Maximizing Kolmogorov Complexity for accurate and robust bright field cell segmentation

BACKGROUND

Analysis of cellular processes with microscopic bright field defocused imaging has the advantage of low phototoxicity and minimal sample preparation. However bright field images lack the contrast and nuclei reporting available with florescent approaches and therefore present a challenge to methods that segment and track the live cells. Moreover, such methods must be robust to systemic and random noise, variability in experimental configuration, and the multiple unknowns in the biological system under study.

RESULTS

A new method called maximal-information is introduced that applies a non-parametric information theoretic approach to segment bright field defocused images. The method utilizes a combinatorial optimization strategy to select specific defocused images from each image stack such that set complexity, a Kolmogorov complexity measure, is maximized. Differences among these selected images are then applied to initialize and guide a level set based segmentation algorithm. The performance of the method is compared with a recent approach that uses a fixed defocused image selection strategy over an image data set of embryonic kidney cells (HEK 293T) from multiple experiments. Results demonstrate that the adaptive maximal-information approach significantly improves precision and recall of segmentation over the diversity of data sets.

CONCLUSIONS

Integrating combinatorial optimization with non-parametric Kolmogorov complexity has been shown to be effective in extracting information from microscopic bright field defocused images. The approach is application independent and has the potential to be effective in processing a diversity of noisy and redundant high throughput biological data.

Stomatal responses to humidity in isolated epidermes

The ability of guard cells to hydrate and dehydrate from the surrounding air was investigated using isolated epidermes of Tradescantia pallida and Vicia faba. Stomata were found to respond to the water vapour pressure on the outside and inside of the epidermis, but the response was more sensitive to the inside vapour pressure, and occurred in the presence or absence of living, turgid epidermal cells. Experiments using helium-oxygen air showed that guard cells hydrated and dehydrated entirely from water vapour, suggesting that there was no significant transfer of water from the epidermal tissue to the guard cells. The stomatal aperture achieved at any given vapour pressure was shown to be consistent with water potential equilibrium between the guard cells and the air near the bottom of the stomatal pore, and water vapour exchange through the external cuticle appeared to be unimportant for the responses. Although stomatal responses to humidity in isolated epidermes are the result of water potential equilibrium between the guard cells and the air near the bottom of the stomatal pore, stomatal responses to humidity in leaves are unlikely to be the result of a similar equilibrium.

The role of the mesophyll in stomatal responses to light and CO2

Stomatal responses to light and CO(2) were investigated using isolated epidermes of Tradescantia pallida, Vicia faba and Pisum sativum. Stomata in leaves of T. pallida and P. sativum responded to light and CO(2), but those from V. faba did not. Stomata in isolated epidermes of all three species could be opened on KCl solutions, but they showed no response to light or CO(2). However, when isolated epidermes of T. pallida and P. sativum were placed on an exposed mesophyll from a leaf of the same species or a different species, they regained responsiveness to light and CO(2). Stomatal responses in these epidermes were similar to those in leaves in that they responded rapidly and reversibly to changes in light and CO(2). Epidermes from V. faba did not respond to light or CO(2) when placed on mesophyll from any of the three species. Experiments with single optic fibres suggest that stomata were being regulated via signals from the mesophyll produced in response to light and CO(2) rather than being sensitized to light and CO(2) by the mesophyll. The data suggest that most of the stomatal response to CO(2) and light occurs in response to a signal generated by the mesophyll.

Membrane trafficking and osmotically induced volume changes in guard cells

Guard cells rapidly adjust their plasma membrane surface area while responding to osmotically induced volume changes. Previous studies have shown that this process is associated with membrane internalization and remobilization. To investigate how guard cells maintain membrane integrity during rapid volume changes, the effects of two membrane trafficking inhibitors on the response of intact guard cells of Vicia faba to osmotic treatments were studied. Using confocal microscopy and epidermal peels, the relationship between the area of a medial paradermal guard-cell section and guard-cell volume was determined. This allowed estimates of guard-cell volume to be made from single paradermal confocal images, and therefore allowed rapid determination of volume as cells responded to osmotic treatments. Volume changes in control cells showed exponential kinetics, and it was possible to calculate an apparent value for guard-cell hydraulic conductivity from these kinetics. Wortmannin and cytochalasin D inhibited the rate of volume loss following a 0-1.5 MPa osmotic treatment. Cytochalasin D also inhibited volume increases following a change from 1.5 MPa to 0 MPa, but wortmannin had no effect. Previous studies showing that treatment with arabinanase inhibits changes in guard-cell volume in response to osmotic treatments were confirmed. However, pressure volume curves show that the effects of arabinanase and the cytochalasin D were not due to changes in cell wall elasticity. It is suggested that arabinanase, cytochalasin D, and wortmannin cause reductions in the hydraulic conductivity of the plasma membrane, possibly via gating of aquaporins. A possible role for aquaporins in co-ordinating volume changes with membrane trafficking is discussed.

