Antimicrobial treatment ameliorates delirium-like phenotypes in a murine model of urinary tract infection

Abstract

Practice guidelines recommend withholding antimicrobial therapy (ABX) in delirious patients with suspected urinary tract infection (UTI) who do not endorse classic genitourinary symptoms, citing both a lack of a causal relationship between bacteriuria and delirium and benefit from ABX. In this study, we tested the hypothesis that UTI induces delirium-like phenotypes that are mitigated by ABX. Escherichia coli (CFT073) UTI was induced in female C57BL/6 J mice aged 4–5 months. Mice were randomized to receive one dose of ceftriaxone 600 mg/kg (n = 23) or saline (n = 21) one day after induction of UTI. Delirium-like behaviors were assessed using Open Field, Elevated Plus Maze, and Y-maze, while neurostructural changes were evaluated using neuronal cleaved caspase-3 (CC3) and interleukin-6 (IL-6) via immunohistochemistry. Plasma IL-6 was quantified using ELISA. Compared to vehicle-treated mice, ABX mice with UTI demonstrated: 1) decreased time in the periphery of the Open Field maze (p = 0.017), 2) decreased time in the closed arms of the Elevated Plus Maze (p = 0.013), and 3) increased spontaneous alternations in Y-maze (p = 0.015). These behavioral changes were accompanied by significantly lower frontal/hippocampal CC3 (p = 0.0038, p = 0.0003, respectively) and IL-6 (p = 0.015) levels in ABX- compared to vehicle-treated UTI mice. ABX significantly lowered plasma IL-6 compared to vehicle-treated UTI mice (p < 0.01). This study suggests a causal relationship between UTI and functional/neurostructural delirium-like phenotypes that are attenuated with ABX. These findings provide strong rationale for a randomized clinical trial to evaluate the role of ABX in patients with delirium as the isolated presumed sign of UTI.

Introduction

Urinary tract infection (UTI) is a highly prevalent clinical condition with more than 400 million cases per year worldwide [1, 2]. Delirium, which presents with acute confusion, occurs in up to one-third of patients and is associated with increased mortality, prolonged hospitalization, and an increased risk of incident or progressive Alzheimer’s disease and related dementias (ADRD) [3–9].

UTI is distinguished from asymptomatic bacteriuria by the presence of classic focal genitourinary symptoms such as urinary frequency, urgency, dysuria, and costovertebral angle tenderness. However, delirious patients, who may be suspected to have UTI, often demonstrate bacteriuria without complaints of focal genitourinary symptoms and are thus considered asymptomatic [6, 10]. The current Infectious Diseases Society of America (IDSA) practice guidelines recommend in cognitively impaired patients “assessment for other causes and careful observation rather than antimicrobial treatment” (ABX), citing both a lack of a causal relationship between bacteriuria and delirium and an apparent lack of benefit with ABX in patients who do not complain of focal genitourinary symptoms [11].

These IDSA guidelines, however, acknowledge the low-quality evidence that inform these recommendations and call for studies to clarify potential causal relationships between UTI and delirium and evaluate putative benefits of ABX on mental status. Recent developments in animal models now facilitate these investigations - our mouse model of UTI-induced delirium recapitulates a common clinical scenario where acute behavioral changes, i.e., delirium-like phenotypes, are considered the only presenting sign of UTI where cognitive impairment due to delirium may not allow elicitation of the presence of focal genitourinary symptoms.

In recent studies, we demonstrated a central pathological role for the interleukin-6 (IL-6) trans-signaling pathway in orthogonal etiologies of delirium, which have been subsequently validated in clinical populations [2, 12, 13]. Although these studies demonstrate the feasibility to leverage pharmacology to reverse delirium-like phenotypes after acute systemic injuries, it remains unknown whether ABX alone may ameliorate UTI-induced delirium-like phenotypes.

Accordingly, we hypothesize the existence of a causal relationship between UTI and delirium-like phenotypes that is mitigated by ABX.

Methods

Animals

Forty-four female C57BL/6 J (#0664, Jackson Laboratory, Bar Harbor, ME, USA) mice aged 4–5 months had UTI induced as described below, over 4 weeks with 5–6 animals/group/week. One day later, mice were injected with a single dose of intraperitoneal (IP) ceftriaxone 600 mg/kg (ABX, n = 23) or saline (vehicle, n = 21) (Fig. 1A).

