Alan S. L. Yu

This study provides proof-of-concept for a CRISPR activation (CRISPRa)-based approach to ADPKD. CRISPRa targeting the PKD1 promoter increased polycystin-1 levels in mouse and human cells. Upregulation of a hypomorphic Pkd1 allele increased functional polycystin-1, corrected dysregulated signaling pathways, and suppressed cell proliferation and in vitro cyst formation. These results establish CRISPRa as a promising therapeutic approach to restore polycystin levels above a critical threshold and suppress cystogenesis in ADPKD.

Background: Acute kidney injury (AKI) is a common, life-threatening complication among hospitalized patients. Timely prediction is crucial for prevention, while accurate prognosis following onset is essential for guiding care. Although electronic health records (EHRs) provide rich longitudinal data, their high dimensionality, heterogeneity, and temporal complexity pose challenges for traditional machine learning. Objective: The aim of this study was to develop and evaluate a transformer-based model for generalizable AKI prediction and prognosis by modeling long-range temporal dependencies and complex relationships in heterogeneous EHR data and to examine its interpretability. Methods: We developed the Biomarker-Aware Transformer for AKI (BAT-AKI), which integrates multidimensional embeddings capturing temporal dynamics, textual descriptions, hierarchical ontology structures, and semantic relationships in EHRs. To enhance sensitivity to AKI-related biomarker fluctuations, BAT-AKI combines masked language modeling with a novel biomarker abnormality-aware pretraining task. The model was pretrained and fine-tuned for 4 tasks: AKI onset, early reversal, recovery at discharge, and 90-day post-AKI mortality. Pretraining and internal validation used data from the National Patient-Centered Clinical Research Network (PCORnet) site, with external validation across 3 additional sites. Results: BAT-AKI achieved the strongest overall discrimination, outperforming baseline models in area under the receiver operating characteristic curve (AUROC) and area under the precision-recall curve (AUPRC) on most tasks. For AKI onset prediction, it achieved an AUROC of 0.8549 (95% CI 0.8433-0.8658) and an AUPRC of 0.6041 (95% CI 0.5786-0.6292), exceeding the second-best baseline by 2.8% in AUROC and 8.1% in AUPRC. Across prognostic tasks, BAT-AKI achieved AUROCs of 0.6006 (95% CI 0.5703-0.6318), 0.6080 (95% CI 0.5745-0.6397), and 0.8017 (95% CI 0.7708-0.8311) and AUPRCs of 0.7036 (95% CI 0.6690-0.7388), 0.7417 (95% CI 0.7082-0.7765), and 0.5063 (95% CI 0.4422-0.5684) for early reversal, recovery at discharge, and 90-day post-AKI mortality, respectively, corresponding to relative gains of 2.4% to 6% in AUROC and 3.7% to 4.1% in AUPRC over the second-best baseline. Performance remained stable across external cohorts, with AKI onset AUROCs ranging from 0.7754 (95% CI 0.7677-0.7827) to 0.7989 (95% CI 0.7885-0.8092) and AUPRCs ranging from 0.4777 (95% CI 0.4631-0.4924) to 0.5687 (95% CI 0.5495-0.5871), corresponding to relative improvements of 3.7% to 5.3% in AUROC and 16.8% to 25.3% in AUPRC over the transformer baseline. In a low-resource setting using only 2% of the fine-tuning data, BAT-AKI achieved an AUROC of 0.7898 (95% CI 0.7759-0.8036) and an AUPRC of 0.4903 (95% CI 0.4640-0.5178) for AKI onset prediction, exceeding the second-best baseline by 4.2% in AUROC and 10.9% in AUPRC. Visualization of embeddings, attention patterns, and integrated-gradient feature attributions highlighted clinically meaningful relationships and patient-specific trajectories. Conclusions: BAT-AKI provides generalizable AKI prediction and prognosis with consistent discrimination across institutions, while offering interpretability insights into its predictions, supporting its potential for early warning and AKI care optimization.

