Hereditary multiple exostoses: are there new plausible treatment strategies?

Abstract

Introduction:

Hereditary multiple exostoses (HME) is a rare congenital pediatric disorder characterized by osteochondromas forming next to the growth plates in young patients. The osteochondromas cause multiple health problems that include skeletal deformities and chronic pain. Surgery is used to remove the most symptomatic osteochondromas but because of their large number, many are left in place, causing life-long problems and increasing the probability of malignant transformation. There is no other treatment to prevent or reduce osteochondromas formation at present.

Areas covered:

Recent studies reviewable through PubMed are providing new insights into cellular and molecular mechanisms of osteochondroma development. The resulting data are suggesting rational and plausible new therapeutic strategies for osteochondroma prevention some of which are being tested in HME animal models and one of which is part of a just announced clinical trial.

Expert Commentary:

This section summarizes and evaluates such strategies and points also to possible future alternatives.

Brief introduction to HME

HME, also known as Multiple Osteochondromas (MO), is a rare congenital pediatric disorder characterized by benign tumors called exostoses or osteochondromas that form next to the growth plates in ribs, long bones, vertebrae and other skeletal sites in young patients [1, 2]. The osteochondromas are initially cartilaginous but with time, undergo endochondral ossification in their proximal region and become connect to the adjacent osseous elements. Because of their large number and location, osteochondromas can cause a variety of health problems that include skeletal deformities, growth retardation, chronic pain and impingement of nerves, vessels and muscles [3]. Surgery is currently used to remove the most symptomatic osteochondromas and alleviate and correct physical problems [3]. However, because of their number and difficult anatomical location, many osteochondromas are left in place, causing life-long problems and increasing the probability of malignant transformation into chondrosarcomas [4]. At present, there is no treatment to prevent, reduce or reverse the formation of osteochondromas [5, 6].

Most HME cases are linked to heterozygous loss-of-function mutations in the tumor suppressor genes EXT1 or EXT2 that encode Golgi glycosyl-polymerases responsible for the synthesis of heparan sulfate (HS) [7, 8]. Because both allele products are needed for function, HME patients display a systemic decrease of about 50% in their HS levels [9]. HS is a component of cell surface-and matrix associated proteoglycans such as syndecans, glypicans and perlecan that regulate a multitude of developmental and physiologic processes [10]. One mechanism of action involves HS’s ability to interact with, and regulate the distribution and activities of, key signaling proteins including members of the bone morphogenetic protein (BMP), hedgehog and fibroblast growth factor (FGF) families [11, 12]. The partial systemic HS deficiency in HME patients may on its own affect certain physiologic functions such as postprandial lipid metabolism and clearance [13] and could directly or indirectly lead to other HME health complications [14], but is not sufficient to trigger osteochondroma formation. In line with Knudson’s law of tumorigenicity [15], osteochondroma initiation requires a “second hit” that would result in steeper decreases in -or even total loss of-local HS and/or changes in other mechanisms [6]. Possible second hits include loss-of-heterozygosity (LOH), aneuploidy or mutations in other genes [5, 16, 17]. Experimental studies have shown that stochastic local loss of both Ext1 alleles did in fact cause formation of multiple osteochondromas in conditional transgenic mice, while loss of one allele only did not [18, 19]. We obtained similar results in conditional mutant mice in which both Ext1 alleles were deleted in cells within perichondrium only or in growth plate plus flanking perichondrium [20, 21]. Our studies revealed for the first time that Ext1 deletion in perichondrium is sufficient to induce osteochondroma formation and that the tumors largely if not exclusively originate from progenitor cells in perichondrium itself [20]. This is at variance with previous studies concluding that the osteochondromas develop from growth plate chondrocytes (reviewed in [22]). Together, the above studies and other studies have provided new insights and a better understanding of the mechanisms regulating the induction and growth of osteochondromas in HME (schematically represented in Fig. 1) and in doing so, have suggested targets of therapeutic intervention. In this Editorial, I will highlight the most plausible and possible targets and strategies and will discuss additional, but more speculative, treatment options.

