A new and improved algorithm for the quantification of chromatin condensation from microscopic data shows decreased chromatin condensation in regenerating axolotl limb cells

Abstract

The nuclear landscape plays an important role in the regulation of tissue and positional specific genes in embryonic and developing cells. Changes in this landscape can be dynamic, and are associated with the differentiation of cells during embryogenesis, and the de-differentiation of cells during induced pluripotent stem cell (iPSC) formation and in many cancers. However, tools to quantitatively characterize these changes are limited, especially in the in vivo context, where numerous tissue types are present and cells are arranged in multiple layers. Previous tools have been optimized for the monolayer nature of cultured cells. Therefore, we present a new algorithm to quantify the condensation of chromatin in two in vivo systems. We first developed this algorithm to quantify changes in chromatin compaction and validated it in differentiating spermatids in zebrafish testes. Our algorithm successfully detected the typical increase in chromatin compaction as these cells differentiate. We then employed the algorithm to quantify the changes that occur in amphibian limb cells as they participate in a regenerative response. We observed that the chromatin in the limb cells de-compacts as they contribute to the regenerating organ. We present this new tool as an open sourced software that can be readily accessed and optimized to quantify chromatin compaction in complex multi-layered samples.

Fig 1. Improved computational efficiency of new chromatin quantification algorithm.

Plots represent the amount of time (in seconds) it takes for the new algorithm (Sosnik et al.) and the previously published algorithm (Irianto et al.) to process different amounts of data. (A) Images ranging from 2048 by 2048 pixels to 128 by 128 pixels in 8 bits of data depth were analyzed with both algorithms. The Irianto et al. was tested using the reported algorithm (Irianto et al.) and a modified algorithm in which we eliminated the pixel reduction (compression) (Irianto et al. uncompressed). n = 15 images per size. Bars represent standard error. Analysis of variance followed by Tukey’s multiple comparison tests revealed significant to very significant differences (p < 0.05–0.01) between the Sosnik et al. algorithm and the Irianto et al. algorithm for small images (128–512). The comparison of the Sosnik et al. algorithm and the Irianto et al. uncompressed algorithm for larger images was not significantly different. (B) Images ranging from 2048 by 2048 pixels to 128 by 128 pixels in 16 bits of data depth were analyzed with both algorithms. n = 15 images per size. Bars represent standard error. Student t test for each image size revealed very to extremely significant (p = 0.0028 - <0.0001) differences between the Sosnik’s and Irianto’s algorithms.

Fig 2. Representative images of chromatin compaction in developing spermatogonia.

Images were obtained of histologically or DAPI stained tissue sections of adult zebrafish testes. (A-D) Representative fluorescent images of DAPI stained nuclei of B type spermatogonia (B), spermatocytes (C), and spermatids (D). Representative bright field images of B type spermatogonia (B’), spermatocytes (C’), and spermatids (D’) that have been stained with hematoxalin and eosin to characterize their basic cellular morphology. Scale bar in A is 10 μm, and in B-D’ is 5 μm.

Fig 3. Quantification of the chromatin compaction index during spermatogenesis.

The average compaction index of the DAPI stained nuclei of Spermatogonia B (n = 15 samples), spermatocytes (n = 20 samples), and spermatids (n = 15 samples) was quantified using the algorithm described in the materials and methods section. The histogram represents the average compaction index of the nuclei that were quantified in each cell type. Error bars are standard error. Analysis of variance followed by Bonfarroni’s multiple comparison test indicates that each bar is significantly different from all other samples (range of p = 0.01–0.001).

Fig 4. Use of the new quantification algorithm to characterize chromatin modifications in regenerating Axolotl limb cells.

Fluorescent images were obtained of the DAPI stained nuclei in tissue sections of regenerating axolotl limbs at the early-blastema stage (A), and the chromatin compaction index was quantified. The red line indicates the amputation plane to indicate where the mature tissue ends and the blastema tissue begins, the red arrow points to the wound epithelium covering the blastema mesenchyme, and the blue arrow points to the dermal layer below the epidermis in the mature skin. Representative image of a DAPI stained nucleus of a mature (uninjured) cell in the dermis (B) and in the early regenerating blastema (C). (D) The chromatin compaction index was quantified as described in the materials and methods section and the average compaction index was plotted for mature (n = 30 samples) and regenerating (n = 30 samples) limb cells. Error bars are standard error. Analysis of variance was performed for statistical analysis (p<0.0001).