Skip to main content
NCBI home page
PLOS One logoLink to PLOS One
. 2020 Sep 30;15(9):e0238955. doi: 10.1371/journal.pone.0238955

Improved 3D cellular morphometry of Caenorhabditis elegans embryos using a refractive index matching medium

Rain Xiong 1,2, Kenji Sugioka 1,2,*
Editor: Bob Goldstein3
PMCID: PMC7526913  PMID: 32997668

Abstract

Cell shape change is one of the driving forces of animal morphogenesis, and the model organism Caenorhabditis elegans has played a significant role in analyzing the underlying mechanisms involved. The analysis of cell shape change requires quantification of cellular shape descriptors, a method known as cellular morphometry. However, standard C. elegans live imaging methods limit the capability of cellular morphometry in 3D, as spherical aberrations generated by samples and the surrounding medium misalign optical paths. Here, we report a 3D live imaging method for C. elegans embryos that minimized spherical aberrations caused by refractive index (RI) mismatch. We determined the composition of a refractive index matching medium (RIMM) for C. elegans live imaging. The 3D live imaging with the RIMM resulted in a higher signal intensity in the deeper cell layers. We also found that the obtained images improved the 3D cell segmentation quality. Furthermore, our 3D cellular morphometry and 2D cell shape simulation indicated that the germ cell precursor P4 had exceptionally high cortical tension. Our results demonstrate that the RIMM is a cost-effective solution to improve the 3D cellular morphometry of C. elegans. The application of this method should facilitate understanding of C. elegans morphogenesis from the perspective of cell shape changes.

Introduction

Live imaging of Caenorhabditis elegans (C. elegans) embryos plays a paramount role in dissecting questions pertaining to both cell and developmental biology. However, widely used C. elegans live imaging methods with conventional confocal microscopes fails to capture entire embryos and embryonic cell sets at high resolution presumably due to spherical aberrations. Spherical aberrations misalign the optical paths, resulting in the loss of signal intensity and resolution. This effect increases with sample thickness, limiting the capacity of 3D cellular morphometry [1]. A major source of spherical aberration is the refractive index (RI) mismatch among three different materials: the lens immersion medium, samples, and the medium surrounding the samples. In typical C. elegans live imaging, samples are mounted onto 2%-4% agarose and are bathed in M9 or egg salt buffer [2, 3], or bathed in these buffers without agarose [4]. These commonly used aqueous media (RI: ~1.33) should create the RI mismatch with both lens immersion media (RI: 1.52 for oil immersion) and C. elegans embryos (RI: 1.33–1.37; [5]).

In this study, we focused on the cost-effective method to reduce the RI mismatch that can be immediately adopted by the broad C. elegans research community. It is known that the use of water (RI: 1.33) and silicone (RI: 1.4) as lens-immersion media reduces the RI mismatch, but high-numerical aperture (NA) objective lenses designed for water- and silicone-immersion typically cost more than 10,000 USD. The use of non-conventional microscopy systems such as dual-view plane illumination microscopy [6] and lattice light-sheet microscopy [7] allow rapid and high-resolution volumetric imaging. However, these are also currently not cost-effective solutions.

We hypothesized that a refractive index matching medium (RIMM), which has a RI close to that of the specimen, reduces spherical aberration during C. elegans 3D live imaging with conventional confocal microscopy. Recently, Boothe et al. identified iodixanol solution as a RIMM suitable for live imaging of cultured cells, zebrafish embryos, and planarians [8]. Iodixanol is a transparent solution with a high RI of 1.429 at a 60% stock solution. Booth et al. diluted 60% iodixanol solution to identify the RIMM for each sample. However, the RIMM for C. elegans has not been determined.

In this study, we determined the composition of a RIMM for C. elegans live imaging. To analyze the RI mismatch between C. elegans samples and the surrounding medium, we imaged embryos and adult heads in different iodixanol concentrations using Nomarski differential interference contrast microscopy (DIC), a common transmitted light microscopy method used by worm researchers. Using the obtained RIMM, we performed 3D volumetric imaging and 3D watershed segmentation to generate cellular surface models. These models were then used to perform 3D cellular morphometry of individual embryonic blastomeres.

Materials and methods

C. elegans culture and strains

All strains used in this study were cultured using standard methods [9]. All worms were grown at 25°C and imaged at 22.5°C. The following integrated transgenic lines were used: cpIs56 for TagRFP::PH [10] and ruIs32 [11] for GFP::H2B.

Microscopy

All embryos were dissected in an egg salt buffer from gravid hermaphrodites [12]. For live imaging, embryos were transferred by a mouth pipette into the egg salt buffer containing different iodixanol concentrations [13]. We obtained similar results by directly dissecting gravid hermaphrodites in the iodixanol solution. Owing to the high density of iodixanol solutions, embryos tend to float in the medium. To mitigate this problem and situate embryos as close as the coverslips as possible, we used polystyrene beads with embryo thickness as spacers between coverslips and glass slides (30 μm diameter, Kisker Biotech GmbH & Co). C. elegans adult samples were also prepared with the beads. Nomarski DIC imaging was performed with generic CCD cameras mounted on either AxioPlan or AxioSkop compound microscopes (Zeiss). Confocal images used to generate a right graph in Fig 2D and a graph in Fig 2E were captured with a spinning-disk confocal unit CSU-W1 (Yokogawa) and a scientific CMOS camera Prime 95B (Photometrics) mounted on an inverted microscope Olympus IX83 (Olympus), controlled by Cellsense Dimension (Olympus). Samples were illuminated by a diode-pumped laser with 561 nm wavelength. Z-stacks were obtained with a silicone-immersion UPLSAPO60XS2 objective lens (60X, NA = 1.3; Olympus) at 0.5-μm Z-step and with a silicone-immersion oil (RI: 1.41 at 23°C; Z81114, Olympus). The rest of confocal images were captured with a spinning-disk confocal unit CSU-W1 with Borealis (Andor Technology) and dual EMCCD camera iXon Ultra 897 (Andor Technology) mounted on an inverted microscope Leica DMi8 (Leica Microsystems), controlled by Metamorph (Molecular Devices). Samples were illuminated by diode-pumped lasers with 488 nm and 561 nm wavelengths. Z-stacks were obtained with an oil-immersion PL APO objective lens (63X, NA = 1.4; Leica) at 0.5-μm Z-step and with an immersion oil (type F, RI = 1.51; Leica).

Fig 2. The refractive index matching medium improves the signal intensity in the deeper cell layers.

Fig 2

Open in a new tab

(A) Schematic drawings of the imaging setup. Embryos and embryo-sized polystyrene beads were sandwiched between coverslips and glass slides to prevent embryo flotation due to the higher density of iodixanol solution. (B, C) Depth-induced blurring of the cell-cell boundary. Spinning-disk confocal images of 24-cell stage embryos (B). Magenta, TagRFP-PH (plasma membrane); green, histone H2B-GFP (chromosomes). Plasma membrane signal at the cell-cell boundary (C). Each raw image in C, which captures a cell-cell boundary in 4.96 μm x 4.96 μm area, was converted into binary images using Otsu’s method. (D) Intensity profiles of cell boundaries. A line passing a cell-boundary at 25-μm depth was used to measure intensity profiles of TagRFP-PH. The background level was adjusted to zero (see S4A Fig for detail). Error bars are standard errors of means. (E) Peak Tag RFP-PH signal intensity at 25-μm depth in different iodixanol concentrations. Error bars are 95% confidence intervals. Statistical test was performed by comparing samples treated with 30% iodixanol and others. P-values were calculated by one-way ANOVA with the Holm-Sidak’s multiple comparison test. n.s. (P > 0.05), ** (P < 0.001), and *** (P < 0.0001). All depths shown are relative to the cell surface closest to the objective lens. Scale bars, 10 μm.

Determination of a RIMM for live C. elegans

A 60% iodixanol solution (Optiprep; D1556, Millipore Sigma) was sequentially diluted with egg salt buffer. Embryos at mixed stages and adult head regions in different iodixanol concentrations were then imaged with a 10X Plan-Apochromat objective lens (NA = 0.45; Zeiss). Using Fiji [14], we drew the regions of interest (ROI) corresponding to the whole embryos and adult heads and quantified the signal intensities. The maximum gray values were divided by the minimum gray values to calculate the image contrast. The RI of commonly used C. elegans buffers were measured at 22.5°C using a digital refractometer (13940000, Reichert) at a wavelength of 598 nm. Egg salt buffer, M9 buffer, and Shelton’s growth medium were prepared as described before [12, 15].

