Publication of optimized multicolor immunofluorescence panels

The past 15 years have seen a dramatic advance in the multiparameter capabilities of flow cytometry and fluorescence microscopy. Commercially available flow cytometers are capable of measuring as many as 18 distinct fluorochromes simultaneously (1, 2); reagent vendors can supply conjugates (or custom-synthesize conjugates) to achieve this level of multiplexing. However, designing and optimizing multicolor panels remains as a time-consuming and challenging endeavor. There are a wide range of hurdles that need to be overcome, including qualification (e.g., titration) of each individual reagent, assessing potential interactions between reagents used in the same panel, choosing the best combination of fluorochrome-antibody pairs, and overcoming the loss in sensitivity that accompanies multicolor fluorescence experiments (3). In our laboratory, the design and optimization of a typical 12 color panel can require 1–4 months of work before we can begin using it experimentally. Many immunofluorescence panels will be useful to numerous laboratories; there should be no reason that each group should independently optimize it. However, currently there is no mechanism by which such information can be shared in a rigorous, consistent, and peer-reviewed fashion. Moreover, there has been no satisfactory way to recognize the significant effort needed to generate such panels, such as that provided by citation of research results. Publication and citation of research is the de facto standard for dissemination and attribution of scientific information; development and optimization of a multicolor panel is sufficiently novel and complex that it warrants such attribution. Therefore, we propose the creation of a new publication type in Cytometry A, the “Optimized Multicolor Immunofluorescence Panel” (OMIP). The goals of this publication type are: (1) to alleviate the development time for researchers who wish to use the same (or highly similar) panels, (2) to provide a starting point for the creation of novel OMIPs, and (3) to provide a mechanism for attribution to the developers of the panel via citation of the publication. We also propose that, the publisher of Cytometry A, Wiley-Blackwell, provide an interface to a database of the published OMIPs. This will provide an online method that researchers can use to search the existing OMIPs for combinations of reagents that they would like to use in their experiment, thereby quickly linking to the original submission and associated information. A number of criteria will be used to judge the suitability of an OMIP for publication. Preliminary submissions will be reviewed by a Cytometry A “Panels Editor” against these criteria to decide whether or not the OMIP should be sent out for peer review. Such peer review will follow standard mechanisms; reviewers will ensure that the publication has met the requirements for acceptability and all of the requirements outlined below. The criteria for acceptance of a new OMIP for review include: (1) Novelty–the OMIP should represent a unique combination of reagents designed to address a specific question that other panels do not address or do so demonstrably less well; (2) Thoroughly optimized–the publication must provide documentation for why particular reagents were selected, and how the final combination of reagents and fluorochromes was devised; (3) Utility–the panel should be useful to a range of researchers. The submission cover letter must address all three of these points in detail. Authors must also ensure that all required parts of an OMIP, as specified by the publication guide, have been provided. An OMIP publication will consist of two equally important parts: a two page print publication in Cytometry A and an extended online publication that contains additional key information, examples, protocols, as well as a link to updates or comments added after publication. In principle, OMIPs can be dynamic, reflecting ongoing development efforts that include new reagents, procedures, and information. While the original published document will exist unmodified, it should be easy to navigate online to such recent added material. We must strike a balance between too many and too few published OMIPs. If every panel is published, then the value of this series will be diluted, and researchers will have difficulty identifying the best of many panels that are only slightly different from each other. If publication is too restrictive, then the series becomes of limited value. To try to achieve this balance, among other reasons, we propose that the following criteria must be met. 1. An OMIP must comprise five or more colors. Panels comprising fewer colors are generally easy enough to construct that publication is not worthwhile. In addition, many combinations of fewer reagents will already be part of published OMIPs; whereas they will not be optimized for smaller combinations, it should be straightforward to do so. In very specific cases, OMIPs might comprise fewer than five colors, for example, for fluorescence microscopy applications or functional measurements using dyes with considerable spectral overlap, where optimization is particularly difficult. In the case of duplicate submissions, priority would generally be given to more complex OMIPs. 2. By definition, an OMIP has been thoroughly optimized. Considerable detail must be shown to ensure that optimization has been achieved. Optimization means different things to different researchers. It is necessary that extensive optimization has been achieved and documented. This includes detailed titration information and a detailed justification of why each particular reagent was chosen. Often, it will also be useful to show comparable panels using other conjugates and illustrating why they are inferior in preference of the final panel. A fundamental point is that it is not acceptable to publish the first panel that “worked” for the purposes of the authors' experiments. Many panels can be assembled fairly quickly and may be adequate; but these are not optimized. Optimization includes a comparison of many different available reagents, choosing among the best of each, and then generating an appropriate combination of those reagents to provide the best possible identification of the cells of interest. 