Visualization of relative 11β-hydroxylase (CYP11B1) activity in human cortisol-producing adrenocortical adenoma (CPA) using mass spectrometry imaging-based enzyme histochemistry

Abstract

Enzymes play essential roles in numerous biochemical reactions, making it necessary to evaluate both their localization and activity. While conventional enzyme activity detection methods require specific antibodies and colorimetric reactions for each target compound, a recent method using mass spectrometry imaging enables the visualization of enzyme distribution and activity by applying substrates directly to tissue sections and detecting the resulting enzyme reaction products generated on these sections. Notably, the 11β-hydroxylase (CYP11B1, EC 1.14.15.4) reaction plays a key role in stress response, metabolism, and immune function regulation. Consequently, the regulation of CYP11B1 activity can significantly impact these physiological processes, enabling improved drug development. Therefore, controlling this enzyme activity will facilitate the development of new therapies for related diseases. This study investigated a method for visualizing the distribution and relative activity of CYP11B1 in cortisol-producing adrenocortical adenoma (CPA) tissue sections derived from patients with mild autonomous cortisol secretion (MACS). By optimizing key parameters such as the substrate and coenzyme concentrations, together with reaction times, the method successfully revealed the relative CYP11B1 activity in human MACS. Additionally, visualization was achieved in aldosterone-producing adrenocortical adenoma (APA) derived from patients with primary aldosteronism and in non-functioning adrenal adenoma (NFA) derived from patients with adrenal cortical adenoma. The relative CYP11B1 activities of CPA, APA, and NFA samples obtained in this study correlated with the clinical information and were consistent with the previously reported cortisol, adrenocorticotropic hormone, and dexamethasone suppression test profiles in blood and urine, thereby validating the proposed method.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s00216-026-06518-x.

Introduction

Enzymes play important roles in a wide range of biochemical processes. To better understand these processes, it is essential to evaluate both the distribution and activity of these enzymes within biological systems. Traditional approaches, including immunohistochemistry (IHC) methods and in situ hybridization, have been used to investigate enzyme localization [1, 2]. Additionally, transcriptome and genetic analyses have been employed to evaluate enzyme protein and mRNA expression levels and to identify genetic mutations [3, 4]. However, these methods do not provide information on the localization of enzyme activity.

Enzyme histochemistry is a technique used to visualize enzyme activity in tissues and cells [5, 6]. This technique has been used to localize 17β-hydroxysteroid dehydrogenase (17β-HSDs, EC 1.1.1.51), 3β-HSD (EC 1.1.1.145), and aromatase (EC 1.14.14.14) activities in mouse tests [7] and NADH-methemoglobin reductase (EC 1.6.5.2) activity in breast cancer [8]. Although enzyme distribution does not necessarily correlate with the site of enzyme activity, the localization of enzyme activity is strongly correlated with the in vivo metabolic mapping of cells and tissues within the same organism [5]. Therefore, the in vivo analysis of enzyme activity serves as a powerful tool for understanding biological phenomena.

Matrix-assisted laser desorption/ionization (MALDI) mass spectrometry imaging (MSI) [9] enables the visualization of molecular localization in tissue sections. In enzyme histochemistry using MALDI-MSI, labeled substrates are applied to tissue sections to distinguish between endogenous substrates and products. Enzymatic reactions are then carried out under optimal conditions, and the resulting labeled products are examined to generate images reflecting the enzyme activity distribution in tissues. This technique has been used to study the distribution of cholinesterase enzymatic activity in the mouse brain [10, 11] and that of glutamine decarboxylase activity in soybeans [12]. Additionally, studies have detected multiple enzyme activities simultaneously in mouse kidneys [13].

The mitochondrial enzyme 11β-hydroxylase (CYP11B1, EC 1.14.15.4) is localized in the adrenocortical bundle junction and brain. It primarily functions as a convertase, catalyzing the conversion of 11β-deoxycortisol (11-DOC) to cortisol in humans and 11-deoxycorticosterone to corticosterone in mice and rats. This reaction is involved in stress response [14], metabolism [15], and immune function regulation [16, 17]. Regulating CYP11B1 activity can significantly impact these physiological processes, making it an attractive target for drug development. For example, suppression of cortisol levels in cases of adrenocorticotropic hormone (ACTH) overproduction and Cushing’s syndrome (CS, a condition characterized by prolonged exposure to inappropriately high levels of cortisol) helps alleviate associated symptoms such as obesity, hypertension, diabetes, and osteoporosis [18, 19]. This enzyme also aids in restoring hormonal balance in conditions involving aldosterone overproduction and cortisol deficiency, such as congenital adrenal hyperplasia [20]. Additionally, it may affect the tumor microenvironment and immune response [21]. However, urinary cortisol levels are not specific enough to indicate the disease, and blood cortisol levels fluctuate according to the diurnal cycle, making them unsuitable as an indicator of CYP11B1 activity. Conversely, suppression of serum ACTH levels indicates active negative feedback due to excessive cortisol secretion (ACTH-independent somatic gene mutation), whereas elevated ACTH levels indicate that the negative feedback is not functioning (ACTH-dependent), resulting in excessive cortisol secretion [22]. Furthermore, while most ACTH-independent mild autonomous cortisol secretion (MACS, a condition characterized by excessive cortisol production by the adrenal glands, but with no physical symptoms or only mild ones) cases are known to harbor somatic mutations affecting the cortisol synthesis pathway, MACS has been reported in cases without somatic mutations owing to CYP11B1 overexpression, increased activity within the cortisol-producing adrenocortical adenoma (CPA, a condition in which the body produces and secretes excessive amounts of cortisol without regulation by ACTH from the pituitary gland), or abnormal localization/expression of other steroid synthases [23–26]. Conversely, aldosterone synthase (CYP11B2, EC 1.14.15.5) expression is often reduced [27–29], suggesting that cortisol production predominates. Furthermore, inadequate suppression of cortisol levels following a dexamethasone suppression test (DST) suggests the occurrence of autonomous cortisol secretion (ACS). In addition, the cortisol levels of patients with primary aldosteronism (PA, a condition in which aldosterone continues to be secreted in excess despite low levels of renin, which regulates blood pressure) are within or slightly outside the normal range. However, some patients exhibit a condition accompanied by ACS [30]. This condition, in which PA coexists with CS (particularly MACS), is known as Conn’s syndrome [31], and cortisol levels may be normal or slightly elevated in such cases [32]. Furthermore, findings indicative of ACS have been reported after DST, including decreased blood ACTH levels and increased blood cortisol levels. As mentioned earlier, MACS diagnosis involves several tests, including somatic mutation testing, cortisol level measurement in blood and urine, imaging, and endocrinological testing [33]. Treatment decisions are made based on test results. Therefore, MACS pathogenesis is complex, and both molecular and functional data need to be considered when evaluating enzyme activity in clinical settings.

