IL-8 as Antibody Therapeutic Target in Inflammatory Diseases: Reduction of Clinical Activity in Palmoplantar Pustulosis¹

Chronic inflammatory conditions often result from aberrant production of proinflammatory factors such as chemokines. One chemokine implicated in chronic inflammation is IL-8 (1). IL-8 is a member of the CXC chemokine family and was initially identified as a neutrophil chemotactic and activating factor (2, 3, 4, 5, 6). Since then, a variety of other proinflammatory activities have been attributed to IL-8 (7), including immune cell activation (8) and promotion of angiogenesis (9, 10). IL-8 can be produced by a variety of cell types involved in inflammation, including monocytes and endothelial cells. Activation of cells occurs after binding to the IL-8 receptors, CXCR1 and CXCR2, expressed on neutrophils, monocytes, endothelial cells, astrocytes, and microglia (8, 11, 12).

Aberrant IL-8 production can lead to chronic inflammatory conditions, as suggested for inflammatory diseases such as rheumatoid arthritis (13, 14, 15), inflammatory bowel disease (16), psoriasis (17, 18), and palmoplantar pustulosis (PPP)⁴ (19, 20). Accumulation of activated neutrophils in lesional areas and elevated IL-8 production is observed in all these diseases. In several animal models of acute inflammatory disease, neutralizing Abs against IL-8 inhibit neutrophil function and resolve inflammation (21, 22, 23, 24). Hence, abrogation of IL-8 activity represents a candidate therapeutic strategy for chronic inflammatory diseases. We aimed to establish an in vivo proof of this concept in PPP, which is a rare chronic inflammation of the skin that is clinically characterized by recurrent eruptions of sterile pustules on the palms and/or soles. PPP is more often observed in females than in males, the onset is between 20 and 60 years of age, and the disease is positively correlated with smoking and diseases of the thyroid gland. Novel therapeutic strategies for PPP are needed, because this disease is very resistant to current treatments (25, 26).

We generated a fully human anti-IL-8 mAb, characterized its IL-8 inhibitory capacity and binding epitope, and performed a single dose escalation study followed by a 4-wk multiple dose extension in patients with PPP. The marked reduction in disease activity following treatment as reported herein demonstrates a central role for IL-8 in the pathogenesis of PPP and highlights the potential of HuMab 10F8 for immunotherapy of pathologic conditions relating to IL-8 overexpression.

Human Ig transgenic mice, strain HCo7 (Medarex), were immunized by i.p. injection of recombinant human IL-8 (rhIL-8; Biosite Diagnostics) dispersed in complete or IFA (Sigma-Aldrich). Mice developing human anti-IL-8 Ab, as detected in serum by ELISA, were boosted by i.v. injection with rhIL-8. Experiments were reviewed and approved by the local Animal Ethics Committee. Spleen and lymph node cells were fused with SP2/0 cells (27), and specific hybridoma clones were selected and subjected to limiting dilution at least twice to ensure monoclonality. Three human mAbs were obtained that were specific for human IL-8 and were of the IgG1κ subclass. Of this panel, HuMab 10F8 (HuMax-Inflam; HuMax-IL8) was selected for further evaluation.

RhIL-8 was immobilized on a CM-5 sensor chip (Biacore) via amine chemistry. HuMab 10F8 whole IgG or Fab fragments (prepared using the ImmunoPure Fab preparation kit, Pierce Biotechnology) were injected, and binding was detected using surface plasmon resonance on a Biacore 3000 (Biacore).

ELISA plates were coated with endothelial cell-derived rhIL-8_77aa, monocyte-derived rhIL-8_72aa Gro-α, Gro-β, or IP-10 (Strathmann Biotech). Plates were washed with PBS-Tween (PBS supplemented with 0.05% (v/v) Tween 20) and nonspecific binding was blocked with PBS-Tween supplemented with 2% (v/v) chicken serum (Sigma-Aldrich). Plates were then incubated with HuMab 10F8, mouse-anti-human IL-8 (MAK-IL-8, Strathmann Biotech), or with control human IgG1 (Genmab). Bound primary Ab was detected by HRP-conjugated rabbit-anti-human IgG F(ab′)2 (DakoCytomation) or rabbit-anti-mouse IgG F(ab′)2 (Jackson ImmunoResearch Laboratories) and addition of ABTS substrate. Absorbance was measured at 405 nm.

