Cancer Therapy Vol 2, 353-364, 2004

Department of Pharmaceutical Sciences, Bouve College of Health Sciences, Northeastern University, 360 Huntington Avenue, Boston MA 02115, USA

__________________________________________________________________________________

*Correspondence: Vladimir Torchilin, Ph.D., D.Sc., Department of Pharmaceutical Sciences, Northeastern University, Mugar Building, Room 312, 360 Huntington Avenue, Boston, MA 02115, USA; Tel: 617-373-3206; Fax: 617-373-8886; e-mail: v.torchilin@neu.edu

Key words: cancer immunotherapy; human tumors; nude mice; monoclonal antibody; anti-nuclear autoantibody; nucleosome-specific antibody

Abbreviations: antibody-dependent cellular cytotoxicity, (ADCC); antinuclear autoantibodies, (ANAs); counts per minute, (CPM); Diethylene triamine pentaacetic acid anhydride, (DTPA); Hank's balanced salt solution, (HBSS); nucleosome, (NS)

VPT and LZI hold a US patent on antitumor antibodies described in this paper. The patent is assigned to Procyon Biopharma, Inc. (Dorval, Canada) for commercial development.

Recent success with mAbs as cancer therapeutics (Bell, 2002; Boye et al, 2003; Mendelsohn, 2003; Zelenetz, 2003) has renewed the attention to the anti-cancer role of humoral immunity, including natural humoral immunity. Natural anti-tumor antibodies are common for healthy individuals (Colnaghi et al, 1982), and experimental in vivo data on their strong tumor-suppressive activity (David et al, 1996) support their possible role in tumor immunosurveillance. Such antibodies usually belong to the IgM isotype (Brandlein et al, 2003), which might limit their infiltration into solid tumors (compared to smaller IgGs) and restrict their anti-tumor effector function since no antibody-dependent cellular cytotoxicity (ADCC) is known for IgMs. In addition, the applicability of natural anticancer antibodies may be limited by their restricted specificity against only certain tumor types (Cote et al, 1983). We have identified a subset of natural IgG class antibodies capable of binding to the surface of a broad spectrum of various cancer cells but not normal cells (Iakoubov, Torchilin, 1997). These antibodies are natural antinuclear autoantibodies (ANAs) that are present in a proportion of healthy rodents and humans, especially in the aged (Xavier et al, 1995). Our hypothesis that these ANAs participate in tumor immunosurveillance (Torchilin et al, 2001) was strongly supported by the results of pre-clinical in vivo experiments with mAb 2C5, a monoclonal tumor cell surface-reactive nucleosome (NS)-specific ANA of IgG2a isotype obtained from a non-immunized healthy aged Balb/c mouse. Within syngeneic mouse models, mAb 2C5 suppressed the growth of EL4 T lymphoma and led to a prolonged survival time in B16 melanoma-bearing mice (Iakoubov and Torchilin, 1997). Existing data on the tumor reactivity and anti-tumor properties of certain ANAs related to autoimmune pathology or induced by immunization also favor this hypothesis (Brinkman et al, 1993; Johnson, Shin, 1983; Sorace, Johnson, 1990). Numerous additional data supporting the hypothesis on ANAs' role in tumor immunosurveillance have been reviewed recently (Torchilin et al, 2003).

ADCC was considered as a most probable in vivo mechanism for mAb 2C5 anti-tumor activity because of significant ADCC effect of this antibody in vitro (Iakoubov and Torchilin, 1997). Though in vitro experiments did reveal neither complement-mediated cytotoxicity, nor direct inhibition of tumor cell proliferation, some additional mechanisms might be involved in vivo. By forming immune complexes with free NSs in the circulation, mAb 2C5 and similar antibodies might induce the enhanced release of inflammatory cytokines and proteolytic enzymes from certain immune cells (Lucisana and Mantovani, 1984), as well as toll receptor-involving events (Leadbetter et al, 2002), and/or empowerment of dendritic cell (Schuurhuis et al, 2002).

