| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Experimental Therapeutics, Preclinical Pharmacology |
Cancer Immunology Research Laboratory [A. W. T., M. H. P., D. T. A., G. O., J. M. L., M. J. S.] and Department of Pathology [G. N.], Baylor-Sammons Cancer Center, Baylor University Medical Center, Dallas, Texas 75246
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, IL-6, and transforming growth factor ß inhibit BrCa cell growth in vitro (3, 4, 5)
. Expression of an IL-2 transgene also induced breast cancer cell necrosis, an outcome that was correlated with the extent of lymphocytic infiltration (6)
. Perturbation of the surface HER-2/neu receptor produced objective clinical responses in a recent clinical trial (7)
. These findings indicate that breast cancer cells are subject to biological and immunological modulation by virtue of specific receptor expression. CD40 (Mr 45,000–50,000; 277 amino acids) is a type I transmembrane glycoprotein receptor of the TNF-receptor superfamily. CD40 is best known as a growth signal receptor for B lymphocytes (8, 9, 10) . Interaction of the CD40+ B cell with an activated, CD40L-expressing T cell (or with sCD40L) produces multifaceted regulatory signals, ranging from T-cell-dependent B-cell proliferation, immunoglobulin production, and immunoglobulin class-switching, to the induction of apoptosis (8 , 10, 11, 12) . The biological outcome depends on maturational stage of the B cell (13, 14, 15) and mode of CD40 binding (16, 17, 18) . Similarly, ligation of CD40 produces diverse physiological functions in a multitude of CD40+ cells that include monocyte/macrophages, T cells, dendritic cells, synovial fibroblasts, and renal tubular epithelial cells (19 , 20) .
Although CD40 was originally identified in a bladder carcinoma cell line (21) , information is limited with regard to its pathophysiological role in epithelial tumors. CD40 is expressed in breast and lung carcinomas and carcinomas of the urinary bladder, nasopharynx, and colon, whereas normal nonproliferating tissues are CD40-negative (22) . CD40 ligation induced in vitro growth inhibition and/or apoptosis in bladder, ovarian, cervical, and squamous epithelial carcinoma cell lines, as well as in viral-transformed keratinocyte and fibroblast lines (23 , 24) . Malignant melanoma cells coexpressed CD40 and CD40L, although the pathophysiological role of CD40-CD40L interaction is unclear (25) . CD40L inhibited melanoma cell proliferation and induced apoptosis in vitro (26) , whereas an earlier report suggested that CD40 expression was correlated with a shorter tumor-free survival for malignant melanoma (25) .
Recently, we (27) and others (28 , 29) demonstrated that CD40 binding modulates human breast cancer cell growth, in which binding with a recombinant sCD40L molecule or with a CD40-reactive MAb produced a direct growth-inhibitory effect. This report confirms our initial findings and demonstrates that multiple forms of sCD40L and membrane-bound CD40L produced similar antitumor outcomes. We also document the expression of CD40L in human breast cancer biopsies. Interestingly, we found that breast cancer TILs, unlike normal activated T lymphocytes, rarely express membrane CD40L and, hence, are limited in their capacity to down-regulate CD40+ breast cancer growth through the CD40 pathway.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
HBL-100 is an epithelial cell line obtained from American Type Culture Collection that was derived from the milk of a nursing mother and contained a tandemly integrated SV40 genome and a type D retrovirus genome. We identified that this cell line expressed CD40 and responded to CD40L growth modulation (see "Results"). However, no additional studies were performed with this virally transformed cell line.
Induction of CD40L Expression in PBLs.
For the induction of CD40L expression, peripheral blood mononuclear cells from mismatched donors were obtained by Hypaque-ficoll gradient centrifugation (34)
and then treated with PMA (10 ng/ml) and ionomycin (1 µg/ml) at 37° C for 4 h (35)
. The activated cells were washed (250x g for 7 min), fixed in 1% paraformaldehyde in PBS for 10 min, and washed again before use in coculture studies. CD40L expression was confirmed by flow cytometric direct immunofluorescence analysis, using the CD40L-reactive MAb (clone 89-76) conjugated to PE (Becton Dickinson; San Jose, CA). Two color immunofluorescence analysis with CD3-FITC and CD19-PerCP (Becton Dickinson) indicated that 99% of CD40L+ cells coexpressed the T-cell surface marker CD3.
[3H]Thymidine Uptake Proliferative Assay.
[3H]thymidine uptake was used to determine the proliferative activity of BrCa cells after treatment with gp39, huCD40LT, PBLs, or CD40L-transfected L cells. The optimal concentration for each lot of gp39 and huCD40LT was predetermined by B-cell costimulatory and myeloma growth inhibitory assays (16)
. BrCa cells growing at logarithmic growth phase were trypsinized and resuspended in growth media at 2 x 105cells/ml. One hundred µl of the cell suspension were dispensed onto a 96-well plate. Replicate (three each) samples were incubated with 100 µl of media alone, gp39 (1:8), or huCD40LT (3.0 µg/ml) with or without the leucine-zipper cross-linking antibody M15 (10 µg/ml; Immunex). Alternatively, T47D cells (20,000 cells/well) were dispensed onto a 96-well plate and coincubated with PMA+ionomycin-activated, CD40L-expressing, PBLs seeded at 2 x 104 cells/well. In separate studies, T47D cells were coincubated with irradiated CD40L-L cells (10
, 16)
or untransfected L-cell adherent monolayers (5 x 103 cells/well). For the neutralization of CD40L response, the CD40L blocking MAb LL48 was added to CD40L-L cells or untransfected L cells at the time of culture. PMA+ionomycin-activated PBLs were treated with LL48 for 30 min (23°C) prior to the start of culture. The preoptimized LL48 concentration of 10 µg/ml was used in all of the studies (14)
. In this and previous studies, it has been found that further increase in LL48 (20 µg/ml) did not improve its biological effect. For initial studies with gp39, [3H]thymidine (0.1 µCi/well) was added after 24 or 48 h (37°C, 5% CO2/95% air). In subsequent studies, breast cancer cells were incubated with huCD40LT for 24 h. Thymidine uptake was determined by liquid scintillation counting after overnight incubation. Growth inhibition was determined by comparing the level of radioisotope uptake (counts per min, or CPM) with untreated cultures by the formula:
 |
Immunohistochemical Analysis of CD40 and CD40L Expression.
