Locally Delivered CD40 Agonist Antibody Accumulates in Secondary Lymphoid Organs and Eradicates Experimental Disseminated Bladder Cancer

In 2008, approximately 380,000 patients worldwide were diagnosed with bladder cancer and 150,000 succumbed to the disease (1). The majority of these patients present with superficial transitional cell carcinoma (Tis, Ta, or T1) that is treated with transurethral resection followed by intravesical chemotherapy, or for high-risk tumors local Bacillus Calmette-Guérin (BCG) immunotherapy (2, 3). BCG treatment induces remission in a majority of patients but is associated with adverse events and 30% to 50% of the patients ultimately fail to respond (4). Radical cystectomy for patients with muscle-invasive bladder cancer results in initial tumor control, but provides a 5-year survival rate of only 40% to 60% due to the presence of micrometastatic disease (2). We have previously shown that local adenoviral CD40L (AdCD40L) as well as CpG therapy can induce systemic antitumor responses in experimental bladder cancer (5, 6). AdCD40L therapy has proven efficient and safe in both humans (7) and dogs (8). Consequently, we wanted to investigate whether local low-dose agonistic CD40 antibody injection could clear experimental bladder cancer. CD40 is expressed by various cell types in the myeloid cell lineage, i.e., dendritic cells, macrophages and monocytes, B cells as well as endothelial cells (9), and dendritic cells are recognized as important targets for anti-CD40 antibody in cancer immunotherapy. Activated dendritic cells can efficiently engulf, process, and present tumor antigen to T cells, resulting in powerful antitumor responses (10, 11). In addition to the antigen-presenting cell (APC)–activating properties of anti-CD40 antibody, CD40 ligation on CD40-expressing tumor cells initiates programmed cell death (12), antibody-dependent cellular cytotoxicity (13), antibody-dependent cellular phagocytosis (14), or complement-mediated cytotoxicity (15), depending on the immunoglobulin G (IgG) subclass.

Several CD40 agonistic antibodies have been developed for clinical use by the intravenous administration route (16–18). The resulting systemic immune activation has been associated with considerable toxicity, including grade 1 to 2 cytokine release syndrome and transient lymphopenia (16, 19). This study extends the observations of Fransen and colleagues, who compared local low-dose anti-CD40 therapy in combination with the slow-release adjuvant Montanide of virally transformed murine tumors. Local low-dose administration was as effective as systemic high-dose therapy, but with reduced liver toxicity (20). In the current study, we use an experimental bladder cancer model, a tumor well suited for local immunotherapy for which novel therapies are needed to target disseminated disease. We wanted to investigate whether locally administered anti-CD40 antibody could be as efficient as an equivalent dose delivered systemically, and to establish whether the alternative administration route could reduce toxicity. Subsequently, we aimed to pinpoint the effector mechanisms behind the successful local CD40-stimulating immunotherapy for bladder cancer and, lastly, to compare the in vivo biodistribution between the two administration routes.

The murine bladder transitional cell carcinoma cancer cell line mouse bladder-49 (MB49; a kind gift from Dr. K. Esuvaranathan, National University of Singapore, Singapore in 1996) was cultured in DMEM+ GlutaMax supplemented with 10% FBS, 0.1 mmol/L sodium pyruvate, 100 U/mL penicillin–streptomycin (PEST) at 37°C and 5% CO₂. Lewis lung cell carcinoma-1 (LLC-1;American Type Culture Collection) was kept in the same culture medium as MB49. The D1 cell line (21) is a growth factor–dependent immature splenic mouse dendritic cell line cultured in Iscove's modified Dulbecco's medium (IMDM) supplemented with 10% FBS, 100 U/mL PEST, 100 μmol/L β-mercaptoethanol (Invitrogen), and 20 ng/mL recombinant murine granulocyte macrophage colony–stimulating factor (GM-CSF; Nordic Biosite) at 37°C and 5% CO₂. All cell lines tested negative for Mycoplasma but were not authenticated in our laboratory. Agonistic rat-anti-mouse CD40 (clone: FGK4.5) and rat IgG2a (clone: 2A3) were purchased from BioXCell and diluted in PBS. Depletion of CD8⁺ cells was performed by injecting 20 μg/g body weight rat-anti-mouse CD8a (clone: 53.6.72, BioXCell) intraperitoneally on days 0, 1, 2, 6, 10, and 14. CD8⁺ T-cell depletion was confirmed by flow cytometry by staining with clone 53-5.8 (data not shown).

