Platelet Refractoriness: Strategies for Identifying Compatible Platelets

$Platelet Refractoriness: Strategies for Identifying Compatible Platelets$

Mark K Fung, MD PhD Medical Director, HLA/Tissue Typing Laboratory UVM Medical Center Professor and Vice Chair, Population Health Dept of Pathology and Laboratory Medicine University of Vermont Health Network

Oct 16, 2018 Financial Disclosure/ Conflicts

• I am on the data safety monitoring board for study sponsored by Cerus Corporation.

Vermont (where I live) You are here (Boston)

2 Learning Objectives

• Become familiar with different techniques for testing for HLA antibodies • Be able to apply different strategies for identifying HLA-compatible platelets

3 Platelet compatibility: Three things the Laboratory can determine (and improve transfusion response of the patient)

1. Is ABO incompatibility a possible cause for refractoriness to platelet transfusions? 2. Is HLA incompatibility a possible cause for refractoriness to platelet transfusions? 3. Is non ABO, non HLA incompatibility a possible cause for refractoriness to platelet transfusions?

FYI: Platelet refractoriness can occur without any prior platelet transfusion (can be due to pregnancy or RBC transfusions).

Example: 5 month old infant with hypoplastic left heart on ECMO: Schmidt AE et al. ASHI 2018 meeting 4 ABO incompatibility is usually not a problem: On average, largest difference b/t ABO-identical and non- identical platelet transfusion increment was 4,000 / uL -Shehata et al. Transfusion 2009; 49: 2442-2453

Except high titer antibodies: Anti-A or –B titer of ≥ 64 is associated with positive platelet crossmatches (At least twice as many radiolabeled IgG, IgM, or C3d molecules bound to platelets compared to controls) -Ellinger et al. Transfusion 1989; 29:134-138

5 Antibody screens for HLA antibodies

• Lymphocytoxic antibody screen – not used as much anymore • Solid phase antibody assays (two types) – ELISA vs. bead-based (Luminex) – Antigens can be pooled (for screening) – Antigens can expressed singly (one bead-one antigen to determine specificity of antibody)

6 Lymphocytotoxicity Assay – basic principle

2. Complement activation 3. Lysis of cell membrane by complement

1. Antibody binding 4. Trypan Blue (viability stain) Performed by using microwells on a plastic tray that contain lymphocytes of enters cells different HLA types and looking for a pattern of cytotoxicity against a consistent HLA type – like RBC antibody identification panels

7 ELISA and bead-based assays

Plastic wells or beads are coated with purified or recombinantly expressed HLA antigens Y

Y Y Multiple HLA types in one well or bead, useful for

screening for HLA antibodies.

Y Y

Single HLA types in one well Y or bead, optimal for identifying all HLA antibody specificities.

8 HLA antibody detection using micro-bead Micro beads are labeled with a ratio of two different array assay colored fluorescent markers. Patient’s Beads of specific HLA types are assigned a specific ratio Alloantibody of red vs. infrared fluorescent dye. (assigned specific X,Y coordinates of color intensity) – up to 100 positions in this PE anti-IgG example. Fluorescent- Anti-IgG PE is used to identify antibodies binding to tagged certain beads. HLA antigen Bound to color coded bead

Adopted from OneLamba 9 Example of micro bead array analysis

A69 A68 A2 A1

10 Courtesy of OneLambda, Inc. Lymphocytotoxic vs. Solid Phase Antibody Assays

• Lymphocytoxic assay – Less sensitive (this could be good!) – False positive from non-HLA-related cytotoxic substances in patient’s serum – Detects IgG and IgM antibodies – Cannot identify all antibody specificities if there are too many specificities (overlap of different HLA types on the lymphocytes). – Not commonly used, but useful to understand

11 Lymphocytotoxic vs. Solid Phase Antibody Assays (continued) • Solid phase assay (ELISA, Flow beads, Luminex) – More sensitive (too sensitive?) – Detects only HLA antibodies – Detects IgG antibodies only – Can identify all antibody specificities • Using high definition ELISA or single antigen Luminex beads • Can be very powerful method for platelet selection in highly sensitized patients

