Article pubs.acs.org/jmc
Proteolytically Stable Cyclic Decapeptide for Breast Cancer Cell Targeting Yogita Raghuwanshi,† Hashem Etayash,† Rania Soudy,† Igor Paiva,† Afsaneh Lavasanifar,† and Kamaljit Kaur*,†,‡ †
Faculty of Pharmacy and Pharmaceutical Sciences, University of Alberta, Edmonton, Alberta T6G 2E1, Canada Chapman University School of Pharmacy (CUSP), Harry and Diane Rinker Health Science Campus, Chapman University, Irvine, California 92618-1908, United States
‡
S Supporting Information *
ABSTRACT: Starting with a previously reported linear breast cancer targeting decapeptide WxEAAYQkFL, here we report the synthesis of a novel cyclic peptide analogue cyclic WXEAAYQkFL. The N- to C-terminus amide cyclized peptide with one D-amino acid (k) displayed higher uptake by breast cancer cells, with minimal uptake by the noncancerous cells compared to the linear peptide with two D-amino acids (x and k), and was stable toward proteolytic degradation. When immobilized on gold microcantilever surface, the cyclic peptide was able to capture breast cancer cells specifically and sense samples with ≥25 cancer cells/mL. Animal studies using mice carrying orthotopic breast MDA-MB-231 tumors showed that the cyclic peptide preferentially accumulates in tumor (2 h after injection) and is rapidly cleared from all other organs except kidneys and liver. The study highlights the discovery of a novel proteolytically stable cyclic peptide that can be used for targeted drug delivery or for enumerating circulating breast tumor cells.
■
INTRODUCTION Selective delivery of chemotherapeutic (CT) drugs to cancer cells is needed to avoid toxic side effects on the healthy cells and tissues. Among different strategies the use of ligand molecules such as antibodies, tumor homing peptides, and aptamers that target specific receptors on particular type of cancer cells has been particularly effective for selective drug delivery.1−3 These targeting ligands do not often exhibit anticancer properties; however their conjugation to anticancer drugs or drug carriers such as micelles and liposomes enhances the efficacy and therapeutic index of the chemotherapeutics.4−6 Peptides as drug carriers are believed to be one of the effective approaches to deliver chemotherapeutic agents to cancer site.7,8 Despite lacking sufficient stability due to rapid renal clearance, peptides have unique advantages as drugs carriers. Numerous peptides have been identified by phage display for targeting breast cancer cells and have been reported to show promising outcome for targeted delivery of drugs to tumors.8−11 One such peptide is a dodecapeptide 1 (p160) that was identified by random peptide phage display by Zhang et al. in 2001 (Table 1).12 Specifically, peptide 1 was isolated via selection rounds of a phage library on the human WAC 2 neuroblastoma cell line. Authors demonstrated high affinity to cancer cell lines MDA-MB-435 and WAC2 cells and low binding to primary HUVEC cells. In another study it was shown that when intravenously injected, 131I-labeled 1 preferentially accumulated in tumors than in normal organs © 2017 American Chemical Society
Table 1. Amino Acid Sequence of Breast Cancer Cell Targeting Peptidesa peptide
peptide sequence
peptide design
1 2 3 4 5 6 7 8
VPWMEPAYQRFL WXEAAYQRFL WxEAAYQrFL WxEAAYQkFL WxEAAYQKFL WXEAAYQkFL cWXEAAYQkFL cWXEAAYQKFL
phage display12 synthetic library (1st generation)14 synthetic library (2nd generation)15 synthetic library (2nd generation)15 new analogue new analogue new analogue new analogue
a
Lower case letters x, r, and k are D-norleucine, D-arginine, and Dlysine, respectively. c stands for cyclic peptide.
like heart, liver, spleen, lung, kidney, muscle, and brain.13 The stability studies of peptide 1 in human serum, however, revealed complete degradation by serum proteases after 4 h. In order to improve peptide 1 specific binding to breast cancer cells, a peptide array−whole cell binding assay was developed to screen for better analogues of 1.14 A library of 70 peptide sequences was synthesized on cellulose membrane and screened against human breast cancer cell lines that led to the identification of a decapeptide 2 (or peptide 18)14 that showed Received: January 31, 2017 Published: May 18, 2017 4893
DOI: 10.1021/acs.jmedchem.7b00163 J. Med. Chem. 2017, 60, 4893−4903
Journal of Medicinal Chemistry
Article
Figure 1. Structures of peptide 4 analogues, namely, linear 5 and 6, and cyclic 7 and 8. D-amino acids are shown in red and with lower case letters. c stands for cyclic peptide. X is Nle (norleucine).