Eggshell penetration assay

To test whether molecules with the size of iodixanol can pass the eggshell, we mounted wild-type embryos in the 30% iodixanol containing 0.05 U/μl Alexa Fluor 568 Phalloidin (1590 Da; A12380, Invitrogen) and polystyrene beads. Immediately after the preparation of the sample slide as described above, confocal and brightfield images were captured.

Quantification of the signal intensity

For the generation of the binary images shown in Fig 2C, we used Otsu’s thresholding method with Fiji [16]. To calculate the signal intensity of TagRFP-PH in each image, we drew a 4.96 μm line passing the cell-cell boundary and generated a plot profile of gray value (S4A Fig). Mean signal intensity of the first 1 μm length, corresponding to the cytoplasmic region, was used as a background value. Signal intensity was then calculated by the following equation.

[Signalintensity]=[grayvalue][backgroundvalue]−1

Each dot in Fig 2E represents the maximum signal intensity within a cell-cell boundary. Note that four to five cell-cell boundaries were measured for each embryo.

3D segmentation and generation of cellular surface models

To perform 3D segmentation of embryos, source images (TagRFP::PH; 0.16 μm/pixel and 0.5-μm voxel depth) were processed using Fiji. First, images were processed with the Attenuation Correction plug-in ([17]; opening radius = 3.0, reference slice was set to midplane) to normalize the signal intensity across different depths. Second, noise was attenuated with Gaussian blur 3D to facilitate the detection of cell-cell boundaries (X sigma = 2, Y sigma = 2, and Z sigma = 0.6). Third, images were processed using the morphological segmentation function of the MorphoLibJ plugin ([18]; image type is “border,” tolerance = 22). Obtained catchment basin images (right panels in Fig 3B and 3C) were then analyzed for cellular sphericity using the “analyze regions” function of the MorphoLibJ plugin. Finally, catchment basin images were used to generate 3D surface models of the embryos using an ImageJ 3D viewer macro [18]. To generate images shown in Figs 4A, 4C–4E and 3D surface models were exported as STL files and rendered using Adobe Photoshop (Adobe).

Fig 3. A refractive index matching medium improves 3D segmentation quality.

Fig 3

Open in a new tab

(A) Image processing pipeline used in this study. See text. (B, C) 3D watershed segmentation of C. elegans 24-cell stage embryos using 0% (B) and 30% (C) iodixanol. The left and right panels show the images after the 3D Gaussian blur and 3D watershed segmentation in the pipeline, respectively. Note that each segmented cell was labeled with a unique color. Arrowheads indicate cells with segmentation errors. (D) Quantified segmentation quality: by using the number of cells manually counted as a ground truth, segmentation quality α was determined. P-values were calculated using the Welch’s t-test. Scale bars, 10 μm.

Fig 4. Cellular morphometry analysis predicts exceptionally high cortical tension in the interphase P4 blastomere.

Fig 4

Open in a new tab

(A) Cell surface models reconstructed using 3D segmentation data. Note that in 0% iodixanol, segmentation failed, and cells elongated in the deeper cell layers. (B) Cellular sphericity of 413 blastomeres from 18 embryos is shown. Error bars indicate 95% confidence interval (CI). P-values were calculated by one-way ANOVA with the Holm-Sidak’s multiple comparison test. ** (P < 0.001) and *** (P < 0.0001). (C) Lateral and ventral views of surface model of 24-cell stage embryos obtained with 30% iodixanol. (D, E) Surface model of P4 and E.p cells from four different embryos. (F) 2D simulation of cellular shape changes. Each cell-cell boundary contains more than two contractile and viscoelastic elements to mimic active force generation and cell deformability of the cell cortex (see text). Cortical tension in P4 cells gradually increased relative to that in E.p and D cells.

2D cell shape simulation

To perform 2D cell shape simulation, we have used the previously reported model [19] and a simulation tool called Morphogenie (http://www.celldynamics.org/celldynamics/downloads/morphogenie/index.html). First, we defined the position and size of thirteen cells by specifying the vertices and junctions in a two-dimensional space. Next, posteriorly located three cells were specified as “E.p”, “P4”, and “D” in a pattern resembles that of wild-type embryos. Other cells were specified as the same cell type (“Others”). Cells’ cortical tension (“outer boundary tension”) was set to 0.1 for all the cell types except for the P4 cell. Tension at the cell-cell boundary (“inner boundary tension”) was set to 0.1 for all cell-boundaries. In order to test the effects of P4 cortical tension on the P4 internalization, we used the P4 cortical tension 0.1, 0.15, and 0.2.

Statistical analysis

For multiple comparisons, a one-way ANOVA with the Holm-Sidak’s method was used. Other analyses were performed using the Welch’s t-test. Statistical tests were performed using Prism 8 (Graph Pad). No statistical method was used to predetermine the sample size. The experiments were not randomized. The investigators were not blinded to the study.

Results

30% iodixanol is a RIMM for live C. elegans embryos

To determine the composition of a RIMM for C. elegans embryos, we imaged embryos bathed in egg salt buffer containing 0%, 10%, 20%, 25%, 30%, 40%, 50%, and 60% iodixanol (Fig 1 and S1 Fig). A previous study used phase-contrast microscopy to determine the optimal iodixanol concentration, because it converts the local differences in refractive indices of transparent specimens into the image contrast [8]. The loss of image contrast indicates the RI matching between samples and the surrounding medium. In this study, we used Nomarski DIC microscopy as it is widely used by C. elegans researchers. DIC microscopy converts the local differences in thickness and refractive indices of transparent specimens into the image contrast [20]. C. elegans embryos are similar in thickness (30 μm; [21]), such that the changes in DIC image contrast should reflect the degree of RI mismatch between embryos and the surrounding medium. Thus, we sought a condition where the image contrast, calculated by the ratio of maximum to minimum signal intensity, were at their lowest (Fig 1A and 1C). As a reference for the DIC shear direction, and as spacers against compression, we added 30 μm polystyrene beads to the medium (spheres in Fig 1A). We found that embryos imaged at 20%, 25%, and 30% iodixanol showed the lowest image contrast (Fig 1A and 1C). The 30% iodixanol solution did not affect embryo viability (0% iodixanol: 98% hatched, n = 154; 30% iodixanol: 97% hatched, n = 148).

Fig 1. Determination of a refractive index matching medium for live C. elegans.

Fig 1

Open in a new tab

(A, B) C. elegans embryos (A) and adult heads (B) imaged at different iodixanol concentrations. Images were obtained by Nomarski DIC microscopy. LUT images are intensity-based pseudocolored DIC images (pseudocolor was applied using the look-up table called “physics”). Arrows indicate embryos. Note that the spheres in some images are polystyrene beads used as a reference (see Materials and methods). (C, D) The extent of image contrast was calculated by dividing the maximum signal intensity in the sample area by that of the minimum. As both curves show that the average image contrast is the lowest at 30% iodixanol, statistical tests were performed by comparing samples treated with 30% iodixanol and others. P-values were calculated by one-way ANOVA with the Holm-Sidak’s multiple comparison test. n.s. (P > 0.05), ** (P < 0.001), and *** (P < 0.0001). Scale bars, 50 μm.

We assessed whether iodixanol could pass the C. elegans eggshell. C. elegans embryo is surrounded by the chitinous eggshell and the permeability barrier [22]. To reduce the RI mismatch between embryos and the surrounding media, 30% iodixanol may need to penetrate into the perivitelline space, a space between the eggshell and the permeability barrier (S2 Fig; a space between a white solid line and a yellow dotted line). A previous study reported that 389 Da and 1300 Da fluorescent dye passed the eggshell while that of 3000 Da did not [22]. The molecular weight of iodixanol is 1550 Da, which is only slightly higher than the reported limit of 1300 Da, such that iodixanol likely passes the eggshell. To directly test if molecules with the size of iodixanol can pass the eggshell, we bathed embryos into 30% iodixanol containing Alexa 568-phalloidin (1590 Da). We found that Alexa 568-phalloidin passed the eggshell (S2 Fig). Although the fluorescent molecules did not pass the permeability barrier, most areas of the embryo were in very close proximity to the surrounding perivitelline space (S2 Fig).

