In vivo imaging in experimental spinal cord injury - Techniques and trends.

Introduction: Traumatic Spinal Cord Injury (SCI) is one of the leading causes of disability in the world. Treatment is limited to supportive care and no curative therapy exists. Experimental research to understand the complex pathophysiology and potential mediators of spinal cord regeneration is essential to develop innovative translational therapies. A multitude of experimental imaging methods to monitor spinal cord regeneration in vivo have developed over the last years. However, little literature exists to deal with advanced imaging methods specifically available in SCI research. Research Question: This systematic literature review examines the current standards in experimental imaging in SCI allowing for in vivo imaging of spinal cord regeneration on a neuronal, vascular, and cellular basis. Material and
Methods: Articles were included meeting the following criteria: experimental research, original studies, rodent subjects, and intravital imaging. Reviewed in detail are microstructural and functional Magnetic Resonance Imaging, Micro-Computed Tomography, Laser Speckle Imaging, Very High Resolution Ultrasound, and in vivo microscopy techniques.
Results: Following the PRISMA guidelines for systematic reviews, 689 articles were identified for review, of which 492 were sorted out after screening and an additional 104 after detailed review. For qualitative synthesis 93 articles were included in this publication. Discussion and
Conclusion: With this study we give an up-to-date overview about modern experimental imaging techniques with the potential to advance the knowledge on spinal cord regeneration following SCI. A thorough knowledge of the strengths and limitations of the reviewed techniques will help to optimally exploit our current experimental armamentarium in the field.

Introduction

Traumatic spinal cord injury (SCI) is one of the world's leading causes of disability (Singh et al., 2014; Cripps et al., 2010; Majdan et al., 2017; Badhiwala et al., 2018). No curative therapy exists, and treatment is limited to supportive care. In the past decades, experimental research brought tremendous progress to the basic knowledge of the complex pathophysiology underlying SCI, as well as to the endogenous regenerative responses of the spinal cord. With advancing knowledge, research is shifting from the general characterization of SCI pathophysiology to the potential amelioration of spinal cord regeneration via targeting endogenous repair mechanisms (Monje, 2021). Experimental studies using longitudinal in vivo imaging play a significant role in the characterization of potential targets for advancing spinal cord regeneration.

Experimental imaging of CNS pathologies has evolved widely over the past decades. In comparison to classical histological tissue analysis, longitudinal in vivo imaging allows for an increasingly accurate morphological and functional tissue analysis in the same subject at different time points. With high dynamics and a high spatial resolution, longitudinal in vivo imaging can be performed non-invasively and therefore repeatedly. This does not only increase comparability, but also reduces the necessary group sizes in animal experimental research. In SCI, in vivo imaging evolved over the last years from the mere morphological assessment of trauma size and regeneration to the possibility of live tracking of spinal cord injury and regeneration on a neuronal, vascular, and even on a cellular level. These techniques allow furthermore for real-time tracking and longitudinal follow-up of the regenerative capacities of experimental therapies.

However, little literature exists to deal with the multitude of advanced experimental imaging methods specifically available in SCI research, and their capacities to enhance the understanding of spinal cord regeneration.

Material and methods

Approach to the systematic literature review

We performed a systematic review of the literature. The Medline databases Pubmed and Pubmed central, as well as CINAHL, Embase, Google Scholar and Science Direct were searched. The main search terms included Spinal Cord Injury, Intravital Imaging, Intravital Microscopy, Magnetic Resonance Imaging (MRI), Ultrasound, Photoacoustic Imaging, Laser Speckle Imaging, and Micro-Computed Tomography (μCT), in combination with Spinal Cord Injury and rodent subjects. The search included no time limit. Articles in English were included, meeting the following criteria: experimental research, original studies, rodent subjects, intravital imaging including MRI, μCT, Intravital Microscopy, Laser Speckle Imaging or Ultrasound techniques. The systematic literature search was conducted independently by four researchers from March 2020–November 2021. Data analysis was conducted according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Obtained for review were 689 articles, of which 492 were sorted out after screening due to not meeting the inclusion criteria. An additional 104 articles were excluded due to not meeting the inclusion criteria after full-text assessment for eligibility. For qualitative synthesis 93 articles were assessed in detail and included in this publication (Fig. 1: PRISMA flow chart).

Fig. 1

Flow Chart of the conducted literature review and the systematic synthesis according to the PRISMA guidelines for systematic reviews and meta-analyses. (PRISMA – Preferred Reporting Items for Systematic Reviews and Meta-Analyses).

Relevant experimental in vivo imaging methods according to the literature search

According to the systematic literature search and the inclusion of 93 articles for qualitative data synthesis, five main in vivo imaging techniques in use in experimental SCI research were identified: 1. Magnetic Resonance Imaging with functional (f)MRI and microstructural MRI, 2. Micro-CT (μCT), 3. Laser Speckle Contrast Imaging (LSCI), 4. Ultrasonography with Very High Resolution Ultrasound (VHRUS) and Photoacoustic Imaging (PA), and 5. Intravital Microscopy including Epifluorescence Videomicroscopy (IVM) and 2- or Multiphoton-Microscopy (TPEF). In the following, we discuss these topics in detail, including history, technical development, area of applications, strengths and weaknesses and potential outlook.

