A novel minor histocompatibility antigen recognized by HLA ... - Nature

Bone Marrow Transplantation, (1999) 24, 129–137  1999 Stockton Press All rights reserved 0268–3369/99 $12.00 http://www.stockton-press.co.uk/bmt

A novel minor histocompatibility antigen recognized by HLA-A31 restricted cytotoxic T lymphocytes generated from HLA-identical bone marrow donor lymphocytes M Yazaki1, T Takahashi2, M Andho1, Y Akatsuka3, T Ito1, Y Miyake1, Y Ito1, S Nakamura4 and Y Wada1 1

Department of Pediatrics, Nagoya City University Medical School, Nagoya; 2Laboratory of Immunology, Aichi Cancer Center Research Institute, Nagoya, Japan; 3Department of Human Immunogenetics, Fred Hutchinson Cancer Research Center, Seattle, WA, USA; and 4Department of Clinical Laboratory, Aichi Cancer Center Hospital, Nagoya, Japan

Summary: Bulk cytotoxic T lymphocytes (CTL) were generated by in vitro stimulation of BMT donor lymphocytes with Philadelphia chromosome (Ph)-positive leukemic cells from an HLA-identical sibling patient. CTL were cytotoxic against the patient’s leukemic cells as well as the EBV-lymphoblastoid cell line (EBV-LCL) generated from the patient’s cells, suggesting that they recognize a minor histocompatibility antigen (mHAg). Subsequently, several CTL lines were established by a limiting dilution method and analyzed. One of these CTL lines, 16C12 CTL which used a single TCR␤V3S1 for CD8 cells, lysed HLA-A31-positive leukemic cells and EBV-LCL, but not fibroblasts. The cytotoxicity against the patient’s leukemic cells and EBV-LCL was blocked by anti-HLA-A31 moAb, anti-HLA-class I moAb, and anti-CD8 moAb, suggesting that this mHAg was presented with HLA-A31. The antigen recognized by 16C12 CTL seemed to be a novel mHAg, since HLAA31 restricted antigen has not been reported to date and 16C12 CTL showed no cytotoxicity against EBVLCL which probably express known mHAgs. CTL detecting this mHAg may play an important role in the GVL effect in HLA-A31-positive BMT patients. Keywords: minor histocompatibility antigen; GVL; CTL; BMT

The GVL effect has been considered to be very important in preventing leukemia relapse after bone marrow transplantation (BMT).1–5 From this point of view, donor leukocyte infusion (DLI) was introduced to CML patients and ALL patients with relapse after BMT, with promising results.6–8 However, DLI was often complicated by GVHD. Therefore, generation of CTL with selective specificity against leukemic cells from donor PBL is important. As several reports have suggested, minor histocompatibility antigens (mHAg) predominantly expressed on hematopoCorrespondence: Dr M Yazaki, Department of Pediatrics, Nagoya City University Medical School, 1-Kawasumi, Mizuho-cho, Mizuho-ku, Nagoya, 467-8601, Japan Received 21 December 1998; accepted 17 February 1999

ietic cells are the target antigens for the GVL effect in HLA-identical BMT.9–17 We report here HLA-A31-restricted CTL generated from an HLA-identical sibling donor which recognizes a new mHAg on leukemic cells, EBVlymphoblastoid cell lines, (EBV-LCL) and PHA-stimulated blasts.

Materials and methods Patient The patient was a 12-year-old girl with Philadelphia chromosome (Ph)-positive (minor BCR-ABL) ALL. Cell surface markers on the leukemic cells were CD10, CD13, CD19, CD34 and HLA-DR. She (HLA-A31,A24; B35,B52;Cw3,-; DR15,DR15; DRB1 *1501,*1502) received an allogeneic BMT from an HLA-identical brother, TW without GVHD. She successfully received several DLIs from TW to prevent relapse after BMT without GVHD or myelosuppression, because of the persistent BCR-ABL chimeric mRNA detected by the reverse transcriptase (RT)PCR method after BMT.18 Bulk CTL culture A bulk CTL culture was established by stimulating 1 × 105 BMT donor PBL with 1 × 105 of 35 Gy irradiated patient’s leukemic cells in a 96-well round-bottomed microtiter plate and suspended in RPMI 1640 plus 15% pooled human AB serum and 10 mm HEPES. On day 6, 1 × 105 viable cells were restimulated with 1 × 106 irradiated leukemic cells, and 2% highly purified IL-2 (Biotest, Dreieich, Germany) in a 24-well flat-bottomed culture plate. From day 14, cultures were restimulated and expanded weekly with 1 × 106 irradiated leukemic cells and 20% T cell growth factor (TCGF; Biotest). Allogeneic CTL were induced from unrelated PBL, which shared no HLA with the patient, stimulated with patient’s leukemic cells and maintained using the same method as described above. Limiting dilution of bulk CTL A bulk CTL culture was cloned by a limiting dilution method on day 6 after the 5th stimulation. One, 2, 4, 8,

CTL reactive with new mHAg M Yazaki et al

130

and 16 blastic cells/well were expanded in RPMI 1640 plus 15% human serum, 10 mm HEPES, 100 IU/ml human recombinant IL-2 (Takeda Chemical Industries, Osaka, Japan), and 10% TCGF (Biotest) in a 96-well roundbottomed mictotiter plate. Feeder cells contained 2 × 104 of 20 Gy irradiated donor’s PBL and 2 × 104 of 35 Gy irradiated patient’s leukemic cells. From day 8, cultures from each well were expanded and restimulated weekly with 2 × 104 irradiated donor’s PBL and 2 × 104 irradiated patient’s leukemic cells under the same conditions as described above.