Heparinized peripheral whole blood from healthy volunteers was diluted 1/1 in culture medium (IMDM supplemented with 100 IU/ml penicillin, 100 μg/ml streptomycin (all from Invitrogen), and 1% CCS (HyClone)). Diluted blood was stimulated with 100 μg/ml LPS (E. coli serotype 026:B6, Sigma-Aldrich) for 2 h at 37°C, together with GolgiStop (monensin, Cytofix/Cytoperm Kit, BD Pharmingen) to accumulate IL-8 intracellularly. Cells were lysed in BD lysing solution (BD Pharmingen). Cells were washed by centrifugation and resuspended in Cytofix/Cytoperm solution (BD Pharmingen). Intracellular IL-8 was detected by staining with Alexa Fluor 488-conjugated HuMab 10F8 in Perm/Wash buffer (BD Pharmingen), and HuMab-KLH-Alexa Fluor 488 (Genmab) was used as isotype control. Cell-associated fluorescence was analyzed using a FACSCanto II and FACSDiva software (BD Biosciences).

Pepscan Chemically Linked Peptides on Scaffolds technology (28, 29) was used to determine the IL-8 epitope to which HuMab 10F8 is directed. Approximately 2000 different peptides were designed based on the amino acid sequence of IL-8_77aa (endothelial cell-derived) and used in Pepscan screening either as linear peptides or linked to scaffolds to yield 30-mer single-, double-, or triple-looped or sheetlike peptides. Binding of HuMab 10F8 (1 μg/ml) to the resulting peptides was determined using anti-human-peroxidase (1/1000) as detection agent.

Neutrophils were enriched from heparinized peripheral blood from healthy volunteers by Ficoll density gradient centrifugation (room temperature, 400 × g, 30 min) and erythrocytes were removed by hypotonic shock. Enriched neutrophils were resuspended in PBS containing 0.1% BSA and 0.02% sodium azide (PBA buffer). The IL-8 inhibition of binding assay was performed as previously described (30) with 4 × 10⁵ neutrophils per well using [¹²⁵I]rhIL-8 (Amersham Life Sciences) and HuMab 10F8, mouse-anti-human IL-8 mAb (clone 6712.111, R&D Systems), or human IgG1 (Medarex). IC₅₀ values were determined using a nonlinear regression (allowing variable slopes) of the separate experiments (mean IC₅₀ ± SEM).

Heparinized peripheral blood from healthy volunteers was diluted 1/5 (v/v) in culture medium and incubated with serial dilutions of HuMab 10F8 in the absence or presence of 25 ng/ml rhIL-8 (PeproTech) (37°C, 2 h). Cells were then incubated with PE-conjugated mouse-anti-human CD11b or FITC-conjugated mouse-anti-human CD62L (BD Pharmingen) (4°C, 30 min). Erythrocytes were lysed with FACS lysing solution (BD Pharmingen), and cell-associated fluorescence was analyzed by flow cytometry.

Neutrophil chemotaxis was studied using the Boyden chamber system. Neutrophils enriched from heparinized peripheral blood from normal volunteers (4 × 10⁵ cells) were transferred to the upper compartment of a Boyden chamber. RhIL-8 (PeproTech) and HuMab 10F8 were added to the lower compartment, chambers were incubated for 1 h at 37°C, and cell migration was quantified by microscopic inspection of the filter (pore width 8 μm, thickness 150 μm; Sartorius). The synthetic neutrophil chemoattractant fMLP (Sigma-Aldrich) was used as a positive control. At 15 intervals of 10 μm, the total cell number and the total distance migrated were scored in the Z-direction of the filter. Cells in layer 1 were excluded for calculation of the total number of migrated cells, the total migrated distance, and the mean migration per cell. Alternatively, neutrophil chemotaxis was studied using a transwell sytem. RhIL-8 and HuMab 10F8, mouse-anti-human IL-8 mAb (clone 6712.111, R&D Systems), and control human IgG1 (Medarex) were added to one compartment on a transwell plate. Neutrophils were transferred to the other compartment. Plates were incubated for 2 h at 37°C, and cell migration was quantified by microscopic inspection (data not shown).

