ILA24-A Fluorescence Calibration and Quantitative Measurement of Fluorescence Intensity; Approved Guideline

I/LA24-A Vol. 24 No. 26 Replaces I/LA24-P Vol. 23 No. 31 Fluorescence Calibration and Quantitative Measurement of Fluorescence Intensity; Approved Guideline

This guideline describes the basic principles, reference materials, and laboratory procedures upon which quantitative fluorescence calibration is based. A guideline for global application developed through the NCCLS consensus process.

NCCLS... Global Consensus Standardization for Health Technologies

NCCLS is an international, interdisciplinary, nonprofit, Most NCCLS documents are subject to two levels of standards-developing, and educational organization that consensus—“proposed” and “approved.” Depending on the promotes the development and use of voluntary need for field evaluation or data collection, documents may consensus standards and guidelines within the healthcare also be made available for review at an intermediate community. It is recognized worldwide for the consensus level. application of its unique consensus process in the Proposed An NCCLS consensus document undergoes the development of standards and guidelines for patient first stage of review by the healthcare community as a testing and related healthcare issues. NCCLS is based on proposed standard or guideline. The document should the principle that consensus is an effective and cost- receive a wide and thorough technical review, including an effective way to improve patient testing and healthcare overall review of its scope, approach, and utility, and a line- services. by-line review of its technical and editorial content. In addition to developing and promoting the use of Approved An approved standard or guideline has achieved voluntary consensus standards and guidelines, NCCLS consensus within the healthcare community. It should be provides an open and unbiased forum to address critical reviewed to assess the utility of the final document, to issues affecting the quality of patient testing and health ensure attainment of consensus (i.e., that comments on care. earlier versions have been satisfactorily addressed), and to PUBLICATIONS identify the need for additional consensus documents. An NCCLS document is published as a standard, NCCLS standards and guidelines represent a consensus guideline, or committee report. opinion on good practices and reflect the substantial agreement by materially affected, competent, and interested Standard A document developed through the consensus parties obtained by following NCCLS’s established process that clearly identifies specific, essential consensus procedures. Provisions in NCCLS standards and requirements for materials, methods, or practices for use guidelines may be more or less stringent than applicable in an unmodified form. A standard may, in addition, regulations. Consequently, conformance to this voluntary contain discretionary elements, which are clearly consensus document does not relieve the user of identified. responsibility for compliance with applicable regulations. Guideline A document developed through the COMMENTS consensus process describing criteria for a general operating practice, procedure, or material for voluntary The comments of users are essential to the consensus use. A guideline may be used as written or modified by process. Anyone may submit a comment, and all comments the user to fit specific needs. are addressed, according to the consensus process, by the NCCLS committee that wrote the document. All comments, Report A document that has not been subjected to including those that result in a change to the document when consensus review and is released by the Board of published at the next consensus level and those that do not Directors. result in a change, are responded to by the committee in an CONSENSUS PROCESS appendix to the document. Readers are strongly encouraged to comment in any form and at any time on any NCCLS The NCCLS voluntary consensus process is a protocol document. Address comments to the NCCLS Executive establishing formal criteria for: Offices, 940 West Valley Road, Suite 1400, Wayne, PA • the authorization of a project 19087, USA. • the development and open review of documents VOLUNTEER PARTICIPATION • the revision of documents in response to comments Healthcare professionals in all specialties are urged to by users volunteer for participation in NCCLS projects. Please contact the NCCLS Executive Offices for additional • the acceptance of a document as a consensus information on committee participation. standard or guideline.

I/LA24-A ISBN 1-56238-543-7 Volume 24 Number 26 ISSN 0273-3099 Fluorescence Calibration and Quantitative Measurement of Fluorescence Intensity; Approved Guideline

Gerald E. Marti, M.D., Ph.D. Robert F. Vogt, Jr., Ph.D. Adolfas K. Gaigalas, Ph.D. Craig S. Hixson, Ph.D. Robert A. Hoffman, Ph.D. Rodica Lenkei, M.D., Ph.D. Louise E. Magruder, B.S. Norman B. Purvis, Jr., Ph.D. Abe Schwartz, Ph.D. Howard M. Shapiro, M.D. Alan Waggoner, Ph.D.

Abstract

Quantitative fluorescence calibration (QFC) is an empiric system to calibrate fluorescence intensity in a way that preserves stoichiometry between the concentration of fluorochrome in solutions and the equivalent molar quantity of fluorochrome on stained measurands such as cells, gels, microspheres, and microdots. This guideline describes the basic principles, reference materials, and laboratory procedures upon which QFC is based. This guideline is intended for use with reference materials and procedures developed under the National Institute of Standards and Technology (NIST) Fluorescence Intensity Standards program. While the general principles of QFC apply to any fluorescence measurement, this guideline specifically addresses analysis of cells and microspheres by flow cytometry, including cellular immunophenotyping and suspension array technology. The current and emerging uses of these laboratory methods will have an increasing impact on public health and primary care, from large-scale screening of populations to the individual profiling of each patient’s disease.

NCCLS. Fluorescence Calibration and Quantitative Measurement of Fluorescence Intensity; Approved Guideline. NCCLS document I/LA24-A (ISBN 1-56238-543-7). NCCLS, 940 West Valley Road, Suite 1400, Wayne, Pennsylvania 19087-1898 USA, 2004.

THE NCCLS consensus process, which is the mechanism for moving a document through two or more levels of review by the healthcare community, is an ongoing process. Users should expect revised editions of any given document. Because rapid changes in technology may affect the procedures, methods, and protocols in a standard or guideline, users should replace outdated editions with the current editions of NCCLS documents. Current editions are listed in the NCCLS Catalog, which is distributed to member organizations, and to nonmembers on request. If your organization is not a member and would like to become one, and to request a copy of the NCCLS Catalog, contact the NCCLS Executive Offices. Telephone: 610.688.0100; Fax: 610.688.0700; E-Mail: [email protected]; Website: www.nccls.org

Number 26 NCCLS

This publication is protected by copyright. No part of it may be reproduced, stored in a retrieval system, transmitted, or made available in any form or by any means (electronic, mechanical, photocopying, recording, or otherwise) without prior written permission from NCCLS, except as stated below.

NCCLS hereby grants permission to reproduce limited portions of this publication for use in laboratory procedure manuals at a single site, for interlibrary loan, or for use in educational programs provided that multiple copies of such reproduction shall include the following notice, be distributed without charge, and, in no event, contain more than 20% of the document’s text.

Reproduced with permission, from NCCLS publication I/LA24-A—Fluorescence Calibration and Quantitative Measurement of Fluorescence Intensity; Approved Guideline (ISBN 1-56238-543-7). Copies of the current edition may be obtained from NCCLS, 940 West Valley Road, Suite 1400, Wayne, Pennsylvania 19087-1898, USA.

Permission to reproduce or otherwise use the text of this document to an extent that exceeds the exemptions granted here or under the Copyright Law must be obtained from NCCLS by written request. To request such permission, address inquiries to the Executive Director, NCCLS, 940 West Valley Road, Suite 1400, Wayne, Pennsylvania 19087-1898, USA.

Suggested Citation

(NCCLS. Fluorescence Calibration and Quantitative Measurement of Fluorescence Intensity; Approved Guideline. NCCLS document I/LA24-A [ISBN 1-56238-543-7]. NCCLS, 940 West Valley Road, Suite 1400, Wayne, Pennsylvania 19087-1898 USA, 2004.)

Proposed Guideline November 2003

Approved Guideline August 2004

ISBN 1-56238-543-7 ISSN 0273-3099 ii

Volume 24 I/LA24-A

Committee Membership

Area Committee on Immunology and Ligand Assay

Dorothy J. Ball, Ph.D. Ronald J. Whitley, Ph.D. Robert F. Ritchie, M.D. Chairholder University of Kentucky Med. Ctr. Foundation for Blood Research Abbott Laboratories Lexington, Kentucky Scarborough, Maine Irving, Texas Advisors Donald R. Tourville, Ph.D. W. Harry Hannon, Ph.D. Zeus Scientific, Inc. Vice-Chairholder Kaiser J. Aziz, Ph.D. Raritan, New Jersey Centers for Disease Control and Food and Drug Administration, CDRH Prevention Rockville, Maryland Daniel Tripodi, Ph.D. Atlanta, Georgia The Sage Group Linda Ivor Bridgewater, New Jersey Joan H. Howanitz, M.D. Gen-Probe Inc. SUNY Brooklyn San Diego, California Robert W. Veltri, Ph.D. Brooklyn, New York Johns Hopkins Hospital Gerald E. Marti, M.D., Ph.D. Baltimore, Maryland Marilyn M. Lightfoote, M.D., Ph.D. Food and Drug Administration, CBER Food and Drug Administration, CDRH Bethesda, Maryland Robert F. Vogt, Jr., Ph.D. Silver Spring, Maryland Centers for Disease Control and Robert M. Nakamura, M.D. Prevention Robin G. Lorenz, M.D., Ph.D. Scripps Clinic & Research Foundation Atlanta, Georgia University of Alabama at Birmingham La Jolla, California Birmingham, Alabama Philip R. Wyatt, M.D., Ph.D. Thomas A. O’Brien, Ph.D. North York General Hospital Per N. J. Matsson, Ph.D. Ortho Biotech Products LP North York, Ontario, Canada Pharmacia and Upjohn Diagnostics Bridgewater, New Jersey Uppsala, Sweden

Subcommittee on Fluorescence Calibration and Quantitative Measurement of Fluorescence

Gerald E. Marti, M.D., Ph.D. Louise E. Magruder, B.S. Michael Borowitz, M.D., Ph.D. Co-Chairholder Food and Drug Administration, Johns Hopkins Medical Institutions Food and Drug Administration, CDRH Baltimore, Maryland CBER Rockville, Maryland Bethesda, Maryland Charles W. Caldwell, M.D., Ph.D. Norman B. Purvis, Jr., Ph.D. University of Missouri Robert F. Vogt, Jr., Ph.D. Esoterix, Incorporated Columbia, Missouri Co-Chairholder Brentwood, Tennessee Centers for Disease Control and Nina M. Chace, M.S. Prevention Abe Schwartz, Ph.D. Food and Drug Administration, Atlanta, Georgia Center for Quantitative Cytometry CDRH San Juan, Puerto Rico Rockville, Maryland Adolfas K. Gaigalas, Ph.D. National Institute of Standards and Howard M. Shapiro, M.D. Lauren A. Ernst, Ph.D. Technology West Newton, Massachusetts Carnegie Mellon University Gaithersburg, Maryland Pittsburgh, Pennsylvania Alan Waggoner, Ph.D. Craig S. Hixson, Ph.D. Carnegie Mellon University Shawn P. Fay, Ph.D. Bio-Rad Laboratories Pittsburgh, Pennsylvania Princeton, New Jersey Benicia, California Advisors Jan W. Gratama, M.D., Ph.D. Robert A. Hoffman, Ph.D. Daniel den Hoed Cancer Center BD Biosciences C. Bruce Bagwell, M.D., Ph.D. Rotterdam, The Netherlands San Jose, California Verity Software House, Inc. Topsham, Maine L. Omar Henderson, Ph.D. Rodica Lenkei, M.D., Ph.D. Centers for Disease Control and CALAB Research Dorothy J. Ball, Ph.D. Prevention Stockholm, Sweden Abbott Laboratories Atlanta, Georgia Irving, Texas

iii Number 26 NCCLS

Advisors (Continued) Phillip Poncelet, Ph.D. James Weaver, Ph.D. Biocytex Food and Drug Administration, Ben Hunsberger Marseille, France CDER Verity Software House, Inc. Laurel, Maryland Topsham, Maine Jorge Quintana, Ph.D. Beckman Coulter, Inc. Vince Zenger, Ph.D. Anne A. Hurley, Ph.D., CCRA Miami, Florida Food and Drug Administration, Comprehensive Cytometric Consulting CDER Hanover, Massachusetts Larry Seamer, M.T.(ASCP) Washington, District of Columbia Bio-Rad Laboratories Louis A. Kamenstky, Ph.D. Hercules, California Staff Compucyte Corporation Cambridge, Massachusetts Maryalice Stetler-Stevenson, M.D., Lois M. Schmidt, D.A. Ph.D. Staff Liaison Francis Mandy, Ph.D. National Institutes of Health NCCLS Bureau of Laboratories and Research Bethesda, Maryland Wayne, Pennsylvania Services, Health Canada Ottawa, Ontario, Canada Carleton C. Stewart, Ph.D. Donna M. Wilhelm Roswell Park Cancer Institute Editor Katharine A. Muirhead, Ph.D. Buffalo, New York NCCLS SciGro, Inc. Wayne, Pennsylvania Malvern, Pennsylvania Robin Thorpe National Institute for Biological Melissa A. Lewis Janet K.A. Nicholson, Ph.D. Standards and Control Assistant Editor Centers for Disease Control and Herts, United Kingdom NCCLS Prevention Wayne, Pennsylvania Atlanta, Georgia

Acknowledgement

NCCLS and the Subcommittee on Fluorescence Calibration and Quantitative Measurement of Fluorescence Intensity gratefully acknowledge the following colleagues for their important contributions to the development of this NCCLS document: Dr. Lili Wang (National Institute of Standards and Technology) for her role in developing much of the method and theory for fundamental fluorescence calibration included herein; Dr. Michael Edidin (Johns Hopkins University) for his incisive review of the proposed guideline; Dr. Bruce Davis (Trillium Diagnostics, LLC, and the Maine Medical Center Research Institute) for the original methods and data in Section 11.4.6; the late Dr. Janis Giorgi (UCLA School of Medicine) for her unwavering commitment to developing this technology; and Dr. Harry Hannon (Centers for Disease Control and Prevention) for his leadership and patience as Chairholder of the NCCLS Area Committee on Immunology and Ligand Assay during the evolution of this guideline.

Volume 24 I/LA24-A

Contents

Abstract...... i

Committee Membership...... iii

Foreword...... vii

1 Scope...... 1

2 Introduction...... 1

3 Standard Precautions...... 2

4 Terminology...... 2 4.1 Definitions ...... 2 4.2 Acronyms and Abbreviations ...... 7 5 Basic Principles of Measuring Fluorescence Intensity ...... 8 5.1 Optical Radiation, Fluorescence, and Radiometry...... 8 5.2 Factors That Influence Fluorescence Intensity Measurements ...... 9 6 Instruments for Measuring Fluorescence Intensity...... 16 6.1 General Components...... 16 6.2 Specific Platforms for Measuring Fluorescence ...... 18 6.3 Fluorometer Evaluation and Performance Criteria for FI Measurements...... 22 7 Quantitative Fluorescence Calibration for Conjugates and Particles Under Varying Measurement Conditions...... 25 7.1 Comparisons Based on the Fluorescence Yield...... 25 7.2 Fluorescence Yield of Particle Suspensions (Microspheres and Cells)...... 26 8 Solution-Based Fluorochrome Reference Materials for QFC...... 27 8.1 Purpose and Criteria for Fluorochrome Reference Solutions ...... 27 8.2 Property Values for Fluorochrome Reference Solutions ...... 28 8.3 Other Specifications for Fluorochrome Reference Solutions ...... 28 8.4 General Chemical Classes of Fluorochromes ...... 29 8.5 Reference Solutions for Organic Dye Fluorochromes...... 29 8.6 Reference Solutions for Protein Fluorochromes...... 31 8.7 Reference Solutions for Specialty Fluorochromes...... 33 9 Solid Phase Fluorochrome Reference Materials for QFC...... 34 9.1 Purpose and Criteria for Solid Phase Fluorochrome Reference Materials ...... 34 9.2 Microspheres as Solid Phase Fluorochrome Reference Materials...... 35 9.3 Property Values for Fluorochrome Reference Microspheres ...... 36 9.4 Other Specifications for Fluorochrome Reference Microspheres...... 38 9.5 Sources of Fluorochrome Reference Microspheres...... 38 10 Quantitative Ligand-Binding Assays Using Fluorochrome-Ligand Conjugates (FLC) ...... 39 10.1 The Relationship Between Equivalent Fluorescence and FLC Binding Values...... 39 10.2 Molecular Properties That Influence Fluorochrome-Ligand Conjugate Binding ...... 39 10.3 Quantifying the Molar Quantity of Fluorochrome-Ligand Conjugates Bound to Particles...... 40

v Number 26 NCCLS

Contents (Continued)

10.4 Sources of Variability in Measuring FLC Binding Values...... 41 10.5 Using QFC to Measure Binding Constants and Affinities...... 42 11 Practical Methods for Quantitative Fluorescence Calibration in Clinical and Research Applications of Cytometry...... 43 11.1 Basic Reference Method for Constructing the FI Calibration Curve...... 43 11.2 The FI Calibration Plot ...... 43 11.3 Evaluation of Calibration Parameters ...... 47 11.4 Laboratory Approaches Using QFC for Evaluating Instrument Performance and Measuring Expression of Cellular Receptors ...... 50 References...... 56

Appendix A. An Overview of Molecular Fluorescence...... 61

Appendix B. Selected Excerpts From Papers in the Journal of Research of the National Institute of Standards and Technology...... 63

Appendix C. NIST Certificate for Standard Reference Material SRM 1932...... 68

Appendix D. NIST Fluorescein Microsphere Reference Material RM 8640 ...... 73

Appendix E. Measurement of Fluorescence Intensity Detection Efficiency (Q) and Optical Background (B) on a Flow Cytometer...... 77

Appendix F. A Classification System of Microsphere Fluorescence Standards Used for Flow Cytometry ...... 81

Summary of Consensus/Delegate Comments and Committee Responses...... 82

The Quality System Approach...... 86

Related NCCLS Publications...... 87

Volume 24 I/LA24-A

Foreword

Over the last four decades, fluorescence has emerged as the most useful signal in biomedical science. Fluorochromes are the only labels that combine the sensitivity of radioactive tracers with the capacity for direct visualization at the submicron level. Fluorescence measurements can be made on bands, microspheres, microdots or cells in gels, suspensions, planar arrays, or three-dimensional tissue slices. The fluorescent antibody technique first developed by Coons in the 1940s1,2 is today the basis of the most commonly used assays in the biomedical laboratory. Fluorescence microscopy remains an important tool in biology, medicine, and public health. Its heir apparent, fluorescence cytometry, is a continuing wellspring of technical advances and biomedical applications. Fluorescence imaging techniques have been harnessed to analyze high-density microarrays and “chips,” each with tens of thousands of targets. As of the publication of this guideline, multiplexed fluorescence measurements are capable of resolving dozens of populations in a single suspension of microspheres, each with its own target; emerging platforms will allow simultaneous analysis of suspensions with hundreds or thousands of targets. The high-capacity, low-cost capability of these technologies for genetic and proteomic analysis presents a profound opportunity for improving human health, from large-scale screening of populations to the individual profiling of each patient’s disease.

Because of its many advantages (including biosafety), fluorescence has largely supplanted radioactivity as the label of choice. However, fluorescence does not provide the same direct measure of molar quantity that radioactivity does. Units of radioactivity such as curies convert readily to molar quantities because nuclear emission is spontaneous, its decay is predictable, and it is unaffected by the chemical environment. In contrast, fluorescence requires a source of excitation, its decay is variable, and its intensity is influenced by many environmental factors including temperature, hydrophobicity, concentration, and photodegradation. While this dependence provides valuable probes for the biomolecular environment, it complicates the relationship between fluorescence signals and the molar quantity of fluorochrome. The problem of molar equivalence is even more difficult when fluorescence from fluorochrome solutions must be related to fluorescence from stained particulates like cells, microspheres, and microdots.

Despite these difficulties, an empiric system of fluorescence calibration that preserves stoichiometry between the concentration of fluorochrome solutions and their equivalent molar quantity on stained microparticles has been utilized in research applications of flow cytometry for many years.3-7 Recent efforts show that it can be formalized and implemented in a fully standardized fashion across all fluorescence detection systems.8 We refer to this system as quantitative fluorescence calibration (QFC).9 The challenge and opportunity of this guideline is to establish the basic principles and the practical basis of QFC.

The earliest efforts toward QFC originated with immunologists, pathologists, and cell biologists who wanted to convert the subjective descriptions of “bright” or “dim” cell staining that reflected “high” or “low” levels of receptor expression into objective units that could be compared longitudinally across samples, experiments, and instruments.10-19 A number of ensuing reports addressed technical issues and expanded the clinical application utility of QFC,20-47 and an “emerging consensus” regarding the methods and uses of QFC48 was evident from a compilation of papers published in 1998.9,49-69 However, the lack of a formal theory and of authoritative reference materials hindered its full maturation, until the National Institute of Standards and Technology (NIST) initiated the Fluorescence Intensity Standards (FIS) program in 1999 (see Appendices B, C, and D).8

The NIST FIS initiative addressed both the practical and theoretical aspects of QFC. The practical need for authoritative standards led to the development of two reference materials: a fluorescein solution (see SRM 1932 in Appendix C) and a calibrated set of fluorescein-labeled microspheres (see RM 8640 in Appendix D). The parallel development of these reference materials bridges the gap between solution and particle fluorescence and serves as the model for developing future FI reference materials. While these materials have been in development, a series of papers in the NIST Journal of Research8,70-72 (see vii Number 26 NCCLS

Appendix B) has delineated the formal theory that supports and augments a fundamental approach to QFC first described in the cytometry literature.33 The formalization is based upon a well-defined property of fluorochrome solutions called the fluorescence yield, which relates the molar quantity of fluorochrome and its quantum efficiency to the measured intensity of fluorescence emission. Ironically, the term fluorescence yield had been applied much earlier in fluorescence microscopy to explain the variance in staining intensity caused by environmental influences in fluorescence-stained tissues.73

The fluorescence yield provides the stoichiometric link between fluorochrome solutions and labeled particles. It allows fluorescence intensity measurements on such particles to be converted into equivalent molar quantity units called molecules of equivalent soluble fluorochrome (MESF).9 The fluorescence yield of fluorochrome-labeled conjugate solutions, expressed as MESF per conjugate molecule (the “effective F/P ratio”), is analogous to the specific activity of radiolabeled tracers expressed as curies per mole. Thus, the molar quantity of conjugate on labeled particles can be determined by measuring the fluorescence yield of the conjugate solution and the MESF of the particles.

Having established the theoretical basis for QFC, this guideline addresses the operational steps required for implementing it, from characterizing primary reference materials to constructing calibration curves. Two types of fluorochromes are presented as models for QFC: organic dyes such as fluorescein, and proteins such as phycoerythrins and green fluorescent protein. These are among the most often used fluorochromes, and they pose somewhat different considerations for QFC that are representative of their different molecular properties.

Fluorescein is a small molecule which can be accurately characterized for mass purity by standard physical methods. With the availability of NIST fluorescein reference materials SRM 1932 and RM 8640, it becomes the first fluorochrome truly standardized for QFC. Its environmental sensitivities present technical difficulties for QFC, but these very properties make it a good test model for MESF calibration.

Phycoerythrin and green fluorescent protein are macromolecules that cannot be accurately characterized for mass purity by standard physical methods. This makes it difficult to establish a generic standard applicable to the wide range of possible variants. While their internalized chromophores are somewhat insulated from direct effects of the environment, these proteins are subject to changes in peptide folding that can alter fluorescence intensity. The approaches in this guideline should help bring better characterization and uniformity to the fluorescence measurements from these more complex molecules.

The subcommittee has focused on the use of QFC in ligand-binding assays in which measured fluorescence comes from a fluorochrome-conjugated ligand bound to cells or microspheres. This configuration encompasses the methods most often used in cellular immunophenotyping, suspension arrays, and microarrays. The goal of measuring fluorescence intensity (FI) from a cell-bound ligand is often to quantify the expression of its corresponding receptor, however indirect the relationship may be. Receptor binding values for unknown analytes may also be quantified by directly calibrating FI readings with cells or microspheres that have predetermined binding capacities. This approach obviates the need for MESF-calibrated fluorescence measurements. Since either approach should give the same answer, the chemistry of specific receptor-ligand interactions provides an independent reference frame for assessing the accuracy of FI measurements. The issues involved with quantifying receptor expression through QFC are detailed in Section 11.

The contents of this guideline are directed at both the suppliers and the users of fluorescence technology. The guideline should be immediately relevant to companies, researchers, and clinical specialists involved with flow and image-based cytometry and assays employing microspheres and microarrays.

Because the theoretical basis for QFC has been developed only recently, there is more background material than with most other NCCLS documents. This guideline will be most useful if it is used in concert with other NCCLS guidelines that address flow cytometry and fluorescence-based assays on other viii

Volume 24 I/LA24-A platforms (see the most current editions of NCCLS documents H42—Clinical Applications of Flow Cytometry: Quality Assurance and Immunophenotyping of Lymphocytes; H43—Clinical Applications of Flow Cytometry: Immunophenotyping of Leukemic Cells; H44—Methods for Reticulocyte Counting (Automated Blood Cell Counters, Flow Cytometry, and Supravital Dyes); and H52—Fetal Red Cell Detection).

Trueness in all fluorescence-based assays can be improved through QFC, and this quality assurance alone is sufficient reason for this guideline. However, the real benefit of QFC will be the direct translation of relative fluorescence measurements into standardized molar quantities, opening the way for accurate measurements of cellular expression in health and disease.74

A Note on Terminology

NCCLS, as a global leader in standardization, is firmly committed to achieving global harmonization in terminology wherever possible. Harmonization is a process of recognizing, understanding, and explaining differences in terms while taking steps to achieve worldwide uniformity. NCCLS recognizes that medical conventions in the global metrological community have evolved differently in the United States, Europe, and elsewhere; that these differences are reflected in NCCLS, ISO, and CEN documents; and that legally required use of terms, regional usage, and different consensus timelines are all obstacles to harmonization. In light of this, NCCLS recognizes that harmonization of terms facilitates the global application of standards and deserves immediate attention. Implementation of this policy must be an evolutionary and educational process that begins with new projects and revisions of existing documents.

In keeping with NCCLS’s commitment to align terminology with that of ISO, the following describes the metrological terms and their uses in I/LA24-A:

The term accuracy refers to the “closeness of the agreement between the result of a (single) measurement and a true value of a measurand” and comprises both random and systematic effects. Trueness is used in this document when referring to the “closeness of the agreement between the average value from a large series of measurements and to a true value of a measurand”; the measurement of trueness is usually expressed in terms of bias. Precision is defined as the “closeness of agreement between independent test/measurement results obtained under stipulated conditions.” As such, it cannot have a numerical value, but may be determined qualitatively as high, medium, or low. For its numerical expression, the term imprecision is used, which is the “dispersion of results of measurements obtained under specified conditions.”

Users of I/LA24-A should understand, however, that the fundamental meanings of the terms are identical in many cases, and to facilitate understanding, terms are defined in the Definitions section of this guideline (see Section 4.1).

All terms and definitions will be reviewed again for consistency with international use, and revised appropriately during the next scheduled revision of this document.

Key Words

Antibody binding capacity (ABC), calibration, effective F/P ratio, flow cytometry, fluorescence, fluorescence yield, fluorochrome, fluorochrome-ligand conjugate, fluorometer, immunophenotyping, ligand binding assays, molecules of equivalent soluble fluorochrome (MESF), reference materials, spectrophotofluorometer, standardization

ix Number 26 NCCLS

Volume 24 I/LA24-A

Fluorescence Calibration and Quantitative Measurement of Fluorescence Intensity; Approved Guideline

1 Scope

The scope of this document comprises the basic principles and essential procedures for quantitative fluorescence calibration (QFC): standardizing the measurement of fluorescence intensity (FI) on solutions and particles in a way that correctly maps the stoichiometry of chemical reactions. To elucidate the basic principles of QFC, the document includes background on the properties of fluorescence, the general features of instruments used to measure FI, and the assignment of molar equivalent values based on fluorescence yield. The document addresses the characterization and use of fluorochrome reference solutions, reference microparticles, and fluorochrome-labeled ligands. The document is directed toward suppliers and users of fluorescence-based technology for cellular analysis and for ligand binding assays employing microspheres, microarrays, microtiter plates, and image cytometry.

Although the general principles of QFC presented in this guideline apply to all FI measurements, the primary focus of this document is the use of suspended microparticles (microspheres and cells) to calibrate FI signals on flow cytometers. Practical methods for implementing QFC on flow cytometers are presented in the final section. QFC on imaging and scanning instruments will require additional consideration.

2 Introduction

Despite the widespread use of fluorescence in biomedical research and clinical laboratory assays, no uniform system of fluorescence calibration that maps fluorescence intensity (FI) measurements onto a scale of standardized units has yet been adopted. This overall process is referred to as quantitative fluorescence calibration (QFC).9

This guideline outlines the basic principles and laboratory practices involved in QFC. Because the theoretical basis of QFC has been formalized only recently, the guideline contains more detailed background material, appendices, and references than typical guidelines. It is intended for use with reference materials and procedures developed under the National Institute of Standards and Technology (NIST) Fluorescence Intensity Standards program.

At this time of rapid development in biotechnology and the need for expedient translation of research methods to the clinical laboratory, a consistent system of QFC will improve the characterization of all instruments and reagents that employ fluorochrome labels. The information provided by QFC enhances good laboratory practices (GLP)75-79 and helps comply with U.S. CLIA regulations,80-82 recommendations from the Centers for Disease Control and Prevention and from Health Canada for enumerating CD4 lymphocytes,83,84 and the ISO 07025 Guidelines.85

Looking ahead to future applications in biomedical science and public health, QFC formalizes a property value first identified by cell biologists: the specific binding capacity for fluorochrome-labeled ligands. The accurate measurement of binding capacities to quantify cellular expression will open a whole new window to prevention and treatment through predictive medicine and disease profiling.

3 Standard Precautions

Because it is often impossible to know what might be infectious, all patient and laboratory specimens are treated as infectious and handled according to “standard precautions.” Standard precautions are guidelines that combine the major features of “universal precautions and body substance isolation” practices. Standard precautions cover the transmission of all infectious agents and thus are more comprehensive than universal precautions which are intended to apply only to transmission of blood-borne pathogens. Standard and universal precaution guidelines are available from the U.S. Centers for Disease Control and Prevention (Guideline for Isolation Precautions in Hospitals. Infection Control and Hospital Epidemiology. CDC. 1996;17(1):53-80 and MMWR 1988;37:377-388). For specific precautions for preventing the laboratory transmission of all infectious agents from laboratory instruments and materials and for recommendations for the management of exposure to all infectious disease, refer to the most current edition of NCCLS document M29—Protection of Laboratory Workers from Occupationally Acquired Infections.

4 Terminology

4.1 Definitions

Absorptivity//Molecular absorption coefficient//Molar extinction coefficient – A measure of the absorption of radiant energy having a given wavelength and/or frequency as it passes through a material of substance concentration of one mol/L.

Accuracy (of measurement) – Closeness of the agreement between the result of a measurement and a true value of the measurand (VIM93)86; NOTE: See the definition of Measurand, below.

Analyte – Component represented in the name of a measurable quantity (ISO 17511); NOTE: As used in this document, the pure molecular or cellular form of the substance to be detected or quantified, independent of the sample matrix in which it is present.

Antibody binding capacity (ABC) – As used in this document, the number of antibody molecules specifically bound to the homologous antigen (receptor) in a cell or microparticle under saturating or near- saturating conditions; NOTES: a) Antibody binding capacity is often used as an indirect measure of expression of the homologous antigen, which is usually a cellular receptor or capture antibody on a microsphere. It is in this context that near-saturation binding is required. However, some results given in ABC units (such as titrations) clearly do not imply saturation; b) “ABC” has become a general term that does not distinguish between binding capacities for native antibody and fluorochrome-antibody conjugates. It is even used mistakenly to describe the binding capacity for ligands other than antibodies. While terms like “Ligand Binding Capacity,” and “Conjugate Binding Capacity” would be more precise, they are rarely used; c) Some reports use the variant term “AB/C” to stand for “Antibodies Bound per Cell.” This term presumably does not imply a requirement for saturation.

Autofluorescence – Background fluorescence arising from intrinsic sources in unstained measurands (particularly cells) under conditions used to detect desired fluorochromes.

Boundary – As used in this document, an absolute limit to the measurement response reading from an instrument; NOTES: a) Boundaries apply particularly to digitized data expressed in histograms, where the zero channel and the maximum channel define absolute limits to the readings obtained from the instrument; b) Boundaries for fluorescence intensity calibration curves are often expressed in units of analyte dose as extrapolated from a calibration curve to the minimal or maximal possible reading on the instrument scale. Lower boundary – The analyte dose extrapolated from a calibration curve to the lowest possible reading from an instrument; NOTE: The lower boundary is a theoretical value and should not be taken as a true

2 An NCCLS global consensus guideline. ©NCCLS. All rights reserved. Volume 24 I/LA24-A measurement of instrument sensitivity. However, the actual sensitivity of the instrument can never be less than the lower boundary.

Capture microsphere calibrator – As used in this document, a microsphere that has been surface-labeled with a reagent, usually an antibody, that binds fluorochrome-labeled ligands.

Digital signal processors (DSP) – Electronic components of instruments that change or analyze in real time information that was digitized from analog signals.

Effective F/P ratio – The fluorescence yield of a fluorochrome-ligand conjugate (FLC) expressed as moles of equivalent soluble fluorochrome per mole of FLC (in solutions), or as molecules of equivalent soluble fluorochrome (MESF) per molecule of FLC (on stained particles); NOTES: a) The word “ratio” is optional since it is implied by the term “F/P”; b) The term “F/P” originally meant “fluorochrome to protein” ratio when it was applied to molar quantities of fluorochrome conjugated to the carrier protein. Since most ligands used in QFC are proteins, the term has been retained. However, FLC reagents often contain an excess of carrier protein for stability, and carrier protein should not be included in determining either molar or effective F/P ratios. For that reason, “F/P” is sometimes taken to mean “fluorochrome to probe” ratio, where the term “probe” implies an active species such as the ligand of a receptor, whether or not it is protein.

Excitation – The transition of a fluorochrome molecule from its lowest energy state (i.e., ground state) to a higher energy state following absorption of incident light (see Appendix A).

FITC//Fluorescein isothiocyanate – The form of fluorescein most commonly conjugated to ligand molecules; NOTE: This is often used to describe any covalent fluorescein conjugate regardless of the actual conjugation chemistry used.

Fluorescence – Brief electromagnetic radiation emitted as a result of absorption of radiation (photons) by an atom, molecule, or ion; NOTE: Generally, fluorescent radiation has a longer wavelength than the absorbed radiation.

Fluorescence intensity (FI) – 1) The reading on an instrument response scale caused by detection of a portion of the fluorescence emission from excited fluorochromes; 2) a measure of fluorescence radiant power.

Fluorescence quenching – Any interaction of the fluorescence molecule with a solvent, solutes, or other environmental factors that lowers the fluorescence quantum yield; NOTE: As used in this document, fluorescence quenching may also be due to decreased absorptivity of the fluorochrome.

Fluorescence yield – Product of the concentration of a fluorochrome in solution or suspension and the quantum yield of the fluorochrome molecule.

Fluorochrome – A substance that fluoresces when excited by electromagnetic radiation; NOTE: This term is synonymous with, and in this document supersedes, the terms “fluorescer,” “fluorophore,” and “fluorophor.”

