CD 1-reactive leukemic cells in bone marrow: Presence of Langerhans cell marker on leukemic monocytic cells Misery L, Campos L, Dezutter-Dambuyant C, Guyotat D, Treille D, Schmitt D, Thivolet J. CD1-reactive leukemic cells in bone marrow: Presence of Langerhans cell marker on leukemic monocytic cells. Eur J Haematol 1992: 48: 27-32.
Laurent Misery', Lydia Campos', Colette Dezutter-Dambuyant Denis Guyotat3, Danielle Treille', Daniel Schmitt' and Jean Thivolet
',
'
' Laboratoire de Recherche Dermatologiqueet
I
Abstract: Langerhans cells originate in bone marrow and probably belong to the monocyte-macrophagelineage. CDl is a specific marker of Langerhans cells. By immunofluorescence and immunoelectron microscopy, CDla antigen and myeloid markers (CD11, CD13, CD14, CD15, CD33, HLA-DR) were studied in 53 cases of acute myeloid leukemias (AML) and 3 acute lymphoblastic leukemias (ALL). The 11 ANLL without monocytic component were CDla negative. 2/5 of acute myelomonocytic leukemias (AML4) and 9/37 of acute monocytic leukemias (AMLS) were positive. All 3 ALL were negative. No correlation was found between CDla and myeloid markers. CDla' AML did not differ from CDla- AML with regard to cytogenetics or response to therapy. The CDla positive cells may originate from an abnormal proliferation of CDla positive cells which are present in bone marrow and which may differentiate into Langerhans cell precursors.
Introduction
Langerhans cells (LC) are present in the skin and in the mucous membranes. Their main function is antigen presentation to lymphocytes (1, 2, 3). These cells belong to the mononuclear phagocyte system (4). They originate in the bone marrow and probably have common precursors with monocytic lineage. After allogeneic bone marrow transplantation, the skin is colonized by the donor's Langerhans cells in a few months (6,7). On the other hand, there are CDla' cells among monocytic progenitors (1/1000) (8,9, 10). Lastly, among blood monocytes, there are CDla cells (1 1, 12) and LC have common receptors with monocytes (13). Certain arguments lead us to think that there is a gradual transformation of monocytes into Langerhans cells in the epidermis, with the aspect of Birbeck granules (14). CDla marker is specific to LC in the epidermis (15). Elsewhere, it is only present on putative precursors of LC, on cortical thymocytes (16) and in a small number of chronic lymphoid leukemias (17). We have studied the expression of CDla antigen on acute non-lymphoblastic leukemic cells obtained from bone marrow samples.
Immunologie, Unite lnserm 209, Hopital Edouard Herriot (Pavillon R), Laboratoire de Cryobiologie, Centre Regional de Transfusion Sanguine, Lyon, Departement d'Hematologie, Hopital Nord, Saint-Priest en Jarez, France
Key words: acute myeloid leukemia - monocyte
- Langerhans cell - differentiation antigen CD 1 Correspondence: Laurent Misery, Cryobiology, CRTS, 1-3 rue du Vercors, BP 7015, 69342 Lyon Cedex 0 7 , France Accepted for publication 7 August 1991
Among 53 samples, 11 have more than 20% of CDla' cells. These leukemias were of the monocytic subtypes (AML4 or AML5). Material and methods Patients
56 patients with newly diagnosed acute leukemias were entered in this study (53 with AML and 3 with ALL). The median age of the patients was 55.5 years (range 21-83). The median number of leukocytes was 65.5.109/1. Cytological study
The 53 AML were classified as: 4 acute myeloblastic leukemia (AMLl), 4 AML2, 3 AML3, 5 AML4, 37 AML5. 3 patients with CALL were used as negative control. The diagnosis was made in all cases by a study of bone marrow smears stained by May-GrunwaldGiemsa. The cytochemical study included stainings for the following activities: myeloperoxidase, chloroacetate esterase and alpha naphtyl acetate esterase with and without fluoride inhibition. The FAB subtype was determined by standard criteria (18, 19). 21
Misery et al. Immunological study
lmmunogold labeling
Mononuclear cells were recovered from heparinized bone marrow by sedimentation on Ficoll-Hypaque, washed three times and resuspended in phosphatebuffered-saline (PB s).The final suspension always contained more than 80% blast cells, assessed by cytocentrifbge smears. The cells were cryopreserved after addition of fetal calf serum (final concentration 20%) and dimethylsulfoxide (lo%), and the immunological study was performed after quick thawing. Surfacemarkerswere analyzedby indirect immunofluorescence using the following antibodies: M 0 1 (IgG1, Coulter), directed to C D l l b recognized C3bi receptor and mainly stains monocytes and granulocytes. My7 (IgG1, Coulter), directed to CD13, recognizes AML cells with granulocytic and monocytic differentiation and normal precursor cells. CRIS 6 (IgG1, provided by Dr Vilella, Hospital Clinic, Barcelona, Spain) is directed to CD14, a monocytic antigen. 8.27 (IgM, Biosys, France) is directed to CD15 and recognizes leukemic cells with granulocytic differentiation and the normal granulocytic lineage. My9 (IgG2b, Coulter) is directed to CD33 and recognizes myeloid progenitors, and then monocytic progenitors. BL2 (IgG2b, provided by J. Brochier, Uniti: INSERM 80, Lyon, France), is directed to HLADR. T6 (IgG2b, Coulter) is directed to C Dl a and recognizes cortical thymocytes and Langerhans cells. (Third International Workshop on Leucocyte Differentiation Antigens, Oxford, 1986).
