Neurofibromatosis Type 1 (NF1) is an uncommon RASopathy characterized by the occurrence of pleiotropic secondary manifestations affecting diverse organ systems [1]. NF1 is caused by heterozygous loss-of-function mutations in the NF1 gene, which encodes for the neurofibromin protein [[2], [3], [4]]. Mutations can be inherited from affected family members or can occur spontaneously as de novo mutations within the NF1 gene. The neurofibromin protein negatively regulates the activation of the RAS signaling protein, thereby limiting activation of the downstream mitogen-activated protein kinase (MAPK) pathway [5,6]. Therefore, NF1 gene mutations lead to activation of the MAPK pathway, predisposing patients to a variety of secondary manifestations that occur with varying frequencies throughout the NF1 population. Secondary manifestations may include dermal, skeletal, ocular, or oncologic presentations. The clinical diagnostic criteria for NF1 were recently revised to identify affected individuals based on the presence of one (if known family history) or more (if no family history) of these secondary manifestations [7].

Skeletal manifestations in NF1 are collectively common and may include systemic and focal differences compared to age- and sex-matched individuals without NF1. For example, children and adults with NF1 typically have lower bone density and shorter stature compared to their non-NF1 peers [[8], [9], [10], [11]]. Additionally, spine deformities may include dural ectasia and scoliosis, including early-onset scoliosis [12]. Scoliosis in children with NF1 is often classified as either non-dystrophic or dystrophic. Non-dystrophic scoliosis resembles the more common non-NF1 adolescent idiopathic scoliosis, while dystrophic scoliosis presents with a variety of radiographic atypia that contribute to more severe and progressive deformity [13]. Dystrophic scoliosis may also be associated with paraspinal neurofibromas [14]. Finally, plexiform neurofibromas, benign Schwann cell-derived tumors occurring in up to 50% of individuals with NF1 [15], may be associated with skeletal abnormalities, though a causal relationship between the tumor and bone abnormalities is unclear [16].

In the long bones, children with NF1 may present with significant dysplasia, most often affecting the tibia, which is prone to fracture [17]. While tibial dysplasia is rare, its incidence is increased in children with NF1. Treatment of tibial dysplasia focuses on preventing fractures during skeletal growth, often through bracing or, potentially, by correcting mechanical alignment with growth modulation [18,19]. Fractures of the dysplastic tibiae often result in pseudarthrosis at the fracture site, which may be treated with resection, intramedullary or external fixation, grafting, and addition of recombinant human bone morphogenetic protein 2 (rhBMP2), though the clinical efficacy of rhBMP2 remains unclear [[20], [21], [22]]. Interestingly, proximal osteotomies performed during the treatment of the fracture do not result in pseudarthrosis [23], suggesting that biologic factors contributing to pseudarthrosis are localized to the fracture site.

The purpose of this paper is to review the current state of knowledge regarding the biologic understanding of NF1-associated fracture pseudarthrosis, informed in part by analyses of genetically engineered mouse models (GEMMs) and patient-derived research specimens. Results from these studies are beginning to implicate novel therapies that, when combined with current surgical approaches, may have potential to alter the natural history of this skeletal NF1 manifestation.

Pre-clinical studies, primarily utilizing genetically engineered mouse strains, facilitate mechanistic investigations addressing key molecular and cellular questions in the field. For example, delineation of the somatic cell of origin and the associated molecular dysregulation leading to dysplasia and pseudarthrosis remains incomplete. Complete loss of the Nf1 gene in mice (Nf1^−/−) is incompatible with survival, and these knock-out mice display multiple congenital defects leading to embryonic lethality [32]. Loss of a single gene copy (Nf1^+/−), however, increases the risk of spontaneous tumor formation in aged mice [33], although mice fail to display other manifestations evident in the human condition. To circumvent the lethality inherent in Nf1^−/− mice, “Nf1 floxed” (Nf1^flox) mice were engineered [34], wherein two loxP sequences were inserted into the gene, which would then facilitate deletion of the intervening sequence (and hence gene inactivation) within cells expressing the gene encoding the Cre recombinase enzyme. The “Cre/Lox” system enables the selective deletion of the Nf1 gene in specific cells by crossing Nf1^flox mice with mice engineered to express the Cre gene in targeted cell types. In the presence of Cre expression, the Nf1^flox gene is deleted, but the Nf1 gene is expressed at normal levels (and encodes for a normal neurofibromin protein) in cells that do not express the Cre gene. The Cre/Lox system has been further developed to enable temporal control of Cre expression in mice or to induce Cre expression using viral vectors. Likewise, numerous mouse strains have been developed to express Cre in specific skeletal cell populations [35], though it is essential to recognize that Cre may also be expressed in other extraskeletal tissues, which may limit their usefulness in specific experimental contexts [36]. Here, we highlight how pre-clinical mouse studies have shaped the field's understanding of skeletal disease in NF1.

