MRI Metrics of Cerebral Endothelial Cell–Derived Exosomes for the Treatment of Cognitive Dysfunction Induced in Aging Rats Subjected to Type 2 Diabetes

Nearly one in three older adults meet current American Diabetes Association criteria for diabetes, with the large majority (90–95%) of patients having type 2 diabetes mellitus (T2DM), which affects 24 million Americans (1). T2DM induces neurovascular dysfunction and is a risk factor for cognitive deficits (2), and neurovascular dysfunction in midlife is highly related to the development of Alzheimer disease and dementia later in life (2–4). We thus propose that early interventions to improve neurovascular function could lead to a reduction of evolving cognitive deficits in T2DM.

Exosomes are small extracellular vesicles secreted by most cell types in all living systems (5). Exosomes play critical roles in intercellular communications under physiological and pathophysiological conditions by transferring their cargo of biological materials to recipient cells, leading to change of recipient cell biological function (5). We and others have demonstrated that exosomes derived from mesenchymal stromal cells improve neurological outcome in rodent models of stroke and traumatic brain injury (6,7). The parent cells determine the exosome cargo profile, which affects recipient cell biological function (8,9). In the current study, we tested the hypothesis that exosomes derived from healthy cerebral endothelial cells (CEC-Exos) improve neurovascular function in rats with T2DM.

To test our hypothesis, we used MRI to noninvasively examine the effect of CEC-Exos on cerebral blood flow (CBF), blood-brain barrier (BBB) permeability, and relaxation time constants T₁ and T₂ of cerebral tissues. The associated selected MRI metrics and region of interest (ROI) are based on previous investigations (10). Among the current routinely measured MRI metrics, diffusion-related measurements are suitable for studying white matter (11). In addition, we have previously demonstrated that T₁ and T₂ measurements are sensitive to aging and diabetic pathophysiological alterations in white and gray matter (10). CBF and BBB are important cerebrovascular parameters, and their MRI metrics exhibit differential features in white and gray matter (10). Therefore, MRI-measured brain tissues in the current study included CBF, BBB permeability, T₁ and T₂ of the corpus callosum as pure white matter, cortex as relative pure gray matter, and hippocampus because of its sensitivity to glucose homeostasis (12) and long-term memory and cognitive changes (13). Only performing a single time point MRI measurement is a caveat of the current study; however, the MRI abnormalities of the 2-month diabetic brain in rats have been previously investigated (10). In the current study, 2-month T2DM is set as a baseline where the MRI of T2DM rats was performed at the 3-month T2DM time point with or without exosome treatments.

All experimental procedures were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the institutional animal care and use committee of Henry Ford Health System.

T2DM was induced in male Wistar rats at 13 months of age (Charles River Laboratories, Wilmington, MA) by a single intraperitoneal injection of both 210 mg/kg nicotinamide (NTM) (Sigma-Aldrich, St. Louis, MO), a water-soluble vitamin with anti-inflammatory actions to prevent apoptosis in cells exposed to agents that induce oxidative stress, and 60 mg/kg streptozotocin (STZ) (Zanosar; Sigma-Aldrich), a naturally occurring chemical that is particularly toxic to insulin-producing β-cells of mammalian pancreas. This rat model of middle-aged T2DM has been demonstrated to produce non–insulin-dependent diabetic syndromes that resemble human T2DM (14,15). The rats with nonfasting plasma glucose (Accu-Chek; Roche Diagnostics, Indianapolis, IN) concentrations >250 mg/dL at 2 weeks after STZ-NTM administration were selected for the MRI scan later. A group of T2DM rats (n = 10) that randomly received twice-weekly intravenous administration (via a tail vein) of exosomes (1 × 10¹¹ particles per injection) for 4 weeks starting 2 months after diabetes induction were selected for MRI scanning at 3 months after STZ-NTM administration. These rats served as the CEC-Exos treated rats in experiments where exosomes were isolated from the supernatant of cultured CEC-Exos harvested from healthy young adult male rats (3–4 months) by means of ultracentrifugation (7). MRI was also performed on a control group of age- and T2DM-matched rats (n = 10) at 3.0 months after STZ-NTM administration without exosome treatment, where saline as a vehicle was substituted for exosomes.