Fluorochrome-ligand conjugate (FLC) – As used in this document, a reagent for staining receptor-bearing microparticles in which a fluorochrome is covalently attached to a ligand for that particular receptor; NOTE: The ligand is most often an antibody raised against the receptor.

Fluorometer – A generic term for any instrument used to measure fluorescence intensity and possibly other qualities of fluorescence emission such as spectral distribution or anisotropy; NOTE: This term is synonymous with, and in this document supersedes, the term “fluorimeter.”

Flow cytometer – A microphotometer in which the sample consists of a stream of cells or other particles, ideally flowing in single file through a sensing region from which optical signals are collected as each particle is illuminated one at a time.

Histogram channel number – The ordinal number beginning at zero that represents the relative position of a particular bin in a histogram; NOTE: When analog signals are digitized, the histogram channel number represents the relative strength of the original analog signal.

Illuminate//Illumination – As used in this document, to supply a measurand with a source of light for purposes of exciting fluorochrome molecules with which the measurand may be labeled; NOTES: a) Fluorochrome molecules in the measurand are illuminated by incident light from the fluorometer; b) These terms are often used interchangeably with the terms “excite/excitation.” However, illumination is predetermined by instrument conditions alone, whereas excitation depends further on fluorochrome and environmental factors.

Image cytometer – Any of a group of instruments comprising fluorescence microscopes and microphotometers, image and scanning cytometers, and confocal microscopes that illuminate and collect optical signals from a microscopic field; NOTE: Image cytometers generally use microscope objectives (lenses with relatively high numerical aperture (N.A.) and magnification) to maximize illumination and detection efficiency.

Incident light – As used in this document, as used in the document, the light supplied by a fluorometer to illuminate a measurand for purposes of exciting fluorochrome molecules with which the measurand may be labeled.

Inner filter effect – An apparent decrease in emission quantum yield and/or distortion of bandshape as a result of reabsorption of emitted radiation, or absorption of incident radiation by a species other than the intended primary absorber (IUPAC Compendium of Chemical Technology, 2nd ed: 1997).

Linearity – The ability (within a given range) to provide results that are directly proportional to the concentration (amount) of the analyte in the test sample (WHO-BS/95.1793); NOTE: Linearity typically refers to the overall system response rather than the raw instrument output, but in this guideline the definition applies to both and is considered especially pertinent to instrument output.

Matrix – All components of a material system, except the analyte.

Measurand – Particular quantity subject to measurement (VIM93)86; NOTES: a) This term and definition encompass all quantities, while the commonly used term “analyte” refers to a tangible entity subject to measurement. For example, “substance” concentration is a quantity that may be related to a particular analyte; b) As used in this document, the measurand is the sample containing the analyte to be detected or quantified along with the matrix in which it is found.

Microparticle – As used in this document, a discrete physical body such as a biological cell or plastic bead or liquid droplet having a volume between about 1 and 10 000 cubic microns (i.e., an equivalent spherical diameter of about 1 to 30 microns).

Microphotometry – The measurement of light emanating from an illuminated microscopic field.

Molar quantity –See Quantity.

Molar F/P ratio – The molarity of fluorochrome per mole of fluorochrome-ligand conjugate (FLC); NOTE: the ratio is equivalent to the average number of molecules of fluorochrome per molecule of FLC.

Mole – The amount of a substance containing 6.023x1023 basic relatively homogenous units (e.g., atoms, molecules, cells, or particles) of that substance.

Molecules of equivalent soluble fluorochrome (MESF) – The number of native (unconjugated) fluorochrome molecules in solution that gives the same fluorescence yield as a solution of fluorochrome conjugated to other molecules or a suspension of fluorochrome bound to microspheres or cells.

NIST traceable reference material (NTRM) – A reference material produced by a commercial supplier with a well-defined traceability linkage to the National Institute of Standards and Technology (NIST); NOTE: This linkage is established via criteria and protocols defined by NIST that are tailored to meet the needs of the metrological community to be served. Reference material producers adhering to these requirements will be allowed the NIST “NTRM” trademark.

Photodegradation – Irreversible photochemically-induced structural changes and accelerated decay of fluorochromes upon illumination that reduce or ablate fluorescence emissions; NOTE: This term is synonymous with and in this document supersedes the terms “photobleaching” and “photodecomposition.”

Precision (of measurement) – Closeness of agreement between independent results of measurements obtained under stipulated conditions; NOTE: Precision is not typically represented as a numerical value but is expressed quantitatively in terms of imprecision—the standard deviation (SD) or the coefficient of variation (CV) of the results in a set of replicate measurements(ISO 3534-1).87

Quantity (measurable) – attribute of a phenomenon, body, or substance that may be distinguished qualitatively and determined quantitatively (VIM93-1.1).86 In this document, the term Molar Quantity is used to refer to the amount of a substance expressed as the number of moles, number of molecules, or molar concentration.

Quantum yield – The probability that absorption of an excitation photon results in emission of a fluorescence photon.

Radiant power – The time rate of flow of electromagnetic radiant energy; NOTES: a) Radiant power is usually expressed in watts, i.e., joules per second; b) Radiant Flux is an equivalent deprecated term still used in the context of light as a photon stream.

Radiometry – The science of measuring optical radiation; NOTE: Optical radiation is electromagnetic energy commonly referred to as light, encompassing wavelengths from 10 to 1 000 000 nm and divided into regions called ultraviolet, visible, and infrared.

Receptor – As used in this document in reference to solid phase ligand binding assays, the molecular species of the solid phase (e.g., cell or microsphere) to which the corresponding fluorochrome-labeled ligand shows saturable binding.

Relative fluorescence intensity (RFI) – An arbitrary value reported by a fluorometer that expresses the magnitude of an FI measurement as a relative proportion of other FI measurements taken under identical conditions.

Residual – The difference between a given data point and its predicted value; NOTES: a) As used in this guideline for evaluating FI calibration curves, the residual is calculated from the fluorochrome dose as: [Observed Value - Reference Value], where the observed value has been interpolated from the calibration curve and the reference value is the assigned value in relative or absolute units. Absolute residual – The absolute value of the residual. Percent residual – The residual expressed as a percentage, positive or negative, of the reference value, calculated as: [100 X ({Observed Value - Reference Value}/Reference Value)]. Absolute percent residual – The absolute value of the percent residual. Average absolute

An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 5 Number 26 NCCLS percent residual – The sum of the absolute percent residuals for all calibrator data points in a calibration curve divided by the number of data points, i.e., the average percent difference, positive or negative, between the observed and expected values of all standards used to generate the calibration curve; b) This value gives the most basic evaluation of linearity for the entire FI calibration curve.

Resolution – Degree to which an instrument can detect and report small changes in the signal being measured.

Sensing volume – In fluorometry, the intersection of the illumination volume (created by the incident light) and the detection volume (determined by the optical detection path); NOTE: Only fluorochrome molecules in the sensing volume can contribute to the fluorescence signal.

Sensitivity – The change in response of a measuring system or instrument divided by the corresponding change in the stimulus (modified from VIM93)86; NOTES: a) The sensitivity may depend on the value of the stimulus (VIM93)86; b) The sensitivity depends on the imprecision of the measurements of the sample; c) Sensitivity measures an instrument’s ability to distinguish a true analytical signal from background.

Singlet state – Excited state of a molecule in which the electrons in the higher-energy and lower-energy orbitals have paired spins; NOTES: a) Transitions to and from singlet states, which do not require changes of spin, have a high probability of occurrence; b) Fluorescence is the emission resulting from transition from a singlet state to a ground state.

Spectrophotofluorometer – A fluorometer that uses monochromators to allow a continuum of wavelengths to be scanned for illumination or detection.

Standard reference material (SRM) – Certified reference materials issued by NIST; NOTES: a) These are well-characterized materials produced in quantity to improve measurement science; b) SRM are certified for specific chemical or physical properties, and are issued by NIST with certificates that report the results of the characterization and indicate the intended use of the material.

Steradian – Solid angle at the center of a sphere subtending a section on the surface equal in area to the square of the radius of the sphere; NOTE: A complete sphere comprises 4π steradians.

Stokes’ law – As used in this document, the observation by Sir George Stokes that fluorescence emission has a longer wavelength than the illuminating light that causes excitation. Stokes’ shift – The difference between the wavelength of the excitation maximum and the wavelength of the emission maximum of a fluorescer; NOTE: This may also be referred to in the possessive (Stokes’ law/shift), but Stokes’ law/shift is incorrect.

Suspension array technology – As used in this document, the use of microspheres in suspension as the solid phase of ligand binding assays; NOTE: Such suspensions are usually analyzed by flow cytometry and are often multiplexed, i.e., they contain multiple populations of microspheres identified by unique fluorescent or light scatter properties that can be differentiated in a single analytical pathway.

Traceability – Property of the result of a measurement or the value of a standard whereby it can be related to stated references, usually national or international standards, through an unbroken chain of comparisons all having stated uncertainties.

Triplet state – Excited state of a molecule in which the electrons in the higher-energy and lower-energy orbitals have the same spin; NOTES: a) Transitions to and from triplet states, which require changes of spin, have a low probability of occurrence; b) Phosphorescence is the emission resulting from transition from a triplet state to a ground state.

Trueness – Closeness of agreement between the average value obtained from a large series of test results and an accepted reference value (ISO 3534-1).87 NOTE: [ISO 3534-1] The measure of trueness is usually expressed in terms of bias.87

Uncertainty of measurement – Parameter, associated with the result of a measurement, that characterizes the dispersion of the values that could reasonably be attributed to the measurand (VIM93)86; NOTE: The parameter may be, for example, a standard deviation (ISO/FDIS 15195).88

Value transfer protocol – A formal plan for carrying out the assignment of analyte values to a reference material from a reference material of higher order.

4.2 Acronyms and Abbreviations

ABC antibody binding capacity APD avalanche photodiode BPE B-phycoerythrin CCD charge-coupled devices CD cluster of differentiation CE capillary electrophoresis CE-LIF capillary electrophoresis with laser-induced fluorescence CLIA Clinical Laboratories Improvement Act CV coefficient of variation DSP digital signal processor/processing FI fluorescence intensity FIS fluorescence intensity standards FITC fluorescein isothiocyanate FLC fluorochrome-ligand conjugate FRET fluorescence resonance energy transfer GFP green fluorescence protein GLP good laboratory practices HEMA hydroxyethyl methacrylate HPLC high pressure liquid chromatography ISO International Organization for Standardization LED light-emitting diode LLL lower limit of linearity MESF molecules of equivalent soluble fluorochrome MTP microtiter plate NA numerical aperture ND neutral density NIST National Institute of Standards and Technology NTRM NIST-Traceable Reference Material®a PE phycoerythrin PMMA polymethyl methacrylate PMT photomultiplier tube QFC quantitative fluorescence calibration QSE quality system essentials RFI relative fluorescence intensity RM reference material (used as a formal designation by NIST) RPE R-phycoerythrin SAT suspension array technology SPF spectrophotofluorometer SRM Standard Reference Material®a a Registered trademark of National Institute of Standards and Technology. An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 7 Number 26 NCCLS

YAG yttrium-aluminum-garnet

5 Basic Principles of Measuring Fluorescence Intensity

5.1 Optical Radiation, Fluorescence, and Radiometry

Optical radiation is electromagnetic energy commonly referred to as light, encompassing wavelengths from 10 to 1 000 000 nm and divided into regions called ultraviolet, visible, and infrared. Fluorescence is the prompt emission of optical radiation after an atom, molecule, or ion absorbs optical radiation of a lower wavelength. A simplified overview of molecular fluorescence is given in Appendix A.

Stokes’ law states that the fluorescence emitted by an excited fluorochrome has a longer wavelength than the light that excited it. The Stokes shift is the difference between the wavelength of the excitation maximum and the wavelength of the emission maximum. This shift allows the excitation light to be filtered out from the light received by the emission detectors, creating a dark background against which individual emission photons can be seen. This accounts for the exquisite sensitivity of fluorescence detection, which, in theory, is greater than even that of radioactivity.

Radiometry is the science of measuring optical radiation. The basic properties of light are its spectral distribution and its energy. The radiant power of a light source, which can be viewed as a stream of photons (radiant flux) or a series of waves, is the energy of the light stream per unit time. Radiance is a measurement scale for radiant power expressed in watts and normalized to a source area of one square meter and a solid angle of one steradian (see Appendix B). Spectral radiant power (spectral radiance) is the radiant power within a given bandwidth (usually 1 nm).

Fluorescence intensity (FI) is the radiant power from fluorescence emission. In practice, FI generally refers to the arbitrary instrument reading from the fluorescence detector, expressed in whatever units the detector reports. Quantitative fluorescence calibration (QFC) converts these instrument-dependent units to a standardized platform-independent scale for FI measurements.

5.1.1 The Measurement Equation for Fluorescence

Measurements involving optical radiation are best characterized by reference to a measurement equation, which mathematically relates the signal measured by the instrument to the properties of the sample, the incident light, and the detection system. The formal derivation of a measurement equation for fluorescence is given in Appendix B2. Four properties of a fluorochrome solution appear in the equation: its molar concentration, its absorptivity, its quantum yield, and its normalized emission spectral function. The remaining factors that influence the measured FI depend on the incident light and the detection system of the measurement platform. The fluorochrome and instrument factors can be combined into a constant to give a linear relationship between the measured FI and the fluorochrome concentration in the sample:

if = [c] K

The components of the constant K represent 1) the fluorochrome factors that must be invariant between calibrators and analytes; and 2) the instrument factors that determine the reported FI. The goal of QFC is to retain direct proportionality between fluorochrome molar quantity and FI measurements even when instrument and fluorochrome factors differ, (see Section 6). Therefore, the first requirement for QFC is to calibrate within a linear response range to ensure that a stoichiometric relationship between fluorochrome concentration and FI measurement is maintained.

The major components of the constant K are delineated in the following simplified version of the equation and accompanying table of variables, where λ represents a given wavelength, and the contribution to fluorescence intensity is integrated across all wavelengths detected:

(λ) (λ) (λ) if =[c][ε][φ][gΩ ∫ T Q s dλ]

DEPENDENT SYMBOL TERM UNIT EXPLANATION ON Instrument, Varies with The fluorescence signal (instrument Fluorescence fluorochrome, i instrument reading) evoked by the radiant power of f intensity and and scale the integrated spectral fluorescence environment Molarity of the solution or number of c Concentration Molarity Fluorochrome molecules in the sensing volume Log-transformed proportion of illumination absorbed by fluorochrome. Molar Molar Fluorochrome ε Environmental effects are generally absorptivity absorbance (? Environment) small but technically difficult to measure on particle-conjugated fluorochromes. Fluorochrome Probability that absorption of an Quantum Unitless φ & environment excitation photon results in emission of a yield ratio fluorescence photon Unitless Gain of detection system (e.g., PMT and g Gain Instrument function associated signal processors) Illumination photon flux and emission No single term; Photons integrates photon detection efficiency integrated per m2 Instrument Ω illumination & across all positions in the sensing per sec collection volume Influence of emission wavelength on the Spectral Unitless T(λ) Instrument probability of detecting an emitted responsivity ratio photon Probability of detector signaling the Quantum Unitless Q(λ) Instrument arrival of an emitted photon at a given efficiency ratio wavelength Normalized Unitless Fluorochrome Relative proportion of photons emitted at S(λ) spectral ratio & environment a given wavelength function

This measurement equation shows that, under the simplifying assumptions of QFC, when FI is measured using a stable instrument under fixed settings, the calibration curve that maps instrument response to the molar quantity of fluorochrome will be a straight line if and only if all fluorochrome molecules have the same molecular extinction and quantum yield. These conditions are generally true when the fluorochromes are compared in dilute solutions with identical solvents. They are generally not true when fluorochromes are highly concentrated in solution, when they have been conjugated to ligands or bound to particles, or when the measurements are made in different matrices.

5.2 Factors That Influence Fluorescence Intensity Measurements

A detailed conceptual basis for measuring FI is presented in Appendices A and B. This section gives an overview of the major considerations involved.

5.2.1 Instrument Factors

5.2.1.1 Illumination and Detection Volumes

Given a source of photons, the excitation optics of any fluorometer select a set of wavelengths that illuminate a specific volume called the illumination volume. In the case of laser excitation, the illumination volume is the waist of the focused laser beam in the cuvette. The extended image of the monochromator entrance slit in the cuvette is the detection volume (see Figure B1, Appendix B).

5.2.1.2 Sensing Volume, Responsivity, and Detection Throughput

The total fluorescence radiant power delivered by the collection optics to the entrance slit of the monochromator depends on the overlap of the illumination and detection volumes in the cuvette. This overlap volume is called the sensing volume. Only fluorochrome molecules in the sensing volume can contribute to the fluorescence signal.

To compare the fluorescence radiance to the radiance of a standard source (usually a surface source), the fluorescent radiant power from the sensing volume must be mathematically converted into a radiant power from an effective surface source. The details of this conversion are presented in Appendix B. The comparison to a standard source corrects for differences in fluorometer responsivity, which is influenced by the grating efficiency and the photomultiplier (or CCD) efficiency, and detection throughput, which is influenced by the collection optics and the area of the entrance slit.

5.2.2 Fluorochrome Factors

5.2.2.1 Concentration and Molar Quantity

The fluorochrome concentration and the sensing volume determine the actual molar quantity (number of molecules) responsible for the FI signal.

5.2.2.2 Environmental Effects on Absorptivity, Quantum Yield, and Spectral Characteristics

Absorptivity reflects the probability that a photon of a given wavelength will be absorbed, while quantum yield reflects the probability that the excited state will decay by emitting a photon. Both can be influenced by the molecular environment of a fluorochrome, but the environmental effects on quantum yield are generally larger than the effects on absorptivity. The environment can also influence the shape of the emission spectrum and, to a lesser extent, the shape of the excitation spectrum. Since variances in quantum yields and emission spectra are likely to be much larger than variances in absorptivity and excitation spectra, the first approximation of QFC accounts for the former and assumes that the latter contributes only negligible error.

5.2.2.3 Inner Filter Effect

In sufficiently dilute solutions or suspensions of a fluorochrome, FI is directly proportional to the fluorochrome concentration. At higher concentrations (absorptivity above 0.1), light cannot be completely transmitted through the sample because it is partially absorbed by the fluorochrome molecules. Under these conditions, the fluorescence intensity becomes nonlinear with respect to fluorochrome concentration because not all the incident (excitation) light can penetrate the solution to the point where the collection optics are focused. The details of this phenomenon, sometimes called the inner filter effect or concentration quenching, depend on the geometric relationship between the excitation and emission detection paths, and on the path length of the sample. QFC avoids the inner filter effect by constraining the FI calibration curve to the range of linearity.

5.2.2.4 Polarization

Fluorescence polarization (FP) is a complex phenomenon dependent on the electromagnetic wave-like properties of light and its preferential absorption and emission by fluorochrome molecules oriented in different directions. The net effect of fluorescence polarization is that the measured FI is not uniform in all directions, a phenomenon called polarization anisotropy. This effect is not taken into consideration in the fluorescence measurement equation (see Section 5.1.1).

For purposes of QFC, the most important consequence of polarization anisotropy lies in the difference between fluorochrome molecules in solution and those bound to particles. Unbound fluorochrome molecules in aqueous solution are free to rotate, thus they will have negligible polarization anisotropy. However, fluorochrome molecules immobilized on particles will be constrained in their rotational motion and therefore may have significant polarization anisotropy. In assigning MESF units to labeled particles, polarization anisotropy could lead to bias between instruments because of differences in their optical geometries and in the polarization response of their optical elements.89

Even if MESF values on fluorochrome-labeled reference particles are corrected for polarization anisotropy when they are assigned by comparison with solutions, any significant differences in fluorochrome mobility between reference particles and labeled measurands such as cells will contribute some degree of inaccuracy in QFC. Polarization anisotropy has generally been ignored in flow cytometry, but it may be detected with careful analysis.90 These difficulties are likely to be amplified in microscopy and other imaging techniques. One proposed solution is to place a polarizing filter at a so-called “magic angle” (57.4 degrees for linearly- polarized source emission) in the fluorescence detection pathway, which obviates the dependence of FI on polarization anisotropy.90 Such a filter would, however, decrease the FI signal to some extent and might therefore compromise sensitivity.

Regardless of the general principles behind polarization anisotropy, its contribution to the overall error in QFC must be assessed on a case-by-case basis. Recent studies at NIST suggest that it will be a relatively small contributor to error when particles stained when fluorescein-labeled antibody conjugates are calibrated against microsphere standards that have been labeled with fluorescein by attachment to the microsphere surface through an extended alkyl chain. In this situation, both species of fluorochrome are relatively mobile and anisotropy is minimized. In particular, the error caused by neglecting polarization anisotropy appears to be less than the error caused by neglecting differences in absorptivity.

5.2.3 Quenching

Interaction of the fluorochrome with a solvent, solutes, or other environmental factors can lower the fluorescence radiance. This guideline will use the term quenching to refer to any reversible reduction in fluorescence radiance, whether it occurs because of altered absorptivity or decreased quantum yield. This is a more general definition of quenching than the one used in NCCLS terminology, which limits the effect to decreased quantum yield.

External factors can produce quenching by providing additional paths for the transfer of energy from the excited state of the fluorochrome. Collisional quenching involves the interaction of a transient excited-state fluorochrome with other molecules that result in the loss of excitation energy as heat instead of emitted light. This process is always present to some degree and is particularly noted in solutions that contain collisional quenchers such as iodide ions, molecular oxygen, and the nitroxide radical.

5.2.3.1 Static and Dynamic Quenching

Collisional quenching is an example of dynamic quenching, which occurs due to the interaction of the excited fluorochrome with a quencher. Dynamic quenching is characterized by a decrease in both the intensity and the lifetime. Static quenching involves the interaction of the fluorescent molecule in the ground

An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 11 Number 26 NCCLS state with a quencher that forms a transient nonfluorescent complex. A special case of static quenching is self-quenching where the fluorochrome and the quencher molecule are the same species. Self-quenching is most evident in highly concentrated solutions of the dye. Static quenching is characterized by a decreased intensity and an unchanged lifetime.

5.2.3.2 Fluorescence Resonance Energy Transfer (FRET)

Fluorescence resonance energy transfer (FRET) occurs when the energy of an excited fluorochrome is transferred to an acceptor species by electric coupling (dipole-dipole interaction). FRET requires overlap between the emission spectrum of the excited fluorochrome and the excitation spectra of the acceptor species. The result of this energy exchange is an excited acceptor molecule and a ground state donor fluorochrome. The donor and acceptor molecules do not need to collide, but merely approach to within about 100 angstroms. The interaction results in either a shift of fluorescence to longer wavelengths if the acceptor molecule fluoresces, or the complete absence of fluorescence if it does not.

FRET is the basis of some very useful techniques for measuring molecular conformation and molecular proximity.91 It has been used extensively with flow cytometry to elucidate membrane protein associations.92 FRET is also the basis of so-called “tandem dyes” that provide an extended Stokes shift by using fluorescence emission from one fluorochrome to excite a second fluorochrome (see Section 8.7.1).

Despite its utility as a probe of the molecular environment, FRET is a confounding factor in QFC. At best, FRET complicates the fluorescence measurement equation (see Section 5.1.1) by adding additional terms for absorption and emission by the affecter molecule. At worst, it invalidates the relationship between fluorochrome-labeled MESF calibrators that are not affected by FRET and fluorochrome-labeled measurands that are affected.

5.2.3.3 Dimer and Excimer Formation

As the concentration of a fluorochrome increases, the excitation and fluorescence spectra may change as a result of dimer formation. Formation of dimers in the ground state may occur with dyes and may lead to static quenching if the absorptivity of the dimer is less than the monomer. The term excimer is used to describe the formation of dimers in the excited state. As the concentration of solute is increased, there may be a gradual disappearance of the initial fluorescence spectrum associated with a single fluorochrome molecule and the appearance of a new emission spectrum associated with dimers or excimers. If there is no change in the excitation spectrum, then the resulting emission is from excimer formation in the electronically excited state.

5.2.3.4 Temperature

Quantum yields of some fluorochromes are very sensitive to temperature, while others are relatively insensitive. FI increases with decreased temperature. Elevated temperature results in an increase in the number of molecular collisions, which may result in loss of excitation energy.

The problem of changing emission intensities as a function of temperature is easily compensated in QFC by ensuring that all samples, blanks, calibrators, and controls are at the same temperature when measured. A more advanced calibration technique uses two fluorochromes with different emission spectra and known to have the same relationship of quantum yield to temperature. One such compound may be added to the samples as a reference temperature-normalizing fluorochrome. The emissions of both dyes may be determined, and a ratio of emission intensities can be calculated to remove temperature as a variable in the fluorescence readings.

5.2.3.5 Other Environmental Factors

Other potential influences on fluorescence intensity include pH, ionic strength, solvent type, covalent coupling to a ligand, and noncovalent interactions. Some fluorochromes undergo ion-induced structural changes that are evidenced by changes in the excitation or emission spectra, or in the fluorescence intensity. Some fluorochromes display characteristics that are highly solvent dependent, especially those with large excited-state dipole moments resulting in fluorescence spectral shifts to longer wavelengths in polar solvents.

Conjugation of fluorochromes to macromolecules or even to small molecules can significantly affect their fluorescence properties. Many fluorochromes are partially quenched when coupled to other molecules, including small molecules like drugs or macromolecules like proteins. Other fluorochromes are much less fluorescent as free molecules in solution than when they are bound to a large molecule. The chemical nature of the link between the coupled molecules as well as their electronic structures determines whether fluorescence will be quenched or enhanced.

Extrinsic agents such as oxygen and heavy metal ions quench some fluorescence compounds. Proteins quench some fluorochromes such as fluorescein and the BODIPY dyes due to noncovalent interactions. Such quenching may be due to charge-transfer interactions with aromatic amino acid groups in proteins. Some fluorochromes that contain one or more ionizable groups are useful as pH indicators, including fluorescein, hydroxycoumarins, and others.

5.2.4 Photodegradation

Absorption of light by a fluorochrome molecule can lead not only to fluorescence emission and return to the ground state, but also to permanent photochemically induced structural changes. Since the excited state of a fluorochrome is more chemically reactive than its ground state, excitation increases the probability of chemical reactions that result in such structural changes. When these changes reduce absorptivity or quantum yield, they cause an irreversible decrease in FI. This process is known as photodegradation, also called photobleaching and photodecomposition.

The rate of photodegradation during an FI measurement depends on instrument, fluorochrome, and environmental factors. The chemical nature of the fluorochrome is the most important factor, and the sensitivity of different fluorochromes to photodegradation is highly variable. For instance, rhodamine is much less light sensitive than fluorescein, and fluorescein is much less light sensitive than phycoerythrin.

The instrument and environmental factors that influence photodegradation include the illumination intensity, the sensing volume, the molecular environment of the fluorochrome, its dwell time in the excited state, and the number of times it has been excited. A number of precautions can be taken to reduce the effect of these factors. Fluorometers can be designed to minimize the required excitation energy by maximizing detection sensitivity (e.g., using the widest emission band pass filters compatible with the fluorochrome, low light detectors such as CCD cameras, and high numerical aperture collection lenses). Ambient light can be minimized and dark or opaque containers and tubing used to store and transfer fluorochrome-labeled material. The use of antioxidants such as phenylalanine, azide, and vitamin C in the sample matrix can protect fluorochromes, since the reaction of molecular oxygen with the triplet state of a fluorochrome often produces highly-reactive singlet oxygen that causes photodegradation.

Instrument, fluorochrome, and environmental factors can all influence the extent to which photodegradation causes departure from the assumptions of QFC. In the approximations of the basic reference method for QFC, these influences are assumed to be comparable in the measurands and MESF calibrators and therefore, are not taken into account. However, photodegradation should be accounted for when assigning values for authoritative MESF reference materials, which can be done in specialized flow fluorometers.72,93

5.2.5 Interferences

Signals that interfere with the FI measurement degrade the overall sensitivity of detecting fluorochromes and cause failure of the proportionate relationship between measured FI and the molar quantity of fluorochrome, upon which QFC depends. Since these interferences may not be the same for measurands and calibrators, they can cause systematic errors in QFC measurements, particularly in the lower calibration range. Every effort should be made to minimize such interferences for accurate QFC measurements.

5.2.5.1 Background Fluorescence

The sensitivity of specific fluorescence detection can be significantly reduced by background fluorescence signals. Extrinsic sources of background fluorescence include constituents in the matrix of the measurand other than the fluorochrome of interest that also fluoresce under the particular measurement conditions. Background fluorescence in ligand-binding assays can also be caused by residual fluorochrome from the labeled ligand that was unwashed or nonspecifically bound.

Background fluorescence may also arise from intrinsic sources, particularly in cells, where it is often called autofluorescence. A number of common cellular biomolecules such as flavins can cause intrinsic background fluorescence. The contribution of such autofluorescence to the measured FI can be reduced by selection of optical filters that reduce the excitation or detection of the background fluorescence relative to the desired fluorochrome. However, optical filters usually reduce fluorescence signal from the desired fluorochrome as well. A better approach is to select probes that excite and emit at wavelengths shorter or longer than those that cause autofluorescence. The autofluorescence of many animal cells and bacteria can be reduced by using fluorochromes that excite at wavelengths above 500 nm.

5.2.5.2 Elastic Light Scattering

When a beam of light passes through a solution or suspension, some photons are absorbed, some are transmitted, and some are scattered. Elastic scattering occurs whenever there are interfaces between materials with different indexes of refraction. Elastic scattering is characterized by the absence of any wavelength shift in the scattered light. However, the elastically scattered light can be detected in a fluorometer due to the scattering of the light from various surfaces in the monochromator. This type of scattering presents a problem in fluorescence measurements when the Stokes shift is very small, because the scattered excitation light is difficult to remove effectively with conventional optical filters. In this case, unless a judicious choice of light filters is made, some of the scattered incident light will be detected in the fluorescence measurement.

Elastic light scattering is of great concern when fluorescence is measured in suspensions of particles such as cells and microspheres. Quantitative FI measurements are possible in these highly scattering suspensions provided that absorption is not significant at the excitation and emission wavelengths. However, FI will be reduced since the particles scatter incident and emitted light. The decrease in FI due to scattering is characterized by turbidity (analogous to absorptivity). Interferences due to turbidity can be minimized by differentially subtracting the measured spectrum from a “blank” source, which uses an identical particle suspension without the presence of fluorochrome.

Scattered light from bubbles in the sample or adhered to the cuvette surface can be problematic in fluorescence readings. Such readings can be erratic and fluctuate with a change in the number, size, and position of the bubbles relative to the excitation beam. Bubbles may be removed by degassing, stirring, or simply waiting before a measurement is taken. In flow spectrofluorometers, care should be taken to minimize the introduction of bubbles; otherwise, a bubble trap or filtering system may be used. Sample intake that is routed through the bottom of the flow cell with outflow at the top is less likely to trap bubbles in the sensing region.

5.2.5.3 Raman Scattering

In addition to elastic light scattering, a very small amount of light transmitted through optically clear solutions or suspensions scatters and shifts to longer wavelengths. This effect, known as Raman scattering, is the result of inelastic vibrations of the solvent molecules. Because its low intensity and wavelength shift is similar to fluorescence emission, Raman scattering is commonly used as an independent calibrator for fluorometers.

If the wavelength shift of incident light due to Raman scattering is comparable to the Stokes shift of the fluorochrome being measured, the Raman scattered light cannot be separated from fluorescence emission with optical filters and becomes a source of background in the measurement of FI. Because the amount of Raman-scattering is so small, the background it causes is generally significant only when a monochromatic light source (e.g., laser) is used and FI from the fluorochrome of interest is weak. One situation where this can occur is the measurement of phycoerythrin-stained measurands excited with an argon ion laser at 488 nm, a common configuration on flow cytometers, where the Raman scatter from 488 nm light overlaps the fluorescence emission peak from phycoerythrin, typically measured around 560 to 590 nm.

Since scattered light is highly polarized, polarizing optical elements may diminish the relative proportion of background light scattering. However, such filters will also reduce FI to some extent and may compromise sensitivity more than the background caused by Raman scattering. They can also cause bias in QFC because the polarization anisotropy of fluorochrome-labeled microsphere calibrators may differ from that of labeled cellular measurands.

5.2.5.4 Simultaneous Measurement of Multiple Fluorochromes

The increasing use of multiple fluorochromes detected simultaneously presents a challenge in measuring each of the specific emissions, since most fluorochromes have rather broad emission spectra which may overlap each other. Contamination of FI signals by spectral overlap may be minimized by using narrow band-pass optical filters and by choosing fluorochromes with narrow emission spectra distributed over as wide a range as possible. Even with such optimization, some spectral overlap may be expected, especially as the number of different fluorochromes detected simultaneously increases.

Some analytical methods (notably flow cytometry and cell sorting) employ compensation hardware and/or software to correct for overlapping emissions by electronically subtracting a proportion of the overlapping FI signal from the desired FI signal. Hardware adjustments are useful for two-color compensation, but they become difficult and inaccurate when three or more colors must be compensated. Spectral overlap may also be compensated mathematically using a full matrix spillover correction applied through software, and this is the preferred approach for analyzing measurands labeled with three or more fluorochromes. The matrix coefficients required for this mathematical manipulation are best obtained by analyzing individual preparations of the measurand stained with only one fluorochrome at a time.

Accurate compensation is required for QFC, since inconsistencies can cause artifactual differences in FI calibration between instruments, measurement conditions, and measurands stained with different combinations of fluorochromes. Weak FI from fluorochromes that are heavily influenced by spectral overlap from other fluorochromes are most likely to be subject to significant error, so QFC should be applied cautiously under these conditions.

6 Instruments for Measuring Fluorescence Intensity

6.1 General Components

All instruments for fluorescence intensity (FI) measurements require one or more illumination sources for excitation, one or more detectors for emission, and the associated optics required to focus illumination and detection upon the sensing volume (see Appendix B). Additional optical components are usually present to restrict illumination and detection to the desired wavelength regions. Electronic and/or electromechanical elements provide an indication of the detector signal strength to the operator.

6.1.1 Illumination Sources

Illumination may be provided by spectrally and spatially extended sources such as incandescent filament and arc lamps and light-emitting diodes (LED), or quasi-monochromatic sources such as lasers. When extended sources are used, optical filters and/or prism or grating monochromators are generally required to restrict the spectral bandwidth of illumination, but these components are often omitted from instruments in which lasers are used.

6.1.2 Emission Detection Components

Detectors include solid state devices such as photodiodes, avalanche photodiodes (APD), and charge coupled devices (CCD); vacuum tubes such as photomultiplier tubes (PMT); and camera tubes (e.g., vidicons). Arrays of each type of device, except camera tubes, have been made and may be used for fluorescence imaging or, with a grating or prism to provide spectral dispersion, for multiple wavelength measurements. The interaction of photons with a detector sensing element (photocathode) yields electrons (current). Some electronic amplification and conversion of the current to voltage is typically necessary in the course of signal processing.

Although the photocathode quantum yield of solid state detectors is typically at least several times as high as that of tubes such as PMT, the latter incorporate a series of intermediate electrodes (dynodes) maintained at progressively higher DC bias voltages to produce largely noise-free current gain, which can exceed 10 000 000, while solid state devices, with the exception of APD, do not provide gain. However, CCD integrate the electrons over specified time intervals, providing great sensitivity.

