Histomathematical analysis of clinical specimens: challenges and progress.

Proteomic analysis of clinical tissue specimens is a difficult undertaking. Described here is a multiplex study of protein expression levels in histological sections of human prostate that addresses many of the associated challenges. Whole-mount sections from 10 prostatectomy specimens were studied using 15 antibodies, immunohistochemical staining, digital imaging, and mathematical analysis of the data sets. The approach was successful in stratifying cell lineages present in the samples based on proteomic patterns, including differentiating normal epithelium from cancer. This strategy likely will be a useful method for extending the number of proteins that can be analyzed in clinical cancer specimens using currently available laboratory techniques.     

Studying multiple protein profiles over time to assess biomarker validity

Protein profile analysis is increasingly used for identification of disease biomarkers. The approaches vary from surface-enhanced laser desorption/ionization to protein arrays. Newer platforms are constantly being developed. Almost all are based on matrix-assisted laser desorption/ionization time-of-flight mass spectrometry and are often coupled with sophisticated software tools. Protein profiling has been applied to a variety of samples including plasma, urine, cerebrospinal fluid, saliva and solid tissue. This article focuses on those instances where it is possible to obtain sequential samples from the same individual. In the authors use of a profile method, many protein changes with highly significant correlations to disease have been found. The main challenge lies in the validation of the marker to demonstrate its adequacy for use in the clinical setting. The latter requires a methodology that is robust and amenable to high-throughput. One problem is that interindividual variability among the healthy population can mask major changes that occur on an intraindividual basis. Often, a large change for an individual may remain within the range of healthy individuals. Thus, one strategy to optimize biomarker discovery is to examine serial samples from a given individual, where a disease biomarker is established by comparison with the individual's own baseline sample. The focus of this review is to illustrate the principle and value of serial protein profiling using a rapid protein extraction method.     

Paraneoplastic antigen Ma2 autoantibodies as specific blood biomarkers for detection of early recurrence of small intestine neuroendocrine tumors

Background

Small intestine neuroendocrine tumors (SI-NETs) belong to a rare group of cancers. Most patients have developed metastatic disease at the time of diagnosis, for which there is currently no cure. The delay in diagnosis is a major issue in the clinical management of the patients and new markers are urgently needed. We have previously identified paraneoplastic antigen Ma2 (PNMA2) as a novel SI-NET tissue biomarker. Therefore, we evaluated whether Ma2 autoantibodies detection in the blood stream is useful for the clinical diagnosis and recurrence of SI-NETs.

Methodology/Principal Findings

A novel indirect ELISA was set up to detect Ma2 autoantibodies in blood samples of patients with SI-NET at different stages of disease. The analysis was extended to include typical and atypical lung carcinoids (TLC and ALC), to evaluate whether Ma2 autoantibodies in the blood stream become a general biomarker for NETs. In total, 124 blood samples of SI-NET patients at different stages of disease were included in the study. The novel Ma2 autoantibody ELISA showed high sensitivity, specificity and accuracy with ROC curve analysis underlying an area between 0.734 and 0.816. Ma2 autoantibodies in the blood from SI-NET patients were verified by western blot and sequential immunoprecipitation. Serum antibodies of patients stain Ma2 in the tumor tissue and neurons. We observed that SI-NET patients expressing Ma2 autoantibody levels below the cutoff had a longer progression and recurrence-free survival compared to those with higher titer. We also detected higher levels of Ma2 autoantibodies in blood samples from TLC and ALC patients than from healthy controls, as previously shown in small cell lung carcinoma samples.

Conclusion

Here we show that high Ma2 autoantibody titer in the blood of SI-NET patients is a sensitive and specific biomarker, superior to chromogranin A (CgA) for the risk of recurrence after radical operation of these tumors.     

