Troglitazone Improves Psoriasis and Normalizes Models of Proliferative Skin Disease: Title and subTitle BreakLigands for Peroxisome Proliferator-Activated Receptor-γ Inhibit Keratinocyte Proliferation FREE

PSORIASIS IS a common and often debilitating skin disorder that affects 1% to 2% of the US population.¹ Disordered differentiation and hyperproliferation of keratinocytes with inflammation are the hallmarks of psoriasis; however, the pathophysiological features of this disorder are complex and its etiology is unknown.^{2 - 3} A variety of therapies are available for psoriasis.⁴ Nevertheless, because of problems with adverse effects and variability in clinical response, there is a desire for new treatments.

Nuclear hormone receptors have proved to be useful targets in the development of antipsoriatic drugs.⁵ Although ligands for such receptors (including glucocorticoid, retinoid, and vitamin D receptors) are commonly used as antipsoriatic agents, each of these treatments has limitations, and new therapeutic options in psoriasis are needed.⁴

The thiazolidinediones are a novel class of insulin-sensitizing agents used for the treatment of type 2 diabetes mellitus.^{6 - 7} Thiazolidinediones act as ligands for the peroxisome proliferator-activated receptor-γ (PPARγ), a member of the nuclear hormone receptor superfamily that includes the retinoic acid receptor and the vitamin D receptor. Ligand activation of PPARγ by thiazolidinediones can inhibit proliferation and promote differentiation in a variety of malignant and nonmalignant tissues,^{8 - 13} although the effects on skin are unknown.

In view of the antiproliferative effects of ligands for PPARγ, we gave troglitazone, a thiazolidinedione marketed in the United States, orally to patients with psoriasis. With this description of 2 normoglycemic patients, we now have treated 5 patients with troglitazone in an open fashion. We also report our investigations into the effects of ligands for PPARγ in experimental cellular and animal models of psoriasis.

Five patients with moderate to severe psoriasis provided informed consent for the new use of troglitazone for the treatment of psoriasis. Ten different untreated patients with psoriasis and 10 control subjects provided skin samples after giving informed consent. The University of Michigan Health System institutional review board approved the procedures. We used skin from any individual in only one of the studies described herein.

Human epidermal keratinocytes (Clonetics Inc, San Diego, Calif) were grown in monolayer cultures.¹⁴ Peroxisome proliferator-activated receptor-γ messenger RNA (mRNA) was detected in human keratinocytes by reverse transcriptase–polymerase chain reaction (RT-PCR) analysis; PPARγ protein was detected by immunoblotting and immunocytochemical studies. For RT-PCR analysis, RNA was prepared from quiescent human keratinocytes and reverse transcribed to generate complementary DNA, which was amplified by PCR to detect message for PPARγ1 and PPARγ2. The primers used for PCR amplification of PPARγ1 were 5-′CTC GAG GAC ACC GGA GAG-3′ (upstream) and 5′-GTC ATT TCT GCG GCC ACG-3′ (downstream). These primers specifically amplify a 50–base pair fragment of the 5′ untranslated region of PPARγ1 (Genbank accession number X90563). The primers used to amplify PPARγ2 were 5′-GGT GAA ACT CTG GGA GAT TCT-3′ (upstream) and 5′-TGT AAT CTG CAA CCA CTG GAT-3′ (downstream).These primers amplify a 277–base pair fragment from the N-terminal region of PPARγ2 (Genbank accession number U63415). The PCR products were analyzed by 2.5% agarose gel electrophoresis.

Immunoblotting was performed after sodium dodecyl sulfate–polyacrylamide gel electrophoresis separation of keratinocyte protein extracts using a rabbit anti-PPARγ antibody and an enhanced chemiluminescence detection system (Amersham Corp, Arlington Heights, Ill). Immunofluorescence cytochemical analysis was performed on methanol-fixed and permeabilized keratinocytes using rabbit anti-PPARγ antibody and fluorescein isothiocyanate–labeled goat antirabbit IgG.

