1.      Introduction

Tissue heterogeneity is the major problem in the analysis of cancer tissues and presents a main difficulty in identifying specific cancer biomarkers. Many techniques have been developed to overcome this problem and enrich for the cells of interest. One of these technique is Laser Capture Microdissection (LCM) [1,2]. The technique consists of 3 steps:(I) preparation of tissue section, (II) tissue sections staining with a suitable stain, and(III) manual or automated dissection using system such Arcturus^XT(Molecular device, CA) and Leica (Leica, Milton Keynes, UK).

In cancer research, LCM in conjunction with many other techniques such as genomic and proteomic has been used to determine the gene or protein expression profiles of tumor cells without the effect of normal or surrounding stroma cells. In the field of proteomic, it has been coupled with reverse phase array, antibody array surface enhanced laser desorption/ionization-time of flight mass spectrometry, and two-dimensional gel electrophoresis (2D-PAGE). 2D-PAGE allows hundreds to thousands of proteins to be analyzed simultaneously. These proteins can be isolated in pure form from the resultant spots which later can be quantified and further analyzed by mass spectrometry.

Although LCM in combination with 2D-PAGE analysis of cancer tissue is promising technology especially in cancer biomarkers research, there are certain limitations to this technology. Firstly, collectingthe adequate number of cells to obtain enough protein for 2D-PAGE analysis is time-consuming. Secondly, there is a limited sensitivity of different stains to visualize the protein spots in the gel.

From our experience, we found that even though enough cells were collected and there was enough protein for 2D-PAGE analysis; the number of spots in the gel was less than that obtained with pure E. coli proteins when using Coomassie blue stain (the commonly used stain in laboratory due to its affordable cost compared to other stain). However, these differences were not seen with SYPRO^® Ruby stain. Although, SYPRO Ruby is a very sensitive fluorescent stain which has many advantages over other stains [3,4] and is relatively expensive for daily laboratory use [3]. These advantages could not exclusively explain the huge differences seen between the two stains. Thus, in this study we have examined the effect of Haematoxylin (the commonly used histological stain for tissue preparation for LCM)on second stains such as Coomassie blue and SYPRO Ruby, which are used to visualize protein spots in 2D-PAGE.

2.      Materials and Methods

2.1.  Study Design

 In this study, we have used pancreatic cancer tissue to examine the effect of Haematoxylin stain on other stains as mentioned above. Also,1D-PAGE was used instead of 2D-PAGE analysis, due to the ease of determining the bands intensities.Tissues from pancreatic cancer without sectioning or staining were used as control and compared to tissues sections that were stained with Haematoxylin or unstained.

2.2.  Tissue and Protein Preparation

(a)     Protein preparation from whole pancreatic tissues

Pancreatic cancer tissues were homogenized using lysis buffer (7Murea, 2M thiourea, 1% MEGA-10,1% Octyl-β-D-glucopyranoside, 40mM Tris, 50mMDTT, 2mMtributylphosphine). The homogenate was centrifuged (1000rpm for 5 min);and protein concentration in the supernatant was determined using Amido black protein assay (sample # 1).

(b)     Protein preparation from stained and unstained tissue section

Sections(8µM thick) were prepared from pancreatic tissues using a cryostat and mounted in a clean glass slides. Slides were stainedwith hematoxylin (Sample # 2) then scraped or scraped without staining (Sample # 3) in Eppendorf tubes for protein preparation. The tissue sections for sample #2were stained with Haematoxylin before scraping as follow: 100% acetone- (15s), Haematoxylin(15s), protease inhibitor containing water (15s), 75% ethanol (15s), 95% ethanol (15s), absolute alcohol(15s), Xylene(15s). All solutions were prepared in water containing protease inhibitor. Both scraped stained and unstained tissue sections weremixed in lysing buffer, homogenized and vortexed at 37^oC for 30min.The protein concentrations in the supernatants were determined with Amido black assay.

2.3.  Gel Electrophoresis

The proteins extracted from whole, not sectioned or stained tissues (sample # 1); sectioned and stained with hematoxylin tissues (sample # 2); or section but not stained tissues (sample # 3) were separated intoa 10% SDS-PAGE. Twenty micrograms of the BovineSerum Albumin (Pierce Chemical Company, Rockford, IL, USA) were used as the standard marker. The gels were run at 200V and stained with CBB R-250 for 2hrs and then de-stained for 4-6 hrs or were stained with SYPRO^® Ruby. They were visualized through Kodak-Digital Science Image Station 440-CF (Eastman Kodak Company, Scientific Imaging Systems, Rochester, NY, USA) and the intensities of some of the clear bands were measured (by selecting the band area).A graph was drawn by using the concentrations of whole tissues extracted protein series as the x-axis and the bands intensities as the y-axis.The bands’ intensities obtained from Haematoxylin stained and unstained gels were plotted in the graph for thecomparison [5].

