Keywords

Introduction

In 1995, television viewers around the world followed a criminal trial, People of the State of California v. Orenthal James Simpson, which introduced many viewers to some of the nuances of DNA testing. More recently, the film and television industries have incorporated identity testing or DNA analysis in many productions, further familiarizing the lay public with the concept of “DNA testing” for identification purposes. In 1997, one of the first case reports on the use of forensic DNA analysis for specimen identification was published [1]. While analysis of identity through DNA typing has its origins in parentage testing and forensic identity testing, similar techniques have also found clinical utility in pathology laboratories for a variety of sample identification applications [2–5]. Such applications include identification of the origin of mislabeled specimens [2], exclusion of a potential source for a surreptitious histologic “floater” in an anatomic pathology specimen [2, 4, 6], diagnosis of hydatidiform moles [7, 8], detection of maternal cell contamination in prenatal specimens [9, 10], identification of cell lines or research specimens [11], diagnosis of graft-vs-host disease (GVHD) [12], and identification of the origin of tumor cells in a transplant recipient [13].

Specimen misidentification is not a rare event in clinical practice. For prostate biopsies, the combined rate of type 1 (complete transposition between patients) and type 2 (contamination of a patient’s tissue with that of other patients) errors has been estimated as 0.93 % or nearly 1 in 100 [14]. Recently, DNA testing helped to resolve biopsy misidentifications in a large multicenter clinical trial (The Reduction by Dutasteride of Prostate Cancer Events clinical trial) [15], and The Dark Report (http://www.darkreport.com/) has reported that a large Urology Group in New England approached a commercial laboratory to perform DNA testing on cheek swabs to confirm patient identity on all patients with a positive prostate biopsy [16]. On occasion a patient who questions a new cancer diagnosis requests DNA identification analysis of the tumor to confirm that the diagnostic specimen originated from the patient’s body. A full description of the scientific principles and technologies for identity testing appears in Chap. 54. The practice of specimen identification through DNA analysis is well-established, but the cost-effectiveness of the use of these techniques must be determined on a case-by-case basis. This chapter focuses on identity testing issues and result interpretation specific to patient identification and related applications of identity testing in the clinical laboratory.

Methods for Specimen Identification

DNA typing exploits the polymorphic differences in DNA between individuals to resolve questions of identity. The first polymerase chain reaction (PCR) based DNA typing system to become commercially available was the AmpliType PM + DQA1 system from Applied Biosystems (formerly Perkin Elmer, Foster City, CA). This kit typed multiple loci simultaneously using a reverse dot blot procedure with an array of immobilized allele specific probes in a dot blot pattern on a strip of nylon membrane. The AmpliType PM+DQA1 amplification and typing kit was used routinely by many forensic laboratories, and it was applied to cases of clinical specimen identification as it could be used with many different specimen types [1, 17–19]. In recent decades, short tandem repeat (STR)-based methods have supplanted reverse dot blot methods, in part due to their greater discriminatory power. Single nucleotide polymorphism (SNP) analysis and mitochondrial DNA polymorphism analysis have some applications.

Short Tandem Repeat Analysis

Among methods of specimen identity testing, STR analysis has become the method of choice in most laboratories due to the convenience of several commercial assays, accessibility of equipment and genotyping software to assist with the analyses, and availability of technical resources and experience to facilitate the interpretation of results. The forensics community took the lead in defining panels of STRs (also known as microsatellites) leading to the development of the Combined DNA Index System (CODIS) loci originally selected by the US Federal Bureau of Investigation (FBI) for criminal identification investigations [20]. These loci have proven extremely useful because they exhibit not only polyallelism but also wide distribution of the different alleles across various racial and ethnic groups. Combined, these features allow STR-based testing to offer a great degree of discriminatory power for specimen identification.

For both forensic analysis and specimen identification, the ability to work with minute samples is important, as illustrated by the application of microdissection in resolution of tissue section “floater” cases (see below). STR-based methods employing PCR amplification of specific STR regions of the genome are applicable to small samples, permitting the interrogation of nucleic acids extracted even from samples with only a few cells as well as from partially degraded samples such as from formalin-fixed, paraffin-embedded (FFPE) tissue. Many STR polymorphisms are present in the human genome, and specific subsets are commercially available for identity testing purposes.

Multiplex primer sets are commercially available for amplification of STR loci (http://www.cstl.nist.gov/strbase/multiplx.htm) [20]. Such primers co-amplify multiple STR loci in a single reaction. Fluorescent labeling of the PCR primers at overlapping loci with different fluorophores allows multiplexing of STR loci which may have alleles that fall in the same size range. After amplification of the STR loci, fragment size analysis, usually by capillary electrophoresis, enables precise determination of the size of the polymorphisms at each locus for a specimen [2]. In general, two specimens are considered different when the alleles differ for at least one locus. The calculated probability that two individuals will share a set of STR alleles by chance depends upon the number of loci tested, as well as the ethnicity of the individual. Different alleles at as few as eight loci can quickly push the probability of two specimens matching by chance to less than 1 in 100,000,000, well below the number of patient specimens encountered in an entire year in a histology laboratory. Detailed information on STRs is available at http://www.cstl.nist.gov/strbase/.

