Oncogene.
`v. 33, no. 4 (Jan.23 2014)
`General Collection
`W1 ON102H
`2014—02—11 06:56:30
`
`Ullb
`
`23 January 2014
`
`Volume 33 Number 4
`
`MEFZC in hematopoiesis and leukemia
`
`878163842
`
`Nedd9 regulation of mammary luminal
`progenitors
`»
`xn»’4
`
`l
`m,
`
`MDMZ
`
`* nature publishing group IQ?
`
`IMMUNOGEN 2173, Pg. 1
`Phigenix v. Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2173, pg. 1
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`Oncogene (2014) 33, 429–439
`& 2014 Macmillan Publishers Limited All rights reserved 0950-9232/14
`
`www.nature.com/onc
`
`ORIGINAL ARTICLE
`Design optimization and characterization of Her2/neu-targeted
`immunotoxins: comparative in vitro and in vivo efficacy
`studies
`
`Y Cao, JW Marks, Z Liu, LH Cheung, WN Hittelman and MG Rosenblum
`
`Targeted therapeutics are potential therapeutic agents because of their selectivity and efficacy against tumors resistant to
`conventional therapy. The goal of this study was to determine the comparative activity of monovalent, engineered anti-Her2/neu
`immunotoxins fused to recombinant gelonin (rGel) to the activity of bivalent IgG-containing immunoconjugates. Utilizing Herceptin
`and its derived humanized single-chain antibody (single-chain fragment variable, designated 4D5), we generated bivalent chemical
`Herceptin/rGel conjugate, and the corresponding monovalent recombinant immunotoxins in two orientations, 4D5/rGel and rGel/
`4D5. All the constructs showed similar affinity to Her2/neu-overexpressing cancer cells, but significantly different antitumor
`activities. The rGel/4D5 orientation construct and Herceptin/rGel conjugate were superior to 4D5/rGel construct in in vitro and
`in vivo efficacy. The enhanced activity was attributed to improved intracellular toxin uptake into target cells and efficient
`downregulation of Her2/neu-related signaling pathways. The Her2/neu-targeted immunotoxins effectively targeted cells with Her2/
`neu expression level 41.5 105 sites per cell. Cells resistant to Herceptin or chemotherapeutic agents were not cross-resistant to
`rGel-based immunotoxins. Against SK-OV-3 tumor xenografts, the rGel/4D5 construct with excellent tumor penetration showed
`impressive tumor inhibition. Although Herceptin/rGel conjugate demonstrated comparatively longer serum half-life, the in vivo
`efficacy of the conjugate was similar to the rGel/4D5 fusion. These comparative studies demonstrate that the monovalent,
`engineered rGel/4D5 construct displayed comparable in vitro and in vivo antitumor efficacy as bivalent Herceptin/rGel conjugate.
`Immunotoxin orientation can significantly impact the overall functionality and performance of these agents. The recombinant rGel/
`4D5 construct with excellent tumor penetration and rapid blood clearance may reduce the unwanted toxicity when administrating
`to patients, and warrants consideration for further clinical evaluation.
`
`Oncogene (2014) 33, 429–439; doi:10.1038/onc.2012.612; published online 4 February 2013
`
`Keywords: immunotoxin; Her2/neu; gelonin; valency; design optimization
`
`INTRODUCTION
`Numerous studies have revealed that Her2/neu overexpression by
`tumors is a threshold event leading to a highly aggressive cellular
`phenotype, and therapeutic strategies directed against Her2/neu
`have rapidly gained recognition.1,2 Her2/neu-targeted therapies,
`including Herceptin and Lapatinib, have significantly improved
`outcomes in Her2/neu-positive cancers; however, use of these
`agents can be limited by resistance and tolerability issues.3,4
`Therefore, there is a need for novel and improved therapeutic
`approaches targeting Her2/neu.
`Immunotoxins are exquisitely powerful cytotoxic proteins.5,6
`Initial studies
`focused on constructs created by chemically
`conjugating an antibody to a protein toxin.7,8 With advances in
`recombinant DNA technology, engineered antibody fragments
`have been employed to deliver various toxins to Her2/neu-positive
`tumor cells.9,10 There have been numerous studies examining the
`impact of construct size, antibody affinity and the valency of
`constructs on overall efficacy.11,12 Wels and colleagues13 suggested
`that higher avidity and longer
`residence time of
`IgG-based
`immunoconjugates may outweigh the improved tumor pene-
`tration of single-chain fragment variable (scFv)-based constructs.
`However, immunoconjugate development has been hampered by
`
`nonspecific toxicity and vascular leak syndrome.14 In addition, tight
`junctions between tumor cells and high interstitial tumor pressures
`could limit successful use of IgG conjugates.12
`Recombinant engineering of immunotoxins has the potential
`to overcome many of these problems.12,15 To improve Her2/
`neu-targeting properties, Adams et al.16 suggested that (scFv)2
`constructs display improved tumor
`targeting and retention
`compared with
`scFv monomers. However,
`Pastan
`and
`colleagues17 described diabody-based immunotoxins (designated
`e23 (dsFv)2/PE) showing 410-fold greater in vitro cytotoxicity than
`their monovalent counterparts, but only about 2-fold greater
`activity in vivo than the monovalent analogs. The high affinity of a
`diabody may result in formation of a binding-site barrier at the
`periphery of tumors, which impedes immunotoxin penetration into
`the tumor mass.18 Thus, the therapeutic window for Her2/neu
`targeting may be optimized on design changes instead of
`confinement to the valency argument.
