1. Introduction

Megakaryopoiesis and platelet production is an intricate biological phenomenon, requiring a tightly regulated sequence of cellular transformations. Megakaryocytes (MKs) are cells of hematopoietic origin that reside primarily in the bone marrow [1] and are responsible for maintaining and renewing the pool of circulating platelets. Platelet function is fundamental for hemostasis and thrombosis but is also implicated in inflammation and cancer [2]. Hematopoietic Stem Cell (HSC) differentiation into MKs requires several successive lineage commitment steps and is actively driven by biochemical and mechanical signals triggered by multiple cytokines and extracellular matrix components, among which, thrombopoietin (Tpo) and fibronectin, play protagonistic roles [3, 4]. Early committed MK progenitors lose proliferative potential and undergo a series of transformational stages to prepare for platelet production [1]. In the latest phases of differentiation, MKs migrate proximally to bone marrow capillaries where, in response to specific stimuli, they convert their cytoplasm into long, branched extensions (proplatelets) into the circulation. Under shear, platelets are released from the tip of the proplatelets into the blood stream [5]. Both in vivo and in vitro, MKs in various transformational stages coexist, creating an inherently complex and heterogeneous biological system spatially and temporally. Despite major advancements in the understanding of MK biology, the exact mechanisms activated during MK differentiation that drive or interfere with differentiation progression and ultimately with platelet formation remain elusive. Consequently, the pathogenesis of many related diseases and corresponding targeted therapies remain unknown, resulting in palliative treatments. Elucidating the metabolic behavior of MKs dynamically during their tightly controlled maturation, could further advance our understanding of platelet generation and thus be exploited for new therapeutic strategies and the improvement of in vitro platelet production protocols to achieve clinical-grade standards [6, 7].

Two photon excited fluorescence (TPEF) microscopy is a non-destructive imaging modality that offers subcellular resolution and can be utilized to assess the metabolic state of living cells within heterogeneous cellular populations by exploiting the natural fluorescence detected from the metabolic co-enzymes nicotinamide adenine dinucleotide (phosphate) (NAD(P)H) and flavin adenine dinucleotide (FAD) [8]. The relative fluorescence intensities of NAD(P)H and FAD can be quantified and described as a normalized measure of cell redox state in the form of the optical redox ratio (RR), defined as FAD/(FAD + NAD(P)H). The latter has been previously shown to correlate with liquid chromatography/mass spectrometry (LC-MS) measurements of NAD⁺/(NADH + NAD⁺) and/or FAD/ (FAD + NADH) [9, 10]. Further, evaluation of NAD(P)H fluorescence lifetime can provide complementary information to the traditional ratiometric fluorescence intensity outcomes, by probing microenvironmental parameters related to the metabolic co-enzymes’ functional states, such as their enzyme-bound and unbound contributions [11, 12]. Lastly, as the NAD(P)H TPEF images are predominantly sensitive to the bound NAD(P)H form [13], which along with FAD [14] resides primarily within mitochondria, fluorescence intensity and fluorescence lifetime images can be used to provide insight into the cellular mitochondrial activity. Mitochondria are key organelles for the regulation of the bioenergetic homeostasis and for other cellular processes including that of cellular differentiation, an event necessary for platelet generation [5].

In this study, we utilize intrinsic cellular auto-fluorescence contrast generated by TPEF imaging to evaluate megakaryocytic differentiation in vitro. We show that TPEF microscopy allows for non-invasive structural identification of the individual megakaryocytic maturation stages at the single cell level and we further extract quantitative metabolic outcomes for each of the identified differentiation steps. Our results demonstrate that high-contrast, label-free two photon imaging of megakaryocytic development holds great potential in studying and elucidating the underlying physiological processes controlling and driving the intricate process of platelet production.

2. Materials and methods

2.1 Cell culture

Human MKs were differentiated from cord blood-derived hematopoietic stem cells (HSCs), according to an established protocol [15]. CD34⁺ hematopoietic stem cells were isolated from cord blood collected from two healthy donors by density gradient (Lymphoprep, Cedarlane), followed by immunomagnetic cell separation (CD34⁺ MACS selection kit, Miltenyi, Germany). The HSCs were cultured for 13 total days in serum-free media (StemSpan SFEM, StemCell Technologies, Canada), in the presence of 10ng/mL IL-6, IL-11, Tpo (Peprotech, NJ) to promote MK differentiation. At day 13, the megakaryocyte population, characterized by the expression of the Fibrinogen receptor (CD61/CD41), reaches the maximum expansion. The CD61⁺ expressing MK populations were purified by immunomagnetic selection (CD61⁺ MACS selection kit, Miltenyi) and seeded on 4 independent glass bottom petri dishes per donor, coated with 50 µg/mL fibronectin (Sigma-Aldrich, MO), to enhance proplatelet extension. TPEF intensity and lifetime images were acquired 3h and 24h after seeding, denoted hereafter as Day 0 and Day 1 of fibronectin stimulation respectively. According to our previously published data [15], the 3h (Day 0) time-point was chosen as a baseline condition and the 24h (Day 1) one as the proplatelet-extension peak prior to extension shedding and apoptotic activation. 12 fields on average were imaged per dish and timepoint yielding a total of 191 fields (2 donors x 2 timepoints x 12 fields x 4 dishes).

