WO2021250323A1 - Anti-cox-2 autoantibody as a diagnostic marker, and methods, kits and uses related thereto - Google Patents

Abstract

The present invention relates to the fields of life sciences, medicine, autoantibodies and in vitro diagnostics. Specifically, the invention relates to a method for determining a bone marrow failure disorder of a subject or the absence thereof, or subclassifying a subject having a bone marrow failure disorder, or to a method for determining a specific autoantibody or a polynucleotide encoding said autoantibody or a fragment thereof of a sample. Still, the present invention relates to a specific autoantibody or a polynucleotide encoding said autoantibody or a fragment thereof e.g. for use as a diagnostic marker or for determining a bone marrow failure disorder of a subject. Also, the present invention relates to a test kit for determining a bone marrow failure disorder or for determining a specific autoantibody or a polynucleotide encoding said autoantibody or a fragment thereof of a sample. And still, the present invention relates to use of the specific autoantibody or a polynucleotide encoding said autoantibody or a fragment thereof as a diagnostic marker or use of the test kit or the specific autoantibody or a polynucleotide encoding said autoantibody or a fragment thereof for determining a bone marrow failure disorder of a subject.

Description

Anti-COX-2 autoantibody as a diagnostic marker, and methods, kits and uses related thereto

FIELD OF THE INVENTION

The present invention relates to the fields of life sciences, medicine, autoantibod ies and in vitro diagnostics. Specifically, the invention relates to a method for de termining a bone marrow failure disorder of a subject or the absence thereof, or subclassifying a subject having a bone marrow failure disorder, or to a method for determining a specific autoantibody or a polynucleotide encoding said autoanti body or a fragment thereof of a sample. Still, the present invention relates to a specific autoantibody or a polynucleotide encoding said autoantibody or a frag ment thereof e.g. for use as a diagnostic marker or for determining a bone marrow failure disorder of a subject. Also, the present invention relates to a test kit for de termining a bone marrow failure disorder or for determining a specific autoantibody or a polynucleotide encoding said autoantibody or a fragment thereof of a sample. And still, the present invention relates to use of the specific autoantibody or a pol ynucleotide encoding said autoantibody or a fragment thereof as a diagnostic marker, or use of the test kit or the specific autoantibody or a polynucleotide en coding said autoantibody or a fragment thereof for determining a bone marrow failure disorder of a subject.

BACKGROUND OF THE INVENTION

Bone marrow failure (BMF) disorders are rare diseases characterized by an inabil ity to make enough blood - either red cells, which carry oxygen; white cells, which fight infection; or platelets, which help the blood clot. Bone marrow failure disor ders may be either inherited or acquired.

Aplastic anemia (AA) is a rare bone marrow failure (BMF) disorder that is charac terized by loss of all hematopoietic lineages (pancytopenia) (Young NS. The New England journal of medicine, 2018;379(17):1643-1656). The disease can develop via three alternate routes: chemical /physical insults (including radiation and toxic agents), hereditary genetic defects or via immune mediated mechanisms. Sporad ic AA cases without family history of without documented chemical exposure are considered as idiopathic, this group constituting the largest patient group. AA treatments include the use of immunosuppressive agents (anti-thymocyte globu lins, cyclosporin A, eltrombopag), transfusions and allogenic bone marrow trans- plantation. In addition to the life threatening cytopenias, AA patients have signifi cantly elevated risk of developing myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML) (Young NS. The New England journal of medicine, 2018;379(17):1643-1656).

The immunological pathomechanisms of idiopathic i.e. immune mediated AA are not known in detail. Cytotoxic CD8+ T cells have been shown to display an acti vated phenotype in AA patients (Zoumbos NC et al. The New England journal of medicine, 1985;312(5):257-265; Hosokawa K et al. Journal of immunology (Balti more, Md.: 1950), 2016; 196(4):1568-1578). Furthermore, CD4+ regulatory T cell signatures can be used to predict therapy responses in idiopathic AA thus assign ing Tregs as important disease regulators in AA (Kordasti S et al. Blood, 2016;128(9):1193-1205). The importance of non-hematopoietic factors in AA is highlighted by the recent finding that bone marrow mesenchymal stem cells have the capacity to regulate Treg/ Th17 balance in AA (Li H et al. Scientific Reports, 2017;7(1):42488). Repeated autoantibody findings point to the importance of the humoral immunity and more specifically to the role of B-cells and their supporting CD4+ T helper cells in the induction of organ specific autoimmunity in AA. Autoan tibodies have been detected against kinectin (Hirano N. Blood, 2003;102(13):4567-4575; Hirano N et al. British Journal of Haematology, 2005;128(2):221-223), DRS-1 (Feng X. Blood, 2004;104(8):2425-2431) CA-1 (Jankovicova B et al. Immunology letters, 2013;153(1-2):47-49; Lakota J et al. Acta haematologica, 2012;128(3):190-194), hnRNP K (Qi Z et al. Annals of hema tology, 2010;89(12):1255-1263), CLIC1, HSBP11, RSP27 (Goto M et al. British Journal of Haematology, 2013;160(3):359-362) Moesin (Takamatsu H et al. Blood, 2007;109(6):2514-2520) and PMS-1 (Hirano N et al. British Journal of Haematology, 2005;128(2):221-223). The majority of these antibodies are not specific for AA or BMF diseases.

BMF disorders are challenging to diagnose due to overlapping findings and gradu al transitions from one condition to another. For example, in clinical praxis AA di agnosis is set based on bone marrow morphology, peripheral blood counts, cyto genetics and genetic deep sequencing analyses including both germ line and so matic variant analysis. This is an expensive and time-consuming process. Fur thermore, none of the earlier reported autoantibodies with BMF or AA association have been adapted to routine diagnostic use. There remains a significant unmet need for simple, effective and specific diagnos tic markers and methods for determining BMF disorders.

BRIEF DESCRIPTION OF THE INVENTION

It has now been found a novel diagnostic biomarker, an autoantibody, which is able to reveal BMF disorders. Said autoantibody is surprisingly an anti prostaglandin G/H synthase 2 or cyclooxygenase-2 (COX-2) binding autoantibody.

The present invention, namely an anti-COX-2 autoantibody and methods and uses related thereto, is able to overcome the defects of the prior art including but not limited to a lack of specific and/or sensitive, and excellent diagnostic biomarkers and methods for determining a BMF disorder.

Surprisingly, in addition to AA, positive anti-COX-2 autoantibody findings were identified in patients with other BMF diseases. Indeed, the inventors of the present disclosure have been able to identify an anti-COX-2 autoantibody that is highly specific for BMF disorders. Interestingly, autoantibody positivity associated tightly with the MHC II HLA DRB1 15:01 haplotype, linking the novel autoantibody to the function of CD4+ T helper cells and thus suggesting a potential association to BMF pathogenesis.

Anti-COX-2 autoantibody, and simple and cost-effective methods and diagnostic kits of the present invention enable determination of a BMF disorder from a sam ple obtained from a subject e.g. within less than 4 hours e.g. within less than one hour.

Patients with a BMF disorder can have an excellent chance of living out a normal life or getting accurate treatment if properly diagnosed. Therefore, in addition to providing tools and kits for diagnostic purposes the present invention further ena bles fast and effective monitoring of the BMF disorder as well as timely and effec tive treatment of patients thereby lengthening the life and improving the way of life of BMF patients. One example of an enhanced diagnostic pathway enabled by the present invention is illustrated in Figure 9.

Specifically, the present invention relates to a method for determining a bone mar row failure disorder of a subject or the absence thereof, or subclassifying a subject having a bone marrow failure disorder, wherein the method comprises determining an anti-COX-2 autoantibody or a polynucleotide encoding said anti-COX-2 autoan tibody or a fragment of said autoantibody or polynucleotide from a sample of a subject.

Also, the present invention relates to an anti-COX-2 autoantibody or a polynucleo tide encoding said anti-COX-2 autoantibody or a fragment of said autoantibody or polynucleotide as a diagnostic marker, for determining a bone marrow failure dis order of a subject, and/or for use as a diagnostic marker.

Also, the present invention relates to an anti-COX-2 autoantibody or a polynucleo tide encoding said anti-COX-2 autoantibody or a fragment of said autoantibody or polynucleotide, wherein said autoantibody is capable of binding one or more ami no acids at positions 1 - 604, 56 - 604, 100 - 604, 200 - 604, 300 - 604, 350 - 604, 442 - 604, 490 - 604, 530 - 604 or 490 - 600 as shown in SEQ ID NO: 1.

Still, the present invention relates to a test kit for determining a bone marrow fail ure disorder, or for determining the anti-COX-2 autoantibody or a polynucleotide encoding said anti-COX-2 autoantibody or a fragment of said autoantibody or pol ynucleotide of a sample, wherein the kit comprises one or more tools for determin ing the anti-COX-2 autoantibody of claim 4 or a polynucleotide encoding said anti- COX-2 autoantibody or a fragment of said autoantibody or polynucleotide from a sample of a subject, and optionally reagents for the test; and/or the kit comprises a capture complex for binding an anti-COX-2 autoantibody, or a polynucleotide encoding said anti-COX-2 autoantibody or a fragment of said auto antibody or polynucleotide.

Still furthermore, the present invention relates to a method for determining an anti- COX-2 autoantibody or a polynucleotide encoding said anti-COX-2 autoantibody or a fragment of said autoantibody or polynucleotide from a sample of a subject, wherein the method comprises allowing a capture complex to bind with an anti- COX-2 autoantibody or a polynucleotide encoding said anti-COX-2 autoantibody or a fragment of said autoantibody or polynucleotide, and determining or detecting said anti-COX-2 autoantibody, polynucleotide or fragment or amount thereof.

Still furthermore, the present invention relates to use of the test kit of the present invention or anti-COX-2 autoantibody (e.g. of the present invention) or a polynu cleotide encoding said anti-COX-2 autoantibody or a fragment of said autoanti body or polynucleotide for determining a bone marrow failure disorder of a subject. And still furthermore, the present invention relates to use of an anti-COX-2 autoan tibody (e.g. of the present invention) or a polynucleotide encoding said anti-COX-2 autoantibody or a fragment of said autoantibody or polynucleotide as a diagnostic marker.

Other objects, details and advantages of the present invention will become appar ent from the following drawings, detailed description and examples. The objects of the invention are achieved by methods, biological compounds, kits and uses characterized by what is stated in the independent claims. The preferred embodiments of the invention are disclosed in the dependent claims.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 shows autoantibody screening results. Autoantibody screening results from patients with large granular lymphocyte (LGL) leukemia (n=12), immune aplastic anemia (n=7) and rheumatoid arthritis (n=10) were compared to the mean values from HC (n= 5) and all values that were more than 10 times higher than the mean of FC were considered as positive. Heat map presents data from all individ ual antibodies that were positive in at least two individual AA patients.

