Cell Volume Measurements by Optical Transmission Microscopy
UNIT 12.39
Michael A. Model1 1
Department of Biological Sciences, Kent State University, Kent, Ohio
Cell volume is an important parameter in cell adaptation to anisosmotic stress, in the development of apoptosis and necrosis, and in the pathogenesis of several diseases. This unit describes a method for measuring the volume of adherent cells using a standard light microscope. A coverslip with attached cells is placed in a shallow chamber in a medium containing a strongly absorbing and cell-impermeant dye, Acid Blue 9. When such a sample is imaged in transmitted light at a wavelength of maximum dye absorption (630 nm), the resulting contrast quantitatively reflects cell thickness. Once the thickness is known at every point, the volume can be computed as well. Technical details, interpretation of data, and possible artifacts are discussed. Measurements in absolute units require knowledge of the absorption coefficient, and a similar C 2015 procedure for the measurement of absorption coefficient is described. by John Wiley & Sons, Inc. Keywords: optical microscopy r transmission-through-dye microscopy r cell volume r cell topography r absorption r acid blue 9
How to cite this article: Model, M.A. 2015. Cell volume measurements by optical transmission microscopy. Curr. Protoc. Cytom. 72:12.39.1-12.39.9. doi: 10.1002/0471142956.cy1239s72
INTRODUCTION The volumes of cells in suspension are usually measured by electronic sizing or, less accurately, by light scattering (Nash et al., 1979; Sloot et al., 1988; UNIT 1.13). Absolute volume measurements of adherent cells have been achieved by a variety of techniques. The list, by no means complete, includes confocal sectioning (Satoh et al., 1996), atomic force microscopy (Hessler et al., 2005), side viewing (Boudreault and Grygorczyk, 2004), digital holography (Boss et al., 2013), dye exclusion in fluorescence imaging (Bottier et al., 2011), and dye exclusion in transmission (Gregg et al., 2010). The latter technique, also referred to as transmission-through-dye (TTD) microscopy, is described in this unit. It uses either a standard widefield or a laser scanning microscope in transmission mode. TTD is best suited for cells growing on a solid and transparent surface. No cell staining is required, but the method is compatible with fluorescence imaging.
IMAGE ACQUISITION The TTD method is based on exclusion of an external dye by the cell plasma membrane. Cells attached to a flat transparent surface, such as a glass coverslip, are placed in a thin chamber, and a strongly absorbing dye is added to the extracellular medium. The sample is imaged in transmitted light at a wavelength of maximal dye absorption. Because the depth of the absorbing layer is complementary to cell thickness, thicker cells or parts of cells appear brighter (Fig. 12.39.1). Image contrast thus obtained can be quantitatively converted into cell thickness and volume.
Current Protocols in Cytometry 12.39.1-12.39.9, April 2015 Published online April 2015 in Wiley Online Library (wileyonlinelibrary.com). doi: 10.1002/0471142956.cy1239s72 C 2015 John Wiley & Sons, Inc. Copyright
BASIC PROTOCOL
Cellular and Molecular Imaging
12.39.1 Supplement 72
Figure 12.39.1 (A) The principle of TTD imaging. (B) Low-resolution (×10/0.4) image of live T24 cells. (C) Highresolution (×60/1.2) image of the membrane of a formaldehyde-fixed T24 cell. Panel A is reproduced from Gregg et al. (2010) and panel C from Pelts et al. (2011), with permission.
Materials Cells Appropriate cell medium Acid Blue 9 (AB9) in sufficiently pure form can be obtained from TCI America (if using the chemical from other suppliers, verify that it is nontoxic for cells at the working concentration and that it does not change the pH of the medium) Glass slides and coverslips can be obtained from numerous vendors, such as Fisher Scientific (http://www.fishersci.com), Tedd Pella (http://www.tedpella.com/), or Electron Microscopy Sciences (https://www.emsdiasum.com/) Wet cotton swabs Standard upright or inverted microscope Bandpass filters: 630/10 (Andover, Salem, NH); 485/10 nm (Omega Optical) or similar (for strongly scattering samples, filters with antireflective coating are preferred, such as ET630-10 bp (Chroma Technology) or Brightline 632/22 from Semrock Sample preparation 1. Grow cells on a glass coverslip to ࣘ50% confluency, according to your laboratory’s protocol. 2. Completely replace normal cell medium with one containing 7 mg/ml AB9. This can be done before or after fixing the coverslip onto a slide (see step 3). One way to do it is to hold the coverslip with forceps over a paper towel and run the AB9-containing medium down its surface.
