WO2011015602A2 - Lung cancer biomarkers - Google Patents

Abstract

The present invention relates to compositions and methods for detecting, managing or monitoring cancer. The invention also relates to antibodies specific for cancer markers, compositions and chips containing the same, as well as their uses for cancer detection, managing, monitoring, imaging or treatment, as well as for drug development. The invention is particularly suited for detecting, managing or monitoring lung cancer in human subjects.

Description

Lung cancer biomarkers

The present invention relates to compositions and methods for detecting, managing or monitoring cancer. The invention also relates to antibodies specific for cancer markers, compositions and chips containing the same, as well as their uses for cancer detection, managing, monitoring, imaging or treatment, as well as for drug development. The invention is particularly suited for detecting, managing or monitoring lung cancer in human subjects. Background

Lung cancer is the most common cause of death from cancer and each year 1.4 million new cases are diagnosed worldwide. More than two-thirds of lung cancers are diagnosed at a late stage, when clinical symptoms appear. The overall survival rate after diagnosis ranges from 14 % in the USA to 1.1 % in some regions of Asia and is currently very low due to the late diagnosis of the disease. Current diagnostics methods are based on imaging techniques and invasive procedures such as bronchoscopy or biopsy ². The few known plasma biomarkers for lung cancer, such as carcinoembryionic antigen (CEA), cytokeratin- 19, squamous cell carcinoma antigen (SCC) and neuron-specific enolase (NSE) lack sufficient sensitivity and specificity ³ to be used as early diagnostics tools. Ongoing efforts using high throughput discovery technologies are still struggling to provide reliable and easily accessible lung cancer biomarkers to enter the clinic '⁵.

Global proteome analysis has been hampered by a variety of methodological problems including the fact that current mass spectrometry based proteome profiling technologies are in general far from covering the necessary dynamic range; are not adequately sensitive and lack sufficient reproducibility and throughput . Polyclonal and monoclonal antibodies

(mAbs) provide appropriately validated tools for the characterization and quantitative analysis of proteins but have primarily been targeted at a set of epitopes with low complexity ' . Global, antibody proteomics approaches aim to generate libraries of antibodies to cover most or all individual proteins and their immunogenic epitops in any complex proteome. Recombinant phage, bacterial and viral display represent approaches to global antibody generation but have had limited success in resulting of libraries capable to detect complex proteomes with sufficient quality affinity reagents and coverage^{9 10} . The Human Protein Atlas and other large initiatives such as the NCI Clinical Proteomics Technology Initiative ¹² are targeted at the generation of comprehensive libraries to the far more complex human proteome with both approaches using recombinant proteins as immunogens. The problem with recombinant proteins is that they do not represent the protein natural state, lack post-translational modifications and correct folding, therefore limiting the potential of the obtained libraries to profile natural proteomes.

Summary of the invention

An object of the invention relates to particular antibodies, fragments or derivatives thereof, which bind cancer biomarkers, particularly lung cancer biomarkers. These antibodies, either alone or in combination, can be used to detect, manage or monitor cancer in a subject, particularly lung cancer. The invention also relates to kits or devices containing such antibodies, suitable for immunologic detection or reaction from any biological sample.

The invention also relates to nucleic acids, vectors or cells, including hybridomas or recombinant cells, which produce antibodies of the invention.

The invention is particularly suited to detect or monitor lung cancer in human subjects.

Legend to the figures

Figure 1. A: The major steps of the monoclonal antibody proteomics process. B

Application of the mAb proteomics for biomarker discovery in lung cancer. Pooled plasma from a lung cancer cohort (collection I; see Table 1) was processed to remove the most abundant proteins and then either the remaining proteins or the glycosylated fraction was used for normalization of protein representational differences. The glycosylated and the normalized fractions were used to immunize mice for the generation of a total of 3848 nascent hybridomas. Of these, 1051 contained sufficient concentration of IgG (>50 ng/ml) in the cell supernatant to warrant high throughput ELISA screening against pooled biotinylated plasma tracers from the cancer cohort and from matching controls (collection I). From these, 181 hybridomas were identified as candidates by showing a ratio higher than 1.5 between the two pools. These candidates were then tested with individual biotinylated depleted plasmas from a second small cohort (collection II) and the obtained candidates were further qualified on two larger cohorts (collection III and IV) of lung cancer samples and controls including non-related cancers and other lung diseases. These qualification steps led to the identification of 13 cloned hybridoma candidates suitable for further development of diagnostics tools.

Figure 2. Normalisation of plasma proteins. (A) The number of identified peptides for each protein identified in a shotgun proteomics experiment was plotted against the reported protein concentration in human plasma from the literature for depleted plasma (■) and three different conditions of normalization: (A T *) - loading of 1 mg protein at a flow lml/min (B) Complexity of protein fractions obtained from plasma shown by SDS-PAGE: total plasma (lane 1); depleted plasma using the affinity column Hu7 (lane 2) and Y12 (lane 4); the normalized plasma (lane 3) and glycosylated proteins enriched fraction fromY12 depleted plasma (lane 5).

Figure 3. Screening for lung cancer specific antibodies. A. Results from HT ELISA of the generated hybridomas with pooled plasma from lung cancer patients (y axis) are plotted against the results obtained with pooled plasma from the matched controls (collection I). Hybridomas considered as primary hits are shown as large circles; colour coding is according to the fusion performed to obtain the hybridomas. B. Results from the individual screening of one of the validated hits with plasmas from collection III. The results are plotted on the y axis according to their disease status (Control vs Lung Cancer) on the x axis and the descriptive statistics of each group are shown as box-plots on the right side of the group. C. Statistical analysis of the screening results of the validated hits. The p-value from each analysis Colour code: black - comparison between LC and control groups; red - comparison between men and women in the control population; blue - comparison between COPD patients, control group and other non-cancerous pulmonary diseases; dark green - comparison between controls and other non-lung cancers; green comparison between smokers in the control group (triangle) and in the LC group (inverted triangle). Open symbols are for measurements performed with purified IgG; closed symbols are for measurements performed with IgG containing supernatant from a monoclonal hybridoma line.

Figure 4. Receiver operating curves analysis for a single antibody (panel A) and the best panel determined from the data generated with 61 hybridoma supematants (panel B). The ROC for the panel was determined either on the testing data set only (50 % of all samples) shown as straight line, or for the entire dataset (including the training seta) shown as dotted line. The area under the curve which is a measure for the diagnostics capacity is shown for each curve.

Figure 5. High-throughput screening with nascent hybridomas. A. The same pool of depleted plasma from healthy donors was biotinylated and a HT-ELISA screening experiment with a panel of 88 mAb hybridoma supematants was performed 6 times on the same day and on 6 different days. The average and standard deviation from the results of the six experiments performed on the same day are plotted on the Y axis versus the average and standard deviations from the six experiments performed over 6 days (X-axis). B. The results of the screening experiment performed with two different purifications of the same antibody each spotted in four adjacent wells of the plate with 581 tracers (two tracers per plate) prepared from individual plasma samples. C. Plasma sample was split in two parts and two individual tracers were independently prepared (depletion and biotinylation) and tested with a panel of 88 mAb hybridoma supematants. Figure 6. Raw and transformed data from a HT-ELISA screening experiment. All screening data from an experiment performed with plasma collection III and 61 hybridoma supernatants (primary candidates) are plotted as the averaged Vmax (panel A) for each reaction. The results are presented per plate (in this case each plate is reacted with one plasma tracer); the positive controls are shown as red squares and the negative controls as blue squares. Panel B: The results after normalization using positive and negative controls as described in Methods section "Screening data analysis".

Figure 7. Algorithm used for building predictive model and determining best panel of biomarker able to discriminate lung cancer and control plasma.

Figure 8: Heavy chain V-region sequence of mABs. Bsi0033: SEQ ID NO: 77; Bsi0068:

SEQ ID NO: 78; Bsi0070: SEQ ID NO: 79; Bsi0072: SEQ ID NO: 80; Bsi0071: SEQ ID

NO: 81; Bsi0076: SEQ ID NO: 82; Bsi0077: SEQ ID NO: 83; Bsi0080: SEQ ID NO: 84;

Bsi0270: SEQ ID NO: 85; BsiO272: SEQ ID NO: 86; BsiO349: SEQ ID NO: 87; BsiO351 : SEQ ID NO: 88; BsiO352: SEQ ID NO: 89; BsiO358: SEQ ID NO: 90; BsiO359: SEQ ID

NO: 91; Bsi0392: SEQ ID NO: 92.

Figure 9. BSI0392 mAB. Relative individual (•) and descriptive statistics (box plots) of levels of cognate antigen recognized by mAB BSI0392 in the plasma of control subjects and LC patients (A). SDS PAGE (reduced) followed by Western blotting (B).

Figure 10. BSI0352 mAB. Relative individual (•) and descriptive statistics (box plots) of levels of cognate antigen recognized by mAB BSI0352 in the plasma of control subjects and LC patients.

Figure 11. BSI0351 mAB. Relative individual (•) and descriptive statistics (box plots) of levels of cognate antigen recognized by mAB BSI0351 in the plasma of control subjects and LC patients. Figure 12. BSI0358 mAB. Relative individual (•) and descriptive statistics (box plots) of levels of cognate antigen recognized by mAB BSI0358 in the plasma of control subjects and LC patients . SDS PAGE (reduced) followed by Western blotting of normal control and lung cancer plasma samples (B). Immunohistological staining of paraffin embedded formaldehyde fixed tissue samples after antigen retrieval. Inserts are from HE stained tissue (2Ox primary objectives). Adenocarcinoma (upper) and squamous cell carcinoma (lower) panel. (C).

Figure 13. BSI0359 mAB. Relative individual (•) and descriptive statistics (box plots) of levels of cognate antigen recognized by mAB BSI0359 in the plasma of control subjects and LC patients (A). Two dimensional dot plot of relative cognate antigen (Alpha- 1- antichymotrypsin) levels recognized be mAB BSI0358 and BSI0359 in control subjects (red dots) and in lung cancer patients (blue dots). (B). Figure 14. BSI0271 mAB. Relative individual (•) and descriptive statistics (box plots) of levels of cognate antigen recognized by mAB BSI0271 in the plasma of control subjects and LC patients (A). SDS PAGE (reduced) followed by Western blotting of total and abundant protein depleted plasma (B) Immunohistological staining of paraffin embedded formaldehyde fixed tissue samples after antigen retrieval. Adenocarcinoma from the lung, 4Ox primary objective (C).

Figure 15. BSI0033 mAB. Relative individual (•) and descriptive statistics (box plots) of levels of cognate antigen recognized by mAB BSI0033 in the plasma of control subjects and LC patients (A). SDS PAGE (reduced) followed by Western blotting of a non reduced plasma sample. Note the haptoglobin specific ladder, which is due to polymerization of haptoglobin in the plasma (B).

Figure 16. BSI0071 mAB. Relative individual (•) and descriptive statistics (box plots) of levels of cognate antigen recognized by mAB BSI0071 in the plasma of control subjects and LC patients (A). Figure 17. Histogram distribution of accuracy (AUC from ROC curves). Data from 408 patients (189 Ctrl & 219 LC) and 10 mAbs used in the analysis. LDA function with Leave- one-out cross validation. All possible combinations (1013) of the 10 mABs were tested. Figure 18. Frequency of individual accuracies above 0.8 for each mABs (data from the comparisons shown on Figure 17.

Figure 19. ROC curve and formula classifiers for the panel of BSI0272, BSI0358 and BSI0392 mABs. Accuracy (AUC) is 0.89. Two thresholds; (i) with 0.876 sensitivity and 0.71 specificity and (ii) with 0.78 sensitivity and 0.88 specificity are marked.

