This study demonstrates that SDs can be elicited following cerebral venous occlusion in the gyrencephalic brain. This may implicate SDs in the pathophysiological progression of cerebral venous sinus thrombosis. Our findings illustrate for the first time within a reproducible swine model that occluding the middle third of the SSS can initiate SDs. Notably, the highest incidence of SDs occurred within the first-hour post-occlusion, diminishing subsequently. IOS imaging facilitated the detection of SD onset and propagation, revealing various morphologies and their hemodynamic profiles. Additionally, this study introduces measurements of OxyHb and deOxyHb during SDs using LSCI with the novel MoorO2Flo device. Following the experimental protocol, no evidence of venous infarction was observed.
Most of the experimental studies of venous occlusion have been carried out in rodents. However, the translation of findings from lissencephalic rodents to gyrencephalic human brains is complex [22]. Our choice of a swine model is justified by its closer resemblance to the human cerebrum. Translational swine models of experimental sinus occlusion using different techniques such as endovascular flow blockage, thrombin injection, distal thrombus formation, and photothrombosis have been previously described [23–27]. They report occlusion-dependent hemodynamic changes ranging from a well-tolerated partial thrombosis to a complete thrombosis, where no collateralization is possible, and successive infarct formation [23–27]. In our model, with the exception of one animal, we did not observe an apparent hemodynamic and perfusion change in IOS imaging immediately after clipping. Neither were we able to document venous infarction development, at least during the monitoring time. This would speak in support of a well-tolerated occlusion with possible collateral out-flow development. In this regard, local and global disturbances have been reported as critical players of brain damage after cerebral venous occlusion. Development of brain edema, increase in intracranial pressure, focal hemorrhage, hypoperfusion, and hypoxia/ischemia leading to venous infarction are some of the pathophysiological consequences that can develop depending on the degree, site, and duration of the occlusion [28]. The metabolic disruption ensuing from these occlusions precipitates an excitotoxic environment [1].
Anatomical differences and changes of out-flow due to possible vein collateralization after the sinus occlusion might help to explain our present findings. Unlike humans, the venous system is configured differently in swine, which might account for the absence of venous infarction post-SSS occlusion attempts in previous studies [9]. Besides, in swine, the sagittal sinus communicates with the confluence sinus and transverse sinuses, with the system draining into the spinal epidural venous plexus. In comparison, in humans, the spinal venous plexus connects with the internal jugular vein [23, 27], and the occlusion of the first third of the SSS generally causes no clinical problems [29]. However, the occlusion of the mid-SSS, as was done in the current study, or posterior third, is not recommended because of the high risk of ischemic complications. The anatomical differences in the gyrencephalic swine brain might help to elucidate why prior efforts at experimental production of venous infarction through SSS thrombosis have generally been unsuccessful. Two successful alternatives for inducing infarctions after cerebral thrombosis in swine are bilateral occlusion or massive injection of prothrombotic material into the jugular vein [23] and the occlusion of two adjacent bridging veins to produce a focal venous infarction [7].
It is known that SDs can be caused by hypoxia and ischemia after arterial occlusion. After cerebral arterial ischemia, SDs occur at high incidences, both experimentally and clinically. They arise spontaneously from functionally and metabolically compromised tissue in the periphery of the ischemic insult, the so-called penumbra [30, 31]. It is also known that under these conditions, SDs are triggered by notable fluctuations in metabolic supply [32–34]. Previous studies on rat retina have demonstrated successful induction of SDs after photothrombosis of both arterioles and venules occlusion [35]. Nevertheless, the pathophysiology underlying the arterial insult may mask the specific consequences caused solely by the venous injury. SD induction from cerebral venous occlusions has also been documented in the lissencephalic mice brain [36]. SD development with tissue hypoxia and infarct formation has been demonstrated in the rat lissencephalic brain after the occlusion of adjacent superficial cortical veins [7, 37]. The present work, however, demonstrates that SDs can be induced in the gyrencephalic brain after venous occlusion by clipping the middle third of the SSS. Still, compared to our previous experiments, where an arterial stroke model was performed by clipping the middle cerebral arteries of the swine [9–11]. This study found a lower SD incidence and absence of infarction formation. Moreover, SD mainly developed during the first experimental hour with a decrease in incidence in the subsequent hours, whereas in our arterial stroke model, SD incidence increased gradually with time, and most SDs presented as clusters. Additionally, in the middle cerebral artery occlusion model, a mean total incidence as high as 6.4 SDs/h (STD ± 2.9) was reported [11], while after the sinus venous occlusion, we found in IOS a mean total incidence of 0.4 SDs/h (STD ± 0.7).
