Introduction
Matrix-assisted laser desorption/ionization mass spectrometry imaging
(MALDI-MSI) is a technique that can directly identify and localize hundreds to
thousands of molecules in a tissue sample simultaneously, including metabolites,
lipids, glycans, peptides, proteins, glycolipids, and drugs and their metabolites
[1]. The molecular distribution can then
be correlated with the histomorphology. Unlike conventional molecular analysis
methods of tissue samples [e. g., immunohistochemical (IHC) and
histochemical techniques], which specifically detect individual compound
localizations via antibody applications or compound classes such as polysaccharides
[2], MSI-based techniques provide a
simultaneous label-free detection of various types of species in the same tissue
section. The cell-specific molecular signatures accessible through MSI are hugely
complementary to histopathological, IHC and other established molecular
(e. g., genomics-related) data [1]
[3] and therefore offer a major advantage for
synergizing molecular information and morphology. Consequently, MSI has rapidly and
substantially impacted basic and clinical research by identifying molecular changes
associated with disease. MSI is a promising tool to expand the field of molecular
pathology to improve diagnostics, prognostics, and theranostics for various tumor
types [4]
[5]
[6].
Metabolomics can be broadly defined as the comprehensive analysis of all small
molecule metabolites (i. e.,≤1500 Da) in a biological system. These
small molecular weight compounds, including small peptides, oligonucleotides,
sugars, nucleosides, organic acids, ketones, aldehydes, amines, amino acids, lipids
and steroids, are present in cells, tissues and/or organisms [2]. Metabolism alteration occurs in various
human diseases, and metabolic reprogramming is a core hallmark of cancer [7]
[8]
[9]. Qualitative and quantitative analysis of
metabolites in biological systems are facilitated by various analytical methods,
particularly MS (e. g., MALDI-MS, liquid chromatography-MS and gas
chromatography-MS) [9]. Among these and other
frequently used analytical assays, which usually homogenize and extract the species,
MSI is the leading technique for conducting spatially resolved metabolome
measurement either in fresh frozen (FF) tissue sections or formalin-fixed
paraffin-embedded (FFPE) tissue [10]
[11]. The preparation of FF samples can be
particularly difficult (i. e., collection, handling and storage in the
routine setting), whereas the vast majority of FFPE specimens (biopsies and surgical
resections) are prepared in pathology laboratories. Several investigators have
conducted MALDI-MSI analysis of FFPE tissue specimens (reviewed in [2]
[10]). A proof-of-principal study of the
spatial and chemical conservation of metabolites in archived FFPE tissue samples
reports that 72% of the m/z species overlap between
FFPE and FF tissue sections [11]. Based on
this work, a detailed analytical protocol for measuring molecular spatial
distributions in FFPE samples using MALDI Fourier-transform ion cyclotron resonance
(FTICR) MSI platform has been published [12].
The validity of the MSI protocol for FFPE tissues was demonstrated by a multicenter
interlaboratory round robin study which showed a high level of between-center
reproducibility of FFPE tissue metabolite data [13].
The adrenal gland plays a key role in the regulation of metabolic, cardiovascular,
immune, neuronal, and mental processes through the action of two major types of
hormones, steroids and catecholamines, which are produced in the outer cortex and
inner medulla regions, respectively. Although both tissues of the adrenal gland
descend from different origins during embryonic development and produce different
types of hormones, the adrenal cortex and medulla interact with and influence each
other under both healthy and pathological conditions in multiple ways, including
through the blood flow that delivers steroids collected in the cortex region to the
medulla and through contact between different cells at the cortical-medullary
junction region [14]. In addition, the adrenal
gland is in bidirectional crosstalk with the major systemic functions in the context
of systemic disease, such as obesity and inflammation. Metabolites can interact
directly or indirectly with molecular targets, influencing the risk and
complications associated with adrenal gland disease or with adrenal dysfunction. The
functional integrity of this organ system therefore is essential for hormonal
homeostasis, and adrenal dysfunction may contribute to various metabolic and
cardiovascular disorders and stress-related diseases [15]
[16]
[17]
[18].
