Advances in sporadic brain arteriovenous malformations: Novel genetic insights, innovative animal models and emerging therapeutic approaches

Abstract

Brain arteriovenous malformations (bAVMs) are a notable cause of intracranial hemorrhage, strongly associated with severe morbidity and mortality. Contemporary treatment options include surgery, stereotactic radiosurgery, and endovascular embolization, each of which has limitations. Hence, development of pharmacological interventions is urgently needed. The recent discovery of the presence of activating Kirsten rat sarcoma (KRAS) viral oncogene homologue mutations in most sporadic bAVMs has opened the door for a more comprehensive understanding of the pathogenesis of bAVMs and has pointed to entirely novel possible therapeutic targets. Herein, we review the status quo of genetics, animal models, and therapeutic approaches in bAVMs.

Introduction

Brain arteriovenous malformations (bAVMs) involve abnormal direct connections between cerebral arteries and veins via a vascular nidus that bypasses normal capillary circulation and forms a high-flow shunt of arterial blood to a low-pressure venous system.

The management of bAVMs, particularly for unruptured bAVMs,^1,2 is controversial, with surgical resection being the gold standard for treatment. However, the safety and efficacy of surgical resection are related to the size of the bAVMs, associated brain eloquence, and/or venous drainage pattern, ³ and only a subset of bAVMs are amenable to safe surgical resection. ⁴ Stereotactic radiosurgery (SRS) is another potentially curative treatment option that leads to progressive thrombosis and AVM obliteration. Nonetheless, this process occurs over months to years, and the risk of hemorrhage during the latency period is unchanged as compared to that in the natural history.^5,6 Finally, endovascular embolization can occasionally be used with curative intent, but has low cure and high complication rates.^7,8 Currently, no pharmacological intervention exists for bAVMs, which represents a major gap in our armamentarium for treating such lesions. The lack of pharmacologic interventions for bAVMs is due to a lack of knowledge of the underlying pathogenesis of bAVMs.

Most bAVMs are sporadic, without a family history. ⁹ The recent discovery that most sporadic bAVM cases have oncogenic Kirsten rat sarcoma (KRAS) viral oncogene homologue mutations has opened new research avenues, offering insights into shared pathways and potential targeted therapies, and improving our understanding of sporadic bAVMs. ¹⁰ The discovery of KRAS mutations has led to the development of new animal models of bAVMs, ¹¹ such as genetically engineered mouse models. These models closely mimic the genetic alterations found in human bAVMs, providing a valuable tool for studying the pathogenesis of the disease. By studying these animal models, researchers can gain insights into how KRAS mutations contribute to bAVM formation and progression, facilitating the development of novel therapeutic approaches.

By providing a comprehensive overview of the current genetically engineered animal models of sporadic bAVMs, we hope to inspire further research and contribute to developing effective treatments for this devastating condition.

KRAS introduction

KRAS was originally discovered in Kirsten sarcoma virus ¹² and acts as a molecular switch that finely controls a variety of signaling pathways. ¹³ Upstream stimuli, such as vascular endothelial growth factor (VEGF), can activate KRAS, which then mediates various signaling pathways involved in cell proliferation, cell migration, and angiogenesis, including the MAPK (mitogen-activated protein kinase)-ERK (extracellular signal-regulated kinase) pathway and PI3K (phosphoinositide 3 kinase)-AKT (protein kinase B)-mTOR (mammalian target of rapamycin) pathway. ¹³