Physiologically, it is recognized that sinus thrombosis raises venous pressure and hinders CBF absorption, leading to cerebral edema and an increase in intracranial pressure [38–40]. Therefore, we hypothesize that after the occlusion of a major vein, such as the SSS, a reverse flow occurs along with subsequent vascular hypertension, which is sufficient to induce SDs. A possible collateral out-flow could explain the gradual decrease in SD incidence. Similarly to our previous findings in normoxic and arterial ischemic swine models, we also detected different hemodynamic responses of SDs in IOS, albeit with less heterogeneity. In comparison to our previous observations, in the present work, we identified four distinct morphologies of SDs, characterized primarily by peak hyperemia (component III) either alone or together with late hyperemia (component IV). Such differences in the vascular responses can be attributed to the properties of the underlying tissue microenvironment, which are influenced by physiological tissue states. Likewise, various vascular segments and compartments, such as arteries, capillaries, and veins, have been shown to exhibit different responses to SDs, thereby influencing the vasomotor components [41]. Hence, venous occlusion is anticipated to result in a unique hemodynamic signature with varying proportions of hemodynamic morphologies since intravascular perfusion pressure is a significant determinant of the hemodynamic response to SD. More studies are needed to understand the physiology and pathophysiology of the hemodynamic response in the gyrencephalic brain.
Notably, technical limitations due to movement artifacts constrained our ability to extract robust data from the LSCI device in all but one experiment. Despite these challenges, the congruence between IOS and LSCI data strengthens our findings. The patterns of OxyHb and deOxyHb responses during SD indicate a near-normal level of perfusion. This is supported by the fact that the magnitude of oxygen supply-demand mismatch in brain tissue caused by SD did not show strong hypoperfusion/ischemic responses. In ischemic scenarios caused by an arterial insult, the initial hypoperfusion (component II) and subsequent post-SD oligemia (component V) have been associated with marked increases in O2 extraction fraction and reductions in O2 saturation [42]. In the current SSS occlusion model, we posit that the availability of O2 was adequately compensated, as evidenced by the predominance of hyperemic reactions (components III and IV) since the promotion of hyperemic responses has been associated with vascular collateralization [43].
Clinical implication of SDs in venous occlusion
Over the past twenty years, extensive literature substantiates the occurrence of SDs within the human cerebrum, implicating their significance in cerebrovascular pathologies. Numerous studies delineate a correlation between SD episodes and neuronal functional impairment, neurodegeneration, and suboptimal clinical outcomes [44–49]. Notably, the demonstration of SDs post-venous occlusion within a translational gyrencephalic model underscores their potential clinical significance. As previously mentioned, cerebral venous sinus thrombosis is a rare type of stroke. Similarly, instances of postoperative venous occlusion can occur, and in certain cases, occlusion of the SSS may be necessary. In both scenarios, clinical manifestations differ from those typically observed in arterial strokes. Clinical presentations are heterogeneous, with a spectrum ranging from asymptomatic cases to transient neurological deficits, and in a subset, culminating in irreversible disability or mortality [50]. SDs may partially elucidate the transient neurological impairments observed in these cases. Notably, SSS occlusion frequently precipitates epileptogenic phenomena [51–53]. While epileptic discharges have the propensity to trigger SDs, evidence suggests SDs concurrently possess an inhibitory effect on seizure propagation by hindering neuronal synchrony and interrupting epileptiform cascades [54–56]. This dichotomy posits a potentially protective role for SDs in mitigating epileptiform propagation across cortical strata and stymying widespread seizure activity. Further investigations are indispensable to ascertain the precise clinical effects of SDs in the context of cerebral venous occlusive phenomena.
Study limitations
Prospective longitudinal research is imperative to unravel the cascade of events that eventuate in venous infarction after complete venous occlusion and to articulate the symptomatic nexus to SD phenomena. The study at hand, however, was of an exploratory genus, concentrating on the immediate aftermath of SSS occlusive events. Hence, we employed a clipping technique, which conferred the disadvantage of high exposure of the brain cortex and manipulation of the dura mater, which incidentally may have occluded smaller bridge veins and the drainage of the glymphatic system. Control group (shams animals) experiments were omitted, as prior research within our working group consistently indicates that the surgical procedure performed, except arterial clipping, does not notably predispose to the development of SD [11].