DSpace at VNU: Surface-Enhanced Raman Spectroscopy Study of 4-ATP on Gold Nanoparticles for Basal Cell Carcinoma Fingerprint Detection

4.—e-mail: namnh@hus.edu.vn The surface-enhanced Raman signals of 4-aminothiophenol 4-ATP attached to the surface of colloidal gold nanoparticles with size distribution of 2 to 5 nm were

Surface-Enhanced Raman Spectroscopy Study of 4-ATP

on Gold Nanoparticles for Basal Cell Carcinoma

Fingerprint Detection

LUU MANH QUYNH,1NGUYEN HOANG NAM,1,2,4K KONG,3 NGUYEN THI NHUNG,1I NOTINGHER,3M HENINI,3

1.—Faculty of Physics, Hanoi University of Science, Vietnam National University, Hanoi, 334 Nguyen Trai, Hanoi, Vietnam 2.—Nano and Energy Center, Hanoi University of Science, Viet-nam National University, Hanoi, 334 Nguyen Trai, Hanoi, VietViet-nam 3.—School of Physics and Astronomy, Nottingham University, University Park, Nottingham NG7 2RD, UK 4.—e-mail:

namnh@hus.edu.vn

The surface-enhanced Raman signals of 4-aminothiophenol (4-ATP) attached

to the surface of colloidal gold nanoparticles with size distribution of 2 to 5 nm were used as a labeling agent to detect basal cell carcinoma (BCC) of the skin

The enhanced Raman band at 1075 cm 1corresponding to the C-S stretching vibration in 4-ATP was observed during attachment to the surface of the gold nanoparticles The frequency and intensity of this band did not change when the colloids were conjugated with BerEP4 antibody, which specifically binds to BCC We show the feasibility of imaging BCC by surface-enhanced Raman spectroscopy, scanning the 1075 cm 1 band to detect the distribution of 4-ATP-coated gold nanoparticles attached to skin tissue ex vivo

Key words: Skin cancer, basal cell carcinoma, surface-enhanced Raman

scattering, gold nanoparticles

INTRODUCTION Skin cancer is the most common type of cancer in

humans, and its incidence is increasing.1Basal cell

carcinomas (BCCs) constitute approximately 74% of

skin cancer cases worldwide.2 The most efficient

treatment for ‘‘high-risk’’ BCCs (i.e BCCs on the

face and neck or recurrent BCCs) is Mohs

micro-graphic surgery (MMS).3This procedure maximizes

the removal of tumor cells while sparing as much

healthy tissue as possible Although MMS provides

improved outcomes compared to other treatment

options, the need for a pathologist or specialized

surgeon to diagnose frozen sections during surgery

has limited the widespread use of this approach,

leading to cases of inappropriate inferior treatment

Frozen-section histopathology also requires

labori-ous and time-consuming procedures, resulting in

increased costs compared to standard excision of BCC

Raman spectroscopic imaging is a promising technique for the diagnosis of skin cancers, given its high sensitivity to molecular and structural changes associated with cancer The use of Raman spectroscopy to detect biochemical alteration in skin tissue caused by BCC was first demonstrated by Gniadecka et al.4 Raster-scanning Raman spectral mapping has been used to image BCC in tissue samples ex vivo in MMS.5,6 However, raster-scan-ning Raman mapping requires long data acquisition times, typically days for tissue specimens of

1 cm 9 1 cm More recently, multimodal spectral imaging based on tissue autofluorescence and Raman spectroscopy has been used to reduce the time for diagnosis of BCC to only 30–60 min, which becomes feasible for use during MMS.7,8

An alternative technique that can reduce data acquisition and BCC diagnosis time during MMS is surface-enhanced Raman spectroscopy (SERS) It

(Received October 11, 2015; accepted February 20, 2016)