CD40 expression was determined by the immunoperoxidase technique (36)
. Briefly, the cytospins were fixed in 1% hydrogen peroxide/methanol followed by blocking with normal horse serum. Immunoperoxidase staining involved sequential 60-min incubations at room temperature with the MAb G28.5 (10 µg/ml) or control mouse immunoglobulin, secondary antibody (biotinylated horse antimouse IgG; for 30 min), avidin-biotin-conjugated horseradish peroxidase (Vectastain ABC Elite kit; Vector Laboratories, Burlingame CA; for 35 min), and substrate [3-amino-9-ethylcarbazole (AEC; Biomeda, Foster City CA) or DAB (Vector Laboratories)] with appropriate washings in between. The preparations were counterstained with Meyer’s hematoxylin. The frequency of CD40-expressing cells was determined by light microscopy based on the enumeration of 200 cells. For determinations with patient tumor biopsies, formalin-fixed, paraffin-embedded specimens were dewaxed, rehydrated, and quenched in 1% hydrogen peroxide/methanol. The tissue was blocked with normal horse serum followed by incubation with the CD40-reactive MAb89 (1 µg/ml) or control mouse immunoglobulin for 1 h. Subsequent staining steps were performed according to the procedure described above for cytospin samples. The sections were then dehydrated and mounted in nonaqueous mounting media.
CD40L expression in primary tumor specimens was determined by the highly sensitive TSA immunohistochemistry technique (TSA-Indirect kit; NEN Life Science Products, Boston MA). The paraffin-embedded tissue was dewaxed, rehydrated, and quenched in 3% hydrogen peroxide/methanol, then blocked for 30 min. Incubation was carried out for 60 min with the CD40L antibody D-19 (0.5 µg/ml; Santa Cruz Biotechnology, Santa Cruz CA), known to react with CD40L in paraffin-embedded tissues. The tissue was then reacted with a biotinylated rabbit antigoat secondary antibody (Dako, Carpinteria CA; 2 µg/ml, 30 min), streptavidin-horseradish peroxidase (SA-HRP; 30 min), biotinyl tyramide (5 min), and another incubation with SA-HRP (30 min) with washings in between. The substrate DAB (Vector Laboratories) was added (3–5 min), followed by counterstaining with hematoxylin. The tissue was dehydrated and mounted in nonaqueous mounting media. Controls included the use of normal goat IgG as a primary antibody, or CD40L staining without biotinyl tyramide amplification. In addition, PMA+ionomycin-activated PBLs were used as positive controls in initial evaluations. Under optimized conditions, membrane and cytoplasmic staining with the D-19 was demonstrated in 69% of PMA+ionomycin-activated peripheral blood mononuclear cells, whereas control IgG-reacted samples were uniformly negative. Reactivity with CD40 or CD40L was graded by light microscopy by two independent, trained observers (A. W. T., G. N.) based on staining intensity (neg, no staining; +, weak staining; 2+, moderate staining; or 3+, strong staining) and the frequency of reactive cells (rare, <5%; focal, 5–50%; or diffuse, >50%).
RT-PCR.
CD40 and CD40L cellular mRNA was detected by the RT-PCR reaction (16)
and reverse-transcribed cDNA. Cellular RNA was extracted from cells with the TRIZOL reagent (Life Technologies, Rockville MD). Total RNA was annealed to a sequence-specific 3' primer and reverse-transcribed with MuLV reverse transcriptase (GeneAmp RNA PCR kit; Perkin-Elmer, Foster City, CA). The reverse-transcription products and PCR reagents were kept at the primer annealing temperature before mixing and the addition of the 5' and 3' oligodeoxy-nucleotide primers. Amplification was carried out with AmpliTaq DNA polymerase (denaturing: at 95°C for 1 min and at 94°C for 1 min; annealing: at 56° or 60°C for 45 s; elongation: at 72°C for 1 min; all for 35 cycles). Amplified DNA of the expected size was identified after agarose (1.5%) electrophoresis and ethidium bromide staining. The primer sequences (upstream, 5'-AGAAGGCTGGCACTGTACGA-3'; downstream, 5'-CAGTGTTGGAGCCAGGAAGA-3') corresponded to residues 365–384 and 769–788, respectively, of the CD40 message and produced a 425-bp CD40-specific DNA amplification product. Concomitant RT and amplification of mRNA for the housekeeping gene ß-actin mRNA were carried out as an internal positive control.
