Imaging and spatial omics of kidney injury: significance, challenges, advances and perspectives

Diagnosis

In the clinical field, four major demands for full achievement of early diagnosis, etiological diagnosis, prognosis evaluation, and the right timing of treatment decisions are increasing. However, the existing methods and strategies for AKI diagnosis, such as clinical imaging and biomarkers, are far from developed. Renal pathology is still the leading strategy and gold standard of diagnosis. However, clinical kidney biopsy is highly traumatic and limited in sampling. Excepting minor complications of transient bleeding, fever, hematuria, proteinuria and transient hypertension, major complications such as puncture of nearby organs or structures, infection near the biopsy site, development of arteriovenous fistula or pseudoaneurysm happen occasionally. Cases of nephrectomy or even death after renal needle biopsy are also factors that cannot be ignored in some extreme cases. As a result, it is not feasible for dynamic assessment of kidney injury or being performed in patients with severe conditions or with contraindications. Even after a successful and safe renal puncture, the low depth and small volume of biopsied samples disabled the pathologists from having a global view of the kidney and therefore may miss important disease information. These gave rise to a relatively low diagnosis rate and a limited precision of etiology identification [4]. In this case, it is in urgent medical need and of great clinical significance to develop a novel clinically practicable diagnosis strategy by imaging with high standard of non-invasion, dynamics and continuity.

Treatment

In the aspect of treatment, another sacred realm of kidney injury research is to interpret the elaborate mechanisms underlying the development and progression of long-lasting injury, since no successful drug therapy exists for AKI or reversing kidney fibrosis after so many years of mechanism study [1]. Besides the novel molecular targets are required to be identified urgently, their function in the time-series from the beginning to the end should also be elucidated clearly for a proper decision of treatment window. It is most meaningful to depict the mechanisms of kidney injury systematically, panoramically and comprehensively.

X-ray imaging

X-rays are less commonly used in kidney disease due to low soft tissue contrast and radiation concerns for CT, yet their non-invasive and dynamic advantages make them a potentially strong competitor. The application of micro-CT in kidney imaging still relies on the various contrast agents for better soft tissue contrast now. With the help of blood vessel contrast Microfil, it is possible to reconstruct vascular tree structure with 3 μm resolution and reveal the reduction of vascular density in vitro following injury. Besides contrast agents, another way of enhancing contrast is to use synchrotron radiation X-ray (SRX). With characteristics of strong penetration, high sensitivity, multi-scale resolution and low radiation dose, it is widely used in multi-scale, high-sensitivity, low-dose and non-destructive imaging of complete organs of humans and model animals. The 3rd generation of SRX by European Synchrotron Radiation Facility (ESRF) enabled hierarchical phase-contrast tomography (HiP-CT), which further gave rise to the first cross-scale scan of an intact human kidney [7]. Through combining the breakthrough of virtual histology (VH) image digital conversion technology, it allowed construction a panoramic nondestructive 3D pathology map of kidney. The utility of SRX-based HiP-CT may play a vital role in revealing the physiological microstructure, as well as its pathophysiological alteration in time series of kidney injury. Thus, X-ray imaging has the potential of construction a brand-new and revolutionary 3D pathology diagnosis system for kidney diseases, as well as a microstructure atlas for mechanism research.

Molecular imaging

Real-time and non-invasive in vivo monitoring of the molecular-level pathophysiological status within renal tissues is crucial for understanding diseases and developing therapeutic targets. Currently, commonly used clinical methods, such as non-invasive measurements of glomerular filtration rate (GFR), urine sediment analysis as well as invasive renal biopsy fail to achieve the goal. Molecular imaging techniques offer the potential to observe molecular processes and blood supply inside the kidney by injecting specific radioactive substances, contrast agents, specific molecular probes or material tracers which are expected to provide more accurate and personalized functional information for the diagnosis and treatment of kidney diseases. Molecular imaging can utilize various imaging modalities such as MRI, PET-CT, single-photon emission computed tomography (SPECT) and fluorescence molecular tomography (FMT). Molecular renal probes (MRPs) are used for labeling early biomarkers of AKI, which enable long-term real-time imaging of multiple molecular events in the kidneys of live mice and further monitor the onset, progress and outcome of kidney injury [8]. Detailed molecular imaging probes in previous studies have been addressed elsewhere [9]. Besides molecular imaging, imaging based on advanced material is also rising and playing more roles in AKI diagnosis. Novel microbubbles tracked by super-resolution ultrasound (SRUS) imaging as they flow with the blood, also allowed for non-invasive in vivo evaluation of vascular function and structure following ischemia-induced AKI or various types of glomerular injury [10]. Focusing on spatial heterogeneity, cationized ferritin-enhanced magnetic resonance imaging (CFE-MRI) was used to examined the three-dimensional spatial distribution of nephron density and calculation of single nephron glomerular filtration rate (snGFR) [11]. Thus, the advanced molecular imaging facilitated the detection of focal lesions that are distributed irregularly, and make possible of the mechanism research behind. The booming of probe and material development have made molecular imaging expandable and more powerful in monitoring kidney injury precisely.

