https://orcid.org/0000-0002-9073-3173
Abstract
Platelet-transfusion refractoriness (PTR) is common in patients with hematological malignancies. The etiology of immune PTR is typically human leukocyte antigen (HLA) antibodies (Abs) from pregnancy or previous transfusion. Herein, we report PTR in the setting of induction chemotherapy for acute myelogenous leukemia (AML) from Abs against CD36/glycoprotein (GP)IV. A 66-year-old African American woman presented with anemia and thrombocytopenia. She was found to have transfusion-dependent AML, and a 7 + 3 regimen (7 days of standard-dose cytarabine and 3 days of an anthracycline antibiotic or an anthracenedione, most often daunorubicin) was initiated. The patient developed profound thrombocytopenia, with platelet nadir of 0 by day 13. The results of HLA antibody screening were negative. However, the results of a screening test for platelet-specific antibodies screen showed Abs against cluster of differentiation (CD)36. The platelets of the patient lacked expression of CD36, and DNA analysis showed mutations in the CD36 gene. HLA Ab–mediated PTR is common in patients with hematological malignancies. However, once HLA Abs are excluded, other less-frequent Abs should be considered, particularly in patient populations of Asian, African, or Middle Eastern descent.
Platelet-transfusion refractoriness (PTR) occurs in approximately 17% to 45% of patients treated by hematology and oncology physicians.1,2 The causes of platelet refractoriness can be grouped as immune mediated or non–immune mediated. The latter is more common and mostly occurs due to sepsis/infection, fever, splenomegaly, medications, bleeding, and/or disseminated intravascular coagulation (DIC). The remaining cases of platelet refractoriness are attributed to immune causes such as antibodies to class I human leukocyte antigens (HLAs), antibodies to human platelet-specific alloantigens (HPAs), platelet autoantibodies, and drug-dependent platelet antibodies. There is also a small role for ABO incompatibility in immune-platelet refractoriness. A large portion of immune-mediated platelet refractoriness is due to HLA class I antibodies. Notably, only class I (not class II) HLA antigens are expressed on platelets, and of the class I antigens, only HLA-A and -B are well expressed.
Once HLA antibodies have been excluded, other causes of immune-mediated platelet refractoriness should be evaluated, such as antibodies to HPA. HPA antibodies can be directed against a variety of different antigens, and the presence of some is based on different HPA-antigen frequencies in different ethnic groups.3 In some populations of Asian and African descent, a significant number of people lack expression of glycoprotein (GP)IV or cluster of differentiation (CD)36 on their platelets and monocytes; these individuals can produce CD36 antibodies when immunized by pregnancy or blood transfusions.4 Herein, we describe a patient with CD36 deficiency who developed CD36 antibodies and PTR that may have impeded her ability to obtain optimal treatment for acute myelogenous leukemia (AML).
Clinical History
A 66-year-old African American woman with a history of endometrial cancer, cerebral vascular accident without neurological deficit (~10 years ago) and hypertension presented to the emergency department with gum bleeding and fatigue. She was found to have severe anemia and thrombocytopenia. She had had a single pregnancy and no history of blood transfusions. Her white blood cell (WBC) count was 21,000 per μL (reference range [RR], 4.000–10.000), her hematocrit was 11% (RR, 34%–45%), and her platelet count was 71,000 per μL (RR, 160,000–370,000/μL). A peripheral smear showed 70% blasts with monocytic features.
The patient was admitted to the hospital and subsequently received 4 units of red blood cells (RBCs). Later, she was diagnosed as having AML with monosomy 7 and started on 7 + 3 regimen (7 days of standard-dose cytarabine and 3 days of an anthracycline antibiotic or an anthracenedione, most often daunorubicin). By day 7, she had developed profound thrombocytopenia and by day 13, her platelet count was essentially 0. She received numerous transfusions of single-donor apheresis platelets with little to no increase in platelet count. Before transfusion, her platelet count was 7000 per μL, and 1 hour after transfusion of a single-donor apheresis unit of platelets, her count was 6000 per μL The following day, she received another unit of single-donor apheresis platelets and had a post-transfusion platelet count of 5000 per μL several hours after transfusion. PTR was suspected.
The platelets of the patient were studied, and the patient was found to have a unique compilation of CD36 variants. All of the variants found in the patient have been previously described5 in individuals with African ancestry at a frequency of 1% to 8%. This patient was unique in having a total of 3 variants in the coding region of CD36. The first variant in exon 4, present in a heterozygous state, is a missense variant with a frequency of 1.6% in the ethnic African population. The second variant in the same exon and present in a homozygous state caused a frameshift resulting in a premature stop codon at amino acid 24. It has been documented in a previous report5 of PTR due to CD36 antibodies that transfusion of CD36-deficient donor platelets is successful in maintaining the platelet counts of affected patients. Because it was not possible to find and obtain compatible platelets for our patient, the chemotherapy regimen for the patient was therefore changed to a less aggressive one, to reduce her degree of thrombocytopenia. The patient received 14 cycles of decitiabine, which maintained her AML at a 10% to 20% peripheral blast level for longer than 1 year, which allowed her to maintain a better quality of life. However, she later developed pulmonary aspergillosis and died.