MAPK-ERK pathway

The MAPK-ERK pathway is the canonical downstream signaling pathway of activated KRAS. Activated KRAS localizes RAF (rapidly accelerated fibrosarcoma) to the plasma membrane, resulting in RAF activation. ¹³ Then, activated RAF phosphorylates and activates MEK (mitogen-activated protein kinase kinase), which phosphorylates and activates ERK. ¹³ Activated ERK enters the nucleus and activates multiple transcription factors that control the gene expression involved in cell proliferation, migration, and angiogenesis. ¹³ Overactivation of the RAS-MAPK-ERK pathway has been involved in a variety of cancers, such as lung, colorectal, and pancreatic cancers. ¹⁴ KRAS mutations in bAVMs may be associated with abnormal AVM-driving functions, such as dysregulated angiogenesis and vascular development. Hence, elucidating the mechanisms that control KRAS signaling and the MAPK-ERK pathway is crucial for the development of targeted therapies for bAVM treatment. The RAS-MAPK-ERK pathway involved in sporadic bAVM pathogenesis, and potential therapeutic approaches for this condition are summarized in Figure 1.^{15 –17}

Figure 1.

Genetic mutations, signaling pathways, and potential therapeutic targets for sporadic brain arteriovenous malformations (bAVMs). The RAS (rat sarcoma)–MAPK (mitogen-activated protein kinase)–ERK (extracellular signal-regulated kinase) signaling pathway is involved in sporadic bAVMs. Ligand binding and activation of receptor tyrosine kinases, such as VEGFR2 (vascular endothelial growth factor receptor 2), activate Kirsten rat sarcoma (KRAS) and the MAPK–ERK signaling pathway leads to endothelial cell proliferation, cell migration, and angiogenesis. Targeted therapies for sporadic bAVMs include bevacizumab, sotorasib, dabrafenib, and trametinib. Created with BioRender.com. VEGF: vascular endothelial growth factor; RAF: rapidly accelerated fibrosarcoma; MEK: mitogen-activated protein kinase kinase.

Somatic mutations in sporadic bAVMs

Nikolaev et al. ¹⁰ were the first to identify the presence of somatic activating KRAS mutations (G12V, G12D, and Q61H) in the majority of bAVM tissue samples. Since then, several investigators have identified somatic mutations involved in the RAS-MAPK-ERK pathway in sporadic AVMs.^{18 –22} Hong et al. ¹⁸ detected activating BRAF (v-Raf murine sarcoma viral oncogene homolog B) mutations and two novel mutations in KRAS (G12A and S65 A66insDS) in spinal AVMs as well as bAVMs for the first time. Priemer et al. ²⁰ demonstrated the first reported case of a KRAS G12C mutation in a bAVM. These results suggest that somatic KRAS/BRAF mutations and the MAPK–ERK signaling pathway play a key role in the pathogenesis of central nervous system AVMs. Goss et al. ²¹ and Gao et al. ²² each investigated the association of somatic mutations including KRAS and BRAF with phenotypes of sporadic bAVMs. However, the authors did not detect significant associations between the presence of KRAS/BRAF mutations and bAVM phenotypes including age, sex, presenting symptoms, lesion location, and lesion size. Bameri et al. ²³ performed a systematic review and meta-analysis to evaluate the prevalence of KRAS/BRAF mutations in bAVMs. The prevalence of KRAS mutations among 1,726 patients with AVM in 6 studies was 55%, whereas the prevalence of BRAF mutation was 7.5%.

Animal models for sporadic bAVMs

Several clinical studies have discovered somatic KRAS/BRAF mutations in human bAVMs,^{10,18 –22} yet it remains unclear how these mutations are involved in bAVMs. Establishing a clinically relevant animal model is essential for investigating the causative role of KRAS/BRAF mutations in bAVM development and to discover a potential pharmacological intervention for human bAVMs. Mouse and zebrafish models that simulate sporadic bAVM characteristics in humans have recently been developed by engineering somatic endothelial cell (EC)-specific gain-of-function mutations in KRAS/BRAF.^{11,16,17,24,25}

Fish et al. ¹¹ developed the first KRAS-induced transgenic bAVM mouse and zebrafish models by introducing constitutively active KRAS mutations into ECs. Endothelial-specific Cre drivers were used to induce KRAS G12D expression in brain ECs in postnatal and adult mice. Almost 50% of mice in both age groups developed bAVMs by 8 weeks post-introduction. The brain vasculature of adult mice was fully formed, and bAVM formation may not necessarily require physiological angiogenesis during early development. Almost 50% of the KRAS G12V-induced transgenic embryonic zebrafish also developed AV shunts. Established AV shunts in zebrafish were reversed by pharmacological inhibition of activity of MEK but were refractory to inhibition of activity of PI3K. These findings demonstrated that KRAS-activating mutations in ECs may cause the bAVM phenotype via the MAPK–ERK, but not via the PI3K–AKT pathway.

