Aging is a primary risk factor for a number of diseases and pathologies, including stroke, which is recognized as one of the leading causes of death and disability globally (Johnson et al., 2016; Benjamin et al., 2019). Ischemic stroke is particularly devastating as the long-term phenotypic deficits in cognitive behavior and sensorimotor function build in the days following the insult, even if blood flow to the ischemic tissue is restored [as reviewed by George and Steinberg (2015)]. Ischemia precipitates cellular damage in the brain within minutes, as the loss of blood flow induces metabolic and oxidative stress in the energy-demanding neurons, and nearby glial and cerebrovascular cells. Immediate restoration of blood flow is critical, as the length of ischemia correlates with the extent of tissue damage (Terasaki et al., 2014). To date, the most effective ischemic stroke therapy is fibrinolytic breakdown of the clot with administration of tissue plasminogen activator. Unfortunately, this therapy is time-restricted and requires confirmation of ischemic vs hemorrhagic manifestation, which restricts clinical utility and has limited its use to around 5% of stroke patients each year (National Institute of Neurological Disorders and Stroke rt-Pa Stroke Study Group, 1995). Longitudinal observation-based clinical assessments have identified multiple possible biomarkers of stroke risk and severity, including insulin-like growth factor-1 (IGF-1), which may serve as a point of therapeutic intervention in the days surrounding insult (Katzan et al., 2000; California Acute Stroke Pilot Registry Investigators, 2005; Qureshi et al., 2005; Kleindorfer et al., 2008).

Insulin-like growth factor-1 is a pleiotropic neuroendocrine modulator that has been assessed in both clinical and preclinical studies of stroke [as reviewed by Hayes et al. (2021)]. Both humans and rodents exhibit age-related circulating declines of IGF-1 (O’Connor et al., 1998; Sonntag et al., 2013). IGF-1 levels decline in advanced age, and the extent of this decline is associated with increased risk of stroke incidence, increased mortality, and worsened functional outcomes post-stroke in humans (Johnsen et al., 2004; De Smedt et al., 2011; Tang et al., 2014; Armbrust et al., 2017; Saber et al., 2017). Moreover, altered IGF-1 levels are also associated with increased risk of multiple age-related neuropathologies marked by neurovascular distress, excitotoxicity, and oxidative stress (Watanabe et al., 2005; Sonntag et al., 2013; Ghazi Sherbaf et al., 2018; Gubbi et al., 2018). While many studies associate reduced IGF-1 with increased neurodegeneration, some clinical studies of dementias and stroke see increased IGF-1 levels at the start of pathology, which may be an attempted compensatory mechanism designed to protect the brain (Gubbi et al., 2018; Castilla-Cortazar et al., 2020). Preclinical studies suggest that IGF-1 plays a causal role in the stroke risk and recovery processes by directly regulating cell damage during ischemia [as reviewed by Hayes et al. (2021)]. IGF-1 supplementation in rodent models of middle cerebral artery occlusion and photothrombotic stroke reduces the size of infarcted tissue and the accompanying sensorimotor deficits, even when administered in the days after insult (Liu et al., 2001; Bake et al., 2014, 2016; Parker et al., 2017; Serhan et al., 2019; Hayes et al., 2021). Because IGF-1 regulates numerous cells throughout the body, the cellular mechanism by which IGF-1 protects viability and function in the brain remains unclear. It is commonly inferred that the beneficial effects of IGF-1 are mediated by direct regulation of neurons. However, neurons, astrocytes, endothelial cells, and perivascular cells each express the receptor for IGF-1, IGFR. Together, these cells work to compose the neurovascular unit (also known as neuro-glio-vascular unit) which tightly coordinates physiological responses within the nervous system and is a central target for dysfunction in neurodegenerative diseases and pathologies (Stanimirovic and Friedman, 2012). Additional work is needed to tease apart the impact of each cell type when exogenous IGF-1 is administered to protect against the damage induced by stroke and neurodegenerative diseases.