Target cells Leukemic cells were collected and frozen in liquid nitrogen until use. Frozen leukemic cells were thawed 1 day before a chromium release assay and cultured with 10 ng/ml human tumor necrosis factor-␣ (TNF-␣; Boehringer Mannheim, Mannheim, Germany) overnight in order to increase the sensitivity to a chromium release assay as reported by Dolstra et al.16 EBV-LCL from the patient was established from PBL at diagnosis by infection with EBV of B95–8 in the presence of 800 ng/ml cyclosporin A (Sandoz Industries, Basel, Switzerland). PHA blasts were generated from PBL at diagnosis by stimulating PBL with 1% PHA-M (Difco Laboratories, Detroit, MI, USA) for 3 days. PHA blasts were further cultured with 10% TCGF for 3–4 days and were frozen in liquid nitrogen. HLA-A31positive and/or HLA-B35-positive EBV-LCLs were obtained from Mr Kenichi Ohya, Aichi Red Cross Blood Center, Seto, Japan. The patient’s bone marrow fibroblasts were established from the bone marrow at diagnosis.

Chromium release assay Target cells except fibroblasts were incubated with 100 ␮Ci of Na251CrO4 (CSJ-11; Amersham Japan, Tokyo, Japan) at 37°C for 1 h. Fibroblasts were labelled with 200 ␮Ci of Na251CrO4 for 18 h.19 Labelled cells were washed three times with RPMI 1640 with 2% FCS. After spinning down at 50 g for 1 min, 5 × 103 leukemic cells, EBV-LCL, and PHA blasts were incubated with CTL line at different effector to target (E/T) ratios for 4 h. Fibroblasts were incubated with CTL for 8 h. After incubation, the supernatant was collected using the Supernatant Harvesting System (Skatron Instruments, Tranby, Norway) and the radioactivity of supernatant was counted for 1 min using a ␥ scintillation counter. Spontaneous release was measured by incubation of target cells in the absence of effector cells, and maximum release was determined by lysing the target cells in 2.5% Triton-X 100 (Sigma, St Louis, MO, USA). The percentage of specific lysis was calculated as follows: 100 × (experimental mean c.p.m. ⫺ spontaneous release mean c.p.m.)/(maximum release mean c.p.m. ⫺ spontaneous release mean c.p.m.). In a cold target cell inhibition experiment, mixtures of non-radioactive (cold) and 51Cr-labelled (hot) target cells at different ratios were incubated with CTL line at fixed E/T ratio.

Blocking of cytotoxicity by moAbs moAbs (anti-CD3, -CD4, -CD8) were added and incubated for 30 min at 37°C until target cells were added. Anti-HLA class I and anti-HLA class II moAbs were added just before the addition of target cells and incubated for 20 min at 37°C. Thereafter, a chromium release assay was conducted as described above. moAbs used were anti-CD3 (OKT3; American Tissue Culture Collection, Rockville, MD, USA), anti-CD4 (OKT4; American Tissue Culture Collection), anti-CD8 (CTL5–11a; Aichi Cancer Center Research Institute, Nagoya, Japan), anti-HLA-class I (wb/32–3a; Aichi Cancer Center Research Institute), and anti-HLA-class II (Ia-91–5a; Aich Cancer Center Research Institute). Mouse anti-HLA-A31 moAb (2D12), reacting with HLA-A31 but not with HLA-A33, -A2, -A11, -A24, -A26, -B35, -B46, -B55, -B61, -B62, -Cw1, and -Cw9, was kindly provided by Drs Hiroeki Sahara and Noriyuki Sato, 1st Department of Pathology, Sapporo Medical University School of Medicine, Sapporo, Japan.20 Immunofluorescence To establish the phenotype of CTL lines, viable cells were incubated with various moAbs for 30 min at 4°C and analyzed with FACScan flow cytometer (Becton Dickinson, Mountain View, CA, USA). moAbs used for immunofluorescence analysis were anti-TCR␣␤ (Becton Dickinson), anti-TCR␥␦ (Becton Dickinson), anti-CD3 (Leu4; Becton Dickinson), anti-CD4 (Leu3; Becton Dickinson), anti-CD8 (Leu2; Becton Dickinson), anti-CD2 (NUTER; Becton Dickinson), anti-CD28 (KOLT2; Becton Dickinson), antiCD11b (MAC-1; Nichirei, Tokyo, Japan) and anti-HLADR (HLA-DP,DQ,DR; DAKO; Glostrup, Demmark). TCR␤V gene expression on 16C12 CTL line Total RNA was prepared from 16C12 CTL according to the method of Chomczynski and Sacchi21 using a commercially available kit (Stratagene, La Jolla, CA, USA). The entire RNA content was subjected to synthesis of cDNA by MMLV RT (GibcoBRL, Gaithersburg, MD, USA) and primer p(dT)15 (Boehringer Mannheim). PCR reactions were carried out on a GeneAmp PCR system 9600 (Perkin Elmer, Foster City, CA, USA) in a 20 ␮l of volume with appropriate amount of cDNA from the line as previously described.22 Briefly, each reaction mixture contained one of 24 TCR␤V family-specific primers, 500 ␮m dNTPs, 1.5 mm MCl2, 1 unit AmpliTaq Gold DNA polymerase (Perkin Elmer) in 1 × buffer. The PCR conditions were as follows: one cycle of 94°C for 9 min, followed by 35 cycles of 94°C for 30 s; 58°C for 30 s; 72°C for 30 s, and finally one cycle of 72°C for 10 min. The amplified PCR products were size fractionated on 2.2% mini-agarose gels (Mupid II; Cosmo Bio, Tokyo, Japan). The DNA fragments were excised and purified with a kit (QlAquick Gel Extraction Kit; Qiagen, Santa Clara, CA, USA). Direct DNA sequencing was performed using BC primer used in the PCR with Thermo Sequenase (Amersham, Cleveland, OH, USA). For TCR expression, 16C12 CTL was analyzed by FACScan using anti-CD4-PE (Becton Dickinson), anti-CD8-PE (Becton