Figure 2 shows measurements of aCOX-2 antibodies in plasma samples by DELFIA assay in different cohorts. Dashed line denotes the statistical cutoff value determined from combined cohort data. Under each diagnosis group are annotat ed the number of positive patients, the total number of patients and the positive percentage. (A) Results in the Finnish study cohort. (B) Combined results in the two United States study cohorts. (C) Combined results in the two Japanese study cohorts. (D) Results in the control cohort obtained from Helsinki Biobank. IAA = immune aplastic anemia, ITP = immune thrombocytopenic purpura, PNH = parox ysmal nocturnal hemoglobinuria, ICUS/CHIP = idiopathic cytopenia of undeter mined significance / clonal hematopoiesis of indeterminate potential, GVHD = graft versus host disease, PRCA = pure red cell aplasia. Figure 3 shows measurements of aCOX-2 antibodies in patient plasma by DELFIA. (A) ROC curve including all IAA patients. (B) ROC curve including only HLA-DRB1^*15 -positive IAA patients. (C) Results from combined Finnish, United States and Japanese study cohorts, excluding the Helsinki Biobank control cohort. Under each diagnosis group the number of positive patients, the total number of patients and the positive percentage are annotated. Dashed line denotes the sta tistical cutoff value determined from combined cohort data. (D) Follow-up sample results from n = 9 AA patients. Each line denotes a separate patient. AUC = area under curve, IAA = immune aplastic anemia, MDS = myelodysplastic syndrome, ITP = idiopathic thrombocytopenia, PNH = paroxysmal nocturnal hemoglobinuria, ICUS/CHIP = idiopathic cytopenia of undetermined significance / clonal hemato poiesis of indeterminate potential, LGLL = large granular lymphocyte leukemia, RA = rheumatoid arthritis, GVHD = graft versus host disease, PPRCA = pure red cell aplasia.

Figure 4 reveals immunoglobulin isotypes in IAA patients. Each isotype was de termined separately by DELFIA for a selection of IAA patients, columns are indi vidual cases. Each isotype level was compared to corresponding subclass mean FIC in standard deviations (SD). Heatmap color ramp starts at a set level of mean HC + 5 SDs and is capped at +20 SDs. (A) Comparison of IgG, IgM and IgA in n = 24 IAA patients. Healthy controls n = 17. (B) IgG isotypes 1 - 4 in n = 38 aCOX-2 Ab positive cases and one Ab negative control patient (total n = 39). Healthy con trols n = 30.

Figure 5 shows epitope mapping. High resolution epitope mapping was performed to identify antigen epitopes from ten (n=10) COX-2 autoantibody positive and from two (n=2) autoantibody negative AA plasma samples. (A) Linear 15 amino acid peptides with 14 amino acid overlap covering the whole COX-2 protein were spot ted on a microchip. Antibody binding to the linear peptide was detected with Goat anti-human IgG (Fc) DyLight6 and LI-COR Odyssey Imaging System. Scanning values are reported as fluorescence intensities (a.u.). (B) Autoantibody samples from which no clear epitope could be identified using linear peptides, were sub jected for conformational peptide mapping. Here all peptides were cyclized by us ing a thioether linkage linking the amino and carboxy terminals. The peptides were spotted on a microarray as 10-mers (7 and 13 -mers) with a peptide-peptide over lap of n-1 amino acids. Signal intensities are reported as fluorescence intensities (a.u.). (C) Visual summary of all epitope findings reveals a shared antigenic pro tein sequence that is localized between amino acids 490 and 596 in the C-terminal part of COX-2.

Figure 6 shows comparison of clinical findings between aCOX-2 Ab negative and positive IAA patients. (A) Difference between blood measurement levels around the time of diagnosis. ALC = absolute lymphocyte count, ANC = absolute neutro phil count, ARC = absolute reticulocyte count, Hb = hemoglobin, Pit = platelets, WBC = total white blood cells. Boxes present the 25-75-percentile range (inter quantile range, IQR) with median. Whiskers show the furthest value within 1.5x IQR of the 25- or 75-percentile limits respectively, points outside these limits can be considered outliers. Unadjusted p-values for Mann-Whitney U test between means are shown. (B) Distributions of patient age at time of diagnosis. Distribu tions were tested for similar distribution using two-sided two-sample Kolmogorov- Smirnov test. (C) Prevalence of HLA-DRB1 ^*15:01 or 15:02 in AA patients (n = 144 for aCOX-2 Ab negative and n = 90 for positive). The p-value for Pearson’s Chi- squared test is shown.

Figure 7 shows COX-2 expression in healthy controls and in BMF patients’ bone marrow. A: Single cell analysis of healthy bone marrow reveals clusters of all main cell lineages. B: COX-2 expression over the clusters. C: Clustering of BMF bone marrow CD34+ cells D: The main genes differentiating the clusters. E: Original identity of the cells in each cluster indicated by disease group. F: Sample level original identity of the cells in each cluster. G: COX-2 expression in each cluster, healthy controls compared to BMF diseases. H: COX-2 expression in FISC cluster. BMF = bone marrow failure.

Figure 8 shows an example of tools for detecting anti-COX-2 antibody in the method or kit of the present invention. Anti-COX-2 antibody can be bound to a capture complex (e.g. a (recombinant) target protein shown in the figure) optionally immobilized on the surface. The combination of an anti-COX-2 antibody and cap ture complex can be determined via an indicator, such as a detection antibody, which is optionally labelled.

Figure 9 shows enhanced diagnostic pathway enabled by the present invention.

Figure 10 shows the measurements of aCOX-2 antibodies in plasma samples by DELFIA assay in different cohorts of adult (over 18 years old) IAA patients. Re sults of the United States (USA), the Japanese (JPN) and the Finnish (FIN) patient groups are annotated with the number of positive patients, the total number of pa tients and the percentage of aCOX-2 Ab positive patients in the cohort.

Figure 11 shows the measurements of aCOX-2 antibodies in plasma samples by DELFIA assay in different control groups. Under each group are annotated the number of positive patients, the total number of patients and the percentage of aCOX-2 Ab positive individuals in each group. IAA = immune aplastic anemia, HBP = Helsinki Biobank, PNH = paroxysmal nocturnal hemoglobinuria, MDS = myelodysplastic syndrome, ICUS/CHIP = idiopathic cytopenia of undetermined significance / clonal hematopoiesis of indeterminate potential, ITP = immune thrombocytopenic purpura, PRCA = pure red cell aplasia, LGLL=large granular lymphocyte leukemia, RA = rheumatoid arthritis, MA = multiple sclerosis, DM1 = diabetes mellitus, Misc Al = miscellaneous autoimmune diseases, GVHD = graft versus host disease, NON-AI = non-autoimmune diseases.

Figure 12 shows aCOX-2 Ab levels in a series of follow-up samples from n = 21 diagnostic phase IAA patients. Each line denotes the autoantibody levels in a sep arate patient. Measurements of aCOX-2 antibodies in patient plasma have per formed by the DELFIA assay. Dg = diagnosis, mo = month.

Figure 13 shows the percentages of aCOX-2 autoantibody positive and negative IAA patients in different age groups. The age groups are divided in intervals of 10 years starting from the age of 20 years. The number of aCOX-2 positive individu als in each age group is indicated under the age group information.

Figure 14 shows the genotype distribution of both aCOX-2 Ab positive and nega tive IAA patients in regard of the HLA-DRB1 ^*15:01 locus.

Figure 15 shows the platelet accounts at diagnosis of aCOX-2 Ab negative and positive IAA patients.

SEQUENCE LISTING

SEQ ID NO: 1 shows the amino acid sequence of the target polypeptide PGH2JHUMAN Prostaglandin G/H synthase 2 (COX-2) bound by the anti-COX-2 autoantibody (uniprot P35354, https://www.uniprot.org/uniprot/P35354)

SEQ ID NO: 2 shows the fragment of the target polypeptide (COX-2), which can be bound by the anti-COX-2 autoantibody (PALLVE, e.g. starting from amino acid 491 of SEQ ID NO: 1).

SEQ ID NO: 3 shows the fragment of the target polypeptide (COX-2), which can be bound by the anti-COX-2 autoantibody (STFGGEV, e.g. starting from amino ac id 534 of SEQ ID NO: 1). SEQ ID NO: 4 shows the fragment of the target polypeptide (COX-2), which can be bound by the anti-COX-2 autoantibody (GFQIINT, e.g. starting from amino acid 540 of SEQ ID NO: 1).

SEQ ID NO: 5 shows the fragment of the target polypeptide (COX-2), which can be bound by the anti-COX-2 autoantibody (QIINTAS, e.g. starting from amino acid 542 of SEQ ID NO: 1). SEQ ID NO: 6 shows the fragment of the target polypeptide (COX-2), which can be bound by the anti-COX-2 autoantibody (VTIN, e.g. starting from amino acid 577 of SEQ ID NO: 1).

SEQ ID NO: 7 shows the fragment of the target polypeptide (COX-2), which can be bound by the anti-COX-2 autoantibody (GLDDIN, e.g. starting from amino acid 587 of SEQ ID NO: 1)

SEQ ID NO: 8 shows the fragment of the target polypeptide (COX-2), which can be bound by the anti-COX-2 autoantibody (DINPTVL, e.g. starting from amino acid 590 of SEQ ID NO: 1).

SEQ ID NO: 9 shows the fragment of the target polypeptide (COX-2) (NASSSRSGLDDINPT, e.g. starting from amino acid 580 of SEQ ID NO: 1). The anti-COX-2 autoantibody can bind one or more amino acids of said fragment.

DETAILED DESCRIPTION OF THE INVENTION

The present invention concerns a method for determining a BMF disorder of a subject or the absence thereof, or subclassifying a subject having a BMF disorder. Anti-COX-2 autoantibodies have now been found to associate with BMF disorders and therefore, the method comprises determining an anti-COX-2 autoantibody (or a polynucleotide encoding said anti-COX-2 autoantibody or a fragment of said au toantibody or polynucleotide) from a sample of a subject. In the present disclosure “an/the anti-COX-2 autoantibody” or “(the) anti-COX-2 autoantibodies” can also be used for “a/the polynucleotide(s) encoding said anti- COX-2 autoantibody or a/the fragment(s) of said autoantibody or polynucleotide”. Bone marrow failure is the reduction or cessation of blood cell production affecting one or more cell lineages. Bone marrow failure disorders can be defined as a group of disorders characterized by an inability to make enough blood - either red cells, which carry oxygen; white cells, which fight infection; or platelets, which help the blood clot. In one embodiment the bone marrow failure disorder is inherited or acquired.

As used herein “subclassifying a subject having a BMF disorder” refers to classify ing a subject having a BMF disorder to a more specific subgroup within the group of patients having a BMF disorder. Said more specific subgroup can be for exam ple a specific genetic group or a very specific disease subgroup (e.g. (immune) aplastic anemia or any other specific disorder belonging to BMF disorders).