Washing fluid of six PPP patients was obtained as described previously (17) from the most severely afflicted hand or foot based on pustule counts (i.e., the number of fresh pustules on the plantar and palmar side of hand/foot). As a control, washing fluid was obtained from six patients with hand eczema and from six healthy controls. Briefly, the hand or foot was inserted into a plastic bag with 2 ml of saline for 30 min, washing fluid was aspirated, the volume was determined, and the fluid was stored at −20°C until analysis. Washing fluid was examined for the presence of eotaxin, G-CSF, GM-CSF, IFN-γ, IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-8, IL-10, IL-12p40, IL-12p40/p70, IL-12p70, IL-13, IL-15, IP-10, MCP-1, MCP-3, monokine induced by IFN-γ (MIG), sCD23, sCD95, sICAM-1, TGF-β, TNF-α, TNF-β, and TRAIL using the human cytokine Protein Profiling Biochip technology (Zyomyx). IL-8, Gro-α, C5a, and ENA-78 concentrations were also determined by ELISA using commercially available ELISA kits (IL-8 PeliKine Compact (Sanquin); Quantikine human IL-8 kit, Quantikine human Gro-α kit, and Quantikine ENA-78 kit (R&D Systems); and C5a kit (OptEIA via BD Pharmingen)). Normal human serum was used as a negative control.

An open-label multicenter study was performed with a single dose dose-escalation setup followed by a 4-wk multiple dose extension. The clinical study was performed at three university hospitals in Denmark (Copenhagen University Hospital, Gentofte; County Hospital, Roskilde; Aarhus University Hospital, Aarhus), one in Finland (Hospital for Skin and Allergic Diseases, Helsinki), and one private dermatology practice in Denmark (Aalborg Hudklinik, Aalborg). The study was conducted according to the Declaration of Helsinki, and the study protocol was identical for all study sites and was reviewed and approved by the human research ethics committees for each of the five study sites. All patients gave written informed consent before inclusion in the study. The study was performed as a single dose dose-escalation including single dose levels of 0.15, 0.5, 1, 2, 4, and 8 mg/kg. The next dose level was initiated after safety evaluation of the previous dose level, including dose-limiting toxicity for 2 wk. Four weeks after the single dose, patients entered the repeated-dose extension comprising four administrations once per week, including dose levels of 0.15, 0.5, 1, 2, and 4 mg/kg. Patients receiving 8 mg/kg as a single dose received 4 mg/kg as a repeating dose. All other patients continued repeat dosing of the initial single dose level. Patients were followed for an additional 4 wk. HuMab 10F8 was administered by i.v. infusion during 60 min in one arm.

Of the 32 PPP patients planned for inclusion, 31 were enrolled from April 2003 until May 2004. All patients, recruited in Denmark and Finland, 18 years or older, were clinically diagnosed with palmoplantar pustulosis of at least 6 mo duration by a board-certified dermatologist (patients with psoriasis were excluded) and were suffering from active disease (defined as having at least a total of 20 fresh pustules on the soles or palms at screening). None of the patients had received systemic therapies for PPP or UV therapies within 4 wk before the start of the study or any topical therapies or devices (e.g., hydrocolloid occlusion) within 2 wk before the start of the study. Use of nonsteroid antiinflammatory drugs (NSAID) was allowed, but patients receiving systemic immunosuppressive or antiinflammatory therapies (other than NSAID) or having received prior treatment with anti-IL-8 mAb were excluded. Patients were excluded from the study for any of the following reasons: immunodeficiency, past or current malignancy, acute or chronic infection, and clinically significant cardiac or cerebrovascular disease. Pregnant or breast-feeding women were not enrolled in the study, and women with childbearing potential were to use a contraceptive.

The efficacy endpoints were: 1) number of fresh pustules (white to yellow pustules of at least 1 mm); 2) clinical response (≥50% reduction in the number of fresh pustules compared with baseline (i.e., week 0) for the single dose period, or to week 4 for the multiple dose period); 3) PPP composite severity index taking into account extent of involvement (25, 26) (i.e., area of involved skin, rated from 0 to 100% of the entire skin surface of the region), severity of erythema, infiltration, and scaling (rated from 0 to 3); 4) Physician’s Global Assessment (visual analog scale); and 5) Patient’s Symptom Assessment (visual analog scale).

