Treatment of aluimmunization and refractoriness to platelet transfusion by protein A column therapy

Abstract

The invention provides a therapeutic method for treating refractoriness to platelet transfusion by isolating blood serum from an alloimmunized patient undergoing platelet transfusion therapy, passing the serum through a bed comprising staphococcal protein A coupled to a solid support and returning the treated plasma to the patient. The method can be conducted by batch-type procedure or by continuously conducting the steps so that blood is withdrawn, passed through the bed, and returned to the patient as a continuous stream. The method is useful to treat alloimmunized patients suffering from leukemia, aplastic anemia, myelofibrosis, myelodysplastic syndrome, or in a bone marrow transplant patient.

Description

Refractoriness to platelet transfusion (RPT) is a serious complication of leukemia therapy, bone marrow transplantation, and other disorders where multiple infusions of platelets are required to prevent bleeding. Alloimmunization to HLA class I antigens is an important cause of RPT, with some estimates as high as 50-70% of patients receiving multiple platelet transfusions becoming alloimmunized to HLA antigens. (P. Daly et al., JAMA, 253:435 (1980); D. Hogge et al., Am. J. Hematol., 14:363 (1983); R. Kakaiya et al., Transfusion, 24:35 (1984); J. Eernisse et al., Exp. Hematol., 9:77 (1981); M. Murphy et al., Brit. J. Haematol., 60:409 (1985); and M. Murphy et al., Brit. J. Haematol., 62:529 (1986).) Additionally, antibodies directed against platelet-specific antigens, such as HPA-1a(=P1.sup.A1)*, HPA-1b(=P1.sup.A2) and HPA-3a(=Bak.sup.a), HPA-2b(=Sib.sup.a), Nak.sup.a, and Gov.sup.a /Gov.sup.b have also been implicated in platelet transfusion failure. (R. Kickler et al., Transfusion, 30:622 (1990); F. Langenscheidt et al., Transfusion, 28:597 (1988); H. Saji et al., Vox Sang, 56:283 (1989), H. Ikeda et al., Vox Sang, 57:213 (1989); and J. Kelton et al., Blood, 75:2172 (1990).) Moreover, drug-dependent antibodies induced by vancomycin were recently found in association with RPT in two patients with leukemia. (D. Christie et al., Blood, 75:518 (1990).) Thus, it is apparent that a wide variety of different antibodies may contribute to shortened survival of transfused platelets.

Standard transfusion strategies generally recommend that patients first receive pooled random donor platelets, followed by random single donor and HLA-matched single donor platelets once alloimmunization occurs. (J. Herman et al., Am. J. Ped. Hematol./Oncol., 9:272 (1987)). Although steroids, splenectomy, immunosuppressive therapy, IV IgG, and plasmapheresis have often proved successful in treating autoimmune thrombocytopenia (ATP), they have for the most part been ineffective in reducing RPT due to alloimmunization. (J. Herman et al., Am. J. Ped. Hematol./Oncol., 9:272 (1987); D. Hogge et al., Blood, 64:253 (1984); C. Schiffer et al., Blood, 64:937 (1984); and W. Bensinger et al., Transplantation, 41:602 (1986).) Several reports have shown marked increases in platelet counts in patients with ATP following reinfusion of autologous plasma that had been absorbed by immobilized protein A. (T. Guthrie et al, Sem. Hematol., 26(Suppl 1):3 (1989); K. Muroi et al., Sem. Hematol., 26(Suppl 1):10 (1989); A. Mittleman et al., Sem. Hematol., 26(Suppl 1):15 (1989); and H. Snyder, Jr. et al., Artif. Organ, 13:71 (1989).) In the present study, we investigated the effectiveness of protein A column therapy for improving platelet counts and transfusion refractoriness in 12 thrombocytopenic patients with histories of being unresponsive to platelet transfusion therapy.

SUMMARY OF THE INVENTION

A therapeutic method is provided for treating refractoriness to platelet transfusion (RTP) in alloimmunized patients undergoing platelet transfusion therapy which permits the patients to sustain higher average daily platelet counts (2-10X higher than pretreatment) and to exhibit improved responsiveness to both random and HLA-matched platelet transfusion. The present method also decreases circulating platelet antibodies, which are responsible for alloimmunization.

Generally, the present method comprises:

(a) isolating a portion of blood serum, i.e., about 500-2500 ml of plasma, from an alloimmunized patient undergoing platelet transfusion therapy;

(b) passing the portion of serum through a bed comprising Staphylococcal protein A immobilized on a solid support (i.e., coupled to silica gel or to another particulate inorganic oxide) so that immunoglobulin (IgG) and, preferably, IgG-associated immune complexes are removed from the serum, to yield a portion of treated plasma; and

(c) returning said portion of treated plasma to the patient.

