To assess the role of CD4+CD25+Foxp3+ regulatory T (Treg) cells in overcoming immunosurveillance of Erbb2 (HER-2/neu) mammary lesions, we studied the effects of their sustained removal in BALB/c female mice made transgenic for the rat Erbb2 (r-Erbb2) oncogene (BALB-neuT mice), which develop multiple mammary carcinomas. During the progression of these lesions, Treg cells expand in the spleen, tumor draining lymph nodes, and tumors. Repeated administration of anti-CD25 antibodies extends tumor-free survival, reduces carcinoma multiplicity, and leads to the manifestation of a natural antibody and CTL-mediated reactivity against r-Erbb2. Loss of Foxp3+ Treg cells during anti-CD25 treatment remarkably caused the disappearance of Gr1+ immature myeloid cells, suggesting a cross-talk between these two inhibitory immune cell types. Treg cell expansion associated with r-Erbb2 overexpression may be seen as a physiologic response to dampen the immune reaction elicited by local anomalous overexpression of a self-antigen. (Cancer Res 2006; 66(15): 7734-40)
Overexpression of the Erbb2 (HER-2/neu) oncogene by a significant percentage of human carcinomas is associated with aggressive tumor growth, greater invasiveness, enhanced metastatic potential, and increased resistance to therapy (1). The protein product coded by Erbb2, p185, is a member of the epidermal growth factor receptor family endowed with a potent tyrosine kinase activity that plays crucial roles in physiologic processes, such as embryogenesis, cell proliferation, and apoptosis (2). However, as p185 is a self-antigen poorly expressed by the cells of adult individuals, patients with clinically evident carcinomas overexpressing Erbb2 may display a natural antibody and cell-mediated response to p185 (3, 4). These are too small and too late to enhance resistance to tumor expansion. Their induction, however, suggests that p185 overexpression triggers an immune response. This is an important issue, because elicitation of an immune response to p185 (5) and administration of high-avidity monoclonal antibodies (mAb; ref. 6) are effective against Erbb2 carcinomas.
The inability to naturally mount an effective immune response to p185 is the result of immune tolerance to self-antigens. Clonal deletion of T and B cells recognizing p185 with high avidity irreversibly impairs the immune repertoire. Lower-affinity responses and responses to subdominant epitopes of self p185 that were not deleted may be inhibited by regulatory T (Treg) cells (7), interleukin (IL)-13–producing natural killer (NK) T cells (8), and immature myeloid cells (9) as well as additional mechanisms providing a negative regulation of the immune response of autoreactive T cells.
To study the surmounting of natural immunosurveillance during Erbb2 carcinogenesis, we used female BALB/c mice made transgenic for the rat Erbb2 (r-Erbb2) transforming oncogene (BALB-neuT). All these female mice develop a multifocal carcinoma in each of their 10 mammary glands (10) with a stepwise progression that mimics a few typical features of human Erbb2 carcinogenesis (11). This progression can be exploited to study the natural expansion of CD4+CD25+ Treg cells expressing the Foxp3 transcription factor and the glucocorticoid-inducible tumor necrosis factor receptor (GITR; ref. 12).
Transgenic Erbb2 is the only genetic difference between BALB-neuT and wild-type BALB/c mice. They can thus be compared with assess the consequences of progressive overexpression of a self-antigen in the expansion of Treg cells, because these are physiologically involved in inhibiting the immune response and maintaining homeostatic tolerance to self-antigens (12, 13). As Treg cell removal with antibodies or defects in their maturation may result in various forms of autoimmunity (12–15), reduction of their expansion may disclose the presence of a natural immunosurveillance to overexpressed rat p185 (r-p185) in BALB-neuT mice and may pose the basis for the design of more effective immunologic maneuvers in tumor prevention and treatment.
The results of the present study show that Treg cells expand in BALB-neuT mice during the lengthy progression of Erbb2-driven mammary carcinogenesis. Their sustained removal discloses an antibody and CTL-mediated natural immunosurveillance able to hamper the progression of autochthonous Erbb2 lesions.