3.      Results

3.1.  Effect of Hematoxylin in Protein Detection Using CBB R-250

The band’s intensities of the Haematoxylin stained tissue and the unstained tissue sections with respect to the whole tissue standard series are shown in Figure 1.The proteins were stained with CBB R-250.Thirty micrograms of Haematoxylin stained tissue gave lower intensity than the 30 µg of the whole tissue. The calibration curve for the intensity of the bands of the protein from whole tissuevs. concentration series of the proteins from whole tissue is shown in Figure 2. The intensities of the 30µg of proteins obtained from Haematoxylin stained and unstained tissues (10µg) were compared with the calibration curve. Proteins from the Haematoxylin stained tissue (30 µg) gave a lower intensity compared to the unstained and whole tissues.

3.2.  Effect of Haematoxylin in Protein Detection Using SYPRO Ruby

Figure 3 shows a comparison between theintensities of the proteins from Haematoxylin stained tissues(lane 7) and the proteins from unstained tissues(lane 8) with respect to proteins from whole tissue standard series (lane 1-6). There was no significant difference between the band’s intensities of the lanes 5, 7 and 8 (each lane was loaded with 30 µg of protein). The calibration curve of the intensity of the protein bandsvs. concentrations of the whole homogenized tissue when stained with Syproruby is shown in Figure 4.The Intensities of the proteins (30 µg) from the homogenized Haematoxylin stained and the homogenized unstained tissues were compared with the calibration curve of the homogenized whole tissue concentration series. There was no significant difference between the intensities of the whole tissue homogenized proteins, homogenized Haematoxylin stained tissue and the homogenized unstained tissue sections [6].

Table 1 gives the mean value of the ratiosbetween the band’s intensities taken from the two gels (CBB R-250 and Sypro ruby). This value was calculated to determine the effect of Haematoxylin on the band intensities of the proteins with respect to CBB R-250 and syproruby. The mean ratio between the intensities of the protein bands obtained from the whole tissue and the Haematoxylin stained tissue was close to 4. The ratio between the homogenized unstained and the homogenized Haematoxylin tissue was 2.55. But when syproruby was used, the mean ratio for both was around 1. The intensity ratio between the unstained/whole tissue with respect to the CBB R-250 and syproruby was around 1.

4.      Discussion

The data from this study showed that Haematoxylin stain affects proteins when it was visualized in gel using CBB R-250 stain.However, this effect was not seen with SYPRO Ruby. We hypothesized that the observed effect of Haematoxylin may be due to structure and reactivity of these stains towards [7,6].

Pure Haematoxylinis a colorless solution [8] and it is activated to Haematein by a mordant. The Haematoxylin-mordant complexes are positively charged and behave as a cationic dye at low pH. The most common mordents are aluminum, iron or tungsten.The cationic stain-mordant complexes are attracted to negatively charged sites in the staining process [9]. Under suitable conditions, the complexes will stain cationic proteins such as histones associated with nucleicacids [10-13] and nucleoproteins that carry lysine and arginine residues [12,14-16]

Anions of CBB formed in acidic conditions combines with the protonated amino groups of the proteins by electrostatic attraction (arginine, lysine, tyrosine, and histidine residues) [17,18]. Sypro ruby dye interacts strongly with lysine, arginine, and histidine residues. It closely resembles CBBstaining through basic amino acids side chains [19]. Although the binding of sypro ruby and CBB are the same, the sypro ruby has a significant sensitivity towards protein staining against the CBB because of its structure. Syproruby has a ruthenium atom. This can react with amino acids such as cysteine through ligand exchange. This atom is also able to form η6-complexes with derivatives oftryptophan amino acids (Figure 1)[18].

When staining, CBB R-250 binds through arginine, lysine, tyrosine, and histidine residues in proteins [17,18]. Meanwhile,  sypro ruby binds through the same amino acids[19].When staining with Haematoxylin,it binds to nucleoproteins with lysine, arginine residues or other parts having suitably spaced carboxyl and hydroxyl groups[11,16]. Haematoxylin masked these binding sites; hence, the results presented in this paper show that the band’s intensities for the Haematoxylin stained homogenized tissue and the unstained homogenized tissue are not similar (Figures 2&3). According to Table 1,the ratio between the whole tissue homogenized proteins and the homogenized Haematoxylin stained tissue band’s intensitieswas around 4. The ratio betweenthe protein band’s intensities of the homogenized unstained tissue sections and the homogenized Haematoxylin stained tissue sections was 2.55. When detecting with CBB R-250, the Haematoxylin stained tissue had a low intensity compared to the whole homogenized tissue and the homogenized unstained tissue. However, syproruby had no effect on the Haematoxylin stain tissue. The ratio between the intensities of the protein bands of the homogenized whole tissue and the homogenized Haematoxylin stained tissue was close to 1. The ratio between the homogenized unstained tissue sections and the Haematoxylin stain tissue was also close to 1 (Table 1). This may be due to the special type of interaction between the syproruby and the amino acids through metal-ligand interaction. These interactions did not interfere with the Haematoxylin staining. Thus, the effect of Haematoxylin on syproruby detection was much lower than the CBB detection. Finally,it should be emphasized that when the tissues were stained with Haematoxylin before running the gels, it was better to use sypro ruby to visualize the proteins instead of CBB because it allowedthe visualization of more proteins and trace amounts of proteins in 1-D.