3. Apply the coverslip to a slide. To prevent cells from being squeezed, use a small amount of silicon grease as a spacer (Fig. 12.39.2). Gently press on the coverslip, at the same time removing extra liquid that may seep from under the glass, until the color of the coverslip, initially dark blue, turns lighter (with some practice, one can easily adjust the gap by eye). At this point, the sample will be sufficiently transparent to red light. If the medium needs to be added or replaced, it can be done by adding fresh liquid from the edge of the coverslip and wicking it off from the opposite side (Fig. 12.39.2).
Cell Volume Measurements
12.39.2 Supplement 72
Current Protocols in Cytometry
Figure 12.39.2 A shallow space for TTD imaging can be created by fixing a coverslip onto a slide. The cells are attached to the coverslip; the distance between the coverslip and the slide is kept stable using small spots of silicone grease. Liquid can be added or removed from the open sides.
4. Clean the top face of the coverslip with a wet cotton swab. Because the presence of AB9 slightly increases the osmolarity of the medium, allow some time (which can be determined empirically) for volume equilibration before collecting an image. NOTE: For repeated observations, cells can be kept under a coverslip. Put some extra medium on the top of the slide so that it would be touching the coverslip and keep the sample in a humidified chamber at desired temperature. In this way, drying or starvation of the cells is prevented, and they will remain healthy for many hours. Replace the medium before each observation to make sure the dye concentration in the sample is the same every time. The main drawback of this method is that it is difficult to automate. Instead of mounting a coverslip on a slide, one can use a perfusion chamber (Bioptechs) with a 30-μm spacer (Gregg et al., 2010). To enable TTD imaging in standard petri dishes, a waterimmersion objective on an upright microscope would be ideal; unfortunately, there are no commercially available objectives with sufficiently short working distances. Instead, one can use an inverted microscope and build an attachment for the condenser with a watertight horizontal window; it is slowly lowered into the dish until the gap becomes thin enough to make the cells visible under red light. Because the vertical drive for the condenser may be too coarse for precise positioning, an additional mechanism (such as a fine-pitch thread) would be helpful (Model and Schonbrun, 2013). Shallow and flat channels can also be prepared by chemical etching (Schonbrun et al., 2012).
Microscopic imaging The sample is now ready to be imaged on a microscope. The latter can be a standard upright or inverted microscope equipped additionally with a 630-nm bandpass filter. It may be useful (see Critical Parameters and Equation 4) to additionally install a 480-nm bandpass filter to enable true brightfield imaging simultaneously with TTD (AB9 does not absorb in that range). The filter can be mounted anywhere between the lamp and detector. The condenser wheel with slots for phase contrast or DIC components is an obvious choice; however, for some critical quantitative applications the filter should be placed in the “infinity space” between the objective and the tube lens, where the beam is collimated (for example, in a fluorescence filter cube). TTD imaging on a laser-scanning confocal microscope can be achieved by using a 633-nm or 635-nm laser line and a transmission detector. 5. First, focus on the cells in white or blue light; adjust the condenser (with moderate NA, not exceeding 0.5) for K¨ohler illumination, as for regular brightfield imaging; then move the 630-nm filter into the path. If the field is too dark, even at maximal lamp brightness, it may be necessary to push the coverslip closer to the slide. The sample will have a dark red background with bright red cells (Fig. 12.39.1A,B). 6. Collect an image of the sample, as well as the dark image (without light exposure).
Cellular and Molecular Imaging
12.39.3 Current Protocols in Cytometry
Supplement 72
Image processing Conversion of image intensity to thickness and volume is done using the following relationships:
h ij =
ln(Iij − Id ) ; α
V =
(h ij − h bkg ) = A(h¯ − h bkg )
A
Equation 12.39.1
where Iij is the gray level at pixel ij of the original image, Id is the dark count, and α is the absorption of the dye solution. The difference between logarithmic intensity hij and its value at the background hbkg represents local cell thickness. Summation of thicknesses over area A encompassing the cell of interest gives cell volume V; this operation is more conveniently performed by multiplying the area by the average thickness h¯ − h bkg . The specific steps needed to accomplish this computation are described below for ImageJ (http://imagej.nih.gov/ij/). 7. Convert the image to a floating 32-bit format (Image – Type – 32-bit). 8. Subtract the dark level Id (the average gray level obtained without light exposure) from the image (Image – Math – Subtract). 9. Apply the logarithm to the image (Image – Math – Log). 10. Readjust the display (Image – Adjust – Brightness/Contrast). 11. If measurements in absolute units are desired, perform two additional steps: set the correct scale (Analyze – Set Scale) and divide the image by the absorption coefficient α (Image – Math – Divide). Determination of α is described in the next section. 12. Now the gray level difference between the cells and the background represents cell thickness, and the integrated intensity over the cell area gives cell volume. To find the latter, draw a contour around the cell of interest, measure its area (Analyze – Set Measurements; choose Area and Mean gray value; then Analyze – Measure). Determine the mean gray level at the background in the vicinity of the cell. Export the results to a spreadsheet. To find the volume, multiply the area by the difference between the mean gray value of the cell and the background. The accuracy of cell segmentation (drawing the line in step 12) is not critical as long as the entire cell is selected; if some extra background area is included, it will be corrected at the background subtraction step. For sufficiently spread cells, one can select them by intensity threshold. SUPPORT PROTOCOL
MEASUREMENT OF THE ABSORPTION COEFFICIENT α Solutions containing 0.5% to 1% AB9 are optically too dense to be measured in a spectrophotometric cuvette. However, a “reverse” TTD method can be used to determine α (Model et al., 2008). While in regular TTD, a known α is used to characterize an unknown surface, one can start out with a known geometry to measure α. The following procedure can be implemented on an inverted microscope.
