flow cytometry

NBS versus FCS versus BSA

NBS... Newborn Calf Serum... originates from 3 to 10 day old calves. Various suppliers claim that NCS  contains a somewhat higher proportion of proteins and immunoglobulin, but fewer growth factors. FCS... Fetal Calf Serum... originates from bovine fetus. Supposedly contains less antibodies but more growth factors. BSA... Bovine Serum Albumin... nicknamed "Fraction V" as per the Cohn process for separation of albumin from blood plasma. 

I could have sworn we used BSA in the wash buffer at the Emory Vaccine Center, but Altman's protocol calls for NCS. Tenang and McCaldin looked at the differential effect of NCS versus FBS in cell culture. Bockman describes a method to discern fetal serum from other sera, via Western blotting and ConA staining. Pollard and Walker, in Animal Cell Culture, write that the plating of bone marrow cells is better in NCS rather than FCS. A cursory search for NCS as used in flow seems to appear only in the context of functional cellular analysis: here and here. 

For those not interested in mucking about with "baby cows", Johansonn et all describe using platelet lysate to replace FBS. As with all things manufactured... there is an industry association for serum.

Staining Index

As per Maecker and Trotter...

...is defined as D/W, where D is the difference between positive and negative populations and W is equal to 2 SD of the negative population. And since FCS Express 3 does not offer SD*, we'll have to go commando, using Shapiro's instructions on page 235 of Practical Flow 4th Ed.

Staining Index = (geoMFI pos - geoMFI neg) / 2* (arithmetic mean*CV)

 

* which is completely fine and reasonable since Version 4 is available.

 

CD117

CD117... aka C-Kit... cytokine receptor for stem cell factor (SCF).

According to Canonico B, Felici C, Papa S.... c-kit expression in acute leukemia can be considered as an indicator of myeloid origin of tumor cells as it is rarely expressed in lymphoid blast cells. Ortolani in Flow Cytometry of Hematological Malignancies describes CD117 staining as "less likely" in AML-M4, "can be expressed in" more immature AML-5a, and "often CD117-" in more mature AML-5a. In AML-5b, Dunphy/Orton/Mantel show flow data with dim CD117 expression. This article had a nice series of plots for CD117 vs CD64, excerpts of which are below with ommision of other panel components:

healthy normal	CD117- / CD64+	monos
CMML	CD117- / CD64+	"25% mature monos"
CML-2	CD117 very dim / CD64+	"2:3 ratio immature to mature"
CML-1	CD117 dim / CDE64-	"predominantly mature monos"
AML-5b	CD117 int / CD64+	"mostly immature monocytic"

 From the BD TDS for #550412... clone YB5.B8... listed in the references is Ashman et al. describing epitope maping and function studies, in a 1994 article in J Cell Physiol. Mast cell tumors can be diagnosed via CD117 staining, according to Leong et al. in the second edition of "Manual of Diagnostic Cytology". Ashman's review of the biology of c-Kit is fine but doesn't address the issue of a biological basis for transient incidents of high background staining by flow.

FGFR3

Fibroblast growth factor receptor 3...

AKA CD333.
R&D Systems seems to have the only clone specifically marketed for flow cytometry.
Santa Cruz Biotech has SU5402  in RUO format.

Chesi et al. posit that it's an oncogene for multiple myeloma: "Myeloma cells carrying a t(4:14) translocation express a functional FGFR3 that in some cases is constitutively activated...". Zhihua Li et al. declare that circulating tumor cells from mice with FGFR3-TD expression in the bone marrow "revealed a pre-B-cell pheotype... although immature T-lymphoid or mature myelome populations also predominated in some animals". Jun-Hyeog Jang et al. identified "two novel mutations of FGFR3 in colorectal carcinomas." Joshua L Patterson et al. looked at FGFR3 inhibitor SU5402. It seems to inhibit receptor-dependent growth of human MM cell lines via cell cycle arrest and increased apoptosis. Patterson laments that "given the difficulty in establishing [primary myeloma cell cultures] under physiologically relevant conditions, we believe that only clinical trials of FGFR3 kinase inhibitors will formally establish the role of such drugs in MM therapy."

And now for some flow:
http://onlinelibrary.wiley.com/doi/10.1111/j.1365-2141.2006.06479.x/pdf
http://clincancerres.aacrjournals.org/content/17/13/4389.full

 From the Chandesris paper, plotting of FGFR3 data.

LIMS with Flow Cytometry

There has been more and more talk over at FlowJo about various third-party system integrations - with LabArchives, Bioconductor/R via GenePattern, CytoGenie/Evertrieve and others.

Here's a general list of more LIMS from CAP Today. News about LIMS systems can also be found at the LIMS Software Directory  and LIMSource. Details about which packages play nice with flow cytometry data are soon to be posted to FlowJo's blog, the Daily Dongle.

Biexponential Transformation settings

 Mucking about in FlowJo is fun, especially when you can get immediate visual feedback from changing a transformation setting. For the below, I've chosen to plot only in Pseudocolor and vary only one setting at a time in FlowJo Mac version 9.3.1.  For a series of lovely movies that manipulate all three settings at once, check out the Daily Dongle's Width Basis Titration. 

In short, reducing the width basis value compresses the linear space around zero. See here for settings at -1, -10, -100, and -1000. It results in more or less plotspace devoted to the Negative and Dim populations. I find that customizing this setting to be most useful for investigating dim events.

     

For the positive decades, the setting reflects the first decade that is displayed  after the linear space ends. It feels like the plot zooms-in on the intermediately bright populations. See here for settings as applied at 3, 3.5, 4, 4.5, and 5.

When it comes to the additional negative decades, settings result in more and more space devoted in plotspace for negative events. See here for 0, 0.5, 1, 1.5. I don't find altering them very useful.

 

 

 

Update on Reference Ranges

A new round of bead gating is ready, this time with gating directions updated for doublet discrimination. If you would like to participate, you can download the bead files here. To read the details in the original posting about this, follow this link.

Doublet discrimination gating

I found that using a positive control for douplets helps figure out the best gating method for doublet discrimination. Use this Doublet Discrimination on the FACSAria PDF to see plotting of Area, Height and Width for chicken erythrocyte nuclei from BD's DNA QC Particle Kit. I'd be happy to hear your comments. 

With thanks to the Emory Vaccine Center's  flow cytometry core and Dr. John Altman.

Who to talk to about Assay Validation

 Julie Wilkinson at Immunosite Technologies.

David Favre at National Immune Monitoring.

Reference ranges for clustering

In a collaborative effort with FlowJo, Steve Perfetto, and Ryan Brinkman's group,  I'm collecting statistics from manual gating on bead files. The pupose is to define a reference range of statistics that adequately describes the variation from one manual analyst to another. The files in the analysis set are completely artificial, and eliminate any experience bias that may occur if real-world samples were provided to a variety of individuals possessing different levels of exposure to real population gating. The context for this collection is development of an evaluation method for automated gating algorithm. More on automated cluster detection and the related Q and B platform in development can be found on the FlowJo and FlowDx websites. 

Members of the Brinkman lab are well-known in the open-source community of Bioconductor/R for their flow cytometry packages, and the exercise is expected to have significant benefit to developers on any platform.  Would you like to get involved in cluster development too? Or you can download the files and instructions here. Send an email to my FlowJo account: suzanne at treestar dot com and I can walk you through the analysis. The gist of it is in the pic above.