Illustrative examples of in vivo imaging

In Fig. 2, Fig. 3, Fig. 4, we display exemplary images of in vivo imaging methods performed at our institutions in murine specimens (C57BL/6J) with SCI or sham injury. The illustrative images were produced as by-products during the performance of experimental SCI studies conducted at our local institutions and were not previously published. All animal procedures were approved by the local governmental institutions (G0314/17). A continuous and close monitoring protocol was followed, and potent pain medication was applied as described in detail before (Soubeyrand et al., 2014a; Figley et al., 2014; Forgione et al., 2017). For SCI induction, the Clip Compression Contusion Injury model was used and for sham injury a two-level laminectomy without SCI was performed at the thoracic level (T6/7 or T10/11) (Joshi and Fehlings, 2002). For imaging procedures, volatile anesthesia (isoflurane; for magnetic resonance imaging, MRI) or intravenous anesthesia (Ketamine, Xylazine; for Epifluorescence videomicroscopy and very high resolution ultrasound, VHRUS) was used. MRI was performed using a 7 ​T small-animal MRI scanner (7 ​T PharmaScan 70/20USR, Bruker corp.). For VHRUS, the VEVO 770 (FUJIFILM VisualSonics) ultrasound system was used (Soubeyrand et al., 2014a; Badner et al., 2016). For longitudinal in vivo microscopy, an implanted spinal window chamber adapted from Farrar et al. was implanted at the thoracic level following SCI or sham injury (Th10/11) and epifluorescence videomicroscopy was performed in the spinal cord as described before in the brain (Farrar et al., 2012; Vajkoczy et al., 1999; Uhl et al., 2018).

Fig. 2

In vivo T2 TurboRARE 7 ​T Magnetic Resonance Imaging in mice with Sham-injury and SCI for axial and sagittal imaging of the spinal cord (spinal cord swelling after SCI, green arrows). Ex vivo MRI (spinal cord in saccharose) following sacrifice is also possible with exact volumetry of grey and white matter as well as injury volumetry (in blue/green, right). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3

A Technical Setting of in vivo Very High Resolution Ultrasound (VHRUS) in a murine model of experimental SCI. B Example of VHRUS of the intact (Sham) and injured (SCI) murine spinal cord in B-mode and Power Doppler mode for structural, functional and volumetrical analysis of SCI. The white arrows show structural alterations in a specimen with SCI. The images of VHRUS in B-Mode and Power Doppler Mode are printed with the permission of Michael G. Fehlings and Anna Badner.

Fig. 4

Experimental setup of in vivo microscopy of the spinal cord via implanted spinal window chamber, adapted by Farrar et al. (A ​+ ​B) In vivo video-epifluorescence microscopy via the implanted spinal window chamber at 7 days post implantation in a healthy specimen, with the arrow showing intact spinal cord vessels (C). Schematic setup of TPEF (Two-photon excitation fluorescence microscopy) in a murine model of experimental SCI using the spinal window chamber for longitudinal in vivo imaging (D).

Theory

With this study we aim to comprehensively review experimental setups using in vivo imaging of different kind in SCI, to give an up-to-date overview to experimental researchers dealing with spinal cord regeneration and repair following SCI. In the following sections, we will review in vivo imaging techniques used in modern experimental settings to analyze SCI and spinal cord regeneration in detail.

Results

Structural and functional magnetic resonance imaging (MRI/fMRI)

4.1.1 History: Magnetic resonance imaging (MRI) and its derived techniques are the oldest of the imaging techniques discussed, with the underlying methods originating between the 1980s and early 2000s (Falconer et al., 1994; Basser et al., 1994; Malisza and Stroman, 2002; Ford et al., 1994; Ramu et al., 2006). The clinical use of MRI predates its use in animal experimental research. MRI as a tool to image experimental SCI in rodents was first established by Hackney et al., in 1986 (Ford et al., 1994). Falconer et al. in 1994 were the first to perform a longitudinal in vivo imaging study to quantify SCI pathology using MRI by constructing 3D-volume rendered images (Falconer et al., 1994). Within the first 5–10 years following the establishment of MRI in experimental SCI, the additional use of diffusion tensor imaging (DTI) appeared, though largely applied ex vivo (Cohen et al., 2009; Patel et al., 2016). Since 2002, functional MRI (fMRI) is established in experimental SCI, with its first application in the cervical spine in 2006 (Malisza and Stroman, 2002). Since the beginning of its use in animal experiments, MRI quality improved from 1,4 ​T to about 7–9 ​T, while in certain cases even ultra-high field resolution MR imaging is available at 16,4 ​T (Ford et al., 1994; Schwartz and Hackney, 2003; Brennan et al., 2013).

4.1.2 Technical: MRI scanners used for experimental imaging are not notably different to those used in the clinical setting, though with smaller bore sizes (30–90 ​cm) and an appropriately smaller surface coil. For improving image quality during in vivo imaging, implanted coils appeared at the time, but were abandoned later for their invasiveness with the technical evolvement of modern MRI scanners (Ford et al., 1994; Sundberg et al., 2010). Modern experimental MRI scanners can monitor breathing, and use respiratory and cardiac gating to reduce imaging artifacts produced by movement (Ford et al., 1994; Sundberg et al., 2010). At the beginning mostly T1 weighted imaging with or without contrast-enhancing and T2 or Rapid Acquisition with Relaxation Enhancement (RARE) sequences were used (Malisza and Stroman, 2002; Song et al., 2018; Sundberg et al., 2011). During the technical evolvement, Diffusion Tensor Imaging (DTI) became predominant, using computer assistance to reconstruct axonal bundles on the diffusion of present water molecules in the tissue (Basser et al., 1994; Soares et al., 2013). Functional MRI (fMRI) uses Blood Oxygen Level Dependent (BOLD) contrast changes to detect active regions containing ascending and descending sensorimotor pathways in resting state or upon stimulation (Malisza and Stroman, 2002; Ramu et al., 2006; Ghosh et al., 2009a; Ogawa et al., 1990). Specialized devices for small animals have been improved over the years to generate excellent resolution images with high contrast between grey and white matter. A higher magnetic field strength than in the clinical setting allows a stronger signal and reduces voxel size (Bilgen, 2012), with field strengths up to 16.4T available (Brennan et al., 2013). Additionally, size-adjusted volume and surface coils achieve noise reduction and decrease the field of view for a higher signal to noise ratio (Bilgen, 2012). However, with varying field strength and for each imaging modality mentioned above the common acquisition parameters must be adjusted, which include but are not limited to repetition time, echo time, field of view and slice thickness (Bilgen, 2012).

