Plant-parasitic nematodes are a major threat to many plant cultures worldwide, and root-knot nematodes (RKN) are responsible for extensive economic losses caused by phytonematodes (Chitwood, 2003). Cyst nematodes (CN) are also a major group of plant-parasitic nematodes causing great economic losses worldwide (Turner and Subbotin, 2013). The major genus that belongs to CN are Heterodera and Globodera.

Feeding sites induced by RKN are composed of outsized feeding cells named “giant cells” (GCs) which hold a dense cytoplasm filled with organelles and enlarged nuclei, representative of high metabolic activity. These giant-feeding cells are surrounded by mitotically active vascular tissue cells characterized by asymmetric cell wall positioning. This peculiar division pattern gives rise to a multi-layered shell consisting of small cells neighboring the GCs. This finally outlines the typical knot shape in susceptible roots called a “gall.” On the other hand, CN initiate a syncytium through elongation of a single initial feeding cell, accompanied by division of neighboring cells which will fuse as syncytia expands (Golinowski et al., 1997). Unlike RKN-induced GCs, the multinucleated state of CN-induced syncytia is achieved by cell wall dissolution of these neighboring cells (Rodiuc et al., 2014; Bohlmann, 2015) rather than by mitotic activity of the syncytia itself. Moreover, first steps of gall formation involve the redifferentiation of vascular cells into binucleate GCs which become multinucleate via succeeding nuclear divisions (Figure 1A). Nematode-induced GCs undergo multiple and often synchronous acytokinetic mitotic events (Figure 1B) (Jones and Payne, 1978; Banora et al., 2011; Vieira et al., 2013).

Giant cells of Arabidopsis hold up to 60 nuclei which become particularly enlarged, often grouped, presenting peculiar shapes with variable sizes and irregular surfaces (Figures 1C,D). Cytogenetic studies revealed that giant cell nuclei are polyploid or aneuploid (Wiggers et al., 1990; Starr, 1993). Confocal images of three-dimensional reconstruction of gall nuclei at different time points after nematode infection confirmed variable nuclei morphology and size within GCs and the multiplicity of often synchronous mitotic events (Vieira et al., 2012). DAPI stained nuclei in diploid root cells of Arabidopsis present approximately 10 chromocenters (Figure 1E) while giant cell nuclei present profuse chromocenters differently sized suggestive of nuclei polyploidization (Figure 1F). Chromocenters are densely stained heterochromatic regions matching to centromeres and nuclear organizing regions (Fransz et al., 2002). Endoreduplication is recognized to affect chromatin dynamics and the number and size of these centromeric heterochromatin domains in interphase nuclei can be indicative of endocycle activity. This type of DNA amplification is the major process involved in the increase of ploidy in GCs (de Almeida Engler et al., 1999, 2012; de Almeida Engler and Gheysen, 2013). The latter is illustrated by the increase of DNA replication cycles occurring in GCs as shown by ³H-thymidine incorporation studies (Rubinstein and Owens, 1964; Rohde and McClure, 1975; de Almeida Engler et al., 1999, 2015). Functional studies on cell cycle genes (e.g., AtCCS52B gene family) also support the implication of the endocycle in giant cell development (de Almeida Engler et al., 2012).

Up to now, there is no evidence that endomitosis may take place in GCs considering that chromosome condensation has never been observed within the nuclear membrane. Therefore, it is most likely that polyploidy in GCs is mainly achieved via the endocycle by raising global cell cycle gene expression levels within GCs. This might contribute to the overall increase in metabolic activity required for GC growth and an increased cytoplasm content essential to provide sufficient resources for nematode feeding (de Almeida Engler et al., 2012, 2015; de Almeida Engler and Gheysen, 2013).