`Recombinant gelonin (rGel), a 29-kDa single-chain ribosome-
`inactivating protein, has been well established as a highly
`cytotoxic payload of chemical conjugates or fusion constructs
`for the treatment of many tumor types.19–21 In this study, we
`utilized Herceptin and its humanized scFv (designated 4D5) to
`
`Immunopharmacology and Targeted Therapy Laboratory, Department of Experimental Therapeutics, M.D. Anderson Cancer Center, Houston, TX, USA. Correspondence: Professor
`MG Rosenblum, Department of Experimental Therapeutics, University of Texas, MD. Anderson Cancer Center, Houston, TX 77054, USA.
`E-mail: mrosenbl@mdanderson.org
`Received 14 June 2012; revised 5 October 2012; accepted 14 November 2012; published online 4 February 2013
`
`IMMUNOGEN 2173, pg. 2
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`430
`
`optimization of anti-Her2/neu immunotoxins
`Y Cao et al
`
`generate a conventional Herceptin/rGel conjugate and corres-
`ponding recombinant immunotoxins in two orientations: 4D5/rGel
`and rGel/4D5. Further characterization studies were performed,
`including examining the impact of valency and construct
`orientation on in vitro selectivity, specificity and efficacy of these
`agents, as well as comparison of their pharmacokinetics, tumor
`penetration and tumor-targeting efficacy in vivo against tumor
`xenografts.
`
`RESULTS
`Preparation of rGel-based immunotoxins
`Antibody–toxin conjugates were generated with a disulfide-based
`succinimidyl 3-(2-pyridyldithio)propionate (SPDP) linker for facile
`release of toxin from the antibody carrier (Figure 1a). As shown in
`Figure 1b, the final product contained a mixture of immunocon-
`jugates containing one rGel molecule (major) and two rGel
`molecules (minor; average molar ratio of 1.21 rGel molecules per
`antibody). No free Herceptin or free rGel were detected.
`The monovalent immunotoxins were generated by fusing scFv
`4D5 to the rGel, using the flexible GGGGS linker
`in two
`orientations (4D5/rGel and rGel/4D5, Figure 1a). Both immunotox-
`ins were expressed in E.coli AD494 (DE3) pLysS. Following
`purification, the immunotoxins were shown to migrate at the
`expected molecular weight (55 kDa under nonreducing condition)
`with a purity 495% (Figure 1b).
`
`Analysis of binding affinity
`fusion constructs and
`The binding affinities of monovalent
`bivalent chemical conjugates were assessed by enzyme-linked
`
`immunosorbent assay using Her2/neu extracellular domain
`(Figure 2a). The apparent binding affinities (Kd) were determined
`by calculating the concentration of immunotoxins that produced
`half-maximal specific binding. The monovalent 4D5/rGel and rGel/
`4D5 demonstrated apparent affinities of 0.106 and 0.142 nM,
`respectively, and the bivalent Herceptin/rGel conjugate had an
`apparent affinity of 0.201 nM. These results are in agreement with
`the published affinity values for native Herceptin to the Her2/neu
`receptor (Kd 0.15 nM).22
`We next tested the cellular Her2/neu-binding activities of these
`immunotoxins by flow cytometry. As shown in Figure 2b, all the
`immunotoxins produced higher staining intensities with the
`Her2/neu-positive SK-OV-3 and BT-474-M1 cells, and displayed a
`significantly high specificity based on negative MDA-MB-468 cells.
`These studies confirmed that monovalent
`fusion constructs
`display virtually identical binding affinities compared with their
`original bivalent antibody-based conjugates.
`
`Cell-free protein synthesis inhibitory activity
`rGel component of
`To examine the n-glycosidic activity of
`immunotoxins, these materials were tested by cell-free protein
`synthesis assay.
`Inhibition curves for 4D5/rGel, rGel/4D5 and
`Herceptin/rGel conjugate were compared with that of native rGel
`(Supplementary Figure S1). The calculated IC50 (half-maximal
`inhibitory concentration) values for each immunotoxin were
`(55.2 pM, 48.9 pM, 38.3 pM vs
`found to be virtually identical
`70.3 pM for rGel, respectively). The consistency of rGel-based
`immunotoxins clearly demonstrated that no loss of
`toxin
`enzymatic activity occurred in the different constructs.
`
`a
`
`4D5/rGel
`
`Herceptin/rGel conjugate
`
`Fab
`
`Fv
`
`Fc
`
`V H
`
`VL
`
`C H1
`
`CL
`
`CH2
`
`CH3
`
`rGel
`
`rGel
`
`V H
`
`VL
`
`rGel/4D5
`
`V H
`
`VL
`
`rGel
`
`4D5/rGel
`
`rGel/4D5
`
`4D5/rGel
`
`rGel/4D5
`
`b
`
`kDa
`
`150
`100
`75
`
`50
`
`35
`
`25
`
`Herceptin/rGel
`conjugate
`
`Herceptin
`
`kDa
`
`212
`
`158
`
`116
`
`97
`
`Reducing
`
`Non-reducing
`
`Non-reducing
`
`Figure 1. Construction and preparation of Herceptin-based immunotoxins. (a) Schematic diagram of immunotoxin constructs containing scFv
`4D5 or full-length antibody Herceptin and rGel. (b) Purified immunotoxins were analyzed by SDS–polyacrylamide gel electrophoresis under
`nonreducing condition.