Studies were performed with keratinocytes (obtained from Clonetics Inc or prepared from biopsy samples of skin of 8 controls and lesional skin of 5 patients with psoriasis) maintained in either keratinocyte basal medium (KBM) or keratinocyte growth medium (HyClone Laboratories, Logan, Utah). Each is a low Ca⁺⁺ (0.15 mmol/L) modification of MCDB-153 medium; KBM contains no serum or exogenous growth factors, and keratinocyte growth medium is supplemented with several growth factors, including human recombinant epidermal growth factor (0.1 ng/mL), insulin (2.5 µg/mL), and pituitary extract (2% vol/vol). For proliferation assays, cells were plated in 24-well dishes at 2 × 10⁴ cells per well. Twenty-four hours after attachment, troglitazone (Parke-Davis Pharmaceuticals, Ann Arbor, Mich), ciglitazone (Upjohn Co, Kalamazoo, Mich), 15-deoxy-Δ-12,14-prostaglandin J₂ (Cayman Chemical Co, Ann Arbor), or dimethyl sulfoxide vehicle was added at indicated concentrations. Control and treated keratinocytes were incubated for 3 to 8 days and then counted. Cell numbers were determined in 6 replicates of each experimental condition using the MTS assay 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-H-tetrazolium as described in the CellTiter 96 Aqueous Non-Radioactive Cell Proliferation Assay (Promega Corp, Madison, Wis) or the neutral red dye assay. The effect of thiazolidinediones on the growth of human dermal fibroblasts in KBM supplemented with calcium, 1.4 mmol/L, was assessed using fibroblasts prepared from biopsy specimens of skin of 5 controls.

Under serum-free conditions, normal and psoriatic human skin can be maintained in organ culture for days and abnormal histological features of psoriasis persist.^{15 - 18} Pieces of skin from 2 controls and from psoriasis lesions from 2 patients were placed into wells of a 96-well dish containing 250 µL of culture medium. The culture medium consisted of KBM supplemented with calcium chloride to a final Ca⁺⁺ concentration of 1.4 mmol/L. The tissue pieces were incubated at 37°C in 5% carbon dioxide and 95% air for 8 days with fresh culture medium as previously described.^{15 - 16 ,18} Various concentrations of troglitazone or dimethyl sulfoxide vehicle alone were added at days 2, 4, and 6. At the end of the incubation period, tissue pieces were fixed in 10% buffered formalin and embedded in paraffin. Sectioned tissues were stained with hematoxylin-eosin and evaluated by light microscopy.

Human psoriasis skin engrafted onto severe combined immunodeficient (SCID) mice maintains its histological characteristics.¹⁹ Lesional psoriasis skin from 3 patients was obtained and divided into portions. One portion from each patient was reserved; the remaining tissue was divided and transplanted onto CB17 SCID mice as previously described.¹⁹ Each SCID mouse received 2 pieces of psoriatic skin from one of the human patients. The transplanted tissue was allowed to heal for 4 weeks, after which 11 animals (with 21 pieces of psoriasis; 1 piece did not survive) were fed once daily for 6 weeks with 200 µg of troglitazone dissolved in 300 µL of sugar-free gelatin dessert; 3 animals (with 6 pieces of psoriasis) fed plain gelatin dessert served as controls. At the end of treatment, the human skin pieces were removed, fixed in 10% buffered formalin, embedded in paraffin, sectioned, and stained with hematoxylin-eosin. Under light microscopy, we traced the epidermal boundaries and determined the area of the epidermis (in square micrometers) across 3 nonoverlapping regions (NIH Image software, version 1.61; National Institutes of Health, Bethesda, Md). The average of the 3 measurements is an index of epidermal hyperplasia of the psoriatic epidermis for each piece of skin.

We compared counts of keratinocytes in culture at all concentrations for each experimental condition with counts of keratinocytes maintained under similar conditions (but without experimental agents added) by analysis of variance; we calculated the average of the epidermal hyperplasia indices for each patient for each "treatment" assignment in SCID mice (nontransplanted skin, transplanted skin on untreated mice, and transplanted skin on treated mice) and compared the means across treatment assignments by repeated-measures analysis of variance with the Tukey studentized range for pairwise comparisons (SigmaStat; SSPS Inc, Chicago, Ill). Differences were considered statistically significant at P<.05; data are given as mean±SE.