Studying multiple protein profiles over time to assess biomarker validity

Protein profile analysis is increasingly used for identification of disease biomarkers. The approaches vary from surface-enhanced laser desorption/ionization to protein arrays. Newer platforms are constantly being developed. Almost all are based on matrix-assisted laser desorption/ionization time-of-flight mass spectrometry and are often coupled with sophisticated software tools. Protein profiling has been applied to a variety of samples including plasma, urine, cerebrospinal fluid, saliva and solid tissue. This article focuses on those instances where it is possible to obtain sequential samples from the same individual. In the authors use of a profile method, many protein changes with highly significant correlations to disease have been found. The main challenge lies in the validation of the marker to demonstrate its adequacy for use in the clinical setting. The latter requires a methodology that is robust and amenable to high-throughput. One problem is that interindividual variability among the healthy population can mask major changes that occur on an intraindividual basis. Often, a large change for an individual may remain within the range of healthy individuals. Thus, one strategy to optimize biomarker discovery is to examine serial samples from a given individual, where a disease biomarker is established by comparison with the individual’s own baseline sample. The focus of this review is to illustrate the principle and value of serial protein profiling using a rapid protein extraction method.     

Selection of cell‐type specific antibodies on tissue‐sections using phage display

With the advent of modern technologies enabling single cell analysis, it has become clear that small sub‐populations of cells or even single cells can drive the phenotypic appearance of tissue, both diseased and normal. Nucleic acid based technologies allowing single cell analysis has been faster to mature, while technologies aimed at analysing the proteome at a single cell level is still lacking behind, especially technologies which allow single cell analysis in tissue. Introducing methods, that allows such analysis, will pave the way for discovering new biomarkers with more clinical relevance, as these may be unique for microenvironments only present in tissue and will avoid artifacts introduced by in vitro studies. Here, we introduce a technology enabling biomarker identification on small sub‐populations of cells within a tissue section. Phage antibody libraries are applied to the tissue sections, followed by washing to remove non‐bound phage particles. To eliminate phage antibodies binding to antigens ubiquitously expressed and retrieve phage antibodies binding specifically to antigens expressed by the sub‐population of cells, the area of interest is protected by a ‘shadow stick’. The phage antibodies on the remaining areas on the slide are exposed to UV light, which introduces cross‐links in the phage genome, thus rendering them non‐replicable. In this work we applied the technology, guided by CD31 expressing endothelial cells, to isolate recombinant antibodies specifically binding biomarkers expressed either by the cell or in the microenvironment surrounding the endothelial cell.     

Methylation-specific MLPA (MS-MLPA): simultaneous detection of CpG methylation and copy number changes of up to 40 sequences

Copy number changes and CpG methylation of various genes are hallmarks of tumor development but are not yet widely used in diagnostic settings. The recently developed multiplex ligation-dependent probe amplification (MLPA) method has increased the possibilities for multiplex detection of gene copy number aberrations in a routine laboratory. Here we describe a novel robust method: the methylation-specific MLPA (MS-MLPA) that can detect changes in both CpG methylation as well as copy number of up to 40 chromosomal sequences in a simple reaction. In MS-MLPA, the ligation of MLPA probe oligonucleotides is combined with digestion of the genomic DNA–probe hybrid complexes with methylation-sensitive endonucleases. Digestion of the genomic DNA–probe complex, rather than double-stranded genomic DNA, allowed the use of DNA derived from the formalin treated paraffin-embedded tissue samples, enabling retrospective studies. To validate this novel method, we used MS-MLPA to detect aberrant methylation in DNA samples of patients with Prader–Willy syndrome, Angelman syndrome or acute myeloid leukemia.     

Combined beta-galactosidase and immunogold/silver staining for immunohistochemistry and DNA in situ hybridization.

A combination of beta-galactosidase enzyme and the immunogold/silver staining method was studied for evaluation of double-staining experiments. Applications are shown for immunohistochemical double staining using two monoclonal antibodies and for combined immunohistochemistry and DNA in situ hybridization in one tissue section. The following advantages for the present double-staining method were evaluated: superior sensitivity of the immunogold/silver staining method for at least one epitope, which also allows detection of biotinylated DNA probes. The structure of the indolyl precipitate after revelation of beta-galactosidase activity did not show a concealing effect during a sequential double-staining method, as compared with the visualization of peroxidase with diaminobenzidine. These factors, and the sharply contrasting colored reaction products of beta-galactosidase (blue-green) and the immunogold/silver staining method including silver enhancement (brown-black), allow clear distinction of mixed-stained cell constituents.     

