Taboos and opportunities in sonothrombolysis for stroke

Abstract

Systemic thrombolysis with tissue plasminogen activator (tPA) is the only approved treatment for acute ischaemic stroke that improves functional outcome if given up to 4.5 h from symptom onset. At least half of treated patients have unfavourable outcomes long-term though, emphasising the need to amplify the only approved acute stroke therapy. Ultrasound targeting of an intra-arterial occlusive clot and delivering mechanical pressure to its surrounding fluids (referred to as sonothrombolysis) accelerates the thrombolytic effect of tPA. Higher recanalisation rates produce a trend towards better functional outcomes that could be safely achieved with the combination of 2 MHz frequency ultrasound and systemic tPA. To further accelerate the clot-dissolving effect of ultrasound, a variety of frequencies and intensities as well as other adjuvant treatment elements are being studied. However, literature reports argue efficacy and safety of these novel approaches doubting promptly translation into the clinical practice. This review will summarise our current knowledge about potentially harmful (taboos) directions and what we think are promising avenues for these future stroke therapies. We also give a prospect for novel technologies such as operator-independent devices that aim to further spread the use of sonothrombolysis for stroke.

• Energy
• sonothrombolysis
• stroke

Introduction: Energy and stroke

Our body receives and produces energy, storing, transmitting and dissipating it constantly. Arterial blood flow maintains cerebral metabolism by delivering nutrients through capillaries while venous outflow removes by-products. A thrombo-embolic occlusion disrupts this balance leaving oxygen-dependent neurons at the mercy of the residual flow and recruitment of collaterals. As time passes, damage becomes complete and the infarcted area grows. In addition to pharmacological clot dissolving treatment, energy transmission to the clot-residual flow interface and to the ischaemic areas can be beneficial [1]. Ultrasound can promote fibrinolysis and reduce brain oedema through a variety of mechanisms [2–7] while lasers are now one step away from becoming neuro-regenerative treatment presumably due to recharging mitochondria and other yet-to-be confirmed effects of energy absorption in the brain [8–10]. Targeting tissues in need of energy, delivering the right amount, and monitoring the response are exciting new areas where sonothrombolysis for stroke is making constant progress despite a few setbacks. This review will cover what we learned to be harmful (taboos) and what we think are promising avenues for future stroke therapies.

Primum non nocere: Taboos in sonothrombolysis

Temperature increase

Ischaemic brain is sensitive to temperature: hyperthermia could be harmful [11] while lowering temperature is cytoprotective (hence hypothermia is now tested in a phase II–III sequential trial for acute ischaemic stroke as an adjunctive treatment following the infusion of tissue plasminogen activator (tPA)) [12], [13]. Stroke clinicians’ need for fever control in acute stroke management creates perceived limitations or concerns if any temperature rise in the ischaemic brain can be caused by exposure to ultrasound. Thus, in the setting of sonothrombolysis for acute ischaemic stroke, ultrasound-associated thermal effects are important since temperature is linked to tissue metabolism and ultimately the fate of ischaemic brain tissues. Due to the need to hold the ultrasound beam stationary and the high acoustic absorption coefficient of the transtemporal bone, transcranial pulsed spectral Doppler ultrasound may locally induce thermal-mediated bio-effects [14]. Bone attenuates a significant amount of energy that penetrates into deep brain tissue and the absorbed acoustic energy is rapidly transformed into heat in the bone and the nearby soft tissues. Despite clear success of enzymatic activity enhancement with 1–2 MHz [5–7] in experiments, an observation of local heating produced by these frequencies led pioneers such as Charles Francis, Robert Siegel and Hiroshi Furuhata to choose kHz ranges for further ultrasound-enhanced thrombolysis experiments [15–17]. As a result, two low frequency systems were built for clinical testing, and others are in development.

