The granzyme B inhibitor proteinase inhibitor 9 (PI9) is expressed by human mast cells


  • Bellinda A. Bladergroen,

    Corresponding author
    1. Department of Tumorimmunology, Radboud University Nijmegen Medical Centre, Nijmegen Centre for the Molecular Life Sciences, Nijmegen, The Netherlands
    2. Department of Clinical Chemistry, VU University Medical Center, Amsterdam, The Netherlands
    • UMC St. Radboud, NCMLS, Department of Tumorimmunology, Geert Grooteplein 28, 6525 GA Nijmegen, The Netherlands, Fax: +31-24-354-0339
    Search for more papers by this author
  • Merel C. M. Strik,

    1. Department of Immunopathology, Sanquin Research at CLB, and Landsteiner Laboratory, Academical Medical Center, Amsterdam, The Netherlands
    2. Department of Clinical Chemistry, VU University Medical Center, Amsterdam, The Netherlands
    Search for more papers by this author
  • Angela M. Wolbink,

    1. Department of Immunopathology, Sanquin Research at CLB, and Landsteiner Laboratory, Academical Medical Center, Amsterdam, The Netherlands
    Search for more papers by this author
  • Dorine Wouters,

    1. Department of Clinical Chemistry, VU University Medical Center, Amsterdam, The Netherlands
    Search for more papers by this author
  • Roel Broekhuizen,

    1. Department of Pathology, University Medical Center Utrecht, Utrecht, The Netherlands
    2. Department of Pathology, VU University Medical Center, Amsterdam, The Netherlands
    Search for more papers by this author
  • J. Alain Kummer,

    1. Department of Pathology, University Medical Center Utrecht, Utrecht, The Netherlands
    Search for more papers by this author
  • C. Erik Hack

    1. Department of Immunopathology, Sanquin Research at CLB, and Landsteiner Laboratory, Academical Medical Center, Amsterdam, The Netherlands
    2. Department of Clinical Chemistry, VU University Medical Center, Amsterdam, The Netherlands
    Search for more papers by this author


The activity of granzyme B, a main effector molecule of cytotoxic T lymphocytes (CTL) and natural killer cells, is regulated by the human intracellular serpin proteinase inhibitor 9 (PI9). This inhibitor is particularly expressed by CTL and dendritic cells, in which it serves to protect these cells against endogenous and locally released granzyme B. Moreover, PI9 expression by neoplastic cells may constitute one of the mechanisms for tumors to escape immune surveillance. Here we show that PI9 is also expressed by human mast cells. In immunohistochemical studies using a PI9-specific monoclonal antibody, strong cytoplasmic staining for PI9 was found in normal mast cells in various tissues throughout the body. In addition, in 80% of all cases of cutaneous and systemic mastocytosis tested the majority of the mast cells expressed PI9. As an in vitro model for PI9 expression by mast cells, we studied expression by the human mast cell line HMC-1. Stimulation of HMC-1 with PMA and the calcium ionophore A23187 resulted in a marked increase of PI9 expression. Thus, PI9 is expressed by activated mast cells. We suggest that this expression serves to protect these cells against apoptosis induced by granzyme B released during initiation of the local inflammatory response.


Proteinase inhibitor


Human mast cell line


Proteinase inhibitor 9 (PI9) is a member of the large superfamily of serine proteinase inhibitors (serpins). Serpins are widely dispersed in nature and regulate the activity of proteases in diverse (patho)physiological processes such as coagulation, fibrinolysis, inflammation, cell migration, tumorigenesis and apoptosis 13. Although they are functionally diverse proteins, serpins have a similar 3D-structure and inhibition mechanism 3.

PI9 belongs to a subfamily of serpins, called the ovalbumin- or intracellular serpins 4. Compared to other members of the serpin family, intracellular serpins share a higher degree of homology (about 30% and >50% homology, respectively) and they lack a typical cleavable N-terminal signal sequence, so that they reside mainly intracellularly. Human intracellular serpins include plasminogen activator inhibitor-2 (PAI-2), monocyte neutrophil elastase inhibitor (MNEI), PI5 (maspin), PI6, PI8, PI9, PI10 (bomapin), PI13 (hurpin), and squamous cell carcinoma antigen-1 and -2 (SCCA1/2) 2. They are widely expressed in human tissues, and several intracellular serpins, including PI9, PI8, PI6, MNEI, and PAI-2, show a nucleocytoplasmic subcellular distribution 5. Although different serpins are present in various human tissues, each individual serpin has its own restricted expression pattern, reflecting its unique physiological function.

