Ramı́rez-Alvarado, Marina

Abstract Background: Immunoglobulin free light chains (FLC) form the substrate for synthesis of amyloid fibrils in patients with AL amyloidosis. Development of FLC assay (Freelite) has allowed us to better assess the clonal cell burden in patients with AL amyloid. The relationship of the light chain types and their levels in serum to the clinical feature at presentation and the eventual outcome has not been systematically studied. Methods: We identified 730 patients with biopsy proven AL amyloidosis, who were seen at Mayo Clinic between January 1980 and July 2006, who were seen within 90 days of their diagnosis and in whom FLC levels were performed or had stored serum available for analysis. Cardiac biomarkers (cTnT, NT ProBNP) were also performed on stored serum for some of the patients as part of previous studies. Clinical data and follow up status are prospectively collected into the Dysproteinemia database, which was used for the study. Results: The median age of the study population was 63 years (range; 32–90) and 463 patients (63%) were male. The median estimated survival for the entire group was 37 months from diagnosis and the median follow up for the 263 patients (37%) alive at the time of the analysis was 27.8 months (range; 1–177 months). The plasma cell clone was lambda in 528 patients (72%) and kappa in 202 (28%) patients. The median absolute difference between the FLCs was 19.6 mg/dL for the entire group (range; 0.01–2478); for the kappa patients was 29.4 mg/dL (range; 0.01–1359) and for the lambda patients was 18.2 mg/dL (range; 0.03–2478). Kappa AL was associated with more involvement of the GI tract and liver as indicated by higher alkaline phosphatase (1.8 vs. 1.3 fold; P=0.03), higher total bilirubin (0.95 vs. 0.78 mg/dL, P = 0.04) and lower serum carotene (119 vs. 155 ug/dL; P = 0.0002). On the contrary, renal involvement was more in the lambda AL with the 24 hour urinary albumin higher in the lambda AL (2.4 gm vs. 1.4 gm; P = 0.0002) and lower serum albumin (2.77 vs. 2.97 mg/dL; P = 0.001). While there was no difference in the overall survival (OS) between the kappa and lambda AL, the median OS for those without a heavy chain was significantly shorter (12.6 months vs. 29.9 months; P= 0.01) compared to those with a heavy chain identified. There was also a correlation between high FLC difference and degree of organ involvement, especially cardiac. Among the 202 kappa AL patients, the median survival for those with a high FLC difference ( > 29.4 mg/dL) was 13 months vs. 48.8 months for those with low difference (P = 0.001). Similarly among the 528 patients with lambda AL, those with high difference (> 18.2 mg/dL) had a median OS of 9.3 months compared to 33 months for those with low difference (P < 0.0001) (Figure). The FLC difference was independent of the cTnT, NT Pro BNP, uric acid and fold-increase in alkaline phosphatase, all of which were significant in the multivariable analysis. Conclusions: The type of FLC impacts the spectrum of organ involvement, with gastrointestinal involvement observed more frequently with kappa, and renal involvement with lambda FLC. The FLC burden correlates with the degree of organ involvement, and is a significant predictor of overall survival in AL amyloidosis. Lambda FLC exerts similar effect as kappa FLCs, but at lower levels confirming the increased “amyloidogenicity” of lambda FLC. Finally, lack of a chaperone heavy chain is associated with a poorer outcome. Figure 1. Overall survival form diagnosis in patients with high vs. low FLC difference (using median values for cutoffs for kappa and lambda FLC) Figure 1. Overall survival form diagnosis in patients with high vs. low FLC difference (using median values for cutoffs for kappa and lambda FLC)

Introduction: The clinical presentation and frequency of organ involvement are quite distinct in patients with IgM amyloidosis compared to patients with non-IgM amyloidosis, with less heart involvement and more nerve, soft tissue and lung involvement in IgM patients. It has been observed that immunoglobulin light chain variable region (IGVL) gene and gene family of the plasma cell clone impact phenotypic manifestations in AL amyloidosis, particularly organ tropism. (Kourelis et al. Blood. 