Fig. 1. UTI does not induce fever in mice and is resolved by antimicrobial treatment.

A Timeline for main experiment. B Bacterial concentration in urine shows no difference between ABX- or vehicle-treated mice before antibiotic treatment, but there was a significant difference after a single-dose of ceftriaxone (n = 24, p < 0.01). C There was a significant difference in bacterial concentration in the bladder after a single-dose of ceftriaxone (n = 44, p < 0.01). D Rectal temperatures were not significantly different between UTI and non-UTI mice.

In a follow up experiment, UTI was induced in one cohort of 12 female C57BL/6 J mice aged 4–5 months and immediately injected with ceftriaxone 300 mg/kg IP (ABX, n = 6) or saline (Vehicle, n = 6) and given an additional dose of ceftriaxone 300 mg/kg or saline IP a day later. Based on our prior work in this model, we established a minimum of 10–12 animals/group to detect significant differences [12, 14].

All mice were housed in Cedars-Sinai’s AAALAC accredited animal facility under standard conditions (kept in ventilated cages at approximately 21 °C, 40–70% humidity, a 12-h light/dark cycle, with food and water available to the animals ad libitum). All procedures herein follow the recommendations in the ARRIVE 2.0 guidelines for research involving the use of animals [15]. All experiments were conducted in accordance with Cedars-Sinai Medical Center Institutional Animal Care and Use Committee guidelines under an approved protocol (#7914). All assessors were unaware of group allocations during analyses. No mice were individually housed. UTI and non-UTI mice were housed separately to prevent cross-contamination of E. coli and ABX-treated mice were separated from vehicle-treated mice to prevent possible cross-treatment of ABX.

UTI induction and antimicrobial treatment

The following methods have been published previously [12, 16] and are restated here. Cultures of uropathogenic E. coli (Migula) Castellani and Chalmers (#700928, ATCC, Manassas, VA, USA) strain (CFT073) were grown for at least 24 h at 37 °C, pelleted by centrifugation (10 min, 3700 rpm, 4 °C), and resuspended in sterile phosphate-buffered saline (PBS) at a concentration of 1 × 10⁹ colony forming unit/mL (CFU/mL). Prior to induction of UTI, mice underwent baseline behavioral testing as described below. The following day, mice were anesthetized with 2% USP isoflurane delivered by a vaporizer (VetEquip Inc., Pleasanton, CA, USA) at 0.8 liters per minute of oxygen. Mice cages were randomly assigned to injury and treatment groups. A 0.5 inch long, 30-gauge needle was inserted into 0.61 mm internal diameter polyethylene tubing (PE10, Becton-Dickinson, Sparks, MD, USA) and inserted into the urethra using water-based lubricant to deliver 100 μL of inoculum containing 1 × 10⁸ CFUs of E. coli or equal volume of sterile PBS. ABX was prepared by dissolving ceftriaxone in saline at 100 mg/mL and injected intraperitoneally into mice at a corresponding dose. Vehicle treatment was USP saline and given at equal volumes by weight. For the main experiment, mice received a single treatment dose 1-day post-inoculation of 600 mg/kg ceftriaxone or saline, administered IP. In the follow up experiment, mice were given two treatment doses on the day-of and 1-day post-inoculation of 300 mg/kg ceftriaxone IP. Post-induction behavioral assessments are performed at 2-days post-inoculation, and at 3-days post-inoculation the mice are deeply anesthetized with 75 mg/kg ketamine and 0.5 mg/kg dexmedetomidine and then euthanized by perfusion. Urine is collected at 1- and 3-days post-inoculation and plated on agar for colony counting.

Behavioral testing

To evaluate delirium-like behaviors, individual mice were assessed using the Open field, Elevated Plus Maze, and Y-maze assays one day prior (baseline) and 2 days after induction of UTI. Each of the following tests are described below:

Open field

The Open Field test assesses exploratory behavior and locomotor activity. Inducing delirium may exhibit behaviors such as altered level of consciousness, anxiety, or cognitive impairment [17, 18]. Time spent immobile has been shown to be increased in anxiety-related conditions [19] and also could reflect hypoactive delirium states [20]. Each mouse was placed in opaque arenas (40 × 40 × 40 cm) and monitored for 45 min through an automated camera (Digital USB 2.0 CMOS Camera, Stoelting Co., Wood Dale, IL, USA). 18 × 18 cm center zone and surrounding periphery zone for analysis. ANY-maze Video Tracking Software (Stoelting Co.) was used to track total distance traveled (meters), distance traveled within individual zones (meters), time spent immobile (seconds), and time spent in each zone (seconds).