Kilohertz (kHz) frequency range

The first system, the Walnut device (Walnut Technologies, Boston, MA), gave us several important pieces of information. With its initial design, patients did not tolerate frequencies starting at 37 kHz and engineers had to increase the frequency up to 300 kHz at which human subjects stopped experiencing tinnitus. The 300 kHz system tested in the multicentre, randomised, controlled Transcranial Low Frequency Ultrasound-Mediated Thrombolysis in Brain Ischemia (TRUMBI) trial produced excessive bleeding complications (such as symptomatic intracerebral haemorrhage (sICH) and subarachnoid haemorrhage) when co-administered with intravenous tPA for acute ischaemic stroke leading to the discontinuation of the trial [18]. Given the lower sICH rates with diagnostic ultrasound frequencies, these facts strongly suggest differential bioeffects of low and high frequency ranges in sonothrombolysis (Figure 1). The recent findings of Reinhard et al. showing that abnormal permeability of the human blood–brain barrier can be induced by wide-field low frequency insonation are in line with the hypothesis that the observed excessive bleeding rate with low frequency sonothrombolysis might be attributable to primary blood–brain barrier disruption by ultrasound [19]. The second system intermittently exposes patients to kHz and MHz frequencies [20] and is further discussed in the opportunities for sonothrombolysis section below.

Figure 1. Attenuation per unit length plotted against ultrasound carrier frequency in human temporal bone as given in recent publications (blue diamonds and line) [36], [72], [73]. Symptomatic intracerebral haemorrhage rates in sonothrombolysis trials (controlled and uncontrolled) plotted against ultrasound carrier frequency (red dots) [1]. Grey-filled squares illustrate acoustic cavitation thresholds from an experimental measurement in different blood sample combinations at 120 kHz pulsed ultrasound [27]. It should be mentioned that acoustic cavitation thresholds vary depending on media, emitting frequencies and acoustic parameters (e.g. with lower frequency and continuous wave ultrasound, the threshold for any cavitation in blood or any other biological media will decrease). Higher attenuation through the skull with high frequency ultrasound may lead to lower acoustic pressures inside the targeted region (e.g. predicted acoustic pressure for the CLOTBUST (combined lysis of thrombus in brain ischemia using transcranial ultrasound and systemic TPA), trial: 0.07 MPa) [43], whereas lower absorption of the skull with low frequency ultrasound may yield acoustic pressures well above the inertial cavitation threshold putting the brain at higher risk of bleeding (e.g. predicted acoustic pressure for the TRUMBI trial: 0.27–1.2 MPa) [57].

It should be mentioned at this point that lack of standardisation regarding the reporting of acoustic properties applied in sonothrombolysis studies make direct comparisons between different studies difficult, if not impossible. Ultrasound beams with lower frequencies should be better characterised, and pulse lengths shorter than the ones in TRUMBI may still be re-evaluated in tolerability studies prior to clinical trials. Thus, one of the major taboos in experimental and clinical sonothrombolysis studies is that researchers still do not characterise ultrasound that is used for sonothrombolysis (be it low or high frequency ultrasound). We summarise certain ultrasound parameters proposed by Mark Schafer that should be determined and reported by researchers conducting sonothrombolysis studies in Table 1.

Table I.  Proposed guidelines for reporting certain ultrasound parameters in experimental and clinical sonothrombolysis studies (if applicable) (adapted from Mark Schafer [74]).

Ultrasound parameter	Unit
Peak rarefaction pressure	MPa
Frequency, and any modulation	Hz (MHz or kHz)
ISPTA intensity (for comparison to FDA limits)	W/cm^2
Total power	W
Pulse duration or CW	s (and number of cycles)
Pulse repetition frequency	Hz
Beam width −6/−12 dB	mm
Aperture size and shape	mm
Focal distance (or unfocused)	mm
Distance from transducer to region of interest	mm
Mechanical/thermal index	number

Continuous wave ultrasound

Several experiments have shown that continuous wave (CW) produces more heating of tissues and does not enhance lysis as efficiently as pulsed wave (PW) ultrasound transmission [21], [22]. In addition, CW ultrasound caused potentially harmful effects including brain oedema and subarachnoid haemorrhages in rat animal models of cerebral ischaemia [23], [24]. Overall, CW configurations are not deployed in current clinical studies of sonothrombolysis.

Standing waves

Higher pulse repetition frequencies, longer duty cycles and longer spatial pulse lengths can lead to formation of standing waves. The latter were implicated as the mechanism by which the Walnut device caused haemorrhages including remote ones in the non-affected side of the brain [18], [19].

