Anal Bioanal Chem DOI 10.1007/s00216-014-7967-5
RESEARCH PAPER
The use of Au@SiO2 shell-isolated nanoparticle-enhanced Raman spectroscopy for human breast cancer detection Chao Zheng & Lijia Liang & Shuping Xu & Haipeng Zhang & Chengxu Hu & Lirong Bi & Zhimin Fan & Bing Han & Weiqing Xu
Received: 4 May 2014 / Revised: 26 May 2014 / Accepted: 11 June 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract This study uses the powerful fingerprint features of Raman spectroscopy to distinguish different types of breast tissues including normal breast tissues (NB), fibroadenoma (FD), atypical ductal hyperplasia (ADH), ductal carcinoma in situ (DCIS), and invasive ductal carcinoma (IDC). Thin frozen tissue sections of fresh breast tissues were measured by Raman spectroscopy. Due to the inherent low sensitivity of Raman spectra, Au@SiO 2 shell-isolated nanoparticleenhanced Raman spectroscopy (SHINERS) technique was utilized to provide supplementary and more informative spectral features. A total of 619 Raman spectra were acquired and compared to 654 SHINERS spectra. The maximum enhancement effect of distinct and specific bands was characterized for different tissue types. When applying the new criteria, excellent separation of FD, DCIS, and IDC was obtained for all tissue types. Most importantly, we were able to distinguish ADH from DCIS. Although only a preliminary distinction was characterized between ADH and NB, the results provided a good foundation of criteria to further discriminate ADH from NB and shed more light toward a better understanding of the mechanism of ADH formation. This is the first report to detect the premalignant (ADH and DCIS) breast tissue frozen sections and also the first report exploiting SHINERS to detect Chao Zheng and Lijia Liang contributed equally to this work. C. Zheng : H. Zhang : Z. Fan : B. Han (*) Department of Breast Surgery, The First Hospital of Jilin University, Changchun 130021, China e-mail: [email protected] L. Liang : S. Xu : C. Hu : W. Xu (*) State Key Laboratory for Supramolecular Structure and Materials, Jilin University, Changchun 130012, China e-mail: [email protected] L. Bi Pathology Department, The First Hospital of Jilin University, Changchun 130021, China
and distinguish breast tissues. The results presented in this study show that SHINERS can be applied to accurately and efficiently identify breast lesions. Further, the spectra can be acquired in a minimally invasive procedure and analyzed rapidly facilitating early and accurate diagnosis in vivo/in situ. Keywords Breast cancer . Frozen sections . SHINERS . Shell-isolated nanoparticles
Introduction Breast cancer is the most frequently diagnosed female cancer in China. It is responsible for 16.81 % of all female cancer and 7.54 % of all female cancer mortalities [1]. The 5-year relative survival rate of women diagnosed with localized breast cancer (LBC) is 98.6 %, survival declines to 83.8 % for regional stages and to 23.3 % for distant stages [2]. Early and accurate diagnosis of this disease is imperative for efficiently treating patients. Presently, although a routine examination includes ultrasound, mammography, and magnetic resonance imaging (MRI), it has been shown that 70–90 % of biopsies of detected lesions are benign [3]. So the need to develop a rapid, early, and accurate diagnosis tool has become imperative in medical science [4, 5]. The last decade has been marked by numerous efforts to explore Raman spectroscopy and to develop new efficient methods that could provide detailed and meaningful information about the biochemical composition and molecular structures of breast tissues [6–10]. Although Raman spectroscopy offers highly specific information on the molecular features of tissues, its use has remained limited because of its relatively low ability to produce distinguishable signals. Surfaceenhanced Raman scattering (SERS) was first applied in the 1970s with an electrochemically roughened Ag electrode [11, 12] and largely improved the low sensitivity limitation.