Most detectors are operated in a linear mode, in which output current is proportional to the number of photons reaching the photocathode (neglecting the additive effect of thermionic emission, which produces a “dark current” even when no photons from the sample reach the detector). In this linear mode of operation, PMT are the detectors of choice for measurement of low-level fluorescence, since they are capable of detecting signals over a dynamic range of four or more decades. Cooling of both vacuum tubes and solid state detectors reduces dark current.

Like PMT, APD require a DC bias voltage, typically ranging from a few hundred to a few thousand volts; gain increases with the applied voltage. However, when APD are operated above breakdown voltage (so- called “Geiger mode”), gain increases dramatically, providing a large burst of current at the anode in response to a photon striking the cathode, but also introducing a dead time in which there will be no response to a second photon, a condition not present when APD are run in linear mode. APD in Geiger mode may be used for single photon counting as long as the photon flux is not so large that a substantial fraction are missed due to dead time. APD in Geiger mode can provide higher sensitivity than a PMT, but generally over a much more restricted range of pulse rates than a PMT can provide.

6.1.3 Optical Components

Illumination and collection are generally focused on the sensing volume by combinations of lenses and mirrors. In some instruments, it is feasible to use fiber optics or equivalent waveguides in the optical path. Polarization of incident light can introduce polarization in the fluorescence emission. However, fiber optics can change the polarization of light. In theory, polarizing components such as multimode optical fibers could be used to randomize polarization between the fluorescence emitter and the detector. The basic reference method for QFC assumes that polarization does not bias the relationship between calibrators and measurands, an assumption that should be tested for specific calibrators on specific platforms.

Selection of appropriate spectral regions for illumination and for fluorescence emission measurements can be accomplished with prism or grating monochromators or with optical filters that have transmission and blocking characteristics selected to minimize stray illumination in the FI signal. Many laser sources are sufficiently monochromatic as to make the use of filters or monochromators in the illumination path unnecessary; however, some lasers (e.g., red diodes) produce enough emission at wavelengths longer than the laser wavelength to require the insertion of a band pass filter centered on the laser wavelength in the illumination path.

6.1.4 Signal Processing

Typical fluorescence measurement devices now utilize both analog and digital electronics in their signal processing (See Figure 1). While final output of almost all instruments is in digital form, the point in the signal processing chain at which analog-to-digital conversion occurs varies from instrument to instrument. In a device that uses single photon counting techniques, the output of the detector module itself is a series of digital logic pulses, and digital circuitry and/or software calculate the pulse count rate, from which an intensity value is derived. In other systems, analog electronics are used to capture a steady-state signal value, or the value(s) of the peak amplitude, area, and/or duration of pulses representing individual measurement events, from the output of the detector/preamplifier. In newer pulse-processing instruments, characteristics of pulses may be derived by digital circuitry and/or software after the detector output is digitized at a high sampling rate.

In order to accommodate measurements that vary in value over a wide dynamic range, it is necessary for the analog-to-digital converter (ADC) to be able to get the smallest signal value on scale. The minimum range of numerical outputs that can adequately represent a four-decade dynamic range is 1 to 10 000, requiring at least a 14-bit ADC, but precision for low values may be inadequate. Before high-resolution ADCs became available, high dynamic range measurements were often made by using analog logarithmic amplifiers to convert data from a linear to a logarithmic scale; it is still possible to make accurate measurements using such devices if their deviations from ideal response are monitored and corrected after data are converted to digital form.

TRADITIONAL SEMIDIGITAL FULLY DIGITAL DESIGN DESIGN DESIGN (1960s-1990s) (1990s- current) (1990s-current)

FLUORESCENCE EMISSION (OPTICAL RADIATIVE FLUX )

DETECTOR DETECTOR DETECTOR

PRE-AMP PRE- AMP PRE-AMP

COMPENSATION INTEGRATOR/ 14-16 BIT ADC PEAK DETECTOR (HIGH SPEED)

LOG AMP 16-20 BIT ADC

INTEGRATOR / INTEGRATION/ PEAK DETECTOR PEAK DETECTION

RECORDING RECORDING COMPENSATION COMPENSATION 8-10 BIT ADC LOG CONVERSION LOG CONVERSION DATA DISPLAYS DATA DISPLAYS DATA ANALYSIS DATA ANALYSIS RECORDING DATA DISPLAYS DATA ANALYSIS

Figure 1. Block Diagrams Comparing Signal Processing in Traditional, Semidigital, and Fully Digital Flow Cytometers. ADC = analog-to-digital converter; LOG AMP = logarithmic amplifier. Solid outlines indicate hardware/analog components, while dashed outlines indicate digital processing and software components. In the traditional models, signals were often amplified by logarithmic amplifiers, analog devices in which the output signal amplitude is proportionate to the logarithm of the input signal amplitude. This approach accommodated the wide dynamic range (>1000-fold) required for many fluorescence intensity measurements on fluorochrome-stained cells. The availability of affordable high- resolution ADCs allows linear data to be digitized directly across a wide dynamic range, obviating the need for log amps. High-speed ADCs further allow peak detection and integration to be handled by digital processing.

6.2 Specific Platforms for Measuring Fluorescence

6.2.1 Fluorometers and Spectrophotofluorometers

The simplest fluorometers use only light sources and optical filters to select discrete wavelengths for illuminating measurands and detecting their fluorescence emissions. Spectrophotofluorometers (SPF) use monochromators to allow a continuum of wavelengths to be scanned for illumination or detection. Typical SPF use a xenon arc lamp as a source because its radiant power is relatively constant from the near ultraviolet to the near infrared region (a so-called “flat illumination spectrum”). Collimating optics direct illuminating light from the slit of the monochromator in the illumination path through the sample cuvette,

© 18 An NCCLS global consensus guideline. NCCLS. All rights reserved. Volume 24 I/LA24-A and collection optics with an optical axis perpendicular to that of the illumination optics direct light emitted from the sample through the entrance slit of the monochromator in the collection path. Light that passes through the exit slit of this monochromator is directed to the photocathode of a PMT. For spectral scanning, the illumination and/or emission wavelengths are changed, typically at a constant rate.

MONOCHROMATOR EXCITATION LIGHT CONTROLLERS MONOCHROMATOR SOURCE

EMISSION SAMPLE MONOCHROMATOR CHAMBER

DETECTOR (PMT, CCD)

DATA ACQUISITION, ANALYSIS & OUTPUT

Figure 2. Simplified Block Diagram of a Spectrophotofluorometer. (Adapted from Lakowicz JR. Principles of Fluorescence Spectroscopy. New York: Plenum Press; 1983. Used with permission from Kluwer Academic Publishers.) Solid lines indicate hardware components; dotted lines indicate components usually under computer control on modern instruments.

SPF are designed to operate over a relatively wide range of wavelengths, and to measure a wide range of signal intensities. Once measurement parameters have been defined for a particular application, the technique may be routinely implemented in a far less expensive fluorometer, in which expensive monochromators may be replaced by optical filters, the arc lamp source by a filament lamp or LED, and the PMT detector by an APD or photodiode.

Most SPF and fluorometers measure samples in cuvettes with a 1-cm square cross section. Cuvettes with a smaller cross section and which permit measurements to be made on a flowing fluid stream (running at right angles to both the illumination and collection optics) are also available.

6.2.2 Microtiter Plate Readers

Microtiter plate (MTP) reading technologies span a broad range of applications and detection methods. MTP readers are often used in high-throughput screening for pharmaceutical drug discovery, as they provide the capability of rapidly screening hundreds of thousands of compounds through miniaturization. Microtiter plate readers for fluorescence employ the same basic components as all fluorometers, with optical paths that focus on individual MTP wells.

There are several special considerations for MTP fluorescence readers. The intrinsic fluorescence from different microtiter plates can be highly variable across vendors and between lots from a single vendor. The

An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 19 Number 26 NCCLS illumination, detection, and sensing volumes depend on the size of the MTP wells and the fluorometer optics. The method of reading each well will influence the excitation time and other parameters of the intensity measurement. Data analysis and management for high-throughput systems are also critical features.

6.2.3 Capillary Electrophoresis and High Pressure Liquid Chromatography

Capillary electrophoresis (CE) and high pressure liquid chromatography (HPLC) use small diameter capillaries to reduce separation times and improve resolution. Typical capillaries are approximately 50 µm in diameter and about two meters in length. CE uses very high electric fields to induce movement through the capillary, while HPLC uses pressure. Components may be detected by fluorescence, absorptivity, electrochemical signals, and mass spectrometry. Capillary electrophoresis with laser-induced fluorescence (CE-LIF) has been developed for detection and quantification of small molecules and macromolecules in biologic fluids. The CE-LIF format eliminates antigen immobilization and avoids many solid phase- associated problems. Since laser-induced fluorescence detection is used, nonlabeled proteins and peptides that do not interact with the complex are not detected, giving clean baselines and easily interpreted electropherograms.

6.2.4 Fluorescence Microscopes and Image Cytometers

This group comprises fluorescence microscopes and microphotometers, image and scanning cytometers, and confocal microscopes. These devices use microscope objectives to illuminate the microscopic field and collect the fluorescence emission, (see Figure 3). Microscope objectives, which are lenses with relatively high numerical aperture (NA) and magnification, maximize illumination and collection efficiency. Most modern fluorescence microscopes use a single objective for both illumination and collection; the illumination beam is derived from an arc or filament lamp using a condenser lens and optical filters with appropriate band pass characteristics, and then reflected into the objective by a dichroic mirror, which reflects the illumination wavelength and transmits the longer emission wavelength(s). Fluorescence emission collected from the sample passes through the dichroic mirror and forms an image, which is visualized (or photographed) with the aid of an eyepiece or eyepieces.

Figure 3. Schematic of a Fluorescence Microscope. (From Shapiro HM. Practical Flow Cytometry. 4th ed., © 2003 Wiley-Liss, Inc. This material is used by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.)

Fluorescence imaging (image cytometry when the subject is a cell or cells) can be done with a CCD or other video camera device mounted in place of the usual microscope camera; cooled CCD are often used to permit detection of low-intensity fluorescence. Laser illumination is not generally useful for fluorescence imaging with wide field illumination, because of the pronounced speckle typically noted when a coherent source is used to illuminate a relatively large area.

In contrast to wide field illumination, both scanning laser cytometers and confocal microscopes typically use electromechanical components, such as galvanometer mirrors, to move an illuminating laser beam in one or two dimensions across a field of view, collecting light from a region of space much smaller than the field at any one time. Scanning may be implemented in various ways; one of the simplest involves the use of a motorized stage allowing the slide to be moved past the objective in a raster pattern, with suitably small illumination and collection regions defined by apertures and field stops in the illumination and collection paths. A single detector, typically a PMT, may be placed in the camera position; alternatively, a series of dichroic mirrors and optical filters may be used to separate signals from several spectral ranges, diverting each to a different detector. If the stage is not moved, i.e., if measurements are made of one suitably sized segment of the field of view at a time, the process is referred to simply as microphotometry. In a scanning laser cytometer, the goal is to produce a relatively low resolution image using a relatively large beam diameter with a large depth of focus. In a confocal microscope, the laser beam is focused to a smaller spot, yielding a smaller depth of focus, and a matched field stop in the image plane is used to restrict the volume from which fluorescence emission is sent to the detector; this produces a high resolution image. Variations on the fluorescence scanning theme are used in a variety of devices, ranging from plate readers to microarray readers.

6.2.5 Flow Cytometers

A flow cytometer is essentially a microphotometer in which the sample consists of a stream of cells or other particles, ideally flowing in single file. In most modern flow cytometers, the sample or “core” stream is injected into a larger volume of “sheath” fluid, allowing the sample to be confined to the central region of the overall fluid stream. The stream may be observed in a miniaturized (typical cross section less than 0.5 mm) cuvette, or (particularly in cell sorters, which allow selected cells or particles to be diverted from the sample stream) in a stream in air. Many early flow cytometers were built around fluorescence microscopes, and used arc lamps for illumination. Most recent models use one or more lasers as illumination sources, and follow the layout described for fluorescence spectrophotometers, with the optical axes of the illumination and collection optics perpendicular to each other and to the direction of sample flow (orthogonal geometry, see Figure 4). Microscope objectives or their functional equivalents are used for collection; dichroics and mirrors are used for spectral separation, and PMT are typically used for detection, except in the case of small angle scatter measurements, where photodiodes usually provide adequate sensitivity.

Although many different properties of cells and microparticles can be measured by flow cytometry, fluorescence is clearly the most useful. The best-known application is fluorescence-activated cell sorting (FACS), 94 which remains a major tool in cell biology and is emerging as a preparative technique for clinical biologic therapeutics such as stem cells (see http://www.fda.gov/cber/minutes/flocyt042001p1.htm). However, clinical laboratory applications of flow cytometry require only analysis without sorting, and most clinical instruments do not have sorting capability. Typical flow cytometers measure fluorescence emission in three or more spectral regions, with advanced research instruments capable of simultaneously detecting emission from as many as twelve different fluorochromes. Most flow cytometers also measure light scattered at small (0.5 to 10 degrees) and large (roughly 90 degrees) angles from the cells or particles being analyzed.

Modern commercial instruments incorporate personal computer-based systems for data acquisition, analysis, and archiving. A consensus format for raw data (list mode) files is observed by most manufacturers,95 and third-party analysis software is readily available. A greater diversity of flow cytometry platforms has recently emerged, including multi-laser high-speed desktop cell sorters, clinical instruments for cellular analysis, and instruments to conduct multiplexed assays on microsphere suspensions (suspension array technology, or SAT). Miniaturized platforms such as microchip-based flow cytometers are rapidly coming into use.

Figure 4. Schematic of the Optical System of a Fluorescence Flow Cytometer. (From Shapiro HM. Practical Flow Cytometry. 4th ed., © 2003 Wiley-Liss, Inc. This material is used by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.)

6.3 Fluorometer Evaluation and Performance Criteria for FI Measurements

Since the ability to use QFC depends on the direct proportionate relationship between the FI reading and molar quantity of fluorochrome, fluorometric measurements on any platform used for QFC should be evaluated for the linearity of response. Theoretically, the ideal dose-response relationship is a straight line in log space with unity slope, reflecting linearity in both log space and linear space. Practically, slight variations from unity slope may be found, but any substantial departure indicates a failure of either instrumentation or of the assumptions around which QFC is based.

In addition to linearity, the precision and sensitivity of FI measurements should be evaluated. At the level of the end user, all three parameters can be monitored by continuing assessment of dose-response calibration curves using authoritative reference materials (see Section 10). At the level of the manufacturers and suppliers of instruments and reagents (especially those assigning values to standards and fluorochrome conjugates), a more thorough evaluation is appropriate. However, the general considerations in performing such evaluations should be considered by both suppliers and users.

NOTE: This guideline will not detail the evaluation of wavelength accuracy. For cuvette spectrophotofluorometers with monochromators, wavelength calibration can be performed using emission lines from lamps, and spectral response can be evaluated using quinine sulfate (SRM 936a, available from NIST (see http://patapsco.nist.gov/srmcatalog/sp_publications/documents/SP260-64.pdf)). Band pass and interference filters should be evaluated by spectral scans. Interference filters in particular are subject to degradation over time and should be periodically checked for spectral characteristics by users as well as suppliers.

6.3.1 Linearity

For instruments with removable filters, NIST-traceable calibrated neutral density (ND) filters can be used for an accurate assessment of linearity, although this approach does not account for the effects of optical noise before the filter. The FI signal is provided by a relatively stable fluorochrome at a sufficient concentration to provide a reading at or above the highest desired range. ND filters are inserted for successive readings; three to four ND values per decade provide a good density of measurement points. The dose-response curve may be constructed by the same simple least-squares method used for the QFC calibration curve (see Section 11.2 and Figure 5). The log %T is the independent (X) variable and the log FI reading is the dependent (Y) variable (response values from log-amplified data require a conversion factor based on the number of channels per decade). The best-fit line can be evaluated for linearity and responsivity by comparison with typical values. For a fluorometer to show acceptable linearity over the range evaluated by certified ND filters, it should have very high correlation and very low residual differences between observed and assigned values. Typical values for these parameters are:

Correlation coefficient > 0.995 Average % Residual < 3.0% Slope (Response Coefficient) 0.97 - 1.03

Measurement points that force these parameters outside typical ranges lie outside of the linear dynamic range. A practical guide for the lower limit of linearity (LLL) is the value of the lowest calibrator that falls on the best-fit line with a percent residual less than 10%. In a more formal approach, objective statistical methods may be used to identify the linear region within a broad range of measurements.96

Linearity Evaluation using Neutral Density (ND) Filters 2.00

Cuvette Fluorometer 2 1.50 r = 0.9999 Slope = 0.9797 Cuvette Fluorometer Flow Cytometer AvgRes = 0.02% AbsRes = 1.59% 1.00 Flow Cytometer r2 = 0.9997 0.50 Slope = 1.0040 AvgRes = 0.04% AbsRes = 2.18% Log Observed %T Log Observed 0.00

-0.50 -0.50 0.00 0.50 1.00 1.50 2.00 Log Assigned %T

Figure 5. Linearity of Fluorescence Intensity Measurements Evaluated on a Cuvette Fluorometer and on a Flow Cytometer Using NIST-Traceable Neutral Density (ND) Filters. Both instruments had optics designed for using exchangeable filters. The FI signal in the absence of any ND filter was assigned the value 100% transmission (%T). Assigned ND values (expressed in absorbance units) were converted to %T values for the independent (X) variable. The observed FI response for each ND value was expressed as the percent of the response observed with no ND filter present. The regression was performed in log space, and regression parameters were determined in the same manner as for the basic QFC calibration curve (see Section 11).

Although the ND filter assessment uses authoritative standards, it has drawbacks. It cannot be used on instruments without exchangeable filters, and it will overestimate the lower limit of the linear range since optical noise, which contributes to nonlinearity at low concentrations of fluorochrome, is filtered out to the same extent as the fluorescence signal. Therefore, this method is of limited practical use.

As a practical method for cuvette fluorometry, dilutions of a fluorochrome solution may be used to provide the relative dose values. Equal-volume serial dilutions are easily prepared and give the most precise relative values from volumetric dilution, while gravimetric dilutions weighed on certified balances are more difficult to prepare but give the highest accuracy for relative values. For particle-based fluorometry, fluorescent reference particles with assigned relative fluorescence values (see Section 8) can provide the relative dose values. The linearity of electronics following the optical detector can also be assessed by measuring the difference between two reference particles with slightly differing FI as the detection electronics are adjusted to move the particles across the entire response range.20 The difference between the log of the response value for the two particles (or the values themselves if the signals are log-amplified) will remain constant within the linear range. An ingenious variant of this method analyzes a single fluorescent microsphere population and discriminates the singlet and doublet populations by their different light scatter positions. In the linear response range of the cytometer, the doublet population should have exactly twice the measured FI (a constant log difference of 0.30) as the singlet population.

6.3.2 Precision

Because fluorescence measurements are subject to so many sources of variability, each fluorometer used for QFC should have a stable operating configuration that can be assessed for consistent operation independent of any single fluorochrome preparation. Precision over time can be assessed by monitoring the response to stable fluorescence sources such as YAG crystals or plastic-embedded dyes. If the stable source is used to adjust gain settings on the fluorometer to preset target conditions, variation in the gain settings provides a record of consistency. Alternatively, if the gains are held constant, the response to the stable fluorochrome provides that record. Radiance from the stable surrogate fluorochrome should have comparable flux and spectral characteristics to that of analytes of interest. Whether gains or responses are monitored, variability should be small and random over time.97

6.3.3 Sensitivity

For practical purposes of this guideline, the best parameter of fluorometer sensitivity is the lower limit of linearity (LLL): the lowest observed point meeting criteria for linearity on a dose-response curve as (see Section 6.3.1). The reading from a blank analyte, sometimes called the detection threshold (see Section 11.3.5), provides an indication of background. The detection threshold will generally be lower than the LLL (i.e., the presence of a fluorochrome can usually be detected as levels below which its equivalent molar quantity can be reliably quantified).

A definitive assessment of sensitivity in flow cytometers and other event-based fluorescence measuring systems is complex and requires platform-dependent procedures and analysis. Two parameters impact sensitivity: the background (B), and the detection efficiency (Q). Both must be measured for a complete assessment, as B will determine the threshold above which a true signal can be detected, and Q and B will determine the ability to resolve events, especially those with low FI. The MESF unit supplies a standardized way of expressing these parameters: background is expressed in MESF, and detection efficiency in photoelectrons per MESF. A method for measuring these parameters on flow cytometers is presented in Appendix E and could be generalized to other platforms.66-68

7 Quantitative Fluorescence Calibration for Conjugates and Particles Under Varying Measurement Conditions

Analytes are generally quantified by comparison with reference materials in the same matrix. In dilute fluorochrome solutions with identical solvents, fluorescence radiance from each solution is directly proportional to fluorochrome concentration, and fluorescence intensity (FI) readings can be compared directly. However, when fluorochromes are used to label molecules and particles, their absorptivity and quantum yield are often changed by their environmental conditions. When these factors differ, equality of fluorescence radiance does not imply equality of fluorochrome molar quantities, and FI readings cannot be meaningfully compared.

Quantitative fluorescence calibration (QFC) allows the molar quantity of fluorochromes to be measured from fluorescence intensity comparisons where absorptivity and quantum yields differ. In the most general sense, QFC allows comparison of FI from any two fluorochromes in any environment, as long as the MESF calibrators have the same environmental responsivity as the labeled measurand.

7.1 Comparisons Based on the Fluorescence Yield

The fluorescence yield of a solution is the product of the molar concentration of fluorochrome and its quantum yield.70 The fluorescence yield is a measurable property of free fluorochromes in solution, solutions of fluorochrome conjugates, or suspensions of fluorochrome-labeled particles.

Just as it is valid to compare the quantum yield of any two fluorochromes, it is also valid to compare the fluorescence yield of any two fluorochrome solutions. If the relative molar absorptivities and relative quantum yields are known, an FI calibration plot of any fluorochrome can in principle be used to determine the concentration of any other fluorochrome excited under the same conditions.

In principle, fluorescence yields can serve as the basis for an absolute FI unit which could be used to compare any two fluorochromes so long as their absorptivities are taken into account and the integrated emission spectra are measured. For simplification, the basic reference method for QFC assumes that MESF units will be derived from solutions of the same fluorochrome used in the labeled reagent.

7.1.1 MESF Values as a Measure of Relative Fluorescence Yield

In practice, QFC uses fluorescence yield as the basis of a relative unit that allows comparison between different forms of the same fluorochrome: free fluorochrome molecules in solution, fluorochrome molecules conjugated to other molecules, or fluorochrome molecules bound to particles. Assuming that the molar absorptivities of the different forms are equal, a fluorescence intensity reading can be related directly to the fluorescence yield.

For a given excitation and detection system as defined in the fluorescence measurement equation, the fluorescence yield of a fluorochrome-labeled ligand or particle is equivalent to the number of free fluorochrome molecules in the sensing volume of a solution that gives the same spectral fluorescent radiance (assuming identical molar absorptivity). The fluorescence yield of labeled ligands and particles can therefore be expressed as molecular equivalent values of free fluorochrome in solution. This expression is most often given as “molecules of equivalent soluble fluorochrome,” or MESF. Molecular units are used because numbers of molecules are more convenient than molarities for expressing ligand- binding values for cells and microspheres.

7.1.2 Fluorescence Yield of Fluorochrome Conjugates (Effective F/P Ratio)

The fluorescence yield of a fluorochrome-ligand conjugate (FLC) solution can be expressed in moles of equivalent soluble fluorochrome by comparison with free fluorochrome reference solutions of known

An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 25 Number 26 NCCLS molarity. When the FLC molarity is also known, a value for average MESF per FLC molecule can be calculated. This average value, known as the “effective F/P ratio” or “effective F/P,” is the most useful parameter in characterizing FLC for use in QFC. The effective F/P will often be less than the molar F/P because of quenching effects on the conjugated fluorochrome. Practical methods for determining the effective F/P are presented later (see Sections 11.4.1 and 11.4.2).

7.2 Fluorescence Yield of Particle Suspensions (Microspheres and Cells)

To measure the fluorescence yield of a particle suspension, the particle concentration is measured with a particle counter, and the FI readings of particle suspensions are measured under the same conditions as the FI readings of fluorochrome reference solutions. The fluorescence yield of the particle suspension is interpolated from the FI calibration curve of the soluble fluorochrome. This value is divided by the particle concentration to determine the average MESF value per particle, which represents the number of soluble fluorochrome molecules required to produce the same FI reading as the fluorochrome molecules immobilized on a single particle.

7.2.1 Practical Aspects of Assigning MESF Values to Particles

The procedure for assigning MESF values to particles requires accurate measurement of the particle concentration and the fluorescence intensity across the entire emission spectra. The emission spectra are difficult to measure for suspensions of microspheres with a low number of immobilized fluorochrome molecules because of the elastic and Raman scatter of the excitation light by the beads and solution components. Elastic scattering can be reduced to acceptable levels by using two holographic notch filters in the apparatus. However, since Raman scattering in the fluorescence detector band pass region cannot be selectively removed by optical filters, the background contribution from a suspension of blank microspheres still must be subtracted.

To determine the concentration of microspheres in suspension, a particle counter using electronic resistivity is recommended. When using this technology, coincident events and particle aggregates can be sources of error. Most resistivity counters have coincidence circuitry that determines and corrects for the occurrence of two particles passing through the detector orifice at the same time. To reduce coincident counting, the concentration of microspheres should be within the recommended range for the particle counter. Aggregates should be minimized to the greatest extent possible before counting and in general, should not exceed 2% of the microsphere concentration. A small proportion of doublets is acceptable if they are accurately identified and the microsphere count is corrected accordingly.

7.2.2 Using MESF Particles as Calibrators for QFC

A series of particles with varying amounts of immobilized fluorochrome exposed to the matrix can be calibrated against fluorochrome reference solutions to prepare a set of reference particles, each with an assigned value for average MESF per particle, that encompasses a useful range of FI. These MESF values can then be used as the dose (X) values of the calibration curve that preserves stoichiometry across varying conditions of FI measurements.

7.2.2.1 Differences in Instruments and Environments

Once microspheres have been properly calibrated in MESF units, as long as the fluorochrome molecules are available to the chemical environment, the MESF values will remain valid in any fluorometer, even if the environmental conditions are different from the original calibration. These microspheres can be used to calibrate the response of flow cytometers, and, if polarization effects are taken into account, they can be used to calibrate image cytometers.

7.2.2.2 Differences in Emission Spectra

Because the proper assignment of MESF is based on comparison of total fluorescent radiance between solution and particles, MESF calibrators may be used in any spectral region. However, in measurements that do not include the entire range of emission, spectral matching between calibrators and analytes is essential. Differences in emission spectra could create substantial bias in comparing FI measurements made across different bandwidths.

7.2.2.3 Differences in Fluorochrome Absorptivity

The assumption that immobilized fluorochrome molecules have the same absorptivity as free fluorochrome molecules can be a source of error. However, unlike differences in emission spectra, this error would not be instrument-specific, but rather a systematic error in assigned MESF values that could be corrected by factoring the ratio of absorptivities into the MESF values obtained by the calibration of any fluorometer.

8 Solution-Based Fluorochrome Reference Materials for QFC

This chapter presents a general approach to characterizing and using reference fluorochrome solutions for QFC. Fluorescein, phycoerythrin, and green fluorescent protein are used as specific examples.

8.1 Purpose and Criteria for Fluorochrome Reference Solutions

Reference solutions of fluorochromes are needed in QFC as standards for determining the fluorescence yields of conjugate solutions and the MESF values of labeled particles such as cells and microspheres.

In general, reference materials must have certain property values that are sufficiently homogenous and well- established so as to be used for the calibration of an apparatus, the assessment of a measurement method, or assigning values to other reference materials.9 This is a broad definition that encompasses many materials.

8.1.1 Primary Standards and NIST Reference Materials

The most authoritative reference materials are primary standards—those having the highest metrological qualities and the assigned values of which are accepted without reference to other standards of the same quantity.9

The term Standard Reference Material® (SRM) is trademarked and can only be applied to preparations certified and distributed by NIST. The reference values assigned to SRM are certified by NIST and include estimates of uncertainty. NIST SRM are generally considered primary standards.

NIST also provides Reference Materials (RM) with assigned values that represent the best estimate of true values but do not meet the criteria for certification. Their estimates of uncertainty may reflect only measurement imprecision and may not include all other sources of uncertainty. NIST RM may also serve as primary standards since they are often the best available materials.

A NIST Traceable Reference Material® (NTRM) is a reference material produced by a commercial supplier with well-defined traceability to specifications determined by NIST. This traceability is established via criteria and protocols defined by NIST that are tailored to meet the needs of the end users. NTRM was established to allow NIST to help commercial supplies respond to the increasing needs for high quality reference materials. Reference material producers adhering to these requirements are allowed to use the NIST “NTRM” trademark to identify their product. NTRM is also often used as a primary standard.

8.1.2 Other Standards and Calibrators

A formal nomenclature for reference materials has been presented in the context of QFC.9 As of the publication of this guideline, only one fluorochrome solution (NIST fluorescein SRM 1932—see Appendix C) and one fluorochrome-labeled microsphere calibrator (NIST RM 8640) are available as primary standards for QFC. Other materials that might serve as standards must be evaluated from the documentation provided with them and whatever additional information (spectral properties, stability, etc.) can be obtained in the laboratory.

8.2 Property Values for Fluorochrome Reference Solutions

8.2.1 Mass Concentration and Purity

For QFC, the essential property value of a fluorochrome reference solution is the mass concentration of fluorochrome. Mass/mass concentrations are preferable to mass/volume concentrations, which cannot be measured as accurately.

Purity must be accurately determined to assign mass values. The highest available purity is desirable to avoid impurities that might interfere with fluorescence.

8.2.2 Absorptivity

The absorptivity of a fluorochrome reference solution must be known for it to serve as a standard for fluorescence yield.

8.2.3 Relative Quantum Yield

The relative quantum yield of any two fluorochromes excited under the same conditions is readily measured by comparing their fluorescence yield. Therefore, the availability of a primary reference fluorochrome solution allows the standardized determination of relative quantum yield for any other fluorochrome that can be excited at the same wavelengths.

8.2.4 Fluorescence Lifetime

The fluorescence lifetime is the average time that a molecule spends in the excited state before emitting a photon and returning to the ground state. Most fluorescence lifetimes fall within the range of hundreds of picoseconds to hundreds of nanoseconds. The fluorescence lifetime may be readily measured with the proper equipment.

8.3 Other Specifications for Fluorochrome Reference Solutions

8.3.1 Fluorochrome Concentration

The fluorochrome concentration in a stock reference solution should be sufficiently high to allow dilution into working solutions so that the fluorochrome environment is comparable to those encountered for specific applications. Higher concentrations also reduce loss by adsorption to the storage container walls.

8.3.2 Well-Defined Solvent/Matrix

Because most fluorochromes are environmentally sensitive, the matrix of a fluorochrome standard solution is critical. The exact composition of the solvent should be specified, including buffer components, protein additives, and any other materials. Since reference solutions will be diluted into various matrices used for different applications, the simplest stable solvent/matrix is advisable to avoid incompatibilities.

8.3.3 Storage, Stability, and Expiration

Fluorochrome reference solutions should be properly packaged and accompanied by recommendations for optimal storage. Amber or light-free containers are preferred to minimize photodecomposition. Losses by adsorption onto glass, plastic, or bottle-cap material must be minimized. Single-use sealed glass ampules are recommended for reference solutions intended to serve as primary standards.

Heat and light are the major sources of instability to fluorochromes. Long-term stability is generally increased by refrigerated storage. In some cases, temperatures below freezing are optimal, but freezing destroys the fluorescence of some protein fluorochromes. Thermal stability is a desirable attribute for reference materials, and some fluorochromes that are particularly stable may be better reference materials than those commonly used in the biomedical laboratory.

Extended or accelerated time stability studies that would justify specific expiration dates are generally not available for fluorochrome solutions. The stability of fluorochrome reference solutions can be monitored by ongoing comparison with stable radiant signals such as those from atomic fluorescence (e.g., YAG crystals) or Raman scattering. Specific expiration dates should be assigned by suppliers based on real time (shelf life) studies when fluorochrome reference solutions are held under recommended conditions.

8.3.4 Biohazards

Many DNA-binding fluorochromes are potential mutagens. Theoretically, this mutagenic potential could be enhanced by ionizing radiation such as that from a laser. Most other fluorochromes used in flow cytometry are not biohazardous. Ideally, material safety data sheets should be included with fluorochrome reference solutions.

8.4 General Chemical Classes of Fluorochromes

The fluorochromes in most common use today are either organic dyes (small aromatic molecules like fluorescein) or proteins (large biomolecules like phycoerythrin and green fluorescent protein). Specialized fluorochromes include the tandem conjugates formed from organic acid and protein molecules, lanthanide chelates used for time-resolved fluorescence, and novel molecular superstructures called nanocrystals (“quantum dots”).

8.5 Reference Solutions for Organic Dye Fluorochromes

These fluorochromes are generally available at reagent-grade specifications from chemical supply houses or specialty companies. Their mass and purity are relatively easy to document using standard analytical techniques. Many are proprietary designer molecules engineered for spectral characteristics, but generic fluorescein remains the prototype example of this class.

8.5.1 Fluorescein

Fluorescein was first used for immunofluorescence microscopy by Coons over 50 years ago,2,98 and it remains the most commonly used fluorochrome for this purpose. A major attraction of fluorescein is the relative stability under dry storage of its activated isothiocyanate derivative (FITC), an off-the-shelf reagent which can be used directly to produce stable protein conjugates. Lyophilized fluorescein-protein conjugates are particularly stable.

8.5.1.1 Chemical Structure and Fluorescence Properties

Fluorescein can exist in at least six different ionic species with widely variant absorptivity and fluorescence.99,100 Therefore, fluorescein FI is highly dependent on pH, with a maximum above pH 9 and a 90% reduction at pH 5.

Fluorescein has many benefits as a fluorochrome: relatively high maximum absorptivity (ε approximately 87 000 in the dianion form), excellent fluorescence quantum yield (>0.9), and good water solubility. Its excitation maximum (approximately 494 nm) is close to the 488 nm spectral line of the argon-ion laser most often used on flow cytometers and confocal laser scanning instruments. The green emission (approximately 520 nm) is readily visible, easily detected, and leaves room to discriminate additional longer-wavelength labels in multiparameter/multiplexed applications. However, fluorescein also has drawback: such as its environmental sensitivity (particularly to pH), a relatively high rate of photobleaching, a broad fluorescence emission spectrum, and a tendency toward quenching on conjugation to biopolymers. These drawbacks have led to the development of more stable dyes with similar spectral characteristics to fluorescein, which allows their use in existing instruments.

8.5.1.2 Storage and Stability

Fluorescein is highly stable when stored in borate buffer above pH 9 in the dark at 4 °C. Once fluorescein solutions have been excited, they should be discarded.