As an emerging tool and leading technique for spatially resolved molecular
assessment, MSI facilitates the study of the critical role of the adrenal gland in
physiological and pathophysiological conditions. This technology fills the gap for
visualizing the distributions of hormones, metabolites and drugs in the cortex and
medulla of the adrenal gland and other organs or tissues from animal models to
patients. Mapping metabolites and hormones to their respective pathways provides a
means to image the activities of pathways in tissues [19]
[20]
[21]. Molecular and metabolomic pathway
changes associated with adrenal diseases can be discovered under the context of
their highly structured histology and interactive functionality among diverse
cortical and medullary cell components by this methodology. Importantly, the tissue
section remains intact during MALDI analysis, which allows for further
co-registration analyses to be conducted after the measurement has been taken. This
feature therefore provides great potential for addressing highly interesting and
unanswered questions in the study of the adrenal gland. The precise localization of
species within specific cells may enable the study of the aforementioned spatial and
functional interaction between the cortex and medulla regions. In this review, we
described the recent applications of in situ mass spectrometry imaging in adrenal
gland study, including molecular imaging of adrenal gland physiology as well as
pathophysiological molecular changes of adrenal diseases. MSI provides an advanced
methodology for gaining a deeper insight into adrenal function, as well as
clinically relevant tissue molecular and functional adaptation in adrenal-related
disease.
In situ MALDI-MSI used in tissue studies
MSI was developed from secondary ion mass spectrometry (SIMS), which introduces a
pulsed ion beam moving across the sample’s surface to generate a
secondary ion to be directed to the mass spectrometer inlet [22]
[23]. Among various sample introduction
methods, matrix-assisted laser desorption/ionization (MALDI) is one of
the most commonly applied ionization techniques coupled with MSI analysis, and
was applied to biological tissue sections in the mid-to-late 1990s. MALDI-MSI
was first used to visualize proteins and then quickly was extended to other
endogenous and exogenous molecular classes by utilizing other matrices and
sample preparation optimization [23]
[24]
[25]
[26]. MALDI-MSI has been a powerful tool
for metabolome study in tissue samples. In a typical MALDI-MSI approach, tissue
sections are mounted on a MALDI-compatible target carrier and then coated with a
matrix [27]. The laser beam probes a
specific location on the matrix-coated tissue in a predefined raster. The matrix
absorbs and transfers laser energy to the analytes, thus enabling desorption and
ionization. A mass spectrum is therefore generated at each measurement spot. The
generated MS data can then be co-registered to the histologically stained
section to enable visualization of the molecular content within a tissue
morphological context [12]. Depending on
the molecular classes of interest, pre-treatment of samples may be necessary to
detect specific compound/compound classes or to improve sensitivity.
These pre-treatments include washing to attenuate the influence of high
abundance compounds on low abundant molecules, on-tissue digestion for protein
or glycan analysis, and on-tissue derivatization to improve sensitivity for
specific drugs or metabolites [28]
[29]
[30].
Post-measurement data processing and statistical analysis are a key step in
MALDI-MSI analysis [23]
[31]. Processing of MALDI-MSI data must
consider the nature of spatially resolved mass spectra for each sample, with
each mass spectrum containing hundreds to thousands of mass signals [31]
[32]. The basic data processing pipeline
typically involves raw data exporting, pick alignment, pick peaking, and peak
annotation. Peak annotation can be done either by MS/MS experiments
directly from tissue and/or by a database search of measured
m/z values [23].
The commonly used public databases for MSI annotation of small molecules are
HMDB (http://www.hmdb.ca/), METLIN
(https://metlin.scripps.edu), lipid maps
(http://www.lipidmaps.org/), and/or METASPACE
(http://annot ate.metaspace2020.eu) [33]
[34], Accurate mass measurement, coupled
with tandem MS, enables small endogenous compounds to be unambiguously
identified directly from the tissue [28].
Identification of metabolites is a continually developing field in metabolomics.
Comparison of the mass spectral data against literature and databases is
important for metabolite annotation. The continual development of metabolite
databases, bioinformatics tools and MS instrumentation lead to improvements in
metabolite annotations and structural identifications [35]. It is a challenge to identify and
characterize the structures of unknown metabolites that are not available in
databases. In silico analyses have been developed to assist with identification
of metabolites for which MS data are not available in databases. In these cases,
analysis of the fine isotope pattern of the m/z of interest using Bruker
DataAnalysis and SmartFormula can be useful to predict molecular formula and
possible unknown metabolite structures [11]
[12]. Furthermore, it may be necessary to
identify unknown metabolites using nuclear magnetic resonance (NMR) analysis or
extensive chemical synthesis to enable structural comparisons using tandem mass
spectrometry (MS/MS) [36]
[37].
Due to the amount of MSI data (i. e., in the gigabyte range), the
processing and handling of the data have become computationally demanding.
Commercial software packages, such as Multimaging (Imabiotech, France) or SCiLS
(Bruker Daltonics, Germany), can handle terabyte-sized, multi-sample data sets
and include many statistical tools for biomarker discovery [23]. In addition to classical statistical
tests, supervised and/or unsupervised statistical analyses have been
applied to MSI data analysis [31]
[38]
[39]
[40]
[41]
[42]. A schematic of the MALDI-MSI concept
with post-measurement data handling is provided in [Fig. 1], [43].
