New Users of Sodium-Glucose Cotransporter 2 Inhibitors Are at Low Risk of Prostate Cancer: A Nationwide Cohort Study
Article information
Abstract
Background
Preclinical studies have reported anticancer properties of sodium-glucose cotransporter 2 inhibitors (SGLT2is). We aimed to elucidate the association between the use of SGLT2is and the risk of prostate cancer among male patients with type 2 diabetes mellitus (T2DM).
Methods
An active-comparator, new-user cohort design using a nationwide database between September 2014 and June 2020 was conducted on 45,601 new SGLT2i users and 205,395 new users of other glucose-lowering medications (oGLMs). In the following 1:1 propensity score matched (PSM) analysis, 35,371 SGLT2i users matched with an equivalent number of oGLM users were assessed. The hazard ratios (HRs) and 95% confidence intervals (CIs) for prostate cancer were calculated.
Results
Among the cohort, prostate cancer was diagnosed in 210 out of 45,601 SGLT2i users, corresponding to a cumulative incidence of 1.0%, in contrast to 1,880 cases among 205,395 users of oGLMs, with a cumulative incidence of 1.5%. The use of SGLT2is was significantly correlated with a reduced risk of prostate cancer based on a multivariable-adjusted HR of 0.83 (95% CI, 0.71 to 0.98). PSM analysis affirmed 18% reduction in prostate cancer risk associated with SGLT2i use (HR, 0.82; 95% CI, 0.67 to 0.99). Subgroup analyses revealed that body mass index (BMI) significantly influenced the effect of SGLT2i on prostate cancer risk, with a more pronounced reduction in the subgroup with a BMI <25 kg/m2 (P=0.037).
Conclusion
The use of SGLT2is in Korean male patients with T2DM is associated with a lower risk of prostate cancer.
Highlights
• This propensity score-matched analysis included Korean men with T2DM aged 50–84 years.
• Use of SGLT2 inhibitors was associated with a reduced risk of prostate cancer.
• SGLT2 inhibitor use was also linked to lower all-cause mortality during follow-up.
• The reduction in prostate cancer risk was more pronounced in nonobese individuals.
• These findings support a potential anticancer benefit of SGLT2 inhibitors in T2DM.
INTRODUCTION
Sodium-glucose cotransporter 2 inhibitors (SGLT2is) are oral antidiabetic medications primarily prescribed for the management of type 2 diabetes mellitus (T2DM). SGLT2i targets SGLT2, the main protein responsible for glucose reabsorption in the proximal tubules of the kidneys, thereby reducing glucose reabsorption and increasing glucose excretion via urine [1]. Beyond their hypoglycemic effects, SGLT2is exhibit significant therapeutic potential in managing cardiovascular diseases and kidney outcomes. As a result, SGLT2is are currently approved for the treatment of heart failure and chronic kidney disease. The excellent safety profile of SGLT2is in humans has spurred interest in their potential as treatments for various non-metabolic disorders.
Recently, the antitumor effects of SGLT2is have garnered increasing interest. Although the primary mechanism of SGLT2is involves the inhibition of glucose uptake, SGLT2is also leverage a multifaceted approach in combating cancer. This approach includes inducing mitochondrial membrane instability, inhibiting the β-catenin and phosphoinositide 3-kinase (PI3K)-Akt pathways, promoting cell cycle arrest and apoptosis, and downregulating oxidative phosphorylation [2]. These findings suggest that SGLT2is have diverse molecular actions and may exhibit anticancer effects, extending beyond their conventional role in diabetes management. The use of SGLT2is has been associated with lower risks of all-cause mortality, cancer-related mortality, and new overall cancers [3]. In addition, the anticancer activity of SGLT2is in various types of cancers, including hepatocellular, pancreatic, colon, lung, and breast carcinomas, has been revealed [3].
Prostate cancer predominantly affects middle-aged men, particularly those between 45 and 60 years old. Prostate cancer is the leading cause of cancer-related deaths in Western countries [4,5]. The incidence of prostate cancer has significantly increased in South Korea. This increase has been attributed to the increase in life expectancy, the adoption of Western dietary patterns, and increased awareness of prostate cancer screening. Notably, the number of prostate cancer cases has increased from 1,455 in 2000 to 16,815 in 2020, which indicates an approximately 11-fold increase [6,7]. According to recent preclinical studies, SGLT2is could be employed as therapeutic agents for prostate cancer [8-10]. However, the effect of SGLT2i use on the incidence of prostate cancer in patients with T2DM has not been determined.
