Jump to: Authorized Access | Attribution | Authorized Requests

Study Description

The purpose of this study is to characterize patients with Bone Marrow Failure Syndromes (BMFS) and to track the clinical course of patients with these diseases over time. An exact and comprehensive diagnosis at presentation is essential. Due to the sporadic nature of BMFS (aplastic anemia, myelodysplastic syndrome, paroxysmal nocturnal hemoglobinuria, pure red cell aplasia, amegakaryocytic thrombocytopenic purpura and large granular lymphocyte leukemia) most diseases have not undergone systematic long-term outcome studies.

Individual diseases, combined under the collective entity of bone marrow failure syndromes, frequently overlap and evolve from each other. The frequency of complications, associations with other diseases, and genetic factors such as allelic polymorphisms are not well studied. Some of these syndromes are likely mediated by the cellular immune system. However, while target antigens are likely to be expressed on early hematopoietic cells, their identification has been unsuccessful.

Hypothesis: Clinical and epidemiologic studies may, through etiologic clues and intricate laboratory testing, facilitate the understanding of the pathophysiology of BMFS and ultimately lead to implementation of better diagnostic tools and more targeted and rational therapies.

Authorized Access
Publicly Available Data (Public ftp)
Study Inclusion/Exclusion Criteria

Eligibility: Patients 18 years of age or older with aplastic anemia, Myelodysplastic Syndrome, Paroxysmal nocturnal hemoglobinuria, Idiopathic pure red cell aplasia, Amegakaryocytic thrombocytopenic purpura and Large granular lymphocyte leukemia who are capable of giving a truly informed consent are eligible for enrollment in the Study. New and previously diagnosed patients can participate. Children ages 11-17 years of age are eligible with consent of their parent/legal guardian.

Study History

Population studied includes patients with one or more BMFS diagnosis of:

  • Aplastic Anemia (AA)
  • Myelodysplastic Syndrome (MDS)
  • Paroxysmal nocturnal hemoglobinuria (PNH)
  • Pure red cell aplasia (PRCA)
  • Amegakaryocytic thrombocytopenic purpura (ATP)
  • Large granular lymphocyte leukemia (LGL leukemia)

Clinical and pathophysiologic aspects of BMFS

BMFS result from a global or lineage-restricted failure of hematopoietic progenitor or stem cells. As in many orphan diseases, assignment of specific etiologies and application of therapeutic trials have been difficult due to their sporadic occurrence. Apart from these characteristics, certain rare diseases are very instructive with regard to specific pathophysiologic mechanisms and genetic factors.

In hematology, AA demonstrates the efficacy of the cellular immune system capable, under pathologic conditions, of destruction of an entire organ (bone marrow) without collateral damage to other organ systems. The extraordinary specificity of immune attack is illustrated by the lineage-restriction of hematopoietic suppression in LGL leukemia. Similarly, restoration of normal blood counts after successful therapy of AA suggests the regenerative capacity of hematopoietic stem cells. The replacement of most of the normal blood cells with those of a PNH phenotype, and their persistence for years, suggest a single hematopoietic stem cell can repopulate the bone marrow and provide sustained blood cell production for years. These simple examples demonstrate that the study of distinct rare syndromes may reveal physiologic mechanisms of pathophysiologic and physiologic importance for other, perhaps more common, conditions.

The need for a database and systematic longitudinal studies and multicenter cooperative research is well illustrated by the history of clinical research in AA. Before the use of antithymocyte globulin (ATG) for severe AA, this disease had a high mortality and a median survival of approximately 6 months. Today medical therapy results in a median survival exceeding 10 years1. The value of ATG therapy for AA was first demonstrated in a randomized controlled trial conducted at UCLA by Richard Champlin more than 20 years ago 2. The therapeutic benefits of ATG were confirmed in severe AA, and extended to patients with moderate AA in a subsequent US multicenter trial 3. Since that time, the major studies evaluating medical therapies for AA have been made by European cooperative groups, primarily because no single US center can enroll enough patients to conduct such a trial, and there is no mechanism by which multicenter studies can be performed. The German Aplastic Anemia Study Group demonstrated that adding cyclosporine A (CsA) to ATG significantly improved the response rate compared to ATG alone in the treatment of AA 4. The European Blood and Marrow Transplantation (EBMT) Severe Aplastic Anemia Working Party showed that concurrent use of ATG and CsA induced a significantly higher response rate than CsA alone in patients with moderate AA 5. The development of many new immunosuppressive and immunomodulatory drugs in the last 10 years holds particular promise for patients with bone marrow failure. However, the absence of a cooperative group in the US committed to the systematic evaluation of these drugs in rare diseases such as idiopathic BMFS is largely responsible for the recent lack of treatment advances.


