Macroencapsulation is a powerful approach to increase the efficiency of extrahepatic pancreatic islet transplant. FTY720, a small molecule that activates signaling through sphingosine-1-phosphate receptors, is immunomodulatory and pro-angiogenic upon sustained delivery from biomaterials. While FTY720 (fingolimod, Gilenya) has been explored for organ transplantation, in the present work the effect of locally released FTY720 from novel nanofiber-based macroencapsulation membranes is explored for islet transplantation. We screened islet viability during culture with FTY720 and various biodegradable polymers. Islet viability is significantly reduced by the addition of high doses (≥500 ng/mL) of soluble FTY720. Among the polymers screened, islets have the highest viability when cultured with poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV). Therefore, PHBV was blended with polycaprolactone (PCL) for mechanical stability and electrospun into nanofibers. Islets had no detectable function ex vivo following 5 days or 12 h of subcutaneous implantation within our engineered device. Subsequently, we explored a preconditioning scheme in which islets are transplanted 2 weeks after FTY720-loaded nanofibers are implanted. This allows FTY720 to orchestrate a local regenerative milieu while preventing premature transplantation into avascular sites that contain high concentrations of FTY720. These results provide a foundation and motivation for further investigation into the use of FTY720 in preconditioning sites for efficacious islet transplantation. © 2017 Wiley Periodicals, Inc. J Biomed Mater Res Part B: Appl Biomater, 106B: 555-568, 2018.

BackgroundIvabradine selectively inhibits the pacemaker current of the sinoatrial node, slowing heart rate. Few studies have examined the effects of ivabradine on the mechanical properties of the heart after reperfused myocardial infarction (MI). Advances in ultrasound speckle‐tracking allow strain analyses to be performed in small‐animal models, enabling the assessment of regional mechanical function. Methods and ResultsAfter 1 hour of coronary occlusion followed by reperfusion, mice received 10 mg/kg per day of ivabradine dissolved in drinking water (n=10), or were treated as infarcted controls (n=9). Three‐dimensional high‐frequency echocardiography was performed at baseline and at days 2, 7, 14, and 28 post‐MI. Speckle‐tracking software was used to calculate intramural longitudinal myocardial strain (Ell) and strain rate. Standard deviation time to peak radial strain (SD Tpeak Err) and temporal uniformity of strain were calculated from short‐axis cines acquired in the left ventricular remote zone. Ivabradine reduced heart rate by 8% to 16% over the course of 28 days compared to controls (P<0.001). On day 28 post–MI, the ivabradine group was found to have significantly smaller end‐systolic volumes, greater ejection fraction, reduced wall thinning, and greater peak Ell and Ell rate in the remote zone, as well as globally. Temporal uniformity of strain and SD Tpeak Err were significantly smaller in the ivabradine‐treated group by day 28 (P<0.05). ConclusionsHigh‐frequency ultrasound speckle‐tracking demonstrated decreased left ventricular remodeling and dyssynchrony, as well as improved mechanical performance in remote myocardium after heart rate reduction with ivabradine.

Microfluidics-based production of stable microbubbles for ultrasound contrast enhancement or drug/gene delivery allows for precise control over microbubble diameter but at the cost of a low production rate. In situ microfluidic production of microbubbles directly in the vasculature may eliminate the necessity for high microbubble production rates, long stability, or small diameters. Towards this goal, we investigated whether microfluidic-produced microbubbles directly administered into a mouse tail vein could provide sufficient ultrasound contrast. Microbubbles composed of nitrogen gas and stabilized with 3 % bovine serum albumin and 10 % dextrose were injected for 10 seconds into wild type C57BL/6 mice, via a tail-vein catheter. Short-axis images of the right and left ventricle were acquired at 12.5 MHz and image intensity over time was analyzed. Microbubbles were produced on the order of 105 microbubbles/s and were observed in both the right and left ventricles. The median rise time, duration, and decay time within the right ventricle were 2.9, 21.3, and 14.3 s, respectively. All mice survived the procedure with no observable respiratory or heart rate distress despite microbubble diameters as large as 19 μm.

