NANO LETTERS
Neurotoxin Quantum Dot Conjugates Detect Endogenous Targets Expressed in Live Cancer Cells
2009 Vol. 9, No. 7 2589-2599
Rebecca L. Orndorff† and Sandra J. Rosenthal*,†,‡,§,| Department of Chemistry, Department of Pharmacology, Vanderbilt School of Medicine, Department of Physics, Department of Chemical and Biomolecular Engineering, Vanderbilt UniVersity, NashVille, Tennessee 37235-1822 Received March 12, 2009; Revised Manuscript Received May 26, 2009
ABSTRACT High affinity peptide neurotoxins are effective agents for integrating technological advances with biological inquiries. Both chlorotoxin (CTX) and dendrotoxin-1 (DTX-1) are peptide neurotoxins demonstrated to bind targets expressed by glioma cancer cells and are suitable ligands for quantum dot (QD) live cell investigations. Here, we present dual labeling of endogenously expressed cellular proteins within living cells utilizing high affinity peptide neurotoxins conjugated to QDs. Multiplexing experiments reveal quantifiable evidence that CTX and DTX-1 conjugated QDs may potentially be used as a live assessment of markers toward identification of cancer cell presence.
Peptide-mediated quantum dot (QD) detection has been demonstrated to be an effective methodology for probing biological systems with high specificity. Targeted neurotoxin approaches have been employed to selectively label proteins for monitoring, detection, and destruction. Although designed to induce disruption of its targeting system, each toxin provides the possibility to be utilized as a therapeutic vehicle for disease detection and destruction, and as a means to study its target in greater depth. Typically, peptide neurotoxins inhibit ion channels by acting as antagonists and result in the cessation of neurotransmission propagation and loss of channel functionality. The high affinity characteristic of peptide neurotoxins enables the toxins to be used as ligands for thorough study of their biological targets,1-4 as well as targets lacking adequate ligands.5 Previous studies have demonstrated that a well-characterized ion channel may be detected with QDs and a high affinity peptide neurotoxin.6 QD cellular targeting may be mediated with use of toxins that interact and bind to endogenously expressed cell membrane proteins and ion channels. QDs exhibit characteristics that allow for multiplexing analyses;7-12 thus, the fluorescent nanocrystals may be utilized to detect multiple endogenous targets simultaneously with adequate high affinity ligands. Since endogenous expression levels vary not only between cell types, but also between cellular targets * To whom correspondence should be addressed. E-mail: Sandra.J.Rosenthal@ Vanderbilt.edu. Phone: +1 615 322 2633. Fax: + 1 615 343 1234. † Department of Chemistry. ‡ Department of Pharmacology, Vanderbilt School of Medicine. § Department of Physics. | Department of Chemical and Biomolecular Engineering. 10.1021/nl900789e CCC: $40.75 Published on Web 06/09/2009
2009 American Chemical Society
being examined, visualization of dye conjugate detection may be hindered because of low emission intensity; however, near 100% QD quantum yield13 provides the possibility to visually detect low expression levels due to brighter emission.14-16 Since high affinity peptide toxins show marked specificity for their targets,17,18 minimal cross-talk and nonspecific binding is anticipated. This indicates that there is the potential to use more than one peptide toxin on a cellular assay simultaneously, although toxin exposure at high concentrations may have adverse effects on the assays. QD photoemission properties enable lower concentrations to be utilized for the studies because visualization may be achieved at lower expression levels. Previous studies support the use of toxins as ligands with QDs toward detection of endogenous cellular proteins for applications involving multispectral analyses.12,19 One toxin of interest toward these applications is an insect venom toxin isolated from the scorpion, Leiurus quinquestriatus, known as chlorotoxin (CTX).20 CTX is 36 amino acids in length with four disulfide bonds, 2Cys-19Cys, 5 Cys-28Cys, 16Cys-33Cys, and 20Cys-35Cys.21 The toxin appears to bind matrix metalloproteinase II (MMP-2),22,23 an extracellular matrix enzyme that exhibits gelatinase activity. Additionally, CTX preferentially binds to cancer cells over non-cancer cells.22,24 Of particular biological interest is the primary brain glial cell cancer, glioma. Glioma cells are malignant cancer cells that infiltrate healthy brain tissue and evade complete surgical resection through not yet fully understood mechanisms.25 CTX binds effectively to MMP-2 endogenously expressed by glioma cells,22,23,25 and exposure results in loss of gelatinase activity, disruption in
chloride channel currents, reduction in both MMP-2 and chloride channel expressions, and internalization of chloride channels.22-24,26,27 In 2003, Sontheimer and colleagues determined that CTX targets MMP-2 through recombinant His tagging methodologies coupled with analytical techniques including mass spectrometry.22 Olson and co-workers provided further evidence in 2007 that MMP-2 is the CTX receptor with colocalization of MMP-2 antibody fluorescence with CTX:Cy5.5 conjugate fluorescence in MMP-2 transfected MCF7 cells.23 These studies garner evidence that MMP-2 is influential in glioma cell proliferation and raise questions regarding the nature of the chloride channel interactions with MMP-2. This property indicates that CTX has the potential to be used as a therapy for gliomas and affords insights into the understanding of all cancer metastasis and into the poorly characterized chloride channel superfamily.28 A greater understanding regarding the mechanistic aspects of glioma metastasis through the narrow cavities of the brain may afford insight into the fundamental basis of most cancer metastasis mechanisms. CTX conjugates are currently in phase I/II clinical trials for tagging glioma cells for surgical resection.23 These previous studies indicate that CTX may be a suitable peptide for QD conjugation toward use in live cell detection of endogenous targets. Unlike previous studies that have used dye conjugates with CTX, specifically Cy5.5,23 QDs are fluorescent probes that emit within the visible region of the electromagnetic spectrum. They may also be excited using a single excitation source when present with additional fluorescent probes. CTX has been shown to be amenable to conjugation to fluorescent dyes, which thus indicates it may be suitable for incorporation into QD high affinity probe development. By conjugating CTX to QDs, which exhibit greater photostability than their fluorescent dye counterparts, the CTX:QD fluorescent probe has the potential to be utilized as a visible spectrum fluorescent marker to examine the metastatic movement of glioma cells through various environments, as well as used to monitor the viability of the cells during therapeutic testing with other anticancer agents. A second toxin suited for QD applications is a member of the dendroxotins (DTXs), a family of high affinity cation channel antagonists. DTXs are isolated from the venoms of the mamba snake family and act by blocking various subtypes of voltage-gated potassium channels. The peptides are approximately 7000 Da and range in size from 57-60 amino acids in length.29 Their mechanism of action results in depletion of acetylcholine at the neuromuscular junction through extension of the action potential. DTXs interfere with the ability of potassium channels to perform their primary responsibilities: (1) cell membrane potential maintenance, and (2) cell membrane repolarization following action potential propagation. Interference with potassium channel functions results in muscle convulsions and eventual death due to hyperexcitability of the muscles.30 Dendrotoxin-1 (DTX-1), a peptide isolated from the Black Mamba, has been used to target potassium channels on C6 glioma cells. It was found by Allen et al.31 that unstimulated C6 glioma cells express potassium channels of the subtype, Kv1.1, which 2590
