Lysosomal Trafficking and Cysteine Protease Metabolism Confer Target-specific Cytotoxicity by Peptide-linked Anti-CD30-Auristatin Conjugates*

The chimeric anti-CD30 monoclonal antibody cAC10, linked to the antimitotic agents monomethyl auristatin E (MMAE) or F (MMAF), produces potent and highly CD30-selective anti-tumor activity in vitro and in vivo. These drugs are appended via a valine-citrulline (vc) dipeptide linkage designed for high stability in serum and conditional cleavage and putative release of fully active drugs by lysosomal cathepsins. To characterize the biochemical processes leading to effective drug delivery, we examined the intracellular trafficking, internalization, and metabolism of the parent antibody and two antibody-drug conjugates, cAC10vc-MMAE and cAC10vc-MMAF, following CD30 surface antigen interaction with target cells. Both cAC10 and its conjugates bound to target cells and internalized in a similar manner. Subcellular fractionation and immunofluorescence studies demonstrated that the antibody and antibody-drug conjugates entering target cells migrated to the lysosomes. Trafficking of both species was blocked by inhibitors of clathrin-mediated endocytosis, suggesting that drug conjugation does not alter the fate of antibody-antigen complexes. Incubation of cAC10vc-MMAE or cAC10vc-MMAF with purified cathepsin B or with enriched lysosomal fractions prepared by subcellular fractionation resulted in the release of active, free drug. Cysteine protease inhibitors, but not aspartic or serine protease inhibitors, blocked antibody-drug conjugate metabolism and the ensuing cytotoxicity of target cells and yielded enhanced intracellular levels of the intact conjugates. These findings suggest that in addition to trafficking to the lysosomes, cathepsin B and perhaps other lysosomal cysteine proteases are requisite for drug release and provide a mechanistic basis for developing antibody-drug conjugates cleavable by intracellular proteases for the targeted delivery of anti-cancer therapeutics.

elimination of antigen-positive target cells. The first such clinically approved agent, Gemtuzumab ozogamicin (Mylotarg), an anti-CD33 mAb linked to the potent DNA damaging agent calicheamicin, is used for the treatment of patients with relapsed acute myeloid leukemia (1). Mylotarg and multiple other mAb-drug conjugates (ADCs) and mAbtoxin conjugates currently in development use primarily disulfide or hydrazone linkers sensitive to the reductive or acidic environment of the tumor cell (2)(3)(4)(5)(6). These linkers readily deliver and liberate free drug or toxin within the target cell and yet are relatively unstable in circulation compared with the circulating half-life of the mAb (7), resulting in premature drug release.
An alternative approach is to employ ADC linkers of protease-cleavable dipeptides. These combine qualities of high stability in serum or plasma with efficient drug release potentially by lysosomal proteases (8). Using this strategy, we recently described a new ADC of the anti-CD30 mAb cAC10 (9), appended to the anti-tubulin agent, monomethyl auristatin E (MMAE), via a cathepsin-cleavable valine-citrulline (vc) linker (10,11). This drug linker system was shown to be highly stable in vitro and in vivo (7,12), and when applied to multiple mAbs, the resulting ADCs were selectively potent and effective against cognate antigen-positive tumor cells and tumor xenografts (LeY (8), CD30 (11), TMEFF2 (13), CD20 (14), and EphB2 (15)). One premise of this drug delivery technology is that the mAb-antigen complex on the cell surface will internalize, traffic to the lysosomes, and be metabolized by lysosomal proteases to release free drug.
ADC efficacy therefore depends in part on mAb-antigen interaction at the cell surface triggering the internalization, trafficking, and subsequent release of the active cytotoxic payload. Thus, conjugates comprised of different drug linkers or with different mAbs against the same target can vary significantly in their utility. For example, anti-CD20 ADC incorporating doxorubicin (16,17) or ricin-A (18) were ineffective as anti-tumor agents, whereas anti-CD20 conjugates of vcMMAE were highly effective against CD20 ϩ tumors (14). Interestingly, the anti-CD20 mAb remained on the cell surface, whereas the anti-CD20vc-MMAE conjugate was readily internalized, resulting in rapid cell cycle arrest and apoptosis (14). Alternatively, in targeting CD30, both mAb and anti-CD30 ADC demonstrated comparable binding and internalization rates in CD30 ϩ tumor cells (11), with the ADC inducing rapid mitotic arrest cell and apoptosis (11). Just what governs optimal ADC internalization and trafficking critical for effective drug release is not known.