RTP is commonly exhibited by alloimmunized patients suffering from leukemia, aplastic anemia, myelofibrosis, myelodysplasia or bone marrow transplant patients. Typically, the bed of immobilized protein A is packed into a chromatography column, and about 500-2500 ml of plasma per treatment, is passed through the column. The 500-2500 ml portion, or other portion of blood plasma is typically obtained by withdrawal of serial units of blood from the patient undergoing transfusion therapy, centrifuging the units to separate the blood into the plasma and the cellular portion, pooling the plasma to yield said plasma portion, and returning the cellular component to the patient. The method can be operated in a "batch" fashion, i.e., in discrete steps, or steps (a) to (c) above can be performed continuously. For example, a stream of blood is removed from the patient, centrifuged to separate a plasma stream from the cellular component, the plasma stream can be passed through a column packed with SiO.sub.2 -immobilized protein A, and the resultant treated plasma stream can be recombined with the cellular component and reinfused into the patient.

BRIEF DESCRIPTION OF THE FIGURES

FIGS. 1A and 1B: Average daily best platelet counts (circles) and posttransfusion CCI from random donor platelets (open squares) or HLA matched platelets (closed squares) measured at 10-120 min (or occasionally >2 h and <24 h) before, during, and after protein A column treatments (arrows) in patients No. 4 and 7. CCI <7500 (horizontal line) were below levels acceptable for successful transfusions. All CCI measured at >2 h were <4500. Each CCI value represents the best response to a single platelet transfusion for that day, although other transfusions may have been given on the same day. Patient No. 4 received a five-day course of cyclosporin A (CSA) beginning with the fourth protein A column treatment.

FIG. 2: Antibody titers for Patient No. 5 measured with serum collected prior (closed symbols) to and following (open symbols) 12 protein A column treatments. Antibodies were detected using the MAIPA assay as described in Methods. Antibodies are against HPA-1a (circles) and HPA-3a (squares).

DETAILED DESCRIPTION OF THE INVENTION

Twelve thrombocytopenic patients (platelets <10-24.times.10.sup.9 /L) were investigated for their responsiveness to staphylococcal protein A column therapy. All patients had bone marrow failure with diagnoses including leukemia, aplastic anemia, myelofibrosis, and myelodysplastic syndrome. Ten patients had previously been treated with steroids, intravenous gammaglobulin (IV IgG), and/or other forms of immunosuppressive therapy without response. All patients were receiving multiple platelet transfusions without achieving one hour corrected count increments (CCI) .gtoreq.7500. Ten patients had platelet antibodies directed against HLA class I antigens, ABO antigens, and/or platelet-specific antigens as detected by immunofluorescence, monoclonal antibody antigen-capture ELISA, and/or lymphocytotoxicity. The other two patients had no detectable platelet antibodies by these methods.

Plasma (500-2000 mL) was passed over a protein A silica gel column (Prosorba.RTM., IMRE Corp., Seattle, Wash.) then reinfused into the patient. Patients typically received six treatments (range=1-14). A positive response to protein A therapy was defined as at least a doubling of the pretreatment platelet count and/or two successive 10 to 120-minute posttransfusion CCI.gtoreq.7500. Following plasma treatments, 7/12 patients responded with daily platelet counts that averaged 62.+-.21.times.10.sup.9 /L (range=46 to >100.times.10.sup.9 /L) as compared with 14.+-.8.times.10.sup.9 /L (range=<10 to 24.times.10.sup.9 /L) before treatment. Posttransfusion CCI values determined among four of these patients averaged 1900.+-.700 and 11,800.+-.3900 before and after treatment, respectively. In contrast, among the five nonresponders pre- and post-treatment platelet counts averaged 9.+-.8.times.10.sup.9 /L (range=<10 to 23.times.10.sup.9 /L) and 11.+-.9.times.10.sup.9 /L (range=<10 to 25.times.10.sup.9 /L), respectively, while posttransfusion CCI were 600.+-.1300 and 1100.+-.2200, respectively.

Following protein A column treatments, seven of these patients achieved transient (1-2 week) to long term (>6 month) increases in platelet counts that in six patients (Patient No. 5 no longer being transfused) were accompanied by improved responsiveness to platelet transfusions. For most of these patients, the improvement in post-transfusion platelet increments was accompanied by decreased platelet usage, although two non-responders also utilized fewer platelets following treatment. All six responders with detectable platelet antibodies had measurable reductions in the titers and/or specificities of these antibodies after protein A therapy while two non-responders demonstrated similar declines in these antibodies.

Other factors investigated included the effect of protein A treatments on platelet usage and platelet antibody levels. Protein A column treatments also were associated with decreased platelet usage of 15-65% in 5/7 responders and 1/5 nonresponders. For most of these patients the improvement in posttransfusion platelet increments was accompanied by decreased platelet usage, although two nonresponders also utilized fewer platelets following treatment. Of the six responders with detectable platelet antibodies, all had a reduction of 40 to >95% in the level of these antibodies following protein A treatment. Only one nonresponder had a similar reduction in platelet antibody level. These findings suggest that in certain patients with refractoriness to platelet transfusion due to platelet alloantibodies, protein A column treatments may be an effective means of increasing platelet counts and posttransfusion CCI, reducing platelet usage, and reducing levels of circulating platelet antibodies.