To assess the efficacy of HuMab 10F8 in situ, the IL-8 concentration in hand or foot washing fluid was determined by ELISA. Washing fluid was obtained at baseline and weeks 1, 4, and 8 from the most severely afflicted hand or foot as described above; in case of hand washing fluid, the hand at the noninfused arm was used. On the treatment days (baseline and week 4), washing fluid was obtained during infusion of HuMab 10F8. Patients were also assessed for dose-limiting toxicity and adverse events in general. At all visits, patients were asked for occurrence of adverse events, and vital signs and ECG were evaluated. Blood was sampled for standard hematology and clinical chemistry at all visits except two. Samples to assess potential development of human anti-human antibodies (HAHA) were collected at baseline and weeks 4 and 10 and analyzed by ELISA. IL-8, Gro-α, C5a, and ENA-78 concentrations in washing fluid were determined by ELISA as described above. Also, the concentration of free HuMab 10F8 in the same washing fluid samples was quantified by ELISA. Briefly, rhIL-8 (0.5 μg/ml in PBS; PeproTech) was coated onto ELISA plates. After washing with PBS-Tween and blocking of nonspecific binding with PBS supplemented with 2% (v/v) chicken serum (Invitrogen), plates were incubated with serial dilutions of PPP samples. HuMab 10F8 was detected using goat-anti-human IgG-HRP (Jackson ImmunoResearch Laboratories) and ABTS substrate. Absorbance was measured at 405 nm. The concentration of free HuMab 10F8 was calculated using a HuMab 10F8 standard curve. All samples for a specific laboratory parameter were assessed by a central laboratory. Note that one patient in the 4 mg/kg group displayed a PK profile indicating erroneous infusion administration or measuring.

The differences in chemokine levels between PPP and eczema samples were analyzed using a t test or a Wilcoxon test for comparisons with healthy control samples. In the analysis of clinical trial data, change from baseline in pustule count was analyzed using a paired t test of log pustule count. A p value <0.05 was considered to be significant. For all data, descriptive statistics were applied.

A fully human mAb against human IL-8 was generated using human Ig transgenic mice (31). Mice were immunized with IL-8, and hybridomas were obtained using somatic cell fusion. HuMab 10F8 was selected for further studies because of its excellent IL-8-binding characteristics and efficient inhibition of IL-8 binding to neutrophils and neutrophil chemotaxis. The binding constant of HuMab 10F8 was determined by surface plasmon resonance in Biacore. The affinity of HuMab 10F8 Fab fragments for rhIL-8 was 372 pM at 37°C (k_a of 1.06 × 10⁵ M⁻¹ sec⁻¹, k_d of 3.94 × 10⁻⁵ sec⁻¹). When testing HuMab 10F8 whole IgG1, the avidity was 20 pM at 37°C (K_a of 2.75 × 10⁵ M/sec; K_d of 0.54 × 10⁻⁵/sec). The binding of HuMab 10F8 to rhIL-8_72aa (72 aa form of IL-8 as derived from monocytes) and rhIL-8_77aa (77 aa form of IL-8 as derived from endothelial cells) and the homologous chemokines rhGro-α, rhGro-β, and rhIP-10 was assessed by ELISA. HuMab 10F8 bound to plate-bound, monocyte-derived, and endothelial cell-derived IL-8 (Fig. 1,A) and did not cross-react with Gro-α, Gro-β, or IP-10 (data not shown). Binding of HuMab 10F8 to native human IL-8 was confirmed by staining of intracellular IL-8 in LPS-activated human peripheral blood cells and analysis by flow cytometry (Fig. 1 B) and immunocytochemistry (data not shown).