Previous studies have confirmed that CYP11B1 is expressed in non-functioning adrenal adenoma (NFA, a condition in which a benign adenoma develops in the adrenal cortex, but it does not produce excessive amounts of hormones associated with conditions such as CS or PA. It also does not result in any clinical endocrine symptoms), aldosterone-producing adrenocortical adenoma (APA, a condition characterized by autonomous, i.e., independent of the renin-angiotensin system, overproduction of aldosterone), and CPA [34]. CYP11B1 is primarily localized in the zona fasciculata of NFA and is known to convert into either CPA or APA owing to adenoma growth or chronic instability of hormone secretion [35]. These changes may alter the expression and localization pattern of CYP11B1. However, the differences in CYP11B1 localization among these adenomas and the implications of these differences for their pathology have not yet been fully elucidated.

In this study, we developed a method for visualizing the localization of the relative CYP11B1 activity in CPA, APA, and NFA using enzyme histochemistry combined with MSI.

To visualize the relative CYP11B1 activity, several prerequisites were established. By establishing these preconditions, we clarify the limitations of optimizing the CYP11B1 enzyme reaction and interpreting the relative CYP11B1 activity results, thereby ensuring this method’s validity. First, it is assumed that the substrate is in excess of the enzyme on the tissue section. In this case, the rate of the enzyme reaction does not depend on the concentration of the substrate and can be interpreted as being dependent on the localized enzyme activity. Additionally, measuring both the substrate and product avoided enzyme saturation. Second, it is assumed that cortisol-d5 is produced exclusively through the CYP11B1 activity. However, because other steroid-metabolizing enzymes and pathways exist within tissues, not all cortisol-d5 may be generated from CYP11B1 reactions. To minimize this possibility, substrate selection and reaction time optimization were performed. The third assumption is that the variability in the extraction and ionization efficiencies at each measurement position—challenging for MALDI-MSI—can be corrected through normalization. However, because the peaks obtained through this correction cannot account for the differences in the compound-specific ionization efficiency or suppression effects within the sample, its application is limited to relative spatial comparisons. Fourth, it is assumed that cortisol-d5 will remain localized from the start of the enzymatic reaction until it is measured. The effects of diffusion were minimized in this study by reducing the time taken for the enzymatic reaction and sample preparation. This method is based on certain assumptions; therefore, the CYP11B1 activity distribution obtained using it should be interpreted as a relative measure. Furthermore, comparative analyses conducted under unified conditions reveal it to be an effective method for evaluating the relative distribution of the enzyme activity within tissues.

“Relative CYP11B1 activity” in MSI was calculated as the ratio of the amount of product to that of the substrate. Initially, CYP11B1-mediated substrate conversion was confirmed. Substrate, cofactor, and incubation time parameters were then optimized. Finally, the localization of the relative CYP11B1 activity in human CPA sections was visualized using MSI. This same procedure was applied to human NFA, APA, and CPA sections, both with and without somatic mutations. The results were evaluated using IHC. Additionally, we compared the relative CYP11B1 activity obtained using MSI in the human NFA, APA, and CPA tissues. We confirmed that the observed trend was consistent with previously reported cortisol concentration profiles.

Experimental

Chemicals and reagents

Frozen human adrenal gland blocks were provided by Kyushu University (approval number: 21025-03). 11-Deoxycortisol-2,2,4,6,6-d5 (11DOC-d5), 11-deoxycortisol (11DOC), cortisol, β-nicotinamide adenine dinucleotide 2′-phosphate–reduced tetrasodium salt hydrate (NADPH), Girard T reagent (GirT), 2,5-dihydroxybenzoic acid (DHB), α-cyano-4-hydroxycinnamic acid (CHCA, 98%), PathoMount, and an anti-CYP11B1 antibody (clone 80-7) were purchased from Merk (Darmstadt, Germany). Disodium hydrogen phosphate 12-water, sodium dihydrogen phosphate, acetonitrile, acetic acid, methanol, trifluoroacetic acid (TFA), Mayer’s hematoxylin solution, ethanol, xylene, and 1% eosin Y solution were purchased from FUJIFILM Wako Pure Chemical (Osaka, Japan). A high-pH antibody–antigen activation solution (50×) was purchased from Agilent (California, USA). Rabbit Anti-Rat IgG Antibody (H + L) Biotinylated and Universal Elite ABC Standard Kit were purchased from Vector Laboratories (California, USA). Ultrapure water was obtained using an ultrapure water purification system (GenPure UV-TOC xCAD Plus, Thermo Fisher Scientific, Massachusetts, USA).

