Longitudinal three-dimensional visualisation of autoimmune diabetes by functional optical coherence imaging
- 1.3k Downloads
It is generally accepted that structural and functional quantitative imaging of individual islets would be beneficial to elucidate the pathogenesis of type 1 diabetes. We here introduce functional optical coherence imaging (FOCI) for fast, label-free monitoring of beta cell destruction and associated alterations of islet vascularisation.
NOD mouse and human islets transplanted into the anterior chamber of the eye (ACE) were imaged with FOCI, in which the optical contrast of FOCI is based on intrinsic variations of the index of refraction resulting in a faster tomographic acquisition. In addition, the phase sensitivity allows simultaneous label-free acquisition of vascularisation.
We demonstrate that FOCI allows longitudinal quantification of progressive autoimmune insulitis, including the three-dimensional quantification of beta cell volume, inflammation and vascularisation. The substantially increased backscattering of islets is dominated by the insulin–zinc nanocrystals in the beta cell granules. This translates into a high specificity for the functional beta cell volume of islets. Applying FOCI to a spontaneous mouse model of type 1 diabetes, we quantify the modifications of the pancreatic microvasculature accompanying the progression of diabetes and reveal a strong correlation between increasing insulitis and density of the vascular network of the islet.
FOCI provides a novel imaging technique for investigating functional and structural diabetes-induced alterations of the islets. The label-free detection of beta cell volume and infiltration together with vascularisation offers a unique extension to study ACE-transplanted human islets. These results are contributing to a deeper understanding of human islet transplant rejection and label-free in vivo monitoring of drug efficacy.
Keywords3D visualisation Beta cell volume Human islets Inflammation Label-free Longitudinal NOD mouse Non-invasive OCM Quantification Vasculature
Anterior chamber of the eye
Functional optical coherence imaging
Forkhead box P3
Green fluorescent protein
High endothelial venule
Mucosal addressin cell adhesion molecule 1
Optical coherence microscopy
Optical projection tomography
Diabetes mellitus develops as a functional impairment in insulin production, sometimes in association with insulin resistance. In both major types of diabetes mellitus, the progressive dysfunction of the beta cell causes the disease development. While type 1 diabetes is the result of an autoimmune attack on beta cells, type 2 diabetes is considered to be driven by metabolic factors associated with sedentary lifestyle and obesity, albeit with accumulating evidence of low-grade inflammation. Modifications of the pancreatic microvasculature are likely to accompany the progression of both types of diabetes. Alterations in vascular variables, such as transient vasoconstriction, vasodilation, increased blood flow and vascular leakage, are necessary preludes to inflammation as they orchestrate the influx of diverse cell types and affect local homeostasis [1, 2, 3, 4, 5, 6]. To fully appreciate how these events contribute to the pathogenesis, improved methods allowing longitudinal, high-resolution monitoring of vascularisation and affected tissues close to the cellular level during its natural progression are warranted and constitute the key objectives of this study.
To date, longitudinal non-invasive, intra-vital imaging of the pancreas has been mainly restricted to non-optical imaging modalities such as MRI and computed tomography (CT). While the medical potential of these modalities is indisputable, they are limited by a restricted set of reagents targeting specific tissues and by a relatively low spatial resolution [4, 7, 8]. Optical microscopy on the other hand is mainly limited to preclinical studies because of a reduced penetration depth [9, 10, 11, 12]. Optical projection tomography (OPT) has been used to image the adult mouse pancreas and retrieve the three-dimensional (3D) and undistorted structure of the tissue . While OPT can quantify inflammation at different stages of disease [14, 15], it is limited by being an ex vivo technique that requires sample fixation and immunolabelling.
In accordance with these objectives, we opted for a non-invasive imaging method, extended-focus optical coherence microscopy (xfOCM) , which allows label-free longitudinal visualisation and quantification of pancreatic ducts, blood vessels and individual islets of Langerhans ex vivo and in vivo. The technique requires only a two-dimensional scan to obtain a 3D image. This intrinsic multiplex advantage results in a fast parallel acquisition of entire depth profiles without depth scanning, with an in vivo penetration depth of ∼300 μm in the pancreas. However, optical pancreas imaging requires laparotomy, restricting this technique to transversal studies [17, 18, 19]. Recently, the anterior chamber of the eye (ACE) has been used to study transplanted pancreatic islets [20, 21, 22]. Engrafted on the iris, islets could now be repeatedly imaged using the mouse eye as a natural body window. Using this approach, we elucidate important structural and functional alterations induced during the progression of autoimmune diabetes.
Mice of genotype C57BL/6 (B6).Rag2−/− were backcrossed with NOD mice for the generation of NOD.Rag2−/− mice, as previously described . NOD.Foxp3-green fluorescent protein (GFP) mice were generated by speed-congenic backcrossing of the B6.Foxp3-GFP (Jax stock number 006772) with NOD mice, as previously described . BALB/c mice were purchased from Taconic (Ejby, Denmark). All animals were bred and maintained in a specific pathogen-free environment at the animal facilities at Lund University. Guinea pigs were purchased from Harlan Laboratories (Indianapolis, IN, USA). Mice with ZnT8 (also known as Slc30a8) knockout (KO) and littermates were generated as previously described . The Ethics Committee at Lund University and the Swiss cantonal veterinary authorities approved all experimental animal procedures. Experiments were performed after mice were genotyped, without randomisation or blinding. No data were specifically included or excluded.