Figure 20. ROC curves from of exemplary monoclonal antibody panels numbers correspond to Table 5.

Detailed description of the invention

A challenge in the treatment of lung cancer is the lack of early, pre-symptomatic detection as lung cancer symptoms generally present at advanced stages. Here, we describe the discovery of cancer specific biomarkers. We produced and characterized nascent monoclonal antibody libraries directed to the natural form of protein antigens present in the plasma of lung cancer patients. Differential plasma profiling of normal and cancer plasma proteomes of four clinical cohorts, totaling 370 patients with lung cancer and 146 controls via high throughput ELISA screening, identified several lung cancer specific (p<0.01) monoclonal antibodies. Further tests with plasma from patients with other inflammatory lung diseases and non-related cancers confirmed the lung cancer specificity of the monoclonal antibodies. The cognate antigen was identified for these antibodies, as well as binding peptides thereof. These antibodies were tested, alone and in combinations, on clinical samples, and their ability to discriminate cancer and control subjects was confirmed. These antibodies, as well as their manufacture, devices and uses represent specific objects of the present invention.

We have used mAB proteomics (see Fig. IA) to address the need for better plasma biomarkers for lung cancer. We have started with multiple-step plasma processing, aimed at depleting the most abundant proteins and accessing low-abundant proteins to use as complex antigen (Fig. IB). The most abundant proteins, representing -90% (w/w) of the pooled plasma from twenty lung cancer patients (collection I - Table 1), were removed by means of commercially available immunoaffinity depletion column (see Methods section). Two different fractionation strategies were then used to obtain protein fractions of high complexity. Glycosylated proteins, known as a source of potential cancer biomarkers ^H, were separated using multi-lectin affinity chromatography ¹⁵. Both the depleted plasma and glycosylated protein enriched fractions were then subject to immuno-affinity chromatography based normalization (see Methods). This key step is necessary to reduce representational differences among the remaining plasma proteins and to maximize the number of protein species reaching the immunogenic threshold during immunization. Shotgun mass spectrometry coupled to spectral counting analysis indicated that normalization using moderate loading flow (see Methods) on the immunoaffinity column adequately reduced representational differences (Fig. 2a). Consequently, the apparent complexity of the normalised samples, measured as the number of visible bands and/or smears on a SDS-PAGE increased as compared to normal and depleted plasma samples (Fig. 2b). The final yield after normalisation represented 1.2 - 2.5% (w/w) of the starting plasma protein input thus increasing the probability of low and medium abundant proteins to stimulate immune response. The obtained protein fractions were used to immunize two groups of Balb/c mice using standard protocols.

Hybridomas were obtained by PEG mediated fusion of spleen cells from the immunized mice with the Sp2/0-Ag-14 mouse myeloma cell line . Fused cells were seeded in microwells and their IgG production was estimated as the IgG concentration measured in the cell supernatants. IgG containing wells with at least 50 ng/ml IgG (27% of all fused cells) were then tested with biotinylated plasma tracers using direct ELISA in a series of screening experiments (Fig. IB) designed to identify antibodies capable of detecting lung cancer specific biomarkers. The screening experiments were based on direct ELISA using kinetic reading and the reproducibility of our assay (see Figure 5) allowed us to measure differences in the concentration of the antibody-antigen complex as low as 1.5. In the range of antigen concentrations below the saturation of the available antibody, the observed differences in the majority of the cases would be directly reflecting a difference in the biomarker concentration. The first three steps of the screening strategy aimed at identifying the antibodies with a good potential to discriminate cancer patients from control subjects. In the initial step, all IgG producing hybridomas from the two generated libraries were screened with pooled depleted and biotinylated plasma from twenty, apparently healthy control subjects and from twenty non-small cell lung-cancer patients (collection I) (Fig 3a). The plasma pooling was necessary to reduce the need for biological material (hybridoma supernatants and quantity of plasma) and to reduce the biological variability at this step ⁷. The 184 nascent hybridoma supernatants selected to detect a differential ratio between the two pools higher than 1.5 were in their majority (87%) cancer specific. The next steps of screening were aimed at better qualifying the capacity of the selected antibodies to distinguish plasma from lung cancer patients from control subjects. This required screening experiments using tracers obtained from the individual plasmas. The plasma processing (depletion of highly abundant proteins and labelling with biotin) is highly reproducible (Figure 5C) and allowed us to compare signals obtained with the antibody candidates on hundreds of plasmas. Parallel to the qualification process we cloned the selected antibodies, therefore some of the screening steps included different number of supernatants from hybridoma mixes, from cloned cell lines and also purified IgG (Fig. IB). The three successive screenings allowed us to select 16 candidates which showed statistically significant (p-value estimated from Mann-Whitney U test < 0.05) remarkable discrimination between the groups of control and lung cancer patients (Table 2). Biological specificity of the markers was tested by assaying plasma samples from patients with other than lung cancer disease indications (collection IV). We have included patients with inflammatory lung diseases such as COPD (42), pneumonia (23), fibrosis (29) and sarcoidosis (25) as well patients with cancers other than lung cancer with and without lung metastasis (30). The capacity of each one of the candidate antibodies to discriminate these disease groups from the controls was estimated (Fig. 3C lower panel). As smoking is a strong risk factor for lung cancer, and lung cancer develops more often in males than in females, we included these groups in the analysis of potential confounding factors. The results (Fig. 4) show that the majority of the markers are specific for lung cancer, and neither smoking habits nor sex affect the biomarker levels. It should be noted the lower specificity of BSI0077 in collection IV as compared with collection III and the p-value of the comparison between the control and non-lung cancer patients with this antibody suggesting that it may recognize a rather general cancer biomarker. Redundancy of the group of specific antibodies was assessed by V-region sequencing of the IgG, and the results confirmed independent clones for all of our candidates. The analysis of the functional redundancy showed several pairs of antibodies (e.g., BSI0070/BSI0072 and BSI0351/BSI0352) with similar reactivity and the results was confirmed by competitive ELISA experiments between these pairs that indicated the recognition of distinct epitopes on the same antigen by antibodies from each pair (data not shown).

Statistical analysis therefore clearly shows that the levels of the biomarkers detected by the 10 mABs are different in healthy population as compared to the lung cancer patients (Figures 9-16). In order to further improve the diagnostic performance of the biomarkers, different combinations of these candidates were tested using leave-one out cross validation and linear discriminant analysis to identify for each mAb panel the optimal function capable of separating the controls from the patients. Indeed from the standpoint of application, an early cancer detection predictive assay would be possible if its sensitivity and the specificity are sufficiently high. The current tests using CT-scan as screening tool have to cope with the high cost of the test as well as its low specificity (64%). Both limitations can be overcome with a blood based test with sufficiently high specificity as the multiple biomarker panel presented here.

1013 combinations between the 10 mAbs were tested in panels with size ranging from 2 to 10 mAbs. 60% of these panels have accuracy higher than 0.8 (Figure 17), with panels achieving an accuracy of 0.84. All 10 mAbs are represented in these panels with similar frequency (Figure 18), although two antibodies have slightly higher occurrence (BsiO272 and BsiO392). As shown in the examples, a panel of six antibodies showed an increased diagnostics accuracy (Figure 4B), with a specificity higher than 82 % and a sensitivity higher than 80%. Results for further antibody combinations will be disclosed in the present application.

These experiments show that 10 mABs of the invention are sufficient for optimal performance.

The cognate antigens for these 10 mAB have been identified, as well as binding peptide sequences thereof. Three mABs are specific for LRGl (Leucine rich glycoprotein 1), two mABs recognize Haptoglobin and Haptoglobin related protein (HRP), one mAB recognizes C9 (complement factor 9), two mABs recognize CHF (complement factor H), and two bind Alpha- 1-antichymo trypsin (ACT, Serpin A3). These antigens, especially in combinations, thus represent novel valuable targets for lung cancer detection, as illustrated in the experimental section.

Furthermore, some of the mABs also react with cancer cells in-situ, as detected by immunohistology, allowing their use for histopathological classification, immunological or immunohisto logical staining, and/or imaging of cancer.

Altogether, the invention thus provides novel antibodies which represent highly valuable reagents for cancer detection, management, diagnosis, monitoring, histopathological classification, staining, or imaging. In a particular embodiment, the invention therefore relates to an antibody which binds a peptide having a sequence selected from SEQ ID NOs: 1 to 76 (see Table 4 below), or a fragment or derivative of such an antibody having the same antigen specificity.

Binding peptides recognized by antibodies of the invention are disclosed below. These peptides have been tested in phage display ELISA. Bold peptide sequences also bind to the corresponding mABs in biotinylated form, immobilized to avidin coated 96 well plates, in direct ELISA experiments. Binding competition, displacement with human plasma and specific purified antigen was also confirmed for some peptides.

Table 4

In a further particular embodiment, the invention relates to an antibody which binds a polypeptide comprising a peptide sequence selected from SEQ ID NOs: 1 to 76, or a fragment or derivative of such an antibody having the same antigen specificity.

In a particular preferred embodiment, the invention relates to an antibody which binds a peptide selected from SEQ ID NO: 26-35 and 57-63 and which also binds a human LRGl (Leucine Rich Alpha -2 Glycoprotein 1) protein, or a fragment or derivative of such an antibody having the same antigen specificity. The invention particularly relates to an antibody which binds a peptide selected from SEQ ID NO: 26-35 and 57-63 and wherein said binding is at least partially displaced by a human LRGl protein, or a fragment or derivative of such an antibody having the same antigen specificity.

Lung cancer specific antibodies binding to other peptide sequences have also been isolated. Corresponding antibodies and peptides are listed in table 6 (SEQ ID NO: 93-135).

Accordingly, the invention also relates to an antibody which binds a peptide having a sequence selected from SEQ ID NOs: 93-135 (see Table 6), or a fragment or derivative of such an antibody having the same antigen specificity.

Another particular preferred embodiment of the invention relates to an antibody which binds a peptide selected from SEQ ID NOs: 14-25 and which also binds a human C9

(Complement component 9) protein, or a fragment or derivative of such an antibody having the same antigen specificity. The invention particularly relates to an antibody which binds a peptide selected from SEQ ID NOs: 14-25 and wherein said binding is at least partially displaced by a human C9 protein, or a fragment or derivative of such an antibody having the same antigen specificity. A particularly preferred antibody of the invention binds a peptide selected from SEQ ID NOs: 14-22 and 24 and binds a human C9 protein, or a fragment or derivative of such an antibody having the same antigen specificity.

Another particular preferred embodiment of the invention relates to an antibody which binds a peptide selected from SEQ ID NOs: 36-38 and which also binds a human haptoglobin or haptoglobin-related protein, or a fragment or derivative of such an antibody having the same antigen specificity. The invention particularly relates to an antibody which binds a peptide selected from SEQ ID NOs: 36-38 and wherein said binding is at least partially displaced by a human haptoglobin or haptoglobin-related protein, or a fragment or derivative of such an antibody having the same antigen specificity.

Another particular preferred embodiment of the invention relates to an antibody which binds a peptide selected from SEQ ID NOs: 43-56 and 64-76 and which also binds a human CFH (Complement factor H, formerly known as Beta IH globulin) protein, or a fragment or derivative of such an antibody having the same antigen specificity. The invention particularly relates to an antibody which binds a peptide selected from SEQ ID NOs: 43-56 and 64-76 and wherein said binding is at least partially displaced by a human CFH protein, or a fragment or derivative of such an antibody having the same antigen specificity. A particularly preferred antibody of the invention binds a peptide selected from SEQ ID NOs: 64-71, 73 and 75-76 and also binds a human CFH protein, or a fragment or derivative of such an antibody having the same antigen specificity.