Molecularly targeted ultrasound contrast agents have been shown to identify previously-ischemic regions of the myocardium by targeting cell adhesion molecules that are displayed on activated endothelial surfaces in response to ischemia/reperfusion injury. An improved method for targeted microbubble (MB) delivery using constant infusion and dual targeted microbubbles is presented to enable better characterization of microbubble interaction with activated endothelium. Ten male C57BL/6 mice underwent 15-minute Left Anterior Descending (LAD) coronary occlusion followed by 2-hour reperfusion. Immediately after the 2-hour reperfusion, mouse hearts were imaged with a Sequoia scanner and 15L8 transducer. All mice received three of five MB preparations in random order: Sialyl Lewis X MB (MBX), anti-P-Selectin (MBP), anti-VCAM MB (MBV), Sialyl Lewis X + anti-VCAM MB (MBD), and isotype control MB (MBI). Subsequently, mouse hearts were excised and stained with TTC & Phthalo blue dyes to assess infarct size and area at risk, respectively. Video data indicated that the binding of targeted MB to ischemic regions was significantly higher for MBD, MBX, MBP, and MBV, compared to MBI (p<;0.05). Furthermore, normalized intensity for MBD was higher than both MBX and MBV (p<;0.05). Using Phthalo blue dye as the gold standard, receiver operating curve analysis demonstrated accurate detection of both ischemic and non-ischemic regions using MBD (AUC = 0.90), MBX (AUC = 0.89), MBP (AUC = 0.87), and MBV (AUC = 0.97).

Microbubble production by microfluidic devices for ultrasound contrast enhancement allows for precise control over microbubble diameter but at the cost of low production rate and poor microbubble stability. In this work, we investigated whether microbubbles produced by a microfluidic device could provide sufficient ultrasound contrast enhancement when directly injected into the mouse tail vein. Microfluidic-produced microbubbles composed of nitrogen gas and stabilized with 10% dextrose and 3% bovine serum albumin were injected for 10 seconds into the tail vein of wild type C57BL/6 mice. Short-axis ultrasound images of the right and left ventricle were acquired at 12.5 MHz and image intensity over time was analyzed. Microbubble production rates ranged between 2.5×105 and 8.3×105 microbubbles/s, and microbubble diameters were between 9.1 and 19 μm. In all cases, microbubbles were observed in both the right and left ventricle, although the average contrast enhancement was approximately 13.5 dB lower in the left ventricle than in the right ventricle. All mice survived the procedure with no observable respiratory or heart rate distress. The results of this work suggest that on-site production and immediate administration to the murine vasculature may eliminate the necessity for high microbubble production rates, long-term stability, or small microbubble diameters.

Introduction: Echocardiography using targeted microbubbles (MB) has been shown to detect previously-ischemic regions of myocardium by molecularly targeting cell adhesion molecules that are displayed on the activated endothelial surface after ischemia/reperfusion injury. We hypothesized that the accuracy and sensitivity of detection could be improved by constant infusion and dual-targeting, respectively. Methods: Ischemia/reperfusion injury was induced in 8 male C57BL/6 mice with 15-min LAD occlusions. After 2h of reperfusion, hearts of the closed-chest mice were imaged in long-axis with a Sequoia 15L8 transducer at 14 MHz. Each mouse received 3 of 4 MB preparations in random order: Sialyl Lewis X MB (MBX), anti-VCAM MB (MBV), Sialyl Lewis X + anti-VCAM MB (MBD), and isotype control MB (MBI). MB concentrations were approximately 200x10^6/mL, and 100 μL were infused at a constant rate of 25 μL/min via tail vein catheter. After imaging, mouse hearts were excised and stained with TTC/Phthalo blue dye to assess area at risk, and compared to regional time intensity curve (RTIC) retrieved from video data (Panel A). Results: Example RTIC for MBD shows enhanced MB signals during and after MB infusion in the anterior apical region of the heart (Panel A). In Panel B, normalized intensities are significantly higher for MBD, MBX, and MBV, compared to MBI (p<0.05). Furthermore, normalized intensity for MBD is higher than both MBX and MBV (p<0.05). Receiver operating curve analysis using blue dye as the gold standard showed accurate detection of both ischemic and non-ischemic regions using MBD (AUC = 0.90), MBX (AUC = 0.89), and MBV (AUC = 0.97). Conclusions: Molecular imaging of endothelial cell adhesion molecules in closed-chest mice is possible using constant infusion of MB. Dual-targeted MB yield greater MB signals than single-targeted MB when delivered using constant infusion, improving both the sensitivity and accuracy of this technique to measure risk regions in intact mice.

