2.1.

Identification of the Intrinsic Biomarkers in Human Embryonic Stem-Cell Colonies

We separate and identify different intrinsic fluorescent metabolites in the hESC colonies by detecting clusters with different FLIM signatures within the phasor plot. After label-free FLIM imaging we perform retrospective in vitro staining or immunostaining after fixation to colocalize the FLIM signature of intrinsic fluorescent biomarkers with specific hESC compartments.

Figure 1 shows a representative image of the autofluorescence from a colony of undifferentiated H9 hESCs plated on a mouse-embryonic-fibroblast (MEF) feeder substrate. We perform the phasor transformation of the FLIM image of the hESC colony (see Sec. 4 and Ref. 32). Every pixel of the FLIM image is transformed into a pixel in the phasor plot. Figure 1(a) displays the phasor histogram distribution of the FLIM image of an undifferentiated hESC colony that is located inside the universal circle of the phasor plot, indicating the multiexponential characteristic of its decay.³² Within the phasor distribution we can identify three main clusters that correspond to different cell types and cell compartments [Fig. 1(a) and 1(d)]. hESC nucleus and cytoplasm, bright granules within the hESCs and the MEFs are selected by the green, red and blue cluster respectively [Fig. 1(a) and 1(d)]. After label-free FLIM imaging, we fixed the cells and we performed retrospective immunostaining imaging (see Sec. 4) for the phenotypic identification of undifferentiated hESCs and MEFs. Undifferentiated hESCs are identified by the expression of the pluripotency transcription factor OCT4 [Fig. 1(e)–1(g), while MEFs are identified by DAPI staining and the absence of OCT4 expression [arrows in Fig. 1(f) and 1(g)]. The specific bright granules within the hESCs are identified by the red cluster in Fig. 1(a) and are homogeneous in dimensions and typically have a diameter of $\sim 1\mu \mathrm{m}$. [Fig. 2(a) and 2(b)]. They are characterized by a long lifetime distribution [Figs. 1(a)–1(d) and 2(c)] very close to the universal circle (defined in the Appendix) and by a single exponential of about 10 ns. Although the hESC granules have a very similar lifetime to protoporphyrin IX,³² we exclude its presence because the hESC granules emission spectrum [Fig. 2(d)] is very different with respect to protoporphyrin IX spectrum that has a peak at 630 nm.⁴⁹ In fact the emission spectrum of the hESC granules is very broad and it has a peak at 500 nm [Fig. 2(d)]. The lipid-droplet-associated-granules (LDAG) emission spectrum is similar to the spectrum of flavins. However we exclude presence of flavins both in the granules and in the rest of the hESCs, because we do not detect any fluorescence at 800-, 850-, and 900-nm excitations (Fig. 3).

Fig. 1

Label-free identification of hESCs: (a) FLIM phasor histogram of the FLIM image excited at 760 nm from one H9 hESC colony cocultured with mouse embryonic fibroblasts (MEFs). The color scale (from blue to purple) corresponds to the 64 levels of the contours that indicate the percent occurrence in the phasor histogram of the pixels of the image. Different clusters within the phasor distribution correspond to bright lipid droplet-associated granules within hESCs (red), the hESCs themselves (green) and the MEF feeders (blue). Transmission image (a) and two-photon fluorescence intensity image (c) of a undifferentiated hESC colony grown on MEF feeders. (d) Phasor color map. Pixels of different colors correspond to the color of the cluster in the phasor plot A. Arrow indicates a MEF. (e) Expression of the pluripotency marker OCT4 in the same colony of B after cell fixation and immunostaining. (f) DAPI staining. (g) Merge of DAPI and OCT4 staining. Arrow indicates a MEF whose nucleus does not express OCT4.

Fig. 2

Spectroscopic characteristics of granules in hESCs: two-photon fluorescence intensity image (a) and transmission image (b) of a single undifferentiated H9 hESCs colony. (c) FLIM phasor plot of the FLIM image excited at 760 nm of the hESCs colony area in (a). (d) Emission spectrum from hESCs granules in (a).