Figure 1.

Schematic illustrates multiple phenotypic changes that occur during the development of osteochondromas. (A) Formation of osteochondromas (OC, yellow) invariably occurs along the border between growth plate (cells in blue) and perichondrium (PC). The osteochondroma cells exhibit a germline heterozygous EXT1 or EXT2 mutation in addition to a “second hit” and thus are genetically different from neighboring growth plate or perichondrial cells. It is also well established that osteochondroma tissue masses contain a variable number of neighboring cells (in blue) recruited into the tumor-formation process and are thus heterogeneous in terms of cell composition. (B) The mutant phenotype of cells along the growth plate-perichondrium border is complex and involves multiple regulatory reciprocal changes that converge into promoting osteochondroma development. The decreases in EXT expression and HS levels (in combination with a second hit) would decrease the activities of such anti-chondrogenic mechanisms as FGF/ERK/MEK signaling and Noggin and Gremlin expression [76]. Concurrently, there would be reciprocal increases in pro-chondrogenic mechanisms including BMP and hedgehog signaling and heparanase expression. It is possible of course that there could be other changes providing further support to osteochondroma growth (illustrated by the question mark).

Targets and strategies for osteochondroma prevention or inhibition

Osteochondroma formation in HME is a very complex and multistep process. The tumors form exclusively next to, but never within, the growth plates, thus obeying strict pathogenic and topographical mechanisms. The osteochondromas largely originate from progenitor cells in perichondrium, implying that their onset must involve, and be dependent on, a re-programming of local perichondrial cells from their normal mesenchyme-fibroblastic phenotype to a chondrogenic cell lineage phenotype. This reprogramming must require powerful mechanisms that are ectopically activated in perichondrium and redirect the differentiation potentials and fate of resident cells. Plausible treatment strategies discussed below are largely based on these data and observations and related considerations and insights.

Canonical BMP signaling inhibitors –

One powerful mechanism possibly involved in the reprogramming of perichondrial cells from a mesenchymal to chondrogenic lineage could be the BMP signaling pathway, given its well known chondrogenic potency [23]. Indeed, studies showed that canonical BMP signaling was ectopically activated in Ext1-null and HS-poor perichondrium in HME mouse models (see Fig. 1 schematic) and was followed by osteochondroma formation and growth [20, 24]. The findings are in keeping with the fact that HS and HS-rich proteoglycans normally retrain and limit BMP signaling [25, 26], thus making this pathway to act ectopically and exuberantly during HS deficiency. The findings provided a specific therapeutic target, and Sinha et al. showed for the first time that systemic oral administration of a BMP signaling antagonist did inhibit osteochondroma formation in mouse models [24]. Data were confirmed by another group subsequently [27]. The treatment reduced osteochondroma size and volume by over 60% by 6 to 8 weeks, did so at every anatomical site studied including long bones, ribs and cranial base, and acted primarily by inhibiting chondrogenesis. The drug utilized in the study was LDN-193189 that is an inhibitor of canonical BMP signaling mediated by phosphorylated SMAD1/5/8 (pSMAD) proteins [28]. The study was and remains very promising, but the outcome needs to be considered with caution. LDN-193189 is a powerful drug, affects other signaling pathways including Akt and p38 [29], could elicit side effects, and may thus not represent a viable clinical drug candidate. Happily, several groups and industries are actively engaged in creating related drugs with far more selective and specific action [30, 31] that could have lower and perhaps minor side effects and represent safer and hopeful candidates for therapy.

