Toward Improving the Proteomic Analysis of Formalin-fixed, Paraffin-embedded Tissue

Carol B Fowler; Timothy J O'Leary; Jeffrey T Mason


Expert Rev Proteomics. 2013;10(4):389-400. 

In This Article

Effect of Analytical Factors on Proteomic Profiling

Incomplete solubilization of protein from FFPE tissues can preclude accurate protein quantitative analysis while incomplete reversal of formaldehyde adducts and cross-links can result in either a failure to identify constituent proteins in the extract or an excessively high false protein identification rate. Formaldehyde-induced protein modifications severely complicate the calculation of the protein content of FFPE tissues,[28] thus making the determination of protein extraction efficiency (protein recovery) equally difficult. To circumvent this problem, the authors' research group studied 'tissue surrogates' containing from one to several proteins.[32–34] A tissue surrogate is a concentrated solution of proteins in formalin that gel upon fixation to yield an opaque plug with sufficient physical integrity to be processed using standard histological methods. This model system enables the rapid evaluation of tissue extraction protocols and to more easily identify formaldehyde-induced protein modifications and their reversal. The authors evaluated a wide range of protein extraction solutions and conditions reported in the literature and found protein extraction efficiencies ranging from 5 to 90%.[32] The results demonstrated that heat, a detergent and a protein denaturant were required for efficient extraction of proteins from FFPE tissues. The role of various charged and neutral detergents on extraction efficiency have been studied by several groups.[32,34–37] The detergents examined included CHAPS, β-octylglucoside, Tween-20, Triton X-100 and SDS among others. SDS was found to be the most effective detergent in promoting the solubilization of proteins from FFPE tissues when included as a component of the antigen retrieval buffer at a level of 2–4% (w/v). SDS, which serves a dual role as detergent and protein denaturant, was also found to be the most effective detergent for promoting protein cross-link reversal. A modification of the extraction buffer and conditions described by Shi et al. yielded the best results, with protein recoveries of 80–90%. This extraction method involves heating the FFPE tissue section at 100°C for 20 min followed by incubation at 60°C for 2 h using a recovery buffer consisting of 20 mM Tris-HCl, 2% (w/v) SDS, 200 mM glycine, pH 4–9.[2] Subsequent studies where the protein extraction was performed under a pressure of 40,000 psi resulted in 100% protein recovery.[33] Extraction buffers reported in the literature have included many additional additives, including alternate protein denaturants, detergents and buffering compounds in addition to reducing agents, protease inhibitors and organic solvents.[38] Although the role of many of these additional additives remains unclear, one such additive, trifluoroethanol, does appear to improve the recovery of membrane proteins.[39] The role of pH in optimizing protein extraction efficiency has not been clearly established. Studies based upon IHC staining suggest that epitope recovery depends upon each protein's isoelectric point (pI).[40] Most proteomic studies, however, have found that neutral to alkaline pH values are optimal for protein extraction.[37]

Heat is required to provide the energy for the reversal of formaldehyde-induced protein adducts and cross-links. However, the high temperature used for protein recovery can cause pH-dependent protein degradation. At acidic pH, deamination of glutamine and asparagine residues and peptide bond hydrolysis on the carboxylate side of aspartic acid residues are observed.[41] At neutral pH, deamination of glutamine and asparagine residues and thiol-catalyzed disulfide exchange can occur. At alkaline pH, β-elimination of cysteine residues can occur by heterolytic cleavage of the disulfide bond to form dehydroalanine and thiocysteine residues. Dehydroalanine can then react with the €-amino groups of lysine residues to form a lysino-alanine cross-link.[42] Many of these thermally induced pH-dependent modifications can be included in the proteomic database used for protein identification. However, deamination of glutamine and asparagine yields glutamate and aspartate, respectively; changes that would not be detected by a peptide database search.

Incomplete solubilization of protein from FFPE tissue can lead to extraction bias such that the composition of the extract does not represent the proteome of the tissue. The authors' laboratory investigated this possibility using a multi-protein tissue surrogate comprising hen egg-white lysozyme, bovine carbonic anhydrase, bovine ribonuclease A, bovine serum albumin and equine myoglobin (55:15:15:10:5 wt%).[34] When the tissue surrogate was extracted in Tris buffer, pH 4 or 8 with 2% (w/v) SDS, according to the method of Shi et al., the protein extraction efficiency was only 26%.[2] When heated at atmospheric pressure, there was an extraction bias in which three of the tissue surrogate proteins were not detected and the remaining two proteins were significantly underrepresented relative to their percentage in the surrogate. In addition, the false protein identification rate was 42% for the tissue surrogate extracted at pH 4. No tissue surrogate proteins were detected when extracted in pH 8 buffer. By contrast, when the extraction was carried out under heat and a pressure of 40,000 psi the extraction bias was significantly reduced along with the false identification rate (7.8 and 5.6% for the pH 4 and 8 buffers, respectively). These results suggest that standardized protein extraction protocols that ensure the complete solubilization of FFPE tissue is one of the key unmet needs in the proteomic analysis of FFPE tissues.[8]