2017;129). Given the variability in clinical features amongst patients with IgM vs. non IgM AL amyloidosis, we hypothesized that IGVL gene usage would be different between these two groups. Methods: Patients with newly diagnosed IgM and non-IgM amyloidosis seen at our institution from 01/2006 to 12/2015 were identified. IGVL gene usage was assessed by liquid chromatography tandem mass spectrometry (LC/MS/MS) as previously described. (Vrana et al Blood 2009;Dasari et al J of Proteome Res; Kourelis et al. Blood. 2017). Mass spectrometry data for IGVL analysis was obtained from clinically available data. LC/MS/MS was performed on available archival specimens when such data was not available clinically. Results: Patients with newly diagnosed amyloidosis who had sufficient mass spectrometry data on fat aspirate or tissue biopsy for IGVL gene usage identification were included (IgM AL, N=44 and Non-IgM AL, N=391). Clinical Features: As expected, differences were noted in baseline characteristics of patients with IgM vs. non-IgM AL amyloidosis. Patients with IgM amyloidosis were older (68 vs. 64 years, p=0.04) and more likely to have kappa as the involved light chain (34% vs. 25%, p=0.2), though the difference did not reach statistical significance. IgM amyloidosis patients had lower difference in involved and uninvolved free light chains (dFLC) (15.5 vs. 25.8 mg/dL, p=0.01) and lower NTProBNP levels (median: 1987 vs. 2655 pg/mL, p=0.02) compared with non-IgM amyloid patients. Presence of t(11;14) was less common by FISH (30% vs. 48%, p=0.1), though the difference did not reach statistical significance. Patterns of organ involvement were different, with heart involvement being less frequent in IgM patients (64% vs. 76%), while peripheral nerve (30% vs. 18%, p=0.05), soft tissue (41% vs. 21%, p=0.005) and lung involvement (7% vs. 3%, p=0.1) were more common in patients with IgM vs. non-IgM AL amyloidosis. There was no difference in liver (9% vs. 15%), kidney (50% vs. 51%) GI (23% vs. 25%) or autonomic nervous system involvement (18% vs. 14%). IGVL Gene Usage: As shown in Figure 1, IGVL usage differed across the two groups. In the lambda group, LV2-08 (14% vs. 2%, p<0.001) and LV2-14 (36% vs. 10%, p<0.001) were more common in IgM amyloid patients, while LV1-44 (0% vs. 10%, p=0.02) and LV6-57 (2% vs. 18%, p=0.004) were less common in the IgM group. There was also a trend towards less frequent involvement of LV3-01 (2% vs. 12%, p=0.07) in IgM amyloid patients. In the kappa group, KV4-01 (11% vs. 4%, p=0.04) was more common in patients with IgM amyloidosis. Correlation Between IGVL Gene usage and Clinical Features: IGVL gene usage correlated with disease characteristics and differences in the clinical presentation of patients across the two groups. LV2-14 was more commonly seen in IgM amyloid patients (above). It has been previously associated with a higher frequency of peripheral nerve involvement (Kourelis et al. Blood 2017), which was more frequent in IgM group. LV2-14 is also associated with low dFLC levels, which were noted to be lower in the IgM AL cohort. LV1-44 has been associated with higher likelihood of cardiac involvement. LV1-44 was seen less frequently in the IgM amyloid patients, who also had lower rates of cardiac involvement. LV6-57 has been historically associated with higher rates of t(11;14). It was less common in IgM AL patients, who also had lower rates of t(11;14). Conclusions: IGVL gene usage differed significantly in patients with IgM vs. non-IgM amyloidosis, correlating with differences in disease characteristics and organ involvement. This is the first study to our knowledge to report IGVL gene usage differences in IgM vs. non-IgM AL amyloidosis. Differences in IGVL gene usage may explain the distinct clinical presentation seen in IgM amyloidosis, the biology of which has so far remained poorly understood. *SS and SD contributed equally Disclosures Dasari: The Binding Site: Patents & Royalties: US Patent Rights on Mass Spectroscopy Licensing agreement with The Binding Site, Research Funding. Dispenzieri:Alnylam: Research Funding; Celgene: Research Funding; Takeda: Research Funding; Pfizer: Research Funding; Janssen: Consultancy; Akcea: Consultancy; Intellia: Consultancy. Murray:The Binding Site: Patents & Royalties: US Patent Rights on Mass Spectroscopy Licensing agreement with The Binding Site, Research Funding. Kumar:Janssen: Consultancy, Research Funding; Celgene: Consultancy, Research Funding; Takeda: Research Funding. Gertz:Prothena: Honoraria; Alnylam: Honoraria; Ionis: Honoraria; Janssen: Honoraria; Spectrum: Honoraria, Research Funding; Celgene: Honoraria.

