Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2018 Jan;17(1):169-182.
doi: 10.1158/1535-7163.MCT-17-0092. Epub 2017 Sep 22.

Targeting Phosphatidylserine With Calcium-Dependent Protein-Drug Conjugates for the Treatment of Cancer

Affiliations
Free PMC article

Targeting Phosphatidylserine With Calcium-Dependent Protein-Drug Conjugates for the Treatment of Cancer

Ran Li et al. Mol Cancer Ther. .
Free PMC article

Abstract

In response to cellular stress, phosphatidylserine is exposed on the outer membrane leaflet of tumor blood vessels and cancer cells, motivating the development of phosphatidylserine-specific therapies. The generation of drug-conjugated phosphatidylserine-targeting agents represents an unexplored therapeutic approach, for which antitumor effects are critically dependent on efficient internalization and lysosomal delivery of the cytotoxic drug. In the current study, we have generated phosphatidylserine-targeting agents by fusing phosphatidylserine-binding domains to a human IgG1-derived Fc fragment. The tumor localization and pharmacokinetics of several phosphatidylserine-specific Fc fusions have been analyzed in mice and demonstrate that Fc-Syt1, a fusion containing the synaptotagmin 1 C2A domain, effectively targets tumor tissue. Conjugation of Fc-Syt1 to the cytotoxic drug monomethyl auristatin E results in a protein-drug conjugate (PDC) that is internalized into target cells and, due to the Ca2+ dependence of phosphatidylserine binding, dissociates from phosphatidylserine in early endosomes. The released PDC is efficiently delivered to lysosomes and has potent antitumor effects in mouse xenograft tumor models. Interestingly, although an engineered, tetravalent Fc-Syt1 fusion shows increased binding to target cells, this higher avidity variant demonstrates reduced persistence and therapeutic effects compared with bivalent Fc-Syt1. Collectively, these studies show that finely tuned, Ca2+-switched phosphatidylserine-targeting agents can be therapeutically efficacious. Mol Cancer Ther; 17(1); 169-82. ©2017 AACR.