CECs were isolated and cultured following our previously reported protocol (16). Briefly, the cerebral cortex and subcortex areas were separated from rat brains and homogenized with RPMI medium (11875093; Thermo Fisher Scientific, Waltham, MA). CECs were isolated and then cultured in endothelial cell culture medium (R819-500; Cell Applications, San Diego, CA). Passage 2–4 CECs were used in the current study. The isolation and characterization of CEC-Exos were performed using our previously published protocol (17). Briefly, exosomes were isolated from culture medium of primary rat CECs using the differential ultracentrifugation method. The isolated exosomes had an average size of ∼78 nm and contained exosome markers as assayed by Nanoparticle Tracking Analyzer (NS300; Malvern Panalytical, Malvern, U.K.), transmission electron microscopy, and Western blots. Exosome number was determined based on particle numbers identified by the Nanoparticle Tracking Analyzer. The number of CEC-Exos used in the current study was selected on the basis of our published studies (17,18).

MRI scans were performed using a 7T system (Bruker BioSpin, Ettlingen, Germany). During MRI scans, rats were anesthetized using medical air (1.0 L/min) with isoflurane (1.0–1.5%). Stereotactic ear bars were used to minimize movement, and rectal temperature was maintained at 37.0 ± 0.5°C using a feedback controlled water bath (YSI Inc., Yellow Springs, OH). A birdcage transmitter and a quadrature half-volume receiver were used in the MRI scans for all rats.

Three-dimensional variable flip angle (2°, 5°, 10°, 15°, 20°, and 25°) spoiled gradient-recalled echo sequence was selected for in vivo T₁ mapping (19). A 128 × 64 × 16 matrix was fitted for a 32 × 32 × 16-mm³ field of view (FOV), with time of repetition (TR) and time of echo (TE) as 30 ms and 3.8 ms, respectively. T₂ mapping was acquired by using a multislice (13 slices) and multiecho (six echoes: TE as 15 ms and equally to 90 ms) T₂-weighted spin echo imaging sequence, with TR = 4.5 s, 128 × 96 matrix fitting 32 × 32-mm² FOV, and 1-mm slice thickness. Contrast-enhanced (CE) T₁-weighted imaging (T1WI) was composed of two T1WIs before and 6 min after tail vein injection of imaging contrast agent gadolinium-diethylenetriamine penta-acetic acid (Gd-DTPA) (Magnevist; Berlex Inc., Montville, NJ) at a dose of 0.4 mL/kg, where TR = 500 ms and TE = 8 ms for the T1WIs and with the same geometric parameters as in T₂-weighted imaging (T2WI).

Regional CBF was estimated by perfusion MRI using the pulsed arterial spin labeling technique (20,21). Single-shot echo-planar readout was applied for each 1.6-mm-thick slice for a total of five slices perpendicular to the magnetic field direction, with a 0.4-mm gap between each neighboring slice and caudal 0.5 mm off center to match slices of T2WI. Complementary pulsed arterial spin labeling imaging parameters were FOV = 32 × 32 mm², matrix = 128 × 128, flip angle = 90°, frames = 201, TR = 3,000 ms, TE = 18 ms, start of label saturation = 500 ms, end of label saturation = 800 ms, and inversion time = 1,000 ms. Bipolar crusher gradients were used with a cutoff speed of 60 cm/s. The 80-mm labeling slab was positioned with a 6.1-mm gap caudal to the imaging slab.

The T₂ and CBF maps were automatically generated by the postprocessing software of the MRI system. The CE-T1WI map was produced by subtracting the pre–Gd-DTPA T1WI image from the corresponding post–Gd-DTPA T1WI (22), performed with Eigentool, an in-house–customized software program (23). All T₁ fittings performed to produce T₁ maps via nonlinear least squares regression (24) were executed using MATLAB (MathWorks, Natick, MA). Final T₁ maps were equivalent to a 128 × 128 matrix in a plane of 16 slices with a 1-mm slice thickness.

T₁, CBF, and CE-T1WI maps were then coregistered slice by slice with the T₂ map and were finally transformed into the same image format for Eigentool. Since the geometric parameters of other pulse sequences in the experiment were set to match the T2WI, only scan-to-scan misregistration caused by head movement needed to be corrected by rigid body alignment of each scan to the T₂ map (25,26).

The distortions in echo planar readout CBF maps were corrected by registration. For converting units of the CBF map (mL/100 g/min), the conversion coefficient was 0.24 on the basis of current sequence parameters, with a blood water partition coefficient of 0.9 mL/g in the cerebral tissue, inversion efficiency of equilibrium magnetization of 0.95, and T₁ of the water in arterial blood of 2,200 ms.

For evaluation of cognitive impairments, the modified Morris water maze (MWM) test and the odor recognition test were performed at 4 months after induction of T2DM in rats before sacrifice (27). MWM mainly assesses hippocampal-dependent spatial learning function (28), while a social odor recognition test evaluates olfactory learning and memory (29).