Mathematical models of varying complexity have proved useful in fitting and interpreting regional cardiac displacements obtained from imaging methods such as ultrasound speckle tracking or MRI tagging. Simpler models, such as the classic thick-walled cylinder model of the left ventricle (LV), can be solved quickly and are easy to implement, but they ignore regional geometric variations and are difficult to adapt to the study of regional pathologies like myocardial infarctions. Complex, anatomically accurate finite-element models work well, but are computationally intensive and require specialized expertise to implement. We developed a kinematic model that offers a compromise between these two traditional approaches, assuming only that displacements in the left ventricle are polynomial functions of initial position and that the myocardium is nearly incompressible, while allowing myocardial motion to vary spatially as would be expected in an ischemic or dyssynchronous LV. Model parameters were determined using an objective function with adjustable weights to account for confidence in individual displacement components and desired strength of the incompressibility constraint. The model accurately represented the motion of both normal and infarcted mouse LVs during the cardiac cycle, with normalized root mean square errors in predicted deformed positions of 8.2 ± 2.3% and 7.4 ± 2.1% for normal and infarcted hearts, respectively.

Regional tissue motion estimation using medical ultrasound images is an important first step in solving many problems with significant clinical applications that include non-invasive assessments of myocardial function in echocardiography. Many existing techniques for estimating motion are based on block-matching using B-mode image intensity, and phase information is often neglected. In digital image processing, it is widely accepted that the information carried in the image phase is more important than information carried by its amplitude. In this work, we implemented a phase-based tracking technique using the monogenic signal to estimate cardiac motion in both synthetic and in vivo mouse heart data. In vivo mouse heart data were acquired with a 30MHz transducer array and synthetic data were simulated with Field II using comparable parameters. The tracking results were compared with a reference speckle tracking analysis method: minimum sum of absolute different (MSAD). The tracking errors for in vivo mouse heart data were 5.3±1.6% and 7.1±2.1% for phase-based and MSAD, respectively. Phase-based displacement estimation via monogenic signal is an improved alternative to conventional block-matching techniques as it makes use of image signal phase information. Phase-based motion tracking using the monogenic signal is accurate and computationally inexpensive.

Arterial stiffness has been shown to be a predictor of cardiovascular events and pulse wave velocity, a measure of arterial stiffness, has been shown to be correlated with clinical outcomes. The ability to measure changes in the mechanical properties of an arterial vessel wall may therefore enable detection of early manifestations of cardiovascular disease. In this study, high frequency (30MHz) ultrasound was used to scan human common carotid arteries, in order to determine arterial wall motion as a measure of arterial stiffness. A 2D RF speckle tracking algorithm was used to calculate wall motion and a mathematical model, based on a viscoelastic tube was developed to model radial direction motion. No significant difference in radial motion was observed between the intima and adventitia of the imaged vessel wall. Simulations using the model accurately predicted arterial radial motion as determined by speckle tracking - exhibiting less than a 5% mean error over 2 cardiac cycles.