Another particular preferred embodiment of the invention relates to an antibody which binds a peptide selected from SEQ ID NOs: 1-13 and 39-42 and which also binds a human Alpha- 1-antichymo trypsin protein, or a fragment or derivative of such an antibody having the same antigen specificity. The invention particularly relates to an antibody which binds a peptide selected from SEQ ID NOs: 1-13 and 39-42 and wherein said binding is at least partially displaced by a human Alpha- 1-antichymo trypsin protein, or a fragment or derivative of such an antibody having the same antigen specificity. A particularly preferred antibody of the invention binds a peptide selected from SEQ ID NOs: 1-5 and 7-12 and also binds a human Alpha- 1 -antichymo trypsin protein, or a fragment or derivative of such an antibody having the same antigen specificity.

Most preferred examples of antibodies of this invention comprise all or part of a heavy chain variable region sequence selected from SEQ ID NOs: 77-92 (Figure 8).

The polypeptide may be or may comprise a complete variable region of an antibody, or may comprise only a part thereof, such part preferably comprising at least 5 consecutive amino acid residues, more preferably at least 6, 7 or 10 consecutive amino acid residues. A preferred polypeptide of the invention is a polypeptide comprising at least a CDR or FR domain of any one of SEQ ID NOs: 77-92. The FR domains are represented in blue (or grey) on Figure 8, while the CDR domains are represented in white. Variant sequences wherein from 1 to 10 amino acid residues have been replaced, deleted or inserted are also included in the present invention, as long as the modification does not substantially alter antigen binding capacity or specificity of the sequence or domain. Such variants typically contain 1, 2 or 3 amino acid modifications, typically consisting of replacement with amino acid of same nature. Illustrative and preferred examples of antibodies of the invention include monoclonal antibodies BsiO358, BsiO359, BsiO272, BsiO392, Bsi0080, BsiO352, Bsi0077, BsiO349, Bsi0270, BsiO271, Bsi0072, Bsi0076, Bsi0068, Bsi0033, Bsi0071, Bsi351, or Bsi0070, or a fragment or derivative thereof having the same antigen specificity. These antibodies comprise a variable region sequence as disclosed in SEQ ID NOs: 77-92, respectively.

A further object of the invention is a peptide consisting of an amino acid sequence selected from SEQ ID NOs: 1 to 76. A further object of the invention is an isolated nucleic acid encoding an antibody of the invention, or the fragment or derivative of said antibody. The nucleic acid typically encodes at least a portion of a variable region of the antibody, e.g., a portion of the heavy or light chain including a variable domain, such as a CDR or FR domain. The nucleic acid may be DNA or RNA, including cDNA, gDNA, recombinant DNA, synthetic or semi-synthetic DNA, which may be single- or double-stranded. A particular object of the invention is a nucleic acid encoding a polypeptide comprising at least a CDR or FR domain of any one of SEQ ID NOs: 77-92. The nucleic acid may be fused to other regions (e.g., constant regions, hinge regions, etc) to create synthetic antibodies, humanized antibodies, chimeric antibodies, etc., according to techniques well known per se in the art. They may also be used to produce single chain antibodies.

As mentioned, the invention relates to antibodies (or fragments or derivatives thereof) which "bind" peptides or polypeptides. In the context of the invention, such binding should be specific or selective, meaning that the binding to the reference peptide or polypeptide can be discriminated from (e.g., occurs with higher affinity or avidity than) possible non specific binding to other antigens. Preferred antibodies do not bind, under selective condition, to any other unrelated human blood protein but the reference protein. Binding of an antibody to the above reference peptide can be tested as disclosed in the examples. Binding to a peptide can be verified with either the isolated peptide (e.g., immobilized on a support) or with the peptide included in a larger polypeptide sequence. In a particular embodiment, the peptide is in isolated form and immobilized on a support (e.g., a plate) and the candidate antibody is incubated with the immobilized peptide. Binding may then be revealed using known techniques. Binding to a protein may be tested by incubating any sample containing the protein in solution, and verifying the formation of an immune complex. In a particular embodiment, the term "binding" to a peptide indicates the antibody can bind the corresponding peptide in biotinylated form, immobilized to avidin coated well plates in direct ELISA experiments.

The antibody may be a polyclonal or a monoclonal antibody, most preferably a monoclonal. It may be of various classes (e.g., IgG, IgE, IgM, etc.). The antibody may be of various animal origin, or human or synthetic or recombinant. Furthermore, the term antibody also includes fragments and derivatives thereof, in particular fragments and derivatives of said monoclonal or polyclonal antibodies having substantially the same antigenic specificity. Antibody fragments include e.g., Fab, Fab '2, CDRs, etc). Derivatives include humanized antibodies, human antibodies, chimeric antibodies, poly- functional antibodies, Single Chain antibodies (ScFv), etc. These may be produced according to conventional methods, including immunization of an animal and collection of serum (polyclonal) or spleen cells (to produce hybridomas by fusion with appropriate cell lines). Methods of producing polyclonal antibodies from various species, including mice, rodents, primates, horses, pigs, rabbits, poultry, etc. may be found, for instance, in Vaitukaitis et al.

[29]. Briefly, the antigen is combined with an adjuvant (e.g., Freud's adjuvant) and administered to an animal, typically by sub-cutaneous injection. Repeated injections may be performed. Blood samples are collected and immunoglobulins or serum are separared.

Methods of producing monoclonal antibodies from various species as listed above may be found, for instance, in Harlow et al (Antibodies: A laboratory Manual, CSH Press, 1988) or in Kohler et al (Nature 256 (1975) 495), incorporated therein by reference. Briefly, these methods comprise immunizing an animal with the antigen, followed by a recovery of spleen cells which are then fused with immortalized cells, such as myeloma cells. The resulting hybridomas produce the monoclonal antibodies and can be selected by limit dilutions to isolate individual clones. Antibodies may also be produced by selection of combinatorial libraries of immunoglobulins, as disclosed for instance in Ward et al (Nature 341 (1989) 544).

Recombinant antibodies of the invention, or fragments or derivatives thereof, may be produced by methods known per se in the art, for example by recombination in a host cell, transformed with one or more vectors enabling the expression and/or secretion of the nucleotide sequences encoding the heavy chain or the light chain of the antibody. The vector generally contains a promoter, translation initiation and termination signals, and suitable transcriptional regulatory regions. It is stably maintained in the host cell and may optionally possess specific signals for secretion of the translated protein. These different components are selected and optimized by one of skill in the art according to the host cell used.

Another object of the invention is an expression vector, for example a viral or plasmid vector, comprising a nucleic acid of the invention. The vector may replicate autonomously in the chosen host cell, or it may be an integrative vector. Also useful is an expression vector comprising a nucleic acid coding for the light chain of the antibody.

Another object of the invention is an expression vector comprising a nucleic acid coding for the heavy chain of an antibody of the invention.

Such vectors are prepared by methods known per se in the art, and the resulting clones may be introduced into a suitable host cell by standard methods, such as lipofection, electroporation, use of polycationic agents, heat shock, or chemical methods.

Another object of the invention is a host cell transfected with said vector or vectors. The host cell may be selected from among prokaryotic or eukaryotic systems, for example bacterial cells but also yeast cells or animal cells, in particular mammalian cells. Insect cells or plant cells may also be used.

Another object of the invention is a hybridoma cell producing an antibody of the invention.

In another aspect, the invention relates to a method for producing an antibody of the invention, said method comprising the following steps : a) culturing in a suitable culture medium a host cell expressing a heavy chain and/or a light chain such as defined herein; and b) recovering said antibodies so produced from the culture medium or from said cultured cells. To allow the simultaneous expression of the heavy chain (H chain) and the light chain (L chain) and their reassociation to form the recombinant antibody molecule, one may use two cassettes in a same expression vector. In this manner for example a double-recombinant vector may be prepared in which the sequence encoding each of the H and L chains is under the control of a strong promoter.

A particular example of a production method is production in an insect cell, as described for example in international patent application WO 96/07740. To this end, an expression cassette is used comprising a sequence coding for the variable region of the monoclonal antibody light chain, or a sequence coding for the variable region of the monoclonal antibody heavy chain, said sequence is placed under transcriptional control of a suitable promoter, for example a baculovirus promoter.

Another example of a production method is the use of a viral or plasmid expression vector for expressing the monoclonal antibody in a mammalian cell. Preferred mammalian cells for expressing the monoclonal antibody are the rat YB2/0 line, the hamster CHO line, in particular the lines CHO dhfr- and CHO Led 3, PER.C6TM (Crucell), 293, K562, NSO, SP2/0, BHK or COS, C0S7.

A further production method is the expression of the recombinant antibody in transgenic organisms, for example in plants (Ayala M, Gavilondo J, Rodriguez M, Fuentes A, Enriquez G, Perez L, Cremata J, Pujol M. Production of plantibodies in Nicotiana plants. Methods MoI. Biol. 2009; 483: 103-34) or else in the milk of transgenic animals such as rabbit, goat or pig (Pollock, D.P., J.P. Kutzko, E. Birck- Wilson, J.L. Williams, Y. Echelard and H. M. Meade. (1999) Transgenic milk as a method for the production of recombinant antibodies. Journal of Immunological Methods. 231: 147-157).

The antibodies of the invention may be coupled to heterologous moieties, such as toxins, labels, drugs or other therapeutic agents, covalently or not, either directly or through the use of coupling agents or linkers. Labels include radiolabels, enzymes, fluorescent labels, magnetic particles and the like. Toxins include diphteria toxins, botulinum toxin, etc. Drugs or therapeutic agents include lymphokines, antibiotics, antisense RNA or antisense nucleic acid, modified or not, growth factors, etc. Methods of using such heterologous moieties are illustrated, for instance, in US4,277,149 and US3,996,345. The antibodies of this invention have various applications, including, diagnostic, purification, detection, therapeutic, prophylactic, etc.

In vitro, they can be used as screening agents or to purify the antigen from various samples, including various biological samples (e.g., blood samples). As demonstrated in the examples, these antibodies have the remarkable property of binding to antigens which are differentially expressed between cancerous and control human subjects.

The invention thus relates to a diagnostic composition comprising an antibody (or a fragment or derivative thereof) as defined above. The composition may comprise any excipient or solid support.

The invention also relates to a method for detecting lung cancer in a subject, the method comprising contacting a sample from said subject, preferably a blood sample, with at least one antibody as defined above and determining the presence of an antigen bound to said at least one antibody, said presence being indicative of lung cancer. The method preferably comprises contacting said sample from said subject with at least two antibodies as defined above, preferably at least 3, in combination. As disclosed in the examples, the invention discloses particular antibody combinations which allow specific and sensitive determination of lung cancer in human subjects. These combinations include, without limitation:

- BsiO392, BsiO358, and BsiO272; or

- Bsi0071, Bsi0077, BsiO272, BsiO358, and BsiO392; or

- Bsi0071, Bsi0271, and Bsi0392; or

- BsiO271, BsiO272, and BsiO392; or

- Bsi0077, BsiO358, and BsiO392; or

- Bsi0071 , Bsi0077, and BsiO351 , BsiO358; or

- Bsi0033, Bsi0071, Bsi0077, BsiO271, BsiO272, BsiO351 , BsiO352, BsiO358, BsiO359, and BsiO392; or

- Bsi0077, BsiO272, BsiO352, and BsiO358; or

- Bsi0077, BsiO272, BsiO352, BsiO358, and BsiO392; or

- Bsi0071, Bsi0077, BsiO272, BsiO352, and BsiO358.