Spatial imaging omics

Single-cell omics such as RNA sequencing have tremendously accelerated the mechanism studies of kidney injury in the last decade, yet the failure of retaining spatial information of cells impeded the precise interpretation of microenvironment, spatial neighborhood and niche networks of kidney injury. The first spatial omics atlas of kidney injury was generated by the transcriptomics solution from the 10× Genomics Visium Spatial Gene Expression. Spatial molecular expression patterns of repair and injury response to ischemic insult could easily be defined and uncovered by spatial clustering analysis and identification of spatially variable features. Deconvolution of spatial atlas also provided a preliminary solution to detection of key cell-cell interaction pattern in the elaborate injured micro-environment [12]. Another spatial atlas by matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MSI) gave the first metabolomics landscape of AKI. Combined with multiplexed immunofluorescence staining to label specific tubular segment or state, metabolic signature of region-specific cell types related to repair and maladaptive repair could be identified, and the trajectories of metabolic pathway alteration along the kidney injury progression were able to be determined [13]. However, in exchange, implementation of the large field of Visium and MALDI-MSI sacrifices the resolution of single cell, which is most precious for precise depiction of molecular signature, fate trajectories and interaction patterns between spatially variable cell types or even subgroups. At present, advances in spatial technologies are breaking the borderline between imaging and sequencing-based omics, which truly catalyze the birth of spatially single-cell resolved detection. Technologies of novel single-cell resolved spatial imaging and omics that are promising in kidney injury research were summarized (Table 1). As was displayed, without losing spatial resolution or large field of view (FOV), these technologies are gradually developed to have higher throughput and better compatibility with proteomics or metabolomics, no matter within the same slice or by sequential slices. Better compatibility with formalin fixation and paraffin embedding (FFPE) slices also makes possible of establishment of kidney injury cohort. This will greatly fill the huge gap of prognostic research in the field. Outcome-related cell populations, cell-cell interaction patterns and micro-environment signature will be uncovered to provide a clearer direction for mechanism research. In conclusion, with a better balance of depth, scale, throughput, field and resolution, application of spatial imaging omics in kidney injury will overcome the challenge of heterogeneity and complexity with less hindrance.

3D tissue imaging

3D visualization of renal microstructure is primarily limited to confocal microscopy. However, the refractive properties of protein and lipid components within the kidney scatter light and significantly degrade image quality with increasing depth. Optical tissue clearing methods have been employed to address this issue by removing pigments, lipids, and other light-scattering substances from tissues, allowing light to better penetrate the kidney. Combined with deep tissue labeling strategies and the development of optical projection tomography and multiple fluorescence microscopic imaging technologies such as confocal, multiphoton or light sheet, 3D scalable renal studies of cells and molecules have become practicable. Besides traditional tissue clearing methods such as the 3D imaging of solvent-cleared organs (DISCO) family, clear lipid-exchanged acrylamide-hybridized rigid imaging/immunostaining/in situ hybridization-compatible tissue-hydrogel (CLARITY) and clear, unobstructed brain/body imaging cocktails and computational analysis (CUBIC), novel strategies like small-micelle-mediated human organ efficient clearing and labeling (SHANEL), raman dye imaging and tissue clearing (RADIANT) and nanobody (V_HH)-boosted 3D imaging of solvent-cleared organs (vDISCO) are developed to be compatible with depth and multiplexity [25]. Now it is feasible to achieve multiplex co-localization with about 30 channels within an intact human kidney or within an entire mouse. Technical compatibility with subsequent pathology and transmission electron microscopy (TEM) allows for integration structural information [26]. By combining molecules indicating segment or cell type with markers of injury-induced heterogenous pathological function, future 3D multiplex renal imaging will be vastly helpful in resolving heterogeneity obstacles caused by segment and spatial location in mechanism studies of kidney injury.

Multimodal integration of imaging and omics

Implementation of multi-modal imaging and omics integration of tissue structure, renal function and molecular information enables the possibility to interpret intricate microenvironments of kidney injury. However, the methodology and technology roadmap for multi-modal integration are spiny and bumpy. With this initiative, an attempt was made and provided us with precious practices of organizing multimodal imaging and omics to achieve a systematic description of nephropathy mechanisms. The Kidney Precision Medicine Project (KPMP) focuses on AKI and CKD, with the goal of establishing a human kidney tissue map with spatial localization at single-cell resolution in both healthy and diseased conditions. KPMP integrated various single-cell sequencing, single-nucleus sequencing, and spatial imaging technologies such as multiplexed protein imaging and metabolic mass spectrometry imaging in the same renal sample to construct the most comprehensive human kidney map to date [27]. A total of 51 major cell types were identified, and 28 cell states during the process of kidney injury were defined in this research. Distinct spatial localization of protein forms in the vasculature, medulla, cortex regions as well as injury-related cellular neighborhoods and co-expression network within the kidney were sufficiently revealed. Spatial imaging and omics techniques were employed to localize these states within the vicinity of the injury. For example, 3D imaging techniques such as light sheet fluorescence microscopy (LSFM) were utilized to provide information about the immune microenvironment in the injured kidney. This study serves as a paradigm and provides tools for multi-modal imaging analysis of the kidney.

Creating a comprehensive, 3D spatially resolved cellular and molecular atlas is required for achieving the final goal of establishing and developing systematic diagnosis and evaluation system of kidney injury. Though not with flawless integration, KPMP is an encouraging attempt of great value in providing us with the first paradigm and many opportunities for reaching this landmark. Indeed, overcoming the above challenges is still highly required for the fulfillment of this goal.