Materials and Methods
Mice. Severe combined immunodeficient (SCID) and BALB/c (H-2d; 6-8 weeks old) female mice were obtained from Charles River Italia SpA (Calco, Italy). Mammary cancer-prone BALB-neuT female mice (H-2d) overexpressing the r-Erbb2 transforming oncogene under the control of the mouse mammary tumor virus promoter (10) were bred for us under specific pathogen-free conditions at Charles River Italia. These mice were randomly assigned to control and treatment groups and concurrently treated. Mammary glands were inspected weekly to note tumor appearance. Progressively growing masses >1 mm mean diameter were regarded as tumors. Tumor multiplicity was calculated as the cumulative number of incident tumors divided by total number of mice and is reported as mean ± SE (10). Each neoplastic mass was measured with calipers in two perpendicular diameters and its volume was calculated as (X2 × Y) / 2, where X and Y represent the short and long diameters, respectively. Total tumor volume is the sum of individual tumor volumes of each mouse and is reported as mean ± SD. In all operations, mice were treated in accordance with the European Community guidelines.
Production and administration of anti-CD25 antibodies. The PC61 hybridoma-secreting IgG1 mAbs to the α-chain of murine IL-2 receptor (CD25; ref. 16) was purchased from the American Type Culture Collection and cultured in vitro in DMEM (BioWhittaker Europe, Verviers, Belgium) supplemented with 5% fetal bovine serum, 0.5 mmol/L sodium pyruvate, 1 mmol/L nonessential amino acid, 1.25 g/L bicarbonate, and 25.7 mmol/L β-mercaptoethanol and then grown as ascites in SCID mice (15). The titer of IgG1 in the ascite fluids passed through 0.45 μm membrane filters (BD Biosciences, Erembodegem, Belgium) was determined with radial immunodiffusion kits (The Binding Site Ltd., Birmingham, United Kingdom). The fluid was diluted in PBS to obtain a concentration of 2.5 mg IgG1/mL. At the sixth week of age, BALB-neuT mice twice received 500 μg anti-CD25 IgG or control normal rat IgG (rIgG; Sigma-Aldrich, St. Louis, MO) followed by weekly i.p. repeats until the 24th week.
Morphologic analyses. Histologic evaluation of mammary carcinogenesis and whole mounts of mammary glands were done as described previously in detail on groups of five BALB-neuT mice of progressive age (11, 12). To evaluate the presence of Foxp3+ cells in mammary glands, spleen, and lymph nodes during tumor progression and Treg cell depletion, groups of five BALB-neuT mice untreated or receiving normal rIgG or anti-CD25 IgG were sacrificed at ages 7, 13, 19, and 25 weeks. For immunohistochemistry, pyridoxal phosphate–fixed tissues were embedded in OCT and acetone-fixed cryostat sections were incubated for 60 minutes with anti-Foxp3 (clone MF333F, Alexis Italia, Vinci, Florence, Italy). Microwave antigen retrieval was done with 1 mol/L urea for 3 minutes. After washing, sections were overlaid with biotinylated goat anti-rIgG (Vector Laboratories, Burlingame, CA) for 30 minutes, incubated with streptavidin ABC/alkaline phosphatase. Staining was developed with fuxin (DakoCytomation, Heverlee, Belgium). For each mouse, Foxp3+ lymphoid cells were counted in a blind fashion independently by three pathologists in 10 high-power ×400 fields.
Cytometric identification of Treg cells and CD11b+Gr1+ immature myeloid cells. The relative numbers of CD4+CD25+Foxp3+GITR+ Treg and CD11b+Gr1+ immature myeloid cells in the spleen and lymph nodes draining the mammary pad were evaluated by flow cytometry. Spleen cells (Spc; 1 × 106) and cells from lymph nodes draining the mammary pad were treated with Fc receptor blocker (CD16/CD32; PharMingen, San Diego, CA) for 15 minutes at 4°C. For Treg cell detection, directly conjugated phycoerythrin (PE) anti-mouse GITR (clone DTA-1, eBioscience, San Diego, CA), PE/Cy7 anti-mouse CD4 (clone GK1.5, BioLegend, San Diego, CA), and allophycocyanin anti-mouse CD25 (clone PC61.5, eBioscience) were incubated for 30 minutes at 4°C. The cells were then washed in PBS with 0.1% sodium azide and 2% fetal bovine serum. Cell pellets were resuspended in 1 mL Fix/Perm (eBioscience) and the samples were incubated overnight at 4°C. After two washes with permeabilization buffer (eBioscience), the samples were incubated with 2 μL Fc receptor blocker for 15 minutes at 4°C and then with FITC anti-mouse/rat Foxp3 (FJK-16s, eBioscience) for 30 minutes at 4°C and washed twice before analysis. For immature myeloid cell detection, cells were incubated with 1 μL PE anti-mouse Ly-6G and Ly-6C (Gr-1, clone RB6-8C5, PharMingen) diluted 1:20 in PBS and 3 μL FITC anti-mouse CD11b (Mac1α, clone M1/70, PharMingen) at 4°C for 30 minutes. The samples were washed twice with PBS containing 0.1% sodium azide and 2% calf serum and analyzed on the CyAn ADP (DakoCytomation) through Summit 4.2 (DakoCytomation) software.