Single Nucleotide Polymorphism Analysis

SNPs are the most common genetic variants in the human genome and occur approximately every 100–300 bases. Each individual has an allelic profile at these sites. Several studies related to specimen identification have been published using SNP profiling by real-time PCR as an alternative to STR analysis [21, 22]. SNP analysis has advantages over STR analysis, including a lower cost per reaction, smaller target sequence which is more amenable to amplification in degraded specimens, and no requirement for capillary electrophoresis equipment. A disadvantage of SNP analysis is a lower power of discrimination for each marker thus requiring use of a larger number of markers. Which SNPs to detect and their population frequencies must be carefully considered. Currently, SNP analysis is not frequently used for specimen identification except in degraded specimens.

Mitochondrial DNA Polymorphism Analysis

Mitochondria contain a circular genome that is distinct from the nuclear genome. Within this genome are two noncoding regions, hypervariable regions I and II, which vary in their sequences and contain known polymorphisms. Indeed, for unrelated individuals, the mitochondrial DNA sequence will differ at multiple nucleotide positions. However, maternal relatives share mitochondrial sequences except for the presence of new mutations. Mitochondrial DNA polymorphism analysis therefore can be used both to exclude the possibility that an unknown sample matches a reference sample or that the sample is derived from a maternal relative of the reference sample. Because mitochondrial genomes are naturally amplified due to the presence of multiple mitochondria per cell and the circular DNA is more resistant to degradation, sequencing of mitochondrial DNA may prove useful as an alternative to STR typing of archival specimens with low DNA content and high DNA degradation [23, 24]. Mitochondrial DNA polymorphism analysis is not commonly used for clinical laboratory specimen identity analysis.

Clinical Applications

DNA analysis of polymorphic markers used in forensic identity and relationship testing can aid in the identification of specimens that are essential to proper management of a patient when the possibility of specimen misidentification has arisen. As the cost of STR-based identity testing decreases, it has been suggested that routine verification of the source of certain cancer-positive diagnostic specimens (e.g., prostate biopsies), may become a cost-effective approach for preventing medical errors resulting from a misidentified patient sample [25].

Specimen Identification

For patient safety and quality assurance purposes, clinical specimens are required to be labeled with two patient identifiers to assure the proper identification of the patient specimen and to associate each specimen with a correct patient. Examples of identifying information include patient name, date of birth, medical record number, demographic data, date/time of specimen collection, and laboratory identifiers such as a unique accession number. Despite written policies, training of personnel, and the careful attention of personnel collecting and handling specimens, sample mislabeling or switches occasionally occur. Tissue specimens may be mislabeled, co-mingled, or interchanged at the time of collection (i.e., multiple biopsies from the same or different patients) or during the various stages of tissue processing such as labeling of the blocks containing the tissue or the glass slides with sections of the tissue [1, 3, 15, 17, 26]. In a Q-probe study performed by the College of American Pathologists (CAP), mislabeling occurred in approximately 1 per 1,000 cases [26]. Sometimes, a patient will not believe a tissue diagnosis made on a sample submitted under the patient’s name and will challenge a physician to prove that the tested and reported specimen was truly the patient’s specimen.

Mislabeled or unlabeled clinical specimens remain familiar identification issues in clinical laboratories. According to regulatory standards, these improperly labeled or unlabeled specimens should not be tested by a clinical laboratory. The strict adherence to this standard varies by specific circumstances. On occasion, when the clinician can definitively document the identity of such a specimen and the specimen is irreplaceable, a laboratory makes an exception. Ideally, though, one rejects a mislabeled specimen in favor of a new, equivalent specimen. Unfortunately, some specimens are difficult or impossible to replace. For example, a replacement blood sample is rather easily obtained in most circumstances, unless the specimen was a timed or pretreatment specimen. On the other hand, in the anatomic pathology laboratory, tissue specimen recollection may be difficult or impossible and typically results in increased morbidity for a patient who must undergo a repeat procedure. Occasionally, the specimen itself may be so unique that it suggests its likely source. In the right circumstances, available DNA identity testing methods can confirm or exclude the source of a sample.

Histologic “Floaters”

Unfortunate occurrences that become critical challenges for the anatomic pathologist are tissue fragments referred to as “floaters.” These tissue contaminants do not originate from the patient’s specimen but become intermingled with it at some point between specimen collection and pathologist interpretation. Indeed, these small fragments of loose tissue may be “carried over” from one case to another during various tissue handling or processing steps in the laboratory. During gross dissection, a tissue fragment from one dissecting field can be brought to another by a scalpel blade, forceps, or improper cleaning between grossing of specimens from different patients. Later, during histologic processing (i.e., fixation, dehydration, impregnation, sectioning, or staining), a tissue fragment may literally or figuratively float away from its source and become associated with another specimen. Such specimen contamination may occur despite rigorous quality control procedures. In one study, cross contamination to blank slides occurred in 8 % of cases (most commonly during the staining process) [27].

In many cases, during the course of careful histological examination of a slide, an anatomic pathologist is able to confidently identify when an unexpected fragment of tissue is present and dismiss it from diagnostic consideration. This is more easily accomplished when the “floater” does not resemble the histology of the remaining tissue, is positively recognized as coming from another case, is physically distinct from the main mass of tissue on the slide, or presents incongruous histology [2, 4]. However, when the “floater” is a type of cell or tissue compatible with the expected specimen or is closely intermingled with the expected cells or tissue, it is difficult to exclude the possibility that it is part of the sample originating from the patient. In addition, a “floater” may be a few cells or a cluster of cells that are difficult to evaluate and definitively classify. A common scenario is a tissue section that includes a small number of malignant cells suspected to be “floaters” by a pathologist. The pathologist cannot simply ignore such cells, but definitive interpretation and reporting of their significance is not straightforward. In the past, when diagnostic certainty could not be assured because of potential “floater” contamination, clinical recommendations were limited to a cautionary note such as “advise close follow-up and short term rebiopsy if possible” or a similar disclaimer. Using identity testing methods and processing techniques to carefully separate regions of tissue on a slide, the relevance of a potential “floater” to the diagnosis can be resolved.

