Resolving Low-Expression Cell Surface Antigens by Time-Gated

Oct 25, 2012 - (7) The revolution that CD4+T cell quantification provided for HIV monitoring in the 1980s is a classic example. ... provides sharp sig...
0 downloads 13 Views 3MB Size
Letter pubs.acs.org/ac

Resolving Low-Expression Cell Surface Antigens by Time-Gated Orthogonal Scanning Automated Microscopy Jie Lu,† Jody Martin,‡ Yiqing Lu,† Jiangbo Zhao,† Jingli Yuan,§ Martin Ostrowski,⊥ Ian Paulsen,⊥ James A. Piper,† and Dayong Jin*,† †

Advanced Cytometry Laboratories, MQ Biofocus Research Centre, Faculty of Science, Macquarie University, NSW 2109, Sydney, Australia ‡ BD Bioscience, I11077 North Torrey Pines Road, La Jolla, California 92037, United States § State Key Laboratory of Fine Chemicals, School of Chemistry, Dalian University of Technology, Dalian, 116024, P. R. China ⊥ Biomolecular Frontiers Research Centre, Faculty of Science, Macquarie University, NSW 2109, Sydney, Australia S Supporting Information *

ABSTRACT: We report a highly sensitive method for rapid identification and quantification of rare-event cells carrying low-abundance surface biomarkers. The method applies lanthanide bioprobes and time-gated detection to effectively eliminate both nontarget organisms and background noise and utilizes the europium containing nanoparticles to further amplify the signal strength by a factor of ∼20. Of interest is that these nanoparticles did not correspondingly enhance the intensity of nonspecific binding. Thus, the dramatically improved signal-to-background ratio enables the low-expression surface antigens on single cells to be quantified. Furthermore, we applied an orthogonal scanning automated microscopy (OSAM) technique to rapidly process a large population of targetonly cells on microscopy slides, leading to quantitative statistical data with high certainty. Thus, the techniques together resolved nearly all false-negative events from the interfering crowd including many false-positive events. which produce a fluctuating baseline, making it impractical to read absolute signal levels from individual single cells. As a solution to the autofluorescence problem, time-gated luminescence bioassays have been developed using luminescent lanthanide bioprobes (mainly Eu3+ and Tb3+) which display long-lived luminescent emissions.15−20 Short-lived autofluorescence backgrounds from raw biological samples or scattering from nearby optics can be effectively eliminated to provide highly sensitive detections of analytes, achieving an improved detection limit by a factor of more than 2 orders of magnitude. There are now available commercial spectroscopy systems for microsecond time-resolved luminescence measurements, such as DELFIA (dissociation enhanced lanthanide fluoroimmunoassay) system,21 CyberFluor (or FIAgen) system,22 enzymeamplified time-resolved fluorometric system,23 and TRACE (time-resolved amplified cryptate emission) system.24 Coupling of high throughput methods, such as cytometry, with the enhanced sensitivity of time-gated fluorescence approaches may provide a solution to detect low-expression surface antigens on rare-event cells. In this report, we applied a novel combination of techniques to maximize the signal-to-background ratio as

D

etection of disease biomarkers at the single cell level holds great promise for early stage diagnostics.1 However, sensitivity is a critical issue since only trace amounts of biomolecules carried by rare-event cells need to be identified and analyzed.2,3 This is particularly true when the target biomarkers are expressed in very low abundance in the early stages of disease development or in post-therapy monitoring.4,5 This situation truly poses a “needle-in-a-haystack” problem.6 Cell surface biomarkers are well characterized in the field of immunology, with hundreds of in vitro diagnostic tests covering major fields of medicine.7 The revolution that CD4+ T cell quantification provided for HIV monitoring in the 1980s is a classic example. In recent years, surface biomarkers have been recognized as important indicators in cancer diagnosis.8 For example, the prostate-specific membrane antigen was found to be elevated in malignant prostatic epithelium,9 and mesothelin, a 40-kD cell-surface glycoprotein, is overexpressed in tissues of epithelial ovarian cancer.10 Flow cytometry, capable of analyzing single cells in a high throughput fashion, is typically used for detection of cell surface antigens.11,12 However, it remains challenging to discriminate weak signal fluorescence from background autofluorescence when surface antigens are expressed in low abundance. Background autofluorescence is mainly generated by instrument fluidics, optical scattering, and cellular molecules,13,14 © 2012 American Chemical Society