In two previous studies, an alloimmunized patient with aplastic anemia and one with acute myelogenous leukemia, demonstrated improved response to platelet transfusion following protein A column treatments, although three other alloimmunized patients with RPT failed to respond to platelet transfusion following similar therapy. (K. Muroi et al., Sem. Hematol, 26(Suppl 1):10 (1989); R. Branda et al., Transfusion, 26:471 (1986); and K. Muroi et al., Sem. Hematol, 26(Suppl 1):10 (1989).) Other than these reports, we are aware of no other studies that have investigated the potential medical benefit of protein A column therapy in the treatment of alloimmunized patients with RPT.

In this study, patients received from 1-14 protein A column treatments. This study does not address the optimum number of treatments that will result in a clinically meaningful benefit to the patient. However, it is clear from our results that as few as one or two columns was associated with markedly increased platelet counts and posttransfusion CCI (Table 2), which in some patients (FIG. 1) may have involved increased endogenous platelet production and/or increased survival of transfused platelets. Moreover, 25-50% of 96 patients with ATP who presented with average platelet counts <30.times.10.sup.9 /L responded after only one or two Prosorba.RTM. treatments with increased platelet counts averaging >130.times.10.sup.9 /L (personal communication, Dr. Harry Snyder, Seattle, Wash., February 1991).

Due to the use of concurrent alternate forms of immunosuppressive therapy in five of the responders (Table 2), we cannot conclude with certainty that protein A column therapy alone is effective in the treatment of alloimmunized patients with RPT. However, none had previously responded to these therapies with increased platelet counts, increased posttransfusion CCIs, or decreased transfusion requirements. Moreover, only one responder (No. 4, Tables 1 and 2) received additional immunosuppressive agents during protein A treatment that had not been administered before hand. This patient was given ATG 25 days before protein A therapy followed by cyclosporin between the 4th and 6th protein A column treatments that appeared to correlate with an almost immediate (24-48 h) improvement in posttransfusion CCI values (Table 2). However, it seems unlikely that this was a direct effect of either the ATG or the cyclosporin because the response time for these agents in aplastic anemia is 2-9 months. (W. Miller et al., Brit. J. Haematol., 55:17 (1983); M. Litzow et al., Am. J. Hematol., 32:226 (1989); and M. Lazzarino et al., Bone Marrow Transplant, 4(Suppl 4):165 (1989).) The present study strongly suggests that protein A therapy played a role in the improved platelet transfusion outcome in these patients.

The mechanism of action of protein A column therapy is poorly understood. However, it does not appear to be related solely to the removal from plasma of circulating platelet antibodies because removal of <5% of circulating IgG by protein A treatments is associated with remission in patients with ATP and malignancies. (F. Jones et al., Plasma Ther. Transfus. Technol., 7:333 (1987) and G. Messerschmidt et al., J. Clin. Oncol., 6:203 (1988).) Patients No. 2 and 10 lacked detectable platelet antibodies and yet failed in response to platelet transfusion despite lack of evidence for obvious nonimmunologic factors, such as disseminated intravascular coagulation, splenomegaly, or sepsis, to account for platelet consumption. Others have similarly reported transfusion failure in apparently antibody-negative patients with RPT, and conversely, occasionally observed patients responding to platelet transfusion despite the presence of circulating platelet-reactive antibodies. (D. Hogge, Am. J. Hematol., 14:363 (1983) and D. Hogge et al., Blood, 64:253 (1984).) Nevertheless, after protein A treatment Patient No. 2 achieved increased platelet counts, while Patient No. 10 did not (Table 2). Recent observations in 3/13 apparently antibody-negative patients with ATP who responded to protein A therapy support our findings (personal communication, Dr. Harry Snyder, Seattle, Wash., May 1991). However, all patients with detectable platelet antibodies who achieved increased platelet counts and responsiveness to transfusion after protein A treatments had decreased titers in platelet alloantibodies, although two patients (Nos. 8 and 12) who had a similar reduction in platelet antibodies failed to respond to transfusion. Thus, the mechanism of protein A action may not directly involve reduction in the level of circulating platelet antibodies.

Recently, HIV infected patients with ATP, who were treated with protein A therapy, were found to possess anti-F(ab').sub.2 antibodies directed against platelet autoantibodies. (H. Snyder Jr. et al., Artif. Organ, 13:71 (1989).) The presence of anti-F(ab').sub.2 antibodies in these patients was associated with reduced levels of circulating platelet antibodies and platelet-associated immunoglobulins and increased endogenous platelet production. It is of interest to note that the presence of circulating antiidiotypic antibodies to HLA antigens has been associated with graft survival among renal transplant patients, while the presence of antibodies potentiating HLA antibodies correlated with graft rejection. (E. Reed et al., N. Enq. J. Med., 316:1450 (1987).) One speculation then may be that in patients who have been sensitized to HLA and platelet-specific antigens and who are refractory to platelet transfusion, protein A column therapy promotes production of antiidiotypic antibodies to these platelet-reactive antibodies, thus neutralizing their anti-platelet activity.

