Purification and amine-activation of murine α₂M was performed as previously described [7,8,10,11]. Human α₂M was purified from fresh, frozen human plasma, obtained from the American Red Cross (Durham, NC, USA), according to a previously published protocol [12].

Native human α₂M was converted to the “amine-activated” form (α₂M^*) by incubation in 200 mM ammonium bicarbonate at room temperature for 1 h. Buffer exchange into PBS was achieved using a 5 mL disposable de-salting column. Native human α₂M was converted to the “trypsin-activated” form (α₂M-T) by incubation with a 5-fold molar excess of trypsin (Worthington Biochemical, Lakewood, NH) for 1 h at room temperature. The Halt protease inhibitor cocktail (Thermo Scientific, Rockford, IL) was then added to inhibit proteolytic activity. Excess trypsin and protease inhibitors were removed by consecutive passes over a 100 kDa centrifugal concentrator (Pall Life Sciences, Ann Arbor, MI). Purified protein contained less than 10 pg of endotoxin per mg of protein, as determined by a commercial assay kit (Limulus Amebocyte Lysate Kinetic-QCL by Cambrex, Walkersville, MD).

Complexes of amine-activated α₂M^* and Alexa Fluor^Ò 647-conjugated ovalbumin (OVA) (Molecular Probes, Eugene, OR) were prepared as previously described [7]. Molar ratio of incorporation was approximately 3:1 OVA: α₂M^*, as determined by fluorescence quantification.

A 3-fold molar excess of OVA was incubated at 37˚C with either α₂M^* or α₂M-T for 0.5 to 24 h. Samples were then immediately analyzed by native PAGE.

The MO5 cell line, an OVA-transfected subclone of B16 melanoma, was a kind gift from Dr. Kenneth Rock (University of Massachusetts Medical School, Worcester, MA). MO5 cells were cultured in complete media supplemented with 2 mg/mL G418. Murine splenocytes were harvested and cultured as previously described [7].

Female C57/BL6 mice were obtained from Charles River Laboratories (Raleigh, NC). All mice were housed in the Duke University Animal Facility, an AAALAC approved facility. All experiments were conducted under an Institutional Animal Care and Use Committee-approved protocol.

Prior to tumor challenge, C57/BL6 mice were immunized by intradermal injection into the right ear pinna with 10 μL antigen or PBS, with or without the addition of α₂M^* or CpG 1826, 5’-TCCATGACGTTCCTGACGTT-3’ (Midland Certified Reagent Co., Midland, TX, USA). The treatment groups (n = 5; each receiving 1.35 μg OVA/ injection) included the following: OVA alone; OVA administered with 10 μg CpG 1826; α₂M^*-OVA; α₂M^*-OVA administered with 10 μg CpG 1826; α₂M^* co-administered with unbound OVA; and α₂M^* co-administered with unbound OVA with 10 μg CpG 1826. The quantity of α₂M^* administered with unbound OVA was equivalent to the amount of α₂M^* present in the α₂M^*-OVA preparations (6 μg).

Preparations of α₂M^* and OVA were kept in separate containers on ice until immediately prior to injection in order to minimize the possibility of in vitro α₂M^*-encapsulation. Mice were subsequently boosted at days 35 and 63, consistent with our previously published immunization protocol [7]. Control groups (n = 5) received intradermal injections of PBS, CpG 1826, or α₂M^* alone. Serum anti-OVA IgG was monitored every 2 weeks by ELISA, as previously described [7]. Mice were injected s.c. in the left flank with 10⁴ MO5 tumor cells in Matrigel™ basement membrane matrix (BD Biosciences PharMingen) at week 14. Staining of mouse PBLs with iTAgÔ MHC Tetramer H2-K^b SIINFEKL-PE (Beckman Coulter, Fullerton, CA) and CaltagÔ FITC-conjugated Rat anti-Mouse CD8a antibody (Invitrogen Corp., Carlsbad, CA) was performed 2 weeks following tumor implantation, as previously described [7], in order to quantify the proportion of CD8+ T cells specific for the H2- K^b-restricted CTL epitope of OVA, the SIINFEKL peptide (OVA_257-264). Tumor diameters were measured using digital calipers, and tumor volume was calculated using the equation V = 0.4 ab², where a and b are the longest and shortest diameters, respectively. Mice were euthanized when tumor volume reached 2 cm³.