Modulation of FGF/ERK/MEK signaling –

Perichondrium is endowed with several mechanisms that enable it to normally maintain its innate mesenchyme and fibroblastic phenotype [32, 33]. One of these mechanisms is the FGF/ERK/MEK signaling pathway which is strongly anti-chondrogenic [34, 35]. In an interesting study, Matsushita et al. first showed that conditional ablation of ERK1 and ERK2 in mice caused ectopic chondrogenesis and chondroma-like cartilage formation in perichondrium [36]. This was accompanied by reduced expression of β-catenin which itself is part of anti-chondrogenic mechanisms and whose ablation causes ectopic cartilage formation in perichondrium as well [37, 38]. A similar development of ectopic cartilage and osteochondroma-like masses was observed after conditional ablation of FGFR3 in limb mesenchyme [39] and conditional ERK1/ERK2 deletion in CD4-expressing cells [40]. These interesting data related quite well to our observations in conditional Ext1-null mice showing that phosphorylated ERK1/ERK2 (pERK1/ERK2) levels were markedly reduced in perichondrium prior to ectopic cartilage and osteochondroma formation [24]. Because HS is normally required for FGF/ERK signaling [41], its deficiency may have directly inhibited that pathway, thus favoring osteochondroma initiation. Is this signaling pathway amenable to therapeutic manipulations and would its stimulation prevent osteochondroma formation? This possibility has not been tested yet. Certain recombinant FGF proteins are used or being tested in the clinics for other conditions [42]. There are also peptides and small compounds such as Ceramide C6 that stimulate FGF signaling and pERK1/2 activity [43, 44]. Potentially then, the FGF/ERK/MEK pathway could be targeted, but given the multiple roles it plays, a major challenge would be to elicit selective effects on osteochondroma formation while minimizing side effects.

Hedgehog signaling inhibitors –

Another powerful pro-chondrogenic mechanism is the hedgehog signaling pathway that plays critical roles in skeletogenesis [45]. This pathway is active in the growth plate, regulates the rates of chondrocyte proliferation and maturation, and also induces bone collar formation around the hypertrophic zone [45, 46]. HS and HS-rich proteoglycans control and restrict the range and boundaries of action by this pathway within the growth plate, a reflection of tight regulatory interactions of HS with hedgehog proteins via their HS-binding domains [47]. It is of interest that in the FGFR3-deficient mice described above, osteochondroma-like tissue development was preceded by up-regulation of Indian hedgehog (Ihh) gene expression and signaling [39], and the same was observed in Kif3a-deficient mice [48] in which the ectopic presence of Ihh in perichondrium was linked, and probably caused by, concurrent decreases in HS proteoglycan gene expression. Importantly, osteochondroma formation in the FGFR3-deficient mice was reduced by systemic treatment with GDC-0449 [39], a drug that inhibits the activity of the hedgehog transducing receptor Smoothened [49]. Mundy et al. showed that treatment with the hedgehog inhibitor HhAntag blocked both basal chondrogenesis as well as chondrogenesis that had been stimulated by HS depletion in vitro [50]. Together, these and other studies point to the possibility that the hedgehog pathway may be a reasonable target of therapeutic intervention in HME, and it is worth noting that hedgehog antagonists are already in clinical use and testing for other cancer conditions with acceptable safety profiles [51].

Heparanase roles and targeting –

Heparanase is an extracellular multifunctional protein that has roles in physiologic and pathologic process and represents the only entity of its kind encoded by the mammalian genome [52, 53]. When released and active, the enzyme can cleave the HS chains present in cell surface and matrix proteoglycans and in so doing, is found to release HS-bound cytokines, growth factors and signaling proteins, enhancing their bioavailability and effects on target cells [54]. About 10 years ago, Trebicz-Geffen and colleagues reported that heparanase was up-regulated at the protein and RNA levels in osteochondromas from HME patients [55]. On the surface, this finding was very surprising and counter-intuitive since the osteochondromas are already HS-deficient. On the other hand, it agreed with studies in other tumors indicating that heparanase expression was inversely correlated to EXT expression [56, 57]. In HME, the enzyme could play a role in further decreasing local HS levels and may even represent a “second hit” in osteochondroma pathogenesis. In a recent study, Huegel et al. confirmed that heparanase is broadly and abundantly distributed in human osteochondromas; they found also that heparanase expression was up-regulated during chondrogenesis in vitro and even more so when chondrogenesis was further stimulated by Ext1 ablation [58]. Thus, the authors tested the effects of the powerful heparanase inhibitor SST0001 (Roneparastat), a modified heparin chain without anticoagulant activity [59], and found that SST0001 treatment markedly suppressed chondrogenesis in vitro [58]. While the drug was not tested in HME mouse models, it was previously shown to inhibit myeloma and sarcoma growth in vivo [59, 60], thus representing a possible HME therapeutic. One caveat is that because of its large size and its heparin backbone, the drug may not easily reach and affect every target tissue in vivo and could also have side effects [58]. Nonetheless, heparanase remains an attractive and promising therapeutic target in HME, one that merits testing with current or future selective drugs.