Identifying Donor Cells in a Transplant Recipient

Chapter 56 details the utility of identity testing in the setting of hematopoietic stem cell transplantation (HSCT). Identifying cells of donor origin in a transplant recipient can be useful in additional settings, such as a new post-transplant neoplasm or suspected graft-vs-host disease in the recipient.

Donor-Transmitted Malignancies

Donor origin cancer, including malignancies transmitted with the graft as well as tumors that develop within a graft, have been estimated to occur in 0.02–0.06 % of recipients [28–30]. Donor-transmitted malignancies are more common in solid organ transplants than in HSCT [13]. When a transplant recipient develops a malignancy in the transplanted organ or in another location, identification of the origin of the tumor (recipient or donor) is important for patient management and for identification of other individuals who may be at risk because they received other tissues or organs from the same donor. Donors with a history of cancer are usually excluded; however, in some cases the cancer is remote or is not recognized at the time of tissue donation.

To assess the origin of a tumor sample one must have DNA from the recipient, the donor, and the tumor; the tumor being of “unknown” identity. The STR allele genotypes of these samples are compared. Usually the explanted paraffin-embedded tissue, if available, is a good source of recipient DNA; alternatively, recipient buccal cells may be used. Donor DNA or blood may be banked in an HLA laboratory or a donor tissue sample may be available (e.g., donor gallbladder is usually available in liver transplant cases). Alternatively, the transplanted organ can serve this function with the caveat that recipient white blood cells may be admixed, which can complicate the assignment of donor alleles. Assessment of the tumor specimen depends on whether the tumor is in the transplanted organ or in another location because this will determine the expected origin of the DNA of the adjacent normal tissue. The tumor specimen (either a biopsy or excision) should be evaluated as a paraffin-embedded tissue (unless it is a blood-based malignancy) in order to separate tumor from adjacent normal tissue, if possible. When the tumor is within the transplanted organ, the presence of recipient-origin white blood cells will complicate the interpretation. In such cases, a judgment must be made based on the relative abundance of donor and recipient DNA; therefore, it is preferable to isolate the tumor for DNA extraction by macrodissection or microdissection. A tumor that is not in the transplanted organ is easier to interpret as the presence of donor DNA is not expected, and the presence of a significant amount of donor DNA will indicate a tumor of donor origin. Assessment of recipient blood to rule out the presence of circulating donor DNA is a helpful control.

In a sex mismatched transplant, origin of a tumor can be inferred by traditional karyotyping when one has a viable tissue specimen with proliferating cells. Fluorescent in situ hybridization (FISH) can play a role in the evaluation of sex mismatched transplant recipients if the available sample is fresh frozen tissue, a cytologic preparation, or formalin-fixed paraffin-embedded tissue as the presence of two X chromosomes vs one X and one Y chromosome in the tumor cells can exclude a donor of the opposite sex as the source. Limitations of FISH include nuclear truncation artifacts, hybridization failures, and a reliance on single regions of chromosomes that may be altered in a malignancy.

Graft-vs-Host Disease

Graft-vs-host disease (GVHD) results when donor T cells recognize alloantigens expressed on host antigen-presenting cells and initiate an attack on host epithelial cells, damaging host tissue in a variety of possible anatomic locations including the skin, mouth, eyes, gastrointestinal tract, or liver. As such, skin, gastrointestinal tract, or liver abnormalities in the setting of an organ transplant or even certain types of blood transfusion can point to the possible diagnosis of GVHD. A National Institutes of Health working group recognized two main categories of GVHD: acute and chronic [31]. Diagnostic manifestations in the skin, mouth, eyes, female genitalia, esophagus, lungs, and connective tissues exist for both categories. Although not always mandatory for the diagnosis of GVHD, a biopsy can confirm the diagnosis.

While the donor T cells are nonneoplastic, they may be abundant in GVHD. Although low levels of donor T cell chimerism may be observed in the peripheral blood of liver transplant recipients during the first month after transplantation [32], a study of liver transplant associated GVHD found T cell macrochimerism [33] of at least 4 % in the peripheral blood of patients with histologically diagnosed GVHD after transplant. Hence, DNA-based identity testing of a specimen with T cells can confirm donor origin and provide support for a diagnosis of GVHD.

Specimens for confirmation of GVHD are similar to specimens for assessment of a donor-associated malignancy, with selection of a recipient specimen, the “unknown” specimen suspected of containing the inflammatory cells of GVHD (usually blood or epithelium), and a donor specimen. Results supporting a diagnosis of GVHD would consist of the presence of donor DNA at a significant level in a specimen where only recipient DNA is expected, whereas detection of only recipient DNA indicates that GVHD is less likely. A negative result (no donor DNA detected) requires careful interpretation when only minimal inflammatory cells are present in the “unknown” specimen, as this can represent a false-negative result.

