Introduction

Cysteine proteases, including cathepsin L (CTSL), hold pivotal roles in human pathobiology due to their multifaceted activities within and outside cells. Emerging evidence links cathepsins, notably CTSL, to metabolic disorders like obesity and diabetes, as well as diabetic complications (Crawford et al, 2022; Ding et al, 2020; Limonte et al, 2022). Previous studies have associated CTSL with proteinuria in podocytes (Reiser et al, 2004) and intraocular angiogenesis (Shimada et al, 2010), suggesting its potential as a therapeutic target for diabetic nephropathy and vision-threatening conditions such as proliferative diabetic retinopathy (Shimada et al., 2010).

Recent investigations highlight CTSL’s involvement in the cleavage and processing of the SARS-CoV-2 spike protein, critical for viral entry and replication within host cells, as reported by our group (Zhao et al, 2021b) and others (Jackson et al, 2022; Muralidar et al, 2021). Our previous studies show that elevated CTSL levels correlate with disease severity and CTSL is crucial for activation of all emerging SARS-CoV-2 variants, making it a potential drug target for future mutation-resistant therapy (Zhao et al., 2021b; Zhao et al, 2022). However, the specific role of CTSL in COVID-19 infection among diabetic patients remains unexplored.

Patients with diabetes face heightened risks of severe COVID-19 outcomes and increased mortality rates (Khunti et al, 2021). Studies report a 1.23 to 5.87 times higher likelihood of severe COVID-19 and death among diabetic individuals compared to non-diabetic counterparts (Dennis et al, 2021; Shi et al, 2020; Williamson et al, 2020). Diabetes ranks as the second most prevalent chronic comorbidity contributing to COVID-19 fatalities, following hypertension, according to data from the American Centers for Disease Control and Prevention (CDC). Notably, within the diabetic cohort, individuals with elevated blood glucose levels (HbA1c ≥ 7.5%) exhibit a higher hazard ratio for adverse outcomes compared to those with lower glucose levels (HbA1c < 7.5%) (Williamson et al., 2020).

This study delves into the impact of high glucose levels on CTSL maturation and its implications for diabetic comorbidities, complications, and susceptibility to SARS-CoV-2 infection. Our findings unveil the role of high glucose in promoting CTSL maturation and translocation from the endoplasmic reticulum to the lysosome, potentially exacerbating diabetic complications and contributing to COVID-19 susceptibility among diabetic individuals.

Results

Diabetic COVID-19 patients have severe conditions and elevated CTSL

In COVID-19 patients, we conducted a case-control study to examine the association of diabetes and COVID-19 severity in 207 COVID-19 inpatients from two hospitals. We matched 62 patients by gender and age, 31 with diabetes and 31 without (Fig. 1a). Supplementary Table 1 summarizes the demographic and clinical characteristics of these diabetic and non-diabetic COVID-19 patients. We found that diabetic patients had a significantly higher risk of developing severe COVID-19 than non-diabetic patients according to the clinical classification criteria (http://www.nhc.gov.cn/), and had more symptoms such as fever, cough, fatigue, and dyspnea (Fig. 1b). Diabetic COVID-19 patients showed higher levels of inflammation and infection markers (Supplementary Table 1). These results suggest that diabetes is strongly associated with severity of COVID-19.

Fig. 1

To explore the mechanisms of hyperglycemia and SARS-CoV-2 infection, we collected plasma samples from Beijing Youan Hospital on Day 0 (the admission day), Day 14, and Day 28 after discharge from the hospital (Fig. 1c). SARS-CoV-2 infects host cells through the virus spike protein binding with ACE2 receptor. It uses host proteases, such as CTSL and cathepsin B (CTSB) to activate its spike protein by cleavage, which enhances its cell entry (Jackson et al., 2022; Muralidar et al., 2021). We measured the plasma levels of COVID-19 related proteins, ACE2, CTSL, and CTSB in diabetic and non-diabetic COVID-19 patients. Only CTSL levels were significantly higher in diabetic patients than in non-diabetic patients and changed with the course of COVID-19. CTSL peaked on admission day and decreased significantly after discharge from the hospital (Fig. 1d-f). These results indicate that CTSL is strongly associated with COVID-19, as previously reported (Zhao et al., 2022), and may be involved in diabetes in COVID-19 patients.