Ó2016 The Minerals, Metals & Materials Society

was discovered that in the very close vicinity of

metal nanostructures, strongly increased Raman

scattering signals could be obtained, due mainly to

resonances between optical fields and the collective

oscillations of the free electrons in a metal Since the

discovery of this surface-enhanced Raman (SER)

scattering in 1974,9 it has been recognized as a

powerful technique for biomedical applications SER

scattering has been studied for cancer

detec-tion,10–12 and it has been widely used in molecular

structure analysis.13–16 For non-labeling agent

probes, Raman spectra were analyzed by measuring

the intrinsic signals to distinguish between healthy

and diseased regions.10,11 In these studies, SER

signals of cancer-specific biomolecules were

reported as effective indicators of the presence of

cancer genes10 and cancer expression.11 However,

the signals were still broadened, thus posing a

challenge in distinguishing the cancerous from

non-cancerous areas It was noted that the SER peaks of

some linkages that were close to the metal surface

were strong, individually sharp, and did not change,

as the metal–organic complex was attached to other

organisms or molecules In the present study, we

investigated the SER signal of 4-aminothiophenol

(4-ATP, sometimes called p-aminothiophenol

[PATP]) linked to the surface of gold nanoparticles

conjugated with the skin carcinoma cell antibody

BerEP4 With this BCC-specific antibody

conjuga-tion, SER signals of some linkages from the 4-ATP

organic molecules were noted to be stable and

potentially to allow detection of the tumor regions

Here, we investigate the usefulness of these SERS

probes for the detection of BCC in ex vivo specimens

EXPERIMENTS AND METHODS

Synthesis of Gold Nanoparticles Coated with

4-ATP (Au-4ATP)

Gold nanoparticles ranging in size from 2 to 5 nm

were prepared by a wet chemical process using

cetyltrimethylammonium bromide (CTAB; Merck,

99%) Specifically, ion Au3+ from chloroauric acid

(HAuCl4; Merck, 99%) was prepared in

double-distilled water We placed 75 ml of CTAB 0.2 M and

0.2 ml of HAuCl40.5 M in a 200-ml flask, which was

then diluted with double-distilled water to obtain

100 ml of 1 mM HAuCl4 in 0.15 M CTAB We used

sodium borohydride (NaBH4; Merck, 99%) 0.1 M to

reduce the dark yellow Au3+ion-containing solution

to a dark brown After 12 h, the solution had

changed to a dark red Next, 4-ATP 10 3M (Merck,

99%) was injected into the solution at a 1:40 volume

rate After 12 h, the solution was washed several

times by centrifugation The resulting solution is

referred to as the Au-4ATP solution

The structural and morphological properties of

the Au-4ATP sample were investigated using a

Bruker D5005 x-ray diffractometer (XRD) and

JEOL JEM-1010 transmission electron microscope

(TEM)

Conjugation of Au-4ATP with BerEP4 (Au-4ATP-Antibody)

For antibody conjugation, 3 mg 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) was mixed with 1 mg BerEP4, and was then added to the Au-4ATP solution The mixture was incubated for a minimum of 20 min until the Au-4ATP had completely reacted with the BerEP4 molecules Skin Tissue Samples

Skin tissue sections were obtained from the Nottingham University Hospitals National Health Service Trust Tissue sections were cut from blocks removed during surgical procedures and sliced into one-third thickness Two of the three tissue slides were investigated using an optical microscope with bright-field imaging, with and without conventional hematoxylin and eosin (H&E) staining The third slide was treated with the Au-4ATP-antibody com-plex before subjection to Raman microspectroscopy Raman Spectroscopic Measurements

Raman spectroscopic measurements were carried out using a custom-made Raman microspectrometer built by Notingher’s group.6The laser power was set

to 20 mW to avoid sample damage, the scanning interval was set from 600 cm 1 to 1700 cm 1, and the integration time was set to 0.1 s

SER Spectra of Au-4ATP and Au-4ATP-Antibody

One drop each of the Au-4ATP and Au-4ATP-antibody colloidal solutions were deposited on the sample holder surface The spectra of the samples were observed separately and were then drawn in one image to compare the differences

Scanning Measurement of Au-4ATP-Antibody-Treated Tissue

For the third tissue sample, 1 lL of Au-4ATP-antibody-containing solution was deposited onto the surface of the sample After conjugation of the antibody with the cells for a period of 5 min, the scanning measurement was initiated SER scatter-ing signals were collected for every 1 lm 9 1 lm on

a 40 lm 9 40 lm region of the tissue sample There were 1600 spots in total All 1600 spectra from the 1600 spots were collected and analyzed using two methods In the first, principal component analysis was employed.17In the second method, we consider the peak at 1075 cm 1corresponding to a stretching band of C-S linkage from the 4-ATP molecules Peak heights at 1075 cm 1were mapped, depending on the position of the single spots This landscape image was then examined as a finger-printed image in comparison with normal bright-field images of the first two samples and with the image obtained by principal component analysis

