Protein profile study of Pap smear and tissue of cervix by High Performance Liquid Chromatography- Laser Induced Fluorescence Sujatha1, Lavanya Rai2, Pratap Kumar2,B.R.Krishnanand3, K K Mahato1, Sajan D.George1, V. B Kartha1 and Santhosh C1 1 Centre for Laser Spectroscopy, MAHE Life Sciences Center, 2 Department of OBG, 3 Department of Pathology, Kasturba Medical College, MAHE, Manipal, India.
ABSTRACT HPLC combined with laser induced fluorescence provides a very sensitive method for the separation and identification of the many proteins present in clinical samples. Protein profiles of clinical samples like Pap smear and tissue samples, from subjects with cervical cancer and normal volunteers, were recorded using HPLC-LIF. The protein profiles were analyzed by Principal Component Analysis (PCA). The profiles were characterized by parameters like scores of the factors, sum of squared residuals, and Mahalanobis Distance, derived from PCA. Parameters of each sample were compared with those of a standard set and Match/ No Match results were generated. Good discrimination between normal and malignant samples was achieved with high sensitivity and specificity. Key words: HPLC-LIF, Pap smear, Biopsy tissue, Cancer of the cervix and PCA 1. INTRODUCTION Cervical cancer is one of the most common neoplastic diseases affecting women, with a combined worldwide incidence of almost half a million new cases annually, second only to breast cancer 1. Cervical cancer is the most common cancer of the female genital tract in India, with approximately 100,000 new cases occurring each year. This accounts for almost 20% of all new cases diagnosed in the world annually 2. The incidence is higher in rural areas, where prevention and screening programs are not widely established 3, 4. The Human Papilloma Virus (HPV) has been found to be the primary risk factor for cervical cancer and plays a central role in cervical carcinogenesis 5. Early detection provides the most effective method for successful therapy in cancer. At present, the Pap smear technique is the standard screening tool for this purpose in cervical cancer. This test looks for the presence of irregular cells that could become cancerous. A regular Pap smear check provides an opportunity to detect pre-cancerous cells in the cervix. The introduction of this method has been associated with a 70% decrease of cervical cancer deaths. A positive Pap smear test is followed by Colposcopy and biopsy for pathological examination. Pap smear, and biopsy tissue examination depend on morphological studies, which need experienced personals. False negative results are frequent due to limited experience and fatigue factor associated with microscopy screening. Even screened populations judged to be safe, often continue to have a significant incident of cervical cancer. Specificity of these methods are low, and hence they generally require repetitive sampling. The abnormal Papanicolaou smear cases are usually evaluated with colposcopy, which is a more accurate diagnostic method, but require highly skilled personnel and is also more expensive 6,7. In recent years, laser fluorescence spectroscopy, due to its high sensitivity to biochemical changes, has emerged as an attractive alternate technique for early diagnosis of cancer 8-16. Laser induced fluorescence spectroscopy is an optical technique which has the capability to quickly probe the biochemical and morphological changes that occur as tissue becomes neoplastic. The altered biochemical and morphological state of the neoplastic tissue is reflected in the spectral characteristics of the fluorescence. This spectral information can be correlated to tissue histo-pathology, the current “gold standard” technique. The mathematical data obtained will be analyzed by software, potentially enabling automated, fast and accurate pre-cancer detection in the hands of non-experts.
Imaging, Manipulation, and Analysis of Biomolecules, Cells, and Tissues V, edited by Daniel L. Farkas, Robert C. Leif, Dan V. Nicolau, Proc. of SPIE Vol. 6441, 64410B, (2007) · 1605-7422/07/$18 · doi: 10.1117/12.699711 Proc. of SPIE Vol. 6441 64410B-1
High Performance Liquid Chromatography with Laser Induced Fluorescence is a very sensitive technique for the study of changes in the protein profiles of the clinical samples when they change from normal to malignant condition 17-19. The technique can detect the proteins even when they are present in femto-molar concentration. In the present work protein profiles of clinical samples were recorded by HPLC-LIF. The profiles were studied by multivariate analytical method, Principle Component Analysis (PCA). The discrimination of the individual classes of the samples was carried out by Scores of factors, Sum of squared residuals, and Mahalanobis Distance. Match /No Match results, giving high sensitivity and specificity, were obtained when the individual protein profiles were compared to calibration sets of pathologically certified standard samples.
