Mass Spectrometry-based Proteomic Studies of Human Anaplastic Large Cell Lymphoma*

Malignant lymphomas are a diverse group of malignant neoplasms that arise as a result of a complex interplay of multiple factors including genetic aberrations, immunosuppression, and exposure to noxious agents such as ionizing radiation and chemical agents. Anaplastic large cell lymphoma (ALCL) is an aggressive T-lineage lymphoma harboring chromosomal translocations involving the anaplastic lymphoma kinase (ALK) tyrosine kinase. The most common translocation in ALCL is the t(2;5)(p23;q35). This results in the formation of a chimeric fusion kinase, nucleophosmin/ALK. Nucleophosmin/ALK activates numerous downstream signaling pathways resulting in enhanced survival and proliferation. Using a variety of mass spectrometry-driven proteomic strategies, we have studied several aspects of the ALCL proteome. In this review, we provide a summary of mass spectrometry-based proteomic studies that expands the current understanding of the molecular pathogenesis of ALCL and provides the basis for the identification of biomarkers and targets for novel therapeutic agents.

Human malignant lymphomas are clonal proliferations of lymphocytes that arise from B-or T-cells and are heterogeneous with respect to their clinical presentation, pathogenesis, and biologic behavior. Hodgkin lymphoma and non-Hodgkin lymphoma (NHL) 1 are the two major categories of malignant lymphomas recognized (1). The overall incidence of NHL increased by about 75% between 1973 and 1991 and now accounts for ϳ54,370 new malignancies diagnosed annually in the United States (1). Important advances in the understanding of the molecular biology and immunology of the lymphocytes and related cells have culminated in the World Health Organization classification scheme that incorporates the clinical, histopathologic, immunophenotypic, and molecular genetic characteristics of the lymphomas and defines distinct pathobiologic entities.
NHLs arise as a result of multiple factors, including genetic aberrations, congenital or acquired immunosuppression, exposure to ionizing radiation, and chemical agents. In many cases, the primary abnormality is a cytogenetic aberration that is consistently associated with a particular subtype of malignant lymphoma. The cytogenetic abnormality is frequently a nonrandom chromosomal translocation that leads to juxtaposition of a proto-oncogene (e.g. c-myc) to one of the antigen receptor genes (immunoglobulin heavy chain or T-cell receptor). Alternatively chromosomal translocations may lead to the formation of fusion proteins with oncogenic activity that is often central to the tumor pathogenesis. In addition to proto-oncogene activation, other mechanisms implicated in lymphoma genesis and progression include deregulation of antiapoptosis genes (e.g. bcl-2) and tumor suppressor gene inactivation (e.g. p53).
Anaplastic large cell lymphoma (ALCL) is a distinct subtype of aggressive peripheral T-cell lymphomas harboring chromosomal translocations involving the ALK tyrosine kinase (2,3). Morphologically they are characterized by large cells with pleomorphic cytology with multilobated "horseshoe-shaped" nuclei (see Fig. 1). The t(2;5)(p23;q35) chromosomal aberration resulting in overexpression of a chimeric oncogene, nucleophosmin/anaplastic lymphoma kinase (NPM/ALK) (4,5), is the most common translocation found in these tumors (see Fig. 1). Native NPM is a nucleolar phosphoprotein whose functions include ribosomal RNA assembly, chaperone activities, and nuclease activity (6). Anaplastic lymphoma kinase (CD236) is a receptor tyrosine kinase of the insulin receptor subfamily (7,8) with homology to leukocyte tyrosine kinase. It is a highly conserved receptor tyrosine kinase and is critical for Drosophila muscle development (9,10). In humans, its expression is highly tissue-specific and is normally restricted to cells of the nervous system where it is thought to play a role in neural development (11,12) via its ligands pleiotrophin (13) and midkine (14). An N-terminal oligomerization domain within NPM facilitates the dimerization of NPM/ALK, leading to autophosphorylation and constitutive activation of the ALK tyro-sine kinase (15), which is the causative oncogene in t(2;5)positive ALCLs (8, 16) (see Fig. 1).
