Contrasting Quiescent G0 Phase with Mitotic Cell Cycling in the Mouse Immune System

A transgenic mouse line expressing Fucci (fluorescent ubiquitination-based cell-cycle indicator) probes allows us to monitor the cell cycle in the hematopoietic system. Two populations with high and low intensities of Fucci signals for Cdt1(30/120) accumulation were identified by FACS analysis, and these correspond to quiescent G0 and cycling G1 cells, respectively. We observed the transition of immune cells between quiescent and proliferative phases in lymphoid organs during differentiation and immune responses.


Introduction
In addition to the four conventional phases of the cell cycle (G 1 , S, G 2 , and M), there is a fifth phase, G 0 , which denotes the nonproliferating or quiescent state of cells that have withdrawn from the active cell cycle [1,2]. At a certain point during G 1 , a cell decides whether it will remain in G 1 or retreat from the active cell cycle into G 0 .
Many cells in the adult animal body stay in G 0 . However, the regulation of the G 1 /G 0 transition varies among different cell types. Whereas terminally differentiated cells, such as neurons and muscle cells, rarely divide, most lymphocytes are assumed to withdraw from and reenter the cell cycle repeatedly throughout their lifetime. We thus planned to study dynamic transition between quiescence and proliferation of lymphocytes using Fucci transgenic mice. Although the line #596/#504 has been useful for studying relationships between cell-cycle progression and morphogenesis in many organs, we noticed that neither mKO2-hCdt1(30/120) nor mAG-hGem(1/110) was expressed in the hematopoietic system of this line. Thus, we screened a pool of Fucci transgenic mouse lines constructed with the CAG promoter, and found that #639 and #474 exhibit hematopoietic gene expression of mKO2-hCdt1(30/120) and mAG-hGem(1/110), respectively. We then investigated Fucci signals in immune cells from these two lines, which are hereafter referred to as FucciG 1 -#639 and FucciS/G 2 /M-#474.

Ethics Statement
The experimental procedures and housing conditions for animals were approved by the Animal Experimental Committees at the Institutes of Physical and Chemical Research (RIKEN) -Research Center for Allergy and Immunology (RCAI) and -Brain Science Institute (BSI), and Kyoto University school of medicine, and all animals were cared for and treated humanely in accordance with the Institutional Guidelines for Experiments using Animals.

Cell Culture and Imaging
NMuMG/Fucci cells were grown in DMEM (high glucose) supplemented with 10% fetal bovine serum (FBS), penicillin/ streptomycin, and 10 mg/ml insulin (Sigma: I0516). Cells were fixed with 1% PFA for 1 hour at room temperature and then with 70% ethanol for overnight. This procedure was sufficient for effective fixation while avoiding the quencing of fluorescent proteins. After being washed, cells were stained with Alexa Fluor 647-conjugated anti-Ki-67 monoclonal antibody (mAb) (BD Pharmingen) and DAPI, then analyzed using a FACSAria II (BD Biosciences). Data were analyzed using FlowJo software (Tree star). Time-lapse imaging and data analysis were performed as described previously [3].

Flow Cytometry Analysis
Antibodies used in this study were purchased from BD Pharmingen, eBioScience, or BioLegend. After being washed with Dulbecco's phosphate-buffered saline (PBS) containing 2% fetal calf serum (FCS) and 0.02% sodium azide (staining buffer), cells were treated with culture supernatant from the 2.4G2 hybridoma for blocking Fc binding, and subsequently with appropriate fluorochrome-conjugated antibodies. For T and B cell gating, splenocytes were stanined with allophycocyanin (APC)-conjuagted anti-CD3 mAb and APC-conjuagted anti-CD19 mAb, respectively. Thymocytes were stained with APC-conjuagted anti-CD3 mAb, APC-Cy7-conjuagted anti-CD8 mAb, and Pacific blueconjugated anti-CD4 mAb. For gating differentiating B cell populations, bone marrow (BM) cells were stained with PE-Cy7or APC-conjugated anti-IgM mAb, biotinylated anti-B220, and Pacific blue-conjugated anti-CD43 mAb, followed by Pacific orange-conjugated streptavidin (Invitrogen). For detection of Ki-67, cells were treated with anti-surface antigen antibodies and fixed with 1% PFA at 4u C for 15 min, followed by 75% ethanol at 220uC for 2 hours, then stained with Alexa 647-conjugated anti-Ki-67 mAb or Alexa 647-conjugated control mAb. Stained cells were analyzed using a FACS Calibur (BD Biosciences) and a Fortessa (BD Biosciences). mAG and mKO2 signals were detected via the FITC and PE channels, respectively. Flow cytometry data were analyzed using the FlowJo software (Tree Star, Inc.).

