FFPE Sample Availability and US Standard Care

Researchers who work with human FFPE samples and other human biospecimens can sometimes lose track of the fact that the material they work with really does come from either living or dead human beings. As such, the availability of ethically collected samples in academic and commercial biorepositories is closely tied to the practical details of prevalence within the population of that disease indication, and US standard care procedures in clinical environments. Researchers who don’t remember this when they design their studies are setting themselves up for failure. It’s also important to be realistic about the availability of normal samples. In many cases, normal control samples of human tissue are harder and more expensive to procure than diseased samples. Why? Because normal tissue is not typically removed from living donors. This may sound obvious, but it’s amazing how many researchers forget this. For some types of tissue, fully consented postmortem samples may be the only way to procure “normal” examples, and for these samples it’s important to recognize that biobanks must rely on generous donations by donors or their families and that the average age of people who die tends to be much higher than the population at large. Additionally, and sometimes crucially, researchers need to also be realistic about postmortem Interval (PMI) times and the possible affects of agonal state and sparse clinical information for non-lethal diseases in postmortem donors. 

To minimize procurement times and maximize the number of cases received, researchers need to “do their homework” when it comes to placing realistic requests. There is no point in asking a biobank for >100 cases of an indication that only occurs in 1 in 1,000,000 individuals, and similarly there is no point in asking for whole tumor samples for cancer indications that are not treated surgically as part of US standard care. How can researcher’s find out what currently constitutes standard care? Actually the answer to this question is not as simple as it sounds, especially since a) standard care is always changing, b) not all clinics follow exactly the same guidelines, c) many biobanks contain “CAP graduated” archival specimens that were collected more than 10 years ago and therefore have the characteristics and attached data associated with standard care as it was at the time of collection, and d) it’s often surprisingly difficult for qualified life scientists who are not themselves medical doctors to be able to locate a succinct description of standard care, written in a way that makes sense to them. 

My advice to researchers who may not necessarily be medical doctors is to ALWAYS partner with a medical specialist or pathologist, and talk to them to get a clear idea of what constitutes standard care for the indication you are interested in. If for example, your medical partner tells you that Fine Needle Aspirates (FNAs) are ALWAYS used for that indication, but the indication is never treated surgically, then FNAs might be your best for acquiring samples. If you are interested in only acquiring samples that have a particular biomarker, but this biomarker is not tested for as part of standard care, then you are going to have to make sure that your budget includes the cost of paying for , or performing your OWN screening for ALL of your samples. Even worse news is that if the biomarker of interest only occurs in 10 percent of cases, then you are probably going to have to procure and pay to test at least 10 times as many samples as you actually need in order to find the ones of interest. Similarly, if chemotherapy ALWAYS precedes surgery for a particular indication, then you pretty much need to abandon any idea of procuring samples that were surgically excised before treatment. Remember that standard care, NOT your experimental interests will determine what samples are available.

If you are a life science researcher who is not lucky enough to have an experienced medical partner to help you predict sample availability based on standard care, you may have no choice other than to simply do a lot of googling to see what (or who) you can find to help you, but trust me, this is time well spent compared to the months or even years you might waste waiting for the procurement of sample types or numbers that are simply not compatible with the realities of what biobanks actually receive from clinics. 

Lab-Ally maintains a growing list of descriptions of standard care, and we would certainly appreciate additions, corrections and updates to this list that will help researchers to make meaningful predictions about sample availability.

New technique for improving staining in older FFPE sections.

The small biotech where I am employed frequently works with with FFPE specimens that are at least 10 or more years old, and 20+ year old specimens are not particularly unusual. Any histologist can tell you that the performance of older FFPE blocks and sections used in IHC and nucleic acid hybridization studies gradually degrades with time. The logical assumption is that the target molecules are decaying and losing their affinity for the probes sent to detect them. Various antigen retrieval techniques and signal-boosting protocols have been used to try to get older tissues to take up stain more efficiently, but their usefulness seems limited. We have taken a different approach and have been experimenting with ways to "restore" older FFPE tissue samples so that the specimen stains just as it did when it was first fixed. The process we are developing seems to improve staining in both IHC and ISH studies. It also seems to help optimize IHC studies such that a uniform staining protocol can be used with similar sections of any age. I can't say too much about the details of the proprietary process that we are still developing, but I can show you some before and after images to give you an idea of how effective it is.

The first pair of images below simply illustrates the problem. The image on the left was taken in 1996 and shows HPV in situ analysis. The image on the right is the same tissue analyzed with the same probes in 2012. Nothing controversial here, just a the typical loss of signal strength that most scientists would predict in 16 year old tissue.

(Click on the image to enlarge)

In the following pairs of images we see serial sections taken from tissue blocks of various ages that have been used for  either IHC or ISH studies. In each pair, the image on the left shows the staining one might typically expect from an old section or a section from an old block. The image on the right is a similar section that was pretreated with our proprietary process, then subjected to the exact same analysis as the section shown on the left.