Fig. 1 Representation of the matrix assisted laser
desorption/ionization mass spectrometry imaging (MALDI-MSI) workflow:
matrix deposition, in situ laser desorption mass spectrometry analysis,
hematoxylin and eosin (H & E) staining, and co-registration of
MSI data with the H & E image to achieve H & E-guided
extraction of mass spectra information. (Reprinted from [43] with permission from Elsevier)
[rerif].
The application of MSI to research questions of biological and clinical
significance has increased exponentially since the introduction of the
technology [28]. MSI has been applied by
many studies to investigate complex structures in tissue, including in brains,
eyes, joints, wounds and atherosclerotic plaques. Oncology, however, has been
and still is the driving field for MSI. In this context, MSI has been used to
extract tumor-specific molecular information for the purpose of biomarker
discovery to improve diagnosis and prognosis as well as to study intra-tumor
heterogeneity and classification of tumor margins. Its applications generally
cover all stages of translational research from pre-clinical to clinical studies
[4]
[5]
[6]
[23]
[44]
[45]
[46]
[47].
MSI analysis has enabled access to a variety of molecules in the adrenal gland.
In the healthy human adrenal gland, a wide spectrum of endogenous metabolites,
including catecholamines, sterol and steroid metabolites, nucleotide
derivatives, intermediates of glycolysis and the tricarboxylic acid (TCA) cycle
and lipids and fatty acids, have been measured by MALDI-FTICR-MSI analysis [20]. Under pathophysiological conditions,
endogenous metabolites and various hormones and their derivatives
(e. g., 18-hybrid steroids) also were detected in FFPE specimen from
adrenal tumors [21]
[48]
[49]. These findings have led to deeper
insight into the metabolites and hormone adaptation in the functional and
dysfunctional adrenal gland and adrenal tumors as will be shown in later cases.
Specific to the adrenal gland, a few studies also have focused on the
improvement of the detection and localization of adrenal-specific hormones,
including for catecholamines [50]
[51]
[52] and particularly for poorly ionized
and isomeric steroids, leading to new hypothesis of adrenal disease
pathogenesis, as will be reported later in this review [49]
[52]
[53].
MSI method adaption for visualizing steroids in adrenal cortex (FF
tissue)
Although high-resolution mass analyzers such as FTICR-MS (i. e., 100 000
mass resolution for common tissue metabolites) allow relatively accurate and
confident assignments of ion peaks to chemical formulas, a lack of detection
sensitivity can hinder its application to the molecular imaging due to low
ionization efficiency, ion suppression by abundant species, and background
spectral interference from either the matrix and/or tissue components
[51]
[52]. Low sensitivity in the adrenal gland
MSI study is particularly prominent for neutral steroids as they have relatively
low abundance, poor ionization yields (i. e., lack of hydrogen donor or
acceptor moieties), and susceptibility to ion suppression of lipids and proteins
[53]. Another challenge for MSI
molecular assignment is that it cannot discriminate between structural isomers,
which occur frequently for steroids [52].
Four pairs of structural isomers can be found in human adrenocortical
steroidogenesis pathways, including 18-hydroxyl corticosterone and cortisol,
cortisone and aldosterone, corticosterone and 11-deoxycortisol, and
11-deoxycorticosterone and 17-hydroxyprogesterone. Differentiating these
steroids during in situ on-tissue MSI analysis is challenging but provides the
opportunity to study their biological activity and communication within their
producing context. With regard to these two major challenges, in situ chemical
derivatization in combination with tandem MS techniques has demonstrated its
usefulness in steroid imaging in the adrenal gland. Chemical derivatization
introduces polar functional groups resulting in enhanced ionization yields,
whereas tandem mass fragmentation produces characteristic fragments that are
particularly useful for structure identification and differentiation of
structural isomers.
To enhance the sensitivity of neutral steroid MSI analysis, Cobice et al.
explored derivatization methods and found that on-tissue chemical derivatization
using GirT reagent has the highest sensitivity for 11-dehydrocorticosterone and
corticosterone, achieving low limit of detection and high signal-to-noise ratio
(>100) [53]. Such derivatization
reactions target the α,β-unsaturated ketone at C3 in the steroid
A-ring. Cobice et al. used collision induced dissociation (CID)/liquid
extraction surface analysis to confirm the structure of the generated compounds.
CID of the GirT derivatives produced a series of fragment ions characterized by
the loss of the quaternary amine tag [M – 59]+ and
carbon monoxide [M – 87]+ of the derivatized group.