A preferred type of combinations comprises BsiO392 in combination with at least one or two additional antibodies.

Another preferred type of combinations comprises Bsi0071 in combination with at least one or two additional antibodies. Another preferred type of combinations comprises BsiO272 in combination with at least one or two additional antibodies.

Another preferred type of combinations comprises BsiO358 in combination with at least one or two additional antibodies. Another preferred type of combinations comprises Bsi0077 in combination with at least one or two additional antibodies.

Another preferred type of combinations comprises at least one anti-C9 antibody in combination with at least one or two additional antibodies.

The term "combination" indicates the sample should be tested for antigen binding to all antibodies of the combination, either simultaneously or separately (e.g., sequentially). Preferably, the antibodies are tested simultaneously (e.g., on the same device).

A further object of this invention resides in a device comprising at least one antibody as defined above immobilized on a support. The support may be, e.g., a membrane, a slide, a microarray, a chip or a microbead. Immobilization can be made through techniques known per se in the art (using linkers, cross linking reagents, passive adsorption, etc.).

A further object of the invention resides in a method for detecting lung cancer in a subject, the method comprising contacting a sample from said subject, preferably a blood sample, with at least one antibody that binds a Leucine-Rich alpha-2 glycoprotein (LRGl) and determining the presence of an antigen bound to said at least one antibody, said presence being indicative of lung cancer.

A further object of the invention is a method for detecting lung cancer in a subject, the method comprising contacting a sample from said subject, preferably a blood sample, with at least one antibody that binds a haptoglobin (HP) or HRP protein, and determining the presence of an antigen bound to said at least one antibody, said presence being indicative of lung cancer.

A further object of the invention is a method for detecting lung cancer in a subject, the method comprising contacting a sample from said subject, preferably a blood sample, with at least one antibody that binds a C9 protein and determining the presence of an antigen bound to said at least one antibody, said presence being indicative of lung cancer.

A further object of the invention is a method for detecting lung cancer in a subject, the method comprising contacting a sample from said subject, preferably a blood sample, with at least one antibody that binds a CFH protein and determining the presence of an antigen bound to said at least one antibody, said presence being indicative of lung cancer.

A further object of the invention is a method for detecting lung cancer in a subject, the method comprising contacting a sample from said subject, preferably a blood sample, with at least one antibody that binds an Alpha- 1 -anti chymotrypsin (ACT, SERPINA3) protein and determining the presence of an antigen bound to said at least one antibody, said presence being indicative of lung cancer. Preferred methods comprise contacting said sample from said subject with at least two antibodies which bind a distinct protein selected from Leucine -Rich alpha-2 glycoprotein, haptoglobin, haptoglobin related protein, C9, CFH or Alpha- 1-antichymo trypsin.

Further preferred methods comprise contacting said sample from said subject with at least three distinct antibodies which bind a distinct protein selected from Leucine-Rich alpha-2 glycoprotein, haptoglobin, C9, CFH or Alpha 1 Antichymotrypsin.

As discussed above, the term "antibody" as used therein also includes fragments and derivatives thereof as defined above.

A further object of the invention is a device comprising at least one antibody that binds a protein selected from Leucine-Rich alpha-2 glycoprotein, haptoglobin, haptoglobin related protein, C9, CFH or Alpha- 1 -antichymotrypsin, immobilized on a support. A further object of the invention is a device comprising at least two antibodies that each bind a distinct protein selected from Leucine-Rich alpha-2 glycoprotein, haptoglobin, C9, CFH or Alpha-1-antichymotrypsin, immobilized on a support. The support may be a membrane, a slide, a microarray, a chip or a microbead based detection system.

A further object of the invention is a kit comprising a device as defined above and a reagent to perform or detect (or quantify) an immune reaction, particularly an antibody-antigen complex. Reagents include labels, buffers, substrates, etc. The kits typically comprise containers for the different reagents and products, and may further comprise a support or other device suitable to perform the assay.

The antibodies can be used individually or in combination to measure the level of the cognate antigen (analyte) in biofiuids including serum, plasma, urine, cerebrospinal fluid, bronchoalveolar lavage (BAL) fluid, sputum, tear, sweat, amniotic fluid and inflammatory exudate, using any number of detection technologies or platforms such as, without limitation Capture assay, Sandwich assay, Competition assay, Radio-immuno assays, Enzyme labels with substrates that generate colored, fluorescent, chemiluminescent, or electrochemically-active products, Fluorescence, fluorescent polarization, Chemiluminescence, Optical and colorimetric, Electrochemiluminescence, Time -resolved fluorescence, Surface plasmon resonance, Evanescent wave, Multiwell plate (ELISA), Individual assay, Multiplex assay, Latex bead - multiplex assay, Microarray (Laminar surface) - multiplex assay, Glass, Ceramic (like Randox), Plate based assays, Strip based assays, dipsticks, Closed systems immunoassays. Preferred assay formats include:

Capture assay

An assay carried out using a single immobilized antibody (multiwall plate, latex bead, microarray, etc.) which captures a specific labeled protein from a biofiuid, the detection which is measured using appropriate detection reagents as detailed in the following paragraph.

The antibody is immobilized directly to the support or captured by an affinity reagent such as an anti-mouse IgG antibody coated onto the support. The immobilized antibody is then incubated with any of the above mentioned body fluids in which the proteins have been labeled with a detection molecule such as biotin, with or without pre -treatment to remove abundant proteins. The labeled protein which is bound by the antibody is detected by the addition of an appropriate detection reagent which binds to the label such as avidin or streptavidin which has been modified to be compatible with one of the detection technologies described in the section "detection technology."

The resulting signal provides a quantitative measure of the amount of labeled protein bound by the antibody Sandwich ELISA

An assay using two antibodies, the first which is immobilized on a support (multiwell plate, latex bead, microarray, etc.) which binds a specific protein from a biofluid, the detection which is measured using a labeled second antibody against the same protein and appropriate detection reagents as detailed in the following paragraph.

The first antibody is immobilized directly to the support or captured by an affinity reagent such as an anti-mouse IgG antibody coated onto the support. The immobilized antibody is then incubated with any of the above mentioned body fluids, with or without pre-treatment to remove abundant proteins. The antibody/antigen complex is then incubated with a second antibody, made against the same protein, which has been labeled with a detection molecule such as biotin. The bound antibody is detected by the addition of an appropriate detection reagent which binds to the label such as avidin or streptavidin which has been modified to be compatible with one of the detection technologies described in the section "detection technology." The resulting signal provides a quantitative measure of the amount of protein bound by the antibody

Competitive assay

An assay in which the binding of a labeled tracer protein by a single antibody as described in "capture assay" is inhibited by pre-incubation of a biofiuid to indirectly quantify the analyte.

The antibody is immobilized directly to the support or captured by an affinity reagent such as an anti-mouse IgG antibody coated onto the support. The immobilized antibody is then incubated with any of the above mentioned body fluids. The immobilized antibody/antigen complex is then incubated with a labeled tracer consisting of either (1) any of the above mentioned body fluids in which the proteins have been labeled with a detection molecule such as biotin, with or without pre -treatment to remove abundant proteins, or (2) a purified or recombinant protein recognized (bound) by the monoclonal antibody, or (3) a peptide which is recognized (bound) by the monoclonal antibody. The labeled protein or peptide which is bound by the antibody is detected by the addition of an appropriate detection reagent which binds to the label such as avidin or streptavidin which has been modified to be compatible with one of the detection technologies described in the section "detection technology."

The level of the specific protein in the unlabeled biofiuid is determined as a function of the inhibition of signal. Preferred Detection Technologies include:

Enzyme labels with substrates that generate colored, fluorescent, chemiluminescent, or electrochemically-active products .

The detection reagent (for example steptavidin or avidin, which binds to biotin) is coupled to an enzyme such as horseradish peroxidase which is capable of catalyzing o an appropriate colorimetric substrate of which the product demonstrates maximal absorbance at a given wavelength allowing the quantitative measurement of the labeled protein by measuring the optical density of the final product in the well at or near the wavelength of maximal absorbance. o a chemiluminescent substrate to a sensitized reagent which upon oxidation emits light, providing the quantitative measurement of the labeled protein. o a chemiluminescent substrate to a sensitized reagent which upon the application of an electrical current emits light, providing the quantitative measurement of the labeled protein

Fluorescence

The detection reagent (for example steptavidin or avidin, which binds to biotin) is coupled to a fluorescent tag. Preferred platform Technologies include:

Multiwell Plate

o Single test: one antibody is immobilized per well either directly or indirectly using a capture reagent such as goat anti-mouse antibody.

o Multiplex: 2 or more antibodies are immobilized in a single well by deposition in a pattern

Latex bead

Two or more antibodies are immobilized onto a latex bead between x and y microns

Arrays, microarrays, and nanoarrays

• Two or more antibodies are spotted onto an activated laminar surface with a spot diameter between 100 μm - 5 mm (arrays), 2 μm - 100 μm (microarrays), 10 nm-2 μm (nano-arrays) The surface can be composed of glass, plastic, ceramic, carbon nanotube lattice etc. The method can be performed at any stage of the disease, such as early or late stage, to confirm or reject a prior diagnosis, select patients for surgery, classify cancer type or severity, or monitor patients. The test may also be conducted before disease symptoms, as a first line detection. Clinical diagnostic application of a lung cancer plasma (serum) test will be useful e.g., for patients who are suspected to have lung cancer because of a suspicious nodule that has been detected by imaging of the lung (CT scan). Nodules < 0.5 cm are suspicious and quite frequent in the populations. Nodules > 0.5 cm, but < 1.1 cm represent an increased likelihood of being cancerous, however challenging to find by surgery and invasive endoscopic procedures. It is important to select patients from this group who's nodule are definitively cancerous to reduce the burden of futile surgeries and other invasive diagnostic procedures.

The test will avoid futile thoracotomies and unnecessary and expensive imaging technologies that are not specific enough and expose the patients to potentially harmful irradiation, and missed cures as observed patients could receive the test repeatedly.

Further aspects and advantages of the invention will be disclosed in the following examples, which should be considered illustrative.

Examples

A. Methods

Clinical samples: Plasma specimens were in part purchased from Proteogenex (Culver, CA) and Asterand (Royston, UK). Other samples (see Table 1) were collected at the Department of Pulmonology of the University of Debrecen in Hungary from informed and consented patients and matched (age, sex and smoking habit) apparently healthy individuals by a clinical protocol RKEB/IKEB:2422-2005 approved by the regional ethics committee and the IRB of the clinic.