C57BL/6 mice were obtained from Taconic M&B. CD40^−/− knockout C57BL/6 mice were obtained from The Jackson Laboratory (B6.129P2-Cd40^tm1Kik/J). Animals were housed at the Rudbeck Animal Facility and cared for by the staff according to regional regulation. All animal experiments were approved by the Uppsala Animal Ethics Committee (Dnr: C303/9, C21/10, C11/11, and C38/11).

Several variants of subcutaneous MB49 tumor models were used in this study. In the majority of experiments, 2.5 × 10⁵ cells were injected in the right flank of C57BL/6 mice with therapy conducted on days 7, 10, and 13. For the biodistribution study, 2 × 10⁵ MB49 cells were inoculated in the right and left flank on day 0. Ten days later, 30 μg of radioactively labeled ¹²⁵I-CD40 antibody and ¹³¹I-rat IgG2a were mixed and injected once intravenously or peritumorally at the right side of the tumor. Animals were sacrificed 4, 24, 48, and 72 hours after injection, organs were isolated, and their radioactivity was measured. Radioactivity of ¹²⁵I-CD40 antibody was measured in the energy window of 3 to 6 keV and ¹³¹I-rat IgG2a of 100 to 380 keV. In the last tumor model investigating the systemic effects of low-dose anti-CD40 therapy, 2.5 × 10⁵ MB49 cells were injected in the right flank on day 0 and close to the left shoulder on day 1. Therapy was injected peritumorally at the primary tumor, subcutaneously in the nontumor flank, or intravenously every third day for a total of three times. To keep a similar distance between injection site and the distant tumor as well as to reduce passive diffusion of antibodies in the void space of the skin, animals injected with subcutaneous anti-CD40 antibody in the nontumor flank had both tumors inoculated on the same side of the animal, opposite to the injection side (Fig. 2E). Antibody solution was administered in 100 μL. Tumor growth and survival were monitored throughout the experiment using a caliper and tumor size was calculated by the ellipsoid volume formula: = 4/3×π×a (radius of length) × b (radius of width) × c (radius of depth). Mice were sacrificed if the tumor exceeded 1 cm³ or if ulcers developed. Tumor rechallenge, by injection of 2.5 × 10⁵ MB49 (contralateral flank) and LLC-1 cells (right foreleg), was performed on mice that had been tumor free for over 100 days.

¹²⁵I and ¹³¹I were purchased from PerkinElmer. Chloramine-T and sodium metabisulfite were from Sigma Chemical Company. Chloramine-T and sodium metabisulfite solutions were prepared immediately before use. The radiochemical purity of the labeled antibody construct was analyzed using instant thin layer chromatography (ITLC) on 150-771 DARK GREEN, Tec-Control Chromatography Strips from Biodex Medical System. Distribution of radioactivity along the ITLC strips was measured using a Cyclone Storage phosphor system and analyzed with the OptiQuant image analysis software (PerkinElmer). Size-exclusion chromatography was performed on disposable NAP-5 columns (Amersham Pharmacia Biotech AB) according to the manufacturer's instructions. The radioactivity was measured using an automated gamma-counter with a 3-inch NaI (Tl) detector (1480 Wizard; Wallac Oy). Monoclonal antibodies were labeled using Chloramine-T as an oxidant according to the following protocol. Anti-CD40 antibody was labeled with ¹²⁵I. An antibody solution in PBS (40 μg, 4–2 μL) was mixed with radioiodine stock solution (2–5 μL, 5–10 MBq) and 40 μL PBS. The reaction was initiated by adding Chloramine-T (20 μg, 1 mg/mL in PBS). After 2-minute incubation at ambient temperature, the reaction was terminated by adding sodium metabisulfite (40 μg, 2 mg/mL in PBS). Rat IgG2a was labeled with ¹³¹I in a similar way. Radiolabeled antibodies were purified from unreacted radioiodine and low-molecular-weight components of the reactive mixture using size-exclusion chromatography on disposable NAP-5 columns. Radiochemical purity of the labeled antibodies was determined by radio-ITLC eluted with acetone:water (8:2) mixture. The yields were in the range of 65% to 75%. After size-exclusion purification, the radiochemical purity of the antibodies was more than 99.5%.