This method was applied to image 11-dehydrocorticosterone and corticosterone
distribution in the mouse adrenal gland and brain. MSI analysis of the mouse
brain with glucocorticoid-amplifying enzyme 11β-HSD1 deficiency or
inhibition revealed sub-regional corticosterone/11-dehydrocorticosterone
ratio changes, and accumulation of 7-ketocholesterol as the alternative
11β-HSD1 substrate. This MSI data was validated by liquid chromatography
(LC)-MS/MS in whole brain homogenates with high correlation.
Recently, GirT derivatization followed by tandem MSI (MS/MS:
MS2 and MS/MS/MS: MS3) successfully
differentiated aldosterone from its isomer cortisone, which shows a different
distribution pattern. MS3 was necessary for generating characteristic
fragments to distinguish aldosterone from cortisone, as shown in [Fig. 2], [52]. As the first studies to visualize the localization of
aldosterone on a frozen rat adrenal section, MSI analysis confirmed its
production in the ZG layer following a sodium-deficient diet [49]. GirT derivatization has also enabled
the visualization of other steroids of GirT-corticosterone,
GirT-18-hydroxycorticosterone and GirT-progesterone [49]. This method has been applied to the
study of aldosterone-producing cell clusters (APCCs), which occur in more than
80% of adult adrenals (>30 years old), and their relation to
APAs as will be discussed later [49]
[54]. Subsequently, this derivatization
and MS3-based method was extended to separate not only aldosterone
but also cortisol from their isomers, cortisone and 18-oxocorticosterone,
respectively. Cortisol distribution was visualized for the first time in the
human adrenal gland [52].
Fig. 2 Product ion spectra of GirT-aldosterone and GirT-cortisone
(GirT-E) standards. MS2 spectra of m/z 474 and
presumed structures show similar fragmentation patterns for a
GirT-aldosterone and b GirT-cortisone. MS3 spectra of
m/z 415 show characteristic fragmentation patterns for
c GirT-aldosterone and d GirT-cortisone. (Reprinted
from [52] with permission from
American Chemical Society) [rerif].
For the detection of steroids, publications are mainly focused on chemical
derivatization using FF tissue samples which could increase ionization
efficiency of steroids. Nevertheless, previous studies have demonstrated that
certain steroids can be detected in FFPE tissue without specific chemical
derivatization [21]
[48].
Molecular imaging of adrenal gland physiology by MSI (FF tissue)
The normal functional anatomy of the adrenal gland has been studied using MSI
techniques to look into their metabolites and hormone profile in situ.
Localization of cholesterol and corticosterone with a sub-cellular spatial
resolution was demonstrated in the rat adrenal cortex using imaging
time-of-flight secondary ion mass spectrometry (TOF-SIMS) [55]. Wu et al. applied desorption
electrospray ionization (DESI) imaging (positive ion mode, raster size 200
μm) to porcine and rabbit adrenal gland tissue and determined that the
distribution of the small hormones epinephrine and norepinephrine enriched in
the medulla layer [56]. Using
MALDI-FTICR-MSI in both the positive and negative ion modes (raster size 200
μm), Wang et al. determined the spatial distribution pattern of 544
lipids and 11 non-lipid metabolites in porcine adrenal glands, and showed that
many classes of lipids had distinct distribution patterns in different
functional zones of the adrenal gland [57]. Another study has increased the molecular specificity and
sensitivity in the rabbit adrenal gland using sodium-doped gold-assisted
LDI-enhanced MSI (high lateral resolution of 25 μm) to show
distributions of cholesterol, different cholesterol esters (CE) and
triacylglycerols (TAGs) [58].
Normal human adrenal gland molecular imaging has been performed by Sun et al.
using a comprehensive MADLI-FTICR-MSI [20]. As mentioned before, a broad spectrum of metabolites and hormones,
including epinephrine, norepinephrine, dehydroepiandrosterone sulfate,
pregnenolone sulfate, cholesterol sulfate, AMP, ADP, ATP, UDP-glucose,
UDP-N-acetylglucosamine, glucose monophosphate, citrate, palmitic
acid, adrenic acid, and phosphatidylinositol, can be measured via this method.
Functional adrenal zonation assessed by IHC demonstrates an exact
co-localization of sulfated steroid hormone with zone specific IHC markers in
the human adrenal gland. A clear separation of the cortex and medulla was
observed at the level of individual metabolites by heat map-based clustering
analysis of discriminative metabolites. At the comprehensive metabolic pathway
level, pathway enrichment analysis revealed enhanced steroid hormone and fatty
acid synthesis in the adrenal cortex, which is in line with the known steroid
hormone production in the cortex. Elevated activity of the pentose phosphate
pathway and glycolysis and nucleotide metabolism in the medulla was found,
reflecting the high energy consumption in the medulla. A spatial segmentation
map based on unsupervised hierarchical clustering generated not only a clear
separation between the cortex and medulla ([Fig. 3], level 1) but also a virtually molecular-based map that
matched perfectly with the histological morphological features of adrenal gland,
including the zona glomerulosa (ZG), zona fasciculata (ZF), zona reticularis
(ZR), and medulla ([Fig. 3], level 2).