Most cases of autosomal dominant polycystic kidney disease (ADPKD) are caused by mutations in PKD1, which reduce polycystin-1 (PC1) levels below a critical functional threshold. Normalizing PC1 dosage mitigates disease progression; therefore, we sought to develop a CRISPR activation (CRISPRa) strategy to transcriptionally upregulate endogenous PKD1 . We systematically screened multiple single-guide RNAs using an EGFP-reporter platform and identified potent candidates targeting the proximal PKD1 promoter in mouse and human cell models. Our results demonstrate that CRISPRa effectively increased endogenous Pkd1 mRNA in the mouse collecting duct-derived Pkd1 ^RC/- cell model and in the primary renal epithelial cells from PKD mice. In Pkd1 ^RC/- cells, CRISPRa of Pkd1 increased PC1 protein levels and significantly reduced cell proliferation and in vitro cyst formation in 3D cultures. Mechanistically, Pkd1 activation improved mitochondrial membrane potential, reduced dependency on aerobic glycolysis, and corrected signaling pathways involved in cystogenesis, specifically reducing intracellular cAMP, cMyc, pCreb, and pErk levels, while increasing pYap1 levels. We confirmed the translational potential of this platform by successfully activating PKD1 in primary renal epithelial cells from human kidneys. We observed a heterogeneous response across both normal and ADPKD patient-derived donor lines, with significant upregulation achieved in two of the tested cell preparations. These findings provide a compelling proof-of-concept that CRISPRa-mediated gene augmentation can increase PC1 levels, establishing a foundation for promising gene therapies aimed at successfully suppressing the pathogenic features of ADPKD. New and noteworthy: This study provides proof-of-concept for a CRISPR activation (CRISPRa)-based approach to ADPKD. CRISPRa targeting the PKD1 promoter, increased polycystin-1 levels in mouse and human cells. Upregulation of a hypomorphic Pkd1 allele increased functional polycystin-1, corrected dysregulated signaling pathways, suppressed cell proliferation and in vitro cyst formation. These results establish CRISPRa as a promising therapeutic approach to restore polycystin levels above a critical threshold and suppress cystogenesis in ADPKD.

Background Autosomal Dominant Polycystic Kidney Disease is caused by loss-of-function mutations in PKD1 or PKD2 genes, leading to reduced polycystin protein levels. Increasing PKD1 expression via CRISPR activation (CRISPRa) represents a promising therapeutic strategy; however, delivery of large CRISPRa plasmids into renal epithelial cells, and particularly primary cells, remains inefficient due to size-related barriers. We aimed to enable Pkd1 transactivation by miniaturizing CRISPRa plasmids into ~ 6 kb vectors using a one-pot method to enhance cellular uptake in mouse kidney epithelial cells. Methods and Results Using type IIS restriction enzymes, we excised the mammalian expression cassette from full-length large 9–11 kB plasmids. The excised cassette was engineered to have complimentary overhangs. Thermocycling with T4 DNA ligase promoted circularization of the excised cassette (forming ~ 6kB mini-CRISPRa vectors), and T5 exonuclease digestion removed residual backbone fragments. These mini vectors substantially enhanced nucleofection efficiency from 16.10% ± 0.53 to 54.17% ± 2.10 in Pkd1^RC/− cells, and from 10.14% ± 1.40 to 31.27% ± 0.12 in primary Pkd1^RC/Cond; Pkhd1^Cre+ cells. Functionally, the mini-CRISPRa plasmid (mdCas9-VPR) with Pkd1-targeting sgRNAs induced robust endogenous Pkd1 upregulation compared with non-targeting controls: a 4.1-fold increase in Pkd1^RC/− cells (p < 0.001) and a 2.9-fold increase in primary cells (p < 0.001). Full-length plasmids produced no significant activation in either cell type. Conclusions Miniaturization of CRISPRa vectors with this one-pot approach overcomes delivery limitations in hard-to-transfect renal epithelial cells and enables efficient, functional Pkd1 activation, in vitro.