Trichloroacetic acid (TCA) precipitates high-molecular-weight molecules. Precipitation was carried out according to the protocol described previously (22). Briefly, on top of 300 μL of carrier solution (PBS/0.02% BSA, w/v), 200 μL of plasma was added. Then, 500 μL of ice-cold 20% TCA/H₂O (w/v) solution was added to precipitate high-molecular-weight molecules in the plasma. Tubes were centrifuged and separated before individual measurement for radioactivity using an automated gamma-counter with a 3-inch NaI (Tl) detector.

Anti-CD40 antibody was conjugated with Alexa Fluor 488 protein-labeling kit according to the manufacturer's protocol (Invitrogen). The functionality of the labeled antibody was verified by staining of A20 cells (data not shown). D1 cells were seeded in 96-well plates and stimulated overnight with 1 μg/mL lipopolysaccharide (LPS; Sigma) to upregulate the expression of the CD40 receptor. Cells were incubated with 0.1 μg/mL Alexa Fluor 488-conjugated anti-CD40 antibody for 15 or 60 minutes at 37°C or 4°C or left untreated before being fixed in 3%PFA/PBS on ice, and then washed twice in 0.5%PFA/PBS. Trypan blue (Invitrogen) was added just before flow-cytometric analysis to quench any surface-bound antibody. Samples were run in triplicates.

To evaluate which cell populations could potentially act as targets for anti-CD40 therapy, naïve animals were injected subcutaneously with 30 μg of anti-CD40 one, two, or three times at 3-day intervals. Four hours postinjection the spleen and (pooled) inguinal lymph nodes were harvested and digested with Liberase TL (Roche) for 15 minutes at 37°C. The Liberase-treated tissue was passed through a MESH membrane, blocked for FcγR (TruStain fcX), and subsequently stained for CD11c (clone: N418), CD11b (clone: M1/70), F4/80 (CI:A3-1), B220 (RA3-6B2), and CD40 (3/23; all antibodies are from BioLegend). Staining with antibody clone 3/23 was not hampered by the presence of the therapeutic clone FGK4.5 on cells (data not shown). The following cell types were investigated: B cells (B220⁺, CD11b⁻, CD11c⁻, and F4/80⁻), conventional dendritic cells (cDC; CD11c^hi, CD11b^+/−, and B220⁻), medullary macrophage (MΦ; CD11c^low/int, CD11b⁺, B220⁻, and F4/80⁺), and subcapsular sinus MΦ (CD11c^low/int, CD11b⁺, B220⁻, and F4/80⁻). Samples were analyzed in a FACSCanto II cytometer (BD Biosciences). Data analysis was performed with FlowJo software (TreeStar).

