Molecular Biology in the Clinical Management of Musculoskeletal Neoplasia

Michael J. Monument, MD, MSc, FRCSC, FAAOS

Kevin B. Jones, MD, FAAOS

Neither of the following authors nor any immediate family member has received anything of value from or has stock or stock options held in a commercial company or institution related directly or indirectly to the subject of this chapter: Dr. Monument and Dr. Jones.

ABSTRACT

The confluence of modern molecular technologies with a growing understanding of the genetic and cellular mechanisms associated with musculoskeletal neoplasia provides an evolving framework to diagnose, delineate, and therapeutically target these conditions. Mutations, gene and protein expression changes, and gene fusions are the focus of modern histology and genomics-based diagnostics. The increasing adoption of next-generation sequencing continues to unearth novel genetic, epigenetic, and molecular fingerprints of musculoskeletal neoplasia. This has refined disease categorization; improved diagnostic accuracy; and galvanized systemic treatments that target neoplastic drivers, cellular or metabolic vulnerabilities, and the tumor microenvironment.

INTRODUCTION

The past 30 years have witnessed a steady infusion of molecular biology principles and techniques into medicine. Perhaps no other disease space has been so fully overtaken by molecular medicine as oncology. Personalized medicine, the buzz term of the past 15 years, has promoted the application of molecular biology to specific diagnostic steps and therapeutic decision making for many cancers. Molecular diagnosis has led to both splitting and lumping functions among the binned entities of neoplasia. For example, identification of many molecular subtypes within common cancers, such as breast and prostate cancer, has shifted increasing numbers of these patients into rare disease categories; this distinction has always been the difficult reality for patients with musculoskeletal neoplasms. Other molecular subtypes of cancers arising in different tissues of origin have linked disparate groups of cancers for a potential shared targeted therapy. Such has been the case with some Langerhans cell histiocytoses that share BRAF V600E mutations with lung adenocarcinomas and melanomas, all of which can benefit from the selective BRAF inhibitor, vemurafenib. Inflammatory myofibroblastic tumors (IMTs) and lung carcinomas with anaplastic lymphoma kinase (ALK) fusions are being managed with ALK inhibitors. A broad variety of cancers, including infantile fibrosarcomas (IFSs), bear neurotrophic receptor tyrosine kinase (NTRK) fusions, often experiencing complete responses to larotrectinib.

Molecular biology principles initially were applied through research to determine the drivers of musculoskeletal neoplasia. Later, certain molecular features were associated with specific diagnoses, some of which developed a depth of association that these entities became defined by their molecular diagnostic findings. Considering the approximate order in which the varied technologies were developed, the types of molecular tests that have become dominant diagnostic approaches, as well as which molecular targets for therapy have progressed to clinical strategies used in the care of musculoskeletal neoplasia, are reviewed.

DIAGNOSIS

Immunohistochemistry

The first broadly used molecular diagnostic tool applied to sarcomas was immunohistochemistry (IHC), or the identification of specific protein epitopes by antibodies that are either conjugated to chemical staining or fluorescent probes or matched with secondary antibodies that contain the marking apparatus, as described in a 2021 study.¹ Initially, IHC was broadly used for the identification of differentiation lineages of cells in tissue sections or aspirates smeared onto slides. The principal limitation of IHC is that one antibody can identify the presence, or even relative abundance, of only a single protein. Panels of IHC probes have been used for decades to screen soft-tissue tumor samples for the presence of mesenchymal lineage cells, hematopoietic cells, epithelial cells from organ-based tumors, or melanomas. For example, vimentin, a type III intermediate filament protein, is used to detect cells of mesenchymal lineage, although it is not exclusively diagnostic of sarcoma and is expressed in some carcinomas and hematopoietic lineages.² Pancytokeratin antibodies such as AE1/AE3 are used to detect cells of epithelial origin, and diffuse positivity can aid in the exclusion of sarcoma diagnosis; however, some specific sarcoma subtypes do express focal positivity of epithelial keratin markers, including synovial sarcoma, desmoplastic round cell sarcoma, Ewing sarcoma, chordoma, and myxofibrosarcoma.³