In many somatic plant cells, DNA C-values continue to increase by endoreduplication in which nuclear DNA replication is irreversibly uncoupled from mitosis. The early literature describes that high mitotic activity in GCs may result in chromosome fusion of adjacent metaphases unequal distribution to the daughter nuclei or to incomplete mitotic events (Wiggers et al., 1990; de Almeida Engler et al., 2004). DNA content in GCs can further increase for at least 2 weeks after mitosis has ceased and flow cytometric ploidy measurements confirmed the occurrence of extensive endoreduplication in GCs (Vieira et al., 2013, 2014).

Mitotic defects and consequently nuclear phenotype changes are often observed to occur in GCs ectopically expressing cell cycle genes. Recently, it has been reported that severe morphological changes like convoluted and apparently connected nuclei are formed in lines overexpressing members of the KRP gene family known as inhibitors of cyclin-dependent kinases (CDKs) and their regulatory subunits (CYCs) [see Vieira and de Almeida Engler in same issue and (Vieira et al., 2012, 2013, 2014; Coelho et al., 2017)]. The appearance of complex nuclear phenotypes prompted us to apply other techniques besides flow cytometry to get a more accurate picture of the relation nuclear volume/shape and ploidy. Indeed, although it is generally accepted that endoreduplication leads to an increase in nuclear volume, the outsized and highly irregular nuclear shapes observed in the lines studied here, suggest that other mechanisms might be responsible for the nuclear phenotypes observed. Therefore, we opted to accurately measure nuclear volumes in GCs induced by the RKN Meloidogyne incognita using thick slices of nematode infection sites and confocal microscopy imaging. Optical sections of galls were subjected to nuclear volume measurements that were performed by two independent softwares; the public domain Volumest plugin from ImageJ and the AMIRA 3D software for visualization and analysis. Comparative volumetric measurements of GC and neighboring cells here named non-giant cells (NGC) nuclei present in wild-type (Col-0) and KRP3^OE and KRP5^OE Arabidopsis transgenic lines (Coelho et al., 2017) were performed and results are presented herein.

In plant cells, polyploidization is linked with cessation of cell division and initiation of terminal differentiation. Previous data indicate that nematodes maneuver the plant host cell cycle pushing feeding cells to undergo multiple mitotic and endoreduplication cycles resulting in polyploidy. Therefore, deciphering the plant cell cycle machinery in nematode feeding sites will contribute to better understand how nematodes manage to steer the plant cell cycle in their favor.

Classical methods like light- and transmission electron microscopy have been employed to observe nuclei and chromosomes from RKN-induced GCs (Bird, 1961, 1973; Wiggers et al., 1990; Starr, 1993). Confocal microscopy of fixed and in vivo specimens allowed us to determine the nuclear 3D shape and to measure nuclear sizes and volumes in nematode-induced GC (Vieira et al., 2012; Coelho et al., 2017). In addition, flow cytometric measurements revealed the high ploidy levels state of GC in our model host A. thaliana (Vieira et al., 2012, 2014; Coelho et al., 2017).

Volumetric information can be very helpful when studying individual nuclei with respect to the plant cell cycle in giant-feeding cells induced by RKNs. Ours and other studies have described methodologies to visualize nuclei in vivo and/or in fixed GCs in nematode feeding sites using confocal microscopy but none of these studies report GC nuclear volumetric measurements (Vieira et al., 2012, 2013, 2014; Dinh et al., 2014; Hasegawa et al., 2016; Coelho et al., 2017). Different methods have been applied to measure nematode-induced GC area to deduce corresponding volumes (Vieira et al., 2013, 2014; Cabrera et al., 2015) and purely their nuclear area (Vieira et al., 2013, 2014), but there are no reports on GC nuclear volumes. Surface measurements of GC nuclei were reported by Vieira et al. (2013, 2014) using DAPI-stained plastic embedded and sectioned galls. Measuring nuclear surfaces, however, is very labor intensive and strongly influenced by the orientation of nuclei in the section. To avoid tedious histology, Vieira et al. (2012) described a method based on the clearing of whole galls and confocal microscopy imaging. Although very informative, this method is still limited since most galls are covered by multiple layers of neighboring cells (herein named NGC) perturbing an optimal 3D reconstruction of the multiple nuclei dispersed in GCs. We adapted this protocol as outlined in Figure 2. Fixed galls were mounted in agarose, sectioned with a vibroslicer, cleared and DAPI stained and finally mounted between two large coverslips while avoiding tissue compression. Thicker samples can, therefore, be imaged at both sides generating high-quality images even for very large galls. Applying the same methodology, galls in other hosts, as well as other complex plant organs for, e.g., in tobacco, rice, soybean, and cotton, can be analyzed. Here, optical sections were generated according to Coelho et al. (2017) by confocal microscopy and image datasets were processed for volumetric measurements using two different softwares: the free Volumest plugin tool from ImageJ and the AMIRA 3D analysis package. These approaches enabled us to compare nuclear volume between galls at different time points, among two KRP^OE transgenic lines compared to wild-type.