Blood was collected by tail vein incision and serum was stored at −80°C. Mouse haptoglobin was detected by ELISA according to the manufacturer's protocol (Life Diagnostics, Inc.). Quantification of circulating anti-CD40 antibody was performed by ELISA. Briefly, wells were coated with 1.25 μg/mL anti-rat IgG2a (clone: MARG2a-1; Serotec) and blocked with 3% milk powder before the addition of samples (1:10–1:75 dilution). Horseradish peroxidase (HRP)–conjugated mouse-α-rat κ/λ light chain (MARK-1/MARL-15; Serotec) was used for detection. Substrate (Super Signal pico chemoluminescence; Pierce) was added before the luminescence was measured with an ELISA reader (Fluostar optimal; LabVision). Samples were run in duplicates.

Survival data were plotted by the Kaplan–Meier method and analyzed by the log-rank test. When applicable, the D'Agostino and Pearson omnibus normality test was used before selection of statistical test. Where indicated, the difference between groups was evaluated using unpaired t test or paired t test. P values less than 0.05 were considered significant. Asterisks indicate the confidence interval (*, P < 0.05; **, P < 0.01; ***, P < 0.001). Statistical analyses were performed using GraphPad prism software (GraphPad Software, Inc.).

To our knowledge, a direct comparison between local and systemic administration of equal doses of anti-CD40 therapy has not been performed. Ten and 30 μg of peritumorally delivered antibody prolonged survival compared with control animals, whereas systemic therapy of the same doses did not (Fig. 1A, top and middle). Only the highest systemic dose (100 μg) resulted in prolonged survival that was similar to that in the local therapy-treated animals (Fig. 1A, bottom).

Several immune-related side effects have been observed with systemic anti-CD40 antibody therapy in both preclinical models (23) and patients (19). In our current study, levels of circulating therapeutic antibody were measured 4 hours after each injection (Fig. 1B). For all doses and administration routes, the serum maximum level of anti-CD40 antibody occurred after the second injection, with the highest concentrations in animals treated systemically with 30 or 100 μg of antibodies. Local immunotherapy led to reduced serum levels of antibody in all groups compared with that of the equivalent systemic dose. The initial reduction in systemic antibody levels could translate to reduced toxicity. In the current studies, we used haptoglobin, an acute phase protein produced mainly by the liver, as a marker for systemic inflammation (24–26). Serum levels of haptoglobin were measured 24 hours after the first dose, and elevated levels were detected in mice treated with intravenous injections (Fig. 1C). Local injection of the same dose did not elevate haptoglobin levels to the same degree, and neither peritumorally nor intravenously administered irrelevant antibody affected haptoglobin levels (data not shown).

The local low-dose strategy was also assessed using the fast-growing B16-F10 tumor model where tumor growth inhibition was evident compared with control animals (Supplementary Fig. S1A). Furthermore, weekly dosing with peritumorally low-dose anti-CD40 was evaluated in a high MB49-tumor burden model. Both the weekly and the 3-day interval treatment schedule restrained tumor growth. However, elevated levels of ADA were detected in the serum already after the second injection of anti-CD40 antibody on a weekly schedule, which was not seen using the 3-day interval protocol (data not shown).

CD40 is widely expressed on different cell types in vivo, and a variety of cells have been proposed to be effectors in CD40 agonistic therapy. Macrophages (27, 28), B cells (29), natural killer (NK) cells indirectly (30), and T cells (20) via dendritic cell activation have been suggested to play crucial roles in tumor eradication. We asked whether our local low-dose administration of anti-CD40 antibody would depend on the presence of CD8⁺ T cells by using cell-depletion experiments. CD8⁺ cells were necessary for optimal antitumor effects as depleted mice showed no survival benefit compared with controls (Fig. 2A). Furthermore, depletion of CD4⁺ T cells yielded variable results as 6 of 10 animals experienced rapidly growing tumors and 3 of 10 had recurrence of tumors after initial remission (data not shown). Removal of regulatory T cells (Treg) cannot be excluded. Previous studies have shown that the CD40–CD40L interaction not only activates APCs but generates a direct antitumor effect in CD40-expressing tumors (12, 31). Throughout this study, MB49 cells presented a negligible amount of CD40 expression when staining with the therapeutic antibody clone (data not shown). Despite this result, we wanted to exclude tumor cell death induced directly by the antibody and therefore subjected tumor-bearing C57BL/6 wild-type (WT) and CD40^−/− knockout mice to local or systemic treatment. Figure 2B shows that host CD40 expression is required for effective anti-CD40 therapy regardless of administration route. The knockout mice had elevated levels of circulating anti-CD40 antibody 4 hours after the last treatment compared with WT animals (Figs. 2C and 1B).