4.1.3 Areas of application: MRI in SCI research is mostly used for the morphologic assessment of white matter pathology, including volumetric calculations of the lesion, and labeling of hyper- and hypointense areas as edema, hemorrhage or necrosis (Brennan et al., 2013; Sundberg et al., 2010; Song et al., 2018). The progression of the lesion volume over time can be assessed, displaying pathophysiological changes at both the epicenter and at areas rostral and caudal of the lesion, allowing for a prediction correlation to neurologic function (Wilkins et al., 2020). Following SCI, the integrity and regeneration of the Blood-Spinal-Cord-Barrier (BSCB) can also be assessed with MRI (Sundberg et al., 2011; Bilgen et al., 2001; Patel et al., 2009). Using Diffusion Tensor Imaging (DTI) technique, axonal integrity can be assessed (Tu et al., 2013). Another interesting area of application is the assessment of cortical plasticity following SCI using functional MRI (fMRI) (Ghosh et al., 2009a; Matsubayashi et al., 2018). This technique holds the ability to display the redistribution of somatosensory neurons in the cortex featuring the rewiring of hindlimb motoneurons (Ghosh et al., 2010). As a correlating technique Transcranial Magnetic Stimulation (TMS) has been shown to influence cortical plasticity, with both electromyography (EMG) and high-resolution fMRI detecting cortical responses to hindlimb stimulation in SCI animals, delivering promising results on enhancing neuronal plasticity and functional recovery (Krishnan et al., 2019). Additionally, MRI allows for the dynamic and non-invasive monitoring of stem cell migration after transplantation (Filippi et al., 2016; Gonzalez-Lara et al., 2011; Hu et al., 2012a; Shen et al., 2009). A selection of contrast agents to render stem cells MRI-detectable have been tested in rodent SCI models, most frequently superparamagnetic iron-oxide particles (SPIO) (Hu et al., 2012a; Shen et al., 2009; Gonzalez-Lara et al., 2009) and more recently Gadolinium-based (Gd) (Filippi et al., 2016; Yahyapour et al., 2018). While both show promising images of initial cell movements, signal loss is observed after about ten days with Gd agents, likely due to dilution from continuous cell division, as all contrast agents are administered in vitro (Filippi et al., 2016). SPIO signals were detectable up to three weeks post injection by Hu et al. (2012), but more susceptible to artifacts (Hu et al., 2012a). Guzman et al. (2007) have managed to trace SPIO-labeled stem cells for up to 18 weeks, however in the brain (Guzman et al., 2007).

4.1.4 Strengths and weaknesses: A good correlation of white matter integrity and regeneration with functional recovery could be demonstrated in studies using MRI, neurobehavioral analysis and histopathology (Wilkins et al., 2020). The rodent's physiological movements remain a problem, causing longer examination times as compared to humans (Schilling et al., 2019). Rodent studies therefore often use a small sample size, due to high cost and effort (Ramu et al., 2006; Filippi et al., 2016; Tu et al., 2010). To increase availability and reduce material costs, the use of clinical 1.5 –3 T MRI scanners in rodents has been investigated, most recently by Derksen et al. (2021) The authors performed fMRI on rats with a custom-designed coil in a clinical scanner. However, resulting voxel size and repetition time were inferior to specialized animal scanners (Derksen et al., 2021). Small animal scanners with a higher range of field strengths allow the imaging of spinal cord tissue in growing detail. Thus, contrast-labeled stem cells in the literature were traceable with 4.7 and 7 T scanners (Filippi et al., 2016; Gonzalez-Lara et al., 2011), whereas fMRI of neuronal pathways could be carried out at 9.4 T (Ghosh et al., 2010), and the exact location of responses to stimulation using targeted TMS therapy was evaluable by 11.1 T MRI (Krishnan et al., 2019). Clear differentiation of white and grey matter in the rodent spinal cord, including pathological changes, is indeed possible in clinical 2 T scanners - as was demonstrated as early as 1994 by Falconer et al. in the first longitudinal in vivo MRI study of murine SCI (Falconer et al., 1994). However, DTI allows quantitative assessment of white matter integrity, and Brennan et al. suggest that 16.4 T DTI detects changes in radial diffusivity better than a similar study carried out at 4.7 T, proposing a higher field strength beneficial (Brennan et al., 2013). MR images are stationary and cannot monitor dynamic changes in real-time, like in imaging methods discussed below (Brieu et al., 2010; Richards et al., 2017; Najafzadeh et al., 2020; Mallidi et al., 2011; Luke et al., 2012; Vawda et al., 2019; Badner et al., 2019). However, MRI still remains one of the most familiar and well-established imaging modality and is widely available (Farrar et al., 2012).

4.1.5 Outlook: Longitudinal in vivo rodent MRI provides the potential to explore microstructural damage and regeneration of the spinal cord following SCI (Duval et al., 2015). With DTI, axonal integrity and remyelination can potentially be assessed (Tu et al., 2013), while cortical plasticity following SCI can be examined using fMRI (Ghosh et al., 2009a; Matsubayashi et al., 2018) (Study overview: Table 1 (Patel et al., 2016; Brennan et al., 2013; Ghosh et al., 2010; Filippi et al., 2016; Endo et al., 2007; Ghosh et al., 2009b), Fig. 2: Examples of in and ex vivo MRI and volumetry in SCI.).

Table 1

Exemplary articles to the experimental usage of MRI and fMRI, Ultrasound, and in vivo Microscopy after Spinal Cord Injury.

