Depicting the spatial distribution of proteins in human tumor tissue combining SELDI and MALDI imaging and immunohistochemistry

Abstract

Carcinoma tissue consists of not only tumor cells but also fibroblasts, endothelial cells or vascular structures, and inflammatory cells forming the supportive tumor stroma. Therefore, the spatial distribution of proteins that promote growth and proliferation in these complex functional units is of high interest. Matrix-assisted laser desorption/ionization imaging mass spectrometry is a newly developed technique that generates spatially resolved profiles of protein signals directly from thin tissue sections. Surface-enhanced laser desorption/ionization mass spectrometry (MS)combined with tissue microdissection allows analysis of defined parts of the tissue with a higher sensitivity and a broader mass range. Nevertheless, both MS-based techniques have a limited spatial resolution. IHC is a technique that allows a resolution down to the subcellular level. However, the detection and measurement of a specific protein expression level is possible only by semiquantitative methods. Moreover, prior knowledge about the identity of the proteins of interest is necessary. In this study, we combined all three techniques to gain highest spatial resolution, sensitivity, and quantitative information. We used frozen tissue from head and neck tumors and chose two exemplary proteins (HNP1-3 and S100A8) to highlight the advantages and disadvantages of each technique. It could be shown that the combination of these three techniques results in congruent but also synergetic data.

Figure 1

(A) Microdissection of an unstained tissue section. The red lines demonstrate the course of the laser shots during the run. Necrotic tumor tissue (area between I and II) and native tumor tissue (area between II and III and area encircled by I) were microdissected. (B) Surface-enhanced laser desorption/ionization (SELDI) analysis of the microdissected tissue regions. The SELDI spectrum in the range of 3000–4000 Da of the tissue shows characteristic peaks of 3368, 3440, and 3485 Da in the necrotic tumor tissue (area between I and II). (C) The SELDI spectrum in the range of 10,600–11,200 Da of the necrotic tumor tissue (tissue area between I and II) and the native tumor tissue (tissue areas between II and III and area encircled by I). The peak of 10,860 Da shows a higher expression in the native tumor tissue compared with the necrotic tumor tissue. (D) Matrix-assisted laser desorption/ionization (MALDI) imaging (1 and 3) and IHC (2 and 4) of human defensins (1 and 2) as well as S100A8 (3 and 4). The human defensins were almost expressed only in the necrotic tumor tissue (1 and 2). S100A8 shows a higher expression in the native epithelial tumor tissue than in the necrotic tumor tissue (3 and 4). N, necrotic tumor tissue; Tu, native epithelial tumor tissue (IHC does not provide exactly the same detail as MALDI image).

Figure 2

MALDI–imaging mass spectrometry (IMS): simultaneously expressed proteins of different m/z ratio (Da) showing the same detail of a head and neck carcinoma. Top left: HE-stained tissue after MALDI–IMS. Panels with red frame show defensin and S100A8. Scale 0–100% represents semiquantitative estimation of protein expression and protein gradients.

Figure 3

IHC for defensins (A) and S100A8 (B). S, tumor stroma.

Figure 4

IHC for S100A8.