Introduction

Ischemic stroke, characterized by the sudden loss of blood circulation to an area of the brain due to an embolism or occlusion of an artery, stands as a leading cause of morbidity and mortality worldwide. The hyper-acute phase, particularly the first 8 hours following stroke onset, is critical for therapeutic intervention. During this period, the penumbra—the brain tissue surrounding the occlusion that is at risk yet potentially salvageable—underscores the importance of early intervention1,2.

The standard of care in stroke management involves thrombolysis, such as the administration of tissue Plasminogen Activator (tPA), and mechanical thrombectomy to restore blood flow3,4. However, these methods have limitations, including a narrow therapeutic window and potential hemorrhagic risks. In this context, hyper-acute cerebral flow augmentation strategies, including NEH (a combination of norepinephrine and hydralazine) and Sanguinate (pegylated bovine carboxyhemoglobin), have emerged as promising adjunct or “bridge” therapies during this critical phase5,6.

Our study hypothesizes that these flow augmentation therapies, when administered in conjunction with tPA or thrombectomy, could enhance cerebral blood flow to ischemic regions, thereby preserving brain tissue and function. NEH acts as a potent vasopressor and vasodilator, increasing systemic blood pressure and dilating cerebral arterioles to enhance collateral flow to ischemic regions. Sanguinate, an oxygen carrier, facilitates oxygen delivery to oxygen-deprived cells and tissues and possesses vasodilatory and anti-inflammatory properties that enhance collateral recruitment into ischemic tissue7,8.

In assessing the effectiveness of these flow augmentation therapies, key neuroimaging metrics such as functional connectivity, global mean signal (GMS), and blood oxygen level-dependent (BOLD) time delay are crucial9,10,11. Functional connectivity indicates the temporal correlation between spatially remote neurophysiological events, reflecting functional integration between separate brain regions12. GMS provides insights into the global state of brain function, and BOLD time delay measures the lag between neural activity and the subsequent hemodynamic response, indicating the efficiency and health of neural processing13.

This study, utilizing a pre-clinical canine model, investigates how hyper-acute cerebral flow augmentation influences these neuroimaging parameters in the context of ischemic stroke. We hypothesize that NEH and Sanguinate can preserve functional connectivity, maintain GMS, and minimize BOLD time delay during the hyper-acute phase, thereby offering neuroprotective benefits. Our research aims to contribute to the stroke treatment field by potentially integrating these novel interventions into standard stroke treatment protocols. Overall, this study sought to establish baseline findings for a permanent MCAO in a large animal. Future work will seek to establish the benefit of flow augmentation therapy in a setting of reversible infarct as is anticipated with successful recanalization.

Methods

Animal model and induction of ischemic stroke

The study utilized a pre-clinical canine model, chosen for its neuroanatomical similarities to humans, ensuring translational relevance to human stroke pathology. Ischemic stroke was induced using a permanent Middle Cerebral Artery (MCA) occlusion technique, involving the deployment of occlusion coils at the M1 segment and carotid terminus14,15. This occlusion technique was carefully randomized between the left and right MCAs on a case-by-case basis to minimize potential bias and ensure consistent and reproducible ischemic regions. Specifically, for subjects with right hemisphere lesions, we mirrored the imaging data across the mid-sagittal plane, effectively flipping the lesion to the left hemisphere to facilitate group-level analyses and comparisons. The detailed procedures for inducing ischemic stroke were meticulously followed to create a reliable model aligning with the targeted brain region, as critical for the study’s objectives.

Experimental animals and grouping

The study involved a total of 40 mongrel canines (25–30 kg, 27 females, 13 males, 4.22 ± 3.3 years old). These animals were randomly allocated into three groups: Control/Natural History group (n = 12), NEH treatment group (n = 14), and Sanguinate treatment group (n = 14).

The inclusion criteria required animals to be in good general health and without neurological deficits. Exclusion criteria included any pre-existing conditions or abnormalities that could potentially affect cerebrovascular function. Randomization of animals to treatment groups and the randomization of lesion laterality (left or right MCA occlusion) were performed using a computer-generated sequence. All imaging data were analyzed in a blinded fashion, with the analysts unaware of the group allocations during the analysis process.