Cellular calcium mobilization in response to phosphoinositide delivery

Intracellular calcium [Ca(2+)](i) is mobilized in many cell types in response to activation of phosphoinositide (PIP(n)) signaling pathways involving PtdIns(4,5)P(2) or PtdIns(3,4,5)P(3). To further explore the relationship between increases in intracellular PIP(n) concentrations and mobilization of [Ca(2+)](i), each of the seven phosphorylated phosphoinositides (PIP(n)s) were delivered into cells and the metabolism and physiological effects of the exogenously administered PIP(n)s were determined. The efficient cellular delivery of fluorophore-tagged and native PIP(n)s was accomplished using histone protein, neomycin, and dendrimeric polyamines. PtdIns(4,5)P(2) fluorophore-tagged analogs with short- and long-acyl chains were substrates for cellular enzymes in vitro and for phospholipases in stimulated fibroblasts. PtdIns(4)P, PtdIns(3,4)P(2) and PtdIns(4,5)P(2), each induced calcium mobilization rapidly after exogenous addition to fibroblasts. PtdIns(3,4,5)P(3) induced a significant, but smaller increase in intracellular calcium. These observations suggest that PIP(n)s other than PtdIns(4,5)P(2) or PtdIns(3,4,5)P(3) may have direct roles in signaling involving [Ca(2+)](i).

Metastasis Suppression by Breast Cancer Metastasis Suppressor 1 Involves Reduction of Phosphoinositide Signaling in MDA-MB-435 Breast Carcinoma Cells

Several molecules that suppress metastasis without suppressing tumorigenicity have been identified, but their mechanisms of action have not yet been determined. Many block growth at the secondary site, suggesting involvement in how cells respond to signals from the extracellular milieu. Breast cancer metastasis suppressor 1 (BRMS1)–transfected MDA-MB-435 cells were examined for modifications of phosphoinositide signaling as a potential mechanism for metastasis suppression. 435/BRMS1 cells expressed <10% of phosphatidylinositol-4, 5-bisphosphate compared with parental cells, whereas levels of the PtdIns(4)P and phosphatidylinositol-3-phosphate were unchanged. Inositol (1,4,5)-trisphosphate [Ins(1,4,5)P3] were decreased in 435/BRMS1 cells by ∼50%. Phosphatidylinositol-3,4,5-trisphosphate levels were undetectable in 435/BRMS1 cells, even when stimulated by exogenous insulin or platelet-derived growth factor. Immunofluorescence microscopy to examine cellular distribution confirmed that phosphatidylinositol-4,5-bisphosphate distribution with cells was unchanged but was uniformly decreased throughout the cell. Although the gross morphology of 435/BRMS1 cells is similar to the parent, filamentous actin was more readily apparent in 435/BRMS1. Intracellular calcium, measured using Fluo-3 and Fura-2 fluorescent calcium indicator dyes, was somewhat lower, but not statistically different in 435/BRMS1 compared with parental cell. However, when stimulated with platelet-derived growth factor, MDA-MB-435 cells, but not 435/BRMS1 cells mobilized intracellular calcium. Taken together, these results implicate signaling through phosphoinositides in the regulation of breast cancer metastasis, specifically metastasis that can be suppressed by BRMS1.

Metastasis suppression by breast cancer metastasis suppressor 1 involves reduction of phosphoinositide signaling in MDA-MB-435 breast carcinoma cells

Several molecules that suppress metastasis without suppressing tumorigenicity have been identified, but their mechanisms of action have not yet been determined. Many block growth at the secondary site, suggesting involvement in how cells respond to signals from the extracellular milieu. Breast cancer metastasis suppressor 1 (BRMS1)-transfected MDA-MB-435 cells were examined for modifications of phosphoinositide signaling as a potential mechanism for metastasis suppression. 435/BRMS1 cells expressed <10% of phosphatidylinositol-4, 5-bisphosphate compared with parental cells, whereas levels of the PtdIns(4)P and phosphatidylinositol-3-phosphate were unchanged. Inositol (1,4,5)-trisphosphate [Ins(1,4,5)P(3)] were decreased in 435/BRMS1 cells by approximately 50%. Phosphatidylinositol-3,4,5-trisphosphate levels were undetectable in 435/BRMS1 cells, even when stimulated by exogenous insulin or platelet-derived growth factor. Immunofluorescence microscopy to examine cellular distribution confirmed that phosphatidylinositol-4,5-bisphosphate distribution with cells was unchanged but was uniformly decreased throughout the cell. Although the gross morphology of 435/BRMS1 cells is similar to the parent, filamentous actin was more readily apparent in 435/BRMS1. Intracellular calcium, measured using Fluo-3 and Fura-2 fluorescent calcium indicator dyes, was somewhat lower, but not statistically different in 435/BRMS1 compared with parental cell. However, when stimulated with platelet-derived growth factor, MDA-MB-435 cells, but not 435/BRMS1 cells mobilized intracellular calcium. Taken together, these results implicate signaling through phosphoinositides in the regulation of breast cancer metastasis, specifically metastasis that can be suppressed by BRMS1.