Clinical and Laboratory Information
HLA and HPA Antibody Screening
We used the LIFECODES Class I ID Class I HLA antibody detection kit (Immucor Inc) to screen the serum of the patient for HLA antibodies and the PakPlus ELISA (Immucor, Inc) to screen for HPA antibodies, both according to manufacturer instructions. The platelet antibody bead array (PABA) assay was used to confirm platelet-antigen specificity of antibodies and was performed as previously described.6
Analysis of CD36 Expression and Sequencing of the CD36 Gene
CD36 expression on platelets was measured by flow cytometry using the CD36‐specific MoAb MBC131.7, as previously described.4 Genomic DNA was isolated from leukocytes and buccal epithelial cells by using the Gentra Puregene DNA isolation kit from Qiagen. CD36 variants were identified by bidirectional Sanger sequencing CD36 after amplification of genomic DNA by polymerase chain reaction (PCR).
Results
No HLA antibodies were detected in the serum of the patient; however, platelet-reactive antibodies with specificity for CD36 were detected, and CD36 antibody specificity was confirmed by PABA assay (Figure 1). No platelet-reactive antibodies were detected by PABA against antigens on GPIIb/IIIa, GPIa/IIa, GPIb/IX, or HLA class I proteins (Figure 1). Flow-cytometry analysis results demonstrated absence of CD36 expression on the platelets of the patient (Figure 2). Although CD36 expression was not measured on the monocytes of the patient, the lack of CD36 on her platelets, considered together with the presence of CD36 antibodies in her serum, supports that the patient had type 1 CD36 deficiency. Sanger sequencing revealed 3 variants in CD36. The first 2 variants were located in exon 4; a heterozygous missense variant, c.157A>G (p.Asn53Ser) and immediately 3ʹ, a homozygous frameshift variant, c.158delA (Asn53Ilefs*24). Together, these variants cause a stop codon 24 amino acids downstream (c.157_158delAAinsG [p.Asn53Valfs*24]). Both of these variants have been reported independently of each other at a frequency of approximately 1% in the African population (gnomAD).7 The third variant, c.975T>G (p.Y325*), is a well-characterized nonsense variant in exon 10. Approximately 9% of the ethnic African population has this variant in a heterozygous state (gnomAD),7 and African Americans who are homozygous for this variant have been reported to have CD36 deficiency.8
Reactions of patient serum (solid black bars) and normal control serum (white bars) against platelet glycoprotein (GP)IV/cluster of differentiation (CD)36, GPIIb/IIIa, GPIb/IX, GPIa/IIa, and Class I human leukocyte antigen (HLA) antigens, as determined via platelet antibody bead array (PABA) assay. The serum from the patient contained immunoglobulin G (IgG) antibodies that were reactive only with CD36 (solid black bars). The results shown are from original PABA clinical tests performed one at a time. The dashed line represents the fluorescence ratio (FLR) cutoff value (≥4.2) for positive test results. FLR is determined by dividing the anti-IgG fluorescence value obtained with the test specimen by the negative-control-serum fluorescence value.
A, Reactions of patient platelets (solid black bars) with glycoprotein-specific monoclonal antibodies against glycoprotein (GP)IIb/IIIa, GPIV/cluster of differentiation (CD)36, and mouse monoclonal isotype control (negative control). The platelets of the patient showed positive reactivity with monoclonal against GPIIb/IIIa but no detectable reactivity with monoclonal antibody against CD36, which indicated that the patient has deficiency in platelet expression of CD36. The results shown are from original clinical tests performed one at a time. The dashed line represents the fluorescence ratio (FLR) cutoff value (>2.0) for positive test results. FLR is determined by dividing the fluorescence value obtained with the test specimen by the isotype control serum-fluorescence value. B, Reactions of patient (solid black bars) and normal control (white bars) platelets with glycoprotein-specific monoclonal antibodies against GPIIb/IIIa, GPIV/CD36, and mouse monoclonal isotype control (negative control). The platelets of the patient showed positive reactivity with monoclonal against GPIIb/IIIa but no detectable reactivity with monoclonal antibody against CD36, indicating that the patient has deficiency in platelet expression of CD36. The results shown are from original clinical tests performed 1 at a time, and the numbers are FLR values. FLR is determined by dividing the fluorescence value obtained with the GP-specific monoclonal antibodies by that of the isotype control value. The positive GP expression is FLR >2.0 (dashed line).