2.2 Imaging

MK cultures sustained in glass bottom dishes were imaged with a Leica TCS SP8 confocal microscope (Wetzlar, Germany) equipped with a Ti:sapphire laser (Spectra Physics, Mountain View, CA) and time-correlated single-photon counting electronics (Picoquant, Germany). Samples were excited with 755 nm and 860 nm and imaged using a 40x/1.1 NA water immersion objective (Leica Microsystems, Germany). TPEF (512x512 pixels; 2.6x zoom; 145μm FOV) redox ratio and fluorescence lifetime (FLIM) images were acquired at 460 ± 20 nm and 525 ± 25 nm. Incident laser power at the focal plane during the acquisition was ~20mW, pixel dwell time was 1.67µsec and FLIM integration time was 30 sec. Samples were kept in a humid incubator stage chamber maintained at 37 ° (Leica Microsystems) and 20 mM HEPES (Thermo Fisher Scientific) was added to the media prior to imaging for pH buffering. After the endogenous fluorescence imaging session was completed, cells were stained with Mitotracker orange (200nM) and reimaged. 2P NAD(P)H autofluorescence images were first collected and then at the same field location confocal images of the mitotracker staining were obtained using the 543 nm excitation at an emission range of 550–625 nm (Appendix Fig. 5).

2.3 Image processing and analysis

Image processing was performed in Matlab. All acquired images were normalized for illumination power and detector gain. For redox analysis, for a given field, the 755nm excitation, 460nm emission (NAD(P)H channel) and 860nm excitation, 525nm emission (FAD channel) were spatially co-registered by determining the shift maximizing correlation between the two channels (Appendix Fig. 6(A)). Redox ratio image maps were generated on a pixel-by-pixel basis as the normalized fluorescence intensity contributions from FAD over the sum of the intensity contributions from NAD(P)H and FAD. Regions of interest (ROIs) were then manually selected to classify individual cells according to their differentiation stages (Appendix Fig. 6(B)). A custom batch processing routine was created and utilized, that allowed the experts to trace the cells through a graphical user interface (GUI) and to automatically sort through the selected cells and assign the class the cells belong to. This minimized other laborious aspects, such as manual image loading and data saving; however, approximately 10 hours were still dedicated for the cellular selection and classification aspects of our study. The cell classification was performed in this study by two trained experts (A.B and L.T). The allocation was based on standard morphological criteria, routinely used in clinical practice, which include cell size and circularity, nuclear morphology, size, and position (Appendix Fig. 7) [17, 18]. For the terminal differentiation stages, classification was assigned based on the presence and shape of the well recognizable cytoplasmic extensions, with wide and short extensions (podosomes) defining class V, and long and thin extensions (proplatelets) defining class VI respectively. The classification mismatch between the experts was below ~5% of the total cells examined and mismatched cells were excluded from the analysis. To isolate cells and remove background and saturated pixels, low and high intensity thresholds were utilized to create binary masks that were based on the lowest 10th and highest 70th intensity percentile. These masks captured the entire cellular area including the nuclei. To isolate intracellular image features attributed to NAD(P)H fluorescence, while also removing dark nuclear and vesicular features, custom digital bandpass filters were employed as described previously [16] that created stricter binary masks (Appendix Fig. 6(C)). A ratio between the nuclear contribution (the difference of the entire cell area minus the stricter NAD(P)H-related mask) over the entire cell, provides information about the nuclear contribution with respect to the cellular area, in a manner similar to the traditional histopathological marker of nuclear to cytoplasmic (N:C) ratio. Using batch-analysis custom software, a mean normalized RR value was computed for each selected cell as the normalized fluorescence intensity contributions from FAD over the sum of the intensity contributions from NAD(P)H and FAD within each respective cellular ROI. Time-resolved fluorescence decay data were analyzed by the phasor approach [19] as described previously [16, 20, 21]. Briefly, sine and cosine transforms were applied to decay profiles at each pixel and an instrument response function calibration was applied to the resulting phasors based on an umbelliferone standard. For each cellular differentiation stage a density map was plotted from all respective pixel phasors and a phasor centroid location was calculated. Further a line fitted to the phasor distributions identified component lifetimes, assuming an underlying bi-exponential decay profile. In phasor space, a mono-exponential decay is represented by a point along the “universal circle”, whereas a bi-exponential decay (i.e. unbound and enzyme-bound NAD(P)H) will fall within the area circumscribed by the “universal circle” and its location will be along a linear trajectory defined by the “universal circle” positions of the individual single exponential decay components. The relative position towards the shorter or the longer lifetime component signifies the components’ respective linear contributions to the emitted fluorescence intensity and that relative distance can be correlated with the bound fraction ratio [20]. Briefly, for the bound fraction calculation, an average fitted line was calculated incorporating all data to determine a standard curve connecting the short and long lifetime components contributing to the fluorescence signal captured. Then the projection of each pixel’s phasor and its bound fraction value was computed, by calculating the ratio of the projection’s absolute distance from the short lifetime component over the absolute distance between the points on the universal semi-circle representing the free (short) and bound (long) NAD(P)H lifetimes. On the bound fraction scale, the fluorescence lifetime increases as the bound fraction approaches 1. Bound Fraction (BF) image maps were calculated on a per pixel basis and a mean Bound Fraction value was computed per each respective cellular ROI by averaging the pixels’ BF values. The overall procedure for extracting the bound fraction information is much faster and simpler than performing bi-exponential fitting of the corresponding lifetime spectra of each pixel. A total of 191 fields and 1,818 cells were evaluated. Cells corresponding to the blast-undifferentiated phase were not considered in the analysis as they are morphologically undistinguishable from other lineage progenitors that may have inadvertently escaped the immunomagnetic separation.