AA is a rare idiopathic disease of the hematopoietic stem cells with an incidence in the Western World of around 4/106 (see 6 for review of the topic). Typically, AA is a disease of young adults and can affect children or even infants with a possible second peak between the ages of 60-70 years. The etiologic factors in AA are unknown and may include infection with an unknown viral pathogen, idiosyncratic reaction to chemicals or alteration of the normal proteins by mutations or chemical modification 7;8. AA is believed to be induced by an immune-mediated attack on early hematopoietic cells. Clinically, AA can be classified according to its severity and acuity. Its diagnosis is based on bone marrow biopsy and peripheral blood counts criteria of the International Study of AA and Agranulocytosis. Severity can be classified by the criteria of Camitta et al 9. A moderate form of AA, often referred to as hypoplastic anemia, may also exist and is characterized by a less than severe depression of at least two hematopoietic lineages and hypocellular marrow. This disease may be a stage of severe AA, but it may also represent a separate disease with a chronic and non-progressive moderate cytopenia. Clearly, the incidence and natural history of this more indolent form of AA is less well defined. Other clinically distinct forms of AA include AA/hepatitis syndrome occurring after often fulminant seronegative hepatitis 10. AA/hepatitis syndrome constitutes a model of virally mediated bone marrow failure.

Prior to the advent of modern supportive care and specific therapies, severe AA was a very deadly disease. Currently, immunosuppressive therapy (IS) with ATG, steroids and cyclosporine A (CsA) can induce remission in a significant proportion of patients (50-70%)11. For younger patients with a matched sibling donor, allogeneic bone marrow transplantation (BMT) is a treatment of choice but in older patients, due to the higher treatment-related mortality, intense immunosuppression and BMT produce comparable results 7;11. However, a significant proportion of patients do not respond to immunosuppressive therapy and allogeneic BMT is not an option due to the lack of a matched sibling donor. These patients show a poor long-term survival.

Despite progress, AA remains a serious and often fatal disease and may represent a significant diagnostic challenge. Hypocellularity of the bone marrow may hamper the attainment of informative chromosomal analysis and morphologic diagnosis of dysplasia in the setting of aplasia may be difficult 6. Consequently, a number of patients with hypocellular MDS may be mistakenly diagnosed as AA. The relapse rate of AA in patients treated with immunosuppressive therapy is notoriously high, but a repeated course of immunosuppression appears to induce remission in a majority of these patients. However, AA shows a number of long-term complications 12. The most significant complication of AA is the evolution of clonal disease including PNH, MDS and AML. The evolution rate of MDS approaches 20% at 10 years. Among those with MDS, trisomy-8 and monosomy-7 are the most frequent abnormalities. Particularly, monosomy-7 and complex cytogenetic abnormalities carry a very serious prognosis 13.


MDS is a heterogeneous group of clonal stem cell disorders characterized by the paradox of a failure to produce blood cells in the setting of a hyperproliferative bone marrow. Although past classification schemes have been based on the commonality of morphologic dysplasia of hematopoietic cells, individual disorders can be subcategorized based on the clinical similarities or specific molecular defects, and to some extent this is reflected in newer classification schemes 14;15. Unlike typical AA, MDS is a disease of older adults and its incidence increases with age. MDS is associated with severe morbidity related to the complications of cytopenias. Except for bone marrow transplantation, an option available to a minority of patients (estimated to be < 5%) due to age and lack of matched sibling donors, no curative therapies exist and most of the currently available therapies have a low efficacy or unacceptable toxicities. In some patients, MDS progresses steadily to AML, while in others the disease follows a chronic course with persistence of even high numbers of blasts for long periods of time without progression to a more aggressive form 14;16;17. The heterogeneity of clinical courses, diverse prognosis, and the wide spectrum of genetic defects observed strongly suggest that this syndrome is comprised of various disease entities, and traditional classifications do not address the need for individualized treatment of distinct sub-entities.