Impact of chronic and acute hyperglycemia on CTSL activity

In non-COVID-19 participants, we investigated the correlation of CTSL with chronic and acute hyperglycemia using two studies. First, to examine the impact of chronic hyperglycemia on CTSL, we performed a case-control study in 61 patients with type 2 diabetes and 61 euglycemic subjects, matched for sex and age (Supplementary Table 2). We found that plasma CTSL activity was strongly positively correlated with chronic hyperglycemia indicated by HbA1c, and was significantly higher in diabetic patients than in euglycemic individuals (Fig. 2a, c). Additionally, plasma CTSL concentration showed a positive trend with chronic hyperglycemia indicated by HbA1c (Fig. 2b, d). Second, to examine the impact of acute hyperglycemia on CTSL, we performed a hyperglycemic clamp study in 15 healthy male subjects (Fig. 2e-g). We observed that CTSL activity increased parallelly with blood glucose levels (Fig. 2h, i). However, CTSL concentration did not change with blood glucose levels (Fig. 2j). Therefore, chronic hyperglycemia is strongly associated with both CTSL concentration and activity, while acute hyperglycemia only affects CTSL activity.

Fig. 2

We also observed a strongly correlation between CTSL activity/concentration and other chronic comorbidities like hypertensin and coronary heart disease (CHD), except for diabetes (Supplementary Table 3). Additionally, elevated blood glucose levels also accompanied with an increase in insulin and proinsulin C-peptide levels in acute hyperglycemic individuals (Fig. 2f, g) and diabetic patients (insulin resistance) (Gerich, 2003). It is unclear whether the increased CTSL activity/concentration is solely a result of hyperglycemia or the corresponding hyperinsulinemia, and further clarification is needed.

Hyperglycemia enhances SARS-CoV-2 infection through CTSL

To validate our hypothesis that diabetic patients are more susceptible to SARS-CoV-2 infection results from higher CTSL levels, we conducted in vitro experiments using a SARS-CoV-2 pseudovirus system, which indicates the virus invasion rate instead of replication, and utilized the SARS-CoV-2 susceptible human hepatoma cell line Huh7, given that the liver is a primary target of SARS-CoV-2 (Gupta et al, 2020).

First, to examine if cells cultured with serum from diabetic patients are more susceptible to SARS-CoV-2 infection, we compared the infectivity of cells cultured with healthy and diabetic sera exposed to SARS-CoV-2 using a luciferase assay. The results showed that cells cultured with diabetic serum had higher infection rate than cultured with healthy serum (Fig. 3a), indicating diabetic serum facilitated SARS-CoV-2 entry into host cells.

Fig. 3

Next, we investigated whether elevated blood glucose or insulin level promoted SARS-CoV-2 infection by culturing Huh7 cells in cell media with various concentrations of glucose or insulin. Consistent with our clinical data, the results showed that infection was more severe in Huh7 cells at high glucose levels, while insulin levels had a minimal impact on SARS-CoV-2 infection (Fig. 3b, c).

Given that hyperglycemia increased CTSL levels (Fig. 2) and facilitated SARS-CoV-2 entry into host cells (Fig. 3b), we hypothesized that high blood glucose promoted SARS-CoV-2 infection through CTSL. To investigate the requirement of CTSL in SARS-CoV-2 infection, we used the CRISPR-Cas9 system to establish a stable CTSL knockout (KO) Huh7 cell line. The knockout efficiency of CTSL protein of CTSL KO Huh7 cell line was confirmed in Fig. 3d. Compared with wild-type (WT) cells, knockout of CTSL led to a significant reduction in SARS-CoV-2 infection (Fig. 3e), suggesting that CTSL is crucial for SARS-CoV-2 infection as we previously reported (Zhao et al., 2022).

We then conducted a CTSL KO Huh7 cell infection experiment under different glucose conditions, to illustrate the impact of glucose level on SARS-CoV-2 infection via CTSL. The results showed that WT Huh7 cell cultured in high glucose medium exhibited a much higher infective rate than those in low glucose medium. However, CTSL KO Huh7 cells maintained a low infective rate of SARS-CoV-2 regardless of glucose or insulin levels (Fig. 3f-h). Therefore, hyperglycemia enhanced SARS-CoV-2 infection through CTSL. Considering that CTSL realizes its proteolytic function through its enzyme activity and protein concentration, subsequent studies aimed to reveal whether the CTSL activity or concentration changed under high glucose condition.