RESULTS AND DISCUSSION

Structure and Morphology of Gold

Nanoparticles

Figure1 illustrates the XRD pattern and TEM

image of the as-prepared gold nanoparticles The XRD

peaks at 38.2°, 44.4° and 64.7° indicate the (111), (200)

and (220) reflection phases of the fcc structure,

respectively The calculated lattice parameter was

4.08 ± 0.05 A˚´ , which agreed with earlier works.18 , 19

Using the Debye–Scherrer formula, a particle size of

4.2 nm was determined, in good agreement with the

results observed in the TEM image, where most of the

particles were distributed at 4 nm

Gold Nanoparticles Surface Modification

Figure2 shows a schematic of the reaction

between the Au-4ATP and the carboxyl group

(–COOH) from the antibody BerEP4 Under the

catalytic effect of EDC, the free amino group (–NH2)

of the Au-4ATP colloid reacted with the carboxyl

group, and a peptide group (–NH–CO) was created

This reaction created a stable covalent linkage

binding the gold nanoparticles and antibody to form

the Au-4ATP-antibody

The Raman spectra of raw 4-ATP 10 3M and

Au-4ATP were observed (data not shown) Slight shifts

of the peaks were experienced as the 4-ATP

molecules were deposited on the surface of the Au

nanoparticles, which corresponds to the linking of

the molecules with the metal particles via the Au-S

bond In addition, significant magnification of

Raman intensity was detected in the Au-4ATP

spectrum in comparison with that of the raw

4-ATP Our results show the greatest intensity

enhancement of approximately 105 times Due to

the reaction shown in Fig.2, some vibration modes

corresponding to the –NH2 disappeared and were

replaced by vibrations of the peptide linkage, with

the majority occurring in the BerEP4 molecules The electromagnetic field surrounding the metal nanoparticles was enhanced from the surface plas-mon resonance (SPR) effect, which increased the Raman signal of the vibrations near the particle surfaces.20,21 With the significantly increased Raman signal in the SER scattering, it can then

be used to detect changes to the surface of each colloid solution after linking the antibody with the gold nanoparticles When one –NH linkage from the free NH2 was exchanged, and large BerEP4 mole-cules then attached to the surface of the gold nanoparticles, some SER peaks containing –NH vibrations disappeared, and peaks characterizing the peptide link appeared, as shown in Fig 3 The SER spectra of and Au-4ATP-antibody-containing samples are shown in Fig.3

We can clearly see that the Raman peaks of Au-4ATP measured at 1495 cm 1, 1432 cm 1 and

1134 cm 1 disappeared after conjugation of the

Fig 1 (a) X-ray diffraction pattern of as-prepared gold nanoparticles The black pattern shows the measurement data and the red vertical lines show the standard diffraction positions of the (111), (200) and (220) planes of Au bulk material (pattern 4-784) (b) TEM image of as-prepared gold nanoparticles The dark gray and black dots show the presence of the nanoparticles in the sample Inset: size distribution of the nanoparticles calculated from the TEM image.

Fig 2 Schematic graph of peptide link created by the reaction The formation of the Au–S covalent bond is a well-known phenomenon, linking the 4-ATP molecules to the surface of the gold nanoparticle surfaces, and allowing the amino group (-NH 2 ) to freely dissolve in solution After the reaction of the carboxyl groups (-COOH) from the antibody BerEP4 molecules with the present catalyst EDC, peptide (-NH-CO-) binding occurs Here, R COOH denotes the whole anti-body, of which we consider the reaction of only one carboxyl group, with R remaining.