Park et al. ²⁴ developed a KRAS-induced sporadic bAVM mouse model using an adeno-associated virus targeting brain endothelium-KRAS G12V (AAV-BR1-KRAS G12V). The mice were administered AAV-BR1-KRAS G12V via retro-orbital sinus injection. All mice treated with AAV-BR1-KRAS G12V occurred one or more bAVM by 9 weeks after induction, and this bAVM model showed high reproducibility. Remarkably, the bAVMs induced by this approach demonstrated a tangled nidus composed of dilated vessels, feeding arteries, and draining veins, simulating the morphological features of human bAVMs. The formation of KRAS G12V-induced bAVM was inhibited by trametinib, a MEK inhibitor, at the early stages of development. However, it was unknown whether trametinib may be effective for already established bAVMs.

Nguyen et al. ²⁵ tested whether trametinib could alleviate the progression of established vascular defects such as vessel dilations and focal lesion formation. By creating LoxP-STOP-LoxP-Kras (G12D);Cdh5 (PAC)-CreERT2 mice [noted as iEC-Kras (G12D∗)], KRAS activation was induced within all vascular ECs in mutant mice. As vascular defects were observed by PN day 8, trametinib treatment was started at this timepoint. The survival rate of vehicle-treated iEC-Kras (G12D∗) mice was 21.5% by PN day 14, whereas trametinib treatment improved the survival of iEC-Kras (G12D∗) mice to 76.9% at the same point. Furthermore, the vascular defects in iEC-Kras (G12D∗) mice treated with trametinib were milder than with the vehicle, although still present. Park et al. ²⁴ and Nguyen et al. ²⁵ each demonstrated that trametinib is preventative and curative for KRAS-induced AVM mouse models. These results suggest that trametinib could potentially treat bAVMs in human patients, emphasizing the necessity for clinical trials.

Fraissenon et al. ¹⁶ established KRAS G12C-Cdh5 and KRAS G12C-CAGG mice, two mouse models of KRAS G12C-driven vascular malformations to evaluate the effectiveness of sotorasib, a specific KRAS G12C inhibitor. The KRAS G12C-Cdh5 mouse model was slowly progressive, and the mortality was low, whereas the KRAS G12C-CAGG mouse model showed a more severe phenotype. The necropsy on the mutant mice revealed brain hemorrhages in all of them, along with brain vascular malformations. Histologic examination in these mouse models confirmed the presence of typical vascular malformations in the brain tissue, with tortuous arteries, loss of capillary structure, and dilated veins. Next, the authors investigated whether sotorasib could decrease the volume of vascular malformations in the two mouse models. KRAS G12C-Cdh5 and KRAS G12C-CAG mice received a daily oral dose of sotorasib 2 days after Cre induction, and the mice were treated continuously for 4 months. In a KRAS G12C-Cdh5 mouse model, sotorasib-treated mice showed no brain vascular malformations. In a KRAS G12C-CAG mouse model, all vehicle-treated mice died 4 weeks after Cre recombination. Whereas the sotorasib-treated mice remained alive with no apparent phenotype up to 70 days after tamoxifen induction. No vascular malformations were found in the brain in the sotorasib-treated mice on necropsy. These findings revealed that sotorasib has a rationale for use in patients with bAVMs with the KRAS G12C mutation.