drug.15,18 Recently we found that peptide 1 and 1-derived peptides bind to 67 kDa keratin 1 (KRT1) on breast cancer cell surface.19 Affinity column chromatography followed by tandem mass spectrometry and proteomics was used to identify the target receptor for 1 peptide(s) that is highly expressed on breast cancer cells. Surface plasmon resonance (SPR) was used to confirm the binding specificity of the peptide to a fragment of KRT1 (387−496 aa, Mwt = 38 kDa), and the Kd values found were ∼1.1 μM and 0.98 μM for peptides 1 and 3, respectively.19 Here we have designed cyclic analogues of peptide 4 to enhance the affinity and specificity toward breast cancer cell lines while maintaining the proteolytic stability. In addition, we have reduced the number of D-amino acids from two to one or
better binding to breast cancer cells than 1 (Table 1). A proteolytically stable analogue of peptide 2, namely, peptide 3 (18-4 or WxEAAYQrFL), was obtained by substitution of two labile amino acid residues with D-amino acids.15 In addition, 3 was found to be safe with minimal cellular toxicity.15−17 Further, an analogue to 3 with r8k substitution (4 or 18-4a, WxEAAYQkFL) was designed to facilitate side chain attachment of doxorubicin (Dox) and synthesize peptide−Dox conjugate.18 The conjugate functions as a breast cancer prodrug to selectively target and deliver Dox to breast cancerous cells with reduced delivery to normal cells. Peptide−Dox conjugate (4−Dox), where peptide was attached to Dox via an ester linkage, enhanced Dox selectivity 40 times to breast cancerous cells than the noncancerous cells compared to the free 4894
DOI: 10.1021/acs.jmedchem.7b00163 J. Med. Chem. 2017, 60, 4893−4903
Journal of Medicinal Chemistry
Article
Figure 2. Solid phase synthesis of FITC labeled cyclic peptide 7.
The peptides were synthesized following stepwise solid phase peptide synthesis on Rink amide resin as a polymeric support. For the synthesis of cyclic peptides, orthogonally protected Nα-Fmoc-L-glutamic acid α-allyl ester (Fmoc-Glu-OAll) was used that served as the cyclization position for the peptides. Activated Fmoc-Glu-OAll was coupled to Rink amide resin from the side chain of Glu, followed by stepwise coupling of all amino acids (Figure 2). Removal of the allyl ester using Pd (PPh3)4 and PhSiH3 and the N-terminal Fmoc group using 20% piperidine allowed on resin peptide cyclization (see Experimental Section). Next, the Dde protecting group of the lysine side chain was removed using hydrazine monohydrate followed by coupling of β-alanine linker and FITC to the lysine side chain to give the final labeled cyclic peptide. For the preparation of labeled linear peptides, the N-terminal amino group of Trp residue was used to attach β-alanine and FITC. All peptides were purified using reversed phase (RP) HPLC and were obtained in good yields ranging from 50% to 75% (Table S1, Figures S1 and S2, Supporting Information). In Vitro Cell Uptake for Screening Peptides. Cell binding and uptake studies of designed analogues (5, 6, and 7) were investigated using flow cytometry assay. The lead sequence 4, reported previously, was used as a control peptide. The tumor targeting ability of these peptides (at 1 μM) was studied using breast cancer cell lines, MDA-MB-435, MDAMB-231 and MCF-7, while noncancerous cell lines HUVEC and MCF-10A were used as controls. Among the four peptides, the cyclic peptide analogue 7 showed highest binding to breast cancer cell lines as evidenced by the increase in the fluorescence of cells treated with FITC-7 relative to the untreated cells or cells treated with other peptides (Figure 3). The linear peptide analogue 6 with one D-amino acid (D-Lys8) showed comparable uptake to the lead 4 peptide, and the conversion of this linear sequence into conformational constrained cyclic sequence resulted in remarkable increase (2-fold or higher) in binding
none, as it is hypothesized that cyclization may impart sufficient stability to the peptide structure. The results show that the cyclic peptide analogues display higher uptake by breast cancer cells than all other analogues tested so far, and show minimal uptake by the noncancerous cells. One of the cyclic analogues with one D-amino acid (7) was sufficiently stable toward proteolytic degradation. When immobilized on gold microcantilever surface, the peptide was able to capture breast cancer cells specifically. Preliminary animal studies using mice carrying orthotopic breast MDA-MB-231 tumors were also performed to track the Cy5.5 labeled peptide 7 in live mice. The study highlights the discovery of a novel proteolytically stable cyclic peptide 7 for breast cancer targeting that can be used for targeted drug delivery or for capturing circulating breast tumor cells from human blood samples.
