**3.4 Clinical experience**

To date, PXS TPI1100 has not been dosed in human subjects, however, a review of the current literature on clinical study designs and using the Daxas/Daliresp background as guidance, the projected clinical path for PXS TPI1100 has been defined. Furthermore, reviewing its pharmacology profile, there are potential advantages for PXS TPI1100 which may be manifested in the clinic. In this section, we will first outline some general challenges facing COPD clinical study design, then capture some of the salient points from the Daxas/Daliresp experience in clinical study design.

#### **3.4.1 Challenges in COPD clinical studies**

Typically COPD clinical studies measure as a primary outcome lung function by spirometry, either through improvement in postbronchodilator FEV1 or in cases where assessing the efficacy of nonbronchodilators is preferred, the measure is of the change from baseline in prebronchodilator FEV1 (Giembycz &Field 2010). As a procedure for detection of airflow obstructions spirometry is a reliable, simple, non-invasive, safe, and non-expensive (Soriano et al. 2009). The test is relatively standardized with most COPD guidelines accepting the threshold to define a positive bronchodilation test as suggested by the Global Initiative for asthma (increase in FEV1 larger than 12% and 200 mL from the prebronchodilator value) (Bateman et al. 2008) with the variation of suggesting minimum limits of 300 or 400 mL.

Besides the use of these spirometry measures, there have been attempts to determine a relevant easily accessible and rapid assessed biomarker as a measure of improvement. In preclinical pharmacology studies, animal models of neutrophil inflammation are routinely used for efficacy measures in an attempt to mimic the disease state in humans. It is known that in COPD patients the percentage of sputum neutrophils are increased with each GOLD stage, are also raised in COPD exacerbations (Caramori et al. 2003) (Papi et al. 2006), and that neutrophils are involved in the pathogenesis of emphysema through the secretion of proteases and elastases (Cowburn et al. 2008) (Sharafkhaneh et al. 2008). Taken together, these observations would suggest that sputum neutrophils have the potential to be a biomarker predictor of the degree of airflow obstruction, however the reality is far from clear. Reports with small cohorts of patients suggest a relationship between sputum neutrophils measures and FEV1 (% predicted) (O'Donnell et al. 2004), however a larger cohort study by Singh et al. (Singh et al. 2010) demonstrated that this relationship is only weakly associated. A similar finding was shown with regard to sputum neutrophils

injection (approximately 5 min) and these levels were greatly reduced by 4 h post-dose and

The pharmacokinetics properties following pulmonary delivery has been well characterized (Templin et al. 2000; Ali et al. 2001; Guimond et al. 2008) and confers a significant advantage of AON over small molecule drugs. For example, orally-delivered Daxas/Daliresp has a bioavailabilty of 79%(David 2004) and with an elimination half-life of 14-18 h there is a greater opportunity for this drug to act upon PDE4 outside of the lung and for a long period of time. In comparison, PXS TPI1100 has reduced systemic bioavailability and based on results in mouse lung, the half-life of PXS TPI 1100 has been shown to be relatively short (<5h) suggesting a potentially safer drug that would work locally at the site of action in the lung.

To date, PXS TPI1100 has not been dosed in human subjects, however, a review of the current literature on clinical study designs and using the Daxas/Daliresp background as guidance, the projected clinical path for PXS TPI1100 has been defined. Furthermore, reviewing its pharmacology profile, there are potential advantages for PXS TPI1100 which may be manifested in the clinic. In this section, we will first outline some general challenges facing COPD clinical study design, then capture some of the salient points from the

Typically COPD clinical studies measure as a primary outcome lung function by spirometry, either through improvement in postbronchodilator FEV1 or in cases where assessing the efficacy of nonbronchodilators is preferred, the measure is of the change from baseline in prebronchodilator FEV1 (Giembycz &Field 2010). As a procedure for detection of airflow obstructions spirometry is a reliable, simple, non-invasive, safe, and non-expensive (Soriano et al. 2009). The test is relatively standardized with most COPD guidelines accepting the threshold to define a positive bronchodilation test as suggested by the Global Initiative for asthma (increase in FEV1 larger than 12% and 200 mL from the prebronchodilator value) (Bateman et al. 2008) with the variation of suggesting minimum limits of 300 or 400 mL.