PI9 efficiently inhibits the serine protease granzyme B, which induces DNA-fragmentation and apoptosis in target cells when released from degranulating cytotoxic cells, such as cytotoxic T lymphocytes (CTL) and natural killer (NK) cells, upon target cell recognition. PI9 has also been reported to inhibit caspase 1 6, 7 and caspase 4 8in vitro. Although far less efficient than that of granzyme B, this inhibition endows PI9 with additional features to interfere with the apoptotic process. In vitro, PI9-transfected cells are protected against granzyme B-induced apoptosis 8, 9. In normal human tissues, PI9 is mainly expressed by dendritic cells, CTL, cells at immune privileged sites, endothelial and mesothelial cells 10, 11. This in vivo distribution pattern supports the hypothesis that PI9 is expressed at sites where degranulation of CTL or NK cells is potentially deleterious. By scavenging granzyme B PI9 serves to protect CTL from death induced by their own, misdirected granzyme B 8 and other cells from granzyme B released from neighboring or circulating CTL during the immune response 10, 11.

PI9 is also present in tumors of different origin. In particular, PI9 is expressed by neoplastic cells in some cases of Hodgkin lymphoma and in a substantial number of T and B cell non-Hodgkin lymphoma (NHL), implicating a novel mechanism for these tumors to escape immune surveillance by tumor-infiltrating CTL or NK cells 12. Although less pronounced, PI9 is also expressed by several carcinomas such as nasopharyngeal carcinoma 13 and carcinomas of breast, colon, and cervix 14. In anaplastic large cell lymphoma (ALCL) high numbers of PI9-positive neoplastic cells appeared to be strongly related to poor prognosis 15. These tumors, in general, are less susceptible to chemotherapy, and it can be speculated that this results from the ability of PI9 to inhibit some caspases 68. Whether PI9 also predicts poor clinical outcome in tumors other than ALCL is not known yet.

To obtain further clues about the function of PI9 we investigated the expression of PI9 in neoplastic and normal tissues in more detail. Here we report abundant expression of PI9 by normal as well as (pre-)malignant mast cells. As a model for mast cells in vitro, we also analyzed PI9 expression in human mast cell line (HMC)-1 cells, and show that these cells markedly increase PI9 expression upon stimulation.


PI9 is expressed in normal mast cells

Analysis of PI9 expression by immunohistochemistry with mAb PI9–17 revealed cytoplasmic staining for PI9 in mast cells in various tissues throughout the body (Fig. 1). For example, PI9 expression was observed in mast cells present in tonsil, muscle, and skin (Fig. 1A, C, and D, respectively). The identity of the PI9-positive mast cells was confirmed by sequential staining with antibodies specific for mast cell marker proteins such as mast cell tryptase (Fig. 1B) and CD117 (c-Kit; not shown). Detection of PI9 in CTL or dendritic cells served as a positive internal control for the staining procedure in all cases (results not shown). No PI9 staining was observed when isotype control antibodies were used (results not shown).

Figure 1.

PI9 is expressed by normal mast cells. Tissue sections of tonsil (A, B), muscle (C) and skin (D) were stained with anti-PI9 mAb PI9–17 (A, C, D), or anti-tryptase mAb AA1 (B). Original magnification, x400 (A); x630 (C, D); x1000 (B).

(Pre-)malignant cells in cutaneous and systemic mastocytosis express PI9

PI9 expression by mast cells was found in all of 2 cases of urticaria pigmentosa, 6 cases of diffuse cutaneous mastocytosis, 9 cases of mastocytoma, and 7 cases of systemic mastocytosis (Table 1). In the majority of mastocytosis lesions (16 out of 24 cases) as much as 75–100% of the mast cells were positive for PI9. Fig. 2 shows PI9-positive cells in a case of mastocytoma (Fig. 2A), urticaria pigmentosa (Fig. 2C) and systemic mastocytosis (Fig. 2D).