Here we examine the trafficking and intracellular fate of anti-CD30 mAb cAC10 and two cAC10-drug conjugates, cAC10vc-MMAE and cAC10vc-MMAF, in CD30 ϩ T-cell lymphoma cell lines. Using flow cytometry, immunofluorescence, subcellular fractionation, and chemical inhibitors of trafficking and processing, we demonstrate that the cytotoxicity of peptide-linked auristatin ADCs is contingent upon their delivery to the lysosome and the activity of lysosomal cysteine proteases. These studies demonstrate the functional basis of drug delivery by ADCs with protease-cleavable linkers for anti-cancer therapeutics.

MATERIALS AND METHODS
Flow Cytometry for Antibody and Antibody-Drug Conjugate Internalization and Trafficking-The chimeric anti-CD30 antibody cAC10 was produced as described previously (9) and conjugated to MMAE (11) and monomethyl auristatin F (MMAF) (19) to yield cAC10vc-MMAE and cAC10vc-MMAF, respectively. The CD30 ϩ Hodgkin disease L540cy cell line, a derivative of the L540 Hodgkin disease cell line adapted for xenograft growth, was provided by Dr. Harald Stein (Institut fur Pathologie, University of Veinikum Benjamin Franklin, Berlin, Germany). L540cy cells were grown in RPMI 1640 supplemented with 20% heat-inactivated fetal calf sera and antibiotics.
L540cy cells (1 ϫ 10 6 cells/ml) were incubated with 2 g/ml cAC10, cAC10vc-MMAE, or cAC10vc-MMAF for 30 min on ice, rinsed with ice-cold PBS, and then incubated for 30 min in the presence (crosslinking) or absence (no cross-linking) of 2 g/ml goat anti-human IgG (Jackson Immunoresearch, West Grove, PA) (20). The cells were rinsed with cold PBS, resuspended in growth media, and incubated at 37°C. The samples were harvested at various times and processed for flow cytometry. To detect surface-bound ADC, the cells were incubated with 10 g/ml mouse anti-id antibody to cAC10 for 30 min at 4°C, washed, and then incubated with 10 g/ml goat anti-mouse IgG-fluorescein isothiocyanate with minimal cross-reactivity to human IgG (Fc␥-specific, F(abЈ) 2 fragment; Jackson Immunoresearch). To detect internalized antibody or ADC, the cells were washed with cold PBS, incubated with proteinase K (5 g/ml for 10 min at 37°C), washed to remove cell surface-bound antibody, and treated with Cytofix/Cytoperm (BD Biosciences, San Jose, CA) prior to incubation with the anti-id antibody as described above. The cells were assessed by flow cytometry using a Becton Dickinson FACScan.
Immunofluorescence for Antibody and Antibody-Drug Conjugate Trafficking-L540cy cells (5 ϫ 10 5 cells/ml in regular medium) were incubated with 1 g/ml cAC10, cAC10vc-MMAE, or cAC10vc-MMAF for 30 min on ice or for 16 h at 37°C. After the incubation, the cells were washed with cold PBS to remove unbound antibody and drug conjugates. The cells were fixed and permeabilized with Cytofix/Cytoperm and stained as described previously (14). cAC10 and its conjugates were detected following incubation with Alexa Fluor 488-labeled goat antihuman IgG (HϩL) with minimal cross-reactivity to mouse IgG (Molecular Probes, Eugene, OR). Lysosomal compartments were visualized by staining with Lamp-1 (mouse CD107a antibody, BD Biosciences) and a secondary antibody, Alexa Fluor 568-conjugated goat anti-mouse IgG (HϩL) with minimal cross-reactivity to human IgG (Molecular Probes). Nuclear compartments were stained with 4Ј,6Ј-diamidino-2-phenylindole (Roche Applied Science). Fluorescence images were acquired with a Leitz Orthoplan epifluorescence microscope. In other experiments, the cAC10 antibody was linked to Alexa Fluor 488 reactive dye (Molecular Probes) and incubated with L540cy cells (2 ϫ 10 5 cells/ml) at 200 ng/ml for 3 h at 37°C in the presence or absence of trafficking inhibitors (40 M cytochalasin D or 3 mM ammonium chloride). Lysosomal and nuclear compartments were visualized by staining with LysoTracker and Hoescht DNA dyes, respectively (Molecular Probes). Fluorescence images were taken on fixed cells with a Carl Zeiss Axiovert 200M microscope.