are targeted by DTX-1. DTX-1 also targets two other Shaker family potassium channels, Kv1.2 and Kv1.6. In the study by Allen et al.,31 only Kv1.1 was detected to be present within the cells. DTX-1 blockade caused significant reduction in potassium currents, and not all cells recovered completely from exposure.31 Significant reduction in potassium currents is indicative of impact on potassium channel functionality. Loss of potassium channel functionality results in the possibility that this channel subclass contributes to glioma metastasis through motility mechanisms. In addition to Kv1.1 interaction with MMP-2, other cellular proteins that may interact with potassium channels include sodium channels, calcium channels, and chloride channels. These responses to DTX-1 exposure enable the peptide to be utilized as a high affinity ligand for QD studies of the intricate interplay between potassium channels and other cellular targets. QD labeling utilizing high affinity peptide toxins has been demonstrated to provide better sensitivity for probing biological systems in their native environment.6 Multiplexing experiments using two high affinity peptide toxins, CTX and DTX-1, conjugated to QDs are presented. These two toxins were chosen based upon their potential application toward analyses of the metastasis motility hypothesis.22-24,26,32,33 For example, an interesting property of CTX is that it is believed to severely reduce cell motility, if not stop it entirely. When taken in context with a glioma, which undergoes metastasis through motility mechanisms unlike other cancers, reduction in cell motility results in loss of metastasis. Additionally, DTX-1 targets Kv1.1, a member of the Shaker family of potassium channels.34 Ion channels are implicated in a variety of diseases. This makes them interesting targets to examine independently and with other cellular proteins for in depth analyses. Probe development utilizing fluorescent nanocrystals with high affinity ligands, such as the peptide toxins chosen, may provide the capacity to monitor these cellular targets for greater time periods and in conditions unfavorable for common organic dyes. Each toxin was conjugated directly to QDs and used to visualize endogenous expression of MMP-2 and Kv1.1 in C6 glioma cells, along with expression of MMP-2 and Kv1.1 in a series of representative cell types. QD conjugates for CTX and DTX-1 were synthesized according to the reaction scheme in Figure 1A-C.33 The toxins were cross-linked to both QD525 and QD655 for CTX and DTX-1, respectively, to produce CTX:QD525 and DTX1:QD655 for labeling. QD conjugate detection of endogenous cellular targets was determined through antibody colocalization analyses. Live C6 glioma cells were incubated with either 10nM CTX:QD525 or QD525 and either 5nM DTX1:QD655 or QD655 at 37 °C and 5% CO2 for 120 and 180 min, respectively. This was to provide maximum binding saturation prior to antibody detection. Following incubation cells were postfixed in 4% paraformaldehyde and permeabilized in 0.1% Triton X-100 prior to primary antibody exposure and subsequent secondary Alexafluor conjugated antibody labeling. Figure 2A,B illustrates the colocalization of CTX:QD525 conjugates (Figure 2A) with anti-MMP-2 antibodies (Sigma, Inc.) detected with secondary-Alexa594 Nano Lett., Vol. 9, No. 7, 2009
Figure 1. Toxin conjugation to QDs reaction scheme and gel electrophoresis characterization. ITK amino(PEG) QD525s and QD655s were reacted with SIA (A), and cross-linked with CTX and DTX-1 (C), which were previously reacted with Traut’s reagent (B). Conjugates were isolated using centrifugation through a 7000 MWCO Zeba desalting column, followed by concentration with a Millipore Ultrafree-0.5 Biomax column (C). Once isolated, conjugation was confirmed via a 1% agarose gel run at continuous current of 80 V for 120 min (D). The gel was imaged using Bio-Rad Chemidox XRS System with Quantity One software. Both CTX:QD525 and DTX-1:QD655 migrations are greater than QD525 and QD655 migrations, respectively.