The transthyretin (TTR) and immunoglobulin light chain proteins do not exist in isolation in cardiac amyloid (CA) plaques. Laser microdissection followed by mass spectrometry-based proteomic (LMD-MS) typing techniques can identify hundreds of co-deposited proteins in the "nano-environment" of amyloid deposits. Some, such as the "amyloid signature" proteins (SAP, APOE, APOA4, vitronectin, clusterin), are specific for amyloid plaques across different tissues and amyloid types. However, many of the remaining proteins are also abundant in the serum or normal tissues making it difficult to discern if they are unique to the amyloid plaque or part of the normal tissue background. We included 288 patients/samples with TTR CA (of which 53 were mutated) diagnosed by endomyocardial biopsy and 5 normal control samples obtained from 3 patients during autopsy. At the time of the meeting updated analyses including 168 cardiac light chain amyloid patients and 15 controls from patients with restrictive and hypertensive cardiomyopathy will be presented. Normalized spectral counts were used as a semi-quantitative measure of abundance. Proteins were considered part of the proteome if their abundance in the amyloid plaque was increased by 1.5-fold compared to normal controls (FDR p<0.05). Of 2240 unique proteins identified across all samples, 42 were differentially expressed in amyloid plaques. These included a) signature proteins: SAP, APOE, APOA4, vitronectin, clusterin; b) serum proteins: albumin, b2-glycoprotein, 7 complement pathway proteins; c) matrix proteins: 2 collagen isoforms, fibulin 1, decorin, PCOLCE2, TIMP3; d) cytosketal and contractile proteins: 2 actin isoforms, MYL2, 8 myosin isoforms, 3 ankyrin isofroms, desmin; and "other" proteins: serpine2, MAGE-like protein-2, pleiotrophin, AMBP, SRPX and QSOX1). The signature and cytoskeletal proteins were the most abundant overall, however, proteins that were most increased compared to normal included (log2 fold change in parentheses): fibulin (24.5), PCOLCE2 (23.5), b2-glycoprotein (23.5), factor H related peptide-2 (24.2), serpine2 (24.7), mage-like protein-2 (22.8), pleiotrophin (23.8), QSOX1 (22.9), SRPX (23.5). B2 glycoprotein was more abundant in wild type TTR samples(p=0.003). APOE (p=0.002), fibulin (p=0.008) and serpine2 (p<0.0001) were more abundant in UK stage (Gilmore JD et al, Eur Heart J 2018) 2/3 patients. Unsupervised hierarchical clustering of patients by normalized protein abundance revealed two major distinct patterns of co-deposited proteins (figure). Patients with high levels of contractility proteins (myosin isoforms, desmin, actins) were characterized by low levels of most other proteins (clusters on the left). Most complement proteins co-deposited along with higher levels of signature proteins (clusters on the right). When comparing clinical and laboratory characteristics between those 2 groups of patients only cardiac strain was found to be significantly higher (worse) in patients with lower levels of contractility proteins (p=0.01). Our data shows that the TTR CA proteome includes signature, serum and cardiac tissue proteins with variable expression according to TTR mutation status and cardiac functional status, which could inform mechanisms of disease pathogenesis. Comparison with diseased controls may further clarify if this proteomic pattern is specific to CA. Figure Disclosures Dasari: The Binding Site: Patents & Royalties: US Patent Rights on Mass Spectroscopy Licensing agreement with The Binding Site, Research Funding. Dispenzieri:Akcea: Consultancy; Intellia: Consultancy; Janssen: Consultancy; Pfizer: Research Funding; Takeda: Research Funding; Celgene: Research Funding; Alnylam: Research Funding.