Further improved spatial segmentation created 10 individual molecular-defined
zones corresponding to functionally distinct parts that are not completely
reflected by the histological layers of the adrenal gland ([Fig. 3], level 3) [20]. Among these identified zones,
molecular zones 1–6 represented the cortex area and molecular zones
7–10 corresponded to the medulla area. The relevance of molecular-based
zonation, although not well-elucidated, is likely to reflect functional
adaptation to the interplay between the adrenal medulla and cortex.
Fig. 3 Unsupervised hierarchical clustering reveals the molecular
structures of the adrenal gland. These structures match well the
histology of the adrenal gland in level 1 (two classes separate the
cortex and medulla) and level 2 [four classes separate the zona
glomerulosa (ZG), zona fasciculata (ZF), zona reticularis (ZR), and
medulla], and identified additional distinct molecular layers in level 3
(10 molecular classes). Scale bars: 1 mm. ZG: Zona glomerulosa; ZF: Zona
fasciculate; ZR: Zona reticularis; H & E: Hematoxylin and eosin.
(Reprinted from [20] by permission
of Oxford University Press) [rerif].
Molecular Imaging of whole-body animal tissue section (FF tissue)
MSI has also been used for whole-body sections from different model animals such
as rats, mice [59]
[60]. Molecular structures enriched in the
adrenal gland could be accessible via whole-body molecular imaging. As an
example, MALDI-MSI of mice has identified one phospholipid
(m/z 810.60) enriched in the adrenal gland [61]. Whole-body MSI simultaneously detects
the presence of molecules in different organs, thus providing a way for studying
the relation of adrenal gland with other organs. Another study observed the
presence of the antitumor candidate drug
(S)-(+)-deoxytylophorinidine (CAT) after drug dosing in a
whole-body molecular image via air flow-assisted ionization in the DESI mode
(AFADESI)-imaging mass spectrometry (IMS) [62]. High drug concentration in the adrenal gland was found 20 min
after dosing and was steadily conserved 2 hours after dosing. The visualization
of exogenous drug disposition in multiple organs in the whole-body section
through MSI allows the direct discovery of the organ targeted by drug candidate.
The prediction of pharmacological activity and potential toxicity could be one
potential application of such method.
Clinicopathologic application of MSI in the study of adrenal tumors
The dysfunction of the adrenal gland could present as an excess or insufficiency
of adrenal hormones in conjunction with relevant symptoms. For example,
overproduction of glucocorticoids causes Cushing’s syndrome, whereas
excessive aldosterone results in hyperaldosteronism [63]
[64]
[65]. The most common benign adrenal tumor
is adrenocortical adenoma, most of which are aldosterone-producing or
cortisol-producing adrenocortical adenoma. Adrenocortical carcinoma (ACC) is a
rare endocrine tumor with great clinical significance due to its high malignancy
potential [66]. Pheochromocytomas are
tumors of chromaffin cells, which most commonly arise in the adrenal medulla and
are usually not metastatic [67]
[68]. In all cases, adrenal diseases and
tumors have the potential to impair quality of life and may be
life-threatening.
In situ metabolomics of aldosterone-producing adenomas (APAs) by MSI and its
clinical implications (FFPE tissue)
MALDI-FTICR-MSI was applied to a cohort of 136 FFPE tissue cores of patients with
a unilateral primary aldosteronism, which allowed for a comprehensive metabolite
profile analysis and association with genotype/phenotype, IHC of
steroidogenic enzymes and clinical features/outcomes [48]. This study identified 137
significantly different metabolites between KCNJ5- and CACNA1D-mutated APAs,
including metabolites associated with steroidogenesis and purine metabolism. An
increase in purine synthesis in APAs with KCNJ5 mutations could reflect the
promotion of cell proliferation, which is also connected to the larger tumor
size observed in KCNJ5-mutated APAs ([Fig.
4]). High levels of 18-oxocortisol and 18-hydroxylcortisol were found
in KCNJ5-mutated APAs ([Fig. 4]).
Accordingly, high plasma levels of 18-oxocortisol and 18-hydroxylcortisol were
identified in patients with KCNJ5-mutated APA and were shown to be predictive of
KCNJ5 mutation. Distinct metabolome profiles between these two gene mutations
were further supported by orthogonal partial least squares discriminant analysis
(ortho-PLSDA) clustering ([Fig. 4]).