For detection of the CD40L message in patient tumor biopsies, RNA was extracted from formalin-fixed, paraffin-embedded tissue blocks with the Paraffin Block RNA Isolation kit (Ambion, Austin TX). Tissue sections (20 µm each; 2 sections/case) were deparafinized with xylene (1 ml, 20 min), washed with 100% ethanol (three changes), air-dried, and then digested with proteinase (50°C, 4 h). The samples were washed, suspended in RNA extraction buffer (from kit) for 5 min, treated with acid phenol:chloroform (5 min), and pelleted by centrifugation (12,000 x g, 5 min). The aqueous phase was transferred and mixed with 1 µl of linear acrylamide and an equal volume of isopropanol (overnight incubation, -20°C). The RNA was repelleted (13,000 x g, 18 min, 4°C) and washed with 75% alcohol (500 µl, 13,000 x g, 8 min, 4°C), then air-dried and resuspended in 10 µl of RNA storage solution. To eliminate DNA, the sample was treated with 1 µl of DNase (2 units/µl) in nuclease-free H2O and DNase I reaction buffer (from kit) for 15 min at 37°C. RNA extraction was repeated with acid phenol:chloroform. The RNA was reprecipitated with isopropanol, linear acrylamide, and 3 M sodium acetate (pH 4.5) for 30 min at -20°C, washed with 75% alcohol, air-dried, then resuspended in 6 µl of RNA storage solution and stored at -20°C. For the RT-PCR reaction, total RNA was annealed to random hexamers and reverse transcribed with MuLV reverse transcriptase (GeneAmp RNA PCR kit; Perkin-Elmer). The reverse-transcription products and PCR reagents were kept at 4°C. Amplification was carried out with AmpliTaq DNA polymerase with an initial denaturation step at 94°C (3 min), followed by 45 cycles of denaturation (94°C, 30 s), annealing (57°C, 30 s), and elongation (72°C, 30 s). The reaction was terminated after a final elongation step (72°C, 7 min). The amplification product was resolved by agarose (2%) electrophoresis. The primer sequences (upstream, 5'-AATTGCGGCACATGTCATAA-3'; downstream, 5'-GTTTCCCATTTTCCAGGGTT-3') corresponded to residues 384–403 and 480–499, respectively, of the CD40L message and produced a 116-bp CD40L-specific amplification product. Concomitant RT and amplification of mRNA for the housekeeping gene ß-actin mRNA were carried out as internal positive control. Dideoxy cycle sequencing reactions were carried out as described previously (16) to verify the molecular identity of the RT-PCR amplification product for the CD40L message. The samples were run on a 310 Genetic Analyzer (Perkin-Elmer). The sequencing data were analyzed bidirectionally and compared with the known sequence for human CD40L cDNA (30) .
Quantification of Apoptosis by the TUNEL Assay.
A TUNEL method was used to detect DNA fragmentation, with the use of the 96-well apoptosis detection kit TiterTACS (R&D Systems, Minneapolis MN). One hundred µl of T47D cells (1 x 106 cells/ml) were dispensed to each well of a 96-well plate and were reacted with Fas antibody CH-11 (4 µg/ml) or gp39 (1:8) for 24 h. The reactants were washed with PBS (1000 x g, 6 min), fixed with 200 µl of 3.7% buffered formaldehyde solution (10 min), then treated with 100% methanol (20 min) with washings in between. The cells were permeated (CytoPore, 50 µl; 30 min, 23°C), followed by washing with distilled water. Endogenous peroxidase activity was quenched by the addition of H2O2 (2.5%, 5 min), washed with distilled H2O, then treated with Tdt (5 min) and the labeling reaction mix (1 h, 37°C). The cells were treated with a stop buffer (5 min), then streptomycin-horseradish peroxidase (10 min, 23°C), and washed with PBS-0.1% Tween 20. One hundred µl of TACS-Sapphire were added (5 min in the dark), and the reaction was stopped with the addition of 2 N HCl. The colorimetric reaction was quantified as a function of absorbance at 450 nm (SpectraMax 340, Molecular Devices, Sunnyvale, CA). The maximal level of apoptosis was determined by adding 50 µl of TACS-nuclease to T47D cells (1 h, 37°C). Cells treated with the labeling reaction mix without the Tdt enzyme served as a negative control.
Flow Cytometric Detection of CD40 Expression and Annexin V Binding.
Direct immunofluorescence analysis was used to determine the frequency of CD40-expressing breast cancer cell lines. Breast cancer cell lines at logarithmic growth phase were trypsinized for 15 (BT-20, HBL-100) or 30 min (T47D, ZR-75–1, MCF-7), washed, and resuspended at 2.5 x 105 cells/ml in PBS containing 1% BSA. Five x 105 cells were pelleted, resuspended in 80 µl of PBS, and incubated with 20 µl of the PE-conjugated CD40 antibody MAb89 (12.5 µg/ml) or control mouse MAb for 40 min at 4°C in the dark. The reactants were washed three times with cold PBS+1% BSA, and resuspended with 500 µl of 1% paraformaldehyde in PBS. The frequency of CD40-positive cells was determined by fluorescence-activated cell sorting (FACS) flow cytometric analyses of 5000 events (16)
and CELLQuest software analysis (Becton Dickinson). Concomitant analysis of the multiple myeloma cell line RPMI 8226 was carried out as a positive control, in which >90% cells were found to be CD40 positive (16)
.
Confirmation of apoptotic activity was carried out by quantifying annexin V binding cells (Apoptosis Detection kit; R&D Systems). T47D cells were seeded in a 6-well plate (1 x 106 cells/well) overnight, then treated with medium alone, gp39 (1:4), and/or cyclohexamide (50 µg/ml; Sigma, St. Louis, MO) for 24 h. The cells were trypsinized, resuspended with medium, washed twice in cold PBS, then reconstituted to 1 x 106 cells/ml in binding buffer (from kit). One x 105 cells (100 µl) were treated with prediluted fluorescein-labeled annexin V (ANN) and propidium iodide (PI; 10 µl each, 15 min at 23°C) in the dark. Binding buffer (400 µl) was added to each sample prior to two-color immunofluorescence analysis (Becton Dickinson FACScan) with an excitation wavelength at 488 nm. The frequency of ANN+, PI+, and double-positive cell fractions was determined by the proportion of cells with green, red, and double-fluorescent cells in 5000 events, based on CELLQuest software analysis (Becton Dickinson).