The present study was performed to determine whether the use of SGLT2is is associated with a decreased risk of prostate cancer in patients with T2DM. This investigation was conducted using data from a comprehensive, nationwide, population-based cohort. To address potential methodological biases inherent in this observational study, we compared the incidence and risk of prostate cancer among new users of SGLT2is and other glucose-lowering medications (oGLMs) and validated the results using propensity score matching (PSM) to achieve balance among various confounding factors.
METHODS
Data source and study population
An active-comparator, new-user cohort design was employed in this study and data were obtained from the National Health Insurance Service (NHIS), a representative population-based sample cohort from South Korea [11]. A comprehensive description of the study cohort is available in a prior publication, which provides details of the NHIS study database [11]. Initially, we enrolled male patients aged 50 to 84 years with T2DM who began using either SGLT2i or oGLMs from September 2014 to June 2020. Patients undergoing dialysis (n=2,756) and individuals previously diagnosed with any cancer (n=27,866) were excluded from this group. The final selection comprised 45,601 new SGLT2i users and 205,395 new oGLM users. ‘New use’ was defined as the initiation of use of either drug category with no prior usage of these drugs in recent years and continued intake of these drugs for at least 180 days. The initial prescribed drug was termed the ‘index drug,’ and its first prescription date was the ‘index date.’ The SGLT2is included dapagliflozin, empagliflozin, ipragliflozin, and ertugliflozin.
Using propensity scores, we matched each new SGLT2i user with a new oGLM user at a 1:1 ratio. The study group comprised 35,371 SGLT2i participants while the control group comprised 35,371 oGLM participants. The inaugural claim date for either SGLT2i or oGLM was termed the ‘drug index date.’ Supplementary Fig. 1 depicts the study enrolment flowchart.
Measurements and definitions of covariates
Standardized methods were employed for measuring height, weight, waist circumference, and blood pressure (BP) during routine medical check-ups. BP was measured thrice; however, the average of the latter two readings was used. Body mass index (BMI) was calculated as weight in kg divided by height in m2. Blood specimens for measuring fasting plasma glucose (FPG), hemoglobin, total cholesterol, high-density lipoprotein cholesterol (HDL-C), and triglyceride levels were collected after an overnight fast. Estimates of estimated glomerular filtration rate (eGFR) were determined using the abbreviated Modification of Diet in Renal Disease (MDRD) equation. Patients were classified as nonsmokers, ex-smokers, or current smokers. Regular exercise was defined as strenuous physical activity more than three times weekly or moderate activity at least five times weekly. Heavy drinkers were defined as those consuming seven or more drinks per occasion for >5 days weekly; mild and moderate drinkers consumed less than seven drinks daily and drank 1–4 days weekly.
Patients prescribed antidiabetic drugs and assigned International Classification of Diseases, 10th Revision (ICD-10) codes E11 (noninsulin-dependent diabetes mellitus), E12 (malnutrition-related diabetes mellitus), E13 (other specified diabetes mellitus), and E14 (unspecified diabetes mellitus) as either a principal or secondary diagnosis were considered to have T2DM. During the performance of this study, nine classes of antidiabetic drugs were distributed by pharmacies in Korea: SGLT2i, sulfonylureas, biguanides, α-glucosidase inhibitors, thiazolidinediones, meglitinide, glucagon-like peptide-1 receptor agonists, dipeptidyl peptidase 4 (DPP-4) inhibitors, and insulin. Hypertension was defined according to a prescription of antihypertensive medications with ICD-10 codes I10 (essential [primary] hypertension), I11 (hypertensive heart disease), I12 (hypertensive chronic kidney disease), I13 (hypertensive heart and chronic kidney disease), and I15 (secondary hypertension) assigned as either a principal or secondary diagnosis. Dyslipidemia was defined according to the prescription of lipid-lowering medication based on a diagnosis with the ICD-10 code E78 (disorders of lipoprotein metabolism and other lipidemias) as either a principal or secondary diagnosis.
