**4.2.3 Accumulation of cells**

Under the influence of chemotactic and other immunoregulatory signals, many different inflammatory and reparative cells migrate through local tissues regionally and systemically

In addition to pro-inflammatory cytokines, chemo-attractive cytokines are also produced by different cell types involved in the inflammatory response at the bone-implant interface including: polymorphonuclear leukocytes, monocytes, macrophages, osteoblasts, fibroblasts and synovial cells. Monocyte chemoattractant protein-1 (MCP-1) is one of the most abundantly released chemokines and is an immediate early stress-responsive factor (Tuan et al. 2008). Macrophage inhibitory protein-1 (MIP-1 or CCL2) and interleukin 8 (IL-8) are also

*MCP-1* is a C-C chemokine ( chemokine subfamily) and its gene location in humans is 17q11.2. MCP-1 is primarily secreted by monocytes, macrophages and dendritic cells. With regards to bone, MCP-1 is also expressed by osteoblasts and osteoclasts. MCP-1 can bind to several receptors (CCR2, CCR4, CCR5) but binds preferentially to the CCR2. CCR2 has two isoforms: CCR2A and CCR2B; the expression of the different isoforms is restricted to specific cell types. Mononuclear cells and vascular smooth muscle cells express CCR2A whereas CCR2B is mainly expressed by monocytes and activated NK cells (Deshmane et al. 2009). MCP-1 regulates the migration and infiltration of monocytes, dendritic cells, memory T-lymphocytes and NK cells. Production of MCP-1 is under the direct control of the

*MIP-1* belongs to the same subfamily of chemokines as MCP-1 (the C-C chemokine family), and is secreted primarily by activated macrophages and T-lymphocytes. Other cells may also secrete MIP-1. MIP-1 has two main isoforms: MIP-1 (CCL3) and MIP-1 (CCL4). In mice, MIP-1 and MIP-1 are encoded by one gene each, however in humans, MIP-1 is encoded by at least three genes and MIP-1 by two, all located in chromosome 17 (Menten et al. 2002). Inducers of MIP-1 release are IL-1, IL-7, LPS and specific adhesion molecules (e.g. intercellular adhesion molecule-1, ICAM-1). High levels of MIP-1 have been observed when macrophages from retrieved periprosthetic tissues have been challenged with titanium alloy and polymethylmethacrylate particles (Nakashima et al. 1999). The release of MIP-1 increases the systemic recruitment of macrophages and has a paracrine effect enhancing IL-1 and IL-6 release (Cook 1996; Menten et al. 2002). Although MCP-1 has a specific receptor (CCR2), MIP-1

*Interleukin-8* (IL-8) is a member of the CXC chemokine family, and also has the designation CXCL8. It is an early stress responsive chemokine and the main cellular sources are: macrophages, endothelial cells, epithelial cells, and cells of the mesenchymal lineage such as mesenchymal stem cells and osteoblasts. IL-8 is a potent chemokine that induces the chemotaxis of neutrophils and monocytes/macrophages, and increases the mobility of osteoclasts (Baggiolini and Clark-Lewis 1992). Il-8 is also an angiogenic factor. IL-8 does not have a specific receptor but acts through the coupled G protein receptors CXCR1 and CXCR2, which are receptors for other chemokines (e.g. GCP-2, NAP-2, Gro- and others), (Proudfoot 2002). When osteoblasts are challenged with phagocytosable titanium particles, the IL-8 peak in gene expression occurs within 1-2 hours after particle exposure (Fritz et al. 2002). The biological mechanism of IL-8 expression involves first a protein tyrosine phosphorylation after contact with Ti particles and then activation of the pro-inflammatory NF-B through two subunits, p50 (NF-B1) and p65 (RelA) to bind the IL-8 gene promoter.

Under the influence of chemotactic and other immunoregulatory signals, many different inflammatory and reparative cells migrate through local tissues regionally and systemically

acts through several receptors (CCR1, CCR5), (Proudfoot 2002).

**4.2.2 Accumulation of chemokines** 

involved in these events.

transcription factor NFB.

**4.2.3 Accumulation of cells** 

through the blood stream to the areas of particle generation. Recent murine studies involving continuous intra-osseous infusion of ultra high molecular weight polyethylene (UHMWPE) particles have demonstrated an increase in the systemic trafficking of reporter macrophages to the particle infusion site using non-invasive imaging (bioluminescence) and immunofluorescence microscopy (Ren et al. 2010). Furthermore, a decrease in bone mineral density has been observed within the UHMWPE particle-infused femora (Ren et al. 2011). This suggests that wear particles stimulate a systemic response, mediated by chemokines released locally. Recent unpublished studies in our laboratory have demonstrated that inhibition of the MCP-1-CCR2 chemokine-receptor axis can mitigate this response, including reversing in part, particle-induced bone loss.

Fig. 3. Biologic reaction induced by orthopaedic wear particles on host cells; note the complex interaction of different cell types and factors to the presence of wear particles. Chemokines play a central role in cell recruitment. This diagram does not include the type 4 immune reaction to metallic wear particles.

*IL = interleukin, TNF = tumor necrosis factor , GM-CSF = granulocyte macrophage colonystimulating factor, MCP-1 = monocyte chemoattractant protein-1, MIP-1 = macrophage inhibitory protein-1, ROS = reactive oxygen species, NO = nitric oxide, VEGF = vascular endothelial growth factor, NF-B = nuclear factor B, IFN = interferon, RANKL = receptor activator of nuclear factor B ligand, TLR = toll-like receptor, FGF = fibroblast growth factor, OPG = osteoprotegerin, TGF = transforming growth factor , PGE2 = prostaglandin E2* 

Aseptic Loosening of Total Hip Arthroplasty as a Result of Local Failure of Tissue Homeostasis 333

the activation and proper polarization of adaptive immunity, development of chronic

Currently already more than 200 different cytokines are known. Cytokines can be classified into several broad categories by their functional properties. Such classification is at best a simplification as the biological functions of specific cytokines vary considerably depending on the context. For example, cell type and cells activation state as well as the presence of other activating or inhibiting signals derived from e.g. cell-to-cell and cell-to-ECM contacts and other cytokines can alter the final outcome of the cytokine signalling. This phenomenon is known as *cytokine pleiotropy*: one cytokine typically has multiple functions. Considerable *redundancy* exists in cytokine signaling, different cytokines can activate very similar intracellular signalling pathways and transcription factors and thus have very similar effects on the target cell. Cytokines can have *synergy* so that they enhance each other's signalling and the net sum of the effect is far greater than their additive effects. As cytokines typically regulate the secretion of other cytokines and the expression of each other's receptors, a

At the beginning of the inflammation reaction resident tissue macrophages are activated by exo- or endogenous danger signals recognized by different pattern recognition receptors (e.g. TLRs) to produce *pro-inflammatory cytokines* e.g. tumor necrosis factor alpha (TNFα), Il-1β and IL-6. These cytokines activate and enhance the innate immune response, induce the secretion of anti-microbial molecules and other cytokines that refine the immune response, evoke systemic inflammatory response discussed above and importantly induce the expression of adhesion molecules in local microvascular beds in the process known as endothelial activation. Activated endothelium of post-capillary venules expresses selectin and integrin ligands, namely adhesion molecules (e.g. glycam-1, VCAM-1, ICAM-1), to which circulating and marginalized leukocytes effectively stick (capture and rolling), adhere and finally transmigrate from vascular compartment into the inflamed tissue. In unison with the pro-inflammatory cytokine production, macrophages, activated endothelial cells and other cells start to produce *chemokines* that further guide and activate transmigrated leukocytes to the site of inflammation along their increasing concentration gradient. The chemokines are discussed in more detail in *part 4.2* of this chapter. Also *interferons* are secreted by the cells of the innate immunity at the early stages of innate immune reaction. They are traditionally classified in two groups having very distinct functions. Type I interferons ("anti-viral interferons"), including IFNα (13 subtypes), IFNβ and IFNω (3 subtypes), are produced by macrophages after e.g. TLR recognition of viral structures (double stranded RNA or DNA). IFNγ, the only member of type II interferons ("immune interferon"), is secreted to some extend by macrophages and in greater magnitude by activated Th1 and NK-cells and is an important activating signal for macrophages which enhances their ability to secrete pro-inflammatory cytokines and kill phagocytosed bacteria. The initial strong pro-inflammatory response must be well regulated and typically proinflammatory cytokines are rapidly degraded, their actions are inhibited by soluble or solubilised receptor proteins (e.g. IL-1RA) and further counterbalanced by production of *anti-inflammatory cytokines* e.g. IL-4, IL-10, transforming growth factor-β (TGF-β), which efficiently inhibit the production of pro-inflammatory cytokines and function of innate immunity. After the initial inflammatory reaction is resolved the tissue regenerative processes are rapidly engaged. *Growth factors* represent a diverse group of cytokines that

inflammation as well as tissue regeneration and healing.

complex and integrated network of cytokine functions is formed.