C. Zheng et al.
Subsequently, SERS attracted considerable interest as a potentially efficient clinical tool because of its nondestructive characteristics and its capacity to diagnose diseases [13–17] such as breast cancer in real-time [18–21]. However, unsatisfactory substrate universality and poor measurement reproducibility still hinder a more universal use of SERS. These limitations are linked by the fact that the substrate for SERS should have rough surfaces or particular metal nanostructures. It is known that hot spots (aggregated colloidal nanoparticles) provide giant SERS enhancement at the gaps, but it is difficult to ascertain that the probed molecules are located at the hot spots, especially when the biomolecules possess a large mass [22]. As a result, some SERS substrates have poor measurement reproducibility. To overcome these drawbacks, Tian’s group of Xiamen University [22] designed a shell-isolated nanoparticle-enhanced Raman spectroscopy (SHINERS). With SHINERS, the Raman signal amplification is provided by a gold core coated with an ultrathin silica or alumina shell (thickness from 2 to 20 nm). While the gold core of the Au@SiO2 shell-isolated nanoparticles (SHINs) allows obvious SERS enhancement, the shell is used to protect the gold core from degradation of the minutely fabricated nanostructures and protect the bare gold nanoparticles from probed adsorbates. These improvements extended significantly the versatility of Raman spectroscopy which could then be applied to surface and spectroscopic sciences, electrochemistry, environment protection, drug, food safety inspection, and even in the usage of industrial semiconductor to living cells [23]. However, the use of SHINERS in the diagnosis of any cancers, most particularly breast cancer has been very limited. In this study, SHINERS technique is applied to study breast tissues and characterize specific spectral information to further distinguish between normal tissues and the pathological states of fibroadenoma (FD), atypical ductal hyperplasia (ADH), ductal carcinoma in situ (DCIS), and invasive ductal carcinoma (IDC). Our work focused on three major parts. Firstly, we acquired the Raman spectra and SHINERS spectra of all tissue types; we assigned the main vibrational modes of the mean spectra and analyzed all spectral features for reproducible distinguishable characteristics. In the second stage, we showed that SHINs provide characteristic enhancement effects or specific Raman bands with various types of tissues. Using these new criteria, we were able to efficiently and reproducibly separated FD, DCIS, and IDC. Thirdly, in order to differentiate normal breast tissues (NB) from ADH, the enhancement effect of SHINs was utilized in combination with the mean ratios of the intensity of nucleic acids and the characteristic bands of proteins. Finally, we show that the SHINERS technique can successfully be applied to characterize breast tissues and accurately determine five types of breast tissues FD, DCIS, IDC, NB, and ADH.
Materials and methods Patients and samples Frozen sections were collected from 56 patients who underwent surgical resection or mammotome biopsy in the Department of Breast Surgery at the First Hospital of Jilin University. All patients agreed to participate in this research project, and the project and methodology were accepted by the Ethics Committee of Jilin University. The samples were immediately frozen after the operation at −20∼−25 °C. Two contiguous 6-μm sections were cut from each sample with a freezing microtome (LEICA-CM3050S, Germany). One section was stained with hematoxylin and eosin (HE) for routine histopathological analysis, and the other was transported to the research laboratory in liquid nitrogen. All breast slices were placed on a slide. Prior to analysis, the frozen section was thawed at 22 °C for 10 min. Multiple spectra were collected from each patient as some tissue samples contained both normal and diseased breast tissues. Characteristics of the patients are summarized in Table 1. Raman spectrometer A confocal Raman system (LabRAM ARAMIS, HORIBA Jobin Yvon, Edison, NJ, USA) with a ∼0.7-μm spatial resolution and a 5-mW/633-nm HeNe laser as excitation source was used for the collection of Raman spectra. The detection of Raman signal was carried out with a Synapse Thermoelectric cooled charge-coupled device (CCD) camera (HORIBA Jobin Yvon, Edison, NJ, USA). Raman scattering light was collected with a ×50 microscope objective lens (0.50 NA, LMPLFLN, Olympus, Japan) that was also used for focusing the excitation laser light. The laser beam focusing on the tissue formed a spot with a diameter of 1.5 μm. The strong Rayleighscattered lights were then blocked by a four-notch filter (HORIBA Jobin Yvon, Edison, NJ, USA). Extended scan spectra with a spectral range of 600–2,000 cm−1 were acquired using an integration time of 60 s and three accumulations. The wave number calibration was set by reference to the 520.7 cm−1 vibrational band of a silicon wafer. These settings were kept constant for all spectral measurements (Scheme 1). Data collection and process At first, tumor locations are located from HE sections and then relocated by Raman spectroscopy in the contiguous frozen sections. Ten to twelve spectra were collected from different locations in each sample to ensure representative sampling and incorporate spot-to-spot variability in the signal. Once all needed spectra were obtained, the SHINs were immediately added to the surface of the frozen sections. Finally, SHINERS spectra were collected.