antibodies versus exposure to QD525 in lieu of the toxin conjugate (Figure 2B). The presence of colocalization between CTX:QD525 and secondary-Alexa594 antibodies (Invitrogen, Inc.) indicates that CTX:QD525 has labeled MMP-2 in the live cell and demonstrates probe suitability for the specified target. DTX-1:QD655 and secondaryAlexafluor antibody colocalization (Figure 2C) is similarly observed with fluorescence colocalization achieved using an anti-Kv1.1 (Sigma, Inc.) antibody and secondary-Alexa488 (Invitrogen, Inc.) antibody system and compared with exposure to QD655s only (Figure 2D). Comparable to CTX: QD525, secondary-Alexa488 fluorescence colocalization with DTX-1:QD655 emission is observed (Figure 2C) without the presence of QD655 nonspecific binding (Figure 2D). This supports DTX-1:QD655 specificity for the target, Kv1.1, making the conjugate a suitable probe for endogenous potassium channel detection in living cells. The fluorescence patterning of the colocalization reveals that internalization of the toxin:QD conjugates and their respective targets may Nano Lett., Vol. 9, No. 7, 2009
have occurred over time; however, the majority of toxin: QD conjugates colocalized with their cellular targets appear to be membrane bound. Colocalization of secondary-Alexafluor fluorescence with CTX:QD525 and DTX-1:QD655 emission enabled labeling efficacy to be determined as specific to the cellular targets, MMP-2 and Kv1.1. Additionally, lack of secondary-Alexafluor colocalization when present in the system exposed only to QD525 or QD655 further substantiates probe specificity via CTX:QD525 and DTX-1:QD655 affinity for the targets, MMP-2 and Kv1.1, and supports that fluorescence is not the result of QD525 or QD655 nonspecific binding within C6 glioma cells. It is also significant to note that detection of endogenous targets was achieved using QD conjugates at concentrations approximately ten times less than reported for Alexafluor labeling in fixed cell culture. Following examination utilizing antibody colocalization analyses for positive determination of target detection, verification that CTX:QD525 binds to the specified target 2591
Figure 2. Antibody colocalization of endogenous MMP-2 and Kv1.1 expression in live C6 glioma cells. CTX:QD525 (A) and QD525 (B) were exposed to live C6 glioma cells for 2 h at 37 °C at 5% CO2. DTX-1:QD655 (C) and QD655 (D) were exposed to live C6 glioma cells for 3 h at 37 °C at 5% CO2. Cells were postfixed in 4% paraformaldehyde and permeabilized in 0.1% TritonX-100. CTX:QD525 and QD525 exposed cells were incubated with MMP-2 (1:500) primary antibody in 1% NGS for 20 h at 4 °C and then exposed to Alexa594 secondary antibody (1:500) for 1 h at room temperature. DTX-1:QD655 and QD655 exposed cells were incubated with Kv1.1 (1:500) primary antibody in 1% NGS for 24 h at 4 °C and then exposed to Alexa488 secondary antibody (1:500) for 1 h at room temperature. Imaging was performed on an LSM510 Meta inverted confocal microscope following mounting with Aqua Polymount. CTX:QD525 labels endogenous MMP-2 expression in live cell culture, indicating effectiveness as a probe for live detection studies. Kv1.1 expression is effectively detected using DTX-1:QD655 in live cell culture, which indicates probe effectiveness for live detection studies of endogenously expressed targets. Images represent z-stack confocal microscopy acquisition and are at the center of the z-stack.
was further tested through time-dependent blocking assays. C6 glioma cells were incubated in media containing 200 nM unconjugated CTX overnight. The cultures were then treated with 10 nM CTX:QD525 or QD525 for time intervals of 30, 60, and 120 min at 37 °C and 5% CO2. Figure 3 illustrates representative images of effective CTX blocking of CTX: 2592
QD525 labeling for the designated time intervals. Comparatively, CTX:QD525 detection of MMP-2 increases with time in C6 glioma cultures left unexposed to unconjugated CTX, which further substantiates CTX:QD525 probe specificity for the target. There is a significant increase at 120 min in CTX: QD525 labeling of native MMP-2 in live culture, which was Nano Lett., Vol. 9, No. 7, 2009
Figure 3. CTX preincubation blocks CTX:QD525 detection in live culture for up to 120 min. C6 glioma cells were exposed to 200 nM CTX in complete media overnight at 37 °C at 5% CO2 prior to exposure to 10 nM CTX:QD525 (B1-B3) and 10 nM QD525 (B4), and 10 nM CTX:QD525 only (A1-A3) and 10 nM QD525 only (A4) for time intervals of 30, 60, and 120 min. CTX effectively blocked CTX:QD525 endogenous detection in live culture for up to 120 min. (C) (P < 0.001, analysis of variance, values are the means ( SEM) and further indicates probe suitability for live assays. Each time interval was performed in triplicate with at least six images acquired per replicate. Additionally, flow cytometry (D) supports CTX:QD525 detection of MMP-2 in C6 glioma cells. The percentage of cells labeled with CTX:QD525 only within the cell population is statistically significant (P < 0.001, analysis of variance, values are the means ( SEM), as compared with cells incubated with unconjugated CTX prior to CTX:QD525 exposure at 120 min. Each time interval was performed in triplicate.