**4.3.1 Functional classification of cytokines** 

After systemic and local cell trafficking to the inflammatory site, mononuclear macrophages may undergo a process called *macrophage fusion*. This results in polykaryons known as multinucleated "foreign body" giant cells which can be small (and contain 3-6 nuclei) or large (7-50 nuclei). Macrophage fusion is complex and requires the presence of specific cytokines and other factors (*Part 4.4*). The presence of foreign body giant cells has been widely demonstrated within the bone-implant interface (Goodman et al. 1989). Multinucleated giant cells have the capability to strongly express transforming growth factor- (TGF-) as shown by Al-Saffar et al in their tissue retrieval study (Al-Saffar and Revell 2000). TGF- is a key factor involved in angiogenesis, neovascularization and has been shown to impact both osteoclastic bone resorption and osteoclast-like cell growth and differentiation (Takahashi et al. 1986).

In vitro studies using murine bone marrow macrophages have shown that polymethylmethacrylate (PMMA) particles stimulate both RANKL expression, increase the number of osteoclasts and induce the formation of multinucleated giant cells (Clohisy et al. 2003). The authors also observed that PMMA particles activate the NF-B and c-jun/AP-1 transcription factors, both of which mediate the osteoclastogenic effect of RANKL. Thus, *after contact with orthopaedic wear debris, resident macrophages become activated and release proinflammatory cytokines, chemokines and other factors*. MCP-1, MIP-1 and IL-8 release evokes the systemic recruitment of macrophages to the site of inflammation. Macrophage accumulation maintains the inflammatory process with release of pro-inflammatory cytokines that have paracrine and autocrine functions. Macrophage fusion induces the local release of TGF- by multinucleated foreign body giant cells that facilitates the growth and differentiation of osteoclasts. Further cytokine release (TNF-, IL-1, RANKL and others) by resident macrophages, fibroblasts, and cells of the osteoprogenitor cell line maintains a heightened inflammatory state and promotes the growth and differentiation of osteoclasts. These interlinked pathways are summarized in figure 3.

### **4.3 Cytokines and orchestrating of local adverse reaction to prosthetic particles**  *(Pajarinen, Takakubo, Mackiewicz, Takagi, Konttinen)*

Cytokines are small (15-25 kd) protein or glycoprotein messenger molecules that mediate intercellular (or even intracellular) communication and modulate immune reactions. They are produced, to some extent, by virtually all types of cells but by cells of the epithelial and hematopoietic origin as well as the cells of the immune system in particular. Typically cytokines are, after cell activation, secreted as soluble mediators into the extracellular space where they exert their function by binding to their specific high-affinity receptors located on the target cell membrane. Cytokine binding to its receptors induces receptor conformational changes and receptor dimerization or polymerization, binding of co-receptors in some instances and binding of adaptor proteins to the cytoplasmic tails, which leads to the activation of intracellular signalling cascades, activation or inactivation of specific transcriptional factors and finally up- or down regulation of gene expression. Some cytokines are not secreted but act in an intracrine (inside the cell) or juxtacrine (bound to cell membrane and requiring a direct cell-to-cell contact) mode of action while others are confined to the extracellular matrix (ECM) and released only when ECM proteins are degraded or damaged.

Cytokines control a very wide variety of different cell functions including cell growth, differentiation, migration, activation and survival. They are essentially involved into the development of blood cells (hematopoiesis) and not only in to the development and coordination of innate immune reaction and initial acute inflammation reaction but also to the activation and proper polarization of adaptive immunity, development of chronic inflammation as well as tissue regeneration and healing.

### **4.3.1 Functional classification of cytokines**

332 Recent Advances in Arthroplasty

After systemic and local cell trafficking to the inflammatory site, mononuclear macrophages may undergo a process called *macrophage fusion*. This results in polykaryons known as multinucleated "foreign body" giant cells which can be small (and contain 3-6 nuclei) or large (7-50 nuclei). Macrophage fusion is complex and requires the presence of specific cytokines and other factors (*Part 4.4*). The presence of foreign body giant cells has been widely demonstrated within the bone-implant interface (Goodman et al. 1989). Multinucleated giant cells have the capability to strongly express transforming growth factor- (TGF-) as shown by Al-Saffar et al in their tissue retrieval study (Al-Saffar and Revell 2000). TGF- is a key factor involved in angiogenesis, neovascularization and has been shown to impact both osteoclastic bone resorption and osteoclast-like cell growth and

In vitro studies using murine bone marrow macrophages have shown that polymethylmethacrylate (PMMA) particles stimulate both RANKL expression, increase the number of osteoclasts and induce the formation of multinucleated giant cells (Clohisy et al. 2003). The authors also observed that PMMA particles activate the NF-B and c-jun/AP-1 transcription factors, both of which mediate the osteoclastogenic effect of RANKL. Thus, *after contact with orthopaedic wear debris, resident macrophages become activated and release proinflammatory cytokines, chemokines and other factors*. MCP-1, MIP-1 and IL-8 release evokes the systemic recruitment of macrophages to the site of inflammation. Macrophage accumulation maintains the inflammatory process with release of pro-inflammatory cytokines that have paracrine and autocrine functions. Macrophage fusion induces the local release of TGF- by multinucleated foreign body giant cells that facilitates the growth and differentiation of osteoclasts. Further cytokine release (TNF-, IL-1, RANKL and others) by resident macrophages, fibroblasts, and cells of the osteoprogenitor cell line maintains a heightened inflammatory state and promotes the growth and differentiation of osteoclasts. These

**4.3 Cytokines and orchestrating of local adverse reaction to prosthetic particles** 

Cytokines are small (15-25 kd) protein or glycoprotein messenger molecules that mediate intercellular (or even intracellular) communication and modulate immune reactions. They are produced, to some extent, by virtually all types of cells but by cells of the epithelial and hematopoietic origin as well as the cells of the immune system in particular. Typically cytokines are, after cell activation, secreted as soluble mediators into the extracellular space where they exert their function by binding to their specific high-affinity receptors located on the target cell membrane. Cytokine binding to its receptors induces receptor conformational changes and receptor dimerization or polymerization, binding of co-receptors in some instances and binding of adaptor proteins to the cytoplasmic tails, which leads to the activation of intracellular signalling cascades, activation or inactivation of specific transcriptional factors and finally up- or down regulation of gene expression. Some cytokines are not secreted but act in an intracrine (inside the cell) or juxtacrine (bound to cell membrane and requiring a direct cell-to-cell contact) mode of action while others are confined to the extracellular matrix (ECM) and released only when ECM proteins are

Cytokines control a very wide variety of different cell functions including cell growth, differentiation, migration, activation and survival. They are essentially involved into the development of blood cells (hematopoiesis) and not only in to the development and coordination of innate immune reaction and initial acute inflammation reaction but also to

differentiation (Takahashi et al. 1986).

interlinked pathways are summarized in figure 3.

degraded or damaged.

*(Pajarinen, Takakubo, Mackiewicz, Takagi, Konttinen)* 

Currently already more than 200 different cytokines are known. Cytokines can be classified into several broad categories by their functional properties. Such classification is at best a simplification as the biological functions of specific cytokines vary considerably depending on the context. For example, cell type and cells activation state as well as the presence of other activating or inhibiting signals derived from e.g. cell-to-cell and cell-to-ECM contacts and other cytokines can alter the final outcome of the cytokine signalling. This phenomenon is known as *cytokine pleiotropy*: one cytokine typically has multiple functions. Considerable *redundancy* exists in cytokine signaling, different cytokines can activate very similar intracellular signalling pathways and transcription factors and thus have very similar effects on the target cell. Cytokines can have *synergy* so that they enhance each other's signalling and the net sum of the effect is far greater than their additive effects. As cytokines typically regulate the secretion of other cytokines and the expression of each other's receptors, a complex and integrated network of cytokine functions is formed.