Regarding steroids and steroidogenesis enzymes, 18-oxocortisol and
18-hydroxylcortisol were inversely correlated with CYP11B1 but not CYP11B2 in
APA tissues, which might be explained by an overall high expression of
CYP11B2above a threshold critical for detecting 18 steroid level differences.
CYP11B2 has been shown to convert 11-deoxycortisol to 18-oxocortisol and
18-hydroxylcortisol, whereas CYP11B1 only produces 18-hydroxylcortisol [69]. Moreover, both 18-oxocortisol
intensity and CYP11B1 expression levels in tumor tissue are predictive of the
clinical outcome, independent of other clinical parameters [48]. For example, complete clinical success
after surgery was linked to high intensity of 18-oxocortisol or low staining of
CYP11B1, demonstrating the importance of CYP11B1 over CYP11B2 in tumor tissue
with regard to the clinical outcome. This association between 18 steroid levels,
CYP11B1 and clinical outcome also may be related to the glucocorticoid
co-secretion observed in APAs, and also may be related to
“Connshing” syndrome [70]
[71].
Fig. 4
a The results of orthogonal partial least squares discriminant
analysis (ortho-PLSDA), which identified separate clusters for KCNJ5-
and CACNA1D-mutated aldosterone-producing adenomas (APAs) based on
metabolome profiles. b Significant difference in metabolites
related to purine metabolism were found between KCNJ5- and
CACNA1D-mutated APAs (adjusted p=0.010, 0.003, and 0.047,
respectively). c Significant changes in the intensities of
18-oxocoritsol (p=0.020) and 18-hydroxycortisol (p
<0.001) were identified between KCNJ5- and CACNA1D-mutated APAs.
Mann–Whitney U-tests were used for statistical analysis.
* p <0.05; ** p <0.01.
(Reprinted from [48] with
permission from American Society for Clinical Investigation)
[rerif].
MSI imaging in the adrenal cortex suggests pathophysiological progression
from aldosterone-producing cell clusters (APCCs) to APA (FF and FFPE
tissue)
APCCs can be identified by IHC staining of strong aldosterone synthase (CYP11B2)
expression without expression of steroid 11β-hydroxylase
(cortisol-synthesizing enzyme, CYP11B1), even in the non-primary aldosteronism
(AP) producing adult cortex ([Fig. 5a],
cases 1–4). Thus, APCCs seem to produce aldosterone autonomously. The
accumulation of aldosterone and 18-oxocortisol in CYP11B2-expressive lesions
(APCC) was confirmed by Sugiyama et al. by visualizing GirT derivatized
aldosterone on FF adrenal samples [49].
The distribution patterns of 18-oxocortisol and aldosterone in APCC and APA
tissues ([Fig. 5b]) have led to a
progressive development hypothesis of APA generation in which APCCs develop into
APAs via APCC-to-APA transitional lesions (pAATLs: [Fig. 5b], case 5). These pAATLs can be
divided into a sub-capsular APCC-like region and an inner APA-like region. The
APCC-like region of a pAATL consists of aldosterone-producing cells like an
APCC, whereas the APA-like region of a pAATL contains both aldosterone- and
cortisol-producing cells like a small size APA. In one case with a large
CYP11B2-positive APA ([Fig. 5b], case 8),
both an aldosterone-producing area and a nonfunctional area without aldosterone
were identified via imaging; the latter nonfunctional area may have been due to
a lack of precursor steroids, including progesterone. Both plasma 18-oxocortisol
and aldosterone levels are increased in patients with APAs. This study found
accumulation of 18-oxocortisol in the CYP11B2-positive APCC regions, as well as
in the APA regions, with either heterogeneous or relatively homogenous
distributions ([Fig. 5b]) [49]. By comparing the steroid distributions
imaged via MSI analysis ([Fig. 5b]) with
the distributions of steroidogenesis enzymes identified via IHC staining ([Fig. 5a]), this study generated the
hypothesis that 18-oxocortisol in the APCC region results from 18-oxidization of
cortisol by CYP11B2, and that the cortisol diffused to the APCC region from the
surrounding ZF. The hypothesis that cortisol moves from CYPB11-active region to
APCCs and that cortisol is further oxidized by CYP11B2 into 18-oxocortisonl and
18-hydroxycortisol were further supported by visualizing cortisol in the human
adrenal gland as mentioned before [52].
Cortisol were found to be absent in all APCCs where aldosterone was
accumulated.