MRI and fMRI
Author (Date)	Animal	Injury model	Injury location	Imaging method	Primarily analyzed issue
Brennan et al. (2013)	C57BL/6 mice	Contusion	Thoracic, Th9	MRI, DTI	White matter pathology and regeneration
Endo et al. (2007)	Sprague-Dawley rats	Transection	Thoracic, Th9	BOLD fMRI	Cortical rewiring
Filippi et al. (2016)	Balb/c mice	Lateral hemisection	Lumbar, L2	MRI, DTI	Tracking of Gadoteridol-labeled Mesenchymal stem cells
Ghosh et al., 2009a, Ghosh et al., 2009b	Lewis rats	Lateral hemisection	Cervical, C4	BOLD fMRI	Cortical rewiring
Ghosh et al. (2010)	Lewis rats	Transection	Thoracic, Th8	BOLD fMRI	Cortical rewiring
Patel et al. (2016)	Sprague-Dawley rats	Contusion	Thoracic, Th7	MRI	BSCB permeability
(Very High Resolution) Ultrasound
Khaing et al., 2018	Sprague-Dawley rats	Compression	Thoracic, Th6-Th10	Ultrafast CEU	Hemodynamic changes
Soubeyrand et al., 2014a, Soubeyrand et al., 2014b	Wistar rats	Clip Compression	Thoracic, Th10-Th12	VHRUS	Vascular injury and regeneration
Figley et al., 2013a, Figley et al., 2013b	Athymic nude/C57BL/6 mice	Irradiation	Lumbar, L2-L3	VHRUS and PA	Vascular injury and regeneration
In vivo microscopy
Kerschensteiner et al. (2005)	C57BL/6 mice	Transection	Cervical, C3–C6	WFFM	Axonal degeneration and regeneration
Farrar et al. (2012)	C57BL/6 mice (YFP-H)	Laser ablation	Thoracic, Th12	TPEF	Axonal degeneration and scar formation
Fenrich et al. (2013)	C57BL/6 mice	Unilateral Pinprick	Thoracic – Lumbar, Th12 – L2	TPEF	Infiltrating and resident myelomonocytic cells
Horiuchi et al. (2015)	B6.Cg-Tg (Thy1-YFP) H2Jrs/J mice	Contusion	Thoracic, Th11	TPEF	Axons of the dorsal funiculi
Chen et al., 2017a, Chen et al., 2017b	Sprague Dawley rats	Contusion	Cervical, C7	TPEF	Vascular changes

Abbreviations: BOLD = Blood Oxygen Saturation Level, BSCB = Blood-Spinal-Cord-Barrier, C = Cervical vertebra, CEU = Contrast-Enhanced Ultrasound, DTI ​= ​Diffusion Tensor Imaging, fMRI ​= ​functional Magnetic Resonance Imaging, L ​= ​Lumbar vertebra, MRI ​= ​Magnetic Resonance Imaging, PA = Photoacoustic Imaging, Th = Thoracic vertebra, TPEF ​= ​Two Photon Excitation Fluorescence Microscopy, VHRUS = Very High Resolution Ultrasound, WFFM = Widefield Fluorescence Microscopy, YFP = Yellow Fluorescent Protein.

4.1.6 Potential for clinical translation: As clinical scanners predate specialized small animal scanners, availability does not pose an obstacle. Monitoring the development of the BSCB in patients may provide insight into an opportune moment for therapeutic intervention, while fMRI will help in understanding changes to ascending and descending pathways that can be targeted (Bakhsheshian et al., 2021).

Micro-Computed Tomography

4.2.1 History: In 1994, Micro-Computed Tomography (μCT) appeared in experimental use. Initially mostly used for high-resolution bone imaging ex vivo, the method found its application in experimental SCI rather recently (Xu et al., 2017; Hu et al., 2012b, 2015; Tschuchnig et al., 2021; Cao et al., 2017; Zambrano-Rodríguez et al., 2019). Although in SCI research μCT is mostly used ex vivo too (Xu et al., 2017; Hu et al., 2012b, 2015; Cao et al., 2017; Strotton et al., 2021), recent studies started with in vivo protocols to assess Spinal Subarachnoid Space (SSAS), or to monitor microangioarchitecture and posttraumatic tissue integrity, as well as stenosis of the spinal canal (Zambrano-Rodríguez et al., 2019, 2021a; Lee et al., 2012; Huang et al., 2020).

4.2.2 Technical: With μCT, a high-speed full-body scan can be performed in a small animal model with ultra-low X-ray radiation doses, with less than 2 ​mGy whole-body in a mouse. The technical background is the illumination of a sample with a micro-focus X-ray source. Through rotation and serial imaging acquiring 2D images, later 3D reconstruction and thus the complete imaging of a sample at high resolution can be achieved. This way in SCI, 3D-imaging of the micro-neurovascular anatomy or imaging of the SSAS can be achieved at a very high resolution, in vivo reported by Zambrano-Rodríguez et al. (2019) as 100 ​μm in rats (Hu et al., 2012b, 2015; Zambrano-Rodríguez et al., 2019, 2021a, 2021b; Cheng et al., 2015).

4.2.3 Areas of application: In SCI research, μCT is most frequently used ex vivo to 3D image explanted specimen (Hu et al., 2012b, 2015; Cao et al., 2017). One large field of use is the high-resolution imaging of bone quality in SCI induced sarcopenia (Otzel et al., 2019; Wu et al., 2021; Yarrow et al., 2014; Debaud et al., 2017). In vivo, Zambrano-Rodríguez et al. were 2019 able to combine myelography with contrast-enhanced intravital μCT to image the SSAS in the rat and were later able to provide high-resolution in vivo images of changes in the SSAS following SCI (Zambrano-Rodríguez et al., 2019, 2021b). Moreover, μCT is used for the assessment of spinal canal width in a rodent model of spinal spondylotic myelopathy (Lee et al., 2012) and for the imaging of spinal cord integrity in therapy studies following SCI (Huang et al., 2020).