8.5.1.3 NIST Fluorescein SRM 1932

This solution consists of highly-purified fluorescein (>98%) which has been characterized by mass spectroscopy, capillary electrophoresis, atomic absorption, and fluorescence spectroscopy (see Appendix D). The solvent is a borate buffer, pH 9.1, 0.1 M. The certified mass/mass concentration is equivalent to about 60 micromolar, which allows dilution into working buffers that will equalize the fluorochrome microenvironment with any particular analytical matrix. SRM 1932 is packaged in kits of five sealed ampules, each containing 2 mL and intended for one-time use to create secondary standards. Other property values such as absorptivity and fluorescence lifetime may be given as background information, but are not certified.

8.5.1.4 Other Fluorescein Solutions

Fluorescein is used for imaging blood vessel leakage in ophthalmology, and formulations meeting United States Pharmacopeia (USP) criteria are available by prescription. They have not been evaluated as laboratory reference materials to date.

8.5.1.5 Fluorescein Isothiocyanate (FITC)

FITC, the most commonly used derivative of fluorescein, is an unstable derivative that slowly hydrolyzes in water to form fluorescein. Although it is not strictly accurate from a chemical point of view, the terms “fluorescein” and “FITC” are often used interchangeably and may be applied as well to other related derivatives.

The spectral properties of FITC and fluorescein are very similar, but the fluorescence yield of FITC is lower because of electron-withdrawing effects by the thiocyanate group. The inherent instability of FITC makes it a poor reference material. Since the spectra of fluorescein and FITC are nearly identical, fluorescein is used as the FI standard for FITC and FITC conjugates.

FITC reacts rapidly with free amines, labeling proteins at their lysine and N-terminal residues. The fluorescein fluorochrome is further quenched after covalent linkage to protein, so that the fluorescence yield of FITC conjugates is lower than that of FITC or free fluorescein.

8.6 Reference Solutions for Protein Fluorochromes

No primary reference materials for protein fluorochromes have been developed at the time of this guideline’s publication, in part because they present unique problems. The mass concentration and purity of protein solutions are customarily estimated through biochemical assays rather than rigorous analytical methods. Moreover, the fluorescence properties of protein fluorochromes may be altered by subtle changes in peptide folding such as those induced by conjugation and binding.

Given these difficulties, protein fluorochrome solutions intended as reference materials for QFC should be characterized as completely as possible to establish their molar concentration (which requires accurate values for mass concentration and for molecular weight), purity, and conformational stability. Mass spectroscopy probably provides the highest metrological criteria for protein reference standards.101 In some cases, recombinant proteins may offer advantages as reference materials. The spectral properties of reference protein fluorochromes should also be characterized and monitored as part of stability studies.

This section discusses characterization of phycoerythrin and green fluorescent protein solutions by way of examples.

8.6.1 Phycoerythrin

Phycoerythrin (PE) is widely used in flow cytometry because its high absorptivity and quantum yield produce a strong FI signal. On a molar basis, the fluorescence intensity of PE is about the same as 30 unquenched fluorescein molecules at comparable wavelengths.102

8.6.1.1 Chemical Structure and Fluorescence Properties

The phycoerythrins comprise several phycobiliproteins found in red algae (seaweed) that absorb light between 490 and 570 nm. The most commonly used PE is R-phycoerythrin (RPE), a cylindrically-shaped protein molecule (12 nm diameter and 6 nm length) with a molecular weight near 240 000. RPE contains three major subunits: alpha (20 000 daltons), beta (20 000 daltons), and gamma (30 000 daltons) with the quaternary structure [(alpha-beta)6-(gamma)1]. It has 34 distinct chromophores attached to its polypeptide backbone.103,104 The alpha subunit contains only phycoerythrobilin chromophores, while the beta and gamma subunits also contain phycourobilin chromophores. B-phycoerythrin (BPE) is very similar to RPE and is actually a brighter fluorochrome when excited at its absorbance maximum, but it is not used as widely because it does not excite as efficiently at 488 nm, the argon laser line commonly used to excite both fluorescein and PE on flow cytometers.

Variation in the absorbance spectra of RPE from different sources reflects differences in the chromophore ratios of the beta and gamma subunits. The major absorbance peak is always near 566 nm, with varying contributions by secondary peaks at 545 nm and 496 nm. Fluorescence emits maximally near 575 nm. The interactions between chromophores of all the phycobiliproteins including RPE and BPE are particularly important to their fluorescence properties. These interactions may be influenced by changes in the tertiary and quaternary structure of the polypeptide chains, and sufficiently denatured biliproteins are devoid of fluorescence.103

Because of its broad absorbance spectrum, PE can be excited efficiently by argon lasers (488 nm or 514.5 nm bands), solid state lasers (532 nm), and helium-neon lasers (543.5 nm). The Stokes shift as defined by excitation and emission maxima is not large (about 10 nm), but the broad absorption region allows excitation at illumination wavelengths well below those of fluorescence emission. Depending on the

An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 31 Number 26 NCCLS excitation wavelength used, PE fluorescence can be readily discriminated from incident light scattering and background fluorescence from cells (autofluorescence). However, the Raman scatter from 488 nm laser light in water, which peaks at 593 nm, may cause a noticeable background signal in the PE fluorescence emission region.

PE fluorescence is generally insensitive to pH, temperature, and quenching by other biomolecules. It is negatively-charged at physiologic pH, which reduces nonspecific binding to cell membranes, and it is highly water soluble.105

8.6.1.2 Storage and Stability

Phycobiliproteins display reasonable long-term stability when stored refrigerated as ammonium sulfate precipitates in the presence of an antimicrobial agent such as sodium azide. In solution, PE is most stable near neutral pH and at protein concentrations higher than 0.1 mg/mL. At very low protein concentrations (< 1 pM), the multi-subunit structure of PE dissociates with concomitant loss of fluorescence intensity. To improve stability, bulk proteins such as protease-free gelatin or bovine serum albumin may be added to PE solutions.106 Acidic (pH < 5) and basic (pH > 9) conditions also promote dissociation. Spectral shifts may occur upon dissociation of the subunits and as a function of protein concentration.

8.6.1.3 Characterization of Purity and Spectral Properties

Phycobiliproteins can be harvested from natural or cultured seaweed. Some commercial suppliers of PE describe in detail their criteria for its purity and spectral properties. For RPE, the A566/A280 ratio is commonly used as an indicator of purity; this ratio is also affected by the condition of the pigment, which may be altered during harvest and purification. The A566/A496 ratio will increase with increasing contamination by B-phycoerythrin (BPE), as BPE has only a slight absorbance shoulder at 496 nm. The A620/A566 ratio can be used as a rough indicator of contamination by R-phycocyanin, another phycobiliprotein found in seaweed. Polyacrylamide gel electrophoresis can be used to assess contamination by other proteins. Nominal values for molar absorptivity (molecular extinction) can be found in some product literature, suggesting that lot-specific values would be available. The fluorescence yield of PE can be monitored for consistency by comparison with rhodamine, a more stable fluorochrome available in high purity that can be measured under the same instrument conditions as PE.

8.6.2 Green Fluorescent Protein (GFP)

Green fluorescent protein (GFP), originally purified from jellyfish, quickly became a valuable marker for gene expression and subcellular localization.107 GFP is most often used as a gene “tag” to measure in vivo expression. Since the GFP gene was first cloned, a number of variants have been derived from random and directed mutations. These variants provide a wide variety of fluorescent colors and much greater intensity, some over 300-fold brighter than wild-type GFP.

8.6.2.1 Chemical Structure and Fluorescence Properties

GFP purified from the jellyfish Aequorea victoria contains 238 amino acids arranged as an 11-stranded beta-barrel into which an alpha helix is threaded. The chromophore is attached to the alpha helix and is buried in the center of the cylinder, which has been called a “beta-can.” The chromophore itself is a p- hydroxybenzyl-ideneimidazolinone, formed from residues 65 to 67 (Ser-Tyr-Gly in the native protein). The posttranslational, autocatalytic modification of this tripeptide occurs in a sequence of nucleophilic attack, dehydration, oxidation, double-bond formation, and ring formation (cyclization). The end product is a chromophore with visible absorptivity and fluorescence.

Random and directed mutageneses have led to many GFP variants which can be grouped into different classes based on the components of their chromophores.107 Each class has a distinct set of multiple excitation

32 An NCCLS global consensus guideline. ©NCCLS. All rights reserved. Volume 24 I/LA24-A and emission wavelengths. Class 2 variants can be measured with the same optics used for fluorescein. Therefore, under identical measurement conditions, the relative fluorescence yield of Class 2 GFP solutions can be calibrated to NIST fluorescein standard solution (SRM 1932), and NIST fluorescein-labeled MESF microsphere standards (RM 8640) can be used to calibrate GFP fluorescence on particulates.

Protein folding, photoisomerization of the complex, and dimerization can all influence the fluorescence properties of GFP variants. Of particular concern for QFC is that variation in folding affects fluorescence intensity, and fluorescence from expressed GFP does not report the fraction of molecules that have not yet folded to the fluorescent state.

GFP has been somewhat disappointing in flow cytometry because of an apparent lack of sensitivity. However, sensitivity is actually limited by the relatively high autofluorescence of mammalian cells. Subcellular resolution through image analysis rather than flow cytometry can greatly reduce the number of molecules required for GFP detection. As GFP becomes highly concentrated, the surrounding unlabeled region of the cell provides an internal reference for the autofluorescence background. In imaging systems, as few as 300 to 3000 GFP molecules packed into a centrosome are readily visible inside a cell.

8.6.2.2 Preparation, Characterization, Stability, and Storage

Since most applications of GFP involve in situ expression of GFP genes in cells of interest, the protein itself has received relatively little attention as a reagent. However, various forms of GFP are commercially available, usually recombinant GFP (rGFP) purified from bacteria transfected with the gene coding for the particular GFP variant. The isolation of rGFP from bacteria can involve sonication, repeated freeze-thawing, and/or organic solvent extraction. Extraction and purification can alter the fluorescence properties if the protein is denatured during these steps.108 Absorbance and fluorescence emission at wavelengths appropriate for the particular variant are used to monitor the specific activity of the preparations; these spectral characteristics do not necessarily distinguish purity from variations in fluorescence properties. Nominal values for molar absorptivity (or molecular extinction) and quantum yield are provided by some commercial suppliers; these appear to vary widely between GFP variants.

The stability of purified rGFP depends to some extent on the particular variant and preparative methods, so documentation of the stability and storage conditions should be obtained from each particular source. Most rGFP variants are resistant to the common sources of degradation: room light and temperature (from below freezing to above 60 °C), neutral to moderately alkaline pH (7 to 11), mild denaturants (e.g., 8M urea), and most common proteases (except pronase). For many GFP variants, the rate of photodegradation is less than that of fluorescein and may be minimized by using the longest possible wavelength for excitation. Since GFP must be in an oxidized state to fluoresce, strong reducing agents (e.g., 2 mM FeSO4) may ablate fluorescence entirely, but weaker reducing agents (e.g., 10 mM dithiothreitol) generally do not affect it. Many of these effects are reversible, and GFP fluorescence will recover if the matrix is adjusted to a more optimal condition. However, hydrogen peroxide and sulfhydryl reagents can irreversibly damage its ability to fluoresce.

8.7 Reference Solutions for Specialty Fluorochromes

Novel and specialty fluorochromes are often proprietary and may only be available from one source. Therefore, it is particularly incumbent on the manufacturer to provide well-characterized materials when such fluorochromes are sold for use in QFC.

8.7.1 Tandem Fluorochrome Conjugates

Tandem fluorochromes increase the Stokes shift by using the fluorescent radiance from one as excitation energy for the other. The most common tandems combine phycoerythrin with an organic dye such as Texas red. Tandem conjugates have not been commonly used or systemically evaluated for QFC applications,

An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 33 Number 26 NCCLS and their dependence on fluorescent resonance energy transfer (FRET) makes them relatively unstable (so-called “leakage,” due to suboptimal covalent linkage of the two fluorochromes) and subject to spectral changes upon aging.109

8.7.2 Lanthanide Chelates

These chelated rare-earth elements such as europium are used as fluorochromes in time-resolved fluorescence systems. They have long decay times, and the instruments used to measure them excite with microsecond-pulsed light while recording fluorescence emission only during the dark phase of each pulse. The elements themselves are highly stable, but the chelated forms used to label proteins can dissociate and lose fluorescence. No authoritative reference solutions for these fluorochromes were available at the time of this guideline publication.

8.7.3 Nanocrystals

Nanocrystals are nanometer-sized particulate semiconductors that exhibit intriguing optical properties.110,111 To be useful in biological applications, nanocrystals are coated with molecular shells that eliminate non- radiative relaxation pathways, thereby improving quantum yield and reducing photodegradation. The solubility of the nanocrystals can be tuned using further coating layers.

Nanocrystals are fairly efficient fluorochromes, but their long fluorescence lifetime reduces their usefulness in flow cytometry, where the short period in the sensing volume does not allow the collection of many emission photons. However, the long fluorescence lifetime combined with resistance to photodegradation make nanocrystals especially well-suited for scanning platforms and time-resolved fluorescence measurements.

The emission spectrum of nanocrystals is quite narrow compared to organic fluorochromes; it may be as little as 12 nm bandwidth for a single nanocrystal. Since the emission spectrum depends on the size of the nanocrystal, it can be tuned by controlling the size. Furthermore, a nanocrystal sample can be prepared to be highly monodisperse (5% variation in size), so the overall spectral width of nanocrystal emission can be very narrow (typically 30 nm or less). Conversely, the excitation spectrum of nanocrystals is very broad, so they can be excited by a wide range of wavelengths. A mixture of nanocrystals with different emission peaks may be excited efficiently by light of a single wavelength, which facilitates simultaneous detection, imaging, or quantification. This is of particular benefit where a laser or spectral line source is being used for excitation, allowing separation of the incident and emission light and reducing background due to scattering. However, the ultraviolet and blue wavelengths that most efficiently excite nanocrystals also increase the risk of photodamage to cells, particularly in imaging techniques that use repetitive scanning.

Nanocrystals are typically very stable. Composed of simple inorganic compounds, they are chemically inert, and those with outer shells are also very resistant to photodegradation. The outer shell further instills resistance to changes in fluorescence caused by environmental conditions such as pH and protein concentration. No authoritative reference solutions for nanocrystals are available at the time of this guideline’s publication. See http:/www.npl.co.uk/biotech/fluorescence.html (National Physical Laboratory, United Kingdom). Nanoparticles supplied with reactive groups for conjugating to proteins are becoming commercially available.

9 Solid Phase Fluorochrome Reference Materials for QFC

9.1 Purpose and Criteria for Solid Phase Fluorochrome Reference Materials

As with all reference materials, solid phase fluorochrome materials must have certain property values that are sufficiently homogenous and well established to be used for the calibration of an apparatus, the assessment of a measurement method, or assigning values to other reference materials. Solid phase

34 An NCCLS global consensus guideline. ©NCCLS. All rights reserved. Volume 24 I/LA24-A fluorochrome reference materials are used to calibrate instruments that measure the fluorescence intensity (FI) of particulate objects rather than solutions. They can be used to determine the linear dynamic range and assess the sensitivity of FI measurements. If they are properly calibrated in MESF units, they can be used to convert arbitrary relative fluorescence intensity (RFI) readings into standardized molecular equivalent values (molecules of soluble fluorochrome, or MESF) that preserve stoichiometry. The use of MESF- calibrated reference materials is the basis of quantitative fluorescence calibration (QFC).

This section is intended to provide guidance for manufacturers and suppliers and to help end users select appropriate reference materials for their QFC applications.

9.2 Microspheres as Solid Phase Fluorochrome Reference Materials

Uniform cell-sized spherical microparticles (microspheres, also called microbeads) containing one or more fluorochromes are the most often used calibrators for stained cells. They are so widely used in flow cytometry that a special classification scheme for the different types and applications has been presented (see Appendix F).50 Fluorescent reference microspheres can also serve as calibrators for image-based FI measurements, although correction for polarization effects may be required. The remainder of this section will therefore, focus on fluorochrome-labeled microspheres as reference materials for QFC.

9.2.1 Chemical Composition and Fluorochrome Environment

Microparticles may be synthesized in a range of sizes from a variety of polymeric materials. They are fundamentally hydrophilic or hydrophobic. Hydrophilic microparticles are synthesized from monomers such as hydroxyethyl methacrylate (HEMA). They are not well-suited as reference materials because they are highly variant in size and shape.

The majority of fluorescent reference microspheres have been produced from hydrophobic materials such as polystyrene and polymethyl methacrylate (PMMA). Using specialized swelling and solvent mixing techniques, highly-uniform microspheres embedded with fluorochromes can be produced in sizes appropriate for cellular standards (2 to 20 microns in diameter) with less than 2% CV in fluorescence intensity. Fluorochrome molecules may also be immobilized by adsorption to the surface of hydrophobic microspheres. Finally, fluorochrome molecules may also be covalently attached to functional groups on the surface of hydrophobic microspheres, and fluorochrome-conjugated proteins can be bound to molecular receptors (usually antibodies) similarly attached.

9.2.1.1 Microspheres with Fluorochrome Molecules Embedded in the Polymer (Hard-Dyed)

Microspheres with embedded fluorochromes tend to give more stable and uniform FI than microspheres with surface-bound fluorochromes. Since embedded fluorochrome is sequestered from the aqueous chemical environmental, it is unaffected by many potential sources of variation and degradation. Microspheres made by embedding mixed fluorochromes can be calibrated in relative fluorescence intensity (RFI) values that apply across a wide range of emission wavelengths for given excitation conditions. Such microspheres are well-suited for aligning instruments and for monitoring the linear dynamic range of FI measurements at different wavelengths with just one calibrator set. However, RFI values assigned under particular excitation conditions (wavelength and incident power) may not be applicable under different excitation conditions, so the excitation conditions used to assign RFI values should be specified.

9.2.1.2 Microspheres with Fluorochrome Molecules Covalently Attached to the Surface

Microspheres with fluorochrome molecules attached to the surface are more difficult to prepare uniformly than the hard-dyed microspheres, and the distribution of FI in a given population may be much broader. In addition to the intrinsic variance, further variance may arise from influences of the aqueous environment on the exposed fluorochrome molecules. This source of variance is an essential property of reference

An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 35 Number 26 NCCLS microspheres calibrated in MESF units, which are used in QFC as the standards for comparative fluorescence yield measurements. The assumptions implicit in this system are that absorptivities of fluorochromes on the calibrators and their respective measurands are comparable and that changes in quantum yield caused by the environment are also comparable (see Section 5.2.2.2). Thus, MESF reference microspheres should have the same environmental responsivity as the stained measurands.

MESF values assigned to fluorochrome reference microspheres can be used to produce the same parameters obtained from an RFI calibration curve (see Section 11.4). In addition, the MESF calibration curve allows arbitrary FI readings to be converted into standardized units that can quantify fluorochrome-ligand conjugate binding on labeled cells and particles.

9.2.1.3 Microspheres with Fluorochrome Conjugates Bound to Surface Receptors

Microspheres may also be labeled by binding fluorochrome-ligand conjugates to appropriate receptors (usually antibodies) absorbed or covalently attached to the microsphere surface. These capture microspheres can serve as calibrators for fluorescence yield, since the fluorochrome molecules on the ligands they bind are exposed to the aqueous environment. In addition, spectral matching between capture microspheres and conjugate-labeled measurands such as cells may be better than the spectral matching with microspheres that have been directly labeled with fluorochromes. However, the precision and stability of FI is generally much lower with capture microspheres than with direct fluorochrome-labeled microspheres.

MESF values cannot be assigned to capture microspheres, since their FI will depend on the particular fluorochrome-ligand conjugate which they have bound. Capture microspheres can be assigned values in molecular units of antibody binding capacity (ABC). However, due to variance in the true binding capacity for different antibody-fluorochrome conjugates,58,59 these values to date have not given consistent results. While the properties of capture microspheres are not ideally suited for instrument evaluation and monitoring, the FI calibration curve they produce contains, in theory, all the information available from the other types of fluorochrome reference microspheres.

9.3 Property Values for Fluorochrome Reference Microspheres

9.3.1 Relative Fluorescence Intensity (RFI)

Multiple populations of reference microspheres containing different amounts of fluorochrome may be calibrated relative to each other with values assigned in arbitrary units based on their proportionate fluorescence intensity. By analogy with transmittance measurements, the microsphere with the highest FI in a set can be assigned a value of 100% (or some other arbitrary value), and lower FI values assigned proportionately. RFI values for fluorochrome-labeled reference microspheres are generally assigned using flow cytometry. The accuracy of these assigned RFI values depends on the linearity of the FI response, and instruments used to assign RFI values should meet the criteria for linearity and dynamic range given in Section 6. Assignments should be made only within the documented range of linearity.

9.3.2 Comparative Fluorescence Yield (in MESF) and Emission Spectrum

For purposes of QFC, the most essential property value of fluorochrome-labeled reference microspheres is their fluorescence yield compared to a fluorochrome solution. The comparative fluorescence yield is expressed as molecules of equivalent soluble fluorochrome (MESF). MESF values for primary reference microspheres should be assigned by direct comparison with fluorochrome reference solutions.71,72

It should be noted that the assignment of MESF values to microspheres by comparison to fluorochrome solutions is based upon comparing the entire emission spectrum of each. Therefore, spectral matching is not required. However, when cells are analyzed on a flow cytometer or other instrument where barrier filters are used to capture a segment of the emission spectrum, spectral matching between the microspheres and the

36 An NCCLS global consensus guideline. ©NCCLS. All rights reserved. Volume 24 I/LA24-A measurand should be as close as possible. Similarly, primary fluorochrome reference microspheres may be used to assign MESF values to secondary reference microspheres only if their emission spectra match closely. The extent to which spectral mismatch contributes to the overall variance in QFC has been discussed.71

9.3.3 Coefficient of Variation (CV)

Whether calibrated in RFI or MESF, a population of fluorochrome reference microspheres with a single assigned value will show variance in the FI of individual events. The coefficient of variation expressed as a percentage (% CV) is the best indicator of this variance. (NOTE: The % CV should not be calculated from log-transformed results.

In microspheres with embedded fluorochromes, less than 2% CV is desirable. Microspheres labeled with surface fluorochromes give higher variance, and up to 15% CV is acceptable. Capture microspheres stained with fluorochrome-labeled conjugates often show even higher within-population variance, which can be minimized by adhering to a consistent staining protocol. The broader CV of surface-labeled and capture microspheres should not contribute inaccuracy to their assigned values, but it may complicate the identification of individual populations when microspheres with different values are pooled.

9.3.4 Environmental Sensitivity of MESF-Calibrated Microspheres

Environmental influences such as pH should have the same effects on the spectral fluorescence of solid- phase fluorochrome reference materials calibrated in MESF units as they do on fluorochrome solutions and labeled cells.

9.3.5 Physical Properties

Since reference microspheres must be detectable as individual events by the instrument measuring them, their physical properties that are relevant to measuring FI should be as close to those of the measurand (typically cells or other microspheres) as possible.

9.3.6 Polarization

The degree of polarization anisotropy of a microsphere-bound fluorochrome should be determined as described in Section 5.2.2.4, using appropriately polarized illumination and detection.

9.3.7 Absorptivity and Relative Quantum Yield

For the basic reference method of QFC, the absorptivity of a fluorochrome on labeled reference microspheres is presumed to be comparable to that of the soluble fluorochrome. For NIST reference material RM 8640, a fluorescein-labeled microsphere (see Section 9.5), preliminary studies suggest a maximum difference in absorptivity on the order of 10% between soluble and immobilized fluorescein. A method for the exact measurement of the absorptivity of immobilized fluorochromes is in development at NIST. Since any difference in absorptivities would create a constant proportionate bias, assigned MESF values can be adjusted retrospectively to take them into account.

The relative quantum yield is implicit in the MESF value assigned to solid phase material. The relative quantum yields can be determined from measured fluorescence lifetimes.112,113

9.4 Other Specifications for Fluorochrome Reference Microspheres

9.4.1 Intrinsic Fluorescence (Autofluorescence)

Many unstained microspheres emit background fluorescence when illuminated, especially by lasers. The FI signal from unstained microspheres should be as low as possible to provide a true indicator of instrument background when they are used as a “blank” and to avoid the complications of mixed fluorescence signals that are not spectrally matched when they are labeled. A reasonable criterion for microsphere background fluorescence is that it is not more than the intrinsic fluorescence of the measurand (e.g., cellular autofluorescence).

9.4.2 Suspension Media

Like the matrix of fluorochrome reference solutions, the matrix in which fluorochrome reference microspheres is suspended is critical. The exact composition of the solvent should be specified, including buffer components, protein additives, and any other materials. Since microsphere suspensions will be diluted into various media for different applications, the simplest stable media are advisable to avoid incompatibilities. The microsphere concentration should be sufficiently high for use in working dilutions where the environmental properties of the assay media are the only significant influence on fluorescence and the count rate is optimal.

9.4.3 Storage, Stability, and Expiration

Fluorochrome reference microspheres should be properly packaged and accompanied by recommendations for optimal storage. Heat and light are the major sources of instability to most of the commonly used fluorochromes. Amber or light-free containers are preferred to minimize photodecomposition. Long-term stability is generally increased by refrigerated storage. Freezing should be avoided, as it often fractures microspheres.

Reference microspheres with embedded fluorochromes will generally be more stable than those labeled with surface fluorochromes. With dark refrigerated storage, microspheres containing embedded fluorochromes can remain stable for over a decade. Specific expiration dates should be assigned by suppliers based on real time (shelf life) studies, but the end user should also monitor stability by ongoing comparison with other materials and instrument target settings.

9.4.4 Biohazards

Microsphere suspensions of commonly used fluorochromes other than those labeled with dyes that bind to DNA do not present a health hazard. Constituents of the suspension matrix that may be biohazardous (such as azide) should be documented.

9.5 Sources of Fluorochrome Reference Microspheres

The only authoritative fluorochrome reference microsphere available at the time of this guideline’s publication is RM 8640, a set of calibration microspheres surface-labeled with fluorescein (see Appendix D). The MESF values of RM 8640 were assigned at NIST by comparison with SRM 1932 fluorescein reference solution.

A variety of other fluorochrome reference microspheres may be obtained from commercial sources.50 In general, the values assigned to them may be considered reasonable estimates of RFI or MESF values to the extent that they have been tested against other calibrators or evaluated in multicenter studies.26,39,54,56,59,61,62,64,114,115 However, as of the publication date of this guideline, none of these microsphere standards have been authenticated independently or compared to SRM or RM fluorochrome

38 An NCCLS global consensus guideline. ©NCCLS. All rights reserved. Volume 24 I/LA24-A reference solutions. It is anticipated that additional reference microspheres traceable to authoritative standards will become available through partnerships between the NIST fluorescence intensity standards program and those who supply and use such materials.

10 Quantitative Ligand-Binding Assays Using Fluorochrome-Ligand Conjugates (FLC)

Fluorochrome-ligand conjugates (FLC) are commonly used in ligand-binding assays. Microspheres or cells may serve as a solid phase, separating bound and free ligands. The staining reactions used to phenotype cells by antibody binding to specific receptors are a particular type of solid phase ligand-binding assay. Microspheres may be coated with specific receptors (usually antibodies that bind a particular antigen) and used as the solid phase in ligand-binding assays to measure the concentration of soluble ligand molecules, either by direct binding (usually with the conjugate as the final component in a sandwich assay), or by competition assays in which the native ligand in the analyte or standard displaces the conjugate.

Two major goals have driven the development of methods to quantify FLC binding. The first objective is to quantify the expression of cellular proteins, a major attribute of cell lineage, differentiation, activation, and transformation. The number of FLC specifically bound by a cell can often provide an indirect but highly useful measure of protein expression. The second objective is to characterize the conjugates and microspheres used for receptor-ligand assays in suspension arrays.

This section describes the general considerations for reagents and methods required for ligand-binding assays using QFC. Practical methods for the end user to perform such assays are described in Section 11.

10.1 The Relationship Between Equivalent Fluorescence and FLC Binding Values

The MESF unit provides an equivalent fluorescence scale that relates the measured FI values of stained particles to the molar quantity of bound FLC. The quantitative relationship depends on the effective F/P ratio of the FLC solution (see Section 10.2.1). Alternatively, capture reference particles binding known molar quantities of FLC can be used to establish a direct relationship between the particle FI and bound FLC. The agreement in the results of these two approaches is perhaps the best overall indicator of the accuracy of the binding values obtained from each.

10.2 Molecular Properties That Influence Fluorochrome-Ligand Conjugate Binding

Both the fluorescence and binding properties of FLC are critical to the performance of binding assays, and each requires characterization for accurate quantitative results.

10.2.1 Fluorescence Properties of Fluorochrome-Ligand Conjugates

Conjugation of a fluorochrome to a ligand (usually a protein) often changes its fluorescence properties, such as reducing its quantum yield or causing a shift in its emission spectrum. Using the principle of fluorescence yield (see Section 7), the equivalent fluorescence of FLC solutions can be expressed in MESF units which closely approximate the MESF values obtained from properly-calibrated microsphere standards. The equivalent fluorescence of an FLC solution is usually expressed as a ratio (moles of equivalent soluble fluorochrome per mole of FLC, or MESF per molecule of FLC). This value is called the effective F/P ratio, or effective F/P. It will differ from the molar F/P ratio unless the absorptivity and quantum yield of the conjugated fluorochrome molecules are identical to those of free fluorochrome molecules. In practice, the effective F/P can be determined by comparison with reference fluorochrome solutions using cuvette fluorometry (see Sections 11.4.1 and 11.4.2).71

10.2.2 Binding Properties of Fluorochrome-Ligand Conjugates

Conjugation can also change the binding properties of the ligand for its homologous receptor or to unrelated molecules due to denaturation and steric hindrance. FLC may have lower or (uncommonly) higher affinity than native ligand. In extreme cases, FLC may fail to bind at all or may bind nonspecifically to a variety of targets.

10.2.3 Microheterogeneity in the Molecular Properties of Fluorochrome-Ligand Conjugates

Ideally, FLC are fairly homogeneous at the molecular level and can be displaced by free ligand in an approximately equimolar fashion (i.e., the equilibrium constant for free-ligand binding is similar to that of the conjugated ligand). In fact, most FLC solutions contain mixtures of molecular species with fluorochrome molecules covalently attached at varying ratios to a diversity of sites on the ligand. Both the binding and fluorescence properties of these heterogeneous molecules may differ from one to another.

10.2.4 Unimolar Fluorochrome-Ligand Conjugates

The use of purified FLC in which nearly all the species have one fluorochrome molecule attached to one ligand molecule can greatly improve the consistency of ligand-binding assays.14,60,61 Such conjugates are probably the best candidates for use in ligand-binding assays based on QFC.

10.3 Quantifying the Molar Quantity of Fluorochrome-Ligand Conjugates Bound to Particles

10.3.1 Quantifying FLC Binding by MESF Calibration

An MESF calibration curve derived from fluorochrome-labeled reference particles can be used to determine the MESF of a particle with bound FLC. With the assumption that the fluorescence yield of FLC does not change appreciably upon binding to a particle, the molar quantity of bound ligand can be determined from the MESF of the particle and the equivalent fluorescence of the FLC:

MESF(particle) = (MESF per FLC molecule)* (N(FLC molecules)) or N(FLC molecules) = (MESF(particle) ) / (MESF per FLC molecule).

Thus, QFC provides a way to convert FI measurements on particles to the molar quantity of bound FLC, usually expressed as FLC molecules per particle.

10.3.2 Quantifying FLC Binding by Calibrating FI Measurements Directly in Binding Units

Under the tenets of QFC, the FI of particles with bound FLC should be directly proportional to the number of FLC molecules bound. Therefore, a dose-response curve of FLC bound vs. FI should be parallel to an MESF vs. FI dose-response curve; that is, both curves should be linear with unity slope in log space. If a known FLC molar quantity (number of molecules) is bound to a particle, the FI reading of that particle can be used to translate the MESF vs. FI curve into FLC binding values (this is mathematically equivalent to dividing the MESF values by the effective F/P of the FLC). If a series of such particles binding varying amounts of FLC is available, a calibration curve can be constructed in units of FLC molecules per particle without any need for MESF units.50,56-59

The primary drawback to this approach is that accurate binding values for microparticle calibrators are difficult to determine and may not apply to different combinations of ligands and fluorochromes.54-59 Even lot-to-lot variation can affect binding values for a given FLC (unpublished observations by participants on the subcommittee).

10.3.3 Biologic Calibrators for FLC Binding

Biologic calibrators may be a practical approach for standardizing certain FLC binding measurements. One of the earliest approaches to QFC used cell lines expressing CD5 as a biologic calibrator (see Section 11.4.5).12 More recently, freshly stained human CD4 T-cells obtained from peripheral blood have been used to adjust the relative FI calibration curve to a CD4-FLC binding curve (see Section 11.4.3),61 and CD64 on neutrophils and monocytes have been used in the same manner (see Section 11.4.6).124 One advantage of biologic calibrators is that they more nearly approximate the binding properties and microenvironment of cellular measurands. Concerns over biologic variability and biosafety remain impediments to their widespread use in the clinical laboratory, but they will have a continuing role in research and development.

10.4 Sources of Variability in Measuring FLC Binding Values

Differences in the effective F/P ratios, valencies of binding to cells and capture microspheres, affinity or avidity, and the extent of nonspecific binding and nonspecific fluorescence and all influence the measurement of FLC binding values. Each reaction between FLC and a particular receptor is unique, and the relative contributions of these and other potential factors must be evaluated individually.

10.4.1 Concentration and Binding Activity of Fluorochrome-Ligand Conjugates

Most ligands for cellular receptors and many soluble biologic analytes are peptides or proteins. Their molar concentration must be determined from the total protein, the proportion of protein represented by the FLC, and its molecular weight. While this guideline does not address these measurements in detail, careful determinations are essential to obtaining an accurate molar fluorescence.

A molar fluorescence value based on the FLC concentration assumes that all the FLC can specifically bind to their homologous receptors. However, conjugation with a fluorochrome can disrupt protein conformation or cause steric hindrance, reducing or inactivating the ligand-binding properties. Erroneous binding results may be obtained if a significant proportion of the measured ligand concentration is not active or has reduced affinity for the receptor binding site. Competition assays can be used to compare the binding affinity of native ligands with the affinity of FLC.116 The measurement of the effective F/P ratio (MESF per FLC molecule) by comparison with free fluorochrome reference solutions presumes that all the fluorochrome molecules in the conjugate solution are bound to conjugate molecules. The presence of free fluorochrome or fluorochrome bound to nonligand molecules will cause overestimation of the effective F/P ratio.

10.4.2 Fluorochrome Reference Solutions

When FLC solutions are compared with fluorochrome reference solutions to obtain their effective F/P ratios, inaccuracy in the fluorochrome concentrations of the reference solutions will cause error in these effective F/P values. Fluorochrome reference solutions should be characterized as completely as possible and preferably traceable to an authoritative source such as a NIST reference material.

10.4.3 Assigned Values for FLC Binding Calibrators

Synthetic (microspheres) and biologic (cells) calibrators with preassigned values for FLC binding may show different values for different FLC, including different lots of the same FLC.