12 Microbead assays that detect complement activating HLA antibodies: best of both worlds? • C1q is part of complement activation cascade after antibody binding • Assay exists that uses fluorescently tagged C1q to identify antibodies binding to HLA antigen coated microbeads • Hypothetically could identify the more clinically relevant HLA antibodies? • C1q binding HLA antibodies associated with worse outcomes in kidney transplants • Mixed results for platelet refractoriness

Fontaine MJ et al. Transfusion 2011; 51: 2611-2618 Jackman RP et al. Transfusion 2016; 56: 1442-1450. 13 Strategies for identifying compatible platelets for alloimmunized patients

• HLA matched platelets • Acceptable HLA mismatches (CREG/HLAMatchmaker) • Platelet crossmatching • Avoidance of antibody specificities

14 HLA matched platelets

• Requires HLA type of patient and platelet donor • Attempt is made to identify platelet donors or platelet products already in Blood Bank inventory which have same HLA type as patient (HLA-A, -B, -C) – i.e. Patient is A2, A24; B27, B44; Cw2, Cw6 – Look for A2, A24; B27, B44; Cw2, Cw6 donors

15 Difficulties with obtaining HLA matched platelets • Only a few donors in entire donor database completely match patient’s HLA type • Donors may be ABO-incompatible and CMV-positive • Donors may be unavailable or temporarily ineligible to donate platelets

• Requests for daily use of HLA matched platelets that are complete matches for several weeks may be difficult to impossible to support

16 Finding HLA matched platelets: the reality

• Most “HLA matched” platelets are not perfect matches • Often product shipped is the best partial HLA match available in Blood Center’s inventory • Even if a perfectly matched donor is collected immediately, it will take 24-36 hrs before it is available – infectious disease testing, labeling, shipping

17 Solutions to finding HLA matched platelets • Find more HLA matched platelet donors: – Ask for platelet donations from other blood collection centers that maintain an apheresis platelet donor database – Determine if patient’s siblings are HLA matches (controversial) • Expand the definition of “acceptable” platelet donors: – Donors with partial HLA match and “acceptable”* mismatches – Donors with HLA types that avoids patient’s HLA antibody specificity – Randomly compatible platelet crossmatched units Possible acceptable mismatches: *similar to patient’s own HLA type *does not correspond to any patient antibodies *corresponds to only “weak” antibodies

18 Some HLA types are related to each other: Using the HLA “clan” tree or CREGs

CREG = Cross reactive antigen group HLA types can be grouped into serologically similar “families” When using partial HLA matches, look for those donors where the mismatched HLA type at least is in a related CREG

1. Deshpande A. Glob J Transfus Med 2017;2:77-88 2. Rodey GE, et al. Hum Immunol. 1994;39:272–280. 3. Rodey GE. 2nd ed. Durango CO: De Novo; 2000. HLA Beyond Tears.

19 Some common CREGs and the HLA types within the group

Cross Reactive Group No. HLA types within this group

When a perfect HLA match is not possible, the next best alternative may be to use HLA types where the mismatched antigen is at least “similar” or cross-reactive

20 Grading of HLA matching for mismatched platelets

21 Relative effectiveness of different levels of HLA matching of platelets in highly sensitized patients

Duquesnoy et al. (1976) 22 Finding HLA matched platelets: The reality (cont’d) • Sometimes it is better to ask for platelets that avoid certain antibody specificities instead of getting a partial HLA match • For example: Patient is: A2, A24; B7, B44 with antibodies against B27 Possible choices for platelets: Donor #1: A2, A24; B27, B44 (grade B1X match!) Donor #2: A1, A3; B35, B70 (grade D match but B27 negative)

23 CREG-based selection of platelets (A, BU, BX, etc) is outdated • Still used by a number laboratories to choose platelets • Recommend using the following methods to select HLA compatible platelets (to be described later): – HLAMatchmaker (free Excel-based program) – Antibody avoidance using HLA antibody results from single antigen bead (Luminex) or single antigen ELISA wells (high definition trays) • Avoid “stronger” antibodies if not antibodies can be avoided – Platelet crossmatching

24 Important fact: HLA types are based on recombination of a limited number of polymorphisms at fixed locations on the HLA molecule

For example, how different is a donor’s HLA B8 from a patient who is HLA B42 and B53?