To establish tumor-specific memory after local anti-CD40 therapy, animals exhibiting complete tumor regression were rechallenged with MB49 cells and LLC-1 (Fig. 2D). LLC-1 tumors grew progressively in both naïve and cured mice, whereas progressive growth of MB49 tumors was only seen in naïve animals.

Because low but detectable levels of anti-CD40 antibody were found in the blood after local administration, we wanted to investigate whether distant tumors could be cleared and whether tumor antigens are required at the injection site, as demonstrated earlier by Fransen and colleagues (20). To minimize systemic antibody dissemination, 15 μg of anti-CD40 antibody was used but without a slow-release formulation. A twin-tumor model was set up by engrafting MB49 cells in the right flank on day 0 and close to the left shoulder on day 1 (Fig. 2E). As visualized in Fig. 2F, anti-CD40 antibody needs to be injected in close proximity to a tumor (or systemically) to affect both the treated site as well as the distant tumor. Therapy injected solely in the skin (nontumor side) did not inhibit tumor growth. For individual tumor growth, see Supplementary Fig. S1B.

We determined the level of effector cytokines in the serum and registered a dose-dependent elevation of serum IFN-γ, IP-10, and TNF-α 4 hours after the second dose using all administration routes (data not shown). In agreement with the clinical observation using the humanized CP-870,893 antibody (16, 19), we registered a transient B-cell depletion after a single injection of anti-CD40 antibody (data not shown), which was less pronounced after local administration.

To investigate the in vivo biodistribution of anti-CD40 agonist or control antibody upon local versus systemic administration, we measured antibody accumulation in target organs after a single injection. For this purpose, anti-CD40 antibody was labeled with ¹²⁵I and the irrelevant isotype with ¹³¹I to allow simultaneous measurement in the same animal by gamma-spectrometry. The binding-specificity of the ¹²⁵I-CD40-specific antibody was confirmed before experimental initiation (Supplementary Fig. S2A). An equal dosage mixture of ¹²⁵I-CD40–specific and ¹³¹I-rat IgG2a control antibodies was injected into mice carrying a tumor in each flank either peritumorally at the right side tumor or intravenously. At 4, 24, 48, and 72 hours after injection, animals were sacrificed, organs isolated, and radioactivity measured. A high accumulation of ¹²⁵I-CD40–specific antibody was detected in the spleen 4 hours after injection followed by a decrease over time (Fig. 3A). ¹³¹I-rat IgG2a control antibody uptake in the spleen was significantly lower.

The highest amount of locally administered anti-CD40 antibody was found in the spleen at 24 hours, followed by a decline. The uptake of ¹³¹I-rat IgG2a control antibody in the same animal was significantly lower than that of the ¹²⁵I-CD40–specific antibody (Fig. 3A). The spleen accumulation was significantly lower after peritumoral administration than after intravenous injection at 4 and 24 hours. At subsequent time points, the radioactivity uptakes were equal for both administration routes. Anti-CD40 antibody uptake in the right tumor-draining lymph node (TDLN) after peritumoral injection was much higher at 4 hours as compared with that of systemic administration (Fig. 3B). Systemic delivery of anti-CD40 antibody induced a more pronounced accumulation in the liver compared with peritumoral injection at 4 and 24 hours (Fig. 3C). Furthermore, local administration resulted in a lower initial antibody concentration in the blood at 4 and 24 hours (Fig. 3D), in agreement with our ELISA results. We found that the concentration of radioactivity in the plasma was higher than in whole blood, indicating minor (if any) binding of ¹²⁵I-CD40 to blood cells (Supplementary Fig. S2B). In summary, peritumoral delivery significantly reduced anti-CD40 antibody concentrations in the spleen, liver, and blood compared with intravenous injection at both 4 and 24 hours after injection.