Materials Cell Volume Measurements
Solution of dye to be measured Glass slide Pipets
12.39.4 Supplement 72
Current Protocols in Cytometry
Figure 12.39.3 (A) Transmission image of a half-ball lens immersed in a solution of AB9. (B) To obtain a radial intensity distribution, a circle is drawn on the image (dark line) and aligned with intensity distribution. (C) A plot of ln(I-Idark ) against h and a best-fit straight line.
Half-ball lens with d = 10 mm (Edmund Optics, cat. no. 45-937) Acquire the image 1. Place a small drop (20 to 50 μl) of solution on a clean slide using a pipet. 2. Place the half-ball lens into the liquid with its spherical surface facing the slide. 3. Ensure K¨ohler illumination; focus an objective with sufficiently large working distance on the top of the slide where it touches the lens, and capture the image at 630 nm (Fig. 12.39.3A).
Process the image 4. Draw a circle and align it as best as possible with an intensity isoline by using an appropriate contrast stretch (Fig. 12.39.3B). Apply Radial Intensity plugin (http://imagej.nih.gov/ij/plugins/radial-profile.html). 5. Export the data to Excel. Convert the radial distance x to depth h using the formula:
h= R−
R2 − x 2
Equation 12.39.2
where R = 5000 μm (the radius of the half-ball lens). Convert intensity into ln(I − Idark ); Idark is determined by collecting an image without opening the shutter. Construct a scatter plot of the logarithm of intensity versus depth and fit the data to a linear function (may have to exclude points corresponding to very small depths, where errors can be large due to imperfections of the surfaces). The slope of the line gives α (Fig. 12.39.3C). The absorption of a 7 mg/ml solution of AB9 at 630 nm should be close to 0.155 μm−1 .
COMMENTARY Background Information There can be numerous reasons to measure cell volume. Cell volume is an important player in monovalent ion balance and in cell adaptation to anisosmotic environment (Chamberlin and Strange, 1989; Okada, 2004). Strong evidence exists that it is actively regulated and supplies feedback into the
signaling chain (Schliess et al., 2007; Hoffmann et al., 2009). Disturbances in the cell volume are linked to the pathogenesis of various health disorders, such as edema, glaucoma, complications of diabetes, hypertension and liver diseases (McManus et al., 1995; Lang et al., 1998). Shrinkage is an essential determinant of apoptotic cell death (G´omez-Angelats
Cellular and Molecular Imaging
12.39.5 Current Protocols in Cytometry
Supplement 72
Cell Volume Measurements
and Cidlowski, 2002) and, moreover, the volume behavior constitutes probably the main difference between apoptosis and necrosis: necrotic cells swell and eventually burst, while apoptotic cells shrink, split into fragments and become phagocytosed. Valuable information can also be obtained by combining volume measurements with other techniques. The average concentration of intracellular fluorophore is a major parameter in nonratiometric ion imaging, and the most straightforward way to obtain it is to divide the total wide-field fluorescence signal by the cell volume (UNIT 10.14). While in some instances (e.g., swelling or shrinkage in anisosmotic medium), changes in the cell volume can be safely attributed to redistribution of intracellular water, the relationship is not always that clear. Cell growth or division is one obvious example (Grebien et al., 2005); apoptotic shrinkage, too, can be caused either by dehydration or by dissociation of apoptotic bodies (Model, 2014). To distinguish between these possibilities, intracellular water must be measured along with the volume. Water can be evaluated through the average refractive index of the cell by quantitative phase microscopy combined with volume measurements (Model and Schonbrun, 2013). The TTD method described in this protocol can be assessed as follows. Its advantages include good accuracy and vertical resolution (see below), minimal cost and insensitivity to lamp instability (the reason for the latter is that fluctuations in the lamp would affect the cell and the background identically, and that would be automatically corrected at the background subtraction step). Because all the volume information is contained in a single image, image acquisition is fast. The samples are not subjected to intense illumination, and therefore the method is well suited for prolonged time-lapse experiments. The presence of the dye in the medium does not preclude fluorescence observation from the side of the chamber to which the cells are adhered (however, the amount of green excitation from the mercury lamp should be limited because it may cause heating of the sample). Additionally, the TTD method adds a strong and quantitative contrast to transparent specimens, which can be used for detailed characterization of cellular morphology. TTD is limited to cells with intact plasma membranes (Pelts et al., 2011); cells with leaky membranes are easily identifiable on images, as they appear darker than the background.