The ability to recognize tumor but not normal cells was observed for mAb 2C5 and another similarly obtained ANA, mAb 1G3. Both mAbs demonstrated NS-restricted specificity. Thus, tumor cell surface-bound NSs (supramolecular constituents of nuclear material consisting of well-characterized individual monoNS composed of DNA and four pairs of histones arranged in a characteristic pattern) were proposed to be their target on tumor cells (Iakoubov, Torchilin, 1997). The binding of extracellular NSs to tumor cell surface might be mediated by specific NS receptors that have been repeatedly reported by several investigators to be present on the surface of tumor cells (Jacob et al, 1989; Koutouzov et al, 1996; Seddiki et al, 2001). As for the origin of tumor cell-bound NSs, extracellular NSs were found in tumor cell cultures (Bell, Morrison, 1991) and in patients with tumors (Le Lann-Terrisse et al, 1994), where they arise from apoptotic tumor cells present in every in vivo developing tumor (Wyllie, 1993). In cell culture experiments, NSs released from apoptotic S49 T lymphoma cells after sub-optimal dexamethasone treatment were able to attach to the surface of surviving cells. This was accompanied by a 50-fold increase in monoclonal ANA 2C5 binding to these cells (Iakoubov and Torchilin, 1998). Elevated free extracellular nucleochromatin has also been observed under other, non-cancerous conditions accompanied by massive apoptotic death, such as lupus erythematosis and AIDS (Licht et al, 2001). It is especially important, that NSs are exposed on the cell surface of apoptotic cells (Radic et al, 2004).

Functionally, extracellular chromatin fragments have been shown to inhibit the tumor cell killing by NK cells in vitro (Le Lann-Terisse et al, 1994, 1997). Such data allow considering the NS release by dying tumor cells as a tumor self-defense mechanism that protects the surviving tumor cells from host immune attack. In this case, the increased production of NS-specific cytotoxic autoantibodies by a tumor-bearing organism may be contemplated a response that acts to overcome the tumor self-defense. The data demonstrating a prolonged time to disease progression and an increased survival rate in cancer patients with elevated serum ANAs (Blaes et al, 2000; Syrigos et al, 2000), and a higher remission rate in chronic myelogenous leukemia patients who develop autoimmune phenomena (known to be accompanied with a rise in ANAs) as a result of alpha-interferon therapy vs those who do not (Sacchi et al, 1995) support a possibility of anti-tumor activity of ANAs. Here, we present the first experimental data on strong anti-tumor activity of mAb 2C5 against diverse human tumors in in vivo models.

In addition, the specificity of mAb 2C5 and similar antibodies against a broad variety of tumors is a prerequisite for their broad applicability as vehicles for targeting different therapeutic and diagnostic agents to various tumors (Torchilin et al, 2003). Supporting data on mAb 2C5's ability for intratumoral accumulation are also presented.

All cell culture media, heat-inactivated fetal bovine serum, and other cell culture reagents were obtained from Cellgro (Herndon, VA). Isotype-matching control antibody UPC10 and fluorescein-conjugated goat anti-mouse antibody were purchased from ICN Pharmaceuticals, Inc. (Cosa Mesa, CA). Diethylene triamine pentaacetic acid anhydride (DTPA) for radiolabeling mAb 2C5 with ¹¹¹In and all other chemicals and buffer solution components were obtained from Sigma (St. Louis, MO) and were of analytical grade. ¹¹¹In with specific radioactivity of 395 Ci/mg was from Perkin-Elmer (Boston, MA).

mAb 2C5 [IgG2a] (Iakoubov, Torchilin, 1997) was purified from the supernatant of 2C5 hybridoma cells grown in the CM10 culture medium by using a standard saturated ammonium sulfate precipitation and affinity chromatography on a protein G column (Exalpha Biologicals, Inc., Watertown, MA). Non-reducing SDS-PAGE of the purified antibody (Laemmli, 1970), and ELISA with purified monoNSs (Li et al, 1995) were performed to test mAb 2C5 purity and immunoreactivity. The purified antibody was stored at -80° C as a 5 mg/ml solution.