Fig. 3

Transmission image (a) and two-photon fluorescence intensity images of an undifferentiated H9 hESC colony area excited at 720 nm (b), 760 nm (c), 800 (d), 850 (e) and 900 (f)

Figure 4 shows the colocalization of hESC granules with lipid droplets (LDs) labeled in vitro with 4,4-difluoro-1,3,5,7,8-pentamethyl-4-bora-3a,4a-diaza-s-indacene (BODIPY493/503). Although granules with similar morphology and which positively stain for lipids have been identifies as lipofuscin in several tissue and cells,^28,50 we exclude that the LDAG we observe could be related to lipofuscin. LDAG absorption spectrum (Fig. 3) is not as wide as the one of the lipofuscin which strongly fluoresces under excitation ranging from 360 to 660 nm.^28,50 The lifetime of the granules that we observe [Fig. 2(c)] is much longer than the lifetime of the lipofuscin reported in the literature.^28,51–53 Moreover the LDAG do not colocalize with the lysosomes (Fig. 5), which are known to be the structures involved in the storage of lipofuscin^54,55 and which have been previously colocalized with lipofuscin.^30,52,56

Fig. 4

Colocalization of lipid droplets with granules in hESCs: (a) FLIM phasor plot of the FLIM image excited at 760 nm of a single undifferentiated H9 hESC colony area. The red cluster in the phasor plot specifically selects some bright granules within the hESCs. Transmission image (b) and two-photon fluorescence intensity image (c) of the hESCs colony area. (d) Phasor color map. Red pixels have a specific lifetime signature that is selected by the red cluster in the phasor plot A. (e) In vitro staining of lipid droplets with Bodipy 493/503 shows colocalization with the hESCs granules identified by FLIM in D and A.

Fig. 5

Nonlocalization of hESC granules with lysosomes: transmission image (a) and two-photon fluorescence-intensity image (b) of an undifferentiated H9 hESC colony area. (c) Phasor color map. hESC granules are highlighted by the red pixels which correspond to the red cluster in the phasor plot D. (d) In vitro FLIM phasor plot of the undifferentiated hESC colony area. The red cluster in the phasor plot specifically selects hESC granules. (e) In vitro staining of lysosomes with Lysotracker.

We hypothesize that the formation of LDAGs in hESC could be generated by the combination of high ROS level, associated to high oxidative-phosphorylation rate⁴⁶ and the abundance of unsaturated metabolic precursors, such as arachidonic acid, that are highly reactive under oxidative conditions.³⁴ Linoleic acid and arachidonic acid peroxides can form fluorescent products from the interaction with polypeptides and free amino acid groups of proteins.^57,58 LDAGs [Fig. 4(d) and 4(e)] are surrounded by endoplasmic reticulum membranes (Fig. 6). We tested and excluded colocalization between these LDAGs and CONNEXIN-43 (Fig. 7 of the Appendix) and Alkaline phosphatase (Fig. 8 of the Appendix).

Fig. 6

hESC granules are surrounded by endoplasmic reticulum: (a) transmission image of an undifferentiated H9 hESC colony area. (b) In vitro staining of endoplasmic reticulum with ER-Tracker Red. (c) Merge of transmission image and ER staining.

Fig. 7

Nonlocalization of hESC granules with CONNEXIN-43: (a) In vitro FLIM phasor plot of an undifferentiated H9 hESC colony. The red cluster selects the bright granules within hESCs. Transmission image (b) and two-photon fluorescence intensity image (c) of the undifferentiated hESC colony. (d) Phasor color map. hESC granules are highlighted by the red pixels that correspond to the red cluster in the phasor plot A. (e) DAPI staining. Expression of the pluripotency marker OCT4 (f) and CONNEXIN-43 (g) after cell fixation and immunostaining.

Fig. 8

Nonlocalization of hESC granules with alkaline phosphatase: transmission image (a) and two-photon fluorescence intensity image (b) of an undifferentiated H9 hESC colony area. (c) In vitro FLIM phasor plot of an undifferentiated hESC colony area. The red cluster specifically selects hESC granules. (d) Phasor color map. hESC granules are highlighted by the red pixels that correspond to the red cluster in the phasor plot C. (e) Alkaline phosphatase staining performed after cell fixation.

To identify the contribution of NADH to the FLIM phasor signature of hESCs, we labeled the mitochondria with tetramethylrhodamine ethyl ester perchlorate (TMRE), a dye-marking active mitochondria (Fig. 9). Our data show that the mitochondria do not specifically localize with the LDAGs, which are indicated by the red clusters; conversely, mitochondria are primarily localized to the blue cluster areas in hESCs at the center of the phasor plot [Fig. 9(a)]. This location corresponds to the typical autofluorescence arising from the NADH within cells and tissues.³²) To confirm the contribution of NADH to the autofluorescence detected in the center of the phasor plot, we treat the hESCs with potassium cyanide (KCN) to block oxidative phosphorylation during cellular respiration and increase the levels of the reduced form of NADH. When hESCs are treated with KCN the FLIM Phasor distribution shifts toward the location of the free reduced NADH [Fig. 10(c)and 10(i)] and the concentration of free NADH increases with respect to bound NADH [Fig. 10(g)and 10(m) of the Appendix].