Figure 4 shows antibody uptake for selected target organs at 4 (Fig. 4A) and 24 hours (Fig. 4B). Results from the 48- and 72-hour time points are visualized in Supplementary Fig. S3. Statistical differences between the anti-CD40 and the control antibody and between the two administration routes in all organs investigated are summarized in Supplementary Tables S1–S4.

The anti-CD40–specific antibody showed a more rapid clearance from the blood than the control antibody. To verify how much of the blood-borne radioactivity was still conjugated, we performed a TCA precipitation assay. Although almost all ¹³¹I was conjugated to rat IgG2a in both treatment groups during the entire study period, a decrease of activity in the high-molecular-weight form was detected for the ¹²⁵I-CD40–specific antibody conjugate (Fig. 5A). This suggests that the anti-CD40 antibody was efficiently sequestered from the circulation due to specific interaction followed by subsequent internalization and catabolism. This could be by active uptake in cells and we therefore incubated fluorophore-conjugated anti-CD40 antibody with CD40-expressing dendritic cells for 15 or 60 minutes and subsequently quenched surface-bound antibodies. Our findings imply that anti-CD40 antibody is internalized within 15 minutes (Fig. 5B). As several cell populations in lymphoid organs, including B cells, CD8⁺, CD8⁻ cDCs, macrophages, NK cells, NKT cells, and activated CD4⁺ T cells, can act as potential targets for CD40-activating therapy, we investigated the expression of CD40 on immune cells from the inguinal lymph nodes (Fig. 5C) and the spleen (Supplementary Fig. S2C) 4 hours after one, two, or three subcutaneous injections of 30 μg anti-CD40. Lymph node B cells, cDCs, and macrophages all increased their surface expression of CD40 in a dose-dependent manner following anti-CD40 injections (Fig. 5C, top). Also, elevated numbers of total cDCs and macrophages were found in response to therapy (Fig. 5C, bottom). For the spleen, the results were slightly more variable; however, there was an apparent drop in the numbers of monocyte-derived macrophage and dendritic cells after the third injection (Supplementary Fig. S2C).

Effective tumor immunotherapy needs to overcome the challenge of poorly immunogenic antigens (32), lack of danger signals (33), and the presence of an immunosuppressive milieu caused by both the tumor and the infiltrating cells (34). CTLs exhibit poor activation by improperly activated or immature dendritic cells (35, 36). The general principle of anti-CD40 antibody in tumor immunotherapy, as demonstrated earlier, is to create a Th-independent activation of CTLs by direct licensing of dendritic cells (10, 11, 37).

We used peritumoral injection as the local administration route in this study because this route confers minimal mechanical stimulation of the tumor. A comparison with intratumoral injection in tumors using the same treatment protocol as in Fig. 1 showed that there was no advantage to this approach over peritumoral delivery (data not shown).

A survival benefit was observed for mice treated with local low-dose anti-CD40 antibody (10 and 30 μg) compared with the same dose delivered systemically, which indicates that systemic injections in this dose range result in a low concentration of antibody in the tumor/TDLN axis that is insufficient for CTL activation. As no difference in survival was noticed at the 100 μg dose between the two administration routes, a high level of circulating antibody could potentially overcome this problem but with increased side effects. We registered elevated levels of serum haptoglobin as a sign of systemic inflammation (24–26) already at low intravenous anti-CD40 antibody doses. Furthermore, liver exposure to the antibody was appreciably higher after intravenous injection compared with peritumoral injection. These data suggest that local low-dose antibody delivery can reduce liver toxicity compared with systemic therapy, which could be valuable in the clinic as patients often display metastatic disease that could be targeted by recruiting immune cells rather than the anti-CD40 antibody itself.