To label the mAb 2C5 with ¹¹¹In, it was first modified with chelating residues DTPA and loaded with ¹¹¹In by transchelation from the ¹¹¹In-citrate complex (Samokhin et al, 2001).

The referenced protocol includes the incubation of the antigen [monoNSs purified as in (Li et al, 1995), in this particular case] at various concentrations with mAb 2C5 at constant concentration until equilibrium is reached and subsequent determination of the free antibody concentration by an indirect ELISA. Samples of monoNSs at various concentrations (0.016 - 16 m g/ml) were mixed with the constant amount of antibody (1 m g/ml) in TBS, 2mM EDTA, pH 7.5, supplemented with 10 mg/ml BSA. After a 15 h incubation at 20° C, 100m l of each mixture was transferred and incubated for 30 min at 20° C into the wells of a microtiter plate coated with monoNSs [100 m l per well, at 10 m g/ml in TBST-Cas (TBS containing 0.05% w/v Tween-20 and 2 mg/ml casein), pH 7.5, for 1 h at 20° C]. After washing with TBS supplemented with 0.05% Tween 20, the wells were incubated for 1 h at room temperature with 100 m l of 1:5000 diluted goat anti-mouse IgG peroxidase conjugate (ICN Biomedicals Inc., OH) in TBST-Cas. After incubation, the wells were washed three times with 200 m l of TBST and each well was incubated with 100m l of enhanced K-blue® TMB peroxidase substrate (Neogen Corporation, KY) for 15 min. Finally, the plate was read at a dual wavelength of 620 nm with the reference filter at 492 nm using a Labsystems Multiskan MCC/340 microplate reader installed with GENESIS-LITE windows based microplate software.

Approximately 10,000 cells of different cell lines in each sample were used to check for mAb 2C5 binding using the isotype-matching non-specific UPC10 antibody as a negative control. Cells were washed with Hank's balanced salt solution (HBSS) and then with 1% BSA in PBS, pH 7.4, to prevent a non-specific binding. Cells were then incubated for 30 min at room temperature with mAb 2C5 or UPC10 and then stained with fluorescein-conjugated goat anti-mouse IgG. Finally, the cells were fixed with 4% paraformaldehyde solution in PBS, live-gated using forward vs. side scatter to exclude debris and dead cells, and analyzed for surface fluorescence using FACScan (Beckton Dickinson, Franklin Lakes, NJ).

All experiments were performed in 6-8 week old female nu/nu mice (Charles River Laboratories, Wilmington, MA) following protocol #011022 approved by the Institutional Animal Care and Use Committee in accordance with Principles of Laboratory Animal Care (NIH publication #85-23, revised in 1985). The animals were allowed free access to food and water. Approximately 8x10⁶ of H69(AR) cells, 10-12x10⁶ of COLO2O5, PC3, C32TG, and NCI-H82 cells, or 40x10⁶ of BT-20 cells in 0.25 ml of RPMI, were inoculated s.c. into the left flanks of the mice. Mice were examined every 2-3 days until palpable tumors were detected.

Mice with tumors 1-2 cm in diameter were injected via the tail vein with ¹¹¹In-labeled mAb 2C5. After 48 hrs, the mice were anaesthetized and body distribution and tumor accumulation of mAb 2C5-associated ¹¹¹In radioactivity was visualized using an Ohio Nuclear 400 g camera equipped with high-energy collimator and Technicare 560 dedicated computer with whole body pictures taken by digital camera. Tumor and adjacent normal tissues were then collected, and ¹¹¹In radioactivity was quantified (as counts per minute, CPM) using a Beckman 5500B g counter (Fullerton, CA). Then tumor-to-normal tissue ratios of the %-injected dose/g of tissue were calculated.