Clinical samples

The clinical samples were provided by Kyushu University. All tissue samples were frozen and stored in a deep freezer immediately after their removal. NFA is an adrenal tumor that does not secrete excess hormones. PA and MACS tissues were subjected to target DNA sequencing using xGen Predesigned Gene Capture Pools (Integrated DNA Technologies, Iowa, USA). The obtained data were analyzed using the Genomon2 pipeline (Human Genome Center, Institute of Medical Science, The University of Tokyo, Tokyo, Japan) to identify mutations. ACS is a condition in which the adrenal glands produce excess cortisol independently of signals from the brain, i.e., cortisol production is ACTH-independent. ACS is the primary cause of CS. MACS is a subset of ACS. Although it does not present with overt symptoms such as CS, it is characterized by an inability to suppress cortisol with 1 mg DST. It is also associated with an increased risk of metabolic syndrome (e.g., dyslipidemia) and neuropsychiatric disorders (e.g., depression and anxiety) compared to NFA. Even when symptoms are mild, it poses health risks, which make it clinically significant and drive efforts to establish diagnostic methods.

Targeted DNA sequencing

Unstained, formalin-fixed, paraffin-embedded (FFPE) MACS samples were sectioned using 17α-hydroxylase (CYP17A1, EC 1.14.14.37) (IHC) slides as guides, and nucleic acids were isolated. DNA isolation, target capture sequencing, and somatic mutation detection were performed using previously reported methods [36]. A total of 36 genes, including MACS-related genes, were selected from a custom bait library.

Tissue preparation

An optimal cutting temperature compound was used to fix each tissue block onto the holder of a cryomicrotome (CM 1950, Leica Microsystems, Nussloch, Germany). All tissue sections were cut to a thickness of 8 µm at − 20 °C using a cryomicrotome. The sections were thaw-mounted onto indium tin oxide–coated glass slides (SI0100N, Matsunami Glass, Osaka, Japan) for MALDI-MSI and allowed to dehydrate in a 50-mL conical tube containing silica gel. The slides were stored at − 80 °C in tubes until analysis.

Tissue enzyme reaction

The substrate solution for CYP11B1 was prepared by mixing final concentrations of 0.05–1 mM 11DOC-d5 and 0.05–10 mM NADPH with ultrapure water or 0.05 M phosphate buffer (pH 7.0). The CYP11B1 substrate solution (20 µL per section) was applied evenly throughout the section using an airbrush (GSI Creos, Tokyo, Japan). After 3 min of treatment with the substrate solution, the tissues were incubated at 37 °C for 0–60 min. Here, 0 min incubation was defined as no incubation at 37 °C. Incubation was carried out in a petri dish lined with ProWipe and impregnated with 2 mL ultrapure water in the control experiments.

Enzyme inactivation

To confirm that the observed products were generated by the endogenous enzymatic activity, the tissue sections underwent an enzyme inactivation treatment. Specifically, the sections were incubated at 100 °C for 30 min to denature the enzymes and stop them from converting the substrate.

Derivatization

Derivatization is essential for the detection of 11DOC-d5 and cortisol-d5 using MALDI-MSI [37]. A total of 10 mg GirT was dissolved in 1 mL acetic acid:ultrapure water (2:8, v/v) solution by vortexing, and then the mixture was treated in an ultrasonic bath for 10 min to ensure complete dissolution. The supernatant obtained after centrifugation was used as the derivatization reagent. As the GirT reagent exists in an acidic environment with a pH of around 2.0, it is considered that the activity of the target enzyme is virtually halted once the reagent has been applied to the tissue sections. The GirT reagent was applied manually using an airbrush with a 0.2-mm nozzle. The room in which manual spraying was carried out was maintained at an approximate temperature of 25 °C and a humidity of 35%–60%. The nozzle was held 5 cm from the tissue surface, and the reagent was sprayed for 5 s, followed by a 20-s drying period. This process was repeated ten times, with 50 µL GirT reagent delivered to each section at a spray rate of approximately 60 µL/min. After spraying, the derivatization reaction was allowed to proceed at room temperature (approximately 25 °C) for 60 min to complete the reaction on the tissue.

Matrix application

After the derivatization reaction was complete, the MALDI matrix was immediately applied to the tissue. In this experiment, CHCA was used as the MALDI matrix and sprayed using an airbrush with a 0.2-mm nozzle. A 10 mg/mL solution of CHCA was prepared with acetonitrile:ultrapure water (1:1, v/v) as the solvent, and TFA was added to achieve a final concentration of 0.1%. The CHCA solution was mixed using a vortex mixer and then sonicated for 10 min. The resulting supernatant was sprayed onto a single human adrenal gland section using an airbrush. The room in which manual spraying was carried out was maintained at an approximate temperature of 25 °C and a humidity of 35%–60%. The nozzle was held 5 cm away from the tissue surface. The first cycle consisted of 5 s of spraying, followed by 120 s of drying. The second cycle consisted of 5 s of spraying, followed by 40 s of drying. From the third cycle onward, spraying continued for 5 s, followed by 20 s of drying. This cycle (5 s of spraying, followed by 20 s of drying) was repeated until 150 µL CHCA matrix was sprayed per section at a spray rate of approximately 60 µL/min. After matrix application, the samples were allowed to stand at room temperature (approximately 25 °C) for 5 min before measurement. To prevent contamination during the matrix-spraying process, separate airbrushes were used for spraying the substrate solution, derivatization reagent, and matrix solution. The uniformity of the manual spraying process was evaluated by controlling the working environment and examining the crystal state on the slide surface with an optical microscope. Reproducibility was assessed by calculating the coefficient of variation (CV) of the total ion current (TIC) obtained from the standard solution and tissue section prepared on the same slide. A CV value of 20% or less was considered acceptable.

