Case report

A sustained response to low dose interferon-a in a case of refractory pure red cell aplasia

Chronic pure red cell aplasia (PRCA) is a raredisorder affecting erythroid precursors in the bonemarrow. In some cases, cell-mediated immunedestruction of erythroid progenitors may occur.Treatment is attempted with immunosuppressiveagents such as anti-thymocyte globulin (ATG),cyclosporin A (CsA), and more recently ale-mtuzumab, with variable responses (1–4). Ale-mtuzumab targets lymphocytes expressing CD52,including natural killer (NK) cells, which may beresponsible for the selective destruction of eryth-roid precursors in the bone marrow of patients withPRCA (5). In these patients, the low level expres-sion of HLA class I antigens on developingerythroid precursors may render them susceptibleto KIR-mediated lysis by auto-reactive NK orcytotoxic T cells. This phenomenon was previouslydemonstrated in malignant melanoma cell lines,and is also thought to operate in PRCA (6–9).

Rapid but transient responses to alemtuzumabtherapy in a patient with PRCA led us to consider acytokine-mediated effect of alemtuzumab in thisdisorder. Alemtuzumab causes release of tumor

necrosis factor-a, interferon-a (INF-a), and inter-leukin-10 in vivo (10). INF-a is of particular interestas it is an immune-modulating agent which increa-ses HLA class I expression on erythroid progenitors(11). Furthermore, Davidovitz et al. (12), observedcorrection of anemia in a patient with PRCA andhepatitis C after IFN-a therapy. We describe hereanother patient with refractory PRCA in whomIFN-a therapy led to a normal hematocrit andtransfusion independence. We also observed in-creased HLA class I expression on bone marrowerythroid precursors after therapy.

Patient and methods

A previously healthy 43-yr-old male presented withfatigue and dyspnea. Physical examination wasunremarkable aside from pallor. Severe anemiawith a hematocrit (hct) of 17%, and a reticulocytecount of 0.3% were noted. Iron studies, vitaminB12 and folate levels, serum protein electropheresis,LDH, bilirubin, and haptoglobin were normal.Hepatitis C antibody, hepatitis B surface antigen,

Nadeau L, Meyerson H, Warren G, Koc ON. A sustained response tolow dose interferon-a in a case of refractory pure red cell aplasia.Eur J Haematol 2004: 73: 300–303. � Blackwell Munksgard 2004.

Abstract: Acquired pure red cell aplasia (PRCA) may be the result of acellular or humoral autoimmune process. One proposed mechanism isthe destruction of erythroid progenitors by self-reactive, cytotoxic Tcells or natural killer (NK) cells. These cells normally express MHC classI receptors (KIR) which inhibit cytotoxicity when the target cellexpresses the HLA class I antigen(s) they bind. Therefore, loss of theseantigens on maturing erythroid progenitors may render them susceptibleto destruction by the pathogenic cells. Interferon-a (INF-a) increasesHLA class I expression on hematopoietic precursor cells. Therefore, weinitiated a trial of INF-a in a patient with refractory PRCA. Followingtreatment, he developed transfusion independence, and a sustainednormal hematocrit. Analysis of bone marrow erythroid cells revealed anincrease in expression of HLA class I molecules. INF-a should be usedin a controlled trial in patients with PRCA to determine its activity andmechanism of action.

Laura Nadeau, Howard Meyerson,Gregory Warren, Omer N. KoÅDivision of Hematology/Oncology, Department ofMedicine, Case Western Reserve University and theUniversity Hospitals of Cleveland, Cleveland, OH, USA

Key words: red cell aplasia; PRCA; refractory anemia

Correspondence: Omer N. KoÅ MD, Division ofHematology/Oncology, Department of Medicine, CaseWestern Reserve University, 10900 Euclid Avenue,Cleveland, OH 44106, USATel: (216) 368-1176Fax: (216) 368-1166e-mail: onk2@po.cwru.edu

