COVID-19 patients

From January 26 to March 4, 2021, we enrolled 11 non-vaccinated male patients deceased from COVID-19 complications, confirmed by SARS-CoV-2 RT-qPCR performed during their hospital stay in the Mater Dei Hospital Intensive Care Unit, Belo Horizonte, Brazil. All 11 patients were admitted to the ICU due to severe pulmonary symptoms (Table 1). The standard treatment at ICU included antibiotics, antimycotics, sedatives, muscular relaxants, analgesics, antihypertensives, inotropes, vasopressors agents, invasive mechanical ventilation, and hemodialysis. The Research Ethics Committee of the Mater Dei Hospital and the National Research Ethics Committee (CONEP) approved this investigation under the number CAAE: 30999320.1.0000.5128.

Postmortem collection of both testicles was performed after a legally responsible family member signed an informed consent document.

Testicles were collected through an incision on the median raphe of the scrotum. Two authors (MHF and YLG) collected all testicles no later than 3 h after the patient’s death. To perform viral and testicular genetic studies, fragments of testicular parenchyma were sampled and immersed into RNAlater® solution (Sigma-Aldrich). To investigate the viral replicative activity and testosterone and angiotensin levels, testis fragments were also sampled and then snap-frozen in liquid nitrogen. The remaining testicular halves were immersed in different fixatives, such as paraformaldehyde 4%, Bouin, methacarn, and glutaraldehyde 4%. Testis fragments were embedded in methacrylate, Epon 812 resin, and Paraplast® for histological, transmission electron microscopy, and immunohistochemistry analyses.

Control patients

The control group was composed of six patients who underwent orchiectomy due to prostate cancer suspicion. These patients did not go through any treatment at the time of orchiectomy. Moreover, they exhibited normal spermatogenesis in seminiferous tubules. The age and hormonal levels of these patients are presented in Additional file 2: Table S1. Testicular fragments were obtained after the study was approved by the Ethics Committee in Research of the Universidade Federal de Minas Gerais COEP/UFMG (COEP ETIC n°117/07). All patients signed the informed consent.

Testicular samples were placed in liquid nitrogen, embedded in methacrylate, Epon, paraplast®, and conserved in RNAlatter. These samples were used for TEM, histological, hormonal, and molecular analyses. The mean age of patients was 58 years old, ranging from 46 to 65 years old.

Detection OF SARS-CoV-2 in testis tissueGenetic assays SARS-CoV-2 detection using standard RT-qPCR

RNA of samples was extracted according to the protocol specified by the extraction kit (QIAamp® Viral RNA Mini Kit). The samples were stored in an ultra-freezer at − 80 °C. The collected samples were tested for the presence of SARS-CoV-2 viral RNA by RT-qPCR with primers to amplify the envelope (E) gene and the human transcript of the gene for RNAseP (RNP) as the endogenous control [44]. Samples were considered positive with a cycle threshold (CT) ≤ 40.

RT-qPCR using specific viral primers

Testes samples were macerated and submitted to RNA extraction using the Viral RNA Kit (Zymo Research, USA), following the manufacturer’s protocol. A two-step RT-qPCR approach was performed to optimize the detection of the viral RNA in the tissue samples and to avoid the possible influence of host cellular RNA. The obtained RNA was first submitted to cDNA synthesis using the CDC’s SARS-CoV-2 specific reverse primer 2019-nCoV_N1-R (TCT GGT TAC TGC CAG TTG AAT CTG) and the SuperScript™ III First-Strand Synthesis System (Invitrogen, Brazil).

The viral cDNAs were then amplified in a qPCR reaction using the GoTaq qPCR Master Mix (Promega, USA). We used both N1 primers from CDC’s SARS-CoV-2 detection protocol (2019-nCoV_N1-F: GAC CCC AAA ATC AGC GAA AT; 2019-nCoV_N1-R: TCT GGT TAC TGC CAG TTG AAT CTG) and followed the cycling recommendation indicated by the enzyme’s supplier in a QuantStudio 3 Real-Time PCR System (Applied Biosystems, USA). To normalize the results, the same process was performed to amplify the human β-actin control.