Assessment of anti-r-p185 antibody. Sera collected at 10 and 25 weeks from BALB-neuT mice receiving normal rIgG or anti-CD25 IgG were diluted 1:50 in PBS/sodium azide/bovine serum albumin (Sigma-Aldrich) and the presence of anti-r-p185 antibodies was determined by flow cytometry using BALB/c NIH3T3 fibroblasts, wild-type or stably cotransfected with the wild-type r-Erbb2, mouse class I H-2Kd and B7.1 genes (BALB/c NIH3T3-NK B cells; ref. 17). FITC-conjugated goat antibodies specific for mouse IgG Fc (DakoCytomation) were used to detect bound primary antibodies. Normal mouse serum was the negative control. The mAb Ab4 (Oncogene Research Products, Cambridge, MA), which recognizes an extracellular domain of r-p185, was used as a positive control. Serial Ab4 dilutions in normal mouse serum were used to generate a standard curve to determine the concentration (μg/mL) of anti-r-p185 antibodies in mouse sera. Flow cytometry was done on the CyAn ADP (17).
In vitro and in vivo CTL assays. Spc were restimulated in vitro by culturing for 6 days 1 × 107 cells with 5 × 105 mitomycin C (Sigma-Aldrich)–treated 3T3NKB cells in the presence of 10 units/mL recombinant IL-2 (Eurocetus, Milan, Italy). In some experiments, Spc were restimulated by adding 2.5 μg/mL of the r-p185 63-71 nonamer peptide (TYVPANASL; ref. 18; InBios Srl, Biotech Products, Naples, Italy) predicted to bind the H-2d glycoproteins with high affinity (http://www.syfpeithi.de/). Restimulated Spc were assayed in a 4-hour Na251CrO4 (51Cr, Perkin-Elmer, Boston, MA) release assay at E:T ratios from 50:1 to 6:1 in round-bottomed, 96-well microtiter plates in triplicate as described previously in detail (19). In vivo cytotoxicity assay was done as described by Ritchie et al. (20), with slight modifications. Briefly, a single-cell suspension of 107 naive Spc/mL was labeled with 0.5 or 5.0 μmol/L of the fluorescent dye CFSE (Molecular Probes, Leiden, the Netherlands). Spc labeled with 5 μmol/L CFSE (CFSEhigh) were also pulsed with r-p185 63-71 nonamer peptide for 1 hour at room temperature. The two Spc populations were mixed together in equal amounts and injected i.v. into control and treated mice. Mice were sacrificed 48 hours later, and single-cell suspensions from spleens were processed individually to evaluate the presence of CFSEhigh and CFSElow cells with the CyAn ADP after adding propidium iodide to exclude dead cells. The specific cytolytic activity was calculated as 100 × (percentage CFSElow cells − percentage CFSEhigh cells) / percentage CFSElow cells.
Statistics. Differences in tumor incidence were evaluated with the Mantel-Haenszel log-rank test, those in tumor multiplicity, number of positive cells at flow cytometry and antibody titer with Student's two-tailed t test.