Maternal Cell Contamination of Prenatal Specimens

Clinical laboratories performing prenatal diagnostic testing of chorionic villus sampling (CVS) or amniocentesis specimens, or of products of conception face the risk of inaccurate prenatal diagnosis due to the presence of maternal cells in the specimen. If the specimen is mistakenly thought to contain only fetal cells, contaminating maternal cells may lead to a misdiagnosis even when the level of contamination is seemingly modest (1–2 %) [34]. Even with only approximately 10 % of fetal cell preparations complicated by maternal cell contamination, the problem is so well recognized that maternal cell contamination assessment has been recommended for all prenatal specimens [35, 36]. Despite the recommendation, all laboratories performing prenatal genetic testing do not exclude maternal cell contamination for every prenatal specimen [37].

Sometimes there is frank maternal blood in an amniotic fluid sample, but low levels of contamination usually go unnoticed. Interestingly, culture conditions favor growth of amniocytes over contaminating maternal cells, resulting in less maternal cell contamination of amniotic fluid cultures compared to direct amniotic fluid specimens [38]. CVS specimens present a higher risk of maternal cell contamination because completely separating the fetal cells from the maternal decidua is difficult to accomplish by gross dissection.

STR analysis can be used to test DNA isolated from a fetal specimen with comparison to maternal DNA tested in parallel to determine if maternal alleles are present and if so, to approximate the percentage of maternal DNA [9, 10]. The percent contaminating maternal DNA is significant depending on the prenatal testing method and its sensitivity. A relatively insensitive method may not be affected by 10 % or less maternal contamination; however a very sensitive method could result in an incorrect interpretation due to the unintended presence of maternal DNA. Therefore, the significance of the level of maternal contamination is assessed in the context of the intended use of the prenatal specimen. Recommendations by the American College of Medical Genetics, the Clinical Laboratory Standards Institute, and the Clinical Molecular Genetics Society confirm the importance of maternal cell contamination testing in prenatal diagnosis. Detailed preanalytical, technical, interpretive, and reporting guidelines for maternal cell contamination testing are available [36].

Evaluation of Hydatidiform Moles

Aberrant fertilization of a normal or abnormal egg by one or more sperm can lead to the formation of a hydatidiform mole. A hydatidiform mole can be classified as either complete or partial. Complete hydatidiform moles occur when one (90 %) or two (10 %) sperm fertilize an anucleate ovum, and the proliferating gestational tissue is entirely paternally derived (i.e., androgenetic diploidy). A complete hydatidiform mole can be either homozygous, when a single sperm fertilizes an empty ovum and that sperm’s haploid genetic material is duplicated, or heterozygous when two sperm fertilize an empty ovum. Partial hydatidiform moles arise from the fertilization of a single egg by two sperm and are triploid with both maternal and paternal genetic material (i.e., diandric triploid; usually 69XXX or 69XXY).

Differentiation of a hydatidiform mole from a non-molar specimen, as well as the type of hydatidiform mole, is important for clinical management of the patient because complete hydatidiform moles have a higher risk of developing into choriocarcinoma than do partial moles. Reliable diagnosis of hydatidiform moles based solely on morphology is challenging especially after early evacuation of a complete mole and in many instances of partial moles [39]. Ancillary techniques such as DNA ploidy analysis, immunohistochemistry, conventional karyotyping, and interphase FISH can inform the diagnosis, but have limitations [40]. Partial moles are triploid, so can be distinguished from a non-molar specimen by DNA ploidy analysis. In contrast, complete moles with a higher risk of choriocarcinoma are diploid and harder to distinguish from a non-molar specimen by DNA ploidy analysis. STR analysis definitively distinguishes complete moles, partial moles, and non-molar specimens.

As with other applications of identity testing in the clinical laboratory, STR analysis is the most commonly applied approach to compare the genotypes of the hydropic chorionic villi to those of the decidua (or maternal blood) to determine the genetic origin of the villi. For identity testing, chorionic villi are microdissected from maternal decidua. Significant contamination with decidua tissue will complicate analysis since a complete mole has no maternal DNA while a partial mole does. The maternal DNA for comparison can be obtained from maternal blood if available, or alternatively from microdissected decidua. Additional testing of paternal DNA, while not required, can help confirm the informative alleles in the molar tissue. Analysis of complete moles will identify only non-maternally derived homozygous or heterozygous alleles at multiple loci in the hydropic villi consistent with fertilization of an empty ovum [7, 8]. Analysis of partial moles will show three different alleles at multiple loci, one of maternal origin and two of paternal origin consistent with the fertilization of a normal egg by two sperm. Non-molar specimens are usually diploid with one maternal and one paternal allele (biparental diploidy). In many cases, if the maternal and molar tissue is adequately separated, visual inspection of the results of STR analysis can be diagnostic of a complete or partial hydatidiform mole.

Specimen Considerations

Specimen Selection and Documentation

The critical initial step in identity testing is selection of appropriate specimens to test, which are determined in consideration of the purpose for the testing and the questions needing to be answered. At a minimum there will be an unknown sample and a known sample. In some cases additional known samples for comparison from one or more individuals may be required (Table 57.1).

Table 57.1 Specimens frequently used for identity testing
Full size table

For specimen identification testing, the “unknown” DNA from the specimen is compared to “known” DNA from a specimen verified to be from the individual, such as a blood sample or a tissue block from a previous procedure. For mislabeled specimens, additional documentation of specimen handling and processing is useful. In contrast, for the identification of suspected floaters that are not consistent with the remainder of the specimen section, “unknown” DNA from microdissected floater tissue will be tested and compared to “known” DNA from an area of tissue containing tissue consistent with what is expected for the specimen. In this scenario it may also be prudent to analyze a verified alternative sample from the same individual as the current specimen in question as a further confirmation that the “known” DNA did, in fact, originate from the individual whose sample was contaminated by the floater. Confirmation of the actual source of the floater is unnecessary in most cases. In a case of two possibly switched samples, both are “unknown” and can be compared with two “known” samples from the two potential source individuals.