Conversion of a Capture ELISA to a Luminex xMAP Assay using a Multiplex Antibody Screening Method

The enzyme-linked immunosorbent assay (ELISA) has long been the primary tool for detection of analytes of interest in biological samples for both life science research and clinical diagnostics. However, ELISA has limitations. It is typically performed in a 96-well microplate, and the wells are coated with capture antibody, requiring a relatively large amount of sample to capture an antigen of interest . The large surface area of the wells and the hydrophobic binding of capture antibody can also lead to non-specific binding and increased background. Additionally, most ELISAs rely upon enzyme-mediated amplification of signal in order to achieve reasonable sensitivity. Such amplification is not always linear and can thus skew results.In the past 15 years, a new technology has emerged that offers the benefits of the ELISA, but also enables higher throughput, increased flexibility, reduced sample volume, and lower cost, with a similar workflow ^{1, 2}. Luminex xMAP Technology is a microsphere (bead) array platform enabling both monoplex and multiplex assays that can be applied to both protein and nucleic acid applications ^3-5. The beads have the capture antibody covalently immobilized on a smaller surface area, requiring less capture antibody and smaller sample volumes, compared to ELISA, and non-specific binding is significantly reduced. Smaller sample volumes are important when working with limiting samples such as cerebrospinal fluid, synovial fluid, etc. ⁶. Multiplexing the assay further reduces sample volume requirements, enabling multiple results from a single sample.Recent improvements by Luminex include: the new MAGPIX system, a smaller, less expensive, easier-to-use analyzer; Low-Concentration Magnetic MagPlex Microspheres which eliminate the need for expensive filter plates and come in a working concentration better suited for assay development and low-throughput applications; and the xMAP Antibody Coupling (AbC) Kit, which includes a protocol, reagents, and consumables necessary for coupling beads to the capture antibody of interest. (See Materials section for a detailed list of kit contents.)In this experiment, we convert a pre-optimized ELISA assay for TNF-alpha cytokine to the xMAP platform and compare the performance of the two methods ^7-11. TNF-alpha is a biomarker used in the measurement of inflammatory responses in patients with autoimmune disorders.We begin by coupling four candidate capture antibodies to four different microsphere sets or regions. When mixed together, these four sets allow for the simultaneous testing of all four candidates with four separate detection antibodies to determine the best antibody pair, saving reagents, sample and time. Two xMAP assays are then constructed with the two most optimal antibody pairs and their performance is compared to that of the original ELISA assay in regards to signal strength, dynamic range, and sensitivity.     

Multiplex bead array assays: performance evaluation and comparison of sensitivity to ELISA

The measurement of soluble cytokines and other analytes in serum and plasma is becoming increasingly important in the study and management of many diseases. As a result, there is a growing demand for rapid, precise, and cost-effective measurement of such analytes in both clinical and research laboratories. Multiplex bead array assays provide quantitative measurement of large numbers of analytes using an automated 96-well plate format. Enzyme-linked immunosorbent assay (ELISAs) have long been the standard for quantitative analysis of cytokines and other biomarkers, but are not well suited for high throughput multiplex analyses. However, prior to replacement of ELISA assays with multiplex bead array assays, there is a need to know how comparable these two methods are for quantitative analyses. A number of published studies have compared these two methods and it is apparent that certain elements of these assays, such as the clones of monoclonal antibodies used for detection and reporting, are pivotal in obtaining similar results from both assays. By careful consideration of these variables, it should be possible to utilize multiplex bead array assays in lieu of ELISAs for studies requiring high throughput analysis of numerous analytes.     

Application of immunohistochemical staining to detect antigen destruction as a measure of tissue damage

Electrocautery and directed energy devices (DEDs) such as lasers, which are used in surgery, result in tissue damage that cannot be readily detected by traditional histological methods, such as hematoxylin and eosin staining. Alternative staining methods, including 3-(4,5-dimethyl-thiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) to stain live tissue, have been reported. Despite providing superior detection of damaged tissue relative to the hematoxylin and eosin (H&E) method, the MTT method possesses a number of drawbacks, most notably that it must be carried out on live tissue samples. Herein, we report the development of a novel staining method, "antigen destruction immunohistochemistry" (ADI), which can be carried out on paraffin-embedded tissue. The ADI method takes advantage of epitope loss to define the area of tissue damage and provides many of the benefits of live tissue MTT staining without the drawbacks inherent to that method. In addition, the authors provide data to support the use of antibodies directed at a number of gene products for use in animal tissue for which there are no species-specific antibodies commercially available, as well as an example of a species-specific direct antibody. Data are provided that support the use of this method in many tissue models, as well as evidence that ADI is comparable to the live tissue MTT method.     