For the detection of fluorescently-labeled OVA encapsulated by α₂M^* in vivo, three mice were injected intradermally in the right ear pinnae with 3:1 OVA:α₂M^* (30 μg:10 μg) in PBS. For comparison, one mouse was injected with only OVA (30 μg) and another was injected with only α₂M^* (10 μg). After 1.5 h, mice were euthanized, and ear pinnae were flushed with 3 × 20 μL PBS. The collected fluid samples were analyzed by native PAGE, and incorporation of antigen was measured directly by fluorescence imaging using an Odyssey infrared imaging system (LI-COR Biosciences, Lincoln, NE, USA).

Splenocytes harvested from mice that had been previously immunized with OVA and CpG 1826 (as above) were pulsed for 6 h with 2.5 μM OVA, either free or codelivered with amine-activated α₂M^* or trypsin-activated α₂M-T (7.5 μM), in serum-free media. As controls, cells were also pulsed with Con A (5 μg/mL) and α₂M^* containing no antigen. Cells were then washed, resuspended in complete media, loaded at 2.5 × 10⁵ cells per well onto a 96-well flat-bottom plate and cultured for 3 d at 37˚C in a humidified 5% CO₂ incubator. Cultured cells were treated with 1 μCi/well [methyl-³H]thymidine (PerkinElmer, Waltham, MA) 18 h prior to harvesting. Incorporation of [³H]thymidine was measured using a Tri-Carb 2100 TR liquid scintillation counter (PerkinElmer, Boston, MA).

For in vitro studies, the Student’s t-test was performed to determine P values and ascertain statistical significance between two treatments. For in vivo studies (antibody titers, tetramer staining, and tumor growth), ANOVA was performed, followed by multiple comparison procedures (Tukey) to determine differences between groups. Significance between Kaplan-Meier survival curves was determined by log-rank Mantel-Cox analysis. The level of significance used was 0.05.

Consistent with a previous report [8], maximal incorporation of antigen by α₂M^* occurs after 24 h of incubation at 37˚C (Figure 1). However, some association of antigen with α₂M^* can be observed at 30 min. This association likely represents covalent antigen incorporation by α₂M^* because it is not observed with proteolyticallyactivated α₂M-T. Although proteolytic thiol ester cleavage in α₂M is irreversible, thiol ester cleavage by primary amines can be reversed with heat, allowing the incorporation of new antigens by α₂M [11]. A small association between OVA and α₂M-T was observed following 24 h of incubation.

To determine if co-delivery of antigen with unbound α₂M^* could enhance immune responses in vivo, groups of naïve C57/BL6 mice (n = 5) were immunized intradermally with OVA and unbound α₂M^*, with or without the addition of an immunostimulatory adjuvant, CpG 1826. This study was performed concurrently with a previously reported experiment investigating immune responses to prepared α₂M^*-OVA complexes [7]. Separate unbound α₂M^* and OVA preparations were kept on ice and combined immediately prior to injection in order to minimize the possibility of α₂M^*-encapsulation occurring in vitro. Following two booster injections (days 35 and 63), mice were challenged with a subcutaneously implanted OVAexpressing B16 melanoma flank tumor.

For both conjugated α₂M^*-OVA and unconjugated α₂M^* + OVA groups, the development of anti-OVA IgG antibody was first observed 8 weeks following initial injection (result not shown). End-point titers (antibody titers at the time of tumor implantation; week 14) are shown in Figure 2(a). OVA administered either with unbound α₂M^* or the well-characterized adjuvant CpG 1826 produced similar antibody titers in immunized mice.

