The Evolution of Pre-analytic Factors in Anatomic Pathology

Abstract

Anatomic Pathology has continuously evolved since launch by Virchow in Berlin. The era from 1990 to 2010 saw the rise of immunohistochemistry and its application for diagnosis, prognosis, and prediction of response to therapy. Currently the next wave of evolution is ongoing; molecular pathology, with emphasis on alterations to DNA, and expression of mRNA as biomarkers. The interrogation of biomolecules by specific probes is more demanding on specimens than the traditional application of histologic stains to tissue. This issue is juxtaposed to the fact that the majority of specimens are purely evaluated by histomorphology, for which current specimen practices are adequate. The capacity to identify a priori which cassette of tissue is appropriate for molecular analysis is difficult, if not impossible, the goal is to improve the quality of all pathology specimens in an economically viable model to enable advanced assay, when applicable.

1. Introduction

Anatomic Pathology has continuously evolved since launch by Virchow in Berlin. The era from 1990 to 2010 saw the rise of immunohistochemistry and its application for diagnosis, prognosis, and prediction of response to therapy. Currently the next wave of evolution is ongoing; molecular pathology, with emphasis on alterations to DNA, and expression of mRNA as biomarkers. The interrogation of biomolecules by specific probes is more demanding on specimens than the traditional application of histologic stains to tissue. This issue is juxtaposed to the fact that the majority of specimens are purely evaluated by histomorphology, for which current specimen practices are adequate. The capacity to identify a priori which cassette of tissue is appropriate for molecular analysis is difficult, if not impossible, the goal is to improve the quality of all pathology specimens in an economically viable model to enable advanced assay, when applicable.

2. Specification for Fixed Tissue

The specification of tissue preparation for pathologic examination is relatively simple. The primary goals are:

1. Preservation of tissue in such a manner that it is a permanent record, available for re-evaluation with new diagnostic modalities.

2. Economics are critical, as a large volume of tissue is collected and stored. The pathology laboratory is paid once, at time of diagnosis, but must store the material for a minimum of a decade. Volume of specimens drives low cost for preservation, and payment model drives low cost for storage.

3. The collection and preservation of tissue is a labor-intensive process with multiple staff involved across sites from patient care to the laboratory. Safety must be a primary concern in handling reagents.

4. The fixative is an anaseptic, preventing the growth of microorganisms in the tissue.

5. The methods must be applicable worldwide, to support the fund-of-knowledge employed by pathologist, as well as allow consultation.

Although far from perfect, the use of formalin fixation and paraffin embedding has evolved as the preferred method of tissue preservation. Alternative fixatives abound, most commonly containing acids, alcohols, and/or glycols. In the past these fixatives have been limited to specialty uses, or limited geographic distributions, but have seen diminished use with the introduction of large tissue processors, and the demands for immunohistochemistry. With formalin now acknowledged as a (likely) carcinogen, it remains unclear what the fixative in widespread use 25–50 years from now will be (http://www.cancer.gov/cancertopics/factsheet/Risk/formaldehyde).

3. Evolution/Fit-for-Purpose

Although the application of formalin fixed, paraffin embedded tissue to diagnostic histopathology appears static, and relatively unchanged over the last 120 years, this is not accurate. The fixation of tissue has been carried out for centuries. Prior to the introduction of formaldehyde as a fixative in 1893 [1], alcohols and other organic solvent based fixatives predominated. The introduction of buffers to formaldehyde solutions dates from the mid-twentieth century, and were applied to reduce the formation of “formalin-pigment”, iron-formaldehyde precipitates. These first buffers were commonly formulations of calcium. Multiple buffers have been applied since that time, largely consolidating into the use of phosphate buffers, contributing both buffering of pH as well as modification of osmolarity of the buffer. As demonstrated by Chung et al. the evolution of buffers for formalin supported the nascent development of molecular pathology, with improved RNA preservation compared to other buffers, and undoubtedly improved immunohistochemical assays, although the benefit was obscure to most investigators [2].

Concurrent with the evolution of fixation, impregnation evolved, with the introduction of the first instrumentation for the serial dehydration, clearing, and impregnation of tissue introduced in first half of the twentieth century, and the development of vacuum processing by Lillie in the 1940s [3] (SMH Library). Although beneficial, vacuum impregnation is not universally used. Many low volume laboratories, as laboratories in Africa, and other underdeveloped nations continue to rely on rotary “dip and dunk” tissue processors. Concurrent with this, reagents became more refined and standardized. Paraffins underwent a substantial advancement from preparations derived from beeswax to synthetic, lower-melting point paraffins which provide better impregnation, and result in better section quality with microtomy [4].