MALDI-MSI analysis

All MALDI-MSI experiments [38] were conducted using a MALDI ion trap time-of-flight mass spectrometer (iMScope™ TRIO; Shimadzu, Kyoto, Japan) equipped with an Nd:YAG laser (λ = 355 nm; laser diameter = 25 µm). Data were collected at 60-µm intervals. The tissue surface was irradiated 80 times per pixel at a repetition rate of 1 kHz. All data were acquired using an external calibration method in the positive ion detection mode. The laser power and collision energy (dimensionless, as defined in iMScope™ TRIO) were adjusted to 50.0 and 75.0, respectively, and the voltage of the microchannel plate detector was set to 2.1 kV. Following sample analysis, ion images were reconstructed based on the peak intensities extracted from the m/z range of the target ion at m/z 0.02 Da using IMAGEREVEAL™ MS version.1.30.1.12682 SP1 (Shimadzu, Kyoto, Japan). The MS/MS product ion peaks at m/z 406.27 and 422.27 were used to generate the 11DOC-d5 and cortisol-d5 ion images, respectively. To confirm the structure of GirT-cortisol-d5, MS³ measurements were performed on the product ion at m/z 406.27 detected in the tissue (mass imaging was not performed). The collision energy was adjusted to 65.0. The other measurement parameters were the same as those for MS/MS measurements. The target peak at m/z 404.26 was detected in the MS³ spectrum.

Estimation of the ionization efficiency ratio of cortisol-d5 to 11DOC-d5

To determine the ionization efficiency ratio of cortisol-d5 to 11DOC-d5, cortisol, 11DOC, and 11DOC-d5 standards were dissolved in methanol:ultrapure water (1:1, v/v) to a final concentration of 100 µM. A volume of 0.3 µL of each standard solution was spotted onto human adrenal sections. The matrix was applied after the derivatives were analyzed. The average total peak intensity was obtained from the mass spectra of GirT-cortisol (m/z 417.24), GirT-11DOC (m/z 401.24), and GirT-11DOC-d5 (m/z 406.27) in MS/MS (Figure S3). The ionization efficiency ratio of cortisol-d5 to 11DOC-d5 was estimated using the following three-step method: (1) the ionization efficiency ratio of cortisol to 11DOC was calculated as the mean intensity of 11DOC divided by that of cortisol; (2) the ionization efficiency ratio of 11DOC-d5 to 11DOC was calculated as the mean intensity of 11DOC divided by that of 11DOC-d5; (3) the ionization efficiency ratio of cortisol-d5 to 11DOC-d5 was calculated by dividing the value obtained in step (2) by that obtained in step (1).

Data analysis

The acquired raw data were converted to the IMDX format using the IMDX Converter software. The converted IMDX files were processed using IMAGEREVEAL™ MS to extract the mass spectra from the region of interest (ROI) defined by the intersection of the TIC image. Subsequently, TIC normalization was applied to all ROIs to correct for technical bias and obtain the mass spectra independent of the TIC variations. The raw MS images were normalized, and the relative CYP11B1 activity was calculated using the following equation:

$$RelativeCYP11B1activity(\%)=100\times [(\epsilon \times I_{cortisol-d5})/(I_{11DOC-d5}+\epsilon \times I_{cortisol-d5})]$$ 1

where ε is the ionization efficiency ratio of cortisol-d5 to 11DOC-d5. A one-way ANOVA was conducted using GraphPad Prism (GraphPad Software, Boston, MA, USA) to analyze the substrate conversion efficiency or relative CYP11B1 activity among the three groups (NFA, APA, and CPA). Additionally, a t-test was carried out on the two groups (those with and without somatic mutations). The error bars represent standard deviations to compare the data variability. The differences between samples were considered significant at a significance level of less than 5%.

Hematoxylin and eosin (H&E) staining

The same sections used for MSI analysis were stained with H&E following the analysis. The tissues were stained with Mayer’s hematoxylin solution for 1 min, lightly rinsed with tap water, and washed with new tap water for 1 min. They were then immersed in a 1% eosin Y solution for 2 s and washed using the method described above. The sections were dehydrated by soaking them thrice for 1 min each in 100% ethanol, followed by 1-min permeations in 100% xylene twice. After permeation, the sections were mounted onto anti-exfoliation-coated glass slides (Matsunami Glass Industry), with the tissue sections facing upward, sealed with PathoMount, and covered with a coverslip (Matsunami Glass Industry). The samples were dried overnight, and their images were obtained using an all-in-one microscope (BZ-X710; Keyence, Osaka, Japan).

Immunohistochemical (IHC) staining

After the frozen tissue sections were washed in water, they were fixed in 5% formalin at room temperature for 20 min. After fixation, the sections were washed and subjected to antigen retrieval using a high-pH antigen retrieval solution at 95 °C for 10 min. The sections were then rewashed and treated with 0.3% hydrogen peroxide at room temperature for 10 min to inactivate endogenous enzymes. After two washes with Tris-buffered saline (TBS), the primary antibody CYP11B1 (rat monoclonal; dilution = 1:300) was added, and the sections were incubated at room temperature for 60 min. After the reaction, the sections were washed twice with TBS. The secondary antibody, anti-Rat IgG Antibody (H + L) Biotinylated (rabbit polyclonal), was then added, and the sections were incubated at room temperature for 30 min. The sections were washed twice more with TBS. VECTASTAIN Elite ABC Reagent (horseradish peroxidase; HRP) was added as the labeled enzyme, and the sections were incubated at room temperature for 30 min. After the reaction, the sections were washed twice with TBS, and a 3,3′-diaminobenzidine (DAB) substrate solution was added to develop color. Following color development, the sections were washed twice with TBS, after which nuclear staining was conducted using hematoxylin. Dehydration, clearing, and mounting were then carried out similarly to H&E staining, and the sections were observed under a microscope.

