Tissue Cytokine Patterns in Patients with Polymyalgia Rheumatica and Giant Cell Arteritis

Methods

Patients and Normal Temporal Artery Samples

Forty-eight temporal artery specimens from a consecutive case series of 34 patients having biopsies were studied. Specimens from 15 patients showed an inflammatory infiltrate by standard histologic techniques, confirming the diagnosis of giant cell arteritis [3]. Nine patients without histologic evidence for giant cell arteritis had polymyalgia rheumatica based on the following criteria: at least a 4-week history of aching and morning stiffness involving two of three areas (shoulders and proximal arms, hips and proximal thighs, and neck and torso) and an erythrocyte sedimentation rate of more than 40 mm/h in the absence of any other underlying disease [7]. Ten patients did not have any inflammatory lesions suggestive of giant cell arteritis or evidence of polymyalgia rheumatica, and these patients served as controls. The ultimate diagnoses in these patients included fever of unknown origin, the myelodysplastic syndrome, diplopia, and ischemic optic neuropathy. None of the patients who had negative biopsy results had clinical findings strongly suggestive of giant cell arteritis. The demographic and clinical data of the study patients are shown in Table 1.

Complementary DNA Synthesis and Polymerase Chain Reaction Amplification of Cytokines

Ten 5- to 10-µmthick sections of temporal artery biopsy specimens were lysed using 5 µL of a 1% Nonidet P-40 nonionic detergent lysis solution (1% weight by volume Nonidet P-40 nonionic detergent, 10 mmol/L Tris HCl [pH 8.0], 10 mmol/L NaCl, 3 mmol/L MgCl₂, and 40 units ribonuclease inhibitor [Boehringer Mannheim, Indianapolis, Indiana]). Oligo-deoxyribothymidylic acid (15mer; 50 pmol) and 8 µL of diethyl pyrocarbonate-treated distilled H₂O were added to the lysate. Lysates were incubated at 65 °C for 5 minutes, briefly microfuged at 4 °C, and placed on ice. Supernatants were moved to a new 0.5-mL Eppendorf tube leaving pelleted solid material behind. 4 µL of 5 x concentrated reverse transcriptase buffer (250 mmol/L Tris HCl; 30 mmol/L MgCl₂; 500 mmol/L NaCl; pH 8.2 [at 37 °C]), 2 µL of 0.1 mol/L 1,4-dithiothreitol, 2 µL of 10 mmol/L deoxyribo-nucleotide triphosphates, 40 units of ribonuclease inhibitor, and 12 units of reverse transcriptase from avian myeloblastosis virus were all added to the lysate samples. Samples were incubated for 2 hours at 37 °C. The following oligonucleotide primers were used to analyze for cytokine mRNA expression by polymerase chain reaction amplification:

1. Interleukin-1 β (GACACATGGGATAACGAGGC, ACGCAGGACAGGTACAGATT).

2. Interleukin-2 (ACTCACCAGGATGCTCACAT, AGGTAATCCATCTGTTCAGA).

3. Interleukin-4 (CTTCCCCCTCTGTTCTTCCT, TTCCTGTCGAGCCGTTTCAG).

4. Interleukin-5 (ATGAGGATGCTTCTGCATTTG, TCAACTTTCTATTATCCACTCGGTGTTCATTAC).

5. Interleukin-6 (GATGTAGCCGCCCCACACAGACAG, CCTCAAACTCCAAAAGACCAGTGATG).

6. Interferon-γ (AGTTATATCTTGGCTTTTCA, ACCGAATAATTAGTCAGCTT).

7. Transforming growth factor-β 1 (GCCCTGGACACCAACTATTGC, GCTGCACTTGCAGGAGCGCAC).

8. Tumor necrosis factor-α (TCTCGAACCCCGAGTGACAA, TATCTCTCAGCTCCACGCCATT).

9. Granulocyte-macrophage-colony stimulating factor (TGGCTGCAGAGCCTGCTGCTC, TCACTCCTGGACTGGCTCCC).

10. β-actin (ATCATGTTTGAGACCTTCAACACC, CATGGTGGTGCCGCCAGACAG).