4. High-Throughput Pathology—A Turning Point

Until the early twenty-first century the impact of pre-analytic variables was largely obscure to pathologist. Although immunohistochemistry had entered clinical practice, antigen retrieval was still viewed more as magic than science, and little standardization or systematic approach was applied. Histology laboratory practice was derived from a limited number of manuals, and local practice generally followed local variations based on pathologist preference and availability of equipment and reagents.

The tissue microarray (TMA) was described by Kononnen in 1998 [5]. Although a refinement on the work of Battifore [6] (Can find in a SMH paper), it represented the turning point. The modern TMA was high-throughput, with hundreds of samples per recipient block, and used primarily for immunohistochemical analysis. Studies jumped from tens of samples from a single site and unit numbers of immunohistochemical markers to hundreds of samples collected from multiple sources, or over longer periods of time with tens of immunohistochemical markers. The direct impact was that suddenly variables that had been invisible to investigators because of the small scope of studies now became glaring, when they could be visualized within a single experiment, and not simply dismissed as a failure of the investigator to accurately replicate the assay [4] (SMH review).

5. Quality—Subjective

The challenge remains how to quantify quality in tissue. The historic, and still most commonly used approach is cyto- and histomorphology observed on a Hematoxylin and Eosin (H&E) stain. This process is inherently subjective, based on the experience and preferences of the pathologist as well as the tissue evaluated. Furthermore, breaking and H&E is very difficult—when a section can be obtained from a block, it is generally of acceptable quality, and in fact it is the difficulty of obtaining the section that is a far better metric [2].

Other metrics remains as complex. Protein quality, generally evaluated by immunohistochemistry, results in sample bias, especially when an immunohistochemical assay is optimized for an undefined pre-analytic matrix [7]. The evaluation of these assays remains manual and qualitative, with limited dynamic range. Evaluation of nucleic acids is equally challenging—either relying of fragment length and quantity recovered, or dependent on specific assays for which the two metrics demonstrate poor correlations [8].

6. Immunohistochemistry as the Gatekeeper of Quality

Over the last decade, with the increased use of immunohistochemistry, the baseline for the evaluation of tissue, and specification on which pre-analytic variables are defined has shifted from producing a specimen adequate for diagnosis by histologic stains, and is stable over time, to a specimen adequate for immunohistochemistry of prognostic/predictive biomarkers. Although this shift appears small, it is in fact substantial, as it adds critical specifications to those already described. Until this shift, it was not uncommon for histology laboratories to utilize different fixatives, based on preferences based on cyto- and histo-morphology to shift everything to a single fixative to support the validation of IHC. Phosphate-buffered formalin became essentially universal [2].

Immunohistochemistry also rose to dominance ahead of the other molecular assays, and because of the breadth of specific targets that can be measured, the simplicity of the assay, and the economics of the assay has become a gate-keeper before the application of other molecular assays (including FISH and PCR-based assays, but also sequencing applications and advanced proteomic assays).

7. Steps Forward

With the unmasking of the lack of pre-analytic standardization and a lack of science to support more advanced rigorous specifications, a number of studies have been carried out to better define the impact of pre-analytic variables. Below are some of the advances, however more glaring are the “black spaces” that have yet to be investigated.

8. Quantity of RNA

In a paper published in 2006, Chung was the first to show quantitative data on RNA recovery from tissue, comparing frozen, fixed, and paraffin embedded tissues [9]. This data was a substantial step forward in appreciating how to leverage RNA from FFPE tissue. The authors demonstrated that the fixation process was the period during which RNA “loss” occurred, and that tissue processing had minimal impact on RNA quantity. The authors also showed that ethanol fixation, and its absence of crosslinking retained 70 % of the RNA in tissue, compared to frozen tissue, and formalin fixation resulted in a pool of 30 % of the RNA as measured by quantity per mg of starting tissue.

9. Fixation Time, Buffers, and Processing Time

Carrying this effort forward, the authors evaluated three variables and demonstrated at the RNA level, but by evaluation of total RNA, as well as transcript-specific measurements optimal conditions for the length of formalin fixation, the impact and choice of buffers in formalin, and the impact of time (isolated as a single variable) in the process of impregnation of tissue [2]. This study resulted in a new model of tissue fixation. The authors, using PCR were able to demonstrate that in formalin fixed paraffin embedded tissue, mRNA integrity was preserved mid-gene, with degradation originating from both the 5′ and 3′, while frozen tissue, RNA integrity was relatively flat across the gene, when reverse transcription is mediated by random hexamers. Although not altering the chemical models of fixation, the authors put for a model in which the tissue suffers a terminal hypoxia event, when immersed in fixative, and the artifact embodied by the RNA was a result of the cells response to hypoxia.

