Prospects of Deep Raman Spectroscopy for ... - ACS Publications

Apr 16, 2010 - Nicholas Stone,* Karen Faulds,† Duncan Graham,† and Pavel Matousek‡. Biophotonics Research Unit, Gloucestershire Hospitals NHS ...
0 downloads 0 Views 2MB Size
Anal. Chem. 2010, 82, 3969–3973

Letters to Analytical Chemistry Prospects of Deep Raman Spectroscopy for Noninvasive Detection of Conjugated Surface Enhanced Resonance Raman Scattering Nanoparticles Buried within 25 mm of Mammalian Tissue Nicholas Stone,* Karen Faulds,† Duncan Graham,† and Pavel Matousek‡ Biophotonics Research Unit, Gloucestershire Hospitals NHS Foundation Trust, Great Western Road, Gloucester, GL1 3NN, U.K., Centre for Molecular Nanometrology, WestCHEM, Department of Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow, G1 1XL, U.K., and Central Laser Facility, Science and Technology Facilities Council, Rutherford Appleton Laboratory, Oxfordshire, OX11 0QX, U.K. This letter discusses the potential of deep Raman spectroscopy, surface enhanced spatially offset Raman spectroscopy (SESORS and its variants), for noninvasively detecting small, deeply buried lesions using surface enhanced resonance Raman scattering (SERRS) active nanoparticles. An experimental demonstration of this concept is performed in transmission Raman geometry. This method opens prospects for in vivo, noninvasive, specific detection of molecular changes associated with disease up to depths of several centimeters representing significant improvement over traditionally detected Raman signals by 2 orders of magnitude. The disease specific signals can be achieved using uniquely tagged nanoparticles conjugated to target molecules, e.g., antibodies for production of the SERRS signal. This provides the molecular specific signal which is many orders of magnitude greater than normal biological Raman signals and can be easily multiplexed. To date, there have been no studies demonstrating the viability of deep Raman spectroscopy coupled to surface enhanced techniques for detecting low concentrations of molecules of interest at depths of greater than 5.5 mm in tissue. Such a breakthrough would open a host of new applications in medical diagnoses. Here we propose to facilitate such capability by combining SERRS (as a probe for disease specific changes) with deep Raman spectroscopy techniques. This permits noninvasive measurement of Raman signatures from conjugated SERRS nanoparticles at clinically relevant concentrations through tissues of between 15 and 25 mm thick. * To whom correspondence should be addressed. E-mail: n.stone@ medical-research-centre.com. † University of Strathclyde. ‡ Rutherford Appleton Laboratory. 10.1021/ac100039c  2010 American Chemical Society Published on Web 04/16/2010

Recently developed deep Raman techniques have demonstrated the detection of Raman signals from calcifications1 and cancerous tissues2 deeply buried within areas of soft tissue but at much higher concentrations than those required for a SERRS signal to be detected. This study demonstrates that conjugated SERRS nanoparticles located within tissues at depths of several centimeters can be detected by this method at extremely low, clinically relevant concentration levels. The detection limit of the technique for the configuration used in this demonstration will be outlined. There are numerous clinical needs where the application of an in vivo molecule-specific test would significantly aid an accurate, early, and rapid diagnosis. These include early cancer detection and staging, treatment monitoring, and chemo-sensitivity which may potentially be aided in this way. There have been a number of methodologies proposed to achieve this goal in tandem with a specific molecular- based targeting strategy. These include functionalized quantum dots and fluorescently labeled antibodies in addition to surface enhanced Raman probes. These fluorescence/luminescence approaches can provide strong signals, although they can be difficult to multiplex in a clinical environment. A common problem to the majority of these is the unavailability of an effective method capable of reading these signals noninvasively, safely and rapidly from deep areas of soft tissue. Surface enhanced Raman scattering (SERS) can provide molecular specific enhancement of Raman signals3,4 by bringing the target molecule into close proximity with a roughened (nanometer scale) noble metal surface. Huge enhancement factors (1) Stone, N.; Matousek, P. Cancer Res. 2008, 68, 4424–4430. (2) Keller, M. D.; Majumder, S. K.; Mahadevan-Jansen, A. Opt. Lett. 2009, 34, 7. (3) Fleischmann, M.; Hendra, P. J.; McQuillan, A. J. Chem. Phys. Lett. 1974, 26, 163–166. (4) Jeanmaire, D. L.; Van Duyne, R. P. J. Electroanal. Chem. 1977, 84 (2), 1–20.

Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

3969

of the order of 109 are possible, and single molecule detection has been reported.5 However, it has proved difficult with SERS to achieve reliable and reproducible results, a key requirement for clinical use; major improvements have been made with recent developments of novel substrates, such as encapsulated nanoparticles, promising to overcome some of these difficulties.6 These nanoparticles can be easily tagged with antibodies to enable molecular specific detection of disease. Furthermore, reproducibility can be achieved with the use of photonic crystals which by careful manufacture can provide reliable substrates that can be tuned to specific resonance with excitation wavelengths.7,8 Further developments have included the use of SERRS, a resonance SERS technique pioneered by Stacy and Van Duyne9 and developed for use as a clinical tool by Graham et al.,10 which is able to provide equivalent detection limits to fluorescence labeled dyes.11 In the cancer environment, tagged nanoparticles enhancing specific signals from malignant markers are either being used in vivo12-14 (safety issues to be resolved for human use) or as molecular specific stains for histopathology;15-17 with the possibility of numerous multiplexed SERS/SERRS stains providing hyperspectral images of locations of molecules of interest from the same spectral acquisition and tissue slice.18 An in vivo study demonstrating the use of multiplexed tags within a mouse model was able to demonstrate the detection of 1.6 × 107 nanoparticles subcutaneously and 3.1 × 107 nanoparticles following intravenous injection.13 Furthermore, this study demonstrated the maximum penetration depth achieved to date with SERS and traditional Raman microscopy was of the order of 5.5 mm with a bolus of 1.6 × 1011 nanoparticles. Recent developments have led to the demonstration of Raman signal collection from orders of magnitude deeper than previously thought possible. These include novel developments of spatially (5) Kneipp, K.; Wang, Y.; Kneipp, H.; Perelman, L. T.; Itzkan, I.; Dasari, R. R.; Feld, M. S. Phys. Rev. Lett. 1997, 78, 1667. (6) Doering, W. E.; Piotti, M. E.; Natan, M. J.; Freeman, R. G. Adv. Mater. 2007, 19, 3100–3108. (7) Cintra, S.; Abdelsalam, M. E.; Bartlett, P. N.; Baumberg, J. J.; Kelf, T. A.; Sugawara, Y.; Russell, A. E. Faraday Discuss. 2006, 132, 191–199. (8) Stacy, A. A.; Van Duyne, R. P. Phys. Chem. Chem. Phys. 2007, 9 (1), 104– 109. (9) Stacy, A. A.; Van Duyne, R. P. Chem. Phys. Lett. 1983, 102 (4), 365–370. (10) Graham, D.; Smith, W. E.; Linacre, A. M. T.; Munro, C. H.; Watson, N. D.; White, P. C. Anal. Chem., 1997, 69 (22), 4703–4707. (11) Sabatte, G.; Keir, R.; Lawlor, M.; Black, M.; Graham, D.; Smith, W. E. Anal. Chem. 2008, 80 (7), 2351–2356. (12) Pal, A.; Isola, N. R.; Alarie, J. P.; Stokes, D. L.; Vo-Dinh, T. Faraday Discuss. 2006, 132, 293–301. (13) Keren, S.; Zavaleta, C.; Cheng, Z.; De La Zerda, A.; Gheysens, O.; Gambhir, S. S. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 5844–5849. (14) Qian, X.; Peng, X.-H.; Ansari, D. O.; Yin-Goen, Q.; Chen, G. Z.; Shin, D. M.; Yang, L.; Young, A. N.; Wang, M. D.; Nie, S. Nat. Biotechnol. 2008, 26, 83–90. (15) Sun, L.; Sung, K.-B.; Dentinger, C.; Lutz, B.; Nguyen, L.; Zhang, J.; Qin, H.; Yamakawa, M.; Cao, M.; Lu, Y.; Chmura, A. J.; Zhu, J.; Su, X.; Berlin, A. A.; Chan, S.; Knudsen, B. Nano Lett. 2007, 7, 351–356. (16) Lutz, B.; Dentinger, C.; Sun, L.; Nguyen, L.; Zhang, J.; Chmura, A. J.; Allen, A.; Chan, S.; Knudsen, B. Histochem. Cytochem. 2008, 56 (4), 371–379. (17) Liu, Y. N.; Zou, Z. O.; Liu, Y. O.; Xu, X. X.; Yu, G.; Zhang, C. Z. Spectrosc. Spectral Anal. 2007, 27, 2045–2048. (18) Faulds, K.; Jarvis, R.; Smith, W. E.; Graham, D.; Goodacre, R. Analyst 2008, 133, 1505–1512.

