We examined autopsied human adrenal glands. Adrenocortical steroid hormones are synthesized from cholesterol, which is mostly esterified and stored in intracellular lipid droplets until steroidogenesis (16). Adrenocortical tissues obtained from autopsied patients who died after PCI usually show a diffuse pattern of cholesterol ester depletion (hereinafter lipid depletion) (16, 17). In contrast to death from PCI, death occurring within 24 h of symptom onset is referred to as sudden death (18). Adrenocortical tissues obtained from autopsied patients who died suddenly are considered under near-normal physiological conditions. In the current study, we defined the control group as patients who died suddenly and the PCI group as patients with adrenal lipid depletion who died after PCI.

Regarding the PCI group, adrenal lipid depletion can be histologically identified using hematoxylin and eosin (H&E) staining. To obtain adrenal glands with lipid depletion, we assessed the adrenal glands obtained from 262 serial autopsies conducted between September 2015 and December 2020 at the Tokyo Metropolitan Institute for Geriatrics and Gerontology, Tokyo, Japan. Immediately after harvesting during the autopsy, the adrenal glands were routinely fixed with 10% buffered formalin and sectioned at 5-mm intervals in a plane vertical to the long axis. Each section was processed for microscopic examination. A 2-μm-thick section was cut from each paraffin-embedded tissue block and stained with H&E. Three pathologists (K. N., Y. Y., and H. S.) examined H&E-stained sections independently and selected 26 patients with adrenal lipid depletion. The exclusion criteria were as follows: (i) a history of endocrine disorders such as pituitary adenomas and adrenocortical disorders; (ii) use of any form of corticosteroids or drugs that possibly affect the hypothalamic–pituitary–adrenal axis, at least during the last month before death; (iii) primary or metastatic adrenal tumors; and (iv) diffuse necrosis, autolysis, or hemorrhage in the adrenal tissue. In accordance with these criteria, eight patients were withdrawn from this study. Overall, the PCI group included 18 elderly patients with adrenal lipid depletion (10 men and 8 women). All patients in the PCI group were Japanese and died of prolonged critical illness.

Regarding the control group, 18 age- and sex-matched specimens of autopsied patients who died a sudden death (from January 2002 to December 2020) were retrieved from the pathology files of the Tokyo Metropolitan Institute for Geriatrics and Gerontology, Tokyo, Japan. Adrenal H&E-stained sections (cut in a central plane vertical to the long axis) were examined independently by three pathologists (K. N., Y. Y., and H. S.), and four patients were withdrawn from the study (based on the exclusion criteria). Adrenal lipid depletion was not observed in any of these patients. Overall, the control group included 14 elderly patients with adrenals showing near-normal conditions (7 men and 7 women). All patients in the control group were Japanese.