10.4.4 Binding Equilibrium and Saturability

The extent to which binding has reached equilibrium and to which receptor sites are saturated can vary with reagent, staining protocol, and other preparative steps, and nature of the receptor on the target (cell or microsphere) for FLC binding. Most methods to quantify cellular receptors presume saturation binding by the cellular measurand (so the binding capacity truly reflects the number of receptors) and saturation binding

An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 41 Number 26 NCCLS by the calibrators (so predetermined binding values can be reproduced accurately). To ensure the validity of this assumption, the binding kinetics of each FLC receptor on measurands and calibrators should be characterized before use in clinical applications, and the staining protocol should be standardized on the basis of the kinetics.

10.4.5 Multiple Interactions Between Fluorochrome-Ligand Conjugates and Binding Targets

In systems where more than one FLC (or primary antibody in indirect assays) is present during the binding reaction, reagent interactions between the various ligand-receptor reactions may cause alteration in the extent of binding. This effect can complicate both multicolor cellular immunophenotyping and multiplexed microsphere suspension array assays.

The most common problem in multicolor cellular immunophenotyping is variation in staining by a labeled ligand when it is conjugated to different fluorochromes. Variation can even occur when the same FLC is used in different mixtures containing other FLC. This effect is seen most often with PE FLC and may be caused by several different factors. The most obvious variation is diminished or absent staining, probably caused by steric hindrance from the other FLC binding to cellular receptors in close physical proximity to the PE FLC. Also, nonspecific binding of a particular FLC can be altered by the presence of other FLC and may, therefore, differ between subsets, complicating the identification of clusters and strategies for gating. A third reason for variation in staining is the presence of donor immunoglobulin that may contribute to nonspecific binding when unwashed whole blood is used, particularly with PE FLC and EDTA blood samples.117

To minimize the above difficulties, FLC oversaturation should be avoided by titering to the lowest saturating dilution. Since saturated binding is desired in binding capacity measurements made by QFC, the working dilution of FLC used for this purpose should be carefully optimized. Ideally, the results of an FLC binding assay in the absence of other reagents would serve as a reference point for comparing results when multiple FLC are present. However, this may not always be possible, since some mixture of other FLC may be required to identify the subpopulation of cells on which the FLC of interest is being titered. Moreover, different cellular analytes (such as normal cells and malignant cells from different cancers) may show different discrepancies. To reduce this problem, the particular mixtures (so-called cocktails) of FLC used for staining clinical samples should be tested and optimized for measurands of interest (e.g., normal and malignant cells).

The most common problem caused by multiple interactions in multiplexed suspension arrays is an overall increase in background as the number of analytes increases, which can cause a decrease in sensitivity for most or even all of the analytes. These assays are generally multiple-layer indirect formats in which the same FLC (often an avidin-PE conjugate) is used as the final ligand for the assembled complexes on all the microsphere subpopulations. Although assay development is often aided by evaluating reactions individually, the final performance characteristics of such assays should be determined in their multiplexed configuration.

10.5 Using QFC to Measure Binding Constants and Affinities

The approach to using QFC for measuring thermodynamic parameters of ligand-binding assays is entirely analogous to those using radioactive tracers. Knowledge of the free (unbound) FLC concentration and the average FLC molar quantity bound per particle for each dose of a titration provides the necessary parameters for Scatchard-type analysis. These parameters may also be obtained entirely from flow cytometric measurements of the bound FLC using the method of isoparametric titration, which is a model- independent analysis based on a checkerboard titration where FLC concentration is varied in one dimension and particle concentration in the other.14

11 Practical Methods for Quantitative Fluorescence Calibration in Clinical and Research Applications of Cytometry

The goal of QFC in flow cytometry is to convert relative fluorescence intensity (RFI) readings into standardized units that serve three basic purposes. First, they provide an objective scale for displaying FI results in a quantitative window of analysis that can be compared across different instruments. Second, they allow an objective assessment of the FI measurement. Finally, they allow the molar quantity of fluorochrome-labeled conjugate (FLC) to be determined from the FI readings on calibrated instruments. The FLC binding capacity is often used as a measure of receptor expression, an increasingly important determination with the emergence of therapeutics targeted at specific cellular receptors.74

This final section presents practical methods for generating and evaluating the calibration curves, establishing an instrument-independent scale in equivalent fluorescence units, and converting equivalent fluorescence units to molar quantity units of FLC in solution or bound to particles. These methods have been adapted from published reports and presented in the overall context of this guideline.

11.1 Basic Reference Method for Constructing the FI Calibration Curve

The calibration process maps a relationship, historically called the calibration curve or the dose-response curve, between the instrument reading and the amount of the substance being measured. The proper construction and interpretation of a calibration curve for QFC requires the correct selection of standards as well as the correct transformation of units used to report the FI reading.

The simplest FI standards are those that have been assigned relative fluorescence intensity (RFI) values. An RFI calibration curve reveals the linearity and dynamic range of the FI measurement on a particular instrument, but it provides no information about the molar quantity of fluorochrome or fluorochrome- labeled conjugates (FLC) on stained particles. Standards which have been assigned values in molecules of equivalent soluble fluorochrome (MESF) do establish an instrument-independent scale that accurately depicts the molar quantity of fluorochromes on stained particles. MESF calibration also determines whether the dose-response assumptions behind QFC (see Section 5) are valid approximations in the range bracketed by the standards. Standards which have been assigned binding capacity units also provide an instrument- independent molar quantification scale with the same criteria for validity as the MESF scale, but the scale is applicable only for the particular FLC used to stain the standards.

The QFC model defines an ideal calibration curve that is linear in log space (log dose vs. log response) and has unity slope. Therefore, when a graph depicts the FI reading against the assigned FI value in log space, the region that is not significantly different from a straight line with unity slope is the range over which the assumptions behind QFC are valid approximations. Since these approximations fail at the extremes of fluorescence intensity, the basic method requires that QFC results be interpolated from within a demonstrably linear range of the instrument response.

Since the basic reference method for QFC is based upon a linear calibration curve (in log space), it can be implemented formally using the ISO standard for linear calibration.118 The rigorous approaches in this standard have been simplified in a working system widely accepted for practical flow cytometry.7,89 This section of the guideline describes the simplified approach and indicates where it departs from the more formal ISO standard.

11.2 The FI Calibration Plot

The log-log calibration plot (see Figures 6 and 7) provides an objective assessment of the relevant parameters of the FI measurement as well as the capacity to measure labeled analytes. Calibration parameters provide the most objective way to validate instrument performance and compare measurements

An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 43 Number 26 NCCLS across time, instruments, and platforms. Monitoring calibration parameters provides the most comprehensive instrument quality control and assurance of measurement accuracy.

Figure 6. Graphic Depiction of Data and the Best-fit Line from Flow Cytometric Analysis of Seven Populations of Fluorescent Microspheres, Each of Which had an Assigned Value for Relative Fluorescence Intensity (RFI). The independent (dose) variable is shown on the X-axis and the dependent variable (response) on the Y-axis. This depiction is consistent with the linear regression, which uses a simple least-squares method to determine the best-fit line. The accompanying table shows the parameters of the regression as performed with all seven standards and of other regressions in which the lowest and/or highest standards have been removed. The regression using standards 2 to 7 had the best combination of fit (low residuals), responsivity (slope closest to unity), and linear dynamic range (about 200-fold).

100

10 H

i s

)

to Ev

F g

0 e

R r

1 a

t t m

e C

s C o

u h

n a

n t

H n

( e

0.1 l

-2 1 10 100 1000 10,000 0 255 511 767 1023 Measured RFI (Linear Units 1-10,000 or Histogram Channel Number 10-bit 4-decade log)

Figure 7. The Same Data Shown in Figure 6 are Depicted with the X-axis and Y-axis Exchanged, so that the X-axis Depicts Instrument Response, and the Y-axis Depicts Both the Dose (Assigned RFI Value) and the Event Count. This allows the histogram event-count peaks to align with the observed points on the calibration curve. Despite this depiction, the least-squares line was still regressed as in Figure 6, with dose (assigned RFI value) as the X-variable and response (cytometer RFI reading) as the Y-variable.

11.2.1 Constructing the Calibration Curve

Ideally, the FI calibration curve should be derived from reference standards that are evenly distributed in logarithmic space across the desired range of the measurement such that each half-decade (about a three- fold range) has at least one observation point. The curve should be constructed by simple best-fit linear regression of the log values for fluorochrome dose as the independent (X) variable and the corresponding log values (electronically or computationally transformed) of the instrument readings as the dependent (Y) values.7 In advanced applications such as the assignment of MESF values to secondary standards, the uncertainty (error) in both the assigned values for the independent variable and the response values for the dependent value should be taken into account when performing the linear regression. 118,119

The regression should be performed with dose as the X and response as the Y values (see Figure 5), even though the curve is sometimes depicted with the axes switched so that the calibration curve can be superimposed on the histogram (see Figure 6). In this depiction, the response is on the X-axis (RFI reading or histogram channel number), and there are essentially two Y-axes, one for event count of the histogram channels and the other for dose value of the calibration.

11.2.2 Units for the Independent (Dose) Variable

The dose values for a QFC calibration curve using microparticles as standards are assigned by the supplier or user and assumed to be accurate.118 The values can be scaled in arbitrary units of relative fluorescence intensity (RFI values), molecules of equivalent soluble fluorochrome (MESF values), or molar quantities of bound fluorochrome-ligand conjugates (FLC). The value assigned to microsphere standards that bind FLC is often called antibody binding capacity (ABC) of the standard, which is taken to mean the amount of FLC bound in the presence of excess unbound FLC (i.e., near saturability).

RFI units are applicable only to a set of standards comprising multiple populations, each with a different FI. The RFI value is often assigned as a proportion or percentage of the highest standard, although any scale may be used as long as the proportions are accurate. MESF units should only be assigned to standards that are environmentally-responsive and spectrally-matched to a specified fluorochrome. Binding units such as ABC should only be assigned on the basis of an independent measure of the molar amount of FLC bound.

Multiple populations of microsphere standards calibrated in MESF or FLC binding units can also be assigned RFI units by normalizing all values to the value of a selected standard (e.g., the highest). Conversely, an RFI calibration curve (calibrators not spectrally matched or environmentally responsive) can be converted into MESF or FLC values using a single MESF or FLC calibrator if and only if their respective calibration curves are parallel.

11.2.3 Units for the Dependent (Response) Variable

Different instruments and platforms use a variety of scales and units to report their response to fluorescence emission, including photon counts, relative values, and histogram distributions. If instrument readings are discrete numerical values from linear data (e.g., photoelectron counts), the numerical values should be used as the response variables for the regression. If the readings are distributions of individual measurements (e.g., a histogram of many microparticle measurements binned into histogram channels), a single parameter of the distribution must be selected to use as the response variable. Of the common parameters used to represent central tendencies of distributions, the median is probably the most practical choice for the end user as the response value. It makes no presumption about the shape of the distribution, it is suitable for either linear or log-transformed data, and it is valid as long as at least half the events in the distribution lie within the calibration range. However, it should be noted that reference values assigned to microspheres in bulk (e.g., by comparing the FI of microsphere suspensions with the FI of fluorochrome solutions) are actually mean values (the total FI of the suspension divided by the concentration of microspheres). The mean of the histogram distribution (in linear space) may, therefore, be a better parameter for transferring values between different populations of microspheres.

Instrument readings from linear data should be converted to logarithmic values for the regression. Instrument readings from log-amplified signals can be used directly, since they have been electronically log transformed. However, most cytometers mathematically convert these measured logarithmic values into a presumed linear scale, often called the “relative fluorescence intensity” or “relative linear channel.” The converted results are not true instrument readings, but rather arithmetic conversions based on the presumption of perfect logarithmic amplification across the nominal dynamic range of the instrument. The mean value reported from these linearized conversions of log-transformed histogram channel distributions may be either geometric or arithmetic. If the mean of the original histogram channel distribution is computed and then converted, the result is a presumed geometric mean. If the mean value is calculated after a channel-by-channel conversion of the histogram, the result is a presumed arithmetic mean. The median channel value will be the same whether it is determined from the original histogram then converted or determined after channel-by-channel conversion. Thus, the most consistent choice for a response value is to use the median channel.

Newer fluorescence detection instruments increasingly use digital signal processors (DSP) in lieu of analog devices such as log amplifiers and pulse analyzers. The readings reported by such instruments depend heavily upon the resolution of the digitizers, upon algorithmic signal manipulations in the DSP firmware, and upon the software used for data analysis and display. Although experience with formal FI calibration on these platforms is limited, the same general principles apply. The readings from DSP-based instruments should be used in the same manner as those from analog systems, and significant departure from ideal or nominal values should be further investigated with respect to both hardware and software.

It should be noted that this FI calibration method does not preclude instrument-specific software conversion of signals into corrected response values based on independent calibration or modeling of electronic

46 An NCCLS global consensus guideline. ©NCCLS. All rights reserved. Volume 24 I/LA24-A components such as baseline restoration circuitry. Such corrections may allow extensions of the calibration curve in the regions that deviate from linear behavior in a predictable fashion. However, they should be used cautiously in QFC because such deviations may also be due to failure of the basic approximations in the measurement equation.

11.3 Evaluation of Calibration Parameters

The formal statistical evaluation of a linear calibration curve involves analysis of variance and residuals in the response (Y) variable.118 The practical method most commonly used in flow cytometry7 evaluates the coefficient of determination (R-squared), the residuals in the dose (X) variable, the slope, the lower and upper measurement boundaries, and the detection threshold.

11.3.1 Coefficient of Determination (R-squared)

The coefficient of determination is the square of the correlation coefficient. Although it is often used as an indicator of linearity, it is not very informative for QFC because the inherent linearity of the FI measurement and the wide measurement range mean that the correlation coefficient will be close to unity even in instruments with unacceptable linearity. However, the correlation coefficient does establish an objective criterion for the suitability of using simple least squares as the best-fit procedure. If the correlation coefficient is less than 0.99, the estimates of slope and intercepts will be biased by greater than 1%, and least squares is not appropriate.120 Virtually any fluorometer should be sufficiently linear over the working range of the assay to use least squares calculations for linear regression. If the correlation coefficient is below 0.99, the instrument and standards should be further evaluated.

11.3.2 Linearity and Residuals

Linearity is the ability (within a given range) to provide results that are directly proportional to the concentration (dose) of the analyte. The formal method of evaluating linearity uses the standard deviation of the Y-residuals (response scale) and the “runs” test118. The simplified method for flow cytometry uses the X-residual (dose scale), since it allows consistent evaluation across the wide variety of FI response scales.7 The residual is commonly expressed as a percentage above (positive residual) or below (negative residual) the assigned value. The absolute value of the percent residual indicates the proportion of disagreement without regard to positive or negative bias. The average absolute percent residual gives the average percent disagreement between the assigned and best-fit dose values for all data points and is the simplest aggregate parameter of overall linearity. It corresponds well with the standard deviation of the Y-residual (see Figure 5). With valid FI standards, the calibration curve on any instrument should have less than 5% average absolute residual.7,33,54,56,59

11.3.3 Slope and Coefficient of Response

The slope of the linear regression line is expressed in units of log instrument response per log fluorochrome dose. If the X and Y variables are in linear units transformed to log values before regression, the slope has an ideal value of unity, reflecting the direct proportionate relationship between dose and measured FI values. QFC curves on flow cytometers generally have values within 5% of ideal (0.95 to 1.05).7,54

When response values are histogram channel numbers binned from log-amplified signals, the slope is expressed in channels per ten-fold increase in fluorochrome dose, which has been termed the coefficient of response.7,54 With these units, the dynamic range of the log amplifier can be determined by dividing the response coefficient into the total number of histogram channels. This observed value for dynamic range can be checked against the nominal value (e.g., 4 decades). Historically, the simplified approach used in flow cytometry standardized the response coefficient from different histogram resolutions and reading scales by expressing all response variables as their equivalent value in a 256-channel histogram (eight-bit resolution) of a log-amplified distribution. Thus, a four-decade log amplifier would have an ideal response coefficient

An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 47 Number 26 NCCLS of 64 channels per decade. With the advent of high-resolution histograms that accrue data without logarithmic amplification, this approach is expected to become mostly of historic interest.

However it is expressed, the slope of the QFC curve from valid standards should be close to the nominal or ideal value and should show only small, random variation over time. Trends or sudden shifts suggest deterioration of electronic components or standards.

11.3.4 Fluorescence Measurement Boundaries (Window of Analysis)

When the instrument reading is expressed as a histogram channel number, the overall measurement range is specifically limited to signals that fall between the lowest (minimum) channel and the highest (maximum) channel value of the histogram. Extrapolating the calibration curve to the zero and maximum channels allows these boundaries to be expressed in MESF units, thus defining the quantitative window of analysis.50,55,56,59 These parameters are used only to indicate the position of the calibration curve in the overall space of the dose-response relationship, and they are not meant to imply true measurement values corresponding to the extreme channel numbers.

The windows of analysis for fluorescein and phycoerythrin commonly used on clinical flow cytometers have lower boundaries in the range of 10 to 100 MESF and upper boundaries in the range of 1 million to 10 million MESF (see Figure 8). Since these boundaries determine whether various cell phenotypes appear on scale and how they cluster, a consistent window of analysis should be a major component of quality control in clinical applications.

While the window of analysis applies particularly to histogram-bounded measurements, all reporting scales have a lower boundary set by the minimum possible reading on their response scale. The extrapolated fluorochrome dose corresponding to the lower boundary provides important information about the measurement range of all types of fluorometers. Since a valid FI calibration curve must be a straight line with approximately unity slope, this single point is sufficient to fix the position of the curve in dose- response space. The lower boundary can also be used for comparisons of measurement ranges over time or across different instruments and platforms.

Note that the lower boundary for histograms of log-amplified signals should be taken as zero, while the lower boundary for true linear data such as counts and linear histogram arrays should be taken as unity (log value equals zero). The lower boundary for data arithmetically converted to a linear scale from histograms of log-amplified signals depends on the formula used for the conversion and must be determined from the particular scale.

The lower boundary and other parameters of the calibration curve are important tools in optimizing the FI measurement. If the gain is not high enough, the lower boundary will be higher than the inherent sensitivity of the instrument. When maximum sensitivity is important to the application, gains should be set high enough to make full use of the lower measurement range. However, lower gains allow higher fluorochrome doses to fall in the linear calibration range, which may be preferable for certain applications. Once the optimal conditions for a given assay have been determined, they can be objectively characterized by the parameters of the MESF calibration curve and reliably reproduced in other instruments.

CD38 Plasma Cells CD45 CD4 CD4 Lymphocytes Unstained Monocytes Lymphocytes Lymphocytes

Blank MESFMicrosphere Standards Microsphere

Histogram Channel Number (8-bit, 4-decade log) 4-decade (8-bit, Number Channel Histogram

1000 191

100 127

10 64

. Relative Fluorescence Intensity Reading (1-10,000) 200 2000 20,000 200,000 2,000,000 Molecules of Equivalent Soluble Fluorochrome (MESF)

Figure 8. A Quantitative Window of Analysis Showing Typical Distributions of Normal Peripheral Blood Leukocytes Stained with FITC Conjugates (Upper Region) and MESF Microsphere Standards Labeled with FITC (Lower Region). (From Vogt RF, Schwartz A, Marti GE, Whitfield WE, Henderson O. Quantitative fluorescence cytometry. In: Faguet G, ed. Hematologic Malignancies: Methods and Techniques. Humana Press. 2000;12(55):255-273. Adapted with permission from Humana Press.) The minimum-channel value is about 200 MESF, and the maximum channel value is about 2 000 000 MESF. Lymphocyte autofluorescence causes unstained lymphocytes to occupy most of the first decade, and increased levels of specific staining are seen in the remaining populations. The brightest staining is shown by CD38-stained plasma cells, many of which accumulate in the maximum channel because their fluorescence intensity is brighter than two million MESF. The MESF calibration curve is depicted as the regression is performed, with the independent variable (assigned MESF value) on the X-axis and the dependent variable (instrument response expressed as a histogram channel number or a converted RFI value) on the Y-axis.

11.3.5 Detection Threshold

The FI reading of a truly blank analyte in a flow cytometer represents the total background in the fluorescence measurement. The position of this background relative to the MESF calibration curve has been called the fluorescence detection threshold.7,68 Like the measurement boundaries, the detection threshold provides a convenient reference frame to monitor and compare instruments and platforms. As an extrapolation, it should not be taken as an actual measurement.

As long as the background signal is on scale, the ability to detect signals near the background level depends more on the variance in the background measurement than on the position of the background on the scale. This influence of variance limits the general applicability of the detection threshold as a measure of the ability to resolve weak fluorescence from background. One approach to quantifying the background variance is to express the background in terms of the equivalent number of fluorochrome MESF that would produce the observed background variance.67,68 Expressing the background in this way allows a direct connection with the underlying physics of the optical measurement. Background may be measured directly68 from a blank or indirectly67 from the effect of background on dimly fluorescent particles. (See Appendix E for an approach for the indirect measurement of background.) A complete assessment of sensitivity also requires measurement of the detection efficiency (see Appendix E).

11.4 Laboratory Approaches Using QFC for Evaluating Instrument Performance and Measuring Expression of Cellular Receptors

11.4.1 Basic MESF Calibration with Binding Values Derived from the Effective F/P Ratio

In this system, the FI calibration curve is derived from microspheres labeled with varying amounts of environmentally-responsive fluorochrome spectrally matched with the fluorochrome-labeled conjugates used for staining measurands. The independent variable is the MESF value of each labeled population, assigned by (or traceable to) comparison with fluorochrome reference solutions.6,7,57

MESF calibration is commonly used to standardize FI measurements on different flow cytometers through the parameters of the calibration curve (see Section 11.3). If the effective F/P ratio of the fluorochrome- labeled conjugate (FLC) is known, MESF values may also be used to quantify FLC binding on stained measurands.

The FITC-MESF calibration system has been widely applied in a variety of studies within and between laboratories.26,54-56 Historically, values derived from one source of FITC-labeled MESF microsphere standards constitute the most extensive QFC database to date. The assigned FITC-MESF values from this source are reasonably precise (values assigned to successive lots have been relatively stable over many years), and independent methods confirm their trueness within about a two-fold range of uncertainty.3,4,14 Measurement uncertainties were quantified in assigning MESF values to NIST FITC-labeled microsphere reference material RM 8640 by comparison to NIST fluorescein reference solution SRM 1932 (see Appendix D). These paired reference materials provide the basis for uniform FITC-MESF results to be obtained from any measurement that lies within the constraints of the basic QFC model.

MESF microsphere standards have also been available for PE, but they have not been as well standardized as FITC, partly because of the inherent difficulties in standardizing complex protein macromolecules. As of the publication of this guideline, other approaches to standardizing FI measurements from PE conjugates may be preferred (see Sections 11.4.3 and 11.4.4). Authoritative MESF reference materials for PE and other fluorochromes may become available within the useful lifetime of this guideline.

MESF units are fluorochrome-specific, so MESF values for materials stained by different fluorochromes cannot be meaningfully compared. MESF values can be compared if materials are stained by different conjugates that are labeled with the same fluorochrome (presuming reasonable spectral matching).

However, this is a comparison only of fluorescence intensity (FI) and not of FLC binding, since the effective F/P ratios of different FLC vary depending on their molar F/P ratios and the microenvironments of their conjugated fluorochrome molecules. It should be noted that different methods of chemical conjugation may result in different fluorescence yields that influence the effective F/P.

When different measurands are stained with the same FLC solution, a comparison of MESF values will give an accurate value for the relative amounts of FLC bound by each measurand. However, for absolute binding determinations, the effective F/P of the FLC must be known so that MESF units can be converted into binding values:

FLCmolecules per particle = (MESFper particle) / (effective F/P)

The effective F/P, which accounts for differences in fluorescence yield, is commonly determined by one of two methods. The first method, using only a flow cytometer, measures the MESF of capture (conjugate- binding) microspheres stained to saturation with a particular FLC. To obtain the effective F/P, this MESF value is divided by a predetermined microsphere binding capacity:

effective F/P = (MESFper particle) / (FLCmolecules per particle)

Valid results from this method depend on 1) near-saturation of the stained microspheres; 2) absence of nonspecific binding 3) accurate values for the microsphere binding capacity and 4) complete capture of FLC without steric hindrance effects (see Section 11.4.2).

The second method, using only solution fluorometry, measures the MESF of an FLC solution with a standard curve constructed from fluorochrome reference solutions of known molarity. To obtain the effective F/P, this MESF value is divided by the molar concentration of the FLC solution:

effective F/P = (MESFEquiv Solution molarity) / (FLCmolarity)

Valid results from the cuvette fluorometry method depend on 1) accurate FLC molar concentration values; and 2) absence of residual unbound fluorochrome in the FLC solution.

Since the cuvette fluorometry method does not require an accurate predetermined value for conjugate binding capacity on microspheres, it is more direct and should produce a more precise and accurate determination of the effective F/P. Ideally, the effective F/P would be determined by this method when the FLC is first prepared, before addition of carrier proteins. It could then be provided to the end user as a reagent specification.

Neither the microsphere nor the solution method addresses potential complications caused by different binding constants among conjugate molecules with different molar F/P ratios. The presence of unconjugated ligand with stronger binding activity or of highly conjugated ligand with essentially no binding activity could cause significant bias in this approach to QFC calibration.

11.4.2 Binding Calibration with Capture Microsphere Standards

In this system, FI values are converted to molecules of bound FLC using capture microspheres with predetermined antibody binding capacity (ABC) values. In one approach, capture microspheres with varying predetermined ABC values are used directly as multilevel calibrators, and a standard curve of ABC vs RFI is constructed. The same methods that apply to constructing RFI and MESF calibration curves are used for ABC curves, but the evaluation criteria may need to be relaxed since the preparative staining step often results in a less robust linear fit. An alternative approach uses a single population of capture microspheres having a predetermined ABC value as a single-point calibrator that allows transformation of values extrapolated from an RFI or MESF calibration curve into ABC values.

Either of these approaches can give consistent results in idealized model systems. 6,121 However, in practical application, considerable variation is seen between reagents and preparative staining methods for both cells and microsphere standards. In two studies of CD4 FLC binding on T-cells in which capture microsphere standards were used with different conjugates,58,59 results differed over a six-fold range (30 000 to 180 000 FLC bound per cell). These and other studies show that it is impractical to presume universal binding capacity values for capture microspheres, since they will not be true for all FLC. However, if the binding capacity values are accurately assigned for a particular conjugate under standardized staining protocols, this method can be useful.

A standardized staining protocol is critical for consistent results. The following procedure has given optimal results in one system that has been widely used on a variety of FLC 56:

1) Add about 10 000 capture microbeads to FLC diluted in 250 µL PBS-azide. 2) Incubate for 72 hour at 4 °C. 3) Wash once with PBS-azide and aspirate immediately. 4) Resuspend in 250 µL PBS-azide and analyze immediately.

The prolonged incubation period is required to reach equilibrium binding, and the subsequent wash, resuspension, and analysis should be completed in the shortest possible time to avoid dissociation. Protein is not required in the matrix and may cause artifacts. All other properties (particularly pH) should be identical to those of the matrix in which the measurands are suspended.

To determine the optimal concentration of an FLC for near-saturation binding, capture microspheres should be stained with a series of equal-volume (two-fold) dilutions of FLC. The titration should begin at least two- fold above the recommended amount for routine staining and should continue down to where the lower plateau approaches the blank microsphere. The resulting titration curve, constructed as a log dilution vs. log RFI plot in the same manner as the QFC calibration curve (see Section 11.2), should show a background plateau at the lowest FLC concentrations, a linear increase through the intermediate dilutions, and a plateau at the highest concentrations. The optimal working concentration is generally represented by the first point on the high plateau, (i.e., the lowest concentration where near-saturation is achieved). Since some FLC tend to bind nonspecifically to microspheres, a blank microsphere should be included at each dilution. The fluorescence from the blank microsphere is usually not subtracted from the stained microspheres, but high levels of nonspecific binding preclude the use of this calibration method.

The most comprehensive approach to optimizing near-saturation binding is to titer pooled multilevel calibrators that bracket the range of binding values expected for the measurands. The optimal FLC concentration will produce near-saturation of the microsphere population with the highest binding capacity while maintaining the plateau endpoint for the lower binding capacity populations. This concentration should give a good linear fit (i.e., low residuals) when ABC values are regressed against their respective FI values.

11.4.3 RFI Calibration and Translation to Binding Calibration Using a Biologic Calibrator

In this system, which was originally developed for quantifying CD38 expression, RFI readings are converted to molecules of PE using a constant (the RFI multiplier) that represents PE molecules per RFI unit (see http://cyto.mednet.ucla.edu/protocols/cd38.pdf). 52 RFI microsphere calibrators (see Figure 6) are used as setup and qualifying standards. The conversion is based on a single-point FI determination of CD4 T- cells from a pool of normal donors stained with presumed unimolar PE conjugates. The FI of the CD4 T-cell is converted to a molecular value of FLC binding based on a presumed expression of 50 000 FLC binding sites per cell. The binding of a unimolar CD38-PE FLC is then quantified by applying the same RFI conversion factor.

The benefit of this approach is apparent from interlaboratory studies: although interlaboratory variation for the RFI multiplier is large (14 to 25%), the variation in calculated CD4 FLC binding is much lower (3 to 7%). This method makes the implicit assumption that the effective F/P of the CD4 unimolar PE-FLC is the same as the effective F/P of the CD38 unimolar PE-FLC, which was later confirmed independently (see Section 11.4.4). The presumed value of 50 000 binding sites for CD4 fluorochrome-antibody conjugates on CD4 T-cells has been repeatedly documented within experimental error.6,60,122 Further studies showed that the variation of CD4 expression over time is stable, and within-individual variation is 4 to 5%,123 making it a suitable biologic calibrator. The method could presumably be applied to any measurand stained with unimolar PE conjugates, or more generally to those stained with any FLC shown to have the same effective F/P as a CD4 reference FLC. It is now of mostly historical interest, since the use of the freshly-stained CD4 biologic calibrator has been largely superseded by the availability of PE microsphere standards that give essentially equivalent results (see Section 11.4.4).

Other biological materials, including some stabilized cellular materials used for staining controls, may offer certain advantages as QFC calibrators. Their binding properties, based on recognition of actual cellular receptors, most closely approach those of cellular measurands. Their nonspecific binding and environmental influences on FI may also be more nearly matched to those of stained cells.

11.4.4 PE-Labeled Microsphere Calibrators and Unimolar PE-Ligand Conjugates

This system bypasses the need for MESF determinations by using unimolar PE-FLC and assuming that the PE molecules on the microsphere standards have the same fluorescence yield as PE molecules on FLC bound to stained cells. In one carefully evaluated implementation of this approach, the assigned values of PE molecules per microsphere were established by three independent methods: fluorometric, enzymatic, and parametric binding.60 This approach was validated for measuring CD38 expression by comparing it with the CD4 biologic calibration method.61 Results for CD4 ABC values were equivalent on two models of flow cytometers from different manufacturers. In a multisite study,114 the CD38 PE-FLC binding on CD8 T-cells measured using PE microsphere calibrators was comparable to that measured by the CD4 biologic calibrator method (see Section 11.4.3). Although statistically significant differences were seen, they were small compared to the range of biologic variability, and in two out of four sites, the results were nearly identical.

The assumption of equivalent fluorescence yield may not be universally true for PE-labeled microspheres and conjugates even if they are unimolar. The fluorescent yields of unimolar FLC obtained from different sources may differ significantly, and determination of the effective F/P is suggested for every manufactured FLC lot. Verification should be performed using an appropriate FLC-specific biologic calibrator of known FLC binding capacity. The end user can check the validity of the calibration for at least one FLC-receptor system by staining T-cells with unimolar CD4 PE-FLC: the interpolated value of the CD4-positive cells should be about 50 000 PE molecules.

11.4.5 Indirect Immunofluorescence and Biologic Calibrators

In this system, a radiobinding assay for cell-surface CD5 antigen was used to measure expression on T-cell cultures, and CD5 FLC were used to stain the same cell lines using indirect immunofluorescence.12 Sub- cultures were used to obtain T-cell lines with differing CD5 expression. When these cell lines were allowed to bind the CD5 conjugates under saturating conditions, a direct correlation between radiobinding and FI was established. Reference values for FLC binding per cell were obtained from the radiobinding results and used to calibrate the FI measurements. The use of a common second antibody as the FLC allows the same FI calibration to be applied to different primary antibodies recognizing different cell receptors. Assays based on this approach have been developed for several clinical analytes.115

The binding values obtained with this indirect approach appear to be consistent with consensus values obtained by direct immunofluorescence using the methods described above. Since the FLC is a second antibody, the method in principal can be applied without recalibration to any marker detected by a primary

An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 53 Number 26 NCCLS antibody to which the second antibody binds or is bound. Although staining by indirect immunofluorescence is not commonly used for multiparameter immunophenotyping of cells, this method is widely used in suspension array technology.

11.4.6 Microsphere Calibration with Software and Biologic Value Assignment

This system has features of both the MESF and biologic calibration systems above. One implementation for this approach is for the quantitation of CD64 on neutrophils, which is an indicator of a systemic acute inflammatory response to infection, sepsis, or tissue injury.124-126

Like the MESF system, this FI calibration for CD64 is derived from microspheres labeled with a spectrally matched environmentally responsive fluorochrome and assigned FI values by comparison with fluorochrome reference solutions. These reference microspheres are fully compatible with NIST FITC- labeled microsphere standards (RM 8640) having MESF values assigned by comparison to NIST fluorescein solution standard (SRM 1932). Therefore, values from the Leuko64 calibration system can be converted to authoritative MESF values, and accurate CD64 binding values can be derived if the effective F/P ratio of the CD64 FITC conjugate is known.

The reference microspheres employed in this system are additionally labeled with a second fluorochrome (starfire red), which allows identification of the beads by the software supplied with the Leuko64 kit using cluster-finding algorithms. The FITC signal of the beads is used to calibrate the FITC scale within the software that carries a lot-specific value that scales the FITC signal in terms of a CD64 index value. Through the use of a lot-specific FI value assigned to the reference microsphere and bundled in the accompanying software, this system compensates for lot-to-lot differences in microsphere FI as well as variations in the monoclonal antibody effective F/P ratio due to differences in conjugation efficiencies.

The CD64 index value assignments to the beads in each lot are derived through a biologic calibration procedure based on the measured FI of CD64-FITC conjugate binding on multiple blood samples with varying levels of neutrophil and monocyte CD64 expression. A regression analysis between different lots of the Leuko64 assay on these different biologic samples allows for the value assignment of the CD64 index to the bead, thereby allowing for a high degree of lot to lot correlation. The software within this system uses cluster-finding algorithm methods to identify beads, lymphocytes, monocytes, and granulocytes, which improves reproducibility (CV< 5%) between laboratories and instruments.

The other function of the calibration beads in this assay is to direct the instrument setup of all light scatter and fluorescence signals of the flow cytometer, thus ensuring the appropriate window of analysis for all parameters is established. The assay instructions direct the user to adjust the instrument gains to place the beads in specific positions on the fluorescence and scatter signal scales. Additionally, the assay system provides a cell lysis buffer with a controlled pH of 7.4, thus controlling an important variable of FITC fluorescence, into which the calibration beads and blood sample are suspended prior to flow cytometric analysis. This system thereby provides an integrated link between instrument setup, fluorescence calibration, and cellular analysis through a coordination of calibration beads, monoclonal reagents, cell lysis buffer, and data analysis through assay-specific software.