In the current study, the MRI measurements of brain white matter or gray matter were obtained, respectively, from the corpus callosum or cortex. As indicated in Fig. 1, according to a T2WI (Fig. 1A) and its complementary diffusion fractional anisotropy map (Fig. 1B) (10), the ROIs of corpus callosum (Fig. 1C) and cortex (Fig. 1D) were outlined by thresholds in corresponding maps, where selected pixels along the edge of brain parenchyma close to the skull would be withdrawn from the cortex ROI because of their apparent abnormal values due to edge and partial volume effects. An additional ROI in the current study was the cerebral tissue of hippocampus (Fig. 1E), a mixture of cerebral tissue white and gray matters. The hippocampal area was outlined by its increased values in the T2WI (Fig. 1F).

BBB permeability was indicated by the changes of CE-T1WI intensities between pre- and postinjection of Gd-DTPA (22). A dimensionless quantity, percentage, was induced as the ratio of the differences from the subtracted T1WIs to the T1WI without contrast injection.

Data analysis was performed in a blinded fashion. Measurements are presented as mean and SD. Student t test was then applied to pairs of animal groups of every MRI measurement for individual statistical differences. The effect was detected at the 0.05 level.

The data sets generated and/or analyzed during the current study are included in the published article and are available from the corresponding author upon reasonable request.

Although the MRIs from the representative saline- and CEC-Exos–treated animals, i.e., T₁, T₂, CBF maps (Fig. 2), and CE-T1WIs (Fig. 3) do not identify the therapeutic benefits of CEC-Exos in middle-aged T2DM rats, the quantitative MRI neurovascular metrics (Fig. 4) revealed that CEC-Exos induce heterogeneously improved neurovascular changes in gray and white matter as well as in the hippocampus.

MRI analysis showed that T₁ values in the corpus callosum tissue were 1,444 ± 22 and 1,424 ± 47 ms between T2DM rats treated with CEC-Exos and saline, respectively (P > 0.2) (Fig. 4). However, the CEC-Exo treatment significantly (P < 0.02) increased the mean value of T₂ (48.2 ± 3.0 ms), compared with the saline treatment (44.8 ± 2.6 ms). The CEC-Exo treatment also significantly (P < 0.02) increased the CBF (92.9 ± 5.9 mL/100 g/min) compared with the saline (77.6 ± 12.4 mL/100 g/min). CE-T1WI with Gd-DTPA analysis showed no significant (P > 0.2) differences of BBB integrity between the saline (3.0 ± 2.0%) and CEC-Exo (2.3 ± 2.5%) groups. These data suggest that CEC-Exos change white matter of the neurovascular function in white matter by augmentation of CBF and transverse relaxation time T₂.

In contrast to the corpus callosum, the CEC-Exo treatment resulted in a significant (P < 0.02) increase of the cortical T₁ value (1,611 ± 65 ms) compared with saline (1,513 ± 53 ms); however, there were no significant (P > 0.2) differences of the cortical T₂ values between the CEC-Exos (59.1 ± 2.9 ms) and saline (58.6 ± 0.8 ms) groups (Fig. 4). The CEC-Exo treatment significantly (P < 0.02) reduced the cortical intensity enhancement value (1.5 ± 2.5%) because of a decrease of parenchymal Gd-DTPA concentration leaked from blood vessels compared with the saline group (3.5 ± 1.5%), whereas treatment with CEC-Exos did not substantially (P > 0.2) alter cortical CBF versus saline (208.9 ± 32.3 vs 197.9 ± 51.5 mL/100 g/min, respectively). These data suggest that the neurovascular unit in gray matter has distinct responses to CEC-Exos from white matter, with decreased neurovascular leakage of GD-DTPA and increased longitudinal relaxation time T₁.

In the hippocampus (Fig. 4), CEC-Exos versus saline treatment significantly (P < 0.02) increased T₁ values (1,603 ± 43 vs. 1,518 ± 36 ms, respectively) but did not alter T₂ values (63.9 ± 1.9 vs. 63.0 ± 1.4 ms; P > 0.2). Moreover, CEC-Exos significantly augmented CBF (190.2 ± 38.2 vs. 151.8 ± 42.0 mL/100 g/min; P < 0.05) but did not result in a substantial change in Gd-DTPA-enhanced values (2.0 ± 3.8 vs. 3.0 ± 2.2%; P > 0.2). These data suggest that the response of the neurovascular unit in the hippocampus to the CEC-Exo treatment has mixed MRI metrics of gray and white matter.