In one embodiment the bone marrow failure disorder or a subgroup of the bone marrow failure disorder is selected from the group comprising or consisting of aplastic anemia (AA), paroxysmal nocturnal hemoglobinuria (PNFI), polycythemia vera (PV), Diamond-Blackfan anemia (DBA), essential thrombocytosis (ET), Fan- coni anemia (FA), large granular lymphocyte (LGL) leukemia, myelodysplastic syndrome (MDS), telomere syndromes, dyskeratosis congenita, immune thrombo cytopenic purpura (ITP), idiopathic cytopenia of undetermined significance (ICUS), clonal hematopoiesis of indeterminate potential (CHIP) and idiopathic bone mar row dysplasia of undetermined significance (IDUS). In a specific embodiment the bone marrow failure disorder or the subgroup of the bone marrow failure disorder is aplastic anemia (AA).

Aplastic anemia (AA) is a form of bone marrow failure. The bone marrow is the factory for all blood cell lines i.e. red cells, white cells, and platelets. These cells are critical to supplying oxygen and nourishment to other tissues and organs, fighting infection, and clotting the blood. Typically, as old blood cells die off natu rally, they are replaced by new blood cells formed in the bone marrow. In aplastic anemia, the bone marrow does not produce new cells, leaving the body suscepti ble to bleeding and infection. The most common cause of aplastic anemia is that the immune system attacks the stem cells of the bone marrow (immune aplastic anemia). Other factors that can injure bone marrow and affect blood cell produc tion include but are not limited to radiation and chemotherapy treatments, expo sure to toxic chemicals, use of certain drugs, autoimmune disorders, viral infec tions (e.g. hepatitis, Epstein-Barr, cytomegalovirus, parvovirus B19, FI IV), and un known factors. Paroxysmal nocturnal hemoglobinuria (PNH) is a rare, acquired, disease that is caused by a mutation in bone marrow stem cells. The disease is characterized by destruction of red blood cells (hemolytic anemia), blood clots (thrombosis), im- paired bone marrow function, and a 3% to 5% lifetime risk of developing leukemia. PNH affects only one or two people per million of the population.

Polycythemia vera (PV) is a rare blood disorder in which there is an increase in all blood cells, particularly red blood cells. The increase in blood cells makes the blood thicker, leading to strokes or tissue and organ damage.

Diamond-Blackfan anemia (DBA) is a rare blood disorder in which the bone mar row does not make enough red blood cells to carry oxygen throughout the body. It is associated with birth defects or abnormal features.

Essential thrombocytosis (ET), or primary thrombocythemia, is a rare disorder in which the body produces too many platelets for unknown reasons. This can cause abnormal blood clotting or bleeding. Fanconi anemia (FA) is a rare, inherited blood disorder that prevents the bone marrow from producing enough new blood cells for the body to function properly, or that causes the bone marrow to make faulty blood cells.

Large granular lymphocyte (LGL) leukemia is a rare cancer of white blood cells called lymphocytes, which originate in the lymph system and bone marrow and help fight infection. In people with the disease, the lymphocytes are enlarged and contain granules, which can be seen when the blood is examined under the micro scope. There are two types of LGL leukemia: T-cell (T-LGL) and natural killer cell (NK-LGL). Each type may be chronic (slow-growing) or aggressive (fast- growing).

Myelodysplastic Syndrome (MDS) refers to a group of disorders in which the bone marrow produces too few mature and/or functioning red blood cells, white blood cells or platelets. It begins with a change to a normal stem cell in the bone marrow.

Telomere syndromes are inherited conditions that can cause bone marrow failure and lung disease. These syndromes vary in severity and can affect children and adults. In rare cases, a patient’s telomere syndrome may appear as a condition called dyskeratosis congenita. Said condition is characterized by abnormal find ings in the skin, mouth and nails.

Immune thrombocytopenic purpura (ITP) is a bleeding disorder in which the im mune system destroys platelets, which are necessary for normal blood clotting.

A patient with an idiopathic cytopenia of undetermined significance (ICUS) has a decrease in peripheral blood counts (e.g. causing either anemia, leucopenia and thrombocytopenia) and age-related hematopoietic clones driven by mutations of genes that are recurrently mutated in myeloid neoplasms and associated with in crease in the risk of hematologic cancer. Clonal hematopoiesis (CH) is a term most commonly used to refer to a population of related myeloid cells with an ac quired gene mutation. CH is a characteristic of myelodysplastic syndromes and leukemias, but it may also be found in some individuals who have no detectable hematologic malignancy; in such cases it may be referred to as clonal hematopoi esis of indeterminate potential (CHIP). Clonal hematopoiesis of indeterminate po tential does not include monoclonal B cell lymphocytosis or other lymphoprol itera tive disorders.

Patients with dysplastic bone marrow features with or without a karyotype, who have only mild if any cytopenia can be diagnosed to have an idiopathic bone mar row dysplasia of undetermined significance (IDUS).

In one embodiment of the invention the presence or an increased level of the anti- COX-2 autoantibody in the sample of the subject indicates the bone marrow failure disorder of the subject, or the presence or an increased level of the anti-COX-2 autoantibody in the sample of the subject indicates a subgroup of bone marrow failure disorders; and/or the absence of the anti-COX-2 autoantibody in the sample of the subject indicates lack of the bone marrow failure disorder of said subject. As used herein “an increased level” refers to a situation, wherein the level of anti- COX-2 autoantibodies in a sample is increased compared to a normal sample without a BMF disorder or to a level of normal controls (without a BMF disorder). An increase of the level of anti-COX-2 autoantibodies is preferably a significant in crease. In one embodiment the anti-COX-2 autoantibody levels are significantly different between patients and controls. In some cases, very low levels of anti-COX-2 autoantibodies can be present in samples of subjects without a BMF disorder. In those cases, the anti-COX-2 auto antibody levels can however differ significantly from the patient levels.

In one embodiment the levels of anti-COX-2 autoantibodies in the sample of the subject are compared to the level or level range of said antibodies in a control sample; or the levels of said anti-COX-2 autoantibodies in the sample of the sub ject are compared to the normal levels or normal level ranges of said antibodies determined from a set of controls. For example, the absolute values of autoanti bodies in a sample can be considered, e.g. against a control value or range. In one embodiment the obtained autoantibody level(s) of the sample(s) differ(s) sig nificantly from the control or normal levels or ranges thereby indicating a BMF dis order; or the obtained autoantibody level(s) of the sample(s) differ(s) do not differ significantly from the control or normal levels or ranges thereby indicating the ab sence of a BMF disorder.

In one embodiment a cut-off value to differentiate positive and negative samples can be selected e.g. from cut-off for sample positivity FC5 - 20 such as FC8 - 15 or FC10 compared to the average of healthy controls. In one embodiment the cut off value can be calculated according to Findcutoffs (Chang C et al. 2017, PLoS ONE. 12(4):e0176231) and/or OptimalCutPoint (Lopez-Raton M et al. OptimalCut- points: An R Package for Selecting Optimal Outpoints in Diagnostic Tests. 2014).

As used herein “significant” refers to statistically significant i.e. p< 0.05. Statistical methods suitable for the present invention are any common statistical methods known to a person skilled in the art. In a specific embodiment of the invention the statistical method for determining an increase, significant increase, decrease or significant decrease in the autoantibody level or amount includes but is not limited to a t-test, modified t-test, Shrinkage t-test, Fischer’s exact test, one-way ANOVA and Dunnett’s multiple comparison test.

In one embodiment the method comprises determining the presence, absence, amount or level of anti-COX-2 autoantibodies. The presence, absence, amount or level of the autoantibodies in a sample can be detected or measured by any suita ble method known in the art. Determinations, detections or measurements suitable for the present invention can either directly or indirectly reveal the autoantibodies. In addition to studying the presence, absence, amount or level of the specific poly peptides (anti-COX-2 autoantibodies), or as an alternative to it, for example, ex- pression of the polynucleotides encoding said autoantibodies or expression of said polypeptides (autoantibodies) can be utilized for determining the presence of anti- COX-2 autoantibodies or anti-COX-2 antibody producing B cells. Non-limiting ex amples of suitable detection methods include polypeptide-based methods (e.g. polypeptides capable of binding an anti-COX-2 autoantibody or a fragment there of), polynucleotide- or nucleic acid-based methods (e.g. polynucleotides capable of binding a polynucleotide encoding an anti-COX-2 autoantibody of a fragment thereof), staining methods (e.g., fluorescent staining), hybridization methods, and any combination thereof. For example, RNA and/or DNA -based methods are suit able nucleic acid methods for the present invention and include but are not limited to PCR methods (e.g., PCR, qPCR, RT-PCR), sequencing methods (e.g. basic sequencing methods, large-scale sequencing, high-throughput methods) and hy bridization methods (e.g. microarrays).

In the present disclosure, the term “polypeptide” refers to polymers of amino acids of any length. Furthermore, as used herein “a polynucleotide” refers to any poly nucleotide, such as single or double-stranded DNA (genomic DNA or cDNA) or RNA (e.g. mRNA or rRNA), including but not limited to a nucleic acid sequence encoding a polypeptide in question or a conservative sequence variant thereof. Conservative nucleotide sequence variants (i.e. nucleotide sequence modifica tions, which do not significantly alter biological properties of the encoded polypep tide) include variants arising from the degeneration of the genetic code and from silent mutations. “A fragment of a polypeptide or polynucleotide” refers to a frag ment of any length, e.g. any part of a polypeptide or polynucleotide.

The present invention concerns an anti-COX-2 autoantibody. Also, the methods and kits of present invention concern an anti-COX-2 autoantibody. In one embod iment the anti-COX-2 autoantibody is an isolated antibody. Said autoantibody can be used as a diagnostic marker or for determining a bone marrow failure disorder of a subject. In one embodiment said autoantibody associates with the FILA II cod ing DRB1^*15 locus and/or altered peripheral blood platelet. Therefore, in addition to serving as a novel biomarker, this COX-2 binding autoantibody also participates in the pathogenesis of BMF disorders.

COX-2 (cyclooxygenase-2, prostaglandin-endoperoxide synthase 2, prostaglandin G/H synthase and cyclooxygenase, PTGS2) (e.g. EC number 1.14.99.1) is an en zyme involved in the conversion of arachidonic acid to prostaglandin H2, an im portant precursor of prostacyclin, which is expressed in inflammation. An example of a human COX-2 polypeptide is presented by Uniprot P35354 or the sequence with the accession number NP_000954. An example of a human COX-2 mRNA reference sequence is presented by the sequence with the GenBank accession number NM_000963.