`
`Oncogene (2014) 429 – 439
`
`& 2014 Macmillan Publishers Limited
`
`IMMUNOGEN 2173, pg. 3
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`optimization of anti-Her2/neu immunotoxins
`Y Cao et al
`
`431
`
`SK-OV-3
`
`BT-474-M1
`
`MDA-MB-468
`
`500
`
`400
`
`300
`
`200
`
`100
`
`0
`100 101 102 103 104
`Log Fluorescence Intensity
`
`Relative Cell Number
`
`500
`
`400
`
`300
`
`200
`
`100
`
`0
`100 101 102 103 104
`Log Fluorescence Intensity
`
`Relative Cell Number
`
`500
`
`400
`
`300
`
`200
`
`100
`
`0
`100 101 102 103 104
`Log Fluorescene Intensity
`
`Relative Cell Number
`
`4D5/rGel
`4Gel/4D5
`Herceptin/rGel conjugate
`rGel
`
`100
`101
`10–2 10–1
`Concentration (nM)
`
`102
`
`2.0
`
`1.5
`
`1.0
`
`0.5
`
`0.0
`10–3
`
`Absorbance at 405nm
`
`4D5/rGel
`
`rGel/4D5
`
`Herceptin/rGel conjugate
`
`rGel
`
`4D5/rGel
`
`rGel/4D5
`
`Herceptin/rGel
`conjugate
`
`rGel
`
`SK-OV-3
`
`4D5/rGel
`
`rGel/4D5
`
`Herceptin/rGel
`conjugate
`
`SK-OV-3
`
`BT-474-M1
`
`MDA-MB-468
`
`1 2 4 8 24 h 1 2 4 8 24 h 1 2 4 8 24 h
`
`kDa
`80
`58
`46
`
`30
`25
`
`Full-length
`immunotoxin
`
`Free rGel
`
`B-actin
`Relative Density:
`Total rGel Signal
`0.2
`Free rGel Signal
`0
`
`0.2
`0.3
`0.5
`0.1
`0 0.1 0.3 0.1
`
`0.6
`0
`
`0.6
`0.8
`1.5
`1.0
`0 0.3 1.2 0.9
`
`1.5 1.5 1.5 1.2 0
`1.5 1.5 1.5 1.2 0
`
`BT-474-M1
`
`4D5/rGel
`
`rGel/4D5
`
`Herceptin/rGel
`conjugate
`
`kDa
`
`0 1 2 4 8 12 18 24 30 36 42 48 h
`
`0 1 2 4 8 12 18 24 30 36 42 48 h
`
`0 1 2 4 8 12 18 24 30 36 42 48 h
`
`80
`58
`46
`
`30
`25
`
`Β-actin
`Relative Density:
`Total rGel Signal
`Free rGel Signal
`
`0 0.1 0.3 0.4 1.0 1.3 1.5 1.4 1.7 1.5 1.3 0.1
`0 0 0 0 0 0.3 0.5 0.5 0.7 0.7
`0.9
`0.1
`
`0 0.6 0.8 1.0 1.3 1.5 1.6 1.8 2.2 2.3 2.4 2.4
`0
`0
`0
`0 0.1 0.3 0.4 0.7 0.7 0.8
`0.9
`0.9
`
`0 2.4 4.9 5.3 3.5 3.5 2.1 0.7 0
`0 2.4 4.9 5.3 3.5 3.5 2.1 0.7 0
`
`0
`0
`
`0
`0
`
`0
`0
`
`Full-length
`immunotoxin
`
`Free rGel
`
`Figure 2. Characterization and comparison of anti-Her2/neu immunotoxins. (a) Binding curves of immunotoxins to Her2/neu extracellular
`domain by enzyme-linked immunosorbent assay. (b) Binding affinity analysis of 25 nM constructs on Her2/neu-positive (SK-OV-3 and BT-474-
`M1) and -negative (MDA-MB-468) cells by flow cytometry. (c) Internalization analysis on Her2/neu-positive and -negative cells after 4 h
`treatment of 25 nM immunotoxin. Cells were subjected to immunofluorescent staining with anti-rGel antibody (fluorescein isothiocyanate-
`conjugated secondary), with propidium iodine nuclear counterstaining. (d, e) Western blot analysis of intracellular behavior of 25 nM
`immunotoxin in SK-OV-3 and BT-474-M1 cells. Relative density of total rGel signal and free rGel signal normalized to the b-actin protein-
`loading control.
`
`Cellular uptake and toxin delivery of immunotoxins
`We next examined the comparative ability of the immunotoxins to
`internalize into SK-OV-3, BT-474-M1 and MDA-MB-468 cells. As
`shown in Figure 2c, after 4 h of exposure, the rGel moiety of all the
`immunotoxins was observed primarily in the cytosol after
`treatment of SK-OV-3 or BT-474-M1 cells, but not MDA-MB-468
`cells. This demonstrated that all constructs were comparable in
`efficient internalization after exposure to Her2/neu-positive cells.