2.4 Morphometric analysis

For morphometric analysis, a subset of the total fields (n = 50 representative images/ ~25% of total fields) were analyzed in ImJ and the perimeter of ~5 cells per image was carefully traced yielding a total of 263 cells (N_II = 91, N_III = 69, N_IV = 55, N_V = 19, N_VI = 29). For classes V and VI the traced perimeter incorporated the cellular cytoplasmic extensions. The traced shape was fit with an ellipse and parameters including the area, major and minor axes along with circularity were calculated. The cellular diameter is reported as the mean of the major and minor axes per cell. (Table 1).

Table 1. Morphometric analysis

View Table

2.5 Statistics

Statistical analysis was performed in JMP 13 (SAS). To assess statistical significance of the redox ratio differences among groups, we used an one-way ANOVA and Tukey post hoc test. The average redox ratio of each cell was computed and then the values from all cells per differentiation stage were analyzed in the ANOVA. Similarly, one-way ANOVAs and Tukey post hoc tests were also used for the analysis of the bound fraction and N:C calculations.

3. Results

3.1 In vitro label free evaluation of megakaryocytic development

Traditionally, nuclear size, location and polyploidy assessments require exogenous staining or invasive approaches [17, 18]. In TPEF images though, nuclear features appear dark, facilitating therefore label free cellular classification of megakaryocytic development, especially in the stages prior to proplatelet extension. The morphological stage of megakaryocytic development was determined by visual assessment of cell size, nuclear-cytoplasmic ratio and nuclear configuration [17, 18] (Fig. 1(A), Table 1), using the transmission, and the NAD(P)H and FAD channels displayed as redox image maps (Fig. 1(B)-1(C)).

 

Fig. 1 Morphological stages of megakaryocytic development and platelet generation. A. Schematic of morphological stages of megakaryocytic development, indicating the nuclear configurations and cytoplasmic transformations. Cell examples for each stage are also shown enlarged as redox ratio images. Scale bar is 15 µm. B. Representative transmission images displaying typical cellular diversity within image fields and C. Corresponding redox ratio maps. Examples of cells in several differentiation stages are noted (II-VI). Black arrows in B point to cytoplasmic extensions from MKs during terminal differentiation. The advantage of TPEF images in non-invasively distinguishing nuclear morphology and localization is evident. Scale bar is 20µm for all images.