The use of Au@SiO2 SHINERS for human breast cancer detection Table 1 Clinicopathological parameters Type
Number of patients
Ages (range)
Number of Raman spectra
Number of SHINERS spectra
Normal breast tissues (NB) Fibroadenoma (FD) Atypical ductal hyperplasia (ADH) Ductal carcinoma in situ (DCIS) Infiltrating ductal carcinoma (IDC) Total
9 17 7 8 15 56
40.3 (27–59) 37.1 (19–56) 44.8 (39–56) 43.0 (34–51) 51.6 (31–74) 44.1 (19–74)
103 144 89 123 160 619
92 181 85 139 157 654
A total of 619 Raman and 654 SHINERS spectra were obtained from all different tissues. The detailed information is presented in Table 1. All the spectra were dealt with baseline corrected by fitting and subtracting a third-order polynomial using the NGSLabSpec software (HORIBA Jobin Yvon, Edison, NJ, USA). The spectra were then smoothed with a 15point adjacent averaging. Statistical analysis was done using PASW Statistics Base 18 statistical software (SPSS Inc., Chicago, USA), using Kruskal Wallis test for the difference of the enhancement ratios of characteristic bands between five tissues and using Mann–Whitney test for the difference of characteristic band intensities before and after enhancement. p
RESEARCH PAPER
The use of Au@SiO2 shell-isolated nanoparticle-enhanced Raman spectroscopy for human breast cancer detection Chao Zheng & Lijia Liang & Shuping Xu & Haipeng Zhang & Chengxu Hu & Lirong Bi & Zhimin Fan & Bing Han & Weiqing Xu
Received: 4 May 2014 / Revised: 26 May 2014 / Accepted: 11 June 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract This study uses the powerful fingerprint features of Raman spectroscopy to distinguish different types of breast tissues including normal breast tissues (NB), fibroadenoma (FD), atypical ductal hyperplasia (ADH), ductal carcinoma in situ (DCIS), and invasive ductal carcinoma (IDC). Thin frozen tissue sections of fresh breast tissues were measured by Raman spectroscopy. Due to the inherent low sensitivity of Raman spectra, Au@SiO 2 shell-isolated nanoparticleenhanced Raman spectroscopy (SHINERS) technique was utilized to provide supplementary and more informative spectral features. A total of 619 Raman spectra were acquired and compared to 654 SHINERS spectra. The maximum enhancement effect of distinct and specific bands was characterized for different tissue types. When applying the new criteria, excellent separation of FD, DCIS, and IDC was obtained for all tissue types. Most importantly, we were able to distinguish ADH from DCIS. Although only a preliminary distinction was characterized between ADH and NB, the results provided a good foundation of criteria to further discriminate ADH from NB and shed more light toward a better understanding of the mechanism of ADH formation. This is the first report to detect the premalignant (ADH and DCIS) breast tissue frozen sections and also the first report exploiting SHINERS to detect Chao Zheng and Lijia Liang contributed equally to this work. C. Zheng : H. Zhang : Z. Fan : B. Han (*) Department of Breast Surgery, The First Hospital of Jilin University, Changchun 130021, China e-mail: [email protected] L. Liang : S. Xu : C. Hu : W. Xu (*) State Key Laboratory for Supramolecular Structure and Materials, Jilin University, Changchun 130012, China e-mail: [email protected] L. Bi Pathology Department, The First Hospital of Jilin University, Changchun 130021, China
and distinguish breast tissues. The results presented in this study show that SHINERS can be applied to accurately and efficiently identify breast lesions. Further, the spectra can be acquired in a minimally invasive procedure and analyzed rapidly facilitating early and accurate diagnosis in vivo/in situ. Keywords Breast cancer . Frozen sections . SHINERS . Shell-isolated nanoparticles
