CCL2 Blockade Augments Cancer Immunotherapy

Current immunotherapies are primarily aimed at initiating or boosting T-cell responses to tumors and their antigens. However, the effectiveness of these therapies may be limited by systemic and local tumor-induced immunosuppression (1). It is therefore becoming more widely accepted that successful immunotherapy will require a second approach to alter tumor microenvironment and/or decrease immunosuppression (2). Several approaches have been used, such as blockade of transforming growth factor-β (TGF-β) or TGF-β signaling (3), use of cyclooxygenase-2 inhibitors (4), depletion of T regulatory (Treg) cells (5), or blocking CTLA-4 (6).

Another immunomodulatory factor secreted from tumor cells and the associated tumor stromal cells is monocyte chemoattractant protein (MCP-1, CCL2), a CC (β) member of the cytokine/chemokine superfamily. Although first identified as a chemokine that could induce the migration of monocytes (7), CCL2 has several other chemotactic properties that include attraction of subsets of lymphocytes (including T regulatory cells) and endothelial cells into sites of inflammation (7–9). Importantly, it has also been observed to directly affect T-cell function, specifically inhibiting CD8⁺ T-cell effector functions (10–12).

Because of these immunosuppressive properties, we hypothesized that CCL2 was acting as an inhibitor of the effect of cancer immunotherapy, and its blockade might thus be benficial. In the mouse, there are two human CCL2 orthologues: CCL2 (MCP-1) and CCL12 (MCP-5). We initially evaluated the effect of blocking either one of these orthologues with antibodies that specifically neutralize these chemokines (8, 13) and found that monoclonal antibodies (mAb) against each one of them had a modest effect alone on tumor growth. We therefore used a mixture of the two mAbs in further clinical and mechanistic experiments (which we will heretofore refer to as “α-CCL2”). Our data suggest that CCL2/CCL12 is an endogenous barrier to cancer immunotherapy and that blockade could be a promising approach to augment CD8⁺ T-cell–mediated immunotherapy.

Female C57BL/6 mice were purchased from Charles River Laboratories. Female C57BL/6J × 129P3/J hybrids (B6-129/J1) were purchased from The Jackson Laboratory. The Animal Use Committee of the University of Pennsylvania approved all protocols in compliance with the Guide for the Care and Use of Laboratory Animals.

TC1 cells were derived from mouse lung epithelial cells of a C57BL/6 mouse, immortalized with human papillomavirus (HPV)-16 E6 and E7, and transformed with the c-Ha-ras oncogene (14). The murine lung cancer line LKR was derived from an explant of a pulmonary tumor from an activated Kras G12D mutant mouse that had been induced in an F1 hybrid of 129Sv.J and C57BL/6 (15). The murine malignant mesothelioma cell line AE17 was derived from the peritoneal cavity of C57BL/6J mice injected with asbestos (crocidolite) fibers and given to us by Dr. Delia Nelson from the University of Western Australia, Perth, Australia (16). Human mesothelin was transfected into the AE17 cell line using a lentiviral construct (AE17.hmeso).

C1142 is a rat/mouse chimeric mAb that neutralizes mouse CCL2/JE (MCP-1) and C1450 is a human/mouse chimeric mAb that neutralizes the second mouse homologue CCL12 (MCP-5; refs. 8, 13, 17). Both mAbs were produced at Centocor, Inc. In most experiments, mice were treated with a mixture of 250 μg per mouse of each mAb (α-CCL2) in a total volume of 200 μL normal saline i.p., twice per week. Control mice were treated with an equal volume of normal saline.

We used three different immunotherapy models. For the TC1 tumor model, we used an E1/E3-deleted type 5 adenoviral vector expressing the HPV-E7 protein under control of a cytomegalovirus promoter as previously described (Ad.E7; ref. 4). Animals bearing TC1 tumors were vaccinated s.c. with 1 × 10⁹ plaque-forming units (pfu) of Ad.E7 vector followed by a booster after 7 d. For the LKR cell line, we used an adenovirus expressing a hybrid IFN-α2α1 (Ad.IFNα) with activity in mice, received from Schering-Plough, Inc. (17). One dose of 1 × 10⁹ pfu of virus was injected intratumorally. For the AE17.hmeso tumor model, we used a modified live-attenuated Listeria monocytogenes vector expressing human mesothelin (Lm.Meso) provided by Drs. Dirk Brockstedt and Thomas Dubinsky (Anza Corp.). Mesothelin is a tumor-associated antigen highly expressed in human malignant mesotheliomas (18). Lm.Meso was constructed by inserting a mesothelin expression cassette integrated at the inlB locus.