Elevation of glucose level boosts CTSL activity

Hyperglycemia can lead to metabolic acidosis and alter blood pH. However, the normal range for blood pH in humans is relatively narrow, typically ranging from 7.35 to 7.45. In our study, blood pH remained within this normal range for both diabetic and healthy control samples. Fig. 4a demonstrates consistent CTSL activity despite pH variations. Then we conducted a series of experiments to investigate the impact of hyperglycemia on CTSL activity. Our findings showed that elevated glucose levels significantly stimulated both intracellular and extracellular CTSL activity in a dose-dependent manner in Huh7 cells (Fig. 4b, c), which was consistent with our clinical data (Fig. 2a, h). In contrast, insulin levels had no effect on CTSL activity in Huh7 cells (Fig. 4d, e), indicating that it was hyperglycemia, rather than hyperinsulinemia, that boosted CTSL activity in diabetic patients.

Fig. 4

We then evaluated CTSL activity in human and mouse biopsy tissues under high blood glucose condition. The demographic and clinical information for the human lung tissue samples donors in this study can be found in Supplementary Table 4. Diabetic mice (db/db mice) had significantly increased glucose, body weight and fasting insulin levels compared to control (db/m mice) group (Fig. 4f-h). We observed an elevation of CTSL activity in both diabetic mice and human lung tissues (Fig. 4i, j), suggesting that high glucose condition may increase CTSL activity in the respiratory system in diabetic patients and mice in vivo. The increase of CTSL activity in diabetic mice liver tissue (Fig. 4i) was consistent with the results of Huh7 hepatocytes (Fig. 4b).

High glucose level stimulates CTSL maturation

To investigate the impact of high blood glucose on CTSL expression, we first measured the mRNA levels of CTSL under different glucose concentrations using D-glucose and L-glucose (Supplementary Fig. 1). While D-glucose is commonly used as a major energy source, L-glucose cannot be absorbed by cells but has similar physical properties to D-glucose, making it an ideal control. Surprisingly, our results showed that glucose or insulin levels did not affect CTSL mRNA levels in Huh7 cells, mouse tissues or human tissues (Supplementary Fig. 1), indicating that glucose did not influence CTSL transcription.

CTSL undergoes several forms during its translation and post-translational maturation process. The immature pro-cathepsin L (proCTSL, 39 kDa) possesses an N-terminal proregion, which acts as an autoinhibitor. The protein is translocated through the endoplasmic reticulum (ER)-Golgi apparatus-lysosome axis, and the proregion is removed in the acidic environment in lysosome results in either single chain mature cathepsin L (sc-mCTSL, 31 kDa) or double chain mature cathepsin L (dc-mCTSL, 24 kDa) (Fig. 5a) (Coutinho et al, 2012; Ishidoh & Kominami, 2002; Reiser et al, 2010). Only mature CTSL in lysosome has catalytic activity against specific substrates, while proCTSL does not (Ishidoh & Kominami, 2002).

Fig. 5

Interestingly, we found that high glucose levels promoted CTSL maturation, whereas insulin had no effect on this process, as shown in Fig. 5b. Our results indicated that high D-glucose levels reduced proCTSL and increased sc-mCTSL and dc-mCTSL in a glucose dose-dependent manner (Fig. 5b, Supplementary Fig. 2). The results also confirmed that only D-glucose induced CTSL maturation from proCTSL to mCTSL, while L-glucose had no such effects (Fig. 5b, c). Similar effects of high glucose on CTSL maturation were also observed in diabetic mice and human tissues compared with their healthy counterparts (Fig. 5d, e). The differences observed in the processing of CTSL between cells (Fig. 5b) and tissues (Fig. 5d-e) may be attributed to the complexities inherent in tissue samples, which can impact the clarity of the images. Furthermore, in human tissue samples, it is pertinent to consider that patients in the diabetes group had their blood glucose levels controlled within or near the normal range prior to lung surgery. As a result, the evidence supporting CTSL maturation in human lung tissue blotting images may be less compelling. These results suggested that hyperglycemia rather than hyperinsulinemia or other physical parameters associated with CTSL maturation. This was consistent with our previous data that elevated glucose levels enhanced CTSL activity (Fig. 4b), since only mature CTSL has enzymatic activity.