Accepted for publication 4 May 2004

Eur J Haematol 2004: 73: 300–303Printed in UK. All rights reserved

Copyright � Blackwell Munksgaard 2004

EUROPEANJOURNAL OF HAEMATOLOGY

300

and hepatitis B surface antibody titers were negat-ive. Peripheral blood smear and flow cytometry didnot reveal an increase in large granular lympho-cytes (LGL). The bone marrow aspirate revealed80% cellularity, severely diminished erythroid pre-cursors, and an M : E ratio of 8 : 1. There were noviral inclusions, and karyotypic analysis was nor-mal. CT of the chest revealed no thymoma. HIVserology was negative, and Parvovirus titersshowed elevated IgG but absent IgM, consistentwith remote infection. PCR for Parvovirus B19 wasnegative. He was diagnosed with PRCA andgiven bi-monthly transfusion support, Prednisone,1 mg/kg/d, and weekly erythropoietin, 40 000 U.Because of lack of a response, intravenous immu-noglobulin, 0.5 g/kg, weekly for 4 wk, was added.Despite these measures, the patient required 2 unitsof red cells every other week. Therefore, ATG(20 mg/kg/d · 8 doses), and CsA, 300 mg bymouth twice daily, were given, while continuingerythropoietin and prednisone. Despite this, thereticulocyte count remained at 0–0.1%.Six months after presentation, the patient began

alemtuzumab therapy with a total cumulativedose of 120 mg given over 2 wk. Within 2 wk,the hct incremented to 30% and reticulocytes to3.1%. However, during the ensuing 4 wk, thereticulocyte count decreased to 0.3% with acorresponding decrement in hct. The patient wasthen re-treated with alemtuzumab (cumulativedose 120 mg over 2 wk). One week later, thereticulocyte count again increased to 3.2%, andhct to 32.5%, but 4 wk later, response was lost(Fig. 1).

In light of these brief responses to alemtuzumab,and prior reports documenting secondary INF-asecretion with this agent, we postulated that itseffect was mediated by increases in INF-a, which isknown to upregulate HLA class I protein expres-sion (10, 11). Previous reports have suggested thatincreased HLA class I expression on erythroidprecursors may protect them from KIR-mediatedlysis by NK cells (10, 11), so we initiated INF-a,3 million units s.c., three times per week. Within1 wk, the reticulocyte count and hct incremented to2.4% and 31.8%, respectively, leading to transfu-sion independence. After 5 wk of therapy, the hctreached 42% and he has remained transfusionindependent with 11 months follow-up (Fig. 1).

We also investigated HLA class I protein expres-sion on bone marrow erythroid precursors(CD45) CD71+) before and after INF-a therapy,and found a statistically significant increase in classI expression. Following lysis of the non-nucleatedred cells, approximately 5 · 105 total bone marrowcells were incubated with antibodies to CD71 (cloneL01.1, FITC conjugated), CD45 (clone 2D1, PerCPconjugated), and either HLA class I (clone G46–2.6, PE conjugated), HLA class II (clone L243, PEconjugated) or a PE-conjugated isotype control(BD Biosciences, San Jose, CA, USA). Fiftythousand to 100 000 total events were collectedand analyzed on a FACSCalibur flow cytometer(BD Biosciences) and analyzed using Cellquestsoftware (BD Biosciences). To determine molecularequivalents of surface fluorescence (MESF) of classI and II expression on the erythroid precursors,gates were applied on the CD45-negative and CD71

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

100 20 30 40 50 60 70

Weeks

% R

etic

uloc

yte

IFNA A

Fig. 1. Reticulocyte count over time: blood reticulocyte concentrations plotted against time from diagnosis in weeks. Thedecrease in reticulocyte count after interferon therapy reflects normalization of the patient’s hematocrit. A, alemtuzumabtherapy; IFN, interferon-a therapy given as 3 million units s.c., three times per week.

Red cells

301

bright cells. Prior to each analysis, the flow cytom-eter was calibrated using fluorescent-labelled beads(QuanitBRITE PE; BD Biosciences) with a knownamount of conjugated fluorochrome allowing forthe conversion of the fluorescent units into molecu-lar equivalents. Roughly 2% of all events wereCD45()) CD71 bright in the pretreatment analysisand 20% of all events were CD45()) CD71 brightin the analysis after therapy. Comparison of thehistograms and MESF were performed after nor-malization using Cellquest software’s Kolmogo-rov–Smirnov statistic. We found a mean of824 HLA class I molecules per precursor prior toIFN-a, and a mean of 1146 molecules per precursorafter treatment (P < 0.001).