Targeted proteomics in biomarker validation: detection and quantification of proteins using a multi-dimensional peptide separation strategy

Authentic biomarkers, distilling the essence of a complex, functionally significant process in a mammalian system into a precise, physicochemical measurement have been implicated as a tool of increasing importance for drug discovery and development. However, even in spite of recent technological advances, validating a new biomarker candidate, where generation of suitable antibodies is required, is still a long-lasting task. Methods to accelerate initial validation by MS approaches have been suggested, but all methods described so far are associated with serious drawbacks, finally leading to non-generic methods of detection and quantification. Moreover, when complex body fluids are used as samples, efficient debulking strategies are crucial to open a window of analytical sensitivity in the ng/mL range, where many diagnostically relevant analytes are present. Here we report the proof-of-principle of a multi-dimensional strategy for accelerated initial validation of biomarker candidates by MS, which promises to be generally applicable, sensitive and quantitative. The method presented employs a combination of electrophoretic and chromatographic steps on the peptide level, followed by MS quantification using isotopically labeled synthetic peptides as internal standards. Our proposed workflow includes up to four dimensions, finally resulting in a desired LOD sufficient to detect and quantify diagnostically relevant analytes from complex samples. Although the current state of the method only represents a starting point for further validation and development, it reveals great potential in biomarker validation.     

Quality control of human tissues-experience from the Indiana University Cancer Center-Lilly Research Labs human tissue bank

The success of molecular research and its applications in both the clinical and basic research arenas is strongly dependent on the collection, handling, storage, and quality control of fresh human tissue samples. This tissue bank was set up to bank fresh surgically obtained human tissue using a Clinical Annotated Tissue Database (CATD) in order to capture the associated patient clinical data and demographics using a one way patient encryption scheme to protect patient identification. In this study, we determined that high quality of tissue samples is imperative for both genomic and proteomic molecular research. This paper also contains a brief compilation of the literature involved in the patient ethics, patient informed consent, patient de-identification, tissue collection, processing, and storage as well as basic molecular research generated from the tissue bank using good clinical practices. The current applicable rules, regulations, and guidelines for handling human tissues are briefly discussed. More than 6,610 cancer patients have been consented (97% of those that were contacted by the consenter) and 16,800 tissue specimens have been banked from these patients in 9 years. All samples collected in the bank were QC’d by a pathologist. Approximately 1,550 tissue samples have been requested for use in basic, clinical, and/or biomarker cancer research studies. Each tissue aliquot removed from the bank for a research study were evaluated by a second H&E, if the samples passed the QC, they were submitted for genomic and proteomic molecular analysis/study. Approximately 75% of samples evaluated were of high histologic quality and used for research studies. Since 2003, we changed the patient informed consent to allow the tissue bank to gather more patient clinical follow-up information. Ninety two percent of the patients (1,865 patients) signed the new informed consent form and agreed to be re-contacted for follow-up information on their disease state. In addition, eighty five percent of patients (1,584) agreed to be re-contacted to provide a biological fluid sample to be used for biomarker research.     

Immunoperoxidase staining of carcinoembryonic antigen with monoclonal antibodies in adenocarcinoma of the colon

Summary Mouse monoclonal antibodies to carcinoembryogenic antigen (CEA) obtained by the somatic cell hybridization technique of Köhler and Milstein were used in a modified enzyme bridge immunoperoxidase staining method. Both high and low affinity antibodies were tested and their staining properties compared with those of a commercial polyvalent rabbit antiserum. The staining pattern of neoplastic epithelial cells in all seven antibodies in samples of primary adenocarcinoma of the colon was similar, indicating that no gross differences were found in the exposure of the different antigenic determinants of CEA in formalin fixed tissue. The background staining of the monoclonal antibodies was negligible. It is concluded that monoclonal antibodies are superior to conventional antisera in immunoperoxidase staining of CEA.     