We also determined the composition of a refractive-index-matching medium for C. elegans adults. We used the head region of adult C. elegans to minimize variations in sample thickness. Similar to embryos, using 30% iodixanol resulted in the lowest image contrast (Fig 1B and 1D). L4-stage larva cultured in 30% iodixanol solution with OP50 E. coli produced fertile progenies, suggesting that the RIMM was not particularly detrimental to C. elegans postembryonic development. The RI of the 30% iodixanol solution was 1.379 ± 0.00004 at 22.5°C according to the measurement using a refractometer at 589 nm wavelength (S3 Fig). Selection of buffers should not have a large impact, since the RI of commonly used C. elegans buffers, such as M9 (RI: 1.335 ± 0.0001), egg salt (RI: 1.334 ± 0.00009), and Shelton’s growth medium (RI: 1.339 ± 0.00005) were similar at 22.5°C (S3 Fig). Based on these results, we concluded that 30% iodixanol, diluted in any common C. elegans buffers, is a RIMM for both C. elegans embryos and adult head regions.

The RIMM improves the signal-to-noise ratio in the deeper cell layers

We tested whether the obtained RIMM improved 3D live imaging of C. elegans embryos. One of the challenges in volumetric imaging is the reduced signal-to-noise ratio in deeper cell layers distal to the objective lens. This is mainly caused by the depth- and RI mismatch-induced spherical aberration that leads to the loss of contrast and signal intensity [1]. A previous study reported that an iodixanol-based RIMM improved the signal intensity in the deeper cell layers in other organisms by reducing the RI mismatch between samples and the surrounding media [8]. The same study used a silicone-immersion objective lens to minimize the RI mismatch between 1) the lens-immersion medium and samples and 2) the lens-immersion medium and the medium surrounding samples. However, a silicone-immersion objective lens is not a cost-effective measure for a wide-research community. We hypothesized that a RIMM alone could improve spherical aberrations regardless of the types of objective lenses by contributing to the reduction of the RI mismatches among the lens-immersion medium, samples, and the medium surrounding samples. Thus, we performed live imaging of histone H2B::GFP (chromosome) and TagRFP::PH (plasma membrane) expressed in C. elegans embryos with a spinning-disk confocal microscope equipped with an oil-immersion objective lens. Since embryos float in the 30% iodixanol solution, the distance between the objective lens and the embryos can differ when using 0% and 30% iodixanol. To minimize these differences in spherical aberration, we used polystyrene beads with embryo thickness as spacers (Fig 2A). In this imaging condition, the longitudinal surface of embryos physically contacts with coverslips, and all embryos are situated at a similar depth regardless of iodixanol concentration (Fig 2A and 2B).

We assessed the signal intensity of TagRFP-PH at different depths and different iodixanol concentrations. We collected multiple 4.96 μm x 4.96 μm images at 5-μm, 15-μm, and 25-μm depth, with each capturing a cell-cell boundary in the center (Fig 2C). First, we qualitatively assessed the outcomes of automatic image thresholding using Otsu’s method [16]. At 5-μm and 15-μm depth, cell-cell boundaries were clearly defined in both 0% and 30% iodixanol. On the other hand, cell-cell boundaries at 25-μm retained better defined linear patterns in 30% iodixanol (Fig 2C). Second, we quantitatively analyzed the TagRFP-PH signal intensity. We drew a line that perpendicularly passes the cell-cell boundary and generated intensity profiles (Fig 2D and S4A Fig). The signal intensities at 5-μm and 15-μm were not improved by 30% iodixanol (S4B Fig). At 25-μm depth, the cell-cell boundary showed a higher signal intensity with 30% iodixanol (Fig 2D; left graph). The average peak signal intensity with 30% iodixanol was 0.2566 ± 0.07, which was 1.38-fold higher than that with 0% iodixanol (0.1854 ± 0.070, p = 0.0388 by Welch’s t-test, n = 10 for each condition). These results suggest that the RIMM alone improves the signal intensity in the deeper cell layer in C. elegans embryos.

We also assessed whether the observed improvement could be achieved with a silicone-immersion objective lens. Consistent with the result using an oil-immersion objective lens and the previous study by Boothe et al. [8], we found that the RIMM improved the signal intensity at 25-μm depth (Fig 2D; right graph). In order to confirm the optimal iodixanol concentration for C. elegans fluorescence imaging, we have compared the peak TagRFP-PH signal intensities at 25-μm using different iodixanol concentrations (Fig 2E). The average peak signal intensity was the highest at 30% iodixanol concentration, confirming that the DIC-based RI matching method was successful in determining the RIMM (Fig 1C and Fig 2E). Taken together, these results suggest that a RIMM can improve signal intensity at the deep cell layer regardless of the types of objective lens immersion media.

The RIMM improves the 3D segmentation quality

Next, we tested whether the RIMM improved the 3D segmentation quality of individual cells in embryos. We imaged tagRFP-PH in 24- to 32-cell stage embryos for whole embryo volume (30-μm depth, Z step = 0.5 μm) and performed 3D segmentation of cells using the membrane signals as source images. We processed the source images using algorithms such as signal attenuation correction [17], Gaussian blur, and imposing regional minima [18] (Fig 3A; see Materials and methods). The images after Gaussian blur are shown in the left panels in Fig 3B and 3C for embryos bathed in 0% and 30% iodixanol media, respectively. Processed images were then segmented using a 3D watershed segmentation algorithm [18] (Fig 3A). When the cells were successfully segmented, individual blastomeres were labeled by unique colors, as shown in Fig 3B and 3C (right panels). From the surface plane to the mid-plane, segmentation quality appeared to be similar between 0% and 30% iodixanol media (Fig 3B and 3C; right panels). However, segmentation errors occurred more frequently in the deeper cell layers when using 0% iodixanol (Fig 3B; white arrowheads). To quantify the 3D segmentation quality, we manually counted the number of cells in embryos and used these counts as the ground truth. The manually counted cell number was then divided by the cell number detected by 3D segmentation to calculate the segmentation quality denoted as α (Fig 3D). We found that α obtained for 30% iodixanol was 91% (n = 19), while that for 0% iodixanol was 78% (n = 14; Fig 3D). These results suggest that the RIMM improves 3D segmentation quality.

Improved 3D segmentation allows high-throughput cellular morphometry analysis

We sought to determine the method by which improved 3D segmentation could facilitate the analysis of developmental processes. We generated 3D surface models of individual blastomeres in embryos using the results of 3D watershed segmentation (Fig 4A; see Materials and methods). In addition to the segmentation errors described above, we found that segmented cells in the deeper cell layers were more abnormally elongated in 0% iodixanol, likely due to the spherical aberration (Fig 4A). Thus, the RIMM allowed us to obtain more accurate and comprehensive cellular morphology data.

Next, we examined how meaningful information could be extracted by using cellular morphometry. As a proof of principle, we analyzed cell sphericity in 20- to 32-cell stage embryos. Owing to the improved 3D segmentation, we were able to analyze the sphericity of 413 individual blastomeres from 18 embryos by 3D ellipsoid fitting [18] (Fig 4B). As expected, on average, mitotic cells had higher sphericity than interphase cells due to mitotic rounding (Fig 4B; [23]). However, we found that interphase P4 blastomeres exhibited higher sphericity than mitotic cells (Fig 4B). We confirmed this data by manually inspecting the 3D surface models of blastomeres (Fig 4C–4E). The P4 blastomere is a germline cell surrounded by its sister cell, known as D, and an endoderm precursor cell E.p (Fig 4C and 4D). We confirmed that P4 cells were round and bulged into E.p and D cells, forming curved cell-cell boundaries (Fig 4E). As reported previously, curved cell-cell boundaries are induced by the asymmetry of cortical tension and internal cellular pressure [24]. In C. elegans 2-cell stage embryos, the anterior AB cell bulges into the posterior P1 cell due to the higher cortical tension in AB [24]. Thus, we speculated that interphase P4 cells had higher cortical tension than E.p and D cells. To test this possibility, we employed a 2D simulation of cell shape change using a previously described model [19]. In this simulation, the dynamics of cell shape change were predicted based on the tension values at different cell-cell boundaries. Each cell-cell boundary contains more than two smaller contractile and viscoelastic elements. These elements, with fixed tension and viscosity, mimic actomyosin contraction and resistance to deformation, respectively. Under this condition, we gradually increased the cortical tension in the P4 cell relative to those in E.p and D cells (Fig 4F). We found that P4-E.p and P4-D boundaries started to curve when the P4 cortical tension was 1.5 fold higher than in the other cells, and that P4 cells bulged into E.p and D cells (Fig 4F). These results, obtained by 3D cellular morphometry, suggest that the interphase P4 cell retains unexpectedly high cortical tension.