MWM analysis revealed that T2DM rats treated with CEC-Exos spent significantly (P < 0.01) longer times in the correct quadrant (47.4 ± 4.1% of total time) than T2DM rats treated with saline (40.5 ± 4.7% of total time). In addition, the CEC-Exo treatment, compared with saline, increased the time spent by the T2DM rats exploring a novel object, as assayed by the odor recognition test (53.5 ± 6.4 vs. 46.3 ± 6.9%; P < 0.05). Detailed functional data of the MWM and odor recognition test are presented in Table 1.

The current study demonstrated that treatment of T2DM rats with CEC-Exos significantly increased CBF in the corpus callosum and hippocampus and reduced BBB permeability in the cortex, while CEC-Exos substantially reduced T2DM-induced cognitive deficits. We previously demonstrated that CEC-Exos intravenously administered cross the BBB are internalized by CECs, neurons, and astrocytes (17). Additionally, we demonstrated that CEC-Exos promote axonal growth via a network of miRNAs/proteins that mediates axonal remodeling (18,30,31). In addition to brain, CEC-Exos likely affect the function of other major organs. For example, treatment of diabetic brain with intravenous administration of CD133 exosomes improves cardiac function (32). Exosomes derived from blood plasma of patients with diabetes induce dysfunction of immune cells and human aortic endothelial cells (33,34), which are associated with an increased risk for vascular dysfunction, including cerebrovascular disease (33,34). CEC-Exos reduce cerebral vascular inflammation (17). Together, these data suggest that in addition to their direct therapeutic effect on brain, CEC-Exos may also have a systemic effect, which may feed back to and thereby indirectly affect the diabetic brain (35). Future studies are warranted to identify the contributions of these multiple component effects on the therapeutic benefit of CEC-Exos on diabetic brain and could provide additional information about how a systemic effect of CEC-Exos contributes to the improvement of brain function. Our data demonstrate that noninvasive quantitative MRI neurovascular metrics are able to detect neurovascular functional improvement by CEC-Exos treatment in a regionally specific manner (although the mechanisms underlying the regional preferential effects of CEC-Exos remain to be determined) and that these metrics are associated with a reduction of cognitive deficits in T2DM rats.

Diabetes induces vascular dysfunction (36). Preclinical and clinical studies have shown that hyperglycemia is closely associated with cortical and subcortical regional cerebral hypoperfusion (2,10,37). We previously reported that rats with T2DM for 1.5 months exhibited hypoperfusion in white matter but not in cortical gray matter (10). The current study extends these findings by showing that in addition to gray and white matter, hypoperfuson is present in the hippocampus 3 months after T2DM induction. T2DM rats also exhibited BBB disruption in the cortical gray matter. Importantly, treatment of T2DM rats with CEC-Exos significantly increased CBF in the corpus callosum and hippocampus and robustly reduced BBB permeability in the cortex compared with saline-treated T2DM rats. The increased CBF in the corpus callosum and reduced BBB permeability in the cortex are close to the respective values of the healthy aged rats (10), where the healthy rats (15 months old) were 1 month younger than the 3-month T2DM rats with or without CEC-Exo treatment in the current study. A caveat is the absence of identical age-matched (16-month-old) healthy rats as a control. Given that T2DM rats exhibit reduction of CBF and increased vascular permeability at 1.5 months after induction of T2DM, although the CEC-Exo treatment was initiated at 2 months after T2DM, the present data suggest that CEC-Exos reverses T2DM-induced cerebral vascular dysfunction.

In human, the microvascular densities in white matter are ∼20–49% lower than that in gray matter; for example, microvascular densities were 1,311 ± 326 vessels/mm³ in cortex and 222 ± 147 vessels/mm³ in underlying white matter (38), which may account for the lower CBF in the white matter than in gray matter (39). In the nondiabetic rodent, we found that white and gray CBFs were 91.7 ± 6.3 and 252.8 ± 54.9 mL/100 g/min, respectively (10), in middle-aged rats. T2DM-induced intravascular thrombosis is a major cause of reduced CBF and increased BBB permeability compared with age-match nondiabetic animals (40–42). CEC-Exos may suppress coagulation and inflammatory factors in CECs (5,7), which likely contribute to improvements of cerebrovascular function measured by MRI metrics. Because of the different microvascular densities in white and gray matter, CEC-Exo treatment differentially affects CBF and BBB permeability of white and gray matter. Thus, increased CBF and reduced BBB permeability induced by CEC-Exos likely contribute to the reduction of cognitive deficits in T2DM.