Plasma processing:

-depletion of high abundant proteins for preparation of the immunogen and the pooled plasma tracers:

Pools of 20 mL from both normal and NSCLC subjects (collection I) were made by combining 1.0 mL of plasma from each of the respective individual donors. The pooled plasma was immediately aliquoted into 250 uL per tube, and samples were stored at -75⁰C. Samples were thawed on ice prior to processing. The plasma fractionation was performed. Depletion of abundant proteins was performed using a commercially available SEPPRO IgY12 LC10® (12.7 x 79.0 mm) from Beckman Coulter (Fullerton, CA) on a BioCad chromatography HPLC workstation (Applied Biosystems, Foster City, CA). Chromatography was performed according to the technical protocol supplied by the vendor, with minor buffer modifications. Briefly, to reduce the concentration of the twelve most abundant proteins, a plasma sample (250 uL) was thawed and diluted by addition of 750 uL of buffer A (25 mM Tris, 0. 5M NaCl, ImM MnC12, ImM CaC12 and 0.05% sodium azide, pH 7.4). The diluted plasma was loaded onto the Seppro 12 column at a flow rate of 0.5mL/min for 30 min.; the flow rate was then increased to 2 mL/min for the remainder of the run. The unbound proteins (depleted fraction) were washed off with binding buffer and the depleted fraction was collected into a 15 mL centrifugal filter Amicon with a cut-off at 5kDa. The depleted plasma was concentrated by centrifugation at 3,500 x g. The bound proteins were eluted from the column with stripping buffer (100 mM glycine, pH 2.5) and collected into a separate tube. The column packing material was neutralized with 100 mM Tris-HCl, pH 8.0 for 10 min and re-equilibrated with binding buffer, before performing protein depletion from the next plasma sample. A total of 27 depletion runs for the normal plasma and 20 runs for the lung cancer plasma were performed. The concentrated and depleted fractions from each group were pooled before further processing. - depletion of high abundant proteins from individual plasma samples for tracers preparation

The depletion of the seven most abundant proteins was performed using commercially available Human-7 Multiple Affinity Removal System columns (10X100mm) from Agilent Technologies (Santa Clara, CA). The process was automated using a chromatography system AKTA™purifier 10 - collector F950" from GE Healthcare (Chalfont St. Giles, UK) connected to an autosampler A-900. The individual plasmas (70 uL) were diluted four times with the vendor provided equilibration buffer (buffer A) and filtered using 0.22 μm spin filters at 16000 g for 1 min. The filtered plasmas were loaded on the column at a flow rate of 0. 5 ml/min and the depleted plasmas (flow-through) collected in the equilibration buffer between 9 and 14 min after injection were further concentrated using spin filters Vivaspin 500 uL (Sartorius stedim, FR) with a cut-off of 5 kDa. The bound proteins were removed from the column using low pH stripping buffer provided by the vendor (buffer B) at a flow rate of 3 ml/min for 7 min and the column was re-equilibrated with buffer A for 8 min at a flow rate of 3 ml/min between each injection. Under these conditions the MARS Hu-7 columns were used according to the vendor specifications for maximum of 200 runs and their performance was followed at regular intervals (each 40 runs) for protein leaking (albumin, IgG, IgA and fibrinogen) by an ELISA using commercially available mAbs.

- glycoprotein enrichment

Glycoprotein enrichment was performed using a multi-lectin affinity chromatography (M- LAC) column. The lectin column contains a mixture of 3 lectins: Con A, WGA and JAC from Vectors Laboratories (Burlingame, CA) and was prepared in-house using Aldehyde POROS- 20 AL® (20 μm beads) beads from Applied Biosystems, (Foster City, CA) as previously reported ¹⁵. One mL of the pooled depleted plasma sample was loaded on a 7.8 mL M-LAC column (10 x 100) at a flow rate of 0.5 mL/min for 20 min. The flow rate was increased to 4 mL/min, and the unbound proteins (flow through fraction) were washed-off with M-LAC binding buffer (25 mM Tris, 0. 5OM NaCl, ImM MnCl₂, ImM CaCl₂ and 0.05% sodium azide, pH 7.4) for 10 min. The flow through fraction was collected and stored at -75°C. The proteins which bound to the M-LAC were eluted with 100 mM acetic acid, pH 3.8 at 4 mL/min for 10 min. The eluted glycoproteins were collected directly into a 15 mL Amicon filter device and concentrated as described above. The sample was buffer exchanged into IX PBS by addition of 14 mL of buffer to the filter device; the volume was reduced down to ~ 1.0 mL by centrifugation; this step was repeated twice. The M-LAC column was neutralized with 0.5 M Tris, pH 7.5, IM NaCl, 0.05 % sodium azide and equilibrated prior to the next run. The bound glycoproteins from 5 runs were pooled, aliquoted and stored at -75°C. The total glycoprotein yield was approximately 5.0 mg for the lung cancer and 3.0 mg for the matched control samples.

-normalisation

Protein normalization of the depleted plasma and the glycosylated protein fraction was performed using an in-house immunoaffinity column prepared as follows: rabbit polyclonal antibodies raised against normal human serum from Sigma (Saint Louis, MO) were covalently linked to HiTrap Protein G HP column (4.6 x 100) from GE Healthcare (Chalfont St. Giles, UK) using 15 mM dimethyl pimelimidate and 15 mM dimethyl suberimidate as previously described . For normalization of the depleted plasma and glycosylated protein enriched fraction, one mg of proteins was loaded at a flow rate of 1 ml/min and incubated with the column for 5 min before washing with Ix PBS buffer at pH 7.0. The flow through represented the normalized protein fraction, while the bound proteins were eluted with stripping buffer B from Agilent Technologies (Santa Clara, CA). The column was equilibrated with Ix PBS.

-biotinylation of plasma protein fractions

The pooled or individual depleted plasma was labeled with a bifunctional NHS-biotin having a long alkyl chain as a spacer EZ-Link Sulfo-NHS-LC-Biotin from Pierce (Rockford, IL). Labeling was performed in PBS buffer (pH 7.0) at a 100 time molar excess of biotin assuming an average protein mass of 68 kDa for 30 min. at room temperature. The non-reacted protein was removed using a 5 ml HiTrap desalting column HP from GE Healthcare (Chalfont St. Giles, UK) with PBS (pH 7.0) at a flow rate of 1 ml/min. Immunisation of BALB/c mice

Two groups of four female Balb/c mice of at least 8 weeks of age from Charles River Laboratories (Evry, France) were injected subcutaneously in the rear footpads and at the base of the tail with the two complex antigen protein mixtures. Each mice received 10 μg protein of glycoprotein enriched depleted plasma (group A) and normalized glycoprotein enriched depleted plasma (group B) on days 1, 15 and 29. Complete Freund's adjuvant (Sigma, Saint Louis, MO) was used in all cases. Blood was taken from each mouse by retro- orbital bleed using a Pasteur pipette on days 19 and 33 to monitor antibody production by ELISA. Three weeks minimum after the third immunization (day 52 for group A and day 61 for group B) an additional injection with 10 ug of the complex antigen mixture in PBS pH 7.0 was performed to boost the immune response. Mice were treated in an animal facility according to the French and European laws and regulations regarding animal experimentation.

Fusion, cell growth and cloning

The procedures were done according to published methods. In short; Sp2/0-Ag-14 cells were fused to splenocytes of immunized mice with the help of polyethylene glycol (PEG). Fused cells were seeded to provide quasi clonal distribution of hybrids, which were selected in HAT media. Hybridoma cell lines were cloned before antibody production (in-vitro or in-vivo)

High throughput direct-ELISA Screening:

All chemicals, unless specified, were obtained from Sigma (St. Louis, MO). 384 well high- binding plates (Corning Inc., Lowell, MA) were coated with 20 u_g/ml (13 μl per well) goat anti-mouse Ig gamma chain specific polyclonal antibody (GAM) from Southern Biotechnology Associates, Inc. (Birmingham, AL) in coating buffer at pH 9.6 for 2h at room temperature (RT). The plates were washed four times with 80 μl/well of PBS containing 0.05% (v/v) Tween (washing buffer) and blocked with 40 μl of PBS and 0.5% bovine serum albumine (BSA) at 4 ⁰C overnight. Supernatants from the nascent hybridomas or from the clonal cell lines were added non-diluted to the wells. Each hybridoma supernatant was added to four adjacent wells to provide four independent readings. Mouse anti -human mAb against albumin was used as positive control and spotted in eight wells at 1.2 μg/ml in CM (13 μl/well). CM was used as negative control and also added to eight wells. The plates were incubated and then washed four times with washing buffer . All wells were incubated with biotinylated depleted plasmas (tracers) at 10 μg/ml in PBS with 0.05% (v/v) Tween and 1 % low IgG FBS. Incubation was continued for 90 min. at RT and the unbound proteins were removed by washing the plates four times with washing buffer. HRP-coupled avidin (Vectastain Elite ABC Peroxidase kit from Reactolab SA, Switzerland) was used as specified by the vendor's protocol. After four fold washing with washing buffer, reaction development was carried out by adding 20 μl freshly prepared substrate solution to each well (o-phenylenediamine at 0.4 mg/ml in 0.05 M phosphate/citrate buffer pH 5.0). The kinetics of the reaction development at 37 ⁰C was followed at 450 nm by recording the absorbance multiple times. Liquid handling was performed using Multidrop Combi from Thermo (Waltham, MA), Multimek with 96 pin head from Beckman Coulter (Fullerton, CA) and STAR from Hamilton (Reno, NV). Plate washing was performed using ELX405 from BioTek (Winooski, VT). Absorbance was measured with a microplate reader SpectraMax from Molecular Device (MDS, Toronto, Canada).

Screening data analysis.

Vmax of the chromogenic reactions were calculated from the linear part of the kinetic readings using the software provided with the plate reader SoftMax Pro from Molecular Device (MDS, Toronto, Canada). Each plate had eight positive and negative controls used to calculate Z' factor, a metrics used to quantify the quality of the screening experiment with respect to reproducibility and data scatter . Plates with a Z' factor below 0.5 (usually less than 10% in a screening campaign) were repeated. The positive (PC) and negative controls (NC) were used to normalize the data across plates and according to the following formula:

VmaxN _sampie = (Vmax _sampi_e -Vmax _NC)/ (Vmax _PC-Vmax _NC). Aberrant data (outliers) for each group of replicates (four per hybridoma sample reacted with one tracer and eight per control reacted with one tracer) were removed using automated procedure based on the mean and standard deviation values of the multiple measurements.

For each sample (i), the coefficient of variation (CV) is calculated on the n replicates. CF is a normalized measure of dispersion of the probability distribution and it is defined as the ratio of the standard deviation σ to the mean μ as follows:

CV(i) = σ(i) / μ(i)

If CV > 0.05, then a maximum (T_max) and a minimum (Tm_1n) thresholds are calculated as follows:

T_max (i)= μ (i) + 1.3*σ (i)

Tm_1n (i)= μ (i) - 1.3* σ (i)

The replicates which VmaxN is lower than T_1111n or higher than T_max are removed. In total 9.2 % of all generated data was considered as aberrant which represents the removal of less than one measurement for every two samples. The data obtained after normalizing and averaging the replicates (Figure 6) were further analyzed using statistical methods.

Statistical analysis

The normality of the distribution of the results was estimated using Wilks-Shapiro test ²⁰. The distribution of the results for each hybridomas, mABs was calculated separately for the controls and the lung cancer samples. Nonparametric statistical analyses were applied: differences between two independent groups were determined by means of Mann-Whitney U test; differences between more than two groups were determined by means of Kruskal- Wallis one-way analysis of variance test. A p-value less than 0.05 was considered significant. Algorithm for Panel prediction

The dataset was partitioned randomly into two parts (Figure 7), a training dataset and a testing dataset with a respective size of 50%(75%) and 50%(25%) of the available cancer and control samples from collection III. The classifiers were built and tested on the training set and the best one was finally applied on the testing data set to determine the diagnostic accuracy of the classifier. Selection of hybridomas for building the classifier was carried out by ranking the hybridomas according the p-values obtained for each one of them in comparing the results between the two groups (control vs lung cancer) using the Mann- Whitney test. All combinations between the sixteen best hits were created and the performance of each classifier was evaluated by Leave -one-out Cross Validation (LOOCV) as follows: LDA (Linear Discriminant Analysis) model was used with each classifier on the training set with the first observation (sample) omitted and then a prediction was made for the omitted sample. The correctness for the prediction was recorded and the procedure is repeated for all samples in the training set. For each classifier the area under the receiver operating curve (AUC) was calculated in order to compare the performance of the panels and select the best panel.

