Transcription factor profiling in individual hematopoietic progenitors by digital RTPCR

1
Transcription factor profiling in individual
hematopoietic progenitors by digital RT-PCR

• Luigi Warren, David Bryder, Irving L. Weissman,
  and Stephen R. Quake

2
Background

• Stem cell differentiation
• Chemical state machine
• Sequencing logic implemented by cross-regulating
  transcription factors
• State of the network realized in the abundance
  profile of these regulatory molecules
• Transitions between states
• Instability, stochastic fluctuation, external
  signals
• Transcription factor PU.1
• Cytokine receptor flk2
• Housekeeping transcript GAPDH

3
Early Progenitors in the Hematopoietic Lineage
Tree
4
Goals

• Understand the behavior of transcriptional
  regulatory network for stem cell differentiation
• Leads to understanding of development
• Requires the ability to characterize network
  states quantitatively

5
Problems

• Network states cannot be characterized
  quantitatively
• Current gene profiling methods not sensitive
  enough
• Conventional gene expression assays
• Stem cells not easily isolated in such quantities
• Require thousands of cells worth of RNA as
  analyte
• Population-average expression data provide an
  incomplete picture
• Variations in network state determined by just a
  few phenotypic differences between cell types

6
Conventional PCR

• Quantitation based on number of cycles required
  for dye fluorescence to reach given threshold
• Exponential nature magnifies slight variations in
  amplification efficiency
• Interassay comparisons only valid if
  gene-of-interest measurements are normalized to
  measurements on endogenous controls or synthetic
  standards

7
Solution

• Digital RT-PCR
• Partition individual cDNA molecules into discrete
  reaction chambers before PCR amplification
• Quantitation uses binary, positive/negative calls
  for each subreaction within partitioned analyte
• Flow Cytometry
• Reveals diversity in patterns of surface protein
  expression within populations of superficially
  similar cells

8
Expression Profiling for CMP
9
FACS

• Fractionation of CMP cells into flk2 and flk2-
  subsets

10
Digital Array Chip

• cDNA from individual HSCs
• Green GAPDH
• Red PU.1
• Each well captures 0 or 1 template molecules

11
Results

• Number of individual cells expressing PU.1
• PU.1 expression up-regulated in CMP/flk2
• Down-regulated in CMP/flk2- cells and MEPs
• Higher GAPDH expression in CMPflk2cells.

12
Further Optimizations

• Threshold values
• Reference endogenous controls
• Weighted normalization of data
• mRNA vs protein turnover rate
• Measurement noise

13
Future Application

• Gene expression measurements can be made on an
  absolute, copy-number-per-cell basis
• Sophisticated regulatory network analysis
• Spread of public databases cataloguing
  cell-type-specific expression data
• Refinement of taxonomies through single-cell
  survey approach

14
Thank you

• Questions?