Abstract
Gene expression profiles of five consecutive stages of mouse B cell development were generated with high-density oligonucleotide arrays from as few as 2 x 10(4) ex vivo isolated and flow-cytometrically purified cells. Between 2.8% and 6.8% of all genes change on differentiation from one cellular stage to the next by at least twofold. The entire pathway involves differential expression of 10.7% of all genes. Previously known expression patterns of 15 genes (like surrogate light chain, RAG-1/2, MHC class II, mel-14 antigen) are confirmed. The gene expression patterns of the proliferating pre-BI and large pre-BII cells on the one hand, and the resting immature and mature B cells on the other hand, are most similar to each other. Small pre-BII cells display a pattern that is transitional between these two groups. Most of the genes expressed in early precursors are involved in general processes, like protein folding or cell cycle regulation, whereas more mature precursors express genes involved in more specific molecular programs. Combining knowledge about gene function and expression pattern allows identification of novel candidate genes potentially involved in self-maintenance of pre-BI cells, allelic exclusion and pre-B cell receptor signaling in large pre BII cells, cell-cycle arrest of small pre-BII cells, propensity toward apoptosis or anergization in immature B cells, propensity toward cell division and activation in mature B cells, and stage-specific interactions with stromal cells in the bone marrow.
Introduction
B-cell development is a highly orchestrated process where progenitors in the bone marrow undergo a series of genetic rearrangements and selection checkpoints to become functional, antibody-producing cells. While several "master regulator" genes like RAG-1 and VpreB had been identified, a global view of the molecular program was previously missing due to technical limitations. High-density oligonucleotide arrays typically required millions of cells, but specific developmental subsets are extremely rare in vivo. This study addresses this bottleneck by developing a high-sensitivity profiling method to identify the genes and pathways governing B-cell maturation, allelic exclusion, and cell-cycle exit.
Methods
The researchers combined high-precision cell sorting with a specialized RNA amplification procedure. Five consecutive stages of B-cell development—pre-BI, large pre-BII, small pre-BII, immature B, and mature B—were purified using flow cytometry (FACS) based on surface markers such as c-kit, CD25, IgM, and IgD.
To overcome the low cell count (as few as 20,000 cells), a modified in vitro transcription-based amplification procedure was used to boost the mRNA signal by a factor of 1,000 while maintaining relative transcript proportions. Samples were hybridized to Affymetrix Mu11k arrays, interrogating over 13,000 transcripts. Differentially expressed genes were identified via ANOVA, and Self-Organizing Maps (SOMs) were used to cluster genes into 20 distinct expression signatures.
Key Results and Insights
The study revealed that 10.7% of all monitored genes (1,406 transcripts) changed significantly across the developmental pathway. The most dramatic transcriptomic shift occurred between cycling cells (pre-BI and large pre-BII) and resting cells (immature and mature B), with small pre-BII cells representing a transitional state.
Key findings include:
- Functional Transition: Early precursors primarily express genes for "housekeeping" and growth (protein folding, DNA replication), while mature cells switch to specialized programs (cell surface receptors, adhesion molecules).
- Expression Signatures: The 20 identified SOM clusters highlight specific biological roles. For instance, Cluster 3 contains genes for DNA replication, while Cluster 17 contains signaling molecules specific to mature B cells.
- Novel Candidates: The study identified hundreds of previously unknown genes associated with checkpoints, such as MyD118 and gadd45 in the cell-cycle arrest of small pre-BII cells.
Discussion
This work demonstrates that reliable microarray data can be generated from rare primary cell types using as few as $2 \times 10^4$ cells. While the study provides a wealth of candidate genes, it was limited by the incomplete annotation of the murine genome at the time (many ESTs) and the need for functional validation of these candidates. Furthermore, future single-cell technologies may resolve additional heterogeneity within these defined stages. The primary dataset is available via GEO accession GSE13.
Summary updated