B42

B53

Many of HLA B8’s epitopes or polymorphisms are the same as those found in B42 or B53

25 HLAMatchmaker: Choosing acceptable mismatched HLA types based on degree of variation being limited A list of known possible amino acids at specific locations of the HLA molecule 9 17 45 56 62 66 70 74 76 80 82 90 107 127 F gR kMe G Rn rKv aQs D An gGl aLr D G N Variation in HLA types are due S gS Me R Qe rNv aHs N Vd rIa lRg A W K mainly to rearrangement of amino T E Ee gKv H En acid triplets or eplets where there Y Lq rNm iD Es are amino acid sequence Ge differences.

Some positions are more 142 144 149 151 156 158 163 166 171 184 186 193 248 253 immunogenic than others. I tKr aVh vHa L A R Dg Y dP K Av A Ee T tKh aAh aHv W V T Ew H dA R Pi S Ke • Nambiar A et al. Blood 107: 1690- tQr tQr aAr aHe R dT Va Ge 1697, 2006 tAh aRv Q E • Brooks EG et al. Transfusion 48: 2159-2165, 2008 aRw • Pai SC et al. Transfusion 50: 2318- aRr Different HLA types are just different 2327, 2010 aHa combinations of a limited set of amino acid polymorphisms in the amino acid sequence of the HLA molecule 26 Donor HLA-B8 is mismatched but acceptable for patient with HLA type of: A2,A31; B42,B53; Cw2,Cw7

Patient HLA 9 12 14 17 41 45 56 62 66 70 74 76 80 A2 A*0201 F sV R gR A Me G Ge rKv aHs H Vd gTL A31 A*3101 T sV R gR A Me R Qe rNv aHs iD Vd gTL B42 B*4201 Y sV R gR A Ee G Rn qIy aQa D Es rNl B53 B*5301 Y aM R gR A Te G Rn qIf tNt Y En rIa Cw2 Cw*0202 Y aV R sR A Ge G Re qKy rQa D Vn rKl Cw7 Cw*0701 D aV R gR A Ge G Re qNy rQa aD Vs rNl Donor B8 B*0801 D aM R gR A Ee G Rn qIf tNt D Es rNl

HLA polymorphisms associated with B8 are already present in the patient’s “different” HLA type, just in a different combination of HLA types Website for the free Excel-based program: www.epitopes.net

27 Platelet Crossmatching: No HLA typing of donor or patient required !

Patient’s serum is added Anti-HLA/PLT antibodies bind Donor Platelets coat bottom of well

Indicator RBC’s coated with anti-IgG are added and will bind to anti-HLA/PLT antibodies if present

RBCs not immobilized will be Immobilization of RBCs spun down into the center

AABB Technical Manual

28 Comparing crossmatching and HLA matching using CREG match grade

Note that some, but not many positive XM’s generated adequate platelet responses

“B matched” platelets that gave positive XM results and poor response

Kickler et al. Transfusion 1985; 25: 385-389 29 Defining antibody specificities (to avoid certain HLA types)

• Recommend use of single antigen beads (or high definition ELISA) to identify all HLA antibodies • Requires only HLA type of the donors (patient HLA type not needed) • Often serologically related HLA types based on common CREG are also avoided • Computerized program previously created call the “antibody specificity prediction” (ASP) method, but not necessary with single antigen HLA antibody methods.