The main limitation of the method is the requirement for a thin chamber. Also, high vertical resolution comes at the expense of a limited range of thicknesses resolvable in a single image (discussed below). Because of the need for background correction, TTD cannot be used on tissues or on confluent cultures. In these cases, confocal scanning might be the simplest option; however, to compute the cell volume from a confocal scan one needs an appropriate software (Yu et al., 1994; Hevia et al., 2011) and the axial scaling effect may have to be taken into account as well (Gibson and Lanni, 1992). The dye AB9, although nontoxic, easily stains hands and clothes (it comes off with soap).
Critical Parameters and Troubleshooting The food colorant AB9 has a strong absorption at 630 nm and, at concentrations below 10 mg/ml, is very well tolerated by all cell types tested so far. Even during prolonged incubations, it does not seem to affect cell functioning or accumulate in the cytoplasm in amounts that would significantly affect the measurements. Nevertheless, it would be prudent to check for the dye compatibility when embarking on a new project. In addition to AB9, at least two other dyes have been tried: Patent Blue V (Gregg et al., 2010) and a mixture of tartrazine and Allura Red (Schonbrun et al., 2013). The vertical resolution h of the TTD method can be estimated as h ≈
1 α(S/N )
Equation 12.39.3
where S/N stands for the signal-to-noise ratio of the intensity measurement per pixel. Therefore, if α = 0.16 μm−1 and S/N = 100, one should be able to resolve less than 0.1 μm in the vertical dimension. The main factor contributing to possible errors in thickness measurements is the contrast of brighfield origin, which results from out-of-focus boundaries separating regions with different refractive indices (e.g., intracellular vesicles). Fortunately, intracellular inclusions, while slightly distorting local thickness readings, cause minimal errors in volume measurements because the brighter and darker rings resulting from these artifacts cancel each other out in summation (Eq. 12.39.1). Accurate measurements require that the image background is sufficiently above the dark
12.39.6 Supplement 72
Current Protocols in Cytometry
Figure 12.39.4 Refraction may cause incorrect volume measurements. (A) Double refraction on a loosely attached spherical cell with refractive index higher than that of the medium will result in a shortened path through the dye (d ’ < d) and underestimation of the cell volume. (B) A single refraction on a tightly attached cell will not cause such a problem, even in the presence of a refractive index difference.
level, while the highest point of the cell of interest is still kept below saturation. The height range (typically 10 to 20 μm with 7 mg/ml AB9) can be controlled to some extent by varying the exposure time and the lamp brightness. Very high vertical resolution of the membrane is possible with formaldehydefixed cells (which remain impermeable to AB9 as long as formaldehyde remains in the medium; Pelts et al., 2011) because then the concentrations of AB9 can be increased many fold to obtain detailed membrane topography (Fig. 12.39.1C). Although the calculations of thickness assume vertical light travel, light in the real microscope propagates as a cone. However, for NA 0.4, the marginal rays, after crossing from air into water, will travel the distance only by 5% longer than the vertical line, and this constant factor will be accounted for by using the same optics in absorption measurements. It is important that only the smallest NA between the objective and the condenser plays a role because the method depends on the direct, and not on diffracted wave. Thus, as long as NAobjective > NAcondenser , any objective is expected to produce the same result (Model, 2012). Dehydrated cells with the average refractive index far exceeding that of the medium may look artificially darker due to scattering of light away from the objective. This effect can be corrected in two ways. The easiest one is to collect a control brightfield image through the blue 480-nm filter and perform the same operations 1 to 6 on this image. The “volume”
calculated on the control image (its value will be negative) is subtracted from the volume measured at 630 nm: V = V630 − V480 Equation 12.39.4
This does achieve accurate correction, but only when the magnitude of V480 is significantly less than that of V630 (
UNIT 12.39
Michael A. Model1 1
Department of Biological Sciences, Kent State University, Kent, Ohio