The R statistical software (available from www.cran.r-project.org) was used for the statistical analysis and for the implementation of the algorithm.

Phage display

The Ph.D-12™ Phage Display Peptide Library Kit was obtained from New England Biolabs (Beverly, MA). The phages display random peptide 12-mers are fused to a minor coat protein (pill) of M13 phage. The library consists of approximately 2.7x10 sequences. The experiments were carried out as described in the Instruction Manual of the Phage Display Kit with minor modifications. Phages were propagated using the supplied E. coli (ER2738) host strain. 96-well microtiter plates (Maxisorp Cat. no. 442404, Nunc, Napperville, IL) were coated with the mapped mAbs (10 μg antibody in 100 μl 0.1 M NaHCO₃ pH-8.6/well). 4xlO¹⁰ phages were added in TBS-Tween to the wells and incubated for 1 hour at room temperature. Bound phages were eluted by lowering the pH (0.2 M Glycine -HCl pH 2.2). Recovered phages were amplified and their titer determined. The selection was repeated twice using the same steps described above. After the third round of panning the target binding of selected phages was tested by ELISA. After the third round of panning individual clones were picked up from plates and grown in 96-deepwell plates (Eppendorf, Hamburg, Germany). Phages were isolated from the culture supernatant using PEG precipitation. 96-well plates were coated with the target mAbs as in case of the selection. The isolated phages were diluted in TBS-T ween and incubated for 1 hour at room temperature in the antibody coated microtiter wells. Bound phages were detected with anti- Mi 3 antibody-horseradish peroxidase conjugate using OPD substrate. Results were scanned with a Victor² (PerkinElmer/Wallac, Waltham, MA) microplate reader. ELISA positive clones were selected for DNA sequencing. Phage clones with confirmed binding (ELISA) to the target mAbs were grown in deepwell plates and isolated by PEG precipitation. Single stranded DNA was prepared from them using precipitation with 4 M NaI and ethanol. The sequencing was done by Biomi Ltd. (Godollό, Hungary) on an ABI 3100 Genetic Analyzer (Life Technologies/Applied Biosystems Carlsbad, CA) using the BigDye Terminator 3.1 Cycle Sequencing Kit (Life Technologies/ Applied Biosystems) and the -96 gill sequencing primer suggested by the Instruction Manual of the Phage Display Kit. For each mapped mAb at least 12 independent phage clones were sequenced in order to determine the consensus motif for the mapped epitop. The peptides sequences obtained from the experiments were manually curated to derive a set of unique sequences for each antibody. The obtained datasets were aligned using ClustalW and the matrix of the identity scores for all sequences derived from the experiments were calculated based on the PAM series. Then for each pair of antibodies an epitope redundancy score (EIS) was calculated using the following formula:

EIS =∑ (Xi j)/n*m, where Xi,j is the calculated identity score between the ith sequence from the first dataset and the jth sequence from the second dataset; m and n are the number of unique sequences in the two datasets. Cognate antigen protein (protein ID) identification:

Immunoprecipitation and SDS PAGE electrophoresis was used to specifically precipitate and isolate cognate protein antigens for the mABs. Specific bands were cut from the gels dugested with trypsin and analyzed with MS/MS methodolgy.

Western blotting technology:

Standard methods with chemoluminescent detection were used. B. Isolation and characterization of lung cancer antibodies

This example discloses with more details the isolation, cloning and characterization of the following antibodies:

We produced and characterized nascent monoclonal antibody libraries directed to the natural form of protein antigens present in the plasma of lung cancer patients. Differential plasma profiling of normal and cancer plasma proteomes of four clinical cohorts, totaling 370 patients with lung cancer and 146 controls via high throughput ELISA screening, identified several lung cancer specific (p<0.01) monoclonal antibodies. Further tests with plasma from patients with other inflammatory lung diseases and non-related cancers confirmed the lung cancer specificity of the monoclonal antibodies. The cognate antigen was identified for these antibodies, as well as binding peptides thereof. These antibodies were tested, alone and in combinations, on clinical samples, and their ability to discriminate cancer and control subjects was confirmed. The results are presented below, for each of the above ten antibodies. For each antibody, when available, the sequence of the variable region is presented; differential cancer vs. control binding is illustrated, the cognate antigen is disclosed as well as binding peptides. Bl - LRGl -binding antibodies

The leucine -rich repeat (LRR) family of proteins, including LRGl, have been shown to be involved in protein-protein interaction, signal transduction, and cell adhesion and development. LRGl is expressed during granulocyte differentiation (O'Donnell et al., 2002). Human LRGl was isolated from human serum by Haupt and Baudner, 1977. By sequence analysis, Takahashi et al. (1985) determined that purified LRGl protein has 312 amino acids and an experimentally determined molecular mass of 45 kD. The LRGl polypeptide contains 1 galactosamine and 4 glucosamine oligosaccharides attached and has 2 intrachain disulfide bonds. Leucine comprises 66 of the 312 amino acids, and LRGl contains at least 8 24-amino acid leucine -rich repeats. Increased LRGl expression was detected in GCSF-treated human cells derived from a patient with myeloproliferative disorder. In contrast, decreased LRGl expression was detected after PMA treatment and induction of monocytic differentiation of HL-60 cells A number of proteomics studies focused on plasma and serum have demonstrated an association between elevated levels of leucine rich alpha-2 glycoprotein (LRGl) and the presence of a number of different cancers in multiple preliminary studies. However, none of the findings have been confirmed by clinical validation. All of these studies are based on proteomics and none involved the use of antibodies.

The invention discloses novel antibodies that recognize specific epitopes in LRGl that represent cancer-specific biomarkers.

BsiO392

BsiO392 is an IgG type monoclonal antibody. The heavy chain variable region amino acid sequence is represented in SEQ ID NO: 92 (see Figure 8), which is reproduced below (CDRs are underlined):

GEPTYADDFKGRFAFSLETSATTAYLOINNLKNEDTATYFCARGGYYGNYDYAMD YWGQGTSLTVSS

The difference in biomarker level with BsiO392 is represented Fig 9a, showing a very substantial difference between control and lung cancer. In western blot, BsiO392 recognizes a 4OkDa band in total control plasma samples (see fig 9b). By mass spectrometry analysis of the antigen, various peptide sequences were obtained (see table below), which demonstrate that BsiO392 binds to LRGl.

RGPLQLERL

KDLLLPQPDLRY

KALGHLDLSGNRLR

RTLDLGENQLETLPPDLLRG

RWLQAQKD RTLDLGENQLETLPPDLLRGPLQLERL

RVAAGAFQGLRQ

RCAGPEAVKGQTLLAVAKS

BsiO352 BsiO352 is an IgG type monoclonal antibody. The heavy chain variable amino acid sequence is represented in SEQ ID NO: 89 (see figure 8), which is reproduced below (CDRs are underlined):

EVOLOESGPSLVKPSOTLSLTCSVTGDSITSGSWNWIREFPGNKLEYMGYISYSGST DYSPSLKSRISITRDTSKNOYYLOLNSVTTEDTATYYCATHYYGYLSLDYWGOGTS VTVSS

The difference in biomarker level with BsiO352 is represented Fig 10, showing a very substantial difference between control and lung cancer. In western blot, BsiO352 recognizes a 4OkDa band in total control plasma samples similar to BsiO392. By mass spectrometry analysis of the antigen, various peptide sequences were obtained (see table below), which demonstrate that BsiO352 binds to LRGl .

RGPLQLERL

VAAGAFQGLRQ

KALGHLDLSGNRLRK

RTLDLGENQLETLPPDLLRG

RTLDLGENQLETLPPDLLRGPLQLERL

KLQELHLSSNGLESLSPEFLRPVPQLRV

RCAGPEAVKGQTLLAVAKS

VAAGAFQGLR

TLDLGENQLETLPPDLLR

DLLLPQPDLR

TLDLGENQLETLPPDLLR

YLFLNGNK Peptides bound by BsiO352 have been identified and verified, using either phage display technique or direct Elisa binding. These peptides are presented as SEQ ID NOs: 26-35. BsiO351

BsiO351 is an IgG type monoclonal antibody. The heavy chain variable amino acid sequence is represented in SEQ ID NO: 88 (see fig 8), reproduced below (CDRs underlined).

SIGYKPSLKSRISITRDTSKNQFFLQLNSVTTEDTATYYCARGGFYALDYWGQGTS

The difference in biomarker level with BsiO351 is represented Fig 11, showing a very substantial difference between control and lung cancer. In western blot, BsiO351 recognizes a 4OkDa band in total control plasma samples similar to BsiO392. By mass spectrometry analysis of the antigen, two peptide sequences were obtained (KDLLLPQPDLRY and RTLDLGENQLETLPPDLLR), which demonstrate that BsiO351 binds to LRGl .

Peptides bound by BsiO351 have been identified. These peptides are presented as SEQ ID NOs: 57-63.

B2. Alpha- 1- antichymotrypsin-binding antibodies

Bsi358 BsiO358 is an IgG type monoclonal antibody. The heavy chain variable amino acid sequence is represented in SEQ ID NO: 90 (see fig 8) reproduced below (CDRs underlined).

NNYATYYVDSVKDRFTIYRDDSOSMLYLOMNNLKTEDTAIYYCVREGDWGOGTL VTVSA The difference in biomarker level with BsiO358 is represented Fig 12a, showing a very substantial difference between control and lung cancer. Western blots with BSI0358 on normal and LC plasma samples before and after depletion of abundant proteins by Agilent depletion columns is represented Fig 12b. Molecular mass of reactive band from LC plasma seems to be lower, suggesting that BSI 0358 reacts with a disease-specific form of protein as well.

By mass spectrometry analysis, binding assays and SW assays, we have demonstrated that BsiO358 binds to purified natural Alpha- 1 -anti chymotrypsin and works in SW assays with antibodies that recognize Alpha- 1 -antichymotrypsin.

Peptides bound by BsiO358 have been identified and verified, using either phage display technique or direct Elisa binding. These peptides are presented as SEQ ID NOs: 1-13.

Furthermore, in histology experiments using BsiO358, it was demonstrated that BsiO358 can image non small cell lung cancer (NSCLC), as illustrated fig 12c: >90% of NSCLC is positive and 17.6% of SCLC is positive.

BsiO359

BsiO359 is an IgG type monoclonal antibody. The heavy chain variable amino acid sequence is represented in SEQ ID NO: 91 (see fig 8) reproduced below (CDRs

underlined).

EVOLVESGGGLVOPKGSLKLSCAASGFTFSTSAMNWVROAPGKGLEWISRIRSKTN NYATy

VTVSA The difference in biomarker level with BsiO359 is represented Fig 13a, showing a very substantial difference between control and lung cancer. Importantly, as for BsiO358, BSI 0359 alone or in combination, detects a Lung Cancer specific form of the cognate antigen (see Figure 13b).

By mass spectrometry analysis, binding assays and SW assays, we have demonstrated that BsiO359 binds to purified natural Alpha- 1 -anti chymotrypsin and works in SW assays with antibodies that recognize Alpha- 1 -antichymo trypsin. Peptides bound by BsiO359 have been identified and verified, using either phage display technique or direct Elisa binding. These peptides are presented as SEQ ID NOs: 39-42.