4.2.4 Strengths and weaknesses: In the ex vivo application, one strength of μCT is its non-destructiveness and therefore the preservation of samples for later destructive analyses like histology (Hu et al., 2012b, 2015; Cao et al., 2017). In vivo, the potential for high-resolution SSAS imaging has been shown. Also, tissue integrity following injury and the regeneration potential of SCI therapies was examined (Zambrano-Rodríguez et al., 2019, 2021b). In comparison to MRI, μCT image acquisition can be fast, but high-resolution 3D image reconstruction takes up to 5 ​h (Zambrano-Rodríguez et al., 2019, 2021a) and a complex IT infrastructure for image processing is necessary. Although μCT uses low radiation doses, repeated in vivo full body imaging of small animals can lead to relevant radiation exposure (Zambrano-Rodríguez et al., 2021b).

4.2.5 Outlook and potential for clinical translation: In addition to the beforementioned, interesting is the ability of modern μCT scanners to be combined with PET imaging and thus the multifold of possible experimental applications in vivo. In the clinic, μCT is reserved for high-resolution ex vivo sample analysis, for example in breast cancer and pulmonary diseases (Mai et al., 2017; Dicorpo, Tiwari, Tang, Griffin, Aftreth, Pinky BautistaHughes, Gershenfeld, Michaelson).

Laser Speckle Contrast Imaging

4.3.1 History: The fundamental principle of Laser Speckle Contrast Imaging (LSCI) evolved in the 1960s and 1970s (1976; Briers et al., 2013). Later on, it has been used for real-time imaging of blood flow mostly on the surface of the cortex in neuroscience and clinical research (Briers et al., 2013). Its usage for experimental application in SCI up to date is limited.

4.3.2 Technical: In LSCI, a laser light (coherent light, infrared or near infrared) throws light on an object with an irregular surface which is then reflected. This creates an interference pattern, called a speckle (Briers et al., 2013; Briers, 2001; Davis et al., 2014). The speckle approach was originally a single point measurement method; later scanning techniques were developed to provide a map of velocities (Briers et al., 2013; Briers, 2001). The theory of laser speckle contrast analysis is based on the principle that motion causes alterations in the speckle pattern, as for the movement of blood cells. Fast flow leads to a more indistinct pattern while the contrast is diminished whereas a decrease in flow is associated with higher contrast (Briers, 2001; Hecht et al., 2009; Lesage et al., 2009). Cameras detect the speckle pattern, and contrast can be calculated. Based on that, contrast calculating software is able to create a color-coded map for example of blood flow velocities (Hecht et al., 2009; Senarathna et al., 2013). However, imaging depth remains low, creating a momentary map of surface vascular anatomy (Senarathna et al., 2013; Gallagher et al., 2019). Depending on the wavelength the tissue can be penetrated by less than 1 ​mm and a spatial resolution of about 10 ​μm can be achieved (Senarathna et al., 2013).

4.3.3 Areas of Application: LSCI is mostly used to map superficial vessels in the cortex and less to image deep vessels in the spinal cord (Senarathna et al., 2013). Thus, LSCI is rarely used in experimental spinal cord imaging compared to its wide usage in measuring the blood flow of the cortex (Lesage et al., 2009; Dunn et al., 2005; Luo et al., 2007; Woitzik et al., 2013). Reason for this is mostly the limited depth of this technique and the challenge of respiration artifacts (Lesage et al., 2009). However, some SCI researchers used this tool to measure the restoration of spinal cord blood flow following electrical stimulation, or for imaging the superficial vascular reaction to neuronal stimulation in injured and non-injured rodents (Lesage et al., 2009; Beaumont et al., 2014; Brieu et al., 2010).

4.3.4 Strengths and Weaknesses: LSCI is a non-invasive, real-time technique which enables full-field imaging in 2-dimension. The equipment is well accessible (common laser, optics, and camera) at relatively low cost. An additional contrast agent is not required (Briers, 2001; Richards et al., 2017; Murari et al., 2007). Although the spatial and temporal resolution of this imaging technique is high, imaging depth remains low, making the technique not suitable for imaging deeper vascular networks in the spinal cord (Lesage et al., 2009; Senarathna et al., 2013).

4.3.5 Outlook and potential for clinical translation: Although widely used in the clinic and in the laboratory for cortical surface blood flow analysis, the use of LSCI for clinical and experimental application in SCI remains limited.

Very high resolution ultrasound and Photoacoustic Imaging

4.4.1 History: The use of Very High Resolution Ultrasound (VHRUS) as an in vivo imaging tool in SCI has emerged over the last decade. With a trend towards the use of higher frequency ultrasound, a higher spatial resolution can be achieved. Moreover, a combination with other techniques like contrast-enhanced Ultra-sonography and Photoacoustic Imaging (PA) is available (Xu and Wang, 2006). This makes VHRUS an inexpensive and user-friendly tool for in vivo imaging of the whole spinal cord following SCI (Soubeyrand et al., 2012, 2014b; Figley et al., 2013a).