An antibody is an immunoglobulin molecule and it can belong to any of classes IgA, IgD, IgE, IgG or IgM. In one embodiment an isotype of the anti-COX-2 auto antibody is IgA, IgD, IgE, IgG or IgM, typically IgG, IgA or IgM. In one embodiment an isotype of the autoantibody is IgG and the subclass is IgGi, lgG2, lgG3 and/or lgG4. In one embodiment, the isotype of the autoantibody is IgG and the subclass is IgGi and/or lgG3.

The anti-COX-2 autoantibody or an antigen binding fragment thereof is against a COX-2 polypeptide or capable of binding a COX-2 polypeptide or a fragment thereof. As used herein “an anti-COX-2 autoantibody” refers to an antibody or an antigen binding fragment thereof produced by the immune system of a subject and directed against the subject’s COX-2 polypeptides. E.g. any fragments or single chain antibodies (e.g. Fab, Fv, scFv), complementarity determining regions, heavy chain variable regions, light chain variable regions and any combinations thereof can be included within the scope of “antigen binding fragments”. In antibodies var iable loops between b-strands, three on each light (VL) and heavy (VFI) chain, are responsible for binding to the antigen. These loops are referred to as the comple mentarity determining regions (CDRs). As used herein, the term “hypervariable re gion” refers to the amino acid residues of an antibody or fragment thereof which are responsible for antigen binding. As used herein, the terms “heavy chain varia ble domain” and/or “VH” are used interchangeably and reference the hypervariable region (encompassing both the CDR and framework domains) of the heavy chain of an antibody; the terms “light chain variable domain” and/or “VL” are used inter changeably and reference the hypervariable region (encompassing both the CDR and framework domains) of the heavy chain of an antibody.

In one embodiment, the anti-COX-2 autoantibody is an isolated anti-COX-2 auto antibody. In one embodiment, the anti-COX-2 autoantibody is a human anti-COX- 2 autoantibody. In one embodiment, the anti-COX-2 autoantibody is an isolated human anti-COX-2 autoantibody. In one embodiment, the anti-COX-2 autoanti body is a polyclonal antibody. In one embodiment, the anti-COX-2 autoantibody is a human polyclonal antibody. In one embodiment the anti-COX-2 autoantibody is against a COX-2 polypeptide e.g. as shown in SEQ ID NO:1 or capable of binding a COX-2 polypeptide e.g. as shown in SEQ ID NO:1 or a fragment thereof. The anti-COX-2 autoantibody can be capable of binding one or more amino acids e.g. within amino acid positions 1 - 604, 56 - 604, 100 - 604, 200 - 604, 300 - 604, 350 - 604, 442 - 604, 490 - 604, 530-604 or 490 - 600 as shown in SEQ ID NO: 1.

In one embodiment the anti-COX-2 autoantibody is capable of binding the C- terminal part of a COX-2 polypeptide, e.g. the C-terminal part at one or more ami no acid positions within amino acids 56 - 604, 442 - 604, 490 - 604, 530-604, 577 - 604, 580 - 594, or 490 - 600, as shown in SEQ ID NO: 1. In one embodiment the anti-COX-2 autoantibody is capable of binding at one or more amino acid posi tions within amino acids 530 - 604 as shown in SEQ ID NO: 1. In one embodiment the autoantibody can bind the C-terminal part of a COX-2 polypeptide at positions comprising amino acid 491, 534, 540, 577, 580, 587 and/or 590. In one embodi ment the autoantibody is capable of binding one or more of the following amino ac id fragments of the COX-2 polypeptide: PALLVE (SEQ ID NO: 2; starting from amino acid 491 of SEQ ID NO: 1), STFGGEV (SEQ ID NO: 3; starting from amino acid 534 of SEQ ID NO: 1), GFQIINT (SEQ ID NO: 4; starting from amino acid 540 of SEQ ID NO: 1), QIINTAS (SEQ ID NO: 5; starting from amino acid 542 of SEQ ID NO: 1), VTIN (SEQ ID NO: 6; starting from amino acid 577 of SEQ ID NO: 1), GLDDIN (SEQ ID NO: 7; starting from amino acid 587 of SEQ ID NO: 1), and/or DINPTVL (SEQ ID NO: 8; starting from amino acid 590 of SEQ ID NO: 1). In one embodiment the autoantibody is capable of binding one or more amino acids of the following amino acid fragment of the COX-2 polypeptide: NASSSRSGLDDINPT (SEQ ID NO: 9; positions 580-594 of SEQ ID NO: 1). (See also Figure 5C.)

The method or a (diagnostic) test kit for determining a bone marrow failure disor der or for determining an anti-COX-2 autoantibody of a sample can comprise a capture complex for binding an anti-COX-2 autoantibody. In one embodiment the binding of an anti-COX-2 autoantibody is specific binding. As used herein “a cap ture complex” refers to a binding partner (a first binding partner) capable of binding a diagnostic autoantibody i.e. an anti-COX-2 autoantibody. Actually, a capture complex can be any polypeptide or molecule or comprise any polypeptide or mol ecule, which is able to bind an anti-COX-2 antibody. One or more capture com plexes (identical or different capture complexes) may be used in the method or kit of the present invention. In one embodiment the capture complex is a polypeptide, recombinant polypep tide, or any fragment thereof or a combination of different polypeptides comprising one or more recombinant polypeptides or fragments thereof, for binding an anti- COX-2 autoantibody. In one embodiment the capture complex is a recombinant polypeptide (e.g. a recombinant COX-2 polypeptide or a recombinant mammalian or human COX-2 polypeptide) or a fragment thereof or a combination of different polypeptides comprising one or more recombinant polypeptides (e.g. a recombi nant COX-2 polypeptide or a recombinant mammalian or human COX-2 polypep- tide) or fragments thereof. For example, the recombinant mammalian COX-2 poly peptide or a fragment thereof capable of binding an anti-COX-2 antibody can be selected e.g. from a group comprising or consisting of a human, common chim panzee, monkey, mouse, dog, and cat COX-2 polypeptide or a fragment thereof. In one embodiment the capture complex is a polynucleotide, recombinant polynu cleotide, or any fragment thereof or a combination of different polynucleotides comprising one or more recombinant polynucleotides or fragments thereof, for binding a polynucleotide encoding an anti-COX-2 autoantibody. In one embodi ment the capture complex is a recombinant polynucleotide (e.g. a recombinant COX-2 polynucleotide or a recombinant mammalian or human COX-2 polynucleo tide) or a fragment thereof or a combination of different polynucleotides comprising one or more recombinant polynucleotides (e.g. a recombinant COX-2 polynucleo tide or a recombinant mammalian or human COX-2 polynucleotide) or fragments thereof for binding the polynucleotide encoding the anti-COX-2 autoantibody. For example, the recombinant mammalian COX-2 polynucleotide or a fragment thereof capable of binding a polynucleotide encoding an anti-COX-2 antibody can be se lected e.g. from a group comprising or consisting of a human, common chimpan zee, monkey, mouse, dog, and cat COX-2 polypeptide or a fragment thereof. As used herein, "a recombinant polypeptide" refers to any polypeptide that has been genetically modified to contain different genetic material compared to a cor responding polypeptide of an organism (e.g. comprises a deletion, substitution, disruption or insertion of one or more amino acids, or comprises a deletion, substi tution, disruption or insertion of one or more nucleic acids in a polynucleotide en- coding said polypeptide), or any polypeptide produced with recombinant DNA technology. “A recombinant polynucleotide” refers to any polynucleotide that has been genetically modified to contain different genetic material compared to a cor responding polynucleotide of an organism (e.g. comprises a deletion, substitution, disruption or insertion of one or more nucleic acids, or any polynucleotide pro duced with recombinant DNA technology. A recombinant polynucleotide may en code a recombinant polypeptide.

A polypeptide for binding an anti-COX-2 antibody or a polynucleotide for binding a polynucleotide encoding an anti-COX-2 antibody can be produced by any suitable method known to a person skilled in the art. For example, the polypeptides or pol ynucleotides may be produced synthetically. For example, eukaryotic, prokaryotic, animal, human, mammalian, bacterial, yeast, filamentous fungal, plant and insect cells are widely used expression systems for the production of recombinant poly peptides or polynucleotides e.g. in vitro. The polypeptides or polynucleotides uti lized in the present invention can be developed e.g. by using genetically engi neered genes expressed in an in vitro cell lines. All techniques for producing poly peptides or polynucleotides are known to a person skilled in the art and are de scribed in practical manuals and handbooks for laboratory molecular techniques. In a specific embodiment the polypeptide for binding an anti-COX-2 antibody has been frozen in a liquid before use in the method or kit of the present invention.

The capture complex for binding an anti-COX-2 autoantibody or a polynucleotide encoding said anti-COX-2 autoantibody can optionally be attached or immobilized to a surface (e.g. in a well, optionally in a multi-well plate), optionally through a binding partner (a second binding partner). For example, the capture complex (e.g. the polypeptide, polynucleotide or any fragment thereof) may be tagged, e.g. com prise one or more affinity tags selected from the group consisting of a polyhisti dine-tag, FIA-tag, FIN-tag, FIQ-tag, FLAG-tag, Myc-tag, V5-tag, GST-tag, MBP-tag, Strep-tag and CL7-tag. Then suitable binding partners for attaching the capture complex comprising an affinity tag to a surface include but are not limited to re combinant anti-His-tag antibodies, monoclonal anti-FLAG-antibodies, monoclonal anti-Myc antibodies, an engineered streptavidin, monoclonal anti-Strep antibody and any combination thereof. Tags or affinity tags are selected based on the spe cific binding partners. In one embodiment the capture complex is immobilized on the surface by electrochemical forces.

In one embodiment the method of the present invention comprises allowing a cap ture complex to bind with an anti-COX-2 autoantibody or a polynucleotide encod ing said anti-COX-2 autoantibody or any fragment thereof (thereby forming a com bination of the capture complex and the anti-COX-2 autoantibody or a fragment thereof, or the combination of the capture complex and the polynucleotide or a fragment thereof), and determining said anti-COX-2 autoantibody, polynucleotide or fragment or amount thereof. In one embodiment the capture complex is allowed to bind a surface or to be immobilized on a surface either before or after the cap ture complex is allowed to contact with an anti-COX-2 autoantibody or a polynu cleotide encoding said autoantibody (e.g. with a sample comprising an anti-COX-2 autoantibody or a polynucleotide encoding said autoantibody). Indeed, in one em bodiment, part of a (pre-treated or untreated) sample of a subject is added to a surface comprising an immobilized capture complex thereby allowing binding of the capture complex and the autoantibody or polynucleotide.

In one embodiment of the invention one or more washing steps are carried out af ter allowing capture complexes to contact with anti-COX-2 autoantibodies, polynu cleotides or fragments thereof to remove unbound molecules (e.g. capture com plexes, anti-COX-2 autoantibodies, polynucleotides encoding said autoantibody etc.). In addition, any other optional washing step can be carried out in the method of the present invention. A washing step or steps of the present invention may be carried out with any wash buffer known to a person skilled in the art including but not limited to phosphate buffered saline (PBS) and Tris-HCI buffered saline (TBS).

