Published online before print November 21, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Max-Planck-Institut für Immunbiologie, Freiburg, Germany
1Correspondence at current address: Institute of Chemistry and Biochemistry, University of Salzburg, Hellbrunner Str. 34, A-5020 Salzburg, Austria. E-mail: gschwamberger{at}yahoo.de
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
150 kDa, closely resembling a novel tumoricidal factor of murine macrophages (M
) termed MTC 170 (M tumor cytotoxin, approximate molecular mass 170 kDa), which we have previously proposed to constitute a major effector pathway for the destruction of tumor cells by activated M.
Key Words: tumor immunology • macrophage • cytotoxicity • cytokines
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) [7
], suggesting M to be the primary source of TNF. Thus, TNF was originally purified from culture supernatants of murine as well as human M-like cell lines [8
, 9
] and was characterized as a 17-kDa polypeptide, the active form being a trimer of identical subunits [10
]. Soon thereafter, murine and human TNF were cloned [11
, 12
], and only afterwards, this molecule was also purified from TNS [13
]. This factor, which was shown to be identical to cachectin [14
] and today, is referred to as TNF-
, is regarded as the central mediator of LPS-induced tumor necrosis and regression in vivo as well as one of the key mediators of M-mediated tumor cytotoxicity [15
].
Whereas TNF- is now known to constitute the master proinflammatory cytokine elicited by LPS and a crucial mediator of cachexia and the toxic shock syndrome in bacterial septicaemia [16
], its role as an endogenous antitumor factor has been challenged by the finding that TNF may also act as a strong tumor-promoting factor in various model systems of experimental carcinogenesis and metastasis (reviewed in ref. [17
]) With respect to the antitumor properties of TNF, particularly concerning the role as the key mediator of the antitumor effects of LPS, some unresolved discrepancies still exist. Thus, the antitumor effects of TNF in vivo have been shown to be primarily indirect, destroying the tumor vasculature instead of the tumor cells and thereby provoking hemorrhagic necrosis of the tumor mass [18
, 19
]. However, the antitumor efficacy of TNF with respect to actual tumor regression, as opposed to hemorrhagic necrosis, has been shown to be markedly lower than the one obtained by treatment with LPS or TNS, the limiting factor being the severe systemic toxicity exerted by TNF [20
, 21
]. Conversely, the direct cytotoxic activity of TNF in vitro is restricted to a fairly small number of highly sensitive tumor cell lines [22
], which is in contrast to the results obtained with activated M [23
]. This suggests that other mediators apart from TNF may be operative in the destruction of tumor cells by activated M or TNS. Some years ago, we started to characterize a novel high molecular weight tumoricidal activity distinct from TNF in culture supernatants of murine as well as human M, which we originally termed MTC 170 (M tumor cytotoxin, approximate molecular mass 170 kDa) [24
25
26
]. Surprisingly, however, this activity displays many features highly reminiscent of the original reports on the tumor-necrotizing activity in TNS [4
5
6
]. Given the discrepancies between these reports and what is known as TNF- today, this prompted us to reinvestigate the antitumor activity in TNS with respect to a potential contribution of MTC 170. In this paper, we now describe the induction of an antitumor activity closely resembling MTC 170 in sera of mice pretreated with heat-killed Propionibacterium acnes (P. acnes) and challenged with LPS as well as a comparative analysis of the antitumor activity of these sera in vitro and in vivo using the classical Meth A tumor model system.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(IFN-) was produced by Genentech Inc. (San Francisco, CA) and kindly provided by Boehringer Ingelheim (Vienna, Austria). Recombinant murine TNF- (r-muTNF-) with a specific activity of 107 units/mg was obtained from Boehringer Mannheim (Germany). Neutralizing antibodies against r-muTNF- were purified by high-pressure liquid chromatography (HPLC) ion-exchange chromatography on a Mono Q HR 5/5 column (Pharmacia, Uppsala, Sweden) from a rabbit antiserum purchased from Genzyme (Cambridge, MA). All cell culture media and supplements were obtained from Biochrom (Berlin, Germany), unless stated otherwise.
Mice
Mice raised under specific pathogen-free conditions at the Max-Planck-Institut für Immunbiologie (Freiburg, Germany) were used at 6–12 weeks of age.