Protein assaysNanosensor

Gold nanorods (GNRs) were synthesized by the seed-mediated growth method as previously described [45]. The nanoparticles contained an average aspect ratio of 10 × 38 nm and a light absorbance peak of 713 nm. They were covalently functionalized with the polyclonal antibody anti-Spike protein (Rhea Biotech, Brazil) and a polyclonal antibody anti-Nucleocapsid protein (CTVacinas, Brazil) through a carbodiimide-activated amidation reaction. The binding between the gold surface and the antibodies was mediated by adding a capping layer formed by α-lipoic acid. A 2 mM α-lipoic acid solution (LA; Sigma Aldrich, USA) in ethanol was added to the GNR suspension (0.039 mg/mL). These suspensions were exposed to an ultrasonic bath (UNIQUE model U5C1850, 154W, 25KHz) at 55 °C for 30 min. The suspension was sonicated again for 2 h at 30 °C and left to rest overnight at RT to stabilize the interaction. GNRs were then centrifuged at 5600 g for 10 min and suspended in an aqueous solution. The GNR-LA complexes were kept at 4 °C in the dark. Next, the modified GNRs were re-dispersed in a 10 mM phosphate buffer containing 16 mM EDAC and 4 mM sulfo-NHS (30 min in an ice-bath under sonication). After another centrifugation step, GNR-LA suspensions were blocked with poly (ethylene glycol)-thiolate (5kD mPEG-SH, 10-4 mM, from Nanocs) for 10 min in an ice bath, under stirring (Additional file 1: Fig. S1a-b).

TEM images of the nanosensors were obtained on a 120 kV FEI Technai G2-12 (Spirit BioTwin, USA) microscope (Additional file 1: Fig. S1c). Samples were directly dripped onto a holey carbon film supported on a copper grid (400 mesh) (Pelco®, USA) without any further processing. Zeta potential was obtained with a Zetasizer Nano ZS90 analyzer from Malvern at an angle of 173° at RT, as previously described [7, 46, 47]. The nanoparticle’s size and zeta potential were measured simultaneously three times and in triplicate (Additional file 1: Fig. S1d-e).

Samples’ labeling and measurements were previously described [7]. Briefly, GNR-S were incubated with 1mg/mL of anti-S protein polyclonal antibody (Rhea Biotech, Brazil). Next, the samples were labeled with goat anti-rabbit IgG-CFL 488, 100 μg/mL λex: 488 nm, λem: 520 nm (Santa Cruz Biotechnology, USA) (Additional file 1: Fig. S1f). For GNR-N, samples were incubated with 10 μg of a polyclonal antibody anti-protein N for 60 min under axial shaking. Unbound antibodies were blocked with BSA (1%) for 30 min. Next, samples were labeled with Alexa Fluor® 546 goat anti-rabbit IgG H & L, 5 μg/mL λex: 540 nm, λem: 585 nm (Invitrogen, USA) and incubated under continuous shaking in the dark and at RT for 30 min (Additional file 1: Fig. S1g). Samples were measured using the spectrophotometer Varioskan Flash spectral scanning multimode reader (Thermo Scientific, USA).

To determine the concentration of antibodies in the LSPR-nanosensor, an antibody curve was carried out ranging from 0.25 μg to 10 μg for anti-S and anti-N proteins, respectively (Additional file 1: Fig. S1h-l). The concentration chosen was 0.5 μg for both GNR-S and GNR-N nanosensors. Afterward, the nanosensors were exposed to purified S- and N-proteins, ranging from 10 μg to 10 fg, and their binding affinity was assessed by UV-Vis spectroscopy (Thermo Scientific, USA).

Finally, the GNR-S and GNR-N nanosensors were exposed to macerated testes samples (obtained from the 11 patients with COVID-19), diluted in PBS-1X, for 30 min at 4 °C under sonication, and their respective LSPR were measured using optical plates (Costar, USA), by UV-Vis spectroscopy (Thermo Scientific, USA). Spectra analyses were performed using OriginPro version 9.0. We normalized and smoothed the curves and measured the biosensing event through the X-axis intercept of the derivative of the Gaussian peak for each patient. We compared the COVID-19 patient curves to the GNR and negative control patient curves (Additional file 1: Fig. S1m-q). We focused on the redshift that occurred at point zero (derivative axis) of the longitudinal peak. In this analysis, 5nm or below shifts were not considered significant.