In one embodiment of the invention the method comprises or the kit enables de tecting and/or measuring the presence, absence, amount or level of anti-COX-2 autoantibodies, nucleotides encoding said autoantibodies, or fragments thereof, in a sample for example by enzymatic assays or means, immunological detection methods or means, colorimetric or fluorescent methods or means, paramagnetic or electrochemical methods or means, or any instrumental detection methods or means. In one embodiment the method utilizes, or the test kit comprises a label, agent, molecule or any combination thereof for detecting the anti-COX-2 autoanti body, nucleotide encoding said autoantibody or any fragment thereof. Detectable labels such as isotope, metal, (poly)nucleotide, amino acid, color or fluorescent moieties may be utilized for determination or detection. In one embodiment the molecule for detecting the anti-COX-2 autoantibody is a detection antibody, op tionally labelled. As used herein “a detection antibody” refers to an antibody (e.g. an IgA, IgD, IgE, IgG or IgM detection antibody or a fragment thereof, correspond ing to the anti-COX-2 autoantibody to be determined) or a fragment thereof, which is capable of binding an anti-COX-2 autoantibody. Agents for detecting the anti- COX-2 autoantibody or polynucleotide encoding said autoantibody include but are not limited to suitable buffers and detection pairs (such as biotin/streptavidin). !n one embodiment after the anti-COX-2 autoantibody has bound to the capture complex (optionally followed by a washing step), the detection antibody can be added to form a complex with the anti-COX-2 autoantibody and the capture com plex. The detection antibody can be linked e.g. to an indicator such as an enzyme or can itself be detected e.g. by an indicator such as a label or a second ary antibody that is linked e.g. to a label or an enzyme through bioconjugation. (See e.g. Figure 8.) For example, green fluorescent proteins, color labels, metal labels, fluorescent labels, isotope labels, (poly)nucleotide labels, amino acid labels or any other suitable labels can be used for determining the presence, absence, amount or level of the formed complexes comprising at least the capture complex and the anti-COX-2 autoantibody.

Optionally determination of the anti-COX-2 autoantibody, polynucleotide encoding said autoantibody, or any fragment thereof, may also comprise use of any suitable statistical methods known to a person skilled in the art.

In one embodiment, the method or kit of the present invention may comprise tools for an immunoassay comprising an antibody or an antigen binding fragment e.g. for the binding of an anti-COX-2 antibody. The immunoassay can be either a com petitive or non-competitive immunoassay. Competitive immunoassays include ho mogenous and heterogenous immunoassays. The immunoassay is not limited to but can be selected e.g. from the group consisting of DELFIA, FIA, ELISA, or im- munoPCR. In one embodiment the immunoassay may be for example a conven tional sandwich test in microtiter wells or a lateral flow-test. In a specific embodi ment the test kit comprises reagents for carrying out an (immune)assay. In a spe cific embodiment of the invention, the method comprises or is based on an enzy matic assay and/or immunoassay, or the kit comprises tools for an enzymatic as say and/or immunoassay.

In one embodiment in addition to the capture complex, the kit of the present inven tion comprises at least tools for determining an anti-COX-2 autoantibody, a poly nucleotide encoding said autoantibody, or a fragment thereof from a sample of a subject, and optionally reagents (such as one or more selected from the group consisting of suitable reaction solutions, washing solutions, buffers and enzymes) for carrying out a method for determining anti-COX-2 autoantibodies. In one em bodiment tools for determining may include one or more tools selected from the group consisting of detection means (such as labels or enzyme(s)), and one or more antibodies or antigen binding fragments for binding an anti-COX-2 autoanti- body. In a very specific embodiment of the invention the kit comprises one or more from the group comprising or consisting of tools to determine an anti-COX-2 auto antibody or a polynucleotide encoding said autoantibody or a fragment thereof, re agents for determining said autoantibody or polynucleotide or fragment, the refer ence levels or ranges (e.g. cut off levels) of suitable subjects (e.g. normal subjects and/or subjects with a disease of interest), and instructions for carrying out a method of the present invention, for determining anti-COX-2 autoantibodies or pol ynucleotides or fragments, for carrying out statistical methods and/or determining whether a subject has a BMF disorder.

In one embodiment the test kit is for the method of the present invention.

The method, autoantibody, polynucleotide, fragment or kit of the present invention is very sensitive and/or specific for BMF diseases. In one embodiment, the sensi tivity of the method, autoantibody, polynucleotide, fragment or test kit is more than 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, or 99%; and/or the specificity is more than 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95% or 99%. In one embodiment at least 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45% or 50% of all BMF disease patients or patients with a specific BMF disease (such as AA patients) is detected with the method or kit of the present invention. See also Figure 3.

In a very specific embodiment of the invention ROC (receiver operating character istic) curve i.e. a graphical plot is utilized for illustrating the diagnostic ability, i.e. the sensitivity and/or specificity, of the method or kit for detecting a BMF disorder. ROC may be used for example when the method or kit of the present invention is established.

In one embodiment the method of the present invention is carried out in one or more wells, multi-well plates, tubes, vials, vessels, columns or any combination thereof, or the kit comprises one or more wells, multi-well plates, tubes, vials, ves sels, columns or any combination thereof.

Indeed, the technology of the present invention allows high-throughput determina tion or screening of large amounts of samples.

In one embodiment of the invention the determination of the anti-COX-2 autoanti bodies or polynucleotide encoding said autoantibodies or fragments thereof is car- ried out in vitro. In another embodiment the kit of the present invention is for in vitro method in vitro diagnostics refer to medical and veterinary laboratory tests that are used to diagnose diseases and/or monitor the health or clinical status of subjects using samples obtained from said patients.

A sample utilized in the present invention can be e.g. a blood, plasma, serum, bone marrow aspirate, or any body fluid sample. In one embodiment of the inven tion the sample is selected from the group consisting of a blood sample, plasma sample, serum sample, urine sample, sputum sample, saliva sample, cerebrospi nal fluid sample, and bone marrow aspirate. Samples may be collected with any suitable method known to a person skilled in the art including but not limited to col lecting a liquid sample, needle, and aspiration. Typically, samples are obtained through collection of the body fluid or by inserting a needle into the blood vessel or body cavity and optionally aspirating a portion of the fluid.

In one embodiment of the invention a subject is a human (e.g. an adult, adolescent or child) or an animal (e.g. a mammal), 0 - 10, 11 - 20, 21 - 50 or 51 - 70 or 71 - 100 years old; and/or a female or male. A subject can be in need of determining a BMF disorder or a subgroup thereof. The mammal can be selected from the group consisting of a human, common chimpanzee, monkey, mouse, dog, and cat.

Before classifying a subject as suitable for any method or kit of the present inven tion, the clinician may for example study any symptoms or assay any disease markers of the subject. Symptoms of a BMF can include but are not limited to anemia (in some embodiments including one or more of the following: tiredness, weakness, pallor, breathlessness, tachycardia), neutropenia (in some embodi ments including e.g. recurrent and/or severe bacterial infections), thrombocytope nia (in some embodiments including one or more of the following: easy bruising, petechiae, bleeding from the nose and/or gums), skin or skeletal abnormalities, enlarged liver and/or spleen, and/or lymphadenopathy. The clinician may sug gest the method or kit of the present invention for determining a BMF disorder e.g. based on the results deviating from the normal or when having the suspicion of a BMF disorder.

The tools and methods of the present invention can be utilized as diagnostic tools and methods in patients with symptoms of BMF disorders. As an example, if anti- COX-2 autoantibodies are present or the levels are increased in a sample of a subject, these subjects would typically then be forwarded to clinics for further ex- amination and guidance. For example, platelet counts and/or HLA-DRB1 *15:01 haplotype can be studied from a subject before or after the method of the present invention or before or after utilizing the tools of the present invention. As an exam ple, it was found that aCOX-2-Ab positive patients had statistically significantly lower platelet counts at diagnosis. In addition, it was discovered that aCOX-2-Ab positive patients were more likely to present HLA-DRB1 ^*15:01 genotype. Also, it was found that aCOX-2-Ab positive patients were typically adults, >18 years old, >40 years old, >50 years old, >60 years old or >70 years old.

For example, a specific treatment for the subject having anti-COX-2 autoantibod ies or polynucleotides encoding said autoantibodies or fragments thereof, or hav ing increased levels thereof can be selected based on the found levels and/or the symptoms. The method and kit of the present invention enable identifying subjects that can benefit from specific treatments of BMF disorders, optimization of the specific treatment, and/or are responsive/nonresponsive to a treatment of a BMF disorder. As used herein, the term "treatment" or "treating" refers to administration of at least one therapeutic agent to a subject for purposes which include not only complete cure but also amelioration or alleviation of disorders or symptoms related to a BMF disorder in question. Therapeutically effective amount of an agent refers to an amount with which the harmful effects of a BMF disorder are, at a minimum, ameliorated.

It will be obvious to a person skilled in the art that, as the technology advances, the inventive concept can be implemented in various ways. The invention and its embodiments are not limited to the examples described below but may vary within the scope of the claims.

EXAMPLES

Patients, materials and methods

Patient samples

Plasma samples from patients and healthy controls were collected according to the institutional ethics approvals (Helsinki: 303/12/03/01/2011, 181/13/03/01/12, Cleveland, USA: IRB 5024 CR, Kanazawa, Japan: Kanazawa 2018/4/25 and Shinshu, Japan: Shinshu IRB 581) after informed consent was obtained. The prin ciples of the Declaration of Helsinki were carefully followed. A cohort of 300 plas ma samples was obtained from Helsinki Biobank. Autoantibody screen and analysis

Autoantibodies were screened from plasma samples using the Invitrogen (Carls bad, California, USA) ProtoArray protein microarray v.5.1 (https://www.thermofisher.com/jp/en/home/life-science/protein-biology/protein- assays-analysis/protein-microarrays.html) as previously described in Hamano Y et al. 2017, Sci Rep. 7(1 ): 1 —15. All proteins have been expressed as glutathione-S- transferase (GST) fusion proteins, purified under native conditions and spotted on nitrocellulose-coated glass slides. Slides were blocked with 1% BSA/phosphate- buffered saline/Tween (PBST). Plasma samples were added to the arrays. After washing, anti-human immunoglobulin G (IgG) conjugated to Alexa Fluor 647 dye was added. Arrays were washed and dried (Partnership, Evry, France). Arrays were scanned with a GenePix 4000B Fluorescent Scanner, GenePix. Molecular Devices, Sunnyvale, California, USA. Data were acquired with GenePix Pro soft- ware and processed using ProtoArray Prospector 2.0 (Invitrogen). A panel of val ues was calculated for each protein array, including the Z-score, the Chebyshev inequality precision (CIP) value and the coefficient of variation (CV) value as pre viously described (Yoshizato T et al. 2015, New England Journal of Medicine. 373(1 ):35-47). A Z-score >3.0, a CIP value <0.05 and a CV <0.5 define a positive spot. The cut-off for sample positivity was set to FC10 compared to the average of healthy controls.