Introduction Breast cancer is the most frequently diagnosed female cancer in China. It is responsible for 16.81 % of all female cancer and 7.54 % of all female cancer mortalities [1]. The 5-year relative survival rate of women diagnosed with localized breast cancer (LBC) is 98.6 %, survival declines to 83.8 % for regional stages and to 23.3 % for distant stages [2]. Early and accurate diagnosis of this disease is imperative for efficiently treating patients. Presently, although a routine examination includes ultrasound, mammography, and magnetic resonance imaging (MRI), it has been shown that 70–90 % of biopsies of detected lesions are benign [3]. So the need to develop a rapid, early, and accurate diagnosis tool has become imperative in medical science [4, 5]. The last decade has been marked by numerous efforts to explore Raman spectroscopy and to develop new efficient methods that could provide detailed and meaningful information about the biochemical composition and molecular structures of breast tissues [6–10]. Although Raman spectroscopy offers highly specific information on the molecular features of tissues, its use has remained limited because of its relatively low ability to produce distinguishable signals. Surfaceenhanced Raman scattering (SERS) was first applied in the 1970s with an electrochemically roughened Ag electrode [11, 12] and largely improved the low sensitivity limitation.
C. Zheng et al.
Subsequently, SERS attracted considerable interest as a potentially efficient clinical tool because of its nondestructive characteristics and its capacity to diagnose diseases [13–17] such as breast cancer in real-time [18–21]. However, unsatisfactory substrate universality and poor measurement reproducibility still hinder a more universal use of SERS. These limitations are linked by the fact that the substrate for SERS should have rough surfaces or particular metal nanostructures. It is known that hot spots (aggregated colloidal nanoparticles) provide giant SERS enhancement at the gaps, but it is difficult to ascertain that the probed molecules are located at the hot spots, especially when the biomolecules possess a large mass [22]. As a result, some SERS substrates have poor measurement reproducibility. To overcome these drawbacks, Tian’s group of Xiamen University [22] designed a shell-isolated nanoparticle-enhanced Raman spectroscopy (SHINERS). With SHINERS, the Raman signal amplification is provided by a gold core coated with an ultrathin silica or alumina shell (thickness from 2 to 20 nm). While the gold core of the Au@SiO2 shell-isolated nanoparticles (SHINs) allows obvious SERS enhancement, the shell is used to protect the gold core from degradation of the minutely fabricated nanostructures and protect the bare gold nanoparticles from probed adsorbates. These improvements extended significantly the versatility of Raman spectroscopy which could then be applied to surface and spectroscopic sciences, electrochemistry, environment protection, drug, food safety inspection, and even in the usage of industrial semiconductor to living cells [23]. However, the use of SHINERS in the diagnosis of any cancers, most particularly breast cancer has been very limited. In this study, SHINERS technique is applied to study breast tissues and characterize specific spectral information to further distinguish between normal tissues and the pathological states of fibroadenoma (FD), atypical ductal hyperplasia (ADH), ductal carcinoma in situ (DCIS), and invasive ductal carcinoma (IDC). Our work focused on three major parts. Firstly, we acquired the Raman spectra and SHINERS spectra of all tissue types; we assigned the main vibrational modes of the mean spectra and analyzed all spectral features for reproducible distinguishable characteristics. In the second stage, we showed that SHINs provide characteristic enhancement effects or specific Raman bands with various types of tissues. Using these new criteria, we were able to efficiently and reproducibly separated FD, DCIS, and IDC. Thirdly, in order to differentiate normal breast tissues (NB) from ADH, the enhancement effect of SHINs was utilized in combination with the mean ratios of the intensity of nucleic acids and the characteristic bands of proteins. Finally, we show that the SHINERS technique can successfully be applied to characterize breast tissues and accurately determine five types of breast tissues FD, DCIS, IDC, NB, and ADH.
