Expression of coagulation factor XIII subunit A in acute promyelocytic leukemia


  • How to cite this article: Simon Á, Bagoly Z, Hevessy Z, Csáthy L, Katona E, Vereb G, Ujfalusi A, Szerafin L, Muszbek L, Kappelmayer J. Expression of coagulation factor XIII subunit A in acute promyelocytic leukemia. Cytometry Part B 2012; 82B: 209–216.


Leukemic cells often express markers, which are not characteristic of their particular cell lineage. In this study, we identified the “A” subunit of coagulation factor XIII (FXIII-A) in leukemic promyelocytes in de novo AML M3 cases. The cytoplasmic presence of factor XIII-A has previously been shown only in platelets/megakaryocytes and monocytes/macrophages. Furthermore, more recently we described the presence of FXIII-A in leukemic lymphoblasts.

We studied 14 patients with this rare type of acute leukemia in a period of 4 years and investigated their bone marrow samples by 3-color flow cytometry upon diagnosis, mainly focusing on FXIII-A expression of leukemic cells. We detected FXIII-A also by ELISA, Western-blot, and confocal laser scanning microscopy.

This was a homogenous group of AML M3 patients with translocation t(15;17)(q22;q21) detected by fluorescence in situ hybridization (FISH). In 10 out of 14 samples, FXIII-A was detectable by flow cytometry and was coexpressed with markers characteristic for leukemic promyleocytes (CD45dim/CD13+/CD33+/CD117+/cyMPO+ and HLA-DR-/CD34−/CD14−/CD15−). Staining for the markers GPIIb and GPIX were negative, and FXIII-A was identified in the cytoplasm of the cells by confocal microscopy in a relatively high quantity, as measured by ELISA. By Western blot analysis we could identify FXIII-A in the native 82 kDa form and in cleaved forms corresponding to cleavage products observed when purified FXIII-A was treated by human neutrophil elastase.

This novel expression site of FXIII-A in AML M3 can be considered as a leukemia associated immunophenotype and may have pathophysiological significance. © 2012 International Clinical Cytometry Society


Factor XIII (FXIII) of blood coagulation, a protransglutaminase, becomes activated by the proteolytic action of thrombin in the presence of calcium. The activated form of the enzyme is responsible for crosslinking fibrin strands, thus stabilizing the clot in the final stage of the coagulation process [1, 2]. It is present in two forms in the human body; one of that circulates in the plasma as a heterotetramer comprising of two potentially active A and two carrier/inhibitory B subunits (A2B2). The other one is an intracellular homodimer made up by two A subunits (A2) [3].

Intracellular FXIII-A has first been described in platelets and megakaryocytes [4, 5] and later in monocytes and macrophages [6–8]. Platelets contain huge amounts of FXIII-A, 150-fold more per volume than plasma, while FXIII-A concentration in monocytes is at least one magnitude less than that in platelets [9]. It has been demonstrated that normal bone marrow precursor cells of monocytes and megakaryocytes also express FXIII-A [10]. The function of the intracellular enzyme has not yet been elucidated.

We and others have described that FXIII-A was a sensitive intracellular marker for the monocytic and megakaryocytic series and a useful diagnostic tool in AML immunophenotyping [11–13].

Acute Promyelocytic Leukemia (APL) is the M3 type of acute myeloid leukemia characterized by an accumulation of abnormal promyelocytes in the bone marrow, a severe bleeding tendency and the presence of the chromosomal translocation t(15;17). The immunophenotype of leukemic promyelocytes has been well characterized over the past and they usually coexpress the myeloperoxidase (MPO), CD9, CD117, CD13, and CD33 markers in the absence of reactivity for HLA-DR, CD34, and CD15 [14–16]. In this study, we investigated only de novo APL cases. We wanted to explore the presence of FXIII-A in leukemic promyelocytes. In addition, we did some initial evaluation on FXIII-A expression with survival data. We established, that FXIII-A is not present in normal promyelocytes, but it can be used as a diagnostic marker in APL and its presence may identify a favorable prognostic subgroup within this cytogenetically and morphologically homogenous APL group.


