Jonathan A. Fletcher, MD

Investigations into the molecular alterations and signaling pathways in sarcomas have made
significant progress over the past decade. Genomic and expression-level analyses have identified
candidate genes in many different sarcomas (Table 1). A large group of these genes participate in
signal transduction pathways and represent potential sites of disease intervention with targeted
therapies. This overview will discuss signaling pathway opportunities and challenges in several
types of sarcoma (among many that could have been discussed). The examples include various
sarcomas which display aberrant tyrosine kinase pathway signaling: GIST, inflammatory
myofibroblastic tumor and dermatofibrosarcoma protuberans, and one type of sarcoma with
dysregulation of the RAS pathway: MPNST. Other examples, which will not be discussed due to
time constraints, are numerous: these include congenital fibrosarcoma (NTRK3 oncoprotein);
subsets of Ewing’s sarcoma, osteosarcoma and rhabdomyosarcoma (IGF1R pathway); and clear cell
sarcoma (MET pathway).

Gastrointestinal Stromal Tumors and KIT
Although more than 80% of inoperable GIST patients have dramatic clinical benefits from imatinib,
most of these patients will eventually progress. The imatinib-resistance mechanisms responsible for
clinical progression typically involve KIT (or PDGFRA) kinase domain mutations, and less often
involve KIT amplification, or KIT-independent pathways. Alternate small molecule KIT kinase
inhibitors have shown strong KIT inhibitory effects in vitro but only minor efficacy when used in
GIST patients with secondary imatinib-resistance. The limited clinical success of second-line KIT
kinase inhibitors might result from the heterogeneity of imatinib-resistance mutations within
individual patients. These observations suggest that alternate KIT inhibitors fail to control disease
because any one drug is unlikely to inhibit the broad variety of imatinib-resistance mutations that
are found in patients with progressing GIST. Hence, alternative, synergistic combinations of KIT
inhibitors or inhibitors of crucial KIT-depending signaling intermediates may be needed to
overcome the problem of resistance to KIT kinase inhibitors.

Mutation-driven KIT/PDGFRA autophosphorylation results in activation of various downstream
signaling proteins, including PI3-K, AKT, mTOR, MAPK, and STATs, which regulate GIST
growth and survival.

Based on mouse model and human cell line studies, the PI3-K/AKT pathway appears to be
essential to oncogenic signaling in KIT-mutant GIST, irrespective of whether the mutant KIT is
imatinib-sensitive or imatinib-resistant. The PI3-K/AKT pathway seems relevant for therapeutic
targeting in imatinib-resistant GIST, even in those tumors that have lost their usual dependence
upon mutant KIT expression. We have shown that KIT inhibition results in substantial AKT and S6
deactivation in imatinib-resistant GIST cells, whereas other downstream signaling proteins,
including 4EBP1, were unaffected. PI3-K inhibition by the pan-PI3K inhibitor LY294002 resulted
in inactivation of AKT, S6 and 4EBP1 whereas mTOR inhibition (by RAD001) inactivated S6K but
not 4EBP1: these findings may result from inactivation of PI3K-like kinases by LY294002, which
is a relatively nonspecific PI3-K inhibitor.

Additive clinical effects with imatinib-induced KIT inactivation have been seen after clinical
mTOR inhibitor in GIST, but most patients do not benefit from this combination approach, and the
reasons why some patients respond but most don’t remain to be clarified. Although several key
KIT/PDGFRA signaling intermediates have been characterized, much work remains to identify the
most advantageous intermediates for therapeutic targeting in GIST patients, and equally much work
remains to validate useful combinations and schedules of drugs that can optimally and selectively
inhibit these targets.

Inflammatory myofibroblastic tumor and ALK
Inflammatory myofibroblastic tumors (IMT) are a clinicopathologically heterogeneous group of
tumors, which – until recently – were regarded by many as nonneoplastic lesions. They arise
typically in the soft tissues, but occasionally in bone, and are composed of spindled mesenchymal
cells admixed with a striking inflammatory infiltrate, in which lymphocytes and plasma cells are
predominant. In keeping with the inflammatory component, patients can present with constitutional
signs – including fever, weight loss, and anemia – but in other cases IMT present as painless masses
or as incidental radiological findings.