Global, non-targeted approaches to MDS hampers the research and the development of effective therapies for this disease. However, over time the realization that specific forms of MDS exist has resulted in a definition of a growing number of individually rare syndromes, such as the 5q- syndrome and refractory anemia with ringed sideroblasts. Clearly, other separate disease sub-entities should be investigated separately. In addition to the acquired molecular defects in the hematopoietic stem cells, cytopenias occurring in the course of MDS are difficult to explain solely by a mechanism involving the displacement of normal hematopoiesis, and it is likely that immune mechanisms may also play a role 18. For example, it is possible that inherently abnormal dysplastic hematopoietic clones may carry new antigens that can incite the autoimmune process that also affect normal progenitor cells. Although the determinants of a patient's ability to mount immune responses to dysplastic hematopoiesis are not known, both immunogenetic (e.g., HLA-types) as well as phenotypic factors intrinsic to the target cells can be involved in a fashion similar to the graft-versus leukemia effect after allogeneic bone marrow transplantation,. Development of MDS could be consistent with only a partial ability to mount an immune response to such antigens and inhibit clonal evolution.


PNH is an acquired clonal stem cell disorder characterized by a clinical triad of hemolysis, thrombosis and bone marrow failure. The incidence of PNH in the general population is not known but it is believed that this disease is less frequent than AA. PNH results from a somatic mutation in the PIG-A gene, responsible for the biosynthesis of the GPI-anchor 19-21. Progeny of all hematopoietic lineages can carry identical PIG-A mutations and have the characteristic PNH phenotype. The inability to attach a GPI-anchor to normally GPI-anchored proteins (GPI-AP) results in their deficiency in the membrane of affected cells, a feature that can be easily recognized by immunophenotyping 22-24. Over 100 GPI-AP have been discovered to date 20;25;26, but the deficiency of certain GPI-AP explains some of the clinical manifestations of PNH. For example, susceptibility of erythrocytes to complement-mediated lysis is related to the lack of membrane CD55 and CD59 in PNH red cells. However, the pathogenesis of other symptoms of PNH, such as thrombosis and bone marrow failure, remains elusive. Although absolutely required, the PIG-A mutation does not explain the evolution of PNH to its manifest disease state. The nature of the growth advantage for PNH stem cells remains unclear but some permissive conditions appear to be required for the expansion of the PNH clone because GPI-AP-deficient granulocytes are present in normal individuals 27;28 but manifest PNH is a rare disease. PNH often evolves from AA, but GPI-AP deficient clones are detected in a significant proportion of AA patients at presentation suggesting that an immune attack on early hematopoietic cells as observed in AA may constitute permissive conditions providing for the growth advantage of defective stem cells over their normal counterparts. Supportive measures including anticoagulation in patients with thrombotic complications, iron supplementation and transfusions are the most important components of the therapy for PNH. Patients with cytopenia may be treated with immunosuppressive agents similar to AA. There are no specific therapies and the efficacy of steroids and androgens remain controversial due to the lack of controlled prospective studies. Similarly, the value of antithrombotic therapy remains unclear 20;29;30.

Single lineage cytopenias including PRCA, ATP and LGL leukemia.

Similar to AA, immune pathogenesis has been postulated for single lineage cytopenias such as PRCA and ATP. Many epidemiologic and clinical observations suggest a viral etiology but immunogenetic factors may also play a role. However, it is likely the cytotoxicity in these diseases is restricted to committed progenitor cells of individual hematopoietic lineages. Although their incidence is not known it is likely lower than that of AA. PRCA aplasia may be associated with thymoma or lymphoid malignancies. Some cases may also represent transition stages of evolving AA.

A very peculiar association of red cell aplasia or neutropenia is that of LGL leukemia. LGL leukemia has been observed in the context of different autoimmune conditions and viral infections. In essence, LGL leukemia represents a clonal proliferation of cytotoxic T cells. Although an intrinsic defect within the LGL CTL is likely to facilitate their clonal expansion, clearly this process is "fueled" by a persisting antigenic drive. While the observed pathology may be a result of the specificity of LGL, it is possible that LGL represents an autoimmune T cell clone that, due to its persistent proliferation, is more likely to acquire a genetic defect. LGL can occur in rheumatoid arthritis (RA)31;32, Sjorgren's 33;34, and Felty's syndrome 32;35 and can also be present with immune-mediated cytopenias 26;36-39. The specificity of LGL-mediated effects can be inferred from clinical observation of an isolated neutropenia or PRCA, indicating that the broad cytokine effects are not sufficient to explain the observed pathology. Similar to idiopathic AA, LGL leukemia and other single lineage cytopenias have been shown to respond to immunosuppression 36;40-42. Clinically, a large proportion of patients with PRCA and ATP remain refractory to current therapies and these diseases are associated with a long-term need for transfusions and inherent complications.