The first nuclear volume quantification method we used made use of “Volumest,” an ImageJ plug-in. We measured a user-defined area(s) in a single cross section and, the plug-in generated a three-dimensional volume calculation (Merzin, 2008) based on a series of z-stack images. This plug-in was originally described to be practiced in radiology (Merzin, 2008) and later to measure cell and nuclear volumes of Drosophila brain and epithelium cells (Tamori and Deng, 2013; Doll and Broadie, 2015). Herein, we applied “Volumest” to measure nuclear volumes in nematode-induced GC from wild-type and transgenic Arabidopsis lines overexpressing the cell cycle inhibitors genes KRP3 and KRP5. Since GC nuclei in KRP3^OE and KRP5^OE lines presented peculiar elongated and apparently connected shapes (Figures 5, 8, 9), we compared their volumes with wild-type GC nuclei. Despite the observed variation in nuclear volume amongst GCs in different lines evaluated, nuclei in KRP3^OE and KRP5^OE lines are consistently larger in volume than in wild-type. These data suggest that defective mitotic events occurring in GC may prevent complete nuclear division and thereby lead to the abnormal increase in nuclear volume. This result is in agreement with the fact that even though there is only a slight increase in ploidy level for the KRP3^OE line, a considerable increase in nuclear volume is observed. Moreover, despite the vast increase in GC nuclear volume in the KRP5^OE line, previous flow cytometry data showed no increase in ploidy levels for this line supporting the idea that the nuclear volume increase observed for these mutants is not caused by an increase in DNA content but rather a consequence of mitotic defects (Coelho et al., 2017). Using the “Volumest” method we were unable to observe minor differences in nuclear volumes of NGC for the three different lines.

As a second measuring approach we used AMIRA, a commercial software package which is based on morphological segmentation of the image data set. This segmentation relies on a white top-hat transformation to extract objects (here nuclei) from the data set. The size of the extracted object is conditioned by a self-chosen kernel size for a mathematical operation called “opening.” In digital image processing, an opening is the dilation of the erosion of the image by a structuring element named a kernel matrix. While opening removes small objects from the foreground, it can also be used to extract specific features from the image that fits the kernel matrix. Segmented objects are finally properly labeled and analyzed for volumetrics. Since the latter method involves image processing before measuring, it is expected to be less depending on manual errors.

Interestingly, measurements by AMIRA revealed that GCs nuclei in galls 21 DAI in the KRP3^OE and KRP5^OE lines increased up to three times in volume compared to wild-type GC nuclei. A slight increase of NGC nuclear volume was also observed for both KRP lines within a range of 1.5–2 times for KRP3^OE and KRP5^OE 21 DAI as compared to wild-type NGC nuclei. These results suggest that ectopic KRP3 and KRP5 expression in galls promotes an increase in the nuclear volume of GC and to a lesser extent but also true for NGC nuclei.