Ideally, prevention of alloimmunization should greatly reduce the incidence of RPT in multiply transfused recipients. In fact several reports indicate that reduction in alloimmunization and RPT may be accomplished by transfusing red blood cells and platelets that have been depleted of contaminating leukocytes. (J. Eernisse et al., Exp. Hematol., 9:77 (1981); A. Brand et al., Vox Sang, 54:160 (1988); G. Andreu et al., Blood, 72:964 (1988); I. Sniecinski et al., Blood, 71:1402 (1988); U. Saarinen etc al., Blood, 75:512 (1990); and M. van Marwijk Kooy et al., Blood, 77:201 (1991). Other studies indicate that UV-B irradiation of lymphocytes present in transfused blood components may also be an effective method for reducing alloimmunization. (R. Kahn et al., Transfusion, 25:547 (1985); H. Deeg, Transfusion, 29:450 (1989); and G. Andreu et al., Transfusion, 30:401 (1990). However, as suggested by a recent editorial, these procedures are unlikely to reduce or eliminate the problem of alloimmunization in transfusion recipients previously sensitized to platelet antigens. (C. Schiffer, Blood, 77:1 (1991).) Thus, the need for treatment of alloimmunized patients will be an ongoing requirement.

In summary, use of IV IgG, splenectomy, immunosuppressive agents, and apheresis, often found to be effective in the treatment of refractory ATP, have generally failed or had limited utility in the reduction of RPT due to alloimmunization. (J. Herman et al., Am. J. Ped. Hematol./Oncol., 9:272 (1987); D. Hogge et al., Blood, 64:253 (1984); C. Schiffer et al., Blood, 64:937 (1984); and W. Bensinger et al., Transplantation, 41:602 (1986).) Clearly then, an effective means to manage- the alloimmunized patient who is unresponsive to platelet transfusion is needed. Our findings strongly suggest that some alloimmunized patients with RPT may benefit from treatment of their plasma with immobilized protein A. Specifically, our study indicates that in certain patients protein A column therapy may be safe and effective for increasing platelet counts and posttransfusion increments, reducing platelet usage, and reducing levels of circulating platelet-reactive alloantibodies.

EXAMPLE

Patients

Patients were recruited for this study by the following criteria: 1) platelet count <30.times.10.sup.9 /L;2) refractory to platelet transfusion as defined by failure to achieve a 10-60 min posttransfusion corrected count increment (CCI) .gtoreq.7500 following two successive transfusions; 3) diagnosis consistent with bone marrow failure; and 4) evidence of alloimmunization to platelet antigens. (T. Kickler et al., Transfusion, 25:385 (1985) and B. O'Connell et al., Transfusion, 28:66 (1988).) By this criteria, 12 patients were enrolled in this study between November 1989 and September 1991. All patients gave informed consent in accordance with guidelines established by the University of Minnesota Committee on Use of Human Subjects in Research. Specific diagnoses and other relevant patient information are listed in Table 1.

TABLE 1__________________________________________________________________________PATIENT INFORMATION PlateletPatient Clinical Count Prior.sctn. AntibodyNo. Age Sex Diagnosis* Outcome (.times. 10.sup.9 /L) Treatment Specificity__________________________________________________________________________1 19 M AA discharged 17 ATG/CT .alpha.-HLA2 49 F CML expired (3 wk) 24 CT (22) Not detected**3 20 M MDS/BMT expired (3 days) <10 CT/ATG/ALG/IgG .alpha.-HLA4 70 F AA discharged 23 ATG/CT (12) .alpha.-A(IgG) .alpha.-HLA(IgG)5 61 M CMML discharged 13 None .alpha.-HLA(IgG) .alpha.-HPA-1a(IgG) .alpha.-HPA-3a(IgG)6 12 F AA/BMT discharged <10 CT/IgG (12) .alpha.-HLA(IgG + IgM) .alpha.-GPIb/IX(IgM) .alpha.-GPIIb/IIIa(IgM)7 48 F MDS expired (8 days) 11 CT/IgG (18) .alpha.-HLA(IgG)8 35 M AML/BMT discharged <10 CT/IgG .alpha.-HLA(IgG + IgM)9 54 M MF expired (1 day) <10 CT/IgG .alpha.-HLA10 20 M ALL expired (11 mo) 23 CT (31) Not detected11 68 M MF expired (4 wk) <10 CT/IR (9) .alpha.-HLA(IgG)12 36 F AML expired (7 days) <10 None .alpha.-HLA(IgG)__________________________________________________________________________ *Abbreviations are: AA, aplastic anemia; ALL, acute lymphocytic leukemia; CML, chronic myelogenous leukemia; CMML, chronic myelomonocytic leukemia; MDS, myelodysplastic syndrome; and MF, myelofibrosis. Three patients (Nos 3, 6 and 8) were bone marrow transplant recipients (BMT). Seven patients expired (EXP) from complications of their primary diagnoses at various times following protein A column treatment, which di not contribute to any of the causes of death. Average best daily posttransfusion (10-120 min) platelet count from the onset of RPT during the last admission and before starting protein A column therapy, which ranged from 4 to 23 days. .sctn. For thrombocytopenia and RPT during last admission: ATG, antithymocyte globulin; CT, chemotherapy (steroids, vinca alkaloids, cyclophosphamide, azathioprine); IgG, intravenous gammaglobulin; IR, irradiation of spleen. The number in parentheses indicates the days patient was unresponsive to therapy from the time of the last admission. In some case this information was unavailable. Plateletreactive antibody studies were performed with sera collected immediately prior to the beginning of protein A column therapy. Antibodie reacted with antigens of the ABO system (A), human leukocyte antigen system (HLA), human platelet antigen system [HPA:HPA1a(.dbd.P1.sup.A1) an HPA3a(.dbd.Bak.sup.a)], and unidentified epitopes of various platelet glycoproteins (GP). In most but not all cases, immunoglobulin class was determined and noted in parentheses. **No antibodies were detected by PIFT, MAIPA, or lymphocytotoxicity.