Subcellular Fractionation-L540cy cells (5 ϫ 10 5 cells) were incubated in growth medium with Alexa Fluor 488-labeled cAC10vc-MMAE (8 g/ml) for 18 h at 37°C. The cells were harvested, washed with ice-cold PBS, and pelleted at 500 ϫ g for 5 min. Lysosome-enriched fractions were prepared as described previously (Axis-Shield, Oslo, Norway and Ref. 28). Briefly, the cell pellets (5 ϫ 10 7 cells) were incubated on ice for 30 min in cold homogenization buffer (0.25 M sucrose, 1 mM EDTA, 10 mM HEPES, pH 7.4) and gently Dounce homogenized to break open the cells. The homogenate was spun at 3000 ϫ g for 10 min at 4 C. The supernatant was collected and centrifuged at 17,000 ϫ g for 15 min at 4°C. The resulting pellet was resuspended in homogenization buffer and mixed with 60% OptiPrep (Axis-Shield; final concentration, 20%) before centrifugation at 208,000 ϫ g for 18 h at 4°C. The samples were then fractionated (5 drops/fraction), and the 25 fractions were assayed for the lysosomal marker, ␤-galactosidase activity (metabolism of 4-methylumbelliferyl-␤-D-galactopyranoside; Sigma), the presence of Alexa Fluor-labeled ADC by fluorimetry (Fusion-HT; Packard Instruments, Meridien, CT), and the loss of drug associated with heavy or light chains on the antibody frame by Western blot analyses with the anti-MMAE SG2.15 antibody (14).
Western Blot Analyses-To evaluate the metabolism of cAC10-drug conjugates, purified cathepsin B (2 units/ml; Calbiochem) was incubated for 3 h at 37°C with 5 g/ml cAC10vc-MMAE or cAC10vc-MMAF in buffer (2 mM dithiothreitol, 50 mM sodium acetate, pH 5.0) in the presence or absence of 10 M E64d cysteine protease inhibitor. In other experiments, Karpas and L540cy pooled lysosome-enriched fractions (numbers 10 -19) were incubated with 5 g/ml cAC10vc-MMAE for 18 h in the presence or absence of 20 M E64d or 3 M CA074-OME. The reactions were stopped by quick freezing in a dry ice bath. The digests were mixed with Novex sample buffer (Invitrogen), run on 4 -20% Tris-Gly or 4 -12% Bis-Tris gradient gels (Invitrogen) under reducing conditions, and transferred onto polyvinylidene difluoride membranes (Invitrogen). The membranes were blocked with 2% nonfat dry milk in PBST (PBSϩ 0.1% Tween 20) prior to incubation with the mouse SG2.15 antibody to detect drug. This antibody recognizes both MMAE and MMAF (data not shown). Detection was then performed using a horseradish peroxidase-goat anti-mouse IgG (Fc␥ specific, Jackson Immunoresearch) followed by ECL (SuperSignal West Pico; Pierce). To detect the heavy and light chains of the antibody, detection was performed using horseradish peroxidase-goat anti-human IgG (Fc␥ specific; Jackson Immunoresearch) or horseradish peroxidase-goat anti-human IgG (-specific; Southern Biotech, Birmingham, AL), respectively, followed by ECL.
Cytotoxicity Assay-Cell viability was measured by Alamar Blue (BIOSOURCE International, Camarillo, CA) dye reduction (10,14) or with Celltiter-Glo (Promega, Madison, WI). The results were reported as the IC 50 values, the concentration of compound needed to yield a 50% reduction in viability compared with untreated cells (control ϭ 100%).
In other experiments, Karpas cells (5 ϫ 10 5 cells) were incubated in growth medium for 4 h at 37°C prior to the harvesting of cells to generate lysosomal fractions. The fractions were washed with 0.25 M sucrose buffer containing 10 mM HEPES, pH 7.2. Each fraction was incubated with cAC10vc-MMAE (5 g/ml) for 24 h in reaction buffer (2 mM dithiothreitol, 50 mM sodium acetate, pH 5.0). Afterward, the fractions were diluted 300-fold in RPMI growth medium and added to cul-

RESULTS
MMAE and MMAF belong to the dolastatin 10 family of highly potent anti-mitotic agents that inhibit tubulin polymerization ( Fig. 1) (10,19). In contrast to MMAE, the charged, carboxylic acid terminus of free MMAF can potentially limit passive transit through cell membranes. The cAC10 antibody-drug conjugates of MMAE and MMAF containing a protease-cleavable vc linker were prepared as described previously (10,19).