Pancreatic islet isolation, culture and ACE transplantation
For experiments, including the study of healthy control islet grafts, 6–8-week old BALB/c or B6.Foxp3-GFP recipient mice were ACE transplanted with islets from BALB/c or B6.Rag2 −/− donors, respectively. For inflamed islet grafts, 10-week-old NOD.Foxp3-GFP recipient mice were ACE transplanted with islets from NOD.Rag2 −/− donors. Human islets from non-diabetic cadaver donors were provided by the Nordic Islet Transplantation Program (www.nordicislets.org), coordinated by O. Korsgren, Uppsala University, and were transplanted into 6–8-week-old female NOD.Rag2 −/− mice (see the electronic supplementary materials [ESM] Methods for details). The study was approved by the Ethics Committee at Lund University.
Functional optical coherence imaging
Mouse head and eyeball were restrained as previously described  and anaesthesia was maintained at 1% isoflurane. Functional optical coherence imaging (FOCI) integrates the xfOCM instrument equipped with a Zeiss Neofluar objective × 10; NA 0.3, Carl Zeiss (Jena, Germany) with an improved acquisition (customised optical coherence microscopy [OCM] spectrometer) and optimised scanning modality for longitudinal functional imaging. A fast processing unit allows real-time monitoring (structure and blood flow), with a fully equipped platform for small-animal imaging. Details of the xfOCM instrument, data acquisition, scanning protocols and analysis are given in the ESM Methods.
Graft-bearing eyes or pancreas were isolated from mice after perfusion with 4% paraformaldehyde/PBS, cryosectioned parallel to the iris (section thickness 8–10 μm) and subjected to immunohistochemical staining as previously described  (see ESM Methods).
Ex vivo imaging with dark-field optical coherence microscopy
Groups were compared by the Mann–Whitney test or unpaired Student’s t test, as indicated in the respective figure legend, and data are expressed as means ± SEM.
Beta cell specificity in OCM
Longitudinal quantification of beta cell volume and islet vascularisation using FOCI
Using FOCI, we next visualised and quantified the microvasculature in the transplanted islets over a 2 month period. As illustrated in Fig. 2a–d, f, we found that, in parallel with islet growth, re-vascularisation of the grafted islet occurred rapidly over the first 2 weeks to form a capillary network.
FOCI detects inflammation in pancreatic islets under autoimmune attack
We next applied FOCI to follow the progressive autoimmune destruction of beta cells in the NOD mouse model for type 1 diabetes. To determine if the cells infiltrating the ACE-transplanted islets could be visualised in a label-free manner using FOCI, we next analysed 14-week-old non-diabetic NOD mice transplanted with syngeneic islets 4 weeks prior to analysis.
Insulitis correlates with attenuated vascularisation density
As expected, we could visualise a gradual increase in the total islet vascular network shortly after transplantation (ESM Fig. 7). Counter intuitively, we noted a decrease in the microvasculature density in infiltrated regions of inflamed islets, with only a few big vessels remaining (Fig. 7a, g, h, Fig. 4d and ESM Videos 1–3). This decrease in the microvasculature density was not observed in islets displaying only limited infiltration (ESM Fig. 7 and ESM Video 4), indicating that the observed attenuation of microvasculature correlates with the severity of inflammation. Additionally, an attenuation of the microvascular network was observed in the inflamed islets imaged in situ in the surgically exposed pancreases (Fig. 5a–c), but not in control B6 mice in situ (Fig. 5d) or B6 islet grafts in the ACE (Fig. 2, Fig. 4d).
Our label-free 3D images are based on next-generation OCM, denoted functional optical coherence imaging for its substantial extension into functional imaging. Compared with related imaging techniques such as OCT  or optical frequency domain imaging (OFDI) , FOCI has a micrometric resolution over an extended depth range. In contrast to conventional confocal microscopy, FOCI offers several advantages for islet imaging. First, it allows for simultaneous, label-free imaging of the islet structure and the vasculature. As the FOCI contrast is caused by the intrinsic light scattering of tissue it does not require genetically modified mice or an extrinsic biomarker to detect the beta cell volume and functional vascularisation. In addition, it is not affected by autofluorescence. Second, the coherent amplification of the backscattered light results in a higher signal-to-noise ratio. This increased contrast enables quantitative segmentation of the beta cell volume, vascularisation and infiltration. Third, the multiplex advantage of FOCI translates into a parallel acquisition of depth profiles, circumventing the need for voxel per voxel scanning in depth. The fast acquisition increases the time resolution allowing for visualising the vasculature of whole islets in 30 s. By applying the phase sensitivity of FOCI, blood flow can be quantified, a critical variable to monitor during the progression of diabetes. Finally, FOCI uses a broadband near infra-red (NIR) light source (<5 mW on the cornea) to acquire a full depth, resulting in a substantially lower power exposure and minimises the risk of photo-damage of fragile islet structures and microvasculature.