Elevated plus maze

The Elevated Plus Maze measures behavior associated with anxiety and altered level of consciousness [18, 21]. The maze is elevated 30 cm above ground, comprised of two opposite 310 mm long closed arms surrounded by walls and two opposite 310 mm long open arms without walls resembling a plus sign from above. Mice displaying anxiety-like behavior will spend more time in the closed wall arms than the open wall arms. Time spent in either arm was measured and analyzed.

Y-maze

The Y-maze measures multiple forms of memory such as short-term, attentional, and cognition [22, 23]. The maze is comprised of three walled arms of 360 mm length each placed at angles 120 degrees from each other. Each mouse is inserted into one arm and is monitored for 10 min. Uninfected mice will preferably venture into the unexplored arms of the Y-maze than the arm they are already familiar with, meaning repeated visits of the same arms points to impaired memory. ANY-maze tracking software was used to record videos of the tests, and the number and sequence of arm entries was recorded manually so the percentage of spontaneous alterations could be calculated.

Histology

Brain isolation

Mice were perfused using 20 mL PBS containing 0.5 mM ethylenediaminetetraacetic acid after administering ketamine/dexmedetomidine anesthetic. The right hemisphere was obtained and suspended in PBS buffered 4% paraformaldehyde on ice for 24 h prior to addition of sucrose to cryoprotect the tissue in a 2% paraformaldehyde + 30% sucrose solution. After 2 weeks, 30-µm coronal cryosections were suspended at 4 °C in PBS + 0.02% sodium azide. Plasma IL-6 concentrations were measured using ELISA (#M6000B, R&D Systems, Minneapolis, MN, USA).

Immunohistochemistry and microscopy

Sections were placed in heat-induced epitope retrieval for 10 min in 6.0 pH citrate antigen retrieval buffer prior to permeabilizing and blocking in an animal-free blocking solution (#15019, Cell Signaling Technologies, Danvers, MA, USA) at room temperature. Frontal/hippocampal cleaved caspase-3 (CC3) was quantified as a neurostructural marker of delirium-like behavior, consistent with prior studies demonstrating a dose-dependent relationship between frontal/hippocampal CC3 and delirium-like behaviors [14, 24]. Though well known as an early apoptotic marker, CC3 has also been implicated in various neuronal and synaptic regulatory functions [25]. NeuN, neuronal nuclear protein marker, was utilized to identify CC3-expressing neurons. After, sections were placed in primary antibody solution overnight consisting of antibodies for CC3 (#9664, Cell Signaling Technology, Danvers, MA), IL-6 (#AMC0864, ThermoFisher Scientific, Waltham, MA, USA), NeuN, and antibody diluent (#MA5-33103, ThermoFisher Scientific). Primary antibody solution was then removed, and fluorescent secondary antibodies (AlexaFluor, ThermoFisher Scientific, Waltham, MA, USA) were administered to the tissue. CC3 signal amplification was achieved using HRP-conjugated secondary antibodies (#ab214880, Abcam, Cambridge, UK) in conjunction with the Opal 650 reagent (Akoya Biosciences, Marlborough, MA, USA). Sections were then captured using Zeiss AxioImager Z.2 epi-fluorescence microscope and a 10x objective lens (Zeiss, Oberkochen, Germany) as a tiled image including the frontal cortex and hippocampus. Frontal IL-6 images were acquired using a 20x objective and taking a 9-slice stack (1 um spacing) of a single field of view with the Apotome.

Image analysis

FIJI ImageJ was utilized to analyze images exported as 8-bit TIFF files. Each animal had three coronal sections analyzed in regions of interest (ROIs) around the entire hippocampus or a 1.2 mm² area of the frontal cortex. The pixel intensity histogram was used to select a threshold and record the number of positive pixels as a percentage of the total area in each ROI. The percent areas were averaged for each animal. Frontal IL-6 images were optically sectioned, orthogonally projected in ZEN Blue (Maximum projection), pre-processed with background subtraction (rolling ball radius = 50.0 pixels) as full-size TIFF files before conversion to 8-bit for thresholding.