Results and discussion

Confirmation of substrate conversion by CYP11B1 in human adrenal cortical adenoma

First, we confirmed the conversion of the substrate into the product by the CYP11B1 activity in human adrenal cortical adenoma sections. After applying the 11DOC-d5 and NADPH substrate solutions via an airbrush, the sections were incubated at 37 °C for 10 min. Subsequently, CHCA was applied as the matrix using an airbrush, and the substrate (11DOC-d5) and product (cortisol-d5) were detected. The product ion spectra of GirT-DOC-d5 and cortisol-d5 in human adrenal cortical adenoma sections are shown in Fig. 1a. To further characterize the fragmentation patterns of GirT-cortisol-d5, we conducted triple-stage mass spectrometry (MS³) analysis (Figure S1). The spectrum revealed fragment ions of GirT-11DOC-d5 and cortisol-d5 at m/z 406.27 and 422.27 in MS/MS, respectively, derived from the non-reagent side of GirT. To confirm that this conversion occurred owing to the human adrenal cortical adenoma CYP11B1 activity, the same experiment was repeated on heat-inactivated sections incubated at 100 °C for 30 min. Cortisol-d5 was detected in the non-inactivated section but not in the inactivated one (Fig. 1b). This result verified CYP11B1-induced conversion of 11DOC-d5 to cortisol-d5 in the human CPA section. To confirm the essential role of NADPH as a cofactor in the CYP11B1 enzymatic reaction, the experiment was repeated using substrate solutions with and without NADPH. Cortisol-d5 was detected in the human adrenal sections only when NADPH was included (Fig. 1c). In addition, we evaluated two matrices, namely CHCA and DHB, to determine the one that was most suitable for cortisol detection. The mean cortisol-d5 intensity was significantly higher in CHCA than in DHB (Figure S2). Therefore, CHCA was selected as the matrix for subsequent experiments. Furthermore, when the CV for the peak intensity was calculated using the same amount of 11-DOC as that during pretreatment, all values fell within the generally accepted 10% range. This indicates minimal variation owing to sample processing, matrix deposition, and derivatization efficiency, suggesting that the pretreatment process has a limited influence on the enzyme conversion rate and activity values. However, despite these controls, tissue boundaries were sometimes unclear in the MSI images, and the correspondence with H&E staining was not complete. One possible reason is the difference in the resolution between MSI analysis and H&E staining, and another reason is the difference in the type of information captured by the two methods. While MSI measures the molecular distribution at a scale of tens of micrometers, H&E staining enables fine morphological structures to be observed at cellular and tissue levels. Consequently, if different tissue components (e.g., light-staining cells and dense cells) coexist within a single MSI pixel, tissue boundaries may appear indistinct. Furthermore, during image analysis, normalization processes may average out signals near tissue boundaries, reducing contrast and causing boundaries to appear indistinct. In addition to these limitations of data processing, phenomena such as matrix supply, diffusion during staining, and tissue deformation may be involved. Future research can improve spatial consistency by optimizing tissue section handling, matrix delivery methods, and MSI measurement conditions.

Fig. 1.

Confirmation of substrate conversion by CYP11B1 in human CPA sections. a Chemical structures and the corresponding mass spectra of GirT-11DOC-d5 and GirT-cortisol-d5 formed by the reaction of GirT with 11DOC-d5 and cortisol-d5, respectively. The precursor and product ion peaks were observed at m/z 465.34 and m/z 406.27 for GirT-11DOC-d5 and at m/z 481.34 and m/z 422.27 for GirT-cortisol-d5, respectively. NADPH functions as an electron donor, enabling the hydroxylation of the substrates. b CPA sections were sprayed with 11DOC-d5 and incubated at 37 °C for 10 min. H&E image (left) and cortisol-d5 distribution (MS/MS product ion peak at m/z 422.27; right) in non-inactivated and inactivated sections. Color scale bar shows the intensity (a.u.) range, which is 0–1.9 × 10⁵. Scale bar is 1500 µm. c H&E image (left) and cortisol-d5 distribution (MS/MS product ion peak at m/z 422.27; right) in human CPA sections with (+) and without (−) NADPH. Color scale bar shows the intensity (a.u.) range, which is 0–5.4 × 10⁵. Scale bars: 1500 µm. Dashed lines indicate the organization’s outline

Optimization of CYP11B1 enzyme reaction for MALDI-MSI

We optimized the substrate and NADPH concentrations and incubation times to visualize the relative CYP11B1 activity. Figure 2 presents the conversion rate determined using the following equation:

$$Conversionrate(\%)=100\times [I_{cortisol-d5}/(I_{11DOC-d5}+I_{cortisol-d5})]$$ 2

Fig. 2.

Comparison of substrate conversion rates (NADPH 10 mM) under varying (a) substrate concentrations, b NADPH concentrations with a fixed substrate concentration of 0.1 mM, and (c) incubation times. All experiments were performed in triplicate

