Introduction

Hereditary hemochromatosis (HH) is a relatively common genetic disorder of iron metabolism with variable penetrance. End-organ damage in HH is typically attributed to increased intestinal iron absorption and decreased hepcidin levels resulting in iron deposition in tissues. The liver, pancreas, heart, and pituitary gland are common sites of damage. Risk factors for clinically important manifestations of HH include advanced age, male sex, high dietary iron intake, and pre-existing or concurrent tissue damage. The most common genetic mutation causing HH is the C282Y mutation in the HFE gene. HH is one of the strongest known associations with calcium pyrophosphate crystal deposition (CPPD). CPPD occurs in 20–30% of patients with HH [1, 2]. After several decades of observation and experimental attempts to clarify the process, the link between these two diseases remains unexplained. Here, we will discuss the shared clinical features of hemochromatosis arthropathy (HHA) and CPPD, describe the existing theories of CPPD pathogenesis, and review the current understanding of HHA pathogenesis. The purpose of this review is to delineate potential pathogenic pathways common to both types of arthritis.

Shared Clinical Overlap Between HHA and CPPD

HHA shares important clinical features with CPPD even without evidence of CPP crystal deposition, and both of these types of arthritis have features which distinguish them from the most common type of degenerative arthritis, osteoarthritis (OA) (Table 1). The patterns of joint involvement in HHA and CPPD are nearly identical and involve sites that are not commonly affected by primary OA, such as the 2nd and 3rd metacarpophalangeal (MCP) joints. Larger joints such as wrists and shoulders are also involved. Increased but mild inflammation in the synovium distinguishes HHA synovial tissue from that of OA [3], and low-grade inflammation is also present in CPPD, with higher blood cell counts in synovial fluids of joints with CPPD compared to OA even in the absence of acute CPP crystal arthritis [4, 5]. Most CPPD occurs sporadically in patients over the age of 60 but is also associated with a handful of seemingly unrelated conditions including hypophosphatasia, hyperparathyroidism, and hypomagnesemia. Reversal of these secondary causes in CPPD and iron removal in HHA does not ameliorate the disease. In fact, reduction of iron has been shown to have no beneficial effects on symptoms and may actually worsen joint pain [6, 7]. Treatment of both these forms of arthritis currently focuses on controlling symptoms.

Table 1 Shared clinical features between CPPD, HHA, and OA
Full size table

Interestingly, both diseases share a similar bone phenotype. Patients demonstrate osteopenia and osteoporosis more commonly than age and sex matched controls. Guggenbuhl in 2005 reported that of 38 men with HH and confirmed iron overload by liver biopsy, 79% had osteopenia and 34% had osteoporosis as measured by DXA [8]. Similar results were shown by Valenti et al. in 2009, although Nguyen [2] reported lower frequency of osteoporosis [2, 9]. In CPPD, increased rates of osteopenia and osteoporosis were noted in two large cross-sectional studies of CPPD patients [10, 11]

Current Theories of CPPD Pathogenesis

While the pathogenesis of CPPD is not fully understood, current data support the existence of two overarching and intertwined theories. The first and best experimentally supported theory is that CPPD begins in articular cartilage due to the overproduction of inorganic pyrophosphate (PPi) by chondrocytes. Excess PPi complexes with extracellular calcium from synovial fluid form CPP crystals in the pericellular matrix. CPP crystals then induce synovial inflammation, damage the surrounding cartilage, and alter joint biomechanics to produce arthritis. The best clinical examples of such a mechanism are gain-of-function mutations in the ANKH protein seen in some cases of familial CPPD. ANKH is present in normal chondrocytes, and ANKH gain of function increases levels of the PPi precursor, ATP in cartilage [12].