Animal models of circulating IGF-1 deficiency exhibit impaired neurovascular coupling, compromised blood-brain barrier integrity, and increased prevalence of micro hemorrhages (Toth et al., 2015; Tarantini et al., 2016b, 2017, 2021a; Hayes et al., 2021). Increased reactive oxygen species (ROS) production, increased inflammation, and reduced glutamate and glucose handling machinery accompany the loss of circulating IGF-1 in these models (Toth et al., 2015, 2017). Together, these data suggest that the age-related reduction in IGF-1 likely leads to severe consequences on the structure and function of the neuro-glio-vascular unit. Indeed, in vitro studies have highlighted roles for IGF-1 in promoting specific functions of each cellular component of the neuro-glio-vascular unit. Pharmacological and genetic manipulations of IGF-1 alter the viability and oxidative stress levels of cultured neurons (Wang et al., 2014; Li et al., 2017; Chen et al., 2019). Additionally, IGF-1 signaling in brain vascular endothelial cell cultures promotes angiogenesis and tight-junction formation/integrity (Lopez-Lopez et al., 2004; Bake et al., 2016; Higashi et al., 2020). Genetic reductions of IGFR specifically in endothelial cells impairs stimulation-induced cerebral blood flow, indicating that the loss of IGF-1 signaling in one component of the neurovascular unit results in phenotypic changes commonly observed in the aged brain (Tarantini et al., 2021b). We recently reported that deficiency in IGF-1 signaling in astrocytes impairs glutamate handling in vitro and in vivo, by reducing glutamate transporter expression and availability at the cell surface (Prabhu et al., 2019). Although there are thought to several key drivers to aging pathologies and disease, glutamate over excitation is a central contributor to tissue damage following ischemic stroke (Choi and Rothman, 1990; Moskowitz et al., 2010), it is likely that the age-related loss of IGF-1 may compromise the neuro-glio-vascular unit by weakening the glutamate-buffering capabilities of astrocytes which ultimately exacerbates calcium imbalances, mitochondrial dysfunction, and oxidative stress. While the aforementioned studies highlight the ability of IGF-1 to regulate neurons, astrocytes, and brain-derived endothelial cells, each of these studies was performed under varied conditions. A comparative approach is needed to better-understand how the age-related loss of IGF-1 influences cells within the neuro-glio-vascular unit. Thus, we exposed individual cell cultures and co-cultures to excitotoxic levels of glutamate when IGF-1 signaling was impaired. This approach allows for determination of the cellular mechanism(s) by which IGF-1 exerts its protective effects during ischemic stroke-induced stressors. In addition, it highlights cell-specific responses to a common driver of neuronal dysfunction- glutamate excitotoxicity- which has applications to numerous age-related neurodegenerative disease states.

The age-related loss of IGF-1 has been linked to cognitive impairment, neurodegeneration, and increased susceptibility ischemic stroke and other neurovascular pathologies (Sonntag et al., 2013; Ashpole et al., 2015b). These deficits often come as a disconnect from the enhanced longevity observed in animal models of IGF-1 deficiency. While early-life reductions in circulating IGF-1 do lead to increased lifespan, reduced IGF-1 in advanced age has multiple consequences in animal models as it increases sarcopenia, bone frailty, learning and memory deficits, and cerebrovascular dysfunction (Ashpole et al., 2015a, 2016, 2017; Toth et al., 2015; Tarantini et al., 2016a, 2021a; Farias Quipildor et al., 2019). For example, IGF-1 deficiency in mice decreases stimulation-evoked cerebral blood flow, blood brain barrier integrity, and the strength and flexibility of cerebral arteries. IGF-1 deficiency also impairs the production and release of vasomediator eicosanoids from astrocytes, alters endothelial nitric oxide production, and increases susceptibility to hypertension-induced microhemorrhages (Toth et al., 2015; Tarantini et al., 2017, 2021a; Fulop et al., 2019). Considering this, it is not surprising that IGF-1 has long-been implicated in the risk and severity of ischemic stroke. As described earlier, clinical and preclinical evidence highlight inverse correlation (and causation in animal models) between IGF-1 and the risk/outcome of ischemic stroke (Hayes et al., 2021).