Administration of therapeutic agents

The therapeutic agents, NEH (a combination of norepinephrine and hydralazine) and Sanguinate (pegylated bovine carboxyhemoglobin), were administered during the hyper-acute phase within 30 mins following the induction of MCA occlusion. NEH was administered intravenously via a peripherally inserted catheter, with norepinephrine dosed at 0.1–1.52 µg/kg/min, adjusted to maintain mean arterial pressure (MAP) 25–45 mmHg above baseline, and a 20 mg bolus of hydralazine. Control canines were maintained at a MAP of 80–105 mmHg. Sanguinate subjects received an 8 mL/kg intravenous bolus via a transfemoral venous line with tip in the inferior vena cava16,17.

MRI protocol and data acquisition

The study leveraged resting-state functional MRI (rs-fMRI) and Diffusion Tensor Imaging (DTI) to assess both functional and structural aspects of brain changes before and after the induction of stroke and the administration of therapeutic agents. Rs-fMRI was executed on a 3.0 T system, utilizing a BOLD-sensitive Echo Planar Imaging sequence with a repetition time (TR) of 1400 ms and an echo time (TE) of 20 ms. The voxel size was set at 2.5 mm isotropic, capturing 300 temporal positions. Throughout the imaging, canines were under approximately 1% isoflurane anesthesia, with their physiological parameters continuously monitored by veterinary professionals to ensure stability18,19.

DTI sessions, aimed at examining brain structural integrity, involved the generation of mean diffusivity (MD) maps and facilitated the creation of accurate lesion maps, integral for infarct volume assessment. Lesion volumes were measured using diffusion-weighted imaging (DWI) at 24 h post-occlusion. The mean lesion volume for the control group was 4739 ± 4400 mm3, while for the NEH group, it was 2798 ± 2340 mm3, and for the Sanguinate group, it was 2585 ± 2680 mm3. For a detailed analysis of lesion volumes, please refer to our previous publications10,21. The imaging protocol employed a Spin Echo-Echo Planar Imaging sequence with a TR of 2993 ms, a TE of 83 ms, a slice thickness of 2 mm, and incorporated 32 diffusion directions to ensure comprehensive anisotropy characterization20,21.

Resting-state functional connectivity (rs-FC) analysis

A critical aspect of the methodology was the investigation of resting-state functional connectivity (rs-FC) networks within the brain, serving as a window into the impact of stroke and therapeutic interventions on neural interactions. The study employed a group Independent Component Analysis (ICA) approach using FSL’s MELODIC (FMRIB’s Software Library, version 6.0) to identify 50 viable components22. The rs-FC analysis aimed to visually locate specific networks of interest, namely the sensorimotor and visual networks. These networks were identified based on literature precedence3,4,19, ensuring a robust basis for network selection. The selection of seed regions for the rs-FC analysis was meticulous, considering the ipsilateral and contralateral hemispheres. The use of DTI MD maps ensured that ipsilateral regions of interest (ROIs) were fully within the lesioned region, while contralateral ROIs were fully outside the lesion, minimizing potential contamination of results23.

BOLD time delay analysis

The calculation of voxel-wise BOLD time delay was a crucial component of FC analysis. This calculation involved several intricate steps, elucidating the temporal dynamics of neural activity and BOLD responses at a voxel level. Beginning with the selection of seed regions within the brain and the extraction of BOLD time series data from these regions, the analysis encompassed cross-correlation techniques to quantify temporal relationships between seed regions and voxels across the entire brain. The phase differences between the time series of seed regions and each voxel were determined, representing the temporal offset or time lag between neural activity and BOLD responses at the voxel level. Statistical tests, such as paired T-tests and Mann-Whitney U tests, were applied to assess group differences in voxel-wise time lag, while false discovery rate (FDR) correction was employed to control for multiple comparisons. Data processing and analysis were performed using custom scripts in Matlab (R2020a, MathWorks) and Python (version 3.8)24.

