Abbreviations
AVM — arteriovenous malformations
ANG — angiopoietin
MMP — matrix metalloproteinase
VEGF — vascular endothelial growth factor
It is known that cerebral arteriovenous malformations (AVMs) are not static congenital lesions. Moreover, malformations are able to grow, recur and even appear de novo after complete recovery. Impaired balance in releasing vascular endothelial growth factor (VEGF), angiopoietins (ANG), matrix metalloproteinases and their receptors results a continuous angiogenic process in AVM. High level of expression of these factors correlates with biological activity of pathological process [1—4].
The aim of the study was to analyze the dynamics of angiogenic factors before and after embolization of cerebral AVMs.
Material and methods
The study was conducted at the Vascular Surgery Department of the Polenov National Research Neurosurgery Institute. Serum VEGF, ANG-2 and MMP-9 were assessed in 314 patients with cerebral AVM and 33 healthy ones for the period from 2011 to 2017. All patients underwent endovascular embolization of cerebral AVMs with glue (Histoacryl) and non-adhesive (Onyx) compositions. Embolization was carried out both through transarterial and transvenous approaches. There were several stages of endovascular treatment (2—7) in 249 out of 314 patients. Mean age of patients was 34.1±1.4 years (range 18—64 years), age of healthy volunteers — 33.9±2.5 years (range 25—52 years). The levels of ANG-2, VEGF and MMP-9 were analyzed before surgical treatment (on the day of surgery) and in 1 day after endovascular intervention. Venous blood was sampled on an empty stomach into a vacuum tube with a coagulation activator. Centrifugation was performed in biochemical laboratory (2700 rpm for 20 min) to obtain serum. Subsequent aliquotation was followed by storage of serum at a temperature of –20 °С. Analysis of concentration of VEGF, ANG-2 and MMP-9 was carried out using Personal Lab microplate analyzer ("Adaltis", Italy). Invitrogen test systems (USA) were used to determine the level of VEGF, Research & Diagnostics Systems (USA) test systems — for MMP-9 and ANG-2. The manufacturer proposed reference values in instructions for each test system. These values were determined in analysis of blood of 60 healthy volunteers. In 33 healthy volunteers of our sample, VEGF level was 271.2±41.1 pg/ml, MMP-9 — 432.5±48.4 ng/ml, ANG-2 — 2197.7±273.9 pg/ml. Our values coincided with reference values obtained by the manufacturer of test systems in healthy volunteers. These values were accepted as control. Increase of the level of serum angiogenic factors was determined considering values in healthy volunteers exceeding standard deviation [1, 4].
Results
Mean preoperative level of serum VEGF in patients with cerebral AVM was 349±22.9 pg/ml, MMP-9 — 596.3±30.4 ng/ml, ANG-2 — 2971.4±80 pg/ml. An increased level of at least one of the investigated angiogenic factors was detected in 152 (48.4%) out of 314 patients with AVM. In other patients, preoperative levels of all angiogenic factors were similar to those in healthy volunteers. VEGF augmentation was found in 151 (48%) out of 314 patients, normal values — in 163 patients. Increased level of MMP-9 was detected in 147 (46.8%) patients, increase of ANG-2 in 126 (40,1%) patients.
No correlation between increased level of angiogenic factors and gender, age, comorbidities and localization of AVM (supra- and subtentorial) was found. Drainage into deep cerebral veins was predominant (n=89). In case of this drainage, increased level of angiogenic factors was found in 54 (60.7%) patients (increased level of VEGF — 21 (38.9%) out of 54 patients, ANG-2 in 22 (40.7%) ones, MMP -9 — in 19 (35.2%) patients).
The relationship between clinical course of AVM and the level of angiogenic factors was analyzed. Epileptic course of AVM was observed in 119 patients, hemorrhages from AVMs — in 172 patients. In 23 patients, AVMs were found during examination for headaches. The level of MMP-9 was elevated in 102 (59.3%) out of 172 patients with hemorrhagic course of AVM.