While recent studies are beginning to shed light on the impact of IGF-1 on neurovascular communication in advanced age, there remains a significant gap in our understanding of how individual cell types that normally coordinate signaling together within the neurovascular unit each respond to reduced IGF-1 signaling. In this study, we aimed to compare the effects of IGF-1 deficiency on the cellular responses to high levels of extracellular glutamate. As the predominant excitatory neurotransmitter in the brain, glutamate induces depolarization and initiates multiple calcium signaling cascades. During periods of ischemia or proteotoxic stress, neurons experience energy imbalances and ionic stressors that ultimately lead to terminal depolarization and the dumping of glutamate stores, inflammatory mediators, and oxidative stressors into the extracellular space, which expands the region of distress and malfunction [as recently reviewed (Choi, 2020)]. This excitotoxic cascade is one of the drivers of neuronal loss and glial activation in ischemic stroke and neurodegenerative diseases. Thus, understanding how the loss of IGF-1 alters the glutamate response of neurons, astrocytes, and endothelial cells within the neurovascular unit provides much-needed information on potential cellular mechanisms by which IGF-1 deficiency in advanced age influences health and function of the brain.

Recent evidence from our laboratory indicated that IGFR inhibition reduced the ability of astrocytes to buffer extracellular glutamate by decreasing glutamate transporter availability (Prabhu et al., 2019). Therefore, in this study, we hypothesized that maintenance of IGF-1 signaling in astrocytes is necessary for neuroprotection from glutamate excitotoxicity. We also hypothesized that maintenance of IGF-1 signaling in neurons was essential for limiting overexcitation, as other studies have shown exogenous IGF-1 can protect against excitotoxicity and oxidative stress (Wang et al., 2014; Li et al., 2017; Chen et al., 2019). Our results indicate that while exogenous IGF-1 can protect neurons from glutamate, acute and prolonged loss of IGFR signaling does not exacerbate neuronal death in cultures of pure neurons. Interestingly, inhibition of IGF-1R not only impairs the neuroprotective capabilities of astrocytes, it predisposes astrocytes to glutamate toxicity as well. It is unclear what the mechanism for this enhanced sensitivity may be. Perhaps the lack of glutamate transporter availability with IGF-1 signaling deficiency resulted in increased GluNR activation in the astrocytes as well, since astrocytes are known to express a variety of ionotropic and metabotropic GluNRs (Serrano et al., 2008; Lalo et al., 2011; Bradley and Challiss, 2012; Ceprian and Fulton, 2019). If this were the case, the astrocytes could be experiencing intracellular calcium imbalances and oxidative stress, which have both been shown to increase astrocyte toxicity in response to other stressors.

As mentioned, reductions in IGF-1 in the brain increase ROS levels and mitochondrial dysfunction. We observed significant increases in ROS production in both astrocytes and endothelial cells when IGFR was inhibited at the time of glutamate stimulation. We chose to focus our experimental design on these two cell types and this acute treatment because it did not lead to any changes in observed toxicity. Oxidative stress is expected in neurons that are dying from excitotoxicity, but neither astrocytes nor endothelial cells were sensitive to glutamate when IGFR was acutely inhibited. The increased ROS levels in both cell types indicates that the cells were indeed undergoing a stress response with the combined treatment. While it could be inferred that the ability of astrocytes to return ROS levels back to baseline by 24 h may contribute to their increased resistance to excitotoxic stress, there is a disconnect in the endothelial cells that likely renders this to be more complicated. Perhaps an increase in endothelial cell toxicity would have been observed at later time points. Further studies on the underlying mechanism(s) and consequence(s) of these observations are needed.