Abstract Proteomics with great sensitivity and specificity effectively analyzes proteins by prior tryptic digestion and subsequent analysis by LC-MS/MS of the tryptic digest. We have utilized this approach in the development of a LC-MS/MS method to characterize minimum residual disease in multiple myeloma. The abundant antibodies produced by multiple myeloma plasma cells are identical and appear as a spike (M-spike) upon protein electrophoresis. The M-spike and histological examination of the bone marrow constitute part of the myeloma diagnostic repertoire. Upon treatment, myeloma cells and the antibodies they produce are reduced in number and amount and become increasingly hard to detect. The bone marrow biopsy serves as the “gold standard” test for clinical remission. We have developed a sensitive test for the presence of the monoclonal antibody produced by the plasma cells which may serve as a substitute for invasive bone marrow biopsy. We have focused on tryptic peptides comprising the variable CDR regions of the Ig light chains that are unique to each patients antibody clone. Utilizing 2-5 µL of patient plasma/serum from the initial M-spike sample we have utilized SDS-PAGE to yield a crudely purified immunoglobulin light chain. The light chain band is isolated, reduced, alkylated and trypsin digested with subsequent LC-MS/MS analysis to identify CDR specific tryptic peptide(s). These can be identified as variable peaks (elution time and mass) in a base peak ion chromatogram where constant region peptides of either lambda or kappa light chains (clone dependent) serve as “internal standards” for identifying the CDR tryptic peptides. The peptide’s mass and its corresponding MS/MS spectra are unique to this CDR tryptic peptide and this patient’s clone. The unique diagnostic peptide is isolated in subsequent samples by immunoaffinity purification of the target kappa/lambda clone, SDS-PAGE separation and LC-MS/MS analysis of the light chain gel band. An extracted ion chromatogram is generated based upon the CDR peptides identified in the initial analysis of the M-spike sample. In patients in clinical remission the presence of a significant signal from the targeted peptides indicates that targeting a CDR peptide from the M-spike protein is more sensitive than currently available diagnostic tools including immunofixation. Comparison of this test to the gold standard bone marrow biopsy will be examined. Disclosures: Barnidge: Mayo Foundation: provisional patent application for technology, provisional patent application for technology Patents & Royalties.

Abstract Background Immunoglobulin light chain-associated amyloidosis (AL) is caused by deposition of immunoglobulin light chain molecules with unique, clonotypic variable (V) regions. Detection and identification of the V region, however, has been challenging due heterogeneity inherent in V regions. We developed a new method for detecting IGKV and IGLV region fragments in AL deposits using protein mass spectrometry and novel bioinformatics approach. We report distinct IGKV and IGLV usage in localized versus systemic AL. Methods Shotgun protein mass spectrometry on amyloid plaques was performed as previously described (Blood. 2009; 114: 4957-9). Peptide tandem mass spectra (MS/MS) were searched against a composite sequence database containing SwissProt's human complete proteome augmented with 1764 IG V region sequences obtained from ImMunoGeneTics database and a Mayo Clinic internal database. Reversed sequences were appended to the database for estimating identification false discovery rates (FDRs). Peptide identifications were processed with Scaffold software. Confident protein identifications (probability > 0.9) with at least one unique peptide identifications and five MS/MS matches were considered. Detected variable gene family (if any) with most number of peptide and spectral matches is considered to be present in the deposit. The method was first validated in 8 cases of AL amyloidosis with known IGKV and IGLV region sequences and then applied on 1238 systemic and 393 localized AL cases. Differences between groups were measured by generating hypergeometric p-values. Results In all 8 AL cases with known IGKV and IGLV region usage, gene family identified in the AL deposit via proteomics matched the gene family inferred from bone marrow plasma cells by Sanger sequencing of IGLC genes. Armed with this method, we analyzed the amyloid proteomics data from 1238 patients with systemic AL amyloidosis and 393 patients with localized AL amyloidosis. The anatomical site distribution in systemic cohort was 296 GI tract, 364 heart, 225 kidney, 81 liver and 272 fat aspirate; Localized cohort has 78 bladder, 158 lung and 157 skin cases. Figure 1 shows the normalized frequency of the variable gene families detected between the systemic and localized AL amyloidosis. For AL-kappa, KV1 was more prevalent in systemic cases when compared to localized cases (p=7.2E-12) suggesting KV1 clones as a hallmark for