Background: Many of the existing dyssynchrony metrics rely on myocardial strain measurements, including circumferential uniformity ratio estimate (CURE) and related metrics such as RURE. Since strain is a derivative of displacement, amplification of noise in strain calculations due to noisy displacement data is common, thus compromising strain-based dyssynchrony metrics. We propose a novel displacement-based dyssynchrony metric, called Dyskinesia Index (DI), that is sensitive and consistent in detecting left ventricular (LV) dyssynchrony, particularly dyskinesia. Methods: Long-axis (LA) cines of 7 C57BL/6 mice were captured with a 30MHz ultrasound transducer. Baseline (B) data were acquired prior to myocardial infarction (MI) induced by a 1-hr occlusion of the LAD followed by reperfusion. Thereafter, cines were acquired at 2, 4, 7, 14 and 28 days post-MI. Displacements were quantified using speckle tracking on the LA cines. DI was determined using Fourier analysis of radial displacement distributed along the endocardial wall, and is the square root of the ratio of the zero-order power term to the sum of zero-order and a scaled first order power terms. Results: In Fig. 1, DI is compared to RURE at B and at 5 time-points post-MI. DI values at B are consistently close to 1 (0.96±0.01), and decrease significantly early after MI due to dyskinesia, reaching a minimum at D4 (0.64±0.08). After the first week, motion in the infarct zone becomes more akinetic due to scar formation and DI values partially recover to a plateau late after MI. ANOVA reveals significant differences between B and all post-MI time-points (p < 0.05). In contrast, RURE is less sensitive to dyskinesia at D4 and post-MI estimates are only significant from B for after D4 (p < 0.05). Conclusion: Strain-based measures of dyssynchrony are error-susceptible without proper filtering of displacement data. DI has been shown to work well in ultrasound B-mode images, where image artifacts can degrade displacement data.

Introduction: Ivabradine selectively inhibits the If channel of the sinoatrial node, slowing heart rate. Few studies have examined the effects of ivabradine on the mechanical properties of the hear...

Fine spatial and temporal resolution of 3D displacement field in the mouse left ventricle (LV) is reconstructed from finely sampled stacks of 2D orthogonal images. Using orthogonal 2D image stacks allows for each of X, Y and Z direction displacement vectors to be detected. Serial short-axis (SA) and long-axis (LA) images of both healthy and infarcted mouse left ventricles were acquired at 0.5mm intervals using a linear array transducer operating at 30MHz. Myocardial motion was tracked using a 2D minimum sum of absolute difference (MSAD) speckle tracking technique. In regions experiencing image artifacts or signal dropout, an incompressible tissue mathematical model was employed, and motion was corrected based on a weighted average of tracked motion and model predicted values. Displacement error was computed based on the ratio of final displacement to the length of the trajectory through the entire heart cycle. By incorporating an incompressible LV mathematical model, tracking error was reduced from 8.4±1.4% to 5.6±1.2%. 3D analyses of cardiac motion provide a more comprehensive assessment of post-infarct ventricular function than conventional 2D analyses. For example, after evaluating the relative displacement magnitudes of post-infarct hearts, dysfunctional myocardium was localized to the apical-anterolateral region of the LV. In the dysfunctional tissue, average radial, circumferential, and longitudinal displacements were reduced by 45.1±9.8%, 43.5±8.2%, and 52.4±9.2%, respectively.

A general filtering method, called the singular value filter (SVF), is presented as a framework for principal component analysis (PCA) based filter design in medical ultrasound imaging. The SVF approach operates by projecting the original data onto a new set of bases determined from PCA using singular value decomposition (SVD). The shape of the SVF weighting function, which relates the singular value spectrum of the input data to the filtering coefficients assigned to each basis function, is designed in accordance with a signal model and statistical assumptions regarding the underlying source signals. In this paper, we applied SVF for the specific application of clutter artifact rejection in diagnostic ultrasound imaging. SVF was compared to a conventional PCA-based filtering technique, which we refer to as the blind source separation (BSS) method, as well as a simple frequency-based finite impulse response (FIR) filter used as a baseline for comparison. The performance of each filter was quantified in simulated lesion images as well as experimental cardiac ultrasound data. SVF was demonstrated in both simulation and experimental results, over a wide range of imaging conditions, to outperform the BSS and FIR filtering methods in terms of contrast-to-noise ratio (CNR) and motion tracking performance. In experimental mouse heart data, SVF provided excellent artifact suppression with an average CNR improvement of 1.8 dB (P < 0.05) with over 40% reduction (P < 0.05) in displacement tracking error. It was further demonstrated from simulation and experimental results that SVF provided superior clutter rejection, as reflected in larger CNR values, when filtering was achieved using complex pulse-echo received data and non-binary filter coefficients.