Cell volume is an important parameter in cell adaptation to anisosmotic stress, in the development of apoptosis and necrosis, and in the pathogenesis of several diseases. This unit describes a method for measuring the volume of adherent cells using a standard light microscope. A coverslip with attached cells is placed in a shallow chamber in a medium containing a strongly absorbing and cell-impermeant dye, Acid Blue 9. When such a sample is imaged in transmitted light at a wavelength of maximum dye absorption (630 nm), the resulting contrast quantitatively reflects cell thickness. Once the thickness is known at every point, the volume can be computed as well. Technical details, interpretation of data, and possible artifacts are discussed. Measurements in absolute units require knowledge of the absorption coefficient, and a similar C 2015 procedure for the measurement of absorption coefficient is described. by John Wiley & Sons, Inc. Keywords: optical microscopy r transmission-through-dye microscopy r cell volume r cell topography r absorption r acid blue 9
How to cite this article: Model, M.A. 2015. Cell volume measurements by optical transmission microscopy. Curr. Protoc. Cytom. 72:12.39.1-12.39.9. doi: 10.1002/0471142956.cy1239s72
INTRODUCTION The volumes of cells in suspension are usually measured by electronic sizing or, less accurately, by light scattering (Nash et al., 1979; Sloot et al., 1988; UNIT 1.13). Absolute volume measurements of adherent cells have been achieved by a variety of techniques. The list, by no means complete, includes confocal sectioning (Satoh et al., 1996), atomic force microscopy (Hessler et al., 2005), side viewing (Boudreault and Grygorczyk, 2004), digital holography (Boss et al., 2013), dye exclusion in fluorescence imaging (Bottier et al., 2011), and dye exclusion in transmission (Gregg et al., 2010). The latter technique, also referred to as transmission-through-dye (TTD) microscopy, is described in this unit. It uses either a standard widefield or a laser scanning microscope in transmission mode. TTD is best suited for cells growing on a solid and transparent surface. No cell staining is required, but the method is compatible with fluorescence imaging.
IMAGE ACQUISITION The TTD method is based on exclusion of an external dye by the cell plasma membrane. Cells attached to a flat transparent surface, such as a glass coverslip, are placed in a thin chamber, and a strongly absorbing dye is added to the extracellular medium. The sample is imaged in transmitted light at a wavelength of maximal dye absorption. Because the depth of the absorbing layer is complementary to cell thickness, thicker cells or parts of cells appear brighter (Fig. 12.39.1). Image contrast thus obtained can be quantitatively converted into cell thickness and volume.
Current Protocols in Cytometry 12.39.1-12.39.9, April 2015 Published online April 2015 in Wiley Online Library (wileyonlinelibrary.com). doi: 10.1002/0471142956.cy1239s72 C 2015 John Wiley & Sons, Inc. Copyright
BASIC PROTOCOL
Cellular and Molecular Imaging
12.39.1 Supplement 72
Figure 12.39.1 (A) The principle of TTD imaging. (B) Low-resolution (×10/0.4) image of live T24 cells. (C) Highresolution (×60/1.2) image of the membrane of a formaldehyde-fixed T24 cell. Panel A is reproduced from Gregg et al. (2010) and panel C from Pelts et al. (2011), with permission.
Materials Cells Appropriate cell medium Acid Blue 9 (AB9) in sufficiently pure form can be obtained from TCI America (if using the chemical from other suppliers, verify that it is nontoxic for cells at the working concentration and that it does not change the pH of the medium) Glass slides and coverslips can be obtained from numerous vendors, such as Fisher Scientific (http://www.fishersci.com), Tedd Pella (http://www.tedpella.com/), or Electron Microscopy Sciences (https://www.emsdiasum.com/) Wet cotton swabs Standard upright or inverted microscope Bandpass filters: 630/10 (Andover, Salem, NH); 485/10 nm (Omega Optical) or similar (for strongly scattering samples, filters with antireflective coating are preferred, such as ET630-10 bp (Chroma Technology) or Brightline 632/22 from Semrock Sample preparation 1. Grow cells on a glass coverslip to ࣘ50% confluency, according to your laboratory’s protocol. 2. Completely replace normal cell medium with one containing 7 mg/ml AB9. This can be done before or after fixing the coverslip onto a slide (see step 3). One way to do it is to hold the coverslip with forceps over a paper towel and run the AB9-containing medium down its surface.