Ten percent formalin-fixed paraffin-embedded (FFPE) samples of left or right adrenals were sliced into 2-μm sections for determination of telomere length using Q-FISH. The tissue slides were processed by the Q-FISH method, as described previously (19, 20). Tissue sections were hybridized with peptide nucleic acid probes for the telomere (Telo C-Cy3 probe: 5′-CCCTAACCCTAACCCTAA-3′; catalogue number F1002, Fasmac, Atsugi, Japan) and centromere (Cenp1-fluorescein isothiocyanate or FITC probe: 5′-CTTCGTTGGAAACGGGGT-3′; custom-made, Fasmac), and the nuclei were stained with 40,6-diamidino-2-phenylindole (DAPI) (Molecular Probes, Eugene, OR, USA). In Q-FISH images, we could identify the individual ZG and ZF parenchymal cells and chromaffin cells, based on their respective morphological features: ZG cells had less abundant cytoplasm compared to ZF cells, which were organized into bundles, and chromaffin cells had relatively regular nuclei in size and contour (21–23). ZR parenchymal cells could be identified because they had cytoplasm with prominent lipofuscin, displaying brilliant yellow autofluorescence in Q-FISH images (24). Q-FISH digital images were captured using a charge-coupled device camera (RETIGA R3; Teledyne Photometrics, Surrey, BC, Canada) mounted on an epifluorescence microscope (80i, Nikon, Tokyo, Japan) equipped with a triple band-pass filter set for DAPI/fluorescein isothiocyanate (FITC)/Cy3 (Part #61010, Chroma Technology Corp., Rockingham, VT, USA) and a 40× objective lens (Plan Fluor 40/0.75; Nikon). Microscopic control and image acquisition were performed using the Image-Pro Plus software package (version 7.0; Media Cybernetics Co. Ltd., Silver Spring, MD, USA). As reported previously, the captured images were analyzed with our original tissue analysis software, TissueTelo, which can calculate the telomere to centromere ratio (TCR) as an estimate of the relative telomere length in the individual nucleus (19, 25). As a control for variations during sample preparation, we also performed Q-FISH in a cell block-section of a cultured cell line, TIG-1 (26), placed on the same slides as the adrenal gland sections. TIG-1 is a human fibroblast strain established from fetal lung and commonly used in studies of cellular senescence (26, 27). The TIG-1 cell line was subcultured (34 PDL; population doubling level) and had a telomere length of 8.64 kbp, as measured by Southern blot analysis. Each TCR for the target cells was divided by the mean TCR for the cell block on the same slide to obtain the normalized TCR (NTCR). We measured the NTCR of individual parenchymal cells in the three adrenocortical zones and the adrenomedullary chromaffin cells. In each type of adrenal parenchymal cells of each patient, the number of cells analyzed ranged from 110–120, and we defined the mean NTCR as RTL in this study.

At the Tokyo Metropolitan Institute for Geriatrics and Gerontology, residual serum specimens obtained from patients in a hospital ward or emergency department are stored for one week in laboratory refrigerators for additional tests. When the patients die and autopsies are performed, after obtaining informed consent, the residual serum samples are pipetted into several 1.5-mL plastic vials and stored at -80 °C until assayed. In this study, residual serum samples were obtained from five men and five women in the control group, and eight men and five women in the PCI group. The DHEA-S levels were measured by chemiluminescent enzyme immunoassay using a UniCel DxI 800 Access Immunoassay System (Beckman Coulter Inc., Brea, CA, USA).

We examined the antemortem serum albumin levels by referring to the medical records of each patient in the control and PCI groups.

Patients with PCI experience oxidative stress caused by reactive oxygen and nitrogen species (31, 32). Therefore, we performed immunohistochemical analyses of 8-hydroxydeoxyguanosine (8OHdG) and phosphorylated gamma histone H2AX (γH2AX), which are representative markers of oxidative stress and DNA damage (33, 34).

Statistical analysis was performed using GraphPad Prism version 5.01 for Windows (GraphPad Software Inc., CA, USA). In this study, the Mann–Whitney test was used to compare two groups. When the Mann–Whitney test was not applicable, the Wilcoxon signed–rank test was used to compare the two groups. The Kruskal–Wallis test with Dunn’s multiple comparison test was used to compare three or more groups. Correlations were analyzed using Spearman’s correlation coefficients. Statistical significance was set at P < 0.05.

Table 1 summarizes the clinicopathological characteristics of autopsied patients. The mean age in the control group was 85.3 years in men and 87.7 years in women, whereas that in the PCI group was 78.8 years in men and 85.6 years in women. Disease duration represents the estimated period from symptom development to death ( Table 1 ). Figure 1 shows the representative H&E-stained histological images of the adrenal glands in the control and PCI groups.

Figure S1 shows the representative Q-FISH images of the adrenal glands. Table 2 summarizes the RTLs in the three types of adrenocortical parenchymal and chromaffin cells in the adrenal medulla. In the control group, the RTLs in ZR and ZF cells were the longest and shortest, respectively. In the PCI group, the RTLs in ZR and ZG cells were the shortest and longest, respectively ( Figure 2A ). When we compared the RTLs in the same type of adrenal parenchymal cells between the control and PCI groups, the RTLs in ZR cells were shorter in the PCI group than in the control group for both men and women (P = 0.0001 in men; P = 0.0012 in women; Figure 2B ). However, the RTLs in the remaining three types of adrenal parenchymal cells did not differ between the control and PCI groups ( Figure 2B ).