Fig. 5
a Immunohistochemistry (IHC) for CYP11B2 in human
adrenalectomized samples. Cases 1–4 (blue background), case 5
(orange background) and cases 6–8 (green background) show IHC
results for aldosterone-producing cell clusters (APCCs; arrowheads),
possible APCC-to-aldosterone-producing adenoma (APA) transitional
lesions (pAATLs), and APAs, respectively. In cases 3–8; T: Tumor
area within red dotted lines, NT: Non-tumor area outside of the red
dotted line. Bars represent 1 mm. b Matrix-assisted laser
desorption/ionization (MALDI) imaging of APCCs (cases 1–4, blue
background), pAATLs (case 5, orange background), and APAs (cases
5–8, green background) using Fourier-transform ion cyclotron
resonance mass spectrometry (FT-ICR-MS). GirT-18-oxoF shows
the distribution of derivatized 18-oxocortisol.
GirT-aldo/cortisone demonstrates the distribution of
derivatized aldosterone and cortisone. In cases 1–4, white
arrowheads correspond to black arrowheads indicating APCCs in [Fig. 5a]. In cases 3–8, T
(tumor region within orange dotted lines) and NT (non-tumor region
outside of the orange dotted line) correspond to CYP11B2 staining in
[Fig. 5a]. Bars represent 1
mm. (Reprinted from [49] with
permission from Wolters Kluwer Health, Inc.) [rerif].
The most recent publication of Sun et al. used used in situ metabolic imaging
analysis by MALDI-FT-ICR-MSI of FFPE adrenals resected from patients with
unilateral primary aldosteronism to investigate the potential existence of
diverse metabolic phenotypes of APCCs [72]. The analysis covered a wide range of central metabolic and steroid
hormone biosynthesis pathways and determined heterogeneous patterns of
metabolites associated with APCCs.
Specific distribution patterns of metabolites were associated with APCCs and
classified 2 separate APCC subgroups (subgroups 1 and 2) indistinguishable by
CYP11B2 immunohistochemistry. Metabolic profiles of APCCs in subgroup 1 were
tightly clustered and distinct from subgroup 2 and APAs. Multiple APCCs from the
same adrenal displayed metabolic profiles of the same subgroup ([Fig. 6]). Metabolites of APCC subgroup 2
were highly similar to the APA group and indicated enhanced metabolic pathways
favoring cell proliferation compared with APCC subgroup 1. In conclusion, the
results demonstrate specific subgroups of APCCs with strikingly divergent
distribution patterns of metabolites. One subgroup displays a metabolic
phenotype convergent with APAs and may represent the progression of APCCs to
APAs.
Fig. 6 Hierarchical clustering and component analysis
demonstrated different metabolic profiles in APCC subgroup 1, APCC
subgroup 2 and APAs. a Peak lists with respective intensities
were uploaded to MetaboAnalyst. Each colored cell corresponds to an
intensity value, with samples in rows and features in columns. Euclidean
distance and Ward’s method were applied for cluster analysis.
b Component analysis using sPLSDA identified 3 patterns of
metabolites comprising 2 subgroups of clearly separated APCCs (subgroup
1 and subgroup 2) and the APA group. c CYP11B2 (aldosterone
synthase) immunohistochemistry of adrenal samples included in the
metabolic analyses showing APCCs in subgroup 1, APCCs in subgroup 2 and
APAs. ROI (region of interest) identification numbers are shown for each
APCC analyzed. (Reprinted from [72] with permission from Wolters Kluwer Health, Inc.)
[rerif].
Prognostic relevance of steroid sulfation detected by MSI in adrenocortical
carcinoma (ACC) (FFPE tissue)
Molecules measured by MSI could be used in conjunction with genomic and RNA
expression data to determine the molecular phenotype of tumors. As shown in
[Fig. 7a], MALDI-FTICR-MSI analysis
identified the molecular phenotype of 72 ACCs based on steroid hormone
metabolites [21]. In this study, high
levels of estradiol sulfate (E2S) and estrone-3-sulfate (E1S) were significantly
associated with a good ACC prognosis via both univariate Kaplan-Meier analysis
and after multivariate adjustment for tumor stage, age, and sex ([Fig. 7c] and d). This finding
supports the idea that active steroid sulfation is a marker of a less aggressive
ACC phenotype. Moreover, the expression of sulfotransferase SULT2A1 in the same
tissue samples had similar prognostic power. This finding was validated in two
published data sets, and suggests that SULT2A1 is the enzyme responsible for the
synthesis of E1S and E2S. Beyond the known steroid sulfates, disulfated steroid
estradiol-17β-3,17-disulfate (E2S2) also was found in six samples, first
by searching for the calculated theoretical mass in the measured spectrum and
then validating by MS/MS. Its presence in these six patients was
associated with a particularly poor overall survival ([Fig. 7e]). Multiple membrane-delimited
vacuoles were found only in these tumors with E2S2, supporting the intracellular
accumulation of E2S2 ([Fig. 7b]). This
study thus highlights the potential of steroid hormone sulfation to predict a
patient’s prognosis and to stratify patients by treatment.