Discussion

In this study, we determined the composition of an RIMM for C. elegans embryos and adult heads. The obtained RIMM improved the 3D segmentation quality and enabled more accurate and high throughput 3D cellular morphometry analysis. Based on the cellular morphometry analysis performed, we identified that interphase P4 cells have exceptionally higher cortical tension than the surrounding cells.

From our experiments, the RI suitable for C. elegans live imaging medium is estimated to be 1.379 for embryos and adult heads. In the RI matching method using DIC, only variables were concentrations of iodixanol, such that the method is sensitive to the RI mismatch between samples and the surrounding media, but not to the internal heterogeneity of RI within specimens. The RI value for adult heads, which was determined by the previous tomographic phase microscopy analysis (RI: 1.36–1.38; [25]) and the 3D refractive index microscopy analysis (RI: ~1.35; [5]), were slightly different compared to the value obtained by us (RI: 1.379). Similarly, the RI value for embryos estimated by 3D refractive index microscopy was highly heterogeneous (RI: 1.33–1.37; [5]). However, the reported RI values at the outer edge of adults (all regions, including the head) and embryos were around 1.37, which was similar to the value obtained in our study. We found that iodixanol did not pass the permeability barrier (S2 Fig), suggesting that the physical interaction between embryo and the RIMM is unnecessary for reducing spherical aberrations. As one of the primary sources of depth-dependent spherical aberrations originates from the RI mismatch between the samples and the surrounding media, we conclude that the RI matching between the outer edge of embryos and the RIMM in the perivitelline space was sufficient to improve the spherical aberrations.

The use of a RIMM in conjunction with a water- or a silicone- immersion objective lens would generate the best 3D segmentation results for confocal microscopy. This is because these objective lenses minimize the refractive index mismatch between 1) the objective lens immersion medium and samples and 2) the objective lens immersion medium and the medium surrounding samples. We have not compared the performance between water-immersion and silicone-immersion objectives, but the silicone-immersion objective would be better because of its higher numerical aperture and the very low evaporation rate of silicone. Alternatively, the use of dual-view plane illumination microscopy or lattice light-sheet microscopy would allow generation of isotropic images. However, all measures discussed are costly and is not always accessible to many researchers and institutes. Our study, using conventional spinning-disk confocal microscopy and an oil-immersion lens, demonstrated that a RIMM alone was highly effective at improving 3D segmentation quality. With this low-cost solution, university instructors can also print 3D models and use them for biology curricula. It should be noted that a RIMM improves signal intensity at the image acquisition step, such that it can be coupled with other state-of-the-art post-acquisition 3D cell segmentation algorithms, such as RACE [26], 3DMMS [27], and BCOMS [28].

Cellular morphometry analysis revealed that the interphase P4 cell had an exceptionally high sphericity and bulges into the E.p cell. As confirmed in the 2D cell shape simulation, the high sphericity of P4 is probably due to the high cortical tension reminiscent of mitotic rounding. As mitotic rounding accelerates cell invagination [29], P4 rounding may accelerate its internalization. Indeed, the daughters of P4, primordium germ cells (PGCs), migrate into the interior of embryos by hitchhiking gastrulation [30]. Thus, the cortical tension of P4 and its descendants may be precisely controlled to promote the PGC internalization process. Further analysis is required to test these possibilities, but our study demonstrates that a RIMM facilitates 3D cellular morphometry analysis in C. elegans.

Supporting information

S1 Fig

(TIF)

Click here for additional data file. (6.1MB, tif)
S2 Fig

(TIF)

Click here for additional data file. (9.2MB, tif)
S3 Fig

(EPS)

Click here for additional data file. (1.2MB, eps)
S4 Fig

(EPS)

Click here for additional data file. (1.6MB, eps)

Acknowledgments

We thank Bruce Bowerman for providing valuable advice. We thank Chris Doe for sharing lab equipment. We thank the Caenorhabditis Genetics Center (funded by the NIH Office of Research Infrastructure Programs; P40 OD010440) for C. elegans strains. We thank Farid Jalali for critical reading of the manuscript. This work was supported by the Canadian Institutes of Health Research (Project Grant; PJT-169145) to K.S.

Data Availability

All relevant data are within the manuscript, Supporting Information files, and via Figshare: https://doi.org/10.6084/m9.figshare.c.5108522.v1.

Funding Statement

This work was supported by the Canadian Institutes of Health Research (Project Grant; PJT-169145) to K.S. C. elegans strains were provided by the Caenorhabditis Genetics Center (funded by the NIH Office of Research Infrastructure Programs; P40 OD010440).

References

  • 1.Booth MJ, Neil MA, Wilson T. Aberration correction for confocal imaging in refractive-index-mismatched media. Journal of Microscopy. 1998;192: 90–98. 10.1111/j.1365-2818.1998.99999.x [DOI] [Google Scholar]
  • 2.Green RA, Audhya A, Pozniakovsky A, Dammermann A, Pemble H, Monen J, et al. Expression and Imaging of Fluorescent Proteins in the C. elegans Gonad and Early Embryo. Methods in Cell Biology. Academic Press; 2008. pp. 179–218. 10.1016/S0091-679X(08)85009-1 [DOI] [PubMed] [Google Scholar]
  • 3.Sugioka K, Fielmich LE, Mizumoto K, Bowerman B, van den Heuvel S, Kimura A, et al. Tumor suppressor APC is an attenuator of spindle-pulling forces during C. elegans asymmetric cell division. Proc Natl Acad Sci USA. 2018;115: E954–E963. 10.1073/pnas.1712052115 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Sugioka K, Bowerman B. Combinatorial Contact Cues Specify Cell Division Orientation by Directing Cortical Myosin Flows. Developmental Cell. 2018;46: 257–270.e5. 10.1016/j.devcel.2018.06.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Chowdhury S, Chen M, Eckert R, Ren D, Wu F, Repina N, et al. High-resolution 3D refractive index microscopy of multiple-scattering samples from intensity images. Optica, OPTICA. 2019;6: 1211–1219. 10.1364/OPTICA.6.001211 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Kumar A, Wu Y, Christensen R, Chandris P, Gandler W, McCreedy E, et al. Dual-view plane illumination microscopy for rapid and spatially isotropic imaging. Nat Protoc. 2014;9: 2555–2573. 10.1038/nprot.2014.172 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Chen BC, Legant WR, Wang K, Shao L, Milkie DE, Davidson MW, et al. Lattice Light Sheet Microscopy: Imaging Molecules to Embryos at High Spatiotemporal Resolution. Science. 2014;346: 1257998 10.1126/science.1257998 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Boothe T, Hilbert L, Heide M, Berninger L, Huttner WB, Zaburdaev V, et al. A tunable refractive index matching medium for live imaging cells, tissues and model organisms. Waterman CM, editor. eLife. 2017;6: e27240 10.7554/eLife.27240 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Brenner S. The genetics of Caenorhabditis elegans. Genetics. 1974;77: 71–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Heppert JK, Dickinson DJ, Pani AM, Higgins CD, Steward A, Ahringer J, et al. Comparative assessment of fluorescent proteins for in vivo imaging in an animal model system. Mol Biol Cell. 2016;27: 3385–3394. 10.1091/mbc.E16-01-0063 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Praitis V, Casey E, Collar D, Austin J. Creation of low-copy integrated transgenic lines in Caenorhabditis elegans. Genetics. 2001;157: 1217–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Edgar LG. Blastomere culture and analysis. Methods Cell Biol. 1995;48: 303–321. 10.1016/s0091-679x(08)61393-x [DOI] [PubMed] [Google Scholar]
  • 13.Hsu CR, Xiong R, Sugioka K. In Vitro Reconstitution of Spatial Cell Contact Patterns with Isolated Caenorhabditis elegans Embryo Blastomeres and Adhesive Polystyrene Beads. J Vis Exp. 2019. 10.3791/60422 [DOI] [PubMed] [Google Scholar]
  • 14.Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, et al. Fiji: an open-source platform for biological-image analysis. Nat Methods. 2012;9: 676–682. 10.1038/nmeth.2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Shelton CA, Bowerman B. Time-dependent responses to glp-1-mediated inductions in early C. elegans embryos. Development. 1996;122: 2043–2050. [DOI] [PubMed] [Google Scholar]
  • 16.Otsu N. A Threshold Selection Method from Gray-Level Histograms. IEEE Transactions on Systems, Man, and Cybernetics. 1979;9: 62–66. 10.1109/TSMC.1979.4310076 [DOI] [Google Scholar]
  • 17.Biot E, Crowell E, Hofte H, Maurin Y, Vernhettes S, Andrey P. A new filter for spot extraction in n-dimensional biological imaging. 2008 5th IEEE International Symposium on Biomedical Imaging: From Nano to Macro. 2008. pp. 975–978. 10.1109/ISBI.2008.4541161 [DOI] [Google Scholar]
  • 18.Legland D, Arganda-Carreras I, Andrey P. MorphoLibJ: integrated library and plugins for mathematical morphology with ImageJ. Bioinformatics. 2016;32: 3532–3534. 10.1093/bioinformatics/btw413 [DOI] [PubMed] [Google Scholar]
  • 19.Sherrard K, Robin F, Lemaire P, Munro E. Sequential activation of apical and basolateral contractility drives ascidian endoderm invagination. Curr Biol. 2010;20: 1499–1510. 10.1016/j.cub.2010.06.075 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Allen RD, David GB, Nomarski G. The Zeiss-Nomarski differntial interference equipment for transmitted-light microscopy. Z Wiss Mikrosk. 1969;69: 193–221. [PubMed] [Google Scholar]
  • 21.Riddle DL, Blumenthal T, Meyer BJ, Priess JR. C. elegans II. 2nd ed Cold Spring Harbor Laboratory Press; 1997. [PubMed] [Google Scholar]
  • 22.Olson SK, Greenan G, Desai A, Muller-Reichert T, Oegema K. Hierarchical assembly of the eggshell and permeability barrier in C. elegans. J Cell Biol. 2012;198: 731–748. 10.1083/jcb.201206008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Théry M, Bornens M. Get round and stiff for mitosis. HFSP J. 2008;2: 65–71. 10.2976/1.2895661 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Fujita M, Onami S. Cell-to-Cell Heterogeneity in Cortical Tension Specifies Curvature of Contact Surfaces in Caenorhabditis elegans Embryos. PLOS ONE. 2012;7: e30224 10.1371/journal.pone.0030224 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Choi W, Fang-Yen C, Badizadegan K, Oh S, Lue N, Dasari RR, et al. Tomographic phase microscopy. Nature Methods. 2007;4: 717–719. 10.1038/nmeth1078 [DOI] [PubMed] [Google Scholar]
  • 26.Stegmaier J, Amat F, Lemon WC, McDole K, Wan Y, Teodoro G, et al. Real-Time Three-Dimensional Cell Segmentation in Large-Scale Microscopy Data of Developing Embryos. Developmental Cell. 2016;36: 225–240. 10.1016/j.devcel.2015.12.028 [DOI] [PubMed] [Google Scholar]
  • 27.Cao J, Wong MK, Zhao Z, Yan H. 3DMMS: robust 3D Membrane Morphological Segmentation of C. elegans embryo. BMC Bioinformatics. 2019;20: 176 10.1186/s12859-019-2720-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Azuma Y, Onami S. Biologically constrained optimization based cell membrane segmentation in C. elegans embryos. BMC Bioinformatics. 2017;18: 307 10.1186/s12859-017-1717-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Kondo T, Hayashi S. Mitotic cell rounding accelerates epithelial invagination. Nature. 2013;494: 125–129. 10.1038/nature11792 [DOI] [PubMed] [Google Scholar]
  • 30.Chihara D, Nance J. An E-cadherin-mediated hitchhiking mechanism for C. elegans germ cell internalization during gastrulation. Development (Cambridge, England). 2012;139: 2547–56. 10.1242/dev.079863 [DOI] [PMC free article] [PubMed] [Google Scholar]