The challenge of achieving a stable and durable union for NF1 pseudarthroses has prompted the use of additional off-label therapies to promote healing. While the use of Zoledronic acid (ZA) in combination with surgical cross-union has been previously reported in children with NF1 pseudarthrosis [50], outcomes were similar to those of cross-union without ZA [51], and ZA is not widely used in children. In contrast, rhBMP2 delivered via collagen sponge is frequently used as an off-label device to promote union. The use of rhBMP2 to improve healing rates in children with NF1 pseudarthroses has not been demonstrated in a clinical trial [21]. However, multiple retrospective studies related to its use in pseudarthroses have been published, including those involving children with and without NF1 treated with various surgical techniques [[52], [53], [54], [55], [56], [57]]. A recent systematic analysis evaluated multiple factors associated with primary union and refracture for pseudarthrosis, including patients with and without NF1 [58]. Interestingly, this study found an inverse association between union and the use of BMP in patients with NF1. Taken together, the field lacks consensus on whether rhBMP2 provides a significant benefit in promoting healing in the setting of NF1 pseudarthroses.

While BMP is known to promote bone formation, in the setting of NF1, multiple factors may contribute to reduced osteogenic responses. Chief among these is the hyperactivation of MAPK signaling associated with loss of NF1, which has been demonstrated in patient fracture-derived primary cells compared to cells from the iliac crest [29,30,45]. Consistent with this, spatial transcriptomic analysis demonstrated increased MAPK activation and blunted BMP pathway activation associated with persistent fibro-cartilagenous tissue in a patient's pseudarthrosis fracture [45]. These results align with molecular evidence suggesting that fracture-derived primary cells are primed for a fibrogenic rather than osteogenic fate [29]. Taken together, analyses of patient-derived primary cells suggest inherent deficiencies in the response of NF1-deficient cells to native and/or exogenous (i.e., rhBMP2) osteogenic cues. Therefore, modulation of the underlying dysregulation (MAPK, PIK3CA-AKT-MTOR, etc.) may alter the molecular priming of NF1-deficient cells toward a more osteogenic trajectory, thereby promoting healing.

In support of this concept, MEK inhibition partially rescued fracture healing in a novel mouse model where Nf1 was removed from Prss56-expressing periosteal progenitor cell [29]. Although improvements in the rate of union were modest, it remains unclear whether the addition of rhBMP2, as used in patients, may further increase union rates in this model. While inhibiting MAPK hyperactivation with a MEK inhibitor may reverse dysregulation associated with the loss of NF1 in cells at the fracture site, this may not be sufficient to activate the intrinsic molecular pathways responsible for fracture healing absent an extrinsic osteoanabolic cue, such as rhBMP2. Consistent with this, MEK inhibitor alone or rhBMP2 alone failed to rescue osteogenic differentiation of Nf1-deficient osteochondroprogenitor cells from Col2a1-cre;Nf1^flox mice; however, the combination of a MEK inhibitor with rhBMP2 significantly rescued differentiation and mineralization in vitro [39]. Taken together, current research into the therapeutic potential of targeted therapies is likely to implicate their use in combination with existing treatment paradigms that include internal or external fixation, iliac crest grafts, and rhBMP2. Emerging research, as discussed here, aims to improve fracture healing, reduce the risk of re-fracture, and decrease the incidence of amputation in children with NF1 pseudarthrosis.

Aysha Khalid: Conceptualization, Writing – original draft, Writing – review & editing. Nandina Paria: Conceptualization, Writing – original draft, Writing – review & editing. Jonathan J. Rios: Conceptualization, Writing – original draft, Writing – review & editing.

The author(s) declare that no patient consent was necessary as no images or identifying information are included in the article.

The authors acknowledge funding from Scottish Rite for Children. The funder had no role in this manuscript but provided salary support for the authors.

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Jonathan J Rios reports a relationship with US Department of Defense that includes: funding grants. Aysha Khalid reports a relationship with US Department of Defense that includes: funding grants. Nandina Paria reports a relationship with US Department of Defense that includes: funding grants. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.