had previouslybeen separated from unencapsulated sucrose and stored inbuffer at pH 7.4. The [3H]inulin-containing liposomes provided

a marker that did not leak appreciably and was useful formeasuring endocytosis. After a 2-h incubation at 37°C,25% of

the added liposomes were cell associated, while 8% were internalized, that is, remained cell associated after low pH elution.In contrast, only 8% of the liposomes became cell associated at0°C,and only 1.8% internalized. Timed intracellular measurement of [l4C]sucrose-labeled SPA-liposomes compared to[3H]inulin-labeled liposomes showed failure of accumulation ofthe [l4C]sucrose label over time, again apparently because of

rapid leakage of the label from the liposome or the cells, the[3H]inulin-containing liposomes were therefore used in the sub

sequent studies.The uptake of the [3H]inulin-labeled SPA-liposomes by

AKR/J SL2 cells sensitized with anti-Thy 1.1 was examinedover a 12-h period, by comparing the cell-associated liposomesbefore and after low pH elution (Fig. 6). At 0.5 h and at 2 h,approximately 24% of the liposomes were cell associated, whileless than 6% were internalized, that is, remained after elution.By 8 to 12 h, internalization reached a maximum of 12.8%.The overall cell association and degree of uptake were considerably lower for the other three antibodies against the AKR/JSL2 (Fig. 6), in proportion to their lesser efficacy for drug

Table 4 Effect of SPA-liposome label choice on measurement of cell associationand endocytosis

SPA-liposomes i Id f.\i lipid) labeled with the indicated radioactive tracercompound were incubated for 2 h with AKR/J SL2 cells (2 x 107/ml) sensitizedwith anti-Thy 1.1 antibody (19E12). Free liposomes were removed, and cell-associated liposomes were counted both before and after low pH elution to removeliposomes associated but not internalized.

Incubationtemperature

CO037PH3.07.43.07.4Cell-associated

SPA-liposomes(%)[3H)DPPC5.07.014.517.5[MC]Sucrose0.78.42.722.1[3H]Inulin1.88.18.025.3

Õo

20

10

I '•O o

10

:[•,[11ini

rii

i1

t_19E1216B7l

l19VIIIE81CMRCOX7i

fnA!\\\~i61

iC

iaD1

H0.5 2 4 B 12

Exposure time, hrFig. 6. Effect of different specific antibodies on internalization of SPA-

liposomes by AKR/J SI 2. Cells were pretreated with specific antibody, washed,and then incubated with the [3H]inulin-labeled liposomes (10 /IM) for the designated period of time. Internalized liposomes (¡T])were measured by the radioactivity which remained cell associated after low pH elution; total cell-associatedliposomes (O) were measured by the cell-associated radioactivity as the percentageof total added radioactivity prior to elution. .1, 19E12 (anti-Thy 1.1), IgG2a; H.16B7 (anti-gp70), IgG2a; C, 19VIHE8 (anti-plSE), IgG2b; D, MRC OX7 (anti-Thy 1.1), IgGl.

COOA•oco**auoco<r>

i

30201010lgG2a-

-iIAlgG1

B,a

B S B B0.5 2 6 8

Exposure time, hrFig. 7. Effect of antibody isotype on internalization of SPA-liposomes by

AKR/J SL2. Cells were preincubated with specific antibody of different isotypesbut identical binding sites as before, washed, and then incubated with the | '111inulin-labeled SPA-liposomes (10 *IM)as above. Internalized liposomes (Lì)wereagain compared to total cell-associated liposomes (O) by comparing cell-associated radioactivity before and after low pH elution. A, IgG2a (anti-Thy 1.1); A,IgGl (anti-Thy 1.1).

delivery. The difference in efficacy of the IgGl antibody fromthe other isotypes is further emphasized by the marked difference in endocytosis of the SPA-liposomes over time with theclass-switched anti-Thy 1.1 IgG2a and IgGl (Fig. 7).

Fig. 8 shows the extent of internalization as a function oflipid concentration and demonstrates that endocytosis is notsaturable until high lipid concentrations are reached. Theselipid concentrations are much higher than those used in thegrowth inhibition experiments, where at a drug concentrationof 0.1 MM,the lipid concentration would be only 5 MM(corresponds to 1 nmol of added lipid in Fig. 8).

