Folate-targeted immunotherapy effectively treats established adjuvant and collagen-induced arthritis

Rheumatoid arthritis (RA) is an autoimmune disease characterized by inflammation of the joints and destruction of cartilage and bone. Activated macrophages have been identified as a key mediator of the disease, as numbers and level of macrophage activation correlate with the extent of joint inflammation and bone degradation [1‐4]. In addition, neutralization of inflammatory mediators secreted by activated macrophages suppresses symptoms of the disease [1‐5].

Many therapies for RA are specifically designed to eliminate pro-inflammatory byproducts of activated macrophages [6]. Remicade and etanercept neutralize tumor necrosis factor alpha (TNF-α) [7‐9], whereas anakinra blocks the activity of interleukin (IL)-1 [8‐10]. Celecoxib and Vioxx are inhibitors of cyclooxygenase-2 [11], and anti-oxidants inactivate reactive oxygen species. Yet some patients require other remedies, either because of the ineffectiveness of the above therapies or due to their associated toxicities [12, 13].

Efforts have been made to eliminate the entire population of macrophages [14‐19], but elimination of all mononuclear phagocytes can be harmful, because they are important in fighting infectious diseases and promoting tissue repair [20, 21]. Thus, an attractive alternative to elimination of all macrophages might be to remove only that subpopulation that promotes RA (that is, the activated macrophage).

Activated macrophages from patients and rodents with arthritis over-express a cell surface receptor for folic acid that also binds folate-linked compounds with high affinity (K_D approximately 10^-10 M) [22‐24]. Recent studies have shown that folate-linked radiopharmaceuticals concentrate in arthritic joints, enabling visualization of such tissues by gamma scintigraphy [23‐25]. We hypothesized therefore that selective removal of activated macrophages with folate-linked drugs could be exploited to treat RA with little toxicity to other tissues. Here, we report a test of this strategy in two distinct rodent animal models of the disease.

Heat-killed Mycoplasma butyricum (BD Biosciences, Sparks, MD, USA); light mineral oil, methotrexate (MTX), clodronate, bovine serum albumin, keyhole limpet hemocyanine (KLH), anti-rat immunoglobulin G (IgG)-horseradish peroxidase (HRP), IgG-phycoerythrin conjugates, and alum (Sigma, St. Louis, MO, USA); biotin-conjugated goat anti-rat IgG1, IgG2a, and IgG (H+L), and streptavidin-conjugated HRP (Caltag Laboratories, Burlingame, CA, USA); aminofluorescein (single isomer) and fluorescein isothiocyanate (FITC) (Molecular Probes, Eugene, OR, USA); Microcon-30 membranes (Millipore Corp., Bedford, MA, USA); TiterMax Gold^® adjuvant (CytRx Corporation, Los Angeles, CA, USA); Celecoxib (Pfizer, Ann Arbor, MI, USA); entanercept and anakinra (Amgen, Inc., Thousand Oaks, CA, USA); EC20 (a folate-linked chelator of ^99mTc) and folate-FITC (Endocyte, Inc., West Lafayette, IN, USA) were obtained from commercial sources. The studies were approved by the Purdue University Animal Care and Use Committee.

We have explored a novel approach for the treatment of experimental arthritis (the redirection of the immune system to reduce endogenous activated macrophage, which is known to be involved in maintenance of RA) [1‐5, 32‐34]. Thus, ever since it was observed that activated, but not resting, macrophages express a folate receptor [22], folate-linked drugs have been found to specifically target arthritic joints and sites of inflammation [23‐25]. In this study, we have exploited folate's targeting ability to selectively decorate activated macrophages with immunogenic haptens that can mediate their removal in immunized hosts. Prior vaccination against the hapten was found to be necessary, because non-immunized animals did not benefit from treatment with folate-FITC. Folate targeting of the hapten was shown to be critical, because identical treatment with non-targeted aminofluorescein was totally ineffective. Fluorescein derivatization was also found to be essential, because administration of folic acid without an attached fluorescein was equally impotent. In contrast, vaccination with KLH-FITC followed by treatment with folate-FITC led to significant reduction in both local and systemic inflammation and to measurable diminution in bone and cartilage degradation in two advanced-stage models of RA.