■
RESULTS AND DISCUSSION Design of Peptide 4 Analogues. Peptide 4 is a positively charged (net charge of +1) linear decapeptide with two Damino acids (D-norleucine and D-lysine) in the sequence.18 Here we have designed four analogues of peptide 4, two linear and two cyclic decapeptides (Figure 1). The design strategy involved replacement of either norleucine (x) or lysine (k) amino acid in the lead sequence with the corresponding Lresidues and/or cyclization of the sequence. We hypothesized that a single D-amino acid and/or cyclization of the decapeptide 4 may be sufficient for the proteolytic stability of the peptide while maintaining for breast cancer cell targeting. This will also avoid the use of two D-amino acids in the sequence. The linear peptide analogues consist of a single D-amino acid in the sequence, namely, peptide 5 with D-norleucine and peptide 6 with D-lysine (Table 1). In addition, two cyclic analogues were designed, one with a single D-amino acid in the sequence (7) and the other one with all L-amino acids (8). 4895
DOI: 10.1021/acs.jmedchem.7b00163 J. Med. Chem. 2017, 60, 4893−4903
Journal of Medicinal Chemistry
Article
Figure 3. Peptide uptake by the breast cancer cells (left) MDA-MB-435, MCF-7, and MDA-MB-231 and the control cells (right) HUVEC and MCF-10A measured using BD FACS Canto II flow cytometry. The peptides (10−6 mol/L or 1 μM) FITC-4 (blue), FITC-5 (green), FITC-6 (orange), and FITC-7 (light green) were incubated with the cells for 30 min at 37 °C. Autofluorescence of the cells is shown by shaded area. (A) shows overlays from one representative experiment for the uptake of peptides. (B) shows mean fluorescence intensity (MFI) of FITC positive cells from three independent experiments ± SD.
cancer cells, all peptides showed relatively low binding to noncancerous cells. Next, we compared the breast cancer cell uptake of the two cyclic peptides, 7 and 8 (with all L-amino acids), versus the noncancerous cell uptake to three previously reported lead sequences (linear peptides), namely, 1, 2, and 4 (Table 1). Two peptide concentrations were used, 1 μM and 10 μM. As shown in Figure S3, the magnitude of cell uptake increased considerably at higher peptide concentration (10 μM). The intensity of cell-associated fluorescence at 10 μM peptide concentration was at least 2-fold higher when compared to the cell-associated fluorescence observed at 1 μM peptide concentration. This trend was observed for all breast cancer cell lines used in the study. The mean fluorescence intensities
in all three breast cancer cell lines. Contrary, the cellular uptake of the linear 5 with one D-norleucine (D-Nle2) was much less than that observed for the lead 4 peptide. These results demonstrate that the presence of D-Lys8 is well tolerated and may not affect the interaction of peptide with its target receptor. Furthermore, cyclization of the peptide led to a slight increase in its hydrophobicity, observed by HPLC data with higher retention time (Table S1); this may cause an enhancement in the membrane permeability of the peptide and thereby its cellular uptake. Previous studies have shown that cyclization reduces the hydrogen bonding and hydrodynamic radius in solution and increases the lipophilicity of the peptides, thereby enhancing the membrane permeability, which could ultimately result in high cell uptake.20,21 In comparison to 4896
DOI: 10.1021/acs.jmedchem.7b00163 J. Med. Chem. 2017, 60, 4893−4903
Journal of Medicinal Chemistry
Article
Figure 4. Competitive uptake of FITC labeled 7 peptide in the presence of unlabeled excess peptide using MCF-7 and MDA-MB-231 cells. The cells, seeded in 12-well plates and grown overnight, were preincubated with unlabeled 7 (50 μM) for 15 min at 37 °C. Thereafter, the FITC labeled 7 (1 μM) was added and incubated for 30 min. The cells were subjected to flow cytometry, and data were analyzed using FlowJo software. (A) shows the comparative uptake profile of peptide with or without free 7 peptide. (B) shows bar diagram for percent of FITC positive cells in the absence or presence of free 7 from three independent experiments ± SD. The asterisk (∗) denotes statistically significant difference (p < 0.05).