COX-2 antigen

A recombinant cyclooxygenase-2 (COX-2) protein encoded by gene PTGS2 (Sino Biological Inc., Beijing, China) was used in the binding studies. The protein con sisted of amino acids 1-604 of the human COX-2, was C-terminally Flis-tagged and supplied in frozen solution form.

DELFIA assays Antibody binding studies were carried out with sandwiched Dissociation Enhanced Lanthanide Fluorescence Immunoassays (DELFIA). Each well of a 96-well Nunc- Immuno Maxisorp plate (Nunc A/G, Roskilde, Denmark) was coated with 250 ng of mouse anti-His-tag antibody (Thermo Fisher Scientific, Rockford, IL) in phosphate- buffered saline (PBS, Corning Life Sciences, Oneonta, NY) overnight in room temperature (RT). The plates were washed in 4 cycles with PBS + 0.05 % Tween20 (PBS-T) using DELFIA Platewash (PerkinElmer, Shelton, CT). Wells were blocked against non-specific serum protein binding with 1 % DTPA-purified bovine serum albumin (BSA; PerkinElmer) for 1 h in RT. After washing, wells were incubated with 100 ng/well of recombinant COX-2 protein in a diluting buffer (PBS- T with 0.2 % of DTPA-purified BSA; PBS-T+BSA) for 1 h in RT. After wash cycles, plasma/serum samples were added diluted 1 :100 in the diluting buffer (PBS- T+BSA) in duplicates. Blank controls of diluting buffer were placed in duplicates at the beginning and end of each plate. Each set of plates included a series of 6 standards prepared from a cross-reacting rabbit anti-human-COX-2 antibody (SDIX LLC, Newark, DE). After incubating for 1 h in RT and washing for 4 cycles 100 pL/well of Eu-labelled mouse anti-human-lgG antibody diluted 1 :1000 in DELFIA Assay buffer (both PerkinElmer) was added and incubated for 1 h in RT. The Eu-labelled detection antibody was washed off extensively for 6 cycles, DELFIA Enhancement Solution (PerkinElmer) was added and plates incubated for 5 minutes to activate the label. Fluorescence was measured using Victor X4 plate reader (PerkinElmer) with Time Resolved Fluorometry Europium protocol (excita tion at 340 nm) as per manufacturer recommendations.

IgG subclass isotypes IgGi - lgG4 were determined with the DELFIA method de scribed in detail above. No subclass isotype standards were available. Instead of the Eu-labelled anti-human-lgG antibody, each batch was incubated 1 h in RT with 100 pL/well of subclass-specific biotinylated mouse anti-human-IgGi - lgG4 (Sig- ma, Saint Louis, MO). Anti-lgGi was diluted 1 :1000, anti-lgG2 and anti-lgG3 1 :5000 and anti-lgG4 1 :10 000 in DELFIA Assay buffer. These were followed with incuba tion of 1 h in RT with 100 pL/well Eu-labelled streptavidin (PerkinElmer) 1:1000 in DELFIA Assay buffer. Anti-COX-2 IgA and IgM isotypes were determined in separate batches using the DELFIA method described in detail above. Instead of using the Eu-labelled anti- human-lgG antibody, each batch was incubated 1 h in RT with biotinylated goat anti-human-lgA (a chain) or -IgM (m chain) antibody (Life Technologies, Frederick, MD). Both antibodies were diluted 1 :7500 in DELFIA Assay buffer and used 100 pL/well. These were followed by Eu-labelled streptavidin 1 :1000 in DELFIA Assay buffer as described above for IgG subclass isotypes. Determination of test positivity cut-off

To separate positive cases from negative two R packages Findcutoffs (Chang C et al. 2017, PLoS ONE. 12(4):e0176231) and OptimalCutPoint (Lopez-Raton M et al. OptimalCutpoints: An R Package for Selecting Optimal Outpoints in Diagnostic Tests. 2014) were used. The goal was to set a cutoff level optimizing sensitivity and specificity simultaneously for DELFIA fluorescence counts as a marker for the dichotomized outcome of aplastic anemia as the diagnosed disease. The com bined measurements of 681 patient samples (all cohorts excluding the controls from Helsinki Biobank) were used as training data, and the 300 Helsinki Biobank samples were used as the validation group. Likelihood ratio test for statistical sig nificance and AUC were used for the former package and Youden’s index and AUC for the latter as optimization methods. As the resulting cutoff levels were very close to each other, their mean value was chosen as the cutoff for positivity.

Linear epitope mapping

Both linear and conformational epitopes were mapped using the PEPperPRINT® technology. The sequence of prostaglandin G/H synthase 2 (UniProt ID P35354) was elongated with neutral GSGSGSG linkers at the C- and N-terminus to avoid truncated peptides. The elongated antigen sequence was translated into linear 15 amino acid peptides with a peptide-peptide overlap of 14 amino acids for high- resolution epitope mapping. The prostaglandin G/H synthase 2 peptide microar rays contained 604 different peptides printed in duplicate (1,208 peptide spots) and were framed by additional HA (YPYDVPDYAG) and polio (KEVPALTAVETGAT) control peptides (44 peptide spots each control). Plasma di lutions of 1 :500 and 1 :100 were incubated and the signal was detected with Goat anti-human IgG (Fc) DyLight680 (0.1 pg/ml) secondary antibody. Mouse monoclo nal anti-HA (12CA5) DyLight800 (0.5 pg/ml) was used as control antibody. LI-COR Odyssey Imaging System; scanning offset 0.65 mm, resolution 21 pm, scanning intensities of 7/7 (red = 700 nm/green = 800 nm).

Conformational epitope mapping

The elongated antigen sequence was translated into 7, 10 and 13 amino acid pep tides with a peptide-peptide overlap of 6, 9 and 12 amino acids. After peptide syn thesis, all peptides were cyclized via a thioether linkage between a C-terminal cys teine and an appropriately modified N-terminus. The conformational prostaglandin G/H synthase 2 peptide microarrays contained 1,827 different peptides printed in duplicate (3,654 peptide spots), and were framed by additional HA (YPYDVPDYAG, 64 spots) and polio (KEVPALTAVETGAT, 62 spots) control pep tides. Detection as in linear epitope mapping.

Western blot Recombinant COX-2 was mixed with Laemmli buffer (Bio-Rad Laboratories, Her cules, CA), PBS (Corning Life Sciences) and DTT (Dithiothreitol) to a final concen tration of 2,5pg/mL. 20pL of this sample was applied on a 7,5% SDS-PAGE pre cast gel (Bio-Rad Laboratories) together with WesternSure Pre-stained Chemilu- minescent Protein Ladder (LI-COR Biosciences, Lincoln, NE) and the proteins were transferred to a PVDF membrane (Merck Millipore, Burlington, MA). Odyssey blocking buffer (OBB, LI-COR Biosciences) mixed 1 :1 with PBS was used as blocking solution before incubation with patient plasma diluted (1 : 2000) in 60% PBS, 40% OBB and 0,2% Tween 20. Mouse anti-human IgG-HRP (1 :1000) (Merck Millipore) diluted in 60% PBS, 40% OBB and 0,2% Tween 20 was used to detect the autoantibodies using Clarity Western ECL Blotting Substrates (Bio-Rad Laboratories) and ChemiDoc MP Imaging System (Bio-Rad Laboratories).

Single-cell RNA-sequencing analysis from healthy bone marrow mononucle- ar cells

Droplet-based single-cell RNA sequencing data from bone marrow mononuclear cells from 22 samples from 20 healthy donors (Oetjen KA et al. 2018, JCI Insight. 3(23)) were analyzed. Cells with a high amount of mitochondrial transcripts (>10% of all UMI counts) or ribosomal transcripts (>50%), cells with less than 100 genes or over than 4500 genes expressed, cells expressing low or high (<25% or >60%) amount of housekeeping genes or cells with low or high read depth (<500 or >30 000) were excluded from the analyses. To overcome batch-effect, we used a re cently described probabilistic framework to overcome different nuisance factors of variation in an unsupervised manner with deep generative modelling. Briefly, the transcriptome of each cell is encoded through a nonlinear transformation into a low-dimensional, batch corrected latent embedding. The latent embedding was then used for graph-based clustering implemented in Seurat (3.x. x) and UMAP- dimensionality reduction. Cell types were annotated with recent machine-learning method Singler, where the Blueprint data set was used as a reference data set (Aran D et al. 2019, Nat Immunol. 20(2):163-172).

Single-cell RNA-sequencing analysis from AA, de novo MDS and healthy CD34+ cells

Plate-based single-cell RNA sequencing from CD34+ cells from 3 AA patients, 2 de novo MDS patients and 4 healthy donors were gathered from Zhao et al., Blood 2017 (Zhao X et al. 2017, Blood. 130(25):2762-2773). All cells were previously quality controlled with comparable thresholds and previously normalized. Highly variables genes were determined FindVariableFeatures-function implemented in Seurat, with expression cut-off at 1.0. The first 50 principal components were counted with the highly variable genes, and principal components with standard deviation > 2 were selected to be used in clustering with FindNeighbors and FindClusters-functions and data visualization in UMAP-reduced space. Clusters were annotated using canonical marker genes suggested by the original publica tion and by Van Galen et al. (van Galen P et al. 2019, Cell. 176(6):1265-1281.e24)

Statistical methods

Comparisons between groups on blood counts were made using the nonparamet- ric Mann-Whitney U test (Wilcoxon rank-sum test). Age distributions between groups were tested for equality using the nonparametric Kolmogorov-Smirnov (K- S) test. Presence of HLA allele or PNH clone in patient groups was tested using cross-tabulation with Chi-squared test of independence.

Results

Identification of autoantibodies in IAA patients

Protoarray whole protein microarray analysis was performed to identify potential autoantibodies in seven (n=7) AA patients, five (n=5) healthy controls (HC), twelve (n=12) LGL leukemia and ten (n=10) RA patients. In Figure 1 we present all pro teins, against which there were plasma antibodies in at least two IAA patients with values > FC10 compared to the HC. COX-2 was the only protein target with IAA restricted autoantibody expression with > 20-fold change expression values in all positive cases.

Anti-COX-2 autoantibodies are specific for IAA

To confirm microarray results, we set up a DELFIA immunoassay that detected anti-COX-2 IgG antibodies. Anti-COX-2 autoantibodies were successfully detected in all three index cases, while all control patients were tested negative. As de scribed in materials and methods, the positivity threshold was set using two previ ously described algorithms, and the same threshold was applied on all tested samples and sample cohorts.