Advances in mass spectrometry-driven proteomics have shifted the paradigm of translational cancer research (for a review, see Lim and Elenitoba-Johnson (24)). Although the ultimate goals would be to identify biomarkers for diagnosis, prognosis, and development of novel agents for therapy, significant investments into understanding the basic protein building blocks and the proteomic circuitry at a global level need to be made. To this end, we have focused on mass spectrometry-based proteomic studies to accomplish the following: 1) global proteome analysis and annotation, 2) identification of interacting partners of ALK, 3) comprehensive analysis of deregulated signaling pathways, 4) phosphoproteome analysis, 5) quantitative profiling of proteomic consequences of targeted therapeutic agents, and 6) identification of ALK translocation partners. The approaches outlined in this review, although focused on NPM/ALK-positive ALCL, can be FIG. 1. Anaplastic large cell lymphomas are characterized by the t(2;5)(p23;q35) chromosomal translocation. A, ALCLs are composed of pleomorphic malignant lymphoma cells. B, nucleophosmin located on chromosome 5 is fused to the C-terminal end of anaplastic lymphoma kinase located on chromosome 2. The NPM portion in the fusion protein contains an oligomerization domain (OD), whereas the intracytoplasmic moiety of the fusion protein contains the tyrosine kinase domain (TKD) and tyrosine residues for recruitment of adaptor proteins. The dimerization domain within the NPM leads to the oligomerization of the fusion protein and to autoactivation of the kinase, which can be detected in the malignant cells. AD, activation domain; NLS, nuclear localization signal; TM, transmembrane domain; MB, molybdylate binding. applicable to the study of malignant lymphoma in general. This review summarizes the recently obtained mass spectrometry-based proteomic studies that have enhanced not only the understanding of the molecular pathogenesis of NPM/ALK-positive ALCLs but also identified potential novel therapeutic targets.
The signaling pathways and biological mechanisms involved in the pathogenesis of NPM/ALK-positive ALCL are not completely understood, and a comprehensive profile of its proteome has yet to be performed. Large scale profiling of the proteome of ALCL will provide the initial information required to define the proteomic building blocks and will be of utility in the discovery of pathogenetic mechanisms, diagnostic biomarkers, and treatment strategies.
We have annotated the largest database of proteins comprising the proteome of the NPM/ALK-positive ALCL (25). 2 To identify and catalog a comprehensive list of proteins expressed in the NPM/ALK-positive ALCL, lysates from the SUDHL-1 cell line were used to obtain three subcellular fractions: nuclear, cytoplasmic, and membrane. To enhance the separation of the proteome even further, the subcellular fractions were separated by size using 13% SDS-PAGE after which each of the lanes were cut into 32 slices, subjected to in-gel digestion, and analyzed by LC-MS/MS (see Fig. 3A for experimental design). The MS data were interpreted using SEQUEST TM to search the UniProt database and analyzed by ProteinProphet TM and INTERACT TM to assess error rates. This approach was simple, fast, and effective in simplification of sample complexity and allowed for the identification of proteins with molecular masses ranging from 7 to 620 kDa.
A total of 623 proteins consisting of 210 membrane, 229 cytoplasm, and 184 nuclear proteins were identified with a Յ5% error rate. Extensive annotation and PubMed investigation of a subset of 209 proteins indicated that 19.9% were reported previously to be expressed in T-cells, and 44.7% were reported to have important function in cancers. Notably only 2.4% of the proteins have been reported previously in the literature to be implicated in ALCL pathogenesis. GoMiner TM (27) protein function analysis on the ALCL proteome revealed a highly complex system and identified proteins involved in diverse cellular functions such as catalytic, enzyme, signaling, motor, translation, transcription, and structural functions. Fig.  4 summarizes the functional categories of proteins identified within the cytoplasmic, nuclear, and membrane fractions of ALCL. The results of this study demonstrate the utility of subcellular fractionation combined with one-dimensional SDS-PAGE and LC-MS/MS for the comprehensive, high throughput identification of proteins within a cellular proteome.