Preparation of TNS
TNS were prepared by intravenously (i.v.) injecting female mice with suspensions of heat-killed P. acnes [28
] (25 µg per g body weight). Seven days later, animals were challenged by i.v. injection of varying doses of LPS and exsanguinated at the times indicated. Blood, pooled from at least three to five individual mice receiving identical treatment, was collected into ice-chilled vials, then transferred to 37°C for 15 min for clotting, and subsequently returned to ice temperature for 1 h to allow clot contraction. Afterwards, sera were separated from clots by centrifugation, and aliquots were stored at -70°C until use. Control sera from untreated mice were prepared in an identical manner.
Murine bone marrow-derived M (BMM) culture and preparation of conditioned culture supernatants
BMM were obtained by in vitro differentiation from bone marrow precursor cells under serum-free conditions as described previously [29
]. After 9 days of differentiation, BMM were harvested and activated by sequential treatment with recombinant IFN- (100 U/ml) and LPS (100 ng/ml) in Iscove’s modified Dulbecco’s medium (IMDM) at a final density of 5 x 105 cells/ml. After 24 h, culture supernatants were harvested, cleared by centrifugation at 400 g for 15 min, and concentrated 1000-fold by ultrafiltration through Amicon YM-10 membranes (Amicon, Bedford, MA).
Target cells
As target cells for determination of cytocidal activity of mouse sera, the following murine cells or cell lines were used. Tumor cell-lines: YAC-1(T-lymphoma), P815 (mastocytoma), EL-4 (thymoma), B16 (melanoma), Meth A (fibrosarcoma), S-180 (fibrosarcoma), LL-3 (carcinoma), and L-929 (fibroblastoma). As nontumorigenic control cells, freshly isolated spleen cells and the 3T3 fibroblastoid cell line were used. All target cells were maintained in serum-free medium consisting of 50% IMDM + 50% HAM’S F12 medium, supplemented with bovine insulin (2.5 µg/ml), human holo-transferrin (5 µg/ml), and bovine thyroglobulin (5 µg/ml), all obtained from Sigma. L-929 cells used for determination of TNF bioactivity were maintained in IMDM supplemented with 10% fetal calf serum.
Assay for cytocidal activity
Cytocidal activity of mouse sera, diluted in HEPES-buffered saline (HBS) was assessed after 24 h incubation, except where indicated, with 2 x 105 target cells per ml in IHM medium (70% IMDM, 20% HAM’S F12, 10% 8 mM MgCl2 in HBS), supplemented with bovine insulin (1 µg/ml), human holo-transferrin (2 µg/ml), and bovine thyroglobulin (2 µg/ml) by a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) dye reduction assay [30
] and/or trypan blue exclusion. Assays were performed in triplicates, and cytotoxicity was determined as percentage reduction of MTT absorbance as compared with controls receiving identical concentrations of pooled normal mouse sera (NMS). Cytotoxicity units were calculated by linear regression analysis of serial sample dilutions, 1 U/ml being defined as the concentration required to achieve 50% cytotoxicity. Cytocidal activity of r-muTNF- was assayed under identical conditions as for TNS.
Assay for TNF-
Cytocidal activity of TNF- was assessed using an L-929 bioassay in the absence of actinomycin D as described previously [25
]. One unit/ml is defined as the concentration of TNF- yielding 50% cytotoxicity as compared with untreated controls. Neutralizing capacity of purified rabbit anti-r-muTNF- antibodies was determined by inhibition of cytocidal activity of r-muTNF- using the L-929 bioassay described above. Neutralizing capacity was 
25 µg (equivalent to 3.0x105 U) per mg antibody for recombinant and natural forms of TNF-.
Additives used in cytotoxicity assays
All additives used in cytotoxicity assays of mouse sera were dissolved in HBS and added before the addition of target cells. Cytotoxicity values were calculated with respect to the appropriate controls receiving only additives in HBS. None of the additives alone showed significant effects on the target cells at the concentrations used.