Next, we collected independent patient cohorts from Finland (IAA n= 38), Japan (IAA n= 108) and the US (IAA n=118) (Table 1). An updated dataset with only adult patients (over 18 years old) derived from larger total patient cohorts is pre- sented in Figure 10 (Finland n= 37, Japan n=88, US n=209). In addition to IAA pa tients, all cohorts also included related diseases as control samples.

Table 1. Study cohort demographics.

Cohort n Male/Female Mean age VSAA and Response to 1st

(range) ¹⁾ SAA / Mode- line 1ST (NR/PR/CR) rate

31 %/69% 54%/46% 40%/50%/10%

IAA Finland 38 (11/24) 57 (17-84) (19/16) (8/10/2)

44%/56% 42%/58% 17%/49%/34%

IAA Japan 108 (48/60) 50 (6-94) (45/61) (13/37/26) IAA United 42%/58% 70%/30% 16%/57%/27% States 118 (50/68) 44 (6-86) (45/19) (7/25/12)

Healthy cont- 63%/37%

1 ) At time of diagnosis

2) Finland n=30, Japan n=20, United States n=24

Abbreviations: VSAA (very severe aplastic anemia), SAA (severe aplastic ane mia), 1ST (immunosuppressive therapy), NR (no response), PR (partial response), CR (complete response)

In all study cohorts aCOX-2 antibodies were highly specific to IAA (Figure 2 A-D, Figure 11). Sample positivity for aCOX-2 Ab was set using a cutoff value of 900 000 DELFIA counts. This cutoff level corresponds to the sharp turning point in ROC curve maximizing test sensitivity (0.36) with acceptable specificity (0.98, Figure 3A). The highest aCOX-2 antibody positivity (61%) was noted in the Finn ish cohort of IAA patients (Figure 2A). In the cohorts collected from the United States (Figure 2B) and from the Japan (Figure 2C), 27 % and 37 % of the AA pa tients were tested positive for the aCOX-2-autoantibody. In all cohorts, there were sporadic positive findings in related bone marrow failure diseases, while all RA, LGL leukemia, graft-versus-host disease patients and also all healthy controls were negative for the aCOX-2 antibody. Three hundred (n=300) plasma samples were accessed from the Helsinki Biobank (Figure 2D). There were two (n=2) aCOX-2 positive IAA patients among the six (n=6) available samples. All patients without a documented history of autoimmune disease (n=100) were tested nega tive for the aCOX-2-Ab. From the autoimmunity cohort (total n = 194) only 2 indi viduals with multiple sclerosis (MS, n = 50) and 2 individuals with diabetes mellitus 1 (n = 50) were tested aCOX-2 Ab positive. More controls were analyzed and all control groups are presented in Figure 11. After pooling all cohorts’ data, the overall aCOX-2 prevalence was 36 % (97 patients out of 273) of all IAA patients (Figure 3C). Updated analysis indicated overall prevalence of 32 % in IAA, 37 % in adults (over 18 year old) and 55 % in IAA patients over 40 years old (Table 2). The observed aCOX-2 antibodies persisted in patients’ circulation during follow up (Figure 3D). A larger follow-up cohort of n=21 confirmed the persisting nature of the aCOX-2 antibody in IAA patients (Figure 12). Table 2.

The aCOX-2 Ab isotype profile is dominated by lgG1 and lgG3

In addition to IgG antibodies, the isotypes IgA and IgM were analyzed for the anti- COX-2 Ab positive subset of the Finnish IAA patients (n = 26) and healthy controls (n = 17). Anti-COX-2 Ab response was predominantly IgG type, with three patients presenting elevated levels of both IgA and IgG, and one presenting all three (Figure 4A). This patient presented with an active disease in follow-up, with stable aCOX-2 IgG counts (P8 in Figure 3D). IgA and IgM levels were measured in the latest time point (69 months from diagnosis), not a particularly early stage in dis ease. Next, we further specified the IgG response into its subclass isotypes IgGi - lgG4 in the same subset of Ab positive AA patients (n = 38) against healthy con trols (n = 30) and one (n = 1 ) Ab negative control AA patient. The negative control patient remained negative for all the tested subclasses, while all anti-COX2 Ab positive patients presented IgGi and 55 % presented lgG3. Only singular cases presented lgG2 or lgG4 in addition to the IgGi and lgG3 class antibodies (Figure 4B). aCOX-2 autoantibodies bind the C-terminal part of COX-2

SDS-PAGE electrophoresis followed by western blotting confirmed, that the newly identified autoantibodies bind linearized full-length recombinant COX-2. Microar ray-based linear peptide mapping was performed for ten aCOX-2-Ab positive and two aCOX-2-Ab negative AA patients. We identified an almost identical, DIN ami no acid signature containing C-terminal epitope in five of the tested aCOX-2-Ab positive patients and additionally, other closely mapping C-terminal epitopes were identified for 2 patients (Figure 5A). The linear mapping method could not identify an antigenic epitope for three of the aCOX-2-Ab positive patients, and thus, they were subjected for conformational peptide screen (Figure 5B). Conformational peptide mapping with cyclic peptides revealed two additional antigen epitopes in the C-terminal part of COX-2. All identified epitopes are summarized in Figure 5C.

Anti-COX-2 antibodies correlate with age, lower platelet counts, and HLA type

We compared key clinical findings in IAA patients between aCOX-2 Ab negative and positive patients. Logistic regression analysis was performed with many addi tional IAA patients included in the analysis (Table 3). Logistic regression analysis revealed that aCOX-2 antibody positive IAA patients were older (Figure 13), were more likely to present HLA-DRB1 ^*15:01 genotype (Figure 14) and also had lower platelet counts at diagnosis (Figure 15.) than the aCOX-2 negative IAA patients. There was a statistically significant difference in peripheral blood platelet counts measured at the time of diagnosis (Figure 6A) between aCOX-2 Ab negative and positive patients (n = 126 and n = 84 respectively, Mann-Whitney U test p-value <0.001). However, both groups presented with clinically meaningful low mean platelet counts <50 E9/L at diagnosis. At the time of diagnosis, IAA patients’ ages showed a two-peaked distribution, and aCOX-2 Abs presented more frequently in elderly patients (Two-sided two-sample Kolmogorov-Smirnov test p-value <0.001, Figure 6B). Age distributions varied between country cohorts, where IAA samples from United States included more younger patients. Some patients were sampled closer to diagnosis than others, but this difference in distribution did not change when considering age at time of sampling (K-S test p-value <0.001). Examining the time delay from diagnosis to sampling did not reveal any significant change in younger patients’ serostatus for those patients that had received IAA diagnosis early but were sampled for the antibody at a later age.

Table 3. aCOX-2 Ab nega- aCOX-2 Ab tive positive OR (multivariable. MI)

HLA-DRB1*15:01 Absent 115 (71.4) 18 (18.0) Present 46 (28.6) 82 (82.0) 14.96 (6.40-34.98. p<0.001)

Gender Female 82 (48.5) 65 (60.7)

Male 87 (51.5) 42 (39.3) 0.77 (0.34-1.72. p=0.516)

Age at dg (years) Mean (SD) 38.9 (17.6) 61.4 (13.7) 1.34 (1.16-1.55. p<0.001) Age at dg^A2 Mean (SD) 1821.9 (1636.7) 3956.6 (1612.1) 1.00 (1.00-1.00. p=0.004) PNH clone Absent 78 (53.4) 41 (42.3) Present 68 (46.6) 56 (57.7) 1.18 (0.51-2.73. p=0.698)

Severity Moderate 41 (26.5) 28 (28.0)

Severe 101 (65.2) 61 (61.0) 1.17 (0.37-3.68. p=0.782)

Very se 13 (8.4) 11 (11.0) 1.49 (0.26-8.58. p=0.657) vere

Hb at dg (g/dl) Mean (SD) 8.9 (2.1) 9.5 (10.1) 1.08 (0.90-1.29. p=0.412) WBC at dg (10⁹/1) Mean (SD) 2.4 (1.5) 2.5 (1.3) 1.18 (0.59-2.38. p=0.637) Pit at dg (10⁹/1) Mean (SD) 40.8 (46.4) 21.6 (16.8) 1.34 (1.16-1.55. p<0.001) ANC at dg (10⁹/1) Mean (SD) 0.9 (1.0) 0.9 (0.7) 0.64 (0.26-1.59. p=0.336) ALC at dg (10⁹/1) Mean (SD) 1.2 (0.8) 1.4 (0.7) 0.99 (0.38-2.55. p=0.980) 85.6 % of aCOX-2 Ab -positive IAA patients expressed the HLA-DRB1^*15 allele

(15:01 or 15:02), while it was only 31.2% in Ab negative patients (pos n = 90, neg n = 144, Pearson’s Chi² test p-value <0.001, Figure 6C). Considering only HLA- DRB1^*15 -positive IAA cases, the overall frequency of aCOX-2 antibodies in com bined cohorts is 63 % (See ROC curve in Figure 3C). Frequency of paroxysmal nocturnal hemoglobinuria (PNH) clone in white blood cells (CD15+ FLAER nega tive clone size <0.1 % is considered negative and >0.1 % is positive) did not differ between aCOX-2 Ab positive and negative patients (Figure 6D).

The expression of PTGS2 and biological role of aCOX2-ab

To understand in which cells COX-2 is expressed in human bone marrow, we re analyzed a recently published droplet-based single-cell RNA sequencing (scRNAseq) data set from 22 samples from 20 healthy bone marrow donors (Oet- jen KA et al. 2018, JCI Insight. 3(23)). Of the 22 cell phenotypes found in the bone marrow (Figure 7A), PTGS2 gene encoding the COX-2 protein was mainly ex pressed in hematopoietic stem cells (FISC) and myeloid cells, mostly in monocytes (Figure 7B).

Next, to elucidate whether the PTGS2 expression levels differ between AA and MDS patients and healthy controls we analyzed another plate-based scRNAseq dataset from sorted CD34+ populations (Zhao X et al. 2017, Blood. 130(25):2762- 2773). All 4 healthy controls and 1 MDS patient had sufficient amount for sorting of CD38+ and CD38- populations, but all 3 AA patients and 1 MDS patients had lim ited number of cells due to BMF and thus only CD34+ Lineage- cells were ana lyzed. Of the 10 clusters, 2 were specific for an MDS patient and 1 was specific for a healthy donor, but the other identified populations such HSC, promonocyte and granulocyte/macrophage progenitors (GMP) were seen in all different disease states (Figure 7C - F). For the different clusters, we noted a statistically significant difference between the PTGS2 expression in the HSC cluster, where AA patients had the highest expression (p < 0.05, Mann-Whitney in comparison to healthy, Figure 7G). Thus, our analysis suggests that PTGS2 expression is restricted to HSC and myeloid cell lineages, and COX-2 is upregulated in AA HSCs in the bone marrow poising its potential as an autoantigen.