The CYP11B1 enzyme activity was optimal at pH values of 6.5–7.0 and reaction temperatures of 30–37 °C [39]. Therefore, in this study, the substrate solvent pH was set to 7.0, and the enzymatic reaction was conducted at 37 °C. In addition, a 0.05 M phosphate buffer (pH 7.0) was used as the substrate solvent to maintain the pH. However, no significant difference in the conversion rate was observed compared to that using ultrapure water (Figure S3). Given that pH has a significant impact on enzymatic reactions, a phosphate buffer with a pH of approximately 7.0 was chosen to maintain the activity of the CYP11B1 enzyme and preserve the oxidation state of iron ions, which serve as cofactors. When the coenzyme (NADPH) concentration was fixed at 10 mM, no significant difference in substrate conversion was observed across substrate concentrations of 0.05, 0.1, 0.5, and 1 mM (Fig. 2a). At substrate concentrations of 0.05 mM or higher, the enzyme was considered to have sufficient substrate availability. Moreover, when the substrate concentration was held constant at 0.1 mM, the conversion rate was significantly higher at NADPH concentrations of 5 and 10 mM than at 0.05 and 0.5 mM (Fig. 2b). These results indicate that the enzymatic activity peaks when sufficient coenzyme is present. However, excess coenzyme can reduce the enzyme activity, resulting in inactivation, loss of specificity, and undesired side reactions. The results indicate that at least 5 mM NADPH is sufficient to provide an appropriate cofactor concentration for a substrate concentration of 0.1 mM. In addition, the effect of the incubation time following substrate addition was investigated. The conversion rates at incubation times of 0–60 min are presented in Fig. 2c. Plotting the product intensity against the reaction time shows a high degree of linearity (R² > 0.99) during the first 10 min of the reaction (Figure S4). The product formation rate remained constant within this range, indicating that the initial rate condition was met. Furthermore, substrate consumption at 10 min, within the linear range, was less than 10% of the supplied substrate amount. Based on these findings, a 10-min incubation time was determined to be optimal for the enzyme reaction, as decreases in the substrate concentration and enzyme instability do not affect the reaction rate. At this point, the detected cortisol-d5 intensity was considered to reflect the relative distribution of the activity of the CYP11B1 enzyme under linear reaction conditions. All subsequent evaluations of the relative CYP11B1 activity based on the cortisol-d5 signal intensity were performed under these optimized experimental conditions.

Imaging of 11DOC-d5 and cortisol-d5 and visualization of relative CYP11B1 activity

The enzyme activity in tissues was evaluated under optimized conditions. Exogenous substrates (11DOC-d5) and their metabolites (cortisol-d5) were measured, and the distribution of the relative CYP11B1 activity was visualized using a calculation method that considered the tissue-specific matrix effects and ionization efficiency. MALDI-MSI was used to image the relative CYP11B1 activity in MACS human adrenal gland sections (Fig. 3a). The product ion spectra at m/z 406.27 and m/z 422.27 were obtained by MS/MS from the precursor ions of GirT-11DOC-d5 (m/z 465.35) and GirT-cortisol-d5 (m/z 481.34), respectively. Figure 3b shows the ion images of GirT-11DOC-d5 and GirT-cortisol-d5 in MACS human adrenal gland sections without processing, including normalization. Normalization is important for the proper interpretation of the MALDI-MSI imaging data [40]. The intensity of spectra acquired by MALDI-MSI is affected by factors such as the sample’s molecular ionic strength, ionization efficiency, and the matrix type. Normalization is necessary to eliminate these biases and improve data reliability. Common methods for normalization include TIC normalization, peak area standardization, and maximum intensity normalization. In this study, E1 was used for normalization to account for matrix effects specific to 11DOC-d5 and cortisol-d5, as follows:

$$RelativeCYP11B1activity(\%)=100\times [(\epsilon \times I_{cortisol-d5})/(I_{11DOC-d5}+\epsilon \times I_{cortisol-d5})]$$ 3

Fig. 3.

Imaging of 11DOC-d5 and cortisol-d5 for the visualization of relative CYP11B1 activity. a MS/MS mass spectra of GirT-11DOC-d5 and GirT-cortisol-d5 showing product ion peaks at m/z 406.27 and m/z 422.27, respectively, after the enzymatic reaction. Color scale bar shows the intensity (a.u.) range, which is 0–1.4 × 10⁵. b Ion images of 11DOC-d5 and cortisol-d5 in human CPA without pretreatment or normalization (pitch, 80 µm). Color scale bar shows the intensity (a.u.) range, which is 0–5.5 × 10⁵. c Relative CYP11B1 enzyme activity and H&E image after normalization. Color scale bar shows the relative CYP11B1 activity range, which is 0%–70%. Scale bars: 1500 µm. Dashed lines indicate the organization’s outline

where ε is the ionization efficiency ratio of 11DOC-d5 to cortisol-d5, reflecting that the ionization efficiencies of 11DOC-d5 and cortisol-d5 may differ. This difference is believed to correct for variations in the peak intensity. The peak intensity of cortisol-d5 was then divided by the sum of the peak intensities of 11DOC-d5 and cortisol-d5, which represents the total peak intensity at each spot.

A three-step method was used to estimate the ionization efficiency ratios because the rate for the conversion of 11DOC-d5 to cortisol-d5 by CYP11B1 was less than 100% (i.e., conversion was incomplete). We prepared sections of human adrenal glands and analyzed them for equal amounts of 11DOC, 11DOC-d5, and cortisol within the tissue. The total ion peak intensities of these compounds at each spot are shown in Figure S4.

Given that the average ion intensity ratio of cortisol-d5 to 11DOC was 0.93, we modified it by introducing a correction coefficient ε to correct the ionization efficiency as follows:

$$RelativeCYP11B1activity(\%)=100\times [(0.93\times I_{cortisol-d5})/(I_{11DOC-d5}+0.93\times I_{cortisol-d5})$$ 4

Figure 3c shows the distribution of the relative CYP11B1 activity, obtained by processing the raw data for 11DOC-d5 and cortisol-d5 (Fig. 3b) using the equation given above. The H&E image (left side of Fig. 3c) revealed that the relative CYP11B1 activity was heterogeneously distributed across the tissues.

Imaging of relative CYP11B1 activity in human NFA, APA, and CPA

Next, we measured the relative activity of CYP11B1 in adrenal adenomas with different characteristics and evaluated the usefulness and validity of the proposed method by comparing its results with previously reported clinical findings. The MSI and IHC results showing the relative CYP11B1 activity in human NFA, APA, and CPA are presented in Fig. 4a. IHC analysis revealed that CYP11B1 was localized in NFA but distributed throughout APA and CPA. However, in adrenal cortical adenoma, the expression pattern of CYP11B1 changed, leading to its uneven localization. Furthermore, a comparison of the enzyme distribution observed using IHC and enzyme activity measured by MSI demonstrated that the distribution of CYP11B1 did not correspond to its enzymatic activity. These results suggest that both the enzyme distribution and activity must be evaluated to fully understand biochemical reactions in living organisms.