In addition, other IHCs for markers of cell differentiation have been developed to identify specific subtypes of musculoskeletal neoplasia. In chordomas, a malignancy derived from notochordal remnants, IHC detects the nuclear expression of the notochordal transcription factor, brachyury, which is positive in more than 90% of cases.⁴ Epithelial membrane antigen is frequently used to test spindle cell neoplasms in mesenchymal tissue compartments for synovial sarcoma, which typically expresses some epithelial markers, even in the monophasic histologic variants that have no discrete epithelial gland formation (termed biphasic synovial sarcoma). Markers of smooth muscle differentiation (smooth muscle actin) frequently identify leiomyosarcomas, even when other cell morphologies and tissue architecture features are less determinant. Some pleomorphic sarcomas are now subclassified as exhibiting myogenic differentiation, even with only minimal staining for one or two myogenic IHC markers (myogenin and MyoD, for example), because this group within the prior bin of undifferentiated pleomorphic sarcoma (UPS) was found to carry more aggressive behavior.⁵

Beyond markers of differentiation, some IHC markers are thought to represent specific targets of the driving genetics of the tumor, even if they are not the drivers themselves. CD99 has long been used to identify Ewing sarcoma, because of its membranous decoration pattern of the cancer cells; however, CD99 expression is also detected in synovial sarcoma, rhabdomyosarcoma, and other small round blue cell sarcomas. Others, such as NKX2.2, also highlight Ewing sarcoma cells, whereas nuclear TLE1 identifies synovial sarcoma and MUC4 identifies low-grade fibromyxoid sarcoma and sclerosing epithelioid fibrosarcoma. None of these is considered the specific driver of its respective tumor type, but IHC for the marker is relatively specific for each associated tumor type in the correct histologic context.

Other IHC molecular tests define something that is specifically lost in certain types of musculoskeletal neoplasia. SMARCB1 loss on IHC staining (often termed INI1 IHC) in the nucleus confirms the diagnosis of epithelioid sarcoma, as well as malignant rhabdoid tumors. This gene loss is the driver mutation in these tumors, but it is most often identified diagnostically via IHC, rather than sequencing or more complicated methods.⁶ Similarly, SMARCA4 (BRG1) IHC shows loss of this protein in uterine sarcomas, and SDHB IHC shows its absence in a subset of gastrointestinal stromal tumors (GISTs). Although the drivers of progression to a malignant peripheral nerve sheath tumor from a plexiform neurofibroma usually include EED1 or SUZ12 mutation or loss, what is diagnostically tested for is the loss of H3K27me3, a specific histone trimethylation mark that is the result of the activity of the complex called PRC2 that is defined by those driver genes.⁷ H3K27me3 benefits from clinically developed specific antibodies that will show its complete loss in many malignant peripheral nerve sheath tumor cells.

Some of the most recently developed diagnostic IHCs specifically identify the driver mutations, such as high expression of MDM2 identified by IHC after the gene’s amplification in atypical lipomatous tumor/well-differentiated liposarcoma (ALT/WDL), dedifferentiated liposarcomas (DDL) and low-grade intramedullary osteosarcoma. Diffuse KIT (CD117) expression is detected via IHC in approximately 95% of GISTs, and constitutive activation of the KIT receptor tyrosine kinase by activating point mutations in the C-KIT gene renders these tumors sensitive to tyrosine kinase inhibitors (TKIs) such as imatinib (Gleevec).⁸ Desmoid tumors are commonly associated with stabilizing mutations of the CTNNB1 gene resulting in strong nuclear IHC staining of β-catenin in these benign-aggressive tumors, although nuclear β-catenin IHC positivity is also observed in other mesenchymal tumors such as synovial sarcoma, solitary fibrous tumors, low-grade myofibroblastic sarcomas, and infantile sarcomas.⁹ Specific IHC antibodies have been developed for the point mutations that drive giant cell tumor (GCT) of bone (H3F3A-G34W), as described in a 2021 study,¹⁰ and chondroblastoma (H3F3B-K36M).¹¹ Application of these antibodies has substantially enhanced the diagnosis of these entities, even if they have undergone complete histologic
drift toward secondary aneurysmal bone cyst change. Although not all fusion oncoproteins that drive specific types of musculoskeletal neoplasia have fusion-specific IHC antibodies, a few do, such as the new anti-SS18-SSX antibody, which detects SS18-SSX1, SS18-SSX2, or SS18-SSX4, specific to synovial sarcoma. Other fusions are diagnosed by a particular staining pattern for an antibody against the appropriate amino or carboxy terminus of one of the parent proteins. TFE3 has strong nuclear staining almost exclusively in cells that harbor a TFE3 fusion oncoprotein, such as in alveolar soft-part sarcoma (ASPS), with the ASPSCR1-TFE3 fusion, or perivascular epithelioid cell tumors (PEComas) with SFPQ-TFE3 or DVL2-TFE3. ALK, NTRK, STAT6, or CCNB3 are identified at simply very strong patterns of IHC staining in IMTs, IFSs, solitary fibrous tumors, or BCOR-rearranged small round blue cell tumors, respectively.