Treg cells expand during Erbb2 carcinogenesis. First, we determined whether the progression of carcinogenesis in the mammary glands of BALB-neuT mice triggers the expansion of CD4+CD25+Foxp3+ Treg cells. The atypical mammary hyperplasia generated by cells overexpressing r-p185 first evident at age 4 weeks (11) progresses to multifocal preneoplastic lesions around week 7 (Fig. 1C and H). Multiple carcinomas of ∼200 mm3 are palpable in every mouse by week 25 (Fig. 1A,, F, and K). At the seventh week, the initial Erbb2 overexpression does not lead to detectable Treg cell accumulation in the spleen (Fig. 1A) and the axillary lymph nodes draining the mammary pad (data not shown) or in the mammary lesions, where Foxp3+ lymphoid cells remain 1 ± 2 per ×400 microscopic field (Fig. 2A). However, the subsequent progression of mammary carcinogenesis is accompanied by an increment in CD4+CD25+Foxp3+ Treg cells in the spleen (Fig. 1A) and the mammary tumors (Fig. 2B). We have shown previously that this progression of BALB-neuT carcinogenesis is also accompanied by an expansion of CD11b+Gr1+ immature myeloid cells (21).
Sustained depletion of CD4+CD25+Foxp3+GITR+ Treg cells through chronic infusion of anti-CD25 IgG. Cancer-prone BALB-neuT mice received repeated infusions of normal rIgG or anti-CD25 IgG from the 6th week to the 24th week of age to assess their effect on Treg cells. A dramatic decrease in Treg cells was already evident at week 7 in both the spleen (Fig. 3A) and axillary lymph nodes (Fig. 3B) following the first two administrations of 500 μg anti-CD25 IgG. At week 25, the percentage of Foxp3+GITR+ T cells among total CD4+ cells almost doubled in the spleen and markedly increased in the lymph nodes of normal rIgG-treated BALB-neuT mice. By contrast, repeated injections of anti-CD25 IgG, twice in week 6 and then once weekly until week 24, keeps low their percentage in both spleen and lymph node as shown by flow cytometry (Fig. 3A and B). This reduction is mirrored by immunocytochemistry observations (Fig. 3C versus Fig. 3D).
Treg cell depletion accompanies delayed carcinogenesis. To determine whether chronic removal of Treg cells unveils an immune response able to hamper the progression of Erbb2 lesions, tumor incidences were compared in normal rIgG-treated and anti-CD25 IgG-treated BALB-neuT mice. Both a marked delay in the appearance of the first tumor and a reduction in the number of palpable tumors per mouse (tumor multiplicity; ref. 11) were evident in anti-CD25 IgG-treated mice. At week 20, when all control mice displayed one or more tumors, 80% of mice treated with anti-CD25 IgG were free of palpable tumors (Fig. 4A). The tumor-free survival curve of anti-CD25 IgG-treated mice was significantly delayed (P < 0.0001) compared with that of normal rIgG-treated mice. The tumor multiplicity was also significantly lower (P < 0.04 to P < 0.0004) from weeks 19 to 32 (Fig. 4B). Moreover, at 38 weeks, when the experiment ended, a few mammary glands of BALB-neuT mice receiving anti-CD25 IgG until week 24 did not display a palpable tumor. Depletion of Treg cells through anti-CD25 IgG also leads to both a reduced number of CD11b+Gr1+ immature myeloid cells at week 7 and their drastically hindered expansion at week 25 (Fig. 5).
Treg cell depletion unveils a natural immune response to r-p185. We next determined whether Treg cell expansion conceals an immune response that may be naturally triggered by r-p185 overexpression. A single weekly infusion of anti-CD25 IgG, but not normal rIgG, uncovers both a significant antibody (Fig. 6) and a CTL response to r-p185 (Fig. 7). The presence of anti-r-p185 natural antibodies was assayed in the sera from 10- and 25-week-old BALB-neuT mice repeatedly infused with normal rIgG or anti-CD25 IgG. A significant higher titer of anti-r-p185 antibodies was found in sera from anti-CD25-treated mice at both age 10 weeks (P = 0.0045) and age 25 weeks (P = 0.028) compared with age-matched normal rIgG-treated control mice (Fig. 6). No anti-r-p185 antibodies were detectable in untreated 10- and 25-week-old BALB/c mice (data not shown). The CTL response of Spc from anti-CD25-treated 25-week-old BALB-neuT mice was studied against both target cells pulsed with the r-p185 63-71 dominant H-2Kd restriction element peptide (Fig. 7A and B) and target cells expressing the whole r-p185 (Fig. 7C). The in vivo cytotoxic response against r-p185 63-71 peptide-pulsed cells (Fig. 7A) was significantly higher in anti-CD25-treated mice compared with normal rIgG-treated mice (P < 0.0001). The cytotoxic response of Spc from anti-CD25-treated BALB-neuT mice was confirmed in vitro against cells pulsed with the r-p185 63-71 peptide (Fig. 7B; Spc from anti-CD25 versus normal rIgG-treated mice, 48.5 ± 0.6 versus 14.2 ± 0.4 LU20/107 cells; P < 0.0001) and against 3T3NKB cells expressing r-p185 (Fig. 7C; Spc from anti-CD25 versus normal rIgG-treated mice, 45.1 ± 0.5 versus 6.3 ± 0.3 LU20/107 cells; P < 0.0001). No cytotoxicity against r-p185 was detectable in Spc from untreated 10- and 25-week-old BALB/c mice (data not shown). The presence of a marked cell-mediated cytotoxic response to r-p185 is impressive because it was never observed in tolerant BALB-neuT mice even after repeated anti-r-p185 immunizations (19, 22–25).