Outcome measures
The primary outcome of this study was prostate cancer, as indicated by ICD-10 code C61 (malignant neoplasm of prostate). In Korea, registration with the NHIS is mandatory for malignancies and rare incurable diseases, as patients receive financial assistance for these conditions. As a result, a rigorous definition of cancer is needed based on ICD-10 codes, ensuring accuracy and minimizing potential bias from misclassification. The follow-up for each patient began from the index date (start of SGLT2i or oGLM) and lasted until the onset of prostate cancer, death, or study conclusion (December 31, 2020), whichever occurred first.
Statistical analyses
We adopted an intention-to-treat exposure principle, wherein patients were regarded as continuously exposed to the study drug from the cohort’s inception until its conclusion. To offset treatment selection bias and potential confounding, we employed an active-comparator, new-user cohort approach.
The primary objective of this study was to determine whether SGLT2is are associated with a lower risk of prostate cancer. In the primary analysis, the relative effect of SGLT2i was quantified as crude and multivariable-adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) using Cox regression models. Thereafter, various sensitivity analyses were performed using the PSM cohort. First, the PSM method was employed to mitigate confounding caused by variations in the distribution of the measured covariates between individuals with and without SGLT2i use. These methods approximate the structure of a randomized controlled trial by minimizing the influence of confounding factors, thereby enabling a comparison of the outcomes across groups that are balanced in terms of measured baseline characteristics, except for the exposure. The selection of covariates for the propensity score model was guided by their potential association with the outcomes and included age, index year, BMI, systolic BP, HDL-C, low-density lipoprotein cholesterol, triglyceride, hemoglobin, FPG, alanine aminotransferase, creatinine, eGFR category, heavy alcohol drinking, current smoking, regular physical activity, comorbidities within the previous year, oGLMs prescribed within the previous year, and the number of other concomitant glucose-lowering medications. All variables are comprehensively listed in Supplementary Table 1. Based on the estimated propensity score, patients were matched 1:1 using greedy nearest neighbor matching without replacement, within a caliper of 0.2 standard deviations (SDs) of the logit of the propensity score. Further exploratory subgroup analyses were performed to discern treatment effect variations within the PSM cohort. An interaction term gauged treatment heterogeneity according to subgroup status.
Two-sided P<0.05 was considered to indicate statistical significance. The PSM was performed using the SAS Enterprise Guide software version 7.1 (SAS Institute, Inc., Cary, NC, USA). All other analyses were conducted using R statistical software version 4.1.3 (R Foundation for Statistical Computing, Vienna, Austria).
IRB approval
This study adhered to the tenets of the Declaration of Helsinki and was approved by the Institutional Review Board of Asan Medical Center (protocol no. 2021-1178). The board waived the requirement for written informed consent.
Data availability
The data that support the findings of this study are available from the authors but restrictions apply to the availability of these data, which were used under permission from the National Health Insurance Sharing Service (NHISS) for the current study, and so are not publicly available. Data are, however, available from the authors upon reasonable request and with permission from the NHISS (https://nhiss.nhis.or.kr).
RESULTS
Study population
A total of 45,601 new users of SGLT2i and 205,395 new users of oGLM were analyzed (Table 1, Supplementary Fig. 1). The SGLT2i users were younger, with a mean age (±SD) of 61.7±7.7 years, than the oGLM users (64.4±8.6 years; P<0.001). SGLT2i users also had significantly higher mean BMI, waist circumference, and FPG levels than oGLM users. Patients in the SGLT2i group also had lower mean creatinine levels and better mean eGFR than those in the oGLM group, including a greater proportion of individuals in the normal eGFR range. Regarding medical history, SGLT2i users had a higher incidence of cardiovascular events, including myocardial infarction, congestive heart failure, and peripheral vascular disease, and a lower incidence of cerebrovascular disease and dementia. Hypertension and dyslipidemia were more prevalent among SGLT2i users than oGLM users. Furthermore, in the year preceding the index date, SGLT2i users were more likely to have used insulin, DPP-4 inhibitors, and sulfonylureas. All observed differences between the groups were statistically significant.
Baseline characteristics of new users of SGLT2is and users of other GLMs
Following the 1:1 PSM analysis, a balanced cohort of 70,742 new users was created, which comprised 35,371 individuals from each group (Supplementary Table 1). Post-matching, the baseline characteristics were not found to significantly differ between the two groups (Supplementary Table 1).