(Click on the images to enlarge)

These results are quite striking. So striking, in fact, that we are not sure that we totally believe them ourselves. Although we have had the effectiveness of this process confirmed by at least one third-party laboratory, we are currently actively seeking additional organizations with experience in IHC to help us validate this procedure.

If you are a lab that is struggling with poor staining in older slides, please contact me to see if we can enter into some kind of mutually beneficial collaboration. Ideally, we would set up a carefully controlled blind study where you send us pairs of problematic slides, we would treat one of the pair and send them back to you. We won't tell you which one we treated. You then stain and examine them as normal, and let us know if you see a distinct difference between them and also with controls kept at your location. Naturally there would be no charge for this. Ultimately we may be interested in commercializing this process, but first we need to be absolutely sure that it really works the way we think it does.

Separating out different color wavelengths to better understand target distribution

The nuance multispectral imaging system from PerkinElmer is pretty cool. Basically it's just a camera with some spectral filters that sits on top of your microscope, and some smart software that can make sense of the captured image. The system lets you isolate the color signal of specific stains, filter out the noise that comes from autofluorescence, figure out which cells have taken up which stain, and impartially calculate the strength of the staining response. Best of all it can use some wavelength addition and subtraction calculations to highlight areas that have taken up more than one stain. 

In a recent study conducted in our CLIA certified lab, we used this system to show co-localization of a particular protein and the mRNA transcript from which that protein was translated. The 4 images below are of the same region of a human prostate section. IHC was used to stain a smooth muscle cytoskeletal protein called "smoothelin". ISH was then used to stain the same material so that we can see the mRNA transcripts from which the protein was translated. 

The first image shows the presence of antibody stained protein as brown, the blue stain is the mRNA. Note the localization of the mRNA signal to the stroma (long arrow) and the lack of signal in the epithelia (short arrow). In the first image (brightfield microscopy), it's sort of hard to see if the brown and the blue stains are really tightly co-localized.

In the next pair of images, fluorescence microscopy is used to separately show the smoothelin protein (green) and the mRNA transcript (blue).

Finally, in the last image, the nuance system helps us to see BOTH types of staining simultaneously, with the dual stained regions shown as yellow. This makes the co-expression much easier to see. For more information on the Nuance system here.

In-situ hybridizations and FFPE slides.

Here's an example of the difference between what you could do with an FFPE, H&E slide 15 years ago versus today. This illustrates WHY archives of FFPE tissue have become so much more useful in recent years. In this example we are looking at a sagittal section of an embryonic mouse:

Abbreviations: Br – brain; DMO – dorsal region medulla oblongata; IC – internal capsule; K – kidney; Li – liver with blood vessels; Lu – lung; Ma – mandible; NCh – nasal chamber; NPhD – nasopharangeal duct; PhU – phallic part of urethra; (s) – sense. 

So, you can see that the mouse has cells, tissues, organs, the usual, yup, it's definitely a mouse, and honestly that's about all you can tell from the above image.

Now imagine you are interested in a specific gene, and you want to know where that gene is most actively expressed. Today you can do that using a technique called in situ hybridization (ISH).  The image below was taken from a study of a gene I will simply call "GENE X"  which is believed to code for a transcription factor protein. In humans, GENE X mutations cause severe diseases including neonatal diabetes mellitus.  The purpose of this ISH analysis was to localize GENE X mRNA at the anatomical and cellular levels in order to help scientists learn something of its likely function.  The formaldehyde fixed sections were mounted on glass slides then hybridized with 35S-labeled cRNA probes. Patterns of gene expression were analyzed by x-ray film autoradiography and  darkfield illumination. The image below is of the same section as the one above.

The results provided evidence for the presence of GENE A mRNA at various concentrations in several tissues including specific brain regions, structures related to the nasal chamber, lung, liver blood vessels, kidney, ureter and pancreas. In the pancreatic islets, the hybridization intensity was comparable to that of the kidney cortex. I won't bore you with all the details of what this experiment suggests, but the point here is that now the scientist has the option to ask questions about what specific genes might be doing in specific places and at specific times during the development of the organism. Pretty neat huh?

Why are FFPE specimens in demand?