Finally, IHC can be used on tissue sections, or antibody-based fluorescent probes in flow cytometry, to identify the prevalence of infiltrating immune cells or their interacting surface markers on tumors cells. Programmed cell death protein 1 (PD1) and programmed death-ligand 1 (PD-L1) prevalence by IHC has been used in many studies of immune checkpoint inhibitors to identify likely responders to these therapies.¹²

Detecting Fusion Oncogenes

Although most of the original discoveries of driver fusion oncogenes that associate with specific musculoskeletal neoplasms were made through formal cytogenetics on living tumor cells that can be synchronized for metaphase spreads and banding, such methods have proved to be very cumbersome to apply across many laboratories and require interpretation by cytogenetics experts who are not broadly available. Although formal cytogenetics certainly could be used as a diagnostic test for many of these specific fusions, it is currently only used in a few centers, mostly as a research tool.¹³ As described in a 2022 study, increased utilization of reverse transcriptase polymerase chain reaction (RT-PCR) for the detection of fusion oncogenes expanded this capacity to a few more major centers in the 2000s. However, this required careful fixation and embedding of tissues to avoid degradation of RNA, which is the least stable of the major molecular components of a cell (DNA, RNA, protein).¹⁴ Although RT-PCR is still currently used, its use is mostly limited to a few centers that have developed primer sets to detect specific fusions for which other tests are not available yet. The molecular diagnostic test type used most frequently for fusion oncogene detection is called break-apart fluorescent in situ hybridization (FISH).¹⁵ After the specific probes have been developed, this technology can be applied to fixed and embedded tissues, and requires only DNA (the most stable molecular category of biologic molecules) to remain somewhat intact, which means that specimens can be shipped to reference laboratories to have these tests performed. Break-apart FISH generates hybridization probes of two colors to decorate the genomic DNA to either side of a typical breakpoint in a parent gene for a fusion oncogene. The normal signal in a cell lacking the translocation will demonstrate two alleles in each nucleus with both colors adjacent to each other in space. If one of the alleles has been rearranged by a chromosomal translocation, the two colors will be physically separated from each other for that allele (Figure 1). This only confirms the rearrangement of one gene per probe, but this is sufficient to make the molecular diagnosis in most cases, given the histologic context ascertained before the application of a given break-apart FISH. The first of these break-apart FISH probes that were widely available tested for rearrangement of the EWSR1 locus in the Ewing sarcoma family of tumors and the DDIT3 locus (originally called CHOP) in myxoid and round cell liposarcomas, but shortly thereafter, probes were developed for the SS18 locus (originally called SYT) in synovial sarcoma (Figure 1). They have since been developed for FOXO1, FUS, USP6, and many other translocation-rearranged loci in alveolar rhabdomyosarcoma, myxoid liposarcomas (MLPSs) and low-grade fibromyxoid sarcoma, and aneurysmal bone cyst, respectively. FISH of another type (not break-apart) can be used to identify amplification of a given locus. The most commonly used probe of this variety is for the MDM2 locus, which can be amplified manifold in ALT/WDL and DDL. Low-grade central osteosarcomas are rare subtypes of osteosarcoma that have historically been challenging to distinguish from benign entities such as fibrous dysplasia. MDM2 and CDK4 gene amplifications are common in low-grade central osteosarcomas, and IHC detection of one or both of these amplified protein markers supports the diagnosis of low-grade central osteosarcomas over benign mimickers.¹⁶

FIGURE 1 Break-apart fluorescent in situ hybridization (FISH).

