Spatial mapping of metals in tissue-sections using combination of mass-spectrometry and histology through image registration

Abstract

We describe a new procedure for the parallel mapping of selected metals in histologically characterized tissue samples. Mapping is achieved via image registration of digital data obtained from two neighbouring cryosections by scanning the first as a histological sample and subjecting the second to laser ablation inductively coupled plasma mass spectrometry. This computer supported procedure enables determination of the distribution and content of metals of interest directly in the chosen histological zones and represents a substantial improvement over the standard approach, which determines these values in tissue homogenates or whole tissue sections. The potential of the described procedure was demonstrated in a pilot study that analysed Zn and Cu levels in successive development stages of pig melanoma tissue using MeLiM (Melanoma-bearing-Libechov-Minipig) model. We anticipate that the procedure could be useful for a complex understanding of the role that the spatial distribution of metals plays within tissues affected by pathological states including cancer.

Figure 1. A schematic description of the whole transformation process from cryosections to matched layers representation and statistical evaluation.

(A) The two cryosections were sliced from tissue, (B) the narrow slice is used for histological analysis and to find uniform spots - red - GMT (normally growing melanoma tissue), violet - ESR (early spontaneous regression), yellow - LSR (late spontaneous regression) and green - FT (fibrous tissue). (C) The wide slice is photographed and (D) measured by LA-ICP-MS. (E) The measured elemental maps are normalized with respect to dry weight. The slices are registered in two steps. (F) Firstly, the normalized metal maps are registered with the slice photography and secondary, the result of the first step is registered with the histological scan. The registration is based on silhouette registration. (G) The output of the process is a layered representation of tissue consisting of a histological layer with selected spots and metal layers. (H) The layered representation of all tissues is statistically evaluated. Our data confirm the hypothesis that content of zinc in the zone of growing melanoma tissue (GMT) is significantly greater than in all remaining zones.

Figure 2

Four histologically differing zones identified in haematoxylin-eosin stained skin porcine melanoma: (A) growing melanoma tissue (GMT), (B) melanoma tissue with early destruction of melanoma cells (early spontaneous regression-ESR), (C) melanoma tissue with late destruction of melanoma cells (late spontaneous regression-LSR), (D) fibrous tissue (FT) with a few remaining melanoma cells. Scale bar = 50 μm (* = melanoma cells; + = cellular debris from damaged melanoma cells; x = fibrous tissue).

Figure 3

(A) LA-ICP-MS signal in line scan mode recorded for laser beam pass across one printed line at laser spot diameter of 100 μm, scan speed of 20 μm s⁻¹, laser beam fluence of 8 J cm⁻² and repetition rate of 10 Hz. (B) Duration of scanning of the sample area of 15 × 15 mm at various scan speeds (μm s⁻¹). (C) Relative broadening of a printed line (expressed in %) with width of 800 μm obtained at various scan speeds (μm s⁻¹). (D) Elemental maps of C, Si and Zn obtained at “soft” ablation parameters (2 J/cm⁻²) for tissue K320/1 (12 weeks old). (E) Elemental maps of C, Si and Zn obtained at “hard” ablation parameters (8 J/cm⁻²) for tissue K320/1 (12 weeks old).

Figure 4

Photographs of two neighboring cryosections prepared to be subjected to laser ablation and to histological analysis (A,B). The image C is the result of registering the images (A,B). The blue rectangle indicates in the images (A–C) the minimal rectangle (with sides parallel to axes) the image fits in. The red line in the image (A,C) accentuates orientation of the corresponding borders on both the images. Images (B,C) are compared in the image (D): while the places appearing in both images have black color, the space in blue corresponds to symmetric difference of both images as explained in Supplementary Note 1.

Figure 5. The illustration of the SSD criterion for image registration.

The panels (A,B) show the reference image and the image to be registered. The panel (C) shows overlaying of both the images, the gray part corresponds to the difference between the images. The panel (D) shows the difference on itself. As the patterns are 3 × 3 pixels, the difference part corresponds to SSD of 4.

Figure 6

Average content of Zn and Cu in all samples (A,D) and in samples stratified according to age of minipigs (B,E). Box plots (C,F) visualize dispersion of data corresponding to average Zn and Cu content in specified tissue zones by samples as well as by individual samples as a whole for all 10 considered minipigs (the white box plot).

Figure 7. A schematic depiction of a plausible role of zinc in melanoma progression.

Marginal parts of tumour border with a reactive zone of an adjacent healthy tissue. In these parts of melanoma, frequent necroses, a presence of H₂O₂ and elevation of malondialdehyde (MDA) can be found. (A) High zinc in marginal sections of melanoma may indicate a physiological defence of zinc-containing antioxidant molecules against the potential danger of ongoing oxidative stress. Although intralesional regions of melanomas generally exhibit hypozincemia status, oxidative stress in reactive zones/marginal parts interface can likely result in an oxidative stress-triggered transport of zinc-containing antioxidant molecules from adjacent healthy tissue and temporary accumulation of zinc. (B) Physiologically, an expression of melanosomes is regulated by microphthalmia-associated transcription factor gene (MITF) and phosphorylation (P) of the homonymous MITF-encoded protein (Mitf). (C) Translated melanosomes are the place for entire melanin synthesis, starting by action of tyrosinase, producing L-3,4-dihydroxyphenylalanine (DOPA) from tyrosine (Tyr). Tyrosinase further produces DOPA quinone from DOPA. DOPA quinone can be converted through a sequence of reactions to various types of melanins (black and brown eumelanins or red to yellow pheomelanin). (D) Noteworthy, another plausible reason for the free radicals formation in melanoma is a presence of abnormal and incomplete melanosomes, causing a significant leakage of the reactive melanin prescursors, causing oxidative stress in the pigmented tumours through redox cycling and an accumulation of zinc ions originating from antioxidant molecules from adjacent tissues as a kind of physiological protection.