Dickinson), TCRV␤ 3-FITC (Beckman Coulter, Miami, FL, USA) and TCRV␤ 5.1-FITC (Beckman Coulter).

activity against TNF-␣ pretreated leukemic cells and EBV-LCL. Limiting dilution of bulk CTL

Results Generation of bulk CTL A bulk CTL culture was initiated by stimulating BMT donor PBL with the HLA-identical patient’s leukemic cells and then stimulated weekly with the leukemic cells. The bulk CTL showed only weak cytotoxicity against the patient’s leukemic cells, but strong cytotoxicity against the leukemic cells pretreated with TNF-␣ and against EBVLCL of patient origin, but not against donor EBV-LCL or K562, a standard target for NK activity, suggesting that the antigen detected is a mHAg rather than a leukemia-specific antigen (Figure 1a). The lysis of the patient’s PHA blasts by the bulk CTL was not observed. The cytotoxic activity of bulk CTL was increased by repeated stimulations. The lysis of TNF-␣ pretreated leukemic cells by the bulk CTL was inhibited by anti-CD3, -CD8 and HLA class I moAbs, but not by anti-CD4 and HLA class II moAbs (Figure 1b). The lysis of patient-derived EBV-LCL was also inhibited by anti-CD8 and HLA class I moAbs, suggesting that HLA class I-restricted CD8 cells in the bulk CTL have cytotoxic

a

% Specific lysis

60

40

20

Cytotoxicity of CTL lines CD8 cell dominant, 1F8 and 16E3, lines showed cytotoxicity against the patient’s TNF-␣ pretreated leukemic cells and patient-derived EBV-LCL, but not against donorderived EBV-LCL or K562. These two showed no cytotoxicity against the patient’s PHA blasts (Figure 2a, b). CD4 cell dominant, 8H9 and 16B6, lines killed patient-derived EBV-LCL, but showed weak cytotoxicity against the patient’s TNF-␣ pretreated leukemic cells (Figure 2e, f). 16C12 and 2E12 lines of mixed CD4 and CD8 cells showed strong cytotoxicity against the patient’s TNF-␣ pretreated leukemic cells and patient-derived EBV-LCL. 16C12 line showed cytotoxicity against the patient’s PHA blasts, whereas 2E12 did not (Figure 2c, d). 16C12 line showed constant growth and killing activity, whereas other CTL lines did not. Accordingly, 16C12 line was further analyzed in the subsequent study. TCR␤V gene analysis of 16C12 CTL line

0 10

b

The growth efficiency of of 16, 8, 4, 2, 1 cells/well was 42%, 24%, 13%, 18% and 3%, respectively. About 60% of the lines showed cytotoxicity against both the patient’s leukemic cells pretreated with TNF-␣ and patient-derived EBV-LCL, but not against donor-derived EBV-LCL, similarly to the bulk CTL. Six CTL lines, showing relatively good growth, were analyzed for the phenotype (Table 1). All CTL lines were TCR␣␤ positive and TCR␥␦ negative. 1F8 and 16E3 were CD8 cell dominant, while 8H9 and 16B6 were CD4 cell dominant. 16C12 and 2E12 were mixed with CD4 and CD8 cells.

30 E/T ratio

50

Target cells: TNF- α pretreated patient’s leukemic cells

Blocking moAb: HLA class ll HLA class l

Attempts were made to establish a 16C12 clone consisting of only CD8 cells, but were not successful. So, TCR␤V gene expression on CD8 cells was studied. By RT-PCR, two cDNA clones, TCR␤V3S1 and TCR␤V5S1, were obtained from the 16C12 CTL line consisting of CD4 and CD8 cells. 16C12 CTL cells were analyzed by FACS using PE-conjugated anti-CD8 moAb, PE-conjugated anti-CD4 moAb, FITC-conjugated anti-TCR␤V3S1 moAb, and

CD8 CD4

Table 1

CD3

Phenotypea

Phenotype of CTL lines 1F8

16E3

16C12

2E12

8H9

16B6

85 0 97 8 96 99 3 57 97

94 0 99 2 99 100 8 86 99

82 0 99 39 72 100 28 88 100

86 0 99 40 82 100 34 59 100

82 0 98 97 10 100 64 57 99

95 0 97 99 1 100 50 76 99

Control 0

20

40 60 Specific lysis (%)