Changes in surface area of intact guard cells are correlated with membrane internalization

Guard cells must maintain the integrity of the plasma membrane as they undergo large, rapid changes in volume. It has been assumed that changes in volume are accompanied by changes in surface area, but mechanisms for regulating plasma membrane surface area have not been identified in intact guard cells, and the extent to which surface area of the guard cells changes with volume has never been determined. The alternative hypothesis-that surface area remains approximately constant because of changes in shape-has not been investigated. To address these questions, we determined surface area for intact guard cells of Vicia faba as they underwent changes in volume in response to changes in the external osmotic potential. We also estimated membrane internalization for these cells. Epidermal peels were subjected to external solutions of varying osmotic potential to shrink and swell the guard cells. A membrane-specific fluorescent dye was used to identify the plasma membrane, and confocal microscopy was used to acquire a series of optical paradermal sections of the guard cell pair at each osmotic potential. Solid digital objects representing the guard cells were created from the membrane outlines identified in these paradermal sections, and surface area, volume, and various linear dimensions were determined for these solid objects. Surface area decreased by as much as 40% when external osmotic potential was increased from 0 to 1.5 MPa, and surface area varied linearly with volume. Membrane internalization was approximated by determining the amount of the fluorescence in the cell's interior. This value was shown to increase approximately linearly with decreases in the cell's surface area. The changes in surface area, volume, and membrane internalization were reversible when the guard cells were returned to a buffer solution with an osmotic potential of approximately zero. The data show that intact guard cells undergo changes in surface area that are too large to be accommodated by plasma membrane stretching and shrinkage and suggest that membrane is reversibly internalized to maintain cell integrity.

A real-time fluorogenic phospholipase A(2) assay for biochemical and cellular activity measurements

A fluorogenic analog of the PLA(2) substrate PC, named Dabcyl-BODIPY-PC, or simply DBPC, was synthesized with a fluorescence quencher (Dabcyl, 4-[(4-[N,N-dimethylamino]phenyl)azo]benzoic acid) in the sn-1 acyl chain and a BODIPY fluor in the sn-2 acyl chain. DBPC was recognized by sPLA(2) from each of the four sources examined (bee venom, human synovial fluid, cobra venom, and bovine pancreas). A dramatic and quantifiable fluorescence enhancement of DBPC occurred upon phospholipase digestion both in the presence and absence of excess PC. Both real-time and endpoint assays for PLA(2) were sensitive, consistent, and rapid. Thus, DBPC can be used as a sensitive fluorogenic probe for in vitro high-throughput screening assays for PLA(2) activation and inhibition and would expedite studies of PLA(2) in cellular signaling, in vitro screening for drug discovery, and subcellular localization of enzyme activity.

A monoclonal antibody to visualize PtdIns(3,4,5)P(3) in cells

Phosphatidylinositol 3,4,5-trisphosphate [PtdIns(3,4,5)P(3)] is a second messenger produced in response to agonist stimulation. Traditionally, visualization of phosphoinositide polyphosphates (PtdInsP(n)) in living cells is accomplished using chimeric green fluorescent protein (GFP)-pleckstrin homology (PH) domain proteins, while PtdInsP(n) quantitation is accomplished by extraction and separation of radiolabeled cellular PtdInsP(n)s. Here we describe preparation of a covalent protein-PtdIns(3,4,5)P(3) immunogen, characterization of binding selectivity of an anti-PtdIns(3,4,5)P(3) IgM, and immunodetection of PtdIns(3,4,5)P(3) in stimulated mammalian cells. This antibody has greater than three orders of magnitude selectivity for binding PtdIns(3,4,5)P(3) relative to its precursor, phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P(2)), and is therefore optimal for studies of cell function. The immunodetection in platelet-derived growth factor (PDGF)-stimulated NIH 3T3 cells was benchmarked against HPLC analysis of [3H]-myo-inositol-labeled cellular PtdInsP(n)s. In addition, the changes in subcellular amounts and localizations of both PtdIns(3,4,5)P(3) and PtdIns(4,5)P(2) in stimulated NIH 3T3 fibroblasts and human neutrophils were observed by immunofluorescence. In insulin- or PDGF-stimulated fibroblasts, PtdIns(3,4,5)P(3) levels increased in the cytoplasm, peaking at 10 min. In contrast, increases in the PtdIns(4,5)P(2) levels were detected in nuclei, corresponding to the production of new substrate following depletion by phosphoinositide (PI) 3-kinase.