3. Apply the coverslip to a slide. To prevent cells from being squeezed, use a small amount of silicon grease as a spacer (Fig. 12.39.2). Gently press on the coverslip, at the same time removing extra liquid that may seep from under the glass, until the color of the coverslip, initially dark blue, turns lighter (with some practice, one can easily adjust the gap by eye). At this point, the sample will be sufficiently transparent to red light. If the medium needs to be added or replaced, it can be done by adding fresh liquid from the edge of the coverslip and wicking it off from the opposite side (Fig. 12.39.2).
Cell Volume Measurements
12.39.2 Supplement 72
Current Protocols in Cytometry
Figure 12.39.2 A shallow space for TTD imaging can be created by fixing a coverslip onto a slide. The cells are attached to the coverslip; the distance between the coverslip and the slide is kept stable using small spots of silicone grease. Liquid can be added or removed from the open sides.
4. Clean the top face of the coverslip with a wet cotton swab. Because the presence of AB9 slightly increases the osmolarity of the medium, allow some time (which can be determined empirically) for volume equilibration before collecting an image. NOTE: For repeated observations, cells can be kept under a coverslip. Put some extra medium on the top of the slide so that it would be touching the coverslip and keep the sample in a humidified chamber at desired temperature. In this way, drying or starvation of the cells is prevented, and they will remain healthy for many hours. Replace the medium before each observation to make sure the dye concentration in the sample is the same every time. The main drawback of this method is that it is difficult to automate. Instead of mounting a coverslip on a slide, one can use a perfusion chamber (Bioptechs) with a 30-μm spacer (Gregg et al., 2010). To enable TTD imaging in standard petri dishes, a waterimmersion objective on an upright microscope would be ideal; unfortunately, there are no commercially available objectives with sufficiently short working distances. Instead, one can use an inverted microscope and build an attachment for the condenser with a watertight horizontal window; it is slowly lowered into the dish until the gap becomes thin enough to make the cells visible under red light. Because the vertical drive for the condenser may be too coarse for precise positioning, an additional mechanism (such as a fine-pitch thread) would be helpful (Model and Schonbrun, 2013). Shallow and flat channels can also be prepared by chemical etching (Schonbrun et al., 2012).
Microscopic imaging The sample is now ready to be imaged on a microscope. The latter can be a standard upright or inverted microscope equipped additionally with a 630-nm bandpass filter. It may be useful (see Critical Parameters and Equation 4) to additionally install a 480-nm bandpass filter to enable true brightfield imaging simultaneously with TTD (AB9 does not absorb in that range). The filter can be mounted anywhere between the lamp and detector. The condenser wheel with slots for phase contrast or DIC components is an obvious choice; however, for some critical quantitative applications the filter should be placed in the “infinity space” between the objective and the tube lens, where the beam is collimated (for example, in a fluorescence filter cube). TTD imaging on a laser-scanning confocal microscope can be achieved by using a 633-nm or 635-nm laser line and a transmission detector. 5. First, focus on the cells in white or blue light; adjust the condenser (with moderate NA, not exceeding 0.5) for K¨ohler illumination, as for regular brightfield imaging; then move the 630-nm filter into the path. If the field is too dark, even at maximal lamp brightness, it may be necessary to push the coverslip closer to the slide. The sample will have a dark red background with bright red cells (Fig. 12.39.1A,B). 6. Collect an image of the sample, as well as the dark image (without light exposure).
Cellular and Molecular Imaging
12.39.3 Current Protocols in Cytometry
Supplement 72
Image processing Conversion of image intensity to thickness and volume is done using the following relationships:
h ij =
ln(Iij − Id ) ; α
V =
(h ij − h bkg ) = A(h¯ − h bkg )
A
Equation 12.39.1
where Iij is the gray level at pixel ij of the original image, Id is the dark count, and α is the absorption of the dye solution. The difference between logarithmic intensity hij and its value at the background hbkg represents local cell thickness. Summation of thicknesses over area A encompassing the cell of interest gives cell volume V; this operation is more conveniently performed by multiplying the area by the average thickness h¯ − h bkg . The specific steps needed to accomplish this computation are described below for ImageJ (http://imagej.nih.gov/ij/). 7. Convert the image to a floating 32-bit format (Image – Type – 32-bit). 8. Subtract the dark level Id (the average gray level obtained without light exposure) from the image (Image – Math – Subtract). 9. Apply the logarithm to the image (Image – Math – Log). 10. Readjust the display (Image – Adjust – Brightness/Contrast). 11. If measurements in absolute units are desired, perform two additional steps: set the correct scale (Analyze – Set Scale) and divide the image by the absorption coefficient α (Image – Math – Divide). Determination of α is described in the next section. 12. Now the gray level difference between the cells and the background represents cell thickness, and the integrated intensity over the cell area gives cell volume. To find the latter, draw a contour around the cell of interest, measure its area (Analyze – Set Measurements; choose Area and Mean gray value; then Analyze – Measure). Determine the mean gray level at the background in the vicinity of the cell. Export the results to a spreadsheet. To find the volume, multiply the area by the difference between the mean gray value of the cell and the background. The accuracy of cell segmentation (drawing the line in step 12) is not critical as long as the entire cell is selected; if some extra background area is included, it will be corrected at the background subtraction step. For sufficiently spread cells, one can select them by intensity threshold. SUPPORT PROTOCOL
MEASUREMENT OF THE ABSORPTION COEFFICIENT α Solutions containing 0.5% to 1% AB9 are optically too dense to be measured in a spectrophotometric cuvette. However, a “reverse” TTD method can be used to determine α (Model et al., 2008). While in regular TTD, a known α is used to characterize an unknown surface, one can start out with a known geometry to measure α. The following procedure can be implemented on an inverted microscope.