established to be the optimal incubation time for the CTX: QD525 conjugate for additional studies utilizing the probe. Incubation times greater than 120 min appeared to result in cell death, possibly due to the toxin’s lethal effects. Following exposure to CTX overnight, the C6 glioma cells appeared to be less adherent to the dish surface and had fewer and shorter projections comparatively. At 120 min, CTX blocked MMP-2 remains undetected by CTX:QD525. Quantitative analysis that CTX:QD525 detection in live culture is effectively blocked for 120 min in addition to antibody Nano Lett., Vol. 9, No. 7, 2009
colocalization (Figure 2A) confirms CTX:QD525 target specificity and concurs with qualitative microscopic evidence. Flow cytometry experimentation (Figure 3D, Supporting Information Figure S1) further supports CTX:QD525 specificity by providing quantitative evidence reflecting the number of C6 glioma cells labeled at 120 min, relative to cells exposed to unconjugated CTX prior to CTX:QD525 incubation. The CTX:QD525 labeled cells were present in approximately 80% of the culture, whereas unconjugated CTX incubation resulted in cell labeling levels similar to 2593
the cells exposed only to QD525 (P < 0.001, analysis of variance, values are the means ( SEM). This methodology of unconjugated toxin blocking was repeated using DTX-1:QD655 to ascertain endogenous specificity for Kv1.1 in C6 glioma culture. Cells were treated similarly to the CTX assays and blocked with 100 nM unconjugated DTX-1 overnight. Following blocking with unconjugated DTX-1, C6 glioma cells were incubated with 5nM DTX-1:QD655 for time intervals of 10, 30, 60, 120, and 180 min. Subsequent comparison to native Kv1.1, illustrated in Figure 4, reveals the rapidity of DTX-1:QD655 labeling of native Kv1.1 in live culture. Both quantitative and qualitative analyses reveal that DTX-1 blocks DTX-1: QD655 detection of Kv1.1 for approximately 180 min. At approximately 180 min, DTX-1:QD655 fluorescence in cells not exposed to unconjugated DTX-1 is significantly greater and more widely distributed throughout the culture than at other time points; however, at both 60 and 120 min, labeling is significant for adequate detection. The lack of DTX-1: QD655 labeling in cells incubated with unconjugated DTX-1 denotes that the toxin effectively blocks DTX-1:QD655 detection of Kv1.1 at all time intervals. This is further confirmed by quantitative analysis in Figure 4C. As with CTX exposure, DTX-1 exposure resulted in qualitative morphological changes that may affect cell attachment and overall cellular appearance. DTX-1:QD655 detection of Kv1.1 is also supported by flow cytometry experimentation (Figure 4D, Supporting Information Figure S2). As with CTX:QD525, DTX-1:QD655 labeled cells were present at statistically significant numbers within culture (P < 0.001, analysis of variance, values are the means ( SEM) versus the number of labeled cells previously exposed to unconjugated DTX-1. At 180 min, the number of DTX-1:QD655 labeled cells in culture was approximately