Received: September 7, 2012 Accepted: October 25, 2012 Published: October 25, 2012 9674

dx.doi.org/10.1021/ac302550u | Anal. Chem. 2012, 84, 9674−9678

Analytical Chemistry

Letter

Figure 1. Comparison of conventional UV (epi)fluorescence microscopy imaging (left) and time-gated luminescence microscopy imaging (right) of medium (upper images; the green cell was labeled by SA-Alexa 488 and red cells were labeled by SA-BHHCT-Eu3+) and low (lower images; cells were labeled by SA-BHHCT-Eu3+) expression CD34+ cells (Exposure time: 1 s). Background suppression from 42.177 to 8.323 was measured by Image J (see Table S1, Supporting Information).The intensity difference was displayed using ImageJ 3-D interactive intensity analysis. Scale bar = 10 μm.

Information). In contrast to this result with conventional epifluorescence microscopy, time-gated luminescence (TGL)25 imaging of CD34+ cells stained with biotinylated anti-CD34 antibodies provides sharp signal-to-background ratio through effective suppression of the major sources of autofluorescence background (Figure 1, Figure S3, and Table S1, Supporting Information). TGL not only makes nonspecific targets (SAAlexa 488 labeled cells as a model) invisible but also provides an opportunity for absolute quantification of fluorescence signals in background-free conditions. This is especially true when signals were extremely weak as in our example of low CD34+ cells. In order to resolve the population of low-CD34+ cells, we amplified the europium signal strength by employing europium-complex-rich nanoparticles. Both Zeta potential26,27 and gel electrophoresis,28 shown in Figure 2 A, were used to monitor the conjugation process on nanoparticles from a highly negatively (−40 mV) charged carboxyl surface to a slightly positively (+20 mV) charged SA surface in a pH 6.0 2-(Nmorpholino)ethanesulfonic acid (MES) solution (Figure S4, Supporting Information). The reduced surface charge after SA bioconjugation explains the reason why we observed the particles to be likely to aggregate. The SEM images (Figure 2C,D and Figures S5 and S6B, Supporting Information) also confirm the actual distribution of nanoparticles on the lowexpression CD34+ cell surface. Previous reports have indicated that a range of hundreds to thousands of dye molecules could be entrapped onto one nanoparticle biolabel to significantly amplify the signal strength.29−31 However, there are concerns

well as a novel cytometry platform capable of resolving lowabundance cell surface antigens. The detailed methods for sample preparations were listed as the experimental section in the Supporting Information. In order to evaluate our technique, we engineered HEK293 cells to express medium and low levels of CD34 antigen (herein referred to as medium CD34 + and low CD34+ cells, respectively) by controlling the level of its transient transfection. We demonstrated the CD34 expression levels using flow cytometry with conventional labeling of biotinylated antiCD34 antibody and streptavidin−phycoerythrin (SA-PE) dyes (Figure S1, Supporting Information). While the majority of medium-level CD34+ cells are clearly distinguishable from the negative control samples (HEK293 cells alone and SA-PE secondary staining without the primary antibody), the majority of PE stained low-CD34+ cells (>80%) display fluorescence levels that overlapped with the autofluorescence of the control populations (Figure S1, Supporting Information). In order to demonstrate the effective background suppression of time-gated luminescence compared with conventional UV (epi)fluorescence, both populations of CD34+ cells were stained with a streptavidin−europium complex conjugate and streptavidin−Eu3+-4,4′-bis(1″,1″,1″,2″,2″,3″,3″-heptafluoro4″,6″-hexanedione-6″-yl)-chlorosulfo-o-terphenyl (SABHHCT-Eu3+) complex and imaged under a true-color timegated luminescence microscope (excitation: pulsed ultraviolet light emitting diode). The signal strengths of the medium CD34+ population and partial low CD34+ population were sufficient to be visible to the naked-eye (Figure S2, Supporting 9675

dx.doi.org/10.1021/ac302550u | Anal. Chem. 2012, 84, 9674−9678

Analytical Chemistry

Letter

Figure 2. (A) Zeta potential (buffer: 50 mM MES at pH 6.0) and agarose gel electrophoresis shift before and after streptavidin conjugation to the 40 and 200 nm nanoparticles (buffer: 0.5× TBE at pH 8.0 for 30 min in 50 V, agarose concentration (40 nm: 0.75%, 200 nm: 0.25%)). (B) Luminescence intensity amplifications by 40 and 200 nm of europium containing polystyrene nanoparticles compared to BHHCT-Eu3+ complex labeling; this result was obtained with the medium-abundance CD34+ cells. Exposure times were indicated in images. Scale bar = 10 μm. (C and D) Scanning electron microscopy images of SA-200 nm Eu3+ nanoparticles labeled on low-abundance CD34+ cells. Green arrows show the typical nanoparticles conjugated onto the cell surface (also see Figure S6 B, Supporting Information).

