Multimodal Approaches for Regenerative Stroke Therapies: Combination of Granulocyte Colony-Stimulating Factor with Bone Marrow Mesenchymal Stem Cells is Not Superior to G-CSF Alone

Abstract

Attractive therapeutic strategies to enhance post-stroke recovery of aged brains include methods of cellular therapy that can enhance the endogenous restorative mechanisms of the injured brain. Since stroke afflicts mostly the elderly, it is highly desirable to test the efficacy of cell therapy in the microenvironment of aged brains that is generally refractory to regeneration. In particular, stem cells from the bone marrow allow an autologous transplantation approach that can be translated in the near future to the clinical practice. Such a bone marrow-derived therapy includes the grafting of stem cells as well as the delayed induction of endogenous stem cell mobilization and homing by the stem cell mobilizer granulocyte colony-stimulating factor (G-CSF). We tested the hypothesis that grafting of bone marrow-derived pre-differentiated mesenchymal cells (BM-MSCs) in G-CSF-treated animals improves the long-term functional outcome in aged rodents. To this end, G-CSF alone (50 μg/kg) or in combination with a single dose (10(6) cells) of rat BM MSCs was administered intravenously to Sprague-Dawley rats at 6 h after transient occlusion (90 min) of the middle cerebral artery. Infarct volume was measured by magnetic resonance imaging at 3 and 48 days post-stroke and additionally by immunhistochemistry at day 56. Functional recovery was tested during the entire post-stroke survival period of 56 days. Daily treatment for post-stroke aged rats with G-CSF led to a robust and consistent improvement of neurological function after 28 days. The combination therapy also led to robust angiogenesis in the formerly infarct core and beyond in the "islet of regeneration." However, G-CSF + BM MSCs may not impact at all on the spatial reference-memory task or infarct volume and therefore did not further improve the post-stroke recovery. We suggest that in a real clinical practice involving older post-stroke patients, successful regenerative therapies would have to be carried out for a much longer time.

Keywords: BM MSC; G-CSF; aging; angiogenesis; cell therapy; stroke; translational medicine.

Figure 1

Experimental design and time course of cutaneous sensitivity and sensorimotor integration recovery after stroke therapy. (A) Schematic overview of the experimental design. (B,C) Time course of cutaneous sensitivity and sensorimotor integration recovery after stroke therapy by the adhesive tape removal test. By day 3, post-stroke animals started recuperation and reached significant recovery of function by day 14 in the G-CSF group [(B), filled red squares] as compared to the control group [(B), filled black circles]. The combination of G-CSF and BM MSC showed no significant improvement of recuperation of sensorimotor function [(C), filled green squares] vs controls [(C), filled black circles]. (D,E) Functional recovery on the rotating beam. Control rats began improvement and recovered to 47% by day 56 [(D), (E) filled black circles]. Of the treated groups, best recovery of the bilateral sensorimotor coordination was shown in G-CSF alone that reached 72% of the pre-surgery value [(D), red squares] followed by G-CSF + BM MSC [58%; (E), blue squares]. Data are given as mean ± SEM.

Figure 2

Time course of post-stroke recovery of learning and (spatial) memory by water maze. Representative swim paths are shown in (A–C) and included the start of training (−7d), the pre-surgery path pattern (0d), first testing after stroke (+7d), and the final testing (+56d). The best recovery was seen for the G-CSF group that showed significant improvement of spatial reference-memory between days 21 and 49 in the second quadrant [(D); p = 0.05]. However, in the third quadrant, the performance was temporarily improved between days 14 and 28 in the group treated with G-CSF + BM MSC as compared to the control group [(E); p = 0.019]. Data are given as mean ± SEM.

Figure 3

Edema and stroke volumes by MRI and NeuN immunohistochemistry. (A–C) Perilesional brain edema at day 3 post-stroke, as defined by the region of T2 hyperintensity (A) was not significantly reduced by any treatment (J). (D–F) The second MRI done at day 48 (B) post-stroke revealed much smaller infarcts. (G–I) By immunohistochemistry at day 58, the infarct volumes for controls (G), G-CSF alone (H), and combination treatment (I) were largely similar to those measured by MRI at day 48. The infarct volume after 7–8 weeks post-stroke was not significantly different among the groups (K,L).

Figure 4

Phenotyping of human BMSCs. In the ipsilateral hemisphere, the injected human BMSCs were localized in the corpus callosum as shown for CD166-positive cells [(A), arrows] and CD105-positive cells [(F), arrows]. In our model, the cells most likely entered the injured brain via the lateral ventricle as shown by the CD166-positive cells (B). A fraction (about 1%) of the injected CD166- and CD105-positive cells reached the infarcted area [(C,E), arrows] where they were intermingled with surviving or degenerating neuronal nuclei [(C), arrowheads]. Noteworthy was also the presence of immunopositivity for human nuclei [(D), arrows] that were dispersed between the rat nuclei in the infarcted area [(D), arrowheads]. Cc, corpus callosum; IC, infarct core; LV, lateral ventricle; PI, periinfract.

Figure 5

Post-stroke neurogenesis and angiogenesis. At 8 weeks post-stroke, none of the DCX⁺ cells in the SVZ of control animals co-localized with BrdU-labeled nuclei. Instead, the BrdU-positive nuclei were distributed mainly in the “pinwheel” architecture of the ventricular epithelium (A). The DCX⁺ cells occupied an adjacent, distinct position [(A), arrows]. Some of the DCX⁺ migrated away from the ventricular wall (B). We noted vigorous neurogenesis with many DCX⁺ (arrows) co-localizing with BrdU nuclei in the G-CSF-treated animals [(C); arrowheads] and animals treated with G-CSF + BM MSC [(D), arrows]. (E–G) Post-stroke angiogenesis. In regions adjacent to the infarct scar, we found numerous BrdU⁺ nuclei in the endothelium of newly formed blood vessels in the formerly infarct core [(E), green]. The border to the healthy brain region was abruptly demarcated to the left by NeuN-positive nuclei [(E), red]. Beyond the formerly infarct core, we noted vigorous sprouting angiogenesis as evidenced by RECA/BrdU double positive blood vessels [(F), violet] as well as numerous BrdU⁺ nuclei in the newly formed endothelium [(F), blue] and reconstruction of the basal lamina [(F), green] during the resolution phase of angiogenesis. By number of laminin/BrdU co-localizations, the density of the newly formed blood vessels was significantly higher (threefold, p = 0.01) in the brains of animals treated with the combination G-CSF + BM MSC as compared to controls and G-CSF alone (G). Cc, corpus callosum; IC, infarct core; IR, islet of regeneration; LV, lateral ventricle; PI, periinfract.