We have reported that cAC10 and cAC10-vcMMAE were comparable in binding to CD30 ϩ cells and lacked interaction with CD30-negative cells (11). The activities of the free drugs and the conjugates cAC10vc-MMAE and cAC10vc-MMAF were compared on CD30 ϩ lymphoma cell lines including Karpas-299 and L540cy (Table 1). Com- The structures of each drug appended to mAb sulfhydryl groups via valine-citrulline linkers are depicted. ADCs were prepared by controlled partial reduction of internal mAb disulfides with dithiothreitol, followed by the addition of the linker drugs as described previously (10,19). parison of free drugs showed that the cell-permeable MMAE was 50 -200-fold more effective than MMAF. As an ADC however, cAC10vc-MMAF was significantly more potent than cAC10vc-MMAE (IC 50 Ͻ than 0.11 nM, p Ͻ 0.001). Neither drug conjugate exhibited appreciable activity against multiple antigen-negative cell lines, demonstrating that the potency of these ADCs is antigen-dependent. For example, the IC 50 value against the CD30-negative cell line Ramos following 96 h continuous exposure was Ͼ75 nM (Table 1).
To evaluate the trafficking and internalization of cAC10 and its ADC, CD30 ϩ L540cy cells were preincubated with mAb or ADCs and treated with vehicle or anti-human IgG for cross-linking prior to culture at 37°C. Cross-linked antibodies exhibit increased internalization and cellular clearance (10). At appointed times, the cells were harvested and assessed for surface-bound and intracellular levels of mAb and ADCs by flow cytometry (Fig. 2). Cell surface levels of both mAb and ADCs decreased sharply with time, coincident with increased intracellular levels (Fig. 2), suggesting that they internalized with similar kinetics. By 20 h, the surface levels of mAb and ADCs were ϳ60% of the initial levels ( Fig. 2A). Intracellular levels rose quickly within the first hour, peaked between 2 and 5 h, and maintained lower but steady levels up to 48 h (Fig. 2B). Cross-linking of the cAC10 antibody and ADCs increased intracellular levels of both the mAb and the conjugates by ϳ3-fold over levels observed in the absence of cross-linking (Fig. 2B).
Immunofluorescence microscopy was used to localize the internalized mAb and ADCs within L540cy cells (Fig. 3). The cells were incubated with the mAb or ADCs either on ice or at 37°C for the stated times and imaged immediately or fixed, permeabilized, and then processed for immunofluorescence. The cells incubated on ice and stained for mAb or ADCs (Fig. 3, top panels, green signal) showed diffuse, cell surface-associated staining and no evidence of internalization. Lysosomes, visualized using an antibody to lysosome-associated membrane protein 1 (Lamp-1; Fig. 3, top panels, red signal) or with LysoTracker (bottom panels, red signal) were distinct and punctate. At 37°C, there was capping and punctate staining for both mAb and ADCs within the L540cy cells and reduced staining on the cell surface. The intracellular mAb and ADC signals co-localized with those for Lamp-1 (Fig. 3, arrows, yellow signal, 16 h incubation) or with LysoTracker (Fig. 3, 3-h incubation, yellow signal, Merged, and data not shown), suggesting that  both mAb and ADCs internalized and were transported to the lysosomes. The results were not affected by fixation and sample processing because similar images were obtained using either live or fixed cells.
Trafficking of the ADCs to the lysosomes was also confirmed bio-chemically. L540cy cells were preincubated with Alexa Fluor-labeled cAC10vc-MMAE. The cells were then lysed and organelles fractionated by density gradient as described under "Materials and Methods." Enrichment for lysosomes following ultracentrifugation was confirmed by determining the distribution of phenotypic markers for various subcellular organelles. The markers for nuclei (DNA) and mitochondria (ATP generation) were found in the more dense fractions of the gradient (fractions 1-7), whereas the lysosomal marker, ␤-galactosidase activity, was detected in the middle fractions (fractions 8 -16), and the late endosomes/Golgi may be associated with lighter fractions (fractions 16 -20), as reported elsewhere (29). When gradient fractions were analyzed for the levels of fluorescently labeled ADC by fluorimetry and for ␤-galactosidase activity (Fig. 4A), peak enzyme activity was coincident with that of the fluorescently labeled ADC. Consistent with data from photomicrography (Fig. 3), these findings suggest that the internalized ADCs traffic and concentrate in the lysosomes.