Besides the use of these spirometry measures, there have been attempts to determine a relevant easily accessible and rapid assessed biomarker as a measure of improvement. In preclinical pharmacology studies, animal models of neutrophil inflammation are routinely used for efficacy measures in an attempt to mimic the disease state in humans. It is known that in COPD patients the percentage of sputum neutrophils are increased with each GOLD stage, are also raised in COPD exacerbations (Caramori et al. 2003) (Papi et al. 2006), and that neutrophils are involved in the pathogenesis of emphysema through the secretion of proteases and elastases (Cowburn et al. 2008) (Sharafkhaneh et al. 2008). Taken together, these observations would suggest that sputum neutrophils have the potential to be a biomarker predictor of the degree of airflow obstruction, however the reality is far from clear. Reports with small cohorts of patients suggest a relationship between sputum neutrophils measures and FEV1 (% predicted) (O'Donnell et al. 2004), however a larger cohort study by Singh et al. (Singh et al. 2010) demonstrated that this relationship is only weakly associated. A similar finding was shown with regard to sputum neutrophils

near LLOQ by 24 h demonstrating the clearance of AON from the system.

**3.4 Clinical experience** 

Daxas/Daliresp experience in clinical study design.

**3.4.1 Challenges in COPD clinical studies** 

measures and the relationship to health status as defined by the use of the St. Georges Respiratory Questionnaire (SGRQ) (Singh et al. 2010). Furthermore sputum neutrophil measures in the stable state were shown not be predictive of the future rate of exacerbations (Singh et al. 2010). Lastly, no association between sputum neutrophils measures and emphysema or systemic inflammation as measured by serum levels of IL-6, IL-8, C-reactive protein (CRP) and surfactant protein D was observed (Singh et al. 2010). In short, although there is a plausible assumption for the use of sputum neutrophils as a biomarker, there is little validity in using them in face of the current evidence.

In lieu of the identification and validation of a biomarker that could predict the rate of lung function decline in COPD, most COPD clinical trials attempt to measure relevant changes in exacerbations. Exacerbation frequency has been considered to be an important outcome parameter in COPD as it is associated with increase in mortality (Patil et al. 2003) (Fuso et al. 1995). Measuring exacerbations is not without its challenges. It is difficult among studies to find consensus on what is defined as an exacerbation and to gauge the severity of the exacerbation. Symptom-based definitions include use of diaries, while event-based definitions may refer to hospitalizations or use of antibiotics and/or steroids (Miravitlles et al. 2004). Although a systematic literature review of studies reporting exacerbation frequency in COPD patients showed the relationship between increased exacerbation frequency with decreasing lung function to be borderline significance (p=0.053) (Spencer et al. 2004), exacerbations are still considered to be an important parameter in COPD. Exacerbations are more likely to occur in winter and according to current recommendations (Cazzola et al. 2008) studies need to have at least a 12 month follow up to give reliable estimate of exacerbation frequency, which requires the planning of lengthy clinical trials.

The clinical program of PXS TPI1100 has not been initiated yet we expect its design can follow that of other PDE4 inhibitors. An initial Phase 2 study design does not test in COPD patients but rather in allergic asthmatic patients following inhaled allergen challenge (2009). Another AON drug, ASM8 designed specifically for asthma and as such has targets different from PXS TPI1100, has demonstrated clinical efficacy in this allergen challenge model (Gauvreau et al. 2008; Gauvreau 2010) clearly showing the potential for the AON approach. An advantage of this allergen challenge model is that the studies are generally brief in duration and the fall in FEV1 is a well-recognized response as well as the incorporation of monitoring induced sputum allows for other inflammatory indicators to be measured.