The work from Craven et al. and De Souza et al. showed that there was no effect on the protein recovery when stained with Haematoxylin & Eosin. Nevertheless, there was an effect on the proteins’ quality [1,20]. This finding contradicts our results and may be due to several reasons. In our work we focused on the proteins’ intensity instead of the proteins’ recovery. Craven et al. and De Souza et al. also used silver staining instead of sypro ruby to detect the proteins and the stain differentiation was longer(Craven et al. - 70% ethanol for 30s, 100% ethanol for 1min, xylene for 2 X 5 min), (De Souza- 70% ethanol for 30s, 100% ethanol for 1min, xylene for 5min). In our protocol, we used a shorter stain differentiation time (75% ethanol 15s, 95% ethanol 15s, absolute alcohol 15s, xylene 15s). It is better to use a short differentiation time in LCM for the visualization of proteins.

Mouledous et al. work indicates that the fixing of the tissue had no effect on the recovery of protein. However, syproruby affects Haematoxylin and eosin. This contradicts our results, which indicates that Syproruby does not influence the intensity of the proteins stained with Hematoxylin. Although both works concluded that there was no effect on the fixation, they haddifferent fixation procedures [21].

Ahram et al obtained results like ours showing that Haematoxylin did not influence the qualitative and quantitative determination of proteins. Eosin however had an effect on the proteins’ separation. The detection was done by aminoacyl silver staining [22].Our work is the first work looking at the effect of Haematoxylin on the detection of syproruby& CBB R-250.

In the future, it would be appropriate to test the effect of other types of tissue staining methods like Haematoxylin, eosin, methylene blue, Wright-Giemsa and toluidine blue on the detection methods of syproruby and CBB R-250. Becausesyprorubyis expensive it using other types of staining methods instead of Haematoxylin such as CBB R-250may give better results.This will be more economically advantage. Furthermore, it is important to assess the effect of these stains on the recovery and quantity of proteins.

5.      Conclusion

Tumor tissues staining for LCM with Haematoxylin resulted in the proteins’ bands intensity drop when later visualized through CBB R-250 in gels. However, there was no intensity drop in the proteins’ bands when visualized through SYPRO Ruby. SYPRO Ruby should therefore be used in conjunction with Haematoxylin. This is more important when using LCM because the amount of proteins that could be visualized in 1-D and 2-D gel will be higher, especially in the presence of trace amounts of proteins in the sample. Therefore, the study recommends assessing the effect of tissue stains on the recovery and quantity of proteins for down stream applications such as protein extraction for 2-D analysis.

6.      Acknowledgements: NIH Grant #5r03CA119310-2 supported this work.

Figure 1: 1-DE SDS-PAGE of proteins extracted by different protocols using CBB R-250 as the protein staining. Lane 1: 20µg from whole tissue, lane 2: 30 µg from whole tissue, lane 3: 60 µg from whole tissue, lane 4: 20 µg from Haematoxlin stained tissue slide sections, lane 5: 20 µg from unstained tissue slide sections: lane 6: 20 µg of standard BSA.

Figure 2: Comparison of the band intensities of proteins from whole tissue, stained and unstained Haematoxlin tissue sections using CBB R-250. 1-DE SDS-PAGE of proteins extracted by different protocols using CBB R-250 as the protein staining. Lane 1: 10 µg of protein from unstained tissue sections, lane 2: 30 µg of protein from haemotoxyline stained tissue sections, lane 3: 60 µg of protein from whole tissue, lane 4: 30 µg of protein from whole tissue, lane 5: 20 µg of protein from whole tissue, lane 6: 10 µg of protein from whole tissue, lane 7: 5 µg of protein from whole tissue, lane 8: 0 µg protein.

Figure 3: Calibration curve of the intensity of a particular band for the homogenized whole tissue series vs. the loaded whole tissue proteins to the gel. ■-intensity for the 10 µg unstained tissue, ▲- intensity for the 30 µg Haematoxlin stained tissue, (stained with CBB R -250.).