To assess the prevailing mechanism of cellular uptake, the cells were preincubated with inhibitors of trafficking prior to the addition of the mAb or ADCs and subsequent evaluation by flow cytometry (Fig. 4B) and fluorescence microscopy (Fig. 4C). Ammonium chloride (NH 4 Cl), a lysosomotropic agent, disrupts trafficking and lysosomal processing by neutralizing the acidic environment of the endosomal/lysosomal compartments (30 -32). Treatment of L540cy cells with this reagent significantly decreased the total intracellular levels of the mAb and ADCs (Fig. 4B, p Ͻ 0.001; Fig. 4C, yellow signal) and increased the cell surface levels (Fig. 4C, green signal). Intracellular accumulation of mAb and ADCs was also significantly diminished (40 -65%) in cells pretreated with inhibitors of clathrin-mediated endocytosis (amantadine, phenylarsine oxide, clasto-lactacystin-␤-lactone, and chlorpromazine) (21,25,27) (Fig. 4B, p Ͻ 0.001), including cytochalasin D (23,27), that disrupt actin assembly. Increased levels of the antibody and ADCs were found on the cell surface of cytochalasin D-treated cells (Fig. 4C). Conversely, inhibitors of cathepsin B-mediated proteolysis (CA074-OME and E64d) (24) significantly enhanced the mAb and ADC levels by 50 -65% (Fig. 4B, p Ͻ 0.001). Under the conditions used, compounds known to modulate caveolae-mediated uptake (e.g. nystatin or filipin) (26) or microtubule assembly (e.g. colchicine or nocodazole) (22) had no effect on the total intracellular levels of the mAb and ADCs (Fig. 4B and data not shown). These data collectively suggest that the mAb and ADCs internalize and traffic to the lysosomes with similar magnitude and kinetics via clathrin-mediated endocytosis. Additionally, the activity of lysosomal enzymes such as cathepsin B appeared to modulate the intracellular levels of ADC.
We then investigated the ability of lysosomal metabolism of ADCs to release free drug. In a control study, ADCs were incubated with the purified lysosomal cysteine protease, cathepsin B (33). The reaction products were analyzed by Western blot using antibodies to the drug and to the heavy and light chain components of the ADC. Fig. 4D shows that cathepsin B treatment effectively removed detectable drug associated with the heavy and light chains. Cleavage of drug from the heavy and light chains of cAC10vc-MMAE and cAC10vc-MMAF with purified cathepsin B led to slight mobility shifts (reduced mass) but nonetheless preserved discrete heavy and light chain bands (Fig. 4D). No heterogeneity caused by mAb proteolysis was detected. These findings agree with those reported previously (10). Interestingly, an increase in signal for the light chain was observed in the presence of cathepsin B. This may be due to limited proteolysis augmenting the immunoreactivity of the -specific detection antibody (Fig. 4D). In addition, cysteine protease inhibitors such as E64d and CA074-OME were able to block the metabolism of the antibody and ADCs, restoring the drug-associated signal on Western analysis (Fig. 4D) as well as increasing the intracellular levels of these proteins (Fig. 4B).