We hypothesize that the efficacy of the FTI is related to its ability to promote systemic depletion of FR⁺ macrophages. As noted above, activated macrophages constitute a likely orchestrator of inflammation and joint destruction in experimental arthritis [1‐5, 32‐34]. Not only do activated macrophages release reactive oxygen species, collagenases, cathepsins, and other mediators of joint destruction, but they also recruit other immune cells by releasing pro-inflammatory cytokines (TNF-α, IL-1, and IL-6), leukotrienes, and prostaglandins [25]. Several investigators [22, 23, 25] have demonstrated that activated macrophages express significant numbers of cell-surface FR. Because folate-linked drugs bind avidly to these receptors and localize to inflamed joints of animals and humans with arthritis [23‐25], it seems reasonable to posit that their uptake at sites of inflammation might stem from their high affinity for FR. Others have, in fact, established that synovial macrophages from patients with RA aggressively bind folate-FITC [22], and we have similarly demonstrated that folate-FITC uptake in inflamed joints of rodents with experimental arthritis occurs primarily in cells expressing both FR and macrophage markers [23]. Because treatment of both CIA mice and AIA rats leads directly to a significant reduction in folate-binding macrophages, we speculate that painting of activated macrophage cell surfaces with folate-FITC directly causes their elimination in anti-FITC-immunized rodents by antibody-dependent pathways. Studies are currently under way to test this mechanism.

Comparative experiments revealed that FTI was as effective as CL, MTX, etanercept, anakinra, and celecoxib at alleviating the symptoms of experimental arthritis. Furthermore, FTI appeared to be better than the above therapies at suppressing systemic inflammation, as measured by both accumulation of EC20 in liver and spleen and by reduction of inflammation-mediated spleen enlargement. This capability to reduce systemic inflammation may be important, because patients with RA have a reduced life expectancy that is caused by systemic, rather than articular, inflammation [35].

An attractive feature of the FTI approach lies in its apparent lack of toxicity. Because FRs are also expressed on the apical surfaces of a few normal epithelia (for example, the choroid plexus of the brain, bronchioalveolar cells of the lungs, and proximal tubules of the kidney), it might have been anticipated that immune responses would have been elicited against these tissues also. However, fluorescence micrographs show that sites of FR expression on normal epithelia are inaccessible to circulating antibodies [31], because the receptor invariably resides on the apical surface of the epithelium where it faces a lumen that is usually inaccessible to parenterally administered drugs. Thus, pathological analyses have not revealed any cytotoxicity or immune damage to normal tissues.

One concern that must accompany systemic depletion of activated macrophages relates to the possible decline in defense against pathogens. Although no infections/diseases were observed in rats maintained on the therapy for 15 continuous weeks, nor in RA dogs treated for more than 2.5 years (Paulos CP, Varghese B, Vlashi E, Low PS, Widmer WR, and Breur GJ), unpublished observations), sufficient time has not yet elapsed to conclude that long-term defects will not emerge during continuous treatment. Although resting macrophages and other immune cells might compensate for elimination of most activated macrophages, it will be important to examine chronically treated animals for deficiencies in tasks normally performed by activated macrophages. Where temporary deficiencies do arise, it might be sufficient to remove the patient from the therapy until the deficiency subsides.

A second potential concern might arise from the possible formation of multimeric immune complexes in circulation after injection of the folate-hapten conjugate. However, because folate-FITC contains only one antibody-binding site, it cannot mediate formation of multimeric immune complexes, even though monomeric IgG-folate-FITC complexes are readily observed in circulation. Thus, deposition of macroscopic immune complexes in non-targeted tissues has not constituted a source of observed toxicity.

In summary, we have demonstrated that removal of folate-binding macrophages suppresses established arthritis in two animal models of the disease. Because little measurable toxicity accompanied the therapy, these data suggest that removal or inactivation of folate receptor-expressing macrophages might constitute an alternate approach for treatment of the pathology. Thus, in addition to the existing pantheon of drugs aimed at neutralizing secretion products of activated macrophages (that is, Remicade, Enbrel, Humira, Kineret, Vioxx, Celebrex, and so on), examination of folate-targeted drugs that can block the inflammatory activity of the same cell type would seem to warrant further scrutiny [25].