The HuMab 10F8 epitope on IL-8 was determined via Pepscan Chemically Linked Peptides on Scaffolds technology (28, 29). Approximately 2000 different peptides were designed based on the IL-8 amino acid sequence (77 aa form) and used in Pepscan screening either as linear peptides (15-mers) or linked to scaffolds to yield 30-mer single-, double-, or triple-looped or sheetlike peptides. Three different binding regions were identified: YSKPF (IL-8_18–22), KFIK (IL-8_25–28), and DGRELCL (IL-8_50–56). The region YSKPF with KP as its core region was recognized in both linear and looped topology, and it was found to represent the dominant binding region. Combining YSKPF with DGRELCL in a double-looped topology gave the best binding activity, indicating that DGRELCL contributed to binding and is part of the HuMab 10F8 epitope. The region KFIKELRV was the third region found to have binding activity, independent of the other two binding regions. Taken together, these data indicated that HuMab 10F8 recognizes a discontinuous epitope on human IL-8 comprising all of the three identified regions (Fig. 2, A and bottom panel). The identified HuMab 10F8 epitope overlaps with the binding site of CXCR1 (Fig. 2 B).

The ability of HuMab 10F8 to inhibit IL-8 activity was first assessed in vitro. To analyze inhibition of IL-8 receptor binding, neutrophils were enriched from human peripheral blood and incubated with radiolabeled rhIL-8 in the presence of HuMab 10F8. HuMab 10F8 dose-dependently inhibited binding of IL-8 to human neutrophils (Fig. 3 A) with an IC₅₀ at 0.3 nM (compared with 1.6 nM for a positive control mAb).

To study whether HuMab 10F8 inhibited IL-8-induced neutrophil activation, we incubated whole blood cells with HuMab 10F8 and determined CD11b up-regulation and CD62L down-regulation in the presence/absence of rhIL-8. HuMab 10F8 potently and dose-dependently inhibited both the induction of CD11b expression by IL-8 on neutrophils (Fig. 3,C), as well as the shedding of CD62L (Fig. 3 B) from these cells.

We studied the effect of HuMab 10F8 on neutrophil chemotaxis, because the recruitment of neutrophils to sites of inflammation represents an important IL-8 function. Neutrophil migration was assessed using a Boyden chamber system, with neutrophils in the upper compartment and rhIL-8 alone or with HuMab 10F8 in the lower compartment. Neutrophils readily migrated toward IL-8, which was inhibited in a dose-dependent manner when HuMab 10F8 was present in the lower compartment (Fig. 4). The values for mean migration per cell were 132.7 ± 7.0 μm in the presence of 10⁻⁸ M IL-8 alone, which was decreased to 132.3 ± 1.2, 128.7 ± 4.0, 71.3 ± 23.5, 56.0 ± 43.7, and 57.0 ± 41.6 μm in the presence of 0.03, 0.15, 0.8, 4, or 20 μg/ml HuMab 10F8. These results were confirmed in a trans-well system (data not shown).

PPP represents a chronic inflammatory skin disease characterized by intraepidermal accumulation of neutrophils. The presence of IL-8 in PPP lesional skin biopsies (19, 20) suggested that IL-8 plays a role in PPP pathophysiology. To address the role of IL-8 in relationship to other proinflammatory factors, we measured the levels of IL-8 and other cytokines/chemokines in exudates of PPP lesions. Exudates from PPP patients were obtained as washing fluid by incubating the most severely afflicted hand or foot into a plastic bag with 2 ml of saline for 30 min. As a control, washing fluids from patients with hand eczema and healthy controls were taken in a similar manner. Samples were analyzed for the presence of cytokines and chemokines by protein array and ELISA.

Of all inflammatory mediators tested, only a limited number of cytokines/chemokines were found to be elevated in patients’ washing fluids. The relative expression of IL-12p40 was specifically elevated in PPP patients (Table I), while sCD23 and sICAM-1 were specifically elevated in eczema patients (data not shown). IL-6, IL-8, MCP-1, and MIG levels were elevated both in PPP and eczema patients (mean log concentration_PPP/eczema ≥ mean log concentration_controls + 2×SD log concentration_controls; Table I). Absolute IL-8 concentrations were higher in PPP samples compared with samples from eczema patients and healthy controls: IL-8 protein levels were elevated in all PPP patients tested, with a 1.5-log difference in mean concentration from healthy controls (compared with a 0.9-log difference in eczema patients from healthy controls).