Multiprobe RNase Protection Assay.
The RNase protection assay (RPA III; Ambion) was used to characterize mRNA expression of the Bcl-2 family of genes (bcl-xL/S, bfl-1, bik, bak, bax, bcl-2, and mcl-1). Radiolabeled probe to each specific message was synthesized using the hAPO-2 Multi-Probe Template Set (PharMingen, San Diego, CA) and the MAXIscript T7 Transcription kit (Ambion). T47D cells were seeded in a 60 x 15-mm tissue culture dish and allowed to adhere overnight. The cells were treated with medium alone, gp39 (1:4), or gp39+LL48 (10 µg/ml) for 24 h. Cellular RNA was extracted by TRIZOL (Life Technologies, Rockville, MD), and incubated with the radiolabeled probes (7 x 105 cpm) overnight, followed by digestion of the unprotected fragments with RNase (30°C, 45 min). The protected fragments were precipitated with an inactivation/precipitation solution (provided in kit), diluted in loading buffer (95% formamide, 0.025% xylene and bromophenol blue, 18 mM EDTA, 0.025% SDS), and resolved by agarose gel electrophoresis according to the manufacturer’s protocol (Quick Point gel; Novex, San Diego, CA). mRNA expression was quantified by densitometric gel band analysis of the autoradiograph. The absorbance for each gel band was determined using the software Scion Image (Scion Corporation) and was normalized to the corresponding densitometric reading for the housekeeping gene L32.
Western Blot Analysis.
Western blotting was used to characterize Bcl-2 gene family protein expression. BrCa cells at logarithmic growth phase (2–5 x 106 per treatment) were treated with medium alone, gp39, or gp39+LL48 for 24 h. Total cell protein lysate was extracted with 0.3–0.4 ml RIPA buffer (PBS with 1% NP40, 0.5% sodium dioxycholate, 0.1% SDS) containing the ATPase inhibitor sodium orthovanadate (184 µg/ml; Sigma), the protease inhibitors phenylmethanesulfonyl fluoride (100 µg/ml; Sigma) and aprotinin (30 µl/ml; Sigma) at 4°C. The protein concentration was determined by the Coomassie Blue (Pierce, Rockford IL) spectrophotometry (SpectraMax 340). Protein extract (18.5–12.2 µg) was mixed at 1:2 (v/v) with electrophoresis sample buffer (1.0 ml of glycerol, 0.5 ml of ß-mercaptoethanol, 3.0 ml of 10% SDS, 1–2 mg of bromophenol blue), boiled at 95°C for 3 min, loaded onto a 12% Tris-HCl gel (Bio-Rad, Hercules CA), and electrophoresed for 90–180 min at 100 V (Mini-Protean II; Bio-Rad) alongside a Kaleidoscope molecular size reference marker (Bio-Rad). The samples were transblotted to a nitrocellulose membrane (18 h, 15 V, 4°C), then soaked in 7% milk in TBS and 0.1% Tween 20 for 1 h to block nonspecific binding. Each sample was treated with the appropriate primary antibody (Santa Cruz; 1:10 final dilution, 23°C, 1 h), horseradish peroxidase secondary antibody (1 h; 1:4000 for Bcl 2, and Bax; 1:1000 for ß-actin; suspended in 7% milk, 0.1% Tween 20 in TBS) with washings with TBS, 0.1% Tween 20 in between. The membrane was developed by chemiluminescence (Luminol ECL Reagent; Amersham Life Sciences, Buckinghamshire, England) per manufacturer’s specifications, and visualized by autoradiography. Semiquantitative analysis was based on densitometric measurements of the reactant bands and normalization to the ß-actin reaction in the same lane. Duplicate values differed by 
10%. Gel bands with A ratios (Aprotein:Aß-actin) that differed by 
25% were arbitrarily considered to be significantly different.
BT-20 Xenografts in SCID Mice.
The in vivo antitumorigenic effect of CD40L was examined using a BT-20 xenograft model in SCID mice (C.B-17/IcrHsd-scid; Harlan Sprague Dawley, Indianapolis IN). BT-20 cells at 
70% confluence were treated for 24 h with medium alone, gp39 from transfected CHO cell cultures, gp39 and the blocking CD40L antibody LL48, control culture supernatant from untransfected CHO cell cultures, or a trimeric soluble recombinant CD40L purchased from Alexis Biochemicals (0.1 µg/ml) in absence of any additional cross-linking agent. Cells were collected after treatment, washed once, then resuspended in RPMI 1640 (1 x 108 cells/ml). Each cell suspension (0.2 ml) was injected into the right flank of each mouse with a 1-ml syringe fitted with a 25-gauge needle. The mice were examined twice weekly for tumor emergence. Tumor size was measured for up to 2 months post-tumor emergence with a Vernier caliper. Mean tumor diameter (D) represents the geometric mean [D = 
(D1D2)] of two perpendicular measurements (D1, D2). The mice were killed and examined for overt tumor metastasis. The primary tumors were also weighed.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Similarly, >80% of total cells in T47D and BT-20 lines expressed CD40 on their membrane surface as defined by immunohistochemical analysis, whereas MCF-7 and ZR-75–1 lines were CD40 negative (<5% positive cells). A 425-bp RT-PCR product corresponding to nucleotides 365–788 of the mature CD40 message was detected only in T47D and BT-20 cells (Fig. 1)
. Concomitant analysis indicates that the virally transformed mammary epithelial line HBL-100 also expressed CD40. These findings indicated that T47D, and BT-20 lines constitutively expressed CD40 and were suitable for the study of CD40 growth regulation.