Furthermore, in histology experiments using BsiO359, it was demonstrated that BsiO359 can be used efficiently for cancer histopathology: staining was positive for 84% of NSCLC, 88% adenocarcinoma, 64% squamosus cell carcinoma, and 100% of large cell carcinomas. BsiO359 thus could be useful for imaging.

B3. Complement factor 9-binding antibodies Complement C9 is a component of the complement system, a multi -protein biochemical cascade which aids to clear pathogens. The cascade is activated upon binding of IgG or IgM to pathogen molecules. C9 is one of the terminal components of the cascade and is responsible for forming pores in target cells leading to their destruction. Deficiencies in complement proteins are believed to be linked to auto-immunity and higher sensitivity to infections. The cDNA coding for C9 was sequenced and the protein sequence—537 amino acids in a single polypeptide chain—was derived. The amino -terminal half of C9 is predominantly hydrophilic and the carboxyl-terminal half is more hydrophobic. The amphipathic organization of the primary structure is consistent with the known potential of polymerized C9 to penetrate lipid bilayers and cause the formation of transmembrane channels as part of the lytic action of MAC. Marazziti et al. (1988) compared gene and protein structure of C9 and compared both with low density lipoprotein receptor (606945). The C9 gene is composed of 11 exons with lengths between 100 and 250 bp, except for exon 11 which extends over more than 1 kb, as it includes the 3-prime untranslated region. Witzel-Schlomp et al. (1997) gave revised information on the structure of the C9 gene, especially the exon-intron boundaries

To date, C9 has not been associated with cancer of any type. The present invention discloses specific antibodies against C9 that can be used as lung cancer biomarkers.

BsiO272

BsiO272 is an IgG type monoclonal antibody. The heavy chain variable amino acid sequence is represented in SEQ ID NO: 86 (see fig 8) reproduced below (CDRs

underlined).

OVOLOOPGAELVRPGASVKLSCKASGYSFASYWMNWVKORPGOGLEWIGMIHPS DSGTSLDEKFKDKATLTVDKSSNTAYIOLNSPTSEDSAVYYCAREGYD.PAWFAY WGQGTLVTVSA

The difference in biomarker level with BsiO272 is represented Fig 14a, showing a very substantial difference between control and lung cancer. In western blot, BsiO272 recognizes a band in total control plasma samples which is compatible to a 559aa polypeptide corresponding to mature C9 polypeptide (Fig 14b). By mass spectrometry analysis of the antigen, several peptide sequences were obtained (see table below), which demonstrate that BsiO272 binds to C9.

AIEDYINEFSVR

DGNTLTYYR

DVVLTTTFVDDIK

FEGIACEISK GTVIDVTDFVNWASSINDAPVLISQK

LSPIYNLVPVK

NRDVVLTTTFVDDIK

RPWNVASLIYETK

SIEVFGQFNGK

SIEVFGQFNGKR

TEHYEEQIEAFK

TSNFNAAISLK

VVEESELAR

Peptides bound by BsiO272 have been identified and verified, using either phage display technique or direct Elisa binding. These peptides are presented as SEQ ID NOs: 14-25.

Furthermore, in histology experiments using BsiO272, it was demonstrated that BsiO272 stains non small cell lung cancer (NSCLC), as illustrated fig 14c: >74% of NSCLC is positive.

B4. Haptoglobin (HP)- and Haptoglobin related protein (HRP)-binding antibodies

Haptoglobin (HP), NM 005143 is a tetrameric protein that functions to bind free plasma hemoglobin, thereby allowing degradative enzymes to gain access to the hemoglobin, while at the same time preventing loss of iron through the kidneys and protecting the kidneys from damage by hemoglobin. Mutations in the HP gene and/or its regulatory regions cause ahaptoglobinemia or hypohaptoglobinemia.

An increase in haptoglobin (Hp) levels or changes in Hp glycosylation have been associated with almost major forms of cancer. The present invention discloses specific antibodies against HP or HRP that can be used as lung cancer biomarkers.

Bsi0033

Bsi0033 is an IgG type monoclonal antibody. The heavy chain variable amino acid sequence is represented in SEQ ID NO: 77 (see fig 8) reproduced below (CDRs underlined).

GNSKYDPKFOGKATIT ADTSSNTAYLOLSSLTSEDTAVYFCTKSAGVPF AYWGOG TLVTVSA

The difference in biomarker level with BsiOO33 is represented Fig 15a, showing a very substantial difference between control and lung cancer. In western blot, BsiOO33 recognizes a band which is consistent with reaction of HP under non-reducing conditions (Fig 15b). By mass spectrometry analysis of the antigen, several peptide sequences were obtained (see table below), which demonstrate that BsiOO33 binds to HP and HRP. HP peptides

DIAPTLTLYVGK

SPVGVQPILNEHTFCAGMSK

ILGGHLDAK

DIAPTLTLYVGKK

HYEGSTVPEK

VGYVSGWGR

QKVSVNER

VTSIQDWVQK

VTSIQDWVQK

GSFPWQAK

KQLVEIEK

HRP peptides

DIAPTLTLYVGK

VTSIQDWVQK DIAPTLTLYVGKK

VGYVSGWGQSDNFK

Peptides bound by Bsi0033 have also been identified and verified, using either phage display technique or direct Elisa binding. These peptides are presented as SEQ ID NOs: 36- 38.

Furthermore, in histology experiments, it was demonstrated that Bsi0033 can react with non small cell lung cancer (NSCLC) as well as large cell carcinomas: 81% of NSCLC, 89% adenocarcinoma, 20% squamosus cell carcinoma, and 83% of large cell carcinomas are positive.

Bsi0071

Bsi0071 is an IgG type monoclonal antibody. The heavy chain variable amino acid sequence is represented in SEQ ID NO: 81 (see fig 8) reproduced below (CDRs underlined).

SFTYYPDNLKGRFTVSRDNAKDTLYLOMSSLRSEDTAIYYCAROSLGYYFDSWGO GTTLTVSS

The difference in biomarker level with Bsi0071 is represented Fig 16a, showing a very substantial difference between control and lung cancer. In western blot, Bsi0071 recognizes a band which is consistent with reaction of HP under non-reducing conditions, similar to BsiOO33. By mass spectrometry analysis of the antigen, several peptide sequences were obtained (see table below), which demonstrate that Bsi0071 binds to HP and HRP.

HP peptides

DIAPTLTLYVGK

GSFPWQAK

ILGGHLDAK KQLVEIEK

SCAVAEYGVYVK

VGYVSGWGR

VTSIQDWVQK

YVMLPVADQDQCIR

B5. Complement factor H (CFH)-binding antibodies

Complement Factor H (NM OOO 186) is a member of the Regulator of Complement Activation (RCA) gene cluster. The CFH protein contains twenty short consensus repeat (SCR) domains, is secreted into the bloodstream, and has an essential role in the regulation of complement activation, restricting this innate defense mechanism to microbial infections. Mutations in this gene have been associated with hemolytic-uremic syndrome (HUS) and chronic hypocomplementemic nephropathy. Complement factor H (CFH) is an inhibitor of the alternative complement pathway. The present invention discloses specific antibodies against CFH that can be used as lung cancer biomarkers.

Bsi0077

Bsi0077 is an IgG type monoclonal antibody. The heavy chain variable amino acid sequence is represented in SEQ ID NO: 83 (see fig 8) reproduced below (CDRs

underlined).

EVOLOOSGPVLVKPGASVKISCKTSGYTFTEYTIHWMROSHGKSLEWIGGINPNKG NTNFNOKFKGKATLTVDKSSSTAYMELHSLPSEDSAVFYCARANWDVYAVDSWG QGTSVTVSS Antigen binding was determined by sandwich reaction with BSI0271 and direct binding to purified natural CFH. Peptides bound by Bsi0077 have been identified and verified, using either phage display technique or direct Elisa binding. These peptides are presented as SEQ ID NOs: 64-76.

Furthermore, in histology experiments, it was demonstrated that Bsi0077 stains non small cell lung cancer (NSCLC) as well as carcinomas: 87% of NSCLC, 84% adenocarcinoma, 100% squamous cell carcinoma, and 83% of large cell carcinomas were positive.

BsiO271

BsiO271 is an IgG type monoclonal antibody. Antigen binding was determined by western blot and MS analysis. By mass spectrometry analysis of the antigen, several peptide sequences were obtained (see table below), which demonstrate that Bsi0077 binds to CFH.

K.GEWVALNPLR.K

K.IVSSAMEPDR.E

K.IVSSAMEPDR.E

R.FVCNSGYK.I

R.TGDEITYQCR.N

K.SSNLIILEEHLK.N

R.SSQESYAHGTK.L

R.TGESVEFVCK.R Peptides bound by BsiO271 have been identified and verified, using either phage display technique or direct Elisa binding. These peptides are presented as SEQ ID NOs: 43-56.

Furthermore, in histology experiments, it was demonstrated that Bsi0077 can image non small cell lung cancer (NSCLC) as well as carcinomas: 81% of NSCLC, 84% adenocarcinoma, 60% squamous cell carcinoma, and 67% of large cell carcinomas were positive.

C. Lung Cancer detection Cl - Differential Expression levels

The levels of the biomarkers detected by the 10 mABs was tested in healthy population as compared to lung cancer patients (Fig. 9-16). The performance of the classifiers was further verified, as represented Figure 17.

For each combination of mAbs, the best linear model was calculated using leave-one out cross-validation on a training set, and the accuracy was estimated on a test set of independent samples. The accuracy of the 1013 combinations is plotted as a histogram where the number of panels with specific accuracy is reported on the y-axis (see Fig 18). The number of the occurrences of each antibody in the panels with accuracy higher than 0.8 is plotted on the y-axis.

C2 - Performance of the panel Bsi0272-Bsi0358-Bsi0392

The rationale behind the choice of a panel to be implemented as a diagnostic tool is a combination between its performance in terms of the accuracy as described by the ROC curve of the selected function, as well as the technical robustness of the mAbs in terms of the simplicity of the panel and its reproducibility. The panel shown below answered all these requirements; however is should be stressed again that alternative panels are capable to provide similar performance (see above). The performance of a classifier composed of three antibodies (Bsi0272-Bsi0358-Bsi0392) and a linear model obtained by a linear support vector machine (SVM) with the SMO algorithm (sequential minimal optimization) as classifier are summarized in figure 19.

The optimal threshold (0.115) was calculated from the logistic regression model to provide the optimum combination of sensitivity (80.4%) and specificity (86.2 %). The performance of this classifier to predict patients with different stages of the disease show that its performance increases with the stage of the disease, and even at stage I the sensitivity is 77.3%. Slightly better performance of the classifier should be also noted for patients with squamous cell carcinoma (Table II).

Number of good prediction (total) Sensitivity

LC 176 (219) 80.4 %

Stage I 99 (128) 77.3 %

Stage Il 33 (39) 84.6 %

Stage III 27 (31) 87.1 %

Stage IV 4 (5) 80 %

Adenocarcinoma 66 (85) 77.6 %

Squamous 89 (104) 85.6 %

Other 18 (25) 72 %

C3 - Performance of further panels

The performance of 1013 combinations was tested. Results for 10 illustrative and preferred combinations are disclosed in table 5 below.

Table 5

Table 5: Ten examples of antibody panels, AUC: accuracy, Se: sensitivity, Sp: specificity.

The corresponding ROC curves are presented in Figure 20 where panels are numbered from 1-10 based on this table. The results show the antibodies of the invention differentiate control and lung cancer, either alone or in combinations. These antibodies allow the design of efficient diagnostic products (e.g., devices, kits) to detect, monitor or image cancer in human subjects.