Petz et al. Transfusion 2000;40:1446-56

30 Pros and Cons of the different strategies when perfect HLA match is not available • CREG matching – Pro: Simple grading system of identifying additional acceptable HLA mismatches – Con: Requires HLA typing of donors and patient. – Con: CREG-matched platelets might still be XM incompatible • HLAMatchmaker program – Pro: Will identify those mismatched HLA types that would be viewed as “self” (tolerant) – Con: Expands number of HLA types, but still might not be enough– useful for patients who are 100% alloimmunized – Con: Slight learning curve to use the Excel-sheet program

31 Pros and Cons of the different strategies when perfect HLA match is not available (cont’d) • Avoiding antibody specificities – Pro: Can be cost effective and allow for more platelets to be used. Does not require patient HLA type. – Con: Requires single antigen ELISA or Luminex testing if patient has many HLA antibodies • Platelet crossmatching – Pro: Can identify useful platelets that are not HLA-matched. No HLA typing needed. – Con: Must use ABO-identical or O donors. – Con: May be difficult if patient has many HLA antibodies

32 What to do when no platelet transfusion is effective • Consider selecting some platelets by crossmatching • Test for HPA antibodies (crossmatching is faster) • Repeat the ABO and HLA antibody testing (every 1- 2 weeks may be necessary) • Review the HLA types of the platelets, they might not be compatible despite being HLA-”selected”

33 Thank you for your attention!

• E-mail: [email protected] • Office: +1 802-847-5114

www.aabb.org 34 How do we support platelet refractory patients? Justin Kreuter, MD @KreuterMD

Tuesday, 10/16/2018 Faculty Disclosure (In compliance with ACCME policy, AABB requires the following disclosures to the session audience) • None

www.aabb.org 2 Objectives

• Interpret the results of HLA/HPA antibody testing and platelet cross-matching to recommend platelet selection options • Recognize the benefits of developing an algorithmic approach to supporting the platelet refractory patient • Discuss how aspects of a given platelet pipeline may impact platelet algorithms

www.aabb.org 3 Our Process

(not THE process)

www.aabb.org 4 Clinical Consult Platelet Crossmatch Platelet crossmatch - Platelet Crossmatch - the cause is…

“non-immune” • Return focus to patient Platelet Crossmatch - the cause is…

“non-immune” “immune” • Return focus to patient • Reserve appropriate • Order Luminex SAB If “immune,” how much challenge?

“immune” Yes No Able to find compatible?

1. Antigen-negative HLA-matched 2. Crossmatched If “immune,” then which HLA antibodies? How do I? If things don’t add up… HPA?

HPA If things don’t add up… mucho antibodies?

Neat sample

1:8 dilution Case 1

The red herring

www.aabb.org 14 Patient = O+

www.aabb.org 15 www.aabb.org 16 Case 2

Meat & Potatoes

www.aabb.org 17 Patient = AB+

www.aabb.org 18 www.aabb.org 19 www.aabb.org 20 www.aabb.org 21 Case 3

The Red Carpet

www.aabb.org 22 Patient = A+

www.aabb.org 23 www.aabb.org 24 www.aabb.org 25 www.aabb.org 26 www.aabb.org 27 Note: testing for allo versus auto-antibody is different

www.aabb.org 28 Case 4

The Solution is Dilution

www.aabb.org 29 Patient = B-

www.aabb.org 30 www.aabb.org 31 www.aabb.org 32 Objectives, revisited

www.aabb.org 33 Take Home Points HLA- • Communication • Optimize [availability v success] match XM Antigen- negative www.aabb.org 34 Take Home Points HLA- • Communication • Optimize [availability v success] match • Continuous learning XM Antigen- negative www.aabb.org 35 Our process: [email protected] @KreuterMD My Patient Is Refractory to Platelet Transfusions...Now What?