2. INSTRUMENTATION 2.1. HPLC LIF set up: An HP 1100 (Hewlett Packard, USA) gradient HPLC system with degasser (G1322A), quaternary pump (G1311A), and manual Injector (Model No.7725, Rheodyne) coupled to a biphenyl narrow bore type reversed phase (Vydac 219TP52) column (2.1X 250mm, 5µm, 300Ao ) was used for separation. The effluent from the column was sent into a capillary flow cell made of a quartz capillary (75µm I.D. 300m O.D.). A frequency doubled Ar+ laser (Innova 90C FreD, Coherent) with 15 mW power at 257.5nm was used for excitation. The fluorescence was collected and focused at the entrance slit of a monochromator (DH 10 SPEX) set at 340nm, and detected by a PMT (Hamamatsu R453) operated at 850 V. The fluorescence was chopped with an EG&G Model 651 chopper at 20Hz and the PMT output was given to a Preamplifier (EG&G Model 5113), followed by Lock-in Amplifier (EG &G Model 7265) interfaced to a PC. The schematic diagram of the HPLC –LIF system is shown in figure 1.
C
D
B
A
257 nm
F1 J
I
G F
E
H
K L
M
A. HPLC, B. Injector, C. Reverse phase liquid chromatography Column, D. Frequency Doubled Argon ion laser, E. Quartz capillary, F. Lens, F1 laser focusing lens G. Chopper, H. .Monochromator, I. PMT, J. Preamplifier, K Lock-in Amplifier, L. Computer, M Effluent collector. Figure 1. The HPLC-LIF setup
Proc. of SPIE Vol. 6441 64410B-2
2.2. Data Analysis All the data (protein profiles of the clinical samples) analysis was carried out using Principle Component Analysis. The PCA runs were done using PLS PLUS/IQ from Galactic Corporation Inc., USA 20. Discriminant analysis was performed. Before any PCA runs all the chromatograms were mathematically preprocessed. Background subtraction of all the chromatograms was carried out by multipoint polynomial fit. The background subtracted chromatograms were then subjected to normalization. To reduce the shift in peak position, from run to run, calibration of all the chromatograms were done along the time scale, using protein peaks common to all samples. In our analysis, for discrimination of normal and malignant clinical samples, we have used three parameters derived from PCA. These are the scores of the factors, sum of squared residuals residuals, and the Mahalanobis distance. The Mahalanobis distance is expressed in units of standard deviation. For discrimination of normal and malignant pap smear, we have employed the Mahalanobis distance (M-distance) and the residual error squared sum as the criteria. The M-distance is given by: D2 = (S test) M –1(S test) 1, Where S test is the vector of scores and sum of squared residuals for a given test sample, and M is given by S′S/(n-1), where S contains the corresponding parameters for the calibration set of n standards.