Download Full Size | PDF

Stage I, the megakaryoblast stage, represents early committed MK progenitors that are characterized by very compact nuclear morphology occupying almost the entire cytoplasm. As cells lose proliferative potential and start to differentiate into promegakaryocytes (stage II), the nucleus takes a horseshoe-like morphology, while contralaterally bright autofluorescence starts to arise. In stage III, the process of endomitosis (nuclear multiplication) peaks, leading to the formation of large, multinucleated, mature MKs with a DNA content up to 64N [22]. Cellular size increases (Table 1) and intense multilobe nucleic features are identified occupying significant percentage of the expanded cellular volume. In stage IV, nuclear features become compact and pushed to the edge of the cell. Concurrently, cytoplasmic maturation occurs, which entails the accumulation of organelles, specialized granules (alpha and dense granules) and membrane content, in a highly invaginated system of interconnected tubules and reservoirs, referred to as the Demarcation Membrane System (DMS) [1]. Morphologically, this corresponds to a sharp decrease of the N:C ratio (Fig. 2(D)). Stage V and VI account for cells displaying proplatelet extensions. In stage V, during the initial stages of proplatelet extension the MK starts to spread and its cytoplasm unravels into cytoplasmic extensions, called podosomes [23]. As the cytoplasmic erosion continues, the extensions elongate and are remodeled into thinner strands with a pearl-on-a-string morphology, which will eventually shed into platelets [24]. These morphological changes are evidently witnessed in the TPEF images and the intensely autofluorescent cytoplasmic areas are primarily associated with mitochondrial-derived fluorescent signal as validated by Mitotracker Orange staining (Appendix Fig. 5). As mitochondria are transported from the cellular body to the proplatelet extensions, both structural and metabolic spatiotemporal evaluation is enabled. We performed morphological population analysis by quantifying the percentage of cells in the different maturation stages before (Day 0) and after (Day 1) fibronectin adhesion (Fig. 2(A)-2(C)), as well as their morphological nuclear/cytoplasmic changes (Fig. 2(D)).

 

Fig. 2 Morphological evaluation of megakaryocytic development. A-C. MK population analysis, quantifying the proportional contribution of each differentiation stage before (Day 0) and after (Day1) fibronectin adhesion as well as overall contributions. The raw cell numbers and corresponding percentages are shown as polar area diagrams with the area of a sector representing each stage’s contribution. D. Automated quantification of nuclear to cytoplasmic ratio. * denotes p<0.05 with statistical significance calculated by ANOVA and post-hoc Tukey test.

Download Full Size | PDF

In our cultures, stage II and III megakaryocytes dominate in Day 0 (Fig. 2(A)), as expected. They are characterized by high nuclear contribution to the cellular volume (Fig. 2(D)), caused by the endomitotic processes reaching maximal rate. On Day 1, as fibronectin sustains MK terminal differentiation [25], a significant increase in mature MKs is observed (stages IV-VI) (Fig. 2(A)), with their respective nuclear contribution to the cellular volume substantially decreasing, as cytoplasmic expansion and proplatelet extension dominate in these later phases (Fig. 2(D)).

3.2 Fluorescence lifetime and optical redox ratio imaging capture differences in the metabolism of the differentiating megakaryocytic population

We sought to evaluate whether metabolic changes, as assessed by FLIM and optical redox ratio imaging, showed similar trends accompanying the structural transformations observed during the distinct differentiation stages. Most endogenous fluorophores, due to their conformational heterogeneity i.e unbound or bound states, have decays described by multiexponential components. Phasors provide a simple and fast graphical approach to display a fluorophore’s lifetime decay without requiring any exponential fitting, thus without making any a priori assumptions about the number of contributing components.

The respective pixels’ phasors from all cells belonging to each differentiation stage, as quantified in Fig. 2(C), were compiled (Fig. 3(A)-3(E)) generating each stage’s phasor map. Each stage’s phasor centroid was also individually plotted (Fig. 3(F)) to more clearly show trends of the phasor trajectory among the MK maturation phases. The phasor maps and the respective bound fraction-coded images (Fig. 3(A)-3(E) image inserts) reveal that as differentiation progresses the contribution of longer lifetimes increases (Fig. 3(F)). Note that no fitting is required for this information to be extracted from the phasor maps.

 

Fig. 3 NAD(P)H fluorescence lifetime phasor analysis during megakaryocytic differentiation. A-E Phasor maps compiled for each differentiation stage, pseudocolored based on phasor point density (brighter color hues indicate higher spatial density). Dashed lines depict linear fits to the phasor distributions, with respective distribution centroids indicated by black circles. Image inserts show representative cells from each stage, pseudocolored based on bound NAD(P)H fraction maps. Scale bars are 15um for all images. F. Centroid linear trajectory comparison for all groups.