In the model of prophylactic treatment, nude mice were injected with mAb 2C5 and control PBS, pH 7.4 (in PC3 and NCI-H82 tumor models) or isotype-matched antibody UPC-10 (in COLO2O5 tumor model) prior to and after the inoculation with tumor cells. 150 m g of mAb 2C5 in 100 m l of PBS or the same volume/quantity of control injections per mouse were given via tail vein on days —1, 6, 13 and 19 in COLO2O5- and NCI-H82-bearing mice, and on days —1, 6 and 13 in PC3-bearing mice. In the established tumor treatment model, mAb 2C5 and control injections were initiated after the formation of a visible tumor 7-14 days after the inoculation of tumor cells. Tumor-bearing mice were injected with mAb 2C5 and control PBS [in PC3, NCI-H82, and H69(AR) tumor models] or UPC-10 (in COLO2O5 tumor model). 150 m g of mAb 2C5 in 100 m l of PBS or the same volume/quantity of control injections per mouse were given via tail vein on days 13 and 20 post inoculation in COLO2O5-bearing mice. In case of PC3-, NCI-H82-, and H69(AR)-bearing mice, injections were given on days 7, 11 & 15; 7, 14 & 21; and 9, 14, 20, and 27, respectively.

Anticancer activity of mAb 2C5 was assessed by following the tumor volumes during the treatment and average tumor weights at the completion of treatment in the control and mAb 2C5-treated groups. The apparent tumor volume was calculated on different days during the treatment using the formula (Geran et al, 1972), tumor volume (mm³) = (Length X Width²)/2, length being greater than width. At the end of treatment, the mice were sacrificed and tumors were removed and weighed.

Differences in apparent tumor volumes during the treatment and post-mortem tumor weights were compared using the non-parametric two-tailed Wilcoxon rank sum test for two independent samples. For easy representation of the difference in tumor volumes during the treatment, the average tumor volume (mm³) of the group vs. days after tumor cell inoculation was plotted. Tumor weights were plotted as box plots that represent the group median, quartiles and extremes.

We confirmed the purity of mAb 2C5 by the presence of the single peak in the protein G column antibody elution profile and the single band corresponding to an IgG protein after polyacrylamide gel electrophoresis (Figure 1, A,B). We have also confirmed antibody immunoreactivity against NSs purified from the rat liver (Li et al, 1995) using a standard ELISA protocol (Figure 1, C).

The determination of mAb 2C5 K_D with monoNS is presented on Figure 2 (here, the K_D value characterizes avidity or relative affinity because of bivalent nature of an antibody).

Figure 1. Purification and characterization of mAb 2C5. (A) The elution profile of mAb 2C5 from a protein G column; (B) SDS-PAGE showing purified mAb 2C5 (lane 1) and Bio-Rad high-range MW standards (lane 2, numbers on the right show corresponding MW values); and (C) The immunoreactivity of purified mAb 2C5 with nucleosome as a substrate in ELISA.

Figure 2. Determination of K_D for mAb 2C5 interaction with NSs. (A) Calibration curve for the binding of mAb 2C5 to monoNSs in ELISA; (B) The Scatchard plot for the binding of mAb 2C5 to monoNSs. v is the fraction of the bound antibody and a the concentration of free antigen at equilibrium. The total concentration of the antibody (as NS binding sites) was 6.6 x 10^-9 M.

According to the protocol used for determination of the dissociation constant, we worked in the linear binding concentration range of mAb 2C5 to monoNSs. One hundred per cent saturation was achieved, thus we used only the linear concentration range as shown in the standard curve to determine the dissociation constant. As shown in Figure 2A, the dependence of the absorbance (resulting from the enzymatic activity of the immunoconjugate retained in the wells) on the initial concentration of the mAb 2C5 remains linear in the whole range of concentrations used. This linear dependence allows for the determination of the free antibody concentration at equilibrium, provided that the total antibody concentration is known. Figure 2B shows the results, plotted according to the Scatchard equation for the binding of the pure mAb 2C5 to monoNSs. From the linear regression, the apparent K_D value for the mAb 2C5 is 2.36±0.64 x 10^-10 M (actually, it might be somewhat lower because of the bivalency of the whole antibody and possible presence of more than one antibody-binding site on the monoNS) that shows high specificity of antibody binding to NSs.