Fig. 9

Mitochondria colocalization and contribution of NADH: (a) FLIM phasor plot of the FLIM image excited at 760 nm of a single differentiating H9 hESC colony area. The red cluster in the phasor plot specifically selects some bright granules within the hESCs, while the blue cluster selects the rest of the cellular cytoplasm and cell nuclei. Transmission image (b) and two-photon fluorescence intensity image (c) of the hESC colony area. (d) Phasor color map of the hESC colony area. Pixels of different colors correspond to the color of the cluster in the phasor plot (a). (e) In vitro staining of mitochondria with TMRE.

Fig. 10

Effect of electron-transport-chain inhibition on the hESC FLIM phasor distribution: (a) Phasor plot selection using linear cluster combination that represents all the possible relative concentrations of bound NADH and free NADH. The phasor locations of pure bound and free NADH have been measured in Ref. 32. Each point along the line has a color corresponding to specific relative concentration of bound-to-free NADH. In vitro FLIM phasor plot of an undifferentiated H9 hESC colony (b, c) and a differentiating H9 hESC colony (h, i) before and after the treatment with potassium cyanide (KCN). Two-photon fluorescence-intensity images of the undifferentiated hESC colony (d–e) and differentiating hESC colony (j–k) before and after the KCN treatment. Phasor color map images representing the relative concentrations of bound and free NADH in the undifferentiated hESC colony (f–g) and the differentiating hESC colony (l–m) before and after the KCN treatment. By blocking the respiratory chain in hESCs, the FLIM phasor distribution shifts toward the location of the free reduced NADH (c, i) and the cell concentration of free NADH increases with respect to bound NADH (g, m).

2.2.

Phasor FLIM Discriminates Undifferentiated hESCs from Differentiating hESC Colonies

We measured the metabolic signature of hESC differentiation by monitoring the concentrations of the intrinsic fluorescent biomarkers NADH and LDAGs. In Fig. 11 we identified the differences between the FLIM phasor signature of undifferentiated hESCs and differentiating hESCs. To induce differentiation towards trophectoderm or a neurogenic lineage, we cultured hESCs in the presence of BMP4 or retinoic acid (RA) respectively, while simultaneously removing basic fibroblast growth factor (bFGF) from the culture media, which is critical to the maintenance of hESC pluripotency and self-renewal (see Sec. 4).^59,60 We measure the pluripotency and differentiation of hESC by monitoring the expression of the hESC marker OCT4 [Fig. 11(f) and 11(h)] with retrospective immunostaining after live imaging (see Sec. 4). Undifferentiated hESC have a high expression of OCT4, while differentiating hESC are characterized by lower expression of OCT4 [Fig. 11(f) and 11(h)].

Fig. 11

FLIM phasor separates undifferentiated from differentiated hESC colonies: FLIM phasor plot of the FLIM image excited at 760 nm of a single undifferentiated H9 hESC colony area (a) and a differentiating H9 hESC colony treated with BMP4 medium for four days (b). Phasor plot selection using linear cluster combination that represents all the possible relative contributions of the hESCs granule FLIM signature (purple) identified in Figs. 1 and 2 and the NADH FLIM signature typical of the MEFs (cyan-white), identified in Figs. 1 and 4. Each point along the line has a color that corresponds to a specific relative concentration of the two species. Transmission images (c) and two-photon fluorescence intensity images (d) of the undifferentiated hESC colony and the differentiating hESC colony. (e) Phasor color map images representing the relative concentration hESC granules (purple) and NADH (cyan-white) according to the color scale in Fig. A and B. (f) Expression of the pluripotency marker OCT4. (g) DAPI staining. (h) Oct4 expression in undifferentiated hESC compared to differentiating hES cells with BMP4 medium for four days, measured from image (f). (i)–(j) Scatter plot of the phasor FLIM signature of hESC colonies. Every point represents the average phasor value of an entire hESC colony. Black and cyan squares represent undifferentiated H9 colonies ($N=27$) and undifferentiated H1 colonies respectively ($N=8$). Red circles represent differentiating H9 colonies treated with BMP4 media for four days ($N=24$), purple triangles H9 hESC colonies differentiating in a medium without bFGF ($N=6$), orange triangles H9 hESCs induced to differentiate in RA medium for four days ($N=6$).