Immunofluorescence against Spike protein

Immunofluorescence was performed using a validated primary anti-S protein antibody [48] (Rhea Biotech, Brazil) to detect and corroborate the viral presence in the testicular parenchyma (Additional file 2: Table S2). Reactions were visualized using Alexa-488 (Thermo Fisher Scientific, USA) conjugated secondary antibody, and images were acquired using a Nikon Eclipse Ti fluorescence microscope. As controls, we performed three different assays: (1) we used the testes of controls (prostate cancer patients); (2) we omitted the primary antibody in the testes of COVID-19 patients; and (3) we performed a negative control antigen, incubating the primary antibody with purified Spike protein (1:10; donated by CT-Vacinas-UFMG) before the immunofluorescence assay (Additional file 1: Fig. S2m-p).

SARS-CoV-2 activity and replicationVERO cell exposition to testicular macerates

In the BSL-3 laboratory, testis samples from the 11 COVID-19 patients were macerated using the tissue Lyser (Tissuelyser II Retsch) and metallic beads (Qiagen, USA). All fragments were disrupted in 600 μL of DMEM without serum, and 300 μL was used for infection in vitro. VERO CCL-81 cells were seeded in a 6-well plate (Sarstedt, Germany) under 90% confluence. Culture media was discarded, and testis macerates were placed in the cell culture system for 1 h, at 37 °C and 5% CO2, with periodic homogenization. We did a negative control, exposing VERO cells to DMEM, and a positive control, exposing VERO cells to DMEM with SARS-CoV-2 (Wuhan strain SARS-CoV-2; MOI 0.01; donated by Dr. Edison Luiz Durigon).

After incubation, 2 mL of fresh supplemented DMEM was added, and cells were maintained (at 37 °C and 5% CO2) for 48 h. Next, 300 μL of each sample were adsorbed in another 6-well plate containing Vero CCL-81 cells for 1 h, at 37°C and 5% CO2, with periodic homogenization. Afterward, 2 mL of fresh DMEM was added, and cells were incubated. In parallel, 200 μL of each supernatant was collected (1st and 2nd day; Table 2), inactivated in 200 μL of Isothiocyanate of Guanidine, and frozen at – 80 °C for the RT-qPCR analysis.

Viral detection in the supernatant

To confirm the isolation of SARS-CoV-2, we extracted the total RNA from the supernatant (200 μl) using a using PureLink RNA Mini Kit (Invitrogen, USA), and the RT-qPCR reaction was performed using iTaq Universal Probes One-Step Kit (Bio-Rad). The SARS-CoV nucleocapsid RNA (N1) was measured with the following primer-probe set: F 5′-GACCCCAAAATCAGCGAAAT-3′, R 5′-TCTGGTTACTGCCAGTTGAATCTG-3′, and Probe 5′-FAM-ACCCCGCAT/ZEN/TACGTTTGGTGGACC-3IABkFQ-3′. The human transcript of the gene for RNAseP (RNP) was used as the endogenous control.

SARS-CoV-2 genomic and subgenomic RNA evaluation in VERO cells

Cytopathic effects were monitored, and the results were compared with a Mock-infection and concentrated SARS-CoV-2 virus as a positive control. After two passages, infected and control Vero CCL-81 cells were detached with Trypsin (Gibco, USA), spun down, and the concentrated pellets were kept frozen at – 80 °C for the RT-qPCR analysis. After cell lysis using Tissuelyser II Retsch and metallic beads (Qiagen, USA), 200 μL of each sample was added to 200 μL of lysis buffer with 2-mercaptoethanol. After homogenization, 200 μL of 70% ethanol was added. After washing with buffers, we added 40 μL of RNase-free water, and the samples were centrifuged for 1min 12000 × g to collect the total RNA.