At the beginning of the inflammation reaction resident tissue macrophages are activated by exo- or endogenous danger signals recognized by different pattern recognition receptors (e.g. TLRs) to produce *pro-inflammatory cytokines* e.g. tumor necrosis factor alpha (TNFα), Il-1β and IL-6. These cytokines activate and enhance the innate immune response, induce the secretion of anti-microbial molecules and other cytokines that refine the immune response, evoke systemic inflammatory response discussed above and importantly induce the expression of adhesion molecules in local microvascular beds in the process known as endothelial activation. Activated endothelium of post-capillary venules expresses selectin and integrin ligands, namely adhesion molecules (e.g. glycam-1, VCAM-1, ICAM-1), to which circulating and marginalized leukocytes effectively stick (capture and rolling), adhere and finally transmigrate from vascular compartment into the inflamed tissue. In unison with the pro-inflammatory cytokine production, macrophages, activated endothelial cells and other cells start to produce *chemokines* that further guide and activate transmigrated leukocytes to the site of inflammation along their increasing concentration gradient. The chemokines are discussed in more detail in *part 4.2* of this chapter. Also *interferons* are secreted by the cells of the innate immunity at the early stages of innate immune reaction. They are traditionally classified in two groups having very distinct functions. Type I interferons ("anti-viral interferons"), including IFNα (13 subtypes), IFNβ and IFNω (3 subtypes), are produced by macrophages after e.g. TLR recognition of viral structures (double stranded RNA or DNA). IFNγ, the only member of type II interferons ("immune interferon"), is secreted to some extend by macrophages and in greater magnitude by activated Th1 and NK-cells and is an important activating signal for macrophages which enhances their ability to secrete pro-inflammatory cytokines and kill phagocytosed bacteria. The initial strong pro-inflammatory response must be well regulated and typically proinflammatory cytokines are rapidly degraded, their actions are inhibited by soluble or solubilised receptor proteins (e.g. IL-1RA) and further counterbalanced by production of *anti-inflammatory cytokines* e.g. IL-4, IL-10, transforming growth factor-β (TGF-β), which efficiently inhibit the production of pro-inflammatory cytokines and function of innate immunity. After the initial inflammatory reaction is resolved the tissue regenerative

processes are rapidly engaged. *Growth factors* represent a diverse group of cytokines that

Aseptic Loosening of Total Hip Arthroplasty as a Result of Local Failure of Tissue Homeostasis 335

on the quantity and quality of the cytokines produced. This wear particle induced macrophage activation is possibly mediated by various types of cell surface pattern recognition receptors e.g. Toll-like receptors (Takagi et al. 2007; Tamaki et al. 2009) and likely involves the activation and action of transcription factor NF-kB (Lahdeoja et al. 2010;

The mere wear particle stimulation of pure monocyte/macrophage cultures is not sufficient to drive osteoclastogeneis or foreign body giant cell formation, most likely due to lack of RANKL signalling. Further *in vitro* studies using other relevant cell types of the periprosthetic tissue have demonstrated that, to some extent, wear particles as such and especially pro-inflammatory cytokines effectively *up-regulate RANKL production* from periprosthetic tissue fibroblasts and osteoblasts as well as suppress osteoblast formation and

Consistent with *in vitro* findings, numerous studies of the explanted periprosthetic tissues and pseudosynovial fluid have demonstrated increased production of vast array proinflammatory cytokines, chemokines and growth factors including e.g. TNFα, IL-1β, IL-6, IL-8, M-CSF, GM-CSF, MCP-1, MIP-1α as well as increased RANKL/OPG ratio favouring osteoclast formation (Mandelin et al. 2003; Holt et al. 2007; Purdue et al. 2007; Goodman and Ma 2010). Pseudosynovial fluid from aseptically loosened implants effectively induces formation of osteoclasts in cell culture conditions and OPG diminished this effect (Kim et al. 2001; Mandelin et al. 2005b). It should be noted however, that there is some discrepancy in these findings and, for example, the strong pro-inflammatory cytokines TNFα and IL-1β are not detected so consistently as e.g. IL-6, IL-8 or MCP-1, MIP-1α/β or RANKL possibly reflecting individual variations or different stages of the disease progression (Purdue et al. 2007). Because systemic inflammation reaction (e.g. fever, fatigue, cachexia) is not typically observed in patients with aseptic osteolysis, the observed low levels of TNFα and IL-1 might actually represent poorly understood aspects of particle induced chronic inflammation. Studies conducted in animal models have further elaborated the role of pro-inflammatory cytokines and RANKL in the aseptic osteolysis. Several of these models have shown that implanted wear particles of various nature cause inflammation and *increase RANKL/OPG ratio and osteolysis* (Purdue et al. 2007). This reaction can be prevented or reduced by inhibiting TNFα signalling by using TNFα neutralizing antibody or by deleting TNF receptor (Childs et al. 2001). Similarly blocking of RANK signalling by RANK antagonist or OPG, or using mice genetically lacking RANK prevented experimental osteolysis in murine model of particle induced osteolysis (Childs et al. 2002; Ulrich-Vinther et al. 2002). Contradicting *in vivo* findings do also exist e.g. Taki et al. could not observe decrease in osteolysis in IL-1r and IL-6 or TNF double knock-out mice but concluded that osteolysis was

Pearl et al. 2011).

function (Vermes et al. 2001; Mandelin et al. 2005a).

likely mediated by other pro-inflammatory cytokines (Taki et al. 2007).

Taken together, following speculations can be made based on the detected cytokine profile of periprosthetic tissue. *First*, the definite majority of the cytokines detected are proinflammatory cytokines or chemokines secreted primary by cells of the innate immunity, which supports the essential role of activated macrophages in the process of aseptic loosening. *Second*, RANKL seems to be the most important end-point cytokine driving the osteoclast formation and loosening. *Third*, if only the cytokine profile of peri-implant tissues is considered, the role of adaptive immunity seems unlikely as both Th1 and Th2 signature and effector cytokines IFNγ and IL-4 are not generally detected from the peri-implant tissues. Furthermore both IFNγ and IL-4 effectively suppress the formation of osteoclasts. In this regard the possible role of Th 17 cells in the aseptic osteolysis seems an interesting line

regulate these tissue regenerative processes by modulating the growth and differentiation, of various cell types and tissues. Colony stimulating factors (CSF) typically control the myeloid development of monocytes and granulocytes, while epidermal growth factor (EGF), vascular endothelial growth factor (VEGF) and fibroblast growth factor (FGF) are secreted by macrophages and fibroblasts during regeneration of various tissues and they control e.g. formation of granulation tissue.

If the initial activation of innate immunity does not rapidly lead to the resolution of inflammation, the adaptive immune system is gradually activated by in particular professional antigen capturing, processing and presenting dendritic cells and macrophages that present MHC II-bound antigenic epitopes to the T-cell receptors (TCRs) of naïve CD4+ T-helper (Th) cells. The cytokine milieu that accompanies the antigen presentation process crucially determines the nature of the *developing immune response* (Th type 1, 2, 17, Treg). It is in the field of bone biology useful to distinguish a group of cytokines that control the formation, activation and survival of bone resorbing cells, osteoclasts. The complete list of cytokines that have been shown to *support osteoclastogenesis* or their bone resorptive activity is quite extensive and includes e.g. TNFα, IL-1, IL-6, IL-7, macrophage-CSF (M-CSF), IL-11, IL-17, MIP-1α/β, IFN-γ-induced protein-10 (IP-10) and monokine induced by IFN-γ (MIG), bone (or body) morphogenic proteins 2 and 7 (BMP2, 7) and VEGF among others. Cytokines that *inhibit osteoclastogenesis* include IFNγ, IFNβ, IL-4, IL-10, IL-13 and IL-18. Yet some cytokines play a dual role: low concentrations of TGF-β stimulate osteoclast formation and high concentrations inhibit it. However, none of the above-mentioned cytokines seems to be absolutely necessary for osteoclast development (Theill et al. 2002). The most important cytokine system controlling the osteoclast formation and function is the *RANKL/ RANK/OPG system* (Boyle et al. 2003). Mesenchymal cells, e.g. osteoblasts, mesenchymal stromal cells and fibroblasts as well as activated T cells tightly control osteoclastogenesis by producing receptor activator of nuclear factor kappa B ligand RANKL, which is produced in both cell membrane-bound and solubilized forms. Binding of RANKL to its receptor RANK expressed on osteoclast progenitors or macrophages in the presence of M-CSF, leads to cell fusion and formation of mature bone resorbing osteoclasts. In addition mesencymal cells produce osteoprotegerin (OPG) which is a soluble decoy receptor for RANKL and acts to limit and regulate its function. A high RANKL/OPG ratio is thus considered to drive osteoclastogenesis. RANKL/RANK/OPG system seems to represent the final common pathway in osteoclastogenesis, and many of the above-mentioned cytokines that drive or inhibit osteoclastogensis act by regulating RANKL or OPG levels. In the presence of low levels of M-CSF, RANKL seems to both sufficient and necessary for the complete differentiation of osteoclast precursors into mature osteoclasts.