Figure 4: Comparison of the band intensities of proteins from whole tissue, stained and unstained Haematoxlin tissue sections using syproruby. 1-DE of the proteins obtained from homogenized tissue, Haematoxlin stained tissue and unstained tissue sections using syproruby as the staining method. Lane 1: 0 µg of protein, lane 2: 5 µg of protein from whole tissue, lane 3: 10 µg of protein from whole tissue, lane 4: 20µg of protein from whole tissue, lane 5: 30µgof protein from whole tissue, lane 6: 60µg of protein from whole tissue, lane 7: 30µg of protein from Haematoxlin stained tissue sections, lane 8: 10µgof protein from unstained tissue sections

Table 1: Intensity ratio of the protein bands using CBB R -250 and syproruby for different tissue sections.

1. Craven RA1, Totty N, Harnden P, Selby PJ, Banks RE (2002) Laser capture microdissection and two-dimensional polyacrylamide gel electrophoresis: evaluation of tissue preparation and sample limitations. Am J Pathol 160: 815-822.
2. Espina V, Wulfkuhle JD, Calvert VS, VanMeter A, Zhou W, et al. (2006) Laser-capture microdissection. Nat Protoc1: 586-603.
3. Berggren K1, Chernokalskaya E, Steinberg TH, Kemper C, Lopez MF, et al. (2000) Background-free, high sensitivity staining of proteins in one- and two-dimensional sodium dodecyl sulfate-polyacrylamide gels using a luminescent ruthenium complex. Electrophoresis 21: 2509-2021.
4. Lopez MF, Berggren K, Chernokalskaya E, Lazarev A, Robinson M, et al. (2000) Gel Stain with respect to protein detection in two-dimensional gels and identification by peptide mass profiling. Electrophoresis 21: 3673-3683.
5. Wirth PJ1, Romano A (1995) Staining methods in gel electrophoresis, including the use of multiple detection methods. J Chromatogr A698: 123-143.
6. Bettinger C1, Zimmermann HW (1991) New investigations on hematoxylin, hematein, and hematein-aluminium complexes. II. Hematein-aluminium complexes and hemalum staining. Histochemistry96:215-228.
7. Arshid FM, Connelly RF, Desai JN, Fulton RG, Giles CH, et al. (1954) A Study of Certain Natural Dyes II-The Structure of the Metallic Lakes of the Brazilwood and Logwood Colouring Matters. Journal of the Society of Dyers and Colourists 70: 402-412.
8. Palmer RW, Lillie RD (1965) Spectroscopic and staining studies of the ripening and overripening of aluminum hematoxylins. Histochemie 5: 44-54.
9. Marshall PN, Horobin RW (1973) Measurements of the affinities of basic and "mordant" dyes for various tissue substrates. Histochemie 36: 303-312.
10. Baker JR (1962) Experiments on the action of mordants. 2. Aluminium-hematein. Q. J Microsc Sci 103: 493-517.
11. Puchtler H, Meloan SN, Waldrop FS (1986) Application of current chemical concepts to metal-hematein and -brazilein stains. Histochemistry 85: 353-364.
12. Hrapchak BB (1976) Selective staining with hematoxylin, applications and theory: a review. Am J Med Technol 42: 371-379.
13. Prentø P, Schulte E (1991) Staining Involving Metal Complex Dyes. In: Lyon H. (eds) Theory and Strategy in Histochemistry. Springer, Berlin, Heidelberg.
14. Baker JR (1960) Experiments on the action of mordants. 1. ‘Single-bath’ mordant dyeing. Q. J. Microsc. Sci101: 255-272.
15. Scott JE, Willett IH (1966) Binding of cationic dyes to nucleic acids and their biological polyanions. Nature 209: 985-987.
16. Puchtler H, Sweat F (1964) On the mechanism of sequence iron-hematein stains. Histochemie 4: 197-208.
17. Syrový I1, Hodný Z (1991) Staining and quantification of proteins separated by polyacrylamide gel electrophoresis. J Chromatogr 569: 175-196.
18. Ball MS, Karuso P (2009) Mass spectral compatibility of four proteomics stains. J Proteome Res 6: 4313-4120.
19. Steinberg TH, Chernokalskaya E, Berggren K, Lopez MF, Diwu Z, et al. (2000) Ultrasensitive fluorescence protein detection in isoelectric focusing gels using a ruthenium metal chelate stain. Electrophoresis 21: 486-496.
20. De Souza AI, McGregor E, Dunn MJ, Rose ML (2004) Preparation of human heart for laser microdissection and proteomics. Proteomics 4: 578-586.
21. Mouledous L, Hunt S, Harcourt R, Harry JL, Williams KL, et al. (2002) Lack of compatibility of histological staining methods with proteomic analysis of laser-capture microdissected brain samples. J Biomol Tech13: 258-264.
22. Ahram M, Flaig MJ, Gillespie JW, Duray PH, Linehan WM, et al. (2003) Evaluation of ethanol-fixed, paraffin-embedded tissues for proteomic applications. Proteomics 3: 413-421