References

1. Stewart, B.W. & Kleihues, P. World cancer report. 1, 181-188 (2003).

2. Rivera, M.P., Detterbeck, F. & Mehta, A.C. Diagnosis of lung cancer: the

guidelines. Chest 123, 129S-136S (2003).

3. Kulpa, J., Wojcik, E., Reinfuss, M. & Kolodziejski, L. Carcinoembryonic antigen, squamous cell carcinoma antigen, CYFRA 21-1, and neuron-specific enolase in squamous cell lung cancer patients. Clin Chem 48, 1931-1937 (2002).

4. ZoIg, W. The proteomic search for diagnostic biomarkers: lost in translation? MoI

Cell Proteomics 5, 1720-1726 (2006).

5. Beane, J., Spira, A. & Lenburg, M.E. Clinical impact of high-throughput gene expression studies in lung cancer. J Thorac Oncol 4, 109-118 (2009).

6. Rifai, N., Gillette, M.A. & Carr, S.A. Protein biomarker discovery and validation: the long and uncertain path to clinical utility. Nat Biotechnol 24, 971-983 (2006).

7. Gao, W.M. et al. Distinctive serum protein profiles involving abundant proteins in lung cancer patients based upon antibody microarray analysis. BMC Cancer 5, 110

(2005).

8. Knezevic, V. et al. Proteomic profiling of the cancer microenvironment by antibody arrays. Proteomics 1, 1271-1278 (2001).

9. Maneewatch, S. et al. Human single-chain antibodies that neutralize homologous and heterologous strains and clades of influenza A virus subtype H5N 1. Antivir

Ther 14, 221-230 (2009).

10. Houimel, M. & Dellagi, K. Isolation and characterization of human neutralizing antibodies to rabies virus derived from a recombinant immune antibody library. J Virol Methods (2009).

11. Uhlen, M. et al. A human protein atlas for normal and cancer tissues based on antibody proteomics. MoI Cell Proteomics 4, 1920-1932 (2005).

12. Blow, N. Antibodies: The generation game. Nature 447, 741-744 (2007).

13. Csanky, E. et al. Monoclonal antibody proteomics: discovery and prevalidation of chronic obstructive pulmonary disease biomarkers in a single step. Electrophoresis 28, 4401-4406 (2007). 14. Peracaula, R., Barrabes, S., Sarrats, A., Rudd, P.M. & de Llorens, R. Altered glycosylation in tumours focused to cancer diagnosis. Dis Markers 25, 207-218 (2008).

15. Kullolli, M., Hancock, W.S. & Hincapie, M. Preparation of a high-performance multi-lectin affinity chromatography (HP-M-LAC) adsorbent for the analysis of human plasma glycoproteins. J Sep Sci 31, 2733-2739 (2008).

16. Kohler, G. & Milstein, C. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256, 495-497 (1975).

17. Kendziorski, C, Irizarry, R.A., Chen, K.S., Haag, J.D. & Gould, M.N. On the utility of pooling biological samples in microarray experiments. Proc Natl Acad Sci USA

102, 4252-4257 (2005).

18. Page, M. & Thorpe, R. Hybridoma Production. The Protein Protocols Handbook, 113-115 (2002).

19. Zhang, J.H., Chung, T.D. & Oldenburg, K.R. A Simple Statistical Parameter for Use in Evaluation and Validation of High Throughput Screening Assays. JBiomol

Screen 4, 67-73 (1999).

20. Shapiro, S.S.a.W., M. B. Biometrika 52, 591-611 (1965).

Table 1. Clinical collections used in the studies.

Table 2. Statistical analysis of the screening data obtained from the validated hits with collection III and IV. The predictive value of each candidate Abs is represented by the p- value from the statistical test. The biological material used in each experiment is supernatant from nascent hybridomas (SN/hybridoma), supernatant from clonal cell line (SN/clone) and purified IgG from the monoclonal cell line.

Table 3. Epitope and functional redundancy of the LC candidates estimated from the phage display (upper-right side of the matrix) and screening data (lower left side of the data). clone BSI0033 BSI0068 BSI0070 BS 10071 BSIOO 72 BSI0077 BSI0080 BSIO 086 BSIOO 88 BSI0270 BSI0272 BSI0346 BSI0349 BSI0351 BSI0352 BSI0358 BSI0359

# seq 3 19 10 0 3 14 3 8 4 1 11 8 10 2 10 13 4

BS 10033 3 5% 4% - 1 % 5% 6% 3% 6% 9% 6% 3% 3% 3% 2% 1% 2%

BS 10068 19 -3,90 4% - 7% 13% 7% 3% 10% 5% 5% 3% 6% 3% 3% 4% 2%

BS 10070 10 -1,43 0,13 - 4% 4% 2% 4% 4% 6% 5% 4% 4% 5% 4% 3% 3%

BS 10071 0 0,75 -0,1 1 -0,82 - - - - - - - - - - - - -

BSIOO 72 3 -1,34 -0,07 0,81 -1 ,59 3% 5% 4% 6% 2% 3% 7% 6% 7% 5% 5% 2%

BSIOO 77 14 -3,28 -0,01 -2,19 -3,83 -2,88 9% 3% 10% 7% 6% 2% 5% 1% 2% 3% 2%

BSIOO 80 3 -12,96 -1 ,95 -10,40 -16,32 -12,75 -2,79 7% 41 % 2% 11 % 4% 4% 4% 5% 3% 5%

BSIOO 86 8 -3,42 -0,06 -2,67 -4,31 -3,29 0,69 0,01 5% 8% 5% 5% 2% 26% 21% 5% 3%

BSIOO 88 4 -8,19 -0,88 -6,07 -11 ,02 -8,06 -1 ,94 0,50 -1,76 3% 10% 3% 4% 5% 4% 2% 4%

BSI0270 1 -2,97 -0,08 -3,43 -4,64 -4,44 -0,23 -0,02 0,01 0,02 4% 2% 0% 8% 9% 1% 4%

BSI0272 11 -1,69 0,35 -1 ,1 1 -2,72 -2,00 -0,13 0,00 -0,08 0,04 0,32 4% 3% 3% 5% 3% 6%

BSI0346 8 -3,01 0,35 -1 ,45 -3,85 -2,45 0,51 -0,06 0,51 0,06 0,06 0,01 3% 8% 6% 2% 1 %

BSI0349 10 -3,72 -0,13 -2,91 -4,49 -3,71 0,64 -0,05 0,68 -0,31 0,05 -0,56 0,39 2% 3% 14% 8%

BS 10351 2 -3,46 0,15 -2,67 -4,37 -3,59 -0,17 0,10 -0,17 -0,07 0,08 -0,08 -0,33 0,13 22% 7% 1 %

BSI0352 10 -2,49 0,14 -1 ,67 -3,22 -2,34 -0,09 0,06 0,01 0,10 0,10 0,02 -0,09 0,18 0,77 6% 3%

BSI0358 13 -3,43 -0,14 -2,62 -4,21 -3,63 0,17 0,06 0,09 -0,10 -0,03 -0,11 -0,04 0,52 0,46 0,20 7%

BSI0359 4 -2,64 0,36 -1 ,02 -3,54 -1 ,98 0,31 -0,01 0,30 0,08 0,18 0,30 0,63 0,46 0,33 0,26 0,57

BSI0392 0 0,05 -0,43 0,03 0,00 0,00 -0,20 -3,15 -0,21 -1 ,59 0,25 0,31 -0,12 -0,05 0,36 0,46 0,21 0,26

The epitope redundancy is expressed as percentage of identical residues in the aligned set of peptide sequences between two antibodies (see Methods). The functional redundancy is expressed as the adjusted chi-square of fitting the correlation of the responses of two antibodies with the 610 clinical samples measured in the HT-ELISA screening to a linear function (see

Methods). The adjusted chi-square for screening with two identical antibodies (see Figure 5B) is 0.73.

Table 6

Peptides bound by further antibodies of the invention. % represent peptide sequence occurrence in the tested group.

Claims

1. An antibody which binds a peptide having a sequence selected from SEQ ID NOs:

1-76, or a fragment or derivative of such an antibody having essentially the same antigen specificity.

2. The antibody, fragment or derivative of claim 1, which binds a polypeptide comprising a peptide sequence selected from SEQ ID NOs: 1-76.

3. The antibody, fragment or derivative of claim 1, which binds a peptide having a sequence selected from SEQ ID NOs: 26-35 and 57-63, and which also binds a human LRGl protein.

4. The antibody, fragment or derivative of claim 1, which binds a peptide having a sequence selected from SEQ ID NOs: 14-25, and which also binds a human C9 protein.

5. The antibody, fragment or derivative of claim 1, which binds a peptide having a sequence selected from SEQ ID NOs: 36-38 and which also binds a human haptoglobin protein and/or haptoglobin related protein.

6. The antibody, fragment or derivative of claim 1, which binds a peptide having a sequence selected from SEQ ID NOs: 43-56 and 64-76, and which also binds a human CFH protein.

7. The antibody, fragment or derivative of claim 1, which binds a peptide having a sequence selected from SEQ ID NOs: 1-13 and 39-42 and which also binds a human Alpha- 1 -antichymo trypsin protein.

8. A polypeptide which comprises all or part of a sequence selected from SEQ ID NOs: 77-92, said part comprising at least 5 consecutive amino acid residues of the reference sequence.

9. An antibody of claim 1, selected from BsiO358, BsiO359, BsiO272, BsiO392, BsiOO8O, BsiO352, Bsi0077, BsiO349, Bsi0270, BsiO271, Bsi0072, Bsi0076, BsiOO68, BsiOO33, Bsi0071, Bsi351, or Bsi0070, or a fragment or derivative thereof having essentially the same antigen specificity.

10. The antibody of any one of claims 1 to 8, which is a mouse, human, chimeric or humanized antibody.

11. An isolated nucleic acid encoding a polypeptide of claim 8.

12. An isolated nucleic acid encoding the variable region of an antibody of any one of claims 1 to 10.

13. A hybridoma cell producing an antibody of claim 1.

14. A recombinant cell containing a nucleic acid of claim 11 or 12.

15. A method for detecting lung cancer in a subject, the method comprising contacting a sample from said subject, preferably a blood sample, with at least one antibody of any one of claims 1 to 10 and determining the presence of an antigen bound to said at least one antibody, said presence being indicative of lung cancer.

16. The method of claim 15, which comprises contacting said sample from said subject with at least two antibodies of any one of claims 1 to 10, preferably at least 3, in combination.

17. The use of an antibody of any one of claims 1 to 10 for in vitro or ex vivo histopatho logical classification or immunostaining of cancers.

18. A device comprising at least one antibody of any one of claims 1 to 10 immobilized on a support.

19. The device of claim 18, wherein the support is a membrane, a slide, a microarray, a chip or a microbead.

20. A method for detecting lung cancer in a subject, the method comprising contacting a sample from said subject, preferably a blood sample, with at least one antibody, fragment or derivative thereof, that binds a protein selected from Leucine -Rich alpha-2 glycoprotein, haptoglobin, C9, CFH or Alpha 1 Antichymotrypsin, and determining the presence of a binding, said presence being indicative of lung cancer.

21. A device comprising at least one antibody, fragment or derivative thereof, that binds a protein selected from Leucine-Rich alpha-2 glycoprotein, haptoglobin, C9, CFH or Alpha 1 Antichymotrypsin, immobilized on a support.

22. The device of claim 21 , wherein the support is a membrane, a slide, a microarray, a chip or a microbead.

23. A kit comprising a device of claim 21 or 22 and a reagent to perform or detect (or quantify) an immune reaction, particularly an antibody-antigen complex.