L- 007 Bead A L- 004 Bead B L- 003 Bead C L- 007 Bead A L- 004 Bead B L- 007 Bead C L- 003 Bead B L03- 004 Bead A Index L– 007 Bead C L- 003 Bead B 10.00 3.50 9.00 x 3.00 8.00

2.50 7.00 6.00 2.00 5.00 1.50 4.00 PMN CD64 Index

PMN CD 64 Index 3.00 1.00 2.00 0.50 1.00 0.00 0.00 0. 00 2.00 4. 00 6.00 8. 00 10.00 0. 00 1. 00 2. 00 3. 00 4. 00 5.00 PMN CD64 Index (lot 03 -003 Bead A ) PMN CD64 Index (lot 03- 003 bead A)

Lot Comparison Regression r2 Lot Comparison Regression r2 03-003 A vs 03-003 B y = 0. 7355x + 0.0212 0.9996 03-003 A vs 03-007 A y = x + 6E-05 0.9956 03-003 A vs 03-007 A y = 0. 6867x + 0.0568 0. 9956 03-003 A vs 03-003 B y = x - 4E-06 0.9996 03-003 A vs 03-004 B y = 0. 5159x + 0.0841 0. 9984 03-003 A vs 03-004 B y = x -3E-05 0.9984 03-003 A vs 03-003 C y = 0. 3837x + 0.0325 0. 9958 03-003 A vs 03-003 C y = x + 0.0001 0.9958 03-003 A vs 03-007 C y = 0. 2754x + 0.0438 0. 9884 03-003 A vs 03-007 C y = x - 0.0002 0.9884

Figure 9. Measurement of CD64 Expression by Microsphere Calibration with Software and Biologic Value Assessment. (Contributed by Dr. Bruce Davis, Trillium Diagnostics, LLC, and the Maine Medical Center Research Assessment.) Leuko64 assay utilizes software and calibration beads to adjust for lot-to-lot variation in antibody fluorochrome conjugation efficiencies (effective F/P ratio) and bead fluorescence intensities. The graph and tabular data on the left show the relationship of PMN CD64 Index measurements on 15 clinical blood specimens (left side) with varying expression of PMN CD64 using six different lots of Leuko64 reagent (lots 03-003 to 03-007) and bead (lots A, B, or C) combination. There is a high degree of correlation (r2) in results among the different lots, but as shown in the regression analysis, there is variable bias between the six lots. The interlot bias is nearly totally removed by assigning a lot-specific FITC intensity value or PMN CD64 index value to the beads, as shown (right side) in the graphic and tabular summary of the same 15 specimens following reanalysis with the lot-specific Leuko64 software version.

References

1 Coons AH, Creech HJ, Jones RN. Immunological properties of an antibody containing a fluorescent group. Proc Soc Exptl Biol Med. 1941;47:200.

2 Coons AH. The beginnings of immunofluorescence. J Immunol. 1961;87:499-503.

3 Brown MC, Hoffman RA, Kirchanski SJ. Controls for flow cytometers in hematology and cellular immunology. Ann NY Acad Sci. 1986;468:93-103.

4 Vogt RF Jr., Cross GD, Henderson LO, Phillips DL. Model system evaluating fluorescein-labeled microbeads as internal standards to calibrate fluorescence intensity on flow cytometers. Cytometry. 1989;10:294-302.

5 Hoffman RA, Recktenwald DJ, Vogt R. Cell-associated receptor quantitation. In: Bauer KD, Duque RE, Shankey TV, eds. Clinical Flow Cytometry: Principles and Application. Baltimore, MD: Williams & Wilkins; 1993:469-477.

6 Vogt RF, Marti GE, Schwartz A. Quantitative calibration of fluorochrome intensity for clinical and research applications of leukocyte immunophenotyping by flow cytometry. In: Tyrer HW, ed. Critical Issues in Biotechnology and Bioengineering, Vol. 1. Norwood, NJ: Ablex Publishing Co.; 1995.

7 Schwartz A, Repollet EF, Vogt RF, Gratama JW. Standardizing flow cytometry: construction of a standardized fluorescence calibration plot using matching spectral calibrators. Cytometry. 1996;26:22-31.

8 Gaigalas AK, Li L, Henderson O, et al. The development of fluorescence intensity standards. J Res Natl Inst Stand Technol. 2001;106:381-389.

9 Henderson LO, Marti GE, Gaigalas A, Hannon WH, Vogt RF. Terminology and nomenclature for standardization in quantitative fluorescence cytometry. Cytometry. 1998;33(2):97-105.

10 Cossman J, Neckers LM, Hsu SM, Londo D, Jaffee ES. Low-grade lymphomas. Expression of developmentally regulated B-cell antigens. Am J Pathol. 1984;115:117.

11 Oonishi T, Uyesaka N. A new standard fluorescence microsphere for quantitative flow cytometry. J Immunol Meth. 1985;84:143-154.

12 Poncelet P, Carayon P. Cytofluorometric quantification of cell-surface antigens by indirect immunofluorescence using monoclonal antibodies. J Immunol Meth. 1985;85:65-74.

13 Spack E, Edidin M. The class I MHC antigens of erythrocytes: a serologic and biochemical study. J Immunol. 1986;136(8):2943-2952.

14 Chatelier RC, Ashcroft RG, Lloyd CJ, et al. Binding of fluoresceinated epidermal growth factor to A431 cell sub-populations studied using a model-independent analysis of flow cytometric fluorescence data. EMBO J. 1986;5(6):1181-1186.

15 Oonishi T, Sakashita K, Uyesaka N. Flow cytometric studies of the binding of monoclonal antibodies OKT3, OKT4 and OKT8. J Immunol Meth. 1988;115:159-167.

16 Caldwell CW, Patterson WP. Fluorescence intensity of immunostained cells as a diagnostic aid in lymphoid leukemias. Diagnostic & Clinical Immunology. 1988;5(6):371-376.

17 Marti GE, Zenger V, Caproaso NE, et al. Antigenic expression of B-cell chronic lymphocytic leukemia lymphocytes. Analytical and Quantitative Cytology and Histology. 1989;11(5):315-323.

18 Guyre PM, Granziano RF, Vance BA, Morganelli PM, Fanger MW. Monoclonal antibodies that bind to distinct epitopes on fc-gamma- ri are able to trigger receptor function. J Immunol. 1989;143:1650-1655.

19 Caldwell CW, Maggi J, Henry LB, Taylor HM. Fluorescence intensity as a quality control parameter in clinical flow cytometry. Am J Clin Path. 1987;88:447-456.

20 Muirhead KA, Schmitt TC, Muirhead AR. Determination of linear fluorescence intensities from flow cytometric data accumulated with logarithmic amplifiers. Cytometry. 1983;3(4):251-256.

21 Bohmer RM, Papaioannou J, Ashcroft RG. Flow-cytometric determination of fluorescence ratios between differently stained particles is dependent on excitation intensity. J Histochem Cytochem. 1985;33:(9)974-976.

22 Dawson CD, Scheppler JA, Nicholson JKA, Holman RC. Enumeration of antigen sites on cells by flow cytometry. Pathobiology. 1991; 59:57-61.

23 Fay SP, Posner RG, Swann WN, Sklar LA. Real-time analysis of the assembly of ligand, receptor, and G protein by quantitative fluorescence flow cytometry. Biochemistry. 1991;30(20):5066-5075.

24 Givan AL, Calvert JE, Shenton BK. The effect of improvements in cytometer sensitivity on the detection of cd5-positive B cells with dim fluorescence. Cytometry. 1991;12:360-365.

25 Caldwell CW, Patterson WP. Relationship between CD45 antigen expression and putative stages of differentiation in B-cell malignancies. Am J Hematol. 1991;36(2):111-115.

26 Vogt RF, Cross GD, Phillips DL, Henderson LO. Inter-laboratory study of cellular fluorescence intensity measurements with fluorescein- labeled microbead standards. Cytometry. 1991;12:525-536.

27 Marti GE, Faguit G, Bertin P, et al. CD20 and CD5 expression in B-chronic lymphocytic leukemia. Ann NY Acad Sci. 1992;651:480-483.

28 Labus JM, Peterson BH. Quantitation of human anti-mouse antibody in serum by flow cytometry. Cytometry. 1992;13:275-281.

29 Marti GE, Zenger V, Brown M, et al. Antigenic expression of B-cell chronic lymphocytic leukemic cell lines. Leukemia & Lymphoma. 1992;7:497-504.

30 Almasri NM, Duque RE, Iturraspe J, Everett E, Braylan RC. Reduced expression of CD20 antigen as a characteristic marker for chronic lymphocytic leukemia. Am J Hematol. 1992;40:259-263.

31 Carter PH, Resto-Ruiz S, Washington GC, et al. A flow cytometric analysis of whole blood lysis, three anticoagulants and five cell preparations. Cytometry. 1992;13:68-74.

32 Creamer P, Edmonds J, Sullivan J, Matthews S. Measurement of HLA class I expression in ankylosing spondylitis. Ann Rheum Dis. 1992;51:1138-1142.

33 Schwartz A, Fernandez-Repollet E. Development of clinical standards for flow cytometry. Ann NY Acad Sci. 1993;677:28-39.

34 Vogt RF, Henderson LO, Ethridge SF, Huang EY, White JT, Meredith NK. Lymphocyte immunophenotyping with extended quantitative analysis of list mode files for epidemiologic health studies. Ann NY Acad Sci. 1993;677:462-464.

35 Caldwell CW, Patterson WP, Hakami N. CD45 expression and prognosis in acute lymphoblastic leukemia. Blood. 1993;81(2):562-563.

36 Lavabre-Bertrand T, Exbrayat C, Bourquard P, et al. CD23 antigen density is related to serum gamma globulin level, bone marrow reticulin pattern, and treatment in B chronic lymphocytic leukemia. Leukemia and Lymphoma. 1993;1-6.

37 Lavabre-Bertrand T, Janossy G, Exgrayat C, Bourquard P, Duperray C, Navarro M. Leukemia-associated changes identified by quantitative flow-cytometry. Leukemia. 1994;8:1557-1563.

38 Coder DM, Redelman D, Vogt RF. Computing the central location of immunofluorescence distributions: Logarithmic data transformations are not always appropriate. Cytometry. 1994;18(2):75-78.

39 Marti GE, Vogt R, Zenger V, et al. The binding capacity of CD20 reagents (A quantitative flow cytometry protocol for Fifth International B Cell Workshop on Normal B-PBL and B-CLL Lymhocytes: Results for CD20 on Normal PBL) In: Schlossman S, Boumsell L, Gilks W, et al., eds. Leukocyte Typing V. Oxford University Press. 1994;11:515.

40 Lenkei R, Andersson B. Determination of the antibody binding capacity of lymphocyte membrane antigens by flow cytometry in 58 blood donors. J Immunol Methods. 1995;183:267.

41 Lavabre-Bertrand T, Poncelet P. Antigen density in b-cell chronic lymphocytic leukemia. Am J Clin Path. 1995;103(3):375.

42 Cabezudo E, Matutes E, Ramrattan M, Morilla R, Catovsky D. Analysis of residual disease in chronic lymphocytic leukemia by flow cytometry. Leukemia. 1997;11(11):1909-1914.

43 Stelzer GT, Marti GE, Hurley A, McCoy P, Lovett EL, Schwartz A. U.S. Canadian consensus recommendations on the immunophenotypic analysis of hematologic neoplasia by flow cytometry: Standardization and validation of laboratory procedures. Cytometry. 1997;30:214- 230.

44 Ginaldi L, De Martinis M, Matutes E, Farahat N, Morilla R, Catovsky D. Levels of expression of CD19 and CD20 in chronic B cell leukaemias. J Clin Pathol. 1998;51(5):364-369.

45 D’Arena G, Musto P, Casavilla N, et al. Quantitative flow cytometry for the differential diagnosis of leukemic B-cell chronic lymphoproliferative disorders. Am J Heme. 2000;64:275-281.

46 Vogt RF, Schwartz A, Marti GE, Whitfield WE, Henderson O. Quantitative fluorescence cytometry. In: Faguet G, ed. Hematologic Malignacies: Methods and Techniques. Humana Press. 2000;12(55):255-273.

47 Hendrickx A, Bossuyt X. Quantification of the leukocyte common antigen (CD45) in mature B-cell malignancies. Cytometry. 2001;46(6):336-339.

48 Lenkei R, Mandy F, Marti G, Vogt R. Preface: quantitative flow cytometry: an emerging consensus. Cytometry. 1998;33(2):93-96.

49 Lenkei R, Marti G, Vogt R, Nicholson J, Mandy F. Proceedings from three international conferences on quantitative fluorescence cytometry (QFCM). Cytometry. 1998;33(2):93-96.

50 Schwartz A, Marti GE, Poon R, Gratama JW, Fernandez-Repollet E. Standardizing flow cytometry: A classification system of fluorescence standards used for flow cytometry. Cytometry. 1998;33(2):106-114.

51 Lenkei R, Bratt G, Holmberg V, Muirhead K, Sandstrom E. Indicators of T-cell activation: Correlation between quantitative CD38 expression and soluble CD8 levels in asymptomatic HIV+ individuals and healthy controls. Cytometry. 1998;33(2):115-122.

52 Hultin LE, Matud JL, Giorgi JV. Quantitation of CD38 activation antigen expression on CD8+ T cells in HIV-1 infection using CD4 expression on CD4+ T lymphocytes as a biological calibrator. Cytometry.1998;33(2):123-132.

53 Perfetto SP, Malone JD, Hawkes C, et al. CD38 expression on cryopreserved CD8+ T cells predicts HIV disease progression. Cytometry. 1998;33(2):133-137.

54 Zenger VE, Vogt R, Mandy F, Schwartz A, Marti GE. Quantitative flow cytometry: Inter-laboratory variation. Cytometry. 1998;33(2):138- 145.

55 Bergeron M, Faucher S, Minkus T, et al. Impact of unified procedures as implemented in the Canadian Quality Assurance Program for T- lymphocyte subset enumeration. Participating Flow Cytometry Laboratories of the Canadian Clinical Trials Network for HIV/AIDS Therapies. Cytometry. 1998;33(2):146-155.

56 Purvis N, Stelzer G. Multi-platform, multi-site instrumentation and reagent standardization. Cytometry. 1998;33(2):156-165.

57 Gratama JW, D’hautcourt JL, Mandy F, et al. Flow cytometric quantitation of immunofluorescence intensity: Problems and perspectives. European Working Group on Clinical Cell Analysis. Cytometry. 1998;33(2):166-178.

58 Serke S, van Lessen A, Huhn D. Quantitative fluorescence flow cytometry: A comparison of the three techniques for direct and indirect immunofluorescence. Cytometry. 1998;33(2):179-187.

59 Lenkei R, Gratama JW, Rothe G, et al. Performance of calibration standards for antigen quantitation with flow cytometry. Cytometry. 1998;33(2):188-196.

60 Davis KA, Abrams B, Iyer SB, Hoffman RA, Bishop JE. Determination of CD4 antigen density on cells: Role of antibody valency, avidity, clones, and conjugation. Cytometry. 1998;33(2):197-205.

61 Iyer SB, Hultin LE, Zawadzki JA, Davis KA, Giorgi JV. Quantitation of CD38 expression using QuantiBRITE beads. Cytometry. 1998;33(2):206-212.

62 Waxdal MJ, Monical MC, Palini AG. Inter-laboratory relative fluorescence intensity measurements using FlowCal 575 calibration beads: A baseline study. Cytometry. 1998;33(2):213-218.

63 Powell MK, Whitfield W, Redelman D, Henderson LO, Vogt RF Jr. Titration of a CD45-FITC conjugate to determine the linearity and dynamic range of fluorescence intensity measurements on lymphocytes. Cytometry. 1998;33(2):219-224.

64 Zhang YZ, Kemper C, Bakke A, Haugland RP. Novel flow cytometry compensation standards: Internally stained fluorescent microspheres with matched emission spectra and long-term stability. Cytometry. 1998;33(2):244-248.

65 Collins DP, Luebering BJ, Shaut DM. T-lymphocyte functionality assessed by analysis of cytokine receptor expression, intracellular cytokine expression, and femtomolar detection of cytokine secretion by quantitative flow cytometry. Cytometry. 1998;33(2):249-255.

66 Wood JC, Hoffman RA. Evaluating fluorescence sensitivity on flow cytometers: An overview. Cytometry. 1998;33(2):256-259.

67 Wood JC. Fundamental flow cytometer properties governing sensitivity and resolution. Cytometry. 1998;33(2):260-266.

68 Chase ES, Hoffman RA. Resolution of dimly fluorescent particles: A practical measure of fluorescence sensitivity. Cytometry. 1998;33(2):267-279.

69 Shapiro HM, Perlmutter NG, Stein PG. A flow cytometer designed for fluorescence calibration. Cytometry. 1998;33(2):280-287.

70 Schwartz A, Wang L, Early E, et al. Quantitating fluorescence intensity from fluorophore: The definition of MESF assignment. J Res Natl Inst Stand Technol. 2002;107(1):83-91.

71 Wang L, Gaigalas AK, Abbassi F, Marti GE, Vogt RF, Schwartz A. Quantitating fluorescence intensity from fluorophors: Practical use of MESF values. J Res Natl Inst Stand Technol. 2002;107:339-353.

72 Gailgalas AK, Wang L, Schwartz A, Marti GE, Vogt RF. Quantitating fluorescence intensity from fluorophore: Assignment of MESF values. J Res Natl Inst Stand Technol. in press.

73 Hougaard DM, Larsson LI. Towards microfluorometric quantitation of polyamines in situ. Relationship between cellular polyamine concentration and fluorescence yield of the formaldehyde fluorescamine method. Histochemistry. 1990;93(4):359-362.

74 Marti GE, Vogt RF, Stetler-Stevenson M. Clinical quantitative flow cytometry: Identifying the optimal methods for clinical flow cytometry. Cytometry. 2003;55B:59.

75 21 CFR §58.63. Good Laboratory Practice for Nonclinical Laboratory Studies—Maintenance and Calibration of Equipment; 2002.

76 Swanson BN. Delivery of high-quality biomarker assays. Dis Markers. 2002;18(2):47-56.

77 Baak JP. The framework of pathology: Good laboratory practice by quantitative and molecular methods. J Pathol. 2002;198(3):277-283.

78 Brunetti MM. Critical aspects in the application of the principles of good laboratory practice (GLP). Ann Ist Super Sanita. 2002;38(1):41- 45.

79 Berend K. Good laboratory practice in the European Community. Role of the commission and the member states: External aspects. Ann Ist Super Sanita. 2002;38(1):9-13.

80 Centers for Disease Control and Prevention (CDC) (2) Centers for Medicare & Medicaid Services (CMS), HHS. Medicare, Medicaid, and CLIA programs; laboratory requirements relating to quality systems and certain personnel qualifications. Final rule. Fed Regist. 2003 Jan 24;68(16):3639-714.

81 42 CFR §493.17. Laboratory Requirements, Subpart A. General Provisions. Test Categorization.

82 Lum G, Tholen DW, Floering DA. The usefulness of calibration verification and linearity surveys in predicting acceptable performance in graded proficiency tests. Arch Pathol Lab Med. 1995;119(5):401-408.

83 CDC. 1997 revised guidelines for performing CD4+ T-cell determinations in persons infected with human immunodeficiency virus (HIV). MMWR. Recommendations & Reports. 1997 Jan 10; 46(RR-2):1-29.

84 CDC. Guidelines for Performing Single-Platform CD4+ T-cell Determinations with CD45 Gating for Persons Infected with Human Immunodeficiency Virus. MMWR. Recommendations & Reports. 2003 Jan 31; 52(RR-2):1-11.

85 ISO. General requirements for the competence of testing and calibration laboratories. ISO 17025. Geneva: International Organization for Standardization; 1999.

86 ISO. International Vocabulary of Basic and General Terms in Metrology. Geneva: International Organization for Standardization; 1993.

87 ISO. Statistics - Vocabulary and Symbols – Part 1: Probability and General Statistical Terms. ISO 3534-1. Geneva: International Organization for Standardization; 1993.

88 ISO. Laboratory medicine—requirements for reference measurement laboratories. ISO/FDIS 15195. Geneva: International Organization for Standardization; 2003.

89 Shapiro HM. Practical Flow Cytometry. 4th ed. New York, NY: Wiley-Liss, Inc.; 2003.

90 Asbury CL, Uy JL, van den Engh G. Polarization of scatter and fluorescence signals in flow cytometry. Cytometry 2000; 40:88-101.

91 Kenworthy AK, Edidin M. Imaging fluorescence resonance energy transfer as probe of membrane organization and molecular associations of GPI – anchored proteins. Methods Mol Biol.1999;116:37-49.

92 Damjanovich, Matko J, Matyus L, et al. Supramolecular receptor structures in the plasma membrane of lymphocytes revealed by flow cytometric energy transfer, scanning force- and transmission electron-microscopic analyses. Cytometry. 1998;33(2):225-233.

93 Gaigalas AK, Wang L, Vogt RF. Frequency-domain measurement of the photodegradation process of fluorescein. Photochem Photobiol. 2002;76:22-28.

94 Parks DR, Herzenberg LA. Fluorescence-activated cell sorting: theory, experimental optimization, and applications in lymphoid cell biology. Methods Enzymol. 1984;108:197-241.

95 Seamer LC, Bagwell CB, Barden L, et al. Proposed new data file standard for flow cytometry, version FCS 3.0. Cytometry. 1997;28:118- 122.

96 Kroll MH, Emancipator K, Floering D, Tholen D. An algorithm for finding the linear region in a norlinear data set. Comput Biol Med. 1999;29(5):289-301.

97 Muirhead KA. Quality control for clinical flow cytometry. In: Bauer K, Duque R, Shankey TV, eds. Flow Cytometry: Principles and Clinical Applications. Baltimore, MD: Williams & Wilkins; 1992:177-199.

98 Coons AH, Kaplan MH. Localization of antigen in tissue cells. II. Improvements in a method for the detection of antigen by means of fluorescent antibody. J Exp Med. 1950;91:1.

99 Klonis N, Sawyer WH. Spectral properties of the prototropic forms of fluorescein in aqueous solutions. J Fluoresc. 1996;6(3):147-157.

100 Smith SA, Pretorius WA. Spectrophotometric determination of pKa values for fluorescein using activity coefficient corrections. Water SA. 2002;28(4):395-402.

101 Barr JR, Maggio VL, Patterson DG Jr, et al. Isotopye dilution-mass spectrometric quantification of specific proteins model application with apolipoprotein A-I. Clin Chem. 1996;42(10):1676-1682.

102 Calvin J, Burling K, Blow C, Barnes I, Price CP. Evaluation of fluorescence excitation transfer immunoassay for measurement of specific proteins. J Immunol Methods. 1986;86(2):249-256.

103 Hoyland J. Fluorescent probes in practice – potential artifacts. In: Mason WT, ed. Fluorescent and Luminescent Probes for Biological Activity. London: Academic Press; 1993:223-228.

104 Wilk KE, Harrop SJ, Jankova L, et al. Evolution of a light-harvesting protein by addition of new subunits and rearrangement of conserved elements: crystal structure of a cryptophyte phycoerythrin at 1.63-Å resolution. Proc National Acad Sci. 1999;96:8901-8906.

105 Waggoner A, DeBiasio R, Conrad P, et al. Multiple spectral parameter imaging. Methods Cell Biol. 1989;30:449-478.

106 Fisher HF, McGregor LL. Biochim Biophys Acta. 1960;38:562.

107 Tsien RY. The green fluorescent protein. Ann Rev Biochm. 1998;67:509-544.

108 Penna TC, Ishii M. Selective permeation and organic extraction of recombinant green fluorescent protein from Escherichia coli. BMC Biotechnology. 2002;2:7-18.

109 Stewart CC, Stewart SJ. Four color compensation. Cytometry. 1999;38(4):161-175. 110 Dubertret B, Skourides P, Norris DJ, Noireaux V, Brivanlou AH, Libchaber A. In vivo imaging of quantum dots encapsulated inphospholipid micelles. Science. 2002;298:1759-1762. 111 Bruchez M Jr, Moronne M, Gin P, Weiss S, Alivisatos AP. Semiconductor nanocrystals as fluorescent biological labels. Science. 1998; 281:2013-2016.

112 Demas JN, Crosby GA. The measurement of photoluminescence quantum yield. A Review. J Phys Chem. 1971;75(8):991-1024.

113 Lakowicz JR. Principles of Fluorescence Spectroscopy. 2nd ed. New York, NY: Kluwer Academic/Plenum Publishers; 1999:10.

114 Schmitz JL, Czerniewski MA, Edinger M, et al. Multisite comparison of methods for the quantitation of the surface expression of CD38 on CD8+ T lymphocytes. Cytometry. 2000;42:174-179.

115 Poncelet P, Besson-Faure I, Lavabre-Bertrand T. Clinical Applications of Quantitative Immunophenotyping in Immunophenotyping. Stewart CC, Nicholson JKA, eds. New York, NY: Wiley-Liss, Inc.; 2000.

116 Robins RA, Laxton RR, Garnett M, Price MR, Baldwin RW. Measurement of tumour reactive antibody and antibody conjugate by competition, quantitated by flow cytofluorimetry. J Immunol Methods. 1986;90(2):165-172.

117 Nicholson KA, Rao PE, Calvelli T, Stetler-Stevenson M, Browning SW, Marti G. Artifactual staining of monoclonal antibodies in 2-color combinations is due to an immunoglobulin in the serum and plasma. Communications in Clinical Cytometry. 1994;18:140-146.

118 ISO. Linear calibration using reference materials. ISO 11095. Geneva: International Organization for Standardization; 1996.

119 Eberhardt KR, Mee RW. A comparison of uncertainty criteria for calibration. Technometrics. 1996;38(3):221-229.

120 Wakkers PJ, Hellendoorn B, Op de Weegh GJ, Heerspink W. Applications of statistics in clinical chemistry. A critical evaluation of regression lines. Clin Chim Acta. 1975;64(2):173-184.

121 Marti GE, Zenger VE, Vogt R, Gaigalas A. Quantitative flow cytometry: History, practice, theory, consensus, inter-laboratory variation and present status. Cytotherapy. 2002;4(1):97-98.

122 Poncelet P, Poinas G, Corbeau P, et al. Surface CD4 density remains constant on lymphocytes of HIV-infected patients in the progression of disease. Res Immunol. 1991;142(4):291-298.

123 Pannu KK, Joe ET, Iyer SB. Performance evaluation of Quantibrite phycoerythrin beads. Cytometry. 2000;45:250-258.

124 Davis BH, Bigelow NC, Curnutte JT, Ornvold K: Neutrophil high affinity Fc receptor (CD 64) is a marker of in vivo activation and interferon- gamma therapy. Laboratory Hematology. 1995;1:3-12.

125 Davis BH, Wallace P, Guyre PM: Leukocyte Fc-g receptors in health and disease. In: Robinson JP, Babcock GF, eds. Phagocyte Function: A Guide for Research and Clinical Evaluation. Wiley-Liss, Inc., New York. 1998:47-76.

126 Davis BH. Diagnosis and monitoring of infection and sepsis through quantitative neutrophil CD64 analysis. Cytometry. 2002;Suppl. 11:49-50.

Fluorescence is the result of a three-stage process that occurs in certain molecules (generally polyaromatic hydrocarbons or heterocycles and in some naturally occurring photosynthetic accessory proteins) called fluorophores, fluorochromes, or fluorescent dyes. A fluorescent probe is a fluorochrome designed to localize within a specific region of a biologic specimen or to respond to a specific stimulus. Most fluorescent molecules used for labeling and staining in biologic applications contain unsaturated carbon bonds. Their electronic structure contains π-orbitals, which are responsible for the fluorescence process.

The three stages of the fluorescence process are illustrated by the following simplified electronic-state diagram:

Figure A1. The Three Stages of the Fluorescence Process

Stage 1 – Excitation:

A photon of energy hνEX is supplied by an external source, such as an incandescent lamp or a laser, and absorbed by the fluorophore, creating an excited electronic singlet state (S1'). This process distinguishes fluorescence from chemiluminescence, in which the excited state is populated by a chemical reaction.

Stage 2 – Excited-State Lifetime:

The excited state exists for a finite time (typically one to ten nanoseconds). During this time, the fluorophore undergoes conformational changes and is also subject to a multitude of possible interactions with its molecular environment. These processes have two important consequences. First, the energy of S1' is partially dissipated, yielding a relaxed singlet excited state (S1) from which fluorescence emission originates. Second, not all the molecules initially excited by absorption (Stage 1) return to the ground state (S0) by fluorescence emission. Other processes such as collisional quenching, fluorescence resonance energy transfer (FRET), and intersystem crossing may also depopulate S1. The fluorescence quantum yield, which is the ratio of the number of fluorescence photons emitted (Stage 3) to the number of photons absorbed (Stage 1), is a measure of the relative extent to which these processes occur.

Stage 3 – Fluorescence Emission:

A photon of energy hνEM is emitted, returning the fluorophore to its ground state S0. Due to energy dissipation during the excited-state lifetime, the energy of this photon is lower, and therefore, of longer wavelength, than the excitation photon hνEX. The difference in energy or wavelength represented by (hνEX – hνEM) is called the Stokes shift. The Stokes shift is

Appendix A. (Continued)

fundamental to the sensitivity of fluorescence techniques because it allows emission photons to be detected against a low background, isolated from excitation photons. In contrast, absorption spectrophotometry requires measurement of transmitted light relative to high incident light levels at the same wavelength.

The entire fluorescence process is cyclical. Unless the fluorophore is irreversibly destroyed in the excited state (an important phenomenon known as photodegradation), the same fluorophore can be repeatedly excited and detected. The fact that a single fluorophore can generate many hundreds or thousands of detectable photons is fundamental to the high sensitivity of fluorescence detection techniques. For polyatomic molecules in solution, the discrete electronic transitions represented by hEX and hEM in Figure A1 are replaced by rather broad energy spectra called the fluorescence excitation spectrum and fluorescence emission spectrum, respectively. The bandwidths of these spectra are parameters of particular importance for applications in which two or more different fluorophores are simultaneously detected. With few exceptions, the fluorescence excitation spectrum of a single fluorophore species in dilute solution is identical to its absorption spectrum. Under the same conditions, the fluorescence emission spectrum is independent of the excitation wavelength, due to the partial dissipation of excitation energy during the excited-state lifetime, as illustrated in Figure A2. The emission intensity is proportional to the amplitude of the fluorescence excitation spectrum at the excitation wavelength. This is depicted in the following figure, which shows that excitation of a fluorophore at three different wavelengths (EX 1, EX 2, EX 3) does not change the emission profile but does produce variations in fluorescence emission intensity (EM 1, EM 2, EM 3) that correspond to the amplitude of the excitation spectrum:

Figure A2. Excitation of a Fluorophore at Three Different Wavelengths

Appendix B. Selected Excerpts From Papers in the Journal of Research of the National Institute of Standards and Technology

NOTE: These papers can be downloaded in their entirety as PDF files at no charge from the NIST website (www.nist.gov). The following excerpts were selected to augment material presented in this guideline.

B1. Excerpt from: The Development of Fluorescence Intensity Standards*

The figure below shows a schematic of the geometry of illumination and detection. The shaded volume is the sensing volume and is the volume over which the integral is taken. The sensing volume is the overlap of the illumination volume bounded by I and I´ and the detection volume bounded by D,D´. The geometry and various filters and optical elements give the instrumental factors in the fluorescence intensity measurements as described in Eq. (B3).

Figure B1. Schematic of the Geometry of Illumination and Detection

* Reprinted from Gaigalas AK, Li L, Henderson O, et al. J Res Natl Inst Stand Technol. 2001;106:383.

Appendix B. (Continued)

B2. Excerpt from: Quantitating Fluorescence Intensity from Fluorophore: The Definition of MESF Assignment*

NOTE: the term “flux” as used in the excerpt is synonymous with the term “power” as used in the guideline.

Derivation of the Measurement Equation

The experimental setup, execution, and analysis of measurements involving optical radiation are greatly simplified by using a measurement equation. This equation mathematically describes the measurement, relating the signal from the detector to the optical properties of the incident light, sample, and detection system. The measurement equation for the total fluorescent radiance is derived in this appendix in a four- step process. In the first three steps, equations are derived for the fluorescence radiant flux, the signal from the detection system, and the fluorescent spectral radiance. The last step combines all of the concepts into a measurement model for fluorescence radiance. Much of the derivation is based upon concepts presented by K. D. Mielenz, “Photoluminescence Spectrometry” in Optical Radiation Measurements, Volume 3, Measurement of Photoluminescence, ed. K. D. Mielenz, Academic Press, 1982, pp. 2-87.

1. Fluorescence radiant flux: Any measurement of optical radiation involves spectral fluxes over finite spectral bandwidths. In the case of fluorescence, fluxes of both excitation and emission are present. These are denoted by subscripts x and m, respectively. The radiant flux absorbed by a sample, whether it is fluorescent or not, is given by

−ε (λx )Nl Φ a (λx )∆λx = Φ i (λx )∆λx (1−10 ) , (B1)

where λx [nm] and ∆λx [nm] are the wavelength and spectral bandwidth of the excitation radiant flux, respectively, Φa [W/nm] and Φi [W/nm] are the absorbed and incident spectral fluxes, respectively, ε [cm2] is the absorptivity, N [1/cm3] is the concentration of fluorophore, and l [cm] is the path length. Expanding the expression in parentheses and keeping only the first-order term, Eq. (B1) becomes

Φ a (λx )∆λx = 2.3 N l ε(λx )Φ i (λx )∆λx . (B2)

The radiant flux from fluorescence is a fraction of the excitation radiant flux, and is given by

Φf (λm , λx )∆λm = y(λm , λx )∆λm Φa (λx )∆λx , (B3)

where λm [nm] and ∆λm [nm] are the wavelength and spectral bandwidth of the emission radiant flux, respectively, Φf [W/nm] is the fluorescent spectral flux, and y [1/nm] is the spectral quantum yield.

NOTE: The y relates the absorbed radiant flux at wavelength λx to the fluorescent radiant flux at wavelength λm.

Using the expression for the absorbed radiant flux from Eq. (B2), Eq. (B3) becomes

Φ f (λm , λx )∆λm = 2.3 N l y(λm , λ x )∆λm ε(λ x ) Φ i (λx )∆λx . (B4)

* Reprinted from Schwartz A, Wang L, Early E, et al. J Res Natl Inst Stand Technol. 2002;107(1):88-90.