antibody These peaks were assigned to mCC + dCH,

mCC + dCH and dCH vibrations, respectively, where

m denotes stretching movement and d bending

movement.22,23 We suggest that the wagging

vibra-tion NH2 linkage (pNH2) may occur together with

these vibrations The disappearance of the pNH2

vibration due to the reaction described in Fig 2may

be responsible for the disappearance of peaks at

1495 cm 1, 1432 cm 1 and 1134 cm 1 We also

1382 cm 1 after the antibody conjugation This

peak was assigned to the dCH + mCC vibration

modes.22,23 The change from the –CN– linkage to

the peptide linkage (-NH-CO-) may be responsible

for the disappearance of this peak In addition, new

peaks arise at 1449 cm 1and 1297 cm 1 The peak

at 1449 cm 1is assigned to CH2, CH3deformation,

and the peak at 1297 cm 1 is assigned to the

vibration of the helix structure of amide III

link-age.24 We should note that Raman peaks below

1005 cm 1were not considered because the peaks in

this region may also correspond to the phonon

vibrations of the metal material

Owens et al investigated enhanced Raman

spec-tra of 4-ATP on an Au-subsspec-trate conjugated with

anti-p53 protein.25 The characteristic peak of C-S

linkage close to 1080 cm 1was also employed as a

detection signal When the 4-ATP-modified Au

surface was covalently connected with anti-p53

molecules, the peak position corresponding to the

C-S vibration observed at 1080 cm 1 shifted to a

higher wavenumber within 1 cm 1 After protein–

antibody interaction, the peak position shifted about

1 cm 1, depending on the added protein

concentra-tion We observed the same effect in our Raman

investigation As revealed in Fig 3, the strongest

peak is observed at 1075 cm 1, which is assigned to

a mCS vibration, while a strong peak at 1614 cm 1is

assigned to a mCC vibration.22,23 Interpretation of

the peak at 1614 cm 1may be easily confused with

other organic molecules, because mCC vibrations are quite common for organic systems.6,24 A sharp individual peak at around 1075 cm 1 was detected

by Zheng et al on the SERS spectrum of 4-ATP absorbed on a silver surface,15 by Osawa et al on SERS spectrum of 4-ATP absorbed on a silver film,22 and by Jiao et al on SERS spectrum of 4-ATP on an

Au surface.23,26 However, the peak at 1075 cm 1 was not detected on the SERS spectra of the antibody and/or polypeptide on metal sub-strates.24,27 We propose the enhanced Raman peak

at 1075 cm 1as a strong signal for the detection of the position and concentration of Au-4ATP nanopar-ticles, and hence, of antibody molecules

Fingerprinted Landscape of BCC Tissue

As discussed in the ‘‘Experiments and Methods’’ section, skin tissue sections obtained from the Nottingham University Hospitals National Health Service Trust were cut from blocks removed during surgical procedures, and were sliced into one-third thickness Two of three tissue slides were investi-gated using an optical microscope with bright-field imaging, with and without conventional hema-toxylin and eosin (H&E) staining The third slide was treated with the Au-4ATP-antibody complex as described above, before it was subjected to investi-gation by Raman microspectroscopy The SER scat-tering signal of every 1 lm 9 1 lm spot on a

40 lm 9 40 lm region of the tissue sample was observed and analyzed using two methods The first was principal component analysis,17 in which the SER spectra were compared to the averaged spec-trum, and the difference was then shown in the landscape In the second method, the peak at

1075 cm 1 corresponding to the stretching band of C-S linkage from 4-ATP molecules was considered Peak heights at 1075 cm 1were mapped, depending

on the position of the single spots The fingerprinted landscape of SER signals of the Au-4ATP antibody

on BCC tissue is shown in Fig.4

In this work, simple H&E staining was used as control diagnosis; the color image of the tissue is shown in Fig.4a In this non-specific method employed in Nottingham University Hospitals National Health Service Trust, immunofluorescence labeling has not been used, and only regions of condensation on the tissue sample have been con-sidered, where the areas of cancer cells may be observed as dark-colored regions—for example, the regions marked with the red circles as A1 and A2 in Fig.4 However, the diagnostic result is ultimately the subjective decision of the pathologist, because this method may lead to misinterpretation of non-cancerous areas as cancer cells As can be seen in Fig.4b, which shows the bright-field microscopic result, with this sample, it is easy to confirm that the B1 region corresponds to a hair follicle position, and B2 does not, although B1 and B2 have the same position on the tissue as the regions marked A1 and

Fig 3 SER spectra of Au-4ATP- and Au-4ATP-antibody-containing

samples.