Discussion

We report a striking response to INF-a in a patientwith refractory PRCA. PRCA is felt to be the resultof an underlying autoimmune process, and studieshave demonstrated that about 40% of cases arebecause of an IgG antibody directed againsterythroblasts (13). Other studies have demonstratedantibodies inhibiting burst forming unit-erythroid(BFU-E) or colony forming unit-erythroid (CFU-E)(14) and, rarely, erythropoietin (15). A cell-medi-ated mechanism may be etiologic in patientswithout a demonstrable antibody. This is supportedby the occasional effectiveness of anti-T-cell regi-mens in treating this disease, as well as theoccasional occurrence of PRCA in association withthymomas (16) and T-cell leukemia (17). Severalgroups have also shown an association of LGLswith PRCA (18–20).

One mechanism proposed by Handgretingeret al. (9) is that cytotoxic T or NK cells maycarry inhibitory MHC class I receptors (KIR)which inhibit cytotoxicity when the target cellexpresses the HLA class I antigen(s) they bind. Aserythroid progenitors in the bone marrow loseHLA class I antigens with maturation, dysreglu-ation of the pathogenic cells may lead to selectivedestruction of erythroid precursors in PRCA (11).This led to a trial of interferon in our patient, asINF-a is known to upregulate HLA class Iexpression on erythroid precursors (11). Subse-quently, we observed a low level, but significantincrease in HLA class I protein expression on thered cell progenitors. These results may have beenaffected by a significant difference in the numberof erythroid precursors in the bone marrow beforeand after therapy, because the small number oferythroid progenitors present before IFN-a ther-apy made the analysis difficult. However, becauseas few as 200 HLA class I molecules are thoughtto allow for target cell recognition by KIR-

expressing cytotoxic lymphocytes (9), minorincreases in HLA class I expression may deter-mine counter–receptor recognition. Overall, theseresults suggest a critical role for the level of HLAclass I proteins expressed on erythroid precursorsin the immunopathogenesis of PRCA.

Multiple agents are used with varying success totreat PRCA. Response rates for treatment withsteroids, chemotherapeutics, ATG, and CSA arereported at around 40%, 55%, 50%, and 65%,respectively (5, 21–24). Erythropoietin also hassome efficacy (13). Unfortunately, many patientsare refractory to these agents, and chemothera-peutics may cause secondary cancers. In ourpatient, INF-a induced rapid remission and asustained normal hematocrit. To date, there hasbeen one other case report of PRCA responsive toINF-a therapy, in a patient with hepatitis Cinfection (12). The patient’s PRCA went intolasting remission, and the authors speculated animmune mechanism separate from his hepatitis.There has similarly been one case of LGLleukemia with a durable response to interferon(25). The authors speculated that in this case,INF-a induced apoptosis of the leukemic cells. Asimilar mechanism may be etiologic in our patientbecause, although we could not identify an LGLclonal population, these expansions are typicallyvery small (<5%) (19). In future trials of INF-atherapy of PRCA, the KIR repertoires of NKcells, and the level of NK activity should beevaluated before and after therapy. This informa-tion may further elucidate the mechanism ofaction of IFN-alpha therapy in PRCA.

Finally, there is a report of spontaneousregression in up to 14% of patients with PRCA(26), and although we cannot definitively excludethis in our patient, the temporal relationship ofhis transfusion independence to INF-a therapymake it unlikely.

In conclusion, this is the second report of INF-ashowing therapeutic efficacy in a patient withrefractory PRCA. We propose that the mechanismof action involves induction of HLA class Iexpression on erythroid precursors, rendering themresistant to cytotoxic immune destruction.

Therefore, INF-a should be considered for use ina controlled trial in patients with PRCA to deter-mine its activity and mechanism of action.

References

1. Fisch P, Handgretinger R, Hans-Eckart S. Pure red cellaplasia. Br J Haematol 2000;111:1010–1022.

2. Pangalis G, Dimopoulou M, Angelopoulou M et al.Campath-1H (anti-CD52) monoclonal antibody therapyin lymphoproliferative disorders. Med Oncol 2001;18:99–107.

Nadeau et al.

302

3. Rodon P, Breton P, Courouble G. Treatment of pure redcell aplasia and autoimmune haemolytic anaemia in chroniclymphocytic leukaemia with Campath-1H. Eur J Haematol2003;70:319–321.

4. Willis F, Marsh J, Bevan D et al. The effect of treatmentwith Campath-1H in patients with autoimmune cytopenias.Br J Haematol 2001;114:891–898.