(MFIs) of FITC-7 at 10 μM were 2771 ± 78, 2607 ± 116, and 1761 ± 83 when incubated with MCF-7, MDA-MB-231, and MDA-MB-435, respectively, whereas the MFIs were 179 ± 25 and 247 ± 27, respectively, when incubated with normal HUVEC and MCF10A (Figure S3). In comparison, the MFIs for the linear peptide with two D-amino acids x and k (FITC-4) were 965 ± 44, 1666 ± 97, and 1031 ± 54 when incubated with MCF-7, MDA-MB-231, and MDA-MB-435, respectively. Overall the uptake was observed in the following order 1 < 2 < 4 < 7 < 8 with peptide 8 displaying highest cellular uptake. Noncancerous cell uptake (HUVEC and breast tissue derived MCF-10A cells) of all the peptides was minimal. Even at high peptide concentration (10 μM) the uptake of the peptides was low suggesting the presence of smaller number of putative receptors on these cells. Cellular Uptake through Specific Binding. Flow cytometry was employed to determine the binding specificity of the peptides using a competitive binding assay. Breast cancer
cells MCF-7 or MDA-MB-231 were incubated with FITClabeled 7 in the presence or absence of excess (50-fold) unlabeled 7. After 30 min, the cells were analyzed using flow cytometry and a decrease in % (percent) of FITC positive cells was calculated when the cells were incubated in the presence of unlabeled excess peptide. The percent of FITC positive cells reduced from 95% to 22% (up to ∼76% decrease) for MCF-7 cells and from 98% to 27% (∼72% decrease) for MDA-MB-231 cells (Figure 4), suggesting specific binding of the peptide 7 to the cancer cells. The results support our conjecture that uptake of peptides can be reduced when the putative receptor is preoccupied with excess ligand. The presence of FITC at the εamino group of lysine in the cyclic peptide does not affect binding. We have previously shown that blocking the lysine side chain amino group of the peptide with doxorubicin in the peptide−doxorubicin conjugate maintains specific binding and uptake of the conjugate in breast cancer cells.18 4897
DOI: 10.1021/acs.jmedchem.7b00163 J. Med. Chem. 2017, 60, 4893−4903
Journal of Medicinal Chemistry
Article
Figure 5. Uptake of peptides by MCF-7 (left) and MDA-MB-231 (right) cancer cells at 4 °C. The MCF-7 or MDA-MB-231 cells were grown overnight in 12-well plates in respective growth media and next day incubated with 10 μM FITC labeled peptide analogs for 30 min at 4 °C. (A) The flow cytometry overlays shown are from one representative experiment out of three independent experiments performed. (B) Bar graph shows mean fluorescence intensity (MFI) of three experiments ± SD.