Fractionation of TNS and BMM culture supernatants
TNS and concentrated BMM culture supernatants were fractionated by HPLC size-exclusion chromatography on a Superose 12 HR 10/30 column (Pharmacia) at a flow rate of 0.25 ml/min with ice-cold charcoal-filtered HBS as the eluent. Fractions (0.5 ml) used for in vitro cytotoxicity assays were sterilized by UV irradiation and were immediately tested for cytocidal activity as described above. Cytotoxicity of TNS and BMM culture supernatant fractions was calculated with reference to corresponding fractions of pooled NMS and HBS, respectively. For tumor regression assays in vivo, pooled, corresponding fractions from 20 consecutive runs were concentrated tenfold using Centriplus-10 ultrafiltration devices (Amicon), and the concentrates were filtered through sterile 0.22 µm Spin-X centrifuge filters (Costar, Cambridge, MA) before i.v. injection into tumor-bearing mice.
Propagation of Meth A tumors and treatment of established tumors in vivo
Meth A fibrosarcoma was propagated by weekly injection of 106 tumor cells into the peritoneal cavity of syngeneic BALB/c mice. For analysis of antitumor activity of TNS, 2 x 105 tumor cells, obtained by peritoneal lavage and washed once in charcoal-filtered, sterile phosphate-buffered saline (PBS), were injected s.c. into naive mice. Tumor-bearing mice were treated after 7 days by a single i.v. injection of 200 µl samples of TNS or HPLC fractions of TNS, prepared as described above. Tumor growth was evaluated by measuring tumor diameters with precision calipers, and tumor cross-sections were calculated according to the formula 
/4 x (axb), a and b representing tumor diameters measured in perpendicular directions. Regression was considered complete when no more tumor tissue was visible at the site of implantation. All tumor experiments were performed under specific pathogen-free housing conditions.
Determination of LPS content of sera and HPLC fractions
LPS content of sera and HPLC fractions used for in vivo tumor experiments was estimated by Limulus assay [31
].
Data presentation
All data values shown are representative of at least three independent experiments, except for in vivo tumor-regression studies, performed independently twice. Evaluation of statistical significance was done by Student’s t-test.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (25K): [in a new window] |
Figure 1. Cytotoxic activity against TNF-resistant YAC-1 tumor cells in sera of mice untreated (- > -) or treated with P. acnes (Pa > -), LPS (S. abortus equi, 1 µg; - > LPS), or P. acnes followed by LPS (Pa > LPS). Sera were collected 6 h after LPS challenge. One unit/ml, defined as the concentration required to achieve 50% cytotoxicity. Values represent means ± SEM from triplicate samples.
|
 View larger version (18K): [in a new window] |
Figure 2. Induction of TNF and cytotoxic activity against TNF-resistant YAC-1 tumor cells in sera of mice treated with P. acnes after challenge with various doses of LPS (S. abortus equi). One unit/ml, defined as the concentration required to achieve 50% cytotoxicity in a 48-h TNF bioassay using L-929 target cells or a 24-h cytotoxicity assay using YAC-1 target cells, respectively. Values represent means from triplicate samples; SEM is below 10% of the data values for all data points shown.
|
|
View this table: [in a new window] |
Table 1. Cytotoxic Activity of Late TNS against Nontumorigenic and Tumorigenic Cells of Murine Origin
|
. As expected from the data described so far, no inhibition of tumoricidal activity by anti-TNF antibodies was observed (Table 2
). The efficacy of anti-TNF antibodies in neutralizing serum TNF was verified by inhibition of TNF bioactivity in early TNS (2 h after LPS challenge), as evidenced by complete abrogation of cytotoxicity on TNF-sensitive L-929 cells. |
View this table: [in a new window] |
Table 2. Sensitivity of Tumoricidal Activity of Late TNS to Anti-TNF Antibodies, Heat, or Complement Inactivators
|
Preliminary biochemical characterization of tumoricidal activity
For initial biochemical characterization of the tumoricidal activity, late TNS was fractionated by HPLC size-exclusion chromatography, and the fractions were assayed for cytocidal activity against several murine tumor cell lines in vitro. As shown in Figure 3a
, cytotoxic activity eluted as a single peak with a molecular mass (Mr) of 150 kDa, giving essentially identical profiles on all target cell lines tested (data not shown). Whereas the 150-kDa peak was found consistently, occasionally, a second but only minor peak at a Mr of 50–60 kDa was also observed. This elution profile of tumoricidal activity closely resembles the profile obtained for a novel tumoricidal factor in culture supernatants of activated murine BMM, displayed for comparison in Figure 3b
, which we have described previously and termed MTC 170 [24
, 25
], suggesting both activities to be due to the same factor.
