Therapeutic time window of multipotent adult progenitor therapy after traumatic brain injury

Traumatic brain injury (TBI) affects 1.5 million people in the USA annually, with 50,000 deaths due to acute injury [1]. Heretofore, interventions have been limited to modulating intracranial pressure and cerebral perfusion pressure, while long-term treatments consist of cognitive and behavioral rehabilitation [2]. Cellular therapy presents one of the most promising avenues to improve outcomes by limiting collateral inflammatory damage that follows the immediate trauma. Previously, we utilized cellular therapy to attenuate short-term (3–4 days after injury) and long-term effects (120 days after injury) of TBI in rodents [3‐6]. Specifically, we have demonstrated that multiple doses of multipotent adult progenitor cells [MAPC (2/24 h after TBI)] at 10 million/kg body weight reduced blood-brain barrier (BBB) permeability and reduced the number of activated microglia/macrophages (based on morphology) in the dentate gyrus (no differences were found in the CA1 or CA3). These changes were accompanied by improvement in spatial learning in chronic animals (120 days) [5].

Multipotent adult progenitor cells are patented adult human-derived “off-the-shelf” cellular therapy. Multipotent adult progenitor cells are adherent cell type isolated from the bone marrow and other tissues [7], with their own transcriptome [8], secretome [9], miRNA profile [10, 11], and differentiation capability [8]. After injury, when MAPC are injected intravenously, they get trapped in the lungs and spleen rather than the brain [7]. Intravenous delivery of MAPC after TBI results in an increase in splenocyte and plasma T regulatory cell populations, with an increase in locoregional anti-inflammatory microglia/macrophages [12]. Multipotent adult progenitor cells treatment after cortical contusion injury (CCI) also resulted in a significant increase in anti-inflammatory cytokine interleukin-10 (IL-10) [3]. In vitro experiments demonstrated that direct contact between MAPC and splenocytes is required to modulate activated microglia [12]. In addition, in the absence of spleen, there is no effect of MAPC after TBI on microglia [3], thereby necessitating the involvement of spleen. Thus, these cells do not require honing to the brain to exert paracrine effect. Similar results have been demonstrated with other cell types in TBI [13].

Breakdown of the BBB following TBI results in increased vascular permeability, leading to cerebral edema. At the cellular level, neutrophils are the first to accumulate at about 24 h after injury [14]. Neutrophil accumulation is followed by leukocyte subsets that peak at about 3 days post injury [15]. Monocytes are recruited to the damaged brain in response to local chemokine signals, and once in the brain, they differentiate into macrophages [16]. Inflammatory monocytes are preferentially recruited to the injured brain [17]. Pro-inflammatory mediators peak at 24–36 h after injury [18, 19]. Infiltrating macrophages release pro-inflammatory cytokines such as interleukin-1 (IL-1) and tumor necrosis factor alpha (TNF-α) [18, 20‐22]. The endogenous macrophages of the brain, microglia, are also activated and work to remove necrotic tissue and induce myelin repair [23, 24]. Resident microglia activation depends upon the biochemical milieu [25]. Microglia, in the presence of anti-inflammatory cytokines such as interleukin-4 (IL-4) and IL-10, are generally in a resting/ramified state [26]. In an unperturbed environment, microglia have a distinct morphology, consisting of a small, static cell body with dynamic and branched processes (inactivated). After a CNS injury, microglia are activated and transform into a distinct amoeboid-like morphology (activated). In stroke models, adult progenitor cell therapy leads to decreased tissue levels of pro-inflammatory cytokines such as interleukin-1 alpha (IL1α), interleukin-1 beta (IL1β), and TNF-α and interleukin-6 (IL-6) and increased levels of anti-inflammatory cytokine such as IL-10 [27, 28]. A recent clinical trial in ischemic stroke with MAPC demonstrated an advantage to treating at a time point less than 36 h post injury for MAPC treatment [29]. Similarly, in TBI, we hypothesized that there is an optimum clinically relevant time window of MAPC treatment in order for the therapy to attenuate the chronic harmful effects of TBI. To test this hypothesis, a series of experiments were completed to investigate the critical window of MAPC treatment to preserve the BBB, improve spatial learning, and attenuate activated microglia.

