Soluble FLT1 Gene Therapy Alleviates Brain Arteriovenous Malformation Severity
Background and Purpose—Brain arteriovenous malformation (bAVM) is an important risk factor for intracranial hemorrhage. Current therapies are associated with high morbidities. Excessive vascular endothelial growth factor has been implicated in bAVM pathophysiology. Because soluble FLT1 binds to vascular endothelial growth factor with high affinity, we tested intravenous delivery of an adeno-associated viral vector serotype-9 expressing soluble FLT1 (AAV9-sFLT1) to alleviate the bAVM phenotype.
Methods—Two mouse models were used. In model 1, bAVM was induced in R26CreER;Eng2f/2f mice through global Eng gene deletion and brain focal angiogenic stimulation; AAV2-sFLT02 (an AAV expressing a shorter form of sFLT1) was injected into the brain at the time of model induction, and AAV9-sFLT1, intravenously injected 8 weeks after. In model 2, SM22αCre;Eng2f/2f mice had a 90% occurrence of spontaneous bAVM at 5 weeks of age and 50% mortality at 6 weeks; AAV9-sFLT1 was intravenously delivered into 4- to 5-week-old mice. Tissue samples were collected 4 weeks after AAV9-sFLT1 delivery.
Results—AAV2-sFLT02 inhibited bAVM formation, and AAV9-sFLT1 reduced abnormal vessels in model 1 (GFP versus sFLT1: 3.66±1.58/200 vessels versus 1.98±1.29, P<0.05). AAV9-sFLT1 reduced the occurrence of bAVM (GFP versus sFLT1: 100% versus 36%) and mortality (GFP versus sFLT1: 57% [12/22 mice] versus 24% [4/19 mice], P<0.05) in model 2. Kidney and liver function did not change significantly. Minor liver inflammation was found in 56% of AAV9-sFLT1–treated model 1 mice.
Conclusions—By applying a regulated mechanism to restrict sFLT1 expression to bAVM, AAV9-sFLT1 can potentially be developed into a safer therapy to reduce the bAVM severity.
Brain arteriovenous malformations (bAVMs) tend to rupture spontaneously causing intracranial hemorrhage. Current treatments are associated with high morbidities/mortalities. New, effective, and safe therapies are, therefore, needed.
Vascular endothelial growth factor (VEGF) is abnormally high in bAVM lesions.1 Patients with an autosomal dominant disease, Hereditary Hemorrhagic Telangiectasia, have a higher incidence of AVMs in multiple organs, including the brain. Increased VEGF has also been found in Hereditary Hemorrhagic Telangiectasia patients’ plasma.2 Therefore, inhibiting VEGF may be effective to treat bAVM. However, considerable side effects are associated with commonly used blocking agents: VEGF antibodies3 and tyrosine kinase inhibitors.4
Soluble FMS-like tyrosine kinase 1 (sFLT1) contains only the extracellular domains of FLT1 (VEGFR1) and binds with VEGF.5
The adeno-associated virus (AAV) mediates long-term transgene expression in nondividing cells. AAV serotype-9 (AAV9) passes through the blood–brain barrier around the angiogenic region.6 Intravenous injection of AAV9-sFLT1 inhibits VEGF-induced brain angiogenesis.6
Our study shows that through intrabrain or intravenous injection, AAV-sFLT1 attenuates bAVM phenotypes.
The Institutional Animal Care and Use Committee of the University of California, San Francisco approved protocol/experimental procedures. Institutional Animal Care and Use Committee and Animal Care Facility staff provided animal husbandry.
Eng2f/2f (Endoglin, Hereditary Hemorrhagic Telangiectasia–causative gene) mice7 were crossbred with R26CreER or SM22αCre mice (Jackson Laboratory, Bar Harbor, ME) to produce R26CreER;Eng2f/2f or SM22αCre;Eng2f/2f mice.
In model 1, 8-week-old R26CreER;Eng2f/2f mice were intraperitoneally injected with tamoxifen (2.5 mg/25g of body weight) daily for 3 consecutive days to globally delete the Eng gene, and intrabrain injected with AAV1-VEGF to induce focal angiogenesis when the first dose of tamoxifen was given; bAVM developed 8 weeks later.8 AAV1-sFLT02 was coinjected with AAV1-VEGF at the time of model induction, and AAV9-sFLT1, intravenously injected 8 weeks after.