`The internalization efficiency of all the immunotoxins was further
`examined by time-dependent western blot analysis on total rGel
`signal (full-length immunotoxinþ free rGel; Figures 2d and e). The
`Herceptin/rGel conjugate showed the fastest and highest inter-
`nalization efficiency into target cells. Both 4D5/rGel and rGel/4D5
`internalized rapidly into SK-OV-3 cells.
`In contrast,
`rGel/4D5
`internalized much faster than 4D5/rGel
`into BT-474-M1 cells.
`Compared with 4D5/rGel,
`rGel/4D5 displayed a long-lasting
`increase in cell internalization.
`The intracellular release of rGel after endocytosis of various
`constructs was assessed on free rGel signal (Figures 2d and e). For
`the Herceptin/rGel conjugate, there was a rapid initial delivery of
`
`free rGel to the cytoplasm within the first hour after drug
`exposure. The Herceptin/rGel conjugate delivered the greatest
`amount of rGel to both SK-OV-3 and BT-474-M1 cells. As for
`monovalent constructs, both fusions displayed similar
`initial
`delivery of free rGel
`in target cells, but rGel/4D5 orientation
`construct delivered constantly high levels of rGel compared with
`4D5/rGel.
`
`In vitro cytotoxicity of immunotoxins
`The cytotoxic effects of the constructs were then tested against a
`variety of different tumor cell lines. As shown in Table 1, all the
`immunotoxins demonstrated specific cytotoxicity to cells expres-
`sing þ 3 and þ 4 levels of Her2/neu. Targeting indices ranged
`from 54 to 5120 and from 21 to 2364 for Herceptin/rGel conjugate
`and rGel/4D5, respectively. The 4D5/rGel construct was compara-
`tively less potent with targeting indices only as high as 213.
`Treatment of Herceptin-resistant (HR) BT-474-M1 cells demon-
`strated no cross-resistance to rGel-based immunotoxins compared
`with parental BT-474-M1 cells. In addition, MCF-7 cells transfected
`
`& 2014 Macmillan Publishers Limited
`
`Oncogene (2014) 429 – 439
`
`IMMUNOGEN 2173, pg. 4
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`432
`
`optimization of anti-Her2/neu immunotoxins
`Y Cao et al
`
`Table 1. Comparative IC50 values of fusion constructs against various types of tumor cell lines
`
`Cell line
`
`Type
`
`SK-BR-3
`BT-474-M1
`BT-474-M1 (HR)
`MCF-7/Her2
`NCI-N87
`SK-OV-3
`Calu-3
`MDA-MB-453
`MDA-MB-361
`MDA-MB-435S
`BT-20
`ZR-75-1
`A-431
`MCF-7
`MDA-MB-231
`MDA-MB-468
`
`Breast
`Breast
`Breast
`Breast
`Gastric
`Ovarian
`Lung
`Breast
`Breast
`Breast
`Breast
`Breast
`Epidermoid
`Breast
`Breast
`Breast
`
`Her2//neu
`level
`
`þ þ þ þ
`þ þ þ þ
`þ þ þ þ
`þ þ þ þ
`þ þ þ þ
`þ þ þ þ
`þ þ þ þ
`þ þ þ
`þ þ þ
`þ þ
`þ þ
`þ þ
`þ
`þ
`þ
`–
`
`IC50 (nM)
`
`Targeting indexa
`
`Herceptin
`
`4D5/
`rGel
`
`rGel/
`4D5
`
`Herceptin/
`rGel
`conjugate
`
`rGel
`
`4D5/
`rGel
`
`rGel/
`4D5
`
`Hereceptin/
`rGel
`conjugate
`
`60.52
`50.35
`7930.49
`8128.31
`668.21
`1026.80
`4015.33
`9120.11
`9549.93
`ND
`ND
`ND
`ND
`–
`ND
`ND
`
`5.10
`26.53
`91.92
`32.36
`2.37
`70.86
`21.62
`543.70
`776.25
`120.64
`429.14
`5903.00
`98.31
`595.39
`526.62
`335.66
`
`0.36
`0.74
`0.54
`0.28
`0.16
`0.54
`0.20
`1.62
`31.62
`111.15
`284.97
`6005.10
`106.68
`322.18
`349.70
`364.33
`
`0.17
`0.17
`0.13
`0.11
`0.21
`0.28
`0.18
`0.79
`12.30
`181.68
`174.38
`6555.07
`213.99
`222.89
`571.48
`447.10
`
`851.14
`457.09
`315.28
`263.03
`505.48
`501.19
`457.09
`435.70
`660.69
`364.33
`232.22
`3473.20
`167.38
`197.15
`447.10
`485.18
`
`167
`17
`3
`8
`213
`23
`20
`21
`1
`3
`1
`1
`2
`ND
`1
`1
`
`2364
`618
`582
`939
`3159
`928
`2285
`269
`21
`3
`1
`1
`1
`ND
`1
`1
`
`5120
`2689
`2425
`2391
`2407
`1790
`2539
`552
`54
`2
`1
`1
`1
`1
`1
`1
`
`Abbreviations: IC50, half-maximal inhibitory concentration; ND, not determined. aTargeting index represents IC50 of rGel/IC50 of immunotoxin.