Fig. 4.

Visualization of relative CYP11B1 activity in human adrenal cortical adenoma. The 11DOC-d5 substrate was sprayed onto human adrenal cortical adenoma sections and incubated at 37 °C for 10 min. a H&E images, relative CYP11B1 activity maps, and IHC images of NFA, APA, and CPA sections of human adrenal cortical adenoma. White arrows indicate typical images of clear cells, and black arrows indicate typical images of compact cells. Color scale bar shows the relative CYP11B1 activity range, which is 0%–45%. Scale bar is 1500 µm. b Comparison of relative CYP11B1 activity among NFA, APA, and CPA sections. For each patient with adrenal adenoma (NFA, APA, or CPA), three adjacent sections from a single specimen were analyzed, and the relative CYP11B1 activity was expressed as mean ± standard deviation. Dashed lines indicate the organization’s outline

Although the intensity detected by IHC correlated with the amount of CYP11B1, its expression did not correlate with its activity. Several mechanisms have been proposed to explain the discrepancy between the localization of the CYP11B1 protein and its enzymatic activity. It is possible that CYP11B1 has not undergone the cleavage of the pre-sequence that is essential for its functional expression in the mitochondria or that it is inactivated by post-translational modifications such as phosphorylation. Meanwhile, there are cases in which high enzymatic activity is maintained even when protein expression levels fall below the detection limit of IHC.

Consequently, the CYP11B1 functional activity cannot be assessed based on the IHC results alone. To evaluate activity more accurately, the mean relative activity of CYP11B1 was measured in three sections of each human adrenal cortical adenoma. No significant difference in the relative CYP11B1 activity was observed between NFA and APA; however, the activity tended to be higher in NFA. Meanwhile, the relative CYP11B1 activity is significantly higher in CPA than in both NFA and APA (Fig. 4b). These results suggest that CPA both induces ACS and increases CYP11B1 activity. Furthermore, the relative CYP11B1 activity, as measured by MSI, tended to be lower in PA tissues than in NFA tissues. This indicates a typical PA pathology characterized solely by normal cortisol levels and elevated aldosterone levels [27]. To verify the correspondence between the observed distribution of the CYP11B1 activity and its relative intensity with morphological heterogeneity in tissues, ROIs were set for the distribution of CYP11B1 expression in IHC and that of CYP11B1 activity in MSI. The relative activity values of CYP11B1 were then compared. In NFA, both the distribution of CYP11B1 expression and that of CYP11B1 activity were confined to specific regions. Furthermore, the CYP11B1 phase activity value was higher in the ROI set in the enzyme activity distribution region than that set in the CYP11B1 expression region in MSI. These results suggest that the CYP11B1 enzyme activity may be expressed and functional only in certain regions of NFA. NFA has a structure similar to that of the normal adrenal cortex. Reports indicate that the expression of CYP11B1 and activity of the enzyme do not significantly increase compared to functional adenomas and remain at lower levels [4]. Therefore, it was deemed acceptable for it to remain at a level that does not cause excessive cortisol production. It was thought that the spatial characteristics of the “localized increase in the CYP11B1 activity” observed in NFA differed from those of the “widespread enhancement of the CYP11B1 activity” observed in CPA. Morphological evaluation using H&E staining revealed that APA consisted primarily of pale, lipid-rich cells, while CPA consisted primarily of dense, organelle-rich cells, such as mitochondria. Meanwhile, NFA showed a mixed pattern of both cell types. Examining the relationship between the cell morphology and enzyme activity revealed that the CYP11B1 activity was markedly elevated in the dense cell areas, exceeding expression levels. By contrast, the activity in the pale cell areas was generally low. Morphology and function were also found to be incongruent, as evidenced by morphological changes from pale cells to dense cells, the localization of mitochondria, and the presence of areas of low activity that did not elevate blood cortisol levels. Nevertheless, the content and distribution patterns of these cellular components within each adenoma showed a high correlation with the tumor’s overall CYP11B1 activity. This strongly suggests that the heterogeneous distribution of the CYP11B1 activity directly reflects the morphological heterogeneity within the adenoma.

The validity of this MSI analysis was verified based on the clinical data presented above and the results of the CYP11B1 relative activity. These biological and clinical interpretations strongly support the reliability of the proposed method. Furthermore, visualizing the CYP11B1 activity distribution using MSI is expected to help elucidate the correlation between the morphological characteristics of adrenal adenomas and their cortisol-producing capacity. This should lead to a deeper understanding of functional heterogeneity within individual adenomas.

Imaging of MACS with somatic mutants of the cortisol synthesis pathway

Furthermore, to investigate whether there are differences in the CYP11B1 expression levels and activity between adenomas with and without somatic mutations, we compared the relative CYP11B1 activity in CPA with and without somatic mutations in the cortisol synthesis pathway. Figure 5a presents the H&E images, MSI-based maps of the relative CYP11B1 activity, and IHC images of human CPA from both groups. High enzyme activity was observed in localized areas, regardless of the presence or absence of somatic mutations. As discussed in the previous section, the distribution of CYP11B1 expression did not correspond to its enzymatic activity (Fig. 4a). The mean relative CYP11B1 activity values, calculated from three sections of each human adrenal gland, showed no significant difference between the mutation-positive and mutation-negative groups (Fig. 5b). Furthermore, H&E staining revealed that dense cells predominated in mutation-negative tissues, while clear cells predominated in mutation-positive tissues. For adenomas in which clear cells are predominant, CTNNB1 mutations tend to result in heterogeneous CYP11B1 protein expression patterns, and there is often a significant discrepancy between the predicted activity based on morphological assessment and actual enzymatic activity. In this study, despite the predominance of clear cells, which tended to exhibit low CYP11B1 activity, the relative CYP11B1 activity was high. It is considered that transcriptional promotion of CYP11B1 resulting from constitutive activation of the Wnt signaling pathway associated with the CTNNB1 mutation is involved. This suggests that the CTNNB1 mutation possesses a regulatory mechanism that specifically affects the steroid hormone synthesis pathway and lipid metabolism system.