### **4.3.2 Cytokines in aseptic loosening**

As the monocyte/macrophage and foreign body giant cells (FBGCs) are by far the most dominant cell type in the periprosthetic tissue, the wear particle activated macrophage has long been considered to *play a pivotal role in the development of aseptic loosening*.

Several studies over the years have demonstrated that monocyte/macrophages challenged *in vitro* with wear particles are activated to produce a wide variety of pro-inflammatory cytokines, chemokines and growth factors, including TNFα, IL-1β, IL-6, IL-8, MCP-1 and MIP-1α/β, VEGF, M-CSF, GM-CSF and even more so if particles are contaminated with bacterial products (Holt et al. 2007; Purdue et al. 2007; Goodman and Ma 2010). Also the shape, size, number, and biomaterial composition of wear particles seem to have an effect

regulate these tissue regenerative processes by modulating the growth and differentiation, of various cell types and tissues. Colony stimulating factors (CSF) typically control the myeloid development of monocytes and granulocytes, while epidermal growth factor (EGF), vascular endothelial growth factor (VEGF) and fibroblast growth factor (FGF) are secreted by macrophages and fibroblasts during regeneration of various tissues and they control e.g.

If the initial activation of innate immunity does not rapidly lead to the resolution of inflammation, the adaptive immune system is gradually activated by in particular professional antigen capturing, processing and presenting dendritic cells and macrophages that present MHC II-bound antigenic epitopes to the T-cell receptors (TCRs) of naïve CD4+ T-helper (Th) cells. The cytokine milieu that accompanies the antigen presentation process crucially determines the nature of the *developing immune response* (Th type 1, 2, 17, Treg). It is in the field of bone biology useful to distinguish a group of cytokines that control the formation, activation and survival of bone resorbing cells, osteoclasts. The complete list of cytokines that have been shown to *support osteoclastogenesis* or their bone resorptive activity is quite extensive and includes e.g. TNFα, IL-1, IL-6, IL-7, macrophage-CSF (M-CSF), IL-11, IL-17, MIP-1α/β, IFN-γ-induced protein-10 (IP-10) and monokine induced by IFN-γ (MIG), bone (or body) morphogenic proteins 2 and 7 (BMP2, 7) and VEGF among others. Cytokines that *inhibit osteoclastogenesis* include IFNγ, IFNβ, IL-4, IL-10, IL-13 and IL-18. Yet some cytokines play a dual role: low concentrations of TGF-β stimulate osteoclast formation and high concentrations inhibit it. However, none of the above-mentioned cytokines seems to be absolutely necessary for osteoclast development (Theill et al. 2002). The most important cytokine system controlling the osteoclast formation and function is the *RANKL/ RANK/OPG system* (Boyle et al. 2003). Mesenchymal cells, e.g. osteoblasts, mesenchymal stromal cells and fibroblasts as well as activated T cells tightly control osteoclastogenesis by producing receptor activator of nuclear factor kappa B ligand RANKL, which is produced in both cell membrane-bound and solubilized forms. Binding of RANKL to its receptor RANK expressed on osteoclast progenitors or macrophages in the presence of M-CSF, leads to cell fusion and formation of mature bone resorbing osteoclasts. In addition mesencymal cells produce osteoprotegerin (OPG) which is a soluble decoy receptor for RANKL and acts to limit and regulate its function. A high RANKL/OPG ratio is thus considered to drive osteoclastogenesis. RANKL/RANK/OPG system seems to represent the final common pathway in osteoclastogenesis, and many of the above-mentioned cytokines that drive or inhibit osteoclastogensis act by regulating RANKL or OPG levels. In the presence of low levels of M-CSF, RANKL seems to both sufficient and necessary for the complete

differentiation of osteoclast precursors into mature osteoclasts.

As the monocyte/macrophage and foreign body giant cells (FBGCs) are by far the most dominant cell type in the periprosthetic tissue, the wear particle activated macrophage has

Several studies over the years have demonstrated that monocyte/macrophages challenged *in vitro* with wear particles are activated to produce a wide variety of pro-inflammatory cytokines, chemokines and growth factors, including TNFα, IL-1β, IL-6, IL-8, MCP-1 and MIP-1α/β, VEGF, M-CSF, GM-CSF and even more so if particles are contaminated with bacterial products (Holt et al. 2007; Purdue et al. 2007; Goodman and Ma 2010). Also the shape, size, number, and biomaterial composition of wear particles seem to have an effect

long been considered to *play a pivotal role in the development of aseptic loosening*.

**4.3.2 Cytokines in aseptic loosening** 

formation of granulation tissue.

on the quantity and quality of the cytokines produced. This wear particle induced macrophage activation is possibly mediated by various types of cell surface pattern recognition receptors e.g. Toll-like receptors (Takagi et al. 2007; Tamaki et al. 2009) and likely involves the activation and action of transcription factor NF-kB (Lahdeoja et al. 2010; Pearl et al. 2011).

The mere wear particle stimulation of pure monocyte/macrophage cultures is not sufficient to drive osteoclastogeneis or foreign body giant cell formation, most likely due to lack of RANKL signalling. Further *in vitro* studies using other relevant cell types of the periprosthetic tissue have demonstrated that, to some extent, wear particles as such and especially pro-inflammatory cytokines effectively *up-regulate RANKL production* from periprosthetic tissue fibroblasts and osteoblasts as well as suppress osteoblast formation and function (Vermes et al. 2001; Mandelin et al. 2005a).

Consistent with *in vitro* findings, numerous studies of the explanted periprosthetic tissues and pseudosynovial fluid have demonstrated increased production of vast array proinflammatory cytokines, chemokines and growth factors including e.g. TNFα, IL-1β, IL-6, IL-8, M-CSF, GM-CSF, MCP-1, MIP-1α as well as increased RANKL/OPG ratio favouring osteoclast formation (Mandelin et al. 2003; Holt et al. 2007; Purdue et al. 2007; Goodman and Ma 2010). Pseudosynovial fluid from aseptically loosened implants effectively induces formation of osteoclasts in cell culture conditions and OPG diminished this effect (Kim et al. 2001; Mandelin et al. 2005b). It should be noted however, that there is some discrepancy in these findings and, for example, the strong pro-inflammatory cytokines TNFα and IL-1β are not detected so consistently as e.g. IL-6, IL-8 or MCP-1, MIP-1α/β or RANKL possibly reflecting individual variations or different stages of the disease progression (Purdue et al. 2007). Because systemic inflammation reaction (e.g. fever, fatigue, cachexia) is not typically observed in patients with aseptic osteolysis, the observed low levels of TNFα and IL-1 might actually represent poorly understood aspects of particle induced chronic inflammation.