DISCUSSIONSPA-liposomes containing methotrexate--y-aspartate provide

a versatile and highly effective method of /// vitro tumor cell4908

on April 18, 2018. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ANTIBODY-DIRECTED LIPOSOMES

2.0

1.0

0.6

60 100

Lposomes added, nmol

Fig. 8. Saturation of endocytosis of liposomes with increasing lipid concentration. [3H]Inulin-labeled SPA-liposomes were incubated for 2 h with AKR/J SL2cells that were pre-coated with specific anti-Thy 1.1 antibody (19E12). Internalized liposomes, represented by the shaded curve, were compared to total cell-associated liposomes as in Fig. 7. •,spécifie(I9EI2) antibody; O, nonspecific(SS. 1) antibody.

growth inhibition. These investigations form the basis for an invitro model for bone marrow purging by establishing the efficacyof different antibodies for endocytosis, the time necessary foradequate endocytosis of liposomes, and the necessary concentration of liposomes at shortened incubation times for growthinhibition.

The SPA-liposomes containing methotrexate-7-aspartatemay be suitable for screening large numbers of cell lines fortheir sensitivity to liposome-mediated drug delivery. SPA-liposomes inhibit growth in a target-dependent fashion withseveral different cell lines. SPA-liposomes may also be usefulfor screening large numbers of hybridomas for their potency inpromoting liposome-mediated drug delivery. SPA-liposomesreact with cells coated with small amounts of antibody, whichwill permit testing of hybridoma cell culture supernatants orascites for potency in liposomal drug delivery.

The antibody-directed SPA-liposomes are 9 times more effective against the AKR/J SL2 cells than liposomes conjugatedto anti-Thy 1.1 or to anti-mouse immunoglobulin. The difference between antibody-conjugated liposomes and SPA-liposomes was not so marked in two other cell lines tested usingH-2Kk as the target antigen. L929 fibroblasts and R 1.1 lym-phoma cells showed only a 2- to 3-fold advantage of the SPA-liposomes. The liposomes conjugated to affinity-purified goatanti-mouse immunoglobulin are potentially more versatile thanthe SPA-liposomes, since they will react with more subclassesof antibody. However, this versatility must be weighed againstthe increased cost of the antibody reagent and the lower efficacyagainst some cell types. There are several possible reasons forthe potency of the SPA-liposomes and for the variability withdifferent cells and antigenic targets. The association of thedirect conjugate and the SPA-liposomes with the AKR/J SL2cells is equivalent. Therefore, the difference in potency mustinvolve the endocytosis or subsequent processing of the liposomes. We have previously shown the importance of size foreffective drug delivery (5, 7). SPA-liposomes would be moreeffective than antibody-conjugated liposomes if they were lessprone to aggregation, a phenomenon which we have observedwith antibody-conjugated liposomes (9). Moreover, the dissociation of SPA-liposomes from their antibody-antigen receptorat low pH may be important for the efficient internalizationand processing of the vesicles. The level of expression of thetarget antigen is an important determinant of potency, as shownby the studies with the AKR/J SL2 cells with anti-Thy 1.1compared to anti-pi5E antibody. The lower level of targetantigen expression may also explain the lower potency of the

SPA-liposomes with the human leukemia cells. An additionalvariable of the target antigen may be its rate of internalization(22, 23). The cell type is another variable in the events leadingto liposome-mediated drug delivery, as shown by our previousstudies of optimal liposome size for drug delivery (5, 7).

The SPA-liposomes are useful for studying the sequentialuptake of liposomes by the cells, because Protein A dissociatesfrom immunoglobulin at low pH. We demonstrated the effective use of the aqueous phase marker, pHjinulin, for the investigation of cell association and endocytosis. [l4C]Sucrose was

not a good marker of endocytosis due to excessive leakage fromthe liposome. [3H]DPPC, the lipid marker, was also inappro

priate as a measure of endocytosis of the liposomes, probablydue to lipid exchange with the cell membrane, resulting inapparently nonelutable liposomes under conditions whichshould prohibit endocytosis, such as low temperature. This lipidhas previously been shown to exchange with the cell membranein studies at 2°Cusing nonconjugated liposomes (24). Theassociation studies with the [3H]inulin showed a temperaturedependence of both cell association and endocytosis of the SPA-liposomes. These results are similar to the results of Huang etal. (3) using antibody-conjugated liposomes with a nonex-changeable lipid label (25) and protease elution.

The timed studies of association and endocytosis show that,by 8 h, 0.1710 nmol of lipid per 2 x IO6 cells is internalized(lipid concentration, 10 /¿M).This corresponds to 2.06 x 10~18

mol of methotrexate-7-aspartate per cell. A comparable leukemia cell, 1,1211).has been shown to contain approximately 0.5x 10~18mol of dihydrofolate reducíaseper cell (26). Therefore,

the amount of antifolate internalized is approximately 4 timeshigher than the amount required to saturate the dihydrofolatereductase of the cell. This should be sufficient for growthinhibition and shows a reasonable correlation between cellassociation and drug delivery studies.