IDENTIFICATION OF ALK-INTERACTING PROTEINS BY TANDEM MS
The molecular mechanisms involved in NPM/ALK-mediated transformation are not fully understood. Tyrosine residues located within the cytoplasmic domain of ALK act as docking sites for recruitment of signaling molecules via the SRC homology 2 (SH2) and phosphotyrosine binding (PTB) protein motifs (8). The NPM/ALK fusion protein effects its transforming capacity via several key signaling pathways including the PI3K pathway leading to AKT activation (18) as well as activation of the JAK/STAT pathway (21,28,29). Indeed coimmunoprecipitation studies have demonstrated the interaction of ALK with PLC␥ (17), IRS-1 (30), heat shock protein 90 (Hsp90) (31), Grb, ShcC (32), JAK3 (21), JAK2 (28), STAT5 (29), and STAT3 (21,33). Because ALK plays a critical role in the activation of multiple signaling pathways, determination of its interacting proteins and those that form the complex of proteins within its "interactome" may provide novel insights as to its function as well as reveal potential targets for novel therapies.
A mass spectrometry-based proteomic strategy was used to determine the identity of proteins that interact with NPM/ ALK tyrosine kinase (see Fig. 3B) (34). The ALK interactome was enriched using a monoclonal antibody against ALK, and the components were separated by one-dimensional SDS-PAGE. Distinct bands that were present in the ALK immunocomplex but not in the IgG control immunoprecipitation were digested and subjected to LC and ESI-MS/MS.
Among the over 40 proteins identified in the ALK interactome, nine proteins had been reported previously to be important mediators of the ALK signal pathway and interacted with ALK. These include PLC␥1 (17), PI3K (18,19), JAK2 (28), JAK3 (21), STAT3 (21,33), and IRS (8). More importantly, many proteins not recognized previously to be associated with NPM/ALK but with potential NPM/ALK-interacting pro-tein domains were identified. These included adaptor molecules (suppressors of cytokine signaling, Rho-GTPase-activating protein, and RAB35), kinases (MEK kinases 1 and 4, protein kinase C, myosin light chain kinase, cyclin G-associated kinase, EphA1, EphB, JNK kinase, and mitogen-activated protein kinase 1), phosphatases (meprin, PTPK, and protein phosphatase 2 subunit), and heat shock proteins (Hsp60 precursor). A comprehensive interactome map of NPM/ALK is illustrated in Fig. 5 using MetaCore (GeneGo Inc.), an integrated systems level analysis tool for high throughput MS and genomic data. Importantly a subset of the proteins that were identified by MS was confirmed by Western blotting and reciprocal immunoprecipitation. Interestingly two of the proteins that were identified by MS but not reported in the final report due to their low scores for identification (NIPA (35) and SRC (22)) were subsequently reported by two independent studies to interact with ALK and have important biologic functions in cellular transformation in ALK-positive lymphoma. The significant role of STAT3 in the pathogenesis of ALK-mediated lymphomagenesis has been demonstrated in mice where tumor formation was inhibited by genetic ablation using antisense STAT3 oligonucleotides (36). This study demonstrates the utility of antibody immunoprecipitation and peptide identification by nanoflow ESI-LC/MS/MS for the high throughput identification of proteins within the ALK signaling complex and potential definition of its signaling pathways. Although the specific nature of the protein interactions needs to be functionally validated, these approaches represent a high throughput method for the identification of proteins (37) within a complex and provide insights into the fundamental basis of their biologic activities.
Using an affinity pulldown and MS/MS strategy we analyzed the interactome of the phosphatase and tensin homolog (PTEN) tumor suppressor, a multifunctional protein deregulated in many types of cancer (38) (Fig. 3C). Wild type PTEN cDNA was inserted into pTRC-His2 vector to create a hexahistidine (His 6 )-tagged protein that was expressed in Escherichia coli. Lysate from SUDHL-1 cells was used in pulldown assays, utilizing affinity for nickel-agarose beads. Bound proteins were eluted with imidazole, digested, and analyzed by LC-MS/MS. Acquired data were searched against the National Center for Biotechnology Information (NCBI) nr.fasta non-redundant protein database using the SEQUEST algorithm (39) and screened using INTERACT and ProteinProphet. All experiments were performed in duplicate with His 6 -lacZ serving as control.
A total of 79 proteins were identified in the wild type Although the majority of ALCLs carry chromosomal translocations involving ALK, the remaining 10 -20% do not show evidence of ALK gene deregulation. The NPM/ALK chimeric tyrosine kinase modulates key signaling pathways important for its survival including PI3K/AKT (18) and JAK/STAT (21,28,29); however, the global impact of constitutive NPM/ALK expression in tumor cells is unknown. The proteomic differences between ALK-positive and ALK-negative ALCLs can be analyzed by a variety of approaches that are complementary in their overall efficiency and utility.