Tu et al. ¹⁷ first developed a mouse model of sporadic bAVMs caused by somatic BRAF V600E mutations. BRAF V600E mutations in brain ECs were induced by stereotactic intracerebral injection of AAV-BR1-Cre in mice with homozygous (Braf^flox/flox) or heterozygous (Braf^flox/+) mutations. BAVMs were caused in both types of mice, yet bAVMs in mice with Braf^fl/+ mutations basically showed a course of the disease characterized by a less aggressive growth progression and lower hemorrhage rate, simulating the characteristics of the human bAVMs. Notably, this model, in which the dosage and injection site of AAV can be accurately controlled, demonstrated an advantage that can develop various bAVM phenotypes, including lesion size and location, and hemorrhage severity. The formation of BRAF V600E-induced bAVM was significantly inhibited by pharmacological BRAF inhibition at the early stages of development. However, the therapeutic efficacy was decreased in already established bAVM lesions. These results may suggest though the activation of the RAS-MAPK-ERK pathway may be essential for initial abnormal vessel formation, once arteriovenous shunts have been formed, consequent AVM growth may not necessarily depend on the activation of the RAS-MAPK-ERK pathway.

Somatic KRAS/BRAF activating mutations alone in brain ECs are sufficient to initiate the formation of bAVMs in mouse and zebrafish models, which provides supportive evidence that somatic KRAS/BRAF activating mutations play a causative role in bAVM pathology. These models simulate the pathology of human bAVMs and can become invaluable platforms for elucidating the molecular mechanisms involved in sporadic bAVM pathogenesis and developing novel targeted therapies.

Potential therapeutic targets for sporadic bAVMs

Bevacizumab, a VEGF tyrosine kinase inhibitor, was studied in a clinical trial as treatment for sporadic bAVMs, based on angiogenesis inhibition. ²⁶ This was the first phase 1 clinical trial to assess the safety and efficacy of bevacizumab in patients with sporadic bAVMs, and the primary outcome measure was the change in bAVM volume as compared to pre-treatment magnetic resonance imaging (NCT02314377). No serious adverse events, including intracranial hemorrhage, were observed, although bAVM volume did not change during the study period. Bevacizumab therapy was well tolerated, and further clinical trials are needed to evaluate dose-dependent efficacy. Phase 2/3 clinical trials to evaluate safety and efficacy of bevacizumab in patients with symptomatic bAVMs will be conducted in the future (NCT06264531).

The discovery of KRAS mutations in sporadic bAVMs suggests the potential for targeted therapies of this signaling pathway. However, KRAS is hard to focus on pharmacologically, and so most therapies for KRAS-mutant cancers have focused on its downstream targets, such as MEK.^27,28 Trametinib, an FDA-approved MEK inhibitor, ²⁹ inhibited bAVM formation and alleviated already established vascular defects in KRAS-induced bAVM mouse models.^24,25 A case report described off-label use of trametinib in patient with KRAS-positive extracranial AVM, which resulted in a substantial reduction in blood flow to the AVM after 6 months of treatment. ³⁰ NCT04258046, a phase 2 clinical trial, is currently ongoing for patients with extracranial AVMs to undergo treatment with trametinib. A pilot study (NCT06098872) is also currently investigating the use of trametinib in surgical unruptured bAVMs.

Dabrafenib, an FDA-approved BRAF inhibitor, ³¹ may be also a promising treatment for bAVMs. Dabrafenib inhibited BRAF V600E-induced bAVM formation in mice. ¹⁷ A case report described off-label use of dabrafenib and trametinib in a patient with BRAF V600E-positive extracranial AVM, which resulted in a reduction in the AVM lesion after 3 months of treatment. ³² Combination therapy of trametinib and dabrafenib has the potential to treat AVMs as well as cancers. ³³ Targeting the MAPK–ERK pathway may be a promising treatment for AVMs. However, extracranial AVMs are proliferative, but bAVMs are not, which may affect how bAVMs respond to trametinib or dabrafenib. Further studies are required to evaluate the safety and efficacy of trametinib and dabrafenib for patients with sporadic bAVMs.