|
|
The growth inhibitory outcome with optimized concentrations of gp39 and huCD40LT is shown in Table 1
. For gp39, no inhibition was observed at a final dilution of 1:32. gp39 (1:8) significantly lowered [3H]thymidine uptake of CD40+ T47D cells by 39.8% (P < 0.001; n = 5; Table 1
) and BT-20 cells by 40.3% (P < 0.01; n = 3) after 48 h of culture. There was no significant increase in growth inhibition with a higher concentration (1:2) of gp39 (T47D: 48.3 ± 14.6%, n = 7; BT-20: 47 ± 13.5%, n = 4). Subsequent analysis showed that comparable levels of growth inhibition were attained within 24 h (38.5 ± 7.1% for T47D cells and gp39 at 1:8; P > 0.05 as compared with 48-h cultures). A comparable level of growth inhibition (37%) was observed with the virally transformed, CD40+ mammary epithelial line HBL-100 at 24 h. The same treatment did not significantly alter [3H]thymidine uptake of CD40-negative MCF-7 and ZR-75–1 cells at 24 or 48 h (P > 0.05). The trimeric huCD40LT was used at preoptimized concentration (29
, 32)
, which was validated in this laboratory to be 3.0 µg/ml. Like gp39, huCD40LT selectively inhibited the growth of CD40+ breast cancer cells (28.5% for T47D; 33.8% for BT-20 cells; Table 1
) in the presence of a cross-linking MAb, M15, at 24 h. In the absence of MAb M15, trimeric CD40L nevertheless induced significant, albeit a lower level of, growth inhibition of T47D cells (22.0 ± 7.3% at 24 h).
Growth Inhibition by Membrane-bound CD40L.
The presence of TILs and proinflammatory cells has been correlated with a favorable prognosis for early-onset breast cancer (37
, 38) . The growth inhibitory effect by exogenous sCD40L raises the possibility that activated CD40L-expressing lymphocytes may participate in antitumor activity via ligation of CD40 on breast cancer cells. We examined this hypothesis by characterizing the antitumor efficacy of normal peripheral blood donor lymphocytes after induction of CD40L expression with PMA+ionomycin (35)
. Fifty % of peripheral blood mononuclear cells expressed CD40L after this treatment, of which 99% coexpressed the T-cell marker CD3 as defined by flow immunophenotyping analyses.
Treatment with paraformaldehyde-fixed, CD40L+ PBLs from three separate HLA-mismatched, healthy donors resulted in significantly decreased [3H]thymidine uptake of the CD40+ T47D breast cancer cells by 39.7 ± 3.7% (P < 0.01; n = 3) at 48 h of culture, whereas unactivated lymphocytes from the same donors did not affect breast cancer cell growth (13.8 ± 12.0%; P > 0.05; n = 3). Similar growth inhibition was observed in three separate experiments by coincubation with irradiated murine L cells that constitutively express membrane-bound CD40L but lacked other surface coactivation molecules that may be expressed on PMA+ionomycin-activated lymphocytes. T47D [3H]thymidine uptake was decreased by 55.0 ± 8.8%, (P < 0.001; n = 3), whereas untransfected L cells did not (13.3 ± 11.9%; P > 0.05; n = 3). The growth inhibitory effects by CD40L T lymphocytes or CD40L-expressing L cells were abrogated by pretreatment (for CD40L T cells) or coincubation (for CD40L-L cells) with the CD40L blocking antibody LL48. The results of one of three such experiments are shown in Fig. 2
. LL48 alone did not affect breast cancer cell growth (P > 0.05; n = 3). Together with the finding that multiple HLA-mismatched donors can initiate this CD40-dependent growth-inhibitory event, our observations indicate that peripheral blood T lymphocytes possess the capability of inducing breast cancer growth inhibition via CD40 ligation, an event that is independent of coactivation molecule interaction and HLA compatibility.
|
. Similar studies were carried out with a soluble, trimeric rhsCD40L purchased from Alexis, without further cross-linking per manufacturer’s recommendations. Tumor xenografts developed in all untreated (n = 5) and mock-treated animals (n = 8), and were 10.2 ± 0.7 mm and 6.7 ± 1.5 mm in diameter, respectively, at day 55 postinoculation. By comparison, mean tumor diameter of rhsCD40L- and gp39-treated groups was 3.4 ± 2.4 mm (n = 4) and 3.6 ± 1.4 mm (n = 8), respectively, at the same time point. Xenograft growth was not evident in one of four mice that received rhsCD40L-treated BT-20 inoculates, and in one of eight animals with gp-39-treated cells. Cotreatment with the CD40L-blocking antibody LL48 abrogated the antitumor effect of gp39 treatment, during which five of five animals developed tumors with a mean diameter of 7.1 ± 1.9 mm (day 55) that did not differ significantly from the untreated group.
|
as determined by regression analysis (r, 0.75; P < 0.001; n = 28). The mean tumor weights of sCD40L-treated animals (139.7 mg) and gp39-treated animals (106.5 mg) were significantly reduced (P < 0.01), whereas gp39+LL48-treated tumor weights (362.8 mg) did not differ significantly from untreated controls (536 mg; P > 0.05). At necropsy, local invasion into adjacent tissues was observed in 3 of 18 animals that received either untreated, mock-treated, or gp39+LL48-treated tumor cells. No local invasion was observed in animals that received gp39- or sCD40L-treated xenografts (n = 12). These findings are in keeping with the hypothesis that CD40 ligation produces an antitumorigenic effect, resulting in up to 67% decrease in xenograft growth.
