(A) Immunofluorescent co-staining for NG2 and mitosis marker phospho-histone3 (PH3) showing proliferating NG2⁺ cells in the corpus callosum (CC) of wild type P60 mice.
(B) NG2⁺ cells co-express OPC markers PDGFRα and OLIG2.
(C) Immunofluorescent co-staining for NG2 (asterisks) and O4 expression (white arrows).
(D) Quantification of OPC marker expression of NG2⁺ cells in the adult CC. NG2⁺ cells are PDGFRα/OLIG2/NG2 triple-positive (Rα/OLIG2/NG2; 65%); Rα/NG2 (23%), NG2/OLIG2 (8%) and NG2/O4 (4%) double-positive. 2% of NG2⁺ cells are OPC marker-negative.
(E) Schematic of a hemisphere of a coronal brain section. NG2⁺ cells were acutely isolated from CC tissue and also obtained from subcallosal zone (SCZ)-derived neurospheres.
(F) Immunofluorescent staining for NG2, PDGFRα and OLIG2 on acutely isolated NG2⁺ cells.
(G) BrdU was added to FACS sorted, acutely isolated NG2⁺ OPC cells after they have grown in culture for 48 hrs, and flow cytometry was used to determine the percentage of NG2 and BrdU double-positive (NG2⁺BrdU⁺) cells.
(H) NG2⁺ cells from p1-p3 were immunostained for NG2 and O4 after ten days in differentiation medium.
(I) Percentage of O4⁺ cells remained 23-26% for three passages. Scale bars in (A) and (F) 10μM, (B) 50 μM, (C), (E) and (H) 100μM. Error bars are +/− standard error of the mean (SEM). See also Figure S1.

(A) Schematic of experimental approach to determine differences in proliferative cell fate. Acutely isolated NG2⁺ OPC divide into NG2⁺ and NG2⁻ cells in proliferative medium for 48 hr (division 1). Cells were then treated with BrdU to label those re-entering division within the next 12 hr (division 2). BrdU⁺ cells in the NG2⁺ and NG2⁻ progeny were quantified by flow cytometry (B, C) and in pair assays (D, E).
(B) A representative flow cytometry plot showing overlap between NG2 expression and BrdU incorporation. Grey line indicates border set between BrdU⁻ and BrdU⁺ cells.
(C) Percentage of BrdU⁺ cells in the NG2⁺ and NG2⁻ progeny.
(D) A representative cell pair with NG2 asymmetry immunostained for NG2 and BrdU.
(E) Percentage of BrdU staining in cell pairs with NG2 asymmetry.
(F) Schematic of experimental approach to determine differentiating fate. Freshly isolated OPC kept under proliferative conditions for 48 hr (division 1) were re-sorted into NG2⁺ and NG2⁻cells, switched to differentiation medium (differentiation) and percentage of O4 expression was assessed by flow cytometry 24 hr after sorting (G, H) and pair assays 48 hrs after plating (I, J).
(G) A representative flow cytometry plot for O4 expression in the NG2⁻ (Left Plot) and NG2⁺ (Right Plot) population.
(H) Percentage of O4⁺ cells as determined by flow cytometry.
(I) A representative cell pair with NG2 asymmetry and co-immunostained for O4 and NG2.
(J) Percentage of O4 expression in cell pairs with NG2 asymmetry. Error bars are +/− standard error of the mean (SEM). Scale bars in (D) and (I) 10 μm. See also Figure S3.

(A) A p53^+/− OPC cell pair is stained for NG2 and EGFR.
(B) A verbB p53^+/− OPC cell pair is stained for NG2 and EGFR.
(C) Percentage of NG2 and EGFR asymmetry in control and verbB p53^+/− OPC following their acute isolation from CC tissue.
(D) An OPC cell pair in the CC of verbB p53^+/− mice is stained for NG2, PH3 and BrdU.
(E) Percentages of NG2 asymmetry are 23% in acutely isolated (ex vivo) verbB p53^+/− cells and 13% in verbB p53^+/− tissue (in vivo), compared to 41% in acutely isolated control, p53^+/− OPC and 42% in control, p53^+/− CC tissue.
(F) Percentage of NG2⁺BrdU⁺ cells in verbB p53^+/− OPC and control cells as determined by flow cytometry. 47% and 76% of verbB p53^+/− cells at passage 1 (p1) and passage 3 (p3), respectively, are NG2⁺BrdU⁺.
(G) OPC numbers, as determined by quantifying triple-positive (NG2/Rα/OLIG2) cells, in the verbB p53^+/− and the control CC tissues.
(H) Control and (I) verbB p53^+/− OPC plated for differentiation and stained for NG2 and O4. (J) Percentage of NG2⁺ cells differentiated into O4⁺ cells 10 d after differentiation. O4 staining non-overlapping with NG2 in control cells (white arrow in H). In contrast, verbB p53^+/− O4⁺ cells with overlapping NG2 staining (yellow staining; white arrow in I). O4 and NG2 double-positive cells were counted as O4⁺ cells.
Pair assays (A, B), self-renewal (F) and differentiation (H, I) experiments were done with acutely isolated NG2⁺ cells. Controls are p53^+/−. Error bars are +/− standard error of the mean (SEM). **Student's t-test p<0.01. Scale bar in (A) and (B) 10 μM, in (D) is 50 μM and in (F) is 100 μM. See also Figure S5.