Inertial cavitation

Acoustic stable cavitation is generally acknowledged as playing a major role in efficiency of sonothrombolysis, whereas inertial cavitation induced by higher intensity levels may enhance permeability of the blood–brain barrier putting the brain at higher risk for haemorrhage in the presence of tPA [25], [26]. Too much energy can increase local acoustic pressures and cause collapse of cavitational bubbles [27–29], a potentially harmful bio-effect of ultrasound that should be avoided [30]. Brain vessels are paper-thin and lack the protective muscular layers of the coronary arteries. Therefore, brain vessels become vulnerable to both excessive stretching and any cavitational processes that can disrupt integrity of the vessel wall. Finally, brain parenchyma has very low resistance and once bleeding starts, it usually continues for some time, often leading to neurological deterioration and further haematoma expansion even if the offending factor is no longer in play.

However, besides cavitation-related sonothrombolysis effects, there may be other ultrasound-related (e.g. non-cavitational) effects perhaps even within lower acoustic pressure ranges such as pressure-driven permeation of plasma through the thrombus (equivalent to mild stirring), or improved tissue perfusion likely through a nitric oxide-dependent vasodilation [31–34], both with the ability to influence drug transport through and around the thrombus. In a rat experiment with permanent occlusion, exposure to 2 MHz PW ultrasound reduced brain infarct size and oedema without thrombolytic drug [2].

Intensity (FDA limit)

The imposed upper threshold of ≤0.72 W/cm² for human exposure to diagnostic ultrasound was derived from the analysis of outputs of devices in clinical use that had no reported harmful bio-effects [35]. It refers to the derated spatial peak temporal average intensity (I_SPTA), and does not regulate other parameters of intensity, pulse repetition or beam configuration. Higher I_SPTA intensities were tested in experimental models only [32] and have not been applied in the clinical setting of sonothrombolysis. Of note, the presence of the temporal bone greatly attenuates ultrasound energy [36] perhaps leaving some room for more generous intensities to be tested for therapeutic as opposed to diagnostic ultrasound exposures (Figure 1).

Opportunities for sonothrombolysis

The risk–benefit ratio

Over half of stroke patients receiving the only approved therapy (intravenous tPA) remain permanently disabled or die despite treatment [37]. This is linked to a variety of reasons, with the most important ones being:

1. tPA dose is not based on the weight of the clot [38];

2. tPA is less effective with more proximal or larger/longer thrombi [39];

3. tPA induces slow and partial recanalisation and its efficacy to reverse brain damage decreases over time [40].

Therefore, there is room for improvement, and the significant likelihood for a patient to remain disabled justifies the risks of taking further steps to improve tPA-induced recanalisation as the likely mechanism leading to neurological and functional recovery after stroke.

Another risk associated with tPA is bleeding (sICH), which occurs in 6.4% of treated stroke patients [41]. In a prior study, the risk of haemorrhage was not increased with fast and complete recanalisations, while persisting arterial occlusion after 2 h since tPA bolus emerged as an independent predictor of sICH [42]. Therefore, a very early amplification of enzymatic tPA activity and mechanical separation of loosened fibrin strands with ultrasound is desirable, as it has not increase the risk of sICH with 2 MHz PW beams [43] and perhaps may even lower this risk in the future.

Innovative directions

A variety of novel approaches and technologies are being tested in experiments including various ultrasound frequencies and beam optimisations, microspheres, liposomes, etcetera [20], [44–55]. The targets of sonothrombolysis evolve from just proximal occlusions causing acute ischaemic stroke to neuroprotection [2] and melting blood clots in patients with haemorrhagic strokes [56].

Targeted liposomal drug delivery

Targeted liposomal drug delivery activated by ultrasound may increase local efficacy of a thrombolytic treatment while at the same time systemic exposure is reduced. Several in vitro studies demonstrated that the combination of ultrasound and echogenic liposomes incorporated with tPA significantly enhanced thrombolytic treatment [44], [45]. Since acoustic cavitational effects are the proposed mechanisms to induce drug release from tPA-loaded echogenic liposomes [54], [55], it needs to be determined whether this approach can be sufficiently used in the clinical setting with high frequency ultrasound, where attenuation through the human skull yields ultrasound energy levels way below acoustic cavitation thresholds [36], [57].