`
`with Her2/neu (MCF-7/Her2) showed increased resistance to
`chemotherapeutic agents (Supplementary Figure S2), but showed
`increased sensitivity to these constructs. The cytotoxicity of
`Herceptin/rGel conjugate and rGel/4D5 outperformed that of
`4D5/rGel, which resulted from the improved internalization and
`intracellular rGel delivery. Antigen-negative cells or those expres-
`sing þ 1 and þ 2 levels of Her2/neu were not specifically targeted
`by any of the constructs.
`
`Effects of immunotoxins on Her2/neu-related signaling pathways
`We next examined the mechanistic effects of the constructs on
`Her2/neu-related signaling events in SK-OV-3 cells. As shown in
`Figure 3a, treatment with Herceptin/rGel conjugate and rGel/4D5
`resulted in an impressive inhibition of phosphorylation of
`Her2/neu, epidermal growth factor
`receptor (EGFR), Akt and
`extracellular signal-regulated kinase (ERK), which are critical events
`in the Her2/neu signaling cascade. In contrast, 4D5/rGel showed a
`comparatively reduced effect on these pathways. Treatment with
`these immunotoxins resulted in reduction in the phosphorylation
`of insulin-like growth factor 1 receptor (IGF1R), a crosstalk partner
`of Her2/neu. These results suggest a link between downregulation
`of IGF1R signaling and the antiproliferative effect of anti-Her2/neu
`immunotoxins against SK-OV-3 cells. The rGel/4D5 construct was
`comparatively more cytotoxic to target cells than 4D5/rGel, and
`the improved cytotoxicity coincided with the increased effects on
`signal transduction.
`
`Immunotoxin effects on Her2/neu-driven multidrug resistance
`Cells overexpressing Her2/neu often display an increased require-
`ment for the phosphoinositide-3-kinase/Akt signaling pathway in
`anchorage-independent growth.23 Currently, we compared the
`efficacy of chemotherapeutic agents against MCF-7/Her2 cells and
`parental MCF-7 cells. The MCF-7 cells nominally coexpress Her3
`and transfection results in an increased resistance to multiple
`chemotherapeutic agents (Supplementary Figure S2).24 However,
`we didn’t observe the cross-resistance of these cells to rGel-based
`immunotoxins. As shown in Figure 3b, treatment with Herceptin/
`rGel conjugate and rGel/4D5 caused a dose-dependent inhibition
`of Her2/neu and Her3 phosphorylation, and inhibition of the
`downstream phosphoinositide-3-kinase/Akt and Ras/ERK cascade.
`
`Cytotoxic activity of immunotoxins against HR cells
`Acquired resistance to Herceptin therapy can be mediated by
`concomittent upregulation of Her2/neu downstream signaling
`pathways and involve constitutive Akt activation, or activation
`from extrinsic growth factor stimulation. We developed a model of
`HR variant BT-474-M1 cells. We also demonstrated that addition of
`EGF or neuregulin-1 growth factors to parental BT-474-M1 cells
`can prevent the cytotoxic response to Herceptin (Supplementary
`Figure S3). However, the treatment of rGel-based immunotoxins
`displayed impressive cytotoxic effects against all these HR cells
`(Supplementary Figure S4).
`Mammary BT-474-M1 cells in anchorage-independent three-
`dimensional
`(3D) culture have been shown to organize into
`structures resembling in vivo architecture. We examined the
`growth of BT-474-M1 and the HR cells in response to immunotox-
`ins under the 3D growth on basement membrane. As shown in
`Figure 4a, all the immunotoxins demonstrated impressive growth
`inhibition in these 3D models. Treatment with Herceptin/rGel
`conjugate or rGel/4D5 showed potent cytotoxic effects of both
`Herceptin-sensitive and HR models, whereas 4D5/rGel only
`partially inhibited cell growth (Figure 4b).
`We further investigated the signaling pathway of Her2/neu and
`its crosstalk receptors underlying the inhibitory potential of
`immunotoxins (Figure 4c).
`In either Herceptin-sensitive or HR
`cells, treatment with immunotoxins led to a potent blockade of
`phosphorylation of EGFR, Her2/neu and Her3. Similar results
`were found for inactivation of downstream Akt and ERK in these
`models. As indicated, Herceptin/rGel conjugate and rGel/4D5
`were the most potent at inhibiting phosphorylation of these
`targets, whereas 4D5/rGel showed comparatively reduced effects.