Summary

Taken together, we have identified a novel autoantibody aCOX-2 that is highly specific for IAA. The immunogenic epitopes where the diagnostic aCOX-2 autoan tibodies bind, were mapped to the C-terminal end of the COX-2 protein. The novel aCOX-2 Ab test has great potential as a clinical biomarker as it is cheap and easy to perform using e.g. pre-existing technical platforms in clinical laboratories and al lows the use of plasma /serum as sample material.

36 % of all IAA patients (37 % of all adult IAA patients in the appended cohort) of the present study tested positive for this autoantibody aCOX-2, whereas none of the healthy controls had the antibody. Importantly, aCOX-2-Ab positivity was sig nificantly associated with the HLA-DRB1 ^*15:01 haplotype, with aCOX-2 Ab positiv ity rising to 63 % in HLA-DRB1^*15-positive adult IAA and to 83% of HLA- DRB1 ^*15:01 positive and over 40 years old IAA patients. In addition to IAA, spo radic aCOX-2-Ab positive cases were identified in related bone marrow failures diseases, suggesting a common pathomechanism for these patients.

Importantly, healthy controls and RA patients, that served as controls to exclude the systemic inflammation-associated expression of aCOX-2 antibody, did not give rise to positive aCOX-2-Ab findings. Within the Helsinki biobank cohort including 100 patients with other autoimmune diseases, we identified two aCOX-2-Ab posi tive patients that had been diagnosed with multiple sclerosis (MS). This is intri guing, since MS is also associated with the DRB1 ^*15:01 (Lincoln MR et al. 2005, Nat Genet. 37(10):1108—1112) haplotype.

In the Finnish patient cohort, 61 % of all immune mediated AA patients (62 % of the adult IAA patients) were aCOX-2-Ab positive. The positivity rate was significantly lower in other tested populations; 27% (30 % of the adult IAA patients) in the US and 37% (44 % of the adult IAA patients) in the Japanese cohort. The study pa tients’ ages of receiving AA diagnosis varied between cohorts, with US and in lesser extent Japanese representing more younger cases. These younger cases were shown to be more frequently negative for the antibody. The population-level differences may also, at least in part, be explained by the differences in the clinical practice to assign the diagnosis of IAA. aCOX-2-Ab positivity neither correlated with 1ST treatment responses nor with the presence of PNH clone. However, aCOX-2-Ab positive patients were older, and had statistically significantly lower platelet counts at diagnosis.

Claims

Claims

1. A method for determining a bone marrow failure disorder of a subject or the ab- sence thereof, or subclassifying a subject having a bone marrow failure disorder, wherein the method comprises determining an anti-COX-2 autoantibody or a poly nucleotide encoding said anti-COX-2 autoantibody or a fragment of said autoanti body or polynucleotide from a sample of a subject. 2. The method of claim 1, wherein the presence or an increased level of the anti-

COX-2 autoantibody or the polynucleotide encoding said anti-COX-2 autoantibody or a fragment of said autoantibody or polynucleotide in the sample of the subject indicates the bone marrow failure disorder of the subject, or the presence or an in creased level of the anti-COX-2 autoantibody or the polynucleotide encoding said anti-COX-2 autoantibody or a fragment of said autoantibody or polynucleotide in the sample of the subject indicates a subgroup of bone marrow failure disorders.

3. An anti-COX-2 autoantibody or a polynucleotide encoding said anti-COX-2 au toantibody or a fragment of said autoantibody or polynucleotide for determining a bone marrow failure disorder of a subject, or for use as a diagnostic marker.

4. An anti-COX-2 autoantibody or a polynucleotide encoding said anti-COX-2 au toantibody or a fragment of said autoantibody or polynucleotide, wherein said au toantibody is capable of binding one or more amino acids at positions 1 - 604, 56 - 604, 100 - 604, 200 - 604, 300 - 604, 350 - 604, 442 - 604, 490 - 604, 530 -

604, or 490 - 600 as shown in SEQ ID NO: 1.

5. The anti-COX-2 autoantibody of claim 3 or claim 4, wherein said autoantibody is a human autoantibody, optionally an isolated human autoantibody.

6. The anti-COX-2 autoantibody of any one of claims 3 - 5, wherein said autoanti body is a polyclonal autoantibody.

7. A test kit for determining a bone marrow failure disorder, or for determining the anti-COX-2 autoantibody or a polynucleotide encoding said anti-COX-2 autoanti body or a fragment of said autoantibody or polynucleotide of a sample, wherein the kit comprises one or more tools for determining the anti-COX-2 auto antibody of claim 4 or a polynucleotide encoding said anti-COX-2 autoantibody or a fragment of said autoantibody or polynucleotide from a sample of a subject, and optionally reagents for the test; and/or the kit comprises a capture complex for binding an anti-COX-2 autoantibody, or a polynucleotide encoding said anti-COX-2 autoantibody or a fragment of said auto antibody or polynucleotide.

8. A method for determining an anti-COX-2 autoantibody or a polynucleotide en coding said anti-COX-2 autoantibody or a fragment of said autoantibody or poly nucleotide from a sample of a subject, wherein the method comprises allowing a capture complex to bind with an anti-COX-2 autoantibody or a polynucleotide en coding said anti-COX-2 autoantibody or a fragment of said autoantibody or poly nucleotide, and determining said anti-COX-2 autoantibody, polynucleotide or fragment or amount thereof.

9. The test kit or method of claim 7 or 8, wherein the capture complex is a poly peptide (such as a recombinant polypeptide) or a fragment thereof or a combina tion of different polypeptides comprising one or more polypeptides (such as re combinant polypeptides) or fragments thereof for binding the anti-COX-2 autoanti body; or a polynucleotide (such as a recombinant polynucleotide) or a fragment thereof or a combination of different polynucleotides comprising one or more poly nucleotides (such as recombinant polynucleotides) or fragments thereof for bind ing the polynucleotide encoding the anti-COX-2 autoantibody; or a recombinant human COX-2 polypeptide or any fragment thereof or a combination of different polypeptides comprising one or more recombinant human COX-2 polypeptides or fragments thereof for binding the anti-COX-2 autoantibody; or a recombinant hu man COX-2 polynucleotide or a fragment thereof or a combination of different pol ynucleotides comprising one or more recombinant human COX-2 polynucleotides or fragments thereof for binding the polynucleotide encoding the anti-COX-2 auto antibody.

10. The test kit or method of claim 9, wherein the polypeptide or polynucleotide of the capture complex or any fragment thereof is tagged, optionally with one or more tags selected from the group consisting of a polyhistidine-tag, HA-tag, HN-tag, HQ-tag, FLAG-tag, Myc-tag, V5-tag, GST-tag, MBP-tag, Strep-tag and CL7-tag.

11. The test kit or method of any of claims 7 - 10, wherein the test kit comprises tools for an enzymatic assay and/or immunoassay, or the method is based on an enzymatic assay and/or immunoassay. 12. The test kit or method of any of claims 7 - 11, wherein the test kit comprises a label, agent, molecule or any combination thereof for detecting the anti-COX-2 au toantibody or the polynucleotide encoding said anti-COX-2 autoantibody, or the method comprises utilization of a label, agent, molecule or any combination there of for determining the anti-COX-2 autoantibody or the polynucleotide encoding said anti-COX-2 autoantibody.

113. The test kit or method of claim 12, wherein the molecule for detecting or de- termining the anti-COX-2 autoantibody or the polynucleotide encoding said anti- COX-2 autoantibody is a detection antibody, optionally labelled.

14. The test kit of any of claims 7 - 13, wherein the test kit is for the method of any of claims 1 - 2 or 8 - 13.

15. The method, autoantibody, polynucleotide, fragment or test kit of any preced ing claim, wherein an isotype of the autoantibody is IgA, IgD, IgE, IgG or IgM, typi cally IgG, IgA or IgM; or wherein an isotype of the autoantibody is IgG and the subclass is lgGi, lgG2, lgG3 and/or lgG4.

16. The method, autoantibody, polynucleotide, fragment or test kit of any preced ing claim, wherein the autoantibody is capable of binding the C-terminal part of a COX-2 polypeptide, or the C-terminal part at one or more amino acid positions within amino acids 56 - 604, 442 - 604, 490 - 604, 530 - 604, 577 - 604, 580 - 594, or 490 - 600 as shown in SEQ ID NO: 1.

17. The method, autoantibody, polynucleotide, fragment or test kit of any preced ing claim, wherein the autoantibody is capable of binding one or more of the fol lowing amino acid fragments of the COX-2 polypeptide: PALLVE (SEQ ID NO: 2), STFGGEV (SEQ ID NO: 3), GFQIINT (SEQ ID NO: 4), QIINTAS (SEQ ID NO: 5), VTIN (SEQ ID NO: 6), GLDDIN (SEQ ID NO: 7), and/or DINPTVL (SEQ ID NO: 8).

18. The method, autoantibody, polynucleotide, fragment or test kit of any preced ing claim, wherein the sample is a blood, plasma, serum, bone marrow aspirate, or any body fluid sample.

19. The method, autoantibody, polynucleotide, fragment or test kit of any preced ing claim, wherein the bone marrow failure disorder is inherited or acquired. 20. The method, autoantibody, polynucleotide, fragment or test kit of any preced ing claim, wherein the bone marrow failure disorder or the subgroup of the bone marrow failure disorder is selected from the group consisting of aplastic anemia (AA), paroxysmal nocturnal hemoglobinuria (PNH), polycythemia vera (PV), Dia- mond-Blackfan anemia (DBA), essential thrombocytosis (ET), Fanconi anemia (FA), large granular lymphocyte (LGL) leukemia, myelodysplastic syndrome (MDS), telomere syndromes, dyskeratosis congenita, immune thrombocytopenic purpura (ITP), idiopathic cytopenia of undetermined significance (ICUS), clonal hematopoiesis of indeterminate potential (CHIP) and idiopathic bone marrow dys plasia of undetermined significance (IDUS).

21. The method, autoantibody, polynucleotide, fragment or test kit of any preced ing claim, wherein the sensitivity of the method, autoantibody, polynucleotide, fragment or test kit is more than 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, or 99%; and/or the specificity is more than 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95% or 99%.

22. Use of the test kit of any of claims 7 - 21 or anti-COX-2 autoantibody or a poly- nucleotide encoding said anti-COX-2 autoantibody or a fragment of said autoanti body or polynucleotide for determining a bone marrow failure disorder of a subject.

23. Use of an anti-COX-2 autoantibody or a polynucleotide encoding said anti- COX-2 autoantibody or a fragment of said autoantibody or polynucleotide as a di- agnostic marker.

24. The use of claim 22 or 23, wherein the anti-COX-2 autoantibody or a polynu cleotide encoding said anti-COX-2 autoantibody or a fragment of said autoanti body or polynucleotide is according to any of claims 4 - 6 or 15 - 17.