The true neoplastic nature of IMT became apparent due to cytogenetic evidence of clonality in the
form of translocations and other rearrangements, some of which target the chromosome 2 short arm.
The chromosome 2 cytogenetic abnormalities, which are found in the IMT spindle cells but not in
the inflammatory component, create fusion oncogenes which encode constitutively activated forms
of the ALK receptor tyrosine kinase. The activation mechanisms in the IMT ALK oncoproteins,
similar to those in other translocation-associated tyrosine kinase oncoproteins, stems from
juxtaposition of the ALK cytoplasmic kinase domain with an ectopic, highly expressed,
oligomerizing, N-terminal domain. The latter domain, in some IMT, is contributed by one of the
tropomyosin family of proteins, but other proteins can serve this role. These findings suggest that
biologically rational therapies under development for ALK-mutant ALCL might, in some cases,
have application in the rare setting of IMT that progresses to sarcoma.

Dermatofibrosarcoma protuberans and PDGFΒ
Dermatofibrosarcoma protuberans is a low-grade spindle cell tumor which is generally
subcutaneous and which can undergo progression to high-grade sarcoma. The lower-grade forms
are primarily managed surgically, and metastases are uncommon although local recurrence is not
infrequent. Higher grade, fibrosarcomatous, forms can metastasize widely. Most
dermatofibrosarcoma protuberans contain a translocation of chromosomes 17 and 22, resulting in
oncogenic juxtaposition of the COL1A1 and platelet derived growth factor beta (PDGFΒ) genes.
The oncogenic mechanism is interesting in that the COL1A1 promoter mediates high level
expression of PDGFΒ transcripts whose coding sequence is normal. PDGFΒ serves as ligand for
both types of PDGF receptors (PDGFRΑ and PDGFRΒ), and it remains to be determined whether
PDGFRB alone, or a combination of PDGFRA and PDGFRB, communicates the oncogenic signal
in dermatofibrosarcoma protuberans. Notably, therapeutic inhibition of PDGFR by imatinib has
been accompanied by dramatic clinical responses in patients with inoperable DFSP.

Neurofibromatosis, MPNST, and RAS
Neurofibromatosis type 1 (NF1) is a common tumor predisposition syndrome affecting 1 in 3500
individuals worldwide. Patients with NF1 develop multiple benign peripheral nerve sheath tumors,

termed neurofibromas, which are often debilitating and can progress to malignant peripheral nerve
sheath tumors (MPNSTs). Neurofibromas arise in either a superficial-dermal distribution or
internally along the plexus of major nerves, and the plexiform neurofibromas are uniquely
predisposed to malignant progression. Additionally, NF1 patients are also predisposed to
developing other neoplasms including astrocytic brain tumors, juvenile chronic myelomonocytic
leukemia (JCMML), GIST, and pheochromocytoma, among others.

Pathological and molecular investigations have provided much insight into the development of
MPNST. Neurofibromas are heterogeneous lesions histologically, with an abundance of Schwann
cells and perineurial cells, and entrapped nerve fibers. In contrast, MPNSTs, which arise from a
subset of neurofibromas, are primarily composed of Schwann cells, suggesting and NF1
mutations/deletions have been identified in the Schann-cell component of neurofibromas. The
genetic basis of NF1 was revealed with the identification of the NF1 gene in 1990, in kindreds
affected with this disease. Furthermore, NF1-associated tumors were shown to contain mutations in
the remaining wild-type NF1 allele, suggesting that NF1 functions as a tumor suppressor gene.
Disruption of the normal NF1 allele, accompanied by loss of heterozygosity (LOH), was also
identified in Schwann cells from neurofibromas and MPNSTs, supporting the central role of
Schwann cells in these lesions.

The NF1 gene encodes a large protein, known as neurofibromin, with a region of similarity to the
p120 Ras GTPase Activating Protein (Ras-GAP) and the yeast Ras-GAP proteins IRA1 and IRA2.
Ras-GAP proteins attenuate signaling through the Ras pathway by stimulating the hydrolysis of
GTP in Ras-GTP complexes, producing the inactive Ras-GDP complex. Neurofibromin has been
shown to act as a Ras-GAP by interacting with Ras-GTP in vitro, and by complementation studies
of mutant yeast strains deficient in IRA proteins. Importantly, Ras-GTP levels were found to be
elevated in multiple tumor types containing NF1 mutations. Furthermore, a subset of familial NF1
mutations are point mutations that result in the loss of Ras binding or Ras-GAP activity. Finally, a
subset of JCMML patients have either NF1 mutations or oncogenic mutations of N-ras, suggesting
that dysregulation of the Ras pathway per se is an important step in the pathobiology of JCMML.
Collectively, these results suggest that aberrant signaling through the Ras pathway may be central to
the pathogenesis of NF1.