Incidence of prostate cancer and mortality in the unmatched and matched cohort
Among the entire cohort, 210 of the 45,601 SGLT2i users (cumulative incidence of 1.0%) were diagnosed with prostate cancer during the follow-up period compared to 1,880 of the 205,395 oGLM users (cumulative incidence of 1.5%) diagnosed with prostate cancer (Table 2). Regarding all-cause mortality, 1,026 SGLT2i users (cumulative incidence of 6.1%) died while 14,240 oGLM users (11.9%) died. Patients taking SGLT2i had lower cumulative incidences of pancreatic cancer and overall mortality than those taking oGLM (Table 2). Univariate analyses indicated a HR for prostate cancer risk of 0.69 (95% CI, 0.60 to 0.79). The multivariable-adjusted HR for prostate cancer risk was 0.83 (95% CI, 0.71 to 0.98) in the unmatched cohort (Table 3). The all-cause mortality in the SGLT2i group was also significantly lower than that in the oGLM group, with an unadjusted HR of 0.48 (95% CI, 0.45 to 0.51) and a multivariable-adjusted HR of 0.75 (95% CI, 0.70 to 0.80) (Table 3).
Cumulative incidence of prostate cancer and all-cause mortality in the unmatched cohort and PSM cohort
Univariate and multivariable analyses of the risk of prostate cancer and all-cause mortality in sodium-glucose cotransporter 2 inhibitors users, with nonusers as the reference group
To validate the association between SGLT2i use and reduced incidence of prostate cancer, we conducted a sensitivity analysis within a PSM cohort. In this PSM cohort, SGLT2i users consistently had a reduced incidence of prostate cancer and all-cause mortality (Table 2). In particular, 178 SGLT2i users (1.0%) and 211 oGLM users (1.2%) were diagnosed with prostate cancer over a mean follow-up period of 2.2 years, further reinforcing the initial findings. The PSM HR for prostate cancer risk was 0.82 (95% CI, 0.67 to 0.99; P=0.046), indicating that SGLT2i users had a reduced risk of prostate cancer compared to oGLM users (Table 3). SGLT2i use was also associated with a significant reduction in the risk of death preceding a pancreatic cancer diagnosis (cause-specific HR, 0.74; 95% CI, 0.68 to 0.81) (Table 3). The consistent outcomes in the PSM cohort indicated a significant protective effect of SGLT2i against prostate cancer, independent of the overall mortality rate. The cumulative incidence graph depicted in Fig. 1 illustrates that SGLT2i users had a significantly lower risk of developing prostate cancer and experiencing mortality than oGLM users.
Kaplan–Meier curves of the cumulative incidence of prostate cancer and death according to the use of sodium-glucose cotransporter 2 inhibitors (SGLT2is) in the entire cohort (A, B) and propensity score matched cohort (C, D). HR, hazard ratio; CI, confidence interval.
Subgroup analyses
Major subgroup analyses were conducted based on age, BMI, smoking status, alcohol consumption, underlying diseases (chronic kidney disease, heart failure, or established atherosclerotic cardiovascular disease), and the use of insulin, metformin, thiazolidinediones, DPP-4 inhibitors, and sulfonylureas (Fig. 2). The beneficial effect of SGLT2is on prostate cancer appeared particularly pronounced in the subgroup with a BMI less than 25 kg/m² (HR, 0.63; 95% CI, 0.46 to 0.86 for BMI <25 kg/m²; HR, 0.97; 95% CI, 0.75 to 1.26 for BMI ≥25 kg/m²; P for interaction=0.037). Otherwise, the effect of SGLT2i on prostate cancer risk was consistent across the other pre-specified subgroups, except for that of BMI (Fig. 2).
Subgroup analyses for prostate cancer according to sodium-glucose cotransporter 2 inhibitor (SGLT2i) use in the propensity score matched (PSM) cohort. HR, hazard ratio; CI, confidence interval; BMI, body mass index; CKD, chronic kidney disease; ASCVD, atherosclerotic cardiovascular disease; DPP-4, dipeptidyl peptidase 4.
DISCUSSION
Our study underscored the potential protective benefits of SGLT2i for patients with T2DM in relation to prostate cancer. By analyzing 45,601 new users of SGLT2i and 205,395 new users of oGLM, we found a reduced incidence and HR for prostate cancer and mortality among SGLT2i users. Sensitivity analysis, which involved 1:1 PSM within the matched cohort, confirmed these findings, highlighting a consistent reduction in prostate cancer incidence and mortality. Subgroup analyses revealed that the protective effect of SGLT2i on prostate cancer was particularly pronounced in individuals with a BMI less than 25 kg/m²; however, the effects remained consistent across other subgroups. These results highlight the significant protective effect of SGLT2i against both prostate cancer and overall mortality.