Analysis of maternal cell contamination of a fetal specimen requires a known maternal blood sample in addition to the “unknown” fetal sample. Similarly, diagnosis of a hydatidiform mole requires analysis of both chorionic villi and a known maternal sample. In the typical post-transplant scenario, analysis of the “unknown” tumor specimen is compared with the analysis of “known” specimens from the donor and recipient.

These are only general examples, some of which are illustrated in the cases described later. In fact, identity testing for sample identification is useful in a host of situations, each requiring careful specimen selection on a case by case basis to ensure that the tested specimens reflect the best choices from the available specimens, and that the results of testing those specimens will enable resolution of the clinical question.

Tissue Processing

While specimen handling and processing are major issues for certain types of molecular genetic testing, the use of smaller polymorphic markers allows identity testing to be performed on specimens that would be less than ideal for other clinical molecular tests. Frequently, identity testing for sample identification involves the use of FFPE tissue specimens, which is the routine method for preserving tissue for histopathology in nearly all histology laboratories. Commercial reagents simplify extraction of adequate DNA for identity testing from FFPE tissue sections. However, not all fixatives or chemical treatments used in the histology laboratory are compatible with PCR amplification. Specimens exposed to fixatives containing heavy metals (for example B5 fixative) are not recommended due to inhibition of enzymatic reactions in PCR. Similarly, decalcification solutions are usually acidic as are some fixatives (e.g., Carnoy’s, Zenker’s, and Bouin’s fixatives), which degrade the DNA and render the specimens unsuitable for PCR amplification. Therefore, it is very important to review the tissue processing during selection of tissue blocks for identity testing to avoid incompatible treatments. In extraordinary cases when the only specimen available is suboptimal but the testing will have a significant impact on patient management, testing of the compromised specimen can be considered. In such cases, if there is some amplification of the DNA, it may only be for loci with the smallest PCR target sizes. In some situations, sufficient information may be obtained to provide a limited interpretation, at the risk of over-interpreting potential artifacts that can occur at low DNA concentrations. If critical, testing using SNPs or mitochondrial DNA can be considered since those methods are more tolerant of degraded specimens.

Some staining methods may produce added damage to the nucleic acids. Most hematoxylin and eosin (H&E) stained tissue sections are acceptable [6, 41]. However, because reagents for histochemistry or immunohistochemistry may substantially degrade the nucleic acids or inhibit PCR and the effects of such potential interferences may not have been thoroughly examined during assay validation, due caution should accompany interpretation of STR analysis of previously stained material.

Selection of Tissue for Testing

Identity testing of tissue specimens begins with review of all available material by an anatomic pathologist. After assessing the quantity and distribution of the available material, the best method of isolating DNA from the submitted material is chosen. Available methods range from extraction of DNA from an entire tissue section to extracting DNA from small, isolated areas within a tissue section by macrodissection or microdissection. Regardless of the final method used, the goal of tissue selection is to obtain the appropriate specimens for comparison.

Testing All Tissue Within a Block

When the identity of all of the tissue in a block is in question, several (e.g., three) 10 μm thick section rolls from the FFPE tissue block can be cut and placed directly into microcentrifuge tubes for DNA extraction. This saves time and eliminates the effort of preparing a slide and then scraping the tissue off the slide for DNA extraction.

Testing Areas of Interest Within a Single Block

When different portions of tissue in the same block potentially originate from different sources, macrodissection or microdissection of the tissue from representative slides (see below) is appropriate to separate the tissues for subsequent comparison. Macrodissection after appropriate training and practice can be regularly employed using readily available tools [42]. Microdissection techniques permit more precise selection of cells for testing, but require more expertise and expensive equipment.

When the only slide containing the area of interest has been previously stained and coverslipped, xylenes can be used to remove the coverslip (with a risk of losing the desired cells) for DNA extraction from the stained tissue [6, 41]. Importantly, the scenario of tissue present on only a single level is suggestive of a true “floater” and may not merit further testing because tissue that is truly a part of the specimen is typically present on more than one level of a block.

Macrodissection

For macrodissection, an anatomic pathologist first identifies and outlines representative cellular, nucleated, non-necrotic areas of interest on a stained slide using a permanent marking pen (see Fig. 57.1a). Then, unstained 4–10 μm sections are cut and placed onto glass slides (e.g., 5–15 slides are prepared). The first slide, “cut off the top,” and the last slide “cut off the bottom” are stained with H&E to demonstrate the distribution of the areas of interest across the intervening unstained slides. Identifying the distinct areas of interest is necessary when the tissue on a slide is not homogenous. Regions representing “floaters,” benign, malignant, or other relevant tissues and cells on a slide are delineated with a diamond etching pen (on the non-tissue-containing surface of the slide) or a permanent marking pen on a coverslipped slide (Fig. 57.1a, b). Alternatively, a permanent marking pen can be used to mark areas of interest directly on a faced FFPE block, and this marked area can be manually separated from the remainder of the block. The marked slides are used as guides for identifying and marking the areas of interest on the unstained slides that will be used for dissection and extraction of DNA (Fig. 57.1c). Macrodissection is done by eye or using a low power dissecting microscope and the equivalent circled tissue regions are scraped off the unstained slides (Fig. 57.1d).