systemic AL-kappa amyloidosis. KV3 was more frequently seen in localized AL-kappa cases when compared to systemic AL-kappa cases (p =2.8E-11). For AL-lambda, LV1 and LV2 gene families were more prevalent in localized cases when compared to systemic cases (p= 0.039 and 2.7E-05). LV6 was more prevalent in systemic AL-lambda cases (p=1.4E-07). We also detected a higher incidence of mixed AL/AH type in localized AL (18%) when compared to systemic AL (5%; Odds Ratio=4.2673, p<0.0001). We next turned to the organ specific IGKV and IGLV gene family usage patterns in patients with localized amyloidosis. For 147 localized AL-kappa cases, KV3 was most dominant in lung (64%) and KV1 was dominant in skin (38%). For 246 localized AL-lambda cases, LV2 was most prevalent in lung (34%) and LV3 was most prevalent in skin (36%). For bladder, both LV1 and LV2 had comparable prevalence (42% and 32%). For 361 systemic AL-kappa cases, KV1 consistently ranked at the top of gene families used in GI tract (55%), heart (71%), kidney (59%), liver (75%) and fat aspirate (50%). For 877 systemic AL-lambda cases, LV2 was predominant in GI tract (30%), LV3 in heart (35%) and LV6 in kidney (33%). The prevalence of LV1, LV2 and LV3 in liver is comparable (36%, 22% and 31%). LV6 and LV3 had comparable prevalence (22% and 24%) in AL-lambda fat aspirates. Conclusion The novel proteomics method detects IGLC V family usage in large cohorts of AL patients. It identifies unique profiles in systemic and localized cases, and in different organ sites. This information will be helpful in determining systemic versus localized nature of AL amyloidosis at diagnosis and to assess risk of specific end organ involvement. We found also a strong association between IGKV and IGLV gene usage and organ involvement. Disclosures: No relevant conflicts of interest to declare.

Abstract Abstract 4886 Detection of high molecular weight light chain oligomers in urinary exosomes of patients with AL amyloidosis. Background Exosomes are microvesicles that are part of the multivesicular body (MVB) pathway. They are created by the inward budding of the cell surface membrane and contain both surface bound membrane proteins and cytosolic proteins which can be used to identify the cell of origin. Immunoglobulin light chain amyloidosis (AL) occurs as the result of amyloid formation by the misfolding of monoclonal light chains (LC) and deposition of these amyloid fibrils in various soft tissues. This reaction requires the organization of the monoclonal LC's into protofibrils which are then weave into amyloid fibrils. This study was undertaken to determine whether urinary exosomes of glomerular origin contain intermediate species of amyloid formation. Method Urine samples from patients with AL, light chain deposition disease (LCDD), multiple myeloma (MM) and monoclonal clonal gammopathy of undetermined significance (MGUS) were collected. Urinary exosomes were isolated and separated into fractions by gradient centrifugation. Western blots were performed on the urinary exosome fractions using anti-kappa or anti-lambda antibodies. Glomerular fractions were identified using antibodies directed toward podocin. Results Urine samples were collected from 5 patients with AL, 2 from LCDD, 1 from MM and 1 MGUS. On the Western blot, immunoglobulin LC were seen in all exosomal fractions in patients with AL amyloidosis, LCDD, MM but not MGUS which is similar to normal controls (not shown). In patients with AL, oligomeric species were found in the highest concentrations in fraction 4 and 5 (Figure 1). Fraction 4 and 5 were also stained for podocin, a glomerular protein (not shown). The highest molecular weight species was ∼250 kd which corresponds to a LC decamer. High molecular weight species were also identified in 1 of 2 LCDD patients corresponding to a tetramer. The band was identified in fraction 10 which had polycystin-1 expression suggesting a tubular origin. No high molecular weight LC species was found in patients with MM or MGUS. Conclusion Our study found high molecular weight LC species corresponding to the intermediates involved in protofibril formation in urinary exosomes of patients with AL. Smaller (tetramer) high molecular weight LC species were also found in a patient with LCDD but not in patients with MM and MGUS. Not only were the high molecular weight LC species found exclusively in the diseases characterized by deposition of LC aggregates, they were also found in the segments of the nephron where the deposits were expected: glomerulus for AL and tubular epithelium for LCDD. This is consistent with our current understanding of the pathogenic mechanisms of these diseases. We believe urinary exosomes are a powerful tool in the study of diseases involving self-aggregation of monoclonal proteins. It has tremendous potential in both diagnostic and scientific research in this area. Disclosures Gertz: celgene: Honoraria; millenium: Honoraria, Membership on an entity's Board of Directors or advisory committees.