3. SAMPLES COLLECTION AND PROCESSING 3.1. Sample collection and storage: Normal biopsy tissue samples were obtained from the subjects who underwent hysterectomy and Pap smear samples were collected from healthy volunteers. Biopsy tissue and Pap smear samples of malignant subjects were collected from the Department of Obstetrics and Gynecology, Kasturba Medical College, Manipal, with informed consent. Ethical clearance for the present study was obtained from the Kasturba Medical College Ethical Clearance Committee. The malignant subjects were at different stages of cancer of the cervix. The samples collected were immediately transported to the laboratory. If storage was necessary, samples were snap frozen and stored at -800C in the deep freezer. Exfoliated cells (Pap smear) were washed with saline and pelleted by spinning (3000 rpm for 5 minutes) in a microcentrifuge (Costar mini centrifuge, 10MVSS). The supernatant was discarded and the cell pellet was collected for lysing. Cells were mixed with Tris- EDTA buffer (0.1M 7.4 pH) [2% of the wet weight of the cells] and lysed using a sonicator (Sonic vibra cell model: VC 130PB). Lysed cells were again subjected to centrifugation and 50µl of the supernatant was injected to HPLC. Tissue samples were washed several times with normal saline. Then the samples were weighed and minced with Tris EDTA buffer, 10% the wet weight of the sample. The tissues were then homogenized by using manual homogenizer (T8 blade IKA-WERKE), and centrifuged at 5000rpm for 20 minutes. Supernatant was then passed through the syringe filters to remove cell debris and other particles. 50µl of the collected supernatant was injected in to the HPLC-LIF system. Samples
Number
Age
Pap smear normal
10
32-60
Pap smear malignant
9
37-72
Normal cervical tissue
13
40-60
Malignant cervical tissue
17
37-72
Table 1. Number of samples recorded and analysed
Proc. of SPIE Vol. 6441 64410B-3
3.2. Liquid Chromatography conditions: Water with 0.1 % TFA and Acetonitrile [Merck] with 0.1% TFA (both HPLC grade) are used for gradient runs. Each time a blank was run before the gradient to confirm the stability and residual contamination of the column. 50µl of sample was then injected into the narrow bore biphenyl column fitted with a 20 µl loop. The sample is then eluted with a water – Acetonitrile gradient. The gradient starts with 70% water (0.1%TFA) +30% Acetonitrile (0.1%TFA) and changes to 40% water + 60% Acetonitrile (0.1%TFA) in 60 minutes. The rate of elution was kept at 200 micro liter/minute.
4. RESULTS AND DISCUSSION 4.1. Protein profile of the Pap smear: The protein profile of the exfoliated cells of cervix (Pap smear) of normal subjects shows nearly 15 peaks, out of which 10 are intense (Figure 2). Noticeable differences have been observed when the normal protein profiles of the Pap smears were compared to the malignant. The mean chromatograms of normal and malignant cells are shown in Figure 2. Several intense peaks in case of malignant samples suggest that these peaks correspond to proteins which might have got overexpressed during malignancy.
Intensity
Normal
Malignant
500
1000
1500
2000
2500
3000
3500
Time (seconds) Figure 2 Protein profile of Pap smear from normal and malignant subjects
4.2. PCA analysis of Pap smear protein profiles: The PCA analysis was carried out by using 7 factors, with factor 1 contributing 83% of total variance. The Eigen values and total percentage variance are shown in Table 2 and the corresponding plots in Figure 3. The contributions of these factors to each of the chromatograms give the score for those chromatograms. It can be seen that the scores for factor 1 of one class of the samples, say normal, are very different from that of the malignant (Figure 4a). The scores alone thus seem to give good discrimination between normal and malignant samples.
Proc. of SPIE Vol. 6441 64410B-4
Total % Variance 83.1209279 90.8930327 94.6066716 97.6773735 98.5095904 99.2927685 100
100
11000
T o 94 t a l 88
Total % Variance
Eigenvalue 11592.132 1083.9059 517.90796 428.2433 116.06184 109.22283 98.63125
Eigenvalue
Factor Number 1 2 3 4 5 6 7
8000 5000 2000 -1000
Table2. Eigen values and total percentage variance for seven factors of Pap smear
Ð 1
2.5
4
Factor Factor number
Ð
82
5.5
11
2.5
4
Factor Factor number
5.5
Figure 3. Plot showing Eigen values and total percentage variance for pap smear
The Match/ No Match results were obtained by testing all the samples against the normal standard set. Good discrimination have been achieved by M distance VS Sum of Squared Residuals plot in Figure 4B.