Although the most widely used proteomic approach is the two-dimensional gel-based method followed by mass spectrometry (40), the recent development of LC-LC-MS/MS permits the sensitive detection of low abundance proteins, membrane proteins, and proteins with extreme pI (41)(42)(43). Furthermore the ability to perform global quantitative proteomics has been significantly enhanced by the advent of ICAT TM , which is efficient in simplifying the proteome and in combination with three-dimensional LC-MS/MS permits detection and quantification of proteins and peptides from very complex samples (44,45).
To determine the comprehensive proteomic changes that occur in response to the singular molecular abnormality of NPM/ALK expression, we have performed quantitative proteomic analysis of NPM/ALK-positive and NPM/ALK-negative lymphoid cells using cleavable ICAT and electrospray ionization tandem mass spectrometry (Fig. 3D).
NPM/ALK was ectopically expressed in Jurkat T-cells, and the proteomic expression profile was compared with that of vector-transfected cells. Equivalent quantities of total cell lysates obtained from the NPM-ALK-transfected cells and the vector control cells were labeled with either a light cleavable ICAT reagent ( 12 C 9 ) or a heavy cleavable ICAT reagent ( 13 C 9 ). The cleavable ICAT reagent contains a thiol-reactive group and biotin moiety that is separated by a linker molecule that is either isotopically labeled with nine atoms of 13 C or not labeled. After covalent labeling, the protein mixtures were combined and subjected to avidin affinity chromatography to recover ICAT-labeled proteins. Off-line fractions were collected, digested with trypsin, and analyzed by automated reverse phase nanospray LC coupled on line with ESI-MS/MS. Relative protein expression levels were determined by analysis of ion chromatograms for each of the ICAT-labeled peptides and determining the expression ratios using XPRESS TM software.
Overall 110 proteins showed a 1.5-fold or greater change in the NPM/ALK-positive cells as compared with the vector control cells. Of these, 73 proteins were up-regulated in the NPM/ALK-positive cells, whereas 37 proteins were downregulated. Proteins from nearly every cellular compartment including nucleus, cytoplasm, membrane, and secretome were identified to be differentially expressed. Proteins of both high abundance (cytoskeletal proteins such as myosin heavy polypeptide) and low abundance (transcription factors and growth factor receptors) were observed. Analysis of differentially expressed proteins revealed them to belong to a variety of functional groups. The most common functional group of proteins that were up-regulated were those that belong to kinases (14 of 79); those associated with cell adhesion, migration, and cytoskeletal function (14 of 79); and those associated with proliferation and protein synthesis (10 of 79). The others represent phosphatases (3 of 79), small GTPases (5 of 79), and proteins associated with centrosome assembly (7 of 79). Importantly many were hypothetical proteins and represent ideal targets for functional studies. As shown in Fig. 6, our quantitative proteomic studies reveal that NPM/ALK deregulates the expression of numerous proteins within previously described and novel signaling modules important for cell proliferation, survival, apoptosis evasion, and tumor dissemination. These data show that comprehensive analysis of the proteomic changes produced by a singular molecular alteration such as NPM/ALK facilitates understanding of tumor pathobiology and enables the identification of novel biomarkers and therapeutic targets.
Cussac et al. (46) utilized two-dimensional gel electrophoresis followed by MALDI-TOF MS to compare the proteome of the cytoplasmic soluble proteins of an NPM/ALK-positive cell line (SUDHL-1) with that of NPM/ALK-negative cell line (FE-PD). The spot patterns of the two-dimensional electrophoresis gels were analyzed with an image analysis software that detected 82 proteins that were differentially expressed between the two cell lines. Identification of the proteins demonstrated that they belonged to five broad functional categories including signaling, cytoskeleton and motility, chaperones, control of protein expression and degradation, and homeostasis and metabolism. A comparison of cell lines representative of two morphologic subtypes of NPM/ALK-positive ALCLs, namely the common type (SUDHL-1) and the small cell variant (COST), revealed distinct set of proteins that were selectively differentially expressed within each category. Furthermore the relative overexpression of one protein, carbonic anhydrase II, in the small cell variant was validated in tissue biopsy samples of small cell variant of ALCL. The differential proteomic analysis was extended by performing two-dimensional liquid chromatography, which allowed the separation of a larger panel of proteins and increased the number of differentially expressed proteins to 175.