Statistcal analysis

Prism (GraphPad Software, Boston, MA, USA) was used for statistical analyses. Repeated Measures (RM) ANOVAs were used to analyze bacterial growth and behavioral assays (Figs. 1, 2, 6) with mixed-effects modeling used to adjust for possible missing data, with residuals inspected to confirm normal distribution and absence of outliers. Post-hoc pair-wise comparisons were completed with uncorrected Fisher’s LSD test. Independent sample Welch’s t-tests were used to evaluate differences between groups in histological analyses (Figs. 3, 4). Pearson’s correlation coefficient was used for behavior and histology correlations, as data were determined to be normally distributed (Fig. 5). All differences were considered significant where two-sided p-values were <0.05. Quantitative data are expressed in mean ± SEM. Figure legends list any exclusions from analyses.

Fig. 2. ABX started one day following UTI induction significantly improves delirium-like behaviors.

A The time spent in the periphery, when normalized for distance, during the Open Field test was significantly decreased in ABX compared to vehicle-treated mice with UTI (n = 23 ABX, n = 21 vehicle, p = 0.01), indicating lower anxiety-like behaviors for ABX. B Time spent immobile in the Open Field was significantly lower in ABX compared to vehicle-treated UTI (n = 23 ABX, n = 21 vehicle, p = 0.04), indicating decreased anxiety-related behavior. C The percentage of spontaneous alternations in the Y-maze was significantly higher in ABX compared to vehicle-treated UTI mice (n = 23 ABX, n = 21 vehicle, p = 0.02), indicating improved attention and short-term memory. D Following UTI, ABX mice showed decreased preference to the closed arms of the Elevated Plus Maze compared to vehicle-treated mice, indicating a lower anxiety-like state (n = 23 ABX, n = 21 vehicle, UTIxABX p = 0.02). Quantitative data are expressed in mean ± SEM. Results from Fisher’s LSD are displayed. Residuals were observed and there were no influential outliers.

Fig. 6. Early ABX initiation reduces delirium-like behaviors.

A Timeline of early ABX initiation study. Mice were treated with two doses of 300 mg/kg ceftriaxone (n = 6) or vehicle (n = 6), one day apart, beginning on the day of inoculation. B Compared to vehicle-treated UTI mice, early ABX-treated mice showed significantly reduced time in the periphery of the Open Field, normalized to total distance traveled (p = 0.07 and p = 0.04, respectively). C UTI mice treated with early ABX-treatment exhibited greater spontaneous alternations in the Y-maze (p = 0.01) compared to vehicle-treated mice. D Following UTI, early ABX-treated mice spent less time in the closed arms of the Elevated Plus Maze (p = 0.05, UTIxABX p = 0.01) compared to their vehicle-treated controls. The sample sizes for all the post-UTI comparisons are n = 6 vehicle and n = 5 early ABX. One early ABX animal was excluded from post-behavioral analyses due to a physical injury. Quantitative data are expressed in mean ± SEM. Results from Fisher’s LSD are displayed. Residuals were observed and there were no influential outliers.

Fig. 3. ABX significantly reduces neuronal injury in mice with UTI.

A Frontal and hippocampal neuronal CC3 were significantly lower in ABX compared to vehicle-treated UTI mice (n = 13 ABX and 10 vehicle, p < 0.01, for both). Frontal and hippocampal neuronal CC3 were not significantly different in ABX- compared to vehicle-treated PBS-inoculated control mice. B Representative images of 10X neuronal CC3 in vehicle- and ABX-treated UTI mice, respectively. One vehicle animal was excluded from all analyses due to a technical failure with staining. Quantitative data are expressed in mean ± SEM. Residuals were observed and there were no influential outliers.

Fig. 4. ABX significantly reduces neuronal injury and peripheral IL-6 in mice with UTI.

A Plasma IL-6 is significantly lower in ABX- compared to vehicle-treated UTI mice (n = 24, p < 0.01). B Frontal cortex IL-6 is significantly lower in ABX- compared to vehicle-treated UTI mice (n = 24, p = 0.01) C 20x representative images of CC3 and IL-6 in vehicle- and ABX-treated mice, respectively. One vehicle animal was excluded from Frontal IL-6 analysis due to a technical failure with staining. Quantitative data are expressed in mean ± SEM. Residuals were observed and there were no influential outliers.

Fig. 5. Neuronal injury correlates with delirium-like behaviors following UTI.