Bone marrow cells were harvested and washed in 5 % FCS-5 % normal human serum-RPMI 1640, incubated with a primary specific monoclonal antibody mouse IgGl and revealed by an antimouse goat IgG conjugated with 15 nm gold particles (Goat anti-mouse IgG 1 15 nm, Cambridge Research Biochemicals) applied at a dilution of 1 5 for 1 h at 4°C. After washing, the cells were fixed with 2% glutaraldehyde in cacodylate buffer, then post-fixed in osmium tetraoxide and embedded in epoxy medium. Ultrathin sections were examined, after post-staining with lead citrate and uranyl acetate, using a Jeol 1200 EX electron microscope. All the solutions were filtered through Millipore filters (0.22 pm mesh). As negative isotype control, labelings with the mouse myeloma Igs from clarified ascites (MOPC2 1< IgG 1> and MOPC 141< IgG2b > ) for each monoclonal antibody were performed. This study concerns 6 samples.
Cytofluorometric analysis
Rabbit serum (Biosys, France) or human AB serum was first added to the cell suspension to avoid nonspecific fixation of the monoclonal antibodies on Fc receptors. The cells were then incubated with the monoclonal antibodies (mAb) at 4° C for 30 min and washed with serum-containingmedium. Fluoresceinlabelled goat anti-mouse F(ab')2 fragments (Bioart, France) were added as second layer (4"C, 10 min). Flow cytometry was performed on a cytofluorograph FACSCAN. The evaluation of percentages of labeled cells was done by setting a threshold on the histogram by comparison with background fluorescence of cells treated with normal mouse serum. Determinations were based on 10000 cells per sample.
Statistical analysis
The relationship of the expression of CDl a antigen was compared to the other myeloid antigens by Chisquare test. Mean numbers were compared by Mann-Whitney test.
Results Correlations with initial characteristics, morphology and response to therapy (Table 1)
The median age of the patients was 55.5 years (range 21-83) versus 62.3 yr (range 43-83) for the patients with CDl a + bone marrow samples. The mean count of leukocytes was 92.109/l with CD1' bone marrows and 57.7.109/1 with CD1bone marrows and 65.5.109/1with all bone marrows. However, this difference is not significant. The karyotypes of CD1+ and CD1- leukemias were analyzed. Among the CD1 -t group, only 9/ 11 were available. 3 had cytogenetic abnormalities, 2 with inv(l6) corresponding to M4 eosinophilia phenotype and 1 with inv(9). The others were normal. Among the CD1- group, 34/42 karyotypes were performed and 2 had t(8; 21), 2 had t(15; 17), 6 had a polyploidody 1 had del(7) and 1 del(5). The others were normal. In the CD1' group, 4/11 patients (36%) were in complete remission after chemotherapy and 7/11 had no remission. In the CD1- group, 23/42 patients (54.8%) had a complete remission and 19 patients didn't have remission. No correlation was found between the presence of CD1 antigen on cells and the response to therapy.
CD1-reactive leukemic cells in bone marrow Table 1. Correlations of CD1' cases with initial characteristics, morphology and response to therapy Mean leucocyte count (10~11)
FA8
CDl (1)
Cutaneous infiltration
1 2
M5 M5
40 35
Yes Yes
160 99
3 4 5
M5 M5 M4
70 80 20
No No No
77 120 52
M5 M5 M5 M4 M5 M4
60 66 35 60 74 70
No No NO No No Yes
45 9 150 160 70 102
Patient
Remission ("1
Karyotype
ND 46, XY, inv (911 46, XY ND 46, XX 46, XYl 46, XY, inv (16)/ 48, XY, inv (161, t8, t 2 1 46, XX 46, XY 46, XY 46, XY 46, XX 46, XX del (9) (pll-p22) inv ( 16) (p 13-q221
Survival (*)
-
t
t
+
+
(*)t is for complete remission or survival. (1 1 Percentage of stained cells.