Statistical analysis

Statistical analysis played a pivotal role in validating the observed connectivity changes, assessing the significance of time-lag variations, and determining the effects of therapeutic interventions. The choice of statistical methods was guided by the nature of the data and the study’s objectives. For connectivity changes, paired T-tests and Mann–Whitney U tests were applied to quantitatively validate the results, providing a robust statistical framework. Post-hoc pairwise comparisons were performed using Tukey’s honest significant difference (HSD) test following the initial ANOVA analysis to assess differences between the NEH, Sanguinate, and control groups. To address the challenge of multiple comparisons, FDR correction was utilized. Matlab (R2024a, MathWorks) and Python (version 3.12) were used to conduct these statistical analyses25,26.

Ethical considerations

The study adhered rigorously to ethical standards, particularly in the care and use of animal subjects, in accordance with the University of Chicago’s IACUC (Institutional Animal Care and Use Committee) protocol. The ethical considerations extended to anesthesia protocols, continuous monitoring of physiological parameters, and the compassionate and humane treatment of animals throughout the study. The utmost care was taken to minimize any potential distress or discomfort to the animal subjects, aligning with IACUC standards for animal research.

Use of human subjects and animals

Experiments for the present study were approved by the University of Chicago Institutional Animal Care and Use Committee and reported in compliance with ARRIVE guidelines. The University of Chicago is an AAALAC International accredited institution adhering to the following guidelines, regulations and policies: (a) Guide for the Care and Use of Laboratory Animals (National Research Council), (b) USDA Animal Welfare Act and Animal Welfare Regulations, and (c) Public Health Service Policy on Humane Care and Use of Laboratory Animals.

Results

Group independent component analysis (ICA)

Figure 1 showcases the resting-state networks identified through group ICA conducted on all subjects across all groups (Control, NEH, Sanguinate) both pre- and post-occlusion. These networks include the Default Mode Network (DMN), Visual Network, Sensorimotor Network, and Somatosensory Network. The z-score maps range from 3 to 10, and all maps are overlaid on a T1-weighted template of a canine atlas27.

Separate comparisons for each individual group (Control, NEH-treated, and Sanguinate-treated) were performed to assess changes in the sensorimotor and visual networks pre- and post-occlusion. The dual regression results averaged by group and condition (pre- and post-occlusion) are shown for the Visual Network and Sensorimotor Network. These visualizations allow for the assessment of network function changes specific to each group before and after occlusion28.

Figure 1
figure 1

Identification and Comparison of Resting-State Networks Pre- and Post-Occlusion Across All Groups and Individual Groups. Top Panel: Group ICA results from all subjects across all groups (Control, NEH, Sanguinate) both pre- and post-occlusion. Identified networks include the Default Mode Network (DMN), Visual Network, Sensorimotor Network, and Somatosensory Network. The z-score maps range from 3 to 10, and all maps are overlaid on a T1-weighted template of a canine atlas. Bottom Panel: Dual regression results averaged by group (Control, NEH, Sanguinate) and by condition (pre- and post-occlusion). For the Visual Network and Sensorimotor Network, the z-score maps (ranging from 3 to 10) are shown for each group pre- and post-occlusion. The z-score maps are overlaid on a T1-weighted template of a canine atlas. These visualizations allow for the assessment of network function changes specific to each group before and after occlusion.

Full size image

Voxel-wise functional connectivity analysis

Connectivity patterns across hemispheres Voxel-wise FC analysis delineated significant post-occlusion connectivity alterations within the ipsilesional and contralesional hemispheres, as visualized in Fig. 2. Pre-occlusion, the control group showcased a median FC value of approximately 0.5, with a relatively narrow interquartile range (IQR) indicating homogeneous connectivity. Post-occlusion, a notable increase in the IQR was observed, suggestive of disrupted connectivity patterns, with median FC values decreasing to 0.4529.