The lack of direct, non-invasive technologies to visualise inflammation in the pancreas has been a key obstacle to the early detection of type 1 diabetes and the rapid assessment of the effectiveness of therapeutic intervention. To highlight the usefulness of FOCI applied to the ACE model, we analysed and compared islets from a mouse model of type 1 diabetes. The location and anatomy of the pancreas make the study of this disease at the organ level difficult, requiring compromises on resolution for longitudinal studies in the timescale of disease progression. While direct examination of fixed pancreases has yielded valuable insight into the processes leading to diabetes, such classic attempts only provide a snapshot of the disease. Thus, non-invasive imaging strategies to monitor changes within the islets associated with diabetes are being sought to fill this gap. By applying the FOCI technique, we revealed several surprising features of the disease.
The longitudinal recording of the autoimmune-induced islet alterations in the ACE provided detailed monitoring and quantification of the decreasing beta cell volume and of the impact on the islet vasculature during the inflammatory process. An observed decrease in the vascular density in areas affected by insulitis coincided with a promotion of MAdCAM-1-expressing HEV-like structures. Altered vasculature has previously been associated with type 1 diabetes [2, 3, 4, 36]. More specifically, Villalta et al  have reported an increase rather than a decrease in vasculature when comparing pancreases from NOD mice prior to insulitis onset and NOD mice with advanced insulitis. This apparently conflicting statement may, in part, be due to differences in the overall assessment. We compared the vascularisation of infiltrated and non-infiltrated areas on the same individual islets, whereas Villalta et al  observed the increase in vasculature comparing the whole islets isolated from 4-week-old NOD mice with 18-week-old NOD mice. We therefore argue that the increase in vasculature under those conditions may be part of a developmentally controlled increase in islet vascularisation. HEVs have been linked to the promotion of guidance of lymphocyte trafficking and recruitment into inflammatory sites . The observed overall reduction in the microvasculature is counterintuitive and remains to be understood in the larger framework of its impact on the pathogenesis of autoimmune diabetes.
The label-free detection of beta cell volume and vascularisation offered by the FOCI technique provides the unique possibility of studying ACE-transplanted human islets. Interestingly, the acceptance rate of xenografting human islets to NOD.Rag2−/− mice was similar to that of grafting syngeneic NOD islets. This may be surprising, as recombination activating gene 2 (RAG2)-deficient mice have been shown to retain functional natural killer (NK) cells, which have been suggested to contribute to xenograft rejection . However, the role of NK cells in the rejection of xenotransplanted solid organs remains debated  and we cannot fully explain the mechanisms.
We thank M. Sison (School of Engineering, Ecole Polytechnique Fédérale de Lausanne, Switzerland) for his help with the dfOCM setup.
This work was partly financially supported by the Swiss National Science Foundation Grant (20320L-150191, 206021-139141), the Commission for Technology and Innovation (CTI, Bern, grant 17537.2 PFLS-LS), the Novo Nordisk Foundation, Diabetesförbundet, Barndiabetesförbundet and by the Swedish Research Council. D. Szlag acknowledges support from the Scientific Exchange Programme between Switzerland and the New Member States of the European Union and the project Enhancing Educational Potential of Nicolaus Copernicus University in the Disciplines of Mathematical and Natural Sciences grant.
Duality of interest
The authors declare that there is no duality of interest associated with this manuscript.
CB, AS-C, DS, JE, AG-B, TL and DH conceived and designed the experiments. AS-C, LH, CB, JE, AB and DS performed, acquired and analysed the FOCI data. AS-C performed and analysed the immunohistochemistry experiments. DS implemented the vascularisation extraction. CB performed the analysis and quantification of the FOCI images. CB, JE, JG, MV, AG-B and FS performed, acquired and analysed the Znt8 and guinea pig experiments. CB, AS-C, DS, TL and DH drafted the manuscript. JE, LH, AB, MV, JG, FS and AG-B critically revised the manuscript. All authors approved the final version to be published and agreed to be accountable for all aspects of the work. DH is guarantor of this work.
- 11.Coppieters K, Martinic MM, Kiosses WB, Amirian N, von Herrath M (2010) A novel technique for the in vivo imaging of autoimmune diabetes development in the pancreas by two-photon microscopy. Plos One 5Google Scholar
- 31.Mebius RE, Streeter PR, Michie S, Butcher EC, Weissman IL (1996) A developmental switch in lymphocyte homing receptor and endothelial vascular addressin expression regulates lymphocyte homing and permits CD4+ CD3- cells to colonize lymph nodes. Proc Natl Acad Sci U S A 93:11019–11024PubMedCentralCrossRefPubMedGoogle Scholar