Mice were injected on the right flank with 1 × 10⁶ of TC1, LKR, or AE17.hmeso tumor cells in the appropriate mouse strain. The flank tumors were allowed to reach an average size of 200 to 250 mm³ (approximately 12–15 d). Mice were treated in one of four groups: (a) control untreated, (b) α-CCL2/CCL12 mAbs, (c) immunotherapy alone, or (d) combination of immunotherapy and α-CCL2/CCL12 mAbs. All experiments had at least five mice per group and were repeated at least once. When needed for analysis [i.e., for fluorescence-activated cell sorting (FACS), RNA, and cell subset isolation], flank tumors were harvested from the mice and digested with 2 mg/mL DNase I (Sigma) and 4 mg/mL collagenase type IV (Sigma) at 37°C for 1 h.

Splenocytes, lymph nodes, and tumor cells were studied by FACS analysis as previously described (4). All fluorescently labeled antibodies used were purchased from BD Biosciences, except for CD206-phycoerythrin (PE; obtained from Serotec), 4-1BB (CD137)-PE (obtained from Abcam), and GR1-FITC (obtained from eBioscience). Flow cytometry was done using a FACSCalibur flow cytometer (Becton Dickinson). Data analysis was done using FlowJo software (TreeStar). The allophycocyanin-labeled H-2D^b tetramer loaded with E7 peptide (RAHYNIVTF) was obtained from the National Institute of Allergy and Infectious Diseases tetramer core. Intracellular staining for FoxP3 was done using the PE anti-mouse/rat FoxP3 staining set (eBioscience).

Mice with tumors were treated with either one of the four treatments detailed above, removed 2 d after the Ad.E7 boost vaccine, and flash frozen, and the RNA from each tumor was isolated. For each treatment condition, a pool of RNA was created by adding the same amount of RNA from each of the tumors within the group. cDNA was made from each pool, RNA levels were normalized to β-actin levels, and quantification of tumor mRNA levels was performed as previously described (19). Relative levels of expression of each of the selected genes (fold change versus control) were determined. Each sample was run in quadruplicate and the experiment was repeated at least once. Primer sequences are given in Supplementary Table S1.

Animals bearing flank tumors, treated with each of the treatments as above, were euthanized 2 to 3 d after the booster Ad.E7 vaccine. The tumors were immediately placed in Tissue-Tek OCT compound (Sakura Finetek USA, Inc.) to be stored at −80°C. Staining was done as previously described (4).

Mice with tumors were treated with one of the four treatment options as above. Tumors were removed 2 d after the boost vaccine, cut into pieces of ∼5 × 5 mm, weighed, and placed in a 24-well plate with 800 μL of culture medium. After 24 h, the medium was collected and spun to remove cellular debris (5 min, 1,500 rpm). The amount of tumor necrosis factor-α (TNF-α) secreted by tumors (corrected for weight) was quantified using an ELISA kit according to the instructions of the manufacturer (BD OptEIA ELISA set, BD Biosciences).

For the reverse transcription-PCR (RT-PCR), FACS studies, and flank tumor studies comparing differences between two groups, we used unpaired Student's t tests. For FACS and flank tumor studies comparing more than two groups, we used one-sided ANOVA with appropriate post hoc testing. Differences were considered significant when P < 0.05. Data are presented as mean ± SE.

We initially evaluated the effect of blocking each of the two murine CCL2 orthologues, CCL2 and CCL12, on tumor growth. Using mAb blocking either of the two orthologues alone, there was a nonsignificant trend for slower tumor growth (Fig. 1A). However, the combination of mAb against both CCL2 and CCL12 consistently inhibited tumor growth by 30% to 50% (P < 0.05) in the TC1 non–small cell lung cancer (NSCLC) cell line (Fig. 1A) as well as in several other flank and orthotopic mouse lung cancer and mesothelioma models.⁴

Figure 1B shows the effects of α-CCL2 alone, the Ad.E7 vaccine alone, or the combination of treatments on TC1 tumors. Both the antibodies alone and Ad.E7 alone significantly slowed tumor growth (P < 0.05 versus control for each) without causing regressions. However, combining these agents led to clear tumor regressions and an ∼50% cure rate in two independent experiments. Although some effect of the combination with immunotherapy was noted when we used either anti-CCL2 alone or anti-CCL12 alone (Supplementary Fig. S1), the effect was less than when the two mAbs were combined (Fig. 1B), suggesting that blocking both CCL2 and CCL12 is important in augmenting immunotherapy. Further mechanistic studies were therefore done with the combination of both mAbs.