To determine whether diminished ADC metabolism was correlated with reduced cytotoxicity, we evaluated the effects of cAC10vc-MMAE on the viability of L540cy cells in the presence of the cysteine protease inhibitors, E64d and CA074-OME. In L540cy cells, CA074-OME alone had small effects on the cytotoxic activity of cAC10vc-MMAE (Fig. 5A,  diamonds), shifting the IC 50 ϳ1.7-fold. In comparison, E64d significantly reduced the activity of the ADC by shifting the IC 50 ϳ10-fold (Fig. 5A, triangles). A greater decrease in the activity of the ADC was observed when the cells were pretreated with both inhibitors. With the combination of E64d and CA074-OME, there was a marked increase in the number of viable cells and a dramatic shift in IC 50 to greater than 10 g/ml ADC or 250 nM drug (Fig. 5A, open circles). When E64d was added 6 or 18 h after the start of culture with the ADC, there was minimal blocking effect on the activity of the ADC (data not shown). This was not surprising given that anti-CD30-ADCs appear to work rapidly. The cAC10vc-MMAE drug conjugate induced the growth arrest of L540cy cells within 12 h (11). At sufficient levels (Ͼ60 M for E64d and Ͼ3 M for CA074-OME), the protease inhibitors were cytotoxic. Yet treatment of L540cy cells with subtoxic levels of these inhibitors reduced the activity of the ADC, resulting in an increase in the IC 50 values and an increase in the viability of the lymphoma cells. In comparison, inhibitors of serine proteases (aprotinin and p-aminoethylbenzenesulfonyl fluoride), aspartic proteases (pepstatin A), transglutaminases (cystamine (34)) or cytoskeleton-associated calpains (calpeptin and N-acetyl-Leu-Leu-Nle-CHO (35)) had no significant effects on the activity of cAC10vc-MMAE (Fig. 5B) or cAC10vc-MMAF (data not shown). We then explored the ability of subcellular fractions to process ADCs using lysosome-enriched fractions from another CD30 ϩ cell line, Karpas 299 cells. The fractions were incubated with cAC10vc-MMAE and then analyzed by Western blot using an anti-MMAE antibody to detect drug associated with the heavy and light chains of the cAC10 antibody. Fig. 5C clearly shows a reduced signal corresponding to the loss of drug associated with both mAb heavy and light chains that was coincident with fractions containing the peak lysosomal ␤-galactosidase activity (fractions [13][14][15][16][17]. Drug loss from the mAb chains was blocked in the presence of E64d or CA074-OME (Fig. 5D for cAC10vc-MMAE and data not shown for cAC10vc-MMAF). To determine whether the released drug was biologically active, aliquots of these reaction mixtures were added to the CD30-negative Ramos cell line with cytotoxicity assessed after 96 h. The released drug is freely permeable and cytotoxic to nontarget cells, whereas intact ADC is not. Fig. 5E shows a significant reduction in the viability of Ramos cells coincident with the peak of lysosomal enzyme activity in fractionated Karpas cells (fractions [13][14][15][16][17]. Cytotoxicity was blocked by the presence of an anti-drug antibody, confirming that the effect was due to the presence of active drug in these fractions. Other fractions yielded reduced or negligible effects. The results collectively show that the cAC10-drug conjugates traffic to the lysosomes in target cells and interact with lysosomal enzymes, yielding the release of active free drug.

DISCUSSION
Targeted therapy offers the promise of selective elimination of antigen-positive tumor cells with minimal toxicity to normal tissues. To be therapeutically beneficial, the targeting mAb must efficiently internalize, traffic, and be metabolized to release a fully active agent. The development of a mAb-drug linker that is highly stable to the extracellular environment (7,10,12) yet readily releases fully active drug upon internalization into tumor cells has shown preclinical efficacy with mAbs against multiple targets in several cancers (10,11,13,15). Here we examined the activities and the cellular properties of drug conjugates of the anti-CD30 mAb, cAC10, appended to the anti-mitotic drugs, MMAE or MMAF, via the protease-cleavable valine-citrulline linker. In vitro cytotoxicity assays have shown that both conjugates were highly potent against CD30 ϩ Karpas and L540cy cells.