Abstract Abstract 1909 AL amyloidosis is the most common type of systemic amyloidosis. It is caused by extracellular deposition of circulating immunoglobulin light chains (LC) thought to be secreted by a primarily bone marrow based clonal plasma cell proliferative disorder. Although, invariably, the type of LC secreted by the plasma cell clone is identical to the LC deposited in the amyloid plaque, a comprehensive analysis of the re-arranged clonal LC gene coding secreted LC protein, the protein sequence of the serum and urine monoclonal band and the LC protein deposited in the amyloid plaques have not been previously performed. In this study, we compared the protein sequence predicted by the clonal LC gene sequence with the protein sequence identified in the serum and urine monoclonal bands and protein plaques by mass spectrometry (MS) based proteomics. 10 cases with systemic AL amyloidosis were studied. Bone marrow samples containing the clonal plasma cells were obtained and clonal LC variable gene sequences were determined by PCR amplification, cloning and sequencing as described previously (Abraham et al, Blood 2003, 101: 3801). For protein analysis, monoclonal protein bands identified in serum and/or urine protein electrophoresis and immunofixation were isolated from archival gels. For amyloid plaque protein analysis, the plaques were microdissected from paraffin embedded tissue. To identify the protein amino acid sequence of serum and urine bands and amyloid plaques, the samples were digested into peptides by trypsin and analyzed by MS based proteomic analysis. (see Vrana et al. and Blood 2009; 114: 4957 for methods). The results for amino acid sequence predicted from the clonal LC gene sequence, determined directly by MS from the serum and/or urine monoclonal band and amyloid plaques were then aligned and compared. In all 10 cases, the LC amino acid sequences predicted by sequencing of the clonal LC gene from bone marrow plasma cells, the amino acid sequence obtained directly from the monoclonal serum and/or urine protein and the amino acid sequence obtained from the amyloid plaque were identical. In most cases, both the complete variable region and constant regions were deposited as amyloid plaques. Some amyloid plaques showed cleavage of first 16 amino acids of the variable region suggesting that this part was not required for amyloid deposition. All LC variable region gene sequences were affected by somatic hypermutation, and identical mutations were present in both the serum and urine bands and in the amyloid plaques. In conclusion, these results, for the first time, comprehensively show that the monoclonal protein secreted by the neoplastic plasma cells is the pathogenetic protein causing the amyloid deposits in AL amyloidosis. Amyloid deposits contain virtually all of the LC protein identified in the serum or urine including both variable and constant regions, suggesting that the presence of the constant region may also play a role in amyloidogenesis. The technological capability to determine the complete protein sequence of the amyloidogenic monoclonal serum and urine LC protein isolated from archived gels will create new opportunities to develop diagnostic, predictive and prognostic serum or urine based tests for amyloidosis. Disclosures: Theis: Mayo Clinic: Patents & Royalties. Vrana:Mayo Clinic: Patents & Royalties. Dogan:Mayo Clinic: Patents & Royalties.