Table 1. Patient characteristics and expression of PI9, tryptase, and CD117 in relation to type of mastocytosis
Biopsy no.TypePatient characteristicsPercentage of positive tumor cells
  1. a) ND, not determined.

AgeSexPlace of biopsyPI9TryptaseCD117
 1Urticaria pigmentosa21 yearsMTrunk (back)75–10075–10075–100
 2Urticaria pigmentosa48 yearsFLeg50–7575–100NDa)
 3Diffuse cutaneous mastocytosis 9 monthsFTrunk (back)75–10075–10075–100
 4Diffuse cutaneous mastocytosis 7 monthsMButtock75–10050–7575–100
 5Diffuse cutaneous mastocytosis 9 monthsFTrunk (back)75–10025–5075–100
 6Diffuse cutaneous mastocytosis 6 monthsFTrunk (shoulder)ND25–5075–100
 7Diffuse cutaneous mastocytosis44 yearsMTrunk<575–100ND
 8Diffuse cutaneous mastocytosis18 yearsMTrunk (back)<575–100ND
 9Mastocytoma11 monthsMFoot75–10075–100ND
10Mastocytoma 6 monthsMLeg75–100 5–2575–100
11Mastocytoma 9 monthsMTrunk (back)75–10025–5075–100
12Mastocytoma 6 monthsFTrunk (abdomen)75–10050–7575–100
13Mastocytoma 6 monthsFGroin75–10075–100ND
14Mastocytoma 2 yearsFLeg50–7575–100ND
15Mastocytoma 1 yearMArm75–10075–100ND
16Mastocytoma 2 yearsFHead75–10075–100ND
17Mastocytoma10 monthsMArm75–10075–100ND
18Systemic mastocytosis40 yearsMColon 5–2550–7575–100
19Systemic mastocytosis75 yearsFBone marrow75–10050–7575–100
20Systemic mastocytosis53 yearsFBone marrow75–10075–100ND
21Systemic mastocytosis44 yearsFBone marrow<575–100ND
22Systemic mastocytosis58 yearsFBone marrow75–10075–100ND
23Systemic mastocytosis59 yearsMBone marrow75–10075–100ND
24Systemic mastocytosis39 yearsFBone marrow 5–2575–100ND
Figure 2.

PI9 is expressed by (pre-)malignant cells in cutaneous and systemic mastocytosis. Tissue sections of mastocytoma (A, B), urticaria pigmentosa (C), and systemic mastocytosis (D) were stained with anti-PI9 mAb PI9–17 (A, C, D) or anti-tryptase mAb AA1 (B). Original magnification, x200 (C); x400 (A, B, D).

Notably, in most cases the intensity of staining of PI9 and tryptase was comparable in normal and malignant cells (Figs. 1 and 2, compare A with B). However, as shown in Table 1, the percentage of tryptase-positive mast cells in a number of cases of diffuse cutaneous mastocytosis and mastocytoma, but not in urticaria pigmentosa, was substantially lower than 75–100%, indicating loss of this marker. On the other hand, in all lesions tested nearly all mast cells exhibited positive staining for the other mast cell marker, CD117.

Generation of new specific antibodies and development of an ELISA for soluble PI9

To measure expression of PI9 by cells quantitatively, we developed an ELISA for soluble PI9. Since mAb PI9–17, which had been raised against rPI9 isolated from inclusion bodies 10, did not bind soluble PI9 in ELISA or immunoprecipitation studies (results not shown), new anti-PI9 antibodies were generated. From a fusion experiment of a mouse immunized with yeast-expressed rPI9, four hybridomas producing stable anti-PI9 mAb were selected. Three anti-PI9 mAb (mAb PI9–101, mAb PI9–102, and mAb PI9–104) were of the IgG1 subtype, one mAb (mAb PI9–103) was of the IgG2b subtype, and they all had kappa light chains. In addition, a polyclonal anti-PI9 Ab was produced. In screening experiments, using 125I-radiolabeled recombinant serpins in a RIA, all mAb as well as the polyclonal Ab appeared to be specific for PI9 and did not cross-react with the two most homologous serpins PI6 and PI8 (data not shown).