`
`Activity of immunotoxins against cells expressing intermediate
`levels of Her2/neu
`Previous studies have suggested that targeting of tumors with
`antibodies or immunotoxins to Her2/neu may result in unex-
`pected organ toxicities due to low levels of Her2/neu expressed
`on normal tissues.25,26 We tested the immunotoxins on ZR-75-1
`and BT-20 cells expressing intermediate levels of Her2/neu
`(B1.5 105 sites per cell).27,28 As shown in Supplementary
`Figure S5A, all the immunotoxins were shown to bind to these
`
`Oncogene (2014) 429 – 439
`
`& 2014 Macmillan Publishers Limited
`
`IMMUNOGEN 2173, pg. 5
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`optimization of anti-Her2/neu immunotoxins
`Y Cao et al
`
`433
`
`SK-OV-3
`
`pEGFR (T845)
`
`EGFR
`
`pHer2/neu (Y877)
`
`pHer2/neu (Y1221/1222)
`
`Her2/neu
`
`pIGF1R (Y1165/1166)
`
`IGF1R
`
`♢-actin
`
`pAkt (S473)
`
`Akt
`
`pERK (Y204)
`
`pERK (T177/T160)
`
`ERK
`
`♢-actin
`
`MCF-7
`
`MCF-7/Her2
`
`4D5/rGel
`
`rGel/4D5
`
`Herceptin/rGel
`conjugate
`
`1000 nM
`
`0
`
`1
`
`10
`
`100
`
`1000
`
`0
`
`1
`
`10
`
`100
`
`1000
`
`0
`
`1
`
`10
`
`100
`
`1000
`
`nM
`
`pHer2/neu (Y877)
`
`Her2/neu
`
`pHer3 (Y1328)
`
`Her3
`
`pAkt (S473)
`
`Akt
`
`pERK (Y204)
`
`ERK
`
`♢-actin
`
`Figure 3. Western blot analysis of signaling pathway inhibition of the immunotoxins. (a) Analysis of signal transduction of SK-OV-3 cells
`treated with 100 nM drugs for 48 h. (b) Analysis of the effect of different doses of immunotoxins on the signaling pathways of MCF-7/Her2 and
`MCF-7 cells after 48 h treatment.
`
`cells; however, we were unable to demonstrate the cellular uptake
`of
`immunotoxins (Supplementary Figure S5B). We were also
`unable to demonstrate cytotoxic effects of the immunotoxins on
`Her2/neu-related signaling pathways (Supplementary Figure S5C).
`These results
`indicated that Herceptin-based immunotoxins
`appeared to show specific cytotoxicity only to tumor cells
`expressing high levels of Her2/neu (41.5 105 sites per cell),
`but not cells with Her2/neu expression below this threshold.
`
`Immunotoxin pharmacokinetic studies
`To examine the pharmacokinetic behavior, we utilized fluorescent
`molecular imaging probe (IRDye 800CW) to label the immunotox-
`ins. The pharmacokinetics of each agent in mice was determined
`by quantitating the immunotoxin levels in serum by photolumi-
`nescence intensity analysis (Supplementary Figure S6). As shown
`in Figure 5a, Herceptin/rGel conjugate demonstrated a- and
`b-phase half-lives of 42 min and 42.7 h,
`respectively. These
`clearance kinetics appear to be similar to that of the intact
`Herceptin IgG.29,30 In addition, the monovalent 4D5/rGel and rGel/
`4D5 demonstrated relatively rapid clearance from the circulation
`with monophasic half-lives of 32.3 min and 34.8 min, respectively.
`These were similar to the serum half-lives found for scFv-based
`immunotoxins in mice.13,31
`
`Intra-tumor distribution patterns of immunotoxins
`Dual immunofluorescence studies were performed to evaluate the
`intratumoral distribution patterns of different immunotoxins in
`SK-OV-3 xenografts at different times (mouse CD31, red, and rGel,
`
`Immunofluorescence revealed significant
`green; Figure 5b).
`differences in the penetration of the immunotoxins from blood
`vessels into tumor with an apparent valency difference. The scFv-
`based immunotoxins exhibited the greatest average penetration
`distance from blood vessels as early as 1 h post injection. Optimal,
`diffuse distribution occurred throughout
`the tumor at 24 h
`after administration.
`In contrast, the intratumoral migration of
`Herceptin/rGel appeared to be restricted primarily to the
`perivascular or intravascular areas of the tumor from 1 to 24 h
`after administration. At 72 h, the immunoconjugate had dissemi-
`nated from vascular
`lumen. Quantitative analysis (Figure 5c)
`supports
`the observation that monovalent
`fusions achieve
`maximal tumor distribution by 24 h, and clear from the tumor
`thereafter.
`In contrast, Herceptin/rGel conjugate increases in
`tumors over the 72-h observation period. These findings suggest
`that high interstitial pressures in tumors may limit the diffusion of
`full-length antibodies from the blood vessels into the tumor.32,33
`In contrast, smaller fusion constructs display faster tumor uptake
`and clearance patterns, and provide uniform tumor penetration.