Decision Letter 0

Bob Goldstein

9 Jul 2020

PONE-D-20-18786

Improved 3D cellular morphometry of Caenorhabditis elegans embryos using a refractive index matching medium

PLOS ONE

Dear Dr. Sugioka,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

You will see that Reviewer 1 raised a substantive concern about whether the correction of spherical aberration in deep sections can occur if the iodixanol is not penetrating the eggshell, and noted that the degree of improvement appears small. We would like to give you a chance to address this concern and other issues raised by the reviewers in your manuscript. Please also note that the statistical test used in Figure 2D, the first figure showing the improvement quantitatively, is a test that assumes a normal distribution of the data.

Please submit your revised manuscript by Aug 23 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

  • A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Bob Goldstein

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: No

Reviewer #2: Yes

Reviewer #3: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: N/A

Reviewer #2: Yes

Reviewer #3: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

Reviewer #3: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The manuscript by Xiong & Sugioka describes the use of a refractive index matching medium to increase the visibility of in vivo fluorescence images of C. elegans embryos in order to facilitate the analysis of cellular morphometry. Although there are some potentially interesting aspects to this study, I consider it fundamentally flawed and not worthy of publication in its current form.

The eggshell of a C. elegans embryo is very impermeable and, as such, it is very unlikely that any of the iodixanol refractive index matching agent that was used will penetrate the embryo. This means that the only effect that refractive index matching the external medium will have is to reduce scattered light from the bathing medium/eggshell boundary. The claimed correction of spherical aberration in deep sections cannot occur. In Fig. 2 the authors compare two embryos with different refractive index bathing media. There is very little difference to be seen, and both images exhibit the tell-tale deep section axial expansion typical of spherical aberration. There is also little visible improvement of deep section images to be seen in Fig. 3. Given the intrinsic variability of in vivo fluorescence between embryos and the fact that different embryos were used in the comparisons, I am unconvinced by these data. There is a well-established way of reducing the spherical aberration from in vivo specimens that the authors do not mention, and that is to use water immersion lenses, where there is a reduced refractive index mismatch between cytosol and the lens immersion medium.

I did find the author’s use of DIC microscopy to tune the refractive index of the medium interesting. As they demonstrate, this reduces the contrast extremes at the boundary of the embryo while maintaining the DIC contrast of structures within the embryo. This strategy might improve visibility of cells adjacent to the lateral sides of the embryo and thereby facilitate cell lineage studies. This simple technique might be worth developing into a technical note.

Reviewer #2: Summary

The authors of this study apply refractive index media matching (RIMM), previously reported in Boothe, et al. 2017 for other organisms, to imaging of C. elegans embryos. They empirically and quantitatively determined the RIMM by titrating different percentages of iodixanol into egg salts, a commonly used mounting medium of C. elegans embryos. Importantly, they demonstrate that the addition of iodixanol at the RIMM concentration is non-toxic to embryos. Using fluorescent C. elegans lines and confocal microscopy, the authors showed that imaging C. elegans embryos in RIMM can result in improved signal at deeper z-planes (further from the coverslip). This is presumably because the RIMM decreases spherical aberration or light scattering that occurs when sample and media have differing refractive indices. The authors subsequently demonstrated that images obtained with RIMM were more successfully segmented using 3D segmentation software than those acquired without RIMM. Finally, the authors used cellular morphometry from their images to show that the germ cell precursor P4 has increased sphericity, and therefore, likely increased cortical tension than its surrounding cell neighbors.

Major comments

How does tuning the RIMM under DIC conditions compare to other studies (Boothe et al.,) where phase contrast was used? I like that the authors used DIC imaging, because most C. elegans researchers image using DIC optics. However, the basis of matching the RIMM (according to Boothe et al., 2017) is minimizing the contrast in the phase images, and in DIC microscopy “Contrast can be varied instrumentally to suit the object” (Allen, David, and Nomarski, 1969). Does this present challenge in applying this RIMM across different microscopes with DIC or even on the same scope with the DIC adjusted differently? Due to results in other figures, it appears that it worked to tune the RIMM this way. Can the authors comment on the use of DIC optics vs. phase contrast for tuning RIMM?

I do like the quantification of loss of contrast in the iodixanol series, as opposed to the highly estimated method used in the other papers.

Minor comments

Methods:

In the Methods section, the authors mention that the embryos float in the RIMM (iodixanol) solution. Anyone who has previously worked with C. elegans embryos (which typically sink when mounting) would really want to know this. Thus, maybe mention of this could be made in the main text.

Explain in Methods that Ostu’s method is thresholding (this is explained in the figure, but not in the Methods for some reason).

The description of 3D segmentation using FIJI in the methods is very thorough. However, the 2D cell shape simulation description is incredibly minimal. Were there specific parameters or settings the authors used?

For the statistical analyses, was any software used?

Results:

When reporting “(0% iodixanol: 98%, n = 154; 30% iodixanol: 97%, n = 148)” specify that percentages refer to survival (I assume).

Figure 1:

I understand the desire to have zoomed in views of the embryos, but ideally, there would also be images with many embryos visible in the different iodixanol percentages, not just a single embryo (or two) shown per condition.