In Wistar rats, brain water content is ∼80%, with regional variations (43). The water content may have the largest influence on the in vivo measures of relaxation times (44). Higher water content would increase the volume portion of bulk water with longer relaxation times and increase mobility of surrounding macromolecules of cerebral tissue, resulting in increases of relaxation times T₁ and T₂ of cerebral tissue. White and gray matter have distinct cellular properties and water content (43,44), leading to alterations of the T₁ and T₂ relaxation times in different ways (45). We previously demonstrated that aging and/or diabetes reduce T₁ and T₂ of white and gray matter (10). The current study showed that CEC-Exos treatment reversed the decreased trend of the T2DM-altered T₁ and T₂, where the elevated T₁ of gray matter and T₂ of white matter induced by CEC-Exos treatment are close to the corresponding values of healthy aged rats (10), suggesting that CEC-Exos treatment increases water content of white and gray matter in T2DM rats.

The hippocampus contains white and gray matter, although there is no absolute pure gray or white matter in the hippocampus. Interestingly, our hippocampal MRI data demonstrated that CEC-Exo treatment of T2DM significantly improved CBF without reducing vascular permeability, as in white matter, whereas it significantly increased T₁ but not T₂, as in gray matter. The hippocampus plays an important role in cognition and long-term memory, while cortex and corpus callosum are also involved in cognitive functions (46). Treatment with CEC-Exos significantly reduces T2DM-induced spatial learning deficit and reduces cognitive function impairment, data that are consistent with improved physical indices measured by MRI.

The MRI results of hippocampus in the current study indicate that the response to the CEC-Exo treatment of a neurovascular unit in T2DM rat brain is independent of and not readily predicted from responses of gray or white matter. In the CEC-Exo-treated versus control (without CEC-Exo treatment) 3-month T2DM rats, T₁ values of hippocampus significantly increased, as in gray matter (cortex); meanwhile, and CBF values of hippocampus significantly increased, as in white matter (corpus callosum). Furthermore, the CEC-Exo treatment of T2DM rats did not significantly reduce hippocampal BBB permeability, unlike in gray matter (cortex), and did not significantly elevate hippocampal T₂ values, unlike in white matter (corpus callosum). Such independent responses were also found in other brain regions with mixed white and gray matter (e.g., thalamus; data not shown).

The current study provides a proof of principle that MRI can be used to monitor therapeutic responses to CEC-Exo treatment in T2DM rats. However, this study is limited in multiple ways, such as only male middle-aged rats were used, and these rats were subjected to a very specific CEC-Exo treatment protocol. In addition, a caveat is that there was an absence of direct histological verification of the MRI metrics used. Related to T₁ and T₂ changes, brain water content was not quantitatively measured, and cellular-based histology was not performed on frozen slice or waxed sections. Real-time CBF measurement was difficult to compare with other measurements of CBF, such as transcranial Doppler ultrasound measurements, which are limited to the large basal arteries and can only provide an index of global rather than local CBF velocity (47). BBB permeability is sensitive to the molecular size of the penetrating dye used, and MRI contrast agents usually differ from the histological staining dyes in BBB permeability measurements (22,48). Thus, additional studies on different populations of rats (e.g., females, aged male and female rats) subjected to various treatment protocols, as well as extensive comparisons of MRI indices with histological measures, are warranted.

The parent cells determine exosome cargo profile, which affects recipient cell biological function (8,9). For example, intravenous administration of exosomes derived from healthy Schwann cells remarkably ameliorate diabetic peripheral neuropathy by improving sciatic nerve conduction velocity and increasing thermal and mechanical sensitivity in T2DM db/db mice (8). In contrast, exosomes derived from high-glucose–challenged Schwann cells promote diabetic peripheral neuropathy in db/db mice (9). However, future experiments are warranted to study whether and how the CEC-Exo cargo improves cerebral vascular and neuronal function in diabetic brain.

In summary, treatment of T2DM rats with CEC-Exos diminished T2DM-impaired neurovascular functions, which likely contribute to the significant reduction of cognitive deficits. The quantitative MRI neurovascular metrics are valuable to noninvasively assay neurovascular function in the brain of T2DM rat.

Funding. This work was supported by Foundation for the National Institutes of Health grants R01 NS079612 (Z.Z.), R01 AG068072 (L.Z.), and R56 AG055583 (L.Z.).

The content is solely the responsibility of the authors and does not necessarily represent the official view of the National Institutes of Health.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. G.D. completed the MRI scans, researched data, and wrote the manuscript. L.L. researched the MRI data. L.Z. performed the animal functional tests. M.C. contributed to discussion and reviewed and edited the manuscript. E.D.-B. maintained the image processing programs. Q.L. processed the MRI images. C.L. cultured and isolated cells. M.W. managed animals and treatments. Z.Z. supervised the study and reviewed and edited the manuscript. Q.J. supervised the MRI experiment and edited the manuscript. Q.J. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.