Figure 3. Time-gated luminescence scanning cytometry histograms of resolving the low CD34+ cells from nonspecific control cells (blue bars represent stain negative control and red bars are CD34 positive population) detected by: left, SA-BHHCT-Eu3+ labeling (inset as magnified view); right, SA-200 nm nanoparticles labeling.

two different diameters (40 and 200 nm) were tested on the medium-CD34+ cells (Figure 2) and achieved 3.1- and 20-fold enhancement, respectively, in the luminescence signal, over the BHHCT-Eu3+ complex (shown in Figure 2B). This result is consistent with the results of a previously reported immunoassay that used 107 nm europium nanoparticles (containing over 30 000 dye molecules) and achieved a 70-fold improve-

that the relatively large size of these nanoparticles (>40 nm) may occupy too much space on the cell surface and that low numbers of capture molecules may not be able to anchor the antibody-conjugated nanoparticles, so that signal amplification of this magnitude may not be feasible for actual cell surface labeling. This work provides a direct quantitative comparison for nanoparticle signal amplification. Here, the nanoparticles of 9676

dx.doi.org/10.1021/ac302550u | Anal. Chem. 2012, 84, 9674−9678

Analytical Chemistry

Letter

ment in detection limit.29−31 The reason we utilized the medium-CD34+ population here was due to some concerns over the large variation of low-expression cells and the lowsignal strength from BHHCT-Eu3+ labeling. In order to conduct population studies to provide statistical cytometry data, we used an automated system of time-gated orthogonal scanning automated microscopy (OSAM) to process slides of labeled cells (3 min per 15 mm by 15 mm slide).32 Shown in Figures S7, S8, and S9, Supporting Information, briefly, since the antiphase sequence of pulsed excitation and time-delayed detection is employed, a singleelement photomultiplier tube (PMT) detector only recognizes the long-lived europium luminescence, so that the points of interest can be rapidly spotted by wide-field microscopy optics.33 The system employed a 2-step orthogonal scanning strategy to accurately bring the target cells in the middle of wide field, so that the maximum intensities of target cells could be recorded with optimal coefficient of variation (CV) performance. Under such a system, autofluorescence was suppressed for unstained HEK293 cells, and only the negative stain control transfected cells (without anti-CD34 but stained with SA-BHHCT-Eu3+ or SA-200 nm Eu3+ nanoparticles) displayed low-level luminescence signals due to nonspecific binding backgrounds, shown in Figure 3 (blue bars). The weak luminescence signal detected by SA-BHHCT-Eu3+ could only resolve 28.4% of the low-CD34+ cells out of the nonspecific binding region (left), which was a marginal increase over SA-PE flow cytometry detection (19.4%). In contrast, the labeling of SA-200 nm europium-containing nanoparticles (right) successfully resolved the population of low-expression CD34 cells (98.6%). We also demonstrated an additional advantage of using the nanoparticles since the signal amplification of the CD34 positive cells occurred without a corresponding increase in the signal from nonspecific binding to the CD34 negative population, thereby greatly increasing the signal-to-background ratio. A possible explanation for this result might be that the bulkier nanoparticles destabilize the weaker nonspecific low affinity interactions of the streptavidin−dye conjugate. Moreover, conventional time-gated luminescence techniques using lanthanide bioprobes have been constrained by the fact that lanthanide emission has very long lifetimes (hundreds of microseconds) and much weaker signal than conventional probes (e.g., FITC, PE dyes, lifetimes of less than 5 ns) so that a much longer exposure time is needed to collect detectable signal. Our demonstration shows the feasibility of on-the-fly detection which only requires a few hundred photoelectrons within 100 μs time-delayed detection phase to distinguish target events. The cells with nonspecific binding were clearly detectable (each cell carrying only ∼ tens of 200 nm particles, as estimated from SEM images). This ultrasensitivity is due to the negligible background level in the time-delayed window and our scanning system using a single-element photomultiplier tube detector. This suggests the signal strength was not a limiting factor of lanthanide bioprobes used with the highspeed cell analysis. The use of wide-field microscopy optics (105 μm in diameter) significantly increased scanning speed but also can increase the chance for doublet (two cells in close proximity) or triplet (three cells in close proximity) events to be detected, which can then distort intensity variations. This problem was overcome by retrieving functions to visually inspect the points of interest after rapid scanning (scanning time ∼3 min per slide). Figure S10 A (Supporting Information) illustrates the