Testing these antibodies for their ability to bind fluid phase PI9 in ELISA revealed an optimal combination of polyclonal anti-PI9 as the coating antibody and mAb PI9–101 as detecting antibody. Therefore, this antibody combination was used to develop a highly reproducible ELISA for the detection of fluid phase PI9. Fig. 3 shows a dose-response curve for soluble rPI9 purified from yeast (Fig. 3A) as well as for PI9 present in cell lysate of YT-indy cells or Jurkat cells stably transfected with PI9-cDNA (Fig. 3B). Since no cross-reactivity was observed with PI6 and PI8, either present as recombinant proteins purified from yeast, or in the cell lysate of transfected Jurkat cells (Fig. 3), this ELISA was considered to be specific for PI9. In further experiments YT-indy cell lysate was used as a standard.

Figure 3.

ELISA for fluid phase PI9. A sandwich ELISA for the detection of soluble PI9 was developed with polyclonal anti-PI9 as catching antibody and biotinylated mAb PI9–101 as detecting antibody. (A) Serial dilutions of recombinant PI9, and, as a control, recombinant PI8 and PI6 were tested. (B) Dilutions of YT-indy and Jurkat cell lysate or lysates of Jurkat cells stably transfected with cDNA encoding full-length PI9, PI8 or PI6 were also tested.

Stimulation of mast cells results in up-regulation of PI9 expression

PI9 expression in mast cells was studied in more detail in the human mast cell line HMC-1. In untreated HMC-1 cells the concentration of PI9 was approximately 0.5 ng/106 cells as determined by ELISA. The localization of PI9 in the cytoplasm of mast cells as determined by immunohistochemistry was confirmed by subcellular fractionation of HMC-1 cells. On a sucrose density gradient PI9 co-migrated with the cytosolic enzyme LDH (Fig. 4C). On Western blot, anti-PI9 mAb PI9–17 recognized a 42-kDa band in lysates of HMC-1 cells (Fig. 4D, lane 1). This band is consistent with the expected size of native uncleaved PI9. Notably, anti-PI9 mAb PI9–17 also recognized a protein band of about 67 kDa, of which the identity is yet unknown. In contrast to the 42-kDa band, this higher molecular mass protein band could not be seen after immunoprecipitation with anti-PI9 mAb PI9–101, which was raised against rPI9 expressed in yeast (Fig. 4D, lane 3).

Figure 4.

Stimulation of HMC-1 cells results in up-regulation of PI9 expression. HMC-1 cells were cultured in the absence (control) or presence (PMA/A23187) of PMA (30 ng/ml) and A23187 (350 ng/ml) for varying times. At the time points indicated, the culture medium was collected and cells were lysed as described in Materials and methods. PI9 levels in cell lysate (A) or culture medium (B) were measured by ELISA. Values are presented as the mean ± SEM (n=6). (C) Unstimulated HMC-1 cells were subject to fractionation by centrifugation on a sucrose density gradient. Density increases from right (0.5 M sucrose) to left (2.0 M sucrose). Fractions were analyzed for PI9 (left axis) and LDH activity (right axis). (D) PI9 was immunoprecipitated from cell lysate of control and PMA/A23187 stimulated HMC-1 cells with mAb PI9-101. Cell lysate (10 μl) or 10 μl of immunoprecipitate were separated by 10% SDS-PAGE and analyzed by immunoblotting with mAb PI9–17. The arrow denotes the position of the 42-kDa band of PI9. The position of the molecular mass marker bands is indicated on the left. IP, immunoprecipitation.

Stimulation of HMC-1 cells with PMA and the calcium ionophore A23187 resulted in a dramatic increase in PI9 expression, which reached a maximum 48 h after stimulation (18.3 ng/106 cells; Fig. 4A). This increase in the amount of PI9 could also be demonstrated by Western blot analysis (Fig. 4D, lanes 2 and 4). From lysates of stimulated cells also a protein migrating at approximately 38 kDa was precipitated. This 38-kDa band probably represents PI9, which has been degraded or cleaved in the reactive site loop by a yet unidentified protease.