Figure 6A shows differential nanomechanical deflection (nm) of the microcantilever sensor functionalized with peptide 7 to cancer cell line (MCF7) in comparison to noncancerous cells MCF10A and a negative control peptide sensor. As indicated, the sensing cantilever beam, functionalized with 7, expresses statistically significant (P = 0.0032) nanomechanical bending in response to cancer cells (MCF7) binding compared to the noncancerous cells and the negative control sensor. When cell lines were injected into the sensor array (100 cells mL−1), the peptide 7 coated cantilever beam showed significant deflection for MCF-7 cancer cells (301 ± 19 nm) compared to noncancerous MCF10A cells (90 ± 24 nm). The variation in the peptide 7 binding affinity between cancerous and noncancerous cells can be attributed to the differential expression levels of specific peptide-binding receptors (putative) in cancerous and noncancerous cells. As indicated earlier, peptide 7 is most likely targeting a specific receptor expressed in much higher levels on the surface of cancer cells than that of noncancerous cells. In contrast to peptide 7, the response of the negative control peptide to cancer cells was significantly low. The observed deflection was 18 ± 13 nm for MCF7 and 17 ± 17 nm for noncancerous MCF10A cells. This clearly shows specific binding behavior of 7 to MCF7 breast cancer cells. Sensitivity of peptide 7 toward capturing cancer cells on a microcantilever array sensor was determined by exposing the sensor to various concentrations of MCF7 spiked in buffer solution (Figure 6B). The results show capability of the peptide 7-coated lever to differentiate as low as 25 cancer cells per mL in buffer solution from the background noise. It is also noted that the 7 coated levers deflect linearly with increase in the number of cancer cells in the samples before reaching saturation at ∼600 cells mL−1. These results suggest that the cyclic peptide 7 can be used for capturing breast cancer cells and has a potential to be used for cancer cell diagnostics. Further studies in this direction, including exploration of the
Next the two cyclic peptides 7 and 8 were evaluated for binding at low temperature (4 °C). Breast cancer cells MCF-7 and MDA-MB-231 were kept on ice for 5 min, and FITC labeled peptides (10 μM) dissolved in ice-cold growth media were added. After 30 min of incubation at 4 °C, the cells were analyzed using flow cytometry. The results show that the binding of peptides was drastically reduced compared to that observed at 37 °C (Figure 5). Consistent with the previous reported studies, the results suggest that once the cells are metabolically rendered inactive, the binding of peptides is severely dampened.12,13 Specific Binding of Peptides Immobilized on Microcantilever Surface. The specific binding of the peptides to breast cancer cells was also evaluated using microcantilever based peptide sensors. Microcantilevers as biosensors have been used extensively to observe binding of biomolecules and cells.22−25 Here the peptides (7 or a negative control) were immobilized on a gold microcantilever surface and the interaction between the peptide and the cells was studied using the microcantilever deflection as described previously.23 Briefly, the interaction between the analyte (cell) and ligand (peptide) generates a change in the differential surface stress between the top and bottom interfaces of the microcantilever, causing it to deflect or bend by a certain degree that is expressed by Stoney’s formula.
⎤ ⎡ Et 2 Δσ = ⎢ δ 2⎥ ⎣ 3(1 − v)L ⎦
(1)
where Δσ is the change in surface stress (or surface energy) due to the analyte−ligand interaction, E is the elastic modulus (Young’s modulus) of the cantilever, v is Poisson’s ratio, while L and t are the length and thickness of the cantilever, respectively. This formula (eq 1) is used to determine the differential surface stress upon peptide microcantilever when breast cancer cells are captured. 4898
DOI: 10.1021/acs.jmedchem.7b00163 J. Med. Chem. 2017, 60, 4893−4903
Journal of Medicinal Chemistry
Article
MALDI-TOF mass of 972.6/973.6 was observed which eluted ∼12 min in RP-HPLC. In Vivo Tracking of the Cy5.5 Labeled Peptide 7 in Live Mice. The luciferase expressing MDA-MB-231 cell line was used in the animal studies to be able to visualize the presence of primary and potential metastatic tumors in the animals in this experiment. Bioluminescence imaging of different organs yielded negative results, indicating the absence of noticeable tumor metastasis in other organs under current experimental conditions. The results of our preliminary in vivo imaging studies in live animals showed the labeled peptide 7 to distribute to several organs nonspecifically within 0.5 h but clear very rapidly from these organs (within 2 h) as well. At the 2 h time point, the peptide was mainly observed in the liver, spleen, kidney, and orthotopic breast tumor location (Figure 7A−C). Ex vivo data following excision of the tumor confirmed these results and showed the presence of peptide in the tumor tissue at 1 h after injection (Figure 7D). At 6 h, the peptide was cleared from tumor but was still observed in potential sites of its elimination, i.e., kidneys and liver, up to 24 h after injection (Figure 7A,B). The preferential accumulation of cyclic peptide 7 in tumor at 2 h after injection, in spite of its clearance from other organs, may be attributed to the existence of specific interaction between this peptide and its receptors present only in tumor tissue. A previous study has reported on the biodistribution of peptide 1 (the lead peptide to peptide 7), where 1 h following intravenous injection of 131I-labeled 1 to mice carrying subcutaneous MDA-MB-435 tumors, nonspecific distributions in lung, spleen, liver, kidney, and the tumor were seen.13 In that study, following perfusion of the mice with 0.9% NaCl, peptide uptake in all organs was reduced, but it stayed the same in the tumor pointing to better peptide interaction with the tumor tissue. We have seen the peptide to be washed quickly (