Figure 57.1
figure 1

Macrodissection can be used to enrich for tumor and/or normal tissue from histologic tissue sections for identity testing, depending on the question being asked. Regions of interest on an H&E stained slide are indicated by an anatomic pathologist by circling the tissue areas of interest using a permanent marking pen, as shown in (a) (*, tumor; N, normal). This functions as a guide slide for marking regions of interest on unstained slides of sections adjacent to the tissue on the guide slide (b). A final H&E stained slide cut after the unstained sections is used to confirm that the tissue of interest is present on all the unstained slides. The marked tissue from the unstained slides (c) can then be individually scraped with a scalpel to isolate the tissue of interest (d)

Full size image

Microdissection

Diamond etching pens and permanent marking pens are useful for delineating distinct areas on a slide, and in most cases of identity testing they are satisfactory for marking relevant areas to be collected for DNA extraction and comparison. When greater precision is required, microdissection techniques (i.e., use of a higher power microscope to assist dissection) may be used to isolate discrete cells or clusters of cells from a complex tissue section.

Laser Capture Microdissection

Laser capture microdissection (LCM) permits reliable procurement of pure cell populations from tissue sections. The principle advantage of LCM is the ability to isolate clusters of cells or even single cells of interest as a pure population free from contaminating stromal, inflammatory, and other surrounding cell types. Its precision can also complicate analysis by limiting the choice of analysis methods to those that are amplification-based because the amount of material isolated often is minute. Due to the increased expertise and cost associated with the LCM systems, many laboratories are unable to use LCM routinely for identity testing.

LCM systems such as the ArcturusXT™ LCM System (Life Technologies, Grand Island, NY), Leica LMD6500 and LMD7000 (Leica Microsystems Inc., Buffalo Grove, IL), or PALM MicroBeam (Carl Zeiss AG, Oberkochen, Germany) are available. Each system has unique technologies and techniques for separating the selected areas of interest from tissue sections. In general, the systems include an inverted microscope, an infrared laser, control systems for the laser and the microscope stage, and a digital imaging system to permit the user to view and capture images of the microdissected fragments [43]. The systems differ mostly in terms of the slides and other consumables they require to separate and collect the cells.

Chemical Microdissection

An alternative to LCM is The PinPoint Slide DNA Isolation System™ (Zymo Research, Orange, CA) [17]. With this system a solution is applied to the microscopic area of interest. The solution dries into a thin film that captures the underlying cells. The film is lifted with a scalpel and transferred into a tube for DNA extraction. This method is more time consuming than macrodissection of marked slides and requires a specific kit, but because the area of interest is precisely selected by an anatomic pathologist, the possibility that an incorrect area will be used for analysis is reduced.

Interpretation of Results

Specimen Identification

Sample identification determines whether two samples originated from the same or different individuals. Comparison of a tissue or other specimen with an identified specimen from the potential source patient, such as peripheral blood, can be used to verify the source of the specimen in question. Interpretation of the results involves comparison of the alleles (STR or sequence polymorphisms) between the samples. The identification of non-matching alleles, preferably in two or more loci, provides evidence to exclude a potential individual as the source of the sample, or at least to conclude that both samples did not originate from the same individual. The possibility of microsatellite instability, loss of heterozygosity, or chromosomal abnormalities should be considered when malignant tissues are being compared to normal samples using STR analysis, as described further below. If an exact allele match is observed between two samples, then the likelihood that they are from the same person is high, and the matching probability can be determined from published data or manufacturer data for the loci used. The following case is an example of identity testing for sample identification.

Case 1

Two blocks of tissue (Unknown A and Unknown B) obtained from two different individuals were suspected of being switched and mislabeled. An alternate, confirmed sample from each individual (Confirmed A and Confirmed B) was available. DNA was extracted from all four specimens and each was amplified at 12 STR loci. Comparison of the genotypes indicated that the unknown samples had been switched (Table 57.2). Blocks Unknown A and Unknown B were relabeled to accurately reflect their true source, thereby avoiding the diagnoses of the two individuals being switched.

Table 57.2 Case 1 genotypes confirming that specimens were switched
Full size table

A “confirmed” specimen is partially a subjective designation due to the possibility of specimen misidentification. As such, in interpretive reports language such as “…block number 1 is identified as originating from patient ‘1’ and block number 2 is identified as originating from patient ‘2’…” makes it clear that even for specimens with presumably confirmed identity there remains the possibility of preanalytical errors resulting in a specimen misidentification.

Analysis of Floaters

To assess the origin of an extraneous, frequently small, tissue fragment (floater) in a tissue section, the genotype of the majority of the tissue on the slide is compared with that of the floater. As with other forms of identity testing, differences in two or more markers should be identified in order to exclude the patient as the source of the floater. Generally it is not necessary or cost-effective to identify the source of the extraneous tissue; however, if desired, possible source cases with similar histology that were processed during the same time period can be tested. As always, histopathologic review of the case to verify the nature of the submitted material and to select the appropriate tissues for analysis is essential before initiating identity testing. The histopathologic review also informs the analysis, as macrodissection is frequently imperfect and a minor component of the predominant tissue’s genotype may be detected in what is intended to be only floater tissue. When mixed genotyping result are obtained, correlation with the histologic picture usually will show intermingled “known” and “floater” cells in the specimen with the mixed genotype result.