Alternative thrombolytic agents

Other thrombolytic agents such as glycoproteine (GP) IIb/IIIa inhibitors may be an alternative to the FDA's only approved therapy for acute ischaemic stroke, and ultrasound combination with such agents should be explored. Although the combination of 120 kHz ultrasound, tPA and eptifibatide did not show greater thrombolytic efficacy in an in vitro human clot model than ultrasound and tPA alone [47], other investigators showed promising results using GP IIb/IIIa inhibitors such as abciximab in an in vitro experiment [48]. However, since the clinical trial of GP IIb/IIIa inhibitors in acute ischaemic stroke patients showed safety concerns [58], their future role as an adjuvant to ultrasound remains uncertain in the clinical setting.

High intensity focused ultrasound (HIFU)

High intensity focused ultrasound (HIFU) is widely used in treatment of cancer; however, research regarding its use for thrombolytic treatment is lacking [49]. Frenkel et al. [50] insonated human clots (1 MHz; power 60 W) with and without additional tPA, and observed a significant increase in the degree of thrombolysis in the combined group compared to tPA alone. Meanwhile, these results have been verified in an in vivo experiment applying lower power settings (40 W) [51]. Further experimental work, particularly in terms of safety (since HIFU is associated with a substantial increase in local temperature) and perhaps with cadaver skulls [59], needs to be conducted before clinical trials in humans can be initiated.

Gaseous microspheres

Another promising adjuvant to potentiate sonothrombolysis could be gaseous microspheres (µS) with protective shells that were first engineered as contrast agents for ultrasound imaging [54], [55]. After gaseous µS are compressed by an ultrasonic pressure wave, the gas expands, and the spheres oscillate. Because µS have impedance much higher than red blood cells [60], they act like bright reflectors and send back stronger echoes useful for imaging. With expansion, they also transmit mechanical energy momentum to surrounding fluids, accelerating residual flow and facilitating even faster and more complete recanalisation [61]. The addition of gaseous μS to transcranial Doppler monitoring can also augment end-diastolic flow velocity very early into treatment, thus likely improving reperfusion of the ischaemic tissues since end-diastolic flow velocity is an indicator of distal resistance. Recently, clinical pilot studies of different types of gaseous μS activated by transcranial 2 MHz PW Doppler ultrasound showed promising results in acute ischaemic stroke patients treated with intravenous tPA [52], [53].

Thrombolysis independent mechanisms

Multiple animal studies have demonstrated that ultrasound has a significant vasodilatory effect in both coronary and peripheral arteries, suggesting involvement of vasoactive substances to be released by ultrasound [16], [34], [62–64]. Suchkova and colleagues demonstrated that in rabbits with femoral artery occlusions receiving no thrombolytic drugs, low intensity 40 kHz ultrasound improves tissue perfusion in ischaemic limbs, likely through a nitric oxide (NO)-dependent vasodilation [64]. Interestingly, the most important physiological stimulus in NO synthesis are mechanical forces (shear stress) generated by the streaming blood on the endothelial cell mechanoreceptors [65]. Moreover, a recent study suggested that ultrasound may also activate NO generation independent of NO synthases [66]. Thus, strategies to increase NO production (such as the mechanical momentum applied by an ultrasound wave to the vessel wall during its propagation) may be useful in haemodynamically compromised tissue by promoting microcirculation, collateral and interstitial flow.

These observations are in line with the results of our recent experiment in a permanent middle cerebral artery occlusion rat model exposed to 2 MHz PW ultrasound (such as used with TCD) at two levels of insonation power in the absence of pharmacological fibrinolysis. In this experiment, smaller infarcts were seen with low power US with most robust reduction in brain oedema as compared with controls (no ultrasound) suggesting that low power US monitoring of the brain vasculature may have a fibrinolysis-independent beneficial effect, that could be synergistically applied in addition to sonothrombolytic promotion of tPA-induced recanalisation [2].