 View larger version (14K): [in a new window] |
Figure 3. Analysis of cytotoxic activity against YAC-1 tumor cells by HPLC size exclusion chromatography of late TNS [0.5 µg LPS (E. coli), 8 h]; (a) and conditioned culture supernatant of BMM stimulated by consecutive treatment with IFN- and LPS (b). Final dilution of fractions for cytotoxicity assays was 1:5 and 1:2 for TNS and BMM culture supernatant, respectively. A 280 = absorbance at 280 nm. Markers used for Mr calibration were thyreoglobulin, 670 kDa; bovine IgG, 158 kDa; ovalbumin, 43 kDa; and myoglobin, 17 kDa.
|
|
View this table: [in a new window] |
Table 3. Antitumor Effects of LPS, TNF, and Late TNS on Meth A Fibrosarcoma in Vivo
|
 View larger version (47K): [in a new window] |
Figure 4. Effect of treatment with late TNS [0.5 µg LPS (E. coli), 8 h] as compared with PBS (control) and NMS on growth of s.c. implants of Meth A tumor cells in syngeneic BALB/c mice. Tumor cross-sections, calculated as described in Material and Methods (a), and photographical documentation of tumor development (b). All data points displayed represent means from five to six animals. Effective dosage of late TNS was 20 U in vitro activity per mouse. LPS content of all samples used was  0.1 ng/ml.
|
 View larger version (22K): [in a new window] |
Figure 5. Analysis of cytotoxic activity against Meth A tumor cells in vitro and tumor-regressing activity for Meth A tumors in vivo of pooled late TNS by HPLC size-exclusion chromatography. The cytotoxicity assay on Meth A cells was performed for 24 h with an initial cell density of 105/ml. The final dilution of fractions was 1:2. Tumor size regression was calculated as the percentage of reduction of tumor cross-sections in comparison with animals receiving only HBS on day 14 after treatment. All data points displayed represent means from eight to 10 animals (P value for fraction No. 12<0.01; for all other fractions, 0.4). Effective dosage of fraction No. 12 was 24 U in vitro activity per mouse, yielding complete regression in eight out of nine treated animals. Data points depicted as fractions 2, 4, 19, and 22 represent pooled, neighboring fractions. TNF activity of fractions was below the threshold of detection (0.5 U/ml); LPS content of samples used for tumor regression analysis in vivo was 1 ng/ml. A 280 = absorbance at 280 nm; full scale is 1.5 absorbance units. Markers used for Mr calibration were the same as in Figure 3
.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
described by us previously [24
25
26
], which we have originally termed MTC 170 (although the calculated Mr of the peak fraction appears to be closer to 150 kDa under the modified eluting conditions used here). This factor, which we have previously suggested to play a fundamental role in M-mediated destruction of tumor cells, thus apparently may also constitute a key mediator of LPS-induced tumor regression in vivo. The presumed identity of this M-derived factor with the antitumor activity observed in TNS also suggests M to be the main source of this activity in serum, which, however, does not rule out contributions by other cell types responsive to LPS or LPS-induced proinflammatory cytokines.
So, while this activity is clearly distinct from TNF, our data in several aspects, however, are highly reminiscent of the data published on the antitumor activity in TNS in the late 1970s. Interestingly, that the original reports by Carswell et al. [3
], Green et al. [4
], and others thereafter [5
, 6
] described TNF, mainly based on its antitumor activity in vivo, as having a Mr of 150 kDa, with minor activities detected only by in vitro assays on L-929 cells at 225 and 50 kDa. However, upon purification of TNF activity, which was monitored using this highly TNF-sensitive cell line, cytotoxic activity from M sources [8
] as well as TNS [13
] was recovered at a Mr of 50 kDa only, whereas reports on the 150-kDa peak were discrepant [5
, 6
]. As, however, as shown in this paper, a 150-kDa peak of antitumor activity is also present in late TNS, which is completely devoid of TNF, this clearly demonstrates that this 150-kDa peak constitutes a TNF-independent activity, but most probably identical to the antitumor factor originally observed in classical TNS. In fact, we have observed similar amounts of MTC 170-like activity in classical compared with late TNS preparations, which is incompletely resolved from TNF by size-exclusion chromatography (our unpublished data). Taken together, this suggests that this activity may have been lost upon fractionation, probably due to the fact that L-929 cells were used to monitor purification of TNF activity, which, although exquisitely sensitive to TNF, appear to be highly resistant to the action of MTC 170.