Analysis included 152 patients with elevated preoperative levels of angiogenic factors and no previous surgery. Small AVMs (Spetzler—Martin grade I—II) were noted in 12 (7.9%) patients, Spetzler—Martin grade III — in 98 (64.5%) cases, grade IV—V — in 42 (27,6%) cases. Thus, the largest number of patients (64.5%) with increased levels of growth factors was found among patients with AVM Spetzler — Martin grade III. The number of patients with increased VEGF and MMP-9 was comparable in groups of small (Spetzler—Martin grade I—II) and large/giant AVMs (Spetzler—Martin grade IV—V). At the same time, increased level of ANG-2 was observed in 19 (45.2%) out of 42 patients with large and giant AVMs (Spetzler—Martin grade IV—V). This value was almost 2 times higher than in patients with small AVM (Spetzler—Martin grade I—II) (3 (25%) out of 12 patients). These data are consistent with the results obtained in analysis of the structure of AVM drainage. Increased level of ANG-2 was observed in 40.7% of patients with deep venous drainage as the most common type in large and giant AVMs. We analyzed changes of angiogenic factors after embolization depending on Spetzler—Martin grade of AVM. Each stage of surgery was separately analyzed. Partial embolization was followed by MMP-9 augmentation over preoperative value in 2 (16.7%) out of 12 patients with small AVM (Spetzler—Martin grade I—II). At the same time, in patients with AVM Spetzler—Martin grade III this value was 58.2% (57 out of 98 patients). Postoperative augmentation of MMP-9 was noted in 24 (57.1%) out of 42 patients with AVM Spetzler—Martin grade IV—V.
In patients with small AVMs (Spetzler—Martin grade I—II), partial embolization was followed by VEGF augmentation in 33% (4 out of 12 patients), in patients with AVM grade III — 37.8% (37 out of 98 patients), in patients with large and giant AVM (Spetzler—Martin grade IV—V) — 35.7%.
Thus, partial embolization of AVMs results further increase of MMP-9 level in approximately 50% of patients with large and giant AVMs. VEGF is increased in approximately 30% of patients regardless Spetzler—Martin grade. External carotid artery was involved into blood supply of AVM in 47 (30.9%) out of 152 patients. Elevated level of VEGF was detected in 23 (48.9%) of these patients, MMP-9 — in 19 (40.4%) cases, ANG-2 — in 9 (19%) patients. Further inclusion of external carotid artery into blood supply of AVM between the stages of embolization resulted significant augmentation of the levels of angiogenic factors in other 36 out of 139 patients after partial embolization of AVMs. Total occlusion of AVM within one endovascular stage was performed in 13 out of 152 patients. Occurrence of new afferent of AVM between endovascular stages was followed by further augmentation of VEGF in 21 (58.3%) out of 36 patients, MMP-9 in 15 (41.7%) cases and ANG-2 in 9 (25%) patients. Angiographic picture between endovascular stages was analyzed in 65 patients without new afferents from external carotid artery. VEGF continued to increase even before redo surgery in 18 (27.7%) patients, MMP-9 — in 17 (26.1%) cases, ANG-2 — in 8 (12.3%) cases.
Depending on the quality of embolization, 314 study patients were divided into subgroups: 86 patients underwent total endovascular embolization of AVM with various compositions and their combinations (Onyx, Histoacryl), 228 patients underwent partial endovascular embolization of AVM. The result has been achieved for 1—7 surgical stages.
Changes of angiogenic factors after each surgical stage were analyzed depending on certain composition for embolization in 152 patients with elevation of one or more angiogenic factors. In these patients, total embolization was made in 49 cases, partial occlusion — in 103 patients. Onyx non-adhesive composition was applied in 84 (55.3%) out of 152 patients, Histoacryl adhesive composition – in 68 (44.7%) cases.
Onyx non-adhesive composition ensured one-stage total occlusion of AVM in 9 (10.7%) patients. Moreover, postoperative reduction of the level of all angiogenic factors was observed. Partial embolization of AVM was followed by a further increase of VEGF level in 24 (33.8%) out of 71 patients. No further postoperative augmentation of VEGF was noted in 47 patients. Further increase of elevated level of MMP-9 was detected in 28 (39.4%) out of 71 patients. Stable concentration of this factor was found in 43 cases. Further increase of elevated level of ANG-2 was detected in 13 (18.3%) patients. Stable concentration of this factor was observed in 58 cases.