CDla antigen was not found in the ALL, AML1, AML2, AML3. It was observed in 2/5 AML4 (40%) and in 9/37 AMLS (24.3%), i.e. altogether 11 times in 42 cases (26%). Among CD1- bone marrows, C D l l was observed in 012 AML4 and in 13/29 AMLS. CD13 was observed in 1/2 AML4 and in 22/29 AMLS. CD14 in 1/2 and 16/29. For CD15, there were 1/2 AML4 and 21/29 AMLS. For CD33, 0/2 AML4 and 13/ 29AML5 had more than 20% positive cells. The
myeloid markers were not found in ALL. HLA-DR was present in 1/2 AML4 and 22/29 AMLS. Among CDla' bone marrows (Table2), the CD11' samples were 1/3 in AML4 and 4/8 in AMLS, the CD13' samples were 3/3 in AML4 and 7/8 in AMLS, the CD14' samples were 2/3 and 4/8, the CD15' were 3/3 and 7/8, the CD33' were 3/3 and 3/8 and the HLA-DR+ were 3/3 and 7/8. The most common phenotypes were CD13' (10/11 cases) and CD 15' (lo/ 11 cases) and especially CD13+/HLA-DR+ (9/11 cases) and CD13+/ CD15+ (Sill cases). The T-cell related antigens (CD2, CD3, CD5, CD7) were not present. Double stainings were performed in 11 cases of CDla' monocytic cases. We found that these antigens were expressed by the same cells (Table3): 6/11 were CDl'/CD14+ (the others were CD14-) and 5/11 were CDl'/CD33' (the others were CD33-), 9/11 were CDl'/CD13' (the others
Table 2. Immunologicalmarkers of CD1' cases (percentage of stained cells)
Table 3. Double stainings (percentage of positive cells)
In the CD1' group, 3/11 (27.3%) were alive after 2 yr versus 12/37 (32.4%) in the CD1- group. No difference was found in skin involvement between the two groups. 3/11 (27.3%) patients had cutaneous infiltration in the CD 1 group versus 10/42 (23.8%) in the CD1- group. +
Immunological markers
Patient
FAB
CD1
CDll
CD13
CD14
CD15
CD33
HLA-DR
1 2 3 4 5 6 7 8 9 10 11
M5 M5 M5 M5 M4 M5 M5 M5 M4 M5 M4
40 35 70 80 20 60 66 35 60 74 70
20 29 17 40 1 5 18 16 25 25 7
50 75 6 28 34 64 36 74 85 90 85
8 80 38 58 2 1 14 1 50 90 40
27 1 39 28 26 64 72 75 65 50 50
1 73 10 67 32 9 1 1 5 90 25
35 3 43 87 57 70 50 49 90 40 84
Patient
TAB
CD1
CDl/CD13
CDl/CD14
CDl/CD33
CDl/HLA-DR
1 2 3 4 5 6 7 8 9 10 11
M5 M5 M5 M5 M4 M5 M5 M5 M4 M5 M4
40 35 70 80 20 60 66 35 74 60 70
42 34 6 28 18 62 37 35 74 60 69
8 34 38 58 2 2 14 15 74 50 40
1 31 10 67 20 9 1 1 74 6 25
35 2 40 80 20 58 49 32 39 60 70
29
Misery et al.