Sotorasib, a specific KRAS G12C inhibitor, has been recently approved by the FDA to treat patients with advanced non-small cell lung cancer with a KRAS G12C mutation. Sotorasib was administered to two adult patients with severe KRAS G12C-related extracranial AVMs. ¹⁶ Both patients rapidly improved symptoms and reduced AVM size as well. Targeting KRAS G12C appears to be a promising therapeutic approach for patients with KRAS G12C-related AVMs. This study reveals the importance of genetic testing in vascular malformations to provide targeted therapy. In AVMs, KRAS G12C mutation was previously detected in only one case. ²⁰ The most common KRAS mutation was either KRAS G12V or KRAS G12D mutation.^{10,18 –22} No KRAS G12V and KRAS G12D inhibitors have been approved by the FDA yet, and so further studies are required. Promising novel agents for bAVMs are summarized in Table 1.^{16,17,24 –26,32,34}

Table 1.

Promising novel agents for sporadic brain arteriovenous malformations.

Agent	Target	Updated in vivo and clinical studies
Bevacizumab	VEGF	Bevacizumab attenuated VEGF-induced angiogenesis and vascular malformations in Alk12f/2f mice with Ad-Cre 34 Phase I clinical trial for treatment of bAVMs; no changes in bAVM volumes reported after 52 weeks (NCT02314377) 26 Future phase II/III clinical trial evaluating safety and efficacy in patients with symptomatic bAVMs (NCT06264531)
Trametinib	MEK	Trametinib inhibited bAVM formation and alleviated already established vascular defects in KRAS-induced mouse models24,25 Phase II clinical trial assessing treatment in complicated extracranial AVMs (NCT04258046) Phase II pilot study investigating use in surgical unruptured AVMs of the brain and body (NCT06098872)
Dabrafenib	BRAF	Dabrafenib inhibited bAVM formation in BRAF V600E-induced mouse model 17 A case report described off-label use of dabrafenib and trametinib in a patient with BRAF V600E-positive extracranial AVM, which resulted in a reduction in the AVM lesion after 3 months of treatment 32
Sotorasib	KRAS G12C	Sotorasib inhibited bAVM formation in KRAS G12C-induced mouse model 16 Sotorasib was administered to two adult patients with severe KRAS G12C-related extracranial AVMs; 16 both patients had rapidly improved symptoms and reduced AVM size as well

VEGF: vascular endothelial growth factor; Alk1: activin receptor like kinase 1; Ad-Cre: adenoviral vector expressing Cre recombinase; bAVMs: brain arteriovenous malformations; MEK: mitogen-activated protein kinase kinase; KRAS: Kirsten rat sarcoma; BRAF: v-Raf murine sarcoma viral oncogene homolog B.

Future research directions

It is controversial whether targeted therapies can reverse already established bAVMs. If the continual activity of KRAS and the MAPK-ERK signaling pathway are essential for the persistence of bAVMs like cancer, ³⁵ pharmacological inhibition of these signaling pathways may provide hope to patients suffering from sporadic bAVM. MEK inhibition could reverse already established AV shunts in a KRAS-induced zebrafish model ¹¹ and could alleviate the progression of established vascular defects and improve the survival rate in a KRAS-induced mouse model. ²⁵ These promising results need to be confirmed in further studies. Continuous research using these bAVM animal models could offer new insights into the pathophysiology of sporadic bAVMs and may contribute to the development of novel pharmacologic therapies.

Conclusion

Suitable animal models are essential for achieving better understanding of KRAS-related pathways and in preclinically testing new therapeutic approaches. Continuous research aimed at elucidating the pathogenesis of bAVMs will open new research avenues for effective bAVM therapeutics, potentially leading to improved outcomes for patients suffering from this devastating condition.