Materials Cell Volume Measurements
Solution of dye to be measured Glass slide Pipets
12.39.4 Supplement 72
Current Protocols in Cytometry
Figure 12.39.3 (A) Transmission image of a half-ball lens immersed in a solution of AB9. (B) To obtain a radial intensity distribution, a circle is drawn on the image (dark line) and aligned with intensity distribution. (C) A plot of ln(I-Idark ) against h and a best-fit straight line.
Half-ball lens with d = 10 mm (Edmund Optics, cat. no. 45-937) Acquire the image 1. Place a small drop (20 to 50 μl) of solution on a clean slide using a pipet. 2. Place the half-ball lens into the liquid with its spherical surface facing the slide. 3. Ensure K¨ohler illumination; focus an objective with sufficiently large working distance on the top of the slide where it touches the lens, and capture the image at 630 nm (Fig. 12.39.3A).
Process the image 4. Draw a circle and align it as best as possible with an intensity isoline by using an appropriate contrast stretch (Fig. 12.39.3B). Apply Radial Intensity plugin (http://imagej.nih.gov/ij/plugins/radial-profile.html). 5. Export the data to Excel. Convert the radial distance x to depth h using the formula:
h= R−
R2 − x 2
Equation 12.39.2
where R = 5000 μm (the radius of the half-ball lens). Convert intensity into ln(I − Idark ); Idark is determined by collecting an image without opening the shutter. Construct a scatter plot of the logarithm of intensity versus depth and fit the data to a linear function (may have to exclude points corresponding to very small depths, where errors can be large due to imperfections of the surfaces). The slope of the line gives α (Fig. 12.39.3C). The absorption of a 7 mg/ml solution of AB9 at 630 nm should be close to 0.155 μm−1 .
COMMENTARY Background Information There can be numerous reasons to measure cell volume. Cell volume is an important player in monovalent ion balance and in cell adaptation to anisosmotic environment (Chamberlin and Strange, 1989; Okada, 2004). Strong evidence exists that it is actively regulated and supplies feedback into the
signaling chain (Schliess et al., 2007; Hoffmann et al., 2009). Disturbances in the cell volume are linked to the pathogenesis of various health disorders, such as edema, glaucoma, complications of diabetes, hypertension and liver diseases (McManus et al., 1995; Lang et al., 1998). Shrinkage is an essential determinant of apoptotic cell death (G´omez-Angelats
Cellular and Molecular Imaging
12.39.5 Current Protocols in Cytometry
Supplement 72
Cell Volume Measurements
and Cidlowski, 2002) and, moreover, the volume behavior constitutes probably the main difference between apoptosis and necrosis: necrotic cells swell and eventually burst, while apoptotic cells shrink, split into fragments and become phagocytosed. Valuable information can also be obtained by combining volume measurements with other techniques. The average concentration of intracellular fluorophore is a major parameter in nonratiometric ion imaging, and the most straightforward way to obtain it is to divide the total wide-field fluorescence signal by the cell volume (UNIT 10.14). While in some instances (e.g., swelling or shrinkage in anisosmotic medium), changes in the cell volume can be safely attributed to redistribution of intracellular water, the relationship is not always that clear. Cell growth or division is one obvious example (Grebien et al., 2005); apoptotic shrinkage, too, can be caused either by dehydration or by dissociation of apoptotic bodies (Model, 2014). To distinguish between these possibilities, intracellular water must be measured along with the volume. Water can be evaluated through the average refractive index of the cell by quantitative phase microscopy combined with volume measurements (Model and Schonbrun, 2013). The TTD method described in this protocol can be assessed as follows. Its advantages include good accuracy and vertical resolution (see below), minimal cost and insensitivity to lamp instability (the reason for the latter is that fluctuations in the lamp would affect the cell and the background identically, and that would be automatically corrected at the background subtraction step). Because all the volume information is contained in a single image, image acquisition is fast. The samples are not subjected to intense illumination, and therefore the method is well suited for prolonged time-lapse experiments. The presence of the dye in the medium does not preclude fluorescence observation from the side of the chamber to which the cells are adhered (however, the amount of green excitation from the mercury lamp should be limited because it may cause heating of the sample). Additionally, the TTD method adds a strong and quantitative contrast to transparent specimens, which can be used for detailed characterization of cellular morphology. TTD is limited to cells with intact plasma membranes (Pelts et al., 2011); cells with leaky membranes are easily identifiable on images, as they appear darker than the background.