SARS-CoV-2 genomic RNA was measured with the following primer-probe set: CoV-F3 (5′-TCCTGGTGATTCTTCTTCAGGT-3′), CoV-R3 (5′-TCTGAGAGAGGGTCAAGTGC-3′) and CoV-P3 (5′-AGCTGCAGCACCAGCTGTCCA-3′). SARS-CoV-2 subgenomic RNA was measured with the following primer-probe set: CoV-sgRNA-F (5′-CGATCTCTTGTAGATCTGTTCTC-3′), CoV-sgRNA-R (5′-ATATTGCAGCAGTACGCACACA-3′), and CoV-sgRNA-P (5′-ACACTAGCCATCCTTACTGCGCTTCG-3′).

Immunofluorescence and immunoperoxidase in Vero cells

To confirm viral isolation from the patient’s testes, we performed an immunofluorescence assay using the anti-S protein commercial antibody (Abcam, USA). The cells were seeded in a 96-well Black plate with a transparent bottom (Costar, Corning Incorporated, USA), and 50 μL of each supernatant from the third passage were adsorbed for 1 h, at 37 °C and 5% CO2, with periodic homogenization. Next, 50 μL of supplemented DMEM was added to each well and kept incubated for 4 h.

After the VERO infection, the contents were removed, and cell fixation was carried out using Paraformaldehyde 4% (BioRad, USA) for 15 min. The plates were washed with PBS 1X and blocked with a PBS solution containing 5% of Donkey Serum for 1h. The cells were then incubated with the anti-S antibody (Abcam, USA) for 1 h at 37°C. Afterward, cells were rinsed with PBS 1X and incubated with secondary antibody Alexa Fluor 488 anti-rabbit (Invitrogen, USA) and DAPI 300nM (dilution 1:1000) for 2 h at dark and room temperature. Finally, cells were washed with PBS 1X, and images were taken using EVOS 5000 ML (Invitrogen, USA).

We performed an immunohistochemistry assay using the anti-S commercial antibody (Rhea Biotech, Brazil). After fixation, the plates were blocked with a PBS solution containing 3% FBS for 15 min. Then, the cells were incubated overnight with the anti-S antibody (dilution 1:500) at RT. The cells were also incubated with an anti-rabbit IgG antibody conjugated to horseradish peroxidase (Promega, USA), diluted at 1:2500 at RT for 60 min. Immunoassay was revealed using the KPL TrueBlue Peroxidase Substrate (SeraCare, USA) for 10 min at RT, under gentle stir.

Histomorphometric analysis

All testicular slides were scanned using the Panoramic MIDI II slide scanner (3DHISTECH, Hungary). The histomorphometric analyses were performed using the CaseViewer software (3DHISTECH, Hungary) and the Image J v.1.45s software (Image Processing and Analysis, in Java, USA).

Seminiferous epithelium cell composition

To describe the seminiferous epithelium integrity, at least 50 seminiferous tubule cross-sections were evaluated and classified according to the germ cell composition in the seminiferous epithelium. According to this classification, the patients were then categorized into three phases, as follows: phase 1, cross-sections containing all germ cell layers (spermatogonia, spermatocytes, and spermatids); phase 2, cross-sections containing spermatogonia and spermatocytes; phase 3, cross-sections containing only spermatogonia and/or degenerating/Sertoli cell-only seminiferous tubules. The results are presented as the percentage of seminiferous tubules in each category. After this analysis, we examined the clinical data of COVID-19 patients to understand the evolution of testicular pathogeny.

Seminiferous tubule measurements

Seminiferous tubules were analyzed using computer-assisted image analysis of 30 randomly chosen seminiferous tubules cross-sections per donor. To determine the seminiferous tubule diameter and tunica propria width, the measurements were taken at × 400 magnification, and the results were expressed in micrometers.

Leydig cell volume density

The volume densities (%) of testicular tissue components were obtained after counting 7200 points over testis parenchyma. The intersections that coincided with Leydig cells were counted in 15 randomly chosen fields by horizontal scanning of the histological sections at × 200 magnification [49].