High glucose facilitates CTSL maturation via translocation

We have shown that high glucose promoted CTSL maturation by converting proCTSL into mCTSL (Fig. 5, Supplementary Fig. 2). However, the mechanism underlying this process remained unclear. As mentioned previously, CTSL maturation depends on structural activation via the ER-Golgi apparatus-lysosome axis and acid activation at low pH environment (Fig. 6a) (Ishidoh & Kominami, 2002; Reiser et al., 2010). Only the CTSL in lysosome is processed into mature form and has proteolytic activity, while the CTSL in the ER or Golgi apparatus is immature and does not have enzymatic activity. Based on this, we hypothesized that high blood glucose may drive CTSL maturation via the ER-Golgi apparatus-lysosome axis. We labeled specific proteins in each organelle: calreticulin for ER, GM130 for Golgi apparatus and lamp1 for lysosome (Supplementary Fig. 3). We also confirmed CTSL expression in Huh7 cells (Supplementary Fig. 4). Confocal microscopy results revealed that under low glucose conditions, CTSL tended to co-localize within the ER rather than lysosomes (Fig. 6b, e). This finding suggests that CTSL primarily existed in an immature form under low glucose conditions. As glucose level increased, CTSL in the Golgi apparatus remained unchanged (Fig. 6c, f), while co-localized largely increased in the lysosomes (Fig. 6d, g), suggesting CTSL predominantly existed in mature form under high glucose conditions. Therefore, we concluded that high glucose facilitated CTSL translocation through the ER-Golgi apparatus-lysosome axis.

Fig. 6

Discussion

Almost immediately after the SARS-CoV-2 emerged, it became apparent that individuals with chronic conditions, including diabetes, were disproportionately affected, with a heightened risk of hospitalization and mortality (Williamson et al., 2020). Diabetes may increase susceptibility to severe SARS-CoV-2 infections for various suggested reasons. These include higher viral titer, relatively low functioning T lymphocytes that lead to decreased viral clearance, vulnerability to hyperinflammation and cytokine storm syndrome, and comorbidities associated with type 2 diabetes, such as cardiovascular disease, non-alcoholic fatty liver disease, hypertension, and obesity (Mazucanti & Egan, 2020; Moradi-Marjaneh et al, 2021). Additionally, other risk factors that may contribute to the severity of infection include increased expression of angiotensin-converting enzyme 2 (ACE2) (Rao et al, 2020) and furin (Fernandez et al, 2018). However, most current studies on COVID-19 and diabetes focus on epidemiological evidence and biomarker features, but few investigate the causal link and underlying mechanisms of how hyperglycemia enhances SARS-CoV-2 infection, confirmed by human body fluids, biopsies, and animal models. The underlying mechanisms by which diabetes or hyperglycemia exacerbates COVID-19 remain to be fully elucidated.

This study identified that CTSL maturation induced by hyperglycemia may contribute to the higher mortality and severity of COVID-19 in patients with diabetes. While our initial study involved 62 COVID-19 patients, with 31 having diabetes and 31 without, matching based on gender and age, we acknowledged the challenge of obtaining balanced gender distribution in both groups due to the difficulty of collecting blood samples from COVID-19 patients. To mitigate potential gender bias resulting from small sample sizes, we conducted a supplementary clinical study involving 122 non-COVID-19 volunteers, including 61 individuals with diabetes and 61 without. The percentage of males in the diabetes group was 50.8%, while in the healthy group, males constituted 44.3% (P value = 0.468), indicating no significant gender bias. The clinical data from this study indicated that human plasma CTSL activity and concentration were correlated with acute (in euglycemic participants under high glucose clamp conditions) and chronic (in diabetic patients, both with and without COVID-19) hyperglycemia, respectively.

Using lung tissue samples from diabetic and non-diabetic patients, as well as db/db diabetic and control mice, we found that diabetic conditions increased CTSL activity in both humans and mice. The liver is a significant target organ for COVID-19 (Gupta et al., 2020). Despite potential limitations, such as generalization of liver function abnormalities and lack of tissue specificity in SARS-CoV-2 impact, Huh7 cells offer practical advantages as a mature cell model for studying SARS-CoV-2 infection, including accessibility, susceptibility to infection, and stable proliferation(Nie et al, 2021) (Ciotti et al, 2020). Taking all these factors into consideration, we have ultimately chosen to utilize the hepatoma cell line to investigate how hyperglycemia induces CTSL maturation and subsequently promotes SARS-CoV-2 infection. High glucose promoted SARS-CoV-2 infection in WT cells, while CTSL KO cells showed reduced susceptibility to high glucose promoting effects. Mechanistically, we proposed that hyperglycemia promoted CTSL maturation by accelerating its translocation from the ER to lysosome via Golgi apparatus. This condition increased the functionality of CTSL, which cleaved the spike protein of SARS-CoV-2, promoting virus membrane fusion and infection (Fig. 6h). In our study, the term “diabetes” encompasses the condition of hyperglycemia in a broad sense, rather than specifically indicating type 1 diabetes (T1DM) or type 2 diabetes (T2DM). This broader definition aligns with the scope of our research objectives and findings, particularly observed in the cell experiments conducted.