80

Figure 1 Characteristics of bulk CTL. (a) Cytotoxicity of bulk CTL. The following target cells were used: patient’s leukemic cells (---왖---), TNF-␣ pretreated patient’s leukemic cells (––쐽––), patient’s EBV-LCL (⭈⭈⭈䉬⭈⭈⭈), donor’s EBV-LCL (⭈⭈⭈䊊⭈⭈⭈) and K562 (哹哹). (b) moAb blocking of the cytotoxicity by bulk CTL against TNF-␣ pretreated patient’s leukemic cells. 1/500 ascites (approximately 1 mg moAb/ml) containing each moAb was added to the chromium release assay, as described in Materials and methods. E/T ratio was 20:1.

TCR␣␤ TCR␥␦ CD3 CD4 CD8 CD2 CD28 CD11b HLA-DR a

Percentage of positive cells was detected by a flow cytometer.

131


a

b

1F8 (CD8+)

16E3 (CD8+) 80

60

% Specific lysis

% Specific lysis

20

10

40

20

0

0 0.2

c

0.4 E/T ratio

0.6

0.8

1 E/T ratio

d

16C12 (CD8+,CD4+)

2

2E12 (CD8+,CD4+) 60

60 % Specific lysis

% Specific lysis

80

40

20

1 E/T ratio

e

40

20

0

0

1 E/T ratio

2

f

8H9 (CD4+)

2

16B6 (CD4+) 40

% Specific lysis

40 % Specific lysis

132

10

20

0

0 1 E/T ratio

2

2 E/T ratio

4

Figure 2 Target specificity of CTL lines obtained after limiting dilution method. The target cells used were TNF-␣ pretreated patient’s leukemic cells (–쐽–), patient’s EBV-LCL (%䉬%), donor’s EBV-LCL (%䊊%), patient’s PHA blasts (⭈⭈⭈왖⭈⭈⭈) and K562 (⭈⭈⭈ ⭈⭈⭈).

FITC-conjugated anti-TCR␤V5S1 moAb. The two-color analysis showed that most of CD8-positive cells expressed TCR␤V3S1, suggesting that 16C12 CTL was monoclonal (Figure 3).

inhibited by anti-CD3, -CD8 and HLA class I moAbs, suggesting that CD8, but not CD4, cells of 16C12 CTL lysed TNF-␣ pretreated leukemic cells and EBV-LCL by recognizing an antigen presented by HLA class I.

Antibody blocking experiment of 16C12 CTL

Cytotoxicity of 16C12 CTL against EBV-LCL derived from the patient’s family

The lysis of TNF-␣ pretreated leukemic cells by 16C12 CTL was inhibited by anti-CD3, -CD8, and HLA-class I moAbs, but not by anti-CD4 and HLA class II moAbs (Figure 4). The lysis of patient-derived EBV-LCL was also

We tested the cytotoxicity of 16C12 CTL against EBVLCLs derived from the patient’s family. It lysed patientand father-derived EBV-LCL, but not donor-EBV-LCL,


specificity (Figure 5). Unlabelled patient’s EBV-LCL and father’s EBV-LCL, both of which were sensitive to 16C12 CTL, inhibited the lysis of radiolabelled patient-derived EBV-LCL by 16C12 CTL, but not EBV-LCLs of donor, brother and mother origin.

104

Anti-CD8-PE

103

102

HLA restriction of cytotoxicity of 16C12 CTL 101

Since the cytotoxicity of 16C12 CTL against EBV-LCLs derived from the family members suggested the restriction by HLA-A31, -B35 or -Cw3, it was tested against unrelated EBV-LCLs expressing these HLA allotypes. Besides patient-derived EBV-LCL, 16C12 CTL showed clear cytotoxicity against two EBV-LCLs (SPL, KM) and weak cytotoxicity against three EBV-LCLs (HT, MM, SN). All the EBV-LCLs were found to express HLA-A31, suggesting the HLA-A31 restriction of 16C12 CTL (Table 3).

100 100

101 102 103 Anti-TCR β V3S1-FITC

104

Figure 3 TCR␤V gene analysis of 16C12 CTL line. 16C12 CTL line was analyzed by FACS using anti-CD8-PE and anti-TCR␤V3S1-FITC.

Target cell: TNF- α pretreated patient’s leukemic cells Blocking moAb:

Anti-HLA-A31 moAb blocking experiment of 16C12 CTL HLA class II

Fortunately, mouse anti-HLA-A31 moAb was available20 for an antibody blocking experiment (Figure 6). The lysis of patient’s TNF-␣ pretreated leukemic cells, patientderived EBV-LCL, and an unrelated HLA-A31-positive

HLA class I CD8 CD4 CD3

80

Control 20 40 Specific lysis (%)

60 % Specific lysis

0

Figure 4 moAb blocking of the cytotoxicity by 16C12 CTL against TNF-␣ pretreated patient’s leukemic cells. 1/500 ascites (approximately 1 mg moAb/ml) containing each moAb was added to the chromium release assay, as described in Materials and methods.