Magnetic resonance images of a synovial sarcoma in the proximal medial thigh of a 35-year-old woman. A, Coronal T1-weighted view. B, Short tau inversion recovery view. C, Gadolinium-enhanced fat-saturated T1-weighted view. D, Schematic representation of the chromosomal translocation between chromosomes X and 18 that produces the fusion oncogene SS18-SSX1, which is detectable by break-apart FISH probes designed with two colors flanking the translocation breakpoint in the SS18 gene. E, Image showing FISH probes need not be applied to condensed metaphase chromosomes, as the red and green probe signals will be adjacent (as well as glowing slightly yellow together in proximity) for each normal allele in any chromatin organization or lack thereof, unless the region has been rearranged, splitting the two signals to distant parts of the nucleus.

Sequencing

Commercial technologies are becoming increasingly available to sequence various constituents of the neoplastic genome. These include bulk DNA and RNA sequencing (whole exome, whole genome, RNA transcriptome) and targeted selective sequencing or profiling of DNA and RNA hot spots with custom primers and oligonucleotide hybridization probes. These technologies have been developed to identify specific genetic or molecular alterations that can improve diagnostic accuracy, detect known or novel gene fusions, identify regions of genome amplification or loss, assess epigenetic programs, and identify coding mutations that can prognosticate or influence targeted therapies.

For example, the TruSight RNA PanCancer Panel (Illumina) uses probe hybridization, PCR amplification, and RNA sequencing to detect gene fusions, single nucleotide variants, insertions/deletions, and gene expression changes from a minimum of 20 ng of RNA from FFPE tissue blocks. In a 2021 study of 138 previously diagnosed solid malignancies (mostly sarcomas) evaluated using the Illumina TruSight RNA Fusion Panel on archival FFPE tissue blocks, an additional 30 cases (22%) with gene fusions that were not detected by conventional methods were identified.¹⁷ In seven of these cases, the additional fusion information would have altered the diagnosis and clinical management. This study also identified 19 novel fusion pairs not previously described in the literature.

Multiplex fluorescent molecular barcoding (NanoString), another emerging platform, uses nucleic acid hybridization probes with specific and distinguishable fluorescent bar code tags that allow detection of multiple specific short nucleic acid sequences within a given sample. This avoids reverse transcription or nucleic acid amplification, shortening turnaround time, and mitigating issues associated with low-quality nuclei acids. These probe sets are customizable and can be strategically pooled into code sets curated for optimized investigations into tumor immunology, cell metabolism, tumor signaling programs, gene fusions, and others.
A custom sarcoma fusion probe code set has been developed and validated in a study of 212 sarcomas including 71 fusion-positive and 57 fusion-negative samples detected using routine clinical FISH or RT-PCR testing, and 84 that had not undergone any prior molecular testing.¹⁸ The NanoString RNA fusion assay confirmed fusion gene expression in 89% of cases, including 100% detection for Ewing sarcomas, synovial sarcomas, MLPSs, alveolar rhabdomyosarcomas, and high-grade endometrial stromal sarcomas. The fusion assay also detected an additional 15 fusions in tumors previously identified as negative by clinical FISH/RT-PCR testing. These cases included four cases of Ewing sarcoma, one case of mesenchymal chondrosarcoma (HEY1-NCOA2), one case of angiomatoid fibrous histiocytoma (EWSR1-CREB1), one case of nodular fasciitis (MYH9-USP6), three cases of undifferentiated round cell sarcoma (BCOR-CCNB3), one case of soft-tissue myoepithelioma (IRF2BP2–CDX1), one case of undifferentiated spindle cell sarcoma (ZC3H7B–BCOR), one case of a poorly differentiated spindle cell sarcoma (CIC–DUX4), one case of epithelioid hemangioendothelioma (YAP1–TFE3), and one case of a low-grade sarcoma, type not otherwise specified (COL1A1-PDGFB). For these 15 cases, the fusion confirmations would affect diagnostic and treatment considerations, highlighting the importance of these emerging sequencing platforms and justification to incorporate these tools into clinical practice.