Combined triple cultures of neurons, astrocytes, and endothelial cells highlighted a few distinct responses to glutamate and IGFR inhibition than the individual cultures of each cell type. First, the same concentration of glutamate that led to significant increases in neurotoxicity within pure neuron cultures failed to induce toxicity within the triple culture. This was an expected outcome, as astrocytes are known to buffer glutamate and co-cultures of neurons and astrocytes have been shown to reduce the extent of neurotoxicity (Choi et al., 1987; Rosenberg et al., 1992). Second, and more interestingly, prolonged IGFR inhibition in astrocytes prior to glutamate stimulation did not increase the toxicity of astrocytes nor decrease the total cell number of astrocytes within the triple cultures. These effects were pronounced in pure astrocyte cultures, suggesting that co-culturing astrocytes with neurons and endothelial cells afforded them protection against the stressors of IGFR deficiency/glutamate stimulation. Additional studies examining ROS levels, inflammatory mediators, and growth factor production and release with triple cultures exposed to these same stressors may be warranted.

In our recent study examining glutamate uptake in astrocytes, we utilized similar concentrations of the pharmacological inhibitor PPP and did not observe a significant reduction in total cell number. Thus, we were surprised to see a reduction in astrocyte count here. However, in that study, cells were treated when at or close to confluence, which was evident in DAPI-stained microscopy images accompanying those data (Prabhu et al., 2019). In the current study, we aimed to treat at 75% confluence in order to avoid stress of over-growth, however, microscopic live/dead staining analysis showed that even the vehicle-treated controls were not at this level of confluence even 24 h after treatment. Thus, the reduction in cell number observed in pure cultures of astrocytes and endothelial cells was likely due to changes in cell division and growth with IGF-1 deficiency. Nevertheless, it is interesting that the reduction in astrocyte number with prolonged IGFR inhibition was restricted to the pure cultures, and not the triple cultures.

The complication of cell proliferation also limited our ability to utilize genetic approaches in our in vitro studies. IGF-1 is a primary growth factor with autocrine and paracrine functions, and genetic reductions that alter cell proliferation make it difficult to control studies of neuroendocrine hormones. We originally aimed to include siRNA and/or cell-specific knock-outs using our inducible cell-specific IGFR knock-out lines. However, due to obvious insufficient cell proliferation caused by siRNA transfection/transduction (both approaches were attempted) as well as Cre-recombinase induced knock-outs, we concluded that the significant reductions in cell number would confound results. Astrocyte count following knock-out with siRNA or with 4-OH-tamoxifen-induced recombination was less than 50% of that observed in the control-treated wells (siRNA vector transfected or vehicle-treated). Hence, pharmacological interventions were utilized to address our central question of how the loss of IGF-1 impacts the major cells of the neurogliovascular unit. While exogenous IGF-1 was added in a few studies here, the growth media of the cells contains insulin and IGF-1 at concentrations known to promote growth and development, thus pharmacological inhibitors were selected to more specifically reduce IGFR signaling. Alternative studies could alter the density of seeded cells to account for the loss in proliferation with knock-out, however, compensatory changes in other growth receptors (GH, InsR, etc.) and IGF-1 binding proteins will also need to be examined as they are known to occur with IGF-1 manipulations in vivo (Ashpole et al., 2015b).

Our comparative study is limited by not including additional cells found within the brain and blood-brain barrier. For example, pericytes not only serve as a protective layer of the blood brain barrier, they regulate blood capillary diameter within the brain by dilating and relaxing in response to glutamate (Kisler et al., 2017, 2020; Brown et al., 2019). Because of this, pericytes are important during ischemic reperfusion as they can assist with restoring flow of oxygen and glucose to the ischemic tissue (Yang et al., 2017; Khennouf et al., 2018; Sun J. et al., 2020). Little is known about the influence of IGF-1 on pericytes within the brain, and additional studies on how IGF-1 may influence pericytic glutamate response are needed.