Figures

Figure 1
Figure 1
Generation and characterization of PS-targeting agents. A, schematic representation of PS agents (left panel). Filled circles and rectangles represent the PS-binding domains and IgG1 hinge region, respectively. Right panel shows reducing SDS-PAGE analyses of the PS-specific Fc fusions, with molecular weights (MW) shown in kDa on the left. B, lipid binding profiles of PS-specific Fc fusions using lipid-coated nitrocellulose membranes. Bound proteins were detected with goat anti-human IgG (H+L) antibody conjugated with HRP. C, binding of PS-specific Fc fusions to PS-positive 2H11 and MDA-MB-231 cells using flow cytometry analysis, using Alexa 647-labeled anti-human IgG (H+L) for detection. 2nd and Fc represent negative controls using secondary conjugate or recombinant Fc fragment, respectively. D, pharmacokinetic analyses of PS-specific Fc fusions in BALB/c SCID mice (n = 5 mice/group). Whole body and blood levels of radioactivity were measured at the indicated time points. E, areas under curves in panel D for whole body (upper panel) and blood (lower panel) counts were quantitated. F, nude mice bearing orthotopic human MDA-MB-231 tumors (n = 3 mice/group) were injected with IRDye800CW-labeled PS-specific Fc fusions, and NIR fluorescence images acquired at the indicated time points. G, tumor-associated fluorescence intensities at 48 hours in F normalized to the corresponding tumor volumes were quantitated. H, female BALB/c SCID mice bearing MDA-MB-231 tumors (n = 3 mice/group) were injected with IRDye800CW-labeled PS-specific agents. 48 hours post-injection, tumors were dissected out and NIR images were acquired. I, tumor-associated fluorescence intensities in H normalized to the corresponding tumor weights. Statistically significant differences in E, G and I were analyzed using one-way ANOVA followed by Tukey post hoc test (**, P < 0.01; ***, P < 0.001; ****, P < 0.0001). Error bars in D, E, G and I represent SEM.
Figure 2
Figure 2
Cell binding and internalization of PS-specific Fc fusions containing Syt1 C2A. A, schematic representation of bivalent and tetravalent PS-specific Fc fusions with filled circles representing the Syt1 C2A domain (left panel). Right panel shows reducing SDS-PAGE analyses of the Syt1-Fc fusions, with molecular weights (MW) shown in kDa on the right. B, Fc fusions (25 nM) were incubated with nitrocellulose membranes coated with the indicated amounts of PS. Bound proteins were detected with goat anti-human IgG (H+L) antibody conjugated with HRP. C, lipid binding profiles of Syt1-Fc fusions using lipid-coated nitrocellulose membranes as shown in Figure 1B. Bound proteins were detected using goat anti-human IgG (H+L) antibody conjugated with HRP. D, 2H11 cells were treated with 50 nM docetaxel for 72 hours, or treated with vehicle control (DMSO), and incubated with 50 nM control IgG or PS-specific Fc fusions. Bound Fc fusion was detected using Alexa 488-labeled anti-human IgG (H+L), followed by flow cytometry analyses (MFI, mean fluorescence intensity). E, 2H11 cells were incubated with Alexa 647-labeled PS-specific Fc fusions on ice at optimized concentrations (220 nM for Fc-Syt1 and 40 nM for Syt1-Fc-Syt1) to achieve similar levels of surface binding. Cells were then incubated at 37°C for the indicated time points. Surface bound Fc fusions were stripped using 5 mM EDTA and internalized proteins quantitated by flow cytometry analyses. F and G, 2H11 (F) or MDA-MB-231 (G) cells were incubated with 50 nM control IgG or PS-specific Fc fusions at 37°C for four hours. Cells were fixed, stained with Cy3/Alexa 555-labeled anti-human IgG (H+L) and LAMP-1-specific antibody followed by Alexa 488-labeled secondary antibody for detecting LAMP-1. Fluorescence images were acquired and Cy3/Alexa 555, Alexa 488 and DAPI are pseudo-colored red, green and blue, respectively, in the overlays. Scale bars: 10 μm (F) and 5 μm (G). For D and E, statistically significant differences were analyzed using two-way ANOVA followed by Tukey post hoc test (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001). Error bars in D and E represent SEM.
Figure 3
Figure 3
PS-specific PDCs bind to PS in a Ca2+-dependent way, and dissociate from PS in early endosomes prior to lysosomal delivery and disruption of microtubule networks in target cells. A, schematic representation of PS-specific PDCs (left panel). The hinge cysteines were reduced and conjugated to four molecules of MMAE (small filled circles). Right panel shows reducing SDS-PAGE analyses of the unconjugated or conjugated PS-specific Fc fusions and control IgG, with molecular weights (MW) shown in kDa on the left. B, PS-specific PDCs or MMAE-conjugated control IgG were incubated with PS-coated beads in the presence of the indicated Ca2+ concentrations (left panel) or pH levels (right panel). Bead-associated proteins were analyzed using immunoblotting and detection with goat anti-human IgG (H+L) conjugated with HRP. C, MDA-MB-231 cells were incubated with 100 nM PS-specific PDCs or MMAE-conjugated control IgG for 30 minutes. Cells were fixed, stained with Alexa 555-labeled anti-human IgG (H+L) and EEA1-specific antibody followed by Alexa 488-labeled secondary antibody for detecting EEA1. Fluorescence images were acquired and Alexa 555, Alexa 488 