`
`Antitumor activity of immunotoxins in xenograft models
`We further evaluated the ability of anti-Her2/neu immunotoxins to
`inhibit the growth of SK-OV-3 tumor xenografts. As shown in
`Figure 6a, a dose-dependent inhibition of tumor growth was
`observed. For Herceptin/rGel conjugate, total doses of 10 mg/kg
`caused a 20-day tumor growth delay, whereas 20 or 40 mg/kg
`doses resulted in a 440-day growth delay until the animals were
`killed. The groups treated with 20 or 40 mg/kg rGel/4D5 showed a
`
`& 2014 Macmillan Publishers Limited
`
`Oncogene (2014) 429 – 439
`
`IMMUNOGEN 2173, pg. 6
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`434
`
`optimization of anti-Her2/neu immunotoxins
`Y Cao et al
`
`NT
`
`Herceptin
`
`Herceptin
`(1♯M)
`
`rGel
`
`4D5/rGel
`
`rGel/4D5
`
`Herceptin/rGel
`conjugate
`
`BT-474-M1
`
`BT-474-M1
`+
`EGF
`
`BT-474-M1
`+
`NRG-1
`
`BT-474-M1 (HR)
`
`25 ♯m
`
`BT-474-M1 cells
`
`BT-474-M1 cells + EGF
`
`BT-474-M1 cells + NRG-1
`
`BT-474-M1 (HR) cells
`
`30
`
`20
`
`10
`
`(pixels ×1000)
`Spheroid Area
`
`30
`
`20
`
`10
`
`(pixels ×1000)
`Spheroid Area
`
`30
`
`20
`
`10
`
`(pixels ×1000)
`Spheroid Area
`
`NT
`
`
`rGel
`
`4D5/rGelrGel/4D5
`
`Herceptin
`Herceptin/rGel conjugate
`Herceptin (1♯ M)
`
`0
`
`NT
`
`Herceptin (1♯ M)
`
`rGel
`
`
`4D5/rGelrGel/4D5
`
`Herceptin/rGel conjugate
`
`0
`
`NT
`
`Herceptin (1♯ M)
`
`
`rGel
`
`4D5/rGelrGel/4D5
`
`Herceptin/rGel conjugate
`
`0
`
`NT
`
`Herceptin (1♯ M)
`
`rGel
`
`
`4D5/rGelrGel/4D5
`
`Herceptin/rGel conjugate
`
`30
`
`20
`
`10
`
`0
`
`(pixels ×1000)
`Spheroid Area
`
`BT-474-M1
`
`BT-474-M1 (HR)
`
`BT-474-M1
`
`BT-474-M1
`
`-
`
`++++++
`
`20 ng/mL
`EGF
`
`-
`
`++++++
`
`50 ng/mL
`NRG-1
`
`pHer2/neu (Y877)
`
`pHer2/neu (Y1221/1222)
`
`Her2/neu
`
`pAkt (S473)
`
`Akt
`
`pERK (Y204)
`
`ERK
`♢-actin
`
`pEGFR (T845)
`
`EGFR
`
`pHer2/neu (Y877)
`
`pHer2/neu (Y1221/1222)
`
`Her2/neu
`
`pHer3 (Y1328)
`
`Her3
`
`pAkt (S473)
`
`Akt
`
`pERK (Y204)
`
`ERK
`
`♢-actin
`
`Figure 4. Analysis of the immunotoxins on HR cells. BT-474-M1 HR cells were derived by either continuous presence of Herceptin, or transient
`induction of 20 mg/ml EGF or 50 mg/ml NRG-1. Cells were treated with drugs at the concentration of 50 nM, unless otherwise noted. (a) Three-
`dimensional (3D) Matrigel growth assays with BT-474-M1 parental and HR cells by drug treatment for 12 days. The respective media were
`replenished every 3 days. Shown are representative images taken at  200 resolution. (b) Analysis of 3D Matrigel growth area of each
`spheroid. The calculation was applied with commercially available software ImageJ, and the average area presented as pixels (n425
`spheroids). (c) Western blot analysis of signaling pathways downstream of EGFR, Her2/neu and Her3 after 48 h treatment.
`
`similar pattern of tumor growth delay (B35 days). In contrast, the
`highest doses of 4D5/rGel (40 mg/kg) showed modest (B12 days)
`growth delay, and no significant tumor inhibition was observed
`below this dose. For the rGel or Herceptin treatment groups,
`tumor size showed continuous progressive growth. The experi-
`ments demonstrated that the FcR-dependent mechanisms are
`considered factors responsible for the effects of immunoconju-
`gates;
`likewise, the tumor penetration characterized by fusion
`constructs seemed to be efficient in regard to the tumor uptake
`
`and toxin delivery. There was no drug-induced toxicity observed
`in the mice at the doses of 4D5/rGel and rGel/4D5 used in this
`experiment. Doses of 40 mg/kg of Herceptin/rGel conjugate
`demonstrated a 17% weight loss in the test animals (Supple-
`mentary Figure S7).
`
`Immunohistochemistry studies
`The ability of immunotoxins to downregulate p-Her2/neu, p-Akt
`and phosphor-ERK (p-ERK)
`in vivo was
`investigated by
`
`Oncogene (2014) 429 – 439
`
`& 2014 Macmillan Publishers Limited
`
`IMMUNOGEN 2173, pg. 7
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`optimization of anti-Her2/neu immunotoxins
`Y Cao et al
`
`435
`
`Herceptin/rGel conjugate
`t1/2 (α) = 42.50 min
`t1/2 (β) = 42.67 h
`
`10
`
`1
`
`0.1
`
`0.01
`
`0
`
`12 16 20 24
`8
`4
`hours after injection
`
`Concentration (nM)
`
`rGel/4D5
`t1/2 = 34.75min
`
`12 16 20 24
`8
`4
`hours after injection
`
`10
`
`1
`
`0.1
`
`0.01
`
`0
`
`Concentration (nM)
`
`4D5/rGel
`t1/2 = 32.34min
`
`12 16 20 24
`8
`4
`hours after injection
`
`10
`
`1
`
`0.1
`
`0.01
`
`0
`
`Concentration (nM)
`
`Red: mouse CD31
`Green: rGel-based immunotoxins
`Blue: Hoechst 33342
`
`1h
`
`24h
`
`72h
`
`4D5/rGel
`
`rGel/4D5
`
`Herceptin/rGel
`conjugate
`
`rGel
`
`4D5/rGel
`
`rGel/4D5
`
`800
`
`600
`
`400
`
`200
`
`0
`
`800
`
`600
`
`400
`
`200
`
`72
`24
`1
`hours after injection
`
`rGel
`
`0
`
`Average intensity
`
`Average Intensity
`
`72
`24
`1
`hours after injection
`
`Herceptin/rGel conjugate
`
`800
`
`600
`
`400
`
`200
`
`0
`
`800
`
`600
`
`400
`
`200
`
`0
`
`Average intensity
`
`Average intensity
`
`72
`24
`1
`72
`24
`1
`hours after injection
`hours after injection
`Figure 5.