Abstract Abstract 2797 Poster Board II-773 Background: Primary systemic amyloidosis (AL) is characterized by deposition of light chain derived amyloid fibrils in multiple organs leading to varying degree of dysfunction. Cardiac involvement is a common feature of AL, and when significant is highly predictive of poor outcome. The currently used prognostic classification is based on cardiac biomarkers troponin-T (cTnT) and N-terminal pro B-type Natriuretic peptide (NT-ProBNP) and is effective at predicting outcome in patients with AL. However, this is driven primarily by the degree of cardiac involvement and does not take into account the degree of plasma cell burden and its impact on the disease course. It is possible that some of the heterogeneity in outcome can be explained by the plasma cell characteristics rather than the degree of the end organ damage. Methods: We examined the baseline clinical and laboratory data from 2119 patients with AL who were seen at our institution with in 90 days of their diagnosis. The data were collected from a prospectively maintained data base, and additional testing was done for free light chain levels (FLC) on stored sera from patients seen before the routine introduction of FLC testing. Cox proportional hazards analysis was performed to estimate the prognostic values of different variables. Results: We first examined the impact of plasma cell clone related characteristics namely, difference between involved and uninvolved FLC (FLC-Diff), bone marrow plasma cell (BMPC) %, plasma cell labeling index (PCLI), beta 2 microglobulin (B2M), and presence of circulating plasma cells, on overall survival (OS) from diagnosis. All variables were dichotomized into high and low based on their median value (FLC-diff: 20 mg/dL, BMPC: 10%, PCLI as a continuous variable, B2M: 3 mg/dL and circulating cells: absent or present). While all were found to be prognostic for OS in univariate analysis; in a multivariate analysis incorporating a stepwise selection only B2M and FLC-diff were significant. We then used these two variables along with the cTnT and NT-ProBNP that are used in the current model to develop a staging system. The median values for all four variables were used for developing the scores. Patients were assigned a score of 1 for presence of each characteristic (FLC difference > 20 mg/dL, troponin-T > 0.02, NT-ProBNP > 1000, and B2M > 3) or 0 if the value was at or below the cutoff. The scores were added to obtain a composite prognostic score that grouped the 370 patients (who had all the variables available for analysis) into 5 groups with very different OS (4.4, 9.3, 24.4, 43, and not reached for stages 4, 3, 2, 1, and 0 respectively; Figure A). However, since the NT-ProBNP did not have independent impact on survival in a multivariable model incorporating all the four variables, we also examined a system that only included FLC-diff, cTnT and B2M. This system again allowed grouping of patients (n=450 with all three variables available) into four distinct groups with divergent outcome with OS of 4.6, 10.5, 36.8, and not reached for stages 3, 2, 1, and 0 respectively; (Figure B). Conclusion: Incorporation of plasma cell related measurements into the existing staging system using cardiac biomarkers allows better risk stratification of patients with AL amyloidosis. The system using FLC, B2M and troponin has the advantage of easily available laboratory values and can be widely adopted. Addition of NT-ProBNP into the system allows better refinement of the intermediate risk groups. This system will allow upfront risk stratification and development of risk-adapted therapies for patients with AL. Disclosures: Gertz: Genzyme: Research Funding.

Abstract Immunoglobulin (Ig) light chain (LC)-associated amyloidosis (AL) is a fatal plasma cell (PC) disorder characterized by the overproduction of Ig light chains that deposit in an abnormal conformation as amyloid throughout the body. Lambda LC are involved in amyloid deposition 2–3 times more often than kappa LC, and certain LC Ig variable genes are more frequently involved than others and influence clinical presentation and outcome. AL is a potential complication of any immunoglobulin clonal process, but is most often pathologically associated with minimal clonal expansion of PC, as seen in monoclonal gammopathy of undetermined significance. It is much less commonly observed in patients with multiple myeloma (MM). Because of this, and to our knowledge, there are no cell lines that have been established from AL patients, an experimental tool that would be of great value in studying amyloid formation and the biology of amyloid producing PC. In this study, we have established two cell lines from a 50 yr old female initially diagnosed with AL. Upon initial diagnosis, the BM aspirate consisted of 27% IgG λPC with a PC labeling index (PCLI) of 1.9% and a κ to λratio of <0.1. Amyloid was present in periosteal vessel walls and in the fat aspirate. The first cell line, ALMC-1, was established from BM mononuclear cells isolated from the diagnostic