and DAPI are pseudo-colored red, green and blue, respectively, in the overlays. The endosomes in boxed regions (labeled a and b) are cropped and expanded. Fluorescence intensities along the dotted lines in the overlays for these endosomes are shown in the fluorescence intensity plots. Scale bars: 5 μm (left panels; whole cell images) and 1 μm (right panels; cropped endosomes). D, MDA-MB-231 cells were incubated with 50 nM MMAE-conjugated control IgG or PS-specific Fc fusions at 37°C for four hours. Cells were fixed, stained with Alexa 555-labeled anti-human IgG (H+L) and LAMP-1-specific antibody followed by Alexa 488-labeled secondary antibody for detecting LAMP-1. Fluorescence images were acquired and Alexa 555, Alexa 488 and DAPI are pseudo-colored red, green and blue, respectively, in the overlays. Scale bars: 10 μm. E and F, 2H11 (E) and MDA-MB-231 cells (F) were treated with 100 nM or 50 nM PS-specific PDCs or MMAE-conjugated control IgG for 10 or 20 hours, respectively. Cells were then fixed, stained with tubulin-specific antibody followed by Alexa 555-labeled secondary antibody and imaged. Alexa 555 and DAPI are pseudo-colored red and blue, respectively, in the overlays. Scale bars: 15 μm (E) and 10 μm (F).
Figure 4
Figure 4
PS-specific PDCs inhibit cancer cell growth and survival in vitro. PS-positive cancer endothelial cells (2H11), breast cancer cells (T-47D, SK-BR-3 and MDA-MB-231) and prostate cancer cells (LNCaP and 22Rv1) were treated with antibody/protein-drug conjugates at the indicated concentrations. Cell viability following 72 hours (2H11), 96 hours (SK-BR-3, MDA-MB-231 and 22Rv1) or 120 hours (T-47D) is shown. Representative data from two or three independent experiments for each cell line are presented.
Figure 5
Figure 5
Bivalent Fc-Syt1_MMAE inhibits tumor growth in vivo and targets multiple PS-positive cells in the tumor tissue. A, pharmacokinetic analyses of PS-PDCs in BALB/c SCID mice (n = 5 mice/group). Whole body and blood radioactivities were measured at the indicated time points. B, areas under curves in A were quantitated and statistically significant differences analyzed using unpaired Student’s t-test (****, P < 0.0001). C-E, BALB/c SCID mice were implanted with MDA-MB-231 (C and E) or LNCaP (D) tumors. Mice (n = 5–6 mice/group) were treated (day 33–60 in C, E and day 27–59 in D) with either unconjugated or MMAE-conjugated Fc fusions at a dose of 1 nmole/mouse (4.1 mg/Kg for Fc-Syt1 or Fc-Syt1_MMAE, 5.6 mg/Kg for Syt1-Fc-Syt1 or Syt1-Fc-Syt1_MMAE) twice per week. PBS was delivered as vehicle control. Tumor volumes (C, D) and body weights (E) were measured twice per week. Statistically significant differences (Fc-Syt1_MMAE vs PBS in C; Fc-Syt1_MMAE vs. PBS or Syt1-Fc-Syt1_MMAE in D) at treatment end points were analyzed using one-way ANOVA followed by Bonferroni post hoc test (*, P < 0.05; ***, P < 0.001; ****, P < 0.0001). Error bars in all panels represent SEM. F and G, BALB/c SCID mice bearing MDA-MB-231 tumors were treated (i.p.) with 5 mg/Kg docetaxel 72 and 48 hours before delivery of 1 nmole Fc-Syt1_MMAE. PBS was delivered as vehicle control. Mice were perfused either 1 hour (F) or 24 hours (G) post-injection of Fc-Syt1_MMAE. Tumors were dissected out and tissue sections were fixed and stained with Alexa 555-labeled anti-human IgG (H+L), mouse CD31, F4/80 or human Ki-67-specific antibodies followed by Alexa 488-labeled secondary antibodies for detection of CD31/Ki67 and Alexa 647-labeled secondary antibody for detection of F4/80. Confocal images were acquired and Alexa 555, Alexa 488, Alexa 647 and DAPI are pseudo-colored red, green, white and blue, respectively, in the overlays. Scale bars: 50 μm (F) and 20 μm (G).
Figure 6
Figure 6
Therapeutic effects of Fc-Syt1_MMAE are dependent on PS binding. A, reducing SDS-PAGE analyses of the unconjugated or MMAE-conjugated PS-specific Fc fusions and control IgG1 Fc, with molecular weights (MW) shown in kDa on the left. B, lipid-coated nitrocellulose membranes were incubated with 2 μg/ml Fc-Syt1 or Fc-Syt1(DN), and bound proteins detected using goat anti-human IgG antibody conjugated with HRP. C, 2H11 cells were treated with 50 nM docetaxel for 72 hours, or treated with vehicle control (DMSO), and incubated with 5 μg/ml control Fc, Fc-Syt1 or Fc-Syt1(DN). Bound Fc or Fc fusion was detected using Alexa 488-labeled anti-human IgG (H+L), followed by flow cytometry analyses. Statistically significant differences were analyzed using two-way ANOVA followed by Tukey post hoc test (***, P < 0.001; ****, P < 0.0001). D, female BALB/c SCID mice (n = 6 mice/group) bearing MDA-MB-231 tumors were treated with the indicated agents at a dose of 1 nmole/mouse (4.1 mg/Kg for Fc-Syt1_MMAE or Fc-Syt1(DN)_MMAE, 2.6 mg/Kg for Fc_MMAE) twice per week for four weeks (day 28–56) and tumor sizes were measured for a further 2.5 weeks (day 56–74). Statistically significant differences between Fc-Syt1_MMAE and Fc-Syt1(DN)_MMAE treatment groups at the treatment end point were analyzed using one-way ANOVA followed by Bonferroni post hoc test (***, P < 0.001). E, tumors in each group shown in D were isolated and photographed. Scale bar: 1 cm. Error bars in C and D indicate SEM.

Similar articles

See all similar articles

Cited by 1 article

Publication types

Feedback