In L540cy cells, both mAb and ADC internalized via clathrin-mediated endocytosis and localized to the lysosomes. Inhibitors of caveolinassociated pathways did not appear to affect the CD30 internalization process, although caveolin may play a role in the internalization of other mAb-antigen complexes (14,36,37). As measured by the decline in cell surface quantities and increased intracellular levels of mAb, the internalization of both anti-CD30 mAb and ADCs was gradual and compa-rable. In contrast, vcMMAE appended to the anti-CD20 mAb rituximab significantly increased internalization rates compared with mAb alone (14), suggesting that ADC internalization is both mAb-and antigen-dependent. Approximately 40% of the initial surface-bound anti-CD30 mAb and ADC were lost by 20 h, without a concomitant increase in internalized signal. mAb or ADC within the cell was maximal between 2 and 5 h and remained constant over 48 h, suggesting that surface CD30 shedding into the medium (38) may contribute to a loss in extracellular detection. The repopulation of the cell surface and intracellular degradation of mAb may also play a role in maintaining steady state levels. Extracellular decline caused by ADC instability is unlikely because vc- FIGURE 5. Lysosomal protease activity is necessary for release of active drug. A, cysteine protease inhibitors block the activity of cAC10vc-MMAE in L540cy cells. L540cy cells were treated for 1 h with the cysteine protease inhibitors E64d and CA074-OME prior to the addition of the ADC. Cell viability was assessed 96 h later. The data are expressed as the means Ϯ S.D. of triplicate values (percentage of untreated control) from four separate experiments. B, minimal effects of inhibitors of aspartic proteases, serine proteases, transglutaminase, and calpains on the activity of cAC10vc-MMAE in L540cy cells. L540cy cells were incubated with the ADC in the presence or absence of serine protease inhibitors (aprotinin and p-aminoethylbenzenesulfonyl fluoride), aspartic protease inhibitor (pepstatin A), calpain inhibitors (calpeptin and N-acetyl-Leu-Leu-Nle-CHO), a transglutaminase inhibitor (cystamine), or E64d. Cell viability was assessed 96 h later. Similar results were obtained with cAC10vc-MMAF. The data are expressed as the means Ϯ S.D. of triplicate values (percentage of untreated control) from three separate experiments. C, drug release from cAC10vc-MMAE is coincident with lysosomal activity. CD30-positive Karpas cells were incubated with cAC10vc-MMAE. Lysosome-enriched fractions were prepared by density gradient sedimentation. Fractions from the gradient were analyzed for ␤-galactosidase activity and loss of mAb-associated MMAE drug by Western blot analysis. The first lane in the Western blot represents untreated ADC. The data shown are representative of the results from two independent experiments. D, lysosomal cysteine protease activity is blocked by E64d and CA074-OME. Lysosome-enriched fractions from Karpas cells (fractions 10 -19; C) were pooled and incubated with cAC10vc-MMAE in the presence or absence of E64d and CA074-OME. The loss of mAb-associated MMAE drug was assessed by Western blot analysis. Similar results were obtained with cAC10vc-MMAF and with lysosomal fractions from L540cy cells. E, lysosomal incubation of cAC10vc-MMAE resulted in the release of active free drug. Gradient fractions from Fig. 5C were added to CD30-negative Ramos cells with or without a neutralizing anti-drug antibody. The cells were incubated for 96 h, and viability was assessed with Alamar Blue. Fractions containing peak ␤-galactosidase activity yielded maximum cytotoxic activity (squares). Cytotoxic activity was blocked by the presence of anti-drug antibody (Eϩanti-MMAE mAb, triangles). The data shown are representative of triplicate measurements (percentage of untreated control) from two separate experiments. E ϭ cAC10vc-MMAE. linked conjugates are highly stable in vitro and in vivo (7,10,12). In CD30 ϩ L540cy cells, the blocking of either the internalization process or the activity of lysosomal enzymes, but not the activity of cytosolic enzymes, effectively reduced the intracellular levels of the ADCs and ablation of tumor cells. Within the lysosomes, cysteine proteases including cathepsin B but not aspartic or serine proteases or transglutaminases contributed to the release of active drug. Because released drugs are permeable and active against all cells, the results of the cytotoxicity assay using the CD30-negative Ramos cells clearly correlated with the lysosomal activity of Karpas cells with release of an active agent. Inhibition of this cytotoxicity by a MMAE/MMAF-specific neutralizing antibody suggested that active drug and not other metabolites are released by the lysosomal activity.
Other ADCs have shown a disconnect between trafficking to the lysosomes and ineffectiveness as a cytotoxic agent, suggesting that internalization, production, and concentration of free drug were inadequate to induce toxicity. Both of the anti-CD20 conjugates of vc-doxorubicin and vc-MMAE internalized and localized to the lysosomes of target cells but only the MMAE conjugate actively induced growth arrest and apoptosis (14), indicating that free doxorubicin was not efficiently released or sequestered (14). The current data demonstrate that whereas antigen expression levels may be sufficient to produce effective cell killing by ADC, the combined parameters of internalization, trafficking, lysosomal protease activity, drug release, and intracellular concentration of free drugs are key to determining the activity of ADCs against target tumor cells.