Total cell association of liposomes reaches a plateau after 4to 8 h. However, continuous uptake of liposomes from thesurrounding medium appears to be necessary for optimal drugdelivery, because in wash studies the IC50 of the liposomepreparation falls substantially between 8 and 48 h of exposure.It is possible that accelerated leakage of drug from the liposomes occurs when they attach to the cells (27-29). Theseliposomes would then contain less drug and would lose potencyunless they were constantly replaced by fresh liposomes fromthe surrounding medium. Such leakage induced by heat-aggregated immunoglobulins (30, 31) and antigen-complexed im-munoglobulins (32) has been shown to occur in multilamellarliposomes.

In conclusion, the highly potent SPA-liposomes may providea system for removal of residual tumor cells from autologousbone marrow prior to transplant. In such cases, it is essentialto destroy tumor cells while not harming normal hematopoieticstem cells. The ease with which this system can be used withseveral antibodies is also an advantage, because combinationsmay allow greater cytotoxicity toward heterogeneous populations of tumor cells. In a growth inhibition assay, a combinationof antibodies can be more effective. Future studies using aclonogenic assay should reveal whether liposomes are trulycytotoxic and whether more clonogenic tumor cells are eliminated by targeting against multiple antigens.

ACKNOWLEDGMENTS

We thank A. Abai for liposome sizing and R. Antonucci for secretarial assistance. We thank Dr. B. Sikic, Dr. N. Berenstein, and Dr. M.

4909

on April 18, 2018. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ANTIBODY-DIRECTED LIPOSOMES

Cowan for the human T-leukemia cell lines and appropriate monoclonalantibodies.

REFERENCES

1. Bragman, K. S., Heath, T. I)., and Papahadjopoulos, D. Cytotoxicity ofantibody-directed liposomes that recognize two receptors on KS62 cells. J.Nati. Cancer Inst., 73:127-131, 1984.

2. Hashimoto, Y., Sugawara, M., Masuko, T., and Hojo, H. Antitumor effectof actinomycin D entrapped in liposomes bearing subunits of tumor-specificmonoclonal immunoglobulin M antibody. Cancer Res., 43:5328-5334,1983.

3. Huang, A., Kennel, S. J., and Huang, L. Interactions of immunoliposomeswith target cells. J. Biol. Chem., 258:14034-14040, 1983.

4. Machy, P., and Leserman, L. D. Small liposomes are better than largeliposomes for specific drug delivery in vitro. Biochim. Biophys. Acta, 730:313-320,1983.

5. Matthay, K. K.. Heath, T. D., and Papahadjopoulos, D. Specific enhancementof drug delivery to AKR lymphoma by antibody-targeted small unilamellarvesicles. Cancer Res., 44: 1880-1886, 1984.

6. Heath, T. D., Montgomery, J. A., Piper, J. R., and Papahadjopoulos, D.Antibody-targeted liposomes: increase in specific toxicity of methotrexate--y-aspartate. Proc. Nati. Acad. Sci. USA, 80: 1377-1381, 1983.

7. Heath, T. D., Lopez, N., and Papahadjopoulos, D. Effects of liposome sizeand surface charge on liposome-mediated delivery of methotrexate-i-aspar-tate to cells in vitro. Biochim. Biophys. Acta, «20:74-84, 1985.

8. Leserman, L. D., Barbet, J., and Kourilsky, F. Targeting to cells of fluorescentliposomes covalently coupled with monoclonal antibody or Protein A. Nature(Lond.), 288:602-604, 1980.

9. Martin, F. J., and Papahadjopoulos, D. Irreversible coupling of immunoglobulin fragments to preformed vesicles. J. Biol. Chem., 257: 286-288,1982.

10. Piper, J. R., Montgomery, J. A., Sirotnak, F. M., and Chello, P. L. Synthesisof a- and ^-substituted amides, peptides, and esters of methotrexate and theirevaluation as inhibitors of folate metabolism. J. Med. Chem., 25: 182-187,1982.

11. Oi, V. T., Jones, P. P., Coding, J. W., and Herzenberg, L. A. Properties ofmonoclonal antibodies to mouse Ig allotypes, H-2 and la antigens. Curr.Top. Microbio!. Immunol., 81: 115-120, 1978.

12. Houston, L. L., Nowinski, R. C., and Bernstein, I. D. Specific in vivolocalization of monoclonal antibodies directed against the Thy 1.1 antigen.J. Immunol., 725: 837-843, 1980.

13. Lostrom, M. E., Stone, M. R., Tarn, M., Burnette, W. N., Pinter, A., andNowinski, R. C. Monoclonal antibodies against murine leukemia viruses:identification of six antigenic determinants on the pI5(E) and gp70 envelopeproteins. Virology, 98: 336-350, 1979.

14. Mason, D. W., and Williams, A. F. The kinetics of antibody binding tomembrane antigens in solution and at the cell surface. Biochem. J., ¡87:I20. 1980.