`(a) Blood pharmacokinetic analysis of
`In vivo pharmacokinetics and tumor distribution of anti-Her2/neu immunotoxins.
`immunotoxins in Balb/c mice. IRDye800-labeled immunotoxins were injected (intravenously) to mice, and blood samples were drawn for
`photoluminescence intensity analysis using an IVIS optical imaging system. Results are mean±s.d. (b) Immunofluorescence examination of
`the tumor penetration of immunotoxins relative to the tumor vasculature in SK-OV-3 tumor-bearing mice. Animal was killed and frozen tumor
`sections were prepared and detected by anti-rGel antibody (green) and anti-mouse CD31 antibody (red) 1, 24 and 72 h after injection.
`(c) Quantitative examination of drug intensity in tumor samples with software ImageJ, and the average area presented as views (n¼ 8).
`
`immunohistochemical analysis (Figure 6b). Control groups (phos-
`phate-buffered saline or rGel) had large area with viable SK-OV-3
`cells, with high levels of p-Her2/neu, p-Akt and p-ERK (dark brown
`areas), involving Her2/neu signaling cascade. Examining tumors
`from mice treated with either Herceptin or 4D5/rGel, we also
`
`found large areas of viable tumor cells with only slightly reduced
`levels of phosphorylation. In contrast, treatment with Herceptin/
`rGel conjugate or rGel/4D5 resulted in a significantly reduced level
`of phosphorylation of Her2/neu, Akt and ERK in tumors. These
`results suggest that rGel/4D5 is able to inhibit tumor growth
`
`& 2014 Macmillan Publishers Limited
`
`Oncogene (2014) 429 – 439
`
`IMMUNOGEN 2173, pg. 8
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`436
`
`optimization of anti-Her2/neu immunotoxins
`Y Cao et al
`
`rGel (20 mg/kg)
`rGel/4D5 (10 mg/kg)
`rGel/4D5 (20 mg/kg)
`rGel/4D5 (40 mg/kg)
`
`10
`
`20
`
`30
`
`40
`
`50
`
`60
`Days
`
`rGel (20 mg/kg)
`Herceptin (40 mg/kg)
`Herceptin/rGel conjugate (10 mg/kg)
`Herceptin/rGel conjugate (20 mg/kg)
`Herceptin/rGel conjugate (40 mg/kg)
`
`1500
`
`1200
`
`900
`
`600
`
`300
`
`0
`
`0
`
`Tumor volume (mm3)
`
`rGel (20 mg/kg)
`4D5/rGel (10 mg/kg)
`4D5/rGel (20 mg/kg)
`4D5/rGel (40 mg/kg)
`
`1500
`
`1200
`
`900
`
`600
`
`300
`
`0
`
`0
`
`10
`
`20
`
`30
`
`40
`
`60
`50
`Days
`
`Tumor volume (mm3)
`
`1500
`
`1200
`
`900
`
`600
`
`300
`
`0
`
`0
`
`10
`
`20
`
`30
`
`40
`
`60
`50
`Days
`
`Tumor volume (mm3)
`
`Vehicle
`
`rGel
`
`Herceptin
`
`4D5/rGel
`
`rGel/4D5
`
`Herceptin/rGel
`conjugate
`
`pHer2/neu
`
`pAkt
`
`pERK
`
`50 μm
`
`50 μm
`
`50 μm
`
`50 μm
`
`Figure 6.
`In vivo study of the immunotoxins against SK-OV-3 tumor xenografts in nude mice. (a) Treatment of SK-OV-3 tumors with
`immunotoxins at a dose of 10, 20 or 40 mg/kg, or rGel at a dose of 20 mg/kg, or Herceptin at a dose of 40 mg/kg. Mean tumor volume was
`calculated by W L H as measured by calipers. (b) Immunohistochemistry analysis of phosphorylation level of Her2/neu, Akt and ERK in
`SK-OV-3 tumors treated with immunotoxin. Dark brown color is indicative of p-Her2/neu, p-Akt or p-ERK. Shown are representative images
`taken at  400 resolution.
`
`efficiently as Herceptin/rGel conjugate, and this agent is effective
`in causing the downregulation of Her2/neu signaling pathways in
`tumors after systemic administration.
`
`DISCUSSION
`Initial studies of anti-Her2/neu rGel-based fusion constructs
`focused on the characterization of target specificity of scFv, and
`linker design between scFv and rGel.34,35 However, until now,
`there are no in-depth reports of the relationship of rGel-based
`fusion constructs in regard to large chemical conjugates. To the
`best of our knowledge,
`this
`is
`the first
`comprehensive
`examinations of the characteristics of rGel-based monovalent
`
`immunotoxin design compared with bivalent immunoconjugates
`with respect to the impact of valency and toxin delivery, cytotoxic
`activity,