Minimal amounts of PI9 were detected in the medium of control and stimulated cells (Fig. 4B), although there was a slight increase in the PI9-concentration up to 96 h after stimulation. PI9 present in the medium probably represents protein released from dying cells rather than secreted PI9.


Several proteinase inhibitors, including PI6 16, are expressed by mast cells. In the present study we demonstrate that these cells also express the intracellular serpin PI9. In immunohistochemical studies it was found that PI9 is expressed by both normal and (pre-) malignant mast cells in various human tissues. In the majority of cases of cutaneous- and systemic mastocytosis tested 75–100% of the mast cells expressed PI9. PI9 was also expressed by the human mast cell line HMC-1. This expression was up-regulated upon stimulation of the cells with PMA and the calcium ionophore A23187.

PI9 has previously been described to be present mainly in CTL, dendritic cells, endothelial cells, and at immune privileged sites 911. So, PI9 expression appeared to be restricted to cell types that have to deal with substantial levels of granzyme B either originating from misdirection of own cellular reservoirs, or released from circulating or neighboring CTL and NK cells. For all these cells, resistance to granzyme B-induced apoptosis by expression of PI9 seems to be a prerequisite to keep their function 10, 11.

The function of mast cells in vivo is far from clear. Their location predicts that these cells form an important part of the initial host defense against pathogens that enter the body across epithelial barriers. Upon stimulation of mast cells, granule proteases and other mediators are released, that promote vascular permeability, contribute to tissue remodeling, and recruit inflammatory cells 17. The latter property brings mast cells in close vicinity to degranulating NK cells and CTL. Therefore, it is likely that also in mast cells expression of PI9 serves to protect these cells from the deleterious effects of granzyme B released by cytotoxic cells.

Upon stimulation, mast cells exert their functions by release of several mediators, among others histamine and several serine proteases. The majority of the serine proteases has trypsin-like (tryptase) or chymotrypsin-like (chymase) specificity 17. Since PI9 contains a Glu residue at position P1 in its so-called reactive-site loop, endowing the inhibitor with specificity for granzyme B-like activities, PI9 unlikely is expressed to protect mast cells against misdirected tryptase or chymase. Protection of mast cells from the proteolytic activities of these enzymes, which are released from secretory granules into the cytosol, is at least partially achieved by expression of other intracellular serpins, closely related to PI9. PI6, one of the two most homologous serpins to PI9 (63% amino acid homology), was recently found to bind to β-tryptase in mast cells 16. Furthermore, squamous cell carcinoma antigen 2 has been shown to be an inhibitor of mast cell chymase and cathepsin G 18, and monocyte/neutrophil elastase inhibitor (MNEI) was reported to show a rapid interaction with mast cell chymase as well 19.

Considering the fact that cytotoxic cells express PI9 to be protected against their own, misdirected granzyme B, expression of this inhibitor by mast cells raises the question whether these cells may also express this serine protease. Indeed, in a recent study we observed that granzyme B expression is markedly up-regulated in the mast cell line HMC-1 upon stimulation with PMA and the calcium ionophore A23187 (M. Strik et al., manuscript in preparation). In rodents, activated mast cells in tissues of parasite-infected animals express serine proteases with granzyme B-like activity, like mouse mast cell serine protease 8 (MMCP-8) 20 and the rat mast cell proteases (RMCP)-2, -8, -9, and -10 21. Interestingly, whereas the expression of MMCP-8 was found to be very low in normal mouse tissues, it was high or strongly increased in a mouse mast cell-like tumor line and mast cells in lungs of Schistosoma mansoni-infected BALB/c mice 20. This is in accordance with our observation that the expression of both PI9 and granzyme B was very low or undetectable in unstimulated HMC-1 cells, to be raised dramatically upon stimulation with PMA and A23187. Hence, it might be speculated that PI9 protects mast cells not only against granzyme B released from neighboring cells, but also against the damaging effects of endogenously produced proteases with granzyme B-like activity.