The results obtained from both volumetric measurement methods show that giant cell nuclei are significantly bigger in both KRP3^OE and KRP5^OE lines than in the wild-type. These data together with volume rendering of giant cell nuclei suggest that KRP3^OE and KRP5^OE lines accumulate mitotic defects which can lead to disturbed mitotic events and erroneous nuclear division. These defects may explain the elongated and undulated nuclear phenotype observed in GC of both KRP overexpressing lines (seen in Figure 5). Despite that both analysis point in the same direction, nuclear volume differences are more pronounced when using the AMIRA approach. The latter also enabled to detect minor but significant volumetric variation when comparing nuclei from different plant lines. Several criteria may explain such differences: a lack of reproducibility when using Volumest which relies on the manual delineation of each nucleus determined by the user, and its limited ability to discriminate all shades of gray preventing the accurate delineation of nuclear boundaries. This error may affect the volume output and the observed differences between the two methods. Besides, the highly irregular nuclear shapes observed in the KRP lines may hamper the selection of a single nucleus which actually might consist of two entangled entities in very close proximity. A manual measuring procedure will be more prone to introducing this type of error. Since Volumest seems less accurate, it may also have prevented to observe the increase of NGC nuclear volume in KRP3^OE and KRP5^OE lines during the nematode infection process.

Although our analysis revealed that the Volumest approach seems to provide somewhat less accurate results than AMIRA, the former still features some advantages: -it is a free public domain software, -there is no need for additional mathematical operation and, -it is easy in use relying on hand-made object delineation. Since “Volumest” is supported by ImageJ, it is up to the user to implement additional data processing tools such as the “top-hat” approach prior to “Volumest” analysis.

It is important to mention that recently Ohtsu et al. (2017) have described a method for imaging CN-induced syncytia using the ClearSee clearing method (Kurihara et al., 2015). In this work, the authors have used the software IMARIS (Bitplane, United Kingdom) to assess syncytia nuclear volume. The authors were able to nicely image whole syncytia samples of Astragalus sinicus roots infected with Heterodera glycines. The strategy we propose here and which is based on root vibroslicing combined with nuclear volume measurement (ImageJ and AMIRA) can also be applied for much thicker roots infected with cyst or RKN. Currently, thick vibrosliced samples are being optimized using different clearing approaches for deep tissue imaging.

Summing up, “Volumest” ImageJ and AMIRA softwares have been applied to determine nuclear volumes of GC and NGC nuclei from transgenic lines overexpressing KRP3 and KRP5 compared to wild-type. Both lines presented exceptionally elongated and apparently connected nuclear morphology. These phenotypes are characteristically occurring in some KRP overexpressing lines and were enhanced in mature GC (21 DAI) compared to young feeding cells (5–10 DAI), most likely due to cumulative defective mitotic events (de Almeida Engler et al., 2012; Vieira et al., 2012). Remarkably, differences observed in nuclear ploidy and volume in GC show that an increase in nuclear volume is not necessarily linked to ploidy levels. The evident accumulation of mitotic defects in GC nuclei upon overexpression of KRP3 and KRP5 could to a certain degree prevent optimal cell cycle progression. The two methods used here for volumetric measurements provide valuable information that complements former data obtained by flow cytometry. Previous records combined with the results obtained here suggest that the increase in nuclear volume, as well as the formation of highly extended and undulated nuclei in GCs overexpressing KRP3 and KRP5, may largely result from mitotic defects such as incomplete chromosomes separation rather than endoreduplication.

JDASJ, OP, GE, and JdA have conceived and designed the paper; JDASJ, OP, and RC have performed the experiments; JDASJ, OP, RC, JdA, and GE have analyzed the data; MG-d-S, GE, and JdA have contributed with reagents/materials/analysis tools; JDASJ, JdA, and GE have written the paper; MG-d-S, OP, GE, and JdA, have amended the paper (helped writing legends, M&M, added suggestions).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.