mapping of points of interest and shows an improvement of scanning CV from 55.49% (Figure S10B, Supporting Information) to 31.02% by postretrieving (Figure S10C, Supporting Information). With a very small data volume, the computerized mapping of target cells on slide can provide even more detailed information for cell function, tracking, and analysis in the post-scanning stage. Thus, the retrieving function is a powerful tool to optimize data accuracy for single cell population analysis. In conclusion, we demonstrated a combination of practical approaches to resolve populations of low-expression CD34 cells, which otherwise were not distinguishable in conventional flow cytometry as a result of the autofluorescence backgrounds and signal fluctuations from one cell to another. We applied time-gated luminescence detection to suppress autofluorescence and scattering noise and employed functionalized polystyrene nanoparticles to amply the signal strength (up to 20 times), which together achieved a high signal-to-background and successfully resolved those CD34 cells with low expression. This technique is compatible with our recent development of an orthogonal scanning automated microscopy (OSAM) system,33 and resulted in robust statistical data showing separated target cell population (98%) from stained control cells with an optimized CV of 31%. This indicates that the new time-gated OSAM not only can selectively scan and spot the rare-event target cells but also has enabled the absolute intensity reading and quantitative analysis of cell surface molecules. As the time-gated luminescence technique has already shown its power in achieving much better detection limits in bioassays of molecular suspensions, this work demonstrates the potential for time-gated OSAM, a scanning cytometry, to break through the current detection limits for analysis of rare-event cells expressing antigens at extremely low levels and be further used as an ultrasensitive tool for early stage biomarker diagnosis.



ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors wish to acknowledge Mattias Roth for his assistance in the gel electrophoresis experiments, Australian Research Council (Discovery Project DP 1095465), ISAC Scholar program, and International Macquarie University Research Excellence Scholarship. J.Y. acknowledges the financial support from the National Natural Science Foundation of China (No. 20835001).



REFERENCES

(1) Zhang, Z.; Bast, R. C., Jr.; Yu, Y.; Li, J.; Sokoll, L. J.; Rai, A. J.; Rosenzweig, J. M.; Cameron, B.; Wang, Y. Y.; Meng, X. Y.; Berchuck, A.; Van Haaften-Day, C.; Hacker, N. F.; de Bruijn, H. W.; van der Zee, A. G.; Jacobs, I. J.; Fung, E. T.; Chan, D. W. Cancer Res. 2004, 64, 5882.