Results and Discussion
To examine quiescence-associated cell-cycle dynamics in cultured cells, we used Fucci-expressing, stably transformed normal murine mammary gland (NMuMG) cells (NMuMG/Fucci cells), which show confluence-induced proliferation arrest [3,4]. We analyzed their Fucci signals and Ki-67 immunosignals when cells were exponentially growing (Fig. 1a-c) and after they reached confluency (Fig. 1d-f). We gated 5 populations accroding to the intensities of mKO2 and mAG (Fig. 1a, d). Gates #2 through #5 in growing condition showed Ki-67 positive signals and should represent the mitotic cell cycling phases. Gate #1 contained cells that showed an order of magnitude lower Ki-67 signal than the other gated cells. Remarkably, nearly all of the cells at confluency were collected in gate #1. These cells contained bright red (mKO2 ++ /mAG 2 ) (Fig. 1d), diploid (Fig. 1e), and Ki-67 negative (Fig. 1f) cells, which were supposed to stay in quiescent G 0 phase.
After all cells showed bright red nuclei (mKO2 ++ /mAG 2 , quiescent G 0 phase), a wound was introduced mechanically; cells at the edge of the wound turned green after a latency period of 12-20 hr (Fig. 1g, and Movie S1), reminiscent of the time (8 hr) required for NIH 3T3 cells to reenter the cell cycle from the G 0 state after the onset of proliferation stimuli [6]. The latency time we observed should consist of the 8 hr required to re-enter the cell cycle, plus the time required for cells to proceed through the remainder of G 1 . One cell (indicated by an arrowhead in Fig. 1g) was tracked for 40 hr (Fig. 1h). Temporal profiles of the intensities of red and green signals in this cell are shown in Fig. 1i. It was apparent that the red fluorescence in the G 0 state was several times stronger than in the cycling G 1 state. The differential intensity of red fluorescence between quiescent and proliferating cells was observed previously in the developing cerebral cortex of #596/#504 mice; postmitotic neurons in the cortical plate exhibited much brighter red fluorescence than mitotic neural progenitors in the ventricular zone, presumably due to accumulation of mKO2-hCdt1(30/120) after cell-cycle exit [3,4].
We aimed to analyze both quiescence and proliferation of lymphocytes using Fucci transgenic mice. We crossbred FucciG 1 -#639 and FucciS/G 2 /M-#474 to generate double transgenic mice (#639/#474 mice) for analyzing Fucci-signals of immune cells in various lymphoid organs (Fig. 3). Bone marrow (BM) and thymus, as well as secondary lymphoid organs, such as lymph node (LN), spleen, and Peyer's patch (PP), were isolated and analyzed with regard to mKO2 and mAG fluorescence intensities. Most lymphocytes in spleen, LN, and PP exhibited very strong mKO2 signal but no mAG signal, which should represent the mKO2 ++ / mAG 2 population (gate #1, G 0 population). By contrast, cells in thymus and BM showed substantial heterogeneity; these cells can also be grouped into five types as illustrated in Fig. 2c. While NMuMG/Fucci cells show a high clonality, lymphocytes from #639/#474 mice are heterogenous in terms of gene expression of both mKO2-hCdt1(30/120) and mAG-hGem(1/110). Thus, it is noted that the lymphocytes collected in gate #3 should contain the ones that failed to synthesize either mKO2-hCdt1(30/120) or mAG-hGem(1/110).