Studies conducted in animal models have further elaborated the role of pro-inflammatory cytokines and RANKL in the aseptic osteolysis. Several of these models have shown that implanted wear particles of various nature cause inflammation and *increase RANKL/OPG ratio and osteolysis* (Purdue et al. 2007). This reaction can be prevented or reduced by inhibiting TNFα signalling by using TNFα neutralizing antibody or by deleting TNF receptor (Childs et al. 2001). Similarly blocking of RANK signalling by RANK antagonist or OPG, or using mice genetically lacking RANK prevented experimental osteolysis in murine model of particle induced osteolysis (Childs et al. 2002; Ulrich-Vinther et al. 2002). Contradicting *in vivo* findings do also exist e.g. Taki et al. could not observe decrease in osteolysis in IL-1r and IL-6 or TNF double knock-out mice but concluded that osteolysis was likely mediated by other pro-inflammatory cytokines (Taki et al. 2007).

Taken together, following speculations can be made based on the detected cytokine profile of periprosthetic tissue. *First*, the definite majority of the cytokines detected are proinflammatory cytokines or chemokines secreted primary by cells of the innate immunity, which supports the essential role of activated macrophages in the process of aseptic loosening. *Second*, RANKL seems to be the most important end-point cytokine driving the osteoclast formation and loosening. *Third*, if only the cytokine profile of peri-implant tissues is considered, the role of adaptive immunity seems unlikely as both Th1 and Th2 signature and effector cytokines IFNγ and IL-4 are not generally detected from the peri-implant tissues. Furthermore both IFNγ and IL-4 effectively suppress the formation of osteoclasts. In this regard the possible role of Th 17 cells in the aseptic osteolysis seems an interesting line

Aseptic Loosening of Total Hip Arthroplasty as a Result of Local Failure of Tissue Homeostasis 337

cytokines and orchestrate their microenvironment towards increased osteoclast differentiation and osteolysis. The mechanism by which macrophages recognize particles and become activated as a result is not fully understood. In this context it is interesting to speculate if perhaps excessive biomechanical loading, mast cell activation and/or tissue necrosis can release alarmins which could initiate expression of polarizing cytokines. Another option is that microbes or microbe-derived pathogen-associated molecular patterns could provide the danger signal initiating polarization of M0 to other macrophage subtypes (Tamaki et al. 2009). In fact, early in the adaptation process, tissue macrophages exhibit traits specific for the tissue in which they reside and, in the resolution phase of inflammation, they induce original tissue repair or fibrosis. It has been postulated that during initial or middle phases of tissue response on wear particle load M1 macrophage activation pattern could play a key role while end-stage of aseptic loosening could be associated with alternatively activated macrophages expressing predominantly chemokines

Macrophages have their own weak bone-resorptive capacity but more importantly they are the *cell reservoir for the generation of multinucleated giant cells* (MGCs). Of these the multinuclear bone resorbing osteoclast derived from M0 macrophages is the most important consumer of bone. Induction of monocyte/macrophage-derived multinucleated giant cells has been demonstrated in response to hematopoietic growth factors (e.g. IL-3, granulocyte macrophage colony stimulating factor, GM-CSF), and IL-4. The role of adhesion molecules in polykaryon cell formation has also been reported, and mainly involves intercellular adhesion molecule-1 (ICAM-1/CD54) and the receptor CR3 (CD11b/CD18) expressed by multinucleated giant cells (Anderson et al. 2008). Furthermore, immunohistochemical analysis of tissues retrieved from aseptic loosened total hip and knee replacements has shown high levels of ICAM-1 and its receptor CD11b expressed by giant cells and

Fibrocytes are *mesenchymal cells that arise from monocyte precursors derived from blood*. They express features of both macrophages and fibroblasts dependent on local factors participating in acute response to injury/surgery and also play an important role in regulation/ resolution of chronic inflammation (Reilkoff et al. 2011). Their increased occurrence in periprosthetic tissues long after the healing phase could be associated with

About 30% of periprosthetic membrane is composed of fibroblasts and these cells exhibit a very high proliferation rate pointing on their activation state (Koreny et al. 2006). Fibroblasts are *tissue resident cells* that have the capacity to synthesize and remodel extracellular matrix. In this way they create the local tissue architecture and also develop a milieu in which particular homeostatic regulation (e.g. inflammation, repair) can occur (Buckley 2011). Thus, they could significantly influence the fate of a prosthesis because they govern the switch to resolving postoperative inflammation and formation of fibrous capsule (repair) during the initial postsurgery period (Anderson 2009). In fact, fibroblasts and fibrocytes residing in surface layers of the periprosthetic capsule are the first cell types that come into contact with prosthetic particles. Later, a series of interactions between immune cells (esp. macrophages) and fibroblasts could determine the overall adaptative capacity of periprosthetic tissues for wear debris. For this reason, fibroblasts in concert with other cells greatly influence the grade of periprosthetic inflammation.

such as CHIT1 and CCL18 (Koulouvaris et al. 2008).

phagocytic cells containing metal and polyethylene particles.

**Fibroblasts/ Fibrocytes** 

disturbance of local tissue homeostasis.

of further research as their signature cytokine IL-17 has been shown to favour osteoclastogenesis. *Fourth*, the nature of FBGCs remains somewhat elusive as morphologically sound FBGCs can be made in cell culture conditions using long term IL-4 stimulation, but as stated the IL-4 is not present in the periprosthetic tissue in any considerable amounts.

Based on the *in vitro* cell stimulation experiments and explanted tissue analysis, a following sequence of events can be postulated: wear particles activate macrophages, even more so if particles are coated by microbial products or alarmins. TLRs and other pattern recognition receptors of the innate immunity might mediate this wear particle induced cell activation. Activated macrophages produce pro-inflammatory chemokines, cytokines and growth factors which cause, via local endothelial activation and chemotaxis, further recruitment of monocytes and osteoclast precursors to the periprosthetic tissue. Pro-inflammatory cytokines, and to some extend also wear particles directly, drive the production of RANKL from mesenchymal cells and at the same time suppress the production of OPG and bone formation. Together with the pro-inflammatory cytokines they create an environment which favours ostoclastogenesis, bone resorption and finally prosthesis loosening. As the RANK/RANKL/OPG system seems represent the final common pathway in the proinflammatory cytokine mediated osteolysis, inhibition of RANKL might seem as a rational means to prevent aseptic osteolysis.

### **4.4 Key cell players – macrophages, fibroblasts, lymphocytes, osteoblasts, osteoclasts** *(Gallo, Goodman)*

Histological and immunohistochemical studies of retrieved tissues have shown that when loosening is associated with osteolysis, the cellular profile is dependent on the method of prosthesis fixation (cemented or uncemented). Indeed, in the cemented group, tissue was more plentiful and often composed of highly vascularized fibrous tissue with a preponderance of macrophages and, to a lesser degree, T lymphocytes. With uncemented implants, tissue was less abundant and less cellular and more fibrous in nature. In addition to large numbers of fibroblasts, macrophages were also abundant but with increased numbers of all T-lymphocyte subgroups (Goodman et al. 1998).

### **Macrophages**

Macrophages comprise a *heterogeneous group of cells* that play an important role in both the activation of the immune response and tissue homeostasis. They are *derived from monocytes which arise from myeloid progenitor cells in the bone marrow*. After maturation in bone marrow, monocytes are released into the circulation, then enter the tissues and differentiate into *tissue macrophages (M0)* that can survive for several months. Here, they can be activated by microbial/non-microbial agents (proinflammatory *M1 macrophage activation*), by IL-4, IL-13 etc. produced locally by mast cells and/or TREG lymphocytes (*M2 antiinflammatory tissueremodelling macrophages*), or by other signals (*M2-like immunoregulatory macrophages*). Suppression of cytokine signalling 1 (SOCS1) is a key determinant of differential macrophage activation and function influencing the overall macrophage balance in tissues together with other factors (Ma et al. 2003; Whyte et al. 2011).

Macrophages are considered *the most potent phagocytic cells* in tissues around the THA and there are believed to be able to phagocytise even non-opsonized particles, e.g. polyethylene, metal or polymethylmethacrylate. After phagocytosis of the wear particles (the "fuel of the particle disease"), the macrophages have to become activated to produce pro-inflammatory cytokines and orchestrate their microenvironment towards increased osteoclast differentiation and osteolysis. The mechanism by which macrophages recognize particles and become activated as a result is not fully understood. In this context it is interesting to speculate if perhaps excessive biomechanical loading, mast cell activation and/or tissue necrosis can release alarmins which could initiate expression of polarizing cytokines. Another option is that microbes or microbe-derived pathogen-associated molecular patterns could provide the danger signal initiating polarization of M0 to other macrophage subtypes (Tamaki et al. 2009). In fact, early in the adaptation process, tissue macrophages exhibit traits specific for the tissue in which they reside and, in the resolution phase of inflammation, they induce original tissue repair or fibrosis. It has been postulated that during initial or middle phases of tissue response on wear particle load M1 macrophage activation pattern could play a key role while end-stage of aseptic loosening could be associated with alternatively activated macrophages expressing predominantly chemokines such as CHIT1 and CCL18 (Koulouvaris et al. 2008).