Hemorrhagic scores

As red cell bleeding was a common finding, this pathology was measured in four scores, as follows:

1.

Patients who presented many red blood cells inside the seminiferous tubule lumen and intertubular compartment (score 3);

2.

Patients with vast areas of red blood cells bleeding in the intertubular compartment (score 2);

3.

Patients with small areas of red blood cell bleeding in the intertubular compartment (score 1);

4.

Patients without red blood cell bleeding (score 0).

Histochemistry techniquesMast cell counts

Toluidine-blue staining was used to determine the number of mast cells. We investigated 15 testicular fields (× 20 magnification), and the cells were quantified per patient.

PAS staining

A Schiffs kit (Sigma-Aldrich, USA) was used for PAS staining, as per the manufacturer’s protocol. In brief, the sections were pre-treated with periodic acid for 5 min at RT, slowly rinsed in distilled water, and then stained with Schiffs solution for 10 min at RT in the dark. The nuclei were stained with hematoxylin for 5 min at RT, followed by six dips in 1% hydrochloric alcohol. After dehydration with 70, 90, and 100% graded alcohol, the sections were immersed twice in xylene for 10 min each. Then, the slides were mounted with a coverslip and sealed with Entellan resin (Sigma-Aldrich, USA). Images were captured using light Olympus microscopy (BX-60).

Masson’s Trichrome and Picrosirius red

Tissue samples were stained with Masson’s Trichrome and Picrosirius red to assess fibrosis and collagen types I and III. Images of the two techniques were captured in a Spot Insight Color digital camera adapted for Olympus BX-40, using the Spot software version 3.4.5. To determine the area of fibrosis and differentiate the types of collagens, images of three random areas of each patient were obtained at × 100 magnification. Images were analyzed with Image J v. 1.53c software (National Institutes of Health, USA) using the Color Deconvolution tool, getting the average of the three areas evaluated [50].

Immunostaining

For immunostaining, deparaffinized sections (5 μm thick) were dehydrated and submitted to heat-induced antigenic recovery (water bath) with buffered sodium citrate (pH 6.0) at 90 °C for 40 min. Then, the sections were immersed in BSA 3% (in PBS) solution to block non-specific antibody binding and incubated overnight at 4 °C with primary antibodies (Additional file 2: Table S2). Reactions were visualized using biotin-conjugated secondary antibodies (anti-goat: 1:100 dilution, Abcam, ab6740; anti-mouse: 1:200 dilution, Imuny, IC1M02; anti-rabbit: 1:200 dilution, Abcam, ab6720) combined with Elite ABC Kit (Vector Laboratories, USA). Signal detection was obtained via peroxidase substrate 3,39-diaminobenzidine (DAB; Sigma Aldrich, USA) reaction and counterstaining with Mayer’s hematoxylin (Merck, USA).

To identify targets of SARS-CoV-2 in the testis intertubular compartment, double-immunofluorescence anti-S with anti-CD-68 or anti-Chymase was performed. Reactions were visualized using Alexa-488 (1:100 dilution), Alexa-546 (1:200 dilution), and Alexa-633 (1:200 dilution), all from Thermo Fisher Scientific (USA), and conjugated secondary antibody. For ACE2 evaluation, we used anti-ACE2 (Proteintech, USA) and Alexa-594 (1:100 dilution) from Jackson Immunoresearch (USA). Images were acquired using the Nikon Eclipse Ti fluorescence microscope.

To analyze immunopositive macrophages and T lymphocytes, CD68 and CD3 positive cells were quantified by counting 10 testicular fields at × 400 magnification. Only cells with visible nuclei, brown cytoplasm, and morphology compatible with the evaluated cells were counted. Microvessel density analysis was also performed in 10 fields at × 400 magnification, and only CD31 immunopositive structures with or without lumen were counted (vessels containing muscle walls were not counted).

