Interleukin-11 therapy selectively downregulates type I cytokine proinflammatory pathways in psoriasis lesions (original) (raw)
Clinical and histopathological response to rhIL-11 treatment in psoriatic patients. Twelve patients with psoriasis vulgaris were treated with daily subcutaneous injections of rhIL-11 for 8 weeks. Eleven of 12 patients had reductions in clinical disease severity ranging from 20 to 80%, as judged by the PASI score. Mean reductions in disease severity were statistically significant at all points analyzed from weeks 2 through 8 following rhIL-11 treatment. Figure 1 illustrates 2 patients with marked reductions in erythema, scaling, and induration in psoriatic skin lesions after 8 weeks of treatment with rhIL-11. A 50% drop in the PASI score from 22.6 to 11.7 was observed in patient A following rhIL-11 treatment. Similarly, an 80% reduction in the PASI score from 19.8 to 3.5 was observed in patient B.
Photographs of psoriatic plaques undergoing resolution in 2 patients (A and B) before and after 8 weeks of treatment with rhIL-11.
To rigorously quantify disease improvement, serial biopsies from an indicator plaque were used to measure disease-related pathology in each patient during IL-11 administration. Representative histopathology of a responding patient is shown in Figure 2. Based on this analysis (Figure 2 and Table 1), we found that 7/12 patients had remarkable disease reduction as defined by consistent decreases in (a) epidermal hyperplasia (epidermal thickness, number of proliferating [Ki67+] keratinocytes, and expression of K16 in suprabasal keratinocytes); (b) number of T cells in skin lesions; and (c) ICAM-1 production by epidermal keratinocytes. The 7 patients with consistent improvements in each of these histologic parameters at the end of 8 weeks of rhIL-11 treatment were listed as “responders” in Table 1. In contrast, 5 patients had smaller or less-consistent improvements in disease-defining histopathology, and these patients were labeled as “nonresponders” in Table 1.
Six-millimeter punch biopsies were obtained from lesional and nonlesional skin of a patient before treatment with rhIL-11 (pretreatment) and at weeks 1, 4, and 8 following daily treatment with 2.5 mg/kg of rhIL-11. Biopsies were equally divided for immunohistochemical analysis and RNA preparation. Frozen biopsies from pretreatment, week 4, and week 8 were sectioned and stained with antibodies to Ki67+, K16, CD3, CD8, and ICAM-1. Enlargements of ICAM-1 staining (boxes) show elimination of its production by epidermal keratinocytes during IL-11 administration (large arrows).
Histology measures in responding and nonresponding patients
Responding patients showed marked improvement in several measures of pathological epidermal hyperplasia (epidermal acanthosis, keratinocyte proliferation, and synthesis of keratin) (17). Epidermal thickness was reduced on average by 40%, which represents about a 60% decrease in pathological acanthosis. Keratinocyte proliferation was reduced by a mean of 66% at 8 weeks of treatment. Importantly, synthesis of K16, which is produced only by “regenerative” or hyperplastic keratinocytes, was eliminated in suprabasal keratinocytes in all responding patients (Figure 2 and Table 1). Hence, homeostatic epidermal growth was restored in all 7 responding patients, and quantitative reductions in epidermal thickness and keratinocyte proliferation were statistically significant at 8 weeks (Table 1). Proliferative decreases in epidermal keratinocytes were accompanied by marked reductions in T lymphocytes infiltrating the epidermis by 4 weeks of treatment. Mean reductions in intraepidermal T cells and total lesional T cells averaged 50 to 75% at weeks 4–8 of treatment and were statistically significant (P < 0.01) by 4 weeks of treatment (Table 1). Synthesis of ICAM-1 by epidermal keratinocytes (Figure 2) was eliminated by 8 weeks of treatment in all responders, and, importantly, reduced expression of this adhesion molecule was detected as early as 1 week of treatment with IL-11, at a time when T-cell numbers were not reduced in lesional tissue (Table 1). In contrast, no statistically significant improvements in epidermal growth parameters or T-cell infiltration of lesions were measured in nonresponders (Table 1).