4.4.2 Technical: VHRUS is possible with specialized ultrasound machines (Soubeyrand et al., 2014a, 2014b). Using a high frequency (40–50 ​MHz), high resolution with 20–30 microns per pixel can be achieved (Soubeyrand et al., 2012; Finn-Bodner et al., 1995; Huang et al., 2013; Jones et al., 2012; Dubory et al., 2015). Contrast-enhanced ultrasonography is possible via the intravenous administration of microbubbles serving as a contrast-agent (Soubeyrand et al., 2012; Dubory et al., 2015). Compared to other functional vessel imaging like MR-angiography or CT-angiography, emerging ultrasound techniques allow for a higher resolution and co-registering with Photoacoustic Imaging (PA) is possible (Khaing et al., 2018). PA is based on the physical principle of the photoacoustic effect, initially reported in 1880. With this technique, non-invasive image-based oxygen saturation measuring in spinal cord vessels and spinal cord tissue is possible to assess hypoxia and ischemia (Xu and Wang, 2006; Figley et al., 2013b; Bell, 1880). In PA, absorption by the target molecule leads to a local temperature rise and therefore to thermal expansion of the tissue producing acoustic waves recognized by an ultrasound transducer (Najafzadeh et al., 2020; Mallidi et al., 2011). Distinct absorption properties apply to various optical chromophores (i.e. oxygenated or deoxygenated hemoglobin, melanin, water, and lipid) (Jeon et al., 2016). An additional combination with contrast agents like contrast dyes or nanoparticles is possible to increase the photoacoustic signal (Xu and Wang, 2006; Luke et al., 2012). With appropriate contrast agents, imaging on the cellular and even molecular level is realizable (Najafzadeh et al., 2020; Mallidi et al., 2011; Jeon et al., 2016). Also, the technique allows for 3D tissue imaging (Soubeyrand et al., 2014a; Moonen et al., 2016; Vawda et al., 2019).

4.4.3 Areas of Application: In experimental SCI, VHRUS can be used to longitudinally quantify the posttraumatic morphological alterations as well as changes in hemodynamics in the spinal cord (Moonen et al., 2016; Badner et al., 2019). Parameters thus assessable include lesion volume, cavity volume, parenchymal hemorrhage and alterations in spinal cord blood flow (Soubeyrand et al., 2014b; Dubory et al., 2015; Khaing et al., 2018). Additionally, the migration of infused mesenchymal stem cells can be examined (Vawda et al., 2019; Soubeyrand et al., 2014c). In combination with PA, oxygen saturation of the spinal cord tissue and vasculature can be non-invasively assessed (Figley et al., 2013b; Soubeyrand et al., 2014c). Thus, using VHRUS the effects of potential neuroprotective and regenerative treatments such as norepinephrine and mesenchymal stomal cells could previously be assessed (Vawda et al., 2019; Soubeyrand et al., 2014c) and the role of IL-10 in the vascular pathology of SCI could be examined (Badner et al., 2019).

4.4.4 Strengths and Weaknesses: VHRUS allows for low-invasive, real-time longitudinal in vivo 3D-imaging of the whole spinal cord with high resolution and high depth (Vawda et al., 2019; Soubeyrand et al., 2014c). Some investigators perform repetitive surgery for VHRUS application on top of the dura. However, more modern technique allows the non-invasive application of VHRUS through the skin. Using ultrasound, tissue morphology is preserved and unaltered post mortem tissue processing remains possible (Soubeyrand et al., 2014b; Moonen et al., 2016). Apart from the VHRUS machine, the technique is inexpensive, with manageable recourses necessary. Image acquisition is fast and can easily be adjusted (Soubeyrand et al., 2014b; Dubory et al., 2015). PA and contrast-enhanced imaging can be combined, allowing for accurate and rapid monitoring of effects of prospective treatments targeting spinal cord regeneration (Soubeyrand et al., 2014a, 2014c; Dubory et al., 2015; Vawda et al., 2019).

4.4.5 Experimental outlook: VHRUS is a fast and easily adjustable technique for longitudinal 3D in vivo imaging of the whole spinal cord (Soubeyrand et al., 2014a). The combination with contrast-enhanced ultrasonography allows for the tracking of blood flow and for cell tracking, whereas the combination with PA allows for non-invasive oxygen saturation measurement in the spinal cord and spinal cord vasculature (Soubeyrand et al., 2014a; Figley et al., 2013b). The trend with ultrasound techniques goes towards higher frequencies leading to even higher resolution (Dubory et al., 2015) (Study overview: Table 2 (Soubeyrand et al., 2014b; Khaing et al., 2018; Figley et al., 2013b), Schematic setup of VHRUS and exemplary images in B-Mode and Power Doppler Mode: Fig. 3.).

4.4.6Potential for clinical translation: To monitor spinal cord blood flow following injury in humans with minimally-invasive ultrasound techniques is of interest (Chen et al., 2017a; Khaing et al., 2020), as the hemodynamic stability and an intact perfusion of the spinal cord is associated with an improved outcome regarding functionality (Werndle et al., 2014; Tator, 1991). Another potential translational approach to integrate ultrasound in combination with specific contrast agents could allow molecular imaging in clinical SCI (Khaing et al., 2020). In a preclinical model targeted contrast-enhanced ultrasound (TCEUS) was already described as a promising tool for the future (Volz et al., 2016, 2017).

Spinal cord microscopy

4.5.1 History: The first experimental tool used for in vivo microscopy was wide field fluorescence microscopy (WFFM), also called in vivo Epifluorescence Videomicroscopy (IVM), a basic microscope technique to image fluorescence (Kerschensteiner et al., 2005; Misgeld et al., 2007; Vajkoczy et al., 2001). However, due to limited penetration depth and limited axial resolution, the technique developed further. Two-photon excitation fluorescence microscopy (TPEF) was established in 1990, and first used in animal experiments to image glial cells in the brain (Helmchen and Denk, 2005; Denk et al., 1990). Later, TPEF was used for the tracking of immune cells in various organs. After many ex vivo studies it was first used in vivo 2003 for the imaging of lymphocytes (Cahalan et al., 2003). Originally only used for imaging cell movement, TPEF was later on also used with fluorescent proteins to image cell functions in vivo (Kawakami, 2018). Up to date, TPEF has become one of the superior tools for in vivo imaging.