The second theory is based on the recent discovery that a loss of function mutation in the gene that codes for osteoprotegerin (OPG) accounts for multiple kindred with familial CPPD [13]. However, a neither mutant or wild-type OPG nor its cognate ligands affected CPP crystal formation or cartilage PPi production [14]. This shifted attention to the role of mutant OPG in bone homeostasis. This OPG mutation permits excess RANKL-mediated osteoclast formation in bone [13]. Excess osteoclasts in subchondral bone may release factors into cartilage, such as TGFβ, which mediate joint destruction [15]. TGFβ is also a potent inducer of increased PPi production in cartilage [16], and this results in CPP crystal formation and further exacerbates joint damage. Bone abnormalities in CPPD are further supported by the observations that (1) there is increased risk of osteoporosis in sporadic CPPD; (2) there are abnormal bone phenotypes associated with other risk factors for CPPD such as hypophosphatasia and hyperparathyroidism; and (3) there is abundant evidence for a critical role of bone in other forms of degenerative arthritis [17,18,19]. Aging and injury, which are common risk factors for sporadic CPPD, likely affect both cartilage and bone. For example, ANKH levels increase with age in cartilage, as does osteoclast activity and aggressive bone resorption [20, 21].

Work Directly Exploring the Link Between HHA and CPPD

To date, relatively few studies address the causal links between HH and CPPD. Most of this early work sought to implicate iron in processes important for CPPD. Two groups postulated a direct role for iron in CPP mineral formation. The presence of ferric or ferrous ions in solution or in gel models of CPP crystal formation demonstrated a complex effect. Iron both interfered with [22] and enhanced crystal formation [23]. Early papers speculated that iron might decrease CPP crystal dissolution, but there is little evidence for this experimentally [1, 24], nor is it easily supported clinically as removal of iron does not ameliorate CPPD in HHA. Ryan et al. found no effects of either ferrous or ferric salts on levels of chondrocyte extracellular pyrophosphate (PPi) although significant cell toxicity was observed [25].

Current Theories of HHA Pathogenesis that Share Features with Those of CPPD Pathogenesis

For many years, all types of degenerative arthritis were attributed to primary articular cartilage damage. It follows logically that HHA was felt to occur due to iron detrimental actions on chondrocyte/cartilage homeostasis, and that aging produces parallel processes in cartilage promoting CPPD. Indeed, much of the work in HHA investigated the direct cellular toxicity of iron on cartilage. In studies of HHA patients, there is little consistent evidence that iron deposits in cells or extracellular matrix of articular cartilage [26, 27], and CPP crystals are not physically associated with iron deposits [28]. Murine models of HH based on HFE deficiency show iron deposition in subchondral bone and synovium, but iron is often absent from articular cartilage [29, 30]. Fe has well known toxic effects on cells and increases ROS through the Fenton reaction. Not surprisingly, in vitro exposure of chondrocytes, osteoblasts, and osteocytes to high level of iron decreases cell viability [31,32,33,34]. In CPPD, chondrocyte senescence associated with advanced age may play a similar role to iron [35].

Both CPPD and HHA share moderately increased synovial inflammation compared to OA. Synovial inflammation can certainly contribute to joint damage, impede repair after injury, and contribute to degenerative arthritis [36]. In HHA, human and murine studies confirm the presence of iron deposits in synovial tissue with iron overload. Direct effects of iron on synovial tissues have not been well studied in HHA. In one study of synovial tissue from a hemophilia patient with iron deposition, ex vivo cultures showed increased production of Il-6, TNFα, and other inflammatory cytokines compared to normal synovial tissue [37]. CPPD synovium also shows more inflammation in certain settings [38]. Certainly ATP, a critical precursor of extracellular PPi which is increased in joints affected by CPPD [39] can induce IL-1β production by macrophages through the P2X7 receptor activation [40] and also induces inflammation in other cell types through its action as a DAMP [41].

Joint damage and/or deficient repair processes may contribute to both HHA and CPPD. A longstanding theory of many types of degenerative arthritis is that arthritis stems from a lifetime of minor and major injuries followed by ineffective repair. Injury is a major risk factor for CPPD as illustrated by the dramatic increase in CPPD in damaged menisci compared to undamaged menisci [42]. In a murine joint injury model, Camacho et al. showed that HFE-deficient mice developed more severe injury-induced OA than wild type mice [30]. A critical role for defective repair processes is further supported in HHA by studies of HFE deficient mice which show no spontaneous arthritis up to 12 months of age [29]. Iron may also impede joint repair [43]. Ample evidence exists for deficiencies in joint repair processes with advanced age that may explain the association of age with CPPD [44].