3970

Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

offset Raman spectroscopy (SORS)19 and transmission Raman spectroscopy. In this work, we have used transmission Raman spectroscopy to demonstrate the concept of deep Raman spectroscopy. As discussed earlier,21 this method can be considered a special form of spatially offset Raman spectroscopy where the collection and laser delivery areas are displaced to the extreme, i.e., being on the opposite sides of the sample. This experimental geometry is particularly well suited to breast cancer detection and for this reason has been selected for this study. The results and conclusions reached can be straightforwardly extended to the pure form of spatially offset Raman spectroscopy, where the collection and deposition areas are on the same side of the sample. A feature of a particular distinction between the two variants is that the accessible depth with the pure form of SORS is about a half of that achievable with the transmission concept. This is due to the fact that the photons in the pure form of SORS have to penetrate to the target zone and then travel back to the same surface (i.e., traversing the same distance twice) whereas with the transmission Raman geometry the photons propagate this distance only once and after reaching the target zone they propagate to the opposite side of the sample. The applicability of both the deep Raman approaches and their commonalities were discussed in several earlier publications.20,21 Rather than rely upon confocal techniques able to probe depths of up to around 200 µm in turbid media, deep Raman techniques have been able to distinguish chemical composition of samples through around 2.5 cm thick or 100 times deeper.22 The limits have theoretically been extrapolated to 4-5 cm of tissue thickness with proposed optimization of instrumentation.1 This rapidly expanding field is leading toward many potential applications for in vivo medical diagnostics from probing bone composition through the skin to disease specific breast calcifications and soft tissue lesions for cancer diagnostics.23-27,2 A further possibility includes in vivo monitoring of drug delivery and identification of the location of the drug interactions. SERRS active nanoparticles produced using a specific functional linker to provide a SERRS signal using near infrared (NIR) excitation28 were used for this study. They demonstrate the signal obtainable from a sample when molecule-specific SERRS nanoparticles are conjugated with their target molecule of interest. This could include many targets, such as PSA, p53 proteins, DNAfragments, and cell specific proteins. One of the key benefits of SERRS is that the signals are obtained only from conjugated SERRS nanoparticles, rather than from the target molecule of interest. This allows us to demonstrate its generic performance by utilizing a set of conjugated particles producing the same signal as would be obtained from SERRS particles conjugated directly (19) Matousek, P.; Clark, I. P.; Draper, E. R. C.; Morris, M. D.; Goodship, A. E.; Everall, N.; Towrie, M.; Finney, W. F.; Parker, A. W. Appl. Spectrosc. 2005, 59, 393. (20) Matousek, P. Chem. Soc. Rev. 2007, 36, 1292. (21) Macleod, N. A.; Matousek, P. Appl. Spectrosc. 2008, 62, 291A. (22) Matousek, P.; Stone, N. Analyst 2009, 134, 1058. (23) Schulmerich, M. V. Proc. SPIE 2006, 6093, 60930O. (24) Schulmerich, M. V.; Dooley, K. A.; Morris, M. D.; Vanasse, T. M.; Goldstein, S. A. J. Biomed. Opt. 2006, 11, 060502. (25) Matousek, P.; Draper, E. R. C.; Goodship, A. E.; Clark, I. P.; Ronayne, K. L.; Parker, A. W. Appl. Spectrosc. 2006, 60, 758. (26) Schulmerich, M. V. J. Biomed. Opt. 2008, 13, 020506. (27) Matousek, P.; Stone, N. J. Biomed. Opt. 2007, 12, 024008. (28) McKenzie, F.; Ingram, A.; Stokes, R.; Graham, D. Analyst 2009, 134 (3), 549–556.

as the depth series. All spectra were corrected for spectrometer wavelength sensitivity variations using a luminescent green glass standard. A sixth order polynomial was fit to the spectra to remove any luminescent background found in the spectra.

Figure 1. Schematic of Agc solution preparation and injection into 25 mm × 40 mm × 40 mm porcine tissue sample. Transmission Raman apparatus with arrows indicating illumination and inelastic scattering collection.