The increase of IL-8 in PPP washing fluids was confirmed by ELISA. Both PPP patients and eczema patients showed elevated IL-8 levels compared with healthy controls (Fig. 5,A). Comparison of washing fluids of PPP and eczema patients demonstrated a much higher IL-8 level in PPP washing fluids (10,000-fold elevation in PPP patients over control, 1,200-fold elevation in eczema patients over control; p = 0.002 for PPP from controls, Wilcoxon test, exact two-sided; p = 0.028 for PPP from eczema patients, t test on logarithmic IL-8 values). The levels of Gro-α (Fig. 5 B), complement factor C5a, and ENA-78 (data not shown) were also assessed by ELISA. Of these, only Gro-α was found to be significantly elevated in PPP washing fluids compared with eczema patients (p = 0.175, t test on logarithmic IL-8 values) and to healthy controls (p = 0.002, Wilcoxon test, exact two-sided).

Thirty-one PPP patients with a median number of fresh pustules of 35 at baseline were included in the study. Patients were allocated to six dose groups: four to 0.15 mg/kg, four to 0.5 mg/kg, four to 1 mg/kg, four to 2 mg/kg, eight to 4 mg/kg, and seven to 8 mg/kg as a single dose. Four weeks after the single dose, patients entered the repeated-dose extension, comprising four administrations once per week of a dose equal to the initial dosing, except the 8 mg/kg initial dosing, which was continued as a 4 mg/kg repeating dose. All patients (25 women and 6 men) were Caucasian with a median age of 57, ranging from 41 to 71 years of age. The median time since diagnosis of PPP was ∼8 years, ranging from 6 mo to 43 years. Twenty-six patients were current smokers and the remaining 5 patients were former smokers. Three patients discontinued the study. One refused further participation 14 days after administration of a single dose of 0.15 mg/kg, and one discontinued due to a serious, non-drug-related adverse event 29 days after receiving a single dose of 4 mg/kg. One patient discontinued the study during the multiple dose phase, after two doses of 0.15 mg/kg due to a non-drug-related serious adverse event. Patient baseline characteristics are summarized in Table II.

The numbers of fresh pustules were strongly reduced after both single and multiple dose administrations (Fig. 6). Across all dose groups, there was a reduction of 52.9% from baseline to week 1 (p = 0.003), and a reduction of 55.9% from baseline to week 8 (p = 0.0002; paired t test of log pustule count). The change in numbers of fresh pustules following a single dose (baseline to week 1) and for all data (baseline to week 8) are shown in Fig. 7, A and B. A clinical response (measured as a decrease of 50% or more in numbers of fresh pustules) was observed after both single and multiple doses. The proportion of patients with a clinical response was between 37% at 4 wk and 61% at 8 wk. In the high-dose group, 7 out of 7 patients showed at least a 50% reduction in numbers of fresh pustules, with 4 patients having a reduction of 75% from baseline to week 8. Following multiple dose administrations, the PPP composite severity index was reduced in 82% (23 out of 28) of patients (Fig. 7,D). A similar, yet less pronounced tendency was observed after single dose administration in 58% (18 out of 31) of patients (Fig. 7 C). According to the Physician’s Global Assessment, the disease severity was reduced in 71% of the patients after multiple dose administration and in 48% after a single administration (data not shown). According to the Patient’s Symptom Assessment, the disease severity was reduced in 82% of the patients after multiple dose administration and in 81% after a single administration (data not shown).

HuMab 10F8 therapy was well tolerated. Twenty-five of the 31 patients (81%) had in total 85 adverse events, which were mostly mild or moderate (Table III). There were two serious adverse events, both considered not related to HuMab 10F8 treatment (acute myocardial infarction and syncope). The adverse events were distributed equally between cohorts; that is, ∼80% of patients in each cohort experienced one or more adverse events. Approximately 50% of reported adverse events (45 of 85) were judged attributable to the study drug. The most frequent adverse events were nausea, nasopharyngitis, and headache. There was no increase in frequency of adverse events with increasing dose.

The standard safety panel of serum chemistry and hematology indicated no impact of HuMab 10F8. Samples for HAHA analyses were collected at weeks 0, 4, and 11. No HAHA development was observed as measured by ELISA, and pharmacokinetics were as expected (see next section).