Brian R. Curtis, Ph.D., D(ABMLI), MT(ASCP)SBB

Sr. Director, Versiti PD and Clinical Labs Director, Platelet & Neutrophil Immunology Lab Investigator, Blood Research Institute BloodCenter of Wisconsin, part of Versiti, Milwaukee, WI

October 18, 2018 Faculty Disclosure (In compliance with ACCME policy, AABB requires the following disclosures to the session audience) • Member, Platelet Advisory Board – Ionis Pharmaceuticals

www.aabb.org 2 Platelet Transfusion Refractoriness (PTR)

• Failure to achieved acceptable increment with 2 platelet transfusions in a row

• Platelet counts should be performed 10 to 60 min post transfusion

• CCI < 5000-7500 or PPR of < 20%

• Most cases show no increment or count < 5000 platelets/µL

• Typically patients with hematologic disease or malignancy

www.aabb.org 3 PTR Frequency

• 25-75% of patients with HM experience one or more poor platelet increments

• Trial to Reduce Alloimmunization to Platelets (TRAP) – multi-institutional, randomized, blinded trial to determine if LR and/or UVB platelets prevent alloimmunization, PTR

– 530 patients who were receiving induction chemotherapy for AML

– 16% (NLR), 7-10% (LR) developed PTR – 13% (NLR), 3-5% (LR) developed immune PTR

TRAP TRIAL. NEJM 1997;337(26):1861. www.aabb.org 4 PTR Financial Impact

Hospital Days Median Cost No PTR 14.4 $33,818 PTR 35 $103,956

KR Meehan, et al. Am J Hematol 2000;64:251-256.

N = 40 PTR patients Avg. 14.4 crossmatch plt transfusions over 33 months. Kit costs alone = 173,000 EUR (~$188K)

P Rebulla, et al. Br J Haematol 2004;125:83-9.

www.aabb.org 5 Causes of PTR

• Majority of PTR is non-Immune *~88%! – DIC, fever, sepsis, chemotherapy, splenomegaly

• Immune causes (~12%): – class I HLA antibodies - most common – platelet-specific (HPA) antibodies - uncommon – platelet autoantibodies - probably more common than we know – drug-dependent antibodies - vancomycin, pipericillin, sulfamethoxazole, oxaliplatin – CD8+ T cells – PTP (autoantibodies)

*Doughty HA, et al. Vox Sang 1994;66:200-5.

www.aabb.org 6 Now what? • Non-Immune PTR – Treat the infection, fever, etc. – Patience and transfuse platelets when required – Ideally, transfuse fresh, ABO identical, apheresis platelets

• Immune causes – autoantibodies – transfuse platelets when required, immune suppression? – drug-dependent antibodies – stop the implicated drug – HLA and/or HPA antibodies – transfuse antigen matched platelets

www.aabb.org 7 Platelet Matching HLA-A1/A3, B8/B27

Grade Description Donor Phenotype A 4-antigen match A1/A3, B8/B27 HLA matching B1X 1 cross-reactive group A1/A3, B8/B7 C 1 mismatched antigen A1/A3, B8/B35 D 2 or more mismatched antigens A1/A32, B8/B35 R Random A2/A28, B7/B35

HPA matching Patient: HPA-1a antibodies Donor: HPA-1b/1b

Crossmatching + www.aabb.org 8 Triplet & Eplet HLA Matching

MO Muench, et al. Transfusion 2016;56;1419–1429 www.aabb.org 10 Platelet Recovery

MO Muench, et al. Transfusion 2016;56;1419–1429 www.aabb.org 11 Reduced T Cell Immunization and Partial IT in Mice Receiving UV+R

R Jackman, et al. Transfusion 2017;57:337-48 www.aabb.org Leukofiltration + UV-R prevents alloimmune PTR in dogs

Overall, 31 of 32 (97%) recipient dogs accepted FLR+MPR (UV-R) platelets; none developed antibodies to donor lymphocytes.