9677

dx.doi.org/10.1021/ac302550u | Anal. Chem. 2012, 84, 9674−9678

Analytical Chemistry

Letter

(2) Mor, G.; Visintin, I.; Lai, Y.; Zhao, H.; Schwartz, P.; Rutherford, T.; Yue, L.; Bray-Ward, P.; Ward, D. C. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 7677. (3) Joensson, H. N.; Samuels, M. L.; Brouzes, E. R.; Medkova, M.; Uhlen, M.; Link, D. R.; Andersson-Svahn, H. Angew. Chem., Int. Ed. Engl. 2009, 48, 2518. (4) Hubert, R. S.; Vivanco, I.; Chen, E.; Rastegar, S.; Leong, K.; Mitchell, S. C.; Madraswala, R.; Zhou, Y.; Kuo, J.; Raitano, A. B.; Jakobovits, A.; Saffran, D. C.; Afar, D. E. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 14523. (5) van Dongen, J. J.; Krissansen, G. W.; Wolvers-Tettero, I. L.; Comans-Bitter, W. M.; Adriaansen, H. J.; Hooijkaas, H.; van Wering, E. R.; Terhorst, C. Blood 1988, 71, 603. (6) Abercrombie, M.; Ambrose, E. J. Cancer Res. 1962, 22, 525. (7) Knapp, W.; Strobl, H.; Majdic, O. Cytometry 1994, 18, 187. (8) Mellors, J. W.; Munoz, A.; Giorgi, J. V.; Margolick, J. B.; Tassoni, C. J.; Gupta, P.; Kingsley, L. A.; Todd, J. A.; Saah, A. J.; Detels, R.; Phair, J. P.; Rinaldo, C. R., Jr. Ann. Intern. Med. 1997, 126, 946. (9) Silver, D. A.; Pellicer, I.; Fair, W. R.; Heston, W. D.; CordonCardo, C. Clin. Cancer Res. 1997, 3, 81. (10) Hassan, R.; Viner, J. L.; Wang, Q. C.; Margulies, I.; Kreitman, R. J.; Pastan, I. J. Immunother. 2000, 23, 473. (11) Morkve, O.; Halvorsen, O. J.; Stangeland, L.; Gulsvik, A.; Laerum, O. D. Int. J. Cancer 1992, 52, 851. (12) Ortaldo, J. R.; Sharrow, S. O.; Timonen, T.; Herberman, R. B. J. Immunol. 1981, 127, 2401. (13) Benson, R. C.; Meyer, R. A.; Zaruba, M. E.; McKhann, G. M. J. Histochem. Cytochem. 1979, 27, 44. (14) Andersson; Baechi; Hoechl; Richter. J. Microsc. 1998, 191, 1. (15) Hemmilá, I.; Mukkala, V.-M. Crit. Rev. Clin. Lab. Sci. 2001, 38, 441. (16) Yuan, J.; Wang, G. TrAC, Trends Anal. Chem. 2006, 25, 490. (17) Moore, E. G.; Samuel, A. P.; Raymond, K. N. Acc. Chem. Res. 2009, 42, 542. (18) Bunzli, J. C. Chem. Rev. 2010, 110, 2729. (19) Hagan, A. K.; Zuchner, T. Anal. Bioanal. Chem. 2011, 400, 2847. (20) Gahlaut, N.; Miller, L. W. Cytometry, Part A 2010, 77, 1113. (21) Ogata, A.; Tagoh, H.; Lee, T.; Kuritani, T.; Takahara, Y.; Shimamura, T.; Ikegami, H.; Kurimoto, M.; Yoshizaki, K.; Kishimoto, T. J. Immunol. Methods 1992, 148, 15. (22) Morton, R. C.; Diamandis, E. P. Anal. Chem. 1990, 62, 1841. (23) Templeton, E. F.; Wong, H. E.; Evangelista, R. A.; Granger, T.; Pollak, A. Clin. Chem. 1991, 37, 1506. (24) Bazin, H.; Trinquet, E.; Mathis, G. J. Biotechnol. 2002, 82, 233. (25) Jin, D.; Piper, J. A. Anal. Chem. 2011, 83, 2294. (26) Lockman, P. R.; Koziara, J. M.; Mumper, R. J.; Allen, D. D. J. Drug Targeting 2004, 12, 635. (27) Zhong, W. Anal. Bioanal. Chem. 2009, 394, 47. (28) Wang, S.; Mamedova, N.; Kotov, N. A.; Chen, W.; Studer, J. Nano Lett. 2002, 2, 817. (29) Ow, H.; Larson, D. R.; Srivastava, M.; Baird, B. A.; Webb, W. W.; Wiesner, U. Nano Lett. 2005, 5, 113. (30) Burns, A.; Ow, H.; Wiesner, U. Chem. Soc. Rev. 2006, 35, 1028. (31) Zhao, X.; Hilliard, L. R.; Mechery, S. J.; Wang, Y.; Bagwe, R. P.; Jin, S.; Tan, W. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 15027. (32) Lu, Y.; Xi, P.; Piper, J. A.; Huo, Y.; Jin, D. Sci. Rep. 2012, DOI: 10.1038/srep00837. (33) Lu, Y.; Jin, D.; Leif, R. C.; Deng, W.; Piper, J. A.; Yuan, J.; Duan, Y.; Huo, Y. Cytometry, Part A 2011, 79, 349.

9678

dx.doi.org/10.1021/ac302550u | Anal. Chem. 2012, 84, 9674−9678