In terms of prostate cancer, Scafoglio et al. [8] reported the functional expression of SGLT2 in tumors, opening new research avenues for SGLT2is. In this study, SGLT2 expression was demonstrated in prostate adenocarcinomas and tumor glucose uptake using the SGLT-specific radioactive glucose analog, α-methyl-4-deoxy-4-[18F]fluoro-D-glucopyranoside (Me4FDG). Notably, Me4FDG uptake in prostate acinar adenocarcinoma was blocked by phlorizin or dapagliflozin. In addition, using prostate cancer-3 cell line (PC-3) tumor xenografts in mice, these investigators found SGLT2 expression in vital tumor tissue. Moreover, dapagliflozin was found to inhibit the uptake of Me4FDG into prostate tumors by 40%–50% [8]. This study was the first to demonstrate the functional expression of SGLT2 in prostate carcinomas, inspiring our epidemiological study by highlighting new diagnostic and therapeutic possibilities for prostate cancer.
The following year, another significant study on the anticancer effects of SGLT2i was published [10]. In this study, canagliflozin was revealed to inhibit mitochondrial complex-I, leading to adenosine monophosphate-activated protein kinase (AMPK) activation and exhibiting anticancer activity in prostate cell lines [9,10]. The investigators also demonstrated that clinically achievable concentrations of canagliflozin inhibited cellular proliferation and clonogenic survival of prostate and lung cancer cells, both alone and in combination with ionizing radiation and the chemotherapy drug, docetaxel. Canagliflozin reduced glucose uptake, mitochondrial complex-I-supported respiration, adenosine triphosphate, and lipogenesis, while increasing the activating phosphorylation of AMPK [10]. More recently, Ali et al. [12] revealed that canagliflozin suppresses the proliferation and survival of androgen-sensitive and insensitive human prostate cancer cells and tumors, and sensitizes them to radiotherapy. These investigators found that canagliflozin blocks mitochondrial respiration, promotes AMPK activity, inhibits the MAPK and mammalian target of rapamycin–p70 S6 kinase/eukaryotic translation initiation factor 4E-binding protein 1 (mTOR-p70S6k/4EBP1) pathways, activates cell cycle checkpoints, and inhibits proliferation partly through hypoxia-inducible factor-1 alpha (HIF-1α) suppression [12].
Collectively, these prior studies suggest two possible mechanisms for the anticancer effects of SGLT2is on prostate cancer: reducing glucose uptake and cell growth by inhibiting SGLT activity and targeting mitochondrial metabolism to reduce cancer cell proliferation. Although several preclinical and clinical studies have explored the anticancer effects of SGLT2is on various cancers [3,13], no clinical trials have directly assessed their efficacy against prostate cancer. Notably, our study is the first to demonstrate that SGLT2i use reduces prostate cancer risk and all-cause mortality in Korean males with T2DM via multivariable-adjusted and PSM analyses; however, randomized controlled trials are needed to validate their use as anticancer agents against prostate cancer.
To provide historical context, the 2015 Empagliflozin Cardiovascular Outcome Event Trial in Type 2 Diabetes Mellitus Patients (EMPA-REG OUTCOME trial)—a landmark cardiovascular investigation of SGLT2is—revealed a significant reduction of 32% in overall mortality with empagliflozin compared to a placebo [14]. Subsequently, a growing body of evidence, including a comprehensive meta-analysis of 42 trials involving 61,076 patients with T2DM, highlighted the mortality benefits of this drug class. The meta-analysis revealed that compared to controls, SGLT2is were associated with a significant reduction in overall mortality (odds ratio, 0.85; 95% CI, 0.79 to 0.92; P<0.0001) [15]. Our study corroborates these findings, indicating a 25% reduction in all-cause mortality among SGLT2i users in the unmatched cohort and a 26% reduction in the PSM cohort.