In model 2, SM22αCre;Eng2f/2f mice were used. AAV9-sFLT1 was intravenously injected to 4- to 5-week-old mice.
Random group assignment was applied, and treatment end point selected based on a previous study.6
Sample sizes are shown in the figures. GraphPad Prism 6 was used to analyze data, t test for comparing 2 groups, and 2-way ANOVA with Tukey correction for >2 groups with multiple comparisons. Survival rate was analyzed using log-rank test. Data are presented as mean±SD. Pvalue <0.05 was considered statistically significant. All methods are described in the online-only Data Supplement.
To test whether sFLT1 inhibits bAVM formation, AAV2-sFLT02 containing the VEGF-binding domain 2 of human sFLT1 and a CH3 domain of IgG16 was coinjected with AAV1-VEGF into the brain of model 1 mice when the first dose of tamoxifen was given. Eight weeks later, cerebrovasculature was latex cast (Figure IA and IB in the online-only Data Supplement). Because of size, latex particles enter the veins only when there is an arteriovenous shunt (an AVM hallmark). bAVMs were detected in AAV2-EV (empty vector)–treated mice, but not in AAV2-sFLT02–treated mice (Figure 1A). AAV2-sFLT02–treated mice had a smaller latex-perfused area than AAV2-EV–injected mice (P=0.037; Figure 1B; Figure IC in the online-only Data Supplement), indicating that sFLT02 inhibited bAVM formation.
To reduce risk of intralesion injection, AAV9-sFLT1 expressing a full-length human sFLT16 or AAV9-GFP (green fluorescent protein, control) was intravenously delivered to model 1 mice 8 weeks after model induction when bAVM had developed (Figure IIA and IIB in the online-only Data Supplement).8 Gene expression in bAVM was confirmed via histological analysis for GFP and ELISA for sFLT1 (Figure IIC through IIE in the online-only Data Supplement).
Four weeks later, therapeutic effect was evaluated by analyzing vessel density and dysplasia vessels (dysplasia index: number of vessels larger than 15 μm/200 vessels)8 using fresh-frozen brain sections (Figure IIB in the online-only Data Supplement). Compared with AAV9-GFP–treated mice (3.66±1.58), AAV9-sFLT1–treated mice had a lower dysplasia index (1.98±1.29, P=0.011) and a trend toward lower vessel density (P=0.17; Figure 2), indicating that AAV9-sFLT1 reduced bAVM severity. No GFP signal was detected in the fresh-frozen sections of Ad-GFP–injected brain (Figure III in the online-only Data Supplement). The positive signals of fluorescent-labeled lectin and CD31 antibody staining were completely colocalized (Figure IIIB in the online-only Data Supplement).
To confirm that the effect in model 1 was not because of inhibition of exogenous VEGF used to induce bAVM, we tested AAV9-sFLT1 in model 2, in which bAVMs develop spontaneously without exogenous VEGF stimulation. AAV9-sFLT1 was intravenously injected into 4- to 5-week-old mice (Figure IVA in the online-only Data Supplement) because 90% of model 2 mice had bAVMs by 5 weeks and 50% died by 6 weeks.8 About 57% of AAV9-GFP–treated and 24% of AAV9-sFLT1–treated mice died during the 4-week treatment period (P=0.036; Figure 3A; Table I in the online-only Data Supplement). All AAV9-GFP–treated mice had bAVM; only 36% of AAV9-sFLT1–treated mice had detectable bAVMs (Figure 3B and 3C; Figure IVB through IVD). Therefore, AAV9-sFLT1 treatment also reduced the severity of spontaneously developed bAVM.
Potential adverse effects on the liver and kidney were analyzed. The activities of alkaline phosphatase and alanine transaminase and the levels of creatinine were similar in AAV9-GFP–treated, AAV9-sFLT1–treated, and untreated groups in model 1 (Figure VA through VC in the online-only Data Supplement). Small clusters of inflammatory cells (including monocytes and lymphocytes) were detected in the liver of 58% of AAV9-GFP–treated and 56% of AAV9-sFLT1–treated mice, but not in AAV9-EV–injected mice (Figure VD and VE in the online-only Data Supplement; Table II in the online-only Data Supplement). Body weight of AAV9-sFLT1–treated R26CreER;Eng2f/2f mice did not increase as much as control mice during the treatment period and was lower than the control groups at the end of therapy (P=0.044; Figure VI in the online-only Data Supplement).