Platelet Transfusions

Patients received platelet transfusions in accordance with general practices and guidelines approved by the University of Minnesota Hospital Transfusion Therapeutics Committee. (J. McCullough et al., JAMA, 259:2414 (1988).) Where possible, successful transfusions were judged by the criteria described above, although the posttransfusion time was occasionally extended to 120 min. or longer but not more than 24 hr.

Protein A column therapy

The PROSORBA.RTM. column is marketing approved by the United States Food and Drug Administration (FDA) for use in the therapeutic removal of immunoglobulin G (IgG) and IgG-containing circulating immune complexes (CIC) from plasma of patients with ITP having platelet counts less than 100,000/mm.sup.3. ITP is an autoimmune disease that afflicts about 25,000 Americans with approximately 2,500 new cases annually. ITP occurs in the general population but also is present in approximately 10% of HIV antibody-positive individuals.

Protein A is a major component of certain strains of Staphylococcus aureus (SAC). Protein A has a propensity to bind IgG via the Fc portion and has even greater affinity for IgG bound to an antigen (i.e. CIC). When protein A is a part of a solid matrix, either as a component of the SAC or when it is isolated from SAC and bound to an inert substance, it can be used as an effective reagent for the removal of IgG or CIC from plasma. The use of protein A as an immunological reagent has been studied in great detail, due to its applicability to the specific isolation of immunoglobulins and/or immunoglobulin bound antigens.

The PROSORBA column employs 200 mg of protein A which has been covalently bound to an inert silica matrix. The protein A-silica is contained within a biocompatible, polycarbonate assembly which is designed for the sterile passage of plasma. Plasma can thus be exposed to the protein A with subsequent adsorption of IgG and CIC. Treated plasma which has passed through the device can then be returned to the patient.

This medical treatment device was approved for market distribution by the FDA on Dec. 23, 1987 as follows: The PROSORBA.RTM. column is indicated for use in the therapeutic removal of immunoglobulin G (IgG) and IgG-containing circulating immune complexes (CIC) from plasma in patients with idiopathic thrombocytopenic purpura having platelet numbers less than 100,000/mm.sup.3.

A. Column Preparation

1. Standard aseptic technique must be used for all PROSORBA.RTM. column tubing connections.

2. Fluid flowing through the PROSORBA.RTM. column must always be directed upward (i.e., fluid must enter the PROSORBA.RTM. column at the base inlet and exit at the top outlet).

3. Prime the PROSORBA.RTM. column with a wash of 4.0 liters of 0.9% sterile saline (IV quality). To remove trapped air and to ensure maximum wash efficiency, use the following steps: (a) fill the column with saline; (b) interrupt the flow; and (c) rotate the column slowly and gently to completely wet the matrix prior to passing the rest of the sterile saline through the column. Do not allow air to enter the column.

4. When the wash is complete, anticoagulate the column by passing an additional 1/2 liter of 0.9% sterile sale (IV quality) containing 5,000 units of sodium heparin through the column at 20 ml/min.

5. Do not disturb the PROSORBA.RTM. column during the final 3.5 liters of sterile saline prime or during the anticoagulation procedure.

B. On-Line Preparation

1. Prepare the apheresis apparatus as indicated in the apparatus manufacturer's instructions, including establishment of patient blood access and anticoagulation.

2. Attach the washed, anticoagulated PROSORBA.RTM. column (see Column Preparation--VIII.A. above) in an upright position to the plasma return line of the apheresis apparatus. Plasma flow must be directed upward through the column.

3. Initiate apheresis apparatus operation discarding the first 250 ml of fluid (0.9% saline+5,000 units of sodium heparin) from the column to the waste container.

4. Reestablish plasma flow back into the patient and perfuse the prescribed plasma volume (up to 2,000 ml) through the PROSORBA.RTM. column at 10-20 ml/min.

5. Terminate the plasma perfusion with the following steps: (a) displace the plasma remaining in the column with 250 ml of sterile IV quality 0.9% saline; (b) infuse the displaced plasma into the patient at 10-20 ml./min.; and (c) disconnect the patient from the apheresis apparatus as indicated in the manufacturer's instructions.

6. Vital signs should be monitored at the initiation of the PROSORBA.RTM. column treatment and at least every 30 minutes during the procedure. The patient also should be monitored for a period of time after the treatment has been terminated.