A Pearson’s correlation of hippocampal CC3 and Y-maze performance showed a significant inverse correlation. (n = 23, R = −0.43, p = 0.04). Only mice with both CC3 and Y-Maze data were included. B Pearson’s correlation of hippocampal CC3 and peripheral time in the Open Field Test showed a significant positive correlation (n = 23, R = 0.42, p = 0.04). Only mice with both CC3 and Open Field data were included. The lines on the graphs show the results of a Simple Linear Regression.

Results

Figure 1A shows the main experimental timeline. ABX significantly reduced bacterial CFU count after 3 days following UTI induction (Fig. 1B, C; p < 0.01) while temperatures were not significantly different between UTI and non-UTI controls. (Fig. 1D).

ABX significantly improves delirium-like behaviors in mice with UTI

There were no significant differences in baseline delirium-like behaviors prior to induction of UTI between vehicle- and ABX mice, except in the Elevated Plus Maze (Fig. 2A–D). After induction of UTI, ABX mice spent significantly less time in the periphery of the Open Field when normalized for total distance traveled compared to vehicle-treated mice (Fig. 2A; p = 0.01). In addition, ABX mice with UTI demonstrated significantly lower time immobile in the Open Field compared to vehicle-treated controls (Fig. 2B; p = 0.04).

In the Y-maze, ABX mice showed significantly higher percentage of spontaneous alternation when compared to their vehicle-treated controls after induction of UTI (Fig. 2C, p = 0.02). Similarly, vehicle-treated animals spent significantly more time in the closed arm of the Elevated Plus Maze compared to ABX mice (Fig. 2D, UTIxABX p = 0.02).

ABX significantly reduces neuronal injury and reduces peripheral IL-6 in mice with UTI

Frontal and hippocampal neuronal CC3 levels were significantly lower in ABX compared to vehicle-treated mice with UTI (Fig. 3A, B; p < 0.005, p < 0.001, respectively) while there was no significant difference in frontal/hippocampal CC3 between ABX and vehicle-treated non-UTI controls (Fig. 3A). Representative fluorescence microscopy images for frontal and hippocampal CC3 for both groups are shown in Fig. 3B.

At 3 days post-induction of UTI, plasma and frontal cortex IL-6 levels were significantly reduced in ABX mice compared to vehicle-treated controls (Fig. 4A, B; p < 0.001, p = 0.01, respectively). Representative fluorescence microscopy images for frontal CC3 and IL-6 for both groups are shown in Fig. 4C.

Delirium-like behaviors correlate with hippocampal CC3

There was a significant inverse correlation between hippocampal CC3 and spontaneous alternations on Y-maze (Fig. 5A; R = −0.43, p = 0.04). There was also a significant positive correlation between hippocampal neuronal CC3 and normalized periphery time (Fig. 5B; R = 0.42, p = 0.04).

Early ABX significantly reduces UTI-induced delirium-like behaviors

To further elucidate the temporal relationship between UTI and delirium, we tested whether early initiation of ABX, i.e., concurrent to induction of UTI, enhances the treatment effect on delirium-like behavioral phenotypes (Fig. 6A). In a cohort of 5–6 animals/group, we observed significant improvements in delirium-like behaviors in early ABX compared to vehicle-treated mice with UTI. Specifically, compared to vehicle-treated mice, early ABX-treated mice showed significantly reduced time in the peripheral zone normalized to total distance (Fig. 6B; p = 0.07 and p = 0.04, respectively) in the Open Field. Additionally, mice with UTI treated with early ABX exhibited greater spontaneous alternations in the Y-maze (Fig. 6C; p = 0.01) and spent significantly less time in the closed arms of the Elevated Plus Maze (Fig. 6D; UTIxABX p = 0.01) compared to their vehicle-treated controls.

Early ABX abrogates delirium-like behaviors compared to delayed ABX one day after induction of UTI

To compare the effects of early (Fig. 6) versus delayed (Fig. 2) ABX treatment regimens on delirium-like behaviors, we used mixed modeling to allow for the random effects of each animal and experiment.

Mice with UTI treated with delayed ABX or vehicle spent significantly more time in the periphery of the Open Field, normalized for total distance traveled (p = 0.026 and p < 0.001, respectively), while there was no significant difference in periphery time/total distance traveled for UTI mice who received early ABX (p > 0.999). Similarly, UTI mice treated with delayed ABX or vehicle spent significantly less time in the closed arms of the Elevated Plus Maze (p < 0.001 for both), while there was no significant difference in animals who received early ABX (p = 0.321). Finally, percent spontaneous alternations were significantly greater in UTI mice treated with early ABX compared to vehicle (p = 0.047), while there was no significant difference in spontaneous alternations between UTI mice treated with delayed ABX and vehicle (p = 0.092). These findings are summarizd in Supplemental Table 1.