10. Storage

In their latest studies, this same group has investigated the stability of formalin fixed, paraffin embedded tissue [7]. Although the degradation of RNA in FFPE blocks, and proteins in cut sections has been well documented, no mechanistic models existed, and all prior efforts were empiric, and of dubious value. The authors demonstrated that degradation was dependent largely on impregnation, rather than fixation, and that residual water, from inadequate dehydration, was responsible for a substantial proportion of the degradation. The authors also demonstrated that this process was temperature dependent.

There remains a long list of variable yet to be tested, including for example reagent quality and substitution. Most concerning, no variable yet examined to date, has not been demonstrated not to be important, and new potential variables continue to come to light.

11. New Models

Based on the examples above, a new model for how tissues cease to be vital emerged. The cessation of cellular activity in frozen tissue is fairly obvious—water freezes and cellular activity comes to a halt, including oxidative phosphorylation and glycolysis. With a shift in temperatures that is sufficiently rapid, the cessation of cellular processes is nearly immediate and the cell has little to no time to activate protective or apoptotic programs. There is a period of hypoxia/ischemia between removal of blood supply and freezing that does result in measurable alterations in biomolecules. In the case of fixation in a liquid the process is vastly different. The hypoxia/ischemia functionally has no end until the penetration of the fixative, which in the instance of larger fragments of tissue is long after metabolic activity has ceased. With the use of liquid fixatives, this is a function of diffusion, and with formalin is estimated at 1 mm/h. Fundamentally, the process is that of drowning, during which oxidative phosphorylation comes to an end quickly, as a result of a lack of oxygen, and then quickly switches to glycolysis, again which is limited by a lack of glucose. The cells activate their survival/apoptotic mechanisms. The model is well supported at the protein level by the alterations in the phosphorylation of protein. The original assumptions of RNA degradation in FFPE tissue were that the damage to the RNA would be “democratic” and random, without respect to location within a mRNA transcript, as would the crosslinking of proteins [4]. Rather, as demonstrated by the RT-PCR data, RNA integrity was greatest mid-transcript and diminished at both the 5′ and 3′′ ends of the mRNA due to the hypoxic related signals to remove the 5′ cap and degrade the 3′ poly-A tail [2, 9]. This model was supported by other independent data, and was critical toward a better understanding of optimal probe design for RNA obtained from FFPE tissue.

12. The Future of Formalin

Clearly diagnostic pathology continues to evolve. Although it is not entirely possible to speculate what the next substantial change in the preservation of tissue will be, the subject engenders the greatest investigation remains the replacement of formalin as the primary fixative of diagnostic pathology. The pressures include the fact that it is a carcinogen, costly to dispose of, and results in complex damage to DNA, RNA, and proteins. Bolstering the continued use of formalin is the substantial fund-of-knowledge, and some diagnosis which are dependent on formalin artifact. The concepts of what constitute the specification of formalin have been discussed in the preceding sections. The challenge remains the issues to be defined. Formalin could well live on, applied in a different manner. Alternatively formalin could be replaced with a new solution that offers benefits (less damage to biomolecules) and fewer drawbacks. The path forward will remain chaotic until a new approach arrives via consensus, of which a critical element will be the economic cost of the solution in a total-cost model accounting for reagent cost, disposal cost, staffing cost, and infrastructure demands.

13. Conclusions

The preservation of tissue for diagnostic purposes has followed a “fit-for-purpose” evolution in which changes in process and reagents have been driven by the diagnostic needs. The same basic process is used world-wide, supporting a singularity of diagnostic paradigms. Although the advancement of diagnostic tools and schema is readily apparent to pathologist, the slow evolution of the biospecimens they work with occurs largely unnoticed. The last shift in tissue preservation was the introduction of buffers in formalin, to reduce the presence of formalin pigment, but inadvertently improving the capacity to perform immunohistochemistry on tissue.

With the advancement of molecular biology and the capacity to evaluate DNA, RNA and proteins with greater precision, our understanding of tissue preservation has advanced moving beyond the simple models of chemistry of fixation. These models are driven by the demand that the specimen deliver more in the diagnostic space. Histo- and cyto-morphology, although not replaced, are now the starting point for such ancillary tools as immunohistochemistry, FISH, RNA in situ, and sequencing of RNA and DNA. Formalin fixation remains a subject of much debate—economic, robust, and a link to all pathology supports its continued use. Pressures for a new fixative are driven by formalin’s status as a carcinogen, and the chemical reactions that it undergoes with biomolecules, damaging our capacity to measure these biomolecules.

Concurrent with this, are ongoing research to improve tissue impregnation. Numerous alternative tissue impregnation methods have been demonstrated, but none have become dominant. Currently there are a lack of data on the variables associated with impregnation, other than that the removal of water is critical.

Ultimately tissue preservation will continue its constant, albeit slow, evolution, with changes in both fixation and impregnation, to result in a better biospecimen for diagnosis.