Etiologic links

A similar spectrum of possible inciting events has been hypothesized and investigated for idiopathic BMFS. Possible etiologic factors include viruses or chemical agents but it is likely that immunogenetic factors may also play a role in disease development. For AA, its increased prevalence in certain areas of the world suggests an endemic pathogen but the search for the etiologic agents has been unsuccessful. Certain rare cases of AA have been attributed to EBV 43;44, and an unknown viral pathogen has been strongly suggested in a distinctive sero-negative AA/hepatitis syndrome 10. Clonally polarized proliferation of CTL can occur during certain infections e.g., EBV or HCMV, and per interference, a viral pathogen such as HTLV-1 has been hypothesized to be a causative agent in LGL leukemia 45;46 and also other diseases accompanied by often severe cytopenias. Although large epidemiologic studies demonstrated a clear association of specific drugs, the etiologic fraction that can be assigned to them is small (see 6 for review). Except for agranulocytosis, with clear associations with drugs, for PRCA, and ATP, specific drug exposures have been implicated but none of them is sufficient to explain the pathophysiology of these diseases.

While various inciting events may trigger the pathologic cascade in bone marrow failure, an immune-mediated process is likely the major pathophysiologic mechanism responsible for the observed pathology. Immune-mediated destruction of hematopoiesis in idiopathic AA has been inferred from the success of immunosuppressive therapies 11;47;48. Similar therapies can also produce hematologic improvement in certain forms of MDS and are often applied in the therapy of PRCA and LGL leukemia 18;49;50, suggesting that similar pathophysiologic mechanisms may operate in these diseases.

Theoretically, the inciting antigen may be a foreign protein that either induces breach of tolerance or shows cross-reactivity with certain normal hematopoietic antigens. In addition, cross-reactive antigens could be generated by chemical modification, conjugation with drugs, or could be the product of a mutated gene. It is likely that the nature of the antigen determines the clinical presentation and specific features of bone marrow failure in an individual patient. Experimental evidence supports an immunologic mechanism in AA (for review see 7;8;51). A role for T cells in AA was first suggested by co-culture and depletion experiments, in which inhibition of hematopoietic colony formation was associated with this lymphocyte population 52;53. Later, an inverted CD4/CD8 ratio 54, activated cytotoxic lymphocytes (CTL) as detected by the expression of HLA-DR 55 and CD25 56, and skewing of the variable δ chain (VB)-repertoire of the T cell receptor (TCR) were found, consistent with expansion of autoimmune T cell clones 57-65. T lymphocytes from AA overproduced interferon-γ (IFN-γ) and tumor necrosis factor-α (TNF-α) consistent with a shift in the Th1/Th2 balance 66-68. Similar cytokine-mediated effects have also been described for MDS and LGL leukemia. Specific T cells express activation markers such as HLA-DR, CD25, CD69 and secrete IFN-γ during the immune attack on the hematopoietic progenitor and stem cells. Due to antigen-specific T cell receptor triggering and co-stimulatory signals, effector cells differentiate from a proliferating naive or memory cell pool. This process is associated with CD45 isotype switching and loss of CD28 69-71. Experimental evidence suggests that T cells mediating hematopoietic suppression in AA and MDS are contained within the mature effector cell population. In LGL leukemia, the effector cells show a phenotype typical of mature CTL, but unlike in AA, MDS, ITP or PRCA in which the T cell response is polyclonal, they show an extreme clonal polarity. It is likely these diseases represent less extreme pathophysiologic equivalents of a frank LGL leukemia. Consequently, this disease may represent a good experimental model to investigate more complex clonal responses seen in classical BMFS.

Analysis of TCR repertoire including polymorphisms within the VB CDR3 of the TCR has been utilized for the study of autoimmune diseases and viral infections 72;73. Consequently, CDR3 spectratyping has also been performed in patients with AA, PNH, MDS and LGL 27;45;57-62;74-76.