30
CD1-reactive leukemic cells in bone marrow negative) and lo/ 11 were CD 1'/HLA-DR' (the other were negative). There was no significant correlation between CD 1 and myeloid markers nor between C D l a and HLADR. lmmunogold labeling
In immunoelectron microscopy, the cells presented a promonocytic-monocytic morphology and were as small as monocytes. The nucleus was folded with a large amount of heterochromatin. The nuclear contour showed considerable variation, from slight irregularity to marked indentation. The cytoplasm was large, clear and contained numerous mitochondria, filaments and sparse vacuoles. The cytoplasmic membrane was rather villous. No typical Birbeck granule was observed. Some rare structures of the smooth endoplasmic reticulum could have suggested Birbeck granule-like structure (Fig. 1). CD1' cells were identified by the presence of variable numbers of electron-dense gold (15 nm) particles scattered along the cell surface. No gold particles were observed in the cytoplasm of these cells. CD1+ cells showed a range of density of labeling suggesting a variable expression of membrane antigens (5-19 gold granules/cell section). Discussion
was found between C D l a antigen and monocytic lineage. The association with hyperleukocytosis can probably be explained by these FAB subtypes because the monocytic leukemias (AML4 and 5 ) have more hyperleukocytosis than the other leukemias. The same is no doubt also true for older age of patients CD1'. In this group, 36% of patients obtained complete remission versus 54.8% in the C D 1 group. ~ The most common phenotypes were CD 13, CD 15 and especially CD 13/CD15 and CD 13/HLA-DR. This is normal because monoblastic leukemias are stained with these myeloid markers. C D l a antigen was found before on AML5 cells in the skin, but not in the blood or in the bone marrow (23). In a study of 26 cases of AML5 and AML4, Schwonzen et al. did not find this antigen on blood cells or bone marrow cells (24). This negative result could be due to a smaller number of studied cases (26 versus 42) and to a difference of sensitivity in the procedures used. Our results were confirmed by immunoelectron microscopy and showed the weak density of CD1 antigens on the cellular membrane. Three hypotheses may be put forward to explain the C D l a expression on these leukemic cells: - The leukemic transformation of the cells can deregulate the expression of CDlagene, which is not normally expressed, - CD 1 leukemias may be the leukemic counterpart of the normal LC differentiation. - Leukemic blasts are at an early phase of differentiation common to both monocytes and LC. CD1' cells have been described in granulomonocytic marrow colonies (8). Recently, specific dendritic LC colonies have been observed (1 1). These findings are consistent with the fact that LC probably originate in bone marrow and have common precursors with monocytes. Such a hypothesis would explain the case of these associations between monoblastic leukemia and histiocytosis X (25,26). +
In this study, we searched for the presence of CD l a antigen on 53 acute myeloid leukemia patients. Myeloid markers (CD11, CD13, CD14, CD15, CD33, HLA-DR) observed on bone marrow cells confirm that these cells are from AML4 and AML5 (20, 21), according to cytological criteria of FAB group (18, 19). Our study showed a significant number of samples with more than 20% CDla' cells: 3 1% of AML4 and AML5 samples. A non-specific binding was blocked by saturating monoblastic receptors to Ig Fc fragment with human AB serum (22). The presence of C D l a antigens was confirmed by two procedures: cytofluorometry and electron microscopy. There was no CD l a antigen on cells from non-monocytic leukemias and the monocytic specificity was confirmed by double staining. CDla antigen was found on 11 bone marrow samples composed of 8 AMLS and 3 AML4. There was no association between specific clinical manifestations (and especially cutaneous manifestations) and the presence of CDla, but a direct relationship
References 1. SCHMITT D, DEZUTTER-DAMBUYANT C, STAQUET MJ, THIVOLET J. La cellule de Langerhans, cellule dendritique de I'Bpiderme et des muqueuses. Med/Sci 1989: 5: 103-111. 2. BREATNACH SM. The Langerhans cells. Br J Dermatol 1988: 119: 463469. 3. KATZSI, COOPER KD, IIJIMA M, TSUCHIDA T. The role of Langerhans cells in antigen presentation. J Invest Dermato1 1985: 85: 96 s-98 s. ~
~~
Fig. 1. Ultrastructural morphology of a CDla-positive cell isolated from a human leukemic bone marrow sample. a) This labeled cell presents a typical appearance of a pro-monocytic cell with a lobulated nucleus and a large cytoplasm containing many mitochondria, a few small electron-dense and large clear vacuoles. Bar: 1 pm. b, c, d, e, f) High power views of this labeled cell (a), a few gold granules are regularly bound along the membrane. This cell contains some rare Birbeck granule-like structures localized in the cytoplasm near the plasma membrane (black stars). Bar: 200 nm.