Our finding that CD8⁺ T cells are important for the antitumor effects of anti-CD40 therapy is consistent with data from other groups (23). However, CD40 activation can target many other cell types, such as macrophages, which could participate in tumor eradication. The differences observed may be dose- or animal model-related. CD40-stimulating therapy can have a direct antitumor effect on CD40-expressing tumors by inducing apoptosis (12, 31). Throughout this study, MB49 cells presented a negligible level of CD40 expression based on staining with the therapeutic antibody. Notably, when staining with the anti-CD40 antibody clone 3/23 on MB49 cells, CD40 expression could be visualized. To confirm that locally delivered therapy was fully dependent on host CD40 expression, we made use of CD40^−/− knockout animals. Endogenous CD40 expression was required for full therapeutic effect regardless of injection site. We also confirmed the antitumor effects in the CD40-negative B16 melanoma model in WT mice. In contrast with our MB49 model, human bladder cancer cells usually present CD40 expression compared with the CD40-negative urothelium (38). Also, CD40 ligation in the form of cell-surface bound CD40L or Fcγ receptor cross-linked anti-CD40 antibody induces apoptosis in transitional bladder carcinoma (39). Engaging both the direct- and indirect-killing mechanisms for the treatment of bladder cancer could be even more efficient. These observations, together with the convenient intravesical access of bladder cancer, make this tumor a suitable target for localized anti-CD40 therapy.

The presence of tumor antigen in close proximity to the anti-CD40 antibody injection for systemic tumor eradication was demonstrated by Fransen and colleagues (20). Our data strengthen these results and extended them further by showing substantiated levels of the antibody in the regional lymph nodes, both at the injection site and at the nontreated side. These results suggest that tumor-localized anti-CD40 therapy can be effective for targeting disseminated bladder cancer.

We performed in vivo biodistribution studies and detected high initial levels of anti-CD40 antibody in the spleen after intravenous injection and in the right TDLN after peritumoral delivery. However, at later time points, accumulation of anti-CD40–specific antibody was observed in these lymphoid organs with both administration routes, which may reflect the abundance of CD40-positive target cells. The drainage of both anti-CD40 antibody and tumor antigens to the same TDLN could increase the possibility of tumor-specific T-cell triggering by locally activated APCs as indicated by Fransen and colleagues (20). Therefore, our demonstration of a high concentration of anti-CD40 antibody in the TDLN seems crucial for optimal anti-CD40 antibody therapy. In addition, our in vivo study revealed increased CD40 expression and expansion of target APCs in the lymph node after repeated anti-CD40 injections, likely reflecting that the lymph node is a key player in local anti-CD40 therapy.