There is increasing attention to the role of bone in joint injury and repair which may be important in both CPPD and HHA. The mechanisms through which HH produces bone loss remain unclear. There is literature supporting the notion that HH bone disease results from a direct effect of iron on osteoclastogenesis [45,46,47,48] and subsequent excess bone resorption [46]. A slightly smaller number of studies fail to show enhanced osteoclast differentiation with iron exposure [49, 50] or genetic manipulation [51, 52]. Two seminal studies examining bone in HFE deficient mice showed progressive osteopenia with age [53, 54]. One study attributed this to enhanced osteoclastogenesis while the second attributed it to defective osteoblast function and thus reduced mineral formation. Another recent study by Yang (34) suggested that iron-induced osteocyte apoptosis provided a rich source of RANKL for osteoclast formation and subsequent bone resorption [34].

Increased osteoclast activity can worsen joint injury and impede repair. Multiple studies demonstrate that one of the earliest findings in degenerative arthritis, for example, is an increase in the number of subchondral osteoclasts [18, 44]. It is logical that subchondral bone instability resulting from aggressive destruction of the cartilage/bone interface may contribute to cartilage microcracks and accelerate joint damage. Osteoblasts and osteocytes may also contribute to arthritis either through their ability to regulate osteoclast-mediated bone resorption or through other actions [17]. A loss of function mutation in osteoprotegerin, resulting in increased OC activity, has recently been identified as a cause of familial CPPD as described above [13]. The delicate balance between bone formation and bone resorption in both HHA and CPPD is clearly a shared mechanism that warrants further study.

A key regulator of PPi levels in bone and cartilage is tissue non-specific alkaline phosphatase (TNAP). TNAP hydrolyzes PPi and hypophosphatasia, which results from a congenital deficiency of TNAP, increases the risk of CPPD by raising PPi levels in bone and cartilage. Bone levels of TNAP are dysregulated in both HHA and CPPD. Osteoblasts account for a large percentage of bone TNAP which is essential to normal bone matrix mineralization. Interestingly, multiple studies of osteoblasts show decreased TNAP activity resulting from treatment with iron [55,56,57,58,59]. Iron may also affect osteoblast-mediated bone mineral formation. Human bone marrow stromal cells treated with ferrous sulfate demonstrated attenuation of bone mineralization and osteoblastogenesis in one study [60], while other studies found no such affect [47]. The effects of iron on osteocyte alkaline phosphatase remain unclear.

Other Theories

HFE may have additional functions independent of its iron-regulating properties that produce or are shared with CPPD. HFE is a membrane protein with similarities to major histocompatibility complex (MHC)-class I proteins. Studies of patients carrying HFE mutations but without clinical evidence of iron overload show a propensity for hand OA patterns that mimic HHA and CPPD, such as predominant involvement of the MCPs. This observation has raised questions about whether iron mediates joint damage or whether some other function of HFE may be involved. The best of these studies positively correlates MCP-predominant hand OA with the C282Y HFE mutation [61], but there are limitations in the study design. In another study, patients with idiopathic CPPD were screened for C282Y [62, 63], but no differences in gene frequencies were found in the CPPD group compared to random controls.

HFE binds β2-microglobulin [64], which is present in musculoskeletal tissues and has important effects on bone [65] and anti-anabolic effects on chondrocytes [66]. It is not known whether loss of function mutations in HFE alter β2 microglobulin levels in musculoskeletal tissues. Interestingly, β2-microglobulin is the protein involved in dialysis-related amyloid deposits, and amyloid has been noted around CPP crystals in cartilage [67, 68]. In addition, factors such as FGF23, which can be regulated along with iron and regulate pathologic calcification in many tissues [69], may also play a role in both conditions.

Summary

In summary, current evidence supports the existence of certain shared pathogenic pathways in HHA and CPPD which may explain their clinical overlap. There may be a role for synovial inflammation in both diseases. There is some experimental evidence for abnormal joint damage/repair processes in HHA and good clinical support for similar processes in CPPD. While there is a paucity of evidence for a direct effect of iron exposure on cartilage in HHA, altered bone metabolism particularly related to increased osteoclastogenesis is supported clinically and experimentally in both diseases. Further work in this area is clearly needed, but elucidation of these pathogenic mechanisms may eventually identify potential therapeutic targets for both types of arthritis.