with molecules of diagnostic interest. This reveals the prospects for numerous clinical applications such as early cancer detection and staging, metastatic detection, treatment monitoring, and chemo-sensitivity, provided that the disease specific molecule target molecule is known.29 EXPERIMENTAL SECTION Tissue preparation involved collection of fresh porcine samples from the local abattoir and dissecting them to the required dimensions. Muscle tissue was chosen to represent dense human tissues in this instance. A demonstration of signal obtained through increasing depth was achieved by cutting samples from pork muscle to thicknesses of 15 and 25 mm (other dimensions 40 mm × 40 mm) and mounted in the experimental apparatus. Citrate reduced silver (Agc) conjugated nanoparticles were used for this study. Silver nanoparticles were prepared with a NIR dye tag and stabilized with a linker as previously reported.28 The Agc suspension was injected into the center of the tissue samples. A volume of 100 µL of 50 pM was inserted into the 15 mm thick specimen, and a volume of 280 µL of 35 pM Agc was inserted into the 25 mm specimen. The number of conjugated SERRS nanoparticles was approximately 3.0 × 109 and 5.9 × 109, respectively. One side of the sample (as shown in Figure 1) was illuminated with 200 mW of 830 nm laser light through a dielectric filter tuned to 830 nm. The inelastically scattered photons were collected on the opposite side of the sample as demonstrated previously.1 The resulting spectra were measured in 10 s. An increasing SERRS particle density series was studied using a sample of pork muscle, cut to approximate dimensions of 40 mm × 40 mm × 25 mm. The shortest dimension was used as the transmission axis in this experiment (see Figure 1). The concentration of citrate reduced silver (Agc) NIR conjugated nanoparticles within the sampling volume was increased slowly. Agc suspensions were injected into the center of the tissue in volumes ranging from 40 to 280 µL and particle concentrations ranging from 15 to 50 pM. This produced a series of samples with increasing numbers of SERRS-conjugated nanoparticles from 3.7 × 108 to 5.9 × 109. A transmission Raman spectrum was acquired for 10 s at every step using the same experimental configuration (29) Rao, C. G.; Chianese, D.; Doyle, G. V.; Miller, M. C.; Russell, T.; Sanders, R. A., Jr.; Terstappen, L. W. Int. J. Oncol. 2005, 27 (1), 49–57.

RESULTS AND DISCUSSION This is the first demonstration of our approach to improving depth penetration used porcine tissue and silver nanoparticles functionalized to contain a permanent Raman signal. Porcine tissue was chosen to better mimic human tissue and is preferred to avian tissue due to the more realistic mixture of fats and proteins. Silver nanoparticles were used as they give a stronger Raman enhancement than gold or other metallic nanoparticles. Figure 2 shows the NIR-Raman spectra acquired of conjugated Agc (approximately 3 × 109 and 6 × 109) SERRS nanoparticles injected into the two samples through 15 and 25 mm of tissue, respectively. All spectra were corrected for energy sensitivity of the detection system (see Experimental Section) and background subtracted using a sixth order polynomial. NIR-Raman spectral data were collected in a transmission geometry through a 25 mm thick tissue block with an increasing SERRS nanoparticle density injected into the center. Principal components were calculated using the data (following mean centering).30,31 This enabled the identification of the key spectral load (others not shown) representing the changing SERRS nanoparticle concentration in the data set. It is quite apparent from the loads that the signal from the Agc is almost exclusively manifest in PC4 (Figure 3). The variance described in the first 4 components is approximately 100%. The relative contribution from each principal component to the signal for each spectrum is described by the principal component scores. With these scores plotted for PC4 for each spectrum in the series, it was possible to identify when the Agc signal became distinct from the spectral noise, i.e., became detectable through 25 mm of tissue. The score for PC4 is plotted against the approximate number of conjugated Agc particles injected into the tissue sample, to provide a measure of the relative detected signal across the series. Figure 3 shows this plot and allows the effective detection threshold to be observed for this particular experimental configuration and SERRS conjugates. It also shows the spectrum for Agc plotted with the PC4 load (multiplied by -1) to demonstrate that this load is directly representative of the SERRS signal from the Agc. In the experimental configuration explored here, the threshold for detection was shown to be 1.1 × 109 particles injected into the sampling volume. Within a sample of dimensions 25 mm × 40 mm × 40 mm, the sampling volume can be crudely approximated from photon migration theory by a sphere the diameter of the sample thickness, i.e., 25 mm. This leads us to a sampling volume of 8.2 × 1012 µm3, and therefore the mean particle volume density ranged from 4.5 × 10-4 to 72.4 × 10-4 µm-3 for this experiment (detection threshold ) 1.36 × 10-4 µm-3). For a lesion of diameter 5 mm (∼65 mm3), its detection through 25 mm tissue would require at least 1.1 × 109 particles (30) Pearson, K. Philos. Mag. 1901, 2 (6), 559–572. (31) Martens, H.; Naes, T. Multivariate Calibration; John Wiley & Sons: Chichester, U.K., 1991.

Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

3971

Figure 2. Background subtracted Raman spectra of Agc through 15 mm and 25 mm (×100) of porcine muscle tissue. Acquisition times of 10 s for both; the 25 mm sample had approximately twice as many nanoparticles injected (3.01 × 109 vs 5.93 × 109).