4.5.2 Technical: In vivo imaging of the spinal cord is possible following surgical removal of the laminae (laminectomy), and for longitudinal in vivo imaging can be combined with the implantation of a spinal window chamber (Farrar et al., 2012; Figley et al., 2013b; Fenrich et al., 2012). In Epifluorescence Videomicroscopy (IVM), the specimen is exposed to light while illumination and detection of light cover the whole visual field simultaneously. The light source can be an LED, Mercury or Xenon arc-lamp, combined with an optical filter to choose wavelength. The image is captured by camera. Out of focus light and diffraction-limited optics result in comparably low contrast. Therefore, layers of single cells or organelles are best suited for this imaging modality (Sanderson et al., 2014). Two-photon excitation fluorescence microscopy (TPEF) is a non-linear optical method, based on photon excitation. Two photons of the same wavelength excite a fluorophore simultaneously. Together, they create enough energy for the emission of a single fluorescence photon with a higher level of energy and with shorter wavelength. This represents one pixel in the image. A femtosecond pulsated laser as light source with an optimum range at 800 ​nm (650–1000 ​nm) generates ultrashort impulses (<100 ​fsec) at 100 ​MHz and is pumped with another laser (around 500 ​nm). This way, a tissue penetration depth of around 600 ​nm is created (Helmchen and Denk, 2005; Johannssen and Helmchen, 2013). With even higher impulse energy systems, up to 1000 ​μm depth was previously achieved in the murine cortex (Theer et al., 2003).

4.5.3 Areas of application: Transgenic mice have been frequently used with IVM, for example YFP-labeled (yellow-fluorescent protein, H-line of thy1-YFP) (Misgeld et al., 2007; Skuba et al., 2011). As blood vessels originally appear dark, fluorescent dyes like rhodamine or FITC were also frequently administered intravenously (Dray et al., 2009; Chen et al., 2017b). This makes the tool relevant for the examination of structures rarely covered with tissue, like dorsal root ganglia (DRG) after laminectomy, the imaging of vessels and the continual tracing of axons and their de- and regeneration (Kerschensteiner et al., 2005). TPEF is increasingly used for spinal cord imaging in recent years. Mostly the dorsal funiculi and DRG have been imaged, with a penetration depth of 150–300 ​μm (Laskowski and Bradke, 2013). So far, imaging studies using TPEF in the spinal cord have examined cellular migration with microglia and macrophages (Evans et al., 2014) or bone-marrow derived stromal cells (Cord et al., 2005), dynamic vascular changes (Dray et al., 2009; Chen et al., 2017b; Davalos and Akassoglou, 2012; Tang et al., 2015), vascular damage (Zhang et al., 2014; Yang et al., 2017a) and neuronal and axonal degeneration and regeneration (Dray et al., 2009; Horiuchi et al., 2015; Zhang et al., 2015). Different fluorescent dyes and fluorescent transgenic mouse lines can be combined with this imaging method. The use of fluorescent dyes coupled to molecules of different sizes (e.g. dextran) further allows for the assessment of extravasation kinetics and a detailed analysis of BSCB (blood-spinal-cord-barrier) function and integrity. Frequently used transgenic mouse lines possess promotors coupled to fluorescent proteins YFP to image motor and sensory neurons and/or GFP (green fluorescent protein) to image mononuclear cells (Horiuchi et al., 2015). Moreover, SCa1-IREF can be used for imaging Ca2+-dynamics in live neurons, while Adeno-Associated Virus (AAV) expressing a fluorophore (e.g. GFP) can function as a tracer for axon-regeneration (Liao et al., 2017; Uckermann et al., 2015; Galli et al., 2018).

To assess axonal degeneration and regeneration following SCI, in vivo imaging of YFP-labeled axons allows for the visualization and quantification of acute axonal degeneration and Wallerian degeneration. The combination of TPEF microscopy with Third Harmonic Generation microscopy reveals the spatially and temporally overlapped degeneration of axons and myelin after SCI (Farrar et al., 2011a). Recently, the formation and dynamic change of spheroids has been examined, thereby further characterizing the acute axonal degeneration post injury (Orem et al., 2020a, 2020b). To further understand the pathophysiological processes underlying axonal degeneration, intracellular Ca2+ levels were quantified using mice containing the GFP-positive calcium indicator GCaMP6f in axons (Orem et al., 2020b; Tang, Zhang, Chen, Xinran, Ju, Liu, Gan, He, Zhang, Li, Zhang, Kirby). Some of the parameters mentioned above were also used to show the therapeutic effects of e.g. Methylprednisolone and Progesterone on axonal survival in vivo (Zhang et al., 2015; Orem et al., 2020a; Tang, Zhang, Chen, Xinran, Ju, Liu, Gan, He, Zhang, Li, Zhang, Kirby; Yang et al., 2017b). To examine vascular changes post SCI, most studies imaging vasculature via TPEF microscopy in vivo use the visualization of vessels as orientation and show the interaction with extraluminal cells like microglia or axons (Davalos and Akassoglou, 2012; Oshima et al., 2014; Fenrich et al., 2013). Also, in vivo imaging of cell migration is possible with labeled macrophages (Farrar et al., 2012; Evans et al., 2014; Fenrich et al., 2013; Davalos, Lee, Smith, Brinkman, Ellisman, Zheng, Akassoglou).