Quantification of Apoptotic Activity.
For CD40-transfected carcinoma cells, the ligation of this constitutively expressed growth receptor resulted in growth inhibition and apoptosis (24)
. Quantification of apoptotic activity by an ELISA-based TUNEL assay showed that gp39-treated T47D and BT-20 cultures had an elevated apoptotic activity that corresponded to 31 and 27% of maximum level, respectively, as compared with 56 and 65% of maximum level after treatment with agonistic Fas antibody CH-11, respectively (Table 2)
.
|
Alterations in Bcl-2 Family of Proteins after CD40 Binding.
Breast cancer cells that undergo chemotherapy or hormone-induced apoptosis display unique patterns of proapoptotic and antiapoptotic expressions (41
, 42)
. The involvement of Bcl-2 family of proteins in CD40-induced apoptosis was characterized with a multiprobe RNase protection assay and Western blot analysis. A multiprobe template set allows simultaneous detection of multiple RNA species in a given sample (bcl-xL, bcl-xS, bfl-1, bik, bak, bax, bcl-2, mcl-1, L32, GAPDH; RiboQuant hAPO-2 assay). We identified detectable RNA levels for antiapoptotic elements bcl-xL and mcl1, and the proapoptotic elements bcl-xS, bak, and bax in untreated T47D cells (Fig. 4A)
, whereas endogenous bcl-2 expression was absent. Growth-inhibitory doses of gp39 correspondingly induced a significantly up-regulated level of proapoptotic bax (42% increase as compared with untreated) and bak (33.3% increase), based on gel band densitometric analysis after normalization to the housekeeping gene L32 (Fig. 4A)
. This outcome was not observed in cell cultures cotreated with gp39 and LL48 (8 and 12% increase, respectively, for bax and bak, as compared with control; Table 3
). By comparison, mRNA levels for mcl-1 was not significantly altered. Bcl-xL mRNA was also increased after gp39 treatment, but the increase cannot be attributable to a CD40L-specific effect because of its collaterally increased level in LL48+gp39-treated cultures (Fig. 4A)
.
|
|
. The proapoptotic protein Bax was significantly elevated (95% for T47D; 50% for BT-20) after gp39 treatment but not in cultures that were treated with gp39 in the presence of LL48 (Fig. 4C)
. Consistent with RNase protection assay findings, CD40-positive T47D and BT-20 cells both lacked Bcl-2 expression before or after gp39 treatment, whereas Bcl-2 was expressed endogenously in the CD40-negative line ZR75–1 and was unaffected by gp39 treatment (Fig. 4B)
. These findings indicate that CD40 growth inhibition correlated with an increased apoptotic activity in breast cancer cells and was accompanied by up-regulation of the proapoptotic bax.
CD40 and CD40L Expression in BrCa Patient Biopsies.
The in vivo growth-inhibitory outcome by exogenous CD40L suggests that endogenous expression of this molecule may potentially serve a growth-regulatory role in human breast cancer. Currently, CD40L expression is undefined within the breast tumor microenvironment. Retrospective immunohistochemical evaluation was carried out to characterize CD40 and CD40L expression in formalin-fixed, paraffin-embedded tumor biopsies (Table 4)
. Archived primary or metastatic tumor specimens were selected at random from patients with diagnosed breast cancer in 1995. Of the 12 selected biopsy specimens, 5 contained areas of infiltrating carcinoma as well as carcinoma in situ by histopathological criteria (Table 4)
. CD40 was expressed unequivocally in 12 of 12 cases, including infiltrating ductal (5 of 5 cases tested) and lobular (4 of 4 cases) subtypes, carcinomas in situ (6 of 6 cases), mucinous carcinoma (1 of 1 case), and 1 case of ductal breast cancer metastatic to the lymph node. Ten of 12 cases tested contained >50% of CD40+ breast cancer cells (Table 4)
. There was complete concordance with respect to the pattern of CD40 expression in all five of the cases that coexpressed infiltrating and in situ tumors, although intensity of CD40 staining was noticeably higher in areas of infiltrating carcinoma as compared with regions of in situ carcinoma in two cases (case 958 and 150). CD40 expression was also observed in the cytoplasm of benign ductal epithelium of biopsy specimens containing CD40+ breast cancer cells (10 of 10 cases; Table 4
). In all cases, a focal pattern and lighter staining intensity was observed, which suggested a lower level of CD40 expression as compared with the corresponding malignant cells. A proportion of TILs and endothelial cells were also CD40-positive. These observations extend previous findings (22
, 29) and demonstrate that CD40 is commonly expressed in various histological types of breast cancer.
|
. The pattern of CD40L expression was in complete concordance in areas of infiltrating carcinoma and carcinoma in situ for each of the five cases that coexpressed both tumor cell types. By comparison, CD40L expression in benign ductal epithelium was observed in only 2 of 12 cases examined (Table 4)
.