Patient Specimens

Bone marrow and peripheral blood of 14 newly diagnosed APL patients have been analyzed by three-color flow cytometry and bone marrow samples were anticoagulated with EDTA and heparin. The bone marrow smears contained >70% blast cells. All APL cases were classified as hypergranular form and all were positive for t(15;17) observed by FISH. Ten patients were male and 4 were female with a mean age of 48 years; the range was from 25 to 68 years. All cases were studied upon diagnosis before any treatment was initiated.

Flow Cytometry Immunophenotyping Studies

Generation and labeling of mouse monoclonal antibody against FXIII subunits was carried out as previously described [12] utilizing a fluorescein isothiocyanate (FITC) labeling kit (Sigma, St. Louis, MO). The other monoclonal antibodies used in the three-color panels were the following (PE = phycoerythrin, PE/Cy5[Cyanine5] fluorochrome tandem, PerCP = peridin chlorophyll protein): CD4-FITC, CD8-PE, CD13-PECy5, CD34-PE, CD34-PerCP, CD19-PECy5, CD14-FITC, CD33-PE, HLA-DR-PerCP, glycophorin-FITC, CD45-PerCP, CD45-FITC, CD15-FITC, CD117-PE, CD135-PE, CD7-FITC (all purchased from Becton Dickinson Biosciences San Jose, CA) and CD13-FITC, CD41-PE, MPO-PE (all purchased from Dako Glostrup, Denmark), the Intrastain permeabilizing kit was the product of Dako.

Surface staining of whole blood and bone marrow cells were carried out according to standard procedures. Fifty microliters of whole blood or bone marrow sample adjusted to leukocyte count of 10 G/L by phosphate buffered saline (PBS) was incubated by saturating concentrations of directly conjugated antibodies for 15 min at room temperature in the dark with antibodies against different cell surface epitopes. Red cells were lysed by FACS lysing solution (Becton Dickinson Biosciences San Jose, CA) and samples were washed (300g, 5 min) in PBS and finally resuspended in PBS containing 1% paraformaldehyde (PFA). For intracytoplasmic staining, the procedure described for Intrastain was strictly followed. Surface staining was executed before permeabilization and intracytoplasmic staining. FXIII-A antibody was used at a 2 μg/ml final concentration with appropriately matched isotype control. PFA fixed samples were kept at 4°C for maximum 24 h. Flow cytometric measurement was performed on a FACSCalibur flow cytometer (Becton Dickinson Biosciences San Jose, CA) using the same setting for all investigated samples. Data obtained on 10,000 cells in de novo leukemias were stored in list-mode data files and analyzed by Cell Quest 3.2 software.

FXIII-A expression was tested on normal bone marrow promyelocytes, therefore four patients with iron deficiency were selected and their samples were stained with FXIII-FITC, CD117-PE, HLA-DR-PerCp-Cy5.5, CD33-PE-Cy7, CD15-APC, CD14-APC-H7, CD45-Pacific Orange combination according to the previously described protocol. Normal samples were measured in a FACSCanto II flow cytometer equipped with three lasers (Becton Dickinson Biosciences San Jose, CA).

Cytogenetic and FISH Analysis

Conventional cytogenetic analysis was performed on bone marrow samples cultured for 24 h, prepared using standard procedures. For each patient 20 G-banded metaphases were analyzed. The karyotypes were described according to the International System of Human Cytogenetic Nomenclature (ISCN, 2009). Fluorescence in situ hybridization was carried out on cell suspension originated from chromosome preparation according to the manufacturer's instructions using PML-RARA dual color, single fusion translocation probe (Abbot/Vysis, Downers Grove, IL). Cells were counterstained with DAPI (4,6-diamidino-2-phenylindole). In general, 200 interphase cells were counted in each case. The images were captured by Zeiss Axioplan2 (Carl Zeiss, Zaventem, Brussels) fluorescence microscope and analyzed by ISIS software (Metasystems, Altlussheim, Germany).


