The Burdens of “Do-It-Yourself” Complex Flow Cytometry for Routine Testing

“Difficulty is the excuse history never accepts.”    Edward R. Murrow

“In the middle of difficulty lies opportunity.”    Albert Einstein

“Quality is never an accident; it is always the result of high intention, sincere effort, intelligent direction and skillful execution; it represents the wise choice of many alternatives.”   William A. Foster

Designing a reproducible flow cytometry assay for complex cellular analysis involves a lot of creativity and is both fun and challenging.  To be successful, all projects require good planning for the short and long term.  For example, many people have done the simple DIY home project of painting a room a new color. This is a straightforward project which still requires planning for the short term to get it completed, and for the long term of being able to do the exact color again elsewhere.   The immediate plan is to pick out the color, purchase the proper tools and the paint, prepare the room, paint, clean up, and finally, re-assemble the room.  The project can be completed successfully in a very short time and can be repeated accurately due in large part to the purchase of the commercial tools and pre-mixed colored paint from reputable specialists.  Even craftsmen with the abilities to make their own painting tools and mix their own paint colors insist on purchasing these critical components.  Flow cytometry assays are no different and ImmunoSite Technologies provides you with the critical reagent components so that you avoid the risks and spend your time doing what you really want to be doing.

The difficult challenges to produce reproducible high quality complex flow cytometry data are many, not the least of which are the choices of detection reagents [1, 2].  The specialized testing needs inherent in clinical research trial studies prevent manufacturers from being able to produce standard (off the shelf) complex testing reagents because of the high cost associated with these relatively small markets.  As a result, many manufacturers offer custom reagent programs to help meet the needs of users for special reagents.

Depending on the reagent provider, the custom services range from providing a titrated single color conjugate without any claims, to full design consultation expertise and providing complex mixtures of reagents with stability and performance data.  Ultimately, however, regardless of the manufacturer the end user must adopt the attitude of never trusting their reagents or instruments.  Like a manufacturer they must test carefully and often.  Simply put, end users must take the time to perform the same level of design control and manufacturing quality control that manufacturers of reagents do for their products and that we have grown accustomed to trusting, thus assuring accurate and precise results over time.

There are many instructional websites, textbooks [3] and key publications on how to establish optimal instrument [4, 5, 6, 7 ] and assay performance [8, 9, 10, 11], and data analysis strategies [12, 13].  Also, manufacturers have made increasingly robust instruments and software to help users perform complex setups and track performance, so these topics will not be discussed in detail here.

FLUOROCHROMES:

Historically, reagent design was simple due to the relatively low number of well characterized fluorochromes, dyes and stains.  However, as the technique has grown and progressed, the need for more sensitive and simultaneous measurements has presented many problems.  Several of the dyes were not optimal choices for sensitive fluorescent measurements because of their low quantum yield (e.g. Pacific Orange, Pacific Blue), but users were desperate for choices.

Tandem dye technology was a giant step forward for complex flow cytometry by increasing the possibility of dye choices, but this added a lot of color compensation complexity due to the <100% energy transfer, and large variability between manufacturer’s and potentially between manufactured lots of tandem dye and conjugates.  For years, even the addition of tandem dyes still only provided a relatively small number of molecules with only fairly well defined interactions and idiosyncrasies.  Meanwhile, the analytical instrumentation was gaining rapidly in complexity. The new instruments were capable of much more; they had dramatically increased sensitivity resulting in bringing back the use of previously unappealing dyes.

Nanoparticles (e.g. Qdots) [14] and small derivative dyes (e.g. Alexas, DyeLights, etc.) were a huge leap forward for increasing the wavelength choices for users.  The nanoparticles fluoresce in a very narrow spectrum and are very appealing but are still relative newcomers to the field and need many more years of use to identify interactions and pitfalls.

The current offering of flow cytometry instrumentation is available in many configurations starting with simple 2-color closed systems that operate almost as a “push-button” system. Next we find clinical systems of 2-10 colors that have automated setup, QC, and reporting. On the cutting edge we find very powerful 10-30+ color systems with a high degree of flexibility.

Any instrument platform can be used as a standardized system for testing if, and only if, the users are aware of the many QC components and points, and ruthlessly maintain the system as defined and validated for specific testing needs.  For a defined “Clinical Trial” or “Validated Test or Procedure”; once these points have been adhered to; it is not acceptable to run down the hall to use the “same” kind of instrument if the qualified instrument malfunctions.  If a backup instrument is to be designated, prior to using it for sample testing, the two instruments must be cross-qualified to each other with rigorous testing, and then maintained in that configuration. This is a major problem for a majority of clinical research labs operating on tight budgets.