PHOSPHOPROTEOME ANALYSIS OF NPM/ALK-POSITIVE LYMPHOMAS
The constitutive activation of NPM/ALK provides a critical signal in downstream signaling proteins/pathways that ultimately enhances the survival and growth of ALCLs. Analysis of the phosphotyrosine proteome in ALK-positive ALCL cells may lead to identification of phosphotyrosine-containing proteins that associate with ALK and play a role in tumor cell survival as well as potential drug targets that may reduce ALK signaling. Because of their low abundance, analysis of the phosphotyrosine proteome has been challenging. IMAC has been used to purify phosphopeptides from cell lysate (47), but this technique isolates all phosphorylated residues, including phosphoserine and phosphothreonine as well as the least abundant phosphotyrosine (48). Enrichment of the phosphotyrosine proteome with anti-phosphotyrosine antibodies is essential given the low stoichiometry of tyrosine phosphorylation and sample complexity (49). Furthermore enrichment of the phosphotyrosine proteome may enhance detection of proteins by LC-MS/MS in a more expedient and sensitive manner than IMAC (50).
In an effort to carry out global analysis of the phosphotyrosine proteins in NPM/ALK-positive ALCLs, we have optimized strategies for enrichment of the phosphotyrosine proteome of SUDHL-1 cell lines using antibody-mediated enrichment techniques. 3 We compared the phosphoproteome using immunoaffinity (IA) enrichment with that of immunoprecipitation (IP) in ALK-positive ALCL cells for sub-sequent detection by immunoblot and LC-MS/MS. Immunoprecipitation was more effective at enrichment than immunoaffinity chromatography as determined by the number of phosphotyrosine-immunoreactive bands on Western blot analysis and by subsequent LC-MS/MS. Investigation of immunoprecipitation enrichment using two commercially available antibodies demonstrated improved enrichment and detection of phosphotyrosines using 4G10 over sc-18182. Examination of IP enriched proteins by LC-MS/MS demonstrated a 2.5-fold increase in the numbers of proteins identified over IA enriched proteins. Proteins identified by LC-MS/MS in both IP and IA enriched samples were subjected to GoMiner analysis for cellular localization and molecular functions.
Alternatively peptides containing phosphotyrosines have been directly isolated from multienzyme protease-digested protein extracts with phosphotyrosine-specific antibody and identified by tandem mass spectrometry (52). Using this strategy, 278 phosphopeptides representing 180 phosphotyrosine sites were identified from the SUDHL-1 cell line. Because peptides are analyzed, this approach allows the identification of not only proteins but specific phosphorylation sites. Furthermore nine sites of tyrosine phosphorylation within ALK were identified, five of which were novel. The functional significance of these novel sites of tyrosine phosphorylation remains unknown and may represent novel SH2 or PTB sites through which adaptor proteins may interact with ALK in a phosphospecific manner and pose exciting targets for future research. The phosphorylation sites are annotated in the PhosphoSite database, which is a comprehensive resource of human and mouse in vivo phosphorylation sites (53). In addi- tion, the data have been utilized to determine the protein phosphorylation motifs of cells constitutively expressing activated NPM/ALK by analyzing the distribution of amino acid residues surrounding phosphotyrosine sites using an iterative statistical approach (54). Their analysis was able to extract four novel motif classes; this is of particular interest as they may represent candidate consensus sequences for NPM/ALK.

PROTEOMIC CONSEQUENCES OF TARGETED INHIBITION OF ABERRANT SURVIVAL SIGNALS
Comparative quantitative analysis of cells in response to selective small molecular inhibitors of deregulated signaling proteins/pathways can aid in understanding the biologic mechanisms of its action. Furthermore many small molecule inhibitors have pleiotropic effects and may inhibit multiple proteins and pathways. Many of the drugs are in Phase II/III clinical trials, and thus knowledge of the proteomic consequences of these drugs may provide important clinically relevant information. One such chemical is the benzoquinone ansamycin geldanamycin (GA) and derivative 17-allylamino-17-demethoxygeldanamycin that elicit antitumor activity in ALCL (31,55). This group of ansamycins was initially thought to target tyrosine kinases and was therefore used as tyrosine kinase inhibitors. Subsequent studies identified Hsp90 as the primary target of GA (56). Hsp90 is a highly conserved, ubiquitous molecular chaperone that is required for the stability and conformational maturation of a diverse group of client proteins, including components of signaling pathways exploited by cancer cells for survival and proliferation (57).