Expression patterns of disease-related genes in lesional and uninvolved skin. The designation of responders and nonresponders was created principally to determine how alterations in expression of a variety of inflammation-associated genes that could potentially be regulated by IL-11 would relate to major changes in cellular features of tissue inflammation or clinical disease activity. For example, reduced production of ICAM-1 by epidermal keratinocytes at 1 week after starting rhIL-11 treatment, but no major changes in the number of T cells in lesional tissue (Table 1), suggests that T cells in skin lesions produce less IFN-γ or that keratinocyte responsiveness to this cytokine is altered (31). Accordingly, a comprehensive survey of expression patterns of genes involved in the pathogenesis of this disease and the effect of rhIL-11 on this expression pattern was undertaken using a new quantitative RT-PCR technique.
Punch biopsies of psoriatic plaques and noninvolved skin from the 6 initial patients were obtained and analyzed for gene-expression differences between lesional and nonlesional skin by RT-PCR. Quantitative mRNA levels were determined using an ABI 7700 sequence detection system and all values were normalized to a housekeeping gene, human acidic ribosomal protein (HARP) mRNA, that was coamplified during each PCR run. Based on initial experiments evaluating the reproducibility of the sequence detection system as well as patient-to-patient variability, changes in mRNA ex-pression levels of 2-fold or greater for lesional versus nonlesional skin were deemed to be significant (W. Trepicchio, unpublished observations). Where possible, changes in gene expression were compared with changes in immunohistochemical markers of protein expression (taken from adjacent skin) to further confirm expression data.
A total of 32 potentially disease-associated genes, as well as 3 housekeeping genes, were analyzed. Twenty-four of these 32 genes were found to provide consistent results across all patients and were considered for further analysis in biopsies from rhIL-11–treated patients. Expression levels of many genes known to play or thought to play a role in the disease process, such as proinflammatory markers, leukocyte surface markers, and keratinocyte hyperplasia markers displayed consistently higher expression levels in lesional psoriatic tissue compared with uninvolved (nonlesional) skin (see Figure 3). For example, 2-fold to greater than 10-fold elevations in mRNA levels of the p40 subunit of IL-12, IFN-γ, TNF-α, IL-1β, IL-8, and iNOS were observed in psoriatic lesions compared with nonlesional skin (Figure 3a).
RNA was prepared from nonlesional and lesional skin of 6 untreated psoriasis patients. The mRNA was amplified for the indicated genes by quantitative RT-PCR. Levels of RNA were normalized to HARP. Average expression levels of nonlesional and lesional skin for all 6 patients are presented ± SEM. *Statistically significant differences (P < 0.05) between nonlesional and lesional skin; filled box, lesion; shaded box, nonlesion.
A number of leukocyte markers such as CD4, CD8, CD25, CD80, and IL-12Rβ2 mRNA levels were also elevated from 2- to 5-fold in psoriatic tissue compared with uninvolved skin (Figure 3b). The elevated level of these transcripts correlates with increased numbers of activated CD3+, CD4+, and CD8+ T cells in psoriatic lesions as measured by histologic analysis (Table 1 and Figure 2). In addition, elevated expression levels of genes associated with keratinocyte hyperplasia (K16 and KGF) were observed in lesional skin versus nonlesional skin (Figure 3c). Increased expression of these genes is in agreement with established differences in protein expression of K16 and KGF in psoriatic skin lesions (14, 15).
In contrast to inflammatory mediators, levels of a number of other genetic markers were unchanged. For example, mRNA levels for the Th2 cytokine, IL-4, was not significantly different between lesional and nonlesional skin (Figure 3d). Coupled with the previously demonstrated 10-fold elevation in IFN-γ production, these results indicate that the ratio of IL-4 to IFN-γ mRNA is altered in lesional skin and suggests a dominant Th1 T-cell response is found in the psoriatic tissue. Anti-inflammatory cytokine IL-10, IL-11, and IL-6 mRNA was found at comparable levels in uninvolved and lesional skin (Figure 3d). In addition, mRNA levels for IκB-β, the inhibitor of the proinflammatory transcriptional inducer, NF-κB, were not changed between nonlesional and lesional skin (Figure 3d). An NK cell marker, CD56; a Langerhans/dendritic cell marker, CD1a; and a T-cell costimulatory molecule, CD86, were present in all samples but were not differentially expressed in disease tissue (data not shown).
