Correlation of in vivo imaging to morphomolecular pathology in translational research: challenge accepted

Because of a variety of new technologies during the last years [2, 3], which made it possible to create complex and genetically modulated animal models with the aim to mimic human diseases on a molecular as well as on a morphological level, the number of new animals models, especially mouse models, increases steadily. Besides expected similarities, genetically engineered animal models often show divergences or unexpected alterations resulting in background lesions which easily can falsify results. Therefore, profound knowledge of strain specific background lesions as well as classifications of the occurring pathological alterations in a comparative way is an essential component of translational research. Indeed, there are several examples in the literature where the lack of pathologists from histopathological evaluations of mouse models led to considerable errors in description and diagnosis of histopathological findings [4, 5].

The term “comparative pathology” is defined as pathology of analogous diseases in different species [6]. It can be divided in 3 domains depending on the field of application:

In translational research, comparative pathology is usually used synonymously to experimental animal-based pathology.

Several authors pointed out the looming need of pathologists focusing on comparative aspects for handling the increased use of genetically engineered mice [10‐12]. For the last 10 years, the first units and grant programs were dedicated to this topic of research [13, 14]. Still, the number of comparative pathology units in Europe is small, especially when it comes to a direct embedding into an academic pathology department. Some more units have been established in the USA: several centers for comparative pathology are located at veterinary schools, and also some few units are integrated into departments of pathology of medical faculties.

The main reasons for the poor transferability of results in animal model systems to clinical applications with respect to pathology are: 1. missing or incomplete pathologic description/evaluation, 2. lack of current comparative classifications of animal models, and 3. lack in pathologists with appropriate training in exercise of comparative issues [4]. With respect to the correlation of in vivo imaging methods, one could add also a deficiency in the development of protocols and workflows allowing for an appropriate correlation of in vivo imaging techniques with the underlying histomorphological background. In particular, the adequate correlation of images from different in vivo imaging techniques to histological slides can be challenging and requires a high degree of interdisciplinary cooperation. To enable for an evaluation of corresponding planes and the maintenance of 3D-orientation, often different steps of protocol adaptations are needed to achieve satisfying results.

The focus of CRC824 research groups is on oncology. Translational oncological research uses different systems of animal models resulting in different levels of complexity which influence comparability and reproducibility [6]. Table 1 provides an overview of commonly used types of animal models in cancer research.

Given the fact that there is a lack of surgical pathologists with background knowledge of animal model pathology and in reverse only very few veterinary pathologists are focusing on translational and imaging issues, it became clear that there is an urgent need bringing the two disciplines together. Besides the appropriate planning and evaluation of animal model-based experiments, a major task of comparative facilities lies in the correct and effective training of pathologists. Additionally, with respect to animal protection and ethics, it must be the aim to decrease animal numbers in research. By planning and evaluating experiments in a scientifically sound way, repetitions and false conclusions can be prevented. The ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines as well as the 3R (Replace, Reduce, Refine) principles have to be considered. The majority of animal-based research to date is performed with mice due to their fast reproduction rate and their high availability for genetic engineering methods. There are two main aspects that have to be considered when choosing the mouse as a model organism: 1. characterizing the animal on a morphological and molecular level to ensure that the model used is suitable for the question and 2. correct planning including necropsy and slide preparation as well as consultation of pathologists with expertise in animal model pathology to ensure reproducibility [4, 15].

These circumstances resulted in the establishment of the CEP in 2016, a veterinary pathology unit embedded in the Institute of Pathology at the Technical University of Munich (TUM) consisting of a technical and a scientific part. The CRC824 was the first funding where the CEP unit was involved as a central project (Z-project), meaning a scientific service unit for the research groups included in the CRC824 and focusing on correlation of in vivo imaging techniques to histomorphological evaluation. The CEP rapidly developed and increased within the setting of the CRC824 and became one of the largest comparative pathology facilities in Europe. Started with a basic equipment, it now expanded to a fully equipped and automated histopathological laboratory which enables for processing of the high amount of requests in an appropriate time and way. Its unique feature is the embedding at the Institute of Pathology at the TUM together with a complex network of a broad field of translational research groups focusing on in vivo multimodal molecular imaging. Over the last years, several additional consortial fundings have been added.

Our performance monitoring for the period of 2018 to 2020 showed a total number of processed projects between 500 and 600 per year. Thereof, most projects were mouse studies which accounted for ~ 50% every year (2018: 50.5%, 2019: 48.8% and 2020: 54.0%). Translational projects aiming at comparative evaluation of human tissue specimen also made up a major part of our work: around 40% of projects per year requested comparative human evaluation (2018: 41.6%, 2019: 44.0% and 2020: 39.3%). Rat tissue requests had a lower percentage ranging from 1.7% to 3.2% in the given period. The remaining specimen (around 5% every year) originated from various other species. A detailed analysis of the project requests in 2020 (despite a slight drop due to the Coronavirus pandemic) is shown in Fig. 1. In 2020, in total 521 projects were processed with > 22.500 submissions (ranging from unfixed tissue to already available empty cuts). Nearly 50.000 cuts from formalin-fixed paraffin-embedded (FFPE) and cryo tissue were provided, and in total > 32.000 slides were digitalized. This is not surprising as digitalization in pathology nowadays belongs to standard processes, at least when it comes to research pathology. Technical service was performed in 100% of all projects while scientific support was requested in 75.8%. The technical workup has been performed by 8 technicians employed during the time period analyzed.