Figure S2 shows the representative images of double-staining immunohistochemistry for CytB5 and CgA in the adrenal glands. Figure S3 shows a representative digital image processing of double-staining immunohistochemistry. Table S1 shows the number of CytB5-positive cells (i.e., ZR parenchymal cells) and the total number of adrenocortical parenchymal cells (i.e., the sum of the numbers of ZG, ZF, and ZR parenchymal cells). In men, the number of CytB5-positive cells was higher in the PCI group than in the control group (P < 0.0001), whereas the total number of adrenocortical parenchymal cells did not differ between the two groups ( Figure 3A ). In women, the number of CytB5-positive cells and the total number of adrenocortical parenchymal cells were higher in the PCI group than in the control group (P < 0.0001 and P = 0.0008) ( Figure 3B ). For both men and women, the proportion of the number of CytB5-positive cells to the total number of adrenocortical parenchymal cells was higher in the PCI group than in the control group (P < 0.0001 for both men and women) ( Figures 3A, B ).

In the PCI group, there was no significant correlation between disease duration ( Table 1 ) and RTLs in the ZR cells ( Table 2 ) for both men and women (P = 0.1231 in men; P = 0.6646 in women).

In the PCI group, there was no significant correlation between disease duration ( Table 1 ) and the number of CytB5-positive cells ( Table S1 ) for both men and women (P = 0.5603 in men; P = 0.8401 in women).

Figure S4 shows representative images of double-staining immunohistochemistry for CytB5 and Ki-67 in the adrenal glands. Table S2 summarizes the number of CytB5-positive cells and Ki-67 proliferation index in CytB5-positive cells. For both men and women, the Ki-67 proliferation index in CytB5-positive cells was higher in the PCI group than in the control group (P = 0.0039 in men; P = 0.0063 in women).

Figure S5 shows representative images of double-staining immunohistochemistry for CYP17A1 and CgA in the adrenal glands. For both men and women, the number of CYP17A1-positive cells was higher in the PCI group than in the control group (P < 0.0001 for both men and women) ( Table S3 ). Similarly, for both men and women, the proportion of the number of CYP17A1-positive cells to the total number of adrenocortical parenchymal cells was higher in the PCI group than in the control group (P < 0.0001 for both men and women) ( Table S3 ).

DHEA-S comprises the majority (over 90%) of DHEA in circulation owing to its slow rate of clearance from circulation and long biological half-life (35, 36). Thus, we checked the antemortem serum DHEA-S levels instead of DHEA levels. Table 3 shows a summary of the antemortem serum DHEA-S levels. There were no significant differences in the serum DHEA-S levels between the control and PCI groups for both men and women (P = 0.9416 in men; P = 0.6905 in women).

The vast majority of DHEA-S in the blood is bound to albumin (37), and changes in the serum albumin levels are related to changes in the serum DHEA-S levels (38). Thus, we checked the antemortem serum albumin levels by referring to the medical records of each patient. Table 4 shows a summary of the antemortem serum albumin levels at the last laboratory test of a patient’s life. The serum albumin levels were lower in the PCI group than in the control group for both men and women (P = 0.0007 in men; P = 0.0094 in women) ( Figure 4 ). Figure S6 shows changes in the serum albumin levels during terminal hospitalization in the PCI group.

Figure S7 shows representative images of the immunohistochemistry for 8OHdG and γH2AX in the adrenal glands. In the control group, positive immunoreactivities for both 8OHdG and γH2AX were observed mainly in the ZG and outer ZF for both men and women. However, in the PCI group, positive or weakly positive immunoreactivities for both 8OHdG and γH2AX were diffusely distributed in the adrenal cortex and medulla. Many inflammatory cells also showed positive immunoreactivities for both 8OHdG and γH2AX in the PCI group.