All protocols involving the use of animals were in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the University of Texas Health Science Center Institutional Animal Care and Use Committee (AWC-11-120).

Our data provide evidence that there is a critical window of 24–36 h post injury in which MAPC treatment reduces BBB permeability, improves cognitive behavior, and decreases activated microglia/macrophages, and these changes are independent of neurogenesis. These data parallel the recently completed ischemic stroke phase II clinical trial and should guide clinical testing paradigms. Furthermore, this study replicates previous results in two species [3‐5], adding robustness to the observations.

TBI results in the disruption of the BBB, which allows for an increase in vascular permeability [3]. In the absence of macrophages from the periphery, there is a reduction in microglial activation [32]. Microglia, the resident macrophages of the brain, when activated following injury, remove necrotic tissue and induce myelin repair [33]. Thus, by clearing out the dead cells and debris, activated microglia may serve a neuroprotective role. However, in many neurological diseases, microglia can have sustained secretion of inflammatory cytokines and cytotoxic molecules—such as IL-1, TNF-alpha, and nitric oxide—compounded with excessive neuronophagia, can have deleterious effects [18, 33] . MAPC attenuate BBB permeability by directly interacting with the spleen to alter the systemic immunologic/inflammatory response [3, 4]. Subsequently, pro-inflammatory microglia polarize into an anti-inflammatory phenotype. Though the precise mechanism by which this happens is yet unknown, splenocytes were shown to release anti-inflammatory cytokines such IL-10, IL-4, and others [12]. By reducing BBB permeability, there is a likely reduction of infiltrating peripheral macrophages, and therefore, a reduction in microglial activation, and conversion of pro-inflammatory microglia into anti-inflammatory microglia. After a CNS injury, pro-inflammatory mediators such as neutrophils and cytokines peak at 24–36 h post injury followed by anti-inflammatory cytokines for 96 h post injury [14, 18, 34]. We observed a significant decrease in BBB permeability using Alexa Fluor 680 [35], when MAPC were administered at 2/24 h (Fig. 1a), but not post 24 h after injury. Surprisingly, MAPC treatment at 6/24 did not significantly attenuate BBB, though there was a trend in reduction of BBB at 10–60K (p = 0.06). Treatment with MAPC after 24 h was ineffective in attenuating BBB permeability.

Similar to the BBB data, our cognitive behavioral data also support a critical time window of MAPC treatment of 24 h post injury. We found an improvement in spatial learning as measured by the Morris water maze if MAPC were administered prior to 24 h (Fig. 2). Spatial learning heavily relies on the circuitry of the hippocampus [36]. Acquisition and spatial localization of relevant visual cues are consolidated, retained, and recalled over the training and probe trials [37]. The dentate gyrus is a critical locus in the hippocampus that is involved in the acquisition of spatial memory [38, 39]. Lesions of the hippocampus, dentate gyrus, and ventral subiculum result in poor post training probe trials [38, 40]. In our experiments, we observed improvement in the latency on day 6 of training, when MAPC were administered 24 h post injury. Specifically, we saw improvement with doses administered at 2/24 and 6/24 h after injury on training day 6 (Fig. 2b). These results are similar to our previously observed changes at 2/24 [5]. We observed moderate improvements in latencies when cells were administered at 12/36 and no change when cells were administered at 36/72 h after injury. Both (12/36 and 36/72) were not significantly different than CCI (Fig. 2b). In addition, we observed significant improvement in the recall of the hidden platform during the probe trials. In the probe trial, when the hidden platform is removed, animals that received doses at or prior to 24 h after injury (2/24 and 12/36) demonstrated a spatial bias to the quadrant of the former location of platform (Fig. 2c). Spatial bias to the former location was also observed with platform proximity duration. Significant increases in platform proximity duration were observed in animals that received MAPC treatment at 2/24 only (Fig. 2d). Both, north quadrant duration and platform proximity duration, indicated that MAPC treatment given prior to 24 h is most effective in improving spatial memory. MAPC treatment given at 36/72 h after injury had worse outcomes than CCI (Fig. 2). Recent data from phase II clinical trial in ischemic stroke patients treated with MAPC also demonstrated similar results. Patients given MAPC 36 h after stroke were no different than patients given placebo. Patients that were given MAPC at 24 h showed significant clinical improvement over patients given placebo [29]. Our preclinical data parallels the phase II clinical data (Fig. 2).