Figure 1C shows similar results when we combined α-CCL2 mAb with Lm.Meso in the AE17.hmeso mesothelioma model. Again, either antibodies alone or Lm.Meso alone significantly slowed tumor growth (P < 0.05 versus control for each) without regressions. However, combining these agents led to tumor regressions and two thirds of the animals were cured (Fig. 1C). To examine the effect of blocking CCL2 in a nonimmunogenic tumor, we used the NSCLC cell line LKR, treated with Ad.IFNα (Fig. 1D). As in the other two models examined, the combination of the vaccine with CCL2 blockade significantly slowed tumor growth (P < 0.05 versus all other groups).

For mechanistic studies, we focused on the TC1/Ad.E7 system. We first studied systemic effects by evaluating splenocytes.

To assess the effect of the combination therapy on CD8⁺ T-cell activation, we measured the expression of the surface activation marker 4-1BB (CD137; refs. 3, 20) in CD8⁺ cells from spleens harvested from mice 2 days after the boost Ad.E7 vaccine (Fig. 2A). The percentage of CD8⁺ T cells expressing 4-1BB in the spleens of mice treated with α-CCL2 mAb or Ad.E7 alone was not increased compared with control mice. However, in tumor-bearing mice treated with the combination therapy, we found a 2-fold increase in the percentage of activated CD8⁺ T cells (CD8⁺/4-1BB⁺) out of total splenocytes (P = 0.05; Fig. 2A).

The murine NSCLC line TC1, which expresses the HPV-E7 peptide, enabled us to directly evaluate the reactivity of CD8⁺ T cells to a specific tumor antigen (HPV-E7) by flow cytometry using tetramers (4). CCL2 blockade by itself did not change the percentage of antigen-specific CD8⁺ cells (0.4 ± 0.1%) in the spleen compared with control tumor-bearing mice (0.3 ± 0.1%). Ad.E7 immunotherapy increased the percentage of splenic E7-specific CD8⁺ cells ∼4-fold to 1.1 ± 0.2% of CD8⁺ cells. Addition of CCL2 blockade to Ad.E7 significantly increased the percentage of E7-reactive CD8⁺ cells to 2.3 ± 0.5% of CD8⁺ cells (P < 0.05; representative tracings in Fig. 2B). The total calculated number of specific CTLs showed the same pattern (P < 0.02; Fig. 2C).

We next evaluated the changes in Treg cells in the spleen, defined by flow cytometry as CD4⁺/FoxP3⁺. The percentage of FoxP3⁺ cells out of CD4⁺ cells in the spleen was slightly, but significantly, reduced from 19.2 ± 0.6% in the Ad.E7-treated mice to 17.4 ± 0.6% in the combination-treated mice (n = 12; P < 0.05). When the total number of Treg cells in the spleens was calculated, we did not see significant changes in the number of splenic Treg cells among control, α-CCL2–treated, and Ad.E7-treated mice. However, the combination treatment resulted in a reduction in their number in ∼2-fold (P < 0.01; Fig. 2D).

Finally, we evaluated splenic CD11b⁺/GR1⁺ cells, generally accepted to be myeloid-derived suppressor cells (MDSC). CD11b⁺/GR1⁺ cells made up to 15% of the splenocytes. However, we saw no differences between the four treatment groups in the number of these cells in the spleens. We further evaluated if there were changes in the numbers of the two subsets of CD11b⁺/GR1⁺ as described recently (21). We saw no changes in the CD11b⁺/Ly6G⁺ granulocytic population nor in the CD11b⁺/Ly6C⁺ monocytic population (data not shown).