A 37-year-old nondiabetic man who had chronic, generalized plaque psoriasis since age 14 years presented with 80% of his body surface involved with psoriasis (Figure 1). He began taking oral troglitazone (400 mg once daily for 5 weeks and 600 mg daily thereafter) without any other form of psoriasis therapy. After 5 weeks of troglitazone therapy, the patient showed a marked improvement in all lesions, and by 12 weeks his psoriasis was in nearly complete remission, with involvement of less than 5% of his body surface. This response has now been sustained for longer than 5 months with therapy (Figure 1). There were no adverse effects.

A 42-year-old nondiabetic man had chronic lesions of psoriasis over 30% of his body surface. Five months before presentation he was unresponsive to topical and systemic treatment with glucocorticoids. Oral administration of troglitazone, 200 mg/d for 2 weeks followed by 400 mg/d for 10 weeks, resulted in remission of psoriasis (Figure 2). Troglitazone use was continued for 12 additional weeks at 200 mg/d. Eight weeks after stopping therapy his psoriasis was still in remission. There were no adverse effects.

Patients with psoriasis and concomitant diabetes mellitus had improvement of both conditions under therapy with troglitazone, 400 to 600 mg/d; these patients have been described previously.²⁰

Immunofluorescence analysis of human keratinocytes treated with anti-PPARγ antibody revealed an intense signal in the nucleus and perinuclear region consistent with the expression of an intracellular receptor. Reverse transcriptase–polymerase chain reaction analysis demonstrated the presence of mRNA for both PPARγ1 and PPARγ2 isoforms in human keratinocytes; Western blot analysis of protein extracts using an antibody that reacts with both PPARγ isoforms revealed a distinct band corresponding to a protein of the expected size for PPARγ of 62 kd.

Use of the PPARγ ligand troglitazone inhibited in a dose-dependent manner the growth of normal and psoriatic keratinocytes maintained in serum-free media (P<.001 for each) (Figure 3, A). In normal keratinocytes maintained in supplemented growth media, higher concentrations of troglitazone were required for the antiproliferative effects (P<.001) (Figure 3, B). In all cases, inhibition of keratinocyte proliferation was observed with concentrations of troglitazone that can be attained in plasma during administration of standard oral doses of the drug for the treatment of type 2 diabetes mellitus.²¹ Use of ciglitazone, another thiazolidinedione, also inhibited proliferation of normal and psoriatic keratinocytes in a dose-dependent and saturable fashion (data not shown). In supplemented media, addition of the prostaglandin metabolite 15-deoxy-Δ-12,14-prostaglandin J₂, a natural ligand for PPARγ,^{22 - 23} inhibited proliferation of cultured keratinocytes in a dose-dependent fashion (P<.001) (Figure 3, C). Thus, the inhibition of proliferation in keratinocytes likely represents a general effect of ligands for PPARγ and is not unique to thiazolidinediones. However, troglitazone treatment did not affect the growth of human dermal fibroblasts in serum-free media at any of the concentrations tested (P=.97) (Figure 3, D).

After incubation for 8 days, the histological appearance of cultured normal skin resembled that of freshly biopsied skin, and lesional psoriatic skin in culture continued to express typical abnormal histological features (as seen in previous studies^{16 ,18} ). The abnormal phenotype of the psoriatic skin was qualitatively nearly normal in all organ cultures treated with 2.2 µmol of troglitazone; psoriasis tissue treated with 1.1 µmol of troglitazone showed similar changes, whereas tissue treated with 0.2 µmol of troglitazone did not.