The FLIM phasor distribution of undifferentiated hESCs [Fig. 11(a)], is dominated by the contribution of the hESC granules identified in Figs. 1, 2, and 4. On the other hand, the FLIM phasor distribution of differentiated hESCs treated with BMP4 [Fig. 11(b)], is shifted toward the central region of the phasor plot represented by the FLIM fingerprint of NADH³² (Figs. 9 and 10). During differentiation the lifetime of the LDAGs in hESC does not change, the ratio of free to bound NADH presents a small increase [Fig. 10(f) and 10(l)], and the ratio of LDAG to NADH increases significantly (Fig. 11).

We used a phasor linear cursor [colored bar in Fig. 11(a) and 11(b)] to represent all possible contributions of the LDAGs (purple) and the NADH (cyan-white). Every single color along the line represents a different relative concentration of the two fluorescent metabolic markers. The colored FLIM map of Fig. 11(e) shows that the H9 undifferentiated hESC colony has a higher concentration of LDAGs with respect to NADH, while the differentiating hESC colony has a much lower concentration of LDAGs with respect to NADH.

In Fig. 11(i) and 11(j)we measured the phasor FLIM signature of the entire hESC colony by calculating the average value of its phasor FLIM distribution (see Sec. 4). Figure 11(j) shows that the FLIM signatures of undifferentiated hESCs from two different cell lines, H9 (black squares) and H1 (cyan circles) are localized in the same region of the phasor plot and are not statistically different ($p=0.34$). The phasor FLIM signatures of undifferentiated H9 hESCs (black squares in Fig. 11(i) are statistically different ($p<0.0001$) from differentiating H9 hESCs treated with either BMP4 (-bFGF) medium [red circles in Fig. 11(i)], RA (-bFGF) medium [orange triangles in Fig. 11(j)] or -bFGF only medium [purple triangles in Fig. 11(j)]. While the hESCs treated with RA and BMP4 are not statistically distinguishable ($p=0.25$), colonies cultured in the absence of bFGF are statistically different from both conditions ($p<0.0001$) and have a phasor FLIM signature closer to that of undifferentiated hESCs [Fig. 11(j)].

The trend in Fig. 11(i) and 11(j)of the phasor FLIM signatures of the H9 colonies during differentiation reflects differences in relative concentration of LDAGs and NADH, that is mapped in Fig. 11(e). The increase of NADH concentration (Fig. 11) and the small increase in free-to-bound NADH [Fig. 10(f) and 10(l)] during early hESC differentiation is in agreement with the decrease of oxidative-phosphorylation rate with respect to glycolysis during early hESC differentiation.⁴⁶ We hypothesize that the abundance of LDAGs in undifferentiated hESCs indicates the formation of fluorescent aggregates generated by the interaction of ROS and unsaturated lipids. Hence the decrease in LDAGs concentration during differentiation can reflect a decrease in the abundance of unsaturated eicosanoids, such as linoleic acid and arachidonic acid, that promote pluripotency,³⁴ the activation of oxidation and a decrease in antioxidant enzymes concentration^36,61 or a decrease in ROS level following a decrease in oxidative-phosphorylation rate.⁴⁶

2.3.

Cell Phasor FLIM of Heterogeneity in hESCs

The phasor FLIM analysis at single cell resolution reveals heterogeneity in the metabolic signature and intrinsic biomarker content of hESCs within the same colony. We exploit the cell-phasor concept and image segmentation (Ref. 32 and Sec. 4) to measure the phasor FLIM signature of single hESCs by calculating the average value of the phasor distribution of single hESCs (see Sec. 4).