Figure 2
figure 2

Ipsilesional (left) and Contralesional (right) Boxplots of Mean FC Values for the left roi and right roi respectively across different experimental groups. The groups include control pre, control post, NEH pre, NEH post, Sanguinate pre, and Sanguinate post. The boxplots illustrate the distribution of mean FC values within each group, with the box representing the interquartile range (IQR), the line inside the box indicating the median, and the whiskers extending to 1.5 times the IQR. Outliers are shown as individual points beyond the whiskers. Asterisks (*) indicate statistically significant differences (P< 0.05) between groups based on Tukey’s honest significant difference (HSD) test, highlighting the impact of NEH and Sanguinate treatments on functional connectivity. The control pre group serves as the baseline for comparison. The observed trends demonstrate that NEH treatment led to higher functional connectivity values compared to the control group post-occlusion, whereas Sanguinate treatment resulted in lower connectivity values relative to the control post-occlusion in the ipsilesional hemisphere. In the contralesional hemisphere, the NEH and Sanguinate treatments maintained functional connectivity levels more consistently compared to the control group.

Full size image

Conversely, post-treatment analysis in the NEH group revealed a median FC value in the ipsilesional hemisphere that remained close to the pre-treatment value, shifting marginally from 0.5 to 0.48, and an IQR that suggests a stabilization of connectivity (IQR pre-treatment: 0.1, IQR post-treatment: 0.12). The Sanguinate group exhibited a slight reduction in median FC value from 0.5 pre-treatment to 0.42 post-treatment, with an expanded post-treatment IQR of 0.18 compared to a pre-treatment IQR of 0.1, indicative of a moderate preservation of connectivity30.

In the contralesional hemisphere, similar trends were observed. The control group’s median FC value decreased from 0.5 to 0.43, with an increased post-treatment IQR, reflecting a general decline in connectivity. The NEH and Sanguinate groups demonstrated a protective trend, albeit with varied efficacy. NEH post-treatment FC values remained comparatively steady (median FC value from 0.5 to 0.47), whereas the Sanguinate group showed a median FC decrease to 0.431.

These quantitative findings underscore the potential of NEH and Sanguinate treatments in maintaining neural network integrity under ischemic stress. Quantitative Validation of Connectivity Alterations: The paired T-tests and Mann–Whitney U tests provided a rigorous statistical assessment of the observed connectivity changes. In the ipsilesional hemisphere, the control group’s pre-occlusion versus post-occlusion comparison yielded a T-statistic of – 6.45 with a p-value of 1.16E–10, which, after FDR correction, altered to 1.93E–10, indicating a significant disruption in connectivity post-occlusion. In contrast, the NEH group showed a T-statistic of 3.54 and a p-value of 4.06E–4 (5.08E–4 post-FDR correction), suggesting a significant improvement in connectivity. The Sanguinate group displayed similar trends, with a T-statistic of – 1.66 and a non-significant p-value of 0.0974, both pre- and post-FDR correction32.

In comparisons involving post-occlusion groups, the control versus NEH comparison revealed a Mann–Whitney U statistic of 6.10E8 and a highly significant p-value of 1.58E–23 (3.96E–23 post-FDR correction), emphasizing the stark differences between the groups. Similarly, the control versus Sanguinate comparison showed a U statistic of 6.28E8 and a p-value of 1.02E–62 (5.11E–62 post-FDR), further highlighting the substantial impact of the treatments. These statistical findings reinforce the differential impacts of ischemic stroke on functional connectivity and the significant modulatory effects of the NEH and Sanguinate bridge therapies33.

GMS and BOLD time delay analysis

GMS analysis The analysis of global mean signal (GMS) revealed notable variations between groups and conditions (Fig. 3). For the control group, a marked decrease in GMS post-occlusion was noted, with a significant shift in mean signal intensity. Contrastingly, in the NEH and Sanguinate treatment groups, the GMS remained relatively stable post-occlusion, indicating the effectiveness of these treatments in mitigating the impact of stroke on overall brain activity. Specific statistics showed a reduction in GMS in the control group post-occlusion by an average of 15%, compared to a mere 3% and 4% reduction in the NEH and Sanguinate groups, respectively34.