Detection of FXIII-A in cell lysates from APL samples was carried out as described earlier, by Katona et al. [9] with slight modification. To avoid platelet contamination, promyelocytic cells were washed three times in PBS containing 20 mM of EDTA at 1,100g for 4 min. To inhibit serine and cysteine proteases a protease inhibitor cocktail (Roche Applied Science, Penzberg, Germany) was added to the washing buffer. The exact cell count was measured before sonication in order to calculate the amount of FXIII-A/cell. Solubilization of cells was carried out with sonication at 4°C for 3 × 30 s in PBS.


After determining the FXIII-A content of platelet-free APL blast cells, remaining cells were centrifuged and dissolved in 100 μL of SDS PAGE sample buffer (62.5 mM Tris-HCl, 2% SDS, 10% glycerol, 0.1% bromphenol blue, 4.5% mercaptoethanol amine). Denatured cell suspensions were boiled for 5 min. Samples (containing 18 ng FXIII-A/well, adjusted to 60 μL with sample buffer) were loaded onto 7.5% SDS polyacrylamide gel and electrophoresed under reducing conditions. Western blotting was carried out using Immobilon P membrane (Millipore, Bedford, MA). Nonspecific binding of membranes was blocked using Tris-buffered saline (0.5M NaCl, 20 mM Tris-HCl, pH 7.5; TBS) containing 3% gelatin at room temperature and an overnight incubation at 4°C in the blocking buffer containing 1% of gelatin. Sheep polyclonal anti-human FXIII-A antibody was used as primary antibody (Affinity Biologicals, Ancaster, Canada). The immunoreaction was developed by biotinylated rabbit anti-sheep IgG and avidin-biotinylated peroxidase complex (components of Vectastain ABC kit, Vector, Burlingame, CA) and visualized by enhanced chemiluminescence (ECL Plus+, Amersham, Little Chalfont, UK) according to the manufacturer's instructions. Granulocytes and their cell lysates were used as negative controls, not showing any detectable FXIII-A antigen as investigated by flow cytometry, ELISA or Western blot (data not shown).

Samples demonstrating FXIII-A cleavage by human neutrophil elastase (HNE) were processed by incubating 2 μg/mL purified FXIII-A2 with 2 μg/mL purified HNE in the presence of 2.85 mM Ca2+ for various intervals at 37°C. Reaction was stopped by the addition of equal volume of SDS Laemmli buffer. FXIII-A2 was prepared from human placenta as described previously [17], highly purified HNE (20–22 units/mg) was purchased from Calbiochem (La Jolla, CA). Platelet-free promyelocyte cell lysates purified from peripheral blood of two patients and from the bone marrow of Patient B were denatured; SDS–PAGE was carried out in reducing conditions loading equal amounts (18 ng) of purified and cell lysate FXIII-A2 per lanes.

Confocal Laser Scanning Microscopy (CLSM)

Cytospin preparations were thawed, fixed in 4% PFA for 10 min and washed three times in PBS. FITC-conjugated anti-FXIII-A monoclonal antibody was dissolved at 15 μg/ml in PBS containing 1 mg/ml BSA (Sigma, Schnelldorf, Germany) and 0.1% Triton X-100 and added to the cells for 30 min at room temperature. During the last 5 min of incubation propidium iodide (PI) was added to the labeling solution at 0.5 μg/ml final concentration. Next, cytospins were washed three times with PBS containing 1 mg/ml BSA and 0.05% Triton X-100. Finally, the samples were washed again in PBS, and mounted in 10 μl Mowiol [0.1M Tris-HCl, pH 8.5, 25 (w/v%) glycerol and 10% Mowiol 4-88, Hoechst Pharmaceuticals, Frankfurt, Germany].