We tested an antiangiogenic gene therapy to treat bAVM in (1) VEGF-induced adult onset model and (2) spontaneously developed model. We showed that in situ injection of AAV1-sFLT02 at the time of model induction inhibited bAVM formation in model 1; intravenously delivered AAV9-sFLT1 reduced bAVM severity in both models.
VEGF functions mainly through VEGFR1 (or FLT1) and VEGFR2. Whereas VEGFR2 is known to mediate endothelial cell mitosis and vascular permeability, VEGFR1-mediated signaling is complex and context dependent. VEGF–VEGFR1 signaling also induces monocytes homing to the injured tissue. sFLT1 effect most likely occurs when it binds to excessive VEGF in the brain parenchyma, thus quenching VEGF signaling through membrane-bound receptors. We found that sFLT1 overexpression reduced dysplasia vessels, with no significant reduction of CD68+ cells in the bAVM (data not shown), suggesting that therapeutic effect might be achieved by reducing VEGFR2 signaling. Although we were not able to quantify the number of dysplasia vessels in model 2 because of the unpredictable lesion locations, AAV9-sFLT1 reduced mortality and the presence of bAVM in this model, suggesting that the effect observed in model 1 was not merely through exogenous VEGF inhibition. Future studies will be needed to determine the underlying mechanisms.
Bevacizumab (Avastin) treatment in an adult onset Alk1-deficient model reduced bAVM severity.9 However, bevacizumab causes bilateral pulmonary embolisms, thrombosis, and hypertension in Hereditary Hemorrhagic Telangiectasia patients.3 Although intravenously delivered AAV9-sFLT1 caused weight loss and liver inflammation, AAV9-EV did not. Therefore, the side effects were caused by constitutive sFLT1 expression. An antitumor study showed ascites and kidney damage in mice treated with adenoviral vector expressing sFLT1 constitutively, but not in mice treated with intermittent sFLT1 expression,10 indicating that sFLT1 side effects can be reduced by controlling expression. An AAV gene therapy has been approved recently, and many clinical trials are ongoing.11 AAV gene therapy could be developed into a safer and an effective therapy to treat chronic diseases such as AVM.
Although our mouse models have many key characteristics of human bAVM,12 no hemodynamic changes in the AV fistula and the nearby brain–blood vessels have been studied. However, our findings suggest that AAV-sFLT1 could be developed into a minimally invasive and safe antiangiogenesis gene therapy for bAVM. Minor adverse effects could be minimized by controlling sFLT1 expression.
Sources of Funding
This study was supported by research grants from the National Institutes of Health (R01 NS027713, R01 HL122774, and R21 NS083788), the Michael Ryan Zodda Foundation, and University of California, San Francisco Research Evaluation and Allocation Committee (REAC) to Dr Su and by a Young Investigator Award from Cure HHT (Hereditary Hemorrhagic Telangiectasia) Foundation to Dr Zhu.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.116.015713/-/DC1.
- Received October 12, 2016.
- Revision received January 5, 2017.
- Accepted January 23, 2017.
- © 2017 American Heart Association, Inc.
- Sadick H,
- Riedel F,
- Naim R,
- Goessler U,
- Hörmann K,
- Hafner M,
- et al
- Tabouret T,
- Gregory T,
- Dhooge M,
- Brezault C,
- Mir O,
- Dréanic J,
- et al
- Shen F,
- Mao L,
- Zhu W,
- Lawton MT,
- Pechan P,
- Colosi P,
- et al
- Walker EJ,
- Su H,
- Shen F,
- Degos V,
- Amend G,
- Jun K,
- et al
- Sivanandam VG,
- Stephen SL,
- Hernandez-Alcoceba R,
- Alzuguren P,
- Zabala M,
- van Rooijen N,
- et al
- Pollack A