Thus, each stage of partial embolization of AVMs with Onyx non-adhesive composition was accompanied by further augmentation of one or more angiogenic factors in 52 (77.7%) out of 71 patients.
Total one-stage embolization with Histoacryl adhesive composition was performed in 4 (5.9%) out of 68 patients. Postoperative normalization of all factors was confirmed. Each stage of partial embolization of AVM with Histoacryl adhesive composition was followed by further increase of angiogenic factors. Continued VEGF augmentation was noted in 19 (29.7%) out of 64 patients, stable concentrations were observed in 45 patients. ANG-2 level continued to grow only in 8 (12.5%) out of 64 patients. Among patients with elevated preoperative levels of MMP-9, postoperative augmentation was obtained in 29 (45.3%) out of 64 patients.
Thus, embolization with Histoacryl adhesive composition was followed by elevation of one or more angiogenic factors increased before surgery in 56 (87.5%) out of 64 patients. The greatest tendency to postoperative augmentation regardless of embolization material was demonstrated for MMP-9. These changes were more obvious after embolization of AVM using Histoacryl adhesive composition.
The results of total embolization of cerebral AVMs were analyzed in 86 out of 314 patients. Complete occlusion of AVM required 1—7 stages of embolization. At the same time, AVM Spetzler—Martin grade I—II was observed in 42 (48.8%) patients, Spetzler—Martin grade III — in 40 (46.5%) cases, grade IV — in 4 (4.7%) patients. In 49 (56.9%) out of 86 patients with complete occlusion of AVM, basic increase of one or more growth factors was revealed before the first operation (Fig.).
Thus, control angiography after six months is required to assess quality of embolization if postoperative angiography data on complete occlusion of AVM are associated with elevated VEGF. Abnormal levels of VEGF after embolization may be a criterion of incomplete occlusion of AVM.
Discussion
We analyzed angiogenic factors VEGF, MMP-9 and AGP-2 in patients with AVM before and after endovascular embolization performed within one or several stages. The purpose was to clarify molecular mechanisms of cerebral AVM growth. It was revealed that hemorrhagic course of AVM is often associated with increased level of MMP-9. Our data are confirmed by R.M. Starke et al. These authors showed significantly increased serum MMP-9 in patients with ruptured AVM compared with non-ruptured AVM and further early postoperative augmentation of these values [1]. Inclusion of new afferents from external carotid artery into blood supply of AVM (i.e. afferent and vasculature de novo appeared between the embolization stages) was followed by increase of VEGF between surgical stages in 58.3% of patients (compared with 27.7% of patients without new afferents). AVM Spetzler—Martin grade III were observed in 98 (64.5%) out of 152 patients with preoperative elevated angiogenic factors.
In our study, preoperative increase of the growth factors was absent in 189 out of 314 patients while further examination was not a goal of this research. Nevertheless, it should be noted that the proportion of patients with AVM Spetzler—Martin grade III in this group was 39.2% (74 patients). The greatest number of (64.5%) patients with high VEGF was noted among those with AVM Spetzler—Martin grade III. Most likely, such AVMs are prone to further growth, since VEGF overexpression causes an excessive development of abnormal blood vessels with impaired specificity of endothelial cells (both in arteries and veins) and deterioration of vascular wall structure [2, 3]. The groups of small (Spetzler—Martin grade I—II) and large/giant AVMs (Spetzler—Martin grade IV—V) were characterized by similar number of patients with increased VEGF and MMP-9. At the same time, increased ANG-2 was detected almost 2 times more often (45.2%) in patients with large and giant AVMs (Spetzler—Martin grade IV—V) compared with AVMs Spetzler —Martin grade I—II (25%). Giant malformations are characterized by several dilated veins with drainage into deep cerebral veins. In our opinion, these data correspond to ANG-2 overexpression in perivascular region around large and giant cerebral AVMs. This factor contributes to development of abnormally dilated veins without a mature wall structure [4, 5].