Figure 3
figure 3

Differences in average global mean signals (GMS) over 300 time points for control, NEH, and Sanguinate groups. The plot illustrates the changes in signal intensity across the entire duration of the experiment. The control group is depicted in blue, the NEH group in green, and the Sanguinate group in orange. Each line represents the mean signal intensity difference (pre-treatment minus post-treatment) at each time point, providing a continuous view of the temporal dynamics. The NEH group shows a distinct pattern compared to the control and Sanguinate groups, reflecting the potential impact of NEH treatment on global mean signal stability.

Full size image

BOLD time delay variations post-stroke Analysis of BOLD time delay post-occlusion, as depicted in Fig. 4, revealed significant regional variations in response timing across different brain areas. The unity plots (which serve as a visual representation of BOLD time-lag shifts due to stroke and subsequent treatment effects) in this figure show a notable increase in time delay for the control group, particularly in the sensorimotor and visual network regions, where delays extended up to 200–300 ms in some regions. In contrast, the treatment groups exhibited more consistent and stable time delay patterns, with average increases of only 50–100 ms. This suggests the efficacy of NEH and Sanguinate in maintaining a more normalized hemodynamic response post-occlusion35.

Figure 4
figure 4

Unity plots for BOLD time lag pre- and post-treatment. Each plot illustrates the relationship between pre-treatment and post-treatment BOLD time lag values for individual voxels within the lesion region of the brain. The control group (left plot), NEH group (middle plot), and Sanguinate group (right plot) are shown. Each point represents a voxel’s time lag, with the pre-treatment time lag on the x-axis and the post-treatment time lag on the y-axis. The red dashed line is the unity line (y = x), indicating no change in time lag from pre- to post-treatment. Points above the unity line suggest an increase in time lag post-treatment, while points below indicate a decrease. The distribution and spread of the points around the unity line provide insights into the consistency and variability of the treatment effects: Control Group: Points are scattered around the unity line, indicating significant changes in BOLD time lag pre- and post-occlusion without treatment. This reflects the impact of the stroke on BOLD time lag dynamics within the lesion region. NEH Group: Points show a more clustered pattern around the unity line compared to the control group, indicating that NEH treatment stabilizes BOLD time lag dynamics, reducing the variability caused by the occlusion within the lesion region. Sanguinate Group: The distribution of points shows a more dispersed pattern with an upward shift, indicating notable changes in BOLD time lag post-treatment. This reflects the effect of Sanguinate on modifying the BOLD response, generally increasing the time lag within the lesion region. The unity plots demonstrate that both NEH and Sanguinate treatments result in significant changes in BOLD time lag, indicating their potential to alter neurovascular dynamics in the brain following ischemic stroke.

Full size image

In the control group unity plot, data points are scattered widely from the line of unity (red dashed line), indicating a significant deviation in BOLD response time from pre-stroke to post-stroke, with some regions experiencing up to 100 ms delay post-stroke compared to pre-stroke. This suggests a broad disruption in neural timing following the stroke event36.

In contrast, the NEH treatment group shows a tighter clustering of points around the line of unity, with fewer outliers and a less pronounced spread, implying a more stable BOLD response post-treatment and a mitigated shift in time-lag post-stroke. This suggests that NEH treatment may help to preserve the timing of the BOLD signal post-stroke, reflecting a protective effect on neural timing consistency. Similarly, the Sanguinate group unity plot also demonstrates a tighter clustering around the line of unity but with a slight shift toward increased positive time-lag values. This shift, while present, is much less than in the control group, suggesting that Sanguinate treatment also contributes to maintaining more consistent BOLD signal timing following stroke. Both treatment groups’ unity plots suggest that NEH and Sanguinate therapies contribute to neuroprotection, reflected in the preservation of BOLD signal timing post-stroke37.

Discussion

The therapeutic potential of hyper-acute cerebral flow augmentation strategies, such as NEH and Sanguinate, represents a promising avenue in ischemic stroke intervention. Our findings highlight the ability of these therapies to maintain functional connectivity, stabilize global mean signal (GMS), and preserve the timing of the BOLD signal post-stroke. These effects suggest neuroprotective benefits that could be crucial during the hyper-acute phase of ischemic stroke.