Fig. 2h-j illustrate the impact of acute hyperglycemia on CTSL concentration and activity in 15 healthy male volunteers over a 160-minute period. During this short timeframe, CTSL concentration remained stable, as evidenced by consistent RNA results from cells exposed to varying glucose levels (Supplementary Fig.1). However, there was a significant increase in CTSL activity, indicating that glucose elevation rapidly triggers CTSL maturation through propeptide cleavage. This activation process occurs more rapidly than CTSL protein synthesis. In summary, acute hyperglycemia specifically elevates CTSL activity, while chronic hyperglycemia may impact both CTSL activity and concentration (Fig. 2a-d). To date, there have been limited studies investigating the relationship between cathepsin maturation and glucose. In 1998, Tournu C, et al. reported that D-glucose did not impact mRNA levels for CTSB or CTSL or secretion of proCTSL. However, D-glucose did significantly enhance the amount of mature forms of CTSB and CTSL (Tournu et al, 1998). More recently, Shi Q, et al. found that increased glucose metabolism promotes O-GlcNAcylation of the lysosome-encapsulated protease CTSB, leading to elevated levels of mature CTSB in macrophages and secretion in the tumor microenvironment (Shi et al, 2022). These findings support our evidence that hyperglycemia drives CTSL maturation.

ACE2 has previously been identified as a critical host cell surface receptor that enables SARS-CoV-2 entry into host cells (Wrapp et al, 2020). While some studies have reported that glucose can increase ACE2 expression in cell lines (Hardtner et al, 2013), numerous other studies have found that ACE2 is downregulated in diabetic patients (Mizuiri et al, 2008; Reich et al, 2008). Garreta et al. recently conducted a study using a human kidney organoid system to investigate the impact of diabetes on SARS-CoV-2 infections. The study revealed that hyperglycemia enhanced SARS-CoV-2 infection and hyperglycemic human kidney organoids had elevated ACE2 levels (Garreta et al, 2022). Therefore, it remains controversial whether diabetes results in up- or downregulation of ACE2. In our study, we evaluated plasma levels of ACE2, CTSL, and CTSB in COVID-19 patients with and without diabetes. We found that only CTSL levels were significantly increased in diabetic patients compared to non-diabetic patients and varied during the course of COVID-19.

In addition to CTSL, there may be other bioactive factors involved in mediating SARS-CoV-2 infection in patients with diabetes. A recent study revealed that diabetic patients have lower levels of serum 1,5-anhydro-D-glucitol (1,5-AG), a small-molecule metabolite in human blood that exhibits potent antiviral activity against SARS-CoV-2. The reduced levels of 1,5-AG have been associated with increased viral loads and severe respiratory tissue damage caused by SARS-CoV-2. Mechanistically, the study found that 1,5-AG binds directly to the S2 subunit of the spike protein, which disrupts virus-host membrane fusion and inhibits infection (Tong et al, 2022). Therefore, we propose that diabetes may promote COVID-19 infection through multiple factors, and CTSL is only one of several important factors.

Apart from diabetes, other comorbidities such as hypertension and CHD are also prevalent in COVID-19 patients. Interestingly, our study revealed a strong correlation between CTSL activity and concentration with hypertension and CHD in these patients. Using an angiotensin II-induced hypertension model, researchers observed an increase in blood pressure and CTSL activity (Lu et al, 2020). Whether these chronic comorbidities contribute to increased morbidity and mortality of COVID-19 by increasing CTSL activity and concentration requires further investigation in the future.

In conclusion, our study demonstrates that hyperglycemia drives the maturation and activation of CTSL, for only mature form of CTSL gains its function of proteolysis. Therefore, targeting CTSL may be a promising therapeutic strategy for diabetic comorbidities and complications(Hua Li, 2022).