The results of FACS analysis (Figure 3) demonstrate an effective recognition by the mAb 2C5 of all human cancer cell lines tested. For our studies, we have selected COLO205, PC3, NCI-H82, BT-20, H9, MCF-7, C32TG, and adriamicin-resistant H69(AR) cell lines to represent a broad variety of unrelated tumors. These results are in good agreement with the previous data and confirm the broad anti-tumor specificity of this ANA with NS-restricted specificity (Iakoubov and Torchilin, 1997). We used several of these cell lines in further experiments on mAb 2C5 intra-tumor accumulation and anti-tumor activity.

The investigation of tumor accumulation of the antibody-bound ¹¹¹In radioactivity in several selected human tumors in mice clearly showed an enhanced accumulation of mAb 2C5 in all tested tumors, although the extent of this enhancement depended on the tumor type. Table 1 presents the data on an enhanced tumor accumulation of the antibody-bound ¹¹¹In radioactivity in several selected human tumors in mice with the tumor-to-normal tissue accumulation ratio varying between 2.7 and 13.4. We have additionally confirmed the efficient tumor accumulation of mAb 2C5 by direct gamma-immunoimaging of tumor-bearing mice showing an increased accumulation of the antibody-associated radioactivity in the tumor area (typical images are presented in Figure 4 for an NCI-H82 tumor-bearing mouse). At the same time, non-specific antibodies were repeatedly shown to have very little (if any) accumulation in various tumor models in mice (De Santes et al, 1992; Camera et al, 1993; Xu et al, 1997; Palm et al, 2003). These data strongly support the application of mAb 2C5 and similar antibodies as universal vehicles for in vivo intra-tumor delivery of drugs and imaging agents (Gao et al, 2003; Torchilin et al, 2003).

Our findings, although rather unusual, are supported by the whole set of circumstantial evidence found in the literature. Many publications provide direct and indirect evidence that ANAs may carry certain anti-tumor properties. Thus, the presence of ANAs is a characteristic feature of systemic autoimmunity (von Muhlen and Tan, 1995). At the same time, the mortality rate from cancer in autoimmune patients was long ago found to be significantly less than that in the healthy population (Palo et al, 1977). Recently, it was found that lupus patients are better protected from cancer (Huges, 2001). On the other hand, the experimental suppression of autoimmune manifestations (resulting in a decrease of autoantibody, including ANAs, production) in spontaneously autoimmune mice was found to sharply increase the incidence of spontaneous tumors (Hahn et al, 1975; Walker et al, 1978). In addition, as was already mentioned, circulating ANAs in patients with lung cancer (Blaes et al, 2000) and colorectal carcinoma (Syrigos et al, 2000) were associated with a prolonged survival without disease progression. Actually, the intentional induction of

Table 1. Accumulation of mAb 2C5-Bound ¹¹¹In radioactivity in various human tumors in nude mice (expressed as the tumor-to-normal tissue radioactivity ratio)

Figure 4. Typical gamma-immunoscintigraphic images obtained at different time points post-injection of ¹¹¹In-radiolabeled mAb 2C5 in a NCI-H82 tumor model.

Figure 5. Prophylactic tumor treatment model in COLO2O5, PC3, and NCI-H82 human tumors in athymic mice. Within each model, the tumor growth in mAb 2C5-treated group and control group is illustrated by calculated tumor volumes during the treatment period (left columns) and by post-mortem tumor weights (right columns). Tumor volumes are plotted as average tumor volume (mm³) vs. days after tumor cell inoculation with arrows indicating days of injections. Post-mortem tumor weights are shown as box plots showing the group median, quartiles, and extreme values.

Figure 6. Established tumor treatment model in COLO2O5, PC3, NCI-H82 and multi-drug resistant H69 (AR) human tumors in athymic mice. See details in the legend to Figure 5.

autoimmunity was seriously considered as an antineoplastic therapeutic strategy (Pardoll, 1999). Though the exact autoimmune components with antitumor function are not known, ANAs seem to be a good candidate.