Rapid accumulation of phosphatidylinositol 4,5-bisphosphate and inositol 1,4,5-trisphosphate correlates with calcium mobilization in salt-stressed arabidopsis

The phosphoinositide phosphatidylinositol 4,5-bisphosphate [PtdIns(4,5)P(2)] is a key signaling molecule in animal cells. It can be hydrolyzed to release 1,2-diacyglycerol and inositol 1,4,5-trisphosphate (IP(3)), which in animal cells lead to protein kinase C activation and cellular calcium mobilization, respectively. In addition to its critical roles in constitutive and regulated secretion of proteins, PtdIns(4,5)P(2) binds to proteins that modify cytoskeletal architecture and phospholipid constituents. Herein, we report that Arabidopsis plants grown in liquid media rapidly increase PtdIns(4,5)P(2) synthesis in response to treatment with sodium chloride, potassium chloride, and sorbitol. These results demonstrate that when challenged with salinity and osmotic stress, terrestrial plants respond differently than algae, yeasts, and animal cells that accumulate different species of phosphoinositides. We also show data demonstrating that whole-plant IP(3) levels increase significantly within 1 min of stress initiation, and that IP(3) levels continue to increase for more than 30 min during stress application. Furthermore, using the calcium indicators Fura-2 and Fluo-3 we show that root intracellular calcium concentrations increase in response to stress treatments. Taken together, these results suggest that in response to salt and osmotic stress, Arabidopsis uses a signaling pathway in which a small but significant portion of PtdIns(4,5)P(2) is hydrolyzed to IP(3). The accumulation of IP(3) occurs during a time frame similar to that observed for stress-induced calcium mobilization. These data also suggest that the majority of the PtdIns(4,5)P(2) synthesized in response to salt and osmotic stress may be utilized for cellular signaling events distinct from the canonical IP(3) signaling pathway.

Guard cell volume and pressure measured concurrently by confocal microscopy and the cell pressure probe

Guard cell turgor pressures in epidermal peels of broad bean (Vicia faba) were measured and controlled with a pressure probe. At the same time, images of the guard cell were acquired using confocal microscopy. To obtain a clear image of guard cell volume, a fluorescent dye that labels the plasma membrane was added to the solution bathing the epidermal peel. At each pressure, 17 to 20 optical sections (each 2 microm thick) were acquired. Out-of-focus light in these images was removed using blind deconvolution, and volume was estimated using direct linear integration. As pressure was increased from as low as 0.3 MPa to as high as 5.0 MPa, guard cell volume increased in a saturating fashion. The elastic modulus was calculated from these data and was found to range from approximately 2 to 40 MPa. The data allow inference of guard cell osmotic content from stomatal aperture and facilitate accurate mechanistic modeling of epidermal water relations and stomatal functioning.

Intracellular delivery of phosphoinositides and inositol phosphates using polyamine carriers

Phosphoinositide signaling regulates events in endocytosis and exocytosis, vesicular trafficking of proteins, transduction of extracellular signals, remodeling of the actin cytoskeleton, regulation of calcium flux, and apoptosis. Obtaining mechanistic insights in living cells is impeded by the membrane impermeability of these anionic lipids. We describe a carrier system for intracellular delivery of phosphoinositide polyphosphates (PIP(n)s) and fluorescently labeled PIP(n)s into living cells, such that intracellular localization can be directly observed. Preincubation of PIP(n)s or inositol phosphates with carrier polyamines produced complexes that entered mammalian, plant, yeast, bacterial, and protozoal cells in seconds to minutes via a nonendocytic mechanism. Time-dependent transit of both PIP(n)s and the carrier to specific cytosolic and nuclear compartments was readily visualized by fluorescence microscopy. Platelet-derived growth factor treatment of NIH 3T3 fibroblasts containing carrier-delivered phosphatidylinositol 4,5-bisphosphate [PtdIns(4, 5)P(2)]-7-nitrobenz-2-oxa-1,3-diazole resulted in the redistribution of the fluorescent signal, suggesting that fluorescent PtdIns(4, 5)P(2) was a substrate for phospholipase C. We also observed a calcium flux in NIH 3T3 cells when complexes of carrier and PtdIns(4, 5)P(2) or inositol 1,4,5-trisphosphate were added extracellularly. This simple intracellular delivery system allows for the efficient translocation of biologically active PIP(n)s, inositol phosphates, and their fluorescent derivatives into living cells in a physiologically relevant context.