Materials and methods

Key resources table

Experimental model and study participant details

Patients and clinical samples

The study protocol was approved by the Ethics Committee of Beijing Tongren Hospital, Capital Medical University (TRECKY2020-013, TRECKY2021-202). The retrospective cohort study included 207 COVID-19 patients from two hospitals. 120 adult COVID-19 inpatients admitted to Wuhan Union Hospital, Huazhong University of Science and Technology (Wuhan, China) between January 29 and March 20, 2020. Another 87 consecutive COVID-19 inpatients were hospitalized at Beijing Youan Hospital, Capital Medical University (Beijing, China) between January 21 and April 30, 2020. 31 COVID-19 patients with diabetes and 31 COVID-19 patients without diabetes were matched for gender and age and included in the final analysis. The clinical features are presented in Supplementary Table 1. SARS-CoV-2 was detected in respiratory specimens using real-time RT-PCR, following the protocol recommended by the World Health Organization.

COVID-19 was classified into four categories: mild, moderate, severe and critical, according to the clinical classification criteria (http://www.nhc.gov.cn/). Patients from Beijing Youan Hospital, Capital Medical University were further followed up. They experienced a mean of 14 days of hospitalization and were followed up on the 14th day (Day 14) and 28th day (Day 28) after discharge from the hospital. Blood samples were collected shortly after the admission to the hospital (Day 0) and on Day 14 and Day 28. Demographic, clinical, and laboratory data were extracted from the electronic hospital information system using a standardized form.

Another total of 122 age- and gender-matched diabetic and non-diabetic volunteers without COVID-19 were recruited in Beijing Tongren Hospital, Capital Medical University. Blood samples were collected after overnight fasting for the determination of CTSL activity and concentration and other biochemical parameters. All biochemical measurements have participated in the Chinese Ministry of Health Quality Assessment Program. The demographic and clinical characteristics are shown in Supplementary Table 2.

Human lung tissue samples were obtained from 6 patients who underwent lung surgery at Beijing Tongren Hospital between March 22 and June 22, 2022. The baseline characteristics are presented in Supplementary Table 4.

The plasma samples of hyperglycemic clamp study were from a previously conducted clinical trial (NCT03972215). Fifteen healthy male research subjects were received a 160-min hyperglycemic clamp study with a baseline blood glucose level + 6.9 mmol/L as the target level. Blood samples were obtained at intervals throughout the clamp study. The plasma was collected and stored at −80℃ until use.

Experimental mice

The study used 10-week-old db/db mice as diabetic mode and db/m mice as their healthy control, maintained on a KBS background. All mice were obtained from Vital River Laboratories (Beijing, China). The mice were housed at constant humidity and temperature, with a 12h light/dark cycle. The protocols for the use of mice were approved by the Ethical Review Committee at the Institute of Zoology, Capital Medical University.

Cell lines and reagents

The Huh7 (Homo sapiens, liver) cell line (Cell Resource Center, Chinese Academy of Medical Sciences, Beijing, China), was maintained in high glucose Dulbecco’s modified Eagle’s medium (DMEM) (Sigma-Aldrich, St. Louis, MO, USA) supplemented with streptomycin (100 mg/ml), penicillin (100 units/ml), and fetal bovine serum (10%, Gibco, Carlsbad, CA). The cells were maintained at 37°C in a humidified atmosphere of 5% CO2 and 95% air.

Method details

Detection of SARS-CoV-2 entry related host biomarkers

Plasma samples of patients with COVID-19 at Day 0, Day 14, and Day 28 were collected and stored at −80 °C within 2 h. The samples were analyzed using commercially available enzyme-linked immunosorbent assays (ELISA) following the manufacturer’s instructions. All samples were detected without virus inactivation to retain the original results in a P2 + biosafety laboratory. ACE2 was measured using the Human ACE2 Elisa kit (Cloud-Clone Corp, Cat. No L220216094). CTSL and CTSB were measured using the Human CTSL ELISA Kit (Elabscience, Cat. No E-EL-H0671) and Human CTSB ELISA kit (Elabscience, Cat. No E-EL-H6151). The kits were designed for usage with human serum or plasma samples and showed no cross-reactions.