11. C α (GAACCCTGACCCTGCCGTGTACC, ATCATAAATTGGGGTAGGATCC).

Samples were incubated at 95 °C for 5 minutes and cycled for 1 minute at 94 °C, 2 minutes at 55 °C, and 2 minutes at 72 °C for 30 cycles followed by a 10-minute extension at 72 °C. Polymerase chain reaction products were reamplified for another 30 cycles by using the identical primer set and the identical amplification conditions. The amplified templates were finally separated on 2% agarose gels containing 0.1% ethidium bromide and were visualized using ultraviolet light. The sensitivity of the polymerase chain reaction approach was established on serial dilutions of freshly separated peripheral blood mononuclear cells. The polymerase chain reaction was equally sensitive to show interleukin-2, interferon-γ, interleukin-4, and interleukin-5 specific sequences.

Statistical Analysis

Patients with giant cell arteritis, polymyalgia rheumatica, and unrelated diseases were compared for the tissue expression of cytokine-specific sequences by the chi-square test or the Fisher exact probability test, if appropriate.

Results

Patients with Giant Cell Arteritis

Expression of Macrophage-Derived Cytokines

Immunohistochemical studies have shown that macrophages and CD4+ T lymphocytes are the dominant cell types in vasculitic infiltrates. To define the cytokine profile of tissue-infiltrating macrophages characteristic of giant cell arteritis, 15 temporal artery tissue specimens with typical vasculitic infiltrates were analyzed for cytokine-specific sequences by polymerase chain reaction. Tissue extracts from 16 specimens morphologically negative for inflammatory lesions from 10 patients who had unrelated diseases and who did not have clinical evidence of polymyalgia rheumatica or giant cell arteritis were examined in parallel and served as controls. A positive amplification signal for actin was used to ensure adequate RNA extraction and complementary DNA synthesis. Interleukin-1, interleukin-6, and tumor necrosis factor-α constitute a group of proinflammatory cytokines predominantly synthesized by macrophages (Appendix). Granulocyte-macrophage-colony stimulating factor, also derived from macrophages, can be secreted by multiple cellular sources.

Messenger RNA specific for granulocyte–macrophage-colony stimulating factor was found in tissue extracts of 5 of the 10 control patients, and tumor necrosis factor-α was found in 2 of these patients (Figure 1 and Table 2). None of the control tissues contained interleukin-1 β, and only one contained interleukin-6 transcripts. A different pattern emerged for cytokine transcripts detected in specimens from patients with giant cell arteritis. Although granulocyte–macrophage-colony stimulating factor and tumor necrosis factor-α transcripts were found slightly more frequently than in noninflamed tissues, synthesis of interleukin-1 β and interleukin-6 mRNA was a highly discriminatory finding. Interleukin-1 β mRNA was detected in 93% (14 of the 15 patients with giant cell arteritis) and interleukin-6 mRNA in 80% (12 of the 15 patients with giant cell arteritis) of positive tissue samples (Figure 1 and Table 2). Although interleukin-1 β, interleukin-6, and tumor necrosis factor-α are considered a core group of macrophage-derived cytokines, mRNA for the three cytokines was not invariably expressed together.

Transforming growth factor-β 1 is a polypeptide growth factor that possesses many biological functions and has a major role in inflammation and in connective tissue metabolism (Appendix Table) [9]. Polymerase chain reaction amplification with the transforming growth factor-β 1 primer set yielded a fragment of approximately 350 basepairs (Figure 1). Transforming growth factor-β 1 mRNA was exclusively expressed by arteritic tissue and was present in 87% (13 of 15) of all patients analyzed (Table 2).

Expression of T-Cell-Derived Cytokines

To ascertain whether T-cell specific mRNA can be found in the arterial specimens, polymerase chain reaction amplification with a primer set specific for the T-cell receptor α chain was applied. Identification of an amplification product specific for the T-cell receptor C α gene was limited to biopsy samples from patients with giant cell arteritis. The presence of T lymphocytes was shown in 67% of samples from patients with giant cell arteritis by this method (10 of 15 patients). In none of the artery specimens from control patients could mRNA of the T-cell receptor C α gene be detected.