Figure 12(a) and 12(b) show the plot of single hESCs from an undifferentiated H9 colony (black dots), an undifferentiated H1 colony (cyan dots) and a differentiating H9 colony treated with BMP4 (red dots). The cell phasor cluster of undifferentiated H9 hESCs is smaller in size compared to the cell phasor cluster from differentiating H9 hESCs [Fig. 12(a)] and the standard deviation of the g component (defined in the Appendix) of the cell phasors from an undifferentiated H9 colony is smaller than that of differentiating H9 hESCs [Fig. 12(c)]. This observation indicates that the FLIM signature of single hESCs in a H9 colony becomes more heterogeneous during differentiation. Cell phasor and FLIM signature of single hESCs is determined by the relative concentration of the intrinsic fluorescent metabolites NADH and LDAGs (Figs. 2, 4, and 9). The cell phasor reveals the metabolic signature of the cells that might be associated with the stem cells’ phenotype and be indicative of commitments to different differentiation pathways. The higher degree of cell phasor heterogeneity within the H1 hESC colonies [Fig. 12(c)] might reflect the higher phenotypic heterogeneity and differential expression of hESC markers that are known to be present in H1 hESC-lines.⁶²

Fig. 12

Heterogeneity within hESC colonies: (a) Scatter plot of the phasor FLIM signature of individual hESCs from the two colonies displayed in Fig. 11. Every point represents the average phasor value of a single hES cell. Black squares represent undifferentiated H9 hESCs ($N_{\text{cells}}=123$) and red circles represent differentiating H9 hESCs treated with BMB4 media for four days ($N_{\text{cells}}=136$). (b) Cyan squares represent the cell phasor of H1 hESCs ($N_{\text{cells}}=119$) from the colony and green squares represent MEFs ($N_{\text{cells}}=42$). (c) Standard deviations ($N_{\text{colony}}=3$) of the phasor g coordinates of the cell phasor of single hESCs from an undifferentiated H9 colony, differentiating H9 and H1 colony and MEFs.

Moreover, Fig. 12(c) shows that the cell phasor cluster size of both undifferentiated and differentiating hESC colonies are smaller than the size of the MEF cluster [green dots in Fig. 12(b)]. The heterogeneity in the FLIM/metabolic signatures of hESCs might reflect different cell phenotypes and plasticity that are reduced in a defined fully differentiated cell types such as MEFs.

4.1.

Cell Culture

We used federally approved H9 and H1 human embryonic stem cell lines. hESCs are cultured on a substrate of mouse fibroblast (MEF) feeders (Chemicon Cat #PMEF-CF). Plates are first coated with 0.1% to 0.2% gelatin (Sigma #G-1393). MEFs are then plated with a density of approximately of $\mathrm{15,000}\text{cells}\text{per}{\mathrm{cm}}^2$. hESCs ranging from passage 42 to 57 were used for image analysis. hESCs were grown in hESC cell lines H1 and H9 and were cultured in a medium containing: DMEM-F12 (Invitrogen 12660), 20% KnockOut serum replacement (KOSR, Invitrogen 10828), 1% nonessential amino acids (NEAA, Invitrogen 11140-050), 1 mM Glutamax (Invitrogen 35035), $4\mathrm{ng}/\mathrm{mL}$ basic fibroblast growth factors (bFGF, Invitrogen 13256–029) and 0.1 mM beta-mercaptoethanol, as described by Xu et al.⁶⁶ Medium was changed every day and hESC were maintained in an incubator at 37 uC with 5% CO2. Differentiation was induced by removing bFGF from hESC medium and adding either $100\mathrm{ng}/\mathrm{ml}$ BMP4 (R&D) or RA (1 µM) as described by Pera et al.⁶⁷ We image hESC during their early differentiation stages, after four days of treatment with BMP4 or RA. Treating embryonic stem cells with BMP4 differentiates them towards trophectoderm, while RA differentiates them down the neurogenic lineage.⁵⁹

4.2.

MEFs Medium

Dulbecco’s Modified Eagle Medium (DMEM), 1 mM Glutamax (Invitrogen 10569), 10% fetal bovine serum (FBS, Invitrogen 16000–044) and 1% nonessential amino acids (Invitrogen 11149–035) were used.

4.3.

In Vitro Staining

Lipid droplets were stained with BODIPY 493/503 (Invitrogen #D3922). hESCs mitochondria were stained with $1\mathrm{ul}/1\mathrm{ml}$ TMRE (Sigma #87917, excitation⁄emission: $540/595\mathrm{nm}$) solution. After 30 min of staining cells were washed and imaged. hESCs lysosomes were stained with 50 nM LysoTracker Red (Invitrogen L7528, excitation⁄emission: $577/590\mathrm{nm}$) in phosphate buffered saline (PBS). Endoplasmic reticulum was stained with ER-Tracker™ Red (Invitrogen #E34250).