Production of pseudovirus

The SARS-CoV-2 pseudovirus were generated with the incorporation of SARS-CoV-2 spike protein (SARS-2-S) into vesicular stomatitis virus (VSV)-based pseudovirus system. The pseudoviruses used in the current study have been validated in previous studies (Lv et al, 2020; Whitt, 2010). For this VSV-based pseudovirus system, the backbone was provided by VSV-G pseudotyped virus (G*ΔG-VSV) that packages expression cassettes for firefly luciferase instead of VSV-G in the VSV genome (Nie et al, 2020). Therefore, the luciferase activity of VSV phosphoprotein (VSV-P) were used for indicators of pseudovirus infection.

Pseudovirus infection in vitro

Huh7 cells were plated in 96-well plates and allowed to adhere until they reached 70% confluency. Subsequently, these cells were cultured with different medium or serum obtained from diabetic or euglycemic individuals as indicated. Following this, the cells were infected with SARS-CoV-2 pseudovirus at 1.3 × 10⁴ TCID₅₀/ml at 37 °C. After coincubation with pseudovirus of 24 hours, the activities of firefly luciferases were measured on cell lysates using luciferase substrate, Britelite Plus Kit (Perkinelmer, Cat. No 6066761) according to the manufacturer’s instructions. The firefly luciferase activity was measured rapidly using a luminometer (Turner BioSystems, USA) as described previously (Yang et al, 2017). Cell viability were measured by CCK assay using Cell Counting Kit (Transgen, Cat. No FC101-04). The infection rates were adjusted by cell viability.

Establishment of CTSL-KO Huh7 cell line via CRISPR/Cas9

To produce the CTSL-KO cell line, we utilized CRISPR/Cas9 technology. The single guide RNA (sgRNA) was designed using the Zhang Lab Guide Design Resources (https://zlab.bio/guide-design-resources) tool. The sgRNA scaffold was commercially obtained from Sangon, with the sequence designed as 5’-CTTTGTGGACATCCCTAAGC-3’. For sgRNA and Cas9 protein to enter into Huh7 cell, electroporation-mediated transfection was performed. First, 4 × 10⁵ Huh7 cells were centrifuged and re-suspended in 10 μL of Buffer R following the manufacturer’s instructions (Invitrogen, USA) for electroporation. Next, Cas9 protein (1 μg) and sgRNA (0.2 μg) were added to each sample and mixed gently (Cas9 protein: sgRNA at a 1:1 molar ratio). Huh7 cells were electroporated for 5 times to minimize the CTSL gene on a Neon Transfection device (Invitrogen, USA).

Generation of CTSL-KO monoclonal Huh7 cell line

Cells were isolated from the stable CTSL-KO Huh7 cell pool by trypsinization and any cell clumps were broken up into individual cells. Cells concentration was quantitated in this homogenized cell solution with a cell counter. Then, 100 µL of the 5 cells/mL solution was transferred into each well of a 96-well plate. By doing this, the average density of 0.5 cells/well of the plate was seeded. Seeding an average of 0.5 cells/well ensured that some wells received a single cell, while minimizing the likelihood that any well receives more than one cell. Then we observed and recorded the cell growth in the plate for the following 30 days. Once the cells have expanded but before they become over-confluent, we trypsinized the cells and expanded them to larger culture dishes.

RNA extraction and quantitative real-time PCR analyses

Total RNA was extracted and purified from the cultured Huh7 cells, human and mouse tissues using RNAprep Pure Cell/ Bacteria kit (Tiangen, Cat. No DP430) and RNAprep Pure Tissue kit (Tiangen, Cat. No DP431) according to the manufacturer’s instructions. RNA (0.5 μg) was reverse transcribed to cDNA in a final volume (20 μL) using TransScript First-Strand cDNA Synthesis SuperMix (Transgen, Cat. No AT301). RT-PCR analyses were performed with TransStart Tip Green qPCR SuperMix (Transgen, Cat. No AQ141). Gene expression values were normalized to the control (β-actin) level. Supplementary Table 5 provides a list of all primer sequences. Quantitative real-time PCR (qRT-PCR) and data collection were done on a LightCycler 96 system (Roche, Switzerland).

Intraperitoneal glucose tolerance test (IPGTT)

The db/db and db/m mice were subjected to an overnight fast lasting 16 hours, during which they were permitted unrestricted access to water. Subsequently, they received an intraperitoneal injection of 1 g/kg body weight glucose. Blood samples were obtained at 0, 15-, 30-, 60-, and 120-minutes post-glucose injection. Blood glucose levels were determined using an automatic glucometer (One Touch, LifeScan, USA), while insulin concentrations were evaluated using a highly sensitive mouse insulin ELISA kit (Millipore, Cat. No EZRMI-13K), according to the manufacturer’s instructions.