PXS TPI1100 has an advantage in that the clinical studies performed by Daxas/Daliresp can be used as a guide, as the two drugs share a common target. As Daxas/Daliresp was breaking new ground many studies had to be performed and it is plausible to conclude that for other drugs in the same class fewer studies may be required. In all, six phase 3 clinical trials were undertaken with Daxas/Daliresp which have been excellently reviewed by Giembycz and Field (Giembycz &Field 2010). Key aspects of these trials that can be used for PXS TPI1100's clinical development include criteria for patient selection and parameters selected for primary and secondary outcomes. In the phase 3 study named RECORD, patients with moderate-to-severe COPD (postbronchodilator FEV1 of 30% to 80% predicted and a FEV1/FVC ratio of less than 70%) were randomized to receive either Daxas/Daliresp at 250 µg or 500 µg or placebo (2:2:1 ratio) for 24 weeks (Rabe et al. 2005). Results showed treated patients experienced improvement in postbronchodilator FEV1 (Rabe et al. 2005) and a change in SGRQ but this change did not reach clinical significant threshold. Although direct

A Multi-Targeted Antisense Oligonucleoitde-Based

inflammatories such as ICS by specifically targeting PDE.

significantly more challenging than that of liquid aerosols.

**4. Conclusions** 

Therapy Directed at Phosphodiesterases 4 and 7 for COPD 449

metabolized in the lung with very limited systemic delivery after inhalation (Templin et al. 2000; Ali et al. 2001; Guimond et al. 2008) which leads to reduced systemic bioavailability of the drug. In comparison to Daxas/Daliresp, which is delivered orally and has a high level of bioavailability, the projected low systemic bioavailability of PXS TPI1100 may limit adverse events associated with PDE4 inhibitors, namely the gastrointestinal and neurological side effects. Another consideration is the projected brief half-life of PXS TPI1100. Based on the mouse lung, the half-life of PXS TPI1100 has been shown to be relatively short (<5h), although it is reassuring that this short tissue half-life does not appear to affect the efficacy of the drug as every-other day dosing of PXS TPI1100 in the smoking mouse model was highly effective. Reconciling the short half-life with longer term efficacy may be a reflection of the mechanism of action of the drug, suggesting that inhibition of PDE mRNA has longer term consequences on downstream effects including limitation of inflammatory responses. Besides a projected favorable safety profile resulting from low systemic bioavailability, PXS TPI1100 can be expected to avoid the toxicity associated with the broader approach of anti-

PXS TPI1100 faces challenges, in part of being the first respirable antisense drug product in COPD. As COPD is a chronic disease, it can be expected that patients will be dosed for years. The long-term effects of this drug class have never before been studied. In addition, in pulmonary/respiratory diseases, there is a risk that administration of therapeutic nucleic acids may lead to immune stimulation, inflammation and possibly hypersensitivity and bronchoconstriction of the airways. Except for the latter, these risks are not specific to the lung as they have been observed with other routes of administration. As with any novel inhaled medication, local tolerability and the absence of long-term effects following chronic dosing will require careful evaluation as drug candidate progresses through the later stages of development. The publicly available toxicological data on inhaled AON are not extensive (Guimond et al. 2008), and therefore deriving definitive conclusions on toxicology at this time is not possible. The phase 2a studies that have been performed until now have not shown any of this potential toxicity but longer term studies are needed to confirm these results. Furthermore, to date AON have been delivered via inhalation of a nebulisate to asthma patients (Gauvreau et al. 2008; Gauvreau 2010), but not to COPD patients who have severely decreased FEV1. How well this patient cohort inhales the nebulisate would need to be determined. The range of delivery devices (including newer portable soft-mist inhalers) have increased and permit liquid aerosols to be targeted more effectively to the specific airways of interest (upper or lower airways), improve ease of use by patients and would be expected to improve compliance to therapy. In contrast, the particle processing and formulation of AON for delivery in dry powder inhalers or pressurized metered dose inhalers, which are most commonly used by COPD patients has, however, proven to be

Another challenge facing PXS TPI1100 is the selection of its targets PDE4B/D and PDE7A. While the success of Daxas/Daliresp demonstrates the effectiveness of targeting PDE4 in COPD, to date there is less corrobative clinical evidence for the efficacy of targeting PDE7A isoform. The success or failure of a specific drug development program is determined by a range of different factors, which includes the clinical relevance of the selected drug target.