T lymphocytes have been divided into functional categories (TH1 and TH2) by the profile of cytokines they produce (Appendix Table) [10]. To characterize the T cells accumulated in the inflammatory foci, cytokine mRNA was analyzed by polymerase chain reaction amplification with specific primer sets for interleukin-2, interferon-γ, interleukin-4, and interleukin-5. Amplification products are shown in Figure 2 for five patients with vasculitis and three controls. Results of these experiments for all samples analyzed are summarized in Table 2.

Expression of T-cell-derived cytokines was a characteristic finding in biopsy samples from patients with arteritic lesions. Specimens from control patients did not contain interleukin-2, interferon-γ, interleukin-4, or interleukin-5 mRNA. Interleukin-2 and interferon-γ mRNA were the most frequently identified T-cell products and were found in 67% (10 of 15) of all patients with giant cell arteritis. In 7 of 15 samples from patients with arteritis, interferon-γ mRNA production was associated with interleukin-2 mRNA synthesis. In tissue sections from three patients, a specific product for interleukin-2 was amplified in the absence of interferon-γ transcripts. In contrast to the frequent expression of interleukin-2 and interferon-γ mRNA, detection of interleukin-4 and interleukin-5 mRNA was a rare event. In only 1 of 15 tissue fragments did amplification with interleukin-4-specific primers result in a specific band. Two different specimens had evidence for interleukin-5 mRNA synthesis.

Although the polymerase chain reaction technique does not accurately quantify cytokine mRNA, comparison of the signal intensities suggested that interleukin-2 and interferon-γ mRNA concentrations were multifold higher than were levels of either interleukin-4 or interleukin-5 mRNA. Within the group of 15 patients with giant cell arteritis, the tissue extracts from 2 patients were negative for T-cell receptor C α as well as T-cell-derived cytokine mRNA. Preparations of both arteries contained specific sequences for transforming growth factor-β 1 and interleukin-1 β, suggesting that, in these sections, macrophages were activated in the absence of a detectable T-cell infiltrate.

Patients with Polymyalgia Rheumatica

Polymyalgia rheumatica was diagnosed in patients with aching and morning stiffness in the shoulder and pelvic girdle muscles, combined with systemic features and abnormal laboratory test results indicative of an acute phase reaction [7, 8]. Patients with polymyalgia rheumatica alone had no symptoms or signs of giant cell arteritis and had normal biopsy specimens from the temporal artery. Those with an arteritic lesion in the temporal artery were categorized as having giant cell arteritis with polymyalgia rheumatica. To examine whether patients with polymyalgia rheumatica who did not have evidence of giant cell arteritis have a subclinical inflammation in the arterial wall, tissue extracts from biopsy samples of patients with polymyalgia rheumatica alone were analyzed by polymerase chain reaction for cytokine-specific sequences. Seventeen tissue samples from 9 different patients were available for analysis. As noted, none of these biopsy specimens had visual evidence of an inflammatory infiltrate on histologic examination of serial tissue sections.

Messenger RNA transcripts of cytokines (interleukin-1 β, interleukin-2, and transforming growth factor-β 1) present in patients with giant cell arteritis but not in controls were frequently detected in patients with polymyalgia rheumatica (Table 3). Interleukin-1 β and transforming growth factor-β 1 transcripts were identified in 6 and 5 patients with polymyalgia rheumatica, respectively. Four patients showed evidence for in situ T-cell activation indicated by the presence of interleukin-2 mRNA. However, in contrast to patients with giant cell arteritis, patients with polymyalgia rheumatica synthesized an incomplete profile of T-cell-derived and macrophage-derived cytokines. Tissue samples from patients with polymyalgia rheumatica did not contain interferon-γ-specific sequences, and interleukin-6 was only found in one patient. Both of these cytokines were found in most of the patients with giant cell arteritis (Table 2).