HuMab 10F8 pharmacokinetics were studied by analyzing plasma concentrations by ELISA. The i.v. infusions of HuMab 10F8 yielded little variation in plasma concentrations between patients within each dose group across time (Fig. 8). The maximum plasma concentrations of HuMab 10F8 linearly increased with increasing dose levels, and plasma clearance rates were dose-independent. Both after single dose administration and after repeated dosing, HuMab 10F8 reached maximum plasma concentrations after ∼1 h. Four infusions at intervals of 7 days led to limited accumulation and a maximum plasma concentration that was ∼60% higher than following single dose administration.

Washing fluids were obtained during HuMab 10F8 infusion at baseline and week 4, and at follow-up visits at weeks 1 and 8. Samples were analyzed for the presence of HuMab 10F8, IL-8, Gro-α, ENA-78, and C5a by ELISA. HuMab 10F8 was readily detected in washing fluid samples, and the average concentrations were dependent on the doses given to patients (Fig. 9,A). In line with this, the average IL-8 concentrations in washing fluids decreased with increasing HuMab 10F8 doses (Fig. 9,B), indicating that HuMab 10F8 is capable of neutralizing IL-8 in situ. For Gro-α, C5a, and ENA-78, no differences were observed between high and low HuMab 10F8 dose groups (data not shown). Moreover, when combining means from Fig. 9,A with Fig. 9,B, a monotonic relation with dose level between mean IL-8 concentrations and mean HuMab 10F8 concentrations in washing fluids was observed (Fig. 10).

Overproduction of IL-8 has been proposed to significantly contribute to a number of inflammatory diseases, characterized by accumulation of activated neutrophils in lesional areas. In the present study, we demonstrate that the human IL-8 Ab HuMab 10F8, which efficiently neutralizes biological activity of this chemokine, reduces the clinical disease activity in patients with PPP. This supports a central role for IL-8 in the pathophysiology of PPP. Our study indicates the feasibility of an immunotherapeutic approach and provides in vivo proof of concept for the application of an anti-human IL-8 Ab for treatment of inflammatory diseases.

HuMab 10F8 was shown to bind rhIL-8 with picomolar affinity and bound both predominant forms of IL-8 expressed in humans: the 77 aa endothelial cell-derived form and the 72 aa monocyte form (32). The effective inhibition of IL-8 binding to human neutrophils, which express CXCR1 and CXCR2 at similar levels, indicated HuMab 10F8 to inhibit IL-8 binding to both receptors. The HuMab 10F8 epitope was found to be discontinuous, and it is formed by three regions within the N-terminal loop and the first and third β-sheet of IL-8. The epitope overlaps with the CXCR1 (and presumably also the CXCR2) docking site for IL-8 (33). The epitope mapping results suggest that, consistent with the efficient inhibition of binding of IL-8 to cells by HuMab 10F8, IL-8 cannot bind HuMab 10F8 and receptor simultaneously. This should be favorable for clinical application of HuMab 10F8, as the Ab therefore cannot induce Fc-mediated immune effector functions (such as complement- and cell-mediated cytotoxicity), which are often suggested to play a role in mediating side-effects. In conclusion, HuMab 10F8 represents an effective blocker of IL-8 activity in which efficient IL-8 neutralization through blockade of ligand-receptor interaction is identified as an important mechanism of action.

Previously, the efficacy of murine IL-8 Ab in acute inflammatory diseases has been demonstrated in a number of animal models, suggesting Ab-mediated neutralization of IL-8 can potentially be used for various human inflammatory disorders (21, 23, 24, 34, 35, 36, 37, 38). Few studies have been performed in chronic inflammatory diseases. The anti-IL-8 activity of another human IL-8 mAb, ABX-IL-8 (IgG2/κ), was demonstrated in vitro and in animal models in vivo (30). The result of a placebo-controlled phase IIb clinical trial for treatment of moderate-to-severe psoriasis with this Ab, however, was disappointing (39). This failure might well have resulted from the low and infrequent dosing used in these studies, leading to insufficient Ab concentrations in situ. Additionally, heterogeneity in clearance rates of the human IgG2 Ab resulting from a polymorphism for FcγRIIa that is known to affect IgG2 serum concentrations (40) may have played a role.