(A) Summary of human oligodendrogliomas used in the study.
(B) NG2⁺ cells were acutely isolated from surgically removed tumor and non-neoplastic brain tissue specimens and percentage of cell pairs with NG2 asymmetry is determined using pair assay. Error bars are +/− standard error of the mean (SEM).
(C, D) Representative NG2⁺ cell pairs from tumor (C) and non-neoplastic tissue (D) stained for NG2 and EGFR. Scale bars are 10 μM.
(E) Box plots representing changes in gene expression levels at a p-value < 0.05 of 7 conserved, candidate asymmetry regulator genes in control brain tissue (NT) versus low-grade oligodendroglioma tissue (ODGII). The band in each box represents the median and the bottom and top the 25 and 75 percentile, respectively. The errors bars are the minimum and maximum of data. See also Table S1 and Figure S7.

(A) Schematic of pair assay. Single cells plated under proliferative conditions enter mitosis and form cell pairs, which are fixed and immunostained. Cell pairs in which proteins partition to one daughter cell are considered asymmetrical.
(B) A representative, neurosphere-derived mitotic cell pair with NG2 asymmetry.
(C) A representative acutely isolated cell pair with NG2 asymmetry.
(D) Cell pair treated with latrunculin A shows cytoplasmic (white arrow) rather then polarized NG2 localization.
(E) Percentage of cell pairs with NG2 asymmetry after latrunculin A and DMSO (control) treatment. **p=0.0087 by Student's t test.
(F) Immunofluorescent co-staining for NG2 and gamma (γ)–Tubulin (γ Tub) to label centrosomes. 1 asterisk = 1 centrosome. White arrow points to patchy NG2 staining in G1/S phase.
(G) Immunofluorescence co-staining for NG2 and alpha (α)-Tubulin (α Tub) to label the mitotic spindle.
(H) Representative cell pair in vivo with NG2 asymmetry after immunofluorescent co-staining for NG2, BrdU and PH3. P60 mice were injected with BrdU prior to sacrifice. White arrows point to NG2⁺ cell in pair.
(I) Percentage of cell pairs with NG2 asymmetry from acutely isolated cells (ex vivo) and CC tissue (in vivo).
Scale bars in (C), (D), (F), (G) and (H) 10 μm. Error bars are +/− standard error of the mean (SEM). (D; F; G) were performed with neurosphere-derived cells. See also Figure S2.

(A) Immunofluorescent co-staining for EGFR and a-Tubulin of neurosphere-derived cells at distinct stages of mitosis.
(B) A representative NG2⁺ OPC cell pair co-immunostained for NG2 and EGFR (white arrows).
(C) A representative NG2⁺ OPC cell pair co-immunostained for NG2 and p-EGFR (white arrow). EGF is present in medium.
(D) A representative NG2KO OPC cell pair co-immunostained for PDGFRα⁺ (Rα) and EGFR showing symmetric EGFR.
(E) Percentage of cell pairs with asymmetric (asym) and symmetric (sym) EGFR.
(F) A representative NG2KO OPC cell pair co-immunostained for PDGFRα⁺ (Rα) and p-EGFR showing symmetric, faint p-EGFR signal.
(G) Percentage of cell pairs with asymmetric p-EGFR staining.
(H) EGF-starved neurosphere cells were treated with EGF for 48 hrs and cell numbers were determined.
(I) Percentage of PDGFRα and BrdU double positive (Rα⁺BrdU⁺) cells with (w/EGF) and without (w/o EGF) EGF.
(J) Percentage of PDGFRα and PH3 double-positive (RαPH3⁺) cells. OPC were acutely isolated from the CC of P60 NG2WT and NG2KO mice using PDGFRα as selection marker with the exception of (A, H). There, OPC were isolated from low passage subcallosal zone-derived NG2WT and NG2KO neurospheres. Error bars are +/− standard error of the mean (SEM). **Students t-test p<0.01. *Student's t-test p<0.05. Scale bars in (C-D) and (F) 10 μM. See also Figure S4.

(A) Kaplan-Meier survival curves of mice injected with 100,000 p53^+/−, verb p53^+/− and tumor-derived NG2⁺ cells.
(B, C) Hematoxylin & Eosin (H&E) staining of orthotopic tumors from verb p53^+/− NG2⁺ cells. The white arrow in panel B indicates infiltrative tumor cells and the white arrow in panel C indicates a cell with the “fried egg” appearance, a feature typical for oligodendroglial tumors.
(D, E) Immunofluorescent staining of orthotopic tumors for NG2 and GFAP (D) and for OLIG2 and Nestin (E).
(F) NG2 staining of orthotopic tumor at higher magnification. The white arrow points to a mitotic cell pair identified by nuclear staining with symmetric NG2.
(G) Percentage of cell pairs with symmetric (white area) and asymmetric (black area) NG2. NG2 asymmetry is 23% in acutely isolated verbB p53^+/− cells and is 18% in cells isolated from orthotopic (tumor/o) and 10% in cells isolated from endogenous tumors (tumor/e).
Error bars are +/− standard error of the mean (SEM). Scale bars in (B), (D), (E) are 100 μM, in (C) 200 μM and in (F) 10 μM. Student's t-test, *p<0.05. See also Figure S6.