Thanks to an amazingly productive half century of molecular biology research, sensitive new techniques are now available that allow scientists to extract useful information from small samples of preserved tissue. Despite the fact that most pathology specimens have been chemically altered by fixation in formalin and subjected to a range of other processes including paraffin embedding and staining protocols, it turns out that informative molecules such as DNA and RNA as well as various antigens and their epitopes are often still present and detectable in both FFPE tissue blocks and cut sections. This makes it possible to use use existing biobanks and pathology lab samples as a source of research material that can help scientists test theories about the molecular characteristics and causes of disease. Additionally, thanks to the cumulative size of modern biobanks, scientific findings can be supported or refuted by statistically significant examination of multiple cases, without having to subject living human subjects to intrusive lab work beyond their initial treatment or biopsy. Individually, archival FFPE blocks are rarely valuable for research, but when they are gathered together into meaningful cohorts that represent many cases of specific indications, and supplied along with attached clinical data and similarly sized cohorts of normal cases, they become an incredibly useful research tool. This is especially true when these cohorts are created by specialized organizations that can screen the blocks for quality and de-identify them as required by 45 CFR 46 US regulations.

One particularly active area of research right now is the investigation of biomarkers on the surface of cancer cells. These biomarkers can that tell scientists something about the mechanisms of oncogenesis and metastasis, and when coupled with outcome information from medical records, they can even be used to predict patient prognoses and likely responses to specific treatments. Although the number of these useful biomarkers is growing rapidly, as yet, only a relatively small proportion are actually used to personalize therapeutic strategies in clinical practice.

Making use of old samples

OK, so there's a huge global stockpile of preserved tissue that's been sitting around unused for decades. Surely all those specimens should have some value to the scientific research community? Enter Maurice Wilkins, Rosalind Franklin, James Watson, and Frances Crick. Quite suddenly, the world changed. For anyone who has bothered to read this far, I'm sure it is not necessary to belabor the enormity of the biotechnology revolution that began with Watson and Crick's collaboration and the famous educated guess that revealed to us the exquisite structure of the DNA molecule. The full ramifications of the half century of productive research that followed will be not truly be understood for generations to come, but if we survive, future societies will look back and wonder what it must have been like for us to actually be present, here at this time, witnessing the beginnings of a seismic change in our species, our culture and our understanding of ourselves and our origin. In subsequent posts, I'll ponder what the DNA-driven revolution in molecular and cellular biology might mean for the far older science of histology.

There's one additional DNA-trivia-tidbit-aside I will throw in here that may be of interest to some. In the mid 1980's I completed my undergraduate degree at King's College London. I was very, very lucky to meet and take some classes with Dr Maurice Wilkins who, along with Rosalind Franklin, produced x-ray data that was used by Watson and Crick to inform their guess at the structure of DNA. Dr Wilkins was friendly and approachable with boyish charm and sharp wit. He indulged my naive questions and I spent many hours talking with him about the nature and philosophy of science, the rapid progress of molecular biology and the scientific revolutions of the 20th century, as well as what it all might mean for humanity. At that time, old-school British academics like Maurice worked on a totally different time scale to the rest of us. Even though there were files on his desk dated 1963 (marked "urgent"), I got the sense that his mind was always racing centuries ahead. He seemed to simultaneously live in the past and the future, as if some sort of distortion field isolated him from the passage of normal time. His mind swirled with questions to be answered, and the fact that it might take humanity centuries to answer these questions never seemed to bother him. I was inspired. Maurice taught me that scientists need to have faith that others will follow in their footsteps. Just because a question seems complex and hard to answer, that's no reason not to get started working on it right away. For some, this might might entail spending an entire professional career inventing tools that may some day enable others to continue a journey that can never be completed in one scientist's lifetime. Science is a highly rationale problem-solving strategy, but like any journey into the unknown, it also involves a leap of faith. I was saddened to hear of Dr Wilkin's death in 2004. I have lots fond memories of him I can share if anyone is interested.

 Maurice Wilkins 1916 -2004

FFPE specimens

The process of creating formalin fixed paraffin embedded (FFPE) tissue and and fixed, H&E stained slides has been around for over a hundred years. It's a standard tool used by clinicians, pathologists and researchers in many life-science fields. Fixed material is stable at room temperature and can be conveniently stored for long periods. Since these techniques has been around for a while, hospitals and research labs all over the world have tended to accumulate huge archives of fixed material containing many millions of specimens and dating back many decades. In hospitals, this can include tissue removed as biopsies, resected material removed from patients during surgery and samples collected post-mortem. Hospitals keep these samples as a way to keep track of patient histories, monitor long-term changes of a patient's condition, ensure and improve the accuracy of diagnosis of disease and to meet legal record-keeping obligations. Only a tiny proportion of the massive global stockpile of preserved tissue has ever been used for research purposes, and for smaller community hospitals with no research labs on site, we can assume that virtually none of the tissue collected yields any research benefit to anyone once it's clinical purpose has been fulfilled. After a certain amount of time has elapsed, some organizations simply dispose of old samples, or move them to long-term storage facilities, never again to see the light of day.  Until recently, the research value of old, preserved tissue specimens seemed dubious anyway, since preserved tissue is biologically inactive and samples are of sufficiently small size as to make the extraction of useful biochemicals impractical using traditional approaches.