The discovery of increased Ras activation in NF1 presents several potential avenues for therapeutic
intervention in this disease, as currently there is no means to replace neurofibromin function in NF1
tumors. Active Ras-GTP complexes interact with multiple effector signaling pathways and elicit a
myriad of biological effects. The major downstream effector pathways of Ras-GTP include PI3-
kinase, Raf kinase, and Ral-GDS, and previous work has implicated these pathways in cellular
proliferation, survival and differentiation.

The first potential site of intervention considered in preclinical studies involved targeting Ras itself.
Farnesyl Protein Transferase Inhibitors (FTI’s) are peptidomimetics designed to inhibit the
farnesylation of ras and thus prevent the proper trafficking of Ras to the plasma membrane.
However, Ras proteins may also be geranylated and still be targeted to the membrane. Other anti-
ras strategies, such as ribozyme therapies, are specific for mutant ras, and although N-ras mutation
and NF1 deletion may be interchangeable in JCMML, ras mutations have not been detected in NF1.
The effector pathways of ras offer alternative targets for intervention, and inhibitors of PI3K, RAF,
mTOR and MEK may prove useful for MPNST.


The determination of the molecular mechanisms involved in different sarcomas may now have a
direct consequence in improving patient care throught the addition of molecularly-targeted
therapies. The selectivity and efficacy of small molecules such as imatinib achieves a therapeutic
index far superior to conventional chemotherapy and allows for daily oral administration in GIST.
In addition to GIST, other sarcomas, such as DFSP, can be sensitive to imatinib, or – in the example
of subsets of Ewing’s sarcoma (and osteosarcoma) to IGF1R inhibitors. Furthermore, anti-ras
therapies, or therapies directed against targets regulated by ras, including mTOR and MEK/MAPK,
are rational strategies to evaluate in NF1 and MPNST.

Signal transduction pathway inhibitors represent a significant advance in oncology. The successful
treatment of GIST and DFSP with imatinib supports the notion that therapies directed against the
molecular mechanisms involved in the initiation and maintenance of tumor cells will likely be more
effective than conventional cytotoxic agents, and provides impetus for the discovery and
development of additional agents for use in other sarcomas. Alternative methods to target signal
transduction pathways include monoclonal antibodies directed against extracellular proteins and
anti-sense and ribozyme agents that target specific mRNA species. A combination of approaches
that includes small molecules and other therapies directed against specific signaling pathways will
be useful alternatives and additions to cytotoxic chemotherapy in the future.