For scientific requests, we provided a well-trained team of veterinary pathologists, veterinary pathologists in training and digital pathology expertise. From the surgical pathology site, 9 pathologists (8 consultants, 1 resident) were involved in CEP-projects.

As a central project in the CRC824, technical and scientific support of the CEP mainly focused on correlating in vivo imaging to histomorphology.

Basically, co-registrations of in vivo imaging techniques with histomorphological features include exact embedding in the right plane of interest and appropriate labeling of the relevant sites. Besides, it is obligate to ensure appropriate necropsy and fixation of samples. For this reason, pathologists should be involved before starting in vivo imaging or at least before planning of animal necropsy so that the workflow and processing of samples can be adapted in an appropriate way. Table 2 shows the benefit of interdisciplinary cooperation for both sides, researchers and pathologists. As any other research technique, histomorphological sample processing also requires certain defined restriction with respect to time and has to stick to a schedule. It should be avoided to unnecessarily prolong necropsy (because of relatively fast beginning autolysis of tissue) or inadequate or prolonged formalin fixation (because of affecting results of immunohistochemistry). The recommended standard fixation is 10% formalin at room temperature for 48 h. Besides, the ratio of tissue to formalin should be 1:10. Tissue should then be embedded into paraffin immediately without any delay or further change of fixation solutions. In cases where radiation decay is necessary because of the use of radionuclides for imaging, the pathology unit should be contacted in advance to discuss the most appropriate workflow with respect to ensure for the best tissue preservation and subsequent processing for the read-outs needed. It should be avoided to freeze organs for storage and later fix them in formalin for FFPE diagnostics. If necessary, the tissue should be thawed in 10% formalin at 4 °C.

During the CRC824 funding period, several in vivo–ex vivo co-localization approaches were successfully applied to different issues depending on the complexity of the research question. Basic and quite simple co-localization can be done for easily accessible tumors, like subcutaneous neoplasias, whereas tissue-specific endogenous tumors, like endogenous abdominal neoplasias, are more challenging and require a more complex and difficult workflow for satisfying co-registration results.

From the pathologists point of view, for all co-localization issues which include histomorphology and immunohistochemistry, processing of FFPE tissue should always be preferred to cryo sectioning. In cryo sections, cell morphology is not so well maintained and cryo artifacts can interfere with co-localization results. If cryo sectioning is necessary due to a local setup (i.e., availability of a cryotome in the radiation protection area, but lack of a paraffin embedding possibility there), the cryo tissue should be stored unprocessed in liquid nitrogen or embedded in cryo cassettes with an appropriate embedding material and be stored at − 80 °C.

LSFM allows a three-dimensional fluorescence based visualization of whole tissue samples. Two thin counterpropagating sheets of laser light illuminate a thin layer of the sample [26]. Therefore, the tissue has to be rendered transparent to enable the laser light to penetrate the sample (optical tissue clearing). Different clearing protocols have been developed during the last years [27, 28]. By adding LSFM as a mesoscopic imaging tool, it is possible to close the gap between macroscopic and microscopic imaging. Mechanical sectioning of tissue destroys tissue integrity resulting in loss of important information [29]. LSFM allows three-dimensional imaging of intact tissue specimen with micrometer resolution without microtome slicing [26].

Clearing of samples was done using the 3DISCO (3D imaging of solvent-cleared organs) protocol [29]. After imaging with a Light Sheet Fluorescence Microscope (UltraMicroscope (UM) II, LaVision BioTech, Germany), samples were rehydrated by immersing them in descending concentrations of ethanol solutions and finally dehydrated again and embedded in paraffin, maintaining the plane of interest corresponding to UM imaging. Serial sections were done in 50-µm steps to ensure co-localization with UM images [30]. After hematoxylin and eosin (H&E) staining of sections, the slides were co-localized with single images from the UM stack. Accordingly, adrenals of multiple endocrine neoplasia X (MENX)-affected rats which were first labeled with T-lectin for visualizing vessels with LSFM were rehydrated and embedded for co-localization of CD31 and Ki67 after immunohistochemistry to characterize pheochromocytomas as comparable to the pseudohypoxic cluster in humans [31].

The last decades have shown that genetically engineered animal models are highly adaptable to a great variety of research questions but bring along the difficulty to translate their findings to the human setting. Therefore, the field of comparative pathology together with correlated in vivo imaging became more important. Meanwhile, several authors emphasized the need of pathologists with comparative knowledge focusing on translational research and animal model pathology [4, 10, 32]. As a result, some comparative pathology units developed during the last decade, for example, in Europe the COMparative PATHology platform (COMPATH) at the University of Bern and the Centre for Comparative Pathology at the University of Edinburgh or in the USA the University of Iowa (Comparative Pathology Laboratory) and the Vanderbilt University (Translational Pathology Shared Source—Comparative Pathology). To best of our knowledge, the integrated setup with a comparative pathology unit fully integrated in an Institute of Pathology within medical faculty is still unique in Europe. The CEP unit extended to a fully equipped experimental comparative pathology laboratory for technical and scientific support for research groups with focus on cancer research and therein developed several approaches for correlation of in vivo imaging to histomorphology. Taken together, the embedding of comparative pathology units in university institutes can help to improve and fasten animal model evaluation and can have a major impact on translational research in future (Table 2).