The kinetics of peritumorally and intravenously delivered anti-CD40 antibody in the blood was determined by ELISA and by radiolabeled antibody. Locally injected antibody reached its maximum after 24 hours followed by exponential decrease. A subsequent injection caused even higher serum levels for both administration routes, but after the third treatment (day 13), anti-CD40 antibody levels were low or nondetectable. Drug elimination could be due to the induction of ADA, the inability of ELISA to measure ADA complexes, and/or increased cell/tissue expression of CD40 leading to rapid antibody clearance. ADA-dependent serum clearance has also been suggested in preclinical studies using cynomolgus monkeys (40). Furthermore, elimination of ¹²⁵I-CD40–specific antibody from the circulation was faster than that for the control, which likely reflects specific binding outside the blood (and not ADA-associated elimination) as these animals received only a single injection. We found higher amounts of circulating anti-CD40 antibody in CD40 knockout mice compared with WT indicative of target-specific sequestering of the antibody. These findings correlate well with the clinical findings of CD40-agonistic antibodies, such as the CP-870,893 antibody (19), and the weak CD40 agonist SGN-40 antibody (41), both display relatively short half-life. This suggests that there may be a target-specific clearance from the blood in humans, and it has been speculated that this reflects a large sink of CD40 molecules in vivo (16). Our results demonstrate that, in mice, this sink is primarily located in the secondary lymphoid organs. In the lymph node and spleen, the CD40-expressing cells are B cells, CD8⁺, CD8⁻ cDCs, a variety of macrophages, NK cells, NKT cells, and activated CD4⁺ T cells. Interestingly, we found that macrophages as well as dendritic cells both increased in numbers and upregulated their CD40 expression subsequent to repeated CD40 agonist injections, whereas B cells only upregulated CD40 expression in the lymph node. The result was slightly more variable for spleen cells, but the drop in monocyte-derived macrophages as well as dendritic cells could reflect a recruitment of APCs into the lymph node and/or tissues. Furthermore, our data showed that anti-CD40 antibody was rapidly (15 minutes) internalized by CD40-expressing dendritic cells, most likely contributing to reduced serum levels of the antibody. In our hands, tumor-targeting seems minimal as no major specific accumulation of radiolabeled anti-CD40 could be detected. This is also in agreement with the lack of CD40-staining on MB49 cells by flow-cytometric analyses using the labeled therapeutic antibody. The correlation between blood and tumor, where late time points (48 and 72 hours) demonstrate specific accumulation of irrelevant antibody, further supports the target-specific sequestering of antibodies.

Systemic anti-CD40 therapy has been used extensively in the clinic (16, 17, 19, 41) and in preclinical models, with immune-related adverse events such as systemic cytokine release syndrome and liver toxicity. Lately, the focus has shifted toward local administration, through which doses and consequently toxicity can be reduced, but potentially retaining the efficacy of the systemic antitumor effects (20, 23, 29, 42–44).

In summary, this is to our knowledge the first study exploring local and systemic anti-CD40 therapy side by side in a dose-comparison study and including an in-depth analysis of antibody biodistribution in vivo. Our data support the use of local low-dose delivery of anti-CD40 antibody as an alternative to systemic high-dose therapy. This should be a relevant immunotherapeutic approach in combinatorial treatments with current immunomodulatory agents for localized and disseminated bladder cancer, and possibly other tumor types as well.

L.C. Sandin, P. Ellmark, and S.M. Mangsbo have ownership interest in a patent. T.H. Tötterman has an ownership interest in a patent and is a consultant/advisory board member for Alligator Bioscience, Inc. and Immuvent Inc. No potential conflicts of interest were disclosed by the other authors.

Conception and design: L.C. Sandin, P. Ellmark, V. Tolmachev, T.H. Tötterman, S.M. Mangsbo

Development of methodology: L.C. Sandin, V. Tolmachev, T.H. Tötterman, S.M. Mangsbo

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L.C. Sandin, A. Orlova, E. Gustafsson, V. Tolmachev, T.H. Tötterman, S.M. Mangsbo

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): L.C. Sandin, A. Orlova, E. Gustafsson, V. Tolmachev, T.H. Tötterman, S.M. Mangsbo

Writing, review, and/or revision of the manuscript: L.C. Sandin, P. Ellmark, V. Tolmachev, T.H. Tötterman, S.M. Mangsbo

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L.C. Sandin, A. Orlova, E. Gustafsson, V. Tolmachev, T.H. Tötterman, S.M. Mangsbo

Study supervision: V. Tolmachev, T.H. Tötterman, S.M. Mangsbo

The authors thank Ann-Charlotte Hellström (IGP) and Anna Rosén (Alligator Bioscience, Inc.) for excellent technical assistance.

This study was supported by The Swedish Cancer Society (to T.H.Tötterman), the Swedish Research Council (to T.H. Tötterman), and FP7 MCA-ITN 317445 (to S.M. Mangsbo). P. Ellmark was sponsored by the Swedish Research Council (VR-IFA hosted by Lund University).

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.