The typical histological appearance of psoriasis was maintained in lesional psoriatic skin from 3 patients transplanted onto untreated control SCID mice (as validated previously¹⁹ ). In lesional psoriatic skin transplanted onto SCID mice treated with troglitazone orally, overall skin thickness was reduced, a more normal pattern of epidermal differentiation occurred, the granular layer (which is decreased or lacking in the untreated psoriatic skin) was present in much of the epidermis, and the inflammatory response was reduced (Figure 4). The index of epidermal hyperplasia was 70±6 × 10⁴ µm² for psoriatic skin not transplanted onto SCID mice, 74±2 × 10⁴ µm² for psoriatic skin transplanted onto untreated SCID mice (P>.05 compared with nontransplanted skin), and 49±2 × 10⁴ µm² for psoriatic skin transplanted onto troglitazone-treated SCID mice (P<.05 compared with nontransplanted and transplanted skin on untreated mice).

Ligands for nuclear hormone receptors expressed in skin (eg, retinoic acid, vitamin D, and glucocorticoid receptors) have proved to be of considerable value in the treatment of dermatologic diseases, including psoriasis.^{4 ,24} In the present studies, we found that thiazolidinedione ligands for the nuclear hormone receptor PPARγ might offer a new therapeutic approach to the management of psoriasis.

Peroxisome proliferator-activated receptor-γ was originally believed to be expressed in just a few tissues, including fat. However, recent studies^{25 - 26} in humans have demonstrated the presence of PPARγ in a variety of tissues. In a recent study,²⁷ PPARγ was detected in keratinocytes by gel mobility shift and RT-PCR assays, and herein we demonstrate that human keratinocytes express mRNA and protein for the nuclear hormone receptor PPARγ.

In the present studies, we found that structurally different ligands for PPARγ inhibit human keratinocyte proliferation, a characteristic feature of psoriasis. We also found that troglitazone, a PPARγ ligand used to treat type 2 diabetes mellitus, ameliorates the abnormal histological phenotype of human psoriatic skin in culture and in an animal model in vivo.

In addition, we demonstrated that oral administration of troglitazone improved psoriasis in a few patients. Although troglitazone treatment has been associated with rare cases of hepatic failure, newer PPARγ ligands, such as pioglitazone hydrochloride and rosiglitazone, have not been reported to induce this adverse effect and may be better agents to try as psoriasis therapies.

The occurrence of psoriasis in patients with diabetes mellitus is not uncommon, and these conditions may coexist more often than predicted from the prevalence rates of either disorder alone.²⁸ Three patients with type 2 diabetes mellitus and psoriasis had clinically significant improvement in their psoriatic lesions when we treated them with troglitazone.²⁰

We have now observed antipsoriatic effects of troglitazone in 2 normoglycemic patients with moderate to severe chronic psoriasis. It would be unusual for stable, chronic plaque psoriasis to undergo such complete and sustained remissions as we observed in these patients either spontaneously or because of a placebo effect, suggesting that a troglitazone treatment–induced effect occurred.

The observation that several different ligands for PPARγ inhibit proliferation of keratinocytes, together with the important role of keratinocyte hyperproliferation in the pathogenesis of psoriasis, suggests that the antipsoriatic effects of troglitazone might be mediated at least in part by the antiproliferative effects of PPARγ activation. Certain fatty acids, such as linoleic acid, are weak activators of PPARγ and have also been shown to decrease the replication rate of cultured keratinocytes.^{29 - 30} Moreover, in some patients with psoriasis, dietary supplements of some fatty acids have been associated with lesion improvement.³¹

Thiazolidinediones may exert a host of other actions such as blockade of calcium channels and mitogen-activated protein kinase signaling pathways that could also affect cell proliferation and the pathogenesis of psoriasis independent of effects on PPARγ activation.^{12 ,32 - 33} However, others have provided compelling evidence that, at least in some cell lines, antiproliferative or apoptotic effects of thiazolidinediones are indeed mediated via PPARγ.^{34 - 36} The inhibition of proliferation of keratinocytes is unlikely to represent simply a toxic effect of the drugs because the growth inhibitory effects are reversible on removal of the drugs from the culture medium (data not shown). In a variety of other normal cell lines studied under conditions similar to those used in keratinocytes, we and others^{11 - 13 ,34} showed that the antiproliferative effects of these agents are reversible and are without obvious cytotoxic effects or effects on lactic dehydrogenase release, exclusion of trypan blue, or viability assays of mitochondrial function.