Transmission electron microscopy

Testes fragments were fixed by immersion in 4% glutaraldehyde (EMS, USA). Smaller pieces (1–2 mm thickness) were obtained and postfixed in reduced osmium (1% OsO4 and 1.5% potassium ferrocyanide in distilled water) for 90 min, dehydrated in ethanol, and embedded in Araldite epoxy resin. Ultrathin sections (60 nm thick) were obtained using a diamond knife on a Leica EM UC6 ultramicrotome (Leica Microsystems) and mounted on 200 mesh copper grids (Ted Pella). The ultrathin sections were stained with lead citrate (Merck, USA) and analyzed using a transmission electron microscope (Tecnai G2-12 Thermo Fisher Scientific/FEI, USA). The viral particle was defined when presenting a round to oval shape, a membrane, an electron-dense interior, and within a membrane compartment [51].

Enzymatic and hormone measurements

The enzymatic activity of MPO and NAG and the total concentration of testosterone and angiotensin II were determined in human testis homogenates. For this purpose, 50 mg of snap-frozen testis from deceased COVID-19 patients and controls were homogenized in 450 μL cold PBS supplemented with a protease inhibitor cocktail (Cat n° S8830, Sigma-Aldrich). After three freeze/thaw cycles in a liquid nitrogen/water bath (37 °C), the samples were centrifugated (14.000 g, 10 min, 4 °C), and the supernatants were collected.

For MPO and NAG measurements, 100 μL of tissue homogenates were mixed 1:1 in MPO buffer assay (0,1M Na3PO4, 0.1% [w/v] HETAB, pH 5.0) or NAG buffer assay (0.2M citric acid, 0.2M Na2HPO4, pH 4.5), respectively, just prior the freeze/thaw step. The activity of MPO and NAG, which is an indirect estimation of the abundance of neutrophils and macrophages, was measured in a colorimetric enzymatic assay. Intratesticular testosterone levels were determined in testis homogenates using a chemiluminometric immunoassay run on the Atellica IM Analyzer (Siemens Healthcare Diagnostics). The concentration of Angiotensin II was measured by ELISA, according to the procedures supplied by the manufacturer (MyBioSource, San Diego, CA, USA). The kit applied the sandwich ELISA technique. The sensitivity of the assay was 12 pg/mL.

Testicular gene expressions

Total RNA was isolated from testes using AurumTM Total RNA MiniKit® (BioRad, USA). A Nanodrop spectrophotometer (Thermo Fischer, USA) was used to measure the quantity and integrity of total RNA. RNA (2 μg/sample) was reverse transcribed using the iScript cDNA Synthesis Kit (BioRad, USA). cDNAs (10 ng) were amplified by qPCR with iTaq Universal SYBR Green Supermix (BioRad, USA) in Rotor-Gene Q (Qiagen, USA). The primer sequences used can be found in Additional file 2: Table S2. Relative levels of expression were determined by normalization to RPL19 e HPRT1 using the ∆∆CT method. The testicular gene expressions were displayed as Heat maps (Fig. 5 and Fig. 6) and individual graphs (Additional file 1: Fig. S6 and Additional file 1: Fig. S7).

Statistical analyses

Demographics and clinical characteristics were presented using descriptive statistics: mean and standard deviation (SD) for normally distributed continuous data, median and interquartile range (IQR) for non-normally distributed continuous data, and proportions and frequencies for categorical data. The presence and the strength of a linear relationship between demographics and clinical characteristics and the damage to the testicles were analyzed using Spearman correlation.

All quantitative data were tested for normality and homoscedasticity of the variances following Kolmogorov–Smirnov (Dallal–Wilkinson–Lilliefor) and Bartlett tests. Data from the fluorometry assay analyses were evaluated by one-way ANOVA for comparisons within groups, followed by Newman-Keuls (normal distribution). Histomorphometric and gene expression data were analyzed by unpaired Student’s t-test, comparing COVID-19 groups to controls and COVID-19-P1 to COVID-19-P2 (COVID-19-P3 was not considered for statistics). The data obtained were represented as the mean ± SEM and geometric mean ± SD. Graphs and statistical analyses were conducted using GraphPad PRISM v6.0 (GraphPad Software, Inc.). Differences were considered statistically significant at p < 0.05.The individual raw data are available online [52].