To verify CD40L expression, RT-PCR analysis was carried out with RNA extracted from the paraffin-embedded tissues and sequence-specific primers corresponding to residues 384–403 and 480–499 of CD40L. RNA was successfully extracted from five of six archived tissue blocks tested, as defined by the capacity to generate a 241-bp ß-actin-specific RT-PCR product (Fig. 5)
. An amplification product corresponding to the residues 384–499 of the open reading frame for CD40L mRNA was successfully generated with RNA extracts from all four of the cases with a positive CD40L immunohistochemical reaction, including case 490, 762, and 958 (Lanes 3, 4, and 5; Fig. 5
). For case 511, which did not express the CD40L protein, the CD40L amplification signal (Lane 6) did not exceed background (Lane 3, Fig. 5
). The cDNA product derived from case 958 was further characterized by RT-PCR cycle sequencing. The bidirectionally analyzed sequence of 5'-AATTGCGGCA CATGTCATAA GTGAGGCCAG CAGTAAAACA ACATCTGTGT TACAGTGGGC TGAAAAAGGA TACTACACCA TGAGCAACAA CTTGGTAACC CTGGAAAATG GGAAAC-3' was identical to residues 384–499 of the CD40L cDNA open reading frame (30)
. These findings confirm the expression of CD40L in the limited number of primary breast cancer biopsies tested.
|
). CD40L may represent a previously unidentified activation marker expressed by TILs (43)
. The coexpression of CD40L and CD40 is suggestive of an endogenous CD40-growth regulatory mechanism within the breast tumor cell population. Our in vitro study indicated that PMA+ionomycin-activated, CD40L+ T lymphocytes can produce a CD40-dependent antitumor effect. However, this activity may be restricted among patient TILs in view of their limited CD40L expression. |
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our study independently confirmed the recent findings of Wingett et al. (28)
and Hirano et al. (29)
that CD40L binding directly inhibits human breast cancer cell growth. Wingett et al. (28)
showed that IFN-
preactivation was required for a significant inhibitory effect. Our data are consistent with the findings of Hirano et al. (29)
that IFN- was not a prerequisite for CD40L-induced growth inhibition, although an enhancing effect by this and other cytokines cannot be excluded (29
, 47)
. Different molecules that ligate CD40 may produce different biological outcomes (16
, 29)
. We have used multiple sources of CD40L to validate the direct growth-inhibitory effect, including gp39, a CD40L-CD8 recombinant molecule (30)
, the trimeric human CD40L/leucine-zipper fusion protein huCD40LT (32)
and the rhsCD40L of similar construct without a leucine-zipper (33)
. The use of CD40-binding antibodies was excluded in this study, because these antibodies may bind to epitopes that are distinct from those involved in natural CD40-CD40L interaction and trigger a correspondingly different growth signal (10
, 15
, 48)
. Discrepancies also have been demonstrated with respect to in vitro and in vivo growth-inhibitory effects by CD40 MAbs, which may be explained in part by the collateral activation of antibody-dependent cellular cytotoxicity in vivo (29)
. By comparison, we observed a uniform growth-inhibitory effect on CD40-positive breast cancer cell lines by different constructs of sCD40Ls. Conversely, CD40L was ineffective in altering CD40-negative breast cancer cell growth, supporting the critical role of CD40 ligation in this growth-regulatory event.
Both gp39 and sCD40L trimer produced significant, albeit incomplete, tumor growth inhibition in vitro. This outcome translated into reductions of in vivo human BrCa xenograft growth of 65 and 67%, respectively. Cotreatment with the CD40L-blocking antibody LL48 abrogated this antitumor effect, confirming CD40-dependence of this phenomenon. Although CD40L effectively reduced xenograft progression, it appeared to have a limited capacity to prevent tumor formation. Twenty-five % of animals that received CD40L-treated inoculates (and 12% that received gp39-treated inoculates) failed to develop xenografts. This distribution did not differ significantly from the untreated control group. Additional studies are needed to better characterize the antitumorigenic properties of CD40L. Recent studies have examined the immune-activating effect of CD40L in controlling tumor cell growth (44, 45, 46 , 49) . Our demonstrated direct growth-inhibitory effect in vitro, the washing of CD40L-treated tumor cells prior to inoculation, and the immunocompromised phenotype of the nu/nu mouse argue against functional B- or T-cell involvement in the observed in vivo antitumor response. However, potential CD40L activation of monocyte/macrophages cannot be excluded, which in turn may produce antitumor cytokines and/or activate natural killer cells (38 , 50 , 51) . Additional studies are planned to characterize the effect of CD40L treatment on a CD40-nonexpressing cell line, to identify or exclude the contribution of immune-activation in CD40L-mediated breast cancer cell growth inhibition in vivo.
We found that CD40-dependent growth inhibition triggered events characteristic of apoptosis in breast cancer cells, including exposure of intramembrane phosphotidylserine and increased DNA strand breaks. The level of CD40L-induced apoptosis to be one-half of that induced with the agonistic Fas antibody CH-11 and of a similar order of magnitude as growth inhibition measured by [3H]thymidine uptake. These observations are indirect evidence that apoptosis contribute significantly to the CD40L-induced growth inhibitory process. Studies of Binder et al. (52)
, and Rochaix et al. (53)
showed that 45–75% of BrCas expressed Bax. Our findings of endogenous Bax expression in CD40+ T47-D and BT-20 cells are in keeping with similar observations by others (54)
. Furthermore, there was a significant up-regulation of this proapoptotic molecule at mRNA and protein levels after CD40 ligation, whereas the CD40L blocking antibody abrogated apoptotic induction and the up-regulation of Bax. Thus, CD40-binding may alter the balance of Bcl-2 family members of apoptotic proteins and trigger apoptosis. Our limited study indicates that the CD40-negative ZR-75–1 cells expressed the antiapoptotic molecule Bcl-2, whereas the CD40+ lines T47D and BT-20 did not. Decreased endogenous Bcl-2 expression was correlated with an increased apoptotic index for breast cancer cases (53)
. However, these CD40-expressing tumor lines did not appear to be more prone to apoptotic induction by the agonistic Fas antibody CH-11. Additional studies are needed to further characterize the mechanism of CD40-triggered apoptosis, including the possible downstream up-regulation of other TNF-receptor family members (such as Fas, or TNF-related apoptosis-inducing ligand) and/or corresponding caspase activation(s) (55, 56)
. In view of the considerably stronger antitumorigenic effect of CD40L in vivo (as compared with measured apoptotic level in vitro), other growth-modulatory mechanisms may directly or indirectly contribute to CD40L-induced breast cancer growth inhibition.