comparisons between doses were not made, as it seemed that patients receiving the higher dose had better and earlier responses in most outcomes the daily dose of Daxas/Daliresp of 500 µg was then used in two subsequent identical trials (RATIO and OPUS). In these studies the patients had more severe COPD than in the RECORD study (postbronchodilator FEV1 of 50% or less, FEV1:FVC ratio of 0.7 or less, or FEV1 reversibility of 5% or less). Although completed, the results from the OPUS trial have not been published, however results from the RATIO study showed an improvement for the change from baseline for post bronchodilator FEV1, yet again no effect on the SGRQ (Calverley et al. 2007). A post-hoc analysis of a subgroup of patients with GOLD stage IV disease in the RECORD study showed a significant effect on reduction of exacerbation frequency (Calverley et al. 2007) which then led to the design of two identical studies AURA and HERMES where patients had a diagnosis of clinical COPD (confirmed by postbronchodilator FEV1/FVC of at least 70%, and a FEV1 at least 50% of predicted), had symptoms of chronic bronchitis and a history of exacerbations. Patients experienced an improvement in pre- and postbronchodilator FEV1 and a reduction in exacerbation rate (Calverley et al. 2009) which were independent of LABA use, but no differences in mortality or C-reactive protein levels.

Taken together, the Daxas/Daliresp studies clearly show effects in patients with GOLD stage IV disease, with focus on measuring flow rates and exacerbation reduction as parameter outcomes. The clinical program for PXS TPI1100 can use this information in designing studies so as to sharply define the patient population at the onset and include the key primary outcomes as success measures.

As ICS and LABA have been shown to be more effective at improving lung function, health status and reducing COPD exacerbations when combined than when used individually (Calverley et al. 2007) the effect of combining Daxas/Daliresp with either the long-acting β2 agonist salmeterol (EOS study) or the long-acting inhaled antimuscarinic tiotroprium (HELIOS study) was studied in patients with less severely reduced lung function as compared to the previous studies. Results showed that the pre- and postbronchodilator FEV1 improved in patients treated with Daxas/Daliresp versus placebo when combined with either LABA or LAMA (Fabbri et al. 2009). PXS TPI1100 can be expected to also function in combination with LABA, similar to that demonstrated by Daxas/Daliresp and could potentially replace ICS.

As with any drug, adverse events to Daxas/Daliresp were reported which included weight loss, diarrhea, nausea, headache, influenza and nasopharyngitis as well as certain cancers such as lung and prostate (Giembycz &Field 2010). There was a greater risk of discontinuation of therapy within the first 12 weeks of treatment for those patients taking Daxas/Daliresp than placebo although by the end of the studies, similar numbers of patients withdrew in both groups. In the Daxas/Daliresp treated groups, the most common reason for withdrawal were the gastrointestinal adverse events or headache (Giembycz &Field 2010).

There are aspects of PXS TPI1100 which may lend itself advantages over Daxas/Daliresp. Firstly, as PXS TPI1100 is administered via inhalation, it is delivered directly to the intended site of action of the lung (Ali et al. 2001; Duan et al. 2005; Gauvreau et al. 2008; Guimond et al. 2008) where the drug can enter target cells directly (Zhang et al. 2004; Griesenbach et al. 2006) thus potentially reducing total dose as compared to orally-available treatments. A further advantage of pulmonary administration of AON is that they are principally metabolized in the lung with very limited systemic delivery after inhalation (Templin et al. 2000; Ali et al. 2001; Guimond et al. 2008) which leads to reduced systemic bioavailability of the drug. In comparison to Daxas/Daliresp, which is delivered orally and has a high level of bioavailability, the projected low systemic bioavailability of PXS TPI1100 may limit adverse events associated with PDE4 inhibitors, namely the gastrointestinal and neurological side effects. Another consideration is the projected brief half-life of PXS TPI1100. Based on the mouse lung, the half-life of PXS TPI1100 has been shown to be relatively short (<5h), although it is reassuring that this short tissue half-life does not appear to affect the efficacy of the drug as every-other day dosing of PXS TPI1100 in the smoking mouse model was highly effective. Reconciling the short half-life with longer term efficacy may be a reflection of the mechanism of action of the drug, suggesting that inhibition of PDE mRNA has longer term consequences on downstream effects including limitation of inflammatory responses.

Besides a projected favorable safety profile resulting from low systemic bioavailability, PXS TPI1100 can be expected to avoid the toxicity associated with the broader approach of antiinflammatories such as ICS by specifically targeting PDE.