Materials and methods Patients and samples Frozen sections were collected from 56 patients who underwent surgical resection or mammotome biopsy in the Department of Breast Surgery at the First Hospital of Jilin University. All patients agreed to participate in this research project, and the project and methodology were accepted by the Ethics Committee of Jilin University. The samples were immediately frozen after the operation at −20∼−25 °C. Two contiguous 6-μm sections were cut from each sample with a freezing microtome (LEICA-CM3050S, Germany). One section was stained with hematoxylin and eosin (HE) for routine histopathological analysis, and the other was transported to the research laboratory in liquid nitrogen. All breast slices were placed on a slide. Prior to analysis, the frozen section was thawed at 22 °C for 10 min. Multiple spectra were collected from each patient as some tissue samples contained both normal and diseased breast tissues. Characteristics of the patients are summarized in Table 1. Raman spectrometer A confocal Raman system (LabRAM ARAMIS, HORIBA Jobin Yvon, Edison, NJ, USA) with a ∼0.7-μm spatial resolution and a 5-mW/633-nm HeNe laser as excitation source was used for the collection of Raman spectra. The detection of Raman signal was carried out with a Synapse Thermoelectric cooled charge-coupled device (CCD) camera (HORIBA Jobin Yvon, Edison, NJ, USA). Raman scattering light was collected with a ×50 microscope objective lens (0.50 NA, LMPLFLN, Olympus, Japan) that was also used for focusing the excitation laser light. The laser beam focusing on the tissue formed a spot with a diameter of 1.5 μm. The strong Rayleighscattered lights were then blocked by a four-notch filter (HORIBA Jobin Yvon, Edison, NJ, USA). Extended scan spectra with a spectral range of 600–2,000 cm−1 were acquired using an integration time of 60 s and three accumulations. The wave number calibration was set by reference to the 520.7 cm−1 vibrational band of a silicon wafer. These settings were kept constant for all spectral measurements (Scheme 1). Data collection and process At first, tumor locations are located from HE sections and then relocated by Raman spectroscopy in the contiguous frozen sections. Ten to twelve spectra were collected from different locations in each sample to ensure representative sampling and incorporate spot-to-spot variability in the signal. Once all needed spectra were obtained, the SHINs were immediately added to the surface of the frozen sections. Finally, SHINERS spectra were collected.
The use of Au@SiO2 SHINERS for human breast cancer detection Table 1 Clinicopathological parameters Type
Number of patients
Ages (range)
Number of Raman spectra
Number of SHINERS spectra
Normal breast tissues (NB) Fibroadenoma (FD) Atypical ductal hyperplasia (ADH) Ductal carcinoma in situ (DCIS) Infiltrating ductal carcinoma (IDC) Total
9 17 7 8 15 56
40.3 (27–59) 37.1 (19–56) 44.8 (39–56) 43.0 (34–51) 51.6 (31–74) 44.1 (19–74)
103 144 89 123 160 619
92 181 85 139 157 654
A total of 619 Raman and 654 SHINERS spectra were obtained from all different tissues. The detailed information is presented in Table 1. All the spectra were dealt with baseline corrected by fitting and subtracting a third-order polynomial using the NGSLabSpec software (HORIBA Jobin Yvon, Edison, NJ, USA). The spectra were then smoothed with a 15point adjacent averaging. Statistical analysis was done using PASW Statistics Base 18 statistical software (SPSS Inc., Chicago, USA), using Kruskal Wallis test for the difference of the enhancement ratios of characteristic bands between five tissues and using Mann–Whitney test for the difference of characteristic band intensities before and after enhancement. p