Discussion
CD36 is a highly glycosylated 88-kDa protein that is expressed on a variety of cells, including endothelial, epithelial, and adipocytes; however, platelets and monocytes/macrophages are the only blood cells that express CD36.4,9 Also, CD36 is a member of the class-B scavenger-receptor family and has been shown10,11 to bind collagen, thrombospondin, long-chain fatty acids, β-amyloid, and oxidized low-density lipoprotein. Additionally, CD36 is an important receptor for uptake of pathogens and apoptotic cells, including Plasmodium falciparum–infected erythrocytes. We find it interesting that CD36 deficiency has been shown4,12–14 to be associated with decreased susceptibility to severe malaria infection. Also, we note that CD36 expression on monocytes is protective against malaria because it facilitates binding and engulfment of infected RBCs. However, CD36 expression on platelets is not protective against malaria because it facilitates platelet CD36 binding to infected RBCs and increased cytoadherence. Thus, individuals with decreased CD36 expression on platelets but retained expression on monocytes have increased protection against severe malaria infection.15 Thus, the ratio of CD3 expression on the platelets of an individual relative to that ratio on their monocytes determines the protective effect against malaria. There are 2 types of CD36 deficiency, namely, type I and type II. Individuals with type I deficiency lack CD36 on platelets and monocytes, and can be immunized against CD36 from pregnancy or blood transfusions. Patients with type II CD36 deficiency lack expression of CD36 only on their platelets.16 CD36 deficiency is prevalent at much higher frequencies in ethnic Asians (3%–11%) and individuals with sub-Saharan African ethnicity (8%), compared with Caucasians (0.4%).17–19
We note that CD36 antibodies were first reported20 in an ethnic Japanese patient with AML who developed PTR. There have also been several case reports5,17–19,21–25 of patients affected with posttransfusion purpura (PTP), neonatal alloimmune thrombocytopenia (NAIT), platelet refractoriness, and transfusion-related acute lung injury (TRALI) caused by antibodies against CD36.
In the present report, we document the case of a 66-year-old African American woman who developed an alloantibody against CD36/GPIV, after pregnancy and multiple blood transfusions, that likely resulted in PTR. It has been documented in a previous report of PTR due to CD36 antibodies5 that transfusion of CD36-deficient donor platelets is successful in maintaining the platelet counts of affected patients. Platelet donors who lack certain platelet antigens such as CD36 are extremely difficult to find. Most platelet donors are only tested for HLA antigens and not for HPA antigen expression. Some donors are tested for expression of HPA-1a because HPA-1a antibodies are a frequent cause of NAIT and HPA-1a–negative platelets are commonly requested. Antibodies to other platelet antigens are more uncommon; only a few reference laboratories have the ability to screen platelet donors for HPA expression, to our knowledge. At the time of our research, 3 platelet donors who tested CD36 negative were identified at 1 blood center in the United States. Thus, it was not possible to support our patient with CD36-negative platelets through an intensive chemotherapy regimen.
Because it was not possible to quickly find and obtain compatible platelets for our patient, we switched her to a less-aggressive and less-myelosuppressive chemotherapy regimen. The patient received 14 cycles of decitiabine only, and she responded very well for more than 1 year before she died of pulmonary aspergillosis, demonstrating the potential for using this therapeutic regimen for similar cases when compatible platelet support is not available.
We also found a unique compilation of CD36 variants present in this patient. Although all of the variants we found have been previously described5 in individuals with African ancestry at a frequency of 1% to 8%, this patient was unique in having a total of 3 variants in the coding region of CD36. These mutations resulted in the absence of CD36 on the platelets of the patient and subsequent anti-CD36 antibody formation after CD36 antigen exposure.
Conclusions
PTR can be caused by a variety of factors, and immune PTR is usually attributed to HLA antibodies. However, when HLA testing results are negative, other less-common immune-mediated causes should be investigated, such as platelet-specific antibodies, and the ethnic and cultural background of the patient should be taken into account. Also, it is necessary to identify compatible blood donors who lack the cognate platelet antigens. Moreover, given the differences in frequencies of various HPA antigens among ethnic populations worldwide, mechanisms to make these rare, lifesaving platelets more widely available are required.
Abbreviations
- PTR
platelet-transfusion refractoriness
- DIC
disseminated intravascular coagulation
- HLAs
human leukocyte antigens
- HPAs
human platelet-specific alloantigens
- GP
glycoprotein
- CD
cluster of differentiation
- AML
acute myelogenous leukemia
- WBC
white blood cell
- RR
reference range
- RBCs
red blood cells
- PABA
platelet antibody bead array
- PCR
polymerase chain reaction
- PTP
posttransfusion purpura
- NAIT
neonatal alloimmune thrombocytopenia
- TRALI
transfusion-related acute lung injury
- IgG
immunoglobulin G
- FLR
fluorescence ratio