Several genes did not demonstrate consistent changes in all patients examined and were not evaluated further. A number of T-cell signal transducers, such as NFAT-1, STAT-1, and STAT-4; cell surface markers, such as CD40, CD122, and ICAM-1; the cell cycle and apoptosis regulator, p53; and the inflammatory cytokine, IL-18, were not differentially expressed in a consistent pattern across all patients (data not shown). These results could be due to the small patient sample size or a lack of transcriptional regulation of these genes. For example, the STAT genes are primarily activated by posttranslational phosphorylation events (reviewed in refs. 32, 33).
Pharmacogenomic analysis of skin lesions of rhIL-11 treated patients. From the panel of disease-related genes identified above, we sought to identify genetic markers of psoriatic pathology that change following rhIL-11 treatment. These markers would then be correlated with patient responsiveness as determined by histopathological criteria. Total RNA was prepared from one-half of the biopsy that was taken for immunohistochemical study and analyzed by quantitative RT-PCR. Comparisons were made in the levels of mRNA between normal skin and psoriatic lesions before and during treatment with rhIL-11. Levels of expression of 14 disease-associated genes were analyzed following patient treatment with rhIL-11 and are presented below. Consistent effects of rhIL-11 on expression levels of another 4 genes, CD86, CD56, CD1a, and KGF, were not observed and data are not presented.
Data showing key disease-related genes from a representative responding patient are presented in Figure 4. The levels of TNF-α, IFN-γ, IL-12 p40, IL-8, and iNOS mRNA in the lesional tissue decreased as early as 1 week following rhIL-11 treatment. This observed effect of rhIL-11 on gene expression preceded any significant clinical changes. At the end of 8 weeks of treatment, these inflammatory markers were significantly reduced compared with pretreatment levels (Figure 4). CD8 mRNA levels also decreased following rhIL-11 treatment, and this result correlated with decreased CD8+ cells detected by immunohistochemistry. Levels of IL-4 mRNA increased slightly after rhIL-11 treatment (Figure 4). As a percentage of total T cells present in the lesion (as measured by CD3 staining and CD4 or CD8 expression) IL-4 mRNA levels actually increased following rhIL-11 treatment. Finally, levels of K16 mRNA, which were elevated in lesional versus uninvolved skin, were significantly reduced following rhIL-11 treatment. This change in K16 levels also correlated with changes in immunohistochemical staining following rhIL-11 treatment (Figure 2). These changes in gene expression also correlated with an improvement in the lesion severity score from 8 to 0 in this patient over the 8-week treatment period. In contrast to the responding patients, 5 patients designated as nonresponders by histologic analysis failed to show consistent and sustained reductions in proinflammatory gene transcripts (data summarized in Figure 5). This lack of a response at the genetic level also correlated with the absence of histologic improvement (Table 1).
Six-millimeter punch biopsies were obtained from lesional and nonlesional skin of a responding patient before treatment with rhIL-11 (pretreatment) and at weeks 1, 4, and 8 following daily treatment with 2.5 mg/kg of rhIL-11. Biopsies were equally divided for immunohistochemical analysis and RNA preparation. RNA was prepared from nonlesional skin and lesional skin from pretreatment, week-1, week-4, and week-8 biopsies. The mRNA for IL-12 p40, IFN-γ, TNF-α, iNOS, CD8, IL-8, IL-4, and K16 were amplified by quantitative RT-PCR. Levels of mRNA were normalized to HARP to control for variations in starting RNA amounts.
RNA was prepared from nonlesional skin biopsies and lesional skin biopsies of 12 patients before treatment (pretreatment) and at week 1, week 4, and week 8 following daily treatment with 2.5 or 5.0 mg/kg of rhIL-11. Quantitative RT-PCR was performed on individual samples for the indicated genes. Gene expression levels were normalized to HARP. Levels of gene expression observed in the nonlesional skin of each patient were arbitrarily set to equal 1, and the fold change in expression in lesional skin before and after treatment with rhIL-11 over nonlesional skin was calculated. Average fold change for rhIL-11–responding (n = 7) and rhIL-11–nonresponding (n = 5) patients was calculated. Data are presented as the average fold change over nonlesional skin ± SEM. *Statistically significant differences between pretreatment and rhIL-11 treatment (P < 0.05); #statistically significant differences between responder and nonresponder pretreatment lesions (P < 0.05).