Under physiological conditions mast cells usually are activated through cross-linking of the high-affinity receptor for immunoglobulin E, FcϵRI, with IgE and antigen 22. Aggregation of FcϵRI eventually results in degranulation and release of granule proteases, which occurs within minutes, synthesis and release of lipid mediators, and up-regulation of gene transcription leading to synthesis and secretion of cytokines, chemokines and growth factors. We postulate that during the early phase of the mast cell response, PI6 is a main serpin involved in the protection of the mast cells, since it can bind and inhibit tryptase, which is already present in the cell and is released immediately upon degranulation 16. PI9 is likely to exert a similar function during a later phase after induction of gene transcription, since PI9 levels in unstimulated HMC-1 cells were very low and became significantly elevated not until at least 6 h after stimulation with PMA and A23187 (see Fig. 4A). The requirement for the induction of gene transcription and de novo synthesis of protein in order to express a serpin at significant amounts has also been described for the extracellular serpin plasminogen activator inhibitor 1 (PAI-1) 23. Secretion of PAI-1 was first detected 8 h after stimulation of HMC-1 cells with PMA and A23187, to reach a maximum after 24 h. Secretion of PAI-1 was almost completely abrogated by pretreatment of the cells with cycloheximide 23.

The effects of FcϵRI cross-linking are mediated via several signal transduction pathways, among others those following activation of phosphatidylinositol-3-kinase (PI-3-K), phospholipase C γ (PLCγ), and c-Jun N-terminal kinase (JNK) 22. Activation of these signal transduction molecules is likely to lead to induction of PI9 gene transcription by binding of the appropriate signal transduction factors to the nuclear factor-κB (NF-κB) and activator protein-1 (AP-1) sites in the promoter region 24. Treatment of HMC-1 cells with PMA and A23187 mimicks the PLCγ pathway, and the induction of PI9 gene expression observed in these cells is in agreement with earlier reports describing an increase in PI9 mRNA and protein levels in cells treated with the phorbol esters TPA or PMA 11, 24.

Neoplastic cells in the mastocytoma resemble PMA-stimulated HMC-1 cells in that they both expressed high levels of PI9. However, in vivo also non-tumor mast cells expressed substantial levels of PI9, while in vitro unstimulated HMC-1 cells expressed only minimal PI9-levels. We speculate that PI9-positive mast cells in vivo represent activated mast cells, while PI9-negative mast cells are non-activated counterparts.

In conclusion, PI9 is expressed abundantly in normal and (pre-)malignant mast cells. Invitro, PI9 expression in the human mast cell line HMC-1 is up-regulated after stimulation with PMA and A23187. We suggest, that up-regulation of PI9 by activated mast cells may serve to protect these cells against apoptosis induced by granzyme B released during initiation of the local inflammatory response, or against their own, misdirected granzyme B.

Materials and methods

Materials and cells

Human mast cell line-1 (HMC-1) cells were kindly provided by Dr. J. Butterfield (Mayo Clinic, Rochester, MN). IMDM was obtained from Bio-Whittaker Europe (Verviers, Belgium). Bovine calf serum was purchased from Hyclone (Logan, UT). Anti-tryptase monoclonal antibody (mAb) AA1 (subtype IgG1), anti-CD117 (c-kit) polyclonal Ab, biotinylated rabbit-anti-mouse F(ab′)2 Ig, biotinylated swine-anti-rabbit F(ab′)2, HRP-conjugated rabbit anti-mouse Ig, avidine-biotin-HRP complex (sABC), and biotinylated tyramine were obtained from Dako (Glostrup, Denmark). Phorbol 12-myristate 13-acetate (PMA) and calcium ionophore A23187 were purchased from Sigma (St. Louis, MO). Protein G-Sepharose, protein A-Sepharose, and CNBr-Sepharose beads were from Amersham Pharmacia Biotech (Uppsala, Sweden). 3-Amino-9-ethylcarbazole (AEC) was from Zymed Laboratories Inc. (San Francisco, CA). All other chemicals were of analytical grade.