60

40

20

brother-EBV-LCL or mother-derived EBV-LCL (Table 2), suggesting that 16C12 CTL recognizes a mHAg on the patient’s leukemic cells and EBV-LCL presented by HLAA31, HLA-B35 or HLA-Cw3.

0 0

Cold target cell inhibition experiments were performed with EBV-LCLs from the patient’s family to study the

Origin of EBV-LCL

Patient Brother (donor) Father Mother Brother

20

Figure 5 Cold target cell inhibition experiment. The cytotoxicity by 16C12 CTL against the patient’s EBV-LCL at E/T ratio of 2 was measured in the presence of the following non-radioactive target cells. Patient’s EBV-LCL (–쮿–), donor’s EBV-LCL (%䉫%), father’s EBV-LCL (⭈⭈⭈䊊⭈⭈⭈), mother’s EBV-LCL (---왕---) and brother’s EBV-LCL (--- ---) (see Table 2).

Cold target cell inhibition experiment of 16C12 CTL

Table 2

10 Cold/Hot ratio

Specific lysis of 16C12 CTL against EBV-LCLs derived from the patient’s family Sex

% Specific lysis (E/T ratio)

HLA

2

1

0.5

A

B

C

DR

DRB1 of DRI5

F M

77 ⫾ 3.0 1 ⫾ 2.4

59 ⫾ 7.8 0 ⫾ 1.8

43 ⫾ 7.6 0 ⫾ 2.0

A31/A24 A31/A24

B35/B52 B35/B52

Cw3/Cw3/-

DR15/DR15 DR15/DR15

*1501/*1502 *1501/*1502

M F M

54 ⫾ 4.7 11 ⫾ 2.0 1 ⫾ 4.0

35 ⫾ 9.2 1 ⫾ 2.8 0 ⫾ 2.1

32 ⫾ 6.0 0 ⫾ 1.1 0 ⫾ 0.8

A31/A2 A24/A24 A31/A24

B35/B60 B52/B7 B35/B7

Cw3/Cw7/Cw3/Cw7

DR15/DR4 DR15/DR1 DR15/DR1

*1501 *1502 *1501

133


134

Table 3

Cytotoxicity of 16C12 CTL against HLA-A31 positive EBV-LCLs

EBV-LCL

HLA type

Patient Donor DEU HT MT-14B SPL OLGA JHAF KM MM YN SN AM RH HU HM JO528239 BM9 KOSE WT100BIS KT-17 TISI EHM HS

% Specific lysis (E/T ratio)

HLA-A

HLA-B

HLA-C

HLA-DR

2:1

1:1

A31/A24 A31/A24 A31 A31/A24 A31 A31 A31 A31 A31/A24 A31/A11 A31/A24 A31/A2 A31/A26 A31/A11 A24/A26 A26/A26 A1 A2 A2 A11 A2/A11 A24 A3 A2/A26

B35/B52 B35/B52 B35 B15/B52 B60 B62 B62 B51 B35/B61 B54/B61 B35/B7 B35/B46 B61/B70 B59/B61 B35/B7 B35/B7 B35 B35 B35 B35 B35/B62 B35 B35 B35/B7

Cw3 Cw3 Cw4 ND Cw10 Cw1 Cw1/Cw10 Cw8 Cw3/Cw8 Cw1/Cw8 Cw3/Cw7 Cw3/Cw1 Cw1 Cw3/Cw1 Cw3/Cw7 Cw3/Cw7 Cw4 Cw4 ND Cw4 Cw4/Cw9 Cw4 Cw4 Cw3/Cw7

DR15/DR15 DR15/DR15 DR4 DR2/DR6 DR4 DR8 DR8 DR4 DR4/DR8 DR4/DR8 DR1/DR9 DR8/DR9 DR8/DR14 DR4/DR4 DR15/DR1 DR15/DR1 DR11 DR8 DR13/DR14 DR1 DR4 DR11 DR1 DR1/DR4

62 ⫾ 2.0 2 ⫾ 1.2 3 ⫾ 0.9 9 ⫾ 2.3 0 ⫾ 1.7 72 ⫾ 4.8 0 ⫾ 0.6 2 ⫾ 0.5 26 ⫾ 2.2 9 ⫾ 0.9 1 ⫾ 3.8 8 ⫾ 1.4 0 ⫾ 1.9 0 ⫾ 2.4 0 ⫾ 3.8 0 ⫾ 1.3 0 ⫾ 3.5 0 ⫾ 1.4 0 ⫾ 1.4 1 ⫾ 0.9 0 ⫾ 1.0 0 ⫾ 0.5 0 ⫾ 1.2 0 ⫾ 1.6

32 ⫾ 3.9 2 ⫾ 0.7 0 ⫾ 0.6 5 ⫾ 1.4 0 ⫾ 1.5 57 ⫾ 3.9 0 ⫾ 0.7 1 ⫾ 0.6 11 ⫾ 1.7 4 ⫾ 2.0 0 ⫾ 1.8 8 ⫾ 1.7 0 ⫾ 0.3 0 ⫾ 1.5 0 ⫾ 1.9 0 ⫾ 2.7 0 ⫾ 2.2 0 ⫾ 1.5 0 ⫾ 1.4 1 ⫾ 0.9 0 ⫾ 2.3 0 ⫾ 0.8 0 ⫾ 0.9 0 ⫾ 0.6

ND = not detected.