Macrophages have their own weak bone-resorptive capacity but more importantly they are the *cell reservoir for the generation of multinucleated giant cells* (MGCs). Of these the multinuclear bone resorbing osteoclast derived from M0 macrophages is the most important consumer of bone. Induction of monocyte/macrophage-derived multinucleated giant cells has been demonstrated in response to hematopoietic growth factors (e.g. IL-3, granulocyte macrophage colony stimulating factor, GM-CSF), and IL-4. The role of adhesion molecules in polykaryon cell formation has also been reported, and mainly involves intercellular adhesion molecule-1 (ICAM-1/CD54) and the receptor CR3 (CD11b/CD18) expressed by multinucleated giant cells (Anderson et al. 2008). Furthermore, immunohistochemical analysis of tissues retrieved from aseptic loosened total hip and knee replacements has shown high levels of ICAM-1 and its receptor CD11b expressed by giant cells and phagocytic cells containing metal and polyethylene particles.

### **Fibroblasts/ Fibrocytes**

336 Recent Advances in Arthroplasty

of further research as their signature cytokine IL-17 has been shown to favour osteoclastogenesis. *Fourth*, the nature of FBGCs remains somewhat elusive as morphologically sound FBGCs can be made in cell culture conditions using long term IL-4 stimulation, but as stated the IL-4 is not present in the periprosthetic tissue in any

Based on the *in vitro* cell stimulation experiments and explanted tissue analysis, a following sequence of events can be postulated: wear particles activate macrophages, even more so if particles are coated by microbial products or alarmins. TLRs and other pattern recognition receptors of the innate immunity might mediate this wear particle induced cell activation. Activated macrophages produce pro-inflammatory chemokines, cytokines and growth factors which cause, via local endothelial activation and chemotaxis, further recruitment of monocytes and osteoclast precursors to the periprosthetic tissue. Pro-inflammatory cytokines, and to some extend also wear particles directly, drive the production of RANKL from mesenchymal cells and at the same time suppress the production of OPG and bone formation. Together with the pro-inflammatory cytokines they create an environment which favours ostoclastogenesis, bone resorption and finally prosthesis loosening. As the RANK/RANKL/OPG system seems represent the final common pathway in the proinflammatory cytokine mediated osteolysis, inhibition of RANKL might seem as a rational

**4.4 Key cell players – macrophages, fibroblasts, lymphocytes, osteoblasts,** 

numbers of all T-lymphocyte subgroups (Goodman et al. 1998).

together with other factors (Ma et al. 2003; Whyte et al. 2011).

Histological and immunohistochemical studies of retrieved tissues have shown that when loosening is associated with osteolysis, the cellular profile is dependent on the method of prosthesis fixation (cemented or uncemented). Indeed, in the cemented group, tissue was more plentiful and often composed of highly vascularized fibrous tissue with a preponderance of macrophages and, to a lesser degree, T lymphocytes. With uncemented implants, tissue was less abundant and less cellular and more fibrous in nature. In addition to large numbers of fibroblasts, macrophages were also abundant but with increased

Macrophages comprise a *heterogeneous group of cells* that play an important role in both the activation of the immune response and tissue homeostasis. They are *derived from monocytes which arise from myeloid progenitor cells in the bone marrow*. After maturation in bone marrow, monocytes are released into the circulation, then enter the tissues and differentiate into *tissue macrophages (M0)* that can survive for several months. Here, they can be activated by microbial/non-microbial agents (proinflammatory *M1 macrophage activation*), by IL-4, IL-13 etc. produced locally by mast cells and/or TREG lymphocytes (*M2 antiinflammatory tissueremodelling macrophages*), or by other signals (*M2-like immunoregulatory macrophages*). Suppression of cytokine signalling 1 (SOCS1) is a key determinant of differential macrophage activation and function influencing the overall macrophage balance in tissues

Macrophages are considered *the most potent phagocytic cells* in tissues around the THA and there are believed to be able to phagocytise even non-opsonized particles, e.g. polyethylene, metal or polymethylmethacrylate. After phagocytosis of the wear particles (the "fuel of the particle disease"), the macrophages have to become activated to produce pro-inflammatory

considerable amounts.

means to prevent aseptic osteolysis.

**osteoclasts** *(Gallo, Goodman)* 

**Macrophages** 

Fibrocytes are *mesenchymal cells that arise from monocyte precursors derived from blood*. They express features of both macrophages and fibroblasts dependent on local factors participating in acute response to injury/surgery and also play an important role in regulation/ resolution of chronic inflammation (Reilkoff et al. 2011). Their increased occurrence in periprosthetic tissues long after the healing phase could be associated with disturbance of local tissue homeostasis.

About 30% of periprosthetic membrane is composed of fibroblasts and these cells exhibit a very high proliferation rate pointing on their activation state (Koreny et al. 2006). Fibroblasts are *tissue resident cells* that have the capacity to synthesize and remodel extracellular matrix. In this way they create the local tissue architecture and also develop a milieu in which particular homeostatic regulation (e.g. inflammation, repair) can occur (Buckley 2011). Thus, they could significantly influence the fate of a prosthesis because they govern the switch to resolving postoperative inflammation and formation of fibrous capsule (repair) during the initial postsurgery period (Anderson 2009). In fact, fibroblasts and fibrocytes residing in surface layers of the periprosthetic capsule are the first cell types that come into contact with prosthetic particles. Later, a series of interactions between immune cells (esp. macrophages) and fibroblasts could determine the overall adaptative capacity of periprosthetic tissues for wear debris. For this reason, fibroblasts in concert with other cells greatly influence the grade of periprosthetic inflammation.

Aseptic Loosening of Total Hip Arthroplasty as a Result of Local Failure of Tissue Homeostasis 339

CSF, required for maturation of osteoclasts and also for formation of other giant cells having potentially important implications for periprosthetic osteolysis and aseptic loosening.

Osteoclasts are *multinucleated bone-resorbing cells* that together with osteoblasts, play a pivotal role in bone homeostasis and bone remodeling. Osteoclast precursors *derive from bone marrow as M0 macrophages* circulating in blood until penetrating tissues and binding to the bone surface utilizing cell-specific signalling pathways. The mechanism of recognition of a particular place on the bone surface is not known. On the other hand, the integrins mediating the bonding of pre-osteoclast to the bone site are well described (e.g. αvβ3). At the bone site, pre-osteoclasts fuse into multinucleated cells under the influence of fusion factors (e.g. dendritic cell-specific transmembrane protein; DC-STAMP) and then maturate in a particular milieu with the help at least of macrophage colony-stimulating factor (M-CSF) and receptor activator of nuclear factor kappa B ligand (RANKL), (Fig. 3). TNF-α and IL-1 potentiate this process, whereas a soluble RANKL neutralizer osteoprotegerin (OPG) inhibits it (Anderson et al. 2008). Indeed, RANKL/OPG ratio is considered a key regulator of osteoclastogenesis although *in vivo,* the RANKL-RANK interaction may occur in a direct cell-cell contact, e.g. between mesenchymal stromal cells and osteoclast progenitors or osteoblasts and pre-osteoclasts. A few giant osteoclasts form the cutting cone at basic multicellular units (BMU) in osteons (compact bone) and hemi-osteons (trabecular bone) that undergo resorption over a few weeks. The resorption space (lacuna) in healthy bone is filled in a few months by the closing cone composed of osteoblasts. However, under pathological conditions such as particle-mediated inflammation, the distortion of the coupling mechanism prevents closure of cone by osteoblasts. Hence, increased osteoclast bone resorption along with other mechanisms continues to create larger bone defects.