Hsp90 client proteins include receptor tyrosine kinases such as human epidermal growth factor receptor family kinases, breakpoint cluster region-Abelson (BCR-ABL), and NPM-ALK; cytosolic signaling proteins such as AKT, v-Raf-1 murine leukemia viral oncogene homolog 1 (RAF-1), and inhibitor of nuclear factor B kinase (IKK); and cell cycle regulators including cyclin-dependent kinase 4, polo-like kinase 1, and survivin (56 -58). Inhibition of Hsp90 by ansamycins in ALK-positive ALCL cells results in down-regulation of NPM/ ALK protein kinase activity (55) leading to cellular apoptosis (31). Clearly there are multiple cellular targets of Hsp90, yet the comprehensive effects of Hsp90 inhibition in ALK-positive ALCL cells are unknown.
To assess the quantitative proteomic changes in ALK-positive ALCL cells due to Hsp90 inhibition by GA, equal mounts of lysates from vehicle-treated and cells exposed to 10 M GA were subjected to ICAT labeling after 12 h of incubation followed by LC-MS/MS analysis of proteins. Examination of the data from 35 individual strong cationic exchange fractions resulted in a total of 2921 peptides identified that were matched to 1129 known database entries and 314 unique proteins.
The results demonstrated that GA induced G 2 /M cell cycle arrest followed by caspase-3-mediated apoptosis. Moreover from the 170 ICAT-labeled peptides, 49 proteins were upregulated 1.5-fold or greater, and 70 proteins were downregulated 1.5-fold or greater in GA-treated cells. Importantly several proteins involved in diverse cellular functions, including signal transduction, DNA metabolism, nucleic acid metabolism, protein metabolism, cell growth and maintenance, and energy pathways, were identified. Some of the down-regulated proteins are known to be involved in described signaling cascades such as JAK/STAT and MAPK (mitogen-activated protein kinase) as well as pathways previously unreported in ALK-positive ALCL including WNT, NF-B, and transforming growth factor ␤. These studies demonstrate some of the molecular mechanisms by which Hsp90 inhibition reduced viability of ALK-positive ALCL cells and illustrate the diverse proteins whose expression is changed due to GA inhibition of Hsp90.
ALK is also involved in rearrangement with other partner genes in inflammatory myofibroblastic tumors (IMTs) such as Ran-binding protein 2 (64) and TPM3 and TPM4 (65). The chimeric proteins consist of the intracytoplasmic portion of the ALK protein, which includes the catalytically active kinase domain, although in most instances the fusion partners encode ubiquitously expressed proteins with active promoters that promote elevated transcription of the ALK fusion transcript and protein.
Existing methods for the identification of chromosomal translocation partners are nucleic acid-based. These include direct cloning and sequencing following screening of cDNA libraries using probes complementary to the known translocation partner and PCR-based technologies such as rapid amplification of cDNA ends (66,67), inverse PCR (68), inverse panhandle PCR (69,70), and long distance PCR (71). The development of proteomics-based methods is desirable and would complement nucleic acid-based methods because of the ability of the peptide sequencing-based approach to distinguish fusion peptides from splice variants.