For CLSM, a Zeiss (Göttingen, Germany) LSM 510 systems and a CApochromat 63×/1.25 NA water immersion objective were used. Fluorescein was excited with a 488 nm Ar ion laser and detected through a 505–550 nm band pass filter. PI was excited with a 543 nm HeNe laser and detected through a 560 nm long pass filter. Pinholes were set to obtain 1 μm optical slices and 512 × 521 pixel images were taken with pixel times of 6.4 μs, and 2× line-averaging. All images were obtained in multitrack mode to avoid crosstalk between channels.

Statistical Analysis

Statistical comparison of the FXIII positive and FXIII negative population survival was performed by GraphPad Prism 4.0 software.


Analysis of FXIII-A Expression by Flow Cytometry

Leukemic promyelocytes are large cells that possess numerous granules, thus are characterized with a high side-scatter and enhanced autofluorescence. These characteristics were also observed in samples from our patients (data not shown). In the case of these APL samples staining for myeloid markers (MPO, CD13, CD33, CD14, CD15), blast markers (CD34, CD117), HLA-DR, and FXIII-A expressions were studied. The CD45dim MPO+ and CD33+ leukemic promyelocytes expressed cytoplasmic FXIII-A (Fig. 1). In addition, stainings for GPIIb (CD41) and GPIX (CD42a) platelet markers were studied in selected cases and these promyelocytes did not show any CD41 and CD42a positivity (Fig. 2). In clinical samples 30% value was defined for positivity as a cut off for all markers including FXIII-A expression.

Figure 1.

Representative dot plots of an APL sample. In the peripheral blood sample most cells are abnormal promyelocytes displaying intense CD33 and MPO labeling, CD34 and CD15 negativity and dim CD117 positivity (A–D). The promyelocytes contain FXIII-A (panels E and F). APL samples were labeled by FITC conjugated monoclonal anti FXIII-A antibody.

Figure 2.

To rule out potential platelet contamination CD41 and CD42a platelet markers were investigated. In the APL samples the cells show intracytoplasmic FXIII-A positivity, but based on the lack of CD41 and CD42a, the FXIII-A positive signal is not derived from platelet or platelet microparticle contamination (circles with broken line).

All the 14 APL cases were positive for MPO, and CD33, the mean percent positivity were 83% (41–99%) and 90% (74–99%), respectively while CD13 positivity mean value was somewhat lower (59%). All but one sample were positive for CD117, mean of positivity was 63% (40–96%). One of the 14 APL cases showed expression of CD15 marker, in 13 cases CD15 was absent (mean: 8%). No APL cases expressed HLA-DR and CD34 (Fig. 3A), but 10 out of the 14 APL cases had FXIII-A staining that exceeded the 30% limit (Fig. 3B).

Figure 3.

Expression of surface and intracellular antigens in APL cells. Mean percent positivity and standard errors of means are shown (n = 13, mean values ± standard deviations are displayed). CD14, CD15, CD34, and HLA-DR markers are consistently negative while CD13, CD33, CD117, and cytoplasmic MPO markers are positive (Fig. 3A). FXIII-A was positive in 9 and negative in 4 cases (Fig. 3B).

Detection of FXIII-A in APL Cells by CLSM

To investigate the intracellular localization of FXIII-A in leukemic promyelocytes, 3 FXIII-A positive APL samples were further analyzed by CLSM. APL cells were analyzed on cytospin preparations (Fig. 4). In leukemic promyelocytes, FXIII-A was detected in the cytoplasm of the cells. On the overview pictures (Figs. 4A and 4D) it is evident that a large proportion of the investigated cells were FXIII-A positive. These samples were negative for the platelet markers GPIIb and GPIX. To rule out technical artifacts, blasts were also investigated in an indirect labeling system with the same polyclonal anti-human FXIII-A as used for immunoblotting. The staining characteristics were similar to that previously observed for leukemic lymphoblasts [18].

Figure 4.