Key Points: In autosomal dominant polycystic kidney disease, there is a linear relationship between the log-transformed total kidney volume and eGFR. There is a predictable change in the rate of eGFR decline with reduction in kidney growth rate. This framework can be used to determine treatment effect to support accelerated approval of drugs for polycystic kidney disease. Background: Total kidney volume (TKV) is accepted by the US Food and Drug Administration as a surrogate end point that is reasonably likely to predict clinical benefit in autosomal dominant polycystic kidney disease and the most commonly used response biomarker for proof-of-concept intervention trials. However, the magnitude of treatment effect on TKV that would be predictive of a meaningful improvement in a clinical outcome, such as eGFR, is unknown. Inference of this from observational studies has previously been approached by examining interindividual variance in the relationship between TKV and GFR slopes over time. Methods: We developed a novel approach to modeling the intraindividual relationship between TKV and eGFR. Patients from the Consortium for Radiologic Imaging Studies of Polycystic Kidney Disease and Halt Progression of Polycystic Kidney Disease Study A dataset were stratified by Mayo Imaging Class (MIC). Linear mixed models were fitted to eGFR with a fixed effect of log(TKV) and random intercepts, and the average slope within each MIC was estimated. Results: We found that within each MIC, there is a consistent, linear relationship between log(TKV) and eGFR. The model predicts that within classes 1C–1E, for each 1% point per year reduction in TKV growth rate, the rate of eGFR decline would be reduced by 0.40–0.52 ml/min per 1.73 m ² per year. Conclusions: We have developed a new model that provides a framework for defining the magnitude of treatment effect on TKV that would support accelerated approval of a drug for autosomal dominant polycystic kidney disease. Podcast: This article contains a podcast at https://dts.podtrac.com/redirect.mp3/www.asn-online.org/media/podcast/CJASN/2026_05_18_CJASNMay.21.5.51820.mp3

BACKGROUND Acute kidney injury (AKI) is a common, life-threatening complication among hospitalized patients. Timely prediction is crucial for prevention, while accurate prognosis following onset is essential for guiding care. Although electronic health records (EHRs) provide rich longitudinal data, their high dimensionality, heterogeneity, and temporal complexity pose challenges for traditional machine learning. OBJECTIVE The aim of this study was to develop and evaluate a transformer-based model for generalizable AKI prediction and prognosis by modeling long-range temporal dependencies and complex relationships in heterogeneous EHR data and to examine its interpretability. METHODS We developed the Biomarker-Aware Transformer for AKI (BAT-AKI), which integrates multidimensional embeddings capturing temporal dynamics, textual descriptions, hierarchical ontology structures, and semantic relationships in EHRs. To enhance sensitivity to AKI-related biomarker fluctuations, BAT-AKI combines masked language modeling with a novel biomarker abnormality-aware pretraining task. The model was pretrained and fine-tuned for 4 tasks: AKI onset, early reversal, recovery at discharge, and 90-day post-AKI mortality. Pretraining and internal validation used data from the National Patient-Centered Clinical Research Network (PCORnet) site, with external validation across 3 additional sites. RESULTS BAT-AKI achieved the strongest overall discrimination, outperforming baseline models in area under the receiver operating characteristic curve (AUROC) and area under the precision-recall curve (AUPRC) on most tasks. For AKI onset prediction, it achieved an AUROC of 0.8549 (95% CI 0.8433-0.8658) and an AUPRC of 0.6041 (95% CI 0.5786-0.6292), exceeding the second-best baseline by 2.8% in AUROC and 8.1% in AUPRC. Across prognostic tasks, BAT-AKI achieved AUROCs of 0.6006 (95% CI 0.5703-0.6318), 0.6080 (95% CI 0.5745-0.6397), and 0.8017 (95% CI 0.7708-0.8311) and AUPRCs of 0.7036 (95% CI 0.6690-0.7388), 0.7417 (95% CI 0.7082-0.7765), and 0.5063 (95% CI 0.4422-0.5684) for early reversal, recovery at discharge, and 90-day post-AKI mortality, respectively, corresponding to relative gains of 2.4% to 6% in AUROC and 3.7% to 4.1% in AUPRC over the second-best baseline. Performance remained stable across external cohorts, with AKI onset AUROCs ranging from 0.7754 (95% CI 0.7677-0.7827) to 0.7989 (95% CI 0.7885-0.8092) and AUPRCs ranging from 0.4777 (95% CI 0.4631-0.4924) to 0.5687 (95% CI 0.5495-0.5871), corresponding to relative improvements of 3.7% to 5.3% in AUROC and 16.8% to 25.3% in AUPRC over the transformer baseline. In a low-resource setting using only 2% of the fine-tuning data, BAT-AKI achieved an AUROC of 0.7898 (95% CI 0.7759-0.8036) and an AUPRC of 0.4903 (95% CI 0.4640-0.5178) for AKI onset prediction, exceeding the second-best baseline by 4.2% in AUROC and 10.9% in AUPRC. Visualization of embeddings, attention patterns, and integrated-gradient feature attributions highlighted clinically meaningful relationships and patient-specific trajectories. CONCLUSIONS BAT-AKI provides generalizable AKI prediction and prognosis with consistent discrimination across institutions, while offering interpretability insights into its predictions, supporting its potential for early warning and AKI care optimization. CLINICALTRIAL