The viability of adult worms was not established. Can they tolerate the RIMM? This is relevant for longer-term imaging of adult C. elegans.

Figure 2:

I really liked that the authors quantified how well automated segmentation worked in the different conditions using ground truth cell counts.

Discussion:

“It should be noted that by coupling a RIMM and state-of-the-art 3D cell segmentation algorithms, such as RACE [24], 3DMMS [25], and BCOMS [26], researchers may obtain even better segmentation results than those shown in this study.” This is confusing – why didn’t the authors use these segmentation programs? What would “better” segmentation results look like?

The last sentence really fizzles out - maybe the authors could strongly restate the important findings of the paper.

Reviewer #3: In the manuscript by Xiong and sugioka, the authors systematically determine a refractive index (RI) matching solution for the imaging of C. elegans. This work builds from a previous study that introduced the used of iodixanol (OptiPrep) as a non-toxic method to increase the RI of various medias (Booth et al, eLife) and thus reduce spherical aberrations caused by RI mismatch that impedes fluorescent based imaging methods. Here the authors find that a 30% final solution of iodixanol enhances fluorescence imaging in C. elegans.

Overall this is a well presented work that is useful for many using C. elegans as a model system. I have only a few issues that should be addressed prior to publication.

Major suggestions:

1) The calculated Ri for the final solution is affected by the RI of the egg buffer (fixed at 1.33). Is this the actual RI of egg buffer? How critical is this?

2) On page 9, it is stated that “image quality is accessed. How is this done precisely? It seems to me that the absolute Histone signal (not only signal to noise) should be the same form each nucleus. How much loss is there in control vs. optimized?

3) The work here was completed using oil immersion. Do the authors expect a difference when using non oil (air, water, silicon, etc) optics? It would be helpful to make a comment on this.

Minor suggestions:

1) There are figure legends embedded in the main text, please remove.

2) In figure 2C is would be helpful (more clear) if the images for converted to gra

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: Yes: Paul Maddox

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2020 Sep 30;15(9):e0238955. doi: 10.1371/journal.pone.0238955.r002

Author response to Decision Letter 0

23 Aug 2020

Dear editor and reviewers,

We appreciate your careful reading of our manuscript and insightful comments that greatly improved the overall quality of the revised manuscript. We also would like to thank the editor for giving us enough time to address these issues. We are delighted to inform you that we were able to obtain more data since the research curtailment of our institute has been partially lifted, and our new spinning disk confocal microscope with a silicone-immersion objective lens has been installed. We performed all key experiments that can answer the points raised by the reviewers, as summarized below.

1. We have performed more rigorous and precise analyses of RIMM-dependent improvement of signal intensity by imaging cell-cell boundary at different depths, at different iodixanol concentrations, with different types of objectives lenses.

According to the reviewers’ comments, we realized that the original Figure 2B-D did not successfully present the effect of RIMM. Thus, we have performed the precise depth-dependent analysis of signal intensity, as shown in the revised Figure 2C, 2D, 2E, and Figure S4. Furthermore, we confirmed that the RIMM-dependent improvement of signal intensity can be achieved using a silicone-immersion objective lens (new Figure 2D). Thus, RIMM can improve signal intensity at the deep cell layers regardless of the types of objective lenses.

2. We have analyzed whether molecules with the size of iodixanol can pass the eggshell.

While it has been known that C. elegans eggshell can pass small molecules with less than 1300 Da molecular weight, whether iodixanol (1550 Da) can pass the eggshell was not shown before. Thus, we have assessed and found that the fluorescent molecules with 1590 Da molecular weight penetrated into the eggshell as shown in Figure S2.

3. We have added more description of 2D cell shape simulation in Materials and Methods.

From here, we will answer each point raised by the reviewers (our answers are in blue).

Reviewer #1: The manuscript by Xiong & Sugioka describes the use of a refractive index matching medium to increase the visibility of in vivo fluorescence images of C. elegans embryos in order to facilitate the analysis of cellular morphometry. Although there are some potentially interesting aspects to this study, I consider it fundamentally flawed and not worthy of publication in its current form.

The eggshell of a C. elegans embryo is very impermeable and, as such, it is very unlikely that any of the iodixanol refractive index matching agent that was used will penetrate the embryo. This means that the only effect that refractive index matching the external medium will have is to reduce scattered light from the bathing medium/eggshell boundary. The claimed correction of spherical aberration in deep sections cannot occur.

The eggshell of C. elegans is permeable to small molecules while the thin inner membrane attached to embryos, called permeability barrier, is not. A previous study by Olson et al., used three different sizes of fluorescent dyes to test their eggshell permeability. Olson et al. found that 389 Da and 1300 Da molecules passed the eggshell while that of 3000 Da did not (Olson et al., JCB 2012, PMID: 22908315). These small molecules penetrate into the perivitelline space— a space between eggshell and permeability barrier. The molecular weight of iodixanol is 1550 Da, only slightly higher than the reported limit of 1300 Da, such that it is likely that iodixanol can pass the eggshell.

To test if molecules with the size of iodixanol can indeed pass the eggshell, we bathed embryos into 30% iodixanol containing Alexa 568-phalloidin (1590 Da). As shown in the new Figure S3, Alexa 568-phalloidin passed the eggshell. While it did not pass the permeability barrier, most areas of the embryo were in very close proximity to the surrounding perivitelline space. Iodixanol works without altering the intracellular refractive index (Boothe et al., eLife 2017). Rather, reducing the refractive index mismatch between embryos and the surrounding medium is sufficient to limit the depth-induced spherical aberration (Boothe et al., eLife 2017).

Based on the new data, we have added the text as follows:

Page 9, line 286:

“We assessed whether iodixanol could pass the C. elegans eggshell. C. elegans embryo is surrounded by the chitinous eggshell and the permeability barrier [22]. To reduce the RI mismatch between embryos and the surrounding media, 30% iodixanol may need to penetrate into the perivitelline space, a space between the eggshell and the permeability barrier (Figure S2; a space between a white solid line and a yellow dotted line). A previous study reported that 389 Da and 1300 Da fluorescent dye passed the eggshell while that of 3000 Da did not [22]. The molecular weight of iodixanol is 1550 Da, which is only slightly higher than the reported limit of 1300 Da, such that iodixanol likely passes the eggshell. To directly test if molecules with the size of iodixanol can pass the eggshell, we bathed embryos into 30% iodixanol containing Alexa 568-phalloidin (1590 Da). We found that Alexa 568-phalloidin passed the eggshell (Figure S2). Although the fluorescent molecules did not pass the permeability barrier, most areas of the embryo were in very close proximity to the surrounding perivitelline space (Figure S2). ”

In Fig. 2 the authors compare two embryos with different refractive index bathing media. There is very little difference to be seen, and both images exhibit the tell-tale deep section axial expansion typical of spherical aberration. There is also little visible improvement of deep section images to be seen in Fig. 3. Given the intrinsic variability of in vivo fluorescence between embryos and the fact that different embryos were used in the comparisons, I am unconvinced by these data.

In the revised Figure 2, we performed more rigorous and quantitative analyses to prove that the iodixanol solution improved the signal intensity in the deep cell layers. According to our analysis, the membrane signal intensity was improved about 1.3-fold at 25 µm depth. Human eyes are known to outperform computer vision at recognizing patterns as discussed in the “Computer Vision in Cell Biology” authored by Gaudenz Danuser (Danuser, Cell 2011 PMID ). The following is a quote from this paper.

“Everyone who has programmed computer vision algorithms knows how hard it is to make a computer “see” the things our eye appears to recognize effortlessly. The reason for the ease with which the human vision system identifies relevant events in complex scenes is that our brain continually associates the observable image signals with our memory of previous visual experiences and our best interpretation of the scene.”

Computer vision is essential for stream-lined 3D cellular morphometry. The main aim of this manuscript is to show that iodixanol can let the computer vision, not the human vision, to better detect membrane signals.

There is a well-established way of reducing the spherical aberration from in vivo specimens that the authors do not mention, and that is to use water immersion lenses, where there is a reduced refractive index mismatch between cytosol and the lens immersion medium.

In our original manuscript, we have mentioned the silicone-immersion objective lens as a relatively costly measure to reduce the RI mismatch between samples and the lens immersion medium in the Discussion. Our view is similar for the water-immersion objective lens in terms of cost. To our knowledge, costs of a typical water-immersion objective (60x magnification, 1.2NA) and a 60x 1.3NA silicone immersion objective are around 10,000 USD and 12,000 USD, respectively. Theoretically, 1.2NA lens is 1.3-fold dimmer than the 1.3NA objective lens. Thus, we think that the silicone immersion objective lens is the better option for C. elegans research. We agree that we did not elaborate enough about this established method, and we modified the texts in the Introduction and the Discussion as shown below.