Plasma from each patient was passed over a staphylococcal protein A silica gel column (the PROSORBA.RTM. immunoadsorption treatment column, IMRE Corp., Seattle, Wash.) by either an on-line or off-line technique as directed by the manufacturer. For the on-line procedure, whenever possible, a minimum of 1500 mL of plasma were treated per protein A column utilizing CS 3000 apheresis equipment (Baxter, Deerfield, Ill.) to collect the plasma. For the off-line procedure, a minimum of 500 mL of plasma were treated per protein A column, where plasma was obtained by centrifugation of two units of whole blood. Following protein A treatment, the plasma was returned to the patient who received additional column treatments at intervals of 2-3 days. The other blood components obtained from each procedure were reinfused prior to return of the treated plasma. Standard protocols for protein A column therapy utilize six column treatments. The actual number of treatments received by each patient are shown in Table 2.

TABLE 2__________________________________________________________________________EFFECT OF PROTEIN A COLUMN THERAPY ONPLATELET COUNTS AND POSTTRANSFUSION INCREMENTSPatient Prosorba .RTM. ConcurrentNo. Therapy* Therapy Platelet Pre Counts Post Posttransfusion Pre CCI.sctn. Post__________________________________________________________________________Responders1 6 (off) None 17 46 1610 .+-. 0 >13,0002 1 (off) CT 24 .+-. 9 46 .+-. 18 3000 .+-. 2400 6000 .+-. 25003 14 (off) CT/CSA/IgG <10 100 .+-. 20 CI = 1 u/h CI = 1 u/h4 7 (on) CSA/CT 23 .+-. 12 53 .+-. 21 1800 .+-. 3400 14,300 .+-. 33005 12 (on) None 13 .+-. 8 100 .+-. 10 <1000 NT**6 6 (off) CT/IgG <10 51 .+-. 27 CI = 2 u/h CI < 0.5 u/h7 2 (on) CT 11 .+-. 9 51 .+-. 26 2000 .+-. 2900 14,000 .+-. 8000 x = 14 .+-. 8 x = 64 .+-. 25 x = 1900 .+-. 700 x = 11,800 .+-. 3900Nonresponders8 1 (off)/5 (on) CT <10 15 0 386 .+-. 4609 1 (off) CT/IgG <10 <10 0 010 2 (on) None 23 .+-. 8 25 .+-. 9 2800 .+-. 2400 5100 .+-. 360011 6 (on) CT/IgG <10 <10 0 012 5 (off) None <10 <10 0 0 x = 9 .+-. 8 x = 11 .+-. 9 x = 600 .+-. 1300 x = 1100 .+-. 2200__________________________________________________________________________ *Form of treatment was either online or offline as indicated in parentheses. Symbols are the same as listed in Table 1; CT primarily consisted of steroids; CSA, cyclosporine A. Average best daily posttransfusion platelet count (mean .+-.sd .times. 10.sup.9 /L). Pre: as described in Table 1. Post: after protein A column therapy was begun and up to 4 wk following treatment depending on whether the patient was discharged or expired. In some cases, patients were followed 1-12 mo after discharge. .sctn. Corrected count increment (CCI) based on best average daily platelet count determined at 10-120 min or up to 24 h posttransfusion if earlier measurements were unavailable. Values are means .+-.sd. Patients No. 3 and 6 were receiving nearly continuous platelet infusion (CI) so that posttransfusion CCI values could not be determined and instead the rate of infusion is shown. **No transfusions. Patient had 10-120 min posttransfusion CCI of 5900 and 3300 after the third and fifth columns, respectively.

Effect of Protein A Column Therapy on Platelet Counts and Post Transfusion CCl

The effectiveness of protein A therapy was determined by comparison of pre- and post-treatment platelet counts and response to platelet transfusions using the following criteria:

1) patients with platelets <10.times.10.sup.9 /L had to increase to <20.times.10.sup.9 /L following therapy or patients with platelets between 20-30.times.10.sup.9 /L had to increase by at least a factor of two; and

2) following treatment, patients had to achieve at least two successive one-hour posttransfusion CCI .gtoreq.7500.

(H. Snyder Jr. et al., Artif. Organ, 13:71 (1989).) Patients fulfilling either of these criteria were considered to have a positive response to protein A therapy.

Changes in platelet counts and posttransfusion CCI before, during, and after protein A column therapy are shown for Patients No. 4 and 7 in FIG. 1. Each patient responded to only a single HLA matched platelet component and were otherwise refractory to all other platelet transfusions (10-120 min posttransfusion CCI <7500) for 21 days. Moreover, both patients were unresponsive to previously administered steroids, Iv IgG, and/or ATG. However, each patient sustained higher average daily platelet counts of .about.50.times.10.sup.9 /L during and after protein A treatments that were accompanied by excellent responsiveness to both random and HLA matched platelet transfusions with 10-120 min posttransfusion CCI averaging 14,000.