Key Points: Machine learning models predicted short-term AKI reversal and progression with area under the receiver operating characteristic curves of 0.79–0.93 across four health systems. Top predictors common to both outcomes included serum creatinine, systolic BP, and albumin, with consistent patterns across diverse sites. Early AKI recovery is often transient, with frequent relapse, highlighting the need for continued monitoring after initial improvement. Background: AKI affects 10%–25% of hospitalized patients and is associated with significant morbidity, mortality, and long-term adverse outcomes. Recovery patterns vary greatly, from full reversal to progression toward CKD. Predicting short-term AKI states and understanding their dynamic trajectories could help optimize clinical interventions. Methods: This retrospective study used data from four health care systems in the Greater Plains Collaborative, spanning 2009–2022. A total of 94,531 inpatient encounters (54% male, 80% White patients) from 75,312 adults with admission eGFR ≥15 ml/min per 1.73 m ² and without prior dialysis or kidney failure were included. We developed CatBoost models, a gradient-boosting decision-tree algorithm, to predict 7-day AKI progression and reversal and then performed multistate modeling to estimate transition intensities and covariate effects on AKI state changes. Results: The models achieved strong performance across outcomes and subgroups: area under the receiver operating characteristic curve 0.79–0.89 and accuracy 0.7–0.79 for reversal; area under the receiver operating characteristic curve 0.91–0.93 and accuracy 0.83–0.87 for progression. Serum creatinine, systolic BP (SBP), and albumin were key predictors across both outcomes. Low SBP (≤120 mm Hg) predicted increased likelihood of progression, whereas high SBP (≥140 mm Hg) predicted asymmetric recovery with reduced chances of subsequent reversal. Multistate analysis highlighted the dynamic and transient nature of AKI recovery, with nearly half of AKI-1 patients (instantaneous hazard rate=0.46) transitioning to “no AKI” peaking on the first day postonset. Conclusions: This study presents a two-stage analytical framework linking short-term prediction with multistate modeling to characterize early recovery and progression patterns after hospital-acquired AKI. The approach highlights a small set of routinely measured variables, such as serum creatinine trajectories and BP trends, that consistently track these dynamics across health systems. While the results are hypothesis-generating and require external, prospective validation, they lay groundwork for future studies on refining risk stratification and exploring targeted interventions in AKI care.

Key Points: Data were collected from a cohort of 759 patients with autosomal dominant polycystic kidney disease that were followed for a median of 10.2 years. Prognostic models were developed to predict kidney failure using routinely obtained clinical information and externally validated. Risk predictions had good discrimination and calibration over a time horizon of 15 years. Background: Autosomal dominant polycystic kidney disease (ADPKD) is a common cause of kidney failure. Progression is highly variable, and accurate prognostic information is needed to guide early treatment decisions. The objective of this study was to develop a multivariable predictive model for progression to kidney failure. Methods: We developed prognostic models using Cox regression for the outcome of kidney failure, defined as eGFR 2 , dialysis, or kidney transplantation. The development dataset consisted of participants in the Consortium for Radiologic Imaging Studies of Polycystic Kidney Disease (CRISP) and Halt Progression of Polycystic Kidney Disease study A (HALT-A) studies. The validation dataset consisted of patients with ADPKD in the Mayo clinical registry aged 15–49 years, with eGFR ≥60 ml/min per 1.73 m ² . Predictor variables in the base model were age, sex, creatinine or eGFR, ADPKD genotype, and Mayo Imaging Class. Clinical and laboratory data were evaluated in the full model. Results: The development cohort included 759 patients with baseline eGFR of 91±29 ml/min per 1.73 m ² (mean±SD), of whom 16% reached end point after median (interquartile range) follow-up of 10.2 (5.5–16.7) years. The validation cohort included 535 patients with baseline eGFR of 90±31 ml/min per 1.73 m ² (mean±SD), of whom 11% reached end point after median (interquartile range) follow-up of 5.5 (1.5–13.7) years. The full model, including age, sex, serum creatinine, Mayo Imaging Class, total carbon dioxide, hemoglobin, diastolic BP, and body mass index, had a C-index of 0.81 (95% confidence interval, 0.72 to 0.90) in the validation cohort and 0.75 (95% confidence interval, 0.62 to 0.88) when ADPKD genotype was also included. Risk predictions from base and full models were well-calibrated out to 15 years. Conclusions: In persons with early ADPKD and preserved eGFR, a model using routinely obtained clinical information quantified risk of kidney failure over 15 years.