A2 in Fig.4a Thus we see that the dark-colored A1

region can be misinterpreted The results of

princi-pal component analysis of the SER signals are

illustrated in Fig.4c, which shows a comparison of

the individual SER signals and then the difference

between the SER spectra and average spectrum In

this landscape, the yellow to red areas, such as C1

and C2, can be considered as regions of cancer The

C1 and C2 regions have the same position as the A1

and A2 and the B1 and B2 regions, respectively

Figure4d shows the results of the SER signal

analyzed using only the intensity of SER peaks at

1075 cm 1, with the D1 and D2 regions having the

same position as regions C1 and C2 With this method, the antigen–antibody coupling orients the Au-4ATP-antibody colloids close to the BCC surface, and the carcinoma sections act as a dock at which high concentrations of Au-4ATP-antibody particles are distributed, and thus the SER peak intensity at

1075 cm 1 is higher in these areas In Fig 4d, we can see the results of using the peak height of

1075 cm 1 for mapping the Au-4ATP-antibody areas appearing within the 40 lm 9 40 lm region However, the D1 area in Fig 4d does not show the high intensity of the peak at 1075 cm 1, while the other areas such as D2 indicate very high intensity

Fig 4 Fingerprinted landscape of SER signals of Au-4ATP-antibody on BCC tissue (a) Image of Gram-stained BCC tissue, where regions marked A1 and A2 are the areas of suspected BCC (b) Bright-field microscopy image of BCC tissue, where regions B1 and B2 are in the same position on the tissue as regions A1 and A2, respectively (c) SER signal landscape analyzed by the principal component method, where regions C1 and C2 are in the same position on the tissue as regions A1 and A2, respectively (d) The fingerprinted landscape of intensity of SER peaks at

1075 cm 1 , where regions D1 and D2 are in the same position on the tissue as regions A1 and A2, respectively The difference between D1 and D2 shows that only the red-colored D2 and similar-colored area are diseased, while D1 is not These results show that this method is a better solution for intraoperative diagnosis.

Both Nijisen et al.28 and Notingher et al.7

reported on the use of Raman spectra for

discrim-ination of BCC, in which differences in Raman

spectra were observed between diseased and

healthy tissue The total intensity of the Raman

spectrum of the BCC-infected region was higher

than that of the healthy region, which the authors

reported as corresponding to the higher

accumula-tion of lipids and nucleic acids within the cancer

cells Scan images of the skin tissue were

con-structed from total intensity calculations and were

employed to distinguish the diseased from healthy

tissue Without selective detection, the Raman

spectra were only able to discriminate the regions

of lipid and nucleotide condensation from other

regions, which could lead to misinterpretation if the

healthy cells also have condensed organic

organ-isms, such as the skin follicle region shown in our

experiment In addition, no specific peaks would be

applicable for selective discrimination of BCC tissue

from healthy tissue, leading to long acquisition time

for intraoperative diagnosis (5–20 h/mm2).7

From the results described above, only the regions

marked as A2, B2, C2 and D2 can be confidently

interpreted as cancerous tissue, while the A1, B1,

C1 and D1 regions may be assigned to hair follicles,

where the cell concentration is also higher In

principal component analysis, only those regions

differing from other regions and in which

non-diseased tissue can also be observed were

high-lighted, which may lead to misinterpretation

Fur-thermore, the SER mapping collection process

required more than 2 h, as the collection time for

each spectrum was nearly 5 s, whereas the

finger-printed image using peak height at 1075 cm 1

required only around 5 min, as the acquisition could

be focused only on the narrow band around the

1075 cm 1 peak (e.g narrow filter) rather than

collection of the entire spectrum, and the

integra-tion time for each pixel could thus be reduced to 0.1–

0.2 s Hence, this method may represent a solution

for quick surgical diagnostic imaging

CONCLUSION

In conclusion, we successfully used the SER

signal of the C-S link vibration at 1075 cm 1 on

gold nanoparticles to detect BCC-contaminated

regions of skin tissue samples The 4-ATP-coated

gold nanoparticles were conjugated with the

BerEP4 antibody, which specifically recognizes

BCC With the fingerprint method using the SER

peak at 1075 cm 1, an image of a 40 lm 9 40 lm

skin sample was obtained, and showed the position

of the tumors These SERS probes show promise for

fast and selective diagnosis of BCC through the

collection of the fingerprinted spectral image of skin resections Furthermore, all results can be observed and analyzed automatically, requiring no subjective interpretation by pathologists

REFERENCES

1 J.M Baxter, A.N Patel, and S Varma, BMJ 345, e5342 (2012).

2 National Cancer Intelligence Network (NCIN), Non-me-lanoma Skin Cancer in England, Scotland, Northern Ire-land, and Ireland (London: NCIN, 2013).