We next evaluated the total number of CD8⁺ cells infiltrating the tumors. We evaluated the tumors 2 days following the boost vaccination with Ad.E7, a time point at which there was no significant change in tumor size. CCL2 blockade as monotherapy had no significant effect on the number of intratumoral CD8⁺ cells. In contrast, the Ad.E7 vaccine induced a significant influx of CD8⁺ cells into the tumor. However, the combination of vaccine with CCL2 blockade significantly (P < 0.05) increased intratumoral CD8⁺ cells compared with vaccine alone, as seen by immunohistochemistry (Fig. 3A), flow cytometry (Fig. 3B), and by evaluating the expression levels of CD8 mRNA in the tumors using real-time RT-PCR (Table 1).

The percentage of activated intratumoral CD8⁺ T cells (4-1BB⁺) out of CD8⁺ cells was 2-fold higher following combination therapy compared with mice treated only with mAbs or immunotherapy (P < 0.05; Fig. 3C; Supplementary Fig. S2A).

Immunotherapy or α-CCL2 mAb alone increased the percentage of intratumoral E7-specific CD8⁺ cells by 3- to 4-fold as shown by tetramer staining (Fig. 3D). However, combining α-CCL2 mAb with Ad.E7 vaccine significantly increased the percentage of tetramer-positive CD8⁺ cells up to 7.2-fold compared with control (P < 0.05, compared with all other groups). Representative tracings are shown in Supplementary Fig. S2B.

There was only a small nonsignificant reduction in the percentage of total CD4⁺ cells in the tumors with the combined treatment (data not shown). By normalizing the percentage of Treg cells out of CD4⁺ cells to those of control tumors, we found a small nonsignificant reduction of Treg cells in the α-CCL2–treated mice and no change with Ad.E7. However, the combination treatment significantly reduced the percentage of intratumoral Treg cells to about half of control (P < 0.01, compared with Ad.E7). The data on intratumoral Treg cells are summarized in Fig. 4A, with examples of FACS tracings shown in Fig. 4B. We also compared the percentage of the CD4⁺ cells within the tumors that were negative for FoxP3, suggesting that these are activated CD4⁺ T cells (Fig. 4C). Ad.E7 or CCL2 blockade alone did not change the percentage of these cells. However, the combination of Ad.E7 with α-CCL2 mAb significantly increased the percentage of these CD4⁺/FoxP3⁻ cells by ∼50% (P < 0.05, for the combination treatment versus Ad.E7 alone).

We next sought to look for major changes in the phenotype of tumor-associated macrophages (TAM). Treatment with Ad.E7 increased the total number of TAM (data not shown) and increased the percentage of alternatively activated macrophages (defined as CD11b⁺/F4/80⁺/CD206⁺) out of total tumor cells by ∼50% compared with control tumors (P < 0.05; Fig. 4D). However, Ad.E7 also increased the percentage of the CD206⁻ macrophages (defined as CD11b⁺/F4/80⁺/CD206⁻), which likely represent classically activated “M1” macrophages, by ∼3-fold (Fig. 4D). Although treatment with mAbs alone did reduce the percentage of CD206⁺ macrophages, suggesting polarization of the macrophages toward an M1 phenotype compared with control (data not shown), we did not find a difference in the level of either CD206⁺ or CD206⁻ macrophages in the combined therapy compared with immunotherapy alone (Fig. 4D). We further calculated the ratio of mRNA of typical cytokines secreted from classically and alternatively activated macrophages. The ratio of interleukin (IL)-12/IL-10 was reduced by 16% compared with control in the tumors from mice treated with Ad.E7 alone and by 17% in the mice treated with the combination therapy. The lack of changes in the percentage of CD206⁺ cells and the similar small changes in IL-12/IL-10 ratio suggest that changes in TAM phenotypes were not the main mechanism of the augmentation of immunotherapy by α-CCL2 mAb.

Finally, we evaluated changes in the tumor microenvironment induced by CCL2 blockade that could explain the increased numbers and activation of the CD8⁺ CTLs.

We used real-time RT-PCR of tumor extracts to profile a set of relevant cytokines, chemokines, and cell adhesion molecules (Table 1). CCL2 blockade induced only minor changes in our panel, with the exception of a 2.3-fold increase in CXCL10 (IP-10). In Ad.E7–treated tumors, the mRNA expression levels of TNF-α, CCL5, and CXCL10 all increased ∼2-fold and the IFN-γ level increased 4-fold. However, expression levels of TNF-α, IFN-γ, IL-10, IL-12, ICAM, and CXCL10 were ∼2-fold higher in the combination group compared with the Ad.E7 alone group (P < 0.01). IL-12 and ICAM-1 were only increased in the combination therapy group.