EBV-LCL (SPL), which was sensitive to 16C12 CTL, was strongly inhibited by anti-HLA-A31 moAb, confirming the HLA-A31 restriction of this CTL activity.

Patient’s TNF- α LC

Lysis of leukemic cells by 16C12 CTL We next examined whether or not 16C12 CTL can lyse HLA-A31-positive TNF-␣ pretreated leukemic cells from other patients. Three out of eight HLA-A31-positive ALLs, including the patient’s own leukemic cells, were killed by 16C12 CTL (Table 4), although it showed no cytotoxicity against five AML and one acute monocytic leukemia (AMoL) samples.

Patient’s EBV-LCL

SPL (unrelated HLA-A31+ EBV-LCL)

No lysis of patient’s fibroblasts by 16C12 CTL

0

20

40

60

80

Specific lysis (%)

Figure 6 Anti-HLA-A31 moAb blocking of the cytotoxicity by 16C12 CTL against various target cells. Blocking studies were performed in the presence of the following moAbs. None (쐍), 1/100 ascites (1 mg moAb/ml) anti-HLA class II ( ), 1/100 ascites (1 mg moAb/ml) antiHLA class I ( ), or 6.25 ␮g/ml purified anti-HLA-A31 ( ) moAb were added. E/T ratio of 2. Patient’s TNF-␣ LC, TNF-␣ pretreated patient’s leukemic cells.

To determine whether 16C12 antigen is expressed on fibroblasts, we initiated a fibroblast culture from the patient’s bone marrow aspirate at diagnosis and then examined it. Although allogeneic CTL, which were generated from unrelated PBL which shared no HLA-A, -B, -DR, showed strong killing against patient’s fibroblasts, ie 60% at E/T ratio of 20 and 32% at E/T ratio of 2, 16C12 CTL did not show significant killing against fibroblasts (4.8% at E/T ratio of 2). The susceptibility of fibroblasts to CTL killing was not induced by 10 ng/ml TNF-␣ pretreatment for 48 h (Figure 7).


Table 4

135

Cytotoxicity of 16C12 CTL against HLA-A31 positive TNF-␣ pretreated leukemic cells

Origin of leukemic cells

Sex

Patient YM YS TS TT AK MK SU HM TO MS MY KM HN

Diagnosis

F M M M M F M F F F M F F F

HLA-A

ALL(Ph,C) ALL(C) ALL(Ph,C) ALL(Ph,C) ALL(C) ALL(C) ALL(C) ALL(C) AML AML AML AML AML AMoL

HLA-B

A31/A24 A31/A24 A31/A24 A31/A11 A31/A33 A31/A24 A31/A26 A31/A2 A31/A2 A31/A26 A31/A2 A31/A2 A31/A2 A31/A24

B35/B52 B52/B7 B35/B51 B51/B60 B61/B44 B61/B48 B61 B7/B48 B51/B60 B13/B60 B51 B35/B7 B51/B46 B62/B55

% Specific lysis (E/T ratio) 2:1

1:1

59 ⫾ 7.0 0 ⫾ 1.0 61 ⫾ 6.3 0 ⫾ 1.5 1 ⫾ 0.5 4 ⫾ 1.6 0 ⫾ 1.7 13 ⫾ 3.9 1 ⫾ 1.2 1 ⫾ 1.7 0 ⫾ 1.3 0 ⫾ 0.9 0 ⫾ 0.4 0 ⫾ 1.7

40 ⫾ 3.4 0 ⫾ 0.8 45 ⫾ 3.0 1 ⫾ 0.6 0 ⫾ 0.5 0 ⫾ 3.1 0 ⫾ 3.1 5 ⫾ 0.8 0 ⫾ 1.7 0 ⫾ 2.0 0 ⫾ 0.5 0 ⫾ 0.6 0 ⫾ 0.8 2 ⫾ 1.0

C = common ALL; Ph, Ph-positive ALL.

Target cells: Patient’s PHA blasts Donor’s PHA blasts Patient’s EBV-LCL Donor’s EBV-LCL Patient’s fibroblasts TNF- α pretreated patient’s fibroblasts 0

20

40

60

80

100

Specific lysis (%)

Figure 7 No lysis of patient’s fibroblasts by 16C12 CTL. Cytotoxicity of 16C12 CTL against PHA blasts, EBV-LCL and fibroblasts was examined. Fibroblasts were also examined after treatment with 10 ng/ml TNF-␣ for 48 h. Allogeneic CTL (Allo CTL) were generated as described in Materials and methods, and used as controls. The E/T ratio for 16C12 CTL ( ) was 2:1, and for allogeneic CTL (䊏) was 20:1.