Table 1: Typical genetic aberrations in soft tissue and bone tumors
Histologic findings                                    Cytogenetic events                           Molecular events                   Frequency
Alveolar soft part sarcoma                             t(X;17)(p11;q25)                             ASPL-TFE3 fusion                     >90%
Aneurysmal bone cyst (extraosseous)                    16q22 and 17p13 rearrangements               USP6 fusion genes                    >50%
Angiomatoid fibrous histiocytoma                       t(12;16)(q13;p11)                            EWS-ATF1 or EWS-CREB1 fusion         >75%
                                                                                                    FUS-ATF1 fusion                      <10%
Chondromyoxid fibroma                                  Deletion of 6q                                                                    >75%
                 Skeletal                              Complex*                                                                         >75%
                 Extraskeletal myxoid                  t(9;22)(q22;q12)                             EWS-NR4A3 fusion                    >75%
                                                       t(9;17)(q22;q11)                             TAF2N-NR4A3 fusion                  <10%
                                                       t(9;15)(q22;q21)                             TCF12-NR4A3 fusion                  <10%
Clear cell sarcoma/Melanoma of soft parts              t(12;22)(q13;q12)                            EWS-ATF1 fusion                     >75%
Dermatofibrosarcoma protuberans                        Ring form of chromosomes 17 and 22           COL1A1-PDGFB fusion                 >75%
                                                       t(17;22)(q21;q13)                            COL1A1-PDGFB fusion                  10%
Desmoplastic small round cell tumor                    t(11;22)(p13;q12)                            EWS-WT1 fusion                      >75%
Endometrial stromal tumor                              t(7;17)(p15;q21)                             JAZF1-JJAZ1                          30%
Ewing's sarcoma                                        t(11;22)(q24;q12)                            EWS-FLI1 fusion                     >80%
                                                       t(21;22)(q12;q12)                            EWS-ERG fusion                      5-10%
                                                       t(2;22)(q33;q12)                             EWS-FEV fusion                       <5%
                                                       t(7;22)(p22;q12)                             EWS-ETV1 fusion                      <5%
                                                       t(17;22)(q12;q12)                            EWS-E1AF fusion                      <5%
                                                       t(16;21)(p11;q12)                            FUS-ERG fusion                       <5%
                                                       t(2;16)(q33;p11)                             FUS-FEV fusion                       <5%
                                                       inv(22)(q12q12)                              EWS-ZSG fusion                       <5%
Fibromatosis (Desmoid)                                 Trisomies 8 and 20                                                                30%
                                                       Deletion of 5q                               APC inactivation                     10%
                                                                                                    Beta catenin mutation                70%
Fibromyxoid sarcoma, low-grade                         t(7;16)(q33;p11)                             FUS-BBF2H7 fusion                    50%
Fibrosarcoma, infantile                                t(12;15)(p13;q26)                            ETV6-NTRK3 fusion                   >75%
                                                       Trisomies 8, 11, 17, and 20                                                      >75%
Gastrointestinal stromal tumor                         Monosomies 14 and 22                                                             >75%
                                                       Deletion of 1p                                                                   >25%
                                                                                                    KIT or PDGFRA mutation              >90%
Giant cell tumor
                   bone                                Telomeric associations                                                            >50%
                   tenosynovial                        Trisomies 5 and 7                                                                 >25%
                                                       t(1;2)(p13;q35)                              CSF1-COL6A3 fusion                    25%
Hibernoma                                              11q13 rearrangement                                                               >50%
Inflammatory myofibroblastic tumor                     2p23 rearrangement                           ALK fusion genes                     50%
                   Uterine                              t(12;14)(q15;q24) or Deletion of 7q         HMGIC rearrangement                  40%
                   Extrauterine                         Deletion of 1p                                                                     ?
Leiomyosarcoma                                          Deletion of 1p                                                                   >50%
Lipoblastoma                                            8q12 rearrangement or polysomy 8            PLAG1 oncogenes                      >80%
                   Typical                              12q15 rearrangement                         HMGIC rearrangement                  60%
                   Spindle cell or pleomorphic          Deletion of 13q or 16q                                                           >75%
                   Aytpical (see well-differentiated liposarcoma)
                   Chondroid                            t(11;16)(q13;p12-13)                                                               ?
                   Well-differentiated                  Ring form of chromosome 12                                                       >75%
                   Myxoid/Round cell                    t(12;16)(q13;p11)                           TLS-CHOP fusion                      >75%
                                                        t(12;22)(q13;q12)                           EWS-CHOP fusion                      <5%
                   Pleomorphic                          Complex*                                                                         90%
Malignant fibrous histiocytoma
                   Myxoid                               Ring form of chromosome 12                                                         ?
                   High-grade                           Complex*                                                                         >90%
Myxofibrosarcoma                                        (see Malignant fibrous histiocytoma)
Malignant peripheral nerve sheath tumor                 (see Schwannoma)
Mesothelioma                                            Deletion of 1p                              ?BCL10 inactivation                  >50%
                                                        Deletion of 9p                              p15 , p16 , and p19 inactivation     >75%
                                                        Deletion of 22q                             NF2 inactivation                     >50%
                                                        Deletions of 3p and 6q                                                           >50%
                   Good prognosis                       Hyperdiploid, no 1p deletion                                                      40%
                   Poor prognosis                       1p deletion                                                                       40%
                                                        Double minute chromosomes                   N-myc amplification                  >25%
Osteochondroma                                          Deletion of 8q                              EXT1 inactivation                    >25%
                   Low-grade                            Ring chromosomes                                                                 >50%
                   High-grade                           Complex*                                    Rb and p53 inactivation              >80%
Pericytoma                                              t(7;12)(p22;q13-15)                         GLI-ACTB fusion                        ?
Pigmented villonodular synovitis                        (see Giant cell tumor - tenosynovial)
Primitive neuroectodermal tumor                         (see Ewing's sarcoma)
Rhabdoid tumor                                          Deletion of 22q                             INI1 inactivation                    >90%
                   Alveolar                             t(2;13)(q35;q14)                            PAX3-FKHR fusion                     >75%
                                                        t(1;13)(p36;q14), double minutes            PAX7-FKHR fusion                    10-20%
                   Embryonal                            Trisomies 2q, 8 and 20                                                           >75%
                                                                                                    Loss of heterozygosity at 11p15      >75%
Schwannoma         Benign                              Deletion of 22q                              NF2 inactivation                     >80%
                   Malignant, low-grade                None
                   Malignant, high-grade               Complex*                                                                          >90%
Synovial sarcoma
                   Monophasic                        t(X;18)(p11;q11)                               SYT-SSX1 or SYT-SSX2 fusion          >90%
                   Biphasic                          t(X;18)(p11;q11)                               SYT-SSX1 fusion                      >90%
*Indicates presence of complicated numerical and structural chromosomal aberrations


To top