Key Points: TraceOrg is a web-based tool that automatically labels kidney, liver, and cysts, reporting volumes and Mayo Imaging Classification. External validation showed high performance and good generalizability on Consortium for Radiologic Imaging Studies of Polycystic Kidney Disease, Polycystic Kidney Disease-Research Resource Consortium, and other external datasets. Training on multiple pulse sequences enables TraceOrg to process images from a wide variety of protocols. Background: Kidney, liver, and cyst volumes are important for diagnosis, classification, and management of autosomal dominant polycystic kidney disease (ADPKD) but challenging to measure accurately and reproducibly. Here, we develop a web-based deep learning platform to automatically and robustly measure kidneys, liver, and cyst volumes in ADPKD. Methods: Magnetic resonance imaging (MRI) and computed tomography scans from patients with ADPKD ( n =611) and participants without ADPKD ( n =109) were used to train a 3D hybrid model combining U-Net and transformer elements for segmenting kidneys, liver, and cysts. The model is implemented as a web-based calculator at www.traceorg.com, providing segmentation labels, volumes, and Mayo Clinic Image Classification. Automatic browser anonymization of digital imaging and communications in medicine images ensures privacy. Internal validation was conducted on 70 MRIs for kidney and liver segmentations and 46 MRIs for cyst segmentations, and performance was compared with five open access segmentation models (TotalSegmentator, MRAnnotator, Kim, Woznicki, and Gregory–Kline). External validation was performed on one single-center dataset ( n =58), one multicenter dataset ( n =73), Consortium for Radiologic Imaging Studies of Polycystic Kidney Disease 2 (CRISP, n =30), and Polycystic Kidney Disease-Research Resource Consortium (PKD-RRC, n =115) MRIs with T2-weighted and T1-weighted images. Results: After training on 720 participants (mean age=48±15, eGFR=74±32 ml/min per 1.73 m ² and height-adjusted total kidney volume=826±772 ml/m), TraceOrg internal validation performance achieved high mean Dice scores of 0.97 (kidneys), 0.97 (liver), 0.93 (kidney cysts), and 0.82 (liver cysts) outperforming existing models for ADPKD. External validation showed strong performance with Dice scores of 0.92–0.94 (kidney), 0.87–0.96 (liver), 0.85 (kidney cysts), and 0.76–0.90 (liver cysts) for the single-center dataset and 0.95 (kidney) and 0.81 (kidney cysts) for the multicenter dataset. Compared with CRISP volumes measured by stereology, the mean absolute percent difference was 5.3% (kidneys, n =30), 11% (kidney cysts, n =30), and 5.5% (liver, n =22). Compared with PKD-RRC ( n =115), the mean absolute percent difference in total kidney volume was 4.9%. Conclusions: TraceOrg, a publicly available web-based tool, automatically measured kidney, liver, and cyst volumes from abdominal MRI in ADPKD with high accuracy compared with manual segmentations. Podcast: This article contains a podcast at https://dts.podtrac.com/redirect.mp3/www.asn-online.org/media/podcast/JASN/2026_02_03_ASN0000000904.mp3