To ascertain whether patients with either polymyalgia rheumatica or giant cell arteritis had typical patterns of tissue cytokines, the co-occurrence of selected cytokine transcripts was compared in the three study populations. The coproduction of interferon-γ mRNA with either transforming growth factor-β 1 or interleukin-1 β mRNA was an exclusive finding for patients with giant cell arteritis. Interleukin-1 β or transforming growth factor-β 1 specific sequences were expressed in 8 of 9 patients with polymyalgia rheumatica in the absence of interferon-γ. This pattern was infrequently detected in patients with giant cell arteritis and was not seen in the controls (Table 4). These data suggest that the patterns of T-cell-derived (interferon-γ) and macrophage-derived cytokines (interleukin-1 β and transforming growth factor-β 1) may distinguish the inflammatory response in patients with giant cell arteritis and polymyalgia rheumatica (Table 4).

Discussion

Our data show that activated macrophages and T lymphocytes producing interleukin-2 and interferon-γ characterize the inflammatory lesion in patients with giant cell arteritis. In patients with polymyalgia rheumatica, cytokine mRNA can be detected in temporal artery tissue specimens despite the lack of microscopic evidence of tissue infiltrating cells. Patients who have either polymyalgia rheumatica or giant cell arteritis all have in situ synthesis of the cytokines interleukin-1 β, interleukin-2, and transforming growth factor-β 1; interleukin-6 was only infrequently found, and interferon-γ was not found at all in the tissue samples from patients with polymyalgia rheumatica. The local secretion of interferon-γ may represent an important amplification loop leading to overt arteritis in patients with giant cell arteritis.

Inflammation and damage involving the walls of blood vessels in the vasculitic syndromes have been attributed to immunologic mechanisms. Although immune-complex–mediated mechanisms appear to be operative in some vasculitides, cellular immune mechanisms are more likely to be important in the initiation and perpetuation of lesions in patients with giant cell arteritis [11]. Infiltrating inflammatory cells may induce tissue damage through direct cellular cytotoxicity or through the secretion of cytokines [12-14]. Immunohistochemical studies have shown that CD4+ T cells are the dominant cell population in temporal arteries, whereas CD8+ T cells are less frequent, and B lymphocytes are essentially absent [4]. Although granulocytes are a typical finding in other vasculitides, they are apparently not involved in the arteritic lesions classic for patients with giant cell arteritis [3]. In 50% of patients with giant cell arteritis, intramural granulomatous inflammatory reactions can be identified. Granulomas consist of multinucleated giant cells, epithelioid cells, activated macrophages, and T lymphocytes.

Our study was designed to analyze the local production of cytokines that are important in the recruitment and regulation of inflammatory cells (Appendix Table). Understanding the role of different cytokines in the in vivo situation is likely to provide insights into cellular components and cell-cell interactions capable of driving the tissue-destructive inflammatory response and, thus, represents an indirect approach to define critical mechanisms in giant cell arteritis. Analysis of biopsy samples showed that a highly selected spectrum of lymphokines is produced in situ. The profile of T-cell-derived lymphokines identified the cells of origin as a selected subpopulation of T-helper cells (TH1-helper cells) and documented that a localized immune response is strongly biased toward a specific functional pathway.

Based on the pattern of lymphokines secreted, CD4+ T-helper cells have recently been subdivided into functional subsets, TH1 and TH2 cells [10, 15, 16]. Interleukin-2 and interferon-γ have been described as the marker lymphokines of TH1-helper cells, whereas the production of interleukin-4 and interleukin-5 has been assigned to a second T-helper cell pathway, the TH2 pathway. Data presented here indicate that giant cell arteritis is a TH1-driven disease. The co-occurrence of interleukin-2, interferon-γ, and interleukin-1 in patients with giant cell arteritis suggests a pathogenetic model involving activation of T cells by local antigen recognition and induction of an immune reaction dominated by TH1-helper cells and macrophages.