Appendix B. (Continued)

2. Relation between detector signal and source spectral radiance: The figure below shows a schematic of an aperture system for a measurement system consisting of a wavelength selector and an optical detector (not shown in Figure B2), be it a filter and detector, a monochromator and a single detector, or a monochromator and an array detector. The signal at a given wavelength setting of the detection system is given by

S(λ ) = Φ (λ ) R(λ ,λ )dλ . (B5) 0 ∫ D m 0 m m

Here, λ0 [nm] is the wavelength setting of the detection system, S [A] is the signal, ΦD [W/nm] is the spectral flux entering the detector, R [A/W] is the responsivity whose dependence on both λ0 and λm is shown explicitly, and λm [nm] is the wavelength of the emission radiant flux. For any wavelength selector with a narrow spectral bandwidth, such as a monochromator, R(λ0, λm) is a sharply peaked function about λ0, in which case Eq. (B5) becomes

S(λ0 ) = Φ D (λ0 ) R(λ0 ) ∆λm . (B6)

The spectral flux entering the detector is related to the spectral radiance of the source by

AA Φ=()λλLL () Γ= () λES , (B7) D 00 0D2

2 2 where L [W/(m nm sr)] is the spectral radiance, AE [m ] is the area of the entrance aperture of the 2 2 collection optics, AS [m ] is the area of the image of the entrance slit in the source plane, D [m ] is the distance between the entrance aperture and the source plane, and Γ[m2 sr] is defined by Eq. (B7). Using Eq. (B7), Eq. (B6) becomes

S(λ0 ) = L(λ0 )Γ R(λ0 ) ∆λm . (B8)

3. Fluorescent spectral radiance: The fluorescence source in a cuvette measurement is a volume source with dimensions set by the illuminating beam. Radiance standards are surface sources with known radiance per unit area of the surface. To compare the fluorescence source to the standard source, we need to convert the volume source into an equivalent surface source. The spectral radiance from the fluorescence source can be defined in terms of the spectral flux from fluorescence, Eq. (B4), as

f 1 LFfmx(,)λ λλ∆= m Φ fmx (,) λλλ ∆ m =Φ fmx (,) λλλ ∆ m , (B9) 4π nASB

2 where AB [m ] is the cross-section area of the intersection between the cylindrical volume illuminated by the incident laser beam and the source plane (see Figure B2). The factor 4π is the total steradians into which the fluorescence is emitted and f is an angle-dependent function which describes possible nonisotropic emission. The solution index of refraction, nS, corrects for distortion due to refraction at the cuvette boundary. The factor F [1/m2sr] as defined in Eq. (B9), converts the total fluorescence radiant flux (W/nm) into fluorescence spectral radiance (W/(nm m2 sr). The conversion permits us to compare the fluorescence spectral radiance to standard sources whose properties are always defined in terms of spectral radiance. Using Eq. (B4), Eq (B9) becomes

LFNlyfmx(,)λ λλ∆= m 2.3(,) λλλελλλ mx ∆ m ()() x Φi x ∆ x. (B10)

Appendix B. (Continued)

4. Measurement of the fluorescent radiance:

The responsivity of the detection system is calibrated by a source of known spectral radiance L´ (λ0) [W/(m2 nm sr)]. This source can be either an integrating sphere with known spectral radiance at its exit port or a diffuse reflector, of known reflectance factor, illuminated by a source of known spectral irradiance. The emitting surface of the standard source coincides with the source plane in Figure B1.

Using Eq. (B8), the signal S´ (λ0) resulting from the source of known spectral radiance is

S´ (λo) = L´ (λo) Γ R (λo) ∆ λm . (B11)

Likewise, the signal Sf(λ0) resulting from the fluorescent source is

Sf (λ0 ) = Lf (λ0 )Γ R(λ0 ) ∆λm . (B12)

Taking the ratio of Eqs. (B11) and (B12) and using the expression for the fluorescent spectral radiance from Eq. (B10) yields

Sf (λ 0 ) Lf (λ 0 ) = L´(λ 0 ) = 2.3F N l y(λ m ,λ x )ε(λ x )Φ i (λ x )∆λ x . (B13) S´(λ 0 )

Equation (B13) expresses the measured fluorescent spectral radiance Lf at a given wavelength setting of the detection system λ0 both in terms of measured or known quantities (Sf, S´, and L´) and in terms of properties of the instrument (F, l, Φi(λx), and ∆λx) and the sample (N, y, and ε).

The quantity y(,)λmxλ relates the absorbed radiant flux at wavelength λx to the fluorescent radiant flux at wavelength λm. In other words, the radiant flux from fluorescence is a fraction y of the absorbed radiant flux. The radiant flux can be converted to a photon number flux by dividing the radiant flux by the energy of a single photon. Thus, the quantity y(,)λmxλ can be converted into a relation between fluorescence photon number flux and absorbed photon number flux by multiplying it by the ratio of the respective wavelengths (personal communication from Prof. W.V. Prestwich, Unit of Medical Physics, McMaster University).

λ m γ´ ()λ m ,λ x = γ ()λ m ,λ x (B14) λ x

The quantity γ´ ()λm ,λ x is conveniently separated into a quantum yield φ and a normalized relative ΄ photon emission function s (λm, λx) [1/nm], where

∫ s´ ()λ m ,λ x d λ m =1. (B15)

Eq. (B13) can be rewritten as ΄ ε

S f (λ m )λ m L'(λ m ) = 2.3F N l φs´(λ m ,λ x ) ε(λ x )Φ i (λ x )∆λ x . (B16) S´(λ m )λ x

The total fluorescent radiance Lf [W/(m2 sr)] (converted to photon flux) is given by

Appendix B. (Continued)

Equation (B17) constitutes the measurement model of fluorescence radiance. The measured fluorescence radiance (left side of [Eq. (B17)]) is equated to the fluorescence radiance which results from known solution and measurement properties (right side of Eq B17). The factor Ω describes some of the instrument characteristics.

Figure B2 shows a schematic of the geometry that is used to convert a volume fluorescence source into an equivalent surface source. AB is the cross-sectional area of the illuminating laser beam which is coming out of the plane of the figure. The source plane intersects the illuminating beam at its midsection. AE is the area of the entrance aperture of the collecting optics, and D is the distance between the entrance aperture and the source plane. Finally, AS is the area of the image of the monochromator entrance slit on the source plane. (AS < AB). The refraction at the cuvette wall is not shown in the figure. A radiance standard would replace the cuvette during calibration. The surface of the radiance standard is placed at the source plane.

Figure B2. Schematic of an Aperture System for a Measurement System

Appendix C. NIST Certificate for Standard Reference Material SRM 1932 (Reprinted with permission from NIST.)

This Appendix is excerpted from the NIST certificate for SRM 1932 issued on 4 March 2003. The certification is stipulated to expire on 31 December 2004. This information from this certificate is reproduced solely as an example of the detailed documentation required for an SRM to be used for calibrating@@ fluorescence Where is intensity. intro text? It should not be used in lieu of a current certificate provided with SRM 1932. Please consult NIST at www.nist.gov for the most recent information on this or any other reference material.

National Institute of Standards & Technology

Certificate of Analysis Standard Reference Material® 1932 Fluorescein Solution

This Standard Reference Material (SRM) is intended for use in establishing a reference scale for fluorescence intensity based upon MESF (molecules of equivalent soluble fluorophore) units [1-3]. This SRM is certified for the concentration of fluorescein with a certified purity in a borate buffer solution. The MESF scale is established for a particular set of experimental conditions by measuring the fluorescence intensity of known amounts of this SRM under the specified set of conditions as described in the Instructions for Use section of this certificate.

Each unit of SRM 1932 consists of three sealed amber glass ampoules containing a solution that consists of fluorescein in an aqueous borate buffer. Approximately 2.0 mL of the solution is flame-sealed into each individual ampoule that has been pre-scored for easy opening.

Certified Value: The certified concentration of fluorescein is given below. A NIST certified value is a value for which NIST has the highest confidence in its accuracy in that all known or suspected sources of bias have been investigated or accounted for by NIST [4]. The mass fraction (purity) of the fluorescein material used to prepare this SRM is 97.55 % ± 0.64 %.

Fluorescein: 60.97 µmol·kg-1 ± 0.40 µmol·kg-1

Reference Values: Table 1 gives reference values for the molar absorption coefficient of the SRM 1932 fluorescein solution compensated for photobleaching. NIST reference values are noncertified values, which represent the best estimate of the true values based on available data; however, the values do not meet the NIST criteria for certification [4] and are provided with associated uncertainties that may reflect only measurement precision, may not include all sources of uncertainty, or may reflect a lack of sufficient statistical agreement among multiple analytical methods.

Table 1. Molar Absorption Coefficient of SRM 1932 Fluorescein Solution at 22.5 °C ± 0.1 °C

Wavelength Molar Absorption Coefficient Standard Deviation (nm) (kg·cm−1·mol−1) (kg·cm−1·mol−1)

488.0 8.493 × 104 0.060 × 104 490.0 8.661 × 104 0.059 × 104 490.5 8.670 × 104 0.059 × 104 491.0 8.660 × 104 0.059 × 104

Expiration of Certification: The certification of this SRM is valid until 31 December 2004, within the measurement uncertainties specified, provided the SRM is handled and stored in accordance with the

Appendix C. (Continued) instructions given here (See Instructions for Use). The certification is valid only for unopened ampoules that have been stored in the dark between 2°C to 6 °C.

Information Values: Table 2 gives information values for SRM 1932. A NIST information value is a value that may be of use to the SRM user, but insufficient information is available to assess adequately the uncertainty associated with the value.

Table 2. Information Values for SRM 1932 Fluorescein Solution

pH Buffer Mass Density (25 °C) Concentration (22 °C)

9.48 0.10 mol·L-1 1.003 g·mL-1

Maintenance of SRM Certification: NIST will monitor this SRM over the period of its certification. If substantive changes occur that affect the certification before the expiration of this certificate, NIST will notify the purchaser. Return of the included registration card will facilitate notification.

Source of Material: The fluorescein (2-(6-hydroxy-3-oxo-3H-xanthen-9-yl)benzoic acid, C20H12O5, relative molecular mass 332.311, CAS No. 2321-07-5) powder used in SRM 1932 was prepared especially for this purpose by Molecular Probes, Inc.* (Eugene, OR, U.S.; Leiden, The Netherlands), 71358, MPR, Lot W018073. Boric acid granules were obtained from Mallinckrodt1 (St. Louis, MO, U.S.), Lot 2549 KVTK, relative molecular mass 61.83. Sodium hydroxide pellets were from Mallinckrodt1, Catalog No. 7708, Lot 7708M484721, relative molecular mass 40.00. All water used was NIST deionized water that was then subsequently passed through a second purification system (Millipore Mill-Q A101) to produce water having a resistivity ≥ 18 MΩ·cm.

Determination of Fluorescein Purity: Table 3 summarizes the purity determinations of the fluorescein material used to prepare SRM 1932.

* Certain commercial organizations, services, equipment, or materials are identified in this certificate to specify adequately the experimental procedure. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the organizations, services, materials or equipment identified are necessarily the best available for the purpose. An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 69 Number 26 NCCLS

Appendix C. (Continued)

Table 3. Summary of Fluorescein Purity Determinations and Associated Uncertainties

Constituent Technique Mass Fraction Uncertainty (%) (%)

Fluorescein 1H NMR† 97.55 0.64 DHBBA* 1H NMR† & HPLC‡ 1.04 0.20 Ethanol 1H NMR† 0.21 0.04 Methyl Isobutyl Ketone 1H NMR† 0.02 0.02 HOAc/Acetate 1H NMR† 0.02 0.01 Ethyl Acetate 1H NMR† 0.02 0.01 Water Karl Fisher 0.25 0.02 Total Organics CHO Analysis# 99.19 0.82 Potassium FAES¶ 0.50 0.14 Sodium FAES¶ 0.03 0.03 Chloride Argentimetry§ 0.37 0.03

† 500 MHz Proton Nuclear Magnetic Resonance Spectroscopy * 2-(2΄,4΄ Dihydroxybenzoyl) benzoic acid ‡ High Performance Liquid Chromatography and High Performance Liquid Chromatography-Mass Spectrometry, both normal and reversed phase # Elemental analyses carried out by Schwartzkopf Microanalytical Laboratory, Inc.1 (Woodside, NY, U.S.), Atlantic Microlab, Inc.1 (Norcross, GA, U.S.), and Galbraith Laboratories, Inc.1 (Knoxville, TN, U.S.) ¶ Flame Atomic Emission Spectroscopy § Halide content identified solely as chloride by X-ray Fluorescence and Inductively Coupled Plasma-Mass Spectrometry

Preparation of the SRM: The borate buffer solution was prepared by dissolving 18.56 g of boric acid granules in slightly less than 3 L of water. The pH was adjusted to above 9.1 using 40 mL of an approximately 3.0 mol·L-1 NaOH solution made by dissolving 29.87 g of sodium hydroxide pellets in 0.250 L of water. The mixture was diluted to 3.00 L with water to give a borate buffer with a boric acid concentration of about 0.10 mol· L-1.

The SRM solution was prepared in a darkened room by dissolving 0.05636 g ± 0.00004 g of fluorescein (buoyancy corrected) in 2.7134 kg ± 0.0006 kg of borate buffer solution (buoyancy corrected). The mass of fluorescein used was corrected for the fluorescein purity yielding a final fluorescein concentration of 60.97 µmol·kg-1 ± 0.40 µmol·kg- 1. The fluorescein solution contained in an amber bottle was immediately aliquotted into ampoules that were subsequently flame-sealed. The pH and mass density of the SRM were determined from measurements on four ampoules selected randomly from the lot.

Assignment of Uncertainties: Standard uncertainty components equivalent to the estimated standard deviation were assigned for sample inhomogeneity and measurement uncertainties. These values were then combined with balance accuracy and estimated instrument method uncertainties using the root-sum- of-squares method. An expansion factor of k = 2 was applied so that the expanded uncertainties given in this certificate express an interval within which the true value is expected to fall with a level of confidence of approximately 95 % for a normal distribution [5].

SRM Stability: NIST has monitored the stability of a prototype fluorescein solution similar to SRM 1932 for over 19 months. Within the error of the measurements, the absorbance spectrum and the fluorescence spectral radiance (as indicated by the fluorescence signal spectrum) of the prototype solution did not change. Therefore, SRM 1932, if stored in the dark between 2 °C to 6 °C, is likely to maintain its original optical properties for the duration of its certification period. NIST will validate this conclusion by periodic monitoring of the stability over the lifetime of the SRM (See Maintenance of SRM Certification section).

Appendix C. (Continued)

INSTRUCTIONS FOR USE

CAUTION: This SRM is a solution contained in tip-sealed glass ampoules with pre-scored stems. Therefore, all appropriate safety precautions, including the use of gloves during handling, should be taken to avoid accidental breakage or spillage. Unopened ampoules should be stored in the dark between 2 °C to 6 °C in an upright position. The ampoules should NOT be frozen because of the possibility of breakage during freezing and thawing. Once the ampoule is opened, the solution should be used promptly with minimal exposure to any light (exposure to incandescent lighting is preferable to illumination from daylight or fluorescent lighting). Any unused solution in the ampoule should be discarded properly (See the Material Safety Data Sheet (MSDS) accompanying the SRM).

Opening an Ampoule: When an ampoule is opened, that area of the stem where the pre-scored band is located (around the gold band) should be wiped with a clean, damp cloth and the body of the ampoule wrapped in absorbent material. Then holding the ampoule steady and with the thumb and forefinger grasping the stem above the gold band, minimal thumb pressure should be applied to the stem to snap it. Correctly done, the stem should break easily where pre-scored. The use of a metal file to break the stem is NOT recommended.

The SRM 1932 solution should always be diluted gravimetrically at least one-hundred fold with the same buffer system used for the analyte of interest. Subsequent gravimetric dilutions can then be employed to generate a calibration curve. The calibration curve describes the relationship between fluorescence intensity, as determined by the fluorometer and the concentration of fluorescein, as determined by the gravimetric dilution. Care must be taken in making the gravimetric dilutions because uncertainties expand during this serial process. An error at one level adversely affects determinations at all subsequent levels. The calibration curve measurements should be generated starting with the lowest concentration. For best results, the conditions used for determining the calibration curve such as solution degassing, temperature, ionic strength, pH, etc. should closely match those used in the measurement of the analyte. Because fluorescence is a highly sensitive technique, great attention must be paid to the cleanliness of glassware and any other apparatus that contacts the solutions. Many plastics and gloves can contaminate samples with small amounts of highly fluorescent materials such as release agents and plasticizers. The running of blanks to check for such contamination is highly recommended.

Depending on the solution pH, aqueous fluorescein solutions are complex, rapidly equilibrating mixtures of its several forms (cation, neutral species, monoanion, and dianion). Each species has unique absorbance and fluorescence spectra. Above pH 9, aqueous fluorescein exists almost exclusively as the highly fluorescent dianion. However, as the pH of the solution is reduced, the concentration of the dianion decreases, and the concentrations of the much less fluorescent monoanion and neutral forms increase [6]. Accordingly, the sensitivity of the fluorometric assay for fluorescein also decreases, and quantitation becomes very dependent on knowing or maintaining the precise pH of the solutions during calibration as well during the assay itself. While the calibration strategy described above can work with solutions below pH 9, the uncertainties of such measurements inevitably grow larger as the solution pH is lowered. REFERENCES

[1] Gaigalas, A.K.; Li, L.; Henderson, O.; Vogt, R.; Barr, J.; Marti, G.; Weaver, J.; Schwartz, A.; The Development of Fluorescence Intensity Standards; J. Res. Natl. Inst. Stand. Technol.; Vol. 106, pp. 381-389 (2001). [2] Schwartz, A.; Wang, L.; Early, E.; Gaigalas, A.K.; Zhang, Y-z.; Marti, G.E.; Vogt, R.F.; Quantitating Fluorescence Intensity from Fluorophore: The Definition of MESF Assignment; J. Res. Natl. Inst. Stand. Technol.; Vol. 107, pp. 83-91 (2002). [3] Wang, L.; Gaigalas, A.K.; Abbasi, F.; Marti, G.E.; Vogt, R.F.; Schwartz, A.; Quantitating Fluorescence Intensity From Fluorophores: Practical Use of MESF Values; J. Res. Natl. Inst. Stand. Technol.; Vol. 107, pp. 339-353 (2002). [4] May, W.; Parris, R.; Beck, C.; Fassett, J.; Greenberg, R.; Guenther, F.; Kramer, G.; Wise, S.; Gills, T.; Colbert, J.; Gettings, R.; MacDonald, B.; Definitions of Terms and Modes Used at NIST for Value-Assignment of Reference Materials for Chemical Measurements; NIST Special Publication 260-136 (2000).

Appendix C. (Continued)

[5] Guide to the Expression of Uncertainty in Measurement; ISBN 92-67-10188-9, 1st Ed., ISO, Geneva, Switzerland (1993); see also Taylor, B.N.; Kuyatt, C.E.; Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results; NIST Technical Note 1297, U.S. Government Printing Office; Washington, DC (1994); available at http://physics.nist.gov/Pubs. [6] Diehl, H.; Studies on Fluorescein–VIII; Talanta Vol. 36 (7), pp. 799-802 (1989) and references therein; Sjöback, R.; Nygren, J.; Kubista, M.; Absorption and Fluorescence Properties of Fluorescein; Spectrochemica Acta Part A; Vol. 51, pp. L7-L21 (1995).

Users of this SRM should ensure that the certificate in their possession is current. This can be accomplished by contacting the SRM Program at: telephone (301) 975-6776; fax (301) 926-4751; e-mail [email protected]; or via the Internet http://www.nist.gov/srm.

Appendix D. NIST Fluorescein Microsphere Reference Material RM 8640 (Reprinted with permission from NIST.)

This Appendix is a reproduction of the NIST Report of Investigation for Reference Material RM 8640: Microspheres with Immobilized Fluorescein Isothiocyanate. The reference values in this report are stipulated to expire after January 31, 2005. This report is reproduced solely to illustrate the documentation that accompanies this set of NIST reference microspheres to be used for calibrating fluorescence intensity. It should not be used in lieu of the current report of investigation that is provided when RM 8640 is obtained from NIST. Please consult NIST at www.nist.gov for the most recent information on this or any other reference material.

National Institute of Standards & Technology

Report of Investigation

Reference Material 8640

Microspheres with Immobilized Fluorescein Isothiocyanate

This Reference Material (RM) is intended for use in establishing a reference scale for fluorescence intensity based upon molecules of equivalent soluble fluorophore (MESF) units [1,2,3]. This RM provides reference values for the number of equivalent soluble fluorescein molecules on a microsphere (see Table 1). The MESF scale is established for a particular set of experimental conditions by measuring the fluorescence intensity of known amounts of this RM under the specified set of conditions, as described under the “Instructions for Use” section of this certificate.

Each unit of RM 8640 consists of six sealed plastic bottles containing a suspension that has a nominal microsphere concentration of 106 particles/mL. One bottle (Blank) contains microspheres without immobilized fluorescein isothiocyanate (FITC). The other five bottles contain microspheres with different amounts of immobilized FITC. Approximately 2.0 mL of the suspension is sealed into each individual plastic bottle.

Expiration of Reference Values: The reference values of this RM are valid until 31 January 2005, within the measurement uncertainties specified, provided the RM is handled and stored in accordance with the instructions given here, see “Instructions for Use”. The certification is valid only for unopened ampoules that have been stored in the dark at 4 °C ± 2 °C.

Maintenance of RM Reference Value: NIST will monitor this RM. If substantive changes occur that affect the reference value before the expiration date, NIST will notify the purchaser. Registration (see attached sheet) will facilitate notification.

The overall direction and coordination of technical measurements for this RM were performed by A.K. Gaigalas of the NIST Biotechnology Division.

Production of RM 8640 was performed by L. Wang and A.K. Gaigalas of the NIST Biotechnology Division.

Packaging of RM 8640 was coordinated through the NIST Standard Reference Materials Program by M.P. Cronise of the NIST Measurement Services Division.

Statistical consultation was provided by J. Lu of the NIST Statistical Engineering Division.

The support aspects involved in the issuance of this RM were coordinated through the NIST Standard Reference Materials Program by B.S. MacDonald of the NIST Measurement Services Division.

Vincent L. Vilker, Chief Biotechnology Division

Gaithersburg, MD 20899 John Rumble, Jr., Chief Certificate Issue Date: 19 February 2004 Measurement Services Division

Appendix D. (Continued)

Reference Values: A NIST reference value [4] is a noncertified value that is the best estimate of the true value; however, the value does not meet NIST criteria for certification and is provided with associated uncertainties that may reflect only measurement precision and may not include all sources of uncertainty. Table 1 gives reference values for the MESF of the RM 8640 as determined by averaging four independent measurements obtained with the fluorometer. The uncertainties, uc, were calculated from the standard deviation of these four independent determinations of MESF values [5].

Assignment of Uncertainties: Standard uncertainty components equivalent to the estimated standard deviation for a normal distribution were assigned for measurement repeatability and were combined with balance accuracy uncertainties and estimated instrument method uncertainties using the root-sum-of-squares method to produce the combined uncertainty, uc. An expansion factor of k = 2 has to be applied such that the expanded uncertainties express an interval within which the true value is expected to fall with a level of confidence of approximately 95 %.

Table 1. Average MESF Values of RM 8640 Microsphere Suspensions Assigned with Fluorometer

Bottle Number MESF*

1 1 700 ± 300 2 5 200 ± 600 3 15 400 ± 3 200 4 102 000 ± 19 400 5 250 000 ± 44 000

* Molecules of equivalent soluble fluorophore

Table 2 gives the average linearized MESF values. These were obtained by requiring that the MESF values correlate linearly with mean channels determined in a cytometer. The values are obtained from the best linear fit of the MESF values and the corresponding mean channel in a cytometer measurement. The slope of the fit is constrained to that characterizing the cytometer. These are the reference MESF values on the RM bottles.

Table 2. Average Linearized MESF Values of RM 8640 Microsphere Suspensions

Bottle Number MESF*

1 1 100 ± 300 2 4 700 ± 1 000 3 19 800 ± 4 300 4 113 700 ± 25 300 5 283 400 ± 65 000

* Molecules of equivalent soluble fluorophore

Source of Material: The microsphere suspensions were provided by Bangs† laboratory. All water used was NIST deionized water that was then subsequently passed through a second purification system (Millipore Mill-Q A101) to produce water having a resistivity ≥ 18 MΩ·cm.

Preparation of the RM: The microspheres were manufactured using a swollen emulsion polymerization process. The composition of the polymer is 70 % methyl methacrylate, 20 % glycidyl methacrylate and 10 % ethylene dimethacrylate by mass. The size is 7.56 µm, and the density is approximately 1.14 g/cm3. The fluorescein

† Certain commercial equipment, instruments, or materials are identified in this certificate in order to adequately specify the experimental procedure. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the materials or equipment identified are necessarily the best available for the purpose.

Appendix D. (Continued) isothiocyanate (FITC) molecules are covalently attached to the surface of the microspheres via a linker molecule. The chosen linker is optimized to provide a controlled distance between the microsphere surface and the fluorochrome (tether length). By controlling the tether length of the FITC molecule, the RM exhibits excitation and emission spectra very similar to that of FITC-labeled cell samples.

INSTRUCTIONS FOR USE

Stability and Storage: This RM is a suspension contained in opaque plastic bottles. CAUTION: Unopened and opened bottles should be stored in the dark at 4 °C ± 2 °C in an upright position. The bottles should NOT be frozen because of the possibility of degrading the suspension. Once the bottle is opened, the suspension should be used with minimal exposure to any light (exposure to incandescent lighting is preferable to illumination from daylight or fluorescent lighting). Any unused suspension in the bottle should be discarded properly.

Use: The RM 8640 suspension is intended for use in flow cytometers. The bottles are designed for repeated use of the suspension. Two drops of the suspension from each bottle may be placed in a solution identical to that used for the analyte. The resulting suspension containing all the microsphere populations should be run through the cytometer. The resulting calibration curve describes the relationship between mean channels of the microsphere populations, as determined by the cytometer, and the MESF values assigned to the five populations of microspheres. For best results, the conditions used for determining the calibration curve such as solution degassing, temperature, ionic strength, pH, etc. should closely match those used in the measurement of the analyte. The construction and use of a calibration curve is discussed in detail in reference 6.

Since fluorescence is a highly sensitive technique, great attention must be paid to the cleanliness of glassware and any other apparatus that contacts the suspension. Many plastics and gloves can contaminate samples with small amounts of highly fluorescent materials such as release agents and plasticizers. The running of blanks to check for such contamination is highly recommended.

Depending on the solution pH, aqueous fluorescein solutions are complex equilibrating mixtures of its several forms: cation, neutral species, monoanion, and dianion. Each species has unique absorbance and fluorescence spectra. Above pH 9, aqueous fluorescein exists almost exclusively as the highly fluorescent dianion. However, as the pH of the solution is reduced, the concentration of the dianion decreases, and the concentrations of the much less fluorescent monoanion and neutral forms increase. Accordingly, the sensitivity of the fluorometric assay for fluorescein also decreases, and quantitation becomes very dependent on the knowing or maintaining the precise pH of the solutions during calibration as well during the assay itself. While the calibration strategy described above can work with solutions below pH 9, the uncertainties of such measurements will inevitably grow larger as the solution pH is lowered. REFERENCES

[1] Gaigalas, A.K.; Li, L.; Henderson, O.; Vogt, R.; Barr, J.; Marti, G.; Weaver, J; Schwartz, A;. The Development of Fluorescence Intensity Standards; J. Res. Natl. Inst. Stds. & Tech.; U.S. Department of Commerce: Gaithersburg, MD; Vol. 106, pp. 381-389 (2001). [2] Schwartz, A.; Wang, L.; Early, E.; Gaigalas, A.K.; Zhang, Y–Z.; Marti, G.E.; Vogt, R.F.; Quantitating Fluorescence Intensity from Fluorophores: The Definition of MESF Assignment; J. Res. Natl. Inst. Stds. & Tech.; U.S. Department of Commerce: Gaithersburg, MD; Vol. 107, pp. 83-91 (2002). [3] Wang, L.; Gaigalas, A.K.; Abbasi, F.; Marti, G.E.; Vogt, R.F.; Schwartz, A.; Quantitating Fluorescence Intensity From Fluorophores: Practical Use of MESF Values; J. Res. Natl. Inst. Stds. & Tech.; U.S. Department of Commerce: Gaithersburg, MD; Vol. 107, pp. 339-353 (2002). [4] May, W.; Parris, R.; Beck, C.; Fassett, J.; Greenberg, R.; Guenther, F.; Kramer, G.; Wise, S; Gills, T; Colbert, J.; Gettings, R; MacDonald, B.; Definitions of Terms and Modes Used at NIST for Value-Assignment of Reference Materials for Chemical Measurements, NIST Special Publication 260-136; U.S. Government Printing Office: Washington, DC (2000). [5] ISO Guide to the Expression of Uncertainty in Measurement; ISBN 92-67-10188-9, 1st Ed., ISO: Geneva, Switzerland (1993): see also Taylor, B.N.; Kuyatt, C.E.; Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results; NIST Technical Note 1297, U.S. Government Printing Office: Washington, DC (1994); available at http://physics.nist.gov/Pubs.

Appendix D. (Continued)

[6] NCCLS I/LA24-A; Fluorescence Calibration and Quantitative Measurement of Fluorescence Intensity; Approved Guideline; ISBN 1-56238-543-7; NCCLS: Wayne, PA (2004).

Users of this RM should ensure that the report in their possession is current. This can be accomplished by contacting the SRM Program at: telephone (301) 975-6776; fax (301) 926-4751; e-mail [email protected]; or via the Internet at http://www.nist.gov/srm.

Appendix E. Measurement of Fluorescence Intensity Detection Efficiency (Q) and Optical Background (B) on a Flow Cytometer

Introduction

The ability of a flow cytometer to resolve dimly fluorescent particles from either background or from other dimly fluorescent particles depends primarily on two factors:

1) the efficiency (Q) in converting fluorochrome molecules on the particles into electrons that can be amplified and measured; and

2) the optical background (B) that is present even when no particle is being measured.

Q and B can be related to MESF by expressing Q in units of photoelectrons/MESF, and expressing B in units of MESF. In this formalism, B is measured in units of the number of MESF that would create the optical background when detected with efficiency Q.

Q and B affect the variance of dim fluorescence measurements and thus, the ability to resolve dimly fluorescent populations. The effect of these two factors on the variance of dim fluorescence measurements can be deduced from the physics of optical detection.1-3

A practical approach2 for determining Q and B uses uniform, but dimly fluorescent, beads to create light signals that are intrinsically quite uniform (i.e., low coefficient of variation (CV) or standard deviation (SD)). The measurement process limited by Q and B broadens the intrinsically narrow distributions. The broadening, measured by increasing standard deviation, contains the information needed to determine Q and B. The contribution to variance due to factors such as bead size or uniformity of staining, spatial uniformity of illumination, and fluctuations in the intensity of illumination is estimated from the CV of brightly fluorescent particles otherwise identical to the dimly fluorescent particles. This nonoptical contribution is subtracted from the total variance to give the optical variance due to Q and B. An MESF standard is used to calibrate the measurement scale, so that the MESF of the dimly fluorescent beads can be determined. Typically, the measured CVs of the dimly fluorescent beads will be in the range of 6 to 30% so that the increased CV is primarily due to the Q and B limitations of optical detection.

The method is most reliably done using a set of beads prepared from the same lot of unstained beads, which are then stained uniformly at various intensity levels. It is not necessary that the bead set closely match the emission spectrum of the fluorochromes used for MESF calibration, since an MESF calibrator is used to assign MESF units to the dimly fluorescent beads. A plot of the SD2 (corrected for the bright bead CV) vs the MESF of the dim beads should be a straight line1 with slope 1/Q and an intercept of B/Q.

Materials Required

Flow cytometer to be characterized

Fluorescent Microsphere Standards a) Relative Fluorescence Intensity (RFI) Calibrators with low intrinsic variability

These microspheres should be between two to ten microns in diameter and as uniform in size and fluorescence intensity as possible. The size and staining intensity should be controlled such that the intrinsic CVs of the stained microspheres are approximately the same and less than 3%. The intrinsic CV of the microspheres can be measured65 or preferably certified by the microsphere manufacturer.

Appendix E. (Continued)

The following levels of FI are required: (i) one level of brightly fluorescent microspheres with a measured CV; (ii) three levels of dimly fluorescent microspheres with intrinsic CVs close to that of the brightly stained microsphere, the highest of which should lie within the interpolative range of the MESF calibration curve on the instrument being assessed. For practical purposes, intrinsic CV of the dim fluorescent microspheres should be less than 4% to allow the CV (or SD) caused by photon counting statistics to be more accurately and precisely measured. To optimize the accuracy and precision of measurements for each population, the different levels of microspheres should be kept in their own suspensions and analyzed separately. b) MESF Calibrators

These microspheres should satisfy the requirements for MESF reference material (see Section 9.3.2) and provide a linear MESF calibration curve over the relevant range of FI measurements (see Section 11.2). Procedure

1. Determine the MESF value of the brightest of the three dim RFI calibrators.

Construct an MESF calibration curve using the MESF reference standards (see Section 11.2). Without changing instrument settings, analyze the brightest of the three dim RFI reference microspheres and interpolate its MESF value from the calibration curve.

NOTE: This MESF value applies only to the RFI microsphere when its FI is measured under the identical conditions as the MESF calibrators.

2. Measure the mean RFI and CV values for the RFI calibrators.

In the following procedure, data is gated on singlet microspheres either during acquisition or in subsequent analysis. As of the publication of this guideline, these procedures should be useable on any flow cytometer in the field. Multiparameter gating may be needed to isolate the fluorescence histogram populations into separate histograms if they overlap enough to affect the measurement of their mean and CV.

Though it may seem possible to collect the abovementioned data in a single run using a mixture of the microspheres and logarithmic amplification, the use of the linear data will provide greater precision and accuracy. The nonunitary width channels in a logarithmic histogram make statistical measurements less precise and accurate. Additionally, analog logarithmic amplifiers commonly found in flow cytometers typically exhibit slight deviations from an ideal logarithmic transformation. These deviations can lead to inaccurate measurements of the CV and mean intensities. Flow cytometers that use high resolution (at least 18-bit resolution) can use a logarithmic display of the fluorescence data for analysis as long as the statistical analysis is done on the underlying linear data. This will simplify the procedure by reducing the number of histograms required to complete the analysis.

1) Set up a flow cytometer to collect forward scatter vs. side scatter, and a single parameter linear scale histogram for the fluorescence data. Do not use log amplification for the fluorescence data. (Log display of high resolution linear data may be used.) Use a low sample rate (narrow

Appendix E. (Continued)

sample core stream) to obtain the most uniform illumination of the microspheres and smallest possible CVs. The fluorescence data should be gated on the forward and side scatter to exclude multiple microsphere data. The histograms should have sufficient resolution to detect singlet particles.

2) Adjust the forward scatter gain and side scatter gain and photomultiplier (PMT) voltage so that the microsphere distributions are placed on the scale. Do not change the gains or the side scatter PMT voltage between or while running the microsphere data sets.

3) Place the gate in the forward and side scatter histograms around the singlet microsphere population and collect gated events in the single parameter histogram for the fluorescence parameter. Collect enough gated events so that the histogram peak in the linear fluorescence histogram has at least 5000 events in it.