(a)
Score
0 10
20
-0.4 -0.6
30
40
Spectral residual
10
0.2
-0.2 0
(b)
12
0.4
8 6 4 2 0
Sample number
0
10
20
M Distance
Normal Malignant Figure 4. PCA of Pap smear protein profiles. a). Sample number verses scores of factor 1 and b). Discriminate analysis
Proc. of SPIE Vol. 6441 64410B-5
30
4. 3. PCA of Cervical Tissue Protein Profile: The protein profile of the normal biopsy tissue shows more than 30 peaks, 10 of them intense. The region of 1000 to 2400 seconds retention time was found to be very different for the two classes (Figure 5). The full region has been taken for PCA analysis.
Intensity
Normal
Malignant
500
1000
1500
2000
2500
Time (seconds) Figure 5. Protein profile of cervical tissue
The PCA analysis of the tissue chromatograms was done taking 5 factors. The Eigen values and total percent variation for the 5 factors are shown in Table 3 and their plot, in figure 6. The total contribution of the factor 1 to the data set is about 89%, and this was used to discriminate the normal and malignant samples based on score values. The scores of factor 1 plotted against the sample number is shown in the figure 7a. All the normal samples were clearly discriminated from the malignant samples with positive score values. All the malignant samples showed negative score value. The score values were thus quite sufficient to differentiate normal and malignant samples.
Eigen values
Total % Variance
110000
98
1
124104.76
89.1351227
2
11348.431
97.2858475
3
2605.1123
99.1569039
4
712.87999
99.668912
5
460.98094
100
Table 3: Eigen values and total percentage variance for cervical tissue protein profile.
Eigenvalue
80000 50000 20000
10000
Ð
Total % Variance
Factor Number
Ð
95 92 89
1 1.6 2.2 2.8 3.4 4 4.6 Factor number
1 1.6
2.8 3.4 4 4.6
Factor
Factor number
Figure 6. Plot showing Eigen values and total percentage variance for cervical tissue
Proc. of SPIE Vol. 6441 64410B-6
0.6 0.5 0.4 0.3 0.2 0.1 0 0 -0.1 -0.2
8 M. Distance
Score
The Mahalonobis distance was another parameter used to discriminate the samples. All the normal samples were having M distance value below two and all the malignant samples having M distance value above this .The sensitivity and specificity of the samples taking normal standard set was very high (above 90%). The plot of Mahalonobis distance verses spectral residual is shown in Figure 7b.
20
40
60
6 4 2 0
0
10
Sample number
20
30
40
Sample number
Normal Malignant Figure 7. PCA of cervical tissue protein profiles. a). Sample number verses scores of factor 1 and b). Discriminate analysis
5. CONCLUSION The protein profiles of Pap smear and biopsy tissues can be used for discrimination of Normal and Malignant cases of Cervical cancer with high sensitivity and specificity. It is clear from the figures that even a visual analysis is sufficient to discriminate the different classes of samples. The protein profiles show that many new proteins are highly expressed as the tissue becomes malignant. PCA analysis gives very good classification of the normal and malignant samples of biopsy tissues and Pap smears. Multi parametric limit test based on Mahalonobis distance gives good sensitivity and specificity (above 90%) in case of biopsy tissue samples when we match all the samples by taking normal standard set.
ACKNOWLEDGEMENT The work was done under the project “Study of the protein profile of the clinical samples for the early diagnosis of female cancers.” Department of Science and Technology, Government of India. Project No: SR/S2/LOP/05/2003.
Proc. of SPIE Vol. 6441 64410B-7
REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.
17. 18. 19. 20.