Download Full Size | PDF

We then proceeded to quantify the mean optical redox ratio (RR) (Fig. 1(C), 4(A)) and bound fraction (BF) (Fig. 4(B)) on a per cell basis for each differentiation stage. The mean RR and BF per cell were extracted and then averaged for each MK stage (Fig. 4(A), 4(B)). Subtle yet distinctive trends are observed between the different stages. A significant drop in the RR ratio is detected during stage III, coinciding with the peak of endomitosis, while in later phases the RR recovers to higher values. The mean BF values support the pixel based phasor map trends described previously, with the BF increasing in later MK stages. Interestingly, the RR and BF trends do not fully correlate with each other during the individual differentiation stages examined, implying a differential sensitivity of each optical biomarker to the underlying biological functions occurring during each transformational stage.

 

Fig. 4 Cell based functional quantification. A. Mean Redox Ratio and B. Mean Bound Fraction calculation from MKs in every differentiating stage averaged among all cells belonging in each stage. Means and standard errors presented, * denotes p<0.05 with statistical significance calculated by ANOVA and post-hoc Tukey test. C. Cell based RR and BF histograms. Histogram inserts show component weights based on the extracted fit parameters from all cells examined (Appendix Fig. 8). Reversed colored representative images are also presented with pixels colored based on which distribution component they belong to for each respective biomarker.

Download Full Size | PDF

Lastly, we plotted the mean RR and BF on a per cell basis for the entire cellular population examined (1,818 cells from 191 image fields) (Appendix Fig. 8(A)-8(B)) as well as for the populations belonging to each differentiation stage (Fig. 4(C)) to holistically evaluate the functional changes observed during megakaryocytic development. The merged population distributions were fit with a 2-component normal mixture and the extracted distribution parameters (means and sigmas) (Appendix Fig. 8(C)-8(D)) were utilized to independently fit the distributions of each differentiation stage and calculate the relative component weights (Fig. 4(C)). The extracted means’ semi difference for each biomarker was also utilized as an image intensity threshold to color in a binary fashion the redox ratio and bound fraction image color maps (Fig. 4(C)). This analysis revealed spatiotemporal changes in the localization of low and high RR and BF components during differentiation. Within stage II the RR and BF values seem to compartmentalize highly, with lower RR and BF values clustering around cell nuclei while the higher BF values spatially correlate with the higher RR values where mitochondrial populations are expected to congregate based on histological images of MKs stained with eosin/hematoxylin [17]. During stage III and IV that compartmentalization dissipates, re-emerging during stage V and VI, primarily for the RR images, with higher RR values detected primarily within the proplatelet extensions. In contrast, the BF images appear more homogeneous and are dominated by the long component. The observed spatiotemporal optical biomarker changes and intracellular heterogeneity patterns hint towards varying metabolic tendencies not only between cells of different transformational stage but also within individual cells’ compartments.

4. Discussion and conclusions

By utilizing only intrinsic, cellular auto-fluorescent contrast and applying automated image analysis algorithms, we show that distinct metabolic behaviors can be detected non-invasively in megakaryocytic cellular populations upon the onset of differentiation and platelet formation. As the exact mechanism and trigger of proplatelet formation and platelet release are still unknown, non-invasive techniques that can provide insight in metabolic dynamics could prove invaluable for the study of these finely regulated processes.

To determine the sensitivity of TPEF microscopy of intrinsic fluorophores to label free evaluation of MK differentiation in in vitro cultures, we imaged heterogeneous MK populations dynamically, in a well-defined culture system. Moreover, the addition of an adhesive protein, such as fibronectin, allowed us to capture the extremely transient phases of terminal differentiation. We identified and quantified the various transformational MK stages based on the observed nuclear and cytoplasmic morphological changes. The population analysis after adhesion on fibronectin revealed a population shift to more terminal differentiation stages, validating in a quantitative manner the effect of extracellular matrix components to differentiation progression [5, 15].