Recently we described a proteomic strategy to identify translocation partners encoding oncogenic chimeric fusion proteins using the human ALK oncogene product as a model system (72). We performed immunoaffinity enrichment of chimeric ALK fusion proteins from total cell lysates of an ALCLderived cell line and a primary tissue biopsy of IMT. These tumor samples exhibited the chromosomal aberrations t(2; 5)(p23;q35) and t(1;2)q25;2p23), respectively, by conventional cytogenetics and aberrant expression of the ALK protein by tissue immunohistochemistry. The candidate proteins were visualized by immunoblotting with anti-ALK antibodies, excised from the SDS-PAGE gel, and subjected to parallel digestion using four enzymes with different proteolytic cleavage specificities (72). For each sample, the aliquots from each specific enzyme digestion were processed and subjected to microelectrospray LC-MS/MS separately with the redundant data from each analysis superimposed to generate a composite sequence map of the chimeric fusion peptides. The combination of data from multiple digests was advantageous in that it produced simpler peptide mixtures than combinations of multiple enzyme digests in one tube. Thus, whereas individual digests yielded between 25% and 41% coverage for both NPM/ALK and TPM3/ALK, the combination of all four enzymes produced an overall coverage of 82.6 and 84.8%, respectively. Tandem mass spectrometry revealed multiple overlapping peptides, identifying the NPM and ALK proteins in the ALCL cell line and the TPM3 and ALK proteins in the biopsy specimen of the IMT. In each case, fusion peptides representing hybrid sequences from both translocation partners were identified that conclusively established the presence of aberrant fusions of NPM and TPM3 to the truncated ALK protein. The appropriate ALK chimeric fusions in both samples were confirmed by bidirectional DNA sequencing. This report demonstrates the utility of a mass spectrometrybased proteomic approach for the identification of a fusion partner of a chimeric protein, and the proteomic approach is applicable to the identification of the participating members of any fusion protein encoded from chromosomal translocation wherein one of the partners is known and an antibody is available to the known partner. CONCLUSIONS/FUTURE DIRECTIONS Over 11 variant translocations involving the ALK gene have been described in lymphomas and a subset of pediatric tumors including inflammatory myofibroblastic tumor and rhabdomyosarcoma indicating that the transforming potential of ALK tyrosine kinase is manifest in more than one cell type. Furthermore the full-length ALK receptor protein, normally limited to cells of neuronal origin, is expressed in a variety of non-hematopoietic tumors including rhabdomyosarcoma, neuroblastoma, glioblastoma, breast carcinoma, and melanomas. The signaling cascades induced by native full-length ALK receptor and variant ALK chimeric proteins commonly found in ALCLs and IMTs remain largely unknown. It will be important to characterize the specific signaling pathways that are activated by stimulation of the full-length ALK receptor. In this regard, quantitative proteomic approaches analyzing global phosphoproteomic changes will facilitate comprehensive identification of signaling pathways that are affected by fulllength ALK stimulation.
Although studies have assessed inhibition of ALK by both small interfering RNA (73), short hairpin RNA (74), tyrosine kinase inhibitors (75), and more recently small molecule inhibitors with selective activity against ALK (76,77), the proteomic effects of these agents have yet to be evaluated. By determining changes in protein expression in response to the specific inhibition of ALK, we may uncover novel proteins required for NPM/ALK survival as well as discover potential therapeutic targets that may be used in combination with ALK inhibitors. Chemical proteomic approaches using affinity columns with covalently linked drugs (78) may provide a broad view of potential substrates/targets of "specific" inhibitors of ALK.
Although the advances in biotechnology have permitted the global analysis of proteins expressed in various cellular tissues, the limited capacity to extract intact proteins has largely precluded the large scale proteomic analysis of formalin-fixed and paraffin-embedded tissues. Recent work from our laboratory (26) and others (51) has demonstrated the ability to utilize archived formalin-fixed and paraffinembedded tissues to identify hundreds to thousands of proteins from whole and laser capture microdissected tissues from as little as 10 5 cells from blocks stored for over 10 years. The ability to utilize formalin-fixed, paraffin-embedded tissue material for mass spectrometry-based proteomics will provide significant advances in analysis of human lymphoma samples for diagnostic and therapeutic protein discovery.
The most critical but challenging step in mass spectrometry-based proteomic studies for human malignant lymphoma as for other disease models remains the functional validation of proteins and signaling pathways using serum and clinical tissue samples. Recent developments in protein microarrays that allow for high throughput analysis of a nanoliter volume of serum against tens of thousands of peptides/proteins provide an important approach for identification of secreted biomarkers and potential secretomic signature of patients with ALCL. Tissue microarrays constructed from biopsy samples obtained from well characterized clinical cohorts/clinical trials are also a critical component in the translation of protein identification to validation of biomarkers for diagnosis and prognosis.