HS replacement –

Given that poor HS synthesis and low HS levels characterize HME, one may surmise that the HS levels could be replenished by systemic administration of exogenous HS. Though seemingly simple and doable, this proposition is actually difficult and quite challenging. The HS chains do not normally occur as free polysaccharides but, as pointed out above, are always covalently linked to the proteoglycan core protein [10]. There is a considerable number of HS-rich proteoglycans that are diverse and are expressed in distinct manners, and the HS chains themselves have significant structural diversity in different tissues [12]. Thus, these various features and patterns would be difficult to reproduce by systemic HS administration, but such treatment strategy cannot be discounted entirely at this point. Esko and colleagues recently described preliminary studies in HME mice treated with certain doses of low molecular weight heparin (to reduce possible interference with coagulation) [61]. While they did not observe major effects on osteochondroma formation in those experiments, they did not test modified heparins without anticoagulant properties such as Roneparastat that could be given at higher doses. An alternative would be to use synthetic HS oligosaccharides prepared with novel chemoenzymatic methods that combine and exploit glycosyltransferases, epimerase and sulfotransferases [62]. The resulting oliosaccharide libraries in this and other studies are being used to define the sugar sequences with preferential interactions with given protein ligands and characterize the resulting antagonistic or agonistic effects [62, 63, 64]. It is conceivable that these approaches could eventually identify the proper combination of oligosaccharides needed to contain and restrain the aberrant behavior of signaling and growth factor proteins inciting osteochondroma formation.

Retinoids as antichondrogenic antagonists –

The putative treatment strategies above are directed at specific pathways and mechanisms that appear to be at the base of osteochondroma formation. A final strategy to be considered here is one that is directed against the initial chondrogenic step in osteochondroma formation, regardless of inciting chondrogenic factors. The retinoid signaling pathway may offer one such strategy. This pathway involves active retinoid ligands that are produced in the cytoplasm of given cells and are transported to the nucleus and activate the transcriptional activity of retinoid acid receptors (RARα, RARβ and RARγ) [65, 66]. In the absence of ligand production, the unliganded RARs still interact with target genes but exert transcriptional repressor activity. Importantly, Underhill and colleagues established several years ago that chondrogenesis normally requires a drop in retinoid ligand production and unliganded RAR repressor function [67, 68]. Indeed, treatment with active retinoids such as all-trans-retinoic acid powerfully suppresses chondrogenesis in vitro [69] and treatment with synthetic retinoid agonists such as Palovarotene inhibits ectopic chondrogenesis and endochondral bone formation in animal models of the severe congenital pediatric disorder Fibrodysplasia Ossificans Progressiva [70], leading to an ongoing phase 2 clinical trial for FOP sponsored by Clementia Pharmaceuticals. Those studies raised the possibility at the time that the retinoid agonists could block ectopic cartilage and bone formation in other diseases such as HME and Ossification of the Posterior Longitudinal Ligament (OPLL). Palovarotene treatment was recently shown to block the excessive level of chondrogenesis occurring in HS-deficient cultures [71] and a concurrent study found that the drug reduced ectopic cartilage formation in a mouse model of HME [72]. It is thus possible that this retinoid agonist approach, by circumventing the specific regulatory changes incited by HS deficiency, may turn out to be an effective treatment for osteochondroma prevention in HME. Happily, Clementia Pharmaceuticals has just announced a phase 2 clinical trial of Palovarotene in HME patients (ClinicalTrials.gov identifier NCT03442985).