In addition to significant morphological differences between differentiation changes, we observed distinct metabolic trends both inter- and intracellularly, as evaluated by single cell RR and BF measurements. Intercellularly, an initial drop in the average RR ratio was observed in stage III followed by a later phase recovery, while BF values displayed increasing trends overall in later MK development stages. Intracellularly, heterogeneous compartmentalization patterns were observed during the individual differentiation stages examined that did not always spatially correlate between the two optical biomarkers. Megakaryocytic development entails a plethora of underlying biosynthetic and catabolic energy producing metabolic pathways that are swiftly regulated to facilitate the differentiation progression. During the early immature stages a substantial need for biosynthesis exists, and glucose metabolites are expected to be diverted from the glycolytic pathway and the TCA cycle to facilitate glycoprotein, lipid, and nucleotide synthesis, in order to support nuclear multiplication, formation of the demarcation membrane system and accumulation of the necessary structures that will eventually be distributed into the nascent platelets (granules, organelles) [5]. This intense biosynthetic need in early stages could be the driving force for the detected optical changes in the early MK stages. As MKs further mature, the DMS, which functions as a membrane reservoir, driven by energy consuming cytoskeletal dynamics [26] is utilized to create cytoplasmic expansions that later fragment and lead to the formation of platelets. Cytoskeletal reorganization for proplatelet extension [26] and cargo transportation through the cytoplasmic extensions create a rising demand for ATP availability. The latter in combination with a switch to metabolic pathways found in resting platelets, like de novo fatty acid synthesis and oxidation, could potentially explain the increasing RR and BF trends and the compartmentalized metabolic differences detected during the terminal differentiation stages. Such a metabolic switch could be also the driving force for the small and gradual change in the slope of the line fit to the phasor distributions that is observed as differentiation progresses, and becomes more evident at the terminal differentiation stages (V, VI) (Fig. 3(F)). Changes in the relative rates of the underlying biochemical pathways may lead to redistribution of bound NAD(P)H to the mitochondrial binding enzymes and substrates and thus increase the contribution of longer lifetimes [27–29]. Additional work is necessary to understand the underlying relative contributions of the metabolic pathways activated during MK differentiation and their suggested association with the detected RR and BF spatiotemporal changes. Perhaps the utilization of optics components enabling enhanced spatial resolution could further facilitate the investigation of the observed intracellular heterogeneity in more detail and especially within the finer cytoplasmic extensions of the terminal differentiation stages. This could be useful, for instance, to identify functionally heterogeneous mitochondrial populations within the cells and potentially understand in which developmental stage such populations start to acquire a functional bias and resulting spatial polarization. Automated segmentation and classification algorithms may be also developed in the future, perhaps through machine learning routines that evaluate and utilize combinations of morphological and metabolic features (e.g. cellular size, circularity, BF or RR FWHM as a means to characterize intracellular spatial metabolic heterogeneity). Nonetheless, we have shown that TPEF imaging in combination with our analytical approach enables noninvasive spatiotemporal classification of MK differentiation at the single cell level and can be utilized as a powerful tool for rapid, high-content functional characterization of heterogeneous megakaryocytic cultures.

Author contributions

D.P., L.T, A.B., D.L.K and I.G. were responsible for conceiving and designing the study, D.P., L.T were involved in data acquisition, D.P. performed the data analysis, D.P., C.A.A., Z.L. developed the data analysis algorithms, A.B. and I.G. supervised all aspects of data acquisition and analysis.

Disclosures

The authors declare that there are no conflicts of interest related to this article.

6 Appendix

Appendix figures, Figs. 5–8.

 

Fig. 5 Mitochondria support MK differentiation. 2P NAD(P)H autofluorescence and confocal mitotracker orange images from MKs at various stages of differentiation.

Download Full Size | PDF

 

Fig. 6 Image analysis steps. A. Overlay of the co-registered NAD(P)H (green) and FAD (red) channels, B. Cellular tracing and classification overlaid on RR respective map C. Automated segmentation of the selected cells after removal of background and saturated pixels. Both the overall cellular masks incorporating nuclei (magenta & white pixels) and the stricter intracellular feature masks (white pixels only) are shown. Scale bar is 20 µm for all images.

Download Full Size | PDF

 

Fig. 7 Flowchart displaying the decision-making steps utilized for the assignment of the MK differentiation classes. Small cells belonging to categories III and IV were very sparsely observed in our cultures (<5%). In those instances, as nuclear morphology and N:C ratio combinations are different between these classes and class II, misclassification is usually avoided.

Download Full Size | PDF

 

Fig. 8 A-B. Cell based RR and BF histograms from all cells examined. Box plots with outliers as dark dots outside the whiskers are shown. C-D. Cell based RR and BF binormal fitted histograms from all cells examined after outlier exclusion. The respective extracted 2-component normal mixture parameters are shown below each distribution.

Download Full Size | PDF

Funding

US National Institutes of Health (R01 EB016041), the NIH Research Infrastructure (NIH S10 OD021624) and from the Alexander S. Onassis Public Benefit Foundation and the Gerondelis Foundation.