Page 3, line 46:

“Spherical aberrations misalign the optical paths, resulting in the loss of signal intensity and resolution. This effect increases with sample thickness, limiting the capacity of 3D cellular morphometry [1]. A major source of spherical aberration is the refractive index (RI) mismatch among three different materials: the lens immersion medium, samples, and the medium surrounding the samples. In typical C. elegans live imaging, samples are mounted onto 2%-4% agarose and are bathed in M9 or egg salt buffer [2][3], or bathed in these buffers without agarose [4]. These commonly used aqueous media (RI: ~1.33) should create the RI mismatch with both lens immersion media (RI: 1.52 for oil immersion) and C. elegans embryos (RI: 1.33-1.37; [5]).

In this study, we focused on the cost-effective method to reduce the RI mismatch that can be immediately adopted by the broad C. elegans research community. It is known that the use of water (RI: 1.33) and silicone (RI: 1.4) as lens-immersion media reduces the RI mismatch, but high-numerical aperture (NA) objective lenses designed for water- and silicone-immersion typically cost more than 10,000 USD.”

Page 18, line 554:

“The use of a RIMM in conjunction with a water- or a silicone- immersion objective lens would generate the best 3D segmentation results for confocal microscopy. This is because these objective lenses minimize the refractive index mismatch between 1) the objective lens immersion medium and samples and 2) the objective lens immersion medium and the medium surrounding samples. We have not compared the performance between water-immersion and silicone-immersion objectives, but the silicone-immersion objective would be better because of its higher numerical aperture and the very low evaporation rate of silicone.”

In addition, by using a silicone-immersion objective lens, we showed that a RIMM improved the depth-induced spherical aberration even when the RI mismatch between sample and the lens immersion medium was effectively minimal (revised Figures 2D and 2E). The result is consistent with the study by Boothe et al., which used both silicone-immersion objective lens and RIMM (Boothe et al., eLife 2017). These results together suggest that the benefit of RIMM is independent of the types of objective immersion media, and RIMM is a cost-effective solution to improve the spherical aberration.

I did find the author’s use of DIC microscopy to tune the refractive index of the medium interesting. As they demonstrate, this reduces the contrast extremes at the boundary of the embryo while maintaining the DIC contrast of structures within the embryo. This strategy might improve visibility of cells adjacent to the lateral sides of the embryo and thereby facilitate cell lineage studies. This simple technique might be worth developing into a technical note.

Thank you for your comment.

Reviewer #2: Summary

The authors of this study apply refractive index media matching (RIMM), previously reported in Boothe, et al. 2017 for other organisms, to imaging of C. elegans embryos. They empirically and quantitatively determined the RIMM by titrating different percentages of iodixanol into egg salts, a commonly used mounting medium of C. elegans embryos. Importantly, they demonstrate that the addition of iodixanol at the RIMM concentration is non-toxic to embryos. Using fluorescent C. elegans lines and confocal microscopy, the authors showed that imaging C. elegans embryos in RIMM can result in improved signal at deeper z-planes (further from the coverslip). This is presumably because the RIMM decreases spherical aberration or light scattering that occurs when sample and media have differing refractive indices. The authors subsequently demonstrated that images obtained with RIMM were more successfully segmented using 3D segmentation software than those acquired without RIMM. Finally, the authors used cellular morphometry from their images to show that the germ cell precursor P4 has increased sphericity, and therefore, likely increased cortical tension than its surrounding cell neighbors.

Major comments

How does tuning the RIMM under DIC conditions compare to other studies (Boothe et al.,) where phase contrast was used? I like that the authors used DIC imaging, because most C. elegans researchers image using DIC optics. However, the basis of matching the RIMM (according to Boothe et al., 2017) is minimizing the contrast in the phase images, and in DIC microscopy “Contrast can be varied instrumentally to suit the object” (Allen, David, and Nomarski, 1969). Does this present challenge in applying this RIMM across different microscopes with DIC or even on the same scope with the DIC adjusted differently? Due to results in other figures, it appears that it worked to tune the RIMM this way. Can the authors comment on the use of DIC optics vs. phase contrast for tuning RIMM?

Phase-contrast microscopy converts differences in refractive index within the local area into image contrast. On the other hand, DIC converts differences in refractive index and sample thickness into image contrast. Boothe et al., made the right choice by selecting phase contrast for simplicity of analysis. We agree with the reviewer 2 that it is important to use DIC for C. elegans research community. As far as using samples with a similar thickness (e.g., C. elegans embryos), DIC image contrast should reflect the differences of refractive index. As the reviewer 2 pointed out, actual image contrast can be varied in different DIC settings, optical systems, and light intensities. Nevertheless, one should be able to reproduce the trend of our graph shown in Figure 1C to identify the iodixanol concentration which gives the lowest image contrast based on the principles described above. We elaborated this point in the text as follows:

Page 8, line 250:

“To determine the composition of a RIMM for C. elegans embryos, we imaged embryos bathed in egg salt buffer containing 0%, 10%, 20%, 25%, 30%, 40%, 50%, and 60% iodixanol (Figure 1 and Figure S1). A previous study used phase-contrast microscopy to determine the optimal iodixanol concentration, because it converts the local differences in refractive indices of transparent specimens into the image contrast [8]. The loss of image contrast indicates the RI matching between samples and the surrounding medium. In this study, we used Nomarski DIC microscopy as it is widely used by C. elegans researchers. DIC microscopy converts the local differences in thickness and refractive indices of transparent specimens into the image contrast [20]. C. elegans embryos are similar in thickness (30 µm; [21]), such that the changes in DIC image contrast should reflect the degree of RI mismatch between embryos and the surrounding medium. Thus, we sought a condition where the image contrast, calculated by the ratio of maximum to minimum signal intensity, were at their lowest (Figure 1A and 1C).”

We could not perform an experiment using a different DIC microscope because it is difficult to quantitatively say how two DIC settings are different. Instead, we have measured the signal intensity of TagRFP-PH at different iodixanol concentrations, as shown in Figure 2E. The graphs in Figure 1C and 2E show clear negative correlation. As fluorescent images were obtained without DIC optics, retrospectively one can assume that the graph in Figure 1C was effective in reporting the refractive index mismatch between samples and the surrounding media.

I do like the quantification of loss of contrast in the iodixanol series, as opposed to the highly estimated method used in the other papers.

Thank you for your comment. In the revised manuscript, we have directly measured the RI of these solutions and commonly used aqueous buffers using a refractometer to further make this study as a useful resource for the C. elegans research community (Figure S3). We have modified the text as follows:

Page 10, line 305:

“The RI of the 30% iodixanol solution was 1.379 ± 0.00004 at 22.5˚C according to the measurement using a refractometer at 589 nm wavelength (Figure S3). Selection of buffers should not have a large impact, since the RI of commonly used C. elegans buffers, such as M9 (RI: 1.335 ± 0.0001), egg salt (RI: 1.334 ± 0.00009), and Shelton’s growth medium (RI: 1.339 ± 0.00005) were similar at 22.5˚C (Figure S3). Based on these results, we concluded that 30% iodixanol, diluted in any common C. elegans buffers, is a RIMM for both C. elegans embryos and adult head regions.”

Minor comments

Methods:

In the Methods section, the authors mention that the embryos float in the RIMM (iodixanol) solution. Anyone who has previously worked with C. elegans embryos (which typically sink when mounting) would really want to know this. Thus, maybe mention of this could be made in the main text.

In the revised main text, the description can be found on page 11, line 338.

“Since embryos float in the 30% iodixanol solution, the distance between the objective lens and the embryos can differ when using 0% and 30% iodixanol. To minimize these differences in spherical aberration, we used polystyrene beads with embryo thickness as spacers (Figure 2A).”

Explain in Methods that Ostu’s method is thresholding (this is explained in the figure, but not in the Methods for some reason).

We have modified the description in the Materials and Methods as follows:

Page 6, line 191:

“For the generation of the binary images shown in Figure 2C, we used Otsu’s thresholding method with Fiji.”

The description of 3D segmentation using FIJI in the methods is very thorough. However, the 2D cell shape simulation description is incredibly minimal. Were there specific parameters or settings the authors used?