Tissue sections

Formalin-fixed, paraffin-embedded biopsies from patients with cutaneous (n=17) and systemic (n=7) mastocytosis were selected from the tissue bank of the Department of Pathology, VU University Medical Center, Amsterdam (n=11) and the Department of Pathology, UMC Utrecht (n=13). The mastocytosis lesions consisted of 2 cases of urticaria pigmentosa, 6 cases of diffuse cutaneous mastocytosis, 9 cases of mastocytoma, and 7 cases of systemic mastocytosis (1 colon and 6 bone marrow biopsies). Patient characteristics are further shown in Table 1.

Healthy control tissues such as skin, tonsil, muscle and colon from regular surgical pathology archival material were used as controls. All tissue samples were prepared from surgical specimens within 2 h of resection. Tissues were processed routinely by fixation in 10% formalin for 18 h and subsequently embedded in paraffin.


Sections (3 μm thick) from the paraffin-embedded biopsies were mounted on slides coated with poly-L-lysine and immunohistochemistry was essentially performed as described previously for PI9 and PI6 10, 16. Tissue sections were subjected to antigen retrieval by boiling in 10 mM sodiumcitrate (pH 6) for 10 min., which was performed in a pressure cooker in case sections were stained for PI9 or in a microwave oven in all other cases. The following primary antibodies were used: anti-PI9 mAb PI9–17 (at 2.5 μg/ml), anti-tryptase mAb AA1 (diluted 1:200) and pAb anti-CD 117 (c-kit, diluted 1:50). Negative control slides were stained with IgG of the appropriate subclass. Bound antibodies were visualized with 3-amino-9-ethylcarbazole (AEC), counterstained with hematoxylin and mounted.

Cell culture and stimulation of HMC-1 cells

SP2/0-Ag-14 mouse myeloma cells were cultured in IMDM containing 5% (v/v) FBS, 50 U/ml penicillin, 50 μg/ml streptomycin and 0.0003% (v/v) β-mercaptoethanol. Hybridomas were cultured in SP 2/0 medium further supplemented with IL-6 (0.5 ng/ml, final concentration). YT-indy cells were cultured and lysed as described 10. Jurkat cells were stably transfected with constructs comprising the cDNA encoding full-length PI9, PI8 or PI6 and cultured as described previously 16. HMC-1 was cultured as described 25. HMC-1 cells were stimulated by incubation in culture medium supplemented with PMA (30 ng/ml, final concentration) and the calcium ionophore A23187 (350 ng/ml, final concentration) for varying times. All cells were maintained under an atmosphere of 5% CO2/95% air in a humidified incubator at 37°C.

After incubation of HMC-1 cells, the cells were spun down and the culture medium was collected and stored at –20°C until further use. To obtain cell lysates, the cell pellet was washed twice with phosphate buffered saline, pH 7.4 (PBS), and resuspended in lysis buffer (150 mM NaCl, 20 mM Tris-HCl pH 7.5, 10% (v/v) glycerol and 1% (v/v) Triton X-100) at a concentration of 40×106 cells/ml. Cells were lysed for 10 min on ice, whereafter the lysate was centrifuged for 5 min at 3,000 rpm at 4°C to remove cell debris and DNA. The supernatant (cell lysate) was stored at –20°C until further use.

Expression and purification of recombinant human serpins

The yeast plasmids pHIL-D2-PI9, pPIC3-PI8, and pHIL-D2-PI6, coding for histidine-tagged human recombinant serpins, were a kind gift from Dr. W. Kisiel (Department of Pathology, New Mexico School of Medicine, Albuquerque, New Mexico) 26. The recombinant serpins were expressed and purified as described previously 16.

Production and purification of anti-PI9 polyclonal antibody

Rabbits were immunized with 50 μg of purified yeast-expressed histidine-tagged rPI9 emulsified in complete Freund's adjuvant, followed by three monthly booster injections with rPI9 emulsified in incomplete adjuvant. Serum antibody responses were screened using a radioimmunoassay (RIA) in which anti-rabbit immunoglobuline-Sepharose beads and 125I-radiolabeled rPI9 were used, similarly as described previously for anti-C3 antibodies 27. Four days after the last booster injection, citrated blood was collected and processed into plasma. IgG antibodies were purified from plasma by protein G-Sepharose affinity chromatography.