Case 2

An isolated fragment of adenocarcinoma was identified in a single block of prostate needle biopsies. The remaining five blocks were completely benign. A “floater” was suspected and the possible source was another prostate biopsy case processed immediately prior to the case with the suspected floater. The adenocarcinoma floater was macrodissected, as well as the majority tissue from the benign case with the floater and the prior case with adenocarcinoma. The three tissue DNAs were tested at 12 STR loci and the results confirmed that the tissue fragment was a “floater” (Table 57.3). As expected, testing of the amelogenin locus confirmed that both samples were from male patients.

Table 57.3 Case 2 genotypes confirming the presence of a floater
Full size table

Cells of Possible Donor Origin in Transplant Recipients

Case 3

GVHD was suspected in a patient who was critically ill and pancytopenic one month after an orthotopic liver transplant. A bone marrow biopsy showed that his bone marrow was aplastic. Identity testing was requested to determine whether donor cells were circulating in the recipient, which would support a diagnosis of GVHD [44]. Twenty STR loci were amplified using recipient DNA (FFPE tissue from the explanted native liver), donor DNA (FFPE tissue from the donor gallbladder that was removed and processed at the time of transplant), and three recipient-derived specimens: peripheral blood, CD3-positive cells (T cells) isolated from peripheral blood, and bone marrow. The results confirmed the presence of donor cells in each specimen. The relative amounts of recipient and donor DNA were calculated from the allele peak areas of four representative loci (Fig. 57.2), which demonstrated 62 % donor alleles in the recipient’s bone marrow and supported the diagnosis of GVHD. GVHD after solid organ transplant has been reported, and in several cases macrochimerism (>1 % donor DNA) was observed. Thus, the finding of significant donor DNA in this case was consistent with the clinical impression of GVHD [45]. The percent of donor DNA in the blood may be tracked over time using this same method.

Figure 57.2
figure 2

Electropherograms from Case 3. Amplification results for four STR loci are shown for DNA extracted from recipient native liver tissue (a), donor gallbladder tissue (b), and post-transplant recipient bone marrow (c). The bone marrow specimen shows the presence of both donor (green D) and recipient (R) alleles at all four loci. Alleles shared by the donor and recipient are indicated (D/R). The percent of donor DNA is calculated from the peak areas as described in Chap. 56. In this sample, the donor DNA was 62 %. This result was interpreted as macrochimerism supportive of a diagnosis of GVHD. al allele, ar area of peak, GVHD graft-vs-host disease, STR short tandem repeat

Full size image

Case 4

Six months after receiving a double lung transplant, a patient was diagnosed with metastatic squamous cell carcinoma in a subcarinal lymph node. Identity testing was requested to ascertain whether the squamous cell carcinoma was of recipient or donor origin and to guide treatment of the patient and other potential recipients of organs from the same donor. Identity testing using 12 STR loci was performed using DNA from an FFPE cell block from an a fine needle aspiration (FNA) of the subcarinal lymph node with squamous cell carcinoma, an FFPE tissue section from the native right lung, and a stored peripheral blood specimen from the donor. Test results showed that the lymph node with the squamous cell carcinoma DNA had the same alleles as DNA extracted from the recipient’s native lung specimen at 11 informative STR loci (Fig. 57.3 and Table 57.4). In interpretation of the test results, consideration was given to the surgical pathologist’s note that the FNA only contained 5–10 % tumor cells. The relative amounts of neoplastic and nonneoplastic cells can help in interpretation of relative peak heights in an electropherogram and must be compared with an assay’s ability to detect low levels of a minor DNA component. There was no evidence of donor allele peaks in the FNA specimen’s electropherogram and the assay was validated to detect at least a 5 % component of DNA. In this context, the results were interpreted as consistent with a recipient origin of the squamous cell carcinoma cells in the lymph node.

Figure 57.3
figure 3

Electropherograms from Case 4. Amplification results for four STR loci are shown for DNA isolated from tumor-containing specimen (a), recipient native lung (b), and donor blood (c). Results indicate no evidence of donor-specific alleles in the tumor-containing specimen (a) at the level of sensitivity of the assay (5 %), supporting a recipient origin for the tumor. al allele, ar area of peak, STR short tandem repeat

Full size image
Table 57.4 Case 4 genotypes demonstrating a match between the tumor and the recipient and confirming the tumor originates from recipient not donor cells
Full size table

Maternal Cell Contamination of a Prenatal Specimen

Case 5

A 37 year old pregnant woman underwent CVS. A limited amount of villi were collected, cultured, and karyotyped with a result of 46, XX. Identity testing was requested to determine whether the karyotype reflected only chromosomes of the fetus or contaminating maternal cells. Identity testing using 12 STR loci and the amelogenin locus was performed using DNA from the cultured chorionic villus sample and DNA from a peripheral blood sample from the mother. The identity testing results demonstrate that the CVS contained fetal DNA with evidence of Mendelian inheritance, as well as low-level maternal contamination. The relative amounts of maternal and fetal DNA were calculated from the allele peak areas at informative loci with approximately 4 % maternal DNA present (Fig. 57.4).