CLOTBUSTER

Sonothrombolysis for acute ischaemic stroke enters testing in the pivotal efficacy multi-national trial called CLOTBUSTER (combined lysis of thrombus using 2 MHz PW ultrasound and systemic TPA for emergent revascularisation) [67]. Briefly, all patients will receive 0.9 mg/kg intravenous tPA therapy (10% bolus, 90% continuous infusion over 1 h, maximum dose 90 mg) as standard of care according to national labels within 4.5 h from symptom onset. All patients with National Institutes of Health Stroke Scale (NIHSS) scores ≥10 points are eligible and after signing a written informed consent they will wear an operator-independent ultrasound emitting device for 2 h. The proprietary device (Cerevast Therapeutics, Redmond, WA) exposes traditional transcranial Doppler bone windows for sequential insonation of the 12 proximal intracranial segments that most commonly contain thrombo-embolic occlusions causing disabling strokes. Patients will be randomised 1:1 to continuous exposure to 2 MHz PW ultrasound versus sham exposure. No pre-treatment proof of a proximal arterial occlusion would be required since angiography is not a standard of care for evaluation of tPA-eligible patients at most institutions. Furthermore, NIHSS ≥10 points identify severe cerebral ischaemia caused by proximal occlusions in >80% of patients [68], [69]. Safety will be determined by the incidence of sICH within 24 h of treatment. Functional recovery will be determined by modified ranking scores (primary end point mRS 0–1) at 3 months. CLOTBUSTER is a large simple efficacy clinical trial, the first of its kind for sonothrombolysis.

Once CLOTBUSTER establishes safety and efficacy of an operator-independent 2 MHz PW ultrasound device, the next phase clinical trials can commence combining experimental μS with regulatory-approved tPA therapy and safe ultrasound exposure. This exposure is needed to activate μS, but there should be a proof of safety of ultrasound before a complex combinatory treatment with or without tPA can be tested according to current regulatory requirements.

Dual mode ultrasonic thrombolysis (DMUT)

Another device which seems promising intermittently exposes patients to MHz and kHz frequencies [20]. Dual mode ultrasonic thrombolysis was developed as an alternative approach to LFUS wherein diagnostic power M-mode Doppler (PMD) ultrasound is combined with LFUS, with a goal of increased safety. The effectiveness of DMUT was successfully explored in vitro. The DMUT system exposed occlusive clots in flow models to PMD (2 MHz) and LFUS (550 kHz) beams in alternating fashion from a small 12 mm diameter probe in combination with the fibrinolytic agent monteplase. Recanalisation times were then compared among three groups: the control (monteplase alone), PMD (monteplase plus PMD) and DMUT (monteplase plus PMD plus LFUT). Recanalisation times were significantly faster in both the PMD and the DMUT groups compared with the control groups. The authors hypothesised that DMUT presents a safer and more efficient approach than normal LFUS and it is currently undergoing pilot clinical testing (H. Furuhata, personal communication).

Conclusion

Intravenous thrombolytic treatment with tPA is still the only approved treatment that improves functional outcomes if given within 4.5 h from symptom onset [70]. However, even after early thrombolytic therapy the majority of stroke patients frequently have persistent major arterial occlusions and unfavourable functional outcomes [71], emphasising the need to explore further ways to augment early reperfusion and improve outcomes. Besides endovascular therapies and novel therapeutic strategies currently under clinical investigation, sonothrombolysis presents an opportunity to improve the only effective stroke therapy and to extend the timeframe for reperfusion therapies for stroke.

Sonothrombolytic treatment will need to be fine-tuned by rational choice of patient exposure to energy levels as well as creation of new devices allowing operator-independent targeted delivery of this energy. Drug-device combinations will enter future clinical trials, and a fresh look at study designs as well as pathways to regulatory approvals will be needed once these complex technologies are combined with existing or experimental drugs.

Declaration of interest: Andrei V. Alexandrov serves as consultant to Cerevast Therapeutics and holds a US patent 6733450 ‘Therapeutic Method and Apparatus for Use of Sonication to Enhance Perfusion of Tissues’, assignee – Texas Board of Regents; licensed to Cerevast Therapeutics. Kristian Barlinn was supported through NINDS SPOTRIAS grant (principal investigator (PI) – James Grotta, University of Texas, Houston), project CLOTBUST-Hands Free, a phase II study of an operator-independent device for sonothrombolysis in stroke. The authors alone are responsible for the content and writing of the paper.