Average daily platelet counts and posttransfusion CCI are summarized for all patients in Table 2. Seven of 12 patients maintained average daily platelet counts two- to ten-fold higher following protein A treatments. These higher platelet counts correlated well with improved posttransfusion CCI for Patients No. 1, 4, and 7 who had average pre-column CCI.ltoreq.2000 and post-column CCI.gtoreq.13,000. Patient No. 2, although not achieving clinically significant CCI for 10-120 min (.gtoreq.7500), nonetheless experienced a two-fold increase in CCI from 3000 to 6000 following protein A column therapy. Patients No. 3 and 6 had been receiving continuous infusions of platelets before treatment. In Patient No. 6, this declined from 2 units/h to <0.5 unit/h following six column treatments, while in Patient No. 3 there was no change following 14 treatments. The former patient was discharged with a platelet count of .about.40.times.10.sup.9 /L and at six weeks following her last column treatment was transfusion independent with a platelet count of 116.times.10.sup.9 /L, while the latter patient expired but had a platelet count of >100.times.10.sup.9 /L at the time of his death.

Effect of Protein A Column Therapy on Platelet Usage and Platelet Antibody Levels

Predictably, increases in platelet counts were generally accompanied by fewer transfusions (Table 3). The 12 patients had been receiving between 4 and 32 units of platelets/day before protein A treatments. Of the seven responders, five had decreased platelet usage of 15-65% after treatment, while two had no change in platelet usage. Of the five nonresponders, Patients No. 10 and 11 experienced a 59% and 47% decrease in platelet usage, respectively.

Platelet Antibody Studies

Serum or plasma was collected from each patient prior to their first protein A column treatment and thereafter following each treatment. Additional serum was collected two weeks after the last column treatment on each surviving patient. Serum was tested for the presence of platelet-reactive antibodies by the monoclonal antibody-specific immobilization of platelet antigens (MAIPA) assay with minor modifications. (D. Christie et al., Sem. Thrombos. Hemostas., in press; V. Kiefel et al., Blood, 70:1722 (1987).) In the MAIPA procedure, solubilized platelet antigens are captured in microtiter wells by murine monoclonal antibodies (MoAbs) directed against specific platelet glycoproteins (GPs) or class I antigens for identification of the site to which the patients' antibodies combine. (D. Christie et al., Sem. Thrombos. Hemostas., in press.) Attached human alloantibodies are then detected by a standard ELISA technique and color development monitored at 405 nm with a microplate reader (Biokinetics Reader EL 340, Bio-Tek Instruments, Winooski, Vt.). Relative changes in the levels of circulating antibodies possessed by patients before and after protein A column therapy were determined by comparing the optical densities of serial dilutions of serum samples. Only serum dilutions giving optical densities between 0.5 and 1.5 were used, which is in the linear range of the reader.

Using the MAIPA assay, the relative amount of antibody present in serum collected immediately before the first treatment and that collected after the last treatment could be determined by comparing titers of the paired serum samples. An example of this is shown for Patient No. 5 in FIG. 2. Among the six responders with detectable platelet antibodies, five demonstrated decreases in circulating platelet antibodies of 65 to >95% following protein A column therapy. One other patient (No. 1) showed a decline of 37% in antibody by lymphocytotoxicity assay, but not by the MAIPA assay (Table 3). Of the four nonresponders with detectable platelet antibodies, Patients No. 8 and 12 had decreases of 75-80% and 40%, respectively.

In addition, sera from some of the patients were tested by immunofluorescence (PIFT) for platelet antibodies (performed by the University of Minnesota Platelet Serology Laboratory) and/or by lymphocytotoxicity (performed by the University of Minnesota Immunology Laboratory) for HLA antibodies. In those patients receiving drugs suspected of causing antibody-mediated platelet destruction, serum was tested for the presence of drug-dependent antibodies by PIFT and the protein A rosette assay, which was performed as previously described. (D. Christie et al., Blood, 75:518 (1990) and D. Christie et al., Transfusion, 28:322 (1988).)

Alloimmunization to platelet and/or HLA class I antigens was assessed by PIFT, lymphocytotoxicity, and/or the MAIPA assay. Using these methods, sera collected immediately prior to the first protein A column treatment tested positive for platelet-reactive antibodies in all but two patients (Nos. 2 and 10) (Table 1). Seven patients (Nos. 1, 3, 7-9, 10, and 11) had only HLA antibodies, while three had HLA antibodies that were coincident either with anti-A (No. 4), anti-HPA-1a and -3a (No. 5), or anti-GPIIb/IIIa and -GPIb/IX (No. 6). Although the anti-GPIIb/IIIa antibody in Patient No. 6 reacted differentially with platelets typed for HPA-1a, -1b, -3a, -3b, and -4a that are known to reside on GPIIb/IIIa (31-33), no clear specificity(ies) could be discerned. The anti-GPIb/IX antibody appears to have been an autoantibody because it reacted with the patient's own platelets and all other platelets studied.

During the period of thrombocytopenia and RPT, Patients No. 2-5, 7, 9 and 11 and were receiving various medications, including vancomycin, ranitidine, and/or sulfamethoxazole/trimethoprim that are known to be associated with drug-dependent antibody formation. (D. Christie et al., Blood, 75:518 (1990); D. Christie et al., Transfusion, 28:322 (1988); and P. Ciimo et al., Am. J. Hematol., 2:65 (1977).) By PIFT and/or the protein A rosette assay, no drug-dependent antibodies were detected in the serum of these patients immediately before protein A column therapy.