We then examined cell-cycle re-entry of lymphocytes in intact lymph nodes (LNs) of #639/#474 mice by two-photon excitation microscopy. On the one hand, intact inguinal LNs of #639/#474 mice were observed to be mostly filled with lymphocytes that contained bright red nuclei (Fig. 5a-c), confirming that most lymphocytes in an LN remain in G 0 (Fig. 3). On the other hand, after the rear footpad of a #639/#474 mouse was injected subcutaneously with CFA/TNP-KLH, a considerable number of cells with green nuclei emerged in the popliteal draining LN (Fig. 5d-f and Movie S2, S3).
Heterogeneity of Fucci signals in thymocytes (Fig. 3, 6a) inspired us to analyze cell-cycle status during their differentiation in thymus. Immature CD4 CD8 double negative (DN) thymocytes differentiate to CD4 CD8 double positive (DP) thymocytes via CD8 + CD3 low (DN to DP transition cells) cells. Then the DP thymocytes differentiate to CD4 SP thymocytes or CD8 SP thymocytes [7][8][9]. As the signal for CD3 is upregulated along with the thymocyte differentiation, the mature CD8 SP thymocytes were characterized as CD8 + CD3 high and distinguished from the DN to DP transition cells (CD8 + CD3 low ) cells. All these thymocyte populations were prepared and analyzed for the Fucci signals (Fig. 6b). While the most immature DN thymocytes were distributed evenly among the 5 gates, the terminally differentiated thymocytes were collected principally in gate #1. Interestingly, no accumulation of mKO2 signal was observed in DN to DP transition thymocytes, which may suggest very rapid cell-cycling of the immature thymocytes.
Unlike neurons and skeletal muscle cells that exist in a terminally differentiated G 0 state, most lymphocytes withdraw from the cell cycle only transiently and reenter the cycle in response to various stimuli. It is important to directly monitor the dynamic transition between cycling G 1 and quiescent G 0 cells. The Fucci technique is based on the cell-cycle2dependent degradation of proteins; thus, the signal intensity may be proportional to the length of the respective cell-cycle phase. There is evidence that transcription driven by the CAG promoter occurs in Ki-67negative cells [13], which are assumed to be in G 0 [5,14]. Thus, mKO2-hCdt1(30/120) may gradually accumulate in G 0 cells. In this study, we demonstrate that the intensity profile of the G 1 (G 0 ) marker signal (the mKO2 fluorescence) exhibits two peaks: mKO2 ++ and mKO2 + , which reflect quiescent G 0 cells and cycling G 1 cells, respectively. Here, we propose that FucciG 1 -#639 and/or FucciS/G 2 /M-#474 transgenic mice can provide reliable readouts of the cell-cycle regulation of lymphocytes, both in vitro and in vivo.

Supporting Information
Movie S1 Time-lapse imaging of Fucci-expressing NMuMG cells response to wound. Fucci-expressing NMuMG cells were grown on a glass-bottom dish to reach the state of confluent (contact inhibition). One hour after a scratch of the monolayer cells, time-lapse imaging was performed using an LCV100 (Olympus). Images were acquired every 19 minutes. Movie was processed every 114 minutes for size reduction. Total imaging time = 85 hours. Playback speed is 38,2506real time.

(AVI)
Movie S2 Time-lapse observation of cells with mAG + and mKO2 + nuclei in a draining LN in a #639/#474 mouse. Movie was processed from the same observation area of Fig. 2e. A region was time-lapse imaged with the z step size of 5 mm every 30 sec for 30 min. Z stacked images (10 mm thick) are shown in this movie.

(MOV)
Movie S3 Time-lapse observation of cells with mAG + and mKO2 + nuclei in a draining LN in a #639/#474 mouse. Movie was processed from the same observation area of Fig. 2f. A region was time-lapse imaged with the z step size of 5 mm every 30 sec for 30 min. Z stacked images (10 mm thick) are shown in this movie. (MOV)