Discussion In the present study, we succeeded in generating CTL detecting a mHAg on hematopoietic cells by stimulating BMT donor lymphocytes with HLA-identical leukemic cells. As a responder cell population, the BM mononuclear cells from the BMT donor or the donor-derived PBL in the recipient have been used by various investigators,12–17 but in this study, we showed that the naive PBL from BMT donor can be used to generate primary in vitro CTL responses against leukemic cells in HLA-identical sibling pairs. The primary purpose of this study was to generate CTL detecting a leukemia-specific antigen and to use the CTL thus generated for the treatment of the patient with minimal residual disease and possibly overt relapse after BMT. Obviously, it is very important to generate such CTL from donor PBL before BMT. Smit et al23 also reported that leukemic-reactive CTL were generated from HLA-identical donors PBL of patients with CML using a limiting dilution

assay, although they did not analyze the mHAg possibly detected by the CTL. We assumed that donor T cells are better than the donor-derived T cells in the recipient because the former is an immunologically healthy population and that the latter may be induced to be tolerant to a leukemia antigen in the recipient under immunosuppressive treatment. It is, however, also conceivable that the latter may be sensitized by a leukemia antigen in the recipient, leading to the effective generation of CTL against leukemic cells. Further study is necessary to elucidate this interesting question. In this study, we reported the mHAg detected by HLAA31-restricted 16C12 CTL. HLA-A31, with which this mHAg was presented, is one of the A3-like alleles, and the phenotypic frequency in Japanese, Caucasian, Black, Chinese and Hispanic is 14.8%, 4.4%, 3.8%, 9.6% and 10.1%, respectively.24 So far, more than 30 human mHAgs detected by CD8 or CD4 T cells have been reported and among these, H-Y, HA-1 and HA-2 antigens have been defined with respect to their genes and antigenic peptides.25–27 At present, it is difficult to compare the 16C12 antigen in this study with other reported mHAgs, but the following evidence supports that 16C12 antigen is new: (1) no subtype of HLA-A31 has been found,28 suggesting that 16C12 antigen is not related to a subtype of HLA-A31, but a mHAg presented by HLA-A31. (2) HLA-A31 restricted mHAg has not been reported. (3) H-Y antigen should not be detectable in the combination of male donor responder cells and female recipient leukemia cells as we employed. (4) 16C12 CTL was not cytotoxic against EBV-LCL expressing HLA-A1(0/4), -A2.1(0/5), -B7(0/6) and/or B35(1/14), -B38(0/2), -B44(0/3); many of these lines probably express a variety of known mHAgs when their population frequency is high.9 (5) Maruya et al29 recently reported that the incompatibility of CD62L and CD62L + CD49b was associated with acute GVHD in Japanese patients with HLA-A3-like superfamily including HLAA31. A preliminary study by Maruya and Saji showed that patient-derived EBV-LCL and donor-derived EBV-LCL had the same polymorphisms of CD49b and CD62L, sug-


136

gesting that these adhesion molecules are not likely to be the target antigens which 16C12 CTL detected. With regard to tissue distribution, the 16C12 antigen is expressed on leukemia cells, EBV-LCL and PHA blasts, but not on fibroblasts, like many of the mHAgs so far reported.12–17 The 16C12 antigen is expressed in 4/15 HLAA31+ EBV-LCLs including those of the patient’s family members and 3/8 HLA-A31+ common ALL, suggesting the population frequency is not high, although further testing of PHA blasts from volunteers is necessary to estimate this. In addition, various types of HLA-A31+ normal and malignant hematopoietic cells, and non-hematopoietic cells need to be tested in order to determine the tissue distribution of this antigen. Since technologies are now available for defining the genes and antigenic peptides corresponding to the antigens detected by CTL, as has been reported for mHAgs and tumor-specific antigens, a similar study will be initiated with 16C12 CTL. Leukemia relapse after allogeneic BMT is still a serious problem. If we can consistently generate mHAg-specific CTL from PBL of BMT donors before BMT, adoptive immunotherapy using such CTLs to mHAg which are selectively or predominantly expressed on leukemic cells, may be useful for the treatment and protection from relapse of high risk leukemia patients with HLA-A31 after BMT.

Acknowledgements We thank Dr Elis Goulmy for sending the protocol regarding the generation of mHAg-specific CTL. We also thank Dr Yoshihiro Komada for providing the moAbs against CD3 and CD4 and Dr Hirokazu Komatsu, Dr Koji Kato, Dr Keizo Horibe, Dr Mitsune Tanimoto, Dr Shigeru Tsuchiya and Dr Shinichiro Nishimura for kindly giving HLA-A31-positive leukemic cells. We thank Mrs Etsuko Maruya for analysis of polymorphism of adhesion molecules in EBV-LCLs and Mr Dee Lynn Johnson for his advice on this manuscript. This work was supported in part by a Grantin-Aid for Scientific Research on Priority Areas, from the Ministry of Education, Science, Sports and Culture, Japan, and a BristolMyers-Squibb Biomedical Research Grant.

References 1 Weiden PL, Flournoy N, Thomas ED et al. Antileukemic effect of graft-versus-host disease in human recipients of allogeneic-marrow grafts. New Engl J Med 1979; 300: 1068– 1073. 2 Weiden PL, Sullivan KM, Flournoy N et al. Antileukemic effect of chronic graft-versus-host disease. New Engl J Med 1981; 304: 1529–1533. 3 Sullivan KM, Weiden PL, Storb R et al. Influence of acute and chronic graft-versus-host disease on relapse and survival after bone marrow transplantation from HLA-identical siblings as treatment of acute and chronic leukemia. Blood 1989; 73: 1720–1728. 4 Horowitz MM, Gale RP, Sondel PM et al. Graft-versusleukemia reactions after bone marrow transplantation. Blood 1990; 75: 555–562. 5 Butturini A, Bortin MM, Gale RP. Graft-versus-leukemia following bone marrow transplantation. Bone Marrow Transplant 1987; 2: 233–242.