Identification of novel biomarker candidates for immunohistochemical diagnosis to distinguish low‐grade chondrosarcoma from enchondroma

Chondrosarcoma is the third most common primary bone cancer, requiring surgical resection. However, differentiation of low‐grade chondrosarcoma (grade 1) from enchondroma that is benign and only requires regular follow‐up is one of the most frequent diagnostic dilemmas facing orthopedic oncologists in clinical management. Although multiple techniques are applied to make the distinction, immunohistochemistry is an important ancillary technique, especially when a histopathological stain of specimen must be obtained in order to guarantee an accurate confirmation. Currently, no adequate immunohistochemical diagnostic protein biomarkers are available to distinguish low‐grade chondrosarcoma from enchondroma. To discover novel protein biomarker candidates, an LC‐MS/MS approach was applied to directly compare formalin‐fixed, paraffin‐embedded low‐grade chondrosarcoma with enchondroma tissue samples. The proteomics analysis revealed 17 protein biomarker candidates. A principle was developed to prioritize the candidates using category and ranking. An algorithm, prioritization index of biomarker candidates for immunohistochemistry on tissue specimens, was developed to rank the candidates inside each category. Using the proteomics data and bioinformatics results, the prioritization index of biomarker candidates for immunohistochemistry on tissue revealed periostin as a top candidate. Immunohistochemical staining of periostin in 23 low‐grade chondrosarcoma and 31 enchondroma tissue specimens disclosed 87% specificity and 70% sensitivity.     

Beta-tubulin isotype classes II and V expression patterns in nonsmall cell lung carcinomas

Previous studies suggest that beta-tubulin isotype protein levels could be useful as indicators of nonsmall cell lung cancer (NSCLC) aggressiveness. However, measurement of protein amounts in tissue samples by staining techniques is semiquantitative at best. Since technologies for measuring mRNA levels have become more efficient and quantitative, we wanted to determine whether beta-tubulin message levels may be useful as biomarkers. Quantitative real-time RT-PCR was used to measure the seven classes of beta-tubulin isotypes, stathmin and MAP4 mRNA levels in 64 NSCLC and 12 normal lung tissue samples. We found significantly higher fractions of beta-tubulin classes II and V mRNA compared to the other isotypes in all lung tumor samples (P < 0.05). In addition, the ratio of beta-tubulin classes II/V mRNA was significantly higher in NSCLCs compared to normal lung tissues (P < 0.001). The data suggest that the ratio of beta-tubulin classes II and V mRNA could be useful as a biomarker for NSCLC tumor differentiation and/or NSCLC aggressiveness. Furthermore, the ratio of MAP4 to stathmin mRNA was found to be higher in diseased lung tissues compared to normal lung tissues, suggesting this ratio might also be used as a clinically relevant biomarker for NSCLCs.     

A Flexible Mouse-On-Mouse Immunohistochemical Staining Technique Adaptable to Biotin-Free Reagents,Immunofluorescence, and Multiple Antibody Staining

Immunohistochemistry on mouse tissue utilizing mouse monoclonal antibodies presents a challenge. Secondary antibodies directed against the mouse monoclonal primary antibody of interest will also detect endogenous mouse immunoglobulin in the tissue. This can lead to significant spurious staining. Therefore, a “mouse-on-mouse” staining strategy is needed to yield credible data. This paper presents a method that is easy to use and highly flexible to accommodate both an avidin-biotin detection system as well as a biotin-free polymer detection system. The mouse primary antibody is first combined with an Fab fragment of an anti-mouse antibody in a tube and allowed sufficient time to form an antibody complex. Any non-complexed secondary antibody is bound up with mouse serum. The mixture is then applied to the tissue. The flexibility of this method is confirmed with the use of different anti-mouse antibodies followed by a variety of detection reagents. These techniques can be used for immunohistochemistry (IHC), immunofluorescence (IF), as well as staining with multiple primary antibodies. This method has also been adapted to other models, such as using human antibodies on human tissue and using multiple rabbit antibodies in dual immunofluorescence.