Production and purification of anti-PI9 mAb

mAb PI9–17, specific for PI9, was produced and purified as described previously 10. New anti-PI9 mAb were produced and purified according to the protocol described for anti-PI6 mAb 16. In short, BALB/c mice were immunized both intravenously and intraperitoneally with 10 μg yeast-expressed rPI9 emulsified in complete Freund's adjuvant, followed by booster injections every three weeks with rPI9 emulsified in incomplete adjuvant (five-to-seven times). Serum antibody titers were screened with RIA using anti-mouse Ig-Sepharose beads and 125I-radiolabeled rPI9. Spleen and lymph node cells of the animal with the highest serum antibody titer were used for fusion with SP2/0-Ag-14 mouse myeloma cells according to standard procedures. Hybridoma culture supernatants were screened for the presence of anti-PI9 antibodies using a RIA with 125I-PI9 as described above. The specificity and cross-reactivity of the antibodies were assessed from a RIA with 125I-PI6 and 125I-PI8. Hybridomas producing specific anti-PI9 mAb were selected and subcloned by several cycles of limiting dilution. Anti-PI9 mAb were purified from conditioned medium by protein A-affinity chromatography. mAb were biotinylated using long chain biotinyl-N-hydroxysuccinimide ester sulfonic acid (Pierce Chemical Co, Rockford, IL) according to the manufacturer's instructions.


The procedure of the PI9 Elisa was essentially the same as described previously for PI6 16. Polyclonal anti-PI9 antibody (1 μg/ml) was used for coating to the microtiter plates, and biotinylated mAb PI9-101 (about 0.5 μg/ml) for detection. Lysate of YT-indy cells, a natural killer cell line expressing PI9, served as a standard.

Subcellular fractionation of HMC-1 cells

HMC-1 cells (40×106 cells) were washed and homogenized in a buffer containing 250 mM sucrose and 20 mM Tris-HCl, pH 7.4, with a syringe (25G x ½"). After centrifugation of the homogenate for 2 min at 2,400×g the supernatant was applied to a preformed sucrose density 0.1 M step gradient (0.4 M–2.0 M in 20 mM Tris-HCl, pH 7.4) and centrifuged overnight in an ultracentifuge at 100,000×g at 4°C. Subsequently, 0.5-ml fractions were collected and assayed for PI9 with ELISA as described above, and lactate dehydrogenase (LDH) activity 28.


PI9 was precipitated from 750-μl lysate of unstimulated or stimulated HMC-1 cells by head over head rotation overnight at 4°C with anti-PI9 mAb PI9-101 coupled to CNBr-Sepharose beads. Subsequently, the beads were pelleted and washed seven times with PBS/0.1% (v/v) Tween-20. Bound proteins were dissolved in 40 μl twice concentrated non-reducing SDS-sample buffer supplemented with 8 M urea and the pellet was stored at -20°C.

Before analysis by SDS-PAGE and Western blotting pellets were boiled and Sepharose beads were briefly spun down. The supernatant, containing immunoprecipitate, was collected and DTT was added (10 mM, final concentration).


Ten microliters of cell lysate from control or stimulated (t =2 days) HMC-1 cells was resolved by electrophoresis on a 10% SDS-polyacrylamide gel under reducing conditions. In case of immunoprecipitation, 10 μl of the resolved pellets were loaded. After electrophoresis, proteins were transferred to nitrocellulose membranes by electrophoretic blotting. The membranes were then blocked for 1 h in blocking buffer [5% (w/v) skim milk powder, 0.5% (w/v) bovine serum albumin (BSA), and 0.1% (v/v) Tween-20 in PBS]. Subsequently, membranes were incubated overnight with 1.8 μg/ml mAb PI9–17 diluted in blocking buffer, followed by HRP-conjugated rabbit-anti-mouse Ig. Bound antibodies were visualized with a chemiluminescence development reagent (ECL-system; Amersham) according to the manufacturer's instructions.


We acknowledge financial support from VUMC Verruiming Middelenkader.


  1. 1


  2. 2


  3. 3


  4. 4