3 S.V Mohan and A.L.S Chang, Curr Dermatol Rep 3, 40 (2014).

4 M Gniadecka, H.C Wulf, O.F Nielsen, D.H Christensen, and J Hercogova, Photochem Photobiol 66, 418 (1997).

5 A Nijssen, T.C Bakker Schut, F Heule, P.J Caspers, D.P Hayes, M.H.A Neumann, and G.J Puppels, J Investig Dermatol 119, 64 (2002).

6 M Larraona-Puy, A Ghita, A Zoladek, W Perkins, S Varma, I.H Leach, A.A Koloydenko, H Williams, H Wil-liams, and I Notingher, J Biomed Opt 14, 054031 (2009).

7 K Kong, C.J Rowlands, S Varma, W Perkins, I.H Leach, A.A Koloydenko, H.C Williams, and I Notingher, Proc Natl Acad Sci USA 110, 15189 (2013).

8 S Takamori, K Kong, S Varma, I Leach, H.C Williams, and I Notingher, Biomed Opt Express 6, 98 (2015).

9 M Fleischmann, P.J Hendra, and A.J McQuillan, Chem Phys Lett 26, 163 (1974).

10 T Vo-Dinh, L.R Allain, and D.L Stokes, J Raman Spec-trosc 33, 511 (2002).

11 P.M Kasili, M.B Wabuyele, and T Vo-Dinh, NanoBiotechnology 2, 29 (2006).

12 L.R Allain and T Vo-Dinh, Analyt Chim Acta 469, 149 (2002).

13 N.J Kim, J Phys Chem C 114, 13979 (2010).

14 S Ye, L Fang, and Y Lu, J Raman Spectrosc 41, 1119 (2010).

15 J Zheng, Y Zhou, X Li, Y Ji, T Lu, and R Gu, Langmuir

19, 632 (2003).

16 Y.C Liu, Langmuir 18, 174 (2002).

17 C.M Stellman, K.S Booksh, A.R Muroski, M.P Nelson, and M.L Myrick, Sci Eng Comp Mater 7, 51 (1998).

18 N.N Long, L.V Vu, C.D Kiem, S.C Doanh, C.T Nguyet, P.T Hang, N.D Thien, and L.M Quynh, J Phys 187,

012026 (2009).

19 X Huang, I.H El-Sayed, and M.A El-Sayed, J Am Chem Soc 128, 2115 (2006).

20 H.Y Jung, Y.K Park, S Park, and S.K Kim, Anal Chim Acta 602, 236 (2007).

21 E.C Le Ru, E Blackie, M Meyer, and P.G Etchegoin, J Phys Chem C 111, 13794 (2007).

22 M Osawa, N Matsuda, K Yoshii, and I Uchida, J Phys Chem 98, 12702 (1994).

23 L.S Jiao, L Niu, J Shen, T You, S Dong, and A Ivaska, Electrochem Commun 7, 219 (2005).

24 N.C Maiti, M.M Apetri, M.G Zagorski, P.R Carey, and V.E Anderson, J Am Chem Soc 126, 2399 (2004).

25 P Owens, N Phillipson, J Perumal, G.M O’Connor, and

M Olivo, Biosensors 5, 664 (2015).

26 L.S Jiao, Z Wang, L Niu, J Shen, T You, S Dong, and A Ivaska, J Solid State Electrochem 10, 886 (2006).

27 T.M Herne, A.M Ahern, and R.L Garrell, Anal Chim Acta 246, 75 (1991).

28 A Nijssen, T.C Bakker Schut, F Heule, P.J Caspers, D.P Hayes, M.H.A Neumann, and G.J Puppels, J Invest Dermatol 119, 64 (2002).