| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
Department of Neurological Surgery, University of Washington School of Medicine, 1959 NE Pacific Street, Box 356470, Seattle, WA 98195-6470, e-mail: mesiwala@u.washington.edu
Editor:
We read with interest the article by Hynynen and colleagues in the September 2001 issue of Radiology (1). Their study of focused ultrasound–mediated blood-brain barrier opening is important, since targeted reversible blood-brain barrier disruption has been a major goal in the neurosciences. In the article, the authors conclude that "...low-power focused ultrasound applied after intravenous administration of [ultrasonographic] US contrast agent can be used to induce reproducible, consistent, focal blood-brain barrier opening without obvious permanent damage to the brain tissue." Unfortunately, findings of this study do not demonstrate a consistent nondestructive blood-brain barrier opening and raise several concerns regarding the authors’ experimental design, interpretation of results, discussion, and conclusions.
First, the experimental design outlined in the study is incomplete. The authors describe only one treatment group and one control group. On the basis of their results, they conclude that focused ultrasound combined with a US contrast agent leads to blood-brain barrier opening. However, use of a US contrast agent alone has been shown (2) to cause blood-brain barrier disruption. While the result was obtained at doses higher than those used in the present protocol, the localization pulse of focused ultrasound used in this study may allow the lower doses of US contrast agent to disrupt the blood-brain barrier. Thus, the authors need two additional treatment groups: (a) US contrast agent alone, without ultrasound localization or treatment; and (b) ultrasound localization followed by US contrast agent, without subsequent ultrasound treatment. Unless these additional studies are performed, it is premature to make the conclusions presented in the article.
Second, the interpretation of histologc findings is problematic. In all specimens, including those that the authors claim show no neuronal damage, hemorrhage and damage to the cerebrovasculature are seen without any reported assay of the brain parenchyma. For example, glia are much more than passive structures (3), and their status must also be assessed. Furthermore, complete histologic analysis of all components of the brain parenchyma is warranted, since the hemorrhagic damage seen in this experiment is analogous to cerebral microbleeds seen in other conditions (6–9), which typically lead to local inflammation, release of various chemotactants, vascular permeability mediators, and local death of glia, microglia, and neurons. The damage is permanent and appears as a scar or infarct several weeks later. If the tissue downstream from the ruptured or damaged capillary depends on this blood vessel for oxygen and nutrients, then a lacunar infarct may result.
Moreover, if one looks at just one of the factors that is released or produced at the site of hemorrhagic damage, for example, bradykinin, a cascade of reactions ensues that eventually leads to neuronal, glial, and microglial injury and death (9), which can take days to weeks to evolve. To justify their claim that hemorrhage does not yield neuronal or parenchymal damage to the brain, the authors cite their previous conclusions that the immediate parenchymal damage seen with hematoxylin-eosin staining of thermally induced US lesions is identical to the damage revealed with cresyl violet and TUNEL stains days following the thermal injury (4). Besides the fact that hematoxylin-eosin staining in other contexts is sufficient to show hemorrhagic damage, thermal lesions (4), unlike cavitation lesions (5), are not characterized by hemorrhage. Therefore, we are puzzled that the authors do not see parenchymal damage with hematoxylin-eosin staining, and we predict that if cresyl violet or TUNEL staining was performed for these hemorrhagic lesions, parenchymal damage would be seen.
Finally, the authors’ claim that "the presence of red blood cells in the extravascular space demonstrates more pronounced temporary damage of the capillary walls" is difficult for us to understand. The authors have not shown that the capillaries reestablish their normal architecture, strength, and blood-brain barrier functions after their rupture, other than to show lack of flux on magnetic resonance (MR) images some days after the study. These damaged capillaries may never reestablish a normal endothelial lining, may be more prone to spontaneous rupture, or may have a permanently altered blood-brain barrier, all in ways that are apparent only weeks after ultrasound treatment.
Third, the authors’ contention that the hemorrhagic damage is temporary and of little functional consequence, since their experimental animals acted normally after treatment, is troubling. The authors have not performed standardized functional tests (10) in these animals, leaving the functional consequences of their protocol unexplored. This is of special concern, since it appears that the US lesions in this experiment are located in the thalamus or brainstem, although this is unclear. If such a microbleed were to occur in the thalamus or brainstem of a human, severe neurologic deficits would result that seem out of proportion to the size or volume of the hemorrhage. Thus, without detailed functional tests that are specific for the animals used in this experiment, the authors can only conclude that hemorrhage in the targeted areas of brain in this study did not yield gross obvious neurologic deficits.