The finding of interleukin-2 mRNA strongly suggests that T cells recruited to the temporal artery wall were recently, and, therefore, probably locally activated. Interleukin-2 is an important T-cell growth and activation factor, and interleukin-2 mRNA is strictly controlled [17]. Perturbation of the T-cell receptor by antigen induces a rapid induction of interleukin-2 mRNA; its half-life is estimated at about one-half hour. The presence of interleukin-2 mRNA in preparations of inflamed arteries indicates recent perturbation of the T-cell receptor through antigen contact. Interleukin-2 mainly regulates the function of other lymphocytes, interferon-γ acts primarily on macrophages [14]. In our study, interferon-γ was the major T-cell-derived lymphokine besides interleukin-2. Interferon-γ is instrumental in macrophage activation. Interferon-γ might indeed represent the stimulus that activates macrophages to release monokines, including the proinflammatory cytokines interleukin-1, interleukin-6, and tumor necrosis factor-α.

Induction of a granulomatous reaction has been linked to the presence of foreign antigen that cannot be degraded [18]. Granuloma formation is characteristic for infections with numerous pathogenic microbes, including bacteria, fungi, and parasites, such as Mycobacterium tuberculosis, Mycobacterium leprae, Leishmania, Histoplasma capsulatum, and Schistosoma mansoni. Formation of granulomas appears to be a T-cell–dependentprocess that requires the action of monocyte-derived cytokines, in particular interleukin-1 and tumor necrosis factor-α [19, 20]. Granulomas induced by a Schistosoma mansoni egg contain high numbers of eosinophils, suggesting the action of a TH2-like helper cell releasing interleukin-5 [18]. In contrast, the granuloma formation in tuberculoid leprosy has been associated with the TH1-like pathway [15]. Formation of giant cells that arise from the fusion of monocytes is also a T-cell–dependentprocess. Interferon-γ and interleukin-4 have been found to induce cultured monocytes to form mononucleated giant cells [21-23]. These data suggest that TH1- and TH2-like cells can be involved in granuloma and giant cell formation. Our results indicate that the granulomatous reaction in patients with giant cell arteritis is associated with a localized immune response of exclusively the TH1 pathway. A TH1-like immunopathologic result in patients with giant cell arteritis could explain the usual absence of eosinophils in the inflammatory lesion and the absence of hypergammaglobulinemia and autoantibodies despite a chronic autodestructive immune response. Support of immunoglobulin production by B lymphocytes has been linked to the action of interleukin-4 and interleukin-6, both TH2-derived lymphokines [24].

Clinical experience has shown that it is often difficult to document the presence or absence of vasculitis in patients with polymyalgia rheumatica. We have recently shown that polymyalgia rheumatica and giant cell arteritis share multiple pathogenetic features in addition to similarities in the clinical presentation. Both diseases have in common an association with selected HLA-DRB1 alleles, in particular the HLA-DRB1*04 alleles [25, 26]. Patients with polymyalgia rheumatica and giant cell arteritis have highly elevated levels of serum interleukin-6 [27]. After initiating corticosteroid therapy, interleukin-6 concentrations abruptly return to normal and remain suppressed as long as steroid therapy is continued. In both diseases, peripheral monocytes are activated and produce interleukin-1 β and interleukin-6 [28]. For all of these findings, polymyalgia rheumatica and giant cell arteritis are indistinguishable, suggesting that none of these factors is important in the progression of polymyalgia rheumatica to giant cell arteritis.

Our current results show that cytokine mRNA can be detected in temporal arteries from patients with polymyalgia rheumatica, indicating that these patients have subclinical vascular involvement. Polymerase chain reaction amplification is a highly sensitive method, and amplification of cytokine-specific transcripts allows the identification of as few as 10 activated cells (data not shown). The data described here do not allow any firm conclusions about the relative amounts of cytokine mRNA present but only show that a low number of activated cells must infiltrate the vessel wall in patients with polymyalgia rheumatica to produce the cytokines that were detected. Thus, considering the usual histologic findings, polymyalgia rheumatica and giant cell arteritis are likely to differ in the amount of cytokines produced in situ. In addition to these quantitative differences, patients with giant cell arteritis have a different cytokine pattern that might be crucial for the development of frank vasculitis.