Local incompressibility can be used to improve fitting and analysis of ultrasound-based displacement data using a heart model. An analytic mathematical model incorporating inflation, torsion, and axial extension was generalized for the left ventricle. Short-axis and long-axis images of mouse left ventricles were acquired using high frequency B-mode ultrasound and myocardial displacements were determined using speckle tracking. Deformation gradient components in the circumferential and longitudinal directions were fitted using linear regressions. The slopes of these lines were then used to predict motion in the radial directions. The optimized kinematic model accurately predicted the motion of mouse left ventricle during filling with normalized root mean square error of 4.4±1.2%.

A singular value filter (SVF) is proposed for rejection of stationary clutter artifact in medical ultrasound. The SVF approach operates by projecting the original data, consisting of ensembles of complex echo data, onto a new set of bases determined from principal component analysis (PCA) using singular value decomposition (SVD). The efficacy of SVF is based on the principle that a stationary clutter signal, with perfect correlation through ensemble length, can be characterized by only the first PCA basis function, whereas significant energy contribution in the secondary PCA basis functions is necessary to describe motion and decorrelation attributed to underlying tissue structures. In contrast to many other PCA-based filtering approaches, SVF determines filter coefficients adaptively from the singular value spectrum of the original data. It is demonstrated that complex echo data is critical to the efficacy of SVF as it provides singular values that exhibit a monotonic relationship with motion complexity, and thus, provide a good means of identifying local regions of clutter. SVF is compared to a separate PCA-based technique, referred to as the blind source separation (BSS) method, as well as a frequency-based finite impulse response (FIR) clutter filter. Performance is quantified in simulated lesion images and SVF is applied to experimental mouse heart imaging data acquired from a Vevo2100 scanner (VisualSonics, Toronto, Canada) at approximately 30MHz center frequency. In simulation with levels of echo correlation expected in mouse heart imaging (0.70 correlation coefficient), SVF provided superior performance (CNR = 4.5dB) over the standard B-mode image (CNR = 2.3dB), BSS-filtered image (CNR = 3.9dB), and FIR-filtered image (CNR = 3.1dB). When SVF was applied to echo data from mouse heart images, stationary artifacts were reduced or eliminated, which enabled myocardium displacement estimates of the underlying tissue structures.

Echocardiography plays a vital role in the evaluation of patients with suspected cardiovascular disease.Heart disease is the leading cause of death worldwide, with myocardial infarction being the major contributing factor to the high morbidity and mortality rates.Preclinical research on heart attack and heart failure is frequently conducted in mouse models of myocardial ischemia.The mouse species is preferred due to its low-cost, short reproductive cycle, and its utility in studying the role of specific genes in the pathophysiology of myocardial ischemia via transgenic and knockout mice.High frequency ultrasound is well suited to in vivo imaging of the mouse heart due to its high spatial and temporal resolution, lack of ionizing radiation, versatility, low-cost, ease of use, and noninvasive nature.To demonstrate the utility of high frequency ultrasound in small animal imaging, I explore several novel approaches to quantify cardiac function in mouse models of myocardial ischemia and infarction.In Chapters 2 and 3, reconstruction of 3D motion in the mouse left ventricle is demonstrated by combining orthogonal 2D displacement fields.A finely sampled matrix of 3D motion vectors is then used to build a kinematic model of the left ventricle using polynomial functions and by assuming that the myocardium is approximately incompressible.In Chapter 4, novel metrics for quantification of left ventricular dyssynchrony and infarct size using ultrasound displacement and strain data, respectively, are presented.Using these metrics, myocardial contractile dysfunction in inducible nitric oxide synthase knockout mice exhibits improved function after myocardial infarction compared to control, wild type mice.Finally, in Chapter 5, a method for identifying previously ischemic regions of the myocardium by using molecularly targeted ultrasound contrast agents is presented.