Figure 57.4
figure 4

Electropherograms from Case 5. Amplification result for the TPOX locus is shown for DNA from chorionic villus sample (CVS) (a) and maternal blood (b). The CVS specimen shows a small amount (3.8 %) of maternal contamination indicated by the presence of the maternal allele of 6 repeats (red arrow). Mother and fetus share the 11 repeat allele. The percent of maternal contamination can be calculated using the peak areas as described in Chap. 56. al allele, ar area of peak

Full size image

Analysis of Hydatidiform Mole

Identity testing is used to characterize hydatidiform moles as complete or partial moles, as well as homozygous or heterozygous complete moles.

Case 6

A 25-year-old pregnant woman underwent a dilation and evacuation procedure after a missed abortion. The surgical pathologist who examined the specimen noted degenerated hydropic chorionic villi suggestive of a molar pregnancy. Tissue was submitted for cytogenetics and the resulting karyotype was 46, XX. Her physician was concerned about persistent trophoblastic disease or early choriocarcinoma based on serial measurements of human chorionic gonadotropin (hCG) in the patient’s blood. Most cases of complete mole or choriocarcinoma have 46 chromosomes which are all of paternal origin. The genetic origin of the villi from the products of conception was questioned and DNA genotyping analysis was requested. Macrodissection was used to isolate the hydropic chorionic villi from the remaining tissue on the section (selected regions shown in Fig. 57.5).

Figure 57.5
figure 5

H&E-stained FFPE tissue section of a chorionic villus sample from a suspected molar pregnancy. Regions of hydropic villi were circled (green circles) with a marking pen to create a guide slide used as a reference for macrodissection of tissue on adjacent unstained slides. DNA was extracted from the macrodissected tissue for identity testing (fig. 57.6)

Full size image

Identity testing at 12 STR loci was performed on DNA from three specimens: macrodissected hydropic chorionic villi (Fig. 57.5) and maternal and paternal peripheral blood specimens. The chorionic villus genotype (Fig. 57.6 and Table 57.5) showed a mixed pattern attributable to incomplete separation of abnormal villus tissue from adjacent normal decidua by macrodissection, which is a common problem due to the difficulty in separating villus tissue from decidua using macrodissection rather than microdissection. Although the pattern appears triploid (three alleles at some loci), each locus has a predominant allele peak matching a single paternal allele and smaller adjacent peaks attributable to contaminating maternal alleles. This case demonstrates the importance of careful microdissection, as well as careful interpretation of the results in the context of the histology. A higher level of maternal contamination would be difficult to interpret. Based on a comparison with the paternal genotype, the dominant allelic pattern in the hydropic villi sample is most consistent with paternal isodisomy in which only a single paternally derived allele is present at each informative locus. The reported histologic features and identity testing interpretation support the diagnosis of a complete hydatidiform mole.

Figure 57.6
figure 6

Electropherograms from Case 6. Amplification results for four STR loci are shown for DNA extracted from maternal blood (a), paternal blood (b), and macrodissected hydropic chorionic villi (c). The dominant alleles in the villus sample are composed of only a single paternal allele most consistent with paternal isodisomy. In addition, at all informative loci, both maternal alleles (red arrows) are present as a result of incomplete macrodissection with inclusion of maternal decidua. This result was interpreted as a complete hydatidiform mole. al allele, ar area of peak

Full size image
Table 57.5 Case 6 alleles identified at each locus tested with bracketed alleles present at lower amplitude and derived from contaminating maternal tissue due to incomplete macrodissection of the villi from the decidua
Full size table

Identity Testing Results for Neoplastic Tissue with Mutations

The presence of a genotype mismatch at two or more loci is usually sufficient to conclude that two samples are from two different individuals. However, when one of the tissues is neoplastic, the differences at a single locus or few loci may be due to mutations in the tumor tissue. Nucleotide polymorphisms are often conserved between the normal and tumor tissue from the same patient, but microsatellite instability (MSI), loss of heterozygosity (LOH), and chromosomal alterations occur in tumors and should be considered as a possible cause for allelic differences. MSI in a tumor results when errors in DNA replication of microsatellites (simple repetitive sequences) are unable to be repaired by DNA mismatch repair (MMR) proteins, either due to mismatch repair gene mutation or promoter hypermethylation. Since identity testing relies on polymorphic STR sequences, MSI at a locus being tested is a theoretical confounding factor in the interpretation of identity testing results. LOH at a tested locus is also possible. If a deletion or other mutational event involving a tested allele renders the tumor cells either hemizygous or homozygous at that allele compared to the heterozygous benign cells, then that discrepancy may mistakenly be interpreted as an indication that the tumor did not originate from the patient. Loci known to be frequently compromised by MSI, LOH, or chromosomal alterations and loci compromised in a specific tumor should not be utilized for an identity testing application. Therefore, when analyzing neoplastic tissue, examination of a number of loci rather than relying on the results at a single locus is necessary, and a cytogenetic karyotype, if available, can be used to assist with interpretation.

Conclusion

Identity testing methods are well established for use in a variety of scenarios encountered in a clinical laboratory. Beyond the routine use for monitoring engraftment for HSCT, the applications described in this chapter demonstrate wide-ranging applicability of identity testing for clinical specimen identification. These applications include suspected specimen misidentifications, histologic floaters, or concern over the attribution of a diagnosis for a patient. Identity testing also has a demonstrable diagnostic role in the classification of hydatidiform moles, assessment of maternal cell contamination in prenatal specimens, determination of the origin of a tumor in a transplant recipient, and diagnosis of GVHD. In a vast majority of cases, polymorphic DNA marker analysis combined with clinical information and histologic correlation provides the answer to questions which are otherwise difficult to resolve.