Our subgroup analyses revealed a significant interaction between BMI and the effect of SGLT2is on prostate cancer (P for interaction=0.037) (Fig. 2). Although obesity, as measured by BMI, is a well-established risk factor for various cancers [16], evidence regarding its impact on prostate cancer risk and mortality remains inconsistent. Historically, obesity has been associated with poor prostate cancer prognosis. In fact, a prospective study comprising 404,576 United States men reported a significant trend of increasing risk and mortality from prostate and stomach cancers with higher BMI values [17]. A meta-analysis in 2022, which pooled data from 15 studies, concluded that obesity is associated with a higher risk of prostate cancer (HR, 1.03; 95% CI, 1.01 to 1.05) [18]. Conversely, a more recent large, multicenter, randomized controlled trial found an inverse association between higher BMI and prostate cancer incidence, including early and advanced-stage disease [19]. However, this study also noted that individuals with higher BMI were less likely to obtain a positive result in the prostate-specific antigen test and/or digital rectal exam and were more likely to have inadequate screening. Importantly, a higher BMI was positively associated with prostate cancer mortality, an association unaffected by screening outcomes [19]. Thus, despite no evidence of delayed detection, the relationship between obesity and prostate cancer mortality persisted, suggesting that delayed detection alone does not explain this link. Instead, these findings suggest that obesity may contribute to a fundamentally more aggressive form of prostate cancer [20]. Accordingly, based on our subgroup analysis, we postulated that the detrimental effect of obesity on prostate cancer may offset the protective effect of SGLT2is in obese populations. Further studies are needed to substantiate this explanation.
Our study had several limitations. Owing to the observational nature of this study, it is susceptible to biases, such as coding errors, missing data, and under-reporting. The reliance on recorded prescriptions introduces potential adherence issues, which could skew the results toward no effect. Although a new-user cohort study design was employed to address confounding by indication, residual and unmeasured confounding could not be completely eliminated. In addition, as a retrospective approach was employed, correlational rather than causal conclusions were drawn in this study. Larger-scale trials will be essential to validate the protective role of SGLT2is in prostate cancer. In supplementary analyses, we explored the associations between SGLT2i use and the incidence of kidney and bladder cancers. However, after multivariable adjustment and PSM, neither outcome showed a statistically significant association with SGLT2i exposure (data not shown). Given the relatively low incidence of these malignancies compared to prostate cancer, our study may have lacked sufficient statistical power to detect meaningful differences. Alternative designs, such as case–control studies, may be more appropriate for evaluating these rarer outcomes. Furthermore, as our cohort was limited to individuals with T2DM, the generalizability of our findings to non-diabetic populations remains uncertain. Additionally, this study did not examine the effects of SGLT2is in individuals with pre-existing prostate cancer. Future research is warranted to evaluate the therapeutic and preventive roles of SGLT2is in other cancer types and populations, including their potential impact on cancer progression and mortality.
In summary, our findings support the notion that SGLT2is may reduce the risk of prostate cancer, warranting further investigation in this area. The presence and physiological role of SGLT2 in prostate cancer provide a plausible biological mechanism for this effect. Moreover, accumulating evidence of the anticancer properties of SGLT2i has already highlighted their promising use as potential prostate cancer treatments. Based on the evidence gathered to date, a randomized trial is warranted to explore the protective role of SGLT2i, particularly in individuals at increased risk of developing prostate cancer.
SUPPLEMENTARY MATERIALS
Supplementary materials related to this article can be found online https://doi.org/10.4093/dmj.2024.0693.
Balance assessment between new users of sodium-glucose cotransporter 2 inhibitors and users of other GLMs before and after propensity score matching
Flow diagram of patient inclusion in the study cohort. SGLT2i, sodium-glucose cotransporter 2 inhibitor; GLM, glucose-lowering medication.
Notes
CONFLICTS OF INTEREST
Chang Hee Jung has been an associate editor of the Diabetes & Metabolism Journal since 2022. He was not involved in the review process of this article. Otherwise, there was no conflict of interest.
AUTHOR CONTRIBUTIONS
Conception or design: Y.J.K., C.H.J.
Acquisition, analysis, or interpretation of data: Y.K.C., S.K., Y.J.K.
Drafting the work or revising: Y.K.C., S.K., M.J.K., W.J.L., Y.J.K.
Final approval of the manuscript: Y.J.K., C.H.J.
FUNDING
This work was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Education (grant number NRF-2020R1A2C1101977: Chang Hee Jung). These funding sources had no roles in the writing of the article or the decision to submit the article for publication.
ACKNOWLEDGMENTS
None