The most important results were obtained in assessment of angiogenic status depending on the quality of endovascular interventions.
Partial embolization of AVMs results a further increase of MMP-9 in patients — with large and giant AVMs regardless of embolization material. Continued augmentation of VEGF is observed in approximately 30% of patients with any type of AVM. Elevated level of ANG is being increased slightly after surgery. The phenomenon of postoperative AVM growth after partial embolization is explained, on the one hand, by X-ray negative hidden compartment of AVM. On the other hand, approximately 50% of patients with AVM can have genetic mechanisms predisposing to development of new vessels involved into blood supply of malformation [6].
Total embolization of AVM led to the opposite result. There were no patients with elevated levels of angiogenic factors after total embolization. All these patients underwent control angiography in 6—11 months after treatment. Complete occlusion of AVM was confirmed. Postoperative angiography confirmed complete occlusion of AVM in 6 patients while VEGF remained elevated one day after surgery. Control angiography 6 months later revealed residual functioning of malformation that required redo embolization. Thus, abnormal angiogenic factors after embolization may be a criterion of incomplete occlusion of AVM.
AVMs are characterized by dynamic postoperative vascular remodeling due to neoangiogenesis. The last one is regulated by postoperative hemodynamic restructuring, hypoxia and inflammation [6]. Analysis of pathogenesis of cerebral AVM and elucidation of molecular mechanisms of their growth are necessary. Research of dynamic models (endothelial cell cultures from resected AVMs) and analysis of serum angiogenic factors throughout the treatment are of the greatest interest. Our data confirm an important role of angiogenesis in pathophysiology of brain AVM since an imbalance of growth factors contributes to progression and growth of malformations [7]. High serum VEGF, ANG-2 and MMP-9 in patients with cerebral AVM may be important for predicting the course of the disease, risk of hemorrhage and choosing the optimal surgical strategy.
Identification of factors predisposing to further growth of arteriovenous malformations should serve as a signal for the most radical and early resection of these lesions. Moreover, a combined technique (complete resection of malformation after preliminary embolization) may be applied [8].
Conclusion
Further analysis of molecular mechanisms of AVM growth will ensure a new view at the fundamental biology of these vascular malformations and optimal treatment. Perhaps, additional knowledge about the role of angiogenesis in pathogenesis of arteriovenous malformations will allow combining endovascular embolization with the use of antiangiogenic drugs or embolization with material coated with antiangiogenic factors.
The authors declare no conflicts of interest.
Commentary
The report is devoted to analysis of angiogenic factors in patients with arteriovenous malformations before and after endovascular treatment. Previously, the authors published literature review on this issue. The relevance of the study is obvious since it concerns the fundamental problems of development and biological behavior of AVM. The last one is rare but causes significant cerebral disruption. Understanding the molecular mechanisms underlying arteriovenous malformations is valuable to modify and improve the quality of treatment. It should be emphasized that there are no similar reports in national literature, and the authors are pioneers in Russia.
The research is based on a large clinical data (314 patients with cerebral arteriovenous malformations). All patients underwent the same survey. A variety of arteriovenous malformations and their clinical course made it possible to compare the results in different groups of patients.
Analysis of 3 factors of angiogenesis revealed increase of at least one of them in approximately 50% of patients prior to surgery. Certain correlations of changes of each factor with morphological characteristics of AVM were obtained. There were multidirectional postoperative changes of these factors (both augmentation and normalization). The authors tried to determine the causes of these changes. The most definite conclusion was made regarding changes before and after embolization. Thus, complete occlusion of AVM is followed by early normalization of all factors.
Undoubtedly, the report raises many questions, first of all, regarding evidence level of these data. Statistical significance of differences is given only for VEGF before and after embolization. In this regard, some conclusions are perceived as overly categorical. There are numerous objectives in this research. Therefore, tables would be desirable to facilitate perception of information.
Nevertheless, the manuscript is of undoubted interest not only for biologists, but also for neurosurgeons and neurologists.
O.B. Belousova (Moscow, Russia)