The inflammatory response in giant cell arteritis consists of activated macrophages producing interleukin-1 β, interleukin-6, and transforming growth factor-β 1 and consists of activated T lymphocytes synthesizing interleukin-2 and interferon-γ mRNA. Using in situ hybridization and immunohistochemical techniques, we have shown that interleukin-1 β, interleukin-6, and transforming growth factor-β 1 are produced by CD68+ macrophages [28]. The inflammatory response in giant cell arteritis can thus be described as a T-cell driven response selectively involving TH1 cells that secrete interleukin-2 and interferon-γ. In tissue samples from patients with polymyalgia rheumatica, we detected macrophage-derived cytokines, in particular transforming growth factor-β 1 and interleukin-1 β. The finding of the low frequency of interleukin-6 in the tissue of patients with polymyalgia rheumatica may be explained by the low number of tissue-infiltrating macrophages and the sensitivity of the polymerase chain reaction. Under the condition of a limiting number of activated macrophages, interleukin-1 β-specific sequences are easier to detect by polymerase chain reaction than are interleukin-6-specific sequences (data not shown).

In contrast to the macrophage activation, the T-cell response appears to be quantitatively, but also qualitatively, different in the two diseases. Although patients with giant cell arteritis and those with polymyalgia did not differ in their in situ production of interleukin-2, the presence of interferon-γ sequences was significantly different (P = 0.001, Table 3). Interferon-γ mRNA is more easily detected in activated T cells than is interleukin-2 mRNA (data not shown), indicating that the absence of interferon-γ in the tissue of patients with polymyalgia rheumatica is of biological significance and is not a result of insufficient sensitivity. Interferon-γ is crucial for macrophage activation and for granuloma formation. Thus, the production of interferon-γ may be essential for the development of vasculitis. Patients with polymyalgia rheumatica may lack an important amplification mechanism in their local immune response in the vasculitic lesion. In this model, the finding that not all tissues from patients with giant cell arteritis had interferon-γ transcripts would not indicate true heterogeneity of the disease but would be caused by the patchy nature of the inflammation. We analyzed the cytokine profile in 5 to 10 consecutive sections of the artery, and we may have missed areas with a full-blown vasculitic picture in some of the patients.

Clinical studies have shown that only 15% of patients with polymyalgia rheumatica eventually develop vasculitic complications. Many patients with polymyalgia rheumatica may not be able to generate a local TH1 response in the arterial wall and may therefore be protected from the development of the vasculitic complications. Genetic factors and the microenvironment may contribute to whether the patients develop a local immune response characterized by few interleukin-2-secreting T cells or by many tissue-infiltrating T cells cosecreting interleukin-2 and interferon-γ. Lack of differentiation of interleukin-2-producing T cells to typical TH1 cells, caused by differences in antigen concentration triggering the immune response, represents an alternative possibility. In such a model, we predict that the triggering antigen is not of endogenous origin.

Management of polymyalgia rheumatica requires low-dose steroid therapy. High doses of steroids are given to patients with giant cell arteritis to prevent the consequences of arteritis. Currently, we are unable to identify patients with polymyalgia rheumatica threatened by progression of the disease. Potential side effects from long-term steroid therapy require the restriction of high-dose therapy to patients who need it. Identifying a pathogenetic mechanism that distinguishes polymyalgia rheumatica and giant cell arteritis not only raises the possibility of developing treatment approaches to prevent progression of polymyalgia rheumatica to giant cell arteritis but also of targeting a crucial biological pathway in the inflammatory response. Selective suppression of TH1-derived cytokines may represent new avenues of therapy. The production of TH1-derived cytokines, including interferon-γ, depends on the intracellular cyclic adenosine monophosphate concentration [29, 30]. Concentrations of cyclic adenosine monophosphate sufficient to suppress the production of interferon-γ are within the physiologic range and can be achieved by the administration of different pharmacological agents [31]. Such agents are therefore prime candidates to be explored for a potential benefit in the management of patients with giant cell arteritis.