Detection of FXIII-A protein by CLSM. FXIII-A(FITC) appears in green, nuclei are stained with PI in red. In the bone marrow (A) and peripheral blood sample (D) numerous FXIII-A positive promyelocytes can be seen. The enlarged pictures (B, E) and 3D images (C, F) show cytoplasmic localization of FXIII-A. [Color figure can be viewed in the online issue, which is available at]

Detection of FXIII-A by ELISA and Immunoblotting

Western blot analysis with a highly sensitive chemiluminescent developing system was utilized to detect FXIII-A antigen in blast cells. A single FXIII-A positive band at 82 kDa was detected in APL blasts that comigrated with FXIII-A in the positive control (platelet lysate) and with the FXIII-A isolated from human placenta macrophages. The negative control cells—isolated human neutrophil granulocytes—contained no FXIII-A protein (data not shown).

Bands, which represent FXIII-A and cleaved products of FXIII-A in promyelocyte cell lysates of APL patients, are similar to bands that can be detected on the Western blot image of purified FXIII-A2 proteolytically degraded by HNE (Fig. 5).

Figure 5.

Western blotting of FXIII-A and FXIII-A cleavage products in promyelocyte cell lysates of AML M3 patients as compared to purified FXIII-A2 proteolytically degraded by human neutrophil elastase (HNE). Solid arrow indicates intact, uncleaved FXIII-A, dotted arrows indicate FXIII-A split products cleaved by HNE. Mr: molecular weight marker, FXIII-A st.: 18 ng FXIII-A standard purified from placenta, b.m.: sample purified from bone marrow. From “patient A” a bone marrow sample, from “patient B” both peripheral blood and bone marrow samples were analyzed.

These two APL samples were also analyzed by ELISA in order to measure the amount of FXIII-A as an antigen in cell lysates. These samples contained over 90% promyelocytes and no measurable monocytes at diagnosis by both flow cytometry and hematology analyzer, therefore, the possibility that the measured FXIII-A originated from monocytes was excluded. In these selected APL samples the following amounts of FXIII-A were measured: 29 and 80 fg/cell; thus in these cases leukemic promyelocytes contained FXIII-A antigen in amounts similar to that described previously in platelets (mean: 60 ± 10 fg/platelet) [9]. These values corresponded to the mean fluorescence intensity values obtained in flow cytometric labeling (data not shown).

Whether normal promyelocytes are a source of cellular FXIII-A was investigated by a seven-color technique using the BD FacsCantoII 3 laser flow cytometer. Three bone marrow samples with no leukemia associated changes were included in this study. Normal promyelocytes were identified by the CD15, CD33, and CD117dim positive staining and HLA-DR negativity. As can be seen compared to normal monocytes these promyelocytes lack FXIII-A expression.


The intracellular presence of FXIII-A has been known for more than 50 years in platelets and megakaryocytes [4] and it is also known for decades that monocytes and macrophages express FXIII-A as well [6]. In later studies, this cytoplasmic enzyme proved useful in the identification of leukemic samples and it differentiated myeloblasts from monoblasts and promonocytes [12]. More recently, we have also identified leukemic lymphoblasts as a source of FXIII-A [18]. Whether this marker is a prognostic factor in childhood B-ALL is presently under investigation. In this study FXIII-A was identified in leukemic promyelocytes derived from de novo AML M3 cases. The presence of FXIII-A antigen was proved by flow cytometry and confocal microscopy. It was also important to establish the amount of FXIII-A since in our previous studies the FXIII-A content of various cell types was largely different [9, 18]. We obtained sufficient cell number from two clinical cases that showed a high FXIII-A content, on average the amount of FXIII-A was in the range of blood platelets and 10-fold higher than that described previously for leukemic lymphoblasts [18]. Although, it should be noted that APL blasts are much larger than platelets thus the intracellular concentration of FXIII-A is considerably lower. The intracellular A subunit structure has long been investigated by gel filtration and electrophoresis techniques and has been shown to occur in a dimeric (A2) form in platelets [19]. In later studies, subsequent cellular expression sites were identified and a similar A2 structure was suggested but was never directly investigated.