Pitfalls:

The very large number of fluorescent molecules currently available in the market place, along with friendlier flow cytometers, means that many labs feel comfortable putting together reagents for their own use.  Often, this practice has led to data that cannot be reconciled across a single location study let alone across multiple sites.  One of the problems is that we have grown comfortable with the high quality single color reagent products, and trust that when they are combined, the high quality will remain.

There is enormous complexity inherent with designing reagent combinations, and there are restrictions on manufactured products.  Manufacturers make very specific claims about a product’s specificity, performance and stability which is clearly described in the product package inserts.  Nowhere will you find a statement on a package insert that indicates a reagent can be cocktailed with “x” reagent and kept for up to “x” period of time.  If they could easily make those claims, then they would manufacture the combination themselves.

In reality, once the user removes a reagent from the bottle in which it was delivered, all bets are off for stability and performance.  One cannot and must not make the assumption that a single home-brew lot that has been tested for specificity, accuracy, precision, linearity, and sensitivity, will be the same as the next lot of home-brew if made from the same or different lot numbers of the component reagents.  Manufacturers have a very specific QC test for each reagent prior to releasing it to the market, however it most likely has nothing to do with the final users intended use (e.g. a QC release test could be on a cell line without lysing reagent, and not diseased whole blood with a particular lyse).  Pay attention to the claims.  A second problem is the misconception that PBS is a suitable dilution buffer when creating cocktails of reagents.  Every reagent company uses proprietary buffers for their reagents to control for non-specific binding and maintaining stability of the antibody and dye performance.

Batch mode testing of study samples can be done very well at well run medical centers in academic (non-industry) settings and has been shown to provide very good results because the testing is completed in a relatively short time period (less than 6 months) with a very defined set of reagents, one set of dedicated and calibrated equipment, and one set of trained operators.  Major problems arise for long term studies that either cannot be batch analyzed or must be analyzed in real-time, unless enormous effort and funding are put into exceptional standardization of all aspects of testing [15].

Maintaining quality over time without drift is the key reason for the highly regulated state of testing performed in hospital and clinical diagnostic labs and CROs.  Performance testing against standards and other labs is critical to maintain the test results as guaranteed.

As an example of consistent data over time, would anyone doubt the data from a series of hematology CBC test results performed on a person from birth through their 80’s?  Probably not, since hematology testing is a tried and true technology which is highly regulated, defined, controlled, and all labs participate in Inter-lab Quality Assurance Programs.  Can the same be said for absolute count of CD4 cells?  Not quite yet because of the variety of non-regulated products in use in many labs.  What about a functional cellular assay measuring the suppressive effect of T-reg cells?   Yes, these test results can be full of doubt because of the enormous complexity in standardizing functional assays across platforms and labs, and the current lack of any large IQAP programs.

Use and Design:

Using panels of poly-color marker combinations requires careful attention to the basic concerns that all users are trained in performing (dose, fluorochromes, instrument configuration, antigen density, steric factors).  In designing any cocktailed reagents there is a high degree of expertise needed to consider dye interactions during the actual staining and analysis, as well as the dye stability in mixtures pre- and post-staining.  Some labs prefer to make the cocktails fresh each day, and this may appear to suffice for a short term study.  It is, however, impractical because it introduces errors, adds time, and requires lot-to-lot testing of every reagent in combination with others which can exceed the volume available of the actual sample to be tested.

Color Compensation:

Proper color compensation requires the largest effort and is the most often overlooked for flow cytometric analysis.  “One size may not fit all.”

The potential lot-to-lot variations in the not-tried-and-true fluorochromes, homemade conjugates, and manufacturer differences in tandem dye efficiencies leads to mandatory compensation on a tube-by-tube (well-by-well) basis.  The ideal, and never (rarely) used, color compensation procedure requires the use of the single color stains using the same reagent, as used in the actual combination, for every fluorochrome and marker.  With advanced planning and careful characterization of the reagents, one might be able to use a handful of single color reagents that are representative of the actual reagents in the test samples.  This still may necessitate off-line compensation of the samples which needs operator intervention.  This increases error and decreases throughput.

The best solution to routine testing is to use rigorously tested commercial products that are manufactured under FDA regulations.  This isn’t possible in most clinical research testing since the reagents desired are “new” and specialty markers haven’t been through the 2-5 year process of FDA approval, and for which cocktails certainly are not available in pre-formulated vials from producers.

So Many Reagent Choices

So what is the solution for clinical research testing?  The minimum is a carefully controlled Stability Testing procedure on the Concentrated Conjugates and the pre-mixed Cocktails.  This includes testing of at least the Closed Vial (i.e. never opened) and the Opened Vial (exposure to everyday testing conditions of time, temperature, and light) in real time for all performance criteria including color compensation requirements.

Also, the incoming specifications for every reagent must be defined and every incoming lot tested against the specifications prior to use in a study.  In addition, reagents must be routinely tested on known standards to maintain quality assurance as to all performance parameters.