In exploring the pathophysiological relevance of CD40L-induced growth inhibition, we have examined CD40 and CD40L expression in patient primary tumor biopsies. We found that all of the breast cancer subtypes tested expressed CD40, including infiltrating ductal and lobular carcinomas and carcinomas in situ. It is of interest to note that benign epithelial tissues of these biopsies also exhibited focal, albeit weaker, expression of CD40. These CD40+ benign ductal epithelial tissues are amid or surrounding the malignant tissue and are likely to represent proliferative, albeit benign, epithelial tissues that have been shown previously to express CD40 (22) . Additional evaluations of normal, fibrocystic, and hyperplastic specimens are needed to properly define the pattern of CD40 expression in breast epithelium in various states of tumor pathogenesis.
Focal areas of the tumor populations studied also expressed CD40L, which was rarely expressed in benign ductal epithelial tissues in the same cases examined. Only CD40L-positive cases produced a RT-PCR amplification product corresponding to residues 384–499 of CD40L, as verified by dideoxy cycle sequencing. To our knowledge, this is the first report that documents endogenous CD40L expression in BrCa cells. Because only a limited number of cases were examined, further analysis is needed to characterize the relationship, if any, between CD40L expression and tumor grade. The coexpression of CD40L and CD40 is suggestive of a feedback growth loop within the breast cancer population. CD40L was detected primarily within the cancer cell cytoplasm with rare membrane staining, which suggests that expression of the membrane-bound CD40L may be uncommon or occurs transiently (and, hence, escapes detection). Alternatively, cytoplasmic CD40L may represent a precursor of a soluble, biologically active CD40L (57) , which has been detected in patients with hematological malignancies (58) . In view of the presence of malignant disease despite the potential growth-inhibitory effect of CD40L, we speculate that the CD40L-expressing breast cancer cells may participate in a dynamic, localized growth interactive process. Alternatively, CD40L may represent the outcome of cytokine cascade activation in the tumor microenvironment (51) . Prior to the testing of these hypotheses, studies are needed to establish the biological activity of CD40L from patient breast cancer cells.
Our in vitro findings indicate that growth inhibition by CD40L+ activated peripheral blood T lymphocytes was mediated at least in part by CD40L-CD40 interactions. Coincubation with LL48 partially abrogated growth inhibition by PMA+ionomycin-activated PBLs. The less than complete reversal may be attributable to the collateral induction of other members of the TNF family, or other molecules that can regulate breast cancer cell growth (59 , 60) . Thus, it is of interest that CD40L was infrequently expressed among TILs in the majority of breast cancer cases tested. These findings suggest that TILs in established breast cancers may lack the ability to down-regulate tumor cell growth via CD40L-CD40 interaction, hence favoring tumor progression. Our current analysis does not allow us to further define the cell type expressing CD40 and/or CD40L among the tumor infiltrating mononuclear cells, although the CD40+ subset is likely to represent infiltrating B lymphocytes, macrophages, and dendritic cells (29 , 61) . The direct antitumor effect by CD40L supports the hypothesis that the stimulation of CD40L expression on breast cancer cells and/or TILs can potentially down-regulate breast cancer cell growth. These approaches are currently being examined in our laboratory. We also plan additional studies with preexisting tumor xenografts that are propagated in a physiologically relevant site such as the mammary fat pad, to confirm the immunotherapeutic potential of sCD40L and/or membrane-bound CD40L. Finally, the relationship of CD40 and CD40L expression with other known surrogate biological prognosticators such as tumor grade and hormonal receptor status (62) will be examined.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported in part by the Tri Delta Cancer Research Fund, the Robert Shanbaum Memorial Fund, and the Edward and Ruth Wilkof Foundation. 
2 To whom requests for reprints should be addressed, at the Cancer Immunology Research Laboratory, Baylor-Sammons Cancer Center, 3500 Gaston Avenue, Dallas, TX 75246. Phone: (214) 820-4123; Fax: (214) 820-2780; E-mail: aw.tong{at}baylordallas.edu
3 The abbreviations used are: TNF, tumor necrosis factor; IL, interleukin; CD40L, CD40 ligand; sCD40L, soluble CD40L; rhsCD40L, recombinant human sCD40L; MAb, monoclonal antibody; PMA, phorbol myristate acetate; PE, phycoerythrin; PBL, peripheral blood lymphocyte; DAB, 3,3'-diaminobenzidine; TSA, tyramide signal amplification; RT, reverse transcription; Tdt, terminal deoxytransferase; TUNEL, Tdt-mediated dUTP nick end labeling; TBS, Tris-buffered saline; SCID, severe combined immunodeficient; BrCa, breast carcinoma; TIL, tumor-infiltrating lymphocyte.
Received 8/29/00; revised 11/27/00; accepted 12/ 8/00.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-induced apoptosis among MCF-7 breast cancer cell variants. Cancer Res., 58: 4940-4946, 1998.
B pathway involving TNF receptor-associated factors. Oncogene, 14: 2889-2916, 1997.
B/Rel expression. Cell Growth Differ., 10: 537-544, 1999.
), gp39 from transfected CHO cell cultures (1:4, 
), gp39+LL48 (10 µg/ml; 
), or the sCD40L trimer (0.1 µg/ml, 
). A, mean tumor diameter versus time post-tumor emergence; tumor diameter at any given time point was calculated as the geometric mean of two perpendicular caliper measurements. SEs were 