We then measured the amount of TNF-α secreted from whole tumor explants. Although no significant change in the secretion of TNF-α was noted in tumors from mice treated with α-CCL2 mAbs or Ad.E7 alone, tumors from mice treated with the combination secreted >3-fold the amount of TNF-α (P = 0.017; Table 1, bottom) compared with the other treatments.

In recent years, immunotherapy strategies have been aggressively pursued to enhance antitumor immune responses and many phase II and III clinical trials have been conducted (22). Although a variety of immunotherapeutic approaches have been shown to generate active CTLs, success in patients has been limited. It has become increasingly clear that the generation of CTL is necessary but not sufficient for an effective response (22, 23). There may be several reasons for this. First, in addition to inducing immune stimulation, cancer immunotherapies also seem to trigger counterregulatory immunosuppressive mechanisms such as upregulation of inhibitory surface molecules on T cells (such as CTLA-4 or PD1) or production of Treg cells (24, 25). Second, tumors are known to produce inhibitory cytokines and chemokines (26, 27) as well as induce populations of suppressor cells (28). Thus, it is becoming increasingly apparent that in addition to the generation of CTL, successful immunotherapy will also require “inhibiting the inhibitors.” The studies presented here, using specific anti-murine CCL2 and CCL12 mAb in three different models of immunotherapy, suggest that MCPs may be additional unrecognized key proximal cytokines able to block the immune responses elicited by immunotherapy.

Human CCL2 has two murine orthologues: CCL2 (MCP-1) and CCL12 (MCP-5). Both bind to the CCR2 receptor, although CCL2 is a better agonist of murine CCR2 (29). Most functions described for CCL12 are similar to those found for CCL2 (13). We found that each of these mAbs had some effect on tumor growth but saw significantly more growth inhibition when the two mAbs were combined by themselves (Fig. 1A) or in combination with immunotherapy (Supplementary Fig. S1). To most accurately model potential effects in humans, we therefore used a mixture of both mAbs for all of our experiments.

There has been some controversy in the literature about the role of CCL2 in tumor development. CCL2, originally identified as a potent chemoattractant for monocytes (7, 8), can also function as a T-cell chemoattractant and induce T-cell tumor tropism, including memory T cells (30–32). It seemed reasonable that CCL2 would thus function to inhibit tumor growth. Indeed, early work showed that transfection of tumor cells that secreted high levels of CCL2 resulted in massive monocyte/macrophage infiltration into the tumor mass, leading to its destruction (33). However, in patients, CCL2 has been found at high levels in multiple tumor types, including NSCLC (7, 34, 35), and high levels usually correlate with poor clinical outcome (36). Studies, such as those by Loberg and colleagues (8), showed that systemic administration of anti–CCL2-neutralizing antibodies significantly retarded tumor growth. The use of α-CCL2 mAb in mice has been recently shown to reduce tumorigenesis and metastasis in prostate cancer xenograft models (37).

These observations support mounting evidence suggesting that most of the effects of CCL2 in nontransduced tumors are actually protumorigenic (7). First, it is now recognized that most monocytes recruited into tumors do not kill tumor cells but are subverted to an M2 phenotype where they actually support tumor growth (38). Second, CCL2 seems to directly augment the growth and invasiveness of certain tumor cells that express the CCR2 receptor (7, 39). Third, CCR2 is expressed by endothelial cells and CCL2 seems to promote angiogenesis (40). Fourth, it has been observed that CCL2 can also serve as a chemoattractant for Treg cells (41, 42). Finally, it is now recognized that CCL2 also has direct immunoinhibitory (protumorigenic) effects on T-cell function (10, 11), such as inhibiting T-cell effector functions and switching T-cell differentiation toward T helper 2 (Th2)–like cells (12).