Neuronal function and subsequent behavior can be affected negatively by neuroinflammation, which in part is mediated by activated microglia [41]. Case analyses of TBI patients have shown presence of activated microglia with white matter degeneration ranging from months up to 47 years [42, 43]. Activation states of microglia can be categorized by morphological features and/or expression of surface or intracellular markers, as previously described [23, 25, 44]. Twenty-four to 120 h after TBI, MAPC treatment alters the ratio of pro-inflammatory (CD86+: activated) microglia/macrophages to anti-inflammatory (CD206+: inactivated) microglia/macrophages in favor of an inactivated state [12]. In the absence of treatment after injury, we observe an increase in pro-inflammatory (activated) vs anti-inflammatory (inactivated) microglia/macrophages [45]. Using IBA1 to identify microglia/macrophages and subsequently categorizing them based on morphology, we observed a decrease in activated microglia/macrophages in the dentate gyrus ipsilateral to the injury when given doses at 2/24 h after injury when measured at a chronic time point of 120 days after injury [5].

Similarly, in our current experiments, we observed a significant decrease in activated microglia/macrophages in the dentate gyri ipsilateral to the injury when MAPC were given at 2/24 (Fig. 3). Interestingly, we found a significant decrease in activated microglia/macrophages in the contralateral dentate gyri as well (Fig. 4). However, there were no significant changes in the inactivated phenotype of microglia/macrophage at 2/24. There were decreases in activated microglia/macrophages with two doses (6/24 and 12/36) when compared to CCI, but these changes were not significant (Fig. 5a–d). When animals were treated with MAPC at 36/72 h after the injury, there was an increase in activated microglia/macrophages (Fig. 5e, f). These results indicate that MAPC treatment is effective at reducing activated microglia/macrophages when administered 2/24 h after injury, but when administered 36/72 h after injury, there is an increase in activated microglia/macrophages. Future studies will correlate pro- and anti-inflammatory surface markers with activated/inactivated microglial/macrophage morphology. Since neuronal function and behavior can be negatively affected by neuroinflammation [41], a reduction in activated microglia/macrophages might indeed contribute to the improvement in spatial learning observed in the Morris water maze (Fig. 2). Another possibility for the improvement is spatial learning after MAPC treatment might be due to neurogenesis. Mammalian neurogenesis can be triggered due to a proliferation of endogenous neural stem cells [46] which can replace populations of damaged neurons. Recent evidence by Kota et al. demonstrated an increase in neurogenesis with cellular therapy in the short term only (7 days after injury) [47]. However, we observed an increase in neurogenesis in the contralateral hippocampus at 12/36 only in which we observed changes in the time spent in the north quadrant (Fig. 2c). Alternatively, there were significant reductions of DCX-positive cells in the contralateral hippocampi at 2/24 and 6/24. Perhaps, MAPC treatment does not affect neurogenesis. It is also possible that neurogenesis occurs acutely after injury [48] and measured at 4 months post injury in this experiment does not capture those findings.

Intravenous injections of MAPC at or before 24 h resulted in improvement of the BBB, improved cognitive behavior, and attenuated activated microglia/macrophages in the dentate gyrus. In addition, neurogenesis does not play a role in the improved behavior with MAPC treatment after injury. These data will be important in guiding future clinical dosing strategies for treating TBI with MAPC.