¨ 6 Kolb HJ, Mittenmuller J, Clemn C et al. Donor leukocyte transfusions for treatment of recurrent chronic myelogenous leukemia in marrow transplant patients. Blood 1990; 76: 2462–2465. 7 Porter DL, Roth MS, McGarigle C et al. Induction of graftversus-host disease as immunotherapy for relapsed chronic myeloid leukemia. New Engl J Med 1994; 330: 100–106. 8 Slavin S, Naparastek E, Nagler A et al. Allogeneic cell therapy with donor peripheral blood cells and recombinant human interleukin-2 to treat leukemia relapse after allogeneic bone marrow transplantation. Blood 1996; 87: 2195–2204. 9 Goulmy E. Human minor histocompatibility antigens. Curr Opin Immunol 1996; 8: 75–81. 10 Perreault C, Roy DC, Fortin C. Immunodominant minor histocompatibility antigens: the major ones. Immunol Today 1998; 19: 69–74. 11 Simpson E, Roopenian D, Goulmy E. Much ado about minor histocompatibility antigens. Immunol Today 1998; 19: 108– 112. 12 De Bueger M, Bakker A, Van Rood JJ et al. Tissue distribution of human minor histocompatibility antigens: ubiquitous versus restricted tissue distribution indicates heterogeneity among human cytotoxic T lymphocyte-defined non-MHC antigens. J Immunol 1992; 149: 1788–1794. ¨ 13 Niederwieser D, Grassegger A, Aubock J et al. Correlation of minor histocompatibility antigen-specific cytotoxic T lymphocytes with graft-versus-host disease status and analyses of tissue distribution of their target antigens. Blood 1993; 81: 2200–2208. 14 Van der Harst D, Goulmy E, Falkenburg JHF et al. Recognition of minor histocompatibility antigens on lymphocytic and myeloid leukemic cells by cytotoxic T-cell clones. Blood 1994; 83: 1060–1066. 15 Goulmy E, Schipper R, Pool J et al. Mismatches of minor histocompatibility antigens between HLA-identical donors and recipients and the development of graft-versus-host disease after bone marrow transplantation. New Engl J Med 1996; 334: 281–285. 16 Dolstra H, Fredrix H, Preijers F et al. Recognition of a B cell leukemia-associated minor histocompatibility antigen by CTL. J Immunol 1997; 158: 560–565. 17 Warren EH, Greenberg PH, Riddell SR. Cytotoxic T-lymphocyte-defined human minor histocompatibility antigens with a restricted tissue distribution. Blood 1998; 91: 2197–2207. 18 Yazaki M, Andoh M, Ito T et al. Successful prevention of hematological relapse for a patient with Philadelphia chromosome-positive acute lymphoblastic leukemia after allogeneic bone marrow transplantation by donor leukocyte infusion. Bone Marrow Transplant 1997; 19: 393–394. 19 Riddell SR, Rabin M, Greballe AP et al. Class I MHC-restricted cytotoxic T lymphocyte recognition of cells infected with human cytomegalovirus does not require endogenous viral gene expression. J Immunol 1991; 146: 2795–2804. 20 Sato N, Okada Y, Suzuki K et al. Natural antigenic peptides of gastric signet ring cell carcinomas. In: Kikuchi K, van den Eynde B (eds). Recent Advances of Human Tumor Immunology and Immunotherapy. Gann Monogr Jap J Cancer Res: Tokyo (in press). 21 Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate–phenol–chloroform extraction. Anal Biochem 1987; 162: 156–159. 22 Akatsuka Y, Cerveny C, Hansen JA. T cell receptor clonal diversity following allogeneic marrow grafting. Hum Immunol 1996; 48: 125–134. 23 Smit WM, Rijinbeek M, van Bergen CAN et al. Generation of leukemia-reactive cytotoxic T lymphocytes from HLAidentical donors of patients with chronic myeloid leukemia


using modifications of a limiting dilution assay. Bone Marrow Transplant 1998; 21: 553–560. 24 Sidney J, Grey HM, Kubo RT, Sette A. Practical, biochemical and evolutionry implications of the discovery of HLA class I supermotifs. Immunol Today 1996; 17: 261–266. 25 Wang W, Meadows LR, den Haan JMM et al. Human HY: a male-specific histocompatibility antigen derived from the SMCY protein. Science 1995; 269: 1588–1590. 26 den Haan JMM, Sherman NE, Blokland E et al. Identification of a graft versus host disease-associated human minor histiocompatibility antigen. Science 1995; 268: 1476–1480.

27 den Haan JMM, Meadows LM, Wang W et al. The minor histocompatibility antigen HA-1: a diallelic gene with a single amino acid polymorphism. Science 1998; 279: 1054–1057. 28 Bodmer JG, Marsh SGE, Albert ED et al. Nomenclature for factors of HLA system, 1996. Tissue Antigens 1997; 49: 297–321. 29 Maruya E, Saji H, Seki S et al. Evidence that CD31, CD49b, and CD62L are immunodominant minor histocompatibility antigens in HLA identical sibling bone marrow transplants. Blood 1998; 92: 2169–2176.

137