15. Edwards, P. A. W. Monoclonal antibodies that bind to the human erythro-cyte-membrane glycoproteins glycophorin A and Band 3. Biochem. Sue.Trans., 8:334-335,1980.

16. Denkers, E., Badger, C. C., Ledbetter, J. A., and Bernstein, I. D. Influenceof antibody isotype on passive serotherapy of lymphoma. J. Immunol., 135:2183-2186, 1985.

17. Hyman, R., Ralph, P., and Sarkar, S. Cell-specific antigens and immunoglobulin synthesis of murine myeloma cells and their variants. J. Nati. CancerInst., 48: 173-184, 1972.

18. Nowinski, R. C., Hays, E. F., Doyle, T., Linkhart, S., Medeiros, E., andPickering, R. Oncornaviruses produced by murine leukemia cells in culture.Virology, 81: 363-370, 1977.

19. Heath, T. D., Macher, B. A., and Papahadjopoulos, D. Covalent attachmentof immunoglobulins to liposomes via glycosphingolipids. Biochim. Biophys.Acta, 640:66-81, 1981.

20. Selser, J. C., Yeh, Y., and Baskin, R. J. A light-scattering characterizationof membrane vesicles. Biophys. J., 16: 337-356, 1976.

21. Ey, P. L., Prowse, S. J., and Jenkin, C. R. Isolation of pure IgGl, IgG2a,and IgG2b immunoglobulin from mouse serum using Protein A-Sepharose.Immunochemistry, 15:429-436, 1978.

22. Leserman, L. D., Machy, P., and Barbel, J. Cell-specific drug transfer fromliposomes bearing monoclonal antibodies. Nature (Lond.), 293: 226-228,1981.

23. Machy, P., Barbet, J., and Leserman, L. D. Differential endocytosis of T-and B-lymphocyte surface molecules evaluated with antibody-bearing fluorescent liposomes containing methotrexate. Proc. Nati. Acad. Sci. USA, 79:4148-4152, 1982.

24. Pagano, R. E., Sandra, A., and Takeichi, M. Interactions of phospholipidvesicles with mammalian cells. Ann. NY Acad. Sci., 308: 185-199, 1978.

25. Pool, G. L., French, M. E., Edwards, R. A., Huang, L., and Lumb, R. H.Use of radiolabeled hexadecyl cholesteryl ether as a liposome marker. Lipids,77:448-452, 1982.

26. Sirotnak, F. M., and Donsbach, R. C. Kinetic correlates of methotrexatetransport and therapeutic responsiveness in murine tumors. Cancer Res., 36:1151-1158,1976.

27. Scherphof, G., Roerdink, F., Hoekstra, D., Zborowski, J., and Wisse, E.Stability of liposomes in presence of blood constituents: consequences foruptake of liposomal lipid and entrapped compounds by rat liver cells. In: G.Gregoriadis and A. C. Allison (eds.), Liposomes in Biological Systems, pp.179-207. London: John Wiley & Sons, Ltd., 1980.

28. Szoka, F. C., Jr., Jacobson, k.. and Papahadjopoulos, D. The use of aqueousspace markers to determine the mechanism of interaction between phospholipid vesicles and cells. Biochim. Biophys. Acta, 551: 295-303, 1979.

29. Van Renswonde, A. J. B. M., Westenberg, P., and Scherphof, G. In vitrointeractions of Zajdela ascites hepatoma cells with lipid vesicles. Biochim.Biophys. Acta, 55«:22-40, 1979.

30. Schieren, H., Weissmann, G., Seligman, M., and Coleman, P. Interactionsof immunoglobulins with liposomes: an ESR and diffusion study demonstrating protection by hydrocortisone. Biochem. Biophys. Res. Commun., 82:1160-1167, 1978.

31. Weissmann, G., Brand, A., and Franklin, E. C. Interaction of immunoglobulins with liposomes. J. Clin. Invest., 53: 536-543, 1974.

32. Tagesson, C., Magnusson, K., and Stendahl, D. Physicochemical consequences of opsonization and perturbation of liposomal membranes by Salmonella typhimurium 395 MS. J. Immunol., 119:609-613,1977.

4910

on April 18, 2018. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1986;46:4904-4910. Cancer Res   Katherine K. Matthay, Timothy D. Heath, Christopher C. Badger, et al.   for Association, Endocytosis, and Drug DeliveryAntibody-directed Liposomes: Comparison of Various Ligands

  Updated version

  http://cancerres.aacrjournals.org/content/46/10/4904

Access the most recent version of this article at:

   

   

   

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

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

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

  Permissions

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

.http://cancerres.aacrjournals.org/content/46/10/4904To request permission to re-use all or part of this article, use this link

on April 18, 2018. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from