4.4.

KCN Treatment

We block the respiratory chain by means of KCN to inhibit the oxidative phosphorylation and increase the mitochondrial concentration of NADH. KCN in PBS was added to the culture medium with a final concentration of 4 mM. Cells were imaged immediately after the addition of KCN.

4.5.

Alkaline Phosphatase Staining

Once the media are removed, cells are washed with PBS minus and fixed for 20 min using 4% paraformaldehyde (PFA). Cells were washed three times with deionized water. $1\mathrm{mg}/\mathrm{ml}$ of FastRed is dissolved in deionized water right before use, then combined with $40\mathrm{\mu l}/\mathrm{ml}$ of Napthol. hESCs were stained for 20 min and then washed with deionized water.

4.6.

Immunostaining

Medium was removed from the stem-cell plate and washed with PBS. Cells were fixed in 4% PFA. hESCs were washed three times with PBS, permeabilized with 1 ml cold Methanol for 5 min at room temperature and then washed three times with PBS. The cells were blocked for an hour at room temperature using 10% Donkey serum (Sigma D9663) in PBS. Primary antibodies OCT4 (R&D AF1759-1∶100) and CONNEXIN-43 (Cell Signaling 3512-1∶50) were diluted in 1% Donkey serum and incubated with the cells for 1 hr at room temperature. Cells were washed three times with PBS. The secondary antibody Donkey anti-Goat Alexa 568 (Invitrogen A-11057) was used at a $1:400$ and incubated with the hESCs for 1 h at room temperature. DAPI solution was added to the cell to stain DNA. Finally cells were washed with PBS.

4.7.

Confocal Imaging

DAPI was excited with 405-nm laser. TMRE, LysoTracker Red, ER-Tracker Red and anti-Goat Alexa 568 were excited with the 561-nm laser. BODIPY 493/503 was excited with the 488-nm laser.

4.8.

Multiphoton and Fluorescence Lifetime Imaging

Fluorescence lifetime images were acquired with a Zeiss 710 microscope coupled to a Ti:Sapphire laser system (Spectra-Physics Mai Tai) and an ISS A320 FastFLIM.⁴⁷ A $40\times 1.2\mathrm{NA}$ water immersion objective (Zeiss Korr C-Apochromat) was used. For image acquisition the following settings were used: image size of $256\times 256\text{pixels}$ and scan speed of $25\mu \mathrm{s}/\text{pixel}$. A dichroic filter (690 nm) was used to separate the fluorescence signal from the laser light and the fluorescence. For the acquisition of FLIM images, fluorescence was detected by a photomultiplier (H7422P-40 of Hamamatsu) and a 610-nm short-pass filter was placed in front of the detector. FLIM data were acquired and processed by the SimFCS software developed at the Laboratory of Fluorescence Dynamics (LFD), University of California Irvine. The excitation wavelength was 760 nm if not differently specified. An average power of about 5 mW was used to excite the live hESCs. FLIM calibration of the system was performed by measuring the known lifetime of the fluorescein with a single exponential of 4.04 ns. FLIM data were collected until 100 counts in the brightest pixel of the image were acquired. Typically the acquisition time was less than 1 min. We verified that the autofluorescence of all MEF and hESC culture media does not interfere with the FLIM and spectral signature of cellular autofluorescence. We verified that the FLIM signature of undifferentiated hESC does not change with number of cell passage and size of the colony.

4.9.

FLIM Data Analysis

Every pixel of the FLIM image was transformed in one pixel in the phasor plot as previously described and reported^32,48 in detail in the Appendix. The coordinates g and s in the phasor plot were calculated from the fluorescence-intensity decay of each pixel of the image by using the transformations defined in the Appendix. The analysis of the phasor distribution was performed by cluster identification. Clusters of pixel values were detected in specific regions of the phasor plot. Fractional intensities of chemical species in every pixel of the image were evaluated with a graphical analysis in the phasor plot (see Appendix). We performed image segmentation on the FLIM data by selecting the region of interest within the sample. The region of interest of single hESCs or hESC colonies was selected by using a cursor of arbitrary shape. We calculate the phasor average value within these regions of interest. When measuring the cell phasor, all pixels of the cell (about 1000) were taken in account and the signal-to-noise ratio of the FLIM signature of cells is higher than in single pixels. All phasor transformation and the data analysis of FLIM data are performed using SimFCS software developed at the LFD.