**4.5 Contribution of allergic and other metal ion mediated responses to loosening of a** 

Allergy and sensitization are frequently used terms, but their definitions should be clarified. Sensitization may be defined as "the process in which exposure to an antigen results in the development of hypersensitivity" whereas a definition of allergy could be "an altered body reaction, usually hypersensitivity, as a response to exposure to a specific substance". It is generally accepted that all metal prostheses release metal debris following wear and corrosion. This has been shown by measuring significantly elevated level of metal ions in hip arthroplasty patients; e.g capsular and periprosthetic tissues, the liver, spleen and lymph nodes, serum and urine (Hallab and Jacobs 2009). Metal ions may potentially activate the immune system, a response that is mainly adaptive rather than innate (Hallab et al. 2008). Typically, large particulate wear debris particles (more than 150 nm to 10 μm) are phagocytosed by cells including macrophages, which present antigens to circulating Tlymphocytes resulting in delayed type hypersensitivity reactions (Hallab and Jacobs 2009). Allergic complications following insertion of metallic orthopaedic implants are plenty and summarized in table 1. In this section, we will discuss the possible association between

Immune reactions described adjacent to metal implants *include T-cell mediated type IV reactions but likely are multifactorial*. Peri-implant reactions seem to be Th1 dominant. Hence, increased levels of IFN-γ and IL-6 in metal hypersensitive joint arthroplasty patients (Hallab et al. 2008) as well as low IFN-γ and elevated levels of IL-17 in nickel allergic patients with

**total hip arthroplasty** *(Thyssen, Schalock)* 

metal allergy and loosening of hip arthroplasties.

**Osteoclasts** 

Aberrant expression of adhesion molecules, chemokines, cytokines, and their receptors could increase the inflammatory status, and vice-versa.

### **Lymphocytes**

Local perivascular infiltrates of lymphocytes and even diffuse lymphocytes have been identified in specimens of periprosthetic tissue from both MoM and non-MoM implants (Fujishiro et al. 2011). *Lymphocytes are a kind of leukocyte circulating in the blood*. Precursors of both T and B cells are cloned in bone marrow but T cells require in addition, maturation in the thymus. The surface of each T cell displays thousands of identical T cell receptors and according to the presence of one or the other of two surface glycoproteins, they are designated as CD4 and CD8 T lymphocytes. The recruitment of lymphocytes into peripheral tissues where there is inflammation and/or damage is mediated by chemokines and adhesion molecules. In tissues, each B and T cell develops a specific response for a particular antigen interplaying complexly with innate immunity. As a result, different classes of pathogens stimulate differentiation of CD4+ T cells into effector T cells (Th1, Th2, Th17 subsets). The aim is to induce the best response to a particular danger signal without vast tissue damage. Therefore, an important role is played by the activity of regulatory T (Treg) cells which suppress inappropriate adaptive immune response via several regulatory mechanisms (e.g. TGF-β, IL-10).

T lymphocytes are key players in the process of local and systemic bone loss associated with inflammation (Djaafar et al. 2010). Importantly, direct cell-to-cell contact between activated T cells and pre-osteoclasts is critical for their maturation in osteoclasts via RANKL and other pro-inflammatory cytokines pathways. Natural killer cells (lymphocytes of the innate immune system) have been shown to be a principal tissue-infiltrating lymphocyte subset in patients with particle-mediated inflammation (Huss et al. 2010). However, despite the above-mentioned consequences, the role of the specific immune system in particle disease remains controversial.

### **Osteoblasts**

The osteoblast is a *mesenchymal cell with a primary role in bone formation and remodelling*. To do this, they come in contact with other cell groups of different origins (including osteoclasts, macrophages, lymphocytes, bone marrow cells) and synthesize a variety of cytokines and chemokines. Such interactions likely play an important role under both physiologic and pathological conditions.

Given the problem is increased bone resorption around the implant, the interplay between osteoblasts and other cell groups likely creates an environment favouring increased osteoclast maturation and/or suppressed osteoblast activity. Investigated have been the following possible explanations. First, some types of prosthetic particles could be toxic to osteoprogenitor cells leading to slowing of their maturation. Second, wear debris could suppress the function of mature osteoblasts inhibiting their osteogenic capacity. Third, osteoblasts could phagocytose prosthetic particles and express pro-inflammatory and proosteoclastogenic molecules on their surface, contributing in this way to the severity of bone resorption. The same can be achieved by chronic inflammatory microenvironment that upregulates osteoblast synthesis of pro-osteoclastogenic molecules (Granchi et al. 2005; Fujii et al. 2011). Osteoblasts can contribute to the predominance of bone resorption at implant-bone interface also by specific interaction between them and T cells. Osteoblasts can express immunoreactive structures on their surfaces (superantigens) which directly stimulate T cells in their surroundings (Stanley et al. 2006). Activated T cells could express RANKL and M-

CSF, required for maturation of osteoclasts and also for formation of other giant cells having potentially important implications for periprosthetic osteolysis and aseptic loosening.

### **Osteoclasts**

338 Recent Advances in Arthroplasty

Aberrant expression of adhesion molecules, chemokines, cytokines, and their receptors

Local perivascular infiltrates of lymphocytes and even diffuse lymphocytes have been identified in specimens of periprosthetic tissue from both MoM and non-MoM implants (Fujishiro et al. 2011). *Lymphocytes are a kind of leukocyte circulating in the blood*. Precursors of both T and B cells are cloned in bone marrow but T cells require in addition, maturation in the thymus. The surface of each T cell displays thousands of identical T cell receptors and according to the presence of one or the other of two surface glycoproteins, they are designated as CD4 and CD8 T lymphocytes. The recruitment of lymphocytes into peripheral tissues where there is inflammation and/or damage is mediated by chemokines and adhesion molecules. In tissues, each B and T cell develops a specific response for a particular antigen interplaying complexly with innate immunity. As a result, different classes of pathogens stimulate differentiation of CD4+ T cells into effector T cells (Th1, Th2, Th17 subsets). The aim is to induce the best response to a particular danger signal without vast tissue damage. Therefore, an important role is played by the activity of regulatory T (Treg) cells which suppress inappropriate adaptive immune response via several regulatory

T lymphocytes are key players in the process of local and systemic bone loss associated with inflammation (Djaafar et al. 2010). Importantly, direct cell-to-cell contact between activated T cells and pre-osteoclasts is critical for their maturation in osteoclasts via RANKL and other pro-inflammatory cytokines pathways. Natural killer cells (lymphocytes of the innate immune system) have been shown to be a principal tissue-infiltrating lymphocyte subset in patients with particle-mediated inflammation (Huss et al. 2010). However, despite the above-mentioned consequences, the role of the specific immune system in particle disease

The osteoblast is a *mesenchymal cell with a primary role in bone formation and remodelling*. To do this, they come in contact with other cell groups of different origins (including osteoclasts, macrophages, lymphocytes, bone marrow cells) and synthesize a variety of cytokines and chemokines. Such interactions likely play an important role under both physiologic and

Given the problem is increased bone resorption around the implant, the interplay between osteoblasts and other cell groups likely creates an environment favouring increased osteoclast maturation and/or suppressed osteoblast activity. Investigated have been the following possible explanations. First, some types of prosthetic particles could be toxic to osteoprogenitor cells leading to slowing of their maturation. Second, wear debris could suppress the function of mature osteoblasts inhibiting their osteogenic capacity. Third, osteoblasts could phagocytose prosthetic particles and express pro-inflammatory and proosteoclastogenic molecules on their surface, contributing in this way to the severity of bone resorption. The same can be achieved by chronic inflammatory microenvironment that upregulates osteoblast synthesis of pro-osteoclastogenic molecules (Granchi et al. 2005; Fujii et al. 2011). Osteoblasts can contribute to the predominance of bone resorption at implant-bone interface also by specific interaction between them and T cells. Osteoblasts can express immunoreactive structures on their surfaces (superantigens) which directly stimulate T cells in their surroundings (Stanley et al. 2006). Activated T cells could express RANKL and M-

could increase the inflammatory status, and vice-versa.

**Lymphocytes** 

mechanisms (e.g. TGF-β, IL-10).

remains controversial.

pathological conditions.