The modulation of FXIII activity is an important aspect whether related to plasma or cytoplasmic factor XIII. In the case of plasma FXIII, usually thrombin and calcium activates the zymogen and their action results in the proteolytic cleavage of an activation peptide. The activation of cellular FXIII is a much slower process and it does not require proteolytic splitting, since the increase of cellular Ca++ is sufficient to induce the activation process [1, 17]. The inactivation of FXIII is also not very well known since—unlike many other clotting factors—no down-regulating mechanisms have been known in the past. More recently, we have shown that the activity of FXIII-A can be modulated by HNE [20, 21]. In this recent study, we used similar concentrations of HNE as previously that can temporarily activate and then downregulate cellular FXIII-A activity. We think that the degraded FXIII-A observed in APL samples was due to intracellular proteolytic degradation by HNE and the extent to which these leukemic promyleocytes may contribute to FXIII-A activity when APL cells are lysed upon treatment, may depend largely on their degradation by HNE.

In our previous studies [12] we found that the myeloblastic PLB-985 cell line never expressed FXIII-A throughout its development to mature neutrophilic granulocyte, thus we hypothetized that the presence of FXIII-A in leukemic promyelocytes is a leukemia associated immunophenotype. This notion was confirmed when FXIII-A was undetectable in normal promyelocytes in our series of measurements on nonleukemic bone marrow samples (Fig. 6).

Figure 6.

Representative dot plots from normal bone marrow samples analyzed with a seven-color combination. Blue dots encircled with continuous line show FXIII-A, CD33, HLA-DR expressing monocytes, whereas red dots encircled with dashed line represent CD15, CD33, CD117dim positive, and HLA-DR negative promyelocytes, which lack FXIII-A expression. [Color figure can be viewed in the online issue, which is available at]

The origin of APL cells is not yet settled. There is a growing body of evidence that APL cells have a basophilic affiliation. Indeed in previous studies, some monocyte-associated antigens (CD9 and CD68) but not others (CD14) were found in APL cells [22–24]. Identifying marker positivity in APL is also of importance since it is now widely accepted that APL is mostly characterized by negativity for stainings instead of well-defined positivities. The best such negative predictive markers that were established are CD11a, CD18, and HLA-DR. Because of the low patient number we do not speculate on the potential association of FXIII-A expression with disseminated intravascular coagulation and subsequent bleeding that are attributes of AML M3. Bleeding is multicausal in this disorder and one contributing factor is elastase [25]; however, judging its effect on bleeding is hampered by the fact that several patients are also severely thrombocytopenic. An interesting phenomenon however, that may need attention is that all FXIII-A negative patients were nonresponsive to treatment or relapsed, while all FXIII-A positive APL patients are still alive (Fig. 7).

Figure 7.

Survival plot of FXIII-A positive and negative APL cases showed significant difference (P < 0.0001 as compared by logrank test, median survival of FXIII negative patients was 1 month).

In conclusion, we have provided immunochemical and morphological evidence for the presence of cellular FXIII-A in leukemic promyelocytes in cleaved and uncleaved forms. Further studies are required to evaluate whether the cellular FXIII-A also have prognostic implications in APL.


The authors would like to thank, Valéria Sziráki Kiss, Gizella Haramura, and Ildikó Beke Debreceni for technical assistance and Csaba Antal and Ildikó Kópis for administrative help. This study was supported by grants OTKA K-75199 (JK), OTKA-NKTH NI-69238 (LM) and the János Bolyai Research Scholarship of the Hungarian Academy of Sciences (ZsH and ZsB). Grant sponsor: University of Debrecen, MHSC, Grant number: Mec-1/2011 and Lajos Szodoray Prize (ZsB). Grant sponsor: New Hungary Development Plan, co-financed by the European Social Fund. Grant number: TÁMOP 4.2.1./B-09/1/KONV-2010-0007 and TÁMOP 4.2.2./B-10/1-2010-0024.