Enhancement of immunosurveillance through Treg cell removal has been mostly reported in transplantable tumor models (26–30) but also with tumors growing in transgenic mice (17, 30) and chemically induced tumor development (31). Here, we show that during Erbb2-driven mammary carcinogenesis CD4+CD25+Foxp3+GITR+ Treg cells expand in the spleen, tumor draining lymph nodes, and mammary lesions. The progression of BALB-neuT carcinogenesis is also accompanied by an expansion of CD11b+Gr1+ immature myeloid cells in both the blood (21) and the spleen. Chronic infusion of anti-CD25 IgG into BALB-neuT mice does not simply result in the inactivation of CD4+CD25+ cells (32) but also leads to a sustained physical depletion of CD4+CD25+Foxp3+GITR+ Treg cells. Such removal unveils a natural immunosurveillance against Erbb2-driven autochthonous carcinogenesis.
Because of the mammary overexpression of membrane r-p185, transgenic BALB-neuT mice are genetically predestined to develop multiple invasive and metastasizing mammary carcinomas (10). Many features of their progression, including gene expression profiles, closely mimic what happens in human mammary cancer (11, 33). In these mice, the chronic removal of CD4+CD25+Foxp3+GITR+ Treg cells extends tumor-free survival, reduces carcinoma multiplicity, and leads to the manifestation of a natural antibody and CTL-mediated reactivity against r-p185. It also hinders the expansion of CD11b+Gr1+ immature myeloid cells that goes along with tumor progression (21).
Because the r-Erbb2 transgene is the genetic difference between wild-type BALB/c mice and transgenic BALB-neuT mice, comparison of the immune response in these two lines allows direct assessment of the tolerance to r-p185 as an overexpressed tumor-associated antigen. BALB/c mice do not express r-p185, which is thus a xenogeneic antigen differing in several epitopes from mouse p185 (18). Following immunization, BALB/c mice develop a strong immune response to r-p185, and CTL are a significant component of such response (34). The reaction triggered by the vaccine (34) or after Treg cell removal (15) is strong enough to bring about the rejection of large transplanted r-p185+ tumors. By contrast, in BALB-neuT mice, r-p185 is expressed in the thymus at birth and is progressively increasingly overexpressed by the cells of hyperplastic mammary lesions starting from the fourth week of age (11). Because this r-p185 overexpression, immunoscope analysis of the T-cell repertoire shows that in BALB-neuT mice CTL clones reacting with high affinity with r-p185 peptides are depleted.5
Rolla et al., submitted for publication.
Very little information is available regarding how tumor-specific Treg cells develop in tumor-bearing hosts. Present data in BALB-neuT mice show that by comparison with age-matched BALB/c mice no major increase in Treg cells is evident during the early stages of mammary hyperplasia. This is not surprising because r-p185 is but one of the innumerable self-antigens against which the autoimmune response is prevented by Treg cells (36). However, as mammary lesions progress and many more cells overexpress r-p185, an expansion of Treg cells becomes evident in the spleen and particularly in the tumors. This late infiltration of Treg cells in Erbb2 carcinomas fits in well with what has been described with transplantable tumors (29) and human cancers (37), where Treg cell accumulation is a late event in well-established tumors. The Treg cell ability to localize to the tumor site seems to permit the close contacts with effector CTL required for interference with their functions. By contrast, Treg cells in the peripheral organs may inhibit CD4 helper function and thus the elicitation of a significant and long-lasting antibody-mediated response (38).