SJ Slichter, et al. Blood 2017;130:1052 www.aabb.org 13 Alloimmunization Not Prevented in Humans Using Pathogen-Inactivated Platelets

• PREPAReS study – Netherlands, Canada, Norway

• Control arm - 6 of 177 HM patients developed HLA class I antibodies

• Intervention arm - 7 of 179 patients developed HLA class I antibodies

• “In contrast to recently published animal studies, pathogen inactivation treatment did not result in a reduction of HLA class I alloimmunization.” (% immunized low, untreated rbcs given)

P.F. van der Meer et al. Blood 2018;132:223-231 www.aabb.org 14 Alloimmunization Increased in Humans Transfused Pathogen-Inactivated Platelets

• HM patients (n = 397) randomized to receive either untreated (n = 205) or UV-R pathogen reduced (n = 192) platelet transfusions (PREPAReS trial)

Of patients antibody negative at enrollment:

• Untreated platelet arm - 11 (5.4%) develop class I HLA antibodies

• UV-R treated platelet arm – 25 (13%) developed class I HLA antibodies (p=0.006)

R Saris, et al. ISBT 2018 abstract 5A-S37-03 www.aabb.org 15 Dendritic Cells Phagocytose Stored, Apoptotic, PR Platelets

**Animal Studies Used Fresh Platelets

A. Saris, et al. Front Immunol 2018;9:1251. www.aabb.org 16 Human Platelet Alloantigens Frequencies Glycoprotein Amino Acid Antigens (Caucasians) Location Substitution

HPA -1a/a 98% GP IIIa p.Leu33Pro HPA -1b/b 2%

HPA - 2a/a 99% GP Iba p.Thr145Met HPA - 2b/b 1%

HPA - 3a/a 85% GP IIb p.Ile843Ser HPA - 3b/b 15%

HPA - 4a/a 99.9% GP IIIa p.Arg143Gln HPA - 4b/b <0.1%

HPA - 5a/a 99% GP Ia p.Glu505Lys HPA – 5b/b 1% CD36 (GPIV) Glycoprotein

• 88 kD (reduced/nonreduced)

• Expressed on: monocytes, capillary endothelium, erythroblasts, adipocytes, and platelets

• Receptor for: Types I & IV collagen, malaria-infected rbcs, thrombospondin, oxidized LDL, long-chain fatty acids

• Platelets of some people lack expression of CD36 CD36 Deficiency Platelets: • Asians: 3-11% • African Americans: 2.5% • Sub-Saharan Africans: 7.8% • Middle East: ? • Caucasians: 0.06%

Type I: Absent from Platelets and Monocytes – risk of immunization against CD36 by transfusion or pregnancy

Type II: Absent from Platelets only Immune Platelet Disorders Caused by Anti-CD36

• Neonatal Alloimmune Thrombocytopenia (NAIT)

• Post-Transfusion Purpura (PTP)

• Platelet Transfusion Refractoriness (PTR) CD 36-Neg Platelet Transfusions Work

180 Pre 160 1 hr Post CD36- Plts 140 120

100/L 9

8010

Platelets x x Platelets 60 40 20 0 PatientPatient 11 Patient 2 Patient 2 (Lebanese) (Chinese)

Saw, CL, Syzkoluk H, Curtis BR, Nickerson P, Goldman M. Transfusion 2010;50(12):2638-42. www.aabb.org 21 My Patient Is Refractory to Platelet Transfusions...Now What?

10/16/2018 Faculty Disclosures

The following faculty have no The following faculty have a relevant financial relationships to relevant financial relationship: disclose: – Brian Curtis PhD – Jeffrey Papiernik MD Ionis Pharmaceuticals: – Justin Kreuter MD Consultant – Mark Fung MD, PhD Cerus Corporation: Consultant

www.aabb.org 2 Learning Objectives

• Differentiate the various contributors to platelet transfusion refractoriness (PTR), the majority of which are NOT immune mediated • Demonstrate an understanding of the technological principles used for HLA/HPA antibody testing and platelet cross-matching • Interpret the results of HLA/HPA antibody testing and platelet crossmatching to recommend platelet selection options • Recognize the benefits of developing an algorithmic approach to supporting the platelet refractory patient • Discuss how aspects of a given platelet pipeline may impact platelet algorithms