Notch3 marks clonogenic mammary luminal progenitor cells in vivo

Abstract

The identity of mammary stem and progenitor cells remains poorly understood, mainly as a result of the lack of robust markers. The Notch signaling pathway has been implicated in mammary gland development as well as in tumorigenesis in this tissue. Elevated expression of the Notch3 receptor has been correlated to the highly aggressive "triple negative" human breast cancer. However, the specific cells expressing this Notch paralogue in the mammary gland remain unknown. Using a conditionally inducible Notch3-CreERT2(SAT) transgenic mouse, we genetically marked Notch3-expressing cells throughout mammary gland development and followed their lineage in vivo. We demonstrate that Notch3 is expressed in a highly clonogenic and transiently quiescent luminal progenitor population that gives rise to a ductal lineage. These cells are capable of surviving multiple successive pregnancies, suggesting a capacity to self-renew. Our results also uncover a role for the Notch3 receptor in restricting the proliferation and consequent clonal expansion of these cells.

Figure 1.

Notch3 is specifically expressed in luminal cells at all developmental stages. (A–J) Notch3-CreERT2^SAT/R26^mTmG mice were induced with 4-OHT at different developmental stages and analyzed 24 h later. Notch3-expressing cells (marked in green) are labeled by the luminal marker CK8 and not by the myoepithelial marker CK5. Representative sections of ducts from prepubertal mice (3 wk of age; A and B), pubertal mice (6 wk of age; C–F), adult mice (9 wk of age; G and H), and pregnant mice (I and J). Bar, 50 µm.

Figure 2.

Notch3-expressing cells give rise to luminal lineages. Notch3-CreERT2^SAT/R26^mTmG mice were induced with a single dose of 4-OHT at different developmental stages, and the derived lineages were analyzed at different time points after induction. The basal layer is marked by CK5 staining in A–D and J–L, and nuclei are stained with DAPI. (A–C) Sections of glands of mice induced at 3 wk of age and analyzed after the following times: 24 h (A), 3 wk (B), and 6 wk (C). p.i., postinjection. D shows sections of glands from mice induced with 4-OHT as adults (9 wk of age) and analyzed after 3 mo. (E–H) GFP⁺ progeny of Notch3-expressing cells contained cells that were both positive and negative for the progesterone receptor (PR; E and F) and the estrogen receptor α (ER-α; G and H). (I) Quantification of the rate of expansion of GFP⁺ cells during puberty (4–8 wk of age). Mean (±SD) percentage of clones over the total number of clones counted per sample. Single cells, small clones of two to four cells, and large clones of more than five cells were counted. (J–L) Representative sections of mammary ducts from midpregnant mice induced at 3 wk of age (J), 4 wk after involution from mice induced at 6 wk of age after one pregnancy cycle (K), and after three consecutive pregnancies (L). Bars: (A–D and J–L) 50 µm; (E–H) 30 µm.

Figure 3.

Notch3-expressing cells present a high clonogenic potential. (A) Number of colonies/well (±SD) formed by GFP⁺ and GFP⁻ luminal cells. (B) Distribution of GFP⁺ cells (bottom plots) from 6-wk-old Notch3-CreERT2^SAT/R26^mTmG mice 24 h after induction (n = 3) among the three luminal subpopulations resolved by Sca1 and CD49b expression (top plots). The two left graphs represent controls without antibodies. The lines on the FACS plots correspond to the gates chosen to select the indicated cell populations: CD49b⁻/Sca1⁺, CD49b⁺/Sca1⁺, and CD49b⁺/Sca1⁻. (C) qRT-PCR showing the relative Notch3 RNA expression in Sca1⁺/CD49b⁻, Sca1⁺/CD49b⁺, and Sca1⁻/CD49b⁺ luminal fractions. Error bars represent the SDs of at least three independent experiments. (D) Number of colonies/well (±SD) formed by GFP⁺ and GFP⁻ cells in clonogenic (CD49b⁺) and nonclonogenic (CD49b⁻) luminal cell subpopulations. *, P ≤ 0.05; **, P ≤ 0.005. The p-values were calculated using the Student’s t test.

Figure 4.