Although the mean antibody titers produced by conjugated α₂M^*-OVA were greater than those elicited with OVA and unbound α₂M^*, the difference between these groups was not found to be statistically significant.

Tetramer staining of PBLs was performed 2 weeks following tumor implantation in order to observe expansion of the antigen-specific CD8+ T cell population in these mice (Figure 2(b)). Mice treated with OVA and unbound α₂M^* with the addition of CpG, however, did elicit expansion of the OVA-specific CD8+ T cell population, which was significant compared to the control groups. Although the OVA with unbound α₂M^* without CpG group appeared to elicit some degree of expansion of the OVA-specific CD8+ T cell population, the tetramer staining population for this group was not significantly greater than that of the control groups (PBS, CpG, or α₂M^* alone). Conjugated α₂M^*-OVA treatment groups appeared to elicit greater expansion of antigen-specific CTLs than the OVA with unbound α₂M^* groups, the differences between these treatment groups were not found to be statistically significant. The greatest expansion of OVA-specific CD8+ T Cell population was observed in the α₂M conjugated-OVA with the addition of CPG.

To investigate the anti-tumor response of co-administered α₂M^* vaccinated mice were challenged with OVA

expressing BIG melanoma flank tumors. Anti-tumor responses elicited by the co-delivery of α₂M^* with unbound OVA were found to be similar to those observed with α₂M^*-OVA. Observable growth of OVA-expressing tumors over time is shown in Figure 2(c). Mice immunized with OVA and unbound α₂M^*, with or without CpG, demonstrated delayed tumor growth compared to each of the control groups, tumor growth for these groups was not significantly different from the α₂M^*-OVA treatment groups. Survival of mice treated with co-delivered α₂M^* and unbound OVA, either with or without CpG, was significantly prolonged (P < 0.005) compared to OVA treatment alone (Figure 2(d)). However, survival of these mice did not differ significantly from that of mice treated with conjugated α₂M^*-OVA.

We hypothesized that enhanced in vivo immune responses with α₂M^* and OVA co-delivery were the result of in vivo encapsulation of antigen. The conditions of antigen delivery, including the depot effect caused by intradermal injection and the 37˚C environment of the mouse, are similar to the conditions used to successfully incorporate antigen into α₂M^* in vitro [8]. To determine if such in vivo encapsulation of antigen could occur in this setting, we similarly intradermally injected the ear pinnae of mice with a 3:1 molar ratio of OVA:α₂M^*. For comparison, mice were injected with either OVA or α₂M^* alone. After 1.5 h, the mice were euthanized, and the ear pinnae were flushed with 3 × 20 μL PBS. The fluid that was recovered was analyzed by native PAGE (Figure 3). The detection of fluorescently labeled OVA co-migrating with α₂M^* dimers in these mice confirmed the occurrence of in vivo encapsulation of OVA into α₂M^*.

The detection of in vivo α₂M^* encapsulation suggests a mechanism for the enhanced in vivo immune responses discussed above, it was also possible that ligation of the α₂M^* receptor, low-density lipoprotein receptor-related protein 1 (LRP-1)/CD91, in the absence of antigen encapsulation, may also contribute to this enhanced immunity. To investigate this possibility, splenocytes harvested from OVA-immunized mice were treated for 6 h with OVA and either unconjugated amine-activated α₂M^* or proteolytically-activated α₂M-T. After 3 days, cell proliferation was measured by [³H]thymidine incorporation (Figure 4). Cell proliferation was increased approximately two-fold following co-delivery of α₂M^* with antigen. However, co-delivery of α₂M-T, which is receptor-recognized but incapable of incorporating new antigen, did not enhance proliferation. Therefore, we concluded that this enhanced response is secondary to α₂M^*-encapsulation and not ligation of the α₂M^* receptor, (LRP-1)/CD91.