Western blot analysis

Total protein was extracted from Huh7 cells and human and mouse tissues. The protein amount was assessed using the BCA protein assay kit (Thermo, Cat. No WH333441). Samples of 30-50 µg of protein were separated by SDS-PAGE, transferred to PVDF membrane (Millipore, Cat. No 0000167358), and detected using enhanced chemiluminescent reaction (Zhao et al, 2021a). The antibody information was summarized in Supplementary Table 6.

Analysis of CTSL activity

The activity of CTSL in plasma, human and mouse tissues, Huh7 cells and cell medium was evaluated using its specific substrate, Ac-FR-AFC (R&D, Cat. No ES009). Prior to measurement, the cell medium was concentrated using the ultra-tube (Milipore, Cat. No UFC501096). The test samples were evaluated in the presence of a reaction buffer (100mM NaAc, 5mM EDTA, pH 5.3). The reaction was conducted in 100 µL system (50µL sample containing CTSL protein+ 47µL reaction buffer + 1 µL DTT (1 mM) + 2 µL substrate Ac-FR-AFC (10 mM)) in 96-well black plates. The plated was cultured at 37℃ for 2 h in light avoidance incubator. The fluorescence emitted from the samples was then measured using a fluorescence plate reader (Infinite 200, TECAN, China) at the Ex = 380 nm, Em = 460 nm wavelengths.

Immunofluorescence assay

The CTSL, calreticulin, Golgi membrane protein 130 (GM130) and lysosomal associated membrane protein 1 (Lamp 1) distributions in the Huh7 cells were visualized by immunofluorescent staining. The CTSL, calreticulin, GM130 and Lamp1 antibody species source IgG were used as negative control (Supplementary Fig. 3). Briefly, Huh7 cells was cultured in 35mm confocal dishes after poly-l-lysine coating. After high/low glucose treatment for 96 h, Huh7 cells were then fixed by 4% paraformaldehyde (PFA) and permeabilized by a detergent 0.25% triton X-100. A specific primary antibody is applied on the Huh7 cell surface at 4 ℃ overnight. After wash out, the secondary antibody is applied at room temperature for 1 hour avoid from light. All pictures were captured under Laser Scanning Confocal Microscopy (FV3000RS, Olympus, Japan). The antibodies information was summarized in Supplementary Table 6.

Quantification and statistical analysis

Clinical data are shown as percentage or median, as appropriate. Comparison of continuous data between two independent groups was performed using the Mann–Whitney U-test. An unpaired t test was used for comparing the averages/means of two independent or unrelated groups. A paired t-test was used to test whether the mean difference between pairs of measurements is existing. Analysis of variance (ANOVA) was used for checking if the means of two or more categories are significantly different from each other. Spearman’s rho test (two-tailed) was used to analyze nonparametric correlations of parameters correlated with CTSL levels and diabetes. Fluorescence intensity was calculated by Plot Profile tool in Image J software. Graphpad prism 7.0 software and SPSS for Windows 17.0 were used for statistical analysis, with statistical significance set at two-sided. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Acknowledgements

We thank participants and staff of the case-control studies for their valuable contributions. This work was supported by grants from National Natural Science Foundation of China (82341076; 81930019). This work was also supported by grants from National Natural Science Foundation of China (82300917), Beijing Municipal Administration of Hospitals Incubating Program (PX20240203) and Foundation of Beijing Tongren Hospital to the Outstanding Youths to M.M.Z.

Author contributions

J.K.Y. conceived the idea for the study, designed the experiments and wrote the final manuscript. Q.H. and M.M.Z. designed and performed the experiments and wrote the first draft of the manuscript. M.J.L. and X.Y.L. performed pseudovirus assays and immunofluorescence assays. J.M.J., Y.M.F., F.Y.Y. and M.M.Z. collected the clinical data. L.Z. and W.J.H. partially participated the pseudovirus assays.

Lead Contact

The data that support the findings of this study are available from the Leading Contact, Dr. Jin-Kui Yang (jkyang@ccmu.edu.cn).

Conflict of interests

The authors declare no competing interests.

Materials Availability

All the unique reagents generated in this study are available from the Lead Contact with a completed Materials Transfer Agreement.

Supplementary information

Figure S1 to S4

Table S1 to S6