We have added the parameters used in the 2D simulation in the Materials and Methods as follows:

Page 7, line 227:

“First, we defined the position and size of thirteen cells by specifying the vertices and junctions in a two-dimensional space. Next, posteriorly located three cells were specified as “E.p”, “P4”, and “D” in a pattern resembles that of wild-type embryos. Other cells were specified as the same cell type (“Others”). Cells’ cortical tension (“outer boundary tension”) was set to 0.1 for all the cell types except for the P4 cell. Tension at the cell-cell boundary (“inner boundary tension”) was set to 0.1 for all cell-boundaries. In order to test the effects of P4 cortical tension on the P4 internalization, we used the P4 cortical tension 0.1, 0.15, and 0.2.”

For the statistical analyses, was any software used?

Thank you for pointing this out. We have added information about the software as follows:

Page 8, line 243:

“Statistical tests were performed using Prism 8 (Graph Pad).”

Results:

When reporting “(0% iodixanol: 98%, n = 154; 30% iodixanol: 97%, n = 148)” specify that percentages refer to survival (I assume).

We have modified the description as follows (changes underlined):

“(0% iodixanol: 98% hatched, n = 154; 30% iodixanol: 97% hatched, n = 148)”

Figure 1:

I understand the desire to have zoomed in views of the embryos, but ideally, there would also be images with many embryos visible in the different iodixanol percentages, not just a single embryo (or two) shown per condition.

We have added more embryo images in the revised Figure S1.

The viability of adult worms was not established. Can they tolerate the RIMM? This is relevant for longer-term imaging of adult C. elegans.

We have cultured L4 larva in 30% iodixanol solution for more than 24 hours. They normally produced hatching embryos. We have added the following description.

Page 10, line 303:

“L4-stage larva cultured in 30% iodixanol solution with OP50 E. coli produced fertile progenies, suggesting that the RIMM was not particularly detrimental to C. elegans postembryonic development.”

Figure 2:

I really liked that the authors quantified how well automated segmentation worked in the different conditions using ground truth cell counts.

Thank you for your comment.

Discussion:

“It should be noted that by coupling a RIMM and state-of-the-art 3D cell segmentation algorithms, such as RACE [24], 3DMMS [25], and BCOMS [26], researchers may obtain even better segmentation results than those shown in this study.” This is confusing – why didn’t the authors use these segmentation programs? What would “better” segmentation results look like?

The last sentence really fizzles out - maybe the authors could strongly restate the important findings of the paper.

These algorithms perform post-acquisition image processing such that they do not improve image quality at image-acquisition step. Thus, we consider that impact of our manuscript is independent of these methods. We have clarified this point as follows:

Page 18, line 565:

“With this low-cost solution, university instructors can also print 3D models and use them for biology curricula. It should be noted that a RIMM improves signal intensity at the image acquisition step, such that it can be coupled with other state-of-the-art post-acquisition 3D cell segmentation algorithms, such as RACE [26], 3DMMS [27], and BCOMS [28].”

Reviewer #3: In the manuscript by Xiong and sugioka, the authors systematically determine a refractive index (RI) matching solution for the imaging of C. elegans. This work builds from a previous study that introduced the used of iodixanol (OptiPrep) as a non-toxic method to increase the RI of various medias (Booth et al, eLife) and thus reduce spherical aberrations caused by RI mismatch that impedes fluorescent based imaging methods. Here the authors find that a 30% final solution of iodixanol enhances fluorescence imaging in C. elegans.

Overall this is a well presented work that is useful for many using C. elegans as a model system. I have only a few issues that should be addressed prior to publication.

Major suggestions:

1) The calculated Ri for the final solution is affected by the RI of the egg buffer (fixed at 1.33). Is this the actual RI of egg buffer? How critical is this?

The buffer we have used is the “egg salt buffer” which contains 118 mM NaCl and 48 mM KCl. We have used this buffer for imaging since it is more physiologically isotonic to embryos than M9 buffer (Edgar LG, Methods Cell Biol. 1995; PMID: 22226523).

The effects of salts used in typical physiological buffers were supposed to be very minimal according to the previous study that estimated the RI of NaCl solution (Kamal and Esmail, Optical Materials 1993, pp.195-199.). Nevertheless, we have directly measured the RI of iodixanol solutions and commonly used C. elegans buffers, using a refractometer, since estimation of accurate RI of complex solution was not trivial (Figure S3). We have added relevant information as follows:

Page 10, line 305:

“The RI of the 30% iodixanol solution was 1.379 ± 0.00004 at 22.5˚C according to the measurement using a refractometer at 589 nm wavelength (Figure S3). Selection of buffers should not have a large impact, since the RI of commonly used C. elegans buffers, such as M9 (RI: 1.335 ± 0.0001), egg salt (RI: 1.334 ± 0.00009), and Shelton’s growth medium (RI: 1.339 ± 0.00005) were similar at 22.5˚C (Figure S3). Based on these results, we concluded that 30% iodixanol, diluted in any common C. elegans buffers, is a RIMM for both C. elegans embryos and adult head regions.”

2) On page 9, it is stated that “image quality is accessed. How is this done precisely? It seems to me that the absolute Histone signal (not only signal to noise) should be the same form each nucleus. How much loss is there in control vs. optimized?

In the original analysis, we have quantified the ratio between peak histone-gfp signal intensity and the mean background signal intensity. However, these signals were measured in a range of 15 µm and 30 µm depths. Since nuclear position and sizes are different among images used for measurement, we realized that the original analysis might not be precise, as the reviewer 2 pointed out. Thus, in the revised Figure 2, we measured the TagRFP-PH membrane signal at 5-µm, 15-µm, and 25-µm depth. We have provided more detailed explanation in Figure S3 as to how this measurement was performed.

3) The work here was completed using oil immersion. Do the authors expect a difference when using non oil (air, water, silicon, etc) optics? It would be helpful to make a comment on this.

We were interested in performing the suggested comparison before our original submission, but were not able to do due to COVID-19. Recently our microscope equipped with a silicone-immersion objective lens has been installed, and we performed the same analysis we have done with the oil-immersion objective. As shown in the revised Figure 2D, we observed the similar extent of improvement in terms of signal intensity. Please note that the configuration of microscopes used for the original analysis (a Leica microscope equipped with an oil-immersion objective and EMCCD cameras) and the revised analysis (an Olympus microscope equipped with a silicone-immersion objective and a sCMOS camera) were slightly different. Thus, we cannot directly compare the graphs in the revised Figure 2D.

Minor suggestions:

1) There are figure legends embedded in the main text, please remove.

According to our understanding, Plos One’s author instruction requires us to insert these figure legends in the first-referred position. We will correct this if it were wrong.

2) In figure 2C is would be helpful (more clear) if the images for converted to gra

We agree that the image may be better in grayscale. In the revised Figure 2C, we have provided more clear evidence of RIMM’s role in reducing the depth-induced spherical aberration. And we have replaced the images in Figure 3B and 3C as grayscale to better show the differences in signal and noise.

Attachment

Submitted filename: review-response2.docx

Click here for additional data file. (37.2KB, docx)

Decision Letter 1

Bob Goldstein

27 Aug 2020

Improved 3D cellular morphometry of Caenorhabditis elegans embryos using a refractive index matching medium

PONE-D-20-18786R1

Dear Dr. Sugioka,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements. Thank you for sending this interesting story to PLOS ONE.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication. In addition to any required amendments you receive, please be sure to subscript the numbers 1 and 4 in "P1" and "P4" in the final manuscript, because non-subscripted P1 and P4 are postembryonic P blast cells in C. elegans.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Bob Goldstein

Academic Editor

PLOS ONE

Acceptance letter

Bob Goldstein

16 Sep 2020

PONE-D-20-18786R1

Improved 3D cellular morphometry of Caenorhabditis elegans embryos using a refractive index matching medium

Dear Dr. Sugioka:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Bob Goldstein

Academic Editor

PLOS ONE

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    S1 Fig

    (TIF)

    Click here for additional data file. (6.1MB, tif)
    S2 Fig

    (TIF)

    Click here for additional data file. (9.2MB, tif)
    S3 Fig

    (EPS)

    Click here for additional data file. (1.2MB, eps)
    S4 Fig

    (EPS)

    Click here for additional data file. (1.6MB, eps)
    Attachment

    Submitted filename: review-response2.docx

    Click here for additional data file. (37.2KB, docx)

    Data Availability Statement

    All relevant data are within the manuscript, Supporting Information files, and via Figshare: https://doi.org/10.6084/m9.figshare.c.5108522.v1.

    Articles from PLoS ONE are provided here courtesy of PLOS

    RESOURCES