Additionally, microglia are known to respond to excitotoxic insult, ischemia, and reperfusion. In fact, microglia have been implicated in the exacerbation of damage following glutamate excitotoxicity and ischemic stroke, due to increased production of ROS and pro-inflammatory mediators (Ma et al., 2017; Qin et al., 2019; Lian et al., 2020). At the same time, microglia also play an important protective role following tissue stroke damage, so the duality of their contribution to physiology/pathophysiology cannot be over-simplified to just a negative contribution to damage (Ma et al., 2017; Qin et al., 2019; Marino Lee et al., 2021). The age-related loss of circulating IGF-1 is associated with increased microglial activation, neuroinflammation, and susceptibility to a compromised blood-brain barrier. The relationship between IGF-1 and microglial activation is not limited to temporal correlations, as there are numerous studies that highlight IGF-1 directly regulates the structure and function of microglia. Exogenous IGF-1 alters phenotypic activation states and reduces microglial-associated production and release of tumor necrosis factor-alpha (TNF-α), interleukin-1 beta (IL-1β), inducible nitric oxide synthase (iNOS), and ROS (Park et al., 2011; Grinberg et al., 2013). Administration of IGF-1 following ischemic and/or hemorrhagic stroke attenuates microglial activation, inflammation, and ROS (Serhan et al., 2020; Sun Z. et al., 2020). Considering this, we expect that reductions in IGF-1 signaling within our cell culture system would also promote pro-inflammatory and oxidative stress phenotypes, which would further neuronal death. However, one study highlights that microglial presence in tri-cultures are neuroprotective and reduces neurons loss and astrocyte hypertrophy even though the presence of microglia increased inflammatory cytokine production compared to cell culture without microglia (Goshi et al., 2020). Future studies of ischemic and excitotoxic stressors should take microglial contributions into account while controlling for the natural migratory phenotype microglia display following activation, as results may be confounded by increased basal levels of neuroinflammatory stress (which often show an activated phenotype in culture).

Transitioning to an in vivo system would circumvent concerns regarding the lack of including all cell types present in the cerebral tissue within our co-culture models; however, the intricacies of timed, cell-specific manipulations make in vivo models difficult, particularly when trying to assess effects without developmental compensation interference. Inducible transgenic animals for each cell type are needed to rigorously assess the effects of the loss of IGF-1 in adulthood on each component of the neuro-glio-vascular unit. We previously established an inducible astrocyte-specific transgenic mouse model, and identified that glutamate handling machinery was reduced in the brains of these mice, which served as a rationale for studying glutamate excitotoxicity here (Prabhu et al., 2019). A recent study highlighted changes in neurovascular coupling when IGFR signaling was targeted in adult microvascular endothelial cells (Tarantini et al., 2021b). Thus, inducible models of astrocytic and endothelial cell IGFR knock-outs are available. Our singular and co-culture in vitro studies allow us to begin to delineate and compare the responses of specific cells without the confounds of other variables like paracrine compensation or pharmacological drug delivery or metabolism in one cell type over another. Further side-by-side comparisons of excitotoxic or ischemic damage in mice with inducible IGFR knock-out in neurons, astrocytes, endothelial cells, pericytes, and microglia would confirm the role for IGF-1 regulation of each cell, while also revealing other potential changes in physiology and pathophysiological responses in the intact brain.

Together, our study highlights that cell types within the neurovascular unit differentially respond to IGF-1 signaling. Despite differences in baseline susceptibility to glutamate, both neurons and astrocytes are both protected from excitotoxicity by IGF-1. Growth and division of endothelial cells and astrocytes are both influenced by IGF-1, however, endothelial cells remain resistant to glutamate toxicity even when IGF-1 signaling is reduced. The resistance of these supporting cells does not mean there are no consequences of reduced IGF-1 signaling in the time of glutamate stress, as both astrocytes and endothelial cells show signs of oxidative stress in the hours following exposure. The combination of neuron, astrocyte, and endothelial cells in culture afforded astrocytic protection from IGF-1 deficiency/glutamate stress but also highlighted a failure in the ability of astrocytes to protect the nearby neurons. Thus, the age-related loss of IGF-1 likely impairs the function and vitality of the entire neurovascular unit by differentially exerting stressors on neurons, endothelial cells, and astrocytes.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was reviewed and approved by Institutional Animal Use and Care Committee of the University of Mississippi.

CH and NA were responsible for conceptualization, investigation, project administration, and formal statistical analyses. CH, JM, and NA developed and designed methodology. BA, AV, and NA performed blinded image analyses and curated data. CH prepared visualizations. CH, BA, and NA prepared the original draft. All authors reviewed and edited the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.