3. An OMIP cannot be a trivial extension of a previously published panel. There is no black-and-white definition of what constitutes a trivial extension. Certainly, changing clones or titration would not constitute a new panel (but would be worthy of a modification of the existing panel). Adding a single reagent to an existing OMIP could be considered significant enough for publication and this will largely depend on the field (i.e., it would need to be an “important” reagent that significantly extends the power of the panel); however, it will be necessary to demonstrate reoptimization of the entire panel. Some OMIPs will be constructed with the express purpose of adding future reagents in a slot (for example, leaving the PE channel open for exploratory purposes); in that case, the new panel would not be publishable. Finally, a new OMIP that is related to an existing one must be demonstrably better. For example, OMIPs with only “dump” (negative) channel differences will be considered trivial extensions. Any OMIP submission must include a complete search on published OMIPs, with a cross-reference to any related OMIPs. The cover letter will indicate the differences from these OMIPs that warrant publication. 4. All reagents must be publicly available. All conjugates must be either commercially available, or must be synthesizable using published protocols, commercially available unconjugated antibodies, and commercially available fluorochromes. 5. OMIP publications may comprise no more than three authors. The optimization of a panel is best performed by one or two individuals who keep track of all details. It is rare that more than two individuals will provide sufficient intellectual input to the optimization of a single panel that warrants co-authorship. Authorship must be restricted to those participating in the optimization only and should not be extended to those who (e.g.) perform reagent manufacture, provide cell samples, or other core services that would be common to multiple OMIPs. We have provided a detailed guide as to what should comprise a published OMIP (see the Appendix). This guide should be considered a draft; we hope the community will provide input and suggestions before it becomes finalized. There are two components to a published OMIP, one that appears in print in Cytometry A, and the other as accompanying, detailed information available online. Both will adhere to rigorous formatting to allow for ease of understanding and to ensure that all required information is present. The component appearing in print will be a two-page publication that contains information such as a brief narrative on the use of the panel, a reagent table, and an example staining figure The bulk of the OMIP will be available online. This component provides technical details required for the execution of the panel, details on the optimization, hints, and comments about use. By design, the details of the implementation (such as clone and titration information) are not provided in the print version so that readers must consult the online information. This encourages researchers to refer to the detailed information as well as any revisions and updates to the panel seen. An important facet of OMIPs is their dynamic nature enabled by deposition of modifications online, easily accessible from the online presentation of the original publication. Changes brought about by discontinued or improved reagents, new fluorochrome options, or new information can be added to this portion by the authors; additional information from other researchers may be added (in an Editor- and possibly peer-reviewed manner) to provide additional information. Such additional information will be presented on a separate web page within the OMIP, thus preserving the original contribution but allowing for easy access to revisions and comments. If new OMIPs are developed, which are related to (or supplant) an existing OMIP, then appropriate cross-references will be inserted to ensure that researchers are aware of the alternatives. The first two OMIPs (4, 5) are published in this issue of Cytometry A and can serve as a template for future submissions. The Editor-in-Chief of Cytometry A has also issued a “call for submissions” for OMIPs to comprise a special issue to be published soon (6). We have proposed a new publication within Cytometry A designed to distribute detailed information about the creation, optimization, and use of multicolor immunophenotyping panels. Associated with this would be an online database that can be used to search through the library of published OMIPS for ones most relevant to a given application. OMIPs will be peer-reviewed and must rigorously adhere to strict publication guidelines. These publications will serve to distribute information about panels to the greater scientific community in a manner that can be attributable through standard citation in publications, thereby recognizing the large amount of work that they represent. There are two components to an OMIP: the printed version appearing in Cytometry A and the online portion containing considerable detailed information. The contents for both are strictly defined, although there is freedom in the supplementary, online section to provide as much supporting and additional data as necessary. Note that we specifically avoid the use of “supplemental” to describe the information provided electronically online. This content contains critical information and represents the majority of the OMIP. The title: The title will be suggested by the authors; it must be short and descriptive. It will be prefixed by a unique incrementing OMIP number generated by Cytometry. The unique number is necessary, as we expect many OMIP titles to be similar if not identical. The title should indicate the species for which the OMIP is designed (e.g., human, mouse, etc.). A summary table: The summary table will be supplied as a template for authors to fill in. It will contain information about the purpose of the panel (one sentence); the species; the cell types applicable (special preparation conditions); and a cross-reference of related, previously-published OMIPs. A brief narrative: The narrative provides some background on the need for the panel. In addition, the narrative discusses the overall approach that was used for developing the panel, and any important issues regarding the limitations or specific uses of the panel. The narrative will general describe the use of the panel (e.g., gating strategies, types of controls that the panel may require, general incubation conditions). A reagent table: The reagent table will list all conjugates (and other fluorescent reagents) used in the OMIP. Only the reagent specificity, clone, and fluorochrome will be listed. The online reagent table will be far more complete (see below). The reagent table should list primary and secondary reagents, if used, in separate rows; the order of the rows is up to the author. The fluorochrome entry will use an abbreviated fluorochrome name from a standard list supplied in the User Guide. This is to eliminate any confusion and ensure ease of comparison with the online database entries. An example staining figure: The example figure will