TABLE 3______________________________________EFFECT OF PROTEIN A COLUMN THERAPYON BLOOD COMPONENT USAGE ANDIMMUNOLOGICAL FACTORSPatient Platelet Usage* AntibodyNo. Pre Post Reduction ** %______________________________________Responders1 .sup..about. 2.6 .sup..about. 2.6 02 5.9 .+-. 2.7 5.0 .+-. 4.2 --3 25-30 25-30 >95 (anti-HLA)4 3.9 .+-. 3.0 1.4 .+-. 2.8 65 (anti-HLA) 85 (anti-A)5 1.7 .+-. 3.9 NT.sup..sctn. 85 (anti-HPA-1a) 90 (anti-HPA-3a) 75 (anti-HLA)6 31.6 .+-. 7.0 10.8 .+-. 5.7 60 (anti-GPIb/IX) 90 (anti-GPIIb/IIIa) 80 (anti-HLA)7 12.3 .+-. 5.5 10.5 .+-. 3.3 90 (anti-HLA)Nonresponders8 5.5 .+-. 1.1 5 .+-. 0 80 (anti-HLA)9 12.6 .+-. 8.9 14.5 .+-. 13 010 6.4 .+-. 5.4 2.6 .+-. 2.9 --11 14.2 .+-. 3.0 7.5 .+-. 5.1 0 15.0 .+-. 3.212 .sup..about. 5 .sup..about. 5 40 (anti-HLA)______________________________________ **Platelet-specific alloantigens were designated according to the nomenclature recently adopted by the Expert Panel on Serology of the International Society of Blood Transfusion and the International Committe for Standardization in Haematology (Transfusion, 30:477, 1990). *Units/day (mean .+-.sd); platelets were either pooled random donor units or platelet concentrates (single random donor or HLA matched) collected b apheresis, which were each counted as 4.7 units. Based on comparison of antibody titers performed with serum collected before the first protein A column treatment with that collected immediately after the last treatment. Patient had no detectable change in HLA antibodies as measured in the MAIPA assay; however, by lymphocytotoxicity assay the patient's serum reacted with 25/48 cells before protein A treatment and 16/48 cells after treatment. .sup..sctn. No transfusions. Platelet usage for protein A column treatments #1-5 (top line, 11 days) and beginning the day after the 5th column treatment until the patient expired (bottom line, 18 days). On the day after the 5th column treatment the patient fell and hit his head; the increased platelet usage was a precautionary measure against possible intracranial hemorrhage.

Side Effects Associated with Protein A Column Therapy

Previously reported side effects (e.g. chills, urticaria, dyspnea, and hypotension) associated with protein A column therapy occurred in 4 of the 12 patients (Nos. 3, 5, 10 and 12). (T. Guthrie et al., Sem. Hematol., 26(Suppl 1):3 (1989) and A. Mittleman et al., Sem. Hematol., 26(Suppl 1):15 (1989).) Patient No. 3 developed mild nausea during four of 14 treatments and mild chills/fever during one treatment. Patient No. 5 experienced some hypotension in association with the first column treatment, but no further problems were encountered with five additional treatments. Two months later the patient received an additional six protein A treatments and experienced mild hypotension and dermatologic reactions that did not prevent completion of therapy. Patient No. 10 elected to discontinue therapy after two treatments following the development of a rash. Patient No. 12 experienced one mild episode of dyspnea during her first column treatment, but completed four subsequent treatments with no further complications. None of the side effects observed in these patients were life-threatening.

The references are incorporated by reference herein. The invention has been described with reference to various specific and preferred embodiments and techniques. However, it should be understood that many variations and modifications may be made while remaining within the spirit and scope of the invention.

Claims

1. A therapeutic method for treating refractoriness to platelet transfusion due to alloimmunization, comprising:

(a) isolating a portion of blood serum from an alloimmunized patient undergoing platelet transfusion therapy;

(b) passing the portion of serum through a bed comprising Staphylococcal protein A coupled to a finely divided solid support, so that IgG and IgG-associated immune complexes are removed from the serum, to yield a portion of treated plasma; and

(c) returning said portion of treated plasma to said patient to achieved an increase in average daily platelet counts over pretreatment levels for at least one week posttransfusion.

2. The method of claim 1 wherein said alloimmunized patient suffers from leukemia, aplastic anemia, myelofibrosis, myelodysplasia or is a bone marrow transplant patient.

3. The method of claim 1 wherein the bed is contained within a chromatography column.

4. The method of claim 1 wherein about 500-2500 ml of plasma is passed through the bed and returned to the patient.

5. The method of claim 1 wherein, in step (a), the portion of plasma is isolated by withdrawing serial units of blood from the patient, centrifuging the serial units of blood to separate the cellular portion from the plasma, pooling the plasma, and returning the cellular component to the patient.

6. The method of claim 1 wherein the solid support is silica gel.

7. The method of claim 1 wherein steps (a) to (c) are performed continuously, so that a stream of blood is withdrawn from the patient in step (a), and a stream of treated plasma is returned to the patient in step (c).