The group ICA revealed notable resilience in the sensorimotor and visual networks of the treatment groups compared to the control group. This resilience underscores the neuroprotective potential of NEH and Sanguinate, as these networks are critical for functional recovery post-stroke. The significant disruptions observed in the control group highlight the acute impact of ischemic events on these critical areas. Previous studies have similarly reported disruptions in these networks following stroke, emphasizing the importance of preserving these connections for better functional outcomes27,28. Our findings extend this knowledge by demonstrating that NEH and Sanguinate treatments can mitigate these disruptions, highlighting the potential of these therapies in preserving essential neural networks during the hyper-acute phase.

Voxel-wise functional connectivity analysis further demonstrated the efficacy of these treatments. The NEH and Sanguinate groups displayed only modest alterations in connectivity patterns, unlike the pronounced disruptions seen in the control group. This suggests that NEH and Sanguinate may facilitate the maintenance of neural communication pathways, potentially supporting better clinical outcomes. The improvement in connectivity in the NEH group, as evidenced by significant statistical findings, supports the hypothesis that these therapies can restore neural network integrity29. This aligns with studies showing that maintaining connectivity in critical brain networks correlates with better functional recovery and reduced neurological deficits post-stroke14,33,34.

The stability of the GMS in the NEH and Sanguinate groups post-occlusion, contrasted with the significant decrease in the control group, points to sustained overall brain function facilitated by these treatments. GMS reflects the overall activity of the brain, and its stability suggests that NEH and Sanguinate help maintain neural activity at levels conducive to recovery. This finding is consistent with the role of GMS as an indicator of brain health and functional integrity30. Comparatively, previous studies have reported that significant decreases in GMS post-stroke are associated with worse outcomes, reinforcing the importance of maintaining GMS for recovery35.

BOLD time-lag unity plots provided additional insights into the impact of NEH and Sanguinate. The tighter clustering around the unity line in the treatment groups, compared to the control group, suggests a more stable hemodynamic response. This stability is crucial for effective neural processing and recovery post-stroke. Disruptions in BOLD signal timing, as observed in the control group, can indicate impaired neurovascular coupling, which is detrimental to brain function31. By preserving this timing, NEH and Sanguinate appear to support neurovascular health, enhancing the brain’s ability to recover from ischemic insult32. These findings are in agreement with studies suggesting that stable neurovascular coupling is critical for the recovery of brain function after stroke14,36,37.

Our study aligns with existing literature on the neuroprotective effects of early intervention in stroke. Maintaining functional connectivity and stable hemodynamic responses can significantly improve recovery outcomes33. The observed preservation of sensorimotor and visual networks in our study is particularly relevant as these networks are critical for post-stroke rehabilitation and functional recovery34,35. Moreover, our findings on BOLD signal timing are supported by research indicating that disruptions in neurovascular coupling are associated with poor stroke outcomes36. The ability of NEH and Sanguinate to stabilize BOLD signal timing suggests that these therapies can support neurovascular health and enhance the efficiency of neural processing during recovery37.

However, it is important to consider the potential impact of reverse electron transport (RET) and associated oxidative stress during ischemia-reperfusion injury. RET can exacerbate oxidative damage, potentially undermining the benefits of increased cerebral blood flow2,40. Future studies should investigate the extent to which NEH and Sanguinate influence RET and oxidative stress, and explore strategies to mitigate these effects, such as combining these therapies with antioxidants or other neuroprotective agents2,39,40 .

In conclusion, our research provides compelling evidence supporting the use of NEH and Sanguinate in the acute phase of ischemic stroke. These therapies’ ability to preserve functional connectivity, stabilize GMS, and maintain BOLD signal timing paves the way for their inclusion in stroke treatment protocols. Future studies should replicate these findings in larger cohorts, explore the underlying mechanisms, and assess long-term behavioral outcomes alongside functional connectivity effects, particularly when recombinant tPA and/or thrombectomy are employed38.