Discussion

This study suggests a causal relationship between UTI and functional/neurostructural delirium-like phenotypes, and a beneficial role of ABX in mitigating these changes. These data are consistent with prior studies implicating peripheral IL-6 in the pathogenesis of delirium and demonstrate that ABX reduces the host IL-6 response. These results call for randomized clinical trials to evaluate the role of ABX for patients with delirium as the isolated presumed sign of UTI.

There are widespread clinical implications for our findings. Apart from the potential to alleviate risk of mortality and duration of hospital stay, delirium is associated with a three-fold increased risk of incident ADRD and doubling in the rate of cognitive decline in patients with ADRD. Indeed, delirium mitigation is now suggested as a potentially modifiable risk factor for development or progression of ADRD [4, 7]. Accordingly, clinical studies focused on patients with suspected UTI and delirium should be designed to evaluate both short-term delirium as well as long-term ADRD symptoms. However, it is also worth noting that the implications of this study’s findings extend beyond delirious patients to other populations who may not be able to endorse classic genitourinary symptoms due to dementia, acute stroke, or acute seizure.

Our findings support the already common practice of using ABX to treat presumed UTI-related delirium [26] and further suggest that earlier initiation of ABX may provide the best opportunity to mitigate delirium. However, any beneficial effects of liberal ABX use must be balanced with the increased risk of Clostridioides difficile infection. Future clinical studies may be designed to include delirious patients with suspected UTI and an additional systemic inflammatory response marker, such as elevated plasma IL-6, which is thought to play a central role in delirium pathogenesis [27]. Additional clinical studies are also indicated to evaluate whether a pathophysiological definition of UTI that includes the presence of active inflammation, e.g., elevated urinary IL-6, and bacteria may more effectively guide antibiotic treatment decisions.There is precedence for using ABX to treat pregnant patients with bacteriuria without classic UTI symptoms, justified by improved pregnancy outcomes with ABX compared to untreated women [28–33]. Clinical studies are needed to evaluate whether ABX directed towards delirium as a presumed sign of infection may improve clinical outcomes.

Our study has several notable strengths including the use of a robust animal model of delirium that encompasses both functional, i.e., behavioral, and structural, i.e., neuronal CC3 and IL-6, phenotypes. Recent advances in preclinical models of delirium now enable assessment of the key delirium-like phenotypes described in this manuscript. Hallmark features of delirium that recapitulate the human condition are evaluated to include an acute change in behavior in response to a systemic stressor, impairments in attention or short-term memory on Y-maze, and evidence of anxiety-like behavior on Open Field and Elevated Plus Maze [12, 14, 20, 24, 34]. An important limitation to consider is that we used a single pathogen, E. coli (CFT073) and a single ABX, ceftriaxone – thus, it is not known whether these findings can be generalized to other antibiotics or organisms. Second, to reflect the population that is disproportionately affected by UTI, we used only female mice -- future studies are needed to evaluate these findings in male subjects. Third, although our model does not induce fever, we were not able to evaluate whether these mice had classic UTI symptoms such as urinary frequency, urgency, dysuria, or costovertebral angle tenderness. However, in the clinical scenario patients with suspected UTI frequently present with delirium as their only sign and do not report or endorse classic UTI symptoms.

In summary, we demonstrate amelioration of delirium-like phenotypes in ABX-treated mice with UTI. These findings provide a strong rationale for a randomized clinical trial to evaluate the role of ABX in patients with delirium as the isolated presumed sign of UTI.

Author contributions

KDW, LS, DM, and WGT contributed to the acquisition of data. KDW, LS, CB, JV, FS, SAK, and SL did the analysis and interpretation of the data. KDW, LS, KHS, TSI, SL drafted the manuscript. SL contributed to the study concept and design. SL did the study supervision. All authors read and approved the final manuscript.

Funding

F. Widjaja Foundation Research Grant (SL).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

All experiments were conducted in accordance with Cedars-Sinai Medical Center Institutional Animal Care and Use Committee (IACUC) guidelines under an approved protocol (#7914) and complied with current US law.

Supplementary information

The online version contains supplementary material available at 10.1038/s41398-025-03624-9.