Fig. 5.

Imaging of human adrenal cortical adenoma with MACS somatic mutants in the cortisol synthesis pathway. a H&E images, relative CYP11B1 activity maps, and IHC images of CPA sections with and without somatic mutations. White arrows indicate typical images of clear cells, and black arrows indicate typical images of compact cells. Color scale bar shows the relative CYP11B1 activity range, which is 0%–60%. Scale bars: 1500 µm. b Comparison of relative CYP11B1 activity in CPA sections with and without somatic mutations. The relative activity of CYP11B1 was measured in three adjacent sections for one patient with a somatic mutation and for one patient without a mutation. The results are expressed as mean ± standard deviation. Dashed lines indicate the organization’s outline

These results indicate that multiple signaling pathways may contribute to CYP11B1 activation, regardless of the presence or absence of somatic mutations. Actually, multiple pathways are involved in CYP11B1 activation [41]. The relative CYP11B1 activity obtained in this study was consistent with previous results, which suggests that this analytical method is valid. However, constraints in obtaining clinical specimens and related data prevented sufficient comparative validation for evaluation using these specimens. Furthermore, the number of reproducibility experiments was limited, as were comparisons with other enzyme activity measurement methods necessary to ensure the performance of this analytical method. Moving forward, the precision, sensitivity, and application scope of this analytical method must be clarified by advancing reproducibility assessments and conducting systematic comparative validation with standard enzyme activity measurement methods.

Conclusion

The enzyme systems involved in the biosynthesis and metabolism of steroid hormones are complex, consisting of various hydroxylases, dehydrogenases, and conjugating enzymes, including the CYP11B1 enzyme. This study revealed the tissue distribution of the relative CYP11B1 activity in adrenal cortical adenomas for the first time, providing important insights into the physiological regulation of its activity. This method is based on delivering labeled substrates to tissue sections and using MSI to spatially visualize the metabolites generated through enzymatic reactions. Provided that the enzymatic reaction products can be detected by mass spectrometry, this method can be applied to other cytochrome P450 enzymes and those involved in steroid biosynthesis. Indeed, we confirmed its applicability to 5α-reductase (EC 1.3.1.30), which converts testosterone to dihydrotestosterone, and aromatase, which converts testosterone to estradiol. However, factors such as substrate specificity, product ionization efficiency, and metabolic complexity within tissues may affect its application scope and sensitivity, so their effects cannot be ruled out. Nevertheless, as this method enables the spatial evaluation of specific enzyme activity within tissues, it is expected to be useful in future research comparing the localized activity of multiple steroid-synthesizing enzymes. Furthermore, it is anticipated to become a powerful tool for elucidating the root causes of metabolic abnormalities and disorders in steroid hormone biosynthesis.

Deeper insights may contribute to the development of therapeutic strategies in the future. While the results of this study suggest clinical significance, this method is currently intended primarily for research purposes. Further standardization of the method as well as validation using external cohorts is considered necessary before it can be applied in routine clinical practice.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

11DOC

11-Deoxycortisol

11DOC-d5

11-Deoxycortisol-2,2,4,6,6-d5

17β-HSD

17β-Hydroxysteroid dehydrogenase

3β-HSD

3β-Hydroxysteroid dehydrogenase

ACS

Autonomous cortisol secretion

ACTH

Adrenocorticotropic hormone

APA

Aldosterone-producing adrenocortical adenoma

CHCA

α-Cyano-4-hydroxycinnamic acid

CPA

Cortisol-producing adrenocortical adenoma

CS

Cushing’s syndrome

CV

Coefficient of variation

CYP11B1

11β-Hydroxylase

CYP11B2

Aldosterone synthase

CYP17A1

17α-Hydroxylase

DAB

3,3′-Diaminobenzidine

DHB

2,5-Dihydroxybenzoic acid

DST

Dexamethasone suppression test

FFPE

Formalin-fixed, paraffin-embedded

GirT

Girard T reagent

H&E

Hematoxylin and eosin

HRP

Horseradish peroxidase

IHC

Immunohistochemistry

MACS

Mild autonomous cortisol secretion

MALDI

Matrix-assisted laser desorption/ionization

MSI

Mass spectrometry imaging

NADPH

β-Nicotinamide adenine dinucleotide 2′-phosphate reduced

NFA

Non-functioning adrenal adenoma

PA

Primary aldosteronism

ROI

Region of interest

TBS

Tris-buffered saline

TFA

Trifluoroacetic acid

TIC

Total ion chromatogram

Author contribution

E.N., M.F., and S.S. performed the measurements; H.U. and S.S. were involved in planning and supervising the study; E.N., H.U., M.F., K.O., and S.S. processed the experimental data, performed the analyses, drafted the manuscript, and designed the figures. Sample collection was E.N., M.F., H.U., T.O., Y.O., S.O., M.E, and K.O., and SS discussed the results, commented on the manuscript, and read and approved the final version of the manuscript.

Funding

Open Access funding provided by The University of Osaka

Data availability

The data that support the findings of this study are available on request from the corresponding author, S.S.

Declarations

Ethics approval

The study protocol was reviewed and approved by the Ethics Committees of both Kyushu University and the University of Osaka (Reference number: 21025-03).

Conflict of interest

The authors declare no conflicts of interest.