Expert Opinion

In their insightful review articles, Esko and colleagues provided thorough and thoughtful descriptions and assessments of the multiple complex roles and mechanisms of action that HS and HS-rich macromolecules normally exert, and concluded that “heparan sulfate proteoglycans fine-tune mammalian physiology” [10, 12]. This is not an overstatement, and there is indeed overwhelming and clear evidence accumulated over the years that HS chains and their proteoglycans are in fact critical in the regulation of a broad spectrum of developmental, homeostatic and modulatory processes. The deficiency of HS in HME needs to be placed within this frame of reference and as a consequence, it is not surprising that a host of physiologic changes occur in HME patients, some subtle and some more severe [6]. It is also not surprising that multiple and reciprocal phenotypic changes occur in cells along the perichondrium-growth plate border that incite osteochondroma initiation and propel their growth (Fig. 1). Given such complexities, we could surmise that the most effective way to treat HME would be to rectify the HS deficiency and overcome all the physiologic changes caused by it. This could theoretically be achieved by CRISPR/Cas9 to correct EXT point mutations that most HME patients have, but this approach is not efficient and safe enough at present (for example, see [73]). An alternative would be to identify a drug that over-stimulates the expression of the normal EXT allele, compensating for the activity of the heterozygous mutant allele most HME patients have [61]. Such a drug would be particularly effective to ameliorate or even reverse the physiologic changes resulting from the partial systemic HS deficiency seen in HME patients such as lipid metabolism [13]. Such putative therapeutic strategy is certainly attractive, but remains a goal for the future. In the meantime then, the various and specific therapeutic targets described above remain reasonable and implementable goals for the present and should be –and are being-pursued.

Osteochondromas represent the most severe trait of the disease and as indicated above, their induction requires a second hit in addition to heterozygous germline mutation in EXT1 or EXT2 present in most HME patients. There is evidence that LOH may be one such second hit [16], but it has not been universally observed in osteochondromas [74]. Mutations in other genes also lead to the development of osteochondroma-like structures in mutant mice that in addition to ERK1/2 and FGFR3, include NFATc1 and NFATc2 [75]. At the moment, there is no evidence that these genes are mutated in HME patients and may serve as second hits, but they deserve close consideration and attention, providing insights into genetic interactions and pathogenic cross-talk in osteochondroma development and identifying additional therapeutic targets. For instance, if FGF/ERK signaling and activity were downregulated during human osteochondroma development as they are in mouse models, it would then make sense to target such pathway singly or in combination therapy. At this regard, heparanase has already been shown to be up-regulated in the osteochondromas of HME patients [55], and could certainly contribute to local HS loss and tumor induction and represent a second hit on its own [58]. It will thus be very important to directly test the role of this extracellular enzyme in osteochondroma formation. Because heparanase is the focus of research activity in several fields, various means to block it are being developed and could be eventually used in HME as well. Lastly, it is important to underline the first ever phase 2 clinical trial for HME mentioned above that will test the effectiveness of Palovarotene against osteochondroma formation. The trial provides concrete evidence that progress is being made toward the creation and testing of pharmacological treatments for this pediatric disorder that continues to affect and afflict many thousands of patients and their families worldwide.

Article Highlights.

• Hereditary Multiple Exostoses is a rare pediatric disorder characterized by growth plate-associated osteochondromas that cause a number of health problems

• HME is linked to EXT mutations and ensuing deficiency in heparan sulfate (HS), a component of cell surface and matrix proteoglycans that regulates many fundamental processes

• This editorial summarizes and analyzes recent studies on pathogenic changes caused by the HS deficiency and leading to osteochondroma formation

• Those studies suggest plausible new and specific treatment strategies by which osteochondroma formation could be prevented or reduced