Hematopoietic progenitors and stem cells are cellular targets for an immune attack in BMFS 8. While in single lineage cytopenias, committed progenitor cells are likely to be targets (e.g., erythroid precursors in PRCA) in AA. The immune inhibition also includes multipotent progenitor cells. We have demonstrated a profound hematopoietic depletion of CD34+ cells 77 and of long-term colony forming cells (LTCIC), the most immature hematopoietic cells that can be measured in vitro 78;79 implying that these populations contain the target antigen for the immune attack. CD34+ cells from AA and MDS show increased expression of Fas 80 and contain a high proportion of apoptotic cells 81. In AA/PNH syndrome, the immune attack appears to selectively spare GPI-AP- deficient hematopoietic progenitor cells, a process that could potentially explain the expansion of the PNH clone in patients 82;83. We have demonstrated that residual, phenotypically normal CD34+ fraction contains a much higher proportion of apoptotic cells than GPI-deficient counterparts 82.

Pathophysiologic relationships

The close pathophysiologic association between individual BMFS and cytopenias has been concluded from clinical observation. For example, MDS often evolves from AA, and hypoplastic MDS shows many clinical similarities to AA 84;85. Analogous to typical AA, MDS cases responsive to immunosuppression are mostly hypocellular and are associated with the presence of a paroxysmal nocturnal hemoglobinuria (PNH)22 clone and a high frequency of HLA-DR2 22;86. PNH often evolves in the setting of AA and immune mechanisms in both disease may be closely related 19;22. In a large portion of AA patients, GPI-AP deficient clones are present at initial presentation 22. Such patients show good responsiveness to immunosuppressive therapy and a high prevalence of HLA-DR15 86;87. We often refer to this frequent presentation of AA as AA/PNH syndrome. Interestingly, analogous to AA/PNH, MDS cases responsive to immunosuppression are mostly hypocellular and are also associated with the presence of a PNH clone and HLA-DR15 86;87. The frequent coexistence of large granular lymphocytosis (LGL) and PNH, MDS, AA or uni-lineage cytopenias also indicates that a T cell-mediated process may be part of the pathophysiology operating in a large proportion of bone marrow failure cases 41;75;88. In general, the overlap between BMFS and LGL suggests a common pathophysiology characterized by a T cell-mediated process directed against various hematopoietic targets 37;38;41;75;88;89. It is likely that a pathophysiologic continuum exists from frank aplasia, hypoplastic MDS to refractory anemia with blast excess in which immune mechanisms are ineffective or selection pressure has facilitated clonal escape 65. Characterization of immune response in these diseases may provide not only information about the genesis of bone marrow failure but also about the immune control of leukemia. It is likely, that common clinical features reflect the similarity of the pathogenic factors including inciting antigens.

Clinical relationships between BMFS are also reflected by the spectrum of therapies showing efficacy in these diseases. Immunosuppressive agents including ATG, CsA and cytoxan (Ctx) are the most frequent therapies and have proven effective in PRCA, LGL leukemia and certain forms of MDS. While intense IS with high doses of Ctx have many toxicities it shows a high efficacy 90. Regimens involving oral agents have been effective in less severe cytopenias. MDS constitutes a significant therapeutic challenge with many very indolent but refractory forms such as refractory anemia with ring sideroblasts (RARS) or 5q- syndrome. Other forms, including certain specific chromosomal aberrations show an aggressive course with a rapid evolution of leukemia. Clearly, the rational therapies are the least developed for MDS due to the lack of pathophysiologic clues.

Clinical relationships between BMFS are also reflected by the spectrum of therapies showing efficacy in these diseases. Immunosuppressive agents including ATG, CsA and cytoxan (Ctx) are the most frequent therapies and have proven effective in PRCA, LGL leukemia and certain forms of MDS. While intense IS with high doses of Ctx have many toxicities it shows a high efficacy 90. Regimens involving oral agents have been effective in less severe cytopenias. MDS constitutes a significant therapeutic challenge with many very indolent but refractory forms such as refractory anemia with ring sideroblasts (RARS) or 5q- syndrome. Other forms, including certain specific chromosomal aberrations show an aggressive course with a rapid evolution of leukemia. Clearly, the rational therapies are the least developed for MDS due to the lack of pathophysiologic clues.

Selected Publications
Diseases/Traits Related to Study (MeSH terms)
Links to Related Resources
Authorized Data Access Requests
See research articles citing use of the data from this study
Study Attribution
  • Principal Investigator
    • Jaroslaw P. Maciejewski, MD, PhD. Cleveland Clinic Foundation Cleveland, Ohio, USA.
  • Funding Source
    • National Institutes of Health, Bethesda, MD, USA.