Date Published: December 14, 2016
Publisher: Springer Berlin Heidelberg
Author(s): Maximilian Bonta, Szilvia Török, Balazs Hegedus, Balazs Döme, Andreas Limbeck.
Laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) is one of the most commonly applied methods for lateral trace element distribution analysis in medical studies. Many improvements of the technique regarding quantification and achievable lateral resolution have been achieved in the last years. Nevertheless, sample preparation is also of major importance and the optimal sample preparation strategy still has not been defined. While conventional histology knows a number of sample pre-treatment strategies, little is known about the effect of these approaches on the lateral distributions of elements and/or their quantities in tissues. The technique of formalin fixation and paraffin embedding (FFPE) has emerged as the gold standard in tissue preparation. However, the potential use for elemental distribution studies is questionable due to a large number of sample preparation steps. In this work, LA-ICP-MS was used to examine the applicability of the FFPE sample preparation approach for elemental distribution studies. Qualitative elemental distributions as well as quantitative concentrations in cryo-cut tissues as well as FFPE samples were compared. Results showed that some metals (especially Na and K) are severely affected by the FFPE process, whereas others (e.g., Mn, Ni) are less influenced. Based on these results, a general recommendation can be given: FFPE samples are completely unsuitable for the analysis of alkaline metals. When analyzing transition metals, FFPE samples can give comparable results to snap-frozen tissues.
Laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) is a tool nowadays widely used for the laterally resolved assessment of trace elements in biological tissues . Exceptional limits of detection for a wide range of metallic analytes and good lateral resolutions combined with reasonable acquisition times make this technique a valuable tool for such imaging applications [2, 3]. During the last decades, a variety of works on bio-imaging using LA-ICP-MS have been presented and the technique has already been included in clinical research in some examples. The tissue type investigations have been performed on (e.g., liver , brain [5, 6], kidney , or tumor tissues [8, 9]) are as numerous as the elements that have been analyzed already; here, naturally occurring minor (e.g., Na, K) and trace elements (e.g., Ni, Cu, Zn), as well as elements artificially introduced during chemotherapy (e.g., Pt [8, 9]) or being used as contrasting agents (e.g., Gd ), have to be named in this context. LA-ICP-MS does not only give the opportunity of acquiring qualitative distribution images but also quantification is possible [11, 12]. With regard to the quantification strategy, a wide variety of approaches has been presented in the past, as LA-ICP-MS suffers from severe matrix effects and slight variations in tissue composition may already induce unwanted signal changes that do not reflect the actual composition of the sample. Thus, the use of appropriate standard materials, as well as employing a suitable internal standard, is imperative . As certified reference materials are scarcely available for biological tissues, alternative approaches have to be used ranging from the preparation of gel standards , over printed patterns on paper , to the use of polymeric layers , or the online addition of aqueous standards . Also concerning internal standardization, possibilities are various, including not only the sample inherent carbon but also additionally applied materials/reagents such as thin gold layers [17, 18], protein-metal tags , or polymer thin films . Each of the named approaches has their advantages and weaknesses. Their description is, however, not within the scope of this work.
With six tissue types, each measured in three replicates, and two different sample preparation strategies, a total of 36 tissue samples were investigated in this comparative study between FFPE and snap-frozen tissue preparation. Dimensions of the tissues ranged from approx. 3 × 3 to 8 × 8 mm2. LA-ICP-MS measurements provided signals above background level for all selected elements on all samples. Normalized signal intensities for each pixel were within the signal range obtained for the dried droplet calibrations. Quantification of the obtained signal intensities was performed by transforming the data matrix of each element using the calibration functions from the dried droplet calibration. Derived element concentrations ranged between 0.1 μg g−1 (Mn) to 3000 μg g−1 (K). Whenever possible, two isotopes of each element were monitored and quantification was performed. The isotope with the higher natural abundance was always used for data evaluation, the other one for quality control purposes. In all cases, the distribution images of two isotopes were well correlated; calculated analyte concentrations did not differ significantly. Repeatability of the measurements was validated using consecutive tissue slices.
FFPE is a widely used preparation strategy for tissue specimen. Due to the large availability of samples in archives, it would be interesting if these samples could also be used for elemental distribution studies. In the comparison with snap-frozen tissues which were used as reference samples, elemental distributions as well as average metal concentrations showed to be partly altered in the FFPE samples, which restricts the general applicability of FFPE sample preparation in metallomics studies. FFPE samples showed to be completely unsuitable, if the investigation of alkaline metals is required; distributions as well as absolute amounts were heavily influenced. Especially the relative distributions of some transition metals as well as their bulk concentrations showed to be altered to a lesser extent in the FFPE samples. These results indicate that if only the analysis of transition metals is required, also FFPE samples can be used; even if finer structures appeared to be blurred, the general elemental distributions were comparable. For Ca and Zn, contaminations introduced during the embedding process could be identified, highlighting that metal analysis in FFPE samples still has to be carefully considered in order not to obtain results which are biased by the sample preparation process. However, even if snap-frozen samples without fixation definitely provide the most accurate way for analysis of metals in tissue samples, FFPE samples can also deliver valid quantitative elemental distribution results.