aspirate. The patient received a peripheral blood stem cell transplant (PBSCT) 2 months later, but relapsed within 100 days post-PBSCT with symptomatic myeloma. At relapse, the patient’s BM aspirate consisted of 70% λ+ PC and a PCLI of 20%; the second cell line, ALMC-2 was established from this aspirate. IgVL and IgVH analysis revealed that both cell lines expressed identical sequences and used IgVλ 6–57*01, IgλJ3*02, and IgλC3*03: both used the IgVH VH3–21 gene and the extent of somatic mutation was approximately 4%. Both cell lines produce significant FLC, and studies are currently underway to characterize in vitro amyloid production. We next used fluorescence in situ hybridization (FISH) to identify the genetic defects in this patient’s tumor population before and after transition to symptomatic MM. The initial BM aspirate at time of diagnosis of AL revealed approximately 30% of the PC had clear evidence of c-MYC gene amplification. In the subsequent draws as well as in both cell lines, 100% c-MYC amplification was identified, consistent with clonal selection. FISH analysis also revealed that 40% of cells had p53 deletion upon initial diagnosis, whereas the subsequent samples and cell lines were 100% for p53 deletion. All cells also had an IgH translocation that did not involve any commonly observed chromosome partners. Lastly, we have characterized the cytokine responsiveness of both cell lines. Although some differences are observed between ALMC-1 and ALMC-2, IL-6 and IGF-1 stimulated growth in both cell lines to varying degrees and both lines expressed autocrine IGF-I. In summary, our initial characterization of ALMC-1 and ALMC-2 predicts that these unique cell lines will prove to be an invaluable tool to better understand AL, from the combined perspectives of amyloidogenic protein structure and amyloid formation, genetics, and cell biology. Insight gained from this model system may eventually have an impact in the clinical arena by providing a better understanding of this incurable disease.

Abstract Abstract 3138 Background: Several recent reports have suggested that doxycycline can inhibit amyloid fibril formation in different in vitro and in vivo model systems. Doxycycline is the antibiotic of choice for infectious disease prophylaxis in patients with penicillin allergy, following autologous peripheral blood stem cell transplantation for various hematological malignancies. We examined the outcome of patients with light chain amyloidosis (AL) undergoing autologous stem cell transplantation to study the impact of doxycycline used as antibacterial prophylaxis. Patients and methods: We studied 455 patients who underwent autologous stem cell transplant (ASCT) for AL between March 1996 and December 2011. We excluded 44 patients who died of transplant related complications or did not survive beyond 100 days after transplant. Patient records were reviewed to ascertain the prophylactic antibiotic used. Survival was estimated using Kaplan Meier method and the survival curves were compared using log-rank test. Cox proportional hazard model was used to estimate the prognostic impact of different variables. Results: The median age of the patients was 58 years (range, 26–74); 290 (67%) were male. The median estimated follow up for the entire cohort was 73 months (95% CI; 66, 79) and 314 (69%) were alive at last follow up. Doxycycline was used a prophylaxis in 106 (23%) of the patients, with most of the remaining receiving penicillin. The usual reason for using doxycycline was a history of penicillin allergy. Hematological response to transplant occurred in 377 (87%) patients, and there was no significant difference between patients receiving doxycycline prophylaxis compared with the rest. The median OS for the entire cohort was 161 months (95% CI; 101, NR). The median OS for those receiving doxycycline prophylaxis was not reached compared with 113 months for the rest (P=0.09). Looking specifically at those patients with a hematological response, the median OS was not reached for the doxycycline group compared with 161 months for the rest (P=0.04, figure). There was no difference in the OS between the groups among those with no hematological response. The impact of the prophylaxis was independent of the number of organs involved as well as organ response to SCT. Conclusions: The results of the current study appear to demonstrate a survival advantage for patients with AL who receives doxycycline prophylaxis, in the setting of hematological response from SCT. The results are consistent with the data regarding the effect of doxycycline on amyloid fibril formation. These initial findings should form the basis of clinical exploration of doxycycline as an adjunctive therapy in patients receiving other standard therapies for AL. Such a study can further clarify if use of doxycycline for longer duration can be of additional benefit. Disclosures: Kumar: Cephalon: Research Funding; Novartis: Research Funding; Millennium: Research Funding; Celgene: Research Funding; Merck: Consultancy, Honoraria; Genzyme: Research Funding.