The main limitation of the method is the requirement for a thin chamber. Also, high vertical resolution comes at the expense of a limited range of thicknesses resolvable in a single image (discussed below). Because of the need for background correction, TTD cannot be used on tissues or on confluent cultures. In these cases, confocal scanning might be the simplest option; however, to compute the cell volume from a confocal scan one needs an appropriate software (Yu et al., 1994; Hevia et al., 2011) and the axial scaling effect may have to be taken into account as well (Gibson and Lanni, 1992). The dye AB9, although nontoxic, easily stains hands and clothes (it comes off with soap).
Critical Parameters and Troubleshooting The food colorant AB9 has a strong absorption at 630 nm and, at concentrations below 10 mg/ml, is very well tolerated by all cell types tested so far. Even during prolonged incubations, it does not seem to affect cell functioning or accumulate in the cytoplasm in amounts that would significantly affect the measurements. Nevertheless, it would be prudent to check for the dye compatibility when embarking on a new project. In addition to AB9, at least two other dyes have been tried: Patent Blue V (Gregg et al., 2010) and a mixture of tartrazine and Allura Red (Schonbrun et al., 2013). The vertical resolution h of the TTD method can be estimated as h ≈
1 α(S/N )
Equation 12.39.3
where S/N stands for the signal-to-noise ratio of the intensity measurement per pixel. Therefore, if α = 0.16 μm−1 and S/N = 100, one should be able to resolve less than 0.1 μm in the vertical dimension. The main factor contributing to possible errors in thickness measurements is the contrast of brighfield origin, which results from out-of-focus boundaries separating regions with different refractive indices (e.g., intracellular vesicles). Fortunately, intracellular inclusions, while slightly distorting local thickness readings, cause minimal errors in volume measurements because the brighter and darker rings resulting from these artifacts cancel each other out in summation (Eq. 12.39.1). Accurate measurements require that the image background is sufficiently above the dark
12.39.6 Supplement 72
Current Protocols in Cytometry
Figure 12.39.4 Refraction may cause incorrect volume measurements. (A) Double refraction on a loosely attached spherical cell with refractive index higher than that of the medium will result in a shortened path through the dye (d ’ < d) and underestimation of the cell volume. (B) A single refraction on a tightly attached cell will not cause such a problem, even in the presence of a refractive index difference.
level, while the highest point of the cell of interest is still kept below saturation. The height range (typically 10 to 20 μm with 7 mg/ml AB9) can be controlled to some extent by varying the exposure time and the lamp brightness. Very high vertical resolution of the membrane is possible with formaldehydefixed cells (which remain impermeable to AB9 as long as formaldehyde remains in the medium; Pelts et al., 2011) because then the concentrations of AB9 can be increased many fold to obtain detailed membrane topography (Fig. 12.39.1C). Although the calculations of thickness assume vertical light travel, light in the real microscope propagates as a cone. However, for NA 0.4, the marginal rays, after crossing from air into water, will travel the distance only by 5% longer than the vertical line, and this constant factor will be accounted for by using the same optics in absorption measurements. It is important that only the smallest NA between the objective and the condenser plays a role because the method depends on the direct, and not on diffracted wave. Thus, as long as NAobjective > NAcondenser , any objective is expected to produce the same result (Model, 2012). Dehydrated cells with the average refractive index far exceeding that of the medium may look artificially darker due to scattering of light away from the objective. This effect can be corrected in two ways. The easiest one is to collect a control brightfield image through the blue 480-nm filter and perform the same operations 1 to 6 on this image. The “volume”
calculated on the control image (its value will be negative) is subtracted from the volume measured at 630 nm: V = V630 − V480 Equation 12.39.4
This does achieve accurate correction, but only when the magnitude of V480 is significantly less than that of V630 (