We noted that vaccines induced an influx of macrophages, compared with control, which was not prevented with CCL2 blockade. It was somewhat surprising that neutralization of a major chemokine attracting monocytes to the tumor did not reduce the total number of monocytes. We have no definitive answer to this question, but we speculate that in tumors there are many other agents (including CCL5, CCL7, CCL8, CXCL8, CXCL12, CXCL1, macrophage colony-stimulating factor, and vascular endothelial growth factor) that can replace MCP-1 in terms of chemoattraction of monocytes (38). Another possibility is that inducing CD8⁺ activity and reducing Treg cells changed the total balance of chemokines in tumor microenvironment, allowing for the influx of monocytes to continue based on other chemokines.

Our mechanistic studies in the TC1 model show that CCL2 blockade in combination with the Ad.E7 vaccine clearly results in increased numbers, activity, and antigen specificity of CD8⁺ T cells and in the percentage of activated CD4⁺ T cells in the spleens (Fig. 2) and, even more importantly, intratumorally (Fig. 3). These data are consistent with those of Peng and colleagues (10) in adoptive transfer studies, who used neutralizing α-CCL2 antibodies and showed that this led to the generation of T cells that were substantially more active and more vigorous at eliminating tumor, suggesting increased tumor specificity, and later studies, showing that prevention of tumor secretion of CCL2 had a positive effect on CTLs that were more active in both spleen and draining lymph nodes (43, 44).

It is likely that some of these activating effects on CD8⁺ T cells were due to blockade of the direct effects of CCL2 on the T cells; however, we also explored other factors that might be involved. First, we did not note any changes in the levels of CD11b⁺/GR1⁺ cells, generally accepted as MDSC. In our studies, we did find some effects of α-CCL2 mAb alone on TAM phenotype;⁴ however, we could not implicate these changes in explaining the augmented effects we saw in combination therapy because there were no obvious differences in macrophage phenotype populations when we compared Ad.E7 treatment alone with combination treatment (Fig. 4D).

In contrast, we found significant reductions in Treg cells in the spleens (Fig. 2D) and tumors (Fig. 4A and B), of mice treated with the combination therapy versus vaccine alone suggesting an important possible mechanism for reduction of immune suppression. The idea that Treg cells are important inhibitors of antitumor immune responses is well established (45), and their presence correlates with poor prognosis of cancer patients (46). The finding that CCL2 blockade could inhibit Treg recruitment is consistent with previous studies showing that (a) CD4⁺ Treg cells selectively overexpress the CCL2/CCL12 receptors CCR2 (47) and CCR4 (48, 49), (b) CCL2 has been shown to specifically chemoattract Treg cells in vitro (41, 42), and (c) blocking CCR2 in vivo reduced the influx of Treg cells to disease sites in a model of arthritis (47). Interestingly, in the TC1 tumor model, depletion of CD4 cells using a specific mAb leads to slower growth (data not shown), suggesting that these tumors do induce Treg cells that then augment their growth.

Given their strong immunoinhibitory properties, reduction of Treg cells has been a goal of many groups. To date, most attempts to reduce Treg cells have used nonspecific agents such as low-dose cyclophosphamide or antibodies/antibody toxins directed toward the IL-2 receptor (CD25). Targeting CD25 may have disadvantages, however, because it is also expressed on activated CD8⁺ CTLs (45, 46). Our data suggest that a novel and possibly safer way to prevent the influx of Treg cells into the tumor microenvironment may be via CCL2 blockade. This may be particularly important when the strong immune reaction induced by vaccines is also accompanied by a strong induction of Treg cells.

Finally, we also found that the tumor microenvironment was altered in the combination-treated tumors with increased mRNA levels of Th1 type mediators, such as TNF-α, IFN-γ, CXCL10, and ICAM-1 (Table 1), and protein levels of TNF-α. It is currently uncertain if this is a direct result of CCL2 blockade leading to enhanced T-cell activation or whether increased numbers of activated CD8⁺ T cells result in a more immunostimulatory microenvironment.

In summary, we showed here that blocking CCL2 dramatically augmented the effect of immunotherapy for NSCLC and mesothelioma in a multifactorial immunologic mechanism. Our observations suggest that combining CCL2 neutralization with vaccines should be considered in future immunotherapy trials.

L.A. Snyder is an employee of Centocor, Inc. S.M. Albelda: commercial research support, Centocor, Inc. The other authors disclosed no potential conflicts of interest.

We thank Luana Pereira and Aaron Blouin for technical assistance in the experiments.

Grant Support: National Cancer Institute grant PO1 CA 66726 and Centocor, Inc.

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.