Most chronic inflammatory diseases such as psoriasis have multifactorial causes (reviewed in Ref. (41)), which potentially complicate treatment by targeting a single disease-related factor. For psoriasis in particular, the presence of other CXC chemokines with redundant activities, such as MIG (42) and Gro-α (43), may further compromise the success of solely targeting IL-8. However, single Ab treatments can be highly effective when targeting a critical and essential pathway in chronic inflammation, as shown by the success of anti-TNF-α treatments in inflammatory diseases, including refractory psoriaris (44).

We focused on an inflammatory disorder with a strong neutrophil involvement implying an essential role for IL-8 in the pathophysiology for establishment of an in vivo proof of concept for the therapeutic potential of IL-8 targeting. In PPP, accumulation of activated neutrophils and elevated levels of IL-8 in lesional areas are prominent (19, 20, 45). We herein confirm increased production of IL-8, as well as the closely related CXC chemokine Gro-α, to be typical for PPP compared with a number of other cytokines and chemokines (such as MIG and IL-6), which are also elevated in eczema, and TNF-α, being only moderately elevated in PPP. Very high IL-8 levels were detected in washing fluids of PPP lesional areas, compared with those from patients with eczema. Indeed, eczema is described as an inflammatory disease lacking strong neutrophil involvement (46).

Targeting IL-8 in PPP, through treatment with HuMab 10F8, demonstrated strongly reduced disease activity in five different clinical efficacy endpoints: numbers of fresh pustules, clinical response (≥50% reduction in numbers of fresh pustules), PPP composite severity index, Physician’s Global Assessment, and Patient’s Symptom Assessment. The reduction in disease activity was mainly observed following multiple dose administration, and to a lesser extent also upon single dose administration. The effects were most pronounced in groups treated with high doses of Ab, yet were also apparent at lower doses. These effects, however, were not unexpected, as the mean C_max of the lowest dose achieved was 10 times higher than the IC₅₀ for chemotaxis in vitro (IC₅₀ of 0.3 μg/ml).

PPP treatment with HuMab 10F8 resulted in dose-dependent decreases in IL-8 concentrations in washing fluids. Notably, the Ab was readily detected in situ in samples obtained during mAb infusion. A priori, we had not anticipated such a rapid passage of mAb from the circulation through the skin into PPP lesions, which represent the actual site of action for IL-8. The in situ neutralization of IL-8 was confirmed by the clinical efficacy of HuMab 10F8 in PPP patients, which was demonstrated for all efficacy endpoints in this clinical trial. Altogether, these data confirmed that IL-8 plays an essential role in PPP and identified HuMab 10F8 as a candidate for treatment of PPP. The presence of the therapeutic Ab in washing fluids implies that the barrier for larger molecules normally existing in the epidermis is compromised in lesional inflamed skin, and indicates mAb-mediated treatments may also be feasible for other skin inflammatory disorders.

In conclusion, we show IL-8 to play a central role in PPP. Using an IL-8-neutralizing Ab, we demonstrate that targeting a single critical factor in this disease characterized by high IL-8 overexpression leads to clinically relevant reductions in disease activity. This observation bears promise for the treatment of other diseases characterized by IL-8 overexpression.

We thank Marleen Voorhorst, Wendy Jansen, and Mischa Houtkamp for technical assistance, Leo Koenderman for guidance and provision of laboratory facilities to conduct neutrophil chemotaxis studies, Jerry Slootstra for peptide-guided epitope mapping, Mohan Srinivasan for affinity measurements, Joost Bakker for graphics, and Sigrid Ruuls for critically reviewing the manuscript. We are grateful to the members of the HuMab 10F8 study group and to Drs. Gregor Jemec (Roskilde, Denmark), Niels Veien (Aalborg, Denmark), and Lars Iversen (Aarhus, Denmark) for contributions to this work.

Frank J. Beurskens, Jessica Teeling, David Satijn, Elmieke P. J. Boot, Paul W. H. I. Parren, and Jan G. J. van de Winkel are employees of Genmab BV. Kim M. Knudsen and Ole Baadsgaard are employees of Genmab A/S. D.H. is an employee of Medarex.