Acknowledgements

Portions of this work were presented at the OSA Biophotonics Congress: Optics in the Life Sciences in 2017, Control ID: 2689819

References and links

1. A. E. Geddis, “Megakaryopoiesis,” Semin. Hematol. 47(3), 212–219 (2010). [CrossRef]   [PubMed]  

2. A. T. Franco, A. Corken, and J. Ware, “Platelets at the interface of thrombosis, inflammation, and cancer,” Blood 126(5), 582–588 (2015). [CrossRef]   [PubMed]  

3. K. Kaushansky, “The molecular mechanisms that control thrombopoiesis,” J. Clin. Invest. 115(12), 3339–3347 (2005). [CrossRef]   [PubMed]  

4. A. Malara, C. Gruppi, P. Rebuzzini, L. Visai, C. Perotti, R. Moratti, C. Balduini, M. E. Tira, and A. Balduini, “Megakaryocyte-matrix interaction within bone marrow: new roles for fibronectin and factor XIII-A,” Blood 117(8), 2476–2483 (2011). [CrossRef]   [PubMed]  

5. K. R. Machlus and J. E. Italiano Jr., “The incredible journey: From megakaryocyte development to platelet formation,” J. Cell Biol. 201(6), 785–796 (2013). [CrossRef]   [PubMed]  

6. S. Nakamura, N. Takayama, S. Hirata, H. Seo, H. Endo, K. Ochi, K. Fujita, T. Koike, K. Harimoto, T. Dohda, A. Watanabe, K. Okita, N. Takahashi, A. Sawaguchi, S. Yamanaka, H. Nakauchi, S. Nishimura, and K. Eto, “Expandable megakaryocyte cell lines enable clinically applicable generation of platelets from human induced pluripotent stem cells,” Cell Stem Cell 14(4), 535–548 (2014). [CrossRef]   [PubMed]  

7. J. N. Thon, D. A. Medvetz, S. M. Karlsson, and J. E. Italiano Jr., “Road blocks in making platelets for transfusion,” J. Thromb. Haemost. 13(Suppl 1), S55–S62 (2015). [CrossRef]   [PubMed]  

8. I. Georgakoudi and K. P. Quinn, “Optical imaging using endogenous contrast to assess metabolic state,” Annu. Rev. Biomed. Eng. 14(1), 351–367 (2012). [CrossRef]   [PubMed]  

9. K. P. Quinn, G. V. Sridharan, R. S. Hayden, D. L. Kaplan, K. Lee, and I. Georgakoudi, “Quantitative metabolic imaging using endogenous fluorescence to detect stem cell differentiation,” Sci. Rep. 3(1), 3432 (2013). [CrossRef]   [PubMed]  

10. A. Varone, J. Xylas, K. P. Quinn, D. Pouli, G. Sridharan, M. E. McLaughlin-Drubin, C. Alonzo, K. Lee, K. Münger, and I. Georgakoudi, “Endogenous two-photon fluorescence imaging elucidates metabolic changes related to enhanced glycolysis and glutamine consumption in precancerous epithelial tissues,” Cancer Res. 74(11), 3067–3075 (2014). [CrossRef]   [PubMed]  

11. M. C. Skala, K. M. Riching, A. Gendron-Fitzpatrick, J. Eickhoff, K. W. Eliceiri, J. G. White, and N. Ramanujam, “In vivo multiphoton microscopy of NADH and FAD redox states, fluorescence lifetimes, and cellular morphology in precancerous epithelia,” Proc. Natl. Acad. Sci. U.S.A. 104(49), 19494–19499 (2007). [CrossRef]   [PubMed]  

12. M. Y. Berezin and S. Achilefu, “Fluorescence lifetime measurements and biological imaging,” Chem. Rev. 110(5), 2641–2684 (2010). [CrossRef]   [PubMed]  

13. K. Blinova, R. L. Levine, E. S. Boja, G. L. Griffiths, Z. D. Shi, B. Ruddy, and R. S. Balaban, “Mitochondrial NADH fluorescence is enhanced by complex I binding,” Biochemistry 47(36), 9636–9645 (2008). [CrossRef]   [PubMed]  

14. T. A. Giancaspero, G. Busco, C. Panebianco, C. Carmone, A. Miccolis, G. M. Liuzzi, M. Colella, and M. Barile, “FAD synthesis and degradation in the nucleus create a local flavin cofactor pool,” J. Biol. Chem. 288(40), 29069–29080 (2013). [CrossRef]   [PubMed]  

15. A. Balduini, I. Pallotta, A. Malara, P. Lova, A. Pecci, G. Viarengo, C. L. Balduini, and M. Torti, “Adhesive receptors, extracellular proteins and myosin IIA orchestrate proplatelet formation by human megakaryocytes,” J. Thromb. Haemost. 6(11), 1900–1907 (2008). [CrossRef]   [PubMed]  

16. D. Pouli, M. Balu, C. A. Alonzo, Z. Liu, K. P. Quinn, F. Rius-Diaz, R. M. Harris, K. M. Kelly, B. J. Tromberg, and I. Georgakoudi, “Imaging mitochondrial dynamics in human skin reveals depth-dependent hypoxia and malignant potential for diagnosis,” Sci. Transl. Med. 8(367), 367ra169 (2016). [CrossRef]   [PubMed]  