5. Cheson B. Hematologic malignancies: new developmentsand future treatments. Semin Oncol 2002;29:33–45.

6. Maio M, Altomonte M, Tatake R, Zeff RA, Ferrone S.Reduction in susceptibility to natural killer cell-mediatedlysis of human FO-1 melanoma cells after induction ofHLA class I antigen expression by transfection with B2mgene. J Clin Invest 1991;88:282–289.

7. Raulet D. Does a low level of expression of HLA mole-cules engender autoimmunity? NEJM 1999;340:314–315.

8. Liao N, Bix M, Z M, Jaenisch R, Raulet D. MHC class Ideficiency: susceptibility to natural killer (NK) cells andimpaired NK activity. Science 1991;253:199–202.

9. Handgretinger R, Geiselhart A, Moris A, Grau R,Teuffel O, Bethge W, Kanz L, Fisch P. Pure red-cellaplasia associated with clonal expansion of granularlymphocytes expressing killer-cell inhibitory receptors.NEJM 1999;340:278–284.

10. Hale G. The CD52 antigen and development of theCAMPATH antibodies. Cytotherapy 2001;3:137–143.

11. Schmidt H, Gekeler V, Haas H, Enler-Blum G, SteiertI,ProbstH,MullerC.Differnetial regulation ofHLA classI genes by interferon. Immunogenetics 1990;31:245–252.

12. Davidovitz Y, Halpern Z, Vardi J, Ballin A, Meytes D.

Pure red cell aplasia responsive to interferon-alpha in apatient with hepatitis C virus infection. Acta Haematol1998;100:213–215.

13. Krantz S. Pure red cell aplasia: biology and treatment. In:Feig S, Freedman MH, eds. Clinical Disorders andExperimental Models of Erythropoietic Failure. BocaRaton, FL: CRC Press, 1993:85–124.

14. Dessypris E, Krantz S, Roloff J et al. Mode of action ofthe IgG inhibitor of erythropoiesis in transient erythro-blastopenia of childhood. Blood 1982;59:114–123.

15. Peschle C, Marmont A, Marone G et al. Pure red cellaplasia: studies on an IgG serum inhibitor neutralizingerythropoietin. Br J Haematol 1975;30:411–417.

16. Schmid J, Kiely J, Harrison E et al. Thymoma associatedwith red cell agenesis: review of literature and report ofcases. Cancer 1965;18:216–230.

17. Hoffman R, Kopel S, Hsu S et al. T cell chronic lymph-ocytic leukemia: presence in bone marrow and peripheralblood of cells that suppress erythropoiesis in vitro. Blood1978;52:255–260.

18. Abkowitz J,Kadin M, Powell J, Adamson J. Pure red cellaplasia: lymphocyte inhibition of erythropoiesis. Br JHaematol 1986;63:59–67.

19. Lacy M, Kurtin P, Tefferi A. Pure red cell aplasia:association with large granular lymphocyte leukemia andthe prognostic value of cytogenetic abnormalities. Blood1996;87:3000–3006.

20. Teffari A, Li C, Witzig T, Dhodapkar M, Okuno S,Phyliky R. Chronic natural killer cell lymphocytosis: adescriptive clinical study. Blood 1994;84:2721–2725.

21. Erslev A, Soltan A. Pure red-cell aplasia: a review. BloodRev 1996;10:20–28.

22. Firkin T, Maker D. Cytotoxic immunosuppressive drugtreatment strategy in pure red cell aplasia. Eur J Haematol1988;41:212–217.

23. Raghavachar A. Pure red cell aplasia: Review of treatmentand proposal for a treatment strategy. Blut 1990;61:47–51.

24. Gascon P, Zoumbos NC, Scala G, Djeu JY, Moore JG,Young NS. Lymphokine abnormalities in aplastic anemia:implications for the mechanism of action of antithymocyteglobulin. Blood 1985;65:407–413.

25. Metzger R, Heckl-Ostreicher B, Nerl C, Schond-

elmaier S, Kabelitz D. Immune studies of cd + T cells ina case of large granular lymphocyte (LGL) leukemia:leukemic cd + T cells are resistant to growth stimulationin vitro but respond to interferon-a treatment in vivo. LeukRes 1992;16:1087–1095.

26. Clark D, Dessypris E, Krantz S. Studies on pure red cellaplasia. XI. Results of immunosuppressive treatment of 37patients. Blood 1984;63:227–286.

Red cells

303