Fourth, it is unclear how the authors were able to localize these lesions for histologic analysis for as many as 7 days afer treatment without the use of a visible tracer, such as trypan or Evans blue stains, or without lesion placement relative to anatomic landmarks. One possibility is that the sections of the brain they examined histologically were identified as the US targets on the basis of the appearance of hemorrhage. Since hemorrhagic brain is a damaged brain, this would contradict their claim that the procedure does not cause damage to brain tissue. Alternatively, histologic analysis of one 6-µm thick section every 0.3 mm leaves the possibility that parenchymal damage in addition to hemorrhage was missed. We have found that focused ultrasound–mediated blood-brain barrier disruption, without acoustic contrast agents, can be associated with parenchymal damage approximately 40% of the time (in 60% of cases, no damage of any sort is seen with light and electron microscopy), so that the extent of this damage can be less than 0.1 mm in its largest dimension (11–14).
Last, the conclusion that cavitation alone is responsible for blood-brain barrier opening is premature. Since hemorrhage near and damage to capillaries were reported for the ultrasound applications that resulted in blood-brain barrier opening in this experiment, bradykinin and other inflammatory mediators may have ultimately contributed to the blood-brain barrier disruption. Also, US contrast agent independent of cavitation may have contributed to blood-brain barrier disruption in this setting.
We applaud the authors’ development of a sophisticated apparatus for monitoring the biologic effects of focused ultrasound beams and US contrast agents with MR imaging. These techniques may also ultimately yield a means of treating the brain with ultrasound beams applied transcranially. However, given the concerns we have raised, we believe that one may best characterize the results thus far as being similar to those in previously published reports that demonstrate that focused ultrasound alone can disrupt the blood-brain barrier while causing simultaneous damage (15–19).
REFERENCES
Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis Street, Boston, MA 02115
We would like to respond to the letter of Drs Mesiwala and Mourad concerning our recent article (1). We appreciate their interest in our study. The current article is a continuation of our prior research in which we investigated the effect of high-intensity ultrasound pulses (2) and continuous wave focal heating (3–5) in the brain. In those studies and the studies cited by Drs Mesiwala and Mourad, blood-brain barrier disruption was sometimes accompanied by brain tissue damage over all ultrasound exposure parameters. In our recent study, we reported parameters that produced a more consistent blood-brain barrier opening when preformed gas bubbles were systemically introduced. We consistently detected no neuronal damage with use of a standard histologic method, which was not the case in the previous studies. We also demonstrated that temporal average peak power levels could be reduced by at least two orders of magnitude below those used in thermal treatments to produce biologic effects in the brain. This reduction will be enormously important when transcranial treatments of the brain (not just blood-brain barrier opening) are considered.
We agree that our histologic study, although a promising first step for a parametric screening study, is not all inclusive. Now that exposure conditions that show promise have been found, more studies need to be performed to verify the findings and search for other potential cellular or subcellular structural changes in the brain tissue that are not detectable with the method we used. Many more investigations are also warranted in the search for vascular damage and its effect on brain tissue. In an ongoing unpublished study, we have addressed several of the issues mentioned by Drs Mesiwala and Mourad. Since their letter was submitted, we have completed a study in which we looked for both apoptotic and ischemic cells that are associated with ultrasound exposures in the presence of the preformed gas bubbles. We used different MR imaging contrast agents to evaluate the size of the molecule that can pass through the blood-brain barrier opening and its extent, and we have performed, to our knowledge, the first experimental series to test this method for gene therapy. Moreover, we used electron microscopy to investigate the ultrastructural changes in the capillary walls that could be connected with the blood-brain barrier disruption. Finally, functional studies with sonications through the intact skull in a primate model are planned to evaluate potential short- and long-term physiologic effects of the opening of blood-brain barrier. The results of these studies will appear in future submissions.
We agree that more studies need to be performed to establish the exact biologic mechanism responsible for the blood-brain barrier disruption. However, we believe that cavitation is the primary physical cause of the blood-brain barrier opening. In our experiments, we clearly demonstrated that bulk temperature elevation was not the cause of the blood-brain barrier opening. It was also clear that neither the sonications nor the contrast agent alone were responsible for the focal leakage of the MR contrast agent. The opening was related to the combination of the systemic introduction of the preformed gas bubbles and the effect of sonication localization; it was the result of an interaction between the ultrasound beam and the gas bubbles. Such an interaction is defined (6) as cavitation. This cavitation activity is associated with multiple microscopic high-energy effects, including shear forces, shock waves, streaming, radiation pressure, temperature elevations close to the bubbles, temporary stasis of the blood flow, and the like. These physical stresses can cause any number of biologic-tissue reactions that may be responsible for the actual disruption of the blood-brain barrier. We are actively investigating the biologic mechanism of this promising phenomenon.
We are happy to see that Drs Mesiwala and Mourad are working on the same problem, and we look forward to seeing their yet unpublished article.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