17. N. Williams and R. F. Levine, “The origin, development and regulation of megakaryocytes,” Br. J. Haematol. 52(2), 173–180 (1982). [CrossRef]   [PubMed]  

18. I. Pallotta, M. Lovett, W. Rice, D. L. Kaplan, and A. Balduini, “Bone marrow osteoblastic niche: a new model to study physiological regulation of megakaryopoiesis,” PLoS One 4(12), e8359 (2009). [CrossRef]   [PubMed]  

19. M. A. Digman, V. R. Caiolfa, M. Zamai, and E. Gratton, “The phasor approach to fluorescence lifetime imaging analysis,” Biophys. J. 94(2), L14–L16 (2008). [CrossRef]   [PubMed]  

20. C. A. Alonzo, S. Karaliota, D. Pouli, Z. Liu, K. P. Karalis, and I. Georgakoudi, “Two-photon excited fluorescence of intrinsic fluorophores enables label-free assessment of adipose tissue function,” Sci. Rep. 6(1), 31012 (2016). [CrossRef]   [PubMed]  

21. E. Stuntz, Y. Gong, D. Sood, V. Liaudanskaya, D. Pouli, K. P. Quinn, C. Alonzo, Z. Liu, D. L. Kaplan, and I. Georgakoudi, “Endogenous Two-Photon Excited Fluorescence Imaging Characterizes Neuron and Astrocyte Metabolic Responses to Manganese Toxicity,” Sci. Rep. 7(1), 1041 (2017). [CrossRef]   [PubMed]  

22. V. Abbonante, C. A. Di Buduo, C. Gruppi, A. Malara, U. Gianelli, G. Celesti, A. Anselmo, L. Laghi, M. Vercellino, L. Visai, A. Iurlo, R. Moratti, G. Barosi, V. Rosti, and A. Balduini, “Thrombopoietin/TGF-β1 Loop Regulates Megakaryocyte Extracellular Matrix Component Synthesis,” Stem Cells 34(4), 1123–1133 (2016). [CrossRef]   [PubMed]  

23. H. Schachtner, S. D. Calaminus, A. Sinclair, J. Monypenny, M. P. Blundell, C. Leon, T. L. Holyoake, A. J. Thrasher, A. M. Michie, M. Vukovic, C. Gachet, G. E. Jones, S. G. Thomas, S. P. Watson, and L. M. Machesky, “Megakaryocytes assemble podosomes that degrade matrix and protrude through basement membrane,” Blood 121(13), 2542–2552 (2013). [CrossRef]   [PubMed]  

24. J. E. Italiano Jr, P. Lecine, R. A. Shivdasani, and J. H. Hartwig, “Blood platelets are assembled principally at the ends of proplatelet processes produced by differentiated megakaryocytes,” J. Cell Biol. 147(6), 1299–1312 (1999). [CrossRef]   [PubMed]  

25. A. Malara, C. Gruppi, I. Pallotta, E. Spedden, R. Tenni, M. Raspanti, D. Kaplan, M. E. Tira, C. Staii, and A. Balduini, “Extracellular matrix structure and nano-mechanics determine megakaryocyte function,” Blood 118(16), 4449–4453 (2011). [CrossRef]   [PubMed]  

26. M. Bender, J. N. Thon, A. J. Ehrlicher, S. Wu, L. Mazutis, E. Deschmann, M. Sola-Visner, J. E. Italiano, and J. H. Hartwig, “Microtubule sliding drives proplatelet elongation and is dependent on cytoplasmic dynein,” Blood 125(5), 860–868 (2015). [CrossRef]   [PubMed]  

27. D. M. Jameson, V. Thomas, and D. M. Zhou, “Time-resolved fluorescence studies on NADH bound to mitochondrial malate dehydrogenase,” Biochim. Biophys. Acta 994(2), 187–190 (1989). [CrossRef]   [PubMed]  

28. J. R. Lakowicz, H. Szmacinski, K. Nowaczyk, and M. L. Johnson, “Fluorescence lifetime imaging of free and protein-bound NADH,” Proc. Natl. Acad. Sci. U.S.A. 89(4), 1271–1275 (1992). [CrossRef]   [PubMed]  

29. C. Stringari, A. Cinquin, O. Cinquin, M. A. Digman, P. J. Donovan, and E. Gratton, “Phasor approach to fluorescence lifetime microscopy distinguishes different metabolic states of germ cells in a live tissue,” Proc. Natl. Acad. Sci. U.S.A. 108(33), 13582–13587 (2011). [CrossRef]   [PubMed]