The mechanisms by which thiazolidinediones affect the cell cycle or programmed cell death may vary among different cell types. We did not find an effect of troglitazone on fibroblasts. This suggests that troglitazone's therapeutic benefit in psoriasis is not the result of interference with any putative dermal effect on the epidermis. We do not know whether the antiproliferative effects of PPARγ ligands in keratinocytes are mediated through the same intracellular pathways involved in the growth effects of these agents in breast, prostate, colon, vascular smooth muscle, and fat cells.

The pathophysiological features of psoriasis are complex, and the antipsoriatic effects of thiazolidinediones may involve more than just their effects on cell cycle control and proliferation of keratinocytes. Disordered cellular immunity involving inflammatory cytokines (eg, tumor necrosis factor α, interleukin [IL] 1, and IL-6) and proinflammatory transcription factors (eg, nuclear factor-κB, signal transducer and activator of transcription [STAT], and activator protein-1) have been implicated in the pathogenesis of psoriatic epidermal inflammation.^{3 ,33 ,37 - 38} Macrophages may interact with basal keratinocytes in the pathogenesis of psoriasis, and keratinocytes and monocytes may elaborate inflammatory cytokines, including IL-1, IL-2, IL-6, and tumor necrosis factor α.^{39 - 41}

Studies^{42 - 43} demonstrating anti-inflammatory effects of PPARγ ligands raise the possibility that the antipsoriatic effects of thiazolidinediones might also be mediated through inhibition of skin inflammation. Peroxisome proliferator-activated receptor-γ is markedly up-regulated in activated macrophages, and ligands for PPARγ can inhibit the activities of nuclear factor-κB, STAT, and activator protein-1, and down-regulate inflammatory cytokines such as tumor necrosis factor α.^{42 - 43}

In our SCID mouse model, the inflammatory infiltrate in transplanted psoriatic skin was reduced during oral troglitazone therapy. Nevertheless, this may not have been a primary mechanism of the salutary effect of troglitazone on the skin transplant; our studies were not designed to determine whether ligands for PPARγ affect the immunologic determinants of psoriasis. Future studies of the anti-inflammatory effects of thiazolidinediones are warranted.

For ligands of PPARγ to affect the transcription of specific proteins and generate a biologic effect, PPARγ must associate with (ie, form a heterodimer with) another nuclear hormone receptor, the retinoid X receptor (RXR),^{44 - 45} which is also expressed in human keratinocytes.⁴⁶ Peroxisome proliferator-activated receptor-γ–RXR heterodimers, together with one or both of their ligands (such as a thiazolidinedione for PPARγ or 9-cis retinoic acid for RXR), activate peroxisome proliferator response elements in DNA, leading to a change in protein transcription.

The cooperative interaction of PPARγ and RXR at the molecular level suggests that ligands for each might share several therapeutic effects. Indeed, retinoids have efficacy in dermatologic disorders, certain forms of cancer, and type 2 diabetes mellitus.^{47 - 48} In a similar fashion, thiazolidinediones, originally developed for the treatment of type 2 diabetes mellitus, are now being investigated for treatment of various skin disorders and selected malignancies. Our findings of antipsoriatic effects of PPARγ ligands in cell and animal models and patients suggest that ligands for PPARγ may be of clinical value in the treatment of psoriasis.

Since the manuscript was accepted for publication, others^{49 - 50} have reported that ligands of PPARγ inhibit T-cell activation and the production of IL-2. These findings support an immunomodulatory mechanism of action for ligands of PPARγ in the treatment of psoriasis.

In March 2000, the manufacturer of troglitazone voluntarily discontinued its sale in the United States because of reports of liver toxicity.

Accepted for publication September 30, 1999.

This work was supported by grant R41 AR44767 from the National Institutes of Health, Bethesda, Md, and by Bethesda Pharmaceuticals.

We thank Navin M. Amin, MD, for assistance and Ted A. Hamilton, MS(Biostat), for statistical support.

Reprints: Charles N. Ellis, MD, Department of Dermatology, University of Michigan Medical School, 1910 A. Alfred Taubman Center, 1500 E Medical Center Dr, Ann Arbor, MI 48109-0314.