THERAPEUTICS

New targeted molecular therapies can be directed against specific drivers of oncogenesis (eg, KIT, MDM2 amplification, NTRK fusions), whereas other targeted therapies are focused on the molecular vulnerabilities intrinsic to the neoplastic cells (eg, cell cycle regulation, chromatin remodeling, tumor-associated antigens [TAAs]), but not the specific driving mutation or genetic alteration. In both instances, selective pressures within neoplastic cells may result in cell escape from these targeted pathways, and thus treatment resistance. This adaptive neoplastic biology has been robustly documented using targeted therapies for driver mutations and signaling cascades in the treatment of carcinomas and melanomas (eg, epidermal growth factor receptor inhibitors in lung cancer, HER2+ breast cancers, BRAF inhibitors in lung cancers and melanoma).¹⁹^,²⁰ In GISTs, additional acquired exon mutations in the C-KIT gene have been attributed to imatinib resistance following chronic therapy.²¹ Similarly, in NTRK fusion-driven cancers (such as IFS; see later discussion), acquired resistance to first-generation NTRK inhibitors can develop following additional mutations in the kinase domain of this target.²² Thus, as the field increasingly uses many of these targeted therapies for musculoskeletal neoplasia, inhibition via single-agent therapy may not be sufficient to inhibit the oncologic process in some patients. Combination strategies or second-generation and third-generation inhibitors will likely be required to extend therapeutic responsiveness. Regardless, advances in molecular profiling of sarcomas have ushered in a new era of predicted sensitivities and vulnerabilities to targeted anticancer therapies that will become of increasing importance for patients with advanced disease or unresectable disease, or as a neoadjuvant treatment to lessen surgical morbidities.

Tyrosine Kinase Inhibitors

The very first molecularly targeted therapy for cancer, imatinib, which was developed just before the turn of the century, has been field-changing with respect to it being the first of a new class of enzymatic inhibitors focused on tyrosine kinases, many of which are involved in the signal cascades that drive proliferation of cancer cells, beginning with receptor-ligand interactions at the cell surface. Imatinib itself was rapidly adopted in the musculoskeletal neoplasia families targeting the constitutively active driver mutations in the KIT receptor tyrosine kinase in GIST²³ and later targeting the COL1A1-PDGFB fusion protein activation of the PDGFB receptor tyrosine kinase in dermatofibrosarcoma protuberans.²⁴ This has led to an explosion of similar drugs that target more specific tyrosine kinases.

One of the only targeted therapy drugs approved for the management of multiple soft-tissue sarcoma (STS) subtypes is pazopanib, a partly specific TKI that targets vascular endothelial growth factor receptors [VEGFR]1, -2, and -3; PDGFR-α and -β; and KIT. These tyrosine kinase targets are activated via mutations, gene amplifications, or gene fusions in numerous STS subtypes,²⁵^,²⁶ although they are not considered driver mutations. Results from a phase III trial demonstrated efficacy of pazopanib against synovial sarcoma and leiomyosarcoma.²⁷ Other studies suggest similar efficacy for pazopanib against angiosarcoma, ASPS, desmoid fibromatosis, desmoplastic small round cell tumors, and synovial sarcoma.²⁸^,²⁹ Other TKIs such as regorafenib (targets VEGFR-1, -2, and -3; FGFR1; PDGFR-α and -β; colony-stimulating factor 1 [CSF1] receptor; and c-KIT)³⁰ have shown encouraging efficacy against advanced STS and adult bone sarcomas.

The NTRK genes consist of the neurotrophic factor receptors TRKA, TRKB, and TRKC, which normally function in neurodevelopment but when dysregulated, can become oncogenic drivers.²² Recurrent chromosomal fusion events involving the carboxy-terminal kinase domain of TRK, and various upstream amino-terminal partners have been identified across diverse solid cancer histologies in children and adults. In the late 1990s, the
ETV6-NTRK3 gene fusion was identified in IFS,³¹ resulting in a hybrid gene encoding the helix-loop-helix protein dimerization domain of the ETV6 transcription factor fused with the kinase domain of TRKC. The ETV6-NTRK3 fusion oncoprotein results in constitutive activation of the signaling cascades normally activated by TRKC, namely, mitogen-activated protein kinase and phosphatidylinositol 3-kinase pathways. Larotrectinib is a highly selective inhibitor of all three TRK proteins and is now approved for the systemic treatment of NTRK-fusion cancers in children and adults (Figure 2). In a phase II study of 55 NTRK fusion cancers, including IFS, the objective response rate was 75%, and complete responses were observed in 13% of patients.³² In this study, all seven patients with IFS responded, with two of seven having complete responses. Furthermore, an additional 10 of 11 NTRK-fusion positive STSs also showed an objective response to treatment. The availability of TKIs that target NTRK highlights the importance of diagnostic fusion-detecting FISH and RNA sequencing technologies described earlier in this chapter.

Only gold members can continue reading. Log In or Register to continue