GFP-expressing cells are mostly nonproliferative. (A–D) Notch3-CreERT2^SAT/R26^mTmG mice induced with 4-OHT at the indicated developmental stages (n = 3 for each developmental stage) and analyzed 24 h later for their expression of the proliferation marker Ki67. Single Notch3⁺ cells (in green) are not labeled by Ki67, with few exceptions in the terminal end buds (TEB; in D). The inset in D is a magnification of the area delineated by a dashed box. (E and F) Mice induced at 3 wk of age and immunostained with Ki67 after a 3-wk (E) or a 6-wk (F) chase present marked lineages that contain Ki67⁺ cells. (G) Mean (±SD) percentage of GFP⁺ cells stained for Ki67 over the total number of GFP⁺ cells counted. 98% of GFP⁺ cells are negative for Ki67 3 d after induction (A–D and G). (H) Quantification of the percentage of Ki67⁺ luminal cells in TEBs and ducts (n = 3). The percentage of Ki67⁺/GFP⁺ cells 24 h after induction is shown compared with the percentage of Ki67⁺ cells within the total luminal population. (I) Schematic representation of the BrdU incorporation experiments for both the short pulse and the long incorporation experiments. (J and K) Immunostaining with anti-BrdU antibodies shows that most GFP-expressing cells are negative for BrdU after a short pulse of BrdU (J), again with few exceptions in the TEBs (K). The progeny of Notch3-epressing cells (in green in L and M) contains both BrdU⁻ (L) and BrdU⁺ (M) clones after a long incorporation of BrdU. (N) Quantification of the percentage of BrdU⁺ total luminal or GFP⁺ cells in TEBs and ducts after a short BrdU pulse shows a striking reduction in the number of cycling GFP⁺ cells, as it is the case for Ki67 (H). (O) Relative expression to the 18S housekeeping gene of Ink4/Arf and p21/Waf1 in GFP⁻ and GFP⁺ luminal cells indicates elevated expression of cell cycle inhibitors in GFP⁺ cells. (N and O) Error bars represent the SDs of at least three independent experiments. *, P ≤ 0.05; **, P ≤ 0.005. The p-values were calculated using the Student’s t test. Bars: (A–C, E, F, and J–M) 40 µm; (D, main image) 100 µm; (D, inset) 25 µm.

Figure 5.

Notch3 activation retains cells in a nonproliferative state. (A) Schematic diagram of the N3IC targeting vector used for the generation of R26-N3IC^SAT knockin mice. The intracellular domain of the Notch3 receptor (N3IC) was knocked in to the ROSA26 locus followed by an internal ribosome entry site and a YFP gene. A lox-STOP-lox cassette blocks the expression of the transgene in the absence of Cre recombinase. (B) Quantification of clonal expansion upon constitutive Notch3 activation. Mean (±SD) percentage of clones (single cells, two to four cell clones, and clones bigger than five cells) over the total number of clones counted from Notch3-CreERT2^SAT/R26^mTmG/R26-N3IC^SAT mice (4wk8wk N3IC) compared with Notch3-CreERT2^SAT/R26^mTmG mice (4wk8wk). Mice were induced with a single dose of 4-OHT at 4 wk of age and analyzed at 8 wk of age. *, P ≤ 0.05. The p-values were calculated using the Student’s t test. (C) Genomic DNA PCR showing the unfloxed N3IC transgene (top) and the floxed N3IC allele (bottom, indicated by a white arrow). The floxed allele is amplified only upon 4-OHT induction. Lanes: (1) 1-kb molecular weight marker; (2) N3IC mouse1 no 4-OHT; (3) N3IC mouse1 + 4-OHT; (4) N3IC mouse2 no 4-OHT; (5) N3IC mouse2 + 4-OHT; and (6) nontransgenic control mouse. (D) qRT-PCR with specific oligos recognizing the extracellular (N3EC) or intracellular (N3IC) domain of Notch3 indicates sustained N3IC expression in GFP⁺ cells sorted from Notch3-CreERT2^SAT/R26^mTmG/R26-N3IC^SAT mice 4 wk after induction. Note that the levels of endogenous Notch3 (detected with the N3EC oligos) are unchanged between GFP⁺ and GFP⁻ cells because the analysis has been performed on the progeny of Notch3-expressing cells 4 wk after induction, containing both Notch3⁺ and Notch3⁻ cells. (E) qRT-PCR analysis of sorted GFP⁻ and GFP⁺ cells from Notch3-CreERT2^SAT/R26^mTmG/R26N3IC^SAT mice 4 wk after induction shows elevated expression levels of Arf and p21 in N3IC-expressing GFP⁺ cells. (D and E) Error bars represent the SD of at least three independent experiments. (F) Proposed model for the observed behavior of GFP-labeled cells. A luminal cell expressing high levels of Notch3 receptor (N3^high) is kept in a resting nonproliferative state until it down-regulates Notch3 expression (N3^low/−), allowing it to enter the cell cycle and give rise to a luminal lineage. The assumption is that only cells expressing high levels of Notch3 (N3^high) will be marked by Cre recombination, whereas proliferative N3^low/− cells will not be labeled by GFP in this system.