**Osteoblasts** 

Osteoclasts are *multinucleated bone-resorbing cells* that together with osteoblasts, play a pivotal role in bone homeostasis and bone remodeling. Osteoclast precursors *derive from bone marrow as M0 macrophages* circulating in blood until penetrating tissues and binding to the bone surface utilizing cell-specific signalling pathways. The mechanism of recognition of a particular place on the bone surface is not known. On the other hand, the integrins mediating the bonding of pre-osteoclast to the bone site are well described (e.g. αvβ3). At the bone site, pre-osteoclasts fuse into multinucleated cells under the influence of fusion factors (e.g. dendritic cell-specific transmembrane protein; DC-STAMP) and then maturate in a particular milieu with the help at least of macrophage colony-stimulating factor (M-CSF) and receptor activator of nuclear factor kappa B ligand (RANKL), (Fig. 3). TNF-α and IL-1 potentiate this process, whereas a soluble RANKL neutralizer osteoprotegerin (OPG) inhibits it (Anderson et al. 2008). Indeed, RANKL/OPG ratio is considered a key regulator of osteoclastogenesis although *in vivo,* the RANKL-RANK interaction may occur in a direct cell-cell contact, e.g. between mesenchymal stromal cells and osteoclast progenitors or osteoblasts and pre-osteoclasts. A few giant osteoclasts form the cutting cone at basic multicellular units (BMU) in osteons (compact bone) and hemi-osteons (trabecular bone) that undergo resorption over a few weeks. The resorption space (lacuna) in healthy bone is filled in a few months by the closing cone composed of osteoblasts. However, under pathological conditions such as particle-mediated inflammation, the distortion of the coupling mechanism prevents closure of cone by osteoblasts. Hence, increased osteoclast bone resorption along with other mechanisms continues to create larger bone defects.

### **4.5 Contribution of allergic and other metal ion mediated responses to loosening of a total hip arthroplasty** *(Thyssen, Schalock)*

Allergy and sensitization are frequently used terms, but their definitions should be clarified. Sensitization may be defined as "the process in which exposure to an antigen results in the development of hypersensitivity" whereas a definition of allergy could be "an altered body reaction, usually hypersensitivity, as a response to exposure to a specific substance". It is generally accepted that all metal prostheses release metal debris following wear and corrosion. This has been shown by measuring significantly elevated level of metal ions in hip arthroplasty patients; e.g capsular and periprosthetic tissues, the liver, spleen and lymph nodes, serum and urine (Hallab and Jacobs 2009). Metal ions may potentially activate the immune system, a response that is mainly adaptive rather than innate (Hallab et al. 2008). Typically, large particulate wear debris particles (more than 150 nm to 10 μm) are phagocytosed by cells including macrophages, which present antigens to circulating Tlymphocytes resulting in delayed type hypersensitivity reactions (Hallab and Jacobs 2009). Allergic complications following insertion of metallic orthopaedic implants are plenty and summarized in table 1. In this section, we will discuss the possible association between metal allergy and loosening of hip arthroplasties.

Immune reactions described adjacent to metal implants *include T-cell mediated type IV reactions but likely are multifactorial*. Peri-implant reactions seem to be Th1 dominant. Hence, increased levels of IFN-γ and IL-6 in metal hypersensitive joint arthroplasty patients (Hallab et al. 2008) as well as low IFN-γ and elevated levels of IL-17 in nickel allergic patients with

Aseptic Loosening of Total Hip Arthroplasty as a Result of Local Failure of Tissue Homeostasis 341

It has often been debated whether metal allergy increases the risk of developing delayed type hypersensitivity reactions, e.g. loosening of THA. The literature was recently reviewed and it was concluded that metal allergy might in a minority increase the risk of complications caused by a delayed type hypersensitivity reaction (Thyssen et al. 2011a). Also, we do not know how to identify the subgroups of metal contact allergic patients with a potentially increased risk of complications following insertion of a metal implant. However, it seems to be certain that insertion of metal bearings increase the prevalence of metal allergy as a review showed it was 25% among patients with well-functioning hip arthroplasties and 60% among patients with a failed or a poorly functioning implants (Hallab et al. 2001); considerably higher frequencies than general population estimates (Thyssen et al. 2007). Despite the higher prevalence in the latter patient group, it remains unknown whether the higher prevalence is caused by implant loosening or whether metal

In addition to metallic devices and prostheses, metal exposure is also related to intake of food and water, dental work, tattooing, and prolonged skin contact with metal objects. These exposures may all result in metal sensitization although oral intake sometimes may lead to tolerance. Oral intake of food and water as well as implantation seems rarely to result in allergic complications or clinical disease. However, skin contact with metal objects e.g. jewellery often results in skin sensitization and dermatitis at the site of skin contact. Several field studies have recently demonstrated that nickel release in concentrations that result in sensitization and dermatitis remains common in Europe and the United States

Patch testing is widely used to establish a diagnosis of delayed type hypersensitivity to metals. The reproducibility of the patch test is high but allergen dependent (Brasch et al. 1994). Patch tests are typically applied to the upper back and occluded for 48 hr. Readings should be performed at least on day 3 or 4 and if possible on more than one occasion (Wahlberg 2006). Patch test studies have suggested that 24-34.5% of positive patch test reactions potentially are missed when readings are not performed beyond day 2 (Uter et al. 1996). Ready-to-use test systems such as the Thin-layer Rapid Use Epicutaneous (TRUE) test® generally have a good concordance with conventional patch test systems using e.g. Finn Chambers®, except for cobalt (Lazarov et al. 2007). Patch testing intends to identify contact sensitized subjects by distinguishing between negative, irritant and allergic reactions. Thus, a valid positive patch test reaction typically requires a trained and experienced person that adheres to a set of valid criteria. Currently, the recommendations from the International Contact Dermatitis Research Group (ICDRG) dictate that homogeneous redness and infiltration in the entire test area is scored as a 1+ reaction, homogeneous redness, infiltration, and vesicles in the test area are scored as a 2+ reaction, and homogeneous redness, infiltration, and coalescing vesicles in the test area as a 3+ reaction (Wilkinson et al. 1970). 1+, 2+, or 3+ readings should be interpreted as positive responses indicating contact sensitization whereas irritant responses, doubtful (+?) responses, or negative readings should be interpreted as negative responses (Fig. 4). Clinicians should be aware that false positive and negative reactions may be encountered

hypersensitivity results in loosening.

(Thyssen and Menne 2010; Thyssen et al. 2011b).

**4.5.2 Metal exposure** 

**4.5.3 Patch testing** 

symptomatic joint implants but not in Ni allergic patients with well-functioning joint implants (Summer et al. 2010) have been reported. Tissues near metal devices in those with metal hypersensitivity may have elevated immune cells/markers including: CD3þ Tlymphocytes, CD4þ cells, CD11cþ macrophages/dendritic cells, and cells with abundant MHC class II (HLA-DR) expression (dendritic cells), (Cadosch et al. 2009). Foreign body giant cell formation is often noted when phagocytosis of foreign particles by macrophages, including metals. Ingesting macrophages then secrete the proinflammatory cytokines TNFα, IL-6 and IL1-α and –β (Cadosch et al. 2009). At high concentrations, Ti and V cause production of superoxide anions in neutrophils and nickel ions break down neutrophil cell membranes at high levels (Kumazawa et al. 2002).

A novel mechanism for aseptic loosening of metal implants was suggested recently (Cadosch et al. 2010a). Osteoclasts mature and grow on titanium metals, leading to degradation, uptake and eventual release of the metal ions by the osteoclasts. This might explain increases in metal ion in systemic circulation, increased recruitment of osteoclast precursors via synthesis of specific chemokines, and also it can contribute to osteoclast differentiation. CCL17/TARC, CCL22/MDC, RANK-L, M-CSF and pro-inflammatory cytokines (CCR4) are elevated in the peri-implant environment (Cadosch et al. 2010b). This leads to loosening of the device as supporting surrounding bone is resorbed. In addition to being present in joint loosening, CCR4 is involved in the inflammatory reaction in cutaneous allergic contact dermatitis reactions (Vocanson et al. 2009). The link between cutaneous dermatitis and CCR4 is not proven, but is an area for further inquiry. While this evidence specifically address Ti metal, these same mechanisms could be relevant for other metal ions as well. Below, a description of the epidemiology of metal allergy and recommendation for clinical practice will be provided.


Table 1. Selected clinical manifestations of delayed type hypersensitivity reactions