4) Collect a histogram of the linearly amplified fluorescence signal from the bright microsphere. Set the linear gain to one and adjust the PMT voltage to place the population peak in the upper portion of the histogram. Measure and record the CV of the population. Only the CV of the brightly stained microsphere needs to be measured. This measure is used to estimate the contribution to CV other than counting error: the intrinsic CV of the microspheres and the additional variance in incident light flux.

5) Measure the mean intensity and CV of each of the three dimmer fluorescently stained microspheres. The PMT voltage should be set so the population peak of the brightest of the dim microspheres (calibrated in MESF units) is in the upper portion of the histogram when the linear gain is set to one. Do not change the PMT voltage for the subsequent measurements of the dim microspheres and/or calibration microsphere. Adjust the linear gain to place the population peak in the histogram so the mean intensity and CV can be measured accurately. Measure the mean intensity and CV. Record the mean intensity divided by the linear gain, and the CV. Repeat this step for each of the other dimly stained microspheres.

3. Calculate detection efficiency (Q) and background light level (B).

1) If necessary, normalize the mean intensity of each two-dimmer microsphere to the gain range used for the highest microsphere that was calibrated in MESF units.

2) Convert the mean intensity for each of the dim microspheres from intensity units to MESF units using the following equation:

Calibrated Microsphere MESF Value MESF value = Normalized Mean Intensity × Calibrated Microsphere Mean Intensity

3) Convert the CVs for each of the dim microspheres to SD2 values using the following equation:

2 2 2 2 SD = MESF value × (CVdim − CVbright ).

Appendix E. (Continued)

4) Perform a linear regression correlating SD2 (y-values) with MESF intensity (x-values). See the figure below.

Plot SD-squared vs. MESF

1000000 y = 80.42x + 18928.26 R2 = 1.00

750000

500000 Beads 1-3

SD Squared

250000 Linear fit to data

0 0 2000 4000 6000 8000 10000 12000 MESF value

The linear regression equation for this plot is equivalent to:

SD2 = (1/Q)(MESF value) + B/Q

5) Calculate Q. 1/Q equals the slope of the line calculated by the linear regression.

6) Calculate B. B/Q equals the y-intercept of the line; thus, B is the y-intercept (B/Q) divided by the slope (1/Q) of the line calculated by the linear regression.

References for Appendix E:

1) Wood JCS. Fundamental flow cytometer properties governing sensitivity and resolution. Cytometry. 1998;33:260-266.

2) Chase ES, Hoffman RA. Resolution of dimly fluorescent particles: a practical measure of fluorescence sensitivity. Cytometry. 1998;33:267-279.

3) Steen HR. Noise, sensitivity, and resolution of flow cytometers. Cytometry. 1992;13:822-830.

Appendix F. A Classification System of Microsphere Fluorescence Standards Used for Flow Cytometry. (Schwartz A, Marti GE, Poon R, et al. Standardizing flow cytometry: A classification system of fluorescence standards used for flow cytometry. Cytometry. 1998;33(2):106-114. Copyright© 1998 Wiley & Sons, Inc. Reprinted by permission from Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.)

Physical Characteristics of Various Types of Standards for Instrument Setup and Calibration

Type O— Type I— Type II— Type III— certified alignment reference calibration Characteristics blank Ia Ib IIA IIB IIC IIIA IIIB IIIC

Size relative to Equivalent Smaller Equivalent Equivalent Equivalent Equivalent Equivalent Equivalent Equivalent lymphocytes

Size uniformity No Yes Yes No No No No No No (i.e., CV<2%)

Fluorescence Very low Very Very Bright Bright Bright* Dim to Dim to Dim to intensity level bright bright bright bright bright*

Number of Single Single Single Single Single Single* Multiple Multiple Multiple* fluorescence intensity levels

Excitation and NA† No No No Yes Yes* No Yes Yes emission spectra matching with sample

Environmentally No No No No Yes Yes* No Yes Yes* responsive

Antibody No No No No No Yes No No Yes binding capacity

*After binding of fluorescently labeled antibodies. †NA, not applicable.

NCCLS consensus procedures include an appeals process that is described in detail in Section 8 of the Administrative Procedures. For further information, contact the Executive Offices or visit our website at www.nccls.org.

Summary of Consensus/Delegate Comments and Committee Responses

I/LA24-P: Fluorescence Calibration and Quantitative Measurement of Fluorescence Intensity; Proposed Guideline

General

1. Fluorescence calibration and standardization of fluorescence measurement should be a CAP requirement, if not of high importance, particularly if patient results are reported using these instruments (flow cytometry). Immunology checks linearity, precision, and accuracy on our spectophotometers quarterly using a calibrated microtiter plate. Before this method, we prepared a color solution and made serial dilutions. Curve was plotted, slope. Y-interrupt and R2 values were calculated. Our fluorescent microscopes have PMs performed every six months, and 4+ and 1+ controls are used with each run to check fluorescence. Bulbs are also changed after 200 hours of usage. Using a NIST reference material 8640 for calibration and for MESF value makes sense and should be implemented. Since flow cytometry is not my area of expertise, I don’t understand all the calculation performed here. I am assuming that these have been performed by experts; therefore, I accept this proposal.

• The subcommittee appreciates and agrees with these comments.

2. There are many tests that can be performed using fluorescence. Since the patent has just expired, many companies are turning to this methodology for precise and accurate results. Like all laboratory procedures, there must be parameters ensuring repeatability and normals for specific populations, and ability to exclude interfering substances.

• The subcommittee appreciates and agrees with these comments.

3. Interesting reading – very detailed but aids in understanding. The Terminology and Definitions section will be particularly useful as an educational aid.

• The subcommittee appreciates and agrees with these comments.

Foreword

4. In the fourth paragraph, the first sentence: should read, “...bright or dim cell staining that reflected high or low levels of receptor expression...”

• The subcommittee agrees with the commenter, and the Foreword has been revised to read as suggested.

5. In the ninth paragraph, the third sentence, I suggest defining “FI” before using the abbreviation.

• The text has been revised to read: “the goal of measuring fluorescence intensity (FI) from a cell-bound ligand…”

Section 4.1, Definitions (Formerly Section 3.1)

6. Fluorochrome and Fluorometer definitions: Is “supercede” the English spelling? Don’t we use “supersede”? (Maybe for an international document it doesn’t matter.)

• The spelling has been changed to “supersede.”

Section 5.2.2.4, Polarization (Formerly Section 4.2.2.4)

7. In the second paragraph, third sentence, no hyphen is required between “relatively” and “large.”

• The correction has been made in the text.

Section 5.2.3.5, Other Environmental Factors (Formerly Section 4.2.3.5)

8. In the second paragraph, first sentence, I suggest the sentence read, “...affect their fluorescence properties” to make grammatically consistent with plural “fluorochromes.”

• The text has been revised as suggested.

Section 5.2.5.1, Background Fluorescence (Formerly Section 4.2.5.1)

9. In the first paragraph, the second sentence: should read, “…constituents…that fluoresce…” to make it grammatically consistent.

• The text has been revised as suggested.

Section 6.2.1, Fluorometers and Spectrophotofluorometers (Formerly Section 5.2.1)

10. In the first paragraph, next to last sentence: the grammar is cumbersome. Should “Emission leaving ... is directed...” read “Emissions ... are directed...”?

• For clarity, the sentence has been revised to read: “light that passes through the exit slit of this monochromator is directed to the photocathode of a PMT.”

Section 6.2.3, Capillary Electrophoresis and High Pressure Liquid Chromatography (Formerly Section 5.2.3)

11. In the second sentence I suggest using symbol “µ,” not the letter “u" for µm (or use the term “micron” or “micrometer”).

• The symbol “µ” is used in the text.

Section 7.2.1, Practical Aspects of Assigning MESF Values to Particles (Formerly Section 6.2.1)

12. The second sentence should read “...microspheres with low numbers of immobilized fluorochrome molecules...”

• The sentence was revised to read: “... microspheres with a low number of ...”

Section 8.6.1.2, Storage and Stability (Formerly Section 7.6.1.2)

13. Fifth sentence: I suggest using consistent phrasing for “< pH 5” and “pH > 9”; i.e., use “pH < 5” and “pH > 9.”

• The text has been revised as suggested.

Section 9.5, Sources of Fluorochrome Reference Microspheres (Formerly Section 8.5)

14. I suggest including the phrase ‘currently available’ – (as of 01/01/04).

• All uses of the word “currently” have been replaced with more appropriate wording such as: “as of the publication of this guideline...”

15. In the second paragraph, the second sentence should read, “...to the extent that they have been tested...”

• The text has been revised as suggested.

Section 10.2.3, Microheterogeneity in the Molecular Properties of Fluorochrome-Ligand Conjugates (Formerly Section 9.2.3)

16. In the first sentence, “homogenous” should be “homogeneous.”

• The spelling of the word has been changed as suggested.

Section 10.3.3, Biologic Calibrators for FLC Binding (Formerly Section 9.3.3)

17. In the second sentence, “lead” should be “led.”

• For clarity, this paragraph has been revised.

Section 10.4, Sources of Variability in Measuring FLC Binding Values (Formerly Section 9.4)

18. In the first sentence: the semicolon (;) after “F/P ratios” should be a comma (,).

• The text has been revised as suggested.

Section 10.4.1, Concentration and Binding Activity of Fluorochrome-Ligand Conjugates (Formerly Section 9.4.1)

19. In the second paragraph, second sentence, “confirmation” should be “conformation.”

• The text has been revised as suggested.

20. In the second paragraph, the fifth sentence should read, “...all the fluorochrome molecules...are indeed bound...” to make it grammatically consistent.

• The text has been revised as suggested.

Section 10.4.5, Multiple Interactions Between Fluorochrome-Ligand Conjugates and Binding Targets (Formerly Section 9.4.5)

21. In the fourth paragraph, the fourth sentence is cumbersome. The sentence should read, “...the subpopulation...” or perhaps “...subpopulations...”

• The sentence has been revised to read: “However, this may not always be possible, since some mixture of other FLC may be required to identify the subpopulation of cells on which the FLC of interest is being titered.”

Section 11.4.1, Basic MESF Calibration with Binding Values Derived from the Effective F/P Ratio (Formerly Section 10.4.1, Basic Calibration Using MESF Microsphere Standards)

22. In the third paragraph, fourth sentence “effect” should be “affect.”

• The text has been corrected as suggested.

Section 11.4.2, Binding Calibration with Capture Microsphere Standards (Formerly Section 10.4.2, FLC Binding Calibration with Capture Microsphere Standards)

23. In the second paragraph, last sentence: “regents” should be “reagents.”

• The text has been corrected as suggested.

Appendix A, An Overview of Molecular Fluorescence (Formerly Appendix A, An Overview of Fluorescence)

24. In the second paragraph, the fourth sentence should refer to “hvEx” and “hvEM,” from Figure A1, rather than “hEx” and “hEM.” 84 An NCCLS global consensus guideline. ©NCCLS. All rights reserved. Volume 24 I/LA24-A

• The printed document has the correct symbolism. This discrepancy might be related to differences in the program version used to print the document.

25. In Figure A2 should “EX 2” have a spectrum of its own?

• Only one excitation spectrum is present in this theoretical example. The designations labeled “EX 1,” “EX 2,” and “EX 3” are examples of different excitation points along that same spectrum.

Appendix B2, Excerpt from: Quantitating Fluorescence Intensity from Fluorophore: The Definition of MESF Assignment

26. Title: Should “Fluorphore” be “Fluorophore”?

• The word has been corrected as suggested.

2 27. After Eq. (B9): I suggest the paragraph begin: “where AB [m ] is the...”

• The text has been revised as suggested.

28. Eq. (B10) should have a period ending it.

• The period has been added at the end of the equation.

Appendix C, NIST Certificate for Standard Reference Material 1932

29. Ancillary measurements reference: Suggest correcting the sentence to read, “...Centers for Disease Control and Prevention...”

• This paragraph has been removed from the appendix, because it was not directly related to the characterization performed at NIST.

Appendix E, Measurement of Fluorescence Intensity Detection Efficiency (Q) and Optical Background (B) on a Flow Cytometer

30. Introduction, (1): The sentence should read, “…fluorochrome molecules…” instead of “…fluorochromes molecules…”

• The text has been revised as suggested.

31. Introduction, second paragraph: Suggest adding spaces around hyphen between “MESF” and “Q,” so the sentence reads more clearly as “...MESF - Q...” rather than “...MESF-Q...”

• For clarity, the sentence has been revised to read: “Q and B can be related to MESF by expressing Q in units of photoelectrons/MESF, and expressing B in units of MESF.”

32. Introduction, fourth paragraph, fourth sentence: “flutuations” should be “fluctuations.”

• The word has been corrected as suggested.

33. I suggest defining “RFI” before using the abbreviation.

• The text has been revised with the addition of “Relative Fluorescence Intensity” before the acronym (RFI).

The Quality System Approach

NCCLS subscribes to a quality system approach in the development of standards and guidelines, which facilitates project management; defines a document structure via a template; and provides a process to identify needed documents through a gap analysis. The approach is based on the model presented in the most current edition of NCCLS document HS1—A Quality System Model for Health Care. The quality system approach applies a core set of “quality system essentials” (QSEs), basic to any organization, to all operations in any healthcare service’s path of workflow. The QSEs provide the framework for delivery of any type of product or service, serving as a manager’s guide. The quality system essentials (QSEs) are:

Documents & Records Equipment Information Management Process Improvement Organization Purchasing & Inventory Occurrence Management Service & Satisfaction Personnel Process Control Assessment Facilities & Safety

I/LA24-A addresses the following quality system essentials (QSEs): Documents & Records Organization Personnel Equipment & Purchasing Inventory Process Control Information Management Occurrence Management Assessment Process Improvement Service & Satisfaction Facilities & Safety X X X X H42 Adapted from NCCLS document HS1—A Quality System Model for Health Care.

Path of Workflow

A path of workflow is the description of the necessary steps to deliver the particular product or service that the organization or entity provides. For example, GP26-A2 defines a clinical laboratory path of workflow which consists of three sequential processes: preanalytic, analytic, and postanalytic. All clinical laboratories follow these processes to deliver the laboratory’s services, namely quality laboratory information.

The NCCLS documents listed in the grid address the clinical laboratory path of workflow steps. For a description of the NCCLS documents listed in the grid, please refer to the Related NCCLS Publications section on the following page.

Preanalytic Analytic Postanalytic Patient Assessment Test Request Specimen Collection Specimen Transport Specimen Receipt Testing Review Laboratory Interpretation Results Report Post-test Specimen Management H42 H42 H42 H42 H42 H42 H43 H43 H43 H43 H43 H43 H44 H44 H44 H44 H44 H44 H44 H52 H52 H52 H52 H52 H52 Adapted from NCCLS document HS1—A Quality System Model for Health Care.

Related NCCLS Publications*

H42-A Clinical Applications of Flow Cytometry: Quality Assurance and Immunophenotyping of Lymphocytes; Approved Guideline (1998). This document contains guidance for the immunophenotypic analysis of non-neoplastic lymphocytes by immunofluorescence-based flow cytometry; guidelines for sample and instrument quality control; and precautions for data acquisition for lymphocytes.

H43-A Clinical Applications of Flow Cytometry: Immunophenotyping of Leukemic Cells; Approved Guideline (1998). This document provides performance guidelines for the immunophenotypic analysis of leukemic and lymphoma cells using immunofluorescence-based flow cytometry; guidelines for sample and instrument quality control; and precautions for data acquisition from leukemic cells.

H44-A2 Methods for Reticulocyte Counting (Flow Cytometry and Supravital Dyes); Approved Guideline – Second Edition (2004). This document provides guidance for the performance of reticulocyte counting by flow cytometry. It includes methods for determining the accuracy and precision of the reticulocyte flow cytometry instrument and a recommended reference procedure. An NCCLS-ICSH joint publication.

H52-A Fetal Red Cell Detection; Approved Guideline (2001). This document provides guidance for the quantitation of fetal red blood cells in blood and other biologic fluids. The performance characteristics of various flow cytometric and microscopic assays are reviewed, recommendations are made for control usage, and principles for distinction of F cells and fetal red cells are discussed.

* Proposed- and tentative-level documents are being advanced through the NCCLS consensus process; therefore, readers should refer to the most recent editions. An NCCLS global consensus guideline. ©NCCLS. All rights reserved. 87 Number 26 NCCLS

NOTES

Sustaining Members Commonwealth of Pennsylvania Gen-Probe Azienda Ospedale Di Lecco (Italy) Bureau of Laboratories GenVault Barnes-Jewish Hospital (MO) Abbott Laboratories Department of Veterans Affairs GlaxoSmithKline Baxter Regional Medical Center (AR) American Association for Clinical Deutsches Institut für Normung Greiner Bio-One Inc. Baystate Medical Center (MA) Chemistry (DIN) Immunicon Corporation Bbaguas Duzen Laboratories Bayer Corporation FDA Center for Devices and ImmunoSite, Inc. (Turkey) BD Radiological Health Instrumentation Laboratory BC Biomedical Laboratories (Surrey, Beckman Coulter, Inc. FDA Center for Veterinary Medicine International Technidyne BC, Canada) bioMérieux, Inc. FDA Division of Anti-Infective Corporation Bermuda Hospitals Board CLMA Drug Products I-STAT Corporation Boulder Community Hospital (CO) College of American Pathologists Iowa State Hygienic Laboratory Johnson and Johnson Pharmaceutical Brazosport Memorial Hospital (TX) GlaxoSmithKline Massachusetts Department of Public Research and Development, L.L.C. Broward General Medical Center Ortho-Clinical Diagnostics, Inc. Health Laboratories K.C.J. Enterprises (FL) Pfizer Inc National Center of Infectious and LAB-Interlink, Inc. Cadham Provincial Laboratory Roche Diagnostics, Inc. Parasitic Diseases (Bulgaria) LifeScan, Inc. (a Johnson & Johnson (Winnipeg, MB, Canada) National Health Laboratory Service Company) Calgary Laboratory Services (Calgary, Professional Members (South Africa) LUZ, Inc. AB, Canada) National Institute of Standards and Machaon Diagnostics California Pacific Medical Center American Academy of Family Technology Medical Device Consultants, Inc. Canterbury Health Laboratories (New Physicians National Pathology Accreditation Merck & Company, Inc. Zealand) American Association for Clinical Advisory Council (Australia) Minigrip/Zip-Pak Cape Breton Healthcare Complex Chemistry New York State Department of mvi Sciences (MA) (Nova Scotia, Canada) American Association for Health National Pathology Accreditation Carilion Consolidated Laboratory Respiratory Care Ontario Ministry of Health Advisory Council (Australia) (VA) American Chemical Society Pennsylvania Dept. of Health Nippon Becton Dickinson Co., Ltd. Carolinas Medical Center (NC) American Medical Technologists Saskatchewan Health-Provincial Nissui Pharmaceutical Co., Ltd. Cathay General Hospital (Taiwan) American Society for Clinical Laboratory Norfolk Associates, Inc. Central Texas Veterans Health Care Laboratory Science Scientific Institute of Public Health; Novartis Pharmaceuticals System American Society for Microbiology Belgium Ministry of Social Corporation Centro Diagnostico Italiano (Milano, American Society of Hematology Affairs, Public Health and the Olympus America, Inc. Italy) American Type Culture Collection, Environment Ortho-Clinical Diagnostics, Inc. Champlain Valley Physicians Inc. Swedish Institute for Infectious (Rochester, NY) Hospital (NY) Asociacion Mexicana de Bioquimica Disease Control Ortho-McNeil Pharmaceutical Chang Gung Memorial Hospital Clinica A.C. (Raritan, NJ) (Taiwan) Assn. of Public Health Laboratories Industry Members Oxoid Inc. Changi General Hospital (Singapore) Assoc. Micro. Clinici Italiani- Paratek Pharmaceuticals Children’s Hospital (NE) A.M.C.L.I. AB Biodisk Pfizer Animal Health Children’s Hospital & Clinics (MN) British Society for Antimicrobial Abbott Laboratories Pfizer Inc Children’s Hospital Medical Center Chemotherapy Abbott Diabetes Care Pfizer Italia Srl (Akron, OH) Canadian Society for Medical Acrometrix Corporation Powers Consulting Services Children’s Medical Center of Dallas Laboratory Science - Société Advancis Pharmaceutical Premier Inc. (TX) Canadienne de Science de Corporation Procter & Gamble Pharmaceuticals, CHR St. Joseph Warquignies Laboratoire Médical Affymetrix, Inc. Inc. (Belgium) Canadian Standards Association Alifax S.P.A. QSE Consulting Christus St. John Hospital (TX) Clinical Laboratory Management Ammirati Regulatory Consulting Quintiles, Inc. Clarian Health - Methodist Hospital Association Anna Longwell, PC Radiometer America, Inc. (IN) COLA A/S ROSCO Radiometer Medical A/S CLSI Laboratories (PA) College of American Pathologists AstraZeneca Pharmaceuticals Replidyne Community Hospital of Lancaster College of Medical Laboratory Aventis Roche Diagnostics GmbH (PA) Technologists of Ontario Axis-Shield POC AS Roche Diagnostics, Inc. Community Hospital of the College of Physicians and Surgeons Bayer Corporation - Elkhart, IN Roche Laboratories (Div. Hoffmann- Monterey Peninsula (CA) of Saskatchewan Bayer Corporation - Tarrytown, NY La Roche Inc.) CompuNet Clinical Laboratories (OH) ESCMID Bayer Corporation - West Haven, Sarstedt, Inc. Cook Children’s Medical Center (TX) International Council for CT Schering Corporation Cook County Hospital (IL) Standardization in Haematology BD Schleicher & Schuell, Inc. Covance Central Laboratory International Federation of BD Consumer Products Second Opinion Services (IN) Biomedical Laboratory Science BD Diagnostic Systems Seraphim Life Sciences Consulting Creighton University Medical Center International Federation of Clinical BD Thailand Ltd. LLC (NE) Chemistry BD VACUTAINER Systems SFBC Anapharm Danish Veterinary Laboratory Italian Society of Clinical Beckman Coulter, Inc. Streck Laboratories, Inc. (Denmark) Biochemistry and Clinical Beckman Coulter K.K. (Japan) SYN X Pharma Inc. Detroit Health Department (MI) Molecular Biology Bio-Development SRL Sysmex Corporation (Japan) DFS/CLIA Certification (NC) Japan Society of Clinical Chemistry Bio-Inova Life Sciences Sysmex Corporation (Long Grove, Diagnósticos da América S/A Japanese Committee for Clinical International IL) (Brazil) Laboratory Standards Biomedia Laboratories SDN BHD TheraDoc Dr. Everett Chalmers Hospital (New Joint Commission on Accreditation bioMérieux, Inc. (MO) Theravance Inc. Brunswick, Canada) of Healthcare Organizations Biometrology Consultants THYMED GmbH Duke University Medical Center National Academy of Clinical Bio-Rad Laboratories, Inc. Transasia Engineers (NC) Biochemistry Bio-Rad Laboratories, Inc. – France Trek Diagnostic Systems, Inc. Dwight David Eisenhower Army National Association of Testing Bio-Rad Laboratories, Inc. – Plano, Tyco Kendall Healthcare Medical Center (GA) Authorities - Australia TX Vetoquinol S.A. Eastern Health Pathology (Australia) National Society for BioVeris Corporation Vicuron Pharmaceuticals Inc. EMH Regional Medical Center (OH) Histotechnology, Inc. Blaine Healthcare Associates, Inc. Vysis, Inc. Emory University Hospital (GA) New Zealand Association of Bristol-Myers Squibb Company Wyeth Research Enzo Clinical Labs (NY) Phlebotomy Canadian External Quality XDX, Inc. Evangelical Community Hospital (PA) Ontario Medical Association Quality Assessment Laboratory YD Consultant Fairview-University Medical Center Management Program-Laboratory Cepheid YD Diagnostics (Seoul, Korea) (MN) Service Chen & Chen, LLC Florida Hospital East Orlando RCPA Quality Assurance Programs Chiron Corporation Trade Associations Focus Technologies (CA) PTY Limited ChromaVision Medical Systems, Focus Technologies (VA) Sociedad Espanola de Bioquimica Inc. AdvaMed Foothills Hospital (Calgary, AB, Clinica y Patologia Molecular Clinical Micro Sensors Japan Association of Clinical Canada) Sociedade Brasileira de Analises The Clinical Microbiology Institute Reagents Industries (Tokyo, Japan) Franciscan Shared Laboratory (WI) Clinicas Cognigen Fresno Community Hospital and Taiwanese Committee for Clinical CONOSCO Associate Active Members Medical Center Laboratory Standards (TCCLS) Copan Diagnostics Inc. Gamma Dynacare Medical Turkish Society of Microbiology Cosmetic Ingredient Review Academisch Ziekenhuis -VUB Laboratories (Ontario, Canada) Cubist Pharmaceuticals (Belgium) Geisinger Medical Center (PA) Government Members Dade Behring Inc. - Cupertino, CA Alfred I. duPont Hospital for Guthrie Clinic Laboratories (PA) Dade Behring Inc. - Deerfield, IL Children (DE) Hagerstown Medical Laboratory (MD) Armed Forces Institute of Pathology Dade Behring Inc. - Glasgow, DE Allina Health System (MN) Harris Methodist Fort Worth (TX) Association of Public Health Dade Behring Inc. - Marburg, American University of Beirut Hartford Hospital (CT) Laboratories Germany Medical Center (NY) Headwaters Health Authority BC Centre for Disease Control Dade Behring Inc. - Sacramento, CA Anne Arundel Medical Center (MD) (Alberta, Canada) Caribbean Epidemiology Centre David G. Rhoads Associates, Inc. Antwerp University Hospital Health Network Lab (PA) Centers for Disease Control and Diagnostic Products Corporation (Belgium) Health Partners Laboratories (VA) Prevention Digene Corporation Arkansas Department of Health Highlands Regional Medical Center Centers for Medicare & Medicaid Eiken Chemical Company, Ltd. ARUP at University Hospital (UT) (FL) Services Elanco Animal Health Associated Regional & University Hoag Memorial Hospital Centers for Medicare & Medicaid Electa Lab s.r.l. Pathologists (UT) Presbyterian (CA) Services/CLIA Program Enterprise Analysis Corporation Atlantic Health System (NJ) Holy Cross Hospital (MD) Chinese Committee for Clinical EXPERTech Associates, Inc. Aurora Consolidated Laboratories Hôpital du Sacré-Coeur de Montreal Laboratory Standards F. Hoffman-La Roche AG (WI) (Montreal, Quebec, Canada) Fort Dodge Animal Health AZ Sint-Jan (Belgium)

Hôpital Maisonneuve - Rosemont Marion County Health Department Providence Health Care (Vancouver, Stormont-Vail Regional Medical (Montreal, Canada) (IN) BC, Canada) Center (KS) Hôpital Saint-Luc (Montreal, Martin Luther King/Drew Medical Provincial Laboratory for Public Sun Health-Boswell Hospital (AZ) Quebec, Canada) Center (CA) Health (Edmonton, AB, Canada) Sunnybrook Health Science Center Hospital Consolidated Laboratories Massachusetts General Hospital Queen Elizabeth Hospital (Prince (ON, Canada) (MI) (Microbiology Laboratory) Edward Island, Canada) Swedish Medical Center - Providence Hospital for Sick Children (Toronto, MDS Metro Laboratory Services Queensland Health Pathology Campus (WA) ON, Canada) (Burnaby, BC, Canada) Services (Australia) Temple University Hospital (PA) Hospital Sousa Martins (Portugal) Medical College of Virginia Quest Diagnostics Incorporated Tenet Odessa Regional Hospital (TX) Hotel Dieu Grace Hospital (Windsor, Hospital (CA) The Toledo Hospital (OH) ON, Canada) Memorial Medical Center Quintiles Laboratories, Ltd. (GA) Touro Infirmary (LA) Huddinge University Hospital (Napoleon Avenue, New Orleans, Regions Hospital Tripler Army Medical Center (HI) (Sweden) LA) Rex Healthcare (NC) Truman Medical Center (MO) Hunter Area Health Service Memorial Medical Center (Jefferson Rhode Island Department of Health UCLA Medical Center (CA) (Australia) Davis Pkwy., New Orleans, LA) Laboratories UCSF Medical Center (CA) Indiana University Methodist Hospital (TX) Riverside Medical Center (IL) UNC Hospitals (NC) Innova Fairfax Hospital (VA) Michigan Department of Robert Wood Johnson University Unidad de Patologia Clinica Institute of Medical and Veterinary Community Health Hospital (NJ) (Mexico) Science (Australia) Mid America Clinical Laboratories, Royal Columbian Hospital (New Union Clinical Laboratory (Taiwan) International Health Management LLC (IN) Westminster, BC, Canada) Universita Campus Bio-Medico (Italy) Associates, Inc. (IL) Middlesex Hospital (CT) Sahlgrenska Universitetssjukhuset University College Hospital Jackson Memorial Hospital (FL) Monmouth Medical Center (NJ) (Sweden) (Galway, Ireland) Jacobi Medical Center (NY) Montreal Children’s Hospital St. Alexius Medical Center (ND) University Hospitals of Cleveland (OH) John C. Lincoln Hospital (AZ) (Canada) St. Anthony Hospital (CO) University of Alabama-Birmingham Johns Hopkins Medical Institutions Montreal General Hospital (Canada) St. Anthony’s Hospital (FL) Hospital (MD) National Serology Reference St. Barnabas Medical Center (NJ) University of Chicago Hospitals (IL) Kadlec Medical Center (WA) Laboratory (Australia) St-Eustache Hospital (Quebec, University of Colorado Hospital Kaiser Permanente (MD) NB Department of Health & Canada) University of Debrecen Medical Health Kangnam St. Mary’s Hospital Wellness (New Brunswick, St. Francis Medical Ctr. (CA) and Science Center (Hungary) (Korea) Canada) St. John Hospital and Medical University of Illinois Medical Center Kantonsspital (Switzerland) The Nebraska Medical Center Center (MI) University of the Ryukyus (Japan) Kenora-Rainy River Regional New Britain General Hospital (CT) St. John’s Hospital & Health Center The University of Texas Medical Laboratory Program (Ontario, New England Fertility Institute (CT) (CA) Branch Canada) New England Medical Center (MA) St. Joseph Mercy Hospital (MI) The University of the West Indies Kimball Medical Center (NJ) New York University Medical St. Joseph’s Hospital - Marshfield University of Virginia Medical King Faisal Specialist Hospital Center Clinic (WI) Center (Saudi Arabia) NorDx (ME) St. Joseph’s Hospital & Medical University of Washington LabCorp (NC) North Carolina State Laboratory of Center (AZ) UZ-KUL Medical Center (Belgium) Laboratoire de Santé Publique du Public Health St. Jude Children’s Research VA (Hampton) Medical Center (VA) Quebec (Canada) North Central Medical Center (TX) Hospital (TN) VA (Hines) Medical Center (IL) Laboratorio Dr. Echevarne (Spain) North Shore - Long Island Jewish St. Mary of the Plains Hospital VA (Tuskegee) Medical Center (AL) Laboratório Fleury S/C Ltda. Health System Laboratories (NY) (TX) Valley Children’s Hospital (CA) (Brazil) North Shore University Hospital St. Michael’s Hospital (Toronto, Vejle Hospital (Denmark) Laboratorio Manlab (Argentina) (NY) ON, Canada) Virginia Beach General Hospital (VA) Laboratory Corporation of America Northwestern Memorial Hospital Ste. Justine Hospital (Montreal, PQ, Virginia Department of Health (NJ) (IL) Canada) Virginia Regional Medical Center LAC and USC Healthcare Ochsner Clinic Foundation (LA) Salem Clinic (OR) (MN) Network (CA) O.L. Vrouwziekenhuis (Belgium) San Francisco General Hospital ViroMed Laboratories (MN) Lakeland Regional Medical Center Ordre professionnel des (CA) Washington Adventist Hospital (MD) (FL) technologists médicaux du Santa Clara Valley Medical Center Washoe Medical Center Landstuhl Regional Medical Center Québec (CA) Laboratory (NV) (APO AE) Orlando Regional Healthcare System Seoul Nat’l University Hospital Waterford Regional Hospital LeBonheur Children’s Medical (FL) (Korea) (Ireland) Center (TN) Ospedali Riuniti (Italy) Shands at the University of Florida Wellstar Health Systems (GA) Lewis-Gale Medical Center (VA) The Ottawa Hospital South Bend Medical Foundation West Jefferson Medical Center (LA) L'Hotel-Dieu de Quebec (Canada) (Ottawa, ON, Canada) (IN) Wilford Hall Medical Center (TX) Libero Instituto Univ. Campus OU Medical Center (OK) South Western Area Pathology William Beaumont Army Medical BioMedico (Italy) Our Lady of the Resurrection Service (Australia) Center (TX) Loma Linda Mercantile (CA) Medical Center (IL) Southern Maine Medical Center William Beaumont Hospital (MI) Long Beach Memorial Medical Pathology and Cytology Southwest Texas Methodist Hospital William Osler Health Centre Center (CA) Laboratories, Inc. (KY) Spartanburg Regional Medical (Brampton, ON, Canada) Louisiana State University Pathology Associates Medical Center (SC) Winn Army Community Hospital (GA) Medical Center Laboratories (WA) Specialty Laboratories, Inc. (CA) Winnipeg Regional Health Lourdes Hospital (KY) The Permanente Medical Group State of Washington Department of Authority (Winnipeg, Canada) Maccabi Medical Care and Health (CA) Health Wishard Memorial Hospital (IN) Fund (Israel) Peking University Shenzhen Stony Brook University Hospital Yonsei University College of Magnolia Regional Health Center Hospital (China) (NY) Medicine (Korea) (MS) Piedmont Hospital (GA) State of Connecticut Dept. of Public York Hospital (PA) Maimonides Medical Center (NY) Pocono Medical Center (PA) Health

OFFICERS BOARD OF DIRECTORS

Thomas L. Hearn, Ph.D., Susan Blonshine, RRT, RPFT, FAARC Willie E. May, Ph.D. President TechEd National Institute of Standards and Technology Centers for Disease Control and Prevention Kurt H. Davis, FCSMLS, CAE Gary L. Myers, Ph.D. Robert L. Habig, Ph.D., Canadian Society for Medical Laboratory Science Centers for Disease Control and Prevention President Elect Abbott Laboratories Mary Lou Gantzer, Ph.D. Klaus E. Stinshoff, Dr.rer.nat. Dade Behring Inc. Digene (Switzerland) Sàrl Wayne Brinster, Secretary Lillian J. Gill, M.S. Kiyoaki Watanabe, M.D. BD FDA Center for Devices and Radiological Health Keio University School of Medicine

Gerald A. Hoeltge, M.D., Carolyn D. Jones, J.D., M.P.H. Judith A. Yost, M.A., M.T.(ASCP) Treasurer AdvaMed Centers for Medicare & Medicaid Services The Cleveland Clinic Foundation J. Stephen Kroger, M.D., MACP Donna M. Meyer, Ph.D., COLA Immediate Past President CHRISTUS Health

John V. Bergen, Ph.D., Executive Vice President

NCCLS T 940 West Valley Road T Suite 1400 T Wayne, PA 19087 T USA T PHONE 610.688.0100

FAX 610.688.0700 T E-MAIL: [email protected] T WEBSITE: www.nccls.org T ISBN 1-56238-543-7