illustrate all fluorescent channels, perhaps in the form of a sample “gating tree.” It should show the salient features of the panel. In general, the Figure should occupy one-quarter to one-half of a page, with figure legend. Similar OMIPs: If similar OMIPs have been published, this section would detail important differences from those existing panels. The online portion of the OMIP is far more important than a typical supplementary addendum to printed research papers. It provides technical details required for the execution of the panel, details on the optimization, hints and comments about use, and a link to any updates to the panel. By design, the details of the implementation (such as clone and titration information) are not provided in the print version in order that readers must consult the online information. This also enables easy, linked access to the revisions and updates to the OMIP. (Required) Panel development strategy: The development and optimization of a multicolor panel is a complex, multistep process. The online material must reflect this process in order to provide the greatest degree of confidence that the work of the authors is complete. The authors will strive to balance between showing too much detail (losing focus on the optimization), and showing too little detail (possibly leaving out important steps that justify a given panel arrangement or optimization step). In general, this strategy should identify the range of conjugates that were tested, a justification for why any given set of reagents were chosen over any other (even if this was arbitrary or economic). Special emphasis must be given to tertiary and critical reagents. The authors should anticipate reviewer and editorial requests for comparisons of one or more reagents to other comparable commercially-available reagents. (Required) Cross-references to related panels: Any highly related panels already published must be listed here. For pre-existing panels, a brief discussion of differences should be given. For panels that are essentially overlapping (information content), a direct comparison must be shown on replicates of a set of samples. This cross-reference will be updated to refer to new panels published later. (Required) Exact staining protocol: To reproduce the panel, readers must be able to reproduce the staining conditions. Details such as incubation time, temperature, buffers, and so forth must be provided. (Required) Instrument configuration (Online Table 1): This will allow readers to evaluate whether or not a given panel can work on their instrument. Information in this table should include, for each fluorochrome that is listed in the reagent tables: the laser used to excite the fluorochrome and the bandpass filters that directly regulate the spectrum of light reaching the detector (often, this includes three filters: two dichroic filters and a bandpass filter). The Table notes should include the laser type, power, and wavelength for each laser. This table will always be Online Table 1. (Required) Reagent Information (Online Tables 2–4): The online reagent tables will contain the following information for each reagent. There are 1–3 such tables: Online Table 2 will list commercially-purchased detection reagents. Online Table 3 will list fluorescent detection reagents manufactured in-house (from commercially-available resources). Online Table 4 will list fluorescent probes manufactured in-house for any tandem dye reagents in Online Table 3. Online Tables 3 and 4 are only provided if necessary. Online Table 2 will list: Specificity of the conjugated reagent. Fluorochrome. Clone. Vendor and catalog number. Titration information, given as a final dilution (e.g., 1:100) or a concentration (e.g., 1 μg/ml). Online Table 3 will list Specificity of the unconjugated reagent. Clone. Vendor and catalog number. Fluorochrome. Vendor and catalog number of the fluorochrome. Nominal Dye:Protein ratio used during conjugation. Titration. The Table notes should indicate the conditions used for titration (incubation time, temperature, and, if applicable, which reagents were used after fixation). Composite Fluorochrome name. Donor fluorochrome. Donor fluorochrome vendor and catalog number. Acceptor fluorochrome. Acceptor fluorochrome vendor and catalog number. Nominal acceptor to donor ratio used during conjugation. (Required) Titrations of each individual component: For each reagent, a titration against cells must be shown (for nonantibody conjugates, such as propidium iodide or other viability markers, titrations are not necessary but are welcome). The titration curves should demonstrate a range of concentrations justifying the final choice: that is, ranging at least as high as the used concentration (ideally showing saturation characteristics), and ranging at least several-fold below the used concentration. General, two- or three-fold dilutions should be used in titrations. If the choice of concentration is not based on obvious criteria such as the lowest saturating dose, then a statement should be included as to what these criteria were. (Optional) Brief description of sample data files and analysis that have been uploaded electronically. At least one, and preferably a few, sample FCS data files must be uploaded for public access. At a minimum, a file with data from a fully-stained sample must be present; additional samples or FMO controls to illustrate gating could also be included. A “workspace” document that references the data file and shows sample gating hierarchy is highly recommended. Ideally, the data file(s) uploaded would be those used for the Figure in the Print portion of the OMIP. (Optional) FMO controls for components: FMO controls are an excellent way to illustrate the use of reagents that are difficult to gate on (e.g., are dimly-expressed); they can be used to justify the appropriateness of the choice of a reagent titration. For many reagents (e.g., CD3, CD4, CD8), FMOs may not be very informative and should only be included if there is a good reason. (Optional) Biological controls (multiple samples): If there is a wide range of expression patterns revealed by a panel (e.g., between cells from different individuals), or if the pattern is variable on different tissues used by the authors, then such examples are highly useful to show. If a panel was specifically optimized on a given tissue, but could be used on another tissue (e.g. PBMC) but is not optimized for those cells, illustrative data and discussion on this point would be useful. (Publisher-supplied) Publications citing this OMIP: A list of publications that use this OMIP. This will provide researchers with knowledge of how useful the OMIP has been in other laboratories; how widely-used it has become, etc. The Publisher will maintain this section.