The Treg cell expansion that accompanies r-p185 overexpression may be seen as a physiologic response to dampen the immune reaction elicited by local anomalous overexpression of a self-antigen, a major source of spontaneous autoimmunity (39). However, r-p185 is not only a self-antigen overexpressed on the cell membrane as carcinogenesis progresses but also a signaling receptor that delivers signals triggering the proliferation and survival of normal and tumor cells and whose anomalous overexpression plays a causal role in the promotion of carcinogenesis (1, 2). This double role of r-p185, a self-tolerated antigen playing important physiologic roles and a tumor antigen causally involved in the neoplastic progression, paradigmatically illustrates what may happen with most tumor-associated antigens. These, in fact, are self-antigens and thus display a natural immune recognition and immunosurveillance counterbalanced by a dominant immune tolerance (40, 41).
In several cases, the antigen presented by autochthonous tumors does not promote the dendritic cell activation necessary for proper arousal of effector CD4+ and CD8+ T-cell responses and results in the induction of tolerance (42, 43). By contrast, present data as well studies in patients (3, 4) suggest that tumor overexpression of p185 is enough to overcome self-tolerance and arouse antibody and cell-mediated immune responses. In the clinical setting, these responses are too small and too late to influence tumor progression. In BALB-neuT mice, they are meaningless as they are buried by Treg cells. Following Treg cell removal, a significant natural surveillance against Erbb2 carcinogenesis is evident. It is, however, only temporarily effective and insufficient to ultimately eradicate the tumor. This failure may rest on the central deletion in BALB-neuT mice of effector T cells recognizing r-p185 with high avidity. Moreover, additional physiologic immunoregulatory mechanisms (8, 9, 21) can be brought into play by the continuous onset of new neoplastic cells in transgenic BALB-neuT mice (19). Treg and myeloid immature cells are intimately associated with the immune suppression mediated by spontaneous tumors (7, 21). It has been shown that the accumulation of myeloid immature cells driven by tumor expansion of a transplantable tumor favors the expansion of Treg cells (44). On the other hand, present data show that Treg cell depletion avoids the accumulation of myeloid immature cells. Although this may depend on the delayed carcinogenesis due to Treg cell removal, a cross-talk between these two regulatory cells cannot be ruled out, and further studies may elucidate the pathways of their interaction.
In conclusion, present data show that r-p185 overexpression by mammary lesions naturally activates an antibody- and CTL-mediated immunity able to counteract initial stages of carcinogenesis. This, however, is dampened by Treg cells and possibly by other regulatory mechanisms. There are several reasons why Treg cell activity becomes dominant during Erbb2 carcinogenesis. Overexpression of r-p185 by the BALB-neuT mammary lesions may build the right conditions leading to Treg cell expansion or the conversion of naive CD4+ T cells into Treg cells (38). Besides the peculiar cytokines produced by the BALB-neuT carcinomas and their microenviroment, the dominant action of Treg cells may rest on the higher avidity with which they recognize self-antigen compared with effector T cells that escape deletional tolerance (45). It is evident that the risk of a rampant autoimmunity to an overexpressed self-antigen is a more effective evolutionary pressure than the production of a crippled immunosurveillance. However, the coexistence of dominant regulatory mechanisms and autoimmune-based immunosurveillance is both intriguing and alarming, because maneuvers leading to Treg cell removal may uncover significant antitumor reactivity but also trigger significant autoimmunity to self-antigens (15). In the specific case of Erbb2, interference with regulatory mechanisms may improve the effectiveness of immunosurveillance and immunotherapy treatments along with the activation of autoimmune reactions. However, the low avidity of autoimmune effector T cells that escape deletional tolerance (45) will react mostly if not solely against target cells that overexpress p185. In adult life, such overexpression is confined to neoplastic cells.
Note: Present address for E. Ambrosino: Molecular Immunogenetics and Vaccine Research Section, Vaccine Branch, National Cancer Institute, NIH, Bethesda, MD 20892.
E. Ambrosino and M. Spadaro contributed equally to this work.
Grant support: Italian Association for Cancer Research, Italian Ministries for the Universities and Health, University of Torino, Regione Piemonte, and Center of Excellence on Aging, University of Chieti, Italy.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Irene Merighi for excellent technical assistance.