31
Misery et al. 4. VANFURTHR. Monocytes. Their origin and density. Transplantation and clinical immunology. Lyon: 1979. Amsterdam: Experta Medica, 1979: 247-254. 5. KATZSI, TOMASKI KI, SACHSDH. Epidermal Langerhans cells are derived from cells which originate in bone marrow. Nature 1979: 282: 324-326. 6. FRELINGER JG, HOODL, HILLS, FRELINGER JA. Mouse epidermal Ia molecules have a bone marrow origin. Nature 1979: 282: 321-323. 7. PERREAULT C, PELLETIER M, LANDRYD, GYGERM. Study of Langerhans cells after allogeneic bone marrow transplantation. Blood 1984: 63: 807-811. 8. DE FRAISSINETTEA, SCHMITT D, DEZUTTERDAMBUYANT C, GUYOTAT D, ZABOTMT, THIVOLET J. Culture of putative Langerhans cells bone marrow precursors: characterization oftheir phenotype. Exp Haematoll988: 16: 764-768. MJ. Origin and precursors of Langerhans cells. In: 9. STAQUET THIVOLET J, SCHMITT D, eds. The Langerhans Cells. John Libbey/INSERM, 1988: 172: 2-3. 10. ORTALDO JR, SHARROW SO, TIMONEN T, HEVERMANN RB. Determination of surface antigens on highly purified human NK cells by flow cytometry with monoclonal antibodies. J Immunol 1981: 127: 2401-2409. 11. REIDCDL, FRYERPR, CLIFFORD C, KIRKA, TIKERPAE J, KNIGHTSC. Identification of hematopoietic progenitors of macrophages and dendritic Langerhans cells (DL-CFU) in human bone marrow and peripheral blood. Blood 1990: 76: 1139-1 149. 12. GOTHELF Y, HANAUD, TSURH, et al. T6 positive cells in the peripheral blood of burn patients: are they Langerhans cells precursors? J Invest Dermatol 1988: 90: 142-148. 13. DE FRAISSINETTE A, DEZUTTER-DAMBUYANT C, STAQUET MJ, SCHMITT D. Common surface molecules (CD14 and CD33) shared by monocytes and normal human epidermal Langerhans cells. In: 4th International Conference on Human Leukocyte Differentiation Antigens. Vienna, February 21-25. Oxford University Press. 14. MURPHY GF, MESSADI D, FONFERKOE, HANCOCK WW. Phenotypic transformation of macrophages to Langerhans cells in the skin. Am J Pathol 1986: 123: 401-406. 15. FITHIANE. Reactivity of Langerhans cells with hybridoma antibody. Proc Natl Acad Sci USA 1981: 78: 2541-2544. 16. MAC MICHAEL AJ, PILCHJR, GALFREG, MASONDY,
32
FABREJW, MILSTEIN C. A human thymocyte antigen definied by hybrid myeloma monoclonal antibody. Eur J Immuno1 1979: 9: 205-210. 17. REINHERZ EC, KUNGPC, GOLDSTEIN G, LEVEYRH, SCHLOSSMAN SF. Discrete stages of intrathymic differentiation: analysis of normal thymocytes and leukemic lymphoblasts of T lineage. Proc Natl Acad Sci USA 1980: 77: 15881592. 18. BENNET JM, CATOWSKY D, DANIELMT, et al. Proposals for the classification of acute leukzmias: French-AmericanBritish (FAB) cooperative group Br J Hemato1 1976: 33: 451-455. 19. BENNETT JM, CATOWSKY D, DANIELMT, et al. Proposed revised criteria for the classification of acute myeloid leukemias: a report of the French-American-British cooperative group. Ann Intern Med 1985: 103: 620. 20. CAMPOSL, GUYOTAT D, ARCHIMBAUD E, et al. Surface marker expression in adult acute myeloid leukzmia: correlations with initial characteristics, morphology and response to therapy. Br J Hzmatol 1989: 72: 161-166. 21. DREXLER HG. Classification of acute myeloid leukemia - a comparison of FAB and immunophenotyping. Leukemia 1987: 1: 697-705. 22. LAWLORE, FINN T, BLANEYC, TEMPERLEY IJ. MAC CANNSR. Monocytic differentiation in acute myeloid leukaemia: a cause of Fc binding of monoclonal antibodies. Br J Hzmatol 1986: 64: 339. 23. GIANOTTIR, BERTI E, ALESSIE, FINAL, CECCAE, CAPUTOR. Myelomonocytic leukemia with cutaneous leJ, sions expressing C D l a and SlOO protein. In: THIVOLET SCHMITTD, eds. The Langerhans Cell. John Libbey/ INSERM, 1988: 172: 473. 24. SHWONZEN M, KUEHNN, VETTENB, DIEHLV, PFREuNDSCHUH M. Phenotyping of acute myelomonocytic (AMMOL) and monocytic leukemia (AMOL): association of T-cell-related-antigens and skin infiltration. leukemia Res 1989: 10: 893-898. 25. CLAUDY AL, LARBREB, COLOMB M, LEVIGNEV, DEVILLE V. Letterer-Siwe disease and subacute monocytic leukemia. J Am Acad Dermatol 1989: 21: 1105-1 106. F, MISERYL, KANITAKIS J, ARCHIMBAUD 26. CAMBAZARD D, HERMIERC, THIVOLET J. Acute monocytic leukemia and histiocytosis X. Cancer (submitted).