Fig. 7
a Hierarchical cluster analysis of molecular m/z
separates adrenocortical carcinoma (ACC) patient cohorts into three
phenotypes, namely phenotype 1, phenotype 2, and phenotype 3. b
Electron microscopy of phenotype 1 tissue (left panel) and phenotype 3
tissue (center panel), and vacuoles in phenotype 3 (highlighted in blue
in the center panel) are enlarged as regions of interest (ROI, right
panel). c,d,e Survival probabilities based on
Kaplan–Meier analyses for patients with high (red) and low
(blue) estrone-3-sulfate (E1S) and estradiol-17β-3-sulfate (E2S)
levels, and patients with (red) or without (blue) disulfated steroid
estradiol-17β-3,17-disulfate (E2S2). (Reprinted from [21] by permission of Oxford
University Press) [rerif].
MSI of catecholamines in the adrenal medulla (FF tissue)
Although catecholamines in the medulla region are accessible via different MSI
techniques such as DESI-MSI or high-resolution MALDI-FTICR-MSI, as previously
mentioned [20]
[56], a few studies have also applied
derivatization for catecholamines visualization. Manier et al. [50] improved the sensitivity of the medulla
origin hormones dopamine, epinephrine and norepinephrine by pre-coating the
target with 4-hydroxy-3-methoxycinnamaldehyde (CA) as a derivatization reagent
in a MALDI ion trap MSI analysis. The signals for these three target compounds
from underivatized analyses were very low with a non-specific distribution in
their analysis, indicating insufficient sensitivity for distinguishing target
compounds from the chemical background or matrix interference. In contrast,
derivatization followed by MALDI-MSI with MSn analysis resulted in
substantially improved sensitivity, allowing a clear visualization of the
respective hormones in situ in the medulla region. Particularly, MS3
was necessary to produce the specific fragments to enable visualization of
derivatized norepinephrine. A more accurate localization of dopamine,
epinephrine and norepinephrine in the modular region also was achieved in
another study [51]. By applying
4-(N-methyl)pyridiniumboronic acid
[4-(NMe)PyB(OH)2] derivatization, both LDI-TOF-MSI and gas
cluster ion beam (GCIB)-TOF-SIMS analysis yielded a similar characteristic
spatial distribution of epinephrine, norepinephrine and dopamine in porcine
adrenal gland tissue sections; these compounds localize in the peripheral,
central and entire medulla region, respectively, as shown in [Fig. 8], [51]. In another study, the derivatization reagent
p-N,N,N-trimethylammonioanilyl
N′-hydroxysuccinimidyl carbamate iodide (TAHS) was applied to
successfully discriminate enriched catecholamine species in the pheochromocytoma
region [52].
Fig. 8 The distribution of 4-(N-methyl)pyridinium boronic
acid [4-(N-Me)PyB(OH)2] derivatized catecholamines in
the porcine adrenal gland was analyzed by GCIB-ToF-SIMS (30 µm
per pixel). a Hematoxylin and eosin (H & E) staining of
adrenal cortex and medulla. b An unidentified peak with
m/z 753.5. c
Norepinephrine-4-(N-Me)PyB(OH)2 with
m/z 271.1. d
Dopamine-4-(N-Me)PyB(OH)2 with m/z
255.1. e Epinephrine-4-(N-Me)PyB(OH)2 with
m/z 285.1. f Fragment ion of
epinephrine/norepinephrine-4-(N-Me)PyB(OH)2.
g and h Merged ion images of
norepinephrine-4-(N-Me)PyB(OH)2 (red),
epinephrine-4-(N-Me)PyB(OH)2 (green),
dopamine-4-(N-Me)PyB(OH)2 (blue, g) or
unidentified peak with m/z 753.5 (blue, h).
(Reprinted from [51] with
permission from American Chemical Society) [rerif].
MSI for utilizing the pheochromocytoma (PC12) cell model (FF tissue)
A few studies have applied MSI techniques, mainly secondary ion (SI) MSI, for
measuring the pheochromocytoma (PC12) cell model, which is widely used as a
model system for studying exocytosis [73].
PC12 cells release catecholamines by exocytosis and differentiate into
neuron-like cells upon treatment with nerve growth factor [74]. TOF-SIMS was used to image lipids
fragments in freeze fractured rat pheochromocytoma (PC12) cells on a single-cell
level [73]
[74]
[75]. Dopamine localization was identified
in the vesicular compartments of pheochromocytoma (PC12) cells by correlating
TEM and NanoSIMS images of single vesicles [76]. In a drug-related study, a decrease of lipids in PC12 cells
treated with cisplatin was observed, suggesting that cisplatin affects
exocytotic release by alternating cell membrane lipids [77].