Two major mechanisms of action are presently being considered to explain anti-tumor activity of therapeutic monoclonal antibodies (Houghton and Scheinberg, 2000). The first is based on the antibody-mediated killing of tumor cells by various immune effectors such as complement or cytotoxic immune cells. The second involves antibody bioactivity un-related to immunologic function, such as blocking the access of growth factors to tumor cells or tumor vasculature. A convincing case for the in vivo involvement of immune effectors in monoclonal antibody-based tumor suppression was obtained recently in immunocompetent as well as athymic mice (Clynes et al, 2000). However, for each particular antibody, the conclusion on the involvement of immune killing is traditionally based on the indirect evidence, where the ability of the antibody to mediate complement- or cytotoxic cell-dependent killing is demonstrated in vitro (Herlyn et al 1980). In addition, according to this in vitro criteria, effector cells from nude mice are known to mediate levels of ADCC of equal or even greater magnitude than cells from normal mice (Kohl et al, 1984). Our in vitro data on the ability of mAb 2C5 to mediate ADCC serve as indirect evidence for this mechanism of tumor suppression in both syngeneic tumor models in normal mice (Iakoubov and Torchilin, 1997) and xenogeneic models in nude mice.

We can also reasonably expect at least one additional activity for the mAb 2C5. Since most tumors are expected to bear NSs on the surfaces of apoptotic cells (Radic et al, 2004) and result in elevated blood NSs (Holdenrieder et al, 2001), external anti-NS antibodies should lead to elevated immune complexes known to activate the immune system (Lucisano, Mantovani, 1984). This activation could be of a significant importance in opposing the tumor-mediated immunosuppression characteristic of cancer patients (Pollock, Roth, 1989). Thus, the efficiency of mAb 2C5 and similar antibodies might be in part based on their ability to turn on some additional immune mechanisms not characteristic for conventional immunization-induced monoclonal antibodies. In addition to their probable ability to block to a certain extent the earlier mentioned NK-inhibiting action of extracellular NSs (chromatin) (Le Lann-Terisse et al, 1994, 1997), these antibodies can also interfere with the recently reported neoangiogenic activity of NSs (Tanner, 2003). The competition between NS-specific antibodies and C-reactive protein (Robey et al, 1985) for the binding with chromatin may exist in the blood, however the results of our experiments clearly demonstrate that in vivo the activity of mAb 2C5 is not blocked by this hypothetical competition.

Earlier, monoclonal antibodies belonging to IgG2a and IgG3 isotypes have been described demonstrating ADCC against tumor cells over-expressing corresponding antigens, such as anti-ganglioside antibodies (Lin et al, 2001; Nakamura et al, 2001; Parajuli et al, 2001). However, these antibodies have been obtained by immunization and, to the best of our knowledge, none of them was ever shown to demonstrate anti-tumor activity against a variety of unrelated human tumors.

As potential anti-tumor agents, natural ANAs may have a number of advantages compared to conventional anti-tumor antibodies. ANAs appear to be effective against a broad spectrum of various tumor types. Their anti-tumor mechanisms are probably multimodal and hence more efficient. Their side-effects should be minimal, since their natural presence in the blood does not seem to be harmful to the host. In addition, assuming that the ADCC and other immune effectors possibly activated by circulating mAb 2C5/NS immunocomplexes are independent of MDR mechanisms, such as the p-glycoprotein pump, mAb 2C5 and similar antibodies might serve as a treatment of choice against MDR tumors. This assumption is well supported by our results with the H69(AR) tumor. One can also speculate that such antibodies may be used together with apoptosis-inducing agents required to convert the tumor cell chromatin into NSs and subsequently provoke the release of these NSs and their binding to the surface of living tumor cells to make them better targets for the antibodies.

The question if the mAb 2C5 (and similar antibodies) only inhibit tumor growth and have to be used as a component of the combination therapy, or, if optimized administration protocols are established, they can completely eliminate tumors, as well as the exact nature of NS-binding sites on the surface of tumor cells and major mechanisms of anti-tumor activity of NS-restricted ANAs are currently under investigation in our laboratory.

This study was funded in part by the NIH grant R01 EB02995 to VPT. The authors would like also to acknowledge the researcher-friendly policy of partial overhead return at Northeastern University.