Figure 3. Plot of PC4 score versus number of conjugated SERRS nanoparticles injected into the 25 mm thick pork tissue. PC loads can indicate negative and positive spectral contributions and the load for PC4 is negative. The score plot represents an increasingly negative amount of PC4 with greater particle numbers. This indicates an increasing signal from the SERRS particles. The insert shows the load for PC4 multiplied by -1 for ease of visualization with the spectrum of the Agc through 15 mm of tissue. Table 1. Numbers of Agc SERRS Nanoparticles Required Per Cell for Detection versus Possible Lesion Size, In the Transmission Configuration Used with 25 mm Thick Tissue Samples diameter of lesion/mm no. of particles per cell for detection

10 3.5

5 27.9

accumulated within the volume of the lesion. For epithelial cells which may be the target of interest, cell volume is of the order of 520 µm3 (for 10 µm diameter cells as an example). The detection would therefore require that at least 28 particles are attached or absorbed into each cell. This is well within a clinically viable range. See Table 1 for the number of Agc SERRS 3972

Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

4 54.6

3 129.3

2 436.5

1 3 492.0

particles required per cell for detection of various lesion sizes in the experimental configuration used here. A study exploring the epithelial cell adhesion molecule (EpCAM) on circulating tumor cells from 100 metastatic carcinoma patients showed EpCAM expression ranging from 9 900 to 246 000 (mean 49 700) antigens per cell.29 This

demonstrates that if we were labeling the SERRS nanoparticles for EpCAM, we would be able to identify lesions of 1 mm or less in diameter (assuming 7% of the mean number of adhesion molecules can be conjugated with around 3 500 particles per cell required (see Table 1)). The maximum depth of penetration shown by Keren et al. using a traditional Raman microscope setup, evaluated using a tissue mimicking phantom, was 5.5 mm for 1.3 nM SERS nanoparticles in 200 µL volume.13 This is equivalent to around 100 times as many nanoparticles and a fifth of the depth achieved in this study. In the example shown here, which demonstrates the feasibility of combining SERRS nanoparticles and deep Raman sampling methods, it was apparent that the SERRS conjugates had a strong additional fluorescent background. This indicates that some of the nanoparticles within the solution were not fully conjugated, and therefore the number of particles contributing to the SERRS signal was actually less than expected for the number of nanoparticles in the suspension. Hence the signal-to-noise achieved in this study would have been lower than the ideal. This leads to the conclusion that the results demonstrated here are less than optimal and significant further increases in depth and a reduction in the detection limit is perfectly feasible with additional optimization of solutions and illumination/collection configurations. Note that the fluorescent background, from the slightly dissociated label dye, is spectrally extremely broad compared to the sharp features of the SERRS signals from the dye. It is this characteristic which enables the multiplexing to be trivial with SERRS but extremely (32) Macaskill, A.; Chernonosov, A. A.; Koval, V. V.; Lukyanets, E. A.; Fedorova, O. S.; Smith, W. E.; Faulds, K.; Graham, D. Nucleic Acids Res. 2007, 35 (6), e42. (33) Faulds, K.; McKenzie, F.; Smith, W. E.; Graham, D. Angew. Chem., Int. Ed. 2007, 46 (11), 1829–1831.

difficult with fluorescence techniques; as well as making it feasible to achieve a high penetration depth and reliable identification of the detected analyte. Previous work by Graham et al. and others have demonstrated the ease of multiplexing with labeled DNA probe sequences. These probes can have emissive Raman yields on par with, and in some cases exceeding those of, fluorescence.18,32,33 In combination with our results using deep Raman techniques, this leads us to conclude that this technique holds a considerable promise for multiplexing a selection of targets at depths of several centimeters in real time. Further exploration of the safety and biocompatibility of injecting these particles will need to be performed prior to in vivo study. CONCLUSIONS The new technique of SESORS has been proposed and discussed here. The concept was demonstrated in transmission Raman geometry. The method has potential to provide multiplexed read out of molecule-specific SERS particles deeply buried within mammalian tissues. The prospect of utilizing the technique for identification of small tumors has been proposed and shown to be technologically feasible based on the results shown here. More work is required to deliver an in vivo demonstration of the technique; however, with numerous groups exploring in vivo SERS the extension to SESORS should be trivial. ACKNOWLEDGMENT Nick Stone holds a Senior (Career Scientist) Research Fellowship funded by the U.K. National Institute of Health Research. Received for review January 7, 2010. Accepted April 2, 2010. AC100039C

Analytical Chemistry, Vol. 82, No. 10, May 15, 2010

3973