CD133 (Prominin) negative human neural stem cells are clonogenic and tripotent

Abstract

Background: CD133 (Prominin) is widely used as a marker for the identification and isolation of neural precursor cells from normal brain or tumor tissue. However, the assumption that CD133 is expressed constitutively in neural precursor cells has not been examined.

Methodology/principal findings: In this study, we demonstrate that CD133 and a second marker CD15 are expressed heterogeneously in uniformly undifferentiated human neural stem (NS) cell cultures. After fractionation by flow cytometry, clonogenic tripotent cells are found in populations negative or positive for either marker. We further show that CD133 is down-regulated at the mRNA level in cells lacking CD133 immunoreactivity. Cell cycle profiling reveals that CD133 negative cells largely reside in G1/G0, while CD133 positive cells are predominantly in S, G2, or M phase. A similar pattern is apparent in mouse NS cell lines. Compared to mouse NS cells, however, human NS cell cultures harbour an increased proportion of CD133 negative cells and display a longer doubling time. This may in part reflect a sub-population of slow- or non-cycling cells amongst human NS cells because we find that around 5% of cells do not take up BrdU over a 14-day labelling period. Non-proliferating NS cells remain undifferentiated and at least some of them are capable of re-entry into the cell cycle and subsequent continuous expansion.

Conclusions: The finding that a significant fraction of clonogenic neural stem cells lack the established markers CD133 and CD15, and that some of these cells may be dormant or slow-cycling, has implications for approaches to identify and isolate neural stem cells and brain cancer stem cells. Our data also suggest the possibility that CD133 may be specifically down-regulated during G0/G1, and this should be considered when this marker is used to identify and isolate other tissue and cancer stem cells.

Figure 1. Human NS cells express CD133 and CD15 heterogeneously.

Human NS cells homogeneously express the neural precursor marker Nestin (A and B, red), but exhibit heterogeneous CD133 (A, green) and CD15 (A, green) expression. Live cell staining (C) and subsequent flow cytometry analysis (D) reveal four sub-populations of human NS cells: CD133+/CD15⁻, CD133+/CD15+, CD133⁻/CD15+, and CD133⁻/CD15⁻. The proportions of each cell population are listed in Table 1.

Figure 2. CD133 and CD15 expression varies in human NS cells but is not linked to stem cell potency.

Human NS cells were flow sorted into four populations using the gates illustrated in (A) and then cultured in NS cell expansion conditions. One week later, each sorted cell population exhibited heterogeneous CD133 and CD15 expression (B). Four weeks later, all four cell populations displayed near-indistinguishable CD133 and CD15 distributions (B). Clonal cultures could also be derived from all four purified cell populations after cell sorting. The cloned cells retained tripotent, being able to generate neurons (Tuj1+) (C), astrocytes (high level GFAP with flattened morphology) (D), and oligodendrocytes (O4+) (E).

Figure 3. CD133 expression is regulated at the mRNA level and is reduced in G0/G1 phase.

(A) RT-PCR (Aa) and real-time PCR (Ab) indicate that the expression of CD133 mRNA in CD133^−/lo human NS cells is approximately 30 fold lower in CD133^+/hi cells. CD133^−/lo and CD133^+/hi NS cells were purified by cell sorting using the gate illustrated in Ba below. (B) Cell cycle analysis of human NS cells stained with Hoechst 33342, anti-CD133-APC, and anti-CD15-FITC. Analysis gates were set as illustrated in Ba and Be. The CD133^+/hi, CD133^−/lo, CD15^+/hi, and CD15^−/lo cells are colored red, green, pink, and blue respectively (Ba-h). Ungated cells are colored black. The cell cycle profile of the entire population is illustrated by dashed lines in Bb and Bf. The majority of CD133^−/lo (green) cells were in G1/G0 phase of the cell cycle (Bb and Bc), whereas over half of the CD133^+/hi (red) cells were in S, G2, or M phase (Bb and Bd). Although human NS cells express CD15 heterogeneously (Be), the CD15+^/hi (pink) and CD15^−/lo (blue) cell populations exhibited similar cell cycle profiles, consistent with the whole cell population (Bf-Bh). (C) To test the indication that CD133 may be preferentially down-regulated in cells in G0/G1 phase, we purified CD133^+/hi and CD133^−/lo cells using the gates illustrated in Ca and applied PI staining after fixation. Flow cytometry analysis indicated that the majority of CD133^−/lo cells were in G1/G0 (Cb), whereas over half of CD133^+/hi cells had >2N DNA content (Cc).

Figure 4. Mouse NS cells express CD133 heterogeneously.

Live mouse NS cells exhibit heterogeneous CD133 expression (A). Flow cytometry indicated that 50% of CD133^−/lo cells (green) were in G1/G0 phase, whereas over half (51%) of the CD133+^/hi cells (red) were in S, G2, or M phase (B, C).

Figure 5. Human NS cell cultures propagate more slowly than mouse counterparts.

Compared to mouse NS cell cultures, human NS cell cultures exhibit a lower percentage of Ki67 expressing cells under the same expansion conditions (A). Over 99% of mouse NS cells incorporated BrdU after 5 days incubation (B, C). Approximately 5% of human NS cells remained BrdU negative after prolonged incubation (B, D). Yellow arrows in Figure 5D indicate BrdU negative human NS cells after 10 days incubation.

Figure 6. Human NS cell cultures harbour slow-cycling or dormant cells.

Approximately 5% of human NS cells remained viable after 5 days exposure to the antimitotic drug Ara-c. The viable cells retained Nestin and Sox2 expression but did not express Ki67 or CD133 (A, B). When these cells were re-plated into medium without Ara-c for 10 days, approximately 14.6% of cells expressed Ki67 and 8.7% expressed CD133 (C, D). Yellow arrows in (C) indicate Ki67 positive cells.