Franco EL, Schlecht NF, Saslow D. The epidemiology of cervical cancer. Cancer J. 9,5,348-59. 2003 Shanta V, Krishnamurthi S, Gajalakshmi CK, Swaminathan R, Ravichandran K. Epidemiology of cancer of the cervix: global and national perspective. Indian Med Assoc, 98,2, 49-52, 2000. Ghim SJ, Basu PS, Jenson A. Cervical cancer: etiology, pathogenesis, treatment and future vaccines. Asian Pac J Cancer Prev;3:207,2002 Radhakrishnan Pillai M, Sreevidya S, Pollock BH, Jayaprakash PG, Herman B. Human Papilloma type 16 E6 and E7 gene variation in Indian cervical cancer. Gynecol. Oncol; 87:268, 2002. Duttagupta C, Sengupta S, Roy M et al. Oncogenic Human Papilloma Virus infection and uterine cervical cancer: a screening strategy in the perspective of rural India. European J of cancer Prev,11, 4, 2002. J. Monsonego et al, Human papillomavirus testing improves the accuracy of colposcopy in detection of cervical intraepithelial neoplasia, Int. J Gynecol. Cancer, 16, 591–598, 2006. Charvet I, Meda P, Genet M et al, Optical dignosis of cervical dysplasia, Bull cancer,91,1, 45-53, 2004. Ramanujam, N, Mitchell, M. F, Mahadevan, A, Thomsen, S, Malpica, A, Wright, T,Atkinson, N, and RichardsKortum, R. R. Photochemistry and Photobiology, 64, 4, 720, 1996. Carrie Brookner, Urs Utzinger et.al. Effects of biographical variables on cervical fluorescence emission spectra, Journal of Biomedical Optics 8,3, 479–483,2003 Ann Gillenwater, Rhonda Jacob, Rebecca Richards-Kortum, Fluorescence Spectroscopy: A technique with potential to improve the early detection of Aerodigestive tract neoplasia, Head & Neck , 556-562,1998. Erin M. Gill, Anais Malpica, Raphael E. Alford, Audrey R. Nath, Michele Follen, Rebecca R. RichardsKortum, and Nirmala Ramanujam, Relationship Between Collagen Autofluorescence of the Human Cervix and Menopausal Status, Photochemistry and Photobiology, 77,6, 653–658, 2003. N. Ramanujam, Fluorescence spectroscopy of neoplastic and non-neoplastic tissues, Neoplasia 2, 89-117, 2000. G. A. Wagnieres, W. M. Star and B. C. Wilson, In vivo fluorescence spectroscopy and imaging for oncological applications, Photochem. Photobiol. 68, 603-632, 1998. Sung K. Chang, Michele Follen, Anais Malpica, Urs Utzinger, Gregg Staerkel, Dennis Cox, E. Neely Atkinson, Calum MacAulay, and Rebecca Richards-Kortum, Optimal Excitation Wavelengths for Discrimination of cervical neoplasia, IEEE Transactions on Biomedical Engineering, Vol. 49, No. 10, 2002. Kortum, Optimal Excitation Wavelengths for Discrimination of cervical neoplasia, IEEE Transactions on Biomedical Engineering, Vol. 49, No. 10, 2002. Urs Utzinger, E. Vanessa Trujillo, E. Neely Atkinson, Michele F. Mitchell, Scott B. Cantor, and Rebecca Richards-Kortum, Performance estimation of diagnostic tests for cervical precancer based on fluorescence spectroscopy: Effects of tissue type, sample size, population, and signal-to-noise ratio IEEE Transactions on Biomedical Engineering, Vol. 46, No. 11, 1293-1303, 1999. Sujatha, Lavanya Rai, B.R.Krishnanand, K K Mahato, V. B Kartha and Santhosh C, Protein profile study of the cervical cancer using HPLC- LIF, Proc. of SPIE Vol. 6092,206-215,2006. Sameer Kumar Singh, Remila L. Martis, Sujatha, Rani A. Bhat, Pralhad Kushtagi, Lavanya Rai,V.B. Kartha, C. Santhosh, Protein profile study of clinical samples of ovarian cancer using High Performance Liquid Chromatography-Laser Induced Fluorescence (HPLC-LIF), Proc. of SPIE Vol. 6092, 149-157,2006. Venkatakrishna K, Kartha VB, Keerthilatha MP et al, HPLC- LIF for early detection of oral cancer, Current Science, 84, 1, 2003. PLS PLUS/IQ software from Galactic corporation Inc
Proc. of SPIE Vol. 6441 64410B-8