A summary of gene-expression profiles for key genes in all responders and nonresponders before and during rhIL-11 treatment is presented in Figure 5 and Table 2. Data are represented as the average fold change of lesional skin over nonlesional skin so that a comparison across all patients can be made. Selective downregulation of iNOS, IFN-γ, IL-8, IL-12 p40, CD8, and K16 genes occurred in 7 responding patients during rhIL-11 treatment (Figure 5). Importantly, reduced production of mRNA for proinflammatory cytokines such as IFN-γ, IL-12 p40, and TNF-α was detected as early as 1 week after starting IL-11 treatment, a time point when no significant overall reductions in T lymphocytes in skin lesions had occurred (compare Table 1 with Figure 5 and Table 2). Statistically significant reductions in iNOS, IFN-γ, IL-8, IL-12 p40, K16, CD80, and IL-12Rβ2 mRNAs were measured after 4–8 weeks of treatment with rhIL-11. In contrast, levels of IL-4 and IL-10 mRNA remained constant over the course of treatment in responding patients. Of particular interest, levels of IL-11 mRNA actually increased in responding patients (Table 2).
These data suggest that rhIL-11 decreases production of proinflammatory molecules by T cells and other leukocytes soon after initiation of treatment, whereas continued administration of rhIL-11 for several weeks also decreases T-cell trafficking into psoriatic skin lesions. Progressive reductions in epidermal hyperplasia, as well as decreased synthesis of ICAM-1 and K16 by keratinocytes, during 8 weeks of treatment with IL-11 are probably best explained by the combination of reduced levels of proinflammatory cytokines in skin and the presence of fewer T cells, especially those within the epidermis. After 4–8 weeks of treatment, there is clearly an altered balance of type 1 versus type 2 T cells within skin lesions, as judged by a change in the IFN-γ/IL-4 ratio. Although this change could reflect outward trafficking of type 1 T cells, we noted a decrease in peripheral T cells producing IFN-γ and an increase in cells producing IL-4 by intracellular staining in a number of patients during rhIL-11 treatment (data not shown). Hence, cytokine mRNA expression profiles may also indicate an effect of rhIL-11 on differentiation of type 1 versus type 2 T cells.
To compare the effects of rhIL-11 to another immunomodulatory drug, gene-expression analysis of psoriatic lesions before and during cyclosporin A treatment was also performed. Lesions from 3 cyclosporin A–treated patients were analyzed before and at 1 and 4 weeks of therapy. On the basis of clinical and histological analysis, 2 patients were considered responding to therapy and 1 was considered nonresponding (data not shown). Similar to rhIL-11 treatment, patients responding to cyclosporin A treatment demonstrated significant reduction in levels of IFN-γ, iNOS, IL-8, and K16, stable expression of CD8 and elevated levels of IL-4 mRNA in resolving lesions compared with pretreatment levels (Table 3). These changes were not observed in the nonresponding patient (Table 3).
A comparison of gene-expression levels in lesional tissue of responding versus nonresponding populations before rhIL-11 treatment allows a retrospective analysis of differences in the patient population that may explain differential responsiveness to therapy. For example, differential patient responsiveness to rhIL-11 could be due to low-level expression of the IL-11 receptor measured in nonresponding patients On average, lower levels of IL-11R α chain mRNA was observed in nonresponding patients, but this difference was not statistically significant. In addition, lower iNOS and IL-8 mRNA levels and higher IL-12 p40 and K16 levels are observed at baseline in the lesional skin of responding patients versus the lesional skin of nonresponding patients, but only K16 and IL-8 differences were statistically significant. Further analysis of this patient population using broader gene expression arrays is in progress to identify patterns of expression correlating with responsiveness to rhIL-11.