4.5.4 Strengths and Weaknesses: Although IVM is limited in penetration depth, its easy technique allows for fast noninvasive in vivo imaging of the Spinal Cord (Shi et al., 2019). SCI as in lesion models (transection, crush, or compression contusion injury) can be thus directly confirmed. Fast camera speed and recording is possible almost without respiratory interruptions and with low phototoxicity (Misgeld et al., 2007). Due to the limited penetration depth of this technique, imaging of the spinal cord vasculature and spinal cord neurons is possible in superficial areas of the dorsal spinal cord. In vivo microscopy enables longitudinal imaging over longer time periods. In combination with the sophisticated implantation of a spinal window chamber, repetitive surgery to expose the spinal cord has become obsolete (Farrar et al., 2012; Figley et al., 2013b; Fenrich et al., 2012). In TPEF, enhanced imaging depth is enabled by longer wavelength, which results in less scattering artifacts and less absorption within the tissue, compared to other in vivo microscopy techniques. The technique also further reduces the probability of phototoxicity, with no out-of-focus tissue bleaching. It allows for the differentiation between spared axons and regenerated sprouts as well as between spared and regenerated vessels and enables a direct evaluation of therapeutic effects. Alterations however are caused by breathing and pulse, though many attempts exist to deal with these artifacts (Davalos and Akassoglou, 2012).

4.5.5 Experimental outlook: TPEF has become one of the superior tools for experimental in vivo imaging, thanks to its high resolution and the possibility for longitudinal imaging without repetitive surgery with the usage of sophisticated spinal window chambers. A magnitude of different fluorescent dyes and transgenic mouse lines enables real-time in vivo imaging of complex pathophysiological processes like axonal de- and regeneration, neuronal Ca2+-signaling, and vessel restoration and function and much more following SCI. TPEF can be integrated with further nonlinear contrast imaging modalities for more specific purposes like Second Harmonic generation microscopy (SHG), used for the imaging of neuronal repair and fibrotic scar formation (Horiuchi et al., 2015; Liao et al., 2017). Third harmonic generation microscopy (THG) and Coherent Anti stokes Raman Scattering (CARS) have been used to image lipids and myelin structures to examine de- and remyelination (Farrar et al., 2011b; Hu et al., 2014; Galli et al., 2012) (Study overview: Table 3 (Farrar et al., 2012; Kerschensteiner et al., 2005; Horiuchi et al., 2015; Fenrich et al., 2013; Wang et al., 2017), Fig. 4: In vivo IVM via implanted spinal window chamber and schematic setup of TPEF).

4.5.6 Potential for clinical translation: Even if TPEF microscopy is most often used in an experimental setting, it already contributes to clinical diagnostics. In the clinical setting it can be used to image human tissue ex vivo for example in the field of oncology (Li et al., 2018). The clinical intravital use of TPEF microscopy is only well established in dermatology. Here it is a useful tool to image skin disease like melanoma (Dimitrow et al., 2009). Intravital TPEF microscopy has also been used in single cases in other surgical fields like in brain tumor surgery (Kantelhardt et al., 2016). Wide-field fluorescence microscopy is not fully translated to the clinic but is tested as a diagnostic tool in the point of care of e.g. Barret's neoplasia (Joshi et al., 2016) or colon cancer (Sensarn et al., 2016). For this application classical point of care white light endoscopes are modified to detect the specific fluorescence signals.

Discussion and limitations

Knowing about the different in vivo imaging methods available in an experimental setting today to monitor spinal cord injury and spinal cord regeneration helps with an overview about the potentials these methods hold to improve pathophysiological knowledge and thus to enhance regenerative therapies. However, for detailed experimental planning it is important to look at this multifold of imaging methods and techniques available regarding the intended study goal to monitor specific pathophysiological changes and regeneration. Thus, not all the mentioned methods hold the same potentials and limitations, as mentioned in the sections above. Moreover, between different techniques of the same imaging method, e.g. different MRI scanners or different MRI sequences, or different VHRUS machines of different generations there exist relevant differences. In the planning of an experiment, these details need to be assessed and dealt with. The goal of this review is to give an overview about the multiple options available, but one possible limitation is that a more detailed review of the different subsets of every discussed imaging technique is not achieved.

Moreover, there exist multiple established experimental models for experimental SCI induction in rodents, and even between rats and mice, there exist relevant differences in SCI pathophysiology. The reviewed studies include different animal models, like weight drop injury, impactor model, clip compression contusion injury, forceps injury, hemisection, or irradiation, in mice and rats. The one specific imaging method for the one injury model or animal model is not existent and methodological heterogeneity in the analyzed studies is high. In Table 1, exemplary studies included in the data synthesis are displayed, with the information which specimen and which SCI model was used. In experimental planning, the planned use of specimen, SCI model and combined imaging method is important regarding the intended study goal. The lack of differentiation thus is another potential limitation of this review.

Conclusions

Experimental studies using longitudinal in vivo imaging play a significant role in the characterization of potential targets for advancing spinal cord regeneration. Techniques at hand for in vivo imaging in SCI have dynamically evolved over the past decades. With the use of advanced imaging techniques like high-resolution MRI and fMRI combined with TMS, μCT, Very High Resolution Ultrasound combined with Contrast-Enhanced or Photoacoustic Imaging, and Two-Photon Excitation Fluorescence Microscopy in combination with various fluorescent dyes and transgenic mouse lines, longitudinal real-time in vivo imaging of neuronal, vascular and cellular regeneration of the spinal cord is possible non-invasively. With this study we comprehensively review experimental setups using in vivo imaging of different kind in SCI, to give an up-to-date overview to experimental researchers thriving to advance the knowledge on spinal cord regeneration and repair following SCI. A thorough knowledge of the strengths and limitations of these techniques will help to optimally exploit our current experimental armamentarium in the field.

Author's disclosure statement

No conflicts of interest exist in the submission of the manuscript and the manuscript is approved by all authors for publication.

Funding

No special funding was received for this study. VH was supported by the 2019 Integra EANS research fund and is currently funded by the BIH Charité Clinician Scientist program.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.