That TNF may not be the sole mediator of the antitumor effects of TNS is also suggested by another discrepancy concerning the antitumor properties of TNF versus TNS. As put forward in a short review by Old [44 ], summarizing the discovery and characterization of TNF in 1985, an unexplained difference between TNF and the original observations of Carswell et al. [3 ], apart from the Mr, is the severe systemic toxicity of purified or r-TNF in contrast to TNS. This discrepancy is also highlighted by the work of North and colleagues in 1988 [20 , 21 ], which critically investigates the differences in the antitumor responses elicited by LPS, TNS, and TNF in a fibrosarcoma model system in vivo. Their findings clearly demonstrate that although TNF is responsible for the hemorrhagic necrosis of the tumors observed after LPS treatment, its efficacy to induce actual regression of the tumors, in contrast to LPS and TNS, is rather low, the limiting factor being severe toxicity. Interestingly they also showed that whereas the hemorrhagic response was almost completely inhibited by administration of anti-TNF antibodies together with LPS or TNS, the effect of this treatment on actual regression was only partial, suggesting these phenomena to be distinct in nature. This is in perfect accordance with the data provided here, showing that whereas TNF induced the expected hemorrhagic response along with a cure in our model system, late TNS caused slow but complete regression of tumors in the absence of a significant hemorrhagic response, demonstrating that TNF may not be required for actual tumor regression.
This, however, does not preclude a vital function of TNF in antitumor responses evocated by LPS. Although our data so far have not provided evidence for a direct synergism between TNF and MTC 170 in tumor cell killing in vitro in general, such a synergism may still occur in certain cases, as previously shown by us for the human lymphoma cell line JMP [26
]. Probably more important, however, we recently have obtained definitive evidence for a crucial involvement of TNF in the induction of MTC 170, by murine M and in sera of P. acnes-primed mice (manuscripts in preparation), suggesting TNF to be an endogenous mediator of M activation for MTC 170 secretion. This may also account for part of the antitumor activity of TNF in vivo, thus tightly linking TNF and MTC 170 in terms of a coordinated antitumor response.
In conclusion, our findings demonstrate the induction of a TNF-independent antitumor principle in sera of mice challenged with LPS, which based on all features analyzed so far, is virtually indistinguishable from the M-derived tumoricidal activity MTC 170, although the molecular identity of the responsible factor(s) remains to be formally proven yet. This strongly suggests that what has originally been termed "tumor necrotizing factor" in the very beginning of the TNF story in 1975, actually comprised two entirely different factors with antitumor activity: One, which has come to be known as TNF-, causing the collapse of the tumor vasculature, and a second, independent principle with direct cytotoxic activity on various types of tumor cells, which most likely act in a concerted manner to bring about the phenomenon of LPS-induced tumor regression. While these data may explain the conflicting results concerning the role of TNF in LPS-induced antitumor responses, we are still at the very beginning of understanding the molecular and cellular basis of this novel antitumor activity, as well as potential cross-links to other factors, which will probably have to wait for the availability of larger amounts of purified MTC 170. We are currently trying to further characterize this activity on a molecular basis as a prerequisite for cloning the responsible factor.
was produced by Genentech Inc. and kindly provided by Dr. G. Adolf, Boehringer Ingelheim, Vienna. We thank Hella Stübig, Nadja Goos, and Alexandra Fehrenbach for excellent technical assistance and Dr. H. Mossmann for support during in vivo tumor studies. We also thank Dr. P. Nielsen for helpful discussions and critically reading the manuscript. 
Received January 24, 2003; accepted July 24, 2003.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  I. N. Buhtoiarov, H. D. Lum, G. Berke, P. M. Sondel, and A. L. Rakhmilevich Synergistic Activation of Macrophages via CD40 and TLR9 Results in T Cell Independent Antitumor Effects J. Immunol., January 1, 2006; 176(1): 309 - 318. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
